From 3e11e530415027a57936545957126aff49267b76 Mon Sep 17 00:00:00 2001 From: Benjamin Marzinski Date: Wed, 23 Mar 2016 14:29:59 -0400 Subject: GFS2: ignore unlock failures after withdraw After gfs2 has withdrawn the filesystem, it may still have many locks not in the unlocked state. If it is using lock_dlm, it will failed trying the unlocks since it has already unmounted the lock manager. Instead, it should set the SDF_SKIP_DLM_UNLOCK flag on withdraw, to signal that it can skip the lock_manager on unlocks, and failback to lock_nolock style unlocking. Signed-off-by: Benjamin Marzinski Signed-off-by: Bob Peterson --- fs/gfs2/glock.c | 9 ++++++++- fs/gfs2/util.c | 1 + 2 files changed, 9 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c index 6539131c52a2..2897ced5fca0 100644 --- a/fs/gfs2/glock.c +++ b/fs/gfs2/glock.c @@ -475,7 +475,14 @@ __acquires(&gl->gl_lockref.lock) if (sdp->sd_lockstruct.ls_ops->lm_lock) { /* lock_dlm */ ret = sdp->sd_lockstruct.ls_ops->lm_lock(gl, target, lck_flags); - if (ret) { + if (ret == -EINVAL && gl->gl_target == LM_ST_UNLOCKED && + target == LM_ST_UNLOCKED && + test_bit(SDF_SKIP_DLM_UNLOCK, &sdp->sd_flags)) { + finish_xmote(gl, target); + if (queue_delayed_work(glock_workqueue, &gl->gl_work, 0) == 0) + gfs2_glock_put(gl); + } + else if (ret) { pr_err("lm_lock ret %d\n", ret); GLOCK_BUG_ON(gl, 1); } diff --git a/fs/gfs2/util.c b/fs/gfs2/util.c index cf645835710f..aee4485ad8a9 100644 --- a/fs/gfs2/util.c +++ b/fs/gfs2/util.c @@ -68,6 +68,7 @@ int gfs2_lm_withdraw(struct gfs2_sbd *sdp, const char *fmt, ...) fs_err(sdp, "telling LM to unmount\n"); lm->lm_unmount(sdp); } + set_bit(SDF_SKIP_DLM_UNLOCK, &sdp->sd_flags); fs_err(sdp, "withdrawn\n"); dump_stack(); } -- cgit v1.3-8-gc7d7 From 798434bda36e357af9ccaf68a7ba1129658c8332 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Thu, 24 Mar 2016 20:38:43 -0400 Subject: __d_alloc(): treat NULL name as QSTR("/", 1) Signed-off-by: Al Viro --- fs/dcache.c | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/dcache.c b/fs/dcache.c index 32ceae3e6112..3cb98dfd9f96 100644 --- a/fs/dcache.c +++ b/fs/dcache.c @@ -1558,7 +1558,11 @@ struct dentry *__d_alloc(struct super_block *sb, const struct qstr *name) * be overwriting an internal NUL character */ dentry->d_iname[DNAME_INLINE_LEN-1] = 0; - if (name->len > DNAME_INLINE_LEN-1) { + if (unlikely(!name)) { + static const struct qstr anon = QSTR_INIT("/", 1); + name = &anon; + dname = dentry->d_iname; + } else if (name->len > DNAME_INLINE_LEN-1) { size_t size = offsetof(struct external_name, name[1]); struct external_name *p = kmalloc(size + name->len, GFP_KERNEL_ACCOUNT); @@ -1812,9 +1816,7 @@ struct dentry *d_make_root(struct inode *root_inode) struct dentry *res = NULL; if (root_inode) { - static const struct qstr name = QSTR_INIT("/", 1); - - res = __d_alloc(root_inode->i_sb, &name); + res = __d_alloc(root_inode->i_sb, NULL); if (res) d_instantiate(res, root_inode); else @@ -1855,7 +1857,6 @@ EXPORT_SYMBOL(d_find_any_alias); static struct dentry *__d_obtain_alias(struct inode *inode, int disconnected) { - static const struct qstr anonstring = QSTR_INIT("/", 1); struct dentry *tmp; struct dentry *res; unsigned add_flags; @@ -1869,7 +1870,7 @@ static struct dentry *__d_obtain_alias(struct inode *inode, int disconnected) if (res) goto out_iput; - tmp = __d_alloc(inode->i_sb, &anonstring); + tmp = __d_alloc(inode->i_sb, NULL); if (!tmp) { res = ERR_PTR(-ENOMEM); goto out_iput; -- cgit v1.3-8-gc7d7 From 16c568efff82e4a6a75d2bd86576e648fad8a7fe Mon Sep 17 00:00:00 2001 From: Al Viro Date: Thu, 12 Nov 2015 22:46:49 -0500 Subject: cifs: merge the hash calculation helpers three practically identical copies... Signed-off-by: Al Viro --- fs/cifs/cifsencrypt.c | 97 ++++++++++++++++++++++++------------------- fs/cifs/cifsproto.h | 3 ++ fs/cifs/smb2transport.c | 107 +++++------------------------------------------- 3 files changed, 67 insertions(+), 140 deletions(-) (limited to 'fs') diff --git a/fs/cifs/cifsencrypt.c b/fs/cifs/cifsencrypt.c index 4897dacf8944..6aeb8d4616a4 100644 --- a/fs/cifs/cifsencrypt.c +++ b/fs/cifs/cifsencrypt.c @@ -66,45 +66,15 @@ cifs_crypto_shash_md5_allocate(struct TCP_Server_Info *server) return 0; } -/* - * Calculate and return the CIFS signature based on the mac key and SMB PDU. - * The 16 byte signature must be allocated by the caller. Note we only use the - * 1st eight bytes and that the smb header signature field on input contains - * the sequence number before this function is called. Also, this function - * should be called with the server->srv_mutex held. - */ -static int cifs_calc_signature(struct smb_rqst *rqst, - struct TCP_Server_Info *server, char *signature) +int __cifs_calc_signature(struct smb_rqst *rqst, + struct TCP_Server_Info *server, char *signature, + struct shash_desc *shash) { int i; int rc; struct kvec *iov = rqst->rq_iov; int n_vec = rqst->rq_nvec; - if (iov == NULL || signature == NULL || server == NULL) - return -EINVAL; - - if (!server->secmech.sdescmd5) { - rc = cifs_crypto_shash_md5_allocate(server); - if (rc) { - cifs_dbg(VFS, "%s: Can't alloc md5 crypto\n", __func__); - return -1; - } - } - - rc = crypto_shash_init(&server->secmech.sdescmd5->shash); - if (rc) { - cifs_dbg(VFS, "%s: Could not init md5\n", __func__); - return rc; - } - - rc = crypto_shash_update(&server->secmech.sdescmd5->shash, - server->session_key.response, server->session_key.len); - if (rc) { - cifs_dbg(VFS, "%s: Could not update with response\n", __func__); - return rc; - } - for (i = 0; i < n_vec; i++) { if (iov[i].iov_len == 0) continue; @@ -117,12 +87,10 @@ static int cifs_calc_signature(struct smb_rqst *rqst, if (i == 0) { if (iov[0].iov_len <= 8) /* cmd field at offset 9 */ break; /* nothing to sign or corrupt header */ - rc = - crypto_shash_update(&server->secmech.sdescmd5->shash, + rc = crypto_shash_update(shash, iov[i].iov_base + 4, iov[i].iov_len - 4); } else { - rc = - crypto_shash_update(&server->secmech.sdescmd5->shash, + rc = crypto_shash_update(shash, iov[i].iov_base, iov[i].iov_len); } if (rc) { @@ -134,21 +102,64 @@ static int cifs_calc_signature(struct smb_rqst *rqst, /* now hash over the rq_pages array */ for (i = 0; i < rqst->rq_npages; i++) { - struct kvec p_iov; + void *kaddr = kmap(rqst->rq_pages[i]); + size_t len = rqst->rq_pagesz; + + if (i == rqst->rq_npages - 1) + len = rqst->rq_tailsz; + + crypto_shash_update(shash, kaddr, len); - cifs_rqst_page_to_kvec(rqst, i, &p_iov); - crypto_shash_update(&server->secmech.sdescmd5->shash, - p_iov.iov_base, p_iov.iov_len); kunmap(rqst->rq_pages[i]); } - rc = crypto_shash_final(&server->secmech.sdescmd5->shash, signature); + rc = crypto_shash_final(shash, signature); if (rc) - cifs_dbg(VFS, "%s: Could not generate md5 hash\n", __func__); + cifs_dbg(VFS, "%s: Could not generate hash\n", __func__); return rc; } +/* + * Calculate and return the CIFS signature based on the mac key and SMB PDU. + * The 16 byte signature must be allocated by the caller. Note we only use the + * 1st eight bytes and that the smb header signature field on input contains + * the sequence number before this function is called. Also, this function + * should be called with the server->srv_mutex held. + */ +static int cifs_calc_signature(struct smb_rqst *rqst, + struct TCP_Server_Info *server, char *signature) +{ + int rc; + + if (!rqst->rq_iov || !signature || !server) + return -EINVAL; + + if (!server->secmech.sdescmd5) { + rc = cifs_crypto_shash_md5_allocate(server); + if (rc) { + cifs_dbg(VFS, "%s: Can't alloc md5 crypto\n", __func__); + return -1; + } + } + + rc = crypto_shash_init(&server->secmech.sdescmd5->shash); + if (rc) { + cifs_dbg(VFS, "%s: Could not init md5\n", __func__); + return rc; + } + + rc = crypto_shash_update(&server->secmech.sdescmd5->shash, + server->session_key.response, server->session_key.len); + if (rc) { + cifs_dbg(VFS, "%s: Could not update with response\n", __func__); + return rc; + } + + return __cifs_calc_signature(rqst, server, signature, + &server->secmech.sdescmd5->shash); +} + /* must be called with server->srv_mutex held */ int cifs_sign_rqst(struct smb_rqst *rqst, struct TCP_Server_Info *server, __u32 *pexpected_response_sequence_number) diff --git a/fs/cifs/cifsproto.h b/fs/cifs/cifsproto.h index eed7ff50faf0..d9b4f444fdf9 100644 --- a/fs/cifs/cifsproto.h +++ b/fs/cifs/cifsproto.h @@ -512,4 +512,7 @@ int cifs_create_mf_symlink(unsigned int xid, struct cifs_tcon *tcon, struct cifs_sb_info *cifs_sb, const unsigned char *path, char *pbuf, unsigned int *pbytes_written); +int __cifs_calc_signature(struct smb_rqst *rqst, + struct TCP_Server_Info *server, char *signature, + struct shash_desc *shash); #endif /* _CIFSPROTO_H */ diff --git a/fs/cifs/smb2transport.c b/fs/cifs/smb2transport.c index 8732a43b1008..bc9a7b634643 100644 --- a/fs/cifs/smb2transport.c +++ b/fs/cifs/smb2transport.c @@ -135,11 +135,10 @@ smb2_find_smb_ses(struct smb2_hdr *smb2hdr, struct TCP_Server_Info *server) int smb2_calc_signature(struct smb_rqst *rqst, struct TCP_Server_Info *server) { - int i, rc; + int rc; unsigned char smb2_signature[SMB2_HMACSHA256_SIZE]; unsigned char *sigptr = smb2_signature; struct kvec *iov = rqst->rq_iov; - int n_vec = rqst->rq_nvec; struct smb2_hdr *smb2_pdu = (struct smb2_hdr *)iov[0].iov_base; struct cifs_ses *ses; @@ -171,53 +170,11 @@ smb2_calc_signature(struct smb_rqst *rqst, struct TCP_Server_Info *server) return rc; } - for (i = 0; i < n_vec; i++) { - if (iov[i].iov_len == 0) - continue; - if (iov[i].iov_base == NULL) { - cifs_dbg(VFS, "null iovec entry\n"); - return -EIO; - } - /* - * The first entry includes a length field (which does not get - * signed that occupies the first 4 bytes before the header). - */ - if (i == 0) { - if (iov[0].iov_len <= 8) /* cmd field at offset 9 */ - break; /* nothing to sign or corrupt header */ - rc = - crypto_shash_update( - &server->secmech.sdeschmacsha256->shash, - iov[i].iov_base + 4, iov[i].iov_len - 4); - } else { - rc = - crypto_shash_update( - &server->secmech.sdeschmacsha256->shash, - iov[i].iov_base, iov[i].iov_len); - } - if (rc) { - cifs_dbg(VFS, "%s: Could not update with payload\n", - __func__); - return rc; - } - } - - /* now hash over the rq_pages array */ - for (i = 0; i < rqst->rq_npages; i++) { - struct kvec p_iov; - - cifs_rqst_page_to_kvec(rqst, i, &p_iov); - crypto_shash_update(&server->secmech.sdeschmacsha256->shash, - p_iov.iov_base, p_iov.iov_len); - kunmap(rqst->rq_pages[i]); - } - - rc = crypto_shash_final(&server->secmech.sdeschmacsha256->shash, - sigptr); - if (rc) - cifs_dbg(VFS, "%s: Could not generate sha256 hash\n", __func__); + rc = __cifs_calc_signature(rqst, server, sigptr, + &server->secmech.sdeschmacsha256->shash); - memcpy(smb2_pdu->Signature, sigptr, SMB2_SIGNATURE_SIZE); + if (!rc) + memcpy(smb2_pdu->Signature, sigptr, SMB2_SIGNATURE_SIZE); return rc; } @@ -395,12 +352,10 @@ generate_smb311signingkey(struct cifs_ses *ses) int smb3_calc_signature(struct smb_rqst *rqst, struct TCP_Server_Info *server) { - int i; int rc = 0; unsigned char smb3_signature[SMB2_CMACAES_SIZE]; unsigned char *sigptr = smb3_signature; struct kvec *iov = rqst->rq_iov; - int n_vec = rqst->rq_nvec; struct smb2_hdr *smb2_pdu = (struct smb2_hdr *)iov[0].iov_base; struct cifs_ses *ses; @@ -431,54 +386,12 @@ smb3_calc_signature(struct smb_rqst *rqst, struct TCP_Server_Info *server) cifs_dbg(VFS, "%s: Could not init cmac aes\n", __func__); return rc; } + + rc = __cifs_calc_signature(rqst, server, sigptr, + &server->secmech.sdesccmacaes->shash); - for (i = 0; i < n_vec; i++) { - if (iov[i].iov_len == 0) - continue; - if (iov[i].iov_base == NULL) { - cifs_dbg(VFS, "null iovec entry"); - return -EIO; - } - /* - * The first entry includes a length field (which does not get - * signed that occupies the first 4 bytes before the header). - */ - if (i == 0) { - if (iov[0].iov_len <= 8) /* cmd field at offset 9 */ - break; /* nothing to sign or corrupt header */ - rc = - crypto_shash_update( - &server->secmech.sdesccmacaes->shash, - iov[i].iov_base + 4, iov[i].iov_len - 4); - } else { - rc = - crypto_shash_update( - &server->secmech.sdesccmacaes->shash, - iov[i].iov_base, iov[i].iov_len); - } - if (rc) { - cifs_dbg(VFS, "%s: Couldn't update cmac aes with payload\n", - __func__); - return rc; - } - } - - /* now hash over the rq_pages array */ - for (i = 0; i < rqst->rq_npages; i++) { - struct kvec p_iov; - - cifs_rqst_page_to_kvec(rqst, i, &p_iov); - crypto_shash_update(&server->secmech.sdesccmacaes->shash, - p_iov.iov_base, p_iov.iov_len); - kunmap(rqst->rq_pages[i]); - } - - rc = crypto_shash_final(&server->secmech.sdesccmacaes->shash, - sigptr); - if (rc) - cifs_dbg(VFS, "%s: Could not generate cmac aes\n", __func__); - - memcpy(smb2_pdu->Signature, sigptr, SMB2_SIGNATURE_SIZE); + if (!rc) + memcpy(smb2_pdu->Signature, sigptr, SMB2_SIGNATURE_SIZE); return rc; } -- cgit v1.3-8-gc7d7 From 3ab3f2a1fee45f5788ddccd8b3ece6f74ae905bd Mon Sep 17 00:00:00 2001 From: Al Viro Date: Fri, 13 Nov 2015 02:36:04 -0500 Subject: cifs: quit playing games with draining iovecs ... and use ITER_BVEC for the page part of request to send Signed-off-by: Al Viro --- fs/cifs/cifsproto.h | 2 - fs/cifs/transport.c | 141 +++++++++++++++------------------------------------- 2 files changed, 41 insertions(+), 102 deletions(-) (limited to 'fs') diff --git a/fs/cifs/cifsproto.h b/fs/cifs/cifsproto.h index d9b4f444fdf9..7d5f53a01922 100644 --- a/fs/cifs/cifsproto.h +++ b/fs/cifs/cifsproto.h @@ -37,8 +37,6 @@ extern void cifs_buf_release(void *); extern struct smb_hdr *cifs_small_buf_get(void); extern void cifs_small_buf_release(void *); extern void free_rsp_buf(int, void *); -extern void cifs_rqst_page_to_kvec(struct smb_rqst *rqst, unsigned int idx, - struct kvec *iov); extern int smb_send(struct TCP_Server_Info *, struct smb_hdr *, unsigned int /* length */); extern unsigned int _get_xid(void); diff --git a/fs/cifs/transport.c b/fs/cifs/transport.c index 87abe8ed074c..206a597b2293 100644 --- a/fs/cifs/transport.c +++ b/fs/cifs/transport.c @@ -124,41 +124,32 @@ cifs_delete_mid(struct mid_q_entry *mid) /* * smb_send_kvec - send an array of kvecs to the server * @server: Server to send the data to - * @iov: Pointer to array of kvecs - * @n_vec: length of kvec array + * @smb_msg: Message to send * @sent: amount of data sent on socket is stored here * * Our basic "send data to server" function. Should be called with srv_mutex * held. The caller is responsible for handling the results. */ static int -smb_send_kvec(struct TCP_Server_Info *server, struct kvec *iov, size_t n_vec, - size_t *sent) +smb_send_kvec(struct TCP_Server_Info *server, struct msghdr *smb_msg, + size_t *sent) { int rc = 0; - int i = 0; - struct msghdr smb_msg; - unsigned int remaining; - size_t first_vec = 0; + int retries = 0; struct socket *ssocket = server->ssocket; *sent = 0; - smb_msg.msg_name = (struct sockaddr *) &server->dstaddr; - smb_msg.msg_namelen = sizeof(struct sockaddr); - smb_msg.msg_control = NULL; - smb_msg.msg_controllen = 0; + smb_msg->msg_name = (struct sockaddr *) &server->dstaddr; + smb_msg->msg_namelen = sizeof(struct sockaddr); + smb_msg->msg_control = NULL; + smb_msg->msg_controllen = 0; if (server->noblocksnd) - smb_msg.msg_flags = MSG_DONTWAIT + MSG_NOSIGNAL; + smb_msg->msg_flags = MSG_DONTWAIT + MSG_NOSIGNAL; else - smb_msg.msg_flags = MSG_NOSIGNAL; - - remaining = 0; - for (i = 0; i < n_vec; i++) - remaining += iov[i].iov_len; + smb_msg->msg_flags = MSG_NOSIGNAL; - i = 0; - while (remaining) { + while (msg_data_left(smb_msg)) { /* * If blocking send, we try 3 times, since each can block * for 5 seconds. For nonblocking we have to try more @@ -177,35 +168,21 @@ smb_send_kvec(struct TCP_Server_Info *server, struct kvec *iov, size_t n_vec, * after the retries we will kill the socket and * reconnect which may clear the network problem. */ - rc = kernel_sendmsg(ssocket, &smb_msg, &iov[first_vec], - n_vec - first_vec, remaining); + rc = sock_sendmsg(ssocket, smb_msg); if (rc == -EAGAIN) { - i++; - if (i >= 14 || (!server->noblocksnd && (i > 2))) { + retries++; + if (retries >= 14 || + (!server->noblocksnd && (retries > 2))) { cifs_dbg(VFS, "sends on sock %p stuck for 15 seconds\n", ssocket); - rc = -EAGAIN; - break; + return -EAGAIN; } - msleep(1 << i); + msleep(1 << retries); continue; } if (rc < 0) - break; - - /* send was at least partially successful */ - *sent += rc; - - if (rc == remaining) { - remaining = 0; - break; - } - - if (rc > remaining) { - cifs_dbg(VFS, "sent %d requested %d\n", rc, remaining); - break; - } + return rc; if (rc == 0) { /* should never happen, letting socket clear before @@ -215,59 +192,11 @@ smb_send_kvec(struct TCP_Server_Info *server, struct kvec *iov, size_t n_vec, continue; } - remaining -= rc; - - /* the line below resets i */ - for (i = first_vec; i < n_vec; i++) { - if (iov[i].iov_len) { - if (rc > iov[i].iov_len) { - rc -= iov[i].iov_len; - iov[i].iov_len = 0; - } else { - iov[i].iov_base += rc; - iov[i].iov_len -= rc; - first_vec = i; - break; - } - } - } - - i = 0; /* in case we get ENOSPC on the next send */ - rc = 0; + /* send was at least partially successful */ + *sent += rc; + retries = 0; /* in case we get ENOSPC on the next send */ } - return rc; -} - -/** - * rqst_page_to_kvec - Turn a slot in the smb_rqst page array into a kvec - * @rqst: pointer to smb_rqst - * @idx: index into the array of the page - * @iov: pointer to struct kvec that will hold the result - * - * Helper function to convert a slot in the rqst->rq_pages array into a kvec. - * The page will be kmapped and the address placed into iov_base. The length - * will then be adjusted according to the ptailoff. - */ -void -cifs_rqst_page_to_kvec(struct smb_rqst *rqst, unsigned int idx, - struct kvec *iov) -{ - /* - * FIXME: We could avoid this kmap altogether if we used - * kernel_sendpage instead of kernel_sendmsg. That will only - * work if signing is disabled though as sendpage inlines the - * page directly into the fraglist. If userspace modifies the - * page after we calculate the signature, then the server will - * reject it and may break the connection. kernel_sendmsg does - * an extra copy of the data and avoids that issue. - */ - iov->iov_base = kmap(rqst->rq_pages[idx]); - - /* if last page, don't send beyond this offset into page */ - if (idx == (rqst->rq_npages - 1)) - iov->iov_len = rqst->rq_tailsz; - else - iov->iov_len = rqst->rq_pagesz; + return 0; } static unsigned long @@ -299,8 +228,9 @@ smb_send_rqst(struct TCP_Server_Info *server, struct smb_rqst *rqst) unsigned int smb_buf_length = get_rfc1002_length(iov[0].iov_base); unsigned long send_length; unsigned int i; - size_t total_len = 0, sent; + size_t total_len = 0, sent, size; struct socket *ssocket = server->ssocket; + struct msghdr smb_msg; int val = 1; if (ssocket == NULL) @@ -321,7 +251,13 @@ smb_send_rqst(struct TCP_Server_Info *server, struct smb_rqst *rqst) kernel_setsockopt(ssocket, SOL_TCP, TCP_CORK, (char *)&val, sizeof(val)); - rc = smb_send_kvec(server, iov, n_vec, &sent); + size = 0; + for (i = 0; i < n_vec; i++) + size += iov[i].iov_len; + + iov_iter_kvec(&smb_msg.msg_iter, WRITE | ITER_KVEC, iov, n_vec, size); + + rc = smb_send_kvec(server, &smb_msg, &sent); if (rc < 0) goto uncork; @@ -329,11 +265,16 @@ smb_send_rqst(struct TCP_Server_Info *server, struct smb_rqst *rqst) /* now walk the page array and send each page in it */ for (i = 0; i < rqst->rq_npages; i++) { - struct kvec p_iov; - - cifs_rqst_page_to_kvec(rqst, i, &p_iov); - rc = smb_send_kvec(server, &p_iov, 1, &sent); - kunmap(rqst->rq_pages[i]); + size_t len = i == rqst->rq_npages - 1 + ? rqst->rq_tailsz + : rqst->rq_pagesz; + struct bio_vec bvec = { + .bv_page = rqst->rq_pages[i], + .bv_len = len + }; + iov_iter_bvec(&smb_msg.msg_iter, WRITE | ITER_BVEC, + &bvec, 1, len); + rc = smb_send_kvec(server, &smb_msg, &sent); if (rc < 0) break; -- cgit v1.3-8-gc7d7 From 09aab880f7c5ac31e9db18a63353b980bcb52261 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Fri, 13 Nov 2015 03:00:17 -0500 Subject: cifs: no need to wank with copying and advancing iovec on recvmsg side either Signed-off-by: Al Viro --- fs/cifs/cifsglob.h | 2 -- fs/cifs/connect.c | 72 ++++-------------------------------------------------- 2 files changed, 5 insertions(+), 69 deletions(-) (limited to 'fs') diff --git a/fs/cifs/cifsglob.h b/fs/cifs/cifsglob.h index d21da9f05bae..df03c5e7d633 100644 --- a/fs/cifs/cifsglob.h +++ b/fs/cifs/cifsglob.h @@ -615,8 +615,6 @@ struct TCP_Server_Info { bool sec_mskerberos; /* supports legacy MS Kerberos */ bool large_buf; /* is current buffer large? */ struct delayed_work echo; /* echo ping workqueue job */ - struct kvec *iov; /* reusable kvec array for receives */ - unsigned int nr_iov; /* number of kvecs in array */ char *smallbuf; /* pointer to current "small" buffer */ char *bigbuf; /* pointer to current "big" buffer */ unsigned int total_read; /* total amount of data read in this pass */ diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c index a763cd3d9e7c..eb426658566d 100644 --- a/fs/cifs/connect.c +++ b/fs/cifs/connect.c @@ -501,77 +501,20 @@ server_unresponsive(struct TCP_Server_Info *server) return false; } -/* - * kvec_array_init - clone a kvec array, and advance into it - * @new: pointer to memory for cloned array - * @iov: pointer to original array - * @nr_segs: number of members in original array - * @bytes: number of bytes to advance into the cloned array - * - * This function will copy the array provided in iov to a section of memory - * and advance the specified number of bytes into the new array. It returns - * the number of segments in the new array. "new" must be at least as big as - * the original iov array. - */ -static unsigned int -kvec_array_init(struct kvec *new, struct kvec *iov, unsigned int nr_segs, - size_t bytes) -{ - size_t base = 0; - - while (bytes || !iov->iov_len) { - int copy = min(bytes, iov->iov_len); - - bytes -= copy; - base += copy; - if (iov->iov_len == base) { - iov++; - nr_segs--; - base = 0; - } - } - memcpy(new, iov, sizeof(*iov) * nr_segs); - new->iov_base += base; - new->iov_len -= base; - return nr_segs; -} - -static struct kvec * -get_server_iovec(struct TCP_Server_Info *server, unsigned int nr_segs) -{ - struct kvec *new_iov; - - if (server->iov && nr_segs <= server->nr_iov) - return server->iov; - - /* not big enough -- allocate a new one and release the old */ - new_iov = kmalloc(sizeof(*new_iov) * nr_segs, GFP_NOFS); - if (new_iov) { - kfree(server->iov); - server->iov = new_iov; - server->nr_iov = nr_segs; - } - return new_iov; -} - int cifs_readv_from_socket(struct TCP_Server_Info *server, struct kvec *iov_orig, unsigned int nr_segs, unsigned int to_read) { int length = 0; int total_read; - unsigned int segs; struct msghdr smb_msg; - struct kvec *iov; - - iov = get_server_iovec(server, nr_segs); - if (!iov) - return -ENOMEM; smb_msg.msg_control = NULL; smb_msg.msg_controllen = 0; + iov_iter_kvec(&smb_msg.msg_iter, READ | ITER_KVEC, + iov_orig, nr_segs, to_read); - for (total_read = 0; to_read; total_read += length, to_read -= length) { + for (total_read = 0; msg_data_left(&smb_msg); total_read += length) { try_to_freeze(); if (server_unresponsive(server)) { @@ -579,10 +522,7 @@ cifs_readv_from_socket(struct TCP_Server_Info *server, struct kvec *iov_orig, break; } - segs = kvec_array_init(iov, iov_orig, nr_segs, total_read); - - length = kernel_recvmsg(server->ssocket, &smb_msg, - iov, segs, to_read, 0); + length = sock_recvmsg(server->ssocket, &smb_msg, 0); if (server->tcpStatus == CifsExiting) { total_read = -ESHUTDOWN; @@ -603,8 +543,7 @@ cifs_readv_from_socket(struct TCP_Server_Info *server, struct kvec *iov_orig, length = 0; continue; } else if (length <= 0) { - cifs_dbg(FYI, "Received no data or error: expecting %d\n" - "got %d", to_read, length); + cifs_dbg(FYI, "Received no data or error: %d\n", length); cifs_reconnect(server); total_read = -ECONNABORTED; break; @@ -783,7 +722,6 @@ static void clean_demultiplex_info(struct TCP_Server_Info *server) } kfree(server->hostname); - kfree(server->iov); kfree(server); length = atomic_dec_return(&tcpSesAllocCount); -- cgit v1.3-8-gc7d7 From a6137305a8c47fa92ab1a8efcfe76f0e9fa96ab7 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 9 Jan 2016 19:37:16 -0500 Subject: cifs_readv_receive: use cifs_read_from_socket() Signed-off-by: Al Viro --- fs/cifs/cifssmb.c | 11 ++++------- 1 file changed, 4 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/cifs/cifssmb.c b/fs/cifs/cifssmb.c index 76fcb50295a3..3da077afad1f 100644 --- a/fs/cifs/cifssmb.c +++ b/fs/cifs/cifssmb.c @@ -1447,10 +1447,8 @@ cifs_readv_receive(struct TCP_Server_Info *server, struct mid_q_entry *mid) len = min_t(unsigned int, buflen, server->vals->read_rsp_size) - HEADER_SIZE(server) + 1; - rdata->iov.iov_base = buf + HEADER_SIZE(server) - 1; - rdata->iov.iov_len = len; - - length = cifs_readv_from_socket(server, &rdata->iov, 1, len); + length = cifs_read_from_socket(server, + buf + HEADER_SIZE(server) - 1, len); if (length < 0) return length; server->total_read += length; @@ -1502,9 +1500,8 @@ cifs_readv_receive(struct TCP_Server_Info *server, struct mid_q_entry *mid) len = data_offset - server->total_read; if (len > 0) { /* read any junk before data into the rest of smallbuf */ - rdata->iov.iov_base = buf + server->total_read; - rdata->iov.iov_len = len; - length = cifs_readv_from_socket(server, &rdata->iov, 1, len); + length = cifs_read_from_socket(server, + buf + server->total_read, len); if (length < 0) return length; server->total_read += length; -- cgit v1.3-8-gc7d7 From 71335664c38f03de10d7cf1d82705fe55a130b33 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 9 Jan 2016 19:54:50 -0500 Subject: cifs: don't bother with kmap on read_pages side just do ITER_BVEC recvmsg Signed-off-by: Al Viro --- fs/cifs/cifsproto.h | 7 +++--- fs/cifs/connect.c | 65 ++++++++++++++++++++++++++++------------------------- fs/cifs/file.c | 53 ++++++++++++++----------------------------- 3 files changed, 55 insertions(+), 70 deletions(-) (limited to 'fs') diff --git a/fs/cifs/cifsproto.h b/fs/cifs/cifsproto.h index 7d5f53a01922..0f9a6bc4ba43 100644 --- a/fs/cifs/cifsproto.h +++ b/fs/cifs/cifsproto.h @@ -179,10 +179,9 @@ extern int set_cifs_acl(struct cifs_ntsd *, __u32, struct inode *, extern void dequeue_mid(struct mid_q_entry *mid, bool malformed); extern int cifs_read_from_socket(struct TCP_Server_Info *server, char *buf, - unsigned int to_read); -extern int cifs_readv_from_socket(struct TCP_Server_Info *server, - struct kvec *iov_orig, unsigned int nr_segs, - unsigned int to_read); + unsigned int to_read); +extern int cifs_read_page_from_socket(struct TCP_Server_Info *server, + struct page *page, unsigned int to_read); extern void cifs_setup_cifs_sb(struct smb_vol *pvolume_info, struct cifs_sb_info *cifs_sb); extern int cifs_match_super(struct super_block *, void *); diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c index eb426658566d..e33c5e0ecfd0 100644 --- a/fs/cifs/connect.c +++ b/fs/cifs/connect.c @@ -501,39 +501,34 @@ server_unresponsive(struct TCP_Server_Info *server) return false; } -int -cifs_readv_from_socket(struct TCP_Server_Info *server, struct kvec *iov_orig, - unsigned int nr_segs, unsigned int to_read) +static int +cifs_readv_from_socket(struct TCP_Server_Info *server, struct msghdr *smb_msg) { int length = 0; int total_read; - struct msghdr smb_msg; - smb_msg.msg_control = NULL; - smb_msg.msg_controllen = 0; - iov_iter_kvec(&smb_msg.msg_iter, READ | ITER_KVEC, - iov_orig, nr_segs, to_read); + smb_msg->msg_control = NULL; + smb_msg->msg_controllen = 0; - for (total_read = 0; msg_data_left(&smb_msg); total_read += length) { + for (total_read = 0; msg_data_left(smb_msg); total_read += length) { try_to_freeze(); - if (server_unresponsive(server)) { - total_read = -ECONNABORTED; - break; - } + if (server_unresponsive(server)) + return -ECONNABORTED; - length = sock_recvmsg(server->ssocket, &smb_msg, 0); + length = sock_recvmsg(server->ssocket, smb_msg, 0); - if (server->tcpStatus == CifsExiting) { - total_read = -ESHUTDOWN; - break; - } else if (server->tcpStatus == CifsNeedReconnect) { + if (server->tcpStatus == CifsExiting) + return -ESHUTDOWN; + + if (server->tcpStatus == CifsNeedReconnect) { cifs_reconnect(server); - total_read = -ECONNABORTED; - break; - } else if (length == -ERESTARTSYS || - length == -EAGAIN || - length == -EINTR) { + return -ECONNABORTED; + } + + if (length == -ERESTARTSYS || + length == -EAGAIN || + length == -EINTR) { /* * Minimum sleep to prevent looping, allowing socket * to clear and app threads to set tcpStatus @@ -542,11 +537,12 @@ cifs_readv_from_socket(struct TCP_Server_Info *server, struct kvec *iov_orig, usleep_range(1000, 2000); length = 0; continue; - } else if (length <= 0) { + } + + if (length <= 0) { cifs_dbg(FYI, "Received no data or error: %d\n", length); cifs_reconnect(server); - total_read = -ECONNABORTED; - break; + return -ECONNABORTED; } } return total_read; @@ -556,12 +552,21 @@ int cifs_read_from_socket(struct TCP_Server_Info *server, char *buf, unsigned int to_read) { - struct kvec iov; + struct msghdr smb_msg; + struct kvec iov = {.iov_base = buf, .iov_len = to_read}; + iov_iter_kvec(&smb_msg.msg_iter, READ | ITER_KVEC, &iov, 1, to_read); - iov.iov_base = buf; - iov.iov_len = to_read; + return cifs_readv_from_socket(server, &smb_msg); +} - return cifs_readv_from_socket(server, &iov, 1, to_read); +int +cifs_read_page_from_socket(struct TCP_Server_Info *server, struct page *page, + unsigned int to_read) +{ + struct msghdr smb_msg; + struct bio_vec bv = {.bv_page = page, .bv_len = to_read}; + iov_iter_bvec(&smb_msg.msg_iter, READ | ITER_BVEC, &bv, 1, to_read); + return cifs_readv_from_socket(server, &smb_msg); } static bool diff --git a/fs/cifs/file.c b/fs/cifs/file.c index ff882aeaccc6..0f718679186e 100644 --- a/fs/cifs/file.c +++ b/fs/cifs/file.c @@ -2855,39 +2855,31 @@ cifs_uncached_read_into_pages(struct TCP_Server_Info *server, int result = 0; unsigned int i; unsigned int nr_pages = rdata->nr_pages; - struct kvec iov; rdata->got_bytes = 0; rdata->tailsz = PAGE_SIZE; for (i = 0; i < nr_pages; i++) { struct page *page = rdata->pages[i]; + size_t n; - if (len >= PAGE_SIZE) { - /* enough data to fill the page */ - iov.iov_base = kmap(page); - iov.iov_len = PAGE_SIZE; - cifs_dbg(FYI, "%u: iov_base=%p iov_len=%zu\n", - i, iov.iov_base, iov.iov_len); - len -= PAGE_SIZE; - } else if (len > 0) { - /* enough for partial page, fill and zero the rest */ - iov.iov_base = kmap(page); - iov.iov_len = len; - cifs_dbg(FYI, "%u: iov_base=%p iov_len=%zu\n", - i, iov.iov_base, iov.iov_len); - memset(iov.iov_base + len, '\0', PAGE_SIZE - len); - rdata->tailsz = len; - len = 0; - } else { + if (len <= 0) { /* no need to hold page hostage */ rdata->pages[i] = NULL; rdata->nr_pages--; put_page(page); continue; } - - result = cifs_readv_from_socket(server, &iov, 1, iov.iov_len); - kunmap(page); + n = len; + if (len >= PAGE_SIZE) { + /* enough data to fill the page */ + n = PAGE_SIZE; + len -= n; + } else { + zero_user(page, len, PAGE_SIZE - len); + rdata->tailsz = len; + len = 0; + } + result = cifs_read_page_from_socket(server, page, n); if (result < 0) break; @@ -3303,7 +3295,6 @@ cifs_readpages_read_into_pages(struct TCP_Server_Info *server, u64 eof; pgoff_t eof_index; unsigned int nr_pages = rdata->nr_pages; - struct kvec iov; /* determine the eof that the server (probably) has */ eof = CIFS_I(rdata->mapping->host)->server_eof; @@ -3314,23 +3305,14 @@ cifs_readpages_read_into_pages(struct TCP_Server_Info *server, rdata->tailsz = PAGE_CACHE_SIZE; for (i = 0; i < nr_pages; i++) { struct page *page = rdata->pages[i]; + size_t n = PAGE_CACHE_SIZE; if (len >= PAGE_CACHE_SIZE) { - /* enough data to fill the page */ - iov.iov_base = kmap(page); - iov.iov_len = PAGE_CACHE_SIZE; - cifs_dbg(FYI, "%u: idx=%lu iov_base=%p iov_len=%zu\n", - i, page->index, iov.iov_base, iov.iov_len); len -= PAGE_CACHE_SIZE; } else if (len > 0) { /* enough for partial page, fill and zero the rest */ - iov.iov_base = kmap(page); - iov.iov_len = len; - cifs_dbg(FYI, "%u: idx=%lu iov_base=%p iov_len=%zu\n", - i, page->index, iov.iov_base, iov.iov_len); - memset(iov.iov_base + len, - '\0', PAGE_CACHE_SIZE - len); - rdata->tailsz = len; + zero_user(page, len, PAGE_CACHE_SIZE - len); + n = rdata->tailsz = len; len = 0; } else if (page->index > eof_index) { /* @@ -3360,8 +3342,7 @@ cifs_readpages_read_into_pages(struct TCP_Server_Info *server, continue; } - result = cifs_readv_from_socket(server, &iov, 1, iov.iov_len); - kunmap(page); + result = cifs_read_page_from_socket(server, page, n); if (result < 0) break; -- cgit v1.3-8-gc7d7 From 49db08c358873af11ba3c25401de88156fa5d365 Mon Sep 17 00:00:00 2001 From: Linus Walleij Date: Fri, 19 Feb 2016 15:36:07 +0100 Subject: chrdev: emit a warning when we go below dynamic major range Currently a dynamically allocated character device major is taken from 254 and downward. This mechanism is used for RTC, IIO and a few other subsystems. The kernel currently has no check prevening these dynamic allocations from eating into the assigned numbers at 233 and downward. In a recent test it was reported that so many dynamic device majors were used on a test server, that the major number for infiniband (231) was stolen. This occurred when allocating a new major number for GPIO chips. The error messages from the kernel were not helpful. (See: https://lkml.org/lkml/2016/2/14/124) This patch adds a defined lower limit of the dynamic major allocation region will henceforth emit a warning if we start to eat into the assigned numbers. It does not do any semantic changes and will not change the kernels behaviour: numbers will still continue to be stolen, but we will know from dmesg what is going on. This also updates the Documentation/devices.txt to clearly reflect that we are using this range of major numbers for dynamic allocation. Reported-by: Ying Huang Cc: Linus Torvalds Cc: Greg Kroah-Hartman Cc: Alan Cox Cc: Arnd Bergmann Signed-off-by: Linus Walleij Signed-off-by: Greg Kroah-Hartman --- Documentation/devices.txt | 6 +++--- fs/char_dev.c | 4 ++++ include/linux/fs.h | 2 ++ 3 files changed, 9 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/Documentation/devices.txt b/Documentation/devices.txt index 87b4c5e82d39..0a3588a9798d 100644 --- a/Documentation/devices.txt +++ b/Documentation/devices.txt @@ -3099,9 +3099,9 @@ Your cooperation is appreciated. 129 = /dev/ipath_sma Device used by Subnet Management Agent 130 = /dev/ipath_diag Device used by diagnostics programs -234-239 UNASSIGNED - -240-254 char LOCAL/EXPERIMENTAL USE +234-254 char RESERVED FOR DYNAMIC ASSIGNMENT + Character devices that request a dynamic allocation of major number will + take numbers starting from 254 and downward. 240-254 block LOCAL/EXPERIMENTAL USE Allocated for local/experimental use. For devices not diff --git a/fs/char_dev.c b/fs/char_dev.c index 24b142569ca9..687471dc04a0 100644 --- a/fs/char_dev.c +++ b/fs/char_dev.c @@ -91,6 +91,10 @@ __register_chrdev_region(unsigned int major, unsigned int baseminor, break; } + if (i < CHRDEV_MAJOR_DYN_END) + pr_warn("CHRDEV \"%s\" major number %d goes below the dynamic allocation range", + name, i); + if (i == 0) { ret = -EBUSY; goto out; diff --git a/include/linux/fs.h b/include/linux/fs.h index 14a97194b34b..60082be96de8 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -2385,6 +2385,8 @@ static inline void bd_unlink_disk_holder(struct block_device *bdev, /* fs/char_dev.c */ #define CHRDEV_MAJOR_HASH_SIZE 255 +/* Marks the bottom of the first segment of free char majors */ +#define CHRDEV_MAJOR_DYN_END 234 extern int alloc_chrdev_region(dev_t *, unsigned, unsigned, const char *); extern int register_chrdev_region(dev_t, unsigned, const char *); extern int __register_chrdev(unsigned int major, unsigned int baseminor, -- cgit v1.3-8-gc7d7 From a8f324a46fbe5473633af00039e81821be0ce51b Mon Sep 17 00:00:00 2001 From: Roman Pen Date: Tue, 9 Feb 2016 11:30:29 +0100 Subject: debugfs: fix inode i_nlink references for automount dentry Directory inodes should start off with i_nlink == 2 (one extra ref for "." entry). debugfs_create_automount() increases neither the i_nlink reference for current inode nor for parent inode. On attempt to remove the automount dentry, kernel complains: [ 86.288070] WARNING: CPU: 1 PID: 3616 at fs/inode.c:273 drop_nlink+0x3e/0x50() [ 86.288461] Modules linked in: debugfs_example2(O-) [ 86.288745] CPU: 1 PID: 3616 Comm: rmmod Tainted: G O 4.4.0-rc3-next-20151207+ #135 [ 86.289197] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.8.2-20150617_082717-anatol 04/01/2014 [ 86.289696] ffffffff81be05c9 ffff8800b9e6fda0 ffffffff81352e2c 0000000000000000 [ 86.290110] ffff8800b9e6fdd8 ffffffff81065142 ffff8801399175e8 ffff8800bb78b240 [ 86.290507] ffff8801399175e8 ffff8800b73d7898 ffff8800b73d7840 ffff8800b9e6fde8 [ 86.290933] Call Trace: [ 86.291080] [] dump_stack+0x4e/0x82 [ 86.291340] [] warn_slowpath_common+0x82/0xc0 [ 86.291640] [] warn_slowpath_null+0x1a/0x20 [ 86.291932] [] drop_nlink+0x3e/0x50 [ 86.292208] [] simple_unlink+0x4b/0x60 [ 86.292481] [] simple_rmdir+0x37/0x50 [ 86.292748] [] __debugfs_remove.part.16+0xa8/0xd0 [ 86.293082] [] debugfs_remove_recursive+0xdb/0x1c0 [ 86.293406] [] cleanup_module+0x2d/0x3b [debugfs_example2] [ 86.293762] [] SyS_delete_module+0x16b/0x220 [ 86.294077] [] entry_SYSCALL_64_fastpath+0x12/0x6a [ 86.294405] ---[ end trace c9fc53353fe14a36 ]--- [ 86.294639] ------------[ cut here ]------------ To reproduce the issue it is enough to invoke these lines: autom = debugfs_create_automount("automount", NULL, vfsmount_cb, data); BUG_ON(IS_ERR_OR_NULL(autom)); debugfs_remove(autom); The issue is fixed by increasing inode i_nlink references for current and parent inodes. Signed-off-by: Roman Pen Signed-off-by: Greg Kroah-Hartman --- fs/debugfs/inode.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'fs') diff --git a/fs/debugfs/inode.c b/fs/debugfs/inode.c index bece948b363d..13d65f8a7b12 100644 --- a/fs/debugfs/inode.c +++ b/fs/debugfs/inode.c @@ -461,7 +461,11 @@ struct dentry *debugfs_create_automount(const char *name, inode->i_flags |= S_AUTOMOUNT; inode->i_private = data; dentry->d_fsdata = (void *)f; + /* directory inodes start off with i_nlink == 2 (for "." entry) */ + inc_nlink(inode); d_instantiate(dentry, inode); + inc_nlink(d_inode(dentry->d_parent)); + fsnotify_mkdir(d_inode(dentry->d_parent), dentry); return end_creating(dentry); } EXPORT_SYMBOL(debugfs_create_automount); -- cgit v1.3-8-gc7d7 From 1b48b530dac3e600d92854d98a2a519243661f6c Mon Sep 17 00:00:00 2001 From: Deepa Dinamani Date: Mon, 22 Feb 2016 07:17:47 -0800 Subject: fs: debugfs: Replace CURRENT_TIME by current_fs_time() CURRENT_TIME macro is not appropriate for filesystems as it doesn't use the right granularity for filesystem timestamps. Use current_fs_time() instead. Signed-off-by: Deepa Dinamani Signed-off-by: Greg Kroah-Hartman --- fs/debugfs/inode.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/debugfs/inode.c b/fs/debugfs/inode.c index 13d65f8a7b12..b1e7f35f3cd4 100644 --- a/fs/debugfs/inode.c +++ b/fs/debugfs/inode.c @@ -39,7 +39,8 @@ static struct inode *debugfs_get_inode(struct super_block *sb) struct inode *inode = new_inode(sb); if (inode) { inode->i_ino = get_next_ino(); - inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME; + inode->i_atime = inode->i_mtime = + inode->i_ctime = current_fs_time(sb); } return inode; } -- cgit v1.3-8-gc7d7 From 3a3a5fece6f28c14d3d05c74fb7696412e53a067 Mon Sep 17 00:00:00 2001 From: Deepa Dinamani Date: Mon, 22 Feb 2016 07:17:53 -0800 Subject: fs: kernfs: Replace CURRENT_TIME by current_fs_time() This is in preparation for the series that transitions filesystem timestamps to use 64 bit time and hence make them y2038 safe. CURRENT_TIME macro will be deleted before merging the aforementioned series. Use current_fs_time() instead of CURRENT_TIME for inode timestamps. struct kernfs_node is associated with a sysfs file/ directory. Truncate the values to appropriate time granularity when writing to inode timestamps of the files. ktime_get_real_ts() is used to obtain times for struct kernfs_iattrs. Since these times are later assigned to inode times using timespec_truncate() for all filesystem based operations, we can save the supers list traversal time here by using ktime_get_real_ts() directly. Signed-off-by: Deepa Dinamani Signed-off-by: Greg Kroah-Hartman --- fs/kernfs/dir.c | 8 +++++--- fs/kernfs/inode.c | 15 ++++++++++----- 2 files changed, 15 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/kernfs/dir.c b/fs/kernfs/dir.c index 03b688d19f69..eb2c58732bcf 100644 --- a/fs/kernfs/dir.c +++ b/fs/kernfs/dir.c @@ -753,7 +753,8 @@ int kernfs_add_one(struct kernfs_node *kn) ps_iattr = parent->iattr; if (ps_iattr) { struct iattr *ps_iattrs = &ps_iattr->ia_iattr; - ps_iattrs->ia_ctime = ps_iattrs->ia_mtime = CURRENT_TIME; + ktime_get_real_ts(&ps_iattrs->ia_ctime); + ps_iattrs->ia_mtime = ps_iattrs->ia_ctime; } mutex_unlock(&kernfs_mutex); @@ -1279,8 +1280,9 @@ static void __kernfs_remove(struct kernfs_node *kn) /* update timestamps on the parent */ if (ps_iattr) { - ps_iattr->ia_iattr.ia_ctime = CURRENT_TIME; - ps_iattr->ia_iattr.ia_mtime = CURRENT_TIME; + ktime_get_real_ts(&ps_iattr->ia_iattr.ia_ctime); + ps_iattr->ia_iattr.ia_mtime = + ps_iattr->ia_iattr.ia_ctime; } kernfs_put(pos); diff --git a/fs/kernfs/inode.c b/fs/kernfs/inode.c index 16405ae88d2d..1ac1dbae5aba 100644 --- a/fs/kernfs/inode.c +++ b/fs/kernfs/inode.c @@ -54,7 +54,10 @@ static struct kernfs_iattrs *kernfs_iattrs(struct kernfs_node *kn) iattrs->ia_mode = kn->mode; iattrs->ia_uid = GLOBAL_ROOT_UID; iattrs->ia_gid = GLOBAL_ROOT_GID; - iattrs->ia_atime = iattrs->ia_mtime = iattrs->ia_ctime = CURRENT_TIME; + + ktime_get_real_ts(&iattrs->ia_atime); + iattrs->ia_mtime = iattrs->ia_atime; + iattrs->ia_ctime = iattrs->ia_atime; simple_xattrs_init(&kn->iattr->xattrs); out_unlock: @@ -236,16 +239,18 @@ ssize_t kernfs_iop_listxattr(struct dentry *dentry, char *buf, size_t size) static inline void set_default_inode_attr(struct inode *inode, umode_t mode) { inode->i_mode = mode; - inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME; + inode->i_atime = inode->i_mtime = + inode->i_ctime = current_fs_time(inode->i_sb); } static inline void set_inode_attr(struct inode *inode, struct iattr *iattr) { + struct super_block *sb = inode->i_sb; inode->i_uid = iattr->ia_uid; inode->i_gid = iattr->ia_gid; - inode->i_atime = iattr->ia_atime; - inode->i_mtime = iattr->ia_mtime; - inode->i_ctime = iattr->ia_ctime; + inode->i_atime = timespec_trunc(iattr->ia_atime, sb->s_time_gran); + inode->i_mtime = timespec_trunc(iattr->ia_mtime, sb->s_time_gran); + inode->i_ctime = timespec_trunc(iattr->ia_ctime, sb->s_time_gran); } static void kernfs_refresh_inode(struct kernfs_node *kn, struct inode *inode) -- cgit v1.3-8-gc7d7 From b1168a928277149e1c606763d76ff5c728988755 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Mon, 28 Mar 2016 00:30:35 -0400 Subject: ecryptfs: avoid multiple aliases for directories ecryptfs_lookup_interpose should use d_splice_alias(), not d_add() (and return struct dentry * rather than int). Get rid of redundant dir_inode argument, while we are touching it... Signed-off-by: Al Viro --- fs/ecryptfs/inode.c | 40 ++++++++++++++++++++-------------------- 1 file changed, 20 insertions(+), 20 deletions(-) (limited to 'fs') diff --git a/fs/ecryptfs/inode.c b/fs/ecryptfs/inode.c index 121114e9a464..91ebc8f18e20 100644 --- a/fs/ecryptfs/inode.c +++ b/fs/ecryptfs/inode.c @@ -324,9 +324,8 @@ static int ecryptfs_i_size_read(struct dentry *dentry, struct inode *inode) /** * ecryptfs_lookup_interpose - Dentry interposition for a lookup */ -static int ecryptfs_lookup_interpose(struct dentry *dentry, - struct dentry *lower_dentry, - struct inode *dir_inode) +static struct dentry *ecryptfs_lookup_interpose(struct dentry *dentry, + struct dentry *lower_dentry) { struct inode *inode, *lower_inode = d_inode(lower_dentry); struct ecryptfs_dentry_info *dentry_info; @@ -339,11 +338,12 @@ static int ecryptfs_lookup_interpose(struct dentry *dentry, "to allocate ecryptfs_dentry_info struct\n", __func__); dput(lower_dentry); - return -ENOMEM; + return ERR_PTR(-ENOMEM); } lower_mnt = mntget(ecryptfs_dentry_to_lower_mnt(dentry->d_parent)); - fsstack_copy_attr_atime(dir_inode, d_inode(lower_dentry->d_parent)); + fsstack_copy_attr_atime(d_inode(dentry->d_parent), + d_inode(lower_dentry->d_parent)); BUG_ON(!d_count(lower_dentry)); ecryptfs_set_dentry_private(dentry, dentry_info); @@ -353,27 +353,25 @@ static int ecryptfs_lookup_interpose(struct dentry *dentry, if (d_really_is_negative(lower_dentry)) { /* We want to add because we couldn't find in lower */ d_add(dentry, NULL); - return 0; + return NULL; } - inode = __ecryptfs_get_inode(lower_inode, dir_inode->i_sb); + inode = __ecryptfs_get_inode(lower_inode, dentry->d_sb); if (IS_ERR(inode)) { printk(KERN_ERR "%s: Error interposing; rc = [%ld]\n", __func__, PTR_ERR(inode)); - return PTR_ERR(inode); + return ERR_CAST(inode); } if (S_ISREG(inode->i_mode)) { rc = ecryptfs_i_size_read(dentry, inode); if (rc) { make_bad_inode(inode); - return rc; + return ERR_PTR(rc); } } if (inode->i_state & I_NEW) unlock_new_inode(inode); - d_add(dentry, inode); - - return rc; + return d_splice_alias(inode, dentry); } /** @@ -393,6 +391,7 @@ static struct dentry *ecryptfs_lookup(struct inode *ecryptfs_dir_inode, size_t encrypted_and_encoded_name_size; struct ecryptfs_mount_crypt_stat *mount_crypt_stat = NULL; struct dentry *lower_dir_dentry, *lower_dentry; + struct dentry *res; int rc = 0; lower_dir_dentry = ecryptfs_dentry_to_lower(ecryptfs_dentry->d_parent); @@ -400,10 +399,10 @@ static struct dentry *ecryptfs_lookup(struct inode *ecryptfs_dir_inode, lower_dir_dentry, ecryptfs_dentry->d_name.len); if (IS_ERR(lower_dentry)) { - rc = PTR_ERR(lower_dentry); ecryptfs_printk(KERN_DEBUG, "%s: lookup_one_len() returned " - "[%d] on lower_dentry = [%pd]\n", __func__, rc, - ecryptfs_dentry); + "[%ld] on lower_dentry = [%pd]\n", __func__, + PTR_ERR(lower_dentry), ecryptfs_dentry); + res = ERR_CAST(lower_dentry); goto out; } if (d_really_is_positive(lower_dentry)) @@ -421,24 +420,25 @@ static struct dentry *ecryptfs_lookup(struct inode *ecryptfs_dir_inode, if (rc) { printk(KERN_ERR "%s: Error attempting to encrypt and encode " "filename; rc = [%d]\n", __func__, rc); + res = ERR_PTR(rc); goto out; } lower_dentry = lookup_one_len_unlocked(encrypted_and_encoded_name, lower_dir_dentry, encrypted_and_encoded_name_size); if (IS_ERR(lower_dentry)) { - rc = PTR_ERR(lower_dentry); ecryptfs_printk(KERN_DEBUG, "%s: lookup_one_len() returned " - "[%d] on lower_dentry = [%s]\n", __func__, rc, + "[%ld] on lower_dentry = [%s]\n", __func__, + PTR_ERR(lower_dentry), encrypted_and_encoded_name); + res = ERR_CAST(lower_dentry); goto out; } interpose: - rc = ecryptfs_lookup_interpose(ecryptfs_dentry, lower_dentry, - ecryptfs_dir_inode); + res = ecryptfs_lookup_interpose(ecryptfs_dentry, lower_dentry); out: kfree(encrypted_and_encoded_name); - return ERR_PTR(rc); + return res; } static int ecryptfs_link(struct dentry *old_dentry, struct inode *dir, -- cgit v1.3-8-gc7d7 From 88ae4ab9802eaa8e400e611f85883143d89d6b61 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Mon, 28 Mar 2016 00:43:29 -0400 Subject: ecryptfs_lookup(): try either only encrypted or plaintext name Signed-off-by: Al Viro --- fs/ecryptfs/inode.c | 55 +++++++++++++++++++---------------------------------- 1 file changed, 20 insertions(+), 35 deletions(-) (limited to 'fs') diff --git a/fs/ecryptfs/inode.c b/fs/ecryptfs/inode.c index 91ebc8f18e20..2b8fc9b6b570 100644 --- a/fs/ecryptfs/inode.c +++ b/fs/ecryptfs/inode.c @@ -388,55 +388,40 @@ static struct dentry *ecryptfs_lookup(struct inode *ecryptfs_dir_inode, unsigned int flags) { char *encrypted_and_encoded_name = NULL; - size_t encrypted_and_encoded_name_size; - struct ecryptfs_mount_crypt_stat *mount_crypt_stat = NULL; + struct ecryptfs_mount_crypt_stat *mount_crypt_stat; struct dentry *lower_dir_dentry, *lower_dentry; + const char *name = ecryptfs_dentry->d_name.name; + size_t len = ecryptfs_dentry->d_name.len; struct dentry *res; int rc = 0; lower_dir_dentry = ecryptfs_dentry_to_lower(ecryptfs_dentry->d_parent); - lower_dentry = lookup_one_len_unlocked(ecryptfs_dentry->d_name.name, - lower_dir_dentry, - ecryptfs_dentry->d_name.len); - if (IS_ERR(lower_dentry)) { - ecryptfs_printk(KERN_DEBUG, "%s: lookup_one_len() returned " - "[%ld] on lower_dentry = [%pd]\n", __func__, - PTR_ERR(lower_dentry), ecryptfs_dentry); - res = ERR_CAST(lower_dentry); - goto out; - } - if (d_really_is_positive(lower_dentry)) - goto interpose; + mount_crypt_stat = &ecryptfs_superblock_to_private( ecryptfs_dentry->d_sb)->mount_crypt_stat; - if (!(mount_crypt_stat - && (mount_crypt_stat->flags & ECRYPTFS_GLOBAL_ENCRYPT_FILENAMES))) - goto interpose; - dput(lower_dentry); - rc = ecryptfs_encrypt_and_encode_filename( - &encrypted_and_encoded_name, &encrypted_and_encoded_name_size, - mount_crypt_stat, ecryptfs_dentry->d_name.name, - ecryptfs_dentry->d_name.len); - if (rc) { - printk(KERN_ERR "%s: Error attempting to encrypt and encode " - "filename; rc = [%d]\n", __func__, rc); - res = ERR_PTR(rc); - goto out; + if (mount_crypt_stat + && (mount_crypt_stat->flags & ECRYPTFS_GLOBAL_ENCRYPT_FILENAMES)) { + rc = ecryptfs_encrypt_and_encode_filename( + &encrypted_and_encoded_name, &len, + mount_crypt_stat, name, len); + if (rc) { + printk(KERN_ERR "%s: Error attempting to encrypt and encode " + "filename; rc = [%d]\n", __func__, rc); + return ERR_PTR(rc); + } + name = encrypted_and_encoded_name; } - lower_dentry = lookup_one_len_unlocked(encrypted_and_encoded_name, - lower_dir_dentry, - encrypted_and_encoded_name_size); + + lower_dentry = lookup_one_len_unlocked(name, lower_dir_dentry, len); if (IS_ERR(lower_dentry)) { ecryptfs_printk(KERN_DEBUG, "%s: lookup_one_len() returned " "[%ld] on lower_dentry = [%s]\n", __func__, PTR_ERR(lower_dentry), - encrypted_and_encoded_name); + name); res = ERR_CAST(lower_dentry); - goto out; + } else { + res = ecryptfs_lookup_interpose(ecryptfs_dentry, lower_dentry); } -interpose: - res = ecryptfs_lookup_interpose(ecryptfs_dentry, lower_dentry); -out: kfree(encrypted_and_encoded_name); return res; } -- cgit v1.3-8-gc7d7 From 2958ec177e400be1e26fc37e1759f84fa2c1761c Mon Sep 17 00:00:00 2001 From: Al Viro Date: Thu, 31 Mar 2016 21:34:25 -0400 Subject: aio: remove a pointless assignment the value is never used after that point Signed-off-by: Al Viro --- fs/aio.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'fs') diff --git a/fs/aio.c b/fs/aio.c index 155f84253f33..a6deaa78326d 100644 --- a/fs/aio.c +++ b/fs/aio.c @@ -1447,8 +1447,6 @@ rw_common: return ret; } - len = ret; - if (rw == WRITE) file_start_write(file); -- cgit v1.3-8-gc7d7 From bc61384dcdd82a6faabafecdcd80040625db5e40 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Thu, 31 Mar 2016 21:48:20 -0400 Subject: rw_verify_area(): saner calling conventions Lift length capping into the few callers that care about it. Most of them treat all non-negatives as "success" and ignore the capped value, and with good reasons. Make rw_verify_area() return 0 on success. Signed-off-by: Al Viro --- fs/read_write.c | 33 ++++++++++++++------------------- 1 file changed, 14 insertions(+), 19 deletions(-) (limited to 'fs') diff --git a/fs/read_write.c b/fs/read_write.c index cf377cf9dfe3..b1a0e6ca218f 100644 --- a/fs/read_write.c +++ b/fs/read_write.c @@ -410,11 +410,6 @@ ssize_t vfs_iter_write(struct file *file, struct iov_iter *iter, loff_t *ppos) } EXPORT_SYMBOL(vfs_iter_write); -/* - * rw_verify_area doesn't like huge counts. We limit - * them to something that fits in "int" so that others - * won't have to do range checks all the time. - */ int rw_verify_area(int read_write, struct file *file, const loff_t *ppos, size_t count) { struct inode *inode; @@ -441,11 +436,8 @@ int rw_verify_area(int read_write, struct file *file, const loff_t *ppos, size_t if (retval < 0) return retval; } - retval = security_file_permission(file, + return security_file_permission(file, read_write == READ ? MAY_READ : MAY_WRITE); - if (retval) - return retval; - return count > MAX_RW_COUNT ? MAX_RW_COUNT : count; } static ssize_t new_sync_read(struct file *filp, char __user *buf, size_t len, loff_t *ppos) @@ -489,8 +481,9 @@ ssize_t vfs_read(struct file *file, char __user *buf, size_t count, loff_t *pos) return -EFAULT; ret = rw_verify_area(READ, file, pos, count); - if (ret >= 0) { - count = ret; + if (!ret) { + if (count > MAX_RW_COUNT) + count = MAX_RW_COUNT; ret = __vfs_read(file, buf, count, pos); if (ret > 0) { fsnotify_access(file); @@ -572,8 +565,9 @@ ssize_t vfs_write(struct file *file, const char __user *buf, size_t count, loff_ return -EFAULT; ret = rw_verify_area(WRITE, file, pos, count); - if (ret >= 0) { - count = ret; + if (!ret) { + if (count > MAX_RW_COUNT) + count = MAX_RW_COUNT; file_start_write(file); ret = __vfs_write(file, buf, count, pos); if (ret > 0) { @@ -1323,7 +1317,8 @@ static ssize_t do_sendfile(int out_fd, int in_fd, loff_t *ppos, retval = rw_verify_area(READ, in.file, &pos, count); if (retval < 0) goto fput_in; - count = retval; + if (count > MAX_RW_COUNT) + count = MAX_RW_COUNT; /* * Get output file, and verify that it is ok.. @@ -1341,7 +1336,6 @@ static ssize_t do_sendfile(int out_fd, int in_fd, loff_t *ppos, retval = rw_verify_area(WRITE, out.file, &out_pos, count); if (retval < 0) goto fput_out; - count = retval; if (!max) max = min(in_inode->i_sb->s_maxbytes, out_inode->i_sb->s_maxbytes); @@ -1485,11 +1479,12 @@ ssize_t vfs_copy_file_range(struct file *file_in, loff_t pos_in, if (flags != 0) return -EINVAL; - /* copy_file_range allows full ssize_t len, ignoring MAX_RW_COUNT */ ret = rw_verify_area(READ, file_in, &pos_in, len); - if (ret >= 0) - ret = rw_verify_area(WRITE, file_out, &pos_out, len); - if (ret < 0) + if (unlikely(ret)) + return ret; + + ret = rw_verify_area(WRITE, file_out, &pos_out, len); + if (unlikely(ret)) return ret; if (!(file_in->f_mode & FMODE_READ) || -- cgit v1.3-8-gc7d7 From 73cc86252bd8a547349d2856497eceaf4f2d1766 Mon Sep 17 00:00:00 2001 From: Bob Peterson Date: Fri, 1 Apr 2016 13:18:38 -0400 Subject: GFS2: Get rid of dead code in inode_go_demote_ok Function inode_go_demote_ok had some code that was only executed if gl_holders was not empty. However, if gl_holders was not empty, the only caller, demote_ok(), returns before inode_go_demote_ok would ever be called. Therefore, it's dead code, so I removed it. Signed-off-by: Bob Peterson Acked-by: Steven Whitehouse --- fs/gfs2/glops.c | 7 ------- 1 file changed, 7 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/glops.c b/fs/gfs2/glops.c index 437fd73e381e..5db59d444838 100644 --- a/fs/gfs2/glops.c +++ b/fs/gfs2/glops.c @@ -286,17 +286,10 @@ static void inode_go_inval(struct gfs2_glock *gl, int flags) static int inode_go_demote_ok(const struct gfs2_glock *gl) { struct gfs2_sbd *sdp = gl->gl_name.ln_sbd; - struct gfs2_holder *gh; if (sdp->sd_jindex == gl->gl_object || sdp->sd_rindex == gl->gl_object) return 0; - if (!list_empty(&gl->gl_holders)) { - gh = list_entry(gl->gl_holders.next, struct gfs2_holder, gh_list); - if (gh->gh_list.next != &gl->gl_holders) - return 0; - } - return 1; } -- cgit v1.3-8-gc7d7 From 611526756a3d0fa7f4d1c519397ba8898f40443f Mon Sep 17 00:00:00 2001 From: Abhi Das Date: Tue, 5 Apr 2016 12:06:15 -0400 Subject: gfs2: Use gfs2 wrapper to sync inode before calling generic_file_splice_read() gfs2_file_splice_read() f_op grabs and releases the cluster-wide inode glock to sync the inode size to the latest. Without this, generic_file_splice_read() uses an older i_size value and can return EOF for valid offsets in the inode. Signed-off-by: Abhi Das Signed-off-by: Bob Peterson --- fs/gfs2/file.c | 28 ++++++++++++++++++++++++++-- 1 file changed, 26 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/file.c b/fs/gfs2/file.c index c9384f932975..fd9a10e4518d 100644 --- a/fs/gfs2/file.c +++ b/fs/gfs2/file.c @@ -950,6 +950,30 @@ out_uninit: return ret; } +static ssize_t gfs2_file_splice_read(struct file *in, loff_t *ppos, + struct pipe_inode_info *pipe, size_t len, + unsigned int flags) +{ + struct inode *inode = in->f_mapping->host; + struct gfs2_inode *ip = GFS2_I(inode); + struct gfs2_holder gh; + int ret; + + mutex_lock(&inode->i_mutex); + + ret = gfs2_glock_nq_init(ip->i_gl, LM_ST_SHARED, 0, &gh); + if (ret) { + mutex_unlock(&inode->i_mutex); + return ret; + } + + gfs2_glock_dq_uninit(&gh); + mutex_unlock(&inode->i_mutex); + + return generic_file_splice_read(in, ppos, pipe, len, flags); +} + + static ssize_t gfs2_file_splice_write(struct pipe_inode_info *pipe, struct file *out, loff_t *ppos, size_t len, unsigned int flags) @@ -1112,7 +1136,7 @@ const struct file_operations gfs2_file_fops = { .fsync = gfs2_fsync, .lock = gfs2_lock, .flock = gfs2_flock, - .splice_read = generic_file_splice_read, + .splice_read = gfs2_file_splice_read, .splice_write = gfs2_file_splice_write, .setlease = simple_nosetlease, .fallocate = gfs2_fallocate, @@ -1140,7 +1164,7 @@ const struct file_operations gfs2_file_fops_nolock = { .open = gfs2_open, .release = gfs2_release, .fsync = gfs2_fsync, - .splice_read = generic_file_splice_read, + .splice_read = gfs2_file_splice_read, .splice_write = gfs2_file_splice_write, .setlease = generic_setlease, .fallocate = gfs2_fallocate, -- cgit v1.3-8-gc7d7 From d0a58e833931234c44e515b5b8bede32bd4e6eed Mon Sep 17 00:00:00 2001 From: Eric Sandeen Date: Wed, 6 Apr 2016 07:05:41 +1000 Subject: xfs: disallow rw remount on fs with unknown ro-compat features Today, a kernel which refuses to mount a filesystem read-write due to unknown ro-compat features can still transition to read-write via the remount path. The old kernel is most likely none the wiser, because it's unaware of the new feature, and isn't using it. However, writing to the filesystem may well corrupt metadata related to that new feature, and moving to a newer kernel which understand the feature will have problems. Right now the only ro-compat feature we have is the free inode btree, which showed up in v3.16. It would be good to push this back to all the active stable kernels, I think, so that if anyone is using newer mkfs (which enables the finobt feature) with older kernel releases, they'll be protected. Cc: # 3.10.x- Signed-off-by: Eric Sandeen Reviewed-by: Bill O'Donnell Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner --- fs/xfs/xfs_super.c | 10 ++++++++++ 1 file changed, 10 insertions(+) (limited to 'fs') diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index d760934109b5..ca058a153c15 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -1276,6 +1276,16 @@ xfs_fs_remount( return -EINVAL; } + if (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_5 && + xfs_sb_has_ro_compat_feature(sbp, + XFS_SB_FEAT_RO_COMPAT_UNKNOWN)) { + xfs_warn(mp, +"ro->rw transition prohibited on unknown (0x%x) ro-compat filesystem", + (sbp->sb_features_ro_compat & + XFS_SB_FEAT_RO_COMPAT_UNKNOWN)); + return -EINVAL; + } + mp->m_flags &= ~XFS_MOUNT_RDONLY; /* -- cgit v1.3-8-gc7d7 From ad747e3b299671e1a53db74963cc6c5f6cdb9f6d Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Wed, 6 Apr 2016 07:06:20 +1000 Subject: xfs: Don't wrap growfs AGFL indexes Commit 96f859d ("libxfs: pack the agfl header structure so XFS_AGFL_SIZE is correct") allowed the freelist to use the empty slot at the end of the freelist on 64 bit systems that was not being used due to sizeof() rounding up the structure size. This has caused versions of xfs_repair prior to 4.5.0 (which also has the fix) to report this as a corruption once the filesystem has been grown. Older kernels can also have problems (seen from a whacky container/vm management environment) mounting filesystems grown on a system with a newer kernel than the vm/container it is deployed on. To avoid this problem, change the initial free list indexes not to wrap across the end of the AGFL, hence avoiding the initialisation of agf_fllast to the last index in the AGFL. cc: # 4.4-4.5 Signed-off-by: Dave Chinner Reviewed-by: Carlos Maiolino Signed-off-by: Dave Chinner --- fs/xfs/xfs_fsops.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_fsops.c b/fs/xfs/xfs_fsops.c index ee3aaa0a5317..ca0d3eb44925 100644 --- a/fs/xfs/xfs_fsops.c +++ b/fs/xfs/xfs_fsops.c @@ -243,8 +243,8 @@ xfs_growfs_data_private( agf->agf_roots[XFS_BTNUM_CNTi] = cpu_to_be32(XFS_CNT_BLOCK(mp)); agf->agf_levels[XFS_BTNUM_BNOi] = cpu_to_be32(1); agf->agf_levels[XFS_BTNUM_CNTi] = cpu_to_be32(1); - agf->agf_flfirst = 0; - agf->agf_fllast = cpu_to_be32(XFS_AGFL_SIZE(mp) - 1); + agf->agf_flfirst = cpu_to_be32(1); + agf->agf_fllast = 0; agf->agf_flcount = 0; tmpsize = agsize - XFS_PREALLOC_BLOCKS(mp); agf->agf_freeblks = cpu_to_be32(tmpsize); -- cgit v1.3-8-gc7d7 From ce5c767db079649db88a9f189798839f9c544981 Mon Sep 17 00:00:00 2001 From: Eryu Guan Date: Wed, 6 Apr 2016 07:19:40 +1000 Subject: xfs: add missing break in xfs_parseargs() Commit 2e74af0e1189 ("xfs: convert mount option parsing to tokens") missed a 'break;' in xfs_parseargs() which causes mount to fail with "-o pqnoenforce" option when mounting a v4 filesystem. xfs/050 catches this failure: XFS (vda6): Super block does not support project and group quota together Fixes: 2e74af0e1189 ("xfs: convert mount option parsing to tokens") Signed-off-by: Eryu Guan Reviewed-by: Brian Foster Reviewed-by: Eric Sandeen Signed-off-by: Dave Chinner --- fs/xfs/xfs_super.c | 1 + 1 file changed, 1 insertion(+) (limited to 'fs') diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index ca058a153c15..a9ea12d46b74 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -350,6 +350,7 @@ xfs_parseargs( case Opt_pqnoenforce: mp->m_qflags |= (XFS_PQUOTA_ACCT | XFS_PQUOTA_ACTIVE); mp->m_qflags &= ~XFS_PQUOTA_ENFD; + break; case Opt_gquota: case Opt_grpquota: mp->m_qflags |= (XFS_GQUOTA_ACCT | XFS_GQUOTA_ACTIVE | -- cgit v1.3-8-gc7d7 From 143f4aede7fb25b9198b15660d6f9830936394a8 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 6 Apr 2016 07:41:43 +1000 Subject: xfs: factor out a helper to initialize a local format inode fork Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_dir2_sf.c | 9 +++----- fs/xfs/libxfs/xfs_inode_fork.c | 48 ++++++++++++++++++++++++++---------------- fs/xfs/libxfs/xfs_inode_fork.h | 1 + fs/xfs/xfs_symlink.c | 12 ++--------- 4 files changed, 36 insertions(+), 34 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_dir2_sf.c b/fs/xfs/libxfs/xfs_dir2_sf.c index 974d62e677f4..e5bb9cc3b243 100644 --- a/fs/xfs/libxfs/xfs_dir2_sf.c +++ b/fs/xfs/libxfs/xfs_dir2_sf.c @@ -257,15 +257,12 @@ xfs_dir2_block_to_sf( * * Convert the inode to local format and copy the data in. */ - dp->i_df.if_flags &= ~XFS_IFEXTENTS; - dp->i_df.if_flags |= XFS_IFINLINE; - dp->i_d.di_format = XFS_DINODE_FMT_LOCAL; ASSERT(dp->i_df.if_bytes == 0); - xfs_idata_realloc(dp, size, XFS_DATA_FORK); + xfs_init_local_fork(dp, XFS_DATA_FORK, dst, size); + dp->i_d.di_format = XFS_DINODE_FMT_LOCAL; + dp->i_d.di_size = size; logflags |= XFS_ILOG_DDATA; - memcpy(dp->i_df.if_u1.if_data, dst, size); - dp->i_d.di_size = size; xfs_dir2_sf_check(args); out: xfs_trans_log_inode(args->trans, dp, logflags); diff --git a/fs/xfs/libxfs/xfs_inode_fork.c b/fs/xfs/libxfs/xfs_inode_fork.c index 11faf7df14c8..86a97f8a9de3 100644 --- a/fs/xfs/libxfs/xfs_inode_fork.c +++ b/fs/xfs/libxfs/xfs_inode_fork.c @@ -231,6 +231,34 @@ xfs_iformat_fork( return error; } +void +xfs_init_local_fork( + struct xfs_inode *ip, + int whichfork, + const void *data, + int size) +{ + struct xfs_ifork *ifp = XFS_IFORK_PTR(ip, whichfork); + int real_size = 0; + + if (size == 0) + ifp->if_u1.if_data = NULL; + else if (size <= sizeof(ifp->if_u2.if_inline_data)) + ifp->if_u1.if_data = ifp->if_u2.if_inline_data; + else { + real_size = roundup(size, 4); + ifp->if_u1.if_data = kmem_alloc(real_size, KM_SLEEP | KM_NOFS); + } + + if (size) + memcpy(ifp->if_u1.if_data, data, size); + + ifp->if_bytes = size; + ifp->if_real_bytes = real_size; + ifp->if_flags &= ~(XFS_IFEXTENTS | XFS_IFBROOT); + ifp->if_flags |= XFS_IFINLINE; +} + /* * The file is in-lined in the on-disk inode. * If it fits into if_inline_data, then copy @@ -248,8 +276,6 @@ xfs_iformat_local( int whichfork, int size) { - xfs_ifork_t *ifp; - int real_size; /* * If the size is unreasonable, then something @@ -265,22 +291,8 @@ xfs_iformat_local( ip->i_mount, dip); return -EFSCORRUPTED; } - ifp = XFS_IFORK_PTR(ip, whichfork); - real_size = 0; - if (size == 0) - ifp->if_u1.if_data = NULL; - else if (size <= sizeof(ifp->if_u2.if_inline_data)) - ifp->if_u1.if_data = ifp->if_u2.if_inline_data; - else { - real_size = roundup(size, 4); - ifp->if_u1.if_data = kmem_alloc(real_size, KM_SLEEP | KM_NOFS); - } - ifp->if_bytes = size; - ifp->if_real_bytes = real_size; - if (size) - memcpy(ifp->if_u1.if_data, XFS_DFORK_PTR(dip, whichfork), size); - ifp->if_flags &= ~XFS_IFEXTENTS; - ifp->if_flags |= XFS_IFINLINE; + + xfs_init_local_fork(ip, whichfork, XFS_DFORK_PTR(dip, whichfork), size); return 0; } diff --git a/fs/xfs/libxfs/xfs_inode_fork.h b/fs/xfs/libxfs/xfs_inode_fork.h index 7d3b1ed6dcbe..f95e072ae646 100644 --- a/fs/xfs/libxfs/xfs_inode_fork.h +++ b/fs/xfs/libxfs/xfs_inode_fork.h @@ -134,6 +134,7 @@ void xfs_iroot_realloc(struct xfs_inode *, int, int); int xfs_iread_extents(struct xfs_trans *, struct xfs_inode *, int); int xfs_iextents_copy(struct xfs_inode *, struct xfs_bmbt_rec *, int); +void xfs_init_local_fork(struct xfs_inode *, int, const void *, int); struct xfs_bmbt_rec_host * xfs_iext_get_ext(struct xfs_ifork *, xfs_extnum_t); diff --git a/fs/xfs/xfs_symlink.c b/fs/xfs/xfs_symlink.c index b44284c1adda..b69f4a770fc9 100644 --- a/fs/xfs/xfs_symlink.c +++ b/fs/xfs/xfs_symlink.c @@ -302,19 +302,11 @@ xfs_symlink( * If the symlink will fit into the inode, write it inline. */ if (pathlen <= XFS_IFORK_DSIZE(ip)) { - xfs_idata_realloc(ip, pathlen, XFS_DATA_FORK); - memcpy(ip->i_df.if_u1.if_data, target_path, pathlen); - ip->i_d.di_size = pathlen; - - /* - * The inode was initially created in extent format. - */ - ip->i_df.if_flags &= ~(XFS_IFEXTENTS | XFS_IFBROOT); - ip->i_df.if_flags |= XFS_IFINLINE; + xfs_init_local_fork(ip, XFS_DATA_FORK, target_path, pathlen); + ip->i_d.di_size = pathlen; ip->i_d.di_format = XFS_DINODE_FMT_LOCAL; xfs_trans_log_inode(tp, ip, XFS_ILOG_DDATA | XFS_ILOG_CORE); - } else { int offset; -- cgit v1.3-8-gc7d7 From 2b3d1d41b4d96c3b074096ae57b27cd191969643 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 6 Apr 2016 07:48:27 +1000 Subject: xfs: set up inode operation vectors later In the next patch we'll set up different inode operations for inline vs out of line symlinks, for that we need to make sure the flags are already set up properly. [dchinner: added xfs_setup_iops() call to xfs_rename_alloc_whiteout()] Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner --- fs/xfs/xfs_inode.c | 1 + fs/xfs/xfs_inode.h | 5 ++++- fs/xfs/xfs_iops.c | 59 ++++++++++++++++++++++++++++++++++-------------------- 3 files changed, 42 insertions(+), 23 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index 96f606deee31..4fcd6a722213 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -2855,6 +2855,7 @@ xfs_rename_alloc_whiteout( * and flag it as linkable. */ drop_nlink(VFS_I(tmpfile)); + xfs_setup_iops(tmpfile); xfs_finish_inode_setup(tmpfile); VFS_I(tmpfile)->i_state |= I_LINKABLE; diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h index 43e1d51b15eb..e52d7c7aeb5b 100644 --- a/fs/xfs/xfs_inode.h +++ b/fs/xfs/xfs_inode.h @@ -440,6 +440,9 @@ loff_t __xfs_seek_hole_data(struct inode *inode, loff_t start, /* from xfs_iops.c */ +extern void xfs_setup_inode(struct xfs_inode *ip); +extern void xfs_setup_iops(struct xfs_inode *ip); + /* * When setting up a newly allocated inode, we need to call * xfs_finish_inode_setup() once the inode is fully instantiated at @@ -447,7 +450,6 @@ loff_t __xfs_seek_hole_data(struct inode *inode, loff_t start, * before we've completed instantiation. Otherwise we can do it * the moment the inode lookup is complete. */ -extern void xfs_setup_inode(struct xfs_inode *ip); static inline void xfs_finish_inode_setup(struct xfs_inode *ip) { xfs_iflags_clear(ip, XFS_INEW); @@ -458,6 +460,7 @@ static inline void xfs_finish_inode_setup(struct xfs_inode *ip) static inline void xfs_setup_existing_inode(struct xfs_inode *ip) { xfs_setup_inode(ip); + xfs_setup_iops(ip); xfs_finish_inode_setup(ip); } diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c index fb7dc61f4a29..f08d91c51b7f 100644 --- a/fs/xfs/xfs_iops.c +++ b/fs/xfs/xfs_iops.c @@ -181,6 +181,8 @@ xfs_generic_create( } #endif + xfs_setup_iops(ip); + if (tmpfile) d_tmpfile(dentry, inode); else @@ -368,6 +370,8 @@ xfs_vn_symlink( if (unlikely(error)) goto out_cleanup_inode; + xfs_setup_iops(cip); + d_instantiate(dentry, inode); xfs_finish_inode_setup(cip); return 0; @@ -1193,7 +1197,7 @@ xfs_diflags_to_iflags( } /* - * Initialize the Linux inode and set up the operation vectors. + * Initialize the Linux inode. * * When reading existing inodes from disk this is called directly from xfs_iget, * when creating a new inode it is called from xfs_ialloc after setting up the @@ -1232,8 +1236,38 @@ xfs_setup_inode( i_size_write(inode, ip->i_d.di_size); xfs_diflags_to_iflags(inode, ip); - ip->d_ops = ip->i_mount->m_nondir_inode_ops; - lockdep_set_class(&ip->i_lock.mr_lock, &xfs_nondir_ilock_class); + if (S_ISDIR(inode->i_mode)) { + lockdep_set_class(&ip->i_lock.mr_lock, &xfs_dir_ilock_class); + ip->d_ops = ip->i_mount->m_dir_inode_ops; + } else { + ip->d_ops = ip->i_mount->m_nondir_inode_ops; + lockdep_set_class(&ip->i_lock.mr_lock, &xfs_nondir_ilock_class); + } + + /* + * Ensure all page cache allocations are done from GFP_NOFS context to + * prevent direct reclaim recursion back into the filesystem and blowing + * stacks or deadlocking. + */ + gfp_mask = mapping_gfp_mask(inode->i_mapping); + mapping_set_gfp_mask(inode->i_mapping, (gfp_mask & ~(__GFP_FS))); + + /* + * If there is no attribute fork no ACL can exist on this inode, + * and it can't have any file capabilities attached to it either. + */ + if (!XFS_IFORK_Q(ip)) { + inode_has_no_xattr(inode); + cache_no_acl(inode); + } +} + +void +xfs_setup_iops( + struct xfs_inode *ip) +{ + struct inode *inode = &ip->i_vnode; + switch (inode->i_mode & S_IFMT) { case S_IFREG: inode->i_op = &xfs_inode_operations; @@ -1241,13 +1275,11 @@ xfs_setup_inode( inode->i_mapping->a_ops = &xfs_address_space_operations; break; case S_IFDIR: - lockdep_set_class(&ip->i_lock.mr_lock, &xfs_dir_ilock_class); if (xfs_sb_version_hasasciici(&XFS_M(inode->i_sb)->m_sb)) inode->i_op = &xfs_dir_ci_inode_operations; else inode->i_op = &xfs_dir_inode_operations; inode->i_fop = &xfs_dir_file_operations; - ip->d_ops = ip->i_mount->m_dir_inode_ops; break; case S_IFLNK: inode->i_op = &xfs_symlink_inode_operations; @@ -1259,21 +1291,4 @@ xfs_setup_inode( init_special_inode(inode, inode->i_mode, inode->i_rdev); break; } - - /* - * Ensure all page cache allocations are done from GFP_NOFS context to - * prevent direct reclaim recursion back into the filesystem and blowing - * stacks or deadlocking. - */ - gfp_mask = mapping_gfp_mask(inode->i_mapping); - mapping_set_gfp_mask(inode->i_mapping, (gfp_mask & ~(__GFP_FS))); - - /* - * If there is no attribute fork no ACL can exist on this inode, - * and it can't have any file capabilities attached to it either. - */ - if (!XFS_IFORK_Q(ip)) { - inode_has_no_xattr(inode); - cache_no_acl(inode); - } } -- cgit v1.3-8-gc7d7 From bfe8804d908a791b16e3686c101f0d7eca9fb5b9 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 6 Apr 2016 07:50:54 +1000 Subject: xfs: use ->readlink to implement the readlink_by_handle ioctl Also drop the now unused readlink_copy export. [dchinner: use d_inode(dentry) rather than dentry->d_inode] Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner --- fs/namei.c | 1 - fs/xfs/xfs_ioctl.c | 18 ++---------------- 2 files changed, 2 insertions(+), 17 deletions(-) (limited to 'fs') diff --git a/fs/namei.c b/fs/namei.c index 794f81dce766..cdd041985929 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -4515,7 +4515,6 @@ int readlink_copy(char __user *buffer, int buflen, const char *link) out: return len; } -EXPORT_SYMBOL(readlink_copy); /* * A helper for ->readlink(). This should be used *ONLY* for symlinks that diff --git a/fs/xfs/xfs_ioctl.c b/fs/xfs/xfs_ioctl.c index bcb6c19ce3ea..5414bcaf2e73 100644 --- a/fs/xfs/xfs_ioctl.c +++ b/fs/xfs/xfs_ioctl.c @@ -277,7 +277,6 @@ xfs_readlink_by_handle( { struct dentry *dentry; __u32 olen; - void *link; int error; if (!capable(CAP_SYS_ADMIN)) @@ -288,7 +287,7 @@ xfs_readlink_by_handle( return PTR_ERR(dentry); /* Restrict this handle operation to symlinks only. */ - if (!d_is_symlink(dentry)) { + if (!d_inode(dentry)->i_op->readlink) { error = -EINVAL; goto out_dput; } @@ -298,21 +297,8 @@ xfs_readlink_by_handle( goto out_dput; } - link = kmalloc(MAXPATHLEN+1, GFP_KERNEL); - if (!link) { - error = -ENOMEM; - goto out_dput; - } - - error = xfs_readlink(XFS_I(d_inode(dentry)), link); - if (error) - goto out_kfree; - error = readlink_copy(hreq->ohandle, olen, link); - if (error) - goto out_kfree; + error = d_inode(dentry)->i_op->readlink(dentry, hreq->ohandle, olen); - out_kfree: - kfree(link); out_dput: dput(dentry); return error; -- cgit v1.3-8-gc7d7 From 30ee052e12b97c190b27fe6f20e3ac3047df7b5c Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 6 Apr 2016 07:53:29 +1000 Subject: xfs: optimize inline symlinks By overallocating the in-core inode fork data buffer and zero terminating the link target in xfs_init_local_fork we can avoid the memory allocation in ->follow_link. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_inode_fork.c | 22 ++++++++++++++++++---- fs/xfs/xfs_inode_item.c | 4 ++-- fs/xfs/xfs_iops.c | 29 ++++++++++++++++++++++++++--- fs/xfs/xfs_symlink.c | 9 +++------ 4 files changed, 49 insertions(+), 15 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_inode_fork.c b/fs/xfs/libxfs/xfs_inode_fork.c index 86a97f8a9de3..4fbe2263c1fc 100644 --- a/fs/xfs/libxfs/xfs_inode_fork.c +++ b/fs/xfs/libxfs/xfs_inode_fork.c @@ -239,19 +239,33 @@ xfs_init_local_fork( int size) { struct xfs_ifork *ifp = XFS_IFORK_PTR(ip, whichfork); - int real_size = 0; + int mem_size = size, real_size = 0; + bool zero_terminate; + + /* + * If we are using the local fork to store a symlink body we need to + * zero-terminate it so that we can pass it back to the VFS directly. + * Overallocate the in-memory fork by one for that and add a zero + * to terminate it below. + */ + zero_terminate = S_ISLNK(VFS_I(ip)->i_mode); + if (zero_terminate) + mem_size++; if (size == 0) ifp->if_u1.if_data = NULL; - else if (size <= sizeof(ifp->if_u2.if_inline_data)) + else if (mem_size <= sizeof(ifp->if_u2.if_inline_data)) ifp->if_u1.if_data = ifp->if_u2.if_inline_data; else { - real_size = roundup(size, 4); + real_size = roundup(mem_size, 4); ifp->if_u1.if_data = kmem_alloc(real_size, KM_SLEEP | KM_NOFS); } - if (size) + if (size) { memcpy(ifp->if_u1.if_data, data, size); + if (zero_terminate) + ifp->if_u1.if_data[size] = '\0'; + } ifp->if_bytes = size; ifp->if_real_bytes = real_size; diff --git a/fs/xfs/xfs_inode_item.c b/fs/xfs/xfs_inode_item.c index c48b5b18d771..37e23c7a5684 100644 --- a/fs/xfs/xfs_inode_item.c +++ b/fs/xfs/xfs_inode_item.c @@ -210,7 +210,7 @@ xfs_inode_item_format_data_fork( */ data_bytes = roundup(ip->i_df.if_bytes, 4); ASSERT(ip->i_df.if_real_bytes == 0 || - ip->i_df.if_real_bytes == data_bytes); + ip->i_df.if_real_bytes >= data_bytes); ASSERT(ip->i_df.if_u1.if_data != NULL); ASSERT(ip->i_d.di_size > 0); xlog_copy_iovec(lv, vecp, XLOG_REG_TYPE_ILOCAL, @@ -305,7 +305,7 @@ xfs_inode_item_format_attr_fork( */ data_bytes = roundup(ip->i_afp->if_bytes, 4); ASSERT(ip->i_afp->if_real_bytes == 0 || - ip->i_afp->if_real_bytes == data_bytes); + ip->i_afp->if_real_bytes >= data_bytes); ASSERT(ip->i_afp->if_u1.if_data != NULL); xlog_copy_iovec(lv, vecp, XLOG_REG_TYPE_IATTR_LOCAL, ip->i_afp->if_u1.if_data, diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c index f08d91c51b7f..aee06d9a7b6c 100644 --- a/fs/xfs/xfs_iops.c +++ b/fs/xfs/xfs_iops.c @@ -446,6 +446,16 @@ xfs_vn_get_link( return ERR_PTR(error); } +STATIC const char * +xfs_vn_get_link_inline( + struct dentry *dentry, + struct inode *inode, + struct delayed_call *done) +{ + ASSERT(XFS_I(inode)->i_df.if_flags & XFS_IFINLINE); + return XFS_I(inode)->i_df.if_u1.if_data; +} + STATIC int xfs_vn_getattr( struct vfsmount *mnt, @@ -1171,6 +1181,18 @@ static const struct inode_operations xfs_symlink_inode_operations = { .update_time = xfs_vn_update_time, }; +static const struct inode_operations xfs_inline_symlink_inode_operations = { + .readlink = generic_readlink, + .get_link = xfs_vn_get_link_inline, + .getattr = xfs_vn_getattr, + .setattr = xfs_vn_setattr, + .setxattr = generic_setxattr, + .getxattr = generic_getxattr, + .removexattr = generic_removexattr, + .listxattr = xfs_vn_listxattr, + .update_time = xfs_vn_update_time, +}; + STATIC void xfs_diflags_to_iflags( struct inode *inode, @@ -1282,9 +1304,10 @@ xfs_setup_iops( inode->i_fop = &xfs_dir_file_operations; break; case S_IFLNK: - inode->i_op = &xfs_symlink_inode_operations; - if (!(ip->i_df.if_flags & XFS_IFINLINE)) - inode->i_mapping->a_ops = &xfs_address_space_operations; + if (ip->i_df.if_flags & XFS_IFINLINE) + inode->i_op = &xfs_inline_symlink_inode_operations; + else + inode->i_op = &xfs_symlink_inode_operations; break; default: inode->i_op = &xfs_inode_operations; diff --git a/fs/xfs/xfs_symlink.c b/fs/xfs/xfs_symlink.c index b69f4a770fc9..5961c1e880c2 100644 --- a/fs/xfs/xfs_symlink.c +++ b/fs/xfs/xfs_symlink.c @@ -131,6 +131,8 @@ xfs_readlink( trace_xfs_readlink(ip); + ASSERT(!(ip->i_df.if_flags & XFS_IFINLINE)); + if (XFS_FORCED_SHUTDOWN(mp)) return -EIO; @@ -150,12 +152,7 @@ xfs_readlink( } - if (ip->i_df.if_flags & XFS_IFINLINE) { - memcpy(link, ip->i_df.if_u1.if_data, pathlen); - link[pathlen] = '\0'; - } else { - error = xfs_readlink_bmap(ip, link); - } + error = xfs_readlink_bmap(ip, link); out: xfs_iunlock(ip, XFS_ILOCK_SHARED); -- cgit v1.3-8-gc7d7 From 2a6fba6d2311151598abaa1e7c9abd5f8d024a43 Mon Sep 17 00:00:00 2001 From: Eric Sandeen Date: Wed, 6 Apr 2016 07:57:18 +1000 Subject: xfs: only return -errno or success from attr ->put_listent Today, the put_listent formatters return either 1 or 0; if they return 1, some callers treat this as an error and return it up the stack, despite "1" not being a valid (negative) error code. The intent seems to be that if the input buffer is full, we set seen_enough or set count = -1, and return 1; but some callers check the return before checking the seen_enough or count fields of the context. Fix this by only returning non-zero for actual errors encountered, and rely on the caller to first check the return value, then check the values in the context to decide what to do. Signed-off-by: Eric Sandeen Reviewed-by: Christoph Hellwig Signed-off-by: Dave Chinner --- fs/xfs/xfs_attr.h | 1 + fs/xfs/xfs_attr_list.c | 8 +++----- fs/xfs/xfs_xattr.c | 14 ++++++++++---- 3 files changed, 14 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_attr.h b/fs/xfs/xfs_attr.h index dd4824589470..234331227c0c 100644 --- a/fs/xfs/xfs_attr.h +++ b/fs/xfs/xfs_attr.h @@ -112,6 +112,7 @@ typedef struct attrlist_cursor_kern { *========================================================================*/ +/* Return 0 on success, or -errno; other state communicated via *context */ typedef int (*put_listent_func_t)(struct xfs_attr_list_context *, int, unsigned char *, int, int, unsigned char *); diff --git a/fs/xfs/xfs_attr_list.c b/fs/xfs/xfs_attr_list.c index 4fa14820e2e2..c8be331a3196 100644 --- a/fs/xfs/xfs_attr_list.c +++ b/fs/xfs/xfs_attr_list.c @@ -108,16 +108,14 @@ xfs_attr_shortform_list(xfs_attr_list_context_t *context) (int)sfe->namelen, (int)sfe->valuelen, &sfe->nameval[sfe->namelen]); - + if (error) + return error; /* * Either search callback finished early or * didn't fit it all in the buffer after all. */ if (context->seen_enough) break; - - if (error) - return error; sfe = XFS_ATTR_SF_NEXTENTRY(sfe); } trace_xfs_attr_list_sf_all(context); @@ -581,7 +579,7 @@ xfs_attr_put_listent( trace_xfs_attr_list_full(context); alist->al_more = 1; context->seen_enough = 1; - return 1; + return 0; } aep = (attrlist_ent_t *)&context->alist[context->firstu]; diff --git a/fs/xfs/xfs_xattr.c b/fs/xfs/xfs_xattr.c index 110f1d7d86b0..f2201299311f 100644 --- a/fs/xfs/xfs_xattr.c +++ b/fs/xfs/xfs_xattr.c @@ -146,7 +146,7 @@ __xfs_xattr_put_listent( arraytop = context->count + prefix_len + namelen + 1; if (arraytop > context->firstu) { context->count = -1; /* insufficient space */ - return 1; + return 0; } offset = (char *)context->alist + context->count; strncpy(offset, prefix, prefix_len); @@ -221,11 +221,15 @@ xfs_xattr_put_listent( } ssize_t -xfs_vn_listxattr(struct dentry *dentry, char *data, size_t size) +xfs_vn_listxattr( + struct dentry *dentry, + char *data, + size_t size) { struct xfs_attr_list_context context; struct attrlist_cursor_kern cursor = { 0 }; - struct inode *inode = d_inode(dentry); + struct inode *inode = d_inode(dentry); + int error; /* * First read the regular on-disk attributes. @@ -239,7 +243,9 @@ xfs_vn_listxattr(struct dentry *dentry, char *data, size_t size) context.firstu = context.bufsize; context.put_listent = xfs_xattr_put_listent; - xfs_attr_list_int(&context); + error = xfs_attr_list_int(&context); + if (error) + return error; if (context.count < 0) return -ERANGE; -- cgit v1.3-8-gc7d7 From e5bd12bfea60af455f4cbad494e4ac1082e3abd6 Mon Sep 17 00:00:00 2001 From: Eric Sandeen Date: Wed, 6 Apr 2016 07:57:32 +1000 Subject: xfs: don't pass value into attr ->put_listent The value is not used; only names and value lengths are returned. Remove the argument. Signed-off-by: Eric Sandeen Reviewed-by: Christoph Hellwig Signed-off-by: Dave Chinner --- fs/xfs/xfs_attr.h | 2 +- fs/xfs/xfs_attr_list.c | 18 ++++++------------ fs/xfs/xfs_xattr.c | 3 +-- 3 files changed, 8 insertions(+), 15 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_attr.h b/fs/xfs/xfs_attr.h index 234331227c0c..dab4f41de278 100644 --- a/fs/xfs/xfs_attr.h +++ b/fs/xfs/xfs_attr.h @@ -114,7 +114,7 @@ typedef struct attrlist_cursor_kern { /* Return 0 on success, or -errno; other state communicated via *context */ typedef int (*put_listent_func_t)(struct xfs_attr_list_context *, int, - unsigned char *, int, int, unsigned char *); + unsigned char *, int, int); typedef struct xfs_attr_list_context { struct xfs_inode *dp; /* inode */ diff --git a/fs/xfs/xfs_attr_list.c b/fs/xfs/xfs_attr_list.c index c8be331a3196..d30dbfae05fe 100644 --- a/fs/xfs/xfs_attr_list.c +++ b/fs/xfs/xfs_attr_list.c @@ -106,8 +106,7 @@ xfs_attr_shortform_list(xfs_attr_list_context_t *context) sfe->flags, sfe->nameval, (int)sfe->namelen, - (int)sfe->valuelen, - &sfe->nameval[sfe->namelen]); + (int)sfe->valuelen); if (error) return error; /* @@ -198,8 +197,7 @@ xfs_attr_shortform_list(xfs_attr_list_context_t *context) sbp->flags, sbp->name, sbp->namelen, - sbp->valuelen, - &sbp->name[sbp->namelen]); + sbp->valuelen); if (error) { kmem_free(sbuf); return error; @@ -430,8 +428,7 @@ xfs_attr3_leaf_list_int( entry->flags, name_loc->nameval, (int)name_loc->namelen, - be16_to_cpu(name_loc->valuelen), - &name_loc->nameval[name_loc->namelen]); + be16_to_cpu(name_loc->valuelen)); if (retval) return retval; } else { @@ -459,16 +456,14 @@ xfs_attr3_leaf_list_int( entry->flags, name_rmt->name, (int)name_rmt->namelen, - valuelen, - args.value); + valuelen); kmem_free(args.value); } else { retval = context->put_listent(context, entry->flags, name_rmt->name, (int)name_rmt->namelen, - valuelen, - NULL); + valuelen); } if (retval) return retval; @@ -549,8 +544,7 @@ xfs_attr_put_listent( int flags, unsigned char *name, int namelen, - int valuelen, - unsigned char *value) + int valuelen) { struct attrlist *alist = (struct attrlist *)context->alist; attrlist_ent_t *aep; diff --git a/fs/xfs/xfs_xattr.c b/fs/xfs/xfs_xattr.c index f2201299311f..7fdcf3327652 100644 --- a/fs/xfs/xfs_xattr.c +++ b/fs/xfs/xfs_xattr.c @@ -166,8 +166,7 @@ xfs_xattr_put_listent( int flags, unsigned char *name, int namelen, - int valuelen, - unsigned char *value) + int valuelen) { char *prefix; int prefix_len; -- cgit v1.3-8-gc7d7 From 7af5ad28a603f2d1ef4c579b8ab0a9d4767a348e Mon Sep 17 00:00:00 2001 From: Eric Sandeen Date: Wed, 6 Apr 2016 07:57:45 +1000 Subject: xfs: remove put_value from attr ->put_listent context The put_value context member is never set; remove it and the conditional test in xfs_attr3_leaf_list_int(). Signed-off-by: Eric Sandeen Reviewed-by: Christoph Hellwig Signed-off-by: Dave Chinner --- fs/xfs/xfs_attr.h | 1 - fs/xfs/xfs_attr_list.c | 31 +++---------------------------- 2 files changed, 3 insertions(+), 29 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_attr.h b/fs/xfs/xfs_attr.h index dab4f41de278..e3da5d448bcf 100644 --- a/fs/xfs/xfs_attr.h +++ b/fs/xfs/xfs_attr.h @@ -127,7 +127,6 @@ typedef struct xfs_attr_list_context { int firstu; /* first used byte in buffer */ int flags; /* from VOP call */ int resynch; /* T/F: resynch with cursor */ - int put_value; /* T/F: need value for listent */ put_listent_func_t put_listent; /* list output fmt function */ int index; /* index into output buffer */ } xfs_attr_list_context_t; diff --git a/fs/xfs/xfs_attr_list.c b/fs/xfs/xfs_attr_list.c index d30dbfae05fe..cbf4f5d072f6 100644 --- a/fs/xfs/xfs_attr_list.c +++ b/fs/xfs/xfs_attr_list.c @@ -429,45 +429,20 @@ xfs_attr3_leaf_list_int( name_loc->nameval, (int)name_loc->namelen, be16_to_cpu(name_loc->valuelen)); - if (retval) - return retval; } else { xfs_attr_leaf_name_remote_t *name_rmt = xfs_attr3_leaf_name_remote(leaf, i); int valuelen = be32_to_cpu(name_rmt->valuelen); - if (context->put_value) { - xfs_da_args_t args; - - memset((char *)&args, 0, sizeof(args)); - args.geo = context->dp->i_mount->m_attr_geo; - args.dp = context->dp; - args.whichfork = XFS_ATTR_FORK; - args.valuelen = valuelen; - args.rmtvaluelen = valuelen; - args.value = kmem_alloc(valuelen, KM_SLEEP | KM_NOFS); - args.rmtblkno = be32_to_cpu(name_rmt->valueblk); - args.rmtblkcnt = xfs_attr3_rmt_blocks( - args.dp->i_mount, valuelen); - retval = xfs_attr_rmtval_get(&args); - if (!retval) - retval = context->put_listent(context, - entry->flags, - name_rmt->name, - (int)name_rmt->namelen, - valuelen); - kmem_free(args.value); - } else { - retval = context->put_listent(context, + retval = context->put_listent(context, entry->flags, name_rmt->name, (int)name_rmt->namelen, valuelen); - } - if (retval) - return retval; } + if (retval) + break; if (context->seen_enough) break; cursor->offset++; -- cgit v1.3-8-gc7d7 From 3ab3ffcaca99e0b77480d77bd393fc227b09069f Mon Sep 17 00:00:00 2001 From: Eric Sandeen Date: Wed, 6 Apr 2016 07:57:47 +1000 Subject: xfs: collapse cases in xfs_attr3_leaf_list_int Consolidate the 2 calls to ->put_listent in xfs_attr3_leaf_list_int(), by setting up name, namelen, and valuelen for the local vs remote cases, then call ->put_listent and do the error handling all in one spot. Signed-off-by: Eric Sandeen --- fs/xfs/xfs_attr_list.c | 34 +++++++++++++++++----------------- 1 file changed, 17 insertions(+), 17 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_attr_list.c b/fs/xfs/xfs_attr_list.c index cbf4f5d072f6..d25f26b22ac9 100644 --- a/fs/xfs/xfs_attr_list.c +++ b/fs/xfs/xfs_attr_list.c @@ -412,6 +412,9 @@ xfs_attr3_leaf_list_int( */ retval = 0; for (; i < ichdr.count; entry++, i++) { + char *name; + int namelen, valuelen; + if (be32_to_cpu(entry->hashval) != cursor->hashval) { cursor->hashval = be32_to_cpu(entry->hashval); cursor->offset = 0; @@ -421,26 +424,23 @@ xfs_attr3_leaf_list_int( continue; /* skip incomplete entries */ if (entry->flags & XFS_ATTR_LOCAL) { - xfs_attr_leaf_name_local_t *name_loc = - xfs_attr3_leaf_name_local(leaf, i); - - retval = context->put_listent(context, - entry->flags, - name_loc->nameval, - (int)name_loc->namelen, - be16_to_cpu(name_loc->valuelen)); - } else { - xfs_attr_leaf_name_remote_t *name_rmt = - xfs_attr3_leaf_name_remote(leaf, i); + xfs_attr_leaf_name_local_t *name_loc; - int valuelen = be32_to_cpu(name_rmt->valuelen); + name_loc = xfs_attr3_leaf_name_local(leaf, i); + name = name_loc->nameval; + namelen = name_loc->namelen; + valuelen = be16_to_cpu(name_loc->valuelen); + } else { + xfs_attr_leaf_name_remote_t *name_rmt; - retval = context->put_listent(context, - entry->flags, - name_rmt->name, - (int)name_rmt->namelen, - valuelen); + name_rmt = xfs_attr3_leaf_name_remote(leaf, i); + name = name_rmt->name; + namelen = name_rmt->namelen; + valuelen = be32_to_cpu(name_rmt->valuelen); } + + retval = context->put_listent(context, entry->flags, + name, namelen, valuelen); if (retval) break; if (context->seen_enough) -- cgit v1.3-8-gc7d7 From bb18782aa47d8cde90fed5cb0af312642e98a4fa Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Wed, 6 Apr 2016 08:11:25 +1000 Subject: xfs: build bios directly in xfs_add_to_ioend Currently adding a buffer to the ioend and then building a bio from the buffer list are two separate operations. We don't build the bios and submit them until the ioend is submitted, and this places a fixed dependency on bufferhead chaining in the ioend. The first step to removing the bufferhead chaining in the ioend is on the IO submission side. We can build the bio directly as we add the buffers to the ioend chain, thereby removing the need for a latter "buffer-to-bio" submission loop. This allows us to submit bios on large ioends as soon as we cannot add more data to the bio. These bios then get captured by the active plug, and hence will be dispatched as soon as either the plug overflows or we schedule away from the writeback context. This will reduce submission latency for large IOs, but will also allow more timely request queue based writeback blocking when the device becomes congested. Signed-off-by: Dave Chinner [hch: various small updates] Signed-off-by: Christoph Hellwig Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/xfs_aops.c | 69 +++++++++++++++++++++++++------------------------------ fs/xfs/xfs_aops.h | 1 + 2 files changed, 32 insertions(+), 38 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c index d445a64b979e..d03946719992 100644 --- a/fs/xfs/xfs_aops.c +++ b/fs/xfs/xfs_aops.c @@ -274,6 +274,7 @@ xfs_alloc_ioend( xfs_ioend_t *ioend; ioend = mempool_alloc(xfs_ioend_pool, GFP_NOFS); + memset(ioend, 0, sizeof(*ioend)); /* * Set the count to 1 initially, which will prevent an I/O @@ -281,16 +282,9 @@ xfs_alloc_ioend( * all the I/O from calling the completion routine too early. */ atomic_set(&ioend->io_remaining, 1); - ioend->io_error = 0; INIT_LIST_HEAD(&ioend->io_list); ioend->io_type = type; ioend->io_inode = inode; - ioend->io_buffer_head = NULL; - ioend->io_buffer_tail = NULL; - ioend->io_offset = 0; - ioend->io_size = 0; - ioend->io_append_trans = NULL; - INIT_WORK(&ioend->io_work, xfs_end_io); return ioend; } @@ -452,13 +446,18 @@ static inline int xfs_bio_add_buffer(struct bio *bio, struct buffer_head *bh) } /* - * Submit all of the bios for an ioend. We are only passed a single ioend at a - * time; the caller is responsible for chaining prior to submission. + * Submit the bio for an ioend. We are passed an ioend with a bio attached to + * it, and we submit that bio. The ioend may be used for multiple bio + * submissions, so we only want to allocate an append transaction for the ioend + * once. In the case of multiple bio submission, each bio will take an IO + * reference to the ioend to ensure that the ioend completion is only done once + * all bios have been submitted and the ioend is really done. * * If @fail is non-zero, it means that we have a situation where some part of * the submission process has failed after we have marked paged for writeback - * and unlocked them. In this situation, we need to fail the ioend chain rather - * than submit it to IO. This typically only happens on a filesystem shutdown. + * and unlocked them. In this situation, we need to fail the bio and ioend + * rather than submit it to IO. This typically only happens on a filesystem + * shutdown. */ STATIC int xfs_submit_ioend( @@ -466,14 +465,13 @@ xfs_submit_ioend( xfs_ioend_t *ioend, int status) { - struct buffer_head *bh; - struct bio *bio; - sector_t lastblock = 0; - /* Reserve log space if we might write beyond the on-disk inode size. */ if (!status && - ioend->io_type != XFS_IO_UNWRITTEN && xfs_ioend_is_append(ioend)) + ioend->io_bio && ioend->io_type != XFS_IO_UNWRITTEN && + xfs_ioend_is_append(ioend) && + !ioend->io_append_trans) status = xfs_setfilesize_trans_alloc(ioend); + /* * If we are failing the IO now, just mark the ioend with an * error and finish it. This will run IO completion immediately @@ -481,31 +479,15 @@ xfs_submit_ioend( * time. */ if (status) { + if (ioend->io_bio) + bio_put(ioend->io_bio); ioend->io_error = status; xfs_finish_ioend(ioend); return status; } - bio = NULL; - for (bh = ioend->io_buffer_head; bh; bh = bh->b_private) { - - if (!bio) { -retry: - bio = xfs_alloc_ioend_bio(bh); - } else if (bh->b_blocknr != lastblock + 1) { - xfs_submit_ioend_bio(wbc, ioend, bio); - goto retry; - } - - if (xfs_bio_add_buffer(bio, bh) != bh->b_size) { - xfs_submit_ioend_bio(wbc, ioend, bio); - goto retry; - } - - lastblock = bh->b_blocknr; - } - if (bio) - xfs_submit_ioend_bio(wbc, ioend, bio); + xfs_submit_ioend_bio(wbc, ioend, ioend->io_bio); + ioend->io_bio = NULL; xfs_finish_ioend(ioend); return 0; } @@ -523,6 +505,7 @@ xfs_add_to_ioend( struct buffer_head *bh, xfs_off_t offset, struct xfs_writepage_ctx *wpc, + struct writeback_control *wbc, struct list_head *iolist) { if (!wpc->ioend || wpc->io_type != wpc->ioend->io_type || @@ -542,8 +525,18 @@ xfs_add_to_ioend( wpc->ioend->io_buffer_tail->b_private = bh; wpc->ioend->io_buffer_tail = bh; } - bh->b_private = NULL; + +retry: + if (!wpc->ioend->io_bio) + wpc->ioend->io_bio = xfs_alloc_ioend_bio(bh); + + if (xfs_bio_add_buffer(wpc->ioend->io_bio, bh) != bh->b_size) { + xfs_submit_ioend_bio(wbc, wpc->ioend, wpc->ioend->io_bio); + wpc->ioend->io_bio = NULL; + goto retry; + } + wpc->ioend->io_size += bh->b_size; wpc->last_block = bh->b_blocknr; xfs_start_buffer_writeback(bh); @@ -803,7 +796,7 @@ xfs_writepage_map( lock_buffer(bh); if (wpc->io_type != XFS_IO_OVERWRITE) xfs_map_at_offset(inode, bh, &wpc->imap, offset); - xfs_add_to_ioend(inode, bh, offset, wpc, &submit_list); + xfs_add_to_ioend(inode, bh, offset, wpc, wbc, &submit_list); count++; } diff --git a/fs/xfs/xfs_aops.h b/fs/xfs/xfs_aops.h index b4421177b68d..8947991e0990 100644 --- a/fs/xfs/xfs_aops.h +++ b/fs/xfs/xfs_aops.h @@ -52,6 +52,7 @@ typedef struct xfs_ioend { xfs_off_t io_offset; /* offset in the file */ struct work_struct io_work; /* xfsdatad work queue */ struct xfs_trans *io_append_trans;/* xact. for size update */ + struct bio *io_bio; /* bio being built */ } xfs_ioend_t; extern const struct address_space_operations xfs_address_space_operations; -- cgit v1.3-8-gc7d7 From 37992c18bba3f578860c6448b7bae18a14e535d3 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Wed, 6 Apr 2016 08:12:28 +1000 Subject: xfs: don't release bios on completion immediately Completion of an ioend requires us to walk the bufferhead list to end writback on all the bufferheads. This, in turn, is needed so that we can end writeback on all the pages we just did IO on. To remove our dependency on bufferheads in writeback, we need to turn this around the other way - we need to walk the pages we've just completed IO on, and then walk the buffers attached to the pages and complete their IO. In doing this, we remove the requirement for the ioend to track bufferheads directly. To enable IO completion to walk all the pages we've submitted IO on, we need to keep the bios that we used for IO around until the ioend has been completed. We can do this simply by chaining the bios to the ioend at completion time, and then walking their pages directly just before destroying the ioend. Signed-off-by: Dave Chinner [hch: changed the xfs_finish_page_writeback calling convention] Signed-off-by: Christoph Hellwig Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/xfs_aops.c | 95 +++++++++++++++++++++++++++++++++++++++---------------- fs/xfs/xfs_aops.h | 5 +-- 2 files changed, 71 insertions(+), 29 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c index d03946719992..9d9a01b50078 100644 --- a/fs/xfs/xfs_aops.c +++ b/fs/xfs/xfs_aops.c @@ -84,20 +84,64 @@ xfs_find_bdev_for_inode( } /* - * We're now finished for good with this ioend structure. - * Update the page state via the associated buffer_heads, - * release holds on the inode and bio, and finally free - * up memory. Do not use the ioend after this. + * We're now finished for good with this page. Update the page state via the + * associated buffer_heads, paying attention to the start and end offsets that + * we need to process on the page. + */ +static void +xfs_finish_page_writeback( + struct inode *inode, + struct bio_vec *bvec, + int error) +{ + unsigned int blockmask = (1 << inode->i_blkbits) - 1; + unsigned int end = bvec->bv_offset + bvec->bv_len - 1; + struct buffer_head *head, *bh; + unsigned int off = 0; + + ASSERT(bvec->bv_offset < PAGE_SIZE); + ASSERT((bvec->bv_offset & blockmask) == 0); + ASSERT(end < PAGE_SIZE); + ASSERT((bvec->bv_len & blockmask) == 0); + + bh = head = page_buffers(bvec->bv_page); + + do { + if (off < bvec->bv_offset) + goto next_bh; + if (off > end) + break; + bh->b_end_io(bh, !error); +next_bh: + off += bh->b_size; + } while ((bh = bh->b_this_page) != head); +} + +/* + * We're now finished for good with this ioend structure. Update the page + * state, release holds on bios, and finally free up memory. Do not use the + * ioend after this. */ STATIC void xfs_destroy_ioend( - xfs_ioend_t *ioend) + struct xfs_ioend *ioend) { - struct buffer_head *bh, *next; + struct inode *inode = ioend->io_inode; + int error = ioend->io_error; + struct bio *bio, *next; + + for (bio = ioend->io_bio_done; bio; bio = next) { + struct bio_vec *bvec; + int i; + + next = bio->bi_private; + bio->bi_private = NULL; - for (bh = ioend->io_buffer_head; bh; bh = next) { - next = bh->b_private; - bh->b_end_io(bh, !ioend->io_error); + /* walk each page on bio, ending page IO on them */ + bio_for_each_segment_all(bvec, bio, i) + xfs_finish_page_writeback(inode, bvec, error); + + bio_put(bio); } mempool_free(ioend, xfs_ioend_pool); @@ -286,6 +330,7 @@ xfs_alloc_ioend( ioend->io_type = type; ioend->io_inode = inode; INIT_WORK(&ioend->io_work, xfs_end_io); + spin_lock_init(&ioend->io_lock); return ioend; } @@ -365,15 +410,21 @@ STATIC void xfs_end_bio( struct bio *bio) { - xfs_ioend_t *ioend = bio->bi_private; - - if (!ioend->io_error) - ioend->io_error = bio->bi_error; + struct xfs_ioend *ioend = bio->bi_private; + unsigned long flags; - /* Toss bio and pass work off to an xfsdatad thread */ bio->bi_private = NULL; bio->bi_end_io = NULL; - bio_put(bio); + + spin_lock_irqsave(&ioend->io_lock, flags); + if (!ioend->io_error) + ioend->io_error = bio->bi_error; + if (!ioend->io_bio_done) + ioend->io_bio_done = bio; + else + ioend->io_bio_done_tail->bi_private = bio; + ioend->io_bio_done_tail = bio; + spin_unlock_irqrestore(&ioend->io_lock, flags); xfs_finish_ioend(ioend); } @@ -511,21 +562,11 @@ xfs_add_to_ioend( if (!wpc->ioend || wpc->io_type != wpc->ioend->io_type || bh->b_blocknr != wpc->last_block + 1 || offset != wpc->ioend->io_offset + wpc->ioend->io_size) { - struct xfs_ioend *new; - if (wpc->ioend) list_add(&wpc->ioend->io_list, iolist); - - new = xfs_alloc_ioend(inode, wpc->io_type); - new->io_offset = offset; - new->io_buffer_head = bh; - new->io_buffer_tail = bh; - wpc->ioend = new; - } else { - wpc->ioend->io_buffer_tail->b_private = bh; - wpc->ioend->io_buffer_tail = bh; + wpc->ioend = xfs_alloc_ioend(inode, wpc->io_type); + wpc->ioend->io_offset = offset; } - bh->b_private = NULL; retry: if (!wpc->ioend->io_bio) diff --git a/fs/xfs/xfs_aops.h b/fs/xfs/xfs_aops.h index 8947991e0990..61a3dc3dbdf8 100644 --- a/fs/xfs/xfs_aops.h +++ b/fs/xfs/xfs_aops.h @@ -46,13 +46,14 @@ typedef struct xfs_ioend { int io_error; /* I/O error code */ atomic_t io_remaining; /* hold count */ struct inode *io_inode; /* file being written to */ - struct buffer_head *io_buffer_head;/* buffer linked list head */ - struct buffer_head *io_buffer_tail;/* buffer linked list tail */ size_t io_size; /* size of the extent */ xfs_off_t io_offset; /* offset in the file */ struct work_struct io_work; /* xfsdatad work queue */ struct xfs_trans *io_append_trans;/* xact. for size update */ struct bio *io_bio; /* bio being built */ + struct bio *io_bio_done; /* bios completed */ + struct bio *io_bio_done_tail; /* bios completed */ + spinlock_t io_lock; /* for bio completion list */ } xfs_ioend_t; extern const struct address_space_operations xfs_address_space_operations; -- cgit v1.3-8-gc7d7 From 0e51a8e191dbd9b9c7b7bb0a1c28d57cd2be8e6a Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 6 Apr 2016 08:34:30 +1000 Subject: xfs: optimize bio handling in the buffer writeback path This patch implements two closely related changes: First it embeds a bio the ioend structure so that we don't have to allocate one separately. Second it uses the block layer bio chaining mechanism to chain additional bios off this first one if needed instead of manually accounting for multiple bio completions in the ioend structure. Together this removes a memory allocation per ioend and greatly simplifies the ioend setup and I/O completion path. Signed-off-by: Christoph Hellwig Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/xfs_aops.c | 247 ++++++++++++++++++++++++----------------------------- fs/xfs/xfs_aops.h | 15 ++-- fs/xfs/xfs_super.c | 26 ++---- 3 files changed, 123 insertions(+), 165 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c index 9d9a01b50078..b5f1c66bbb58 100644 --- a/fs/xfs/xfs_aops.c +++ b/fs/xfs/xfs_aops.c @@ -124,18 +124,25 @@ next_bh: */ STATIC void xfs_destroy_ioend( - struct xfs_ioend *ioend) + struct xfs_ioend *ioend, + int error) { struct inode *inode = ioend->io_inode; - int error = ioend->io_error; + struct bio *last = ioend->io_bio; struct bio *bio, *next; - for (bio = ioend->io_bio_done; bio; bio = next) { + for (bio = &ioend->io_inline_bio; bio; bio = next) { struct bio_vec *bvec; int i; - next = bio->bi_private; - bio->bi_private = NULL; + /* + * For the last bio, bi_private points to the ioend, so we + * need to explicitly end the iteration here. + */ + if (bio == last) + next = NULL; + else + next = bio->bi_private; /* walk each page on bio, ending page IO on them */ bio_for_each_segment_all(bvec, bio, i) @@ -143,8 +150,6 @@ xfs_destroy_ioend( bio_put(bio); } - - mempool_free(ioend, xfs_ioend_pool); } /* @@ -218,7 +223,8 @@ xfs_setfilesize( STATIC int xfs_setfilesize_ioend( - struct xfs_ioend *ioend) + struct xfs_ioend *ioend, + int error) { struct xfs_inode *ip = XFS_I(ioend->io_inode); struct xfs_trans *tp = ioend->io_append_trans; @@ -232,36 +238,14 @@ xfs_setfilesize_ioend( __sb_writers_acquired(VFS_I(ip)->i_sb, SB_FREEZE_FS); /* we abort the update if there was an IO error */ - if (ioend->io_error) { + if (error) { xfs_trans_cancel(tp); - return ioend->io_error; + return error; } return xfs_setfilesize(ip, tp, ioend->io_offset, ioend->io_size); } -/* - * Schedule IO completion handling on the final put of an ioend. - * - * If there is no work to do we might as well call it a day and free the - * ioend right now. - */ -STATIC void -xfs_finish_ioend( - struct xfs_ioend *ioend) -{ - if (atomic_dec_and_test(&ioend->io_remaining)) { - struct xfs_mount *mp = XFS_I(ioend->io_inode)->i_mount; - - if (ioend->io_type == XFS_IO_UNWRITTEN) - queue_work(mp->m_unwritten_workqueue, &ioend->io_work); - else if (ioend->io_append_trans) - queue_work(mp->m_data_workqueue, &ioend->io_work); - else - xfs_destroy_ioend(ioend); - } -} - /* * IO write completion. */ @@ -269,16 +253,17 @@ STATIC void xfs_end_io( struct work_struct *work) { - xfs_ioend_t *ioend = container_of(work, xfs_ioend_t, io_work); - struct xfs_inode *ip = XFS_I(ioend->io_inode); - int error = 0; + struct xfs_ioend *ioend = + container_of(work, struct xfs_ioend, io_work); + struct xfs_inode *ip = XFS_I(ioend->io_inode); + int error = ioend->io_bio->bi_error; /* * Set an error if the mount has shut down and proceed with end I/O * processing so it can perform whatever cleanups are necessary. */ if (XFS_FORCED_SHUTDOWN(ip->i_mount)) - ioend->io_error = -EIO; + error = -EIO; /* * For unwritten extents we need to issue transactions to convert a @@ -288,50 +273,33 @@ xfs_end_io( * on error. */ if (ioend->io_type == XFS_IO_UNWRITTEN) { - if (ioend->io_error) + if (error) goto done; error = xfs_iomap_write_unwritten(ip, ioend->io_offset, ioend->io_size); } else if (ioend->io_append_trans) { - error = xfs_setfilesize_ioend(ioend); + error = xfs_setfilesize_ioend(ioend, error); } else { ASSERT(!xfs_ioend_is_append(ioend)); } done: - if (error) - ioend->io_error = error; - xfs_destroy_ioend(ioend); + xfs_destroy_ioend(ioend, error); } -/* - * Allocate and initialise an IO completion structure. - * We need to track unwritten extent write completion here initially. - * We'll need to extend this for updating the ondisk inode size later - * (vs. incore size). - */ -STATIC xfs_ioend_t * -xfs_alloc_ioend( - struct inode *inode, - unsigned int type) +STATIC void +xfs_end_bio( + struct bio *bio) { - xfs_ioend_t *ioend; - - ioend = mempool_alloc(xfs_ioend_pool, GFP_NOFS); - memset(ioend, 0, sizeof(*ioend)); + struct xfs_ioend *ioend = bio->bi_private; + struct xfs_mount *mp = XFS_I(ioend->io_inode)->i_mount; - /* - * Set the count to 1 initially, which will prevent an I/O - * completion callback from happening before we have started - * all the I/O from calling the completion routine too early. - */ - atomic_set(&ioend->io_remaining, 1); - INIT_LIST_HEAD(&ioend->io_list); - ioend->io_type = type; - ioend->io_inode = inode; - INIT_WORK(&ioend->io_work, xfs_end_io); - spin_lock_init(&ioend->io_lock); - return ioend; + if (ioend->io_type == XFS_IO_UNWRITTEN) + queue_work(mp->m_unwritten_workqueue, &ioend->io_work); + else if (ioend->io_append_trans) + queue_work(mp->m_data_workqueue, &ioend->io_work); + else + xfs_destroy_ioend(ioend, bio->bi_error); } STATIC int @@ -403,56 +371,6 @@ xfs_imap_valid( offset < imap->br_startoff + imap->br_blockcount; } -/* - * BIO completion handler for buffered IO. - */ -STATIC void -xfs_end_bio( - struct bio *bio) -{ - struct xfs_ioend *ioend = bio->bi_private; - unsigned long flags; - - bio->bi_private = NULL; - bio->bi_end_io = NULL; - - spin_lock_irqsave(&ioend->io_lock, flags); - if (!ioend->io_error) - ioend->io_error = bio->bi_error; - if (!ioend->io_bio_done) - ioend->io_bio_done = bio; - else - ioend->io_bio_done_tail->bi_private = bio; - ioend->io_bio_done_tail = bio; - spin_unlock_irqrestore(&ioend->io_lock, flags); - - xfs_finish_ioend(ioend); -} - -STATIC void -xfs_submit_ioend_bio( - struct writeback_control *wbc, - xfs_ioend_t *ioend, - struct bio *bio) -{ - atomic_inc(&ioend->io_remaining); - bio->bi_private = ioend; - bio->bi_end_io = xfs_end_bio; - submit_bio(wbc->sync_mode == WB_SYNC_ALL ? WRITE_SYNC : WRITE, bio); -} - -STATIC struct bio * -xfs_alloc_ioend_bio( - struct buffer_head *bh) -{ - struct bio *bio = bio_alloc(GFP_NOIO, BIO_MAX_PAGES); - - ASSERT(bio->bi_private == NULL); - bio->bi_iter.bi_sector = bh->b_blocknr * (bh->b_size >> 9); - bio->bi_bdev = bh->b_bdev; - return bio; -} - STATIC void xfs_start_buffer_writeback( struct buffer_head *bh) @@ -513,16 +431,19 @@ static inline int xfs_bio_add_buffer(struct bio *bio, struct buffer_head *bh) STATIC int xfs_submit_ioend( struct writeback_control *wbc, - xfs_ioend_t *ioend, + struct xfs_ioend *ioend, int status) { /* Reserve log space if we might write beyond the on-disk inode size. */ if (!status && - ioend->io_bio && ioend->io_type != XFS_IO_UNWRITTEN && + ioend->io_type != XFS_IO_UNWRITTEN && xfs_ioend_is_append(ioend) && !ioend->io_append_trans) status = xfs_setfilesize_trans_alloc(ioend); + ioend->io_bio->bi_private = ioend; + ioend->io_bio->bi_end_io = xfs_end_bio; + /* * If we are failing the IO now, just mark the ioend with an * error and finish it. This will run IO completion immediately @@ -530,19 +451,75 @@ xfs_submit_ioend( * time. */ if (status) { - if (ioend->io_bio) - bio_put(ioend->io_bio); - ioend->io_error = status; - xfs_finish_ioend(ioend); + ioend->io_bio->bi_error = status; + bio_endio(ioend->io_bio); return status; } - xfs_submit_ioend_bio(wbc, ioend, ioend->io_bio); - ioend->io_bio = NULL; - xfs_finish_ioend(ioend); + submit_bio(wbc->sync_mode == WB_SYNC_ALL ? WRITE_SYNC : WRITE, + ioend->io_bio); return 0; } +static void +xfs_init_bio_from_bh( + struct bio *bio, + struct buffer_head *bh) +{ + bio->bi_iter.bi_sector = bh->b_blocknr * (bh->b_size >> 9); + bio->bi_bdev = bh->b_bdev; +} + +static struct xfs_ioend * +xfs_alloc_ioend( + struct inode *inode, + unsigned int type, + xfs_off_t offset, + struct buffer_head *bh) +{ + struct xfs_ioend *ioend; + struct bio *bio; + + bio = bio_alloc_bioset(GFP_NOFS, BIO_MAX_PAGES, xfs_ioend_bioset); + xfs_init_bio_from_bh(bio, bh); + + ioend = container_of(bio, struct xfs_ioend, io_inline_bio); + INIT_LIST_HEAD(&ioend->io_list); + ioend->io_type = type; + ioend->io_inode = inode; + ioend->io_size = 0; + ioend->io_offset = offset; + INIT_WORK(&ioend->io_work, xfs_end_io); + ioend->io_append_trans = NULL; + ioend->io_bio = bio; + return ioend; +} + +/* + * Allocate a new bio, and chain the old bio to the new one. + * + * Note that we have to do perform the chaining in this unintuitive order + * so that the bi_private linkage is set up in the right direction for the + * traversal in xfs_destroy_ioend(). + */ +static void +xfs_chain_bio( + struct xfs_ioend *ioend, + struct writeback_control *wbc, + struct buffer_head *bh) +{ + struct bio *new; + + new = bio_alloc(GFP_NOFS, BIO_MAX_PAGES); + xfs_init_bio_from_bh(new, bh); + + bio_chain(ioend->io_bio, new); + bio_get(ioend->io_bio); /* for xfs_destroy_ioend */ + submit_bio(wbc->sync_mode == WB_SYNC_ALL ? WRITE_SYNC : WRITE, + ioend->io_bio); + ioend->io_bio = new; +} + /* * Test to see if we've been building up a completion structure for * earlier buffers -- if so, we try to append to this ioend if we @@ -564,19 +541,15 @@ xfs_add_to_ioend( offset != wpc->ioend->io_offset + wpc->ioend->io_size) { if (wpc->ioend) list_add(&wpc->ioend->io_list, iolist); - wpc->ioend = xfs_alloc_ioend(inode, wpc->io_type); - wpc->ioend->io_offset = offset; + wpc->ioend = xfs_alloc_ioend(inode, wpc->io_type, offset, bh); } -retry: - if (!wpc->ioend->io_bio) - wpc->ioend->io_bio = xfs_alloc_ioend_bio(bh); - - if (xfs_bio_add_buffer(wpc->ioend->io_bio, bh) != bh->b_size) { - xfs_submit_ioend_bio(wbc, wpc->ioend, wpc->ioend->io_bio); - wpc->ioend->io_bio = NULL; - goto retry; - } + /* + * If the buffer doesn't fit into the bio we need to allocate a new + * one. This shouldn't happen more than once for a given buffer. + */ + while (xfs_bio_add_buffer(wpc->ioend->io_bio, bh) != bh->b_size) + xfs_chain_bio(wpc->ioend, wbc, bh); wpc->ioend->io_size += bh->b_size; wpc->last_block = bh->b_blocknr; diff --git a/fs/xfs/xfs_aops.h b/fs/xfs/xfs_aops.h index 61a3dc3dbdf8..814aab790713 100644 --- a/fs/xfs/xfs_aops.h +++ b/fs/xfs/xfs_aops.h @@ -18,7 +18,7 @@ #ifndef __XFS_AOPS_H__ #define __XFS_AOPS_H__ -extern mempool_t *xfs_ioend_pool; +extern struct bio_set *xfs_ioend_bioset; /* * Types of I/O for bmap clustering and I/O completion tracking. @@ -37,24 +37,19 @@ enum { { XFS_IO_OVERWRITE, "overwrite" } /* - * xfs_ioend struct manages large extent writes for XFS. - * It can manage several multi-page bio's at once. + * Structure for buffered I/O completions. */ -typedef struct xfs_ioend { +struct xfs_ioend { struct list_head io_list; /* next ioend in chain */ unsigned int io_type; /* delalloc / unwritten */ - int io_error; /* I/O error code */ - atomic_t io_remaining; /* hold count */ struct inode *io_inode; /* file being written to */ size_t io_size; /* size of the extent */ xfs_off_t io_offset; /* offset in the file */ struct work_struct io_work; /* xfsdatad work queue */ struct xfs_trans *io_append_trans;/* xact. for size update */ struct bio *io_bio; /* bio being built */ - struct bio *io_bio_done; /* bios completed */ - struct bio *io_bio_done_tail; /* bios completed */ - spinlock_t io_lock; /* for bio completion list */ -} xfs_ioend_t; + struct bio io_inline_bio; /* MUST BE LAST! */ +}; extern const struct address_space_operations xfs_address_space_operations; diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index d760934109b5..e52e9c1fd933 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -58,8 +58,7 @@ #include static const struct super_operations xfs_super_operations; -static kmem_zone_t *xfs_ioend_zone; -mempool_t *xfs_ioend_pool; +struct bio_set *xfs_ioend_bioset; static struct kset *xfs_kset; /* top-level xfs sysfs dir */ #ifdef DEBUG @@ -1688,20 +1687,15 @@ MODULE_ALIAS_FS("xfs"); STATIC int __init xfs_init_zones(void) { - - xfs_ioend_zone = kmem_zone_init(sizeof(xfs_ioend_t), "xfs_ioend"); - if (!xfs_ioend_zone) + xfs_ioend_bioset = bioset_create(4 * MAX_BUF_PER_PAGE, + offsetof(struct xfs_ioend, io_inline_bio)); + if (!xfs_ioend_bioset) goto out; - xfs_ioend_pool = mempool_create_slab_pool(4 * MAX_BUF_PER_PAGE, - xfs_ioend_zone); - if (!xfs_ioend_pool) - goto out_destroy_ioend_zone; - xfs_log_ticket_zone = kmem_zone_init(sizeof(xlog_ticket_t), "xfs_log_ticket"); if (!xfs_log_ticket_zone) - goto out_destroy_ioend_pool; + goto out_free_ioend_bioset; xfs_bmap_free_item_zone = kmem_zone_init(sizeof(xfs_bmap_free_item_t), "xfs_bmap_free_item"); @@ -1797,10 +1791,8 @@ xfs_init_zones(void) kmem_zone_destroy(xfs_bmap_free_item_zone); out_destroy_log_ticket_zone: kmem_zone_destroy(xfs_log_ticket_zone); - out_destroy_ioend_pool: - mempool_destroy(xfs_ioend_pool); - out_destroy_ioend_zone: - kmem_zone_destroy(xfs_ioend_zone); + out_free_ioend_bioset: + bioset_free(xfs_ioend_bioset); out: return -ENOMEM; } @@ -1826,9 +1818,7 @@ xfs_destroy_zones(void) kmem_zone_destroy(xfs_btree_cur_zone); kmem_zone_destroy(xfs_bmap_free_item_zone); kmem_zone_destroy(xfs_log_ticket_zone); - mempool_destroy(xfs_ioend_pool); - kmem_zone_destroy(xfs_ioend_zone); - + bioset_free(xfs_ioend_bioset); } STATIC int __init -- cgit v1.3-8-gc7d7 From 253f4911f297b83745938b7f2c5649b94730b002 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 6 Apr 2016 09:19:55 +1000 Subject: xfs: better xfs_trans_alloc interface Merge xfs_trans_reserve and xfs_trans_alloc into a single function call that returns a transaction with all the required log and block reservations, and which allows passing transaction flags directly to avoid the cumbersome _xfs_trans_alloc interface. While we're at it we also get rid of the transaction type argument that has been superflous since we stopped supporting the non-CIL logging mode. The guts of it will be removed in another patch. [dchinner: fixed transaction leak in error path in xfs_setattr_nonsize] Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_attr.c | 58 +++++++----------------------- fs/xfs/libxfs/xfs_bmap.c | 22 +++++------- fs/xfs/libxfs/xfs_sb.c | 8 ++--- fs/xfs/libxfs/xfs_shared.h | 5 +-- fs/xfs/xfs_aops.c | 19 ++++------ fs/xfs/xfs_attr_inactive.c | 16 ++------- fs/xfs/xfs_bmap_util.c | 45 +++++++----------------- fs/xfs/xfs_dquot.c | 7 ++-- fs/xfs/xfs_file.c | 8 ++--- fs/xfs/xfs_fsops.c | 10 ++---- fs/xfs/xfs_inode.c | 60 ++++++++++++------------------- fs/xfs/xfs_ioctl.c | 13 +++---- fs/xfs/xfs_iomap.c | 53 +++++++++------------------- fs/xfs/xfs_iops.c | 29 ++++++--------- fs/xfs/xfs_log_recover.c | 10 +++--- fs/xfs/xfs_pnfs.c | 7 ++-- fs/xfs/xfs_qm.c | 9 ++--- fs/xfs/xfs_qm_syscalls.c | 26 ++++---------- fs/xfs/xfs_rtalloc.c | 21 +++++------ fs/xfs/xfs_symlink.c | 16 ++++----- fs/xfs/xfs_trans.c | 88 +++++++++++++++++++++------------------------- fs/xfs/xfs_trans.h | 8 ++--- 22 files changed, 191 insertions(+), 347 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_attr.c b/fs/xfs/libxfs/xfs_attr.c index fa3b948ef9c2..4e126f41a0aa 100644 --- a/fs/xfs/libxfs/xfs_attr.c +++ b/fs/xfs/libxfs/xfs_attr.c @@ -242,37 +242,21 @@ xfs_attr_set( return error; } - /* - * Start our first transaction of the day. - * - * All future transactions during this code must be "chained" off - * this one via the trans_dup() call. All transactions will contain - * the inode, and the inode will always be marked with trans_ihold(). - * Since the inode will be locked in all transactions, we must log - * the inode in every transaction to let it float upward through - * the log. - */ - args.trans = xfs_trans_alloc(mp, XFS_TRANS_ATTR_SET); + tres.tr_logres = M_RES(mp)->tr_attrsetm.tr_logres + + M_RES(mp)->tr_attrsetrt.tr_logres * args.total; + tres.tr_logcount = XFS_ATTRSET_LOG_COUNT; + tres.tr_logflags = XFS_TRANS_PERM_LOG_RES; /* * Root fork attributes can use reserved data blocks for this * operation if necessary */ - - if (rsvd) - args.trans->t_flags |= XFS_TRANS_RESERVE; - - tres.tr_logres = M_RES(mp)->tr_attrsetm.tr_logres + - M_RES(mp)->tr_attrsetrt.tr_logres * args.total; - tres.tr_logcount = XFS_ATTRSET_LOG_COUNT; - tres.tr_logflags = XFS_TRANS_PERM_LOG_RES; - error = xfs_trans_reserve(args.trans, &tres, args.total, 0); - if (error) { - xfs_trans_cancel(args.trans); + error = xfs_trans_alloc(mp, &tres, args.total, 0, + rsvd ? XFS_TRANS_RESERVE : 0, &args.trans); + if (error) return error; - } - xfs_ilock(dp, XFS_ILOCK_EXCL); + xfs_ilock(dp, XFS_ILOCK_EXCL); error = xfs_trans_reserve_quota_nblks(args.trans, dp, args.total, 0, rsvd ? XFS_QMOPT_RES_REGBLKS | XFS_QMOPT_FORCE_RES : XFS_QMOPT_RES_REGBLKS); @@ -428,32 +412,16 @@ xfs_attr_remove( if (error) return error; - /* - * Start our first transaction of the day. - * - * All future transactions during this code must be "chained" off - * this one via the trans_dup() call. All transactions will contain - * the inode, and the inode will always be marked with trans_ihold(). - * Since the inode will be locked in all transactions, we must log - * the inode in every transaction to let it float upward through - * the log. - */ - args.trans = xfs_trans_alloc(mp, XFS_TRANS_ATTR_RM); - /* * Root fork attributes can use reserved data blocks for this * operation if necessary */ - - if (flags & ATTR_ROOT) - args.trans->t_flags |= XFS_TRANS_RESERVE; - - error = xfs_trans_reserve(args.trans, &M_RES(mp)->tr_attrrm, - XFS_ATTRRM_SPACE_RES(mp), 0); - if (error) { - xfs_trans_cancel(args.trans); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_attrrm, + XFS_ATTRRM_SPACE_RES(mp), 0, + (flags & ATTR_ROOT) ? XFS_TRANS_RESERVE : 0, + &args.trans); + if (error) return error; - } xfs_ilock(dp, XFS_ILOCK_EXCL); /* diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c index 041b6948aecc..e7ec8ccd3d64 100644 --- a/fs/xfs/libxfs/xfs_bmap.c +++ b/fs/xfs/libxfs/xfs_bmap.c @@ -1121,15 +1121,14 @@ xfs_bmap_add_attrfork( mp = ip->i_mount; ASSERT(!XFS_NOT_DQATTACHED(mp, ip)); - tp = xfs_trans_alloc(mp, XFS_TRANS_ADDAFORK); + blks = XFS_ADDAFORK_SPACE_RES(mp); - if (rsvd) - tp->t_flags |= XFS_TRANS_RESERVE; - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_addafork, blks, 0); - if (error) { - xfs_trans_cancel(tp); + + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_addafork, blks, 0, + rsvd ? XFS_TRANS_RESERVE : 0, &tp); + if (error) return error; - } + xfs_ilock(ip, XFS_ILOCK_EXCL); error = xfs_trans_reserve_quota_nblks(tp, ip, blks, 0, rsvd ? XFS_QMOPT_RES_REGBLKS | XFS_QMOPT_FORCE_RES : @@ -6026,13 +6025,10 @@ xfs_bmap_split_extent( xfs_fsblock_t firstfsb; int error; - tp = xfs_trans_alloc(mp, XFS_TRANS_DIOSTRAT); - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_write, - XFS_DIOSTRAT_SPACE_RES(mp, 0), 0); - if (error) { - xfs_trans_cancel(tp); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_write, + XFS_DIOSTRAT_SPACE_RES(mp, 0), 0, 0, &tp); + if (error) return error; - } xfs_ilock(ip, XFS_ILOCK_EXCL); xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL); diff --git a/fs/xfs/libxfs/xfs_sb.c b/fs/xfs/libxfs/xfs_sb.c index 8a53eaa349f4..12ca86778e02 100644 --- a/fs/xfs/libxfs/xfs_sb.c +++ b/fs/xfs/libxfs/xfs_sb.c @@ -838,12 +838,10 @@ xfs_sync_sb( struct xfs_trans *tp; int error; - tp = _xfs_trans_alloc(mp, XFS_TRANS_SB_CHANGE, KM_SLEEP); - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_sb, 0, 0); - if (error) { - xfs_trans_cancel(tp); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_sb, 0, 0, + XFS_TRANS_NO_WRITECOUNT, &tp); + if (error) return error; - } xfs_log_sb(tp); if (wait) diff --git a/fs/xfs/libxfs/xfs_shared.h b/fs/xfs/libxfs/xfs_shared.h index 81ac870834da..7d4ab6439081 100644 --- a/fs/xfs/libxfs/xfs_shared.h +++ b/fs/xfs/libxfs/xfs_shared.h @@ -181,8 +181,9 @@ int xfs_log_calc_minimum_size(struct xfs_mount *); #define XFS_TRANS_SYNC 0x08 /* make commit synchronous */ #define XFS_TRANS_DQ_DIRTY 0x10 /* at least one dquot in trx dirty */ #define XFS_TRANS_RESERVE 0x20 /* OK to use reserved data blocks */ -#define XFS_TRANS_FREEZE_PROT 0x40 /* Transaction has elevated writer - count in superblock */ +#define XFS_TRANS_NO_WRITECOUNT 0x40 /* do not elevate SB writecount */ +#define XFS_TRANS_NOFS 0x80 /* pass KM_NOFS to kmem_alloc */ + /* * Field values for xfs_trans_mod_sb. */ diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c index d445a64b979e..3311db6ecbbd 100644 --- a/fs/xfs/xfs_aops.c +++ b/fs/xfs/xfs_aops.c @@ -120,13 +120,9 @@ xfs_setfilesize_trans_alloc( struct xfs_trans *tp; int error; - tp = xfs_trans_alloc(mp, XFS_TRANS_FSYNC_TS); - - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_fsyncts, 0, 0); - if (error) { - xfs_trans_cancel(tp); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_fsyncts, 0, 0, 0, &tp); + if (error) return error; - } ioend->io_append_trans = tp; @@ -1391,13 +1387,10 @@ xfs_end_io_direct_write( trace_xfs_end_io_direct_write_append(ip, offset, size); - tp = xfs_trans_alloc(mp, XFS_TRANS_FSYNC_TS); - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_fsyncts, 0, 0); - if (error) { - xfs_trans_cancel(tp); - return error; - } - error = xfs_setfilesize(ip, tp, offset, size); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_fsyncts, 0, 0, 0, + &tp); + if (!error) + error = xfs_setfilesize(ip, tp, offset, size); } return error; diff --git a/fs/xfs/xfs_attr_inactive.c b/fs/xfs/xfs_attr_inactive.c index 2bb959ada45b..55d214981ed2 100644 --- a/fs/xfs/xfs_attr_inactive.c +++ b/fs/xfs/xfs_attr_inactive.c @@ -405,21 +405,11 @@ xfs_attr_inactive( goto out_destroy_fork; xfs_iunlock(dp, lock_mode); - /* - * Start our first transaction of the day. - * - * All future transactions during this code must be "chained" off - * this one via the trans_dup() call. All transactions will contain - * the inode, and the inode will always be marked with trans_ihold(). - * Since the inode will be locked in all transactions, we must log - * the inode in every transaction to let it float upward through - * the log. - */ lock_mode = 0; - trans = xfs_trans_alloc(mp, XFS_TRANS_ATTRINVAL); - error = xfs_trans_reserve(trans, &M_RES(mp)->tr_attrinval, 0, 0); + + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_attrinval, 0, 0, 0, &trans); if (error) - goto out_cancel; + goto out_destroy_fork; lock_mode = XFS_ILOCK_EXCL; xfs_ilock(dp, lock_mode); diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c index a32c1dcae2ff..3246ebc7cbd0 100644 --- a/fs/xfs/xfs_bmap_util.c +++ b/fs/xfs/xfs_bmap_util.c @@ -900,19 +900,15 @@ xfs_free_eofblocks( * Free them up now by truncating the file to * its current size. */ - tp = xfs_trans_alloc(mp, XFS_TRANS_INACTIVE); - if (need_iolock) { - if (!xfs_ilock_nowait(ip, XFS_IOLOCK_EXCL)) { - xfs_trans_cancel(tp); + if (!xfs_ilock_nowait(ip, XFS_IOLOCK_EXCL)) return -EAGAIN; - } } - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_itruncate, 0, 0); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_itruncate, 0, 0, 0, + &tp); if (error) { ASSERT(XFS_FORCED_SHUTDOWN(mp)); - xfs_trans_cancel(tp); if (need_iolock) xfs_iunlock(ip, XFS_IOLOCK_EXCL); return error; @@ -1037,9 +1033,9 @@ xfs_alloc_file_space( /* * Allocate and setup the transaction. */ - tp = xfs_trans_alloc(mp, XFS_TRANS_DIOSTRAT); - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_write, - resblks, resrtextents); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_write, resblks, + resrtextents, 0, &tp); + /* * Check for running out of space */ @@ -1048,7 +1044,6 @@ xfs_alloc_file_space( * Free the transaction structure. */ ASSERT(error == -ENOSPC || XFS_FORCED_SHUTDOWN(mp)); - xfs_trans_cancel(tp); break; } xfs_ilock(ip, XFS_ILOCK_EXCL); @@ -1311,18 +1306,10 @@ xfs_free_file_space( * transaction to dip into the reserve blocks to ensure * the freeing of the space succeeds at ENOSPC. */ - tp = xfs_trans_alloc(mp, XFS_TRANS_DIOSTRAT); - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_write, resblks, 0); - - /* - * check for running out of space - */ + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_write, resblks, 0, 0, + &tp); if (error) { - /* - * Free the transaction structure. - */ ASSERT(error == -ENOSPC || XFS_FORCED_SHUTDOWN(mp)); - xfs_trans_cancel(tp); break; } xfs_ilock(ip, XFS_ILOCK_EXCL); @@ -1482,19 +1469,16 @@ xfs_shift_file_space( } while (!error && !done) { - tp = xfs_trans_alloc(mp, XFS_TRANS_DIOSTRAT); /* * We would need to reserve permanent block for transaction. * This will come into picture when after shifting extent into * hole we found that adjacent extents can be merged which * may lead to freeing of a block during record update. */ - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_write, - XFS_DIOSTRAT_SPACE_RES(mp, 0), 0); - if (error) { - xfs_trans_cancel(tp); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_write, + XFS_DIOSTRAT_SPACE_RES(mp, 0), 0, 0, &tp); + if (error) break; - } xfs_ilock(ip, XFS_ILOCK_EXCL); error = xfs_trans_reserve_quota(tp, mp, ip->i_udquot, @@ -1747,12 +1731,9 @@ xfs_swap_extents( if (error) goto out_unlock; - tp = xfs_trans_alloc(mp, XFS_TRANS_SWAPEXT); - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_ichange, 0, 0); - if (error) { - xfs_trans_cancel(tp); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_ichange, 0, 0, 0, &tp); + if (error) goto out_unlock; - } /* * Lock and join the inodes to the tansaction so that transaction commit diff --git a/fs/xfs/xfs_dquot.c b/fs/xfs/xfs_dquot.c index 316b2a1bdba5..23e24329b132 100644 --- a/fs/xfs/xfs_dquot.c +++ b/fs/xfs/xfs_dquot.c @@ -614,11 +614,10 @@ xfs_qm_dqread( trace_xfs_dqread(dqp); if (flags & XFS_QMOPT_DQALLOC) { - tp = xfs_trans_alloc(mp, XFS_TRANS_QM_DQALLOC); - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_qm_dqalloc, - XFS_QM_DQALLOC_SPACE_RES(mp), 0); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_qm_dqalloc, + XFS_QM_DQALLOC_SPACE_RES(mp), 0, 0, &tp); if (error) - goto error1; + goto error0; } /* diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c index ac0fd32de31e..98bbd8f84c76 100644 --- a/fs/xfs/xfs_file.c +++ b/fs/xfs/xfs_file.c @@ -145,12 +145,10 @@ xfs_update_prealloc_flags( struct xfs_trans *tp; int error; - tp = xfs_trans_alloc(ip->i_mount, XFS_TRANS_WRITEID); - error = xfs_trans_reserve(tp, &M_RES(ip->i_mount)->tr_writeid, 0, 0); - if (error) { - xfs_trans_cancel(tp); + error = xfs_trans_alloc(ip->i_mount, &M_RES(ip->i_mount)->tr_writeid, + 0, 0, 0, &tp); + if (error) return error; - } xfs_ilock(ip, XFS_ILOCK_EXCL); xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL); diff --git a/fs/xfs/xfs_fsops.c b/fs/xfs/xfs_fsops.c index ee3aaa0a5317..9c563d480b3c 100644 --- a/fs/xfs/xfs_fsops.c +++ b/fs/xfs/xfs_fsops.c @@ -198,14 +198,10 @@ xfs_growfs_data_private( return error; } - tp = xfs_trans_alloc(mp, XFS_TRANS_GROWFS); - tp->t_flags |= XFS_TRANS_RESERVE; - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_growdata, - XFS_GROWFS_SPACE_RES(mp), 0); - if (error) { - xfs_trans_cancel(tp); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_growdata, + XFS_GROWFS_SPACE_RES(mp), 0, XFS_TRANS_RESERVE, &tp); + if (error) return error; - } /* * Write new AG headers to disk. Non-transactional, but written diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index 96f606deee31..b82c729634f6 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -1161,11 +1161,9 @@ xfs_create( rdev = 0; resblks = XFS_MKDIR_SPACE_RES(mp, name->len); tres = &M_RES(mp)->tr_mkdir; - tp = xfs_trans_alloc(mp, XFS_TRANS_MKDIR); } else { resblks = XFS_CREATE_SPACE_RES(mp, name->len); tres = &M_RES(mp)->tr_create; - tp = xfs_trans_alloc(mp, XFS_TRANS_CREATE); } /* @@ -1174,20 +1172,19 @@ xfs_create( * the case we'll drop the one we have and get a more * appropriate transaction later. */ - error = xfs_trans_reserve(tp, tres, resblks, 0); + error = xfs_trans_alloc(mp, tres, resblks, 0, 0, &tp); if (error == -ENOSPC) { /* flush outstanding delalloc blocks and retry */ xfs_flush_inodes(mp); - error = xfs_trans_reserve(tp, tres, resblks, 0); + error = xfs_trans_alloc(mp, tres, resblks, 0, 0, &tp); } if (error == -ENOSPC) { /* No space at all so try a "no-allocation" reservation */ resblks = 0; - error = xfs_trans_reserve(tp, tres, 0, 0); + error = xfs_trans_alloc(mp, tres, 0, 0, 0, &tp); } if (error) - goto out_trans_cancel; - + goto out_release_inode; xfs_ilock(dp, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL | XFS_IOLOCK_PARENT | XFS_ILOCK_PARENT); @@ -1337,17 +1334,16 @@ xfs_create_tmpfile( return error; resblks = XFS_IALLOC_SPACE_RES(mp); - tp = xfs_trans_alloc(mp, XFS_TRANS_CREATE_TMPFILE); - tres = &M_RES(mp)->tr_create_tmpfile; - error = xfs_trans_reserve(tp, tres, resblks, 0); + + error = xfs_trans_alloc(mp, tres, resblks, 0, 0, &tp); if (error == -ENOSPC) { /* No space at all so try a "no-allocation" reservation */ resblks = 0; - error = xfs_trans_reserve(tp, tres, 0, 0); + error = xfs_trans_alloc(mp, tres, 0, 0, 0, &tp); } if (error) - goto out_trans_cancel; + goto out_release_inode; error = xfs_trans_reserve_quota(tp, mp, udqp, gdqp, pdqp, resblks, 1, 0); @@ -1432,15 +1428,14 @@ xfs_link( if (error) goto std_return; - tp = xfs_trans_alloc(mp, XFS_TRANS_LINK); resblks = XFS_LINK_SPACE_RES(mp, target_name->len); - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_link, resblks, 0); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_link, resblks, 0, 0, &tp); if (error == -ENOSPC) { resblks = 0; - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_link, 0, 0); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_link, 0, 0, 0, &tp); } if (error) - goto error_return; + goto std_return; xfs_ilock(tdp, XFS_IOLOCK_EXCL | XFS_IOLOCK_PARENT); xfs_lock_two_inodes(sip, tdp, XFS_ILOCK_EXCL); @@ -1710,11 +1705,9 @@ xfs_inactive_truncate( struct xfs_trans *tp; int error; - tp = xfs_trans_alloc(mp, XFS_TRANS_INACTIVE); - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_itruncate, 0, 0); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_itruncate, 0, 0, 0, &tp); if (error) { ASSERT(XFS_FORCED_SHUTDOWN(mp)); - xfs_trans_cancel(tp); return error; } @@ -1764,8 +1757,6 @@ xfs_inactive_ifree( struct xfs_trans *tp; int error; - tp = xfs_trans_alloc(mp, XFS_TRANS_INACTIVE); - /* * The ifree transaction might need to allocate blocks for record * insertion to the finobt. We don't want to fail here at ENOSPC, so @@ -1781,9 +1772,8 @@ xfs_inactive_ifree( * now remains allocated and sits on the unlinked list until the fs is * repaired. */ - tp->t_flags |= XFS_TRANS_RESERVE; - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_ifree, - XFS_IFREE_SPACE_RES(mp), 0); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_ifree, + XFS_IFREE_SPACE_RES(mp), 0, XFS_TRANS_RESERVE, &tp); if (error) { if (error == -ENOSPC) { xfs_warn_ratelimited(mp, @@ -1792,7 +1782,6 @@ xfs_inactive_ifree( } else { ASSERT(XFS_FORCED_SHUTDOWN(mp)); } - xfs_trans_cancel(tp); return error; } @@ -2525,11 +2514,6 @@ xfs_remove( if (error) goto std_return; - if (is_dir) - tp = xfs_trans_alloc(mp, XFS_TRANS_RMDIR); - else - tp = xfs_trans_alloc(mp, XFS_TRANS_REMOVE); - /* * We try to get the real space reservation first, * allowing for directory btree deletion(s) implying @@ -2540,14 +2524,15 @@ xfs_remove( * block from the directory. */ resblks = XFS_REMOVE_SPACE_RES(mp); - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_remove, resblks, 0); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_remove, resblks, 0, 0, &tp); if (error == -ENOSPC) { resblks = 0; - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_remove, 0, 0); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_remove, 0, 0, 0, + &tp); } if (error) { ASSERT(error != -ENOSPC); - goto out_trans_cancel; + goto std_return; } xfs_ilock(dp, XFS_IOLOCK_EXCL | XFS_IOLOCK_PARENT); @@ -2910,15 +2895,15 @@ xfs_rename( xfs_sort_for_rename(src_dp, target_dp, src_ip, target_ip, wip, inodes, &num_inodes); - tp = xfs_trans_alloc(mp, XFS_TRANS_RENAME); spaceres = XFS_RENAME_SPACE_RES(mp, target_name->len); - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_rename, spaceres, 0); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_rename, spaceres, 0, 0, &tp); if (error == -ENOSPC) { spaceres = 0; - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_rename, 0, 0); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_rename, 0, 0, 0, + &tp); } if (error) - goto out_trans_cancel; + goto out_release_wip; /* * Attach the dquots to the inodes @@ -3155,6 +3140,7 @@ out_bmap_cancel: xfs_bmap_cancel(&free_list); out_trans_cancel: xfs_trans_cancel(tp); +out_release_wip: if (wip) IRELE(wip); return error; diff --git a/fs/xfs/xfs_ioctl.c b/fs/xfs/xfs_ioctl.c index bcb6c19ce3ea..6f82187ee297 100644 --- a/fs/xfs/xfs_ioctl.c +++ b/fs/xfs/xfs_ioctl.c @@ -334,12 +334,10 @@ xfs_set_dmattrs( if (XFS_FORCED_SHUTDOWN(mp)) return -EIO; - tp = xfs_trans_alloc(mp, XFS_TRANS_SET_DMATTRS); - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_ichange, 0, 0); - if (error) { - xfs_trans_cancel(tp); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_ichange, 0, 0, 0, &tp); + if (error) return error; - } + xfs_ilock(ip, XFS_ILOCK_EXCL); xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL); @@ -1141,10 +1139,9 @@ xfs_ioctl_setattr_get_trans( if (XFS_FORCED_SHUTDOWN(mp)) goto out_unlock; - tp = xfs_trans_alloc(mp, XFS_TRANS_SETATTR_NOT_SIZE); - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_ichange, 0, 0); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_ichange, 0, 0, 0, &tp); if (error) - goto out_cancel; + return ERR_PTR(error); xfs_ilock(ip, XFS_ILOCK_EXCL); xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL | join_flags); diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c index d81bdc080370..58391355a44d 100644 --- a/fs/xfs/xfs_iomap.c +++ b/fs/xfs/xfs_iomap.c @@ -132,6 +132,7 @@ xfs_iomap_write_direct( int error; int lockmode; int bmapi_flags = XFS_BMAPI_PREALLOC; + uint tflags = 0; rt = XFS_IS_REALTIME_INODE(ip); extsz = xfs_get_extsz_hint(ip); @@ -191,11 +192,6 @@ xfs_iomap_write_direct( if (error) return error; - /* - * Allocate and setup the transaction - */ - tp = xfs_trans_alloc(mp, XFS_TRANS_DIOSTRAT); - /* * For DAX, we do not allocate unwritten extents, but instead we zero * the block before we commit the transaction. Ideally we'd like to do @@ -209,23 +205,17 @@ xfs_iomap_write_direct( * the reserve block pool for bmbt block allocation if there is no space * left but we need to do unwritten extent conversion. */ - if (IS_DAX(VFS_I(ip))) { bmapi_flags = XFS_BMAPI_CONVERT | XFS_BMAPI_ZERO; if (ISUNWRITTEN(imap)) { - tp->t_flags |= XFS_TRANS_RESERVE; + tflags |= XFS_TRANS_RESERVE; resblks = XFS_DIOSTRAT_SPACE_RES(mp, 0) << 1; } } - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_write, - resblks, resrtextents); - /* - * Check for running out of space, note: need lock to return - */ - if (error) { - xfs_trans_cancel(tp); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_write, resblks, resrtextents, + tflags, &tp); + if (error) return error; - } lockmode = XFS_ILOCK_EXCL; xfs_ilock(ip, lockmode); @@ -726,15 +716,13 @@ xfs_iomap_write_allocate( nimaps = 0; while (nimaps == 0) { - tp = xfs_trans_alloc(mp, XFS_TRANS_STRAT_WRITE); - tp->t_flags |= XFS_TRANS_RESERVE; nres = XFS_EXTENTADD_SPACE_RES(mp, XFS_DATA_FORK); - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_write, - nres, 0); - if (error) { - xfs_trans_cancel(tp); + + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_write, nres, + 0, XFS_TRANS_RESERVE, &tp); + if (error) return error; - } + xfs_ilock(ip, XFS_ILOCK_EXCL); xfs_trans_ijoin(tp, ip, 0); @@ -878,25 +866,18 @@ xfs_iomap_write_unwritten( do { /* - * set up a transaction to convert the range of extents + * Set up a transaction to convert the range of extents * from unwritten to real. Do allocations in a loop until * we have covered the range passed in. * - * Note that we open code the transaction allocation here - * to pass KM_NOFS--we can't risk to recursing back into - * the filesystem here as we might be asked to write out - * the same inode that we complete here and might deadlock - * on the iolock. + * Note that we can't risk to recursing back into the filesystem + * here as we might be asked to write out the same inode that we + * complete here and might deadlock on the iolock. */ - sb_start_intwrite(mp->m_super); - tp = _xfs_trans_alloc(mp, XFS_TRANS_STRAT_WRITE, KM_NOFS); - tp->t_flags |= XFS_TRANS_RESERVE | XFS_TRANS_FREEZE_PROT; - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_write, - resblks, 0); - if (error) { - xfs_trans_cancel(tp); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_write, resblks, 0, + XFS_TRANS_RESERVE | XFS_TRANS_NOFS, &tp); + if (error) return error; - } xfs_ilock(ip, XFS_ILOCK_EXCL); xfs_trans_ijoin(tp, ip, 0); diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c index fb7dc61f4a29..fc7766164dc9 100644 --- a/fs/xfs/xfs_iops.c +++ b/fs/xfs/xfs_iops.c @@ -599,12 +599,12 @@ xfs_setattr_nonsize( return error; } - tp = xfs_trans_alloc(mp, XFS_TRANS_SETATTR_NOT_SIZE); - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_ichange, 0, 0); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_ichange, 0, 0, 0, &tp); if (error) - goto out_trans_cancel; + goto out_dqrele; xfs_ilock(ip, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, ip, 0); /* * Change file ownership. Must be the owner or privileged. @@ -633,12 +633,10 @@ xfs_setattr_nonsize( NULL, capable(CAP_FOWNER) ? XFS_QMOPT_FORCE_RES : 0); if (error) /* out of quota */ - goto out_unlock; + goto out_cancel; } } - xfs_trans_ijoin(tp, ip, 0); - /* * Change file ownership. Must be the owner or privileged. */ @@ -722,10 +720,9 @@ xfs_setattr_nonsize( return 0; -out_unlock: - xfs_iunlock(ip, XFS_ILOCK_EXCL); -out_trans_cancel: +out_cancel: xfs_trans_cancel(tp); +out_dqrele: xfs_qm_dqrele(udqp); xfs_qm_dqrele(gdqp); return error; @@ -834,7 +831,7 @@ xfs_setattr_size( * We have to do all the page cache truncate work outside the * transaction context as the "lock" order is page lock->log space * reservation as defined by extent allocation in the writeback path. - * Hence a truncate can fail with ENOMEM from xfs_trans_reserve(), but + * Hence a truncate can fail with ENOMEM from xfs_trans_alloc(), but * having already truncated the in-memory version of the file (i.e. made * user visible changes). There's not much we can do about this, except * to hope that the caller sees ENOMEM and retries the truncate @@ -849,10 +846,9 @@ xfs_setattr_size( return error; truncate_setsize(inode, newsize); - tp = xfs_trans_alloc(mp, XFS_TRANS_SETATTR_SIZE); - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_itruncate, 0, 0); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_itruncate, 0, 0, 0, &tp); if (error) - goto out_trans_cancel; + return error; lock_flags |= XFS_ILOCK_EXCL; xfs_ilock(ip, XFS_ILOCK_EXCL); @@ -971,12 +967,9 @@ xfs_vn_update_time( trace_xfs_update_time(ip); - tp = xfs_trans_alloc(mp, XFS_TRANS_FSYNC_TS); - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_fsyncts, 0, 0); - if (error) { - xfs_trans_cancel(tp); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_fsyncts, 0, 0, 0, &tp); + if (error) return error; - } xfs_ilock(ip, XFS_ILOCK_EXCL); if (flags & S_CTIME) diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index 396565f43247..558f3d1d91ad 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -4205,10 +4205,9 @@ xlog_recover_process_efi( } } - tp = xfs_trans_alloc(mp, 0); - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_itruncate, 0, 0); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_itruncate, 0, 0, 0, &tp); if (error) - goto abort_error; + return error; efdp = xfs_trans_get_efd(tp, efip, efip->efi_format.efi_nextents); for (i = 0; i < efip->efi_format.efi_nextents; i++) { @@ -4355,10 +4354,9 @@ xlog_recover_clear_agi_bucket( int offset; int error; - tp = xfs_trans_alloc(mp, XFS_TRANS_CLEAR_AGI_BUCKET); - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_clearagi, 0, 0); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_clearagi, 0, 0, 0, &tp); if (error) - goto out_abort; + goto out_error; error = xfs_read_agi(mp, tp, agno, &agibp); if (error) diff --git a/fs/xfs/xfs_pnfs.c b/fs/xfs/xfs_pnfs.c index ade236e90bb3..3332baeac582 100644 --- a/fs/xfs/xfs_pnfs.c +++ b/fs/xfs/xfs_pnfs.c @@ -308,12 +308,9 @@ xfs_fs_commit_blocks( goto out_drop_iolock; } - tp = xfs_trans_alloc(mp, XFS_TRANS_SETATTR_NOT_SIZE); - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_ichange, 0, 0); - if (error) { - xfs_trans_cancel(tp); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_ichange, 0, 0, 0, &tp); + if (error) goto out_drop_iolock; - } xfs_ilock(ip, XFS_ILOCK_EXCL); xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL); diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c index be125e1758c1..a60d9e2739d1 100644 --- a/fs/xfs/xfs_qm.c +++ b/fs/xfs/xfs_qm.c @@ -783,13 +783,10 @@ xfs_qm_qino_alloc( } } - tp = xfs_trans_alloc(mp, XFS_TRANS_QM_QINOCREATE); - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_create, - XFS_QM_QINOCREATE_SPACE_RES(mp), 0); - if (error) { - xfs_trans_cancel(tp); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_create, + XFS_QM_QINOCREATE_SPACE_RES(mp), 0, 0, &tp); + if (error) return error; - } if (need_alloc) { error = xfs_dir_ialloc(&tp, NULL, S_IFREG, 1, 0, 0, 1, ip, diff --git a/fs/xfs/xfs_qm_syscalls.c b/fs/xfs/xfs_qm_syscalls.c index f4d0e0a8f517..475a3882a81f 100644 --- a/fs/xfs/xfs_qm_syscalls.c +++ b/fs/xfs/xfs_qm_syscalls.c @@ -236,10 +236,8 @@ xfs_qm_scall_trunc_qfile( xfs_ilock(ip, XFS_IOLOCK_EXCL); - tp = xfs_trans_alloc(mp, XFS_TRANS_TRUNCATE_FILE); - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_itruncate, 0, 0); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_itruncate, 0, 0, 0, &tp); if (error) { - xfs_trans_cancel(tp); xfs_iunlock(ip, XFS_IOLOCK_EXCL); goto out_put; } @@ -436,12 +434,9 @@ xfs_qm_scall_setqlim( defq = xfs_get_defquota(dqp, q); xfs_dqunlock(dqp); - tp = xfs_trans_alloc(mp, XFS_TRANS_QM_SETQLIM); - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_qm_setqlim, 0, 0); - if (error) { - xfs_trans_cancel(tp); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_qm_setqlim, 0, 0, 0, &tp); + if (error) goto out_rele; - } xfs_dqlock(dqp); xfs_trans_dqjoin(tp, dqp); @@ -569,13 +564,9 @@ xfs_qm_log_quotaoff_end( int error; xfs_qoff_logitem_t *qoffi; - tp = xfs_trans_alloc(mp, XFS_TRANS_QM_QUOTAOFF_END); - - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_qm_equotaoff, 0, 0); - if (error) { - xfs_trans_cancel(tp); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_qm_equotaoff, 0, 0, 0, &tp); + if (error) return error; - } qoffi = xfs_trans_get_qoff_item(tp, startqoff, flags & XFS_ALL_QUOTA_ACCT); @@ -603,12 +594,9 @@ xfs_qm_log_quotaoff( *qoffstartp = NULL; - tp = xfs_trans_alloc(mp, XFS_TRANS_QM_QUOTAOFF); - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_qm_quotaoff, 0, 0); - if (error) { - xfs_trans_cancel(tp); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_qm_quotaoff, 0, 0, 0, &tp); + if (error) goto out; - } qoffi = xfs_trans_get_qoff_item(tp, NULL, flags & XFS_ALL_QUOTA_ACCT); xfs_trans_log_quotaoff_item(tp, qoffi); diff --git a/fs/xfs/xfs_rtalloc.c b/fs/xfs/xfs_rtalloc.c index abf44435d04a..3938b37d1043 100644 --- a/fs/xfs/xfs_rtalloc.c +++ b/fs/xfs/xfs_rtalloc.c @@ -780,15 +780,14 @@ xfs_growfs_rt_alloc( * Allocate space to the file, as necessary. */ while (oblocks < nblocks) { - tp = xfs_trans_alloc(mp, XFS_TRANS_GROWFSRT_ALLOC); resblks = XFS_GROWFSRT_SPACE_RES(mp, nblocks - oblocks); /* * Reserve space & log for one extent added to the file. */ - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_growrtalloc, - resblks, 0); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_growrtalloc, resblks, + 0, 0, &tp); if (error) - goto out_trans_cancel; + return error; /* * Lock the inode. */ @@ -823,14 +822,13 @@ xfs_growfs_rt_alloc( for (bno = map.br_startoff, fsbno = map.br_startblock; bno < map.br_startoff + map.br_blockcount; bno++, fsbno++) { - tp = xfs_trans_alloc(mp, XFS_TRANS_GROWFSRT_ZERO); /* * Reserve log for one block zeroing. */ - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_growrtzero, - 0, 0); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_growrtzero, + 0, 0, 0, &tp); if (error) - goto out_trans_cancel; + return error; /* * Lock the bitmap inode. */ @@ -994,11 +992,10 @@ xfs_growfs_rt( /* * Start a transaction, get the log reservation. */ - tp = xfs_trans_alloc(mp, XFS_TRANS_GROWFSRT_FREE); - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_growrtfree, - 0, 0); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_growrtfree, 0, 0, 0, + &tp); if (error) - goto error_cancel; + break; /* * Lock out other callers by grabbing the bitmap inode lock. */ diff --git a/fs/xfs/xfs_symlink.c b/fs/xfs/xfs_symlink.c index b44284c1adda..c3aeaa884478 100644 --- a/fs/xfs/xfs_symlink.c +++ b/fs/xfs/xfs_symlink.c @@ -221,7 +221,6 @@ xfs_symlink( if (error) return error; - tp = xfs_trans_alloc(mp, XFS_TRANS_SYMLINK); /* * The symlink will fit into the inode data fork? * There can't be any attributes so we get the whole variable part. @@ -231,13 +230,15 @@ xfs_symlink( else fs_blocks = xfs_symlink_blocks(mp, pathlen); resblks = XFS_SYMLINK_SPACE_RES(mp, link_name->len, fs_blocks); - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_symlink, resblks, 0); + + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_symlink, resblks, 0, 0, &tp); if (error == -ENOSPC && fs_blocks == 0) { resblks = 0; - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_symlink, 0, 0); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_symlink, 0, 0, 0, + &tp); } if (error) - goto out_trans_cancel; + goto out_release_inode; xfs_ilock(dp, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL | XFS_IOLOCK_PARENT | XFS_ILOCK_PARENT); @@ -455,12 +456,9 @@ xfs_inactive_symlink_rmt( */ ASSERT(ip->i_d.di_nextents > 0 && ip->i_d.di_nextents <= 2); - tp = xfs_trans_alloc(mp, XFS_TRANS_INACTIVE); - error = xfs_trans_reserve(tp, &M_RES(mp)->tr_itruncate, 0, 0); - if (error) { - xfs_trans_cancel(tp); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_itruncate, 0, 0, 0, &tp); + if (error) return error; - } xfs_ilock(ip, XFS_ILOCK_EXCL); xfs_trans_ijoin(tp, ip, 0); diff --git a/fs/xfs/xfs_trans.c b/fs/xfs/xfs_trans.c index 20c53666cb4b..b3669efb2a0a 100644 --- a/fs/xfs/xfs_trans.c +++ b/fs/xfs/xfs_trans.c @@ -46,47 +46,6 @@ xfs_trans_init( xfs_trans_resv_calc(mp, M_RES(mp)); } -/* - * This routine is called to allocate a transaction structure. - * The type parameter indicates the type of the transaction. These - * are enumerated in xfs_trans.h. - * - * Dynamically allocate the transaction structure from the transaction - * zone, initialize it, and return it to the caller. - */ -xfs_trans_t * -xfs_trans_alloc( - xfs_mount_t *mp, - uint type) -{ - xfs_trans_t *tp; - - sb_start_intwrite(mp->m_super); - tp = _xfs_trans_alloc(mp, type, KM_SLEEP); - tp->t_flags |= XFS_TRANS_FREEZE_PROT; - return tp; -} - -xfs_trans_t * -_xfs_trans_alloc( - xfs_mount_t *mp, - uint type, - xfs_km_flags_t memflags) -{ - xfs_trans_t *tp; - - WARN_ON(mp->m_super->s_writers.frozen == SB_FREEZE_COMPLETE); - atomic_inc(&mp->m_active_trans); - - tp = kmem_zone_zalloc(xfs_trans_zone, memflags); - tp->t_magic = XFS_TRANS_HEADER_MAGIC; - tp->t_type = type; - tp->t_mountp = mp; - INIT_LIST_HEAD(&tp->t_items); - INIT_LIST_HEAD(&tp->t_busy); - return tp; -} - /* * Free the transaction structure. If there is more clean up * to do when the structure is freed, add it here. @@ -99,7 +58,7 @@ xfs_trans_free( xfs_extent_busy_clear(tp->t_mountp, &tp->t_busy, false); atomic_dec(&tp->t_mountp->m_active_trans); - if (tp->t_flags & XFS_TRANS_FREEZE_PROT) + if (!(tp->t_flags & XFS_TRANS_NO_WRITECOUNT)) sb_end_intwrite(tp->t_mountp->m_super); xfs_trans_free_dqinfo(tp); kmem_zone_free(xfs_trans_zone, tp); @@ -125,7 +84,6 @@ xfs_trans_dup( * Initialize the new transaction structure. */ ntp->t_magic = XFS_TRANS_HEADER_MAGIC; - ntp->t_type = tp->t_type; ntp->t_mountp = tp->t_mountp; INIT_LIST_HEAD(&ntp->t_items); INIT_LIST_HEAD(&ntp->t_busy); @@ -135,9 +93,9 @@ xfs_trans_dup( ntp->t_flags = XFS_TRANS_PERM_LOG_RES | (tp->t_flags & XFS_TRANS_RESERVE) | - (tp->t_flags & XFS_TRANS_FREEZE_PROT); + (tp->t_flags & XFS_TRANS_NO_WRITECOUNT); /* We gave our writer reference to the new transaction */ - tp->t_flags &= ~XFS_TRANS_FREEZE_PROT; + tp->t_flags |= XFS_TRANS_NO_WRITECOUNT; ntp->t_ticket = xfs_log_ticket_get(tp->t_ticket); ntp->t_blk_res = tp->t_blk_res - tp->t_blk_res_used; tp->t_blk_res = tp->t_blk_res_used; @@ -165,7 +123,7 @@ xfs_trans_dup( * This does not do quota reservations. That typically is done by the * caller afterwards. */ -int +static int xfs_trans_reserve( struct xfs_trans *tp, struct xfs_trans_res *resp, @@ -219,7 +177,7 @@ xfs_trans_reserve( resp->tr_logres, resp->tr_logcount, &tp->t_ticket, XFS_TRANSACTION, - permanent, tp->t_type); + permanent, 0); } if (error) @@ -268,6 +226,42 @@ undo_blocks: return error; } +int +xfs_trans_alloc( + struct xfs_mount *mp, + struct xfs_trans_res *resp, + uint blocks, + uint rtextents, + uint flags, + struct xfs_trans **tpp) +{ + struct xfs_trans *tp; + int error; + + if (!(flags & XFS_TRANS_NO_WRITECOUNT)) + sb_start_intwrite(mp->m_super); + + WARN_ON(mp->m_super->s_writers.frozen == SB_FREEZE_COMPLETE); + atomic_inc(&mp->m_active_trans); + + tp = kmem_zone_zalloc(xfs_trans_zone, + (flags & XFS_TRANS_NOFS) ? KM_NOFS : KM_SLEEP); + tp->t_magic = XFS_TRANS_HEADER_MAGIC; + tp->t_flags = flags; + tp->t_mountp = mp; + INIT_LIST_HEAD(&tp->t_items); + INIT_LIST_HEAD(&tp->t_busy); + + error = xfs_trans_reserve(tp, resp, blocks, rtextents); + if (error) { + xfs_trans_cancel(tp); + return error; + } + + *tpp = tp; + return 0; +} + /* * Record the indicated change to the given field for application * to the file system's superblock when the transaction commits. diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h index e7c49cf43fbc..9a462e892e4f 100644 --- a/fs/xfs/xfs_trans.h +++ b/fs/xfs/xfs_trans.h @@ -90,7 +90,6 @@ void xfs_log_item_init(struct xfs_mount *mp, struct xfs_log_item *item, */ typedef struct xfs_trans { unsigned int t_magic; /* magic number */ - unsigned int t_type; /* transaction type */ unsigned int t_log_res; /* amt of log space resvd */ unsigned int t_log_count; /* count for perm log res */ unsigned int t_blk_res; /* # of blocks resvd */ @@ -148,10 +147,9 @@ typedef struct xfs_trans { /* * XFS transaction mechanism exported interfaces. */ -xfs_trans_t *xfs_trans_alloc(struct xfs_mount *, uint); -xfs_trans_t *_xfs_trans_alloc(struct xfs_mount *, uint, xfs_km_flags_t); -int xfs_trans_reserve(struct xfs_trans *, struct xfs_trans_res *, - uint, uint); +int xfs_trans_alloc(struct xfs_mount *mp, struct xfs_trans_res *resp, + uint blocks, uint rtextents, uint flags, + struct xfs_trans **tpp); void xfs_trans_mod_sb(xfs_trans_t *, uint, int64_t); struct xfs_buf *xfs_trans_get_buf_map(struct xfs_trans *tp, -- cgit v1.3-8-gc7d7 From 710b1e2c2948c1e5d0499def5273ecbc6472342d Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 6 Apr 2016 09:20:36 +1000 Subject: xfs: remove transaction types These aren't used for CIL-style logging and can be dropped. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_log_format.h | 5 +++ fs/xfs/libxfs/xfs_shared.h | 97 ------------------------------------------ fs/xfs/xfs_log.c | 58 +------------------------ fs/xfs/xfs_log.h | 3 +- fs/xfs/xfs_log_cil.c | 1 - fs/xfs/xfs_log_priv.h | 1 - fs/xfs/xfs_trace.h | 5 +-- fs/xfs/xfs_trans.c | 2 +- 8 files changed, 10 insertions(+), 162 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_log_format.h b/fs/xfs/libxfs/xfs_log_format.h index d54a8018b079..e8f49c029ff0 100644 --- a/fs/xfs/libxfs/xfs_log_format.h +++ b/fs/xfs/libxfs/xfs_log_format.h @@ -211,6 +211,11 @@ typedef struct xfs_trans_header { #define XFS_TRANS_HEADER_MAGIC 0x5452414e /* TRAN */ +/* + * The only type valid for th_type in CIL-enabled file system logs: + */ +#define XFS_TRANS_CHECKPOINT 40 + /* * Log item types. */ diff --git a/fs/xfs/libxfs/xfs_shared.h b/fs/xfs/libxfs/xfs_shared.h index 7d4ab6439081..16002b5ec4eb 100644 --- a/fs/xfs/libxfs/xfs_shared.h +++ b/fs/xfs/libxfs/xfs_shared.h @@ -55,103 +55,6 @@ extern const struct xfs_buf_ops xfs_sb_quiet_buf_ops; extern const struct xfs_buf_ops xfs_symlink_buf_ops; extern const struct xfs_buf_ops xfs_rtbuf_ops; -/* - * Transaction types. Used to distinguish types of buffers. These never reach - * the log. - */ -#define XFS_TRANS_SETATTR_NOT_SIZE 1 -#define XFS_TRANS_SETATTR_SIZE 2 -#define XFS_TRANS_INACTIVE 3 -#define XFS_TRANS_CREATE 4 -#define XFS_TRANS_CREATE_TRUNC 5 -#define XFS_TRANS_TRUNCATE_FILE 6 -#define XFS_TRANS_REMOVE 7 -#define XFS_TRANS_LINK 8 -#define XFS_TRANS_RENAME 9 -#define XFS_TRANS_MKDIR 10 -#define XFS_TRANS_RMDIR 11 -#define XFS_TRANS_SYMLINK 12 -#define XFS_TRANS_SET_DMATTRS 13 -#define XFS_TRANS_GROWFS 14 -#define XFS_TRANS_STRAT_WRITE 15 -#define XFS_TRANS_DIOSTRAT 16 -/* 17 was XFS_TRANS_WRITE_SYNC */ -#define XFS_TRANS_WRITEID 18 -#define XFS_TRANS_ADDAFORK 19 -#define XFS_TRANS_ATTRINVAL 20 -#define XFS_TRANS_ATRUNCATE 21 -#define XFS_TRANS_ATTR_SET 22 -#define XFS_TRANS_ATTR_RM 23 -#define XFS_TRANS_ATTR_FLAG 24 -#define XFS_TRANS_CLEAR_AGI_BUCKET 25 -#define XFS_TRANS_SB_CHANGE 26 -/* - * Dummy entries since we use the transaction type to index into the - * trans_type[] in xlog_recover_print_trans_head() - */ -#define XFS_TRANS_DUMMY1 27 -#define XFS_TRANS_DUMMY2 28 -#define XFS_TRANS_QM_QUOTAOFF 29 -#define XFS_TRANS_QM_DQALLOC 30 -#define XFS_TRANS_QM_SETQLIM 31 -#define XFS_TRANS_QM_DQCLUSTER 32 -#define XFS_TRANS_QM_QINOCREATE 33 -#define XFS_TRANS_QM_QUOTAOFF_END 34 -#define XFS_TRANS_FSYNC_TS 35 -#define XFS_TRANS_GROWFSRT_ALLOC 36 -#define XFS_TRANS_GROWFSRT_ZERO 37 -#define XFS_TRANS_GROWFSRT_FREE 38 -#define XFS_TRANS_SWAPEXT 39 -#define XFS_TRANS_CHECKPOINT 40 -#define XFS_TRANS_ICREATE 41 -#define XFS_TRANS_CREATE_TMPFILE 42 -#define XFS_TRANS_TYPE_MAX 43 -/* new transaction types need to be reflected in xfs_logprint(8) */ - -#define XFS_TRANS_TYPES \ - { XFS_TRANS_SETATTR_NOT_SIZE, "SETATTR_NOT_SIZE" }, \ - { XFS_TRANS_SETATTR_SIZE, "SETATTR_SIZE" }, \ - { XFS_TRANS_INACTIVE, "INACTIVE" }, \ - { XFS_TRANS_CREATE, "CREATE" }, \ - { XFS_TRANS_CREATE_TRUNC, "CREATE_TRUNC" }, \ - { XFS_TRANS_TRUNCATE_FILE, "TRUNCATE_FILE" }, \ - { XFS_TRANS_REMOVE, "REMOVE" }, \ - { XFS_TRANS_LINK, "LINK" }, \ - { XFS_TRANS_RENAME, "RENAME" }, \ - { XFS_TRANS_MKDIR, "MKDIR" }, \ - { XFS_TRANS_RMDIR, "RMDIR" }, \ - { XFS_TRANS_SYMLINK, "SYMLINK" }, \ - { XFS_TRANS_SET_DMATTRS, "SET_DMATTRS" }, \ - { XFS_TRANS_GROWFS, "GROWFS" }, \ - { XFS_TRANS_STRAT_WRITE, "STRAT_WRITE" }, \ - { XFS_TRANS_DIOSTRAT, "DIOSTRAT" }, \ - { XFS_TRANS_WRITEID, "WRITEID" }, \ - { XFS_TRANS_ADDAFORK, "ADDAFORK" }, \ - { XFS_TRANS_ATTRINVAL, "ATTRINVAL" }, \ - { XFS_TRANS_ATRUNCATE, "ATRUNCATE" }, \ - { XFS_TRANS_ATTR_SET, "ATTR_SET" }, \ - { XFS_TRANS_ATTR_RM, "ATTR_RM" }, \ - { XFS_TRANS_ATTR_FLAG, "ATTR_FLAG" }, \ - { XFS_TRANS_CLEAR_AGI_BUCKET, "CLEAR_AGI_BUCKET" }, \ - { XFS_TRANS_SB_CHANGE, "SBCHANGE" }, \ - { XFS_TRANS_DUMMY1, "DUMMY1" }, \ - { XFS_TRANS_DUMMY2, "DUMMY2" }, \ - { XFS_TRANS_QM_QUOTAOFF, "QM_QUOTAOFF" }, \ - { XFS_TRANS_QM_DQALLOC, "QM_DQALLOC" }, \ - { XFS_TRANS_QM_SETQLIM, "QM_SETQLIM" }, \ - { XFS_TRANS_QM_DQCLUSTER, "QM_DQCLUSTER" }, \ - { XFS_TRANS_QM_QINOCREATE, "QM_QINOCREATE" }, \ - { XFS_TRANS_QM_QUOTAOFF_END, "QM_QOFF_END" }, \ - { XFS_TRANS_FSYNC_TS, "FSYNC_TS" }, \ - { XFS_TRANS_GROWFSRT_ALLOC, "GROWFSRT_ALLOC" }, \ - { XFS_TRANS_GROWFSRT_ZERO, "GROWFSRT_ZERO" }, \ - { XFS_TRANS_GROWFSRT_FREE, "GROWFSRT_FREE" }, \ - { XFS_TRANS_SWAPEXT, "SWAPEXT" }, \ - { XFS_TRANS_CHECKPOINT, "CHECKPOINT" }, \ - { XFS_TRANS_ICREATE, "ICREATE" }, \ - { XFS_TRANS_CREATE_TMPFILE, "CREATE_TMPFILE" }, \ - { XLOG_UNMOUNT_REC_TYPE, "UNMOUNT" } - /* * This structure is used to track log items associated with * a transaction. It points to the log item and keeps some diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index b49ccf5c1d75..268f7d65cfd4 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -435,8 +435,7 @@ xfs_log_reserve( int cnt, struct xlog_ticket **ticp, __uint8_t client, - bool permanent, - uint t_type) + bool permanent) { struct xlog *log = mp->m_log; struct xlog_ticket *tic; @@ -456,7 +455,6 @@ xfs_log_reserve( if (!tic) return -ENOMEM; - tic->t_trans_type = t_type; *ticp = tic; xlog_grant_push_ail(log, tic->t_cnt ? tic->t_unit_res * tic->t_cnt @@ -823,8 +821,7 @@ xfs_log_unmount_write(xfs_mount_t *mp) } while (iclog != first_iclog); #endif if (! (XLOG_FORCED_SHUTDOWN(log))) { - error = xfs_log_reserve(mp, 600, 1, &tic, - XFS_LOG, 0, XLOG_UNMOUNT_REC_TYPE); + error = xfs_log_reserve(mp, 600, 1, &tic, XFS_LOG, 0); if (!error) { /* the data section must be 32 bit size aligned */ struct { @@ -2032,58 +2029,8 @@ xlog_print_tic_res( REG_TYPE_STR(ICREATE, "inode create") }; #undef REG_TYPE_STR -#define TRANS_TYPE_STR(type) [XFS_TRANS_##type] = #type - static char *trans_type_str[XFS_TRANS_TYPE_MAX] = { - TRANS_TYPE_STR(SETATTR_NOT_SIZE), - TRANS_TYPE_STR(SETATTR_SIZE), - TRANS_TYPE_STR(INACTIVE), - TRANS_TYPE_STR(CREATE), - TRANS_TYPE_STR(CREATE_TRUNC), - TRANS_TYPE_STR(TRUNCATE_FILE), - TRANS_TYPE_STR(REMOVE), - TRANS_TYPE_STR(LINK), - TRANS_TYPE_STR(RENAME), - TRANS_TYPE_STR(MKDIR), - TRANS_TYPE_STR(RMDIR), - TRANS_TYPE_STR(SYMLINK), - TRANS_TYPE_STR(SET_DMATTRS), - TRANS_TYPE_STR(GROWFS), - TRANS_TYPE_STR(STRAT_WRITE), - TRANS_TYPE_STR(DIOSTRAT), - TRANS_TYPE_STR(WRITEID), - TRANS_TYPE_STR(ADDAFORK), - TRANS_TYPE_STR(ATTRINVAL), - TRANS_TYPE_STR(ATRUNCATE), - TRANS_TYPE_STR(ATTR_SET), - TRANS_TYPE_STR(ATTR_RM), - TRANS_TYPE_STR(ATTR_FLAG), - TRANS_TYPE_STR(CLEAR_AGI_BUCKET), - TRANS_TYPE_STR(SB_CHANGE), - TRANS_TYPE_STR(DUMMY1), - TRANS_TYPE_STR(DUMMY2), - TRANS_TYPE_STR(QM_QUOTAOFF), - TRANS_TYPE_STR(QM_DQALLOC), - TRANS_TYPE_STR(QM_SETQLIM), - TRANS_TYPE_STR(QM_DQCLUSTER), - TRANS_TYPE_STR(QM_QINOCREATE), - TRANS_TYPE_STR(QM_QUOTAOFF_END), - TRANS_TYPE_STR(FSYNC_TS), - TRANS_TYPE_STR(GROWFSRT_ALLOC), - TRANS_TYPE_STR(GROWFSRT_ZERO), - TRANS_TYPE_STR(GROWFSRT_FREE), - TRANS_TYPE_STR(SWAPEXT), - TRANS_TYPE_STR(CHECKPOINT), - TRANS_TYPE_STR(ICREATE), - TRANS_TYPE_STR(CREATE_TMPFILE) - }; -#undef TRANS_TYPE_STR xfs_warn(mp, "xlog_write: reservation summary:"); - xfs_warn(mp, " trans type = %s (%u)", - ((ticket->t_trans_type <= 0 || - ticket->t_trans_type > XFS_TRANS_TYPE_MAX) ? - "bad-trans-type" : trans_type_str[ticket->t_trans_type]), - ticket->t_trans_type); xfs_warn(mp, " unit res = %d bytes", ticket->t_unit_res); xfs_warn(mp, " current res = %d bytes", @@ -3709,7 +3656,6 @@ xlog_ticket_alloc( tic->t_tid = prandom_u32(); tic->t_clientid = client; tic->t_flags = XLOG_TIC_INITED; - tic->t_trans_type = 0; if (permanent) tic->t_flags |= XLOG_TIC_PERM_RESERV; diff --git a/fs/xfs/xfs_log.h b/fs/xfs/xfs_log.h index aa533a7d50f2..80ba0c047090 100644 --- a/fs/xfs/xfs_log.h +++ b/fs/xfs/xfs_log.h @@ -161,8 +161,7 @@ int xfs_log_reserve(struct xfs_mount *mp, int count, struct xlog_ticket **ticket, __uint8_t clientid, - bool permanent, - uint t_type); + bool permanent); int xfs_log_regrant(struct xfs_mount *mp, struct xlog_ticket *tic); int xfs_log_unmount_write(struct xfs_mount *mp); void xfs_log_unmount(struct xfs_mount *mp); diff --git a/fs/xfs/xfs_log_cil.c b/fs/xfs/xfs_log_cil.c index 4e7649351f5a..5e54e7955ea6 100644 --- a/fs/xfs/xfs_log_cil.c +++ b/fs/xfs/xfs_log_cil.c @@ -51,7 +51,6 @@ xlog_cil_ticket_alloc( tic = xlog_ticket_alloc(log, 0, 1, XFS_TRANSACTION, 0, KM_SLEEP|KM_NOFS); - tic->t_trans_type = XFS_TRANS_CHECKPOINT; /* * set the current reservation to zero so we know to steal the basic diff --git a/fs/xfs/xfs_log_priv.h b/fs/xfs/xfs_log_priv.h index ed8896310c00..765f084759b5 100644 --- a/fs/xfs/xfs_log_priv.h +++ b/fs/xfs/xfs_log_priv.h @@ -175,7 +175,6 @@ typedef struct xlog_ticket { char t_cnt; /* current count : 1 */ char t_clientid; /* who does this belong to; : 1 */ char t_flags; /* properties of reservation : 1 */ - uint t_trans_type; /* transaction type : 4 */ /* reservation array fields */ uint t_res_num; /* num in array : 4 */ diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h index c8d58426008e..f08129444280 100644 --- a/fs/xfs/xfs_trace.h +++ b/fs/xfs/xfs_trace.h @@ -944,7 +944,6 @@ DECLARE_EVENT_CLASS(xfs_loggrant_class, TP_ARGS(log, tic), TP_STRUCT__entry( __field(dev_t, dev) - __field(unsigned, trans_type) __field(char, ocnt) __field(char, cnt) __field(int, curr_res) @@ -962,7 +961,6 @@ DECLARE_EVENT_CLASS(xfs_loggrant_class, ), TP_fast_assign( __entry->dev = log->l_mp->m_super->s_dev; - __entry->trans_type = tic->t_trans_type; __entry->ocnt = tic->t_ocnt; __entry->cnt = tic->t_cnt; __entry->curr_res = tic->t_curr_res; @@ -980,14 +978,13 @@ DECLARE_EVENT_CLASS(xfs_loggrant_class, __entry->curr_block = log->l_curr_block; __entry->tail_lsn = atomic64_read(&log->l_tail_lsn); ), - TP_printk("dev %d:%d type %s t_ocnt %u t_cnt %u t_curr_res %u " + TP_printk("dev %d:%d t_ocnt %u t_cnt %u t_curr_res %u " "t_unit_res %u t_flags %s reserveq %s " "writeq %s grant_reserve_cycle %d " "grant_reserve_bytes %d grant_write_cycle %d " "grant_write_bytes %d curr_cycle %d curr_block %d " "tail_cycle %d tail_block %d", MAJOR(__entry->dev), MINOR(__entry->dev), - __print_symbolic(__entry->trans_type, XFS_TRANS_TYPES), __entry->ocnt, __entry->cnt, __entry->curr_res, diff --git a/fs/xfs/xfs_trans.c b/fs/xfs/xfs_trans.c index b3669efb2a0a..5f3d33d16e67 100644 --- a/fs/xfs/xfs_trans.c +++ b/fs/xfs/xfs_trans.c @@ -177,7 +177,7 @@ xfs_trans_reserve( resp->tr_logres, resp->tr_logcount, &tp->t_ticket, XFS_TRANSACTION, - permanent, 0); + permanent); } if (error) -- cgit v1.3-8-gc7d7 From 9f27889f3a96ff356ac92688cc0c4be3935ae3af Mon Sep 17 00:00:00 2001 From: Carlos Maiolino Date: Wed, 6 Apr 2016 09:46:30 +1000 Subject: xfs: Add caller function output to xfs_log_force tracepoint I had sent this patch yesterday, but for some reason it didn't reach xfs list, sending again. Output the caller of xfs_log_force might be useful when tracing log checkpoint problems without the need to build kernel with DEBUG. Signed-off-by: Carlos Maiolino Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/xfs_log.c | 4 ++-- fs/xfs/xfs_trace.h | 10 ++++++---- 2 files changed, 8 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index b49ccf5c1d75..8c08015bc094 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -3378,7 +3378,7 @@ xfs_log_force( { int error; - trace_xfs_log_force(mp, 0); + trace_xfs_log_force(mp, 0, _RET_IP_); error = _xfs_log_force(mp, flags, NULL); if (error) xfs_warn(mp, "%s: error %d returned.", __func__, error); @@ -3527,7 +3527,7 @@ xfs_log_force_lsn( { int error; - trace_xfs_log_force(mp, lsn); + trace_xfs_log_force(mp, lsn, _RET_IP_); error = _xfs_log_force_lsn(mp, lsn, flags, NULL); if (error) xfs_warn(mp, "%s: error %d returned.", __func__, error); diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h index c8d58426008e..384bb1791481 100644 --- a/fs/xfs/xfs_trace.h +++ b/fs/xfs/xfs_trace.h @@ -1053,19 +1053,21 @@ DECLARE_EVENT_CLASS(xfs_log_item_class, ) TRACE_EVENT(xfs_log_force, - TP_PROTO(struct xfs_mount *mp, xfs_lsn_t lsn), - TP_ARGS(mp, lsn), + TP_PROTO(struct xfs_mount *mp, xfs_lsn_t lsn, unsigned long caller_ip), + TP_ARGS(mp, lsn, caller_ip), TP_STRUCT__entry( __field(dev_t, dev) __field(xfs_lsn_t, lsn) + __field(unsigned long, caller_ip) ), TP_fast_assign( __entry->dev = mp->m_super->s_dev; __entry->lsn = lsn; + __entry->caller_ip = caller_ip; ), - TP_printk("dev %d:%d lsn 0x%llx", + TP_printk("dev %d:%d lsn 0x%llx caller %ps", MAJOR(__entry->dev), MINOR(__entry->dev), - __entry->lsn) + __entry->lsn, (void *)__entry->caller_ip) ) #define DEFINE_LOG_ITEM_EVENT(name) \ -- cgit v1.3-8-gc7d7 From 664b60f6babc98ee03c2ff15b9482cc8c5e15a83 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 6 Apr 2016 09:47:01 +1000 Subject: xfs: improve kmem_realloc Use krealloc to implement our realloc function. This helps to avoid new allocations if we are still in the slab bucket. At least for the bmap btree root that's actually the common case. This also allows removing the now unused oldsize argument. Signed-off-by: Christoph Hellwig Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/kmem.c | 26 +++++++++++++++----------- fs/xfs/kmem.h | 2 +- fs/xfs/libxfs/xfs_inode_fork.c | 10 +++------- fs/xfs/xfs_log_recover.c | 2 +- fs/xfs/xfs_mount.c | 1 - 5 files changed, 20 insertions(+), 21 deletions(-) (limited to 'fs') diff --git a/fs/xfs/kmem.c b/fs/xfs/kmem.c index 686ba6fb20dd..339c696bbc01 100644 --- a/fs/xfs/kmem.c +++ b/fs/xfs/kmem.c @@ -93,19 +93,23 @@ kmem_zalloc_large(size_t size, xfs_km_flags_t flags) } void * -kmem_realloc(const void *ptr, size_t newsize, size_t oldsize, - xfs_km_flags_t flags) +kmem_realloc(const void *old, size_t newsize, xfs_km_flags_t flags) { - void *new; + int retries = 0; + gfp_t lflags = kmem_flags_convert(flags); + void *ptr; - new = kmem_alloc(newsize, flags); - if (ptr) { - if (new) - memcpy(new, ptr, - ((oldsize < newsize) ? oldsize : newsize)); - kmem_free(ptr); - } - return new; + do { + ptr = krealloc(old, newsize, lflags); + if (ptr || (flags & (KM_MAYFAIL|KM_NOSLEEP))) + return ptr; + if (!(++retries % 100)) + xfs_err(NULL, + "%s(%u) possible memory allocation deadlock size %zu in %s (mode:0x%x)", + current->comm, current->pid, + newsize, __func__, lflags); + congestion_wait(BLK_RW_ASYNC, HZ/50); + } while (1); } void * diff --git a/fs/xfs/kmem.h b/fs/xfs/kmem.h index d1c66e465ca5..689f746224e7 100644 --- a/fs/xfs/kmem.h +++ b/fs/xfs/kmem.h @@ -62,7 +62,7 @@ kmem_flags_convert(xfs_km_flags_t flags) extern void *kmem_alloc(size_t, xfs_km_flags_t); extern void *kmem_zalloc_large(size_t size, xfs_km_flags_t); -extern void *kmem_realloc(const void *, size_t, size_t, xfs_km_flags_t); +extern void *kmem_realloc(const void *, size_t, xfs_km_flags_t); static inline void kmem_free(const void *ptr) { kvfree(ptr); diff --git a/fs/xfs/libxfs/xfs_inode_fork.c b/fs/xfs/libxfs/xfs_inode_fork.c index 11faf7df14c8..1c842dce44b6 100644 --- a/fs/xfs/libxfs/xfs_inode_fork.c +++ b/fs/xfs/libxfs/xfs_inode_fork.c @@ -516,7 +516,6 @@ xfs_iroot_realloc( new_max = cur_max + rec_diff; new_size = XFS_BMAP_BROOT_SPACE_CALC(mp, new_max); ifp->if_broot = kmem_realloc(ifp->if_broot, new_size, - XFS_BMAP_BROOT_SPACE_CALC(mp, cur_max), KM_SLEEP | KM_NOFS); op = (char *)XFS_BMAP_BROOT_PTR_ADDR(mp, ifp->if_broot, 1, ifp->if_broot_bytes); @@ -660,7 +659,6 @@ xfs_idata_realloc( ifp->if_u1.if_data = kmem_realloc(ifp->if_u1.if_data, real_size, - ifp->if_real_bytes, KM_SLEEP | KM_NOFS); } } else { @@ -1376,8 +1374,7 @@ xfs_iext_realloc_direct( if (rnew_size != ifp->if_real_bytes) { ifp->if_u1.if_extents = kmem_realloc(ifp->if_u1.if_extents, - rnew_size, - ifp->if_real_bytes, KM_NOFS); + rnew_size, KM_NOFS); } if (rnew_size > ifp->if_real_bytes) { memset(&ifp->if_u1.if_extents[ifp->if_bytes / @@ -1461,9 +1458,8 @@ xfs_iext_realloc_indirect( if (new_size == 0) { xfs_iext_destroy(ifp); } else { - ifp->if_u1.if_ext_irec = (xfs_ext_irec_t *) - kmem_realloc(ifp->if_u1.if_ext_irec, - new_size, size, KM_NOFS); + ifp->if_u1.if_ext_irec = + kmem_realloc(ifp->if_u1.if_ext_irec, new_size, KM_NOFS); } } diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index 396565f43247..bf6e80703613 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -3843,7 +3843,7 @@ xlog_recover_add_to_cont_trans( old_ptr = item->ri_buf[item->ri_cnt-1].i_addr; old_len = item->ri_buf[item->ri_cnt-1].i_len; - ptr = kmem_realloc(old_ptr, len+old_len, old_len, KM_SLEEP); + ptr = kmem_realloc(old_ptr, len + old_len, KM_SLEEP); memcpy(&ptr[old_len], dp, len); item->ri_buf[item->ri_cnt-1].i_len += len; item->ri_buf[item->ri_cnt-1].i_addr = ptr; diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index 536a0ee9cd5a..654799f716fc 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -89,7 +89,6 @@ xfs_uuid_mount( if (hole < 0) { xfs_uuid_table = kmem_realloc(xfs_uuid_table, (xfs_uuid_table_size + 1) * sizeof(*xfs_uuid_table), - xfs_uuid_table_size * sizeof(*xfs_uuid_table), KM_SLEEP); hole = xfs_uuid_table_size++; } -- cgit v1.3-8-gc7d7 From 6e3e6d55e51774ec7cfc24975749bbddb28a9051 Mon Sep 17 00:00:00 2001 From: Eryu Guan Date: Wed, 6 Apr 2016 09:47:21 +1000 Subject: xfs: mute some sparse warnings These three warnings are fixed: fs/xfs/xfs_inode.c:1033:44: warning: Using plain integer as NULL pointer fs/xfs/xfs_inode_item.c:525:20: warning: context imbalance in 'xfs_inode_item_push' - unexpected unlock fs/xfs/xfs_dquot.c:696:1: warning: symbol 'xfs_dq_get_next_id' was not declared. Should it be static? Signed-off-by: Eryu Guan Reviewed-by: Christoph Hellwig Signed-off-by: Dave Chinner --- fs/xfs/xfs_dquot.c | 2 +- fs/xfs/xfs_inode.c | 2 +- fs/xfs/xfs_inode_item.c | 2 ++ 3 files changed, 4 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_dquot.c b/fs/xfs/xfs_dquot.c index 316b2a1bdba5..8f51370c95c4 100644 --- a/fs/xfs/xfs_dquot.c +++ b/fs/xfs/xfs_dquot.c @@ -692,7 +692,7 @@ error0: * end of the chunk, skip ahead to first id in next allocated chunk * using the SEEK_DATA interface. */ -int +static int xfs_dq_get_next_id( xfs_mount_t *mp, uint type, diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index 96f606deee31..1445a99a6868 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -1030,7 +1030,7 @@ xfs_dir_ialloc( tp->t_flags &= ~(XFS_TRANS_DQ_DIRTY); } - code = xfs_trans_roll(&tp, 0); + code = xfs_trans_roll(&tp, NULL); if (committed != NULL) *committed = 1; diff --git a/fs/xfs/xfs_inode_item.c b/fs/xfs/xfs_inode_item.c index c48b5b18d771..d02cbab18f45 100644 --- a/fs/xfs/xfs_inode_item.c +++ b/fs/xfs/xfs_inode_item.c @@ -479,6 +479,8 @@ STATIC uint xfs_inode_item_push( struct xfs_log_item *lip, struct list_head *buffer_list) + __releases(&lip->li_ailp->xa_lock) + __acquires(&lip->li_ailp->xa_lock) { struct xfs_inode_log_item *iip = INODE_ITEM(lip); struct xfs_inode *ip = iip->ili_inode; -- cgit v1.3-8-gc7d7 From 5ac7eace2d00eab5ae0e9fdee63e38aee6001f7c Mon Sep 17 00:00:00 2001 From: David Howells Date: Wed, 6 Apr 2016 16:14:24 +0100 Subject: KEYS: Add a facility to restrict new links into a keyring Add a facility whereby proposed new links to be added to a keyring can be vetted, permitting them to be rejected if necessary. This can be used to block public keys from which the signature cannot be verified or for which the signature verification fails. It could also be used to provide blacklisting. This affects operations like add_key(), KEYCTL_LINK and KEYCTL_INSTANTIATE. To this end: (1) A function pointer is added to the key struct that, if set, points to the vetting function. This is called as: int (*restrict_link)(struct key *keyring, const struct key_type *key_type, unsigned long key_flags, const union key_payload *key_payload), where 'keyring' will be the keyring being added to, key_type and key_payload will describe the key being added and key_flags[*] can be AND'ed with KEY_FLAG_TRUSTED. [*] This parameter will be removed in a later patch when KEY_FLAG_TRUSTED is removed. The function should return 0 to allow the link to take place or an error (typically -ENOKEY, -ENOPKG or -EKEYREJECTED) to reject the link. The pointer should not be set directly, but rather should be set through keyring_alloc(). Note that if called during add_key(), preparse is called before this method, but a key isn't actually allocated until after this function is called. (2) KEY_ALLOC_BYPASS_RESTRICTION is added. This can be passed to key_create_or_update() or key_instantiate_and_link() to bypass the restriction check. (3) KEY_FLAG_TRUSTED_ONLY is removed. The entire contents of a keyring with this restriction emplaced can be considered 'trustworthy' by virtue of being in the keyring when that keyring is consulted. (4) key_alloc() and keyring_alloc() take an extra argument that will be used to set restrict_link in the new key. This ensures that the pointer is set before the key is published, thus preventing a window of unrestrictedness. Normally this argument will be NULL. (5) As a temporary affair, keyring_restrict_trusted_only() is added. It should be passed to keyring_alloc() as the extra argument instead of setting KEY_FLAG_TRUSTED_ONLY on a keyring. This will be replaced in a later patch with functions that look in the appropriate places for authoritative keys. Signed-off-by: David Howells Reviewed-by: Mimi Zohar --- Documentation/security/keys.txt | 22 ++++++++++++ certs/system_keyring.c | 8 ++--- fs/cifs/cifsacl.c | 2 +- fs/nfs/nfs4idmap.c | 2 +- include/linux/key.h | 53 +++++++++++++++++++++++------ net/dns_resolver/dns_key.c | 2 +- net/rxrpc/ar-key.c | 4 +-- security/integrity/digsig.c | 7 ++-- security/integrity/ima/ima_mok.c | 8 ++--- security/keys/key.c | 43 +++++++++++++++++++---- security/keys/keyring.c | 73 ++++++++++++++++++++++++++++++++++++---- security/keys/persistent.c | 4 +-- security/keys/process_keys.c | 16 +++++---- security/keys/request_key.c | 4 +-- security/keys/request_key_auth.c | 2 +- 15 files changed, 198 insertions(+), 52 deletions(-) (limited to 'fs') diff --git a/Documentation/security/keys.txt b/Documentation/security/keys.txt index 8c183873b2b7..a6a50b359025 100644 --- a/Documentation/security/keys.txt +++ b/Documentation/security/keys.txt @@ -999,6 +999,10 @@ payload contents" for more information. struct key *keyring_alloc(const char *description, uid_t uid, gid_t gid, const struct cred *cred, key_perm_t perm, + int (*restrict_link)(struct key *, + const struct key_type *, + unsigned long, + const union key_payload *), unsigned long flags, struct key *dest); @@ -1010,6 +1014,24 @@ payload contents" for more information. KEY_ALLOC_NOT_IN_QUOTA in flags if the keyring shouldn't be accounted towards the user's quota). Error ENOMEM can also be returned. + If restrict_link not NULL, it should point to a function that will be + called each time an attempt is made to link a key into the new keyring. + This function is called to check whether a key may be added into the keying + or not. Callers of key_create_or_update() within the kernel can pass + KEY_ALLOC_BYPASS_RESTRICTION to suppress the check. An example of using + this is to manage rings of cryptographic keys that are set up when the + kernel boots where userspace is also permitted to add keys - provided they + can be verified by a key the kernel already has. + + When called, the restriction function will be passed the keyring being + added to, the key flags value and the type and payload of the key being + added. Note that when a new key is being created, this is called between + payload preparsing and actual key creation. The function should return 0 + to allow the link or an error to reject it. + + A convenience function, restrict_link_reject, exists to always return + -EPERM to in this case. + (*) To check the validity of a key, this function can be called: diff --git a/certs/system_keyring.c b/certs/system_keyring.c index dc18869ff680..417d65882870 100644 --- a/certs/system_keyring.c +++ b/certs/system_keyring.c @@ -36,11 +36,10 @@ static __init int system_trusted_keyring_init(void) KUIDT_INIT(0), KGIDT_INIT(0), current_cred(), ((KEY_POS_ALL & ~KEY_POS_SETATTR) | KEY_USR_VIEW | KEY_USR_READ | KEY_USR_SEARCH), - KEY_ALLOC_NOT_IN_QUOTA, NULL); + KEY_ALLOC_NOT_IN_QUOTA, + keyring_restrict_trusted_only, NULL); if (IS_ERR(system_trusted_keyring)) panic("Can't allocate system trusted keyring\n"); - - set_bit(KEY_FLAG_TRUSTED_ONLY, &system_trusted_keyring->flags); return 0; } @@ -85,7 +84,8 @@ static __init int load_system_certificate_list(void) KEY_USR_VIEW | KEY_USR_READ), KEY_ALLOC_NOT_IN_QUOTA | KEY_ALLOC_TRUSTED | - KEY_ALLOC_BUILT_IN); + KEY_ALLOC_BUILT_IN | + KEY_ALLOC_BYPASS_RESTRICTION); if (IS_ERR(key)) { pr_err("Problem loading in-kernel X.509 certificate (%ld)\n", PTR_ERR(key)); diff --git a/fs/cifs/cifsacl.c b/fs/cifs/cifsacl.c index 3f93125916bf..71e8a56e9479 100644 --- a/fs/cifs/cifsacl.c +++ b/fs/cifs/cifsacl.c @@ -360,7 +360,7 @@ init_cifs_idmap(void) GLOBAL_ROOT_UID, GLOBAL_ROOT_GID, cred, (KEY_POS_ALL & ~KEY_POS_SETATTR) | KEY_USR_VIEW | KEY_USR_READ, - KEY_ALLOC_NOT_IN_QUOTA, NULL); + KEY_ALLOC_NOT_IN_QUOTA, NULL, NULL); if (IS_ERR(keyring)) { ret = PTR_ERR(keyring); goto failed_put_cred; diff --git a/fs/nfs/nfs4idmap.c b/fs/nfs/nfs4idmap.c index 5ba22c6b0ffa..c444285bb1b1 100644 --- a/fs/nfs/nfs4idmap.c +++ b/fs/nfs/nfs4idmap.c @@ -201,7 +201,7 @@ int nfs_idmap_init(void) GLOBAL_ROOT_UID, GLOBAL_ROOT_GID, cred, (KEY_POS_ALL & ~KEY_POS_SETATTR) | KEY_USR_VIEW | KEY_USR_READ, - KEY_ALLOC_NOT_IN_QUOTA, NULL); + KEY_ALLOC_NOT_IN_QUOTA, NULL, NULL); if (IS_ERR(keyring)) { ret = PTR_ERR(keyring); goto failed_put_cred; diff --git a/include/linux/key.h b/include/linux/key.h index 5f5b1129dc92..83b603639d2e 100644 --- a/include/linux/key.h +++ b/include/linux/key.h @@ -174,10 +174,9 @@ struct key { #define KEY_FLAG_ROOT_CAN_CLEAR 6 /* set if key can be cleared by root without permission */ #define KEY_FLAG_INVALIDATED 7 /* set if key has been invalidated */ #define KEY_FLAG_TRUSTED 8 /* set if key is trusted */ -#define KEY_FLAG_TRUSTED_ONLY 9 /* set if keyring only accepts links to trusted keys */ -#define KEY_FLAG_BUILTIN 10 /* set if key is builtin */ -#define KEY_FLAG_ROOT_CAN_INVAL 11 /* set if key can be invalidated by root without permission */ -#define KEY_FLAG_KEEP 12 /* set if key should not be removed */ +#define KEY_FLAG_BUILTIN 9 /* set if key is built in to the kernel */ +#define KEY_FLAG_ROOT_CAN_INVAL 10 /* set if key can be invalidated by root without permission */ +#define KEY_FLAG_KEEP 11 /* set if key should not be removed */ /* the key type and key description string * - the desc is used to match a key against search criteria @@ -205,6 +204,21 @@ struct key { }; int reject_error; }; + + /* This is set on a keyring to restrict the addition of a link to a key + * to it. If this method isn't provided then it is assumed that the + * keyring is open to any addition. It is ignored for non-keyring + * keys. + * + * This is intended for use with rings of trusted keys whereby addition + * to the keyring needs to be controlled. KEY_ALLOC_BYPASS_RESTRICTION + * overrides this, allowing the kernel to add extra keys without + * restriction. + */ + int (*restrict_link)(struct key *keyring, + const struct key_type *type, + unsigned long flags, + const union key_payload *payload); }; extern struct key *key_alloc(struct key_type *type, @@ -212,14 +226,19 @@ extern struct key *key_alloc(struct key_type *type, kuid_t uid, kgid_t gid, const struct cred *cred, key_perm_t perm, - unsigned long flags); + unsigned long flags, + int (*restrict_link)(struct key *, + const struct key_type *, + unsigned long, + const union key_payload *)); -#define KEY_ALLOC_IN_QUOTA 0x0000 /* add to quota, reject if would overrun */ -#define KEY_ALLOC_QUOTA_OVERRUN 0x0001 /* add to quota, permit even if overrun */ -#define KEY_ALLOC_NOT_IN_QUOTA 0x0002 /* not in quota */ -#define KEY_ALLOC_TRUSTED 0x0004 /* Key should be flagged as trusted */ -#define KEY_ALLOC_BUILT_IN 0x0008 /* Key is built into kernel */ +#define KEY_ALLOC_IN_QUOTA 0x0000 /* add to quota, reject if would overrun */ +#define KEY_ALLOC_QUOTA_OVERRUN 0x0001 /* add to quota, permit even if overrun */ +#define KEY_ALLOC_NOT_IN_QUOTA 0x0002 /* not in quota */ +#define KEY_ALLOC_TRUSTED 0x0004 /* Key should be flagged as trusted */ +#define KEY_ALLOC_BUILT_IN 0x0008 /* Key is built into kernel */ +#define KEY_ALLOC_BYPASS_RESTRICTION 0x0010 /* Override the check on restricted keyrings */ extern void key_revoke(struct key *key); extern void key_invalidate(struct key *key); @@ -288,8 +307,22 @@ extern struct key *keyring_alloc(const char *description, kuid_t uid, kgid_t gid const struct cred *cred, key_perm_t perm, unsigned long flags, + int (*restrict_link)(struct key *, + const struct key_type *, + unsigned long, + const union key_payload *), struct key *dest); +extern int keyring_restrict_trusted_only(struct key *keyring, + const struct key_type *type, + unsigned long, + const union key_payload *payload); + +extern int restrict_link_reject(struct key *keyring, + const struct key_type *type, + unsigned long flags, + const union key_payload *payload); + extern int keyring_clear(struct key *keyring); extern key_ref_t keyring_search(key_ref_t keyring, diff --git a/net/dns_resolver/dns_key.c b/net/dns_resolver/dns_key.c index c79b85eb4d4c..8737412c7b27 100644 --- a/net/dns_resolver/dns_key.c +++ b/net/dns_resolver/dns_key.c @@ -281,7 +281,7 @@ static int __init init_dns_resolver(void) GLOBAL_ROOT_UID, GLOBAL_ROOT_GID, cred, (KEY_POS_ALL & ~KEY_POS_SETATTR) | KEY_USR_VIEW | KEY_USR_READ, - KEY_ALLOC_NOT_IN_QUOTA, NULL); + KEY_ALLOC_NOT_IN_QUOTA, NULL, NULL); if (IS_ERR(keyring)) { ret = PTR_ERR(keyring); goto failed_put_cred; diff --git a/net/rxrpc/ar-key.c b/net/rxrpc/ar-key.c index 3fb492eedeb9..1021b4c0bdd2 100644 --- a/net/rxrpc/ar-key.c +++ b/net/rxrpc/ar-key.c @@ -965,7 +965,7 @@ int rxrpc_get_server_data_key(struct rxrpc_connection *conn, key = key_alloc(&key_type_rxrpc, "x", GLOBAL_ROOT_UID, GLOBAL_ROOT_GID, cred, 0, - KEY_ALLOC_NOT_IN_QUOTA); + KEY_ALLOC_NOT_IN_QUOTA, NULL); if (IS_ERR(key)) { _leave(" = -ENOMEM [alloc %ld]", PTR_ERR(key)); return -ENOMEM; @@ -1012,7 +1012,7 @@ struct key *rxrpc_get_null_key(const char *keyname) key = key_alloc(&key_type_rxrpc, keyname, GLOBAL_ROOT_UID, GLOBAL_ROOT_GID, cred, - KEY_POS_SEARCH, KEY_ALLOC_NOT_IN_QUOTA); + KEY_POS_SEARCH, KEY_ALLOC_NOT_IN_QUOTA, NULL); if (IS_ERR(key)) return key; diff --git a/security/integrity/digsig.c b/security/integrity/digsig.c index 8ef15118cc78..659566c2200b 100644 --- a/security/integrity/digsig.c +++ b/security/integrity/digsig.c @@ -83,10 +83,9 @@ int __init integrity_init_keyring(const unsigned int id) ((KEY_POS_ALL & ~KEY_POS_SETATTR) | KEY_USR_VIEW | KEY_USR_READ | KEY_USR_WRITE | KEY_USR_SEARCH), - KEY_ALLOC_NOT_IN_QUOTA, NULL); - if (!IS_ERR(keyring[id])) - set_bit(KEY_FLAG_TRUSTED_ONLY, &keyring[id]->flags); - else { + KEY_ALLOC_NOT_IN_QUOTA, + NULL, NULL); + if (IS_ERR(keyring[id])) { err = PTR_ERR(keyring[id]); pr_info("Can't allocate %s keyring (%d)\n", keyring_name[id], err); diff --git a/security/integrity/ima/ima_mok.c b/security/integrity/ima/ima_mok.c index 676885e4320e..ef91248cb934 100644 --- a/security/integrity/ima/ima_mok.c +++ b/security/integrity/ima/ima_mok.c @@ -35,20 +35,20 @@ __init int ima_mok_init(void) (KEY_POS_ALL & ~KEY_POS_SETATTR) | KEY_USR_VIEW | KEY_USR_READ | KEY_USR_WRITE | KEY_USR_SEARCH, - KEY_ALLOC_NOT_IN_QUOTA, NULL); + KEY_ALLOC_NOT_IN_QUOTA, + keyring_restrict_trusted_only, NULL); ima_blacklist_keyring = keyring_alloc(".ima_blacklist", KUIDT_INIT(0), KGIDT_INIT(0), current_cred(), (KEY_POS_ALL & ~KEY_POS_SETATTR) | KEY_USR_VIEW | KEY_USR_READ | KEY_USR_WRITE | KEY_USR_SEARCH, - KEY_ALLOC_NOT_IN_QUOTA, NULL); + KEY_ALLOC_NOT_IN_QUOTA, + keyring_restrict_trusted_only, NULL); if (IS_ERR(ima_mok_keyring) || IS_ERR(ima_blacklist_keyring)) panic("Can't allocate IMA MOK or blacklist keyrings."); - set_bit(KEY_FLAG_TRUSTED_ONLY, &ima_mok_keyring->flags); - set_bit(KEY_FLAG_TRUSTED_ONLY, &ima_blacklist_keyring->flags); set_bit(KEY_FLAG_KEEP, &ima_blacklist_keyring->flags); return 0; } diff --git a/security/keys/key.c b/security/keys/key.c index b28755131687..deb881754e03 100644 --- a/security/keys/key.c +++ b/security/keys/key.c @@ -201,6 +201,7 @@ serial_exists: * @cred: The credentials specifying UID namespace. * @perm: The permissions mask of the new key. * @flags: Flags specifying quota properties. + * @restrict_link: Optional link restriction method for new keyrings. * * Allocate a key of the specified type with the attributes given. The key is * returned in an uninstantiated state and the caller needs to instantiate the @@ -223,7 +224,11 @@ serial_exists: */ struct key *key_alloc(struct key_type *type, const char *desc, kuid_t uid, kgid_t gid, const struct cred *cred, - key_perm_t perm, unsigned long flags) + key_perm_t perm, unsigned long flags, + int (*restrict_link)(struct key *, + const struct key_type *, + unsigned long, + const union key_payload *)) { struct key_user *user = NULL; struct key *key; @@ -291,6 +296,7 @@ struct key *key_alloc(struct key_type *type, const char *desc, key->uid = uid; key->gid = gid; key->perm = perm; + key->restrict_link = restrict_link; if (!(flags & KEY_ALLOC_NOT_IN_QUOTA)) key->flags |= 1 << KEY_FLAG_IN_QUOTA; @@ -496,6 +502,12 @@ int key_instantiate_and_link(struct key *key, } if (keyring) { + if (keyring->restrict_link) { + ret = keyring->restrict_link(keyring, key->type, + key->flags, &prep.payload); + if (ret < 0) + goto error; + } ret = __key_link_begin(keyring, &key->index_key, &edit); if (ret < 0) goto error; @@ -551,8 +563,12 @@ int key_reject_and_link(struct key *key, awaken = 0; ret = -EBUSY; - if (keyring) + if (keyring) { + if (keyring->restrict_link) + return -EPERM; + link_ret = __key_link_begin(keyring, &key->index_key, &edit); + } mutex_lock(&key_construction_mutex); @@ -793,6 +809,10 @@ key_ref_t key_create_or_update(key_ref_t keyring_ref, struct key *keyring, *key = NULL; key_ref_t key_ref; int ret; + int (*restrict_link)(struct key *, + const struct key_type *, + unsigned long, + const union key_payload *) = NULL; /* look up the key type to see if it's one of the registered kernel * types */ @@ -811,6 +831,10 @@ key_ref_t key_create_or_update(key_ref_t keyring_ref, key_check(keyring); + key_ref = ERR_PTR(-EPERM); + if (!(flags & KEY_ALLOC_BYPASS_RESTRICTION)) + restrict_link = keyring->restrict_link; + key_ref = ERR_PTR(-ENOTDIR); if (keyring->type != &key_type_keyring) goto error_put_type; @@ -835,10 +859,15 @@ key_ref_t key_create_or_update(key_ref_t keyring_ref, } index_key.desc_len = strlen(index_key.description); - key_ref = ERR_PTR(-EPERM); - if (!prep.trusted && test_bit(KEY_FLAG_TRUSTED_ONLY, &keyring->flags)) - goto error_free_prep; - flags |= prep.trusted ? KEY_ALLOC_TRUSTED : 0; + if (restrict_link) { + unsigned long kflags = prep.trusted ? KEY_FLAG_TRUSTED : 0; + ret = restrict_link(keyring, + index_key.type, kflags, &prep.payload); + if (ret < 0) { + key_ref = ERR_PTR(ret); + goto error_free_prep; + } + } ret = __key_link_begin(keyring, &index_key, &edit); if (ret < 0) { @@ -879,7 +908,7 @@ key_ref_t key_create_or_update(key_ref_t keyring_ref, /* allocate a new key */ key = key_alloc(index_key.type, index_key.description, - cred->fsuid, cred->fsgid, cred, perm, flags); + cred->fsuid, cred->fsgid, cred, perm, flags, NULL); if (IS_ERR(key)) { key_ref = ERR_CAST(key); goto error_link_end; diff --git a/security/keys/keyring.c b/security/keys/keyring.c index f931ccfeefb0..d2d1f3378008 100644 --- a/security/keys/keyring.c +++ b/security/keys/keyring.c @@ -491,13 +491,18 @@ static long keyring_read(const struct key *keyring, */ struct key *keyring_alloc(const char *description, kuid_t uid, kgid_t gid, const struct cred *cred, key_perm_t perm, - unsigned long flags, struct key *dest) + unsigned long flags, + int (*restrict_link)(struct key *, + const struct key_type *, + unsigned long, + const union key_payload *), + struct key *dest) { struct key *keyring; int ret; keyring = key_alloc(&key_type_keyring, description, - uid, gid, cred, perm, flags); + uid, gid, cred, perm, flags, restrict_link); if (!IS_ERR(keyring)) { ret = key_instantiate_and_link(keyring, NULL, 0, dest, NULL); if (ret < 0) { @@ -510,6 +515,51 @@ struct key *keyring_alloc(const char *description, kuid_t uid, kgid_t gid, } EXPORT_SYMBOL(keyring_alloc); +/** + * keyring_restrict_trusted_only - Restrict additions to a keyring to trusted keys only + * @keyring: The keyring being added to. + * @type: The type of key being added. + * @flags: The key flags. + * @payload: The payload of the key intended to be added. + * + * Reject the addition of any links to a keyring that point to keys that aren't + * marked as being trusted. It can be overridden by passing + * KEY_ALLOC_BYPASS_RESTRICTION to key_instantiate_and_link() when adding a key + * to a keyring. + * + * This is meant to be passed as the restrict_link parameter to + * keyring_alloc(). + */ +int keyring_restrict_trusted_only(struct key *keyring, + const struct key_type *type, + unsigned long flags, + const union key_payload *payload) +{ + return flags & KEY_FLAG_TRUSTED ? 0 : -EPERM; +} + +/** + * restrict_link_reject - Give -EPERM to restrict link + * @keyring: The keyring being added to. + * @type: The type of key being added. + * @flags: The key flags. + * @payload: The payload of the key intended to be added. + * + * Reject the addition of any links to a keyring. It can be overridden by + * passing KEY_ALLOC_BYPASS_RESTRICTION to key_instantiate_and_link() when + * adding a key to a keyring. + * + * This is meant to be passed as the restrict_link parameter to + * keyring_alloc(). + */ +int restrict_link_reject(struct key *keyring, + const struct key_type *type, + unsigned long flags, + const union key_payload *payload) +{ + return -EPERM; +} + /* * By default, we keys found by getting an exact match on their descriptions. */ @@ -1191,6 +1241,17 @@ void __key_link_end(struct key *keyring, up_write(&keyring->sem); } +/* + * Check addition of keys to restricted keyrings. + */ +static int __key_link_check_restriction(struct key *keyring, struct key *key) +{ + if (!keyring->restrict_link) + return 0; + return keyring->restrict_link(keyring, + key->type, key->flags, &key->payload); +} + /** * key_link - Link a key to a keyring * @keyring: The keyring to make the link in. @@ -1221,14 +1282,12 @@ int key_link(struct key *keyring, struct key *key) key_check(keyring); key_check(key); - if (test_bit(KEY_FLAG_TRUSTED_ONLY, &keyring->flags) && - !test_bit(KEY_FLAG_TRUSTED, &key->flags)) - return -EPERM; - ret = __key_link_begin(keyring, &key->index_key, &edit); if (ret == 0) { kdebug("begun {%d,%d}", keyring->serial, atomic_read(&keyring->usage)); - ret = __key_link_check_live_key(keyring, key); + ret = __key_link_check_restriction(keyring, key); + if (ret == 0) + ret = __key_link_check_live_key(keyring, key); if (ret == 0) __key_link(key, &edit); __key_link_end(keyring, &key->index_key, edit); diff --git a/security/keys/persistent.c b/security/keys/persistent.c index c9fae5ea89fe..2ef45b319dd9 100644 --- a/security/keys/persistent.c +++ b/security/keys/persistent.c @@ -26,7 +26,7 @@ static int key_create_persistent_register(struct user_namespace *ns) current_cred(), ((KEY_POS_ALL & ~KEY_POS_SETATTR) | KEY_USR_VIEW | KEY_USR_READ), - KEY_ALLOC_NOT_IN_QUOTA, NULL); + KEY_ALLOC_NOT_IN_QUOTA, NULL, NULL); if (IS_ERR(reg)) return PTR_ERR(reg); @@ -60,7 +60,7 @@ static key_ref_t key_create_persistent(struct user_namespace *ns, kuid_t uid, uid, INVALID_GID, current_cred(), ((KEY_POS_ALL & ~KEY_POS_SETATTR) | KEY_USR_VIEW | KEY_USR_READ), - KEY_ALLOC_NOT_IN_QUOTA, + KEY_ALLOC_NOT_IN_QUOTA, NULL, ns->persistent_keyring_register); if (IS_ERR(persistent)) return ERR_CAST(persistent); diff --git a/security/keys/process_keys.c b/security/keys/process_keys.c index e6d50172872f..40a885239782 100644 --- a/security/keys/process_keys.c +++ b/security/keys/process_keys.c @@ -76,7 +76,8 @@ int install_user_keyrings(void) if (IS_ERR(uid_keyring)) { uid_keyring = keyring_alloc(buf, user->uid, INVALID_GID, cred, user_keyring_perm, - KEY_ALLOC_IN_QUOTA, NULL); + KEY_ALLOC_IN_QUOTA, + NULL, NULL); if (IS_ERR(uid_keyring)) { ret = PTR_ERR(uid_keyring); goto error; @@ -92,7 +93,8 @@ int install_user_keyrings(void) session_keyring = keyring_alloc(buf, user->uid, INVALID_GID, cred, user_keyring_perm, - KEY_ALLOC_IN_QUOTA, NULL); + KEY_ALLOC_IN_QUOTA, + NULL, NULL); if (IS_ERR(session_keyring)) { ret = PTR_ERR(session_keyring); goto error_release; @@ -134,7 +136,8 @@ int install_thread_keyring_to_cred(struct cred *new) keyring = keyring_alloc("_tid", new->uid, new->gid, new, KEY_POS_ALL | KEY_USR_VIEW, - KEY_ALLOC_QUOTA_OVERRUN, NULL); + KEY_ALLOC_QUOTA_OVERRUN, + NULL, NULL); if (IS_ERR(keyring)) return PTR_ERR(keyring); @@ -180,7 +183,8 @@ int install_process_keyring_to_cred(struct cred *new) keyring = keyring_alloc("_pid", new->uid, new->gid, new, KEY_POS_ALL | KEY_USR_VIEW, - KEY_ALLOC_QUOTA_OVERRUN, NULL); + KEY_ALLOC_QUOTA_OVERRUN, + NULL, NULL); if (IS_ERR(keyring)) return PTR_ERR(keyring); @@ -231,7 +235,7 @@ int install_session_keyring_to_cred(struct cred *cred, struct key *keyring) keyring = keyring_alloc("_ses", cred->uid, cred->gid, cred, KEY_POS_ALL | KEY_USR_VIEW | KEY_USR_READ, - flags, NULL); + flags, NULL, NULL); if (IS_ERR(keyring)) return PTR_ERR(keyring); } else { @@ -785,7 +789,7 @@ long join_session_keyring(const char *name) keyring = keyring_alloc( name, old->uid, old->gid, old, KEY_POS_ALL | KEY_USR_VIEW | KEY_USR_READ | KEY_USR_LINK, - KEY_ALLOC_IN_QUOTA, NULL); + KEY_ALLOC_IN_QUOTA, NULL, NULL); if (IS_ERR(keyring)) { ret = PTR_ERR(keyring); goto error2; diff --git a/security/keys/request_key.c b/security/keys/request_key.c index c7a117c9a8f3..a29e3554751e 100644 --- a/security/keys/request_key.c +++ b/security/keys/request_key.c @@ -116,7 +116,7 @@ static int call_sbin_request_key(struct key_construction *cons, cred = get_current_cred(); keyring = keyring_alloc(desc, cred->fsuid, cred->fsgid, cred, KEY_POS_ALL | KEY_USR_VIEW | KEY_USR_READ, - KEY_ALLOC_QUOTA_OVERRUN, NULL); + KEY_ALLOC_QUOTA_OVERRUN, NULL, NULL); put_cred(cred); if (IS_ERR(keyring)) { ret = PTR_ERR(keyring); @@ -355,7 +355,7 @@ static int construct_alloc_key(struct keyring_search_context *ctx, key = key_alloc(ctx->index_key.type, ctx->index_key.description, ctx->cred->fsuid, ctx->cred->fsgid, ctx->cred, - perm, flags); + perm, flags, NULL); if (IS_ERR(key)) goto alloc_failed; diff --git a/security/keys/request_key_auth.c b/security/keys/request_key_auth.c index 4f0f112fe276..9db8b4a82787 100644 --- a/security/keys/request_key_auth.c +++ b/security/keys/request_key_auth.c @@ -202,7 +202,7 @@ struct key *request_key_auth_new(struct key *target, const void *callout_info, authkey = key_alloc(&key_type_request_key_auth, desc, cred->fsuid, cred->fsgid, cred, KEY_POS_VIEW | KEY_POS_READ | KEY_POS_SEARCH | - KEY_USR_VIEW, KEY_ALLOC_NOT_IN_QUOTA); + KEY_USR_VIEW, KEY_ALLOC_NOT_IN_QUOTA, NULL); if (IS_ERR(authkey)) { ret = PTR_ERR(authkey); goto error_alloc; -- cgit v1.3-8-gc7d7 From a527b38e1475211b67eb59b3fadb40689f035529 Mon Sep 17 00:00:00 2001 From: Denys Vlasenko Date: Tue, 12 Apr 2016 12:39:12 -0400 Subject: GFS2: fs/gfs2/glock.c: Deinline do_error, save 1856 bytes This function compiles to 522 bytes of machine code. Error paths are not very time critical. Signed-off-by: Denys Vlasenko Signed-off-by: Bob Peterson --- fs/gfs2/glock.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c index 2897ced5fca0..3910cea1aa51 100644 --- a/fs/gfs2/glock.c +++ b/fs/gfs2/glock.c @@ -218,7 +218,7 @@ static void gfs2_holder_wake(struct gfs2_holder *gh) * */ -static inline void do_error(struct gfs2_glock *gl, const int ret) +static void do_error(struct gfs2_glock *gl, const int ret) { struct gfs2_holder *gh, *tmp; -- cgit v1.3-8-gc7d7 From 9fd4dcece43a53e5a9e65a973df5693702ee6401 Mon Sep 17 00:00:00 2001 From: Nicolai Stange Date: Tue, 22 Mar 2016 14:11:13 +0100 Subject: debugfs: prevent access to possibly dead file_operations at file open Nothing prevents a dentry found by path lookup before a return of __debugfs_remove() to actually get opened after that return. Now, after the return of __debugfs_remove(), there are no guarantees whatsoever regarding the memory the corresponding inode's file_operations object had been kept in. Since __debugfs_remove() is seldomly invoked, usually from module exit handlers only, the race is hard to trigger and the impact is very low. A discussion of the problem outlined above as well as a suggested solution can be found in the (sub-)thread rooted at http://lkml.kernel.org/g/20130401203445.GA20862@ZenIV.linux.org.uk ("Yet another pipe related oops.") Basically, Greg KH suggests to introduce an intermediate fops and Al Viro points out that a pointer to the original ones may be stored in ->d_fsdata. Follow this line of reasoning: - Add SRCU as a reverse dependency of DEBUG_FS. - Introduce a srcu_struct object for the debugfs subsystem. - In debugfs_create_file(), store a pointer to the original file_operations object in ->d_fsdata. - Make debugfs_remove() and debugfs_remove_recursive() wait for a SRCU grace period after the dentry has been delete()'d and before they return to their callers. - Introduce an intermediate file_operations object named "debugfs_open_proxy_file_operations". It's ->open() functions checks, under the protection of a SRCU read lock, whether the dentry is still alive, i.e. has not been d_delete()'d and if so, tries to acquire a reference on the owning module. On success, it sets the file object's ->f_op to the original file_operations and forwards the ongoing open() call to the original ->open(). - For clarity, rename the former debugfs_file_operations to debugfs_noop_file_operations -- they are in no way canonical. The choice of SRCU over "normal" RCU is justified by the fact, that the former may also be used to protect ->i_private data from going away during the execution of a file's readers and writers which may (and do) sleep. Finally, introduce the fs/debugfs/internal.h header containing some declarations internal to the debugfs implementation. Signed-off-by: Nicolai Stange Signed-off-by: Greg Kroah-Hartman --- fs/debugfs/file.c | 91 ++++++++++++++++++++++++++++++++++++++++++++++++- fs/debugfs/inode.c | 13 ++++++- fs/debugfs/internal.h | 24 +++++++++++++ include/linux/debugfs.h | 3 -- lib/Kconfig.debug | 1 + 5 files changed, 127 insertions(+), 5 deletions(-) create mode 100644 fs/debugfs/internal.h (limited to 'fs') diff --git a/fs/debugfs/file.c b/fs/debugfs/file.c index d2ba12e23ed9..736ab3c988f2 100644 --- a/fs/debugfs/file.c +++ b/fs/debugfs/file.c @@ -22,6 +22,9 @@ #include #include #include +#include + +#include "internal.h" static ssize_t default_read_file(struct file *file, char __user *buf, size_t count, loff_t *ppos) @@ -35,13 +38,99 @@ static ssize_t default_write_file(struct file *file, const char __user *buf, return count; } -const struct file_operations debugfs_file_operations = { +const struct file_operations debugfs_noop_file_operations = { .read = default_read_file, .write = default_write_file, .open = simple_open, .llseek = noop_llseek, }; +/** + * debugfs_use_file_start - mark the beginning of file data access + * @dentry: the dentry object whose data is being accessed. + * @srcu_idx: a pointer to some memory to store a SRCU index in. + * + * Up to a matching call to debugfs_use_file_finish(), any + * successive call into the file removing functions debugfs_remove() + * and debugfs_remove_recursive() will block. Since associated private + * file data may only get freed after a successful return of any of + * the removal functions, you may safely access it after a successful + * call to debugfs_use_file_start() without worrying about + * lifetime issues. + * + * If -%EIO is returned, the file has already been removed and thus, + * it is not safe to access any of its data. If, on the other hand, + * it is allowed to access the file data, zero is returned. + * + * Regardless of the return code, any call to + * debugfs_use_file_start() must be followed by a matching call + * to debugfs_use_file_finish(). + */ +static int debugfs_use_file_start(const struct dentry *dentry, int *srcu_idx) + __acquires(&debugfs_srcu) +{ + *srcu_idx = srcu_read_lock(&debugfs_srcu); + barrier(); + if (d_unlinked(dentry)) + return -EIO; + return 0; +} + +/** + * debugfs_use_file_finish - mark the end of file data access + * @srcu_idx: the SRCU index "created" by a former call to + * debugfs_use_file_start(). + * + * Allow any ongoing concurrent call into debugfs_remove() or + * debugfs_remove_recursive() blocked by a former call to + * debugfs_use_file_start() to proceed and return to its caller. + */ +static void debugfs_use_file_finish(int srcu_idx) __releases(&debugfs_srcu) +{ + srcu_read_unlock(&debugfs_srcu, srcu_idx); +} + +#define F_DENTRY(filp) ((filp)->f_path.dentry) + +#define REAL_FOPS_DEREF(dentry) \ + ((const struct file_operations *)(dentry)->d_fsdata) + +static int open_proxy_open(struct inode *inode, struct file *filp) +{ + const struct dentry *dentry = F_DENTRY(filp); + const struct file_operations *real_fops = NULL; + int srcu_idx, r; + + r = debugfs_use_file_start(dentry, &srcu_idx); + if (r) { + r = -ENOENT; + goto out; + } + + real_fops = REAL_FOPS_DEREF(dentry); + real_fops = fops_get(real_fops); + if (!real_fops) { + /* Huh? Module did not clean up after itself at exit? */ + WARN(1, "debugfs file owner did not clean up at exit: %pd", + dentry); + r = -ENXIO; + goto out; + } + replace_fops(filp, real_fops); + + if (real_fops->open) + r = real_fops->open(inode, filp); + +out: + fops_put(real_fops); + debugfs_use_file_finish(srcu_idx); + return r; +} + +const struct file_operations debugfs_open_proxy_file_operations = { + .open = open_proxy_open, +}; + static struct dentry *debugfs_create_mode(const char *name, umode_t mode, struct dentry *parent, void *value, const struct file_operations *fops, diff --git a/fs/debugfs/inode.c b/fs/debugfs/inode.c index b1e7f35f3cd4..2905dd160575 100644 --- a/fs/debugfs/inode.c +++ b/fs/debugfs/inode.c @@ -27,9 +27,14 @@ #include #include #include +#include + +#include "internal.h" #define DEBUGFS_DEFAULT_MODE 0700 +DEFINE_SRCU(debugfs_srcu); + static struct vfsmount *debugfs_mount; static int debugfs_mount_count; static bool debugfs_registered; @@ -341,8 +346,12 @@ struct dentry *debugfs_create_file(const char *name, umode_t mode, return failed_creating(dentry); inode->i_mode = mode; - inode->i_fop = fops ? fops : &debugfs_file_operations; inode->i_private = data; + + inode->i_fop = fops ? &debugfs_open_proxy_file_operations + : &debugfs_noop_file_operations; + dentry->d_fsdata = (void *)fops; + d_instantiate(dentry, inode); fsnotify_create(d_inode(dentry->d_parent), dentry); return end_creating(dentry); @@ -570,6 +579,7 @@ void debugfs_remove(struct dentry *dentry) inode_unlock(d_inode(parent)); if (!ret) simple_release_fs(&debugfs_mount, &debugfs_mount_count); + synchronize_srcu(&debugfs_srcu); } EXPORT_SYMBOL_GPL(debugfs_remove); @@ -647,6 +657,7 @@ void debugfs_remove_recursive(struct dentry *dentry) if (!__debugfs_remove(child, parent)) simple_release_fs(&debugfs_mount, &debugfs_mount_count); inode_unlock(d_inode(parent)); + synchronize_srcu(&debugfs_srcu); } EXPORT_SYMBOL_GPL(debugfs_remove_recursive); diff --git a/fs/debugfs/internal.h b/fs/debugfs/internal.h new file mode 100644 index 000000000000..c7aaa5cb6685 --- /dev/null +++ b/fs/debugfs/internal.h @@ -0,0 +1,24 @@ +/* + * internal.h - declarations internal to debugfs + * + * Copyright (C) 2016 Nicolai Stange + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License version + * 2 as published by the Free Software Foundation. + * + */ + +#ifndef _DEBUGFS_INTERNAL_H_ +#define _DEBUGFS_INTERNAL_H_ + +struct file_operations; +struct srcu_struct; + +/* declared over in file.c */ +extern const struct file_operations debugfs_noop_file_operations; +extern const struct file_operations debugfs_open_proxy_file_operations; + +extern struct srcu_struct debugfs_srcu; + +#endif /* _DEBUGFS_INTERNAL_H_ */ diff --git a/include/linux/debugfs.h b/include/linux/debugfs.h index 981e53ab84e8..fcafe2d389f9 100644 --- a/include/linux/debugfs.h +++ b/include/linux/debugfs.h @@ -43,9 +43,6 @@ extern struct dentry *arch_debugfs_dir; #if defined(CONFIG_DEBUG_FS) -/* declared over in file.c */ -extern const struct file_operations debugfs_file_operations; - struct dentry *debugfs_create_file(const char *name, umode_t mode, struct dentry *parent, void *data, const struct file_operations *fops); diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index 1e9a607534ca..ddb0e8337aae 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -257,6 +257,7 @@ config PAGE_OWNER config DEBUG_FS bool "Debug Filesystem" + select SRCU help debugfs is a virtual file system that kernel developers use to put debugging files into. Enable this option to be able to read and -- cgit v1.3-8-gc7d7 From 49d200deaa680501f19a247b1fffb29301e51d2b Mon Sep 17 00:00:00 2001 From: Nicolai Stange Date: Tue, 22 Mar 2016 14:11:14 +0100 Subject: debugfs: prevent access to removed files' private data Upon return of debugfs_remove()/debugfs_remove_recursive(), it might still be attempted to access associated private file data through previously opened struct file objects. If that data has been freed by the caller of debugfs_remove*() in the meanwhile, the reading/writing process would either encounter a fault or, if the memory address in question has been reassigned again, unrelated data structures could get overwritten. However, since debugfs files are seldomly removed, usually from module exit handlers only, the impact is very low. Currently, there are ~1000 call sites of debugfs_create_file() spread throughout the whole tree and touching all of those struct file_operations in order to make them file removal aware by means of checking the result of debugfs_use_file_start() from within their methods is unfeasible. Instead, wrap the struct file_operations by a lifetime managing proxy at file open: - In debugfs_create_file(), the original fops handed in has got stashed away in ->d_fsdata already. - In debugfs_create_file(), install a proxy file_operations factory, debugfs_full_proxy_file_operations, at ->i_fop. This proxy factory has got an ->open() method only. It carries out some lifetime checks and if successful, dynamically allocates and sets up a new struct file_operations proxy at ->f_op. Afterwards, it forwards to the ->open() of the original struct file_operations in ->d_fsdata, if any. The dynamically set up proxy at ->f_op has got a lifetime managing wrapper set for each of the methods defined in the original struct file_operations in ->d_fsdata. Its ->release()er frees the proxy again and forwards to the original ->release(), if any. In order not to mislead the VFS layer, it is strictly necessary to leave those fields blank in the proxy that have been NULL in the original struct file_operations also, i.e. aren't supported. This is why there is a need for dynamically allocated proxies. The choice made not to allocate a proxy instance for every dentry at file creation, but for every struct file object instantiated thereof is justified by the expected usage pattern of debugfs, namely that in general very few files get opened more than once at a time. The wrapper methods set in the struct file_operations implement lifetime managing by means of the SRCU protection facilities already in place for debugfs: They set up a SRCU read side critical section and check whether the dentry is still alive by means of debugfs_use_file_start(). If so, they forward the call to the original struct file_operation stored in ->d_fsdata, still under the protection of the SRCU read side critical section. This SRCU read side critical section prevents any pending debugfs_remove() and friends to return to their callers. Since a file's private data must only be freed after the return of debugfs_remove(), the ongoing proxied call is guarded against any file removal race. If, on the other hand, the initial call to debugfs_use_file_start() detects that the dentry is dead, the wrapper simply returns -EIO and does not forward the call. Note that the ->poll() wrapper is special in that its signature does not allow for the return of arbitrary -EXXX values and thus, POLLHUP is returned here. In order not to pollute debugfs with wrapper definitions that aren't ever needed, I chose not to define a wrapper for every struct file_operations method possible. Instead, a wrapper is defined only for the subset of methods which are actually set by any debugfs users. Currently, these are: ->llseek() ->read() ->write() ->unlocked_ioctl() ->poll() The ->release() wrapper is special in that it does not protect the original ->release() in any way from dead files in order not to leak resources. Thus, any ->release() handed to debugfs must implement file lifetime management manually, if needed. For only 33 out of a total of 434 releasers handed in to debugfs, it could not be verified immediately whether they access data structures that might have been freed upon a debugfs_remove() return in the meanwhile. Export debugfs_use_file_start() and debugfs_use_file_finish() in order to allow any ->release() to manually implement file lifetime management. For a set of common cases of struct file_operations implemented by the debugfs_core itself, future patches will incorporate file lifetime management directly within those in order to allow for their unproxied operation. Rename the original, non-proxying "debugfs_create_file()" to "debugfs_create_file_unsafe()" and keep it for future internal use by debugfs itself. Factor out code common to both into the new __debugfs_create_file(). Signed-off-by: Nicolai Stange Signed-off-by: Greg Kroah-Hartman --- fs/debugfs/file.c | 157 +++++++++++++++++++++++++++++++++++++++++++++++- fs/debugfs/inode.c | 70 ++++++++++++++------- fs/debugfs/internal.h | 6 +- include/linux/debugfs.h | 20 ++++++ 4 files changed, 226 insertions(+), 27 deletions(-) (limited to 'fs') diff --git a/fs/debugfs/file.c b/fs/debugfs/file.c index 736ab3c988f2..6eb58a8ed03c 100644 --- a/fs/debugfs/file.c +++ b/fs/debugfs/file.c @@ -23,9 +23,12 @@ #include #include #include +#include #include "internal.h" +struct poll_table_struct; + static ssize_t default_read_file(struct file *file, char __user *buf, size_t count, loff_t *ppos) { @@ -66,7 +69,7 @@ const struct file_operations debugfs_noop_file_operations = { * debugfs_use_file_start() must be followed by a matching call * to debugfs_use_file_finish(). */ -static int debugfs_use_file_start(const struct dentry *dentry, int *srcu_idx) +int debugfs_use_file_start(const struct dentry *dentry, int *srcu_idx) __acquires(&debugfs_srcu) { *srcu_idx = srcu_read_lock(&debugfs_srcu); @@ -75,6 +78,7 @@ static int debugfs_use_file_start(const struct dentry *dentry, int *srcu_idx) return -EIO; return 0; } +EXPORT_SYMBOL_GPL(debugfs_use_file_start); /** * debugfs_use_file_finish - mark the end of file data access @@ -85,10 +89,11 @@ static int debugfs_use_file_start(const struct dentry *dentry, int *srcu_idx) * debugfs_remove_recursive() blocked by a former call to * debugfs_use_file_start() to proceed and return to its caller. */ -static void debugfs_use_file_finish(int srcu_idx) __releases(&debugfs_srcu) +void debugfs_use_file_finish(int srcu_idx) __releases(&debugfs_srcu) { srcu_read_unlock(&debugfs_srcu, srcu_idx); } +EXPORT_SYMBOL_GPL(debugfs_use_file_finish); #define F_DENTRY(filp) ((filp)->f_path.dentry) @@ -131,6 +136,154 @@ const struct file_operations debugfs_open_proxy_file_operations = { .open = open_proxy_open, }; +#define PROTO(args...) args +#define ARGS(args...) args + +#define FULL_PROXY_FUNC(name, ret_type, filp, proto, args) \ +static ret_type full_proxy_ ## name(proto) \ +{ \ + const struct dentry *dentry = F_DENTRY(filp); \ + const struct file_operations *real_fops = \ + REAL_FOPS_DEREF(dentry); \ + int srcu_idx; \ + ret_type r; \ + \ + r = debugfs_use_file_start(dentry, &srcu_idx); \ + if (likely(!r)) \ + r = real_fops->name(args); \ + debugfs_use_file_finish(srcu_idx); \ + return r; \ +} + +FULL_PROXY_FUNC(llseek, loff_t, filp, + PROTO(struct file *filp, loff_t offset, int whence), + ARGS(filp, offset, whence)); + +FULL_PROXY_FUNC(read, ssize_t, filp, + PROTO(struct file *filp, char __user *buf, size_t size, + loff_t *ppos), + ARGS(filp, buf, size, ppos)); + +FULL_PROXY_FUNC(write, ssize_t, filp, + PROTO(struct file *filp, const char __user *buf, size_t size, + loff_t *ppos), + ARGS(filp, buf, size, ppos)); + +FULL_PROXY_FUNC(unlocked_ioctl, long, filp, + PROTO(struct file *filp, unsigned int cmd, unsigned long arg), + ARGS(filp, cmd, arg)); + +static unsigned int full_proxy_poll(struct file *filp, + struct poll_table_struct *wait) +{ + const struct dentry *dentry = F_DENTRY(filp); + const struct file_operations *real_fops = REAL_FOPS_DEREF(dentry); + int srcu_idx; + unsigned int r = 0; + + if (debugfs_use_file_start(dentry, &srcu_idx)) { + debugfs_use_file_finish(srcu_idx); + return POLLHUP; + } + + r = real_fops->poll(filp, wait); + debugfs_use_file_finish(srcu_idx); + return r; +} + +static int full_proxy_release(struct inode *inode, struct file *filp) +{ + const struct dentry *dentry = F_DENTRY(filp); + const struct file_operations *real_fops = REAL_FOPS_DEREF(dentry); + const struct file_operations *proxy_fops = filp->f_op; + int r = 0; + + /* + * We must not protect this against removal races here: the + * original releaser should be called unconditionally in order + * not to leak any resources. Releasers must not assume that + * ->i_private is still being meaningful here. + */ + if (real_fops->release) + r = real_fops->release(inode, filp); + + replace_fops(filp, d_inode(dentry)->i_fop); + kfree((void *)proxy_fops); + fops_put(real_fops); + return 0; +} + +static void __full_proxy_fops_init(struct file_operations *proxy_fops, + const struct file_operations *real_fops) +{ + proxy_fops->release = full_proxy_release; + if (real_fops->llseek) + proxy_fops->llseek = full_proxy_llseek; + if (real_fops->read) + proxy_fops->read = full_proxy_read; + if (real_fops->write) + proxy_fops->write = full_proxy_write; + if (real_fops->poll) + proxy_fops->poll = full_proxy_poll; + if (real_fops->unlocked_ioctl) + proxy_fops->unlocked_ioctl = full_proxy_unlocked_ioctl; +} + +static int full_proxy_open(struct inode *inode, struct file *filp) +{ + const struct dentry *dentry = F_DENTRY(filp); + const struct file_operations *real_fops = NULL; + struct file_operations *proxy_fops = NULL; + int srcu_idx, r; + + r = debugfs_use_file_start(dentry, &srcu_idx); + if (r) { + r = -ENOENT; + goto out; + } + + real_fops = REAL_FOPS_DEREF(dentry); + real_fops = fops_get(real_fops); + if (!real_fops) { + /* Huh? Module did not cleanup after itself at exit? */ + WARN(1, "debugfs file owner did not clean up at exit: %pd", + dentry); + r = -ENXIO; + goto out; + } + + proxy_fops = kzalloc(sizeof(*proxy_fops), GFP_KERNEL); + if (!proxy_fops) { + r = -ENOMEM; + goto free_proxy; + } + __full_proxy_fops_init(proxy_fops, real_fops); + replace_fops(filp, proxy_fops); + + if (real_fops->open) { + r = real_fops->open(inode, filp); + + if (filp->f_op != proxy_fops) { + /* No protection against file removal anymore. */ + WARN(1, "debugfs file owner replaced proxy fops: %pd", + dentry); + goto free_proxy; + } + } + + goto out; +free_proxy: + kfree(proxy_fops); + fops_put(real_fops); +out: + debugfs_use_file_finish(srcu_idx); + return r; +} + +const struct file_operations debugfs_full_proxy_file_operations = { + .open = full_proxy_open, +}; + static struct dentry *debugfs_create_mode(const char *name, umode_t mode, struct dentry *parent, void *value, const struct file_operations *fops, diff --git a/fs/debugfs/inode.c b/fs/debugfs/inode.c index 2905dd160575..136f269f01de 100644 --- a/fs/debugfs/inode.c +++ b/fs/debugfs/inode.c @@ -300,6 +300,37 @@ static struct dentry *end_creating(struct dentry *dentry) return dentry; } +static struct dentry *__debugfs_create_file(const char *name, umode_t mode, + struct dentry *parent, void *data, + const struct file_operations *proxy_fops, + const struct file_operations *real_fops) +{ + struct dentry *dentry; + struct inode *inode; + + if (!(mode & S_IFMT)) + mode |= S_IFREG; + BUG_ON(!S_ISREG(mode)); + dentry = start_creating(name, parent); + + if (IS_ERR(dentry)) + return NULL; + + inode = debugfs_get_inode(dentry->d_sb); + if (unlikely(!inode)) + return failed_creating(dentry); + + inode->i_mode = mode; + inode->i_private = data; + + inode->i_fop = proxy_fops; + dentry->d_fsdata = (void *)real_fops; + + d_instantiate(dentry, inode); + fsnotify_create(d_inode(dentry->d_parent), dentry); + return end_creating(dentry); +} + /** * debugfs_create_file - create a file in the debugfs filesystem * @name: a pointer to a string containing the name of the file to create. @@ -330,33 +361,24 @@ struct dentry *debugfs_create_file(const char *name, umode_t mode, struct dentry *parent, void *data, const struct file_operations *fops) { - struct dentry *dentry; - struct inode *inode; - - if (!(mode & S_IFMT)) - mode |= S_IFREG; - BUG_ON(!S_ISREG(mode)); - dentry = start_creating(name, parent); - - if (IS_ERR(dentry)) - return NULL; - - inode = debugfs_get_inode(dentry->d_sb); - if (unlikely(!inode)) - return failed_creating(dentry); - inode->i_mode = mode; - inode->i_private = data; + return __debugfs_create_file(name, mode, parent, data, + fops ? &debugfs_full_proxy_file_operations : + &debugfs_noop_file_operations, + fops); +} +EXPORT_SYMBOL_GPL(debugfs_create_file); - inode->i_fop = fops ? &debugfs_open_proxy_file_operations - : &debugfs_noop_file_operations; - dentry->d_fsdata = (void *)fops; +struct dentry *debugfs_create_file_unsafe(const char *name, umode_t mode, + struct dentry *parent, void *data, + const struct file_operations *fops) +{ - d_instantiate(dentry, inode); - fsnotify_create(d_inode(dentry->d_parent), dentry); - return end_creating(dentry); + return __debugfs_create_file(name, mode, parent, data, + fops ? &debugfs_open_proxy_file_operations : + &debugfs_noop_file_operations, + fops); } -EXPORT_SYMBOL_GPL(debugfs_create_file); /** * debugfs_create_file_size - create a file in the debugfs filesystem @@ -579,6 +601,7 @@ void debugfs_remove(struct dentry *dentry) inode_unlock(d_inode(parent)); if (!ret) simple_release_fs(&debugfs_mount, &debugfs_mount_count); + synchronize_srcu(&debugfs_srcu); } EXPORT_SYMBOL_GPL(debugfs_remove); @@ -657,6 +680,7 @@ void debugfs_remove_recursive(struct dentry *dentry) if (!__debugfs_remove(child, parent)) simple_release_fs(&debugfs_mount, &debugfs_mount_count); inode_unlock(d_inode(parent)); + synchronize_srcu(&debugfs_srcu); } EXPORT_SYMBOL_GPL(debugfs_remove_recursive); diff --git a/fs/debugfs/internal.h b/fs/debugfs/internal.h index c7aaa5cb6685..bba52634b995 100644 --- a/fs/debugfs/internal.h +++ b/fs/debugfs/internal.h @@ -13,12 +13,14 @@ #define _DEBUGFS_INTERNAL_H_ struct file_operations; -struct srcu_struct; /* declared over in file.c */ extern const struct file_operations debugfs_noop_file_operations; extern const struct file_operations debugfs_open_proxy_file_operations; +extern const struct file_operations debugfs_full_proxy_file_operations; -extern struct srcu_struct debugfs_srcu; +struct dentry *debugfs_create_file_unsafe(const char *name, umode_t mode, + struct dentry *parent, void *data, + const struct file_operations *fops); #endif /* _DEBUGFS_INTERNAL_H_ */ diff --git a/include/linux/debugfs.h b/include/linux/debugfs.h index fcafe2d389f9..a63e6ea3321c 100644 --- a/include/linux/debugfs.h +++ b/include/linux/debugfs.h @@ -19,9 +19,11 @@ #include #include +#include struct device; struct file_operations; +struct srcu_struct; struct debugfs_blob_wrapper { void *data; @@ -41,6 +43,8 @@ struct debugfs_regset32 { extern struct dentry *arch_debugfs_dir; +extern struct srcu_struct debugfs_srcu; + #if defined(CONFIG_DEBUG_FS) struct dentry *debugfs_create_file(const char *name, umode_t mode, @@ -65,6 +69,11 @@ struct dentry *debugfs_create_automount(const char *name, void debugfs_remove(struct dentry *dentry); void debugfs_remove_recursive(struct dentry *dentry); +int debugfs_use_file_start(const struct dentry *dentry, int *srcu_idx) + __acquires(&debugfs_srcu); + +void debugfs_use_file_finish(int srcu_idx) __releases(&debugfs_srcu); + struct dentry *debugfs_rename(struct dentry *old_dir, struct dentry *old_dentry, struct dentry *new_dir, const char *new_name); @@ -173,6 +182,17 @@ static inline void debugfs_remove(struct dentry *dentry) static inline void debugfs_remove_recursive(struct dentry *dentry) { } +static inline int debugfs_use_file_start(const struct dentry *dentry, + int *srcu_idx) + __acquires(&debugfs_srcu) +{ + return 0; +} + +static inline void debugfs_use_file_finish(int srcu_idx) + __releases(&debugfs_srcu) +{ } + static inline struct dentry *debugfs_rename(struct dentry *old_dir, struct dentry *old_dentry, struct dentry *new_dir, char *new_name) { -- cgit v1.3-8-gc7d7 From c64688081490321f2d23a292ef24e60bb321f3f1 Mon Sep 17 00:00:00 2001 From: Nicolai Stange Date: Tue, 22 Mar 2016 14:11:15 +0100 Subject: debugfs: add support for self-protecting attribute file fops In order to protect them against file removal issues, debugfs_create_file() creates a lifetime managing proxy around each struct file_operations handed in. In cases where this struct file_operations is able to manage file lifetime by itself already, the proxy created by debugfs is a waste of resources. The most common class of struct file_operations given to debugfs are those defined by means of the DEFINE_SIMPLE_ATTRIBUTE() macro. Introduce a DEFINE_DEBUGFS_ATTRIBUTE() macro to allow any struct file_operations of this class to be easily made file lifetime aware and thus, to be operated unproxied. Specifically, introduce debugfs_attr_read() and debugfs_attr_write() which wrap simple_attr_read() and simple_attr_write() under the protection of a debugfs_use_file_start()/debugfs_use_file_finish() pair. Make DEFINE_DEBUGFS_ATTRIBUTE() set the defined struct file_operations' ->read() and ->write() members to these wrappers. Export debugfs_create_file_unsafe() in order to allow debugfs users to create their files in non-proxying operation mode. Signed-off-by: Nicolai Stange Signed-off-by: Greg Kroah-Hartman --- fs/debugfs/file.c | 28 ++++++++++++++++++++++++++++ fs/debugfs/inode.c | 28 ++++++++++++++++++++++++++++ include/linux/debugfs.h | 26 ++++++++++++++++++++++++++ 3 files changed, 82 insertions(+) (limited to 'fs') diff --git a/fs/debugfs/file.c b/fs/debugfs/file.c index 6eb58a8ed03c..8ef56d9499a4 100644 --- a/fs/debugfs/file.c +++ b/fs/debugfs/file.c @@ -284,6 +284,34 @@ const struct file_operations debugfs_full_proxy_file_operations = { .open = full_proxy_open, }; +ssize_t debugfs_attr_read(struct file *file, char __user *buf, + size_t len, loff_t *ppos) +{ + ssize_t ret; + int srcu_idx; + + ret = debugfs_use_file_start(F_DENTRY(file), &srcu_idx); + if (likely(!ret)) + ret = simple_attr_read(file, buf, len, ppos); + debugfs_use_file_finish(srcu_idx); + return ret; +} +EXPORT_SYMBOL_GPL(debugfs_attr_read); + +ssize_t debugfs_attr_write(struct file *file, const char __user *buf, + size_t len, loff_t *ppos) +{ + ssize_t ret; + int srcu_idx; + + ret = debugfs_use_file_start(F_DENTRY(file), &srcu_idx); + if (likely(!ret)) + ret = simple_attr_write(file, buf, len, ppos); + debugfs_use_file_finish(srcu_idx); + return ret; +} +EXPORT_SYMBOL_GPL(debugfs_attr_write); + static struct dentry *debugfs_create_mode(const char *name, umode_t mode, struct dentry *parent, void *value, const struct file_operations *fops, diff --git a/fs/debugfs/inode.c b/fs/debugfs/inode.c index 136f269f01de..41e079a8da26 100644 --- a/fs/debugfs/inode.c +++ b/fs/debugfs/inode.c @@ -369,6 +369,33 @@ struct dentry *debugfs_create_file(const char *name, umode_t mode, } EXPORT_SYMBOL_GPL(debugfs_create_file); +/** + * debugfs_create_file_unsafe - create a file in the debugfs filesystem + * @name: a pointer to a string containing the name of the file to create. + * @mode: the permission that the file should have. + * @parent: a pointer to the parent dentry for this file. This should be a + * directory dentry if set. If this parameter is NULL, then the + * file will be created in the root of the debugfs filesystem. + * @data: a pointer to something that the caller will want to get to later + * on. The inode.i_private pointer will point to this value on + * the open() call. + * @fops: a pointer to a struct file_operations that should be used for + * this file. + * + * debugfs_create_file_unsafe() is completely analogous to + * debugfs_create_file(), the only difference being that the fops + * handed it will not get protected against file removals by the + * debugfs core. + * + * It is your responsibility to protect your struct file_operation + * methods against file removals by means of debugfs_use_file_start() + * and debugfs_use_file_finish(). ->open() is still protected by + * debugfs though. + * + * Any struct file_operations defined by means of + * DEFINE_DEBUGFS_ATTRIBUTE() is protected against file removals and + * thus, may be used here. + */ struct dentry *debugfs_create_file_unsafe(const char *name, umode_t mode, struct dentry *parent, void *data, const struct file_operations *fops) @@ -379,6 +406,7 @@ struct dentry *debugfs_create_file_unsafe(const char *name, umode_t mode, &debugfs_noop_file_operations, fops); } +EXPORT_SYMBOL_GPL(debugfs_create_file_unsafe); /** * debugfs_create_file_size - create a file in the debugfs filesystem diff --git a/include/linux/debugfs.h b/include/linux/debugfs.h index a63e6ea3321c..1438e2322d5c 100644 --- a/include/linux/debugfs.h +++ b/include/linux/debugfs.h @@ -50,6 +50,9 @@ extern struct srcu_struct debugfs_srcu; struct dentry *debugfs_create_file(const char *name, umode_t mode, struct dentry *parent, void *data, const struct file_operations *fops); +struct dentry *debugfs_create_file_unsafe(const char *name, umode_t mode, + struct dentry *parent, void *data, + const struct file_operations *fops); struct dentry *debugfs_create_file_size(const char *name, umode_t mode, struct dentry *parent, void *data, @@ -74,6 +77,26 @@ int debugfs_use_file_start(const struct dentry *dentry, int *srcu_idx) void debugfs_use_file_finish(int srcu_idx) __releases(&debugfs_srcu); +ssize_t debugfs_attr_read(struct file *file, char __user *buf, + size_t len, loff_t *ppos); +ssize_t debugfs_attr_write(struct file *file, const char __user *buf, + size_t len, loff_t *ppos); + +#define DEFINE_DEBUGFS_ATTRIBUTE(__fops, __get, __set, __fmt) \ +static int __fops ## _open(struct inode *inode, struct file *file) \ +{ \ + __simple_attr_check_format(__fmt, 0ull); \ + return simple_attr_open(inode, file, __get, __set, __fmt); \ +} \ +static const struct file_operations __fops = { \ + .owner = THIS_MODULE, \ + .open = __fops ## _open, \ + .release = simple_attr_release, \ + .read = debugfs_attr_read, \ + .write = debugfs_attr_write, \ + .llseek = generic_file_llseek, \ +} + struct dentry *debugfs_rename(struct dentry *old_dir, struct dentry *old_dentry, struct dentry *new_dir, const char *new_name); @@ -193,6 +216,9 @@ static inline void debugfs_use_file_finish(int srcu_idx) __releases(&debugfs_srcu) { } +#define DEFINE_DEBUGFS_ATTRIBUTE(__fops, __get, __set, __fmt) \ + static const struct file_operations __fops = { 0 } + static inline struct dentry *debugfs_rename(struct dentry *old_dir, struct dentry *old_dentry, struct dentry *new_dir, char *new_name) { -- cgit v1.3-8-gc7d7 From 4909f168104b24f592fb8d502e2a6520346a3927 Mon Sep 17 00:00:00 2001 From: Nicolai Stange Date: Tue, 22 Mar 2016 14:11:17 +0100 Subject: debugfs: unproxify integer attribute files Currently, the struct file_operations associated with the integer attribute style files created through the debugfs_create_*() helpers are not file lifetime aware as they are defined by means of DEFINE_SIMPLE_ATTRIBUTE(). Thus, a lifetime managing proxy is created around the original fops each time such a file is opened which is an unnecessary waste of resources. Migrate all usages of DEFINE_SIMPLE_ATTRIBUTE() within debugfs itself to DEFINE_DEBUGFS_ATTRIBUTE() in order to implement file lifetime managing within the struct file_operations thus defined. Introduce the debugfs_create_mode_unsafe() helper, analogous to debugfs_create_mode(), but distinct in that it creates the files in non-proxying operation mode through debugfs_create_file_unsafe(). Feed all struct file_operations migrated to DEFINE_DEBUGFS_ATTRIBUTE() into debugfs_create_mode_unsafe() instead of former debugfs_create_mode(). Signed-off-by: Nicolai Stange Signed-off-by: Greg Kroah-Hartman --- fs/debugfs/file.c | 123 +++++++++++++++++++++++++++++++++--------------------- 1 file changed, 75 insertions(+), 48 deletions(-) (limited to 'fs') diff --git a/fs/debugfs/file.c b/fs/debugfs/file.c index 8ef56d9499a4..4b3967e86e97 100644 --- a/fs/debugfs/file.c +++ b/fs/debugfs/file.c @@ -328,6 +328,24 @@ static struct dentry *debugfs_create_mode(const char *name, umode_t mode, return debugfs_create_file(name, mode, parent, value, fops); } +static struct dentry *debugfs_create_mode_unsafe(const char *name, umode_t mode, + struct dentry *parent, void *value, + const struct file_operations *fops, + const struct file_operations *fops_ro, + const struct file_operations *fops_wo) +{ + /* if there are no write bits set, make read only */ + if (!(mode & S_IWUGO)) + return debugfs_create_file_unsafe(name, mode, parent, value, + fops_ro); + /* if there are no read bits set, make write only */ + if (!(mode & S_IRUGO)) + return debugfs_create_file_unsafe(name, mode, parent, value, + fops_wo); + + return debugfs_create_file_unsafe(name, mode, parent, value, fops); +} + static int debugfs_u8_set(void *data, u64 val) { *(u8 *)data = val; @@ -338,9 +356,9 @@ static int debugfs_u8_get(void *data, u64 *val) *val = *(u8 *)data; return 0; } -DEFINE_SIMPLE_ATTRIBUTE(fops_u8, debugfs_u8_get, debugfs_u8_set, "%llu\n"); -DEFINE_SIMPLE_ATTRIBUTE(fops_u8_ro, debugfs_u8_get, NULL, "%llu\n"); -DEFINE_SIMPLE_ATTRIBUTE(fops_u8_wo, NULL, debugfs_u8_set, "%llu\n"); +DEFINE_DEBUGFS_ATTRIBUTE(fops_u8, debugfs_u8_get, debugfs_u8_set, "%llu\n"); +DEFINE_DEBUGFS_ATTRIBUTE(fops_u8_ro, debugfs_u8_get, NULL, "%llu\n"); +DEFINE_DEBUGFS_ATTRIBUTE(fops_u8_wo, NULL, debugfs_u8_set, "%llu\n"); /** * debugfs_create_u8 - create a debugfs file that is used to read and write an unsigned 8-bit value @@ -369,7 +387,7 @@ DEFINE_SIMPLE_ATTRIBUTE(fops_u8_wo, NULL, debugfs_u8_set, "%llu\n"); struct dentry *debugfs_create_u8(const char *name, umode_t mode, struct dentry *parent, u8 *value) { - return debugfs_create_mode(name, mode, parent, value, &fops_u8, + return debugfs_create_mode_unsafe(name, mode, parent, value, &fops_u8, &fops_u8_ro, &fops_u8_wo); } EXPORT_SYMBOL_GPL(debugfs_create_u8); @@ -384,9 +402,9 @@ static int debugfs_u16_get(void *data, u64 *val) *val = *(u16 *)data; return 0; } -DEFINE_SIMPLE_ATTRIBUTE(fops_u16, debugfs_u16_get, debugfs_u16_set, "%llu\n"); -DEFINE_SIMPLE_ATTRIBUTE(fops_u16_ro, debugfs_u16_get, NULL, "%llu\n"); -DEFINE_SIMPLE_ATTRIBUTE(fops_u16_wo, NULL, debugfs_u16_set, "%llu\n"); +DEFINE_DEBUGFS_ATTRIBUTE(fops_u16, debugfs_u16_get, debugfs_u16_set, "%llu\n"); +DEFINE_DEBUGFS_ATTRIBUTE(fops_u16_ro, debugfs_u16_get, NULL, "%llu\n"); +DEFINE_DEBUGFS_ATTRIBUTE(fops_u16_wo, NULL, debugfs_u16_set, "%llu\n"); /** * debugfs_create_u16 - create a debugfs file that is used to read and write an unsigned 16-bit value @@ -415,7 +433,7 @@ DEFINE_SIMPLE_ATTRIBUTE(fops_u16_wo, NULL, debugfs_u16_set, "%llu\n"); struct dentry *debugfs_create_u16(const char *name, umode_t mode, struct dentry *parent, u16 *value) { - return debugfs_create_mode(name, mode, parent, value, &fops_u16, + return debugfs_create_mode_unsafe(name, mode, parent, value, &fops_u16, &fops_u16_ro, &fops_u16_wo); } EXPORT_SYMBOL_GPL(debugfs_create_u16); @@ -430,9 +448,9 @@ static int debugfs_u32_get(void *data, u64 *val) *val = *(u32 *)data; return 0; } -DEFINE_SIMPLE_ATTRIBUTE(fops_u32, debugfs_u32_get, debugfs_u32_set, "%llu\n"); -DEFINE_SIMPLE_ATTRIBUTE(fops_u32_ro, debugfs_u32_get, NULL, "%llu\n"); -DEFINE_SIMPLE_ATTRIBUTE(fops_u32_wo, NULL, debugfs_u32_set, "%llu\n"); +DEFINE_DEBUGFS_ATTRIBUTE(fops_u32, debugfs_u32_get, debugfs_u32_set, "%llu\n"); +DEFINE_DEBUGFS_ATTRIBUTE(fops_u32_ro, debugfs_u32_get, NULL, "%llu\n"); +DEFINE_DEBUGFS_ATTRIBUTE(fops_u32_wo, NULL, debugfs_u32_set, "%llu\n"); /** * debugfs_create_u32 - create a debugfs file that is used to read and write an unsigned 32-bit value @@ -461,7 +479,7 @@ DEFINE_SIMPLE_ATTRIBUTE(fops_u32_wo, NULL, debugfs_u32_set, "%llu\n"); struct dentry *debugfs_create_u32(const char *name, umode_t mode, struct dentry *parent, u32 *value) { - return debugfs_create_mode(name, mode, parent, value, &fops_u32, + return debugfs_create_mode_unsafe(name, mode, parent, value, &fops_u32, &fops_u32_ro, &fops_u32_wo); } EXPORT_SYMBOL_GPL(debugfs_create_u32); @@ -477,9 +495,9 @@ static int debugfs_u64_get(void *data, u64 *val) *val = *(u64 *)data; return 0; } -DEFINE_SIMPLE_ATTRIBUTE(fops_u64, debugfs_u64_get, debugfs_u64_set, "%llu\n"); -DEFINE_SIMPLE_ATTRIBUTE(fops_u64_ro, debugfs_u64_get, NULL, "%llu\n"); -DEFINE_SIMPLE_ATTRIBUTE(fops_u64_wo, NULL, debugfs_u64_set, "%llu\n"); +DEFINE_DEBUGFS_ATTRIBUTE(fops_u64, debugfs_u64_get, debugfs_u64_set, "%llu\n"); +DEFINE_DEBUGFS_ATTRIBUTE(fops_u64_ro, debugfs_u64_get, NULL, "%llu\n"); +DEFINE_DEBUGFS_ATTRIBUTE(fops_u64_wo, NULL, debugfs_u64_set, "%llu\n"); /** * debugfs_create_u64 - create a debugfs file that is used to read and write an unsigned 64-bit value @@ -508,7 +526,7 @@ DEFINE_SIMPLE_ATTRIBUTE(fops_u64_wo, NULL, debugfs_u64_set, "%llu\n"); struct dentry *debugfs_create_u64(const char *name, umode_t mode, struct dentry *parent, u64 *value) { - return debugfs_create_mode(name, mode, parent, value, &fops_u64, + return debugfs_create_mode_unsafe(name, mode, parent, value, &fops_u64, &fops_u64_ro, &fops_u64_wo); } EXPORT_SYMBOL_GPL(debugfs_create_u64); @@ -524,9 +542,10 @@ static int debugfs_ulong_get(void *data, u64 *val) *val = *(unsigned long *)data; return 0; } -DEFINE_SIMPLE_ATTRIBUTE(fops_ulong, debugfs_ulong_get, debugfs_ulong_set, "%llu\n"); -DEFINE_SIMPLE_ATTRIBUTE(fops_ulong_ro, debugfs_ulong_get, NULL, "%llu\n"); -DEFINE_SIMPLE_ATTRIBUTE(fops_ulong_wo, NULL, debugfs_ulong_set, "%llu\n"); +DEFINE_DEBUGFS_ATTRIBUTE(fops_ulong, debugfs_ulong_get, debugfs_ulong_set, + "%llu\n"); +DEFINE_DEBUGFS_ATTRIBUTE(fops_ulong_ro, debugfs_ulong_get, NULL, "%llu\n"); +DEFINE_DEBUGFS_ATTRIBUTE(fops_ulong_wo, NULL, debugfs_ulong_set, "%llu\n"); /** * debugfs_create_ulong - create a debugfs file that is used to read and write @@ -556,26 +575,30 @@ DEFINE_SIMPLE_ATTRIBUTE(fops_ulong_wo, NULL, debugfs_ulong_set, "%llu\n"); struct dentry *debugfs_create_ulong(const char *name, umode_t mode, struct dentry *parent, unsigned long *value) { - return debugfs_create_mode(name, mode, parent, value, &fops_ulong, - &fops_ulong_ro, &fops_ulong_wo); + return debugfs_create_mode_unsafe(name, mode, parent, value, + &fops_ulong, &fops_ulong_ro, + &fops_ulong_wo); } EXPORT_SYMBOL_GPL(debugfs_create_ulong); -DEFINE_SIMPLE_ATTRIBUTE(fops_x8, debugfs_u8_get, debugfs_u8_set, "0x%02llx\n"); -DEFINE_SIMPLE_ATTRIBUTE(fops_x8_ro, debugfs_u8_get, NULL, "0x%02llx\n"); -DEFINE_SIMPLE_ATTRIBUTE(fops_x8_wo, NULL, debugfs_u8_set, "0x%02llx\n"); +DEFINE_DEBUGFS_ATTRIBUTE(fops_x8, debugfs_u8_get, debugfs_u8_set, "0x%02llx\n"); +DEFINE_DEBUGFS_ATTRIBUTE(fops_x8_ro, debugfs_u8_get, NULL, "0x%02llx\n"); +DEFINE_DEBUGFS_ATTRIBUTE(fops_x8_wo, NULL, debugfs_u8_set, "0x%02llx\n"); -DEFINE_SIMPLE_ATTRIBUTE(fops_x16, debugfs_u16_get, debugfs_u16_set, "0x%04llx\n"); -DEFINE_SIMPLE_ATTRIBUTE(fops_x16_ro, debugfs_u16_get, NULL, "0x%04llx\n"); -DEFINE_SIMPLE_ATTRIBUTE(fops_x16_wo, NULL, debugfs_u16_set, "0x%04llx\n"); +DEFINE_DEBUGFS_ATTRIBUTE(fops_x16, debugfs_u16_get, debugfs_u16_set, + "0x%04llx\n"); +DEFINE_DEBUGFS_ATTRIBUTE(fops_x16_ro, debugfs_u16_get, NULL, "0x%04llx\n"); +DEFINE_DEBUGFS_ATTRIBUTE(fops_x16_wo, NULL, debugfs_u16_set, "0x%04llx\n"); -DEFINE_SIMPLE_ATTRIBUTE(fops_x32, debugfs_u32_get, debugfs_u32_set, "0x%08llx\n"); -DEFINE_SIMPLE_ATTRIBUTE(fops_x32_ro, debugfs_u32_get, NULL, "0x%08llx\n"); -DEFINE_SIMPLE_ATTRIBUTE(fops_x32_wo, NULL, debugfs_u32_set, "0x%08llx\n"); +DEFINE_DEBUGFS_ATTRIBUTE(fops_x32, debugfs_u32_get, debugfs_u32_set, + "0x%08llx\n"); +DEFINE_DEBUGFS_ATTRIBUTE(fops_x32_ro, debugfs_u32_get, NULL, "0x%08llx\n"); +DEFINE_DEBUGFS_ATTRIBUTE(fops_x32_wo, NULL, debugfs_u32_set, "0x%08llx\n"); -DEFINE_SIMPLE_ATTRIBUTE(fops_x64, debugfs_u64_get, debugfs_u64_set, "0x%016llx\n"); -DEFINE_SIMPLE_ATTRIBUTE(fops_x64_ro, debugfs_u64_get, NULL, "0x%016llx\n"); -DEFINE_SIMPLE_ATTRIBUTE(fops_x64_wo, NULL, debugfs_u64_set, "0x%016llx\n"); +DEFINE_DEBUGFS_ATTRIBUTE(fops_x64, debugfs_u64_get, debugfs_u64_set, + "0x%016llx\n"); +DEFINE_DEBUGFS_ATTRIBUTE(fops_x64_ro, debugfs_u64_get, NULL, "0x%016llx\n"); +DEFINE_DEBUGFS_ATTRIBUTE(fops_x64_wo, NULL, debugfs_u64_set, "0x%016llx\n"); /* * debugfs_create_x{8,16,32,64} - create a debugfs file that is used to read and write an unsigned {8,16,32,64}-bit value @@ -598,7 +621,7 @@ DEFINE_SIMPLE_ATTRIBUTE(fops_x64_wo, NULL, debugfs_u64_set, "0x%016llx\n"); struct dentry *debugfs_create_x8(const char *name, umode_t mode, struct dentry *parent, u8 *value) { - return debugfs_create_mode(name, mode, parent, value, &fops_x8, + return debugfs_create_mode_unsafe(name, mode, parent, value, &fops_x8, &fops_x8_ro, &fops_x8_wo); } EXPORT_SYMBOL_GPL(debugfs_create_x8); @@ -616,7 +639,7 @@ EXPORT_SYMBOL_GPL(debugfs_create_x8); struct dentry *debugfs_create_x16(const char *name, umode_t mode, struct dentry *parent, u16 *value) { - return debugfs_create_mode(name, mode, parent, value, &fops_x16, + return debugfs_create_mode_unsafe(name, mode, parent, value, &fops_x16, &fops_x16_ro, &fops_x16_wo); } EXPORT_SYMBOL_GPL(debugfs_create_x16); @@ -634,7 +657,7 @@ EXPORT_SYMBOL_GPL(debugfs_create_x16); struct dentry *debugfs_create_x32(const char *name, umode_t mode, struct dentry *parent, u32 *value) { - return debugfs_create_mode(name, mode, parent, value, &fops_x32, + return debugfs_create_mode_unsafe(name, mode, parent, value, &fops_x32, &fops_x32_ro, &fops_x32_wo); } EXPORT_SYMBOL_GPL(debugfs_create_x32); @@ -652,7 +675,7 @@ EXPORT_SYMBOL_GPL(debugfs_create_x32); struct dentry *debugfs_create_x64(const char *name, umode_t mode, struct dentry *parent, u64 *value) { - return debugfs_create_mode(name, mode, parent, value, &fops_x64, + return debugfs_create_mode_unsafe(name, mode, parent, value, &fops_x64, &fops_x64_ro, &fops_x64_wo); } EXPORT_SYMBOL_GPL(debugfs_create_x64); @@ -668,10 +691,10 @@ static int debugfs_size_t_get(void *data, u64 *val) *val = *(size_t *)data; return 0; } -DEFINE_SIMPLE_ATTRIBUTE(fops_size_t, debugfs_size_t_get, debugfs_size_t_set, - "%llu\n"); /* %llu and %zu are more or less the same */ -DEFINE_SIMPLE_ATTRIBUTE(fops_size_t_ro, debugfs_size_t_get, NULL, "%llu\n"); -DEFINE_SIMPLE_ATTRIBUTE(fops_size_t_wo, NULL, debugfs_size_t_set, "%llu\n"); +DEFINE_DEBUGFS_ATTRIBUTE(fops_size_t, debugfs_size_t_get, debugfs_size_t_set, + "%llu\n"); /* %llu and %zu are more or less the same */ +DEFINE_DEBUGFS_ATTRIBUTE(fops_size_t_ro, debugfs_size_t_get, NULL, "%llu\n"); +DEFINE_DEBUGFS_ATTRIBUTE(fops_size_t_wo, NULL, debugfs_size_t_set, "%llu\n"); /** * debugfs_create_size_t - create a debugfs file that is used to read and write an size_t value @@ -686,8 +709,9 @@ DEFINE_SIMPLE_ATTRIBUTE(fops_size_t_wo, NULL, debugfs_size_t_set, "%llu\n"); struct dentry *debugfs_create_size_t(const char *name, umode_t mode, struct dentry *parent, size_t *value) { - return debugfs_create_mode(name, mode, parent, value, &fops_size_t, - &fops_size_t_ro, &fops_size_t_wo); + return debugfs_create_mode_unsafe(name, mode, parent, value, + &fops_size_t, &fops_size_t_ro, + &fops_size_t_wo); } EXPORT_SYMBOL_GPL(debugfs_create_size_t); @@ -701,10 +725,12 @@ static int debugfs_atomic_t_get(void *data, u64 *val) *val = atomic_read((atomic_t *)data); return 0; } -DEFINE_SIMPLE_ATTRIBUTE(fops_atomic_t, debugfs_atomic_t_get, +DEFINE_DEBUGFS_ATTRIBUTE(fops_atomic_t, debugfs_atomic_t_get, debugfs_atomic_t_set, "%lld\n"); -DEFINE_SIMPLE_ATTRIBUTE(fops_atomic_t_ro, debugfs_atomic_t_get, NULL, "%lld\n"); -DEFINE_SIMPLE_ATTRIBUTE(fops_atomic_t_wo, NULL, debugfs_atomic_t_set, "%lld\n"); +DEFINE_DEBUGFS_ATTRIBUTE(fops_atomic_t_ro, debugfs_atomic_t_get, NULL, + "%lld\n"); +DEFINE_DEBUGFS_ATTRIBUTE(fops_atomic_t_wo, NULL, debugfs_atomic_t_set, + "%lld\n"); /** * debugfs_create_atomic_t - create a debugfs file that is used to read and @@ -720,8 +746,9 @@ DEFINE_SIMPLE_ATTRIBUTE(fops_atomic_t_wo, NULL, debugfs_atomic_t_set, "%lld\n"); struct dentry *debugfs_create_atomic_t(const char *name, umode_t mode, struct dentry *parent, atomic_t *value) { - return debugfs_create_mode(name, mode, parent, value, &fops_atomic_t, - &fops_atomic_t_ro, &fops_atomic_t_wo); + return debugfs_create_mode_unsafe(name, mode, parent, value, + &fops_atomic_t, &fops_atomic_t_ro, + &fops_atomic_t_wo); } EXPORT_SYMBOL_GPL(debugfs_create_atomic_t); -- cgit v1.3-8-gc7d7 From 4d45f7974ccf0aa783034fef2661573b3a28609e Mon Sep 17 00:00:00 2001 From: Nicolai Stange Date: Tue, 22 Mar 2016 14:11:18 +0100 Subject: debugfs: unproxify files created through debugfs_create_bool() Currently, the struct file_operations fops_bool associated with files created through the debugfs_create_bool() helpers are not file lifetime aware. Thus, a lifetime managing proxy is created around fops_bool each time such a file is opened which is an unnecessary waste of resources. Implement file lifetime management for the fops_bool file_operations. Namely, make debugfs_read_file_bool() and debugfs_write_file_bool() safe against file removals by means of debugfs_use_file_start() and debugfs_use_file_finish(). Make debugfs_create_bool() create its files in non-proxying operation mode through debugfs_create_mode_unsafe(). Finally, purge debugfs_create_mode() as debugfs_create_bool() had been its last user. Signed-off-by: Nicolai Stange Signed-off-by: Greg Kroah-Hartman --- fs/debugfs/file.c | 41 ++++++++++++++++++++--------------------- 1 file changed, 20 insertions(+), 21 deletions(-) (limited to 'fs') diff --git a/fs/debugfs/file.c b/fs/debugfs/file.c index 4b3967e86e97..8a548bee1b3d 100644 --- a/fs/debugfs/file.c +++ b/fs/debugfs/file.c @@ -312,22 +312,6 @@ ssize_t debugfs_attr_write(struct file *file, const char __user *buf, } EXPORT_SYMBOL_GPL(debugfs_attr_write); -static struct dentry *debugfs_create_mode(const char *name, umode_t mode, - struct dentry *parent, void *value, - const struct file_operations *fops, - const struct file_operations *fops_ro, - const struct file_operations *fops_wo) -{ - /* if there are no write bits set, make read only */ - if (!(mode & S_IWUGO)) - return debugfs_create_file(name, mode, parent, value, fops_ro); - /* if there are no read bits set, make write only */ - if (!(mode & S_IRUGO)) - return debugfs_create_file(name, mode, parent, value, fops_wo); - - return debugfs_create_file(name, mode, parent, value, fops); -} - static struct dentry *debugfs_create_mode_unsafe(const char *name, umode_t mode, struct dentry *parent, void *value, const struct file_operations *fops, @@ -756,9 +740,17 @@ ssize_t debugfs_read_file_bool(struct file *file, char __user *user_buf, size_t count, loff_t *ppos) { char buf[3]; - bool *val = file->private_data; + bool val; + int r, srcu_idx; + + r = debugfs_use_file_start(F_DENTRY(file), &srcu_idx); + if (likely(!r)) + val = *(bool *)file->private_data; + debugfs_use_file_finish(srcu_idx); + if (r) + return r; - if (*val) + if (val) buf[0] = 'Y'; else buf[0] = 'N'; @@ -774,6 +766,7 @@ ssize_t debugfs_write_file_bool(struct file *file, const char __user *user_buf, char buf[32]; size_t buf_size; bool bv; + int r, srcu_idx; bool *val = file->private_data; buf_size = min(count, (sizeof(buf)-1)); @@ -781,8 +774,14 @@ ssize_t debugfs_write_file_bool(struct file *file, const char __user *user_buf, return -EFAULT; buf[buf_size] = '\0'; - if (strtobool(buf, &bv) == 0) - *val = bv; + if (strtobool(buf, &bv) == 0) { + r = debugfs_use_file_start(F_DENTRY(file), &srcu_idx); + if (likely(!r)) + *val = bv; + debugfs_use_file_finish(srcu_idx); + if (r) + return r; + } return count; } @@ -834,7 +833,7 @@ static const struct file_operations fops_bool_wo = { struct dentry *debugfs_create_bool(const char *name, umode_t mode, struct dentry *parent, bool *value) { - return debugfs_create_mode(name, mode, parent, value, &fops_bool, + return debugfs_create_mode_unsafe(name, mode, parent, value, &fops_bool, &fops_bool_ro, &fops_bool_wo); } EXPORT_SYMBOL_GPL(debugfs_create_bool); -- cgit v1.3-8-gc7d7 From 83b711cbf4ff42a9996c5f092762b3967d307d73 Mon Sep 17 00:00:00 2001 From: Nicolai Stange Date: Tue, 22 Mar 2016 14:11:19 +0100 Subject: debugfs: unproxify files created through debugfs_create_blob() Currently, the struct file_operations fops_blob associated with files created through the debugfs_create_blob() helpers are not file lifetime aware. Thus, a lifetime managing proxy is created around fops_blob each time such a file is opened which is an unnecessary waste of resources. Implement file lifetime management for the fops_bool file_operations. Namely, make read_file_blob() safe gainst file removals by means of debugfs_use_file_start() and debugfs_use_file_finish(). Make debugfs_create_blob() create its files in non-proxying operation mode by means of debugfs_create_file_unsafe(). Signed-off-by: Nicolai Stange Signed-off-by: Greg Kroah-Hartman --- fs/debugfs/file.c | 13 ++++++++++--- 1 file changed, 10 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/debugfs/file.c b/fs/debugfs/file.c index 8a548bee1b3d..2e86d66f7850 100644 --- a/fs/debugfs/file.c +++ b/fs/debugfs/file.c @@ -842,8 +842,15 @@ static ssize_t read_file_blob(struct file *file, char __user *user_buf, size_t count, loff_t *ppos) { struct debugfs_blob_wrapper *blob = file->private_data; - return simple_read_from_buffer(user_buf, count, ppos, blob->data, - blob->size); + ssize_t r; + int srcu_idx; + + r = debugfs_use_file_start(F_DENTRY(file), &srcu_idx); + if (likely(!r)) + r = simple_read_from_buffer(user_buf, count, ppos, blob->data, + blob->size); + debugfs_use_file_finish(srcu_idx); + return r; } static const struct file_operations fops_blob = { @@ -880,7 +887,7 @@ struct dentry *debugfs_create_blob(const char *name, umode_t mode, struct dentry *parent, struct debugfs_blob_wrapper *blob) { - return debugfs_create_file(name, mode, parent, blob, &fops_blob); + return debugfs_create_file_unsafe(name, mode, parent, blob, &fops_blob); } EXPORT_SYMBOL_GPL(debugfs_create_blob); -- cgit v1.3-8-gc7d7 From c4a74f63dfd2e75e7d40a9aaa4052b0ef26e617c Mon Sep 17 00:00:00 2001 From: Nicolai Stange Date: Tue, 22 Mar 2016 14:11:20 +0100 Subject: debugfs: unproxify files created through debugfs_create_u32_array() The struct file_operations u32_array_fops associated with files created through debugfs_create_u32_array() has been lifetime aware already: everything needed for subsequent operation is copied to a ->f_private buffer at file opening time in u32_array_open(). Now, ->open() is always protected against file removal issues by the debugfs core. There is no need for the debugfs core to wrap the u32_array_fops with a file lifetime managing proxy. Make debugfs_create_u32_array() create its files in non-proxying operation mode by means of debugfs_create_file_unsafe(). Signed-off-by: Nicolai Stange Signed-off-by: Greg Kroah-Hartman --- fs/debugfs/file.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/debugfs/file.c b/fs/debugfs/file.c index 2e86d66f7850..9c1c9a01b7e5 100644 --- a/fs/debugfs/file.c +++ b/fs/debugfs/file.c @@ -992,7 +992,8 @@ struct dentry *debugfs_create_u32_array(const char *name, umode_t mode, data->array = array; data->elements = elements; - return debugfs_create_file(name, mode, parent, data, &u32_array_fops); + return debugfs_create_file_unsafe(name, mode, parent, data, + &u32_array_fops); } EXPORT_SYMBOL_GPL(debugfs_create_u32_array); -- cgit v1.3-8-gc7d7 From e97321fa095f1ea7110d4d2ba446bd6141ed9a03 Mon Sep 17 00:00:00 2001 From: Bob Peterson Date: Tue, 12 Apr 2016 16:14:26 -0400 Subject: GFS2: Don't dereference inode in gfs2_inode_lookup until it's valid Function gfs2_inode_lookup was dereferencing the inode, and after, it checks for the value being NULL. We need to check that first. Signed-off-by: Bob Peterson --- fs/gfs2/inode.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c index bb30f9a72c65..aea002ea94a6 100644 --- a/fs/gfs2/inode.c +++ b/fs/gfs2/inode.c @@ -93,12 +93,12 @@ struct inode *gfs2_inode_lookup(struct super_block *sb, unsigned int type, int error; inode = iget_locked(sb, (unsigned long)no_addr); - ip = GFS2_I(inode); - ip->i_no_addr = no_addr; - if (!inode) return ERR_PTR(-ENOMEM); + ip = GFS2_I(inode); + ip->i_no_addr = no_addr; + if (inode->i_state & I_NEW) { struct gfs2_sbd *sdp = GFS2_SB(inode); ip->i_no_formal_ino = no_formal_ino; -- cgit v1.3-8-gc7d7 From f2353d7bd4221dfe0e230c5be49038d4a57d97e9 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Wed, 23 Mar 2016 10:42:01 -0700 Subject: f2fs: give RO message when recovering superblock When one of superblocks is missing, f2fs recovers it with the valid one. But, even if f2fs is mounted as RO, we'd better notify that too. Signed-off-by: Jaegeuk Kim --- fs/f2fs/super.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 006f87d69921..ffe4616376ca 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -1298,6 +1298,9 @@ int f2fs_commit_super(struct f2fs_sb_info *sbi, bool recover) struct buffer_head *bh; int err; + if (f2fs_readonly(sbi->sb) || bdev_read_only(sbi->sb->s_bdev)) + return -EROFS; + /* write back-up superblock first */ bh = sb_getblk(sbi->sb, sbi->valid_super_block ? 0: 1); if (!bh) @@ -1565,7 +1568,7 @@ try_onemore: kfree(options); /* recover broken superblock */ - if (recovery && !f2fs_readonly(sb) && !bdev_read_only(sb->s_bdev)) { + if (recovery) { err = f2fs_commit_super(sbi, true); f2fs_msg(sb, KERN_INFO, "Try to recover %dth superblock, ret: %ld", -- cgit v1.3-8-gc7d7 From df728b0f6954c38545f4bf12dedeeb9e07469a94 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Wed, 23 Mar 2016 17:05:27 -0700 Subject: f2fs: recover superblock at RW remounts This patch adds a sbi flag, SBI_NEED_SB_WRITE, which indicates it needs to recover superblock when (re)mounting as RW. This is set only when f2fs is mounted as RO. Signed-off-by: Jaegeuk Kim --- fs/f2fs/f2fs.h | 1 + fs/f2fs/super.c | 36 +++++++++++++++++++++++++++--------- 2 files changed, 28 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 7a4558d17f36..7cb360e02109 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -672,6 +672,7 @@ enum { SBI_IS_CLOSE, /* specify unmounting */ SBI_NEED_FSCK, /* need fsck.f2fs to fix */ SBI_POR_DOING, /* recovery is doing or not */ + SBI_NEED_SB_WRITE, /* need to recover superblock */ }; enum { diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index ffe4616376ca..f5fbbfdb3d93 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -796,6 +796,15 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data) set_sbi_flag(sbi, SBI_IS_DIRTY); } + /* recover superblocks we couldn't write due to previous RO mount */ + if (!(*flags & MS_RDONLY) && is_sbi_flag_set(sbi, SBI_NEED_SB_WRITE)) { + err = f2fs_commit_super(sbi, false); + f2fs_msg(sb, KERN_INFO, + "Try to recover all the superblocks, ret: %d", err); + if (!err) + clear_sbi_flag(sbi, SBI_NEED_SB_WRITE); + } + sync_filesystem(sb); sbi->mount_opt.opt = 0; @@ -852,8 +861,9 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data) } skip: /* Update the POSIXACL Flag */ - sb->s_flags = (sb->s_flags & ~MS_POSIXACL) | + sb->s_flags = (sb->s_flags & ~MS_POSIXACL) | (test_opt(sbi, POSIX_ACL) ? MS_POSIXACL : 0); + return 0; restore_gc: if (need_restart_gc) { @@ -998,11 +1008,12 @@ static int __f2fs_commit_super(struct buffer_head *bh, return __sync_dirty_buffer(bh, WRITE_FLUSH_FUA); } -static inline bool sanity_check_area_boundary(struct super_block *sb, +static inline bool sanity_check_area_boundary(struct f2fs_sb_info *sbi, struct buffer_head *bh) { struct f2fs_super_block *raw_super = (struct f2fs_super_block *) (bh->b_data + F2FS_SUPER_OFFSET); + struct super_block *sb = sbi->sb; u32 segment0_blkaddr = le32_to_cpu(raw_super->segment0_blkaddr); u32 cp_blkaddr = le32_to_cpu(raw_super->cp_blkaddr); u32 sit_blkaddr = le32_to_cpu(raw_super->sit_blkaddr); @@ -1081,6 +1092,7 @@ static inline bool sanity_check_area_boundary(struct super_block *sb, segment0_blkaddr) >> log_blocks_per_seg); if (f2fs_readonly(sb) || bdev_read_only(sb->s_bdev)) { + set_sbi_flag(sbi, SBI_NEED_SB_WRITE); res = "internally"; } else { err = __f2fs_commit_super(bh, NULL); @@ -1098,11 +1110,12 @@ static inline bool sanity_check_area_boundary(struct super_block *sb, return false; } -static int sanity_check_raw_super(struct super_block *sb, +static int sanity_check_raw_super(struct f2fs_sb_info *sbi, struct buffer_head *bh) { struct f2fs_super_block *raw_super = (struct f2fs_super_block *) (bh->b_data + F2FS_SUPER_OFFSET); + struct super_block *sb = sbi->sb; unsigned int blocksize; if (F2FS_SUPER_MAGIC != le32_to_cpu(raw_super->magic)) { @@ -1169,7 +1182,7 @@ static int sanity_check_raw_super(struct super_block *sb, } /* check CP/SIT/NAT/SSA/MAIN_AREA area boundary */ - if (sanity_check_area_boundary(sb, bh)) + if (sanity_check_area_boundary(sbi, bh)) return 1; return 0; @@ -1239,10 +1252,11 @@ static void init_sb_info(struct f2fs_sb_info *sbi) * to get the first valid one. If any one of them is broken, we pass * them recovery flag back to the caller. */ -static int read_raw_super_block(struct super_block *sb, +static int read_raw_super_block(struct f2fs_sb_info *sbi, struct f2fs_super_block **raw_super, int *valid_super_block, int *recovery) { + struct super_block *sb = sbi->sb; int block; struct buffer_head *bh; struct f2fs_super_block *super; @@ -1262,7 +1276,7 @@ static int read_raw_super_block(struct super_block *sb, } /* sanity checking of raw super */ - if (sanity_check_raw_super(sb, bh)) { + if (sanity_check_raw_super(sbi, bh)) { f2fs_msg(sb, KERN_ERR, "Can't find valid F2FS filesystem in %dth superblock", block + 1); @@ -1298,8 +1312,11 @@ int f2fs_commit_super(struct f2fs_sb_info *sbi, bool recover) struct buffer_head *bh; int err; - if (f2fs_readonly(sbi->sb) || bdev_read_only(sbi->sb->s_bdev)) + if ((recover && f2fs_readonly(sbi->sb)) || + bdev_read_only(sbi->sb->s_bdev)) { + set_sbi_flag(sbi, SBI_NEED_SB_WRITE); return -EROFS; + } /* write back-up superblock first */ bh = sb_getblk(sbi->sb, sbi->valid_super_block ? 0: 1); @@ -1343,6 +1360,8 @@ try_onemore: if (!sbi) return -ENOMEM; + sbi->sb = sb; + /* Load the checksum driver */ sbi->s_chksum_driver = crypto_alloc_shash("crc32", 0, 0); if (IS_ERR(sbi->s_chksum_driver)) { @@ -1358,7 +1377,7 @@ try_onemore: goto free_sbi; } - err = read_raw_super_block(sb, &raw_super, &valid_super_block, + err = read_raw_super_block(sbi, &raw_super, &valid_super_block, &recovery); if (err) goto free_sbi; @@ -1393,7 +1412,6 @@ try_onemore: memcpy(sb->s_uuid, raw_super->uuid, sizeof(raw_super->uuid)); /* init f2fs-specific super block info */ - sbi->sb = sb; sbi->raw_super = raw_super; sbi->valid_super_block = valid_super_block; mutex_init(&sbi->gc_mutex); -- cgit v1.3-8-gc7d7 From 6781eabba1bdb133eb9125c4acf6704ccbe4df02 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Wed, 23 Mar 2016 16:12:58 -0700 Subject: f2fs: give -EINVAL for norecovery and rw mount Once detecting something to recover, f2fs should stop mounting, given norecovery and rw mount options. Signed-off-by: Jaegeuk Kim --- fs/f2fs/f2fs.h | 2 +- fs/f2fs/recovery.c | 11 +++++++---- fs/f2fs/super.c | 14 ++++++++++++-- 3 files changed, 20 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 7cb360e02109..e1c07b60f301 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -1907,7 +1907,7 @@ void build_gc_manager(struct f2fs_sb_info *); /* * recovery.c */ -int recover_fsync_data(struct f2fs_sb_info *); +int recover_fsync_data(struct f2fs_sb_info *, bool); bool space_for_roll_forward(struct f2fs_sb_info *); /* diff --git a/fs/f2fs/recovery.c b/fs/f2fs/recovery.c index 011942f94d64..2c87c12b6f1c 100644 --- a/fs/f2fs/recovery.c +++ b/fs/f2fs/recovery.c @@ -551,12 +551,13 @@ next: return err; } -int recover_fsync_data(struct f2fs_sb_info *sbi) +int recover_fsync_data(struct f2fs_sb_info *sbi, bool check_only) { struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_WARM_NODE); struct list_head inode_list; block_t blkaddr; int err; + int ret = 0; bool need_writecp = false; fsync_entry_slab = f2fs_kmem_cache_create("f2fs_fsync_inode_entry", @@ -573,11 +574,13 @@ int recover_fsync_data(struct f2fs_sb_info *sbi) /* step #1: find fsynced inode numbers */ err = find_fsync_dnodes(sbi, &inode_list); - if (err) + if (err || list_empty(&inode_list)) goto out; - if (list_empty(&inode_list)) + if (check_only) { + ret = 1; goto out; + } need_writecp = true; @@ -625,5 +628,5 @@ out: } else { mutex_unlock(&sbi->cp_mutex); } - return err; + return ret ? ret: err; } diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index f5fbbfdb3d93..8f9648ffbbc3 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -1562,14 +1562,24 @@ try_onemore: if (need_fsck) set_sbi_flag(sbi, SBI_NEED_FSCK); - err = recover_fsync_data(sbi); - if (err) { + err = recover_fsync_data(sbi, false); + if (err < 0) { need_fsck = true; f2fs_msg(sb, KERN_ERR, "Cannot recover all fsync data errno=%ld", err); goto free_kobj; } + } else { + err = recover_fsync_data(sbi, true); + + if (!f2fs_readonly(sb) && err > 0) { + err = -EINVAL; + f2fs_msg(sb, KERN_ERR, + "Need to recover fsync data"); + goto free_kobj; + } } + /* recover_fsync_data() cleared this already */ clear_sbi_flag(sbi, SBI_POR_DOING); -- cgit v1.3-8-gc7d7 From faa0e55bba54b91c464850dc28376056d26fa310 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Thu, 24 Mar 2016 10:29:39 -0700 Subject: f2fs: treat as a normal umount when remounting ro When user remounts f2fs as read-only, we can mark the checkpoint as umount. Signed-off-by: Jaegeuk Kim --- fs/f2fs/super.c | 18 ++++++++++-------- 1 file changed, 10 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 8f9648ffbbc3..19a85cf83f40 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -791,11 +791,6 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data) org_mount_opt = sbi->mount_opt; active_logs = sbi->active_logs; - if (*flags & MS_RDONLY) { - set_opt(sbi, FASTBOOT); - set_sbi_flag(sbi, SBI_IS_DIRTY); - } - /* recover superblocks we couldn't write due to previous RO mount */ if (!(*flags & MS_RDONLY) && is_sbi_flag_set(sbi, SBI_NEED_SB_WRITE)) { err = f2fs_commit_super(sbi, false); @@ -805,8 +800,6 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data) clear_sbi_flag(sbi, SBI_NEED_SB_WRITE); } - sync_filesystem(sb); - sbi->mount_opt.opt = 0; default_options(sbi); @@ -838,7 +831,6 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data) if ((*flags & MS_RDONLY) || !test_opt(sbi, BG_GC)) { if (sbi->gc_thread) { stop_gc_thread(sbi); - f2fs_sync_fs(sb, 1); need_restart_gc = true; } } else if (!sbi->gc_thread) { @@ -848,6 +840,16 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data) need_stop_gc = true; } + if (*flags & MS_RDONLY) { + writeback_inodes_sb(sb, WB_REASON_SYNC); + sync_inodes_sb(sb); + + set_sbi_flag(sbi, SBI_IS_DIRTY); + set_sbi_flag(sbi, SBI_IS_CLOSE); + f2fs_sync_fs(sb, 1); + clear_sbi_flag(sbi, SBI_IS_CLOSE); + } + /* * We stop issue flush thread if FS is mounted as RO * or if flush_merge is not passed in mount option. -- cgit v1.3-8-gc7d7 From 8c11a53fc2557eb247e3eaa4d554d45c1b55fc98 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Fri, 18 Mar 2016 09:46:10 -0700 Subject: f2fs: show current mount status This patch remains the current mount status to f2fs status info. Signed-off-by: Jaegeuk Kim --- fs/f2fs/debug.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/debug.c b/fs/f2fs/debug.c index f4a61a5ff79f..818064bacfda 100644 --- a/fs/f2fs/debug.c +++ b/fs/f2fs/debug.c @@ -216,8 +216,9 @@ static int stat_show(struct seq_file *s, void *v) list_for_each_entry(si, &f2fs_stat_list, stat_list) { update_general_status(si->sbi); - seq_printf(s, "\n=====[ partition info(%pg). #%d ]=====\n", - si->sbi->sb->s_bdev, i++); + seq_printf(s, "\n=====[ partition info(%pg). #%d, %s]=====\n", + si->sbi->sb->s_bdev, i++, + f2fs_readonly(si->sbi->sb) ? "RO": "RW"); seq_printf(s, "[SB: 1] [CP: 2] [SIT: %d] [NAT: %d] ", si->sit_area_segs, si->nat_area_segs); seq_printf(s, "[SSA: %d] [MAIN: %d", -- cgit v1.3-8-gc7d7 From 675f10bde6cc3874632a8f684df2a8a2a8ace76e Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Mon, 22 Feb 2016 18:29:18 +0800 Subject: f2fs: fix to convert inline directory correctly With below serials, we will lose parts of dirents: 1) mount f2fs with inline_dentry option 2) echo 1 > /sys/fs/f2fs/sdX/dir_level 3) mkdir dir 4) touch 180 files named [1-180] in dir 5) touch 181 in dir 6) echo 3 > /proc/sys/vm/drop_caches 7) ll dir ls: cannot access 2: No such file or directory ls: cannot access 4: No such file or directory ls: cannot access 5: No such file or directory ls: cannot access 6: No such file or directory ls: cannot access 8: No such file or directory ls: cannot access 9: No such file or directory ... total 360 drwxr-xr-x 2 root root 4096 Feb 19 15:12 ./ drwxr-xr-x 3 root root 4096 Feb 19 15:11 ../ -rw-r--r-- 1 root root 0 Feb 19 15:12 1 -rw-r--r-- 1 root root 0 Feb 19 15:12 10 -rw-r--r-- 1 root root 0 Feb 19 15:12 100 -????????? ? ? ? ? ? 101 -????????? ? ? ? ? ? 102 -????????? ? ? ? ? ? 103 ... The reason is: when doing the inline dir conversion, we didn't consider that directory has hierarchical hash structure which can be configured through sysfs interface 'dir_level'. By default, dir_level of directory inode is 0, it means we have one bucket in hash table located in first level, all dirents will be hashed in this bucket, so it has no problem for us to do the duplication simply between inline dentry page and converted normal dentry page. However, if we configured dir_level with the value N (greater than 0), it will expand the bucket number of first level hash table by 2^N - 1, it hashs dirents into different buckets according their hash value, if we still move all dirents to first bucket, it makes incorrent locating for inline dirents, the result is, although we can iterate all dirents through ->readdir, we can't stat some of them in ->lookup which based on hash table searching. This patch fixes this issue by rehashing dirents into correct position when converting inline directory. Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/dir.c | 87 ++++++++++++++++++++++++--------------------- fs/f2fs/f2fs.h | 4 ++- fs/f2fs/inline.c | 94 ++++++++++++++++++++++++++++++++++++++++++++++++- include/linux/f2fs_fs.h | 2 ++ 4 files changed, 144 insertions(+), 43 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/dir.c b/fs/f2fs/dir.c index af819571bce7..e90380d82214 100644 --- a/fs/f2fs/dir.c +++ b/fs/f2fs/dir.c @@ -48,7 +48,6 @@ unsigned char f2fs_filetype_table[F2FS_FT_MAX] = { [F2FS_FT_SYMLINK] = DT_LNK, }; -#define S_SHIFT 12 static unsigned char f2fs_type_by_mode[S_IFMT >> S_SHIFT] = { [S_IFREG >> S_SHIFT] = F2FS_FT_REG_FILE, [S_IFDIR >> S_SHIFT] = F2FS_FT_DIR, @@ -64,6 +63,13 @@ void set_de_type(struct f2fs_dir_entry *de, umode_t mode) de->file_type = f2fs_type_by_mode[(mode & S_IFMT) >> S_SHIFT]; } +unsigned char get_de_type(struct f2fs_dir_entry *de) +{ + if (de->file_type < F2FS_FT_MAX) + return f2fs_filetype_table[de->file_type]; + return DT_UNKNOWN; +} + static unsigned long dir_block_index(unsigned int level, int dir_level, unsigned int idx) { @@ -509,11 +515,7 @@ void f2fs_update_dentry(nid_t ino, umode_t mode, struct f2fs_dentry_ptr *d, } } -/* - * Caller should grab and release a rwsem by calling f2fs_lock_op() and - * f2fs_unlock_op(). - */ -int __f2fs_add_link(struct inode *dir, const struct qstr *name, +int f2fs_add_regular_entry(struct inode *dir, const struct qstr *new_name, struct inode *inode, nid_t ino, umode_t mode) { unsigned int bit_pos; @@ -526,28 +528,11 @@ int __f2fs_add_link(struct inode *dir, const struct qstr *name, struct f2fs_dentry_block *dentry_blk = NULL; struct f2fs_dentry_ptr d; struct page *page = NULL; - struct fscrypt_name fname; - struct qstr new_name; - int slots, err; - - err = fscrypt_setup_filename(dir, name, 0, &fname); - if (err) - return err; - - new_name.name = fname_name(&fname); - new_name.len = fname_len(&fname); - - if (f2fs_has_inline_dentry(dir)) { - err = f2fs_add_inline_entry(dir, &new_name, inode, ino, mode); - if (!err || err != -EAGAIN) - goto out; - else - err = 0; - } + int slots, err = 0; level = 0; - slots = GET_DENTRY_SLOTS(new_name.len); - dentry_hash = f2fs_dentry_hash(&new_name); + slots = GET_DENTRY_SLOTS(new_name->len); + dentry_hash = f2fs_dentry_hash(new_name); current_depth = F2FS_I(dir)->i_current_depth; if (F2FS_I(dir)->chash == dentry_hash) { @@ -556,10 +541,8 @@ int __f2fs_add_link(struct inode *dir, const struct qstr *name, } start: - if (unlikely(current_depth == MAX_DIR_HASH_DEPTH)) { - err = -ENOSPC; - goto out; - } + if (unlikely(current_depth == MAX_DIR_HASH_DEPTH)) + return -ENOSPC; /* Increase the depth, if required */ if (level == current_depth) @@ -573,10 +556,8 @@ start: for (block = bidx; block <= (bidx + nblock - 1); block++) { dentry_page = get_new_data_page(dir, NULL, block, true); - if (IS_ERR(dentry_page)) { - err = PTR_ERR(dentry_page); - goto out; - } + if (IS_ERR(dentry_page)) + return PTR_ERR(dentry_page); dentry_blk = kmap(dentry_page); bit_pos = room_for_filename(&dentry_blk->dentry_bitmap, @@ -596,7 +577,7 @@ add_dentry: if (inode) { down_write(&F2FS_I(inode)->i_sem); - page = init_inode_metadata(inode, dir, &new_name, NULL); + page = init_inode_metadata(inode, dir, new_name, NULL); if (IS_ERR(page)) { err = PTR_ERR(page); goto fail; @@ -606,7 +587,7 @@ add_dentry: } make_dentry_ptr(NULL, &d, (void *)dentry_blk, 1); - f2fs_update_dentry(ino, mode, &d, &new_name, dentry_hash, bit_pos); + f2fs_update_dentry(ino, mode, &d, new_name, dentry_hash, bit_pos); set_page_dirty(dentry_page); @@ -628,7 +609,34 @@ fail: } kunmap(dentry_page); f2fs_put_page(dentry_page, 1); -out: + + return err; +} + +/* + * Caller should grab and release a rwsem by calling f2fs_lock_op() and + * f2fs_unlock_op(). + */ +int __f2fs_add_link(struct inode *dir, const struct qstr *name, + struct inode *inode, nid_t ino, umode_t mode) +{ + struct fscrypt_name fname; + struct qstr new_name; + int err; + + err = fscrypt_setup_filename(dir, name, 0, &fname); + if (err) + return err; + + new_name.name = fname_name(&fname); + new_name.len = fname_len(&fname); + + err = -EAGAIN; + if (f2fs_has_inline_dentry(dir)) + err = f2fs_add_inline_entry(dir, &new_name, inode, ino, mode); + if (err == -EAGAIN) + err = f2fs_add_regular_entry(dir, &new_name, inode, ino, mode); + fscrypt_free_filename(&fname); f2fs_update_time(F2FS_I_SB(dir), REQ_TIME); return err; @@ -792,10 +800,7 @@ bool f2fs_fill_dentries(struct dir_context *ctx, struct f2fs_dentry_ptr *d, continue; } - if (de->file_type < F2FS_FT_MAX) - d_type = f2fs_filetype_table[de->file_type]; - else - d_type = DT_UNKNOWN; + d_type = get_de_type(de); de_name.name = d->filename[bit_pos]; de_name.len = le16_to_cpu(de->name_len); diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index e1c07b60f301..3f1551395244 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -1711,7 +1711,7 @@ struct dentry *f2fs_get_parent(struct dentry *child); */ extern unsigned char f2fs_filetype_table[F2FS_FT_MAX]; void set_de_type(struct f2fs_dir_entry *, umode_t); - +unsigned char get_de_type(struct f2fs_dir_entry *); struct f2fs_dir_entry *find_target_dentry(struct fscrypt_name *, f2fs_hash_t, int *, struct f2fs_dentry_ptr *); bool f2fs_fill_dentries(struct dir_context *, struct f2fs_dentry_ptr *, @@ -1732,6 +1732,8 @@ void f2fs_set_link(struct inode *, struct f2fs_dir_entry *, int update_dent_inode(struct inode *, struct inode *, const struct qstr *); void f2fs_update_dentry(nid_t ino, umode_t mode, struct f2fs_dentry_ptr *, const struct qstr *, f2fs_hash_t , unsigned int); +int f2fs_add_regular_entry(struct inode *, const struct qstr *, + struct inode *, nid_t, umode_t); int __f2fs_add_link(struct inode *, const struct qstr *, struct inode *, nid_t, umode_t); void f2fs_delete_entry(struct f2fs_dir_entry *, struct page *, struct inode *, diff --git a/fs/f2fs/inline.c b/fs/f2fs/inline.c index a2fbe6f427d3..772056587eb9 100644 --- a/fs/f2fs/inline.c +++ b/fs/f2fs/inline.c @@ -355,7 +355,7 @@ int make_empty_inline_dir(struct inode *inode, struct inode *parent, * NOTE: ipage is grabbed by caller, but if any error occurs, we should * release ipage in this function. */ -static int f2fs_convert_inline_dir(struct inode *dir, struct page *ipage, +static int f2fs_move_inline_dirents(struct inode *dir, struct page *ipage, struct f2fs_inline_dentry *inline_dentry) { struct page *page; @@ -416,6 +416,98 @@ out: return err; } +static int f2fs_add_inline_entries(struct inode *dir, + struct f2fs_inline_dentry *inline_dentry) +{ + struct f2fs_dentry_ptr d; + unsigned long bit_pos = 0; + int err = 0; + + make_dentry_ptr(NULL, &d, (void *)inline_dentry, 2); + + while (bit_pos < d.max) { + struct f2fs_dir_entry *de; + struct qstr new_name; + nid_t ino; + umode_t fake_mode; + + if (!test_bit_le(bit_pos, d.bitmap)) { + bit_pos++; + continue; + } + + de = &d.dentry[bit_pos]; + new_name.name = d.filename[bit_pos]; + new_name.len = de->name_len; + + ino = le32_to_cpu(de->ino); + fake_mode = get_de_type(de) << S_SHIFT; + + err = f2fs_add_regular_entry(dir, &new_name, NULL, + ino, fake_mode); + if (err) + goto punch_dentry_pages; + + if (unlikely(!de->name_len)) + d.max = -1; + + bit_pos += GET_DENTRY_SLOTS(le16_to_cpu(de->name_len)); + } + return 0; +punch_dentry_pages: + truncate_inode_pages(&dir->i_data, 0); + truncate_blocks(dir, 0, false); + remove_dirty_inode(dir); + return err; +} + +static int f2fs_move_rehashed_dirents(struct inode *dir, struct page *ipage, + struct f2fs_inline_dentry *inline_dentry) +{ + struct f2fs_inline_dentry *backup_dentry; + int err; + + backup_dentry = kmalloc(sizeof(struct f2fs_inline_dentry), + GFP_F2FS_ZERO); + if (!backup_dentry) + return -ENOMEM; + + memcpy(backup_dentry, inline_dentry, MAX_INLINE_DATA); + truncate_inline_inode(ipage, 0); + + unlock_page(ipage); + + err = f2fs_add_inline_entries(dir, backup_dentry); + if (err) + goto recover; + + lock_page(ipage); + + stat_dec_inline_dir(dir); + clear_inode_flag(F2FS_I(dir), FI_INLINE_DENTRY); + update_inode(dir, ipage); + kfree(backup_dentry); + return 0; +recover: + lock_page(ipage); + memcpy(inline_dentry, backup_dentry, MAX_INLINE_DATA); + i_size_write(dir, MAX_INLINE_DATA); + update_inode(dir, ipage); + f2fs_put_page(ipage, 1); + + kfree(backup_dentry); + return err; +} + +static int f2fs_convert_inline_dir(struct inode *dir, struct page *ipage, + struct f2fs_inline_dentry *inline_dentry) +{ + if (!F2FS_I(dir)->i_dir_level) + return f2fs_move_inline_dirents(dir, ipage, inline_dentry); + else + return f2fs_move_rehashed_dirents(dir, ipage, inline_dentry); +} + int f2fs_add_inline_entry(struct inode *dir, const struct qstr *name, struct inode *inode, nid_t ino, umode_t mode) { diff --git a/include/linux/f2fs_fs.h b/include/linux/f2fs_fs.h index b90e9bdbd1dd..4c02c6521fef 100644 --- a/include/linux/f2fs_fs.h +++ b/include/linux/f2fs_fs.h @@ -508,4 +508,6 @@ enum { F2FS_FT_MAX }; +#define S_SHIFT 12 + #endif /* _LINUX_F2FS_FS_H */ -- cgit v1.3-8-gc7d7 From 4a6de50d5408cdc699588119e2338e580adc2b73 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Wed, 30 Mar 2016 11:25:31 -0700 Subject: f2fs: use PGP_LOCK to check its truncation Previously, after trylock_page is succeeded, it doesn't check its mapping. In order to fix that, we can just give PGP_LOCK to pagecache_get_page. Signed-off-by: Jaegeuk Kim --- fs/f2fs/node.c | 9 ++------- 1 file changed, 2 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c index 1a33de9d84b1..ade221c9756b 100644 --- a/fs/f2fs/node.c +++ b/fs/f2fs/node.c @@ -1202,13 +1202,10 @@ static void flush_inline_data(struct f2fs_sb_info *sbi, nid_t ino) if (!inode) return; - page = pagecache_get_page(inode->i_mapping, 0, FGP_NOWAIT, 0); + page = pagecache_get_page(inode->i_mapping, 0, FGP_LOCK|FGP_NOWAIT, 0); if (!page) goto iput_out; - if (!trylock_page(page)) - goto release_out; - if (!PageUptodate(page)) goto page_out; @@ -1223,9 +1220,7 @@ static void flush_inline_data(struct f2fs_sb_info *sbi, nid_t ino) else set_page_dirty(page); page_out: - unlock_page(page); -release_out: - f2fs_put_page(page, 0); + f2fs_put_page(page, 1); iput_out: iput(inode); } -- cgit v1.3-8-gc7d7 From ff37355886ac2082cee594aa949c08e2cfb33aa0 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Tue, 29 Mar 2016 16:13:45 -0700 Subject: f2fs: add BUG_ON to avoid unnecessary flow This patch adds BUG_ON instead of retrying loop. In the case of node pages, we already got this inode page, but unlocked it. By the fact that we don't truncate any node pages in operations, the page's mapping should be unchangeable. Signed-off-by: Jaegeuk Kim --- fs/f2fs/node.c | 7 ++----- 1 file changed, 2 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c index ade221c9756b..095fc2c96e16 100644 --- a/fs/f2fs/node.c +++ b/fs/f2fs/node.c @@ -832,7 +832,7 @@ int truncate_inode_blocks(struct inode *inode, pgoff_t from) trace_f2fs_truncate_inode_blocks_enter(inode, from); level = get_node_path(inode, from, offset, noffset); -restart: + page = get_node_page(sbi, inode->i_ino); if (IS_ERR(page)) { trace_f2fs_truncate_inode_blocks_exit(inode, PTR_ERR(page)); @@ -896,10 +896,7 @@ skip_partial: if (offset[1] == 0 && ri->i_nid[offset[0] - NODE_DIR1_BLOCK]) { lock_page(page); - if (unlikely(page->mapping != NODE_MAPPING(sbi))) { - f2fs_put_page(page, 1); - goto restart; - } + BUG_ON(page->mapping != NODE_MAPPING(sbi)); f2fs_wait_on_page_writeback(page, NODE, true); ri->i_nid[offset[0] - NODE_DIR1_BLOCK] = 0; set_page_dirty(page); -- cgit v1.3-8-gc7d7 From de5307e46d28aa26ffd6f2048398b6303e134a67 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Mon, 11 Apr 2016 11:51:51 -0700 Subject: f2fs: fix dropping inmemory pages in a wrong time When one reader closes its file while the other writer is doing atomic writes, f2fs_release_file drops atomic data resulting in an empty commit. This patch fixes this wrong commit problem by checking openess of the file. Process0 Process1 open file start atomic write write data read data close file f2fs_release_file() clear atomic data commit atomic write Reported-by: Miao Xie Signed-off-by: Jaegeuk Kim --- fs/f2fs/file.c | 8 ++++++++ 1 file changed, 8 insertions(+) (limited to 'fs') diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index 90d1157a09f9..e10eb61d9f62 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -1254,6 +1254,14 @@ out: static int f2fs_release_file(struct inode *inode, struct file *filp) { + /* + * f2fs_relase_file is called at every close calls. So we should + * not drop any inmemory pages by close called by other process. + */ + if (!(filp->f_mode & FMODE_WRITE) || + atomic_read(&inode->i_writecount) != 1) + return 0; + /* some remained atomic pages should discarded */ if (f2fs_is_atomic_file(inode)) drop_inmem_pages(inode); -- cgit v1.3-8-gc7d7 From 26dc3d4424e9f4764d633bd8f9309a01e6f364dd Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Mon, 11 Apr 2016 13:15:10 -0700 Subject: f2fs: unset atomic/volatile flag in f2fs_release_file The atomic/volatile operation should be done in pair of start and commit ioctl. For example, if a killed process remains open-ended atomic operation, we should drop its flag as well as its atomic data. Otherwise, if sqlite initiates another operation which doesn't require atomic writes, it will lose every data, since f2fs still treats with them as atomic writes; nobody will trigger its commit. Reported-by: Miao Xie Signed-off-by: Jaegeuk Kim --- fs/f2fs/file.c | 5 ++--- fs/f2fs/segment.c | 2 ++ 2 files changed, 4 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index e10eb61d9f62..ed389f6a37b4 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -1266,6 +1266,7 @@ static int f2fs_release_file(struct inode *inode, struct file *filp) if (f2fs_is_atomic_file(inode)) drop_inmem_pages(inode); if (f2fs_is_volatile_file(inode)) { + clear_inode_flag(F2FS_I(inode), FI_VOLATILE_FILE); set_inode_flag(F2FS_I(inode), FI_DROP_CACHE); filemap_fdatawrite(inode->i_mapping); clear_inode_flag(F2FS_I(inode), FI_DROP_CACHE); @@ -1449,10 +1450,8 @@ static int f2fs_ioc_abort_volatile_write(struct file *filp) if (ret) return ret; - if (f2fs_is_atomic_file(inode)) { - clear_inode_flag(F2FS_I(inode), FI_ATOMIC_FILE); + if (f2fs_is_atomic_file(inode)) drop_inmem_pages(inode); - } if (f2fs_is_volatile_file(inode)) { clear_inode_flag(F2FS_I(inode), FI_VOLATILE_FILE); ret = f2fs_sync_file(filp, 0, LLONG_MAX, 0); diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c index 540669d6978e..299c784f5b61 100644 --- a/fs/f2fs/segment.c +++ b/fs/f2fs/segment.c @@ -239,6 +239,8 @@ void drop_inmem_pages(struct inode *inode) { struct f2fs_inode_info *fi = F2FS_I(inode); + clear_inode_flag(F2FS_I(inode), FI_ATOMIC_FILE); + mutex_lock(&fi->inmem_lock); __revoke_inmem_pages(inode, &fi->inmem_pages, true, false); mutex_unlock(&fi->inmem_lock); -- cgit v1.3-8-gc7d7 From 4da7bf5a4345f3ce9699476a8022f66cfb4a8ce9 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Wed, 6 Apr 2016 11:27:03 -0700 Subject: f2fs: remove redundant condition check This patch resolves the redundant condition check reported by David. Reported-by: David Binderman Signed-off-by: Jaegeuk Kim --- fs/f2fs/checkpoint.c | 2 +- fs/f2fs/data.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c index 0955312e5ca0..b92782f40643 100644 --- a/fs/f2fs/checkpoint.c +++ b/fs/f2fs/checkpoint.c @@ -211,7 +211,7 @@ void ra_meta_pages_cond(struct f2fs_sb_info *sbi, pgoff_t index) bool readahead = false; page = find_get_page(META_MAPPING(sbi), index); - if (!page || (page && !PageUptodate(page))) + if (!page || !PageUptodate(page)) readahead = true; f2fs_put_page(page, 0); diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index 5dafb9cef12e..c29bcf4cfca1 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -1496,7 +1496,7 @@ restart: } else { /* hole case */ err = get_dnode_of_data(&dn, index, LOOKUP_NODE); - if (err || (!err && dn.data_blkaddr == NULL_ADDR)) { + if (err || dn.data_blkaddr == NULL_ADDR) { f2fs_put_dnode(&dn); f2fs_lock_op(sbi); locked = true; -- cgit v1.3-8-gc7d7 From 58457f1c355545c468b8aed5c431d8a6bb71d35d Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Tue, 12 Apr 2016 11:52:30 -0700 Subject: f2fs: give -E2BIG for no space in xattr This patch returns -E2BIG if there is no space to add an xattr entry. This should fix generic/026 in xfstests as well. Signed-off-by: Jaegeuk Kim --- fs/f2fs/xattr.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/f2fs/xattr.c b/fs/f2fs/xattr.c index 06a72dc0191a..152971243ad8 100644 --- a/fs/f2fs/xattr.c +++ b/fs/f2fs/xattr.c @@ -500,7 +500,7 @@ static int __f2fs_setxattr(struct inode *inode, int index, free = free + ENTRY_SIZE(here); if (unlikely(free < newsize)) { - error = -ENOSPC; + error = -E2BIG; goto exit; } } -- cgit v1.3-8-gc7d7 From 63c52d7878903a014fa4c9075afd051b1e77597b Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Tue, 12 Apr 2016 14:11:03 -0700 Subject: f2fs: don't invalidate atomic page if successful If we committed atomic write successfully, we don't need to invalidate pages. Signed-off-by: Jaegeuk Kim --- fs/f2fs/segment.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c index 299c784f5b61..770cdc95120f 100644 --- a/fs/f2fs/segment.c +++ b/fs/f2fs/segment.c @@ -223,9 +223,10 @@ static int __revoke_inmem_pages(struct inode *inode, f2fs_put_dnode(&dn); } next: - ClearPageUptodate(page); + /* we don't need to invalidate this in the sccessful status */ + if (drop || recover) + ClearPageUptodate(page); set_page_private(page, 0); - ClearPageUptodate(page); f2fs_put_page(page, 1); list_del(&cur->list); -- cgit v1.3-8-gc7d7 From c27753d675fccd3b15a78621a2a66616507693d4 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Tue, 12 Apr 2016 14:36:11 -0700 Subject: f2fs: flush dirty pages before starting atomic writes If somebody wrote some data before atomic writes, we should flush them in order to handle atomic data in a right period. Signed-off-by: Jaegeuk Kim --- fs/f2fs/file.c | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index ed389f6a37b4..7de90e60abd1 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -1369,7 +1369,16 @@ static int f2fs_ioc_start_atomic_write(struct file *filp) set_inode_flag(F2FS_I(inode), FI_ATOMIC_FILE); f2fs_update_time(F2FS_I_SB(inode), REQ_TIME); - return 0; + if (!get_dirty_pages(inode)) + return 0; + + f2fs_msg(F2FS_I_SB(inode)->sb, KERN_WARNING, + "Unexpected flush for atomic writes: ino=%lu, npages=%u", + inode->i_ino, get_dirty_pages(inode)); + ret = filemap_write_and_wait_range(inode->i_mapping, 0, LLONG_MAX); + if (ret) + clear_inode_flag(F2FS_I(inode), FI_ATOMIC_FILE); + return ret; } static int f2fs_ioc_commit_atomic_write(struct file *filp) -- cgit v1.3-8-gc7d7 From 9c7fe83530a351845719acf1dda0587e8c743588 Mon Sep 17 00:00:00 2001 From: Daniel DeFreez Date: Tue, 19 Apr 2016 19:57:45 -0400 Subject: GFS2: Add calls to gfs2_holder_uninit in two error handlers This patch fixes two locations that do not call gfs2_holder_uninit if gfs2_glock_nq returns an error. Signed-off-by: Daniel DeFreez Signed-off-by: Bob Peterson --- fs/gfs2/aops.c | 3 ++- fs/gfs2/file.c | 3 ++- 2 files changed, 4 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/aops.c b/fs/gfs2/aops.c index aa016e4b8bec..58dd0061134b 100644 --- a/fs/gfs2/aops.c +++ b/fs/gfs2/aops.c @@ -1063,7 +1063,7 @@ static ssize_t gfs2_direct_IO(struct kiocb *iocb, struct iov_iter *iter, gfs2_holder_init(ip->i_gl, LM_ST_DEFERRED, 0, &gh); rv = gfs2_glock_nq(&gh); if (rv) - return rv; + goto out_uninit; rv = gfs2_ok_for_dio(ip, offset); if (rv != 1) goto out; /* dio not valid, fall back to buffered i/o */ @@ -1102,6 +1102,7 @@ static ssize_t gfs2_direct_IO(struct kiocb *iocb, struct iov_iter *iter, offset, gfs2_get_block_direct, NULL, NULL, 0); out: gfs2_glock_dq(&gh); +out_uninit: gfs2_holder_uninit(&gh); return rv; } diff --git a/fs/gfs2/file.c b/fs/gfs2/file.c index fd9a10e4518d..f33fd92e5f49 100644 --- a/fs/gfs2/file.c +++ b/fs/gfs2/file.c @@ -160,7 +160,7 @@ static int gfs2_get_flags(struct file *filp, u32 __user *ptr) gfs2_holder_init(ip->i_gl, LM_ST_SHARED, 0, &gh); error = gfs2_glock_nq(&gh); if (error) - return error; + goto out_uninit; fsflags = fsflags_cvt(gfs2_to_fsflags, ip->i_diskflags); if (!S_ISDIR(inode->i_mode) && ip->i_diskflags & GFS2_DIF_JDATA) @@ -169,6 +169,7 @@ static int gfs2_get_flags(struct file *filp, u32 __user *ptr) error = -EFAULT; gfs2_glock_dq(&gh); +out_uninit: gfs2_holder_uninit(&gh); return error; } -- cgit v1.3-8-gc7d7 From a26feccaba296bd0ae410eabce79cb3443c8a701 Mon Sep 17 00:00:00 2001 From: Andreas Gruenbacher Date: Thu, 14 Apr 2016 00:30:16 +0200 Subject: ceph: Get rid of d_find_alias in ceph_set_acl Create a variant of ceph_setattr that takes an inode instead of a dentry. Change __ceph_setxattr (and also __ceph_removexattr) to take an inode instead of a dentry. Use those in ceph_set_acl so that we no longer need a dentry there. Signed-off-by: Andreas Gruenbacher Signed-off-by: "Yan, Zheng" Signed-off-by: Al Viro --- fs/ceph/acl.c | 14 +++++--------- fs/ceph/inode.c | 16 ++++++++++------ fs/ceph/super.h | 4 ++-- fs/ceph/xattr.c | 28 ++++++++++++---------------- 4 files changed, 29 insertions(+), 33 deletions(-) (limited to 'fs') diff --git a/fs/ceph/acl.c b/fs/ceph/acl.c index 5457f216e2e5..4f67227f69a5 100644 --- a/fs/ceph/acl.c +++ b/fs/ceph/acl.c @@ -90,7 +90,6 @@ int ceph_set_acl(struct inode *inode, struct posix_acl *acl, int type) char *value = NULL; struct iattr newattrs; umode_t new_mode = inode->i_mode, old_mode = inode->i_mode; - struct dentry *dentry; switch (type) { case ACL_TYPE_ACCESS: @@ -128,29 +127,26 @@ int ceph_set_acl(struct inode *inode, struct posix_acl *acl, int type) goto out_free; } - dentry = d_find_alias(inode); if (new_mode != old_mode) { newattrs.ia_mode = new_mode; newattrs.ia_valid = ATTR_MODE; - ret = ceph_setattr(dentry, &newattrs); + ret = __ceph_setattr(inode, &newattrs); if (ret) - goto out_dput; + goto out_free; } - ret = __ceph_setxattr(dentry, name, value, size, 0); + ret = __ceph_setxattr(inode, name, value, size, 0); if (ret) { if (new_mode != old_mode) { newattrs.ia_mode = old_mode; newattrs.ia_valid = ATTR_MODE; - ceph_setattr(dentry, &newattrs); + __ceph_setattr(inode, &newattrs); } - goto out_dput; + goto out_free; } ceph_set_cached_acl(inode, type, acl); -out_dput: - dput(dentry); out_free: kfree(value); out: diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c index ed58b168904a..cadb6aee7f70 100644 --- a/fs/ceph/inode.c +++ b/fs/ceph/inode.c @@ -1776,16 +1776,12 @@ static const struct inode_operations ceph_symlink_iops = { .removexattr = ceph_removexattr, }; -/* - * setattr - */ -int ceph_setattr(struct dentry *dentry, struct iattr *attr) +int __ceph_setattr(struct inode *inode, struct iattr *attr) { - struct inode *inode = d_inode(dentry); struct ceph_inode_info *ci = ceph_inode(inode); const unsigned int ia_valid = attr->ia_valid; struct ceph_mds_request *req; - struct ceph_mds_client *mdsc = ceph_sb_to_client(dentry->d_sb)->mdsc; + struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; struct ceph_cap_flush *prealloc_cf; int issued; int release = 0, dirtied = 0; @@ -2009,6 +2005,14 @@ out_put: return err; } +/* + * setattr + */ +int ceph_setattr(struct dentry *dentry, struct iattr *attr) +{ + return __ceph_setattr(d_inode(dentry), attr); +} + /* * Verify that we have a lease on the given mask. If not, * do a getattr against an mds. diff --git a/fs/ceph/super.h b/fs/ceph/super.h index beb893bb234f..1e41bc2eb2a8 100644 --- a/fs/ceph/super.h +++ b/fs/ceph/super.h @@ -785,6 +785,7 @@ static inline int ceph_do_getattr(struct inode *inode, int mask, bool force) return __ceph_do_getattr(inode, NULL, mask, force); } extern int ceph_permission(struct inode *inode, int mask); +extern int __ceph_setattr(struct inode *inode, struct iattr *attr); extern int ceph_setattr(struct dentry *dentry, struct iattr *attr); extern int ceph_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat); @@ -792,9 +793,8 @@ extern int ceph_getattr(struct vfsmount *mnt, struct dentry *dentry, /* xattr.c */ extern int ceph_setxattr(struct dentry *, const char *, const void *, size_t, int); -int __ceph_setxattr(struct dentry *, const char *, const void *, size_t, int); +int __ceph_setxattr(struct inode *, const char *, const void *, size_t, int); ssize_t __ceph_getxattr(struct inode *, const char *, void *, size_t); -int __ceph_removexattr(struct dentry *, const char *); extern ssize_t ceph_getxattr(struct dentry *, struct inode *, const char *, void *, size_t); extern ssize_t ceph_listxattr(struct dentry *, char *, size_t); extern int ceph_removexattr(struct dentry *, const char *); diff --git a/fs/ceph/xattr.c b/fs/ceph/xattr.c index c6e917d360f7..248e32e3ae7d 100644 --- a/fs/ceph/xattr.c +++ b/fs/ceph/xattr.c @@ -877,11 +877,10 @@ out: return err; } -static int ceph_sync_setxattr(struct dentry *dentry, const char *name, +static int ceph_sync_setxattr(struct inode *inode, const char *name, const char *value, size_t size, int flags) { - struct ceph_fs_client *fsc = ceph_sb_to_client(dentry->d_sb); - struct inode *inode = d_inode(dentry); + struct ceph_fs_client *fsc = ceph_sb_to_client(inode->i_sb); struct ceph_inode_info *ci = ceph_inode(inode); struct ceph_mds_request *req; struct ceph_mds_client *mdsc = fsc->mdsc; @@ -939,13 +938,12 @@ out: return err; } -int __ceph_setxattr(struct dentry *dentry, const char *name, +int __ceph_setxattr(struct inode *inode, const char *name, const void *value, size_t size, int flags) { - struct inode *inode = d_inode(dentry); struct ceph_vxattr *vxattr; struct ceph_inode_info *ci = ceph_inode(inode); - struct ceph_mds_client *mdsc = ceph_sb_to_client(dentry->d_sb)->mdsc; + struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; struct ceph_cap_flush *prealloc_cf = NULL; int issued; int err; @@ -1056,7 +1054,7 @@ do_sync_unlocked: "during filling trace\n", inode); err = -EBUSY; } else { - err = ceph_sync_setxattr(dentry, name, value, size, flags); + err = ceph_sync_setxattr(inode, name, value, size, flags); } out: ceph_free_cap_flush(prealloc_cf); @@ -1078,14 +1076,13 @@ int ceph_setxattr(struct dentry *dentry, const char *name, if (size == 0) value = ""; /* empty EA, do not remove */ - return __ceph_setxattr(dentry, name, value, size, flags); + return __ceph_setxattr(d_inode(dentry), name, value, size, flags); } -static int ceph_send_removexattr(struct dentry *dentry, const char *name) +static int ceph_send_removexattr(struct inode *inode, const char *name) { - struct ceph_fs_client *fsc = ceph_sb_to_client(dentry->d_sb); + struct ceph_fs_client *fsc = ceph_sb_to_client(inode->i_sb); struct ceph_mds_client *mdsc = fsc->mdsc; - struct inode *inode = d_inode(dentry); struct ceph_mds_request *req; int err; @@ -1106,12 +1103,11 @@ static int ceph_send_removexattr(struct dentry *dentry, const char *name) return err; } -int __ceph_removexattr(struct dentry *dentry, const char *name) +static int __ceph_removexattr(struct inode *inode, const char *name) { - struct inode *inode = d_inode(dentry); struct ceph_vxattr *vxattr; struct ceph_inode_info *ci = ceph_inode(inode); - struct ceph_mds_client *mdsc = ceph_sb_to_client(dentry->d_sb)->mdsc; + struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; struct ceph_cap_flush *prealloc_cf = NULL; int issued; int err; @@ -1192,7 +1188,7 @@ do_sync_unlocked: if (lock_snap_rwsem) up_read(&mdsc->snap_rwsem); ceph_free_cap_flush(prealloc_cf); - err = ceph_send_removexattr(dentry, name); + err = ceph_send_removexattr(inode, name); return err; } @@ -1204,7 +1200,7 @@ int ceph_removexattr(struct dentry *dentry, const char *name) if (!strncmp(name, XATTR_SYSTEM_PREFIX, XATTR_SYSTEM_PREFIX_LEN)) return generic_removexattr(dentry, name); - return __ceph_removexattr(dentry, name); + return __ceph_removexattr(d_inode(dentry), name); } #ifdef CONFIG_SECURITY -- cgit v1.3-8-gc7d7 From 2cdeb1e472cf03dec4dc614623fd2e6bd8e5f271 Mon Sep 17 00:00:00 2001 From: Andreas Gruenbacher Date: Thu, 14 Apr 2016 00:30:17 +0200 Subject: ceph: Switch to generic xattr handlers Add a catch-all xattr handler at the end of ceph_xattr_handlers. Check for valid attribute names there, and remove those checks from __ceph_{get,set,remove}xattr instead. No "system.*" xattrs need to be handled by the catch-all handler anymore. The set xattr handler is called with a NULL value to indicate that the attribute should be removed; __ceph_setxattr already handles that case correctly (ceph_set_acl could already calling __ceph_setxattr with a NULL value). Move the check for snapshots from ceph_{set,remove}xattr into __ceph_{set,remove}xattr. With that, ceph_{get,set,remove}xattr can be replaced with the generic iops. Signed-off-by: Andreas Gruenbacher Signed-off-by: "Yan, Zheng" Signed-off-by: Al Viro --- fs/ceph/dir.c | 7 +++--- fs/ceph/inode.c | 13 ++++++------ fs/ceph/super.h | 4 ---- fs/ceph/xattr.c | 66 +++++++++++++++++++++++---------------------------------- 4 files changed, 38 insertions(+), 52 deletions(-) (limited to 'fs') diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c index fadc243dfb28..2e0afdc1da9e 100644 --- a/fs/ceph/dir.c +++ b/fs/ceph/dir.c @@ -5,6 +5,7 @@ #include #include #include +#include #include "super.h" #include "mds_client.h" @@ -1342,10 +1343,10 @@ const struct inode_operations ceph_dir_iops = { .permission = ceph_permission, .getattr = ceph_getattr, .setattr = ceph_setattr, - .setxattr = ceph_setxattr, - .getxattr = ceph_getxattr, + .setxattr = generic_setxattr, + .getxattr = generic_getxattr, .listxattr = ceph_listxattr, - .removexattr = ceph_removexattr, + .removexattr = generic_removexattr, .get_acl = ceph_get_acl, .set_acl = ceph_set_acl, .mknod = ceph_mknod, diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c index cadb6aee7f70..addf8efb842d 100644 --- a/fs/ceph/inode.c +++ b/fs/ceph/inode.c @@ -8,6 +8,7 @@ #include #include #include +#include #include #include @@ -92,10 +93,10 @@ const struct inode_operations ceph_file_iops = { .permission = ceph_permission, .setattr = ceph_setattr, .getattr = ceph_getattr, - .setxattr = ceph_setxattr, - .getxattr = ceph_getxattr, + .setxattr = generic_setxattr, + .getxattr = generic_getxattr, .listxattr = ceph_listxattr, - .removexattr = ceph_removexattr, + .removexattr = generic_removexattr, .get_acl = ceph_get_acl, .set_acl = ceph_set_acl, }; @@ -1770,10 +1771,10 @@ static const struct inode_operations ceph_symlink_iops = { .get_link = simple_get_link, .setattr = ceph_setattr, .getattr = ceph_getattr, - .setxattr = ceph_setxattr, - .getxattr = ceph_getxattr, + .setxattr = generic_setxattr, + .getxattr = generic_getxattr, .listxattr = ceph_listxattr, - .removexattr = ceph_removexattr, + .removexattr = generic_removexattr, }; int __ceph_setattr(struct inode *inode, struct iattr *attr) diff --git a/fs/ceph/super.h b/fs/ceph/super.h index 1e41bc2eb2a8..7b99eb756477 100644 --- a/fs/ceph/super.h +++ b/fs/ceph/super.h @@ -791,13 +791,9 @@ extern int ceph_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat); /* xattr.c */ -extern int ceph_setxattr(struct dentry *, const char *, const void *, - size_t, int); int __ceph_setxattr(struct inode *, const char *, const void *, size_t, int); ssize_t __ceph_getxattr(struct inode *, const char *, void *, size_t); -extern ssize_t ceph_getxattr(struct dentry *, struct inode *, const char *, void *, size_t); extern ssize_t ceph_listxattr(struct dentry *, char *, size_t); -extern int ceph_removexattr(struct dentry *, const char *); extern void __ceph_build_xattrs_blob(struct ceph_inode_info *ci); extern void __ceph_destroy_xattrs(struct ceph_inode_info *ci); extern void __init ceph_xattr_init(void); diff --git a/fs/ceph/xattr.c b/fs/ceph/xattr.c index 248e32e3ae7d..8966e9d96bdb 100644 --- a/fs/ceph/xattr.c +++ b/fs/ceph/xattr.c @@ -16,6 +16,8 @@ static int __remove_xattr(struct ceph_inode_info *ci, struct ceph_inode_xattr *xattr); +const struct xattr_handler ceph_other_xattr_handler; + /* * List of handlers for synthetic system.* attributes. Other * attributes are handled directly. @@ -25,6 +27,7 @@ const struct xattr_handler *ceph_xattr_handlers[] = { &posix_acl_access_xattr_handler, &posix_acl_default_xattr_handler, #endif + &ceph_other_xattr_handler, NULL, }; @@ -33,7 +36,6 @@ static bool ceph_is_valid_xattr(const char *name) return !strncmp(name, XATTR_CEPH_PREFIX, XATTR_CEPH_PREFIX_LEN) || !strncmp(name, XATTR_SECURITY_PREFIX, XATTR_SECURITY_PREFIX_LEN) || - !strncmp(name, XATTR_SYSTEM_PREFIX, XATTR_SYSTEM_PREFIX_LEN) || !strncmp(name, XATTR_TRUSTED_PREFIX, XATTR_TRUSTED_PREFIX_LEN) || !strncmp(name, XATTR_USER_PREFIX, XATTR_USER_PREFIX_LEN); } @@ -740,9 +742,6 @@ ssize_t __ceph_getxattr(struct inode *inode, const char *name, void *value, int req_mask; int err; - if (!ceph_is_valid_xattr(name)) - return -ENODATA; - /* let's see if a virtual xattr was requested */ vxattr = ceph_match_vxattr(inode, name); if (vxattr) { @@ -804,15 +803,6 @@ out: return err; } -ssize_t ceph_getxattr(struct dentry *dentry, struct inode *inode, - const char *name, void *value, size_t size) -{ - if (!strncmp(name, XATTR_SYSTEM_PREFIX, XATTR_SYSTEM_PREFIX_LEN)) - return generic_getxattr(dentry, inode, name, value, size); - - return __ceph_getxattr(inode, name, value, size); -} - ssize_t ceph_listxattr(struct dentry *dentry, char *names, size_t size) { struct inode *inode = d_inode(dentry); @@ -956,8 +946,8 @@ int __ceph_setxattr(struct inode *inode, const char *name, int required_blob_size; bool lock_snap_rwsem = false; - if (!ceph_is_valid_xattr(name)) - return -EOPNOTSUPP; + if (ceph_snap(inode) != CEPH_NOSNAP) + return -EROFS; vxattr = ceph_match_vxattr(inode, name); if (vxattr && vxattr->readonly) @@ -1064,21 +1054,6 @@ out: return err; } -int ceph_setxattr(struct dentry *dentry, const char *name, - const void *value, size_t size, int flags) -{ - if (ceph_snap(d_inode(dentry)) != CEPH_NOSNAP) - return -EROFS; - - if (!strncmp(name, XATTR_SYSTEM_PREFIX, XATTR_SYSTEM_PREFIX_LEN)) - return generic_setxattr(dentry, name, value, size, flags); - - if (size == 0) - value = ""; /* empty EA, do not remove */ - - return __ceph_setxattr(d_inode(dentry), name, value, size, flags); -} - static int ceph_send_removexattr(struct inode *inode, const char *name) { struct ceph_fs_client *fsc = ceph_sb_to_client(inode->i_sb); @@ -1115,8 +1090,8 @@ static int __ceph_removexattr(struct inode *inode, const char *name) int dirty; bool lock_snap_rwsem = false; - if (!ceph_is_valid_xattr(name)) - return -EOPNOTSUPP; + if (ceph_snap(inode) != CEPH_NOSNAP) + return -EROFS; vxattr = ceph_match_vxattr(inode, name); if (vxattr && vxattr->readonly) @@ -1192,17 +1167,30 @@ do_sync_unlocked: return err; } -int ceph_removexattr(struct dentry *dentry, const char *name) +static int ceph_get_xattr_handler(const struct xattr_handler *handler, + struct dentry *dentry, struct inode *inode, + const char *name, void *value, size_t size) { - if (ceph_snap(d_inode(dentry)) != CEPH_NOSNAP) - return -EROFS; - - if (!strncmp(name, XATTR_SYSTEM_PREFIX, XATTR_SYSTEM_PREFIX_LEN)) - return generic_removexattr(dentry, name); + if (!ceph_is_valid_xattr(name)) + return -EOPNOTSUPP; + return __ceph_getxattr(inode, name, value, size); +} - return __ceph_removexattr(d_inode(dentry), name); +static int ceph_set_xattr_handler(const struct xattr_handler *handler, + struct dentry *dentry, const char *name, + const void *value, size_t size, int flags) +{ + if (!ceph_is_valid_xattr(name)) + return -EOPNOTSUPP; + return __ceph_setxattr(d_inode(dentry), name, value, size, flags); } +const struct xattr_handler ceph_other_xattr_handler = { + .prefix = "", /* match any name => handlers called with full name */ + .get = ceph_get_xattr_handler, + .set = ceph_set_xattr_handler, +}; + #ifdef CONFIG_SECURITY bool ceph_security_xattr_wanted(struct inode *in) { -- cgit v1.3-8-gc7d7 From b971e94e8f4c09ff775cfb2c4f846b4431a00598 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Thu, 14 Apr 2016 00:30:18 +0200 Subject: ceph: kill __ceph_removexattr() when removing a xattr, generic_removexattr() calls __ceph_setxattr() with NULL value and XATTR_REPLACE flag. __ceph_removexattr() is not used any more. Signed-off-by: "Yan, Zheng" Signed-off-by: Al Viro --- fs/ceph/xattr.c | 126 -------------------------------------------------------- 1 file changed, 126 deletions(-) (limited to 'fs') diff --git a/fs/ceph/xattr.c b/fs/ceph/xattr.c index 8966e9d96bdb..0d66722c6a52 100644 --- a/fs/ceph/xattr.c +++ b/fs/ceph/xattr.c @@ -498,19 +498,6 @@ static int __remove_xattr(struct ceph_inode_info *ci, return 0; } -static int __remove_xattr_by_name(struct ceph_inode_info *ci, - const char *name) -{ - struct rb_node **p; - struct ceph_inode_xattr *xattr; - int err; - - p = &ci->i_xattrs.index.rb_node; - xattr = __get_xattr(ci, name); - err = __remove_xattr(ci, xattr); - return err; -} - static char *__copy_xattr_names(struct ceph_inode_info *ci, char *dest) { @@ -1054,119 +1041,6 @@ out: return err; } -static int ceph_send_removexattr(struct inode *inode, const char *name) -{ - struct ceph_fs_client *fsc = ceph_sb_to_client(inode->i_sb); - struct ceph_mds_client *mdsc = fsc->mdsc; - struct ceph_mds_request *req; - int err; - - req = ceph_mdsc_create_request(mdsc, CEPH_MDS_OP_RMXATTR, - USE_AUTH_MDS); - if (IS_ERR(req)) - return PTR_ERR(req); - req->r_path2 = kstrdup(name, GFP_NOFS); - if (!req->r_path2) - return -ENOMEM; - - req->r_inode = inode; - ihold(inode); - req->r_num_caps = 1; - req->r_inode_drop = CEPH_CAP_XATTR_SHARED; - err = ceph_mdsc_do_request(mdsc, NULL, req); - ceph_mdsc_put_request(req); - return err; -} - -static int __ceph_removexattr(struct inode *inode, const char *name) -{ - struct ceph_vxattr *vxattr; - struct ceph_inode_info *ci = ceph_inode(inode); - struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; - struct ceph_cap_flush *prealloc_cf = NULL; - int issued; - int err; - int required_blob_size; - int dirty; - bool lock_snap_rwsem = false; - - if (ceph_snap(inode) != CEPH_NOSNAP) - return -EROFS; - - vxattr = ceph_match_vxattr(inode, name); - if (vxattr && vxattr->readonly) - return -EOPNOTSUPP; - - /* pass any unhandled ceph.* xattrs through to the MDS */ - if (!strncmp(name, XATTR_CEPH_PREFIX, XATTR_CEPH_PREFIX_LEN)) - goto do_sync_unlocked; - - prealloc_cf = ceph_alloc_cap_flush(); - if (!prealloc_cf) - return -ENOMEM; - - err = -ENOMEM; - spin_lock(&ci->i_ceph_lock); -retry: - issued = __ceph_caps_issued(ci, NULL); - if (ci->i_xattrs.version == 0 || !(issued & CEPH_CAP_XATTR_EXCL)) - goto do_sync; - - if (!lock_snap_rwsem && !ci->i_head_snapc) { - lock_snap_rwsem = true; - if (!down_read_trylock(&mdsc->snap_rwsem)) { - spin_unlock(&ci->i_ceph_lock); - down_read(&mdsc->snap_rwsem); - spin_lock(&ci->i_ceph_lock); - goto retry; - } - } - - dout("removexattr %p issued %s\n", inode, ceph_cap_string(issued)); - - __build_xattrs(inode); - - required_blob_size = __get_required_blob_size(ci, 0, 0); - - if (!ci->i_xattrs.prealloc_blob || - required_blob_size > ci->i_xattrs.prealloc_blob->alloc_len) { - struct ceph_buffer *blob; - - spin_unlock(&ci->i_ceph_lock); - dout(" preaallocating new blob size=%d\n", required_blob_size); - blob = ceph_buffer_new(required_blob_size, GFP_NOFS); - if (!blob) - goto do_sync_unlocked; - spin_lock(&ci->i_ceph_lock); - if (ci->i_xattrs.prealloc_blob) - ceph_buffer_put(ci->i_xattrs.prealloc_blob); - ci->i_xattrs.prealloc_blob = blob; - goto retry; - } - - err = __remove_xattr_by_name(ceph_inode(inode), name); - - dirty = __ceph_mark_dirty_caps(ci, CEPH_CAP_XATTR_EXCL, - &prealloc_cf); - ci->i_xattrs.dirty = true; - inode->i_ctime = current_fs_time(inode->i_sb); - spin_unlock(&ci->i_ceph_lock); - if (lock_snap_rwsem) - up_read(&mdsc->snap_rwsem); - if (dirty) - __mark_inode_dirty(inode, dirty); - ceph_free_cap_flush(prealloc_cf); - return err; -do_sync: - spin_unlock(&ci->i_ceph_lock); -do_sync_unlocked: - if (lock_snap_rwsem) - up_read(&mdsc->snap_rwsem); - ceph_free_cap_flush(prealloc_cf); - err = ceph_send_removexattr(inode, name); - return err; -} - static int ceph_get_xattr_handler(const struct xattr_handler *handler, struct dentry *dentry, struct inode *inode, const char *name, void *value, size_t size) -- cgit v1.3-8-gc7d7 From 1f60fbe7274918adb8db2f616e321890730ab7e3 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Sat, 23 Apr 2016 22:50:07 -0400 Subject: ext4: allow readdir()'s of large empty directories to be interrupted If a directory has a large number of empty blocks, iterating over all of them can take a long time, leading to scheduler warnings and users getting irritated when they can't kill a process in the middle of one of these long-running readdir operations. Fix this by adding checks to ext4_readdir() and ext4_htree_fill_tree(). This was reverted earlier due to a typo in the original commit where I experimented with using signal_pending() instead of fatal_signal_pending(). The test was in the wrong place if we were going to return signal_pending() since we would end up returning duplicant entries. See 9f2394c9be47 for a more detailed explanation. Added fix as suggested by Linus to check for signal_pending() in in the filldir() functions. Reported-by: Benjamin LaHaise Google-Bug-Id: 27880676 Signed-off-by: Theodore Ts'o --- fs/compat.c | 4 ++++ fs/ext4/dir.c | 5 +++++ fs/ext4/namei.c | 5 +++++ fs/readdir.c | 4 ++++ 4 files changed, 18 insertions(+) (limited to 'fs') diff --git a/fs/compat.c b/fs/compat.c index a71936a3f4cb..f940cb20562f 100644 --- a/fs/compat.c +++ b/fs/compat.c @@ -936,6 +936,8 @@ static int compat_filldir(struct dir_context *ctx, const char *name, int namlen, } dirent = buf->previous; if (dirent) { + if (signal_pending(current)) + return -EINTR; if (__put_user(offset, &dirent->d_off)) goto efault; } @@ -1020,6 +1022,8 @@ static int compat_filldir64(struct dir_context *ctx, const char *name, dirent = buf->previous; if (dirent) { + if (signal_pending(current)) + return -EINTR; if (__put_user_unaligned(offset, &dirent->d_off)) goto efault; } diff --git a/fs/ext4/dir.c b/fs/ext4/dir.c index 561d7308b393..4173bfe21114 100644 --- a/fs/ext4/dir.c +++ b/fs/ext4/dir.c @@ -150,6 +150,11 @@ static int ext4_readdir(struct file *file, struct dir_context *ctx) while (ctx->pos < inode->i_size) { struct ext4_map_blocks map; + if (fatal_signal_pending(current)) { + err = -ERESTARTSYS; + goto errout; + } + cond_resched(); map.m_lblk = ctx->pos >> EXT4_BLOCK_SIZE_BITS(sb); map.m_len = 1; err = ext4_map_blocks(NULL, inode, &map, 0); diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c index 48e4b8907826..c07422d254b6 100644 --- a/fs/ext4/namei.c +++ b/fs/ext4/namei.c @@ -1107,6 +1107,11 @@ int ext4_htree_fill_tree(struct file *dir_file, __u32 start_hash, } while (1) { + if (fatal_signal_pending(current)) { + err = -ERESTARTSYS; + goto errout; + } + cond_resched(); block = dx_get_block(frame->at); ret = htree_dirblock_to_tree(dir_file, dir, block, &hinfo, start_hash, start_minor_hash); diff --git a/fs/readdir.c b/fs/readdir.c index e69ef3b79787..5f2d4bee5a73 100644 --- a/fs/readdir.c +++ b/fs/readdir.c @@ -169,6 +169,8 @@ static int filldir(struct dir_context *ctx, const char *name, int namlen, } dirent = buf->previous; if (dirent) { + if (signal_pending(current)) + return -EINTR; if (__put_user(offset, &dirent->d_off)) goto efault; } @@ -248,6 +250,8 @@ static int filldir64(struct dir_context *ctx, const char *name, int namlen, return -EINVAL; dirent = buf->previous; if (dirent) { + if (signal_pending(current)) + return -EINTR; if (__put_user(offset, &dirent->d_off)) goto efault; } -- cgit v1.3-8-gc7d7 From 06bd3c36a733ac27962fea7d6f47168841376824 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Sun, 24 Apr 2016 00:56:03 -0400 Subject: ext4: fix data exposure after a crash Huang has reported that in his powerfail testing he is seeing stale block contents in some of recently allocated blocks although he mounts ext4 in data=ordered mode. After some investigation I have found out that indeed when delayed allocation is used, we don't add inode to transaction's list of inodes needing flushing before commit. Originally we were doing that but commit f3b59291a69d removed the logic with a flawed argument that it is not needed. The problem is that although for delayed allocated blocks we write their contents immediately after allocating them, there is no guarantee that the IO scheduler or device doesn't reorder things and thus transaction allocating blocks and attaching them to inode can reach stable storage before actual block contents. Actually whenever we attach freshly allocated blocks to inode using a written extent, we should add inode to transaction's ordered inode list to make sure we properly wait for block contents to be written before committing the transaction. So that is what we do in this patch. This also handles other cases where stale data exposure was possible - like filling hole via mmap in data=ordered,nodelalloc mode. The only exception to the above rule are extending direct IO writes where blkdev_direct_IO() waits for IO to complete before increasing i_size and thus stale data exposure is not possible. For now we don't complicate the code with optimizing this special case since the overhead is pretty low. In case this is observed to be a performance problem we can always handle it using a special flag to ext4_map_blocks(). CC: stable@vger.kernel.org Fixes: f3b59291a69d0b734be1fc8be489fef2dd846d3d Reported-by: "HUANG Weller (CM/ESW12-CN)" Tested-by: "HUANG Weller (CM/ESW12-CN)" Signed-off-by: Jan Kara Signed-off-by: Theodore Ts'o --- fs/ext4/inode.c | 24 +++++++++++++++--------- 1 file changed, 15 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 981a1fc30eaa..250c2df04a92 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -684,6 +684,21 @@ out_sem: ret = check_block_validity(inode, map); if (ret != 0) return ret; + + /* + * Inodes with freshly allocated blocks where contents will be + * visible after transaction commit must be on transaction's + * ordered data list. + */ + if (map->m_flags & EXT4_MAP_NEW && + !(map->m_flags & EXT4_MAP_UNWRITTEN) && + !(flags & EXT4_GET_BLOCKS_ZERO) && + !IS_NOQUOTA(inode) && + ext4_should_order_data(inode)) { + ret = ext4_jbd2_file_inode(handle, inode); + if (ret) + return ret; + } } return retval; } @@ -1289,15 +1304,6 @@ static int ext4_write_end(struct file *file, int i_size_changed = 0; trace_ext4_write_end(inode, pos, len, copied); - if (ext4_test_inode_state(inode, EXT4_STATE_ORDERED_MODE)) { - ret = ext4_jbd2_file_inode(handle, inode); - if (ret) { - unlock_page(page); - put_page(page); - goto errout; - } - } - if (ext4_has_inline_data(inode)) { ret = ext4_write_inline_data_end(inode, pos, len, copied, page); -- cgit v1.3-8-gc7d7 From 3957ef53a5033bd519b19cf375061be1929bdb5f Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Sun, 24 Apr 2016 00:56:05 -0400 Subject: ext4: remove EXT4_STATE_ORDERED_MODE This flag is just duplicating what ext4_should_order_data() tells you and is used in a single place. Furthermore it doesn't reflect changes to inode data journalling flag so it may be possibly misleading. Just remove it. Signed-off-by: Jan Kara Signed-off-by: Theodore Ts'o --- fs/ext4/ext4.h | 1 - fs/ext4/inode.c | 5 +---- 2 files changed, 1 insertion(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index 349afebe21ee..f75e9ebd4ca2 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -1549,7 +1549,6 @@ enum { EXT4_STATE_DIOREAD_LOCK, /* Disable support for dio read nolocking */ EXT4_STATE_MAY_INLINE_DATA, /* may have in-inode data */ - EXT4_STATE_ORDERED_MODE, /* data=ordered mode */ EXT4_STATE_EXT_PRECACHED, /* extents have been precached */ }; diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 250c2df04a92..8ba46ad06aed 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -3540,10 +3540,7 @@ void ext4_set_aops(struct inode *inode) { switch (ext4_inode_journal_mode(inode)) { case EXT4_INODE_ORDERED_DATA_MODE: - ext4_set_inode_state(inode, EXT4_STATE_ORDERED_MODE); - break; case EXT4_INODE_WRITEBACK_DATA_MODE: - ext4_clear_inode_state(inode, EXT4_STATE_ORDERED_MODE); break; case EXT4_INODE_JOURNAL_DATA_MODE: inode->i_mapping->a_ops = &ext4_journalled_aops; @@ -3636,7 +3633,7 @@ static int __ext4_block_zero_page_range(handle_t *handle, } else { err = 0; mark_buffer_dirty(bh); - if (ext4_test_inode_state(inode, EXT4_STATE_ORDERED_MODE)) + if (ext4_should_order_data(inode)) err = ext4_jbd2_file_inode(handle, inode); } -- cgit v1.3-8-gc7d7 From 41617e1a8dec9fe082ba5dec26bacb154eb55482 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Sun, 24 Apr 2016 00:56:07 -0400 Subject: jbd2: add support for avoiding data writes during transaction commits Currently when filesystem needs to make sure data is on permanent storage before committing a transaction it adds inode to transaction's inode list. During transaction commit, jbd2 writes back all dirty buffers that have allocated underlying blocks and waits for the IO to finish. However when doing writeback for delayed allocated data, we allocate blocks and immediately submit the data. Thus asking jbd2 to write dirty pages just unnecessarily adds more work to jbd2 possibly writing back other redirtied blocks. Add support to jbd2 to allow filesystem to ask jbd2 to only wait for outstanding data writes before committing a transaction and thus avoid unnecessary writes. Signed-off-by: Jan Kara Signed-off-by: Theodore Ts'o --- fs/ext4/ext4_jbd2.h | 3 ++- fs/jbd2/commit.c | 4 ++++ fs/jbd2/journal.c | 3 ++- fs/jbd2/transaction.c | 22 ++++++++++++++++++---- fs/ocfs2/journal.h | 2 +- include/linux/jbd2.h | 13 +++++++++++-- 6 files changed, 38 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/ext4/ext4_jbd2.h b/fs/ext4/ext4_jbd2.h index 5f5846211095..f1c940b38b30 100644 --- a/fs/ext4/ext4_jbd2.h +++ b/fs/ext4/ext4_jbd2.h @@ -362,7 +362,8 @@ static inline int ext4_journal_force_commit(journal_t *journal) static inline int ext4_jbd2_file_inode(handle_t *handle, struct inode *inode) { if (ext4_handle_valid(handle)) - return jbd2_journal_file_inode(handle, EXT4_I(inode)->jinode); + return jbd2_journal_inode_add_write(handle, + EXT4_I(inode)->jinode); return 0; } diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c index 2ad98d6e19f4..70078096117d 100644 --- a/fs/jbd2/commit.c +++ b/fs/jbd2/commit.c @@ -219,6 +219,8 @@ static int journal_submit_data_buffers(journal_t *journal, spin_lock(&journal->j_list_lock); list_for_each_entry(jinode, &commit_transaction->t_inode_list, i_list) { + if (!(jinode->i_flags & JI_WRITE_DATA)) + continue; mapping = jinode->i_vfs_inode->i_mapping; jinode->i_flags |= JI_COMMIT_RUNNING; spin_unlock(&journal->j_list_lock); @@ -256,6 +258,8 @@ static int journal_finish_inode_data_buffers(journal_t *journal, /* For locking, see the comment in journal_submit_data_buffers() */ spin_lock(&journal->j_list_lock); list_for_each_entry(jinode, &commit_transaction->t_inode_list, i_list) { + if (!(jinode->i_flags & JI_WAIT_DATA)) + continue; jinode->i_flags |= JI_COMMIT_RUNNING; spin_unlock(&journal->j_list_lock); err = filemap_fdatawait(jinode->i_vfs_inode->i_mapping); diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c index 435f0b26ac20..b31852f76f46 100644 --- a/fs/jbd2/journal.c +++ b/fs/jbd2/journal.c @@ -94,7 +94,8 @@ EXPORT_SYMBOL(jbd2_journal_blocks_per_page); EXPORT_SYMBOL(jbd2_journal_invalidatepage); EXPORT_SYMBOL(jbd2_journal_try_to_free_buffers); EXPORT_SYMBOL(jbd2_journal_force_commit); -EXPORT_SYMBOL(jbd2_journal_file_inode); +EXPORT_SYMBOL(jbd2_journal_inode_add_write); +EXPORT_SYMBOL(jbd2_journal_inode_add_wait); EXPORT_SYMBOL(jbd2_journal_init_jbd_inode); EXPORT_SYMBOL(jbd2_journal_release_jbd_inode); EXPORT_SYMBOL(jbd2_journal_begin_ordered_truncate); diff --git a/fs/jbd2/transaction.c b/fs/jbd2/transaction.c index 67c103867bf8..be56c8ca34c2 100644 --- a/fs/jbd2/transaction.c +++ b/fs/jbd2/transaction.c @@ -2462,7 +2462,8 @@ void jbd2_journal_refile_buffer(journal_t *journal, struct journal_head *jh) /* * File inode in the inode list of the handle's transaction */ -int jbd2_journal_file_inode(handle_t *handle, struct jbd2_inode *jinode) +static int jbd2_journal_file_inode(handle_t *handle, struct jbd2_inode *jinode, + unsigned long flags) { transaction_t *transaction = handle->h_transaction; journal_t *journal; @@ -2487,12 +2488,14 @@ int jbd2_journal_file_inode(handle_t *handle, struct jbd2_inode *jinode) * and if jinode->i_next_transaction == transaction, commit code * will only file the inode where we want it. */ - if (jinode->i_transaction == transaction || - jinode->i_next_transaction == transaction) + if ((jinode->i_transaction == transaction || + jinode->i_next_transaction == transaction) && + (jinode->i_flags & flags) == flags) return 0; spin_lock(&journal->j_list_lock); - + jinode->i_flags |= flags; + /* Is inode already attached where we need it? */ if (jinode->i_transaction == transaction || jinode->i_next_transaction == transaction) goto done; @@ -2523,6 +2526,17 @@ done: return 0; } +int jbd2_journal_inode_add_write(handle_t *handle, struct jbd2_inode *jinode) +{ + return jbd2_journal_file_inode(handle, jinode, + JI_WRITE_DATA | JI_WAIT_DATA); +} + +int jbd2_journal_inode_add_wait(handle_t *handle, struct jbd2_inode *jinode) +{ + return jbd2_journal_file_inode(handle, jinode, JI_WAIT_DATA); +} + /* * File truncate and transaction commit interact with each other in a * non-trivial way. If a transaction writing data block A is diff --git a/fs/ocfs2/journal.h b/fs/ocfs2/journal.h index f4cd3c3e9fb7..497a4171ef61 100644 --- a/fs/ocfs2/journal.h +++ b/fs/ocfs2/journal.h @@ -619,7 +619,7 @@ static inline int ocfs2_calc_tree_trunc_credits(struct super_block *sb, static inline int ocfs2_jbd2_file_inode(handle_t *handle, struct inode *inode) { - return jbd2_journal_file_inode(handle, &OCFS2_I(inode)->ip_jinode); + return jbd2_journal_inode_add_write(handle, &OCFS2_I(inode)->ip_jinode); } static inline int ocfs2_begin_ordered_truncate(struct inode *inode, diff --git a/include/linux/jbd2.h b/include/linux/jbd2.h index fd1083c46c61..39511484ad10 100644 --- a/include/linux/jbd2.h +++ b/include/linux/jbd2.h @@ -403,11 +403,19 @@ static inline void jbd_unlock_bh_journal_head(struct buffer_head *bh) /* Flags in jbd_inode->i_flags */ #define __JI_COMMIT_RUNNING 0 -/* Commit of the inode data in progress. We use this flag to protect us from +#define __JI_WRITE_DATA 1 +#define __JI_WAIT_DATA 2 + +/* + * Commit of the inode data in progress. We use this flag to protect us from * concurrent deletion of inode. We cannot use reference to inode for this * since we cannot afford doing last iput() on behalf of kjournald */ #define JI_COMMIT_RUNNING (1 << __JI_COMMIT_RUNNING) +/* Write allocated dirty buffers in this inode before commit */ +#define JI_WRITE_DATA (1 << __JI_WRITE_DATA) +/* Wait for outstanding data writes for this inode before commit */ +#define JI_WAIT_DATA (1 << __JI_WAIT_DATA) /** * struct jbd_inode is the structure linking inodes in ordered mode @@ -1270,7 +1278,8 @@ extern int jbd2_journal_clear_err (journal_t *); extern int jbd2_journal_bmap(journal_t *, unsigned long, unsigned long long *); extern int jbd2_journal_force_commit(journal_t *); extern int jbd2_journal_force_commit_nested(journal_t *); -extern int jbd2_journal_file_inode(handle_t *handle, struct jbd2_inode *inode); +extern int jbd2_journal_inode_add_write(handle_t *handle, struct jbd2_inode *inode); +extern int jbd2_journal_inode_add_wait(handle_t *handle, struct jbd2_inode *inode); extern int jbd2_journal_begin_ordered_truncate(journal_t *journal, struct jbd2_inode *inode, loff_t new_size); extern void jbd2_journal_init_jbd_inode(struct jbd2_inode *jinode, struct inode *inode); -- cgit v1.3-8-gc7d7 From ee0876bc69ee8d24d524fb2e9e41e3682aaebb11 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Sun, 24 Apr 2016 00:56:08 -0400 Subject: ext4: do not ask jbd2 to write data for delalloc buffers Currently we ask jbd2 to write all dirty allocated buffers before committing a transaction when doing writeback of delay allocated blocks. However this is unnecessary since we move all pages to writeback state before dropping a transaction handle and then submit all the necessary IO. We still need the transaction commit to wait for all the outstanding writeback before flushing disk caches during transaction commit to avoid data exposure issues though. Use the new jbd2 capability and ask it to only wait for outstanding writeback during transaction commit when writing back data in ext4_writepages(). Tested-by: "HUANG Weller (CM/ESW12-CN)" Signed-off-by: Jan Kara Signed-off-by: Theodore Ts'o --- fs/ext4/ext4.h | 3 +++ fs/ext4/ext4_jbd2.h | 12 +++++++++++- fs/ext4/inode.c | 10 +++++++--- fs/ext4/move_extent.c | 2 +- 4 files changed, 22 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index f75e9ebd4ca2..cb00b1119ec9 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -581,6 +581,9 @@ enum { #define EXT4_GET_BLOCKS_ZERO 0x0200 #define EXT4_GET_BLOCKS_CREATE_ZERO (EXT4_GET_BLOCKS_CREATE |\ EXT4_GET_BLOCKS_ZERO) + /* Caller will submit data before dropping transaction handle. This + * allows jbd2 to avoid submitting data before commit. */ +#define EXT4_GET_BLOCKS_IO_SUBMIT 0x0400 /* * The bit position of these flags must not overlap with any of the diff --git a/fs/ext4/ext4_jbd2.h b/fs/ext4/ext4_jbd2.h index f1c940b38b30..09c1ef38cbe6 100644 --- a/fs/ext4/ext4_jbd2.h +++ b/fs/ext4/ext4_jbd2.h @@ -359,7 +359,8 @@ static inline int ext4_journal_force_commit(journal_t *journal) return 0; } -static inline int ext4_jbd2_file_inode(handle_t *handle, struct inode *inode) +static inline int ext4_jbd2_inode_add_write(handle_t *handle, + struct inode *inode) { if (ext4_handle_valid(handle)) return jbd2_journal_inode_add_write(handle, @@ -367,6 +368,15 @@ static inline int ext4_jbd2_file_inode(handle_t *handle, struct inode *inode) return 0; } +static inline int ext4_jbd2_inode_add_wait(handle_t *handle, + struct inode *inode) +{ + if (ext4_handle_valid(handle)) + return jbd2_journal_inode_add_wait(handle, + EXT4_I(inode)->jinode); + return 0; +} + static inline void ext4_update_inode_fsync_trans(handle_t *handle, struct inode *inode, int datasync) diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 8ba46ad06aed..17bfa42ac971 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -695,7 +695,10 @@ out_sem: !(flags & EXT4_GET_BLOCKS_ZERO) && !IS_NOQUOTA(inode) && ext4_should_order_data(inode)) { - ret = ext4_jbd2_file_inode(handle, inode); + if (flags & EXT4_GET_BLOCKS_IO_SUBMIT) + ret = ext4_jbd2_inode_add_wait(handle, inode); + else + ret = ext4_jbd2_inode_add_write(handle, inode); if (ret) return ret; } @@ -2319,7 +2322,8 @@ static int mpage_map_one_extent(handle_t *handle, struct mpage_da_data *mpd) * the data was copied into the page cache. */ get_blocks_flags = EXT4_GET_BLOCKS_CREATE | - EXT4_GET_BLOCKS_METADATA_NOFAIL; + EXT4_GET_BLOCKS_METADATA_NOFAIL | + EXT4_GET_BLOCKS_IO_SUBMIT; dioread_nolock = ext4_should_dioread_nolock(inode); if (dioread_nolock) get_blocks_flags |= EXT4_GET_BLOCKS_IO_CREATE_EXT; @@ -3634,7 +3638,7 @@ static int __ext4_block_zero_page_range(handle_t *handle, err = 0; mark_buffer_dirty(bh); if (ext4_should_order_data(inode)) - err = ext4_jbd2_file_inode(handle, inode); + err = ext4_jbd2_inode_add_write(handle, inode); } unlock: diff --git a/fs/ext4/move_extent.c b/fs/ext4/move_extent.c index 325cef48b39a..a920c5d29fac 100644 --- a/fs/ext4/move_extent.c +++ b/fs/ext4/move_extent.c @@ -400,7 +400,7 @@ data_copy: /* Even in case of data=writeback it is reasonable to pin * inode to transaction, to prevent unexpected data loss */ - *err = ext4_jbd2_file_inode(handle, orig_inode); + *err = ext4_jbd2_inode_add_write(handle, orig_inode); unlock_pages: unlock_page(pagep[0]); -- cgit v1.3-8-gc7d7 From 7b8081912d75df1d910d6969f0a374b66ef242bf Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Mon, 25 Apr 2016 23:13:17 -0400 Subject: ext4: fix jbd2 handle extension in ext4_ext_truncate_extend_restart() The function jbd2_journal_extend() takes as its argument the number of new credits to be added to the handle. We weren't taking into account the currently unused handle credits; worse, we would try to extend the handle by N credits when it had N credits available. In the case where jbd2_journal_extend() fails because the transaction is too large, when jbd2_journal_restart() gets called, the N credits owned by the handle gets returned to the transaction, and the transaction commit is asynchronously requested, and then start_this_handle() will be able to successfully attach the handle to the current transaction since the required credits are now available. This is mostly harmless, but since ext4_ext_truncate_extend_restart() returns EAGAIN, the truncate machinery will once again try to call ext4_ext_truncate_extend_restart(), which will do the above sequence over and over again until the transaction has committed. This was found while I was debugging a lockup in caused by running xfstests generic/074 in the data=journal case. I'm still not sure why we ended up looping forever, which suggests there may still be another bug hiding in the transaction accounting machinery, but this commit prevents us from looping in the first place. Signed-off-by: Theodore Ts'o --- fs/ext4/extents.c | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c index 95bf4679ac54..ba2be53a61d2 100644 --- a/fs/ext4/extents.c +++ b/fs/ext4/extents.c @@ -120,9 +120,14 @@ static int ext4_ext_truncate_extend_restart(handle_t *handle, if (!ext4_handle_valid(handle)) return 0; - if (handle->h_buffer_credits > needed) + if (handle->h_buffer_credits >= needed) return 0; - err = ext4_journal_extend(handle, needed); + /* + * If we need to extend the journal get a few extra blocks + * while we're at it for efficiency's sake. + */ + needed += 3; + err = ext4_journal_extend(handle, needed - handle->h_buffer_credits); if (err <= 0) return err; err = ext4_truncate_restart_trans(handle, inode, needed); -- cgit v1.3-8-gc7d7 From 4c54659269ecb799133758330e7ea2a6fa4c65ca Mon Sep 17 00:00:00 2001 From: Daeho Jeong Date: Mon, 25 Apr 2016 23:21:00 -0400 Subject: ext4: handle unwritten or delalloc buffers before enabling data journaling We already allocate delalloc blocks before changing the inode mode into "per-file data journal" mode to prevent delalloc blocks from remaining not allocated, but another issue concerned with "BH_Unwritten" status still exists. For example, by fallocate(), several buffers' status change into "BH_Unwritten", but these buffers cannot be processed by ext4_alloc_da_blocks(). So, they still remain in unwritten status after per-file data journaling is enabled and they cannot be changed into written status any more and, if they are journaled and eventually checkpointed, these unwritten buffer will cause a kernel panic by the below BUG_ON() function of submit_bh_wbc() when they are submitted during checkpointing. static int submit_bh_wbc(int rw, struct buffer_head *bh,... { ... BUG_ON(buffer_unwritten(bh)); Moreover, when "dioread_nolock" option is enabled, the status of a buffer is changed into "BH_Unwritten" after write_begin() completes and the "BH_Unwritten" status will be cleared after I/O is done. Therefore, if a buffer's status is changed into unwrutten but the buffer's I/O is not submitted and completed, it can cause the same problem after enabling per-file data journaling. You can easily generate this bug by executing the following command. ./kvm-xfstests -C 10000 -m nodelalloc,dioread_nolock generic/269 To resolve these problems and define a boundary between the previous mode and per-file data journaling mode, we need to flush and wait all the I/O of buffers of a file before enabling per-file data journaling of the file. Signed-off-by: Daeho Jeong Signed-off-by: Theodore Ts'o Reviewed-by: Jan Kara --- fs/ext4/inode.c | 31 ++++++++++++++++++++----------- 1 file changed, 20 insertions(+), 11 deletions(-) (limited to 'fs') diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 17bfa42ac971..779ef4c11bc1 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -5452,22 +5452,29 @@ int ext4_change_inode_journal_flag(struct inode *inode, int val) return 0; if (is_journal_aborted(journal)) return -EROFS; - /* We have to allocate physical blocks for delalloc blocks - * before flushing journal. otherwise delalloc blocks can not - * be allocated any more. even more truncate on delalloc blocks - * could trigger BUG by flushing delalloc blocks in journal. - * There is no delalloc block in non-journal data mode. - */ - if (val && test_opt(inode->i_sb, DELALLOC)) { - err = ext4_alloc_da_blocks(inode); - if (err < 0) - return err; - } /* Wait for all existing dio workers */ ext4_inode_block_unlocked_dio(inode); inode_dio_wait(inode); + /* + * Before flushing the journal and switching inode's aops, we have + * to flush all dirty data the inode has. There can be outstanding + * delayed allocations, there can be unwritten extents created by + * fallocate or buffered writes in dioread_nolock mode covered by + * dirty data which can be converted only after flushing the dirty + * data (and journalled aops don't know how to handle these cases). + */ + if (val) { + down_write(&EXT4_I(inode)->i_mmap_sem); + err = filemap_write_and_wait(inode->i_mapping); + if (err < 0) { + up_write(&EXT4_I(inode)->i_mmap_sem); + ext4_inode_resume_unlocked_dio(inode); + return err; + } + } + jbd2_journal_lock_updates(journal); /* @@ -5492,6 +5499,8 @@ int ext4_change_inode_journal_flag(struct inode *inode, int val) ext4_set_aops(inode); jbd2_journal_unlock_updates(journal); + if (val) + up_write(&EXT4_I(inode)->i_mmap_sem); ext4_inode_resume_unlocked_dio(inode); /* Finally we can mark the inode as dirty. */ -- cgit v1.3-8-gc7d7 From c8585c6fcaf2011de54c3592e80a634a2b9e1a7f Mon Sep 17 00:00:00 2001 From: Daeho Jeong Date: Mon, 25 Apr 2016 23:22:35 -0400 Subject: ext4: fix races between changing inode journal mode and ext4_writepages In ext4, there is a race condition between changing inode journal mode and ext4_writepages(). While ext4_writepages() is executed on a non-journalled mode inode, the inode's journal mode could be enabled by ioctl() and then, some pages dirtied after switching the journal mode will be still exposed to ext4_writepages() in non-journaled mode. To resolve this problem, we use fs-wide per-cpu rw semaphore by Jan Kara's suggestion because we don't want to waste ext4_inode_info's space for this extra rare case. Signed-off-by: Daeho Jeong Signed-off-by: Theodore Ts'o Reviewed-by: Jan Kara --- fs/ext4/ext4.h | 4 ++++ fs/ext4/inode.c | 15 ++++++++++++--- fs/ext4/super.c | 4 ++++ kernel/locking/percpu-rwsem.c | 1 + 4 files changed, 21 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index cb00b1119ec9..ba5aecc07fbc 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -33,6 +33,7 @@ #include #include #include +#include #ifdef __KERNEL__ #include #endif @@ -1508,6 +1509,9 @@ struct ext4_sb_info { struct ratelimit_state s_err_ratelimit_state; struct ratelimit_state s_warning_ratelimit_state; struct ratelimit_state s_msg_ratelimit_state; + + /* Barrier between changing inodes' journal flags and writepages ops. */ + struct percpu_rw_semaphore s_journal_flag_rwsem; }; static inline struct ext4_sb_info *EXT4_SB(struct super_block *sb) diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 779ef4c11bc1..4d8ebbe00456 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -2612,11 +2612,14 @@ static int ext4_writepages(struct address_space *mapping, struct blk_plug plug; bool give_up_on_write = false; + percpu_down_read(&sbi->s_journal_flag_rwsem); trace_ext4_writepages(inode, wbc); - if (dax_mapping(mapping)) - return dax_writeback_mapping_range(mapping, inode->i_sb->s_bdev, - wbc); + if (dax_mapping(mapping)) { + ret = dax_writeback_mapping_range(mapping, inode->i_sb->s_bdev, + wbc); + goto out_writepages; + } /* * No pages to write? This is mainly a kludge to avoid starting @@ -2786,6 +2789,7 @@ retry: out_writepages: trace_ext4_writepages_result(inode, wbc, ret, nr_to_write - wbc->nr_to_write); + percpu_up_read(&sbi->s_journal_flag_rwsem); return ret; } @@ -5436,6 +5440,7 @@ int ext4_change_inode_journal_flag(struct inode *inode, int val) journal_t *journal; handle_t *handle; int err; + struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb); /* * We have to be very careful here: changing a data block's @@ -5475,6 +5480,7 @@ int ext4_change_inode_journal_flag(struct inode *inode, int val) } } + percpu_down_write(&sbi->s_journal_flag_rwsem); jbd2_journal_lock_updates(journal); /* @@ -5491,6 +5497,7 @@ int ext4_change_inode_journal_flag(struct inode *inode, int val) err = jbd2_journal_flush(journal); if (err < 0) { jbd2_journal_unlock_updates(journal); + percpu_up_write(&sbi->s_journal_flag_rwsem); ext4_inode_resume_unlocked_dio(inode); return err; } @@ -5499,6 +5506,8 @@ int ext4_change_inode_journal_flag(struct inode *inode, int val) ext4_set_aops(inode); jbd2_journal_unlock_updates(journal); + percpu_up_write(&sbi->s_journal_flag_rwsem); + if (val) up_write(&EXT4_I(inode)->i_mmap_sem); ext4_inode_resume_unlocked_dio(inode); diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 304c712dbe12..20c5d52253b4 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -859,6 +859,7 @@ static void ext4_put_super(struct super_block *sb) percpu_counter_destroy(&sbi->s_freeinodes_counter); percpu_counter_destroy(&sbi->s_dirs_counter); percpu_counter_destroy(&sbi->s_dirtyclusters_counter); + percpu_free_rwsem(&sbi->s_journal_flag_rwsem); brelse(sbi->s_sbh); #ifdef CONFIG_QUOTA for (i = 0; i < EXT4_MAXQUOTAS; i++) @@ -3930,6 +3931,9 @@ no_journal: if (!err) err = percpu_counter_init(&sbi->s_dirtyclusters_counter, 0, GFP_KERNEL); + if (!err) + err = percpu_init_rwsem(&sbi->s_journal_flag_rwsem); + if (err) { ext4_msg(sb, KERN_ERR, "insufficient memory"); goto failed_mount6; diff --git a/kernel/locking/percpu-rwsem.c b/kernel/locking/percpu-rwsem.c index f231e0bb311c..bec0b647f9cc 100644 --- a/kernel/locking/percpu-rwsem.c +++ b/kernel/locking/percpu-rwsem.c @@ -37,6 +37,7 @@ void percpu_free_rwsem(struct percpu_rw_semaphore *brw) free_percpu(brw->fast_read_ctr); brw->fast_read_ctr = NULL; /* catch use after free bugs */ } +EXPORT_SYMBOL_GPL(percpu_free_rwsem); /* * This is the fast-path for down_read/up_read. If it succeeds we rely -- cgit v1.3-8-gc7d7 From eca76e783cf5970db36edfda7e66487d897ea222 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Wed, 13 Apr 2016 16:14:38 -0700 Subject: f2fs: avoid needless lock for node pages when fsyncing a file When fsync is called, sync_node_pages finds a proper direct node pages to flush. But, it locks unrelated direct node pages together unnecessarily. Acked-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/node.c | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c index 095fc2c96e16..cccee5006cdd 100644 --- a/fs/f2fs/node.c +++ b/fs/f2fs/node.c @@ -1272,10 +1272,14 @@ next_step: * we should not skip writing node pages. */ lock_node: - if (ino && ino_of_node(page) == ino) - lock_page(page); - else if (!trylock_page(page)) + if (ino) { + if (ino_of_node(page) == ino) + lock_page(page); + else + continue; + } else if (!trylock_page(page)) { continue; + } if (unlikely(page->mapping != NODE_MAPPING(sbi))) { continue_unlock: -- cgit v1.3-8-gc7d7 From e6e5f5610d585551785ec654b6db9277b19a0664 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Thu, 14 Apr 2016 16:48:52 -0700 Subject: f2fs: avoid writing 0'th page in volatile writes The first page of volatile writes usually contains a sort of header information which will be used for recovery. (e.g., journal header of sqlite) If this is written without other journal data, user needs to handle the stale journal information. Acked-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/data.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index c29bcf4cfca1..e54489b970ae 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -1177,8 +1177,10 @@ write: goto redirty_out; if (f2fs_is_drop_cache(inode)) goto out; - if (f2fs_is_volatile_file(inode) && !wbc->for_reclaim && - available_free_memory(sbi, BASE_CHECK)) + /* we should not write 0'th page having journal header */ + if (f2fs_is_volatile_file(inode) && (!page->index || + (!wbc->for_reclaim && + available_free_memory(sbi, BASE_CHECK)))) goto redirty_out; /* Dentry blocks are controlled by checkpoint */ -- cgit v1.3-8-gc7d7 From 5268137564920843e581304d9bfb06fb9502cf24 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Wed, 13 Apr 2016 16:24:44 -0700 Subject: f2fs: split sync_node_pages with fsync_node_pages This patch splits the existing sync_node_pages into (f)sync_node_pages. The fsync_node_pages is used for f2fs_sync_file only. Acked-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/checkpoint.c | 2 +- fs/f2fs/f2fs.h | 3 +- fs/f2fs/file.c | 2 +- fs/f2fs/gc.c | 2 +- fs/f2fs/node.c | 108 +++++++++++++++++++++++++++++++++++++-------------- 5 files changed, 84 insertions(+), 33 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c index b92782f40643..bf040b51989d 100644 --- a/fs/f2fs/checkpoint.c +++ b/fs/f2fs/checkpoint.c @@ -892,7 +892,7 @@ retry_flush_nodes: if (get_pages(sbi, F2FS_DIRTY_NODES)) { up_write(&sbi->node_write); - err = sync_node_pages(sbi, 0, &wbc); + err = sync_node_pages(sbi, &wbc); if (err) { f2fs_unlock_all(sbi); goto out; diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 3f1551395244..269abe5959e7 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -1784,7 +1784,8 @@ void ra_node_page(struct f2fs_sb_info *, nid_t); struct page *get_node_page(struct f2fs_sb_info *, pgoff_t); struct page *get_node_page_ra(struct page *, int); void sync_inode_page(struct dnode_of_data *); -int sync_node_pages(struct f2fs_sb_info *, nid_t, struct writeback_control *); +int fsync_node_pages(struct f2fs_sb_info *, nid_t, struct writeback_control *); +int sync_node_pages(struct f2fs_sb_info *, struct writeback_control *); bool alloc_nid(struct f2fs_sb_info *, nid_t *); void alloc_nid_done(struct f2fs_sb_info *, nid_t); void alloc_nid_failed(struct f2fs_sb_info *, nid_t); diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index 7de90e60abd1..3d53ee058aae 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -256,7 +256,7 @@ go_write: goto out; } sync_nodes: - sync_node_pages(sbi, ino, &wbc); + fsync_node_pages(sbi, ino, &wbc); /* if cp_error was enabled, we should avoid infinite loop */ if (unlikely(f2fs_cp_error(sbi))) { diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c index b0051a97824c..e82046523186 100644 --- a/fs/f2fs/gc.c +++ b/fs/f2fs/gc.c @@ -841,7 +841,7 @@ static int do_garbage_collect(struct f2fs_sb_info *sbi, .nr_to_write = LONG_MAX, .for_reclaim = 0, }; - sync_node_pages(sbi, 0, &wbc); + sync_node_pages(sbi, &wbc); } else { f2fs_submit_merged_bio(sbi, DATA, WRITE); } diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c index cccee5006cdd..675b7304c02a 100644 --- a/fs/f2fs/node.c +++ b/fs/f2fs/node.c @@ -1222,12 +1222,84 @@ iput_out: iput(inode); } -int sync_node_pages(struct f2fs_sb_info *sbi, nid_t ino, +int fsync_node_pages(struct f2fs_sb_info *sbi, nid_t ino, struct writeback_control *wbc) { pgoff_t index, end; struct pagevec pvec; - int step = ino ? 2 : 0; + int nwritten = 0; + + pagevec_init(&pvec, 0); + index = 0; + end = ULONG_MAX; + + while (index <= end) { + int i, nr_pages; + nr_pages = pagevec_lookup_tag(&pvec, NODE_MAPPING(sbi), &index, + PAGECACHE_TAG_DIRTY, + min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1); + if (nr_pages == 0) + break; + + for (i = 0; i < nr_pages; i++) { + struct page *page = pvec.pages[i]; + + if (unlikely(f2fs_cp_error(sbi))) { + pagevec_release(&pvec); + return -EIO; + } + + if (!IS_DNODE(page) || !is_cold_node(page)) + continue; + if (ino_of_node(page) != ino) + continue; + + lock_page(page); + + if (unlikely(page->mapping != NODE_MAPPING(sbi))) { +continue_unlock: + unlock_page(page); + continue; + } + if (ino_of_node(page) != ino) + goto continue_unlock; + + if (!PageDirty(page)) { + /* someone wrote it for us */ + goto continue_unlock; + } + + f2fs_wait_on_page_writeback(page, NODE, true); + BUG_ON(PageWriteback(page)); + if (!clear_page_dirty_for_io(page)) + goto continue_unlock; + + set_fsync_mark(page, 1); + if (IS_INODE(page)) + set_dentry_mark(page, + need_dentry_mark(sbi, ino)); + nwritten++; + + if (NODE_MAPPING(sbi)->a_ops->writepage(page, wbc)) + unlock_page(page); + + if (--wbc->nr_to_write == 0) + break; + } + pagevec_release(&pvec); + cond_resched(); + + if (wbc->nr_to_write == 0) + break; + } + return nwritten; +} + +int sync_node_pages(struct f2fs_sb_info *sbi, struct writeback_control *wbc) +{ + pgoff_t index, end; + struct pagevec pvec; + int step = 0; int nwritten = 0; pagevec_init(&pvec, 0); @@ -1266,28 +1338,15 @@ next_step: if (step == 2 && (!IS_DNODE(page) || !is_cold_node(page))) continue; - - /* - * If an fsync mode, - * we should not skip writing node pages. - */ lock_node: - if (ino) { - if (ino_of_node(page) == ino) - lock_page(page); - else - continue; - } else if (!trylock_page(page)) { + if (!trylock_page(page)) continue; - } if (unlikely(page->mapping != NODE_MAPPING(sbi))) { continue_unlock: unlock_page(page); continue; } - if (ino && ino_of_node(page) != ino) - goto continue_unlock; if (!PageDirty(page)) { /* someone wrote it for us */ @@ -1295,7 +1354,7 @@ continue_unlock: } /* flush inline_data */ - if (!ino && is_inline_node(page)) { + if (is_inline_node(page)) { clear_inline_node(page); unlock_page(page); flush_inline_data(sbi, ino_of_node(page)); @@ -1308,17 +1367,8 @@ continue_unlock: if (!clear_page_dirty_for_io(page)) goto continue_unlock; - /* called by fsync() */ - if (ino && IS_DNODE(page)) { - set_fsync_mark(page, 1); - if (IS_INODE(page)) - set_dentry_mark(page, - need_dentry_mark(sbi, ino)); - nwritten++; - } else { - set_fsync_mark(page, 0); - set_dentry_mark(page, 0); - } + set_fsync_mark(page, 0); + set_dentry_mark(page, 0); if (NODE_MAPPING(sbi)->a_ops->writepage(page, wbc)) unlock_page(page); @@ -1466,7 +1516,7 @@ static int f2fs_write_node_pages(struct address_space *mapping, diff = nr_pages_to_write(sbi, NODE, wbc); wbc->sync_mode = WB_SYNC_NONE; - sync_node_pages(sbi, 0, wbc); + sync_node_pages(sbi, wbc); wbc->nr_to_write = max((long)0, wbc->nr_to_write - diff); return 0; -- cgit v1.3-8-gc7d7 From c267ec1526da2d3b12c80a89ebd8eb9b6a01d636 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Fri, 15 Apr 2016 09:25:04 -0700 Subject: f2fs: report unwritten status in fsync_node_pages The fsync_node_pages should return pass or failure so that user could know fsync is completed or not. Acked-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/file.c | 4 +++- fs/f2fs/node.c | 13 ++++++------- 2 files changed, 9 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index 3d53ee058aae..60fd64c59cce 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -256,7 +256,9 @@ go_write: goto out; } sync_nodes: - fsync_node_pages(sbi, ino, &wbc); + ret = fsync_node_pages(sbi, ino, &wbc); + if (ret) + goto out; /* if cp_error was enabled, we should avoid infinite loop */ if (unlikely(f2fs_cp_error(sbi))) { diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c index 675b7304c02a..8a1e21144ecb 100644 --- a/fs/f2fs/node.c +++ b/fs/f2fs/node.c @@ -1227,7 +1227,7 @@ int fsync_node_pages(struct f2fs_sb_info *sbi, nid_t ino, { pgoff_t index, end; struct pagevec pvec; - int nwritten = 0; + int ret = 0; pagevec_init(&pvec, 0); index = 0; @@ -1278,21 +1278,20 @@ continue_unlock: if (IS_INODE(page)) set_dentry_mark(page, need_dentry_mark(sbi, ino)); - nwritten++; - if (NODE_MAPPING(sbi)->a_ops->writepage(page, wbc)) + ret = NODE_MAPPING(sbi)->a_ops->writepage(page, wbc); + if (ret) { unlock_page(page); - - if (--wbc->nr_to_write == 0) break; + } } pagevec_release(&pvec); cond_resched(); - if (wbc->nr_to_write == 0) + if (ret) break; } - return nwritten; + return ret ? -EIO: 0; } int sync_node_pages(struct f2fs_sb_info *sbi, struct writeback_control *wbc) -- cgit v1.3-8-gc7d7 From 608514deba38c8611ad330d6a3c8e2b9a1f68e4b Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Fri, 15 Apr 2016 09:43:17 -0700 Subject: f2fs: set fsync mark only for the last dnode In order to give atomic writes, we should consider power failure during sync_node_pages in fsync. So, this patch marks fsync flag only in the last dnode block. Acked-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/f2fs.h | 4 +- fs/f2fs/file.c | 14 +++++-- fs/f2fs/node.c | 106 +++++++++++++++++++++++++++++++++++++++++++++++++---- fs/f2fs/recovery.c | 9 ++--- 4 files changed, 113 insertions(+), 20 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 269abe5959e7..ca828b0e7d6d 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -159,7 +159,6 @@ struct fsync_inode_entry { struct inode *inode; /* vfs inode pointer */ block_t blkaddr; /* block address locating the last fsync */ block_t last_dentry; /* block address locating the last dentry */ - block_t last_inode; /* block address locating the last inode */ }; #define nats_in_cursum(jnl) (le16_to_cpu(jnl->n_nats)) @@ -1784,7 +1783,8 @@ void ra_node_page(struct f2fs_sb_info *, nid_t); struct page *get_node_page(struct f2fs_sb_info *, pgoff_t); struct page *get_node_page_ra(struct page *, int); void sync_inode_page(struct dnode_of_data *); -int fsync_node_pages(struct f2fs_sb_info *, nid_t, struct writeback_control *); +int fsync_node_pages(struct f2fs_sb_info *, nid_t, struct writeback_control *, + bool); int sync_node_pages(struct f2fs_sb_info *, struct writeback_control *); bool alloc_nid(struct f2fs_sb_info *, nid_t *); void alloc_nid_done(struct f2fs_sb_info *, nid_t); diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index 60fd64c59cce..dc47d5c7b882 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -182,7 +182,8 @@ static void try_to_fix_pino(struct inode *inode) } } -int f2fs_sync_file(struct file *file, loff_t start, loff_t end, int datasync) +static int f2fs_do_sync_file(struct file *file, loff_t start, loff_t end, + int datasync, bool atomic) { struct inode *inode = file->f_mapping->host; struct f2fs_inode_info *fi = F2FS_I(inode); @@ -256,7 +257,7 @@ go_write: goto out; } sync_nodes: - ret = fsync_node_pages(sbi, ino, &wbc); + ret = fsync_node_pages(sbi, ino, &wbc, atomic); if (ret) goto out; @@ -290,6 +291,11 @@ out: return ret; } +int f2fs_sync_file(struct file *file, loff_t start, loff_t end, int datasync) +{ + return f2fs_do_sync_file(file, start, end, datasync, false); +} + static pgoff_t __get_first_dirty_index(struct address_space *mapping, pgoff_t pgofs, int whence) { @@ -1407,7 +1413,7 @@ static int f2fs_ioc_commit_atomic_write(struct file *filp) } } - ret = f2fs_sync_file(filp, 0, LLONG_MAX, 0); + ret = f2fs_do_sync_file(filp, 0, LLONG_MAX, 0, true); err_out: mnt_drop_write_file(filp); return ret; @@ -1465,7 +1471,7 @@ static int f2fs_ioc_abort_volatile_write(struct file *filp) drop_inmem_pages(inode); if (f2fs_is_volatile_file(inode)) { clear_inode_flag(F2FS_I(inode), FI_VOLATILE_FILE); - ret = f2fs_sync_file(filp, 0, LLONG_MAX, 0); + ret = f2fs_do_sync_file(filp, 0, LLONG_MAX, 0, true); } mnt_drop_write_file(filp); diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c index 8a1e21144ecb..de070a524fd2 100644 --- a/fs/f2fs/node.c +++ b/fs/f2fs/node.c @@ -1222,13 +1222,81 @@ iput_out: iput(inode); } +static struct page *last_fsync_dnode(struct f2fs_sb_info *sbi, nid_t ino) +{ + pgoff_t index, end; + struct pagevec pvec; + struct page *last_page = NULL; + + pagevec_init(&pvec, 0); + index = 0; + end = ULONG_MAX; + + while (index <= end) { + int i, nr_pages; + nr_pages = pagevec_lookup_tag(&pvec, NODE_MAPPING(sbi), &index, + PAGECACHE_TAG_DIRTY, + min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1); + if (nr_pages == 0) + break; + + for (i = 0; i < nr_pages; i++) { + struct page *page = pvec.pages[i]; + + if (unlikely(f2fs_cp_error(sbi))) { + f2fs_put_page(last_page, 0); + pagevec_release(&pvec); + return ERR_PTR(-EIO); + } + + if (!IS_DNODE(page) || !is_cold_node(page)) + continue; + if (ino_of_node(page) != ino) + continue; + + lock_page(page); + + if (unlikely(page->mapping != NODE_MAPPING(sbi))) { +continue_unlock: + unlock_page(page); + continue; + } + if (ino_of_node(page) != ino) + goto continue_unlock; + + if (!PageDirty(page)) { + /* someone wrote it for us */ + goto continue_unlock; + } + + if (last_page) + f2fs_put_page(last_page, 0); + + get_page(page); + last_page = page; + unlock_page(page); + } + pagevec_release(&pvec); + cond_resched(); + } + return last_page; +} + int fsync_node_pages(struct f2fs_sb_info *sbi, nid_t ino, - struct writeback_control *wbc) + struct writeback_control *wbc, bool atomic) { pgoff_t index, end; struct pagevec pvec; int ret = 0; + struct page *last_page = NULL; + bool marked = false; + if (atomic) { + last_page = last_fsync_dnode(sbi, ino); + if (IS_ERR_OR_NULL(last_page)) + return PTR_ERR_OR_ZERO(last_page); + } +retry: pagevec_init(&pvec, 0); index = 0; end = ULONG_MAX; @@ -1245,6 +1313,7 @@ int fsync_node_pages(struct f2fs_sb_info *sbi, nid_t ino, struct page *page = pvec.pages[i]; if (unlikely(f2fs_cp_error(sbi))) { + f2fs_put_page(last_page, 0); pagevec_release(&pvec); return -EIO; } @@ -1264,33 +1333,54 @@ continue_unlock: if (ino_of_node(page) != ino) goto continue_unlock; - if (!PageDirty(page)) { + if (!PageDirty(page) && page != last_page) { /* someone wrote it for us */ goto continue_unlock; } f2fs_wait_on_page_writeback(page, NODE, true); BUG_ON(PageWriteback(page)); - if (!clear_page_dirty_for_io(page)) - goto continue_unlock; - set_fsync_mark(page, 1); - if (IS_INODE(page)) - set_dentry_mark(page, + if (!atomic || page == last_page) { + set_fsync_mark(page, 1); + if (IS_INODE(page)) + set_dentry_mark(page, need_dentry_mark(sbi, ino)); + /* may be written by other thread */ + if (!PageDirty(page)) + set_page_dirty(page); + } + + if (!clear_page_dirty_for_io(page)) + goto continue_unlock; ret = NODE_MAPPING(sbi)->a_ops->writepage(page, wbc); if (ret) { unlock_page(page); + f2fs_put_page(last_page, 0); + break; + } + if (page == last_page) { + f2fs_put_page(page, 0); + marked = true; break; } } pagevec_release(&pvec); cond_resched(); - if (ret) + if (ret || marked) break; } + if (!ret && atomic && !marked) { + f2fs_msg(sbi->sb, KERN_DEBUG, + "Retry to write fsync mark: ino=%u, idx=%lx", + ino, last_page->index); + lock_page(last_page); + set_page_dirty(last_page); + unlock_page(last_page); + goto retry; + } return ret ? -EIO: 0; } diff --git a/fs/f2fs/recovery.c b/fs/f2fs/recovery.c index 2c87c12b6f1c..a646d3ba3b25 100644 --- a/fs/f2fs/recovery.c +++ b/fs/f2fs/recovery.c @@ -257,11 +257,8 @@ static int find_fsync_dnodes(struct f2fs_sb_info *sbi, struct list_head *head) } entry->blkaddr = blkaddr; - if (IS_INODE(page)) { - entry->last_inode = blkaddr; - if (is_dent_dnode(page)) - entry->last_dentry = blkaddr; - } + if (IS_INODE(page) && is_dent_dnode(page)) + entry->last_dentry = blkaddr; next: /* check next segment */ blkaddr = next_blkaddr_of_node(page); @@ -521,7 +518,7 @@ static int recover_data(struct f2fs_sb_info *sbi, struct list_head *head) * In this case, we can lose the latest inode(x). * So, call recover_inode for the inode update. */ - if (entry->last_inode == blkaddr) + if (IS_INODE(page)) recover_inode(entry->inode, page); if (entry->last_dentry == blkaddr) { err = recover_dentry(entry->inode, page); -- cgit v1.3-8-gc7d7 From 6bfc49197eba070b799ab4ca8755d3a9fd1700da Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Mon, 18 Apr 2016 17:07:44 -0400 Subject: f2fs: issue cache flush on direct IO Under direct IO path with O_(D)SYNC, it needs to set proper APPEND or UPDATE flags, so taht f2fs_sync_file can make its data safe. Acked-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/data.c | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index e54489b970ae..38ce5d6f8583 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -673,6 +673,9 @@ next_block: err = reserve_new_block(&dn); } else { err = __allocate_data_block(&dn); + if (!err) + set_inode_flag(F2FS_I(inode), + FI_APPEND_WRITE); } if (err) goto sync_out; @@ -1685,8 +1688,12 @@ static ssize_t f2fs_direct_IO(struct kiocb *iocb, struct iov_iter *iter, trace_f2fs_direct_IO_enter(inode, offset, count, iov_iter_rw(iter)); err = blockdev_direct_IO(iocb, inode, iter, offset, get_data_block_dio); - if (err < 0 && iov_iter_rw(iter) == WRITE) - f2fs_write_failed(mapping, offset + count); + if (iov_iter_rw(iter) == WRITE) { + if (err > 0) + set_inode_flag(F2FS_I(inode), FI_UPDATE_WRITE); + else if (err < 0) + f2fs_write_failed(mapping, offset + count); + } trace_f2fs_direct_IO_exit(inode, offset, count, iov_iter_rw(iter), err); -- cgit v1.3-8-gc7d7 From 8d2ae1cbe8a984d7a755755fb53955de2f60a2f9 Mon Sep 17 00:00:00 2001 From: Jakub Wilk Date: Wed, 27 Apr 2016 01:11:21 -0400 Subject: ext4: remove trailing \n from ext4_warning/ext4_error calls Messages passed to ext4_warning() or ext4_error() don't need trailing newlines, because these function add the newlines themselves. Signed-off-by: Jakub Wilk --- fs/ext4/extents.c | 4 ++-- fs/ext4/extents_status.c | 2 +- fs/ext4/file.c | 2 +- fs/ext4/inline.c | 2 +- fs/ext4/mballoc.c | 2 +- fs/ext4/mmp.c | 4 ++-- fs/ext4/namei.c | 2 +- fs/ext4/resize.c | 2 +- 8 files changed, 10 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c index ba2be53a61d2..c53d5a8d2a79 100644 --- a/fs/ext4/extents.c +++ b/fs/ext4/extents.c @@ -2588,7 +2588,7 @@ static int ext4_remove_blocks(handle_t *handle, struct inode *inode, } } else ext4_error(sbi->s_sb, "strange request: removal(2) " - "%u-%u from %u:%u\n", + "%u-%u from %u:%u", from, to, le32_to_cpu(ex->ee_block), ee_len); return 0; } @@ -3743,7 +3743,7 @@ static int ext4_convert_unwritten_extents_endio(handle_t *handle, if (ee_block != map->m_lblk || ee_len > map->m_len) { #ifdef EXT4_DEBUG ext4_warning("Inode (%ld) finished: extent logical block %llu," - " len %u; IO logical block %llu, len %u\n", + " len %u; IO logical block %llu, len %u", inode->i_ino, (unsigned long long)ee_block, ee_len, (unsigned long long)map->m_lblk, map->m_len); #endif diff --git a/fs/ext4/extents_status.c b/fs/ext4/extents_status.c index e38b987ac7f5..37e059202cd2 100644 --- a/fs/ext4/extents_status.c +++ b/fs/ext4/extents_status.c @@ -707,7 +707,7 @@ int ext4_es_insert_extent(struct inode *inode, ext4_lblk_t lblk, (status & EXTENT_STATUS_WRITTEN)) { ext4_warning(inode->i_sb, "Inserting extent [%u/%u] as " " delayed and written which can potentially " - " cause data loss.\n", lblk, len); + " cause data loss.", lblk, len); WARN_ON(1); } diff --git a/fs/ext4/file.c b/fs/ext4/file.c index fa2208bae2e1..3e850b988923 100644 --- a/fs/ext4/file.c +++ b/fs/ext4/file.c @@ -378,7 +378,7 @@ static int ext4_file_open(struct inode * inode, struct file * filp) if (ext4_encrypted_inode(d_inode(dir)) && !ext4_is_child_context_consistent_with_parent(d_inode(dir), inode)) { ext4_warning(inode->i_sb, - "Inconsistent encryption contexts: %lu/%lu\n", + "Inconsistent encryption contexts: %lu/%lu", (unsigned long) d_inode(dir)->i_ino, (unsigned long) inode->i_ino); dput(dir); diff --git a/fs/ext4/inline.c b/fs/ext4/inline.c index 7bc6c855cc18..ff7538c26992 100644 --- a/fs/ext4/inline.c +++ b/fs/ext4/inline.c @@ -1780,7 +1780,7 @@ int empty_inline_dir(struct inode *dir, int *has_inline_data) ext4_warning(dir->i_sb, "bad inline directory (dir #%lu) - " "inode %u, rec_len %u, name_len %d" - "inline size %d\n", + "inline size %d", dir->i_ino, le32_to_cpu(de->inode), le16_to_cpu(de->rec_len), de->name_len, inline_size); diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c index eeeade76012e..efa111a7606d 100644 --- a/fs/ext4/mballoc.c +++ b/fs/ext4/mballoc.c @@ -4935,7 +4935,7 @@ int ext4_group_add_blocks(handle_t *handle, struct super_block *sb, * boundary. */ if (bit + count > EXT4_BLOCKS_PER_GROUP(sb)) { - ext4_warning(sb, "too much blocks added to group %u\n", + ext4_warning(sb, "too much blocks added to group %u", block_group); err = -EINVAL; goto error_return; diff --git a/fs/ext4/mmp.c b/fs/ext4/mmp.c index 24445275d330..23d436d6f8b8 100644 --- a/fs/ext4/mmp.c +++ b/fs/ext4/mmp.c @@ -121,7 +121,7 @@ void __dump_mmp_msg(struct super_block *sb, struct mmp_struct *mmp, __ext4_warning(sb, function, line, "%s", msg); __ext4_warning(sb, function, line, "MMP failure info: last update time: %llu, last update " - "node: %s, last update device: %s\n", + "node: %s, last update device: %s", (long long unsigned int) le64_to_cpu(mmp->mmp_time), mmp->mmp_nodename, mmp->mmp_bdevname); } @@ -353,7 +353,7 @@ skip: * wait for MMP interval and check mmp_seq. */ if (schedule_timeout_interruptible(HZ * wait_time) != 0) { - ext4_warning(sb, "MMP startup interrupted, failing mount\n"); + ext4_warning(sb, "MMP startup interrupted, failing mount"); goto failed; } diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c index c07422d254b6..6e6b3230ee45 100644 --- a/fs/ext4/namei.c +++ b/fs/ext4/namei.c @@ -1618,7 +1618,7 @@ static struct dentry *ext4_lookup(struct inode *dir, struct dentry *dentry, unsi if (nokey) return ERR_PTR(-ENOKEY); ext4_warning(inode->i_sb, - "Inconsistent encryption contexts: %lu/%lu\n", + "Inconsistent encryption contexts: %lu/%lu", (unsigned long) dir->i_ino, (unsigned long) inode->i_ino); return ERR_PTR(-EPERM); diff --git a/fs/ext4/resize.c b/fs/ext4/resize.c index 34038e3598d5..cf681004b196 100644 --- a/fs/ext4/resize.c +++ b/fs/ext4/resize.c @@ -41,7 +41,7 @@ int ext4_resize_begin(struct super_block *sb) */ if (EXT4_SB(sb)->s_mount_state & EXT4_ERROR_FS) { ext4_warning(sb, "There are errors in the filesystem, " - "so online resizing is not allowed\n"); + "so online resizing is not allowed"); return -EPERM; } -- cgit v1.3-8-gc7d7 From a4a13f582c6d36b78b1c0459ee0b28f17bb2fb06 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Wed, 27 Apr 2016 22:22:20 +0800 Subject: f2fs: be aware of invalid filename length The filename length in dirent of may become zero-sized after random junk data injection, once encounter such dirent, find_target_dentry or f2fs_add_inline_entries will run into an infinite loop. So let f2fs being aware of that to avoid deadloop. Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/dir.c | 14 +++++--------- fs/f2fs/inline.c | 14 ++++++-------- 2 files changed, 11 insertions(+), 17 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/dir.c b/fs/f2fs/dir.c index e90380d82214..3b1c14e4eeea 100644 --- a/fs/f2fs/dir.c +++ b/fs/f2fs/dir.c @@ -101,11 +101,6 @@ static struct f2fs_dir_entry *find_in_block(struct page *dentry_page, else kunmap(dentry_page); - /* - * For the most part, it should be a bug when name_len is zero. - * We stop here for figuring out where the bugs has occurred. - */ - f2fs_bug_on(F2FS_P_SB(dentry_page), d.max < 0); return de; } @@ -130,6 +125,11 @@ struct f2fs_dir_entry *find_target_dentry(struct fscrypt_name *fname, de = &d->dentry[bit_pos]; + if (unlikely(!de->name_len)) { + bit_pos++; + continue; + } + /* encrypted case */ de_name.name = d->filename[bit_pos]; de_name.len = le16_to_cpu(de->name_len); @@ -147,10 +147,6 @@ struct f2fs_dir_entry *find_target_dentry(struct fscrypt_name *fname, *max_slots = max_len; max_len = 0; - /* remain bug on condition */ - if (unlikely(!de->name_len)) - d->max = -1; - bit_pos += GET_DENTRY_SLOTS(le16_to_cpu(de->name_len)); } diff --git a/fs/f2fs/inline.c b/fs/f2fs/inline.c index 772056587eb9..e61084cac5f1 100644 --- a/fs/f2fs/inline.c +++ b/fs/f2fs/inline.c @@ -303,11 +303,6 @@ struct f2fs_dir_entry *find_in_inline_dir(struct inode *dir, else f2fs_put_page(ipage, 0); - /* - * For the most part, it should be a bug when name_len is zero. - * We stop here for figuring out where the bugs has occurred. - */ - f2fs_bug_on(sbi, d.max < 0); return de; } @@ -437,6 +432,12 @@ static int f2fs_add_inline_entries(struct inode *dir, } de = &d.dentry[bit_pos]; + + if (unlikely(!de->name_len)) { + bit_pos++; + continue; + } + new_name.name = d.filename[bit_pos]; new_name.len = de->name_len; @@ -448,9 +449,6 @@ static int f2fs_add_inline_entries(struct inode *dir, if (err) goto punch_dentry_pages; - if (unlikely(!de->name_len)) - d.max = -1; - bit_pos += GET_DENTRY_SLOTS(le16_to_cpu(de->name_len)); } return 0; -- cgit v1.3-8-gc7d7 From da011cc0da8cf4a60ddf4d2ae8b42902a3d71e5f Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Wed, 27 Apr 2016 21:40:15 +0800 Subject: f2fs: move node pages only in victim section during GC For foreground GC, we cache node blocks in victim section and set them dirty, then we call sync_node_pages to flush these node pages, but meanwhile, those node pages which does not locate in victim section will be flushed together, so more bandwidth and continuous free space would be occupied. So for this condition, it's better to leave those unrelated node page in cache for further write hit, and let CP or VM to flush them afterward. Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/f2fs.h | 1 + fs/f2fs/gc.c | 25 ++++--------------------- fs/f2fs/node.c | 31 +++++++++++++++++++++++++++++++ 3 files changed, 36 insertions(+), 21 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index ca828b0e7d6d..6a482411e6d4 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -1783,6 +1783,7 @@ void ra_node_page(struct f2fs_sb_info *, nid_t); struct page *get_node_page(struct f2fs_sb_info *, pgoff_t); struct page *get_node_page_ra(struct page *, int); void sync_inode_page(struct dnode_of_data *); +void move_node_page(struct page *, int); int fsync_node_pages(struct f2fs_sb_info *, nid_t, struct writeback_control *, bool); int sync_node_pages(struct f2fs_sb_info *, struct writeback_control *); diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c index e82046523186..19e9fafc5c70 100644 --- a/fs/f2fs/gc.c +++ b/fs/f2fs/gc.c @@ -465,15 +465,7 @@ next_step: continue; } - /* set page dirty and write it */ - if (gc_type == FG_GC) { - f2fs_wait_on_page_writeback(node_page, NODE, true); - set_page_dirty(node_page); - } else { - if (!PageWriteback(node_page)) - set_page_dirty(node_page); - } - f2fs_put_page(node_page, 1); + move_node_page(node_page, gc_type); stat_inc_node_blk_count(sbi, 1, gc_type); } @@ -834,18 +826,9 @@ static int do_garbage_collect(struct f2fs_sb_info *sbi, f2fs_put_page(sum_page, 0); } - if (gc_type == FG_GC) { - if (type == SUM_TYPE_NODE) { - struct writeback_control wbc = { - .sync_mode = WB_SYNC_ALL, - .nr_to_write = LONG_MAX, - .for_reclaim = 0, - }; - sync_node_pages(sbi, &wbc); - } else { - f2fs_submit_merged_bio(sbi, DATA, WRITE); - } - } + if (gc_type == FG_GC) + f2fs_submit_merged_bio(sbi, + (type == SUM_TYPE_NODE) ? NODE : DATA, WRITE); blk_finish_plug(&plug); diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c index de070a524fd2..f80cfb6beac6 100644 --- a/fs/f2fs/node.c +++ b/fs/f2fs/node.c @@ -1222,6 +1222,37 @@ iput_out: iput(inode); } +void move_node_page(struct page *node_page, int gc_type) +{ + if (gc_type == FG_GC) { + struct f2fs_sb_info *sbi = F2FS_P_SB(node_page); + struct writeback_control wbc = { + .sync_mode = WB_SYNC_ALL, + .nr_to_write = 1, + .for_reclaim = 0, + }; + + set_page_dirty(node_page); + f2fs_wait_on_page_writeback(node_page, NODE, true); + + f2fs_bug_on(sbi, PageWriteback(node_page)); + if (!clear_page_dirty_for_io(node_page)) + goto out_page; + + if (NODE_MAPPING(sbi)->a_ops->writepage(node_page, &wbc)) + unlock_page(node_page); + goto release_page; + } else { + /* set page dirty and write it */ + if (!PageWriteback(node_page)) + set_page_dirty(node_page); + } +out_page: + unlock_page(node_page); +release_page: + f2fs_put_page(node_page, 0); +} + static struct page *last_fsync_dnode(struct f2fs_sb_info *sbi, nid_t ino) { pgoff_t index, end; -- cgit v1.3-8-gc7d7 From fe216c7a0ff993cdac885109d8544ba02e6f9127 Mon Sep 17 00:00:00 2001 From: Yunlong Song Date: Wed, 27 Apr 2016 20:32:37 +0800 Subject: f2fs: fix to return 0 if err == -ENOENT in f2fs_readdir Commit 57b62d29ad5b384775974973087d47755a8c6fcc ("f2fs: fix to report error in f2fs_readdir") causes f2fs_readdir to return -ENOENT when get_lock_data_page returns -ENOENT. However, the original logic is to continue when get_lock_data_page returns -ENOENT, but it forgets to reset err to 0. This will cause getdents64 incorretly return -ENOENT when lastdirent is NULL in getdents64. This will lead to a wrong return value for syscall caller. Signed-off-by: Yunlong Song Reviewed-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/dir.c | 1 + 1 file changed, 1 insertion(+) (limited to 'fs') diff --git a/fs/f2fs/dir.c b/fs/f2fs/dir.c index 3b1c14e4eeea..dbfc1d1375a0 100644 --- a/fs/f2fs/dir.c +++ b/fs/f2fs/dir.c @@ -888,6 +888,7 @@ static int f2fs_readdir(struct file *file, struct dir_context *ctx) kunmap(dentry_page); f2fs_put_page(dentry_page, 1); } + err = 0; out: fscrypt_fname_free_buffer(&fstr); return err; -- cgit v1.3-8-gc7d7 From 34d9700702f4042ce10d68a092ab7f79575e7a3b Mon Sep 17 00:00:00 2001 From: Anand Jain Date: Wed, 16 Mar 2016 16:43:06 +0800 Subject: btrfs: rename btrfs_std_error to btrfs_handle_fs_error btrfs_std_error() handles errors, puts FS into readonly mode (as of now). So its good idea to rename it to btrfs_handle_fs_error(). Signed-off-by: Anand Jain Reviewed-by: David Sterba [ edit changelog ] Signed-off-by: David Sterba --- fs/btrfs/ctree.c | 6 +++--- fs/btrfs/ctree.h | 6 +++--- fs/btrfs/disk-io.c | 8 ++++---- fs/btrfs/extent-tree.c | 2 +- fs/btrfs/inode-item.c | 2 +- fs/btrfs/ioctl.c | 2 +- fs/btrfs/relocation.c | 2 +- fs/btrfs/root-tree.c | 4 ++-- fs/btrfs/super.c | 6 +++--- fs/btrfs/transaction.c | 2 +- fs/btrfs/tree-log.c | 8 ++++---- fs/btrfs/volumes.c | 14 +++++++------- 12 files changed, 31 insertions(+), 31 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c index ec7928a27aaa..decd0a3f5d61 100644 --- a/fs/btrfs/ctree.c +++ b/fs/btrfs/ctree.c @@ -1011,7 +1011,7 @@ static noinline int update_ref_for_cow(struct btrfs_trans_handle *trans, return ret; if (refs == 0) { ret = -EROFS; - btrfs_std_error(root->fs_info, ret, NULL); + btrfs_handle_fs_error(root->fs_info, ret, NULL); return ret; } } else { @@ -1928,7 +1928,7 @@ static noinline int balance_level(struct btrfs_trans_handle *trans, child = read_node_slot(root, mid, 0); if (!child) { ret = -EROFS; - btrfs_std_error(root->fs_info, ret, NULL); + btrfs_handle_fs_error(root->fs_info, ret, NULL); goto enospc; } @@ -2031,7 +2031,7 @@ static noinline int balance_level(struct btrfs_trans_handle *trans, */ if (!left) { ret = -EROFS; - btrfs_std_error(root->fs_info, ret, NULL); + btrfs_handle_fs_error(root->fs_info, ret, NULL); goto enospc; } wret = balance_node_right(trans, root, mid, left); diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 84a6a5b3384a..12f80cd230ab 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -4329,7 +4329,7 @@ static inline void assfail(char *expr, char *file, int line) #define btrfs_assert() __printf(5, 6) __cold -void __btrfs_std_error(struct btrfs_fs_info *fs_info, const char *function, +void __btrfs_handle_fs_error(struct btrfs_fs_info *fs_info, const char *function, unsigned int line, int errno, const char *fmt, ...); const char *btrfs_decode_error(int errno); @@ -4472,9 +4472,9 @@ do { \ __LINE__, (errno)); \ } while (0) -#define btrfs_std_error(fs_info, errno, fmt, args...) \ +#define btrfs_handle_fs_error(fs_info, errno, fmt, args...) \ do { \ - __btrfs_std_error((fs_info), __func__, __LINE__, \ + __btrfs_handle_fs_error((fs_info), __func__, __LINE__, \ (errno), fmt, ##args); \ } while (0) diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 4e47849d7427..aeb090583b78 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -2417,7 +2417,7 @@ static int btrfs_replay_log(struct btrfs_fs_info *fs_info, /* returns with log_tree_root freed on success */ ret = btrfs_recover_log_trees(log_tree_root); if (ret) { - btrfs_std_error(tree_root->fs_info, ret, + btrfs_handle_fs_error(tree_root->fs_info, ret, "Failed to recover log tree"); free_extent_buffer(log_tree_root->node); kfree(log_tree_root); @@ -3646,7 +3646,7 @@ static int write_all_supers(struct btrfs_root *root, int max_mirrors) if (ret) { mutex_unlock( &root->fs_info->fs_devices->device_list_mutex); - btrfs_std_error(root->fs_info, ret, + btrfs_handle_fs_error(root->fs_info, ret, "errors while submitting device barriers."); return ret; } @@ -3686,7 +3686,7 @@ static int write_all_supers(struct btrfs_root *root, int max_mirrors) mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); /* FUA is masked off if unsupported and can't be the reason */ - btrfs_std_error(root->fs_info, -EIO, + btrfs_handle_fs_error(root->fs_info, -EIO, "%d errors while writing supers", total_errors); return -EIO; } @@ -3704,7 +3704,7 @@ static int write_all_supers(struct btrfs_root *root, int max_mirrors) } mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); if (total_errors > max_errors) { - btrfs_std_error(root->fs_info, -EIO, + btrfs_handle_fs_error(root->fs_info, -EIO, "%d errors while writing supers", total_errors); return -EIO; } diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 84e060eb0de8..c0520a94a333 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -9058,7 +9058,7 @@ out: if (!for_reloc && root_dropped == false) btrfs_add_dead_root(root); if (err && err != -EAGAIN) - btrfs_std_error(root->fs_info, err, NULL); + btrfs_handle_fs_error(root->fs_info, err, NULL); return err; } diff --git a/fs/btrfs/inode-item.c b/fs/btrfs/inode-item.c index be4d22a5022f..b8acc07ac6c2 100644 --- a/fs/btrfs/inode-item.c +++ b/fs/btrfs/inode-item.c @@ -157,7 +157,7 @@ static int btrfs_del_inode_extref(struct btrfs_trans_handle *trans, */ if (!btrfs_find_name_in_ext_backref(path, ref_objectid, name, name_len, &extref)) { - btrfs_std_error(root->fs_info, -ENOENT, NULL); + btrfs_handle_fs_error(root->fs_info, -ENOENT, NULL); ret = -EROFS; goto out; } diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index 5a23806ae418..06fcc448108b 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -4847,7 +4847,7 @@ static long btrfs_ioctl_qgroup_assign(struct file *file, void __user *arg) /* update qgroup status and info */ err = btrfs_run_qgroups(trans, root->fs_info); if (err < 0) - btrfs_std_error(root->fs_info, ret, + btrfs_handle_fs_error(root->fs_info, ret, "failed to update qgroup status and info\n"); err = btrfs_end_transaction(trans, root); if (err && !ret) diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c index 08ef890deca6..1c29514d8aff 100644 --- a/fs/btrfs/relocation.c +++ b/fs/btrfs/relocation.c @@ -2418,7 +2418,7 @@ again: } out: if (ret) { - btrfs_std_error(root->fs_info, ret, NULL); + btrfs_handle_fs_error(root->fs_info, ret, NULL); if (!list_empty(&reloc_roots)) free_reloc_roots(&reloc_roots); diff --git a/fs/btrfs/root-tree.c b/fs/btrfs/root-tree.c index 9fcd6dfc3266..b2b14e7115f1 100644 --- a/fs/btrfs/root-tree.c +++ b/fs/btrfs/root-tree.c @@ -284,7 +284,7 @@ int btrfs_find_orphan_roots(struct btrfs_root *tree_root) trans = btrfs_join_transaction(tree_root); if (IS_ERR(trans)) { err = PTR_ERR(trans); - btrfs_std_error(tree_root->fs_info, err, + btrfs_handle_fs_error(tree_root->fs_info, err, "Failed to start trans to delete " "orphan item"); break; @@ -293,7 +293,7 @@ int btrfs_find_orphan_roots(struct btrfs_root *tree_root) root_key.objectid); btrfs_end_transaction(trans, tree_root); if (err) { - btrfs_std_error(tree_root->fs_info, err, + btrfs_handle_fs_error(tree_root->fs_info, err, "Failed to delete root orphan " "item"); break; diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index 00b8f37cc306..cc077887bd9d 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -131,11 +131,11 @@ static void btrfs_handle_error(struct btrfs_fs_info *fs_info) } /* - * __btrfs_std_error decodes expected errors from the caller and + * __btrfs_handle_fs_error decodes expected errors from the caller and * invokes the approciate error response. */ __cold -void __btrfs_std_error(struct btrfs_fs_info *fs_info, const char *function, +void __btrfs_handle_fs_error(struct btrfs_fs_info *fs_info, const char *function, unsigned int line, int errno, const char *fmt, ...) { struct super_block *sb = fs_info->sb; @@ -252,7 +252,7 @@ void __btrfs_abort_transaction(struct btrfs_trans_handle *trans, /* Wake up anybody who may be waiting on this transaction */ wake_up(&root->fs_info->transaction_wait); wake_up(&root->fs_info->transaction_blocked_wait); - __btrfs_std_error(root->fs_info, function, line, errno, NULL); + __btrfs_handle_fs_error(root->fs_info, function, line, errno, NULL); } /* * __btrfs_panic decodes unexpected, fatal errors from the caller, diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c index 43885e51b882..a469b55a7ac1 100644 --- a/fs/btrfs/transaction.c +++ b/fs/btrfs/transaction.c @@ -2145,7 +2145,7 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans, ret = btrfs_write_and_wait_transaction(trans, root); if (ret) { - btrfs_std_error(root->fs_info, ret, + btrfs_handle_fs_error(root->fs_info, ret, "Error while writing out transaction"); mutex_unlock(&root->fs_info->tree_log_mutex); goto scrub_continue; diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c index 517d0ccb351e..16a74d1a2720 100644 --- a/fs/btrfs/tree-log.c +++ b/fs/btrfs/tree-log.c @@ -5519,7 +5519,7 @@ int btrfs_recover_log_trees(struct btrfs_root *log_root_tree) ret = walk_log_tree(trans, log_root_tree, &wc); if (ret) { - btrfs_std_error(fs_info, ret, "Failed to pin buffers while " + btrfs_handle_fs_error(fs_info, ret, "Failed to pin buffers while " "recovering log root tree."); goto error; } @@ -5533,7 +5533,7 @@ again: ret = btrfs_search_slot(NULL, log_root_tree, &key, path, 0, 0); if (ret < 0) { - btrfs_std_error(fs_info, ret, + btrfs_handle_fs_error(fs_info, ret, "Couldn't find tree log root."); goto error; } @@ -5551,7 +5551,7 @@ again: log = btrfs_read_fs_root(log_root_tree, &found_key); if (IS_ERR(log)) { ret = PTR_ERR(log); - btrfs_std_error(fs_info, ret, + btrfs_handle_fs_error(fs_info, ret, "Couldn't read tree log root."); goto error; } @@ -5566,7 +5566,7 @@ again: free_extent_buffer(log->node); free_extent_buffer(log->commit_root); kfree(log); - btrfs_std_error(fs_info, ret, "Couldn't read target root " + btrfs_handle_fs_error(fs_info, ret, "Couldn't read target root " "for tree log recovery."); goto error; } diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index bd0f45fb38c4..0d3dd3e10e31 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -1454,7 +1454,7 @@ again: extent = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_dev_extent); } else { - btrfs_std_error(root->fs_info, ret, "Slot search failed"); + btrfs_handle_fs_error(root->fs_info, ret, "Slot search failed"); goto out; } @@ -1462,7 +1462,7 @@ again: ret = btrfs_del_item(trans, root, path); if (ret) { - btrfs_std_error(root->fs_info, ret, + btrfs_handle_fs_error(root->fs_info, ret, "Failed to remove dev extent item"); } else { set_bit(BTRFS_TRANS_HAVE_FREE_BGS, &trans->transaction->flags); @@ -2418,7 +2418,7 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path) ret = btrfs_relocate_sys_chunks(root); if (ret < 0) - btrfs_std_error(root->fs_info, ret, + btrfs_handle_fs_error(root->fs_info, ret, "Failed to relocate sys chunks after " "device initialization. This can be fixed " "using the \"btrfs balance\" command."); @@ -2663,7 +2663,7 @@ static int btrfs_free_chunk(struct btrfs_trans_handle *trans, if (ret < 0) goto out; else if (ret > 0) { /* Logic error or corruption */ - btrfs_std_error(root->fs_info, -ENOENT, + btrfs_handle_fs_error(root->fs_info, -ENOENT, "Failed lookup while freeing chunk."); ret = -ENOENT; goto out; @@ -2671,7 +2671,7 @@ static int btrfs_free_chunk(struct btrfs_trans_handle *trans, ret = btrfs_del_item(trans, root, path); if (ret < 0) - btrfs_std_error(root->fs_info, ret, + btrfs_handle_fs_error(root->fs_info, ret, "Failed to delete chunk item."); out: btrfs_free_path(path); @@ -2857,7 +2857,7 @@ static int btrfs_relocate_chunk(struct btrfs_root *root, u64 chunk_offset) chunk_offset); if (IS_ERR(trans)) { ret = PTR_ERR(trans); - btrfs_std_error(root->fs_info, ret, NULL); + btrfs_handle_fs_error(root->fs_info, ret, NULL); return ret; } @@ -3632,7 +3632,7 @@ static void __cancel_balance(struct btrfs_fs_info *fs_info) unset_balance_control(fs_info); ret = del_balance_item(fs_info->tree_root); if (ret) - btrfs_std_error(fs_info, ret, NULL); + btrfs_handle_fs_error(fs_info, ret, NULL); atomic_set(&fs_info->mutually_exclusive_operation_running, 0); } -- cgit v1.3-8-gc7d7 From 2351d743f6b69d289f04246127c9d92516621974 Mon Sep 17 00:00:00 2001 From: Anand Jain Date: Wed, 16 Mar 2016 16:43:07 +0800 Subject: btrfs: remove unused function btrfs_assert() Apparently looks like ASSERT does the same intended job, as intended btrfs_assert(). Signed-off-by: Anand Jain Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/ctree.h | 1 - 1 file changed, 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 12f80cd230ab..4f37e3766626 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -4326,7 +4326,6 @@ static inline void assfail(char *expr, char *file, int line) #define ASSERT(expr) ((void)0) #endif -#define btrfs_assert() __printf(5, 6) __cold void __btrfs_handle_fs_error(struct btrfs_fs_info *fs_info, const char *function, -- cgit v1.3-8-gc7d7 From c5f4ccb2f77355ac44433feb362fecd7f1a7d03e Mon Sep 17 00:00:00 2001 From: Anand Jain Date: Wed, 16 Mar 2016 16:43:08 +0800 Subject: btrfs: move error handling code together in ctree.h Looks like we added the incompatible defines in between the error handling defines in the file ctree.h. Now group them back. Signed-off-by: Anand Jain Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/ctree.h | 78 +++++++++++++++++++++++++++++--------------------------- 1 file changed, 40 insertions(+), 38 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 4f37e3766626..e22568a54545 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -4338,6 +4338,46 @@ void __btrfs_abort_transaction(struct btrfs_trans_handle *trans, struct btrfs_root *root, const char *function, unsigned int line, int errno); +/* + * Call btrfs_abort_transaction as early as possible when an error condition is + * detected, that way the exact line number is reported. + */ +#define btrfs_abort_transaction(trans, root, errno) \ +do { \ + /* Report first abort since mount */ \ + if (!test_and_set_bit(BTRFS_FS_STATE_TRANS_ABORTED, \ + &((root)->fs_info->fs_state))) { \ + WARN(1, KERN_DEBUG \ + "BTRFS: Transaction aborted (error %d)\n", \ + (errno)); \ + } \ + __btrfs_abort_transaction((trans), (root), __func__, \ + __LINE__, (errno)); \ +} while (0) + +#define btrfs_handle_fs_error(fs_info, errno, fmt, args...) \ +do { \ + __btrfs_handle_fs_error((fs_info), __func__, __LINE__, \ + (errno), fmt, ##args); \ +} while (0) + +__printf(5, 6) +__cold +void __btrfs_panic(struct btrfs_fs_info *fs_info, const char *function, + unsigned int line, int errno, const char *fmt, ...); +/* + * If BTRFS_MOUNT_PANIC_ON_FATAL_ERROR is in mount_opt, __btrfs_panic + * will panic(). Otherwise we BUG() here. + */ +#define btrfs_panic(fs_info, errno, fmt, args...) \ +do { \ + __btrfs_panic(fs_info, __func__, __LINE__, errno, fmt, ##args); \ + BUG(); \ +} while (0) + + +/* compatibility and incompatibility defines */ + #define btrfs_set_fs_incompat(__fs_info, opt) \ __btrfs_set_fs_incompat((__fs_info), BTRFS_FEATURE_INCOMPAT_##opt) @@ -4454,44 +4494,6 @@ static inline int __btrfs_fs_compat_ro(struct btrfs_fs_info *fs_info, u64 flag) return !!(btrfs_super_compat_ro_flags(disk_super) & flag); } -/* - * Call btrfs_abort_transaction as early as possible when an error condition is - * detected, that way the exact line number is reported. - */ -#define btrfs_abort_transaction(trans, root, errno) \ -do { \ - /* Report first abort since mount */ \ - if (!test_and_set_bit(BTRFS_FS_STATE_TRANS_ABORTED, \ - &((root)->fs_info->fs_state))) { \ - WARN(1, KERN_DEBUG \ - "BTRFS: Transaction aborted (error %d)\n", \ - (errno)); \ - } \ - __btrfs_abort_transaction((trans), (root), __func__, \ - __LINE__, (errno)); \ -} while (0) - -#define btrfs_handle_fs_error(fs_info, errno, fmt, args...) \ -do { \ - __btrfs_handle_fs_error((fs_info), __func__, __LINE__, \ - (errno), fmt, ##args); \ -} while (0) - -__printf(5, 6) -__cold -void __btrfs_panic(struct btrfs_fs_info *fs_info, const char *function, - unsigned int line, int errno, const char *fmt, ...); - -/* - * If BTRFS_MOUNT_PANIC_ON_FATAL_ERROR is in mount_opt, __btrfs_panic - * will panic(). Otherwise we BUG() here. - */ -#define btrfs_panic(fs_info, errno, fmt, args...) \ -do { \ - __btrfs_panic(fs_info, __func__, __LINE__, errno, fmt, ##args); \ - BUG(); \ -} while (0) - /* acl.c */ #ifdef CONFIG_BTRFS_FS_POSIX_ACL struct posix_acl *btrfs_get_acl(struct inode *inode, int type); -- cgit v1.3-8-gc7d7 From 13f48dc9094b56c5bffd8d57349a0a01a1926b2d Mon Sep 17 00:00:00 2001 From: Satoru Takeuchi Date: Tue, 15 Mar 2016 09:09:59 +0900 Subject: btrfs: Simplify conditions about compress while mapping btrfs flags to inode flags Signed-off-by: Satoru Takeuchi Signed-off-by: David Sterba --- fs/btrfs/ioctl.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index 06fcc448108b..0f76aea6e398 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -125,10 +125,10 @@ static unsigned int btrfs_flags_to_ioctl(unsigned int flags) if (flags & BTRFS_INODE_NODATACOW) iflags |= FS_NOCOW_FL; - if ((flags & BTRFS_INODE_COMPRESS) && !(flags & BTRFS_INODE_NOCOMPRESS)) - iflags |= FS_COMPR_FL; - else if (flags & BTRFS_INODE_NOCOMPRESS) + if (flags & BTRFS_INODE_NOCOMPRESS) iflags |= FS_NOCOMP_FL; + else if (flags & BTRFS_INODE_COMPRESS) + iflags |= FS_COMPR_FL; return iflags; } -- cgit v1.3-8-gc7d7 From 0713d90c75745dc6148f6346d490e9ef63a4e8b4 Mon Sep 17 00:00:00 2001 From: Anand Jain Date: Thu, 17 Mar 2016 10:38:57 +0800 Subject: btrfs: remove save_error_info() Actually save_error_info() sets the FS state to error and nothing else. Further the word save doesn't induce caffeine when compared to the word set in what actually it does. So to make it better understandable move save_error_info() code to its only consumer itself. Signed-off-by: Anand Jain Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/super.c | 16 ++++++---------- 1 file changed, 6 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index cc077887bd9d..dab51118b972 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -97,15 +97,6 @@ const char *btrfs_decode_error(int errno) return errstr; } -static void save_error_info(struct btrfs_fs_info *fs_info) -{ - /* - * today we only save the error info into ram. Long term we'll - * also send it down to the disk - */ - set_bit(BTRFS_FS_STATE_ERROR, &fs_info->fs_state); -} - /* btrfs handle error by forcing the filesystem readonly */ static void btrfs_handle_error(struct btrfs_fs_info *fs_info) { @@ -170,8 +161,13 @@ void __btrfs_handle_fs_error(struct btrfs_fs_info *fs_info, const char *function } #endif + /* + * Today we only save the error info to memory. Long term we'll + * also send it down to the disk + */ + set_bit(BTRFS_FS_STATE_ERROR, &fs_info->fs_state); + /* Don't go through full error handling during mount */ - save_error_info(fs_info); if (sb->s_flags & MS_BORN) btrfs_handle_error(fs_info); } -- cgit v1.3-8-gc7d7 From 6719afdcf808b38145095b2e485c0f90bfc29722 Mon Sep 17 00:00:00 2001 From: Geert Uytterhoeven Date: Sun, 22 Jun 2014 14:30:09 +0200 Subject: Btrfs: Refactor btrfs_lock_cluster() to kill compiler warning MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit fs/btrfs/extent-tree.c: In function ‘btrfs_lock_cluster’: fs/btrfs/extent-tree.c:6399: warning: ‘used_bg’ may be used uninitialized in this function - Replace "again: ... goto again;" by standard C "while (1) { ... }", - Move block not processed during the first iteration of the loop to the end of the loop, which allows to kill the "locked" variable, Signed-off-by: Geert Uytterhoeven Reviewed-and-Tested-by: Miao Xie [ the compilation warning has been fixed by other patch, now we want to clean up the function ] Signed-off-by: David Sterba --- fs/btrfs/extent-tree.c | 41 ++++++++++++++++++++--------------------- 1 file changed, 20 insertions(+), 21 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index c0520a94a333..b66f460f779a 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -7025,36 +7025,35 @@ btrfs_lock_cluster(struct btrfs_block_group_cache *block_group, int delalloc) { struct btrfs_block_group_cache *used_bg = NULL; - bool locked = false; -again: + spin_lock(&cluster->refill_lock); - if (locked) { - if (used_bg == cluster->block_group) + while (1) { + used_bg = cluster->block_group; + if (!used_bg) + return NULL; + + if (used_bg == block_group) return used_bg; - up_read(&used_bg->data_rwsem); - btrfs_put_block_group(used_bg); - } + btrfs_get_block_group(used_bg); - used_bg = cluster->block_group; - if (!used_bg) - return NULL; + if (!delalloc) + return used_bg; - if (used_bg == block_group) - return used_bg; + if (down_read_trylock(&used_bg->data_rwsem)) + return used_bg; - btrfs_get_block_group(used_bg); + spin_unlock(&cluster->refill_lock); - if (!delalloc) - return used_bg; + down_read(&used_bg->data_rwsem); - if (down_read_trylock(&used_bg->data_rwsem)) - return used_bg; + spin_lock(&cluster->refill_lock); + if (used_bg == cluster->block_group) + return used_bg; - spin_unlock(&cluster->refill_lock); - down_read(&used_bg->data_rwsem); - locked = true; - goto again; + up_read(&used_bg->data_rwsem); + btrfs_put_block_group(used_bg); + } } static inline void -- cgit v1.3-8-gc7d7 From 180e4d4700b1b7bfdffd9e58ae95220ae9482d17 Mon Sep 17 00:00:00 2001 From: Luis de Bethencourt Date: Mon, 4 Apr 2016 15:31:22 +0100 Subject: btrfs: fix typos in comments Correct a typo in the chunk_mutex name to make it grepable. Since it is better to fix several typos at once, fixing the 2 more in the same file. Signed-off-by: Luis de Bethencourt Signed-off-by: David Sterba --- fs/btrfs/super.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index dab51118b972..957976d9d627 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -1484,10 +1484,10 @@ static int setup_security_options(struct btrfs_fs_info *fs_info, memcpy(&fs_info->security_opts, sec_opts, sizeof(*sec_opts)); } else { /* - * Since SELinux(the only one supports security_mnt_opts) does - * NOT support changing context during remount/mount same sb, - * This must be the same or part of the same security options, - * just free it. + * Since SELinux (the only one supporting security_mnt_opts) + * does NOT support changing context during remount/mount of + * the same sb, this must be the same or part of the same + * security options, just free it. */ security_free_mnt_opts(sec_opts); } @@ -1665,8 +1665,8 @@ static inline void btrfs_remount_cleanup(struct btrfs_fs_info *fs_info, unsigned long old_opts) { /* - * We need cleanup all defragable inodes if the autodefragment is - * close or the fs is R/O. + * We need to cleanup all defragable inodes if the autodefragment is + * close or the filesystem is read only. */ if (btrfs_raw_test_opt(old_opts, AUTO_DEFRAG) && (!btrfs_raw_test_opt(fs_info->mount_opt, AUTO_DEFRAG) || @@ -2049,7 +2049,7 @@ static int btrfs_statfs(struct dentry *dentry, struct kstatfs *buf) u64 thresh = 0; /* - * holding chunk_muext to avoid allocating new chunks, holding + * holding chunk_mutex to avoid allocating new chunks, holding * device_list_mutex to avoid the device being removed */ rcu_read_lock(); -- cgit v1.3-8-gc7d7 From 4c63c2454eff996c5e27991221106eb511f7db38 Mon Sep 17 00:00:00 2001 From: Luke Dashjr Date: Thu, 29 Oct 2015 08:22:21 +0000 Subject: btrfs: bugfix: handle FS_IOC32_{GETFLAGS,SETFLAGS,GETVERSION} in btrfs_ioctl 32-bit ioctl uses these rather than the regular FS_IOC_* versions. They can be handled in btrfs using the same code. Without this, 32-bit {ch,ls}attr fail. Signed-off-by: Luke Dashjr Cc: stable@vger.kernel.org Reviewed-by: Josef Bacik Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/ctree.h | 1 + fs/btrfs/file.c | 2 +- fs/btrfs/inode.c | 2 +- fs/btrfs/ioctl.c | 21 +++++++++++++++++++++ 4 files changed, 24 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 84a6a5b3384a..208d19938fdf 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -4122,6 +4122,7 @@ void btrfs_test_inode_set_ops(struct inode *inode); /* ioctl.c */ long btrfs_ioctl(struct file *file, unsigned int cmd, unsigned long arg); +long btrfs_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg); int btrfs_ioctl_get_supported_features(void __user *arg); void btrfs_update_iflags(struct inode *inode); void btrfs_inherit_iflags(struct inode *inode, struct inode *dir); diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index 8d7b5a45c005..751daacd268d 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -2956,7 +2956,7 @@ const struct file_operations btrfs_file_operations = { .fallocate = btrfs_fallocate, .unlocked_ioctl = btrfs_ioctl, #ifdef CONFIG_COMPAT - .compat_ioctl = btrfs_ioctl, + .compat_ioctl = btrfs_compat_ioctl, #endif .copy_file_range = btrfs_copy_file_range, .clone_file_range = btrfs_clone_file_range, diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 2aaba58b4856..167fc3d49450 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -10184,7 +10184,7 @@ static const struct file_operations btrfs_dir_file_operations = { .iterate = btrfs_real_readdir, .unlocked_ioctl = btrfs_ioctl, #ifdef CONFIG_COMPAT - .compat_ioctl = btrfs_ioctl, + .compat_ioctl = btrfs_compat_ioctl, #endif .release = btrfs_release_file, .fsync = btrfs_sync_file, diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index 5a23806ae418..f545f81f642d 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -5552,3 +5552,24 @@ long btrfs_ioctl(struct file *file, unsigned int return -ENOTTY; } + +#ifdef CONFIG_COMPAT +long btrfs_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg) +{ + switch (cmd) { + case FS_IOC32_GETFLAGS: + cmd = FS_IOC_GETFLAGS; + break; + case FS_IOC32_SETFLAGS: + cmd = FS_IOC_SETFLAGS; + break; + case FS_IOC32_GETVERSION: + cmd = FS_IOC_GETVERSION; + break; + default: + return -ENOIOCTLCMD; + } + + return btrfs_ioctl(file, cmd, (unsigned long) compat_ptr(arg)); +} +#endif -- cgit v1.3-8-gc7d7 From a91326679f2a0a4c239cd643674fdcda28ee86be Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Mon, 7 Mar 2016 16:56:21 -0800 Subject: Btrfs: make mapping->writeback_index point to the last written page MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit If sequential writer is writing in the middle of the page and it just redirties the last written page by continuing from it. In the above case this can end up with seeking back to that firstly redirtied page after writing all the pages at the end of file because btrfs updates mapping->writeback_index to 1 past the current one. For non-cow filesystems, the cost is only about extra seek, while for cow filesystems such as btrfs, it means unnecessary fragments. To avoid it, we just need to continue writeback from the last written page. This also updates btrfs to behave like what write_cache_pages() does, ie, bail out immediately if there is an error in writepage(). Reported-by: Holger Hoffstätte Signed-off-by: Liu Bo Signed-off-by: David Sterba --- fs/btrfs/extent_io.c | 32 ++++++++++++++++++++++++++------ 1 file changed, 26 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index d247fc0eea19..fd41a730d239 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -3200,14 +3200,10 @@ int extent_read_full_page(struct extent_io_tree *tree, struct page *page, return ret; } -static noinline void update_nr_written(struct page *page, - struct writeback_control *wbc, - unsigned long nr_written) +static void update_nr_written(struct page *page, struct writeback_control *wbc, + unsigned long nr_written) { wbc->nr_to_write -= nr_written; - if (wbc->range_cyclic || (wbc->nr_to_write > 0 && - wbc->range_start == 0 && wbc->range_end == LLONG_MAX)) - page->mapping->writeback_index = page->index + nr_written; } /* @@ -3926,6 +3922,8 @@ static int extent_write_cache_pages(struct extent_io_tree *tree, int nr_pages; pgoff_t index; pgoff_t end; /* Inclusive */ + pgoff_t done_index; + int range_whole = 0; int scanned = 0; int tag; @@ -3948,6 +3946,8 @@ static int extent_write_cache_pages(struct extent_io_tree *tree, } else { index = wbc->range_start >> PAGE_SHIFT; end = wbc->range_end >> PAGE_SHIFT; + if (wbc->range_start == 0 && wbc->range_end == LLONG_MAX) + range_whole = 1; scanned = 1; } if (wbc->sync_mode == WB_SYNC_ALL) @@ -3957,6 +3957,7 @@ static int extent_write_cache_pages(struct extent_io_tree *tree, retry: if (wbc->sync_mode == WB_SYNC_ALL) tag_pages_for_writeback(mapping, index, end); + done_index = index; while (!done && !nr_to_write_done && (index <= end) && (nr_pages = pagevec_lookup_tag(&pvec, mapping, &index, tag, min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1))) { @@ -3966,6 +3967,7 @@ retry: for (i = 0; i < nr_pages; i++) { struct page *page = pvec.pages[i]; + done_index = page->index; /* * At this point we hold neither mapping->tree_lock nor * lock on the page itself: the page may be truncated or @@ -4009,6 +4011,20 @@ retry: } if (!err && ret < 0) err = ret; + if (ret < 0) { + /* + * done_index is set past this page, + * so media errors will not choke + * background writeout for the entire + * file. This has consequences for + * range_cyclic semantics (ie. it may + * not be suitable for data integrity + * writeout). + */ + done_index = page->index + 1; + done = 1; + break; + } /* * the filesystem may choose to bump up nr_to_write. @@ -4029,6 +4045,10 @@ retry: index = 0; goto retry; } + + if (wbc->range_cyclic || (wbc->nr_to_write > 0 && range_whole)) + mapping->writeback_index = done_index; + btrfs_add_delayed_iput(inode); return err; } -- cgit v1.3-8-gc7d7 From 894b36e35ae01186b77b083f3f67569a349062a6 Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Mon, 7 Mar 2016 16:56:22 -0800 Subject: Btrfs: cleanup error handling in extent_write_cached_pages Now that we bail out immediately if ->writepage() returns an error, we don't need an extra error to retain the error code. Signed-off-by: Liu Bo Signed-off-by: David Sterba --- fs/btrfs/extent_io.c | 7 ++----- 1 file changed, 2 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index fd41a730d239..b67d6d24440b 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -3916,7 +3916,6 @@ static int extent_write_cache_pages(struct extent_io_tree *tree, struct inode *inode = mapping->host; int ret = 0; int done = 0; - int err = 0; int nr_to_write_done = 0; struct pagevec pvec; int nr_pages; @@ -4009,8 +4008,6 @@ retry: unlock_page(page); ret = 0; } - if (!err && ret < 0) - err = ret; if (ret < 0) { /* * done_index is set past this page, @@ -4036,7 +4033,7 @@ retry: pagevec_release(&pvec); cond_resched(); } - if (!scanned && !done && !err) { + if (!scanned && !done) { /* * We hit the last page and there is more work to be done: wrap * back to the start of the file @@ -4050,7 +4047,7 @@ retry: mapping->writeback_index = done_index; btrfs_add_delayed_iput(inode); - return err; + return ret; } static void flush_epd_write_bio(struct extent_page_data *epd) -- cgit v1.3-8-gc7d7 From a2af23b7d7cb0de89570e97da84f6fb642e990a4 Mon Sep 17 00:00:00 2001 From: Chandan Rajendra Date: Mon, 4 Apr 2016 02:53:06 +0530 Subject: Btrfs: __btrfs_buffered_write: Pass valid file offset when releasing delalloc space The delalloc reserved space is calculated in terms of number of bytes used by an integral number of blocks. This is done by rounding down the value of 'pos' to the nearest multiple of sectorsize. The file offset value held by 'pos' variable may not be aligned to sectorsize and hence when passing it as an argument to btrfs_delalloc_release_space(), we may end up releasing larger delalloc space than we originally had reserved. Signed-off-by: Chandan Rajendra Signed-off-by: David Sterba --- fs/btrfs/file.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index 751daacd268d..af059c44684d 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -1696,7 +1696,9 @@ again: btrfs_end_write_no_snapshoting(root); btrfs_delalloc_release_metadata(inode, release_bytes); } else { - btrfs_delalloc_release_space(inode, pos, release_bytes); + btrfs_delalloc_release_space(inode, + round_down(pos, root->sectorsize), + release_bytes); } } -- cgit v1.3-8-gc7d7 From cf25ce518e8ef9d59b292e51193bed2b023a32da Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Mon, 14 Dec 2015 18:29:32 -0800 Subject: Btrfs: do not create empty block group if we have allocated data Now we force to create empty block group to keep data profile alive, however, in the below example, we eventually get an empty block group while we're trying to get more space for other types (metadata/system), - Before, block group "A": size=2G, used=1.2G block group "B": size=2G, used=512M - After "btrfs balance start -dusage=50 mount_point", block group "A": size=2G, used=(1.2+0.5)G block group "C": size=2G, used=0 Since there is no data in block group C, it won't be deleted automatically and we have to get the unused 2G until the next mount. Balance itself just moves data and doesn't remove data, so it's safe to not create such a empty block group if we already have data allocated in other block groups. Signed-off-by: Liu Bo Signed-off-by: David Sterba --- fs/btrfs/volumes.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index bd0f45fb38c4..745a619241c7 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -3402,6 +3402,7 @@ static int __btrfs_balance(struct btrfs_fs_info *fs_info) u32 count_meta = 0; u32 count_sys = 0; int chunk_reserved = 0; + u64 bytes_used = 0; /* step one make some room on all the devices */ devices = &fs_info->fs_devices->devices; @@ -3540,7 +3541,13 @@ again: goto loop; } - if ((chunk_type & BTRFS_BLOCK_GROUP_DATA) && !chunk_reserved) { + ASSERT(fs_info->data_sinfo); + spin_lock(&fs_info->data_sinfo->lock); + bytes_used = fs_info->data_sinfo->bytes_used; + spin_unlock(&fs_info->data_sinfo->lock); + + if ((chunk_type & BTRFS_BLOCK_GROUP_DATA) && + !chunk_reserved && !bytes_used) { trans = btrfs_start_transaction(chunk_root, 0); if (IS_ERR(trans)) { mutex_unlock(&fs_info->delete_unused_bgs_mutex); -- cgit v1.3-8-gc7d7 From 6cf86a006be9f2a77434d5a06ea289719815ad7c Mon Sep 17 00:00:00 2001 From: Anand Jain Date: Sat, 13 Feb 2016 10:01:29 +0800 Subject: btrfs: create a helper function to read the disk super A part of code from btrfs_scan_one_device() is moved to a new function btrfs_read_disk_super(), so that former function looks cleaner. (In this process it also moves the code which ensures null terminating label). So this creates easy opportunity to merge various duplicate codes on read disk super. Earlier attempt to merge duplicate codes highlighted that there were some issues for which there are duplicate codes (to read disk super), however it was not clear what was the issue. So until we figure that out, its better to keep them in a separate functions. Signed-off-by: Anand Jain [ use GFP_KERNEL, PAGE_CACHE_ removal related fixups ] Signed-off-by: David Sterba --- fs/btrfs/volumes.c | 87 ++++++++++++++++++++++++++++++++---------------------- 1 file changed, 52 insertions(+), 35 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index bd0f45fb38c4..5006683283ae 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -988,6 +988,56 @@ int btrfs_open_devices(struct btrfs_fs_devices *fs_devices, return ret; } +void btrfs_release_disk_super(struct page *page) +{ + kunmap(page); + put_page(page); +} + +int btrfs_read_disk_super(struct block_device *bdev, u64 bytenr, + struct page **page, struct btrfs_super_block **disk_super) +{ + void *p; + pgoff_t index; + + /* make sure our super fits in the device */ + if (bytenr + PAGE_SIZE >= i_size_read(bdev->bd_inode)) + return 1; + + /* make sure our super fits in the page */ + if (sizeof(**disk_super) > PAGE_SIZE) + return 1; + + /* make sure our super doesn't straddle pages on disk */ + index = bytenr >> PAGE_SHIFT; + if ((bytenr + sizeof(**disk_super) - 1) >> PAGE_SHIFT != index) + return 1; + + /* pull in the page with our super */ + *page = read_cache_page_gfp(bdev->bd_inode->i_mapping, + index, GFP_KERNEL); + + if (IS_ERR_OR_NULL(*page)) + return 1; + + p = kmap(*page); + + /* align our pointer to the offset of the super block */ + *disk_super = p + (bytenr & ~PAGE_MASK); + + if (btrfs_super_bytenr(*disk_super) != bytenr || + btrfs_super_magic(*disk_super) != BTRFS_MAGIC) { + btrfs_release_disk_super(*page); + return 1; + } + + if ((*disk_super)->label[0] && + (*disk_super)->label[BTRFS_LABEL_SIZE - 1]) + (*disk_super)->label[BTRFS_LABEL_SIZE - 1] = '\0'; + + return 0; +} + /* * Look for a btrfs signature on a device. This may be called out of the mount path * and we are not allowed to call set_blocksize during the scan. The superblock @@ -999,13 +1049,11 @@ int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder, struct btrfs_super_block *disk_super; struct block_device *bdev; struct page *page; - void *p; int ret = -EINVAL; u64 devid; u64 transid; u64 total_devices; u64 bytenr; - pgoff_t index; /* * we would like to check all the supers, but that would make @@ -1018,41 +1066,14 @@ int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder, mutex_lock(&uuid_mutex); bdev = blkdev_get_by_path(path, flags, holder); - if (IS_ERR(bdev)) { ret = PTR_ERR(bdev); goto error; } - /* make sure our super fits in the device */ - if (bytenr + PAGE_SIZE >= i_size_read(bdev->bd_inode)) - goto error_bdev_put; - - /* make sure our super fits in the page */ - if (sizeof(*disk_super) > PAGE_SIZE) - goto error_bdev_put; - - /* make sure our super doesn't straddle pages on disk */ - index = bytenr >> PAGE_SHIFT; - if ((bytenr + sizeof(*disk_super) - 1) >> PAGE_SHIFT != index) - goto error_bdev_put; - - /* pull in the page with our super */ - page = read_cache_page_gfp(bdev->bd_inode->i_mapping, - index, GFP_NOFS); - - if (IS_ERR_OR_NULL(page)) + if (btrfs_read_disk_super(bdev, bytenr, &page, &disk_super)) goto error_bdev_put; - p = kmap(page); - - /* align our pointer to the offset of the super block */ - disk_super = p + (bytenr & ~PAGE_MASK); - - if (btrfs_super_bytenr(disk_super) != bytenr || - btrfs_super_magic(disk_super) != BTRFS_MAGIC) - goto error_unmap; - devid = btrfs_stack_device_id(&disk_super->dev_item); transid = btrfs_super_generation(disk_super); total_devices = btrfs_super_num_devices(disk_super); @@ -1060,8 +1081,6 @@ int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder, ret = device_list_add(path, disk_super, devid, fs_devices_ret); if (ret > 0) { if (disk_super->label[0]) { - if (disk_super->label[BTRFS_LABEL_SIZE - 1]) - disk_super->label[BTRFS_LABEL_SIZE - 1] = '\0'; printk(KERN_INFO "BTRFS: device label %s ", disk_super->label); } else { printk(KERN_INFO "BTRFS: device fsid %pU ", disk_super->fsid); @@ -1073,9 +1092,7 @@ int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder, if (!ret && fs_devices_ret) (*fs_devices_ret)->total_devices = total_devices; -error_unmap: - kunmap(page); - put_page(page); + btrfs_release_disk_super(page); error_bdev_put: blkdev_put(bdev, flags); -- cgit v1.3-8-gc7d7 From f1fa7f264250f2cb60119aca3fd114c3f47070c2 Mon Sep 17 00:00:00 2001 From: Anand Jain Date: Sat, 13 Feb 2016 10:01:33 +0800 Subject: btrfs: create helper function __check_raid_min_devices() move a section of btrfs_rm_device() code to check for min number of the devices into the function __check_raid_min_devices() Signed-off-by: Anand Jain Signed-off-by: David Sterba --- fs/btrfs/volumes.c | 51 ++++++++++++++++++++++++++++++++------------------- 1 file changed, 32 insertions(+), 19 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 5006683283ae..db781d0d3678 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -1705,23 +1705,20 @@ out: return ret; } -int btrfs_rm_device(struct btrfs_root *root, char *device_path) +static int __check_raid_min_devices(struct btrfs_root *root) { - struct btrfs_device *device; - struct btrfs_device *next_device; - struct block_device *bdev; - struct buffer_head *bh = NULL; - struct btrfs_super_block *disk_super; - struct btrfs_fs_devices *cur_devices; u64 all_avail; - u64 devid; u64 num_devices; - u8 *dev_uuid; unsigned seq; int ret = 0; - bool clear_super = false; - mutex_lock(&uuid_mutex); + num_devices = root->fs_info->fs_devices->num_devices; + btrfs_dev_replace_lock(&root->fs_info->dev_replace, 0); + if (btrfs_dev_replace_is_ongoing(&root->fs_info->dev_replace)) { + WARN_ON(num_devices < 1); + num_devices--; + } + btrfs_dev_replace_unlock(&root->fs_info->dev_replace, 0); do { seq = read_seqbegin(&root->fs_info->profiles_lock); @@ -1731,14 +1728,6 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) root->fs_info->avail_metadata_alloc_bits; } while (read_seqretry(&root->fs_info->profiles_lock, seq)); - num_devices = root->fs_info->fs_devices->num_devices; - btrfs_dev_replace_lock(&root->fs_info->dev_replace, 0); - if (btrfs_dev_replace_is_ongoing(&root->fs_info->dev_replace)) { - WARN_ON(num_devices < 1); - num_devices--; - } - btrfs_dev_replace_unlock(&root->fs_info->dev_replace, 0); - if ((all_avail & BTRFS_BLOCK_GROUP_RAID10) && num_devices <= 4) { ret = BTRFS_ERROR_DEV_RAID10_MIN_NOT_MET; goto out; @@ -1760,6 +1749,30 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) goto out; } +out: + return ret; +} + +int btrfs_rm_device(struct btrfs_root *root, char *device_path) +{ + struct btrfs_device *device; + struct btrfs_device *next_device; + struct block_device *bdev; + struct buffer_head *bh = NULL; + struct btrfs_super_block *disk_super; + struct btrfs_fs_devices *cur_devices; + u64 devid; + u64 num_devices; + u8 *dev_uuid; + int ret = 0; + bool clear_super = false; + + mutex_lock(&uuid_mutex); + + ret = __check_raid_min_devices(root); + if (ret) + goto out; + if (strcmp(device_path, "missing") == 0) { struct list_head *devices; struct btrfs_device *tmp; -- cgit v1.3-8-gc7d7 From bd45ffbcb1f082e3c5a0bd56b1d7310d8b707ffb Mon Sep 17 00:00:00 2001 From: Anand Jain Date: Sat, 13 Feb 2016 10:01:34 +0800 Subject: btrfs: clean up and optimize __check_raid_min_device() __check_raid_min_device() which was pealed from btrfs_rm_device() maintianed its original code to show the block move. This patch cleans up __check_raid_min_device(). Signed-off-by: Anand Jain Signed-off-by: David Sterba --- fs/btrfs/volumes.c | 43 +++++++++++++++++++------------------------ 1 file changed, 19 insertions(+), 24 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index db781d0d3678..5538dc78c72e 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -1705,52 +1705,47 @@ out: return ret; } -static int __check_raid_min_devices(struct btrfs_root *root) +static int __check_raid_min_devices(struct btrfs_fs_info *fs_info) { u64 all_avail; u64 num_devices; unsigned seq; - int ret = 0; - num_devices = root->fs_info->fs_devices->num_devices; - btrfs_dev_replace_lock(&root->fs_info->dev_replace, 0); - if (btrfs_dev_replace_is_ongoing(&root->fs_info->dev_replace)) { + num_devices = fs_info->fs_devices->num_devices; + btrfs_dev_replace_lock(&fs_info->dev_replace, 0); + if (btrfs_dev_replace_is_ongoing(&fs_info->dev_replace)) { WARN_ON(num_devices < 1); num_devices--; } - btrfs_dev_replace_unlock(&root->fs_info->dev_replace, 0); + btrfs_dev_replace_unlock(&fs_info->dev_replace, 0); do { - seq = read_seqbegin(&root->fs_info->profiles_lock); + seq = read_seqbegin(&fs_info->profiles_lock); - all_avail = root->fs_info->avail_data_alloc_bits | - root->fs_info->avail_system_alloc_bits | - root->fs_info->avail_metadata_alloc_bits; - } while (read_seqretry(&root->fs_info->profiles_lock, seq)); + all_avail = fs_info->avail_data_alloc_bits | + fs_info->avail_system_alloc_bits | + fs_info->avail_metadata_alloc_bits; + } while (read_seqretry(&fs_info->profiles_lock, seq)); if ((all_avail & BTRFS_BLOCK_GROUP_RAID10) && num_devices <= 4) { - ret = BTRFS_ERROR_DEV_RAID10_MIN_NOT_MET; - goto out; + return BTRFS_ERROR_DEV_RAID10_MIN_NOT_MET; } if ((all_avail & BTRFS_BLOCK_GROUP_RAID1) && num_devices <= 2) { - ret = BTRFS_ERROR_DEV_RAID1_MIN_NOT_MET; - goto out; + return BTRFS_ERROR_DEV_RAID1_MIN_NOT_MET; } if ((all_avail & BTRFS_BLOCK_GROUP_RAID5) && - root->fs_info->fs_devices->rw_devices <= 2) { - ret = BTRFS_ERROR_DEV_RAID5_MIN_NOT_MET; - goto out; + fs_info->fs_devices->rw_devices <= 2) { + return BTRFS_ERROR_DEV_RAID5_MIN_NOT_MET; } + if ((all_avail & BTRFS_BLOCK_GROUP_RAID6) && - root->fs_info->fs_devices->rw_devices <= 3) { - ret = BTRFS_ERROR_DEV_RAID6_MIN_NOT_MET; - goto out; + fs_info->fs_devices->rw_devices <= 3) { + return BTRFS_ERROR_DEV_RAID6_MIN_NOT_MET; } -out: - return ret; + return 0; } int btrfs_rm_device(struct btrfs_root *root, char *device_path) @@ -1769,7 +1764,7 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) mutex_lock(&uuid_mutex); - ret = __check_raid_min_devices(root); + ret = __check_raid_min_devices(root->fs_info); if (ret) goto out; -- cgit v1.3-8-gc7d7 From 24e0474b59538cdb9d2b7318ec7c7ae9f6faf85d Mon Sep 17 00:00:00 2001 From: Anand Jain Date: Sat, 13 Feb 2016 10:01:35 +0800 Subject: btrfs: create helper btrfs_find_device_by_user_input() The patch renames btrfs_dev_replace_find_srcdev() to btrfs_find_device_by_user_input() and moves it to volumes.c, so that delete device can use it. Signed-off-by: Anand Jain Signed-off-by: David Sterba --- fs/btrfs/dev-replace.c | 24 +----------------------- fs/btrfs/volumes.c | 19 +++++++++++++++++++ fs/btrfs/volumes.h | 3 +++ 3 files changed, 23 insertions(+), 23 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/dev-replace.c b/fs/btrfs/dev-replace.c index 26bcb487f958..7ad8ae994ca0 100644 --- a/fs/btrfs/dev-replace.c +++ b/fs/btrfs/dev-replace.c @@ -44,9 +44,6 @@ static void btrfs_dev_replace_update_device_in_mapping_tree( struct btrfs_fs_info *fs_info, struct btrfs_device *srcdev, struct btrfs_device *tgtdev); -static int btrfs_dev_replace_find_srcdev(struct btrfs_root *root, u64 srcdevid, - char *srcdev_name, - struct btrfs_device **device); static u64 __btrfs_dev_replace_cancel(struct btrfs_fs_info *fs_info); static int btrfs_dev_replace_kthread(void *data); static int btrfs_dev_replace_continue_on_mount(struct btrfs_fs_info *fs_info); @@ -329,7 +326,7 @@ int btrfs_dev_replace_start(struct btrfs_root *root, /* the disk copy procedure reuses the scrub code */ mutex_lock(&fs_info->volume_mutex); - ret = btrfs_dev_replace_find_srcdev(root, args->start.srcdevid, + ret = btrfs_find_device_by_user_input(root, args->start.srcdevid, args->start.srcdev_name, &src_device); if (ret) { @@ -626,25 +623,6 @@ static void btrfs_dev_replace_update_device_in_mapping_tree( write_unlock(&em_tree->lock); } -static int btrfs_dev_replace_find_srcdev(struct btrfs_root *root, u64 srcdevid, - char *srcdev_name, - struct btrfs_device **device) -{ - int ret; - - if (srcdevid) { - ret = 0; - *device = btrfs_find_device(root->fs_info, srcdevid, NULL, - NULL); - if (!*device) - ret = -ENOENT; - } else { - ret = btrfs_find_device_missing_or_by_path(root, srcdev_name, - device); - } - return ret; -} - void btrfs_dev_replace_status(struct btrfs_fs_info *fs_info, struct btrfs_ioctl_dev_replace_args *args) { diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 5538dc78c72e..ab28d9dc1b0d 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -2126,6 +2126,25 @@ int btrfs_find_device_missing_or_by_path(struct btrfs_root *root, } } +int btrfs_find_device_by_user_input(struct btrfs_root *root, u64 srcdevid, + char *srcdev_name, + struct btrfs_device **device) +{ + int ret; + + if (srcdevid) { + ret = 0; + *device = btrfs_find_device(root->fs_info, srcdevid, NULL, + NULL); + if (!*device) + ret = -ENOENT; + } else { + ret = btrfs_find_device_missing_or_by_path(root, srcdev_name, + device); + } + return ret; +} + /* * does all the dirty work required for changing file system's UUID. */ diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h index 1939ebde63df..508739314e43 100644 --- a/fs/btrfs/volumes.h +++ b/fs/btrfs/volumes.h @@ -448,6 +448,9 @@ void btrfs_close_extra_devices(struct btrfs_fs_devices *fs_devices, int step); int btrfs_find_device_missing_or_by_path(struct btrfs_root *root, char *device_path, struct btrfs_device **device); +int btrfs_find_device_by_user_input(struct btrfs_root *root, u64 srcdevid, + char *srcdev_name, + struct btrfs_device **device); struct btrfs_device *btrfs_alloc_device(struct btrfs_fs_info *fs_info, const u64 *devid, const u8 *uuid); -- cgit v1.3-8-gc7d7 From 24fc572fe456c02ff4136c07861a3edd4b8de683 Mon Sep 17 00:00:00 2001 From: Anand Jain Date: Sat, 13 Feb 2016 10:01:36 +0800 Subject: btrfs: make use of btrfs_find_device_by_user_input() btrfs_rm_device() has a section of the code which can be replaced btrfs_find_device_by_user_input() Signed-off-by: Anand Jain Signed-off-by: David Sterba --- fs/btrfs/volumes.c | 100 ++++++++++++++++++++--------------------------------- 1 file changed, 37 insertions(+), 63 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index ab28d9dc1b0d..40bbe0a2715b 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -1752,13 +1752,11 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) { struct btrfs_device *device; struct btrfs_device *next_device; - struct block_device *bdev; + struct block_device *bdev = NULL; struct buffer_head *bh = NULL; - struct btrfs_super_block *disk_super; + struct btrfs_super_block *disk_super = NULL; struct btrfs_fs_devices *cur_devices; - u64 devid; u64 num_devices; - u8 *dev_uuid; int ret = 0; bool clear_super = false; @@ -1768,57 +1766,19 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) if (ret) goto out; - if (strcmp(device_path, "missing") == 0) { - struct list_head *devices; - struct btrfs_device *tmp; - - device = NULL; - devices = &root->fs_info->fs_devices->devices; - /* - * It is safe to read the devices since the volume_mutex - * is held. - */ - list_for_each_entry(tmp, devices, dev_list) { - if (tmp->in_fs_metadata && - !tmp->is_tgtdev_for_dev_replace && - !tmp->bdev) { - device = tmp; - break; - } - } - bdev = NULL; - bh = NULL; - disk_super = NULL; - if (!device) { - ret = BTRFS_ERROR_DEV_MISSING_NOT_FOUND; - goto out; - } - } else { - ret = btrfs_get_bdev_and_sb(device_path, - FMODE_WRITE | FMODE_EXCL, - root->fs_info->bdev_holder, 0, - &bdev, &bh); - if (ret) - goto out; - disk_super = (struct btrfs_super_block *)bh->b_data; - devid = btrfs_stack_device_id(&disk_super->dev_item); - dev_uuid = disk_super->dev_item.uuid; - device = btrfs_find_device(root->fs_info, devid, dev_uuid, - disk_super->fsid); - if (!device) { - ret = -ENOENT; - goto error_brelse; - } - } + ret = btrfs_find_device_by_user_input(root, 0, device_path, + &device); + if (ret) + goto out; if (device->is_tgtdev_for_dev_replace) { ret = BTRFS_ERROR_DEV_TGT_REPLACE; - goto error_brelse; + goto out; } if (device->writeable && root->fs_info->fs_devices->rw_devices == 1) { ret = BTRFS_ERROR_DEV_ONLY_WRITABLE; - goto error_brelse; + goto out; } if (device->writeable) { @@ -1908,16 +1868,33 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) * at this point, the device is zero sized. We want to * remove it from the devices list and zero out the old super */ - if (clear_super && disk_super) { + if (clear_super) { u64 bytenr; int i; + if (!disk_super) { + ret = btrfs_get_bdev_and_sb(device_path, + FMODE_WRITE | FMODE_EXCL, + root->fs_info->bdev_holder, 0, + &bdev, &bh); + if (ret) { + /* + * It could be a failed device ok for clear_super + * to fail. So return success + */ + ret = 0; + goto out; + } + + disk_super = (struct btrfs_super_block *)bh->b_data; + } /* make sure this device isn't detected as part of * the FS anymore */ memset(&disk_super->magic, 0, sizeof(disk_super->magic)); set_buffer_dirty(bh); sync_dirty_buffer(bh); + brelse(bh); /* clear the mirror copies of super block on the disk * being removed, 0th copy is been taken care above and @@ -1929,7 +1906,6 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) i_size_read(bdev->bd_inode)) break; - brelse(bh); bh = __bread(bdev, bytenr / 4096, BTRFS_SUPER_INFO_SIZE); if (!bh) @@ -1939,32 +1915,30 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) if (btrfs_super_bytenr(disk_super) != bytenr || btrfs_super_magic(disk_super) != BTRFS_MAGIC) { + brelse(bh); continue; } memset(&disk_super->magic, 0, sizeof(disk_super->magic)); set_buffer_dirty(bh); sync_dirty_buffer(bh); + brelse(bh); } - } - ret = 0; - - if (bdev) { - /* Notify udev that device has changed */ - btrfs_kobject_uevent(bdev, KOBJ_CHANGE); + if (bdev) { + /* Notify udev that device has changed */ + btrfs_kobject_uevent(bdev, KOBJ_CHANGE); - /* Update ctime/mtime for device path for libblkid */ - update_dev_time(device_path); + /* Update ctime/mtime for device path for libblkid */ + update_dev_time(device_path); + blkdev_put(bdev, FMODE_READ | FMODE_EXCL); + } } -error_brelse: - brelse(bh); - if (bdev) - blkdev_put(bdev, FMODE_READ | FMODE_EXCL); out: mutex_unlock(&uuid_mutex); return ret; + error_undo: if (device->writeable) { lock_chunks(root); @@ -1973,7 +1947,7 @@ error_undo: device->fs_devices->rw_devices++; unlock_chunks(root); } - goto error_brelse; + goto out; } void btrfs_rm_dev_replace_remove_srcdev(struct btrfs_fs_info *fs_info, -- cgit v1.3-8-gc7d7 From b3d1b1532ff9620ff5dba891a96f3e912005eb10 Mon Sep 17 00:00:00 2001 From: Anand Jain Date: Sat, 13 Feb 2016 10:01:37 +0800 Subject: btrfs: enhance btrfs_find_device_by_user_input() to check device path The operation of device replace and device delete follows same steps upto some depth with in btrfs kernel, however they don't share codes. This enhancement will help replace and delete to share codes. Signed-off-by: Anand Jain Signed-off-by: David Sterba --- fs/btrfs/dev-replace.c | 4 ---- fs/btrfs/volumes.c | 3 +++ 2 files changed, 3 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/dev-replace.c b/fs/btrfs/dev-replace.c index 7ad8ae994ca0..fdd2880707df 100644 --- a/fs/btrfs/dev-replace.c +++ b/fs/btrfs/dev-replace.c @@ -320,10 +320,6 @@ int btrfs_dev_replace_start(struct btrfs_root *root, return -EINVAL; } - if ((args->start.srcdevid == 0 && args->start.srcdev_name[0] == '\0') || - args->start.tgtdev_name[0] == '\0') - return -EINVAL; - /* the disk copy procedure reuses the scrub code */ mutex_lock(&fs_info->volume_mutex); ret = btrfs_find_device_by_user_input(root, args->start.srcdevid, diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 40bbe0a2715b..d74260567bea 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -2113,6 +2113,9 @@ int btrfs_find_device_by_user_input(struct btrfs_root *root, u64 srcdevid, if (!*device) ret = -ENOENT; } else { + if (!srcdev_name || !srcdev_name[0]) + return -EINVAL; + ret = btrfs_find_device_missing_or_by_path(root, srcdev_name, device); } -- cgit v1.3-8-gc7d7 From 42b674271566c26692c32f4ec1f4b4cf14d5046f Mon Sep 17 00:00:00 2001 From: Anand Jain Date: Sat, 13 Feb 2016 10:01:38 +0800 Subject: btrfs: make use of btrfs_scratch_superblocks() in btrfs_rm_device() With the previous patches now the btrfs_scratch_superblocks() is ready to be used in btrfs_rm_device() so use it. Signed-off-by: Anand Jain [ use GFP_KERNEL ] Signed-off-by: David Sterba --- fs/btrfs/volumes.c | 77 +++++++++--------------------------------------------- 1 file changed, 13 insertions(+), 64 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index d74260567bea..8264b06756e6 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -1752,13 +1752,11 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) { struct btrfs_device *device; struct btrfs_device *next_device; - struct block_device *bdev = NULL; - struct buffer_head *bh = NULL; - struct btrfs_super_block *disk_super = NULL; struct btrfs_fs_devices *cur_devices; u64 num_devices; int ret = 0; bool clear_super = false; + char *dev_name = NULL; mutex_lock(&uuid_mutex); @@ -1786,6 +1784,11 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) list_del_init(&device->dev_alloc_list); device->fs_devices->rw_devices--; unlock_chunks(root); + dev_name = kstrdup(device->name->str, GFP_KERNEL); + if (!dev_name) { + ret = -ENOMEM; + goto error_undo; + } clear_super = true; } @@ -1869,73 +1872,19 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) * remove it from the devices list and zero out the old super */ if (clear_super) { - u64 bytenr; - int i; - - if (!disk_super) { - ret = btrfs_get_bdev_and_sb(device_path, - FMODE_WRITE | FMODE_EXCL, - root->fs_info->bdev_holder, 0, - &bdev, &bh); - if (ret) { - /* - * It could be a failed device ok for clear_super - * to fail. So return success - */ - ret = 0; - goto out; - } - - disk_super = (struct btrfs_super_block *)bh->b_data; - } - /* make sure this device isn't detected as part of - * the FS anymore - */ - memset(&disk_super->magic, 0, sizeof(disk_super->magic)); - set_buffer_dirty(bh); - sync_dirty_buffer(bh); - brelse(bh); - - /* clear the mirror copies of super block on the disk - * being removed, 0th copy is been taken care above and - * the below would take of the rest - */ - for (i = 1; i < BTRFS_SUPER_MIRROR_MAX; i++) { - bytenr = btrfs_sb_offset(i); - if (bytenr + BTRFS_SUPER_INFO_SIZE >= - i_size_read(bdev->bd_inode)) - break; - - bh = __bread(bdev, bytenr / 4096, - BTRFS_SUPER_INFO_SIZE); - if (!bh) - continue; + struct block_device *bdev; - disk_super = (struct btrfs_super_block *)bh->b_data; - - if (btrfs_super_bytenr(disk_super) != bytenr || - btrfs_super_magic(disk_super) != BTRFS_MAGIC) { - brelse(bh); - continue; - } - memset(&disk_super->magic, 0, - sizeof(disk_super->magic)); - set_buffer_dirty(bh); - sync_dirty_buffer(bh); - brelse(bh); - } - - if (bdev) { - /* Notify udev that device has changed */ - btrfs_kobject_uevent(bdev, KOBJ_CHANGE); - - /* Update ctime/mtime for device path for libblkid */ - update_dev_time(device_path); + bdev = blkdev_get_by_path(dev_name, FMODE_READ | FMODE_EXCL, + root->fs_info->bdev_holder); + if (!IS_ERR(bdev)) { + btrfs_scratch_superblocks(bdev, dev_name); blkdev_put(bdev, FMODE_READ | FMODE_EXCL); } } out: + kfree(dev_name); + mutex_unlock(&uuid_mutex); return ret; -- cgit v1.3-8-gc7d7 From 6b526ed70cf189660d009ea6f17af77a9cca0f38 Mon Sep 17 00:00:00 2001 From: Anand Jain Date: Sat, 13 Feb 2016 10:01:39 +0800 Subject: btrfs: introduce device delete by devid This introduces new ioctl BTRFS_IOC_RM_DEV_V2, which uses enhanced struct btrfs_ioctl_vol_args_v2 to carry devid as an user argument. The patch won't delete the old ioctl interface and so kernel remains backward compatible with user land progs. Test case/script: echo "0 $(blockdev --getsz /dev/sdf) linear /dev/sdf 0" | dmsetup create bad_disk mkfs.btrfs -f -d raid1 -m raid1 /dev/sdd /dev/sde /dev/mapper/bad_disk mount /dev/sdd /btrfs dmsetup suspend bad_disk echo "0 $(blockdev --getsz /dev/sdf) error /dev/sdf 0" | dmsetup load bad_disk dmsetup resume bad_disk echo "bad disk failed. now deleting/replacing" btrfs dev del 3 /btrfs echo $? btrfs fi show /btrfs umount /btrfs btrfs-show-super /dev/sdd | egrep num_device dmsetup remove bad_disk wipefs -a /dev/sdf Signed-off-by: Anand Jain Reported-by: Martin [ adjust messages, s/disk/device/ ] Signed-off-by: David Sterba --- fs/btrfs/ioctl.c | 58 +++++++++++++++++++++++++++++++++++++++++++++- fs/btrfs/volumes.c | 4 ++-- fs/btrfs/volumes.h | 2 +- include/uapi/linux/btrfs.h | 14 ++++++++++- 4 files changed, 73 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index 5a23806ae418..7cf0f6328d99 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -2667,6 +2667,60 @@ out: return ret; } +static long btrfs_ioctl_rm_dev_v2(struct file *file, void __user *arg) +{ + struct btrfs_root *root = BTRFS_I(file_inode(file))->root; + struct btrfs_ioctl_vol_args_v2 *vol_args; + int ret; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + ret = mnt_want_write_file(file); + if (ret) + return ret; + + vol_args = memdup_user(arg, sizeof(*vol_args)); + if (IS_ERR(vol_args)) { + ret = PTR_ERR(vol_args); + goto err_drop; + } + + /* Check for compatibility reject unknown flags */ + if (vol_args->flags & ~BTRFS_VOL_ARG_V2_FLAGS) + return -ENOTTY; + + if (atomic_xchg(&root->fs_info->mutually_exclusive_operation_running, + 1)) { + ret = BTRFS_ERROR_DEV_EXCL_RUN_IN_PROGRESS; + goto out; + } + + mutex_lock(&root->fs_info->volume_mutex); + if (vol_args->flags & BTRFS_DEVICE_BY_ID) { + ret = btrfs_rm_device(root, NULL, vol_args->devid); + } else { + vol_args->name[BTRFS_SUBVOL_NAME_MAX] = '\0'; + ret = btrfs_rm_device(root, vol_args->name, 0); + } + mutex_unlock(&root->fs_info->volume_mutex); + atomic_set(&root->fs_info->mutually_exclusive_operation_running, 0); + + if (!ret) { + if (vol_args->flags & BTRFS_DEVICE_BY_ID) + btrfs_info(root->fs_info, "device deleted: id %llu", + vol_args->devid); + else + btrfs_info(root->fs_info, "device deleted: %s", + vol_args->name); + } +out: + kfree(vol_args); +err_drop: + mnt_drop_write_file(file); + return ret; +} + static long btrfs_ioctl_rm_dev(struct file *file, void __user *arg) { struct btrfs_root *root = BTRFS_I(file_inode(file))->root; @@ -2695,7 +2749,7 @@ static long btrfs_ioctl_rm_dev(struct file *file, void __user *arg) } mutex_lock(&root->fs_info->volume_mutex); - ret = btrfs_rm_device(root, vol_args->name); + ret = btrfs_rm_device(root, vol_args->name, 0); mutex_unlock(&root->fs_info->volume_mutex); atomic_set(&root->fs_info->mutually_exclusive_operation_running, 0); @@ -5459,6 +5513,8 @@ long btrfs_ioctl(struct file *file, unsigned int return btrfs_ioctl_add_dev(root, argp); case BTRFS_IOC_RM_DEV: return btrfs_ioctl_rm_dev(file, argp); + case BTRFS_IOC_RM_DEV_V2: + return btrfs_ioctl_rm_dev_v2(file, argp); case BTRFS_IOC_FS_INFO: return btrfs_ioctl_fs_info(root, argp); case BTRFS_IOC_DEV_INFO: diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 8264b06756e6..1421d711629f 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -1748,7 +1748,7 @@ static int __check_raid_min_devices(struct btrfs_fs_info *fs_info) return 0; } -int btrfs_rm_device(struct btrfs_root *root, char *device_path) +int btrfs_rm_device(struct btrfs_root *root, char *device_path, u64 devid) { struct btrfs_device *device; struct btrfs_device *next_device; @@ -1764,7 +1764,7 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) if (ret) goto out; - ret = btrfs_find_device_by_user_input(root, 0, device_path, + ret = btrfs_find_device_by_user_input(root, devid, device_path, &device); if (ret) goto out; diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h index 508739314e43..c73d027e2f8b 100644 --- a/fs/btrfs/volumes.h +++ b/fs/btrfs/volumes.h @@ -454,7 +454,7 @@ int btrfs_find_device_by_user_input(struct btrfs_root *root, u64 srcdevid, struct btrfs_device *btrfs_alloc_device(struct btrfs_fs_info *fs_info, const u64 *devid, const u8 *uuid); -int btrfs_rm_device(struct btrfs_root *root, char *device_path); +int btrfs_rm_device(struct btrfs_root *root, char *device_path, u64 devid); void btrfs_cleanup_fs_uuids(void); int btrfs_num_copies(struct btrfs_fs_info *fs_info, u64 logical, u64 len); int btrfs_grow_device(struct btrfs_trans_handle *trans, diff --git a/include/uapi/linux/btrfs.h b/include/uapi/linux/btrfs.h index dea893199257..396a4efca775 100644 --- a/include/uapi/linux/btrfs.h +++ b/include/uapi/linux/btrfs.h @@ -36,6 +36,13 @@ struct btrfs_ioctl_vol_args { #define BTRFS_SUBVOL_CREATE_ASYNC (1ULL << 0) #define BTRFS_SUBVOL_RDONLY (1ULL << 1) #define BTRFS_SUBVOL_QGROUP_INHERIT (1ULL << 2) +#define BTRFS_DEVICE_BY_ID (1ULL << 3) +#define BTRFS_VOL_ARG_V2_FLAGS \ + (BTRFS_SUBVOL_CREATE_ASYNC | \ + BTRFS_SUBVOL_RDONLY | \ + BTRFS_SUBVOL_QGROUP_INHERIT | \ + BTRFS_DEVICE_BY_ID) + #define BTRFS_FSID_SIZE 16 #define BTRFS_UUID_SIZE 16 #define BTRFS_UUID_UNPARSED_SIZE 37 @@ -76,7 +83,10 @@ struct btrfs_ioctl_vol_args_v2 { }; __u64 unused[4]; }; - char name[BTRFS_SUBVOL_NAME_MAX + 1]; + union { + char name[BTRFS_SUBVOL_NAME_MAX + 1]; + u64 devid; + }; }; /* @@ -659,5 +669,7 @@ static inline char *btrfs_err_str(enum btrfs_err_code err_code) struct btrfs_ioctl_feature_flags[2]) #define BTRFS_IOC_GET_SUPPORTED_FEATURES _IOR(BTRFS_IOCTL_MAGIC, 57, \ struct btrfs_ioctl_feature_flags[3]) +#define BTRFS_IOC_RM_DEV_V2 _IOW(BTRFS_IOCTL_MAGIC, 58, \ + struct btrfs_ioctl_vol_args_v2) #endif /* _UAPI_LINUX_BTRFS_H */ -- cgit v1.3-8-gc7d7 From 02feae3c5525771878461b90edd2ba38fd3f5359 Mon Sep 17 00:00:00 2001 From: Anand Jain Date: Sat, 13 Feb 2016 10:01:40 +0800 Subject: btrfs: optimize check for stale device Optimize check for stale device to only be checked when there is device added or changed. If there is no update to the device, there is no need to call btrfs_free_stale_device(). Signed-off-by: Anand Jain Signed-off-by: David Sterba --- fs/btrfs/volumes.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 1421d711629f..2fdd2d3322aa 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -699,7 +699,8 @@ static noinline int device_list_add(const char *path, * if there is new btrfs on an already registered device, * then remove the stale device entry. */ - btrfs_free_stale_device(device); + if (ret > 0) + btrfs_free_stale_device(device); *fs_devices_ret = fs_devices; -- cgit v1.3-8-gc7d7 From f47ab2588e424cb4898b75ace1e2323ddd18b990 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Mon, 15 Feb 2016 15:28:48 +0100 Subject: btrfs: rename __check_raid_min_devices Underscores are for special functions, use the full prefix for better stacktrace recognition. Reviewed-by: Anand Jain Signed-off-by: David Sterba --- fs/btrfs/volumes.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 2fdd2d3322aa..9051b4472bea 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -1706,7 +1706,7 @@ out: return ret; } -static int __check_raid_min_devices(struct btrfs_fs_info *fs_info) +static int btrfs_check_raid_min_devices(struct btrfs_fs_info *fs_info) { u64 all_avail; u64 num_devices; @@ -1761,7 +1761,7 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path, u64 devid) mutex_lock(&uuid_mutex); - ret = __check_raid_min_devices(root->fs_info); + ret = btrfs_check_raid_min_devices(root->fs_info); if (ret) goto out; -- cgit v1.3-8-gc7d7 From 3cc31a0d5bc77e35229671307eba5ea5c4c4cb9d Mon Sep 17 00:00:00 2001 From: David Sterba Date: Mon, 15 Feb 2016 16:00:26 +0100 Subject: btrfs: pass number of devices to btrfs_check_raid_min_devices Before this patch, btrfs_check_raid_min_devices would do an off-by-one check of the constraints and not the miminmum check, as its name suggests. This is not a problem if the only caller is device remove, but would be confusing for others. Add an argument with the exact number and let the caller(s) decide if this needs any adjustments, like when device replace is running. Reviewed-by: Anand Jain Tested-by: Anand Jain Signed-off-by: David Sterba --- fs/btrfs/volumes.c | 35 ++++++++++++++++++++--------------- 1 file changed, 20 insertions(+), 15 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 9051b4472bea..e966f0ce07f8 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -1706,20 +1706,17 @@ out: return ret; } -static int btrfs_check_raid_min_devices(struct btrfs_fs_info *fs_info) +/* + * Verify that @num_devices satisfies the RAID profile constraints in the whole + * filesystem. It's up to the caller to adjust that number regarding eg. device + * replace. + */ +static int btrfs_check_raid_min_devices(struct btrfs_fs_info *fs_info, + u64 num_devices) { u64 all_avail; - u64 num_devices; unsigned seq; - num_devices = fs_info->fs_devices->num_devices; - btrfs_dev_replace_lock(&fs_info->dev_replace, 0); - if (btrfs_dev_replace_is_ongoing(&fs_info->dev_replace)) { - WARN_ON(num_devices < 1); - num_devices--; - } - btrfs_dev_replace_unlock(&fs_info->dev_replace, 0); - do { seq = read_seqbegin(&fs_info->profiles_lock); @@ -1728,21 +1725,21 @@ static int btrfs_check_raid_min_devices(struct btrfs_fs_info *fs_info) fs_info->avail_metadata_alloc_bits; } while (read_seqretry(&fs_info->profiles_lock, seq)); - if ((all_avail & BTRFS_BLOCK_GROUP_RAID10) && num_devices <= 4) { + if ((all_avail & BTRFS_BLOCK_GROUP_RAID10) && num_devices < 4) { return BTRFS_ERROR_DEV_RAID10_MIN_NOT_MET; } - if ((all_avail & BTRFS_BLOCK_GROUP_RAID1) && num_devices <= 2) { + if ((all_avail & BTRFS_BLOCK_GROUP_RAID1) && num_devices < 2) { return BTRFS_ERROR_DEV_RAID1_MIN_NOT_MET; } if ((all_avail & BTRFS_BLOCK_GROUP_RAID5) && - fs_info->fs_devices->rw_devices <= 2) { + fs_info->fs_devices->rw_devices < 2) { return BTRFS_ERROR_DEV_RAID5_MIN_NOT_MET; } if ((all_avail & BTRFS_BLOCK_GROUP_RAID6) && - fs_info->fs_devices->rw_devices <= 3) { + fs_info->fs_devices->rw_devices < 3) { return BTRFS_ERROR_DEV_RAID6_MIN_NOT_MET; } @@ -1761,7 +1758,15 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path, u64 devid) mutex_lock(&uuid_mutex); - ret = btrfs_check_raid_min_devices(root->fs_info); + num_devices = root->fs_info->fs_devices->num_devices; + btrfs_dev_replace_lock(&root->fs_info->dev_replace, 0); + if (btrfs_dev_replace_is_ongoing(&root->fs_info->dev_replace)) { + WARN_ON(num_devices < 1); + num_devices--; + } + btrfs_dev_replace_unlock(&root->fs_info->dev_replace, 0); + + ret = btrfs_check_raid_min_devices(root->fs_info, num_devices - 1); if (ret) goto out; -- cgit v1.3-8-gc7d7 From 621292bae6ae3b25cd3124b63603d01df4cccfb6 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Mon, 15 Feb 2016 16:28:03 +0100 Subject: btrfs: introduce raid-type to error-code table, for minimum device constraint Reviewed-by: Anand Jain Signed-off-by: David Sterba --- fs/btrfs/volumes.c | 15 +++++++++++++++ fs/btrfs/volumes.h | 2 +- 2 files changed, 16 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index e966f0ce07f8..6b38b72d678b 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -118,6 +118,21 @@ const u64 btrfs_raid_group[BTRFS_NR_RAID_TYPES] = { [BTRFS_RAID_RAID6] = BTRFS_BLOCK_GROUP_RAID6, }; +/* + * Table to convert BTRFS_RAID_* to the error code if minimum number of devices + * condition is not met. Zero means there's no corresponding + * BTRFS_ERROR_DEV_*_NOT_MET value. + */ +const int btrfs_raid_mindev_error[BTRFS_NR_RAID_TYPES] = { + [BTRFS_RAID_RAID10] = BTRFS_ERROR_DEV_RAID10_MIN_NOT_MET, + [BTRFS_RAID_RAID1] = BTRFS_ERROR_DEV_RAID1_MIN_NOT_MET, + [BTRFS_RAID_DUP] = 0, + [BTRFS_RAID_RAID0] = 0, + [BTRFS_RAID_SINGLE] = 0, + [BTRFS_RAID_RAID5] = BTRFS_ERROR_DEV_RAID5_MIN_NOT_MET, + [BTRFS_RAID_RAID6] = BTRFS_ERROR_DEV_RAID6_MIN_NOT_MET, +}; + static int init_first_rw_device(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct btrfs_device *device); diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h index c73d027e2f8b..a13a538cb01e 100644 --- a/fs/btrfs/volumes.h +++ b/fs/btrfs/volumes.h @@ -340,7 +340,7 @@ struct btrfs_raid_attr { }; extern const struct btrfs_raid_attr btrfs_raid_array[BTRFS_NR_RAID_TYPES]; - +extern const int btrfs_raid_mindev_error[BTRFS_NR_RAID_TYPES]; extern const u64 btrfs_raid_group[BTRFS_NR_RAID_TYPES]; struct map_lookup { -- cgit v1.3-8-gc7d7 From 418775a22b4c67bd15915e043c3a8f29816799bd Mon Sep 17 00:00:00 2001 From: David Sterba Date: Mon, 15 Feb 2016 16:28:14 +0100 Subject: btrfs: use existing device constraints table btrfs_raid_array We should avoid duplicating the device constraints, let's use the btrfs_raid_array in btrfs_check_raid_min_devices. Reviewed-by: Anand Jain Signed-off-by: David Sterba --- fs/btrfs/volumes.c | 23 +++++++++-------------- 1 file changed, 9 insertions(+), 14 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 6b38b72d678b..055fbebea199 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -1731,6 +1731,7 @@ static int btrfs_check_raid_min_devices(struct btrfs_fs_info *fs_info, { u64 all_avail; unsigned seq; + int i; do { seq = read_seqbegin(&fs_info->profiles_lock); @@ -1740,22 +1741,16 @@ static int btrfs_check_raid_min_devices(struct btrfs_fs_info *fs_info, fs_info->avail_metadata_alloc_bits; } while (read_seqretry(&fs_info->profiles_lock, seq)); - if ((all_avail & BTRFS_BLOCK_GROUP_RAID10) && num_devices < 4) { - return BTRFS_ERROR_DEV_RAID10_MIN_NOT_MET; - } - - if ((all_avail & BTRFS_BLOCK_GROUP_RAID1) && num_devices < 2) { - return BTRFS_ERROR_DEV_RAID1_MIN_NOT_MET; - } + for (i = 0; i < BTRFS_NR_RAID_TYPES; i++) { + if (!(all_avail & btrfs_raid_group[i])) + continue; - if ((all_avail & BTRFS_BLOCK_GROUP_RAID5) && - fs_info->fs_devices->rw_devices < 2) { - return BTRFS_ERROR_DEV_RAID5_MIN_NOT_MET; - } + if (num_devices < btrfs_raid_array[i].devs_min) { + int ret = btrfs_raid_mindev_error[i]; - if ((all_avail & BTRFS_BLOCK_GROUP_RAID6) && - fs_info->fs_devices->rw_devices < 3) { - return BTRFS_ERROR_DEV_RAID6_MIN_NOT_MET; + if (ret) + return ret; + } } return 0; -- cgit v1.3-8-gc7d7 From 5c5c0df05deaebcdcc9bb31bdca3812a7c22230f Mon Sep 17 00:00:00 2001 From: David Sterba Date: Mon, 15 Feb 2016 16:39:55 +0100 Subject: btrfs: rename btrfs_find_device_by_user_input For clarity how we are going to find the device, let's call it a device specifier, devspec for short. Also rename the arguments that are a leftover from previous function purpose. Reviewed-by: Anand Jain Signed-off-by: David Sterba --- fs/btrfs/dev-replace.c | 2 +- fs/btrfs/volumes.c | 17 ++++++++++------- fs/btrfs/volumes.h | 4 ++-- 3 files changed, 13 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/dev-replace.c b/fs/btrfs/dev-replace.c index fdd2880707df..b7f5f4aa6e86 100644 --- a/fs/btrfs/dev-replace.c +++ b/fs/btrfs/dev-replace.c @@ -322,7 +322,7 @@ int btrfs_dev_replace_start(struct btrfs_root *root, /* the disk copy procedure reuses the scrub code */ mutex_lock(&fs_info->volume_mutex); - ret = btrfs_find_device_by_user_input(root, args->start.srcdevid, + ret = btrfs_find_device_by_devspec(root, args->start.srcdevid, args->start.srcdev_name, &src_device); if (ret) { diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 055fbebea199..e43b01e438b1 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -1780,7 +1780,7 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path, u64 devid) if (ret) goto out; - ret = btrfs_find_device_by_user_input(root, devid, device_path, + ret = btrfs_find_device_by_devspec(root, devid, device_path, &device); if (ret) goto out; @@ -2065,23 +2065,26 @@ int btrfs_find_device_missing_or_by_path(struct btrfs_root *root, } } -int btrfs_find_device_by_user_input(struct btrfs_root *root, u64 srcdevid, - char *srcdev_name, +/* + * Lookup a device given by device id, or the path if the id is 0. + */ +int btrfs_find_device_by_devspec(struct btrfs_root *root, u64 devid, + char *devpath, struct btrfs_device **device) { int ret; - if (srcdevid) { + if (devid) { ret = 0; - *device = btrfs_find_device(root->fs_info, srcdevid, NULL, + *device = btrfs_find_device(root->fs_info, devid, NULL, NULL); if (!*device) ret = -ENOENT; } else { - if (!srcdev_name || !srcdev_name[0]) + if (!devpath || !devpath[0]) return -EINVAL; - ret = btrfs_find_device_missing_or_by_path(root, srcdev_name, + ret = btrfs_find_device_missing_or_by_path(root, devpath, device); } return ret; diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h index a13a538cb01e..febdb7bc9370 100644 --- a/fs/btrfs/volumes.h +++ b/fs/btrfs/volumes.h @@ -448,8 +448,8 @@ void btrfs_close_extra_devices(struct btrfs_fs_devices *fs_devices, int step); int btrfs_find_device_missing_or_by_path(struct btrfs_root *root, char *device_path, struct btrfs_device **device); -int btrfs_find_device_by_user_input(struct btrfs_root *root, u64 srcdevid, - char *srcdev_name, +int btrfs_find_device_by_devspec(struct btrfs_root *root, u64 devid, + char *devpath, struct btrfs_device **device); struct btrfs_device *btrfs_alloc_device(struct btrfs_fs_info *fs_info, const u64 *devid, -- cgit v1.3-8-gc7d7 From 735654ea91a06a30bfe05fdfd09c8895abf6c1bf Mon Sep 17 00:00:00 2001 From: David Sterba Date: Mon, 15 Feb 2016 18:15:21 +0100 Subject: btrfs: rename flags for vol args v2 Rename BTRFS_DEVICE_BY_ID so it's more descriptive that we specify the device by id, it'll be part of the public API. The mask of supported flags is also renamed, only for internal use. The error code for unknown flags is EOPNOTSUPP, fixed. Reviewed-by: Anand Jain Signed-off-by: David Sterba --- fs/btrfs/ioctl.c | 8 ++++---- include/uapi/linux/btrfs.h | 7 ++++--- 2 files changed, 8 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index 7cf0f6328d99..65903ec5e1c0 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -2687,8 +2687,8 @@ static long btrfs_ioctl_rm_dev_v2(struct file *file, void __user *arg) } /* Check for compatibility reject unknown flags */ - if (vol_args->flags & ~BTRFS_VOL_ARG_V2_FLAGS) - return -ENOTTY; + if (vol_args->flags & ~BTRFS_VOL_ARG_V2_FLAGS_SUPPORTED) + return -EOPNOTSUPP; if (atomic_xchg(&root->fs_info->mutually_exclusive_operation_running, 1)) { @@ -2697,7 +2697,7 @@ static long btrfs_ioctl_rm_dev_v2(struct file *file, void __user *arg) } mutex_lock(&root->fs_info->volume_mutex); - if (vol_args->flags & BTRFS_DEVICE_BY_ID) { + if (vol_args->flags & BTRFS_DEVICE_SPEC_BY_ID) { ret = btrfs_rm_device(root, NULL, vol_args->devid); } else { vol_args->name[BTRFS_SUBVOL_NAME_MAX] = '\0'; @@ -2707,7 +2707,7 @@ static long btrfs_ioctl_rm_dev_v2(struct file *file, void __user *arg) atomic_set(&root->fs_info->mutually_exclusive_operation_running, 0); if (!ret) { - if (vol_args->flags & BTRFS_DEVICE_BY_ID) + if (vol_args->flags & BTRFS_DEVICE_SPEC_BY_ID) btrfs_info(root->fs_info, "device deleted: id %llu", vol_args->devid); else diff --git a/include/uapi/linux/btrfs.h b/include/uapi/linux/btrfs.h index 396a4efca775..3975e683af72 100644 --- a/include/uapi/linux/btrfs.h +++ b/include/uapi/linux/btrfs.h @@ -36,12 +36,13 @@ struct btrfs_ioctl_vol_args { #define BTRFS_SUBVOL_CREATE_ASYNC (1ULL << 0) #define BTRFS_SUBVOL_RDONLY (1ULL << 1) #define BTRFS_SUBVOL_QGROUP_INHERIT (1ULL << 2) -#define BTRFS_DEVICE_BY_ID (1ULL << 3) -#define BTRFS_VOL_ARG_V2_FLAGS \ +#define BTRFS_DEVICE_SPEC_BY_ID (1ULL << 3) + +#define BTRFS_VOL_ARG_V2_FLAGS_SUPPORTED \ (BTRFS_SUBVOL_CREATE_ASYNC | \ BTRFS_SUBVOL_RDONLY | \ BTRFS_SUBVOL_QGROUP_INHERIT | \ - BTRFS_DEVICE_BY_ID) + BTRFS_DEVICE_SPEC_BY_ID) #define BTRFS_FSID_SIZE 16 #define BTRFS_UUID_SIZE 16 -- cgit v1.3-8-gc7d7 From fc23c246d72d21385be115305d1cb85fcc34acad Mon Sep 17 00:00:00 2001 From: Anand Jain Date: Thu, 24 Mar 2016 18:48:12 +0800 Subject: btrfs: use fs_info directly Local variable fs_info, contains root->fs_info, use it. Signed-off-by: Anand Jain Signed-off-by: David Sterba --- fs/btrfs/dev-replace.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/dev-replace.c b/fs/btrfs/dev-replace.c index b7f5f4aa6e86..1a3d24592859 100644 --- a/fs/btrfs/dev-replace.c +++ b/fs/btrfs/dev-replace.c @@ -368,7 +368,7 @@ int btrfs_dev_replace_start(struct btrfs_root *root, WARN_ON(!tgt_device); dev_replace->tgtdev = tgt_device; - btrfs_info_in_rcu(root->fs_info, + btrfs_info_in_rcu(fs_info, "dev_replace from %s (devid %llu) to %s started", src_device->missing ? "" : rcu_str_deref(src_device->name), @@ -394,9 +394,9 @@ int btrfs_dev_replace_start(struct btrfs_root *root, ret = btrfs_sysfs_add_device_link(tgt_device->fs_devices, tgt_device); if (ret) - btrfs_err(root->fs_info, "kobj add dev failed %d\n", ret); + btrfs_err(fs_info, "kobj add dev failed %d\n", ret); - btrfs_wait_ordered_roots(root->fs_info, -1); + btrfs_wait_ordered_roots(fs_info, -1); /* force writing the updated state information to disk */ trans = btrfs_start_transaction(root, 0); @@ -414,7 +414,7 @@ int btrfs_dev_replace_start(struct btrfs_root *root, btrfs_device_get_total_bytes(src_device), &dev_replace->scrub_progress, 0, 1); - ret = btrfs_dev_replace_finishing(root->fs_info, ret); + ret = btrfs_dev_replace_finishing(fs_info, ret); /* don't warn if EINPROGRESS, someone else might be running scrub */ if (ret == -EINPROGRESS) { args->result = BTRFS_IOCTL_DEV_REPLACE_RESULT_SCRUB_INPROGRESS; -- cgit v1.3-8-gc7d7 From b5255456c529155730c837f8cfcea47e8feb85ca Mon Sep 17 00:00:00 2001 From: Anand Jain Date: Thu, 24 Mar 2016 18:48:14 +0800 Subject: btrfs: refactor btrfs_dev_replace_start for reuse A refactor patch, and avoids user input verification in the btrfs_dev_replace_start(), and so this function can be reused. Signed-off-by: Anand Jain Signed-off-by: David Sterba --- fs/btrfs/dev-replace.c | 58 ++++++++++++++++++++++++++++++++------------------ fs/btrfs/dev-replace.h | 4 +++- fs/btrfs/ioctl.c | 2 +- 3 files changed, 41 insertions(+), 23 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/dev-replace.c b/fs/btrfs/dev-replace.c index 1a3d24592859..5aebedf12b5f 100644 --- a/fs/btrfs/dev-replace.c +++ b/fs/btrfs/dev-replace.c @@ -302,8 +302,8 @@ void btrfs_after_dev_replace_commit(struct btrfs_fs_info *fs_info) dev_replace->cursor_left_last_write_of_item; } -int btrfs_dev_replace_start(struct btrfs_root *root, - struct btrfs_ioctl_dev_replace_args *args) +int btrfs_dev_replace_start(struct btrfs_root *root, char *tgtdev_name, + u64 srcdevid, char *srcdev_name, int read_src) { struct btrfs_trans_handle *trans; struct btrfs_fs_info *fs_info = root->fs_info; @@ -312,25 +312,16 @@ int btrfs_dev_replace_start(struct btrfs_root *root, struct btrfs_device *tgt_device = NULL; struct btrfs_device *src_device = NULL; - switch (args->start.cont_reading_from_srcdev_mode) { - case BTRFS_IOCTL_DEV_REPLACE_CONT_READING_FROM_SRCDEV_MODE_ALWAYS: - case BTRFS_IOCTL_DEV_REPLACE_CONT_READING_FROM_SRCDEV_MODE_AVOID: - break; - default: - return -EINVAL; - } - /* the disk copy procedure reuses the scrub code */ mutex_lock(&fs_info->volume_mutex); - ret = btrfs_find_device_by_devspec(root, args->start.srcdevid, - args->start.srcdev_name, - &src_device); + ret = btrfs_find_device_by_devspec(root, srcdevid, + srcdev_name, &src_device); if (ret) { mutex_unlock(&fs_info->volume_mutex); return ret; } - ret = btrfs_init_dev_replace_tgtdev(root, args->start.tgtdev_name, + ret = btrfs_init_dev_replace_tgtdev(root, tgtdev_name, src_device, &tgt_device); mutex_unlock(&fs_info->volume_mutex); if (ret) @@ -357,12 +348,11 @@ int btrfs_dev_replace_start(struct btrfs_root *root, break; case BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED: case BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED: - args->result = BTRFS_IOCTL_DEV_REPLACE_RESULT_ALREADY_STARTED; + ret = BTRFS_IOCTL_DEV_REPLACE_RESULT_ALREADY_STARTED; goto leave; } - dev_replace->cont_reading_from_srcdev_mode = - args->start.cont_reading_from_srcdev_mode; + dev_replace->cont_reading_from_srcdev_mode = read_src; WARN_ON(!src_device); dev_replace->srcdev = src_device; WARN_ON(!tgt_device); @@ -389,7 +379,6 @@ int btrfs_dev_replace_start(struct btrfs_root *root, dev_replace->item_needs_writeback = 1; atomic64_set(&dev_replace->num_write_errors, 0); atomic64_set(&dev_replace->num_uncorrectable_read_errors, 0); - args->result = BTRFS_IOCTL_DEV_REPLACE_RESULT_NO_ERROR; btrfs_dev_replace_unlock(dev_replace, 1); ret = btrfs_sysfs_add_device_link(tgt_device->fs_devices, tgt_device); @@ -415,10 +404,8 @@ int btrfs_dev_replace_start(struct btrfs_root *root, &dev_replace->scrub_progress, 0, 1); ret = btrfs_dev_replace_finishing(fs_info, ret); - /* don't warn if EINPROGRESS, someone else might be running scrub */ if (ret == -EINPROGRESS) { - args->result = BTRFS_IOCTL_DEV_REPLACE_RESULT_SCRUB_INPROGRESS; - ret = 0; + ret = BTRFS_IOCTL_DEV_REPLACE_RESULT_SCRUB_INPROGRESS; } else { WARN_ON(ret); } @@ -433,6 +420,35 @@ leave: return ret; } +int btrfs_dev_replace_by_ioctl(struct btrfs_root *root, + struct btrfs_ioctl_dev_replace_args *args) +{ + int ret; + + switch (args->start.cont_reading_from_srcdev_mode) { + case BTRFS_IOCTL_DEV_REPLACE_CONT_READING_FROM_SRCDEV_MODE_ALWAYS: + case BTRFS_IOCTL_DEV_REPLACE_CONT_READING_FROM_SRCDEV_MODE_AVOID: + break; + default: + return -EINVAL; + } + + if ((args->start.srcdevid == 0 && args->start.srcdev_name[0] == '\0') || + args->start.tgtdev_name[0] == '\0') + return -EINVAL; + + ret = btrfs_dev_replace_start(root, args->start.tgtdev_name, + args->start.srcdevid, + args->start.srcdev_name, + args->start.cont_reading_from_srcdev_mode); + args->result = ret; + /* don't warn if EINPROGRESS, someone else might be running scrub */ + if (ret == BTRFS_IOCTL_DEV_REPLACE_RESULT_SCRUB_INPROGRESS) + ret = 0; + + return ret; +} + /* * blocked until all flighting bios are finished. */ diff --git a/fs/btrfs/dev-replace.h b/fs/btrfs/dev-replace.h index 29e3ef5f96bd..e922b42d91df 100644 --- a/fs/btrfs/dev-replace.h +++ b/fs/btrfs/dev-replace.h @@ -25,8 +25,10 @@ int btrfs_init_dev_replace(struct btrfs_fs_info *fs_info); int btrfs_run_dev_replace(struct btrfs_trans_handle *trans, struct btrfs_fs_info *fs_info); void btrfs_after_dev_replace_commit(struct btrfs_fs_info *fs_info); -int btrfs_dev_replace_start(struct btrfs_root *root, +int btrfs_dev_replace_by_ioctl(struct btrfs_root *root, struct btrfs_ioctl_dev_replace_args *args); +int btrfs_dev_replace_start(struct btrfs_root *root, char *tgtdev_name, + u64 srcdevid, char *srcdev_name, int read_src); void btrfs_dev_replace_status(struct btrfs_fs_info *fs_info, struct btrfs_ioctl_dev_replace_args *args); int btrfs_dev_replace_cancel(struct btrfs_fs_info *fs_info, diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index 65903ec5e1c0..36b1ed223509 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -4430,7 +4430,7 @@ static long btrfs_ioctl_dev_replace(struct btrfs_root *root, void __user *arg) 1)) { ret = BTRFS_ERROR_DEV_EXCL_RUN_IN_PROGRESS; } else { - ret = btrfs_dev_replace_start(root, p); + ret = btrfs_dev_replace_by_ioctl(root, p); atomic_set( &root->fs_info->mutually_exclusive_operation_running, 0); -- cgit v1.3-8-gc7d7 From d4ae133b2d195d88cf5394072724adfa6ccdd64b Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Fri, 1 Apr 2016 16:14:23 -0400 Subject: btrfs: uapi/linux/btrfs.h migration, move BTRFS_LABEL_SIZE BTRFS_LABEL_SIZE is required to define the BTRFS_IOC_GET_FSLABEL and BTRFS_IOC_SET_FSLABEL ioctls. Signed-off-by: Jeff Mahoney Reviewed-by: Liu Bo Reviewed-by: Josef Bacik Signed-off-by: David Sterba --- fs/btrfs/ctree.h | 1 - include/uapi/linux/btrfs.h | 1 + 2 files changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 84a6a5b3384a..3beaa24aff9b 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -410,7 +410,6 @@ struct btrfs_header { * room to translate 14 chunks with 3 stripes each. */ #define BTRFS_SYSTEM_CHUNK_ARRAY_SIZE 2048 -#define BTRFS_LABEL_SIZE 256 /* * just in case we somehow lose the roots and are not able to mount, diff --git a/include/uapi/linux/btrfs.h b/include/uapi/linux/btrfs.h index dea893199257..11eee3439994 100644 --- a/include/uapi/linux/btrfs.h +++ b/include/uapi/linux/btrfs.h @@ -23,6 +23,7 @@ #define BTRFS_IOCTL_MAGIC 0x94 #define BTRFS_VOL_NAME_MAX 255 +#define BTRFS_LABEL_SIZE 256 /* this should be 4k */ #define BTRFS_PATH_NAME_MAX 4087 -- cgit v1.3-8-gc7d7 From 83288b60bf6668933689078973136e0c9d387b38 Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Fri, 1 Apr 2016 16:14:24 -0400 Subject: btrfs: uapi/linux/btrfs.h migration, qgroup limit flags The BTRFS_QGROUP_LIMIT_* flags are required to tell the kernel which fields are valid when using the BTRFS_IOC_QGROUP_LIMIT ioctl. Signed-off-by: Jeff Mahoney Reviewed-by: Liu Bo Reviewed-by: Josef Bacik Signed-off-by: David Sterba --- fs/btrfs/ctree.h | 8 -------- include/uapi/linux/btrfs.h | 22 +++++++++++++++++++++- 2 files changed, 21 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 3beaa24aff9b..c228b39fbd15 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -1154,14 +1154,6 @@ struct btrfs_qgroup_info_item { __le64 excl_cmpr; } __attribute__ ((__packed__)); -/* flags definition for qgroup limits */ -#define BTRFS_QGROUP_LIMIT_MAX_RFER (1ULL << 0) -#define BTRFS_QGROUP_LIMIT_MAX_EXCL (1ULL << 1) -#define BTRFS_QGROUP_LIMIT_RSV_RFER (1ULL << 2) -#define BTRFS_QGROUP_LIMIT_RSV_EXCL (1ULL << 3) -#define BTRFS_QGROUP_LIMIT_RFER_CMPR (1ULL << 4) -#define BTRFS_QGROUP_LIMIT_EXCL_CMPR (1ULL << 5) - struct btrfs_qgroup_limit_item { /* * only updated when any of the other values change diff --git a/include/uapi/linux/btrfs.h b/include/uapi/linux/btrfs.h index 11eee3439994..9651af3c46a1 100644 --- a/include/uapi/linux/btrfs.h +++ b/include/uapi/linux/btrfs.h @@ -41,7 +41,19 @@ struct btrfs_ioctl_vol_args { #define BTRFS_UUID_SIZE 16 #define BTRFS_UUID_UNPARSED_SIZE 37 -#define BTRFS_QGROUP_INHERIT_SET_LIMITS (1ULL << 0) +/* + * flags definition for qgroup limits + * + * Used by: + * struct btrfs_qgroup_limit.flags + * struct btrfs_qgroup_limit_item.flags + */ +#define BTRFS_QGROUP_LIMIT_MAX_RFER (1ULL << 0) +#define BTRFS_QGROUP_LIMIT_MAX_EXCL (1ULL << 1) +#define BTRFS_QGROUP_LIMIT_RSV_RFER (1ULL << 2) +#define BTRFS_QGROUP_LIMIT_RSV_EXCL (1ULL << 3) +#define BTRFS_QGROUP_LIMIT_RFER_CMPR (1ULL << 4) +#define BTRFS_QGROUP_LIMIT_EXCL_CMPR (1ULL << 5) struct btrfs_qgroup_limit { __u64 flags; @@ -51,6 +63,14 @@ struct btrfs_qgroup_limit { __u64 rsv_excl; }; +/* + * flags definition for qgroup inheritance + * + * Used by: + * struct btrfs_qgroup_inherit.flags + */ +#define BTRFS_QGROUP_INHERIT_SET_LIMITS (1ULL << 0) + struct btrfs_qgroup_inherit { __u64 flags; __u64 num_qgroups; -- cgit v1.3-8-gc7d7 From 18db9ac644badcb948a623791e599672edade6dd Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Fri, 1 Apr 2016 16:14:26 -0400 Subject: btrfs: uapi/linux/btrfs.h migration, move feature flags The compat/compat_ro/incompat feature flags are used by the feature set/get ioctls. Signed-off-by: Jeff Mahoney Reviewed-by: Liu Bo Reviewed-by: Josef Bacik Signed-off-by: David Sterba --- fs/btrfs/ctree.h | 25 ------------------------- include/uapi/linux/btrfs.h | 31 +++++++++++++++++++++++++++++++ 2 files changed, 31 insertions(+), 25 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index c228b39fbd15..378482c705c3 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -506,31 +506,6 @@ struct btrfs_super_block { * Compat flags that we support. If any incompat flags are set other than the * ones specified below then we will fail to mount */ -#define BTRFS_FEATURE_COMPAT_RO_FREE_SPACE_TREE (1ULL << 0) - -#define BTRFS_FEATURE_INCOMPAT_MIXED_BACKREF (1ULL << 0) -#define BTRFS_FEATURE_INCOMPAT_DEFAULT_SUBVOL (1ULL << 1) -#define BTRFS_FEATURE_INCOMPAT_MIXED_GROUPS (1ULL << 2) -#define BTRFS_FEATURE_INCOMPAT_COMPRESS_LZO (1ULL << 3) -/* - * some patches floated around with a second compression method - * lets save that incompat here for when they do get in - * Note we don't actually support it, we're just reserving the - * number - */ -#define BTRFS_FEATURE_INCOMPAT_COMPRESS_LZOv2 (1ULL << 4) - -/* - * older kernels tried to do bigger metadata blocks, but the - * code was pretty buggy. Lets not let them try anymore. - */ -#define BTRFS_FEATURE_INCOMPAT_BIG_METADATA (1ULL << 5) - -#define BTRFS_FEATURE_INCOMPAT_EXTENDED_IREF (1ULL << 6) -#define BTRFS_FEATURE_INCOMPAT_RAID56 (1ULL << 7) -#define BTRFS_FEATURE_INCOMPAT_SKINNY_METADATA (1ULL << 8) -#define BTRFS_FEATURE_INCOMPAT_NO_HOLES (1ULL << 9) - #define BTRFS_FEATURE_COMPAT_SUPP 0ULL #define BTRFS_FEATURE_COMPAT_SAFE_SET 0ULL #define BTRFS_FEATURE_COMPAT_SAFE_CLEAR 0ULL diff --git a/include/uapi/linux/btrfs.h b/include/uapi/linux/btrfs.h index 0316e23b32cc..de98717386ee 100644 --- a/include/uapi/linux/btrfs.h +++ b/include/uapi/linux/btrfs.h @@ -222,6 +222,37 @@ struct btrfs_ioctl_fs_info_args { __u64 reserved[122]; /* pad to 1k */ }; +/* + * feature flags + * + * Used by: + * struct btrfs_ioctl_feature_flags + */ +#define BTRFS_FEATURE_COMPAT_RO_FREE_SPACE_TREE (1ULL << 0) + +#define BTRFS_FEATURE_INCOMPAT_MIXED_BACKREF (1ULL << 0) +#define BTRFS_FEATURE_INCOMPAT_DEFAULT_SUBVOL (1ULL << 1) +#define BTRFS_FEATURE_INCOMPAT_MIXED_GROUPS (1ULL << 2) +#define BTRFS_FEATURE_INCOMPAT_COMPRESS_LZO (1ULL << 3) +/* + * some patches floated around with a second compression method + * lets save that incompat here for when they do get in + * Note we don't actually support it, we're just reserving the + * number + */ +#define BTRFS_FEATURE_INCOMPAT_COMPRESS_LZOv2 (1ULL << 4) + +/* + * older kernels tried to do bigger metadata blocks, but the + * code was pretty buggy. Lets not let them try anymore. + */ +#define BTRFS_FEATURE_INCOMPAT_BIG_METADATA (1ULL << 5) + +#define BTRFS_FEATURE_INCOMPAT_EXTENDED_IREF (1ULL << 6) +#define BTRFS_FEATURE_INCOMPAT_RAID56 (1ULL << 7) +#define BTRFS_FEATURE_INCOMPAT_SKINNY_METADATA (1ULL << 8) +#define BTRFS_FEATURE_INCOMPAT_NO_HOLES (1ULL << 9) + struct btrfs_ioctl_feature_flags { __u64 compat_flags; __u64 compat_ro_flags; -- cgit v1.3-8-gc7d7 From 04cd01dffbf9be14ccc51595280c8dc8c318a9c9 Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Fri, 1 Apr 2016 16:14:27 -0400 Subject: btrfs: uapi/linux/btrfs.h migration, move balance flags The BTRFS_BALANCE_* flags are used by struct btrfs_ioctl_balance_args.flags and btrfs_ioctl_balance_args.{data,meta,sys}.flags in the BTRFS_IOC_BALANCE ioctl. Signed-off-by: Jeff Mahoney Reviewed-by: Liu Bo Reviewed-by: Josef Bacik Signed-off-by: David Sterba --- fs/btrfs/volumes.h | 46 --------------------------------- include/uapi/linux/btrfs.h | 64 ++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 64 insertions(+), 46 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h index 1939ebde63df..144cec33357a 100644 --- a/fs/btrfs/volumes.h +++ b/fs/btrfs/volumes.h @@ -357,52 +357,6 @@ struct map_lookup { #define map_lookup_size(n) (sizeof(struct map_lookup) + \ (sizeof(struct btrfs_bio_stripe) * (n))) -/* - * Restriper's general type filter - */ -#define BTRFS_BALANCE_DATA (1ULL << 0) -#define BTRFS_BALANCE_SYSTEM (1ULL << 1) -#define BTRFS_BALANCE_METADATA (1ULL << 2) - -#define BTRFS_BALANCE_TYPE_MASK (BTRFS_BALANCE_DATA | \ - BTRFS_BALANCE_SYSTEM | \ - BTRFS_BALANCE_METADATA) - -#define BTRFS_BALANCE_FORCE (1ULL << 3) -#define BTRFS_BALANCE_RESUME (1ULL << 4) - -/* - * Balance filters - */ -#define BTRFS_BALANCE_ARGS_PROFILES (1ULL << 0) -#define BTRFS_BALANCE_ARGS_USAGE (1ULL << 1) -#define BTRFS_BALANCE_ARGS_DEVID (1ULL << 2) -#define BTRFS_BALANCE_ARGS_DRANGE (1ULL << 3) -#define BTRFS_BALANCE_ARGS_VRANGE (1ULL << 4) -#define BTRFS_BALANCE_ARGS_LIMIT (1ULL << 5) -#define BTRFS_BALANCE_ARGS_LIMIT_RANGE (1ULL << 6) -#define BTRFS_BALANCE_ARGS_STRIPES_RANGE (1ULL << 7) -#define BTRFS_BALANCE_ARGS_USAGE_RANGE (1ULL << 10) - -#define BTRFS_BALANCE_ARGS_MASK \ - (BTRFS_BALANCE_ARGS_PROFILES | \ - BTRFS_BALANCE_ARGS_USAGE | \ - BTRFS_BALANCE_ARGS_DEVID | \ - BTRFS_BALANCE_ARGS_DRANGE | \ - BTRFS_BALANCE_ARGS_VRANGE | \ - BTRFS_BALANCE_ARGS_LIMIT | \ - BTRFS_BALANCE_ARGS_LIMIT_RANGE | \ - BTRFS_BALANCE_ARGS_STRIPES_RANGE | \ - BTRFS_BALANCE_ARGS_USAGE_RANGE) - -/* - * Profile changing flags. When SOFT is set we won't relocate chunk if - * it already has the target profile (even though it may be - * half-filled). - */ -#define BTRFS_BALANCE_ARGS_CONVERT (1ULL << 8) -#define BTRFS_BALANCE_ARGS_SOFT (1ULL << 9) - struct btrfs_balance_args; struct btrfs_balance_progress; struct btrfs_balance_control { diff --git a/include/uapi/linux/btrfs.h b/include/uapi/linux/btrfs.h index de98717386ee..abae362d4ec7 100644 --- a/include/uapi/linux/btrfs.h +++ b/include/uapi/linux/btrfs.h @@ -317,6 +317,70 @@ struct btrfs_balance_progress { __u64 completed; /* # of chunks relocated so far */ }; +/* + * flags definition for balance + * + * Restriper's general type filter + * + * Used by: + * btrfs_ioctl_balance_args.flags + * btrfs_balance_control.flags (internal) + */ +#define BTRFS_BALANCE_DATA (1ULL << 0) +#define BTRFS_BALANCE_SYSTEM (1ULL << 1) +#define BTRFS_BALANCE_METADATA (1ULL << 2) + +#define BTRFS_BALANCE_TYPE_MASK (BTRFS_BALANCE_DATA | \ + BTRFS_BALANCE_SYSTEM | \ + BTRFS_BALANCE_METADATA) + +#define BTRFS_BALANCE_FORCE (1ULL << 3) +#define BTRFS_BALANCE_RESUME (1ULL << 4) + +/* + * flags definitions for per-type balance args + * + * Balance filters + * + * Used by: + * struct btrfs_balance_args + */ +#define BTRFS_BALANCE_ARGS_PROFILES (1ULL << 0) +#define BTRFS_BALANCE_ARGS_USAGE (1ULL << 1) +#define BTRFS_BALANCE_ARGS_DEVID (1ULL << 2) +#define BTRFS_BALANCE_ARGS_DRANGE (1ULL << 3) +#define BTRFS_BALANCE_ARGS_VRANGE (1ULL << 4) +#define BTRFS_BALANCE_ARGS_LIMIT (1ULL << 5) +#define BTRFS_BALANCE_ARGS_LIMIT_RANGE (1ULL << 6) +#define BTRFS_BALANCE_ARGS_STRIPES_RANGE (1ULL << 7) +#define BTRFS_BALANCE_ARGS_USAGE_RANGE (1ULL << 10) + +#define BTRFS_BALANCE_ARGS_MASK \ + (BTRFS_BALANCE_ARGS_PROFILES | \ + BTRFS_BALANCE_ARGS_USAGE | \ + BTRFS_BALANCE_ARGS_DEVID | \ + BTRFS_BALANCE_ARGS_DRANGE | \ + BTRFS_BALANCE_ARGS_VRANGE | \ + BTRFS_BALANCE_ARGS_LIMIT | \ + BTRFS_BALANCE_ARGS_LIMIT_RANGE | \ + BTRFS_BALANCE_ARGS_STRIPES_RANGE | \ + BTRFS_BALANCE_ARGS_USAGE_RANGE) + +/* + * Profile changing flags. When SOFT is set we won't relocate chunk if + * it already has the target profile (even though it may be + * half-filled). + */ +#define BTRFS_BALANCE_ARGS_CONVERT (1ULL << 8) +#define BTRFS_BALANCE_ARGS_SOFT (1ULL << 9) + + +/* + * flags definition for balance state + * + * Used by: + * struct btrfs_ioctl_balance_args.state + */ #define BTRFS_BALANCE_STATE_RUNNING (1ULL << 0) #define BTRFS_BALANCE_STATE_PAUSE_REQ (1ULL << 1) #define BTRFS_BALANCE_STATE_CANCEL_REQ (1ULL << 2) -- cgit v1.3-8-gc7d7 From 33ca913349962208e13e894ada99b9ae6e0080ee Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Fri, 1 Apr 2016 16:14:28 -0400 Subject: btrfs: uapi/linux/btrfs.h migration, move struct btrfs_ioctl_defrag_range_args struct btrfs_ioctl_defrag_range_args is used by the BTRFS_IOC_DEFRAG_RANGE ioctl. Signed-off-by: Jeff Mahoney Reviewed-by: Liu Bo Reviewed-by: Josef Bacik Signed-off-by: David Sterba --- fs/btrfs/ctree.h | 31 ------------------------------- include/uapi/linux/btrfs.h | 38 +++++++++++++++++++++++++++++++++++++- 2 files changed, 37 insertions(+), 32 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 378482c705c3..89f36b6176b9 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -1992,37 +1992,6 @@ struct btrfs_root { atomic_t qgroup_meta_rsv; }; -struct btrfs_ioctl_defrag_range_args { - /* start of the defrag operation */ - __u64 start; - - /* number of bytes to defrag, use (u64)-1 to say all */ - __u64 len; - - /* - * flags for the operation, which can include turning - * on compression for this one defrag - */ - __u64 flags; - - /* - * any extent bigger than this will be considered - * already defragged. Use 0 to take the kernel default - * Use 1 to say every single extent must be rewritten - */ - __u32 extent_thresh; - - /* - * which compression method to use if turning on compression - * for this defrag operation. If unspecified, zlib will - * be used - */ - __u32 compress_type; - - /* spare for later */ - __u32 unused[4]; -}; - /* * inode items have the data typically returned from stat and store other diff --git a/include/uapi/linux/btrfs.h b/include/uapi/linux/btrfs.h index abae362d4ec7..98aff38a70d3 100644 --- a/include/uapi/linux/btrfs.h +++ b/include/uapi/linux/btrfs.h @@ -474,9 +474,45 @@ struct btrfs_ioctl_clone_range_args { __u64 dest_offset; }; -/* flags for the defrag range ioctl */ +/* + * flags definition for the defrag range ioctl + * + * Used by: + * struct btrfs_ioctl_defrag_range_args.flags + */ #define BTRFS_DEFRAG_RANGE_COMPRESS 1 #define BTRFS_DEFRAG_RANGE_START_IO 2 +struct btrfs_ioctl_defrag_range_args { + /* start of the defrag operation */ + __u64 start; + + /* number of bytes to defrag, use (u64)-1 to say all */ + __u64 len; + + /* + * flags for the operation, which can include turning + * on compression for this one defrag + */ + __u64 flags; + + /* + * any extent bigger than this will be considered + * already defragged. Use 0 to take the kernel default + * Use 1 to say every single extent must be rewritten + */ + __u32 extent_thresh; + + /* + * which compression method to use if turning on compression + * for this defrag operation. If unspecified, zlib will + * be used + */ + __u32 compress_type; + + /* spare for later */ + __u32 unused[4]; +}; + #define BTRFS_SAME_DATA_DIFFERS 1 /* For extent-same ioctl */ -- cgit v1.3-8-gc7d7 From db6711600e27c885aed89751f04e727f3af26715 Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Fri, 1 Apr 2016 16:14:29 -0400 Subject: btrfs: uapi/linux/btrfs_tree.h migration, item types and defines The BTRFS_IOC_SEARCH_TREE ioctl returns file system items directly to userspace. In order to decode them, full type information is required. Create a new header, btrfs_tree to contain these since most users won't need them. Signed-off-by: Jeff Mahoney Reviewed-by: Liu Bo Reviewed-by: Josef Bacik Signed-off-by: David Sterba --- fs/btrfs/ctree.h | 949 +-------------------------------------- include/uapi/linux/btrfs_tree.h | 966 ++++++++++++++++++++++++++++++++++++++++ 2 files changed, 967 insertions(+), 948 deletions(-) create mode 100644 include/uapi/linux/btrfs_tree.h (limited to 'fs') diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 89f36b6176b9..cf34fb58874c 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -33,6 +33,7 @@ #include #include #include +#include #include #include #include @@ -64,98 +65,6 @@ struct btrfs_ordered_sum; #define BTRFS_COMPAT_EXTENT_TREE_V0 -/* holds pointers to all of the tree roots */ -#define BTRFS_ROOT_TREE_OBJECTID 1ULL - -/* stores information about which extents are in use, and reference counts */ -#define BTRFS_EXTENT_TREE_OBJECTID 2ULL - -/* - * chunk tree stores translations from logical -> physical block numbering - * the super block points to the chunk tree - */ -#define BTRFS_CHUNK_TREE_OBJECTID 3ULL - -/* - * stores information about which areas of a given device are in use. - * one per device. The tree of tree roots points to the device tree - */ -#define BTRFS_DEV_TREE_OBJECTID 4ULL - -/* one per subvolume, storing files and directories */ -#define BTRFS_FS_TREE_OBJECTID 5ULL - -/* directory objectid inside the root tree */ -#define BTRFS_ROOT_TREE_DIR_OBJECTID 6ULL - -/* holds checksums of all the data extents */ -#define BTRFS_CSUM_TREE_OBJECTID 7ULL - -/* holds quota configuration and tracking */ -#define BTRFS_QUOTA_TREE_OBJECTID 8ULL - -/* for storing items that use the BTRFS_UUID_KEY* types */ -#define BTRFS_UUID_TREE_OBJECTID 9ULL - -/* tracks free space in block groups. */ -#define BTRFS_FREE_SPACE_TREE_OBJECTID 10ULL - -/* device stats in the device tree */ -#define BTRFS_DEV_STATS_OBJECTID 0ULL - -/* for storing balance parameters in the root tree */ -#define BTRFS_BALANCE_OBJECTID -4ULL - -/* orhpan objectid for tracking unlinked/truncated files */ -#define BTRFS_ORPHAN_OBJECTID -5ULL - -/* does write ahead logging to speed up fsyncs */ -#define BTRFS_TREE_LOG_OBJECTID -6ULL -#define BTRFS_TREE_LOG_FIXUP_OBJECTID -7ULL - -/* for space balancing */ -#define BTRFS_TREE_RELOC_OBJECTID -8ULL -#define BTRFS_DATA_RELOC_TREE_OBJECTID -9ULL - -/* - * extent checksums all have this objectid - * this allows them to share the logging tree - * for fsyncs - */ -#define BTRFS_EXTENT_CSUM_OBJECTID -10ULL - -/* For storing free space cache */ -#define BTRFS_FREE_SPACE_OBJECTID -11ULL - -/* - * The inode number assigned to the special inode for storing - * free ino cache - */ -#define BTRFS_FREE_INO_OBJECTID -12ULL - -/* dummy objectid represents multiple objectids */ -#define BTRFS_MULTIPLE_OBJECTIDS -255ULL - -/* - * All files have objectids in this range. - */ -#define BTRFS_FIRST_FREE_OBJECTID 256ULL -#define BTRFS_LAST_FREE_OBJECTID -256ULL -#define BTRFS_FIRST_CHUNK_TREE_OBJECTID 256ULL - - -/* - * the device items go into the chunk tree. The key is in the form - * [ 1 BTRFS_DEV_ITEM_KEY device_id ] - */ -#define BTRFS_DEV_ITEMS_OBJECTID 1ULL - -#define BTRFS_BTREE_INODE_OBJECTID 1 - -#define BTRFS_EMPTY_SUBVOL_DIR_OBJECTID 2 - -#define BTRFS_DEV_REPLACE_DEVID 0ULL - /* * the max metadata block size. This limit is somewhat artificial, * but the memmove costs go through the roof for larger blocks. @@ -175,12 +84,6 @@ struct btrfs_ordered_sum; */ #define BTRFS_LINK_MAX 65535U -/* 32 bytes in various csum fields */ -#define BTRFS_CSUM_SIZE 32 - -/* csum types */ -#define BTRFS_CSUM_TYPE_CRC32 0 - static const int btrfs_csum_sizes[] = { 4 }; /* four bytes for CRC32 */ @@ -189,17 +92,6 @@ static const int btrfs_csum_sizes[] = { 4 }; /* spefic to btrfs_map_block(), therefore not in include/linux/blk_types.h */ #define REQ_GET_READ_MIRRORS (1 << 30) -#define BTRFS_FT_UNKNOWN 0 -#define BTRFS_FT_REG_FILE 1 -#define BTRFS_FT_DIR 2 -#define BTRFS_FT_CHRDEV 3 -#define BTRFS_FT_BLKDEV 4 -#define BTRFS_FT_FIFO 5 -#define BTRFS_FT_SOCK 6 -#define BTRFS_FT_SYMLINK 7 -#define BTRFS_FT_XATTR 8 -#define BTRFS_FT_MAX 9 - /* ioprio of readahead is set to idle */ #define BTRFS_IOPRIO_READA (IOPRIO_PRIO_VALUE(IOPRIO_CLASS_IDLE, 0)) @@ -207,138 +99,10 @@ static const int btrfs_csum_sizes[] = { 4 }; #define BTRFS_MAX_EXTENT_SIZE SZ_128M -/* - * The key defines the order in the tree, and so it also defines (optimal) - * block layout. - * - * objectid corresponds to the inode number. - * - * type tells us things about the object, and is a kind of stream selector. - * so for a given inode, keys with type of 1 might refer to the inode data, - * type of 2 may point to file data in the btree and type == 3 may point to - * extents. - * - * offset is the starting byte offset for this key in the stream. - * - * btrfs_disk_key is in disk byte order. struct btrfs_key is always - * in cpu native order. Otherwise they are identical and their sizes - * should be the same (ie both packed) - */ -struct btrfs_disk_key { - __le64 objectid; - u8 type; - __le64 offset; -} __attribute__ ((__packed__)); - -struct btrfs_key { - u64 objectid; - u8 type; - u64 offset; -} __attribute__ ((__packed__)); - struct btrfs_mapping_tree { struct extent_map_tree map_tree; }; -struct btrfs_dev_item { - /* the internal btrfs device id */ - __le64 devid; - - /* size of the device */ - __le64 total_bytes; - - /* bytes used */ - __le64 bytes_used; - - /* optimal io alignment for this device */ - __le32 io_align; - - /* optimal io width for this device */ - __le32 io_width; - - /* minimal io size for this device */ - __le32 sector_size; - - /* type and info about this device */ - __le64 type; - - /* expected generation for this device */ - __le64 generation; - - /* - * starting byte of this partition on the device, - * to allow for stripe alignment in the future - */ - __le64 start_offset; - - /* grouping information for allocation decisions */ - __le32 dev_group; - - /* seek speed 0-100 where 100 is fastest */ - u8 seek_speed; - - /* bandwidth 0-100 where 100 is fastest */ - u8 bandwidth; - - /* btrfs generated uuid for this device */ - u8 uuid[BTRFS_UUID_SIZE]; - - /* uuid of FS who owns this device */ - u8 fsid[BTRFS_UUID_SIZE]; -} __attribute__ ((__packed__)); - -struct btrfs_stripe { - __le64 devid; - __le64 offset; - u8 dev_uuid[BTRFS_UUID_SIZE]; -} __attribute__ ((__packed__)); - -struct btrfs_chunk { - /* size of this chunk in bytes */ - __le64 length; - - /* objectid of the root referencing this chunk */ - __le64 owner; - - __le64 stripe_len; - __le64 type; - - /* optimal io alignment for this chunk */ - __le32 io_align; - - /* optimal io width for this chunk */ - __le32 io_width; - - /* minimal io size for this chunk */ - __le32 sector_size; - - /* 2^16 stripes is quite a lot, a second limit is the size of a single - * item in the btree - */ - __le16 num_stripes; - - /* sub stripes only matter for raid10 */ - __le16 sub_stripes; - struct btrfs_stripe stripe; - /* additional stripes go here */ -} __attribute__ ((__packed__)); - -#define BTRFS_FREE_SPACE_EXTENT 1 -#define BTRFS_FREE_SPACE_BITMAP 2 - -struct btrfs_free_space_entry { - __le64 offset; - __le64 bytes; - u8 type; -} __attribute__ ((__packed__)); - -struct btrfs_free_space_header { - struct btrfs_disk_key location; - __le64 generation; - __le64 num_entries; - __le64 num_bitmaps; -} __attribute__ ((__packed__)); - static inline unsigned long btrfs_chunk_item_size(int num_stripes) { BUG_ON(num_stripes == 0); @@ -346,9 +110,6 @@ static inline unsigned long btrfs_chunk_item_size(int num_stripes) sizeof(struct btrfs_stripe) * (num_stripes - 1); } -#define BTRFS_HEADER_FLAG_WRITTEN (1ULL << 0) -#define BTRFS_HEADER_FLAG_RELOC (1ULL << 1) - /* * File system states */ @@ -357,13 +118,6 @@ static inline unsigned long btrfs_chunk_item_size(int num_stripes) #define BTRFS_FS_STATE_TRANS_ABORTED 2 #define BTRFS_FS_STATE_DEV_REPLACING 3 -/* Super block flags */ -/* Errors detected */ -#define BTRFS_SUPER_FLAG_ERROR (1ULL << 2) - -#define BTRFS_SUPER_FLAG_SEEDING (1ULL << 32) -#define BTRFS_SUPER_FLAG_METADUMP (1ULL << 33) - #define BTRFS_BACKREF_REV_MAX 256 #define BTRFS_BACKREF_REV_SHIFT 56 #define BTRFS_BACKREF_REV_MASK (((u64)BTRFS_BACKREF_REV_MAX - 1) << \ @@ -598,357 +352,8 @@ struct btrfs_path { unsigned int need_commit_sem:1; unsigned int skip_release_on_error:1; }; - -/* - * items in the extent btree are used to record the objectid of the - * owner of the block and the number of references - */ - -struct btrfs_extent_item { - __le64 refs; - __le64 generation; - __le64 flags; -} __attribute__ ((__packed__)); - -struct btrfs_extent_item_v0 { - __le32 refs; -} __attribute__ ((__packed__)); - #define BTRFS_MAX_EXTENT_ITEM_SIZE(r) ((BTRFS_LEAF_DATA_SIZE(r) >> 4) - \ sizeof(struct btrfs_item)) - -#define BTRFS_EXTENT_FLAG_DATA (1ULL << 0) -#define BTRFS_EXTENT_FLAG_TREE_BLOCK (1ULL << 1) - -/* following flags only apply to tree blocks */ - -/* use full backrefs for extent pointers in the block */ -#define BTRFS_BLOCK_FLAG_FULL_BACKREF (1ULL << 8) - -/* - * this flag is only used internally by scrub and may be changed at any time - * it is only declared here to avoid collisions - */ -#define BTRFS_EXTENT_FLAG_SUPER (1ULL << 48) - -struct btrfs_tree_block_info { - struct btrfs_disk_key key; - u8 level; -} __attribute__ ((__packed__)); - -struct btrfs_extent_data_ref { - __le64 root; - __le64 objectid; - __le64 offset; - __le32 count; -} __attribute__ ((__packed__)); - -struct btrfs_shared_data_ref { - __le32 count; -} __attribute__ ((__packed__)); - -struct btrfs_extent_inline_ref { - u8 type; - __le64 offset; -} __attribute__ ((__packed__)); - -/* old style backrefs item */ -struct btrfs_extent_ref_v0 { - __le64 root; - __le64 generation; - __le64 objectid; - __le32 count; -} __attribute__ ((__packed__)); - - -/* dev extents record free space on individual devices. The owner - * field points back to the chunk allocation mapping tree that allocated - * the extent. The chunk tree uuid field is a way to double check the owner - */ -struct btrfs_dev_extent { - __le64 chunk_tree; - __le64 chunk_objectid; - __le64 chunk_offset; - __le64 length; - u8 chunk_tree_uuid[BTRFS_UUID_SIZE]; -} __attribute__ ((__packed__)); - -struct btrfs_inode_ref { - __le64 index; - __le16 name_len; - /* name goes here */ -} __attribute__ ((__packed__)); - -struct btrfs_inode_extref { - __le64 parent_objectid; - __le64 index; - __le16 name_len; - __u8 name[0]; - /* name goes here */ -} __attribute__ ((__packed__)); - -struct btrfs_timespec { - __le64 sec; - __le32 nsec; -} __attribute__ ((__packed__)); - -struct btrfs_inode_item { - /* nfs style generation number */ - __le64 generation; - /* transid that last touched this inode */ - __le64 transid; - __le64 size; - __le64 nbytes; - __le64 block_group; - __le32 nlink; - __le32 uid; - __le32 gid; - __le32 mode; - __le64 rdev; - __le64 flags; - - /* modification sequence number for NFS */ - __le64 sequence; - - /* - * a little future expansion, for more than this we can - * just grow the inode item and version it - */ - __le64 reserved[4]; - struct btrfs_timespec atime; - struct btrfs_timespec ctime; - struct btrfs_timespec mtime; - struct btrfs_timespec otime; -} __attribute__ ((__packed__)); - -struct btrfs_dir_log_item { - __le64 end; -} __attribute__ ((__packed__)); - -struct btrfs_dir_item { - struct btrfs_disk_key location; - __le64 transid; - __le16 data_len; - __le16 name_len; - u8 type; -} __attribute__ ((__packed__)); - -#define BTRFS_ROOT_SUBVOL_RDONLY (1ULL << 0) - -/* - * Internal in-memory flag that a subvolume has been marked for deletion but - * still visible as a directory - */ -#define BTRFS_ROOT_SUBVOL_DEAD (1ULL << 48) - -struct btrfs_root_item { - struct btrfs_inode_item inode; - __le64 generation; - __le64 root_dirid; - __le64 bytenr; - __le64 byte_limit; - __le64 bytes_used; - __le64 last_snapshot; - __le64 flags; - __le32 refs; - struct btrfs_disk_key drop_progress; - u8 drop_level; - u8 level; - - /* - * The following fields appear after subvol_uuids+subvol_times - * were introduced. - */ - - /* - * This generation number is used to test if the new fields are valid - * and up to date while reading the root item. Every time the root item - * is written out, the "generation" field is copied into this field. If - * anyone ever mounted the fs with an older kernel, we will have - * mismatching generation values here and thus must invalidate the - * new fields. See btrfs_update_root and btrfs_find_last_root for - * details. - * the offset of generation_v2 is also used as the start for the memset - * when invalidating the fields. - */ - __le64 generation_v2; - u8 uuid[BTRFS_UUID_SIZE]; - u8 parent_uuid[BTRFS_UUID_SIZE]; - u8 received_uuid[BTRFS_UUID_SIZE]; - __le64 ctransid; /* updated when an inode changes */ - __le64 otransid; /* trans when created */ - __le64 stransid; /* trans when sent. non-zero for received subvol */ - __le64 rtransid; /* trans when received. non-zero for received subvol */ - struct btrfs_timespec ctime; - struct btrfs_timespec otime; - struct btrfs_timespec stime; - struct btrfs_timespec rtime; - __le64 reserved[8]; /* for future */ -} __attribute__ ((__packed__)); - -/* - * this is used for both forward and backward root refs - */ -struct btrfs_root_ref { - __le64 dirid; - __le64 sequence; - __le16 name_len; -} __attribute__ ((__packed__)); - -struct btrfs_disk_balance_args { - /* - * profiles to operate on, single is denoted by - * BTRFS_AVAIL_ALLOC_BIT_SINGLE - */ - __le64 profiles; - - /* - * usage filter - * BTRFS_BALANCE_ARGS_USAGE with a single value means '0..N' - * BTRFS_BALANCE_ARGS_USAGE_RANGE - range syntax, min..max - */ - union { - __le64 usage; - struct { - __le32 usage_min; - __le32 usage_max; - }; - }; - - /* devid filter */ - __le64 devid; - - /* devid subset filter [pstart..pend) */ - __le64 pstart; - __le64 pend; - - /* btrfs virtual address space subset filter [vstart..vend) */ - __le64 vstart; - __le64 vend; - - /* - * profile to convert to, single is denoted by - * BTRFS_AVAIL_ALLOC_BIT_SINGLE - */ - __le64 target; - - /* BTRFS_BALANCE_ARGS_* */ - __le64 flags; - - /* - * BTRFS_BALANCE_ARGS_LIMIT with value 'limit' - * BTRFS_BALANCE_ARGS_LIMIT_RANGE - the extend version can use minimum - * and maximum - */ - union { - __le64 limit; - struct { - __le32 limit_min; - __le32 limit_max; - }; - }; - - /* - * Process chunks that cross stripes_min..stripes_max devices, - * BTRFS_BALANCE_ARGS_STRIPES_RANGE - */ - __le32 stripes_min; - __le32 stripes_max; - - __le64 unused[6]; -} __attribute__ ((__packed__)); - -/* - * store balance parameters to disk so that balance can be properly - * resumed after crash or unmount - */ -struct btrfs_balance_item { - /* BTRFS_BALANCE_* */ - __le64 flags; - - struct btrfs_disk_balance_args data; - struct btrfs_disk_balance_args meta; - struct btrfs_disk_balance_args sys; - - __le64 unused[4]; -} __attribute__ ((__packed__)); - -#define BTRFS_FILE_EXTENT_INLINE 0 -#define BTRFS_FILE_EXTENT_REG 1 -#define BTRFS_FILE_EXTENT_PREALLOC 2 - -struct btrfs_file_extent_item { - /* - * transaction id that created this extent - */ - __le64 generation; - /* - * max number of bytes to hold this extent in ram - * when we split a compressed extent we can't know how big - * each of the resulting pieces will be. So, this is - * an upper limit on the size of the extent in ram instead of - * an exact limit. - */ - __le64 ram_bytes; - - /* - * 32 bits for the various ways we might encode the data, - * including compression and encryption. If any of these - * are set to something a given disk format doesn't understand - * it is treated like an incompat flag for reading and writing, - * but not for stat. - */ - u8 compression; - u8 encryption; - __le16 other_encoding; /* spare for later use */ - - /* are we inline data or a real extent? */ - u8 type; - - /* - * disk space consumed by the extent, checksum blocks are included - * in these numbers - * - * At this offset in the structure, the inline extent data start. - */ - __le64 disk_bytenr; - __le64 disk_num_bytes; - /* - * the logical offset in file blocks (no csums) - * this extent record is for. This allows a file extent to point - * into the middle of an existing extent on disk, sharing it - * between two snapshots (useful if some bytes in the middle of the - * extent have changed - */ - __le64 offset; - /* - * the logical number of file blocks (no csums included). This - * always reflects the size uncompressed and without encoding. - */ - __le64 num_bytes; - -} __attribute__ ((__packed__)); - -struct btrfs_csum_item { - u8 csum; -} __attribute__ ((__packed__)); - -struct btrfs_dev_stats_item { - /* - * grow this item struct at the end for future enhancements and keep - * the existing values unchanged - */ - __le64 values[BTRFS_DEV_STAT_VALUES_MAX]; -} __attribute__ ((__packed__)); - -#define BTRFS_DEV_REPLACE_ITEM_CONT_READING_FROM_SRCDEV_MODE_ALWAYS 0 -#define BTRFS_DEV_REPLACE_ITEM_CONT_READING_FROM_SRCDEV_MODE_AVOID 1 -#define BTRFS_DEV_REPLACE_ITEM_STATE_NEVER_STARTED 0 -#define BTRFS_DEV_REPLACE_ITEM_STATE_STARTED 1 -#define BTRFS_DEV_REPLACE_ITEM_STATE_SUSPENDED 2 -#define BTRFS_DEV_REPLACE_ITEM_STATE_FINISHED 3 -#define BTRFS_DEV_REPLACE_ITEM_STATE_CANCELED 4 - struct btrfs_dev_replace { u64 replace_state; /* see #define above */ u64 time_started; /* seconds since 1-Jan-1970 */ @@ -979,167 +384,6 @@ struct btrfs_dev_replace { struct btrfs_scrub_progress scrub_progress; }; -struct btrfs_dev_replace_item { - /* - * grow this item struct at the end for future enhancements and keep - * the existing values unchanged - */ - __le64 src_devid; - __le64 cursor_left; - __le64 cursor_right; - __le64 cont_reading_from_srcdev_mode; - - __le64 replace_state; - __le64 time_started; - __le64 time_stopped; - __le64 num_write_errors; - __le64 num_uncorrectable_read_errors; -} __attribute__ ((__packed__)); - -/* different types of block groups (and chunks) */ -#define BTRFS_BLOCK_GROUP_DATA (1ULL << 0) -#define BTRFS_BLOCK_GROUP_SYSTEM (1ULL << 1) -#define BTRFS_BLOCK_GROUP_METADATA (1ULL << 2) -#define BTRFS_BLOCK_GROUP_RAID0 (1ULL << 3) -#define BTRFS_BLOCK_GROUP_RAID1 (1ULL << 4) -#define BTRFS_BLOCK_GROUP_DUP (1ULL << 5) -#define BTRFS_BLOCK_GROUP_RAID10 (1ULL << 6) -#define BTRFS_BLOCK_GROUP_RAID5 (1ULL << 7) -#define BTRFS_BLOCK_GROUP_RAID6 (1ULL << 8) -#define BTRFS_BLOCK_GROUP_RESERVED (BTRFS_AVAIL_ALLOC_BIT_SINGLE | \ - BTRFS_SPACE_INFO_GLOBAL_RSV) - -enum btrfs_raid_types { - BTRFS_RAID_RAID10, - BTRFS_RAID_RAID1, - BTRFS_RAID_DUP, - BTRFS_RAID_RAID0, - BTRFS_RAID_SINGLE, - BTRFS_RAID_RAID5, - BTRFS_RAID_RAID6, - BTRFS_NR_RAID_TYPES -}; - -#define BTRFS_BLOCK_GROUP_TYPE_MASK (BTRFS_BLOCK_GROUP_DATA | \ - BTRFS_BLOCK_GROUP_SYSTEM | \ - BTRFS_BLOCK_GROUP_METADATA) - -#define BTRFS_BLOCK_GROUP_PROFILE_MASK (BTRFS_BLOCK_GROUP_RAID0 | \ - BTRFS_BLOCK_GROUP_RAID1 | \ - BTRFS_BLOCK_GROUP_RAID5 | \ - BTRFS_BLOCK_GROUP_RAID6 | \ - BTRFS_BLOCK_GROUP_DUP | \ - BTRFS_BLOCK_GROUP_RAID10) -#define BTRFS_BLOCK_GROUP_RAID56_MASK (BTRFS_BLOCK_GROUP_RAID5 | \ - BTRFS_BLOCK_GROUP_RAID6) - -/* - * We need a bit for restriper to be able to tell when chunks of type - * SINGLE are available. This "extended" profile format is used in - * fs_info->avail_*_alloc_bits (in-memory) and balance item fields - * (on-disk). The corresponding on-disk bit in chunk.type is reserved - * to avoid remappings between two formats in future. - */ -#define BTRFS_AVAIL_ALLOC_BIT_SINGLE (1ULL << 48) - -/* - * A fake block group type that is used to communicate global block reserve - * size to userspace via the SPACE_INFO ioctl. - */ -#define BTRFS_SPACE_INFO_GLOBAL_RSV (1ULL << 49) - -#define BTRFS_EXTENDED_PROFILE_MASK (BTRFS_BLOCK_GROUP_PROFILE_MASK | \ - BTRFS_AVAIL_ALLOC_BIT_SINGLE) - -static inline u64 chunk_to_extended(u64 flags) -{ - if ((flags & BTRFS_BLOCK_GROUP_PROFILE_MASK) == 0) - flags |= BTRFS_AVAIL_ALLOC_BIT_SINGLE; - - return flags; -} -static inline u64 extended_to_chunk(u64 flags) -{ - return flags & ~BTRFS_AVAIL_ALLOC_BIT_SINGLE; -} - -struct btrfs_block_group_item { - __le64 used; - __le64 chunk_objectid; - __le64 flags; -} __attribute__ ((__packed__)); - -struct btrfs_free_space_info { - __le32 extent_count; - __le32 flags; -} __attribute__ ((__packed__)); - -#define BTRFS_FREE_SPACE_USING_BITMAPS (1ULL << 0) - -#define BTRFS_QGROUP_LEVEL_SHIFT 48 -static inline u64 btrfs_qgroup_level(u64 qgroupid) -{ - return qgroupid >> BTRFS_QGROUP_LEVEL_SHIFT; -} - -/* - * is subvolume quota turned on? - */ -#define BTRFS_QGROUP_STATUS_FLAG_ON (1ULL << 0) -/* - * RESCAN is set during the initialization phase - */ -#define BTRFS_QGROUP_STATUS_FLAG_RESCAN (1ULL << 1) -/* - * Some qgroup entries are known to be out of date, - * either because the configuration has changed in a way that - * makes a rescan necessary, or because the fs has been mounted - * with a non-qgroup-aware version. - * Turning qouta off and on again makes it inconsistent, too. - */ -#define BTRFS_QGROUP_STATUS_FLAG_INCONSISTENT (1ULL << 2) - -#define BTRFS_QGROUP_STATUS_VERSION 1 - -struct btrfs_qgroup_status_item { - __le64 version; - /* - * the generation is updated during every commit. As older - * versions of btrfs are not aware of qgroups, it will be - * possible to detect inconsistencies by checking the - * generation on mount time - */ - __le64 generation; - - /* flag definitions see above */ - __le64 flags; - - /* - * only used during scanning to record the progress - * of the scan. It contains a logical address - */ - __le64 rescan; -} __attribute__ ((__packed__)); - -struct btrfs_qgroup_info_item { - __le64 generation; - __le64 rfer; - __le64 rfer_cmpr; - __le64 excl; - __le64 excl_cmpr; -} __attribute__ ((__packed__)); - -struct btrfs_qgroup_limit_item { - /* - * only updated when any of the other values change - */ - __le64 flags; - __le64 max_rfer; - __le64 max_excl; - __le64 rsv_rfer; - __le64 rsv_excl; -} __attribute__ ((__packed__)); - /* For raid type sysfs entries */ struct raid_kobject { int raid_type; @@ -1992,197 +1236,6 @@ struct btrfs_root { atomic_t qgroup_meta_rsv; }; - -/* - * inode items have the data typically returned from stat and store other - * info about object characteristics. There is one for every file and dir in - * the FS - */ -#define BTRFS_INODE_ITEM_KEY 1 -#define BTRFS_INODE_REF_KEY 12 -#define BTRFS_INODE_EXTREF_KEY 13 -#define BTRFS_XATTR_ITEM_KEY 24 -#define BTRFS_ORPHAN_ITEM_KEY 48 -/* reserve 2-15 close to the inode for later flexibility */ - -/* - * dir items are the name -> inode pointers in a directory. There is one - * for every name in a directory. - */ -#define BTRFS_DIR_LOG_ITEM_KEY 60 -#define BTRFS_DIR_LOG_INDEX_KEY 72 -#define BTRFS_DIR_ITEM_KEY 84 -#define BTRFS_DIR_INDEX_KEY 96 -/* - * extent data is for file data - */ -#define BTRFS_EXTENT_DATA_KEY 108 - -/* - * extent csums are stored in a separate tree and hold csums for - * an entire extent on disk. - */ -#define BTRFS_EXTENT_CSUM_KEY 128 - -/* - * root items point to tree roots. They are typically in the root - * tree used by the super block to find all the other trees - */ -#define BTRFS_ROOT_ITEM_KEY 132 - -/* - * root backrefs tie subvols and snapshots to the directory entries that - * reference them - */ -#define BTRFS_ROOT_BACKREF_KEY 144 - -/* - * root refs make a fast index for listing all of the snapshots and - * subvolumes referenced by a given root. They point directly to the - * directory item in the root that references the subvol - */ -#define BTRFS_ROOT_REF_KEY 156 - -/* - * extent items are in the extent map tree. These record which blocks - * are used, and how many references there are to each block - */ -#define BTRFS_EXTENT_ITEM_KEY 168 - -/* - * The same as the BTRFS_EXTENT_ITEM_KEY, except it's metadata we already know - * the length, so we save the level in key->offset instead of the length. - */ -#define BTRFS_METADATA_ITEM_KEY 169 - -#define BTRFS_TREE_BLOCK_REF_KEY 176 - -#define BTRFS_EXTENT_DATA_REF_KEY 178 - -#define BTRFS_EXTENT_REF_V0_KEY 180 - -#define BTRFS_SHARED_BLOCK_REF_KEY 182 - -#define BTRFS_SHARED_DATA_REF_KEY 184 - -/* - * block groups give us hints into the extent allocation trees. Which - * blocks are free etc etc - */ -#define BTRFS_BLOCK_GROUP_ITEM_KEY 192 - -/* - * Every block group is represented in the free space tree by a free space info - * item, which stores some accounting information. It is keyed on - * (block_group_start, FREE_SPACE_INFO, block_group_length). - */ -#define BTRFS_FREE_SPACE_INFO_KEY 198 - -/* - * A free space extent tracks an extent of space that is free in a block group. - * It is keyed on (start, FREE_SPACE_EXTENT, length). - */ -#define BTRFS_FREE_SPACE_EXTENT_KEY 199 - -/* - * When a block group becomes very fragmented, we convert it to use bitmaps - * instead of extents. A free space bitmap is keyed on - * (start, FREE_SPACE_BITMAP, length); the corresponding item is a bitmap with - * (length / sectorsize) bits. - */ -#define BTRFS_FREE_SPACE_BITMAP_KEY 200 - -#define BTRFS_DEV_EXTENT_KEY 204 -#define BTRFS_DEV_ITEM_KEY 216 -#define BTRFS_CHUNK_ITEM_KEY 228 - -/* - * Records the overall state of the qgroups. - * There's only one instance of this key present, - * (0, BTRFS_QGROUP_STATUS_KEY, 0) - */ -#define BTRFS_QGROUP_STATUS_KEY 240 -/* - * Records the currently used space of the qgroup. - * One key per qgroup, (0, BTRFS_QGROUP_INFO_KEY, qgroupid). - */ -#define BTRFS_QGROUP_INFO_KEY 242 -/* - * Contains the user configured limits for the qgroup. - * One key per qgroup, (0, BTRFS_QGROUP_LIMIT_KEY, qgroupid). - */ -#define BTRFS_QGROUP_LIMIT_KEY 244 -/* - * Records the child-parent relationship of qgroups. For - * each relation, 2 keys are present: - * (childid, BTRFS_QGROUP_RELATION_KEY, parentid) - * (parentid, BTRFS_QGROUP_RELATION_KEY, childid) - */ -#define BTRFS_QGROUP_RELATION_KEY 246 - -/* - * Obsolete name, see BTRFS_TEMPORARY_ITEM_KEY. - */ -#define BTRFS_BALANCE_ITEM_KEY 248 - -/* - * The key type for tree items that are stored persistently, but do not need to - * exist for extended period of time. The items can exist in any tree. - * - * [subtype, BTRFS_TEMPORARY_ITEM_KEY, data] - * - * Existing items: - * - * - balance status item - * (BTRFS_BALANCE_OBJECTID, BTRFS_TEMPORARY_ITEM_KEY, 0) - */ -#define BTRFS_TEMPORARY_ITEM_KEY 248 - -/* - * Obsolete name, see BTRFS_PERSISTENT_ITEM_KEY - */ -#define BTRFS_DEV_STATS_KEY 249 - -/* - * The key type for tree items that are stored persistently and usually exist - * for a long period, eg. filesystem lifetime. The item kinds can be status - * information, stats or preference values. The item can exist in any tree. - * - * [subtype, BTRFS_PERSISTENT_ITEM_KEY, data] - * - * Existing items: - * - * - device statistics, store IO stats in the device tree, one key for all - * stats - * (BTRFS_DEV_STATS_OBJECTID, BTRFS_DEV_STATS_KEY, 0) - */ -#define BTRFS_PERSISTENT_ITEM_KEY 249 - -/* - * Persistantly stores the device replace state in the device tree. - * The key is built like this: (0, BTRFS_DEV_REPLACE_KEY, 0). - */ -#define BTRFS_DEV_REPLACE_KEY 250 - -/* - * Stores items that allow to quickly map UUIDs to something else. - * These items are part of the filesystem UUID tree. - * The key is built like this: - * (UUID_upper_64_bits, BTRFS_UUID_KEY*, UUID_lower_64_bits). - */ -#if BTRFS_UUID_SIZE != 16 -#error "UUID items require BTRFS_UUID_SIZE == 16!" -#endif -#define BTRFS_UUID_KEY_SUBVOL 251 /* for UUIDs assigned to subvols */ -#define BTRFS_UUID_KEY_RECEIVED_SUBVOL 252 /* for UUIDs assigned to - * received subvols */ - -/* - * string items are for debugging. They just store a short string of - * data in the FS - */ -#define BTRFS_STRING_ITEM_KEY 253 - /* * Flags for mount options. * diff --git a/include/uapi/linux/btrfs_tree.h b/include/uapi/linux/btrfs_tree.h new file mode 100644 index 000000000000..1e875057d6b2 --- /dev/null +++ b/include/uapi/linux/btrfs_tree.h @@ -0,0 +1,966 @@ +#ifndef _BTRFS_CTREE_H_ +#define _BTRFS_CTREE_H_ + +/* + * This header contains the structure definitions and constants used + * by file system objects that can be retrieved using + * the BTRFS_IOC_SEARCH_TREE ioctl. That means basically anything that + * is needed to describe a leaf node's key or item contents. + */ + +/* holds pointers to all of the tree roots */ +#define BTRFS_ROOT_TREE_OBJECTID 1ULL + +/* stores information about which extents are in use, and reference counts */ +#define BTRFS_EXTENT_TREE_OBJECTID 2ULL + +/* + * chunk tree stores translations from logical -> physical block numbering + * the super block points to the chunk tree + */ +#define BTRFS_CHUNK_TREE_OBJECTID 3ULL + +/* + * stores information about which areas of a given device are in use. + * one per device. The tree of tree roots points to the device tree + */ +#define BTRFS_DEV_TREE_OBJECTID 4ULL + +/* one per subvolume, storing files and directories */ +#define BTRFS_FS_TREE_OBJECTID 5ULL + +/* directory objectid inside the root tree */ +#define BTRFS_ROOT_TREE_DIR_OBJECTID 6ULL + +/* holds checksums of all the data extents */ +#define BTRFS_CSUM_TREE_OBJECTID 7ULL + +/* holds quota configuration and tracking */ +#define BTRFS_QUOTA_TREE_OBJECTID 8ULL + +/* for storing items that use the BTRFS_UUID_KEY* types */ +#define BTRFS_UUID_TREE_OBJECTID 9ULL + +/* tracks free space in block groups. */ +#define BTRFS_FREE_SPACE_TREE_OBJECTID 10ULL + +/* device stats in the device tree */ +#define BTRFS_DEV_STATS_OBJECTID 0ULL + +/* for storing balance parameters in the root tree */ +#define BTRFS_BALANCE_OBJECTID -4ULL + +/* orhpan objectid for tracking unlinked/truncated files */ +#define BTRFS_ORPHAN_OBJECTID -5ULL + +/* does write ahead logging to speed up fsyncs */ +#define BTRFS_TREE_LOG_OBJECTID -6ULL +#define BTRFS_TREE_LOG_FIXUP_OBJECTID -7ULL + +/* for space balancing */ +#define BTRFS_TREE_RELOC_OBJECTID -8ULL +#define BTRFS_DATA_RELOC_TREE_OBJECTID -9ULL + +/* + * extent checksums all have this objectid + * this allows them to share the logging tree + * for fsyncs + */ +#define BTRFS_EXTENT_CSUM_OBJECTID -10ULL + +/* For storing free space cache */ +#define BTRFS_FREE_SPACE_OBJECTID -11ULL + +/* + * The inode number assigned to the special inode for storing + * free ino cache + */ +#define BTRFS_FREE_INO_OBJECTID -12ULL + +/* dummy objectid represents multiple objectids */ +#define BTRFS_MULTIPLE_OBJECTIDS -255ULL + +/* + * All files have objectids in this range. + */ +#define BTRFS_FIRST_FREE_OBJECTID 256ULL +#define BTRFS_LAST_FREE_OBJECTID -256ULL +#define BTRFS_FIRST_CHUNK_TREE_OBJECTID 256ULL + + +/* + * the device items go into the chunk tree. The key is in the form + * [ 1 BTRFS_DEV_ITEM_KEY device_id ] + */ +#define BTRFS_DEV_ITEMS_OBJECTID 1ULL + +#define BTRFS_BTREE_INODE_OBJECTID 1 + +#define BTRFS_EMPTY_SUBVOL_DIR_OBJECTID 2 + +#define BTRFS_DEV_REPLACE_DEVID 0ULL + +/* + * inode items have the data typically returned from stat and store other + * info about object characteristics. There is one for every file and dir in + * the FS + */ +#define BTRFS_INODE_ITEM_KEY 1 +#define BTRFS_INODE_REF_KEY 12 +#define BTRFS_INODE_EXTREF_KEY 13 +#define BTRFS_XATTR_ITEM_KEY 24 +#define BTRFS_ORPHAN_ITEM_KEY 48 +/* reserve 2-15 close to the inode for later flexibility */ + +/* + * dir items are the name -> inode pointers in a directory. There is one + * for every name in a directory. + */ +#define BTRFS_DIR_LOG_ITEM_KEY 60 +#define BTRFS_DIR_LOG_INDEX_KEY 72 +#define BTRFS_DIR_ITEM_KEY 84 +#define BTRFS_DIR_INDEX_KEY 96 +/* + * extent data is for file data + */ +#define BTRFS_EXTENT_DATA_KEY 108 + +/* + * extent csums are stored in a separate tree and hold csums for + * an entire extent on disk. + */ +#define BTRFS_EXTENT_CSUM_KEY 128 + +/* + * root items point to tree roots. They are typically in the root + * tree used by the super block to find all the other trees + */ +#define BTRFS_ROOT_ITEM_KEY 132 + +/* + * root backrefs tie subvols and snapshots to the directory entries that + * reference them + */ +#define BTRFS_ROOT_BACKREF_KEY 144 + +/* + * root refs make a fast index for listing all of the snapshots and + * subvolumes referenced by a given root. They point directly to the + * directory item in the root that references the subvol + */ +#define BTRFS_ROOT_REF_KEY 156 + +/* + * extent items are in the extent map tree. These record which blocks + * are used, and how many references there are to each block + */ +#define BTRFS_EXTENT_ITEM_KEY 168 + +/* + * The same as the BTRFS_EXTENT_ITEM_KEY, except it's metadata we already know + * the length, so we save the level in key->offset instead of the length. + */ +#define BTRFS_METADATA_ITEM_KEY 169 + +#define BTRFS_TREE_BLOCK_REF_KEY 176 + +#define BTRFS_EXTENT_DATA_REF_KEY 178 + +#define BTRFS_EXTENT_REF_V0_KEY 180 + +#define BTRFS_SHARED_BLOCK_REF_KEY 182 + +#define BTRFS_SHARED_DATA_REF_KEY 184 + +/* + * block groups give us hints into the extent allocation trees. Which + * blocks are free etc etc + */ +#define BTRFS_BLOCK_GROUP_ITEM_KEY 192 + +/* + * Every block group is represented in the free space tree by a free space info + * item, which stores some accounting information. It is keyed on + * (block_group_start, FREE_SPACE_INFO, block_group_length). + */ +#define BTRFS_FREE_SPACE_INFO_KEY 198 + +/* + * A free space extent tracks an extent of space that is free in a block group. + * It is keyed on (start, FREE_SPACE_EXTENT, length). + */ +#define BTRFS_FREE_SPACE_EXTENT_KEY 199 + +/* + * When a block group becomes very fragmented, we convert it to use bitmaps + * instead of extents. A free space bitmap is keyed on + * (start, FREE_SPACE_BITMAP, length); the corresponding item is a bitmap with + * (length / sectorsize) bits. + */ +#define BTRFS_FREE_SPACE_BITMAP_KEY 200 + +#define BTRFS_DEV_EXTENT_KEY 204 +#define BTRFS_DEV_ITEM_KEY 216 +#define BTRFS_CHUNK_ITEM_KEY 228 + +/* + * Records the overall state of the qgroups. + * There's only one instance of this key present, + * (0, BTRFS_QGROUP_STATUS_KEY, 0) + */ +#define BTRFS_QGROUP_STATUS_KEY 240 +/* + * Records the currently used space of the qgroup. + * One key per qgroup, (0, BTRFS_QGROUP_INFO_KEY, qgroupid). + */ +#define BTRFS_QGROUP_INFO_KEY 242 +/* + * Contains the user configured limits for the qgroup. + * One key per qgroup, (0, BTRFS_QGROUP_LIMIT_KEY, qgroupid). + */ +#define BTRFS_QGROUP_LIMIT_KEY 244 +/* + * Records the child-parent relationship of qgroups. For + * each relation, 2 keys are present: + * (childid, BTRFS_QGROUP_RELATION_KEY, parentid) + * (parentid, BTRFS_QGROUP_RELATION_KEY, childid) + */ +#define BTRFS_QGROUP_RELATION_KEY 246 + +/* + * Obsolete name, see BTRFS_TEMPORARY_ITEM_KEY. + */ +#define BTRFS_BALANCE_ITEM_KEY 248 + +/* + * The key type for tree items that are stored persistently, but do not need to + * exist for extended period of time. The items can exist in any tree. + * + * [subtype, BTRFS_TEMPORARY_ITEM_KEY, data] + * + * Existing items: + * + * - balance status item + * (BTRFS_BALANCE_OBJECTID, BTRFS_TEMPORARY_ITEM_KEY, 0) + */ +#define BTRFS_TEMPORARY_ITEM_KEY 248 + +/* + * Obsolete name, see BTRFS_PERSISTENT_ITEM_KEY + */ +#define BTRFS_DEV_STATS_KEY 249 + +/* + * The key type for tree items that are stored persistently and usually exist + * for a long period, eg. filesystem lifetime. The item kinds can be status + * information, stats or preference values. The item can exist in any tree. + * + * [subtype, BTRFS_PERSISTENT_ITEM_KEY, data] + * + * Existing items: + * + * - device statistics, store IO stats in the device tree, one key for all + * stats + * (BTRFS_DEV_STATS_OBJECTID, BTRFS_DEV_STATS_KEY, 0) + */ +#define BTRFS_PERSISTENT_ITEM_KEY 249 + +/* + * Persistantly stores the device replace state in the device tree. + * The key is built like this: (0, BTRFS_DEV_REPLACE_KEY, 0). + */ +#define BTRFS_DEV_REPLACE_KEY 250 + +/* + * Stores items that allow to quickly map UUIDs to something else. + * These items are part of the filesystem UUID tree. + * The key is built like this: + * (UUID_upper_64_bits, BTRFS_UUID_KEY*, UUID_lower_64_bits). + */ +#if BTRFS_UUID_SIZE != 16 +#error "UUID items require BTRFS_UUID_SIZE == 16!" +#endif +#define BTRFS_UUID_KEY_SUBVOL 251 /* for UUIDs assigned to subvols */ +#define BTRFS_UUID_KEY_RECEIVED_SUBVOL 252 /* for UUIDs assigned to + * received subvols */ + +/* + * string items are for debugging. They just store a short string of + * data in the FS + */ +#define BTRFS_STRING_ITEM_KEY 253 + + + +/* 32 bytes in various csum fields */ +#define BTRFS_CSUM_SIZE 32 + +/* csum types */ +#define BTRFS_CSUM_TYPE_CRC32 0 + +/* + * flags definitions for directory entry item type + * + * Used by: + * struct btrfs_dir_item.type + */ +#define BTRFS_FT_UNKNOWN 0 +#define BTRFS_FT_REG_FILE 1 +#define BTRFS_FT_DIR 2 +#define BTRFS_FT_CHRDEV 3 +#define BTRFS_FT_BLKDEV 4 +#define BTRFS_FT_FIFO 5 +#define BTRFS_FT_SOCK 6 +#define BTRFS_FT_SYMLINK 7 +#define BTRFS_FT_XATTR 8 +#define BTRFS_FT_MAX 9 + +/* + * The key defines the order in the tree, and so it also defines (optimal) + * block layout. + * + * objectid corresponds to the inode number. + * + * type tells us things about the object, and is a kind of stream selector. + * so for a given inode, keys with type of 1 might refer to the inode data, + * type of 2 may point to file data in the btree and type == 3 may point to + * extents. + * + * offset is the starting byte offset for this key in the stream. + * + * btrfs_disk_key is in disk byte order. struct btrfs_key is always + * in cpu native order. Otherwise they are identical and their sizes + * should be the same (ie both packed) + */ +struct btrfs_disk_key { + __le64 objectid; + u8 type; + __le64 offset; +} __attribute__ ((__packed__)); + +struct btrfs_key { + u64 objectid; + u8 type; + u64 offset; +} __attribute__ ((__packed__)); + +struct btrfs_dev_item { + /* the internal btrfs device id */ + __le64 devid; + + /* size of the device */ + __le64 total_bytes; + + /* bytes used */ + __le64 bytes_used; + + /* optimal io alignment for this device */ + __le32 io_align; + + /* optimal io width for this device */ + __le32 io_width; + + /* minimal io size for this device */ + __le32 sector_size; + + /* type and info about this device */ + __le64 type; + + /* expected generation for this device */ + __le64 generation; + + /* + * starting byte of this partition on the device, + * to allow for stripe alignment in the future + */ + __le64 start_offset; + + /* grouping information for allocation decisions */ + __le32 dev_group; + + /* seek speed 0-100 where 100 is fastest */ + u8 seek_speed; + + /* bandwidth 0-100 where 100 is fastest */ + u8 bandwidth; + + /* btrfs generated uuid for this device */ + u8 uuid[BTRFS_UUID_SIZE]; + + /* uuid of FS who owns this device */ + u8 fsid[BTRFS_UUID_SIZE]; +} __attribute__ ((__packed__)); + +struct btrfs_stripe { + __le64 devid; + __le64 offset; + u8 dev_uuid[BTRFS_UUID_SIZE]; +} __attribute__ ((__packed__)); + +struct btrfs_chunk { + /* size of this chunk in bytes */ + __le64 length; + + /* objectid of the root referencing this chunk */ + __le64 owner; + + __le64 stripe_len; + __le64 type; + + /* optimal io alignment for this chunk */ + __le32 io_align; + + /* optimal io width for this chunk */ + __le32 io_width; + + /* minimal io size for this chunk */ + __le32 sector_size; + + /* 2^16 stripes is quite a lot, a second limit is the size of a single + * item in the btree + */ + __le16 num_stripes; + + /* sub stripes only matter for raid10 */ + __le16 sub_stripes; + struct btrfs_stripe stripe; + /* additional stripes go here */ +} __attribute__ ((__packed__)); + +#define BTRFS_FREE_SPACE_EXTENT 1 +#define BTRFS_FREE_SPACE_BITMAP 2 + +struct btrfs_free_space_entry { + __le64 offset; + __le64 bytes; + u8 type; +} __attribute__ ((__packed__)); + +struct btrfs_free_space_header { + struct btrfs_disk_key location; + __le64 generation; + __le64 num_entries; + __le64 num_bitmaps; +} __attribute__ ((__packed__)); + +#define BTRFS_HEADER_FLAG_WRITTEN (1ULL << 0) +#define BTRFS_HEADER_FLAG_RELOC (1ULL << 1) + +/* Super block flags */ +/* Errors detected */ +#define BTRFS_SUPER_FLAG_ERROR (1ULL << 2) + +#define BTRFS_SUPER_FLAG_SEEDING (1ULL << 32) +#define BTRFS_SUPER_FLAG_METADUMP (1ULL << 33) + + +/* + * items in the extent btree are used to record the objectid of the + * owner of the block and the number of references + */ + +struct btrfs_extent_item { + __le64 refs; + __le64 generation; + __le64 flags; +} __attribute__ ((__packed__)); + +struct btrfs_extent_item_v0 { + __le32 refs; +} __attribute__ ((__packed__)); + + +#define BTRFS_EXTENT_FLAG_DATA (1ULL << 0) +#define BTRFS_EXTENT_FLAG_TREE_BLOCK (1ULL << 1) + +/* following flags only apply to tree blocks */ + +/* use full backrefs for extent pointers in the block */ +#define BTRFS_BLOCK_FLAG_FULL_BACKREF (1ULL << 8) + +/* + * this flag is only used internally by scrub and may be changed at any time + * it is only declared here to avoid collisions + */ +#define BTRFS_EXTENT_FLAG_SUPER (1ULL << 48) + +struct btrfs_tree_block_info { + struct btrfs_disk_key key; + u8 level; +} __attribute__ ((__packed__)); + +struct btrfs_extent_data_ref { + __le64 root; + __le64 objectid; + __le64 offset; + __le32 count; +} __attribute__ ((__packed__)); + +struct btrfs_shared_data_ref { + __le32 count; +} __attribute__ ((__packed__)); + +struct btrfs_extent_inline_ref { + u8 type; + __le64 offset; +} __attribute__ ((__packed__)); + +/* old style backrefs item */ +struct btrfs_extent_ref_v0 { + __le64 root; + __le64 generation; + __le64 objectid; + __le32 count; +} __attribute__ ((__packed__)); + + +/* dev extents record free space on individual devices. The owner + * field points back to the chunk allocation mapping tree that allocated + * the extent. The chunk tree uuid field is a way to double check the owner + */ +struct btrfs_dev_extent { + __le64 chunk_tree; + __le64 chunk_objectid; + __le64 chunk_offset; + __le64 length; + u8 chunk_tree_uuid[BTRFS_UUID_SIZE]; +} __attribute__ ((__packed__)); + +struct btrfs_inode_ref { + __le64 index; + __le16 name_len; + /* name goes here */ +} __attribute__ ((__packed__)); + +struct btrfs_inode_extref { + __le64 parent_objectid; + __le64 index; + __le16 name_len; + __u8 name[0]; + /* name goes here */ +} __attribute__ ((__packed__)); + +struct btrfs_timespec { + __le64 sec; + __le32 nsec; +} __attribute__ ((__packed__)); + +struct btrfs_inode_item { + /* nfs style generation number */ + __le64 generation; + /* transid that last touched this inode */ + __le64 transid; + __le64 size; + __le64 nbytes; + __le64 block_group; + __le32 nlink; + __le32 uid; + __le32 gid; + __le32 mode; + __le64 rdev; + __le64 flags; + + /* modification sequence number for NFS */ + __le64 sequence; + + /* + * a little future expansion, for more than this we can + * just grow the inode item and version it + */ + __le64 reserved[4]; + struct btrfs_timespec atime; + struct btrfs_timespec ctime; + struct btrfs_timespec mtime; + struct btrfs_timespec otime; +} __attribute__ ((__packed__)); + +struct btrfs_dir_log_item { + __le64 end; +} __attribute__ ((__packed__)); + +struct btrfs_dir_item { + struct btrfs_disk_key location; + __le64 transid; + __le16 data_len; + __le16 name_len; + u8 type; +} __attribute__ ((__packed__)); + +#define BTRFS_ROOT_SUBVOL_RDONLY (1ULL << 0) + +/* + * Internal in-memory flag that a subvolume has been marked for deletion but + * still visible as a directory + */ +#define BTRFS_ROOT_SUBVOL_DEAD (1ULL << 48) + +struct btrfs_root_item { + struct btrfs_inode_item inode; + __le64 generation; + __le64 root_dirid; + __le64 bytenr; + __le64 byte_limit; + __le64 bytes_used; + __le64 last_snapshot; + __le64 flags; + __le32 refs; + struct btrfs_disk_key drop_progress; + u8 drop_level; + u8 level; + + /* + * The following fields appear after subvol_uuids+subvol_times + * were introduced. + */ + + /* + * This generation number is used to test if the new fields are valid + * and up to date while reading the root item. Every time the root item + * is written out, the "generation" field is copied into this field. If + * anyone ever mounted the fs with an older kernel, we will have + * mismatching generation values here and thus must invalidate the + * new fields. See btrfs_update_root and btrfs_find_last_root for + * details. + * the offset of generation_v2 is also used as the start for the memset + * when invalidating the fields. + */ + __le64 generation_v2; + u8 uuid[BTRFS_UUID_SIZE]; + u8 parent_uuid[BTRFS_UUID_SIZE]; + u8 received_uuid[BTRFS_UUID_SIZE]; + __le64 ctransid; /* updated when an inode changes */ + __le64 otransid; /* trans when created */ + __le64 stransid; /* trans when sent. non-zero for received subvol */ + __le64 rtransid; /* trans when received. non-zero for received subvol */ + struct btrfs_timespec ctime; + struct btrfs_timespec otime; + struct btrfs_timespec stime; + struct btrfs_timespec rtime; + __le64 reserved[8]; /* for future */ +} __attribute__ ((__packed__)); + +/* + * this is used for both forward and backward root refs + */ +struct btrfs_root_ref { + __le64 dirid; + __le64 sequence; + __le16 name_len; +} __attribute__ ((__packed__)); + +struct btrfs_disk_balance_args { + /* + * profiles to operate on, single is denoted by + * BTRFS_AVAIL_ALLOC_BIT_SINGLE + */ + __le64 profiles; + + /* + * usage filter + * BTRFS_BALANCE_ARGS_USAGE with a single value means '0..N' + * BTRFS_BALANCE_ARGS_USAGE_RANGE - range syntax, min..max + */ + union { + __le64 usage; + struct { + __le32 usage_min; + __le32 usage_max; + }; + }; + + /* devid filter */ + __le64 devid; + + /* devid subset filter [pstart..pend) */ + __le64 pstart; + __le64 pend; + + /* btrfs virtual address space subset filter [vstart..vend) */ + __le64 vstart; + __le64 vend; + + /* + * profile to convert to, single is denoted by + * BTRFS_AVAIL_ALLOC_BIT_SINGLE + */ + __le64 target; + + /* BTRFS_BALANCE_ARGS_* */ + __le64 flags; + + /* + * BTRFS_BALANCE_ARGS_LIMIT with value 'limit' + * BTRFS_BALANCE_ARGS_LIMIT_RANGE - the extend version can use minimum + * and maximum + */ + union { + __le64 limit; + struct { + __le32 limit_min; + __le32 limit_max; + }; + }; + + /* + * Process chunks that cross stripes_min..stripes_max devices, + * BTRFS_BALANCE_ARGS_STRIPES_RANGE + */ + __le32 stripes_min; + __le32 stripes_max; + + __le64 unused[6]; +} __attribute__ ((__packed__)); + +/* + * store balance parameters to disk so that balance can be properly + * resumed after crash or unmount + */ +struct btrfs_balance_item { + /* BTRFS_BALANCE_* */ + __le64 flags; + + struct btrfs_disk_balance_args data; + struct btrfs_disk_balance_args meta; + struct btrfs_disk_balance_args sys; + + __le64 unused[4]; +} __attribute__ ((__packed__)); + +#define BTRFS_FILE_EXTENT_INLINE 0 +#define BTRFS_FILE_EXTENT_REG 1 +#define BTRFS_FILE_EXTENT_PREALLOC 2 + +struct btrfs_file_extent_item { + /* + * transaction id that created this extent + */ + __le64 generation; + /* + * max number of bytes to hold this extent in ram + * when we split a compressed extent we can't know how big + * each of the resulting pieces will be. So, this is + * an upper limit on the size of the extent in ram instead of + * an exact limit. + */ + __le64 ram_bytes; + + /* + * 32 bits for the various ways we might encode the data, + * including compression and encryption. If any of these + * are set to something a given disk format doesn't understand + * it is treated like an incompat flag for reading and writing, + * but not for stat. + */ + u8 compression; + u8 encryption; + __le16 other_encoding; /* spare for later use */ + + /* are we inline data or a real extent? */ + u8 type; + + /* + * disk space consumed by the extent, checksum blocks are included + * in these numbers + * + * At this offset in the structure, the inline extent data start. + */ + __le64 disk_bytenr; + __le64 disk_num_bytes; + /* + * the logical offset in file blocks (no csums) + * this extent record is for. This allows a file extent to point + * into the middle of an existing extent on disk, sharing it + * between two snapshots (useful if some bytes in the middle of the + * extent have changed + */ + __le64 offset; + /* + * the logical number of file blocks (no csums included). This + * always reflects the size uncompressed and without encoding. + */ + __le64 num_bytes; + +} __attribute__ ((__packed__)); + +struct btrfs_csum_item { + u8 csum; +} __attribute__ ((__packed__)); + +struct btrfs_dev_stats_item { + /* + * grow this item struct at the end for future enhancements and keep + * the existing values unchanged + */ + __le64 values[BTRFS_DEV_STAT_VALUES_MAX]; +} __attribute__ ((__packed__)); + +#define BTRFS_DEV_REPLACE_ITEM_CONT_READING_FROM_SRCDEV_MODE_ALWAYS 0 +#define BTRFS_DEV_REPLACE_ITEM_CONT_READING_FROM_SRCDEV_MODE_AVOID 1 +#define BTRFS_DEV_REPLACE_ITEM_STATE_NEVER_STARTED 0 +#define BTRFS_DEV_REPLACE_ITEM_STATE_STARTED 1 +#define BTRFS_DEV_REPLACE_ITEM_STATE_SUSPENDED 2 +#define BTRFS_DEV_REPLACE_ITEM_STATE_FINISHED 3 +#define BTRFS_DEV_REPLACE_ITEM_STATE_CANCELED 4 + +struct btrfs_dev_replace_item { + /* + * grow this item struct at the end for future enhancements and keep + * the existing values unchanged + */ + __le64 src_devid; + __le64 cursor_left; + __le64 cursor_right; + __le64 cont_reading_from_srcdev_mode; + + __le64 replace_state; + __le64 time_started; + __le64 time_stopped; + __le64 num_write_errors; + __le64 num_uncorrectable_read_errors; +} __attribute__ ((__packed__)); + +/* different types of block groups (and chunks) */ +#define BTRFS_BLOCK_GROUP_DATA (1ULL << 0) +#define BTRFS_BLOCK_GROUP_SYSTEM (1ULL << 1) +#define BTRFS_BLOCK_GROUP_METADATA (1ULL << 2) +#define BTRFS_BLOCK_GROUP_RAID0 (1ULL << 3) +#define BTRFS_BLOCK_GROUP_RAID1 (1ULL << 4) +#define BTRFS_BLOCK_GROUP_DUP (1ULL << 5) +#define BTRFS_BLOCK_GROUP_RAID10 (1ULL << 6) +#define BTRFS_BLOCK_GROUP_RAID5 (1ULL << 7) +#define BTRFS_BLOCK_GROUP_RAID6 (1ULL << 8) +#define BTRFS_BLOCK_GROUP_RESERVED (BTRFS_AVAIL_ALLOC_BIT_SINGLE | \ + BTRFS_SPACE_INFO_GLOBAL_RSV) + +enum btrfs_raid_types { + BTRFS_RAID_RAID10, + BTRFS_RAID_RAID1, + BTRFS_RAID_DUP, + BTRFS_RAID_RAID0, + BTRFS_RAID_SINGLE, + BTRFS_RAID_RAID5, + BTRFS_RAID_RAID6, + BTRFS_NR_RAID_TYPES +}; + +#define BTRFS_BLOCK_GROUP_TYPE_MASK (BTRFS_BLOCK_GROUP_DATA | \ + BTRFS_BLOCK_GROUP_SYSTEM | \ + BTRFS_BLOCK_GROUP_METADATA) + +#define BTRFS_BLOCK_GROUP_PROFILE_MASK (BTRFS_BLOCK_GROUP_RAID0 | \ + BTRFS_BLOCK_GROUP_RAID1 | \ + BTRFS_BLOCK_GROUP_RAID5 | \ + BTRFS_BLOCK_GROUP_RAID6 | \ + BTRFS_BLOCK_GROUP_DUP | \ + BTRFS_BLOCK_GROUP_RAID10) +#define BTRFS_BLOCK_GROUP_RAID56_MASK (BTRFS_BLOCK_GROUP_RAID5 | \ + BTRFS_BLOCK_GROUP_RAID6) + +/* + * We need a bit for restriper to be able to tell when chunks of type + * SINGLE are available. This "extended" profile format is used in + * fs_info->avail_*_alloc_bits (in-memory) and balance item fields + * (on-disk). The corresponding on-disk bit in chunk.type is reserved + * to avoid remappings between two formats in future. + */ +#define BTRFS_AVAIL_ALLOC_BIT_SINGLE (1ULL << 48) + +/* + * A fake block group type that is used to communicate global block reserve + * size to userspace via the SPACE_INFO ioctl. + */ +#define BTRFS_SPACE_INFO_GLOBAL_RSV (1ULL << 49) + +#define BTRFS_EXTENDED_PROFILE_MASK (BTRFS_BLOCK_GROUP_PROFILE_MASK | \ + BTRFS_AVAIL_ALLOC_BIT_SINGLE) + +static inline u64 chunk_to_extended(u64 flags) +{ + if ((flags & BTRFS_BLOCK_GROUP_PROFILE_MASK) == 0) + flags |= BTRFS_AVAIL_ALLOC_BIT_SINGLE; + + return flags; +} +static inline u64 extended_to_chunk(u64 flags) +{ + return flags & ~BTRFS_AVAIL_ALLOC_BIT_SINGLE; +} + +struct btrfs_block_group_item { + __le64 used; + __le64 chunk_objectid; + __le64 flags; +} __attribute__ ((__packed__)); + +struct btrfs_free_space_info { + __le32 extent_count; + __le32 flags; +} __attribute__ ((__packed__)); + +#define BTRFS_FREE_SPACE_USING_BITMAPS (1ULL << 0) + +#define BTRFS_QGROUP_LEVEL_SHIFT 48 +static inline u64 btrfs_qgroup_level(u64 qgroupid) +{ + return qgroupid >> BTRFS_QGROUP_LEVEL_SHIFT; +} + +/* + * is subvolume quota turned on? + */ +#define BTRFS_QGROUP_STATUS_FLAG_ON (1ULL << 0) +/* + * RESCAN is set during the initialization phase + */ +#define BTRFS_QGROUP_STATUS_FLAG_RESCAN (1ULL << 1) +/* + * Some qgroup entries are known to be out of date, + * either because the configuration has changed in a way that + * makes a rescan necessary, or because the fs has been mounted + * with a non-qgroup-aware version. + * Turning qouta off and on again makes it inconsistent, too. + */ +#define BTRFS_QGROUP_STATUS_FLAG_INCONSISTENT (1ULL << 2) + +#define BTRFS_QGROUP_STATUS_VERSION 1 + +struct btrfs_qgroup_status_item { + __le64 version; + /* + * the generation is updated during every commit. As older + * versions of btrfs are not aware of qgroups, it will be + * possible to detect inconsistencies by checking the + * generation on mount time + */ + __le64 generation; + + /* flag definitions see above */ + __le64 flags; + + /* + * only used during scanning to record the progress + * of the scan. It contains a logical address + */ + __le64 rescan; +} __attribute__ ((__packed__)); + +struct btrfs_qgroup_info_item { + __le64 generation; + __le64 rfer; + __le64 rfer_cmpr; + __le64 excl; + __le64 excl_cmpr; +} __attribute__ ((__packed__)); + +struct btrfs_qgroup_limit_item { + /* + * only updated when any of the other values change + */ + __le64 flags; + __le64 max_rfer; + __le64 max_excl; + __le64 rsv_rfer; + __le64 rsv_excl; +} __attribute__ ((__packed__)); + +#endif /* _BTRFS_CTREE_H_ */ -- cgit v1.3-8-gc7d7 From ceeb0ae7bf42a3deaaaee981d2da02e5d3ad2b0f Mon Sep 17 00:00:00 2001 From: David Sterba Date: Tue, 26 Apr 2016 23:54:39 +0200 Subject: btrfs: sink gfp parameter to set_extent_bits All callers pass GFP_NOFS. Signed-off-by: David Sterba --- fs/btrfs/extent-tree.c | 4 ++-- fs/btrfs/extent_io.c | 5 ++--- fs/btrfs/extent_io.h | 4 ++-- fs/btrfs/file-item.c | 2 +- fs/btrfs/relocation.c | 4 ++-- fs/btrfs/scrub.c | 2 +- 6 files changed, 10 insertions(+), 11 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 84e060eb0de8..0cfddbe8abba 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -231,9 +231,9 @@ static int add_excluded_extent(struct btrfs_root *root, { u64 end = start + num_bytes - 1; set_extent_bits(&root->fs_info->freed_extents[0], - start, end, EXTENT_UPTODATE, GFP_NOFS); + start, end, EXTENT_UPTODATE); set_extent_bits(&root->fs_info->freed_extents[1], - start, end, EXTENT_UPTODATE, GFP_NOFS); + start, end, EXTENT_UPTODATE); return 0; } diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index d247fc0eea19..3b53d217066c 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -2232,13 +2232,12 @@ int btrfs_get_io_failure_record(struct inode *inode, u64 start, u64 end, /* set the bits in the private failure tree */ ret = set_extent_bits(failure_tree, start, end, - EXTENT_LOCKED | EXTENT_DIRTY, GFP_NOFS); + EXTENT_LOCKED | EXTENT_DIRTY); if (ret >= 0) ret = set_state_failrec(failure_tree, start, failrec); /* set the bits in the inode's tree */ if (ret >= 0) - ret = set_extent_bits(tree, start, end, EXTENT_DAMAGED, - GFP_NOFS); + ret = set_extent_bits(tree, start, end, EXTENT_DAMAGED); if (ret < 0) { kfree(failrec); return ret; diff --git a/fs/btrfs/extent_io.h b/fs/btrfs/extent_io.h index b5e0ade90e88..447c6e6ff88c 100644 --- a/fs/btrfs/extent_io.h +++ b/fs/btrfs/extent_io.h @@ -259,9 +259,9 @@ int set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, struct extent_state **cached_state, gfp_t mask); static inline int set_extent_bits(struct extent_io_tree *tree, u64 start, - u64 end, unsigned bits, gfp_t mask) + u64 end, unsigned bits) { - return set_extent_bit(tree, start, end, bits, NULL, NULL, mask); + return set_extent_bit(tree, start, end, bits, NULL, NULL, GFP_NOFS); } static inline int clear_extent_uptodate(struct extent_io_tree *tree, u64 start, diff --git a/fs/btrfs/file-item.c b/fs/btrfs/file-item.c index 7a7d6e253cfc..62a81ee13a5f 100644 --- a/fs/btrfs/file-item.c +++ b/fs/btrfs/file-item.c @@ -248,7 +248,7 @@ static int __btrfs_lookup_bio_sums(struct btrfs_root *root, BTRFS_DATA_RELOC_TREE_OBJECTID) { set_extent_bits(io_tree, offset, offset + root->sectorsize - 1, - EXTENT_NODATASUM, GFP_NOFS); + EXTENT_NODATASUM); } else { btrfs_info(BTRFS_I(inode)->root->fs_info, "no csum found for inode %llu start %llu", diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c index 08ef890deca6..78cbfb530de6 100644 --- a/fs/btrfs/relocation.c +++ b/fs/btrfs/relocation.c @@ -2814,7 +2814,7 @@ static void mark_block_processed(struct reloc_control *rc, u64 bytenr, u32 blocksize) { set_extent_bits(&rc->processed_blocks, bytenr, bytenr + blocksize - 1, - EXTENT_DIRTY, GFP_NOFS); + EXTENT_DIRTY); } static void __mark_block_processed(struct reloc_control *rc, @@ -3182,7 +3182,7 @@ static int relocate_file_extent_cluster(struct inode *inode, page_start + offset == cluster->boundary[nr]) { set_extent_bits(&BTRFS_I(inode)->io_tree, page_start, page_end, - EXTENT_BOUNDARY, GFP_NOFS); + EXTENT_BOUNDARY); nr++; } diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c index 4678f03e878e..9d24375eed1f 100644 --- a/fs/btrfs/scrub.c +++ b/fs/btrfs/scrub.c @@ -745,7 +745,7 @@ static int scrub_fixup_readpage(u64 inum, u64 offset, u64 root, void *fixup_ctx) * sure we read the bad mirror. */ ret = set_extent_bits(&BTRFS_I(inode)->io_tree, offset, end, - EXTENT_DAMAGED, GFP_NOFS); + EXTENT_DAMAGED); if (ret) { /* set_extent_bits should give proper error */ WARN_ON(ret > 0); -- cgit v1.3-8-gc7d7 From 91166212e0ffbb4db50aa8a238548e967008b33d Mon Sep 17 00:00:00 2001 From: David Sterba Date: Tue, 26 Apr 2016 23:54:39 +0200 Subject: btrfs: sink gfp parameter to clear_extent_bits Callers pass GFP_NOFS and GFP_KERNEL. No need to pass the flags around. Signed-off-by: David Sterba --- fs/btrfs/disk-io.c | 2 +- fs/btrfs/extent-tree.c | 8 ++++---- fs/btrfs/extent_io.c | 4 ++-- fs/btrfs/extent_io.h | 5 +++-- fs/btrfs/inode.c | 3 +-- fs/btrfs/relocation.c | 3 +-- fs/btrfs/scrub.c | 2 +- fs/btrfs/tests/extent-io-tests.c | 2 +- fs/btrfs/tree-log.c | 2 +- 9 files changed, 15 insertions(+), 16 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 4e47849d7427..3bc88b4b0032 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -4357,7 +4357,7 @@ static int btrfs_destroy_marked_extents(struct btrfs_root *root, if (ret) break; - clear_extent_bits(dirty_pages, start, end, mark, GFP_NOFS); + clear_extent_bits(dirty_pages, start, end, mark); while (start <= end) { eb = btrfs_find_tree_block(root->fs_info, start); start += root->nodesize; diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 0cfddbe8abba..b1c6d7634e88 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -246,9 +246,9 @@ static void free_excluded_extents(struct btrfs_root *root, end = start + cache->key.offset - 1; clear_extent_bits(&root->fs_info->freed_extents[0], - start, end, EXTENT_UPTODATE, GFP_NOFS); + start, end, EXTENT_UPTODATE); clear_extent_bits(&root->fs_info->freed_extents[1], - start, end, EXTENT_UPTODATE, GFP_NOFS); + start, end, EXTENT_UPTODATE); } static int exclude_super_stripes(struct btrfs_root *root, @@ -10526,14 +10526,14 @@ void btrfs_delete_unused_bgs(struct btrfs_fs_info *fs_info) */ mutex_lock(&fs_info->unused_bg_unpin_mutex); ret = clear_extent_bits(&fs_info->freed_extents[0], start, end, - EXTENT_DIRTY, GFP_NOFS); + EXTENT_DIRTY); if (ret) { mutex_unlock(&fs_info->unused_bg_unpin_mutex); btrfs_dec_block_group_ro(root, block_group); goto end_trans; } ret = clear_extent_bits(&fs_info->freed_extents[1], start, end, - EXTENT_DIRTY, GFP_NOFS); + EXTENT_DIRTY); if (ret) { mutex_unlock(&fs_info->unused_bg_unpin_mutex); btrfs_dec_block_group_ro(root, block_group); diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 3b53d217066c..2946c1522870 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -1975,13 +1975,13 @@ int free_io_failure(struct inode *inode, struct io_failure_record *rec) set_state_failrec(failure_tree, rec->start, NULL); ret = clear_extent_bits(failure_tree, rec->start, rec->start + rec->len - 1, - EXTENT_LOCKED | EXTENT_DIRTY, GFP_NOFS); + EXTENT_LOCKED | EXTENT_DIRTY); if (ret) err = ret; ret = clear_extent_bits(&BTRFS_I(inode)->io_tree, rec->start, rec->start + rec->len - 1, - EXTENT_DAMAGED, GFP_NOFS); + EXTENT_DAMAGED); if (ret && !err) err = ret; diff --git a/fs/btrfs/extent_io.h b/fs/btrfs/extent_io.h index 447c6e6ff88c..90c5c8176162 100644 --- a/fs/btrfs/extent_io.h +++ b/fs/btrfs/extent_io.h @@ -241,14 +241,15 @@ static inline int unlock_extent_cached(struct extent_io_tree *tree, u64 start, } static inline int clear_extent_bits(struct extent_io_tree *tree, u64 start, - u64 end, unsigned bits, gfp_t mask) + u64 end, unsigned bits) { int wake = 0; if (bits & EXTENT_LOCKED) wake = 1; - return clear_extent_bit(tree, start, end, bits, wake, 0, NULL, mask); + return clear_extent_bit(tree, start, end, bits, wake, 0, NULL, + GFP_NOFS); } int set_record_extent_bits(struct extent_io_tree *tree, u64 start, u64 end, diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 2aaba58b4856..d37997f4eb98 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -3103,8 +3103,7 @@ static int btrfs_readpage_end_io_hook(struct btrfs_io_bio *io_bio, if (root->root_key.objectid == BTRFS_DATA_RELOC_TREE_OBJECTID && test_range_bit(io_tree, start, end, EXTENT_NODATASUM, 1, NULL)) { - clear_extent_bits(io_tree, start, end, EXTENT_NODATASUM, - GFP_NOFS); + clear_extent_bits(io_tree, start, end, EXTENT_NODATASUM); return 0; } diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c index 78cbfb530de6..bd5ea1a8a9f1 100644 --- a/fs/btrfs/relocation.c +++ b/fs/btrfs/relocation.c @@ -4059,8 +4059,7 @@ restart: } btrfs_release_path(path); - clear_extent_bits(&rc->processed_blocks, 0, (u64)-1, EXTENT_DIRTY, - GFP_NOFS); + clear_extent_bits(&rc->processed_blocks, 0, (u64)-1, EXTENT_DIRTY); if (trans) { btrfs_end_transaction_throttle(trans, rc->extent_root); diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c index 9d24375eed1f..841dae874a21 100644 --- a/fs/btrfs/scrub.c +++ b/fs/btrfs/scrub.c @@ -763,7 +763,7 @@ static int scrub_fixup_readpage(u64 inum, u64 offset, u64 root, void *fixup_ctx) end, EXTENT_DAMAGED, 0, NULL); if (!corrected) clear_extent_bits(&BTRFS_I(inode)->io_tree, offset, end, - EXTENT_DAMAGED, GFP_NOFS); + EXTENT_DAMAGED); } out: diff --git a/fs/btrfs/tests/extent-io-tests.c b/fs/btrfs/tests/extent-io-tests.c index 70948b13bc81..88e28e5d2c38 100644 --- a/fs/btrfs/tests/extent-io-tests.c +++ b/fs/btrfs/tests/extent-io-tests.c @@ -262,7 +262,7 @@ static int test_find_delalloc(void) } ret = 0; out_bits: - clear_extent_bits(&tmp, 0, total_dirty - 1, (unsigned)-1, GFP_KERNEL); + clear_extent_bits(&tmp, 0, total_dirty - 1, (unsigned)-1); out: if (locked_page) put_page(locked_page); diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c index 517d0ccb351e..a111e275ab9e 100644 --- a/fs/btrfs/tree-log.c +++ b/fs/btrfs/tree-log.c @@ -3001,7 +3001,7 @@ static void free_log_tree(struct btrfs_trans_handle *trans, break; clear_extent_bits(&log->dirty_log_pages, start, end, - EXTENT_DIRTY | EXTENT_NEW, GFP_NOFS); + EXTENT_DIRTY | EXTENT_NEW); } /* -- cgit v1.3-8-gc7d7 From f734c44a1bfffd762f6f5829cb41224d267b80d9 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Tue, 26 Apr 2016 23:54:39 +0200 Subject: btrfs: sink gfp parameter to clear_record_extent_bits Callers pass GFP_NOFS. No need to pass the flags around. Signed-off-by: David Sterba --- fs/btrfs/extent_io.c | 5 ++--- fs/btrfs/extent_io.h | 3 +-- fs/btrfs/qgroup.c | 5 ++--- 3 files changed, 5 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 2946c1522870..164bda63c5ac 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -1308,8 +1308,7 @@ int clear_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, } int clear_record_extent_bits(struct extent_io_tree *tree, u64 start, u64 end, - unsigned bits, gfp_t mask, - struct extent_changeset *changeset) + unsigned bits, struct extent_changeset *changeset) { /* * Don't support EXTENT_LOCKED case, same reason as @@ -1317,7 +1316,7 @@ int clear_record_extent_bits(struct extent_io_tree *tree, u64 start, u64 end, */ BUG_ON(bits & EXTENT_LOCKED); - return __clear_extent_bit(tree, start, end, bits, 0, 0, NULL, mask, + return __clear_extent_bit(tree, start, end, bits, 0, 0, NULL, GFP_NOFS, changeset); } diff --git a/fs/btrfs/extent_io.h b/fs/btrfs/extent_io.h index 90c5c8176162..b7c258c9fa2d 100644 --- a/fs/btrfs/extent_io.h +++ b/fs/btrfs/extent_io.h @@ -221,8 +221,7 @@ int test_range_bit(struct extent_io_tree *tree, u64 start, u64 end, unsigned bits, int filled, struct extent_state *cached_state); int clear_record_extent_bits(struct extent_io_tree *tree, u64 start, u64 end, - unsigned bits, gfp_t mask, - struct extent_changeset *changeset); + unsigned bits, struct extent_changeset *changeset); int clear_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, unsigned bits, int wake, int delete, struct extent_state **cached, gfp_t mask); diff --git a/fs/btrfs/qgroup.c b/fs/btrfs/qgroup.c index 9e119552ed32..129392a78986 100644 --- a/fs/btrfs/qgroup.c +++ b/fs/btrfs/qgroup.c @@ -2580,8 +2580,7 @@ static int __btrfs_qgroup_release_data(struct inode *inode, u64 start, u64 len, return -ENOMEM; ret = clear_record_extent_bits(&BTRFS_I(inode)->io_tree, start, - start + len -1, EXTENT_QGROUP_RESERVED, GFP_NOFS, - &changeset); + start + len -1, EXTENT_QGROUP_RESERVED, &changeset); if (ret < 0) goto out; @@ -2688,7 +2687,7 @@ void btrfs_qgroup_check_reserved_leak(struct inode *inode) return; ret = clear_record_extent_bits(&BTRFS_I(inode)->io_tree, 0, (u64)-1, - EXTENT_QGROUP_RESERVED, GFP_NOFS, &changeset); + EXTENT_QGROUP_RESERVED, &changeset); WARN_ON(ret < 0); if (WARN_ON(changeset.bytes_changed)) { -- cgit v1.3-8-gc7d7 From af6f8f604d44e05f98ed45a69830547ed133adf8 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Tue, 26 Apr 2016 23:54:39 +0200 Subject: btrfs: sink gfp parameter to clear_extent_dirty Callers pass GFP_NOFS. No need to pass the flags around. Signed-off-by: David Sterba --- fs/btrfs/disk-io.c | 2 +- fs/btrfs/extent-tree.c | 2 +- fs/btrfs/extent_io.h | 4 ++-- 3 files changed, 4 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 3bc88b4b0032..263823f28ba4 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -4392,7 +4392,7 @@ again: if (ret) break; - clear_extent_dirty(unpin, start, end, GFP_NOFS); + clear_extent_dirty(unpin, start, end); btrfs_error_unpin_extent_range(root, start, end); cond_resched(); } diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index b1c6d7634e88..61b3dd25ba4c 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -6408,7 +6408,7 @@ int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans, ret = btrfs_discard_extent(root, start, end + 1 - start, NULL); - clear_extent_dirty(unpin, start, end, GFP_NOFS); + clear_extent_dirty(unpin, start, end); unpin_extent_range(root, start, end, true); mutex_unlock(&fs_info->unused_bg_unpin_mutex); cond_resched(); diff --git a/fs/btrfs/extent_io.h b/fs/btrfs/extent_io.h index b7c258c9fa2d..9e987ee03361 100644 --- a/fs/btrfs/extent_io.h +++ b/fs/btrfs/extent_io.h @@ -279,11 +279,11 @@ static inline int set_extent_dirty(struct extent_io_tree *tree, u64 start, } static inline int clear_extent_dirty(struct extent_io_tree *tree, u64 start, - u64 end, gfp_t mask) + u64 end) { return clear_extent_bit(tree, start, end, EXTENT_DIRTY | EXTENT_DELALLOC | - EXTENT_DO_ACCOUNTING, 0, 0, NULL, mask); + EXTENT_DO_ACCOUNTING, 0, 0, NULL, GFP_NOFS); } int convert_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, -- cgit v1.3-8-gc7d7 From 7cd8c7527cb3ad3b397fecfa1bdc8eec9fa33dd8 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Tue, 26 Apr 2016 23:54:39 +0200 Subject: btrfs: sink gfp parameter to set_extent_delalloc Callers pass GFP_NOFS and tests pass GFP_KERNEL, but using NOFS there does not hurt. No need to pass the flags around. Signed-off-by: David Sterba --- fs/btrfs/extent_io.h | 4 ++-- fs/btrfs/inode.c | 2 +- fs/btrfs/tests/extent-io-tests.c | 6 +++--- 3 files changed, 6 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent_io.h b/fs/btrfs/extent_io.h index 9e987ee03361..e99a6befdbf5 100644 --- a/fs/btrfs/extent_io.h +++ b/fs/btrfs/extent_io.h @@ -291,11 +291,11 @@ int convert_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, struct extent_state **cached_state, gfp_t mask); static inline int set_extent_delalloc(struct extent_io_tree *tree, u64 start, - u64 end, struct extent_state **cached_state, gfp_t mask) + u64 end, struct extent_state **cached_state) { return set_extent_bit(tree, start, end, EXTENT_DELALLOC | EXTENT_UPTODATE, - NULL, cached_state, mask); + NULL, cached_state, GFP_NOFS); } static inline int set_extent_defrag(struct extent_io_tree *tree, u64 start, diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index d37997f4eb98..1a4bd193962c 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -1962,7 +1962,7 @@ int btrfs_set_extent_delalloc(struct inode *inode, u64 start, u64 end, { WARN_ON((end & (PAGE_SIZE - 1)) == 0); return set_extent_delalloc(&BTRFS_I(inode)->io_tree, start, end, - cached_state, GFP_NOFS); + cached_state); } /* see btrfs_writepage_start_hook for details on why this is required */ diff --git a/fs/btrfs/tests/extent-io-tests.c b/fs/btrfs/tests/extent-io-tests.c index 88e28e5d2c38..c190d2a520d1 100644 --- a/fs/btrfs/tests/extent-io-tests.c +++ b/fs/btrfs/tests/extent-io-tests.c @@ -113,7 +113,7 @@ static int test_find_delalloc(void) * |--- delalloc ---| * |--- search ---| */ - set_extent_delalloc(&tmp, 0, 4095, NULL, GFP_KERNEL); + set_extent_delalloc(&tmp, 0, 4095, NULL); start = 0; end = 0; found = find_lock_delalloc_range(inode, &tmp, locked_page, &start, @@ -144,7 +144,7 @@ static int test_find_delalloc(void) test_msg("Couldn't find the locked page\n"); goto out_bits; } - set_extent_delalloc(&tmp, 4096, max_bytes - 1, NULL, GFP_KERNEL); + set_extent_delalloc(&tmp, 4096, max_bytes - 1, NULL); start = test_start; end = 0; found = find_lock_delalloc_range(inode, &tmp, locked_page, &start, @@ -199,7 +199,7 @@ static int test_find_delalloc(void) * * We are re-using our test_start from above since it works out well. */ - set_extent_delalloc(&tmp, max_bytes, total_dirty - 1, NULL, GFP_KERNEL); + set_extent_delalloc(&tmp, max_bytes, total_dirty - 1, NULL); start = test_start; end = 0; found = find_lock_delalloc_range(inode, &tmp, locked_page, &start, -- cgit v1.3-8-gc7d7 From 018ed4f788870fdf6e625d04aa287080bdc8a54f Mon Sep 17 00:00:00 2001 From: David Sterba Date: Tue, 26 Apr 2016 23:54:39 +0200 Subject: btrfs: sink gfp parameter to set_extent_defrag Single caller passes GFP_NOFS. Signed-off-by: David Sterba --- fs/btrfs/extent_io.h | 4 ++-- fs/btrfs/ioctl.c | 2 +- 2 files changed, 3 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent_io.h b/fs/btrfs/extent_io.h index e99a6befdbf5..a1cf4b1dd290 100644 --- a/fs/btrfs/extent_io.h +++ b/fs/btrfs/extent_io.h @@ -299,11 +299,11 @@ static inline int set_extent_delalloc(struct extent_io_tree *tree, u64 start, } static inline int set_extent_defrag(struct extent_io_tree *tree, u64 start, - u64 end, struct extent_state **cached_state, gfp_t mask) + u64 end, struct extent_state **cached_state) { return set_extent_bit(tree, start, end, EXTENT_DELALLOC | EXTENT_UPTODATE | EXTENT_DEFRAG, - NULL, cached_state, mask); + NULL, cached_state, GFP_NOFS); } static inline int set_extent_new(struct extent_io_tree *tree, u64 start, diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index 5a23806ae418..72f48252a55d 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -1228,7 +1228,7 @@ again: set_extent_defrag(&BTRFS_I(inode)->io_tree, page_start, page_end - 1, - &cached_state, GFP_NOFS); + &cached_state); unlock_extent_cached(&BTRFS_I(inode)->io_tree, page_start, page_end - 1, &cached_state, -- cgit v1.3-8-gc7d7 From 3744dbeb7033825e53b919ae0887e08e924841a9 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Tue, 26 Apr 2016 23:54:39 +0200 Subject: btrfs: sink gfp parameter to set_extent_new Single caller passes GFP_NOFS. Signed-off-by: David Sterba --- fs/btrfs/extent-tree.c | 2 +- fs/btrfs/extent_io.h | 5 +++-- 2 files changed, 4 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 61b3dd25ba4c..0ddc996a40e1 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -7923,7 +7923,7 @@ btrfs_init_new_buffer(struct btrfs_trans_handle *trans, struct btrfs_root *root, buf->start + buf->len - 1, GFP_NOFS); else set_extent_new(&root->dirty_log_pages, buf->start, - buf->start + buf->len - 1, GFP_NOFS); + buf->start + buf->len - 1); } else { buf->log_index = -1; set_extent_dirty(&trans->transaction->dirty_pages, buf->start, diff --git a/fs/btrfs/extent_io.h b/fs/btrfs/extent_io.h index a1cf4b1dd290..566cf86d7e63 100644 --- a/fs/btrfs/extent_io.h +++ b/fs/btrfs/extent_io.h @@ -307,9 +307,10 @@ static inline int set_extent_defrag(struct extent_io_tree *tree, u64 start, } static inline int set_extent_new(struct extent_io_tree *tree, u64 start, - u64 end, gfp_t mask) + u64 end) { - return set_extent_bit(tree, start, end, EXTENT_NEW, NULL, NULL, mask); + return set_extent_bit(tree, start, end, EXTENT_NEW, NULL, NULL, + GFP_NOFS); } static inline int set_extent_uptodate(struct extent_io_tree *tree, u64 start, -- cgit v1.3-8-gc7d7 From 2c53b912ae317e560bce1fc446e76915a5b30587 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Tue, 26 Apr 2016 23:54:39 +0200 Subject: btrfs: sink gfp parameter to set_record_extent_bits Single caller passes GFP_NOFS. Signed-off-by: David Sterba --- fs/btrfs/extent_io.c | 5 ++--- fs/btrfs/extent_io.h | 3 +-- fs/btrfs/qgroup.c | 3 +-- 3 files changed, 4 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 164bda63c5ac..0c9b11924f74 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -1284,8 +1284,7 @@ search_again: /* wrappers around set/clear extent bit */ int set_record_extent_bits(struct extent_io_tree *tree, u64 start, u64 end, - unsigned bits, gfp_t mask, - struct extent_changeset *changeset) + unsigned bits, struct extent_changeset *changeset) { /* * We don't support EXTENT_LOCKED yet, as current changeset will @@ -1295,7 +1294,7 @@ int set_record_extent_bits(struct extent_io_tree *tree, u64 start, u64 end, */ BUG_ON(bits & EXTENT_LOCKED); - return __set_extent_bit(tree, start, end, bits, 0, NULL, NULL, mask, + return __set_extent_bit(tree, start, end, bits, 0, NULL, NULL, GFP_NOFS, changeset); } diff --git a/fs/btrfs/extent_io.h b/fs/btrfs/extent_io.h index 566cf86d7e63..94b376446042 100644 --- a/fs/btrfs/extent_io.h +++ b/fs/btrfs/extent_io.h @@ -252,8 +252,7 @@ static inline int clear_extent_bits(struct extent_io_tree *tree, u64 start, } int set_record_extent_bits(struct extent_io_tree *tree, u64 start, u64 end, - unsigned bits, gfp_t mask, - struct extent_changeset *changeset); + unsigned bits, struct extent_changeset *changeset); int set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, unsigned bits, u64 *failed_start, struct extent_state **cached_state, gfp_t mask); diff --git a/fs/btrfs/qgroup.c b/fs/btrfs/qgroup.c index 129392a78986..34af959b4e0f 100644 --- a/fs/btrfs/qgroup.c +++ b/fs/btrfs/qgroup.c @@ -2542,8 +2542,7 @@ int btrfs_qgroup_reserve_data(struct inode *inode, u64 start, u64 len) changeset.bytes_changed = 0; changeset.range_changed = ulist_alloc(GFP_NOFS); ret = set_record_extent_bits(&BTRFS_I(inode)->io_tree, start, - start + len -1, EXTENT_QGROUP_RESERVED, GFP_NOFS, - &changeset); + start + len -1, EXTENT_QGROUP_RESERVED, &changeset); trace_btrfs_qgroup_reserve_data(inode, start, len, changeset.bytes_changed, QGROUP_RESERVE); -- cgit v1.3-8-gc7d7 From b5a4ba14e0caa9da1b3b98939274bdeda71859e3 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Wed, 27 Apr 2016 01:02:15 +0200 Subject: btrfs: untangle gotos a bit in __set_extent_bit Signed-off-by: David Sterba --- fs/btrfs/extent_io.c | 15 +++++++-------- 1 file changed, 7 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 0c9b11924f74..f5eda54e225f 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -1037,7 +1037,13 @@ hit_next: goto out; } - goto search_again; +search_again: + if (start > end) + goto out; + spin_unlock(&tree->lock); + if (gfpflags_allow_blocking(mask)) + cond_resched(); + goto again; out: spin_unlock(&tree->lock); @@ -1046,13 +1052,6 @@ out: return err; -search_again: - if (start > end) - goto out; - spin_unlock(&tree->lock); - if (gfpflags_allow_blocking(mask)) - cond_resched(); - goto again; } int set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, -- cgit v1.3-8-gc7d7 From 7ab5cb2a9e3161de3a288bafd100515b5f44f27e Mon Sep 17 00:00:00 2001 From: David Sterba Date: Wed, 27 Apr 2016 01:02:15 +0200 Subject: btrfs: untangle gotos a bit in __clear_extent_bit Signed-off-by: David Sterba --- fs/btrfs/extent_io.c | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index f5eda54e225f..3ecd8b1f239f 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -726,14 +726,6 @@ next: start = last_end + 1; if (start <= end && state && !need_resched()) goto hit_next; - goto search_again; - -out: - spin_unlock(&tree->lock); - if (prealloc) - free_extent_state(prealloc); - - return 0; search_again: if (start > end) @@ -742,6 +734,14 @@ search_again: if (gfpflags_allow_blocking(mask)) cond_resched(); goto again; + +out: + spin_unlock(&tree->lock); + if (prealloc) + free_extent_state(prealloc); + + return 0; + } static void wait_on_state(struct extent_io_tree *tree, -- cgit v1.3-8-gc7d7 From 03bf5387708069194fbe5072fb57f9542e7a8b82 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Wed, 27 Apr 2016 01:02:15 +0200 Subject: btrfs: untangle gotos a bit in convert_extent_bit Signed-off-by: David Sterba --- fs/btrfs/extent_io.c | 16 +++++++--------- 1 file changed, 7 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 3ecd8b1f239f..d39e2241ceb7 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -1262,15 +1262,6 @@ hit_next: goto out; } - goto search_again; - -out: - spin_unlock(&tree->lock); - if (prealloc) - free_extent_state(prealloc); - - return err; - search_again: if (start > end) goto out; @@ -1279,6 +1270,13 @@ search_again: cond_resched(); first_iteration = false; goto again; + +out: + spin_unlock(&tree->lock); + if (prealloc) + free_extent_state(prealloc); + + return err; } /* wrappers around set/clear extent bit */ -- cgit v1.3-8-gc7d7 From 059f791c6bbaba72dc3c1bd6e2657aacc8552849 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Wed, 27 Apr 2016 01:03:45 +0200 Subject: btrfs: make state preallocation more speculative in __set_extent_bit Similar to __clear_extent_bit, do not fail if the state preallocation fails as we might not need it. One less BUG_ON. Signed-off-by: David Sterba --- fs/btrfs/extent_io.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index d39e2241ceb7..8707bcc615ff 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -873,8 +873,14 @@ __set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, bits |= EXTENT_FIRST_DELALLOC; again: if (!prealloc && gfpflags_allow_blocking(mask)) { + /* + * Don't care for allocation failure here because we might end + * up not needing the pre-allocated extent state at all, which + * is the case if we only have in the tree extent states that + * cover our input range and don't cover too any other range. + * If we end up needing a new extent state we allocate it later. + */ prealloc = alloc_extent_state(mask); - BUG_ON(!prealloc); } spin_lock(&tree->lock); -- cgit v1.3-8-gc7d7 From 210aa27768bec4297a9d6ad0e5cab45935c775e9 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Tue, 26 Apr 2016 23:54:39 +0200 Subject: btrfs: sink gfp parameter to convert_extent_bit Single caller passes GFP_NOFS. We can get rid of the gfpflags_allow_blocking checks as NOFS can block but does not recurse to filesystem through reclaim. Signed-off-by: David Sterba --- fs/btrfs/extent_io.c | 12 ++++++------ fs/btrfs/extent_io.h | 2 +- fs/btrfs/transaction.c | 2 +- 3 files changed, 8 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 8707bcc615ff..c1139bcf8870 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -1078,17 +1078,18 @@ int set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, * @bits: the bits to set in this range * @clear_bits: the bits to clear in this range * @cached_state: state that we're going to cache - * @mask: the allocation mask * * This will go through and set bits for the given range. If any states exist * already in this range they are set with the given bit and cleared of the * clear_bits. This is only meant to be used by things that are mergeable, ie * converting from say DELALLOC to DIRTY. This is not meant to be used with * boundary bits like LOCK. + * + * All allocations are done with GFP_NOFS. */ int convert_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, unsigned bits, unsigned clear_bits, - struct extent_state **cached_state, gfp_t mask) + struct extent_state **cached_state) { struct extent_state *state; struct extent_state *prealloc = NULL; @@ -1103,7 +1104,7 @@ int convert_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, btrfs_debug_check_extent_io_range(tree, start, end); again: - if (!prealloc && gfpflags_allow_blocking(mask)) { + if (!prealloc) { /* * Best effort, don't worry if extent state allocation fails * here for the first iteration. We might have a cached state @@ -1111,7 +1112,7 @@ again: * extent state allocations are needed. We'll only know this * after locking the tree. */ - prealloc = alloc_extent_state(mask); + prealloc = alloc_extent_state(GFP_NOFS); if (!prealloc && !first_iteration) return -ENOMEM; } @@ -1272,8 +1273,7 @@ search_again: if (start > end) goto out; spin_unlock(&tree->lock); - if (gfpflags_allow_blocking(mask)) - cond_resched(); + cond_resched(); first_iteration = false; goto again; diff --git a/fs/btrfs/extent_io.h b/fs/btrfs/extent_io.h index 94b376446042..9c1f160c5984 100644 --- a/fs/btrfs/extent_io.h +++ b/fs/btrfs/extent_io.h @@ -287,7 +287,7 @@ static inline int clear_extent_dirty(struct extent_io_tree *tree, u64 start, int convert_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, unsigned bits, unsigned clear_bits, - struct extent_state **cached_state, gfp_t mask); + struct extent_state **cached_state); static inline int set_extent_delalloc(struct extent_io_tree *tree, u64 start, u64 end, struct extent_state **cached_state) diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c index 43885e51b882..98b93d9c6213 100644 --- a/fs/btrfs/transaction.c +++ b/fs/btrfs/transaction.c @@ -943,7 +943,7 @@ int btrfs_write_marked_extents(struct btrfs_root *root, err = convert_extent_bit(dirty_pages, start, end, EXTENT_NEED_WAIT, - mark, &cached_state, GFP_NOFS); + mark, &cached_state); /* * convert_extent_bit can return -ENOMEM, which is most of the * time a temporary error. So when it happens, ignore the error -- cgit v1.3-8-gc7d7 From c9eb13a9105e2e418f72e46a2b6da3f49e696902 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Sat, 30 Apr 2016 00:48:54 -0400 Subject: ext4: fix hang when processing corrupted orphaned inode list If the orphaned inode list contains inode #5, ext4_iget() returns a bad inode (since the bootloader inode should never be referenced directly). Because of the bad inode, we end up processing the inode repeatedly and this hangs the machine. This can be reproduced via: mke2fs -t ext4 /tmp/foo.img 100 debugfs -w -R "ssv last_orphan 5" /tmp/foo.img mount -o loop /tmp/foo.img /mnt (But don't do this if you are using an unpatched kernel if you care about the system staying functional. :-) This bug was found by the port of American Fuzzy Lop into the kernel to find file system problems[1]. (Since it *only* happens if inode #5 shows up on the orphan list --- 3, 7, 8, etc. won't do it, it's not surprising that AFL needed two hours before it found it.) [1] http://events.linuxfoundation.org/sites/events/files/slides/AFL%20filesystem%20fuzzing%2C%20Vault%202016_0.pdf Cc: stable@vger.kernel.org Reported by: Vegard Nossum Signed-off-by: Theodore Ts'o --- fs/ext4/ialloc.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/ext4/ialloc.c b/fs/ext4/ialloc.c index 237b877d316d..c2caf2df3695 100644 --- a/fs/ext4/ialloc.c +++ b/fs/ext4/ialloc.c @@ -1183,11 +1183,13 @@ struct inode *ext4_orphan_get(struct super_block *sb, unsigned long ino) goto iget_failed; /* - * If the orphans has i_nlinks > 0 then it should be able to be - * truncated, otherwise it won't be removed from the orphan list - * during processing and an infinite loop will result. + * If the orphans has i_nlinks > 0 then it should be able to + * be truncated, otherwise it won't be removed from the orphan + * list during processing and an infinite loop will result. + * Similarly, it must not be a bad inode. */ - if (inode->i_nlink && !ext4_can_truncate(inode)) + if ((inode->i_nlink && !ext4_can_truncate(inode)) || + is_bad_inode(inode)) goto bad_orphan; if (NEXT_ORPHAN(inode) > max_ino) -- cgit v1.3-8-gc7d7 From 7827a7f6ebfcb7f388dc47fddd48567a314701ba Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Sat, 30 Apr 2016 00:49:54 -0400 Subject: ext4: clean up error handling when orphan list is corrupted Instead of just printing warning messages, if the orphan list is corrupted, declare the file system is corrupted. If there are any reserved inodes in the orphaned inode list, declare the file system corrupted and stop right away to avoid doing more potential damage to the file system. Cc: stable@vger.kernel.org Signed-off-by: Theodore Ts'o --- fs/ext4/ialloc.c | 49 ++++++++++++++++++++++--------------------------- 1 file changed, 22 insertions(+), 27 deletions(-) (limited to 'fs') diff --git a/fs/ext4/ialloc.c b/fs/ext4/ialloc.c index c2caf2df3695..3da4cf8d18b6 100644 --- a/fs/ext4/ialloc.c +++ b/fs/ext4/ialloc.c @@ -1150,25 +1150,20 @@ struct inode *ext4_orphan_get(struct super_block *sb, unsigned long ino) unsigned long max_ino = le32_to_cpu(EXT4_SB(sb)->s_es->s_inodes_count); ext4_group_t block_group; int bit; - struct buffer_head *bitmap_bh; + struct buffer_head *bitmap_bh = NULL; struct inode *inode = NULL; - long err = -EIO; + int err = -EFSCORRUPTED; - /* Error cases - e2fsck has already cleaned up for us */ - if (ino > max_ino) { - ext4_warning(sb, "bad orphan ino %lu! e2fsck was run?", ino); - err = -EFSCORRUPTED; - goto error; - } + if (ino < EXT4_FIRST_INO(sb) || ino > max_ino) + goto bad_orphan; block_group = (ino - 1) / EXT4_INODES_PER_GROUP(sb); bit = (ino - 1) % EXT4_INODES_PER_GROUP(sb); bitmap_bh = ext4_read_inode_bitmap(sb, block_group); if (IS_ERR(bitmap_bh)) { - err = PTR_ERR(bitmap_bh); - ext4_warning(sb, "inode bitmap error %ld for orphan %lu", - ino, err); - goto error; + ext4_error(sb, "inode bitmap error %ld for orphan %lu", + ino, PTR_ERR(bitmap_bh)); + return (struct inode *) bitmap_bh; } /* Having the inode bit set should be a 100% indicator that this @@ -1179,8 +1174,12 @@ struct inode *ext4_orphan_get(struct super_block *sb, unsigned long ino) goto bad_orphan; inode = ext4_iget(sb, ino); - if (IS_ERR(inode)) - goto iget_failed; + if (IS_ERR(inode)) { + err = PTR_ERR(inode); + ext4_error(sb, "couldn't read orphan inode %lu (err %d)", + ino, err); + return inode; + } /* * If the orphans has i_nlinks > 0 then it should be able to @@ -1197,29 +1196,25 @@ struct inode *ext4_orphan_get(struct super_block *sb, unsigned long ino) brelse(bitmap_bh); return inode; -iget_failed: - err = PTR_ERR(inode); - inode = NULL; bad_orphan: - ext4_warning(sb, "bad orphan inode %lu! e2fsck was run?", ino); - printk(KERN_WARNING "ext4_test_bit(bit=%d, block=%llu) = %d\n", - bit, (unsigned long long)bitmap_bh->b_blocknr, - ext4_test_bit(bit, bitmap_bh->b_data)); - printk(KERN_WARNING "inode=%p\n", inode); + ext4_error(sb, "bad orphan inode %lu", ino); + if (bitmap_bh) + printk(KERN_ERR "ext4_test_bit(bit=%d, block=%llu) = %d\n", + bit, (unsigned long long)bitmap_bh->b_blocknr, + ext4_test_bit(bit, bitmap_bh->b_data)); if (inode) { - printk(KERN_WARNING "is_bad_inode(inode)=%d\n", + printk(KERN_ERR "is_bad_inode(inode)=%d\n", is_bad_inode(inode)); - printk(KERN_WARNING "NEXT_ORPHAN(inode)=%u\n", + printk(KERN_ERR "NEXT_ORPHAN(inode)=%u\n", NEXT_ORPHAN(inode)); - printk(KERN_WARNING "max_ino=%lu\n", max_ino); - printk(KERN_WARNING "i_nlink=%u\n", inode->i_nlink); + printk(KERN_ERR "max_ino=%lu\n", max_ino); + printk(KERN_ERR "i_nlink=%u\n", inode->i_nlink); /* Avoid freeing blocks if we got a bad deleted inode */ if (inode->i_nlink == 0) inode->i_blocks = 0; iput(inode); } brelse(bitmap_bh); -error: return ERR_PTR(err); } -- cgit v1.3-8-gc7d7 From e4234a1fc343ca35f852bc527fae56fade879d4a Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Thu, 31 Mar 2016 11:45:06 +0100 Subject: kernfs: Move faulting copy_user operations outside of the mutex MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit A fault in a user provided buffer may lead anywhere, and lockdep warns that we have a potential deadlock between the mm->mmap_sem and the kernfs file mutex: [ 82.811702] ====================================================== [ 82.811705] [ INFO: possible circular locking dependency detected ] [ 82.811709] 4.5.0-rc4-gfxbench+ #1 Not tainted [ 82.811711] ------------------------------------------------------- [ 82.811714] kms_setmode/5859 is trying to acquire lock: [ 82.811717] (&dev->struct_mutex){+.+.+.}, at: [] drm_gem_mmap+0x1a1/0x270 [ 82.811731] but task is already holding lock: [ 82.811734] (&mm->mmap_sem){++++++}, at: [] vm_mmap_pgoff+0x44/0xa0 [ 82.811745] which lock already depends on the new lock. [ 82.811749] the existing dependency chain (in reverse order) is: [ 82.811752] -> #3 (&mm->mmap_sem){++++++}: [ 82.811761] [] lock_acquire+0xc3/0x1d0 [ 82.811766] [] __might_fault+0x75/0xa0 [ 82.811771] [] kernfs_fop_write+0x8a/0x180 [ 82.811787] [] __vfs_write+0x23/0xe0 [ 82.811792] [] vfs_write+0xa4/0x190 [ 82.811797] [] SyS_write+0x44/0xb0 [ 82.811801] [] entry_SYSCALL_64_fastpath+0x16/0x73 [ 82.811807] -> #2 (s_active#6){++++.+}: [ 82.811814] [] lock_acquire+0xc3/0x1d0 [ 82.811819] [] __kernfs_remove+0x210/0x2f0 [ 82.811823] [] kernfs_remove_by_name_ns+0x40/0xa0 [ 82.811828] [] sysfs_remove_file_ns+0x10/0x20 [ 82.811832] [] device_del+0x124/0x250 [ 82.811837] [] device_unregister+0x19/0x60 [ 82.811841] [] cpu_cache_sysfs_exit+0x51/0xb0 [ 82.811846] [] cacheinfo_cpu_callback+0x38/0x70 [ 82.811851] [] notifier_call_chain+0x39/0xa0 [ 82.811856] [] __raw_notifier_call_chain+0x9/0x10 [ 82.811860] [] cpu_notify+0x1e/0x40 [ 82.811865] [] cpu_notify_nofail+0x9/0x20 [ 82.811869] [] _cpu_down+0x233/0x340 [ 82.811874] [] disable_nonboot_cpus+0xc9/0x350 [ 82.811878] [] suspend_devices_and_enter+0x5a1/0xb50 [ 82.811883] [] pm_suspend+0x543/0x8d0 [ 82.811888] [] state_store+0x77/0xe0 [ 82.811892] [] kobj_attr_store+0xf/0x20 [ 82.811897] [] sysfs_kf_write+0x40/0x50 [ 82.811902] [] kernfs_fop_write+0x13c/0x180 [ 82.811906] [] __vfs_write+0x23/0xe0 [ 82.811910] [] vfs_write+0xa4/0x190 [ 82.811914] [] SyS_write+0x44/0xb0 [ 82.811918] [] entry_SYSCALL_64_fastpath+0x16/0x73 [ 82.811923] -> #1 (cpu_hotplug.lock){+.+.+.}: [ 82.811929] [] lock_acquire+0xc3/0x1d0 [ 82.811933] [] mutex_lock_nested+0x62/0x3b0 [ 82.811940] [] get_online_cpus+0x61/0x80 [ 82.811944] [] stop_machine+0x1b/0xe0 [ 82.811949] [] gen8_ggtt_insert_entries__BKL+0x2d/0x30 [i915] [ 82.812009] [] ggtt_bind_vma+0x46/0x70 [i915] [ 82.812045] [] i915_vma_bind+0x140/0x290 [i915] [ 82.812081] [] i915_gem_object_do_pin+0x899/0xb00 [i915] [ 82.812117] [] i915_gem_object_pin+0x35/0x40 [i915] [ 82.812154] [] intel_init_pipe_control+0xbe/0x210 [i915] [ 82.812192] [] intel_logical_rings_init+0xe2/0xde0 [i915] [ 82.812232] [] i915_gem_init+0xf3/0x130 [i915] [ 82.812278] [] i915_driver_load+0xf2d/0x1770 [i915] [ 82.812318] [] drm_dev_register+0xa4/0xb0 [ 82.812323] [] drm_get_pci_dev+0xce/0x1e0 [ 82.812328] [] i915_pci_probe+0x2f/0x50 [i915] [ 82.812360] [] pci_device_probe+0x87/0xf0 [ 82.812366] [] driver_probe_device+0x229/0x450 [ 82.812371] [] __driver_attach+0x83/0x90 [ 82.812375] [] bus_for_each_dev+0x61/0xa0 [ 82.812380] [] driver_attach+0x19/0x20 [ 82.812384] [] bus_add_driver+0x1ef/0x290 [ 82.812388] [] driver_register+0x5b/0xe0 [ 82.812393] [] __pci_register_driver+0x5b/0x60 [ 82.812398] [] drm_pci_init+0xd6/0x100 [ 82.812402] [] 0xffffffffa027c094 [ 82.812406] [] do_one_initcall+0xae/0x1d0 [ 82.812412] [] do_init_module+0x5b/0x1cb [ 82.812417] [] load_module+0x1c20/0x2480 [ 82.812422] [] SyS_finit_module+0x7e/0xa0 [ 82.812428] [] entry_SYSCALL_64_fastpath+0x16/0x73 [ 82.812433] -> #0 (&dev->struct_mutex){+.+.+.}: [ 82.812439] [] __lock_acquire+0x1fc9/0x20f0 [ 82.812443] [] lock_acquire+0xc3/0x1d0 [ 82.812456] [] drm_gem_mmap+0x1c7/0x270 [ 82.812460] [] mmap_region+0x334/0x580 [ 82.812466] [] do_mmap+0x364/0x410 [ 82.812470] [] vm_mmap_pgoff+0x6d/0xa0 [ 82.812474] [] SyS_mmap_pgoff+0x184/0x220 [ 82.812479] [] SyS_mmap+0x1d/0x20 [ 82.812484] [] entry_SYSCALL_64_fastpath+0x16/0x73 [ 82.812489] other info that might help us debug this: [ 82.812493] Chain exists of: &dev->struct_mutex --> s_active#6 --> &mm->mmap_sem [ 82.812502] Possible unsafe locking scenario: [ 82.812506] CPU0 CPU1 [ 82.812508] ---- ---- [ 82.812510] lock(&mm->mmap_sem); [ 82.812514] lock(s_active#6); [ 82.812519] lock(&mm->mmap_sem); [ 82.812522] lock(&dev->struct_mutex); [ 82.812526] *** DEADLOCK *** [ 82.812531] 1 lock held by kms_setmode/5859: [ 82.812533] #0: (&mm->mmap_sem){++++++}, at: [] vm_mmap_pgoff+0x44/0xa0 [ 82.812541] stack backtrace: [ 82.812547] CPU: 0 PID: 5859 Comm: kms_setmode Not tainted 4.5.0-rc4-gfxbench+ #1 [ 82.812550] Hardware name: /NUC5CPYB, BIOS PYBSWCEL.86A.0040.2015.0814.1353 08/14/2015 [ 82.812553] 0000000000000000 ffff880079407bf0 ffffffff813f8505 ffffffff825fb270 [ 82.812560] ffffffff825c4190 ffff880079407c30 ffffffff810c84ac ffff880079407c90 [ 82.812566] ffff8800797ed328 ffff8800797ecb00 0000000000000001 ffff8800797ed350 [ 82.812573] Call Trace: [ 82.812578] [] dump_stack+0x67/0x92 [ 82.812582] [] print_circular_bug+0x1fc/0x310 [ 82.812586] [] __lock_acquire+0x1fc9/0x20f0 [ 82.812590] [] lock_acquire+0xc3/0x1d0 [ 82.812594] [] ? drm_gem_mmap+0x1a1/0x270 [ 82.812599] [] drm_gem_mmap+0x1c7/0x270 [ 82.812603] [] ? drm_gem_mmap+0x1a1/0x270 [ 82.812608] [] mmap_region+0x334/0x580 [ 82.812612] [] do_mmap+0x364/0x410 [ 82.812616] [] vm_mmap_pgoff+0x6d/0xa0 [ 82.812629] [] SyS_mmap_pgoff+0x184/0x220 [ 82.812633] [] SyS_mmap+0x1d/0x20 [ 82.812637] [] entry_SYSCALL_64_fastpath+0x16/0x73 Highly unlikely though this scenario is, we can avoid the issue entirely by moving the copy operation from out under the kernfs_get_active() tracking by assigning the preallocated buffer its own mutex. The temporary buffer allocation doesn't require mutex locking as it is entirely local. The locked section was extended by the addition of the preallocated buf to speed up md user operations in commit 2b75869bba676c248d8d25ae6d2bd9221dfffdb6 Author: NeilBrown Date: Mon Oct 13 16:41:28 2014 +1100 sysfs/kernfs: allow attributes to request write buffer be pre-allocated. Reported-by: Ville Syrjälä Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94350 Signed-off-by: Chris Wilson Reviewed-by: Joonas Lahtinen Cc: Ville Syrjälä Cc: Joonas Lahtinen Cc: NeilBrown Acked-by: Tejun Heo Signed-off-by: Greg Kroah-Hartman --- fs/kernfs/file.c | 51 ++++++++++++++++++++++++++++---------------------- include/linux/kernfs.h | 1 + 2 files changed, 30 insertions(+), 22 deletions(-) (limited to 'fs') diff --git a/fs/kernfs/file.c b/fs/kernfs/file.c index 7247252ee9b1..e1574008adc9 100644 --- a/fs/kernfs/file.c +++ b/fs/kernfs/file.c @@ -190,15 +190,16 @@ static ssize_t kernfs_file_direct_read(struct kernfs_open_file *of, char *buf; buf = of->prealloc_buf; - if (!buf) + if (buf) + mutex_lock(&of->prealloc_mutex); + else buf = kmalloc(len, GFP_KERNEL); if (!buf) return -ENOMEM; /* * @of->mutex nests outside active ref and is used both to ensure that - * the ops aren't called concurrently for the same open file, and - * to provide exclusive access to ->prealloc_buf (when that exists). + * the ops aren't called concurrently for the same open file. */ mutex_lock(&of->mutex); if (!kernfs_get_active(of->kn)) { @@ -214,21 +215,23 @@ static ssize_t kernfs_file_direct_read(struct kernfs_open_file *of, else len = -EINVAL; + kernfs_put_active(of->kn); + mutex_unlock(&of->mutex); + if (len < 0) - goto out_unlock; + goto out_free; if (copy_to_user(user_buf, buf, len)) { len = -EFAULT; - goto out_unlock; + goto out_free; } *ppos += len; - out_unlock: - kernfs_put_active(of->kn); - mutex_unlock(&of->mutex); out_free: - if (buf != of->prealloc_buf) + if (buf == of->prealloc_buf) + mutex_unlock(&of->prealloc_mutex); + else kfree(buf); return len; } @@ -284,15 +287,22 @@ static ssize_t kernfs_fop_write(struct file *file, const char __user *user_buf, } buf = of->prealloc_buf; - if (!buf) + if (buf) + mutex_lock(&of->prealloc_mutex); + else buf = kmalloc(len + 1, GFP_KERNEL); if (!buf) return -ENOMEM; + if (copy_from_user(buf, user_buf, len)) { + len = -EFAULT; + goto out_free; + } + buf[len] = '\0'; /* guarantee string termination */ + /* * @of->mutex nests outside active ref and is used both to ensure that - * the ops aren't called concurrently for the same open file, and - * to provide exclusive access to ->prealloc_buf (when that exists). + * the ops aren't called concurrently for the same open file. */ mutex_lock(&of->mutex); if (!kernfs_get_active(of->kn)) { @@ -301,26 +311,22 @@ static ssize_t kernfs_fop_write(struct file *file, const char __user *user_buf, goto out_free; } - if (copy_from_user(buf, user_buf, len)) { - len = -EFAULT; - goto out_unlock; - } - buf[len] = '\0'; /* guarantee string termination */ - ops = kernfs_ops(of->kn); if (ops->write) len = ops->write(of, buf, len, *ppos); else len = -EINVAL; + kernfs_put_active(of->kn); + mutex_unlock(&of->mutex); + if (len > 0) *ppos += len; -out_unlock: - kernfs_put_active(of->kn); - mutex_unlock(&of->mutex); out_free: - if (buf != of->prealloc_buf) + if (buf == of->prealloc_buf) + mutex_unlock(&of->prealloc_mutex); + else kfree(buf); return len; } @@ -687,6 +693,7 @@ static int kernfs_fop_open(struct inode *inode, struct file *file) error = -ENOMEM; if (!of->prealloc_buf) goto err_free; + mutex_init(&of->prealloc_mutex); } /* diff --git a/include/linux/kernfs.h b/include/linux/kernfs.h index c06c44242f39..d306e282bb1d 100644 --- a/include/linux/kernfs.h +++ b/include/linux/kernfs.h @@ -177,6 +177,7 @@ struct kernfs_open_file { /* private fields, do not use outside kernfs proper */ struct mutex mutex; + struct mutex prealloc_mutex; int event; struct list_head list; char *prealloc_buf; -- cgit v1.3-8-gc7d7 From 39d637af5aa7577f655c58b9e55587566c63a0af Mon Sep 17 00:00:00 2001 From: Dmitry Kasatkin Date: Sun, 26 Oct 2014 12:42:07 +0200 Subject: vfs: forbid write access when reading a file into memory This patch is based on top of the "vfs: support for a common kernel file loader" patch set. In general when the kernel is reading a file into memory it does not want anything else writing to it. The kernel currently only forbids write access to a file being executed. This patch extends this locking to files being read by the kernel. Changelog: - moved function to kernel_read_file() - Mimi - updated patch description - Mimi Signed-off-by: Dmitry Kasatkin Cc: Al Viro Signed-off-by: Mimi Zohar Reviewed-by: Luis R. Rodriguez Acked-by: Kees Cook --- fs/exec.c | 29 +++++++++++++++++++++-------- 1 file changed, 21 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/exec.c b/fs/exec.c index c4010b8207a1..fdecb7615587 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -850,15 +850,25 @@ int kernel_read_file(struct file *file, void **buf, loff_t *size, if (ret) return ret; + ret = deny_write_access(file); + if (ret) + return ret; + i_size = i_size_read(file_inode(file)); - if (max_size > 0 && i_size > max_size) - return -EFBIG; - if (i_size <= 0) - return -EINVAL; + if (max_size > 0 && i_size > max_size) { + ret = -EFBIG; + goto out; + } + if (i_size <= 0) { + ret = -EINVAL; + goto out; + } *buf = vmalloc(i_size); - if (!*buf) - return -ENOMEM; + if (!*buf) { + ret = -ENOMEM; + goto out; + } pos = 0; while (pos < i_size) { @@ -876,18 +886,21 @@ int kernel_read_file(struct file *file, void **buf, loff_t *size, if (pos != i_size) { ret = -EIO; - goto out; + goto out_free; } ret = security_kernel_post_read_file(file, *buf, i_size, id); if (!ret) *size = pos; -out: +out_free: if (ret < 0) { vfree(*buf); *buf = NULL; } + +out: + allow_write_access(file); return ret; } EXPORT_SYMBOL_GPL(kernel_read_file); -- cgit v1.3-8-gc7d7 From 05d1a717ec0430c916a749b94eb90ab74bbfa356 Mon Sep 17 00:00:00 2001 From: Mimi Zohar Date: Mon, 29 Feb 2016 19:52:05 -0500 Subject: ima: add support for creating files using the mknodat syscall Commit 3034a14 "ima: pass 'opened' flag to identify newly created files" stopped identifying empty files as new files. However new empty files can be created using the mknodat syscall. On systems with IMA-appraisal enabled, these empty files are not labeled with security.ima extended attributes properly, preventing them from subsequently being opened in order to write the file data contents. This patch defines a new hook named ima_post_path_mknod() to mark these empty files, created using mknodat, as new in order to allow the file data contents to be written. In addition, files with security.ima xattrs containing a file signature are considered "immutable" and can not be modified. The file contents need to be written, before signing the file. This patch relaxes this requirement for new files, allowing the file signature to be written before the file contents. Changelog: - defer identifying files with signatures stored as security.ima (based on Dmitry Rozhkov's comments) - removing tests (eg. dentry, dentry->d_inode, inode->i_size == 0) (based on Al's review) Signed-off-by: Mimi Zohar Cc: Al Viro < Tested-by: Dmitry Rozhkov --- fs/namei.c | 2 ++ include/linux/ima.h | 6 ++++++ security/integrity/ima/ima_appraise.c | 5 +++++ security/integrity/ima/ima_main.c | 25 ++++++++++++++++++++++++- 4 files changed, 37 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/namei.c b/fs/namei.c index 1d9ca2d5dff6..b4bd06839446 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -3608,6 +3608,8 @@ retry: switch (mode & S_IFMT) { case 0: case S_IFREG: error = vfs_create(path.dentry->d_inode,dentry,mode,true); + if (!error) + ima_post_path_mknod(dentry); break; case S_IFCHR: case S_IFBLK: error = vfs_mknod(path.dentry->d_inode,dentry,mode, diff --git a/include/linux/ima.h b/include/linux/ima.h index e6516cbbe9bf..0eb7c2e7f0d6 100644 --- a/include/linux/ima.h +++ b/include/linux/ima.h @@ -21,6 +21,7 @@ extern int ima_file_mmap(struct file *file, unsigned long prot); extern int ima_read_file(struct file *file, enum kernel_read_file_id id); extern int ima_post_read_file(struct file *file, void *buf, loff_t size, enum kernel_read_file_id id); +extern void ima_post_path_mknod(struct dentry *dentry); #else static inline int ima_bprm_check(struct linux_binprm *bprm) @@ -54,6 +55,11 @@ static inline int ima_post_read_file(struct file *file, void *buf, loff_t size, return 0; } +static inline void ima_post_path_mknod(struct dentry *dentry) +{ + return; +} + #endif /* CONFIG_IMA */ #ifdef CONFIG_IMA_APPRAISE diff --git a/security/integrity/ima/ima_appraise.c b/security/integrity/ima/ima_appraise.c index d2f28a0c8614..1bcbc12e03d9 100644 --- a/security/integrity/ima/ima_appraise.c +++ b/security/integrity/ima/ima_appraise.c @@ -275,6 +275,11 @@ out: xattr_value->type != EVM_IMA_XATTR_DIGSIG)) { if (!ima_fix_xattr(dentry, iint)) status = INTEGRITY_PASS; + } else if ((inode->i_size == 0) && + (iint->flags & IMA_NEW_FILE) && + (xattr_value && + xattr_value->type == EVM_IMA_XATTR_DIGSIG)) { + status = INTEGRITY_PASS; } integrity_audit_msg(AUDIT_INTEGRITY_DATA, inode, filename, op, cause, rc, 0); diff --git a/security/integrity/ima/ima_main.c b/security/integrity/ima/ima_main.c index 391f41751021..68b26c340acd 100644 --- a/security/integrity/ima/ima_main.c +++ b/security/integrity/ima/ima_main.c @@ -246,7 +246,8 @@ static int process_measurement(struct file *file, char *buf, loff_t size, ima_audit_measurement(iint, pathname); out_digsig: - if ((mask & MAY_WRITE) && (iint->flags & IMA_DIGSIG)) + if ((mask & MAY_WRITE) && (iint->flags & IMA_DIGSIG) && + !(iint->flags & IMA_NEW_FILE)) rc = -EACCES; kfree(xattr_value); out_free: @@ -315,6 +316,28 @@ int ima_file_check(struct file *file, int mask, int opened) } EXPORT_SYMBOL_GPL(ima_file_check); +/** + * ima_post_path_mknod - mark as a new inode + * @dentry: newly created dentry + * + * Mark files created via the mknodat syscall as new, so that the + * file data can be written later. + */ +void ima_post_path_mknod(struct dentry *dentry) +{ + struct integrity_iint_cache *iint; + struct inode *inode = dentry->d_inode; + int must_appraise; + + must_appraise = ima_must_appraise(inode, MAY_ACCESS, FILE_CHECK); + if (!must_appraise) + return; + + iint = integrity_inode_get(inode); + if (iint) + iint->flags |= IMA_NEW_FILE; +} + /** * ima_read_file - pre-measure/appraise hook decision based on policy * @file: pointer to the file to be measured/appraised/audit -- cgit v1.3-8-gc7d7 From 80f4781d2c0ba63bf9ab4de90a6829a1368b80a3 Mon Sep 17 00:00:00 2001 From: Abhi Das Date: Mon, 2 May 2016 07:07:01 -0500 Subject: gfs2: use inode_lock/unlock instead of accessing i_mutex directly i_mutex has been replaced by i_rwsem and directly accessing the non-existent i_mutex breaks the kernel build. Signed-off-by: Abhi Das Signed-off-by: Bob Peterson --- fs/gfs2/file.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/file.c b/fs/gfs2/file.c index f33fd92e5f49..374dd5327101 100644 --- a/fs/gfs2/file.c +++ b/fs/gfs2/file.c @@ -960,16 +960,16 @@ static ssize_t gfs2_file_splice_read(struct file *in, loff_t *ppos, struct gfs2_holder gh; int ret; - mutex_lock(&inode->i_mutex); + inode_lock(inode); ret = gfs2_glock_nq_init(ip->i_gl, LM_ST_SHARED, 0, &gh); if (ret) { - mutex_unlock(&inode->i_mutex); + inode_unlock(inode); return ret; } gfs2_glock_dq_uninit(&gh); - mutex_unlock(&inode->i_mutex); + inode_unlock(inode); return generic_file_splice_read(in, ppos, pipe, len, flags); } -- cgit v1.3-8-gc7d7 From 8381e6022755863258a352128a55e375a766f50d Mon Sep 17 00:00:00 2001 From: Bob Peterson Date: Mon, 2 May 2016 09:42:49 -0500 Subject: GFS2: Remove allocation parms from gfs2_rbm_find Struct gfs2_alloc_parms ap is never referenced in function gfs2_rbm_find, so this patch removes it. Signed-off-by: Bob Peterson --- fs/gfs2/rgrp.c | 16 ++++++---------- 1 file changed, 6 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c index 07c0265aa195..8b580e535ad4 100644 --- a/fs/gfs2/rgrp.c +++ b/fs/gfs2/rgrp.c @@ -73,8 +73,7 @@ static const char valid_change[16] = { }; static int gfs2_rbm_find(struct gfs2_rbm *rbm, u8 state, u32 *minext, - const struct gfs2_inode *ip, bool nowrap, - const struct gfs2_alloc_parms *ap); + const struct gfs2_inode *ip, bool nowrap); /** @@ -1512,7 +1511,7 @@ static void rg_mblk_search(struct gfs2_rgrpd *rgd, struct gfs2_inode *ip, if (WARN_ON(gfs2_rbm_from_block(&rbm, goal))) return; - ret = gfs2_rbm_find(&rbm, GFS2_BLKST_FREE, &extlen, ip, true, ap); + ret = gfs2_rbm_find(&rbm, GFS2_BLKST_FREE, &extlen, ip, true); if (ret == 0) { rs->rs_rbm = rbm; rs->rs_free = extlen; @@ -1639,7 +1638,6 @@ fail: * @ip: If set, check for reservations * @nowrap: Stop looking at the end of the rgrp, rather than wrapping * around until we've reached the starting point. - * @ap: the allocation parameters * * Side effects: * - If looking for free blocks, we set GBF_FULL on each bitmap which @@ -1651,8 +1649,7 @@ fail: */ static int gfs2_rbm_find(struct gfs2_rbm *rbm, u8 state, u32 *minext, - const struct gfs2_inode *ip, bool nowrap, - const struct gfs2_alloc_parms *ap) + const struct gfs2_inode *ip, bool nowrap) { struct buffer_head *bh; int initial_bii; @@ -1773,7 +1770,7 @@ static void try_rgrp_unlink(struct gfs2_rgrpd *rgd, u64 *last_unlinked, u64 skip while (1) { down_write(&sdp->sd_log_flush_lock); error = gfs2_rbm_find(&rbm, GFS2_BLKST_UNLINKED, NULL, NULL, - true, NULL); + true); up_write(&sdp->sd_log_flush_lock); if (error == -ENOSPC) break; @@ -2330,12 +2327,11 @@ int gfs2_alloc_blocks(struct gfs2_inode *ip, u64 *bn, unsigned int *nblocks, int error; gfs2_set_alloc_start(&rbm, ip, dinode); - error = gfs2_rbm_find(&rbm, GFS2_BLKST_FREE, NULL, ip, false, NULL); + error = gfs2_rbm_find(&rbm, GFS2_BLKST_FREE, NULL, ip, false); if (error == -ENOSPC) { gfs2_set_alloc_start(&rbm, ip, dinode); - error = gfs2_rbm_find(&rbm, GFS2_BLKST_FREE, NULL, NULL, false, - NULL); + error = gfs2_rbm_find(&rbm, GFS2_BLKST_FREE, NULL, NULL, false); } /* Since all blocks are reserved in advance, this shouldn't happen */ -- cgit v1.3-8-gc7d7 From ad8403df054c9b5dc3175a26c6179571b9cafa4e Mon Sep 17 00:00:00 2001 From: Anand Jain Date: Thu, 10 Mar 2016 12:22:15 +0800 Subject: btrfs: pass the right error code to the btrfs_std_error Also drop the newline from the message. Signed-off-by: Anand Jain Signed-off-by: David Sterba --- fs/btrfs/ioctl.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index 0f76aea6e398..03df0bbb445d 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -4847,8 +4847,8 @@ static long btrfs_ioctl_qgroup_assign(struct file *file, void __user *arg) /* update qgroup status and info */ err = btrfs_run_qgroups(trans, root->fs_info); if (err < 0) - btrfs_handle_fs_error(root->fs_info, ret, - "failed to update qgroup status and info\n"); + btrfs_handle_fs_error(root->fs_info, err, + "failed to update qgroup status and info"); err = btrfs_end_transaction(trans, root); if (err && !ret) ret = err; -- cgit v1.3-8-gc7d7 From 23dc974eed576b2464b222a892272073adf6a92c Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Fri, 29 Apr 2016 20:09:15 +0800 Subject: f2fs: fix to clear private data in page Private data in page should be removed during ->releasepage or ->invalidatepage, otherwise garbage data would be remained in that page. Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/data.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'fs') diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index 38ce5d6f8583..888f1781ae0b 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -1723,6 +1723,7 @@ void f2fs_invalidate_page(struct page *page, unsigned int offset, if (IS_ATOMIC_WRITTEN_PAGE(page)) return; + set_page_private(page, 0); ClearPagePrivate(page); } @@ -1736,6 +1737,7 @@ int f2fs_release_page(struct page *page, gfp_t wait) if (IS_ATOMIC_WRITTEN_PAGE(page)) return 0; + set_page_private(page, 0); ClearPagePrivate(page); return 1; } -- cgit v1.3-8-gc7d7 From c81ced05b2c7fc3c1d6db753a55b55733ffb1524 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Fri, 29 Apr 2016 20:13:36 +0800 Subject: f2fs: fix to clear page private flag Commit 28bc106b2346 ("f2fs: support revoking atomic written pages") forgot to clear page private flag correctly, fix it. Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/segment.c | 1 + 1 file changed, 1 insertion(+) (limited to 'fs') diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c index 770cdc95120f..2e6f537a0e7d 100644 --- a/fs/f2fs/segment.c +++ b/fs/f2fs/segment.c @@ -227,6 +227,7 @@ next: if (drop || recover) ClearPageUptodate(page); set_page_private(page, 0); + ClearPagePrivate(page); f2fs_put_page(page, 1); list_del(&cur->list); -- cgit v1.3-8-gc7d7 From 3f8ab270855b0b461995da5dc48dce9461c85d94 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Fri, 29 Apr 2016 20:13:37 +0800 Subject: f2fs: factor out fsync inode entry operations Factor out fsync inode entry operations into {add,del}_fsync_inode. Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/recovery.c | 59 +++++++++++++++++++++++++++++++++++------------------- 1 file changed, 38 insertions(+), 21 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/recovery.c b/fs/f2fs/recovery.c index a646d3ba3b25..58275a29e87d 100644 --- a/fs/f2fs/recovery.c +++ b/fs/f2fs/recovery.c @@ -67,6 +67,28 @@ static struct fsync_inode_entry *get_fsync_inode(struct list_head *head, return NULL; } +static struct fsync_inode_entry *add_fsync_inode(struct list_head *head, + struct inode *inode) +{ + struct fsync_inode_entry *entry; + + entry = kmem_cache_alloc(fsync_entry_slab, GFP_F2FS_ZERO); + if (!entry) + return NULL; + + entry->inode = inode; + list_add_tail(&entry->list, head); + + return entry; +} + +static void del_fsync_inode(struct fsync_inode_entry *entry) +{ + iput(entry->inode); + list_del(&entry->list); + kmem_cache_free(fsync_entry_slab, entry); +} + static int recover_dentry(struct inode *inode, struct page *ipage) { struct f2fs_inode *raw_inode = F2FS_INODE(ipage); @@ -198,6 +220,7 @@ static int find_fsync_dnodes(struct f2fs_sb_info *sbi, struct list_head *head) { unsigned long long cp_ver = cur_cp_version(F2FS_CKPT(sbi)); struct curseg_info *curseg; + struct inode *inode; struct page *page = NULL; block_t blkaddr; int err = 0; @@ -233,27 +256,27 @@ static int find_fsync_dnodes(struct f2fs_sb_info *sbi, struct list_head *head) break; } - /* add this fsync inode to the list */ - entry = kmem_cache_alloc(fsync_entry_slab, GFP_F2FS_ZERO); - if (!entry) { - err = -ENOMEM; - break; - } /* * CP | dnode(F) | inode(DF) * For this case, we should not give up now. */ - entry->inode = f2fs_iget(sbi->sb, ino_of_node(page)); - if (IS_ERR(entry->inode)) { - err = PTR_ERR(entry->inode); - kmem_cache_free(fsync_entry_slab, entry); + inode = f2fs_iget(sbi->sb, ino_of_node(page)); + if (IS_ERR(inode)) { + err = PTR_ERR(inode); if (err == -ENOENT) { err = 0; goto next; } break; } - list_add_tail(&entry->list, head); + + /* add this fsync inode to the list */ + entry = add_fsync_inode(head, inode); + if (!entry) { + err = -ENOMEM; + iput(inode); + break; + } } entry->blkaddr = blkaddr; @@ -274,11 +297,8 @@ static void destroy_fsync_dnodes(struct list_head *head) { struct fsync_inode_entry *entry, *tmp; - list_for_each_entry_safe(entry, tmp, head, list) { - iput(entry->inode); - list_del(&entry->list); - kmem_cache_free(fsync_entry_slab, entry); - } + list_for_each_entry_safe(entry, tmp, head, list) + del_fsync_inode(entry); } static int check_index_in_prev_nodes(struct f2fs_sb_info *sbi, @@ -533,11 +553,8 @@ static int recover_data(struct f2fs_sb_info *sbi, struct list_head *head) break; } - if (entry->blkaddr == blkaddr) { - iput(entry->inode); - list_del(&entry->list); - kmem_cache_free(fsync_entry_slab, entry); - } + if (entry->blkaddr == blkaddr) + del_fsync_inode(entry); next: /* check next segment */ blkaddr = next_blkaddr_of_node(page); -- cgit v1.3-8-gc7d7 From 8ed01abe7da6af62f2089da9a4c1fe839dc638f5 Mon Sep 17 00:00:00 2001 From: Anand Jain Date: Thu, 14 Apr 2016 18:24:10 +0800 Subject: btrfs: s_bdev is not null after missing replace Yauhen reported in the ML that s_bdev is null at mount, and s_bdev gets updated to some device when missing device is replaced, as because bdev is null for missing device, things gets matched up. Fix this by checking if s_bdev is set. I didn't want to completely remove updating s_bdev because the future multi device support at vfs layer may need it. Signed-off-by: Anand Jain Reported-by: Yauhen Kharuzhy Signed-off-by: David Sterba --- fs/btrfs/dev-replace.c | 3 ++- fs/btrfs/volumes.c | 6 ++++-- 2 files changed, 6 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/dev-replace.c b/fs/btrfs/dev-replace.c index 5aebedf12b5f..c93eec9ca433 100644 --- a/fs/btrfs/dev-replace.c +++ b/fs/btrfs/dev-replace.c @@ -569,7 +569,8 @@ static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info, ASSERT(list_empty(&src_device->resized_list)); tgt_device->commit_total_bytes = src_device->commit_total_bytes; tgt_device->commit_bytes_used = src_device->bytes_used; - if (fs_info->sb->s_bdev == src_device->bdev) + if (fs_info->sb->s_bdev && + (fs_info->sb->s_bdev == src_device->bdev)) fs_info->sb->s_bdev = tgt_device->bdev; if (fs_info->fs_devices->latest_bdev == src_device->bdev) fs_info->fs_devices->latest_bdev = tgt_device->bdev; diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index e43b01e438b1..41bc384f3483 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -1848,7 +1848,8 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path, u64 devid) next_device = list_entry(root->fs_info->fs_devices->devices.next, struct btrfs_device, dev_list); - if (device->bdev == root->fs_info->sb->s_bdev) + if (root->fs_info->sb->s_bdev && + (root->fs_info->sb->s_bdev == device->bdev)) root->fs_info->sb->s_bdev = next_device->bdev; if (device->bdev == root->fs_info->fs_devices->latest_bdev) root->fs_info->fs_devices->latest_bdev = next_device->bdev; @@ -1996,7 +1997,8 @@ void btrfs_destroy_dev_replace_tgtdev(struct btrfs_fs_info *fs_info, next_device = list_entry(fs_info->fs_devices->devices.next, struct btrfs_device, dev_list); - if (tgtdev->bdev == fs_info->sb->s_bdev) + if (fs_info->sb->s_bdev && + (tgtdev->bdev == fs_info->sb->s_bdev)) fs_info->sb->s_bdev = next_device->bdev; if (tgtdev->bdev == fs_info->fs_devices->latest_bdev) fs_info->fs_devices->latest_bdev = next_device->bdev; -- cgit v1.3-8-gc7d7 From 88acff64c621aaeee2a4fe0ed124c77358069bce Mon Sep 17 00:00:00 2001 From: Anand Jain Date: Tue, 3 May 2016 17:44:43 +0800 Subject: btrfs: cleanup assigning next active device with a check Creates helper fucntion as needed by the device delete and replace operations. Also now it checks if the next device being assigned is an active device. Signed-off-by: Anand Jain Signed-off-by: David Sterba --- fs/btrfs/dev-replace.c | 8 +++---- fs/btrfs/volumes.c | 60 ++++++++++++++++++++++++++++++++++++-------------- fs/btrfs/volumes.h | 2 ++ 3 files changed, 48 insertions(+), 22 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/dev-replace.c b/fs/btrfs/dev-replace.c index c93eec9ca433..1f193f742f21 100644 --- a/fs/btrfs/dev-replace.c +++ b/fs/btrfs/dev-replace.c @@ -569,11 +569,9 @@ static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info, ASSERT(list_empty(&src_device->resized_list)); tgt_device->commit_total_bytes = src_device->commit_total_bytes; tgt_device->commit_bytes_used = src_device->bytes_used; - if (fs_info->sb->s_bdev && - (fs_info->sb->s_bdev == src_device->bdev)) - fs_info->sb->s_bdev = tgt_device->bdev; - if (fs_info->fs_devices->latest_bdev == src_device->bdev) - fs_info->fs_devices->latest_bdev = tgt_device->bdev; + + btrfs_assign_next_active_device(fs_info, src_device, tgt_device); + list_add(&tgt_device->dev_alloc_list, &fs_info->fs_devices->alloc_list); fs_info->fs_devices->rw_devices++; diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 41bc384f3483..3fb0786060dd 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -1756,10 +1756,49 @@ static int btrfs_check_raid_min_devices(struct btrfs_fs_info *fs_info, return 0; } +struct btrfs_device *btrfs_find_next_active_device(struct btrfs_fs_devices *fs_devs, + struct btrfs_device *device) +{ + struct btrfs_device *next_device; + + list_for_each_entry(next_device, &fs_devs->devices, dev_list) { + if (next_device != device && + !next_device->missing && next_device->bdev) + return next_device; + } + + return NULL; +} + +/* + * Helper function to check if the given device is part of s_bdev / latest_bdev + * and replace it with the provided or the next active device, in the context + * where this function called, there should be always be another device (or + * this_dev) which is active. + */ +void btrfs_assign_next_active_device(struct btrfs_fs_info *fs_info, + struct btrfs_device *device, struct btrfs_device *this_dev) +{ + struct btrfs_device *next_device; + + if (this_dev) + next_device = this_dev; + else + next_device = btrfs_find_next_active_device(fs_info->fs_devices, + device); + ASSERT(next_device); + + if (fs_info->sb->s_bdev && + (fs_info->sb->s_bdev == device->bdev)) + fs_info->sb->s_bdev = next_device->bdev; + + if (fs_info->fs_devices->latest_bdev == device->bdev) + fs_info->fs_devices->latest_bdev = next_device->bdev; +} + int btrfs_rm_device(struct btrfs_root *root, char *device_path, u64 devid) { struct btrfs_device *device; - struct btrfs_device *next_device; struct btrfs_fs_devices *cur_devices; u64 num_devices; int ret = 0; @@ -1846,13 +1885,7 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path, u64 devid) if (device->missing) device->fs_devices->missing_devices--; - next_device = list_entry(root->fs_info->fs_devices->devices.next, - struct btrfs_device, dev_list); - if (root->fs_info->sb->s_bdev && - (root->fs_info->sb->s_bdev == device->bdev)) - root->fs_info->sb->s_bdev = next_device->bdev; - if (device->bdev == root->fs_info->fs_devices->latest_bdev) - root->fs_info->fs_devices->latest_bdev = next_device->bdev; + btrfs_assign_next_active_device(root->fs_info, device, NULL); if (device->bdev) { device->fs_devices->open_devices--; @@ -1981,8 +2014,6 @@ void btrfs_rm_dev_replace_free_srcdev(struct btrfs_fs_info *fs_info, void btrfs_destroy_dev_replace_tgtdev(struct btrfs_fs_info *fs_info, struct btrfs_device *tgtdev) { - struct btrfs_device *next_device; - mutex_lock(&uuid_mutex); WARN_ON(!tgtdev); mutex_lock(&fs_info->fs_devices->device_list_mutex); @@ -1995,13 +2026,8 @@ void btrfs_destroy_dev_replace_tgtdev(struct btrfs_fs_info *fs_info, } fs_info->fs_devices->num_devices--; - next_device = list_entry(fs_info->fs_devices->devices.next, - struct btrfs_device, dev_list); - if (fs_info->sb->s_bdev && - (tgtdev->bdev == fs_info->sb->s_bdev)) - fs_info->sb->s_bdev = next_device->bdev; - if (tgtdev->bdev == fs_info->fs_devices->latest_bdev) - fs_info->fs_devices->latest_bdev = next_device->bdev; + btrfs_assign_next_active_device(fs_info, tgtdev, NULL); + list_del_rcu(&tgtdev->dev_list); call_rcu(&tgtdev->rcu, free_device); diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h index febdb7bc9370..7038017bf93b 100644 --- a/fs/btrfs/volumes.h +++ b/fs/btrfs/volumes.h @@ -445,6 +445,8 @@ int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder, struct btrfs_fs_devices **fs_devices_ret); int btrfs_close_devices(struct btrfs_fs_devices *fs_devices); void btrfs_close_extra_devices(struct btrfs_fs_devices *fs_devices, int step); +void btrfs_assign_next_active_device(struct btrfs_fs_info *fs_info, + struct btrfs_device *device, struct btrfs_device *this_dev); int btrfs_find_device_missing_or_by_path(struct btrfs_root *root, char *device_path, struct btrfs_device **device); -- cgit v1.3-8-gc7d7 From ff0bc08454917964291f72ee5b8eca66de4bc250 Mon Sep 17 00:00:00 2001 From: Seth Forshee Date: Thu, 5 May 2016 10:52:38 -0400 Subject: ext4: fix check of dqget() return value in ext4_ioctl_setproject() A failed call to dqget() returns an ERR_PTR() and not null. Fix the check in ext4_ioctl_setproject() to handle this correctly. Fixes: 9b7365fc1c82 ("ext4: add FS_IOC_FSSETXATTR/FS_IOC_FSGETXATTR interface support") Cc: stable@vger.kernel.org # v4.5 Signed-off-by: Seth Forshee Signed-off-by: Theodore Ts'o Reviewed-by: Jan Kara --- fs/ext4/ioctl.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ext4/ioctl.c b/fs/ext4/ioctl.c index eae5917c534e..0acf8cacb2be 100644 --- a/fs/ext4/ioctl.c +++ b/fs/ext4/ioctl.c @@ -365,7 +365,7 @@ static int ext4_ioctl_setproject(struct file *filp, __u32 projid) struct dquot *transfer_to[MAXQUOTAS] = { }; transfer_to[PRJQUOTA] = dqget(sb, make_kqid_projid(kprojid)); - if (transfer_to[PRJQUOTA]) { + if (!IS_ERR(transfer_to[PRJQUOTA])) { err = __dquot_transfer(inode, transfer_to); dqput(transfer_to[PRJQUOTA]); if (err) -- cgit v1.3-8-gc7d7 From 74177f55b70e2f2be770dd28684dd6d17106a4ba Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Thu, 5 May 2016 11:10:15 -0400 Subject: ext4: fix oops on corrupted filesystem When filesystem is corrupted in the right way, it can happen ext4_mark_iloc_dirty() in ext4_orphan_add() returns error and we subsequently remove inode from the in-memory orphan list. However this deletion is done with list_del(&EXT4_I(inode)->i_orphan) and thus we leave i_orphan list_head with a stale content. Later we can look at this content causing list corruption, oops, or other issues. The reported trace looked like: WARNING: CPU: 0 PID: 46 at lib/list_debug.c:53 __list_del_entry+0x6b/0x100() list_del corruption, 0000000061c1d6e0->next is LIST_POISON1 0000000000100100) CPU: 0 PID: 46 Comm: ext4.exe Not tainted 4.1.0-rc4+ #250 Stack: 60462947 62219960 602ede24 62219960 602ede24 603ca293 622198f0 602f02eb 62219950 6002c12c 62219900 601b4d6b Call Trace: [<6005769c>] ? vprintk_emit+0x2dc/0x5c0 [<602ede24>] ? printk+0x0/0x94 [<600190bc>] show_stack+0xdc/0x1a0 [<602ede24>] ? printk+0x0/0x94 [<602ede24>] ? printk+0x0/0x94 [<602f02eb>] dump_stack+0x2a/0x2c [<6002c12c>] warn_slowpath_common+0x9c/0xf0 [<601b4d6b>] ? __list_del_entry+0x6b/0x100 [<6002c254>] warn_slowpath_fmt+0x94/0xa0 [<602f4d09>] ? __mutex_lock_slowpath+0x239/0x3a0 [<6002c1c0>] ? warn_slowpath_fmt+0x0/0xa0 [<60023ebf>] ? set_signals+0x3f/0x50 [<600a205a>] ? kmem_cache_free+0x10a/0x180 [<602f4e88>] ? mutex_lock+0x18/0x30 [<601b4d6b>] __list_del_entry+0x6b/0x100 [<601177ec>] ext4_orphan_del+0x22c/0x2f0 [<6012f27c>] ? __ext4_journal_start_sb+0x2c/0xa0 [<6010b973>] ? ext4_truncate+0x383/0x390 [<6010bc8b>] ext4_write_begin+0x30b/0x4b0 [<6001bb50>] ? copy_from_user+0x0/0xb0 [<601aa840>] ? iov_iter_fault_in_readable+0xa0/0xc0 [<60072c4f>] generic_perform_write+0xaf/0x1e0 [<600c4166>] ? file_update_time+0x46/0x110 [<60072f0f>] __generic_file_write_iter+0x18f/0x1b0 [<6010030f>] ext4_file_write_iter+0x15f/0x470 [<60094e10>] ? unlink_file_vma+0x0/0x70 [<6009b180>] ? unlink_anon_vmas+0x0/0x260 [<6008f169>] ? free_pgtables+0xb9/0x100 [<600a6030>] __vfs_write+0xb0/0x130 [<600a61d5>] vfs_write+0xa5/0x170 [<600a63d6>] SyS_write+0x56/0xe0 [<6029fcb0>] ? __libc_waitpid+0x0/0xa0 [<6001b698>] handle_syscall+0x68/0x90 [<6002633d>] userspace+0x4fd/0x600 [<6002274f>] ? save_registers+0x1f/0x40 [<60028bd7>] ? arch_prctl+0x177/0x1b0 [<60017bd5>] fork_handler+0x85/0x90 Fix the problem by using list_del_init() as we always should with i_orphan list. CC: stable@vger.kernel.org Reported-by: Vegard Nossum Signed-off-by: Jan Kara Signed-off-by: Theodore Ts'o --- fs/ext4/namei.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c index 6e6b3230ee45..672c99a2051b 100644 --- a/fs/ext4/namei.c +++ b/fs/ext4/namei.c @@ -2833,7 +2833,7 @@ int ext4_orphan_add(handle_t *handle, struct inode *inode) * list entries can cause panics at unmount time. */ mutex_lock(&sbi->s_orphan_lock); - list_del(&EXT4_I(inode)->i_orphan); + list_del_init(&EXT4_I(inode)->i_orphan); mutex_unlock(&sbi->s_orphan_lock); } } -- cgit v1.3-8-gc7d7 From b5cb316cdf3a3f5f6125412b0f6065185240cfdc Mon Sep 17 00:00:00 2001 From: Nicolai Stange Date: Thu, 5 May 2016 17:38:03 -0400 Subject: ext4: address UBSAN warning in mb_find_order_for_block() Currently, in mb_find_order_for_block(), there's a loop like the following: while (order <= e4b->bd_blkbits + 1) { ... bb += 1 << (e4b->bd_blkbits - order); } Note that the updated bb is used in the loop's next iteration only. However, at the last iteration, that is at order == e4b->bd_blkbits + 1, the shift count becomes negative (c.f. C99 6.5.7(3)) and UBSAN reports UBSAN: Undefined behaviour in fs/ext4/mballoc.c:1281:11 shift exponent -1 is negative [...] Call Trace: [] dump_stack+0xbc/0x117 [] ? _atomic_dec_and_lock+0x169/0x169 [] ubsan_epilogue+0xd/0x4e [] __ubsan_handle_shift_out_of_bounds+0x1fb/0x254 [] ? __ubsan_handle_load_invalid_value+0x158/0x158 [] ? ext4_mb_generate_from_pa+0x590/0x590 [] ? ext4_read_block_bitmap_nowait+0x598/0xe80 [] mb_find_order_for_block+0x1ce/0x240 [...] Unless compilers start to do some fancy transformations (which at least GCC 6.0.0 doesn't currently do), the issue is of cosmetic nature only: the such calculated value of bb is never used again. Silence UBSAN by introducing another variable, bb_incr, holding the next increment to apply to bb and adjust that one by right shifting it by one position per loop iteration. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=114701 Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=112161 Cc: stable@vger.kernel.org Signed-off-by: Nicolai Stange Signed-off-by: Theodore Ts'o --- fs/ext4/mballoc.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c index efa111a7606d..49e444b51a0c 100644 --- a/fs/ext4/mballoc.c +++ b/fs/ext4/mballoc.c @@ -1266,6 +1266,7 @@ static void ext4_mb_unload_buddy(struct ext4_buddy *e4b) static int mb_find_order_for_block(struct ext4_buddy *e4b, int block) { int order = 1; + int bb_incr = 1 << (e4b->bd_blkbits - 1); void *bb; BUG_ON(e4b->bd_bitmap == e4b->bd_buddy); @@ -1278,7 +1279,8 @@ static int mb_find_order_for_block(struct ext4_buddy *e4b, int block) /* this block is part of buddy of order 'order' */ return order; } - bb += 1 << (e4b->bd_blkbits - order); + bb += bb_incr; + bb_incr >>= 1; order++; } return 0; -- cgit v1.3-8-gc7d7 From 935244cd54b86ca46e69bc6604d2adfb1aec2d42 Mon Sep 17 00:00:00 2001 From: Nicolai Stange Date: Thu, 5 May 2016 19:46:19 -0400 Subject: ext4: silence UBSAN in ext4_mb_init() Currently, in ext4_mb_init(), there's a loop like the following: do { ... offset += 1 << (sb->s_blocksize_bits - i); i++; } while (i <= sb->s_blocksize_bits + 1); Note that the updated offset is used in the loop's next iteration only. However, at the last iteration, that is at i == sb->s_blocksize_bits + 1, the shift count becomes equal to (unsigned)-1 > 31 (c.f. C99 6.5.7(3)) and UBSAN reports UBSAN: Undefined behaviour in fs/ext4/mballoc.c:2621:15 shift exponent 4294967295 is too large for 32-bit type 'int' [...] Call Trace: [] dump_stack+0xbc/0x117 [] ? _atomic_dec_and_lock+0x169/0x169 [] ubsan_epilogue+0xd/0x4e [] __ubsan_handle_shift_out_of_bounds+0x1fb/0x254 [] ? __ubsan_handle_load_invalid_value+0x158/0x158 [] ? kmem_cache_alloc+0x101/0x390 [] ? ext4_mb_init+0x13b/0xfd0 [] ? create_cache+0x57/0x1f0 [] ? create_cache+0x11a/0x1f0 [] ? mutex_lock+0x38/0x60 [] ? mutex_unlock+0x1b/0x50 [] ? put_online_mems+0x5b/0xc0 [] ? kmem_cache_create+0x117/0x2c0 [] ext4_mb_init+0xc49/0xfd0 [...] Observe that the mentioned shift exponent, 4294967295, equals (unsigned)-1. Unless compilers start to do some fancy transformations (which at least GCC 6.0.0 doesn't currently do), the issue is of cosmetic nature only: the such calculated value of offset is never used again. Silence UBSAN by introducing another variable, offset_incr, holding the next increment to apply to offset and adjust that one by right shifting it by one position per loop iteration. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=114701 Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=112161 Cc: stable@vger.kernel.org Signed-off-by: Nicolai Stange Signed-off-by: Theodore Ts'o --- fs/ext4/mballoc.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c index 49e444b51a0c..c1ab3ec30423 100644 --- a/fs/ext4/mballoc.c +++ b/fs/ext4/mballoc.c @@ -2585,7 +2585,7 @@ int ext4_mb_init(struct super_block *sb) { struct ext4_sb_info *sbi = EXT4_SB(sb); unsigned i, j; - unsigned offset; + unsigned offset, offset_incr; unsigned max; int ret; @@ -2614,11 +2614,13 @@ int ext4_mb_init(struct super_block *sb) i = 1; offset = 0; + offset_incr = 1 << (sb->s_blocksize_bits - 1); max = sb->s_blocksize << 2; do { sbi->s_mb_offsets[i] = offset; sbi->s_mb_maxs[i] = max; - offset += 1 << (sb->s_blocksize_bits - i); + offset += offset_incr; + offset_incr = offset_incr >> 1; max = max >> 1; i++; } while (i <= sb->s_blocksize_bits + 1); -- cgit v1.3-8-gc7d7 From 32157de29a3140bfaa74c6e0fae18c78fe84a5df Mon Sep 17 00:00:00 2001 From: Jens Axboe Date: Thu, 5 May 2016 22:09:49 -0400 Subject: ext4: remove unnecessary bio get/put ext4_io_submit() used to check for EOPNOTSUPP after bio submission, which is why it had to get an extra reference to the bio before submitting it. But since we no longer touch the bio after submission, get rid of the redundant get/put of the bio. If we do get the extra reference, we enter the slower path of having to flag this bio as now having external references. Signed-off-by: Jens Axboe Signed-off-by: Theodore Ts'o --- fs/ext4/page-io.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'fs') diff --git a/fs/ext4/page-io.c b/fs/ext4/page-io.c index e4fc8ea45d78..2a01df9cc1c3 100644 --- a/fs/ext4/page-io.c +++ b/fs/ext4/page-io.c @@ -342,9 +342,7 @@ void ext4_io_submit(struct ext4_io_submit *io) if (bio) { int io_op = io->io_wbc->sync_mode == WB_SYNC_ALL ? WRITE_SYNC : WRITE; - bio_get(io->io_bio); submit_bio(io_op, io->io_bio); - bio_put(io->io_bio); } io->io_bio = NULL; } -- cgit v1.3-8-gc7d7 From 816cd71b0c723a7296d14aadb8ff1ba42f6181d2 Mon Sep 17 00:00:00 2001 From: Nicolai Stange Date: Thu, 5 May 2016 22:43:04 -0400 Subject: ext4: remove unmeetable inconsisteny check from ext4_find_extent() ext4_find_extent(), stripped down to the parts relevant to this patch, reads as ppos = 0; i = depth; while (i) { --i; ++ppos; if (unlikely(ppos > depth)) { ... ret = -EFSCORRUPTED; goto err; } } Due to the loop's bounds, the condition ppos > depth can never be met. Remove this dead code. Signed-off-by: Nicolai Stange Signed-off-by: Theodore Ts'o --- fs/ext4/extents.c | 7 ------- 1 file changed, 7 deletions(-) (limited to 'fs') diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c index c53d5a8d2a79..2a2eef9c14e4 100644 --- a/fs/ext4/extents.c +++ b/fs/ext4/extents.c @@ -912,13 +912,6 @@ ext4_find_extent(struct inode *inode, ext4_lblk_t block, eh = ext_block_hdr(bh); ppos++; - if (unlikely(ppos > depth)) { - put_bh(bh); - EXT4_ERROR_INODE(inode, - "ppos %d > depth %d", ppos, depth); - ret = -EFSCORRUPTED; - goto err; - } path[ppos].p_bh = bh; path[ppos].p_hdr = eh; } -- cgit v1.3-8-gc7d7 From 58409edd2d5cc24716cb9ce690803696c5118503 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Wed, 4 May 2016 11:46:10 +0200 Subject: btrfs: kill unused writepage_io_hook callback It seems to be long time unused, since 2008 and 6885f308b5570 ("Btrfs: Misc 2.6.25 updates"). Propagating the removal touches some code but has no functional effect. Signed-off-by: David Sterba --- fs/btrfs/extent_io.c | 39 ++++++++++++++++----------------------- fs/btrfs/extent_io.h | 1 - 2 files changed, 16 insertions(+), 24 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index d247fc0eea19..d22b0777c7e0 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -3368,6 +3368,8 @@ static noinline_for_stack int __extent_writepage_io(struct inode *inode, while (cur <= end) { u64 em_end; + unsigned long max_nr; + if (cur >= i_size) { if (tree->ops && tree->ops->writepage_end_io_hook) tree->ops->writepage_end_io_hook(page, cur, @@ -3423,32 +3425,23 @@ static noinline_for_stack int __extent_writepage_io(struct inode *inode, continue; } - if (tree->ops && tree->ops->writepage_io_hook) { - ret = tree->ops->writepage_io_hook(page, cur, - cur + iosize - 1); - } else { - ret = 0; + max_nr = (i_size >> PAGE_SHIFT) + 1; + + set_range_writeback(tree, cur, cur + iosize - 1); + if (!PageWriteback(page)) { + btrfs_err(BTRFS_I(inode)->root->fs_info, + "page %lu not writeback, cur %llu end %llu", + page->index, cur, end); } - if (ret) { - SetPageError(page); - } else { - unsigned long max_nr = (i_size >> PAGE_SHIFT) + 1; - set_range_writeback(tree, cur, cur + iosize - 1); - if (!PageWriteback(page)) { - btrfs_err(BTRFS_I(inode)->root->fs_info, - "page %lu not writeback, cur %llu end %llu", - page->index, cur, end); - } + ret = submit_extent_page(write_flags, tree, wbc, page, + sector, iosize, pg_offset, + bdev, &epd->bio, max_nr, + end_bio_extent_writepage, + 0, 0, 0, false); + if (ret) + SetPageError(page); - ret = submit_extent_page(write_flags, tree, wbc, page, - sector, iosize, pg_offset, - bdev, &epd->bio, max_nr, - end_bio_extent_writepage, - 0, 0, 0, false); - if (ret) - SetPageError(page); - } cur = cur + iosize; pg_offset += iosize; nr++; diff --git a/fs/btrfs/extent_io.h b/fs/btrfs/extent_io.h index b5e0ade90e88..981f402bf754 100644 --- a/fs/btrfs/extent_io.h +++ b/fs/btrfs/extent_io.h @@ -71,7 +71,6 @@ struct extent_io_ops { u64 start, u64 end, int *page_started, unsigned long *nr_written); int (*writepage_start_hook)(struct page *page, u64 start, u64 end); - int (*writepage_io_hook)(struct page *page, u64 start, u64 end); extent_submit_bio_hook_t *submit_bio_hook; int (*merge_bio_hook)(int rw, struct page *page, unsigned long offset, size_t size, struct bio *bio, -- cgit v1.3-8-gc7d7 From 2355ac8495c1fb8ae58bdfe191489682538c697d Mon Sep 17 00:00:00 2001 From: David Sterba Date: Thu, 28 Apr 2016 11:59:34 +0200 Subject: btrfs: ioctl: reorder exclusive op check in RM_DEV Move the op exclusivity check before the other code (same as in ADD_DEV). Signed-off-by: David Sterba --- fs/btrfs/ioctl.c | 23 +++++++++++------------ 1 file changed, 11 insertions(+), 12 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index 03df0bbb445d..45a19c6b2341 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -2680,32 +2680,31 @@ static long btrfs_ioctl_rm_dev(struct file *file, void __user *arg) if (ret) return ret; - vol_args = memdup_user(arg, sizeof(*vol_args)); - if (IS_ERR(vol_args)) { - ret = PTR_ERR(vol_args); - goto err_drop; - } - - vol_args->name[BTRFS_PATH_NAME_MAX] = '\0'; - if (atomic_xchg(&root->fs_info->mutually_exclusive_operation_running, 1)) { ret = BTRFS_ERROR_DEV_EXCL_RUN_IN_PROGRESS; + goto out_drop_write; + } + + vol_args = memdup_user(arg, sizeof(*vol_args)); + if (IS_ERR(vol_args)) { + ret = PTR_ERR(vol_args); goto out; } + vol_args->name[BTRFS_PATH_NAME_MAX] = '\0'; mutex_lock(&root->fs_info->volume_mutex); ret = btrfs_rm_device(root, vol_args->name); mutex_unlock(&root->fs_info->volume_mutex); - atomic_set(&root->fs_info->mutually_exclusive_operation_running, 0); if (!ret) btrfs_info(root->fs_info, "disk deleted %s",vol_args->name); - -out: kfree(vol_args); -err_drop: +out: + atomic_set(&root->fs_info->mutually_exclusive_operation_running, 0); +out_drop_write: mnt_drop_write_file(file); + return ret; } -- cgit v1.3-8-gc7d7 From 88be159c905a2b4f6d59afa352bef410afb6af02 Mon Sep 17 00:00:00 2001 From: "Austin S. Hemmelgarn" Date: Wed, 23 Mar 2016 14:22:59 -0400 Subject: btrfs: allow balancing to dup with multi-device Currently, we don't allow the user to try and rebalance to a dup profile on a multi-device filesystem. In most cases, this is a perfectly sensible restriction as raid1 uses the same amount of space and provides better protection. However, when reshaping a multi-device filesystem down to a single device filesystem, this requires the user to convert metadata and system chunks to single profile before deleting devices, and then convert again to dup, which leaves a period of time where metadata integrity is reduced. This patch removes the single-device-only restriction from converting to dup profile to remove this potential data integrity reduction. Signed-off-by: Austin S. Hemmelgarn Signed-off-by: David Sterba --- fs/btrfs/volumes.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 745a619241c7..83dfed667637 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -3700,10 +3700,8 @@ int btrfs_balance(struct btrfs_balance_control *bctl, num_devices--; } btrfs_dev_replace_unlock(&fs_info->dev_replace, 0); - allowed = BTRFS_AVAIL_ALLOC_BIT_SINGLE; - if (num_devices == 1) - allowed |= BTRFS_BLOCK_GROUP_DUP; - else if (num_devices > 1) + allowed = BTRFS_AVAIL_ALLOC_BIT_SINGLE | BTRFS_BLOCK_GROUP_DUP; + if (num_devices > 1) allowed |= (BTRFS_BLOCK_GROUP_RAID0 | BTRFS_BLOCK_GROUP_RAID1); if (num_devices > 2) allowed |= BTRFS_BLOCK_GROUP_RAID5; -- cgit v1.3-8-gc7d7 From ae02d1bd070767e109f4a6f1bb1f466e9698a355 Mon Sep 17 00:00:00 2001 From: Luis de Bethencourt Date: Wed, 30 Mar 2016 21:53:38 +0100 Subject: btrfs: fix mixed block count of available space Metadata for mixed block is already accounted in total data and should not be counted as part of the free metadata space. Signed-off-by: Luis de Bethencourt Link: https://bugzilla.kernel.org/show_bug.cgi?id=114281 Signed-off-by: David Sterba --- fs/btrfs/super.c | 16 +++++++++++++--- 1 file changed, 13 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index 00b8f37cc306..bdca79ce45f1 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -2051,6 +2051,7 @@ static int btrfs_statfs(struct dentry *dentry, struct kstatfs *buf) struct btrfs_block_rsv *block_rsv = &fs_info->global_block_rsv; int ret; u64 thresh = 0; + int mixed = 0; /* * holding chunk_muext to avoid allocating new chunks, holding @@ -2076,8 +2077,17 @@ static int btrfs_statfs(struct dentry *dentry, struct kstatfs *buf) } } } - if (found->flags & BTRFS_BLOCK_GROUP_METADATA) - total_free_meta += found->disk_total - found->disk_used; + + /* + * Metadata in mixed block goup profiles are accounted in data + */ + if (!mixed && found->flags & BTRFS_BLOCK_GROUP_METADATA) { + if (found->flags & BTRFS_BLOCK_GROUP_DATA) + mixed = 1; + else + total_free_meta += found->disk_total - + found->disk_used; + } total_used += found->disk_used; } @@ -2115,7 +2125,7 @@ static int btrfs_statfs(struct dentry *dentry, struct kstatfs *buf) */ thresh = 4 * 1024 * 1024; - if (total_free_meta - thresh < block_rsv->size) + if (!mixed && total_free_meta - thresh < block_rsv->size) buf->f_bavail = 0; buf->f_type = BTRFS_SUPER_MAGIC; -- cgit v1.3-8-gc7d7 From 41b34accb265e3a20211a7a8ef3625678f1c6ec7 Mon Sep 17 00:00:00 2001 From: Luis de Bethencourt Date: Wed, 30 Mar 2016 23:18:14 +0100 Subject: btrfs: avoid overflowing f_bfree Since mixed block groups accounting isn't byte-accurate and f_bree is an unsigned integer, it could overflow. Avoid this. Signed-off-by: Luis de Bethencourt Suggested-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/super.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index bdca79ce45f1..fe03efb5bec0 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -2100,7 +2100,11 @@ static int btrfs_statfs(struct dentry *dentry, struct kstatfs *buf) /* Account global block reserve as used, it's in logical size already */ spin_lock(&block_rsv->lock); - buf->f_bfree -= block_rsv->size >> bits; + /* Mixed block groups accounting is not byte-accurate, avoid overflow */ + if (buf->f_bfree >= block_rsv->size >> bits) + buf->f_bfree -= block_rsv->size >> bits; + else + buf->f_bfree = 0; spin_unlock(&block_rsv->lock); buf->f_bavail = div_u64(total_free_data, factor); -- cgit v1.3-8-gc7d7 From f5ecec3ce21f706e9e7a330b2e8e5a2941927b46 Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Wed, 13 Apr 2016 09:40:59 +0300 Subject: btrfs: send: silence an integer overflow warning The "sizeof(*arg->clone_sources) * arg->clone_sources_count" expression can overflow. It causes several static checker warnings. It's all under CAP_SYS_ADMIN so it's not that serious but lets silence the warnings. Signed-off-by: Dan Carpenter Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/send.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/send.c b/fs/btrfs/send.c index 8d358c547c59..ec433795fa71 100644 --- a/fs/btrfs/send.c +++ b/fs/btrfs/send.c @@ -5978,6 +5978,12 @@ long btrfs_ioctl_send(struct file *mnt_file, void __user *arg_) goto out; } + if (arg->clone_sources_count > + ULLONG_MAX / sizeof(*arg->clone_sources)) { + ret = -EINVAL; + goto out; + } + if (!access_ok(VERIFY_READ, arg->clone_sources, sizeof(*arg->clone_sources) * arg->clone_sources_count)) { -- cgit v1.3-8-gc7d7 From 779bf3fefa835cb52a07457c8acac6f2f66f2493 Mon Sep 17 00:00:00 2001 From: Anand Jain Date: Mon, 18 Apr 2016 16:51:23 +0800 Subject: btrfs: fix lock dep warning, move scratch dev out of device_list_mutex and uuid_mutex When the replace target fails, the target device will be taken out of fs device list, scratch + update_dev_time and freed. However we could do the scratch + update_dev_time and free part after the device has been taken out of device list, so that we don't have to hold the device_list_mutex and uuid_mutex locks. Reported issue: [ 5375.718845] ====================================================== [ 5375.718846] [ INFO: possible circular locking dependency detected ] [ 5375.718849] 4.4.5-scst31x-debug-11+ #40 Not tainted [ 5375.718849] ------------------------------------------------------- [ 5375.718851] btrfs-health/4662 is trying to acquire lock: [ 5375.718861] (sb_writers){.+.+.+}, at: [] __sb_start_write+0xb7/0xf0 [ 5375.718862] [ 5375.718862] but task is already holding lock: [ 5375.718907] (&fs_devs->device_list_mutex){+.+.+.}, at: [] btrfs_destroy_dev_replace_tgtdev+0x3c/0x150 [btrfs] [ 5375.718907] [ 5375.718907] which lock already depends on the new lock. [ 5375.718907] [ 5375.718908] [ 5375.718908] the existing dependency chain (in reverse order) is: [ 5375.718911] [ 5375.718911] -> #3 (&fs_devs->device_list_mutex){+.+.+.}: [ 5375.718917] [] lock_acquire+0xce/0x1e0 [ 5375.718921] [] mutex_lock_nested+0x69/0x3c0 [ 5375.718940] [] btrfs_show_devname+0x36/0x210 [btrfs] [ 5375.718945] [] show_vfsmnt+0x49/0x150 [ 5375.718948] [] m_show+0x17/0x20 [ 5375.718951] [] seq_read+0x2d8/0x3b0 [ 5375.718955] [] __vfs_read+0x28/0xd0 [ 5375.718959] [] vfs_read+0x86/0x130 [ 5375.718962] [] SyS_read+0x49/0xa0 [ 5375.718966] [] entry_SYSCALL_64_fastpath+0x16/0x7a [ 5375.718968] [ 5375.718968] -> #2 (namespace_sem){+++++.}: [ 5375.718971] [] lock_acquire+0xce/0x1e0 [ 5375.718974] [] down_write+0x49/0x80 [ 5375.718977] [] lock_mount+0x43/0x1c0 [ 5375.718979] [] do_add_mount+0x23/0xd0 [ 5375.718982] [] do_mount+0x27b/0xe30 [ 5375.718985] [] SyS_mount+0x8c/0xd0 [ 5375.718988] [] entry_SYSCALL_64_fastpath+0x16/0x7a [ 5375.718991] [ 5375.718991] -> #1 (&sb->s_type->i_mutex_key#5){+.+.+.}: [ 5375.718994] [] lock_acquire+0xce/0x1e0 [ 5375.718996] [] mutex_lock_nested+0x69/0x3c0 [ 5375.719001] [] path_openat+0x468/0x1360 [ 5375.719004] [] do_filp_open+0x7e/0xe0 [ 5375.719007] [] do_sys_open+0x12b/0x210 [ 5375.719010] [] SyS_open+0x1e/0x20 [ 5375.719013] [] entry_SYSCALL_64_fastpath+0x16/0x7a [ 5375.719015] [ 5375.719015] -> #0 (sb_writers){.+.+.+}: [ 5375.719018] [] __lock_acquire+0x17ba/0x1ae0 [ 5375.719021] [] lock_acquire+0xce/0x1e0 [ 5375.719026] [] percpu_down_read+0x4f/0xa0 [ 5375.719028] [] __sb_start_write+0xb7/0xf0 [ 5375.719031] [] mnt_want_write+0x24/0x50 [ 5375.719035] [] path_openat+0xd32/0x1360 [ 5375.719037] [] do_filp_open+0x7e/0xe0 [ 5375.719040] [] file_open_name+0xe4/0x130 [ 5375.719043] [] filp_open+0x33/0x60 [ 5375.719073] [] update_dev_time+0x16/0x40 [btrfs] [ 5375.719099] [] btrfs_scratch_superblocks+0x4e/0x90 [btrfs] [ 5375.719123] [] btrfs_destroy_dev_replace_tgtdev+0x65/0x150 [btrfs] [ 5375.719150] [] btrfs_dev_replace_finishing+0x6b0/0x990 [btrfs] [ 5375.719175] [] btrfs_dev_replace_start+0x33e/0x540 [btrfs] [ 5375.719199] [] btrfs_auto_replace_start+0xf8/0x140 [btrfs] [ 5375.719222] [] health_kthread+0x246/0x490 [btrfs] [ 5375.719225] [] kthread+0xef/0x110 [ 5375.719229] [] ret_from_fork+0x3f/0x70 [ 5375.719230] [ 5375.719230] other info that might help us debug this: [ 5375.719230] [ 5375.719233] Chain exists of: [ 5375.719233] sb_writers --> namespace_sem --> &fs_devs->device_list_mutex [ 5375.719233] [ 5375.719234] Possible unsafe locking scenario: [ 5375.719234] [ 5375.719234] CPU0 CPU1 [ 5375.719235] ---- ---- [ 5375.719236] lock(&fs_devs->device_list_mutex); [ 5375.719238] lock(namespace_sem); [ 5375.719239] lock(&fs_devs->device_list_mutex); [ 5375.719241] lock(sb_writers); [ 5375.719241] [ 5375.719241] *** DEADLOCK *** [ 5375.719241] [ 5375.719243] 4 locks held by btrfs-health/4662: [ 5375.719266] #0: (&fs_info->health_mutex){+.+.+.}, at: [] health_kthread+0x63/0x490 [btrfs] [ 5375.719293] #1: (&fs_info->dev_replace.lock_finishing_cancel_unmount){+.+.+.}, at: [] btrfs_dev_replace_finishing+0x41/0x990 [btrfs] [ 5375.719319] #2: (uuid_mutex){+.+.+.}, at: [] btrfs_destroy_dev_replace_tgtdev+0x20/0x150 [btrfs] [ 5375.719343] #3: (&fs_devs->device_list_mutex){+.+.+.}, at: [] btrfs_destroy_dev_replace_tgtdev+0x3c/0x150 [btrfs] [ 5375.719343] [ 5375.719343] stack backtrace: [ 5375.719347] CPU: 2 PID: 4662 Comm: btrfs-health Not tainted 4.4.5-scst31x-debug-11+ #40 [ 5375.719348] Hardware name: Supermicro SYS-6018R-WTRT/X10DRW-iT, BIOS 1.0c 01/07/2015 [ 5375.719352] 0000000000000000 ffff880856f73880 ffffffff813529e3 ffffffff826182a0 [ 5375.719354] ffffffff8260c090 ffff880856f738c0 ffffffff810d667c ffff880856f73930 [ 5375.719357] ffff880861f32b40 ffff880861f32b68 0000000000000003 0000000000000004 [ 5375.719357] Call Trace: [ 5375.719363] [] dump_stack+0x85/0xc2 [ 5375.719366] [] print_circular_bug+0x1ec/0x260 [ 5375.719369] [] __lock_acquire+0x17ba/0x1ae0 [ 5375.719373] [] ? debug_lockdep_rcu_enabled+0x1d/0x20 [ 5375.719376] [] lock_acquire+0xce/0x1e0 [ 5375.719378] [] ? __sb_start_write+0xb7/0xf0 [ 5375.719383] [] percpu_down_read+0x4f/0xa0 [ 5375.719385] [] ? __sb_start_write+0xb7/0xf0 [ 5375.719387] [] __sb_start_write+0xb7/0xf0 [ 5375.719389] [] mnt_want_write+0x24/0x50 [ 5375.719393] [] path_openat+0xd32/0x1360 [ 5375.719415] [] ? btrfs_congested_fn+0x180/0x180 [btrfs] [ 5375.719418] [] ? debug_lockdep_rcu_enabled+0x1d/0x20 [ 5375.719420] [] do_filp_open+0x7e/0xe0 [ 5375.719423] [] ? rcu_read_lock_sched_held+0x6d/0x80 [ 5375.719426] [] ? kmem_cache_alloc+0x26b/0x5d0 [ 5375.719430] [] ? getname_kernel+0x34/0x120 [ 5375.719433] [] file_open_name+0xe4/0x130 [ 5375.719436] [] filp_open+0x33/0x60 [ 5375.719462] [] update_dev_time+0x16/0x40 [btrfs] [ 5375.719485] [] btrfs_scratch_superblocks+0x4e/0x90 [btrfs] [ 5375.719506] [] btrfs_destroy_dev_replace_tgtdev+0x65/0x150 [btrfs] [ 5375.719530] [] btrfs_dev_replace_finishing+0x6b0/0x990 [btrfs] [ 5375.719554] [] ? btrfs_dev_replace_finishing+0x553/0x990 [btrfs] [ 5375.719576] [] btrfs_dev_replace_start+0x33e/0x540 [btrfs] [ 5375.719598] [] btrfs_auto_replace_start+0xf8/0x140 [btrfs] [ 5375.719621] [] health_kthread+0x246/0x490 [btrfs] [ 5375.719641] [] ? health_kthread+0x138/0x490 [btrfs] [ 5375.719661] [] ? btrfs_congested_fn+0x180/0x180 [btrfs] [ 5375.719663] [] kthread+0xef/0x110 [ 5375.719666] [] ? kthread_create_on_node+0x200/0x200 [ 5375.719669] [] ret_from_fork+0x3f/0x70 [ 5375.719672] [] ? kthread_create_on_node+0x200/0x200 [ 5375.719697] ------------[ cut here ]------------ Signed-off-by: Anand Jain Reported-by: Yauhen Kharuzhy Signed-off-by: David Sterba --- fs/btrfs/volumes.c | 17 ++++++++++++----- 1 file changed, 12 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 83dfed667637..7bd553d42e0d 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -2024,10 +2024,9 @@ void btrfs_destroy_dev_replace_tgtdev(struct btrfs_fs_info *fs_info, btrfs_sysfs_rm_device_link(fs_info->fs_devices, tgtdev); - if (tgtdev->bdev) { - btrfs_scratch_superblocks(tgtdev->bdev, tgtdev->name->str); + if (tgtdev->bdev) fs_info->fs_devices->open_devices--; - } + fs_info->fs_devices->num_devices--; next_device = list_entry(fs_info->fs_devices->devices.next, @@ -2038,10 +2037,18 @@ void btrfs_destroy_dev_replace_tgtdev(struct btrfs_fs_info *fs_info, fs_info->fs_devices->latest_bdev = next_device->bdev; list_del_rcu(&tgtdev->dev_list); - call_rcu(&tgtdev->rcu, free_device); - mutex_unlock(&fs_info->fs_devices->device_list_mutex); mutex_unlock(&uuid_mutex); + + /* + * The update_dev_time() with in btrfs_scratch_superblocks() + * may lead to a call to btrfs_show_devname() which will try + * to hold device_list_mutex. And here this device + * is already out of device list, so we don't have to hold + * the device_list_mutex lock. + */ + btrfs_scratch_superblocks(tgtdev->bdev, tgtdev->name->str); + call_rcu(&tgtdev->rcu, free_device); } static int btrfs_find_device_by_path(struct btrfs_root *root, char *device_path, -- cgit v1.3-8-gc7d7 From 6ff48ce06b07255a6459cd8b816a110971a81f00 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Mon, 11 Apr 2016 18:40:08 +0200 Subject: btrfs: send: use vmalloc only as fallback for send_buf Signed-off-by: David Sterba --- fs/btrfs/send.c | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/send.c b/fs/btrfs/send.c index ec433795fa71..b6e2c6ec4ee5 100644 --- a/fs/btrfs/send.c +++ b/fs/btrfs/send.c @@ -6028,10 +6028,13 @@ long btrfs_ioctl_send(struct file *mnt_file, void __user *arg_) sctx->clone_roots_cnt = arg->clone_sources_count; sctx->send_max_size = BTRFS_SEND_BUF_SIZE; - sctx->send_buf = vmalloc(sctx->send_max_size); + sctx->send_buf = kmalloc(sctx->send_max_size, GFP_KERNEL | __GFP_NOWARN); if (!sctx->send_buf) { - ret = -ENOMEM; - goto out; + sctx->send_buf = vmalloc(sctx->send_max_size); + if (!sctx->send_buf) { + ret = -ENOMEM; + goto out; + } } sctx->read_buf = vmalloc(BTRFS_SEND_READ_SIZE); @@ -6220,7 +6223,7 @@ out: fput(sctx->send_filp); vfree(sctx->clone_roots); - vfree(sctx->send_buf); + kvfree(sctx->send_buf); vfree(sctx->read_buf); name_cache_free(sctx); -- cgit v1.3-8-gc7d7 From eb5b75fe2e61a9ba907785b70318736112b0cf93 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Mon, 11 Apr 2016 18:40:08 +0200 Subject: btrfs: send: use vmalloc only as fallback for read_buf Signed-off-by: David Sterba --- fs/btrfs/send.c | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/send.c b/fs/btrfs/send.c index b6e2c6ec4ee5..4e950ebbef53 100644 --- a/fs/btrfs/send.c +++ b/fs/btrfs/send.c @@ -6037,10 +6037,13 @@ long btrfs_ioctl_send(struct file *mnt_file, void __user *arg_) } } - sctx->read_buf = vmalloc(BTRFS_SEND_READ_SIZE); + sctx->read_buf = kmalloc(BTRFS_SEND_READ_SIZE, GFP_KERNEL | __GFP_NOWARN); if (!sctx->read_buf) { - ret = -ENOMEM; - goto out; + sctx->read_buf = vmalloc(BTRFS_SEND_READ_SIZE); + if (!sctx->read_buf) { + ret = -ENOMEM; + goto out; + } } sctx->pending_dir_moves = RB_ROOT; @@ -6224,7 +6227,7 @@ out: vfree(sctx->clone_roots); kvfree(sctx->send_buf); - vfree(sctx->read_buf); + kvfree(sctx->read_buf); name_cache_free(sctx); -- cgit v1.3-8-gc7d7 From e55d1153dbf48485a74eb4bf4eefeaedcf1486a9 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Mon, 11 Apr 2016 18:52:02 +0200 Subject: btrfs: send: use temporary variable to store allocation size We're going to use the argument multiple times later. Signed-off-by: David Sterba --- fs/btrfs/send.c | 14 ++++++++------ 1 file changed, 8 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/send.c b/fs/btrfs/send.c index 4e950ebbef53..4f85a47c2f55 100644 --- a/fs/btrfs/send.c +++ b/fs/btrfs/send.c @@ -5939,6 +5939,7 @@ long btrfs_ioctl_send(struct file *mnt_file, void __user *arg_) u32 i; u64 *clone_sources_tmp = NULL; int clone_sources_to_rollback = 0; + unsigned alloc_size; int sort_clone_roots = 0; int index; @@ -6050,24 +6051,25 @@ long btrfs_ioctl_send(struct file *mnt_file, void __user *arg_) sctx->waiting_dir_moves = RB_ROOT; sctx->orphan_dirs = RB_ROOT; - sctx->clone_roots = vzalloc(sizeof(struct clone_root) * - (arg->clone_sources_count + 1)); + alloc_size = sizeof(struct clone_root) * (arg->clone_sources_count + 1); + + sctx->clone_roots = vzalloc(alloc_size); if (!sctx->clone_roots) { ret = -ENOMEM; goto out; } + alloc_size = arg->clone_sources_count * sizeof(*arg->clone_sources); + if (arg->clone_sources_count) { - clone_sources_tmp = vmalloc(arg->clone_sources_count * - sizeof(*arg->clone_sources)); + clone_sources_tmp = vmalloc(alloc_size); if (!clone_sources_tmp) { ret = -ENOMEM; goto out; } ret = copy_from_user(clone_sources_tmp, arg->clone_sources, - arg->clone_sources_count * - sizeof(*arg->clone_sources)); + alloc_size); if (ret) { ret = -EFAULT; goto out; -- cgit v1.3-8-gc7d7 From c03d01f3404282712b9fd280297f133860c91c93 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Mon, 11 Apr 2016 18:40:08 +0200 Subject: btrfs: send: use vmalloc only as fallback for clone_roots Signed-off-by: David Sterba --- fs/btrfs/send.c | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/send.c b/fs/btrfs/send.c index 4f85a47c2f55..5a5d37b37150 100644 --- a/fs/btrfs/send.c +++ b/fs/btrfs/send.c @@ -6053,10 +6053,13 @@ long btrfs_ioctl_send(struct file *mnt_file, void __user *arg_) alloc_size = sizeof(struct clone_root) * (arg->clone_sources_count + 1); - sctx->clone_roots = vzalloc(alloc_size); + sctx->clone_roots = kzalloc(alloc_size, GFP_KERNEL | __GFP_NOWARN); if (!sctx->clone_roots) { - ret = -ENOMEM; - goto out; + sctx->clone_roots = vzalloc(alloc_size); + if (!sctx->clone_roots) { + ret = -ENOMEM; + goto out; + } } alloc_size = arg->clone_sources_count * sizeof(*arg->clone_sources); @@ -6227,7 +6230,7 @@ out: if (sctx->send_filp) fput(sctx->send_filp); - vfree(sctx->clone_roots); + kvfree(sctx->clone_roots); kvfree(sctx->send_buf); kvfree(sctx->read_buf); -- cgit v1.3-8-gc7d7 From 2f91306a37809907474a06c1defdb1ff50be06f0 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Mon, 11 Apr 2016 18:40:08 +0200 Subject: btrfs: send: use vmalloc only as fallback for clone_sources_tmp Signed-off-by: David Sterba --- fs/btrfs/send.c | 13 ++++++++----- 1 file changed, 8 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/send.c b/fs/btrfs/send.c index 5a5d37b37150..6a8c86074aa4 100644 --- a/fs/btrfs/send.c +++ b/fs/btrfs/send.c @@ -6065,10 +6065,13 @@ long btrfs_ioctl_send(struct file *mnt_file, void __user *arg_) alloc_size = arg->clone_sources_count * sizeof(*arg->clone_sources); if (arg->clone_sources_count) { - clone_sources_tmp = vmalloc(alloc_size); + clone_sources_tmp = kmalloc(alloc_size, GFP_KERNEL | __GFP_NOWARN); if (!clone_sources_tmp) { - ret = -ENOMEM; - goto out; + clone_sources_tmp = vmalloc(alloc_size); + if (!clone_sources_tmp) { + ret = -ENOMEM; + goto out; + } } ret = copy_from_user(clone_sources_tmp, arg->clone_sources, @@ -6106,7 +6109,7 @@ long btrfs_ioctl_send(struct file *mnt_file, void __user *arg_) sctx->clone_roots[i].root = clone_root; clone_sources_to_rollback = i + 1; } - vfree(clone_sources_tmp); + kvfree(clone_sources_tmp); clone_sources_tmp = NULL; } @@ -6224,7 +6227,7 @@ out: btrfs_root_dec_send_in_progress(sctx->parent_root); kfree(arg); - vfree(clone_sources_tmp); + kvfree(clone_sources_tmp); if (sctx) { if (sctx->send_filp) -- cgit v1.3-8-gc7d7 From 153519559a39725c5a45269256fec0efb81bcd1f Mon Sep 17 00:00:00 2001 From: David Sterba Date: Mon, 11 Apr 2016 18:40:08 +0200 Subject: btrfs: clone: use vmalloc only as fallback for nodesize bufer Signed-off-by: David Sterba --- fs/btrfs/ioctl.c | 13 ++++++++----- 1 file changed, 8 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index f545f81f642d..cfa7d47b1ba0 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -3468,13 +3468,16 @@ static int btrfs_clone(struct inode *src, struct inode *inode, u64 last_dest_end = destoff; ret = -ENOMEM; - buf = vmalloc(root->nodesize); - if (!buf) - return ret; + buf = kmalloc(root->nodesize, GFP_KERNEL | __GFP_NOWARN); + if (!buf) { + buf = vmalloc(root->nodesize); + if (!buf) + return ret; + } path = btrfs_alloc_path(); if (!path) { - vfree(buf); + kvfree(buf); return ret; } @@ -3775,7 +3778,7 @@ process_slot: out: btrfs_free_path(path); - vfree(buf); + kvfree(buf); return ret; } -- cgit v1.3-8-gc7d7 From 49a3c4d9b64d2e5d6fd268d53153945d04cb5559 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Thu, 24 Mar 2016 17:49:22 +0100 Subject: btrfs: use dynamic allocation for root item in create_subvol The size of root item is more than 400 bytes, which is quite a lot of stack space. As we do IO from inside the subvolume ioctls, we should keep the stack usage low in case the filesystem is on top of other layers (NFS, device mapper, iscsi, etc). Reviewed-by: Tsutomu Itoh Signed-off-by: David Sterba --- fs/btrfs/ioctl.c | 65 ++++++++++++++++++++++++++++++++------------------------ 1 file changed, 37 insertions(+), 28 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index cfa7d47b1ba0..798f58e7338e 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -439,7 +439,7 @@ static noinline int create_subvol(struct inode *dir, { struct btrfs_trans_handle *trans; struct btrfs_key key; - struct btrfs_root_item root_item; + struct btrfs_root_item *root_item; struct btrfs_inode_item *inode_item; struct extent_buffer *leaf; struct btrfs_root *root = BTRFS_I(dir)->root; @@ -455,16 +455,22 @@ static noinline int create_subvol(struct inode *dir, u64 qgroup_reserved; uuid_le new_uuid; + root_item = kzalloc(sizeof(*root_item), GFP_KERNEL); + if (!root_item) + return -ENOMEM; + ret = btrfs_find_free_objectid(root->fs_info->tree_root, &objectid); if (ret) - return ret; + goto fail_free; /* * Don't create subvolume whose level is not zero. Or qgroup will be * screwed up since it assume subvolme qgroup's level to be 0. */ - if (btrfs_qgroup_level(objectid)) - return -ENOSPC; + if (btrfs_qgroup_level(objectid)) { + ret = -ENOSPC; + goto fail_free; + } btrfs_init_block_rsv(&block_rsv, BTRFS_BLOCK_RSV_TEMP); /* @@ -474,14 +480,14 @@ static noinline int create_subvol(struct inode *dir, ret = btrfs_subvolume_reserve_metadata(root, &block_rsv, 8, &qgroup_reserved, false); if (ret) - return ret; + goto fail_free; trans = btrfs_start_transaction(root, 0); if (IS_ERR(trans)) { ret = PTR_ERR(trans); btrfs_subvolume_release_metadata(root, &block_rsv, qgroup_reserved); - return ret; + goto fail_free; } trans->block_rsv = &block_rsv; trans->bytes_reserved = block_rsv.size; @@ -509,47 +515,45 @@ static noinline int create_subvol(struct inode *dir, BTRFS_UUID_SIZE); btrfs_mark_buffer_dirty(leaf); - memset(&root_item, 0, sizeof(root_item)); - - inode_item = &root_item.inode; + inode_item = &root_item->inode; btrfs_set_stack_inode_generation(inode_item, 1); btrfs_set_stack_inode_size(inode_item, 3); btrfs_set_stack_inode_nlink(inode_item, 1); btrfs_set_stack_inode_nbytes(inode_item, root->nodesize); btrfs_set_stack_inode_mode(inode_item, S_IFDIR | 0755); - btrfs_set_root_flags(&root_item, 0); - btrfs_set_root_limit(&root_item, 0); + btrfs_set_root_flags(root_item, 0); + btrfs_set_root_limit(root_item, 0); btrfs_set_stack_inode_flags(inode_item, BTRFS_INODE_ROOT_ITEM_INIT); - btrfs_set_root_bytenr(&root_item, leaf->start); - btrfs_set_root_generation(&root_item, trans->transid); - btrfs_set_root_level(&root_item, 0); - btrfs_set_root_refs(&root_item, 1); - btrfs_set_root_used(&root_item, leaf->len); - btrfs_set_root_last_snapshot(&root_item, 0); + btrfs_set_root_bytenr(root_item, leaf->start); + btrfs_set_root_generation(root_item, trans->transid); + btrfs_set_root_level(root_item, 0); + btrfs_set_root_refs(root_item, 1); + btrfs_set_root_used(root_item, leaf->len); + btrfs_set_root_last_snapshot(root_item, 0); - btrfs_set_root_generation_v2(&root_item, - btrfs_root_generation(&root_item)); + btrfs_set_root_generation_v2(root_item, + btrfs_root_generation(root_item)); uuid_le_gen(&new_uuid); - memcpy(root_item.uuid, new_uuid.b, BTRFS_UUID_SIZE); - btrfs_set_stack_timespec_sec(&root_item.otime, cur_time.tv_sec); - btrfs_set_stack_timespec_nsec(&root_item.otime, cur_time.tv_nsec); - root_item.ctime = root_item.otime; - btrfs_set_root_ctransid(&root_item, trans->transid); - btrfs_set_root_otransid(&root_item, trans->transid); + memcpy(root_item->uuid, new_uuid.b, BTRFS_UUID_SIZE); + btrfs_set_stack_timespec_sec(&root_item->otime, cur_time.tv_sec); + btrfs_set_stack_timespec_nsec(&root_item->otime, cur_time.tv_nsec); + root_item->ctime = root_item->otime; + btrfs_set_root_ctransid(root_item, trans->transid); + btrfs_set_root_otransid(root_item, trans->transid); btrfs_tree_unlock(leaf); free_extent_buffer(leaf); leaf = NULL; - btrfs_set_root_dirid(&root_item, new_dirid); + btrfs_set_root_dirid(root_item, new_dirid); key.objectid = objectid; key.offset = 0; key.type = BTRFS_ROOT_ITEM_KEY; ret = btrfs_insert_root(trans, root->fs_info->tree_root, &key, - &root_item); + root_item); if (ret) goto fail; @@ -601,12 +605,13 @@ static noinline int create_subvol(struct inode *dir, BUG_ON(ret); ret = btrfs_uuid_tree_add(trans, root->fs_info->uuid_root, - root_item.uuid, BTRFS_UUID_KEY_SUBVOL, + root_item->uuid, BTRFS_UUID_KEY_SUBVOL, objectid); if (ret) btrfs_abort_transaction(trans, root, ret); fail: + kfree(root_item); trans->block_rsv = NULL; trans->bytes_reserved = 0; btrfs_subvolume_release_metadata(root, &block_rsv, qgroup_reserved); @@ -629,6 +634,10 @@ fail: d_instantiate(dentry, inode); } return ret; + +fail_free: + kfree(root_item); + return ret; } static void btrfs_wait_for_no_snapshoting_writes(struct btrfs_root *root) -- cgit v1.3-8-gc7d7 From e6c11f9a462e9ef4876d5e1539a6c06eded4c793 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Thu, 24 Mar 2016 18:00:53 +0100 Subject: btrfs: reuse existing variable in scrub_stripe, reduce stack usage The key variable occupies 17 bytes, the key_start is used once, we can simply reuse existing 'key' for that purpose. As the key is not a simple type, compiler doest not do it on itself. Signed-off-by: David Sterba --- fs/btrfs/scrub.c | 19 +++++++++---------- 1 file changed, 9 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c index 4678f03e878e..2ff2876656ba 100644 --- a/fs/btrfs/scrub.c +++ b/fs/btrfs/scrub.c @@ -3070,7 +3070,6 @@ static noinline_for_stack int scrub_stripe(struct scrub_ctx *sctx, int slot; u64 nstripes; struct extent_buffer *l; - struct btrfs_key key; u64 physical; u64 logical; u64 logic_end; @@ -3079,7 +3078,7 @@ static noinline_for_stack int scrub_stripe(struct scrub_ctx *sctx, int mirror_num; struct reada_control *reada1; struct reada_control *reada2; - struct btrfs_key key_start; + struct btrfs_key key; struct btrfs_key key_end; u64 increment = map->stripe_len; u64 offset; @@ -3158,21 +3157,21 @@ static noinline_for_stack int scrub_stripe(struct scrub_ctx *sctx, scrub_blocked_if_needed(fs_info); /* FIXME it might be better to start readahead at commit root */ - key_start.objectid = logical; - key_start.type = BTRFS_EXTENT_ITEM_KEY; - key_start.offset = (u64)0; + key.objectid = logical; + key.type = BTRFS_EXTENT_ITEM_KEY; + key.offset = (u64)0; key_end.objectid = logic_end; key_end.type = BTRFS_METADATA_ITEM_KEY; key_end.offset = (u64)-1; - reada1 = btrfs_reada_add(root, &key_start, &key_end); + reada1 = btrfs_reada_add(root, &key, &key_end); - key_start.objectid = BTRFS_EXTENT_CSUM_OBJECTID; - key_start.type = BTRFS_EXTENT_CSUM_KEY; - key_start.offset = logical; + key.objectid = BTRFS_EXTENT_CSUM_OBJECTID; + key.type = BTRFS_EXTENT_CSUM_KEY; + key.offset = logical; key_end.objectid = BTRFS_EXTENT_CSUM_OBJECTID; key_end.type = BTRFS_EXTENT_CSUM_KEY; key_end.offset = logic_end; - reada2 = btrfs_reada_add(csum_root, &key_start, &key_end); + reada2 = btrfs_reada_add(csum_root, &key, &key_end); if (!IS_ERR(reada1)) btrfs_reada_wait(reada1); -- cgit v1.3-8-gc7d7 From ee6111386a1b304f8bf589d36810d53e3b27ee20 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Fri, 23 Jan 2015 18:43:31 +0100 Subject: btrfs: add read-only check to sysfs handler of features We don't want to trigger the change on a read-only filesystem, similar to what the label handler does. Signed-off-by: David Sterba --- fs/btrfs/sysfs.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/sysfs.c b/fs/btrfs/sysfs.c index 539e7b5e3f86..6a6bb600b1ff 100644 --- a/fs/btrfs/sysfs.c +++ b/fs/btrfs/sysfs.c @@ -120,6 +120,9 @@ static ssize_t btrfs_feature_attr_store(struct kobject *kobj, if (!fs_info) return -EPERM; + if (fs_info->sb->s_flags & MS_RDONLY) + return -EROFS; + ret = kstrtoul(skip_spaces(buf), 0, &val); if (ret) return ret; -- cgit v1.3-8-gc7d7 From 66ac9fe7bacf9fa76c472efc7a7aaa590c7bce6a Mon Sep 17 00:00:00 2001 From: David Sterba Date: Tue, 26 Apr 2016 16:03:57 +0200 Subject: btrfs: add check to sysfs handler of label Add a sanity check for the fs_info as we will dereference it, similar to what the 'store features' handler does. Signed-off-by: David Sterba --- fs/btrfs/sysfs.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/sysfs.c b/fs/btrfs/sysfs.c index 6a6bb600b1ff..3d14618ce54b 100644 --- a/fs/btrfs/sysfs.c +++ b/fs/btrfs/sysfs.c @@ -377,6 +377,9 @@ static ssize_t btrfs_label_store(struct kobject *kobj, struct btrfs_fs_info *fs_info = to_fs_info(kobj); size_t p_len; + if (!fs_info) + return -EPERM; + if (fs_info->sb->s_flags & MS_RDONLY) return -EROFS; -- cgit v1.3-8-gc7d7 From ee17fc8005287d2d6ca7cab6e814e5043d773735 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Tue, 26 Apr 2016 16:22:06 +0200 Subject: btrfs: sysfs: protect reading label by lock If the label setting ioctl races with sysfs label handler, we could get mixed result in the output, part old part new. We should either get the old or new label. The chances to hit this race are low. Signed-off-by: David Sterba --- fs/btrfs/sysfs.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/sysfs.c b/fs/btrfs/sysfs.c index 3d14618ce54b..4879656bda3c 100644 --- a/fs/btrfs/sysfs.c +++ b/fs/btrfs/sysfs.c @@ -367,7 +367,13 @@ static ssize_t btrfs_label_show(struct kobject *kobj, { struct btrfs_fs_info *fs_info = to_fs_info(kobj); char *label = fs_info->super_copy->label; - return snprintf(buf, PAGE_SIZE, label[0] ? "%s\n" : "%s", label); + ssize_t ret; + + spin_lock(&fs_info->super_lock); + ret = snprintf(buf, PAGE_SIZE, label[0] ? "%s\n" : "%s", label); + spin_unlock(&fs_info->super_lock); + + return ret; } static ssize_t btrfs_label_store(struct kobject *kobj, -- cgit v1.3-8-gc7d7 From 3d8da67817606380fdadfa483d4dba5c3a5446c6 Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Tue, 26 Apr 2016 17:53:31 -0700 Subject: Btrfs: fix divide error upon chunk's stripe_len The struct 'map_lookup' uses type int for @stripe_len, while btrfs_chunk_stripe_len() can return a u64 value, and it may end up with @stripe_len being undefined value and it can lead to 'divide error' in __btrfs_map_block(). This changes 'map_lookup' to use type u64 for stripe_len, also right now we only use BTRFS_STRIPE_LEN for stripe_len, so this adds a valid checker for BTRFS_STRIPE_LEN. Reported-by: Vegard Nossum Reported-by: Quentin Casasnovas Signed-off-by: Liu Bo Reviewed-by: David Sterba [ folded division fix to scrub_raid56_parity ] Signed-off-by: David Sterba --- fs/btrfs/scrub.c | 2 +- fs/btrfs/volumes.c | 2 +- fs/btrfs/volumes.h | 2 +- 3 files changed, 3 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c index 2ff2876656ba..96d2a0de35a8 100644 --- a/fs/btrfs/scrub.c +++ b/fs/btrfs/scrub.c @@ -2860,7 +2860,7 @@ static noinline_for_stack int scrub_raid56_parity(struct scrub_ctx *sctx, int extent_mirror_num; int stop_loop = 0; - nsectors = map->stripe_len / root->sectorsize; + nsectors = div_u64(map->stripe_len, root->sectorsize); bitmap_len = scrub_calc_parity_bitmap_len(nsectors); sparity = kzalloc(sizeof(struct scrub_parity) + 2 * bitmap_len, GFP_NOFS); diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 7bd553d42e0d..f13e2bcc1398 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -6254,7 +6254,7 @@ static int read_one_chunk(struct btrfs_root *root, struct btrfs_key *key, "invalid chunk length %llu", length); return -EIO; } - if (!is_power_of_2(stripe_len)) { + if (!is_power_of_2(stripe_len) || stripe_len != BTRFS_STRIPE_LEN) { btrfs_err(root->fs_info, "invalid chunk stripe length: %llu", stripe_len); return -EIO; diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h index 1939ebde63df..7507be74f7da 100644 --- a/fs/btrfs/volumes.h +++ b/fs/btrfs/volumes.h @@ -347,7 +347,7 @@ struct map_lookup { u64 type; int io_align; int io_width; - int stripe_len; + u64 stripe_len; int sector_size; int num_stripes; int sub_stripes; -- cgit v1.3-8-gc7d7 From e042d1ec4417981dfe9331e47b76f17929bc2ffe Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Tue, 12 Apr 2016 12:54:40 -0400 Subject: Btrfs: remove BUG_ON()'s in btrfs_map_block btrfs_map_block can go horribly wrong in the face of fs corruption, lets agree to not be assholes and panic at any possible chance things are all fucked up. Signed-off-by: Josef Bacik [ removed type casts ] Signed-off-by: David Sterba --- fs/btrfs/volumes.c | 18 ++++++++++++++++-- 1 file changed, 16 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index f13e2bcc1398..8d3ceb0fd239 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -5290,7 +5290,15 @@ static int __btrfs_map_block(struct btrfs_fs_info *fs_info, int rw, stripe_nr = div64_u64(stripe_nr, stripe_len); stripe_offset = stripe_nr * stripe_len; - BUG_ON(offset < stripe_offset); + if (offset < stripe_offset) { + btrfs_crit(fs_info, "stripe math has gone wrong, " + "stripe_offset=%llu, offset=%llu, start=%llu, " + "logical=%llu, stripe_len=%llu", + stripe_offset, offset, em->start, logical, + stripe_len); + free_extent_map(em); + return -EINVAL; + } /* stripe_offset is the offset of this block in its stripe*/ stripe_offset = offset - stripe_offset; @@ -5531,7 +5539,13 @@ static int __btrfs_map_block(struct btrfs_fs_info *fs_info, int rw, &stripe_index); mirror_num = stripe_index + 1; } - BUG_ON(stripe_index >= map->num_stripes); + if (stripe_index >= map->num_stripes) { + btrfs_crit(fs_info, "stripe index math went horribly wrong, " + "got stripe_index=%u, num_stripes=%u", + stripe_index, map->num_stripes); + ret = -EINVAL; + goto out; + } num_alloc_stripes = num_stripes; if (dev_replace_is_ongoing) { -- cgit v1.3-8-gc7d7 From 2473114981a36b4f0f57cf6e9548037d547a71b7 Mon Sep 17 00:00:00 2001 From: Ashish Samant Date: Fri, 29 Apr 2016 18:33:59 -0700 Subject: btrfs: Fix BUG_ON condition in scrub_setup_recheck_block() pagev array in scrub_block{} is of size SCRUB_MAX_PAGES_PER_BLOCK. page_index should be checked with the same to trigger BUG_ON(). Signed-off-by: Ashish Samant Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/scrub.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c index 96d2a0de35a8..d270c700ed31 100644 --- a/fs/btrfs/scrub.c +++ b/fs/btrfs/scrub.c @@ -1350,7 +1350,7 @@ static int scrub_setup_recheck_block(struct scrub_block *original_sblock, recover->bbio = bbio; recover->map_length = mapped_length; - BUG_ON(page_index >= SCRUB_PAGES_PER_RD_BIO); + BUG_ON(page_index >= SCRUB_MAX_PAGES_PER_BLOCK); nmirrors = min(scrub_nr_raid_mirrors(bbio), BTRFS_MAX_MIRRORS); -- cgit v1.3-8-gc7d7 From 48b3b9d401ec86899a52003b37331190a35a81a6 Mon Sep 17 00:00:00 2001 From: Anand Jain Date: Tue, 12 Apr 2016 21:36:16 +0800 Subject: btrfs: fix lock dep warning move scratch super outside of chunk_mutex Move scratch super outside of the chunk lock to avoid below lockdep warning. The better place to scratch super is in the function btrfs_rm_dev_replace_free_srcdev() just before free_device, which is outside of the chunk lock as well. To reproduce: (fresh boot) mkfs.btrfs -f -draid5 -mraid5 /dev/sdc /dev/sdd /dev/sde mount /dev/sdc /btrfs dd if=/dev/zero of=/btrfs/tf1 bs=4096 count=100 (get devmgt from https://github.com/asj/devmgt.git) devmgt detach /dev/sde dd if=/dev/zero of=/btrfs/tf1 bs=4096 count=100 sync btrfs replace start -Brf 3 /dev/sdf /btrfs <-- devmgt attach host7 ====================================================== [ INFO: possible circular locking dependency detected ] 4.6.0-rc2asj+ #1 Not tainted --------------------------------------------------- btrfs/2174 is trying to acquire lock: (sb_writers){.+.+.+}, at: [] __sb_start_write+0xb4/0xf0 but task is already holding lock: (&fs_info->chunk_mutex){+.+.+.}, at: [] btrfs_dev_replace_finishing+0x145/0x980 [btrfs] which lock already depends on the new lock. Chain exists of: sb_writers --> &fs_devs->device_list_mutex --> &fs_info->chunk_mutex Possible unsafe locking scenario: CPU0 CPU1 ---- ---- lock(&fs_info->chunk_mutex); lock(&fs_devs->device_list_mutex); lock(&fs_info->chunk_mutex); lock(sb_writers); *** DEADLOCK *** -> #0 (sb_writers){.+.+.+}: [] __lock_acquire+0x1bc5/0x1ee0 [] lock_acquire+0xbe/0x210 [] percpu_down_read+0x4a/0xa0 [] __sb_start_write+0xb4/0xf0 [] mnt_want_write+0x24/0x50 [] path_openat+0x952/0x1190 [] do_filp_open+0x91/0x100 [] file_open_name+0xfc/0x140 [] filp_open+0x33/0x60 [] update_dev_time+0x16/0x40 [btrfs] [] btrfs_scratch_superblocks+0x5d/0xb0 [btrfs] [] btrfs_rm_dev_replace_remove_srcdev+0xae/0xd0 [btrfs] [] btrfs_dev_replace_finishing+0x4b5/0x980 [btrfs] [] btrfs_dev_replace_start+0x358/0x530 [btrfs] Signed-off-by: Anand Jain Signed-off-by: David Sterba --- fs/btrfs/volumes.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 8d3ceb0fd239..c5e0d2d431dc 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -1972,11 +1972,8 @@ void btrfs_rm_dev_replace_remove_srcdev(struct btrfs_fs_info *fs_info, if (srcdev->missing) fs_devices->missing_devices--; - if (srcdev->writeable) { + if (srcdev->writeable) fs_devices->rw_devices--; - /* zero out the old super if it is writable */ - btrfs_scratch_superblocks(srcdev->bdev, srcdev->name->str); - } if (srcdev->bdev) fs_devices->open_devices--; @@ -1987,6 +1984,10 @@ void btrfs_rm_dev_replace_free_srcdev(struct btrfs_fs_info *fs_info, { struct btrfs_fs_devices *fs_devices = srcdev->fs_devices; + if (srcdev->writeable) { + /* zero out the old super if it is writable */ + btrfs_scratch_superblocks(srcdev->bdev, srcdev->name->str); + } call_rcu(&srcdev->rcu, free_device); /* -- cgit v1.3-8-gc7d7 From 7ab19625a911f7568ec85302e3aa7a64186006c8 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Wed, 4 May 2016 11:32:00 +0200 Subject: btrfs: add write protection to SET_FEATURES ioctl Perform the want_write check if we get far enough to do any writes. Signed-off-by: David Sterba --- fs/btrfs/ioctl.c | 16 +++++++++++++--- 1 file changed, 13 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index 798f58e7338e..1e8ce5247a81 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -5406,9 +5406,15 @@ static int btrfs_ioctl_set_features(struct file *file, void __user *arg) if (ret) return ret; + ret = mnt_want_write_file(file); + if (ret) + return ret; + trans = btrfs_start_transaction(root, 0); - if (IS_ERR(trans)) - return PTR_ERR(trans); + if (IS_ERR(trans)) { + ret = PTR_ERR(trans); + goto out_drop_write; + } spin_lock(&root->fs_info->super_lock); newflags = btrfs_super_compat_flags(super_block); @@ -5427,7 +5433,11 @@ static int btrfs_ioctl_set_features(struct file *file, void __user *arg) btrfs_set_super_incompat_flags(super_block, newflags); spin_unlock(&root->fs_info->super_lock); - return btrfs_commit_transaction(trans, root); + ret = btrfs_commit_transaction(trans, root); +out_drop_write: + mnt_drop_write_file(file); + + return ret; } long btrfs_ioctl(struct file *file, unsigned int -- cgit v1.3-8-gc7d7 From 58d7bbf81f54667e36940d5f4b5609606efa597b Mon Sep 17 00:00:00 2001 From: David Sterba Date: Wed, 4 May 2016 14:10:47 +0200 Subject: btrfs: ioctl: reorder exclusive op check in RM_DEV Move the op exclusivity check before the other code (same as in ADD_DEV). Signed-off-by: David Sterba --- fs/btrfs/ioctl.c | 23 +++++++++++------------ 1 file changed, 11 insertions(+), 12 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index 1e8ce5247a81..c81d6daefe74 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -2689,32 +2689,31 @@ static long btrfs_ioctl_rm_dev(struct file *file, void __user *arg) if (ret) return ret; - vol_args = memdup_user(arg, sizeof(*vol_args)); - if (IS_ERR(vol_args)) { - ret = PTR_ERR(vol_args); - goto err_drop; - } - - vol_args->name[BTRFS_PATH_NAME_MAX] = '\0'; - if (atomic_xchg(&root->fs_info->mutually_exclusive_operation_running, 1)) { ret = BTRFS_ERROR_DEV_EXCL_RUN_IN_PROGRESS; + goto out_drop_write; + } + + vol_args = memdup_user(arg, sizeof(*vol_args)); + if (IS_ERR(vol_args)) { + ret = PTR_ERR(vol_args); goto out; } + vol_args->name[BTRFS_PATH_NAME_MAX] = '\0'; mutex_lock(&root->fs_info->volume_mutex); ret = btrfs_rm_device(root, vol_args->name); mutex_unlock(&root->fs_info->volume_mutex); - atomic_set(&root->fs_info->mutually_exclusive_operation_running, 0); if (!ret) btrfs_info(root->fs_info, "disk deleted %s",vol_args->name); - -out: kfree(vol_args); -err_drop: +out: + atomic_set(&root->fs_info->mutually_exclusive_operation_running, 0); +out_drop_write: mnt_drop_write_file(file); + return ret; } -- cgit v1.3-8-gc7d7 From 2f3165ecf103599f82bf0ea254039db335fb5005 Mon Sep 17 00:00:00 2001 From: Zygo Blaxell Date: Thu, 5 May 2016 00:23:49 -0400 Subject: btrfs: don't force mounts to wait for cleaner_kthread to delete one or more subvolumes During a mount, we start the cleaner kthread first because the transaction kthread wants to wake up the cleaner kthread. We start the transaction kthread next because everything in btrfs wants transactions. We do reloc recovery in the thread that was doing the original mount call once the transaction kthread is running. This means that the cleaner kthread could already be running when reloc recovery happens (e.g. if a snapshot delete was started before a crash). Relocation does not play well with the cleaner kthread, so a mutex was added in commit 5f3164813b90f7dbcb5c3ab9006906222ce471b7 "Btrfs: fix race between balance recovery and root deletion" to prevent both from being active at the same time. If the cleaner kthread is already holding the mutex by the time we get to btrfs_recover_relocation, the mount will be blocked until at least one deleted subvolume is cleaned (possibly more if the mount process doesn't get the lock right away). During this time (which could be an arbitrarily long time on a large/slow filesystem), the mount process is stuck and the filesystem is unnecessarily inaccessible. Fix this by locking cleaner_mutex before we start cleaner_kthread, and unlocking the mutex after mount no longer requires it. This ensures that the mounting process will not be blocked by the cleaner kthread. The cleaner kthread is already prepared for mutex contention and will just go to sleep until the mutex is available. Signed-off-by: Zygo Blaxell Reviewed-by: Filipe Manana Signed-off-by: David Sterba --- fs/btrfs/disk-io.c | 18 +++++++++++++++--- 1 file changed, 15 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 4e47849d7427..070c1dad42bd 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -2517,6 +2517,7 @@ int open_ctree(struct super_block *sb, int num_backups_tried = 0; int backup_index = 0; int max_active; + bool cleaner_mutex_locked = false; tree_root = fs_info->tree_root = btrfs_alloc_root(fs_info, GFP_KERNEL); chunk_root = fs_info->chunk_root = btrfs_alloc_root(fs_info, GFP_KERNEL); @@ -2997,6 +2998,13 @@ retry_root_backup: goto fail_sysfs; } + /* + * Hold the cleaner_mutex thread here so that we don't block + * for a long time on btrfs_recover_relocation. cleaner_kthread + * will wait for us to finish mounting the filesystem. + */ + mutex_lock(&fs_info->cleaner_mutex); + cleaner_mutex_locked = true; fs_info->cleaner_kthread = kthread_run(cleaner_kthread, tree_root, "btrfs-cleaner"); if (IS_ERR(fs_info->cleaner_kthread)) @@ -3056,10 +3064,8 @@ retry_root_backup: ret = btrfs_cleanup_fs_roots(fs_info); if (ret) goto fail_qgroup; - - mutex_lock(&fs_info->cleaner_mutex); + /* We locked cleaner_mutex before creating cleaner_kthread. */ ret = btrfs_recover_relocation(tree_root); - mutex_unlock(&fs_info->cleaner_mutex); if (ret < 0) { printk(KERN_WARNING "BTRFS: failed to recover relocation\n"); @@ -3067,6 +3073,8 @@ retry_root_backup: goto fail_qgroup; } } + mutex_unlock(&fs_info->cleaner_mutex); + cleaner_mutex_locked = false; location.objectid = BTRFS_FS_TREE_OBJECTID; location.type = BTRFS_ROOT_ITEM_KEY; @@ -3180,6 +3188,10 @@ fail_cleaner: filemap_write_and_wait(fs_info->btree_inode->i_mapping); fail_sysfs: + if (cleaner_mutex_locked) { + mutex_unlock(&fs_info->cleaner_mutex); + cleaner_mutex_locked = false; + } btrfs_sysfs_remove_mounted(fs_info); fail_fsdev_sysfs: -- cgit v1.3-8-gc7d7 From 68cd4ce2caf22a81833eb1abfa075eb1cc39bfe2 Mon Sep 17 00:00:00 2001 From: Bob Peterson Date: Mon, 2 May 2016 11:53:35 -0500 Subject: GFS2: Refactor gfs2_remove_from_journal This patch makes two simple changes to function gfs2_remove_from_journal. First, it removes the parameter that specifies the transaction. Since it's always passed in as current->journal_info, we might as well set that in the function rather than passing it in. Second, it changes the meta parameter to use an enum to make the code more clear. Signed-off-by: Bob Peterson Acked-by: Steven Whitehouse --- fs/gfs2/aops.c | 2 +- fs/gfs2/meta_io.c | 7 ++++--- fs/gfs2/meta_io.h | 8 ++++++-- 3 files changed, 11 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/aops.c b/fs/gfs2/aops.c index 58dd0061134b..e21a931ced53 100644 --- a/fs/gfs2/aops.c +++ b/fs/gfs2/aops.c @@ -977,7 +977,7 @@ static void gfs2_discard(struct gfs2_sbd *sdp, struct buffer_head *bh) if (!list_empty(&bd->bd_list) && !buffer_pinned(bh)) list_del_init(&bd->bd_list); else - gfs2_remove_from_journal(bh, current->journal_info, 0); + gfs2_remove_from_journal(bh, REMOVE_JDATA); } bh->b_bdev = NULL; clear_buffer_mapped(bh); diff --git a/fs/gfs2/meta_io.c b/fs/gfs2/meta_io.c index e137d96f1b17..8f3f1b863687 100644 --- a/fs/gfs2/meta_io.c +++ b/fs/gfs2/meta_io.c @@ -325,18 +325,19 @@ int gfs2_meta_wait(struct gfs2_sbd *sdp, struct buffer_head *bh) return 0; } -void gfs2_remove_from_journal(struct buffer_head *bh, struct gfs2_trans *tr, int meta) +void gfs2_remove_from_journal(struct buffer_head *bh, int meta) { struct address_space *mapping = bh->b_page->mapping; struct gfs2_sbd *sdp = gfs2_mapping2sbd(mapping); struct gfs2_bufdata *bd = bh->b_private; + struct gfs2_trans *tr = current->journal_info; int was_pinned = 0; if (test_clear_buffer_pinned(bh)) { trace_gfs2_pin(bd, 0); atomic_dec(&sdp->sd_log_pinned); list_del_init(&bd->bd_list); - if (meta) + if (meta == REMOVE_META) tr->tr_num_buf_rm++; else tr->tr_num_databuf_rm++; @@ -376,7 +377,7 @@ void gfs2_meta_wipe(struct gfs2_inode *ip, u64 bstart, u32 blen) if (bh) { lock_buffer(bh); gfs2_log_lock(sdp); - gfs2_remove_from_journal(bh, current->journal_info, 1); + gfs2_remove_from_journal(bh, REMOVE_META); gfs2_log_unlock(sdp); unlock_buffer(bh); brelse(bh); diff --git a/fs/gfs2/meta_io.h b/fs/gfs2/meta_io.h index c5086c8af5ed..ffdf6aa3509d 100644 --- a/fs/gfs2/meta_io.h +++ b/fs/gfs2/meta_io.h @@ -57,8 +57,12 @@ extern int gfs2_meta_read(struct gfs2_glock *gl, u64 blkno, int flags, extern int gfs2_meta_wait(struct gfs2_sbd *sdp, struct buffer_head *bh); extern struct buffer_head *gfs2_getbuf(struct gfs2_glock *gl, u64 blkno, int create); -extern void gfs2_remove_from_journal(struct buffer_head *bh, - struct gfs2_trans *tr, int meta); +enum { + REMOVE_JDATA = 0, + REMOVE_META = 1, +}; + +extern void gfs2_remove_from_journal(struct buffer_head *bh, int meta); extern void gfs2_meta_wipe(struct gfs2_inode *ip, u64 bstart, u32 blen); extern int gfs2_meta_indirect_buffer(struct gfs2_inode *ip, int height, u64 num, struct buffer_head **bhp); -- cgit v1.3-8-gc7d7 From b7a15f3dbe3f3d009360273b62a633e47a1a65d6 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Wed, 27 Apr 2016 16:07:56 -0700 Subject: f2fs: introduce macros for proc entries This adds macros to be used multiple proc entries. Signed-off-by: Jaegeuk Kim --- fs/f2fs/super.c | 25 ++++++++++++++----------- 1 file changed, 14 insertions(+), 11 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 19a85cf83f40..b006de60e99b 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -745,19 +745,22 @@ static int segment_info_seq_show(struct seq_file *seq, void *offset) return 0; } -static int segment_info_open_fs(struct inode *inode, struct file *file) -{ - return single_open(file, segment_info_seq_show, PDE_DATA(inode)); -} - -static const struct file_operations f2fs_seq_segment_info_fops = { - .owner = THIS_MODULE, - .open = segment_info_open_fs, - .read = seq_read, - .llseek = seq_lseek, - .release = single_release, +#define F2FS_PROC_FILE_DEF(_name) \ +static int _name##_open_fs(struct inode *inode, struct file *file) \ +{ \ + return single_open(file, _name##_seq_show, PDE_DATA(inode)); \ +} \ + \ +static const struct file_operations f2fs_seq_##_name##_fops = { \ + .owner = THIS_MODULE, \ + .open = _name##_open_fs, \ + .read = seq_read, \ + .llseek = seq_lseek, \ + .release = single_release, \ }; +F2FS_PROC_FILE_DEF(segment_info); + static void default_options(struct f2fs_sb_info *sbi) { /* init some FS parameters */ -- cgit v1.3-8-gc7d7 From f00d6fa727a4c4977241bd21910cafc139f77989 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Wed, 27 Apr 2016 16:29:05 -0700 Subject: f2fs: add proc entry to show valid block bitmap This patch adds a new proc entry to show segment information in more detail. Signed-off-by: Jaegeuk Kim --- fs/f2fs/super.c | 32 +++++++++++++++++++++++++++++++- 1 file changed, 31 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index b006de60e99b..90d4b864590c 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -539,6 +539,7 @@ static void f2fs_put_super(struct super_block *sb) if (sbi->s_proc) { remove_proc_entry("segment_info", sbi->s_proc); + remove_proc_entry("segment_bits", sbi->s_proc); remove_proc_entry(sb->s_id, f2fs_proc_root); } kobject_del(&sbi->s_kobj); @@ -745,6 +746,30 @@ static int segment_info_seq_show(struct seq_file *seq, void *offset) return 0; } +static int segment_bits_seq_show(struct seq_file *seq, void *offset) +{ + struct super_block *sb = seq->private; + struct f2fs_sb_info *sbi = F2FS_SB(sb); + unsigned int total_segs = + le32_to_cpu(sbi->raw_super->segment_count_main); + int i, j; + + seq_puts(seq, "format: segment_type|valid_blocks|bitmaps\n" + "segment_type(0:HD, 1:WD, 2:CD, 3:HN, 4:WN, 5:CN)\n"); + + for (i = 0; i < total_segs; i++) { + struct seg_entry *se = get_seg_entry(sbi, i); + + seq_printf(seq, "%-10d", i); + seq_printf(seq, "%d|%-3u|", se->type, + get_valid_blocks(sbi, i, 1)); + for (j = 0; j < SIT_VBLOCK_MAP_SIZE; j++) + seq_printf(seq, "%x ", se->cur_valid_map[j]); + seq_putc(seq, '\n'); + } + return 0; +} + #define F2FS_PROC_FILE_DEF(_name) \ static int _name##_open_fs(struct inode *inode, struct file *file) \ { \ @@ -760,6 +785,7 @@ static const struct file_operations f2fs_seq_##_name##_fops = { \ }; F2FS_PROC_FILE_DEF(segment_info); +F2FS_PROC_FILE_DEF(segment_bits); static void default_options(struct f2fs_sb_info *sbi) { @@ -1541,9 +1567,12 @@ try_onemore: if (f2fs_proc_root) sbi->s_proc = proc_mkdir(sb->s_id, f2fs_proc_root); - if (sbi->s_proc) + if (sbi->s_proc) { proc_create_data("segment_info", S_IRUGO, sbi->s_proc, &f2fs_seq_segment_info_fops, sb); + proc_create_data("segment_bits", S_IRUGO, sbi->s_proc, + &f2fs_seq_segment_bits_fops, sb); + } sbi->s_kobj.kset = f2fs_kset; init_completion(&sbi->s_kobj_unregister); @@ -1619,6 +1648,7 @@ free_kobj: free_proc: if (sbi->s_proc) { remove_proc_entry("segment_info", sbi->s_proc); + remove_proc_entry("segment_bits", sbi->s_proc); remove_proc_entry(sb->s_id, f2fs_proc_root); } f2fs_destroy_stats(sbi); -- cgit v1.3-8-gc7d7 From 0414b004a894746921bbc05f05dced1e7907b092 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Fri, 29 Apr 2016 15:16:42 -0700 Subject: f2fs: introduce f2fs_kmalloc to wrap kmalloc This patch adds f2fs_kmalloc. Signed-off-by: Jaegeuk Kim --- fs/f2fs/acl.c | 4 ++-- fs/f2fs/dir.c | 2 +- fs/f2fs/f2fs.h | 5 +++++ fs/f2fs/gc.c | 2 +- fs/f2fs/inline.c | 2 +- 5 files changed, 10 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/acl.c b/fs/f2fs/acl.c index c8f25f7241f0..d757d79a4d71 100644 --- a/fs/f2fs/acl.c +++ b/fs/f2fs/acl.c @@ -115,7 +115,7 @@ static void *f2fs_acl_to_disk(const struct posix_acl *acl, size_t *size) struct f2fs_acl_entry *entry; int i; - f2fs_acl = kmalloc(sizeof(struct f2fs_acl_header) + acl->a_count * + f2fs_acl = f2fs_kmalloc(sizeof(struct f2fs_acl_header) + acl->a_count * sizeof(struct f2fs_acl_entry), GFP_NOFS); if (!f2fs_acl) return ERR_PTR(-ENOMEM); @@ -175,7 +175,7 @@ static struct posix_acl *__f2fs_get_acl(struct inode *inode, int type, retval = f2fs_getxattr(inode, name_index, "", NULL, 0, dpage); if (retval > 0) { - value = kmalloc(retval, GFP_F2FS_ZERO); + value = f2fs_kmalloc(retval, GFP_F2FS_ZERO); if (!value) return ERR_PTR(-ENOMEM); retval = f2fs_getxattr(inode, name_index, "", value, diff --git a/fs/f2fs/dir.c b/fs/f2fs/dir.c index dbfc1d1375a0..6bd059591405 100644 --- a/fs/f2fs/dir.c +++ b/fs/f2fs/dir.c @@ -805,7 +805,7 @@ bool f2fs_fill_dentries(struct dir_context *ctx, struct f2fs_dentry_ptr *d, int save_len = fstr->len; int ret; - de_name.name = kmalloc(de_name.len, GFP_NOFS); + de_name.name = f2fs_kmalloc(de_name.len, GFP_NOFS); if (!de_name.name) return false; diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 6a482411e6d4..879e4d77a1bc 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -1644,6 +1644,11 @@ static inline bool f2fs_may_extent_tree(struct inode *inode) return S_ISREG(inode->i_mode); } +static inline void *f2fs_kmalloc(size_t size, gfp_t flags) +{ + return kmalloc(size, flags); +} + static inline void *f2fs_kvmalloc(size_t size, gfp_t flags) { void *ret; diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c index 19e9fafc5c70..38d56f678912 100644 --- a/fs/f2fs/gc.c +++ b/fs/f2fs/gc.c @@ -96,7 +96,7 @@ int start_gc_thread(struct f2fs_sb_info *sbi) dev_t dev = sbi->sb->s_bdev->bd_dev; int err = 0; - gc_th = kmalloc(sizeof(struct f2fs_gc_kthread), GFP_KERNEL); + gc_th = f2fs_kmalloc(sizeof(struct f2fs_gc_kthread), GFP_KERNEL); if (!gc_th) { err = -ENOMEM; goto out; diff --git a/fs/f2fs/inline.c b/fs/f2fs/inline.c index e61084cac5f1..33830b251426 100644 --- a/fs/f2fs/inline.c +++ b/fs/f2fs/inline.c @@ -465,7 +465,7 @@ static int f2fs_move_rehashed_dirents(struct inode *dir, struct page *ipage, struct f2fs_inline_dentry *backup_dentry; int err; - backup_dentry = kmalloc(sizeof(struct f2fs_inline_dentry), + backup_dentry = f2fs_kmalloc(sizeof(struct f2fs_inline_dentry), GFP_F2FS_ZERO); if (!backup_dentry) return -ENOMEM; -- cgit v1.3-8-gc7d7 From 300e129c15f0ed2f94482900a4cb65b28eb09d94 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Fri, 29 Apr 2016 16:11:53 -0700 Subject: f2fs: use f2fs_grab_cache_page instead of grab_cache_page This patch converts grab_cache_page to f2fs_grab_cache_page. Signed-off-by: Jaegeuk Kim --- fs/f2fs/checkpoint.c | 7 ++++--- fs/f2fs/inline.c | 4 ++-- fs/f2fs/node.c | 8 ++++---- 3 files changed, 10 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c index bf040b51989d..9596d61ca6a8 100644 --- a/fs/f2fs/checkpoint.c +++ b/fs/f2fs/checkpoint.c @@ -34,7 +34,7 @@ struct page *grab_meta_page(struct f2fs_sb_info *sbi, pgoff_t index) struct address_space *mapping = META_MAPPING(sbi); struct page *page = NULL; repeat: - page = grab_cache_page(mapping, index); + page = f2fs_grab_cache_page(mapping, index, false); if (!page) { cond_resched(); goto repeat; @@ -64,7 +64,7 @@ static struct page *__get_meta_page(struct f2fs_sb_info *sbi, pgoff_t index, if (unlikely(!is_meta)) fio.rw &= ~REQ_META; repeat: - page = grab_cache_page(mapping, index); + page = f2fs_grab_cache_page(mapping, index, false); if (!page) { cond_resched(); goto repeat; @@ -186,7 +186,8 @@ int ra_meta_pages(struct f2fs_sb_info *sbi, block_t start, int nrpages, BUG(); } - page = grab_cache_page(META_MAPPING(sbi), fio.new_blkaddr); + page = f2fs_grab_cache_page(META_MAPPING(sbi), + fio.new_blkaddr, false); if (!page) continue; if (PageUptodate(page)) { diff --git a/fs/f2fs/inline.c b/fs/f2fs/inline.c index 33830b251426..8a51955b68aa 100644 --- a/fs/f2fs/inline.c +++ b/fs/f2fs/inline.c @@ -161,7 +161,7 @@ int f2fs_convert_inline_inode(struct inode *inode) if (!f2fs_has_inline_data(inode)) return 0; - page = grab_cache_page(inode->i_mapping, 0); + page = f2fs_grab_cache_page(inode->i_mapping, 0, false); if (!page) return -ENOMEM; @@ -358,7 +358,7 @@ static int f2fs_move_inline_dirents(struct inode *dir, struct page *ipage, struct f2fs_dentry_block *dentry_blk; int err; - page = grab_cache_page(dir->i_mapping, 0); + page = f2fs_grab_cache_page(dir->i_mapping, 0, false); if (!page) { f2fs_put_page(ipage, 1); return -ENOMEM; diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c index f80cfb6beac6..af010f5c4ee1 100644 --- a/fs/f2fs/node.c +++ b/fs/f2fs/node.c @@ -995,7 +995,7 @@ struct page *new_node_page(struct dnode_of_data *dn, if (unlikely(is_inode_flag_set(F2FS_I(dn->inode), FI_NO_ALLOC))) return ERR_PTR(-EPERM); - page = grab_cache_page(NODE_MAPPING(sbi), dn->nid); + page = f2fs_grab_cache_page(NODE_MAPPING(sbi), dn->nid, false); if (!page) return ERR_PTR(-ENOMEM); @@ -1087,7 +1087,7 @@ void ra_node_page(struct f2fs_sb_info *sbi, nid_t nid) if (apage) return; - apage = grab_cache_page(NODE_MAPPING(sbi), nid); + apage = f2fs_grab_cache_page(NODE_MAPPING(sbi), nid, false); if (!apage) return; @@ -1128,7 +1128,7 @@ static struct page *__get_node_page(struct f2fs_sb_info *sbi, pgoff_t nid, return ERR_PTR(-ENOENT); f2fs_bug_on(sbi, check_nid_range(sbi, nid)); repeat: - page = grab_cache_page(NODE_MAPPING(sbi), nid); + page = f2fs_grab_cache_page(NODE_MAPPING(sbi), nid, false); if (!page) return ERR_PTR(-ENOMEM); @@ -2012,7 +2012,7 @@ int recover_inode_page(struct f2fs_sb_info *sbi, struct page *page) if (unlikely(old_ni.blk_addr != NULL_ADDR)) return -EINVAL; - ipage = grab_cache_page(NODE_MAPPING(sbi), ino); + ipage = f2fs_grab_cache_page(NODE_MAPPING(sbi), ino, false); if (!ipage) return -ENOMEM; -- cgit v1.3-8-gc7d7 From 73faec4d99358b79815866dd660ae2f9f6f9110a Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Fri, 29 Apr 2016 15:34:32 -0700 Subject: f2fs: add mount option to select fault injection ratio This patch adds a mount option to select fault ratio. Signed-off-by: Jaegeuk Kim --- fs/f2fs/Kconfig | 8 ++++++++ fs/f2fs/f2fs.h | 1 + fs/f2fs/super.c | 19 +++++++++++++++++++ 3 files changed, 28 insertions(+) (limited to 'fs') diff --git a/fs/f2fs/Kconfig b/fs/f2fs/Kconfig index 1f8982a957f1..378c221d68a9 100644 --- a/fs/f2fs/Kconfig +++ b/fs/f2fs/Kconfig @@ -94,3 +94,11 @@ config F2FS_IO_TRACE information and block IO patterns in the filesystem level. If unsure, say N. + +config F2FS_FAULT_INJECTION + bool "F2FS fault injection facility" + depends on F2FS_FS + help + Test F2FS to inject faults such as ENOMEM, ENOSPC, and so on. + + If unsure, say N. diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 879e4d77a1bc..0cf2a26388e5 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -56,6 +56,7 @@ #define F2FS_MOUNT_EXTENT_CACHE 0x00002000 #define F2FS_MOUNT_FORCE_FG_GC 0x00004000 #define F2FS_MOUNT_DATA_FLUSH 0x00008000 +#define F2FS_MOUNT_FAULT_INJECTION 0x00010000 #define clear_opt(sbi, option) (sbi->mount_opt.opt &= ~F2FS_MOUNT_##option) #define set_opt(sbi, option) (sbi->mount_opt.opt |= F2FS_MOUNT_##option) diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 90d4b864590c..22348791700f 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -39,6 +39,10 @@ static struct proc_dir_entry *f2fs_proc_root; static struct kmem_cache *f2fs_inode_cachep; static struct kset *f2fs_kset; +#ifdef CONFIG_F2FS_FAULT_INJECTION +u32 f2fs_fault_rate = 0; +#endif + /* f2fs-wide shrinker description */ static struct shrinker f2fs_shrinker_info = { .scan_objects = f2fs_shrink_scan, @@ -68,6 +72,7 @@ enum { Opt_noextent_cache, Opt_noinline_data, Opt_data_flush, + Opt_fault_injection, Opt_err, }; @@ -93,6 +98,7 @@ static match_table_t f2fs_tokens = { {Opt_noextent_cache, "noextent_cache"}, {Opt_noinline_data, "noinline_data"}, {Opt_data_flush, "data_flush"}, + {Opt_fault_injection, "fault_injection=%u"}, {Opt_err, NULL}, }; @@ -300,6 +306,9 @@ static int parse_options(struct super_block *sb, char *options) char *p, *name; int arg = 0; +#ifdef CONFIG_F2FS_FAULT_INJECTION + f2fs_fault_rate = 0; +#endif if (!options) return 0; @@ -433,6 +442,16 @@ static int parse_options(struct super_block *sb, char *options) case Opt_data_flush: set_opt(sbi, DATA_FLUSH); break; + case Opt_fault_injection: + if (args->from && match_int(args, &arg)) + return -EINVAL; +#ifdef CONFIG_F2FS_FAULT_INJECTION + f2fs_fault_rate = arg; +#else + f2fs_msg(sb, KERN_INFO, + "FAULT_INJECTION was not selected"); +#endif + break; default: f2fs_msg(sb, KERN_ERR, "Unrecognized mount option \"%s\" or missing value", -- cgit v1.3-8-gc7d7 From 2c63fead9e372b3b65d1883bb174df6c9820f1dd Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Fri, 29 Apr 2016 15:49:56 -0700 Subject: f2fs: inject kmalloc failure This patch injects kmalloc failure given a fault injection rate. Signed-off-by: Jaegeuk Kim --- fs/f2fs/f2fs.h | 29 +++++++++++++++++++++++++++++ fs/f2fs/super.c | 6 ++++++ 2 files changed, 35 insertions(+) (limited to 'fs') diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 0cf2a26388e5..cf03c57d0feb 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -37,6 +37,31 @@ } while (0) #endif +#ifdef CONFIG_F2FS_FAULT_INJECTION +enum { + FAULT_KMALLOC, + FAULT_MAX, +}; + +extern u32 f2fs_fault_rate; +extern atomic_t f2fs_ops; +extern char *fault_name[FAULT_MAX]; + +static inline bool time_to_inject(int type) +{ + atomic_inc(&f2fs_ops); + if (f2fs_fault_rate && (atomic_read(&f2fs_ops) >= f2fs_fault_rate)) { + atomic_set(&f2fs_ops, 0); + printk("%sF2FS-fs : inject %s in %pF\n", + KERN_INFO, + fault_name[type], + __builtin_return_address(0)); + return true; + } + return false; +} +#endif + /* * For mount options */ @@ -1647,6 +1672,10 @@ static inline bool f2fs_may_extent_tree(struct inode *inode) static inline void *f2fs_kmalloc(size_t size, gfp_t flags) { +#ifdef CONFIG_F2FS_FAULT_INJECTION + if (time_to_inject(FAULT_KMALLOC)) + return NULL; +#endif return kmalloc(size, flags); } diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 22348791700f..9a56f544da3e 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -41,6 +41,11 @@ static struct kset *f2fs_kset; #ifdef CONFIG_F2FS_FAULT_INJECTION u32 f2fs_fault_rate = 0; +atomic_t f2fs_ops; + +char *fault_name[FAULT_MAX] = { + [FAULT_KMALLOC] = "kmalloc", +}; #endif /* f2fs-wide shrinker description */ @@ -447,6 +452,7 @@ static int parse_options(struct super_block *sb, char *options) return -EINVAL; #ifdef CONFIG_F2FS_FAULT_INJECTION f2fs_fault_rate = arg; + atomic_set(&f2fs_ops, 0); #else f2fs_msg(sb, KERN_INFO, "FAULT_INJECTION was not selected"); -- cgit v1.3-8-gc7d7 From c41f3cc3ae34acdbcec328084b4b74f686c02f0d Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Fri, 29 Apr 2016 16:17:09 -0700 Subject: f2fs: inject page allocation failures This patch adds page allocation failures. Signed-off-by: Jaegeuk Kim --- fs/f2fs/f2fs.h | 9 +++++++++ fs/f2fs/super.c | 1 + 2 files changed, 10 insertions(+) (limited to 'fs') diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index cf03c57d0feb..d550a95d30f1 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -40,6 +40,7 @@ #ifdef CONFIG_F2FS_FAULT_INJECTION enum { FAULT_KMALLOC, + FAULT_PAGE_ALLOC, FAULT_MAX, }; @@ -1296,6 +1297,14 @@ static inline unsigned int valid_inode_count(struct f2fs_sb_info *sbi) static inline struct page *f2fs_grab_cache_page(struct address_space *mapping, pgoff_t index, bool for_write) { +#ifdef CONFIG_F2FS_FAULT_INJECTION + struct page *page = find_lock_page(mapping, index); + if (page) + return page; + + if (time_to_inject(FAULT_PAGE_ALLOC)) + return NULL; +#endif if (!for_write) return grab_cache_page(mapping, index); return grab_cache_page_write_begin(mapping, index, AOP_FLAG_NOFS); diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 9a56f544da3e..986d0da84e01 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -45,6 +45,7 @@ atomic_t f2fs_ops; char *fault_name[FAULT_MAX] = { [FAULT_KMALLOC] = "kmalloc", + [FAULT_PAGE_ALLOC] = "page alloc", }; #endif -- cgit v1.3-8-gc7d7 From cb78942b821380913e6810375c9ce72858e64c4f Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Fri, 29 Apr 2016 16:29:22 -0700 Subject: f2fs: inject ENOSPC failures This patch injects ENOSPC failures. Signed-off-by: Jaegeuk Kim --- fs/f2fs/checkpoint.c | 7 +++++++ fs/f2fs/dir.c | 4 ++++ fs/f2fs/f2fs.h | 10 ++++++++++ fs/f2fs/node.c | 4 ++++ fs/f2fs/super.c | 4 ++++ 5 files changed, 29 insertions(+) (limited to 'fs') diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c index 9596d61ca6a8..79da86d6cf5c 100644 --- a/fs/f2fs/checkpoint.c +++ b/fs/f2fs/checkpoint.c @@ -474,6 +474,13 @@ int acquire_orphan_inode(struct f2fs_sb_info *sbi) int err = 0; spin_lock(&im->ino_lock); + +#ifdef CONFIG_F2FS_FAULT_INJECTION + if (time_to_inject(FAULT_ORPHAN)) { + spin_unlock(&im->ino_lock); + return -ENOSPC; + } +#endif if (unlikely(im->ino_num >= sbi->max_orphans)) err = -ENOSPC; else diff --git a/fs/f2fs/dir.c b/fs/f2fs/dir.c index 6bd059591405..50f42be4ff1a 100644 --- a/fs/f2fs/dir.c +++ b/fs/f2fs/dir.c @@ -537,6 +537,10 @@ int f2fs_add_regular_entry(struct inode *dir, const struct qstr *new_name, } start: +#ifdef CONFIG_F2FS_FAULT_INJECTION + if (time_to_inject(FAULT_DIR_DEPTH)) + return -ENOSPC; +#endif if (unlikely(current_depth == MAX_DIR_HASH_DEPTH)) return -ENOSPC; diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index d550a95d30f1..8ba2f923d95a 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -41,6 +41,10 @@ enum { FAULT_KMALLOC, FAULT_PAGE_ALLOC, + FAULT_ALLOC_NID, + FAULT_ORPHAN, + FAULT_BLOCK, + FAULT_DIR_DEPTH, FAULT_MAX, }; @@ -1087,6 +1091,12 @@ static inline bool inc_valid_block_count(struct f2fs_sb_info *sbi, block_t valid_block_count; spin_lock(&sbi->stat_lock); +#ifdef CONFIG_F2FS_FAULT_INJECTION + if (time_to_inject(FAULT_BLOCK)) { + spin_unlock(&sbi->stat_lock); + return false; + } +#endif valid_block_count = sbi->total_valid_block_count + (block_t)count; if (unlikely(valid_block_count > sbi->user_block_count)) { diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c index af010f5c4ee1..78b98db48805 100644 --- a/fs/f2fs/node.c +++ b/fs/f2fs/node.c @@ -1838,6 +1838,10 @@ bool alloc_nid(struct f2fs_sb_info *sbi, nid_t *nid) struct f2fs_nm_info *nm_i = NM_I(sbi); struct free_nid *i = NULL; retry: +#ifdef CONFIG_F2FS_FAULT_INJECTION + if (time_to_inject(FAULT_ALLOC_NID)) + return false; +#endif if (unlikely(sbi->total_valid_node_count + 1 > nm_i->available_nids)) return false; diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 986d0da84e01..87654f48c55d 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -46,6 +46,10 @@ atomic_t f2fs_ops; char *fault_name[FAULT_MAX] = { [FAULT_KMALLOC] = "kmalloc", [FAULT_PAGE_ALLOC] = "page alloc", + [FAULT_ALLOC_NID] = "alloc nid", + [FAULT_ORPHAN] = "orphan", + [FAULT_BLOCK] = "no more block", + [FAULT_DIR_DEPTH] = "too big dir depth", }; #endif -- cgit v1.3-8-gc7d7 From 221149c00e64c202e6e172a9c4efad142a6b610d Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Mon, 2 May 2016 12:34:48 -0700 Subject: f2fs: revisit error handling flows This patch fixes a couple of bugs regarding to orphan inodes when handling errors. This tries to - call alloc_nid_done with add_orphan_inode in handle_failed_inode - let truncate blocks in f2fs_evict_inode - not make a bad inode due to i_mode change Signed-off-by: Jaegeuk Kim --- fs/f2fs/dir.c | 20 ++++++++++++-------- fs/f2fs/inode.c | 40 ++++++++++++++++------------------------ 2 files changed, 28 insertions(+), 32 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/dir.c b/fs/f2fs/dir.c index 50f42be4ff1a..5373f333a7d7 100644 --- a/fs/f2fs/dir.c +++ b/fs/f2fs/dir.c @@ -391,9 +391,14 @@ struct page *init_inode_metadata(struct inode *inode, struct inode *dir, return page; if (S_ISDIR(inode->i_mode)) { + /* in order to handle error case */ + get_page(page); err = make_empty_dir(inode, dir, page); - if (err) - goto error; + if (err) { + lock_page(page); + goto put_error; + } + put_page(page); } err = f2fs_init_acl(inode, dir, page, dpage); @@ -437,13 +442,12 @@ struct page *init_inode_metadata(struct inode *inode, struct inode *dir, return page; put_error: - f2fs_put_page(page, 1); -error: - /* once the failed inode becomes a bad inode, i_mode is S_IFREG */ + /* truncate empty dir pages */ truncate_inode_pages(&inode->i_data, 0); - truncate_blocks(inode, 0, false); - remove_dirty_inode(inode); - remove_inode_page(inode); + + clear_nlink(inode); + update_inode(inode, page); + f2fs_put_page(page, 1); return ERR_PTR(err); } diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c index cb269c46ac25..f4ac8512b7ba 100644 --- a/fs/f2fs/inode.c +++ b/fs/f2fs/inode.c @@ -368,10 +368,7 @@ no_delete: if (is_inode_flag_set(fi, FI_UPDATE_WRITE)) add_ino_entry(sbi, inode->i_ino, UPDATE_INO); if (is_inode_flag_set(fi, FI_FREE_NID)) { - if (err && err != -ENOENT) - alloc_nid_done(sbi, inode->i_ino); - else - alloc_nid_failed(sbi, inode->i_ino); + alloc_nid_failed(sbi, inode->i_ino); clear_inode_flag(fi, FI_FREE_NID); } @@ -397,37 +394,32 @@ out_clear: void handle_failed_inode(struct inode *inode) { struct f2fs_sb_info *sbi = F2FS_I_SB(inode); - int err = 0; + struct node_info ni; - clear_nlink(inode); - make_bad_inode(inode); + /* don't make bad inode, since it becomes a regular file. */ unlock_new_inode(inode); - i_size_write(inode, 0); - if (F2FS_HAS_BLOCKS(inode)) - err = f2fs_truncate(inode, false); - - if (!err) - err = remove_inode_page(inode); - /* - * if we skip truncate_node in remove_inode_page bacause we failed - * before, it's better to find another way to release resource of - * this inode (e.g. valid block count, node block or nid). Here we - * choose to add this inode to orphan list, so that we can call iput - * for releasing in orphan recovery flow. - * * Note: we should add inode to orphan list before f2fs_unlock_op() * so we can prevent losing this orphan when encoutering checkpoint * and following suddenly power-off. */ - if (err && err != -ENOENT) { - err = acquire_orphan_inode(sbi); - if (!err) + get_node_info(sbi, inode->i_ino, &ni); + + if (ni.blk_addr != NULL_ADDR) { + int err = acquire_orphan_inode(sbi); + if (err) { + set_sbi_flag(sbi, SBI_NEED_FSCK); + f2fs_msg(sbi->sb, KERN_WARNING, + "Too many orphan inodes, run fsck to fix."); + } else { add_orphan_inode(sbi, inode->i_ino); + } + alloc_nid_done(sbi, inode->i_ino); + } else { + set_inode_flag(F2FS_I(inode), FI_FREE_NID); } - set_inode_flag(F2FS_I(inode), FI_FREE_NID); f2fs_unlock_op(sbi); /* iput will drop the inode object */ -- cgit v1.3-8-gc7d7 From 74ef924167ad9d2897ffc0ef83c7ebf322fcc535 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Mon, 2 May 2016 22:09:56 -0700 Subject: f2fs: fix leak of orphan inode objects When unmounting filesystem, we should release all the ino entries. Signed-off-by: Jaegeuk Kim --- fs/f2fs/checkpoint.c | 6 +++--- fs/f2fs/f2fs.h | 2 +- fs/f2fs/super.c | 2 +- 3 files changed, 5 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c index 79da86d6cf5c..432a6baa61d1 100644 --- a/fs/f2fs/checkpoint.c +++ b/fs/f2fs/checkpoint.c @@ -449,12 +449,12 @@ bool exist_written_data(struct f2fs_sb_info *sbi, nid_t ino, int mode) return e ? true : false; } -void release_ino_entry(struct f2fs_sb_info *sbi) +void release_ino_entry(struct f2fs_sb_info *sbi, bool all) { struct ino_entry *e, *tmp; int i; - for (i = APPEND_INO; i <= UPDATE_INO; i++) { + for (i = all ? ORPHAN_INO: APPEND_INO; i <= UPDATE_INO; i++) { struct inode_management *im = &sbi->im[i]; spin_lock(&im->ino_lock); @@ -1106,7 +1106,7 @@ static int do_checkpoint(struct f2fs_sb_info *sbi, struct cp_control *cpc) invalidate_mapping_pages(META_MAPPING(sbi), discard_blk, discard_blk); - release_ino_entry(sbi); + release_ino_entry(sbi, false); if (unlikely(f2fs_cp_error(sbi))) return -EIO; diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 8ba2f923d95a..ea146a599b4c 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -1911,7 +1911,7 @@ void ra_meta_pages_cond(struct f2fs_sb_info *, pgoff_t); long sync_meta_pages(struct f2fs_sb_info *, enum page_type, long); void add_ino_entry(struct f2fs_sb_info *, nid_t, int type); void remove_ino_entry(struct f2fs_sb_info *, nid_t, int type); -void release_ino_entry(struct f2fs_sb_info *); +void release_ino_entry(struct f2fs_sb_info *, bool); bool exist_written_data(struct f2fs_sb_info *, nid_t, int); int acquire_orphan_inode(struct f2fs_sb_info *); void release_orphan_inode(struct f2fs_sb_info *); diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 87654f48c55d..8a28f799a24f 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -599,7 +599,7 @@ static void f2fs_put_super(struct super_block *sb) * normally superblock is clean, so we need to release this. * In addition, EIO will skip do checkpoint, we need this as well. */ - release_ino_entry(sbi); + release_ino_entry(sbi, true); release_discard_addrs(sbi); f2fs_leave_shrinker(sbi); -- cgit v1.3-8-gc7d7 From 4c0c294934b055aca765737c3b06698f51e9124d Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Tue, 3 May 2016 09:22:18 -0700 Subject: f2fs: retry to truncate blocks in -ENOMEM case This patch modifies to retry truncating node blocks in -ENOMEM case. Signed-off-by: Hou Pengyang Signed-off-by: Jaegeuk Kim --- fs/f2fs/inode.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c index f4ac8512b7ba..baf3a2a1f249 100644 --- a/fs/f2fs/inode.c +++ b/fs/f2fs/inode.c @@ -344,7 +344,7 @@ void f2fs_evict_inode(struct inode *inode) sb_start_intwrite(inode->i_sb); set_inode_flag(fi, FI_NO_ALLOC); i_size_write(inode, 0); - +retry: if (F2FS_HAS_BLOCKS(inode)) err = f2fs_truncate(inode, true); @@ -354,6 +354,12 @@ void f2fs_evict_inode(struct inode *inode) f2fs_unlock_op(sbi); } + /* give more chances, if ENOMEM case */ + if (err == -ENOMEM) { + err = 0; + goto retry; + } + sb_end_intwrite(inode->i_sb); no_delete: stat_dec_inline_xattr(inode); -- cgit v1.3-8-gc7d7 From ae8d1db34ff57b11feceeaaa26940015b14e30fb Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Wed, 4 May 2016 23:17:00 +0800 Subject: f2fs: remove unneeded readahead in find_fsync_dnodes In find_fsync_dnodes, get_tmp_page will read dnode page synchronously, previously, ra_meta_page did the same work, which is redundant, remove it. Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/recovery.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/recovery.c b/fs/f2fs/recovery.c index 58275a29e87d..29a37aae95d4 100644 --- a/fs/f2fs/recovery.c +++ b/fs/f2fs/recovery.c @@ -229,8 +229,6 @@ static int find_fsync_dnodes(struct f2fs_sb_info *sbi, struct list_head *head) curseg = CURSEG_I(sbi, CURSEG_WARM_NODE); blkaddr = NEXT_FREE_BLKADDR(sbi, curseg); - ra_meta_pages(sbi, blkaddr, 1, META_POR, true); - while (1) { struct fsync_inode_entry *entry; -- cgit v1.3-8-gc7d7 From e3bc808ca869f0070d438257e4085602636498da Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Wed, 4 May 2016 23:19:46 +0800 Subject: f2fs: remove unneeded memset when updating xattr Each of fields in struct f2fs_xattr_entry will be assigned later, so previously we don't need to memset the struct. Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/xattr.c | 1 - 1 file changed, 1 deletion(-) (limited to 'fs') diff --git a/fs/f2fs/xattr.c b/fs/f2fs/xattr.c index 152971243ad8..55c69d3d2ce3 100644 --- a/fs/f2fs/xattr.c +++ b/fs/f2fs/xattr.c @@ -528,7 +528,6 @@ static int __f2fs_setxattr(struct inode *inode, int index, * Before we come here, old entry is removed. * We just write new entry. */ - memset(last, 0, newsize); last->e_name_index = index; last->e_name_len = len; memcpy(last->e_name, name, len); -- cgit v1.3-8-gc7d7 From bd933d4faedf3a79034e67e8032d153bf1afcd83 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Wed, 4 May 2016 23:19:47 +0800 Subject: f2fs: reuse get_extent_info Reuse get_extent_info for readability. Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/extent_cache.c | 3 +-- fs/f2fs/f2fs.h | 8 ++++---- 2 files changed, 5 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/extent_cache.c b/fs/f2fs/extent_cache.c index c859bb044728..5bfcdb9b69f2 100644 --- a/fs/f2fs/extent_cache.c +++ b/fs/f2fs/extent_cache.c @@ -196,8 +196,7 @@ bool f2fs_init_extent_tree(struct inode *inode, struct f2fs_extent *i_ext) if (!i_ext || !i_ext->len) return false; - set_extent_info(&ei, le32_to_cpu(i_ext->fofs), - le32_to_cpu(i_ext->blk), le32_to_cpu(i_ext->len)); + get_extent_info(&ei, i_ext); write_lock(&et->lock); if (atomic_read(&et->node_cnt)) diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index ea146a599b4c..ccf8bf4debfc 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -428,11 +428,11 @@ struct f2fs_inode_info { }; static inline void get_extent_info(struct extent_info *ext, - struct f2fs_extent i_ext) + struct f2fs_extent *i_ext) { - ext->fofs = le32_to_cpu(i_ext.fofs); - ext->blk = le32_to_cpu(i_ext.blk); - ext->len = le32_to_cpu(i_ext.len); + ext->fofs = le32_to_cpu(i_ext->fofs); + ext->blk = le32_to_cpu(i_ext->blk); + ext->len = le32_to_cpu(i_ext->len); } static inline void set_raw_extent(struct extent_info *ext, -- cgit v1.3-8-gc7d7 From f51b4ce6c1759a64d68d314545854dc226ab69a0 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Wed, 4 May 2016 23:19:48 +0800 Subject: f2fs: shrink size of struct seg_entry Restructure struct seg_entry to eliminate holes in it, after that, in 32-bits machine, it reduces size from 32 bytes to 24 bytes; in 64-bits machine, it reduces size from 56 bytes to 40 bytes. Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/segment.h | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/segment.h b/fs/f2fs/segment.h index 975c33df65c7..7a756ff5a36d 100644 --- a/fs/f2fs/segment.h +++ b/fs/f2fs/segment.h @@ -158,16 +158,17 @@ struct victim_sel_policy { }; struct seg_entry { - unsigned short valid_blocks; /* # of valid blocks */ + unsigned int type:6; /* segment type like CURSEG_XXX_TYPE */ + unsigned int valid_blocks:10; /* # of valid blocks */ + unsigned int ckpt_valid_blocks:10; /* # of valid blocks last cp */ + unsigned int padding:6; /* padding */ unsigned char *cur_valid_map; /* validity bitmap of blocks */ /* * # of valid blocks and the validity bitmap stored in the the last * checkpoint pack. This information is used by the SSR mode. */ - unsigned short ckpt_valid_blocks; - unsigned char *ckpt_valid_map; + unsigned char *ckpt_valid_map; /* validity bitmap of blocks last cp */ unsigned char *discard_map; - unsigned char type; /* segment type like CURSEG_XXX_TYPE */ unsigned long long mtime; /* modification time of the segment */ }; -- cgit v1.3-8-gc7d7 From 29234b1d6d2d4ecbd8f85831285c20329617859f Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Wed, 4 May 2016 19:48:53 -0700 Subject: f2fs: don't worry about inode leak in evict_inode Even if an inode failed to release its blocks, it should be kept in an orphan inode list, so it will be released later. Signed-off-by: Jaegeuk Kim --- fs/f2fs/inode.c | 16 ++-------------- 1 file changed, 2 insertions(+), 14 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c index baf3a2a1f249..689d6914e2e4 100644 --- a/fs/f2fs/inode.c +++ b/fs/f2fs/inode.c @@ -377,20 +377,8 @@ no_delete: alloc_nid_failed(sbi, inode->i_ino); clear_inode_flag(fi, FI_FREE_NID); } - - if (err && err != -ENOENT) { - if (!exist_written_data(sbi, inode->i_ino, ORPHAN_INO)) { - /* - * get here because we failed to release resource - * of inode previously, reminder our user to run fsck - * for fixing. - */ - set_sbi_flag(sbi, SBI_NEED_FSCK); - f2fs_msg(sbi->sb, KERN_WARNING, - "inode (ino:%lu) resource leak, run fsck " - "to fix this issue!", inode->i_ino); - } - } + f2fs_bug_on(sbi, err && + !exist_written_data(sbi, inode->i_ino, ORPHAN_INO)); out_clear: fscrypt_put_encryption_info(inode, NULL); clear_inode(inode); -- cgit v1.3-8-gc7d7 From fb58ae22067e0595d974e3d856522c1ed6d2d7bf Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Wed, 4 May 2016 09:58:10 -0700 Subject: f2fs: remove an obsolete variable This patch removes an obsolete variable used in add_free_nid. Signed-off-by: Jaegeuk Kim --- fs/f2fs/node.c | 3 --- 1 file changed, 3 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c index 78b98db48805..264abe08bafb 100644 --- a/fs/f2fs/node.c +++ b/fs/f2fs/node.c @@ -1690,7 +1690,6 @@ static int add_free_nid(struct f2fs_sb_info *sbi, nid_t nid, bool build) struct f2fs_nm_info *nm_i = NM_I(sbi); struct free_nid *i; struct nat_entry *ne; - bool allocated = false; if (!available_free_memory(sbi, FREE_NIDS)) return -1; @@ -1704,8 +1703,6 @@ static int add_free_nid(struct f2fs_sb_info *sbi, nid_t nid, bool build) ne = __lookup_nat_cache(nm_i, nid); if (ne && (!get_nat_flag(ne, IS_CHECKPOINTED) || nat_get_blkaddr(ne) != NULL_ADDR)) - allocated = true; - if (allocated) return 0; } -- cgit v1.3-8-gc7d7 From 43473f96453f0b075c480a26ec4fc846d5fb3bd4 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Thu, 5 May 2016 19:13:02 +0800 Subject: f2fs: fix incorrect mapping in ->bmap Currently, generic_block_bmap is used in f2fs_bmap, its semantics is when the mapping is been found, return position of target physical block, otherwise return zero. But, previously, when there is no mapping info for specified logical block, f2fs_bmap will map target physical block to a uninitialized variable, which should be wrong. Fix it. Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/data.c | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index 888f1781ae0b..96b035319889 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -648,6 +648,8 @@ next_dnode: set_new_dnode(&dn, inode, NULL, NULL, 0); err = get_dnode_of_data(&dn, pgofs, mode); if (err) { + if (flag == F2FS_GET_BLOCK_BMAP) + map->m_pblk = 0; if (err == -ENOENT) { err = 0; if (map->m_next_pgofs) @@ -683,17 +685,18 @@ next_block: map->m_flags = F2FS_MAP_NEW; blkaddr = dn.data_blkaddr; } else { + if (flag == F2FS_GET_BLOCK_BMAP) { + map->m_pblk = 0; + goto sync_out; + } if (flag == F2FS_GET_BLOCK_FIEMAP && blkaddr == NULL_ADDR) { if (map->m_next_pgofs) *map->m_next_pgofs = pgofs + 1; } if (flag != F2FS_GET_BLOCK_FIEMAP || - blkaddr != NEW_ADDR) { - if (flag == F2FS_GET_BLOCK_BMAP) - err = -ENOENT; + blkaddr != NEW_ADDR) goto sync_out; - } } } -- cgit v1.3-8-gc7d7 From 09210c973af30320edc03a6325422cdd0f03b580 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Thu, 5 May 2016 19:13:03 +0800 Subject: f2fs: avoid panic when truncating to max filesize The following panic occurs when truncating inode which has inline xattr to max filesize. [] get_dnode_of_data+0x4e/0x580 [f2fs] [] ? read_node_page+0x51/0x90 [f2fs] [] ? get_node_page.part.34+0xb9/0x170 [f2fs] [] truncate_blocks+0x131/0x3f0 [f2fs] [] f2fs_truncate+0x73/0x100 [f2fs] [] f2fs_setattr+0x62/0x2a0 [f2fs] [] notify_change+0x158/0x300 [] do_truncate+0x6b/0xa0 [] ? __sb_start_write+0x49/0x100 [] do_sys_ftruncate.constprop.12+0x118/0x170 [] SyS_ftruncate+0xe/0x10 [] tracesys+0xe1/0xe6 [] get_node_path+0x210/0x220 [f2fs] --[ end trace 5fea664dfbcc6625 ]--- The reason is truncate_blocks tries to truncate all node and data blocks start from specified block offset with value of (max filesize / block size), but actually, our valid max block offset is (max filesize / block size) - 1, so f2fs detects such invalid block offset with BUG_ON in truncation path. This patch lets f2fs skip truncating data which is exceeding max filesize. Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/file.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index dc47d5c7b882..dd50f30a2f57 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -563,6 +563,9 @@ int truncate_blocks(struct inode *inode, u64 from, bool lock) free_from = (pgoff_t)F2FS_BYTES_TO_BLK(from + blocksize - 1); + if (free_from >= sbi->max_file_blocks) + goto free_partial; + if (lock) f2fs_lock_op(sbi); @@ -604,7 +607,7 @@ free_next: out: if (lock) f2fs_unlock_op(sbi); - +free_partial: /* lastly zero out the first data page */ if (!err) err = truncate_partial_data_page(inode, from, truncate_page); -- cgit v1.3-8-gc7d7 From b5a7aef1ef436ec005fef0efe31a676ec5f4ab31 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Wed, 4 May 2016 22:05:01 -0700 Subject: fscrypto/f2fs: allow fs-specific key prefix for fs encryption This patch allows fscrypto to handle a second key prefix given by filesystem. The main reason is to provide backward compatibility, since previously f2fs used "f2fs:" as a crypto prefix instead of "fscrypt:". Later, ext4 should also provide key_prefix() to give "ext4:". One concern decribed by Ted would be kinda double check overhead of prefixes. In x86, for example, validate_user_key consumes 8 ms after boot-up, which turns out derive_key_aes() consumed most of the time to load specific crypto module. After such the cold miss, it shows almost zero latencies, which treats as a negligible overhead. Note that request_key() detects wrong prefix in prior to derive_key_aes() even. Cc: Ted Tso Cc: stable@vger.kernel.org # v4.6 Signed-off-by: Jaegeuk Kim --- fs/crypto/keyinfo.c | 120 ++++++++++++++++++++++++++++++----------------- fs/f2fs/f2fs.h | 8 ++++ fs/f2fs/super.c | 13 +++++ include/linux/fscrypto.h | 1 + 4 files changed, 98 insertions(+), 44 deletions(-) (limited to 'fs') diff --git a/fs/crypto/keyinfo.c b/fs/crypto/keyinfo.c index 06f5aa478bf2..1ac263eddc4e 100644 --- a/fs/crypto/keyinfo.c +++ b/fs/crypto/keyinfo.c @@ -78,6 +78,67 @@ out: return res; } +static int validate_user_key(struct fscrypt_info *crypt_info, + struct fscrypt_context *ctx, u8 *raw_key, + u8 *prefix, int prefix_size) +{ + u8 *full_key_descriptor; + struct key *keyring_key; + struct fscrypt_key *master_key; + const struct user_key_payload *ukp; + int full_key_len = prefix_size + (FS_KEY_DESCRIPTOR_SIZE * 2) + 1; + int res; + + full_key_descriptor = kmalloc(full_key_len, GFP_NOFS); + if (!full_key_descriptor) + return -ENOMEM; + + memcpy(full_key_descriptor, prefix, prefix_size); + sprintf(full_key_descriptor + prefix_size, + "%*phN", FS_KEY_DESCRIPTOR_SIZE, + ctx->master_key_descriptor); + full_key_descriptor[full_key_len - 1] = '\0'; + keyring_key = request_key(&key_type_logon, full_key_descriptor, NULL); + kfree(full_key_descriptor); + if (IS_ERR(keyring_key)) + return PTR_ERR(keyring_key); + + if (keyring_key->type != &key_type_logon) { + printk_once(KERN_WARNING + "%s: key type must be logon\n", __func__); + res = -ENOKEY; + goto out; + } + down_read(&keyring_key->sem); + ukp = user_key_payload(keyring_key); + if (ukp->datalen != sizeof(struct fscrypt_key)) { + res = -EINVAL; + up_read(&keyring_key->sem); + goto out; + } + master_key = (struct fscrypt_key *)ukp->data; + BUILD_BUG_ON(FS_AES_128_ECB_KEY_SIZE != FS_KEY_DERIVATION_NONCE_SIZE); + + if (master_key->size != FS_AES_256_XTS_KEY_SIZE) { + printk_once(KERN_WARNING + "%s: key size incorrect: %d\n", + __func__, master_key->size); + res = -ENOKEY; + up_read(&keyring_key->sem); + goto out; + } + res = derive_key_aes(ctx->nonce, master_key->raw, raw_key); + up_read(&keyring_key->sem); + if (res) + goto out; + + crypt_info->ci_keyring_key = keyring_key; + return 0; +out: + key_put(keyring_key); + return res; +} + static void put_crypt_info(struct fscrypt_info *ci) { if (!ci) @@ -91,12 +152,7 @@ static void put_crypt_info(struct fscrypt_info *ci) int get_crypt_info(struct inode *inode) { struct fscrypt_info *crypt_info; - u8 full_key_descriptor[FS_KEY_DESC_PREFIX_SIZE + - (FS_KEY_DESCRIPTOR_SIZE * 2) + 1]; - struct key *keyring_key = NULL; - struct fscrypt_key *master_key; struct fscrypt_context ctx; - const struct user_key_payload *ukp; struct crypto_skcipher *ctfm; const char *cipher_str; u8 raw_key[FS_MAX_KEY_SIZE]; @@ -167,48 +223,24 @@ retry: memset(raw_key, 0x42, FS_AES_256_XTS_KEY_SIZE); goto got_key; } - memcpy(full_key_descriptor, FS_KEY_DESC_PREFIX, - FS_KEY_DESC_PREFIX_SIZE); - sprintf(full_key_descriptor + FS_KEY_DESC_PREFIX_SIZE, - "%*phN", FS_KEY_DESCRIPTOR_SIZE, - ctx.master_key_descriptor); - full_key_descriptor[FS_KEY_DESC_PREFIX_SIZE + - (2 * FS_KEY_DESCRIPTOR_SIZE)] = '\0'; - keyring_key = request_key(&key_type_logon, full_key_descriptor, NULL); - if (IS_ERR(keyring_key)) { - res = PTR_ERR(keyring_key); - keyring_key = NULL; - goto out; - } - crypt_info->ci_keyring_key = keyring_key; - if (keyring_key->type != &key_type_logon) { - printk_once(KERN_WARNING - "%s: key type must be logon\n", __func__); - res = -ENOKEY; - goto out; - } - down_read(&keyring_key->sem); - ukp = user_key_payload(keyring_key); - if (ukp->datalen != sizeof(struct fscrypt_key)) { - res = -EINVAL; - up_read(&keyring_key->sem); - goto out; - } - master_key = (struct fscrypt_key *)ukp->data; - BUILD_BUG_ON(FS_AES_128_ECB_KEY_SIZE != FS_KEY_DERIVATION_NONCE_SIZE); - if (master_key->size != FS_AES_256_XTS_KEY_SIZE) { - printk_once(KERN_WARNING - "%s: key size incorrect: %d\n", - __func__, master_key->size); - res = -ENOKEY; - up_read(&keyring_key->sem); + res = validate_user_key(crypt_info, &ctx, raw_key, + FS_KEY_DESC_PREFIX, FS_KEY_DESC_PREFIX_SIZE); + if (res && inode->i_sb->s_cop->key_prefix) { + u8 *prefix = NULL; + int prefix_size, res2; + + prefix_size = inode->i_sb->s_cop->key_prefix(inode, &prefix); + res2 = validate_user_key(crypt_info, &ctx, raw_key, + prefix, prefix_size); + if (res2) { + if (res2 == -ENOKEY) + res = -ENOKEY; + goto out; + } + } else if (res) { goto out; } - res = derive_key_aes(ctx.nonce, master_key->raw, raw_key); - up_read(&keyring_key->sem); - if (res) - goto out; got_key: ctfm = crypto_alloc_skcipher(cipher_str, 0, 0); if (!ctfm || IS_ERR(ctfm)) { diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index ccf8bf4debfc..dbd277eb9da7 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -711,6 +711,10 @@ enum { MAX_TIME, }; +#ifdef CONFIG_F2FS_FS_ENCRYPTION +#define F2FS_KEY_DESC_PREFIX "f2fs:" +#define F2FS_KEY_DESC_PREFIX_SIZE 5 +#endif struct f2fs_sb_info { struct super_block *sb; /* pointer to VFS super block */ struct proc_dir_entry *s_proc; /* proc entry */ @@ -718,6 +722,10 @@ struct f2fs_sb_info { int valid_super_block; /* valid super block no */ int s_flag; /* flags for sbi */ +#ifdef CONFIG_F2FS_FS_ENCRYPTION + u8 key_prefix[F2FS_KEY_DESC_PREFIX_SIZE]; + u8 key_prefix_size; +#endif /* for node-related operations */ struct f2fs_nm_info *nm_info; /* node manager */ struct inode *node_inode; /* cache node blocks */ diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 8a28f799a24f..28c8992da6f6 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -964,6 +964,12 @@ static int f2fs_get_context(struct inode *inode, void *ctx, size_t len) ctx, len, NULL); } +static int f2fs_key_prefix(struct inode *inode, u8 **key) +{ + *key = F2FS_I_SB(inode)->key_prefix; + return F2FS_I_SB(inode)->key_prefix_size; +} + static int f2fs_set_context(struct inode *inode, const void *ctx, size_t len, void *fs_data) { @@ -980,6 +986,7 @@ static unsigned f2fs_max_namelen(struct inode *inode) static struct fscrypt_operations f2fs_cryptops = { .get_context = f2fs_get_context, + .key_prefix = f2fs_key_prefix, .set_context = f2fs_set_context, .is_encrypted = f2fs_encrypted_inode, .empty_dir = f2fs_empty_dir, @@ -1305,6 +1312,12 @@ static void init_sb_info(struct f2fs_sb_info *sbi) INIT_LIST_HEAD(&sbi->s_list); mutex_init(&sbi->umount_mutex); + +#ifdef CONFIG_F2FS_FS_ENCRYPTION + memcpy(sbi->key_prefix, F2FS_KEY_DESC_PREFIX, + F2FS_KEY_DESC_PREFIX_SIZE); + sbi->key_prefix_size = F2FS_KEY_DESC_PREFIX_SIZE; +#endif } /* diff --git a/include/linux/fscrypto.h b/include/linux/fscrypto.h index 6027f6bbb061..cfa6cde25f8e 100644 --- a/include/linux/fscrypto.h +++ b/include/linux/fscrypto.h @@ -175,6 +175,7 @@ struct fscrypt_name { */ struct fscrypt_operations { int (*get_context)(struct inode *, void *, size_t); + int (*key_prefix)(struct inode *, u8 **); int (*prepare_context)(struct inode *); int (*set_context)(struct inode *, const void *, size_t, void *); int (*dummy_context)(struct inode *); -- cgit v1.3-8-gc7d7 From f61cce5b81f91ba336184008b24baec84afbb3dd Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Sat, 7 May 2016 16:15:05 +0800 Subject: f2fs: fix inode cache leak When testing f2fs with inline_dentry option, generic/342 reports: VFS: Busy inodes after unmount of dm-0. Self-destruct in 5 seconds. Have a nice day... After rmmod f2fs module, kenrel shows following dmesg: ============================================================================= BUG f2fs_inode_cache (Tainted: G O ): Objects remaining in f2fs_inode_cache on __kmem_cache_shutdown() ----------------------------------------------------------------------------- Disabling lock debugging due to kernel taint INFO: Slab 0xf51ca0e0 objects=22 used=1 fp=0xd1e6fc60 flags=0x40004080 CPU: 3 PID: 7455 Comm: rmmod Tainted: G B O 4.6.0-rc4+ #16 Hardware name: innotek GmbH VirtualBox/VirtualBox, BIOS VirtualBox 12/01/2006 00000086 00000086 d062fe18 c13a83a0 f51ca0e0 d062fe38 d062fea4 c11c7276 c1981040 f51ca0e0 00000016 00000001 d1e6fc60 40004080 656a624f 20737463 616d6572 6e696e69 6e692067 66326620 6e695f73 5f65646f 68636163 6e6f2065 Call Trace: [] dump_stack+0x5f/0x8f [] slab_err+0x76/0x80 [] ? __kmem_cache_shutdown+0x100/0x2f0 [] ? __kmem_cache_shutdown+0x100/0x2f0 [] __kmem_cache_shutdown+0x125/0x2f0 [] kmem_cache_destroy+0x158/0x1f0 [] ? mutex_unlock+0xd/0x10 [] exit_f2fs_fs+0x4b/0x5a8 [f2fs] [] SyS_delete_module+0x16c/0x1d0 [] ? do_fast_syscall_32+0x30/0x1c0 [] ? __this_cpu_preempt_check+0xf/0x20 [] ? trace_hardirqs_on_caller+0xdd/0x210 [] ? trace_hardirqs_off+0xb/0x10 [] do_fast_syscall_32+0xa1/0x1c0 [] sysenter_past_esp+0x45/0x74 INFO: Object 0xd1e6d9e0 @offset=6624 kmem_cache_destroy f2fs_inode_cache: Slab cache still has objects CPU: 3 PID: 7455 Comm: rmmod Tainted: G B O 4.6.0-rc4+ #16 Hardware name: innotek GmbH VirtualBox/VirtualBox, BIOS VirtualBox 12/01/2006 00000286 00000286 d062fef4 c13a83a0 f174b000 d062ff14 d062ff28 c1198ac7 c197fe18 f3c5b980 d062ff20 000d04f2 d062ff0c d062ff0c d062ff14 d062ff14 f8f20dc0 fffffff5 d062e000 d062ff30 f8f15aa3 d062ff7c c10f596c 73663266 Call Trace: [] dump_stack+0x5f/0x8f [] kmem_cache_destroy+0x1e7/0x1f0 [] exit_f2fs_fs+0x4b/0x5a8 [f2fs] [] SyS_delete_module+0x16c/0x1d0 [] ? do_fast_syscall_32+0x30/0x1c0 [] ? __this_cpu_preempt_check+0xf/0x20 [] ? trace_hardirqs_on_caller+0xdd/0x210 [] ? trace_hardirqs_off+0xb/0x10 [] do_fast_syscall_32+0xa1/0x1c0 [] sysenter_past_esp+0x45/0x74 The reason is: in recovery flow, we use delayed iput mechanism for directory which has recovered dentry block. It means the reference of inode will be held until last dirty dentry page being writebacked. But when we mount f2fs with inline_dentry option, during recovery, dirent may only be recovered into dir inode page rather than dentry page, so there are no chance for us to release inode reference in ->writepage when writebacking last dentry page. We can call paired iget/iput explicityly for inline_dentry case, but for non-inline_dentry case, iput will call writeback_single_inode to write all data pages synchronously, but during recovery, ->writepages of f2fs skips writing all pages, result in losing dirent. This patch fixes this issue by obsoleting old mechanism, and introduce a new dir_list to hold all directory inodes which has recovered datas until finishing recovery. Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/checkpoint.c | 16 --------------- fs/f2fs/f2fs.h | 2 -- fs/f2fs/recovery.c | 56 +++++++++++++++++++++++++++++----------------------- 3 files changed, 31 insertions(+), 43 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c index 432a6baa61d1..6402e9372f6c 100644 --- a/fs/f2fs/checkpoint.c +++ b/fs/f2fs/checkpoint.c @@ -794,19 +794,9 @@ void update_dirty_page(struct inode *inode, struct page *page) f2fs_trace_pid(page); } -void add_dirty_dir_inode(struct inode *inode) -{ - struct f2fs_sb_info *sbi = F2FS_I_SB(inode); - - spin_lock(&sbi->inode_lock[DIR_INODE]); - __add_dirty_inode(inode, DIR_INODE); - spin_unlock(&sbi->inode_lock[DIR_INODE]); -} - void remove_dirty_inode(struct inode *inode) { struct f2fs_sb_info *sbi = F2FS_I_SB(inode); - struct f2fs_inode_info *fi = F2FS_I(inode); enum inode_type type = S_ISDIR(inode->i_mode) ? DIR_INODE : FILE_INODE; if (!S_ISDIR(inode->i_mode) && !S_ISREG(inode->i_mode) && @@ -816,12 +806,6 @@ void remove_dirty_inode(struct inode *inode) spin_lock(&sbi->inode_lock[type]); __remove_dirty_inode(inode, type); spin_unlock(&sbi->inode_lock[type]); - - /* Only from the recovery routine */ - if (is_inode_flag_set(fi, FI_DELAY_IPUT)) { - clear_inode_flag(fi, FI_DELAY_IPUT); - iput(inode); - } } int sync_dirty_inodes(struct f2fs_sb_info *sbi, enum inode_type type) diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index dbd277eb9da7..052f5a8c96f1 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -1488,7 +1488,6 @@ enum { FI_NO_ALLOC, /* should not allocate any blocks */ FI_FREE_NID, /* free allocated nide */ FI_UPDATE_DIR, /* should update inode block for consistency */ - FI_DELAY_IPUT, /* used for the recovery */ FI_NO_EXTENT, /* not to use the extent cache */ FI_INLINE_XATTR, /* used for inline xattr */ FI_INLINE_DATA, /* used for inline data*/ @@ -1928,7 +1927,6 @@ void remove_orphan_inode(struct f2fs_sb_info *, nid_t); int recover_orphan_inodes(struct f2fs_sb_info *); int get_valid_checkpoint(struct f2fs_sb_info *); void update_dirty_page(struct inode *, struct page *); -void add_dirty_dir_inode(struct inode *); void remove_dirty_inode(struct inode *); int sync_dirty_inodes(struct f2fs_sb_info *, enum inode_type); int write_checkpoint(struct f2fs_sb_info *, struct cp_control *); diff --git a/fs/f2fs/recovery.c b/fs/f2fs/recovery.c index 29a37aae95d4..2b2532903b43 100644 --- a/fs/f2fs/recovery.c +++ b/fs/f2fs/recovery.c @@ -89,7 +89,8 @@ static void del_fsync_inode(struct fsync_inode_entry *entry) kmem_cache_free(fsync_entry_slab, entry); } -static int recover_dentry(struct inode *inode, struct page *ipage) +static int recover_dentry(struct inode *inode, struct page *ipage, + struct list_head *dir_list) { struct f2fs_inode *raw_inode = F2FS_INODE(ipage); nid_t pino = le32_to_cpu(raw_inode->i_pino); @@ -97,18 +98,29 @@ static int recover_dentry(struct inode *inode, struct page *ipage) struct qstr name; struct page *page; struct inode *dir, *einode; + struct fsync_inode_entry *entry; int err = 0; - dir = f2fs_iget(inode->i_sb, pino); - if (IS_ERR(dir)) { - err = PTR_ERR(dir); - goto out; + entry = get_fsync_inode(dir_list, pino); + if (!entry) { + dir = f2fs_iget(inode->i_sb, pino); + if (IS_ERR(dir)) { + err = PTR_ERR(dir); + goto out; + } + + entry = add_fsync_inode(dir_list, dir); + if (!entry) { + err = -ENOMEM; + iput(dir); + goto out; + } } - if (file_enc_name(inode)) { - iput(dir); + dir = entry->inode; + + if (file_enc_name(inode)) return 0; - } name.len = le32_to_cpu(raw_inode->i_namelen); name.name = raw_inode->i_name; @@ -116,7 +128,7 @@ static int recover_dentry(struct inode *inode, struct page *ipage) if (unlikely(name.len > F2FS_NAME_LEN)) { WARN_ON(1); err = -ENAMETOOLONG; - goto out_err; + goto out; } retry: de = f2fs_find_entry(dir, &name, &page); @@ -142,23 +154,12 @@ retry: goto retry; } err = __f2fs_add_link(dir, &name, inode, inode->i_ino, inode->i_mode); - if (err) - goto out_err; - - if (is_inode_flag_set(F2FS_I(dir), FI_DELAY_IPUT)) { - iput(dir); - } else { - add_dirty_dir_inode(dir); - set_inode_flag(F2FS_I(dir), FI_DELAY_IPUT); - } goto out; out_unmap_put: f2fs_dentry_kunmap(dir, page); f2fs_put_page(page, 0); -out_err: - iput(dir); out: f2fs_msg(inode->i_sb, KERN_NOTICE, "%s: ino = %x, name = %s, dir = %lx, err = %d", @@ -501,7 +502,8 @@ out: return err; } -static int recover_data(struct f2fs_sb_info *sbi, struct list_head *head) +static int recover_data(struct f2fs_sb_info *sbi, struct list_head *inode_list, + struct list_head *dir_list) { unsigned long long cp_ver = cur_cp_version(F2FS_CKPT(sbi)); struct curseg_info *curseg; @@ -528,7 +530,7 @@ static int recover_data(struct f2fs_sb_info *sbi, struct list_head *head) break; } - entry = get_fsync_inode(head, ino_of_node(page)); + entry = get_fsync_inode(inode_list, ino_of_node(page)); if (!entry) goto next; /* @@ -539,7 +541,7 @@ static int recover_data(struct f2fs_sb_info *sbi, struct list_head *head) if (IS_INODE(page)) recover_inode(entry->inode, page); if (entry->last_dentry == blkaddr) { - err = recover_dentry(entry->inode, page); + err = recover_dentry(entry->inode, page, dir_list); if (err) { f2fs_put_page(page, 1); break; @@ -567,6 +569,7 @@ int recover_fsync_data(struct f2fs_sb_info *sbi, bool check_only) { struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_WARM_NODE); struct list_head inode_list; + struct list_head dir_list; block_t blkaddr; int err; int ret = 0; @@ -578,6 +581,7 @@ int recover_fsync_data(struct f2fs_sb_info *sbi, bool check_only) return -ENOMEM; INIT_LIST_HEAD(&inode_list); + INIT_LIST_HEAD(&dir_list); /* prevent checkpoint */ mutex_lock(&sbi->cp_mutex); @@ -597,12 +601,11 @@ int recover_fsync_data(struct f2fs_sb_info *sbi, bool check_only) need_writecp = true; /* step #2: recover data */ - err = recover_data(sbi, &inode_list); + err = recover_data(sbi, &inode_list, &dir_list); if (!err) f2fs_bug_on(sbi, !list_empty(&inode_list)); out: destroy_fsync_dnodes(&inode_list); - kmem_cache_destroy(fsync_entry_slab); /* truncate meta pages to be used by the recovery */ truncate_inode_pages_range(META_MAPPING(sbi), @@ -640,5 +643,8 @@ out: } else { mutex_unlock(&sbi->cp_mutex); } + + destroy_fsync_dnodes(&dir_list); + kmem_cache_destroy(fsync_entry_slab); return ret ? ret: err; } -- cgit v1.3-8-gc7d7 From e12dd7bd874cb1c8658d5e8c1eab5f91a71939dc Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Fri, 6 May 2016 15:30:38 -0700 Subject: f2fs: fallocate data blocks in single locked node page This patch is to improve the expand_inode speed in fallocate by allocating data blocks as many as possible in single locked node page. In SSD, # time fallocate -l 500G $MNT/testfile Before : 1m 33.410 s After : 24.758 s Reported-by: Stephen Bates Signed-off-by: Jaegeuk Kim --- fs/f2fs/file.c | 51 ++++++++++++++++++++++----------------------------- 1 file changed, 22 insertions(+), 29 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index dd50f30a2f57..fb8cc27d4cc6 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -1158,10 +1158,11 @@ static int expand_inode_data(struct inode *inode, loff_t offset, loff_t len, int mode) { struct f2fs_sb_info *sbi = F2FS_I_SB(inode); - pgoff_t index, pg_start, pg_end; + struct f2fs_map_blocks map = { .m_next_pgofs = NULL }; + pgoff_t pg_end; loff_t new_size = i_size_read(inode); - loff_t off_start, off_end; - int ret = 0; + loff_t off_end; + int ret; ret = inode_newsize_ok(inode, (len + offset)); if (ret) @@ -1173,43 +1174,35 @@ static int expand_inode_data(struct inode *inode, loff_t offset, f2fs_balance_fs(sbi, true); - pg_start = ((unsigned long long) offset) >> PAGE_SHIFT; - pg_end = ((unsigned long long) offset + len) >> PAGE_SHIFT; - - off_start = offset & (PAGE_SIZE - 1); + pg_end = ((unsigned long long)offset + len) >> PAGE_SHIFT; off_end = (offset + len) & (PAGE_SIZE - 1); - f2fs_lock_op(sbi); + map.m_lblk = ((unsigned long long)offset) >> PAGE_SHIFT; + map.m_len = pg_end - map.m_lblk; + if (off_end) + map.m_len++; - for (index = pg_start; index <= pg_end; index++) { - struct dnode_of_data dn; + ret = f2fs_map_blocks(inode, &map, 1, F2FS_GET_BLOCK_PRE_AIO); + if (ret) { + pgoff_t last_off; - if (index == pg_end && !off_end) - goto noalloc; + if (!map.m_len) + return ret; - set_new_dnode(&dn, inode, NULL, NULL, 0); - ret = f2fs_reserve_block(&dn, index); - if (ret) - break; -noalloc: - if (pg_start == pg_end) - new_size = offset + len; - else if (index == pg_start && off_start) - new_size = (loff_t)(index + 1) << PAGE_SHIFT; - else if (index == pg_end) - new_size = ((loff_t)index << PAGE_SHIFT) + - off_end; - else - new_size += PAGE_SIZE; + last_off = map.m_lblk + map.m_len - 1; + + /* update new size to the failed position */ + new_size = (last_off == pg_end) ? offset + len: + (loff_t)(last_off + 1) << PAGE_SHIFT; + } else { + new_size = ((loff_t)pg_end << PAGE_SHIFT) + off_end; } - if (!(mode & FALLOC_FL_KEEP_SIZE) && - i_size_read(inode) < new_size) { + if (!(mode & FALLOC_FL_KEEP_SIZE) && i_size_read(inode) < new_size) { i_size_write(inode, new_size); mark_inode_dirty(inode); update_inode_page(inode); } - f2fs_unlock_op(sbi); return ret; } -- cgit v1.3-8-gc7d7 From 79344efb93a26378a91193bed133cee42162cd81 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Fri, 6 May 2016 16:19:43 -0700 Subject: f2fs: read node blocks ahead when truncating blocks This patch enables reading node blocks in advance when truncating large data blocks. > time rm $MNT/testfile (500GB) after drop_cachees Before : 9.422 s After : 4.821 s Reported-by: Stephen Bates Signed-off-by: Jaegeuk Kim --- fs/f2fs/file.c | 2 +- fs/f2fs/node.c | 52 ++++++++++++++++++++++++++++------------------------ 2 files changed, 29 insertions(+), 25 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index fb8cc27d4cc6..314612997721 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -584,7 +584,7 @@ int truncate_blocks(struct inode *inode, u64 from, bool lock) } set_new_dnode(&dn, inode, ipage, NULL, 0); - err = get_dnode_of_data(&dn, free_from, LOOKUP_NODE); + err = get_dnode_of_data(&dn, free_from, LOOKUP_NODE_RA); if (err) { if (err == -ENOENT) goto free_next; diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c index 264abe08bafb..cda7c7ccd4c8 100644 --- a/fs/f2fs/node.c +++ b/fs/f2fs/node.c @@ -407,6 +407,29 @@ cache: up_write(&nm_i->nat_tree_lock); } +/* + * readahead MAX_RA_NODE number of node pages. + */ +static void ra_node_pages(struct page *parent, int start, int n) +{ + struct f2fs_sb_info *sbi = F2FS_P_SB(parent); + struct blk_plug plug; + int i, end; + nid_t nid; + + blk_start_plug(&plug); + + /* Then, try readahead for siblings of the desired node */ + end = start + n; + end = min(end, NIDS_PER_BLOCK); + for (i = start; i < end; i++) { + nid = get_nid(parent, i, false); + ra_node_page(sbi, nid); + } + + blk_finish_plug(&plug); +} + pgoff_t get_next_page_offset(struct dnode_of_data *dn, pgoff_t pgofs) { const long direct_index = ADDRS_PER_INODE(dn->inode); @@ -707,6 +730,8 @@ static int truncate_nodes(struct dnode_of_data *dn, unsigned int nofs, return PTR_ERR(page); } + ra_node_pages(page, ofs, NIDS_PER_BLOCK); + rn = F2FS_NODE(page); if (depth < 3) { for (i = ofs; i < NIDS_PER_BLOCK; i++, freed++) { @@ -784,6 +809,8 @@ static int truncate_partial_nodes(struct dnode_of_data *dn, nid[i + 1] = get_nid(pages[i], offset[i + 1], false); } + ra_node_pages(pages[idx], offset[idx + 1], NIDS_PER_BLOCK); + /* free direct nodes linked to a partial indirect node */ for (i = offset[idx + 1]; i < NIDS_PER_BLOCK; i++) { child_nid = get_nid(pages[idx], i, false); @@ -1095,29 +1122,6 @@ void ra_node_page(struct f2fs_sb_info *sbi, nid_t nid) f2fs_put_page(apage, err ? 1 : 0); } -/* - * readahead MAX_RA_NODE number of node pages. - */ -static void ra_node_pages(struct page *parent, int start) -{ - struct f2fs_sb_info *sbi = F2FS_P_SB(parent); - struct blk_plug plug; - int i, end; - nid_t nid; - - blk_start_plug(&plug); - - /* Then, try readahead for siblings of the desired node */ - end = start + MAX_RA_NODE; - end = min(end, NIDS_PER_BLOCK); - for (i = start; i < end; i++) { - nid = get_nid(parent, i, false); - ra_node_page(sbi, nid); - } - - blk_finish_plug(&plug); -} - static struct page *__get_node_page(struct f2fs_sb_info *sbi, pgoff_t nid, struct page *parent, int start) { @@ -1141,7 +1145,7 @@ repeat: } if (parent) - ra_node_pages(parent, start + 1); + ra_node_pages(parent, start + 1, MAX_RA_NODE); lock_page(page); -- cgit v1.3-8-gc7d7 From 0080c5076409d211fbe28d6f07966f7d39e58bad Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Sat, 7 May 2016 08:52:57 -0700 Subject: f2fs: do not preallocate block unaligned to 4KB Previously f2fs_preallocate_blocks() tries to allocate unaligned blocks. In f2fs_write_begin(), however, prepare_write_begin() does not skip its allocation due to (len != 4KB). So, it needs locking node page twice unexpectedly. Signed-off-by: Jaegeuk Kim --- fs/f2fs/data.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index 96b035319889..369d953bd770 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -582,8 +582,8 @@ ssize_t f2fs_preallocate_blocks(struct kiocb *iocb, struct iov_iter *from) struct f2fs_map_blocks map; ssize_t ret = 0; - map.m_lblk = F2FS_BYTES_TO_BLK(iocb->ki_pos); - map.m_len = F2FS_BLK_ALIGN(iov_iter_count(from)); + map.m_lblk = F2FS_BLK_ALIGN(iocb->ki_pos); + map.m_len = F2FS_BYTES_TO_BLK(iov_iter_count(from)); map.m_next_pgofs = NULL; if (f2fs_encrypted_inode(inode)) -- cgit v1.3-8-gc7d7 From 8eb0dfdbda3f56bf7d248ed87fcc383df114ecbb Mon Sep 17 00:00:00 2001 From: Adam Borowski Date: Sun, 8 May 2016 15:08:00 +0200 Subject: btrfs: fix int32 overflow in shrink_delalloc(). UBSAN: Undefined behaviour in fs/btrfs/extent-tree.c:4623:21 signed integer overflow: 10808 * 262144 cannot be represented in type 'int [8]' If 8192<=items<16384, we request a writeback of an insane number of pages which is benign (everything will be written). But if items>=16384, the space reservation won't be enough. Signed-off-by: Adam Borowski Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/extent-tree.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 84e060eb0de8..391f576789e9 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -4620,7 +4620,7 @@ static void shrink_delalloc(struct btrfs_root *root, u64 to_reclaim, u64 orig, /* Calc the number of the pages we need flush for space reservation */ items = calc_reclaim_items_nr(root, to_reclaim); - to_reclaim = items * EXTENT_SIZE_PER_ITEM; + to_reclaim = (u64)items * EXTENT_SIZE_PER_ITEM; trans = (struct btrfs_trans_handle *)current->journal_info; block_rsv = &root->fs_info->delalloc_block_rsv; -- cgit v1.3-8-gc7d7 From 1b1bc66bb4ebd737fb750b0e1afc753098f4a1b2 Mon Sep 17 00:00:00 2001 From: Weston Andros Adamson Date: Fri, 1 Apr 2016 11:42:28 -0400 Subject: pnfs: set NFS_IOHDR_REDO in pnfs_read_resend_pnfs Like other resend paths, mark the (old) hdr as NFS_IOHDR_REDO. This ensures the hdr completion function will not count the (old) hdr as good bytes. Also, vector the error back through the hdr->task.tk_status like other retry calls. This fixes a bug with the FlexFiles layout where libaio was reporting more bytes read than requested. Signed-off-by: Weston Andros Adamson Signed-off-by: Anna Schumaker --- fs/nfs/pnfs.c | 14 ++++++++------ fs/nfs/pnfs.h | 2 +- 2 files changed, 9 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index 89a5ef4df08a..5b404d926e08 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -2143,12 +2143,15 @@ pnfs_try_to_read_data(struct nfs_pgio_header *hdr, } /* Resend all requests through pnfs. */ -int pnfs_read_resend_pnfs(struct nfs_pgio_header *hdr) +void pnfs_read_resend_pnfs(struct nfs_pgio_header *hdr) { struct nfs_pageio_descriptor pgio; - nfs_pageio_init_read(&pgio, hdr->inode, false, hdr->completion_ops); - return nfs_pageio_resend(&pgio, hdr); + if (!test_and_set_bit(NFS_IOHDR_REDO, &hdr->flags)) { + nfs_pageio_init_read(&pgio, hdr->inode, false, + hdr->completion_ops); + hdr->task.tk_status = nfs_pageio_resend(&pgio, hdr); + } } EXPORT_SYMBOL_GPL(pnfs_read_resend_pnfs); @@ -2158,12 +2161,11 @@ pnfs_do_read(struct nfs_pageio_descriptor *desc, struct nfs_pgio_header *hdr) const struct rpc_call_ops *call_ops = desc->pg_rpc_callops; struct pnfs_layout_segment *lseg = desc->pg_lseg; enum pnfs_try_status trypnfs; - int err = 0; trypnfs = pnfs_try_to_read_data(hdr, call_ops, lseg); if (trypnfs == PNFS_TRY_AGAIN) - err = pnfs_read_resend_pnfs(hdr); - if (trypnfs == PNFS_NOT_ATTEMPTED || err) + pnfs_read_resend_pnfs(hdr); + if (trypnfs == PNFS_NOT_ATTEMPTED || hdr->task.tk_status) pnfs_read_through_mds(desc, hdr); } diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h index 1ac1db5f6dad..7222d3a35439 100644 --- a/fs/nfs/pnfs.h +++ b/fs/nfs/pnfs.h @@ -282,7 +282,7 @@ int _pnfs_return_layout(struct inode *); int pnfs_commit_and_return_layout(struct inode *); void pnfs_ld_write_done(struct nfs_pgio_header *); void pnfs_ld_read_done(struct nfs_pgio_header *); -int pnfs_read_resend_pnfs(struct nfs_pgio_header *); +void pnfs_read_resend_pnfs(struct nfs_pgio_header *); struct pnfs_layout_segment *pnfs_update_layout(struct inode *ino, struct nfs_open_context *ctx, loff_t pos, -- cgit v1.3-8-gc7d7 From ed3743a6d4f309f5278c87a69e3125382b78455c Mon Sep 17 00:00:00 2001 From: Weston Andros Adamson Date: Fri, 1 Apr 2016 11:42:29 -0400 Subject: nfs: add debug to directio "good_bytes" counting This will pop a warning if we count too many good bytes. Signed-off-by: Weston Andros Adamson Signed-off-by: Anna Schumaker --- fs/nfs/direct.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c index c93826e4a8c6..f79d98ae4c10 100644 --- a/fs/nfs/direct.c +++ b/fs/nfs/direct.c @@ -87,6 +87,7 @@ struct nfs_direct_req { int mirror_count; ssize_t count, /* bytes actually processed */ + max_count, /* max expected count */ bytes_left, /* bytes left to be sent */ io_start, /* start of IO */ error; /* any reported error */ @@ -123,6 +124,8 @@ nfs_direct_good_bytes(struct nfs_direct_req *dreq, struct nfs_pgio_header *hdr) int i; ssize_t count; + WARN_ON_ONCE(dreq->count >= dreq->max_count); + if (dreq->mirror_count == 1) { dreq->mirrors[hdr->pgio_mirror_idx].count += hdr->good_bytes; dreq->count += hdr->good_bytes; @@ -593,7 +596,7 @@ ssize_t nfs_file_direct_read(struct kiocb *iocb, struct iov_iter *iter, goto out_unlock; dreq->inode = inode; - dreq->bytes_left = count; + dreq->bytes_left = dreq->max_count = count; dreq->io_start = pos; dreq->ctx = get_nfs_open_context(nfs_file_open_context(iocb->ki_filp)); l_ctx = nfs_get_lock_context(dreq->ctx); @@ -1026,7 +1029,7 @@ ssize_t nfs_file_direct_write(struct kiocb *iocb, struct iov_iter *iter) goto out_unlock; dreq->inode = inode; - dreq->bytes_left = iov_iter_count(iter); + dreq->bytes_left = dreq->max_count = iov_iter_count(iter); dreq->io_start = pos; dreq->ctx = get_nfs_open_context(nfs_file_open_context(iocb->ki_filp)); l_ctx = nfs_get_lock_context(dreq->ctx); -- cgit v1.3-8-gc7d7 From fe238e601d2519f259103ab65caea3b077ed7b39 Mon Sep 17 00:00:00 2001 From: Dave Wysochanski Date: Fri, 1 Apr 2016 13:45:09 -0400 Subject: NFS: Save struct inode * inside nfs_commit_info to clarify usage of i_lock Commit ea2cf22 created nfs_commit_info and saved &inode->i_lock inside this NFS specific structure. This obscures the usage of i_lock. Instead, save struct inode * so later it's clear the spinlock taken is i_lock. Should be no functional change. Signed-off-by: Dave Wysochanski Signed-off-by: Anna Schumaker --- fs/nfs/direct.c | 10 +++++----- fs/nfs/filelayout/filelayout.c | 4 ++-- fs/nfs/flexfilelayout/flexfilelayout.c | 4 ++-- fs/nfs/pnfs_nfs.c | 32 ++++++++++++++++---------------- fs/nfs/write.c | 16 ++++++++-------- include/linux/nfs_xdr.h | 2 +- 6 files changed, 34 insertions(+), 34 deletions(-) (limited to 'fs') diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c index f79d98ae4c10..7f03163b5364 100644 --- a/fs/nfs/direct.c +++ b/fs/nfs/direct.c @@ -278,7 +278,7 @@ static void nfs_direct_release_pages(struct page **pages, unsigned int npages) void nfs_init_cinfo_from_dreq(struct nfs_commit_info *cinfo, struct nfs_direct_req *dreq) { - cinfo->lock = &dreq->inode->i_lock; + cinfo->inode = dreq->inode; cinfo->mds = &dreq->mds_cinfo; cinfo->ds = &dreq->ds_cinfo; cinfo->dreq = dreq; @@ -635,13 +635,13 @@ nfs_direct_write_scan_commit_list(struct inode *inode, struct list_head *list, struct nfs_commit_info *cinfo) { - spin_lock(cinfo->lock); + spin_lock(&cinfo->inode->i_lock); #ifdef CONFIG_NFS_V4_1 if (cinfo->ds != NULL && cinfo->ds->nwritten != 0) NFS_SERVER(inode)->pnfs_curr_ld->recover_commit_reqs(list, cinfo); #endif nfs_scan_commit_list(&cinfo->mds->list, list, cinfo, 0); - spin_unlock(cinfo->lock); + spin_unlock(&cinfo->inode->i_lock); } static void nfs_direct_write_reschedule(struct nfs_direct_req *dreq) @@ -676,13 +676,13 @@ static void nfs_direct_write_reschedule(struct nfs_direct_req *dreq) if (!nfs_pageio_add_request(&desc, req)) { nfs_list_remove_request(req); nfs_list_add_request(req, &failed); - spin_lock(cinfo.lock); + spin_lock(&cinfo.inode->i_lock); dreq->flags = 0; if (desc.pg_error < 0) dreq->error = desc.pg_error; else dreq->error = -EIO; - spin_unlock(cinfo.lock); + spin_unlock(&cinfo.inode->i_lock); } nfs_release_request(req); } diff --git a/fs/nfs/filelayout/filelayout.c b/fs/nfs/filelayout/filelayout.c index 3384dc8e6683..3e50057eeadf 100644 --- a/fs/nfs/filelayout/filelayout.c +++ b/fs/nfs/filelayout/filelayout.c @@ -795,7 +795,7 @@ filelayout_alloc_commit_info(struct pnfs_layout_segment *lseg, buckets[i].direct_verf.committed = NFS_INVALID_STABLE_HOW; } - spin_lock(cinfo->lock); + spin_lock(&cinfo->inode->i_lock); if (cinfo->ds->nbuckets >= size) goto out; for (i = 0; i < cinfo->ds->nbuckets; i++) { @@ -811,7 +811,7 @@ filelayout_alloc_commit_info(struct pnfs_layout_segment *lseg, swap(cinfo->ds->buckets, buckets); cinfo->ds->nbuckets = size; out: - spin_unlock(cinfo->lock); + spin_unlock(&cinfo->inode->i_lock); kfree(buckets); return 0; } diff --git a/fs/nfs/flexfilelayout/flexfilelayout.c b/fs/nfs/flexfilelayout/flexfilelayout.c index 0cb1abd535e3..3b398f7b4637 100644 --- a/fs/nfs/flexfilelayout/flexfilelayout.c +++ b/fs/nfs/flexfilelayout/flexfilelayout.c @@ -745,7 +745,7 @@ ff_layout_alloc_commit_info(struct pnfs_layout_segment *lseg, else { int i; - spin_lock(cinfo->lock); + spin_lock(&cinfo->inode->i_lock); if (cinfo->ds->nbuckets != 0) kfree(buckets); else { @@ -759,7 +759,7 @@ ff_layout_alloc_commit_info(struct pnfs_layout_segment *lseg, NFS_INVALID_STABLE_HOW; } } - spin_unlock(cinfo->lock); + spin_unlock(&cinfo->inode->i_lock); return 0; } } diff --git a/fs/nfs/pnfs_nfs.c b/fs/nfs/pnfs_nfs.c index 4aaed890048f..d2a7c9f7aa94 100644 --- a/fs/nfs/pnfs_nfs.c +++ b/fs/nfs/pnfs_nfs.c @@ -61,7 +61,7 @@ EXPORT_SYMBOL_GPL(pnfs_generic_commit_release); /* The generic layer is about to remove the req from the commit list. * If this will make the bucket empty, it will need to put the lseg reference. - * Note this must be called holding the inode (/cinfo) lock + * Note this must be called holding i_lock */ void pnfs_generic_clear_request_commit(struct nfs_page *req, @@ -98,7 +98,7 @@ pnfs_generic_transfer_commit_list(struct list_head *src, struct list_head *dst, if (!nfs_lock_request(req)) continue; kref_get(&req->wb_kref); - if (cond_resched_lock(cinfo->lock)) + if (cond_resched_lock(&cinfo->inode->i_lock)) list_safe_reset_next(req, tmp, wb_list); nfs_request_remove_commit_list(req, cinfo); clear_bit(PG_COMMIT_TO_DS, &req->wb_flags); @@ -119,7 +119,7 @@ pnfs_generic_scan_ds_commit_list(struct pnfs_commit_bucket *bucket, struct list_head *dst = &bucket->committing; int ret; - lockdep_assert_held(cinfo->lock); + lockdep_assert_held(&cinfo->inode->i_lock); ret = pnfs_generic_transfer_commit_list(src, dst, cinfo, max); if (ret) { cinfo->ds->nwritten -= ret; @@ -142,7 +142,7 @@ int pnfs_generic_scan_commit_lists(struct nfs_commit_info *cinfo, { int i, rv = 0, cnt; - lockdep_assert_held(cinfo->lock); + lockdep_assert_held(&cinfo->inode->i_lock); for (i = 0; i < cinfo->ds->nbuckets && max != 0; i++) { cnt = pnfs_generic_scan_ds_commit_list(&cinfo->ds->buckets[i], cinfo, max); @@ -161,16 +161,16 @@ void pnfs_generic_recover_commit_reqs(struct list_head *dst, struct pnfs_layout_segment *freeme; int i; - lockdep_assert_held(cinfo->lock); + lockdep_assert_held(&cinfo->inode->i_lock); restart: for (i = 0, b = cinfo->ds->buckets; i < cinfo->ds->nbuckets; i++, b++) { if (pnfs_generic_transfer_commit_list(&b->written, dst, cinfo, 0)) { freeme = b->wlseg; b->wlseg = NULL; - spin_unlock(cinfo->lock); + spin_unlock(&cinfo->inode->i_lock); pnfs_put_lseg(freeme); - spin_lock(cinfo->lock); + spin_lock(&cinfo->inode->i_lock); goto restart; } } @@ -186,7 +186,7 @@ static void pnfs_generic_retry_commit(struct nfs_commit_info *cinfo, int idx) LIST_HEAD(pages); int i; - spin_lock(cinfo->lock); + spin_lock(&cinfo->inode->i_lock); for (i = idx; i < fl_cinfo->nbuckets; i++) { bucket = &fl_cinfo->buckets[i]; if (list_empty(&bucket->committing)) @@ -194,12 +194,12 @@ static void pnfs_generic_retry_commit(struct nfs_commit_info *cinfo, int idx) freeme = bucket->clseg; bucket->clseg = NULL; list_splice_init(&bucket->committing, &pages); - spin_unlock(cinfo->lock); + spin_unlock(&cinfo->inode->i_lock); nfs_retry_commit(&pages, freeme, cinfo, i); pnfs_put_lseg(freeme); - spin_lock(cinfo->lock); + spin_lock(&cinfo->inode->i_lock); } - spin_unlock(cinfo->lock); + spin_unlock(&cinfo->inode->i_lock); } static unsigned int @@ -238,11 +238,11 @@ void pnfs_fetch_commit_bucket_list(struct list_head *pages, struct pnfs_commit_bucket *bucket; bucket = &cinfo->ds->buckets[data->ds_commit_index]; - spin_lock(cinfo->lock); + spin_lock(&cinfo->inode->i_lock); list_splice_init(&bucket->committing, pages); data->lseg = bucket->clseg; bucket->clseg = NULL; - spin_unlock(cinfo->lock); + spin_unlock(&cinfo->inode->i_lock); } @@ -874,12 +874,12 @@ pnfs_layout_mark_request_commit(struct nfs_page *req, struct list_head *list; struct pnfs_commit_bucket *buckets; - spin_lock(cinfo->lock); + spin_lock(&cinfo->inode->i_lock); buckets = cinfo->ds->buckets; list = &buckets[ds_commit_idx].written; if (list_empty(list)) { if (!pnfs_is_valid_lseg(lseg)) { - spin_unlock(cinfo->lock); + spin_unlock(&cinfo->inode->i_lock); cinfo->completion_ops->resched_write(cinfo, req); return; } @@ -896,7 +896,7 @@ pnfs_layout_mark_request_commit(struct nfs_page *req, cinfo->ds->nwritten++; nfs_request_add_commit_list_locked(req, list, cinfo); - spin_unlock(cinfo->lock); + spin_unlock(&cinfo->inode->i_lock); nfs_mark_page_unstable(req->wb_page, cinfo); } EXPORT_SYMBOL_GPL(pnfs_layout_mark_request_commit); diff --git a/fs/nfs/write.c b/fs/nfs/write.c index 5f4fd53e5764..9283a96b9fb8 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -804,7 +804,7 @@ nfs_page_search_commits_for_head_request_locked(struct nfs_inode *nfsi, * number of outstanding requests requiring a commit as well as * the MM page stats. * - * The caller must hold the cinfo->lock, and the nfs_page lock. + * The caller must hold cinfo->inode->i_lock, and the nfs_page lock. */ void nfs_request_add_commit_list_locked(struct nfs_page *req, struct list_head *dst, @@ -832,9 +832,9 @@ EXPORT_SYMBOL_GPL(nfs_request_add_commit_list_locked); void nfs_request_add_commit_list(struct nfs_page *req, struct nfs_commit_info *cinfo) { - spin_lock(cinfo->lock); + spin_lock(&cinfo->inode->i_lock); nfs_request_add_commit_list_locked(req, &cinfo->mds->list, cinfo); - spin_unlock(cinfo->lock); + spin_unlock(&cinfo->inode->i_lock); nfs_mark_page_unstable(req->wb_page, cinfo); } EXPORT_SYMBOL_GPL(nfs_request_add_commit_list); @@ -864,7 +864,7 @@ EXPORT_SYMBOL_GPL(nfs_request_remove_commit_list); static void nfs_init_cinfo_from_inode(struct nfs_commit_info *cinfo, struct inode *inode) { - cinfo->lock = &inode->i_lock; + cinfo->inode = inode; cinfo->mds = &NFS_I(inode)->commit_info; cinfo->ds = pnfs_get_ds_info(inode); cinfo->dreq = NULL; @@ -967,7 +967,7 @@ nfs_reqs_to_commit(struct nfs_commit_info *cinfo) return cinfo->mds->ncommit; } -/* cinfo->lock held by caller */ +/* cinfo->inode->i_lock held by caller */ int nfs_scan_commit_list(struct list_head *src, struct list_head *dst, struct nfs_commit_info *cinfo, int max) @@ -979,7 +979,7 @@ nfs_scan_commit_list(struct list_head *src, struct list_head *dst, if (!nfs_lock_request(req)) continue; kref_get(&req->wb_kref); - if (cond_resched_lock(cinfo->lock)) + if (cond_resched_lock(&cinfo->inode->i_lock)) list_safe_reset_next(req, tmp, wb_list); nfs_request_remove_commit_list(req, cinfo); nfs_list_add_request(req, dst); @@ -1005,7 +1005,7 @@ nfs_scan_commit(struct inode *inode, struct list_head *dst, { int ret = 0; - spin_lock(cinfo->lock); + spin_lock(&cinfo->inode->i_lock); if (cinfo->mds->ncommit > 0) { const int max = INT_MAX; @@ -1013,7 +1013,7 @@ nfs_scan_commit(struct inode *inode, struct list_head *dst, cinfo, max); ret += pnfs_scan_commit_lists(inode, cinfo, max - ret); } - spin_unlock(cinfo->lock); + spin_unlock(&cinfo->inode->i_lock); return ret; } diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h index d320906cf13e..cb9982d8f38f 100644 --- a/include/linux/nfs_xdr.h +++ b/include/linux/nfs_xdr.h @@ -1431,7 +1431,7 @@ struct nfs_commit_completion_ops { }; struct nfs_commit_info { - spinlock_t *lock; /* inode->i_lock */ + struct inode *inode; /* Needed for inode->i_lock */ struct nfs_mds_commit_info *mds; struct pnfs_ds_commit_info *ds; struct nfs_direct_req *dreq; /* O_DIRECT request */ -- cgit v1.3-8-gc7d7 From 547a637630c61b9e1dae9abce2b44ce7076244af Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Thu, 21 Apr 2016 20:51:57 -0400 Subject: nfs: don't call nfs4_ff_layout_prepare_ds from ff_layout_get_ds_cred All the callers already call that function before calling into here, so it ends up being a no-op anyway. Signed-off-by: Jeff Layton Signed-off-by: Anna Schumaker --- fs/nfs/flexfilelayout/flexfilelayoutdev.c | 6 +----- 1 file changed, 1 insertion(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/nfs/flexfilelayout/flexfilelayoutdev.c b/fs/nfs/flexfilelayout/flexfilelayoutdev.c index add0e5a70bd6..a0dbf94d15ae 100644 --- a/fs/nfs/flexfilelayout/flexfilelayoutdev.c +++ b/fs/nfs/flexfilelayout/flexfilelayoutdev.c @@ -433,16 +433,12 @@ ff_layout_get_ds_cred(struct pnfs_layout_segment *lseg, u32 ds_idx, struct rpc_cred *mdscred) { struct nfs4_ff_layout_mirror *mirror = FF_LAYOUT_COMP(lseg, ds_idx); - struct rpc_cred *cred = ERR_PTR(-EINVAL); - - if (!nfs4_ff_layout_prepare_ds(lseg, ds_idx, true)) - goto out; + struct rpc_cred *cred; if (mirror && mirror->cred) cred = mirror->cred; else cred = mdscred; -out: return cred; } -- cgit v1.3-8-gc7d7 From 57f3f4c0cd50e90aa92eec20d9c309dd67c594a5 Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Thu, 21 Apr 2016 20:51:58 -0400 Subject: nfs: have ff_layout_get_ds_cred take a reference to the cred In later patches, we're going to want to allow the creds to be updated when we get a new layout with updated creds. Have this function take a reference to the cred that is later put once the call has been dispatched. Also, prepare for this change by ensuring we follow RCU rules when getting a reference to the cred as well. Signed-off-by: Jeff Layton Signed-off-by: Anna Schumaker --- fs/nfs/flexfilelayout/flexfilelayout.c | 15 +++++++++------ fs/nfs/flexfilelayout/flexfilelayoutdev.c | 30 ++++++++++++++++++++++++++---- 2 files changed, 35 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/nfs/flexfilelayout/flexfilelayout.c b/fs/nfs/flexfilelayout/flexfilelayout.c index 3b398f7b4637..1f46043c181e 100644 --- a/fs/nfs/flexfilelayout/flexfilelayout.c +++ b/fs/nfs/flexfilelayout/flexfilelayout.c @@ -1712,7 +1712,7 @@ ff_layout_read_pagelist(struct nfs_pgio_header *hdr) goto out_failed; ds_cred = ff_layout_get_ds_cred(lseg, idx, hdr->cred); - if (IS_ERR(ds_cred)) + if (!ds_cred) goto out_failed; vers = nfs4_ff_layout_ds_version(lseg, idx); @@ -1737,7 +1737,7 @@ ff_layout_read_pagelist(struct nfs_pgio_header *hdr) vers == 3 ? &ff_layout_read_call_ops_v3 : &ff_layout_read_call_ops_v4, 0, RPC_TASK_SOFTCONN); - + put_rpccred(ds_cred); return PNFS_ATTEMPTED; out_failed: @@ -1769,7 +1769,7 @@ ff_layout_write_pagelist(struct nfs_pgio_header *hdr, int sync) return PNFS_NOT_ATTEMPTED; ds_cred = ff_layout_get_ds_cred(lseg, idx, hdr->cred); - if (IS_ERR(ds_cred)) + if (!ds_cred) return PNFS_NOT_ATTEMPTED; vers = nfs4_ff_layout_ds_version(lseg, idx); @@ -1798,6 +1798,7 @@ ff_layout_write_pagelist(struct nfs_pgio_header *hdr, int sync) vers == 3 ? &ff_layout_write_call_ops_v3 : &ff_layout_write_call_ops_v4, sync, RPC_TASK_SOFTCONN); + put_rpccred(ds_cred); return PNFS_ATTEMPTED; } @@ -1824,7 +1825,7 @@ static int ff_layout_initiate_commit(struct nfs_commit_data *data, int how) struct rpc_clnt *ds_clnt; struct rpc_cred *ds_cred; u32 idx; - int vers; + int vers, ret; struct nfs_fh *fh; idx = calc_ds_index_from_commit(lseg, data->ds_commit_index); @@ -1838,7 +1839,7 @@ static int ff_layout_initiate_commit(struct nfs_commit_data *data, int how) goto out_err; ds_cred = ff_layout_get_ds_cred(lseg, idx, data->cred); - if (IS_ERR(ds_cred)) + if (!ds_cred) goto out_err; vers = nfs4_ff_layout_ds_version(lseg, idx); @@ -1854,10 +1855,12 @@ static int ff_layout_initiate_commit(struct nfs_commit_data *data, int how) if (fh) data->args.fh = fh; - return nfs_initiate_commit(ds_clnt, data, ds->ds_clp->rpc_ops, + ret = nfs_initiate_commit(ds_clnt, data, ds->ds_clp->rpc_ops, vers == 3 ? &ff_layout_commit_call_ops_v3 : &ff_layout_commit_call_ops_v4, how, RPC_TASK_SOFTCONN); + put_rpccred(ds_cred); + return ret; out_err: pnfs_generic_prepare_to_resend_writes(data); pnfs_generic_commit_release(data); diff --git a/fs/nfs/flexfilelayout/flexfilelayoutdev.c b/fs/nfs/flexfilelayout/flexfilelayoutdev.c index a0dbf94d15ae..baee22929174 100644 --- a/fs/nfs/flexfilelayout/flexfilelayoutdev.c +++ b/fs/nfs/flexfilelayout/flexfilelayoutdev.c @@ -338,6 +338,25 @@ static int ff_layout_update_mirror_cred(struct nfs4_ff_layout_mirror *mirror, return 0; } +static struct rpc_cred * +ff_layout_get_mirror_cred(struct nfs4_ff_layout_mirror *mirror, u32 iomode) +{ + struct rpc_cred *cred, **pcred; + + pcred = &mirror->cred; + + rcu_read_lock(); + do { + cred = rcu_dereference(*pcred); + if (!cred) + break; + + cred = get_rpccred_rcu(cred); + } while(!cred); + rcu_read_unlock(); + return cred; +} + struct nfs_fh * nfs4_ff_layout_select_ds_fh(struct pnfs_layout_segment *lseg, u32 mirror_idx) { @@ -435,10 +454,13 @@ ff_layout_get_ds_cred(struct pnfs_layout_segment *lseg, u32 ds_idx, struct nfs4_ff_layout_mirror *mirror = FF_LAYOUT_COMP(lseg, ds_idx); struct rpc_cred *cred; - if (mirror && mirror->cred) - cred = mirror->cred; - else - cred = mdscred; + if (mirror) { + cred = ff_layout_get_mirror_cred(mirror, lseg->pls_range.iomode); + if (!cred) + cred = get_rpccred(mdscred); + } else { + cred = get_rpccred(mdscred); + } return cred; } -- cgit v1.3-8-gc7d7 From 90a0be00e9d29d9baac27c0c63a5e71ca1334001 Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Thu, 21 Apr 2016 20:51:59 -0400 Subject: nfs: get a reference to the credential in ff_layout_alloc_lseg We're just as likely to have allocation problems here as we would if we delay looking up the credential like we currently do. Fix the code to get a rpc_cred reference early, as soon as the mirror is set up. This allows us to eliminate the mirror early if there is a problem getting an rpc credential. This also allows us to drop the uid/gid from the layout_mirror struct as well. In the event that we find an existing mirror where this one would go, we swap in the new creds unconditionally, and drop the reference to the old one. Note that the old ff_layout_update_mirror_cred function wouldn't set this pointer unless the DS version was 3, but we don't know what the DS version is at this point. I'm a little unclear on why it did that as you still need creds to talk to v4 servers as well. I have the code set it regardless of the DS version here. Also note the change to using generic creds instead of calling lookup_cred directly. With that change, we also need to populate the group_info pointer in the acred as some functions expect that to never be NULL. Instead of allocating one every time however, we can allocate one when the module is loaded and share it since the group_info is refcounted. Signed-off-by: Jeff Layton Signed-off-by: Anna Schumaker --- fs/nfs/flexfilelayout/flexfilelayout.c | 41 ++++++++++++++++++++++----- fs/nfs/flexfilelayout/flexfilelayout.h | 2 -- fs/nfs/flexfilelayout/flexfilelayoutdev.c | 47 ++----------------------------- 3 files changed, 36 insertions(+), 54 deletions(-) (limited to 'fs') diff --git a/fs/nfs/flexfilelayout/flexfilelayout.c b/fs/nfs/flexfilelayout/flexfilelayout.c index 1f46043c181e..ee87c0f9fd55 100644 --- a/fs/nfs/flexfilelayout/flexfilelayout.c +++ b/fs/nfs/flexfilelayout/flexfilelayout.c @@ -26,6 +26,8 @@ #define FF_LAYOUT_POLL_RETRY_MAX (15*HZ) +static struct group_info *ff_zero_group; + static struct pnfs_layout_hdr * ff_layout_alloc_layout_hdr(struct inode *inode, gfp_t gfp_flags) { @@ -407,8 +409,9 @@ ff_layout_alloc_lseg(struct pnfs_layout_hdr *lh, struct nfs4_ff_layout_mirror *mirror; struct nfs4_deviceid devid; struct nfs4_deviceid_node *idnode; - u32 ds_count; - u32 fh_count; + struct auth_cred acred = { .group_info = ff_zero_group }; + struct rpc_cred *cred; + u32 ds_count, fh_count, id; int j; rc = -EIO; @@ -484,24 +487,39 @@ ff_layout_alloc_lseg(struct pnfs_layout_hdr *lh, fls->mirror_array[i]->fh_versions_cnt = fh_count; /* user */ - rc = decode_name(&stream, &fls->mirror_array[i]->uid); + rc = decode_name(&stream, &id); if (rc) goto out_err_free; + acred.uid = make_kuid(&init_user_ns, id); + /* group */ - rc = decode_name(&stream, &fls->mirror_array[i]->gid); + rc = decode_name(&stream, &id); if (rc) goto out_err_free; + acred.gid = make_kgid(&init_user_ns, id); + + /* find the cred for it */ + cred = rpc_lookup_generic_cred(&acred, 0, gfp_flags); + if (IS_ERR(cred)) { + rc = PTR_ERR(cred); + goto out_err_free; + } + + rcu_assign_pointer(fls->mirror_array[i]->cred, cred); + mirror = ff_layout_add_mirror(lh, fls->mirror_array[i]); if (mirror != fls->mirror_array[i]) { + /* swap cred ptrs so free_mirror will clean up old */ + fls->mirror_array[i]->cred = xchg(&mirror->cred, cred); ff_layout_free_mirror(fls->mirror_array[i]); fls->mirror_array[i] = mirror; } - dprintk("%s: uid %d gid %d\n", __func__, - fls->mirror_array[i]->uid, - fls->mirror_array[i]->gid); + dprintk("%s: uid %u gid %u\n", __func__, + from_kuid(&init_user_ns, acred.uid), + from_kgid(&init_user_ns, acred.gid)); } p = xdr_inline_decode(&stream, 4); @@ -2226,6 +2244,11 @@ static int __init nfs4flexfilelayout_init(void) { printk(KERN_INFO "%s: NFSv4 Flexfile Layout Driver Registering...\n", __func__); + if (!ff_zero_group) { + ff_zero_group = groups_alloc(0); + if (!ff_zero_group) + return -ENOMEM; + } return pnfs_register_layoutdriver(&flexfilelayout_type); } @@ -2234,6 +2257,10 @@ static void __exit nfs4flexfilelayout_exit(void) printk(KERN_INFO "%s: NFSv4 Flexfile Layout Driver Unregistering...\n", __func__); pnfs_unregister_layoutdriver(&flexfilelayout_type); + if (ff_zero_group) { + put_group_info(ff_zero_group); + ff_zero_group = NULL; + } } MODULE_ALIAS("nfs-layouttype4-4"); diff --git a/fs/nfs/flexfilelayout/flexfilelayout.h b/fs/nfs/flexfilelayout/flexfilelayout.h index dd353bb7dc0a..c29fc853ce74 100644 --- a/fs/nfs/flexfilelayout/flexfilelayout.h +++ b/fs/nfs/flexfilelayout/flexfilelayout.h @@ -76,8 +76,6 @@ struct nfs4_ff_layout_mirror { u32 fh_versions_cnt; struct nfs_fh *fh_versions; nfs4_stateid stateid; - u32 uid; - u32 gid; struct rpc_cred *cred; atomic_t ref; spinlock_t lock; diff --git a/fs/nfs/flexfilelayout/flexfilelayoutdev.c b/fs/nfs/flexfilelayout/flexfilelayoutdev.c index baee22929174..6ddd8a5c5ae0 100644 --- a/fs/nfs/flexfilelayout/flexfilelayoutdev.c +++ b/fs/nfs/flexfilelayout/flexfilelayoutdev.c @@ -302,42 +302,6 @@ int ff_layout_track_ds_error(struct nfs4_flexfile_layout *flo, return 0; } -/* currently we only support AUTH_NONE and AUTH_SYS */ -static rpc_authflavor_t -nfs4_ff_layout_choose_authflavor(struct nfs4_ff_layout_mirror *mirror) -{ - if (mirror->uid == (u32)-1) - return RPC_AUTH_NULL; - return RPC_AUTH_UNIX; -} - -/* fetch cred for NFSv3 DS */ -static int ff_layout_update_mirror_cred(struct nfs4_ff_layout_mirror *mirror, - struct nfs4_pnfs_ds *ds) -{ - if (ds->ds_clp && !mirror->cred && - mirror->mirror_ds->ds_versions[0].version == 3) { - struct rpc_auth *auth = ds->ds_clp->cl_rpcclient->cl_auth; - struct rpc_cred *cred; - struct auth_cred acred = { - .uid = make_kuid(&init_user_ns, mirror->uid), - .gid = make_kgid(&init_user_ns, mirror->gid), - }; - - /* AUTH_NULL ignores acred */ - cred = auth->au_ops->lookup_cred(auth, &acred, 0); - if (IS_ERR(cred)) { - dprintk("%s: lookup_cred failed with %ld\n", - __func__, PTR_ERR(cred)); - return PTR_ERR(cred); - } else { - if (cmpxchg(&mirror->cred, NULL, cred)) - put_rpccred(cred); - } - } - return 0; -} - static struct rpc_cred * ff_layout_get_mirror_cred(struct nfs4_ff_layout_mirror *mirror, u32 iomode) { @@ -386,7 +350,6 @@ nfs4_ff_layout_prepare_ds(struct pnfs_layout_segment *lseg, u32 ds_idx, struct inode *ino = lseg->pls_layout->plh_inode; struct nfs_server *s = NFS_SERVER(ino); unsigned int max_payload; - rpc_authflavor_t flavor; if (!ff_layout_mirror_valid(lseg, mirror)) { pr_err_ratelimited("NFS: %s: No data server for offset index %d\n", @@ -402,9 +365,7 @@ nfs4_ff_layout_prepare_ds(struct pnfs_layout_segment *lseg, u32 ds_idx, /* matching smp_wmb() in _nfs4_pnfs_v3/4_ds_connect */ smp_rmb(); if (ds->ds_clp) - goto out_update_creds; - - flavor = nfs4_ff_layout_choose_authflavor(mirror); + goto out; /* FIXME: For now we assume the server sent only one version of NFS * to use for the DS. @@ -413,7 +374,7 @@ nfs4_ff_layout_prepare_ds(struct pnfs_layout_segment *lseg, u32 ds_idx, dataserver_retrans, mirror->mirror_ds->ds_versions[0].version, mirror->mirror_ds->ds_versions[0].minor_version, - flavor); + RPC_AUTH_UNIX); /* connect success, check rsize/wsize limit */ if (ds->ds_clp) { @@ -438,11 +399,7 @@ nfs4_ff_layout_prepare_ds(struct pnfs_layout_segment *lseg, u32 ds_idx, } else pnfs_error_mark_layout_for_return(ino, lseg); ds = NULL; - goto out; } -out_update_creds: - if (ff_layout_update_mirror_cred(mirror, ds)) - ds = NULL; out: return ds; } -- cgit v1.3-8-gc7d7 From 3064b6861d00b7124558cdf79f9387f948361be3 Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Thu, 21 Apr 2016 20:52:00 -0400 Subject: nfs: have flexfiles mirror keep creds for both ro and rw layouts A mirror can be shared between multiple layouts, even with different iomodes. That makes stats gathering simpler, but it causes a problem when we get different creds in READ vs. RW layouts. The current code drops the newer credentials onto the floor when this occurs. That's problematic when you fetch a READ layout first, and then a RW. If the READ layout doesn't have the correct creds to do a write, then writes will fail. We could just overwrite the READ credentials with the RW ones, but that would break the ability for the server to fence the layout for reads if things go awry. We need to be able to revert to the earlier READ creds if the RW layout is returned afterward. The simplest fix is to just keep two sets of creds per mirror. One for READ layouts and one for RW, and then use the appropriate set depending on the iomode of the layout segment. Also fix up some RCU nits that sparse found. Signed-off-by: Jeff Layton Signed-off-by: Anna Schumaker --- fs/nfs/flexfilelayout/flexfilelayout.c | 30 +++++++++++++++++++++++------- fs/nfs/flexfilelayout/flexfilelayout.h | 3 ++- fs/nfs/flexfilelayout/flexfilelayoutdev.c | 7 +++++-- 3 files changed, 30 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/nfs/flexfilelayout/flexfilelayout.c b/fs/nfs/flexfilelayout/flexfilelayout.c index ee87c0f9fd55..60d690dbc947 100644 --- a/fs/nfs/flexfilelayout/flexfilelayout.c +++ b/fs/nfs/flexfilelayout/flexfilelayout.c @@ -213,10 +213,16 @@ static struct nfs4_ff_layout_mirror *ff_layout_alloc_mirror(gfp_t gfp_flags) static void ff_layout_free_mirror(struct nfs4_ff_layout_mirror *mirror) { + struct rpc_cred *cred; + ff_layout_remove_mirror(mirror); kfree(mirror->fh_versions); - if (mirror->cred) - put_rpccred(mirror->cred); + cred = rcu_access_pointer(mirror->ro_cred); + if (cred) + put_rpccred(cred); + cred = rcu_access_pointer(mirror->rw_cred); + if (cred) + put_rpccred(cred); nfs4_ff_layout_put_deviceid(mirror->mirror_ds); kfree(mirror); } @@ -410,7 +416,7 @@ ff_layout_alloc_lseg(struct pnfs_layout_hdr *lh, struct nfs4_deviceid devid; struct nfs4_deviceid_node *idnode; struct auth_cred acred = { .group_info = ff_zero_group }; - struct rpc_cred *cred; + struct rpc_cred __rcu *cred; u32 ds_count, fh_count, id; int j; @@ -501,23 +507,33 @@ ff_layout_alloc_lseg(struct pnfs_layout_hdr *lh, acred.gid = make_kgid(&init_user_ns, id); /* find the cred for it */ - cred = rpc_lookup_generic_cred(&acred, 0, gfp_flags); + rcu_assign_pointer(cred, rpc_lookup_generic_cred(&acred, 0, gfp_flags)); if (IS_ERR(cred)) { rc = PTR_ERR(cred); goto out_err_free; } - rcu_assign_pointer(fls->mirror_array[i]->cred, cred); + if (lgr->range.iomode == IOMODE_READ) + rcu_assign_pointer(fls->mirror_array[i]->ro_cred, cred); + else + rcu_assign_pointer(fls->mirror_array[i]->rw_cred, cred); mirror = ff_layout_add_mirror(lh, fls->mirror_array[i]); if (mirror != fls->mirror_array[i]) { /* swap cred ptrs so free_mirror will clean up old */ - fls->mirror_array[i]->cred = xchg(&mirror->cred, cred); + if (lgr->range.iomode == IOMODE_READ) { + cred = xchg(&mirror->ro_cred, cred); + rcu_assign_pointer(fls->mirror_array[i]->ro_cred, cred); + } else { + cred = xchg(&mirror->rw_cred, cred); + rcu_assign_pointer(fls->mirror_array[i]->rw_cred, cred); + } ff_layout_free_mirror(fls->mirror_array[i]); fls->mirror_array[i] = mirror; } - dprintk("%s: uid %u gid %u\n", __func__, + dprintk("%s: iomode %s uid %u gid %u\n", __func__, + lgr->range.iomode == IOMODE_READ ? "READ" : "RW", from_kuid(&init_user_ns, acred.uid), from_kgid(&init_user_ns, acred.gid)); } diff --git a/fs/nfs/flexfilelayout/flexfilelayout.h b/fs/nfs/flexfilelayout/flexfilelayout.h index c29fc853ce74..1318c77aeb35 100644 --- a/fs/nfs/flexfilelayout/flexfilelayout.h +++ b/fs/nfs/flexfilelayout/flexfilelayout.h @@ -76,7 +76,8 @@ struct nfs4_ff_layout_mirror { u32 fh_versions_cnt; struct nfs_fh *fh_versions; nfs4_stateid stateid; - struct rpc_cred *cred; + struct rpc_cred __rcu *ro_cred; + struct rpc_cred __rcu *rw_cred; atomic_t ref; spinlock_t lock; struct nfs4_ff_layoutstat read_stat; diff --git a/fs/nfs/flexfilelayout/flexfilelayoutdev.c b/fs/nfs/flexfilelayout/flexfilelayoutdev.c index 6ddd8a5c5ae0..56296f3df19c 100644 --- a/fs/nfs/flexfilelayout/flexfilelayoutdev.c +++ b/fs/nfs/flexfilelayout/flexfilelayoutdev.c @@ -305,9 +305,12 @@ int ff_layout_track_ds_error(struct nfs4_flexfile_layout *flo, static struct rpc_cred * ff_layout_get_mirror_cred(struct nfs4_ff_layout_mirror *mirror, u32 iomode) { - struct rpc_cred *cred, **pcred; + struct rpc_cred *cred, __rcu **pcred; - pcred = &mirror->cred; + if (iomode == IOMODE_READ) + pcred = &mirror->ro_cred; + else + pcred = &mirror->rw_cred; rcu_read_lock(); do { -- cgit v1.3-8-gc7d7 From 11476e9dec39d90fe1e9bf12abc6f3efe35a073d Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Mon, 11 Apr 2016 16:20:22 -0400 Subject: NFS: Fix an LOCK/OPEN race when unlinking an open file At Connectathon 2016, we found that recent upstream Linux clients would occasionally send a LOCK operation with a zero stateid. This appeared to happen in close proximity to another thread returning a delegation before unlinking the same file while it remained open. Earlier, the client received a write delegation on this file and returned the open stateid. Now, as it is getting ready to unlink the file, it returns the write delegation. But there is still an open file descriptor on that file, so the client must OPEN the file again before it returns the delegation. Since commit 24311f884189 ('NFSv4: Recovery of recalled read delegations is broken'), nfs_open_delegation_recall() clears the NFS_DELEGATED_STATE flag _before_ it sends the OPEN. This allows a racing LOCK on the same inode to be put on the wire before the OPEN operation has returned a valid open stateid. To eliminate this race, serialize delegation return with the acquisition of a file lock on the same file. Adopt the same approach as is used in the unlock path. This patch also eliminates a similar race seen when sending a LOCK operation at the same time as returning a delegation on the same file. Fixes: 24311f884189 ('NFSv4: Recovery of recalled read ... ') Signed-off-by: Chuck Lever [Anna: Add sentence about LOCK / delegation race] Signed-off-by: Anna Schumaker --- fs/nfs/nfs4proc.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'fs') diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 327b8c34d360..bc2676c95e1b 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -6054,6 +6054,7 @@ static int nfs41_lock_expired(struct nfs4_state *state, struct file_lock *reques static int _nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock *request) { struct nfs_inode *nfsi = NFS_I(state->inode); + struct nfs4_state_owner *sp = state->owner; unsigned char fl_flags = request->fl_flags; int status = -ENOLCK; @@ -6068,6 +6069,7 @@ static int _nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock status = do_vfs_lock(state->inode, request); if (status < 0) goto out; + mutex_lock(&sp->so_delegreturn_mutex); down_read(&nfsi->rwsem); if (test_bit(NFS_DELEGATED_STATE, &state->flags)) { /* Yes: cache locks! */ @@ -6075,9 +6077,11 @@ static int _nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock request->fl_flags = fl_flags & ~FL_SLEEP; status = do_vfs_lock(state->inode, request); up_read(&nfsi->rwsem); + mutex_unlock(&sp->so_delegreturn_mutex); goto out; } up_read(&nfsi->rwsem); + mutex_unlock(&sp->so_delegreturn_mutex); status = _nfs4_do_setlk(state, cmd, request, NFS_LOCK_NEW); out: request->fl_flags = fl_flags; -- cgit v1.3-8-gc7d7 From 7e3fcf61abde92a895533b1c6932ad17e073a49c Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Tue, 3 May 2016 15:16:02 -0400 Subject: nfs: don't share mounts between network namespaces There's no guarantee that an IP address in a different network namespace actually represents the same endpoint. Also, if we allow unprivileged nfs mounts some day then this might allow an unprivileged user in another network namespace to misdirect somebody else's nfs mounts. If sharing between containers is really what's wanted then that could still be arranged explicitly, for example with bind mounts. Reported-by: "Eric W. Biederman" Signed-off-by: J. Bruce Fields Signed-off-by: Anna Schumaker --- fs/nfs/super.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'fs') diff --git a/fs/nfs/super.c b/fs/nfs/super.c index f1268280244e..ff537a205188 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -2408,6 +2408,11 @@ static int nfs_compare_super_address(struct nfs_server *server1, struct nfs_server *server2) { struct sockaddr *sap1, *sap2; + struct rpc_xprt *xprt1 = server1->client->cl_xprt; + struct rpc_xprt *xprt2 = server2->client->cl_xprt; + + if (!net_eq(xprt1->xprt_net, xprt2->xprt_net)) + return 0; sap1 = (struct sockaddr *)&server1->nfs_client->cl_addr; sap2 = (struct sockaddr *)&server2->nfs_client->cl_addr; -- cgit v1.3-8-gc7d7 From 05135f597adcb94dc34fa87b82e68eb55e00f0eb Mon Sep 17 00:00:00 2001 From: David Sterba Date: Mon, 9 May 2016 11:32:39 +0200 Subject: btrfs: switch to common message helpers in open_ctree, adjust messages Currently we lack the identification of the filesystem in most if not all mount messages, done via printk/pr_* functions. We can use the btrfs_* helpers in open_ctree, as the fs_info <-> sb link is established at the beginning of the function. The messages have been updated at the same time to be more consistent: * dropped sb->s_id, as it's not available via btrfs_* * added %d for return code where appropriate * wording changed * %Lx replaced by %llx Signed-off-by: David Sterba --- fs/btrfs/disk-io.c | 102 ++++++++++++++++++++++++++--------------------------- 1 file changed, 50 insertions(+), 52 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index aeb090583b78..123fab55b5a4 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -2713,7 +2713,7 @@ int open_ctree(struct super_block *sb, * Pass the whole disk block of size BTRFS_SUPER_INFO_SIZE (4k). */ if (btrfs_check_super_csum(bh->b_data)) { - printk(KERN_ERR "BTRFS: superblock checksum mismatch\n"); + btrfs_err(fs_info, "superblock checksum mismatch"); err = -EINVAL; brelse(bh); goto fail_alloc; @@ -2733,7 +2733,7 @@ int open_ctree(struct super_block *sb, ret = btrfs_check_super_valid(fs_info, sb->s_flags & MS_RDONLY); if (ret) { - printk(KERN_ERR "BTRFS: superblock contains fatal errors\n"); + btrfs_err(fs_info, "superblock contains fatal errors"); err = -EINVAL; goto fail_alloc; } @@ -2768,9 +2768,9 @@ int open_ctree(struct super_block *sb, features = btrfs_super_incompat_flags(disk_super) & ~BTRFS_FEATURE_INCOMPAT_SUPP; if (features) { - printk(KERN_ERR "BTRFS: couldn't mount because of " - "unsupported optional features (%Lx).\n", - features); + btrfs_err(fs_info, + "cannot mount because of unsupported optional features (%llx)", + features); err = -EINVAL; goto fail_alloc; } @@ -2781,7 +2781,7 @@ int open_ctree(struct super_block *sb, features |= BTRFS_FEATURE_INCOMPAT_COMPRESS_LZO; if (features & BTRFS_FEATURE_INCOMPAT_SKINNY_METADATA) - printk(KERN_INFO "BTRFS: has skinny extents\n"); + btrfs_info(fs_info, "has skinny extents"); /* * flag our filesystem as having big metadata blocks if @@ -2789,7 +2789,8 @@ int open_ctree(struct super_block *sb, */ if (btrfs_super_nodesize(disk_super) > PAGE_SIZE) { if (!(features & BTRFS_FEATURE_INCOMPAT_BIG_METADATA)) - printk(KERN_INFO "BTRFS: flagging fs with big metadata feature\n"); + btrfs_info(fs_info, + "flagging fs with big metadata feature"); features |= BTRFS_FEATURE_INCOMPAT_BIG_METADATA; } @@ -2805,9 +2806,9 @@ int open_ctree(struct super_block *sb, */ if ((features & BTRFS_FEATURE_INCOMPAT_MIXED_GROUPS) && (sectorsize != nodesize)) { - printk(KERN_ERR "BTRFS: unequal leaf/node/sector sizes " - "are not allowed for mixed block groups on %s\n", - sb->s_id); + btrfs_err(fs_info, +"unequal nodesize/sectorsize (%u != %u) are not allowed for mixed block groups", + nodesize, sectorsize); goto fail_alloc; } @@ -2820,8 +2821,8 @@ int open_ctree(struct super_block *sb, features = btrfs_super_compat_ro_flags(disk_super) & ~BTRFS_FEATURE_COMPAT_RO_SUPP; if (!(sb->s_flags & MS_RDONLY) && features) { - printk(KERN_ERR "BTRFS: couldn't mount RDWR because of " - "unsupported option features (%Lx).\n", + btrfs_err(fs_info, + "cannot mount read-write because of unsupported optional features (%llx)", features); err = -EINVAL; goto fail_alloc; @@ -2850,8 +2851,7 @@ int open_ctree(struct super_block *sb, ret = btrfs_read_sys_array(tree_root); mutex_unlock(&fs_info->chunk_mutex); if (ret) { - printk(KERN_ERR "BTRFS: failed to read the system " - "array on %s\n", sb->s_id); + btrfs_err(fs_info, "failed to read the system array: %d", ret); goto fail_sb_buffer; } @@ -2865,8 +2865,7 @@ int open_ctree(struct super_block *sb, generation); if (IS_ERR(chunk_root->node) || !extent_buffer_uptodate(chunk_root->node)) { - printk(KERN_ERR "BTRFS: failed to read chunk root on %s\n", - sb->s_id); + btrfs_err(fs_info, "failed to read chunk root"); if (!IS_ERR(chunk_root->node)) free_extent_buffer(chunk_root->node); chunk_root->node = NULL; @@ -2880,8 +2879,7 @@ int open_ctree(struct super_block *sb, ret = btrfs_read_chunk_tree(chunk_root); if (ret) { - printk(KERN_ERR "BTRFS: failed to read chunk tree on %s\n", - sb->s_id); + btrfs_err(fs_info, "failed to read chunk tree: %d", ret); goto fail_tree_roots; } @@ -2892,8 +2890,7 @@ int open_ctree(struct super_block *sb, btrfs_close_extra_devices(fs_devices, 0); if (!fs_devices->latest_bdev) { - printk(KERN_ERR "BTRFS: failed to read devices on %s\n", - sb->s_id); + btrfs_err(fs_info, "failed to read devices"); goto fail_tree_roots; } @@ -2905,8 +2902,7 @@ retry_root_backup: generation); if (IS_ERR(tree_root->node) || !extent_buffer_uptodate(tree_root->node)) { - printk(KERN_WARNING "BTRFS: failed to read tree root on %s\n", - sb->s_id); + btrfs_warn(fs_info, "failed to read tree root"); if (!IS_ERR(tree_root->node)) free_extent_buffer(tree_root->node); tree_root->node = NULL; @@ -2938,20 +2934,19 @@ retry_root_backup: ret = btrfs_recover_balance(fs_info); if (ret) { - printk(KERN_ERR "BTRFS: failed to recover balance\n"); + btrfs_err(fs_info, "failed to recover balance: %d", ret); goto fail_block_groups; } ret = btrfs_init_dev_stats(fs_info); if (ret) { - printk(KERN_ERR "BTRFS: failed to init dev_stats: %d\n", - ret); + btrfs_err(fs_info, "failed to init dev_stats: %d", ret); goto fail_block_groups; } ret = btrfs_init_dev_replace(fs_info); if (ret) { - pr_err("BTRFS: failed to init dev_replace: %d\n", ret); + btrfs_err(fs_info, "failed to init dev_replace: %d", ret); goto fail_block_groups; } @@ -2959,31 +2954,33 @@ retry_root_backup: ret = btrfs_sysfs_add_fsid(fs_devices, NULL); if (ret) { - pr_err("BTRFS: failed to init sysfs fsid interface: %d\n", ret); + btrfs_err(fs_info, "failed to init sysfs fsid interface: %d", + ret); goto fail_block_groups; } ret = btrfs_sysfs_add_device(fs_devices); if (ret) { - pr_err("BTRFS: failed to init sysfs device interface: %d\n", ret); + btrfs_err(fs_info, "failed to init sysfs device interface: %d", + ret); goto fail_fsdev_sysfs; } ret = btrfs_sysfs_add_mounted(fs_info); if (ret) { - pr_err("BTRFS: failed to init sysfs interface: %d\n", ret); + btrfs_err(fs_info, "failed to init sysfs interface: %d", ret); goto fail_fsdev_sysfs; } ret = btrfs_init_space_info(fs_info); if (ret) { - printk(KERN_ERR "BTRFS: Failed to initial space info: %d\n", ret); + btrfs_err(fs_info, "failed to initialize space info: %d", ret); goto fail_sysfs; } ret = btrfs_read_block_groups(fs_info->extent_root); if (ret) { - printk(KERN_ERR "BTRFS: Failed to read block groups: %d\n", ret); + btrfs_err(fs_info, "failed to read block groups: %d", ret); goto fail_sysfs; } fs_info->num_tolerated_disk_barrier_failures = @@ -2991,7 +2988,8 @@ retry_root_backup: if (fs_info->fs_devices->missing_devices > fs_info->num_tolerated_disk_barrier_failures && !(sb->s_flags & MS_RDONLY)) { - pr_warn("BTRFS: missing devices(%llu) exceeds the limit(%d), writeable mount is not allowed\n", + btrfs_warn(fs_info, +"missing devices (%llu) exceeds the limit (%d), writeable mount is not allowed", fs_info->fs_devices->missing_devices, fs_info->num_tolerated_disk_barrier_failures); goto fail_sysfs; @@ -3011,8 +3009,7 @@ retry_root_backup: if (!btrfs_test_opt(tree_root, SSD) && !btrfs_test_opt(tree_root, NOSSD) && !fs_info->fs_devices->rotating) { - printk(KERN_INFO "BTRFS: detected SSD devices, enabling SSD " - "mode\n"); + btrfs_info(fs_info, "detected SSD devices, enabling SSD mode"); btrfs_set_opt(fs_info->mount_opt, SSD); } @@ -3030,8 +3027,9 @@ retry_root_backup: 1 : 0, fs_info->check_integrity_print_mask); if (ret) - printk(KERN_WARNING "BTRFS: failed to initialize" - " integrity check module %s\n", sb->s_id); + btrfs_warn(fs_info, + "failed to initialize integrity check module: %d", + ret); } #endif ret = btrfs_read_qgroup_config(fs_info); @@ -3061,8 +3059,8 @@ retry_root_backup: ret = btrfs_recover_relocation(tree_root); mutex_unlock(&fs_info->cleaner_mutex); if (ret < 0) { - printk(KERN_WARNING - "BTRFS: failed to recover relocation\n"); + btrfs_warn(fs_info, "failed to recover relocation: %d", + ret); err = -EINVAL; goto fail_qgroup; } @@ -3083,11 +3081,11 @@ retry_root_backup: if (btrfs_test_opt(tree_root, FREE_SPACE_TREE) && !btrfs_fs_compat_ro(fs_info, FREE_SPACE_TREE)) { - pr_info("BTRFS: creating free space tree\n"); + btrfs_info(fs_info, "creating free space tree"); ret = btrfs_create_free_space_tree(fs_info); if (ret) { - pr_warn("BTRFS: failed to create free space tree %d\n", - ret); + btrfs_warn(fs_info, + "failed to create free space tree: %d", ret); close_ctree(tree_root); return ret; } @@ -3104,14 +3102,14 @@ retry_root_backup: ret = btrfs_resume_balance_async(fs_info); if (ret) { - printk(KERN_WARNING "BTRFS: failed to resume balance\n"); + btrfs_warn(fs_info, "failed to resume balance: %d", ret); close_ctree(tree_root); return ret; } ret = btrfs_resume_dev_replace_async(fs_info); if (ret) { - pr_warn("BTRFS: failed to resume dev_replace\n"); + btrfs_warn(fs_info, "failed to resume device replace: %d", ret); close_ctree(tree_root); return ret; } @@ -3120,33 +3118,33 @@ retry_root_backup: if (btrfs_test_opt(tree_root, CLEAR_CACHE) && btrfs_fs_compat_ro(fs_info, FREE_SPACE_TREE)) { - pr_info("BTRFS: clearing free space tree\n"); + btrfs_info(fs_info, "clearing free space tree"); ret = btrfs_clear_free_space_tree(fs_info); if (ret) { - pr_warn("BTRFS: failed to clear free space tree %d\n", - ret); + btrfs_warn(fs_info, + "failed to clear free space tree: %d", ret); close_ctree(tree_root); return ret; } } if (!fs_info->uuid_root) { - pr_info("BTRFS: creating UUID tree\n"); + btrfs_info(fs_info, "creating UUID tree"); ret = btrfs_create_uuid_tree(fs_info); if (ret) { - pr_warn("BTRFS: failed to create the UUID tree %d\n", - ret); + btrfs_warn(fs_info, + "failed to create the UUID tree: %d", ret); close_ctree(tree_root); return ret; } } else if (btrfs_test_opt(tree_root, RESCAN_UUID_TREE) || fs_info->generation != btrfs_super_uuid_tree_generation(disk_super)) { - pr_info("BTRFS: checking UUID tree\n"); + btrfs_info(fs_info, "checking UUID tree"); ret = btrfs_check_uuid_tree(fs_info); if (ret) { - pr_warn("BTRFS: failed to check the UUID tree %d\n", - ret); + btrfs_warn(fs_info, + "failed to check the UUID tree: %d", ret); close_ctree(tree_root); return ret; } -- cgit v1.3-8-gc7d7 From e1860a7724828a341037b010b3f4ff58bad53f95 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Mon, 9 May 2016 14:11:38 +0200 Subject: btrfs: GFP_NOFS does not GFP_HIGHMEM Masking HIGHMEM out of NOFS does not make sense. Signed-off-by: David Sterba --- fs/btrfs/delayed-inode.c | 2 +- fs/btrfs/disk-io.c | 2 +- fs/btrfs/extent_io.c | 4 ++-- 3 files changed, 4 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/delayed-inode.c b/fs/btrfs/delayed-inode.c index 6cef0062f929..61561c2a3f96 100644 --- a/fs/btrfs/delayed-inode.c +++ b/fs/btrfs/delayed-inode.c @@ -134,7 +134,7 @@ again: /* cached in the btrfs inode and can be accessed */ atomic_add(2, &node->refs); - ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM); + ret = radix_tree_preload(GFP_NOFS); if (ret) { kmem_cache_free(delayed_node_cache, node); return ERR_PTR(ret); diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 123fab55b5a4..4931d3856aa3 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1640,7 +1640,7 @@ int btrfs_insert_fs_root(struct btrfs_fs_info *fs_info, { int ret; - ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM); + ret = radix_tree_preload(GFP_NOFS); if (ret) return ret; diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index d22b0777c7e0..d02a637530b1 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -4815,7 +4815,7 @@ struct extent_buffer *alloc_test_extent_buffer(struct btrfs_fs_info *fs_info, return NULL; eb->fs_info = fs_info; again: - ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM); + ret = radix_tree_preload(GFP_NOFS); if (ret) goto free_eb; spin_lock(&fs_info->buffer_lock); @@ -4916,7 +4916,7 @@ struct extent_buffer *alloc_extent_buffer(struct btrfs_fs_info *fs_info, if (uptodate) set_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags); again: - ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM); + ret = radix_tree_preload(GFP_NOFS); if (ret) goto free_eb; -- cgit v1.3-8-gc7d7 From 6ac10a6ac2b11ada24580cc76dcd0c182061c576 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Wed, 27 Apr 2016 02:15:15 +0200 Subject: btrfs: rename and document compression workspace members The names are confusing, pick more fitting names and add comments. Signed-off-by: David Sterba --- fs/btrfs/compression.c | 35 +++++++++++++++++++---------------- 1 file changed, 19 insertions(+), 16 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c index ff61a41ac90b..4d5cd9624bb3 100644 --- a/fs/btrfs/compression.c +++ b/fs/btrfs/compression.c @@ -743,8 +743,11 @@ out: static struct { struct list_head idle_ws; spinlock_t ws_lock; - int num_ws; - atomic_t alloc_ws; + /* Number of free workspaces */ + int free_ws; + /* Total number of allocated workspaces */ + atomic_t total_ws; + /* Waiters for a free workspace */ wait_queue_head_t ws_wait; } btrfs_comp_ws[BTRFS_COMPRESS_TYPES]; @@ -760,7 +763,7 @@ void __init btrfs_init_compress(void) for (i = 0; i < BTRFS_COMPRESS_TYPES; i++) { INIT_LIST_HEAD(&btrfs_comp_ws[i].idle_ws); spin_lock_init(&btrfs_comp_ws[i].ws_lock); - atomic_set(&btrfs_comp_ws[i].alloc_ws, 0); + atomic_set(&btrfs_comp_ws[i].total_ws, 0); init_waitqueue_head(&btrfs_comp_ws[i].ws_wait); } } @@ -777,35 +780,35 @@ static struct list_head *find_workspace(int type) struct list_head *idle_ws = &btrfs_comp_ws[idx].idle_ws; spinlock_t *ws_lock = &btrfs_comp_ws[idx].ws_lock; - atomic_t *alloc_ws = &btrfs_comp_ws[idx].alloc_ws; + atomic_t *total_ws = &btrfs_comp_ws[idx].total_ws; wait_queue_head_t *ws_wait = &btrfs_comp_ws[idx].ws_wait; - int *num_ws = &btrfs_comp_ws[idx].num_ws; + int *free_ws = &btrfs_comp_ws[idx].free_ws; again: spin_lock(ws_lock); if (!list_empty(idle_ws)) { workspace = idle_ws->next; list_del(workspace); - (*num_ws)--; + (*free_ws)--; spin_unlock(ws_lock); return workspace; } - if (atomic_read(alloc_ws) > cpus) { + if (atomic_read(total_ws) > cpus) { DEFINE_WAIT(wait); spin_unlock(ws_lock); prepare_to_wait(ws_wait, &wait, TASK_UNINTERRUPTIBLE); - if (atomic_read(alloc_ws) > cpus && !*num_ws) + if (atomic_read(total_ws) > cpus && !*free_ws) schedule(); finish_wait(ws_wait, &wait); goto again; } - atomic_inc(alloc_ws); + atomic_inc(total_ws); spin_unlock(ws_lock); workspace = btrfs_compress_op[idx]->alloc_workspace(); if (IS_ERR(workspace)) { - atomic_dec(alloc_ws); + atomic_dec(total_ws); wake_up(ws_wait); } return workspace; @@ -820,21 +823,21 @@ static void free_workspace(int type, struct list_head *workspace) int idx = type - 1; struct list_head *idle_ws = &btrfs_comp_ws[idx].idle_ws; spinlock_t *ws_lock = &btrfs_comp_ws[idx].ws_lock; - atomic_t *alloc_ws = &btrfs_comp_ws[idx].alloc_ws; + atomic_t *total_ws = &btrfs_comp_ws[idx].total_ws; wait_queue_head_t *ws_wait = &btrfs_comp_ws[idx].ws_wait; - int *num_ws = &btrfs_comp_ws[idx].num_ws; + int *free_ws = &btrfs_comp_ws[idx].free_ws; spin_lock(ws_lock); - if (*num_ws < num_online_cpus()) { + if (*free_ws < num_online_cpus()) { list_add(workspace, idle_ws); - (*num_ws)++; + (*free_ws)++; spin_unlock(ws_lock); goto wake; } spin_unlock(ws_lock); btrfs_compress_op[idx]->free_workspace(workspace); - atomic_dec(alloc_ws); + atomic_dec(total_ws); wake: /* * Make sure counter is updated before we wake up waiters. @@ -857,7 +860,7 @@ static void free_workspaces(void) workspace = btrfs_comp_ws[i].idle_ws.next; list_del(workspace); btrfs_compress_op[i]->free_workspace(workspace); - atomic_dec(&btrfs_comp_ws[i].alloc_ws); + atomic_dec(&btrfs_comp_ws[i].total_ws); } } } -- cgit v1.3-8-gc7d7 From f77dd0d6b2f0f2cf290cacbd48f5eee18586e52b Mon Sep 17 00:00:00 2001 From: David Sterba Date: Wed, 27 Apr 2016 02:55:15 +0200 Subject: btrfs: preallocate compression workspaces Preallocate one workspace for each compression type so we can guarantee forward progress in the worst case. A failure cannot be a hard error as we might not use compression at all on the filesystem. If we can't allocate the workspaces later when need them, it might actually deadlock, but in such situation the system has effectively not enough memory to operate properly. Signed-off-by: David Sterba --- fs/btrfs/compression.c | 16 ++++++++++++++++ 1 file changed, 16 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c index 4d5cd9624bb3..38c058bcf359 100644 --- a/fs/btrfs/compression.c +++ b/fs/btrfs/compression.c @@ -761,10 +761,26 @@ void __init btrfs_init_compress(void) int i; for (i = 0; i < BTRFS_COMPRESS_TYPES; i++) { + struct list_head *workspace; + INIT_LIST_HEAD(&btrfs_comp_ws[i].idle_ws); spin_lock_init(&btrfs_comp_ws[i].ws_lock); atomic_set(&btrfs_comp_ws[i].total_ws, 0); init_waitqueue_head(&btrfs_comp_ws[i].ws_wait); + + /* + * Preallocate one workspace for each compression type so + * we can guarantee forward progress in the worst case + */ + workspace = btrfs_compress_op[i]->alloc_workspace(); + if (IS_ERR(workspace)) { + printk(KERN_WARNING + "BTRFS: cannot preallocate compression workspace, will try later"); + } else { + atomic_set(&btrfs_comp_ws[i].total_ws, 1); + btrfs_comp_ws[i].free_ws = 1; + list_add(workspace, &btrfs_comp_ws[i].idle_ws); + } } } -- cgit v1.3-8-gc7d7 From e721e49dd1681d45d71919f0561f5e978a34153c Mon Sep 17 00:00:00 2001 From: David Sterba Date: Wed, 27 Apr 2016 02:41:17 +0200 Subject: btrfs: make find_workspace always succeed With just one preallocated workspace we can guarantee forward progress even if there's no memory available for new workspaces. The cost is more waiting but we also get rid of several error paths. On average, there will be several idle workspaces, so the waiting penalty won't be so bad. In the worst case, all cpus will compete for one workspace until there's some memory. Attempts to allocate a new one are done each time the waiters are woken up. Signed-off-by: David Sterba --- fs/btrfs/compression.c | 20 ++++++++++++-------- 1 file changed, 12 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c index 38c058bcf359..c70625560265 100644 --- a/fs/btrfs/compression.c +++ b/fs/btrfs/compression.c @@ -785,8 +785,10 @@ void __init btrfs_init_compress(void) } /* - * this finds an available workspace or allocates a new one - * ERR_PTR is returned if things go bad. + * This finds an available workspace or allocates a new one. + * If it's not possible to allocate a new one, waits until there's one. + * Preallocation makes a forward progress guarantees and we do not return + * errors. */ static struct list_head *find_workspace(int type) { @@ -826,6 +828,14 @@ again: if (IS_ERR(workspace)) { atomic_dec(total_ws); wake_up(ws_wait); + + /* + * Do not return the error but go back to waiting. There's a + * workspace preallocated for each type and the compression + * time is bounded so we get to a workspace eventually. This + * makes our caller's life easier. + */ + goto again; } return workspace; } @@ -913,8 +923,6 @@ int btrfs_compress_pages(int type, struct address_space *mapping, int ret; workspace = find_workspace(type); - if (IS_ERR(workspace)) - return PTR_ERR(workspace); ret = btrfs_compress_op[type-1]->compress_pages(workspace, mapping, start, len, pages, @@ -949,8 +957,6 @@ static int btrfs_decompress_biovec(int type, struct page **pages_in, int ret; workspace = find_workspace(type); - if (IS_ERR(workspace)) - return PTR_ERR(workspace); ret = btrfs_compress_op[type-1]->decompress_biovec(workspace, pages_in, disk_start, @@ -971,8 +977,6 @@ int btrfs_decompress(int type, unsigned char *data_in, struct page *dest_page, int ret; workspace = find_workspace(type); - if (IS_ERR(workspace)) - return PTR_ERR(workspace); ret = btrfs_compress_op[type-1]->decompress(workspace, data_in, dest_page, start_byte, -- cgit v1.3-8-gc7d7 From 523567168da04bae0f88802ddef49d00072c5d58 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Wed, 27 Apr 2016 03:07:39 +0200 Subject: btrfs: make find_workspace warn if there are no workspaces Be verbose if there are no workspaces at all, ie. the module init time preallocation failed. Signed-off-by: David Sterba --- fs/btrfs/compression.c | 14 ++++++++++++++ 1 file changed, 14 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c index c70625560265..658c39b70fba 100644 --- a/fs/btrfs/compression.c +++ b/fs/btrfs/compression.c @@ -834,7 +834,21 @@ again: * workspace preallocated for each type and the compression * time is bounded so we get to a workspace eventually. This * makes our caller's life easier. + * + * To prevent silent and low-probability deadlocks (when the + * initial preallocation fails), check if there are any + * workspaces at all. */ + if (atomic_read(total_ws) == 0) { + static DEFINE_RATELIMIT_STATE(_rs, + /* once per minute */ 60 * HZ, + /* no burst */ 1); + + if (__ratelimit(&_rs)) { + printk(KERN_WARNING + "no compression workspaces, low memory, retrying"); + } + } goto again; } return workspace; -- cgit v1.3-8-gc7d7 From 72928f2476d08c79f132b4f44a17c9a011dd98e3 Mon Sep 17 00:00:00 2001 From: Vincent Stehlé Date: Tue, 10 May 2016 14:56:20 +0200 Subject: Btrfs: fix fspath error deallocation MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Make sure to deallocate fspath with vfree() in case of error in init_ipath(). fspath is allocated with vmalloc() in init_data_container() since commit 425d17a290c0 ("Btrfs: use larger limit for translation of logical to inode"). Signed-off-by: Vincent Stehlé Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/backref.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/backref.c b/fs/btrfs/backref.c index 80e8472d618b..d3090187fd76 100644 --- a/fs/btrfs/backref.c +++ b/fs/btrfs/backref.c @@ -1991,7 +1991,7 @@ struct inode_fs_paths *init_ipath(s32 total_bytes, struct btrfs_root *fs_root, ifp = kmalloc(sizeof(*ifp), GFP_NOFS); if (!ifp) { - kfree(fspath); + vfree(fspath); return ERR_PTR(-ENOMEM); } -- cgit v1.3-8-gc7d7 From 7fb17fe44b70c854d380686d8b7e63217f9f1dcf Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Mon, 9 May 2016 19:56:32 +0800 Subject: f2fs: use mnt_{want,drop}_write_file in ioctl In interfaces of ioctl, mnt_{want,drop}_write_file should be used for: - get exclusion against file system freezing which may used by lvm snapshot. - do telling filesystem that a write is about to be performed on it, and make sure that the writes are permitted. Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/file.c | 117 ++++++++++++++++++++++++++++++++++++++++++--------------- 1 file changed, 87 insertions(+), 30 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index 314612997721..2f8f225ddbc5 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -1307,20 +1307,16 @@ static int f2fs_ioc_setflags(struct file *filp, unsigned long arg) unsigned int oldflags; int ret; + if (!inode_owner_or_capable(inode)) + return -EACCES; + + if (get_user(flags, (int __user *)arg)) + return -EFAULT; + ret = mnt_want_write_file(filp); if (ret) return ret; - if (!inode_owner_or_capable(inode)) { - ret = -EACCES; - goto out; - } - - if (get_user(flags, (int __user *)arg)) { - ret = -EFAULT; - goto out; - } - flags = f2fs_mask_flags(inode->i_mode, flags); inode_lock(inode); @@ -1363,18 +1359,22 @@ static int f2fs_ioc_start_atomic_write(struct file *filp) if (!inode_owner_or_capable(inode)) return -EACCES; + ret = mnt_want_write_file(filp); + if (ret) + return ret; + if (f2fs_is_atomic_file(inode)) - return 0; + goto out; ret = f2fs_convert_inline_inode(inode); if (ret) - return ret; + goto out; set_inode_flag(F2FS_I(inode), FI_ATOMIC_FILE); f2fs_update_time(F2FS_I_SB(inode), REQ_TIME); if (!get_dirty_pages(inode)) - return 0; + goto out; f2fs_msg(F2FS_I_SB(inode)->sb, KERN_WARNING, "Unexpected flush for atomic writes: ino=%lu, npages=%u", @@ -1382,6 +1382,8 @@ static int f2fs_ioc_start_atomic_write(struct file *filp) ret = filemap_write_and_wait_range(inode->i_mapping, 0, LLONG_MAX); if (ret) clear_inode_flag(F2FS_I(inode), FI_ATOMIC_FILE); +out: + mnt_drop_write_file(filp); return ret; } @@ -1393,13 +1395,13 @@ static int f2fs_ioc_commit_atomic_write(struct file *filp) if (!inode_owner_or_capable(inode)) return -EACCES; - if (f2fs_is_volatile_file(inode)) - return 0; - ret = mnt_want_write_file(filp); if (ret) return ret; + if (f2fs_is_volatile_file(inode)) + goto err_out; + if (f2fs_is_atomic_file(inode)) { clear_inode_flag(F2FS_I(inode), FI_ATOMIC_FILE); ret = commit_inmem_pages(inode); @@ -1423,32 +1425,48 @@ static int f2fs_ioc_start_volatile_write(struct file *filp) if (!inode_owner_or_capable(inode)) return -EACCES; + ret = mnt_want_write_file(filp); + if (ret) + return ret; + if (f2fs_is_volatile_file(inode)) - return 0; + goto out; ret = f2fs_convert_inline_inode(inode); if (ret) - return ret; + goto out; set_inode_flag(F2FS_I(inode), FI_VOLATILE_FILE); f2fs_update_time(F2FS_I_SB(inode), REQ_TIME); - return 0; +out: + mnt_drop_write_file(filp); + return ret; } static int f2fs_ioc_release_volatile_write(struct file *filp) { struct inode *inode = file_inode(filp); + int ret; if (!inode_owner_or_capable(inode)) return -EACCES; + ret = mnt_want_write_file(filp); + if (ret) + return ret; + if (!f2fs_is_volatile_file(inode)) - return 0; + goto out; - if (!f2fs_is_first_block_written(inode)) - return truncate_partial_data_page(inode, 0, true); + if (!f2fs_is_first_block_written(inode)) { + ret = truncate_partial_data_page(inode, 0, true); + goto out; + } - return punch_hole(inode, 0, F2FS_BLKSIZE); + ret = punch_hole(inode, 0, F2FS_BLKSIZE); +out: + mnt_drop_write_file(filp); + return ret; } static int f2fs_ioc_abort_volatile_write(struct file *filp) @@ -1481,6 +1499,7 @@ static int f2fs_ioc_shutdown(struct file *filp, unsigned long arg) struct f2fs_sb_info *sbi = F2FS_I_SB(inode); struct super_block *sb = sbi->sb; __u32 in; + int ret; if (!capable(CAP_SYS_ADMIN)) return -EPERM; @@ -1488,6 +1507,10 @@ static int f2fs_ioc_shutdown(struct file *filp, unsigned long arg) if (get_user(in, (__u32 __user *)arg)) return -EFAULT; + ret = mnt_want_write_file(filp); + if (ret) + return ret; + switch (in) { case F2FS_GOING_DOWN_FULLSYNC: sb = freeze_bdev(sb->s_bdev); @@ -1509,10 +1532,13 @@ static int f2fs_ioc_shutdown(struct file *filp, unsigned long arg) f2fs_stop_checkpoint(sbi); break; default: - return -EINVAL; + ret = -EINVAL; + goto out; } f2fs_update_time(sbi, REQ_TIME); - return 0; +out: + mnt_drop_write_file(filp); + return ret; } static int f2fs_ioc_fitrim(struct file *filp, unsigned long arg) @@ -1533,9 +1559,14 @@ static int f2fs_ioc_fitrim(struct file *filp, unsigned long arg) sizeof(range))) return -EFAULT; + ret = mnt_want_write_file(filp); + if (ret) + return ret; + range.minlen = max((unsigned int)range.minlen, q->limits.discard_granularity); ret = f2fs_trim_fs(F2FS_SB(sb), &range); + mnt_drop_write_file(filp); if (ret < 0) return ret; @@ -1560,13 +1591,21 @@ static int f2fs_ioc_set_encryption_policy(struct file *filp, unsigned long arg) { struct fscrypt_policy policy; struct inode *inode = file_inode(filp); + int ret; if (copy_from_user(&policy, (struct fscrypt_policy __user *)arg, sizeof(policy))) return -EFAULT; + ret = mnt_want_write_file(filp); + if (ret) + return ret; + f2fs_update_time(F2FS_I_SB(inode), REQ_TIME); - return fscrypt_process_policy(inode, &policy); + ret = fscrypt_process_policy(inode, &policy); + + mnt_drop_write_file(filp); + return ret; } static int f2fs_ioc_get_encryption_policy(struct file *filp, unsigned long arg) @@ -1623,6 +1662,7 @@ static int f2fs_ioc_gc(struct file *filp, unsigned long arg) struct inode *inode = file_inode(filp); struct f2fs_sb_info *sbi = F2FS_I_SB(inode); __u32 sync; + int ret; if (!capable(CAP_SYS_ADMIN)) return -EPERM; @@ -1633,20 +1673,30 @@ static int f2fs_ioc_gc(struct file *filp, unsigned long arg) if (f2fs_readonly(sbi->sb)) return -EROFS; + ret = mnt_want_write_file(filp); + if (ret) + return ret; + if (!sync) { - if (!mutex_trylock(&sbi->gc_mutex)) - return -EBUSY; + if (!mutex_trylock(&sbi->gc_mutex)) { + ret = -EBUSY; + goto out; + } } else { mutex_lock(&sbi->gc_mutex); } - return f2fs_gc(sbi, sync); + ret = f2fs_gc(sbi, sync); +out: + mnt_drop_write_file(filp); + return ret; } static int f2fs_ioc_write_checkpoint(struct file *filp, unsigned long arg) { struct inode *inode = file_inode(filp); struct f2fs_sb_info *sbi = F2FS_I_SB(inode); + int ret; if (!capable(CAP_SYS_ADMIN)) return -EPERM; @@ -1654,7 +1704,14 @@ static int f2fs_ioc_write_checkpoint(struct file *filp, unsigned long arg) if (f2fs_readonly(sbi->sb)) return -EROFS; - return f2fs_sync_fs(sbi->sb, 1); + ret = mnt_want_write_file(filp); + if (ret) + return ret; + + ret = f2fs_sync_fs(sbi->sb, 1); + + mnt_drop_write_file(filp); + return ret; } static int f2fs_defragment_range(struct f2fs_sb_info *sbi, -- cgit v1.3-8-gc7d7 From 0fac558b96584799876498248020dc49a98bd131 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Mon, 9 May 2016 19:56:33 +0800 Subject: f2fs: make atomic/volatile operation exclusive atomic/volatile ioctl interfaces are exposed to user like other file operation interface, it needs to make them getting exclusion against to each other to avoid potential conflict among these operations in concurrent scenario. Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/file.c | 16 ++++++++++++++++ 1 file changed, 16 insertions(+) (limited to 'fs') diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index 2f8f225ddbc5..f67ca62c4f2a 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -1363,6 +1363,8 @@ static int f2fs_ioc_start_atomic_write(struct file *filp) if (ret) return ret; + inode_lock(inode); + if (f2fs_is_atomic_file(inode)) goto out; @@ -1383,6 +1385,7 @@ static int f2fs_ioc_start_atomic_write(struct file *filp) if (ret) clear_inode_flag(F2FS_I(inode), FI_ATOMIC_FILE); out: + inode_unlock(inode); mnt_drop_write_file(filp); return ret; } @@ -1399,6 +1402,8 @@ static int f2fs_ioc_commit_atomic_write(struct file *filp) if (ret) return ret; + inode_lock(inode); + if (f2fs_is_volatile_file(inode)) goto err_out; @@ -1413,6 +1418,7 @@ static int f2fs_ioc_commit_atomic_write(struct file *filp) ret = f2fs_do_sync_file(filp, 0, LLONG_MAX, 0, true); err_out: + inode_unlock(inode); mnt_drop_write_file(filp); return ret; } @@ -1429,6 +1435,8 @@ static int f2fs_ioc_start_volatile_write(struct file *filp) if (ret) return ret; + inode_lock(inode); + if (f2fs_is_volatile_file(inode)) goto out; @@ -1439,6 +1447,7 @@ static int f2fs_ioc_start_volatile_write(struct file *filp) set_inode_flag(F2FS_I(inode), FI_VOLATILE_FILE); f2fs_update_time(F2FS_I_SB(inode), REQ_TIME); out: + inode_unlock(inode); mnt_drop_write_file(filp); return ret; } @@ -1455,6 +1464,8 @@ static int f2fs_ioc_release_volatile_write(struct file *filp) if (ret) return ret; + inode_lock(inode); + if (!f2fs_is_volatile_file(inode)) goto out; @@ -1465,6 +1476,7 @@ static int f2fs_ioc_release_volatile_write(struct file *filp) ret = punch_hole(inode, 0, F2FS_BLKSIZE); out: + inode_unlock(inode); mnt_drop_write_file(filp); return ret; } @@ -1481,6 +1493,8 @@ static int f2fs_ioc_abort_volatile_write(struct file *filp) if (ret) return ret; + inode_lock(inode); + if (f2fs_is_atomic_file(inode)) drop_inmem_pages(inode); if (f2fs_is_volatile_file(inode)) { @@ -1488,6 +1502,8 @@ static int f2fs_ioc_abort_volatile_write(struct file *filp) ret = f2fs_do_sync_file(filp, 0, LLONG_MAX, 0, true); } + inode_unlock(inode); + mnt_drop_write_file(filp); f2fs_update_time(F2FS_I_SB(inode), REQ_TIME); return ret; -- cgit v1.3-8-gc7d7 From 46008c6d42328710f9beaf5c2b47dc92b1cc1a75 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Mon, 9 May 2016 19:56:30 +0800 Subject: f2fs: support in batch multi blocks preallocation This patch introduces reserve_new_blocks to make preallocation of multi blocks as in batch operation, so it can avoid lots of redundant operation, result in better performance. In virtual machine, with rotational device: time fallocate -l 32G /mnt/f2fs/file Before: real 0m4.584s user 0m0.000s sys 0m4.580s After: real 0m0.292s user 0m0.000s sys 0m0.272s In x86, with SSD: time fallocate -l 500G $MNT/testfile Before : 24.758 s After : 1.604 s Signed-off-by: Chao Yu [Jaegeuk Kim: fix bugs and add performance numbers measured in x86.] Signed-off-by: Jaegeuk Kim --- fs/f2fs/data.c | 132 +++++++++++++++++++++++++++++++------------- fs/f2fs/f2fs.h | 20 ++++--- include/trace/events/f2fs.h | 14 +++-- 3 files changed, 117 insertions(+), 49 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index 369d953bd770..b61e2d40cdfb 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -278,6 +278,16 @@ alloc_new: trace_f2fs_submit_page_mbio(fio->page, fio); } +static void __set_data_blkaddr(struct dnode_of_data *dn) +{ + struct f2fs_node *rn = F2FS_NODE(dn->node_page); + __le32 *addr_array; + + /* Get physical address of data block */ + addr_array = blkaddr_in_node(rn); + addr_array[dn->ofs_in_node] = cpu_to_le32(dn->data_blkaddr); +} + /* * Lock ordering for the change of data block address: * ->data_page @@ -286,19 +296,9 @@ alloc_new: */ void set_data_blkaddr(struct dnode_of_data *dn) { - struct f2fs_node *rn; - __le32 *addr_array; - struct page *node_page = dn->node_page; - unsigned int ofs_in_node = dn->ofs_in_node; - - f2fs_wait_on_page_writeback(node_page, NODE, true); - - rn = F2FS_NODE(node_page); - - /* Get physical address of data block */ - addr_array = blkaddr_in_node(rn); - addr_array[ofs_in_node] = cpu_to_le32(dn->data_blkaddr); - if (set_page_dirty(node_page)) + f2fs_wait_on_page_writeback(dn->node_page, NODE, true); + __set_data_blkaddr(dn); + if (set_page_dirty(dn->node_page)) dn->node_changed = true; } @@ -309,24 +309,53 @@ void f2fs_update_data_blkaddr(struct dnode_of_data *dn, block_t blkaddr) f2fs_update_extent_cache(dn); } -int reserve_new_block(struct dnode_of_data *dn) +/* dn->ofs_in_node will be returned with up-to-date last block pointer */ +int reserve_new_blocks(struct dnode_of_data *dn, blkcnt_t count) { struct f2fs_sb_info *sbi = F2FS_I_SB(dn->inode); + if (!count) + return 0; + if (unlikely(is_inode_flag_set(F2FS_I(dn->inode), FI_NO_ALLOC))) return -EPERM; - if (unlikely(!inc_valid_block_count(sbi, dn->inode, 1))) + if (unlikely(!inc_valid_block_count(sbi, dn->inode, &count))) return -ENOSPC; - trace_f2fs_reserve_new_block(dn->inode, dn->nid, dn->ofs_in_node); + trace_f2fs_reserve_new_blocks(dn->inode, dn->nid, + dn->ofs_in_node, count); + + f2fs_wait_on_page_writeback(dn->node_page, NODE, true); + + for (; count > 0; dn->ofs_in_node++) { + block_t blkaddr = + datablock_addr(dn->node_page, dn->ofs_in_node); + if (blkaddr == NULL_ADDR) { + dn->data_blkaddr = NEW_ADDR; + __set_data_blkaddr(dn); + count--; + } + } + + if (set_page_dirty(dn->node_page)) + dn->node_changed = true; - dn->data_blkaddr = NEW_ADDR; - set_data_blkaddr(dn); mark_inode_dirty(dn->inode); sync_inode_page(dn); return 0; } +/* Should keep dn->ofs_in_node unchanged */ +int reserve_new_block(struct dnode_of_data *dn) +{ + unsigned int ofs_in_node = dn->ofs_in_node; + int ret; + + ret = reserve_new_blocks(dn, 1); + dn->ofs_in_node = ofs_in_node; + return ret; +} + int f2fs_reserve_block(struct dnode_of_data *dn, pgoff_t index) { bool need_put = dn->inode_page ? false : true; @@ -545,6 +574,7 @@ static int __allocate_data_block(struct dnode_of_data *dn) struct node_info ni; int seg = CURSEG_WARM_DATA; pgoff_t fofs; + blkcnt_t count = 1; if (unlikely(is_inode_flag_set(F2FS_I(dn->inode), FI_NO_ALLOC))) return -EPERM; @@ -553,7 +583,7 @@ static int __allocate_data_block(struct dnode_of_data *dn) if (dn->data_blkaddr == NEW_ADDR) goto alloc; - if (unlikely(!inc_valid_block_count(sbi, dn->inode, 1))) + if (unlikely(!inc_valid_block_count(sbi, dn->inode, &count))) return -ENOSPC; alloc: @@ -621,8 +651,10 @@ int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map, struct dnode_of_data dn; struct f2fs_sb_info *sbi = F2FS_I_SB(inode); int mode = create ? ALLOC_NODE : LOOKUP_NODE_RA; - pgoff_t pgofs, end_offset; + pgoff_t pgofs, end_offset, end; int err = 0, ofs = 1; + unsigned int ofs_in_node, last_ofs_in_node; + blkcnt_t prealloc; struct extent_info ei; bool allocated = false; block_t blkaddr; @@ -632,6 +664,7 @@ int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map, /* it only supports block size == page size */ pgofs = (pgoff_t)map->m_lblk; + end = pgofs + maxblocks; if (!create && f2fs_lookup_extent_cache(inode, pgofs, &ei)) { map->m_pblk = ei.blk + pgofs - ei.fofs; @@ -659,6 +692,8 @@ next_dnode: goto unlock_out; } + prealloc = 0; + ofs_in_node = dn.ofs_in_node; end_offset = ADDRS_PER_PAGE(dn.node_page, inode); next_block: @@ -671,17 +706,20 @@ next_block: goto sync_out; } if (flag == F2FS_GET_BLOCK_PRE_AIO) { - if (blkaddr == NULL_ADDR) - err = reserve_new_block(&dn); + if (blkaddr == NULL_ADDR) { + prealloc++; + last_ofs_in_node = dn.ofs_in_node; + } } else { err = __allocate_data_block(&dn); - if (!err) + if (!err) { set_inode_flag(F2FS_I(inode), FI_APPEND_WRITE); + allocated = true; + } } if (err) goto sync_out; - allocated = true; map->m_flags = F2FS_MAP_NEW; blkaddr = dn.data_blkaddr; } else { @@ -700,6 +738,9 @@ next_block: } } + if (flag == F2FS_GET_BLOCK_PRE_AIO) + goto skip; + if (map->m_len == 0) { /* preallocated unwritten block should be mapped for fiemap. */ if (blkaddr == NEW_ADDR) @@ -711,32 +752,49 @@ next_block: } else if ((map->m_pblk != NEW_ADDR && blkaddr == (map->m_pblk + ofs)) || (map->m_pblk == NEW_ADDR && blkaddr == NEW_ADDR) || - flag == F2FS_GET_BLOCK_PRE_DIO || - flag == F2FS_GET_BLOCK_PRE_AIO) { + flag == F2FS_GET_BLOCK_PRE_DIO) { ofs++; map->m_len++; } else { goto sync_out; } +skip: dn.ofs_in_node++; pgofs++; - if (map->m_len < maxblocks) { - if (dn.ofs_in_node < end_offset) - goto next_block; + /* preallocate blocks in batch for one dnode page */ + if (flag == F2FS_GET_BLOCK_PRE_AIO && + (pgofs == end || dn.ofs_in_node == end_offset)) { - if (allocated) - sync_inode_page(&dn); - f2fs_put_dnode(&dn); + dn.ofs_in_node = ofs_in_node; + err = reserve_new_blocks(&dn, prealloc); + if (err) + goto sync_out; - if (create) { - f2fs_unlock_op(sbi); - f2fs_balance_fs(sbi, allocated); + map->m_len += dn.ofs_in_node - ofs_in_node; + if (prealloc && dn.ofs_in_node != last_ofs_in_node + 1) { + err = -ENOSPC; + goto sync_out; } - allocated = false; - goto next_dnode; + dn.ofs_in_node = end_offset; + } + + if (pgofs >= end) + goto sync_out; + else if (dn.ofs_in_node < end_offset) + goto next_block; + + if (allocated) + sync_inode_page(&dn); + f2fs_put_dnode(&dn); + + if (create) { + f2fs_unlock_op(sbi); + f2fs_balance_fs(sbi, allocated); } + allocated = false; + goto next_dnode; sync_out: if (allocated) diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 052f5a8c96f1..1401c96724c6 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -1094,7 +1094,7 @@ static inline bool f2fs_has_xattr_block(unsigned int ofs) } static inline bool inc_valid_block_count(struct f2fs_sb_info *sbi, - struct inode *inode, blkcnt_t count) + struct inode *inode, blkcnt_t *count) { block_t valid_block_count; @@ -1106,14 +1106,19 @@ static inline bool inc_valid_block_count(struct f2fs_sb_info *sbi, } #endif valid_block_count = - sbi->total_valid_block_count + (block_t)count; + sbi->total_valid_block_count + (block_t)(*count); if (unlikely(valid_block_count > sbi->user_block_count)) { - spin_unlock(&sbi->stat_lock); - return false; + *count = sbi->user_block_count - sbi->total_valid_block_count; + if (!*count) { + spin_unlock(&sbi->stat_lock); + return false; + } } - inode->i_blocks += count; - sbi->total_valid_block_count = valid_block_count; - sbi->alloc_valid_block_count += (block_t)count; + /* *count can be recalculated */ + inode->i_blocks += *count; + sbi->total_valid_block_count = + sbi->total_valid_block_count + (block_t)(*count); + sbi->alloc_valid_block_count += (block_t)(*count); spin_unlock(&sbi->stat_lock); return true; } @@ -1945,6 +1950,7 @@ int f2fs_submit_page_bio(struct f2fs_io_info *); void f2fs_submit_page_mbio(struct f2fs_io_info *); void set_data_blkaddr(struct dnode_of_data *); void f2fs_update_data_blkaddr(struct dnode_of_data *, block_t); +int reserve_new_blocks(struct dnode_of_data *, blkcnt_t); int reserve_new_block(struct dnode_of_data *); int f2fs_get_block(struct dnode_of_data *, pgoff_t); ssize_t f2fs_preallocate_blocks(struct kiocb *, struct iov_iter *); diff --git a/include/trace/events/f2fs.h b/include/trace/events/f2fs.h index 0f565845707b..497e6e80fb63 100644 --- a/include/trace/events/f2fs.h +++ b/include/trace/events/f2fs.h @@ -694,28 +694,32 @@ TRACE_EVENT(f2fs_direct_IO_exit, __entry->ret) ); -TRACE_EVENT(f2fs_reserve_new_block, +TRACE_EVENT(f2fs_reserve_new_blocks, - TP_PROTO(struct inode *inode, nid_t nid, unsigned int ofs_in_node), + TP_PROTO(struct inode *inode, nid_t nid, unsigned int ofs_in_node, + blkcnt_t count), - TP_ARGS(inode, nid, ofs_in_node), + TP_ARGS(inode, nid, ofs_in_node, count), TP_STRUCT__entry( __field(dev_t, dev) __field(nid_t, nid) __field(unsigned int, ofs_in_node) + __field(blkcnt_t, count) ), TP_fast_assign( __entry->dev = inode->i_sb->s_dev; __entry->nid = nid; __entry->ofs_in_node = ofs_in_node; + __entry->count = count; ), - TP_printk("dev = (%d,%d), nid = %u, ofs_in_node = %u", + TP_printk("dev = (%d,%d), nid = %u, ofs_in_node = %u, count = %llu", show_dev(__entry), (unsigned int)__entry->nid, - __entry->ofs_in_node) + __entry->ofs_in_node, + (unsigned long long)__entry->count) ); DECLARE_EVENT_CLASS(f2fs__submit_page_bio, -- cgit v1.3-8-gc7d7 From 6e9619499f53b22ead972e476c0e8341c997d929 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Mon, 9 May 2016 19:56:31 +0800 Subject: f2fs: support in batch fzero in dnode page This patch tries to speedup fzero_range by making space preallocation and address removal of blocks in one dnode page as in batch operation. In virtual machine, with zram driver: dd if=/dev/zero of=/mnt/f2fs/file bs=1M count=4096 time xfs_io -f /mnt/f2fs/file -c "fzero 0 4096M" Before: real 0m3.276s user 0m0.008s sys 0m3.260s After: real 0m1.568s user 0m0.000s sys 0m1.564s Signed-off-by: Chao Yu [Jaegeuk Kim: consider ENOSPC case] Signed-off-by: Jaegeuk Kim --- fs/f2fs/file.c | 72 +++++++++++++++++++++++++++++++++++++++++++++------------- 1 file changed, 56 insertions(+), 16 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index f67ca62c4f2a..a9cb3148bda5 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -997,6 +997,49 @@ static int f2fs_collapse_range(struct inode *inode, loff_t offset, loff_t len) return ret; } +static int f2fs_do_zero_range(struct dnode_of_data *dn, pgoff_t start, + pgoff_t end) +{ + struct f2fs_sb_info *sbi = F2FS_I_SB(dn->inode); + pgoff_t index = start; + unsigned int ofs_in_node = dn->ofs_in_node; + blkcnt_t count = 0; + int ret; + + for (; index < end; index++, dn->ofs_in_node++) { + if (datablock_addr(dn->node_page, dn->ofs_in_node) == NULL_ADDR) + count++; + } + + dn->ofs_in_node = ofs_in_node; + ret = reserve_new_blocks(dn, count); + if (ret) + return ret; + + dn->ofs_in_node = ofs_in_node; + for (index = start; index < end; index++, dn->ofs_in_node++) { + dn->data_blkaddr = + datablock_addr(dn->node_page, dn->ofs_in_node); + /* + * reserve_new_blocks will not guarantee entire block + * allocation. + */ + if (dn->data_blkaddr == NULL_ADDR) { + ret = -ENOSPC; + break; + } + if (dn->data_blkaddr != NEW_ADDR) { + invalidate_blocks(sbi, dn->data_blkaddr); + dn->data_blkaddr = NEW_ADDR; + set_data_blkaddr(dn); + } + } + + f2fs_update_extent_cache_range(dn, start, 0, index - start); + + return ret; +} + static int f2fs_zero_range(struct inode *inode, loff_t offset, loff_t len, int mode) { @@ -1047,35 +1090,32 @@ static int f2fs_zero_range(struct inode *inode, loff_t offset, loff_t len, (loff_t)pg_start << PAGE_SHIFT); } - for (index = pg_start; index < pg_end; index++) { + for (index = pg_start; index < pg_end;) { struct dnode_of_data dn; - struct page *ipage; + unsigned int end_offset; + pgoff_t end; f2fs_lock_op(sbi); - ipage = get_node_page(sbi, inode->i_ino); - if (IS_ERR(ipage)) { - ret = PTR_ERR(ipage); - f2fs_unlock_op(sbi); - goto out; - } - - set_new_dnode(&dn, inode, ipage, NULL, 0); - ret = f2fs_reserve_block(&dn, index); + set_new_dnode(&dn, inode, NULL, NULL, 0); + ret = get_dnode_of_data(&dn, index, ALLOC_NODE); if (ret) { f2fs_unlock_op(sbi); goto out; } - if (dn.data_blkaddr != NEW_ADDR) { - invalidate_blocks(sbi, dn.data_blkaddr); - f2fs_update_data_blkaddr(&dn, NEW_ADDR); - } + end_offset = ADDRS_PER_PAGE(dn.node_page, inode); + end = min(pg_end, end_offset - dn.ofs_in_node + index); + + ret = f2fs_do_zero_range(&dn, index, end); f2fs_put_dnode(&dn); f2fs_unlock_op(sbi); + if (ret) + goto out; + index = end; new_size = max_t(loff_t, new_size, - (loff_t)(index + 1) << PAGE_SHIFT); + (loff_t)index << PAGE_SHIFT); } if (off_end) { -- cgit v1.3-8-gc7d7 From 652be55162dcf3eb4440a4f7536ffedbe0352dcf Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Tue, 10 May 2016 19:13:50 -0700 Subject: f2fs: show # of orphan inodes This adds debug information for # of orphan inodes. Signed-off-by: Jaegeuk Kim --- fs/f2fs/debug.c | 5 ++++- fs/f2fs/f2fs.h | 2 +- 2 files changed, 5 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/debug.c b/fs/f2fs/debug.c index 818064bacfda..37615b254581 100644 --- a/fs/f2fs/debug.c +++ b/fs/f2fs/debug.c @@ -58,6 +58,7 @@ static void update_general_status(struct f2fs_sb_info *sbi) si->inline_xattr = atomic_read(&sbi->inline_xattr); si->inline_inode = atomic_read(&sbi->inline_inode); si->inline_dir = atomic_read(&sbi->inline_dir); + si->orphans = sbi->im[ORPHAN_INO].ino_num; si->utilization = utilization(sbi); si->free_segs = free_segments(sbi); @@ -192,7 +193,7 @@ get_cache: si->cache_mem += NM_I(sbi)->dirty_nat_cnt * sizeof(struct nat_entry_set); si->cache_mem += si->inmem_pages * sizeof(struct inmem_pages); - for (i = 0; i <= UPDATE_INO; i++) + for (i = 0; i <= ORPHAN_INO; i++) si->cache_mem += sbi->im[i].ino_num * sizeof(struct ino_entry); si->cache_mem += atomic_read(&sbi->total_ext_tree) * sizeof(struct extent_tree); @@ -238,6 +239,8 @@ static int stat_show(struct seq_file *s, void *v) si->inline_inode); seq_printf(s, " - Inline_dentry Inode: %u\n", si->inline_dir); + seq_printf(s, " - Orphan Inode: %u\n", + si->orphans); seq_printf(s, "\nMain area: %d segs, %d secs %d zones\n", si->main_area_segs, si->main_area_sections, si->main_area_zones); diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 1401c96724c6..ec978b44d546 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -1997,7 +1997,7 @@ struct f2fs_stat_info { int nats, dirty_nats, sits, dirty_sits, fnids; int total_count, utilization; int bg_gc, inmem_pages, wb_pages; - int inline_xattr, inline_inode, inline_dir; + int inline_xattr, inline_inode, inline_dir, orphans; unsigned int valid_count, valid_node_count, valid_inode_count; unsigned int bimodal, avg_vblocks; int util_free, util_valid, util_invalid; -- cgit v1.3-8-gc7d7 From 3b9b10f9ce61b574c63d71ff3eeec9cf5dbe763f Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Wed, 11 May 2016 09:13:13 -0700 Subject: f2fs: avoid f2fs_bug_on during recovery We don't need to use f2fs_bug_on() to treat with any error case when allocating a block during recovery. Signed-off-by: Jaegeuk Kim --- fs/f2fs/recovery.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/recovery.c b/fs/f2fs/recovery.c index 2b2532903b43..6303b2a38c34 100644 --- a/fs/f2fs/recovery.c +++ b/fs/f2fs/recovery.c @@ -460,8 +460,7 @@ static int do_recover_data(struct f2fs_sb_info *sbi, struct inode *inode, */ if (dest == NEW_ADDR) { truncate_data_blocks_range(&dn, 1); - err = reserve_new_block(&dn); - f2fs_bug_on(sbi, err); + reserve_new_block(&dn); continue; } -- cgit v1.3-8-gc7d7 From ab47036d8f7227361cad7894adee8e66ab6f95b2 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Wed, 11 May 2016 19:48:44 +0800 Subject: f2fs: fix deadlock when flush inline data Below backtrace info was reported by Yunlei He: Call Trace: [] schedule+0x35/0x80 [] rwsem_down_read_failed+0xed/0x130 [] call_rwsem_down_read_failed+0x18/0x [] down_read+0x20/0x30 [] f2fs_evict_inode+0x242/0x3a0 [f2fs] [] evict+0xc7/0x1a0 [] iput+0x196/0x200 [] __dentry_kill+0x179/0x1e0 [] dput+0x199/0x1f0 [] __fput+0x18b/0x220 [] ____fput+0xe/0x10 [] task_work_run+0x77/0x90 [] exit_to_usermode_loop+0x73/0xa2 [] do_syscall_64+0xfa/0x110 [] entry_SYSCALL64_slow_path+0x25/0x25 Call Trace: [] schedule+0x35/0x80 [] __wait_on_freeing_inode+0xa3/0xd0 [] ? autoremove_wake_function+0x40/0x4 [] find_inode_fast+0x7d/0xb0 [] ilookup+0x6a/0xd0 [] sync_node_pages+0x210/0x650 [f2fs] [] ? do_fsync+0x70/0x70 [] block_operations+0x9e/0xf0 [f2fs] [] ? bio_endio+0x55/0x60 [] write_checkpoint+0x92/0xba0 [f2fs] [] ? mempool_free_slab+0x17/0x20 [] ? mempool_free+0x2b/0x80 [] ? do_fsync+0x70/0x70 [] f2fs_sync_fs+0x63/0xd0 [f2fs] [] ? ext4_sync_fs+0xbf/0x190 [] sync_fs_one_sb+0x20/0x30 [] iterate_supers+0xb9/0x110 [] sys_sync+0x55/0x90 [] do_syscall_64+0x69/0x110 [] entry_SYSCALL64_slow_path+0x25/0x25 With following excuting serials, we will set inline_node in inode page after inode was unlinked, result in a deadloop described as below: 1. open file 2. write file 3. unlink file 4. write file 5. close file Thread A Thread B - dput - iput_final - inode->i_state |= I_FREEING - evict - f2fs_evict_inode - f2fs_sync_fs - write_checkpoint - block_operations - f2fs_lock_all (down_write(cp_rwsem)) - f2fs_lock_op (down_read(cp_rwsem)) - sync_node_pages - ilookup - find_inode_fast - __wait_on_freeing_inode (wait on I_FREEING clear) Here, we change to set inline_node flag only for linked inode for fixing. Reported-by: Yunlei He Signed-off-by: Chao Yu Tested-by: Jaegeuk Kim Cc: stable@vger.kernel.org # v4.6 Signed-off-by: Jaegeuk Kim --- fs/f2fs/data.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index b61e2d40cdfb..1013836c5b6b 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -1546,7 +1546,8 @@ restart: if (pos + len <= MAX_INLINE_DATA) { read_inline_data(page, ipage); set_inode_flag(F2FS_I(inode), FI_DATA_EXIST); - set_inline_node(ipage); + if (inode->i_nlink) + set_inline_node(ipage); } else { err = f2fs_convert_inline_page(&dn, page); if (err) -- cgit v1.3-8-gc7d7 From 6426c7ad697df4bddf1d222685e6802e7616feaa Mon Sep 17 00:00:00 2001 From: Qu Wenruo Date: Wed, 11 May 2016 12:53:52 -0700 Subject: btrfs: qgroup: Fix qgroup accounting when creating snapshot Current btrfs qgroup design implies a requirement that after calling btrfs_qgroup_account_extents() there must be a commit root switch. Normally this is OK, as btrfs_qgroup_accounting_extents() is only called inside btrfs_commit_transaction() just be commit_cowonly_roots(). However there is a exception at create_pending_snapshot(), which will call btrfs_qgroup_account_extents() but no any commit root switch. In case of creating a snapshot whose parent root is itself (create a snapshot of fs tree), it will corrupt qgroup by the following trace: (skipped unrelated data) ====== btrfs_qgroup_account_extent: bytenr = 29786112, num_bytes = 16384, nr_old_roots = 0, nr_new_roots = 1 qgroup_update_counters: qgid = 5, cur_old_count = 0, cur_new_count = 1, rfer = 0, excl = 0 qgroup_update_counters: qgid = 5, cur_old_count = 0, cur_new_count = 1, rfer = 16384, excl = 16384 btrfs_qgroup_account_extent: bytenr = 29786112, num_bytes = 16384, nr_old_roots = 0, nr_new_roots = 0 ====== The problem here is in first qgroup_account_extent(), the nr_new_roots of the extent is 1, which means its reference got increased, and qgroup increased its rfer and excl. But at second qgroup_account_extent(), its reference got decreased, but between these two qgroup_account_extent(), there is no switch roots. This leads to the same nr_old_roots, and this extent just got ignored by qgroup, which means this extent is wrongly accounted. Fix it by call commit_cowonly_roots() after qgroup_account_extent() in create_pending_snapshot(), with needed preparation. Mark: I added a check at the top of qgroup_account_snapshot() to skip this code if qgroups are turned off. xfstest btrfs/122 exposes this problem. Signed-off-by: Qu Wenruo Reviewed-by: Josef Bacik Signed-off-by: Mark Fasheh Signed-off-by: David Sterba --- fs/btrfs/transaction.c | 129 ++++++++++++++++++++++++++++++++++++++++--------- 1 file changed, 105 insertions(+), 24 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c index 43885e51b882..d7172d7ced5f 100644 --- a/fs/btrfs/transaction.c +++ b/fs/btrfs/transaction.c @@ -311,10 +311,11 @@ loop: * when the transaction commits */ static int record_root_in_trans(struct btrfs_trans_handle *trans, - struct btrfs_root *root) + struct btrfs_root *root, + int force) { - if (test_bit(BTRFS_ROOT_REF_COWS, &root->state) && - root->last_trans < trans->transid) { + if ((test_bit(BTRFS_ROOT_REF_COWS, &root->state) && + root->last_trans < trans->transid) || force) { WARN_ON(root == root->fs_info->extent_root); WARN_ON(root->commit_root != root->node); @@ -331,7 +332,7 @@ static int record_root_in_trans(struct btrfs_trans_handle *trans, smp_wmb(); spin_lock(&root->fs_info->fs_roots_radix_lock); - if (root->last_trans == trans->transid) { + if (root->last_trans == trans->transid && !force) { spin_unlock(&root->fs_info->fs_roots_radix_lock); return 0; } @@ -402,7 +403,7 @@ int btrfs_record_root_in_trans(struct btrfs_trans_handle *trans, return 0; mutex_lock(&root->fs_info->reloc_mutex); - record_root_in_trans(trans, root); + record_root_in_trans(trans, root, 0); mutex_unlock(&root->fs_info->reloc_mutex); return 0; @@ -1310,6 +1311,92 @@ int btrfs_defrag_root(struct btrfs_root *root) return ret; } +/* + * Do all special snapshot related qgroup dirty hack. + * + * Will do all needed qgroup inherit and dirty hack like switch commit + * roots inside one transaction and write all btree into disk, to make + * qgroup works. + */ +static int qgroup_account_snapshot(struct btrfs_trans_handle *trans, + struct btrfs_root *src, + struct btrfs_root *parent, + struct btrfs_qgroup_inherit *inherit, + u64 dst_objectid) +{ + struct btrfs_fs_info *fs_info = src->fs_info; + int ret; + + /* + * Save some performance in the case that qgroups are not + * enabled. If this check races with the ioctl, rescan will + * kick in anyway. + */ + mutex_lock(&fs_info->qgroup_ioctl_lock); + if (!fs_info->quota_enabled) { + mutex_unlock(&fs_info->qgroup_ioctl_lock); + return 0; + } + mutex_unlock(&fs_info->qgroup_ioctl_lock); + + /* + * We are going to commit transaction, see btrfs_commit_transaction() + * comment for reason locking tree_log_mutex + */ + mutex_lock(&fs_info->tree_log_mutex); + + ret = commit_fs_roots(trans, src); + if (ret) + goto out; + ret = btrfs_qgroup_prepare_account_extents(trans, fs_info); + if (ret < 0) + goto out; + ret = btrfs_qgroup_account_extents(trans, fs_info); + if (ret < 0) + goto out; + + /* Now qgroup are all updated, we can inherit it to new qgroups */ + ret = btrfs_qgroup_inherit(trans, fs_info, + src->root_key.objectid, dst_objectid, + inherit); + if (ret < 0) + goto out; + + /* + * Now we do a simplified commit transaction, which will: + * 1) commit all subvolume and extent tree + * To ensure all subvolume and extent tree have a valid + * commit_root to accounting later insert_dir_item() + * 2) write all btree blocks onto disk + * This is to make sure later btree modification will be cowed + * Or commit_root can be populated and cause wrong qgroup numbers + * In this simplified commit, we don't really care about other trees + * like chunk and root tree, as they won't affect qgroup. + * And we don't write super to avoid half committed status. + */ + ret = commit_cowonly_roots(trans, src); + if (ret) + goto out; + switch_commit_roots(trans->transaction, fs_info); + ret = btrfs_write_and_wait_transaction(trans, src); + if (ret) + btrfs_std_error(fs_info, ret, + "Error while writing out transaction for qgroup"); + +out: + mutex_unlock(&fs_info->tree_log_mutex); + + /* + * Force parent root to be updated, as we recorded it before so its + * last_trans == cur_transid. + * Or it won't be committed again onto disk after later + * insert_dir_item() + */ + if (!ret) + record_root_in_trans(trans, parent, 1); + return ret; +} + /* * new snapshots need to be created at a very specific time in the * transaction commit. This does the actual creation. @@ -1383,7 +1470,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, dentry = pending->dentry; parent_inode = pending->dir; parent_root = BTRFS_I(parent_inode)->root; - record_root_in_trans(trans, parent_root); + record_root_in_trans(trans, parent_root, 0); cur_time = current_fs_time(parent_inode->i_sb); @@ -1420,7 +1507,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, goto fail; } - record_root_in_trans(trans, root); + record_root_in_trans(trans, root, 0); btrfs_set_root_last_snapshot(&root->root_item, trans->transid); memcpy(new_root_item, &root->root_item, sizeof(*new_root_item)); btrfs_check_and_init_root_item(new_root_item); @@ -1516,6 +1603,17 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, goto fail; } + /* + * Do special qgroup accounting for snapshot, as we do some qgroup + * snapshot hack to do fast snapshot. + * To co-operate with that hack, we do hack again. + * Or snapshot will be greatly slowed down by a subtree qgroup rescan + */ + ret = qgroup_account_snapshot(trans, root, parent_root, + pending->inherit, objectid); + if (ret < 0) + goto fail; + ret = btrfs_insert_dir_item(trans, parent_root, dentry->d_name.name, dentry->d_name.len, parent_inode, &key, @@ -1559,23 +1657,6 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, goto fail; } - /* - * account qgroup counters before qgroup_inherit() - */ - ret = btrfs_qgroup_prepare_account_extents(trans, fs_info); - if (ret) - goto fail; - ret = btrfs_qgroup_account_extents(trans, fs_info); - if (ret) - goto fail; - ret = btrfs_qgroup_inherit(trans, fs_info, - root->root_key.objectid, - objectid, pending->inherit); - if (ret) { - btrfs_abort_transaction(trans, root, ret); - goto fail; - } - fail: pending->error = ret; dir_item_existed: -- cgit v1.3-8-gc7d7 From 2c1984f244838477aab4e5882f4479491ae1084a Mon Sep 17 00:00:00 2001 From: David Sterba Date: Thu, 12 May 2016 11:05:03 +0200 Subject: btrfs: build fixup for qgroup_account_snapshot The macro btrfs_std_error got renamed to btrfs_handle_fs_error in an independent branch for the same merge target (4.7). To make the code compilable for bisectability reasons, add a temporary stub. Signed-off-by: David Sterba --- fs/btrfs/transaction.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c index d7172d7ced5f..530081388d77 100644 --- a/fs/btrfs/transaction.c +++ b/fs/btrfs/transaction.c @@ -1311,6 +1311,11 @@ int btrfs_defrag_root(struct btrfs_root *root) return ret; } +/* Bisesctability fixup, remove in 4.8 */ +#ifndef btrfs_std_error +#define btrfs_std_error btrfs_handle_fs_error +#endif + /* * Do all special snapshot related qgroup dirty hack. * -- cgit v1.3-8-gc7d7 From a00839395103d5e2d132a6c4a9680256580ed3d1 Mon Sep 17 00:00:00 2001 From: Omar Sandoval Date: Wed, 11 May 2016 15:16:36 -0700 Subject: coredump: get rid of coredump_params->written cprm->written is redundant with cprm->file->f_pos, so use that instead. Signed-off-by: Omar Sandoval Signed-off-by: Al Viro --- arch/powerpc/platforms/cell/spufs/coredump.c | 5 +++-- fs/binfmt_elf.c | 2 +- fs/binfmt_elf_fdpic.c | 2 +- fs/coredump.c | 8 +++----- include/linux/binfmts.h | 1 - 5 files changed, 8 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/arch/powerpc/platforms/cell/spufs/coredump.c b/arch/powerpc/platforms/cell/spufs/coredump.c index be6212ddbf06..84fb984f29c1 100644 --- a/arch/powerpc/platforms/cell/spufs/coredump.c +++ b/arch/powerpc/platforms/cell/spufs/coredump.c @@ -137,6 +137,7 @@ static int spufs_arch_write_note(struct spu_context *ctx, int i, char *name; char fullname[80], *buf; struct elf_note en; + size_t skip; buf = (void *)get_zeroed_page(GFP_KERNEL); if (!buf) @@ -171,8 +172,8 @@ static int spufs_arch_write_note(struct spu_context *ctx, int i, if (rc < 0) goto out; - if (!dump_skip(cprm, - roundup(cprm->written - total + sz, 4) - cprm->written)) + skip = roundup(cprm->file->f_pos - total + sz, 4) - cprm->file->f_pos; + if (!dump_skip(cprm, skip)) goto Eio; out: free_page((unsigned long)buf); diff --git a/fs/binfmt_elf.c b/fs/binfmt_elf.c index 7d914c67a9d0..f5ce8f4e00b0 100644 --- a/fs/binfmt_elf.c +++ b/fs/binfmt_elf.c @@ -2273,7 +2273,7 @@ static int elf_core_dump(struct coredump_params *cprm) goto end_coredump; /* Align to page */ - if (!dump_skip(cprm, dataoff - cprm->written)) + if (!dump_skip(cprm, dataoff - cprm->file->f_pos)) goto end_coredump; for (i = 0, vma = first_vma(current, gate_vma); vma != NULL; diff --git a/fs/binfmt_elf_fdpic.c b/fs/binfmt_elf_fdpic.c index b1adb92e69de..8cfde2482468 100644 --- a/fs/binfmt_elf_fdpic.c +++ b/fs/binfmt_elf_fdpic.c @@ -1787,7 +1787,7 @@ static int elf_fdpic_core_dump(struct coredump_params *cprm) goto end_coredump; } - if (!dump_skip(cprm, dataoff - cprm->written)) + if (!dump_skip(cprm, dataoff - cprm->file->f_pos)) goto end_coredump; if (!elf_fdpic_dump_segments(cprm)) diff --git a/fs/coredump.c b/fs/coredump.c index 47c32c3bfa1d..9db0c514438e 100644 --- a/fs/coredump.c +++ b/fs/coredump.c @@ -782,7 +782,7 @@ int dump_emit(struct coredump_params *cprm, const void *addr, int nr) struct file *file = cprm->file; loff_t pos = file->f_pos; ssize_t n; - if (cprm->written + nr > cprm->limit) + if (pos + nr > cprm->limit) return 0; while (nr) { if (dump_interrupted()) @@ -791,7 +791,6 @@ int dump_emit(struct coredump_params *cprm, const void *addr, int nr) if (n <= 0) return 0; file->f_pos = pos; - cprm->written += n; nr -= n; } return 1; @@ -803,12 +802,11 @@ int dump_skip(struct coredump_params *cprm, size_t nr) static char zeroes[PAGE_SIZE]; struct file *file = cprm->file; if (file->f_op->llseek && file->f_op->llseek != no_llseek) { - if (cprm->written + nr > cprm->limit) + if (file->f_pos + nr > cprm->limit) return 0; if (dump_interrupted() || file->f_op->llseek(file, nr, SEEK_CUR) < 0) return 0; - cprm->written += nr; return 1; } else { while (nr > PAGE_SIZE) { @@ -823,7 +821,7 @@ EXPORT_SYMBOL(dump_skip); int dump_align(struct coredump_params *cprm, int align) { - unsigned mod = cprm->written & (align - 1); + unsigned mod = cprm->file->f_pos & (align - 1); if (align & (align - 1)) return 0; return mod ? dump_skip(cprm, align - mod) : 1; diff --git a/include/linux/binfmts.h b/include/linux/binfmts.h index 576e4639ca60..39c6d6e1234e 100644 --- a/include/linux/binfmts.h +++ b/include/linux/binfmts.h @@ -64,7 +64,6 @@ struct coredump_params { struct file *file; unsigned long limit; unsigned long mm_flags; - loff_t written; }; /* -- cgit v1.3-8-gc7d7 From 2c4cb04300fa160e9d78335c74184c4e66a56437 Mon Sep 17 00:00:00 2001 From: Omar Sandoval Date: Wed, 11 May 2016 15:16:37 -0700 Subject: coredump: only charge written data against RLIMIT_CORE Commit 9b56d54380ad ("dump_skip(): dump_seek() replacement taking coredump_params") introduced a regression with regard to RLIMIT_CORE. Previously, when a core dump was sparse, only the data that was actually written out would count against the limit. Now, the sparse ranges are also included, which leads to truncated core dumps when the actual disk usage is still well below the limit. Restore the old behavior by only counting what gets emitted and ignoring what gets skipped. Signed-off-by: Omar Sandoval Signed-off-by: Al Viro --- fs/coredump.c | 5 ++--- include/linux/binfmts.h | 1 + 2 files changed, 3 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/coredump.c b/fs/coredump.c index 9db0c514438e..492c2db25dc9 100644 --- a/fs/coredump.c +++ b/fs/coredump.c @@ -782,7 +782,7 @@ int dump_emit(struct coredump_params *cprm, const void *addr, int nr) struct file *file = cprm->file; loff_t pos = file->f_pos; ssize_t n; - if (pos + nr > cprm->limit) + if (cprm->written + nr > cprm->limit) return 0; while (nr) { if (dump_interrupted()) @@ -791,6 +791,7 @@ int dump_emit(struct coredump_params *cprm, const void *addr, int nr) if (n <= 0) return 0; file->f_pos = pos; + cprm->written += n; nr -= n; } return 1; @@ -802,8 +803,6 @@ int dump_skip(struct coredump_params *cprm, size_t nr) static char zeroes[PAGE_SIZE]; struct file *file = cprm->file; if (file->f_op->llseek && file->f_op->llseek != no_llseek) { - if (file->f_pos + nr > cprm->limit) - return 0; if (dump_interrupted() || file->f_op->llseek(file, nr, SEEK_CUR) < 0) return 0; diff --git a/include/linux/binfmts.h b/include/linux/binfmts.h index 39c6d6e1234e..576e4639ca60 100644 --- a/include/linux/binfmts.h +++ b/include/linux/binfmts.h @@ -64,6 +64,7 @@ struct coredump_params { struct file *file; unsigned long limit; unsigned long mm_flags; + loff_t written; }; /* -- cgit v1.3-8-gc7d7 From 657ed1aa4898c8304500e0d13f240d5a67e8be5f Mon Sep 17 00:00:00 2001 From: Filipe Manana Date: Wed, 6 Apr 2016 17:11:56 +0100 Subject: Btrfs: fix for incorrect directory entries after fsync log replay MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit If we move a directory to a new parent and later log that parent and don't explicitly log the old parent, when we replay the log we can end up with entries for the moved directory in both the old and new parent directories. Besides being ilegal to have directories with multiple hard links in linux, it also resulted in the leaving the inode item with a link count of 1. A similar issue also happens if we move a regular file - after the log tree is replayed the file has a link in both the old and new parent directories, when it should be only at the new directory. Sample reproducer: $ mkfs.btrfs -f /dev/sdc $ mount /dev/sdc /mnt $ mkdir /mnt/x $ mkdir /mnt/y $ touch /mnt/x/foo $ mkdir /mnt/y/z $ sync $ ln /mnt/x/foo /mnt/x/bar $ mv /mnt/y/z /mnt/x/z < power fail > $ mount /dev/sdc /mnt $ ls -1Ri /mnt /mnt: 257 x 258 y /mnt/x: 259 bar 259 foo 260 z /mnt/x/z: /mnt/y: 260 z /mnt/y/z: $ umount /dev/sdc $ btrfs check /dev/sdc Checking filesystem on /dev/sdc UUID: a67e2c4a-a4b4-4fdc-b015-9d9af1e344be checking extents checking free space cache checking fs roots root 5 inode 260 errors 2000, link count wrong unresolved ref dir 257 index 4 namelen 1 name z filetype 2 errors 0 unresolved ref dir 258 index 2 namelen 1 name z filetype 2 errors 0 (...) Attempting to remove the directory becomes impossible: $ mount /dev/sdc /mnt $ rmdir /mnt/y/z $ ls -lh /mnt/y ls: cannot access /mnt/y/z: No such file or directory total 0 d????????? ? ? ? ? ? z $ rmdir /mnt/x/z rmdir: failed to remove ‘/mnt/x/z’: Stale file handle $ ls -lh /mnt/x ls: cannot access /mnt/x/z: Stale file handle total 0 -rw-r--r-- 2 root root 0 Apr 6 18:06 bar -rw-r--r-- 2 root root 0 Apr 6 18:06 foo d????????? ? ? ? ? ? z So make sure that on rename we set the last_unlink_trans value for our inode, even if it's a directory, to the value of the current transaction's ID and that if the new parent directory is logged that we fallback to a transaction commit. A test case for fstests is being submitted as well. Signed-off-by: Filipe Manana --- fs/btrfs/tree-log.c | 13 ++++++++----- 1 file changed, 8 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c index 517d0ccb351e..4709932c62fb 100644 --- a/fs/btrfs/tree-log.c +++ b/fs/btrfs/tree-log.c @@ -5278,11 +5278,16 @@ static int btrfs_log_all_parents(struct btrfs_trans_handle *trans, if (IS_ERR(dir_inode)) continue; + if (ctx) + ctx->log_new_dentries = false; ret = btrfs_log_inode(trans, root, dir_inode, LOG_INODE_ALL, 0, LLONG_MAX, ctx); if (!ret && btrfs_must_commit_transaction(trans, dir_inode)) ret = 1; + if (!ret && ctx && ctx->log_new_dentries) + ret = log_new_dir_dentries(trans, root, + dir_inode, ctx); iput(dir_inode); if (ret) goto out; @@ -5652,11 +5657,9 @@ void btrfs_record_unlink_dir(struct btrfs_trans_handle *trans, * into the file. When the file is logged we check it and * don't log the parents if the file is fully on disk. */ - if (S_ISREG(inode->i_mode)) { - mutex_lock(&BTRFS_I(inode)->log_mutex); - BTRFS_I(inode)->last_unlink_trans = trans->transid; - mutex_unlock(&BTRFS_I(inode)->log_mutex); - } + mutex_lock(&BTRFS_I(inode)->log_mutex); + BTRFS_I(inode)->last_unlink_trans = trans->transid; + mutex_unlock(&BTRFS_I(inode)->log_mutex); /* * if this directory was already logged any new -- cgit v1.3-8-gc7d7 From 3f9749f6e9edcf8ec569fb542efc3be35e06e84a Mon Sep 17 00:00:00 2001 From: Filipe Manana Date: Mon, 25 Apr 2016 04:45:02 +0100 Subject: Btrfs: fix empty symlink after creating symlink and fsync parent dir If we create a symlink, fsync its parent directory, crash/power fail and mount the filesystem, we end up with an empty symlink, which not only is useless it's also not allowed in linux (the man page symlink(2) is well explicit about that). So we just need to make sure to fully log an inode if it's a symlink, to ensure its inline extent gets logged, ensuring the same behaviour as ext3, ext4, xfs, reiserfs, f2fs, nilfs2, etc. Example reproducer: $ mkfs.btrfs -f /dev/sdb $ mount /dev/sdb /mnt $ mkdir /mnt/testdir $ sync $ ln -s /mnt/foo /mnt/testdir/bar $ xfs_io -c fsync /mnt/testdir $ mount /dev/sdb /mnt $ readlink /mnt/testdir/bar A test case for fstests follows soon. Signed-off-by: Filipe Manana --- fs/btrfs/tree-log.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c index 4709932c62fb..a24a0ba523d6 100644 --- a/fs/btrfs/tree-log.c +++ b/fs/btrfs/tree-log.c @@ -5158,7 +5158,7 @@ process_leaf: } ctx->log_new_dentries = false; - if (type == BTRFS_FT_DIR) + if (type == BTRFS_FT_DIR || type == BTRFS_FT_SYMLINK) log_mode = LOG_INODE_ALL; btrfs_release_path(path); ret = btrfs_log_inode(trans, root, di_inode, -- cgit v1.3-8-gc7d7 From 578def7c50f236432ba140d35bb7ca4ef0a1b20b Mon Sep 17 00:00:00 2001 From: Filipe Manana Date: Tue, 26 Apr 2016 15:36:38 +0100 Subject: Btrfs: don't wait for unrelated IO to finish before relocation Before the relocation process of a block group starts, it sets the block group to readonly mode, then flushes all delalloc writes and then finally it waits for all ordered extents to complete. This last step includes waiting for ordered extents destinated at extents allocated in other block groups, making us waste unecessary time. So improve this by waiting only for ordered extents that fall into the block group's range. Signed-off-by: Filipe Manana Reviewed-by: Josef Bacik Reviewed-by: Liu Bo --- fs/btrfs/dev-replace.c | 4 ++-- fs/btrfs/extent-tree.c | 11 +++++++---- fs/btrfs/ioctl.c | 2 +- fs/btrfs/ordered-data.c | 26 +++++++++++++++++++------- fs/btrfs/ordered-data.h | 6 ++++-- fs/btrfs/relocation.c | 4 +++- fs/btrfs/super.c | 2 +- fs/btrfs/transaction.c | 2 +- 8 files changed, 38 insertions(+), 19 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/dev-replace.c b/fs/btrfs/dev-replace.c index 26bcb487f958..3371f9e546d9 100644 --- a/fs/btrfs/dev-replace.c +++ b/fs/btrfs/dev-replace.c @@ -403,7 +403,7 @@ int btrfs_dev_replace_start(struct btrfs_root *root, if (ret) btrfs_err(root->fs_info, "kobj add dev failed %d\n", ret); - btrfs_wait_ordered_roots(root->fs_info, -1); + btrfs_wait_ordered_roots(root->fs_info, -1, 0, (u64)-1); /* force writing the updated state information to disk */ trans = btrfs_start_transaction(root, 0); @@ -495,7 +495,7 @@ static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info, mutex_unlock(&dev_replace->lock_finishing_cancel_unmount); return ret; } - btrfs_wait_ordered_roots(root->fs_info, -1); + btrfs_wait_ordered_roots(root->fs_info, -1, 0, (u64)-1); trans = btrfs_start_transaction(root, 0); if (IS_ERR(trans)) { diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 84e060eb0de8..251452a2b72c 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -4141,7 +4141,7 @@ commit_trans: if (need_commit > 0) { btrfs_start_delalloc_roots(fs_info, 0, -1); - btrfs_wait_ordered_roots(fs_info, -1); + btrfs_wait_ordered_roots(fs_info, -1, 0, (u64)-1); } trans = btrfs_join_transaction(root); @@ -4583,7 +4583,8 @@ static void btrfs_writeback_inodes_sb_nr(struct btrfs_root *root, */ btrfs_start_delalloc_roots(root->fs_info, 0, nr_items); if (!current->journal_info) - btrfs_wait_ordered_roots(root->fs_info, nr_items); + btrfs_wait_ordered_roots(root->fs_info, nr_items, + 0, (u64)-1); } } @@ -4632,7 +4633,8 @@ static void shrink_delalloc(struct btrfs_root *root, u64 to_reclaim, u64 orig, if (trans) return; if (wait_ordered) - btrfs_wait_ordered_roots(root->fs_info, items); + btrfs_wait_ordered_roots(root->fs_info, items, + 0, (u64)-1); return; } @@ -4671,7 +4673,8 @@ skip_async: loops++; if (wait_ordered && !trans) { - btrfs_wait_ordered_roots(root->fs_info, items); + btrfs_wait_ordered_roots(root->fs_info, items, + 0, (u64)-1); } else { time_left = schedule_timeout_killable(1); if (time_left) diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index 5a23806ae418..697cc336bd1c 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -681,7 +681,7 @@ static int create_snapshot(struct btrfs_root *root, struct inode *dir, if (ret) goto dec_and_free; - btrfs_wait_ordered_extents(root, -1); + btrfs_wait_ordered_extents(root, -1, 0, (u64)-1); btrfs_init_block_rsv(&pending_snapshot->block_rsv, BTRFS_BLOCK_RSV_TEMP); diff --git a/fs/btrfs/ordered-data.c b/fs/btrfs/ordered-data.c index 0de7da5a610d..559170464d7c 100644 --- a/fs/btrfs/ordered-data.c +++ b/fs/btrfs/ordered-data.c @@ -661,14 +661,15 @@ static void btrfs_run_ordered_extent_work(struct btrfs_work *work) * wait for all the ordered extents in a root. This is done when balancing * space between drives. */ -int btrfs_wait_ordered_extents(struct btrfs_root *root, int nr) +int btrfs_wait_ordered_extents(struct btrfs_root *root, int nr, + const u64 range_start, const u64 range_len) { - struct list_head splice, works; + LIST_HEAD(splice); + LIST_HEAD(skipped); + LIST_HEAD(works); struct btrfs_ordered_extent *ordered, *next; int count = 0; - - INIT_LIST_HEAD(&splice); - INIT_LIST_HEAD(&works); + const u64 range_end = range_start + range_len; mutex_lock(&root->ordered_extent_mutex); spin_lock(&root->ordered_extent_lock); @@ -676,6 +677,14 @@ int btrfs_wait_ordered_extents(struct btrfs_root *root, int nr) while (!list_empty(&splice) && nr) { ordered = list_first_entry(&splice, struct btrfs_ordered_extent, root_extent_list); + + if (range_end <= ordered->start || + ordered->start + ordered->disk_len <= range_start) { + list_move_tail(&ordered->root_extent_list, &skipped); + cond_resched_lock(&root->ordered_extent_lock); + continue; + } + list_move_tail(&ordered->root_extent_list, &root->ordered_extents); atomic_inc(&ordered->refs); @@ -694,6 +703,7 @@ int btrfs_wait_ordered_extents(struct btrfs_root *root, int nr) nr--; count++; } + list_splice_tail(&skipped, &root->ordered_extents); list_splice_tail(&splice, &root->ordered_extents); spin_unlock(&root->ordered_extent_lock); @@ -708,7 +718,8 @@ int btrfs_wait_ordered_extents(struct btrfs_root *root, int nr) return count; } -void btrfs_wait_ordered_roots(struct btrfs_fs_info *fs_info, int nr) +void btrfs_wait_ordered_roots(struct btrfs_fs_info *fs_info, int nr, + const u64 range_start, const u64 range_len) { struct btrfs_root *root; struct list_head splice; @@ -728,7 +739,8 @@ void btrfs_wait_ordered_roots(struct btrfs_fs_info *fs_info, int nr) &fs_info->ordered_roots); spin_unlock(&fs_info->ordered_root_lock); - done = btrfs_wait_ordered_extents(root, nr); + done = btrfs_wait_ordered_extents(root, nr, + range_start, range_len); btrfs_put_fs_root(root); spin_lock(&fs_info->ordered_root_lock); diff --git a/fs/btrfs/ordered-data.h b/fs/btrfs/ordered-data.h index 23c96059cef2..8ef12623d65c 100644 --- a/fs/btrfs/ordered-data.h +++ b/fs/btrfs/ordered-data.h @@ -197,8 +197,10 @@ int btrfs_ordered_update_i_size(struct inode *inode, u64 offset, struct btrfs_ordered_extent *ordered); int btrfs_find_ordered_sum(struct inode *inode, u64 offset, u64 disk_bytenr, u32 *sum, int len); -int btrfs_wait_ordered_extents(struct btrfs_root *root, int nr); -void btrfs_wait_ordered_roots(struct btrfs_fs_info *fs_info, int nr); +int btrfs_wait_ordered_extents(struct btrfs_root *root, int nr, + const u64 range_start, const u64 range_len); +void btrfs_wait_ordered_roots(struct btrfs_fs_info *fs_info, int nr, + const u64 range_start, const u64 range_len); void btrfs_get_logged_extents(struct inode *inode, struct list_head *logged_list, const loff_t start, diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c index 08ef890deca6..30f77ed60133 100644 --- a/fs/btrfs/relocation.c +++ b/fs/btrfs/relocation.c @@ -4259,7 +4259,9 @@ int btrfs_relocate_block_group(struct btrfs_root *extent_root, u64 group_start) err = ret; goto out; } - btrfs_wait_ordered_roots(fs_info, -1); + btrfs_wait_ordered_roots(fs_info, -1, + rc->block_group->key.objectid, + rc->block_group->key.offset); while (1) { mutex_lock(&fs_info->cleaner_mutex); diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index 00b8f37cc306..89d134794d47 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -1160,7 +1160,7 @@ int btrfs_sync_fs(struct super_block *sb, int wait) return 0; } - btrfs_wait_ordered_roots(fs_info, -1); + btrfs_wait_ordered_roots(fs_info, -1, 0, (u64)-1); trans = btrfs_attach_transaction_barrier(root); if (IS_ERR(trans)) { diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c index 43885e51b882..f0bb54a77314 100644 --- a/fs/btrfs/transaction.c +++ b/fs/btrfs/transaction.c @@ -1821,7 +1821,7 @@ static inline int btrfs_start_delalloc_flush(struct btrfs_fs_info *fs_info) static inline void btrfs_wait_delalloc_flush(struct btrfs_fs_info *fs_info) { if (btrfs_test_opt(fs_info->tree_root, FLUSHONCOMMIT)) - btrfs_wait_ordered_roots(fs_info, -1); + btrfs_wait_ordered_roots(fs_info, -1, 0, (u64)-1); } static inline void -- cgit v1.3-8-gc7d7 From 9cfa3e34e20e6798a671236000d9e97c8aa5d318 Mon Sep 17 00:00:00 2001 From: Filipe Manana Date: Tue, 26 Apr 2016 15:39:32 +0100 Subject: Btrfs: don't do unnecessary delalloc flushes when relocating Before we start the actual relocation process of a block group, we do calls to flush delalloc of all inodes and then wait for ordered extents to complete. However we do these flush calls just to make sure we don't race with concurrent tasks that have actually already started to run delalloc and have allocated an extent from the block group we want to relocate, right before we set it to readonly mode, but have not yet created the respective ordered extents. The flush calls make us wait for such concurrent tasks because they end up calling filemap_fdatawrite_range() (through btrfs_start_delalloc_roots() -> __start_delalloc_inodes() -> btrfs_alloc_delalloc_work() -> btrfs_run_delalloc_work()) which ends up serializing us with those tasks due to attempts to lock the same pages (and the delalloc flush procedure calls the allocator and creates the ordered extents before unlocking the pages). These flushing calls not only make us waste time (cpu, IO) but also reduce the chances of writing larger extents (applications might be writing to contiguous ranges and we flush before they finish dirtying the whole ranges). So make sure we don't flush delalloc and just wait for concurrent tasks that have already started flushing delalloc and have allocated an extent from the block group we are about to relocate. This change also ends up fixing a race with direct IO writes that makes relocation not wait for direct IO ordered extents. This race is illustrated by the following diagram: CPU 1 CPU 2 btrfs_relocate_block_group(bg X) starts direct IO write, target inode currently has no ordered extents ongoing nor dirty pages (delalloc regions), therefore the root for our inode is not in the list fs_info->ordered_roots btrfs_direct_IO() __blockdev_direct_IO() btrfs_get_blocks_direct() btrfs_lock_extent_direct() locks range in the io tree btrfs_new_extent_direct() btrfs_reserve_extent() --> extent allocated from bg X btrfs_inc_block_group_ro(bg X) btrfs_start_delalloc_roots() __start_delalloc_inodes() --> does nothing, no dealloc ranges in the inode's io tree so the inode's root is not in the list fs_info->delalloc_roots btrfs_wait_ordered_roots() --> does not find the inode's root in the list fs_info->ordered_roots --> ends up not waiting for the direct IO write started by the task at CPU 2 relocate_block_group(rc->stage == MOVE_DATA_EXTENTS) prepare_to_relocate() btrfs_commit_transaction() iterates the extent tree, using its commit root and moves extents into new locations btrfs_add_ordered_extent_dio() --> now a ordered extent is created and added to the list root->ordered_extents and the root added to the list fs_info->ordered_roots --> this is too late and the task at CPU 1 already started the relocation btrfs_commit_transaction() btrfs_finish_ordered_io() btrfs_alloc_reserved_file_extent() --> adds delayed data reference for the extent allocated from bg X relocate_block_group(rc->stage == UPDATE_DATA_PTRS) prepare_to_relocate() btrfs_commit_transaction() --> delayed refs are run, so an extent item for the allocated extent from bg X is added to extent tree --> commit roots are switched, so the next scan in the extent tree will see the extent item sees the extent in the extent tree When this happens the relocation produces the following warning when it finishes: [ 7260.832836] ------------[ cut here ]------------ [ 7260.834653] WARNING: CPU: 5 PID: 6765 at fs/btrfs/relocation.c:4318 btrfs_relocate_block_group+0x245/0x2a1 [btrfs]() [ 7260.838268] Modules linked in: btrfs crc32c_generic xor ppdev raid6_pq psmouse sg acpi_cpufreq evdev i2c_piix4 tpm_tis serio_raw tpm i2c_core pcspkr parport_pc [ 7260.850935] CPU: 5 PID: 6765 Comm: btrfs Not tainted 4.5.0-rc6-btrfs-next-28+ #1 [ 7260.852998] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS by qemu-project.org 04/01/2014 [ 7260.852998] 0000000000000000 ffff88020bf57bc0 ffffffff812648b3 0000000000000000 [ 7260.852998] 0000000000000009 ffff88020bf57bf8 ffffffff81051608 ffffffffa03c1b2d [ 7260.852998] ffff8800b2bbb800 0000000000000000 ffff8800b17bcc58 ffff8800399dd000 [ 7260.852998] Call Trace: [ 7260.852998] [] dump_stack+0x67/0x90 [ 7260.852998] [] warn_slowpath_common+0x99/0xb2 [ 7260.852998] [] ? btrfs_relocate_block_group+0x245/0x2a1 [btrfs] [ 7260.852998] [] warn_slowpath_null+0x1a/0x1c [ 7260.852998] [] btrfs_relocate_block_group+0x245/0x2a1 [btrfs] [ 7260.852998] [] btrfs_relocate_chunk.isra.29+0x66/0xdb [btrfs] [ 7260.852998] [] btrfs_balance+0xde1/0xe4e [btrfs] [ 7260.852998] [] ? debug_smp_processor_id+0x17/0x19 [ 7260.852998] [] btrfs_ioctl_balance+0x255/0x2d3 [btrfs] [ 7260.852998] [] btrfs_ioctl+0x11e0/0x1dff [btrfs] [ 7260.852998] [] ? handle_mm_fault+0x443/0xd63 [ 7260.852998] [] ? _raw_spin_unlock+0x31/0x44 [ 7260.852998] [] ? arch_local_irq_save+0x9/0xc [ 7260.852998] [] vfs_ioctl+0x18/0x34 [ 7260.852998] [] do_vfs_ioctl+0x550/0x5be [ 7260.852998] [] ? __fget_light+0x4d/0x71 [ 7260.852998] [] SyS_ioctl+0x57/0x79 [ 7260.852998] [] entry_SYSCALL_64_fastpath+0x12/0x6b [ 7260.893268] ---[ end trace eb7803b24ebab8ad ]--- This is because at the end of the first stage, in relocate_block_group(), we commit the current transaction, which makes delayed refs run, the commit roots are switched and so the second stage will find the extent item that the ordered extent added to the delayed refs. But this extent was not moved (ordered extent completed after first stage finished), so at the end of the relocation our block group item still has a positive used bytes counter, triggering a warning at the end of btrfs_relocate_block_group(). Later on when trying to read the extent contents from disk we hit a BUG_ON() due to the inability to map a block with a logical address that belongs to the block group we relocated and is no longer valid, resulting in the following trace: [ 7344.885290] BTRFS critical (device sdi): unable to find logical 12845056 len 4096 [ 7344.887518] ------------[ cut here ]------------ [ 7344.888431] kernel BUG at fs/btrfs/inode.c:1833! [ 7344.888431] invalid opcode: 0000 [#1] PREEMPT SMP DEBUG_PAGEALLOC [ 7344.888431] Modules linked in: btrfs crc32c_generic xor ppdev raid6_pq psmouse sg acpi_cpufreq evdev i2c_piix4 tpm_tis serio_raw tpm i2c_core pcspkr parport_pc [ 7344.888431] CPU: 0 PID: 6831 Comm: od Tainted: G W 4.5.0-rc6-btrfs-next-28+ #1 [ 7344.888431] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS by qemu-project.org 04/01/2014 [ 7344.888431] task: ffff880215818600 ti: ffff880204684000 task.ti: ffff880204684000 [ 7344.888431] RIP: 0010:[] [] btrfs_merge_bio_hook+0x54/0x6b [btrfs] [ 7344.888431] RSP: 0018:ffff8802046878f0 EFLAGS: 00010282 [ 7344.888431] RAX: 00000000ffffffea RBX: 0000000000001000 RCX: 0000000000000001 [ 7344.888431] RDX: ffff88023ec0f950 RSI: ffffffff8183b638 RDI: 00000000ffffffff [ 7344.888431] RBP: ffff880204687908 R08: 0000000000000001 R09: 0000000000000000 [ 7344.888431] R10: ffff880204687770 R11: ffffffff82f2d52d R12: 0000000000001000 [ 7344.888431] R13: ffff88021afbfee8 R14: 0000000000006208 R15: ffff88006cd199b0 [ 7344.888431] FS: 00007f1f9e1d6700(0000) GS:ffff88023ec00000(0000) knlGS:0000000000000000 [ 7344.888431] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 7344.888431] CR2: 00007f1f9dc8cb60 CR3: 000000023e3b6000 CR4: 00000000000006f0 [ 7344.888431] Stack: [ 7344.888431] 0000000000001000 0000000000001000 ffff880204687b98 ffff880204687950 [ 7344.888431] ffffffffa0395c8f ffffea0004d64d48 0000000000000000 0000000000001000 [ 7344.888431] ffffea0004d64d48 0000000000001000 0000000000000000 0000000000000000 [ 7344.888431] Call Trace: [ 7344.888431] [] submit_extent_page+0xf5/0x16f [btrfs] [ 7344.888431] [] __do_readpage+0x4a0/0x4f1 [btrfs] [ 7344.888431] [] ? btrfs_create_repair_bio+0xcb/0xcb [btrfs] [ 7344.888431] [] ? btrfs_writepage_start_hook+0xbc/0xbc [btrfs] [ 7344.888431] [] ? trace_hardirqs_on+0xd/0xf [ 7344.888431] [] __do_contiguous_readpages.constprop.26+0xc2/0xe4 [btrfs] [ 7344.888431] [] ? btrfs_writepage_start_hook+0xbc/0xbc [btrfs] [ 7344.888431] [] __extent_readpages.constprop.25+0xed/0x100 [btrfs] [ 7344.888431] [] ? lru_cache_add+0xe/0x10 [ 7344.888431] [] extent_readpages+0x160/0x1aa [btrfs] [ 7344.888431] [] ? btrfs_writepage_start_hook+0xbc/0xbc [btrfs] [ 7344.888431] [] ? alloc_pages_current+0xa9/0xcd [ 7344.888431] [] btrfs_readpages+0x1f/0x21 [btrfs] [ 7344.888431] [] __do_page_cache_readahead+0x168/0x1fc [ 7344.888431] [] ondemand_readahead+0x1f6/0x207 [ 7344.888431] [] ? ondemand_readahead+0x1f6/0x207 [ 7344.888431] [] ? pagecache_get_page+0x2b/0x154 [ 7344.888431] [] page_cache_sync_readahead+0x3d/0x3f [ 7344.888431] [] generic_file_read_iter+0x197/0x4e1 [ 7344.888431] [] __vfs_read+0x79/0x9d [ 7344.888431] [] vfs_read+0x8f/0xd2 [ 7344.888431] [] SyS_read+0x50/0x7e [ 7344.888431] [] entry_SYSCALL_64_fastpath+0x12/0x6b [ 7344.888431] Code: 8d 4d e8 45 31 c9 45 31 c0 48 8b 00 48 c1 e2 09 48 8b 80 80 fc ff ff 4c 89 65 e8 48 8b b8 f0 01 00 00 e8 1d 42 02 00 85 c0 79 02 <0f> 0b 4c 0 [ 7344.888431] RIP [] btrfs_merge_bio_hook+0x54/0x6b [btrfs] [ 7344.888431] RSP [ 7344.970544] ---[ end trace eb7803b24ebab8ae ]--- Signed-off-by: Filipe Manana Reviewed-by: Josef Bacik Reviewed-by: Liu Bo --- fs/btrfs/ctree.h | 14 ++++++++++++ fs/btrfs/extent-tree.c | 58 ++++++++++++++++++++++++++++++++++++++++++++++++-- fs/btrfs/inode.c | 8 +++++++ fs/btrfs/relocation.c | 6 +----- 4 files changed, 79 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 84a6a5b3384a..90e70e21e479 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -1408,6 +1408,17 @@ struct btrfs_block_group_cache { struct btrfs_io_ctl io_ctl; + /* + * Incremented when doing extent allocations and holding a read lock + * on the space_info's groups_sem semaphore. + * Decremented when an ordered extent that represents an IO against this + * block group's range is created (after it's added to its inode's + * root's list of ordered extents) or immediately after the allocation + * if it's a metadata extent or fallocate extent (for these cases we + * don't create ordered extents). + */ + atomic_t reservations; + /* Lock for free space tree operations. */ struct mutex free_space_lock; @@ -3499,6 +3510,9 @@ int btrfs_should_throttle_delayed_refs(struct btrfs_trans_handle *trans, struct btrfs_root *root); int btrfs_check_space_for_delayed_refs(struct btrfs_trans_handle *trans, struct btrfs_root *root); +void btrfs_dec_block_group_reservations(struct btrfs_fs_info *fs_info, + const u64 start); +void btrfs_wait_block_group_reservations(struct btrfs_block_group_cache *bg); void btrfs_put_block_group(struct btrfs_block_group_cache *cache); int btrfs_run_delayed_refs(struct btrfs_trans_handle *trans, struct btrfs_root *root, unsigned long count); diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 251452a2b72c..09aad7b447f5 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -6175,6 +6175,57 @@ int btrfs_exclude_logged_extents(struct btrfs_root *log, return 0; } +static void +btrfs_inc_block_group_reservations(struct btrfs_block_group_cache *bg) +{ + atomic_inc(&bg->reservations); +} + +void btrfs_dec_block_group_reservations(struct btrfs_fs_info *fs_info, + const u64 start) +{ + struct btrfs_block_group_cache *bg; + + bg = btrfs_lookup_block_group(fs_info, start); + ASSERT(bg); + if (atomic_dec_and_test(&bg->reservations)) + wake_up_atomic_t(&bg->reservations); + btrfs_put_block_group(bg); +} + +static int btrfs_wait_bg_reservations_atomic_t(atomic_t *a) +{ + schedule(); + return 0; +} + +void btrfs_wait_block_group_reservations(struct btrfs_block_group_cache *bg) +{ + struct btrfs_space_info *space_info = bg->space_info; + + ASSERT(bg->ro); + + if (!(bg->flags & BTRFS_BLOCK_GROUP_DATA)) + return; + + /* + * Our block group is read only but before we set it to read only, + * some task might have had allocated an extent from it already, but it + * has not yet created a respective ordered extent (and added it to a + * root's list of ordered extents). + * Therefore wait for any task currently allocating extents, since the + * block group's reservations counter is incremented while a read lock + * on the groups' semaphore is held and decremented after releasing + * the read access on that semaphore and creating the ordered extent. + */ + down_write(&space_info->groups_sem); + up_write(&space_info->groups_sem); + + wait_on_atomic_t(&bg->reservations, + btrfs_wait_bg_reservations_atomic_t, + TASK_UNINTERRUPTIBLE); +} + /** * btrfs_update_reserved_bytes - update the block_group and space info counters * @cache: The cache we are manipulating @@ -7434,6 +7485,7 @@ checks: btrfs_add_free_space(block_group, offset, num_bytes); goto loop; } + btrfs_inc_block_group_reservations(block_group); /* we are all good, lets return */ ins->objectid = search_start; @@ -7615,8 +7667,10 @@ again: WARN_ON(num_bytes < root->sectorsize); ret = find_free_extent(root, num_bytes, empty_size, hint_byte, ins, flags, delalloc); - - if (ret == -ENOSPC) { + if (!ret && !is_data) { + btrfs_dec_block_group_reservations(root->fs_info, + ins->objectid); + } else if (ret == -ENOSPC) { if (!final_tried && ins->offset) { num_bytes = min(num_bytes >> 1, ins->offset); num_bytes = round_down(num_bytes, root->sectorsize); diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 2aaba58b4856..3991d8a74f61 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -824,6 +824,7 @@ retry: async_extent->ram_size - 1, 0); goto out_free_reserve; } + btrfs_dec_block_group_reservations(root->fs_info, ins.objectid); /* * clear dirty, set writeback and unlock the pages. @@ -861,6 +862,7 @@ retry: } return; out_free_reserve: + btrfs_dec_block_group_reservations(root->fs_info, ins.objectid); btrfs_free_reserved_extent(root, ins.objectid, ins.offset, 1); out_free: extent_clear_unlock_delalloc(inode, async_extent->start, @@ -1038,6 +1040,8 @@ static noinline int cow_file_range(struct inode *inode, goto out_drop_extent_cache; } + btrfs_dec_block_group_reservations(root->fs_info, ins.objectid); + if (disk_num_bytes < cur_alloc_size) break; @@ -1066,6 +1070,7 @@ out: out_drop_extent_cache: btrfs_drop_extent_cache(inode, start, start + ram_size - 1, 0); out_reserve: + btrfs_dec_block_group_reservations(root->fs_info, ins.objectid); btrfs_free_reserved_extent(root, ins.objectid, ins.offset, 1); out_unlock: extent_clear_unlock_delalloc(inode, start, end, locked_page, @@ -7162,6 +7167,8 @@ static struct extent_map *btrfs_new_extent_direct(struct inode *inode, return ERR_PTR(ret); } + btrfs_dec_block_group_reservations(root->fs_info, ins.objectid); + em = create_pinned_em(inode, start, ins.offset, start, ins.objectid, ins.offset, ins.offset, ins.offset, 0); if (IS_ERR(em)) { @@ -9942,6 +9949,7 @@ static int __btrfs_prealloc_file_range(struct inode *inode, int mode, btrfs_end_transaction(trans, root); break; } + btrfs_dec_block_group_reservations(root->fs_info, ins.objectid); last_alloc = ins.offset; ret = insert_reserved_file_extent(trans, inode, diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c index 30f77ed60133..e78f8e44bd9a 100644 --- a/fs/btrfs/relocation.c +++ b/fs/btrfs/relocation.c @@ -4254,11 +4254,7 @@ int btrfs_relocate_block_group(struct btrfs_root *extent_root, u64 group_start) btrfs_info(extent_root->fs_info, "relocating block group %llu flags %llu", rc->block_group->key.objectid, rc->block_group->flags); - ret = btrfs_start_delalloc_roots(fs_info, 0, -1); - if (ret < 0) { - err = ret; - goto out; - } + btrfs_wait_block_group_reservations(rc->block_group); btrfs_wait_ordered_roots(fs_info, -1, rc->block_group->key.objectid, rc->block_group->key.offset); -- cgit v1.3-8-gc7d7 From 3dc9e8f76720fbbd9c56a11775932733fe13d214 Mon Sep 17 00:00:00 2001 From: Filipe Manana Date: Fri, 29 Apr 2016 11:34:22 +0100 Subject: Btrfs: unpin log if rename operation fails If rename operations fail at some point after we pinned the log, we end up aborting the current transaction but never unpin the log, which leaves concurrent tasks that are trying to sync the log (as part of an fsync request from user space) blocked forever and preventing the filesystem from being unmountable. Fix this by safely unpinning the log. Signed-off-by: Filipe Manana --- fs/btrfs/inode.c | 28 +++++++++++++++++++++++++++- 1 file changed, 27 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 3991d8a74f61..4594a263a5f8 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -9406,6 +9406,7 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, u64 root_objectid; int ret; u64 old_ino = btrfs_ino(old_inode); + bool log_pinned = false; if (btrfs_ino(new_dir) == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID) return -EPERM; @@ -9493,6 +9494,7 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, * we unlink the name but before we add the new name back in. */ btrfs_pin_log_trans(root); + log_pinned = true; } inode_inc_iversion(old_dir); @@ -9559,12 +9561,36 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, if (old_inode->i_nlink == 1) BTRFS_I(old_inode)->dir_index = index; - if (old_ino != BTRFS_FIRST_FREE_OBJECTID) { + if (log_pinned) { struct dentry *parent = new_dentry->d_parent; + btrfs_log_new_name(trans, old_inode, old_dir, parent); btrfs_end_log_trans(root); + log_pinned = false; } out_fail: + /* + * If we have pinned the log and an error happened, we unpin tasks + * trying to sync the log and force them to fallback to a transaction + * commit if the log currently contains any of the inodes involved in + * this rename operation (to ensure we do not persist a log with an + * inconsistent state for any of these inodes or leading to any + * inconsistencies when replayed). If the transaction was aborted, the + * abortion reason is propagated to userspace when attempting to commit + * the transaction. If the log does not contain any of these inodes, we + * allow the tasks to sync it. + */ + if (ret && log_pinned) { + if (btrfs_inode_in_log(old_dir, root->fs_info->generation) || + btrfs_inode_in_log(new_dir, root->fs_info->generation) || + btrfs_inode_in_log(old_inode, root->fs_info->generation) || + (new_inode && + btrfs_inode_in_log(new_inode, root->fs_info->generation))) + btrfs_set_log_full_commit(root->fs_info, trans); + + btrfs_end_log_trans(root); + log_pinned = false; + } btrfs_end_transaction(trans, root); out_notrans: if (old_ino == BTRFS_FIRST_FREE_OBJECTID) -- cgit v1.3-8-gc7d7 From c4aba9545430ed0f842d0833072f990a42da90f0 Mon Sep 17 00:00:00 2001 From: Filipe Manana Date: Fri, 29 Apr 2016 13:14:42 +0100 Subject: Btrfs: pin log earlier when renaming We were pinning the log right after the first step in the rename operation (inserting inode ref for the new name in the destination directory) instead of doing it before. This behaviour was introduced in 2009 for some reason that was not mentioned neither on the changelog nor any comment, with the drawback of a small time window where concurrent log writers can end up logging the new inode reference for the inode we are renaming while the rename operation is in progress (so that we can end up with a log containing both the new and old references). As of today there's no reason to not pin the log before that first step anymore, so just fix this. Signed-off-by: Filipe Manana --- fs/btrfs/inode.c | 11 ++--------- 1 file changed, 2 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 4594a263a5f8..422833e70d00 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -9479,6 +9479,8 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, /* force full log commit if subvolume involved. */ btrfs_set_log_full_commit(root->fs_info, trans); } else { + btrfs_pin_log_trans(root); + log_pinned = true; ret = btrfs_insert_inode_ref(trans, dest, new_dentry->d_name.name, new_dentry->d_name.len, @@ -9486,15 +9488,6 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, btrfs_ino(new_dir), index); if (ret) goto out_fail; - /* - * this is an ugly little race, but the rename is required - * to make sure that if we crash, the inode is either at the - * old name or the new one. pinning the log transaction lets - * us make sure we don't allow a log commit to come in after - * we unlink the name but before we add the new name back in. - */ - btrfs_pin_log_trans(root); - log_pinned = true; } inode_inc_iversion(old_dir); -- cgit v1.3-8-gc7d7 From cdd1fedf8261cd7a73c0596298902ff4f0f04492 Mon Sep 17 00:00:00 2001 From: Dan Fuhry Date: Thu, 17 Mar 2016 15:23:38 +0100 Subject: btrfs: add support for RENAME_EXCHANGE and RENAME_WHITEOUT Two new flags, RENAME_EXCHANGE and RENAME_WHITEOUT, provide for new behavior in the renameat2() syscall. This behavior is primarily used by overlayfs. This patch adds support for these flags to btrfs, enabling it to be used as a fully functional upper layer for overlayfs. RENAME_EXCHANGE support was written by Davide Italiano originally submitted on 2 April 2015. Signed-off-by: Davide Italiano Signed-off-by: Dan Fuhry [ remove unlikely ] Signed-off-by: David Sterba Signed-off-by: Filipe Manana --- fs/btrfs/inode.c | 264 +++++++++++++++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 257 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 422833e70d00..452cfefbf4b8 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -9394,8 +9394,244 @@ static int btrfs_getattr(struct vfsmount *mnt, return 0; } +static int btrfs_rename_exchange(struct inode *old_dir, + struct dentry *old_dentry, + struct inode *new_dir, + struct dentry *new_dentry) +{ + struct btrfs_trans_handle *trans; + struct btrfs_root *root = BTRFS_I(old_dir)->root; + struct btrfs_root *dest = BTRFS_I(new_dir)->root; + struct inode *new_inode = new_dentry->d_inode; + struct inode *old_inode = old_dentry->d_inode; + struct timespec ctime = CURRENT_TIME; + struct dentry *parent; + u64 old_ino = btrfs_ino(old_inode); + u64 new_ino = btrfs_ino(new_inode); + u64 old_idx = 0; + u64 new_idx = 0; + u64 root_objectid; + int ret; + + /* we only allow rename subvolume link between subvolumes */ + if (old_ino != BTRFS_FIRST_FREE_OBJECTID && root != dest) + return -EXDEV; + + /* close the race window with snapshot create/destroy ioctl */ + if (old_ino == BTRFS_FIRST_FREE_OBJECTID) + down_read(&root->fs_info->subvol_sem); + if (new_ino == BTRFS_FIRST_FREE_OBJECTID) + down_read(&dest->fs_info->subvol_sem); + + /* + * We want to reserve the absolute worst case amount of items. So if + * both inodes are subvols and we need to unlink them then that would + * require 4 item modifications, but if they are both normal inodes it + * would require 5 item modifications, so we'll assume their normal + * inodes. So 5 * 2 is 10, plus 2 for the new links, so 12 total items + * should cover the worst case number of items we'll modify. + */ + trans = btrfs_start_transaction(root, 12); + if (IS_ERR(trans)) { + ret = PTR_ERR(trans); + goto out_notrans; + } + + /* + * We need to find a free sequence number both in the source and + * in the destination directory for the exchange. + */ + ret = btrfs_set_inode_index(new_dir, &old_idx); + if (ret) + goto out_fail; + ret = btrfs_set_inode_index(old_dir, &new_idx); + if (ret) + goto out_fail; + + BTRFS_I(old_inode)->dir_index = 0ULL; + BTRFS_I(new_inode)->dir_index = 0ULL; + + /* Reference for the source. */ + if (old_ino == BTRFS_FIRST_FREE_OBJECTID) { + /* force full log commit if subvolume involved. */ + btrfs_set_log_full_commit(root->fs_info, trans); + } else { + ret = btrfs_insert_inode_ref(trans, dest, + new_dentry->d_name.name, + new_dentry->d_name.len, + old_ino, + btrfs_ino(new_dir), old_idx); + if (ret) + goto out_fail; + btrfs_pin_log_trans(root); + } + + /* And now for the dest. */ + if (new_ino == BTRFS_FIRST_FREE_OBJECTID) { + /* force full log commit if subvolume involved. */ + btrfs_set_log_full_commit(dest->fs_info, trans); + } else { + ret = btrfs_insert_inode_ref(trans, root, + old_dentry->d_name.name, + old_dentry->d_name.len, + new_ino, + btrfs_ino(old_dir), new_idx); + if (ret) + goto out_fail; + btrfs_pin_log_trans(dest); + } + + /* Update inode version and ctime/mtime. */ + inode_inc_iversion(old_dir); + inode_inc_iversion(new_dir); + inode_inc_iversion(old_inode); + inode_inc_iversion(new_inode); + old_dir->i_ctime = old_dir->i_mtime = ctime; + new_dir->i_ctime = new_dir->i_mtime = ctime; + old_inode->i_ctime = ctime; + new_inode->i_ctime = ctime; + + if (old_dentry->d_parent != new_dentry->d_parent) { + btrfs_record_unlink_dir(trans, old_dir, old_inode, 1); + btrfs_record_unlink_dir(trans, new_dir, new_inode, 1); + } + + /* src is a subvolume */ + if (old_ino == BTRFS_FIRST_FREE_OBJECTID) { + root_objectid = BTRFS_I(old_inode)->root->root_key.objectid; + ret = btrfs_unlink_subvol(trans, root, old_dir, + root_objectid, + old_dentry->d_name.name, + old_dentry->d_name.len); + } else { /* src is an inode */ + ret = __btrfs_unlink_inode(trans, root, old_dir, + old_dentry->d_inode, + old_dentry->d_name.name, + old_dentry->d_name.len); + if (!ret) + ret = btrfs_update_inode(trans, root, old_inode); + } + if (ret) { + btrfs_abort_transaction(trans, root, ret); + goto out_fail; + } + + /* dest is a subvolume */ + if (new_ino == BTRFS_FIRST_FREE_OBJECTID) { + root_objectid = BTRFS_I(new_inode)->root->root_key.objectid; + ret = btrfs_unlink_subvol(trans, dest, new_dir, + root_objectid, + new_dentry->d_name.name, + new_dentry->d_name.len); + } else { /* dest is an inode */ + ret = __btrfs_unlink_inode(trans, dest, new_dir, + new_dentry->d_inode, + new_dentry->d_name.name, + new_dentry->d_name.len); + if (!ret) + ret = btrfs_update_inode(trans, dest, new_inode); + } + if (ret) { + btrfs_abort_transaction(trans, root, ret); + goto out_fail; + } + + ret = btrfs_add_link(trans, new_dir, old_inode, + new_dentry->d_name.name, + new_dentry->d_name.len, 0, old_idx); + if (ret) { + btrfs_abort_transaction(trans, root, ret); + goto out_fail; + } + + ret = btrfs_add_link(trans, old_dir, new_inode, + old_dentry->d_name.name, + old_dentry->d_name.len, 0, new_idx); + if (ret) { + btrfs_abort_transaction(trans, root, ret); + goto out_fail; + } + + if (old_inode->i_nlink == 1) + BTRFS_I(old_inode)->dir_index = old_idx; + if (new_inode->i_nlink == 1) + BTRFS_I(new_inode)->dir_index = new_idx; + + if (old_ino != BTRFS_FIRST_FREE_OBJECTID) { + parent = new_dentry->d_parent; + btrfs_log_new_name(trans, old_inode, old_dir, parent); + btrfs_end_log_trans(root); + } + if (new_ino != BTRFS_FIRST_FREE_OBJECTID) { + parent = old_dentry->d_parent; + btrfs_log_new_name(trans, new_inode, new_dir, parent); + btrfs_end_log_trans(dest); + } +out_fail: + ret = btrfs_end_transaction(trans, root); +out_notrans: + if (new_ino == BTRFS_FIRST_FREE_OBJECTID) + up_read(&dest->fs_info->subvol_sem); + if (old_ino == BTRFS_FIRST_FREE_OBJECTID) + up_read(&root->fs_info->subvol_sem); + + return ret; +} + +static int btrfs_whiteout_for_rename(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct inode *dir, + struct dentry *dentry) +{ + int ret; + struct inode *inode; + u64 objectid; + u64 index; + + ret = btrfs_find_free_ino(root, &objectid); + if (ret) + return ret; + + inode = btrfs_new_inode(trans, root, dir, + dentry->d_name.name, + dentry->d_name.len, + btrfs_ino(dir), + objectid, + S_IFCHR | WHITEOUT_MODE, + &index); + + if (IS_ERR(inode)) { + ret = PTR_ERR(inode); + return ret; + } + + inode->i_op = &btrfs_special_inode_operations; + init_special_inode(inode, inode->i_mode, + WHITEOUT_DEV); + + ret = btrfs_init_inode_security(trans, inode, dir, + &dentry->d_name); + if (ret) + return ret; + + ret = btrfs_add_nondir(trans, dir, dentry, + inode, 0, index); + if (ret) + return ret; + + ret = btrfs_update_inode(trans, root, inode); + if (ret) + return ret; + + unlock_new_inode(inode); + iput(inode); + + return 0; +} + static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, - struct inode *new_dir, struct dentry *new_dentry) + struct inode *new_dir, struct dentry *new_dentry, + unsigned int flags) { struct btrfs_trans_handle *trans; struct btrfs_root *root = BTRFS_I(old_dir)->root; @@ -9457,15 +9693,15 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, * We want to reserve the absolute worst case amount of items. So if * both inodes are subvols and we need to unlink them then that would * require 4 item modifications, but if they are both normal inodes it - * would require 5 item modifications, so we'll assume their normal + * would require 5 item modifications, so we'll assume they are normal * inodes. So 5 * 2 is 10, plus 1 for the new link, so 11 total items * should cover the worst case number of items we'll modify. */ trans = btrfs_start_transaction(root, 11); if (IS_ERR(trans)) { - ret = PTR_ERR(trans); - goto out_notrans; - } + ret = PTR_ERR(trans); + goto out_notrans; + } if (dest != root) btrfs_record_root_in_trans(trans, dest); @@ -9561,6 +9797,16 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, btrfs_end_log_trans(root); log_pinned = false; } + + if (flags & RENAME_WHITEOUT) { + ret = btrfs_whiteout_for_rename(trans, root, old_dir, + old_dentry); + + if (ret) { + btrfs_abort_transaction(trans, root, ret); + goto out_fail; + } + } out_fail: /* * If we have pinned the log and an error happened, we unpin tasks @@ -9596,10 +9842,14 @@ static int btrfs_rename2(struct inode *old_dir, struct dentry *old_dentry, struct inode *new_dir, struct dentry *new_dentry, unsigned int flags) { - if (flags & ~RENAME_NOREPLACE) + if (flags & ~(RENAME_NOREPLACE | RENAME_EXCHANGE | RENAME_WHITEOUT)) return -EINVAL; - return btrfs_rename(old_dir, old_dentry, new_dir, new_dentry); + if (flags & RENAME_EXCHANGE) + return btrfs_rename_exchange(old_dir, old_dentry, new_dir, + new_dentry); + + return btrfs_rename(old_dir, old_dentry, new_dir, new_dentry, flags); } static void btrfs_run_delalloc_work(struct btrfs_work *work) -- cgit v1.3-8-gc7d7 From c990161888f387db136856337c237aa8d5003292 Mon Sep 17 00:00:00 2001 From: Filipe Manana Date: Thu, 5 May 2016 01:41:57 +0100 Subject: Btrfs: fix inode leak on failure to setup whiteout inode in rename If we failed to fully setup the whiteout inode during a rename operation with the whiteout flag, we ended up leaking the inode, not decrementing its link count nor removing all its items from the fs/subvol tree. Signed-off-by: Filipe Manana --- fs/btrfs/inode.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 452cfefbf4b8..98c119b02916 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -9612,21 +9612,21 @@ static int btrfs_whiteout_for_rename(struct btrfs_trans_handle *trans, ret = btrfs_init_inode_security(trans, inode, dir, &dentry->d_name); if (ret) - return ret; + goto out; ret = btrfs_add_nondir(trans, dir, dentry, inode, 0, index); if (ret) - return ret; + goto out; ret = btrfs_update_inode(trans, root, inode); - if (ret) - return ret; - +out: unlock_new_inode(inode); + if (ret) + inode_dec_link_count(inode); iput(inode); - return 0; + return ret; } static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, -- cgit v1.3-8-gc7d7 From 86e8aa0e772caba5f0e0471d5f836b2b997dcb3e Mon Sep 17 00:00:00 2001 From: Filipe Manana Date: Thu, 5 May 2016 02:02:27 +0100 Subject: Btrfs: unpin logs if rename exchange operation fails If rename exchange operations fail at some point after we pinned any of the logs, we end up aborting the current transaction but never unpin the logs, which leaves concurrent tasks that are trying to sync the logs (as part of an fsync request from user space) blocked forever and preventing the filesystem from being unmountable. Fix this by safely unpinning the log. Signed-off-by: Filipe Manana --- fs/btrfs/inode.c | 38 ++++++++++++++++++++++++++++++++++++-- 1 file changed, 36 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 98c119b02916..c92d9b83bb38 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -9412,6 +9412,8 @@ static int btrfs_rename_exchange(struct inode *old_dir, u64 new_idx = 0; u64 root_objectid; int ret; + bool root_log_pinned = false; + bool dest_log_pinned = false; /* we only allow rename subvolume link between subvolumes */ if (old_ino != BTRFS_FIRST_FREE_OBJECTID && root != dest) @@ -9464,6 +9466,7 @@ static int btrfs_rename_exchange(struct inode *old_dir, if (ret) goto out_fail; btrfs_pin_log_trans(root); + root_log_pinned = true; } /* And now for the dest. */ @@ -9479,6 +9482,7 @@ static int btrfs_rename_exchange(struct inode *old_dir, if (ret) goto out_fail; btrfs_pin_log_trans(dest); + dest_log_pinned = true; } /* Update inode version and ctime/mtime. */ @@ -9557,17 +9561,47 @@ static int btrfs_rename_exchange(struct inode *old_dir, if (new_inode->i_nlink == 1) BTRFS_I(new_inode)->dir_index = new_idx; - if (old_ino != BTRFS_FIRST_FREE_OBJECTID) { + if (root_log_pinned) { parent = new_dentry->d_parent; btrfs_log_new_name(trans, old_inode, old_dir, parent); btrfs_end_log_trans(root); + root_log_pinned = false; } - if (new_ino != BTRFS_FIRST_FREE_OBJECTID) { + if (dest_log_pinned) { parent = old_dentry->d_parent; btrfs_log_new_name(trans, new_inode, new_dir, parent); btrfs_end_log_trans(dest); + dest_log_pinned = false; } out_fail: + /* + * If we have pinned a log and an error happened, we unpin tasks + * trying to sync the log and force them to fallback to a transaction + * commit if the log currently contains any of the inodes involved in + * this rename operation (to ensure we do not persist a log with an + * inconsistent state for any of these inodes or leading to any + * inconsistencies when replayed). If the transaction was aborted, the + * abortion reason is propagated to userspace when attempting to commit + * the transaction. If the log does not contain any of these inodes, we + * allow the tasks to sync it. + */ + if (ret && (root_log_pinned || dest_log_pinned)) { + if (btrfs_inode_in_log(old_dir, root->fs_info->generation) || + btrfs_inode_in_log(new_dir, root->fs_info->generation) || + btrfs_inode_in_log(old_inode, root->fs_info->generation) || + (new_inode && + btrfs_inode_in_log(new_inode, root->fs_info->generation))) + btrfs_set_log_full_commit(root->fs_info, trans); + + if (root_log_pinned) { + btrfs_end_log_trans(root); + root_log_pinned = false; + } + if (dest_log_pinned) { + btrfs_end_log_trans(dest); + dest_log_pinned = false; + } + } ret = btrfs_end_transaction(trans, root); out_notrans: if (new_ino == BTRFS_FIRST_FREE_OBJECTID) -- cgit v1.3-8-gc7d7 From 376e5a57bf7f1466031a957d04bf8b8f6801ee6d Mon Sep 17 00:00:00 2001 From: Filipe Manana Date: Thu, 5 May 2016 02:08:56 +0100 Subject: Btrfs: pin logs earlier when doing a rename exchange operation The btrfs_rename_exchange() started as a copy-paste from btrfs_rename(), which had a race fixed by my previous patch titled "Btrfs: pin log earlier when renaming", and so it suffers from the same problem. We pin the logs of the affected roots after we insert the new inode references, leaving a time window where concurrent tasks logging the inodes can end up logging both the new and old references, resulting in log trees that when replayed can turn the metadata into inconsistent states. This behaviour was added to btrfs_rename() in 2009 without any explanation about why not pinning the logs earlier, just leaving a comment about the posibility for the race. As of today it's perfectly safe and sane to pin the logs before we start doing any of the steps involved in the rename operation. Signed-off-by: Filipe Manana --- fs/btrfs/inode.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index c92d9b83bb38..a7db3ed6ac88 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -9458,6 +9458,8 @@ static int btrfs_rename_exchange(struct inode *old_dir, /* force full log commit if subvolume involved. */ btrfs_set_log_full_commit(root->fs_info, trans); } else { + btrfs_pin_log_trans(root); + root_log_pinned = true; ret = btrfs_insert_inode_ref(trans, dest, new_dentry->d_name.name, new_dentry->d_name.len, @@ -9465,8 +9467,6 @@ static int btrfs_rename_exchange(struct inode *old_dir, btrfs_ino(new_dir), old_idx); if (ret) goto out_fail; - btrfs_pin_log_trans(root); - root_log_pinned = true; } /* And now for the dest. */ @@ -9474,6 +9474,8 @@ static int btrfs_rename_exchange(struct inode *old_dir, /* force full log commit if subvolume involved. */ btrfs_set_log_full_commit(dest->fs_info, trans); } else { + btrfs_pin_log_trans(dest); + dest_log_pinned = true; ret = btrfs_insert_inode_ref(trans, root, old_dentry->d_name.name, old_dentry->d_name.len, @@ -9481,8 +9483,6 @@ static int btrfs_rename_exchange(struct inode *old_dir, btrfs_ino(old_dir), new_idx); if (ret) goto out_fail; - btrfs_pin_log_trans(dest); - dest_log_pinned = true; } /* Update inode version and ctime/mtime. */ -- cgit v1.3-8-gc7d7 From 5062af35c3c6e49110ab1ec99295339259298a3d Mon Sep 17 00:00:00 2001 From: Filipe Manana Date: Thu, 5 May 2016 10:26:26 +0100 Subject: Btrfs: fix number of transaction units for renames with whiteout When we do a rename with the whiteout flag, we need to create the whiteout inode, which in the worst case requires 5 transaction units (1 inode item, 1 inode ref, 2 dir items and 1 xattr if selinux is enabled). So bump the number of transaction units from 11 to 16 if the whiteout flag is set. Signed-off-by: Filipe Manana --- fs/btrfs/inode.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index a7db3ed6ac88..0921d2b07f1d 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -9668,6 +9668,7 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, unsigned int flags) { struct btrfs_trans_handle *trans; + unsigned int trans_num_items; struct btrfs_root *root = BTRFS_I(old_dir)->root; struct btrfs_root *dest = BTRFS_I(new_dir)->root; struct inode *new_inode = d_inode(new_dentry); @@ -9730,8 +9731,14 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, * would require 5 item modifications, so we'll assume they are normal * inodes. So 5 * 2 is 10, plus 1 for the new link, so 11 total items * should cover the worst case number of items we'll modify. + * If our rename has the whiteout flag, we need more 5 units for the + * new inode (1 inode item, 1 inode ref, 2 dir items and 1 xattr item + * when selinux is enabled). */ - trans = btrfs_start_transaction(root, 11); + trans_num_items = 11; + if (flags & RENAME_WHITEOUT) + trans_num_items += 5; + trans = btrfs_start_transaction(root, trans_num_items); if (IS_ERR(trans)) { ret = PTR_ERR(trans); goto out_notrans; -- cgit v1.3-8-gc7d7 From 0b901916a00bc7b14ee83cc8e41c3b0d561a8f22 Mon Sep 17 00:00:00 2001 From: Filipe Manana Date: Mon, 9 May 2016 13:15:27 +0100 Subject: Btrfs: fix race between fsync and direct IO writes for prealloc extents When we do a direct IO write against a preallocated extent (fallocate) that does not go beyond the i_size of the inode, we do the write operation without holding the inode's i_mutex (an optimization that landed in commit 38851cc19adb ("Btrfs: implement unlocked dio write")). This allows for a very tiny time window where a race can happen with a concurrent fsync using the fast code path, as the direct IO write path creates first a new extent map (no longer flagged as a prealloc extent) and then it creates the ordered extent, while the fast fsync path first collects ordered extents and then it collects extent maps. This allows for the possibility of the fast fsync path to collect the new extent map without collecting the new ordered extent, and therefore logging an extent item based on the extent map without waiting for the ordered extent to be created and complete. This can result in a situation where after a log replay we end up with an extent not marked anymore as prealloc but it was only partially written (or not written at all), exposing random, stale or garbage data corresponding to the unwritten pages and without any checksums in the csum tree covering the extent's range. This is an extension of what was done in commit de0ee0edb21f ("Btrfs: fix race between fsync and lockless direct IO writes"). So fix this by creating first the ordered extent and then the extent map, so that this way if the fast fsync patch collects the new extent map it also collects the corresponding ordered extent. Signed-off-by: Filipe Manana Reviewed-by: Josef Bacik --- fs/btrfs/inode.c | 43 +++++++++++++++++++++++++++++++++++++------ 1 file changed, 37 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 0921d2b07f1d..45d0dafbbf40 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -7658,6 +7658,25 @@ static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock, if (can_nocow_extent(inode, start, &len, &orig_start, &orig_block_len, &ram_bytes) == 1) { + + /* + * Create the ordered extent before the extent map. This + * is to avoid races with the fast fsync path because it + * collects ordered extents into a local list and then + * collects all the new extent maps, so we must create + * the ordered extent first and make sure the fast fsync + * path collects any new ordered extents after + * collecting new extent maps as well. The fsync path + * simply can not rely on inode_dio_wait() because it + * causes deadlock with AIO. + */ + ret = btrfs_add_ordered_extent_dio(inode, start, + block_start, len, len, type); + if (ret) { + free_extent_map(em); + goto unlock_err; + } + if (type == BTRFS_ORDERED_PREALLOC) { free_extent_map(em); em = create_pinned_em(inode, start, len, @@ -7666,17 +7685,29 @@ static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock, orig_block_len, ram_bytes, type); if (IS_ERR(em)) { + struct btrfs_ordered_extent *oe; + ret = PTR_ERR(em); + oe = btrfs_lookup_ordered_extent(inode, + start); + ASSERT(oe); + if (WARN_ON(!oe)) + goto unlock_err; + set_bit(BTRFS_ORDERED_IOERR, + &oe->flags); + set_bit(BTRFS_ORDERED_IO_DONE, + &oe->flags); + btrfs_remove_ordered_extent(inode, oe); + /* + * Once for our lookup and once for the + * ordered extents tree. + */ + btrfs_put_ordered_extent(oe); + btrfs_put_ordered_extent(oe); goto unlock_err; } } - ret = btrfs_add_ordered_extent_dio(inode, start, - block_start, len, len, type); - if (ret) { - free_extent_map(em); - goto unlock_err; - } goto unlock; } } -- cgit v1.3-8-gc7d7 From f78c436c3931e7df713688028f2b4faf72bf9f2a Mon Sep 17 00:00:00 2001 From: Filipe Manana Date: Mon, 9 May 2016 13:15:41 +0100 Subject: Btrfs: fix race between block group relocation and nocow writes Relocation of a block group waits for all existing tasks flushing dellaloc, starting direct IO writes and any ordered extents before starting the relocation process. However for direct IO writes that end up doing nocow (inode either has the flag nodatacow set or the write is against a prealloc extent) we have a short time window that allows for a race that makes relocation proceed without waiting for the direct IO write to complete first, resulting in data loss after the relocation finishes. This is illustrated by the following diagram: CPU 1 CPU 2 btrfs_relocate_block_group(bg X) direct IO write starts against an extent in block group X using nocow mode (inode has the nodatacow flag or the write is for a prealloc extent) btrfs_direct_IO() btrfs_get_blocks_direct() --> can_nocow_extent() returns 1 btrfs_inc_block_group_ro(bg X) --> turns block group into RO mode btrfs_wait_ordered_roots() --> returns and does not know about the DIO write happening at CPU 2 (the task there has not created yet an ordered extent) relocate_block_group(bg X) --> rc->stage == MOVE_DATA_EXTENTS find_next_extent() --> returns extent that the DIO write is going to write to relocate_data_extent() relocate_file_extent_cluster() --> reads the extent from disk into pages belonging to the relocation inode and dirties them --> creates DIO ordered extent btrfs_submit_direct() --> submits bio against a location on disk obtained from an extent map before the relocation started btrfs_wait_ordered_range() --> writes all the pages read before to disk (belonging to the relocation inode) relocation finishes bio completes and wrote new data to the old location of the block group So fix this by tracking the number of nocow writers for a block group and make sure relocation waits for that number to go down to 0 before starting to move the extents. The same race can also happen with buffered writes in nocow mode since the patch I recently made titled "Btrfs: don't do unnecessary delalloc flushes when relocating", because we are no longer flushing all delalloc which served as a synchonization mechanism (due to page locking) and ensured the ordered extents for nocow buffered writes were created before we called btrfs_wait_ordered_roots(). The race with direct IO writes in nocow mode existed before that patch (no pages are locked or used during direct IO) and that fixed only races with direct IO writes that do cow. Signed-off-by: Filipe Manana Reviewed-by: Josef Bacik --- fs/btrfs/ctree.h | 13 +++++++++++++ fs/btrfs/extent-tree.c | 53 ++++++++++++++++++++++++++++++++++++++++++++++++++ fs/btrfs/inode.c | 15 +++++++++++++- fs/btrfs/relocation.c | 1 + 4 files changed, 81 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 90e70e21e479..7ae758685c7b 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -1419,6 +1419,16 @@ struct btrfs_block_group_cache { */ atomic_t reservations; + /* + * Incremented while holding the spinlock *lock* by a task checking if + * it can perform a nocow write (incremented if the value for the *ro* + * field is 0). Decremented by such tasks once they create an ordered + * extent or before that if some error happens before reaching that step. + * This is to prevent races between block group relocation and nocow + * writes through direct IO. + */ + atomic_t nocow_writers; + /* Lock for free space tree operations. */ struct mutex free_space_lock; @@ -3513,6 +3523,9 @@ int btrfs_check_space_for_delayed_refs(struct btrfs_trans_handle *trans, void btrfs_dec_block_group_reservations(struct btrfs_fs_info *fs_info, const u64 start); void btrfs_wait_block_group_reservations(struct btrfs_block_group_cache *bg); +bool btrfs_inc_nocow_writers(struct btrfs_fs_info *fs_info, u64 bytenr); +void btrfs_dec_nocow_writers(struct btrfs_fs_info *fs_info, u64 bytenr); +void btrfs_wait_nocow_writers(struct btrfs_block_group_cache *bg); void btrfs_put_block_group(struct btrfs_block_group_cache *cache); int btrfs_run_delayed_refs(struct btrfs_trans_handle *trans, struct btrfs_root *root, unsigned long count); diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 09aad7b447f5..dcf89bfa990d 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -3824,6 +3824,59 @@ int btrfs_extent_readonly(struct btrfs_root *root, u64 bytenr) return readonly; } +bool btrfs_inc_nocow_writers(struct btrfs_fs_info *fs_info, u64 bytenr) +{ + struct btrfs_block_group_cache *bg; + bool ret = true; + + bg = btrfs_lookup_block_group(fs_info, bytenr); + if (!bg) + return false; + + spin_lock(&bg->lock); + if (bg->ro) + ret = false; + else + atomic_inc(&bg->nocow_writers); + spin_unlock(&bg->lock); + + /* no put on block group, done by btrfs_dec_nocow_writers */ + if (!ret) + btrfs_put_block_group(bg); + + return ret; + +} + +void btrfs_dec_nocow_writers(struct btrfs_fs_info *fs_info, u64 bytenr) +{ + struct btrfs_block_group_cache *bg; + + bg = btrfs_lookup_block_group(fs_info, bytenr); + ASSERT(bg); + if (atomic_dec_and_test(&bg->nocow_writers)) + wake_up_atomic_t(&bg->nocow_writers); + /* + * Once for our lookup and once for the lookup done by a previous call + * to btrfs_inc_nocow_writers() + */ + btrfs_put_block_group(bg); + btrfs_put_block_group(bg); +} + +static int btrfs_wait_nocow_writers_atomic_t(atomic_t *a) +{ + schedule(); + return 0; +} + +void btrfs_wait_nocow_writers(struct btrfs_block_group_cache *bg) +{ + wait_on_atomic_t(&bg->nocow_writers, + btrfs_wait_nocow_writers_atomic_t, + TASK_UNINTERRUPTIBLE); +} + static const char *alloc_name(u64 flags) { switch (flags) { diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 45d0dafbbf40..ee9be4199e7c 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -1382,6 +1382,9 @@ next_slot: */ if (csum_exist_in_range(root, disk_bytenr, num_bytes)) goto out_check; + if (!btrfs_inc_nocow_writers(root->fs_info, + disk_bytenr)) + goto out_check; nocow = 1; } else if (extent_type == BTRFS_FILE_EXTENT_INLINE) { extent_end = found_key.offset + @@ -1396,6 +1399,9 @@ out_check: path->slots[0]++; if (!nolock && nocow) btrfs_end_write_no_snapshoting(root); + if (nocow) + btrfs_dec_nocow_writers(root->fs_info, + disk_bytenr); goto next_slot; } if (!nocow) { @@ -1416,6 +1422,9 @@ out_check: if (ret) { if (!nolock && nocow) btrfs_end_write_no_snapshoting(root); + if (nocow) + btrfs_dec_nocow_writers(root->fs_info, + disk_bytenr); goto error; } cow_start = (u64)-1; @@ -1458,6 +1467,8 @@ out_check: ret = btrfs_add_ordered_extent(inode, cur_offset, disk_bytenr, num_bytes, num_bytes, type); + if (nocow) + btrfs_dec_nocow_writers(root->fs_info, disk_bytenr); BUG_ON(ret); /* -ENOMEM */ if (root->root_key.objectid == @@ -7657,7 +7668,8 @@ static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock, block_start = em->block_start + (start - em->start); if (can_nocow_extent(inode, start, &len, &orig_start, - &orig_block_len, &ram_bytes) == 1) { + &orig_block_len, &ram_bytes) == 1 && + btrfs_inc_nocow_writers(root->fs_info, block_start)) { /* * Create the ordered extent before the extent map. This @@ -7672,6 +7684,7 @@ static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock, */ ret = btrfs_add_ordered_extent_dio(inode, start, block_start, len, len, type); + btrfs_dec_nocow_writers(root->fs_info, block_start); if (ret) { free_extent_map(em); goto unlock_err; diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c index e78f8e44bd9a..054d9a80e77e 100644 --- a/fs/btrfs/relocation.c +++ b/fs/btrfs/relocation.c @@ -4255,6 +4255,7 @@ int btrfs_relocate_block_group(struct btrfs_root *extent_root, u64 group_start) rc->block_group->key.objectid, rc->block_group->flags); btrfs_wait_block_group_reservations(rc->block_group); + btrfs_wait_nocow_writers(rc->block_group); btrfs_wait_ordered_roots(fs_info, -1, rc->block_group->key.objectid, rc->block_group->key.offset); -- cgit v1.3-8-gc7d7 From 5f9a8a51d8b95505d8de8b7191ae2ed8c504d4af Mon Sep 17 00:00:00 2001 From: Filipe Manana Date: Thu, 12 May 2016 13:53:36 +0100 Subject: Btrfs: add semaphore to synchronize direct IO writes with fsync Due to the optimization of lockless direct IO writes (the inode's i_mutex is not held) introduced in commit 38851cc19adb ("Btrfs: implement unlocked dio write"), we started having races between such writes with concurrent fsync operations that use the fast fsync path. These races were addressed in the patches titled "Btrfs: fix race between fsync and lockless direct IO writes" and "Btrfs: fix race between fsync and direct IO writes for prealloc extents". The races happened because the direct IO path, like every other write path, does create extent maps followed by the corresponding ordered extents while the fast fsync path collected first ordered extents and then it collected extent maps. This made it possible to log file extent items (based on the collected extent maps) without waiting for the corresponding ordered extents to complete (get their IO done). The two fixes mentioned before added a solution that consists of making the direct IO path create first the ordered extents and then the extent maps, while the fsync path attempts to collect any new ordered extents once it collects the extent maps. This was simple and did not require adding any synchonization primitive to any data structure (struct btrfs_inode for example) but it makes things more fragile for future development endeavours and adds an exceptional approach compared to the other write paths. This change adds a read-write semaphore to the btrfs inode structure and makes the direct IO path create the extent maps and the ordered extents while holding read access on that semaphore, while the fast fsync path collects extent maps and ordered extents while holding write access on that semaphore. The logic for direct IO write path is encapsulated in a new helper function that is used both for cow and nocow direct IO writes. Signed-off-by: Filipe Manana Reviewed-by: Josef Bacik --- fs/btrfs/btrfs_inode.h | 10 ++++ fs/btrfs/inode.c | 134 +++++++++++++++++++------------------------------ fs/btrfs/tree-log.c | 51 ++++++------------- 3 files changed, 77 insertions(+), 118 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/btrfs_inode.h b/fs/btrfs/btrfs_inode.h index 61205e3bbefa..1da5753d886d 100644 --- a/fs/btrfs/btrfs_inode.h +++ b/fs/btrfs/btrfs_inode.h @@ -196,6 +196,16 @@ struct btrfs_inode { struct list_head delayed_iput; long delayed_iput_count; + /* + * To avoid races between lockless (i_mutex not held) direct IO writes + * and concurrent fsync requests. Direct IO writes must acquire read + * access on this semaphore for creating an extent map and its + * corresponding ordered extent. The fast fsync path must acquire write + * access on this semaphore before it collects ordered extents and + * extent maps. + */ + struct rw_semaphore dio_sem; + struct inode vfs_inode; }; diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index ee9be4199e7c..c1ee4ade2d87 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -7145,6 +7145,43 @@ out: return em; } +static struct extent_map *btrfs_create_dio_extent(struct inode *inode, + const u64 start, + const u64 len, + const u64 orig_start, + const u64 block_start, + const u64 block_len, + const u64 orig_block_len, + const u64 ram_bytes, + const int type) +{ + struct extent_map *em = NULL; + int ret; + + down_read(&BTRFS_I(inode)->dio_sem); + if (type != BTRFS_ORDERED_NOCOW) { + em = create_pinned_em(inode, start, len, orig_start, + block_start, block_len, orig_block_len, + ram_bytes, type); + if (IS_ERR(em)) + goto out; + } + ret = btrfs_add_ordered_extent_dio(inode, start, block_start, + len, block_len, type); + if (ret) { + if (em) { + free_extent_map(em); + btrfs_drop_extent_cache(inode, start, + start + len - 1, 0); + } + em = ERR_PTR(ret); + } + out: + up_read(&BTRFS_I(inode)->dio_sem); + + return em; +} + static struct extent_map *btrfs_new_extent_direct(struct inode *inode, u64 start, u64 len) { @@ -7160,43 +7197,13 @@ static struct extent_map *btrfs_new_extent_direct(struct inode *inode, if (ret) return ERR_PTR(ret); - /* - * Create the ordered extent before the extent map. This is to avoid - * races with the fast fsync path that would lead to it logging file - * extent items that point to disk extents that were not yet written to. - * The fast fsync path collects ordered extents into a local list and - * then collects all the new extent maps, so we must create the ordered - * extent first and make sure the fast fsync path collects any new - * ordered extents after collecting new extent maps as well. - * The fsync path simply can not rely on inode_dio_wait() because it - * causes deadlock with AIO. - */ - ret = btrfs_add_ordered_extent_dio(inode, start, ins.objectid, - ins.offset, ins.offset, 0); - if (ret) { - btrfs_free_reserved_extent(root, ins.objectid, ins.offset, 1); - return ERR_PTR(ret); - } - + em = btrfs_create_dio_extent(inode, start, ins.offset, start, + ins.objectid, ins.offset, ins.offset, + ins.offset, 0); btrfs_dec_block_group_reservations(root->fs_info, ins.objectid); - - em = create_pinned_em(inode, start, ins.offset, start, ins.objectid, - ins.offset, ins.offset, ins.offset, 0); - if (IS_ERR(em)) { - struct btrfs_ordered_extent *oe; - + if (IS_ERR(em)) btrfs_free_reserved_extent(root, ins.objectid, ins.offset, 1); - oe = btrfs_lookup_ordered_extent(inode, start); - ASSERT(oe); - if (WARN_ON(!oe)) - return em; - set_bit(BTRFS_ORDERED_IOERR, &oe->flags); - set_bit(BTRFS_ORDERED_IO_DONE, &oe->flags); - btrfs_remove_ordered_extent(inode, oe); - /* Once for our lookup and once for the ordered extents tree. */ - btrfs_put_ordered_extent(oe); - btrfs_put_ordered_extent(oe); - } + return em; } @@ -7670,57 +7677,21 @@ static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock, if (can_nocow_extent(inode, start, &len, &orig_start, &orig_block_len, &ram_bytes) == 1 && btrfs_inc_nocow_writers(root->fs_info, block_start)) { + struct extent_map *em2; - /* - * Create the ordered extent before the extent map. This - * is to avoid races with the fast fsync path because it - * collects ordered extents into a local list and then - * collects all the new extent maps, so we must create - * the ordered extent first and make sure the fast fsync - * path collects any new ordered extents after - * collecting new extent maps as well. The fsync path - * simply can not rely on inode_dio_wait() because it - * causes deadlock with AIO. - */ - ret = btrfs_add_ordered_extent_dio(inode, start, - block_start, len, len, type); + em2 = btrfs_create_dio_extent(inode, start, len, + orig_start, block_start, + len, orig_block_len, + ram_bytes, type); btrfs_dec_nocow_writers(root->fs_info, block_start); - if (ret) { - free_extent_map(em); - goto unlock_err; - } - if (type == BTRFS_ORDERED_PREALLOC) { free_extent_map(em); - em = create_pinned_em(inode, start, len, - orig_start, - block_start, len, - orig_block_len, - ram_bytes, type); - if (IS_ERR(em)) { - struct btrfs_ordered_extent *oe; - - ret = PTR_ERR(em); - oe = btrfs_lookup_ordered_extent(inode, - start); - ASSERT(oe); - if (WARN_ON(!oe)) - goto unlock_err; - set_bit(BTRFS_ORDERED_IOERR, - &oe->flags); - set_bit(BTRFS_ORDERED_IO_DONE, - &oe->flags); - btrfs_remove_ordered_extent(inode, oe); - /* - * Once for our lookup and once for the - * ordered extents tree. - */ - btrfs_put_ordered_extent(oe); - btrfs_put_ordered_extent(oe); - goto unlock_err; - } + em = em2; + } + if (em2 && IS_ERR(em2)) { + ret = PTR_ERR(em2); + goto unlock_err; } - goto unlock; } } @@ -9281,6 +9252,7 @@ struct inode *btrfs_alloc_inode(struct super_block *sb) INIT_LIST_HEAD(&ei->delalloc_inodes); INIT_LIST_HEAD(&ei->delayed_iput); RB_CLEAR_NODE(&ei->rb_node); + init_rwsem(&ei->dio_sem); return inode; } diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c index a24a0ba523d6..003a826f4cff 100644 --- a/fs/btrfs/tree-log.c +++ b/fs/btrfs/tree-log.c @@ -4141,6 +4141,7 @@ static int btrfs_log_changed_extents(struct btrfs_trans_handle *trans, INIT_LIST_HEAD(&extents); + down_write(&BTRFS_I(inode)->dio_sem); write_lock(&tree->lock); test_gen = root->fs_info->last_trans_committed; @@ -4169,13 +4170,20 @@ static int btrfs_log_changed_extents(struct btrfs_trans_handle *trans, } list_sort(NULL, &extents, extent_cmp); + btrfs_get_logged_extents(inode, logged_list, start, end); /* - * Collect any new ordered extents within the range. This is to - * prevent logging file extent items without waiting for the disk - * location they point to being written. We do this only to deal - * with races against concurrent lockless direct IO writes. + * Some ordered extents started by fsync might have completed + * before we could collect them into the list logged_list, which + * means they're gone, not in our logged_list nor in the inode's + * ordered tree. We want the application/user space to know an + * error happened while attempting to persist file data so that + * it can take proper action. If such error happened, we leave + * without writing to the log tree and the fsync must report the + * file data write error and not commit the current transaction. */ - btrfs_get_logged_extents(inode, logged_list, start, end); + ret = btrfs_inode_check_errors(inode); + if (ret) + ctx->io_err = ret; process: while (!list_empty(&extents)) { em = list_entry(extents.next, struct extent_map, list); @@ -4202,6 +4210,7 @@ process: } WARN_ON(!list_empty(&extents)); write_unlock(&tree->lock); + up_write(&BTRFS_I(inode)->dio_sem); btrfs_release_path(path); return ret; @@ -4622,23 +4631,6 @@ static int btrfs_log_inode(struct btrfs_trans_handle *trans, mutex_lock(&BTRFS_I(inode)->log_mutex); - /* - * Collect ordered extents only if we are logging data. This is to - * ensure a subsequent request to log this inode in LOG_INODE_ALL mode - * will process the ordered extents if they still exists at the time, - * because when we collect them we test and set for the flag - * BTRFS_ORDERED_LOGGED to prevent multiple log requests to process the - * same ordered extents. The consequence for the LOG_INODE_ALL log mode - * not processing the ordered extents is that we end up logging the - * corresponding file extent items, based on the extent maps in the - * inode's extent_map_tree's modified_list, without logging the - * respective checksums (since the may still be only attached to the - * ordered extents and have not been inserted in the csum tree by - * btrfs_finish_ordered_io() yet). - */ - if (inode_only == LOG_INODE_ALL) - btrfs_get_logged_extents(inode, &logged_list, start, end); - /* * a brute force approach to making sure we get the most uptodate * copies of everything. @@ -4846,21 +4838,6 @@ log_extents: goto out_unlock; } if (fast_search) { - /* - * Some ordered extents started by fsync might have completed - * before we collected the ordered extents in logged_list, which - * means they're gone, not in our logged_list nor in the inode's - * ordered tree. We want the application/user space to know an - * error happened while attempting to persist file data so that - * it can take proper action. If such error happened, we leave - * without writing to the log tree and the fsync must report the - * file data write error and not commit the current transaction. - */ - err = btrfs_inode_check_errors(inode); - if (err) { - ctx->io_err = err; - goto out_unlock; - } ret = btrfs_log_changed_extents(trans, root, inode, dst_path, &logged_list, ctx, start, end); if (ret) { -- cgit v1.3-8-gc7d7 From 1a39ba99b5d533647c5dac45cd6a3e0baa7cb66a Mon Sep 17 00:00:00 2001 From: Al Viro Date: Fri, 13 May 2016 03:59:17 +0200 Subject: gfs2: Switch to generic xattr handlers Switch to the generic xattr handlers and take the necessary glocks at the layer below. The following are the new xattr "entry points"; they are called with the glock held already in the following cases: gfs2_xattr_get: From SELinux, during lookups. gfs2_xattr_set: The glock is never held. gfs2_get_acl: From gfs2_create_inode -> posix_acl_create and gfs2_setattr -> posix_acl_chmod. gfs2_set_acl: From gfs2_setattr -> posix_acl_chmod. Signed-off-by: Andreas Gruenbacher Signed-off-by: Al Viro --- fs/gfs2/acl.c | 58 ++++++++++++++++++++++++++++++++-------- fs/gfs2/acl.h | 1 + fs/gfs2/inode.c | 82 ++++++++------------------------------------------------- fs/gfs2/xattr.c | 46 +++++++++++++++++++++++++++----- 4 files changed, 99 insertions(+), 88 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/acl.c b/fs/gfs2/acl.c index 791932617d1a..363ba9e9d8d0 100644 --- a/fs/gfs2/acl.c +++ b/fs/gfs2/acl.c @@ -24,6 +24,7 @@ #include "glock.h" #include "inode.h" #include "meta_io.h" +#include "rgrp.h" #include "trans.h" #include "util.h" @@ -38,7 +39,7 @@ static const char *gfs2_acl_name(int type) return NULL; } -struct posix_acl *gfs2_get_acl(struct inode *inode, int type) +static struct posix_acl *__gfs2_get_acl(struct inode *inode, int type) { struct gfs2_inode *ip = GFS2_I(inode); struct posix_acl *acl; @@ -50,29 +51,41 @@ struct posix_acl *gfs2_get_acl(struct inode *inode, int type) return NULL; name = gfs2_acl_name(type); - if (name == NULL) - return ERR_PTR(-EINVAL); - len = gfs2_xattr_acl_get(ip, name, &data); - if (len < 0) + if (len <= 0) return ERR_PTR(len); - if (len == 0) - return NULL; - acl = posix_acl_from_xattr(&init_user_ns, data, len); kfree(data); return acl; } -int gfs2_set_acl(struct inode *inode, struct posix_acl *acl, int type) +struct posix_acl *gfs2_get_acl(struct inode *inode, int type) +{ + struct gfs2_inode *ip = GFS2_I(inode); + struct gfs2_holder gh; + bool need_unlock = false; + struct posix_acl *acl; + + if (!gfs2_glock_is_locked_by_me(ip->i_gl)) { + int ret = gfs2_glock_nq_init(ip->i_gl, LM_ST_SHARED, + LM_FLAG_ANY, &gh); + if (ret) + return ERR_PTR(ret); + need_unlock = true; + } + acl = __gfs2_get_acl(inode, type); + if (need_unlock) + gfs2_glock_dq_uninit(&gh); + return acl; +} + +int __gfs2_set_acl(struct inode *inode, struct posix_acl *acl, int type) { int error; int len; char *data; const char *name = gfs2_acl_name(type); - BUG_ON(name == NULL); - if (acl && acl->a_count > GFS2_ACL_MAX_ENTRIES(GFS2_SB(inode))) return -E2BIG; @@ -115,3 +128,26 @@ out: kfree(data); return error; } + +int gfs2_set_acl(struct inode *inode, struct posix_acl *acl, int type) +{ + struct gfs2_inode *ip = GFS2_I(inode); + struct gfs2_holder gh; + bool need_unlock = false; + int ret; + + ret = gfs2_rsqa_alloc(ip); + if (ret) + return ret; + + if (!gfs2_glock_is_locked_by_me(ip->i_gl)) { + ret = gfs2_glock_nq_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &gh); + if (ret) + return ret; + need_unlock = true; + } + ret = __gfs2_set_acl(inode, acl, type); + if (need_unlock) + gfs2_glock_dq_uninit(&gh); + return ret; +} diff --git a/fs/gfs2/acl.h b/fs/gfs2/acl.h index 3af4f407a483..f674fdd22337 100644 --- a/fs/gfs2/acl.h +++ b/fs/gfs2/acl.h @@ -15,6 +15,7 @@ #define GFS2_ACL_MAX_ENTRIES(sdp) ((300 << (sdp)->sd_sb.sb_bsize_shift) >> 12) extern struct posix_acl *gfs2_get_acl(struct inode *inode, int type); +extern int __gfs2_set_acl(struct inode *inode, struct posix_acl *acl, int type); extern int gfs2_set_acl(struct inode *inode, struct posix_acl *acl, int type); #endif /* __ACL_DOT_H__ */ diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c index 45f516cada78..72e9c64ae371 100644 --- a/fs/gfs2/inode.c +++ b/fs/gfs2/inode.c @@ -692,12 +692,12 @@ static int gfs2_create_inode(struct inode *dir, struct dentry *dentry, considered free. Any failures need to undo the gfs2 structures. */ if (default_acl) { - error = gfs2_set_acl(inode, default_acl, ACL_TYPE_DEFAULT); + error = __gfs2_set_acl(inode, default_acl, ACL_TYPE_DEFAULT); posix_acl_release(default_acl); } if (acl) { if (!error) - error = gfs2_set_acl(inode, acl, ACL_TYPE_ACCESS); + error = __gfs2_set_acl(inode, acl, ACL_TYPE_ACCESS); posix_acl_release(acl); } @@ -1948,66 +1948,6 @@ static int gfs2_getattr(struct vfsmount *mnt, struct dentry *dentry, return 0; } -static int gfs2_setxattr(struct dentry *dentry, const char *name, - const void *data, size_t size, int flags) -{ - struct inode *inode = d_inode(dentry); - struct gfs2_inode *ip = GFS2_I(inode); - struct gfs2_holder gh; - int ret; - - gfs2_holder_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &gh); - ret = gfs2_glock_nq(&gh); - if (ret == 0) { - ret = gfs2_rsqa_alloc(ip); - if (ret == 0) - ret = generic_setxattr(dentry, name, data, size, flags); - gfs2_glock_dq(&gh); - } - gfs2_holder_uninit(&gh); - return ret; -} - -static ssize_t gfs2_getxattr(struct dentry *dentry, struct inode *inode, - const char *name, void *data, size_t size) -{ - struct gfs2_inode *ip = GFS2_I(inode); - struct gfs2_holder gh; - int ret; - - /* For selinux during lookup */ - if (gfs2_glock_is_locked_by_me(ip->i_gl)) - return generic_getxattr(dentry, inode, name, data, size); - - gfs2_holder_init(ip->i_gl, LM_ST_SHARED, LM_FLAG_ANY, &gh); - ret = gfs2_glock_nq(&gh); - if (ret == 0) { - ret = generic_getxattr(dentry, inode, name, data, size); - gfs2_glock_dq(&gh); - } - gfs2_holder_uninit(&gh); - return ret; -} - -static int gfs2_removexattr(struct dentry *dentry, const char *name) -{ - struct inode *inode = d_inode(dentry); - struct gfs2_inode *ip = GFS2_I(inode); - struct gfs2_holder gh; - int ret; - - gfs2_holder_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &gh); - ret = gfs2_glock_nq(&gh); - if (ret == 0) { - ret = gfs2_rsqa_alloc(ip); - if (ret == 0) - ret = generic_removexattr(dentry, name); - gfs2_glock_dq(&gh); - } - gfs2_holder_uninit(&gh); - return ret; -} - static int gfs2_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, u64 start, u64 len) { @@ -2054,10 +1994,10 @@ const struct inode_operations gfs2_file_iops = { .permission = gfs2_permission, .setattr = gfs2_setattr, .getattr = gfs2_getattr, - .setxattr = gfs2_setxattr, - .getxattr = gfs2_getxattr, + .setxattr = generic_setxattr, + .getxattr = generic_getxattr, .listxattr = gfs2_listxattr, - .removexattr = gfs2_removexattr, + .removexattr = generic_removexattr, .fiemap = gfs2_fiemap, .get_acl = gfs2_get_acl, .set_acl = gfs2_set_acl, @@ -2076,10 +2016,10 @@ const struct inode_operations gfs2_dir_iops = { .permission = gfs2_permission, .setattr = gfs2_setattr, .getattr = gfs2_getattr, - .setxattr = gfs2_setxattr, - .getxattr = gfs2_getxattr, + .setxattr = generic_setxattr, + .getxattr = generic_getxattr, .listxattr = gfs2_listxattr, - .removexattr = gfs2_removexattr, + .removexattr = generic_removexattr, .fiemap = gfs2_fiemap, .get_acl = gfs2_get_acl, .set_acl = gfs2_set_acl, @@ -2092,10 +2032,10 @@ const struct inode_operations gfs2_symlink_iops = { .permission = gfs2_permission, .setattr = gfs2_setattr, .getattr = gfs2_getattr, - .setxattr = gfs2_setxattr, - .getxattr = gfs2_getxattr, + .setxattr = generic_setxattr, + .getxattr = generic_getxattr, .listxattr = gfs2_listxattr, - .removexattr = gfs2_removexattr, + .removexattr = generic_removexattr, .fiemap = gfs2_fiemap, }; diff --git a/fs/gfs2/xattr.c b/fs/gfs2/xattr.c index 619886ba6e78..f42ab53bd30d 100644 --- a/fs/gfs2/xattr.c +++ b/fs/gfs2/xattr.c @@ -583,13 +583,11 @@ out: * * Returns: actual size of data on success, -errno on error */ -static int gfs2_xattr_get(const struct xattr_handler *handler, - struct dentry *unused, struct inode *inode, - const char *name, void *buffer, size_t size) +static int __gfs2_xattr_get(struct inode *inode, const char *name, + void *buffer, size_t size, int type) { struct gfs2_inode *ip = GFS2_I(inode); struct gfs2_ea_location el; - int type = handler->flags; int error; if (!ip->i_eattr) @@ -611,6 +609,29 @@ static int gfs2_xattr_get(const struct xattr_handler *handler, return error; } +static int gfs2_xattr_get(const struct xattr_handler *handler, + struct dentry *unused, struct inode *inode, + const char *name, void *buffer, size_t size) +{ + struct gfs2_inode *ip = GFS2_I(inode); + struct gfs2_holder gh; + bool need_unlock = false; + int ret; + + /* During lookup, SELinux calls this function with the glock locked. */ + + if (!gfs2_glock_is_locked_by_me(ip->i_gl)) { + ret = gfs2_glock_nq_init(ip->i_gl, LM_ST_SHARED, LM_FLAG_ANY, &gh); + if (ret) + return ret; + need_unlock = true; + } + ret = __gfs2_xattr_get(inode, name, buffer, size, handler->flags); + if (need_unlock) + gfs2_glock_dq_uninit(&gh); + return ret; +} + /** * ea_alloc_blk - allocates a new block for extended attributes. * @ip: A pointer to the inode that's getting extended attributes @@ -1233,8 +1254,21 @@ static int gfs2_xattr_set(const struct xattr_handler *handler, struct dentry *dentry, const char *name, const void *value, size_t size, int flags) { - return __gfs2_xattr_set(d_inode(dentry), name, value, - size, flags, handler->flags); + struct inode *inode = d_inode(dentry); + struct gfs2_inode *ip = GFS2_I(inode); + struct gfs2_holder gh; + int ret; + + ret = gfs2_rsqa_alloc(ip); + if (ret) + return ret; + + ret = gfs2_glock_nq_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &gh); + if (ret) + return ret; + ret = __gfs2_xattr_set(inode, name, value, size, flags, handler->flags); + gfs2_glock_dq_uninit(&gh); + return ret; } static int ea_dealloc_indirect(struct gfs2_inode *ip) -- cgit v1.3-8-gc7d7 From 6c8f980c75185f8ba897814363d21882f7453f53 Mon Sep 17 00:00:00 2001 From: Andreas Gruenbacher Date: Fri, 22 Apr 2016 14:43:48 +0200 Subject: jfs: Clean up xattr name mapping Instead of stripping "os2." prefixes in __jfs_setxattr, make callers strip them, as __jfs_getxattr already does. With that change, use the same name mapping function in jfs_{get,set,remove}xattr. Signed-off-by: Andreas Gruenbacher Signed-off-by: Al Viro --- fs/jfs/xattr.c | 80 ++++++++++++++++++---------------------------------------- 1 file changed, 25 insertions(+), 55 deletions(-) (limited to 'fs') diff --git a/fs/jfs/xattr.c b/fs/jfs/xattr.c index 5becc6a3ff8c..9cdf7dc4d5cb 100644 --- a/fs/jfs/xattr.c +++ b/fs/jfs/xattr.c @@ -86,6 +86,14 @@ struct ea_buffer { #define EA_MALLOC 0x0008 +/* + * Mapping of on-disk attribute names: for on-disk attribute names with an + * unknown prefix (not "system.", "user.", "security.", or "trusted."), the + * prefix "os2." is prepended. On the way back to disk, "os2." prefixes are + * stripped and we make sure that the remaining name does not start with one + * of the know prefixes. + */ + static int is_known_namespace(const char *name) { if (strncmp(name, XATTR_SYSTEM_PREFIX, XATTR_SYSTEM_PREFIX_LEN) && @@ -97,29 +105,19 @@ static int is_known_namespace(const char *name) return true; } -/* - * These three routines are used to recognize on-disk extended attributes - * that are in a recognized namespace. If the attribute is not recognized, - * "os2." is prepended to the name - */ -static int is_os2_xattr(struct jfs_ea *ea) -{ - return !is_known_namespace(ea->name); -} - static inline int name_size(struct jfs_ea *ea) { - if (is_os2_xattr(ea)) - return ea->namelen + XATTR_OS2_PREFIX_LEN; - else + if (is_known_namespace(ea->name)) return ea->namelen; + else + return ea->namelen + XATTR_OS2_PREFIX_LEN; } static inline int copy_name(char *buffer, struct jfs_ea *ea) { int len = ea->namelen; - if (is_os2_xattr(ea)) { + if (!is_known_namespace(ea->name)) { memcpy(buffer, XATTR_OS2_PREFIX, XATTR_OS2_PREFIX_LEN); buffer += XATTR_OS2_PREFIX_LEN; len += XATTR_OS2_PREFIX_LEN; @@ -669,29 +667,24 @@ static int ea_put(tid_t tid, struct inode *inode, struct ea_buffer *ea_buf, * Most of the permission checking is done by xattr_permission in the vfs. * We also need to verify that this is a namespace that we recognize. */ -static int can_set_xattr(struct inode *inode, const char *name, - const void *value, size_t value_len) +static bool map_name_to_disk(const char **name) { - if (!strncmp(name, XATTR_OS2_PREFIX, XATTR_OS2_PREFIX_LEN)) { + if (!strncmp(*name, XATTR_OS2_PREFIX, XATTR_OS2_PREFIX_LEN)) { /* * This makes sure that we aren't trying to set an * attribute in a different namespace by prefixing it * with "os2." */ - if (is_known_namespace(name + XATTR_OS2_PREFIX_LEN)) - return -EOPNOTSUPP; - return 0; + if (is_known_namespace(*name + XATTR_OS2_PREFIX_LEN)) + return false; + *name += XATTR_OS2_PREFIX_LEN; + return true; } /* * Don't allow setting an attribute in an unknown namespace. */ - if (strncmp(name, XATTR_TRUSTED_PREFIX, XATTR_TRUSTED_PREFIX_LEN) && - strncmp(name, XATTR_SECURITY_PREFIX, XATTR_SECURITY_PREFIX_LEN) && - strncmp(name, XATTR_USER_PREFIX, XATTR_USER_PREFIX_LEN)) - return -EOPNOTSUPP; - - return 0; + return is_known_namespace(*name); } int __jfs_setxattr(tid_t tid, struct inode *inode, const char *name, @@ -704,21 +697,10 @@ int __jfs_setxattr(tid_t tid, struct inode *inode, const char *name, int xattr_size; int new_size; int namelen = strlen(name); - char *os2name = NULL; int found = 0; int rc; int length; - if (strncmp(name, XATTR_OS2_PREFIX, XATTR_OS2_PREFIX_LEN) == 0) { - os2name = kmalloc(namelen - XATTR_OS2_PREFIX_LEN + 1, - GFP_KERNEL); - if (!os2name) - return -ENOMEM; - strcpy(os2name, name + XATTR_OS2_PREFIX_LEN); - name = os2name; - namelen -= XATTR_OS2_PREFIX_LEN; - } - down_write(&JFS_IP(inode)->xattr_sem); xattr_size = ea_get(inode, &ea_buf, 0); @@ -841,8 +823,6 @@ int __jfs_setxattr(tid_t tid, struct inode *inode, const char *name, out: up_write(&JFS_IP(inode)->xattr_sem); - kfree(os2name); - return rc; } @@ -862,8 +842,8 @@ int jfs_setxattr(struct dentry *dentry, const char *name, const void *value, if (!strncmp(name, XATTR_SYSTEM_PREFIX, XATTR_SYSTEM_PREFIX_LEN)) return generic_setxattr(dentry, name, value, value_len, flags); - if ((rc = can_set_xattr(inode, name, value, value_len))) - return rc; + if (!map_name_to_disk(&name)) + return -EOPNOTSUPP; if (value == NULL) { /* empty EA, do not remove */ value = ""; @@ -946,18 +926,8 @@ ssize_t jfs_getxattr(struct dentry *dentry, struct inode *inode, if (!strncmp(name, XATTR_SYSTEM_PREFIX, XATTR_SYSTEM_PREFIX_LEN)) return generic_getxattr(dentry, inode, name, data, buf_size); - if (strncmp(name, XATTR_OS2_PREFIX, XATTR_OS2_PREFIX_LEN) == 0) { - /* - * skip past "os2." prefix - */ - name += XATTR_OS2_PREFIX_LEN; - /* - * Don't allow retrieving properly prefixed attributes - * by prepending them with "os2." - */ - if (is_known_namespace(name)) - return -EOPNOTSUPP; - } + if (!map_name_to_disk(&name)) + return -EOPNOTSUPP; err = __jfs_getxattr(inode, name, data, buf_size); @@ -1042,8 +1012,8 @@ int jfs_removexattr(struct dentry *dentry, const char *name) if (!strncmp(name, XATTR_SYSTEM_PREFIX, XATTR_SYSTEM_PREFIX_LEN)) return generic_removexattr(dentry, name); - if ((rc = can_set_xattr(inode, name, NULL, 0))) - return rc; + if (!map_name_to_disk(&name)) + return -EOPNOTSUPP; tid = txBegin(inode->i_sb, 0); mutex_lock(&ji->commit_mutex); -- cgit v1.3-8-gc7d7 From c8b6056a504fa384f36e7577fc5a1c1684fcf18a Mon Sep 17 00:00:00 2001 From: Andreas Gruenbacher Date: Fri, 22 Apr 2016 14:43:49 +0200 Subject: jfs: Switch to generic xattr handlers This is mostly the same as on other filesystems except for attribute names with an "os2." prefix: for those, the prefix is not stored on disk, and on-attribute names without a prefix have "os2." added. As on several other filesystems, the underlying function for setting/removing xattrs (__jfs_setxattr) removes attributes when the value is NULL, so the set xattr handlers will work as expected. Signed-off-by: Andreas Gruenbacher Signed-off-by: Al Viro --- fs/jfs/file.c | 6 +- fs/jfs/jfs_xattr.h | 6 +- fs/jfs/namei.c | 6 +- fs/jfs/symlink.c | 12 ++-- fs/jfs/xattr.c | 170 ++++++++++++++++++++++------------------------------- 5 files changed, 84 insertions(+), 116 deletions(-) (limited to 'fs') diff --git a/fs/jfs/file.c b/fs/jfs/file.c index 4ce7735dd042..7f1a585a0a94 100644 --- a/fs/jfs/file.c +++ b/fs/jfs/file.c @@ -140,10 +140,10 @@ int jfs_setattr(struct dentry *dentry, struct iattr *iattr) } const struct inode_operations jfs_file_inode_operations = { - .setxattr = jfs_setxattr, - .getxattr = jfs_getxattr, + .setxattr = generic_setxattr, + .getxattr = generic_getxattr, .listxattr = jfs_listxattr, - .removexattr = jfs_removexattr, + .removexattr = generic_removexattr, .setattr = jfs_setattr, #ifdef CONFIG_JFS_POSIX_ACL .get_acl = jfs_get_acl, diff --git a/fs/jfs/jfs_xattr.h b/fs/jfs/jfs_xattr.h index e69e14f3777b..561f6af46288 100644 --- a/fs/jfs/jfs_xattr.h +++ b/fs/jfs/jfs_xattr.h @@ -19,6 +19,8 @@ #ifndef H_JFS_XATTR #define H_JFS_XATTR +#include + /* * jfs_ea_list describe the on-disk format of the extended attributes. * I know the null-terminator is redundant since namelen is stored, but @@ -54,12 +56,8 @@ struct jfs_ea_list { extern int __jfs_setxattr(tid_t, struct inode *, const char *, const void *, size_t, int); -extern int jfs_setxattr(struct dentry *, const char *, const void *, size_t, - int); extern ssize_t __jfs_getxattr(struct inode *, const char *, void *, size_t); -extern ssize_t jfs_getxattr(struct dentry *, struct inode *, const char *, void *, size_t); extern ssize_t jfs_listxattr(struct dentry *, char *, size_t); -extern int jfs_removexattr(struct dentry *, const char *); extern const struct xattr_handler *jfs_xattr_handlers[]; diff --git a/fs/jfs/namei.c b/fs/jfs/namei.c index 8a40941ac9a6..7040062cb050 100644 --- a/fs/jfs/namei.c +++ b/fs/jfs/namei.c @@ -1537,10 +1537,10 @@ const struct inode_operations jfs_dir_inode_operations = { .rmdir = jfs_rmdir, .mknod = jfs_mknod, .rename = jfs_rename, - .setxattr = jfs_setxattr, - .getxattr = jfs_getxattr, + .setxattr = generic_setxattr, + .getxattr = generic_getxattr, .listxattr = jfs_listxattr, - .removexattr = jfs_removexattr, + .removexattr = generic_removexattr, .setattr = jfs_setattr, #ifdef CONFIG_JFS_POSIX_ACL .get_acl = jfs_get_acl, diff --git a/fs/jfs/symlink.c b/fs/jfs/symlink.c index f8db4fde0b0b..c94c7e4a1323 100644 --- a/fs/jfs/symlink.c +++ b/fs/jfs/symlink.c @@ -25,19 +25,19 @@ const struct inode_operations jfs_fast_symlink_inode_operations = { .readlink = generic_readlink, .get_link = simple_get_link, .setattr = jfs_setattr, - .setxattr = jfs_setxattr, - .getxattr = jfs_getxattr, + .setxattr = generic_setxattr, + .getxattr = generic_getxattr, .listxattr = jfs_listxattr, - .removexattr = jfs_removexattr, + .removexattr = generic_removexattr, }; const struct inode_operations jfs_symlink_inode_operations = { .readlink = generic_readlink, .get_link = page_get_link, .setattr = jfs_setattr, - .setxattr = jfs_setxattr, - .getxattr = jfs_getxattr, + .setxattr = generic_setxattr, + .getxattr = generic_getxattr, .listxattr = jfs_listxattr, - .removexattr = jfs_removexattr, + .removexattr = generic_removexattr, }; diff --git a/fs/jfs/xattr.c b/fs/jfs/xattr.c index 9cdf7dc4d5cb..beb182b503b3 100644 --- a/fs/jfs/xattr.c +++ b/fs/jfs/xattr.c @@ -663,30 +663,6 @@ static int ea_put(tid_t tid, struct inode *inode, struct ea_buffer *ea_buf, return 0; } -/* - * Most of the permission checking is done by xattr_permission in the vfs. - * We also need to verify that this is a namespace that we recognize. - */ -static bool map_name_to_disk(const char **name) -{ - if (!strncmp(*name, XATTR_OS2_PREFIX, XATTR_OS2_PREFIX_LEN)) { - /* - * This makes sure that we aren't trying to set an - * attribute in a different namespace by prefixing it - * with "os2." - */ - if (is_known_namespace(*name + XATTR_OS2_PREFIX_LEN)) - return false; - *name += XATTR_OS2_PREFIX_LEN; - return true; - } - - /* - * Don't allow setting an attribute in an unknown namespace. - */ - return is_known_namespace(*name); -} - int __jfs_setxattr(tid_t tid, struct inode *inode, const char *name, const void *value, size_t value_len, int flags) { @@ -826,42 +802,6 @@ int __jfs_setxattr(tid_t tid, struct inode *inode, const char *name, return rc; } -int jfs_setxattr(struct dentry *dentry, const char *name, const void *value, - size_t value_len, int flags) -{ - struct inode *inode = d_inode(dentry); - struct jfs_inode_info *ji = JFS_IP(inode); - int rc; - tid_t tid; - - /* - * If this is a request for a synthetic attribute in the system.* - * namespace use the generic infrastructure to resolve a handler - * for it via sb->s_xattr. - */ - if (!strncmp(name, XATTR_SYSTEM_PREFIX, XATTR_SYSTEM_PREFIX_LEN)) - return generic_setxattr(dentry, name, value, value_len, flags); - - if (!map_name_to_disk(&name)) - return -EOPNOTSUPP; - - if (value == NULL) { /* empty EA, do not remove */ - value = ""; - value_len = 0; - } - - tid = txBegin(inode->i_sb, 0); - mutex_lock(&ji->commit_mutex); - rc = __jfs_setxattr(tid, d_inode(dentry), name, value, value_len, - flags); - if (!rc) - rc = txCommit(tid, 1, &inode, 0); - txEnd(tid); - mutex_unlock(&ji->commit_mutex); - - return rc; -} - ssize_t __jfs_getxattr(struct inode *inode, const char *name, void *data, size_t buf_size) { @@ -913,27 +853,6 @@ ssize_t __jfs_getxattr(struct inode *inode, const char *name, void *data, return size; } -ssize_t jfs_getxattr(struct dentry *dentry, struct inode *inode, - const char *name, void *data, size_t buf_size) -{ - int err; - - /* - * If this is a request for a synthetic attribute in the system.* - * namespace use the generic infrastructure to resolve a handler - * for it via sb->s_xattr. - */ - if (!strncmp(name, XATTR_SYSTEM_PREFIX, XATTR_SYSTEM_PREFIX_LEN)) - return generic_getxattr(dentry, inode, name, data, buf_size); - - if (!map_name_to_disk(&name)) - return -EOPNOTSUPP; - - err = __jfs_getxattr(inode, name, data, buf_size); - - return err; -} - /* * No special permissions are needed to list attributes except for trusted.* */ @@ -997,27 +916,16 @@ ssize_t jfs_listxattr(struct dentry * dentry, char *data, size_t buf_size) return size; } -int jfs_removexattr(struct dentry *dentry, const char *name) +static int __jfs_xattr_set(struct inode *inode, const char *name, + const void *value, size_t size, int flags) { - struct inode *inode = d_inode(dentry); struct jfs_inode_info *ji = JFS_IP(inode); - int rc; tid_t tid; - - /* - * If this is a request for a synthetic attribute in the system.* - * namespace use the generic infrastructure to resolve a handler - * for it via sb->s_xattr. - */ - if (!strncmp(name, XATTR_SYSTEM_PREFIX, XATTR_SYSTEM_PREFIX_LEN)) - return generic_removexattr(dentry, name); - - if (!map_name_to_disk(&name)) - return -EOPNOTSUPP; + int rc; tid = txBegin(inode->i_sb, 0); mutex_lock(&ji->commit_mutex); - rc = __jfs_setxattr(tid, d_inode(dentry), name, NULL, 0, XATTR_REPLACE); + rc = __jfs_setxattr(tid, inode, name, value, size, flags); if (!rc) rc = txCommit(tid, 1, &inode, 0); txEnd(tid); @@ -1026,15 +934,77 @@ int jfs_removexattr(struct dentry *dentry, const char *name) return rc; } -/* - * List of handlers for synthetic system.* attributes. All real ondisk - * attributes are handled directly. - */ +static int jfs_xattr_get(const struct xattr_handler *handler, + struct dentry *unused, struct inode *inode, + const char *name, void *value, size_t size) +{ + name = xattr_full_name(handler, name); + return __jfs_getxattr(inode, name, value, size); +} + +static int jfs_xattr_set(const struct xattr_handler *handler, + struct dentry *dentry, const char *name, + const void *value, size_t size, int flags) +{ + struct inode *inode = d_inode(dentry); + + name = xattr_full_name(handler, name); + return __jfs_xattr_set(inode, name, value, size, flags); +} + +static int jfs_xattr_get_os2(const struct xattr_handler *handler, + struct dentry *unused, struct inode *inode, + const char *name, void *value, size_t size) +{ + if (is_known_namespace(name)) + return -EOPNOTSUPP; + return __jfs_getxattr(inode, name, value, size); +} + +static int jfs_xattr_set_os2(const struct xattr_handler *handler, + struct dentry *dentry, const char *name, + const void *value, size_t size, int flags) +{ + struct inode *inode = d_inode(dentry); + + if (is_known_namespace(name)) + return -EOPNOTSUPP; + return __jfs_xattr_set(inode, name, value, size, flags); +} + +static const struct xattr_handler jfs_user_xattr_handler = { + .prefix = XATTR_USER_PREFIX, + .get = jfs_xattr_get, + .set = jfs_xattr_set, +}; + +static const struct xattr_handler jfs_os2_xattr_handler = { + .prefix = XATTR_OS2_PREFIX, + .get = jfs_xattr_get_os2, + .set = jfs_xattr_set_os2, +}; + +static const struct xattr_handler jfs_security_xattr_handler = { + .prefix = XATTR_SECURITY_PREFIX, + .get = jfs_xattr_get, + .set = jfs_xattr_set, +}; + +static const struct xattr_handler jfs_trusted_xattr_handler = { + .prefix = XATTR_TRUSTED_PREFIX, + .get = jfs_xattr_get, + .set = jfs_xattr_set, +}; + const struct xattr_handler *jfs_xattr_handlers[] = { #ifdef CONFIG_JFS_POSIX_ACL &posix_acl_access_xattr_handler, &posix_acl_default_xattr_handler, #endif + &jfs_os2_xattr_handler, + &jfs_user_xattr_handler, + &jfs_security_xattr_handler, + &jfs_trusted_xattr_handler, NULL, }; -- cgit v1.3-8-gc7d7 From aef39ab1534d1a07061e0ee7e428499e127232eb Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Fri, 13 May 2016 00:38:15 -0400 Subject: dax: call get_blocks() with create == 1 for write faults to unwritten extents Currently, __dax_fault() does not call get_blocks() callback with create argument set, when we got back unwritten extent from the initial get_blocks() call during a write fault. This is because originally filesystems were supposed to convert unwritten extents to written ones using complete_unwritten() callback. Later this was abandoned in favor of using pre-zeroed blocks however the condition whether get_blocks() needs to be called with create == 1 remained. Fix the condition so that filesystems are not forced to zero-out and convert unwritten extents when get_blocks() is called with create == 0 (which introduces unnecessary overhead for read faults and can be problematic as the filesystem may possibly be read-only). Signed-off-by: Jan Kara Signed-off-by: Theodore Ts'o --- fs/dax.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/dax.c b/fs/dax.c index 75ba46d82a76..2494255c5785 100644 --- a/fs/dax.c +++ b/fs/dax.c @@ -667,7 +667,7 @@ int __dax_fault(struct vm_area_struct *vma, struct vm_fault *vmf, if (error) goto unlock_page; - if (!buffer_mapped(&bh) && !buffer_unwritten(&bh) && !vmf->cow_page) { + if (!buffer_mapped(&bh) && !vmf->cow_page) { if (vmf->flags & FAULT_FLAG_WRITE) { error = get_block(inode, block, &bh, 1); count_vm_event(PGMAJFAULT); -- cgit v1.3-8-gc7d7 From 7cb476f834d0dc2092e04eeafe9cbe509f6380eb Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Fri, 13 May 2016 00:38:16 -0400 Subject: ext4: handle transient ENOSPC properly for DAX ext4_dax_get_blocks() was accidentally omitted fixing get blocks handlers to properly handle transient ENOSPC errors. Fix it now to use ext4_get_blocks_trans() helper which takes care of these errors. Signed-off-by: Jan Kara Signed-off-by: Theodore Ts'o --- fs/ext4/inode.c | 75 +++++++++++++++------------------------------------------ 1 file changed, 20 insertions(+), 55 deletions(-) (limited to 'fs') diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 4d8ebbe00456..32825dee81d4 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -3232,72 +3232,37 @@ static int ext4_releasepage(struct page *page, gfp_t wait) int ext4_dax_mmap_get_block(struct inode *inode, sector_t iblock, struct buffer_head *bh_result, int create) { - int ret, err; - int credits; - struct ext4_map_blocks map; - handle_t *handle = NULL; - int flags = 0; + int ret; ext4_debug("ext4_dax_mmap_get_block: inode %lu, create flag %d\n", inode->i_ino, create); - map.m_lblk = iblock; - map.m_len = bh_result->b_size >> inode->i_blkbits; - credits = ext4_chunk_trans_blocks(inode, map.m_len); - if (create) { - flags |= EXT4_GET_BLOCKS_PRE_IO | EXT4_GET_BLOCKS_CREATE_ZERO; - handle = ext4_journal_start(inode, EXT4_HT_MAP_BLOCKS, credits); - if (IS_ERR(handle)) { - ret = PTR_ERR(handle); - return ret; - } - } + if (!create) + return _ext4_get_block(inode, iblock, bh_result, 0); - ret = ext4_map_blocks(handle, inode, &map, flags); - if (create) { - err = ext4_journal_stop(handle); - if (ret >= 0 && err < 0) - ret = err; - } - if (ret <= 0) - goto out; - if (map.m_flags & EXT4_MAP_UNWRITTEN) { - int err2; + ret = ext4_get_block_trans(inode, iblock, bh_result, + EXT4_GET_BLOCKS_PRE_IO | + EXT4_GET_BLOCKS_CREATE_ZERO); + if (ret < 0) + return ret; + if (buffer_unwritten(bh_result)) { /* * We are protected by i_mmap_sem so we know block cannot go * away from under us even though we dropped i_data_sem. * Convert extent to written and write zeros there. - * - * Note: We may get here even when create == 0. */ - handle = ext4_journal_start(inode, EXT4_HT_MAP_BLOCKS, credits); - if (IS_ERR(handle)) { - ret = PTR_ERR(handle); - goto out; - } - - err = ext4_map_blocks(handle, inode, &map, - EXT4_GET_BLOCKS_CONVERT | EXT4_GET_BLOCKS_CREATE_ZERO); - if (err < 0) - ret = err; - err2 = ext4_journal_stop(handle); - if (err2 < 0 && ret > 0) - ret = err2; - } -out: - WARN_ON_ONCE(ret == 0 && create); - if (ret > 0) { - map_bh(bh_result, inode->i_sb, map.m_pblk); - /* - * At least for now we have to clear BH_New so that DAX code - * doesn't attempt to zero blocks again in a racy way. - */ - map.m_flags &= ~EXT4_MAP_NEW; - ext4_update_bh_state(bh_result, map.m_flags); - bh_result->b_size = map.m_len << inode->i_blkbits; - ret = 0; + ret = ext4_get_block_trans(inode, iblock, bh_result, + EXT4_GET_BLOCKS_CONVERT | + EXT4_GET_BLOCKS_CREATE_ZERO); + if (ret < 0) + return ret; } - return ret; + /* + * At least for now we have to clear BH_New so that DAX code + * doesn't attempt to zero blocks again in a racy way. + */ + clear_buffer_new(bh_result); + return 0; } #endif -- cgit v1.3-8-gc7d7 From dbc427ce4028580f1244b5b57ca1cbea31aad1e7 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Fri, 13 May 2016 00:42:40 -0400 Subject: ext4: fix race in transient ENOSPC detection When there are blocks to free in the running transaction, block allocator can return ENOSPC although the filesystem has some blocks to free. We use ext4_should_retry_alloc() to force commit of the current transaction and return whether anything was committed so that it makes sense to retry the allocation. However the transaction may get committed after block allocation fails but before we call ext4_should_retry_alloc(). So ext4_should_retry_alloc() returns false because there is nothing to commit and we wrongly return ENOSPC. Fix the race by unconditionally returning 1 from ext4_should_retry_alloc() when we tried to commit a transaction. This should not add any unnecessary retries since we had a transaction running a while ago when trying to allocate blocks and we want to retry the allocation once that transaction has committed anyway. Signed-off-by: Jan Kara Signed-off-by: Theodore Ts'o --- fs/ext4/balloc.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ext4/balloc.c b/fs/ext4/balloc.c index fe1f50fe764f..3020fd70c392 100644 --- a/fs/ext4/balloc.c +++ b/fs/ext4/balloc.c @@ -610,7 +610,8 @@ int ext4_should_retry_alloc(struct super_block *sb, int *retries) jbd_debug(1, "%s: retrying operation after ENOSPC\n", sb->s_id); - return jbd2_journal_force_commit_nested(EXT4_SB(sb)->s_journal); + jbd2_journal_force_commit_nested(EXT4_SB(sb)->s_journal); + return 1; } /* -- cgit v1.3-8-gc7d7 From 914f82a32d026884743fb3de9f6f0a5908a9d5dd Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Fri, 13 May 2016 00:44:16 -0400 Subject: ext4: refactor direct IO code Currently ext4 direct IO handling is split between ext4_ext_direct_IO() and ext4_ind_direct_IO(). However the extent based function calls into the indirect based one for some cases and for example it is not able to handle file extending. Previously it was not also properly handling retries in case of ENOSPC errors. With DAX things would get even more contrieved so just refactor the direct IO code and instead of indirect / extent split do the split to read vs writes. Signed-off-by: Jan Kara Signed-off-by: Theodore Ts'o --- fs/ext4/ext4.h | 2 - fs/ext4/indirect.c | 127 --------------------------------------------------- fs/ext4/inode.c | 131 ++++++++++++++++++++++++++++++++++++++++++++++------- 3 files changed, 114 insertions(+), 146 deletions(-) (limited to 'fs') diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index ba5aecc07fbc..89e1bcb21341 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -2587,8 +2587,6 @@ extern int ext4_get_next_extent(struct inode *inode, ext4_lblk_t lblk, /* indirect.c */ extern int ext4_ind_map_blocks(handle_t *handle, struct inode *inode, struct ext4_map_blocks *map, int flags); -extern ssize_t ext4_ind_direct_IO(struct kiocb *iocb, struct iov_iter *iter, - loff_t offset); extern int ext4_ind_calc_metadata_amount(struct inode *inode, sector_t lblock); extern int ext4_ind_trans_blocks(struct inode *inode, int nrblocks); extern void ext4_ind_truncate(handle_t *, struct inode *inode); diff --git a/fs/ext4/indirect.c b/fs/ext4/indirect.c index 3027fa681de5..bc15c2c17633 100644 --- a/fs/ext4/indirect.c +++ b/fs/ext4/indirect.c @@ -648,133 +648,6 @@ out: return err; } -/* - * O_DIRECT for ext3 (or indirect map) based files - * - * If the O_DIRECT write will extend the file then add this inode to the - * orphan list. So recovery will truncate it back to the original size - * if the machine crashes during the write. - * - * If the O_DIRECT write is intantiating holes inside i_size and the machine - * crashes then stale disk data _may_ be exposed inside the file. But current - * VFS code falls back into buffered path in that case so we are safe. - */ -ssize_t ext4_ind_direct_IO(struct kiocb *iocb, struct iov_iter *iter, - loff_t offset) -{ - struct file *file = iocb->ki_filp; - struct inode *inode = file->f_mapping->host; - struct ext4_inode_info *ei = EXT4_I(inode); - handle_t *handle; - ssize_t ret; - int orphan = 0; - size_t count = iov_iter_count(iter); - int retries = 0; - - if (iov_iter_rw(iter) == WRITE) { - loff_t final_size = offset + count; - - if (final_size > inode->i_size) { - /* Credits for sb + inode write */ - handle = ext4_journal_start(inode, EXT4_HT_INODE, 2); - if (IS_ERR(handle)) { - ret = PTR_ERR(handle); - goto out; - } - ret = ext4_orphan_add(handle, inode); - if (ret) { - ext4_journal_stop(handle); - goto out; - } - orphan = 1; - ei->i_disksize = inode->i_size; - ext4_journal_stop(handle); - } - } - -retry: - if (iov_iter_rw(iter) == READ && ext4_should_dioread_nolock(inode)) { - /* - * Nolock dioread optimization may be dynamically disabled - * via ext4_inode_block_unlocked_dio(). Check inode's state - * while holding extra i_dio_count ref. - */ - inode_dio_begin(inode); - smp_mb(); - if (unlikely(ext4_test_inode_state(inode, - EXT4_STATE_DIOREAD_LOCK))) { - inode_dio_end(inode); - goto locked; - } - if (IS_DAX(inode)) - ret = dax_do_io(iocb, inode, iter, offset, - ext4_dio_get_block, NULL, 0); - else - ret = __blockdev_direct_IO(iocb, inode, - inode->i_sb->s_bdev, iter, - offset, ext4_dio_get_block, - NULL, NULL, 0); - inode_dio_end(inode); - } else { -locked: - if (IS_DAX(inode)) - ret = dax_do_io(iocb, inode, iter, offset, - ext4_dio_get_block, NULL, DIO_LOCKING); - else - ret = blockdev_direct_IO(iocb, inode, iter, offset, - ext4_dio_get_block); - - if (unlikely(iov_iter_rw(iter) == WRITE && ret < 0)) { - loff_t isize = i_size_read(inode); - loff_t end = offset + count; - - if (end > isize) - ext4_truncate_failed_write(inode); - } - } - if (ret == -ENOSPC && ext4_should_retry_alloc(inode->i_sb, &retries)) - goto retry; - - if (orphan) { - int err; - - /* Credits for sb + inode write */ - handle = ext4_journal_start(inode, EXT4_HT_INODE, 2); - if (IS_ERR(handle)) { - /* This is really bad luck. We've written the data - * but cannot extend i_size. Bail out and pretend - * the write failed... */ - ret = PTR_ERR(handle); - if (inode->i_nlink) - ext4_orphan_del(NULL, inode); - - goto out; - } - if (inode->i_nlink) - ext4_orphan_del(handle, inode); - if (ret > 0) { - loff_t end = offset + ret; - if (end > inode->i_size) { - ei->i_disksize = end; - i_size_write(inode, end); - /* - * We're going to return a positive `ret' - * here due to non-zero-length I/O, so there's - * no way of reporting error returns from - * ext4_mark_inode_dirty() to userspace. So - * ignore it. - */ - ext4_mark_inode_dirty(handle, inode); - } - } - err = ext4_journal_stop(handle); - if (ret == 0) - ret = err; - } -out: - return ret; -} - /* * Calculate the number of metadata blocks need to reserve * to allocate a new block at @lblocks for non extent file based file diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 32825dee81d4..4879e93c91d3 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -3295,7 +3295,9 @@ static int ext4_end_io_dio(struct kiocb *iocb, loff_t offset, } /* - * For ext4 extent files, ext4 will do direct-io write to holes, + * Handling of direct IO writes. + * + * For ext4 extent files, ext4 will do direct-io write even to holes, * preallocated extents, and those write extend the file, no need to * fall back to buffered IO. * @@ -3313,21 +3315,37 @@ static int ext4_end_io_dio(struct kiocb *iocb, loff_t offset, * if the machine crashes during the write. * */ -static ssize_t ext4_ext_direct_IO(struct kiocb *iocb, struct iov_iter *iter, - loff_t offset) +static ssize_t ext4_direct_IO_write(struct kiocb *iocb, struct iov_iter *iter, + loff_t offset) { struct file *file = iocb->ki_filp; struct inode *inode = file->f_mapping->host; + struct ext4_inode_info *ei = EXT4_I(inode); ssize_t ret; size_t count = iov_iter_count(iter); int overwrite = 0; get_block_t *get_block_func = NULL; int dio_flags = 0; loff_t final_size = offset + count; + int orphan = 0; + handle_t *handle; - /* Use the old path for reads and writes beyond i_size. */ - if (iov_iter_rw(iter) != WRITE || final_size > inode->i_size) - return ext4_ind_direct_IO(iocb, iter, offset); + if (final_size > inode->i_size) { + /* Credits for sb + inode write */ + handle = ext4_journal_start(inode, EXT4_HT_INODE, 2); + if (IS_ERR(handle)) { + ret = PTR_ERR(handle); + goto out; + } + ret = ext4_orphan_add(handle, inode); + if (ret) { + ext4_journal_stop(handle); + goto out; + } + orphan = 1; + ei->i_disksize = inode->i_size; + ext4_journal_stop(handle); + } BUG_ON(iocb->private == NULL); @@ -3336,8 +3354,7 @@ static ssize_t ext4_ext_direct_IO(struct kiocb *iocb, struct iov_iter *iter, * conversion. This also disallows race between truncate() and * overwrite DIO as i_dio_count needs to be incremented under i_mutex. */ - if (iov_iter_rw(iter) == WRITE) - inode_dio_begin(inode); + inode_dio_begin(inode); /* If we do a overwrite dio, i_mutex locking can be released */ overwrite = *((int *)iocb->private); @@ -3346,7 +3363,7 @@ static ssize_t ext4_ext_direct_IO(struct kiocb *iocb, struct iov_iter *iter, inode_unlock(inode); /* - * We could direct write to holes and fallocate. + * For extent mapped files we could direct write to holes and fallocate. * * Allocated blocks to fill the hole are marked as unwritten to prevent * parallel buffered read to expose the stale data before DIO complete @@ -3368,7 +3385,11 @@ static ssize_t ext4_ext_direct_IO(struct kiocb *iocb, struct iov_iter *iter, iocb->private = NULL; if (overwrite) get_block_func = ext4_dio_get_block_overwrite; - else if (is_sync_kiocb(iocb)) { + else if (!ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS) || + round_down(offset, 1 << inode->i_blkbits) >= inode->i_size) { + get_block_func = ext4_dio_get_block; + dio_flags = DIO_LOCKING | DIO_SKIP_HOLES; + } else if (is_sync_kiocb(iocb)) { get_block_func = ext4_dio_get_block_unwritten_sync; dio_flags = DIO_LOCKING; } else { @@ -3378,10 +3399,11 @@ static ssize_t ext4_ext_direct_IO(struct kiocb *iocb, struct iov_iter *iter, #ifdef CONFIG_EXT4_FS_ENCRYPTION BUG_ON(ext4_encrypted_inode(inode) && S_ISREG(inode->i_mode)); #endif - if (IS_DAX(inode)) + if (IS_DAX(inode)) { + dio_flags &= ~DIO_SKIP_HOLES; ret = dax_do_io(iocb, inode, iter, offset, get_block_func, ext4_end_io_dio, dio_flags); - else + } else ret = __blockdev_direct_IO(iocb, inode, inode->i_sb->s_bdev, iter, offset, get_block_func, @@ -3401,12 +3423,87 @@ static ssize_t ext4_ext_direct_IO(struct kiocb *iocb, struct iov_iter *iter, ext4_clear_inode_state(inode, EXT4_STATE_DIO_UNWRITTEN); } - if (iov_iter_rw(iter) == WRITE) - inode_dio_end(inode); + inode_dio_end(inode); /* take i_mutex locking again if we do a ovewrite dio */ if (overwrite) inode_lock(inode); + if (ret < 0 && final_size > inode->i_size) + ext4_truncate_failed_write(inode); + + /* Handle extending of i_size after direct IO write */ + if (orphan) { + int err; + + /* Credits for sb + inode write */ + handle = ext4_journal_start(inode, EXT4_HT_INODE, 2); + if (IS_ERR(handle)) { + /* This is really bad luck. We've written the data + * but cannot extend i_size. Bail out and pretend + * the write failed... */ + ret = PTR_ERR(handle); + if (inode->i_nlink) + ext4_orphan_del(NULL, inode); + + goto out; + } + if (inode->i_nlink) + ext4_orphan_del(handle, inode); + if (ret > 0) { + loff_t end = offset + ret; + if (end > inode->i_size) { + ei->i_disksize = end; + i_size_write(inode, end); + /* + * We're going to return a positive `ret' + * here due to non-zero-length I/O, so there's + * no way of reporting error returns from + * ext4_mark_inode_dirty() to userspace. So + * ignore it. + */ + ext4_mark_inode_dirty(handle, inode); + } + } + err = ext4_journal_stop(handle); + if (ret == 0) + ret = err; + } +out: + return ret; +} + +static ssize_t ext4_direct_IO_read(struct kiocb *iocb, struct iov_iter *iter, + loff_t offset) +{ + int unlocked = 0; + struct inode *inode = iocb->ki_filp->f_mapping->host; + ssize_t ret; + + if (ext4_should_dioread_nolock(inode)) { + /* + * Nolock dioread optimization may be dynamically disabled + * via ext4_inode_block_unlocked_dio(). Check inode's state + * while holding extra i_dio_count ref. + */ + inode_dio_begin(inode); + smp_mb(); + if (unlikely(ext4_test_inode_state(inode, + EXT4_STATE_DIOREAD_LOCK))) + inode_dio_end(inode); + else + unlocked = 1; + } + if (IS_DAX(inode)) { + ret = dax_do_io(iocb, inode, iter, offset, ext4_dio_get_block, + NULL, unlocked ? 0 : DIO_LOCKING); + } else { + ret = __blockdev_direct_IO(iocb, inode, inode->i_sb->s_bdev, + iter, offset, ext4_dio_get_block, + NULL, NULL, + unlocked ? 0 : DIO_LOCKING); + } + if (unlocked) + inode_dio_end(inode); return ret; } @@ -3434,10 +3531,10 @@ static ssize_t ext4_direct_IO(struct kiocb *iocb, struct iov_iter *iter, return 0; trace_ext4_direct_IO_enter(inode, offset, count, iov_iter_rw(iter)); - if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)) - ret = ext4_ext_direct_IO(iocb, iter, offset); + if (iov_iter_rw(iter) == READ) + ret = ext4_direct_IO_read(iocb, iter, offset); else - ret = ext4_ind_direct_IO(iocb, iter, offset); + ret = ext4_direct_IO_write(iocb, iter, offset); trace_ext4_direct_IO_exit(inode, offset, count, iov_iter_rw(iter), ret); return ret; } -- cgit v1.3-8-gc7d7 From 12735f881952c32b31bc4e433768f18489f79ec9 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Fri, 13 May 2016 00:51:15 -0400 Subject: ext4: pre-zero allocated blocks for DAX IO Currently ext4 treats DAX IO the same way as direct IO. I.e., it allocates unwritten extents before IO is done and converts unwritten extents afterwards. However this way DAX IO can race with page fault to the same area: ext4_ext_direct_IO() dax_fault() dax_io() get_block() - allocates unwritten extent copy_from_iter_pmem() get_block() - converts unwritten block to written and zeroes it out ext4_convert_unwritten_extents() So data written with DAX IO gets lost. Similarly dax_new_buf() called from dax_io() can overwrite data that has been already written to the block via mmap. Fix the problem by using pre-zeroed blocks for DAX IO the same way as we use them for DAX mmap. The downside of this solution is that every allocating write writes each block twice (once zeros, once data). Fixing the race with locking is possible as well however we would need to lock-out faults for the whole range written to by DAX IO. And that is not easy to do without locking-out faults for the whole file which seems too aggressive. Signed-off-by: Jan Kara Signed-off-by: Theodore Ts'o --- fs/ext4/ext4.h | 11 +++++++++-- fs/ext4/file.c | 4 ++-- fs/ext4/inode.c | 43 +++++++++++++++++++++++++++++++++---------- 3 files changed, 44 insertions(+), 14 deletions(-) (limited to 'fs') diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index 89e1bcb21341..b84aa1ca480a 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -2527,8 +2527,8 @@ struct buffer_head *ext4_getblk(handle_t *, struct inode *, ext4_lblk_t, int); struct buffer_head *ext4_bread(handle_t *, struct inode *, ext4_lblk_t, int); int ext4_get_block_unwritten(struct inode *inode, sector_t iblock, struct buffer_head *bh_result, int create); -int ext4_dax_mmap_get_block(struct inode *inode, sector_t iblock, - struct buffer_head *bh_result, int create); +int ext4_dax_get_block(struct inode *inode, sector_t iblock, + struct buffer_head *bh_result, int create); int ext4_get_block(struct inode *inode, sector_t iblock, struct buffer_head *bh_result, int create); int ext4_dio_get_block(struct inode *inode, sector_t iblock, @@ -3334,6 +3334,13 @@ static inline void ext4_clear_io_unwritten_flag(ext4_io_end_t *io_end) } } +static inline bool ext4_aligned_io(struct inode *inode, loff_t off, loff_t len) +{ + int blksize = 1 << inode->i_blkbits; + + return IS_ALIGNED(off, blksize) && IS_ALIGNED(len, blksize); +} + #endif /* __KERNEL__ */ #define EFSBADCRC EBADMSG /* Bad CRC detected */ diff --git a/fs/ext4/file.c b/fs/ext4/file.c index 3e850b988923..37e28082885a 100644 --- a/fs/ext4/file.c +++ b/fs/ext4/file.c @@ -207,7 +207,7 @@ static int ext4_dax_fault(struct vm_area_struct *vma, struct vm_fault *vmf) if (IS_ERR(handle)) result = VM_FAULT_SIGBUS; else - result = __dax_fault(vma, vmf, ext4_dax_mmap_get_block, NULL); + result = __dax_fault(vma, vmf, ext4_dax_get_block, NULL); if (write) { if (!IS_ERR(handle)) @@ -243,7 +243,7 @@ static int ext4_dax_pmd_fault(struct vm_area_struct *vma, unsigned long addr, result = VM_FAULT_SIGBUS; else result = __dax_pmd_fault(vma, addr, pmd, flags, - ext4_dax_mmap_get_block, NULL); + ext4_dax_get_block, NULL); if (write) { if (!IS_ERR(handle)) diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 4879e93c91d3..f9ab1e8cc416 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -3229,13 +3229,17 @@ static int ext4_releasepage(struct page *page, gfp_t wait) } #ifdef CONFIG_FS_DAX -int ext4_dax_mmap_get_block(struct inode *inode, sector_t iblock, - struct buffer_head *bh_result, int create) +/* + * Get block function for DAX IO and mmap faults. It takes care of converting + * unwritten extents to written ones and initializes new / converted blocks + * to zeros. + */ +int ext4_dax_get_block(struct inode *inode, sector_t iblock, + struct buffer_head *bh_result, int create) { int ret; - ext4_debug("ext4_dax_mmap_get_block: inode %lu, create flag %d\n", - inode->i_ino, create); + ext4_debug("inode %lu, create flag %d\n", inode->i_ino, create); if (!create) return _ext4_get_block(inode, iblock, bh_result, 0); @@ -3247,9 +3251,9 @@ int ext4_dax_mmap_get_block(struct inode *inode, sector_t iblock, if (buffer_unwritten(bh_result)) { /* - * We are protected by i_mmap_sem so we know block cannot go - * away from under us even though we dropped i_data_sem. - * Convert extent to written and write zeros there. + * We are protected by i_mmap_sem or i_mutex so we know block + * cannot go away from under us even though we dropped + * i_data_sem. Convert extent to written and write zeros there. */ ret = ext4_get_block_trans(inode, iblock, bh_result, EXT4_GET_BLOCKS_CONVERT | @@ -3264,6 +3268,14 @@ int ext4_dax_mmap_get_block(struct inode *inode, sector_t iblock, clear_buffer_new(bh_result); return 0; } +#else +/* Just define empty function, it will never get called. */ +int ext4_dax_get_block(struct inode *inode, sector_t iblock, + struct buffer_head *bh_result, int create) +{ + BUG(); + return 0; +} #endif static int ext4_end_io_dio(struct kiocb *iocb, loff_t offset, @@ -3385,8 +3397,20 @@ static ssize_t ext4_direct_IO_write(struct kiocb *iocb, struct iov_iter *iter, iocb->private = NULL; if (overwrite) get_block_func = ext4_dio_get_block_overwrite; - else if (!ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS) || - round_down(offset, 1 << inode->i_blkbits) >= inode->i_size) { + else if (IS_DAX(inode)) { + /* + * We can avoid zeroing for aligned DAX writes beyond EOF. Other + * writes need zeroing either because they can race with page + * faults or because they use partial blocks. + */ + if (round_down(offset, 1<i_blkbits) >= inode->i_size && + ext4_aligned_io(inode, offset, count)) + get_block_func = ext4_dio_get_block; + else + get_block_func = ext4_dax_get_block; + dio_flags = DIO_LOCKING; + } else if (!ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS) || + round_down(offset, 1 << inode->i_blkbits) >= inode->i_size) { get_block_func = ext4_dio_get_block; dio_flags = DIO_LOCKING | DIO_SKIP_HOLES; } else if (is_sync_kiocb(iocb)) { @@ -3400,7 +3424,6 @@ static ssize_t ext4_direct_IO_write(struct kiocb *iocb, struct iov_iter *iter, BUG_ON(ext4_encrypted_inode(inode) && S_ISREG(inode->i_mode)); #endif if (IS_DAX(inode)) { - dio_flags &= ~DIO_SKIP_HOLES; ret = dax_do_io(iocb, inode, iter, offset, get_block_func, ext4_end_io_dio, dio_flags); } else -- cgit v1.3-8-gc7d7 From e55d5312444087eb6bfb34c1cd5f6e0bf626cf26 Mon Sep 17 00:00:00 2001 From: Daniel Wagner Date: Thu, 11 Feb 2016 13:36:54 +0100 Subject: crash_dump: Add vmcore_elf32_check_arch parse_crash_elf{32|64}_headers will check the headers via the elf_check_arch respectively vmcore_elf64_check_arch macro. The MIPS architecture implements those two macros differently. In order to make the differentiation more explicit, let's introduce an vmcore_elf32_check_arch to allow the archs to overwrite it. Signed-off-by: Daniel Wagner Suggested-by: Maciej W. Rozycki Reviewed-by: Maciej W. Rozycki Cc: linux-kernel@vger.kernel.org Cc: linux-mips@linux-mips.org Patchwork: https://patchwork.linux-mips.org/patch/12535/ Signed-off-by: Ralf Baechle --- fs/proc/vmcore.c | 2 +- include/linux/crash_dump.h | 8 ++++++-- 2 files changed, 7 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c index 8afe10cf7df8..8ab782d8b33d 100644 --- a/fs/proc/vmcore.c +++ b/fs/proc/vmcore.c @@ -1071,7 +1071,7 @@ static int __init parse_crash_elf32_headers(void) /* Do some basic Verification. */ if (memcmp(ehdr.e_ident, ELFMAG, SELFMAG) != 0 || (ehdr.e_type != ET_CORE) || - !elf_check_arch(&ehdr) || + !vmcore_elf32_check_arch(&ehdr) || ehdr.e_ident[EI_CLASS] != ELFCLASS32|| ehdr.e_ident[EI_VERSION] != EV_CURRENT || ehdr.e_version != EV_CURRENT || diff --git a/include/linux/crash_dump.h b/include/linux/crash_dump.h index 3849fce7ecfe..3873697ba21c 100644 --- a/include/linux/crash_dump.h +++ b/include/linux/crash_dump.h @@ -34,9 +34,13 @@ void vmcore_cleanup(void); /* * Architecture code can redefine this if there are any special checks - * needed for 64-bit ELF vmcores. In case of 32-bit only architecture, - * this can be set to zero. + * needed for 32-bit ELF or 64-bit ELF vmcores. In case of 32-bit + * only architecture, vmcore_elf64_check_arch can be set to zero. */ +#ifndef vmcore_elf32_check_arch +#define vmcore_elf32_check_arch(x) elf_check_arch(x) +#endif + #ifndef vmcore_elf64_check_arch #define vmcore_elf64_check_arch(x) (elf_check_arch(x) || vmcore_elf_check_arch_cross(x)) #endif -- cgit v1.3-8-gc7d7 From 14b7f4a1edf15808cf52dc38277ffe1d84854839 Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Thu, 5 May 2016 06:53:47 -0400 Subject: nfsd: handle seqid wraparound in nfsd4_preprocess_layout_stateid Move the existing static function to an inline helper, and call it. Reviewed-by: Christoph Hellwig Signed-off-by: Jeff Layton Signed-off-by: J. Bruce Fields --- fs/nfsd/nfs4layouts.c | 2 +- fs/nfsd/nfs4state.c | 8 +------- fs/nfsd/state.h | 5 +++++ 3 files changed, 7 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/nfs4layouts.c b/fs/nfsd/nfs4layouts.c index 825c7bc8d789..953c0755cb37 100644 --- a/fs/nfsd/nfs4layouts.c +++ b/fs/nfsd/nfs4layouts.c @@ -289,7 +289,7 @@ nfsd4_preprocess_layout_stateid(struct svc_rqst *rqstp, status = nfserr_bad_stateid; mutex_lock(&ls->ls_mutex); - if (stateid->si_generation > stid->sc_stateid.si_generation) + if (nfsd4_stateid_generation_after(stateid, &stid->sc_stateid)) goto out_unlock_stid; if (layout_type != ls->ls_layout_type) goto out_unlock_stid; diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index 0462eeddfff9..f5f82e145018 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -4651,12 +4651,6 @@ grace_disallows_io(struct net *net, struct inode *inode) return opens_in_grace(net) && mandatory_lock(inode); } -/* Returns true iff a is later than b: */ -static bool stateid_generation_after(stateid_t *a, stateid_t *b) -{ - return (s32)(a->si_generation - b->si_generation) > 0; -} - static __be32 check_stateid_generation(stateid_t *in, stateid_t *ref, bool has_session) { /* @@ -4670,7 +4664,7 @@ static __be32 check_stateid_generation(stateid_t *in, stateid_t *ref, bool has_s return nfs_ok; /* If the client sends us a stateid from the future, it's buggy: */ - if (stateid_generation_after(in, ref)) + if (nfsd4_stateid_generation_after(in, ref)) return nfserr_bad_stateid; /* * However, we could see a stateid from the past, even from a diff --git a/fs/nfsd/state.h b/fs/nfsd/state.h index c050c53036a6..986e51e5ceac 100644 --- a/fs/nfsd/state.h +++ b/fs/nfsd/state.h @@ -573,6 +573,11 @@ enum nfsd4_cb_op { NFSPROC4_CLNT_CB_SEQUENCE, }; +/* Returns true iff a is later than b: */ +static inline bool nfsd4_stateid_generation_after(stateid_t *a, stateid_t *b) +{ + return (s32)(a->si_generation - b->si_generation) > 0; +} struct nfsd4_compound_state; struct nfsd_net; -- cgit v1.3-8-gc7d7 From 6625d0913771df5f12b9531c8cb8414e55f1c21d Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Wed, 4 May 2016 10:52:55 -0400 Subject: svcrdma: Do not add XDR padding to xdr_buf page vector An xdr_buf has a head, a vector of pages, and a tail. Each RPC request is presented to the NFS server contained in an xdr_buf. The RDMA transport would like to supply the NFS server with only the NFS WRITE payload bytes in the page vector. In some common cases, that would allow the NFS server to swap those pages right into the target file's page cache. Have the transport's RDMA Read logic put XDR pad bytes in the tail iovec, and not in the pages that hold the data payload. The NFSv3 WRITE XDR decoder is finicky about the lengths involved, so make sure it is looking in the correct places when computing the total length of the incoming NFS WRITE request. Signed-off-by: Chuck Lever Signed-off-by: J. Bruce Fields --- fs/nfsd/nfs3xdr.c | 2 +- net/sunrpc/xprtrdma/svc_rdma_recvfrom.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/nfs3xdr.c b/fs/nfsd/nfs3xdr.c index 2246454dec76..c5eff5fde11c 100644 --- a/fs/nfsd/nfs3xdr.c +++ b/fs/nfsd/nfs3xdr.c @@ -379,7 +379,7 @@ nfs3svc_decode_writeargs(struct svc_rqst *rqstp, __be32 *p, */ hdr = (void*)p - rqstp->rq_arg.head[0].iov_base; dlen = rqstp->rq_arg.head[0].iov_len + rqstp->rq_arg.page_len - - hdr; + + rqstp->rq_arg.tail[0].iov_len - hdr; /* * Round the length of the data which was specified up to * the next multiple of XDR units and then compare that diff --git a/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c b/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c index 3b24a646eb46..234be9de6210 100644 --- a/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c +++ b/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c @@ -488,7 +488,7 @@ static int rdma_read_chunks(struct svcxprt_rdma *xprt, if (page_offset & 3) { u32 pad = 4 - (page_offset & 3); - head->arg.page_len += pad; + head->arg.tail[0].iov_len += pad; head->arg.len += pad; head->arg.buflen += pad; page_offset += pad; -- cgit v1.3-8-gc7d7 From 4673272f43ae790ab9ec04e38a7542f82bb8f020 Mon Sep 17 00:00:00 2001 From: Scott Talbert Date: Mon, 9 May 2016 09:14:28 -0400 Subject: btrfs: fix memory leak during RAID 5/6 device replacement A 'struct bio' is allocated in scrub_missing_raid56_pages(), but it was never freed anywhere. Signed-off-by: Scott Talbert Signed-off-by: David Sterba --- fs/btrfs/scrub.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c index d270c700ed31..fa35cdc46494 100644 --- a/fs/btrfs/scrub.c +++ b/fs/btrfs/scrub.c @@ -2127,6 +2127,8 @@ static void scrub_missing_raid56_end_io(struct bio *bio) if (bio->bi_error) sblock->no_io_error_seen = 0; + bio_put(bio); + btrfs_queue_work(fs_info->scrub_workers, &sblock->work); } -- cgit v1.3-8-gc7d7 From 99e3e858a486ccef93da0d4b67a71ed1c171b2fe Mon Sep 17 00:00:00 2001 From: Sheng Yong Date: Wed, 11 May 2016 17:08:14 +0800 Subject: f2fs: correct return value type of f2fs_fill_super Signed-off-by: Sheng Yong Signed-off-by: Jaegeuk Kim --- fs/f2fs/super.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 28c8992da6f6..0e54e84e06e4 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -1417,7 +1417,7 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent) struct f2fs_sb_info *sbi; struct f2fs_super_block *raw_super; struct inode *root; - long err; + int err; bool retry = true, need_fsck = false; char *options = NULL; int recovery, i, valid_super_block; @@ -1643,7 +1643,7 @@ try_onemore: if (err < 0) { need_fsck = true; f2fs_msg(sb, KERN_ERR, - "Cannot recover all fsync data errno=%ld", err); + "Cannot recover all fsync data errno=%d", err); goto free_kobj; } } else { @@ -1676,7 +1676,7 @@ try_onemore: if (recovery) { err = f2fs_commit_super(sbi, true); f2fs_msg(sb, KERN_INFO, - "Try to recover %dth superblock, ret: %ld", + "Try to recover %dth superblock, ret: %d", sbi->valid_super_block ? 1 : 2, err); } -- cgit v1.3-8-gc7d7 From e4103849ba784368e802eec4f15832ac54eb7d39 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Sat, 14 May 2016 19:03:52 +0800 Subject: f2fs: fix i_current_depth during inline dentry conversion With below steps, we will see that dentry page becoming unaccessable later. This is because we forget updating i_current_depth in inode during inline dentry conversion, after that, once we failed at somewhere, it will leave i_current_depth as 0 in non-inline directory. Then, during ->lookup, the current_depth value makes all dentry pages in first level invisible. Fix it. 1) mount f2fs with inline_dentry option 2) mkdir dir 3) touch 180 files named [0-179] in dir 4) touch 180 in dir (fail after inline dir conversion) 5) ll dir ls: cannot access /mnt/f2fs/dir/0: No such file or directory ls: cannot access /mnt/f2fs/dir/1: No such file or directory ls: cannot access /mnt/f2fs/dir/2: No such file or directory ls: cannot access /mnt/f2fs/dir/3: No such file or directory ls: cannot access /mnt/f2fs/dir/4: No such file or directory drwxr-xr-x 2 root root 4096 may 13 21:47 ./ drwxr-xr-x 3 root root 4096 may 13 21:46 ../ -????????? ? ? ? ? ? 0 -????????? ? ? ? ? ? 1 -????????? ? ? ? ? ? 10 -????????? ? ? ? ? ? 100 -????????? ? ? ? ? ? 101 -????????? ? ? ? ? ? 102 Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/inline.c | 1 + 1 file changed, 1 insertion(+) (limited to 'fs') diff --git a/fs/f2fs/inline.c b/fs/f2fs/inline.c index 8a51955b68aa..60ba7ac808b9 100644 --- a/fs/f2fs/inline.c +++ b/fs/f2fs/inline.c @@ -400,6 +400,7 @@ static int f2fs_move_inline_dirents(struct inode *dir, struct page *ipage, stat_dec_inline_dir(dir); clear_inode_flag(F2FS_I(dir), FI_INLINE_DENTRY); + F2FS_I(dir)->i_current_depth = 1; if (i_size_read(dir) < PAGE_SIZE) { i_size_write(dir, PAGE_SIZE); set_inode_flag(F2FS_I(dir), FI_UPDATE_DIR); -- cgit v1.3-8-gc7d7 From 8975bdf48220cca703ad4586528e04b76b40ea1c Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Sat, 14 May 2016 19:03:53 +0800 Subject: f2fs: fix incorrect error path handling in f2fs_move_rehashed_dirents Fix two bugs in error path of f2fs_move_rehashed_dirents: - release dir's inode page if fail to call kmalloc - recover i_current_depth if fail to converting Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/inline.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/inline.c b/fs/f2fs/inline.c index 60ba7ac808b9..a4bb155dd00a 100644 --- a/fs/f2fs/inline.c +++ b/fs/f2fs/inline.c @@ -464,12 +464,15 @@ static int f2fs_move_rehashed_dirents(struct inode *dir, struct page *ipage, struct f2fs_inline_dentry *inline_dentry) { struct f2fs_inline_dentry *backup_dentry; + struct f2fs_inode_info *fi = F2FS_I(dir); int err; backup_dentry = f2fs_kmalloc(sizeof(struct f2fs_inline_dentry), GFP_F2FS_ZERO); - if (!backup_dentry) + if (!backup_dentry) { + f2fs_put_page(ipage, 1); return -ENOMEM; + } memcpy(backup_dentry, inline_dentry, MAX_INLINE_DATA); truncate_inline_inode(ipage, 0); @@ -483,13 +486,14 @@ static int f2fs_move_rehashed_dirents(struct inode *dir, struct page *ipage, lock_page(ipage); stat_dec_inline_dir(dir); - clear_inode_flag(F2FS_I(dir), FI_INLINE_DENTRY); + clear_inode_flag(fi, FI_INLINE_DENTRY); update_inode(dir, ipage); kfree(backup_dentry); return 0; recover: lock_page(ipage); memcpy(inline_dentry, backup_dentry, MAX_INLINE_DATA); + fi->i_current_depth = 0; i_size_write(dir, MAX_INLINE_DATA); update_inode(dir, ipage); f2fs_put_page(ipage, 1); -- cgit v1.3-8-gc7d7 From b951a4ec165af4973b2bd9c80fb5845fbd840435 Mon Sep 17 00:00:00 2001 From: Yunlei He Date: Fri, 13 May 2016 14:57:43 +0800 Subject: f2fs: no need inc dirty pages under inode lock No need inc dirty pages under inode lock Signed-off-by: Yunlei He Signed-off-by: Jaegeuk Kim --- fs/f2fs/checkpoint.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c index 6402e9372f6c..3da6499a9e82 100644 --- a/fs/f2fs/checkpoint.c +++ b/fs/f2fs/checkpoint.c @@ -787,9 +787,9 @@ void update_dirty_page(struct inode *inode, struct page *page) spin_lock(&sbi->inode_lock[type]); __add_dirty_inode(inode, type); - inode_inc_dirty_pages(inode); spin_unlock(&sbi->inode_lock[type]); + inode_inc_dirty_pages(inode); SetPagePrivate(page); f2fs_trace_pid(page); } -- cgit v1.3-8-gc7d7 From 087968974fcd9e8c910bba73f3393a49b7380ca6 Mon Sep 17 00:00:00 2001 From: Sheng Yong Date: Mon, 16 May 2016 12:38:50 +0800 Subject: f2fs: add fault injection to sysfs This patch introduces a new struct f2fs_fault_info and a global f2fs_fault to save fault injection status. Fault injection entries are created in /sys/fs/f2fs/fault_injection/ during initializing f2fs module. Signed-off-by: Sheng Yong Signed-off-by: Jaegeuk Kim --- fs/f2fs/f2fs.h | 31 ++++++++++++++++++++----- fs/f2fs/super.c | 70 ++++++++++++++++++++++++++++++++++++++++++++++++++++----- 2 files changed, 91 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index ec978b44d546..1351178b2b58 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -48,15 +48,36 @@ enum { FAULT_MAX, }; -extern u32 f2fs_fault_rate; -extern atomic_t f2fs_ops; +struct f2fs_fault_info { + atomic_t inject_ops; + unsigned int inject_rate; + unsigned int inject_type; +}; + +extern struct f2fs_fault_info f2fs_fault; extern char *fault_name[FAULT_MAX]; +#define IS_FAULT_SET(type) (f2fs_fault.inject_type & (1 << (type))) static inline bool time_to_inject(int type) { - atomic_inc(&f2fs_ops); - if (f2fs_fault_rate && (atomic_read(&f2fs_ops) >= f2fs_fault_rate)) { - atomic_set(&f2fs_ops, 0); + if (!f2fs_fault.inject_rate) + return false; + if (type == FAULT_KMALLOC && !IS_FAULT_SET(type)) + return false; + else if (type == FAULT_PAGE_ALLOC && !IS_FAULT_SET(type)) + return false; + else if (type == FAULT_ALLOC_NID && !IS_FAULT_SET(type)) + return false; + else if (type == FAULT_ORPHAN && !IS_FAULT_SET(type)) + return false; + else if (type == FAULT_BLOCK && !IS_FAULT_SET(type)) + return false; + else if (type == FAULT_DIR_DEPTH && !IS_FAULT_SET(type)) + return false; + + atomic_inc(&f2fs_fault.inject_ops); + if (atomic_read(&f2fs_fault.inject_ops) >= f2fs_fault.inject_rate) { + atomic_set(&f2fs_fault.inject_ops, 0); printk("%sF2FS-fs : inject %s in %pF\n", KERN_INFO, fault_name[type], diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 0e54e84e06e4..9df6d72ff1a1 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -40,8 +40,7 @@ static struct kmem_cache *f2fs_inode_cachep; static struct kset *f2fs_kset; #ifdef CONFIG_F2FS_FAULT_INJECTION -u32 f2fs_fault_rate = 0; -atomic_t f2fs_ops; +struct f2fs_fault_info f2fs_fault; char *fault_name[FAULT_MAX] = { [FAULT_KMALLOC] = "kmalloc", @@ -51,6 +50,17 @@ char *fault_name[FAULT_MAX] = { [FAULT_BLOCK] = "no more block", [FAULT_DIR_DEPTH] = "too big dir depth", }; + +static void f2fs_build_fault_attr(unsigned int rate) +{ + if (rate) { + atomic_set(&f2fs_fault.inject_ops, 0); + f2fs_fault.inject_rate = rate; + f2fs_fault.inject_type = (1 << FAULT_MAX) - 1; + } else { + memset(&f2fs_fault, 0, sizeof(struct f2fs_fault_info)); + } +} #endif /* f2fs-wide shrinker description */ @@ -118,6 +128,10 @@ enum { SM_INFO, /* struct f2fs_sm_info */ NM_INFO, /* struct f2fs_nm_info */ F2FS_SBI, /* struct f2fs_sb_info */ +#ifdef CONFIG_F2FS_FAULT_INJECTION + FAULT_INFO_RATE, /* struct f2fs_fault_info */ + FAULT_INFO_TYPE, /* struct f2fs_fault_info */ +#endif }; struct f2fs_attr { @@ -139,6 +153,11 @@ static unsigned char *__struct_ptr(struct f2fs_sb_info *sbi, int struct_type) return (unsigned char *)NM_I(sbi); else if (struct_type == F2FS_SBI) return (unsigned char *)sbi; +#ifdef CONFIG_F2FS_FAULT_INJECTION + else if (struct_type == FAULT_INFO_RATE || + struct_type == FAULT_INFO_TYPE) + return (unsigned char *)&f2fs_fault; +#endif return NULL; } @@ -188,6 +207,10 @@ static ssize_t f2fs_sbi_store(struct f2fs_attr *a, ret = kstrtoul(skip_spaces(buf), 0, &t); if (ret < 0) return ret; +#ifdef CONFIG_F2FS_FAULT_INJECTION + if (a->struct_type == FAULT_INFO_TYPE && t >= (1 << FAULT_MAX)) + return -EINVAL; +#endif *ui = t; return count; } @@ -253,6 +276,10 @@ F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, max_victim_search, max_victim_search); F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, dir_level, dir_level); F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, cp_interval, interval_time[CP_TIME]); F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, idle_interval, interval_time[REQ_TIME]); +#ifdef CONFIG_F2FS_FAULT_INJECTION +F2FS_RW_ATTR(FAULT_INFO_RATE, f2fs_fault_info, inject_rate, inject_rate); +F2FS_RW_ATTR(FAULT_INFO_TYPE, f2fs_fault_info, inject_type, inject_type); +#endif F2FS_GENERAL_RO_ATTR(lifetime_write_kbytes); #define ATTR_LIST(name) (&f2fs_attr_##name.attr) @@ -289,6 +316,22 @@ static struct kobj_type f2fs_ktype = { .release = f2fs_sb_release, }; +#ifdef CONFIG_F2FS_FAULT_INJECTION +/* sysfs for f2fs fault injection */ +static struct kobject f2fs_fault_inject; + +static struct attribute *f2fs_fault_attrs[] = { + ATTR_LIST(inject_rate), + ATTR_LIST(inject_type), + NULL +}; + +static struct kobj_type f2fs_fault_ktype = { + .default_attrs = f2fs_fault_attrs, + .sysfs_ops = &f2fs_attr_ops, +}; +#endif + void f2fs_msg(struct super_block *sb, const char *level, const char *fmt, ...) { struct va_format vaf; @@ -317,8 +360,9 @@ static int parse_options(struct super_block *sb, char *options) int arg = 0; #ifdef CONFIG_F2FS_FAULT_INJECTION - f2fs_fault_rate = 0; + f2fs_build_fault_attr(0); #endif + if (!options) return 0; @@ -456,8 +500,7 @@ static int parse_options(struct super_block *sb, char *options) if (args->from && match_int(args, &arg)) return -EINVAL; #ifdef CONFIG_F2FS_FAULT_INJECTION - f2fs_fault_rate = arg; - atomic_set(&f2fs_ops, 0); + f2fs_build_fault_attr(arg); #else f2fs_msg(sb, KERN_INFO, "FAULT_INJECTION was not selected"); @@ -1797,6 +1840,16 @@ static int __init init_f2fs_fs(void) err = -ENOMEM; goto free_extent_cache; } +#ifdef CONFIG_F2FS_FAULT_INJECTION + f2fs_fault_inject.kset = f2fs_kset; + f2fs_build_fault_attr(0); + err = kobject_init_and_add(&f2fs_fault_inject, &f2fs_fault_ktype, + NULL, "fault_injection"); + if (err) { + f2fs_fault_inject.kset = NULL; + goto free_kset; + } +#endif err = register_shrinker(&f2fs_shrinker_info); if (err) goto free_kset; @@ -1815,6 +1868,10 @@ free_filesystem: free_shrinker: unregister_shrinker(&f2fs_shrinker_info); free_kset: +#ifdef CONFIG_F2FS_FAULT_INJECTION + if (f2fs_fault_inject.kset) + kobject_put(&f2fs_fault_inject); +#endif kset_unregister(f2fs_kset); free_extent_cache: destroy_extent_cache(); @@ -1841,6 +1898,9 @@ static void __exit exit_f2fs_fs(void) destroy_segment_manager_caches(); destroy_node_manager_caches(); destroy_inodecache(); +#ifdef CONFIG_F2FS_FAULT_INJECTION + kobject_put(&f2fs_fault_inject); +#endif kset_unregister(f2fs_kset); f2fs_destroy_trace_ios(); } -- cgit v1.3-8-gc7d7 From 10aa97c379cdd1e9f537a00ef2d787989759269d Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Mon, 16 May 2016 10:33:40 -0700 Subject: f2fs: manipulate dirty file inodes when DATA_FLUSH is set It needs to maintain dirty file inodes only if DATA_FLUSH is set. Otherwise, let's avoid its overhead. Signed-off-by: Jaegeuk Kim --- fs/f2fs/checkpoint.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c index 3da6499a9e82..d04113b9ac18 100644 --- a/fs/f2fs/checkpoint.c +++ b/fs/f2fs/checkpoint.c @@ -785,9 +785,11 @@ void update_dirty_page(struct inode *inode, struct page *page) !S_ISLNK(inode->i_mode)) return; - spin_lock(&sbi->inode_lock[type]); - __add_dirty_inode(inode, type); - spin_unlock(&sbi->inode_lock[type]); + if (type != FILE_INODE || test_opt(sbi, DATA_FLUSH)) { + spin_lock(&sbi->inode_lock[type]); + __add_dirty_inode(inode, type); + spin_unlock(&sbi->inode_lock[type]); + } inode_inc_dirty_pages(inode); SetPagePrivate(page); @@ -803,6 +805,9 @@ void remove_dirty_inode(struct inode *inode) !S_ISLNK(inode->i_mode)) return; + if (type == FILE_INODE && !test_opt(sbi, DATA_FLUSH)) + return; + spin_lock(&sbi->inode_lock[type]); __remove_dirty_inode(inode, type); spin_unlock(&sbi->inode_lock[type]); -- cgit v1.3-8-gc7d7 From e4b2749158631e6d74bf14d2ef3804d780e2f770 Mon Sep 17 00:00:00 2001 From: NeilBrown Date: Wed, 11 May 2016 11:58:47 +0200 Subject: DAX: move RADIX_DAX_ definitions to dax.c These don't belong in radix-tree.c any more than PAGECACHE_TAG_* do. Let's try to maintain the idea that radix-tree simply implements an abstract data type. Acked-by: Ross Zwisler Reviewed-by: Matthew Wilcox Signed-off-by: NeilBrown Signed-off-by: Jan Kara Signed-off-by: Vishal Verma --- fs/dax.c | 9 +++++++++ include/linux/radix-tree.h | 9 --------- 2 files changed, 9 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/dax.c b/fs/dax.c index 2494255c5785..7ef5aef78442 100644 --- a/fs/dax.c +++ b/fs/dax.c @@ -32,6 +32,15 @@ #include #include +#define RADIX_DAX_MASK 0xf +#define RADIX_DAX_SHIFT 4 +#define RADIX_DAX_PTE (0x4 | RADIX_TREE_EXCEPTIONAL_ENTRY) +#define RADIX_DAX_PMD (0x8 | RADIX_TREE_EXCEPTIONAL_ENTRY) +#define RADIX_DAX_TYPE(entry) ((unsigned long)entry & RADIX_DAX_MASK) +#define RADIX_DAX_SECTOR(entry) (((unsigned long)entry >> RADIX_DAX_SHIFT)) +#define RADIX_DAX_ENTRY(sector, pmd) ((void *)((unsigned long)sector << \ + RADIX_DAX_SHIFT | (pmd ? RADIX_DAX_PMD : RADIX_DAX_PTE))) + static long dax_map_atomic(struct block_device *bdev, struct blk_dax_ctl *dax) { struct request_queue *q = bdev->bd_queue; diff --git a/include/linux/radix-tree.h b/include/linux/radix-tree.h index 51a97ac8bfbf..d08d6ec3bf53 100644 --- a/include/linux/radix-tree.h +++ b/include/linux/radix-tree.h @@ -52,15 +52,6 @@ #define RADIX_TREE_EXCEPTIONAL_ENTRY 2 #define RADIX_TREE_EXCEPTIONAL_SHIFT 2 -#define RADIX_DAX_MASK 0xf -#define RADIX_DAX_SHIFT 4 -#define RADIX_DAX_PTE (0x4 | RADIX_TREE_EXCEPTIONAL_ENTRY) -#define RADIX_DAX_PMD (0x8 | RADIX_TREE_EXCEPTIONAL_ENTRY) -#define RADIX_DAX_TYPE(entry) ((unsigned long)entry & RADIX_DAX_MASK) -#define RADIX_DAX_SECTOR(entry) (((unsigned long)entry >> RADIX_DAX_SHIFT)) -#define RADIX_DAX_ENTRY(sector, pmd) ((void *)((unsigned long)sector << \ - RADIX_DAX_SHIFT | (pmd ? RADIX_DAX_PMD : RADIX_DAX_PTE))) - static inline int radix_tree_is_indirect_ptr(void *ptr) { return (int)((unsigned long)ptr & RADIX_TREE_INDIRECT_PTR); -- cgit v1.3-8-gc7d7 From 02fbd139759feb1f331cebd858523b5d774082e6 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Wed, 11 May 2016 11:58:48 +0200 Subject: dax: Remove complete_unwritten argument Fault handlers currently take complete_unwritten argument to convert unwritten extents after PTEs are updated. However no filesystem uses this anymore as the code is racy. Remove the unused argument. Reviewed-by: Ross Zwisler Signed-off-by: Jan Kara Signed-off-by: Vishal Verma --- fs/block_dev.c | 4 ++-- fs/dax.c | 43 +++++++++---------------------------------- fs/ext2/file.c | 4 ++-- fs/ext4/file.c | 4 ++-- fs/xfs/xfs_file.c | 7 +++---- include/linux/dax.h | 17 +++++++---------- include/linux/fs.h | 1 - 7 files changed, 25 insertions(+), 55 deletions(-) (limited to 'fs') diff --git a/fs/block_dev.c b/fs/block_dev.c index 20a2c02b77c4..b25bb230b28a 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -1746,7 +1746,7 @@ static const struct address_space_operations def_blk_aops = { */ static int blkdev_dax_fault(struct vm_area_struct *vma, struct vm_fault *vmf) { - return __dax_fault(vma, vmf, blkdev_get_block, NULL); + return __dax_fault(vma, vmf, blkdev_get_block); } static int blkdev_dax_pfn_mkwrite(struct vm_area_struct *vma, @@ -1758,7 +1758,7 @@ static int blkdev_dax_pfn_mkwrite(struct vm_area_struct *vma, static int blkdev_dax_pmd_fault(struct vm_area_struct *vma, unsigned long addr, pmd_t *pmd, unsigned int flags) { - return __dax_pmd_fault(vma, addr, pmd, flags, blkdev_get_block, NULL); + return __dax_pmd_fault(vma, addr, pmd, flags, blkdev_get_block); } static const struct vm_operations_struct blkdev_dax_vm_ops = { diff --git a/fs/dax.c b/fs/dax.c index 7ef5aef78442..83e7894d86d8 100644 --- a/fs/dax.c +++ b/fs/dax.c @@ -612,19 +612,13 @@ static int dax_insert_mapping(struct inode *inode, struct buffer_head *bh, * @vma: The virtual memory area where the fault occurred * @vmf: The description of the fault * @get_block: The filesystem method used to translate file offsets to blocks - * @complete_unwritten: The filesystem method used to convert unwritten blocks - * to written so the data written to them is exposed. This is required for - * required by write faults for filesystems that will return unwritten - * extent mappings from @get_block, but it is optional for reads as - * dax_insert_mapping() will always zero unwritten blocks. If the fs does - * not support unwritten extents, the it should pass NULL. * * When a page fault occurs, filesystems may call this helper in their * fault handler for DAX files. __dax_fault() assumes the caller has done all * the necessary locking for the page fault to proceed successfully. */ int __dax_fault(struct vm_area_struct *vma, struct vm_fault *vmf, - get_block_t get_block, dax_iodone_t complete_unwritten) + get_block_t get_block) { struct file *file = vma->vm_file; struct address_space *mapping = file->f_mapping; @@ -727,23 +721,9 @@ int __dax_fault(struct vm_area_struct *vma, struct vm_fault *vmf, page = NULL; } - /* - * If we successfully insert the new mapping over an unwritten extent, - * we need to ensure we convert the unwritten extent. If there is an - * error inserting the mapping, the filesystem needs to leave it as - * unwritten to prevent exposure of the stale underlying data to - * userspace, but we still need to call the completion function so - * the private resources on the mapping buffer can be released. We - * indicate what the callback should do via the uptodate variable, same - * as for normal BH based IO completions. - */ + /* Filesystem should not return unwritten buffers to us! */ + WARN_ON_ONCE(buffer_unwritten(&bh)); error = dax_insert_mapping(inode, &bh, vma, vmf); - if (buffer_unwritten(&bh)) { - if (complete_unwritten) - complete_unwritten(&bh, !error); - else - WARN_ON_ONCE(!(vmf->flags & FAULT_FLAG_WRITE)); - } out: if (error == -ENOMEM) @@ -772,7 +752,7 @@ EXPORT_SYMBOL(__dax_fault); * fault handler for DAX files. */ int dax_fault(struct vm_area_struct *vma, struct vm_fault *vmf, - get_block_t get_block, dax_iodone_t complete_unwritten) + get_block_t get_block) { int result; struct super_block *sb = file_inode(vma->vm_file)->i_sb; @@ -781,7 +761,7 @@ int dax_fault(struct vm_area_struct *vma, struct vm_fault *vmf, sb_start_pagefault(sb); file_update_time(vma->vm_file); } - result = __dax_fault(vma, vmf, get_block, complete_unwritten); + result = __dax_fault(vma, vmf, get_block); if (vmf->flags & FAULT_FLAG_WRITE) sb_end_pagefault(sb); @@ -815,8 +795,7 @@ static void __dax_dbg(struct buffer_head *bh, unsigned long address, #define dax_pmd_dbg(bh, address, reason) __dax_dbg(bh, address, reason, "dax_pmd") int __dax_pmd_fault(struct vm_area_struct *vma, unsigned long address, - pmd_t *pmd, unsigned int flags, get_block_t get_block, - dax_iodone_t complete_unwritten) + pmd_t *pmd, unsigned int flags, get_block_t get_block) { struct file *file = vma->vm_file; struct address_space *mapping = file->f_mapping; @@ -875,6 +854,7 @@ int __dax_pmd_fault(struct vm_area_struct *vma, unsigned long address, if (get_block(inode, block, &bh, 1) != 0) return VM_FAULT_SIGBUS; alloc = true; + WARN_ON_ONCE(buffer_unwritten(&bh)); } bdev = bh.b_bdev; @@ -1020,9 +1000,6 @@ int __dax_pmd_fault(struct vm_area_struct *vma, unsigned long address, out: i_mmap_unlock_read(mapping); - if (buffer_unwritten(&bh)) - complete_unwritten(&bh, !(result & VM_FAULT_ERROR)); - return result; fallback: @@ -1042,8 +1019,7 @@ EXPORT_SYMBOL_GPL(__dax_pmd_fault); * pmd_fault handler for DAX files. */ int dax_pmd_fault(struct vm_area_struct *vma, unsigned long address, - pmd_t *pmd, unsigned int flags, get_block_t get_block, - dax_iodone_t complete_unwritten) + pmd_t *pmd, unsigned int flags, get_block_t get_block) { int result; struct super_block *sb = file_inode(vma->vm_file)->i_sb; @@ -1052,8 +1028,7 @@ int dax_pmd_fault(struct vm_area_struct *vma, unsigned long address, sb_start_pagefault(sb); file_update_time(vma->vm_file); } - result = __dax_pmd_fault(vma, address, pmd, flags, get_block, - complete_unwritten); + result = __dax_pmd_fault(vma, address, pmd, flags, get_block); if (flags & FAULT_FLAG_WRITE) sb_end_pagefault(sb); diff --git a/fs/ext2/file.c b/fs/ext2/file.c index c1400b109805..868c02317b05 100644 --- a/fs/ext2/file.c +++ b/fs/ext2/file.c @@ -51,7 +51,7 @@ static int ext2_dax_fault(struct vm_area_struct *vma, struct vm_fault *vmf) } down_read(&ei->dax_sem); - ret = __dax_fault(vma, vmf, ext2_get_block, NULL); + ret = __dax_fault(vma, vmf, ext2_get_block); up_read(&ei->dax_sem); if (vmf->flags & FAULT_FLAG_WRITE) @@ -72,7 +72,7 @@ static int ext2_dax_pmd_fault(struct vm_area_struct *vma, unsigned long addr, } down_read(&ei->dax_sem); - ret = __dax_pmd_fault(vma, addr, pmd, flags, ext2_get_block, NULL); + ret = __dax_pmd_fault(vma, addr, pmd, flags, ext2_get_block); up_read(&ei->dax_sem); if (flags & FAULT_FLAG_WRITE) diff --git a/fs/ext4/file.c b/fs/ext4/file.c index 37e28082885a..7a6398867ff2 100644 --- a/fs/ext4/file.c +++ b/fs/ext4/file.c @@ -207,7 +207,7 @@ static int ext4_dax_fault(struct vm_area_struct *vma, struct vm_fault *vmf) if (IS_ERR(handle)) result = VM_FAULT_SIGBUS; else - result = __dax_fault(vma, vmf, ext4_dax_get_block, NULL); + result = __dax_fault(vma, vmf, ext4_dax_get_block); if (write) { if (!IS_ERR(handle)) @@ -243,7 +243,7 @@ static int ext4_dax_pmd_fault(struct vm_area_struct *vma, unsigned long addr, result = VM_FAULT_SIGBUS; else result = __dax_pmd_fault(vma, addr, pmd, flags, - ext4_dax_get_block, NULL); + ext4_dax_get_block); if (write) { if (!IS_ERR(handle)) diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c index 569938a4a357..c2946f436a3a 100644 --- a/fs/xfs/xfs_file.c +++ b/fs/xfs/xfs_file.c @@ -1558,7 +1558,7 @@ xfs_filemap_page_mkwrite( xfs_ilock(XFS_I(inode), XFS_MMAPLOCK_SHARED); if (IS_DAX(inode)) { - ret = __dax_mkwrite(vma, vmf, xfs_get_blocks_dax_fault, NULL); + ret = __dax_mkwrite(vma, vmf, xfs_get_blocks_dax_fault); } else { ret = block_page_mkwrite(vma, vmf, xfs_get_blocks); ret = block_page_mkwrite_return(ret); @@ -1592,7 +1592,7 @@ xfs_filemap_fault( * changes to xfs_get_blocks_direct() to map unwritten extent * ioend for conversion on read-only mappings. */ - ret = __dax_fault(vma, vmf, xfs_get_blocks_dax_fault, NULL); + ret = __dax_fault(vma, vmf, xfs_get_blocks_dax_fault); } else ret = filemap_fault(vma, vmf); xfs_iunlock(XFS_I(inode), XFS_MMAPLOCK_SHARED); @@ -1629,8 +1629,7 @@ xfs_filemap_pmd_fault( } xfs_ilock(XFS_I(inode), XFS_MMAPLOCK_SHARED); - ret = __dax_pmd_fault(vma, addr, pmd, flags, xfs_get_blocks_dax_fault, - NULL); + ret = __dax_pmd_fault(vma, addr, pmd, flags, xfs_get_blocks_dax_fault); xfs_iunlock(XFS_I(inode), XFS_MMAPLOCK_SHARED); if (flags & FAULT_FLAG_WRITE) diff --git a/include/linux/dax.h b/include/linux/dax.h index 636dd59ab505..7c45ac7ea1d1 100644 --- a/include/linux/dax.h +++ b/include/linux/dax.h @@ -10,10 +10,8 @@ ssize_t dax_do_io(struct kiocb *, struct inode *, struct iov_iter *, loff_t, int dax_clear_sectors(struct block_device *bdev, sector_t _sector, long _size); int dax_zero_page_range(struct inode *, loff_t from, unsigned len, get_block_t); int dax_truncate_page(struct inode *, loff_t from, get_block_t); -int dax_fault(struct vm_area_struct *, struct vm_fault *, get_block_t, - dax_iodone_t); -int __dax_fault(struct vm_area_struct *, struct vm_fault *, get_block_t, - dax_iodone_t); +int dax_fault(struct vm_area_struct *, struct vm_fault *, get_block_t); +int __dax_fault(struct vm_area_struct *, struct vm_fault *, get_block_t); #ifdef CONFIG_FS_DAX struct page *read_dax_sector(struct block_device *bdev, sector_t n); @@ -27,21 +25,20 @@ static inline struct page *read_dax_sector(struct block_device *bdev, #ifdef CONFIG_TRANSPARENT_HUGEPAGE int dax_pmd_fault(struct vm_area_struct *, unsigned long addr, pmd_t *, - unsigned int flags, get_block_t, dax_iodone_t); + unsigned int flags, get_block_t); int __dax_pmd_fault(struct vm_area_struct *, unsigned long addr, pmd_t *, - unsigned int flags, get_block_t, dax_iodone_t); + unsigned int flags, get_block_t); #else static inline int dax_pmd_fault(struct vm_area_struct *vma, unsigned long addr, - pmd_t *pmd, unsigned int flags, get_block_t gb, - dax_iodone_t di) + pmd_t *pmd, unsigned int flags, get_block_t gb) { return VM_FAULT_FALLBACK; } #define __dax_pmd_fault dax_pmd_fault #endif int dax_pfn_mkwrite(struct vm_area_struct *, struct vm_fault *); -#define dax_mkwrite(vma, vmf, gb, iod) dax_fault(vma, vmf, gb, iod) -#define __dax_mkwrite(vma, vmf, gb, iod) __dax_fault(vma, vmf, gb, iod) +#define dax_mkwrite(vma, vmf, gb) dax_fault(vma, vmf, gb) +#define __dax_mkwrite(vma, vmf, gb) __dax_fault(vma, vmf, gb) static inline bool vma_is_dax(struct vm_area_struct *vma) { diff --git a/include/linux/fs.h b/include/linux/fs.h index 70e61b58baaf..9f2813090d1b 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -74,7 +74,6 @@ typedef int (get_block_t)(struct inode *inode, sector_t iblock, struct buffer_head *bh_result, int create); typedef int (dio_iodone_t)(struct kiocb *iocb, loff_t offset, ssize_t bytes, void *private); -typedef void (dax_iodone_t)(struct buffer_head *bh_map, int uptodate); #define MAY_EXEC 0x00000001 #define MAY_WRITE 0x00000002 -- cgit v1.3-8-gc7d7 From 9b6cd5f76d60b563d75e55e432e03ed134761432 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 16 May 2016 17:17:04 +0200 Subject: ext2: Fix block zeroing in ext2_get_blocks() for DAX When zeroing allocated blocks for DAX, we accidentally zeroed only the first allocated block instead of all of them. So far this problem is hidden by the fact that page faults always need only a single block and DAX write code zeroes blocks again. But the zeroing in DAX code is racy and needs to be removed so fix the zeroing in ext2 to zero all allocated blocks. Reported-by: Ross Zwisler Signed-off-by: Jan Kara Signed-off-by: Vishal Verma --- fs/ext2/inode.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ext2/inode.c b/fs/ext2/inode.c index 6bd58e6ff038..038d0ed5f565 100644 --- a/fs/ext2/inode.c +++ b/fs/ext2/inode.c @@ -740,7 +740,7 @@ static int ext2_get_blocks(struct inode *inode, err = dax_clear_sectors(inode->i_sb->s_bdev, le32_to_cpu(chain[depth-1].key) << (inode->i_blkbits - 9), - 1 << inode->i_blkbits); + count << inode->i_blkbits); if (err) { mutex_unlock(&ei->truncate_mutex); goto cleanup; -- cgit v1.3-8-gc7d7 From 86b0624e42d03a424e9571b8591d191c436f9af1 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Wed, 11 May 2016 11:58:49 +0200 Subject: ext2: Avoid DAX zeroing to corrupt data Currently ext2 zeroes any data blocks allocated for DAX inode however it still returns them as BH_New. Thus DAX code zeroes them again in dax_insert_mapping() which can possibly overwrite the data that has been already stored to those blocks by a racing dax_io(). Avoid marking pre-zeroed buffers as new. Reviewed-by: Ross Zwisler Signed-off-by: Jan Kara Signed-off-by: Vishal Verma --- fs/ext2/inode.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/ext2/inode.c b/fs/ext2/inode.c index 038d0ed5f565..9a14af3b1a69 100644 --- a/fs/ext2/inode.c +++ b/fs/ext2/inode.c @@ -745,11 +745,11 @@ static int ext2_get_blocks(struct inode *inode, mutex_unlock(&ei->truncate_mutex); goto cleanup; } - } + } else + set_buffer_new(bh_result); ext2_splice_branch(inode, iblock, partial, indirect_blks, count); mutex_unlock(&ei->truncate_mutex); - set_buffer_new(bh_result); got_it: map_bh(bh_result, inode->i_sb, le32_to_cpu(chain[depth-1].key)); if (count > blocks_to_boundary) -- cgit v1.3-8-gc7d7 From 2b10945c532c165a824f541df76a15ed0be04d78 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Wed, 11 May 2016 11:58:50 +0200 Subject: dax: Remove dead zeroing code from fault handlers Now that all filesystems zero out blocks allocated for a fault handler, we can just remove the zeroing from the handler itself. Also add checks that no filesystem returns to us unwritten or new buffer. Reviewed-by: Ross Zwisler Signed-off-by: Jan Kara Signed-off-by: Vishal Verma --- fs/dax.c | 17 ++--------------- 1 file changed, 2 insertions(+), 15 deletions(-) (limited to 'fs') diff --git a/fs/dax.c b/fs/dax.c index 83e7894d86d8..ccb8bc399d78 100644 --- a/fs/dax.c +++ b/fs/dax.c @@ -587,11 +587,6 @@ static int dax_insert_mapping(struct inode *inode, struct buffer_head *bh, error = PTR_ERR(dax.addr); goto out; } - - if (buffer_unwritten(bh) || buffer_new(bh)) { - clear_pmem(dax.addr, PAGE_SIZE); - wmb_pmem(); - } dax_unmap_atomic(bdev, &dax); error = dax_radix_entry(mapping, vmf->pgoff, dax.sector, false, @@ -722,7 +717,7 @@ int __dax_fault(struct vm_area_struct *vma, struct vm_fault *vmf, } /* Filesystem should not return unwritten buffers to us! */ - WARN_ON_ONCE(buffer_unwritten(&bh)); + WARN_ON_ONCE(buffer_unwritten(&bh) || buffer_new(&bh)); error = dax_insert_mapping(inode, &bh, vma, vmf); out: @@ -854,7 +849,7 @@ int __dax_pmd_fault(struct vm_area_struct *vma, unsigned long address, if (get_block(inode, block, &bh, 1) != 0) return VM_FAULT_SIGBUS; alloc = true; - WARN_ON_ONCE(buffer_unwritten(&bh)); + WARN_ON_ONCE(buffer_unwritten(&bh) || buffer_new(&bh)); } bdev = bh.b_bdev; @@ -953,14 +948,6 @@ int __dax_pmd_fault(struct vm_area_struct *vma, unsigned long address, dax_pmd_dbg(&bh, address, "pfn not in memmap"); goto fallback; } - - if (buffer_unwritten(&bh) || buffer_new(&bh)) { - clear_pmem(dax.addr, PMD_SIZE); - wmb_pmem(); - count_vm_event(PGMAJFAULT); - mem_cgroup_count_vm_event(vma->vm_mm, PGMAJFAULT); - result |= VM_FAULT_MAJOR; - } dax_unmap_atomic(bdev, &dax); /* -- cgit v1.3-8-gc7d7 From 069c77bc9eaee70fa9ecbd942372b1693b8cdeb0 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Wed, 11 May 2016 11:58:51 +0200 Subject: dax: Remove zeroing from dax_io() All the filesystems are now zeroing blocks themselves for DAX IO to avoid races between dax_io() and dax_fault(). Remove the zeroing code from dax_io() and add warning to catch the case when somebody unexpectedly returns new or unwritten buffer. Reviewed-by: Ross Zwisler Signed-off-by: Jan Kara Signed-off-by: Vishal Verma --- fs/dax.c | 28 ++++++++++------------------ 1 file changed, 10 insertions(+), 18 deletions(-) (limited to 'fs') diff --git a/fs/dax.c b/fs/dax.c index ccb8bc399d78..7c0036dd1570 100644 --- a/fs/dax.c +++ b/fs/dax.c @@ -119,18 +119,6 @@ int dax_clear_sectors(struct block_device *bdev, sector_t _sector, long _size) } EXPORT_SYMBOL_GPL(dax_clear_sectors); -/* the clear_pmem() calls are ordered by a wmb_pmem() in the caller */ -static void dax_new_buf(void __pmem *addr, unsigned size, unsigned first, - loff_t pos, loff_t end) -{ - loff_t final = end - pos + first; /* The final byte of the buffer */ - - if (first > 0) - clear_pmem(addr, first); - if (final < size) - clear_pmem(addr + final, size - final); -} - static bool buffer_written(struct buffer_head *bh) { return buffer_mapped(bh) && !buffer_unwritten(bh); @@ -169,6 +157,9 @@ static ssize_t dax_io(struct inode *inode, struct iov_iter *iter, struct blk_dax_ctl dax = { .addr = (void __pmem *) ERR_PTR(-EIO), }; + unsigned blkbits = inode->i_blkbits; + sector_t file_blks = (i_size_read(inode) + (1 << blkbits) - 1) + >> blkbits; if (rw == READ) end = min(end, i_size_read(inode)); @@ -176,7 +167,6 @@ static ssize_t dax_io(struct inode *inode, struct iov_iter *iter, while (pos < end) { size_t len; if (pos == max) { - unsigned blkbits = inode->i_blkbits; long page = pos >> PAGE_SHIFT; sector_t block = page << (PAGE_SHIFT - blkbits); unsigned first = pos - (block << blkbits); @@ -192,6 +182,13 @@ static ssize_t dax_io(struct inode *inode, struct iov_iter *iter, bh->b_size = 1 << blkbits; bh_max = pos - first + bh->b_size; bdev = bh->b_bdev; + /* + * We allow uninitialized buffers for writes + * beyond EOF as those cannot race with faults + */ + WARN_ON_ONCE( + (buffer_new(bh) && block < file_blks) || + (rw == WRITE && buffer_unwritten(bh))); } else { unsigned done = bh->b_size - (bh_max - (pos - first)); @@ -211,11 +208,6 @@ static ssize_t dax_io(struct inode *inode, struct iov_iter *iter, rc = map_len; break; } - if (buffer_unwritten(bh) || buffer_new(bh)) { - dax_new_buf(dax.addr, map_len, first, - pos, end); - need_wmb = true; - } dax.addr += first; size = map_len - first; } -- cgit v1.3-8-gc7d7 From c3d98e39d5b37320b15f227686575d58f676e6ef Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Wed, 11 May 2016 11:58:52 +0200 Subject: dax: Remove pointless writeback from dax_do_io() dax_do_io() is calling filemap_write_and_wait() if DIO_LOCKING flags is set. Presumably this was copied over from direct IO code. However DAX inodes have no pagecache pages to write so the call is pointless. Remove it. Reviewed-by: Ross Zwisler Signed-off-by: Jan Kara Signed-off-by: Vishal Verma --- fs/dax.c | 10 +--------- 1 file changed, 1 insertion(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/dax.c b/fs/dax.c index 7c0036dd1570..237581441bc1 100644 --- a/fs/dax.c +++ b/fs/dax.c @@ -268,15 +268,8 @@ ssize_t dax_do_io(struct kiocb *iocb, struct inode *inode, memset(&bh, 0, sizeof(bh)); bh.b_bdev = inode->i_sb->s_bdev; - if ((flags & DIO_LOCKING) && iov_iter_rw(iter) == READ) { - struct address_space *mapping = inode->i_mapping; + if ((flags & DIO_LOCKING) && iov_iter_rw(iter) == READ) inode_lock(inode); - retval = filemap_write_and_wait_range(mapping, pos, end - 1); - if (retval) { - inode_unlock(inode); - goto out; - } - } /* Protects against truncate */ if (!(flags & DIO_SKIP_DIO_COUNT)) @@ -297,7 +290,6 @@ ssize_t dax_do_io(struct kiocb *iocb, struct inode *inode, if (!(flags & DIO_SKIP_DIO_COUNT)) inode_dio_end(inode); - out: return retval; } EXPORT_SYMBOL_GPL(dax_do_io); -- cgit v1.3-8-gc7d7 From 7795bec89ebf927ea3ad9ed5f396c227e5c73271 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Wed, 11 May 2016 11:58:53 +0200 Subject: dax: Remove redundant inode size checks Callers of dax fault handlers must make sure these calls cannot race with truncate. Thus it is enough to check inode size when entering the function and we don't have to recheck it again later in the handler. Note that inode size itself can be decreased while the fault handler runs but filesystem locking prevents against any radix tree or block mapping information changes resulting from the truncate and that is what we really care about. Reviewed-by: Ross Zwisler Signed-off-by: Jan Kara Signed-off-by: Vishal Verma --- fs/dax.c | 60 +----------------------------------------------------------- 1 file changed, 1 insertion(+), 59 deletions(-) (limited to 'fs') diff --git a/fs/dax.c b/fs/dax.c index 237581441bc1..9bc6624251b4 100644 --- a/fs/dax.c +++ b/fs/dax.c @@ -305,20 +305,11 @@ EXPORT_SYMBOL_GPL(dax_do_io); static int dax_load_hole(struct address_space *mapping, struct page *page, struct vm_fault *vmf) { - unsigned long size; - struct inode *inode = mapping->host; if (!page) page = find_or_create_page(mapping, vmf->pgoff, GFP_KERNEL | __GFP_ZERO); if (!page) return VM_FAULT_OOM; - /* Recheck i_size under page lock to avoid truncate race */ - size = (i_size_read(inode) + PAGE_SIZE - 1) >> PAGE_SHIFT; - if (vmf->pgoff >= size) { - unlock_page(page); - put_page(page); - return VM_FAULT_SIGBUS; - } vmf->page = page; return VM_FAULT_LOCKED; @@ -549,24 +540,10 @@ static int dax_insert_mapping(struct inode *inode, struct buffer_head *bh, .sector = to_sector(bh, inode), .size = bh->b_size, }; - pgoff_t size; int error; i_mmap_lock_read(mapping); - /* - * Check truncate didn't happen while we were allocating a block. - * If it did, this block may or may not be still allocated to the - * file. We can't tell the filesystem to free it because we can't - * take i_mutex here. In the worst case, the file still has blocks - * allocated past the end of the file. - */ - size = (i_size_read(inode) + PAGE_SIZE - 1) >> PAGE_SHIFT; - if (unlikely(vmf->pgoff >= size)) { - error = -EIO; - goto out; - } - if (dax_map_atomic(bdev, &dax) < 0) { error = PTR_ERR(dax.addr); goto out; @@ -632,15 +609,6 @@ int __dax_fault(struct vm_area_struct *vma, struct vm_fault *vmf, put_page(page); goto repeat; } - size = (i_size_read(inode) + PAGE_SIZE - 1) >> PAGE_SHIFT; - if (unlikely(vmf->pgoff >= size)) { - /* - * We have a struct page covering a hole in the file - * from a read fault and we've raced with a truncate - */ - error = -EIO; - goto unlock_page; - } } error = get_block(inode, block, &bh, 0); @@ -673,17 +641,8 @@ int __dax_fault(struct vm_area_struct *vma, struct vm_fault *vmf, if (error) goto unlock_page; vmf->page = page; - if (!page) { + if (!page) i_mmap_lock_read(mapping); - /* Check we didn't race with truncate */ - size = (i_size_read(inode) + PAGE_SIZE - 1) >> - PAGE_SHIFT; - if (vmf->pgoff >= size) { - i_mmap_unlock_read(mapping); - error = -EIO; - goto out; - } - } return VM_FAULT_LOCKED; } @@ -861,23 +820,6 @@ int __dax_pmd_fault(struct vm_area_struct *vma, unsigned long address, i_mmap_lock_read(mapping); - /* - * If a truncate happened while we were allocating blocks, we may - * leave blocks allocated to the file that are beyond EOF. We can't - * take i_mutex here, so just leave them hanging; they'll be freed - * when the file is deleted. - */ - size = (i_size_read(inode) + PAGE_SIZE - 1) >> PAGE_SHIFT; - if (pgoff >= size) { - result = VM_FAULT_SIGBUS; - goto out; - } - if ((pgoff | PG_PMD_COLOUR) >= size) { - dax_pmd_dbg(&bh, address, - "offset + huge page size > file size"); - goto fallback; - } - if (!write && !buffer_mapped(&bh) && buffer_uptodate(&bh)) { spinlock_t *ptl; pmd_t entry; -- cgit v1.3-8-gc7d7 From 2af3a8159cd204fc8437ed2f75863f0fb930f0d0 Mon Sep 17 00:00:00 2001 From: Toshi Kani Date: Tue, 10 May 2016 10:23:52 -0600 Subject: block: Add vfs_msg() interface In preparation of moving DAX capability checks to the block layer from filesystem code, add a VFS message interface that aligns with filesystem's message format. For instance, a vfs_msg() message followed by XFS messages in case of a dax mount error may look like: VFS (pmem0p1): error: unaligned partition for dax XFS (pmem0p1): DAX unsupported by block device. Turning off DAX. XFS (pmem0p1): Mounting V5 Filesystem : vfs_msg() is largely based on ext4_msg(). Signed-off-by: Toshi Kani Reviewed-by: Christoph Hellwig Cc: Alexander Viro Cc: Jens Axboe Cc: "Theodore Ts'o" Cc: Andreas Dilger Cc: Jan Kara Cc: Dave Chinner Cc: Dan Williams Cc: Ross Zwisler Cc: Christoph Hellwig Cc: Boaz Harrosh Signed-off-by: Vishal Verma --- fs/block_dev.c | 12 ++++++++++++ include/linux/blkdev.h | 11 +++++++++++ 2 files changed, 23 insertions(+) (limited to 'fs') diff --git a/fs/block_dev.c b/fs/block_dev.c index b25bb230b28a..91e0ec0233c0 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -50,6 +50,18 @@ struct block_device *I_BDEV(struct inode *inode) } EXPORT_SYMBOL(I_BDEV); +void __vfs_msg(struct super_block *sb, const char *prefix, const char *fmt, ...) +{ + struct va_format vaf; + va_list args; + + va_start(args, fmt); + vaf.fmt = fmt; + vaf.va = &args; + printk_ratelimited("%sVFS (%s): %pV\n", prefix, sb->s_id, &vaf); + va_end(args); +} + static void bdev_write_inode(struct block_device *bdev) { struct inode *inode = bdev->bd_inode; diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h index 669e419d6234..78c48ab22f46 100644 --- a/include/linux/blkdev.h +++ b/include/linux/blkdev.h @@ -767,6 +767,17 @@ static inline void rq_flush_dcache_pages(struct request *rq) } #endif +#ifdef CONFIG_PRINTK +#define vfs_msg(sb, level, fmt, ...) \ + __vfs_msg(sb, level, fmt, ##__VA_ARGS__) +#else +#define vfs_msg(sb, level, fmt, ...) \ +do { \ + no_printk(fmt, ##__VA_ARGS__); \ + __vfs_msg(sb, "", " "); \ +} while (0) +#endif + extern int blk_register_queue(struct gendisk *disk); extern void blk_unregister_queue(struct gendisk *disk); extern blk_qc_t generic_make_request(struct bio *bio); -- cgit v1.3-8-gc7d7 From 2d96afc8f70ef86c66a0b5d80c24a27d6dd13df3 Mon Sep 17 00:00:00 2001 From: Toshi Kani Date: Tue, 10 May 2016 10:23:53 -0600 Subject: block: Add bdev_dax_supported() for dax mount checks DAX imposes additional requirements to a device. Add bdev_dax_supported() which performs all the precondition checks necessary for filesystem to mount the device with dax option. Also add a new check to verify if a partition is aligned by 4KB. When a partition is unaligned, any dax read/write access fails, except for metadata update. Signed-off-by: Toshi Kani Reviewed-by: Christoph Hellwig Cc: Alexander Viro Cc: Jens Axboe Cc: "Theodore Ts'o" Cc: Andreas Dilger Cc: Jan Kara Cc: Dave Chinner Cc: Dan Williams Cc: Ross Zwisler Cc: Christoph Hellwig Cc: Boaz Harrosh Signed-off-by: Vishal Verma --- fs/block_dev.c | 45 +++++++++++++++++++++++++++++++++++++++++++++ include/linux/blkdev.h | 1 + 2 files changed, 46 insertions(+) (limited to 'fs') diff --git a/fs/block_dev.c b/fs/block_dev.c index 91e0ec0233c0..518cde62c01c 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -509,6 +509,51 @@ long bdev_direct_access(struct block_device *bdev, struct blk_dax_ctl *dax) } EXPORT_SYMBOL_GPL(bdev_direct_access); +/** + * bdev_dax_supported() - Check if the device supports dax for filesystem + * @sb: The superblock of the device + * @blocksize: The block size of the device + * + * This is a library function for filesystems to check if the block device + * can be mounted with dax option. + * + * Return: negative errno if unsupported, 0 if supported. + */ +int bdev_dax_supported(struct super_block *sb, int blocksize) +{ + struct blk_dax_ctl dax = { + .sector = 0, + .size = PAGE_SIZE, + }; + int err; + + if (blocksize != PAGE_SIZE) { + vfs_msg(sb, KERN_ERR, "error: unsupported blocksize for dax"); + return -EINVAL; + } + + err = bdev_direct_access(sb->s_bdev, &dax); + if (err < 0) { + switch (err) { + case -EOPNOTSUPP: + vfs_msg(sb, KERN_ERR, + "error: device does not support dax"); + break; + case -EINVAL: + vfs_msg(sb, KERN_ERR, + "error: unaligned partition for dax"); + break; + default: + vfs_msg(sb, KERN_ERR, + "error: dax access failed (%d)", err); + } + return err; + } + + return 0; +} +EXPORT_SYMBOL_GPL(bdev_dax_supported); + /* * pseudo-fs */ diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h index 78c48ab22f46..71231a55debd 100644 --- a/include/linux/blkdev.h +++ b/include/linux/blkdev.h @@ -1688,6 +1688,7 @@ extern int bdev_read_page(struct block_device *, sector_t, struct page *); extern int bdev_write_page(struct block_device *, sector_t, struct page *, struct writeback_control *); extern long bdev_direct_access(struct block_device *, struct blk_dax_ctl *); +extern int bdev_dax_supported(struct super_block *, int); #else /* CONFIG_BLOCK */ struct block_device; -- cgit v1.3-8-gc7d7 From 87eefeb4e80b1526be83e670a08d059a003b7d0e Mon Sep 17 00:00:00 2001 From: Toshi Kani Date: Tue, 10 May 2016 10:23:54 -0600 Subject: ext4: Add alignment check for DAX mount When a partition is not aligned by 4KB, mount -o dax succeeds, but any read/write access to the filesystem fails, except for metadata update. Call bdev_dax_supported() to perform proper precondition checks which includes this partition alignment check. Reported-by: Micah Parrish Signed-off-by: Toshi Kani Reviewed-by: Jan Kara Cc: "Theodore Ts'o" Cc: Andreas Dilger Cc: Jan Kara Cc: Dan Williams Cc: Ross Zwisler Cc: Christoph Hellwig Cc: Boaz Harrosh Signed-off-by: Vishal Verma --- fs/ext4/super.c | 11 ++--------- 1 file changed, 2 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 20c5d52253b4..3822a5aedc61 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -3417,16 +3417,9 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent) } if (sbi->s_mount_opt & EXT4_MOUNT_DAX) { - if (blocksize != PAGE_SIZE) { - ext4_msg(sb, KERN_ERR, - "error: unsupported blocksize for dax"); - goto failed_mount; - } - if (!sb->s_bdev->bd_disk->fops->direct_access) { - ext4_msg(sb, KERN_ERR, - "error: device does not support dax"); + err = bdev_dax_supported(sb, blocksize); + if (err) goto failed_mount; - } } if (ext4_has_feature_encrypt(sb) && es->s_encryption_level) { -- cgit v1.3-8-gc7d7 From 284854be2b85e967d17c098a1d4c176b5cd37eb3 Mon Sep 17 00:00:00 2001 From: Toshi Kani Date: Tue, 10 May 2016 10:23:55 -0600 Subject: ext2: Add alignment check for DAX mount When a partition is not aligned by 4KB, mount -o dax succeeds, but any read/write access to the filesystem fails, except for metadata update. Call bdev_dax_supported() to perform proper precondition checks which includes this partition alignment check. Signed-off-by: Toshi Kani Reviewed-by: Jan Kara Cc: Jan Kara Cc: Dan Williams Cc: Ross Zwisler Cc: Christoph Hellwig Cc: Boaz Harrosh Signed-off-by: Vishal Verma --- fs/ext2/super.c | 11 ++--------- 1 file changed, 2 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/ext2/super.c b/fs/ext2/super.c index b78caf25f746..1d9379568aa8 100644 --- a/fs/ext2/super.c +++ b/fs/ext2/super.c @@ -922,16 +922,9 @@ static int ext2_fill_super(struct super_block *sb, void *data, int silent) blocksize = BLOCK_SIZE << le32_to_cpu(sbi->s_es->s_log_block_size); if (sbi->s_mount_opt & EXT2_MOUNT_DAX) { - if (blocksize != PAGE_SIZE) { - ext2_msg(sb, KERN_ERR, - "error: unsupported blocksize for dax"); + err = bdev_dax_supported(sb, blocksize); + if (err) goto failed_mount; - } - if (!sb->s_bdev->bd_disk->fops->direct_access) { - ext2_msg(sb, KERN_ERR, - "error: device does not support dax"); - goto failed_mount; - } } /* If the blocksize doesn't match, re-read the thing.. */ -- cgit v1.3-8-gc7d7 From 1e937cddd1ff4b823fc9d148ebabd443f105211e Mon Sep 17 00:00:00 2001 From: Toshi Kani Date: Tue, 10 May 2016 10:23:56 -0600 Subject: xfs: Add alignment check for DAX mount When a partition is not aligned by 4KB, mount -o dax succeeds, but any read/write access to the filesystem fails, except for metadata update. Call bdev_dax_supported() to perform proper precondition checks which includes this partition alignment check. Signed-off-by: Toshi Kani Reviewed-by: Christoph Hellwig Cc: Dave Chinner Cc: Dan Williams Cc: Ross Zwisler Cc: Christoph Hellwig Cc: Boaz Harrosh Signed-off-by: Vishal Verma --- fs/xfs/xfs_super.c | 12 +++++------- 1 file changed, 5 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index 187e14b696c2..39f4e6ab3d36 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -1558,14 +1558,12 @@ xfs_fs_fill_super( if (mp->m_flags & XFS_MOUNT_DAX) { xfs_warn(mp, - "DAX enabled. Warning: EXPERIMENTAL, use at your own risk"); - if (sb->s_blocksize != PAGE_SIZE) { - xfs_alert(mp, - "Filesystem block size invalid for DAX Turning DAX off."); - mp->m_flags &= ~XFS_MOUNT_DAX; - } else if (!sb->s_bdev->bd_disk->fops->direct_access) { + "DAX enabled. Warning: EXPERIMENTAL, use at your own risk"); + + error = bdev_dax_supported(sb, sb->s_blocksize); + if (error) { xfs_alert(mp, - "Block device does not support DAX Turning DAX off."); + "DAX unsupported by block device. Turning off DAX."); mp->m_flags &= ~XFS_MOUNT_DAX; } } -- cgit v1.3-8-gc7d7 From a8078b1fc616da6112eb95f0063cd34531d4ccf0 Mon Sep 17 00:00:00 2001 From: Toshi Kani Date: Tue, 10 May 2016 10:23:57 -0600 Subject: block: Update blkdev_dax_capable() for consistency blkdev_dax_capable() is similar to bdev_dax_supported(), but needs to remain as a separate interface for checking dax capability of a raw block device. Rename and relocate blkdev_dax_capable() to keep them maintained consistently, and call bdev_direct_access() for the dax capability check. There is no change in the behavior. Link: https://lkml.org/lkml/2016/5/9/950 Signed-off-by: Toshi Kani Reviewed-by: Jan Kara Cc: Alexander Viro Cc: Jens Axboe Cc: Andreas Dilger Cc: Jan Kara Cc: Dave Chinner Cc: Dan Williams Cc: Ross Zwisler Cc: Christoph Hellwig Cc: Boaz Harrosh Signed-off-by: Vishal Verma --- block/ioctl.c | 30 ------------------------------ fs/block_dev.c | 39 +++++++++++++++++++++++++++++++++++++-- include/linux/blkdev.h | 1 + include/linux/fs.h | 8 -------- 4 files changed, 38 insertions(+), 40 deletions(-) (limited to 'fs') diff --git a/block/ioctl.c b/block/ioctl.c index 4ff1f92f89ca..7eeda072dc70 100644 --- a/block/ioctl.c +++ b/block/ioctl.c @@ -4,7 +4,6 @@ #include #include #include -#include #include #include #include @@ -407,35 +406,6 @@ static inline int is_unrecognized_ioctl(int ret) ret == -ENOIOCTLCMD; } -#ifdef CONFIG_FS_DAX -bool blkdev_dax_capable(struct block_device *bdev) -{ - struct gendisk *disk = bdev->bd_disk; - - if (!disk->fops->direct_access) - return false; - - /* - * If the partition is not aligned on a page boundary, we can't - * do dax I/O to it. - */ - if ((bdev->bd_part->start_sect % (PAGE_SIZE / 512)) - || (bdev->bd_part->nr_sects % (PAGE_SIZE / 512))) - return false; - - /* - * If the device has known bad blocks, force all I/O through the - * driver / page cache. - * - * TODO: support finer grained dax error handling - */ - if (disk->bb && disk->bb->count) - return false; - - return true; -} -#endif - static int blkdev_flushbuf(struct block_device *bdev, fmode_t mode, unsigned cmd, unsigned long arg) { diff --git a/fs/block_dev.c b/fs/block_dev.c index 518cde62c01c..8477d4501b1e 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -29,6 +29,7 @@ #include #include #include +#include #include #include "internal.h" @@ -554,6 +555,40 @@ int bdev_dax_supported(struct super_block *sb, int blocksize) } EXPORT_SYMBOL_GPL(bdev_dax_supported); +/** + * bdev_dax_capable() - Return if the raw device is capable for dax + * @bdev: The device for raw block device access + */ +bool bdev_dax_capable(struct block_device *bdev) +{ + struct gendisk *disk = bdev->bd_disk; + struct blk_dax_ctl dax = { + .size = PAGE_SIZE, + }; + + if (!IS_ENABLED(CONFIG_FS_DAX)) + return false; + + dax.sector = 0; + if (bdev_direct_access(bdev, &dax) < 0) + return false; + + dax.sector = bdev->bd_part->nr_sects - (PAGE_SIZE / 512); + if (bdev_direct_access(bdev, &dax) < 0) + return false; + + /* + * If the device has known bad blocks, force all I/O through the + * driver / page cache. + * + * TODO: support finer grained dax error handling + */ + if (disk->bb && disk->bb->count) + return false; + + return true; +} + /* * pseudo-fs */ @@ -1295,7 +1330,7 @@ static int __blkdev_get(struct block_device *bdev, fmode_t mode, int for_part) if (!ret) { bd_set_size(bdev,(loff_t)get_capacity(disk)<<9); - if (!blkdev_dax_capable(bdev)) + if (!bdev_dax_capable(bdev)) bdev->bd_inode->i_flags &= ~S_DAX; } @@ -1332,7 +1367,7 @@ static int __blkdev_get(struct block_device *bdev, fmode_t mode, int for_part) goto out_clear; } bd_set_size(bdev, (loff_t)bdev->bd_part->nr_sects << 9); - if (!blkdev_dax_capable(bdev)) + if (!bdev_dax_capable(bdev)) bdev->bd_inode->i_flags &= ~S_DAX; } } else { diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h index 71231a55debd..27cbefe8c985 100644 --- a/include/linux/blkdev.h +++ b/include/linux/blkdev.h @@ -1689,6 +1689,7 @@ extern int bdev_write_page(struct block_device *, sector_t, struct page *, struct writeback_control *); extern long bdev_direct_access(struct block_device *, struct blk_dax_ctl *); extern int bdev_dax_supported(struct super_block *, int); +extern bool bdev_dax_capable(struct block_device *); #else /* CONFIG_BLOCK */ struct block_device; diff --git a/include/linux/fs.h b/include/linux/fs.h index 9f2813090d1b..17f934fcf564 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -2319,14 +2319,6 @@ extern struct super_block *freeze_bdev(struct block_device *); extern void emergency_thaw_all(void); extern int thaw_bdev(struct block_device *bdev, struct super_block *sb); extern int fsync_bdev(struct block_device *); -#ifdef CONFIG_FS_DAX -extern bool blkdev_dax_capable(struct block_device *bdev); -#else -static inline bool blkdev_dax_capable(struct block_device *bdev) -{ - return false; -} -#endif extern struct super_block *blockdev_superblock; -- cgit v1.3-8-gc7d7 From 8b3db9798c689ba2dc509f6ca1a8c7e1c1a3e62a Mon Sep 17 00:00:00 2001 From: Dan Williams Date: Wed, 24 Feb 2016 14:02:06 -0800 Subject: dax: fallback from pmd to pte on error In preparation for consulting a badblocks list in pmem_direct_access(), teach dax_pmd_fault() to fallback rather than fail immediately upon encountering an error. The thought being that reducing the span of the dax request may avoid the error region. Reviewed-by: Jeff Moyer Reviewed-by: Christoph Hellwig Reviewed-by: Jan Kara Signed-off-by: Dan Williams Signed-off-by: Vishal Verma --- fs/dax.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/dax.c b/fs/dax.c index 9bc6624251b4..d602410d8e52 100644 --- a/fs/dax.c +++ b/fs/dax.c @@ -855,8 +855,8 @@ int __dax_pmd_fault(struct vm_area_struct *vma, unsigned long address, long length = dax_map_atomic(bdev, &dax); if (length < 0) { - result = VM_FAULT_SIGBUS; - goto out; + dax_pmd_dbg(&bh, address, "dax-error fallback"); + goto fallback; } if (length < PMD_SIZE) { dax_pmd_dbg(&bh, address, "dax-length too small"); -- cgit v1.3-8-gc7d7 From 1f1735cb75ab31ed948a7fce65beaaa6c0ed4ed4 Mon Sep 17 00:00:00 2001 From: Goldwyn Rodrigues Date: Mon, 18 Apr 2016 06:41:52 -0500 Subject: cifs: Use file_dentry() CIFS may be used as lower layer of overlayfs and accessing f_path.dentry can lead to a crash. Fix by replacing direct access of file->f_path.dentry with the file_dentry() accessor, which will always return a native object. Signed-off-by: Goldwyn Rodrigues Acked-by: Shirish Pargaonkar Signed-off-by: Steve French --- fs/cifs/file.c | 4 ++-- fs/cifs/readdir.c | 4 ++-- 2 files changed, 4 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/cifs/file.c b/fs/cifs/file.c index c03d0744648b..489ddc797105 100644 --- a/fs/cifs/file.c +++ b/fs/cifs/file.c @@ -271,7 +271,7 @@ struct cifsFileInfo * cifs_new_fileinfo(struct cifs_fid *fid, struct file *file, struct tcon_link *tlink, __u32 oplock) { - struct dentry *dentry = file->f_path.dentry; + struct dentry *dentry = file_dentry(file); struct inode *inode = d_inode(dentry); struct cifsInodeInfo *cinode = CIFS_I(inode); struct cifsFileInfo *cfile; @@ -461,7 +461,7 @@ int cifs_open(struct inode *inode, struct file *file) tcon = tlink_tcon(tlink); server = tcon->ses->server; - full_path = build_path_from_dentry(file->f_path.dentry); + full_path = build_path_from_dentry(file_dentry(file)); if (full_path == NULL) { rc = -ENOMEM; goto out; diff --git a/fs/cifs/readdir.c b/fs/cifs/readdir.c index 867439c21001..65cf85dcda09 100644 --- a/fs/cifs/readdir.c +++ b/fs/cifs/readdir.c @@ -303,7 +303,7 @@ initiate_cifs_search(const unsigned int xid, struct file *file) cifsFile->invalidHandle = true; cifsFile->srch_inf.endOfSearch = false; - full_path = build_path_from_dentry(file->f_path.dentry); + full_path = build_path_from_dentry(file_dentry(file)); if (full_path == NULL) { rc = -ENOMEM; goto error_exit; @@ -762,7 +762,7 @@ static int cifs_filldir(char *find_entry, struct file *file, */ fattr.cf_flags |= CIFS_FATTR_NEED_REVAL; - cifs_prime_dcache(file->f_path.dentry, &name, &fattr); + cifs_prime_dcache(file_dentry(file), &name, &fattr); ino = cifs_uniqueid_to_ino_t(fattr.cf_uniqueid); return !dir_emit(ctx, name.name, name.len, ino, fattr.cf_dtype); -- cgit v1.3-8-gc7d7 From 11e31647c9263185cfd990db656dbb7f06697faf Mon Sep 17 00:00:00 2001 From: Sachin Prabhu Date: Mon, 8 Feb 2016 13:44:01 +0530 Subject: cifs: remove any preceding delimiter from prefix_path We currently do not check if any delimiter exists before the prefix path in cifs_compose_mount_options(). Consequently when building the devname using cifs_build_devname() we can end up with multiple delimiters separating the UNC and the prefix path. An issue was reported by the customer mounting a folder within a DFS share from a Netapp server which uses McAfee antivirus. We have narrowed down the cause to the use of double backslashes in the file name used to open the file. This was determined to be caused because of additional delimiters as a result of the bug. In addition to changes in cifs_build_devname(), we also fix cifs_parse_devname() to ignore any preceding delimiter for the prefix path. The problem was originally reported on RHEL 6 in RHEL bz 1252721. This is the upstream version of the fix. The fix was confirmed by looking at the packet capture of a DFS mount. Signed-off-by: Sachin Prabhu Signed-off-by: Steve French --- fs/cifs/cifs_dfs_ref.c | 6 +++++- fs/cifs/connect.c | 8 ++++++-- 2 files changed, 11 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/cifs/cifs_dfs_ref.c b/fs/cifs/cifs_dfs_ref.c index 94f2c8a9ae6d..ec9dbbcca3b9 100644 --- a/fs/cifs/cifs_dfs_ref.c +++ b/fs/cifs/cifs_dfs_ref.c @@ -151,8 +151,12 @@ char *cifs_compose_mount_options(const char *sb_mountdata, if (sb_mountdata == NULL) return ERR_PTR(-EINVAL); - if (strlen(fullpath) - ref->path_consumed) + if (strlen(fullpath) - ref->path_consumed) { prepath = fullpath + ref->path_consumed; + /* skip initial delimiter */ + if (*prepath == '/' || *prepath == '\\') + prepath++; + } *devname = cifs_build_devname(ref->node_name, prepath); if (IS_ERR(*devname)) { diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c index 6f62ac821a84..078b7311c3fa 100644 --- a/fs/cifs/connect.c +++ b/fs/cifs/connect.c @@ -1196,8 +1196,12 @@ cifs_parse_devname(const char *devname, struct smb_vol *vol) convert_delimiter(vol->UNC, '\\'); - /* If pos is NULL, or is a bogus trailing delimiter then no prepath */ - if (!*pos++ || !*pos) + /* skip any delimiter */ + if (*pos == '/' || *pos == '\\') + pos++; + + /* If pos is NULL then no prepath */ + if (!*pos) return 0; vol->prepath = kstrdup(pos, GFP_KERNEL); -- cgit v1.3-8-gc7d7 From cfda35d98298131bf38fbad3ce4cd5ecb3cf18db Mon Sep 17 00:00:00 2001 From: Stefan Metzmacher Date: Tue, 3 May 2016 10:52:30 +0200 Subject: fs/cifs: correctly to anonymous authentication via NTLMSSP See [MS-NLMP] 3.2.5.1.2 Server Receives an AUTHENTICATE_MESSAGE from the Client: ... Set NullSession to FALSE If (AUTHENTICATE_MESSAGE.UserNameLen == 0 AND AUTHENTICATE_MESSAGE.NtChallengeResponse.Length == 0 AND (AUTHENTICATE_MESSAGE.LmChallengeResponse == Z(1) OR AUTHENTICATE_MESSAGE.LmChallengeResponse.Length == 0)) -- Special case: client requested anonymous authentication Set NullSession to TRUE ... Only server which map unknown users to guest will allow access using a non-null NTChallengeResponse. For Samba it's the "map to guest = bad user" option. BUG: https://bugzilla.samba.org/show_bug.cgi?id=11913 CC: Stable Signed-off-by: Stefan Metzmacher Signed-off-by: Steve French --- fs/cifs/sess.c | 32 ++++++++++++++++++++------------ 1 file changed, 20 insertions(+), 12 deletions(-) (limited to 'fs') diff --git a/fs/cifs/sess.c b/fs/cifs/sess.c index 59727e32ed0f..09b0201d15d0 100644 --- a/fs/cifs/sess.c +++ b/fs/cifs/sess.c @@ -400,19 +400,27 @@ int build_ntlmssp_auth_blob(unsigned char *pbuffer, sec_blob->LmChallengeResponse.MaximumLength = 0; sec_blob->NtChallengeResponse.BufferOffset = cpu_to_le32(tmp - pbuffer); - rc = setup_ntlmv2_rsp(ses, nls_cp); - if (rc) { - cifs_dbg(VFS, "Error %d during NTLMSSP authentication\n", rc); - goto setup_ntlmv2_ret; + if (ses->user_name != NULL) { + rc = setup_ntlmv2_rsp(ses, nls_cp); + if (rc) { + cifs_dbg(VFS, "Error %d during NTLMSSP authentication\n", rc); + goto setup_ntlmv2_ret; + } + memcpy(tmp, ses->auth_key.response + CIFS_SESS_KEY_SIZE, + ses->auth_key.len - CIFS_SESS_KEY_SIZE); + tmp += ses->auth_key.len - CIFS_SESS_KEY_SIZE; + + sec_blob->NtChallengeResponse.Length = + cpu_to_le16(ses->auth_key.len - CIFS_SESS_KEY_SIZE); + sec_blob->NtChallengeResponse.MaximumLength = + cpu_to_le16(ses->auth_key.len - CIFS_SESS_KEY_SIZE); + } else { + /* + * don't send an NT Response for anonymous access + */ + sec_blob->NtChallengeResponse.Length = 0; + sec_blob->NtChallengeResponse.MaximumLength = 0; } - memcpy(tmp, ses->auth_key.response + CIFS_SESS_KEY_SIZE, - ses->auth_key.len - CIFS_SESS_KEY_SIZE); - tmp += ses->auth_key.len - CIFS_SESS_KEY_SIZE; - - sec_blob->NtChallengeResponse.Length = - cpu_to_le16(ses->auth_key.len - CIFS_SESS_KEY_SIZE); - sec_blob->NtChallengeResponse.MaximumLength = - cpu_to_le16(ses->auth_key.len - CIFS_SESS_KEY_SIZE); if (ses->domainName == NULL) { sec_blob->DomainName.BufferOffset = cpu_to_le32(tmp - pbuffer); -- cgit v1.3-8-gc7d7 From fa8f3a354bb775ec586e4475bcb07f7dece97e0c Mon Sep 17 00:00:00 2001 From: Stefan Metzmacher Date: Tue, 3 May 2016 10:52:30 +0200 Subject: fs/cifs: correctly to anonymous authentication for the LANMAN authentication Only server which map unknown users to guest will allow access using a non-null LMChallengeResponse. For Samba it's the "map to guest = bad user" option. BUG: https://bugzilla.samba.org/show_bug.cgi?id=11913 Signed-off-by: Stefan Metzmacher CC: Stable Signed-off-by: Steve French --- fs/cifs/sess.c | 28 ++++++++++++++++------------ 1 file changed, 16 insertions(+), 12 deletions(-) (limited to 'fs') diff --git a/fs/cifs/sess.c b/fs/cifs/sess.c index 09b0201d15d0..b9e2cc1437ef 100644 --- a/fs/cifs/sess.c +++ b/fs/cifs/sess.c @@ -678,20 +678,24 @@ sess_auth_lanman(struct sess_data *sess_data) pSMB->req.hdr.Flags2 &= ~SMBFLG2_UNICODE; - /* no capabilities flags in old lanman negotiation */ - pSMB->old_req.PasswordLength = cpu_to_le16(CIFS_AUTH_RESP_SIZE); + if (ses->user_name != NULL) { + /* no capabilities flags in old lanman negotiation */ + pSMB->old_req.PasswordLength = cpu_to_le16(CIFS_AUTH_RESP_SIZE); - /* Calculate hash with password and copy into bcc_ptr. - * Encryption Key (stored as in cryptkey) gets used if the - * security mode bit in Negottiate Protocol response states - * to use challenge/response method (i.e. Password bit is 1). - */ - rc = calc_lanman_hash(ses->password, ses->server->cryptkey, - ses->server->sec_mode & SECMODE_PW_ENCRYPT ? - true : false, lnm_session_key); + /* Calculate hash with password and copy into bcc_ptr. + * Encryption Key (stored as in cryptkey) gets used if the + * security mode bit in Negottiate Protocol response states + * to use challenge/response method (i.e. Password bit is 1). + */ + rc = calc_lanman_hash(ses->password, ses->server->cryptkey, + ses->server->sec_mode & SECMODE_PW_ENCRYPT ? + true : false, lnm_session_key); - memcpy(bcc_ptr, (char *)lnm_session_key, CIFS_AUTH_RESP_SIZE); - bcc_ptr += CIFS_AUTH_RESP_SIZE; + memcpy(bcc_ptr, (char *)lnm_session_key, CIFS_AUTH_RESP_SIZE); + bcc_ptr += CIFS_AUTH_RESP_SIZE; + } else { + pSMB->old_req.PasswordLength = 0; + } /* * can not sign if LANMAN negotiated so no need -- cgit v1.3-8-gc7d7 From 777f69b8d26bf35ade4a76b08f203c11e048365d Mon Sep 17 00:00:00 2001 From: Stefan Metzmacher Date: Tue, 3 May 2016 10:52:30 +0200 Subject: fs/cifs: correctly to anonymous authentication for the NTLM(v1) authentication Only server which map unknown users to guest will allow access using a non-null NTChallengeResponse. For Samba it's the "map to guest = bad user" option. BUG: https://bugzilla.samba.org/show_bug.cgi?id=11913 Signed-off-by: Stefan Metzmacher CC: Stable Signed-off-by: Steve French --- fs/cifs/sess.c | 41 +++++++++++++++++++++++------------------ 1 file changed, 23 insertions(+), 18 deletions(-) (limited to 'fs') diff --git a/fs/cifs/sess.c b/fs/cifs/sess.c index b9e2cc1437ef..5d5d11ea0251 100644 --- a/fs/cifs/sess.c +++ b/fs/cifs/sess.c @@ -781,26 +781,31 @@ sess_auth_ntlm(struct sess_data *sess_data) capabilities = cifs_ssetup_hdr(ses, pSMB); pSMB->req_no_secext.Capabilities = cpu_to_le32(capabilities); - pSMB->req_no_secext.CaseInsensitivePasswordLength = - cpu_to_le16(CIFS_AUTH_RESP_SIZE); - pSMB->req_no_secext.CaseSensitivePasswordLength = - cpu_to_le16(CIFS_AUTH_RESP_SIZE); + if (ses->user_name != NULL) { + pSMB->req_no_secext.CaseInsensitivePasswordLength = + cpu_to_le16(CIFS_AUTH_RESP_SIZE); + pSMB->req_no_secext.CaseSensitivePasswordLength = + cpu_to_le16(CIFS_AUTH_RESP_SIZE); - /* calculate ntlm response and session key */ - rc = setup_ntlm_response(ses, sess_data->nls_cp); - if (rc) { - cifs_dbg(VFS, "Error %d during NTLM authentication\n", - rc); - goto out; - } + /* calculate ntlm response and session key */ + rc = setup_ntlm_response(ses, sess_data->nls_cp); + if (rc) { + cifs_dbg(VFS, "Error %d during NTLM authentication\n", + rc); + goto out; + } - /* copy ntlm response */ - memcpy(bcc_ptr, ses->auth_key.response + CIFS_SESS_KEY_SIZE, - CIFS_AUTH_RESP_SIZE); - bcc_ptr += CIFS_AUTH_RESP_SIZE; - memcpy(bcc_ptr, ses->auth_key.response + CIFS_SESS_KEY_SIZE, - CIFS_AUTH_RESP_SIZE); - bcc_ptr += CIFS_AUTH_RESP_SIZE; + /* copy ntlm response */ + memcpy(bcc_ptr, ses->auth_key.response + CIFS_SESS_KEY_SIZE, + CIFS_AUTH_RESP_SIZE); + bcc_ptr += CIFS_AUTH_RESP_SIZE; + memcpy(bcc_ptr, ses->auth_key.response + CIFS_SESS_KEY_SIZE, + CIFS_AUTH_RESP_SIZE); + bcc_ptr += CIFS_AUTH_RESP_SIZE; + } else { + pSMB->req_no_secext.CaseInsensitivePasswordLength = 0; + pSMB->req_no_secext.CaseSensitivePasswordLength = 0; + } if (ses->capabilities & CAP_UNICODE) { /* unicode strings must be word aligned */ -- cgit v1.3-8-gc7d7 From 1a967d6c9b39c226be1b45f13acd4d8a5ab3dc44 Mon Sep 17 00:00:00 2001 From: Stefan Metzmacher Date: Tue, 3 May 2016 10:52:30 +0200 Subject: fs/cifs: correctly to anonymous authentication for the NTLM(v2) authentication Only server which map unknown users to guest will allow access using a non-null NTLMv2_Response. For Samba it's the "map to guest = bad user" option. BUG: https://bugzilla.samba.org/show_bug.cgi?id=11913 Signed-off-by: Stefan Metzmacher CC: Stable Signed-off-by: Steve French --- fs/cifs/sess.c | 32 ++++++++++++++++++-------------- 1 file changed, 18 insertions(+), 14 deletions(-) (limited to 'fs') diff --git a/fs/cifs/sess.c b/fs/cifs/sess.c index 5d5d11ea0251..af0ec2d5ad0e 100644 --- a/fs/cifs/sess.c +++ b/fs/cifs/sess.c @@ -895,22 +895,26 @@ sess_auth_ntlmv2(struct sess_data *sess_data) /* LM2 password would be here if we supported it */ pSMB->req_no_secext.CaseInsensitivePasswordLength = 0; - /* calculate nlmv2 response and session key */ - rc = setup_ntlmv2_rsp(ses, sess_data->nls_cp); - if (rc) { - cifs_dbg(VFS, "Error %d during NTLMv2 authentication\n", rc); - goto out; - } + if (ses->user_name != NULL) { + /* calculate nlmv2 response and session key */ + rc = setup_ntlmv2_rsp(ses, sess_data->nls_cp); + if (rc) { + cifs_dbg(VFS, "Error %d during NTLMv2 authentication\n", rc); + goto out; + } - memcpy(bcc_ptr, ses->auth_key.response + CIFS_SESS_KEY_SIZE, - ses->auth_key.len - CIFS_SESS_KEY_SIZE); - bcc_ptr += ses->auth_key.len - CIFS_SESS_KEY_SIZE; + memcpy(bcc_ptr, ses->auth_key.response + CIFS_SESS_KEY_SIZE, + ses->auth_key.len - CIFS_SESS_KEY_SIZE); + bcc_ptr += ses->auth_key.len - CIFS_SESS_KEY_SIZE; - /* set case sensitive password length after tilen may get - * assigned, tilen is 0 otherwise. - */ - pSMB->req_no_secext.CaseSensitivePasswordLength = - cpu_to_le16(ses->auth_key.len - CIFS_SESS_KEY_SIZE); + /* set case sensitive password length after tilen may get + * assigned, tilen is 0 otherwise. + */ + pSMB->req_no_secext.CaseSensitivePasswordLength = + cpu_to_le16(ses->auth_key.len - CIFS_SESS_KEY_SIZE); + } else { + pSMB->req_no_secext.CaseSensitivePasswordLength = 0; + } if (ses->capabilities & CAP_UNICODE) { if (sess_data->iov[0].iov_len % 2) { -- cgit v1.3-8-gc7d7 From 5a4f7e8e7ff5e81f57b08797cca41d307878dfc7 Mon Sep 17 00:00:00 2001 From: Steve French Date: Tue, 10 May 2016 08:24:05 -0500 Subject: Update cifs.ko version to 2.09 Signed-off-by: Steven French --- fs/cifs/cifsfs.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/cifs/cifsfs.h b/fs/cifs/cifsfs.h index c89ecd7a5c39..5d0d2ff67a4b 100644 --- a/fs/cifs/cifsfs.h +++ b/fs/cifs/cifsfs.h @@ -130,5 +130,5 @@ extern long cifs_ioctl(struct file *filep, unsigned int cmd, unsigned long arg); extern const struct export_operations cifs_export_ops; #endif /* CONFIG_CIFS_NFSD_EXPORT */ -#define CIFS_VERSION "2.08" +#define CIFS_VERSION "2.09" #endif /* _CIFSFS_H */ -- cgit v1.3-8-gc7d7 From 897fba1172d637d344f009d700f7eb8a1fa262f1 Mon Sep 17 00:00:00 2001 From: Steve French Date: Thu, 12 May 2016 21:20:36 -0500 Subject: remove directory incorrectly tries to set delete on close on non-empty directories Wrong return code was being returned on SMB3 rmdir of non-empty directory. For SMB3 (unlike for cifs), we attempt to delete a directory by set of delete on close flag on the open. Windows clients set this flag via a set info (SET_FILE_DISPOSITION to set this flag) which properly checks if the directory is empty. With this patch on smb3 mounts we correctly return "DIRECTORY NOT EMPTY" on attempts to remove a non-empty directory. Signed-off-by: Steve French CC: Stable Acked-by: Sachin Prabhu --- fs/cifs/smb2glob.h | 1 + fs/cifs/smb2inode.c | 8 ++++++-- fs/cifs/smb2pdu.c | 16 ++++++++++++++++ fs/cifs/smb2proto.h | 2 ++ 4 files changed, 25 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/cifs/smb2glob.h b/fs/cifs/smb2glob.h index bc0bb9c34f72..0ffa18094335 100644 --- a/fs/cifs/smb2glob.h +++ b/fs/cifs/smb2glob.h @@ -44,6 +44,7 @@ #define SMB2_OP_DELETE 7 #define SMB2_OP_HARDLINK 8 #define SMB2_OP_SET_EOF 9 +#define SMB2_OP_RMDIR 10 /* Used when constructing chained read requests. */ #define CHAINED_REQUEST 1 diff --git a/fs/cifs/smb2inode.c b/fs/cifs/smb2inode.c index 899bbc86f73e..4f0231e685a9 100644 --- a/fs/cifs/smb2inode.c +++ b/fs/cifs/smb2inode.c @@ -80,6 +80,10 @@ smb2_open_op_close(const unsigned int xid, struct cifs_tcon *tcon, * SMB2_open() call. */ break; + case SMB2_OP_RMDIR: + tmprc = SMB2_rmdir(xid, tcon, fid.persistent_fid, + fid.volatile_fid); + break; case SMB2_OP_RENAME: tmprc = SMB2_rename(xid, tcon, fid.persistent_fid, fid.volatile_fid, (__le16 *)data); @@ -191,8 +195,8 @@ smb2_rmdir(const unsigned int xid, struct cifs_tcon *tcon, const char *name, struct cifs_sb_info *cifs_sb) { return smb2_open_op_close(xid, tcon, cifs_sb, name, DELETE, FILE_OPEN, - CREATE_NOT_FILE | CREATE_DELETE_ON_CLOSE, - NULL, SMB2_OP_DELETE); + CREATE_NOT_FILE, + NULL, SMB2_OP_RMDIR); } int diff --git a/fs/cifs/smb2pdu.c b/fs/cifs/smb2pdu.c index 42e1f440eb1e..8f38e33d365b 100644 --- a/fs/cifs/smb2pdu.c +++ b/fs/cifs/smb2pdu.c @@ -2574,6 +2574,22 @@ SMB2_rename(const unsigned int xid, struct cifs_tcon *tcon, return rc; } +int +SMB2_rmdir(const unsigned int xid, struct cifs_tcon *tcon, + u64 persistent_fid, u64 volatile_fid) +{ + __u8 delete_pending = 1; + void *data; + unsigned int size; + + data = &delete_pending; + size = 1; /* sizeof __u8 */ + + return send_set_info(xid, tcon, persistent_fid, volatile_fid, + current->tgid, FILE_DISPOSITION_INFORMATION, 1, &data, + &size); +} + int SMB2_set_hardlink(const unsigned int xid, struct cifs_tcon *tcon, u64 persistent_fid, u64 volatile_fid, __le16 *target_file) diff --git a/fs/cifs/smb2proto.h b/fs/cifs/smb2proto.h index 4f07dc93608d..eb2cde2f64ba 100644 --- a/fs/cifs/smb2proto.h +++ b/fs/cifs/smb2proto.h @@ -141,6 +141,8 @@ extern int SMB2_query_directory(const unsigned int xid, struct cifs_tcon *tcon, extern int SMB2_rename(const unsigned int xid, struct cifs_tcon *tcon, u64 persistent_fid, u64 volatile_fid, __le16 *target_file); +extern int SMB2_rmdir(const unsigned int xid, struct cifs_tcon *tcon, + u64 persistent_fid, u64 volatile_fid); extern int SMB2_set_hardlink(const unsigned int xid, struct cifs_tcon *tcon, u64 persistent_fid, u64 volatile_fid, __le16 *target_file); -- cgit v1.3-8-gc7d7 From c2985d001d2fb77357aeae675545893b61c50044 Mon Sep 17 00:00:00 2001 From: Olga Kornievskaia Date: Tue, 10 May 2016 16:57:41 -0400 Subject: Fixing oops in callback path Commit 80f9642724af5 ("NFSv4.x: Enforce the ca_maxreponsesize_cached on the back channel") causes an oops when it receives a callback with cachethis=yes. [ 109.667378] BUG: unable to handle kernel NULL pointer dereference at 00000000000002c8 [ 109.669476] IP: [] nfs4_callback_compound+0x4f8/0x690 [nfsv4] [ 109.671216] PGD 0 [ 109.671736] Oops: 0000 [#1] SMP [ 109.705427] CPU: 1 PID: 3579 Comm: nfsv4.1-svc Not tainted 4.5.0-rc1+ #1 [ 109.706987] Hardware name: VMware, Inc. VMware Virtual Platform/440BX Desktop Reference Platform, BIOS 6.00 05/20/2014 [ 109.709468] task: ffff8800b4408000 ti: ffff88008448c000 task.ti: ffff88008448c000 [ 109.711207] RIP: 0010:[] [] nfs4_callback_compound+0x4f8/0x690 [nfsv4] [ 109.713521] RSP: 0018:ffff88008448fca0 EFLAGS: 00010286 [ 109.714762] RAX: ffff880081ee202c RBX: ffff8800b7b5b600 RCX: 0000000000000001 [ 109.716427] RDX: 0000000000000008 RSI: 0000000000000008 RDI: 0000000000000000 [ 109.718091] RBP: ffff88008448fda8 R08: 0000000000000000 R09: 000000000b000000 [ 109.719757] R10: ffff880137786000 R11: ffff8800b7b5b600 R12: 0000000001000000 [ 109.721415] R13: 0000000000000002 R14: 0000000053270000 R15: 000000000000000b [ 109.723061] FS: 0000000000000000(0000) GS:ffff880139640000(0000) knlGS:0000000000000000 [ 109.724931] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 109.726278] CR2: 00000000000002c8 CR3: 0000000034d50000 CR4: 00000000001406e0 [ 109.727972] Stack: [ 109.728465] ffff880081ee202c ffff880081ee201c 000000008448fcc0 ffff8800baccb800 [ 109.730349] ffff8800baccc800 ffffffffa08d0380 0000000000000000 0000000000000000 [ 109.732211] ffff8800b7b5b600 0000000000000001 ffffffff81d073c0 ffff880081ee3090 [ 109.734056] Call Trace: [ 109.734657] [] svc_process_common+0x5c4/0x6c0 [sunrpc] [ 109.736267] [] bc_svc_process+0x1fc/0x360 [sunrpc] [ 109.737775] [] nfs41_callback_svc+0x10c/0x1d0 [nfsv4] [ 109.739335] [] ? prepare_to_wait_event+0xf0/0xf0 [ 109.740799] [] ? nfs4_callback_svc+0x50/0x50 [nfsv4] [ 109.742349] [] kthread+0xd8/0xf0 [ 109.743495] [] ? kthread_park+0x60/0x60 [ 109.744776] [] ret_from_fork+0x3f/0x70 [ 109.746037] [] ? kthread_park+0x60/0x60 [ 109.747324] Code: cc 45 31 f6 48 8b 85 00 ff ff ff 44 89 30 48 8b 85 f8 fe ff ff 44 89 20 48 8b 9d 38 ff ff ff 48 8b bd 30 ff ff ff 48 85 db 74 4c <4c> 8b af c8 02 00 00 4d 8d a5 08 02 00 00 49 81 c5 98 02 00 00 [ 109.754361] RIP [] nfs4_callback_compound+0x4f8/0x690 [nfsv4] [ 109.756123] RSP [ 109.756951] CR2: 00000000000002c8 [ 109.757738] ---[ end trace 2b8555511ab5dfb4 ]--- [ 109.758819] Kernel panic - not syncing: Fatal exception [ 109.760126] Kernel Offset: disabled [ 118.938934] ---[ end Kernel panic - not syncing: Fatal exception It doesn't unlock the table nor does it set the cps->clp pointer which is later needed by nfs4_cb_free_slot(). Fixes: 80f9642724af5 ("NFSv4.x: Enforce the ca_maxresponsesize_cached ...") CC: stable@vger.kernel.org Signed-off-by: Olga Kornievskaia Signed-off-by: Anna Schumaker --- fs/nfs/callback_proc.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/nfs/callback_proc.c b/fs/nfs/callback_proc.c index 618ced381a14..7c9fbf504f07 100644 --- a/fs/nfs/callback_proc.c +++ b/fs/nfs/callback_proc.c @@ -500,8 +500,10 @@ __be32 nfs4_callback_sequence(struct cb_sequenceargs *args, cps->slot = slot; /* The ca_maxresponsesize_cached is 0 with no DRC */ - if (args->csa_cachethis != 0) - return htonl(NFS4ERR_REP_TOO_BIG_TO_CACHE); + if (args->csa_cachethis != 0) { + status = htonl(NFS4ERR_REP_TOO_BIG_TO_CACHE); + goto out_unlock; + } /* * Check for pending referring calls. If a match is found, a -- cgit v1.3-8-gc7d7 From 67911c8f18b595040f5c2df0550c69a7c9397891 Mon Sep 17 00:00:00 2001 From: Anna Schumaker Date: Tue, 19 Jan 2016 14:37:30 -0500 Subject: NFS: Add nfs_commit_file() Copy will use this to set up a commit request for a generic range. I don't want to allocate a new pagecache entry for the file, so I needed to change parts of the commit path to handle requests with a null wb_page. Signed-off-by: Anna Schumaker --- fs/nfs/internal.h | 1 + fs/nfs/pagelist.c | 6 ++++-- fs/nfs/write.c | 41 +++++++++++++++++++++++++++++++++++++---- 3 files changed, 42 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h index f1d1d2c472e9..5154fa65a2f2 100644 --- a/fs/nfs/internal.h +++ b/fs/nfs/internal.h @@ -477,6 +477,7 @@ void nfs_mark_request_commit(struct nfs_page *req, u32 ds_commit_idx); int nfs_write_need_commit(struct nfs_pgio_header *); void nfs_writeback_update_inode(struct nfs_pgio_header *hdr); +int nfs_commit_file(struct file *file, struct nfs_write_verifier *verf); int nfs_generic_commit_list(struct inode *inode, struct list_head *head, int how, struct nfs_commit_info *cinfo); void nfs_retry_commit(struct list_head *page_list, diff --git a/fs/nfs/pagelist.c b/fs/nfs/pagelist.c index 1f6db4231057..174dd4cf5747 100644 --- a/fs/nfs/pagelist.c +++ b/fs/nfs/pagelist.c @@ -341,8 +341,10 @@ nfs_create_request(struct nfs_open_context *ctx, struct page *page, * long write-back delay. This will be adjusted in * update_nfs_request below if the region is not locked. */ req->wb_page = page; - req->wb_index = page_file_index(page); - get_page(page); + if (page) { + req->wb_index = page_file_index(page); + get_page(page); + } req->wb_offset = offset; req->wb_pgbase = offset; req->wb_bytes = count; diff --git a/fs/nfs/write.c b/fs/nfs/write.c index 9283a96b9fb8..e18a97aab275 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -737,7 +737,7 @@ static void nfs_inode_remove_request(struct nfs_page *req) head = req->wb_head; spin_lock(&inode->i_lock); - if (likely(!PageSwapCache(head->wb_page))) { + if (likely(head->wb_page && !PageSwapCache(head->wb_page))) { set_page_private(head->wb_page, 0); ClearPagePrivate(head->wb_page); smp_mb__after_atomic(); @@ -759,7 +759,8 @@ static void nfs_inode_remove_request(struct nfs_page *req) static void nfs_mark_request_dirty(struct nfs_page *req) { - __set_page_dirty_nobuffers(req->wb_page); + if (req->wb_page) + __set_page_dirty_nobuffers(req->wb_page); } /* @@ -835,7 +836,8 @@ nfs_request_add_commit_list(struct nfs_page *req, struct nfs_commit_info *cinfo) spin_lock(&cinfo->inode->i_lock); nfs_request_add_commit_list_locked(req, &cinfo->mds->list, cinfo); spin_unlock(&cinfo->inode->i_lock); - nfs_mark_page_unstable(req->wb_page, cinfo); + if (req->wb_page) + nfs_mark_page_unstable(req->wb_page, cinfo); } EXPORT_SYMBOL_GPL(nfs_request_add_commit_list); @@ -1724,6 +1726,36 @@ nfs_commit_list(struct inode *inode, struct list_head *head, int how, return -ENOMEM; } +int nfs_commit_file(struct file *file, struct nfs_write_verifier *verf) +{ + struct inode *inode = file_inode(file); + struct nfs_open_context *open; + struct nfs_commit_info cinfo; + struct nfs_page *req; + int ret; + + open = get_nfs_open_context(nfs_file_open_context(file)); + req = nfs_create_request(open, NULL, NULL, 0, i_size_read(inode)); + if (!req) { + ret = -ENOMEM; + goto out_put; + } + + nfs_init_cinfo_from_inode(&cinfo, inode); + + memcpy(&req->wb_verf, verf, sizeof(struct nfs_write_verifier)); + nfs_request_add_commit_list(req, &cinfo); + ret = nfs_commit_inode(inode, FLUSH_SYNC); + if (ret > 0) + ret = 0; + + nfs_free_request(req); +out_put: + put_nfs_open_context(open); + return ret; +} +EXPORT_SYMBOL_GPL(nfs_commit_file); + /* * COMMIT call returned */ @@ -1748,7 +1780,8 @@ static void nfs_commit_release_pages(struct nfs_commit_data *data) while (!list_empty(&data->pages)) { req = nfs_list_entry(data->pages.next); nfs_list_remove_request(req); - nfs_clear_page_commit(req->wb_page); + if (req->wb_page) + nfs_clear_page_commit(req->wb_page); dprintk("NFS: commit (%s/%llu %d@%lld)", req->wb_context->dentry->d_sb->s_id, -- cgit v1.3-8-gc7d7 From 2e72448b07dc3ff1b7593e9bfff91db182262857 Mon Sep 17 00:00:00 2001 From: Anna Schumaker Date: Tue, 21 May 2013 16:53:03 -0400 Subject: NFS: Add COPY nfs operation This adds the copy_range file_ops function pointer used by the sys_copy_range() function call. This patch only implements sync copies, so if an async copy happens we decode the stateid and ignore it. Signed-off-by: Anna Schumaker --- fs/nfs/nfs42.h | 1 + fs/nfs/nfs42proc.c | 105 +++++++++++++++++++++++++++++++++ fs/nfs/nfs42xdr.c | 146 ++++++++++++++++++++++++++++++++++++++++++++++ fs/nfs/nfs4file.c | 23 ++++++++ fs/nfs/nfs4proc.c | 1 + fs/nfs/nfs4xdr.c | 1 + include/linux/nfs4.h | 1 + include/linux/nfs_fs_sb.h | 1 + include/linux/nfs_xdr.h | 26 +++++++++ 9 files changed, 305 insertions(+) (limited to 'fs') diff --git a/fs/nfs/nfs42.h b/fs/nfs/nfs42.h index b587ccd31083..b6cd15314bab 100644 --- a/fs/nfs/nfs42.h +++ b/fs/nfs/nfs42.h @@ -13,6 +13,7 @@ /* nfs4.2proc.c */ int nfs42_proc_allocate(struct file *, loff_t, loff_t); +ssize_t nfs42_proc_copy(struct file *, loff_t, struct file *, loff_t, size_t); int nfs42_proc_deallocate(struct file *, loff_t, loff_t); loff_t nfs42_proc_llseek(struct file *, loff_t, int); int nfs42_proc_layoutstats_generic(struct nfs_server *, diff --git a/fs/nfs/nfs42proc.c b/fs/nfs/nfs42proc.c index dff83460e5a6..579ee20e4120 100644 --- a/fs/nfs/nfs42proc.c +++ b/fs/nfs/nfs42proc.c @@ -126,6 +126,111 @@ int nfs42_proc_deallocate(struct file *filep, loff_t offset, loff_t len) return err; } +static ssize_t _nfs42_proc_copy(struct file *src, loff_t pos_src, + struct nfs_lock_context *src_lock, + struct file *dst, loff_t pos_dst, + struct nfs_lock_context *dst_lock, + size_t count) +{ + struct nfs42_copy_args args = { + .src_fh = NFS_FH(file_inode(src)), + .src_pos = pos_src, + .dst_fh = NFS_FH(file_inode(dst)), + .dst_pos = pos_dst, + .count = count, + }; + struct nfs42_copy_res res; + struct rpc_message msg = { + .rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_COPY], + .rpc_argp = &args, + .rpc_resp = &res, + }; + struct inode *dst_inode = file_inode(dst); + struct nfs_server *server = NFS_SERVER(dst_inode); + int status; + + status = nfs4_set_rw_stateid(&args.src_stateid, src_lock->open_context, + src_lock, FMODE_READ); + if (status) + return status; + + status = nfs4_set_rw_stateid(&args.dst_stateid, dst_lock->open_context, + dst_lock, FMODE_WRITE); + if (status) + return status; + + status = nfs4_call_sync(server->client, server, &msg, + &args.seq_args, &res.seq_res, 0); + if (status == -ENOTSUPP) + server->caps &= ~NFS_CAP_COPY; + if (status) + return status; + + if (res.write_res.verifier.committed != NFS_FILE_SYNC) { + status = nfs_commit_file(dst, &res.write_res.verifier.verifier); + if (status) + return status; + } + + truncate_pagecache_range(dst_inode, pos_dst, + pos_dst + res.write_res.count); + + return res.write_res.count; +} + +ssize_t nfs42_proc_copy(struct file *src, loff_t pos_src, + struct file *dst, loff_t pos_dst, + size_t count) +{ + struct nfs_server *server = NFS_SERVER(file_inode(dst)); + struct nfs_lock_context *src_lock; + struct nfs_lock_context *dst_lock; + struct nfs4_exception src_exception = { }; + struct nfs4_exception dst_exception = { }; + ssize_t err, err2; + + if (!nfs_server_capable(file_inode(dst), NFS_CAP_COPY)) + return -EOPNOTSUPP; + + src_lock = nfs_get_lock_context(nfs_file_open_context(src)); + if (IS_ERR(src_lock)) + return PTR_ERR(src_lock); + + src_exception.inode = file_inode(src); + src_exception.state = src_lock->open_context->state; + + dst_lock = nfs_get_lock_context(nfs_file_open_context(dst)); + if (IS_ERR(dst_lock)) { + err = PTR_ERR(dst_lock); + goto out_put_src_lock; + } + + dst_exception.inode = file_inode(dst); + dst_exception.state = dst_lock->open_context->state; + + do { + mutex_lock(&file_inode(dst)->i_mutex); + err = _nfs42_proc_copy(src, pos_src, src_lock, + dst, pos_dst, dst_lock, count); + mutex_unlock(&file_inode(dst)->i_mutex); + + if (err == -ENOTSUPP) { + err = -EOPNOTSUPP; + break; + } + + err2 = nfs4_handle_exception(server, err, &src_exception); + err = nfs4_handle_exception(server, err, &dst_exception); + if (!err) + err = err2; + } while (src_exception.retry || dst_exception.retry); + + nfs_put_lock_context(dst_lock); +out_put_src_lock: + nfs_put_lock_context(src_lock); + return err; +} + static loff_t _nfs42_proc_llseek(struct file *filep, struct nfs_lock_context *lock, loff_t offset, int whence) { diff --git a/fs/nfs/nfs42xdr.c b/fs/nfs/nfs42xdr.c index 0ca482a51e53..6dc6f2aea0d6 100644 --- a/fs/nfs/nfs42xdr.c +++ b/fs/nfs/nfs42xdr.c @@ -9,9 +9,22 @@ #define encode_fallocate_maxsz (encode_stateid_maxsz + \ 2 /* offset */ + \ 2 /* length */) +#define NFS42_WRITE_RES_SIZE (1 /* wr_callback_id size */ +\ + XDR_QUADLEN(NFS4_STATEID_SIZE) + \ + 2 /* wr_count */ + \ + 1 /* wr_committed */ + \ + XDR_QUADLEN(NFS4_VERIFIER_SIZE)) #define encode_allocate_maxsz (op_encode_hdr_maxsz + \ encode_fallocate_maxsz) #define decode_allocate_maxsz (op_decode_hdr_maxsz) +#define encode_copy_maxsz (op_encode_hdr_maxsz + \ + XDR_QUADLEN(NFS4_STATEID_SIZE) + \ + XDR_QUADLEN(NFS4_STATEID_SIZE) + \ + 2 + 2 + 2 + 1 + 1 + 1) +#define decode_copy_maxsz (op_decode_hdr_maxsz + \ + NFS42_WRITE_RES_SIZE + \ + 1 /* cr_consecutive */ + \ + 1 /* cr_synchronous */) #define encode_deallocate_maxsz (op_encode_hdr_maxsz + \ encode_fallocate_maxsz) #define decode_deallocate_maxsz (op_decode_hdr_maxsz) @@ -49,6 +62,16 @@ decode_putfh_maxsz + \ decode_allocate_maxsz + \ decode_getattr_maxsz) +#define NFS4_enc_copy_sz (compound_encode_hdr_maxsz + \ + encode_putfh_maxsz + \ + encode_savefh_maxsz + \ + encode_putfh_maxsz + \ + encode_copy_maxsz) +#define NFS4_dec_copy_sz (compound_decode_hdr_maxsz + \ + decode_putfh_maxsz + \ + decode_savefh_maxsz + \ + decode_putfh_maxsz + \ + decode_copy_maxsz) #define NFS4_enc_deallocate_sz (compound_encode_hdr_maxsz + \ encode_putfh_maxsz + \ encode_deallocate_maxsz + \ @@ -102,6 +125,23 @@ static void encode_allocate(struct xdr_stream *xdr, encode_fallocate(xdr, args); } +static void encode_copy(struct xdr_stream *xdr, + struct nfs42_copy_args *args, + struct compound_hdr *hdr) +{ + encode_op_hdr(xdr, OP_COPY, decode_copy_maxsz, hdr); + encode_nfs4_stateid(xdr, &args->src_stateid); + encode_nfs4_stateid(xdr, &args->dst_stateid); + + encode_uint64(xdr, args->src_pos); + encode_uint64(xdr, args->dst_pos); + encode_uint64(xdr, args->count); + + encode_uint32(xdr, 1); /* consecutive = true */ + encode_uint32(xdr, 1); /* synchronous = true */ + encode_uint32(xdr, 0); /* src server list */ +} + static void encode_deallocate(struct xdr_stream *xdr, struct nfs42_falloc_args *args, struct compound_hdr *hdr) @@ -181,6 +221,26 @@ static void nfs4_xdr_enc_allocate(struct rpc_rqst *req, encode_nops(&hdr); } +/* + * Encode COPY request + */ +static void nfs4_xdr_enc_copy(struct rpc_rqst *req, + struct xdr_stream *xdr, + struct nfs42_copy_args *args) +{ + struct compound_hdr hdr = { + .minorversion = nfs4_xdr_minorversion(&args->seq_args), + }; + + encode_compound_hdr(xdr, req, &hdr); + encode_sequence(xdr, &args->seq_args, &hdr); + encode_putfh(xdr, args->src_fh, &hdr); + encode_savefh(xdr, &hdr); + encode_putfh(xdr, args->dst_fh, &hdr); + encode_copy(xdr, args, &hdr); + encode_nops(&hdr); +} + /* * Encode DEALLOCATE request */ @@ -266,6 +326,62 @@ static int decode_allocate(struct xdr_stream *xdr, struct nfs42_falloc_res *res) return decode_op_hdr(xdr, OP_ALLOCATE); } +static int decode_write_response(struct xdr_stream *xdr, + struct nfs42_write_res *res) +{ + __be32 *p; + int stateids; + + p = xdr_inline_decode(xdr, 4 + 8 + 4); + if (unlikely(!p)) + goto out_overflow; + + stateids = be32_to_cpup(p++); + p = xdr_decode_hyper(p, &res->count); + res->verifier.committed = be32_to_cpup(p); + return decode_verifier(xdr, &res->verifier.verifier); + +out_overflow: + print_overflow_msg(__func__, xdr); + return -EIO; +} + +static int decode_copy_requirements(struct xdr_stream *xdr, + struct nfs42_copy_res *res) { + __be32 *p; + + p = xdr_inline_decode(xdr, 4 + 4); + if (unlikely(!p)) + goto out_overflow; + + res->consecutive = be32_to_cpup(p++); + res->synchronous = be32_to_cpup(p++); + return 0; +out_overflow: + print_overflow_msg(__func__, xdr); + return -EIO; +} + +static int decode_copy(struct xdr_stream *xdr, struct nfs42_copy_res *res) +{ + int status; + + status = decode_op_hdr(xdr, OP_COPY); + if (status == NFS4ERR_OFFLOAD_NO_REQS) { + status = decode_copy_requirements(xdr, res); + if (status) + return status; + return NFS4ERR_OFFLOAD_NO_REQS; + } else if (status) + return status; + + status = decode_write_response(xdr, &res->write_res); + if (status) + return status; + + return decode_copy_requirements(xdr, res); +} + static int decode_deallocate(struct xdr_stream *xdr, struct nfs42_falloc_res *res) { return decode_op_hdr(xdr, OP_DEALLOCATE); @@ -330,6 +446,36 @@ out: return status; } +/* + * Decode COPY response + */ +static int nfs4_xdr_dec_copy(struct rpc_rqst *rqstp, + struct xdr_stream *xdr, + struct nfs42_copy_res *res) +{ + struct compound_hdr hdr; + int status; + + status = decode_compound_hdr(xdr, &hdr); + if (status) + goto out; + status = decode_sequence(xdr, &res->seq_res, rqstp); + if (status) + goto out; + status = decode_putfh(xdr); + if (status) + goto out; + status = decode_savefh(xdr); + if (status) + goto out; + status = decode_putfh(xdr); + if (status) + goto out; + status = decode_copy(xdr, res); +out: + return status; +} + /* * Decode DEALLOCATE request */ diff --git a/fs/nfs/nfs4file.c b/fs/nfs/nfs4file.c index d0390516467c..014b0e41ace5 100644 --- a/fs/nfs/nfs4file.c +++ b/fs/nfs/nfs4file.c @@ -129,6 +129,28 @@ nfs4_file_flush(struct file *file, fl_owner_t id) } #ifdef CONFIG_NFS_V4_2 +static ssize_t nfs4_copy_file_range(struct file *file_in, loff_t pos_in, + struct file *file_out, loff_t pos_out, + size_t count, unsigned int flags) +{ + struct inode *in_inode = file_inode(file_in); + struct inode *out_inode = file_inode(file_out); + int ret; + + if (in_inode == out_inode) + return -EINVAL; + + /* flush any pending writes */ + ret = nfs_sync_inode(in_inode); + if (ret) + return ret; + ret = nfs_sync_inode(out_inode); + if (ret) + return ret; + + return nfs42_proc_copy(file_in, pos_in, file_out, pos_out, count); +} + static loff_t nfs4_file_llseek(struct file *filep, loff_t offset, int whence) { loff_t ret; @@ -243,6 +265,7 @@ const struct file_operations nfs4_file_operations = { .check_flags = nfs_check_flags, .setlease = simple_nosetlease, #ifdef CONFIG_NFS_V4_2 + .copy_file_range = nfs4_copy_file_range, .llseek = nfs4_file_llseek, .fallocate = nfs42_fallocate, .clone_file_range = nfs42_clone_file_range, diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index bc2676c95e1b..4e83385ea6a9 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -8797,6 +8797,7 @@ static const struct nfs4_minor_version_ops nfs_v4_2_minor_ops = { | NFS_CAP_STATEID_NFSV41 | NFS_CAP_ATOMIC_OPEN_V1 | NFS_CAP_ALLOCATE + | NFS_CAP_COPY | NFS_CAP_DEALLOCATE | NFS_CAP_SEEK | NFS_CAP_LAYOUTSTATS diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c index 88474a4fc669..d1c96fc62c51 100644 --- a/fs/nfs/nfs4xdr.c +++ b/fs/nfs/nfs4xdr.c @@ -7515,6 +7515,7 @@ struct rpc_procinfo nfs4_procedures[] = { PROC(DEALLOCATE, enc_deallocate, dec_deallocate), PROC(LAYOUTSTATS, enc_layoutstats, dec_layoutstats), PROC(CLONE, enc_clone, dec_clone), + PROC(COPY, enc_copy, dec_copy), #endif /* CONFIG_NFS_V4_2 */ }; diff --git a/include/linux/nfs4.h b/include/linux/nfs4.h index 011433478a14..722509482e1a 100644 --- a/include/linux/nfs4.h +++ b/include/linux/nfs4.h @@ -504,6 +504,7 @@ enum { NFSPROC4_CLNT_DEALLOCATE, NFSPROC4_CLNT_LAYOUTSTATS, NFSPROC4_CLNT_CLONE, + NFSPROC4_CLNT_COPY, }; /* nfs41 types */ diff --git a/include/linux/nfs_fs_sb.h b/include/linux/nfs_fs_sb.h index 7fcc13c8cf1f..14a762d2734d 100644 --- a/include/linux/nfs_fs_sb.h +++ b/include/linux/nfs_fs_sb.h @@ -246,5 +246,6 @@ struct nfs_server { #define NFS_CAP_DEALLOCATE (1U << 21) #define NFS_CAP_LAYOUTSTATS (1U << 22) #define NFS_CAP_CLONE (1U << 23) +#define NFS_CAP_COPY (1U << 24) #endif diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h index cb9982d8f38f..e70ed54dad94 100644 --- a/include/linux/nfs_xdr.h +++ b/include/linux/nfs_xdr.h @@ -1343,6 +1343,32 @@ struct nfs42_falloc_res { const struct nfs_server *falloc_server; }; +struct nfs42_copy_args { + struct nfs4_sequence_args seq_args; + + struct nfs_fh *src_fh; + nfs4_stateid src_stateid; + u64 src_pos; + + struct nfs_fh *dst_fh; + nfs4_stateid dst_stateid; + u64 dst_pos; + + u64 count; +}; + +struct nfs42_write_res { + u64 count; + struct nfs_writeverf verifier; +}; + +struct nfs42_copy_res { + struct nfs4_sequence_res seq_res; + struct nfs42_write_res write_res; + bool consecutive; + bool synchronous; +}; + struct nfs42_seek_args { struct nfs4_sequence_args seq_args; -- cgit v1.3-8-gc7d7 From a1d1c4f11adf8a0036cd40a643543047fc952eb6 Mon Sep 17 00:00:00 2001 From: Tigran Mkrtchyan Date: Thu, 12 May 2016 11:16:38 +0200 Subject: nfs4: client: do not send empty SETATTR after OPEN_CREATE OPEN_CREATE with EXCLUSIVE4_1 sends initial file permission. Ignoring fact, that server have indicated that file mod is set, client will send yet another SETATTR request, but, as mode is already set, new SETATTR will be empty. This is not a problem, nevertheless an extra roundtrip and slow open on high latency networks. This change is aims to skip extra setattr after open if there are no attributes to be set. Signed-off-by: Tigran Mkrtchyan Signed-off-by: Anna Schumaker --- fs/nfs/nfs4proc.c | 34 +++++++++++++++++++++++++--------- 1 file changed, 25 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 4e83385ea6a9..9795725d708c 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -74,6 +74,17 @@ #define NFS4_POLL_RETRY_MIN (HZ/10) #define NFS4_POLL_RETRY_MAX (15*HZ) +/* file attributes which can be mapped to nfs attributes */ +#define NFS4_VALID_ATTRS (ATTR_MODE \ + | ATTR_UID \ + | ATTR_GID \ + | ATTR_SIZE \ + | ATTR_ATIME \ + | ATTR_MTIME \ + | ATTR_CTIME \ + | ATTR_ATIME_SET \ + | ATTR_MTIME_SET) + struct nfs4_opendata; static int _nfs4_proc_open(struct nfs4_opendata *data); static int _nfs4_recover_proc_open(struct nfs4_opendata *data); @@ -2558,15 +2569,20 @@ static int _nfs4_do_open(struct inode *dir, if ((opendata->o_arg.open_flags & (O_CREAT|O_EXCL)) == (O_CREAT|O_EXCL) && (opendata->o_arg.createmode != NFS4_CREATE_GUARDED)) { nfs4_exclusive_attrset(opendata, sattr, &label); - - nfs_fattr_init(opendata->o_res.f_attr); - status = nfs4_do_setattr(state->inode, cred, - opendata->o_res.f_attr, sattr, - state, label, olabel); - if (status == 0) { - nfs_setattr_update_inode(state->inode, sattr, - opendata->o_res.f_attr); - nfs_setsecurity(state->inode, opendata->o_res.f_attr, olabel); + /* + * send create attributes which was not set by open + * with an extra setattr. + */ + if (sattr->ia_valid & NFS4_VALID_ATTRS) { + nfs_fattr_init(opendata->o_res.f_attr); + status = nfs4_do_setattr(state->inode, cred, + opendata->o_res.f_attr, sattr, + state, label, olabel); + if (status == 0) { + nfs_setattr_update_inode(state->inode, sattr, + opendata->o_res.f_attr); + nfs_setsecurity(state->inode, opendata->o_res.f_attr, olabel); + } } } if (opened && opendata->file_created) -- cgit v1.3-8-gc7d7 From 181342c5ebe8cc31f75b80ace18ae8a89a0c145a Mon Sep 17 00:00:00 2001 From: Shirley Ma Date: Mon, 2 May 2016 14:40:23 -0400 Subject: xprtrdma: Add rdma6 option to support NFS/RDMA IPv6 RFC 5666: The "rdma" netid is to be used when IPv4 addressing is employed by the underlying transport, and "rdma6" for IPv6 addressing. Add mount -o proto=rdma6 option to support NFS/RDMA IPv6 addressing. Changes from v2: - Integrated comments from Chuck Level, Anna Schumaker, Trodt Myklebust - Add a little more to the patch description to describe NFS/RDMA IPv6 suggested by Chuck Level and Anna Schumaker - Removed duplicated rdma6 define - Remove Opt_xprt_rdma mountfamily since it doesn't support Signed-off-by: Shirley Ma Reviewed-by: Chuck Lever Signed-off-by: Anna Schumaker --- fs/nfs/super.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'fs') diff --git a/fs/nfs/super.c b/fs/nfs/super.c index ff537a205188..2137e0202f25 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -191,6 +191,7 @@ static const match_table_t nfs_mount_option_tokens = { enum { Opt_xprt_udp, Opt_xprt_udp6, Opt_xprt_tcp, Opt_xprt_tcp6, Opt_xprt_rdma, + Opt_xprt_rdma6, Opt_xprt_err }; @@ -201,6 +202,7 @@ static const match_table_t nfs_xprt_protocol_tokens = { { Opt_xprt_tcp, "tcp" }, { Opt_xprt_tcp6, "tcp6" }, { Opt_xprt_rdma, "rdma" }, + { Opt_xprt_rdma6, "rdma6" }, { Opt_xprt_err, NULL } }; @@ -1456,6 +1458,8 @@ static int nfs_parse_mount_options(char *raw, mnt->flags |= NFS_MOUNT_TCP; mnt->nfs_server.protocol = XPRT_TRANSPORT_TCP; break; + case Opt_xprt_rdma6: + protofamily = AF_INET6; case Opt_xprt_rdma: /* vector side protocols to TCP */ mnt->flags |= NFS_MOUNT_TCP; -- cgit v1.3-8-gc7d7 From 6b26cc8c8ead3636a18bfd9489984983f4ddd6f4 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Mon, 2 May 2016 14:40:40 -0400 Subject: sunrpc: Advertise maximum backchannel payload size RPC-over-RDMA transports have a limit on how large a backward direction (backchannel) RPC message can be. Ensure that the NFSv4.x CREATE_SESSION operation advertises this limit to servers. Signed-off-by: Chuck Lever Tested-by: Steve Wise Signed-off-by: Anna Schumaker --- fs/nfs/nfs4proc.c | 10 ++++++---- include/linux/sunrpc/clnt.h | 1 + include/linux/sunrpc/xprt.h | 1 + net/sunrpc/clnt.c | 17 +++++++++++++++++ net/sunrpc/xprtrdma/backchannel.c | 16 ++++++++++++++++ net/sunrpc/xprtrdma/transport.c | 1 + net/sunrpc/xprtrdma/xprt_rdma.h | 1 + net/sunrpc/xprtsock.c | 6 ++++++ 8 files changed, 49 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 9795725d708c..196e41e12621 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -7371,9 +7371,11 @@ int nfs4_proc_get_lease_time(struct nfs_client *clp, struct nfs_fsinfo *fsinfo) * always set csa_cachethis to FALSE because the current implementation * of the back channel DRC only supports caching the CB_SEQUENCE operation. */ -static void nfs4_init_channel_attrs(struct nfs41_create_session_args *args) +static void nfs4_init_channel_attrs(struct nfs41_create_session_args *args, + struct rpc_clnt *clnt) { unsigned int max_rqst_sz, max_resp_sz; + unsigned int max_bc_payload = rpc_max_bc_payload(clnt); max_rqst_sz = NFS_MAX_FILE_IO_SIZE + nfs41_maxwrite_overhead; max_resp_sz = NFS_MAX_FILE_IO_SIZE + nfs41_maxread_overhead; @@ -7391,8 +7393,8 @@ static void nfs4_init_channel_attrs(struct nfs41_create_session_args *args) args->fc_attrs.max_ops, args->fc_attrs.max_reqs); /* Back channel attributes */ - args->bc_attrs.max_rqst_sz = PAGE_SIZE; - args->bc_attrs.max_resp_sz = PAGE_SIZE; + args->bc_attrs.max_rqst_sz = max_bc_payload; + args->bc_attrs.max_resp_sz = max_bc_payload; args->bc_attrs.max_resp_sz_cached = 0; args->bc_attrs.max_ops = NFS4_MAX_BACK_CHANNEL_OPS; args->bc_attrs.max_reqs = NFS41_BC_MAX_CALLBACKS; @@ -7496,7 +7498,7 @@ static int _nfs4_proc_create_session(struct nfs_client *clp, }; int status; - nfs4_init_channel_attrs(&args); + nfs4_init_channel_attrs(&args, clp->cl_rpcclient); args.flags = (SESSION4_PERSIST | SESSION4_BACK_CHAN); status = rpc_call_sync(session->clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT); diff --git a/include/linux/sunrpc/clnt.h b/include/linux/sunrpc/clnt.h index 9a7ddbaf116e..19c659d1c0f8 100644 --- a/include/linux/sunrpc/clnt.h +++ b/include/linux/sunrpc/clnt.h @@ -176,6 +176,7 @@ void rpc_setbufsize(struct rpc_clnt *, unsigned int, unsigned int); int rpc_protocol(struct rpc_clnt *); struct net * rpc_net_ns(struct rpc_clnt *); size_t rpc_max_payload(struct rpc_clnt *); +size_t rpc_max_bc_payload(struct rpc_clnt *); unsigned long rpc_get_timeout(struct rpc_clnt *clnt); void rpc_force_rebind(struct rpc_clnt *); size_t rpc_peeraddr(struct rpc_clnt *, struct sockaddr *, size_t); diff --git a/include/linux/sunrpc/xprt.h b/include/linux/sunrpc/xprt.h index fb0d212e0d3a..5aa3834619a8 100644 --- a/include/linux/sunrpc/xprt.h +++ b/include/linux/sunrpc/xprt.h @@ -142,6 +142,7 @@ struct rpc_xprt_ops { int (*bc_setup)(struct rpc_xprt *xprt, unsigned int min_reqs); int (*bc_up)(struct svc_serv *serv, struct net *net); + size_t (*bc_maxpayload)(struct rpc_xprt *xprt); void (*bc_free_rqst)(struct rpc_rqst *rqst); void (*bc_destroy)(struct rpc_xprt *xprt, unsigned int max_reqs); diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c index 7e0c9bf22df8..06b4df9faaa1 100644 --- a/net/sunrpc/clnt.c +++ b/net/sunrpc/clnt.c @@ -1413,6 +1413,23 @@ size_t rpc_max_payload(struct rpc_clnt *clnt) } EXPORT_SYMBOL_GPL(rpc_max_payload); +/** + * rpc_max_bc_payload - Get maximum backchannel payload size, in bytes + * @clnt: RPC client to query + */ +size_t rpc_max_bc_payload(struct rpc_clnt *clnt) +{ + struct rpc_xprt *xprt; + size_t ret; + + rcu_read_lock(); + xprt = rcu_dereference(clnt->cl_xprt); + ret = xprt->ops->bc_maxpayload(xprt); + rcu_read_unlock(); + return ret; +} +EXPORT_SYMBOL_GPL(rpc_max_bc_payload); + /** * rpc_get_timeout - Get timeout for transport in units of HZ * @clnt: RPC client to query diff --git a/net/sunrpc/xprtrdma/backchannel.c b/net/sunrpc/xprtrdma/backchannel.c index 2dcd7640eeb5..87762d976b63 100644 --- a/net/sunrpc/xprtrdma/backchannel.c +++ b/net/sunrpc/xprtrdma/backchannel.c @@ -191,6 +191,22 @@ int xprt_rdma_bc_up(struct svc_serv *serv, struct net *net) return 0; } +/** + * xprt_rdma_bc_maxpayload - Return maximum backchannel message size + * @xprt: transport + * + * Returns maximum size, in bytes, of a backchannel message + */ +size_t xprt_rdma_bc_maxpayload(struct rpc_xprt *xprt) +{ + struct rpcrdma_xprt *r_xprt = rpcx_to_rdmax(xprt); + struct rpcrdma_create_data_internal *cdata = &r_xprt->rx_data; + size_t maxmsg; + + maxmsg = min_t(unsigned int, cdata->inline_rsize, cdata->inline_wsize); + return maxmsg - RPCRDMA_HDRLEN_MIN; +} + /** * rpcrdma_bc_marshal_reply - Send backwards direction reply * @rqst: buffer containing RPC reply data diff --git a/net/sunrpc/xprtrdma/transport.c b/net/sunrpc/xprtrdma/transport.c index b1b009f10ea3..9954342924df 100644 --- a/net/sunrpc/xprtrdma/transport.c +++ b/net/sunrpc/xprtrdma/transport.c @@ -707,6 +707,7 @@ static struct rpc_xprt_ops xprt_rdma_procs = { #if defined(CONFIG_SUNRPC_BACKCHANNEL) .bc_setup = xprt_rdma_bc_setup, .bc_up = xprt_rdma_bc_up, + .bc_maxpayload = xprt_rdma_bc_maxpayload, .bc_free_rqst = xprt_rdma_bc_free_rqst, .bc_destroy = xprt_rdma_bc_destroy, #endif diff --git a/net/sunrpc/xprtrdma/xprt_rdma.h b/net/sunrpc/xprtrdma/xprt_rdma.h index 2ebc743cb96f..7723e5faff4d 100644 --- a/net/sunrpc/xprtrdma/xprt_rdma.h +++ b/net/sunrpc/xprtrdma/xprt_rdma.h @@ -534,6 +534,7 @@ void xprt_rdma_cleanup(void); #if defined(CONFIG_SUNRPC_BACKCHANNEL) int xprt_rdma_bc_setup(struct rpc_xprt *, unsigned int); int xprt_rdma_bc_up(struct svc_serv *, struct net *); +size_t xprt_rdma_bc_maxpayload(struct rpc_xprt *); int rpcrdma_bc_post_recv(struct rpcrdma_xprt *, unsigned int); void rpcrdma_bc_receive_call(struct rpcrdma_xprt *, struct rpcrdma_rep *); int rpcrdma_bc_marshal_reply(struct rpc_rqst *); diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c index 65e759569e48..f1faf6b9aaff 100644 --- a/net/sunrpc/xprtsock.c +++ b/net/sunrpc/xprtsock.c @@ -1365,6 +1365,11 @@ static int xs_tcp_bc_up(struct svc_serv *serv, struct net *net) return ret; return 0; } + +static size_t xs_tcp_bc_maxpayload(struct rpc_xprt *xprt) +{ + return PAGE_SIZE; +} #else static inline int _xs_tcp_read_data(struct rpc_xprt *xprt, struct xdr_skb_reader *desc) @@ -2660,6 +2665,7 @@ static struct rpc_xprt_ops xs_tcp_ops = { #ifdef CONFIG_SUNRPC_BACKCHANNEL .bc_setup = xprt_setup_bc, .bc_up = xs_tcp_bc_up, + .bc_maxpayload = xs_tcp_bc_maxpayload, .bc_free_rqst = xprt_free_bc_rqst, .bc_destroy = xprt_destroy_bc, #endif -- cgit v1.3-8-gc7d7 From f538d0ba5be39c52addff01c7ec31f0b90e57dfc Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Mon, 16 May 2016 14:41:14 -0400 Subject: pNFS: Fix a leaked layoutstats flag Signed-off-by: Trond Myklebust Signed-off-by: Anna Schumaker --- fs/nfs/pnfs.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index 5b404d926e08..faea98281c75 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -2407,7 +2407,7 @@ pnfs_report_layoutstat(struct inode *inode, gfp_t gfp_flags) spin_lock(&inode->i_lock); if (!NFS_I(inode)->layout) { spin_unlock(&inode->i_lock); - goto out; + goto out_clear_layoutstats; } hdr = NFS_I(inode)->layout; pnfs_get_layout_hdr(hdr); @@ -2436,6 +2436,7 @@ out_free: kfree(data); out_put: pnfs_put_layout_hdr(hdr); +out_clear_layoutstats: smp_mb__before_atomic(); clear_bit(NFS_INO_LAYOUTSTATS, &nfsi->flags); smp_mb__after_atomic(); -- cgit v1.3-8-gc7d7 From 93b717fd81bf6b9a73c3702e9b079b4de8148b34 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Mon, 16 May 2016 17:42:43 -0400 Subject: NFSv4: Label stateids with the type In order to more easily distinguish what kind of stateid we are dealing with, introduce a type that can be used to label the stateid structure. The label will be useful both for debugging, but also when dealing with operations like SETATTR, READ and WRITE that can take several different types of stateid as arguments. Signed-off-by: Trond Myklebust Signed-off-by: Anna Schumaker --- fs/nfs/callback_xdr.c | 17 ++++++++++--- fs/nfs/flexfilelayout/flexfilelayout.c | 7 +++--- fs/nfs/flexfilelayout/flexfilelayoutdev.c | 3 ++- fs/nfs/nfs4_fs.h | 7 ++++-- fs/nfs/nfs4proc.c | 3 +++ fs/nfs/nfs4state.c | 5 +++- fs/nfs/nfs4xdr.c | 42 ++++++++++++++++++++++++------- include/linux/nfs4.h | 25 ++++++++++++++---- 8 files changed, 85 insertions(+), 24 deletions(-) (limited to 'fs') diff --git a/fs/nfs/callback_xdr.c b/fs/nfs/callback_xdr.c index 976c90608e56..d81f96aacd51 100644 --- a/fs/nfs/callback_xdr.c +++ b/fs/nfs/callback_xdr.c @@ -146,10 +146,16 @@ static __be32 decode_stateid(struct xdr_stream *xdr, nfs4_stateid *stateid) p = read_buf(xdr, NFS4_STATEID_SIZE); if (unlikely(p == NULL)) return htonl(NFS4ERR_RESOURCE); - memcpy(stateid, p, NFS4_STATEID_SIZE); + memcpy(stateid->data, p, NFS4_STATEID_SIZE); return 0; } +static __be32 decode_delegation_stateid(struct xdr_stream *xdr, nfs4_stateid *stateid) +{ + stateid->type = NFS4_DELEGATION_STATEID_TYPE; + return decode_stateid(xdr, stateid); +} + static __be32 decode_compound_hdr_arg(struct xdr_stream *xdr, struct cb_compound_hdr_arg *hdr) { __be32 *p; @@ -211,7 +217,7 @@ static __be32 decode_recall_args(struct svc_rqst *rqstp, struct xdr_stream *xdr, __be32 *p; __be32 status; - status = decode_stateid(xdr, &args->stateid); + status = decode_delegation_stateid(xdr, &args->stateid); if (unlikely(status != 0)) goto out; p = read_buf(xdr, 4); @@ -227,6 +233,11 @@ out: } #if defined(CONFIG_NFS_V4_1) +static __be32 decode_layout_stateid(struct xdr_stream *xdr, nfs4_stateid *stateid) +{ + stateid->type = NFS4_LAYOUT_STATEID_TYPE; + return decode_stateid(xdr, stateid); +} static __be32 decode_layoutrecall_args(struct svc_rqst *rqstp, struct xdr_stream *xdr, @@ -263,7 +274,7 @@ static __be32 decode_layoutrecall_args(struct svc_rqst *rqstp, } p = xdr_decode_hyper(p, &args->cbl_range.offset); p = xdr_decode_hyper(p, &args->cbl_range.length); - status = decode_stateid(xdr, &args->cbl_stateid); + status = decode_layout_stateid(xdr, &args->cbl_stateid); if (unlikely(status != 0)) goto out; } else if (args->cbl_recall_type == RETURN_FSID) { diff --git a/fs/nfs/flexfilelayout/flexfilelayout.c b/fs/nfs/flexfilelayout/flexfilelayout.c index 60d690dbc947..53b6391e0eba 100644 --- a/fs/nfs/flexfilelayout/flexfilelayout.c +++ b/fs/nfs/flexfilelayout/flexfilelayout.c @@ -55,14 +55,15 @@ ff_layout_free_layout_hdr(struct pnfs_layout_hdr *lo) kfree(FF_LAYOUT_FROM_HDR(lo)); } -static int decode_stateid(struct xdr_stream *xdr, nfs4_stateid *stateid) +static int decode_pnfs_stateid(struct xdr_stream *xdr, nfs4_stateid *stateid) { __be32 *p; p = xdr_inline_decode(xdr, NFS4_STATEID_SIZE); if (unlikely(p == NULL)) return -ENOBUFS; - memcpy(stateid, p, NFS4_STATEID_SIZE); + stateid->type = NFS4_PNFS_DS_STATEID_TYPE; + memcpy(stateid->data, p, NFS4_STATEID_SIZE); dprintk("%s: stateid id= [%x%x%x%x]\n", __func__, p[0], p[1], p[2], p[3]); return 0; @@ -465,7 +466,7 @@ ff_layout_alloc_lseg(struct pnfs_layout_hdr *lh, fls->mirror_array[i]->efficiency = be32_to_cpup(p); /* stateid */ - rc = decode_stateid(&stream, &fls->mirror_array[i]->stateid); + rc = decode_pnfs_stateid(&stream, &fls->mirror_array[i]->stateid); if (rc) goto out_err_free; diff --git a/fs/nfs/flexfilelayout/flexfilelayoutdev.c b/fs/nfs/flexfilelayout/flexfilelayoutdev.c index 56296f3df19c..eeef89359ad2 100644 --- a/fs/nfs/flexfilelayout/flexfilelayoutdev.c +++ b/fs/nfs/flexfilelayout/flexfilelayoutdev.c @@ -228,7 +228,8 @@ ff_ds_error_match(const struct nfs4_ff_layout_ds_err *e1, return e1->opnum < e2->opnum ? -1 : 1; if (e1->status != e2->status) return e1->status < e2->status ? -1 : 1; - ret = memcmp(&e1->stateid, &e2->stateid, sizeof(e1->stateid)); + ret = memcmp(e1->stateid.data, e2->stateid.data, + sizeof(e1->stateid.data)); if (ret != 0) return ret; ret = memcmp(&e1->deviceid, &e2->deviceid, sizeof(e1->deviceid)); diff --git a/fs/nfs/nfs4_fs.h b/fs/nfs/nfs4_fs.h index 4afdee420d25..b5d9f345c9f2 100644 --- a/fs/nfs/nfs4_fs.h +++ b/fs/nfs/nfs4_fs.h @@ -496,12 +496,15 @@ extern struct svc_version nfs4_callback_version4; static inline void nfs4_stateid_copy(nfs4_stateid *dst, const nfs4_stateid *src) { - memcpy(dst, src, sizeof(*dst)); + memcpy(dst->data, src->data, sizeof(dst->data)); + dst->type = src->type; } static inline bool nfs4_stateid_match(const nfs4_stateid *dst, const nfs4_stateid *src) { - return memcmp(dst, src, sizeof(*dst)) == 0; + if (dst->type != src->type) + return false; + return memcmp(dst->data, src->data, sizeof(dst->data)) == 0; } static inline bool nfs4_stateid_match_other(const nfs4_stateid *dst, const nfs4_stateid *src) diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 196e41e12621..2516467ff17f 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -8675,6 +8675,9 @@ nfs41_free_lock_state(struct nfs_server *server, struct nfs4_lock_state *lsp) static bool nfs41_match_stateid(const nfs4_stateid *s1, const nfs4_stateid *s2) { + if (s1->type != s2->type) + return false; + if (memcmp(s1->other, s2->other, sizeof(s1->other)) != 0) return false; diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c index d854693a15b0..d630f9cca0f1 100644 --- a/fs/nfs/nfs4state.c +++ b/fs/nfs/nfs4state.c @@ -65,7 +65,10 @@ #define OPENOWNER_POOL_SIZE 8 -const nfs4_stateid zero_stateid; +const nfs4_stateid zero_stateid = { + .data = { 0 }, + .type = NFS4_SPECIAL_STATEID_TYPE, +}; static DEFINE_MUTEX(nfs_clid_init_mutex); int nfs4_init_clientid(struct nfs_client *clp, struct rpc_cred *cred) diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c index d1c96fc62c51..661e753fe1c9 100644 --- a/fs/nfs/nfs4xdr.c +++ b/fs/nfs/nfs4xdr.c @@ -4270,6 +4270,24 @@ static int decode_stateid(struct xdr_stream *xdr, nfs4_stateid *stateid) return decode_opaque_fixed(xdr, stateid, NFS4_STATEID_SIZE); } +static int decode_open_stateid(struct xdr_stream *xdr, nfs4_stateid *stateid) +{ + stateid->type = NFS4_OPEN_STATEID_TYPE; + return decode_stateid(xdr, stateid); +} + +static int decode_lock_stateid(struct xdr_stream *xdr, nfs4_stateid *stateid) +{ + stateid->type = NFS4_LOCK_STATEID_TYPE; + return decode_stateid(xdr, stateid); +} + +static int decode_delegation_stateid(struct xdr_stream *xdr, nfs4_stateid *stateid) +{ + stateid->type = NFS4_DELEGATION_STATEID_TYPE; + return decode_stateid(xdr, stateid); +} + static int decode_close(struct xdr_stream *xdr, struct nfs_closeres *res) { int status; @@ -4278,7 +4296,7 @@ static int decode_close(struct xdr_stream *xdr, struct nfs_closeres *res) if (status != -EIO) nfs_increment_open_seqid(status, res->seqid); if (!status) - status = decode_stateid(xdr, &res->stateid); + status = decode_open_stateid(xdr, &res->stateid); return status; } @@ -4937,7 +4955,7 @@ static int decode_lock(struct xdr_stream *xdr, struct nfs_lock_res *res) if (status == -EIO) goto out; if (status == 0) { - status = decode_stateid(xdr, &res->stateid); + status = decode_lock_stateid(xdr, &res->stateid); if (unlikely(status)) goto out; } else if (status == -NFS4ERR_DENIED) @@ -4966,7 +4984,7 @@ static int decode_locku(struct xdr_stream *xdr, struct nfs_locku_res *res) if (status != -EIO) nfs_increment_lock_seqid(status, res->seqid); if (status == 0) - status = decode_stateid(xdr, &res->stateid); + status = decode_lock_stateid(xdr, &res->stateid); return status; } @@ -5016,7 +5034,7 @@ static int decode_rw_delegation(struct xdr_stream *xdr, __be32 *p; int status; - status = decode_stateid(xdr, &res->delegation); + status = decode_delegation_stateid(xdr, &res->delegation); if (unlikely(status)) return status; p = xdr_inline_decode(xdr, 4); @@ -5096,7 +5114,7 @@ static int decode_open(struct xdr_stream *xdr, struct nfs_openres *res) nfs_increment_open_seqid(status, res->seqid); if (status) return status; - status = decode_stateid(xdr, &res->stateid); + status = decode_open_stateid(xdr, &res->stateid); if (unlikely(status)) return status; @@ -5136,7 +5154,7 @@ static int decode_open_confirm(struct xdr_stream *xdr, struct nfs_open_confirmre if (status != -EIO) nfs_increment_open_seqid(status, res->seqid); if (!status) - status = decode_stateid(xdr, &res->stateid); + status = decode_open_stateid(xdr, &res->stateid); return status; } @@ -5148,7 +5166,7 @@ static int decode_open_downgrade(struct xdr_stream *xdr, struct nfs_closeres *re if (status != -EIO) nfs_increment_open_seqid(status, res->seqid); if (!status) - status = decode_stateid(xdr, &res->stateid); + status = decode_open_stateid(xdr, &res->stateid); return status; } @@ -5838,6 +5856,12 @@ out_overflow: } #if defined(CONFIG_NFS_V4_1) +static int decode_layout_stateid(struct xdr_stream *xdr, nfs4_stateid *stateid) +{ + stateid->type = NFS4_LAYOUT_STATEID_TYPE; + return decode_stateid(xdr, stateid); +} + static int decode_getdeviceinfo(struct xdr_stream *xdr, struct nfs4_getdeviceinfo_res *res) { @@ -5919,7 +5943,7 @@ static int decode_layoutget(struct xdr_stream *xdr, struct rpc_rqst *req, if (unlikely(!p)) goto out_overflow; res->return_on_close = be32_to_cpup(p); - decode_stateid(xdr, &res->stateid); + decode_layout_stateid(xdr, &res->stateid); p = xdr_inline_decode(xdr, 4); if (unlikely(!p)) goto out_overflow; @@ -5985,7 +6009,7 @@ static int decode_layoutreturn(struct xdr_stream *xdr, goto out_overflow; res->lrs_present = be32_to_cpup(p); if (res->lrs_present) - status = decode_stateid(xdr, &res->stateid); + status = decode_layout_stateid(xdr, &res->stateid); return status; out_overflow: print_overflow_msg(__func__, xdr); diff --git a/include/linux/nfs4.h b/include/linux/nfs4.h index 722509482e1a..e1692c96cbc8 100644 --- a/include/linux/nfs4.h +++ b/include/linux/nfs4.h @@ -50,12 +50,27 @@ struct nfs4_label { typedef struct { char data[NFS4_VERIFIER_SIZE]; } nfs4_verifier; -struct nfs_stateid4 { - __be32 seqid; - char other[NFS4_STATEID_OTHER_SIZE]; -} __attribute__ ((packed)); +struct nfs4_stateid_struct { + union { + char data[NFS4_STATEID_SIZE]; + struct { + __be32 seqid; + char other[NFS4_STATEID_OTHER_SIZE]; + } __attribute__ ((packed)); + }; + + enum { + NFS4_INVALID_STATEID_TYPE = 0, + NFS4_SPECIAL_STATEID_TYPE, + NFS4_OPEN_STATEID_TYPE, + NFS4_LOCK_STATEID_TYPE, + NFS4_DELEGATION_STATEID_TYPE, + NFS4_LAYOUT_STATEID_TYPE, + NFS4_PNFS_DS_STATEID_TYPE, + } type; +}; -typedef struct nfs_stateid4 nfs4_stateid; +typedef struct nfs4_stateid_struct nfs4_stateid; enum nfs_opnum4 { OP_ACCESS = 3, -- cgit v1.3-8-gc7d7 From abf4e13cc1e16ed83ec8363d44f76149034b2851 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Mon, 16 May 2016 17:42:44 -0400 Subject: NFSv4: Use the right stateid for delegations in setattr, read and write When we're using a delegation to represent our open state, we should ensure that we use the stateid that was used to create that delegation. Signed-off-by: Trond Myklebust Signed-off-by: Anna Schumaker --- fs/nfs/delegation.c | 9 ++++++--- fs/nfs/delegation.h | 2 +- fs/nfs/nfs4_fs.h | 5 +++-- fs/nfs/nfs4proc.c | 13 +++++++++---- fs/nfs/nfs4state.c | 13 +++++++++---- 5 files changed, 28 insertions(+), 14 deletions(-) (limited to 'fs') diff --git a/fs/nfs/delegation.c b/fs/nfs/delegation.c index 5166adcfc0fb..322c2585bc34 100644 --- a/fs/nfs/delegation.c +++ b/fs/nfs/delegation.c @@ -875,15 +875,16 @@ int nfs_delegations_present(struct nfs_client *clp) /** * nfs4_copy_delegation_stateid - Copy inode's state ID information - * @dst: stateid data structure to fill in * @inode: inode to check * @flags: delegation type requirement + * @dst: stateid data structure to fill in + * @cred: optional argument to retrieve credential * * Returns "true" and fills in "dst->data" * if inode had a delegation, * otherwise "false" is returned. */ -bool nfs4_copy_delegation_stateid(nfs4_stateid *dst, struct inode *inode, - fmode_t flags) +bool nfs4_copy_delegation_stateid(struct inode *inode, fmode_t flags, + nfs4_stateid *dst, struct rpc_cred **cred) { struct nfs_inode *nfsi = NFS_I(inode); struct nfs_delegation *delegation; @@ -896,6 +897,8 @@ bool nfs4_copy_delegation_stateid(nfs4_stateid *dst, struct inode *inode, if (ret) { nfs4_stateid_copy(dst, &delegation->stateid); nfs_mark_delegation_referenced(delegation); + if (cred) + *cred = get_rpccred(delegation->cred); } rcu_read_unlock(); return ret; diff --git a/fs/nfs/delegation.h b/fs/nfs/delegation.h index 333063e032f0..64724d252a79 100644 --- a/fs/nfs/delegation.h +++ b/fs/nfs/delegation.h @@ -56,7 +56,7 @@ void nfs_delegation_reap_unclaimed(struct nfs_client *clp); int nfs4_proc_delegreturn(struct inode *inode, struct rpc_cred *cred, const nfs4_stateid *stateid, int issync); int nfs4_open_delegation_recall(struct nfs_open_context *ctx, struct nfs4_state *state, const nfs4_stateid *stateid, fmode_t type); int nfs4_lock_delegation_recall(struct file_lock *fl, struct nfs4_state *state, const nfs4_stateid *stateid); -bool nfs4_copy_delegation_stateid(nfs4_stateid *dst, struct inode *inode, fmode_t flags); +bool nfs4_copy_delegation_stateid(struct inode *inode, fmode_t flags, nfs4_stateid *dst, struct rpc_cred **cred); void nfs_mark_delegation_referenced(struct nfs_delegation *delegation); int nfs4_have_delegation(struct inode *inode, fmode_t flags); diff --git a/fs/nfs/nfs4_fs.h b/fs/nfs/nfs4_fs.h index b5d9f345c9f2..768456fa1b17 100644 --- a/fs/nfs/nfs4_fs.h +++ b/fs/nfs/nfs4_fs.h @@ -438,8 +438,9 @@ extern void nfs41_handle_server_scope(struct nfs_client *, struct nfs41_server_scope **); extern void nfs4_put_lock_state(struct nfs4_lock_state *lsp); extern int nfs4_set_lock_state(struct nfs4_state *state, struct file_lock *fl); -extern int nfs4_select_rw_stateid(nfs4_stateid *, struct nfs4_state *, - fmode_t, const struct nfs_lockowner *); +extern int nfs4_select_rw_stateid(struct nfs4_state *, fmode_t, + const struct nfs_lockowner *, nfs4_stateid *, + struct rpc_cred **); extern struct nfs_seqid *nfs_alloc_seqid(struct nfs_seqid_counter *counter, gfp_t gfp_mask); extern int nfs_wait_on_sequence(struct nfs_seqid *seqid, struct rpc_task *task); diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 2516467ff17f..c9b66085d392 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -2692,6 +2692,7 @@ static int _nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred, .rpc_resp = &res, .rpc_cred = cred, }; + struct rpc_cred *delegation_cred = NULL; unsigned long timestamp = jiffies; fmode_t fmode; bool truncate; @@ -2707,7 +2708,7 @@ static int _nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred, truncate = (sattr->ia_valid & ATTR_SIZE) ? true : false; fmode = truncate ? FMODE_WRITE : FMODE_READ; - if (nfs4_copy_delegation_stateid(&arg.stateid, inode, fmode)) { + if (nfs4_copy_delegation_stateid(inode, fmode, &arg.stateid, &delegation_cred)) { /* Use that stateid */ } else if (truncate && state != NULL) { struct nfs_lockowner lockowner = { @@ -2716,13 +2717,17 @@ static int _nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred, }; if (!nfs4_valid_open_stateid(state)) return -EBADF; - if (nfs4_select_rw_stateid(&arg.stateid, state, FMODE_WRITE, - &lockowner) == -EIO) + if (nfs4_select_rw_stateid(state, FMODE_WRITE, &lockowner, + &arg.stateid, &delegation_cred) == -EIO) return -EBADF; } else nfs4_stateid_copy(&arg.stateid, &zero_stateid); + if (delegation_cred) + msg.rpc_cred = delegation_cred; status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1); + + put_rpccred(delegation_cred); if (status == 0 && state != NULL) renew_lease(server, timestamp); trace_nfs4_setattr(inode, &arg.stateid, status); @@ -4301,7 +4306,7 @@ int nfs4_set_rw_stateid(nfs4_stateid *stateid, if (l_ctx != NULL) lockowner = &l_ctx->lockowner; - return nfs4_select_rw_stateid(stateid, ctx->state, fmode, lockowner); + return nfs4_select_rw_stateid(ctx->state, fmode, lockowner, stateid, NULL); } EXPORT_SYMBOL_GPL(nfs4_set_rw_stateid); diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c index d630f9cca0f1..5075592df145 100644 --- a/fs/nfs/nfs4state.c +++ b/fs/nfs/nfs4state.c @@ -988,15 +988,20 @@ static void nfs4_copy_open_stateid(nfs4_stateid *dst, struct nfs4_state *state) * Byte-range lock aware utility to initialize the stateid of read/write * requests. */ -int nfs4_select_rw_stateid(nfs4_stateid *dst, struct nfs4_state *state, - fmode_t fmode, const struct nfs_lockowner *lockowner) +int nfs4_select_rw_stateid(struct nfs4_state *state, + fmode_t fmode, const struct nfs_lockowner *lockowner, + nfs4_stateid *dst, struct rpc_cred **cred) { - int ret = nfs4_copy_lock_stateid(dst, state, lockowner); + int ret; + + if (cred != NULL) + *cred = NULL; + ret = nfs4_copy_lock_stateid(dst, state, lockowner); if (ret == -EIO) /* A lost lock - don't even consider delegations */ goto out; /* returns true if delegation stateid found and copied */ - if (nfs4_copy_delegation_stateid(dst, state->inode, fmode)) { + if (nfs4_copy_delegation_stateid(state->inode, fmode, dst, cred)) { ret = 0; goto out; } -- cgit v1.3-8-gc7d7 From cca588d6c835e21fd7a3e02a964ce5fe09fdad26 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Mon, 16 May 2016 17:42:45 -0400 Subject: NFS: Reclaim writes via writepage are opportunistic No need to make them a priority any more, or to make them succeed. Signed-off-by: Trond Myklebust Signed-off-by: Anna Schumaker --- fs/nfs/write.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/nfs/write.c b/fs/nfs/write.c index e18a97aab275..4dac51ba1f7e 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -245,8 +245,7 @@ static void nfs_mark_uptodate(struct nfs_page *req) static int wb_priority(struct writeback_control *wbc) { int ret = 0; - if (wbc->for_reclaim) - return FLUSH_HIGHPRI | FLUSH_COND_STABLE; + if (wbc->sync_mode == WB_SYNC_ALL) ret = FLUSH_COND_STABLE; return ret; -- cgit v1.3-8-gc7d7 From 75bf47ebf6b5349d875fdffe7d64b0fc4e936170 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 17 May 2016 12:28:36 -0400 Subject: pNFS/flexfile: Fix erroneous fall back to read/write through the MDS This patch fixes a problem whereby the pNFS client falls back to doing reads and writes through the metadata server even when the layout flag FF_FLAGS_NO_IO_THRU_MDS is set. Signed-off-by: Trond Myklebust Signed-off-by: Anna Schumaker --- fs/nfs/flexfilelayout/flexfilelayout.c | 23 ++++++----------------- 1 file changed, 6 insertions(+), 17 deletions(-) (limited to 'fs') diff --git a/fs/nfs/flexfilelayout/flexfilelayout.c b/fs/nfs/flexfilelayout/flexfilelayout.c index 53b6391e0eba..b6e0e32b6053 100644 --- a/fs/nfs/flexfilelayout/flexfilelayout.c +++ b/fs/nfs/flexfilelayout/flexfilelayout.c @@ -1295,7 +1295,7 @@ ff_layout_set_layoutcommit(struct nfs_pgio_header *hdr) } static bool -ff_layout_reset_to_mds(struct pnfs_layout_segment *lseg, int idx) +ff_layout_device_unavailable(struct pnfs_layout_segment *lseg, int idx) { /* No mirroring for now */ struct nfs4_deviceid_node *node = FF_LAYOUT_DEVID_NODE(lseg, idx); @@ -1332,16 +1332,10 @@ static int ff_layout_read_prepare_common(struct rpc_task *task, rpc_exit(task, -EIO); return -EIO; } - if (ff_layout_reset_to_mds(hdr->lseg, hdr->pgio_mirror_idx)) { - dprintk("%s task %u reset io to MDS\n", __func__, task->tk_pid); - if (ff_layout_has_available_ds(hdr->lseg)) - pnfs_read_resend_pnfs(hdr); - else - ff_layout_reset_read(hdr); - rpc_exit(task, 0); + if (ff_layout_device_unavailable(hdr->lseg, hdr->pgio_mirror_idx)) { + rpc_exit(task, -EHOSTDOWN); return -EAGAIN; } - hdr->pgio_done_cb = ff_layout_read_done_cb; ff_layout_read_record_layoutstats_start(task, hdr); return 0; @@ -1531,14 +1525,8 @@ static int ff_layout_write_prepare_common(struct rpc_task *task, return -EIO; } - if (ff_layout_reset_to_mds(hdr->lseg, hdr->pgio_mirror_idx)) { - bool retry_pnfs; - - retry_pnfs = ff_layout_has_available_ds(hdr->lseg); - dprintk("%s task %u reset io to %s\n", __func__, - task->tk_pid, retry_pnfs ? "pNFS" : "MDS"); - ff_layout_reset_write(hdr, retry_pnfs); - rpc_exit(task, 0); + if (ff_layout_device_unavailable(hdr->lseg, hdr->pgio_mirror_idx)) { + rpc_exit(task, -EHOSTDOWN); return -EAGAIN; } @@ -1755,6 +1743,7 @@ ff_layout_read_pagelist(struct nfs_pgio_header *hdr) dprintk("%s USE DS: %s cl_count %d vers %d\n", __func__, ds->ds_remotestr, atomic_read(&ds->ds_clp->cl_count), vers); + hdr->pgio_done_cb = ff_layout_read_done_cb; atomic_inc(&ds->ds_clp->cl_count); hdr->ds_clp = ds->ds_clp; fh = nfs4_ff_layout_select_ds_fh(lseg, idx); -- cgit v1.3-8-gc7d7 From 3b13b4b311a508a5899070936a873e30fcc96a2f Mon Sep 17 00:00:00 2001 From: Tom Haynes Date: Tue, 17 May 2016 12:28:37 -0400 Subject: pNFS/flexfiles: When checking for available DSes, conditionally check for MDS io Whenever we check to see if we have the needed number of DSes for the action, we may also have to check to see whether IO is allowed to go to the MDS or not. [jlayton: fix merge conflict due to lack of localio patches here] Signed-off-by: Tom Haynes Signed-off-by: Jeff Layton Signed-off-by: Anna Schumaker --- fs/nfs/flexfilelayout/flexfilelayout.c | 5 ++--- fs/nfs/flexfilelayout/flexfilelayout.h | 1 + fs/nfs/flexfilelayout/flexfilelayoutdev.c | 6 ++++++ 3 files changed, 9 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/nfs/flexfilelayout/flexfilelayout.c b/fs/nfs/flexfilelayout/flexfilelayout.c index b6e0e32b6053..a835b75f85ab 100644 --- a/fs/nfs/flexfilelayout/flexfilelayout.c +++ b/fs/nfs/flexfilelayout/flexfilelayout.c @@ -1102,8 +1102,7 @@ static int ff_layout_async_handle_error_v4(struct rpc_task *task, rpc_wake_up(&tbl->slot_tbl_waitq); /* fall through */ default: - if (ff_layout_no_fallback_to_mds(lseg) || - ff_layout_has_available_ds(lseg)) + if (ff_layout_avoid_mds_available_ds(lseg)) return -NFS4ERR_RESET_TO_PNFS; reset: dprintk("%s Retry through MDS. Error %d\n", __func__, @@ -1765,7 +1764,7 @@ ff_layout_read_pagelist(struct nfs_pgio_header *hdr) return PNFS_ATTEMPTED; out_failed: - if (ff_layout_has_available_ds(lseg)) + if (ff_layout_avoid_mds_available_ds(lseg)) return PNFS_TRY_AGAIN; return PNFS_NOT_ATTEMPTED; } diff --git a/fs/nfs/flexfilelayout/flexfilelayout.h b/fs/nfs/flexfilelayout/flexfilelayout.h index 1318c77aeb35..b54058122647 100644 --- a/fs/nfs/flexfilelayout/flexfilelayout.h +++ b/fs/nfs/flexfilelayout/flexfilelayout.h @@ -191,4 +191,5 @@ nfs4_ff_find_or_create_ds_client(struct pnfs_layout_segment *lseg, struct rpc_cred *ff_layout_get_ds_cred(struct pnfs_layout_segment *lseg, u32 ds_idx, struct rpc_cred *mdscred); bool ff_layout_has_available_ds(struct pnfs_layout_segment *lseg); +bool ff_layout_avoid_mds_available_ds(struct pnfs_layout_segment *lseg); #endif /* FS_NFS_NFS4FLEXFILELAYOUT_H */ diff --git a/fs/nfs/flexfilelayout/flexfilelayoutdev.c b/fs/nfs/flexfilelayout/flexfilelayoutdev.c index eeef89359ad2..8242ab327fee 100644 --- a/fs/nfs/flexfilelayout/flexfilelayoutdev.c +++ b/fs/nfs/flexfilelayout/flexfilelayoutdev.c @@ -541,6 +541,12 @@ bool ff_layout_has_available_ds(struct pnfs_layout_segment *lseg) return ff_rw_layout_has_available_ds(lseg); } +bool ff_layout_avoid_mds_available_ds(struct pnfs_layout_segment *lseg) +{ + return ff_layout_no_fallback_to_mds(lseg) || + ff_layout_has_available_ds(lseg); +} + module_param(dataserver_retrans, uint, 0644); MODULE_PARM_DESC(dataserver_retrans, "The number of times the NFSv4.1 client " "retries a request before it attempts further " -- cgit v1.3-8-gc7d7 From 446ca2195303c550b1cec7dcf9a478910dc84672 Mon Sep 17 00:00:00 2001 From: Tom Haynes Date: Tue, 17 May 2016 12:28:38 -0400 Subject: pNFS/flexfiles: When initing reads or writes, we might have to retry connecting to DSes If we are initializing reads or writes and can not connect to a DS, then check whether or not IO is allowed through the MDS. If it is allowed, reset to the MDS. Else, fail the layout segment and force a retry of a new layout segment. Signed-off-by: Tom Haynes Signed-off-by: Anna Schumaker --- fs/nfs/flexfilelayout/flexfilelayout.c | 29 +++++++++++++++++++++++++---- 1 file changed, 25 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/nfs/flexfilelayout/flexfilelayout.c b/fs/nfs/flexfilelayout/flexfilelayout.c index a835b75f85ab..2911ceda8cd9 100644 --- a/fs/nfs/flexfilelayout/flexfilelayout.c +++ b/fs/nfs/flexfilelayout/flexfilelayout.c @@ -848,8 +848,13 @@ ff_layout_pg_init_read(struct nfs_pageio_descriptor *pgio, goto out_mds; ds = ff_layout_choose_best_ds_for_read(pgio->pg_lseg, 0, &ds_idx); - if (!ds) - goto out_mds; + if (!ds) { + if (ff_layout_no_fallback_to_mds(pgio->pg_lseg)) + goto out_pnfs; + else + goto out_mds; + } + mirror = FF_LAYOUT_COMP(pgio->pg_lseg, ds_idx); pgio->pg_mirror_idx = ds_idx; @@ -863,6 +868,12 @@ out_mds: pnfs_put_lseg(pgio->pg_lseg); pgio->pg_lseg = NULL; nfs_pageio_reset_read_mds(pgio); + return; + +out_pnfs: + pnfs_set_lo_fail(pgio->pg_lseg); + pnfs_put_lseg(pgio->pg_lseg); + pgio->pg_lseg = NULL; } static void @@ -905,8 +916,12 @@ ff_layout_pg_init_write(struct nfs_pageio_descriptor *pgio, for (i = 0; i < pgio->pg_mirror_count; i++) { ds = nfs4_ff_layout_prepare_ds(pgio->pg_lseg, i, true); - if (!ds) - goto out_mds; + if (!ds) { + if (ff_layout_no_fallback_to_mds(pgio->pg_lseg)) + goto out_pnfs; + else + goto out_mds; + } pgm = &pgio->pg_mirrors[i]; mirror = FF_LAYOUT_COMP(pgio->pg_lseg, i); pgm->pg_bsize = mirror->mirror_ds->ds_versions[0].wsize; @@ -918,6 +933,12 @@ out_mds: pnfs_put_lseg(pgio->pg_lseg); pgio->pg_lseg = NULL; nfs_pageio_reset_write_mds(pgio); + return; + +out_pnfs: + pnfs_set_lo_fail(pgio->pg_lseg); + pnfs_put_lseg(pgio->pg_lseg); + pgio->pg_lseg = NULL; } static unsigned int -- cgit v1.3-8-gc7d7 From ee26bdd680aa10df628499fe95a7b88345574248 Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Tue, 17 May 2016 12:28:39 -0400 Subject: pnfs: don't merge new ff lsegs with ones that have LAYOUTRETURN bit set Otherwise, we'll end up returning layouts that we've just received if the client issues a new LAYOUTGET prior to the LAYOUTRETURN. Signed-off-by: Jeff Layton Signed-off-by: Anna Schumaker --- fs/nfs/flexfilelayout/flexfilelayout.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/nfs/flexfilelayout/flexfilelayout.c b/fs/nfs/flexfilelayout/flexfilelayout.c index 2911ceda8cd9..3bac2094c175 100644 --- a/fs/nfs/flexfilelayout/flexfilelayout.c +++ b/fs/nfs/flexfilelayout/flexfilelayout.c @@ -299,6 +299,8 @@ ff_lseg_merge(struct pnfs_layout_segment *new, { u64 new_end, old_end; + if (test_bit(NFS_LSEG_LAYOUTRETURN, &old->pls_flags)) + return false; if (new->pls_range.iomode != old->pls_range.iomode) return false; old_end = pnfs_calc_offset_end(old->pls_range.offset, @@ -319,8 +321,6 @@ ff_lseg_merge(struct pnfs_layout_segment *new, new_end); if (test_bit(NFS_LSEG_ROC, &old->pls_flags)) set_bit(NFS_LSEG_ROC, &new->pls_flags); - if (test_bit(NFS_LSEG_LAYOUTRETURN, &old->pls_flags)) - set_bit(NFS_LSEG_LAYOUTRETURN, &new->pls_flags); return true; } -- cgit v1.3-8-gc7d7 From 6675528380f8b7fffa82047211d8352dc6797321 Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Tue, 17 May 2016 12:28:40 -0400 Subject: pnfs: record sequence in pnfs_layout_segment when it's created In later patches, we're going to teach the client to be more selective about how it returns layouts. This means keeping a record of what the stateid's seqid was at the time that the server handed out a layout segment. Signed-off-by: Jeff Layton Signed-off-by: Anna Schumaker --- fs/nfs/pnfs.c | 1 + fs/nfs/pnfs.h | 1 + 2 files changed, 2 insertions(+) (limited to 'fs') diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index faea98281c75..c09ced88bd38 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -1697,6 +1697,7 @@ pnfs_layout_process(struct nfs4_layoutget *lgp) init_lseg(lo, lseg); lseg->pls_range = res->range; + lseg->pls_seq = be32_to_cpu(res->stateid.seqid); spin_lock(&ino->i_lock); if (pnfs_layoutgets_blocked(lo)) { diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h index 7222d3a35439..361fa5494aa5 100644 --- a/fs/nfs/pnfs.h +++ b/fs/nfs/pnfs.h @@ -64,6 +64,7 @@ struct pnfs_layout_segment { struct list_head pls_lc_list; struct pnfs_layout_range pls_range; atomic_t pls_refcount; + u32 pls_seq; unsigned long pls_flags; struct pnfs_layout_hdr *pls_layout; struct work_struct pls_work; -- cgit v1.3-8-gc7d7 From 3982a6a2d0e64441c938f4e7f93985aff37d8fb3 Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Tue, 17 May 2016 12:28:41 -0400 Subject: pnfs: keep track of the return sequence number in pnfs_layout_hdr When we want to selectively do a LAYOUTRETURN, we need to specify a stateid that represents most recent layout acquisition that is to be returned. When we mark a layout stateid to be returned, we update the return sequence number in the layout header with that value, if it's newer than the existing one. Then, when we go to do a LAYOUTRETURN on layout header put, we overwrite the seqid in the stateid with the saved one, and then zero it out. Signed-off-by: Jeff Layton Signed-off-by: Anna Schumaker --- fs/nfs/pnfs.c | 11 ++++++++--- fs/nfs/pnfs.h | 1 + 2 files changed, 9 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index c09ced88bd38..3ad6bd377d35 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -899,6 +899,7 @@ pnfs_prepare_layoutreturn(struct pnfs_layout_hdr *lo) if (test_and_set_bit(NFS_LAYOUT_RETURN, &lo->plh_flags)) return false; lo->plh_return_iomode = 0; + lo->plh_return_seq = 0; pnfs_get_layout_hdr(lo); clear_bit(NFS_LAYOUT_RETURN_REQUESTED, &lo->plh_flags); return true; @@ -969,6 +970,7 @@ static void pnfs_layoutreturn_before_put_layout_hdr(struct pnfs_layout_hdr *lo) bool send; nfs4_stateid_copy(&stateid, &lo->plh_stateid); + stateid.seqid = cpu_to_be32(lo->plh_return_seq); iomode = lo->plh_return_iomode; send = pnfs_prepare_layoutreturn(lo); spin_unlock(&inode->i_lock); @@ -1747,7 +1749,8 @@ out_forget_reply: } static void -pnfs_set_plh_return_iomode(struct pnfs_layout_hdr *lo, enum pnfs_iomode iomode) +pnfs_set_plh_return_info(struct pnfs_layout_hdr *lo, enum pnfs_iomode iomode, + u32 seq) { if (lo->plh_return_iomode == iomode) return; @@ -1755,6 +1758,8 @@ pnfs_set_plh_return_iomode(struct pnfs_layout_hdr *lo, enum pnfs_iomode iomode) iomode = IOMODE_ANY; lo->plh_return_iomode = iomode; set_bit(NFS_LAYOUT_RETURN_REQUESTED, &lo->plh_flags); + if (!lo->plh_return_seq || pnfs_seqid_is_newer(seq, lo->plh_return_seq)) + lo->plh_return_seq = seq; } /** @@ -1793,7 +1798,7 @@ pnfs_mark_matching_lsegs_return(struct pnfs_layout_hdr *lo, continue; remaining++; set_bit(NFS_LSEG_LAYOUTRETURN, &lseg->pls_flags); - pnfs_set_plh_return_iomode(lo, return_range->iomode); + pnfs_set_plh_return_info(lo, return_range->iomode, lseg->pls_seq); } return remaining; } @@ -1811,7 +1816,7 @@ void pnfs_error_mark_layout_for_return(struct inode *inode, bool return_now = false; spin_lock(&inode->i_lock); - pnfs_set_plh_return_iomode(lo, range.iomode); + pnfs_set_plh_return_info(lo, range.iomode, lseg->pls_seq); /* * mark all matching lsegs so that we are sure to have no live * segments at hand when sending layoutreturn. See pnfs_put_lseg() diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h index 361fa5494aa5..3476c9850678 100644 --- a/fs/nfs/pnfs.h +++ b/fs/nfs/pnfs.h @@ -195,6 +195,7 @@ struct pnfs_layout_hdr { unsigned long plh_flags; nfs4_stateid plh_stateid; u32 plh_barrier; /* ignore lower seqids */ + u32 plh_return_seq; enum pnfs_iomode plh_return_iomode; loff_t plh_lwb; /* last write byte for layoutcommit */ struct rpc_cred *plh_lc_cred; /* layoutcommit cred */ -- cgit v1.3-8-gc7d7 From 6d597e1750126a2c6fd99e4b2cc176955bb9972e Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Tue, 17 May 2016 12:28:42 -0400 Subject: pnfs: only tear down lsegs that precede seqid in LAYOUTRETURN args LAYOUTRETURN is "special" in that servers and clients are expected to work with old stateids. When the client sends a LAYOUTRETURN with an old stateid in it then the server is expected to only tear down layout segments that were present when that seqid was current. Ensure that the client handles its accounting accordingly. Signed-off-by: Jeff Layton Signed-off-by: Anna Schumaker --- fs/nfs/callback_proc.c | 3 ++- fs/nfs/nfs42proc.c | 2 +- fs/nfs/nfs4proc.c | 5 ++-- fs/nfs/pnfs.c | 64 +++++++++++++++++++++++++++++++++----------------- fs/nfs/pnfs.h | 6 +++-- 5 files changed, 52 insertions(+), 28 deletions(-) (limited to 'fs') diff --git a/fs/nfs/callback_proc.c b/fs/nfs/callback_proc.c index 7c9fbf504f07..aaa2e8d3df6f 100644 --- a/fs/nfs/callback_proc.c +++ b/fs/nfs/callback_proc.c @@ -217,7 +217,8 @@ static u32 initiate_file_draining(struct nfs_client *clp, } if (pnfs_mark_matching_lsegs_return(lo, &free_me_list, - &args->cbl_range)) { + &args->cbl_range, + be32_to_cpu(args->cbl_stateid.seqid))) { rv = NFS4_OK; goto unlock; } diff --git a/fs/nfs/nfs42proc.c b/fs/nfs/nfs42proc.c index 579ee20e4120..20c44d1209dc 100644 --- a/fs/nfs/nfs42proc.c +++ b/fs/nfs/nfs42proc.c @@ -337,7 +337,7 @@ nfs42_layoutstat_done(struct rpc_task *task, void *calldata) * with the current stateid. */ set_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags); - pnfs_mark_matching_lsegs_invalid(lo, &head, NULL); + pnfs_mark_matching_lsegs_invalid(lo, &head, NULL, 0); spin_unlock(&inode->i_lock); pnfs_free_lseg_list(&head); } else diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index c9b66085d392..2a29f5d12aeb 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -7953,7 +7953,7 @@ static void nfs4_layoutget_done(struct rpc_task *task, void *calldata) * with the current stateid. */ set_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags); - pnfs_mark_matching_lsegs_invalid(lo, &head, NULL); + pnfs_mark_matching_lsegs_invalid(lo, &head, NULL, 0); spin_unlock(&inode->i_lock); pnfs_free_lseg_list(&head); } else @@ -8145,7 +8145,8 @@ static void nfs4_layoutreturn_release(void *calldata) dprintk("--> %s\n", __func__); spin_lock(&lo->plh_inode->i_lock); - pnfs_mark_matching_lsegs_invalid(lo, &freeme, &lrp->args.range); + pnfs_mark_matching_lsegs_invalid(lo, &freeme, &lrp->args.range, + be32_to_cpu(lrp->args.stateid.seqid)); pnfs_mark_layout_returned_if_empty(lo); if (lrp->res.lrs_present) pnfs_set_layout_stateid(lo, &lrp->res.stateid, true); diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index 3ad6bd377d35..21ff53fe7a9e 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -270,7 +270,7 @@ pnfs_mark_layout_stateid_invalid(struct pnfs_layout_hdr *lo, }; set_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags); - return pnfs_mark_matching_lsegs_invalid(lo, lseg_list, &range); + return pnfs_mark_matching_lsegs_invalid(lo, lseg_list, &range, 0); } static int @@ -308,7 +308,7 @@ pnfs_layout_io_set_failed(struct pnfs_layout_hdr *lo, u32 iomode) spin_lock(&inode->i_lock); pnfs_layout_set_fail_bit(lo, pnfs_iomode_to_fail_bit(iomode)); - pnfs_mark_matching_lsegs_invalid(lo, &head, &range); + pnfs_mark_matching_lsegs_invalid(lo, &head, &range, 0); spin_unlock(&inode->i_lock); pnfs_free_lseg_list(&head); dprintk("%s Setting layout IOMODE_%s fail bit\n", __func__, @@ -522,13 +522,35 @@ static int mark_lseg_invalid(struct pnfs_layout_segment *lseg, return rv; } -/* Returns count of number of matching invalid lsegs remaining in list - * after call. +/* + * Compare 2 layout stateid sequence ids, to see which is newer, + * taking into account wraparound issues. + */ +static bool pnfs_seqid_is_newer(u32 s1, u32 s2) +{ + return (s32)(s1 - s2) > 0; +} + +/** + * pnfs_mark_matching_lsegs_invalid - tear down lsegs or mark them for later + * @lo: layout header containing the lsegs + * @tmp_list: list head where doomed lsegs should go + * @recall_range: optional recall range argument to match (may be NULL) + * @seq: only invalidate lsegs obtained prior to this sequence (may be 0) + * + * Walk the list of lsegs in the layout header, and tear down any that should + * be destroyed. If "recall_range" is specified then the segment must match + * that range. If "seq" is non-zero, then only match segments that were handed + * out at or before that sequence. + * + * Returns number of matching invalid lsegs remaining in list after scanning + * it and purging them. */ int pnfs_mark_matching_lsegs_invalid(struct pnfs_layout_hdr *lo, struct list_head *tmp_list, - const struct pnfs_layout_range *recall_range) + const struct pnfs_layout_range *recall_range, + u32 seq) { struct pnfs_layout_segment *lseg, *next; int remaining = 0; @@ -540,10 +562,12 @@ pnfs_mark_matching_lsegs_invalid(struct pnfs_layout_hdr *lo, list_for_each_entry_safe(lseg, next, &lo->plh_segs, pls_list) if (!recall_range || should_free_lseg(&lseg->pls_range, recall_range)) { - dprintk("%s: freeing lseg %p iomode %d " + if (seq && pnfs_seqid_is_newer(lseg->pls_seq, seq)) + continue; + dprintk("%s: freeing lseg %p iomode %d seq %u" "offset %llu length %llu\n", __func__, - lseg, lseg->pls_range.iomode, lseg->pls_range.offset, - lseg->pls_range.length); + lseg, lseg->pls_range.iomode, lseg->pls_seq, + lseg->pls_range.offset, lseg->pls_range.length); if (!mark_lseg_invalid(lseg, tmp_list)) remaining++; } @@ -730,15 +754,6 @@ pnfs_destroy_all_layouts(struct nfs_client *clp) pnfs_destroy_layouts_byclid(clp, false); } -/* - * Compare 2 layout stateid sequence ids, to see which is newer, - * taking into account wraparound issues. - */ -static bool pnfs_seqid_is_newer(u32 s1, u32 s2) -{ - return (s32)(s1 - s2) > 0; -} - /* update lo->plh_stateid with new if is more recent */ void pnfs_set_layout_stateid(struct pnfs_layout_hdr *lo, const nfs4_stateid *new, @@ -1014,7 +1029,7 @@ _pnfs_return_layout(struct inode *ino) pnfs_get_layout_hdr(lo); empty = list_empty(&lo->plh_segs); pnfs_clear_layoutcommit(ino, &tmp_list); - pnfs_mark_matching_lsegs_invalid(lo, &tmp_list, NULL); + pnfs_mark_matching_lsegs_invalid(lo, &tmp_list, NULL, 0); if (NFS_SERVER(ino)->pnfs_curr_ld->return_range) { struct pnfs_layout_range range = { @@ -1721,7 +1736,7 @@ pnfs_layout_process(struct nfs4_layoutget *lgp) * inode invalid, and don't bother validating the stateid * sequence number. */ - pnfs_mark_matching_lsegs_invalid(lo, &free_me, NULL); + pnfs_mark_matching_lsegs_invalid(lo, &free_me, NULL, 0); nfs4_stateid_copy(&lo->plh_stateid, &res->stateid); lo->plh_barrier = be32_to_cpu(res->stateid.seqid); @@ -1775,7 +1790,8 @@ pnfs_set_plh_return_info(struct pnfs_layout_hdr *lo, enum pnfs_iomode iomode, int pnfs_mark_matching_lsegs_return(struct pnfs_layout_hdr *lo, struct list_head *tmp_list, - const struct pnfs_layout_range *return_range) + const struct pnfs_layout_range *return_range, + u32 seq) { struct pnfs_layout_segment *lseg, *next; int remaining = 0; @@ -1798,8 +1814,11 @@ pnfs_mark_matching_lsegs_return(struct pnfs_layout_hdr *lo, continue; remaining++; set_bit(NFS_LSEG_LAYOUTRETURN, &lseg->pls_flags); - pnfs_set_plh_return_info(lo, return_range->iomode, lseg->pls_seq); } + + if (remaining) + pnfs_set_plh_return_info(lo, return_range->iomode, seq); + return remaining; } @@ -1822,7 +1841,8 @@ void pnfs_error_mark_layout_for_return(struct inode *inode, * segments at hand when sending layoutreturn. See pnfs_put_lseg() * for how it works. */ - if (!pnfs_mark_matching_lsegs_return(lo, &free_me, &range)) { + if (!pnfs_mark_matching_lsegs_return(lo, &free_me, + &range, lseg->pls_seq)) { nfs4_stateid stateid; enum pnfs_iomode iomode = lo->plh_return_iomode; diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h index 3476c9850678..971068b58647 100644 --- a/fs/nfs/pnfs.h +++ b/fs/nfs/pnfs.h @@ -266,10 +266,12 @@ int pnfs_choose_layoutget_stateid(nfs4_stateid *dst, struct nfs4_state *open_state); int pnfs_mark_matching_lsegs_invalid(struct pnfs_layout_hdr *lo, struct list_head *tmp_list, - const struct pnfs_layout_range *recall_range); + const struct pnfs_layout_range *recall_range, + u32 seq); int pnfs_mark_matching_lsegs_return(struct pnfs_layout_hdr *lo, struct list_head *tmp_list, - const struct pnfs_layout_range *recall_range); + const struct pnfs_layout_range *recall_range, + u32 seq); bool pnfs_roc(struct inode *ino); void pnfs_roc_release(struct inode *ino); void pnfs_roc_set_barrier(struct inode *ino, u32 barrier); -- cgit v1.3-8-gc7d7 From 094069f1d96f691637b5c335f5eb79566d2457ff Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Tue, 17 May 2016 12:28:43 -0400 Subject: flexfiles: remove pointless setting of NFS_LAYOUT_RETURN_REQUESTED Setting just the NFS_LAYOUT_RETURN_REQUESTED flag doesn't do anything, unless there are lsegs that are also being marked for return. At the point where that happens this flag is also set, so these set_bit calls don't do anything useful. Signed-off-by: Jeff Layton Signed-off-by: Anna Schumaker --- fs/nfs/flexfilelayout/flexfilelayout.c | 2 -- fs/nfs/flexfilelayout/flexfilelayoutdev.c | 8 +------- 2 files changed, 1 insertion(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/nfs/flexfilelayout/flexfilelayout.c b/fs/nfs/flexfilelayout/flexfilelayout.c index 3bac2094c175..40bccb2c06d1 100644 --- a/fs/nfs/flexfilelayout/flexfilelayout.c +++ b/fs/nfs/flexfilelayout/flexfilelayout.c @@ -1270,8 +1270,6 @@ static int ff_layout_read_done_cb(struct rpc_task *task, hdr->pgio_mirror_idx + 1, &hdr->pgio_mirror_idx)) goto out_eagain; - set_bit(NFS_LAYOUT_RETURN_REQUESTED, - &hdr->lseg->pls_layout->plh_flags); pnfs_read_resend_pnfs(hdr); return task->tk_status; case -NFS4ERR_RESET_TO_MDS: diff --git a/fs/nfs/flexfilelayout/flexfilelayoutdev.c b/fs/nfs/flexfilelayout/flexfilelayoutdev.c index 8242ab327fee..bb598923b2ef 100644 --- a/fs/nfs/flexfilelayout/flexfilelayoutdev.c +++ b/fs/nfs/flexfilelayout/flexfilelayoutdev.c @@ -394,13 +394,7 @@ nfs4_ff_layout_prepare_ds(struct pnfs_layout_segment *lseg, u32 ds_idx, mirror, lseg->pls_range.offset, lseg->pls_range.length, NFS4ERR_NXIO, OP_ILLEGAL, GFP_NOIO); - if (!fail_return) { - if (ff_layout_has_available_ds(lseg)) - set_bit(NFS_LAYOUT_RETURN_REQUESTED, - &lseg->pls_layout->plh_flags); - else - pnfs_error_mark_layout_for_return(ino, lseg); - } else + if (fail_return || !ff_layout_has_available_ds(lseg)) pnfs_error_mark_layout_for_return(ino, lseg); ds = NULL; } -- cgit v1.3-8-gc7d7 From 95e2b7e95d43c5c95620b12355a90713268376d3 Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Tue, 17 May 2016 12:28:44 -0400 Subject: flexfiles: add kerneldoc header to nfs4_ff_layout_prepare_ds Signed-off-by: Jeff Layton Signed-off-by: Anna Schumaker --- fs/nfs/flexfilelayout/flexfilelayoutdev.c | 18 +++++++++++++++++- 1 file changed, 17 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/flexfilelayout/flexfilelayoutdev.c b/fs/nfs/flexfilelayout/flexfilelayoutdev.c index bb598923b2ef..35d84d0c0f21 100644 --- a/fs/nfs/flexfilelayout/flexfilelayoutdev.c +++ b/fs/nfs/flexfilelayout/flexfilelayoutdev.c @@ -343,7 +343,23 @@ out: return fh; } -/* Upon return, either ds is connected, or ds is NULL */ +/** + * nfs4_ff_layout_prepare_ds - prepare a DS connection for an RPC call + * @lseg: the layout segment we're operating on + * @ds_idx: index of the DS to use + * @fail_return: return layout on connect failure? + * + * Try to prepare a DS connection to accept an RPC call. This involves + * selecting a mirror to use and connecting the client to it if it's not + * already connected. + * + * Since we only need a single functioning mirror to satisfy a read, we don't + * want to return the layout if there is one. For writes though, any down + * mirror should result in a LAYOUTRETURN. @fail_return is how we distinguish + * between the two cases. + * + * Returns a pointer to a connected DS object on success or NULL on failure. + */ struct nfs4_pnfs_ds * nfs4_ff_layout_prepare_ds(struct pnfs_layout_segment *lseg, u32 ds_idx, bool fail_return) -- cgit v1.3-8-gc7d7 From d03ab29dbbe905c6c7c5abd78e02547fa954ec07 Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Tue, 17 May 2016 12:28:45 -0400 Subject: pnfs: fix bad error handling in send_layoutget Currently, the code will clear the fail bit if we get back a fatal error. I don't think that's correct -- we want to clear that bit if we do not get a fatal error. Fixes: 0bcbf039f6 (nfs: handle request add failure properly) Signed-off-by: Jeff Layton Signed-off-by: Anna Schumaker --- fs/nfs/pnfs.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index 21ff53fe7a9e..e3e9973b5f76 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -876,11 +876,16 @@ send_layoutget(struct pnfs_layout_hdr *lo, lseg = nfs4_proc_layoutget(lgp, gfp_flags); } while (lseg == ERR_PTR(-EAGAIN)); - if (IS_ERR(lseg) && !nfs_error_is_fatal(PTR_ERR(lseg))) - lseg = NULL; - else + if (IS_ERR(lseg)) { + if (!nfs_error_is_fatal(PTR_ERR(lseg))) { + pnfs_layout_clear_fail_bit(lo, + pnfs_iomode_to_fail_bit(range->iomode)); + lseg = NULL; + } + } else { pnfs_layout_clear_fail_bit(lo, pnfs_iomode_to_fail_bit(range->iomode)); + } return lseg; } -- cgit v1.3-8-gc7d7 From 83026d80a16ea6a4e4f06a994fc7831b1d8d6375 Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Tue, 17 May 2016 12:28:46 -0400 Subject: pnfs: lift retry logic from send_layoutget to pnfs_update_layout If we get back something like NFS4ERR_OLD_STATEID, that will be translated into -EAGAIN, and the do/while loop in send_layoutget will drive the call again. This is not quite what we want, I think. An error like that is a sign that something has changed. That something could have been a concurrent LAYOUTGET that would give us a usable lseg. Lift the retry logic into pnfs_update_layout instead. That allows us to redo the layout search, and may spare us from having to issue an RPC. Signed-off-by: Jeff Layton Signed-off-by: Anna Schumaker --- fs/nfs/pnfs.c | 72 +++++++++++++++++++++++++++++------------------------------ 1 file changed, 36 insertions(+), 36 deletions(-) (limited to 'fs') diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index e3e9973b5f76..2a609af845fe 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -839,7 +839,6 @@ send_layoutget(struct pnfs_layout_hdr *lo, struct inode *ino = lo->plh_inode; struct nfs_server *server = NFS_SERVER(ino); struct nfs4_layoutget *lgp; - struct pnfs_layout_segment *lseg; loff_t i_size; dprintk("--> %s\n", __func__); @@ -849,45 +848,30 @@ send_layoutget(struct pnfs_layout_hdr *lo, * store in lseg. If we race with a concurrent seqid morphing * op, then re-send the LAYOUTGET. */ - do { - lgp = kzalloc(sizeof(*lgp), gfp_flags); - if (lgp == NULL) - return NULL; - - i_size = i_size_read(ino); - - lgp->args.minlength = PAGE_SIZE; - if (lgp->args.minlength > range->length) - lgp->args.minlength = range->length; - if (range->iomode == IOMODE_READ) { - if (range->offset >= i_size) - lgp->args.minlength = 0; - else if (i_size - range->offset < lgp->args.minlength) - lgp->args.minlength = i_size - range->offset; - } - lgp->args.maxcount = PNFS_LAYOUT_MAXSIZE; - pnfs_copy_range(&lgp->args.range, range); - lgp->args.type = server->pnfs_curr_ld->id; - lgp->args.inode = ino; - lgp->args.ctx = get_nfs_open_context(ctx); - lgp->gfp_flags = gfp_flags; - lgp->cred = lo->plh_lc_cred; + lgp = kzalloc(sizeof(*lgp), gfp_flags); + if (lgp == NULL) + return ERR_PTR(-ENOMEM); - lseg = nfs4_proc_layoutget(lgp, gfp_flags); - } while (lseg == ERR_PTR(-EAGAIN)); + i_size = i_size_read(ino); - if (IS_ERR(lseg)) { - if (!nfs_error_is_fatal(PTR_ERR(lseg))) { - pnfs_layout_clear_fail_bit(lo, - pnfs_iomode_to_fail_bit(range->iomode)); - lseg = NULL; - } - } else { - pnfs_layout_clear_fail_bit(lo, - pnfs_iomode_to_fail_bit(range->iomode)); + lgp->args.minlength = PAGE_SIZE; + if (lgp->args.minlength > range->length) + lgp->args.minlength = range->length; + if (range->iomode == IOMODE_READ) { + if (range->offset >= i_size) + lgp->args.minlength = 0; + else if (i_size - range->offset < lgp->args.minlength) + lgp->args.minlength = i_size - range->offset; } + lgp->args.maxcount = PNFS_LAYOUT_MAXSIZE; + pnfs_copy_range(&lgp->args.range, range); + lgp->args.type = server->pnfs_curr_ld->id; + lgp->args.inode = ino; + lgp->args.ctx = get_nfs_open_context(ctx); + lgp->gfp_flags = gfp_flags; + lgp->cred = lo->plh_lc_cred; - return lseg; + return nfs4_proc_layoutget(lgp, gfp_flags); } static void pnfs_clear_layoutcommit(struct inode *inode, @@ -1649,6 +1633,22 @@ lookup_again: arg.length = PAGE_ALIGN(arg.length); lseg = send_layoutget(lo, ctx, &arg, gfp_flags); + if (IS_ERR(lseg)) { + if (lseg == ERR_PTR(-EAGAIN)) { + if (first) + pnfs_clear_first_layoutget(lo); + pnfs_put_layout_hdr(lo); + goto lookup_again; + } + + if (!nfs_error_is_fatal(PTR_ERR(lseg))) { + pnfs_layout_clear_fail_bit(lo, pnfs_iomode_to_fail_bit(iomode)); + lseg = NULL; + } + } else { + pnfs_layout_clear_fail_bit(lo, pnfs_iomode_to_fail_bit(iomode)); + } + atomic_dec(&lo->plh_outstanding); trace_pnfs_update_layout(ino, pos, count, iomode, lo, PNFS_UPDATE_LAYOUT_SEND_LAYOUTGET); -- cgit v1.3-8-gc7d7 From 183d9e7b112aaed0d19c16ffcf0f8c3a86dc71e0 Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Tue, 17 May 2016 12:28:47 -0400 Subject: pnfs: rework LAYOUTGET retry handling There are several problems in the way a stateid is selected for a LAYOUTGET operation: We pick a stateid to use in the RPC prepare op, but that makes it difficult to serialize LAYOUTGETs that use the open stateid. That serialization is done in pnfs_update_layout, which occurs well before the rpc_prepare operation. Between those two events, the i_lock is dropped and reacquired. pnfs_update_layout can find that the list has lsegs in it and not do any serialization, but then later pnfs_choose_layoutget_stateid ends up choosing the open stateid. This patch changes the client to select the stateid to use in the LAYOUTGET earlier, when we're searching for a usable layout segment. This way we can do it all while holding the i_lock the first time, and ensure that we serialize any LAYOUTGET call that uses a non-layout stateid. This also means a rework of how LAYOUTGET replies are handled, as we must now get the latest stateid if we want to retransmit in response to a retryable error. Most of those errors boil down to the fact that the layout state has changed in some fashion. Thus, what we really want to do is to re-search for a layout when it fails with a retryable error, so that we can avoid reissuing the RPC at all if possible. While the LAYOUTGET RPC is async, the initiating thread always waits for it to complete, so it's effectively synchronous anyway. Currently, when we need to retry a LAYOUTGET because of an error, we drive that retry via the rpc state machine. This means that once the call has been submitted, it runs until it completes. So, we must move the error handling for this RPC out of the rpc_call_done operation and into the caller. In order to handle errors like NFS4ERR_DELAY properly, we must also pass a pointer to the sliding timeout, which is now moved to the stack in pnfs_update_layout. The complicating errors are -NFS4ERR_RECALLCONFLICT and -NFS4ERR_LAYOUTTRYLATER, as those involve a timeout after which we give up and return NULL back to the caller. So, there is some special handling for those errors to ensure that the layers driving the retries can handle that appropriately. Signed-off-by: Jeff Layton Signed-off-by: Anna Schumaker --- fs/nfs/nfs4proc.c | 115 ++++++++++++++++---------------------- fs/nfs/nfs4trace.h | 10 +++- fs/nfs/pnfs.c | 144 +++++++++++++++++++++++++----------------------- fs/nfs/pnfs.h | 6 +- include/linux/errno.h | 1 + include/linux/nfs4.h | 2 + include/linux/nfs_xdr.h | 2 - 7 files changed, 136 insertions(+), 144 deletions(-) (limited to 'fs') diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 2a29f5d12aeb..62d67f040906 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -427,6 +427,7 @@ static int nfs4_do_handle_exception(struct nfs_server *server, case -NFS4ERR_DELAY: nfs_inc_server_stats(server, NFSIOS_DELAY); case -NFS4ERR_GRACE: + case -NFS4ERR_RECALLCONFLICT: exception->delay = 1; return 0; @@ -7847,40 +7848,34 @@ nfs4_layoutget_prepare(struct rpc_task *task, void *calldata) struct nfs4_layoutget *lgp = calldata; struct nfs_server *server = NFS_SERVER(lgp->args.inode); struct nfs4_session *session = nfs4_get_session(server); - int ret; dprintk("--> %s\n", __func__); - /* Note the is a race here, where a CB_LAYOUTRECALL can come in - * right now covering the LAYOUTGET we are about to send. - * However, that is not so catastrophic, and there seems - * to be no way to prevent it completely. - */ - if (nfs41_setup_sequence(session, &lgp->args.seq_args, - &lgp->res.seq_res, task)) - return; - ret = pnfs_choose_layoutget_stateid(&lgp->args.stateid, - NFS_I(lgp->args.inode)->layout, - &lgp->args.range, - lgp->args.ctx->state); - if (ret < 0) - rpc_exit(task, ret); + nfs41_setup_sequence(session, &lgp->args.seq_args, + &lgp->res.seq_res, task); + dprintk("<-- %s\n", __func__); } static void nfs4_layoutget_done(struct rpc_task *task, void *calldata) { struct nfs4_layoutget *lgp = calldata; + + dprintk("--> %s\n", __func__); + nfs41_sequence_done(task, &lgp->res.seq_res); + dprintk("<-- %s\n", __func__); +} + +static int +nfs4_layoutget_handle_exception(struct rpc_task *task, + struct nfs4_layoutget *lgp, struct nfs4_exception *exception) +{ struct inode *inode = lgp->args.inode; struct nfs_server *server = NFS_SERVER(inode); struct pnfs_layout_hdr *lo; - struct nfs4_state *state = NULL; - unsigned long timeo, now, giveup; + int status = task->tk_status; dprintk("--> %s tk_status => %d\n", __func__, -task->tk_status); - if (!nfs41_sequence_done(task, &lgp->res.seq_res)) - goto out; - - switch (task->tk_status) { + switch (status) { case 0: goto out; @@ -7890,57 +7885,43 @@ static void nfs4_layoutget_done(struct rpc_task *task, void *calldata) * retry go inband. */ case -NFS4ERR_LAYOUTUNAVAILABLE: - task->tk_status = -ENODATA; + status = -ENODATA; goto out; /* * NFS4ERR_BADLAYOUT means the MDS cannot return a layout of * length lgp->args.minlength != 0 (see RFC5661 section 18.43.3). */ case -NFS4ERR_BADLAYOUT: - goto out_overflow; + status = -EOVERFLOW; + goto out; /* * NFS4ERR_LAYOUTTRYLATER is a conflict with another client * (or clients) writing to the same RAID stripe except when * the minlength argument is 0 (see RFC5661 section 18.43.3). + * + * Treat it like we would RECALLCONFLICT -- we retry for a little + * while, and then eventually give up. */ case -NFS4ERR_LAYOUTTRYLATER: - if (lgp->args.minlength == 0) - goto out_overflow; - /* - * NFS4ERR_RECALLCONFLICT is when conflict with self (must recall - * existing layout before getting a new one). - */ - case -NFS4ERR_RECALLCONFLICT: - timeo = rpc_get_timeout(task->tk_client); - giveup = lgp->args.timestamp + timeo; - now = jiffies; - if (time_after(giveup, now)) { - unsigned long delay; - - /* Delay for: - * - Not less then NFS4_POLL_RETRY_MIN. - * - One last time a jiffie before we give up - * - exponential backoff (time_now minus start_attempt) - */ - delay = max_t(unsigned long, NFS4_POLL_RETRY_MIN, - min((giveup - now - 1), - now - lgp->args.timestamp)); - - dprintk("%s: NFS4ERR_RECALLCONFLICT waiting %lu\n", - __func__, delay); - rpc_delay(task, delay); - /* Do not call nfs4_async_handle_error() */ - goto out_restart; + if (lgp->args.minlength == 0) { + status = -EOVERFLOW; + goto out; } - break; + /* Fallthrough */ + case -NFS4ERR_RECALLCONFLICT: + nfs4_handle_exception(server, -NFS4ERR_RECALLCONFLICT, + exception); + status = -ERECALLCONFLICT; + goto out; case -NFS4ERR_EXPIRED: case -NFS4ERR_BAD_STATEID: + exception->timeout = 0; spin_lock(&inode->i_lock); if (nfs4_stateid_match(&lgp->args.stateid, &lgp->args.ctx->state->stateid)) { spin_unlock(&inode->i_lock); /* If the open stateid was bad, then recover it. */ - state = lgp->args.ctx->state; + exception->state = lgp->args.ctx->state; break; } lo = NFS_I(inode)->layout; @@ -7958,20 +7939,16 @@ static void nfs4_layoutget_done(struct rpc_task *task, void *calldata) pnfs_free_lseg_list(&head); } else spin_unlock(&inode->i_lock); - goto out_restart; + status = -EAGAIN; + goto out; } - if (nfs4_async_handle_error(task, server, state, &lgp->timeout) == -EAGAIN) - goto out_restart; + + status = nfs4_handle_exception(server, status, exception); + if (exception->retry) + status = -EAGAIN; out: dprintk("<-- %s\n", __func__); - return; -out_restart: - task->tk_status = 0; - rpc_restart_call_prepare(task); - return; -out_overflow: - task->tk_status = -EOVERFLOW; - goto out; + return status; } static size_t max_response_pages(struct nfs_server *server) @@ -8040,7 +8017,7 @@ static const struct rpc_call_ops nfs4_layoutget_call_ops = { }; struct pnfs_layout_segment * -nfs4_proc_layoutget(struct nfs4_layoutget *lgp, gfp_t gfp_flags) +nfs4_proc_layoutget(struct nfs4_layoutget *lgp, long *timeout, gfp_t gfp_flags) { struct inode *inode = lgp->args.inode; struct nfs_server *server = NFS_SERVER(inode); @@ -8060,6 +8037,7 @@ nfs4_proc_layoutget(struct nfs4_layoutget *lgp, gfp_t gfp_flags) .flags = RPC_TASK_ASYNC, }; struct pnfs_layout_segment *lseg = NULL; + struct nfs4_exception exception = { .timeout = *timeout }; int status = 0; dprintk("--> %s\n", __func__); @@ -8073,7 +8051,6 @@ nfs4_proc_layoutget(struct nfs4_layoutget *lgp, gfp_t gfp_flags) return ERR_PTR(-ENOMEM); } lgp->args.layout.pglen = max_pages * PAGE_SIZE; - lgp->args.timestamp = jiffies; lgp->res.layoutp = &lgp->args.layout; lgp->res.seq_res.sr_slot = NULL; @@ -8083,13 +8060,17 @@ nfs4_proc_layoutget(struct nfs4_layoutget *lgp, gfp_t gfp_flags) if (IS_ERR(task)) return ERR_CAST(task); status = nfs4_wait_for_completion_rpc_task(task); - if (status == 0) - status = task->tk_status; + if (status == 0) { + status = nfs4_layoutget_handle_exception(task, lgp, &exception); + *timeout = exception.timeout; + } + trace_nfs4_layoutget(lgp->args.ctx, &lgp->args.range, &lgp->res.range, &lgp->res.stateid, status); + /* if layoutp->len is 0, nfs4_layoutget_prepare called rpc_exit */ if (status == 0 && lgp->res.layoutp->len) lseg = pnfs_layout_process(lgp); diff --git a/fs/nfs/nfs4trace.h b/fs/nfs/nfs4trace.h index 2c8d05dae5b1..9c150b153782 100644 --- a/fs/nfs/nfs4trace.h +++ b/fs/nfs/nfs4trace.h @@ -1520,6 +1520,8 @@ DEFINE_NFS4_INODE_EVENT(nfs4_layoutreturn_on_close); { PNFS_UPDATE_LAYOUT_FOUND_CACHED, "found cached" }, \ { PNFS_UPDATE_LAYOUT_RETURN, "layoutreturn" }, \ { PNFS_UPDATE_LAYOUT_BLOCKED, "layouts blocked" }, \ + { PNFS_UPDATE_LAYOUT_INVALID_OPEN, "invalid open" }, \ + { PNFS_UPDATE_LAYOUT_RETRY, "retrying" }, \ { PNFS_UPDATE_LAYOUT_SEND_LAYOUTGET, "sent layoutget" }) TRACE_EVENT(pnfs_update_layout, @@ -1528,9 +1530,10 @@ TRACE_EVENT(pnfs_update_layout, u64 count, enum pnfs_iomode iomode, struct pnfs_layout_hdr *lo, + struct pnfs_layout_segment *lseg, enum pnfs_update_layout_reason reason ), - TP_ARGS(inode, pos, count, iomode, lo, reason), + TP_ARGS(inode, pos, count, iomode, lo, lseg, reason), TP_STRUCT__entry( __field(dev_t, dev) __field(u64, fileid) @@ -1540,6 +1543,7 @@ TRACE_EVENT(pnfs_update_layout, __field(enum pnfs_iomode, iomode) __field(int, layoutstateid_seq) __field(u32, layoutstateid_hash) + __field(long, lseg) __field(enum pnfs_update_layout_reason, reason) ), TP_fast_assign( @@ -1559,11 +1563,12 @@ TRACE_EVENT(pnfs_update_layout, __entry->layoutstateid_seq = 0; __entry->layoutstateid_hash = 0; } + __entry->lseg = (long)lseg; ), TP_printk( "fileid=%02x:%02x:%llu fhandle=0x%08x " "iomode=%s pos=%llu count=%llu " - "layoutstateid=%d:0x%08x (%s)", + "layoutstateid=%d:0x%08x lseg=0x%lx (%s)", MAJOR(__entry->dev), MINOR(__entry->dev), (unsigned long long)__entry->fileid, __entry->fhandle, @@ -1571,6 +1576,7 @@ TRACE_EVENT(pnfs_update_layout, (unsigned long long)__entry->pos, (unsigned long long)__entry->count, __entry->layoutstateid_seq, __entry->layoutstateid_hash, + __entry->lseg, show_pnfs_update_layout_reason(__entry->reason) ) ); diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index 2a609af845fe..46339a7fb191 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -796,45 +796,18 @@ pnfs_layoutgets_blocked(const struct pnfs_layout_hdr *lo) test_bit(NFS_LAYOUT_BULK_RECALL, &lo->plh_flags); } -int -pnfs_choose_layoutget_stateid(nfs4_stateid *dst, struct pnfs_layout_hdr *lo, - const struct pnfs_layout_range *range, - struct nfs4_state *open_state) -{ - int status = 0; - - dprintk("--> %s\n", __func__); - spin_lock(&lo->plh_inode->i_lock); - if (pnfs_layoutgets_blocked(lo)) { - status = -EAGAIN; - } else if (!nfs4_valid_open_stateid(open_state)) { - status = -EBADF; - } else if (list_empty(&lo->plh_segs) || - test_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags)) { - int seq; - - do { - seq = read_seqbegin(&open_state->seqlock); - nfs4_stateid_copy(dst, &open_state->stateid); - } while (read_seqretry(&open_state->seqlock, seq)); - } else - nfs4_stateid_copy(dst, &lo->plh_stateid); - spin_unlock(&lo->plh_inode->i_lock); - dprintk("<-- %s\n", __func__); - return status; -} - /* -* Get layout from server. -* for now, assume that whole file layouts are requested. -* arg->offset: 0 -* arg->length: all ones -*/ + * Get layout from server. + * for now, assume that whole file layouts are requested. + * arg->offset: 0 + * arg->length: all ones + */ static struct pnfs_layout_segment * send_layoutget(struct pnfs_layout_hdr *lo, struct nfs_open_context *ctx, + nfs4_stateid *stateid, const struct pnfs_layout_range *range, - gfp_t gfp_flags) + long *timeout, gfp_t gfp_flags) { struct inode *ino = lo->plh_inode; struct nfs_server *server = NFS_SERVER(ino); @@ -868,10 +841,11 @@ send_layoutget(struct pnfs_layout_hdr *lo, lgp->args.type = server->pnfs_curr_ld->id; lgp->args.inode = ino; lgp->args.ctx = get_nfs_open_context(ctx); + nfs4_stateid_copy(&lgp->args.stateid, stateid); lgp->gfp_flags = gfp_flags; lgp->cred = lo->plh_lc_cred; - return nfs4_proc_layoutget(lgp, gfp_flags); + return nfs4_proc_layoutget(lgp, timeout, gfp_flags); } static void pnfs_clear_layoutcommit(struct inode *inode, @@ -1511,27 +1485,30 @@ pnfs_update_layout(struct inode *ino, .offset = pos, .length = count, }; - unsigned pg_offset; + unsigned pg_offset, seq; struct nfs_server *server = NFS_SERVER(ino); struct nfs_client *clp = server->nfs_client; - struct pnfs_layout_hdr *lo; + struct pnfs_layout_hdr *lo = NULL; struct pnfs_layout_segment *lseg = NULL; + nfs4_stateid stateid; + long timeout = 0; + unsigned long giveup = jiffies + rpc_get_timeout(server->client); bool first; if (!pnfs_enabled_sb(NFS_SERVER(ino))) { - trace_pnfs_update_layout(ino, pos, count, iomode, NULL, + trace_pnfs_update_layout(ino, pos, count, iomode, lo, lseg, PNFS_UPDATE_LAYOUT_NO_PNFS); goto out; } if (iomode == IOMODE_READ && i_size_read(ino) == 0) { - trace_pnfs_update_layout(ino, pos, count, iomode, NULL, + trace_pnfs_update_layout(ino, pos, count, iomode, lo, lseg, PNFS_UPDATE_LAYOUT_RD_ZEROLEN); goto out; } if (pnfs_within_mdsthreshold(ctx, ino, iomode)) { - trace_pnfs_update_layout(ino, pos, count, iomode, NULL, + trace_pnfs_update_layout(ino, pos, count, iomode, lo, lseg, PNFS_UPDATE_LAYOUT_MDSTHRESH); goto out; } @@ -1542,14 +1519,14 @@ lookup_again: lo = pnfs_find_alloc_layout(ino, ctx, gfp_flags); if (lo == NULL) { spin_unlock(&ino->i_lock); - trace_pnfs_update_layout(ino, pos, count, iomode, NULL, + trace_pnfs_update_layout(ino, pos, count, iomode, lo, lseg, PNFS_UPDATE_LAYOUT_NOMEM); goto out; } /* Do we even need to bother with this? */ if (test_bit(NFS_LAYOUT_BULK_RECALL, &lo->plh_flags)) { - trace_pnfs_update_layout(ino, pos, count, iomode, lo, + trace_pnfs_update_layout(ino, pos, count, iomode, lo, lseg, PNFS_UPDATE_LAYOUT_BULK_RECALL); dprintk("%s matches recall, use MDS\n", __func__); goto out_unlock; @@ -1557,14 +1534,34 @@ lookup_again: /* if LAYOUTGET already failed once we don't try again */ if (pnfs_layout_io_test_failed(lo, iomode)) { - trace_pnfs_update_layout(ino, pos, count, iomode, lo, + trace_pnfs_update_layout(ino, pos, count, iomode, lo, lseg, PNFS_UPDATE_LAYOUT_IO_TEST_FAIL); goto out_unlock; } - first = list_empty(&lo->plh_segs); - if (first) { - /* The first layoutget for the file. Need to serialize per + lseg = pnfs_find_lseg(lo, &arg); + if (lseg) { + trace_pnfs_update_layout(ino, pos, count, iomode, lo, lseg, + PNFS_UPDATE_LAYOUT_FOUND_CACHED); + goto out_unlock; + } + + if (!nfs4_valid_open_stateid(ctx->state)) { + trace_pnfs_update_layout(ino, pos, count, iomode, lo, lseg, + PNFS_UPDATE_LAYOUT_INVALID_OPEN); + goto out_unlock; + } + + /* + * Choose a stateid for the LAYOUTGET. If we don't have a layout + * stateid, or it has been invalidated, then we must use the open + * stateid. + */ + if (lo->plh_stateid.seqid == 0 || + test_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags)) { + + /* + * The first layoutget for the file. Need to serialize per * RFC 5661 Errata 3208. */ if (test_and_set_bit(NFS_LAYOUT_FIRST_LAYOUTGET, @@ -1573,18 +1570,17 @@ lookup_again: wait_on_bit(&lo->plh_flags, NFS_LAYOUT_FIRST_LAYOUTGET, TASK_UNINTERRUPTIBLE); pnfs_put_layout_hdr(lo); + dprintk("%s retrying\n", __func__); goto lookup_again; } + + first = true; + do { + seq = read_seqbegin(&ctx->state->seqlock); + nfs4_stateid_copy(&stateid, &ctx->state->stateid); + } while (read_seqretry(&ctx->state->seqlock, seq)); } else { - /* Check to see if the layout for the given range - * already exists - */ - lseg = pnfs_find_lseg(lo, &arg); - if (lseg) { - trace_pnfs_update_layout(ino, pos, count, iomode, lo, - PNFS_UPDATE_LAYOUT_FOUND_CACHED); - goto out_unlock; - } + nfs4_stateid_copy(&stateid, &lo->plh_stateid); } /* @@ -1599,15 +1595,17 @@ lookup_again: pnfs_clear_first_layoutget(lo); pnfs_put_layout_hdr(lo); dprintk("%s retrying\n", __func__); + trace_pnfs_update_layout(ino, pos, count, iomode, lo, + lseg, PNFS_UPDATE_LAYOUT_RETRY); goto lookup_again; } - trace_pnfs_update_layout(ino, pos, count, iomode, lo, + trace_pnfs_update_layout(ino, pos, count, iomode, lo, lseg, PNFS_UPDATE_LAYOUT_RETURN); goto out_put_layout_hdr; } if (pnfs_layoutgets_blocked(lo)) { - trace_pnfs_update_layout(ino, pos, count, iomode, lo, + trace_pnfs_update_layout(ino, pos, count, iomode, lo, lseg, PNFS_UPDATE_LAYOUT_BLOCKED); goto out_unlock; } @@ -1632,26 +1630,36 @@ lookup_again: if (arg.length != NFS4_MAX_UINT64) arg.length = PAGE_ALIGN(arg.length); - lseg = send_layoutget(lo, ctx, &arg, gfp_flags); + lseg = send_layoutget(lo, ctx, &stateid, &arg, &timeout, gfp_flags); + trace_pnfs_update_layout(ino, pos, count, iomode, lo, lseg, + PNFS_UPDATE_LAYOUT_SEND_LAYOUTGET); if (IS_ERR(lseg)) { - if (lseg == ERR_PTR(-EAGAIN)) { + switch(PTR_ERR(lseg)) { + case -ERECALLCONFLICT: + if (time_after(jiffies, giveup)) + lseg = NULL; + /* Fallthrough */ + case -EAGAIN: + pnfs_put_layout_hdr(lo); if (first) pnfs_clear_first_layoutget(lo); - pnfs_put_layout_hdr(lo); - goto lookup_again; - } - - if (!nfs_error_is_fatal(PTR_ERR(lseg))) { - pnfs_layout_clear_fail_bit(lo, pnfs_iomode_to_fail_bit(iomode)); - lseg = NULL; + if (lseg) { + trace_pnfs_update_layout(ino, pos, count, + iomode, lo, lseg, PNFS_UPDATE_LAYOUT_RETRY); + goto lookup_again; + } + /* Fallthrough */ + default: + if (!nfs_error_is_fatal(PTR_ERR(lseg))) { + pnfs_layout_clear_fail_bit(lo, pnfs_iomode_to_fail_bit(iomode)); + lseg = NULL; + } } } else { pnfs_layout_clear_fail_bit(lo, pnfs_iomode_to_fail_bit(iomode)); } atomic_dec(&lo->plh_outstanding); - trace_pnfs_update_layout(ino, pos, count, iomode, lo, - PNFS_UPDATE_LAYOUT_SEND_LAYOUTGET); out_put_layout_hdr: if (first) pnfs_clear_first_layoutget(lo); diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h index 971068b58647..f9f3331bef49 100644 --- a/fs/nfs/pnfs.h +++ b/fs/nfs/pnfs.h @@ -228,7 +228,7 @@ extern void pnfs_unregister_layoutdriver(struct pnfs_layoutdriver_type *); extern int nfs4_proc_getdeviceinfo(struct nfs_server *server, struct pnfs_device *dev, struct rpc_cred *cred); -extern struct pnfs_layout_segment* nfs4_proc_layoutget(struct nfs4_layoutget *lgp, gfp_t gfp_flags); +extern struct pnfs_layout_segment* nfs4_proc_layoutget(struct nfs4_layoutget *lgp, long *timeout, gfp_t gfp_flags); extern int nfs4_proc_layoutreturn(struct nfs4_layoutreturn *lrp, bool sync); /* pnfs.c */ @@ -260,10 +260,6 @@ void pnfs_put_layout_hdr(struct pnfs_layout_hdr *lo); void pnfs_set_layout_stateid(struct pnfs_layout_hdr *lo, const nfs4_stateid *new, bool update_barrier); -int pnfs_choose_layoutget_stateid(nfs4_stateid *dst, - struct pnfs_layout_hdr *lo, - const struct pnfs_layout_range *range, - struct nfs4_state *open_state); int pnfs_mark_matching_lsegs_invalid(struct pnfs_layout_hdr *lo, struct list_head *tmp_list, const struct pnfs_layout_range *recall_range, diff --git a/include/linux/errno.h b/include/linux/errno.h index 89627b9187f9..7ce9fb1b7d28 100644 --- a/include/linux/errno.h +++ b/include/linux/errno.h @@ -28,5 +28,6 @@ #define EBADTYPE 527 /* Type not supported by server */ #define EJUKEBOX 528 /* Request initiated, but will not complete before timeout */ #define EIOCBQUEUED 529 /* iocb queued, will get completion event */ +#define ERECALLCONFLICT 530 /* conflict with recalled state */ #endif diff --git a/include/linux/nfs4.h b/include/linux/nfs4.h index e1692c96cbc8..bfed6b367350 100644 --- a/include/linux/nfs4.h +++ b/include/linux/nfs4.h @@ -637,7 +637,9 @@ enum pnfs_update_layout_reason { PNFS_UPDATE_LAYOUT_IO_TEST_FAIL, PNFS_UPDATE_LAYOUT_FOUND_CACHED, PNFS_UPDATE_LAYOUT_RETURN, + PNFS_UPDATE_LAYOUT_RETRY, PNFS_UPDATE_LAYOUT_BLOCKED, + PNFS_UPDATE_LAYOUT_INVALID_OPEN, PNFS_UPDATE_LAYOUT_SEND_LAYOUTGET, }; diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h index e70ed54dad94..ccb2928a0e64 100644 --- a/include/linux/nfs_xdr.h +++ b/include/linux/nfs_xdr.h @@ -233,7 +233,6 @@ struct nfs4_layoutget_args { struct inode *inode; struct nfs_open_context *ctx; nfs4_stateid stateid; - unsigned long timestamp; struct nfs4_layoutdriver_data layout; }; @@ -251,7 +250,6 @@ struct nfs4_layoutget { struct nfs4_layoutget_res res; struct rpc_cred *cred; gfp_t gfp_flags; - long timeout; }; struct nfs4_getdeviceinfo_args { -- cgit v1.3-8-gc7d7 From 1b3c6d07e29515064aca8a9f86efaea7da4da027 Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Tue, 17 May 2016 12:28:48 -0400 Subject: pnfs: make pnfs_layout_process more robust It can return NULL if layoutgets are blocked currently. Fix it to return -EAGAIN in that case, so we can properly handle it in pnfs_update_layout. Also, clean up and simplify the error handling -- eliminate "status" and just use "lseg". Signed-off-by: Jeff Layton Signed-off-by: Anna Schumaker --- fs/nfs/pnfs.c | 27 +++++++++++---------------- 1 file changed, 11 insertions(+), 16 deletions(-) (limited to 'fs') diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index 46339a7fb191..79ae3049608d 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -1708,21 +1708,19 @@ pnfs_layout_process(struct nfs4_layoutget *lgp) struct pnfs_layout_segment *lseg; struct inode *ino = lo->plh_inode; LIST_HEAD(free_me); - int status = -EINVAL; if (!pnfs_sanity_check_layout_range(&res->range)) - goto out; + return ERR_PTR(-EINVAL); /* Inject layout blob into I/O device driver */ lseg = NFS_SERVER(ino)->pnfs_curr_ld->alloc_lseg(lo, res, lgp->gfp_flags); - if (!lseg || IS_ERR(lseg)) { + if (IS_ERR_OR_NULL(lseg)) { if (!lseg) - status = -ENOMEM; - else - status = PTR_ERR(lseg); - dprintk("%s: Could not allocate layout: error %d\n", - __func__, status); - goto out; + lseg = ERR_PTR(-ENOMEM); + + dprintk("%s: Could not allocate layout: error %ld\n", + __func__, PTR_ERR(lseg)); + return lseg; } init_lseg(lo, lseg); @@ -1732,15 +1730,14 @@ pnfs_layout_process(struct nfs4_layoutget *lgp) spin_lock(&ino->i_lock); if (pnfs_layoutgets_blocked(lo)) { dprintk("%s forget reply due to state\n", __func__); - goto out_forget_reply; + goto out_forget; } if (nfs4_stateid_match_other(&lo->plh_stateid, &res->stateid)) { /* existing state ID, make sure the sequence number matches. */ if (pnfs_layout_stateid_blocked(lo, &res->stateid)) { dprintk("%s forget reply due to sequence\n", __func__); - status = -EAGAIN; - goto out_forget_reply; + goto out_forget; } pnfs_set_layout_stateid(lo, &res->stateid, false); } else { @@ -1766,14 +1763,12 @@ pnfs_layout_process(struct nfs4_layoutget *lgp) spin_unlock(&ino->i_lock); pnfs_free_lseg_list(&free_me); return lseg; -out: - return ERR_PTR(status); -out_forget_reply: +out_forget: spin_unlock(&ino->i_lock); lseg->pls_layout = lo; NFS_SERVER(ino)->pnfs_curr_ld->free_lseg(lseg); - goto out; + return ERR_PTR(-EAGAIN); } static void -- cgit v1.3-8-gc7d7 From 2b88fc21cae91e925bc5e361e6ea67467b471da5 Mon Sep 17 00:00:00 2001 From: Andreas Gruenbacher Date: Fri, 22 Apr 2016 19:14:00 +0200 Subject: ubifs: Switch to generic xattr handlers Ubifs internally uses special inodes for storing xattrs. Those inodes had NULL {get,set,remove}xattr inode operations before this change, so xattr operations on them would fail. The super block's s_xattr field would also apply to those special inodes. However, the inodes are not visible outside of ubifs, and so no xattr operations will ever be carried out on them anyway. Signed-off-by: Andreas Gruenbacher Reviewed-by: Richard Weinberger Signed-off-by: Al Viro --- fs/ubifs/dir.c | 6 +-- fs/ubifs/file.c | 12 ++--- fs/ubifs/super.c | 1 + fs/ubifs/ubifs.h | 7 +-- fs/ubifs/xattr.c | 143 ++++++++++++++++++++++++++----------------------------- 5 files changed, 79 insertions(+), 90 deletions(-) (limited to 'fs') diff --git a/fs/ubifs/dir.c b/fs/ubifs/dir.c index 795992a8321e..4d07d1564780 100644 --- a/fs/ubifs/dir.c +++ b/fs/ubifs/dir.c @@ -1182,10 +1182,10 @@ const struct inode_operations ubifs_dir_inode_operations = { .rename = ubifs_rename, .setattr = ubifs_setattr, .getattr = ubifs_getattr, - .setxattr = ubifs_setxattr, - .getxattr = ubifs_getxattr, + .setxattr = generic_setxattr, + .getxattr = generic_getxattr, .listxattr = ubifs_listxattr, - .removexattr = ubifs_removexattr, + .removexattr = generic_removexattr, #ifdef CONFIG_UBIFS_ATIME_SUPPORT .update_time = ubifs_update_time, #endif diff --git a/fs/ubifs/file.c b/fs/ubifs/file.c index 065c88f8e4b8..e14943736ebf 100644 --- a/fs/ubifs/file.c +++ b/fs/ubifs/file.c @@ -1597,10 +1597,10 @@ const struct address_space_operations ubifs_file_address_operations = { const struct inode_operations ubifs_file_inode_operations = { .setattr = ubifs_setattr, .getattr = ubifs_getattr, - .setxattr = ubifs_setxattr, - .getxattr = ubifs_getxattr, + .setxattr = generic_setxattr, + .getxattr = generic_getxattr, .listxattr = ubifs_listxattr, - .removexattr = ubifs_removexattr, + .removexattr = generic_removexattr, #ifdef CONFIG_UBIFS_ATIME_SUPPORT .update_time = ubifs_update_time, #endif @@ -1611,10 +1611,10 @@ const struct inode_operations ubifs_symlink_inode_operations = { .get_link = simple_get_link, .setattr = ubifs_setattr, .getattr = ubifs_getattr, - .setxattr = ubifs_setxattr, - .getxattr = ubifs_getxattr, + .setxattr = generic_setxattr, + .getxattr = generic_getxattr, .listxattr = ubifs_listxattr, - .removexattr = ubifs_removexattr, + .removexattr = generic_removexattr, #ifdef CONFIG_UBIFS_ATIME_SUPPORT .update_time = ubifs_update_time, #endif diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c index a233ba913be4..f72f6423f082 100644 --- a/fs/ubifs/super.c +++ b/fs/ubifs/super.c @@ -2040,6 +2040,7 @@ static int ubifs_fill_super(struct super_block *sb, void *data, int silent) if (c->max_inode_sz > MAX_LFS_FILESIZE) sb->s_maxbytes = c->max_inode_sz = MAX_LFS_FILESIZE; sb->s_op = &ubifs_super_operations; + sb->s_xattr = ubifs_xattr_handlers; mutex_lock(&c->umount_mutex); err = mount_ubifs(c); diff --git a/fs/ubifs/ubifs.h b/fs/ubifs/ubifs.h index 536fb495f2f1..05d9fe70cab4 100644 --- a/fs/ubifs/ubifs.h +++ b/fs/ubifs/ubifs.h @@ -37,6 +37,7 @@ #include #include #include +#include #include "ubifs-media.h" /* Version of this UBIFS implementation */ @@ -1732,12 +1733,8 @@ int ubifs_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat); /* xattr.c */ -int ubifs_setxattr(struct dentry *dentry, const char *name, - const void *value, size_t size, int flags); -ssize_t ubifs_getxattr(struct dentry *dentry, struct inode *host, - const char *name, void *buf, size_t size); +extern const struct xattr_handler *ubifs_xattr_handlers[]; ssize_t ubifs_listxattr(struct dentry *dentry, char *buffer, size_t size); -int ubifs_removexattr(struct dentry *dentry, const char *name); int ubifs_init_security(struct inode *dentry, struct inode *inode, const struct qstr *qstr); diff --git a/fs/ubifs/xattr.c b/fs/ubifs/xattr.c index 413d650c9476..6c277eb6aef9 100644 --- a/fs/ubifs/xattr.c +++ b/fs/ubifs/xattr.c @@ -249,42 +249,6 @@ out_free: return err; } -/** - * check_namespace - check extended attribute name-space. - * @nm: extended attribute name - * - * This function makes sure the extended attribute name belongs to one of the - * supported extended attribute name-spaces. Returns name-space index in case - * of success and a negative error code in case of failure. - */ -static int check_namespace(const struct qstr *nm) -{ - int type; - - if (nm->len > UBIFS_MAX_NLEN) - return -ENAMETOOLONG; - - if (!strncmp(nm->name, XATTR_TRUSTED_PREFIX, - XATTR_TRUSTED_PREFIX_LEN)) { - if (nm->name[XATTR_TRUSTED_PREFIX_LEN] == '\0') - return -EINVAL; - type = TRUSTED_XATTR; - } else if (!strncmp(nm->name, XATTR_USER_PREFIX, - XATTR_USER_PREFIX_LEN)) { - if (nm->name[XATTR_USER_PREFIX_LEN] == '\0') - return -EINVAL; - type = USER_XATTR; - } else if (!strncmp(nm->name, XATTR_SECURITY_PREFIX, - XATTR_SECURITY_PREFIX_LEN)) { - if (nm->name[XATTR_SECURITY_PREFIX_LEN] == '\0') - return -EINVAL; - type = SECURITY_XATTR; - } else - return -EOPNOTSUPP; - - return type; -} - static struct inode *iget_xattr(struct ubifs_info *c, ino_t inum) { struct inode *inode; @@ -302,24 +266,23 @@ static struct inode *iget_xattr(struct ubifs_info *c, ino_t inum) return ERR_PTR(-EINVAL); } -static int setxattr(struct inode *host, const char *name, const void *value, - size_t size, int flags) +static int __ubifs_setxattr(struct inode *host, const char *name, + const void *value, size_t size, int flags) { struct inode *inode; struct ubifs_info *c = host->i_sb->s_fs_info; struct qstr nm = QSTR_INIT(name, strlen(name)); struct ubifs_dent_node *xent; union ubifs_key key; - int err, type; + int err; ubifs_assert(inode_is_locked(host)); if (size > UBIFS_MAX_INO_DATA) return -ERANGE; - type = check_namespace(&nm); - if (type < 0) - return type; + if (nm.len > UBIFS_MAX_NLEN) + return -ENAMETOOLONG; xent = kmalloc(UBIFS_MAX_XENT_NODE_SZ, GFP_NOFS); if (!xent) @@ -363,17 +326,8 @@ out_free: return err; } -int ubifs_setxattr(struct dentry *dentry, const char *name, - const void *value, size_t size, int flags) -{ - dbg_gen("xattr '%s', host ino %lu ('%pd'), size %zd", - name, d_inode(dentry)->i_ino, dentry, size); - - return setxattr(d_inode(dentry), name, value, size, flags); -} - -ssize_t ubifs_getxattr(struct dentry *dentry, struct inode *host, - const char *name, void *buf, size_t size) +static ssize_t __ubifs_getxattr(struct inode *host, const char *name, + void *buf, size_t size) { struct inode *inode; struct ubifs_info *c = host->i_sb->s_fs_info; @@ -383,12 +337,8 @@ ssize_t ubifs_getxattr(struct dentry *dentry, struct inode *host, union ubifs_key key; int err; - dbg_gen("xattr '%s', ino %lu ('%pd'), buf size %zd", name, - host->i_ino, dentry, size); - - err = check_namespace(&nm); - if (err < 0) - return err; + if (nm.len > UBIFS_MAX_NLEN) + return -ENAMETOOLONG; xent = kmalloc(UBIFS_MAX_XENT_NODE_SZ, GFP_NOFS); if (!xent) @@ -460,8 +410,6 @@ ssize_t ubifs_listxattr(struct dentry *dentry, char *buffer, size_t size) lowest_xent_key(c, &key, host->i_ino); while (1) { - int type; - xent = ubifs_tnc_next_ent(c, &key, &nm); if (IS_ERR(xent)) { err = PTR_ERR(xent); @@ -471,14 +419,10 @@ ssize_t ubifs_listxattr(struct dentry *dentry, char *buffer, size_t size) nm.name = xent->name; nm.len = le16_to_cpu(xent->nlen); - type = check_namespace(&nm); - if (unlikely(type < 0)) { - err = type; - break; - } - /* Show trusted namespace only for "power" users */ - if (type != TRUSTED_XATTR || capable(CAP_SYS_ADMIN)) { + if (strncmp(xent->name, XATTR_TRUSTED_PREFIX, + XATTR_TRUSTED_PREFIX_LEN) || + capable(CAP_SYS_ADMIN)) { memcpy(buffer + written, nm.name, nm.len + 1); written += nm.len + 1; } @@ -538,22 +482,19 @@ out_cancel: return err; } -int ubifs_removexattr(struct dentry *dentry, const char *name) +static int __ubifs_removexattr(struct inode *host, const char *name) { - struct inode *inode, *host = d_inode(dentry); + struct inode *inode; struct ubifs_info *c = host->i_sb->s_fs_info; struct qstr nm = QSTR_INIT(name, strlen(name)); struct ubifs_dent_node *xent; union ubifs_key key; int err; - dbg_gen("xattr '%s', ino %lu ('%pd')", name, - host->i_ino, dentry); ubifs_assert(inode_is_locked(host)); - err = check_namespace(&nm); - if (err < 0) - return err; + if (nm.len > UBIFS_MAX_NLEN) + return -ENAMETOOLONG; xent = kmalloc(UBIFS_MAX_XENT_NODE_SZ, GFP_NOFS); if (!xent) @@ -603,7 +544,7 @@ static int init_xattrs(struct inode *inode, const struct xattr *xattr_array, } strcpy(name, XATTR_SECURITY_PREFIX); strcpy(name + XATTR_SECURITY_PREFIX_LEN, xattr->name); - err = setxattr(inode, name, xattr->value, xattr->value_len, 0); + err = __ubifs_setxattr(inode, name, xattr->value, xattr->value_len, 0); kfree(name); if (err < 0) break; @@ -626,3 +567,53 @@ int ubifs_init_security(struct inode *dentry, struct inode *inode, } return err; } + +static int ubifs_xattr_get(const struct xattr_handler *handler, + struct dentry *dentry, struct inode *inode, + const char *name, void *buffer, size_t size) +{ + dbg_gen("xattr '%s', ino %lu ('%pd'), buf size %zd", name, + inode->i_ino, dentry, size); + + return __ubifs_getxattr(inode, name, buffer, size); +} + +static int ubifs_xattr_set(const struct xattr_handler *handler, + struct dentry *dentry, const char *name, + const void *value, size_t size, int flags) +{ + struct inode *inode = d_inode(dentry); + + dbg_gen("xattr '%s', host ino %lu ('%pd'), size %zd", + name, inode->i_ino, dentry, size); + + if (value) + return __ubifs_setxattr(inode, name, value, size, flags); + else + return __ubifs_removexattr(inode, name); +} + +const struct xattr_handler ubifs_user_xattr_handler = { + .prefix = XATTR_USER_PREFIX, + .get = ubifs_xattr_get, + .set = ubifs_xattr_set, +}; + +const struct xattr_handler ubifs_trusted_xattr_handler = { + .prefix = XATTR_TRUSTED_PREFIX, + .get = ubifs_xattr_get, + .set = ubifs_xattr_set, +}; + +const struct xattr_handler ubifs_security_xattr_handler = { + .prefix = XATTR_SECURITY_PREFIX, + .get = ubifs_xattr_get, + .set = ubifs_xattr_set, +}; + +const struct xattr_handler *ubifs_xattr_handlers[] = { + &ubifs_user_xattr_handler, + &ubifs_trusted_xattr_handler, + &ubifs_security_xattr_handler, + NULL +}; -- cgit v1.3-8-gc7d7 From e0d46f5c6e0ba3a79e64cd60e62b7b7191ed93f3 Mon Sep 17 00:00:00 2001 From: Andreas Gruenbacher Date: Fri, 22 Apr 2016 22:36:44 +0200 Subject: btrfs: Switch to generic xattr handlers The btrfs_{set,remove}xattr inode operations check for a read-only root (btrfs_root_readonly) before calling into generic_{set,remove}xattr. If this check is moved into __btrfs_setxattr, we can get rid of btrfs_{set,remove}xattr. This patch applies to mainline, I would like to keep it together with the other xattr cleanups if possible, though. Could you please review? Thanks, Andreas Signed-off-by: Andreas Gruenbacher Signed-off-by: Al Viro --- fs/btrfs/inode.c | 16 ++++++++-------- fs/btrfs/xattr.c | 22 +++------------------- fs/btrfs/xattr.h | 3 --- 3 files changed, 11 insertions(+), 30 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 41a5688ffdfe..0077b3b49e5e 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -10160,10 +10160,10 @@ static const struct inode_operations btrfs_dir_inode_operations = { .symlink = btrfs_symlink, .setattr = btrfs_setattr, .mknod = btrfs_mknod, - .setxattr = btrfs_setxattr, + .setxattr = generic_setxattr, .getxattr = generic_getxattr, .listxattr = btrfs_listxattr, - .removexattr = btrfs_removexattr, + .removexattr = generic_removexattr, .permission = btrfs_permission, .get_acl = btrfs_get_acl, .set_acl = btrfs_set_acl, @@ -10237,10 +10237,10 @@ static const struct address_space_operations btrfs_symlink_aops = { static const struct inode_operations btrfs_file_inode_operations = { .getattr = btrfs_getattr, .setattr = btrfs_setattr, - .setxattr = btrfs_setxattr, + .setxattr = generic_setxattr, .getxattr = generic_getxattr, .listxattr = btrfs_listxattr, - .removexattr = btrfs_removexattr, + .removexattr = generic_removexattr, .permission = btrfs_permission, .fiemap = btrfs_fiemap, .get_acl = btrfs_get_acl, @@ -10251,10 +10251,10 @@ static const struct inode_operations btrfs_special_inode_operations = { .getattr = btrfs_getattr, .setattr = btrfs_setattr, .permission = btrfs_permission, - .setxattr = btrfs_setxattr, + .setxattr = generic_setxattr, .getxattr = generic_getxattr, .listxattr = btrfs_listxattr, - .removexattr = btrfs_removexattr, + .removexattr = generic_removexattr, .get_acl = btrfs_get_acl, .set_acl = btrfs_set_acl, .update_time = btrfs_update_time, @@ -10265,10 +10265,10 @@ static const struct inode_operations btrfs_symlink_inode_operations = { .getattr = btrfs_getattr, .setattr = btrfs_setattr, .permission = btrfs_permission, - .setxattr = btrfs_setxattr, + .setxattr = generic_setxattr, .getxattr = generic_getxattr, .listxattr = btrfs_listxattr, - .removexattr = btrfs_removexattr, + .removexattr = generic_removexattr, .update_time = btrfs_update_time, }; diff --git a/fs/btrfs/xattr.c b/fs/btrfs/xattr.c index 03224b00ea70..3bfb252206c7 100644 --- a/fs/btrfs/xattr.c +++ b/fs/btrfs/xattr.c @@ -237,6 +237,9 @@ int __btrfs_setxattr(struct btrfs_trans_handle *trans, struct btrfs_root *root = BTRFS_I(inode)->root; int ret; + if (btrfs_root_readonly(root)) + return -EROFS; + if (trans) return do_setxattr(trans, inode, name, value, size, flags); @@ -432,25 +435,6 @@ const struct xattr_handler *btrfs_xattr_handlers[] = { NULL, }; -int btrfs_setxattr(struct dentry *dentry, const char *name, const void *value, - size_t size, int flags) -{ - struct btrfs_root *root = BTRFS_I(d_inode(dentry))->root; - - if (btrfs_root_readonly(root)) - return -EROFS; - return generic_setxattr(dentry, name, value, size, flags); -} - -int btrfs_removexattr(struct dentry *dentry, const char *name) -{ - struct btrfs_root *root = BTRFS_I(d_inode(dentry))->root; - - if (btrfs_root_readonly(root)) - return -EROFS; - return generic_removexattr(dentry, name); -} - static int btrfs_initxattrs(struct inode *inode, const struct xattr *xattr_array, void *fs_info) { diff --git a/fs/btrfs/xattr.h b/fs/btrfs/xattr.h index 96807b3d22f5..15fc4743dc70 100644 --- a/fs/btrfs/xattr.h +++ b/fs/btrfs/xattr.h @@ -28,9 +28,6 @@ extern ssize_t __btrfs_getxattr(struct inode *inode, const char *name, extern int __btrfs_setxattr(struct btrfs_trans_handle *trans, struct inode *inode, const char *name, const void *value, size_t size, int flags); -extern int btrfs_setxattr(struct dentry *dentry, const char *name, - const void *value, size_t size, int flags); -extern int btrfs_removexattr(struct dentry *dentry, const char *name); extern int btrfs_xattr_security_init(struct btrfs_trans_handle *trans, struct inode *inode, struct inode *dir, -- cgit v1.3-8-gc7d7 From 9bdd9bd69b826875531bb1b2efb6aeb8d70e6f72 Mon Sep 17 00:00:00 2001 From: Brian Foster Date: Wed, 18 May 2016 10:56:41 +1000 Subject: xfs: buffer ->bi_end_io function requires irq-safe lock Reports have surfaced of a lockdep splat complaining about an irq-safe -> irq-unsafe locking order in the xfs_buf_bio_end_io() bio completion handler. This only occurs when I/O errors are present because bp->b_lock is only acquired in this context to protect setting an error on the buffer. The problem is that this lock can be acquired with the (request_queue) q->queue_lock held. See scsi_end_request() or ata_qc_schedule_eh(), for example. Replace the locked test/set of b_io_error with a cmpxchg() call. This eliminates the need for the lock and thus the lock ordering problem goes away. Signed-off-by: Brian Foster Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner --- fs/xfs/xfs_buf.c | 12 ++++-------- 1 file changed, 4 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c index 9a2191b91137..e71cfbd5acb3 100644 --- a/fs/xfs/xfs_buf.c +++ b/fs/xfs/xfs_buf.c @@ -1100,22 +1100,18 @@ xfs_bwrite( return error; } -STATIC void +static void xfs_buf_bio_end_io( struct bio *bio) { - xfs_buf_t *bp = (xfs_buf_t *)bio->bi_private; + struct xfs_buf *bp = (struct xfs_buf *)bio->bi_private; /* * don't overwrite existing errors - otherwise we can lose errors on * buffers that require multiple bios to complete. */ - if (bio->bi_error) { - spin_lock(&bp->b_lock); - if (!bp->b_io_error) - bp->b_io_error = bio->bi_error; - spin_unlock(&bp->b_lock); - } + if (bio->bi_error) + cmpxchg(&bp->b_io_error, 0, bio->bi_error); if (!bp->b_error && xfs_buf_is_vmapped(bp) && (bp->b_flags & XBF_READ)) invalidate_kernel_vmap_range(bp->b_addr, xfs_buf_vmap_len(bp)); -- cgit v1.3-8-gc7d7 From 192852be8b5fb14268c2133fe9ce5312e4745963 Mon Sep 17 00:00:00 2001 From: Carlos Maiolino Date: Wed, 18 May 2016 10:58:51 +1000 Subject: xfs: configurable error behavior via sysfs We need to be able to change the way XFS behaviours in error conditions depending on the type of underlying storage. This is necessary for handling non-traditional block devices with extended error cases, such as thin provisioned devices that can return ENOSPC as an IO error. Introduce the basic sysfs infrastructure needed to define and configure error behaviours. This is done to be generic enough to extend to configuring behaviour in other error conditions, such as ENOMEM, which also has different desired behaviours according to machine configuration. Signed-off-by: Dave Chinner Signed-off-by: Carlos Maiolino Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/xfs_mount.c | 10 +++++++++- fs/xfs/xfs_mount.h | 20 ++++++++++++++++++++ fs/xfs/xfs_sysfs.c | 53 ++++++++++++++++++++++++++++++++++++++++++++++++++++- fs/xfs/xfs_sysfs.h | 3 +++ 4 files changed, 84 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index 536a0ee9cd5a..677c3e0da472 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -690,10 +690,15 @@ xfs_mountfs( if (error) goto out_remove_sysfs; - error = xfs_uuid_mount(mp); + error = xfs_error_sysfs_init(mp); if (error) goto out_del_stats; + + error = xfs_uuid_mount(mp); + if (error) + goto out_remove_error_sysfs; + /* * Set the minimum read and write sizes */ @@ -968,6 +973,8 @@ xfs_mountfs( xfs_da_unmount(mp); out_remove_uuid: xfs_uuid_unmount(mp); + out_remove_error_sysfs: + xfs_error_sysfs_del(mp); out_del_stats: xfs_sysfs_del(&mp->m_stats.xs_kobj); out_remove_sysfs: @@ -1056,6 +1063,7 @@ xfs_unmountfs( #endif xfs_free_perag(mp); + xfs_error_sysfs_del(mp); xfs_sysfs_del(&mp->m_stats.xs_kobj); xfs_sysfs_del(&mp->m_kobj); } diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h index bac6b3435591..d639795b0310 100644 --- a/fs/xfs/xfs_mount.h +++ b/fs/xfs/xfs_mount.h @@ -37,6 +37,24 @@ enum { XFS_LOWSP_MAX, }; +/* + * Error Configuration + * + * Error classes define the subsystem the configuration belongs to. + * Error numbers define the errors that are configurable. + */ +enum { + XFS_ERR_CLASS_MAX, +}; +enum { + XFS_ERR_ERRNO_MAX, +}; + +struct xfs_error_cfg { + struct xfs_kobj kobj; + int max_retries; +}; + typedef struct xfs_mount { struct super_block *m_super; xfs_tid_t m_tid; /* next unused tid for fs */ @@ -127,6 +145,8 @@ typedef struct xfs_mount { int64_t m_low_space[XFS_LOWSP_MAX]; /* low free space thresholds */ struct xfs_kobj m_kobj; + struct xfs_kobj m_error_kobj; + struct xfs_error_cfg m_error_cfg[XFS_ERR_CLASS_MAX][XFS_ERR_ERRNO_MAX]; struct xstats m_stats; /* per-fs stats */ struct workqueue_struct *m_buf_workqueue; diff --git a/fs/xfs/xfs_sysfs.c b/fs/xfs/xfs_sysfs.c index 6ced4f143494..74e394071242 100644 --- a/fs/xfs/xfs_sysfs.c +++ b/fs/xfs/xfs_sysfs.c @@ -17,10 +17,11 @@ */ #include "xfs.h" -#include "xfs_sysfs.h" +#include "xfs_shared.h" #include "xfs_format.h" #include "xfs_log_format.h" #include "xfs_trans_resv.h" +#include "xfs_sysfs.h" #include "xfs_log.h" #include "xfs_log_priv.h" #include "xfs_stats.h" @@ -362,3 +363,53 @@ struct kobj_type xfs_log_ktype = { .sysfs_ops = &xfs_sysfs_ops, .default_attrs = xfs_log_attrs, }; + +/* + * Metadata IO error configuration + * + * The sysfs structure here is: + * ...xfs//error/// + * + * where allows us to discriminate between data IO and metadata IO, + * and any other future type of IO (e.g. special inode or directory error + * handling) we care to support. + */ +static struct attribute *xfs_error_attrs[] = { + NULL, +}; + +static inline struct xfs_error_cfg * +to_error_cfg(struct kobject *kobject) +{ + struct xfs_kobj *kobj = to_kobj(kobject); + return container_of(kobj, struct xfs_error_cfg, kobj); +} + +struct kobj_type xfs_error_cfg_ktype = { + .release = xfs_sysfs_release, + .sysfs_ops = &xfs_sysfs_ops, + .default_attrs = xfs_error_attrs, +}; + +struct kobj_type xfs_error_ktype = { + .release = xfs_sysfs_release, +}; + +int +xfs_error_sysfs_init( + struct xfs_mount *mp) +{ + int error; + + /* .../xfs//error/ */ + error = xfs_sysfs_init(&mp->m_error_kobj, &xfs_error_ktype, + &mp->m_kobj, "error"); + return error; +} + +void +xfs_error_sysfs_del( + struct xfs_mount *mp) +{ + xfs_sysfs_del(&mp->m_error_kobj); +} diff --git a/fs/xfs/xfs_sysfs.h b/fs/xfs/xfs_sysfs.h index be692e59938d..d04637181ef2 100644 --- a/fs/xfs/xfs_sysfs.h +++ b/fs/xfs/xfs_sysfs.h @@ -58,4 +58,7 @@ xfs_sysfs_del( wait_for_completion(&kobj->complete); } +int xfs_error_sysfs_init(struct xfs_mount *mp); +void xfs_error_sysfs_del(struct xfs_mount *mp); + #endif /* __XFS_SYSFS_H__ */ -- cgit v1.3-8-gc7d7 From ffd40ef697dfd3e06f44b1bb5fea93079de8c77d Mon Sep 17 00:00:00 2001 From: Carlos Maiolino Date: Wed, 18 May 2016 11:01:00 +1000 Subject: xfs: introduce metadata IO error class Now we have the basic infrastructure, add the first error class so we can build up the infrastructure in a meaningful way. Add the metadata async write IO error class and sysfs entry, and introduce a default configuration that matches the existing "retry forever" behavior for async write metadata buffers. Signed-off-by: Dave Chinner Signed-off-by: Carlos Maiolino Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/xfs_mount.h | 3 +++ fs/xfs/xfs_sysfs.c | 34 ++++++++++++++++++++++++++++++++++ 2 files changed, 37 insertions(+) (limited to 'fs') diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h index d639795b0310..352a5c88d7e9 100644 --- a/fs/xfs/xfs_mount.h +++ b/fs/xfs/xfs_mount.h @@ -44,9 +44,11 @@ enum { * Error numbers define the errors that are configurable. */ enum { + XFS_ERR_METADATA, XFS_ERR_CLASS_MAX, }; enum { + XFS_ERR_DEFAULT, XFS_ERR_ERRNO_MAX, }; @@ -146,6 +148,7 @@ typedef struct xfs_mount { /* low free space thresholds */ struct xfs_kobj m_kobj; struct xfs_kobj m_error_kobj; + struct xfs_kobj m_error_meta_kobj; struct xfs_error_cfg m_error_cfg[XFS_ERR_CLASS_MAX][XFS_ERR_ERRNO_MAX]; struct xstats m_stats; /* per-fs stats */ diff --git a/fs/xfs/xfs_sysfs.c b/fs/xfs/xfs_sysfs.c index 74e394071242..07c95999541e 100644 --- a/fs/xfs/xfs_sysfs.c +++ b/fs/xfs/xfs_sysfs.c @@ -399,11 +399,34 @@ int xfs_error_sysfs_init( struct xfs_mount *mp) { + struct xfs_error_cfg *cfg; int error; /* .../xfs//error/ */ error = xfs_sysfs_init(&mp->m_error_kobj, &xfs_error_ktype, &mp->m_kobj, "error"); + if (error) + return error; + + /* .../xfs//error/metadata/ */ + error = xfs_sysfs_init(&mp->m_error_meta_kobj, &xfs_error_ktype, + &mp->m_error_kobj, "metadata"); + if (error) + goto out_error; + + cfg = &mp->m_error_cfg[XFS_ERR_METADATA][XFS_ERR_DEFAULT]; + error = xfs_sysfs_init(&cfg->kobj, &xfs_error_cfg_ktype, + &mp->m_error_meta_kobj, "default"); + if (error) + goto out_error_meta; + cfg->max_retries = -1; + + return 0; + +out_error_meta: + xfs_sysfs_del(&mp->m_error_meta_kobj); +out_error: + xfs_sysfs_del(&mp->m_error_kobj); return error; } @@ -411,5 +434,16 @@ void xfs_error_sysfs_del( struct xfs_mount *mp) { + struct xfs_error_cfg *cfg; + int i, j; + + for (i = 0; i < XFS_ERR_CLASS_MAX; i++) { + for (j = 0; j < XFS_ERR_ERRNO_MAX; j++) { + cfg = &mp->m_error_cfg[i][j]; + + xfs_sysfs_del(&cfg->kobj); + } + } + xfs_sysfs_del(&mp->m_error_meta_kobj); xfs_sysfs_del(&mp->m_error_kobj); } -- cgit v1.3-8-gc7d7 From df3093907ccc718459c54c99da29dd774af41186 Mon Sep 17 00:00:00 2001 From: Carlos Maiolino Date: Wed, 18 May 2016 11:05:33 +1000 Subject: xfs: add configurable error support to metadata buffers With the error configuration handle for async metadata write errors in place, we can now add initial support to the IO error processing in xfs_buf_iodone_error(). Add an infrastructure function to look up the configuration handle, and rearrange the error handling to prepare the way for different error handling conigurations to be used. Signed-off-by: Dave Chinner Signed-off-by: Carlos Maiolino Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/xfs_buf.h | 1 + fs/xfs/xfs_buf_item.c | 112 ++++++++++++++++++++++++++++++-------------------- fs/xfs/xfs_mount.h | 3 ++ fs/xfs/xfs_sysfs.c | 17 ++++++++ fs/xfs/xfs_trace.h | 1 - 5 files changed, 88 insertions(+), 46 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_buf.h b/fs/xfs/xfs_buf.h index 4eb89bd4ee73..adef116db0c3 100644 --- a/fs/xfs/xfs_buf.h +++ b/fs/xfs/xfs_buf.h @@ -183,6 +183,7 @@ typedef struct xfs_buf { unsigned int b_page_count; /* size of page array */ unsigned int b_offset; /* page offset in first page */ int b_error; /* error code on I/O */ + int b_last_error; /* previous async I/O error */ const struct xfs_buf_ops *b_ops; #ifdef XFS_BUF_LOCK_TRACKING diff --git a/fs/xfs/xfs_buf_item.c b/fs/xfs/xfs_buf_item.c index 99e91a0e554e..b8d0cd4adb81 100644 --- a/fs/xfs/xfs_buf_item.c +++ b/fs/xfs/xfs_buf_item.c @@ -1042,35 +1042,22 @@ xfs_buf_do_callbacks( } } -/* - * This is the iodone() function for buffers which have had callbacks - * attached to them by xfs_buf_attach_iodone(). It should remove each - * log item from the buffer's list and call the callback of each in turn. - * When done, the buffer's fsprivate field is set to NULL and the buffer - * is unlocked with a call to iodone(). - */ -void -xfs_buf_iodone_callbacks( +static bool +xfs_buf_iodone_callback_error( struct xfs_buf *bp) { struct xfs_log_item *lip = bp->b_fspriv; struct xfs_mount *mp = lip->li_mountp; static ulong lasttime; static xfs_buftarg_t *lasttarg; - - if (likely(!bp->b_error)) - goto do_callbacks; + struct xfs_error_cfg *cfg; /* * If we've already decided to shutdown the filesystem because of * I/O errors, there's no point in giving this a retry. */ - if (XFS_FORCED_SHUTDOWN(mp)) { - xfs_buf_stale(bp); - bp->b_flags |= XBF_DONE; - trace_xfs_buf_item_iodone(bp, _RET_IP_); - goto do_callbacks; - } + if (XFS_FORCED_SHUTDOWN(mp)) + goto out_stale; if (bp->b_target != lasttarg || time_after(jiffies, (lasttime + 5*HZ))) { @@ -1079,45 +1066,80 @@ xfs_buf_iodone_callbacks( } lasttarg = bp->b_target; + /* synchronous writes will have callers process the error */ + if (!(bp->b_flags & XBF_ASYNC)) + goto out_stale; + + trace_xfs_buf_item_iodone_async(bp, _RET_IP_); + ASSERT(bp->b_iodone != NULL); + /* * If the write was asynchronous then no one will be looking for the - * error. Clear the error state and write the buffer out again. - * - * XXX: This helps against transient write errors, but we need to find - * a way to shut the filesystem down if the writes keep failing. - * - * In practice we'll shut the filesystem down soon as non-transient - * errors tend to affect the whole device and a failing log write - * will make us give up. But we really ought to do better here. + * error. If this is the first failure of this type, clear the error + * state and write the buffer out again. This means we always retry an + * async write failure at least once, but we also need to set the buffer + * up to behave correctly now for repeated failures. */ - if (bp->b_flags & XBF_ASYNC) { - ASSERT(bp->b_iodone != NULL); - - trace_xfs_buf_item_iodone_async(bp, _RET_IP_); + if (!(bp->b_flags & (XBF_STALE|XBF_WRITE_FAIL)) || + bp->b_last_error != bp->b_error) { + bp->b_flags |= (XBF_WRITE | XBF_ASYNC | + XBF_DONE | XBF_WRITE_FAIL); + bp->b_last_error = bp->b_error; + xfs_buf_ioerror(bp, 0); + xfs_buf_submit(bp); + return true; + } - xfs_buf_ioerror(bp, 0); /* errno of 0 unsets the flag */ + /* + * Repeated failure on an async write. Take action according to the + * error configuration we have been set up to use. + */ + cfg = xfs_error_get_cfg(mp, XFS_ERR_METADATA, bp->b_error); + if (!cfg->max_retries) + goto permanent_error; - if (!(bp->b_flags & (XBF_STALE|XBF_WRITE_FAIL))) { - bp->b_flags |= XBF_WRITE | XBF_ASYNC | - XBF_DONE | XBF_WRITE_FAIL; - xfs_buf_submit(bp); - } else { - xfs_buf_relse(bp); - } - - return; - } + /* still a transient error, higher layers will retry */ + xfs_buf_ioerror(bp, 0); + xfs_buf_relse(bp); + return true; /* - * If the write of the buffer was synchronous, we want to make - * sure to return the error to the caller of xfs_bwrite(). + * Permanent error - we need to trigger a shutdown if we haven't already + * to indicate that inconsistency will result from this action. */ +permanent_error: + xfs_force_shutdown(mp, SHUTDOWN_META_IO_ERROR); +out_stale: xfs_buf_stale(bp); bp->b_flags |= XBF_DONE; - trace_xfs_buf_error_relse(bp, _RET_IP_); + return false; +} + +/* + * This is the iodone() function for buffers which have had callbacks attached + * to them by xfs_buf_attach_iodone(). We need to iterate the items on the + * callback list, mark the buffer as having no more callbacks and then push the + * buffer through IO completion processing. + */ +void +xfs_buf_iodone_callbacks( + struct xfs_buf *bp) +{ + /* + * If there is an error, process it. Some errors require us + * to run callbacks after failure processing is done so we + * detect that and take appropriate action. + */ + if (bp->b_error && xfs_buf_iodone_callback_error(bp)) + return; + + /* + * Successful IO or permanent error. Either way, we can clear the + * retry state here in preparation for the next error that may occur. + */ + bp->b_last_error = 0; -do_callbacks: xfs_buf_do_callbacks(bp); bp->b_fspriv = NULL; bp->b_iodone = NULL; diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h index 352a5c88d7e9..0c5a97644d78 100644 --- a/fs/xfs/xfs_mount.h +++ b/fs/xfs/xfs_mount.h @@ -387,4 +387,7 @@ extern void xfs_set_low_space_thresholds(struct xfs_mount *); int xfs_zero_extent(struct xfs_inode *ip, xfs_fsblock_t start_fsb, xfs_off_t count_fsb); +struct xfs_error_cfg * xfs_error_get_cfg(struct xfs_mount *mp, + int error_class, int error); + #endif /* __XFS_MOUNT_H__ */ diff --git a/fs/xfs/xfs_sysfs.c b/fs/xfs/xfs_sysfs.c index 07c95999541e..1cb5a85409af 100644 --- a/fs/xfs/xfs_sysfs.c +++ b/fs/xfs/xfs_sysfs.c @@ -447,3 +447,20 @@ xfs_error_sysfs_del( xfs_sysfs_del(&mp->m_error_meta_kobj); xfs_sysfs_del(&mp->m_error_kobj); } + +struct xfs_error_cfg * +xfs_error_get_cfg( + struct xfs_mount *mp, + int error_class, + int error) +{ + struct xfs_error_cfg *cfg; + + switch (error) { + default: + cfg = &mp->m_error_cfg[error_class][XFS_ERR_DEFAULT]; + break; + } + + return cfg; +} diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h index c8d58426008e..a133dd4c43bc 100644 --- a/fs/xfs/xfs_trace.h +++ b/fs/xfs/xfs_trace.h @@ -364,7 +364,6 @@ DEFINE_BUF_EVENT(xfs_buf_delwri_split); DEFINE_BUF_EVENT(xfs_buf_get_uncached); DEFINE_BUF_EVENT(xfs_bdstrat_shut); DEFINE_BUF_EVENT(xfs_buf_item_relse); -DEFINE_BUF_EVENT(xfs_buf_item_iodone); DEFINE_BUF_EVENT(xfs_buf_item_iodone_async); DEFINE_BUF_EVENT(xfs_buf_error_relse); DEFINE_BUF_EVENT(xfs_buf_wait_buftarg); -- cgit v1.3-8-gc7d7 From ef6a50fbb1bba7951aa23adcfb40e99ca72dc51c Mon Sep 17 00:00:00 2001 From: Carlos Maiolino Date: Wed, 18 May 2016 11:06:44 +1000 Subject: xfs: introduce table-based init for error behaviors Before we start expanding the number of error classes and errors we can configure behaviour for, we need a simple and clear way to define the default behaviour that we initialized each mount with. Introduce a table based method for keeping the initial configuration in, and apply that to the existing initialization code. Signed-off-by: Dave Chinner Signed-off-by: Carlos Maiolino Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/xfs_sysfs.c | 72 +++++++++++++++++++++++++++++++++++++++++++++--------- 1 file changed, 60 insertions(+), 12 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_sysfs.c b/fs/xfs/xfs_sysfs.c index 1cb5a85409af..71046d904985 100644 --- a/fs/xfs/xfs_sysfs.c +++ b/fs/xfs/xfs_sysfs.c @@ -395,11 +395,67 @@ struct kobj_type xfs_error_ktype = { .release = xfs_sysfs_release, }; +/* + * Error initialization tables. These need to be ordered in the same + * order as the enums used to index the array. All class init tables need to + * define a "default" behaviour as the first entry, all other entries can be + * empty. + */ +struct xfs_error_init { + char *name; + int max_retries; +}; + +static const struct xfs_error_init xfs_error_meta_init[XFS_ERR_ERRNO_MAX] = { + { .name = "default", + .max_retries = -1, + }, +}; + +static int +xfs_error_sysfs_init_class( + struct xfs_mount *mp, + int class, + const char *parent_name, + struct xfs_kobj *parent_kobj, + const struct xfs_error_init init[]) +{ + struct xfs_error_cfg *cfg; + int error; + int i; + + ASSERT(class < XFS_ERR_CLASS_MAX); + + error = xfs_sysfs_init(parent_kobj, &xfs_error_ktype, + &mp->m_error_kobj, parent_name); + if (error) + return error; + + for (i = 0; i < XFS_ERR_ERRNO_MAX; i++) { + cfg = &mp->m_error_cfg[class][i]; + error = xfs_sysfs_init(&cfg->kobj, &xfs_error_cfg_ktype, + parent_kobj, init[i].name); + if (error) + goto out_error; + + cfg->max_retries = init[i].max_retries; + } + return 0; + +out_error: + /* unwind the entries that succeeded */ + for (i--; i >= 0; i--) { + cfg = &mp->m_error_cfg[class][i]; + xfs_sysfs_del(&cfg->kobj); + } + xfs_sysfs_del(parent_kobj); + return error; +} + int xfs_error_sysfs_init( struct xfs_mount *mp) { - struct xfs_error_cfg *cfg; int error; /* .../xfs//error/ */ @@ -409,22 +465,14 @@ xfs_error_sysfs_init( return error; /* .../xfs//error/metadata/ */ - error = xfs_sysfs_init(&mp->m_error_meta_kobj, &xfs_error_ktype, - &mp->m_error_kobj, "metadata"); + error = xfs_error_sysfs_init_class(mp, XFS_ERR_METADATA, + "metadata", &mp->m_error_meta_kobj, + xfs_error_meta_init); if (error) goto out_error; - cfg = &mp->m_error_cfg[XFS_ERR_METADATA][XFS_ERR_DEFAULT]; - error = xfs_sysfs_init(&cfg->kobj, &xfs_error_cfg_ktype, - &mp->m_error_meta_kobj, "default"); - if (error) - goto out_error_meta; - cfg->max_retries = -1; - return 0; -out_error_meta: - xfs_sysfs_del(&mp->m_error_meta_kobj); out_error: xfs_sysfs_del(&mp->m_error_kobj); return error; -- cgit v1.3-8-gc7d7 From a5ea70d25d76950e11690110b526374307d05d81 Mon Sep 17 00:00:00 2001 From: Carlos Maiolino Date: Wed, 18 May 2016 11:08:15 +1000 Subject: xfs: add configuration of error failure speed On reception of an error, we can fail immediately, perform some bound amount of retries or retry indefinitely. The current behaviour we have is to retry forever. However, we'd like the ability to choose how long the filesystem should try after an error, it can either fail immediately, retry a few times, or retry forever. This is implemented by using max_retries sysfs attribute, to hold the amount of times we allow the filesystem to retry after an error. Being -1 a special case where the filesystem will retry indefinitely. Add both a maximum retry count and a retry timeout so that we can bound by time and/or physical IO attempts. Finally, plumb these into xfs_buf_iodone error processing so that the error behaviour follows the selected configuration. Signed-off-by: Dave Chinner Signed-off-by: Carlos Maiolino Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/xfs_buf.h | 21 ++++++++++++- fs/xfs/xfs_buf_item.c | 13 +++++++-- fs/xfs/xfs_mount.h | 3 ++ fs/xfs/xfs_sysfs.c | 81 ++++++++++++++++++++++++++++++++++++++++++++++++--- 4 files changed, 111 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_buf.h b/fs/xfs/xfs_buf.h index adef116db0c3..8bfb974f0772 100644 --- a/fs/xfs/xfs_buf.h +++ b/fs/xfs/xfs_buf.h @@ -183,7 +183,26 @@ typedef struct xfs_buf { unsigned int b_page_count; /* size of page array */ unsigned int b_offset; /* page offset in first page */ int b_error; /* error code on I/O */ - int b_last_error; /* previous async I/O error */ + + /* + * async write failure retry count. Initialised to zero on the first + * failure, then when it exceeds the maximum configured without a + * success the write is considered to be failed permanently and the + * iodone handler will take appropriate action. + * + * For retry timeouts, we record the jiffie of the first failure. This + * means that we can change the retry timeout for buffers already under + * I/O and thus avoid getting stuck in a retry loop with a long timeout. + * + * last_error is used to ensure that we are getting repeated errors, not + * different errors. e.g. a block device might change ENOSPC to EIO when + * a failure timeout occurs, so we want to re-initialise the error + * retry behaviour appropriately when that happens. + */ + int b_retries; + unsigned long b_first_retry_time; /* in jiffies */ + int b_last_error; + const struct xfs_buf_ops *b_ops; #ifdef XFS_BUF_LOCK_TRACKING diff --git a/fs/xfs/xfs_buf_item.c b/fs/xfs/xfs_buf_item.c index b8d0cd4adb81..0d95c59f7c68 100644 --- a/fs/xfs/xfs_buf_item.c +++ b/fs/xfs/xfs_buf_item.c @@ -1085,6 +1085,9 @@ xfs_buf_iodone_callback_error( bp->b_flags |= (XBF_WRITE | XBF_ASYNC | XBF_DONE | XBF_WRITE_FAIL); bp->b_last_error = bp->b_error; + bp->b_retries = 0; + bp->b_first_retry_time = jiffies; + xfs_buf_ioerror(bp, 0); xfs_buf_submit(bp); return true; @@ -1095,8 +1098,13 @@ xfs_buf_iodone_callback_error( * error configuration we have been set up to use. */ cfg = xfs_error_get_cfg(mp, XFS_ERR_METADATA, bp->b_error); - if (!cfg->max_retries) - goto permanent_error; + + if (cfg->max_retries != XFS_ERR_RETRY_FOREVER && + ++bp->b_retries > cfg->max_retries) + goto permanent_error; + if (cfg->retry_timeout && + time_after(jiffies, cfg->retry_timeout + bp->b_first_retry_time)) + goto permanent_error; /* still a transient error, higher layers will retry */ xfs_buf_ioerror(bp, 0); @@ -1139,6 +1147,7 @@ xfs_buf_iodone_callbacks( * retry state here in preparation for the next error that may occur. */ bp->b_last_error = 0; + bp->b_retries = 0; xfs_buf_do_callbacks(bp); bp->b_fspriv = NULL; diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h index 0c5a97644d78..2fafa9438bcf 100644 --- a/fs/xfs/xfs_mount.h +++ b/fs/xfs/xfs_mount.h @@ -52,9 +52,12 @@ enum { XFS_ERR_ERRNO_MAX, }; +#define XFS_ERR_RETRY_FOREVER -1 + struct xfs_error_cfg { struct xfs_kobj kobj; int max_retries; + unsigned long retry_timeout; /* in jiffies, 0 = no timeout */ }; typedef struct xfs_mount { diff --git a/fs/xfs/xfs_sysfs.c b/fs/xfs/xfs_sysfs.c index 71046d904985..918d144febd9 100644 --- a/fs/xfs/xfs_sysfs.c +++ b/fs/xfs/xfs_sysfs.c @@ -374,10 +374,6 @@ struct kobj_type xfs_log_ktype = { * and any other future type of IO (e.g. special inode or directory error * handling) we care to support. */ -static struct attribute *xfs_error_attrs[] = { - NULL, -}; - static inline struct xfs_error_cfg * to_error_cfg(struct kobject *kobject) { @@ -385,6 +381,79 @@ to_error_cfg(struct kobject *kobject) return container_of(kobj, struct xfs_error_cfg, kobj); } +static ssize_t +max_retries_show( + struct kobject *kobject, + char *buf) +{ + struct xfs_error_cfg *cfg = to_error_cfg(kobject); + + return snprintf(buf, PAGE_SIZE, "%d\n", cfg->max_retries); +} + +static ssize_t +max_retries_store( + struct kobject *kobject, + const char *buf, + size_t count) +{ + struct xfs_error_cfg *cfg = to_error_cfg(kobject); + int ret; + int val; + + ret = kstrtoint(buf, 0, &val); + if (ret) + return ret; + + if (val < -1) + return -EINVAL; + + cfg->max_retries = val; + return count; +} +XFS_SYSFS_ATTR_RW(max_retries); + +static ssize_t +retry_timeout_seconds_show( + struct kobject *kobject, + char *buf) +{ + struct xfs_error_cfg *cfg = to_error_cfg(kobject); + + return snprintf(buf, PAGE_SIZE, "%ld\n", + jiffies_to_msecs(cfg->retry_timeout) / MSEC_PER_SEC); +} + +static ssize_t +retry_timeout_seconds_store( + struct kobject *kobject, + const char *buf, + size_t count) +{ + struct xfs_error_cfg *cfg = to_error_cfg(kobject); + int ret; + int val; + + ret = kstrtoint(buf, 0, &val); + if (ret) + return ret; + + /* 1 day timeout maximum */ + if (val < 0 || val > 86400) + return -EINVAL; + + cfg->retry_timeout = msecs_to_jiffies(val * MSEC_PER_SEC); + return count; +} +XFS_SYSFS_ATTR_RW(retry_timeout_seconds); + +static struct attribute *xfs_error_attrs[] = { + ATTR_LIST(max_retries), + ATTR_LIST(retry_timeout_seconds), + NULL, +}; + + struct kobj_type xfs_error_cfg_ktype = { .release = xfs_sysfs_release, .sysfs_ops = &xfs_sysfs_ops, @@ -404,11 +473,13 @@ struct kobj_type xfs_error_ktype = { struct xfs_error_init { char *name; int max_retries; + int retry_timeout; /* in seconds */ }; static const struct xfs_error_init xfs_error_meta_init[XFS_ERR_ERRNO_MAX] = { { .name = "default", .max_retries = -1, + .retry_timeout = 0, }, }; @@ -439,6 +510,8 @@ xfs_error_sysfs_init_class( goto out_error; cfg->max_retries = init[i].max_retries; + cfg->retry_timeout = msecs_to_jiffies( + init[i].retry_timeout * MSEC_PER_SEC); } return 0; -- cgit v1.3-8-gc7d7 From e0a431b3a3cc3d0a4c38ccfca8c7320fde40efb6 Mon Sep 17 00:00:00 2001 From: Carlos Maiolino Date: Wed, 18 May 2016 11:09:28 +1000 Subject: xfs: add configuration handlers for specific errors now most of the infrastructure is in place, we can start adding support for configuring specific errors such as ENODEV, ENOSPC, EIO, etc. Add these error configurations and configure them all to have appropriate behaviours. That is, all will be configured to retry forever by default, except for ENODEV, which is an unrecoverable error, so it will be configured to not retry on error Signed-off-by: Dave Chinner Signed-off-by: Carlos Maiolino Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/xfs_mount.h | 3 +++ fs/xfs/xfs_sysfs.c | 22 +++++++++++++++++++++- 2 files changed, 24 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h index 2fafa9438bcf..72ec3e3c988e 100644 --- a/fs/xfs/xfs_mount.h +++ b/fs/xfs/xfs_mount.h @@ -49,6 +49,9 @@ enum { }; enum { XFS_ERR_DEFAULT, + XFS_ERR_EIO, + XFS_ERR_ENOSPC, + XFS_ERR_ENODEV, XFS_ERR_ERRNO_MAX, }; diff --git a/fs/xfs/xfs_sysfs.c b/fs/xfs/xfs_sysfs.c index 918d144febd9..084a606840a1 100644 --- a/fs/xfs/xfs_sysfs.c +++ b/fs/xfs/xfs_sysfs.c @@ -478,9 +478,20 @@ struct xfs_error_init { static const struct xfs_error_init xfs_error_meta_init[XFS_ERR_ERRNO_MAX] = { { .name = "default", - .max_retries = -1, + .max_retries = XFS_ERR_RETRY_FOREVER, .retry_timeout = 0, }, + { .name = "EIO", + .max_retries = XFS_ERR_RETRY_FOREVER, + .retry_timeout = 0, + }, + { .name = "ENOSPC", + .max_retries = XFS_ERR_RETRY_FOREVER, + .retry_timeout = 0, + }, + { .name = "ENODEV", + .max_retries = 0, + }, }; static int @@ -578,6 +589,15 @@ xfs_error_get_cfg( struct xfs_error_cfg *cfg; switch (error) { + case EIO: + cfg = &mp->m_error_cfg[error_class][XFS_ERR_EIO]; + break; + case ENOSPC: + cfg = &mp->m_error_cfg[error_class][XFS_ERR_ENOSPC]; + break; + case ENODEV: + cfg = &mp->m_error_cfg[error_class][XFS_ERR_ENODEV]; + break; default: cfg = &mp->m_error_cfg[error_class][XFS_ERR_DEFAULT]; break; -- cgit v1.3-8-gc7d7 From e6b3bb78962e65c4ad125598755cfbf2a8779e86 Mon Sep 17 00:00:00 2001 From: Carlos Maiolino Date: Wed, 18 May 2016 11:11:27 +1000 Subject: xfs: add "fail at unmount" error handling configuration If we take "retry forever" literally on metadata IO errors, we can hang at unmount, once it retries those writes forever. This is the default behavior, unfortunately. Add an error configuration option for this behavior and default it to "fail" so that an unmount will trigger actuall errors, a shutdown and allow the unmount to succeed. It will be noisy, though, as it will log the errors and shutdown that occurs. To fix this, we need to mark the filesystem as being in the process of unmounting. Do this with a mount flag that is added at the appropriate time (i.e. before the blocking AIL sync). We also need to add this flag if mount fails after the initial phase of log recovery has been run. Signed-off-by: Dave Chinner Signed-off-by: Carlos Maiolino Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/xfs_buf_item.c | 4 ++++ fs/xfs/xfs_mount.c | 12 ++++++++++++ fs/xfs/xfs_mount.h | 2 ++ fs/xfs/xfs_sysfs.c | 46 ++++++++++++++++++++++++++++++++++++++++++++++ 4 files changed, 64 insertions(+) (limited to 'fs') diff --git a/fs/xfs/xfs_buf_item.c b/fs/xfs/xfs_buf_item.c index 0d95c59f7c68..34257992934c 100644 --- a/fs/xfs/xfs_buf_item.c +++ b/fs/xfs/xfs_buf_item.c @@ -1106,6 +1106,10 @@ xfs_buf_iodone_callback_error( time_after(jiffies, cfg->retry_timeout + bp->b_first_retry_time)) goto permanent_error; + /* At unmount we may treat errors differently */ + if ((mp->m_flags & XFS_MOUNT_UNMOUNTING) && mp->m_fail_unmount) + goto permanent_error; + /* still a transient error, higher layers will retry */ xfs_buf_ioerror(bp, 0); xfs_buf_relse(bp); diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index 677c3e0da472..7c05a22c1a73 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -681,6 +681,9 @@ xfs_mountfs( xfs_set_maxicount(mp); + /* enable fail_at_unmount as default */ + mp->m_fail_unmount = 1; + error = xfs_sysfs_init(&mp->m_kobj, &xfs_mp_ktype, NULL, mp->m_fsname); if (error) goto out; @@ -962,6 +965,7 @@ xfs_mountfs( cancel_delayed_work_sync(&mp->m_reclaim_work); xfs_reclaim_inodes(mp, SYNC_WAIT); out_log_dealloc: + mp->m_flags |= XFS_MOUNT_UNMOUNTING; xfs_log_mount_cancel(mp); out_fail_wait: if (mp->m_logdev_targp && mp->m_logdev_targp != mp->m_ddev_targp) @@ -1012,6 +1016,14 @@ xfs_unmountfs( */ xfs_log_force(mp, XFS_LOG_SYNC); + /* + * We now need to tell the world we are unmounting. This will allow + * us to detect that the filesystem is going away and we should error + * out anything that we have been retrying in the background. This will + * prevent neverending retries in AIL pushing from hanging the unmount. + */ + mp->m_flags |= XFS_MOUNT_UNMOUNTING; + /* * Flush all pending changes from the AIL. */ diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h index 72ec3e3c988e..9063a9c7b2fe 100644 --- a/fs/xfs/xfs_mount.h +++ b/fs/xfs/xfs_mount.h @@ -177,6 +177,7 @@ typedef struct xfs_mount { */ __uint32_t m_generation; + bool m_fail_unmount; #ifdef DEBUG /* * DEBUG mode instrumentation to test and/or trigger delayed allocation @@ -195,6 +196,7 @@ typedef struct xfs_mount { #define XFS_MOUNT_WSYNC (1ULL << 0) /* for nfs - all metadata ops must be synchronous except for space allocations */ +#define XFS_MOUNT_UNMOUNTING (1ULL << 1) /* filesystem is unmounting */ #define XFS_MOUNT_WAS_CLEAN (1ULL << 3) #define XFS_MOUNT_FS_SHUTDOWN (1ULL << 4) /* atomic stop of all filesystem operations, typically for diff --git a/fs/xfs/xfs_sysfs.c b/fs/xfs/xfs_sysfs.c index 084a606840a1..4c2c55086208 100644 --- a/fs/xfs/xfs_sysfs.c +++ b/fs/xfs/xfs_sysfs.c @@ -381,6 +381,13 @@ to_error_cfg(struct kobject *kobject) return container_of(kobj, struct xfs_error_cfg, kobj); } +static inline struct xfs_mount * +err_to_mp(struct kobject *kobject) +{ + struct xfs_kobj *kobj = to_kobj(kobject); + return container_of(kobj, struct xfs_mount, m_error_kobj); +} + static ssize_t max_retries_show( struct kobject *kobject, @@ -447,6 +454,38 @@ retry_timeout_seconds_store( } XFS_SYSFS_ATTR_RW(retry_timeout_seconds); +static ssize_t +fail_at_unmount_show( + struct kobject *kobject, + char *buf) +{ + struct xfs_mount *mp = err_to_mp(kobject); + + return snprintf(buf, PAGE_SIZE, "%d\n", mp->m_fail_unmount); +} + +static ssize_t +fail_at_unmount_store( + struct kobject *kobject, + const char *buf, + size_t count) +{ + struct xfs_mount *mp = err_to_mp(kobject); + int ret; + int val; + + ret = kstrtoint(buf, 0, &val); + if (ret) + return ret; + + if (val < 0 || val > 1) + return -EINVAL; + + mp->m_fail_unmount = val; + return count; +} +XFS_SYSFS_ATTR_RW(fail_at_unmount); + static struct attribute *xfs_error_attrs[] = { ATTR_LIST(max_retries), ATTR_LIST(retry_timeout_seconds), @@ -462,6 +501,7 @@ struct kobj_type xfs_error_cfg_ktype = { struct kobj_type xfs_error_ktype = { .release = xfs_sysfs_release, + .sysfs_ops = &xfs_sysfs_ops, }; /* @@ -548,6 +588,12 @@ xfs_error_sysfs_init( if (error) return error; + error = sysfs_create_file(&mp->m_error_kobj.kobject, + ATTR_LIST(fail_at_unmount)); + + if (error) + goto out_error; + /* .../xfs//error/metadata/ */ error = xfs_error_sysfs_init_class(mp, XFS_ERR_METADATA, "metadata", &mp->m_error_meta_kobj, -- cgit v1.3-8-gc7d7 From 8179c03629de67f515d3ab825b5a9428687d4b85 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Wed, 18 May 2016 13:52:42 +1000 Subject: xfs: remove xfs_fs_evict_inode() Joe Lawrence reported a list_add corruption with 4.6-rc1 when testing some custom md administration code that made it's own block device nodes for the md array. The simple test loop of: for i in {0..100}; do mknod --mode=0600 $tmp/tmp_node b $MAJOR $MINOR mdadm --detail --export $tmp/tmp_node > /dev/null rm -f $tmp/tmp_node done Would produce this warning in bd_acquire() when mdadm opened the device node: list_add double add: new=ffff88043831c7b8, prev=ffff8804380287d8, next=ffff88043831c7b8. And then produce this from bd_forget from kdevtmpfs evicting a block dev inode: list_del corruption. prev->next should be ffff8800bb83eb10, but was ffff88043831c7b8 This is a regression caused by commit c19b3b05 ("xfs: mode di_mode to vfs inode"). The issue is that xfs_inactive() frees the unlinked inode, and the above commit meant that this freeing zeroed the mode in the struct inode. The problem is that after evict() has called ->evict_inode, it expects the i_mode to be intact so that it can call bd_forget() or cd_forget() to drop the reference to the block device inode attached to the XFS inode. In reality, the only thing we do in xfs_fs_evict_inode() that is not generic is call xfs_inactive(). We can move the xfs_inactive() call to xfs_fs_destroy_inode() without any problems at all, and this will leave the VFS inode intact until it is completely done with it. So, remove xfs_fs_evict_inode(), and do the work it used to do in ->destroy_inode instead. cc: # 4.6 Reported-by: Joe Lawrence Signed-off-by: Dave Chinner Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/xfs_super.c | 28 +++++++--------------------- 1 file changed, 7 insertions(+), 21 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index d760934109b5..3c39f3ad2e32 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -928,7 +928,7 @@ xfs_fs_alloc_inode( /* * Now that the generic code is guaranteed not to be accessing - * the linux inode, we can reclaim the inode. + * the linux inode, we can inactivate and reclaim the inode. */ STATIC void xfs_fs_destroy_inode( @@ -938,9 +938,14 @@ xfs_fs_destroy_inode( trace_xfs_destroy_inode(ip); - XFS_STATS_INC(ip->i_mount, vn_reclaim); + ASSERT(!rwsem_is_locked(&ip->i_iolock.mr_lock)); + XFS_STATS_INC(ip->i_mount, vn_rele); + XFS_STATS_INC(ip->i_mount, vn_remove); + + xfs_inactive(ip); ASSERT(XFS_FORCED_SHUTDOWN(ip->i_mount) || ip->i_delayed_blks == 0); + XFS_STATS_INC(ip->i_mount, vn_reclaim); /* * We should never get here with one of the reclaim flags already set. @@ -987,24 +992,6 @@ xfs_fs_inode_init_once( "xfsino", ip->i_ino); } -STATIC void -xfs_fs_evict_inode( - struct inode *inode) -{ - xfs_inode_t *ip = XFS_I(inode); - - ASSERT(!rwsem_is_locked(&ip->i_iolock.mr_lock)); - - trace_xfs_evict_inode(ip); - - truncate_inode_pages_final(&inode->i_data); - clear_inode(inode); - XFS_STATS_INC(ip->i_mount, vn_rele); - XFS_STATS_INC(ip->i_mount, vn_remove); - - xfs_inactive(ip); -} - /* * We do an unlocked check for XFS_IDONTCACHE here because we are already * serialised against cache hits here via the inode->i_lock and igrab() in @@ -1663,7 +1650,6 @@ xfs_fs_free_cached_objects( static const struct super_operations xfs_super_operations = { .alloc_inode = xfs_fs_alloc_inode, .destroy_inode = xfs_fs_destroy_inode, - .evict_inode = xfs_fs_evict_inode, .drop_inode = xfs_fs_drop_inode, .put_super = xfs_fs_put_super, .sync_fs = xfs_fs_sync_fs, -- cgit v1.3-8-gc7d7 From b1438f477934f5a4d5a44df26f3079a7575d5946 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Wed, 18 May 2016 13:53:42 +1000 Subject: xfs: xfs_iflush_cluster fails to abort on error When a failure due to an inode buffer occurs, the error handling fails to abort the inode writeback correctly. This can result in the inode being reclaimed whilst still in the AIL, leading to use-after-free situations as well as filesystems that cannot be unmounted as the inode log items left in the AIL never get removed. Fix this by ensuring fatal errors from xfs_imap_to_bp() result in the inode flush being aborted correctly. cc: # 3.10.x- Reported-by: Shyam Kaushik Diagnosed-by: Shyam Kaushik Tested-by: Shyam Kaushik Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Dave Chinner --- fs/xfs/xfs_inode.c | 17 +++++++++++++---- 1 file changed, 13 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index 96f606deee31..696936cad0fa 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -3327,7 +3327,7 @@ xfs_iflush( struct xfs_buf **bpp) { struct xfs_mount *mp = ip->i_mount; - struct xfs_buf *bp; + struct xfs_buf *bp = NULL; struct xfs_dinode *dip; int error; @@ -3369,14 +3369,22 @@ xfs_iflush( } /* - * Get the buffer containing the on-disk inode. + * Get the buffer containing the on-disk inode. We are doing a try-lock + * operation here, so we may get an EAGAIN error. In that case, we + * simply want to return with the inode still dirty. + * + * If we get any other error, we effectively have a corruption situation + * and we cannot flush the inode, so we treat it the same as failing + * xfs_iflush_int(). */ error = xfs_imap_to_bp(mp, NULL, &ip->i_imap, &dip, &bp, XBF_TRYLOCK, 0); - if (error || !bp) { + if (error == -EAGAIN) { xfs_ifunlock(ip); return error; } + if (error) + goto corrupt_out; /* * First flush out the inode that xfs_iflush was called with. @@ -3404,7 +3412,8 @@ xfs_iflush( return 0; corrupt_out: - xfs_buf_relse(bp); + if (bp) + xfs_buf_relse(bp); xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE); cluster_corrupt_out: error = -EFSCORRUPTED; -- cgit v1.3-8-gc7d7 From 51b07f30a71c27405259a0248206ed4e22adbee2 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Wed, 18 May 2016 13:54:22 +1000 Subject: xfs: fix inode validity check in xfs_iflush_cluster Some careless idiot(*) wrote crap code in commit 1a3e8f3 ("xfs: convert inode cache lookups to use RCU locking") back in late 2010, and so xfs_iflush_cluster checks the wrong inode for whether it is still valid under RCU protection. Fix it to lock and check the correct inode. (*) Careless-idiot: Dave Chinner cc: # 3.10.x- Discovered-by: Brain Foster Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Dave Chinner --- fs/xfs/xfs_inode.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index 696936cad0fa..a955b0212453 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -3205,13 +3205,13 @@ xfs_iflush_cluster( * We need to check under the i_flags_lock for a valid inode * here. Skip it if it is not valid or the wrong inode. */ - spin_lock(&ip->i_flags_lock); - if (!ip->i_ino || + spin_lock(&iq->i_flags_lock); + if (!iq->i_ino || (XFS_INO_TO_AGINO(mp, iq->i_ino) & mask) != first_index) { - spin_unlock(&ip->i_flags_lock); + spin_unlock(&iq->i_flags_lock); continue; } - spin_unlock(&ip->i_flags_lock); + spin_unlock(&iq->i_flags_lock); /* * Do an un-protected check to see if the inode is dirty and -- cgit v1.3-8-gc7d7 From 7d3aa7fe970791f1a674b14572a411accf2f4d4e Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Wed, 18 May 2016 13:54:23 +1000 Subject: xfs: skip stale inodes in xfs_iflush_cluster We don't write back stale inodes so we should skip them in xfs_iflush_cluster, too. cc: # 3.10.x- Signed-off-by: Dave Chinner Reviewed-by: Brian Foster Reviewed-by: Christoph Hellwig Signed-off-by: Dave Chinner --- fs/xfs/xfs_inode.c | 1 + 1 file changed, 1 insertion(+) (limited to 'fs') diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index a955b0212453..3cbc9031731b 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -3207,6 +3207,7 @@ xfs_iflush_cluster( */ spin_lock(&iq->i_flags_lock); if (!iq->i_ino || + __xfs_iflags_test(iq, XFS_ISTALE) || (XFS_INO_TO_AGINO(mp, iq->i_ino) & mask) != first_index) { spin_unlock(&iq->i_flags_lock); continue; -- cgit v1.3-8-gc7d7 From 32b43ab6fb983e5a117048443e628c235cd2c5bd Mon Sep 17 00:00:00 2001 From: Alex Lyakas Date: Wed, 18 May 2016 14:01:52 +1000 Subject: xfs: optimise xfs_iext_destroy When unmounting XFS, we call: xfs_inode_free => xfs_idestroy_fork => xfs_iext_destroy This goes over the whole indirection array and calls xfs_iext_irec_remove for each one of the erps (from the last one to the first one). As a result, we keep shrinking (reallocating actually) the indirection array until we shrink out all of its elements. When we have files with huge numbers of extents, umount takes 30-80 sec, depending on the amount of files that XFS loaded and the amount of indirection entries of each file. The unmount stack looks like: [] xfs_iext_realloc_indirect+0x40/0x60 [xfs] [] xfs_iext_irec_remove+0xee/0xf0 [xfs] [] xfs_iext_destroy+0x3d/0xb0 [xfs] [] xfs_idestroy_fork+0xb6/0xf0 [xfs] [] xfs_inode_free+0xb2/0xc0 [xfs] [] xfs_reclaim_inode+0x250/0x340 [xfs] [] xfs_reclaim_inodes_ag+0x233/0x370 [xfs] [] xfs_reclaim_inodes+0x1d/0x20 [xfs] [] xfs_unmountfs+0x7b/0x1a0 [xfs] [] xfs_fs_put_super+0x2d/0x70 [xfs] [] generic_shutdown_super+0x76/0x100 [] kill_block_super+0x27/0x70 [] deactivate_locked_super+0x49/0x60 [] deactivate_super+0x4e/0x70 [] cleanup_mnt+0x43/0x90 [] __cleanup_mnt+0x12/0x20 [] task_work_run+0xa7/0xe0 [] do_notify_resume+0x97/0xb0 [] int_signal+0x12/0x17 Further, this reallocation prevents us from freeing the extent list from a RCU callback as allocation can block. Hence if the extent list is in indirect format, optimise the freeing of the extent list to only use kmem_free calls by freeing entire extent buffer pages at a time, rather than extent by extent. [dchinner: simplified freeing loop based on Christoph's suggestion] Signed-off-by: Alex Lyakas Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_inode_fork.c | 27 +++++++++++++++++++-------- 1 file changed, 19 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_inode_fork.c b/fs/xfs/libxfs/xfs_inode_fork.c index 11faf7df14c8..8fe6617a8653 100644 --- a/fs/xfs/libxfs/xfs_inode_fork.c +++ b/fs/xfs/libxfs/xfs_inode_fork.c @@ -1496,6 +1496,24 @@ xfs_iext_indirect_to_direct( } } +/* + * Remove all records from the indirection array. + */ +STATIC void +xfs_iext_irec_remove_all( + struct xfs_ifork *ifp) +{ + int nlists; + int i; + + ASSERT(ifp->if_flags & XFS_IFEXTIREC); + nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ; + for (i = 0; i < nlists; i++) + kmem_free(ifp->if_u1.if_ext_irec[i].er_extbuf); + kmem_free(ifp->if_u1.if_ext_irec); + ifp->if_flags &= ~XFS_IFEXTIREC; +} + /* * Free incore file extents. */ @@ -1504,14 +1522,7 @@ xfs_iext_destroy( xfs_ifork_t *ifp) /* inode fork pointer */ { if (ifp->if_flags & XFS_IFEXTIREC) { - int erp_idx; - int nlists; - - nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ; - for (erp_idx = nlists - 1; erp_idx >= 0 ; erp_idx--) { - xfs_iext_irec_remove(ifp, erp_idx); - } - ifp->if_flags &= ~XFS_IFEXTIREC; + xfs_iext_irec_remove_all(ifp); } else if (ifp->if_real_bytes) { kmem_free(ifp->if_u1.if_extents); } else if (ifp->if_bytes) { -- cgit v1.3-8-gc7d7 From 1f2dcfe89edac4e3bf5b76c56f745191f921fd2a Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Wed, 18 May 2016 14:01:53 +1000 Subject: xfs: xfs_inode_free() isn't RCU safe The xfs_inode freed in xfs_inode_free() has multiple allocated structures attached to it. We free these in xfs_inode_free() before we mark the inode as invalid, and before we run call_rcu() to queue the structure for freeing. Unfortunately, this freeing can race with other accesses that are in the RCU current grace period that have found the inode in the radix tree with a valid state. This includes xfs_iflush_cluster(), which calls xfs_inode_clean(), and that accesses the inode log item on the xfs_inode. The log item structure is freed in xfs_inode_free(), so there is the possibility we can be accessing freed memory in xfs_iflush_cluster() after validating the xfs_inode structure as being valid for this RCU context. Hence we can get spuriously incorrect clean state returned from such checks. This can lead to use thinking the inode is dirty when it is, in fact, clean, and so incorrectly attaching it to the buffer for IO and completion processing. This then leads to use-after-free situations on the xfs_inode itself if the IO completes after the current RCU grace period expires. The buffer callbacks will access the xfs_inode and try to do all sorts of things it shouldn't with freed memory. IOWs, xfs_iflush_cluster() only works correctly when racing with inode reclaim if the inode log item is present and correctly stating the inode is clean. If the inode is being freed, then reclaim has already made sure the inode is clean, and hence xfs_iflush_cluster can skip it. However, we are accessing the inode inode under RCU read lock protection and so also must ensure that all dynamically allocated memory we reference in this context is not freed until the RCU grace period expires. To fix this, move all the potential memory freeing into xfs_inode_free_callback() so that we are guarantee RCU protected lookup code will always have the memory structures it needs available during the RCU grace period that lookup races can occur in. Discovered-by: Brain Foster Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Dave Chinner --- fs/xfs/xfs_icache.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_icache.c b/fs/xfs/xfs_icache.c index bf2d60749278..0c94cde41016 100644 --- a/fs/xfs/xfs_icache.c +++ b/fs/xfs/xfs_icache.c @@ -94,13 +94,6 @@ xfs_inode_free_callback( struct inode *inode = container_of(head, struct inode, i_rcu); struct xfs_inode *ip = XFS_I(inode); - kmem_zone_free(xfs_inode_zone, ip); -} - -void -xfs_inode_free( - struct xfs_inode *ip) -{ switch (VFS_I(ip)->i_mode & S_IFMT) { case S_IFREG: case S_IFDIR: @@ -118,6 +111,13 @@ xfs_inode_free( ip->i_itemp = NULL; } + kmem_zone_free(xfs_inode_zone, ip); +} + +void +xfs_inode_free( + struct xfs_inode *ip) +{ /* * Because we use RCU freeing we need to ensure the inode always * appears to be reclaimed with an invalid inode number when in the -- cgit v1.3-8-gc7d7 From 8a17d7ddedb4d9031f046ae0e97c40b46aa69db5 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Wed, 18 May 2016 14:09:12 +1000 Subject: xfs: mark reclaimed inodes invalid earlier The last thing we do before using call_rcu() on an xfs_inode to be freed is mark it as invalid. This means there is a window between when we know for certain that the inode is going to be freed and when we do actually mark it as "freed". This is important in the context of RCU lookups - we can look up the inode, find that it is valid, and then use it as such not realising that it is in the final stages of being freed. As such, mark the inode as being invalid the moment we know it is going to be reclaimed. This can be done while we still hold the XFS_ILOCK_EXCL and the flush lock in xfs_inode_reclaim, meaning that it occurs well before we remove it from the radix tree, and that the i_flags_lock, the XFS_ILOCK and the inode flush lock all act as synchronisation points for detecting that an inode is about to go away. For defensive purposes, this allows us to add a further check to xfs_iflush_cluster to ensure we skip inodes that are being freed after we grab the XFS_ILOCK_SHARED and the flush lock - we know that if the inode number if valid while we have these locks held we know that it has not progressed through reclaim to the point where it is clean and is about to be freed. [bfoster: fixed __xfs_inode_clear_reclaim() using ip->i_ino after it had already been zeroed.] Signed-off-by: Dave Chinner Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/xfs_icache.c | 46 ++++++++++++++++++++++++++++++++++------------ fs/xfs/xfs_inode.c | 13 +++++++++++++ 2 files changed, 47 insertions(+), 12 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_icache.c b/fs/xfs/xfs_icache.c index 0c94cde41016..57fcd5917a66 100644 --- a/fs/xfs/xfs_icache.c +++ b/fs/xfs/xfs_icache.c @@ -114,6 +114,18 @@ xfs_inode_free_callback( kmem_zone_free(xfs_inode_zone, ip); } +static void +__xfs_inode_free( + struct xfs_inode *ip) +{ + /* asserts to verify all state is correct here */ + ASSERT(atomic_read(&ip->i_pincount) == 0); + ASSERT(!xfs_isiflocked(ip)); + XFS_STATS_DEC(ip->i_mount, vn_active); + + call_rcu(&VFS_I(ip)->i_rcu, xfs_inode_free_callback); +} + void xfs_inode_free( struct xfs_inode *ip) @@ -129,12 +141,7 @@ xfs_inode_free( ip->i_ino = 0; spin_unlock(&ip->i_flags_lock); - /* asserts to verify all state is correct here */ - ASSERT(atomic_read(&ip->i_pincount) == 0); - ASSERT(!xfs_isiflocked(ip)); - XFS_STATS_DEC(ip->i_mount, vn_active); - - call_rcu(&VFS_I(ip)->i_rcu, xfs_inode_free_callback); + __xfs_inode_free(ip); } /* @@ -772,8 +779,7 @@ __xfs_inode_set_reclaim_tag( if (!pag->pag_ici_reclaimable) { /* propagate the reclaim tag up into the perag radix tree */ spin_lock(&ip->i_mount->m_perag_lock); - radix_tree_tag_set(&ip->i_mount->m_perag_tree, - XFS_INO_TO_AGNO(ip->i_mount, ip->i_ino), + radix_tree_tag_set(&ip->i_mount->m_perag_tree, pag->pag_agno, XFS_ICI_RECLAIM_TAG); spin_unlock(&ip->i_mount->m_perag_lock); @@ -817,8 +823,7 @@ __xfs_inode_clear_reclaim( if (!pag->pag_ici_reclaimable) { /* clear the reclaim tag from the perag radix tree */ spin_lock(&ip->i_mount->m_perag_lock); - radix_tree_tag_clear(&ip->i_mount->m_perag_tree, - XFS_INO_TO_AGNO(ip->i_mount, ip->i_ino), + radix_tree_tag_clear(&ip->i_mount->m_perag_tree, pag->pag_agno, XFS_ICI_RECLAIM_TAG); spin_unlock(&ip->i_mount->m_perag_lock); trace_xfs_perag_clear_reclaim(ip->i_mount, pag->pag_agno, @@ -929,6 +934,7 @@ xfs_reclaim_inode( int sync_mode) { struct xfs_buf *bp = NULL; + xfs_ino_t ino = ip->i_ino; /* for radix_tree_delete */ int error; restart: @@ -993,6 +999,22 @@ restart: xfs_iflock(ip); reclaim: + /* + * Because we use RCU freeing we need to ensure the inode always appears + * to be reclaimed with an invalid inode number when in the free state. + * We do this as early as possible under the ILOCK and flush lock so + * that xfs_iflush_cluster() can be guaranteed to detect races with us + * here. By doing this, we guarantee that once xfs_iflush_cluster has + * locked both the XFS_ILOCK and the flush lock that it will see either + * a valid, flushable inode that will serialise correctly against the + * locks below, or it will see a clean (and invalid) inode that it can + * skip. + */ + spin_lock(&ip->i_flags_lock); + ip->i_flags = XFS_IRECLAIM; + ip->i_ino = 0; + spin_unlock(&ip->i_flags_lock); + xfs_ifunlock(ip); xfs_iunlock(ip, XFS_ILOCK_EXCL); @@ -1006,7 +1028,7 @@ reclaim: */ spin_lock(&pag->pag_ici_lock); if (!radix_tree_delete(&pag->pag_ici_root, - XFS_INO_TO_AGINO(ip->i_mount, ip->i_ino))) + XFS_INO_TO_AGINO(ip->i_mount, ino))) ASSERT(0); __xfs_inode_clear_reclaim(pag, ip); spin_unlock(&pag->pag_ici_lock); @@ -1023,7 +1045,7 @@ reclaim: xfs_qm_dqdetach(ip); xfs_iunlock(ip, XFS_ILOCK_EXCL); - xfs_inode_free(ip); + __xfs_inode_free(ip); return error; out_ifunlock: diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index 3cbc9031731b..e3b27982b3b2 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -3239,6 +3239,19 @@ xfs_iflush_cluster( continue; } + + /* + * Check the inode number again, just to be certain we are not + * racing with freeing in xfs_reclaim_inode(). See the comments + * in that function for more information as to why the initial + * check is not sufficient. + */ + if (!iq->i_ino) { + xfs_ifunlock(iq); + xfs_iunlock(iq, XFS_ILOCK_SHARED); + continue; + } + /* * arriving here means that this inode can be flushed. First * re-check that it's dirty before flushing. -- cgit v1.3-8-gc7d7 From 5a90e53e8124d3ebe4b2a6309fa3c3225c23a62a Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Wed, 18 May 2016 14:09:13 +1000 Subject: xfs: xfs_iflush_cluster has range issues xfs_iflush_cluster() does a gang lookup on the radix tree, meaning it can find inodes beyond the current cluster if there is sparse cache population. gang lookups return results in ascending index order, so stop trying to cluster inodes once the first inode outside the cluster mask is detected. Signed-off-by: Dave Chinner Reviewed-by: Brian Foster Reviewed-by: Christoph Hellwig Signed-off-by: Dave Chinner --- fs/xfs/xfs_inode.c | 13 +++++++++++-- 1 file changed, 11 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index e3b27982b3b2..6c746d79925e 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -3207,11 +3207,20 @@ xfs_iflush_cluster( */ spin_lock(&iq->i_flags_lock); if (!iq->i_ino || - __xfs_iflags_test(iq, XFS_ISTALE) || - (XFS_INO_TO_AGINO(mp, iq->i_ino) & mask) != first_index) { + __xfs_iflags_test(iq, XFS_ISTALE)) { spin_unlock(&iq->i_flags_lock); continue; } + + /* + * Once we fall off the end of the cluster, no point checking + * any more inodes in the list because they will also all be + * outside the cluster. + */ + if ((XFS_INO_TO_AGINO(mp, iq->i_ino) & mask) != first_index) { + spin_unlock(&iq->i_flags_lock); + break; + } spin_unlock(&iq->i_flags_lock); /* -- cgit v1.3-8-gc7d7 From 194293631d009254348f43710a7673bbb84a4172 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Wed, 18 May 2016 14:09:46 +1000 Subject: xfs: rename variables in xfs_iflush_cluster for clarity The cluster inode variable uses unconventional naming - iq - which makes it hard to distinguish it between the inode passed into the function - ip - and that is a vector for mistakes to be made. Rename all the cluster inode variables to use a more conventional prefixes to reduce potential future confusion (cilist, cilist_size, cip). Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Dave Chinner --- fs/xfs/xfs_inode.c | 74 +++++++++++++++++++++++++++--------------------------- 1 file changed, 37 insertions(+), 37 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index 6c746d79925e..54cc7b05708f 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -3162,16 +3162,16 @@ out_trans_cancel: STATIC int xfs_iflush_cluster( - xfs_inode_t *ip, - xfs_buf_t *bp) + struct xfs_inode *ip, + struct xfs_buf *bp) { - xfs_mount_t *mp = ip->i_mount; + struct xfs_mount *mp = ip->i_mount; struct xfs_perag *pag; unsigned long first_index, mask; unsigned long inodes_per_cluster; - int ilist_size; - xfs_inode_t **ilist; - xfs_inode_t *iq; + int cilist_size; + struct xfs_inode **cilist; + struct xfs_inode *cip; int nr_found; int clcount = 0; int bufwasdelwri; @@ -3180,23 +3180,23 @@ xfs_iflush_cluster( pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, ip->i_ino)); inodes_per_cluster = mp->m_inode_cluster_size >> mp->m_sb.sb_inodelog; - ilist_size = inodes_per_cluster * sizeof(xfs_inode_t *); - ilist = kmem_alloc(ilist_size, KM_MAYFAIL|KM_NOFS); - if (!ilist) + cilist_size = inodes_per_cluster * sizeof(xfs_inode_t *); + cilist = kmem_alloc(cilist_size, KM_MAYFAIL|KM_NOFS); + if (!cilist) goto out_put; mask = ~(((mp->m_inode_cluster_size >> mp->m_sb.sb_inodelog)) - 1); first_index = XFS_INO_TO_AGINO(mp, ip->i_ino) & mask; rcu_read_lock(); /* really need a gang lookup range call here */ - nr_found = radix_tree_gang_lookup(&pag->pag_ici_root, (void**)ilist, + nr_found = radix_tree_gang_lookup(&pag->pag_ici_root, (void**)cilist, first_index, inodes_per_cluster); if (nr_found == 0) goto out_free; for (i = 0; i < nr_found; i++) { - iq = ilist[i]; - if (iq == ip) + cip = cilist[i]; + if (cip == ip) continue; /* @@ -3205,10 +3205,10 @@ xfs_iflush_cluster( * We need to check under the i_flags_lock for a valid inode * here. Skip it if it is not valid or the wrong inode. */ - spin_lock(&iq->i_flags_lock); - if (!iq->i_ino || - __xfs_iflags_test(iq, XFS_ISTALE)) { - spin_unlock(&iq->i_flags_lock); + spin_lock(&cip->i_flags_lock); + if (!cip->i_ino || + __xfs_iflags_test(cip, XFS_ISTALE)) { + spin_unlock(&cip->i_flags_lock); continue; } @@ -3217,18 +3217,18 @@ xfs_iflush_cluster( * any more inodes in the list because they will also all be * outside the cluster. */ - if ((XFS_INO_TO_AGINO(mp, iq->i_ino) & mask) != first_index) { - spin_unlock(&iq->i_flags_lock); + if ((XFS_INO_TO_AGINO(mp, cip->i_ino) & mask) != first_index) { + spin_unlock(&cip->i_flags_lock); break; } - spin_unlock(&iq->i_flags_lock); + spin_unlock(&cip->i_flags_lock); /* * Do an un-protected check to see if the inode is dirty and * is a candidate for flushing. These checks will be repeated * later after the appropriate locks are acquired. */ - if (xfs_inode_clean(iq) && xfs_ipincount(iq) == 0) + if (xfs_inode_clean(cip) && xfs_ipincount(cip) == 0) continue; /* @@ -3236,15 +3236,15 @@ xfs_iflush_cluster( * then this inode cannot be flushed and is skipped. */ - if (!xfs_ilock_nowait(iq, XFS_ILOCK_SHARED)) + if (!xfs_ilock_nowait(cip, XFS_ILOCK_SHARED)) continue; - if (!xfs_iflock_nowait(iq)) { - xfs_iunlock(iq, XFS_ILOCK_SHARED); + if (!xfs_iflock_nowait(cip)) { + xfs_iunlock(cip, XFS_ILOCK_SHARED); continue; } - if (xfs_ipincount(iq)) { - xfs_ifunlock(iq); - xfs_iunlock(iq, XFS_ILOCK_SHARED); + if (xfs_ipincount(cip)) { + xfs_ifunlock(cip); + xfs_iunlock(cip, XFS_ILOCK_SHARED); continue; } @@ -3255,9 +3255,9 @@ xfs_iflush_cluster( * in that function for more information as to why the initial * check is not sufficient. */ - if (!iq->i_ino) { - xfs_ifunlock(iq); - xfs_iunlock(iq, XFS_ILOCK_SHARED); + if (!cip->i_ino) { + xfs_ifunlock(cip); + xfs_iunlock(cip, XFS_ILOCK_SHARED); continue; } @@ -3265,18 +3265,18 @@ xfs_iflush_cluster( * arriving here means that this inode can be flushed. First * re-check that it's dirty before flushing. */ - if (!xfs_inode_clean(iq)) { + if (!xfs_inode_clean(cip)) { int error; - error = xfs_iflush_int(iq, bp); + error = xfs_iflush_int(cip, bp); if (error) { - xfs_iunlock(iq, XFS_ILOCK_SHARED); + xfs_iunlock(cip, XFS_ILOCK_SHARED); goto cluster_corrupt_out; } clcount++; } else { - xfs_ifunlock(iq); + xfs_ifunlock(cip); } - xfs_iunlock(iq, XFS_ILOCK_SHARED); + xfs_iunlock(cip, XFS_ILOCK_SHARED); } if (clcount) { @@ -3286,7 +3286,7 @@ xfs_iflush_cluster( out_free: rcu_read_unlock(); - kmem_free(ilist); + kmem_free(cilist); out_put: xfs_perag_put(pag); return 0; @@ -3329,8 +3329,8 @@ cluster_corrupt_out: /* * Unlocks the flush lock */ - xfs_iflush_abort(iq, false); - kmem_free(ilist); + xfs_iflush_abort(cip, false); + kmem_free(cilist); xfs_perag_put(pag); return -EFSCORRUPTED; } -- cgit v1.3-8-gc7d7 From 545c0889d26d47e1139c527002eb131343d13b63 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Wed, 18 May 2016 14:11:41 +1000 Subject: xfs: simplify inode reclaim tagging interfaces Inode radix tree tagging for reclaim passes a lot of unnecessary variables around. Over time the xfs-perag has grown a xfs_mount backpointer, and an internal agno so we don't need to pass other variables into the tagging functions to supply this information. Rework the functions to pass the minimal variable set required and simplify the internal logic and flow. Signed-off-by: Dave Chinner Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/xfs_icache.c | 98 +++++++++++++++++++++++++++-------------------------- 1 file changed, 50 insertions(+), 48 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_icache.c b/fs/xfs/xfs_icache.c index 57fcd5917a66..789f8c32e65f 100644 --- a/fs/xfs/xfs_icache.c +++ b/fs/xfs/xfs_icache.c @@ -37,8 +37,7 @@ #include #include -STATIC void __xfs_inode_clear_reclaim_tag(struct xfs_mount *mp, - struct xfs_perag *pag, struct xfs_inode *ip); +STATIC void xfs_inode_clear_reclaim_tag(struct xfs_perag *pag, xfs_ino_t ino); /* * Allocate and initialise an xfs_inode. @@ -271,7 +270,7 @@ xfs_iget_cache_hit( */ ip->i_flags &= ~XFS_IRECLAIM_RESET_FLAGS; ip->i_flags |= XFS_INEW; - __xfs_inode_clear_reclaim_tag(mp, pag, ip); + xfs_inode_clear_reclaim_tag(pag, ip->i_ino); inode->i_state = I_NEW; ASSERT(!rwsem_is_locked(&ip->i_iolock.mr_lock)); @@ -768,30 +767,46 @@ xfs_reclaim_worker( } static void -__xfs_inode_set_reclaim_tag( - struct xfs_perag *pag, - struct xfs_inode *ip) +xfs_perag_set_reclaim_tag( + struct xfs_perag *pag) { - radix_tree_tag_set(&pag->pag_ici_root, - XFS_INO_TO_AGINO(ip->i_mount, ip->i_ino), + struct xfs_mount *mp = pag->pag_mount; + + ASSERT(spin_is_locked(&pag->pag_ici_lock)); + if (pag->pag_ici_reclaimable++) + return; + + /* propagate the reclaim tag up into the perag radix tree */ + spin_lock(&mp->m_perag_lock); + radix_tree_tag_set(&mp->m_perag_tree, pag->pag_agno, XFS_ICI_RECLAIM_TAG); + spin_unlock(&mp->m_perag_lock); - if (!pag->pag_ici_reclaimable) { - /* propagate the reclaim tag up into the perag radix tree */ - spin_lock(&ip->i_mount->m_perag_lock); - radix_tree_tag_set(&ip->i_mount->m_perag_tree, pag->pag_agno, - XFS_ICI_RECLAIM_TAG); - spin_unlock(&ip->i_mount->m_perag_lock); + /* schedule periodic background inode reclaim */ + xfs_reclaim_work_queue(mp); - /* schedule periodic background inode reclaim */ - xfs_reclaim_work_queue(ip->i_mount); + trace_xfs_perag_set_reclaim(mp, pag->pag_agno, -1, _RET_IP_); +} - trace_xfs_perag_set_reclaim(ip->i_mount, pag->pag_agno, - -1, _RET_IP_); - } - pag->pag_ici_reclaimable++; +static void +xfs_perag_clear_reclaim_tag( + struct xfs_perag *pag) +{ + struct xfs_mount *mp = pag->pag_mount; + + ASSERT(spin_is_locked(&pag->pag_ici_lock)); + if (--pag->pag_ici_reclaimable) + return; + + /* clear the reclaim tag from the perag radix tree */ + spin_lock(&mp->m_perag_lock); + radix_tree_tag_clear(&mp->m_perag_tree, pag->pag_agno, + XFS_ICI_RECLAIM_TAG); + spin_unlock(&mp->m_perag_lock); + trace_xfs_perag_clear_reclaim(mp, pag->pag_agno, -1, _RET_IP_); } + /* * We set the inode flag atomically with the radix tree tag. * Once we get tag lookups on the radix tree, this inode flag @@ -799,47 +814,34 @@ __xfs_inode_set_reclaim_tag( */ void xfs_inode_set_reclaim_tag( - xfs_inode_t *ip) + struct xfs_inode *ip) { - struct xfs_mount *mp = ip->i_mount; - struct xfs_perag *pag; + struct xfs_mount *mp = ip->i_mount; + struct xfs_perag *pag; pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, ip->i_ino)); spin_lock(&pag->pag_ici_lock); spin_lock(&ip->i_flags_lock); - __xfs_inode_set_reclaim_tag(pag, ip); + + radix_tree_tag_set(&pag->pag_ici_root, XFS_INO_TO_AGINO(mp, ip->i_ino), + XFS_ICI_RECLAIM_TAG); + xfs_perag_set_reclaim_tag(pag); __xfs_iflags_set(ip, XFS_IRECLAIMABLE); + spin_unlock(&ip->i_flags_lock); spin_unlock(&pag->pag_ici_lock); xfs_perag_put(pag); } STATIC void -__xfs_inode_clear_reclaim( - xfs_perag_t *pag, - xfs_inode_t *ip) -{ - pag->pag_ici_reclaimable--; - if (!pag->pag_ici_reclaimable) { - /* clear the reclaim tag from the perag radix tree */ - spin_lock(&ip->i_mount->m_perag_lock); - radix_tree_tag_clear(&ip->i_mount->m_perag_tree, pag->pag_agno, - XFS_ICI_RECLAIM_TAG); - spin_unlock(&ip->i_mount->m_perag_lock); - trace_xfs_perag_clear_reclaim(ip->i_mount, pag->pag_agno, - -1, _RET_IP_); - } -} - -STATIC void -__xfs_inode_clear_reclaim_tag( - xfs_mount_t *mp, - xfs_perag_t *pag, - xfs_inode_t *ip) +xfs_inode_clear_reclaim_tag( + struct xfs_perag *pag, + xfs_ino_t ino) { radix_tree_tag_clear(&pag->pag_ici_root, - XFS_INO_TO_AGINO(mp, ip->i_ino), XFS_ICI_RECLAIM_TAG); - __xfs_inode_clear_reclaim(pag, ip); + XFS_INO_TO_AGINO(pag->pag_mount, ino), + XFS_ICI_RECLAIM_TAG); + xfs_perag_clear_reclaim_tag(pag); } /* @@ -1030,7 +1032,7 @@ reclaim: if (!radix_tree_delete(&pag->pag_ici_root, XFS_INO_TO_AGINO(ip->i_mount, ino))) ASSERT(0); - __xfs_inode_clear_reclaim(pag, ip); + xfs_perag_clear_reclaim_tag(pag); spin_unlock(&pag->pag_ici_lock); /* -- cgit v1.3-8-gc7d7 From ad438c4038968e5ca5248f851212634e474983e8 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Wed, 18 May 2016 14:20:08 +1000 Subject: xfs: move reclaim tagging functions Rearrange the inode tagging functions so that they are higher up in xfs_cache.c and so there is no need for forward prototypes to be defined. This is purely code movement, no other change. Signed-off-by: Dave Chinner --- fs/xfs/xfs_icache.c | 234 ++++++++++++++++++++++++++-------------------------- 1 file changed, 116 insertions(+), 118 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_icache.c b/fs/xfs/xfs_icache.c index 789f8c32e65f..99ee6eee5e0b 100644 --- a/fs/xfs/xfs_icache.c +++ b/fs/xfs/xfs_icache.c @@ -37,8 +37,6 @@ #include #include -STATIC void xfs_inode_clear_reclaim_tag(struct xfs_perag *pag, xfs_ino_t ino); - /* * Allocate and initialise an xfs_inode. */ @@ -143,6 +141,122 @@ xfs_inode_free( __xfs_inode_free(ip); } +/* + * Queue a new inode reclaim pass if there are reclaimable inodes and there + * isn't a reclaim pass already in progress. By default it runs every 5s based + * on the xfs periodic sync default of 30s. Perhaps this should have it's own + * tunable, but that can be done if this method proves to be ineffective or too + * aggressive. + */ +static void +xfs_reclaim_work_queue( + struct xfs_mount *mp) +{ + + rcu_read_lock(); + if (radix_tree_tagged(&mp->m_perag_tree, XFS_ICI_RECLAIM_TAG)) { + queue_delayed_work(mp->m_reclaim_workqueue, &mp->m_reclaim_work, + msecs_to_jiffies(xfs_syncd_centisecs / 6 * 10)); + } + rcu_read_unlock(); +} + +/* + * This is a fast pass over the inode cache to try to get reclaim moving on as + * many inodes as possible in a short period of time. It kicks itself every few + * seconds, as well as being kicked by the inode cache shrinker when memory + * goes low. It scans as quickly as possible avoiding locked inodes or those + * already being flushed, and once done schedules a future pass. + */ +void +xfs_reclaim_worker( + struct work_struct *work) +{ + struct xfs_mount *mp = container_of(to_delayed_work(work), + struct xfs_mount, m_reclaim_work); + + xfs_reclaim_inodes(mp, SYNC_TRYLOCK); + xfs_reclaim_work_queue(mp); +} + +static void +xfs_perag_set_reclaim_tag( + struct xfs_perag *pag) +{ + struct xfs_mount *mp = pag->pag_mount; + + ASSERT(spin_is_locked(&pag->pag_ici_lock)); + if (pag->pag_ici_reclaimable++) + return; + + /* propagate the reclaim tag up into the perag radix tree */ + spin_lock(&mp->m_perag_lock); + radix_tree_tag_set(&mp->m_perag_tree, pag->pag_agno, + XFS_ICI_RECLAIM_TAG); + spin_unlock(&mp->m_perag_lock); + + /* schedule periodic background inode reclaim */ + xfs_reclaim_work_queue(mp); + + trace_xfs_perag_set_reclaim(mp, pag->pag_agno, -1, _RET_IP_); +} + +static void +xfs_perag_clear_reclaim_tag( + struct xfs_perag *pag) +{ + struct xfs_mount *mp = pag->pag_mount; + + ASSERT(spin_is_locked(&pag->pag_ici_lock)); + if (--pag->pag_ici_reclaimable) + return; + + /* clear the reclaim tag from the perag radix tree */ + spin_lock(&mp->m_perag_lock); + radix_tree_tag_clear(&mp->m_perag_tree, pag->pag_agno, + XFS_ICI_RECLAIM_TAG); + spin_unlock(&mp->m_perag_lock); + trace_xfs_perag_clear_reclaim(mp, pag->pag_agno, -1, _RET_IP_); +} + + +/* + * We set the inode flag atomically with the radix tree tag. + * Once we get tag lookups on the radix tree, this inode flag + * can go away. + */ +void +xfs_inode_set_reclaim_tag( + struct xfs_inode *ip) +{ + struct xfs_mount *mp = ip->i_mount; + struct xfs_perag *pag; + + pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, ip->i_ino)); + spin_lock(&pag->pag_ici_lock); + spin_lock(&ip->i_flags_lock); + + radix_tree_tag_set(&pag->pag_ici_root, XFS_INO_TO_AGINO(mp, ip->i_ino), + XFS_ICI_RECLAIM_TAG); + xfs_perag_set_reclaim_tag(pag); + __xfs_iflags_set(ip, XFS_IRECLAIMABLE); + + spin_unlock(&ip->i_flags_lock); + spin_unlock(&pag->pag_ici_lock); + xfs_perag_put(pag); +} + +STATIC void +xfs_inode_clear_reclaim_tag( + struct xfs_perag *pag, + xfs_ino_t ino) +{ + radix_tree_tag_clear(&pag->pag_ici_root, + XFS_INO_TO_AGINO(pag->pag_mount, ino), + XFS_ICI_RECLAIM_TAG); + xfs_perag_clear_reclaim_tag(pag); +} + /* * When we recycle a reclaimable inode, we need to re-initialise the VFS inode * part of the structure. This is made more complex by the fact we store @@ -728,122 +842,6 @@ xfs_inode_ag_iterator_tag( return last_error; } -/* - * Queue a new inode reclaim pass if there are reclaimable inodes and there - * isn't a reclaim pass already in progress. By default it runs every 5s based - * on the xfs periodic sync default of 30s. Perhaps this should have it's own - * tunable, but that can be done if this method proves to be ineffective or too - * aggressive. - */ -static void -xfs_reclaim_work_queue( - struct xfs_mount *mp) -{ - - rcu_read_lock(); - if (radix_tree_tagged(&mp->m_perag_tree, XFS_ICI_RECLAIM_TAG)) { - queue_delayed_work(mp->m_reclaim_workqueue, &mp->m_reclaim_work, - msecs_to_jiffies(xfs_syncd_centisecs / 6 * 10)); - } - rcu_read_unlock(); -} - -/* - * This is a fast pass over the inode cache to try to get reclaim moving on as - * many inodes as possible in a short period of time. It kicks itself every few - * seconds, as well as being kicked by the inode cache shrinker when memory - * goes low. It scans as quickly as possible avoiding locked inodes or those - * already being flushed, and once done schedules a future pass. - */ -void -xfs_reclaim_worker( - struct work_struct *work) -{ - struct xfs_mount *mp = container_of(to_delayed_work(work), - struct xfs_mount, m_reclaim_work); - - xfs_reclaim_inodes(mp, SYNC_TRYLOCK); - xfs_reclaim_work_queue(mp); -} - -static void -xfs_perag_set_reclaim_tag( - struct xfs_perag *pag) -{ - struct xfs_mount *mp = pag->pag_mount; - - ASSERT(spin_is_locked(&pag->pag_ici_lock)); - if (pag->pag_ici_reclaimable++) - return; - - /* propagate the reclaim tag up into the perag radix tree */ - spin_lock(&mp->m_perag_lock); - radix_tree_tag_set(&mp->m_perag_tree, pag->pag_agno, - XFS_ICI_RECLAIM_TAG); - spin_unlock(&mp->m_perag_lock); - - /* schedule periodic background inode reclaim */ - xfs_reclaim_work_queue(mp); - - trace_xfs_perag_set_reclaim(mp, pag->pag_agno, -1, _RET_IP_); -} - -static void -xfs_perag_clear_reclaim_tag( - struct xfs_perag *pag) -{ - struct xfs_mount *mp = pag->pag_mount; - - ASSERT(spin_is_locked(&pag->pag_ici_lock)); - if (--pag->pag_ici_reclaimable) - return; - - /* clear the reclaim tag from the perag radix tree */ - spin_lock(&mp->m_perag_lock); - radix_tree_tag_clear(&mp->m_perag_tree, pag->pag_agno, - XFS_ICI_RECLAIM_TAG); - spin_unlock(&mp->m_perag_lock); - trace_xfs_perag_clear_reclaim(mp, pag->pag_agno, -1, _RET_IP_); -} - - -/* - * We set the inode flag atomically with the radix tree tag. - * Once we get tag lookups on the radix tree, this inode flag - * can go away. - */ -void -xfs_inode_set_reclaim_tag( - struct xfs_inode *ip) -{ - struct xfs_mount *mp = ip->i_mount; - struct xfs_perag *pag; - - pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, ip->i_ino)); - spin_lock(&pag->pag_ici_lock); - spin_lock(&ip->i_flags_lock); - - radix_tree_tag_set(&pag->pag_ici_root, XFS_INO_TO_AGINO(mp, ip->i_ino), - XFS_ICI_RECLAIM_TAG); - xfs_perag_set_reclaim_tag(pag); - __xfs_iflags_set(ip, XFS_IRECLAIMABLE); - - spin_unlock(&ip->i_flags_lock); - spin_unlock(&pag->pag_ici_lock); - xfs_perag_put(pag); -} - -STATIC void -xfs_inode_clear_reclaim_tag( - struct xfs_perag *pag, - xfs_ino_t ino) -{ - radix_tree_tag_clear(&pag->pag_ici_root, - XFS_INO_TO_AGINO(pag->pag_mount, ino), - XFS_ICI_RECLAIM_TAG); - xfs_perag_clear_reclaim_tag(pag); -} - /* * Grab the inode for reclaim exclusively. * Return 0 if we grabbed it, non-zero otherwise. -- cgit v1.3-8-gc7d7 From fe742fd4f90fa53cf31296bc5131ae1cdd6d84bb Mon Sep 17 00:00:00 2001 From: Al Viro Date: Wed, 18 May 2016 13:15:05 -0400 Subject: Revert "btrfs: switch to ->iterate_shared()" This reverts commit 972b241f8441dc37a3f89dcd7e71d7f013873d13. Quoth Chris: didn't take the delayed inode stuff into account it got an rbtree of items and it pulls things out so in shared mode, its hugely racey sorry, lets revert and fix it for real inside of btrfs Signed-off-by: Chris Mason Signed-off-by: Al Viro --- fs/btrfs/inode.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 3e2ada1267f3..2aaba58b4856 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -10181,7 +10181,7 @@ static const struct inode_operations btrfs_dir_ro_inode_operations = { static const struct file_operations btrfs_dir_file_operations = { .llseek = generic_file_llseek, .read = generic_read_dir, - .iterate_shared = btrfs_real_readdir, + .iterate = btrfs_real_readdir, .unlocked_ioctl = btrfs_ioctl, #ifdef CONFIG_COMPAT .compat_ioctl = btrfs_ioctl, -- cgit v1.3-8-gc7d7 From 9f5418010940236b2c39ea53b99055ca26ff1279 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Thu, 19 May 2016 00:17:26 +1000 Subject: xfs: concurrent readdir hangs on data buffer locks There's a three-process deadlock involving shared/exclusive barriers and inverted lock orders in the directory readdir implementation. It's a pre-existing problem with lock ordering, exposed by the VFS parallelisation code. process 1 process 2 process 3 --------- --------- --------- readdir iolock(shared) get_leaf_dents iterate entries ilock(shared) map, lock and read buffer iunlock(shared) process entries in buffer ..... readdir iolock(shared) get_leaf_dents iterate entries ilock(shared) map, lock buffer finish ->iterate_shared file_accessed() ->update_time start transaction ilock(excl) ..... finishes processing buffer get next buffer ilock(shared) And that's the deadlock. Fix this by dropping the current buffer lock in process 1 before trying to map the next buffer. This means we keep the lock order of ilock -> buffer lock intact and hence will allow process 3 to make progress and drop it's ilock(shared) once it is done. Reported-by: Xiong Zhou Signed-off-by: Dave Chinner Signed-off-by: Al Viro --- fs/xfs/xfs_dir2_readdir.c | 23 ++++++++++++++--------- 1 file changed, 14 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_dir2_readdir.c b/fs/xfs/xfs_dir2_readdir.c index 93b3ab0c5435..f44f79996978 100644 --- a/fs/xfs/xfs_dir2_readdir.c +++ b/fs/xfs/xfs_dir2_readdir.c @@ -273,10 +273,11 @@ xfs_dir2_leaf_readbuf( size_t bufsize, struct xfs_dir2_leaf_map_info *mip, xfs_dir2_off_t *curoff, - struct xfs_buf **bpp) + struct xfs_buf **bpp, + bool trim_map) { struct xfs_inode *dp = args->dp; - struct xfs_buf *bp = *bpp; + struct xfs_buf *bp = NULL; struct xfs_bmbt_irec *map = mip->map; struct blk_plug plug; int error = 0; @@ -286,13 +287,10 @@ xfs_dir2_leaf_readbuf( struct xfs_da_geometry *geo = args->geo; /* - * If we have a buffer, we need to release it and - * take it out of the mapping. + * If the caller just finished processing a buffer, it will tell us + * we need to trim that block out of the mapping now it is done. */ - - if (bp) { - xfs_trans_brelse(NULL, bp); - bp = NULL; + if (trim_map) { mip->map_blocks -= geo->fsbcount; /* * Loop to get rid of the extents for the @@ -533,10 +531,17 @@ xfs_dir2_leaf_getdents( */ if (!bp || ptr >= (char *)bp->b_addr + geo->blksize) { int lock_mode; + bool trim_map = false; + + if (bp) { + xfs_trans_brelse(NULL, bp); + bp = NULL; + trim_map = true; + } lock_mode = xfs_ilock_data_map_shared(dp); error = xfs_dir2_leaf_readbuf(args, bufsize, map_info, - &curoff, &bp); + &curoff, &bp, trim_map); xfs_iunlock(dp, lock_mode); if (error || !map_info->map_valid) break; -- cgit v1.3-8-gc7d7 From 0a70bd43053331d99881211e1d09f32de531432f Mon Sep 17 00:00:00 2001 From: Dan Williams Date: Wed, 24 Feb 2016 14:02:11 -0800 Subject: dax: enable dax in the presence of known media errors (badblocks) 1/ If a mapping overlaps a bad sector fail the request. 2/ Do not opportunistically report more dax-capable capacity than is requested when errors present. Reviewed-by: Jeff Moyer Reviewed-by: Christoph Hellwig Signed-off-by: Dan Williams [vishal: fix a conflict with system RAM collision patches] [vishal: add a 'size' parameter to ->direct_access] [vishal: fix a conflict with DAX alignment check patches] Signed-off-by: Vishal Verma --- arch/powerpc/sysdev/axonram.c | 2 +- drivers/block/brd.c | 2 +- drivers/nvdimm/pmem.c | 10 +++++++++- drivers/s390/block/dcssblk.c | 4 ++-- fs/block_dev.c | 13 +------------ include/linux/blkdev.h | 2 +- 6 files changed, 15 insertions(+), 18 deletions(-) (limited to 'fs') diff --git a/arch/powerpc/sysdev/axonram.c b/arch/powerpc/sysdev/axonram.c index 0d112b94d91d..ff75d70f7285 100644 --- a/arch/powerpc/sysdev/axonram.c +++ b/arch/powerpc/sysdev/axonram.c @@ -143,7 +143,7 @@ axon_ram_make_request(struct request_queue *queue, struct bio *bio) */ static long axon_ram_direct_access(struct block_device *device, sector_t sector, - void __pmem **kaddr, pfn_t *pfn) + void __pmem **kaddr, pfn_t *pfn, long size) { struct axon_ram_bank *bank = device->bd_disk->private_data; loff_t offset = (loff_t)sector << AXON_RAM_SECTOR_SHIFT; diff --git a/drivers/block/brd.c b/drivers/block/brd.c index 51a071e32221..c04bd9bc39fd 100644 --- a/drivers/block/brd.c +++ b/drivers/block/brd.c @@ -381,7 +381,7 @@ static int brd_rw_page(struct block_device *bdev, sector_t sector, #ifdef CONFIG_BLK_DEV_RAM_DAX static long brd_direct_access(struct block_device *bdev, sector_t sector, - void __pmem **kaddr, pfn_t *pfn) + void __pmem **kaddr, pfn_t *pfn, long size) { struct brd_device *brd = bdev->bd_disk->private_data; struct page *page; diff --git a/drivers/nvdimm/pmem.c b/drivers/nvdimm/pmem.c index f798899338ed..c447579bd853 100644 --- a/drivers/nvdimm/pmem.c +++ b/drivers/nvdimm/pmem.c @@ -182,14 +182,22 @@ static int pmem_rw_page(struct block_device *bdev, sector_t sector, } static long pmem_direct_access(struct block_device *bdev, sector_t sector, - void __pmem **kaddr, pfn_t *pfn) + void __pmem **kaddr, pfn_t *pfn, long size) { struct pmem_device *pmem = bdev->bd_disk->private_data; resource_size_t offset = sector * 512 + pmem->data_offset; + if (unlikely(is_bad_pmem(&pmem->bb, sector, size))) + return -EIO; *kaddr = pmem->virt_addr + offset; *pfn = phys_to_pfn_t(pmem->phys_addr + offset, pmem->pfn_flags); + /* + * If badblocks are present, limit known good range to the + * requested range. + */ + if (unlikely(pmem->bb.count)) + return size; return pmem->size - pmem->pfn_pad - offset; } diff --git a/drivers/s390/block/dcssblk.c b/drivers/s390/block/dcssblk.c index 1bce9cf51b1e..6ac33984bc0f 100644 --- a/drivers/s390/block/dcssblk.c +++ b/drivers/s390/block/dcssblk.c @@ -31,7 +31,7 @@ static void dcssblk_release(struct gendisk *disk, fmode_t mode); static blk_qc_t dcssblk_make_request(struct request_queue *q, struct bio *bio); static long dcssblk_direct_access(struct block_device *bdev, sector_t secnum, - void __pmem **kaddr, pfn_t *pfn); + void __pmem **kaddr, pfn_t *pfn, long size); static char dcssblk_segments[DCSSBLK_PARM_LEN] = "\0"; @@ -883,7 +883,7 @@ fail: static long dcssblk_direct_access (struct block_device *bdev, sector_t secnum, - void __pmem **kaddr, pfn_t *pfn) + void __pmem **kaddr, pfn_t *pfn, long size) { struct dcssblk_dev_info *dev_info; unsigned long offset, dev_sz; diff --git a/fs/block_dev.c b/fs/block_dev.c index 8477d4501b1e..45839b27972c 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -29,7 +29,6 @@ #include #include #include -#include #include #include "internal.h" @@ -501,7 +500,7 @@ long bdev_direct_access(struct block_device *bdev, struct blk_dax_ctl *dax) sector += get_start_sect(bdev); if (sector % (PAGE_SIZE / 512)) return -EINVAL; - avail = ops->direct_access(bdev, sector, &dax->addr, &dax->pfn); + avail = ops->direct_access(bdev, sector, &dax->addr, &dax->pfn, size); if (!avail) return -ERANGE; if (avail > 0 && avail & ~PAGE_MASK) @@ -561,7 +560,6 @@ EXPORT_SYMBOL_GPL(bdev_dax_supported); */ bool bdev_dax_capable(struct block_device *bdev) { - struct gendisk *disk = bdev->bd_disk; struct blk_dax_ctl dax = { .size = PAGE_SIZE, }; @@ -577,15 +575,6 @@ bool bdev_dax_capable(struct block_device *bdev) if (bdev_direct_access(bdev, &dax) < 0) return false; - /* - * If the device has known bad blocks, force all I/O through the - * driver / page cache. - * - * TODO: support finer grained dax error handling - */ - if (disk->bb && disk->bb->count) - return false; - return true; } diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h index 27cbefe8c985..cf7c13c2c38d 100644 --- a/include/linux/blkdev.h +++ b/include/linux/blkdev.h @@ -1668,7 +1668,7 @@ struct block_device_operations { int (*ioctl) (struct block_device *, fmode_t, unsigned, unsigned long); int (*compat_ioctl) (struct block_device *, fmode_t, unsigned, unsigned long); long (*direct_access)(struct block_device *, sector_t, void __pmem **, - pfn_t *); + pfn_t *, long); unsigned int (*check_events) (struct gendisk *disk, unsigned int clearing); /* ->media_changed() is DEPRECATED, use ->check_events() instead */ -- cgit v1.3-8-gc7d7 From 3dc29161070ab14d065554c0ad58988ab77a7bfd Mon Sep 17 00:00:00 2001 From: Matthew Wilcox Date: Tue, 15 Mar 2016 11:20:41 -0600 Subject: dax: use sb_issue_zerout instead of calling dax_clear_sectors dax_clear_sectors() cannot handle poisoned blocks. These must be zeroed using the BIO interface instead. Convert ext2 and XFS to use only sb_issue_zerout(). Reviewed-by: Jeff Moyer Reviewed-by: Christoph Hellwig Reviewed-by: Jan Kara Signed-off-by: Matthew Wilcox [vishal: Also remove the dax_clear_sectors function entirely] Signed-off-by: Vishal Verma --- fs/dax.c | 32 -------------------------------- fs/ext2/inode.c | 8 ++++---- fs/xfs/xfs_bmap_util.c | 15 ++++----------- include/linux/dax.h | 1 - 4 files changed, 8 insertions(+), 48 deletions(-) (limited to 'fs') diff --git a/fs/dax.c b/fs/dax.c index d602410d8e52..0abbbb62981e 100644 --- a/fs/dax.c +++ b/fs/dax.c @@ -87,38 +87,6 @@ struct page *read_dax_sector(struct block_device *bdev, sector_t n) return page; } -/* - * dax_clear_sectors() is called from within transaction context from XFS, - * and hence this means the stack from this point must follow GFP_NOFS - * semantics for all operations. - */ -int dax_clear_sectors(struct block_device *bdev, sector_t _sector, long _size) -{ - struct blk_dax_ctl dax = { - .sector = _sector, - .size = _size, - }; - - might_sleep(); - do { - long count, sz; - - count = dax_map_atomic(bdev, &dax); - if (count < 0) - return count; - sz = min_t(long, count, SZ_128K); - clear_pmem(dax.addr, sz); - dax.size -= sz; - dax.sector += sz / 512; - dax_unmap_atomic(bdev, &dax); - cond_resched(); - } while (dax.size); - - wmb_pmem(); - return 0; -} -EXPORT_SYMBOL_GPL(dax_clear_sectors); - static bool buffer_written(struct buffer_head *bh) { return buffer_mapped(bh) && !buffer_unwritten(bh); diff --git a/fs/ext2/inode.c b/fs/ext2/inode.c index 9a14af3b1a69..17cbd6b696f2 100644 --- a/fs/ext2/inode.c +++ b/fs/ext2/inode.c @@ -26,6 +26,7 @@ #include #include #include +#include #include #include #include @@ -737,10 +738,9 @@ static int ext2_get_blocks(struct inode *inode, * so that it's not found by another thread before it's * initialised */ - err = dax_clear_sectors(inode->i_sb->s_bdev, - le32_to_cpu(chain[depth-1].key) << - (inode->i_blkbits - 9), - count << inode->i_blkbits); + err = sb_issue_zeroout(inode->i_sb, + le32_to_cpu(chain[depth-1].key), count, + GFP_NOFS); if (err) { mutex_unlock(&ei->truncate_mutex); goto cleanup; diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c index 3b6309865c65..930ac6a17ce3 100644 --- a/fs/xfs/xfs_bmap_util.c +++ b/fs/xfs/xfs_bmap_util.c @@ -72,18 +72,11 @@ xfs_zero_extent( struct xfs_mount *mp = ip->i_mount; xfs_daddr_t sector = xfs_fsb_to_db(ip, start_fsb); sector_t block = XFS_BB_TO_FSBT(mp, sector); - ssize_t size = XFS_FSB_TO_B(mp, count_fsb); - - if (IS_DAX(VFS_I(ip))) - return dax_clear_sectors(xfs_find_bdev_for_inode(VFS_I(ip)), - sector, size); - - /* - * let the block layer decide on the fastest method of - * implementing the zeroing. - */ - return sb_issue_zeroout(mp->m_super, block, count_fsb, GFP_NOFS); + return blkdev_issue_zeroout(xfs_find_bdev_for_inode(VFS_I(ip)), + block << (mp->m_super->s_blocksize_bits - 9), + count_fsb << (mp->m_super->s_blocksize_bits - 9), + GFP_NOFS, true); } /* diff --git a/include/linux/dax.h b/include/linux/dax.h index 7c45ac7ea1d1..7f853ffaa987 100644 --- a/include/linux/dax.h +++ b/include/linux/dax.h @@ -7,7 +7,6 @@ ssize_t dax_do_io(struct kiocb *, struct inode *, struct iov_iter *, loff_t, get_block_t, dio_iodone_t, int flags); -int dax_clear_sectors(struct block_device *bdev, sector_t _sector, long _size); int dax_zero_page_range(struct inode *, loff_t from, unsigned len, get_block_t); int dax_truncate_page(struct inode *, loff_t from, get_block_t); int dax_fault(struct vm_area_struct *, struct vm_fault *, get_block_t); -- cgit v1.3-8-gc7d7 From 679c8bd3b29428e736eabb7fc66a978f312f0c86 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 9 May 2016 10:47:04 +0200 Subject: dax: export a low-level __dax_zero_page_range helper This allows XFS to perform zeroing using the iomap infrastructure and avoid buffer heads. Reviewed-by: Jan Kara Signed-off-by: Christoph Hellwig [vishal: fix conflicts with dax-error-handling] Signed-off-by: Vishal Verma --- fs/dax.c | 35 ++++++++++++++++++++--------------- include/linux/dax.h | 7 +++++++ 2 files changed, 27 insertions(+), 15 deletions(-) (limited to 'fs') diff --git a/fs/dax.c b/fs/dax.c index 0abbbb62981e..651d4b18ac29 100644 --- a/fs/dax.c +++ b/fs/dax.c @@ -947,6 +947,23 @@ int dax_pfn_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) } EXPORT_SYMBOL_GPL(dax_pfn_mkwrite); +int __dax_zero_page_range(struct block_device *bdev, sector_t sector, + unsigned int offset, unsigned int length) +{ + struct blk_dax_ctl dax = { + .sector = sector, + .size = PAGE_SIZE, + }; + + if (dax_map_atomic(bdev, &dax) < 0) + return PTR_ERR(dax.addr); + clear_pmem(dax.addr + offset, length); + wmb_pmem(); + dax_unmap_atomic(bdev, &dax); + return 0; +} +EXPORT_SYMBOL_GPL(__dax_zero_page_range); + /** * dax_zero_page_range - zero a range within a page of a DAX file * @inode: The file being truncated @@ -982,23 +999,11 @@ int dax_zero_page_range(struct inode *inode, loff_t from, unsigned length, bh.b_bdev = inode->i_sb->s_bdev; bh.b_size = PAGE_SIZE; err = get_block(inode, index, &bh, 0); - if (err < 0) + if (err < 0 || !buffer_written(&bh)) return err; - if (buffer_written(&bh)) { - struct block_device *bdev = bh.b_bdev; - struct blk_dax_ctl dax = { - .sector = to_sector(&bh, inode), - .size = PAGE_SIZE, - }; - if (dax_map_atomic(bdev, &dax) < 0) - return PTR_ERR(dax.addr); - clear_pmem(dax.addr + offset, length); - wmb_pmem(); - dax_unmap_atomic(bdev, &dax); - } - - return 0; + return __dax_zero_page_range(bh.b_bdev, to_sector(&bh, inode), + offset, length); } EXPORT_SYMBOL_GPL(dax_zero_page_range); diff --git a/include/linux/dax.h b/include/linux/dax.h index 7f853ffaa987..90fbc99e5313 100644 --- a/include/linux/dax.h +++ b/include/linux/dax.h @@ -14,12 +14,19 @@ int __dax_fault(struct vm_area_struct *, struct vm_fault *, get_block_t); #ifdef CONFIG_FS_DAX struct page *read_dax_sector(struct block_device *bdev, sector_t n); +int __dax_zero_page_range(struct block_device *bdev, sector_t sector, + unsigned int offset, unsigned int length); #else static inline struct page *read_dax_sector(struct block_device *bdev, sector_t n) { return ERR_PTR(-ENXIO); } +static inline int __dax_zero_page_range(struct block_device *bdev, + sector_t sector, unsigned int offset, unsigned int length) +{ + return -ENXIO; +} #endif #ifdef CONFIG_TRANSPARENT_HUGEPAGE -- cgit v1.3-8-gc7d7 From 4b0228fa1d753f77fe0e6cf4c41398ec77dfbd2a Mon Sep 17 00:00:00 2001 From: Vishal Verma Date: Thu, 21 Apr 2016 15:13:46 -0400 Subject: dax: for truncate/hole-punch, do zeroing through the driver if possible In the truncate or hole-punch path in dax, we clear out sub-page ranges. If these sub-page ranges are sector aligned and sized, we can do the zeroing through the driver instead so that error-clearing is handled automatically. For sub-sector ranges, we still have to rely on clear_pmem and have the possibility of tripping over errors. Cc: Dan Williams Cc: Ross Zwisler Cc: Jeff Moyer Cc: Christoph Hellwig Cc: Dave Chinner Cc: Jan Kara Reviewed-by: Christoph Hellwig Reviewed-by: Jan Kara Signed-off-by: Vishal Verma --- Documentation/filesystems/dax.txt | 32 ++++++++++++++++++++++++++++++++ fs/dax.c | 30 +++++++++++++++++++++++++----- 2 files changed, 57 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/Documentation/filesystems/dax.txt b/Documentation/filesystems/dax.txt index 7bde64014a89..ce4587d257d2 100644 --- a/Documentation/filesystems/dax.txt +++ b/Documentation/filesystems/dax.txt @@ -79,6 +79,38 @@ These filesystems may be used for inspiration: - ext4: the fourth extended filesystem, see Documentation/filesystems/ext4.txt +Handling Media Errors +--------------------- + +The libnvdimm subsystem stores a record of known media error locations for +each pmem block device (in gendisk->badblocks). If we fault at such location, +or one with a latent error not yet discovered, the application can expect +to receive a SIGBUS. Libnvdimm also allows clearing of these errors by simply +writing the affected sectors (through the pmem driver, and if the underlying +NVDIMM supports the clear_poison DSM defined by ACPI). + +Since DAX IO normally doesn't go through the driver/bio path, applications or +sysadmins have an option to restore the lost data from a prior backup/inbuilt +redundancy in the following ways: + +1. Delete the affected file, and restore from a backup (sysadmin route): + This will free the file system blocks that were being used by the file, + and the next time they're allocated, they will be zeroed first, which + happens through the driver, and will clear bad sectors. + +2. Truncate or hole-punch the part of the file that has a bad-block (at least + an entire aligned sector has to be hole-punched, but not necessarily an + entire filesystem block). + +These are the two basic paths that allow DAX filesystems to continue operating +in the presence of media errors. More robust error recovery mechanisms can be +built on top of this in the future, for example, involving redundancy/mirroring +provided at the block layer through DM, or additionally, at the filesystem +level. These would have to rely on the above two tenets, that error clearing +can happen either by sending an IO through the driver, or zeroing (also through +the driver). + + Shortcomings ------------ diff --git a/fs/dax.c b/fs/dax.c index 651d4b18ac29..0b9a16934017 100644 --- a/fs/dax.c +++ b/fs/dax.c @@ -947,6 +947,19 @@ int dax_pfn_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) } EXPORT_SYMBOL_GPL(dax_pfn_mkwrite); +static bool dax_range_is_aligned(struct block_device *bdev, + unsigned int offset, unsigned int length) +{ + unsigned short sector_size = bdev_logical_block_size(bdev); + + if (!IS_ALIGNED(offset, sector_size)) + return false; + if (!IS_ALIGNED(length, sector_size)) + return false; + + return true; +} + int __dax_zero_page_range(struct block_device *bdev, sector_t sector, unsigned int offset, unsigned int length) { @@ -955,11 +968,18 @@ int __dax_zero_page_range(struct block_device *bdev, sector_t sector, .size = PAGE_SIZE, }; - if (dax_map_atomic(bdev, &dax) < 0) - return PTR_ERR(dax.addr); - clear_pmem(dax.addr + offset, length); - wmb_pmem(); - dax_unmap_atomic(bdev, &dax); + if (dax_range_is_aligned(bdev, offset, length)) { + sector_t start_sector = dax.sector + (offset >> 9); + + return blkdev_issue_zeroout(bdev, start_sector, + length >> 9, GFP_NOFS, true); + } else { + if (dax_map_atomic(bdev, &dax) < 0) + return PTR_ERR(dax.addr); + clear_pmem(dax.addr + offset, length); + wmb_pmem(); + dax_unmap_atomic(bdev, &dax); + } return 0; } EXPORT_SYMBOL_GPL(__dax_zero_page_range); -- cgit v1.3-8-gc7d7 From 40543f62cbdce42633e3fe10923099feee272e1f Mon Sep 17 00:00:00 2001 From: Vishal Verma Date: Thu, 21 Apr 2016 16:45:56 -0400 Subject: dax: fix a comment in dax_zero_page_range and dax_truncate_page The distinction between PAGE_SIZE and PAGE_CACHE_SIZE was removed in 09cbfea mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros The comments for the above functions described a distinction between those, that is now redundant, so remove those paragraphs Cc: Kirill A. Shutemov Reviewed-by: Christoph Hellwig Reviewed-by: Jan Kara Signed-off-by: Vishal Verma --- fs/dax.c | 12 ------------ 1 file changed, 12 deletions(-) (limited to 'fs') diff --git a/fs/dax.c b/fs/dax.c index 0b9a16934017..ea936b3d93dc 100644 --- a/fs/dax.c +++ b/fs/dax.c @@ -995,12 +995,6 @@ EXPORT_SYMBOL_GPL(__dax_zero_page_range); * page in a DAX file. This is intended for hole-punch operations. If * you are truncating a file, the helper function dax_truncate_page() may be * more convenient. - * - * We work in terms of PAGE_SIZE here for commonality with - * block_truncate_page(), but we could go down to PAGE_SIZE if the filesystem - * took care of disposing of the unnecessary blocks. Even if the filesystem - * block size is smaller than PAGE_SIZE, we have to zero the rest of the page - * since the file might be mmapped. */ int dax_zero_page_range(struct inode *inode, loff_t from, unsigned length, get_block_t get_block) @@ -1035,12 +1029,6 @@ EXPORT_SYMBOL_GPL(dax_zero_page_range); * * Similar to block_truncate_page(), this function can be called by a * filesystem when it is truncating a DAX file to handle the partial page. - * - * We work in terms of PAGE_SIZE here for commonality with - * block_truncate_page(), but we could go down to PAGE_SIZE if the filesystem - * took care of disposing of the unnecessary blocks. Even if the filesystem - * block size is smaller than PAGE_SIZE, we have to zero the rest of the page - * since the file might be mmapped. */ int dax_truncate_page(struct inode *inode, loff_t from, get_block_t get_block) { -- cgit v1.3-8-gc7d7 From f573018491fd823e909d587cfe16758f3dd9e6d6 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Tue, 17 May 2016 16:23:36 -0700 Subject: f2fs: use bio count instead of F2FS_WRITEBACK page count This can reduce page counting overhead. Signed-off-by: Jaegeuk Kim --- fs/f2fs/checkpoint.c | 2 +- fs/f2fs/data.c | 26 +++++++++++++++----------- fs/f2fs/debug.c | 6 +++--- fs/f2fs/f2fs.h | 4 ++-- fs/f2fs/super.c | 3 +-- 5 files changed, 22 insertions(+), 19 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c index d04113b9ac18..447e2a986560 100644 --- a/fs/f2fs/checkpoint.c +++ b/fs/f2fs/checkpoint.c @@ -914,7 +914,7 @@ static void wait_on_all_pages_writeback(struct f2fs_sb_info *sbi) for (;;) { prepare_to_wait(&sbi->cp_wait, &wait, TASK_UNINTERRUPTIBLE); - if (!get_pages(sbi, F2FS_WRITEBACK)) + if (!atomic_read(&sbi->nr_wb_bios)) break; io_schedule_timeout(5*HZ); diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index 1013836c5b6b..faef6667ba45 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -71,10 +71,9 @@ static void f2fs_write_end_io(struct bio *bio) f2fs_stop_checkpoint(sbi); } end_page_writeback(page); - dec_page_count(sbi, F2FS_WRITEBACK); } - - if (!get_pages(sbi, F2FS_WRITEBACK) && wq_has_sleeper(&sbi->cp_wait)) + if (atomic_dec_and_test(&sbi->nr_wb_bios) && + wq_has_sleeper(&sbi->cp_wait)) wake_up(&sbi->cp_wait); bio_put(bio); @@ -98,6 +97,14 @@ static struct bio *__bio_alloc(struct f2fs_sb_info *sbi, block_t blk_addr, return bio; } +static inline void __submit_bio(struct f2fs_sb_info *sbi, int rw, + struct bio *bio) +{ + if (!is_read_io(rw)) + atomic_inc(&sbi->nr_wb_bios); + submit_bio(rw, bio); +} + static void __submit_merged_bio(struct f2fs_bio_info *io) { struct f2fs_io_info *fio = &io->fio; @@ -110,7 +117,7 @@ static void __submit_merged_bio(struct f2fs_bio_info *io) else trace_f2fs_submit_write_bio(io->sbi->sb, fio, io->bio); - submit_bio(fio->rw, io->bio); + __submit_bio(io->sbi, fio->rw, io->bio); io->bio = NULL; } @@ -228,7 +235,7 @@ int f2fs_submit_page_bio(struct f2fs_io_info *fio) return -EFAULT; } - submit_bio(fio->rw, bio); + __submit_bio(fio->sbi, fio->rw, bio); return 0; } @@ -248,9 +255,6 @@ void f2fs_submit_page_mbio(struct f2fs_io_info *fio) down_write(&io->io_rwsem); - if (!is_read) - inc_page_count(sbi, F2FS_WRITEBACK); - if (io->bio && (io->last_block_in_bio != fio->new_blkaddr - 1 || io->fio.rw != fio->rw)) __submit_merged_bio(io); @@ -1047,7 +1051,7 @@ got_it: */ if (bio && (last_block_in_bio != block_nr - 1)) { submit_and_realloc: - submit_bio(READ, bio); + __submit_bio(F2FS_I_SB(inode), READ, bio); bio = NULL; } if (bio == NULL) { @@ -1090,7 +1094,7 @@ set_error_page: goto next_page; confused: if (bio) { - submit_bio(READ, bio); + __submit_bio(F2FS_I_SB(inode), READ, bio); bio = NULL; } unlock_page(page); @@ -1100,7 +1104,7 @@ next_page: } BUG_ON(pages && !list_empty(pages)); if (bio) - submit_bio(READ, bio); + __submit_bio(F2FS_I_SB(inode), READ, bio); return 0; } diff --git a/fs/f2fs/debug.c b/fs/f2fs/debug.c index 37615b254581..a18897319347 100644 --- a/fs/f2fs/debug.c +++ b/fs/f2fs/debug.c @@ -48,7 +48,7 @@ static void update_general_status(struct f2fs_sb_info *sbi) si->ndirty_dirs = sbi->ndirty_inode[DIR_INODE]; si->ndirty_files = sbi->ndirty_inode[FILE_INODE]; si->inmem_pages = get_pages(sbi, F2FS_INMEM_PAGES); - si->wb_pages = get_pages(sbi, F2FS_WRITEBACK); + si->wb_bios = atomic_read(&sbi->nr_wb_bios); si->total_count = (int)sbi->user_block_count / sbi->blocks_per_seg; si->rsvd_segs = reserved_segments(sbi); si->overp_segs = overprovision_segments(sbi); @@ -299,8 +299,8 @@ static int stat_show(struct seq_file *s, void *v) seq_printf(s, " - Inner Struct Count: tree: %d(%d), node: %d\n", si->ext_tree, si->zombie_tree, si->ext_node); seq_puts(s, "\nBalancing F2FS Async:\n"); - seq_printf(s, " - inmem: %4d, wb: %4d\n", - si->inmem_pages, si->wb_pages); + seq_printf(s, " - inmem: %4d, wb_bios: %4d\n", + si->inmem_pages, si->wb_bios); seq_printf(s, " - nodes: %4d in %4d\n", si->ndirty_node, si->node_pages); seq_printf(s, " - dents: %4d in dirs:%4d\n", diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 1351178b2b58..bc45a2c61b61 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -650,7 +650,6 @@ struct f2fs_sm_info { * dirty dentry blocks, dirty node blocks, and dirty meta blocks. */ enum count_type { - F2FS_WRITEBACK, F2FS_DIRTY_DENTS, F2FS_DIRTY_DATA, F2FS_DIRTY_NODES, @@ -813,6 +812,7 @@ struct f2fs_sb_info { block_t discard_blks; /* discard command candidats */ block_t last_valid_block_count; /* for recovery */ u32 s_next_generation; /* for NFS support */ + atomic_t nr_wb_bios; /* # of writeback bios */ atomic_t nr_pages[NR_COUNT_TYPE]; /* # of pages, see count_type */ struct f2fs_mount_info mount_opt; /* mount options */ @@ -2017,7 +2017,7 @@ struct f2fs_stat_info { int ndirty_dent, ndirty_dirs, ndirty_data, ndirty_files; int nats, dirty_nats, sits, dirty_sits, fnids; int total_count, utilization; - int bg_gc, inmem_pages, wb_pages; + int bg_gc, inmem_pages, wb_bios; int inline_xattr, inline_inode, inline_dir, orphans; unsigned int valid_count, valid_node_count, valid_inode_count; unsigned int bimodal, avg_vblocks; diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 9df6d72ff1a1..0e175609cf72 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -649,8 +649,7 @@ static void f2fs_put_super(struct super_block *sb) mutex_unlock(&sbi->umount_mutex); /* our cp_error case, we can wait for any writeback page */ - if (get_pages(sbi, F2FS_WRITEBACK)) - f2fs_flush_merged_bios(sbi); + f2fs_flush_merged_bios(sbi); iput(sbi->node_inode); iput(sbi->meta_inode); -- cgit v1.3-8-gc7d7 From 523be8a6b3418eb7e0f0f042fe0490345eb5d516 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Fri, 13 May 2016 12:36:58 -0700 Subject: f2fs: use percpu_counter for page counters This patch substitutes percpu_counter for atomic_counter when counting various types of pages. Signed-off-by: Jaegeuk Kim --- fs/f2fs/debug.c | 11 ++++++----- fs/f2fs/f2fs.h | 24 ++++++++++++++---------- fs/f2fs/super.c | 31 +++++++++++++++++++++++++++---- include/trace/events/f2fs.h | 10 +++++----- 4 files changed, 52 insertions(+), 24 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/debug.c b/fs/f2fs/debug.c index a18897319347..d89a425055d0 100644 --- a/fs/f2fs/debug.c +++ b/fs/f2fs/debug.c @@ -144,6 +144,7 @@ static void update_mem_info(struct f2fs_sb_info *sbi) si->base_mem = sizeof(struct f2fs_sb_info) + sbi->sb->s_blocksize; si->base_mem += 2 * sizeof(struct f2fs_inode_info); si->base_mem += sizeof(*sbi->ckpt); + si->base_mem += sizeof(struct percpu_counter) * NR_COUNT_TYPE; /* build sm */ si->base_mem += sizeof(struct f2fs_sm_info); @@ -299,15 +300,15 @@ static int stat_show(struct seq_file *s, void *v) seq_printf(s, " - Inner Struct Count: tree: %d(%d), node: %d\n", si->ext_tree, si->zombie_tree, si->ext_node); seq_puts(s, "\nBalancing F2FS Async:\n"); - seq_printf(s, " - inmem: %4d, wb_bios: %4d\n", + seq_printf(s, " - inmem: %4lld, wb_bios: %4d\n", si->inmem_pages, si->wb_bios); - seq_printf(s, " - nodes: %4d in %4d\n", + seq_printf(s, " - nodes: %4lld in %4d\n", si->ndirty_node, si->node_pages); - seq_printf(s, " - dents: %4d in dirs:%4d\n", + seq_printf(s, " - dents: %4lld in dirs:%4d\n", si->ndirty_dent, si->ndirty_dirs); - seq_printf(s, " - datas: %4d in files:%4d\n", + seq_printf(s, " - datas: %4lld in files:%4d\n", si->ndirty_data, si->ndirty_files); - seq_printf(s, " - meta: %4d in %4d\n", + seq_printf(s, " - meta: %4lld in %4d\n", si->ndirty_meta, si->meta_pages); seq_printf(s, " - NATs: %9d/%9d\n - SITs: %9d/%9d\n", si->dirty_nats, si->nats, si->dirty_sits, si->sits); diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index bc45a2c61b61..c2d0b24ba6b8 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -813,7 +813,9 @@ struct f2fs_sb_info { block_t last_valid_block_count; /* for recovery */ u32 s_next_generation; /* for NFS support */ atomic_t nr_wb_bios; /* # of writeback bios */ - atomic_t nr_pages[NR_COUNT_TYPE]; /* # of pages, see count_type */ + + /* # of pages, see count_type */ + struct percpu_counter nr_pages[NR_COUNT_TYPE]; struct f2fs_mount_info mount_opt; /* mount options */ @@ -1158,7 +1160,7 @@ static inline void dec_valid_block_count(struct f2fs_sb_info *sbi, static inline void inc_page_count(struct f2fs_sb_info *sbi, int count_type) { - atomic_inc(&sbi->nr_pages[count_type]); + percpu_counter_inc(&sbi->nr_pages[count_type]); set_sbi_flag(sbi, SBI_IS_DIRTY); } @@ -1171,7 +1173,7 @@ static inline void inode_inc_dirty_pages(struct inode *inode) static inline void dec_page_count(struct f2fs_sb_info *sbi, int count_type) { - atomic_dec(&sbi->nr_pages[count_type]); + percpu_counter_dec(&sbi->nr_pages[count_type]); } static inline void inode_dec_dirty_pages(struct inode *inode) @@ -1185,9 +1187,9 @@ static inline void inode_dec_dirty_pages(struct inode *inode) F2FS_DIRTY_DENTS : F2FS_DIRTY_DATA); } -static inline int get_pages(struct f2fs_sb_info *sbi, int count_type) +static inline s64 get_pages(struct f2fs_sb_info *sbi, int count_type) { - return atomic_read(&sbi->nr_pages[count_type]); + return percpu_counter_sum_positive(&sbi->nr_pages[count_type]); } static inline int get_dirty_pages(struct inode *inode) @@ -1198,8 +1200,10 @@ static inline int get_dirty_pages(struct inode *inode) static inline int get_blocktype_secs(struct f2fs_sb_info *sbi, int block_type) { unsigned int pages_per_sec = sbi->segs_per_sec * sbi->blocks_per_seg; - return ((get_pages(sbi, block_type) + pages_per_sec - 1) - >> sbi->log_blocks_per_seg) / sbi->segs_per_sec; + unsigned int segs = (get_pages(sbi, block_type) + pages_per_sec - 1) >> + sbi->log_blocks_per_seg; + + return segs / sbi->segs_per_sec; } static inline block_t valid_user_blocks(struct f2fs_sb_info *sbi) @@ -2013,11 +2017,11 @@ struct f2fs_stat_info { unsigned long long hit_largest, hit_cached, hit_rbtree; unsigned long long hit_total, total_ext; int ext_tree, zombie_tree, ext_node; - int ndirty_node, ndirty_meta; - int ndirty_dent, ndirty_dirs, ndirty_data, ndirty_files; + s64 ndirty_node, ndirty_dent, ndirty_meta, ndirty_data, inmem_pages; + unsigned int ndirty_dirs, ndirty_files; int nats, dirty_nats, sits, dirty_sits, fnids; int total_count, utilization; - int bg_gc, inmem_pages, wb_bios; + int bg_gc, wb_bios; int inline_xattr, inline_inode, inline_dir, orphans; unsigned int valid_count, valid_node_count, valid_inode_count; unsigned int bimodal, avg_vblocks; diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 0e175609cf72..009a9f44205f 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -606,6 +606,14 @@ static void f2fs_destroy_inode(struct inode *inode) call_rcu(&inode->i_rcu, f2fs_i_callback); } +static void destroy_percpu_info(struct f2fs_sb_info *sbi) +{ + int i; + + for (i = 0; i < NR_COUNT_TYPE; i++) + percpu_counter_destroy(&sbi->nr_pages[i]); +} + static void f2fs_put_super(struct super_block *sb) { struct f2fs_sb_info *sbi = F2FS_SB(sb); @@ -666,6 +674,8 @@ static void f2fs_put_super(struct super_block *sb) if (sbi->s_chksum_driver) crypto_free_shash(sbi->s_chksum_driver); kfree(sbi->raw_super); + + destroy_percpu_info(sbi); kfree(sbi); } @@ -1324,7 +1334,6 @@ int sanity_check_ckpt(struct f2fs_sb_info *sbi) static void init_sb_info(struct f2fs_sb_info *sbi) { struct f2fs_super_block *raw_super = sbi->raw_super; - int i; sbi->log_sectors_per_block = le32_to_cpu(raw_super->log_sectors_per_block); @@ -1344,9 +1353,6 @@ static void init_sb_info(struct f2fs_sb_info *sbi) sbi->cur_victim_sec = NULL_SECNO; sbi->max_victim_search = DEF_MAX_VICTIM_SEARCH; - for (i = 0; i < NR_COUNT_TYPE; i++) - atomic_set(&sbi->nr_pages[i], 0); - sbi->dir_level = DEF_DIR_LEVEL; sbi->interval_time[CP_TIME] = DEF_CP_INTERVAL; sbi->interval_time[REQ_TIME] = DEF_IDLE_INTERVAL; @@ -1362,6 +1368,18 @@ static void init_sb_info(struct f2fs_sb_info *sbi) #endif } +static int init_percpu_info(struct f2fs_sb_info *sbi) +{ + int i, err; + + for (i = 0; i < NR_COUNT_TYPE; i++) { + err = percpu_counter_init(&sbi->nr_pages[i], 0, GFP_KERNEL); + if (err) + return err; + } + return 0; +} + /* * Read f2fs raw super block. * Because we have two copies of super block, so read both of them @@ -1552,6 +1570,10 @@ try_onemore: init_waitqueue_head(&sbi->cp_wait); init_sb_info(sbi); + err = init_percpu_info(sbi); + if (err) + goto free_options; + /* get an inode for meta space */ sbi->meta_inode = f2fs_iget(sb, F2FS_META_INO(sbi)); if (IS_ERR(sbi->meta_inode)) { @@ -1754,6 +1776,7 @@ free_meta_inode: make_bad_inode(sbi->meta_inode); iput(sbi->meta_inode); free_options: + destroy_percpu_info(sbi); kfree(options); free_sb_buf: kfree(raw_super); diff --git a/include/trace/events/f2fs.h b/include/trace/events/f2fs.h index 497e6e80fb63..3a09bb4dc3b2 100644 --- a/include/trace/events/f2fs.h +++ b/include/trace/events/f2fs.h @@ -1275,14 +1275,14 @@ TRACE_EVENT(f2fs_destroy_extent_tree, DECLARE_EVENT_CLASS(f2fs_sync_dirty_inodes, - TP_PROTO(struct super_block *sb, int type, int count), + TP_PROTO(struct super_block *sb, int type, s64 count), TP_ARGS(sb, type, count), TP_STRUCT__entry( __field(dev_t, dev) __field(int, type) - __field(int, count) + __field(s64, count) ), TP_fast_assign( @@ -1291,7 +1291,7 @@ DECLARE_EVENT_CLASS(f2fs_sync_dirty_inodes, __entry->count = count; ), - TP_printk("dev = (%d,%d), %s, dirty count = %d", + TP_printk("dev = (%d,%d), %s, dirty count = %lld", show_dev(__entry), show_file_type(__entry->type), __entry->count) @@ -1299,14 +1299,14 @@ DECLARE_EVENT_CLASS(f2fs_sync_dirty_inodes, DEFINE_EVENT(f2fs_sync_dirty_inodes, f2fs_sync_dirty_inodes_enter, - TP_PROTO(struct super_block *sb, int type, int count), + TP_PROTO(struct super_block *sb, int type, s64 count), TP_ARGS(sb, type, count) ); DEFINE_EVENT(f2fs_sync_dirty_inodes, f2fs_sync_dirty_inodes_exit, - TP_PROTO(struct super_block *sb, int type, int count), + TP_PROTO(struct super_block *sb, int type, s64 count), TP_ARGS(sb, type, count) ); -- cgit v1.3-8-gc7d7 From 1beba1b3a953107c3ff5448ab4e4297db4619c76 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Fri, 13 May 2016 12:47:11 -0700 Subject: f2fs: use percpu_counter for # of dirty pages in inode This patch adds percpu_counter for # of dirty pages in inode. Signed-off-by: Jaegeuk Kim --- fs/f2fs/f2fs.h | 10 +++++----- fs/f2fs/file.c | 2 +- fs/f2fs/super.c | 7 ++++++- 3 files changed, 12 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index c2d0b24ba6b8..3f1c71092440 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -436,7 +436,7 @@ struct f2fs_inode_info { /* Use below internally in f2fs*/ unsigned long flags; /* use to pass per-file flags */ struct rw_semaphore i_sem; /* protect fi info */ - atomic_t dirty_pages; /* # of dirty pages */ + struct percpu_counter dirty_pages; /* # of dirty pages */ f2fs_hash_t chash; /* hash value of given file name */ unsigned int clevel; /* maximum level of given file name */ nid_t i_xattr_nid; /* node id that contains xattrs */ @@ -1166,7 +1166,7 @@ static inline void inc_page_count(struct f2fs_sb_info *sbi, int count_type) static inline void inode_inc_dirty_pages(struct inode *inode) { - atomic_inc(&F2FS_I(inode)->dirty_pages); + percpu_counter_inc(&F2FS_I(inode)->dirty_pages); inc_page_count(F2FS_I_SB(inode), S_ISDIR(inode->i_mode) ? F2FS_DIRTY_DENTS : F2FS_DIRTY_DATA); } @@ -1182,7 +1182,7 @@ static inline void inode_dec_dirty_pages(struct inode *inode) !S_ISLNK(inode->i_mode)) return; - atomic_dec(&F2FS_I(inode)->dirty_pages); + percpu_counter_dec(&F2FS_I(inode)->dirty_pages); dec_page_count(F2FS_I_SB(inode), S_ISDIR(inode->i_mode) ? F2FS_DIRTY_DENTS : F2FS_DIRTY_DATA); } @@ -1192,9 +1192,9 @@ static inline s64 get_pages(struct f2fs_sb_info *sbi, int count_type) return percpu_counter_sum_positive(&sbi->nr_pages[count_type]); } -static inline int get_dirty_pages(struct inode *inode) +static inline s64 get_dirty_pages(struct inode *inode) { - return atomic_read(&F2FS_I(inode)->dirty_pages); + return percpu_counter_sum_positive(&F2FS_I(inode)->dirty_pages); } static inline int get_blocktype_secs(struct f2fs_sb_info *sbi, int block_type) diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index a9cb3148bda5..69dd7c9cdd6a 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -1419,7 +1419,7 @@ static int f2fs_ioc_start_atomic_write(struct file *filp) goto out; f2fs_msg(F2FS_I_SB(inode)->sb, KERN_WARNING, - "Unexpected flush for atomic writes: ino=%lu, npages=%u", + "Unexpected flush for atomic writes: ino=%lu, npages=%lld", inode->i_ino, get_dirty_pages(inode)); ret = filemap_write_and_wait_range(inode->i_mapping, 0, LLONG_MAX); if (ret) diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 009a9f44205f..0ff862b08fb9 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -526,9 +526,13 @@ static struct inode *f2fs_alloc_inode(struct super_block *sb) init_once((void *) fi); + if (percpu_counter_init(&fi->dirty_pages, 0, GFP_NOFS)) { + kmem_cache_free(f2fs_inode_cachep, fi); + return NULL; + } + /* Initialize f2fs-specific inode info */ fi->vfs_inode.i_version = 1; - atomic_set(&fi->dirty_pages, 0); fi->i_current_depth = 1; fi->i_advise = 0; init_rwsem(&fi->i_sem); @@ -603,6 +607,7 @@ static void f2fs_i_callback(struct rcu_head *head) static void f2fs_destroy_inode(struct inode *inode) { + percpu_counter_destroy(&F2FS_I(inode)->dirty_pages); call_rcu(&inode->i_rcu, f2fs_i_callback); } -- cgit v1.3-8-gc7d7 From 41382ec43255b502321c3c27f1347efeb3279290 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Mon, 16 May 2016 11:06:50 -0700 Subject: f2fs: use percpu_counter for alloc_valid_block_count This patch uses percpu_count for sbi->alloc_valid_block_count. Signed-off-by: Jaegeuk Kim --- fs/f2fs/checkpoint.c | 2 +- fs/f2fs/f2fs.h | 8 +++++--- fs/f2fs/recovery.c | 5 +++-- fs/f2fs/super.c | 7 +++++-- 4 files changed, 14 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c index 447e2a986560..cf79598aec0e 100644 --- a/fs/f2fs/checkpoint.c +++ b/fs/f2fs/checkpoint.c @@ -1079,7 +1079,7 @@ static int do_checkpoint(struct f2fs_sb_info *sbi, struct cp_control *cpc) /* update user_block_counts */ sbi->last_valid_block_count = sbi->total_valid_block_count; - sbi->alloc_valid_block_count = 0; + percpu_counter_set(&sbi->alloc_valid_block_count, 0); /* Here, we only have one bio having CP pack */ sync_meta_pages(sbi, META_FLUSH, LONG_MAX); diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 3f1c71092440..c8833c802ea9 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -808,7 +808,6 @@ struct f2fs_sb_info { block_t user_block_count; /* # of user blocks */ block_t total_valid_block_count; /* # of valid blocks */ - block_t alloc_valid_block_count; /* # of allocated blocks */ block_t discard_blks; /* discard command candidats */ block_t last_valid_block_count; /* for recovery */ u32 s_next_generation; /* for NFS support */ @@ -816,6 +815,8 @@ struct f2fs_sb_info { /* # of pages, see count_type */ struct percpu_counter nr_pages[NR_COUNT_TYPE]; + /* # of allocated blocks */ + struct percpu_counter alloc_valid_block_count; struct f2fs_mount_info mount_opt; /* mount options */ @@ -1141,8 +1142,9 @@ static inline bool inc_valid_block_count(struct f2fs_sb_info *sbi, inode->i_blocks += *count; sbi->total_valid_block_count = sbi->total_valid_block_count + (block_t)(*count); - sbi->alloc_valid_block_count += (block_t)(*count); spin_unlock(&sbi->stat_lock); + + percpu_counter_add(&sbi->alloc_valid_block_count, (*count)); return true; } @@ -1292,11 +1294,11 @@ static inline bool inc_valid_node_count(struct f2fs_sb_info *sbi, if (inode) inode->i_blocks++; - sbi->alloc_valid_block_count++; sbi->total_valid_node_count++; sbi->total_valid_block_count++; spin_unlock(&sbi->stat_lock); + percpu_counter_inc(&sbi->alloc_valid_block_count); return true; } diff --git a/fs/f2fs/recovery.c b/fs/f2fs/recovery.c index 6303b2a38c34..f89b70e72004 100644 --- a/fs/f2fs/recovery.c +++ b/fs/f2fs/recovery.c @@ -49,8 +49,9 @@ static struct kmem_cache *fsync_entry_slab; bool space_for_roll_forward(struct f2fs_sb_info *sbi) { - if (sbi->last_valid_block_count + sbi->alloc_valid_block_count - > sbi->user_block_count) + s64 nalloc = percpu_counter_sum_positive(&sbi->alloc_valid_block_count); + + if (sbi->last_valid_block_count + nalloc > sbi->user_block_count) return false; return true; } diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 0ff862b08fb9..43e3e9c382d8 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -617,6 +617,7 @@ static void destroy_percpu_info(struct f2fs_sb_info *sbi) for (i = 0; i < NR_COUNT_TYPE; i++) percpu_counter_destroy(&sbi->nr_pages[i]); + percpu_counter_destroy(&sbi->alloc_valid_block_count); } static void f2fs_put_super(struct super_block *sb) @@ -1382,7 +1383,9 @@ static int init_percpu_info(struct f2fs_sb_info *sbi) if (err) return err; } - return 0; + + return percpu_counter_init(&sbi->alloc_valid_block_count, 0, + GFP_KERNEL); } /* @@ -1601,7 +1604,7 @@ try_onemore: sbi->total_valid_block_count = le64_to_cpu(sbi->ckpt->valid_block_count); sbi->last_valid_block_count = sbi->total_valid_block_count; - sbi->alloc_valid_block_count = 0; + for (i = 0; i < NR_INODE_TYPE; i++) { INIT_LIST_HEAD(&sbi->inode_list[i]); spin_lock_init(&sbi->inode_lock[i]); -- cgit v1.3-8-gc7d7 From 513c5f3735a9bd0bd8b58b6cdafbad5ef19f2159 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Mon, 16 May 2016 11:42:32 -0700 Subject: f2fs: use percpu_counter for total_valid_inode_count This patch uses percpu_counter to avoid stat_lock. Signed-off-by: Jaegeuk Kim --- fs/f2fs/f2fs.h | 18 +++++++----------- fs/f2fs/super.c | 11 ++++++++--- 2 files changed, 15 insertions(+), 14 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index c8833c802ea9..02f0656cb2ab 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -801,7 +801,6 @@ struct f2fs_sb_info { unsigned int total_sections; /* total section count */ unsigned int total_node_count; /* total node block count */ unsigned int total_valid_node_count; /* valid node block count */ - unsigned int total_valid_inode_count; /* valid inode count */ loff_t max_file_blocks; /* max block index of file */ int active_logs; /* # of active logs */ int dir_level; /* directory level */ @@ -818,6 +817,9 @@ struct f2fs_sb_info { /* # of allocated blocks */ struct percpu_counter alloc_valid_block_count; + /* valid inode count */ + struct percpu_counter total_valid_inode_count; + struct f2fs_mount_info mount_opt; /* mount options */ /* for cleaning operations */ @@ -1325,23 +1327,17 @@ static inline unsigned int valid_node_count(struct f2fs_sb_info *sbi) static inline void inc_valid_inode_count(struct f2fs_sb_info *sbi) { - spin_lock(&sbi->stat_lock); - f2fs_bug_on(sbi, sbi->total_valid_inode_count == sbi->total_node_count); - sbi->total_valid_inode_count++; - spin_unlock(&sbi->stat_lock); + percpu_counter_inc(&sbi->total_valid_inode_count); } static inline void dec_valid_inode_count(struct f2fs_sb_info *sbi) { - spin_lock(&sbi->stat_lock); - f2fs_bug_on(sbi, !sbi->total_valid_inode_count); - sbi->total_valid_inode_count--; - spin_unlock(&sbi->stat_lock); + percpu_counter_dec(&sbi->total_valid_inode_count); } -static inline unsigned int valid_inode_count(struct f2fs_sb_info *sbi) +static inline s64 valid_inode_count(struct f2fs_sb_info *sbi) { - return sbi->total_valid_inode_count; + return percpu_counter_sum_positive(&sbi->total_valid_inode_count); } static inline struct page *f2fs_grab_cache_page(struct address_space *mapping, diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 43e3e9c382d8..173cf3a1c29d 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -618,6 +618,7 @@ static void destroy_percpu_info(struct f2fs_sb_info *sbi) for (i = 0; i < NR_COUNT_TYPE; i++) percpu_counter_destroy(&sbi->nr_pages[i]); percpu_counter_destroy(&sbi->alloc_valid_block_count); + percpu_counter_destroy(&sbi->total_valid_inode_count); } static void f2fs_put_super(struct super_block *sb) @@ -1384,7 +1385,11 @@ static int init_percpu_info(struct f2fs_sb_info *sbi) return err; } - return percpu_counter_init(&sbi->alloc_valid_block_count, 0, + err = percpu_counter_init(&sbi->alloc_valid_block_count, 0, GFP_KERNEL); + if (err) + return err; + + return percpu_counter_init(&sbi->total_valid_inode_count, 0, GFP_KERNEL); } @@ -1598,8 +1603,8 @@ try_onemore: sbi->total_valid_node_count = le32_to_cpu(sbi->ckpt->valid_node_count); - sbi->total_valid_inode_count = - le32_to_cpu(sbi->ckpt->valid_inode_count); + percpu_counter_set(&sbi->total_valid_inode_count, + le32_to_cpu(sbi->ckpt->valid_inode_count)); sbi->user_block_count = le64_to_cpu(sbi->ckpt->user_block_count); sbi->total_valid_block_count = le64_to_cpu(sbi->ckpt->valid_block_count); -- cgit v1.3-8-gc7d7 From b8bef79df709533416a278c4c3aef3577cefa61a Mon Sep 17 00:00:00 2001 From: Tiezhu Yang Date: Wed, 18 May 2016 08:02:25 +0800 Subject: f2fs: make exit_f2fs_fs more clear init_f2fs_fs does: 1) f2fs_build_trace_ios 2) init_inodecache 3) create_node_manager_caches 4) create_segment_manager_caches 5) create_checkpoint_caches 6) create_extent_cache 7) kset_create_and_add 8) kobject_init_and_add 9) register_shrinker 10) register_filesystem 11) f2fs_create_root_stats 12) proc_mkdir exit_f2fs_fs should do cleanup in the reverse order to make the code more clear. Signed-off-by: Tiezhu Yang Signed-off-by: Jaegeuk Kim --- fs/f2fs/super.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 173cf3a1c29d..74cc8520b8b1 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -1926,17 +1926,17 @@ static void __exit exit_f2fs_fs(void) { remove_proc_entry("fs/f2fs", NULL); f2fs_destroy_root_stats(); - unregister_shrinker(&f2fs_shrinker_info); unregister_filesystem(&f2fs_fs_type); + unregister_shrinker(&f2fs_shrinker_info); +#ifdef CONFIG_F2FS_FAULT_INJECTION + kobject_put(&f2fs_fault_inject); +#endif + kset_unregister(f2fs_kset); destroy_extent_cache(); destroy_checkpoint_caches(); destroy_segment_manager_caches(); destroy_node_manager_caches(); destroy_inodecache(); -#ifdef CONFIG_F2FS_FAULT_INJECTION - kobject_put(&f2fs_fault_inject); -#endif - kset_unregister(f2fs_kset); f2fs_destroy_trace_ios(); } -- cgit v1.3-8-gc7d7 From 585d70006f6e30f42e96d56c6c0933671c516c7b Mon Sep 17 00:00:00 2001 From: Alden Tondettar Date: Wed, 18 May 2016 14:09:17 -0700 Subject: udf: Don't BUG on missing metadata partition descriptor Currently, if a metadata partition map is missing its partition descriptor, then udf_get_pblock_meta25() will BUG() out the first time it is called. This is rather drastic for a corrupted filesystem, so just treat this case as an invalid mapping instead. Signed-off-by: Alden Tondettar Signed-off-by: Jan Kara --- fs/udf/partition.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/udf/partition.c b/fs/udf/partition.c index 5f861ed287c3..e4e9e70c37f2 100644 --- a/fs/udf/partition.c +++ b/fs/udf/partition.c @@ -317,8 +317,9 @@ uint32_t udf_get_pblock_meta25(struct super_block *sb, uint32_t block, mdata = &map->s_type_specific.s_metadata; inode = mdata->s_metadata_fe ? : mdata->s_mirror_fe; - /* We shouldn't mount such media... */ - BUG_ON(!inode); + if (!inode) + return 0xFFFFFFFF; + retblk = udf_try_read_meta(inode, block, partition, offset); if (retblk == 0xFFFFFFFF && mdata->s_metadata_fe) { udf_warn(sb, "error reading from METADATA, trying to read from MIRROR\n"); -- cgit v1.3-8-gc7d7 From 3743a03e72b73b6234768bce06d7bf5a57c47285 Mon Sep 17 00:00:00 2001 From: Alden Tondettar Date: Wed, 18 May 2016 14:09:18 -0700 Subject: udf: Use IS_ERR when loading metadata mirror file entry Currently when udf_get_pblock_meta25() fails to map a block using the primary metadata file, it will attempt to load the mirror file entry by calling udf_find_metadata_inode_efe(). That function will return a ERR_PTR if it fails, but the return value is only checked against NULL. Test the return value using IS_ERR() and change it to NULL if needed. Signed-off-by: Alden Tondettar Signed-off-by: Jan Kara --- fs/udf/partition.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'fs') diff --git a/fs/udf/partition.c b/fs/udf/partition.c index e4e9e70c37f2..ca3cde336324 100644 --- a/fs/udf/partition.c +++ b/fs/udf/partition.c @@ -326,6 +326,8 @@ uint32_t udf_get_pblock_meta25(struct super_block *sb, uint32_t block, if (!(mdata->s_flags & MF_MIRROR_FE_LOADED)) { mdata->s_mirror_fe = udf_find_metadata_inode_efe(sb, mdata->s_mirror_file_loc, map->s_partition_num); + if (IS_ERR(mdata->s_mirror_fe)) + mdata->s_mirror_fe = NULL; mdata->s_flags |= MF_MIRROR_FE_LOADED; } -- cgit v1.3-8-gc7d7 From 7888824b0b1c9c3753d2aedf1d00e7a1c20c18af Mon Sep 17 00:00:00 2001 From: Alden Tondettar Date: Wed, 18 May 2016 14:09:19 -0700 Subject: udf: Use correct partition reference number for metadata UDF/OSTA terminology is confusing. Partition Numbers (PNs) are arbitrary 16-bit values, one for each physical partition in the volume. Partition Reference Numbers (PRNs) are indices into the the Partition Map Table and do not necessarily equal the PN of the mapped partition. The current metadata code mistakenly uses the PN instead of the PRN when mapping metadata blocks to physical/sparable blocks. Windows-created UDF 2.5 discs for some reason use large, arbitrary PNs, resulting in mount failure and KASAN read warnings in udf_read_inode(). For example, a NetBSD UDF 2.5 partition might look like this: PRN PN Type --- -- ---- 0 0 Sparable 1 0 Metadata Since PRN == PN, we are fine. But Windows could gives us: PRN PN Type --- ---- ---- 0 8192 Sparable 1 8192 Metadata So udf_read_inode() will start out by checking the partition length in sbi->s_partmaps[8192], which is obviously out of bounds. Fix this by creating a new field (s_phys_partition_ref) in struct udf_meta_data, referencing whatever physical or sparable map has the same partition number as the metadata partition. [JK: Add comment about s_phys_partition_ref, change its name] Signed-off-by: Alden Tondettar Signed-off-by: Jan Kara --- fs/udf/partition.c | 6 ++++-- fs/udf/super.c | 22 ++++++++++++---------- fs/udf/udf_sb.h | 5 +++++ 3 files changed, 21 insertions(+), 12 deletions(-) (limited to 'fs') diff --git a/fs/udf/partition.c b/fs/udf/partition.c index ca3cde336324..888c364b2fe9 100644 --- a/fs/udf/partition.c +++ b/fs/udf/partition.c @@ -295,7 +295,8 @@ static uint32_t udf_try_read_meta(struct inode *inode, uint32_t block, map = &UDF_SB(sb)->s_partmaps[partition]; /* map to sparable/physical partition desc */ phyblock = udf_get_pblock(sb, eloc.logicalBlockNum, - map->s_partition_num, ext_offset + offset); + map->s_type_specific.s_metadata.s_phys_partition_ref, + ext_offset + offset); } brelse(epos.bh); @@ -325,7 +326,8 @@ uint32_t udf_get_pblock_meta25(struct super_block *sb, uint32_t block, udf_warn(sb, "error reading from METADATA, trying to read from MIRROR\n"); if (!(mdata->s_flags & MF_MIRROR_FE_LOADED)) { mdata->s_mirror_fe = udf_find_metadata_inode_efe(sb, - mdata->s_mirror_file_loc, map->s_partition_num); + mdata->s_mirror_file_loc, + mdata->s_phys_partition_ref); if (IS_ERR(mdata->s_mirror_fe)) mdata->s_mirror_fe = NULL; mdata->s_flags |= MF_MIRROR_FE_LOADED; diff --git a/fs/udf/super.c b/fs/udf/super.c index 5e2c8c814e1b..4942549e7dc8 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -951,13 +951,13 @@ out2: } struct inode *udf_find_metadata_inode_efe(struct super_block *sb, - u32 meta_file_loc, u32 partition_num) + u32 meta_file_loc, u32 partition_ref) { struct kernel_lb_addr addr; struct inode *metadata_fe; addr.logicalBlockNum = meta_file_loc; - addr.partitionReferenceNum = partition_num; + addr.partitionReferenceNum = partition_ref; metadata_fe = udf_iget_special(sb, &addr); @@ -974,7 +974,8 @@ struct inode *udf_find_metadata_inode_efe(struct super_block *sb, return metadata_fe; } -static int udf_load_metadata_files(struct super_block *sb, int partition) +static int udf_load_metadata_files(struct super_block *sb, int partition, + int type1_index) { struct udf_sb_info *sbi = UDF_SB(sb); struct udf_part_map *map; @@ -984,20 +985,21 @@ static int udf_load_metadata_files(struct super_block *sb, int partition) map = &sbi->s_partmaps[partition]; mdata = &map->s_type_specific.s_metadata; + mdata->s_phys_partition_ref = type1_index; /* metadata address */ udf_debug("Metadata file location: block = %d part = %d\n", - mdata->s_meta_file_loc, map->s_partition_num); + mdata->s_meta_file_loc, mdata->s_phys_partition_ref); fe = udf_find_metadata_inode_efe(sb, mdata->s_meta_file_loc, - map->s_partition_num); + mdata->s_phys_partition_ref); if (IS_ERR(fe)) { /* mirror file entry */ udf_debug("Mirror metadata file location: block = %d part = %d\n", - mdata->s_mirror_file_loc, map->s_partition_num); + mdata->s_mirror_file_loc, mdata->s_phys_partition_ref); fe = udf_find_metadata_inode_efe(sb, mdata->s_mirror_file_loc, - map->s_partition_num); + mdata->s_phys_partition_ref); if (IS_ERR(fe)) { udf_err(sb, "Both metadata and mirror metadata inode efe can not found\n"); @@ -1015,7 +1017,7 @@ static int udf_load_metadata_files(struct super_block *sb, int partition) */ if (mdata->s_bitmap_file_loc != 0xFFFFFFFF) { addr.logicalBlockNum = mdata->s_bitmap_file_loc; - addr.partitionReferenceNum = map->s_partition_num; + addr.partitionReferenceNum = mdata->s_phys_partition_ref; udf_debug("Bitmap file location: block = %d part = %d\n", addr.logicalBlockNum, addr.partitionReferenceNum); @@ -1283,7 +1285,7 @@ static int udf_load_partdesc(struct super_block *sb, sector_t block) p = (struct partitionDesc *)bh->b_data; partitionNumber = le16_to_cpu(p->partitionNumber); - /* First scan for TYPE1, SPARABLE and METADATA partitions */ + /* First scan for TYPE1 and SPARABLE partitions */ for (i = 0; i < sbi->s_partitions; i++) { map = &sbi->s_partmaps[i]; udf_debug("Searching map: (%d == %d)\n", @@ -1333,7 +1335,7 @@ static int udf_load_partdesc(struct super_block *sb, sector_t block) goto out_bh; if (map->s_partition_type == UDF_METADATA_MAP25) { - ret = udf_load_metadata_files(sb, i); + ret = udf_load_metadata_files(sb, i, type1_idx); if (ret < 0) { udf_err(sb, "error loading MetaData partition map %d\n", i); diff --git a/fs/udf/udf_sb.h b/fs/udf/udf_sb.h index 27b5335730c9..c13875d669c0 100644 --- a/fs/udf/udf_sb.h +++ b/fs/udf/udf_sb.h @@ -61,6 +61,11 @@ struct udf_meta_data { __u32 s_bitmap_file_loc; __u32 s_alloc_unit_size; __u16 s_align_unit_size; + /* + * Partition Reference Number of the associated physical / sparable + * partition + */ + __u16 s_phys_partition_ref; int s_flags; struct inode *s_metadata_fe; struct inode *s_mirror_fe; -- cgit v1.3-8-gc7d7 From e1daca289a36965d923ec26647b5668e023eb0ac Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Tue, 17 May 2016 17:44:07 -0700 Subject: ocfs2/cluster: block BH in TCP callbacks TCP stack can now run from process context. Use read_lock_bh() variant to restore previous assumption. Fixes: 5413d1babe8f ("net: do not block BH while processing socket backlog") Fixes: d41a69f1d390 ("tcp: make tcp_sendmsg() aware of socket backlog") Signed-off-by: Eric Dumazet Signed-off-by: David S. Miller --- fs/ocfs2/cluster/tcp.c | 17 +++++++++-------- 1 file changed, 9 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/cluster/tcp.c b/fs/ocfs2/cluster/tcp.c index 2d0acd6678fe..4238eb28889f 100644 --- a/fs/ocfs2/cluster/tcp.c +++ b/fs/ocfs2/cluster/tcp.c @@ -600,10 +600,11 @@ static void o2net_set_nn_state(struct o2net_node *nn, static void o2net_data_ready(struct sock *sk) { void (*ready)(struct sock *sk); + struct o2net_sock_container *sc; - read_lock(&sk->sk_callback_lock); - if (sk->sk_user_data) { - struct o2net_sock_container *sc = sk->sk_user_data; + read_lock_bh(&sk->sk_callback_lock); + sc = sk->sk_user_data; + if (sc) { sclog(sc, "data_ready hit\n"); o2net_set_data_ready_time(sc); o2net_sc_queue_work(sc, &sc->sc_rx_work); @@ -611,7 +612,7 @@ static void o2net_data_ready(struct sock *sk) } else { ready = sk->sk_data_ready; } - read_unlock(&sk->sk_callback_lock); + read_unlock_bh(&sk->sk_callback_lock); ready(sk); } @@ -622,7 +623,7 @@ static void o2net_state_change(struct sock *sk) void (*state_change)(struct sock *sk); struct o2net_sock_container *sc; - read_lock(&sk->sk_callback_lock); + read_lock_bh(&sk->sk_callback_lock); sc = sk->sk_user_data; if (sc == NULL) { state_change = sk->sk_state_change; @@ -649,7 +650,7 @@ static void o2net_state_change(struct sock *sk) break; } out: - read_unlock(&sk->sk_callback_lock); + read_unlock_bh(&sk->sk_callback_lock); state_change(sk); } @@ -2012,7 +2013,7 @@ static void o2net_listen_data_ready(struct sock *sk) { void (*ready)(struct sock *sk); - read_lock(&sk->sk_callback_lock); + read_lock_bh(&sk->sk_callback_lock); ready = sk->sk_user_data; if (ready == NULL) { /* check for teardown race */ ready = sk->sk_data_ready; @@ -2039,7 +2040,7 @@ static void o2net_listen_data_ready(struct sock *sk) } out: - read_unlock(&sk->sk_callback_lock); + read_unlock_bh(&sk->sk_callback_lock); if (ready != NULL) ready(sk); } -- cgit v1.3-8-gc7d7 From b9953536c95fd0013695542bfa4694c7468673aa Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Thu, 12 May 2016 18:29:14 +0200 Subject: dax: Fix condition for filling of PMD holes Currently dax_pmd_fault() decides to fill a PMD-sized hole only if returned buffer has BH_Uptodate set. However that doesn't get set for any mapping buffer so that branch is actually a dead code. The BH_Uptodate check doesn't make any sense so just remove it. Signed-off-by: Jan Kara Signed-off-by: Ross Zwisler --- fs/dax.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/dax.c b/fs/dax.c index ea936b3d93dc..bdad05213e4b 100644 --- a/fs/dax.c +++ b/fs/dax.c @@ -788,7 +788,7 @@ int __dax_pmd_fault(struct vm_area_struct *vma, unsigned long address, i_mmap_lock_read(mapping); - if (!write && !buffer_mapped(&bh) && buffer_uptodate(&bh)) { + if (!write && !buffer_mapped(&bh)) { spinlock_t *ptl; pmd_t entry; struct page *zero_page = get_huge_zero_page(); -- cgit v1.3-8-gc7d7 From 348e967ab07c96a9e7a6a194812254a8df2045c0 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Thu, 12 May 2016 18:29:15 +0200 Subject: dax: Make huge page handling depend of CONFIG_BROKEN Currently the handling of huge pages for DAX is racy. For example the following can happen: CPU0 (THP write fault) CPU1 (normal read fault) __dax_pmd_fault() __dax_fault() get_block(inode, block, &bh, 0) -> not mapped get_block(inode, block, &bh, 0) -> not mapped if (!buffer_mapped(&bh) && write) get_block(inode, block, &bh, 1) -> allocates blocks truncate_pagecache_range(inode, lstart, lend); dax_load_hole(); This results in data corruption since process on CPU1 won't see changes into the file done by CPU0. The race can happen even if two normal faults race however with THP the situation is even worse because the two faults don't operate on the same entries in the radix tree and we want to use these entries for serialization. So make THP support in DAX code depend on CONFIG_BROKEN for now. Signed-off-by: Jan Kara Signed-off-by: Ross Zwisler --- fs/Kconfig | 1 + fs/dax.c | 2 +- include/linux/dax.h | 2 +- 3 files changed, 3 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/Kconfig b/fs/Kconfig index 6725f59c18e6..b8fcb416be72 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -52,6 +52,7 @@ config FS_DAX_PMD depends on FS_DAX depends on ZONE_DEVICE depends on TRANSPARENT_HUGEPAGE + depends on BROKEN endif # BLOCK diff --git a/fs/dax.c b/fs/dax.c index bdad05213e4b..0433a2b5e484 100644 --- a/fs/dax.c +++ b/fs/dax.c @@ -675,7 +675,7 @@ int dax_fault(struct vm_area_struct *vma, struct vm_fault *vmf, } EXPORT_SYMBOL_GPL(dax_fault); -#ifdef CONFIG_TRANSPARENT_HUGEPAGE +#if defined(CONFIG_TRANSPARENT_HUGEPAGE) /* * The 'colour' (ie low bits) within a PMD of a page offset. This comes up * more often than one might expect in the below function. diff --git a/include/linux/dax.h b/include/linux/dax.h index 90fbc99e5313..72dc81de3ddb 100644 --- a/include/linux/dax.h +++ b/include/linux/dax.h @@ -29,7 +29,7 @@ static inline int __dax_zero_page_range(struct block_device *bdev, } #endif -#ifdef CONFIG_TRANSPARENT_HUGEPAGE +#if defined(CONFIG_TRANSPARENT_HUGEPAGE) int dax_pmd_fault(struct vm_area_struct *, unsigned long addr, pmd_t *, unsigned int flags, get_block_t); int __dax_pmd_fault(struct vm_area_struct *, unsigned long addr, pmd_t *, -- cgit v1.3-8-gc7d7 From e804315dd0f574b56155c5a2406ab5e0318104f7 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Thu, 12 May 2016 18:29:16 +0200 Subject: dax: Define DAX lock bit for radix tree exceptional entry We will use lowest available bit in the radix tree exceptional entry for locking of the entry. Define it. Also clean up definitions of DAX entry type bits in DAX exceptional entries to use defined constants instead of hardcoding numbers and cleanup checking of these bits to not rely on how other bits in the entry are set. Reviewed-by: Ross Zwisler Signed-off-by: Jan Kara Signed-off-by: Ross Zwisler --- fs/dax.c | 17 +++++++++++------ include/linux/dax.h | 3 +++ 2 files changed, 14 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/dax.c b/fs/dax.c index 0433a2b5e484..351afd3cf8be 100644 --- a/fs/dax.c +++ b/fs/dax.c @@ -32,14 +32,19 @@ #include #include -#define RADIX_DAX_MASK 0xf -#define RADIX_DAX_SHIFT 4 -#define RADIX_DAX_PTE (0x4 | RADIX_TREE_EXCEPTIONAL_ENTRY) -#define RADIX_DAX_PMD (0x8 | RADIX_TREE_EXCEPTIONAL_ENTRY) -#define RADIX_DAX_TYPE(entry) ((unsigned long)entry & RADIX_DAX_MASK) +/* + * We use lowest available bit in exceptional entry for locking, other two + * bits to determine entry type. In total 3 special bits. + */ +#define RADIX_DAX_SHIFT (RADIX_TREE_EXCEPTIONAL_SHIFT + 3) +#define RADIX_DAX_PTE (1 << (RADIX_TREE_EXCEPTIONAL_SHIFT + 1)) +#define RADIX_DAX_PMD (1 << (RADIX_TREE_EXCEPTIONAL_SHIFT + 2)) +#define RADIX_DAX_TYPE_MASK (RADIX_DAX_PTE | RADIX_DAX_PMD) +#define RADIX_DAX_TYPE(entry) ((unsigned long)entry & RADIX_DAX_TYPE_MASK) #define RADIX_DAX_SECTOR(entry) (((unsigned long)entry >> RADIX_DAX_SHIFT)) #define RADIX_DAX_ENTRY(sector, pmd) ((void *)((unsigned long)sector << \ - RADIX_DAX_SHIFT | (pmd ? RADIX_DAX_PMD : RADIX_DAX_PTE))) + RADIX_DAX_SHIFT | (pmd ? RADIX_DAX_PMD : RADIX_DAX_PTE) | \ + RADIX_TREE_EXCEPTIONAL_ENTRY)) static long dax_map_atomic(struct block_device *bdev, struct blk_dax_ctl *dax) { diff --git a/include/linux/dax.h b/include/linux/dax.h index 72dc81de3ddb..70600b63083f 100644 --- a/include/linux/dax.h +++ b/include/linux/dax.h @@ -5,6 +5,9 @@ #include #include +/* We use lowest available exceptional entry bit for locking */ +#define RADIX_DAX_ENTRY_LOCK (1 << RADIX_TREE_EXCEPTIONAL_SHIFT) + ssize_t dax_do_io(struct kiocb *, struct inode *, struct iov_iter *, loff_t, get_block_t, dio_iodone_t, int flags); int dax_zero_page_range(struct inode *, loff_t from, unsigned len, get_block_t); -- cgit v1.3-8-gc7d7 From ac401cc782429cc8560ce4840b1405d603740917 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Thu, 12 May 2016 18:29:18 +0200 Subject: dax: New fault locking Currently DAX page fault locking is racy. CPU0 (write fault) CPU1 (read fault) __dax_fault() __dax_fault() get_block(inode, block, &bh, 0) -> not mapped get_block(inode, block, &bh, 0) -> not mapped if (!buffer_mapped(&bh)) if (vmf->flags & FAULT_FLAG_WRITE) get_block(inode, block, &bh, 1) -> allocates blocks if (page) -> no if (!buffer_mapped(&bh)) if (vmf->flags & FAULT_FLAG_WRITE) { } else { dax_load_hole(); } dax_insert_mapping() And we are in a situation where we fail in dax_radix_entry() with -EIO. Another problem with the current DAX page fault locking is that there is no race-free way to clear dirty tag in the radix tree. We can always end up with clean radix tree and dirty data in CPU cache. We fix the first problem by introducing locking of exceptional radix tree entries in DAX mappings acting very similarly to page lock and thus synchronizing properly faults against the same mapping index. The same lock can later be used to avoid races when clearing radix tree dirty tag. Reviewed-by: NeilBrown Reviewed-by: Ross Zwisler Signed-off-by: Jan Kara Signed-off-by: Ross Zwisler --- fs/dax.c | 553 ++++++++++++++++++++++++++++++++++++++-------------- include/linux/dax.h | 3 + mm/filemap.c | 9 +- mm/truncate.c | 62 +++--- 4 files changed, 447 insertions(+), 180 deletions(-) (limited to 'fs') diff --git a/fs/dax.c b/fs/dax.c index 351afd3cf8be..f43c3d806fb6 100644 --- a/fs/dax.c +++ b/fs/dax.c @@ -46,6 +46,30 @@ RADIX_DAX_SHIFT | (pmd ? RADIX_DAX_PMD : RADIX_DAX_PTE) | \ RADIX_TREE_EXCEPTIONAL_ENTRY)) +/* We choose 4096 entries - same as per-zone page wait tables */ +#define DAX_WAIT_TABLE_BITS 12 +#define DAX_WAIT_TABLE_ENTRIES (1 << DAX_WAIT_TABLE_BITS) + +wait_queue_head_t wait_table[DAX_WAIT_TABLE_ENTRIES]; + +static int __init init_dax_wait_table(void) +{ + int i; + + for (i = 0; i < DAX_WAIT_TABLE_ENTRIES; i++) + init_waitqueue_head(wait_table + i); + return 0; +} +fs_initcall(init_dax_wait_table); + +static wait_queue_head_t *dax_entry_waitqueue(struct address_space *mapping, + pgoff_t index) +{ + unsigned long hash = hash_long((unsigned long)mapping ^ index, + DAX_WAIT_TABLE_BITS); + return wait_table + hash; +} + static long dax_map_atomic(struct block_device *bdev, struct blk_dax_ctl *dax) { struct request_queue *q = bdev->bd_queue; @@ -267,6 +291,263 @@ ssize_t dax_do_io(struct kiocb *iocb, struct inode *inode, } EXPORT_SYMBOL_GPL(dax_do_io); +/* + * DAX radix tree locking + */ +struct exceptional_entry_key { + struct address_space *mapping; + unsigned long index; +}; + +struct wait_exceptional_entry_queue { + wait_queue_t wait; + struct exceptional_entry_key key; +}; + +static int wake_exceptional_entry_func(wait_queue_t *wait, unsigned int mode, + int sync, void *keyp) +{ + struct exceptional_entry_key *key = keyp; + struct wait_exceptional_entry_queue *ewait = + container_of(wait, struct wait_exceptional_entry_queue, wait); + + if (key->mapping != ewait->key.mapping || + key->index != ewait->key.index) + return 0; + return autoremove_wake_function(wait, mode, sync, NULL); +} + +/* + * Check whether the given slot is locked. The function must be called with + * mapping->tree_lock held + */ +static inline int slot_locked(struct address_space *mapping, void **slot) +{ + unsigned long entry = (unsigned long) + radix_tree_deref_slot_protected(slot, &mapping->tree_lock); + return entry & RADIX_DAX_ENTRY_LOCK; +} + +/* + * Mark the given slot is locked. The function must be called with + * mapping->tree_lock held + */ +static inline void *lock_slot(struct address_space *mapping, void **slot) +{ + unsigned long entry = (unsigned long) + radix_tree_deref_slot_protected(slot, &mapping->tree_lock); + + entry |= RADIX_DAX_ENTRY_LOCK; + radix_tree_replace_slot(slot, (void *)entry); + return (void *)entry; +} + +/* + * Mark the given slot is unlocked. The function must be called with + * mapping->tree_lock held + */ +static inline void *unlock_slot(struct address_space *mapping, void **slot) +{ + unsigned long entry = (unsigned long) + radix_tree_deref_slot_protected(slot, &mapping->tree_lock); + + entry &= ~(unsigned long)RADIX_DAX_ENTRY_LOCK; + radix_tree_replace_slot(slot, (void *)entry); + return (void *)entry; +} + +/* + * Lookup entry in radix tree, wait for it to become unlocked if it is + * exceptional entry and return it. The caller must call + * put_unlocked_mapping_entry() when he decided not to lock the entry or + * put_locked_mapping_entry() when he locked the entry and now wants to + * unlock it. + * + * The function must be called with mapping->tree_lock held. + */ +static void *get_unlocked_mapping_entry(struct address_space *mapping, + pgoff_t index, void ***slotp) +{ + void *ret, **slot; + struct wait_exceptional_entry_queue ewait; + wait_queue_head_t *wq = dax_entry_waitqueue(mapping, index); + + init_wait(&ewait.wait); + ewait.wait.func = wake_exceptional_entry_func; + ewait.key.mapping = mapping; + ewait.key.index = index; + + for (;;) { + ret = __radix_tree_lookup(&mapping->page_tree, index, NULL, + &slot); + if (!ret || !radix_tree_exceptional_entry(ret) || + !slot_locked(mapping, slot)) { + if (slotp) + *slotp = slot; + return ret; + } + prepare_to_wait_exclusive(wq, &ewait.wait, + TASK_UNINTERRUPTIBLE); + spin_unlock_irq(&mapping->tree_lock); + schedule(); + finish_wait(wq, &ewait.wait); + spin_lock_irq(&mapping->tree_lock); + } +} + +/* + * Find radix tree entry at given index. If it points to a page, return with + * the page locked. If it points to the exceptional entry, return with the + * radix tree entry locked. If the radix tree doesn't contain given index, + * create empty exceptional entry for the index and return with it locked. + * + * Note: Unlike filemap_fault() we don't honor FAULT_FLAG_RETRY flags. For + * persistent memory the benefit is doubtful. We can add that later if we can + * show it helps. + */ +static void *grab_mapping_entry(struct address_space *mapping, pgoff_t index) +{ + void *ret, **slot; + +restart: + spin_lock_irq(&mapping->tree_lock); + ret = get_unlocked_mapping_entry(mapping, index, &slot); + /* No entry for given index? Make sure radix tree is big enough. */ + if (!ret) { + int err; + + spin_unlock_irq(&mapping->tree_lock); + err = radix_tree_preload( + mapping_gfp_mask(mapping) & ~__GFP_HIGHMEM); + if (err) + return ERR_PTR(err); + ret = (void *)(RADIX_TREE_EXCEPTIONAL_ENTRY | + RADIX_DAX_ENTRY_LOCK); + spin_lock_irq(&mapping->tree_lock); + err = radix_tree_insert(&mapping->page_tree, index, ret); + radix_tree_preload_end(); + if (err) { + spin_unlock_irq(&mapping->tree_lock); + /* Someone already created the entry? */ + if (err == -EEXIST) + goto restart; + return ERR_PTR(err); + } + /* Good, we have inserted empty locked entry into the tree. */ + mapping->nrexceptional++; + spin_unlock_irq(&mapping->tree_lock); + return ret; + } + /* Normal page in radix tree? */ + if (!radix_tree_exceptional_entry(ret)) { + struct page *page = ret; + + get_page(page); + spin_unlock_irq(&mapping->tree_lock); + lock_page(page); + /* Page got truncated? Retry... */ + if (unlikely(page->mapping != mapping)) { + unlock_page(page); + put_page(page); + goto restart; + } + return page; + } + ret = lock_slot(mapping, slot); + spin_unlock_irq(&mapping->tree_lock); + return ret; +} + +void dax_wake_mapping_entry_waiter(struct address_space *mapping, + pgoff_t index, bool wake_all) +{ + wait_queue_head_t *wq = dax_entry_waitqueue(mapping, index); + + /* + * Checking for locked entry and prepare_to_wait_exclusive() happens + * under mapping->tree_lock, ditto for entry handling in our callers. + * So at this point all tasks that could have seen our entry locked + * must be in the waitqueue and the following check will see them. + */ + if (waitqueue_active(wq)) { + struct exceptional_entry_key key; + + key.mapping = mapping; + key.index = index; + __wake_up(wq, TASK_NORMAL, wake_all ? 0 : 1, &key); + } +} + +static void unlock_mapping_entry(struct address_space *mapping, pgoff_t index) +{ + void *ret, **slot; + + spin_lock_irq(&mapping->tree_lock); + ret = __radix_tree_lookup(&mapping->page_tree, index, NULL, &slot); + if (WARN_ON_ONCE(!ret || !radix_tree_exceptional_entry(ret) || + !slot_locked(mapping, slot))) { + spin_unlock_irq(&mapping->tree_lock); + return; + } + unlock_slot(mapping, slot); + spin_unlock_irq(&mapping->tree_lock); + dax_wake_mapping_entry_waiter(mapping, index, false); +} + +static void put_locked_mapping_entry(struct address_space *mapping, + pgoff_t index, void *entry) +{ + if (!radix_tree_exceptional_entry(entry)) { + unlock_page(entry); + put_page(entry); + } else { + unlock_mapping_entry(mapping, index); + } +} + +/* + * Called when we are done with radix tree entry we looked up via + * get_unlocked_mapping_entry() and which we didn't lock in the end. + */ +static void put_unlocked_mapping_entry(struct address_space *mapping, + pgoff_t index, void *entry) +{ + if (!radix_tree_exceptional_entry(entry)) + return; + + /* We have to wake up next waiter for the radix tree entry lock */ + dax_wake_mapping_entry_waiter(mapping, index, false); +} + +/* + * Delete exceptional DAX entry at @index from @mapping. Wait for radix tree + * entry to get unlocked before deleting it. + */ +int dax_delete_mapping_entry(struct address_space *mapping, pgoff_t index) +{ + void *entry; + + spin_lock_irq(&mapping->tree_lock); + entry = get_unlocked_mapping_entry(mapping, index, NULL); + /* + * This gets called from truncate / punch_hole path. As such, the caller + * must hold locks protecting against concurrent modifications of the + * radix tree (usually fs-private i_mmap_sem for writing). Since the + * caller has seen exceptional entry for this index, we better find it + * at that index as well... + */ + if (WARN_ON_ONCE(!entry || !radix_tree_exceptional_entry(entry))) { + spin_unlock_irq(&mapping->tree_lock); + return 0; + } + radix_tree_delete(&mapping->page_tree, index); + mapping->nrexceptional--; + spin_unlock_irq(&mapping->tree_lock); + dax_wake_mapping_entry_waiter(mapping, index, true); + + return 1; +} + /* * The user has performed a load from a hole in the file. Allocating * a new page in the file would cause excessive storage usage for @@ -275,15 +556,24 @@ EXPORT_SYMBOL_GPL(dax_do_io); * otherwise it will simply fall out of the page cache under memory * pressure without ever having been dirtied. */ -static int dax_load_hole(struct address_space *mapping, struct page *page, - struct vm_fault *vmf) +static int dax_load_hole(struct address_space *mapping, void *entry, + struct vm_fault *vmf) { - if (!page) - page = find_or_create_page(mapping, vmf->pgoff, - GFP_KERNEL | __GFP_ZERO); - if (!page) - return VM_FAULT_OOM; + struct page *page; + + /* Hole page already exists? Return it... */ + if (!radix_tree_exceptional_entry(entry)) { + vmf->page = entry; + return VM_FAULT_LOCKED; + } + /* This will replace locked radix tree entry with a hole page */ + page = find_or_create_page(mapping, vmf->pgoff, + vmf->gfp_mask | __GFP_ZERO); + if (!page) { + put_locked_mapping_entry(mapping, vmf->pgoff, entry); + return VM_FAULT_OOM; + } vmf->page = page; return VM_FAULT_LOCKED; } @@ -307,77 +597,72 @@ static int copy_user_bh(struct page *to, struct inode *inode, return 0; } -#define NO_SECTOR -1 #define DAX_PMD_INDEX(page_index) (page_index & (PMD_MASK >> PAGE_SHIFT)) -static int dax_radix_entry(struct address_space *mapping, pgoff_t index, - sector_t sector, bool pmd_entry, bool dirty) +static void *dax_insert_mapping_entry(struct address_space *mapping, + struct vm_fault *vmf, + void *entry, sector_t sector) { struct radix_tree_root *page_tree = &mapping->page_tree; - pgoff_t pmd_index = DAX_PMD_INDEX(index); - int type, error = 0; - void *entry; + int error = 0; + bool hole_fill = false; + void *new_entry; + pgoff_t index = vmf->pgoff; - WARN_ON_ONCE(pmd_entry && !dirty); - if (dirty) + if (vmf->flags & FAULT_FLAG_WRITE) __mark_inode_dirty(mapping->host, I_DIRTY_PAGES); - spin_lock_irq(&mapping->tree_lock); - - entry = radix_tree_lookup(page_tree, pmd_index); - if (entry && RADIX_DAX_TYPE(entry) == RADIX_DAX_PMD) { - index = pmd_index; - goto dirty; + /* Replacing hole page with block mapping? */ + if (!radix_tree_exceptional_entry(entry)) { + hole_fill = true; + /* + * Unmap the page now before we remove it from page cache below. + * The page is locked so it cannot be faulted in again. + */ + unmap_mapping_range(mapping, vmf->pgoff << PAGE_SHIFT, + PAGE_SIZE, 0); + error = radix_tree_preload(vmf->gfp_mask & ~__GFP_HIGHMEM); + if (error) + return ERR_PTR(error); } - entry = radix_tree_lookup(page_tree, index); - if (entry) { - type = RADIX_DAX_TYPE(entry); - if (WARN_ON_ONCE(type != RADIX_DAX_PTE && - type != RADIX_DAX_PMD)) { - error = -EIO; + spin_lock_irq(&mapping->tree_lock); + new_entry = (void *)((unsigned long)RADIX_DAX_ENTRY(sector, false) | + RADIX_DAX_ENTRY_LOCK); + if (hole_fill) { + __delete_from_page_cache(entry, NULL); + /* Drop pagecache reference */ + put_page(entry); + error = radix_tree_insert(page_tree, index, new_entry); + if (error) { + new_entry = ERR_PTR(error); goto unlock; } + mapping->nrexceptional++; + } else { + void **slot; + void *ret; - if (!pmd_entry || type == RADIX_DAX_PMD) - goto dirty; - - /* - * We only insert dirty PMD entries into the radix tree. This - * means we don't need to worry about removing a dirty PTE - * entry and inserting a clean PMD entry, thus reducing the - * range we would flush with a follow-up fsync/msync call. - */ - radix_tree_delete(&mapping->page_tree, index); - mapping->nrexceptional--; - } - - if (sector == NO_SECTOR) { - /* - * This can happen during correct operation if our pfn_mkwrite - * fault raced against a hole punch operation. If this - * happens the pte that was hole punched will have been - * unmapped and the radix tree entry will have been removed by - * the time we are called, but the call will still happen. We - * will return all the way up to wp_pfn_shared(), where the - * pte_same() check will fail, eventually causing page fault - * to be retried by the CPU. - */ - goto unlock; + ret = __radix_tree_lookup(page_tree, index, NULL, &slot); + WARN_ON_ONCE(ret != entry); + radix_tree_replace_slot(slot, new_entry); } - - error = radix_tree_insert(page_tree, index, - RADIX_DAX_ENTRY(sector, pmd_entry)); - if (error) - goto unlock; - - mapping->nrexceptional++; - dirty: - if (dirty) + if (vmf->flags & FAULT_FLAG_WRITE) radix_tree_tag_set(page_tree, index, PAGECACHE_TAG_DIRTY); unlock: spin_unlock_irq(&mapping->tree_lock); - return error; + if (hole_fill) { + radix_tree_preload_end(); + /* + * We don't need hole page anymore, it has been replaced with + * locked radix tree entry now. + */ + if (mapping->a_ops->freepage) + mapping->a_ops->freepage(entry); + unlock_page(entry); + put_page(entry); + } + return new_entry; } static int dax_writeback_one(struct block_device *bdev, @@ -503,17 +788,19 @@ int dax_writeback_mapping_range(struct address_space *mapping, } EXPORT_SYMBOL_GPL(dax_writeback_mapping_range); -static int dax_insert_mapping(struct inode *inode, struct buffer_head *bh, +static int dax_insert_mapping(struct address_space *mapping, + struct buffer_head *bh, void **entryp, struct vm_area_struct *vma, struct vm_fault *vmf) { unsigned long vaddr = (unsigned long)vmf->virtual_address; - struct address_space *mapping = inode->i_mapping; struct block_device *bdev = bh->b_bdev; struct blk_dax_ctl dax = { - .sector = to_sector(bh, inode), + .sector = to_sector(bh, mapping->host), .size = bh->b_size, }; int error; + void *ret; + void *entry = *entryp; i_mmap_lock_read(mapping); @@ -523,16 +810,16 @@ static int dax_insert_mapping(struct inode *inode, struct buffer_head *bh, } dax_unmap_atomic(bdev, &dax); - error = dax_radix_entry(mapping, vmf->pgoff, dax.sector, false, - vmf->flags & FAULT_FLAG_WRITE); - if (error) + ret = dax_insert_mapping_entry(mapping, vmf, entry, dax.sector); + if (IS_ERR(ret)) { + error = PTR_ERR(ret); goto out; + } + *entryp = ret; error = vm_insert_mixed(vma, vaddr, dax.pfn); - out: i_mmap_unlock_read(mapping); - return error; } @@ -552,7 +839,7 @@ int __dax_fault(struct vm_area_struct *vma, struct vm_fault *vmf, struct file *file = vma->vm_file; struct address_space *mapping = file->f_mapping; struct inode *inode = mapping->host; - struct page *page; + void *entry; struct buffer_head bh; unsigned long vaddr = (unsigned long)vmf->virtual_address; unsigned blkbits = inode->i_blkbits; @@ -561,6 +848,11 @@ int __dax_fault(struct vm_area_struct *vma, struct vm_fault *vmf, int error; int major = 0; + /* + * Check whether offset isn't beyond end of file now. Caller is supposed + * to hold locks serializing us with truncate / punch hole so this is + * a reliable test. + */ size = (i_size_read(inode) + PAGE_SIZE - 1) >> PAGE_SHIFT; if (vmf->pgoff >= size) return VM_FAULT_SIGBUS; @@ -570,40 +862,17 @@ int __dax_fault(struct vm_area_struct *vma, struct vm_fault *vmf, bh.b_bdev = inode->i_sb->s_bdev; bh.b_size = PAGE_SIZE; - repeat: - page = find_get_page(mapping, vmf->pgoff); - if (page) { - if (!lock_page_or_retry(page, vma->vm_mm, vmf->flags)) { - put_page(page); - return VM_FAULT_RETRY; - } - if (unlikely(page->mapping != mapping)) { - unlock_page(page); - put_page(page); - goto repeat; - } + entry = grab_mapping_entry(mapping, vmf->pgoff); + if (IS_ERR(entry)) { + error = PTR_ERR(entry); + goto out; } error = get_block(inode, block, &bh, 0); if (!error && (bh.b_size < PAGE_SIZE)) error = -EIO; /* fs corruption? */ if (error) - goto unlock_page; - - if (!buffer_mapped(&bh) && !vmf->cow_page) { - if (vmf->flags & FAULT_FLAG_WRITE) { - error = get_block(inode, block, &bh, 1); - count_vm_event(PGMAJFAULT); - mem_cgroup_count_vm_event(vma->vm_mm, PGMAJFAULT); - major = VM_FAULT_MAJOR; - if (!error && (bh.b_size < PAGE_SIZE)) - error = -EIO; - if (error) - goto unlock_page; - } else { - return dax_load_hole(mapping, page, vmf); - } - } + goto unlock_entry; if (vmf->cow_page) { struct page *new_page = vmf->cow_page; @@ -612,30 +881,37 @@ int __dax_fault(struct vm_area_struct *vma, struct vm_fault *vmf, else clear_user_highpage(new_page, vaddr); if (error) - goto unlock_page; - vmf->page = page; - if (!page) + goto unlock_entry; + if (!radix_tree_exceptional_entry(entry)) { + vmf->page = entry; + } else { + unlock_mapping_entry(mapping, vmf->pgoff); i_mmap_lock_read(mapping); + vmf->page = NULL; + } return VM_FAULT_LOCKED; } - /* Check we didn't race with a read fault installing a new page */ - if (!page && major) - page = find_lock_page(mapping, vmf->pgoff); - - if (page) { - unmap_mapping_range(mapping, vmf->pgoff << PAGE_SHIFT, - PAGE_SIZE, 0); - delete_from_page_cache(page); - unlock_page(page); - put_page(page); - page = NULL; + if (!buffer_mapped(&bh)) { + if (vmf->flags & FAULT_FLAG_WRITE) { + error = get_block(inode, block, &bh, 1); + count_vm_event(PGMAJFAULT); + mem_cgroup_count_vm_event(vma->vm_mm, PGMAJFAULT); + major = VM_FAULT_MAJOR; + if (!error && (bh.b_size < PAGE_SIZE)) + error = -EIO; + if (error) + goto unlock_entry; + } else { + return dax_load_hole(mapping, entry, vmf); + } } /* Filesystem should not return unwritten buffers to us! */ WARN_ON_ONCE(buffer_unwritten(&bh) || buffer_new(&bh)); - error = dax_insert_mapping(inode, &bh, vma, vmf); - + error = dax_insert_mapping(mapping, &bh, &entry, vma, vmf); + unlock_entry: + put_locked_mapping_entry(mapping, vmf->pgoff, entry); out: if (error == -ENOMEM) return VM_FAULT_OOM | major; @@ -643,13 +919,6 @@ int __dax_fault(struct vm_area_struct *vma, struct vm_fault *vmf, if ((error < 0) && (error != -EBUSY)) return VM_FAULT_SIGBUS | major; return VM_FAULT_NOPAGE | major; - - unlock_page: - if (page) { - unlock_page(page); - put_page(page); - } - goto out; } EXPORT_SYMBOL(__dax_fault); @@ -718,7 +987,7 @@ int __dax_pmd_fault(struct vm_area_struct *vma, unsigned long address, struct block_device *bdev; pgoff_t size, pgoff; sector_t block; - int error, result = 0; + int result = 0; bool alloc = false; /* dax pmd mappings require pfn_t_devmap() */ @@ -865,13 +1134,10 @@ int __dax_pmd_fault(struct vm_area_struct *vma, unsigned long address, * the write to insert a dirty entry. */ if (write) { - error = dax_radix_entry(mapping, pgoff, dax.sector, - true, true); - if (error) { - dax_pmd_dbg(&bh, address, - "PMD radix insertion failed"); - goto fallback; - } + /* + * We should insert radix-tree entry and dirty it here. + * For now this is broken... + */ } dev_dbg(part_to_dev(bdev->bd_part), @@ -931,23 +1197,18 @@ EXPORT_SYMBOL_GPL(dax_pmd_fault); int dax_pfn_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) { struct file *file = vma->vm_file; - int error; - - /* - * We pass NO_SECTOR to dax_radix_entry() because we expect that a - * RADIX_DAX_PTE entry already exists in the radix tree from a - * previous call to __dax_fault(). We just want to look up that PTE - * entry using vmf->pgoff and make sure the dirty tag is set. This - * saves us from having to make a call to get_block() here to look - * up the sector. - */ - error = dax_radix_entry(file->f_mapping, vmf->pgoff, NO_SECTOR, false, - true); + struct address_space *mapping = file->f_mapping; + void *entry; + pgoff_t index = vmf->pgoff; - if (error == -ENOMEM) - return VM_FAULT_OOM; - if (error) - return VM_FAULT_SIGBUS; + spin_lock_irq(&mapping->tree_lock); + entry = get_unlocked_mapping_entry(mapping, index, NULL); + if (!entry || !radix_tree_exceptional_entry(entry)) + goto out; + radix_tree_tag_set(&mapping->page_tree, index, PAGECACHE_TAG_DIRTY); + put_unlocked_mapping_entry(mapping, index, entry); +out: + spin_unlock_irq(&mapping->tree_lock); return VM_FAULT_NOPAGE; } EXPORT_SYMBOL_GPL(dax_pfn_mkwrite); diff --git a/include/linux/dax.h b/include/linux/dax.h index aa148937bb3f..756625c6d0dd 100644 --- a/include/linux/dax.h +++ b/include/linux/dax.h @@ -15,6 +15,9 @@ int dax_zero_page_range(struct inode *, loff_t from, unsigned len, get_block_t); int dax_truncate_page(struct inode *, loff_t from, get_block_t); int dax_fault(struct vm_area_struct *, struct vm_fault *, get_block_t); int __dax_fault(struct vm_area_struct *, struct vm_fault *, get_block_t); +int dax_delete_mapping_entry(struct address_space *mapping, pgoff_t index); +void dax_wake_mapping_entry_waiter(struct address_space *mapping, + pgoff_t index, bool wake_all); #ifdef CONFIG_FS_DAX struct page *read_dax_sector(struct block_device *bdev, sector_t n); diff --git a/mm/filemap.c b/mm/filemap.c index dfe55c2cfb34..7b9a4b180cae 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -160,13 +160,15 @@ static void page_cache_tree_delete(struct address_space *mapping, return; /* - * Track node that only contains shadow entries. + * Track node that only contains shadow entries. DAX mappings contain + * no shadow entries and may contain other exceptional entries so skip + * those. * * Avoid acquiring the list_lru lock if already tracked. The * list_empty() test is safe as node->private_list is * protected by mapping->tree_lock. */ - if (!workingset_node_pages(node) && + if (!dax_mapping(mapping) && !workingset_node_pages(node) && list_empty(&node->private_list)) { node->private_data = mapping; list_lru_add(&workingset_shadow_nodes, &node->private_list); @@ -611,6 +613,9 @@ static int page_cache_tree_insert(struct address_space *mapping, /* DAX accounts exceptional entries as normal pages */ if (node) workingset_node_pages_dec(node); + /* Wakeup waiters for exceptional entry lock */ + dax_wake_mapping_entry_waiter(mapping, page->index, + false); } } radix_tree_replace_slot(slot, page); diff --git a/mm/truncate.c b/mm/truncate.c index b00272810871..4064f8f53daa 100644 --- a/mm/truncate.c +++ b/mm/truncate.c @@ -34,40 +34,38 @@ static void clear_exceptional_entry(struct address_space *mapping, if (shmem_mapping(mapping)) return; - spin_lock_irq(&mapping->tree_lock); - if (dax_mapping(mapping)) { - if (radix_tree_delete_item(&mapping->page_tree, index, entry)) - mapping->nrexceptional--; - } else { - /* - * Regular page slots are stabilized by the page lock even - * without the tree itself locked. These unlocked entries - * need verification under the tree lock. - */ - if (!__radix_tree_lookup(&mapping->page_tree, index, &node, - &slot)) - goto unlock; - if (*slot != entry) - goto unlock; - radix_tree_replace_slot(slot, NULL); - mapping->nrexceptional--; - if (!node) - goto unlock; - workingset_node_shadows_dec(node); - /* - * Don't track node without shadow entries. - * - * Avoid acquiring the list_lru lock if already untracked. - * The list_empty() test is safe as node->private_list is - * protected by mapping->tree_lock. - */ - if (!workingset_node_shadows(node) && - !list_empty(&node->private_list)) - list_lru_del(&workingset_shadow_nodes, - &node->private_list); - __radix_tree_delete_node(&mapping->page_tree, node); + dax_delete_mapping_entry(mapping, index); + return; } + spin_lock_irq(&mapping->tree_lock); + /* + * Regular page slots are stabilized by the page lock even + * without the tree itself locked. These unlocked entries + * need verification under the tree lock. + */ + if (!__radix_tree_lookup(&mapping->page_tree, index, &node, + &slot)) + goto unlock; + if (*slot != entry) + goto unlock; + radix_tree_replace_slot(slot, NULL); + mapping->nrexceptional--; + if (!node) + goto unlock; + workingset_node_shadows_dec(node); + /* + * Don't track node without shadow entries. + * + * Avoid acquiring the list_lru lock if already untracked. + * The list_empty() test is safe as node->private_list is + * protected by mapping->tree_lock. + */ + if (!workingset_node_shadows(node) && + !list_empty(&node->private_list)) + list_lru_del(&workingset_shadow_nodes, + &node->private_list); + __radix_tree_delete_node(&mapping->page_tree, node); unlock: spin_unlock_irq(&mapping->tree_lock); } -- cgit v1.3-8-gc7d7 From bc2466e4257369d0ebee2b6265070d323343fa72 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Thu, 12 May 2016 18:29:19 +0200 Subject: dax: Use radix tree entry lock to protect cow faults When doing cow faults, we cannot directly fill in PTE as we do for other faults as we rely on generic code to do proper accounting of the cowed page. We also have no page to lock to protect against races with truncate as other faults have and we need the protection to extend until the moment generic code inserts cowed page into PTE thus at that point we have no protection of fs-specific i_mmap_sem. So far we relied on using i_mmap_lock for the protection however that is completely special to cow faults. To make fault locking more uniform use DAX entry lock instead. Reviewed-by: Ross Zwisler Signed-off-by: Jan Kara Signed-off-by: Ross Zwisler --- fs/dax.c | 12 +++++------- include/linux/dax.h | 7 +++++++ include/linux/mm.h | 7 +++++++ mm/memory.c | 38 ++++++++++++++++++-------------------- 4 files changed, 37 insertions(+), 27 deletions(-) (limited to 'fs') diff --git a/fs/dax.c b/fs/dax.c index f43c3d806fb6..be74635e05a6 100644 --- a/fs/dax.c +++ b/fs/dax.c @@ -478,7 +478,7 @@ void dax_wake_mapping_entry_waiter(struct address_space *mapping, } } -static void unlock_mapping_entry(struct address_space *mapping, pgoff_t index) +void dax_unlock_mapping_entry(struct address_space *mapping, pgoff_t index) { void *ret, **slot; @@ -501,7 +501,7 @@ static void put_locked_mapping_entry(struct address_space *mapping, unlock_page(entry); put_page(entry); } else { - unlock_mapping_entry(mapping, index); + dax_unlock_mapping_entry(mapping, index); } } @@ -884,12 +884,10 @@ int __dax_fault(struct vm_area_struct *vma, struct vm_fault *vmf, goto unlock_entry; if (!radix_tree_exceptional_entry(entry)) { vmf->page = entry; - } else { - unlock_mapping_entry(mapping, vmf->pgoff); - i_mmap_lock_read(mapping); - vmf->page = NULL; + return VM_FAULT_LOCKED; } - return VM_FAULT_LOCKED; + vmf->entry = entry; + return VM_FAULT_DAX_LOCKED; } if (!buffer_mapped(&bh)) { diff --git a/include/linux/dax.h b/include/linux/dax.h index 756625c6d0dd..7bf12277c006 100644 --- a/include/linux/dax.h +++ b/include/linux/dax.h @@ -21,6 +21,7 @@ void dax_wake_mapping_entry_waiter(struct address_space *mapping, #ifdef CONFIG_FS_DAX struct page *read_dax_sector(struct block_device *bdev, sector_t n); +void dax_unlock_mapping_entry(struct address_space *mapping, pgoff_t index); int __dax_zero_page_range(struct block_device *bdev, sector_t sector, unsigned int offset, unsigned int length); #else @@ -29,6 +30,12 @@ static inline struct page *read_dax_sector(struct block_device *bdev, { return ERR_PTR(-ENXIO); } +/* Shouldn't ever be called when dax is disabled. */ +static inline void dax_unlock_mapping_entry(struct address_space *mapping, + pgoff_t index) +{ + BUG(); +} static inline int __dax_zero_page_range(struct block_device *bdev, sector_t sector, unsigned int offset, unsigned int length) { diff --git a/include/linux/mm.h b/include/linux/mm.h index a55e5be0894f..0ef9dc720ec3 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -299,6 +299,12 @@ struct vm_fault { * is set (which is also implied by * VM_FAULT_ERROR). */ + void *entry; /* ->fault handler can alternatively + * return locked DAX entry. In that + * case handler should return + * VM_FAULT_DAX_LOCKED and fill in + * entry here. + */ /* for ->map_pages() only */ pgoff_t max_pgoff; /* map pages for offset from pgoff till * max_pgoff inclusive */ @@ -1084,6 +1090,7 @@ static inline void clear_page_pfmemalloc(struct page *page) #define VM_FAULT_LOCKED 0x0200 /* ->fault locked the returned page */ #define VM_FAULT_RETRY 0x0400 /* ->fault blocked, must retry */ #define VM_FAULT_FALLBACK 0x0800 /* huge page fault failed, fall back to small */ +#define VM_FAULT_DAX_LOCKED 0x1000 /* ->fault has locked DAX entry */ #define VM_FAULT_HWPOISON_LARGE_MASK 0xf000 /* encodes hpage index for large hwpoison */ diff --git a/mm/memory.c b/mm/memory.c index 93897f23cc11..f09cdb8d48fa 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -63,6 +63,7 @@ #include #include #include +#include #include #include @@ -2785,7 +2786,8 @@ oom: */ static int __do_fault(struct vm_area_struct *vma, unsigned long address, pgoff_t pgoff, unsigned int flags, - struct page *cow_page, struct page **page) + struct page *cow_page, struct page **page, + void **entry) { struct vm_fault vmf; int ret; @@ -2800,8 +2802,10 @@ static int __do_fault(struct vm_area_struct *vma, unsigned long address, ret = vma->vm_ops->fault(vma, &vmf); if (unlikely(ret & (VM_FAULT_ERROR | VM_FAULT_NOPAGE | VM_FAULT_RETRY))) return ret; - if (!vmf.page) - goto out; + if (ret & VM_FAULT_DAX_LOCKED) { + *entry = vmf.entry; + return ret; + } if (unlikely(PageHWPoison(vmf.page))) { if (ret & VM_FAULT_LOCKED) @@ -2815,7 +2819,6 @@ static int __do_fault(struct vm_area_struct *vma, unsigned long address, else VM_BUG_ON_PAGE(!PageLocked(vmf.page), vmf.page); - out: *page = vmf.page; return ret; } @@ -2987,7 +2990,7 @@ static int do_read_fault(struct mm_struct *mm, struct vm_area_struct *vma, pte_unmap_unlock(pte, ptl); } - ret = __do_fault(vma, address, pgoff, flags, NULL, &fault_page); + ret = __do_fault(vma, address, pgoff, flags, NULL, &fault_page, NULL); if (unlikely(ret & (VM_FAULT_ERROR | VM_FAULT_NOPAGE | VM_FAULT_RETRY))) return ret; @@ -3010,6 +3013,7 @@ static int do_cow_fault(struct mm_struct *mm, struct vm_area_struct *vma, pgoff_t pgoff, unsigned int flags, pte_t orig_pte) { struct page *fault_page, *new_page; + void *fault_entry; struct mem_cgroup *memcg; spinlock_t *ptl; pte_t *pte; @@ -3027,26 +3031,24 @@ static int do_cow_fault(struct mm_struct *mm, struct vm_area_struct *vma, return VM_FAULT_OOM; } - ret = __do_fault(vma, address, pgoff, flags, new_page, &fault_page); + ret = __do_fault(vma, address, pgoff, flags, new_page, &fault_page, + &fault_entry); if (unlikely(ret & (VM_FAULT_ERROR | VM_FAULT_NOPAGE | VM_FAULT_RETRY))) goto uncharge_out; - if (fault_page) + if (!(ret & VM_FAULT_DAX_LOCKED)) copy_user_highpage(new_page, fault_page, address, vma); __SetPageUptodate(new_page); pte = pte_offset_map_lock(mm, pmd, address, &ptl); if (unlikely(!pte_same(*pte, orig_pte))) { pte_unmap_unlock(pte, ptl); - if (fault_page) { + if (!(ret & VM_FAULT_DAX_LOCKED)) { unlock_page(fault_page); put_page(fault_page); } else { - /* - * The fault handler has no page to lock, so it holds - * i_mmap_lock for read to protect against truncate. - */ - i_mmap_unlock_read(vma->vm_file->f_mapping); + dax_unlock_mapping_entry(vma->vm_file->f_mapping, + pgoff); } goto uncharge_out; } @@ -3054,15 +3056,11 @@ static int do_cow_fault(struct mm_struct *mm, struct vm_area_struct *vma, mem_cgroup_commit_charge(new_page, memcg, false, false); lru_cache_add_active_or_unevictable(new_page, vma); pte_unmap_unlock(pte, ptl); - if (fault_page) { + if (!(ret & VM_FAULT_DAX_LOCKED)) { unlock_page(fault_page); put_page(fault_page); } else { - /* - * The fault handler has no page to lock, so it holds - * i_mmap_lock for read to protect against truncate. - */ - i_mmap_unlock_read(vma->vm_file->f_mapping); + dax_unlock_mapping_entry(vma->vm_file->f_mapping, pgoff); } return ret; uncharge_out: @@ -3082,7 +3080,7 @@ static int do_shared_fault(struct mm_struct *mm, struct vm_area_struct *vma, int dirtied = 0; int ret, tmp; - ret = __do_fault(vma, address, pgoff, flags, NULL, &fault_page); + ret = __do_fault(vma, address, pgoff, flags, NULL, &fault_page, NULL); if (unlikely(ret & (VM_FAULT_ERROR | VM_FAULT_NOPAGE | VM_FAULT_RETRY))) return ret; -- cgit v1.3-8-gc7d7 From 4d9a2c8746671efbb0c27d3ae28c7474597a7aad Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Thu, 12 May 2016 18:29:20 +0200 Subject: dax: Remove i_mmap_lock protection Currently faults are protected against truncate by filesystem specific i_mmap_sem and page lock in case of hole page. Cow faults are protected DAX radix tree entry locking. So there's no need for i_mmap_lock in DAX code. Remove it. Reviewed-by: Ross Zwisler Signed-off-by: Jan Kara Signed-off-by: Ross Zwisler --- fs/dax.c | 24 +++++------------------- mm/memory.c | 2 -- 2 files changed, 5 insertions(+), 21 deletions(-) (limited to 'fs') diff --git a/fs/dax.c b/fs/dax.c index be74635e05a6..6dbe6021cab7 100644 --- a/fs/dax.c +++ b/fs/dax.c @@ -798,29 +798,19 @@ static int dax_insert_mapping(struct address_space *mapping, .sector = to_sector(bh, mapping->host), .size = bh->b_size, }; - int error; void *ret; void *entry = *entryp; - i_mmap_lock_read(mapping); - - if (dax_map_atomic(bdev, &dax) < 0) { - error = PTR_ERR(dax.addr); - goto out; - } + if (dax_map_atomic(bdev, &dax) < 0) + return PTR_ERR(dax.addr); dax_unmap_atomic(bdev, &dax); ret = dax_insert_mapping_entry(mapping, vmf, entry, dax.sector); - if (IS_ERR(ret)) { - error = PTR_ERR(ret); - goto out; - } + if (IS_ERR(ret)) + return PTR_ERR(ret); *entryp = ret; - error = vm_insert_mixed(vma, vaddr, dax.pfn); - out: - i_mmap_unlock_read(mapping); - return error; + return vm_insert_mixed(vma, vaddr, dax.pfn); } /** @@ -1058,8 +1048,6 @@ int __dax_pmd_fault(struct vm_area_struct *vma, unsigned long address, truncate_pagecache_range(inode, lstart, lend); } - i_mmap_lock_read(mapping); - if (!write && !buffer_mapped(&bh)) { spinlock_t *ptl; pmd_t entry; @@ -1148,8 +1136,6 @@ int __dax_pmd_fault(struct vm_area_struct *vma, unsigned long address, } out: - i_mmap_unlock_read(mapping); - return result; fallback: diff --git a/mm/memory.c b/mm/memory.c index f09cdb8d48fa..06f552504e79 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -2453,8 +2453,6 @@ void unmap_mapping_range(struct address_space *mapping, if (details.last_index < details.first_index) details.last_index = ULONG_MAX; - - /* DAX uses i_mmap_lock to serialise file truncate vs page fault */ i_mmap_lock_write(mapping); if (unlikely(!RB_EMPTY_ROOT(&mapping->i_mmap))) unmap_mapping_range_tree(&mapping->i_mmap, &details); -- cgit v1.3-8-gc7d7 From 690a7871225b7f91f8fe13c465a79e183a61cfe3 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 20 May 2016 10:29:15 +1000 Subject: xfs: fix warning in xfs_finish_page_writeback for non-debug builds blockmask is unused if ASSERTs are disabled. Reported-by: kbuild test robot Signed-off-by: Christoph Hellwig --- fs/xfs/xfs_aops.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c index b5f1c66bbb58..736461a861b1 100644 --- a/fs/xfs/xfs_aops.c +++ b/fs/xfs/xfs_aops.c @@ -94,15 +94,14 @@ xfs_finish_page_writeback( struct bio_vec *bvec, int error) { - unsigned int blockmask = (1 << inode->i_blkbits) - 1; unsigned int end = bvec->bv_offset + bvec->bv_len - 1; struct buffer_head *head, *bh; unsigned int off = 0; ASSERT(bvec->bv_offset < PAGE_SIZE); - ASSERT((bvec->bv_offset & blockmask) == 0); + ASSERT((bvec->bv_offset & ((1 << inode->i_blkbits) - 1)) == 0); ASSERT(end < PAGE_SIZE); - ASSERT((bvec->bv_len & blockmask) == 0); + ASSERT((bvec->bv_len & ((1 << inode->i_blkbits) - 1)) == 0); bh = head = page_buffers(bvec->bv_page); -- cgit v1.3-8-gc7d7 From 35e481761cdc688dbee0ef552a13f49af8eba6cc Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Thu, 19 May 2016 17:08:59 -0700 Subject: fsnotify: avoid spurious EMFILE errors from inotify_init() Inotify instance is destroyed when all references to it are dropped. That not only means that the corresponding file descriptor needs to be closed but also that all corresponding instance marks are freed (as each mark holds a reference to the inotify instance). However marks are freed only after SRCU period ends which can take some time and thus if user rapidly creates and frees inotify instances, number of existing inotify instances can exceed max_user_instances limit although from user point of view there is always at most one existing instance. Thus inotify_init() returns EMFILE error which is hard to justify from user point of view. This problem is exposed by LTP inotify06 testcase on some machines. We fix the problem by making sure all group marks are properly freed while destroying inotify instance. We wait for SRCU period to end in that path anyway since we have to make sure there is no event being added to the instance while we are tearing down the instance. So it takes only some plumbing to allow for marks to be destroyed in that path as well and not from a dedicated work item. [akpm@linux-foundation.org: coding-style fixes] Signed-off-by: Jan Kara Reported-by: Xiaoguang Wang Tested-by: Xiaoguang Wang Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/notify/fsnotify.h | 7 ++++ fs/notify/group.c | 17 ++++++--- fs/notify/mark.c | 78 +++++++++++++++++++++++++++++++--------- include/linux/fsnotify_backend.h | 2 -- 4 files changed, 81 insertions(+), 23 deletions(-) (limited to 'fs') diff --git a/fs/notify/fsnotify.h b/fs/notify/fsnotify.h index b44c68a857e7..0a3bc2cf192c 100644 --- a/fs/notify/fsnotify.h +++ b/fs/notify/fsnotify.h @@ -56,6 +56,13 @@ static inline void fsnotify_clear_marks_by_mount(struct vfsmount *mnt) fsnotify_destroy_marks(&real_mount(mnt)->mnt_fsnotify_marks, &mnt->mnt_root->d_lock); } +/* prepare for freeing all marks associated with given group */ +extern void fsnotify_detach_group_marks(struct fsnotify_group *group); +/* + * wait for fsnotify_mark_srcu period to end and free all marks in destroy_list + */ +extern void fsnotify_mark_destroy_list(void); + /* * update the dentry->d_flags of all of inode's children to indicate if inode cares * about events that happen to its children. diff --git a/fs/notify/group.c b/fs/notify/group.c index d16b62cb2854..3e2dd85be5dd 100644 --- a/fs/notify/group.c +++ b/fs/notify/group.c @@ -47,12 +47,21 @@ static void fsnotify_final_destroy_group(struct fsnotify_group *group) */ void fsnotify_destroy_group(struct fsnotify_group *group) { - /* clear all inode marks for this group */ - fsnotify_clear_marks_by_group(group); + /* clear all inode marks for this group, attach them to destroy_list */ + fsnotify_detach_group_marks(group); - synchronize_srcu(&fsnotify_mark_srcu); + /* + * Wait for fsnotify_mark_srcu period to end and free all marks in + * destroy_list + */ + fsnotify_mark_destroy_list(); - /* clear the notification queue of all events */ + /* + * Since we have waited for fsnotify_mark_srcu in + * fsnotify_mark_destroy_list() there can be no outstanding event + * notification against this group. So clearing the notification queue + * of all events is reliable now. + */ fsnotify_flush_notify(group); /* diff --git a/fs/notify/mark.c b/fs/notify/mark.c index 7115c5d7d373..d3fea0bd89e2 100644 --- a/fs/notify/mark.c +++ b/fs/notify/mark.c @@ -97,8 +97,8 @@ struct srcu_struct fsnotify_mark_srcu; static DEFINE_SPINLOCK(destroy_lock); static LIST_HEAD(destroy_list); -static void fsnotify_mark_destroy(struct work_struct *work); -static DECLARE_DELAYED_WORK(reaper_work, fsnotify_mark_destroy); +static void fsnotify_mark_destroy_workfn(struct work_struct *work); +static DECLARE_DELAYED_WORK(reaper_work, fsnotify_mark_destroy_workfn); void fsnotify_get_mark(struct fsnotify_mark *mark) { @@ -173,11 +173,15 @@ void fsnotify_detach_mark(struct fsnotify_mark *mark) } /* - * Free fsnotify mark. The freeing is actually happening from a kthread which - * first waits for srcu period end. Caller must have a reference to the mark - * or be protected by fsnotify_mark_srcu. + * Prepare mark for freeing and add it to the list of marks prepared for + * freeing. The actual freeing must happen after SRCU period ends and the + * caller is responsible for this. + * + * The function returns true if the mark was added to the list of marks for + * freeing. The function returns false if someone else has already called + * __fsnotify_free_mark() for the mark. */ -void fsnotify_free_mark(struct fsnotify_mark *mark) +static bool __fsnotify_free_mark(struct fsnotify_mark *mark) { struct fsnotify_group *group = mark->group; @@ -185,17 +189,11 @@ void fsnotify_free_mark(struct fsnotify_mark *mark) /* something else already called this function on this mark */ if (!(mark->flags & FSNOTIFY_MARK_FLAG_ALIVE)) { spin_unlock(&mark->lock); - return; + return false; } mark->flags &= ~FSNOTIFY_MARK_FLAG_ALIVE; spin_unlock(&mark->lock); - spin_lock(&destroy_lock); - list_add(&mark->g_list, &destroy_list); - spin_unlock(&destroy_lock); - queue_delayed_work(system_unbound_wq, &reaper_work, - FSNOTIFY_REAPER_DELAY); - /* * Some groups like to know that marks are being freed. This is a * callback to the group function to let it know that this mark @@ -203,6 +201,25 @@ void fsnotify_free_mark(struct fsnotify_mark *mark) */ if (group->ops->freeing_mark) group->ops->freeing_mark(mark, group); + + spin_lock(&destroy_lock); + list_add(&mark->g_list, &destroy_list); + spin_unlock(&destroy_lock); + + return true; +} + +/* + * Free fsnotify mark. The freeing is actually happening from a workqueue which + * first waits for srcu period end. Caller must have a reference to the mark + * or be protected by fsnotify_mark_srcu. + */ +void fsnotify_free_mark(struct fsnotify_mark *mark) +{ + if (__fsnotify_free_mark(mark)) { + queue_delayed_work(system_unbound_wq, &reaper_work, + FSNOTIFY_REAPER_DELAY); + } } void fsnotify_destroy_mark(struct fsnotify_mark *mark, @@ -468,11 +485,29 @@ void fsnotify_clear_marks_by_group_flags(struct fsnotify_group *group, } /* - * Given a group, destroy all of the marks associated with that group. + * Given a group, prepare for freeing all the marks associated with that group. + * The marks are attached to the list of marks prepared for destruction, the + * caller is responsible for freeing marks in that list after SRCU period has + * ended. */ -void fsnotify_clear_marks_by_group(struct fsnotify_group *group) +void fsnotify_detach_group_marks(struct fsnotify_group *group) { - fsnotify_clear_marks_by_group_flags(group, (unsigned int)-1); + struct fsnotify_mark *mark; + + while (1) { + mutex_lock_nested(&group->mark_mutex, SINGLE_DEPTH_NESTING); + if (list_empty(&group->marks_list)) { + mutex_unlock(&group->mark_mutex); + break; + } + mark = list_first_entry(&group->marks_list, + struct fsnotify_mark, g_list); + fsnotify_get_mark(mark); + fsnotify_detach_mark(mark); + mutex_unlock(&group->mark_mutex); + __fsnotify_free_mark(mark); + fsnotify_put_mark(mark); + } } void fsnotify_duplicate_mark(struct fsnotify_mark *new, struct fsnotify_mark *old) @@ -499,7 +534,11 @@ void fsnotify_init_mark(struct fsnotify_mark *mark, mark->free_mark = free_mark; } -static void fsnotify_mark_destroy(struct work_struct *work) +/* + * Destroy all marks in destroy_list, waits for SRCU period to finish before + * actually freeing marks. + */ +void fsnotify_mark_destroy_list(void) { struct fsnotify_mark *mark, *next; struct list_head private_destroy_list; @@ -516,3 +555,8 @@ static void fsnotify_mark_destroy(struct work_struct *work) fsnotify_put_mark(mark); } } + +static void fsnotify_mark_destroy_workfn(struct work_struct *work) +{ + fsnotify_mark_destroy_list(); +} diff --git a/include/linux/fsnotify_backend.h b/include/linux/fsnotify_backend.h index 1259e53d9296..29f917517299 100644 --- a/include/linux/fsnotify_backend.h +++ b/include/linux/fsnotify_backend.h @@ -359,8 +359,6 @@ extern void fsnotify_clear_vfsmount_marks_by_group(struct fsnotify_group *group) extern void fsnotify_clear_inode_marks_by_group(struct fsnotify_group *group); /* run all the marks in a group, and clear all of the marks where mark->flags & flags is true*/ extern void fsnotify_clear_marks_by_group_flags(struct fsnotify_group *group, unsigned int flags); -/* run all the marks in a group, and flag them to be freed */ -extern void fsnotify_clear_marks_by_group(struct fsnotify_group *group); extern void fsnotify_get_mark(struct fsnotify_mark *mark); extern void fsnotify_put_mark(struct fsnotify_mark *mark); extern void fsnotify_unmount_inodes(struct super_block *sb); -- cgit v1.3-8-gc7d7 From 766b9f928bd5b9b185d986d40355d1f143484136 Mon Sep 17 00:00:00 2001 From: Deepa Dinamani Date: Thu, 19 May 2016 17:09:05 -0700 Subject: fs: poll/select/recvmmsg: use timespec64 for timeout events struct timespec is not y2038 safe. Even though timespec might be sufficient to represent timeouts, use struct timespec64 here as the plan is to get rid of all timespec reference in the kernel. The patch transitions the common functions: poll_select_set_timeout() and select_estimate_accuracy() to use timespec64. And, all the syscalls that use these functions are transitioned in the same patch. The restart block parameters for poll uses monotonic time. Use timespec64 here as well to assign timeout value. This parameter in the restart block need not change because this only holds the monotonic timestamp at which timeout should occur. And, unsigned long data type should be big enough for this timestamp. The system call interfaces will be handled in a separate series. Compat interfaces need not change as timespec64 is an alias to struct timespec on a 64 bit system. Link: http://lkml.kernel.org/r/1461947989-21926-3-git-send-email-deepa.kernel@gmail.com Signed-off-by: Deepa Dinamani Acked-by: John Stultz Acked-by: David S. Miller Cc: Alexander Viro Cc: Arnd Bergmann Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/eventpoll.c | 12 +++++----- fs/select.c | 67 +++++++++++++++++++++++++++++----------------------- include/linux/poll.h | 11 +++++---- net/socket.c | 8 ++++--- 4 files changed, 54 insertions(+), 44 deletions(-) (limited to 'fs') diff --git a/fs/eventpoll.c b/fs/eventpoll.c index 8a74a2a52e0f..10db91218933 100644 --- a/fs/eventpoll.c +++ b/fs/eventpoll.c @@ -1583,15 +1583,15 @@ static int ep_send_events(struct eventpoll *ep, return ep_scan_ready_list(ep, ep_send_events_proc, &esed, 0, false); } -static inline struct timespec ep_set_mstimeout(long ms) +static inline struct timespec64 ep_set_mstimeout(long ms) { - struct timespec now, ts = { + struct timespec64 now, ts = { .tv_sec = ms / MSEC_PER_SEC, .tv_nsec = NSEC_PER_MSEC * (ms % MSEC_PER_SEC), }; - ktime_get_ts(&now); - return timespec_add_safe(now, ts); + ktime_get_ts64(&now); + return timespec64_add_safe(now, ts); } /** @@ -1621,11 +1621,11 @@ static int ep_poll(struct eventpoll *ep, struct epoll_event __user *events, ktime_t expires, *to = NULL; if (timeout > 0) { - struct timespec end_time = ep_set_mstimeout(timeout); + struct timespec64 end_time = ep_set_mstimeout(timeout); slack = select_estimate_accuracy(&end_time); to = &expires; - *to = timespec_to_ktime(end_time); + *to = timespec64_to_ktime(end_time); } else if (timeout == 0) { /* * Avoid the unnecessary trip to the wait queue loop, if the diff --git a/fs/select.c b/fs/select.c index 869293988c2a..8ed9da50896a 100644 --- a/fs/select.c +++ b/fs/select.c @@ -47,7 +47,7 @@ #define MAX_SLACK (100 * NSEC_PER_MSEC) -static long __estimate_accuracy(struct timespec *tv) +static long __estimate_accuracy(struct timespec64 *tv) { long slack; int divfactor = 1000; @@ -70,10 +70,10 @@ static long __estimate_accuracy(struct timespec *tv) return slack; } -u64 select_estimate_accuracy(struct timespec *tv) +u64 select_estimate_accuracy(struct timespec64 *tv) { u64 ret; - struct timespec now; + struct timespec64 now; /* * Realtime tasks get a slack of 0 for obvious reasons. @@ -82,8 +82,8 @@ u64 select_estimate_accuracy(struct timespec *tv) if (rt_task(current)) return 0; - ktime_get_ts(&now); - now = timespec_sub(*tv, now); + ktime_get_ts64(&now); + now = timespec64_sub(*tv, now); ret = __estimate_accuracy(&now); if (ret < current->timer_slack_ns) return current->timer_slack_ns; @@ -260,7 +260,7 @@ EXPORT_SYMBOL(poll_schedule_timeout); /** * poll_select_set_timeout - helper function to setup the timeout value - * @to: pointer to timespec variable for the final timeout + * @to: pointer to timespec64 variable for the final timeout * @sec: seconds (from user space) * @nsec: nanoseconds (from user space) * @@ -269,26 +269,28 @@ EXPORT_SYMBOL(poll_schedule_timeout); * * Returns -EINVAL if sec/nsec are not normalized. Otherwise 0. */ -int poll_select_set_timeout(struct timespec *to, long sec, long nsec) +int poll_select_set_timeout(struct timespec64 *to, time64_t sec, long nsec) { - struct timespec ts = {.tv_sec = sec, .tv_nsec = nsec}; + struct timespec64 ts = {.tv_sec = sec, .tv_nsec = nsec}; - if (!timespec_valid(&ts)) + if (!timespec64_valid(&ts)) return -EINVAL; /* Optimize for the zero timeout value here */ if (!sec && !nsec) { to->tv_sec = to->tv_nsec = 0; } else { - ktime_get_ts(to); - *to = timespec_add_safe(*to, ts); + ktime_get_ts64(to); + *to = timespec64_add_safe(*to, ts); } return 0; } -static int poll_select_copy_remaining(struct timespec *end_time, void __user *p, +static int poll_select_copy_remaining(struct timespec64 *end_time, + void __user *p, int timeval, int ret) { + struct timespec64 rts64; struct timespec rts; struct timeval rtv; @@ -302,16 +304,18 @@ static int poll_select_copy_remaining(struct timespec *end_time, void __user *p, if (!end_time->tv_sec && !end_time->tv_nsec) return ret; - ktime_get_ts(&rts); - rts = timespec_sub(*end_time, rts); - if (rts.tv_sec < 0) - rts.tv_sec = rts.tv_nsec = 0; + ktime_get_ts64(&rts64); + rts64 = timespec64_sub(*end_time, rts64); + if (rts64.tv_sec < 0) + rts64.tv_sec = rts64.tv_nsec = 0; + + rts = timespec64_to_timespec(rts64); if (timeval) { if (sizeof(rtv) > sizeof(rtv.tv_sec) + sizeof(rtv.tv_usec)) memset(&rtv, 0, sizeof(rtv)); - rtv.tv_sec = rts.tv_sec; - rtv.tv_usec = rts.tv_nsec / NSEC_PER_USEC; + rtv.tv_sec = rts64.tv_sec; + rtv.tv_usec = rts64.tv_nsec / NSEC_PER_USEC; if (!copy_to_user(p, &rtv, sizeof(rtv))) return ret; @@ -396,7 +400,7 @@ static inline void wait_key_set(poll_table *wait, unsigned long in, wait->_key |= POLLOUT_SET; } -int do_select(int n, fd_set_bits *fds, struct timespec *end_time) +int do_select(int n, fd_set_bits *fds, struct timespec64 *end_time) { ktime_t expire, *to = NULL; struct poll_wqueues table; @@ -522,7 +526,7 @@ int do_select(int n, fd_set_bits *fds, struct timespec *end_time) * pointer to the expiry value. */ if (end_time && !to) { - expire = timespec_to_ktime(*end_time); + expire = timespec64_to_ktime(*end_time); to = &expire; } @@ -545,7 +549,7 @@ int do_select(int n, fd_set_bits *fds, struct timespec *end_time) * I'm trying ERESTARTNOHAND which restart only when you want to. */ int core_sys_select(int n, fd_set __user *inp, fd_set __user *outp, - fd_set __user *exp, struct timespec *end_time) + fd_set __user *exp, struct timespec64 *end_time) { fd_set_bits fds; void *bits; @@ -622,7 +626,7 @@ out_nofds: SYSCALL_DEFINE5(select, int, n, fd_set __user *, inp, fd_set __user *, outp, fd_set __user *, exp, struct timeval __user *, tvp) { - struct timespec end_time, *to = NULL; + struct timespec64 end_time, *to = NULL; struct timeval tv; int ret; @@ -648,15 +652,17 @@ static long do_pselect(int n, fd_set __user *inp, fd_set __user *outp, const sigset_t __user *sigmask, size_t sigsetsize) { sigset_t ksigmask, sigsaved; - struct timespec ts, end_time, *to = NULL; + struct timespec ts; + struct timespec64 ts64, end_time, *to = NULL; int ret; if (tsp) { if (copy_from_user(&ts, tsp, sizeof(ts))) return -EFAULT; + ts64 = timespec_to_timespec64(ts); to = &end_time; - if (poll_select_set_timeout(to, ts.tv_sec, ts.tv_nsec)) + if (poll_select_set_timeout(to, ts64.tv_sec, ts64.tv_nsec)) return -EINVAL; } @@ -779,7 +785,7 @@ static inline unsigned int do_pollfd(struct pollfd *pollfd, poll_table *pwait, } static int do_poll(struct poll_list *list, struct poll_wqueues *wait, - struct timespec *end_time) + struct timespec64 *end_time) { poll_table* pt = &wait->pt; ktime_t expire, *to = NULL; @@ -854,7 +860,7 @@ static int do_poll(struct poll_list *list, struct poll_wqueues *wait, * pointer to the expiry value. */ if (end_time && !to) { - expire = timespec_to_ktime(*end_time); + expire = timespec64_to_ktime(*end_time); to = &expire; } @@ -868,7 +874,7 @@ static int do_poll(struct poll_list *list, struct poll_wqueues *wait, sizeof(struct pollfd)) int do_sys_poll(struct pollfd __user *ufds, unsigned int nfds, - struct timespec *end_time) + struct timespec64 *end_time) { struct poll_wqueues table; int err = -EFAULT, fdcount, len, size; @@ -936,7 +942,7 @@ static long do_restart_poll(struct restart_block *restart_block) { struct pollfd __user *ufds = restart_block->poll.ufds; int nfds = restart_block->poll.nfds; - struct timespec *to = NULL, end_time; + struct timespec64 *to = NULL, end_time; int ret; if (restart_block->poll.has_timeout) { @@ -957,7 +963,7 @@ static long do_restart_poll(struct restart_block *restart_block) SYSCALL_DEFINE3(poll, struct pollfd __user *, ufds, unsigned int, nfds, int, timeout_msecs) { - struct timespec end_time, *to = NULL; + struct timespec64 end_time, *to = NULL; int ret; if (timeout_msecs >= 0) { @@ -993,7 +999,8 @@ SYSCALL_DEFINE5(ppoll, struct pollfd __user *, ufds, unsigned int, nfds, size_t, sigsetsize) { sigset_t ksigmask, sigsaved; - struct timespec ts, end_time, *to = NULL; + struct timespec ts; + struct timespec64 end_time, *to = NULL; int ret; if (tsp) { diff --git a/include/linux/poll.h b/include/linux/poll.h index 9fb4f40d9a26..37b057b63b46 100644 --- a/include/linux/poll.h +++ b/include/linux/poll.h @@ -96,7 +96,7 @@ extern void poll_initwait(struct poll_wqueues *pwq); extern void poll_freewait(struct poll_wqueues *pwq); extern int poll_schedule_timeout(struct poll_wqueues *pwq, int state, ktime_t *expires, unsigned long slack); -extern u64 select_estimate_accuracy(struct timespec *tv); +extern u64 select_estimate_accuracy(struct timespec64 *tv); static inline int poll_schedule(struct poll_wqueues *pwq, int state) @@ -153,12 +153,13 @@ void zero_fd_set(unsigned long nr, unsigned long *fdset) #define MAX_INT64_SECONDS (((s64)(~((u64)0)>>1)/HZ)-1) -extern int do_select(int n, fd_set_bits *fds, struct timespec *end_time); +extern int do_select(int n, fd_set_bits *fds, struct timespec64 *end_time); extern int do_sys_poll(struct pollfd __user * ufds, unsigned int nfds, - struct timespec *end_time); + struct timespec64 *end_time); extern int core_sys_select(int n, fd_set __user *inp, fd_set __user *outp, - fd_set __user *exp, struct timespec *end_time); + fd_set __user *exp, struct timespec64 *end_time); -extern int poll_select_set_timeout(struct timespec *to, long sec, long nsec); +extern int poll_select_set_timeout(struct timespec64 *to, time64_t sec, + long nsec); #endif /* _LINUX_POLL_H */ diff --git a/net/socket.c b/net/socket.c index e7793f5601ae..a1bd16106625 100644 --- a/net/socket.c +++ b/net/socket.c @@ -2168,7 +2168,8 @@ int __sys_recvmmsg(int fd, struct mmsghdr __user *mmsg, unsigned int vlen, struct mmsghdr __user *entry; struct compat_mmsghdr __user *compat_entry; struct msghdr msg_sys; - struct timespec end_time; + struct timespec64 end_time; + struct timespec64 timeout64; if (timeout && poll_select_set_timeout(&end_time, timeout->tv_sec, @@ -2220,8 +2221,9 @@ int __sys_recvmmsg(int fd, struct mmsghdr __user *mmsg, unsigned int vlen, flags |= MSG_DONTWAIT; if (timeout) { - ktime_get_ts(timeout); - *timeout = timespec_sub(end_time, *timeout); + ktime_get_ts64(&timeout64); + *timeout = timespec64_to_timespec( + timespec64_sub(end_time, timeout64)); if (timeout->tv_sec < 0) { timeout->tv_sec = timeout->tv_nsec = 0; break; -- cgit v1.3-8-gc7d7 From 8ba442214c332a2a10ac90bc24ebb00aea4ae0ec Mon Sep 17 00:00:00 2001 From: Guozhonghua Date: Thu, 19 May 2016 17:09:44 -0700 Subject: ocfs2: fix comment in struct ocfs2_extended_slot The comment in ocfs2_extended_slot has the offset wrong. Cc: Mark Fasheh Cc: Joel Becker Cc: Junxiao Bi Cc: Joseph Qi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ocfs2/ocfs2_fs.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ocfs2/ocfs2_fs.h b/fs/ocfs2/ocfs2_fs.h index 540ab5b75dbb..44d178b8d1aa 100644 --- a/fs/ocfs2/ocfs2_fs.h +++ b/fs/ocfs2/ocfs2_fs.h @@ -580,7 +580,7 @@ struct ocfs2_extended_slot { /*00*/ __u8 es_valid; __u8 es_reserved1[3]; __le32 es_node_num; -/*10*/ +/*08*/ }; /* -- cgit v1.3-8-gc7d7 From c14688ea248c5ddfb6b8b97151de0a8620ba4a68 Mon Sep 17 00:00:00 2001 From: piaojun Date: Thu, 19 May 2016 17:09:47 -0700 Subject: ocfs2: clean up an unused variable 'wants_rotate' in ocfs2_truncate_rec Clean up an unused variable 'wants_rotate' in ocfs2_truncate_rec. Signed-off-by: Jun Piao Reviewed-by: Joseph Qi Cc: Mark Fasheh Cc: Joel Becker Cc: Junxiao Bi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ocfs2/alloc.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c index e361d1a0ca09..460c0cedab3a 100644 --- a/fs/ocfs2/alloc.c +++ b/fs/ocfs2/alloc.c @@ -5351,7 +5351,7 @@ static int ocfs2_truncate_rec(handle_t *handle, { int ret; u32 left_cpos, rec_range, trunc_range; - int wants_rotate = 0, is_rightmost_tree_rec = 0; + int is_rightmost_tree_rec = 0; struct super_block *sb = ocfs2_metadata_cache_get_super(et->et_ci); struct ocfs2_path *left_path = NULL; struct ocfs2_extent_list *el = path_leaf_el(path); @@ -5457,7 +5457,6 @@ static int ocfs2_truncate_rec(handle_t *handle, memset(rec, 0, sizeof(*rec)); ocfs2_cleanup_merge(el, index); - wants_rotate = 1; next_free = le16_to_cpu(el->l_next_free_rec); if (is_rightmost_tree_rec && next_free > 1) { -- cgit v1.3-8-gc7d7 From aa6913dbd26af386e7703f6c0d4ed2a848b7a052 Mon Sep 17 00:00:00 2001 From: Jun Piao Date: Thu, 19 May 2016 17:09:50 -0700 Subject: ocfs2: clean up unused parameter 'count' in o2hb_read_block_input() Clean up unused parameter 'count' in o2hb_read_block_input(). Signed-off-by: Jun Piao Reviewed-by: Joseph Qi Cc: Mark Fasheh Cc: Joel Becker Cc: Junxiao Bi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ocfs2/cluster/heartbeat.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/cluster/heartbeat.c b/fs/ocfs2/cluster/heartbeat.c index 1934abb6b680..a8d15beee5cb 100644 --- a/fs/ocfs2/cluster/heartbeat.c +++ b/fs/ocfs2/cluster/heartbeat.c @@ -1456,7 +1456,6 @@ static void o2hb_region_release(struct config_item *item) static int o2hb_read_block_input(struct o2hb_region *reg, const char *page, - size_t count, unsigned long *ret_bytes, unsigned int *ret_bits) { @@ -1499,8 +1498,8 @@ static ssize_t o2hb_region_block_bytes_store(struct config_item *item, if (reg->hr_bdev) return -EINVAL; - status = o2hb_read_block_input(reg, page, count, - &block_bytes, &block_bits); + status = o2hb_read_block_input(reg, page, &block_bytes, + &block_bits); if (status) return status; -- cgit v1.3-8-gc7d7 From 8f9b1802c20ace68a7d6603e1fd82a44fedb4078 Mon Sep 17 00:00:00 2001 From: Guozhonghua Date: Thu, 19 May 2016 17:09:53 -0700 Subject: ocfs2: clean up an unneeded goto in ocfs2_put_slot() The goto is not useful in ocfs2_put_slot(), so delete it. Signed-off-by: Guozhonghua Cc: Mark Fasheh Cc: Joel Becker Cc: Junxiao Bi Cc: Joseph Qi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ocfs2/slot_map.c | 6 +----- 1 file changed, 1 insertion(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/slot_map.c b/fs/ocfs2/slot_map.c index 1e09592148ad..d7407994f308 100644 --- a/fs/ocfs2/slot_map.c +++ b/fs/ocfs2/slot_map.c @@ -535,12 +535,8 @@ void ocfs2_put_slot(struct ocfs2_super *osb) spin_unlock(&osb->osb_lock); status = ocfs2_update_disk_slot(osb, si, slot_num); - if (status < 0) { + if (status < 0) mlog_errno(status); - goto bail; - } -bail: ocfs2_free_slot_info(osb); } - -- cgit v1.3-8-gc7d7 From 0139aa7b7fa12ceef095d99dc36606a5b10ab83a Mon Sep 17 00:00:00 2001 From: Joonsoo Kim Date: Thu, 19 May 2016 17:10:49 -0700 Subject: mm: rename _count, field of the struct page, to _refcount Many developers already know that field for reference count of the struct page is _count and atomic type. They would try to handle it directly and this could break the purpose of page reference count tracepoint. To prevent direct _count modification, this patch rename it to _refcount and add warning message on the code. After that, developer who need to handle reference count will find that field should not be accessed directly. [akpm@linux-foundation.org: fix comments, per Vlastimil] [akpm@linux-foundation.org: Documentation/vm/transhuge.txt too] [sfr@canb.auug.org.au: sync ethernet driver changes] Signed-off-by: Joonsoo Kim Signed-off-by: Stephen Rothwell Cc: Vlastimil Babka Cc: Hugh Dickins Cc: Johannes Berg Cc: "David S. Miller" Cc: Sunil Goutham Cc: Chris Metcalf Cc: Manish Chopra Cc: Yuval Mintz Cc: Tariq Toukan Cc: Saeed Mahameed Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- Documentation/vm/transhuge.txt | 10 +++++----- arch/tile/mm/init.c | 2 +- drivers/block/aoe/aoecmd.c | 2 +- drivers/hwtracing/intel_th/msu.c | 2 +- drivers/net/ethernet/mellanox/mlx5/core/en_rx.c | 20 +++++++++---------- drivers/net/ethernet/qlogic/qede/qede_main.c | 4 ++-- fs/proc/page.c | 2 +- include/linux/mm.h | 2 +- include/linux/mm_types.h | 14 ++++++++----- include/linux/page_ref.h | 26 ++++++++++++------------- include/linux/pagemap.h | 8 ++++---- kernel/kexec_core.c | 2 +- mm/huge_memory.c | 4 ++-- mm/internal.h | 2 +- mm/page_alloc.c | 4 ++-- mm/slub.c | 4 ++-- mm/vmscan.c | 4 ++-- 17 files changed, 58 insertions(+), 54 deletions(-) (limited to 'fs') diff --git a/Documentation/vm/transhuge.txt b/Documentation/vm/transhuge.txt index d9cb65cf5cfd..fb0e1f2a19cc 100644 --- a/Documentation/vm/transhuge.txt +++ b/Documentation/vm/transhuge.txt @@ -394,9 +394,9 @@ hugepage natively. Once finished you can drop the page table lock. Refcounting on THP is mostly consistent with refcounting on other compound pages: - - get_page()/put_page() and GUP operate in head page's ->_count. + - get_page()/put_page() and GUP operate in head page's ->_refcount. - - ->_count in tail pages is always zero: get_page_unless_zero() never + - ->_refcount in tail pages is always zero: get_page_unless_zero() never succeed on tail pages. - map/unmap of the pages with PTE entry increment/decrement ->_mapcount @@ -426,15 +426,15 @@ requests to split pinned huge page: it expects page count to be equal to sum of mapcount of all sub-pages plus one (split_huge_page caller must have reference for head page). -split_huge_page uses migration entries to stabilize page->_count and +split_huge_page uses migration entries to stabilize page->_refcount and page->_mapcount. We safe against physical memory scanners too: the only legitimate way scanner can get reference to a page is get_page_unless_zero(). -All tail pages has zero ->_count until atomic_add(). It prevent scanner +All tail pages has zero ->_refcount until atomic_add(). It prevent scanner from geting reference to tail page up to the point. After the atomic_add() -we don't care about ->_count value. We already known how many references +we don't care about ->_refcount value. We already known how many references with should uncharge from head page. For head page get_page_unless_zero() will succeed and we don't mind. It's diff --git a/arch/tile/mm/init.c b/arch/tile/mm/init.c index a0582b7f41d3..adce25462b0d 100644 --- a/arch/tile/mm/init.c +++ b/arch/tile/mm/init.c @@ -679,7 +679,7 @@ static void __init init_free_pfn_range(unsigned long start, unsigned long end) * Hacky direct set to avoid unnecessary * lock take/release for EVERY page here. */ - p->_count.counter = 0; + p->_refcount.counter = 0; p->_mapcount.counter = -1; } init_page_count(page); diff --git a/drivers/block/aoe/aoecmd.c b/drivers/block/aoe/aoecmd.c index 437b3a822f44..d597e432e195 100644 --- a/drivers/block/aoe/aoecmd.c +++ b/drivers/block/aoe/aoecmd.c @@ -861,7 +861,7 @@ rqbiocnt(struct request *r) * discussion. * * We cannot use get_page in the workaround, because it insists on a - * positive page count as a precondition. So we use _count directly. + * positive page count as a precondition. So we use _refcount directly. */ static void bio_pageinc(struct bio *bio) diff --git a/drivers/hwtracing/intel_th/msu.c b/drivers/hwtracing/intel_th/msu.c index d9d6022c5aca..d2209147dc89 100644 --- a/drivers/hwtracing/intel_th/msu.c +++ b/drivers/hwtracing/intel_th/msu.c @@ -1164,7 +1164,7 @@ static void msc_mmap_close(struct vm_area_struct *vma) if (!atomic_dec_and_mutex_lock(&msc->mmap_count, &msc->buf_mutex)) return; - /* drop page _counts */ + /* drop page _refcounts */ for (pg = 0; pg < msc->nr_pages; pg++) { struct page *page = msc_buffer_get_page(msc, pg); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c index f3456798c596..bd947704b59c 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c @@ -433,8 +433,8 @@ static int mlx5e_alloc_rx_fragmented_mpwqe(struct mlx5e_rq *rq, for (i = 0; i < MLX5_MPWRQ_PAGES_PER_WQE; i++) { if (unlikely(mlx5e_alloc_and_map_page(rq, wi, i))) goto err_unmap; - atomic_add(mlx5e_mpwqe_strides_per_page(rq), - &wi->umr.dma_info[i].page->_count); + page_ref_add(wi->umr.dma_info[i].page, + mlx5e_mpwqe_strides_per_page(rq)); wi->skbs_frags[i] = 0; } @@ -452,8 +452,8 @@ err_unmap: while (--i >= 0) { dma_unmap_page(rq->pdev, wi->umr.dma_info[i].addr, PAGE_SIZE, PCI_DMA_FROMDEVICE); - atomic_sub(mlx5e_mpwqe_strides_per_page(rq), - &wi->umr.dma_info[i].page->_count); + page_ref_sub(wi->umr.dma_info[i].page, + mlx5e_mpwqe_strides_per_page(rq)); put_page(wi->umr.dma_info[i].page); } dma_unmap_single(rq->pdev, wi->umr.mtt_addr, mtt_sz, PCI_DMA_TODEVICE); @@ -477,8 +477,8 @@ void mlx5e_free_rx_fragmented_mpwqe(struct mlx5e_rq *rq, for (i = 0; i < MLX5_MPWRQ_PAGES_PER_WQE; i++) { dma_unmap_page(rq->pdev, wi->umr.dma_info[i].addr, PAGE_SIZE, PCI_DMA_FROMDEVICE); - atomic_sub(mlx5e_mpwqe_strides_per_page(rq) - wi->skbs_frags[i], - &wi->umr.dma_info[i].page->_count); + page_ref_sub(wi->umr.dma_info[i].page, + mlx5e_mpwqe_strides_per_page(rq) - wi->skbs_frags[i]); put_page(wi->umr.dma_info[i].page); } dma_unmap_single(rq->pdev, wi->umr.mtt_addr, mtt_sz, PCI_DMA_TODEVICE); @@ -527,8 +527,8 @@ static int mlx5e_alloc_rx_linear_mpwqe(struct mlx5e_rq *rq, */ split_page(wi->dma_info.page, MLX5_MPWRQ_WQE_PAGE_ORDER); for (i = 0; i < MLX5_MPWRQ_PAGES_PER_WQE; i++) { - atomic_add(mlx5e_mpwqe_strides_per_page(rq), - &wi->dma_info.page[i]._count); + page_ref_add(&wi->dma_info.page[i], + mlx5e_mpwqe_strides_per_page(rq)); wi->skbs_frags[i] = 0; } @@ -551,8 +551,8 @@ void mlx5e_free_rx_linear_mpwqe(struct mlx5e_rq *rq, dma_unmap_page(rq->pdev, wi->dma_info.addr, rq->wqe_sz, PCI_DMA_FROMDEVICE); for (i = 0; i < MLX5_MPWRQ_PAGES_PER_WQE; i++) { - atomic_sub(mlx5e_mpwqe_strides_per_page(rq) - wi->skbs_frags[i], - &wi->dma_info.page[i]._count); + page_ref_sub(&wi->dma_info.page[i], + mlx5e_mpwqe_strides_per_page(rq) - wi->skbs_frags[i]); put_page(&wi->dma_info.page[i]); } } diff --git a/drivers/net/ethernet/qlogic/qede/qede_main.c b/drivers/net/ethernet/qlogic/qede/qede_main.c index 3aabfc0adefe..73dd525fbf08 100644 --- a/drivers/net/ethernet/qlogic/qede/qede_main.c +++ b/drivers/net/ethernet/qlogic/qede/qede_main.c @@ -1036,7 +1036,7 @@ static int qede_fill_frag_skb(struct qede_dev *edev, /* Incr page ref count to reuse on allocation failure * so that it doesn't get freed while freeing SKB. */ - atomic_inc(¤t_bd->data->_count); + page_ref_inc(current_bd->data); goto out; } @@ -1487,7 +1487,7 @@ static int qede_rx_int(struct qede_fastpath *fp, int budget) * freeing SKB. */ - atomic_inc(&sw_rx_data->data->_count); + page_ref_inc(sw_rx_data->data); rxq->rx_alloc_errors++; qede_recycle_rx_bd_ring(rxq, edev, fp_cqe->bd_num); diff --git a/fs/proc/page.c b/fs/proc/page.c index 712f1b9992cc..3ecd445e830d 100644 --- a/fs/proc/page.c +++ b/fs/proc/page.c @@ -142,7 +142,7 @@ u64 stable_page_flags(struct page *page) /* - * Caveats on high order pages: page->_count will only be set + * Caveats on high order pages: page->_refcount will only be set * -1 on the head page; SLUB/SLQB do the same for PG_slab; * SLOB won't set PG_slab at all on compound pages. */ diff --git a/include/linux/mm.h b/include/linux/mm.h index 727f799757ab..1193a54ea2b3 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -734,7 +734,7 @@ static inline void get_page(struct page *page) page = compound_head(page); /* * Getting a normal page or the head of a compound page - * requires to already have an elevated page->_count. + * requires to already have an elevated page->_refcount. */ VM_BUG_ON_PAGE(page_ref_count(page) <= 0, page); page_ref_inc(page); diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index c2d75b4fa86c..1fda9c99ef95 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -73,9 +73,9 @@ struct page { unsigned long counters; #else /* - * Keep _count separate from slub cmpxchg_double data. - * As the rest of the double word is protected by - * slab_lock but _count is not. + * Keep _refcount separate from slub cmpxchg_double + * data. As the rest of the double word is protected by + * slab_lock but _refcount is not. */ unsigned counters; #endif @@ -97,7 +97,11 @@ struct page { }; int units; /* SLOB */ }; - atomic_t _count; /* Usage count, see below. */ + /* + * Usage count, *USE WRAPPER FUNCTION* + * when manual accounting. See page_ref.h + */ + atomic_t _refcount; }; unsigned int active; /* SLAB */ }; @@ -248,7 +252,7 @@ struct page_frag_cache { __u32 offset; #endif /* we maintain a pagecount bias, so that we dont dirty cache line - * containing page->_count every time we allocate a fragment. + * containing page->_refcount every time we allocate a fragment. */ unsigned int pagecnt_bias; bool pfmemalloc; diff --git a/include/linux/page_ref.h b/include/linux/page_ref.h index e596d5d9540e..8b5e0a9f2431 100644 --- a/include/linux/page_ref.h +++ b/include/linux/page_ref.h @@ -63,17 +63,17 @@ static inline void __page_ref_unfreeze(struct page *page, int v) static inline int page_ref_count(struct page *page) { - return atomic_read(&page->_count); + return atomic_read(&page->_refcount); } static inline int page_count(struct page *page) { - return atomic_read(&compound_head(page)->_count); + return atomic_read(&compound_head(page)->_refcount); } static inline void set_page_count(struct page *page, int v) { - atomic_set(&page->_count, v); + atomic_set(&page->_refcount, v); if (page_ref_tracepoint_active(__tracepoint_page_ref_set)) __page_ref_set(page, v); } @@ -89,35 +89,35 @@ static inline void init_page_count(struct page *page) static inline void page_ref_add(struct page *page, int nr) { - atomic_add(nr, &page->_count); + atomic_add(nr, &page->_refcount); if (page_ref_tracepoint_active(__tracepoint_page_ref_mod)) __page_ref_mod(page, nr); } static inline void page_ref_sub(struct page *page, int nr) { - atomic_sub(nr, &page->_count); + atomic_sub(nr, &page->_refcount); if (page_ref_tracepoint_active(__tracepoint_page_ref_mod)) __page_ref_mod(page, -nr); } static inline void page_ref_inc(struct page *page) { - atomic_inc(&page->_count); + atomic_inc(&page->_refcount); if (page_ref_tracepoint_active(__tracepoint_page_ref_mod)) __page_ref_mod(page, 1); } static inline void page_ref_dec(struct page *page) { - atomic_dec(&page->_count); + atomic_dec(&page->_refcount); if (page_ref_tracepoint_active(__tracepoint_page_ref_mod)) __page_ref_mod(page, -1); } static inline int page_ref_sub_and_test(struct page *page, int nr) { - int ret = atomic_sub_and_test(nr, &page->_count); + int ret = atomic_sub_and_test(nr, &page->_refcount); if (page_ref_tracepoint_active(__tracepoint_page_ref_mod_and_test)) __page_ref_mod_and_test(page, -nr, ret); @@ -126,7 +126,7 @@ static inline int page_ref_sub_and_test(struct page *page, int nr) static inline int page_ref_dec_and_test(struct page *page) { - int ret = atomic_dec_and_test(&page->_count); + int ret = atomic_dec_and_test(&page->_refcount); if (page_ref_tracepoint_active(__tracepoint_page_ref_mod_and_test)) __page_ref_mod_and_test(page, -1, ret); @@ -135,7 +135,7 @@ static inline int page_ref_dec_and_test(struct page *page) static inline int page_ref_dec_return(struct page *page) { - int ret = atomic_dec_return(&page->_count); + int ret = atomic_dec_return(&page->_refcount); if (page_ref_tracepoint_active(__tracepoint_page_ref_mod_and_return)) __page_ref_mod_and_return(page, -1, ret); @@ -144,7 +144,7 @@ static inline int page_ref_dec_return(struct page *page) static inline int page_ref_add_unless(struct page *page, int nr, int u) { - int ret = atomic_add_unless(&page->_count, nr, u); + int ret = atomic_add_unless(&page->_refcount, nr, u); if (page_ref_tracepoint_active(__tracepoint_page_ref_mod_unless)) __page_ref_mod_unless(page, nr, ret); @@ -153,7 +153,7 @@ static inline int page_ref_add_unless(struct page *page, int nr, int u) static inline int page_ref_freeze(struct page *page, int count) { - int ret = likely(atomic_cmpxchg(&page->_count, count, 0) == count); + int ret = likely(atomic_cmpxchg(&page->_refcount, count, 0) == count); if (page_ref_tracepoint_active(__tracepoint_page_ref_freeze)) __page_ref_freeze(page, count, ret); @@ -165,7 +165,7 @@ static inline void page_ref_unfreeze(struct page *page, int count) VM_BUG_ON_PAGE(page_count(page) != 0, page); VM_BUG_ON(count == 0); - atomic_set(&page->_count, count); + atomic_set(&page->_refcount, count); if (page_ref_tracepoint_active(__tracepoint_page_ref_unfreeze)) __page_ref_unfreeze(page, count); } diff --git a/include/linux/pagemap.h b/include/linux/pagemap.h index 7e1ab155c67c..fe1513ffb7bf 100644 --- a/include/linux/pagemap.h +++ b/include/linux/pagemap.h @@ -90,12 +90,12 @@ void release_pages(struct page **pages, int nr, bool cold); /* * speculatively take a reference to a page. - * If the page is free (_count == 0), then _count is untouched, and 0 - * is returned. Otherwise, _count is incremented by 1 and 1 is returned. + * If the page is free (_refcount == 0), then _refcount is untouched, and 0 + * is returned. Otherwise, _refcount is incremented by 1 and 1 is returned. * * This function must be called inside the same rcu_read_lock() section as has * been used to lookup the page in the pagecache radix-tree (or page table): - * this allows allocators to use a synchronize_rcu() to stabilize _count. + * this allows allocators to use a synchronize_rcu() to stabilize _refcount. * * Unless an RCU grace period has passed, the count of all pages coming out * of the allocator must be considered unstable. page_count may return higher @@ -111,7 +111,7 @@ void release_pages(struct page **pages, int nr, bool cold); * 2. conditionally increment refcount * 3. check the page is still in pagecache (if no, goto 1) * - * Remove-side that cares about stability of _count (eg. reclaim) has the + * Remove-side that cares about stability of _refcount (eg. reclaim) has the * following (with tree_lock held for write): * A. atomically check refcount is correct and set it to 0 (atomic_cmpxchg) * B. remove page from pagecache diff --git a/kernel/kexec_core.c b/kernel/kexec_core.c index 1391d3ee3b86..1c03dfb4abfd 100644 --- a/kernel/kexec_core.c +++ b/kernel/kexec_core.c @@ -1410,7 +1410,7 @@ static int __init crash_save_vmcoreinfo_init(void) VMCOREINFO_STRUCT_SIZE(list_head); VMCOREINFO_SIZE(nodemask_t); VMCOREINFO_OFFSET(page, flags); - VMCOREINFO_OFFSET(page, _count); + VMCOREINFO_OFFSET(page, _refcount); VMCOREINFO_OFFSET(page, mapping); VMCOREINFO_OFFSET(page, lru); VMCOREINFO_OFFSET(page, _mapcount); diff --git a/mm/huge_memory.c b/mm/huge_memory.c index b49ee126d4d1..f8ac8f582fd8 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -3113,7 +3113,7 @@ static void __split_huge_page_tail(struct page *head, int tail, VM_BUG_ON_PAGE(page_ref_count(page_tail) != 0, page_tail); /* - * tail_page->_count is zero and not changing from under us. But + * tail_page->_refcount is zero and not changing from under us. But * get_page_unless_zero() may be running from under us on the * tail_page. If we used atomic_set() below instead of atomic_inc(), we * would then run atomic_set() concurrently with @@ -3340,7 +3340,7 @@ int split_huge_page_to_list(struct page *page, struct list_head *list) if (mlocked) lru_add_drain(); - /* Prevent deferred_split_scan() touching ->_count */ + /* Prevent deferred_split_scan() touching ->_refcount */ spin_lock_irqsave(&pgdata->split_queue_lock, flags); count = page_count(head); mapcount = total_mapcount(head); diff --git a/mm/internal.h b/mm/internal.h index b79abb6721cf..098a89e3b97c 100644 --- a/mm/internal.h +++ b/mm/internal.h @@ -58,7 +58,7 @@ static inline unsigned long ra_submit(struct file_ra_state *ra, } /* - * Turn a non-refcounted page (->_count == 0) into refcounted with + * Turn a non-refcounted page (->_refcount == 0) into refcounted with * a count of one. */ static inline void set_page_refcounted(struct page *page) diff --git a/mm/page_alloc.c b/mm/page_alloc.c index c1069efcc4d7..4ce57f938b7f 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -794,7 +794,7 @@ static inline int free_pages_check(struct page *page) if (unlikely(page->mapping != NULL)) bad_reason = "non-NULL mapping"; if (unlikely(page_ref_count(page) != 0)) - bad_reason = "nonzero _count"; + bad_reason = "nonzero _refcount"; if (unlikely(page->flags & PAGE_FLAGS_CHECK_AT_FREE)) { bad_reason = "PAGE_FLAGS_CHECK_AT_FREE flag(s) set"; bad_flags = PAGE_FLAGS_CHECK_AT_FREE; @@ -6864,7 +6864,7 @@ bool has_unmovable_pages(struct zone *zone, struct page *page, int count, * We can't use page_count without pin a page * because another CPU can free compound page. * This check already skips compound tails of THP - * because their page->_count is zero at all time. + * because their page->_refcount is zero at all time. */ if (!page_ref_count(page)) { if (PageBuddy(page)) diff --git a/mm/slub.c b/mm/slub.c index 8671de2e5b12..cf1faa4d3992 100644 --- a/mm/slub.c +++ b/mm/slub.c @@ -329,8 +329,8 @@ static inline void set_page_slub_counters(struct page *page, unsigned long count tmp.counters = counters_new; /* * page->counters can cover frozen/inuse/objects as well - * as page->_count. If we assign to ->counters directly - * we run the risk of losing updates to page->_count, so + * as page->_refcount. If we assign to ->counters directly + * we run the risk of losing updates to page->_refcount, so * be careful and only assign to the fields we need. */ page->frozen = tmp.frozen; diff --git a/mm/vmscan.c b/mm/vmscan.c index 142cb61f4822..d3a02ac3eed7 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -633,7 +633,7 @@ static int __remove_mapping(struct address_space *mapping, struct page *page, * * Reversing the order of the tests ensures such a situation cannot * escape unnoticed. The smp_rmb is needed to ensure the page->flags - * load is not satisfied before that of page->_count. + * load is not satisfied before that of page->_refcount. * * Note that if SetPageDirty is always performed via set_page_dirty, * and thus under tree_lock, then this ordering is not required. @@ -1720,7 +1720,7 @@ shrink_inactive_list(unsigned long nr_to_scan, struct lruvec *lruvec, * It is safe to rely on PG_active against the non-LRU pages in here because * nobody will play with that bit on a non-LRU page. * - * The downside is that we have to touch page->_count against each page. + * The downside is that we have to touch page->_refcount against each page. * But we had to alter page->flags anyway. */ -- cgit v1.3-8-gc7d7 From c33d6c06f60f710f0305ae792773e1c2560e1e51 Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Thu, 19 May 2016 17:14:10 -0700 Subject: mm, page_alloc: avoid looking up the first zone in a zonelist twice The allocator fast path looks up the first usable zone in a zonelist and then get_page_from_freelist does the same job in the zonelist iterator. This patch preserves the necessary information. 4.6.0-rc2 4.6.0-rc2 fastmark-v1r20 initonce-v1r20 Min alloc-odr0-1 364.00 ( 0.00%) 359.00 ( 1.37%) Min alloc-odr0-2 262.00 ( 0.00%) 260.00 ( 0.76%) Min alloc-odr0-4 214.00 ( 0.00%) 214.00 ( 0.00%) Min alloc-odr0-8 186.00 ( 0.00%) 186.00 ( 0.00%) Min alloc-odr0-16 173.00 ( 0.00%) 173.00 ( 0.00%) Min alloc-odr0-32 165.00 ( 0.00%) 165.00 ( 0.00%) Min alloc-odr0-64 161.00 ( 0.00%) 162.00 ( -0.62%) Min alloc-odr0-128 159.00 ( 0.00%) 161.00 ( -1.26%) Min alloc-odr0-256 168.00 ( 0.00%) 170.00 ( -1.19%) Min alloc-odr0-512 180.00 ( 0.00%) 181.00 ( -0.56%) Min alloc-odr0-1024 190.00 ( 0.00%) 190.00 ( 0.00%) Min alloc-odr0-2048 196.00 ( 0.00%) 196.00 ( 0.00%) Min alloc-odr0-4096 202.00 ( 0.00%) 202.00 ( 0.00%) Min alloc-odr0-8192 206.00 ( 0.00%) 205.00 ( 0.49%) Min alloc-odr0-16384 206.00 ( 0.00%) 205.00 ( 0.49%) The benefit is negligible and the results are within the noise but each cycle counts. Signed-off-by: Mel Gorman Cc: Vlastimil Babka Cc: Jesper Dangaard Brouer Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/buffer.c | 10 +++++----- include/linux/mmzone.h | 18 +++++++++++------- mm/internal.h | 2 +- mm/mempolicy.c | 19 ++++++++++--------- mm/page_alloc.c | 34 ++++++++++++++++------------------ 5 files changed, 43 insertions(+), 40 deletions(-) (limited to 'fs') diff --git a/fs/buffer.c b/fs/buffer.c index af0d9a82a8ed..754813a6962b 100644 --- a/fs/buffer.c +++ b/fs/buffer.c @@ -255,17 +255,17 @@ out: */ static void free_more_memory(void) { - struct zone *zone; + struct zoneref *z; int nid; wakeup_flusher_threads(1024, WB_REASON_FREE_MORE_MEM); yield(); for_each_online_node(nid) { - (void)first_zones_zonelist(node_zonelist(nid, GFP_NOFS), - gfp_zone(GFP_NOFS), NULL, - &zone); - if (zone) + + z = first_zones_zonelist(node_zonelist(nid, GFP_NOFS), + gfp_zone(GFP_NOFS), NULL); + if (z->zone) try_to_free_pages(node_zonelist(nid, GFP_NOFS), 0, GFP_NOFS, NULL); } diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h index 327f0fa1e1ce..4b28d2f8125e 100644 --- a/include/linux/mmzone.h +++ b/include/linux/mmzone.h @@ -959,13 +959,10 @@ static __always_inline struct zoneref *next_zones_zonelist(struct zoneref *z, */ static inline struct zoneref *first_zones_zonelist(struct zonelist *zonelist, enum zone_type highest_zoneidx, - nodemask_t *nodes, - struct zone **zone) + nodemask_t *nodes) { - struct zoneref *z = next_zones_zonelist(zonelist->_zonerefs, + return next_zones_zonelist(zonelist->_zonerefs, highest_zoneidx, nodes); - *zone = zonelist_zone(z); - return z; } /** @@ -980,10 +977,17 @@ static inline struct zoneref *first_zones_zonelist(struct zonelist *zonelist, * within a given nodemask */ #define for_each_zone_zonelist_nodemask(zone, z, zlist, highidx, nodemask) \ - for (z = first_zones_zonelist(zlist, highidx, nodemask, &zone); \ + for (z = first_zones_zonelist(zlist, highidx, nodemask), zone = zonelist_zone(z); \ zone; \ z = next_zones_zonelist(++z, highidx, nodemask), \ - zone = zonelist_zone(z)) \ + zone = zonelist_zone(z)) + +#define for_next_zone_zonelist_nodemask(zone, z, zlist, highidx, nodemask) \ + for (zone = z->zone; \ + zone; \ + z = next_zones_zonelist(++z, highidx, nodemask), \ + zone = zonelist_zone(z)) + /** * for_each_zone_zonelist - helper macro to iterate over valid zones in a zonelist at or below a given zone index diff --git a/mm/internal.h b/mm/internal.h index 114593aab55c..d1ddd71c1bbf 100644 --- a/mm/internal.h +++ b/mm/internal.h @@ -102,7 +102,7 @@ extern pmd_t *mm_find_pmd(struct mm_struct *mm, unsigned long address); struct alloc_context { struct zonelist *zonelist; nodemask_t *nodemask; - struct zone *preferred_zone; + struct zoneref *preferred_zoneref; int classzone_idx; int migratetype; enum zone_type high_zoneidx; diff --git a/mm/mempolicy.c b/mm/mempolicy.c index 7f80ebcd6552..297d6854f849 100644 --- a/mm/mempolicy.c +++ b/mm/mempolicy.c @@ -1739,18 +1739,18 @@ unsigned int mempolicy_slab_node(void) return interleave_nodes(policy); case MPOL_BIND: { + struct zoneref *z; + /* * Follow bind policy behavior and start allocation at the * first node. */ struct zonelist *zonelist; - struct zone *zone; enum zone_type highest_zoneidx = gfp_zone(GFP_KERNEL); zonelist = &NODE_DATA(node)->node_zonelists[0]; - (void)first_zones_zonelist(zonelist, highest_zoneidx, - &policy->v.nodes, - &zone); - return zone ? zone->node : node; + z = first_zones_zonelist(zonelist, highest_zoneidx, + &policy->v.nodes); + return z->zone ? z->zone->node : node; } default: @@ -2266,7 +2266,7 @@ static void sp_free(struct sp_node *n) int mpol_misplaced(struct page *page, struct vm_area_struct *vma, unsigned long addr) { struct mempolicy *pol; - struct zone *zone; + struct zoneref *z; int curnid = page_to_nid(page); unsigned long pgoff; int thiscpu = raw_smp_processor_id(); @@ -2298,6 +2298,7 @@ int mpol_misplaced(struct page *page, struct vm_area_struct *vma, unsigned long break; case MPOL_BIND: + /* * allows binding to multiple nodes. * use current page if in policy nodemask, @@ -2306,11 +2307,11 @@ int mpol_misplaced(struct page *page, struct vm_area_struct *vma, unsigned long */ if (node_isset(curnid, pol->v.nodes)) goto out; - (void)first_zones_zonelist( + z = first_zones_zonelist( node_zonelist(numa_node_id(), GFP_HIGHUSER), gfp_zone(GFP_HIGHUSER), - &pol->v.nodes, &zone); - polnid = zone->node; + &pol->v.nodes); + polnid = z->zone->node; break; default: diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 732875b1bdfb..dba8cfd0b2d6 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -2704,7 +2704,7 @@ static struct page * get_page_from_freelist(gfp_t gfp_mask, unsigned int order, int alloc_flags, const struct alloc_context *ac) { - struct zoneref *z; + struct zoneref *z = ac->preferred_zoneref; struct zone *zone; bool fair_skipped = false; bool apply_fair = (alloc_flags & ALLOC_FAIR); @@ -2714,7 +2714,7 @@ zonelist_scan: * Scan zonelist, looking for a zone with enough free. * See also __cpuset_node_allowed() comment in kernel/cpuset.c. */ - for_each_zone_zonelist_nodemask(zone, z, ac->zonelist, ac->high_zoneidx, + for_next_zone_zonelist_nodemask(zone, z, ac->zonelist, ac->high_zoneidx, ac->nodemask) { struct page *page; unsigned long mark; @@ -2734,7 +2734,7 @@ zonelist_scan: fair_skipped = true; continue; } - if (!zone_local(ac->preferred_zone, zone)) { + if (!zone_local(ac->preferred_zoneref->zone, zone)) { if (fair_skipped) goto reset_fair; apply_fair = false; @@ -2780,7 +2780,7 @@ zonelist_scan: goto try_this_zone; if (zone_reclaim_mode == 0 || - !zone_allows_reclaim(ac->preferred_zone, zone)) + !zone_allows_reclaim(ac->preferred_zoneref->zone, zone)) continue; ret = zone_reclaim(zone, gfp_mask, order); @@ -2802,7 +2802,7 @@ zonelist_scan: } try_this_zone: - page = buffered_rmqueue(ac->preferred_zone, zone, order, + page = buffered_rmqueue(ac->preferred_zoneref->zone, zone, order, gfp_mask, alloc_flags, ac->migratetype); if (page) { if (prep_new_page(page, order, gfp_mask, alloc_flags)) @@ -2831,7 +2831,7 @@ try_this_zone: reset_fair: apply_fair = false; fair_skipped = false; - reset_alloc_batches(ac->preferred_zone); + reset_alloc_batches(ac->preferred_zoneref->zone); goto zonelist_scan; } @@ -3114,7 +3114,7 @@ static void wake_all_kswapds(unsigned int order, const struct alloc_context *ac) for_each_zone_zonelist_nodemask(zone, z, ac->zonelist, ac->high_zoneidx, ac->nodemask) - wakeup_kswapd(zone, order, zone_idx(ac->preferred_zone)); + wakeup_kswapd(zone, order, zonelist_zone_idx(ac->preferred_zoneref)); } static inline unsigned int @@ -3332,7 +3332,7 @@ retry: if ((did_some_progress && order <= PAGE_ALLOC_COSTLY_ORDER) || ((gfp_mask & __GFP_REPEAT) && pages_reclaimed < (1 << order))) { /* Wait for some write requests to complete then retry */ - wait_iff_congested(ac->preferred_zone, BLK_RW_ASYNC, HZ/50); + wait_iff_congested(ac->preferred_zoneref->zone, BLK_RW_ASYNC, HZ/50); goto retry; } @@ -3370,7 +3370,6 @@ struct page * __alloc_pages_nodemask(gfp_t gfp_mask, unsigned int order, struct zonelist *zonelist, nodemask_t *nodemask) { - struct zoneref *preferred_zoneref; struct page *page; unsigned int cpuset_mems_cookie; unsigned int alloc_flags = ALLOC_WMARK_LOW|ALLOC_FAIR; @@ -3416,14 +3415,14 @@ retry_cpuset: ac.spread_dirty_pages = (gfp_mask & __GFP_WRITE); /* The preferred zone is used for statistics later */ - preferred_zoneref = first_zones_zonelist(ac.zonelist, ac.high_zoneidx, - ac.nodemask, &ac.preferred_zone); - if (!ac.preferred_zone) { + ac.preferred_zoneref = first_zones_zonelist(ac.zonelist, + ac.high_zoneidx, ac.nodemask); + if (!ac.preferred_zoneref) { page = NULL; goto no_zone; } - ac.classzone_idx = zonelist_zone_idx(preferred_zoneref); + ac.classzone_idx = zonelist_zone_idx(ac.preferred_zoneref); /* First allocation attempt */ page = get_page_from_freelist(alloc_mask, order, alloc_flags, &ac); @@ -4462,13 +4461,12 @@ static void build_zonelists(pg_data_t *pgdat) */ int local_memory_node(int node) { - struct zone *zone; + struct zoneref *z; - (void)first_zones_zonelist(node_zonelist(node, GFP_KERNEL), + z = first_zones_zonelist(node_zonelist(node, GFP_KERNEL), gfp_zone(GFP_KERNEL), - NULL, - &zone); - return zone->node; + NULL); + return z->zone->node; } #endif -- cgit v1.3-8-gc7d7 From b74cb9a80268be5c80cf4c87c74debf0ff2129ac Mon Sep 17 00:00:00 2001 From: Sachin Prabhu Date: Tue, 17 May 2016 18:20:13 -0500 Subject: cifs: Create dedicated keyring for spnego operations The session key is the default keyring set for request_key operations. This session key is revoked when the user owning the session logs out. Any long running daemon processes started by this session ends up with revoked session keyring which prevents these processes from using the request_key mechanism from obtaining the krb5 keys. The problem has been reported by a large number of autofs users. The problem is also seen with multiuser mounts where the share may be used by processes run by a user who has since logged out. A reproducer using automount is available on the Red Hat bz. The patch creates a new keyring which is used to cache cifs spnego upcalls. Red Hat bz: 1267754 Signed-off-by: Sachin Prabhu Reported-by: Scott Mayhew Reviewed-by: Shirish Pargaonkar CC: Stable Signed-off-by: Steve French --- fs/cifs/cifs_spnego.c | 67 +++++++++++++++++++++++++++++++++++++++++++++++++++ fs/cifs/cifsfs.c | 4 +-- fs/cifs/cifsproto.h | 2 ++ 3 files changed, 71 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/cifs/cifs_spnego.c b/fs/cifs/cifs_spnego.c index 6908080e9b6d..b611fc2e8984 100644 --- a/fs/cifs/cifs_spnego.c +++ b/fs/cifs/cifs_spnego.c @@ -24,10 +24,13 @@ #include #include #include +#include #include #include "cifsglob.h" #include "cifs_spnego.h" #include "cifs_debug.h" +#include "cifsproto.h" +static const struct cred *spnego_cred; /* create a new cifs key */ static int @@ -102,6 +105,7 @@ cifs_get_spnego_key(struct cifs_ses *sesInfo) size_t desc_len; struct key *spnego_key; const char *hostname = server->hostname; + const struct cred *saved_cred; /* length of fields (with semicolons): ver=0xyz ip4=ipaddress host=hostname sec=mechanism uid=0xFF user=username */ @@ -163,7 +167,9 @@ cifs_get_spnego_key(struct cifs_ses *sesInfo) sprintf(dp, ";pid=0x%x", current->pid); cifs_dbg(FYI, "key description = %s\n", description); + saved_cred = override_creds(spnego_cred); spnego_key = request_key(&cifs_spnego_key_type, description, ""); + revert_creds(saved_cred); #ifdef CONFIG_CIFS_DEBUG2 if (cifsFYI && !IS_ERR(spnego_key)) { @@ -177,3 +183,64 @@ out: kfree(description); return spnego_key; } + +int +init_cifs_spnego(void) +{ + struct cred *cred; + struct key *keyring; + int ret; + + cifs_dbg(FYI, "Registering the %s key type\n", + cifs_spnego_key_type.name); + + /* + * Create an override credential set with special thread keyring for + * spnego upcalls. + */ + + cred = prepare_kernel_cred(NULL); + if (!cred) + return -ENOMEM; + + keyring = keyring_alloc(".cifs_spnego", + GLOBAL_ROOT_UID, GLOBAL_ROOT_GID, cred, + (KEY_POS_ALL & ~KEY_POS_SETATTR) | + KEY_USR_VIEW | KEY_USR_READ, + KEY_ALLOC_NOT_IN_QUOTA, NULL, NULL); + if (IS_ERR(keyring)) { + ret = PTR_ERR(keyring); + goto failed_put_cred; + } + + ret = register_key_type(&cifs_spnego_key_type); + if (ret < 0) + goto failed_put_key; + + /* + * instruct request_key() to use this special keyring as a cache for + * the results it looks up + */ + set_bit(KEY_FLAG_ROOT_CAN_CLEAR, &keyring->flags); + cred->thread_keyring = keyring; + cred->jit_keyring = KEY_REQKEY_DEFL_THREAD_KEYRING; + spnego_cred = cred; + + cifs_dbg(FYI, "cifs spnego keyring: %d\n", key_serial(keyring)); + return 0; + +failed_put_key: + key_put(keyring); +failed_put_cred: + put_cred(cred); + return ret; +} + +void +exit_cifs_spnego(void) +{ + key_revoke(spnego_cred->thread_keyring); + unregister_key_type(&cifs_spnego_key_type); + put_cred(spnego_cred); + cifs_dbg(FYI, "Unregistered %s key type\n", cifs_spnego_key_type.name); +} diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c index 08fa36e5b2bc..67f622df0858 100644 --- a/fs/cifs/cifsfs.c +++ b/fs/cifs/cifsfs.c @@ -1303,7 +1303,7 @@ init_cifs(void) goto out_destroy_mids; #ifdef CONFIG_CIFS_UPCALL - rc = register_key_type(&cifs_spnego_key_type); + rc = init_cifs_spnego(); if (rc) goto out_destroy_request_bufs; #endif /* CONFIG_CIFS_UPCALL */ @@ -1326,7 +1326,7 @@ out_init_cifs_idmap: out_register_key_type: #endif #ifdef CONFIG_CIFS_UPCALL - unregister_key_type(&cifs_spnego_key_type); + exit_cifs_spnego(); out_destroy_request_bufs: #endif cifs_destroy_request_bufs(); diff --git a/fs/cifs/cifsproto.h b/fs/cifs/cifsproto.h index 0f9a6bc4ba43..1243bd326591 100644 --- a/fs/cifs/cifsproto.h +++ b/fs/cifs/cifsproto.h @@ -58,6 +58,8 @@ do { \ } while (0) extern int init_cifs_idmap(void); extern void exit_cifs_idmap(void); +extern int init_cifs_spnego(void); +extern void exit_cifs_spnego(void); extern char *build_path_from_dentry(struct dentry *); extern char *cifs_build_path_to_root(struct smb_vol *vol, struct cifs_sb_info *cifs_sb, -- cgit v1.3-8-gc7d7 From 48a77aa7e20557319205f9bd4cc02d4b67b5f761 Mon Sep 17 00:00:00 2001 From: Steve French Date: Wed, 18 May 2016 20:48:32 -0500 Subject: CIFS: Remove some obsolete comments Remove some obsolete comments in the cifs inode_operations structs that were pointed out by Stephen Rothwell. CC: Stephen Rothwell CC: Al Viro Reviewed-by: Sachin Prabhu Signed-off-by: Steve French --- fs/cifs/cifsfs.c | 7 +------ 1 file changed, 1 insertion(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c index 67f622df0858..5d8b7edf8a8f 100644 --- a/fs/cifs/cifsfs.c +++ b/fs/cifs/cifsfs.c @@ -890,7 +890,6 @@ const struct inode_operations cifs_dir_inode_ops = { .rmdir = cifs_rmdir, .rename2 = cifs_rename2, .permission = cifs_permission, -/* revalidate:cifs_revalidate, */ .setattr = cifs_setattr, .symlink = cifs_symlink, .mknod = cifs_mknod, @@ -901,9 +900,8 @@ const struct inode_operations cifs_dir_inode_ops = { }; const struct inode_operations cifs_file_inode_ops = { -/* revalidate:cifs_revalidate, */ .setattr = cifs_setattr, - .getattr = cifs_getattr, /* do we need this anymore? */ + .getattr = cifs_getattr, .permission = cifs_permission, .setxattr = generic_setxattr, .getxattr = generic_getxattr, @@ -915,9 +913,6 @@ const struct inode_operations cifs_symlink_inode_ops = { .readlink = generic_readlink, .get_link = cifs_get_link, .permission = cifs_permission, - /* BB add the following two eventually */ - /* revalidate: cifs_revalidate, - setattr: cifs_notify_change, *//* BB do we need notify change */ .setxattr = generic_setxattr, .getxattr = generic_getxattr, .listxattr = cifs_listxattr, -- cgit v1.3-8-gc7d7 From 975756c41332bc5e523e9f843271ed5ab6aaaaaa Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Thu, 19 May 2016 11:57:21 -0700 Subject: f2fs: avoid ENOSPC fault in the recovery process This patch avoids impossible error injection, ENOSPC, during recovery process. Signed-off-by: Jaegeuk Kim --- fs/f2fs/recovery.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'fs') diff --git a/fs/f2fs/recovery.c b/fs/f2fs/recovery.c index f89b70e72004..3d7216d7a288 100644 --- a/fs/f2fs/recovery.c +++ b/fs/f2fs/recovery.c @@ -470,6 +470,10 @@ static int do_recover_data(struct f2fs_sb_info *sbi, struct inode *inode, if (src == NULL_ADDR) { err = reserve_new_block(&dn); +#ifdef CONFIG_F2FS_FAULT_INJECTION + while (err) + err = reserve_new_block(&dn); +#endif /* We should not get -ENOSPC */ f2fs_bug_on(sbi, err); } -- cgit v1.3-8-gc7d7 From 38f91ca8c0ea69f707c26f592dcc70f937088bcc Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Wed, 18 May 2016 14:07:56 -0700 Subject: f2fs: flush pending bios right away when error occurs Given errors, this patch flushes pending bios as soon as possible. Signed-off-by: Jaegeuk Kim --- fs/f2fs/checkpoint.c | 10 +++++++++- fs/f2fs/data.c | 2 +- fs/f2fs/f2fs.h | 7 +------ fs/f2fs/file.c | 8 ++++---- fs/f2fs/inode.c | 2 +- 5 files changed, 16 insertions(+), 13 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c index cf79598aec0e..389160049993 100644 --- a/fs/f2fs/checkpoint.c +++ b/fs/f2fs/checkpoint.c @@ -26,6 +26,14 @@ static struct kmem_cache *ino_entry_slab; struct kmem_cache *inode_entry_slab; +void f2fs_stop_checkpoint(struct f2fs_sb_info *sbi, bool end_io) +{ + set_ckpt_flags(sbi->ckpt, CP_ERROR_FLAG); + sbi->sb->s_flags |= MS_RDONLY; + if (!end_io) + f2fs_flush_merged_bios(sbi); +} + /* * We guarantee no failure on the returned page. */ @@ -91,7 +99,7 @@ repeat: * meta page. */ if (unlikely(!PageUptodate(page))) - f2fs_stop_checkpoint(sbi); + f2fs_stop_checkpoint(sbi, false); out: return page; } diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index faef6667ba45..105dd3d5cf5e 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -68,7 +68,7 @@ static void f2fs_write_end_io(struct bio *bio) if (unlikely(bio->bi_error)) { set_bit(AS_EIO, &page->mapping->flags); - f2fs_stop_checkpoint(sbi); + f2fs_stop_checkpoint(sbi, true); } end_page_writeback(page); } diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 02f0656cb2ab..916e7c238e3d 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -1698,12 +1698,6 @@ static inline bool f2fs_cp_error(struct f2fs_sb_info *sbi) return is_set_ckpt_flags(sbi->ckpt, CP_ERROR_FLAG); } -static inline void f2fs_stop_checkpoint(struct f2fs_sb_info *sbi) -{ - set_ckpt_flags(sbi->ckpt, CP_ERROR_FLAG); - sbi->sb->s_flags |= MS_RDONLY; -} - static inline bool is_dot_dotdot(const struct qstr *str) { if (str->len == 1 && str->name[0] == '.') @@ -1937,6 +1931,7 @@ void destroy_segment_manager_caches(void); /* * checkpoint.c */ +void f2fs_stop_checkpoint(struct f2fs_sb_info *, bool); struct page *grab_meta_page(struct f2fs_sb_info *, pgoff_t); struct page *get_meta_page(struct f2fs_sb_info *, pgoff_t); struct page *get_tmp_page(struct f2fs_sb_info *, pgoff_t); diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index 69dd7c9cdd6a..4c2125499da0 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -1571,21 +1571,21 @@ static int f2fs_ioc_shutdown(struct file *filp, unsigned long arg) case F2FS_GOING_DOWN_FULLSYNC: sb = freeze_bdev(sb->s_bdev); if (sb && !IS_ERR(sb)) { - f2fs_stop_checkpoint(sbi); + f2fs_stop_checkpoint(sbi, false); thaw_bdev(sb->s_bdev, sb); } break; case F2FS_GOING_DOWN_METASYNC: /* do checkpoint only */ f2fs_sync_fs(sb, 1); - f2fs_stop_checkpoint(sbi); + f2fs_stop_checkpoint(sbi, false); break; case F2FS_GOING_DOWN_NOSYNC: - f2fs_stop_checkpoint(sbi); + f2fs_stop_checkpoint(sbi, false); break; case F2FS_GOING_DOWN_METAFLUSH: sync_meta_pages(sbi, META, LONG_MAX); - f2fs_stop_checkpoint(sbi); + f2fs_stop_checkpoint(sbi, false); break; default: ret = -EINVAL; diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c index 689d6914e2e4..2e68adab0d64 100644 --- a/fs/f2fs/inode.c +++ b/fs/f2fs/inode.c @@ -283,7 +283,7 @@ retry: cond_resched(); goto retry; } else if (err != -ENOENT) { - f2fs_stop_checkpoint(sbi); + f2fs_stop_checkpoint(sbi, false); } return 0; } -- cgit v1.3-8-gc7d7 From 0f3311a8c266b9f4fae4e5cdfcd9a86970e2b9bd Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Sat, 21 May 2016 00:11:09 +0800 Subject: f2fs: fix to update dirty page count correctly Once we failed to merge inline data into inode page during flushing inline inode, we will skip invoking inode_dec_dirty_pages, which makes dirty page count incorrect, result in panic in ->evict_inode, Fix it. ------------[ cut here ]------------ kernel BUG at /home/yuchao/git/devf2fs/inode.c:336! invalid opcode: 0000 [#1] PREEMPT SMP CPU: 3 PID: 10004 Comm: umount Tainted: G O 4.6.0-rc5+ #17 Hardware name: innotek GmbH VirtualBox/VirtualBox, BIOS VirtualBox 12/01/2006 task: f0c33000 ti: c5212000 task.ti: c5212000 EIP: 0060:[] EFLAGS: 00010202 CPU: 3 EIP is at f2fs_evict_inode+0x85/0x490 [f2fs] EAX: 00000001 EBX: c4529ea0 ECX: 00000001 EDX: 00000000 ESI: c0131000 EDI: f89dd0a0 EBP: c5213e9c ESP: c5213e78 DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068 CR0: 80050033 CR2: b75878c0 CR3: 1a36a700 CR4: 000406f0 Stack: c4529ea0 c4529ef4 c5213e8c c176d45c c4529ef4 00000000 c4529ea0 c4529fac f89dd0a0 c5213eb0 c1204a68 c5213ed8 c452a2b4 c6680930 c5213ec0 c1204b64 c6680d44 c6680620 c5213eec c120588d ee84b000 ee84b5c0 c5214000 ee84b5e0 Call Trace: [] ? _raw_spin_unlock+0x2c/0x50 [] evict+0xa8/0x170 [] dispose_list+0x34/0x50 [] evict_inodes+0x10d/0x130 [] generic_shutdown_super+0x41/0xe0 [] ? unregister_shrinker+0x40/0x50 [] ? unregister_shrinker+0x40/0x50 [] kill_block_super+0x22/0x70 [] kill_f2fs_super+0x1e/0x20 [f2fs] [] deactivate_locked_super+0x3d/0x70 [] deactivate_super+0x43/0x60 [] cleanup_mnt+0x39/0x80 [] __cleanup_mnt+0x10/0x20 [] task_work_run+0x71/0x90 [] exit_to_usermode_loop+0x72/0x9e [] do_fast_syscall_32+0x19c/0x1c0 [] sysenter_past_esp+0x45/0x74 EIP: [] f2fs_evict_inode+0x85/0x490 [f2fs] SS:ESP 0068:c5213e78 ---[ end trace d30536330b7fdc58 ]--- Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/node.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c index cda7c7ccd4c8..1f21aae80c40 100644 --- a/fs/f2fs/node.c +++ b/fs/f2fs/node.c @@ -1197,6 +1197,7 @@ static void flush_inline_data(struct f2fs_sb_info *sbi, nid_t ino) { struct inode *inode; struct page *page; + int ret; /* should flush inline_data before evict_inode */ inode = ilookup(sbi->sb, ino); @@ -1216,9 +1217,9 @@ static void flush_inline_data(struct f2fs_sb_info *sbi, nid_t ino) if (!clear_page_dirty_for_io(page)) goto page_out; - if (!f2fs_write_inline_data(inode, page)) - inode_dec_dirty_pages(inode); - else + ret = f2fs_write_inline_data(inode, page); + inode_dec_dirty_pages(inode); + if (ret) set_page_dirty(page); page_out: f2fs_put_page(page, 1); -- cgit v1.3-8-gc7d7 From 63678c32e209bd165f33432bbed72b2954ce5ae4 Mon Sep 17 00:00:00 2001 From: Rich Felker Date: Fri, 20 May 2016 16:57:47 -0700 Subject: tmpfs/ramfs: fix VM_MAYSHARE mappings for NOMMU The nommu do_mmap expects f_op->get_unmapped_area to either succeed or return -ENOSYS for VM_MAYSHARE (e.g. private read-only) mappings. Returning addr in the non-MAP_SHARED case was completely wrong, and only happened to work because addr was 0. However, it prevented VM_MAYSHARE mappings from sharing backing with the fs cache, and forced such mappings (including shareable program text) to be copied whenever the number of mappings transitioned from 0 to 1, impacting performance and memory usage. Subsequent mappings beyond the first still correctly shared memory with the first. Instead, treat VM_MAYSHARE identically to VM_SHARED at the file ops level; do_mmap already handles the semantic differences between them. Signed-off-by: Rich Felker Cc: Michal Hocko Cc: Greg Ungerer Cc: Geert Uytterhoeven Cc: Yoshinori Sato Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ramfs/file-nommu.c | 8 ++------ 1 file changed, 2 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/ramfs/file-nommu.c b/fs/ramfs/file-nommu.c index a586467f6ff6..be3ddd189cd4 100644 --- a/fs/ramfs/file-nommu.c +++ b/fs/ramfs/file-nommu.c @@ -211,14 +211,11 @@ static unsigned long ramfs_nommu_get_unmapped_area(struct file *file, struct page **pages = NULL, **ptr, *page; loff_t isize; - if (!(flags & MAP_SHARED)) - return addr; - /* the mapping mustn't extend beyond the EOF */ lpages = (len + PAGE_SIZE - 1) >> PAGE_SHIFT; isize = i_size_read(inode); - ret = -EINVAL; + ret = -ENOSYS; maxpages = (isize + PAGE_SIZE - 1) >> PAGE_SHIFT; if (pgoff >= maxpages) goto out; @@ -227,7 +224,6 @@ static unsigned long ramfs_nommu_get_unmapped_area(struct file *file, goto out; /* gang-find the pages */ - ret = -ENOMEM; pages = kcalloc(lpages, sizeof(struct page *), GFP_KERNEL); if (!pages) goto out_free; @@ -263,7 +259,7 @@ out: */ static int ramfs_nommu_mmap(struct file *file, struct vm_area_struct *vma) { - if (!(vma->vm_flags & VM_SHARED)) + if (!(vma->vm_flags & (VM_SHARED | VM_MAYSHARE))) return -ENOSYS; file_accessed(file); -- cgit v1.3-8-gc7d7 From 78ebc2f7146156f488083c9e5a7ded9d5c38c58b Mon Sep 17 00:00:00 2001 From: Tetsuo Handa Date: Fri, 20 May 2016 16:58:10 -0700 Subject: mm,writeback: don't use memory reserves for wb_start_writeback When writeback operation cannot make forward progress because memory allocation requests needed for doing I/O cannot be satisfied (e.g. under OOM-livelock situation), we can observe flood of order-0 page allocation failure messages caused by complete depletion of memory reserves. This is caused by unconditionally allocating "struct wb_writeback_work" objects using GFP_ATOMIC from PF_MEMALLOC context. __alloc_pages_nodemask() { __alloc_pages_slowpath() { __alloc_pages_direct_reclaim() { __perform_reclaim() { current->flags |= PF_MEMALLOC; try_to_free_pages() { do_try_to_free_pages() { wakeup_flusher_threads() { wb_start_writeback() { kzalloc(sizeof(*work), GFP_ATOMIC) { /* ALLOC_NO_WATERMARKS via PF_MEMALLOC */ } } } } } current->flags &= ~PF_MEMALLOC; } } } } Since I/O is stalling, allocating writeback requests forever shall deplete memory reserves. Fortunately, since wb_start_writeback() can fall back to wb_wakeup() when allocating "struct wb_writeback_work" failed, we don't need to allow wb_start_writeback() to use memory reserves. Mem-Info: active_anon:289393 inactive_anon:2093 isolated_anon:29 active_file:10838 inactive_file:113013 isolated_file:859 unevictable:0 dirty:108531 writeback:5308 unstable:0 slab_reclaimable:5526 slab_unreclaimable:7077 mapped:9970 shmem:2159 pagetables:2387 bounce:0 free:3042 free_pcp:0 free_cma:0 Node 0 DMA free:6968kB min:44kB low:52kB high:64kB active_anon:6056kB inactive_anon:176kB active_file:712kB inactive_file:744kB unevictable:0kB isolated(anon):0kB isolated(file):0kB present:15988kB managed:15904kB mlocked:0kB dirty:756kB writeback:0kB mapped:736kB shmem:184kB slab_reclaimable:48kB slab_unreclaimable:208kB kernel_stack:160kB pagetables:144kB unstable:0kB bounce:0kB free_pcp:0kB local_pcp:0kB free_cma:0kB writeback_tmp:0kB pages_scanned:9708 all_unreclaimable? yes lowmem_reserve[]: 0 1732 1732 1732 Node 0 DMA32 free:5200kB min:5200kB low:6500kB high:7800kB active_anon:1151516kB inactive_anon:8196kB active_file:42640kB inactive_file:451076kB unevictable:0kB isolated(anon):116kB isolated(file):3564kB present:2080640kB managed:1775332kB mlocked:0kB dirty:433368kB writeback:21232kB mapped:39144kB shmem:8452kB slab_reclaimable:22056kB slab_unreclaimable:28100kB kernel_stack:20976kB pagetables:9404kB unstable:0kB bounce:0kB free_pcp:120kB local_pcp:0kB free_cma:0kB writeback_tmp:0kB pages_scanned:2701604 all_unreclaimable? no lowmem_reserve[]: 0 0 0 0 Node 0 DMA: 25*4kB (UME) 16*8kB (UME) 3*16kB (UE) 5*32kB (UME) 2*64kB (UM) 2*128kB (ME) 2*256kB (ME) 1*512kB (E) 1*1024kB (E) 2*2048kB (ME) 0*4096kB = 6964kB Node 0 DMA32: 925*4kB (UME) 140*8kB (UME) 5*16kB (ME) 5*32kB (M) 0*64kB 0*128kB 0*256kB 0*512kB 0*1024kB 0*2048kB 0*4096kB = 5060kB Node 0 hugepages_total=0 hugepages_free=0 hugepages_surp=0 hugepages_size=1048576kB Node 0 hugepages_total=0 hugepages_free=0 hugepages_surp=0 hugepages_size=2048kB 126847 total pagecache pages 0 pages in swap cache Swap cache stats: add 0, delete 0, find 0/0 Free swap = 0kB Total swap = 0kB 524157 pages RAM 0 pages HighMem/MovableOnly 76348 pages reserved 0 pages hwpoisoned Out of memory: Kill process 4450 (file_io.00) score 998 or sacrifice child Killed process 4450 (file_io.00) total-vm:4308kB, anon-rss:100kB, file-rss:1184kB, shmem-rss:0kB kthreadd: page allocation failure: order:0, mode:0x2200020 file_io.00: page allocation failure: order:0, mode:0x2200020 CPU: 0 PID: 4457 Comm: file_io.00 Not tainted 4.5.0-rc7+ #45 Hardware name: VMware, Inc. VMware Virtual Platform/440BX Desktop Reference Platform, BIOS 6.00 07/31/2013 Call Trace: warn_alloc_failed+0xf7/0x150 __alloc_pages_nodemask+0x23f/0xa60 alloc_pages_current+0x87/0x110 new_slab+0x3a1/0x440 ___slab_alloc+0x3cf/0x590 __slab_alloc.isra.64+0x18/0x1d kmem_cache_alloc+0x11c/0x150 wb_start_writeback+0x39/0x90 wakeup_flusher_threads+0x7f/0xf0 do_try_to_free_pages+0x1f9/0x410 try_to_free_pages+0x94/0xc0 __alloc_pages_nodemask+0x566/0xa60 alloc_pages_current+0x87/0x110 __page_cache_alloc+0xaf/0xc0 pagecache_get_page+0x88/0x260 grab_cache_page_write_begin+0x21/0x40 xfs_vm_write_begin+0x2f/0xf0 generic_perform_write+0xca/0x1c0 xfs_file_buffered_aio_write+0xcc/0x1f0 xfs_file_write_iter+0x84/0x140 __vfs_write+0xc7/0x100 vfs_write+0x9d/0x190 SyS_write+0x50/0xc0 entry_SYSCALL_64_fastpath+0x12/0x6a Mem-Info: active_anon:293335 inactive_anon:2093 isolated_anon:0 active_file:10829 inactive_file:110045 isolated_file:32 unevictable:0 dirty:109275 writeback:822 unstable:0 slab_reclaimable:5489 slab_unreclaimable:10070 mapped:9999 shmem:2159 pagetables:2420 bounce:0 free:3 free_pcp:0 free_cma:0 Node 0 DMA free:12kB min:44kB low:52kB high:64kB active_anon:6060kB inactive_anon:176kB active_file:708kB inactive_file:756kB unevictable:0kB isolated(anon):0kB isolated(file):0kB present:15988kB managed:15904kB mlocked:0kB dirty:756kB writeback:0kB mapped:736kB shmem:184kB slab_reclaimable:48kB slab_unreclaimable:7160kB kernel_stack:160kB pagetables:144kB unstable:0kB bounce:0kB free_pcp:0kB local_pcp:0kB free_cma:0kB writeback_tmp:0kB pages_scanned:9844 all_unreclaimable? yes lowmem_reserve[]: 0 1732 1732 1732 Node 0 DMA32 free:0kB min:5200kB low:6500kB high:7800kB active_anon:1167280kB inactive_anon:8196kB active_file:42608kB inactive_file:439424kB unevictable:0kB isolated(anon):0kB isolated(file):128kB present:2080640kB managed:1775332kB mlocked:0kB dirty:436344kB writeback:3288kB mapped:39260kB shmem:8452kB slab_reclaimable:21908kB slab_unreclaimable:33120kB kernel_stack:20976kB pagetables:9536kB unstable:0kB bounce:0kB free_pcp:0kB local_pcp:0kB free_cma:0kB writeback_tmp:0kB pages_scanned:11073180 all_unreclaimable? yes lowmem_reserve[]: 0 0 0 0 Node 0 DMA: 0*4kB 0*8kB 0*16kB 0*32kB 0*64kB 0*128kB 0*256kB 0*512kB 0*1024kB 0*2048kB 0*4096kB = 0kB Node 0 DMA32: 0*4kB 0*8kB 0*16kB 0*32kB 0*64kB 0*128kB 0*256kB 0*512kB 0*1024kB 0*2048kB 0*4096kB = 0kB Node 0 hugepages_total=0 hugepages_free=0 hugepages_surp=0 hugepages_size=1048576kB Node 0 hugepages_total=0 hugepages_free=0 hugepages_surp=0 hugepages_size=2048kB 123086 total pagecache pages 0 pages in swap cache Swap cache stats: add 0, delete 0, find 0/0 Free swap = 0kB Total swap = 0kB 524157 pages RAM 0 pages HighMem/MovableOnly 76348 pages reserved 0 pages hwpoisoned SLUB: Unable to allocate memory on node -1 (gfp=0x2088020) cache: kmalloc-64, object size: 64, buffer size: 64, default order: 0, min order: 0 node 0: slabs: 3218, objs: 205952, free: 0 file_io.00: page allocation failure: order:0, mode:0x2200020 CPU: 0 PID: 4457 Comm: file_io.00 Not tainted 4.5.0-rc7+ #45 Assuming that somebody will find a better solution, let's apply this patch for now to stop bleeding, for this problem frequently prevents me from testing OOM livelock condition. Link: http://lkml.kernel.org/r/20160318131136.GE7152@quack.suse.cz Signed-off-by: Tetsuo Handa Acked-by: Michal Hocko Cc: Jan Kara Cc: Tejun Heo Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/fs-writeback.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c index 592cea54cea0..989a2cef6b76 100644 --- a/fs/fs-writeback.c +++ b/fs/fs-writeback.c @@ -931,7 +931,8 @@ void wb_start_writeback(struct bdi_writeback *wb, long nr_pages, * This is WB_SYNC_NONE writeback, so if allocation fails just * wakeup the thread for old dirty data writeback */ - work = kzalloc(sizeof(*work), GFP_ATOMIC); + work = kzalloc(sizeof(*work), + GFP_NOWAIT | __GFP_NOMEMALLOC | __GFP_NOWARN); if (!work) { trace_writeback_nowork(wb); wb_wakeup(wb); -- cgit v1.3-8-gc7d7 From d2005e3f41d4f9299e2df6a967c8beb5086967a9 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Fri, 20 May 2016 16:58:36 -0700 Subject: userfaultfd: don't pin the user memory in userfaultfd_file_create() userfaultfd_file_create() increments mm->mm_users; this means that the memory won't be unmapped/freed if mm owner exits/execs, and UFFDIO_COPY after that can populate the orphaned mm more. Change userfaultfd_file_create() and userfaultfd_ctx_put() to use mm->mm_count to pin mm_struct. This means that atomic_inc_not_zero(mm->mm_users) is needed when we are going to actually play with this memory. Except handle_userfault() path doesn't need this, the caller must already have a reference. The patch adds the new trivial helper, mmget_not_zero(), it can have more users. Link: http://lkml.kernel.org/r/20160516172254.GA8595@redhat.com Signed-off-by: Oleg Nesterov Cc: Andrea Arcangeli Cc: Michal Hocko Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/userfaultfd.c | 41 ++++++++++++++++++++++++++++------------- include/linux/sched.h | 7 ++++++- 2 files changed, 34 insertions(+), 14 deletions(-) (limited to 'fs') diff --git a/fs/userfaultfd.c b/fs/userfaultfd.c index 66cdb44616d5..2d97952e341a 100644 --- a/fs/userfaultfd.c +++ b/fs/userfaultfd.c @@ -137,7 +137,7 @@ static void userfaultfd_ctx_put(struct userfaultfd_ctx *ctx) VM_BUG_ON(waitqueue_active(&ctx->fault_wqh)); VM_BUG_ON(spin_is_locked(&ctx->fd_wqh.lock)); VM_BUG_ON(waitqueue_active(&ctx->fd_wqh)); - mmput(ctx->mm); + mmdrop(ctx->mm); kmem_cache_free(userfaultfd_ctx_cachep, ctx); } } @@ -434,6 +434,9 @@ static int userfaultfd_release(struct inode *inode, struct file *file) ACCESS_ONCE(ctx->released) = true; + if (!mmget_not_zero(mm)) + goto wakeup; + /* * Flush page faults out of all CPUs. NOTE: all page faults * must be retried without returning VM_FAULT_SIGBUS if @@ -466,7 +469,8 @@ static int userfaultfd_release(struct inode *inode, struct file *file) vma->vm_userfaultfd_ctx = NULL_VM_UFFD_CTX; } up_write(&mm->mmap_sem); - + mmput(mm); +wakeup: /* * After no new page faults can wait on this fault_*wqh, flush * the last page faults that may have been already waiting on @@ -760,10 +764,12 @@ static int userfaultfd_register(struct userfaultfd_ctx *ctx, start = uffdio_register.range.start; end = start + uffdio_register.range.len; + ret = -ENOMEM; + if (!mmget_not_zero(mm)) + goto out; + down_write(&mm->mmap_sem); vma = find_vma_prev(mm, start, &prev); - - ret = -ENOMEM; if (!vma) goto out_unlock; @@ -864,6 +870,7 @@ static int userfaultfd_register(struct userfaultfd_ctx *ctx, } while (vma && vma->vm_start < end); out_unlock: up_write(&mm->mmap_sem); + mmput(mm); if (!ret) { /* * Now that we scanned all vmas we can already tell @@ -902,10 +909,12 @@ static int userfaultfd_unregister(struct userfaultfd_ctx *ctx, start = uffdio_unregister.start; end = start + uffdio_unregister.len; + ret = -ENOMEM; + if (!mmget_not_zero(mm)) + goto out; + down_write(&mm->mmap_sem); vma = find_vma_prev(mm, start, &prev); - - ret = -ENOMEM; if (!vma) goto out_unlock; @@ -998,6 +1007,7 @@ static int userfaultfd_unregister(struct userfaultfd_ctx *ctx, } while (vma && vma->vm_start < end); out_unlock: up_write(&mm->mmap_sem); + mmput(mm); out: return ret; } @@ -1067,9 +1077,11 @@ static int userfaultfd_copy(struct userfaultfd_ctx *ctx, goto out; if (uffdio_copy.mode & ~UFFDIO_COPY_MODE_DONTWAKE) goto out; - - ret = mcopy_atomic(ctx->mm, uffdio_copy.dst, uffdio_copy.src, - uffdio_copy.len); + if (mmget_not_zero(ctx->mm)) { + ret = mcopy_atomic(ctx->mm, uffdio_copy.dst, uffdio_copy.src, + uffdio_copy.len); + mmput(ctx->mm); + } if (unlikely(put_user(ret, &user_uffdio_copy->copy))) return -EFAULT; if (ret < 0) @@ -1110,8 +1122,11 @@ static int userfaultfd_zeropage(struct userfaultfd_ctx *ctx, if (uffdio_zeropage.mode & ~UFFDIO_ZEROPAGE_MODE_DONTWAKE) goto out; - ret = mfill_zeropage(ctx->mm, uffdio_zeropage.range.start, - uffdio_zeropage.range.len); + if (mmget_not_zero(ctx->mm)) { + ret = mfill_zeropage(ctx->mm, uffdio_zeropage.range.start, + uffdio_zeropage.range.len); + mmput(ctx->mm); + } if (unlikely(put_user(ret, &user_uffdio_zeropage->zeropage))) return -EFAULT; if (ret < 0) @@ -1289,12 +1304,12 @@ static struct file *userfaultfd_file_create(int flags) ctx->released = false; ctx->mm = current->mm; /* prevent the mm struct to be freed */ - atomic_inc(&ctx->mm->mm_users); + atomic_inc(&ctx->mm->mm_count); file = anon_inode_getfile("[userfaultfd]", &userfaultfd_fops, ctx, O_RDWR | (flags & UFFD_SHARED_FCNTL_FLAGS)); if (IS_ERR(file)) { - mmput(ctx->mm); + mmdrop(ctx->mm); kmem_cache_free(userfaultfd_ctx_cachep, ctx); } out: diff --git a/include/linux/sched.h b/include/linux/sched.h index 01fe1bb68754..6b3213d96da6 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -2723,12 +2723,17 @@ extern struct mm_struct * mm_alloc(void); /* mmdrop drops the mm and the page tables */ extern void __mmdrop(struct mm_struct *); -static inline void mmdrop(struct mm_struct * mm) +static inline void mmdrop(struct mm_struct *mm) { if (unlikely(atomic_dec_and_test(&mm->mm_count))) __mmdrop(mm); } +static inline bool mmget_not_zero(struct mm_struct *mm) +{ + return atomic_inc_not_zero(&mm->mm_users); +} + /* mmput gets rid of the mappings and all user-space */ extern void mmput(struct mm_struct *); /* same as above but performs the slow path from the async kontext. Can -- cgit v1.3-8-gc7d7 From 3e42979e65dace1f9268dd5440e5ab096b8dee59 Mon Sep 17 00:00:00 2001 From: "Richard W.M. Jones" Date: Fri, 20 May 2016 17:00:05 -0700 Subject: procfs: expose umask in /proc//status It's not possible to read the process umask without also modifying it, which is what umask(2) does. A library cannot read umask safely, especially if the main program might be multithreaded. Add a new status line ("Umask") in /proc//status. It contains the file mode creation mask (umask) in octal. It is only shown for tasks which have task->fs. This patch is adapted from one originally written by Pierre Carrier. The use case is that we have endless trouble with people setting weird umask() values (usually on the grounds of "security"), and then everything breaking. I'm on the hook to fix these. We'd like to add debugging to our program so we can dump out the umask in debug reports. Previous versions of the patch used a syscall so you could only read your own umask. That's all I need. However there was quite a lot of push-back from those, so this new version exports it in /proc. See: https://lkml.org/lkml/2016/4/13/704 [umask2] https://lkml.org/lkml/2016/4/13/487 [getumask] Signed-off-by: Richard W.M. Jones Acked-by: Konstantin Khlebnikov Acked-by: Jerome Marchand Acked-by: Kees Cook Cc: "Theodore Ts'o" Cc: Michal Hocko Cc: Pierre Carrier Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- Documentation/filesystems/proc.txt | 1 + fs/proc/array.c | 20 +++++++++++++++++++- 2 files changed, 20 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/Documentation/filesystems/proc.txt b/Documentation/filesystems/proc.txt index 7f5607a089b4..e8d00759bfa5 100644 --- a/Documentation/filesystems/proc.txt +++ b/Documentation/filesystems/proc.txt @@ -225,6 +225,7 @@ Table 1-2: Contents of the status files (as of 4.1) TracerPid PID of process tracing this process (0 if not) Uid Real, effective, saved set, and file system UIDs Gid Real, effective, saved set, and file system GIDs + Umask file mode creation mask FDSize number of file descriptor slots currently allocated Groups supplementary group list NStgid descendant namespace thread group ID hierarchy diff --git a/fs/proc/array.c b/fs/proc/array.c index b6c00ce0e29e..88c7de12197b 100644 --- a/fs/proc/array.c +++ b/fs/proc/array.c @@ -83,6 +83,7 @@ #include #include #include +#include #include #include @@ -139,12 +140,25 @@ static inline const char *get_task_state(struct task_struct *tsk) return task_state_array[fls(state)]; } +static inline int get_task_umask(struct task_struct *tsk) +{ + struct fs_struct *fs; + int umask = -ENOENT; + + task_lock(tsk); + fs = tsk->fs; + if (fs) + umask = fs->umask; + task_unlock(tsk); + return umask; +} + static inline void task_state(struct seq_file *m, struct pid_namespace *ns, struct pid *pid, struct task_struct *p) { struct user_namespace *user_ns = seq_user_ns(m); struct group_info *group_info; - int g; + int g, umask; struct task_struct *tracer; const struct cred *cred; pid_t ppid, tpid = 0, tgid, ngid; @@ -162,6 +176,10 @@ static inline void task_state(struct seq_file *m, struct pid_namespace *ns, ngid = task_numa_group_id(p); cred = get_task_cred(p); + umask = get_task_umask(p); + if (umask >= 0) + seq_printf(m, "Umask:\t%#04o\n", umask); + task_lock(p); if (p->files) max_fds = files_fdtable(p->files)->max_fds; -- cgit v1.3-8-gc7d7 From 1b3044e39a89cb1d4d5313da477e8dfea2b5232d Mon Sep 17 00:00:00 2001 From: Janis Danisevskis Date: Fri, 20 May 2016 17:00:08 -0700 Subject: procfs: fix pthread cross-thread naming if !PR_DUMPABLE The PR_DUMPABLE flag causes the pid related paths of the proc file system to be owned by ROOT. The implementation of pthread_set/getname_np however needs access to /proc//task//comm. If PR_DUMPABLE is false this implementation is locked out. This patch installs a special permission function for the file "comm" that grants read and write access to all threads of the same group regardless of the ownership of the inode. For all other threads the function falls back to the generic inode permission check. [akpm@linux-foundation.org: fix spello in comment] Signed-off-by: Janis Danisevskis Acked-by: Kees Cook Cc: Al Viro Cc: Cyrill Gorcunov Cc: Alexey Dobriyan Cc: Colin Ian King Cc: David Rientjes Cc: Minfei Huang Cc: John Stultz Cc: Calvin Owens Cc: Jann Horn Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/proc/base.c | 42 +++++++++++++++++++++++++++++++++++++++++- 1 file changed, 41 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/proc/base.c b/fs/proc/base.c index ff4527dd69b7..a11eb7196ec8 100644 --- a/fs/proc/base.c +++ b/fs/proc/base.c @@ -3162,6 +3162,44 @@ int proc_pid_readdir(struct file *file, struct dir_context *ctx) return 0; } +/* + * proc_tid_comm_permission is a special permission function exclusively + * used for the node /proc//task//comm. + * It bypasses generic permission checks in the case where a task of the same + * task group attempts to access the node. + * The rationale behind this is that glibc and bionic access this node for + * cross thread naming (pthread_set/getname_np(!self)). However, if + * PR_SET_DUMPABLE gets set to 0 this node among others becomes uid=0 gid=0, + * which locks out the cross thread naming implementation. + * This function makes sure that the node is always accessible for members of + * same thread group. + */ +static int proc_tid_comm_permission(struct inode *inode, int mask) +{ + bool is_same_tgroup; + struct task_struct *task; + + task = get_proc_task(inode); + if (!task) + return -ESRCH; + is_same_tgroup = same_thread_group(current, task); + put_task_struct(task); + + if (likely(is_same_tgroup && !(mask & MAY_EXEC))) { + /* This file (/proc//task//comm) can always be + * read or written by the members of the corresponding + * thread group. + */ + return 0; + } + + return generic_permission(inode, mask); +} + +static const struct inode_operations proc_tid_comm_inode_operations = { + .permission = proc_tid_comm_permission, +}; + /* * Tasks */ @@ -3180,7 +3218,9 @@ static const struct pid_entry tid_base_stuff[] = { #ifdef CONFIG_SCHED_DEBUG REG("sched", S_IRUGO|S_IWUSR, proc_pid_sched_operations), #endif - REG("comm", S_IRUGO|S_IWUSR, proc_pid_set_comm_operations), + NOD("comm", S_IFREG|S_IRUGO|S_IWUSR, + &proc_tid_comm_inode_operations, + &proc_pid_set_comm_operations, {}), #ifdef CONFIG_HAVE_ARCH_TRACEHOOK ONE("syscall", S_IRUSR, proc_pid_syscall), #endif -- cgit v1.3-8-gc7d7 From 8da4b8c48e7b43cb16d05e1dbb34ad9f73ab7efd Mon Sep 17 00:00:00 2001 From: Andy Shevchenko Date: Fri, 20 May 2016 17:01:00 -0700 Subject: lib/uuid.c: move generate_random_uuid() to uuid.c Let's gather the UUID related functions under one hood. Signed-off-by: Andy Shevchenko Reviewed-by: Matt Fleming Cc: Dmitry Kasatkin Cc: Mimi Zohar Cc: Rasmus Villemoes Cc: Arnd Bergmann Cc: "Theodore Ts'o" Cc: Al Viro Cc: Jens Axboe Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- drivers/char/random.c | 21 +-------------------- fs/btrfs/volumes.c | 2 +- fs/ext4/ioctl.c | 2 +- fs/f2fs/file.c | 2 +- fs/reiserfs/objectid.c | 2 +- fs/ubifs/sb.c | 2 +- include/linux/random.h | 1 - include/linux/uuid.h | 2 ++ lib/uuid.c | 20 ++++++++++++++++++++ 9 files changed, 28 insertions(+), 26 deletions(-) (limited to 'fs') diff --git a/drivers/char/random.c b/drivers/char/random.c index b583e5336630..0158d3bff7e5 100644 --- a/drivers/char/random.c +++ b/drivers/char/random.c @@ -260,6 +260,7 @@ #include #include #include +#include #include #include @@ -1621,26 +1622,6 @@ SYSCALL_DEFINE3(getrandom, char __user *, buf, size_t, count, return urandom_read(NULL, buf, count, NULL); } -/*************************************************************** - * Random UUID interface - * - * Used here for a Boot ID, but can be useful for other kernel - * drivers. - ***************************************************************/ - -/* - * Generate random UUID - */ -void generate_random_uuid(unsigned char uuid_out[16]) -{ - get_random_bytes(uuid_out, 16); - /* Set UUID version to 4 --- truly random generation */ - uuid_out[6] = (uuid_out[6] & 0x0F) | 0x40; - /* Set the UUID variant to DCE */ - uuid_out[8] = (uuid_out[8] & 0x3F) | 0x80; -} -EXPORT_SYMBOL(generate_random_uuid); - /******************************************************************** * * Sysctl interface diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index bd0f45fb38c4..bfb80da3e6eb 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -20,13 +20,13 @@ #include #include #include -#include #include #include #include #include #include #include +#include #include #include "ctree.h" #include "extent_map.h" diff --git a/fs/ext4/ioctl.c b/fs/ext4/ioctl.c index eae5917c534e..7497f50cb293 100644 --- a/fs/ext4/ioctl.c +++ b/fs/ext4/ioctl.c @@ -13,8 +13,8 @@ #include #include #include -#include #include +#include #include #include "ext4_jbd2.h" #include "ext4.h" diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index eb9d027e5981..c6b14951bef3 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -20,7 +20,7 @@ #include #include #include -#include +#include #include "f2fs.h" #include "node.h" diff --git a/fs/reiserfs/objectid.c b/fs/reiserfs/objectid.c index 99a5d5dae46a..415d66ca87d1 100644 --- a/fs/reiserfs/objectid.c +++ b/fs/reiserfs/objectid.c @@ -3,8 +3,8 @@ */ #include -#include #include +#include #include "reiserfs.h" /* find where objectid map starts */ diff --git a/fs/ubifs/sb.c b/fs/ubifs/sb.c index f4fbc7b6b794..3cbb904a6d7d 100644 --- a/fs/ubifs/sb.c +++ b/fs/ubifs/sb.c @@ -28,8 +28,8 @@ #include "ubifs.h" #include -#include #include +#include /* * Default journal size in logical eraseblocks as a percent of total diff --git a/include/linux/random.h b/include/linux/random.h index 9c29122037f9..e47e533742b5 100644 --- a/include/linux/random.h +++ b/include/linux/random.h @@ -26,7 +26,6 @@ extern void get_random_bytes(void *buf, int nbytes); extern int add_random_ready_callback(struct random_ready_callback *rdy); extern void del_random_ready_callback(struct random_ready_callback *rdy); extern void get_random_bytes_arch(void *buf, int nbytes); -void generate_random_uuid(unsigned char uuid_out[16]); extern int random_int_secret_init(void); #ifndef MODULE diff --git a/include/linux/uuid.h b/include/linux/uuid.h index 6df2509033d7..91c2b6d9cbb7 100644 --- a/include/linux/uuid.h +++ b/include/linux/uuid.h @@ -33,6 +33,8 @@ static inline int uuid_be_cmp(const uuid_be u1, const uuid_be u2) return memcmp(&u1, &u2, sizeof(uuid_be)); } +void generate_random_uuid(unsigned char uuid[16]); + extern void uuid_le_gen(uuid_le *u); extern void uuid_be_gen(uuid_be *u); diff --git a/lib/uuid.c b/lib/uuid.c index 398821e4dce1..6c81c0b0467e 100644 --- a/lib/uuid.c +++ b/lib/uuid.c @@ -23,6 +23,26 @@ #include #include +/*************************************************************** + * Random UUID interface + * + * Used here for a Boot ID, but can be useful for other kernel + * drivers. + ***************************************************************/ + +/* + * Generate random UUID + */ +void generate_random_uuid(unsigned char uuid[16]) +{ + get_random_bytes(uuid, 16); + /* Set UUID version to 4 --- truly random generation */ + uuid[6] = (uuid[6] & 0x0F) | 0x40; + /* Set the UUID variant to DCE */ + uuid[8] = (uuid[8] & 0x3F) | 0x80; +} +EXPORT_SYMBOL(generate_random_uuid); + static void __uuid_gen_common(__u8 b[16]) { prandom_bytes(b, 16); -- cgit v1.3-8-gc7d7 From 8236431d8d09eee70e6cbc506426a7c97778a2e6 Mon Sep 17 00:00:00 2001 From: Andy Shevchenko Date: Fri, 20 May 2016 17:01:21 -0700 Subject: fs/efivarfs/inode.c: use generic UUID library Instead of opencoding let's use generic UUID library functions here. Signed-off-by: Andy Shevchenko Reviewed-by: Matt Fleming Cc: Dmitry Kasatkin Cc: Mimi Zohar Cc: Rasmus Villemoes Cc: Arnd Bergmann Cc: "Theodore Ts'o" Cc: Al Viro Cc: Jens Axboe Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/efivarfs/inode.c | 40 +++------------------------------------- 1 file changed, 3 insertions(+), 37 deletions(-) (limited to 'fs') diff --git a/fs/efivarfs/inode.c b/fs/efivarfs/inode.c index e2ab6d0497f2..1d73fc6dba13 100644 --- a/fs/efivarfs/inode.c +++ b/fs/efivarfs/inode.c @@ -11,6 +11,7 @@ #include #include #include +#include #include "internal.h" @@ -46,11 +47,7 @@ struct inode *efivarfs_get_inode(struct super_block *sb, */ bool efivarfs_valid_name(const char *str, int len) { - static const char dashes[EFI_VARIABLE_GUID_LEN] = { - [8] = 1, [13] = 1, [18] = 1, [23] = 1 - }; const char *s = str + len - EFI_VARIABLE_GUID_LEN; - int i; /* * We need a GUID, plus at least one letter for the variable name, @@ -68,37 +65,7 @@ bool efivarfs_valid_name(const char *str, int len) * * 12345678-1234-1234-1234-123456789abc */ - for (i = 0; i < EFI_VARIABLE_GUID_LEN; i++) { - if (dashes[i]) { - if (*s++ != '-') - return false; - } else { - if (!isxdigit(*s++)) - return false; - } - } - - return true; -} - -static void efivarfs_hex_to_guid(const char *str, efi_guid_t *guid) -{ - guid->b[0] = hex_to_bin(str[6]) << 4 | hex_to_bin(str[7]); - guid->b[1] = hex_to_bin(str[4]) << 4 | hex_to_bin(str[5]); - guid->b[2] = hex_to_bin(str[2]) << 4 | hex_to_bin(str[3]); - guid->b[3] = hex_to_bin(str[0]) << 4 | hex_to_bin(str[1]); - guid->b[4] = hex_to_bin(str[11]) << 4 | hex_to_bin(str[12]); - guid->b[5] = hex_to_bin(str[9]) << 4 | hex_to_bin(str[10]); - guid->b[6] = hex_to_bin(str[16]) << 4 | hex_to_bin(str[17]); - guid->b[7] = hex_to_bin(str[14]) << 4 | hex_to_bin(str[15]); - guid->b[8] = hex_to_bin(str[19]) << 4 | hex_to_bin(str[20]); - guid->b[9] = hex_to_bin(str[21]) << 4 | hex_to_bin(str[22]); - guid->b[10] = hex_to_bin(str[24]) << 4 | hex_to_bin(str[25]); - guid->b[11] = hex_to_bin(str[26]) << 4 | hex_to_bin(str[27]); - guid->b[12] = hex_to_bin(str[28]) << 4 | hex_to_bin(str[29]); - guid->b[13] = hex_to_bin(str[30]) << 4 | hex_to_bin(str[31]); - guid->b[14] = hex_to_bin(str[32]) << 4 | hex_to_bin(str[33]); - guid->b[15] = hex_to_bin(str[34]) << 4 | hex_to_bin(str[35]); + return uuid_is_valid(s); } static int efivarfs_create(struct inode *dir, struct dentry *dentry, @@ -119,8 +86,7 @@ static int efivarfs_create(struct inode *dir, struct dentry *dentry, /* length of the variable name itself: remove GUID and separator */ namelen = dentry->d_name.len - EFI_VARIABLE_GUID_LEN - 1; - efivarfs_hex_to_guid(dentry->d_name.name + namelen + 1, - &var->var.VendorGuid); + uuid_le_to_bin(dentry->d_name.name + namelen + 1, &var->var.VendorGuid); if (efivar_variable_is_removable(var->var.VendorGuid, dentry->d_name.name, namelen)) -- cgit v1.3-8-gc7d7 From 78a9be0a0a3367b94af242632c525d22b26f1a87 Mon Sep 17 00:00:00 2001 From: NeilBrown Date: Fri, 20 May 2016 17:03:51 -0700 Subject: dax: move RADIX_DAX_ definitions to dax.c These don't belong in radix-tree.h any more than PAGECACHE_TAG_* do. Let's try to maintain the idea that radix-tree simply implements an abstract data type. Signed-off-by: NeilBrown Reviewed-by: Ross Zwisler Reviewed-by: Jan Kara Signed-off-by: Matthew Wilcox Cc: Konstantin Khlebnikov Cc: Kirill Shutemov Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/dax.c | 9 +++++++++ include/linux/radix-tree.h | 9 --------- 2 files changed, 9 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/dax.c b/fs/dax.c index 0dbe4e0f16fe..a345c168acaa 100644 --- a/fs/dax.c +++ b/fs/dax.c @@ -32,6 +32,15 @@ #include #include +#define RADIX_DAX_MASK 0xf +#define RADIX_DAX_SHIFT 4 +#define RADIX_DAX_PTE (0x4 | RADIX_TREE_EXCEPTIONAL_ENTRY) +#define RADIX_DAX_PMD (0x8 | RADIX_TREE_EXCEPTIONAL_ENTRY) +#define RADIX_DAX_TYPE(entry) ((unsigned long)entry & RADIX_DAX_MASK) +#define RADIX_DAX_SECTOR(entry) (((unsigned long)entry >> RADIX_DAX_SHIFT)) +#define RADIX_DAX_ENTRY(sector, pmd) ((void *)((unsigned long)sector << \ + RADIX_DAX_SHIFT | (pmd ? RADIX_DAX_PMD : RADIX_DAX_PTE))) + static long dax_map_atomic(struct block_device *bdev, struct blk_dax_ctl *dax) { struct request_queue *q = bdev->bd_queue; diff --git a/include/linux/radix-tree.h b/include/linux/radix-tree.h index 11c8e7cc3920..c2f69e25ba86 100644 --- a/include/linux/radix-tree.h +++ b/include/linux/radix-tree.h @@ -48,15 +48,6 @@ #define RADIX_TREE_EXCEPTIONAL_ENTRY 2 #define RADIX_TREE_EXCEPTIONAL_SHIFT 2 -#define RADIX_DAX_MASK 0xf -#define RADIX_DAX_SHIFT 4 -#define RADIX_DAX_PTE (0x4 | RADIX_TREE_EXCEPTIONAL_ENTRY) -#define RADIX_DAX_PMD (0x8 | RADIX_TREE_EXCEPTIONAL_ENTRY) -#define RADIX_DAX_TYPE(entry) ((unsigned long)entry & RADIX_DAX_MASK) -#define RADIX_DAX_SECTOR(entry) (((unsigned long)entry >> RADIX_DAX_SHIFT)) -#define RADIX_DAX_ENTRY(sector, pmd) ((void *)((unsigned long)sector << \ - RADIX_DAX_SHIFT | (pmd ? RADIX_DAX_PMD : RADIX_DAX_PTE))) - static inline int radix_tree_is_internal_node(void *ptr) { return (int)((unsigned long)ptr & RADIX_TREE_INTERNAL_NODE); -- cgit v1.3-8-gc7d7 From 4108124f5c46efc951c4c6be947598a21d6b7fde Mon Sep 17 00:00:00 2001 From: Heloise Date: Fri, 20 May 2016 17:04:25 -0700 Subject: fs/efs/super.c: fix return value When sb_bread() fails, the return value should be -EIO, fix it. Link: http://lkml.kernel.org/r/1463464943-4142-1-git-send-email-os@iscas.ac.cn Signed-off-by: Heloise Cc: Johannes Weiner Cc: Firo Yang Cc: Vladimir Davydov Cc: Al Viro Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/efs/super.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/efs/super.c b/fs/efs/super.c index cb68dac4f9d3..368f7dd21c61 100644 --- a/fs/efs/super.c +++ b/fs/efs/super.c @@ -275,7 +275,7 @@ static int efs_fill_super(struct super_block *s, void *d, int silent) if (!bh) { pr_err("cannot read volume header\n"); - return -EINVAL; + return -EIO; } /* @@ -293,7 +293,7 @@ static int efs_fill_super(struct super_block *s, void *d, int silent) bh = sb_bread(s, sb->fs_start + EFS_SUPER); if (!bh) { pr_err("cannot read superblock\n"); - return -EINVAL; + return -EIO; } if (efs_validate_super(sb, (struct efs_super *) bh->b_data)) { -- cgit v1.3-8-gc7d7 From acc93d30d7d43f428272c20a047389c4cbca82ba Mon Sep 17 00:00:00 2001 From: Dan Williams Date: Sat, 7 May 2016 11:40:28 -0700 Subject: Revert "block: enable dax for raw block devices" This reverts commit 5a023cdba50c5f5f2bc351783b3131699deb3937. The functionality is superseded by the new "Device DAX" facility. Cc: Jeff Moyer Cc: Christoph Hellwig Cc: Dave Chinner Cc: Andrew Morton Cc: Ross Zwisler Cc: Jan Kara Signed-off-by: Dan Williams --- block/ioctl.c | 32 ----------------- fs/block_dev.c | 96 +++++++++++++++---------------------------------- include/linux/fs.h | 8 ----- include/uapi/linux/fs.h | 1 - 4 files changed, 29 insertions(+), 108 deletions(-) (limited to 'fs') diff --git a/block/ioctl.c b/block/ioctl.c index 4ff1f92f89ca..698c7933d582 100644 --- a/block/ioctl.c +++ b/block/ioctl.c @@ -407,35 +407,6 @@ static inline int is_unrecognized_ioctl(int ret) ret == -ENOIOCTLCMD; } -#ifdef CONFIG_FS_DAX -bool blkdev_dax_capable(struct block_device *bdev) -{ - struct gendisk *disk = bdev->bd_disk; - - if (!disk->fops->direct_access) - return false; - - /* - * If the partition is not aligned on a page boundary, we can't - * do dax I/O to it. - */ - if ((bdev->bd_part->start_sect % (PAGE_SIZE / 512)) - || (bdev->bd_part->nr_sects % (PAGE_SIZE / 512))) - return false; - - /* - * If the device has known bad blocks, force all I/O through the - * driver / page cache. - * - * TODO: support finer grained dax error handling - */ - if (disk->bb && disk->bb->count) - return false; - - return true; -} -#endif - static int blkdev_flushbuf(struct block_device *bdev, fmode_t mode, unsigned cmd, unsigned long arg) { @@ -598,9 +569,6 @@ int blkdev_ioctl(struct block_device *bdev, fmode_t mode, unsigned cmd, case BLKTRACESETUP: case BLKTRACETEARDOWN: return blk_trace_ioctl(bdev, cmd, argp); - case BLKDAXGET: - return put_int(arg, !!(bdev->bd_inode->i_flags & S_DAX)); - break; case IOC_PR_REGISTER: return blkdev_pr_register(bdev, argp); case IOC_PR_RESERVE: diff --git a/fs/block_dev.c b/fs/block_dev.c index 20a2c02b77c4..36ee10ca503e 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -29,6 +29,7 @@ #include #include #include +#include #include #include "internal.h" @@ -1159,6 +1160,33 @@ void bd_set_size(struct block_device *bdev, loff_t size) } EXPORT_SYMBOL(bd_set_size); +static bool blkdev_dax_capable(struct block_device *bdev) +{ + struct gendisk *disk = bdev->bd_disk; + + if (!disk->fops->direct_access || !IS_ENABLED(CONFIG_FS_DAX)) + return false; + + /* + * If the partition is not aligned on a page boundary, we can't + * do dax I/O to it. + */ + if ((bdev->bd_part->start_sect % (PAGE_SIZE / 512)) + || (bdev->bd_part->nr_sects % (PAGE_SIZE / 512))) + return false; + + /* + * If the device has known bad blocks, force all I/O through the + * driver / page cache. + * + * TODO: support finer grained dax error handling + */ + if (disk->bb && disk->bb->count) + return false; + + return true; +} + static void __blkdev_put(struct block_device *bdev, fmode_t mode, int for_part); /* @@ -1724,79 +1752,13 @@ static const struct address_space_operations def_blk_aops = { .is_dirty_writeback = buffer_check_dirty_writeback, }; -#ifdef CONFIG_FS_DAX -/* - * In the raw block case we do not need to contend with truncation nor - * unwritten file extents. Without those concerns there is no need for - * additional locking beyond the mmap_sem context that these routines - * are already executing under. - * - * Note, there is no protection if the block device is dynamically - * resized (partition grow/shrink) during a fault. A stable block device - * size is already not enforced in the blkdev_direct_IO path. - * - * For DAX, it is the responsibility of the block device driver to - * ensure the whole-disk device size is stable while requests are in - * flight. - * - * Finally, unlike the filemap_page_mkwrite() case there is no - * filesystem superblock to sync against freezing. We still include a - * pfn_mkwrite callback for dax drivers to receive write fault - * notifications. - */ -static int blkdev_dax_fault(struct vm_area_struct *vma, struct vm_fault *vmf) -{ - return __dax_fault(vma, vmf, blkdev_get_block, NULL); -} - -static int blkdev_dax_pfn_mkwrite(struct vm_area_struct *vma, - struct vm_fault *vmf) -{ - return dax_pfn_mkwrite(vma, vmf); -} - -static int blkdev_dax_pmd_fault(struct vm_area_struct *vma, unsigned long addr, - pmd_t *pmd, unsigned int flags) -{ - return __dax_pmd_fault(vma, addr, pmd, flags, blkdev_get_block, NULL); -} - -static const struct vm_operations_struct blkdev_dax_vm_ops = { - .fault = blkdev_dax_fault, - .pmd_fault = blkdev_dax_pmd_fault, - .pfn_mkwrite = blkdev_dax_pfn_mkwrite, -}; - -static const struct vm_operations_struct blkdev_default_vm_ops = { - .fault = filemap_fault, - .map_pages = filemap_map_pages, -}; - -static int blkdev_mmap(struct file *file, struct vm_area_struct *vma) -{ - struct inode *bd_inode = bdev_file_inode(file); - - file_accessed(file); - if (IS_DAX(bd_inode)) { - vma->vm_ops = &blkdev_dax_vm_ops; - vma->vm_flags |= VM_MIXEDMAP | VM_HUGEPAGE; - } else { - vma->vm_ops = &blkdev_default_vm_ops; - } - - return 0; -} -#else -#define blkdev_mmap generic_file_mmap -#endif - const struct file_operations def_blk_fops = { .open = blkdev_open, .release = blkdev_close, .llseek = block_llseek, .read_iter = blkdev_read_iter, .write_iter = blkdev_write_iter, - .mmap = blkdev_mmap, + .mmap = generic_file_mmap, .fsync = blkdev_fsync, .unlocked_ioctl = block_ioctl, #ifdef CONFIG_COMPAT diff --git a/include/linux/fs.h b/include/linux/fs.h index 70e61b58baaf..8363a10660f6 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -2320,14 +2320,6 @@ extern struct super_block *freeze_bdev(struct block_device *); extern void emergency_thaw_all(void); extern int thaw_bdev(struct block_device *bdev, struct super_block *sb); extern int fsync_bdev(struct block_device *); -#ifdef CONFIG_FS_DAX -extern bool blkdev_dax_capable(struct block_device *bdev); -#else -static inline bool blkdev_dax_capable(struct block_device *bdev) -{ - return false; -} -#endif extern struct super_block *blockdev_superblock; diff --git a/include/uapi/linux/fs.h b/include/uapi/linux/fs.h index a079d50376e1..fbff8b28aa35 100644 --- a/include/uapi/linux/fs.h +++ b/include/uapi/linux/fs.h @@ -222,7 +222,6 @@ struct fsxattr { #define BLKSECDISCARD _IO(0x12,125) #define BLKROTATIONAL _IO(0x12,126) #define BLKZEROOUT _IO(0x12,127) -#define BLKDAXGET _IO(0x12,129) #define BMAP_IOCTL 1 /* obsolete - kept for compatibility */ #define FIBMAP _IO(0x00,1) /* bmap access */ -- cgit v1.3-8-gc7d7 From f43edca7ed08fc02279f2a62015da5cb6aa0ad61 Mon Sep 17 00:00:00 2001 From: Ralf Baechle Date: Mon, 23 May 2016 16:22:26 -0700 Subject: ELF/MIPS build fix CONFIG_MIPS32_N32=y but CONFIG_BINFMT_ELF disabled results in the following linker errors: arch/mips/built-in.o: In function `elf_core_dump': binfmt_elfn32.c:(.text+0x23dbc): undefined reference to `elf_core_extra_phdrs' binfmt_elfn32.c:(.text+0x246e4): undefined reference to `elf_core_extra_data_size' binfmt_elfn32.c:(.text+0x248d0): undefined reference to `elf_core_write_extra_phdrs' binfmt_elfn32.c:(.text+0x24ac4): undefined reference to `elf_core_write_extra_data' CONFIG_MIPS32_O32=y but CONFIG_BINFMT_ELF disabled results in the following linker errors: arch/mips/built-in.o: In function `elf_core_dump': binfmt_elfo32.c:(.text+0x28a04): undefined reference to `elf_core_extra_phdrs' binfmt_elfo32.c:(.text+0x29330): undefined reference to `elf_core_extra_data_size' binfmt_elfo32.c:(.text+0x2951c): undefined reference to `elf_core_write_extra_phdrs' binfmt_elfo32.c:(.text+0x29710): undefined reference to `elf_core_write_extra_data' This is because binfmt_elfn32 and binfmt_elfo32 are using symbols from elfcore but for these configurations elfcore will not be built. Fixed by making elfcore selectable by a separate config symbol which unlike the current mechanism can also be used from other directories than kernel/, then having each flavor of ELF that relies on elfcore.o, select it in Kconfig, including CONFIG_MIPS32_N32 and CONFIG_MIPS32_O32 which fixes this issue. Link: http://lkml.kernel.org/r/20160520141705.GA1913@linux-mips.org Signed-off-by: Ralf Baechle Reviewed-by: James Hogan Cc: "Maciej W. Rozycki" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/mips/Kconfig | 1 + fs/Kconfig.binfmt | 8 ++++++++ kernel/Makefile | 4 +--- 3 files changed, 10 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/arch/mips/Kconfig b/arch/mips/Kconfig index 8040fb1845b4..46938847e794 100644 --- a/arch/mips/Kconfig +++ b/arch/mips/Kconfig @@ -3117,6 +3117,7 @@ config MIPS32_N32 config BINFMT_ELF32 bool default y if MIPS32_O32 || MIPS32_N32 + select ELFCORE endmenu diff --git a/fs/Kconfig.binfmt b/fs/Kconfig.binfmt index 2d0cbbd14cfc..72c03354c14b 100644 --- a/fs/Kconfig.binfmt +++ b/fs/Kconfig.binfmt @@ -1,6 +1,7 @@ config BINFMT_ELF bool "Kernel support for ELF binaries" depends on MMU && (BROKEN || !FRV) + select ELFCORE default y ---help--- ELF (Executable and Linkable Format) is a format for libraries and @@ -26,6 +27,7 @@ config BINFMT_ELF config COMPAT_BINFMT_ELF bool depends on COMPAT && BINFMT_ELF + select ELFCORE config ARCH_BINFMT_ELF_STATE bool @@ -34,6 +36,7 @@ config BINFMT_ELF_FDPIC bool "Kernel support for FDPIC ELF binaries" default y depends on (FRV || BLACKFIN || (SUPERH32 && !MMU) || C6X) + select ELFCORE help ELF FDPIC binaries are based on ELF, but allow the individual load segments of a binary to be located in memory independently of each @@ -43,6 +46,11 @@ config BINFMT_ELF_FDPIC It is also possible to run FDPIC ELF binaries on MMU linux also. +config ELFCORE + bool + help + This option enables kernel/elfcore.o. + config CORE_DUMP_DEFAULT_ELF_HEADERS bool "Write ELF core dumps with partial segments" default y diff --git a/kernel/Makefile b/kernel/Makefile index f0c40bf49d9f..e2ec54e2b952 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -91,9 +91,7 @@ obj-$(CONFIG_TASK_DELAY_ACCT) += delayacct.o obj-$(CONFIG_TASKSTATS) += taskstats.o tsacct.o obj-$(CONFIG_TRACEPOINTS) += tracepoint.o obj-$(CONFIG_LATENCYTOP) += latencytop.o -obj-$(CONFIG_BINFMT_ELF) += elfcore.o -obj-$(CONFIG_COMPAT_BINFMT_ELF) += elfcore.o -obj-$(CONFIG_BINFMT_ELF_FDPIC) += elfcore.o +obj-$(CONFIG_ELFCORE) += elfcore.o obj-$(CONFIG_FUNCTION_TRACER) += trace/ obj-$(CONFIG_TRACING) += trace/ obj-$(CONFIG_TRACE_CLOCK) += trace/ -- cgit v1.3-8-gc7d7 From ff2fe0aa2e17ff78c676d9c450b819c8b87af801 Mon Sep 17 00:00:00 2001 From: Salah Triki Date: Mon, 23 May 2016 16:22:32 -0700 Subject: fs/befs/datastream.c:befs_read_datastream(): remove unneeded initialization to NULL bh is reinitialized by befs_bread_iaddr() so no need to init it with NULL in the beginning of befs_read_datastream(). Link: http://lkml.kernel.org/r/81e1f70187db34d195c8e42b1ff78be6a71c0060.1462649034.git.salah.triki@acm.org Signed-off-by: Salah Triki Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/befs/datastream.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/befs/datastream.c b/fs/befs/datastream.c index dde0b79f3948..f6ad38e15203 100644 --- a/fs/befs/datastream.c +++ b/fs/befs/datastream.c @@ -48,7 +48,7 @@ struct buffer_head * befs_read_datastream(struct super_block *sb, const befs_data_stream *ds, befs_off_t pos, uint * off) { - struct buffer_head *bh = NULL; + struct buffer_head *bh; befs_block_run run; befs_blocknr_t block; /* block coresponding to pos */ -- cgit v1.3-8-gc7d7 From 3080ea9e5bb4cb337b7f061a085645b9dabf365f Mon Sep 17 00:00:00 2001 From: Salah Triki Date: Mon, 23 May 2016 16:22:35 -0700 Subject: fs/befs/datastream.c:befs_read_lsymlink(): remove unneeded initialization to NULL bh is reinitialized by befs_read_datastream() so no need to init it with NULL in the beginning of befs_read_lsymlink(). Link: http://lkml.kernel.org/r/e22f279bceb8d026af048952e02ba98925b21c92.1462649034.git.salah.triki@acm.org Signed-off-by: Salah Triki Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/befs/datastream.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/befs/datastream.c b/fs/befs/datastream.c index f6ad38e15203..e3b31291e3fa 100644 --- a/fs/befs/datastream.c +++ b/fs/befs/datastream.c @@ -127,7 +127,7 @@ befs_read_lsymlink(struct super_block *sb, const befs_data_stream *ds, { befs_off_t bytes_read = 0; /* bytes readed */ u16 plen; - struct buffer_head *bh = NULL; + struct buffer_head *bh; befs_debug(sb, "---> %s length: %llu", __func__, len); while (bytes_read < len) { -- cgit v1.3-8-gc7d7 From 6c8da24b903851656dc565e8d2336b72a35efe14 Mon Sep 17 00:00:00 2001 From: Salah Triki Date: Mon, 23 May 2016 16:22:38 -0700 Subject: fs/befs/datastream.c:befs_find_brun_dblindirect(): remove unneeded initializations to NULL iaddr_array is unconditionally initialized to NULL in befs_find_brun_dblindirect(). Link: http://lkml.kernel.org/r/940def273e30ef37957fba9da6981a10fb3c9741.1462649034.git.salah.triki@acm.org Signed-off-by: Salah Triki Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/befs/datastream.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/befs/datastream.c b/fs/befs/datastream.c index e3b31291e3fa..af1bc19b7c85 100644 --- a/fs/befs/datastream.c +++ b/fs/befs/datastream.c @@ -429,7 +429,7 @@ befs_find_brun_dblindirect(struct super_block *sb, struct buffer_head *dbl_indir_block; struct buffer_head *indir_block; befs_block_run indir_run; - befs_disk_inode_addr *iaddr_array = NULL; + befs_disk_inode_addr *iaddr_array; struct befs_sb_info *befs_sb = BEFS_SB(sb); befs_blocknr_t indir_start_blk = @@ -488,7 +488,6 @@ befs_find_brun_dblindirect(struct super_block *sb, iaddr_array = (befs_disk_inode_addr *) dbl_indir_block->b_data; indir_run = fsrun_to_cpu(sb, iaddr_array[dbl_block_indx]); brelse(dbl_indir_block); - iaddr_array = NULL; /* Read indirect block */ which_block = indir_indx / befs_iaddrs_per_block(sb); @@ -513,7 +512,6 @@ befs_find_brun_dblindirect(struct super_block *sb, iaddr_array = (befs_disk_inode_addr *) indir_block->b_data; *run = fsrun_to_cpu(sb, iaddr_array[block_indx]); brelse(indir_block); - iaddr_array = NULL; blockno_at_run_start = indir_start_blk; blockno_at_run_start += diblklen * dblindir_indx; -- cgit v1.3-8-gc7d7 From c940876368982afa9c5413f23bf70766dbfa8ac0 Mon Sep 17 00:00:00 2001 From: Salah Triki Date: Mon, 23 May 2016 16:22:40 -0700 Subject: fs/befs/linuxvfs.c:befs_get_block(): remove unneeded initialization to NULL inode is reinitialized by befs_iget() so no need to init it with NULL in the beginning of befs_lookup() Link: http://lkml.kernel.org/r/03d7e46890aef94078130bed97c4f8f8ae9ea2b2.1462842886.git.salah.triki@acm.org Signed-off-by: Salah Triki Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/befs/linuxvfs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/befs/linuxvfs.c b/fs/befs/linuxvfs.c index 71112aa07d84..ed6e712d3972 100644 --- a/fs/befs/linuxvfs.c +++ b/fs/befs/linuxvfs.c @@ -155,7 +155,7 @@ befs_get_block(struct inode *inode, sector_t block, static struct dentry * befs_lookup(struct inode *dir, struct dentry *dentry, unsigned int flags) { - struct inode *inode = NULL; + struct inode *inode; struct super_block *sb = dir->i_sb; const befs_data_stream *ds = &BEFS_I(dir)->i_data.ds; befs_off_t offset; -- cgit v1.3-8-gc7d7 From e4c7f5b10931595097815c9d36e041efb49a42c7 Mon Sep 17 00:00:00 2001 From: Salah Triki Date: Mon, 23 May 2016 16:22:43 -0700 Subject: fs/befs/linuxvfs.c:befs_iget(): remove unneeded initialization to NULL bh is reinitialized by befs_bread() so no need to init it with NULL in the beginning of befs_iget() Link: http://lkml.kernel.org/r/38d62b1469bc3b316ba6b81fd8e26fc66fdd713b.1462842886.git.salah.triki@acm.org Signed-off-by: Salah Triki Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/befs/linuxvfs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/befs/linuxvfs.c b/fs/befs/linuxvfs.c index ed6e712d3972..410315400755 100644 --- a/fs/befs/linuxvfs.c +++ b/fs/befs/linuxvfs.c @@ -294,7 +294,7 @@ static void init_once(void *foo) static struct inode *befs_iget(struct super_block *sb, unsigned long ino) { - struct buffer_head *bh = NULL; + struct buffer_head *bh; befs_inode *raw_inode = NULL; struct befs_sb_info *befs_sb = BEFS_SB(sb); struct befs_inode_info *befs_ino = NULL; -- cgit v1.3-8-gc7d7 From 44ad809865299c7c961161c45094e1bb3a12eaa0 Mon Sep 17 00:00:00 2001 From: Salah Triki Date: Mon, 23 May 2016 16:22:46 -0700 Subject: fs/befs/linuxvfs.c:befs_iget(): remove unneeded raw_inode initialization to NULL raw_inode is reinitialized to bh->b_data so no need to init it with NULL in the beginning of befs_iget() Link: http://lkml.kernel.org/r/0a66baaaacb6b7e5fcea5b31b57b649261152281.1462842887.git.salah.triki@acm.org Signed-off-by: Salah Triki Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/befs/linuxvfs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/befs/linuxvfs.c b/fs/befs/linuxvfs.c index 410315400755..711214064dbb 100644 --- a/fs/befs/linuxvfs.c +++ b/fs/befs/linuxvfs.c @@ -295,7 +295,7 @@ static void init_once(void *foo) static struct inode *befs_iget(struct super_block *sb, unsigned long ino) { struct buffer_head *bh; - befs_inode *raw_inode = NULL; + befs_inode *raw_inode; struct befs_sb_info *befs_sb = BEFS_SB(sb); struct befs_inode_info *befs_ino = NULL; struct inode *inode; -- cgit v1.3-8-gc7d7 From d6e0203369ec2f9425e6d6e7ced7b365f5b97bf3 Mon Sep 17 00:00:00 2001 From: Salah Triki Date: Mon, 23 May 2016 16:22:49 -0700 Subject: fs/befs/linuxvfs.c:befs_iget(): remove unneeded befs_nio initialization to NULL befs_ino is reinitialized by BEFS_I() so no need to init it with NULL in the beginning of befs_iget() Link: http://lkml.kernel.org/r/a5c02445e436629c4d4ba1b65d91501878942f58.1462842887.git.salah.triki@acm.org Signed-off-by: Salah Triki Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/befs/linuxvfs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/befs/linuxvfs.c b/fs/befs/linuxvfs.c index 711214064dbb..7da05b159ade 100644 --- a/fs/befs/linuxvfs.c +++ b/fs/befs/linuxvfs.c @@ -297,7 +297,7 @@ static struct inode *befs_iget(struct super_block *sb, unsigned long ino) struct buffer_head *bh; befs_inode *raw_inode; struct befs_sb_info *befs_sb = BEFS_SB(sb); - struct befs_inode_info *befs_ino = NULL; + struct befs_inode_info *befs_ino; struct inode *inode; long ret = -EIO; -- cgit v1.3-8-gc7d7 From 77169af8b90c96ac40e49ba45da6520dcccf5d8e Mon Sep 17 00:00:00 2001 From: Salah Triki Date: Mon, 23 May 2016 16:22:52 -0700 Subject: fs/befs/io.c:befs_bread_iaddr(): remove unneeded initialization to NULL bh is reinitialized by sb_bread() so no need to init it with NULL in the beginning of befs_bread_iaddr() Link: http://lkml.kernel.org/r/586d2639d729345b9c07aac10ba713d8ceb8745a.1462841692.git.salah.triki@acm.org Signed-off-by: Salah Triki Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/befs/io.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/befs/io.c b/fs/befs/io.c index 7a5b4ec21c56..1341fa362284 100644 --- a/fs/befs/io.c +++ b/fs/befs/io.c @@ -26,7 +26,7 @@ struct buffer_head * befs_bread_iaddr(struct super_block *sb, befs_inode_addr iaddr) { - struct buffer_head *bh = NULL; + struct buffer_head *bh; befs_blocknr_t block = 0; struct befs_sb_info *befs_sb = BEFS_SB(sb); -- cgit v1.3-8-gc7d7 From a46d8c3fc7f520632e773ff51209e98770a56846 Mon Sep 17 00:00:00 2001 From: Salah Triki Date: Mon, 23 May 2016 16:22:55 -0700 Subject: fs/befs/io.c:befs_bread(): remove unneeded initialization to NULL bh is reinitialized by sb_bread() so no need to init it with NULL in the beginning of befs_bread() Link: http://lkml.kernel.org/r/88481760b43226fac16adb1f1e68897e47d8235c.1462841692.git.salah.triki@acm.org Signed-off-by: Salah Triki Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/befs/io.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/befs/io.c b/fs/befs/io.c index 1341fa362284..523c8af2d770 100644 --- a/fs/befs/io.c +++ b/fs/befs/io.c @@ -63,7 +63,7 @@ befs_bread_iaddr(struct super_block *sb, befs_inode_addr iaddr) struct buffer_head * befs_bread(struct super_block *sb, befs_blocknr_t block) { - struct buffer_head *bh = NULL; + struct buffer_head *bh; befs_debug(sb, "---> Enter %s %lu", __func__, (unsigned long)block); -- cgit v1.3-8-gc7d7 From 1c613cb9868359031eb1344a33ac661908804bab Mon Sep 17 00:00:00 2001 From: Julia Lawall Date: Mon, 23 May 2016 16:22:57 -0700 Subject: nilfs2: constify nilfs_sc_operations structures The nilfs_sc_operations structures are never modified, so declare them as const. Done with the help of Coccinelle. Signed-off-by: Julia Lawall Signed-off-by: Ryusuke Konishi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nilfs2/segment.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/nilfs2/segment.c b/fs/nilfs2/segment.c index 4317f72568e6..483e6634466f 100644 --- a/fs/nilfs2/segment.c +++ b/fs/nilfs2/segment.c @@ -617,7 +617,7 @@ static void nilfs_write_file_node_binfo(struct nilfs_sc_info *sci, *vblocknr = binfo->bi_v.bi_vblocknr; } -static struct nilfs_sc_operations nilfs_sc_file_ops = { +static const struct nilfs_sc_operations nilfs_sc_file_ops = { .collect_data = nilfs_collect_file_data, .collect_node = nilfs_collect_file_node, .collect_bmap = nilfs_collect_file_bmap, @@ -666,7 +666,7 @@ static void nilfs_write_dat_node_binfo(struct nilfs_sc_info *sci, *binfo_dat = binfo->bi_dat; } -static struct nilfs_sc_operations nilfs_sc_dat_ops = { +static const struct nilfs_sc_operations nilfs_sc_dat_ops = { .collect_data = nilfs_collect_dat_data, .collect_node = nilfs_collect_file_node, .collect_bmap = nilfs_collect_dat_bmap, @@ -674,7 +674,7 @@ static struct nilfs_sc_operations nilfs_sc_dat_ops = { .write_node_binfo = nilfs_write_dat_node_binfo, }; -static struct nilfs_sc_operations nilfs_sc_dsync_ops = { +static const struct nilfs_sc_operations nilfs_sc_dsync_ops = { .collect_data = nilfs_collect_file_data, .collect_node = NULL, .collect_bmap = NULL, @@ -1043,7 +1043,7 @@ static size_t nilfs_segctor_buffer_rest(struct nilfs_sc_info *sci) static int nilfs_segctor_scan_file(struct nilfs_sc_info *sci, struct inode *inode, - struct nilfs_sc_operations *sc_ops) + const struct nilfs_sc_operations *sc_ops) { LIST_HEAD(data_buffers); LIST_HEAD(node_buffers); @@ -1550,7 +1550,7 @@ nilfs_segctor_update_payload_blocknr(struct nilfs_sc_info *sci, sector_t blocknr; unsigned long nfinfo = segbuf->sb_sum.nfinfo; unsigned long nblocks = 0, ndatablk = 0; - struct nilfs_sc_operations *sc_op = NULL; + const struct nilfs_sc_operations *sc_op = NULL; struct nilfs_segsum_pointer ssp; struct nilfs_finfo *finfo = NULL; union nilfs_binfo binfo; -- cgit v1.3-8-gc7d7 From 8fa7c32094d806c23b35f9f51bcefb3148d3eb54 Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Mon, 23 May 2016 16:23:00 -0700 Subject: nilfs2: fix white space issue in nilfs_mount() Fix the following checkpatch.pl error and warnings: ERROR: code indent should use tabs where possible #1317: FILE: super.c:1317: + ^I^Is_new = true;$ WARNING: please, no space before tabs #1317: FILE: super.c:1317: + ^I^Is_new = true;$ WARNING: please, no spaces at the start of a line #1317: FILE: super.c:1317: + ^I^Is_new = true;$ Link: http://lkml.kernel.org/r/1461935747-10380-2-git-send-email-konishi.ryusuke@lab.ntt.co.jp Signed-off-by: Ryusuke Konishi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nilfs2/super.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nilfs2/super.c b/fs/nilfs2/super.c index 7f5d3d9f1c37..e05d1848164a 100644 --- a/fs/nilfs2/super.c +++ b/fs/nilfs2/super.c @@ -1316,7 +1316,7 @@ nilfs_mount(struct file_system_type *fs_type, int flags, } if (!s->s_root) { - s_new = true; + s_new = true; /* New superblock instance created */ s->s_mode = mode; -- cgit v1.3-8-gc7d7 From f19e78dee9d5adc90af086175f8e3640da3f613c Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Mon, 23 May 2016 16:23:03 -0700 Subject: nilfs2: remove space before comma Fix checkpatch.pl error "ERROR: space prohibited before that ',' (ctx:WxW)" at nilfs_sufile_set_suinfo(). This also fixes checkpatch.pl warning "WARNING: Prefer 'unsigned int' to bare use of 'unsigned'" at nilfs_sufile_set_suinfo() and nilfs_sufile_get_suinfo(). Link: http://lkml.kernel.org/r/1461935747-10380-3-git-send-email-konishi.ryusuke@lab.ntt.co.jp Signed-off-by: Ryusuke Konishi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nilfs2/sufile.c | 4 ++-- fs/nilfs2/sufile.h | 4 ++-- 2 files changed, 4 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/nilfs2/sufile.c b/fs/nilfs2/sufile.c index 52821ffc11f4..5b3720414e34 100644 --- a/fs/nilfs2/sufile.c +++ b/fs/nilfs2/sufile.c @@ -819,7 +819,7 @@ out: * %-ENOMEM - Insufficient amount of memory available. */ ssize_t nilfs_sufile_get_suinfo(struct inode *sufile, __u64 segnum, void *buf, - unsigned sisz, size_t nsi) + unsigned int sisz, size_t nsi) { struct buffer_head *su_bh; struct nilfs_segment_usage *su; @@ -897,7 +897,7 @@ ssize_t nilfs_sufile_get_suinfo(struct inode *sufile, __u64 segnum, void *buf, * %-EINVAL - Invalid values in input (segment number, flags or nblocks) */ ssize_t nilfs_sufile_set_suinfo(struct inode *sufile, void *buf, - unsigned supsz, size_t nsup) + unsigned int supsz, size_t nsup) { struct the_nilfs *nilfs = sufile->i_sb->s_fs_info; struct buffer_head *header_bh, *bh; diff --git a/fs/nilfs2/sufile.h b/fs/nilfs2/sufile.h index b8afd72f2379..35e34c84d4af 100644 --- a/fs/nilfs2/sufile.h +++ b/fs/nilfs2/sufile.h @@ -42,9 +42,9 @@ int nilfs_sufile_mark_dirty(struct inode *sufile, __u64 segnum); int nilfs_sufile_set_segment_usage(struct inode *sufile, __u64 segnum, unsigned long nblocks, time_t modtime); int nilfs_sufile_get_stat(struct inode *, struct nilfs_sustat *); -ssize_t nilfs_sufile_get_suinfo(struct inode *, __u64, void *, unsigned, +ssize_t nilfs_sufile_get_suinfo(struct inode *, __u64, void *, unsigned int, size_t); -ssize_t nilfs_sufile_set_suinfo(struct inode *, void *, unsigned , size_t); +ssize_t nilfs_sufile_set_suinfo(struct inode *, void *, unsigned int, size_t); int nilfs_sufile_updatev(struct inode *, __u64 *, size_t, int, size_t *, void (*dofunc)(struct inode *, __u64, -- cgit v1.3-8-gc7d7 From 5726d0b454614a47e641a04c8106392d67a8e1ad Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Mon, 23 May 2016 16:23:06 -0700 Subject: nilfs2: remove FSF mailing address from GPL notices This removes the extra paragraph which mentions FSF address in GPL notices from source code of nilfs2 and avoids the checkpatch.pl error related to it. Link: http://lkml.kernel.org/r/1461935747-10380-4-git-send-email-konishi.ryusuke@lab.ntt.co.jp Signed-off-by: Ryusuke Konishi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nilfs2/alloc.c | 4 ---- fs/nilfs2/alloc.h | 4 ---- fs/nilfs2/bmap.c | 4 ---- fs/nilfs2/bmap.h | 4 ---- fs/nilfs2/btnode.c | 4 ---- fs/nilfs2/btnode.h | 4 ---- fs/nilfs2/btree.c | 4 ---- fs/nilfs2/btree.h | 4 ---- fs/nilfs2/cpfile.c | 4 ---- fs/nilfs2/cpfile.h | 4 ---- fs/nilfs2/dat.c | 4 ---- fs/nilfs2/dat.h | 4 ---- fs/nilfs2/dir.c | 4 ---- fs/nilfs2/direct.c | 4 ---- fs/nilfs2/direct.h | 4 ---- fs/nilfs2/file.c | 4 ---- fs/nilfs2/gcinode.c | 4 ---- fs/nilfs2/ifile.c | 4 ---- fs/nilfs2/ifile.h | 4 ---- fs/nilfs2/inode.c | 4 ---- fs/nilfs2/ioctl.c | 4 ---- fs/nilfs2/mdt.c | 4 ---- fs/nilfs2/mdt.h | 4 ---- fs/nilfs2/namei.c | 4 ---- fs/nilfs2/nilfs.h | 4 ---- fs/nilfs2/page.c | 4 ---- fs/nilfs2/page.h | 4 ---- fs/nilfs2/recovery.c | 4 ---- fs/nilfs2/segbuf.c | 4 ---- fs/nilfs2/segbuf.h | 4 ---- fs/nilfs2/segment.c | 4 ---- fs/nilfs2/segment.h | 4 ---- fs/nilfs2/sufile.c | 4 ---- fs/nilfs2/sufile.h | 4 ---- fs/nilfs2/super.c | 4 ---- fs/nilfs2/the_nilfs.c | 4 ---- fs/nilfs2/the_nilfs.h | 4 ---- include/linux/nilfs2_fs.h | 4 ---- 38 files changed, 152 deletions(-) (limited to 'fs') diff --git a/fs/nilfs2/alloc.c b/fs/nilfs2/alloc.c index 2ccbf5531554..eaa0c6af80c3 100644 --- a/fs/nilfs2/alloc.c +++ b/fs/nilfs2/alloc.c @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Original code was written by Koji Sato . * Two allocators were unified by Ryusuke Konishi , * Amagai Yoshiji . diff --git a/fs/nilfs2/alloc.h b/fs/nilfs2/alloc.h index 6e6f49aa53df..62982eea63c0 100644 --- a/fs/nilfs2/alloc.h +++ b/fs/nilfs2/alloc.h @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Original code was written by Koji Sato . * Two allocators were unified by Ryusuke Konishi , * Amagai Yoshiji . diff --git a/fs/nilfs2/bmap.c b/fs/nilfs2/bmap.c index a9fb3636c142..fcd79e611c87 100644 --- a/fs/nilfs2/bmap.c +++ b/fs/nilfs2/bmap.c @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Koji Sato . */ diff --git a/fs/nilfs2/bmap.h b/fs/nilfs2/bmap.h index bfa817ce40b3..baa6d404f369 100644 --- a/fs/nilfs2/bmap.h +++ b/fs/nilfs2/bmap.h @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Koji Sato . */ diff --git a/fs/nilfs2/btnode.c b/fs/nilfs2/btnode.c index e0c9daf9aa22..55241effa3c0 100644 --- a/fs/nilfs2/btnode.c +++ b/fs/nilfs2/btnode.c @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * This file was originally written by Seiji Kihara * and fully revised by Ryusuke Konishi for * stabilization and simplification. diff --git a/fs/nilfs2/btnode.h b/fs/nilfs2/btnode.h index d876b565ce64..0727096a8c53 100644 --- a/fs/nilfs2/btnode.h +++ b/fs/nilfs2/btnode.h @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Seiji Kihara * Revised by Ryusuke Konishi */ diff --git a/fs/nilfs2/btree.c b/fs/nilfs2/btree.c index 3a3821b00486..cc748ff3b876 100644 --- a/fs/nilfs2/btree.c +++ b/fs/nilfs2/btree.c @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Koji Sato . */ diff --git a/fs/nilfs2/btree.h b/fs/nilfs2/btree.h index 22c02e35b6ef..9497732b3b1d 100644 --- a/fs/nilfs2/btree.h +++ b/fs/nilfs2/btree.h @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Koji Sato . */ diff --git a/fs/nilfs2/cpfile.c b/fs/nilfs2/cpfile.c index b6596cab9e99..15016eca4dc4 100644 --- a/fs/nilfs2/cpfile.c +++ b/fs/nilfs2/cpfile.c @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Koji Sato . */ diff --git a/fs/nilfs2/cpfile.h b/fs/nilfs2/cpfile.h index a242b9a314f9..3dcaceb66be6 100644 --- a/fs/nilfs2/cpfile.h +++ b/fs/nilfs2/cpfile.h @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Koji Sato . */ diff --git a/fs/nilfs2/dat.c b/fs/nilfs2/dat.c index 7dc23f100e57..7427c114aa11 100644 --- a/fs/nilfs2/dat.c +++ b/fs/nilfs2/dat.c @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Koji Sato . */ diff --git a/fs/nilfs2/dat.h b/fs/nilfs2/dat.h index cbd8e9732503..edd0586220d4 100644 --- a/fs/nilfs2/dat.h +++ b/fs/nilfs2/dat.h @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Koji Sato . */ diff --git a/fs/nilfs2/dir.c b/fs/nilfs2/dir.c index 6723d45a631a..955070bd20a5 100644 --- a/fs/nilfs2/dir.c +++ b/fs/nilfs2/dir.c @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Modified for NILFS by Amagai Yoshiji */ /* diff --git a/fs/nilfs2/direct.c b/fs/nilfs2/direct.c index ebf89fd8ac1a..24700229d6d4 100644 --- a/fs/nilfs2/direct.c +++ b/fs/nilfs2/direct.c @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Koji Sato . */ diff --git a/fs/nilfs2/direct.h b/fs/nilfs2/direct.h index dc643de20a25..e4eb1b7e407c 100644 --- a/fs/nilfs2/direct.h +++ b/fs/nilfs2/direct.h @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Koji Sato . */ diff --git a/fs/nilfs2/file.c b/fs/nilfs2/file.c index 088ba001c6ef..49243ca6e3e9 100644 --- a/fs/nilfs2/file.c +++ b/fs/nilfs2/file.c @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Amagai Yoshiji , * Ryusuke Konishi */ diff --git a/fs/nilfs2/gcinode.c b/fs/nilfs2/gcinode.c index 0224b7826ace..b0321886ce33 100644 --- a/fs/nilfs2/gcinode.c +++ b/fs/nilfs2/gcinode.c @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Seiji Kihara , Amagai Yoshiji , * and Ryusuke Konishi . * Revised by Ryusuke Konishi . diff --git a/fs/nilfs2/ifile.c b/fs/nilfs2/ifile.c index 6548c7851b48..06f9b5aa6175 100644 --- a/fs/nilfs2/ifile.c +++ b/fs/nilfs2/ifile.c @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Amagai Yoshiji . * Revised by Ryusuke Konishi . * diff --git a/fs/nilfs2/ifile.h b/fs/nilfs2/ifile.h index 679674d13372..c61637e444f1 100644 --- a/fs/nilfs2/ifile.h +++ b/fs/nilfs2/ifile.h @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Amagai Yoshiji * Revised by Ryusuke Konishi * diff --git a/fs/nilfs2/inode.c b/fs/nilfs2/inode.c index cfebcd2fc7f3..1bdb17bb01d3 100644 --- a/fs/nilfs2/inode.c +++ b/fs/nilfs2/inode.c @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Ryusuke Konishi * */ diff --git a/fs/nilfs2/ioctl.c b/fs/nilfs2/ioctl.c index e8fe24882b5b..4db0e5ff6c67 100644 --- a/fs/nilfs2/ioctl.c +++ b/fs/nilfs2/ioctl.c @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Koji Sato . */ diff --git a/fs/nilfs2/mdt.c b/fs/nilfs2/mdt.c index f6982b9153d5..0ab93bdfa261 100644 --- a/fs/nilfs2/mdt.c +++ b/fs/nilfs2/mdt.c @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Ryusuke Konishi */ diff --git a/fs/nilfs2/mdt.h b/fs/nilfs2/mdt.h index 03246cac3338..5fc07ecf32a3 100644 --- a/fs/nilfs2/mdt.h +++ b/fs/nilfs2/mdt.h @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Ryusuke Konishi */ diff --git a/fs/nilfs2/namei.c b/fs/nilfs2/namei.c index 3b2af05f9fb4..db0a4bcf6015 100644 --- a/fs/nilfs2/namei.c +++ b/fs/nilfs2/namei.c @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Modified for NILFS by Amagai Yoshiji , * Ryusuke Konishi */ diff --git a/fs/nilfs2/nilfs.h b/fs/nilfs2/nilfs.h index 385704027575..e2089a18b24b 100644 --- a/fs/nilfs2/nilfs.h +++ b/fs/nilfs2/nilfs.h @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Koji Sato * Ryusuke Konishi */ diff --git a/fs/nilfs2/page.c b/fs/nilfs2/page.c index 489391561cda..b10057eb4d8e 100644 --- a/fs/nilfs2/page.c +++ b/fs/nilfs2/page.c @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Ryusuke Konishi , * Seiji Kihara . */ diff --git a/fs/nilfs2/page.h b/fs/nilfs2/page.h index a43b8287d012..6df10e127678 100644 --- a/fs/nilfs2/page.h +++ b/fs/nilfs2/page.h @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Ryusuke Konishi , * Seiji Kihara . */ diff --git a/fs/nilfs2/recovery.c b/fs/nilfs2/recovery.c index 5afa77fadc11..2ae1dfccf6e8 100644 --- a/fs/nilfs2/recovery.c +++ b/fs/nilfs2/recovery.c @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Ryusuke Konishi */ diff --git a/fs/nilfs2/segbuf.c b/fs/nilfs2/segbuf.c index f63620ce3892..8aef62d81553 100644 --- a/fs/nilfs2/segbuf.c +++ b/fs/nilfs2/segbuf.c @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Ryusuke Konishi * */ diff --git a/fs/nilfs2/segbuf.h b/fs/nilfs2/segbuf.h index b04f08cc2397..fa4ed03744bb 100644 --- a/fs/nilfs2/segbuf.h +++ b/fs/nilfs2/segbuf.h @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Ryusuke Konishi * */ diff --git a/fs/nilfs2/segment.c b/fs/nilfs2/segment.c index 483e6634466f..cafb61c32b79 100644 --- a/fs/nilfs2/segment.c +++ b/fs/nilfs2/segment.c @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Ryusuke Konishi * */ diff --git a/fs/nilfs2/segment.h b/fs/nilfs2/segment.h index 0408b9b2814b..a118eabd3ed6 100644 --- a/fs/nilfs2/segment.h +++ b/fs/nilfs2/segment.h @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Ryusuke Konishi * */ diff --git a/fs/nilfs2/sufile.c b/fs/nilfs2/sufile.c index 5b3720414e34..1d602482b51e 100644 --- a/fs/nilfs2/sufile.c +++ b/fs/nilfs2/sufile.c @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Koji Sato . * Revised by Ryusuke Konishi . */ diff --git a/fs/nilfs2/sufile.h b/fs/nilfs2/sufile.h index 35e34c84d4af..c4cf47e14b70 100644 --- a/fs/nilfs2/sufile.h +++ b/fs/nilfs2/sufile.h @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Koji Sato . */ diff --git a/fs/nilfs2/super.c b/fs/nilfs2/super.c index e05d1848164a..d304faa66f49 100644 --- a/fs/nilfs2/super.c +++ b/fs/nilfs2/super.c @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Ryusuke Konishi */ /* diff --git a/fs/nilfs2/the_nilfs.c b/fs/nilfs2/the_nilfs.c index 69bd801afb53..8e270128ae86 100644 --- a/fs/nilfs2/the_nilfs.c +++ b/fs/nilfs2/the_nilfs.c @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Ryusuke Konishi * */ diff --git a/fs/nilfs2/the_nilfs.h b/fs/nilfs2/the_nilfs.h index 23778d385836..426e4acc772b 100644 --- a/fs/nilfs2/the_nilfs.h +++ b/fs/nilfs2/the_nilfs.h @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Ryusuke Konishi * */ diff --git a/include/linux/nilfs2_fs.h b/include/linux/nilfs2_fs.h index e9fcf90b270d..0d7f8141382a 100644 --- a/include/linux/nilfs2_fs.h +++ b/include/linux/nilfs2_fs.h @@ -13,10 +13,6 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU Lesser General Public License for more details. * - * You should have received a copy of the GNU Lesser General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Written by Koji Sato * Ryusuke Konishi */ -- cgit v1.3-8-gc7d7 From 4b420ab4eedc7a816ad0f2278871019de1a8ccef Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Mon, 23 May 2016 16:23:09 -0700 Subject: nilfs2: clean up old e-mail addresses E-mail addresses of osrg.net domain are no longer available. This removes them from authorship notices and prevents reporters from being confused. Link: http://lkml.kernel.org/r/1461935747-10380-5-git-send-email-konishi.ryusuke@lab.ntt.co.jp Signed-off-by: Ryusuke Konishi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nilfs2/alloc.c | 5 ++--- fs/nilfs2/alloc.h | 5 ++--- fs/nilfs2/bmap.c | 2 +- fs/nilfs2/bmap.h | 2 +- fs/nilfs2/btnode.c | 5 ++--- fs/nilfs2/btnode.h | 4 ++-- fs/nilfs2/btree.c | 2 +- fs/nilfs2/btree.h | 2 +- fs/nilfs2/cpfile.c | 2 +- fs/nilfs2/cpfile.h | 2 +- fs/nilfs2/dat.c | 2 +- fs/nilfs2/dat.h | 2 +- fs/nilfs2/dir.c | 2 +- fs/nilfs2/direct.c | 2 +- fs/nilfs2/direct.h | 2 +- fs/nilfs2/file.c | 3 +-- fs/nilfs2/gcinode.c | 5 ++--- fs/nilfs2/ifile.c | 4 ++-- fs/nilfs2/ifile.h | 4 ++-- fs/nilfs2/inode.c | 2 +- fs/nilfs2/ioctl.c | 2 +- fs/nilfs2/mdt.c | 2 +- fs/nilfs2/mdt.h | 2 +- fs/nilfs2/namei.c | 3 +-- fs/nilfs2/nilfs.h | 3 +-- fs/nilfs2/page.c | 3 +-- fs/nilfs2/page.h | 3 +-- fs/nilfs2/recovery.c | 2 +- fs/nilfs2/segbuf.c | 2 +- fs/nilfs2/segbuf.h | 2 +- fs/nilfs2/segment.c | 2 +- fs/nilfs2/segment.h | 2 +- fs/nilfs2/sufile.c | 4 ++-- fs/nilfs2/sufile.h | 2 +- fs/nilfs2/super.c | 2 +- fs/nilfs2/the_nilfs.c | 2 +- fs/nilfs2/the_nilfs.h | 2 +- include/linux/nilfs2_fs.h | 3 +-- 38 files changed, 46 insertions(+), 56 deletions(-) (limited to 'fs') diff --git a/fs/nilfs2/alloc.c b/fs/nilfs2/alloc.c index eaa0c6af80c3..82362a5d81ed 100644 --- a/fs/nilfs2/alloc.c +++ b/fs/nilfs2/alloc.c @@ -13,9 +13,8 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Original code was written by Koji Sato . - * Two allocators were unified by Ryusuke Konishi , - * Amagai Yoshiji . + * Originally written by Koji Sato. + * Two allocators were unified by Ryusuke Konishi and Amagai Yoshiji. */ #include diff --git a/fs/nilfs2/alloc.h b/fs/nilfs2/alloc.h index 62982eea63c0..2bd567d98bc6 100644 --- a/fs/nilfs2/alloc.h +++ b/fs/nilfs2/alloc.h @@ -13,9 +13,8 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Original code was written by Koji Sato . - * Two allocators were unified by Ryusuke Konishi , - * Amagai Yoshiji . + * Originally written by Koji Sato. + * Two allocators were unified by Ryusuke Konishi and Amagai Yoshiji. */ #ifndef _NILFS_ALLOC_H diff --git a/fs/nilfs2/bmap.c b/fs/nilfs2/bmap.c index fcd79e611c87..a5eab798a7a3 100644 --- a/fs/nilfs2/bmap.c +++ b/fs/nilfs2/bmap.c @@ -13,7 +13,7 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Koji Sato . + * Written by Koji Sato. */ #include diff --git a/fs/nilfs2/bmap.h b/fs/nilfs2/bmap.h index baa6d404f369..c14f822682da 100644 --- a/fs/nilfs2/bmap.h +++ b/fs/nilfs2/bmap.h @@ -13,7 +13,7 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Koji Sato . + * Written by Koji Sato. */ #ifndef _NILFS_BMAP_H diff --git a/fs/nilfs2/btnode.c b/fs/nilfs2/btnode.c index 55241effa3c0..0576033699bc 100644 --- a/fs/nilfs2/btnode.c +++ b/fs/nilfs2/btnode.c @@ -13,9 +13,8 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * This file was originally written by Seiji Kihara - * and fully revised by Ryusuke Konishi for - * stabilization and simplification. + * Originally written by Seiji Kihara. + * Fully revised by Ryusuke Konishi for stabilization and simplification. * */ diff --git a/fs/nilfs2/btnode.h b/fs/nilfs2/btnode.h index 0727096a8c53..2cc1b80e18f7 100644 --- a/fs/nilfs2/btnode.h +++ b/fs/nilfs2/btnode.h @@ -13,8 +13,8 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Seiji Kihara - * Revised by Ryusuke Konishi + * Written by Seiji Kihara. + * Revised by Ryusuke Konishi. */ #ifndef _NILFS_BTNODE_H diff --git a/fs/nilfs2/btree.c b/fs/nilfs2/btree.c index cc748ff3b876..8fc73d0a923c 100644 --- a/fs/nilfs2/btree.c +++ b/fs/nilfs2/btree.c @@ -13,7 +13,7 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Koji Sato . + * Written by Koji Sato. */ #include diff --git a/fs/nilfs2/btree.h b/fs/nilfs2/btree.h index 9497732b3b1d..df1a25faa83b 100644 --- a/fs/nilfs2/btree.h +++ b/fs/nilfs2/btree.h @@ -13,7 +13,7 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Koji Sato . + * Written by Koji Sato. */ #ifndef _NILFS_BTREE_H diff --git a/fs/nilfs2/cpfile.c b/fs/nilfs2/cpfile.c index 15016eca4dc4..d192b48df9fb 100644 --- a/fs/nilfs2/cpfile.c +++ b/fs/nilfs2/cpfile.c @@ -13,7 +13,7 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Koji Sato . + * Written by Koji Sato. */ #include diff --git a/fs/nilfs2/cpfile.h b/fs/nilfs2/cpfile.h index 3dcaceb66be6..5bdb8262928b 100644 --- a/fs/nilfs2/cpfile.h +++ b/fs/nilfs2/cpfile.h @@ -13,7 +13,7 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Koji Sato . + * Written by Koji Sato. */ #ifndef _NILFS_CPFILE_H diff --git a/fs/nilfs2/dat.c b/fs/nilfs2/dat.c index 7427c114aa11..e92257bc42ee 100644 --- a/fs/nilfs2/dat.c +++ b/fs/nilfs2/dat.c @@ -13,7 +13,7 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Koji Sato . + * Written by Koji Sato. */ #include diff --git a/fs/nilfs2/dat.h b/fs/nilfs2/dat.h index edd0586220d4..c7035b5b1aed 100644 --- a/fs/nilfs2/dat.h +++ b/fs/nilfs2/dat.h @@ -13,7 +13,7 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Koji Sato . + * Written by Koji Sato. */ #ifndef _NILFS_DAT_H diff --git a/fs/nilfs2/dir.c b/fs/nilfs2/dir.c index 955070bd20a5..5756dda3083f 100644 --- a/fs/nilfs2/dir.c +++ b/fs/nilfs2/dir.c @@ -13,7 +13,7 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Modified for NILFS by Amagai Yoshiji + * Modified for NILFS by Amagai Yoshiji. */ /* * linux/fs/ext2/dir.c diff --git a/fs/nilfs2/direct.c b/fs/nilfs2/direct.c index 24700229d6d4..22058d0b36e9 100644 --- a/fs/nilfs2/direct.c +++ b/fs/nilfs2/direct.c @@ -13,7 +13,7 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Koji Sato . + * Written by Koji Sato. */ #include diff --git a/fs/nilfs2/direct.h b/fs/nilfs2/direct.h index e4eb1b7e407c..3015a6e78724 100644 --- a/fs/nilfs2/direct.h +++ b/fs/nilfs2/direct.h @@ -13,7 +13,7 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Koji Sato . + * Written by Koji Sato. */ #ifndef _NILFS_DIRECT_H diff --git a/fs/nilfs2/file.c b/fs/nilfs2/file.c index 49243ca6e3e9..547381f3ce13 100644 --- a/fs/nilfs2/file.c +++ b/fs/nilfs2/file.c @@ -13,8 +13,7 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Amagai Yoshiji , - * Ryusuke Konishi + * Written by Amagai Yoshiji and Ryusuke Konishi. */ #include diff --git a/fs/nilfs2/gcinode.c b/fs/nilfs2/gcinode.c index b0321886ce33..693aded72498 100644 --- a/fs/nilfs2/gcinode.c +++ b/fs/nilfs2/gcinode.c @@ -13,9 +13,8 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Seiji Kihara , Amagai Yoshiji , - * and Ryusuke Konishi . - * Revised by Ryusuke Konishi . + * Written by Seiji Kihara, Amagai Yoshiji, and Ryusuke Konishi. + * Revised by Ryusuke Konishi. * */ /* diff --git a/fs/nilfs2/ifile.c b/fs/nilfs2/ifile.c index 06f9b5aa6175..d9048f6a7780 100644 --- a/fs/nilfs2/ifile.c +++ b/fs/nilfs2/ifile.c @@ -13,8 +13,8 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Amagai Yoshiji . - * Revised by Ryusuke Konishi . + * Written by Amagai Yoshiji. + * Revised by Ryusuke Konishi. * */ diff --git a/fs/nilfs2/ifile.h b/fs/nilfs2/ifile.h index c61637e444f1..1b7d7afcb35e 100644 --- a/fs/nilfs2/ifile.h +++ b/fs/nilfs2/ifile.h @@ -13,8 +13,8 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Amagai Yoshiji - * Revised by Ryusuke Konishi + * Written by Amagai Yoshiji. + * Revised by Ryusuke Konishi. * */ diff --git a/fs/nilfs2/inode.c b/fs/nilfs2/inode.c index 1bdb17bb01d3..90061151f0c2 100644 --- a/fs/nilfs2/inode.c +++ b/fs/nilfs2/inode.c @@ -13,7 +13,7 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Ryusuke Konishi + * Written by Ryusuke Konishi. * */ diff --git a/fs/nilfs2/ioctl.c b/fs/nilfs2/ioctl.c index 4db0e5ff6c67..e86599801d8b 100644 --- a/fs/nilfs2/ioctl.c +++ b/fs/nilfs2/ioctl.c @@ -13,7 +13,7 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Koji Sato . + * Written by Koji Sato. */ #include diff --git a/fs/nilfs2/mdt.c b/fs/nilfs2/mdt.c index 0ab93bdfa261..36fa9412f850 100644 --- a/fs/nilfs2/mdt.c +++ b/fs/nilfs2/mdt.c @@ -13,7 +13,7 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Ryusuke Konishi + * Written by Ryusuke Konishi. */ #include diff --git a/fs/nilfs2/mdt.h b/fs/nilfs2/mdt.h index 5fc07ecf32a3..54e3b7fd783c 100644 --- a/fs/nilfs2/mdt.h +++ b/fs/nilfs2/mdt.h @@ -13,7 +13,7 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Ryusuke Konishi + * Written by Ryusuke Konishi. */ #ifndef _NILFS_MDT_H diff --git a/fs/nilfs2/namei.c b/fs/nilfs2/namei.c index db0a4bcf6015..8234af99d7d7 100644 --- a/fs/nilfs2/namei.c +++ b/fs/nilfs2/namei.c @@ -13,8 +13,7 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Modified for NILFS by Amagai Yoshiji , - * Ryusuke Konishi + * Modified for NILFS by Amagai Yoshiji and Ryusuke Konishi. */ /* * linux/fs/ext2/namei.c diff --git a/fs/nilfs2/nilfs.h b/fs/nilfs2/nilfs.h index e2089a18b24b..fa179d4e5dba 100644 --- a/fs/nilfs2/nilfs.h +++ b/fs/nilfs2/nilfs.h @@ -13,8 +13,7 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Koji Sato - * Ryusuke Konishi + * Written by Koji Sato and Ryusuke Konishi. */ #ifndef _NILFS_H diff --git a/fs/nilfs2/page.c b/fs/nilfs2/page.c index b10057eb4d8e..19687139f197 100644 --- a/fs/nilfs2/page.c +++ b/fs/nilfs2/page.c @@ -13,8 +13,7 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Ryusuke Konishi , - * Seiji Kihara . + * Written by Ryusuke Konishi and Seiji Kihara. */ #include diff --git a/fs/nilfs2/page.h b/fs/nilfs2/page.h index 6df10e127678..041f2dc5e634 100644 --- a/fs/nilfs2/page.h +++ b/fs/nilfs2/page.h @@ -13,8 +13,7 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Ryusuke Konishi , - * Seiji Kihara . + * Written by Ryusuke Konishi and Seiji Kihara. */ #ifndef _NILFS_PAGE_H diff --git a/fs/nilfs2/recovery.c b/fs/nilfs2/recovery.c index 2ae1dfccf6e8..402a45c2550d 100644 --- a/fs/nilfs2/recovery.c +++ b/fs/nilfs2/recovery.c @@ -13,7 +13,7 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Ryusuke Konishi + * Written by Ryusuke Konishi. */ #include diff --git a/fs/nilfs2/segbuf.c b/fs/nilfs2/segbuf.c index 8aef62d81553..52f6a6c8bab1 100644 --- a/fs/nilfs2/segbuf.c +++ b/fs/nilfs2/segbuf.c @@ -13,7 +13,7 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Ryusuke Konishi + * Written by Ryusuke Konishi. * */ diff --git a/fs/nilfs2/segbuf.h b/fs/nilfs2/segbuf.h index fa4ed03744bb..fc3a0fe3119c 100644 --- a/fs/nilfs2/segbuf.h +++ b/fs/nilfs2/segbuf.h @@ -13,7 +13,7 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Ryusuke Konishi + * Written by Ryusuke Konishi. * */ #ifndef _NILFS_SEGBUF_H diff --git a/fs/nilfs2/segment.c b/fs/nilfs2/segment.c index cafb61c32b79..ddbfb09527cd 100644 --- a/fs/nilfs2/segment.c +++ b/fs/nilfs2/segment.c @@ -13,7 +13,7 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Ryusuke Konishi + * Written by Ryusuke Konishi. * */ diff --git a/fs/nilfs2/segment.h b/fs/nilfs2/segment.h index a118eabd3ed6..6cb12dbee7c1 100644 --- a/fs/nilfs2/segment.h +++ b/fs/nilfs2/segment.h @@ -13,7 +13,7 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Ryusuke Konishi + * Written by Ryusuke Konishi. * */ #ifndef _NILFS_SEGMENT_H diff --git a/fs/nilfs2/sufile.c b/fs/nilfs2/sufile.c index 1d602482b51e..df439da883bc 100644 --- a/fs/nilfs2/sufile.c +++ b/fs/nilfs2/sufile.c @@ -13,8 +13,8 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Koji Sato . - * Revised by Ryusuke Konishi . + * Written by Koji Sato. + * Revised by Ryusuke Konishi. */ #include diff --git a/fs/nilfs2/sufile.h b/fs/nilfs2/sufile.h index c4cf47e14b70..46e89872294c 100644 --- a/fs/nilfs2/sufile.h +++ b/fs/nilfs2/sufile.h @@ -13,7 +13,7 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Koji Sato . + * Written by Koji Sato. */ #ifndef _NILFS_SUFILE_H diff --git a/fs/nilfs2/super.c b/fs/nilfs2/super.c index d304faa66f49..2ed3e4008e72 100644 --- a/fs/nilfs2/super.c +++ b/fs/nilfs2/super.c @@ -13,7 +13,7 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Ryusuke Konishi + * Written by Ryusuke Konishi. */ /* * linux/fs/ext2/super.c diff --git a/fs/nilfs2/the_nilfs.c b/fs/nilfs2/the_nilfs.c index 8e270128ae86..ba4b8189c342 100644 --- a/fs/nilfs2/the_nilfs.c +++ b/fs/nilfs2/the_nilfs.c @@ -13,7 +13,7 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Ryusuke Konishi + * Written by Ryusuke Konishi. * */ diff --git a/fs/nilfs2/the_nilfs.h b/fs/nilfs2/the_nilfs.h index 426e4acc772b..6a262933fed9 100644 --- a/fs/nilfs2/the_nilfs.h +++ b/fs/nilfs2/the_nilfs.h @@ -13,7 +13,7 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * - * Written by Ryusuke Konishi + * Written by Ryusuke Konishi. * */ diff --git a/include/linux/nilfs2_fs.h b/include/linux/nilfs2_fs.h index 0d7f8141382a..823d63d61081 100644 --- a/include/linux/nilfs2_fs.h +++ b/include/linux/nilfs2_fs.h @@ -13,8 +13,7 @@ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU Lesser General Public License for more details. * - * Written by Koji Sato - * Ryusuke Konishi + * Written by Koji Sato and Ryusuke Konishi. */ /* * linux/include/linux/ext2_fs.h -- cgit v1.3-8-gc7d7 From 24e20ead2fb2e1b67a76a04550d29c9027f707eb Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Mon, 23 May 2016 16:23:17 -0700 Subject: nilfs2: get rid of nilfs_mdt_mark_block_dirty() nilfs_mdt_mark_block_dirty() can be replaced with primary functions like nilfs_mdt_get_block() and mark_buffer_dirty(), and it's used only by nilfs_ioctl_mark_blocks_dirty(). This gets rid of the function to simplify the interface of metadata file. Link: http://lkml.kernel.org/r/1461935747-10380-8-git-send-email-konishi.ryusuke@lab.ntt.co.jp Signed-off-by: Ryusuke Konishi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nilfs2/ioctl.c | 11 ++++++++--- fs/nilfs2/mdt.c | 28 ---------------------------- fs/nilfs2/mdt.h | 1 - 3 files changed, 8 insertions(+), 32 deletions(-) (limited to 'fs') diff --git a/fs/nilfs2/ioctl.c b/fs/nilfs2/ioctl.c index e86599801d8b..358b57e2cdf9 100644 --- a/fs/nilfs2/ioctl.c +++ b/fs/nilfs2/ioctl.c @@ -779,6 +779,7 @@ static int nilfs_ioctl_mark_blocks_dirty(struct the_nilfs *nilfs, size_t nmembs = argv->v_nmembs; struct nilfs_bmap *bmap = NILFS_I(nilfs->ns_dat)->i_bmap; struct nilfs_bdesc *bdescs = buf; + struct buffer_head *bh; int ret, i; for (i = 0; i < nmembs; i++) { @@ -796,12 +797,16 @@ static int nilfs_ioctl_mark_blocks_dirty(struct the_nilfs *nilfs, /* skip dead block */ continue; if (bdescs[i].bd_level == 0) { - ret = nilfs_mdt_mark_block_dirty(nilfs->ns_dat, - bdescs[i].bd_offset); - if (ret < 0) { + ret = nilfs_mdt_get_block(nilfs->ns_dat, + bdescs[i].bd_offset, + false, NULL, &bh); + if (unlikely(ret)) { WARN_ON(ret == -ENOENT); return ret; } + mark_buffer_dirty(bh); + nilfs_mdt_mark_dirty(nilfs->ns_dat); + put_bh(bh); } else { ret = nilfs_bmap_mark(bmap, bdescs[i].bd_offset, bdescs[i].bd_level); diff --git a/fs/nilfs2/mdt.c b/fs/nilfs2/mdt.c index 36fa9412f850..c8fe58406a07 100644 --- a/fs/nilfs2/mdt.c +++ b/fs/nilfs2/mdt.c @@ -389,34 +389,6 @@ int nilfs_mdt_forget_block(struct inode *inode, unsigned long block) return ret; } -/** - * nilfs_mdt_mark_block_dirty - mark a block on the meta data file dirty. - * @inode: inode of the meta data file - * @block: block offset - * - * Return Value: On success, it returns 0. On error, the following negative - * error code is returned. - * - * %-ENOMEM - Insufficient memory available. - * - * %-EIO - I/O error - * - * %-ENOENT - the specified block does not exist (hole block) - */ -int nilfs_mdt_mark_block_dirty(struct inode *inode, unsigned long block) -{ - struct buffer_head *bh; - int err; - - err = nilfs_mdt_read_block(inode, block, 0, &bh); - if (unlikely(err)) - return err; - mark_buffer_dirty(bh); - nilfs_mdt_mark_dirty(inode); - brelse(bh); - return 0; -} - int nilfs_mdt_fetch_dirty(struct inode *inode) { struct nilfs_inode_info *ii = NILFS_I(inode); diff --git a/fs/nilfs2/mdt.h b/fs/nilfs2/mdt.h index 54e3b7fd783c..dc88907508cd 100644 --- a/fs/nilfs2/mdt.h +++ b/fs/nilfs2/mdt.h @@ -79,7 +79,6 @@ int nilfs_mdt_find_block(struct inode *inode, unsigned long start, struct buffer_head **out_bh); int nilfs_mdt_delete_block(struct inode *, unsigned long); int nilfs_mdt_forget_block(struct inode *, unsigned long); -int nilfs_mdt_mark_block_dirty(struct inode *, unsigned long); int nilfs_mdt_fetch_dirty(struct inode *); int nilfs_mdt_init(struct inode *inode, gfp_t gfp_mask, size_t objsz); -- cgit v1.3-8-gc7d7 From 2d19961d83524c69e44f179b43aee2daf0906eda Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Mon, 23 May 2016 16:23:20 -0700 Subject: nilfs2: move cleanup code of metadata file from inode routines Refactor nilfs_clear_inode() and nilfs_i_callback() so that cleanup code or resource deallocation related to metadata file will be moved out to mdt.c. Link: http://lkml.kernel.org/r/1461935747-10380-9-git-send-email-konishi.ryusuke@lab.ntt.co.jp Signed-off-by: Ryusuke Konishi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nilfs2/inode.c | 5 ++--- fs/nilfs2/mdt.c | 25 +++++++++++++++++++++++++ fs/nilfs2/mdt.h | 8 ++++++++ fs/nilfs2/super.c | 8 +++----- 4 files changed, 38 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/nilfs2/inode.c b/fs/nilfs2/inode.c index 90061151f0c2..bbebd5ece026 100644 --- a/fs/nilfs2/inode.c +++ b/fs/nilfs2/inode.c @@ -755,7 +755,6 @@ void nilfs_truncate(struct inode *inode) static void nilfs_clear_inode(struct inode *inode) { struct nilfs_inode_info *ii = NILFS_I(inode); - struct nilfs_mdt_info *mdi = NILFS_MDT(inode); /* * Free resources allocated in nilfs_read_inode(), here. @@ -764,8 +763,8 @@ static void nilfs_clear_inode(struct inode *inode) brelse(ii->i_bh); ii->i_bh = NULL; - if (mdi && mdi->mi_palloc_cache) - nilfs_palloc_destroy_cache(inode); + if (nilfs_is_metadata_file_inode(inode)) + nilfs_mdt_clear(inode); if (test_bit(NILFS_I_BMAP, &ii->i_state)) nilfs_bmap_clear(ii->i_bmap); diff --git a/fs/nilfs2/mdt.c b/fs/nilfs2/mdt.c index c8fe58406a07..8a2f8b240f25 100644 --- a/fs/nilfs2/mdt.c +++ b/fs/nilfs2/mdt.c @@ -28,6 +28,7 @@ #include "segment.h" #include "page.h" #include "mdt.h" +#include "alloc.h" /* nilfs_palloc_destroy_cache() */ #include @@ -465,6 +466,30 @@ int nilfs_mdt_init(struct inode *inode, gfp_t gfp_mask, size_t objsz) return 0; } +/** + * nilfs_mdt_clear - do cleanup for the metadata file + * @inode: inode of the metadata file + */ +void nilfs_mdt_clear(struct inode *inode) +{ + struct nilfs_mdt_info *mdi = NILFS_MDT(inode); + + if (mdi->mi_palloc_cache) + nilfs_palloc_destroy_cache(inode); +} + +/** + * nilfs_mdt_destroy - release resources used by the metadata file + * @inode: inode of the metadata file + */ +void nilfs_mdt_destroy(struct inode *inode) +{ + struct nilfs_mdt_info *mdi = NILFS_MDT(inode); + + kfree(mdi->mi_bgl); /* kfree(NULL) is safe */ + kfree(mdi); +} + void nilfs_mdt_set_entry_size(struct inode *inode, unsigned entry_size, unsigned header_size) { diff --git a/fs/nilfs2/mdt.h b/fs/nilfs2/mdt.h index dc88907508cd..ffb876e6efed 100644 --- a/fs/nilfs2/mdt.h +++ b/fs/nilfs2/mdt.h @@ -67,6 +67,11 @@ static inline struct nilfs_mdt_info *NILFS_MDT(const struct inode *inode) return inode->i_private; } +static inline int nilfs_is_metadata_file_inode(const struct inode *inode) +{ + return inode->i_private != NULL; +} + /* Default GFP flags using highmem */ #define NILFS_MDT_GFP (__GFP_RECLAIM | __GFP_IO | __GFP_HIGHMEM) @@ -82,6 +87,9 @@ int nilfs_mdt_forget_block(struct inode *, unsigned long); int nilfs_mdt_fetch_dirty(struct inode *); int nilfs_mdt_init(struct inode *inode, gfp_t gfp_mask, size_t objsz); +void nilfs_mdt_clear(struct inode *inode); +void nilfs_mdt_destroy(struct inode *inode); + void nilfs_mdt_set_entry_size(struct inode *, unsigned, unsigned); int nilfs_mdt_setup_shadow_map(struct inode *inode, diff --git a/fs/nilfs2/super.c b/fs/nilfs2/super.c index 2ed3e4008e72..2cb78b857df6 100644 --- a/fs/nilfs2/super.c +++ b/fs/nilfs2/super.c @@ -169,12 +169,10 @@ struct inode *nilfs_alloc_inode(struct super_block *sb) static void nilfs_i_callback(struct rcu_head *head) { struct inode *inode = container_of(head, struct inode, i_rcu); - struct nilfs_mdt_info *mdi = NILFS_MDT(inode); - if (mdi) { - kfree(mdi->mi_bgl); /* kfree(NULL) is safe */ - kfree(mdi); - } + if (nilfs_is_metadata_file_inode(inode)) + nilfs_mdt_destroy(inode); + kmem_cache_free(nilfs_inode_cachep, NILFS_I(inode)); } -- cgit v1.3-8-gc7d7 From e7a142aaa09fa5db015fd176d6943f888665829f Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Mon, 23 May 2016 16:23:23 -0700 Subject: nilfs2: replace __attribute__((packed)) with __packed This fixes the following checkpatch.pl warning: WARNING: __packed is preferred over __attribute__((packed)) #23: FILE: export.h:23: +} __attribute__ ((packed)); Link: http://lkml.kernel.org/r/1461935747-10380-10-git-send-email-konishi.ryusuke@lab.ntt.co.jp Signed-off-by: Ryusuke Konishi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nilfs2/export.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nilfs2/export.h b/fs/nilfs2/export.h index 19ccbf9522ab..00107fdb9343 100644 --- a/fs/nilfs2/export.h +++ b/fs/nilfs2/export.h @@ -20,6 +20,6 @@ struct nilfs_fid { u32 parent_gen; u64 parent_ino; -} __attribute__ ((packed)); +} __packed; #endif -- cgit v1.3-8-gc7d7 From 4ad364ca1ce3f4f96e04aaeb7974de4380eb6004 Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Mon, 23 May 2016 16:23:25 -0700 Subject: nilfs2: add missing line spacing Clean up checkpatch.pl warnings "WARNING: Missing a blank line after declarations" from nilfs2. Link: http://lkml.kernel.org/r/1461935747-10380-11-git-send-email-konishi.ryusuke@lab.ntt.co.jp Signed-off-by: Ryusuke Konishi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nilfs2/cpfile.c | 2 ++ fs/nilfs2/dir.c | 2 ++ fs/nilfs2/ifile.h | 1 + fs/nilfs2/inode.c | 1 + fs/nilfs2/namei.c | 1 + fs/nilfs2/recovery.c | 1 + fs/nilfs2/segment.c | 1 + fs/nilfs2/sufile.c | 2 ++ fs/nilfs2/super.c | 1 + fs/nilfs2/the_nilfs.h | 2 ++ 10 files changed, 14 insertions(+) (limited to 'fs') diff --git a/fs/nilfs2/cpfile.c b/fs/nilfs2/cpfile.c index d192b48df9fb..16f884bd857c 100644 --- a/fs/nilfs2/cpfile.c +++ b/fs/nilfs2/cpfile.c @@ -37,6 +37,7 @@ static unsigned long nilfs_cpfile_get_blkoff(const struct inode *cpfile, __u64 cno) { __u64 tcno = cno + NILFS_MDT(cpfile)->mi_first_entry_offset - 1; + do_div(tcno, nilfs_cpfile_checkpoints_per_block(cpfile)); return (unsigned long)tcno; } @@ -46,6 +47,7 @@ static unsigned long nilfs_cpfile_get_offset(const struct inode *cpfile, __u64 cno) { __u64 tcno = cno + NILFS_MDT(cpfile)->mi_first_entry_offset - 1; + return do_div(tcno, nilfs_cpfile_checkpoints_per_block(cpfile)); } diff --git a/fs/nilfs2/dir.c b/fs/nilfs2/dir.c index 5756dda3083f..9ca3174f8196 100644 --- a/fs/nilfs2/dir.c +++ b/fs/nilfs2/dir.c @@ -74,6 +74,7 @@ static unsigned nilfs_last_byte(struct inode *inode, unsigned long page_nr) static int nilfs_prepare_chunk(struct page *page, unsigned from, unsigned to) { loff_t pos = page_offset(page) + from; + return __block_write_begin(page, pos, to - from, nilfs_get_block); } @@ -336,6 +337,7 @@ nilfs_find_entry(struct inode *dir, const struct qstr *qstr, n = start; do { char *kaddr; + page = nilfs_get_page(dir, n); if (!IS_ERR(page)) { kaddr = page_address(page); diff --git a/fs/nilfs2/ifile.h b/fs/nilfs2/ifile.h index 1b7d7afcb35e..23ad2f091e76 100644 --- a/fs/nilfs2/ifile.h +++ b/fs/nilfs2/ifile.h @@ -32,6 +32,7 @@ static inline struct nilfs_inode * nilfs_ifile_map_inode(struct inode *ifile, ino_t ino, struct buffer_head *ibh) { void *kaddr = kmap(ibh->b_page); + return nilfs_palloc_block_get_entry(ifile, ino, ibh, kaddr); } diff --git a/fs/nilfs2/inode.c b/fs/nilfs2/inode.c index bbebd5ece026..aaa129398eab 100644 --- a/fs/nilfs2/inode.c +++ b/fs/nilfs2/inode.c @@ -851,6 +851,7 @@ out_err: int nilfs_permission(struct inode *inode, int mask) { struct nilfs_root *root = NILFS_I(inode)->i_root; + if ((mask & MAY_WRITE) && root && root->cno != NILFS_CPTREE_CURRENT_CNO) return -EROFS; /* snapshot is not writable */ diff --git a/fs/nilfs2/namei.c b/fs/nilfs2/namei.c index 8234af99d7d7..47bd989d88ed 100644 --- a/fs/nilfs2/namei.c +++ b/fs/nilfs2/namei.c @@ -44,6 +44,7 @@ static inline int nilfs_add_nondir(struct dentry *dentry, struct inode *inode) { int err = nilfs_add_link(dentry, inode); + if (!err) { d_instantiate(dentry, inode); unlock_new_inode(inode); diff --git a/fs/nilfs2/recovery.c b/fs/nilfs2/recovery.c index 402a45c2550d..685fa73cecd0 100644 --- a/fs/nilfs2/recovery.c +++ b/fs/nilfs2/recovery.c @@ -522,6 +522,7 @@ static int nilfs_recover_dsync_blocks(struct the_nilfs *nilfs, 0, &page, nilfs_get_block); if (unlikely(err)) { loff_t isize = inode->i_size; + if (pos + blocksize > isize) nilfs_write_failed(inode->i_mapping, pos + blocksize); diff --git a/fs/nilfs2/segment.c b/fs/nilfs2/segment.c index ddbfb09527cd..69e9ea8d53c7 100644 --- a/fs/nilfs2/segment.c +++ b/fs/nilfs2/segment.c @@ -2391,6 +2391,7 @@ static int nilfs_segctor_construct(struct nilfs_sc_info *sci, int mode) static void nilfs_construction_timeout(unsigned long data) { struct task_struct *p = (struct task_struct *)data; + wake_up_process(p); } diff --git a/fs/nilfs2/sufile.c b/fs/nilfs2/sufile.c index df439da883bc..1963595a1580 100644 --- a/fs/nilfs2/sufile.c +++ b/fs/nilfs2/sufile.c @@ -57,6 +57,7 @@ static unsigned long nilfs_sufile_get_blkoff(const struct inode *sufile, __u64 segnum) { __u64 t = segnum + NILFS_MDT(sufile)->mi_first_entry_offset; + do_div(t, nilfs_sufile_segment_usages_per_block(sufile)); return (unsigned long)t; } @@ -65,6 +66,7 @@ static unsigned long nilfs_sufile_get_offset(const struct inode *sufile, __u64 segnum) { __u64 t = segnum + NILFS_MDT(sufile)->mi_first_entry_offset; + return do_div(t, nilfs_sufile_segment_usages_per_block(sufile)); } diff --git a/fs/nilfs2/super.c b/fs/nilfs2/super.c index 2cb78b857df6..92acb005b78f 100644 --- a/fs/nilfs2/super.c +++ b/fs/nilfs2/super.c @@ -743,6 +743,7 @@ static int parse_options(char *options, struct super_block *sb, int is_remount) while ((p = strsep(&options, ",")) != NULL) { int token; + if (!*p) continue; diff --git a/fs/nilfs2/the_nilfs.h b/fs/nilfs2/the_nilfs.h index 6a262933fed9..62bd7b10fe43 100644 --- a/fs/nilfs2/the_nilfs.h +++ b/fs/nilfs2/the_nilfs.h @@ -269,6 +269,7 @@ struct nilfs_root { static inline int nilfs_sb_need_update(struct the_nilfs *nilfs) { u64 t = get_seconds(); + return t < nilfs->ns_sbwtime || t > nilfs->ns_sbwtime + nilfs->ns_sb_update_freq; } @@ -276,6 +277,7 @@ static inline int nilfs_sb_need_update(struct the_nilfs *nilfs) static inline int nilfs_sb_will_flip(struct the_nilfs *nilfs) { int flip_bits = nilfs->ns_sbwcount & 0x0FL; + return (flip_bits != 0x08 && flip_bits != 0x0F); } -- cgit v1.3-8-gc7d7 From facb9ec5e6e376c25a00bb7a64332287c7a1403c Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Mon, 23 May 2016 16:23:28 -0700 Subject: nilfs2: clean trailing semicolons in macros Remove trailing semicolons from macros, as suggested by checkpatch.pl. Link: http://lkml.kernel.org/r/1461935747-10380-12-git-send-email-konishi.ryusuke@lab.ntt.co.jp [konishi.ryusuke@lab.ntt.co.jp: fix style issues] Link: http://lkml.kernel.org/r/20160509.231703.1481729973362188932.konishi.ryusuke@lab.ntt.co.jp Signed-off-by: Ryusuke Konishi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nilfs2/sysfs.c | 4 ++-- fs/nilfs2/sysfs.h | 6 +++--- 2 files changed, 5 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/nilfs2/sysfs.c b/fs/nilfs2/sysfs.c index bbb0dcc35905..3e7d85335adf 100644 --- a/fs/nilfs2/sysfs.c +++ b/fs/nilfs2/sysfs.c @@ -68,7 +68,7 @@ static ssize_t nilfs_##name##_attr_store(struct kobject *kobj, \ static const struct sysfs_ops nilfs_##name##_attr_ops = { \ .show = nilfs_##name##_attr_show, \ .store = nilfs_##name##_attr_store, \ -}; +} #define NILFS_DEV_INT_GROUP_TYPE(name, parent_name) \ static void nilfs_##name##_attr_release(struct kobject *kobj) \ @@ -84,7 +84,7 @@ static struct kobj_type nilfs_##name##_ktype = { \ .default_attrs = nilfs_##name##_attrs, \ .sysfs_ops = &nilfs_##name##_attr_ops, \ .release = nilfs_##name##_attr_release, \ -}; +} #define NILFS_DEV_INT_GROUP_FNS(name, parent_name) \ static int nilfs_sysfs_create_##name##_group(struct the_nilfs *nilfs) \ diff --git a/fs/nilfs2/sysfs.h b/fs/nilfs2/sysfs.h index 677e3a1a8370..648cedf9c06e 100644 --- a/fs/nilfs2/sysfs.h +++ b/fs/nilfs2/sysfs.h @@ -66,7 +66,7 @@ struct nilfs_##name##_attr { \ char *); \ ssize_t (*store)(struct kobject *, struct attribute *, \ const char *, size_t); \ -}; +} NILFS_COMMON_ATTR_STRUCT(feature); @@ -77,7 +77,7 @@ struct nilfs_##name##_attr { \ char *); \ ssize_t (*store)(struct nilfs_##name##_attr *, struct the_nilfs *, \ const char *, size_t); \ -}; +} NILFS_DEV_ATTR_STRUCT(dev); NILFS_DEV_ATTR_STRUCT(segments); @@ -93,7 +93,7 @@ struct nilfs_##name##_attr { \ char *); \ ssize_t (*store)(struct nilfs_##name##_attr *, struct nilfs_root *, \ const char *, size_t); \ -}; +} NILFS_CP_ATTR_STRUCT(snapshot); -- cgit v1.3-8-gc7d7 From 06f4abf6ca9364cdbbfac5b6e0fb1a5bf3468195 Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Mon, 23 May 2016 16:23:31 -0700 Subject: nilfs2: do not emit extra newline on nilfs_warning() and nilfs_error() This updates call sites of nilfs_warning() and nilfs_error() so that they don't add a duplicate newline. These output functions are already designed to add a trailing newline to the message. Link: http://lkml.kernel.org/r/1462886671-3521-2-git-send-email-konishi.ryusuke@lab.ntt.co.jp Signed-off-by: Ryusuke Konishi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nilfs2/alloc.c | 10 +++++----- fs/nilfs2/bmap.c | 2 +- fs/nilfs2/dir.c | 4 ++-- fs/nilfs2/inode.c | 6 +++--- fs/nilfs2/namei.c | 2 +- fs/nilfs2/segment.c | 8 ++++---- 6 files changed, 16 insertions(+), 16 deletions(-) (limited to 'fs') diff --git a/fs/nilfs2/alloc.c b/fs/nilfs2/alloc.c index 82362a5d81ed..bdd5ac522904 100644 --- a/fs/nilfs2/alloc.c +++ b/fs/nilfs2/alloc.c @@ -619,7 +619,7 @@ void nilfs_palloc_commit_free_entry(struct inode *inode, if (!nilfs_clear_bit_atomic(lock, group_offset, bitmap)) nilfs_warning(inode->i_sb, __func__, - "entry number %llu already freed: ino=%lu\n", + "entry number %llu already freed: ino=%lu", (unsigned long long)req->pr_entry_nr, (unsigned long)inode->i_ino); else @@ -660,7 +660,7 @@ void nilfs_palloc_abort_alloc_entry(struct inode *inode, if (!nilfs_clear_bit_atomic(lock, group_offset, bitmap)) nilfs_warning(inode->i_sb, __func__, - "entry number %llu already freed: ino=%lu\n", + "entry number %llu already freed: ino=%lu", (unsigned long long)req->pr_entry_nr, (unsigned long)inode->i_ino); else @@ -769,7 +769,7 @@ int nilfs_palloc_freev(struct inode *inode, __u64 *entry_nrs, size_t nitems) if (!nilfs_clear_bit_atomic(lock, group_offset, bitmap)) { nilfs_warning(inode->i_sb, __func__, - "entry number %llu already freed: ino=%lu\n", + "entry number %llu already freed: ino=%lu", (unsigned long long)entry_nrs[j], (unsigned long)inode->i_ino); } else { @@ -814,7 +814,7 @@ int nilfs_palloc_freev(struct inode *inode, __u64 *entry_nrs, size_t nitems) last_nrs[k]); if (ret && ret != -ENOENT) { nilfs_warning(inode->i_sb, __func__, - "failed to delete block of entry %llu: ino=%lu, err=%d\n", + "failed to delete block of entry %llu: ino=%lu, err=%d", (unsigned long long)last_nrs[k], (unsigned long)inode->i_ino, ret); } @@ -833,7 +833,7 @@ int nilfs_palloc_freev(struct inode *inode, __u64 *entry_nrs, size_t nitems) ret = nilfs_palloc_delete_bitmap_block(inode, group); if (ret && ret != -ENOENT) { nilfs_warning(inode->i_sb, __func__, - "failed to delete bitmap block of group %lu: ino=%lu, err=%d\n", + "failed to delete bitmap block of group %lu: ino=%lu, err=%d", group, (unsigned long)inode->i_ino, ret); } diff --git a/fs/nilfs2/bmap.c b/fs/nilfs2/bmap.c index a5eab798a7a3..4976fe3be4a5 100644 --- a/fs/nilfs2/bmap.c +++ b/fs/nilfs2/bmap.c @@ -42,7 +42,7 @@ static int nilfs_bmap_convert_error(struct nilfs_bmap *bmap, if (err == -EINVAL) { nilfs_error(inode->i_sb, fname, - "broken bmap (inode number=%lu)\n", inode->i_ino); + "broken bmap (inode number=%lu)", inode->i_ino); err = -EIO; } return err; diff --git a/fs/nilfs2/dir.c b/fs/nilfs2/dir.c index 9ca3174f8196..bbcc03de1e74 100644 --- a/fs/nilfs2/dir.c +++ b/fs/nilfs2/dir.c @@ -628,8 +628,8 @@ int nilfs_empty_dir(struct inode *inode) while ((char *)de <= kaddr) { if (de->rec_len == 0) { nilfs_error(inode->i_sb, __func__, - "zero-length directory entry " - "(kaddr=%p, de=%p)\n", kaddr, de); + "zero-length directory entry (kaddr=%p, de=%p)", + kaddr, de); goto not_empty; } if (de->inode != 0) { diff --git a/fs/nilfs2/inode.c b/fs/nilfs2/inode.c index aaa129398eab..bbb47e8bde3e 100644 --- a/fs/nilfs2/inode.c +++ b/fs/nilfs2/inode.c @@ -921,7 +921,7 @@ int nilfs_set_file_dirty(struct inode *inode, unsigned nr_dirty) /* This will happen when somebody is freeing this inode. */ nilfs_warning(inode->i_sb, __func__, - "cannot get inode (ino=%lu)\n", + "cannot get inode (ino=%lu)", inode->i_ino); spin_unlock(&nilfs->ns_inode_lock); return -EINVAL; /* NILFS_I_DIRTY may remain for @@ -942,7 +942,7 @@ int __nilfs_mark_inode_dirty(struct inode *inode, int flags) err = nilfs_load_inode_block(inode, &ibh); if (unlikely(err)) { nilfs_warning(inode->i_sb, __func__, - "failed to reget inode block.\n"); + "failed to reget inode block."); return err; } nilfs_update_inode(inode, ibh, flags); @@ -969,7 +969,7 @@ void nilfs_dirty_inode(struct inode *inode, int flags) if (is_bad_inode(inode)) { nilfs_warning(inode->i_sb, __func__, - "tried to mark bad_inode dirty. ignored.\n"); + "tried to mark bad_inode dirty. ignored."); dump_stack(); return; } diff --git a/fs/nilfs2/namei.c b/fs/nilfs2/namei.c index 47bd989d88ed..8f8070cffa58 100644 --- a/fs/nilfs2/namei.c +++ b/fs/nilfs2/namei.c @@ -284,7 +284,7 @@ static int nilfs_do_unlink(struct inode *dir, struct dentry *dentry) if (!inode->i_nlink) { nilfs_warning(inode->i_sb, __func__, - "deleting nonexistent file (%lu), %d\n", + "deleting nonexistent file (%lu), %d", inode->i_ino, inode->i_nlink); set_nlink(inode, 1); } diff --git a/fs/nilfs2/segment.c b/fs/nilfs2/segment.c index 69e9ea8d53c7..a6ef1eb15edb 100644 --- a/fs/nilfs2/segment.c +++ b/fs/nilfs2/segment.c @@ -1938,7 +1938,7 @@ static int nilfs_segctor_collect_dirty_files(struct nilfs_sc_info *sci, ifile, ii->vfs_inode.i_ino, &ibh); if (unlikely(err)) { nilfs_warning(sci->sc_super, __func__, - "failed to get inode block.\n"); + "failed to get inode block."); return err; } mark_buffer_dirty(ibh); @@ -2724,13 +2724,13 @@ static void nilfs_segctor_destroy(struct nilfs_sc_info *sci) if (!list_empty(&sci->sc_dirty_files)) { nilfs_warning(sci->sc_super, __func__, - "dirty file(s) after the final construction\n"); + "dirty file(s) after the final construction"); nilfs_dispose_list(nilfs, &sci->sc_dirty_files, 1); } if (!list_empty(&sci->sc_iput_queue)) { nilfs_warning(sci->sc_super, __func__, - "iput queue is not empty\n"); + "iput queue is not empty"); nilfs_dispose_list(nilfs, &sci->sc_iput_queue, 1); } @@ -2807,7 +2807,7 @@ void nilfs_detach_log_writer(struct super_block *sb) if (!list_empty(&nilfs->ns_dirty_files)) { list_splice_init(&nilfs->ns_dirty_files, &garbage_list); nilfs_warning(sb, __func__, - "Hit dirty file after stopped log writer\n"); + "Hit dirty file after stopped log writer"); } spin_unlock(&nilfs->ns_inode_lock); up_write(&nilfs->ns_segctor_sem); -- cgit v1.3-8-gc7d7 From c9cb9b5c85265acf3e40118727630fd3d57a0bdb Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Mon, 23 May 2016 16:23:34 -0700 Subject: nilfs2: remove space before semicolon This fixes the checkpatch.pl warning "WARNING: space prohibited before semicolon" at nilfs_store_magic_and_option(). Link: http://lkml.kernel.org/r/1462886671-3521-3-git-send-email-konishi.ryusuke@lab.ntt.co.jp Signed-off-by: Ryusuke Konishi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nilfs2/super.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nilfs2/super.c b/fs/nilfs2/super.c index 92acb005b78f..0271784f867b 100644 --- a/fs/nilfs2/super.c +++ b/fs/nilfs2/super.c @@ -886,7 +886,7 @@ int nilfs_store_magic_and_option(struct super_block *sb, nilfs->ns_interval = le32_to_cpu(sbp->s_c_interval); nilfs->ns_watermark = le32_to_cpu(sbp->s_c_block_max); - return !parse_options(data, sb, 0) ? -EINVAL : 0 ; + return !parse_options(data, sb, 0) ? -EINVAL : 0; } int nilfs_check_feature_compatibility(struct super_block *sb, -- cgit v1.3-8-gc7d7 From 7592ecde65f908f082cfd3440888fd6ae99f4cbb Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Mon, 23 May 2016 16:23:36 -0700 Subject: nilfs2: fix code indent coding style issue This fixes checkpatch.pl warning "WARNING: suspect code indent for conditional statements". Link: http://lkml.kernel.org/r/1462886671-3521-4-git-send-email-konishi.ryusuke@lab.ntt.co.jp Signed-off-by: Ryusuke Konishi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nilfs2/super.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nilfs2/super.c b/fs/nilfs2/super.c index 0271784f867b..666107a18a22 100644 --- a/fs/nilfs2/super.c +++ b/fs/nilfs2/super.c @@ -273,7 +273,7 @@ struct nilfs_super_block **nilfs_prepare_super(struct super_block *sb, } } else if (sbp[1] && sbp[1]->s_magic != cpu_to_le16(NILFS_SUPER_MAGIC)) { - memcpy(sbp[1], sbp[0], nilfs->ns_sbsize); + memcpy(sbp[1], sbp[0], nilfs->ns_sbsize); } if (flip && sbp[1]) -- cgit v1.3-8-gc7d7 From 0c6c44cb9f93f7c0ad803b41ae7c0b08cf6942e2 Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Mon, 23 May 2016 16:23:39 -0700 Subject: nilfs2: avoid bare use of 'unsigned' This fixes checkpatch.pl warning "WARNING: Prefer 'unsigned int' to bare use of 'unsigned'". Link: http://lkml.kernel.org/r/1462886671-3521-5-git-send-email-konishi.ryusuke@lab.ntt.co.jp Signed-off-by: Ryusuke Konishi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nilfs2/alloc.c | 8 ++++---- fs/nilfs2/alloc.h | 2 +- fs/nilfs2/bmap.c | 2 +- fs/nilfs2/bmap.h | 4 ++-- fs/nilfs2/btree.c | 3 ++- fs/nilfs2/cpfile.c | 8 +++++--- fs/nilfs2/cpfile.h | 4 ++-- fs/nilfs2/dat.c | 2 +- fs/nilfs2/dat.h | 2 +- fs/nilfs2/dir.c | 47 ++++++++++++++++++++++++----------------------- fs/nilfs2/direct.c | 5 +++-- fs/nilfs2/inode.c | 14 +++++++------- fs/nilfs2/mdt.c | 4 ++-- fs/nilfs2/mdt.h | 6 +++--- fs/nilfs2/namei.c | 2 +- fs/nilfs2/nilfs.h | 2 +- fs/nilfs2/page.c | 8 ++++---- fs/nilfs2/page.h | 3 ++- fs/nilfs2/recovery.c | 4 ++-- fs/nilfs2/segbuf.c | 4 ++-- fs/nilfs2/segbuf.h | 5 +++-- fs/nilfs2/segment.c | 18 +++++++++--------- fs/nilfs2/segment.h | 4 ++-- fs/nilfs2/sysfs.c | 6 +++--- fs/nilfs2/the_nilfs.c | 4 ++-- fs/nilfs2/the_nilfs.h | 10 +++++----- include/linux/nilfs2_fs.h | 6 +++--- 27 files changed, 97 insertions(+), 90 deletions(-) (limited to 'fs') diff --git a/fs/nilfs2/alloc.c b/fs/nilfs2/alloc.c index bdd5ac522904..698f582d69af 100644 --- a/fs/nilfs2/alloc.c +++ b/fs/nilfs2/alloc.c @@ -53,7 +53,7 @@ nilfs_palloc_groups_count(const struct inode *inode) * @inode: inode of metadata file using this allocator * @entry_size: size of the persistent object */ -int nilfs_palloc_init_blockgroup(struct inode *inode, unsigned entry_size) +int nilfs_palloc_init_blockgroup(struct inode *inode, unsigned int entry_size) { struct nilfs_mdt_info *mi = NILFS_MDT(inode); @@ -384,7 +384,7 @@ void *nilfs_palloc_block_get_entry(const struct inode *inode, __u64 nr, */ static int nilfs_palloc_find_available_slot(unsigned char *bitmap, unsigned long target, - unsigned bsize, + unsigned int bsize, spinlock_t *lock) { int pos, end = bsize; @@ -735,8 +735,8 @@ int nilfs_palloc_freev(struct inode *inode, __u64 *entry_nrs, size_t nitems) unsigned long group, group_offset; __u64 group_min_nr, last_nrs[8]; const unsigned long epg = nilfs_palloc_entries_per_group(inode); - const unsigned epb = NILFS_MDT(inode)->mi_entries_per_block; - unsigned entry_start, end, pos; + const unsigned int epb = NILFS_MDT(inode)->mi_entries_per_block; + unsigned int entry_start, end, pos; spinlock_t *lock; int i, j, k, ret; u32 nfree; diff --git a/fs/nilfs2/alloc.h b/fs/nilfs2/alloc.h index 2bd567d98bc6..05149e606a78 100644 --- a/fs/nilfs2/alloc.h +++ b/fs/nilfs2/alloc.h @@ -37,7 +37,7 @@ nilfs_palloc_entries_per_group(const struct inode *inode) return 1UL << (inode->i_blkbits + 3 /* log2(8 = CHAR_BITS) */); } -int nilfs_palloc_init_blockgroup(struct inode *, unsigned); +int nilfs_palloc_init_blockgroup(struct inode *, unsigned int); int nilfs_palloc_get_entry_block(struct inode *, __u64, int, struct buffer_head **); void *nilfs_palloc_block_get_entry(const struct inode *, __u64, diff --git a/fs/nilfs2/bmap.c b/fs/nilfs2/bmap.c index 4976fe3be4a5..f2a7877e0c8c 100644 --- a/fs/nilfs2/bmap.c +++ b/fs/nilfs2/bmap.c @@ -93,7 +93,7 @@ int nilfs_bmap_lookup_at_level(struct nilfs_bmap *bmap, __u64 key, int level, } int nilfs_bmap_lookup_contig(struct nilfs_bmap *bmap, __u64 key, __u64 *ptrp, - unsigned maxblocks) + unsigned int maxblocks) { int ret; diff --git a/fs/nilfs2/bmap.h b/fs/nilfs2/bmap.h index c14f822682da..a6852807b22c 100644 --- a/fs/nilfs2/bmap.h +++ b/fs/nilfs2/bmap.h @@ -57,7 +57,7 @@ struct nilfs_bmap_stats { struct nilfs_bmap_operations { int (*bop_lookup)(const struct nilfs_bmap *, __u64, int, __u64 *); int (*bop_lookup_contig)(const struct nilfs_bmap *, __u64, __u64 *, - unsigned); + unsigned int); int (*bop_insert)(struct nilfs_bmap *, __u64, __u64); int (*bop_delete)(struct nilfs_bmap *, __u64); void (*bop_clear)(struct nilfs_bmap *); @@ -150,7 +150,7 @@ struct nilfs_bmap_store { int nilfs_bmap_test_and_clear_dirty(struct nilfs_bmap *); int nilfs_bmap_read(struct nilfs_bmap *, struct nilfs_inode *); void nilfs_bmap_write(struct nilfs_bmap *, struct nilfs_inode *); -int nilfs_bmap_lookup_contig(struct nilfs_bmap *, __u64, __u64 *, unsigned); +int nilfs_bmap_lookup_contig(struct nilfs_bmap *, __u64, __u64 *, unsigned int); int nilfs_bmap_insert(struct nilfs_bmap *bmap, __u64 key, unsigned long rec); int nilfs_bmap_delete(struct nilfs_bmap *bmap, __u64 key); int nilfs_bmap_seek_key(struct nilfs_bmap *bmap, __u64 start, __u64 *keyp); diff --git a/fs/nilfs2/btree.c b/fs/nilfs2/btree.c index 8fc73d0a923c..57ec6af28b49 100644 --- a/fs/nilfs2/btree.c +++ b/fs/nilfs2/btree.c @@ -685,7 +685,8 @@ static int nilfs_btree_lookup(const struct nilfs_bmap *btree, } static int nilfs_btree_lookup_contig(const struct nilfs_bmap *btree, - __u64 key, __u64 *ptrp, unsigned maxblocks) + __u64 key, __u64 *ptrp, + unsigned int maxblocks) { struct nilfs_btree_path *path; struct nilfs_btree_node *node; diff --git a/fs/nilfs2/cpfile.c b/fs/nilfs2/cpfile.c index 16f884bd857c..b61c3e0eb342 100644 --- a/fs/nilfs2/cpfile.c +++ b/fs/nilfs2/cpfile.c @@ -431,7 +431,8 @@ static void nilfs_cpfile_checkpoint_to_cpinfo(struct inode *cpfile, } static ssize_t nilfs_cpfile_do_get_cpinfo(struct inode *cpfile, __u64 *cnop, - void *buf, unsigned cisz, size_t nci) + void *buf, unsigned int cisz, + size_t nci) { struct nilfs_checkpoint *cp; struct nilfs_cpinfo *ci = buf; @@ -482,7 +483,8 @@ static ssize_t nilfs_cpfile_do_get_cpinfo(struct inode *cpfile, __u64 *cnop, } static ssize_t nilfs_cpfile_do_get_ssinfo(struct inode *cpfile, __u64 *cnop, - void *buf, unsigned cisz, size_t nci) + void *buf, unsigned int cisz, + size_t nci) { struct buffer_head *bh; struct nilfs_cpfile_header *header; @@ -568,7 +570,7 @@ static ssize_t nilfs_cpfile_do_get_ssinfo(struct inode *cpfile, __u64 *cnop, */ ssize_t nilfs_cpfile_get_cpinfo(struct inode *cpfile, __u64 *cnop, int mode, - void *buf, unsigned cisz, size_t nci) + void *buf, unsigned int cisz, size_t nci) { switch (mode) { case NILFS_CHECKPOINT: diff --git a/fs/nilfs2/cpfile.h b/fs/nilfs2/cpfile.h index 5bdb8262928b..0249744ae234 100644 --- a/fs/nilfs2/cpfile.h +++ b/fs/nilfs2/cpfile.h @@ -33,8 +33,8 @@ int nilfs_cpfile_delete_checkpoint(struct inode *, __u64); int nilfs_cpfile_change_cpmode(struct inode *, __u64, int); int nilfs_cpfile_is_snapshot(struct inode *, __u64); int nilfs_cpfile_get_stat(struct inode *, struct nilfs_cpstat *); -ssize_t nilfs_cpfile_get_cpinfo(struct inode *, __u64 *, int, void *, unsigned, - size_t); +ssize_t nilfs_cpfile_get_cpinfo(struct inode *, __u64 *, int, void *, + unsigned int, size_t); int nilfs_cpfile_read(struct super_block *sb, size_t cpsize, struct nilfs_inode *raw_inode, struct inode **inodep); diff --git a/fs/nilfs2/dat.c b/fs/nilfs2/dat.c index e92257bc42ee..7367610ea807 100644 --- a/fs/nilfs2/dat.c +++ b/fs/nilfs2/dat.c @@ -424,7 +424,7 @@ int nilfs_dat_translate(struct inode *dat, __u64 vblocknr, sector_t *blocknrp) return ret; } -ssize_t nilfs_dat_get_vinfo(struct inode *dat, void *buf, unsigned visz, +ssize_t nilfs_dat_get_vinfo(struct inode *dat, void *buf, unsigned int visz, size_t nvi) { struct buffer_head *entry_bh; diff --git a/fs/nilfs2/dat.h b/fs/nilfs2/dat.h index c7035b5b1aed..abbfdabcabea 100644 --- a/fs/nilfs2/dat.h +++ b/fs/nilfs2/dat.h @@ -47,7 +47,7 @@ void nilfs_dat_abort_update(struct inode *, struct nilfs_palloc_req *, int nilfs_dat_mark_dirty(struct inode *, __u64); int nilfs_dat_freev(struct inode *, __u64 *, size_t); int nilfs_dat_move(struct inode *, __u64, sector_t); -ssize_t nilfs_dat_get_vinfo(struct inode *, void *, unsigned, size_t); +ssize_t nilfs_dat_get_vinfo(struct inode *, void *, unsigned int, size_t); int nilfs_dat_read(struct super_block *sb, size_t entry_size, struct nilfs_inode *raw_inode, struct inode **inodep); diff --git a/fs/nilfs2/dir.c b/fs/nilfs2/dir.c index bbcc03de1e74..e506f4f7120a 100644 --- a/fs/nilfs2/dir.c +++ b/fs/nilfs2/dir.c @@ -46,7 +46,7 @@ * nilfs uses block-sized chunks. Arguably, sector-sized ones would be * more robust, but we have what we have */ -static inline unsigned nilfs_chunk_size(struct inode *inode) +static inline unsigned int nilfs_chunk_size(struct inode *inode) { return inode->i_sb->s_blocksize; } @@ -61,9 +61,9 @@ static inline void nilfs_put_page(struct page *page) * Return the offset into page `page_nr' of the last valid * byte in that page, plus one. */ -static unsigned nilfs_last_byte(struct inode *inode, unsigned long page_nr) +static unsigned int nilfs_last_byte(struct inode *inode, unsigned long page_nr) { - unsigned last_byte = inode->i_size; + unsigned int last_byte = inode->i_size; last_byte -= page_nr << PAGE_SHIFT; if (last_byte > PAGE_SIZE) @@ -71,7 +71,8 @@ static unsigned nilfs_last_byte(struct inode *inode, unsigned long page_nr) return last_byte; } -static int nilfs_prepare_chunk(struct page *page, unsigned from, unsigned to) +static int nilfs_prepare_chunk(struct page *page, unsigned int from, + unsigned int to) { loff_t pos = page_offset(page) + from; @@ -80,12 +81,12 @@ static int nilfs_prepare_chunk(struct page *page, unsigned from, unsigned to) static void nilfs_commit_chunk(struct page *page, struct address_space *mapping, - unsigned from, unsigned to) + unsigned int from, unsigned int to) { struct inode *dir = mapping->host; loff_t pos = page_offset(page) + from; - unsigned len = to - from; - unsigned nr_dirty, copied; + unsigned int len = to - from; + unsigned int nr_dirty, copied; int err; nr_dirty = nilfs_page_count_clean_buffers(page, from, to); @@ -103,10 +104,10 @@ static bool nilfs_check_page(struct page *page) { struct inode *dir = page->mapping->host; struct super_block *sb = dir->i_sb; - unsigned chunk_size = nilfs_chunk_size(dir); + unsigned int chunk_size = nilfs_chunk_size(dir); char *kaddr = page_address(page); - unsigned offs, rec_len; - unsigned limit = PAGE_SIZE; + unsigned int offs, rec_len; + unsigned int limit = PAGE_SIZE; struct nilfs_dir_entry *p; char *error; @@ -256,7 +257,6 @@ static int nilfs_readdir(struct file *file, struct dir_context *ctx) unsigned int offset = pos & ~PAGE_MASK; unsigned long n = pos >> PAGE_SHIFT; unsigned long npages = dir_pages(inode); -/* unsigned chunk_mask = ~(nilfs_chunk_size(inode)-1); */ if (pos > inode->i_size - NILFS_DIR_REC_LEN(1)) return 0; @@ -318,7 +318,7 @@ nilfs_find_entry(struct inode *dir, const struct qstr *qstr, { const unsigned char *name = qstr->name; int namelen = qstr->len; - unsigned reclen = NILFS_DIR_REC_LEN(namelen); + unsigned int reclen = NILFS_DIR_REC_LEN(namelen); unsigned long start, n; unsigned long npages = dir_pages(dir); struct page *page = NULL; @@ -408,8 +408,8 @@ ino_t nilfs_inode_by_name(struct inode *dir, const struct qstr *qstr) void nilfs_set_link(struct inode *dir, struct nilfs_dir_entry *de, struct page *page, struct inode *inode) { - unsigned from = (char *) de - (char *) page_address(page); - unsigned to = from + nilfs_rec_len_from_disk(de->rec_len); + unsigned int from = (char *)de - (char *)page_address(page); + unsigned int to = from + nilfs_rec_len_from_disk(de->rec_len); struct address_space *mapping = page->mapping; int err; @@ -431,15 +431,15 @@ int nilfs_add_link(struct dentry *dentry, struct inode *inode) struct inode *dir = d_inode(dentry->d_parent); const unsigned char *name = dentry->d_name.name; int namelen = dentry->d_name.len; - unsigned chunk_size = nilfs_chunk_size(dir); - unsigned reclen = NILFS_DIR_REC_LEN(namelen); + unsigned int chunk_size = nilfs_chunk_size(dir); + unsigned int reclen = NILFS_DIR_REC_LEN(namelen); unsigned short rec_len, name_len; struct page *page = NULL; struct nilfs_dir_entry *de; unsigned long npages = dir_pages(dir); unsigned long n; char *kaddr; - unsigned from, to; + unsigned int from, to; int err; /* @@ -531,13 +531,14 @@ int nilfs_delete_entry(struct nilfs_dir_entry *dir, struct page *page) struct address_space *mapping = page->mapping; struct inode *inode = mapping->host; char *kaddr = page_address(page); - unsigned from = ((char *)dir - kaddr) & ~(nilfs_chunk_size(inode) - 1); - unsigned to = ((char *)dir - kaddr) + - nilfs_rec_len_from_disk(dir->rec_len); - struct nilfs_dir_entry *pde = NULL; - struct nilfs_dir_entry *de = (struct nilfs_dir_entry *)(kaddr + from); + unsigned int from, to; + struct nilfs_dir_entry *de, *pde = NULL; int err; + from = ((char *)dir - kaddr) & ~(nilfs_chunk_size(inode) - 1); + to = ((char *)dir - kaddr) + nilfs_rec_len_from_disk(dir->rec_len); + de = (struct nilfs_dir_entry *)(kaddr + from); + while ((char *)de < (char *)dir) { if (de->rec_len == 0) { nilfs_error(inode->i_sb, __func__, @@ -570,7 +571,7 @@ int nilfs_make_empty(struct inode *inode, struct inode *parent) { struct address_space *mapping = inode->i_mapping; struct page *page = grab_cache_page(mapping, 0); - unsigned chunk_size = nilfs_chunk_size(inode); + unsigned int chunk_size = nilfs_chunk_size(inode); struct nilfs_dir_entry *de; int err; void *kaddr; diff --git a/fs/nilfs2/direct.c b/fs/nilfs2/direct.c index 22058d0b36e9..001068630063 100644 --- a/fs/nilfs2/direct.c +++ b/fs/nilfs2/direct.c @@ -58,7 +58,7 @@ static int nilfs_direct_lookup(const struct nilfs_bmap *direct, static int nilfs_direct_lookup_contig(const struct nilfs_bmap *direct, __u64 key, __u64 *ptrp, - unsigned maxblocks) + unsigned int maxblocks) { struct inode *dat = NULL; __u64 ptr, ptr2; @@ -79,7 +79,8 @@ static int nilfs_direct_lookup_contig(const struct nilfs_bmap *direct, ptr = blocknr; } - maxblocks = min_t(unsigned, maxblocks, NILFS_DIRECT_KEY_MAX - key + 1); + maxblocks = min_t(unsigned int, maxblocks, + NILFS_DIRECT_KEY_MAX - key + 1); for (cnt = 1; cnt < maxblocks && (ptr2 = nilfs_direct_get_ptr(direct, key + cnt)) != NILFS_BMAP_INVALID_PTR; diff --git a/fs/nilfs2/inode.c b/fs/nilfs2/inode.c index bbb47e8bde3e..83d2c485efba 100644 --- a/fs/nilfs2/inode.c +++ b/fs/nilfs2/inode.c @@ -83,7 +83,7 @@ int nilfs_get_block(struct inode *inode, sector_t blkoff, struct the_nilfs *nilfs = inode->i_sb->s_fs_info; __u64 blknum = 0; int err = 0, ret; - unsigned maxblocks = bh_result->b_size >> inode->i_blkbits; + unsigned int maxblocks = bh_result->b_size >> inode->i_blkbits; down_read(&NILFS_MDT(nilfs->ns_dat)->mi_sem); ret = nilfs_bmap_lookup_contig(ii->i_bmap, blkoff, &blknum, maxblocks); @@ -163,7 +163,7 @@ static int nilfs_readpage(struct file *file, struct page *page) * @nr_pages - number of pages to be read */ static int nilfs_readpages(struct file *file, struct address_space *mapping, - struct list_head *pages, unsigned nr_pages) + struct list_head *pages, unsigned int nr_pages) { return mpage_readpages(mapping, pages, nr_pages, nilfs_get_block); } @@ -222,7 +222,7 @@ static int nilfs_set_page_dirty(struct page *page) int ret = __set_page_dirty_nobuffers(page); if (page_has_buffers(page)) { - unsigned nr_dirty = 0; + unsigned int nr_dirty = 0; struct buffer_head *bh, *head; /* @@ -245,7 +245,7 @@ static int nilfs_set_page_dirty(struct page *page) if (nr_dirty) nilfs_set_file_dirty(inode, nr_dirty); } else if (ret) { - unsigned nr_dirty = 1 << (PAGE_SHIFT - inode->i_blkbits); + unsigned int nr_dirty = 1 << (PAGE_SHIFT - inode->i_blkbits); nilfs_set_file_dirty(inode, nr_dirty); } @@ -287,8 +287,8 @@ static int nilfs_write_end(struct file *file, struct address_space *mapping, struct page *page, void *fsdata) { struct inode *inode = mapping->host; - unsigned start = pos & (PAGE_SIZE - 1); - unsigned nr_dirty; + unsigned int start = pos & (PAGE_SIZE - 1); + unsigned int nr_dirty; int err; nr_dirty = nilfs_page_count_clean_buffers(page, start, @@ -902,7 +902,7 @@ int nilfs_inode_dirty(struct inode *inode) return ret; } -int nilfs_set_file_dirty(struct inode *inode, unsigned nr_dirty) +int nilfs_set_file_dirty(struct inode *inode, unsigned int nr_dirty) { struct nilfs_inode_info *ii = NILFS_I(inode); struct the_nilfs *nilfs = inode->i_sb->s_fs_info; diff --git a/fs/nilfs2/mdt.c b/fs/nilfs2/mdt.c index 8a2f8b240f25..3417d859a03c 100644 --- a/fs/nilfs2/mdt.c +++ b/fs/nilfs2/mdt.c @@ -490,8 +490,8 @@ void nilfs_mdt_destroy(struct inode *inode) kfree(mdi); } -void nilfs_mdt_set_entry_size(struct inode *inode, unsigned entry_size, - unsigned header_size) +void nilfs_mdt_set_entry_size(struct inode *inode, unsigned int entry_size, + unsigned int header_size) { struct nilfs_mdt_info *mi = NILFS_MDT(inode); diff --git a/fs/nilfs2/mdt.h b/fs/nilfs2/mdt.h index ffb876e6efed..3f67f3932097 100644 --- a/fs/nilfs2/mdt.h +++ b/fs/nilfs2/mdt.h @@ -53,8 +53,8 @@ struct nilfs_shadow_map { struct nilfs_mdt_info { struct rw_semaphore mi_sem; struct blockgroup_lock *mi_bgl; - unsigned mi_entry_size; - unsigned mi_first_entry_offset; + unsigned int mi_entry_size; + unsigned int mi_first_entry_offset; unsigned long mi_entries_per_block; struct nilfs_palloc_cache *mi_palloc_cache; struct nilfs_shadow_map *mi_shadow; @@ -90,7 +90,7 @@ int nilfs_mdt_init(struct inode *inode, gfp_t gfp_mask, size_t objsz); void nilfs_mdt_clear(struct inode *inode); void nilfs_mdt_destroy(struct inode *inode); -void nilfs_mdt_set_entry_size(struct inode *, unsigned, unsigned); +void nilfs_mdt_set_entry_size(struct inode *, unsigned int, unsigned int); int nilfs_mdt_setup_shadow_map(struct inode *inode, struct nilfs_shadow_map *shadow); diff --git a/fs/nilfs2/namei.c b/fs/nilfs2/namei.c index 8f8070cffa58..1ec8ae5995a5 100644 --- a/fs/nilfs2/namei.c +++ b/fs/nilfs2/namei.c @@ -139,7 +139,7 @@ static int nilfs_symlink(struct inode *dir, struct dentry *dentry, { struct nilfs_transaction_info ti; struct super_block *sb = dir->i_sb; - unsigned l = strlen(symname)+1; + unsigned int l = strlen(symname) + 1; struct inode *inode; int err; diff --git a/fs/nilfs2/nilfs.h b/fs/nilfs2/nilfs.h index fa179d4e5dba..ea320315d557 100644 --- a/fs/nilfs2/nilfs.h +++ b/fs/nilfs2/nilfs.h @@ -274,7 +274,7 @@ extern void nilfs_write_failed(struct address_space *mapping, loff_t to); int nilfs_permission(struct inode *inode, int mask); int nilfs_load_inode_block(struct inode *inode, struct buffer_head **pbh); extern int nilfs_inode_dirty(struct inode *); -int nilfs_set_file_dirty(struct inode *inode, unsigned nr_dirty); +int nilfs_set_file_dirty(struct inode *inode, unsigned int nr_dirty); extern int __nilfs_mark_inode_dirty(struct inode *, int); extern void nilfs_dirty_inode(struct inode *, int flags); int nilfs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, diff --git a/fs/nilfs2/page.c b/fs/nilfs2/page.c index 19687139f197..d97ba5f11b77 100644 --- a/fs/nilfs2/page.c +++ b/fs/nilfs2/page.c @@ -435,12 +435,12 @@ void nilfs_clear_dirty_page(struct page *page, bool silent) __nilfs_clear_page_dirty(page); } -unsigned nilfs_page_count_clean_buffers(struct page *page, - unsigned from, unsigned to) +unsigned int nilfs_page_count_clean_buffers(struct page *page, + unsigned int from, unsigned int to) { - unsigned block_start, block_end; + unsigned int block_start, block_end; struct buffer_head *bh, *head; - unsigned nc = 0; + unsigned int nc = 0; for (bh = head = page_buffers(page), block_start = 0; bh != head || !block_start; diff --git a/fs/nilfs2/page.h b/fs/nilfs2/page.h index 041f2dc5e634..f3687c958fa8 100644 --- a/fs/nilfs2/page.h +++ b/fs/nilfs2/page.h @@ -53,7 +53,8 @@ void nilfs_copy_back_pages(struct address_space *, struct address_space *); void nilfs_clear_dirty_page(struct page *, bool); void nilfs_clear_dirty_pages(struct address_space *, bool); void nilfs_mapping_init(struct address_space *mapping, struct inode *inode); -unsigned nilfs_page_count_clean_buffers(struct page *, unsigned, unsigned); +unsigned int nilfs_page_count_clean_buffers(struct page *, unsigned int, + unsigned int); unsigned long nilfs_find_uncommitted_extent(struct inode *inode, sector_t start_blk, sector_t *blkoff); diff --git a/fs/nilfs2/recovery.c b/fs/nilfs2/recovery.c index 685fa73cecd0..db156a199149 100644 --- a/fs/nilfs2/recovery.c +++ b/fs/nilfs2/recovery.c @@ -152,7 +152,7 @@ int nilfs_read_super_root_block(struct the_nilfs *nilfs, sector_t sr_block, sr = (struct nilfs_super_root *)bh_sr->b_data; if (check) { - unsigned bytes = le16_to_cpu(sr->sr_bytes); + unsigned int bytes = le16_to_cpu(sr->sr_bytes); if (bytes == 0 || bytes > nilfs->ns_blocksize) { ret = NILFS_SEG_FAIL_CHECKSUM_SUPER_ROOT; @@ -504,7 +504,7 @@ static int nilfs_recover_dsync_blocks(struct the_nilfs *nilfs, { struct inode *inode; struct nilfs_recovery_block *rb, *n; - unsigned blocksize = nilfs->ns_blocksize; + unsigned int blocksize = nilfs->ns_blocksize; struct page *page; loff_t pos; int err = 0, err2 = 0; diff --git a/fs/nilfs2/segbuf.c b/fs/nilfs2/segbuf.c index 52f6a6c8bab1..bf36df10540b 100644 --- a/fs/nilfs2/segbuf.c +++ b/fs/nilfs2/segbuf.c @@ -129,7 +129,7 @@ int nilfs_segbuf_extend_payload(struct nilfs_segment_buffer *segbuf, return 0; } -int nilfs_segbuf_reset(struct nilfs_segment_buffer *segbuf, unsigned flags, +int nilfs_segbuf_reset(struct nilfs_segment_buffer *segbuf, unsigned int flags, time_t ctime, __u64 cno) { int err; @@ -236,7 +236,7 @@ nilfs_segbuf_fill_in_super_root_crc(struct nilfs_segment_buffer *segbuf, { struct nilfs_super_root *raw_sr; struct the_nilfs *nilfs = segbuf->sb_super->s_fs_info; - unsigned srsize; + unsigned int srsize; u32 crc; raw_sr = (struct nilfs_super_root *)segbuf->sb_super_root->b_data; diff --git a/fs/nilfs2/segbuf.h b/fs/nilfs2/segbuf.h index fc3a0fe3119c..7bbccc099709 100644 --- a/fs/nilfs2/segbuf.h +++ b/fs/nilfs2/segbuf.h @@ -78,7 +78,7 @@ struct nilfs_segment_buffer { __u64 sb_nextnum; sector_t sb_fseg_start, sb_fseg_end; sector_t sb_pseg_start; - unsigned sb_rest_blocks; + unsigned int sb_rest_blocks; /* Buffers */ struct list_head sb_segsum_buffers; @@ -120,7 +120,8 @@ void nilfs_segbuf_map_cont(struct nilfs_segment_buffer *segbuf, struct nilfs_segment_buffer *prev); void nilfs_segbuf_set_next_segnum(struct nilfs_segment_buffer *, __u64, struct the_nilfs *); -int nilfs_segbuf_reset(struct nilfs_segment_buffer *, unsigned, time_t, __u64); +int nilfs_segbuf_reset(struct nilfs_segment_buffer *, unsigned int, time_t, + __u64); int nilfs_segbuf_extend_segsum(struct nilfs_segment_buffer *); int nilfs_segbuf_extend_payload(struct nilfs_segment_buffer *, struct buffer_head **); diff --git a/fs/nilfs2/segment.c b/fs/nilfs2/segment.c index a6ef1eb15edb..97dee069be83 100644 --- a/fs/nilfs2/segment.c +++ b/fs/nilfs2/segment.c @@ -393,10 +393,10 @@ static void nilfs_transaction_unlock(struct super_block *sb) static void *nilfs_segctor_map_segsum_entry(struct nilfs_sc_info *sci, struct nilfs_segsum_pointer *ssp, - unsigned bytes) + unsigned int bytes) { struct nilfs_segment_buffer *segbuf = sci->sc_curseg; - unsigned blocksize = sci->sc_super->s_blocksize; + unsigned int blocksize = sci->sc_super->s_blocksize; void *p; if (unlikely(ssp->offset + bytes > blocksize)) { @@ -418,8 +418,8 @@ static int nilfs_segctor_reset_segment_buffer(struct nilfs_sc_info *sci) { struct nilfs_segment_buffer *segbuf = sci->sc_curseg; struct buffer_head *sumbh; - unsigned sumbytes; - unsigned flags = 0; + unsigned int sumbytes; + unsigned int flags = 0; int err; if (nilfs_doing_gc()) @@ -468,9 +468,9 @@ static int nilfs_segctor_add_super_root(struct nilfs_sc_info *sci) */ static int nilfs_segctor_segsum_block_required( struct nilfs_sc_info *sci, const struct nilfs_segsum_pointer *ssp, - unsigned binfo_size) + unsigned int binfo_size) { - unsigned blocksize = sci->sc_super->s_blocksize; + unsigned int blocksize = sci->sc_super->s_blocksize; /* Size of finfo and binfo is enough small against blocksize */ return ssp->offset + binfo_size + @@ -529,7 +529,7 @@ static void nilfs_segctor_end_finfo(struct nilfs_sc_info *sci, static int nilfs_segctor_add_file_block(struct nilfs_sc_info *sci, struct buffer_head *bh, struct inode *inode, - unsigned binfo_size) + unsigned int binfo_size) { struct nilfs_segment_buffer *segbuf; int required, err = 0; @@ -773,7 +773,7 @@ static void nilfs_dispose_list(struct the_nilfs *nilfs, { struct nilfs_inode_info *ii, *n; struct nilfs_inode_info *ivec[SC_N_INODEVEC], **pii; - unsigned nv = 0; + unsigned int nv = 0; while (!list_empty(head)) { spin_lock(&nilfs->ns_inode_lock); @@ -954,7 +954,7 @@ static void nilfs_segctor_fill_in_super_root(struct nilfs_sc_info *sci, { struct buffer_head *bh_sr; struct nilfs_super_root *raw_sr; - unsigned isz, srsz; + unsigned int isz, srsz; bh_sr = NILFS_LAST_SEGBUF(&sci->sc_segbufs)->sb_super_root; raw_sr = (struct nilfs_super_root *)bh_sr->b_data; diff --git a/fs/nilfs2/segment.h b/fs/nilfs2/segment.h index 6cb12dbee7c1..27822e760d3f 100644 --- a/fs/nilfs2/segment.h +++ b/fs/nilfs2/segment.h @@ -71,7 +71,7 @@ struct nilfs_recovery_info { */ struct nilfs_cstage { int scnt; - unsigned flags; + unsigned int flags; struct nilfs_inode_info *dirty_file_ptr; struct nilfs_inode_info *gc_inode_ptr; }; @@ -80,7 +80,7 @@ struct nilfs_segment_buffer; struct nilfs_segsum_pointer { struct buffer_head *bh; - unsigned offset; /* offset in bytes */ + unsigned int offset; /* offset in bytes */ }; /** diff --git a/fs/nilfs2/sysfs.c b/fs/nilfs2/sysfs.c index 3e7d85335adf..8ffa42b704d8 100644 --- a/fs/nilfs2/sysfs.c +++ b/fs/nilfs2/sysfs.c @@ -756,7 +756,7 @@ nilfs_superblock_sb_write_count_show(struct nilfs_superblock_attr *attr, struct the_nilfs *nilfs, char *buf) { - unsigned sbwcount; + unsigned int sbwcount; down_read(&nilfs->ns_sem); sbwcount = nilfs->ns_sbwcount; @@ -770,7 +770,7 @@ nilfs_superblock_sb_update_frequency_show(struct nilfs_superblock_attr *attr, struct the_nilfs *nilfs, char *buf) { - unsigned sb_update_freq; + unsigned int sb_update_freq; down_read(&nilfs->ns_sem); sb_update_freq = nilfs->ns_sb_update_freq; @@ -784,7 +784,7 @@ nilfs_superblock_sb_update_frequency_store(struct nilfs_superblock_attr *attr, struct the_nilfs *nilfs, const char *buf, size_t count) { - unsigned val; + unsigned int val; int err; err = kstrtouint(skip_spaces(buf), 0, &val); diff --git a/fs/nilfs2/the_nilfs.c b/fs/nilfs2/the_nilfs.c index ba4b8189c342..b9e19ca3c96e 100644 --- a/fs/nilfs2/the_nilfs.c +++ b/fs/nilfs2/the_nilfs.c @@ -108,8 +108,8 @@ static int nilfs_load_super_root(struct the_nilfs *nilfs, struct nilfs_super_root *raw_sr; struct nilfs_super_block **sbp = nilfs->ns_sbp; struct nilfs_inode *rawi; - unsigned dat_entry_size, segment_usage_size, checkpoint_size; - unsigned inode_size; + unsigned int dat_entry_size, segment_usage_size, checkpoint_size; + unsigned int inode_size; int err; err = nilfs_read_super_root_block(nilfs, sr_block, &bh_sr, 1); diff --git a/fs/nilfs2/the_nilfs.h b/fs/nilfs2/the_nilfs.h index 62bd7b10fe43..06d2548d436d 100644 --- a/fs/nilfs2/the_nilfs.h +++ b/fs/nilfs2/the_nilfs.h @@ -114,10 +114,10 @@ struct the_nilfs { struct buffer_head *ns_sbh[2]; struct nilfs_super_block *ns_sbp[2]; time_t ns_sbwtime; - unsigned ns_sbwcount; - unsigned ns_sbsize; - unsigned ns_mount_state; - unsigned ns_sb_update_freq; + unsigned int ns_sbwcount; + unsigned int ns_sbsize; + unsigned int ns_mount_state; + unsigned int ns_sb_update_freq; /* * Following fields are dedicated to a writable FS-instance. @@ -306,7 +306,7 @@ static inline void nilfs_get_root(struct nilfs_root *root) static inline int nilfs_valid_fs(struct the_nilfs *nilfs) { - unsigned valid_fs; + unsigned int valid_fs; down_read(&nilfs->ns_sem); valid_fs = (nilfs->ns_mount_state & NILFS_VALID_FS); diff --git a/include/linux/nilfs2_fs.h b/include/linux/nilfs2_fs.h index 823d63d61081..3b584925d0e8 100644 --- a/include/linux/nilfs2_fs.h +++ b/include/linux/nilfs2_fs.h @@ -322,9 +322,9 @@ enum { ~NILFS_DIR_ROUND) #define NILFS_MAX_REC_LEN ((1<<16)-1) -static inline unsigned nilfs_rec_len_from_disk(__le16 dlen) +static inline unsigned int nilfs_rec_len_from_disk(__le16 dlen) { - unsigned len = le16_to_cpu(dlen); + unsigned int len = le16_to_cpu(dlen); #if !defined(__KERNEL__) || (PAGE_SIZE >= 65536) if (len == NILFS_MAX_REC_LEN) @@ -333,7 +333,7 @@ static inline unsigned nilfs_rec_len_from_disk(__le16 dlen) return len; } -static inline __le16 nilfs_rec_len_to_disk(unsigned len) +static inline __le16 nilfs_rec_len_to_disk(unsigned int len) { #if !defined(__KERNEL__) || (PAGE_SIZE >= 65536) if (len == (1 << 16)) -- cgit v1.3-8-gc7d7 From 7f00184e9cb534aca1f10f3027cb6137b03f101e Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Mon, 23 May 2016 16:23:42 -0700 Subject: nilfs2: remove unnecessary else after return or break This fixes the checkpatch.pl warning that suggests else is not generally useful after a break or return. Link: http://lkml.kernel.org/r/1462886671-3521-6-git-send-email-konishi.ryusuke@lab.ntt.co.jp Signed-off-by: Ryusuke Konishi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nilfs2/btree.c | 12 ++++++------ fs/nilfs2/direct.c | 6 +++--- fs/nilfs2/segment.c | 26 ++++++++++++-------------- 3 files changed, 21 insertions(+), 23 deletions(-) (limited to 'fs') diff --git a/fs/nilfs2/btree.c b/fs/nilfs2/btree.c index 57ec6af28b49..eccb1c89ccbb 100644 --- a/fs/nilfs2/btree.c +++ b/fs/nilfs2/btree.c @@ -1029,12 +1029,12 @@ static __u64 nilfs_btree_find_target_v(const struct nilfs_bmap *btree, if (ptr != NILFS_BMAP_INVALID_PTR) /* sequential access */ return ptr; - else { - ptr = nilfs_btree_find_near(btree, path); - if (ptr != NILFS_BMAP_INVALID_PTR) - /* near */ - return ptr; - } + + ptr = nilfs_btree_find_near(btree, path); + if (ptr != NILFS_BMAP_INVALID_PTR) + /* near */ + return ptr; + /* block group */ return nilfs_bmap_find_target_in_group(btree); } diff --git a/fs/nilfs2/direct.c b/fs/nilfs2/direct.c index 001068630063..251a44928405 100644 --- a/fs/nilfs2/direct.c +++ b/fs/nilfs2/direct.c @@ -107,9 +107,9 @@ nilfs_direct_find_target_v(const struct nilfs_bmap *direct, __u64 key) if (ptr != NILFS_BMAP_INVALID_PTR) /* sequential access */ return ptr; - else - /* block group */ - return nilfs_bmap_find_target_in_group(direct); + + /* block group */ + return nilfs_bmap_find_target_in_group(direct); } static int nilfs_direct_insert(struct nilfs_bmap *bmap, __u64 key, __u64 ptr) diff --git a/fs/nilfs2/segment.c b/fs/nilfs2/segment.c index 97dee069be83..c9ee03c262ea 100644 --- a/fs/nilfs2/segment.c +++ b/fs/nilfs2/segment.c @@ -150,17 +150,15 @@ static int nilfs_prepare_segment_lock(struct nilfs_transaction_info *ti) if (cur_ti) { if (cur_ti->ti_magic == NILFS_TI_MAGIC) return ++cur_ti->ti_count; - else { - /* - * If journal_info field is occupied by other FS, - * it is saved and will be restored on - * nilfs_transaction_commit(). - */ - printk(KERN_WARNING - "NILFS warning: journal info from a different " - "FS\n"); - save = current->journal_info; - } + + /* + * If journal_info field is occupied by other FS, + * it is saved and will be restored on + * nilfs_transaction_commit(). + */ + printk(KERN_WARNING + "NILFS warning: journal info from a different FS\n"); + save = current->journal_info; } if (!ti) { ti = kmem_cache_alloc(nilfs_transaction_cachep, GFP_NOFS); @@ -2552,10 +2550,10 @@ static int nilfs_segctor_thread(void *arg) if (timeout || sci->sc_seq_request != sci->sc_seq_done) mode = SC_LSEG_SR; - else if (!sci->sc_flush_request) - break; - else + else if (sci->sc_flush_request) mode = nilfs_segctor_flush_mode(sci); + else + break; spin_unlock(&sci->sc_state_lock); nilfs_segctor_thread_construct(sci, mode); -- cgit v1.3-8-gc7d7 From 80d6505232ec2476eef8d988ea98ac81eb2e0975 Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Mon, 23 May 2016 16:23:45 -0700 Subject: nilfs2: remove loops of single statement macros This fixes checkpatch.pl warning "WARNING: Single statement macros should not use a do {} while (0) loop". Link: http://lkml.kernel.org/r/1462886671-3521-7-git-send-email-konishi.ryusuke@lab.ntt.co.jp Signed-off-by: Ryusuke Konishi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nilfs2/the_nilfs.h | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/nilfs2/the_nilfs.h b/fs/nilfs2/the_nilfs.h index 06d2548d436d..79369fd6b13b 100644 --- a/fs/nilfs2/the_nilfs.h +++ b/fs/nilfs2/the_nilfs.h @@ -222,15 +222,14 @@ THE_NILFS_FNS(SB_DIRTY, sb_dirty) * Mount option operations */ #define nilfs_clear_opt(nilfs, opt) \ - do { (nilfs)->ns_mount_opt &= ~NILFS_MOUNT_##opt; } while (0) + ((nilfs)->ns_mount_opt &= ~NILFS_MOUNT_##opt) #define nilfs_set_opt(nilfs, opt) \ - do { (nilfs)->ns_mount_opt |= NILFS_MOUNT_##opt; } while (0) + ((nilfs)->ns_mount_opt |= NILFS_MOUNT_##opt) #define nilfs_test_opt(nilfs, opt) ((nilfs)->ns_mount_opt & NILFS_MOUNT_##opt) #define nilfs_write_opt(nilfs, mask, opt) \ - do { (nilfs)->ns_mount_opt = \ + ((nilfs)->ns_mount_opt = \ (((nilfs)->ns_mount_opt & ~NILFS_MOUNT_##mask) | \ - NILFS_MOUNT_##opt); \ - } while (0) + NILFS_MOUNT_##opt)) \ /** * struct nilfs_root - nilfs root object -- cgit v1.3-8-gc7d7 From 076a378ba6e6b6ddd5f2336aa0876349b7d36409 Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Mon, 23 May 2016 16:23:48 -0700 Subject: nilfs2: fix block comments This fixes block comments with proper formatting to eliminate the following checkpatch.pl warnings: "WARNING: Block comments use * on subsequent lines" "WARNING: Block comments use a trailing */ on a separate line" Link: http://lkml.kernel.org/r/1462886671-3521-8-git-send-email-konishi.ryusuke@lab.ntt.co.jp Signed-off-by: Ryusuke Konishi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nilfs2/alloc.c | 12 +++++--- fs/nilfs2/bmap.h | 12 +++++--- fs/nilfs2/cpfile.c | 6 ++-- fs/nilfs2/ifile.c | 6 ++-- fs/nilfs2/inode.c | 73 ++++++++++++++++++++++++++++++----------------- fs/nilfs2/nilfs.h | 24 ++++++++++------ fs/nilfs2/recovery.c | 14 +++++---- fs/nilfs2/segment.c | 56 ++++++++++++++++++++++++------------ fs/nilfs2/segment.h | 34 ++++++++++++++-------- fs/nilfs2/the_nilfs.c | 6 ++-- include/linux/nilfs2_fs.h | 66 +++++++++++++++++++++++++++--------------- 11 files changed, 202 insertions(+), 107 deletions(-) (limited to 'fs') diff --git a/fs/nilfs2/alloc.c b/fs/nilfs2/alloc.c index 698f582d69af..1a85d94f5b25 100644 --- a/fs/nilfs2/alloc.c +++ b/fs/nilfs2/alloc.c @@ -68,13 +68,17 @@ int nilfs_palloc_init_blockgroup(struct inode *inode, unsigned int entry_size) mi->mi_blocks_per_group = DIV_ROUND_UP(nilfs_palloc_entries_per_group(inode), mi->mi_entries_per_block) + 1; - /* Number of blocks in a group including entry blocks and - a bitmap block */ + /* + * Number of blocks in a group including entry blocks + * and a bitmap block + */ mi->mi_blocks_per_desc_block = nilfs_palloc_groups_per_desc_block(inode) * mi->mi_blocks_per_group + 1; - /* Number of blocks per descriptor including the - descriptor block */ + /* + * Number of blocks per descriptor including the + * descriptor block + */ return 0; } diff --git a/fs/nilfs2/bmap.h b/fs/nilfs2/bmap.h index a6852807b22c..b6a4c8f93ac8 100644 --- a/fs/nilfs2/bmap.h +++ b/fs/nilfs2/bmap.h @@ -122,10 +122,14 @@ struct nilfs_bmap { /* pointer type */ #define NILFS_BMAP_PTR_P 0 /* physical block number (i.e. LBN) */ -#define NILFS_BMAP_PTR_VS 1 /* virtual block number (single - version) */ -#define NILFS_BMAP_PTR_VM 2 /* virtual block number (has multiple - versions) */ +#define NILFS_BMAP_PTR_VS 1 /* + * virtual block number (single + * version) + */ +#define NILFS_BMAP_PTR_VM 2 /* + * virtual block number (has multiple + * versions) + */ #define NILFS_BMAP_PTR_U (-1) /* never perform pointer operations */ #define NILFS_BMAP_USE_VBN(bmap) ((bmap)->b_ptr_type > 0) diff --git a/fs/nilfs2/cpfile.c b/fs/nilfs2/cpfile.c index b61c3e0eb342..8a3d3b65af3f 100644 --- a/fs/nilfs2/cpfile.c +++ b/fs/nilfs2/cpfile.c @@ -870,8 +870,10 @@ int nilfs_cpfile_is_snapshot(struct inode *cpfile, __u64 cno) void *kaddr; int ret; - /* CP number is invalid if it's zero or larger than the - largest exist one.*/ + /* + * CP number is invalid if it's zero or larger than the + * largest existing one. + */ if (cno == 0 || cno >= nilfs_mdt_cno(cpfile)) return -ENOENT; down_read(&NILFS_MDT(cpfile)->mi_sem); diff --git a/fs/nilfs2/ifile.c b/fs/nilfs2/ifile.c index d9048f6a7780..1d2b1805327a 100644 --- a/fs/nilfs2/ifile.c +++ b/fs/nilfs2/ifile.c @@ -64,8 +64,10 @@ int nilfs_ifile_create_inode(struct inode *ifile, ino_t *out_ino, struct nilfs_palloc_req req; int ret; - req.pr_entry_nr = 0; /* 0 says find free inode from beginning of - a group. dull code!! */ + req.pr_entry_nr = 0; /* + * 0 says find free inode from beginning + * of a group. dull code!! + */ req.pr_entry_bh = NULL; ret = nilfs_palloc_prepare_alloc_entry(ifile, &req); diff --git a/fs/nilfs2/inode.c b/fs/nilfs2/inode.c index 83d2c485efba..a0ebdb17e912 100644 --- a/fs/nilfs2/inode.c +++ b/fs/nilfs2/inode.c @@ -129,11 +129,14 @@ int nilfs_get_block(struct inode *inode, sector_t blkoff, /* Error handling should be detailed */ set_buffer_new(bh_result); set_buffer_delay(bh_result); - map_bh(bh_result, inode->i_sb, 0); /* dbn must be changed - to proper value */ + map_bh(bh_result, inode->i_sb, 0); + /* Disk block number must be changed to proper value */ + } else if (ret == -ENOENT) { - /* not found is not error (e.g. hole); must return without - the mapped state flag. */ + /* + * not found is not error (e.g. hole); must return without + * the mapped state flag. + */ ; } else { err = ret; @@ -395,23 +398,26 @@ struct inode *nilfs_new_inode(struct inode *dir, umode_t mode) err = nilfs_init_acl(inode, dir); if (unlikely(err)) - goto failed_after_creation; /* never occur. When supporting - nilfs_init_acl(), proper cancellation of - above jobs should be considered */ + /* + * Never occur. When supporting nilfs_init_acl(), + * proper cancellation of above jobs should be considered. + */ + goto failed_after_creation; return inode; failed_after_creation: clear_nlink(inode); unlock_new_inode(inode); - iput(inode); /* raw_inode will be deleted through - nilfs_evict_inode() */ + iput(inode); /* + * raw_inode will be deleted through + * nilfs_evict_inode(). + */ goto failed; failed_ifile_create_inode: make_bad_inode(inode); - iput(inode); /* if i_nlink == 1, generic_forget_inode() will be - called */ + iput(inode); failed: return ERR_PTR(err); } @@ -662,8 +668,10 @@ void nilfs_write_inode_common(struct inode *inode, else if (S_ISCHR(inode->i_mode) || S_ISBLK(inode->i_mode)) raw_inode->i_device_code = cpu_to_le64(huge_encode_dev(inode->i_rdev)); - /* When extending inode, nilfs->ns_inode_size should be checked - for substitutions of appended fields */ + /* + * When extending inode, nilfs->ns_inode_size should be checked + * for substitutions of appended fields. + */ } void nilfs_update_inode(struct inode *inode, struct buffer_head *ibh, int flags) @@ -681,9 +689,12 @@ void nilfs_update_inode(struct inode *inode, struct buffer_head *ibh, int flags) set_bit(NILFS_I_INODE_SYNC, &ii->i_state); nilfs_write_inode_common(inode, raw_inode, 0); - /* XXX: call with has_bmap = 0 is a workaround to avoid - deadlock of bmap. This delays update of i_bmap to just - before writing */ + /* + * XXX: call with has_bmap = 0 is a workaround to avoid + * deadlock of bmap. This delays update of i_bmap to just + * before writing. + */ + nilfs_ifile_unmap_inode(ifile, ino, ibh); } @@ -748,8 +759,10 @@ void nilfs_truncate(struct inode *inode) nilfs_mark_inode_dirty(inode); nilfs_set_file_dirty(inode, 0); nilfs_transaction_commit(sb); - /* May construct a logical segment and may fail in sync mode. - But truncate has no return value. */ + /* + * May construct a logical segment and may fail in sync mode. + * But truncate has no return value. + */ } static void nilfs_clear_inode(struct inode *inode) @@ -806,8 +819,10 @@ void nilfs_evict_inode(struct inode *inode) if (IS_SYNC(inode)) nilfs_set_transaction_flag(NILFS_TI_SYNC); nilfs_transaction_commit(sb); - /* May construct a logical segment and may fail in sync mode. - But delete_inode has no return value. */ + /* + * May construct a logical segment and may fail in sync mode. + * But delete_inode has no return value. + */ } int nilfs_setattr(struct dentry *dentry, struct iattr *iattr) @@ -915,17 +930,23 @@ int nilfs_set_file_dirty(struct inode *inode, unsigned int nr_dirty) spin_lock(&nilfs->ns_inode_lock); if (!test_bit(NILFS_I_QUEUED, &ii->i_state) && !test_bit(NILFS_I_BUSY, &ii->i_state)) { - /* Because this routine may race with nilfs_dispose_list(), - we have to check NILFS_I_QUEUED here, too. */ + /* + * Because this routine may race with nilfs_dispose_list(), + * we have to check NILFS_I_QUEUED here, too. + */ if (list_empty(&ii->i_dirty) && igrab(inode) == NULL) { - /* This will happen when somebody is freeing - this inode. */ + /* + * This will happen when somebody is freeing + * this inode. + */ nilfs_warning(inode->i_sb, __func__, "cannot get inode (ino=%lu)", inode->i_ino); spin_unlock(&nilfs->ns_inode_lock); - return -EINVAL; /* NILFS_I_DIRTY may remain for - freeing inode */ + return -EINVAL; /* + * NILFS_I_DIRTY may remain for + * freeing inode. + */ } list_move_tail(&ii->i_dirty, &nilfs->ns_dirty_files); set_bit(NILFS_I_QUEUED, &ii->i_state); diff --git a/fs/nilfs2/nilfs.h b/fs/nilfs2/nilfs.h index ea320315d557..b1d48bc0532d 100644 --- a/fs/nilfs2/nilfs.h +++ b/fs/nilfs2/nilfs.h @@ -64,8 +64,10 @@ struct nilfs_inode_info { */ struct rw_semaphore xattr_sem; #endif - struct buffer_head *i_bh; /* i_bh contains a new or dirty - disk inode */ + struct buffer_head *i_bh; /* + * i_bh contains a new or dirty + * disk inode. + */ struct nilfs_root *i_root; struct inode vfs_inode; }; @@ -95,8 +97,10 @@ enum { NILFS_I_NEW = 0, /* Inode is newly created */ NILFS_I_DIRTY, /* The file is dirty */ NILFS_I_QUEUED, /* inode is in dirty_files list */ - NILFS_I_BUSY, /* inode is grabbed by a segment - constructor */ + NILFS_I_BUSY, /* + * Inode is grabbed by a segment + * constructor + */ NILFS_I_COLLECTED, /* All dirty blocks are collected */ NILFS_I_UPDATED, /* The file has been written back */ NILFS_I_INODE_SYNC, /* dsync is not allowed for inode */ @@ -140,8 +144,10 @@ enum { struct nilfs_transaction_info { u32 ti_magic; void *ti_save; - /* This should never used. If this happens, - one of other filesystems has a bug. */ + /* + * This should never be used. If it happens, + * one of other filesystems has a bug. + */ unsigned short ti_flags; unsigned short ti_count; }; @@ -151,8 +157,10 @@ struct nilfs_transaction_info { /* ti_flags */ #define NILFS_TI_DYNAMIC_ALLOC 0x0001 /* Allocated from slab */ -#define NILFS_TI_SYNC 0x0002 /* Force to construct segment at the - end of transaction. */ +#define NILFS_TI_SYNC 0x0002 /* + * Force to construct segment at the + * end of transaction. + */ #define NILFS_TI_GC 0x0004 /* GC context */ #define NILFS_TI_COMMIT 0x0008 /* Change happened or not */ #define NILFS_TI_WRITER 0x0010 /* Constructor context */ diff --git a/fs/nilfs2/recovery.c b/fs/nilfs2/recovery.c index db156a199149..d893dc912b62 100644 --- a/fs/nilfs2/recovery.c +++ b/fs/nilfs2/recovery.c @@ -43,8 +43,10 @@ enum { /* work structure for recovery */ struct nilfs_recovery_block { - ino_t ino; /* Inode number of the file that this block - belongs to */ + ino_t ino; /* + * Inode number of the file that this block + * belongs to + */ sector_t blocknr; /* block number */ __u64 vblocknr; /* virtual block number */ unsigned long blkoff; /* File offset of the data block (per block) */ @@ -869,9 +871,11 @@ int nilfs_search_super_root(struct the_nilfs *nilfs, flags = le16_to_cpu(sum->ss_flags); if (!(flags & NILFS_SS_SR) && !scan_newer) { - /* This will never happen because a superblock - (last_segment) always points to a pseg - having a super root. */ + /* + * This will never happen because a superblock + * (last_segment) always points to a pseg with + * a super root. + */ ret = NILFS_SEG_FAIL_CONSISTENCY; goto failed; } diff --git a/fs/nilfs2/segment.c b/fs/nilfs2/segment.c index c9ee03c262ea..e78b68a81aec 100644 --- a/fs/nilfs2/segment.c +++ b/fs/nilfs2/segment.c @@ -45,18 +45,26 @@ */ #define SC_N_INODEVEC 16 /* Size of locally allocated inode vector */ -#define SC_MAX_SEGDELTA 64 /* Upper limit of the number of segments - appended in collection retry loop */ +#define SC_MAX_SEGDELTA 64 /* + * Upper limit of the number of segments + * appended in collection retry loop + */ /* Construction mode */ enum { SC_LSEG_SR = 1, /* Make a logical segment having a super root */ - SC_LSEG_DSYNC, /* Flush data blocks of a given file and make - a logical segment without a super root */ - SC_FLUSH_FILE, /* Flush data files, leads to segment writes without - creating a checkpoint */ - SC_FLUSH_DAT, /* Flush DAT file. This also creates segments without - a checkpoint */ + SC_LSEG_DSYNC, /* + * Flush data blocks of a given file and make + * a logical segment without a super root. + */ + SC_FLUSH_FILE, /* + * Flush data files, leads to segment writes without + * creating a checkpoint. + */ + SC_FLUSH_DAT, /* + * Flush DAT file. This also creates segments + * without a checkpoint. + */ }; /* Stage numbers of dirty block collection */ @@ -438,8 +446,10 @@ static int nilfs_segctor_feed_segment(struct nilfs_sc_info *sci) { sci->sc_nblk_this_inc += sci->sc_curseg->sb_sum.nblocks; if (NILFS_SEGBUF_IS_LAST(sci->sc_curseg, &sci->sc_segbufs)) - return -E2BIG; /* The current segment is filled up - (internal code) */ + return -E2BIG; /* + * The current segment is filled up + * (internal code) + */ sci->sc_curseg = NILFS_NEXT_SEGBUF(sci->sc_curseg); return nilfs_segctor_reset_segment_buffer(sci); } @@ -869,9 +879,11 @@ static int nilfs_segctor_create_checkpoint(struct nilfs_sc_info *sci) err = nilfs_cpfile_get_checkpoint(nilfs->ns_cpfile, nilfs->ns_cno, 1, &raw_cp, &bh_cp); if (likely(!err)) { - /* The following code is duplicated with cpfile. But, it is - needed to collect the checkpoint even if it was not newly - created */ + /* + * The following code is duplicated with cpfile. But, it is + * needed to collect the checkpoint even if it was not newly + * created. + */ mark_buffer_dirty(bh_cp); nilfs_mdt_mark_dirty(nilfs->ns_cpfile); nilfs_cpfile_put_checkpoint( @@ -1400,8 +1412,10 @@ static void nilfs_free_incomplete_logs(struct list_head *logs, if (atomic_read(&segbuf->sb_err)) { /* Case 1: The first segment failed */ if (segbuf->sb_pseg_start != segbuf->sb_fseg_start) - /* Case 1a: Partial segment appended into an existing - segment */ + /* + * Case 1a: Partial segment appended into an existing + * segment + */ nilfs_terminate_segment(nilfs, segbuf->sb_fseg_start, segbuf->sb_fseg_end); else /* Case 1b: New full segment */ @@ -1625,8 +1639,10 @@ static int nilfs_segctor_assign(struct nilfs_sc_info *sci, int mode) static void nilfs_begin_page_io(struct page *page) { if (!page || PageWriteback(page)) - /* For split b-tree node pages, this function may be called - twice. We ignore the 2nd or later calls by this check. */ + /* + * For split b-tree node pages, this function may be called + * twice. We ignore the 2nd or later calls by this check. + */ return; lock_page(page); @@ -2679,8 +2695,10 @@ static void nilfs_segctor_write_out(struct nilfs_sc_info *sci) { int ret, retrycount = NILFS_SC_CLEANUP_RETRY; - /* The segctord thread was stopped and its timer was removed. - But some tasks remain. */ + /* + * The segctord thread was stopped and its timer was removed. + * But some tasks remain. + */ do { struct nilfs_transaction_info ti; diff --git a/fs/nilfs2/segment.h b/fs/nilfs2/segment.h index 27822e760d3f..6565c10b7b76 100644 --- a/fs/nilfs2/segment.h +++ b/fs/nilfs2/segment.h @@ -189,11 +189,15 @@ enum { NILFS_SC_DIRTY, /* One or more dirty meta-data blocks exist */ NILFS_SC_UNCLOSED, /* Logical segment is not closed */ NILFS_SC_SUPER_ROOT, /* The latest segment has a super root */ - NILFS_SC_PRIOR_FLUSH, /* Requesting immediate flush without making a - checkpoint */ - NILFS_SC_HAVE_DELTA, /* Next checkpoint will have update of files - other than DAT, cpfile, sufile, or files - moved by GC */ + NILFS_SC_PRIOR_FLUSH, /* + * Requesting immediate flush without making a + * checkpoint + */ + NILFS_SC_HAVE_DELTA, /* + * Next checkpoint will have update of files + * other than DAT, cpfile, sufile, or files + * moved by GC. + */ }; /* sc_state */ @@ -203,17 +207,23 @@ enum { /* * Constant parameters */ -#define NILFS_SC_CLEANUP_RETRY 3 /* Retry count of construction when - destroying segctord */ +#define NILFS_SC_CLEANUP_RETRY 3 /* + * Retry count of construction when + * destroying segctord + */ /* * Default values of timeout, in seconds. */ -#define NILFS_SC_DEFAULT_TIMEOUT 5 /* Timeout value of dirty blocks. - It triggers construction of a - logical segment with a super root */ -#define NILFS_SC_DEFAULT_SR_FREQ 30 /* Maximum frequency of super root - creation */ +#define NILFS_SC_DEFAULT_TIMEOUT 5 /* + * Timeout value of dirty blocks. + * It triggers construction of a + * logical segment with a super root. + */ +#define NILFS_SC_DEFAULT_SR_FREQ 30 /* + * Maximum frequency of super root + * creation + */ /* * The default threshold amount of data, in block counts. diff --git a/fs/nilfs2/the_nilfs.c b/fs/nilfs2/the_nilfs.c index b9e19ca3c96e..809bd2de7ad0 100644 --- a/fs/nilfs2/the_nilfs.c +++ b/fs/nilfs2/the_nilfs.c @@ -617,8 +617,10 @@ int init_nilfs(struct the_nilfs *nilfs, struct super_block *sb, char *data) err = nilfs_load_super_block(nilfs, sb, blocksize, &sbp); if (err) goto out; - /* not failed_sbh; sbh is released automatically - when reloading fails. */ + /* + * Not to failed_sbh; sbh is released automatically + * when reloading fails. + */ } nilfs->ns_blocksize_bits = sb->s_blocksize_bits; nilfs->ns_blocksize = blocksize; diff --git a/include/linux/nilfs2_fs.h b/include/linux/nilfs2_fs.h index 3b584925d0e8..5988dd57ba66 100644 --- a/include/linux/nilfs2_fs.h +++ b/include/linux/nilfs2_fs.h @@ -127,10 +127,14 @@ struct nilfs_super_root { #define NILFS_MOUNT_ERRORS_RO 0x0020 /* Remount fs ro on errors */ #define NILFS_MOUNT_ERRORS_PANIC 0x0040 /* Panic on errors */ #define NILFS_MOUNT_BARRIER 0x1000 /* Use block barriers */ -#define NILFS_MOUNT_STRICT_ORDER 0x2000 /* Apply strict in-order - semantics also for data */ -#define NILFS_MOUNT_NORECOVERY 0x4000 /* Disable write access during - mount-time recovery */ +#define NILFS_MOUNT_STRICT_ORDER 0x2000 /* + * Apply strict in-order + * semantics also for data + */ +#define NILFS_MOUNT_NORECOVERY 0x4000 /* + * Disable write access during + * mount-time recovery + */ #define NILFS_MOUNT_DISCARD 0x8000 /* Issue DISCARD requests */ @@ -142,16 +146,20 @@ struct nilfs_super_block { __le16 s_minor_rev_level; /* minor revision level */ __le16 s_magic; /* Magic signature */ - __le16 s_bytes; /* Bytes count of CRC calculation - for this structure. s_reserved - is excluded. */ + __le16 s_bytes; /* + * Bytes count of CRC calculation + * for this structure. s_reserved + * is excluded. + */ __le16 s_flags; /* flags */ __le32 s_crc_seed; /* Seed value of CRC calculation */ /*10*/ __le32 s_sum; /* Check sum of super block */ - __le32 s_log_block_size; /* Block size represented as follows - blocksize = - 1 << (s_log_block_size + 10) */ + __le32 s_log_block_size; /* + * Block size represented as follows + * blocksize = + * 1 << (s_log_block_size + 10) + */ __le64 s_nsegments; /* Number of segments in filesystem */ /*20*/ __le64 s_dev_size; /* block device size in bytes */ __le64 s_first_data_block; /* 1st seg disk block number */ @@ -163,8 +171,10 @@ struct nilfs_super_block { __le64 s_last_seq; /* seq. number of seg written last */ /*50*/ __le64 s_free_blocks_count; /* Free blocks count */ - __le64 s_ctime; /* Creation time (execution time of - newfs) */ + __le64 s_ctime; /* + * Creation time (execution time of + * newfs) + */ /*60*/ __le64 s_mtime; /* Mount time */ __le64 s_wtime; /* Write time */ /*70*/ __le16 s_mnt_count; /* Mount count */ @@ -188,8 +198,10 @@ struct nilfs_super_block { /*A8*/ char s_volume_name[80]; /* volume name */ /*F8*/ __le32 s_c_interval; /* Commit interval of segment */ - __le32 s_c_block_max; /* Threshold of data amount for - the segment construction */ + __le32 s_c_block_max; /* + * Threshold of data amount for + * the segment construction + */ /*100*/ __le64 s_feature_compat; /* Compatible feature set */ __le64 s_feature_compat_ro; /* Read-only compatible feature set */ __le64 s_feature_incompat; /* Incompatible feature set */ @@ -242,12 +254,18 @@ struct nilfs_super_block { #define NILFS_SB_OFFSET_BYTES 1024 /* byte offset of nilfs superblock */ -#define NILFS_SEG_MIN_BLOCKS 16 /* Minimum number of blocks in - a full segment */ -#define NILFS_PSEG_MIN_BLOCKS 2 /* Minimum number of blocks in - a partial segment */ -#define NILFS_MIN_NRSVSEGS 8 /* Minimum number of reserved - segments */ +#define NILFS_SEG_MIN_BLOCKS 16 /* + * Minimum number of blocks in + * a full segment + */ +#define NILFS_PSEG_MIN_BLOCKS 2 /* + * Minimum number of blocks in + * a partial segment + */ +#define NILFS_MIN_NRSVSEGS 8 /* + * Minimum number of reserved + * segments + */ /* * We call DAT, cpfile, and sufile root metadata files. Inodes of @@ -513,9 +531,11 @@ struct nilfs_checkpoint { __le64 cp_inodes_count; __le64 cp_blocks_count; - /* Do not change the byte offset of ifile inode. - To keep the compatibility of the disk format, - additional fields should be added behind cp_ifile_inode. */ + /* + * Do not change the byte offset of ifile inode. + * To keep the compatibility of the disk format, + * additional fields should be added behind cp_ifile_inode. + */ struct nilfs_inode cp_ifile_inode; }; -- cgit v1.3-8-gc7d7 From 9eb8a659dea694b0dcbd6287f6b1fbdc523b80bc Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Mon, 23 May 2016 16:24:08 -0700 Subject: exec: remove the no longer needed remove_arg_zero()->free_arg_page() remove_arg_zero() does free_arg_page() for no reason. This was needed before and only if CONFIG_MMU=y: see commit 4fc75ff4816c ("exec: fix remove_arg_zero"), install_arg_page() was called for every page != NULL in bprm->page[] array. Today install_arg_page() has already gone and free_arg_page() is nop after another commit b6a2fea39318 ("mm: variable length argument support"). CONFIG_MMU=n does free_arg_pages() in free_bprm() and thus it doesn't need remove_arg_zero()->free_arg_page() too; apart from get_arg_page() it never checks if the page in bprm->page[] was allocated or not, so the "extra" non-freed page is fine. OTOH, this free_arg_page() can add the minor pessimization, the caller is going to do copy_strings_kernel() right after remove_arg_zero() which will likely need to re-allocate the same page again. And as Hujunjie pointed out, the "offset == PAGE_SIZE" check is wrong because we are going to increment bprm->p once again before return, so CONFIG_MMU=n "leaks" the page anyway if '0' is the final byte in this page. NOTE: remove_arg_zero() assumes that argv[0] is null-terminated but this is not necessarily true. copy_strings() does "len = strnlen_user(...)", then copy_from_user(len) but another thread or debuger can overwrite the trailing '0' in between. Afaics nothing really bad can happen because we must always have the null-terminated bprm->filename copied by the 1st copy_strings_kernel(), but perhaps we should change this code to check "bprm->p < bprm->exec" anyway, and/or change copy_strings() to ensure that the last byte in string is always zero. Link: http://lkml.kernel.org/r/20160517155335.GA31435@redhat.com Signed-off-by: Oleg Nesterov Reported by: hujunjie Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/exec.c | 7 ------- 1 file changed, 7 deletions(-) (limited to 'fs') diff --git a/fs/exec.c b/fs/exec.c index e92419fd78b3..4f079bd083af 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -243,10 +243,6 @@ static void put_arg_page(struct page *page) put_page(page); } -static void free_arg_page(struct linux_binprm *bprm, int i) -{ -} - static void free_arg_pages(struct linux_binprm *bprm) { } @@ -1499,9 +1495,6 @@ int remove_arg_zero(struct linux_binprm *bprm) kunmap_atomic(kaddr); put_arg_page(page); - - if (offset == PAGE_SIZE) - free_arg_page(bprm, (bprm->p >> PAGE_SHIFT) - 1); } while (offset == PAGE_SIZE); bprm->p++; -- cgit v1.3-8-gc7d7 From 864778b154233d3a673c520aa9606777589ecee7 Mon Sep 17 00:00:00 2001 From: Michal Hocko Date: Mon, 23 May 2016 16:25:36 -0700 Subject: mm, aout: handle vm_brk failures vm_brk is allowed to fail but load_aout_binary simply ignores the error and happily continues. I haven't noticed any problem from that in real life but later patches will make the failure more likely because vm_brk will become killable (resp. mmap_sem for write waiting will become killable) so we should be more careful now. The error handling should be quite straightforward because there are calls to vm_mmap which check the error properly already. The only notable exception is set_brk which is called after beyond_if label. But nothing indicates that we cannot move it above set_binfmt as the two do not depend on each other and fail before we do set_binfmt and alter reference counting. Signed-off-by: Michal Hocko Acked-by: Vlastimil Babka Cc: Thomas Gleixner Cc: Ingo Molnar Cc: "H. Peter Anvin" Cc: Alexander Viro Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/x86/ia32/ia32_aout.c | 22 +++++++++++++++------- fs/binfmt_aout.c | 11 ++++++++--- 2 files changed, 23 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/arch/x86/ia32/ia32_aout.c b/arch/x86/ia32/ia32_aout.c index ae6aad1d24f7..f5e737ff0022 100644 --- a/arch/x86/ia32/ia32_aout.c +++ b/arch/x86/ia32/ia32_aout.c @@ -116,13 +116,13 @@ static struct linux_binfmt aout_format = { .min_coredump = PAGE_SIZE }; -static void set_brk(unsigned long start, unsigned long end) +static unsigned long set_brk(unsigned long start, unsigned long end) { start = PAGE_ALIGN(start); end = PAGE_ALIGN(end); if (end <= start) - return; - vm_brk(start, end - start); + return start; + return vm_brk(start, end - start); } #ifdef CONFIG_COREDUMP @@ -349,7 +349,10 @@ static int load_aout_binary(struct linux_binprm *bprm) #endif if (!bprm->file->f_op->mmap || (fd_offset & ~PAGE_MASK) != 0) { - vm_brk(N_TXTADDR(ex), ex.a_text+ex.a_data); + error = vm_brk(N_TXTADDR(ex), ex.a_text+ex.a_data); + if (IS_ERR_VALUE(error)) + return error; + read_code(bprm->file, N_TXTADDR(ex), fd_offset, ex.a_text+ex.a_data); goto beyond_if; @@ -372,10 +375,13 @@ static int load_aout_binary(struct linux_binprm *bprm) if (error != N_DATADDR(ex)) return error; } + beyond_if: - set_binfmt(&aout_format); + error = set_brk(current->mm->start_brk, current->mm->brk); + if (IS_ERR_VALUE(error)) + return error; - set_brk(current->mm->start_brk, current->mm->brk); + set_binfmt(&aout_format); current->mm->start_stack = (unsigned long)create_aout_tables((char __user *)bprm->p, bprm); @@ -434,7 +440,9 @@ static int load_aout_library(struct file *file) error_time = jiffies; } #endif - vm_brk(start_addr, ex.a_text + ex.a_data + ex.a_bss); + retval = vm_brk(start_addr, ex.a_text + ex.a_data + ex.a_bss); + if (IS_ERR_VALUE(retval)) + goto out; read_code(file, start_addr, N_TXTOFF(ex), ex.a_text + ex.a_data); diff --git a/fs/binfmt_aout.c b/fs/binfmt_aout.c index 4c556680fa74..2fab9f130e51 100644 --- a/fs/binfmt_aout.c +++ b/fs/binfmt_aout.c @@ -297,7 +297,10 @@ static int load_aout_binary(struct linux_binprm * bprm) } if (!bprm->file->f_op->mmap||((fd_offset & ~PAGE_MASK) != 0)) { - vm_brk(N_TXTADDR(ex), ex.a_text+ex.a_data); + error = vm_brk(N_TXTADDR(ex), ex.a_text+ex.a_data); + if (IS_ERR_VALUE(error)) + return error; + read_code(bprm->file, N_TXTADDR(ex), fd_offset, ex.a_text + ex.a_data); goto beyond_if; @@ -378,8 +381,10 @@ static int load_aout_library(struct file *file) "N_TXTOFF is not page aligned. Please convert library: %pD\n", file); } - vm_brk(start_addr, ex.a_text + ex.a_data + ex.a_bss); - + retval = vm_brk(start_addr, ex.a_text + ex.a_data + ex.a_bss); + if (IS_ERR_VALUE(retval)) + goto out; + read_code(file, start_addr, N_TXTOFF(ex), ex.a_text + ex.a_data); retval = 0; -- cgit v1.3-8-gc7d7 From ecc2bc8ac03884266cf73f8a2a42b911465b2fbc Mon Sep 17 00:00:00 2001 From: Michal Hocko Date: Mon, 23 May 2016 16:25:39 -0700 Subject: mm, elf: handle vm_brk error load_elf_library doesn't handle vm_brk failure although nothing really indicates it cannot do that because the function is allowed to fail due to vm_mmap failures already. This might be not a problem now but later patch will make vm_brk killable (resp. mmap_sem for write waiting will become killable) and so the failure will be more probable. Signed-off-by: Michal Hocko Acked-by: Vlastimil Babka Cc: Alexander Viro Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/binfmt_elf.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/binfmt_elf.c b/fs/binfmt_elf.c index 56224ffa94d2..938fc4ede764 100644 --- a/fs/binfmt_elf.c +++ b/fs/binfmt_elf.c @@ -1176,8 +1176,11 @@ static int load_elf_library(struct file *file) len = ELF_PAGESTART(eppnt->p_filesz + eppnt->p_vaddr + ELF_MIN_ALIGN - 1); bss = eppnt->p_memsz + eppnt->p_vaddr; - if (bss > len) - vm_brk(len, bss - len); + if (bss > len) { + error = vm_brk(len, bss - len); + if (BAD_ADDR(error)) + goto out_free_ph; + } error = 0; out_free_ph: -- cgit v1.3-8-gc7d7 From 4e80153a6044bcd9a7d65c1ec4b1d1c44ba0ed6c Mon Sep 17 00:00:00 2001 From: Michal Hocko Date: Mon, 23 May 2016 16:25:45 -0700 Subject: mm, proc: make clear_refs killable CLEAR_REFS_MM_HIWATER_RSS and CLEAR_REFS_SOFT_DIRTY are relying on mmap_sem for write. If the waiting task gets killed by the oom killer and it would operate on the current's mm it would block oom_reaper from asynchronous address space reclaim and reduce the chances of timely OOM resolving. Wait for the lock in the killable mode and return with EINTR if the task got killed while waiting. This will also expedite the return to the userspace and do_exit even if the mm is remote. Signed-off-by: Michal Hocko Acked-by: Oleg Nesterov Acked-by: Vlastimil Babka Cc: Petr Cermak Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/proc/task_mmu.c | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/proc/task_mmu.c b/fs/proc/task_mmu.c index 541583510cfb..4648c7f63ae2 100644 --- a/fs/proc/task_mmu.c +++ b/fs/proc/task_mmu.c @@ -1027,11 +1027,15 @@ static ssize_t clear_refs_write(struct file *file, const char __user *buf, }; if (type == CLEAR_REFS_MM_HIWATER_RSS) { + if (down_write_killable(&mm->mmap_sem)) { + count = -EINTR; + goto out_mm; + } + /* * Writing 5 to /proc/pid/clear_refs resets the peak * resident set size to this mm's current rss value. */ - down_write(&mm->mmap_sem); reset_mm_hiwater_rss(mm); up_write(&mm->mmap_sem); goto out_mm; @@ -1043,7 +1047,10 @@ static ssize_t clear_refs_write(struct file *file, const char __user *buf, if (!(vma->vm_flags & VM_SOFTDIRTY)) continue; up_read(&mm->mmap_sem); - down_write(&mm->mmap_sem); + if (down_write_killable(&mm->mmap_sem)) { + count = -EINTR; + goto out_mm; + } for (vma = mm->mmap; vma; vma = vma->vm_next) { vma->vm_flags &= ~VM_SOFTDIRTY; vma_set_page_prot(vma); -- cgit v1.3-8-gc7d7 From 4136c26b65d12d0c5ef58b7d30dc0aa2a11ed491 Mon Sep 17 00:00:00 2001 From: Michal Hocko Date: Mon, 23 May 2016 16:25:57 -0700 Subject: coredump: make coredump_wait wait for mmap_sem for write killable coredump_wait waits for mmap_sem for write currently which can prevent oom_reaper to reclaim the oom victims address space asynchronously because that requires mmap_sem for read. This might happen if the oom victim is multi threaded and some thread(s) is holding mmap_sem for read (e.g. page fault) and it is stuck in the page allocator while other thread(s) reached coredump_wait already. This patch simply uses down_write_killable and bails out with EINTR if the lock got interrupted by the fatal signal. do_coredump will return right away and do_group_exit will take care to zap the whole thread group. Signed-off-by: Michal Hocko Acked-by: Oleg Nesterov Acked-by: Vlastimil Babka Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/coredump.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/coredump.c b/fs/coredump.c index 492c2db25dc9..38a7ab87e10a 100644 --- a/fs/coredump.c +++ b/fs/coredump.c @@ -413,7 +413,9 @@ static int coredump_wait(int exit_code, struct core_state *core_state) core_state->dumper.task = tsk; core_state->dumper.next = NULL; - down_write(&mm->mmap_sem); + if (down_write_killable(&mm->mmap_sem)) + return -EINTR; + if (!mm->core_state) core_waiters = zap_threads(tsk, mm, core_state, exit_code); up_write(&mm->mmap_sem); -- cgit v1.3-8-gc7d7 From 013373e8b86735086ac86aff782637f7e58ff84d Mon Sep 17 00:00:00 2001 From: Michal Hocko Date: Mon, 23 May 2016 16:25:59 -0700 Subject: aio: make aio_setup_ring killable aio_setup_ring waits for mmap_sem in writable mode. If the waiting task gets killed by the oom killer it would block oom_reaper from asynchronous address space reclaim and reduce the chances of timely OOM resolving. Wait for the lock in the killable mode and return with EINTR if the task got killed while waiting. This will also expedite the return to the userspace and do_exit. Signed-off-by: Michal Hocko Acked-by: Jeff Moyer Acked-by: Vlastimil Babka Cc: Benamin LaHaise Cc: Alexander Viro Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/aio.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/aio.c b/fs/aio.c index a6deaa78326d..fb8e45b88cd4 100644 --- a/fs/aio.c +++ b/fs/aio.c @@ -496,7 +496,12 @@ static int aio_setup_ring(struct kioctx *ctx) ctx->mmap_size = nr_pages * PAGE_SIZE; pr_debug("attempting mmap of %lu bytes\n", ctx->mmap_size); - down_write(&mm->mmap_sem); + if (down_write_killable(&mm->mmap_sem)) { + ctx->mmap_size = 0; + aio_free_ring(ctx); + return -EINTR; + } + ctx->mmap_base = do_mmap_pgoff(ctx->aio_ring_file, 0, ctx->mmap_size, PROT_READ | PROT_WRITE, MAP_SHARED, 0, &unused); -- cgit v1.3-8-gc7d7 From f268dfe905d4682150d4acbb25f59adc04cd9398 Mon Sep 17 00:00:00 2001 From: Michal Hocko Date: Mon, 23 May 2016 16:26:02 -0700 Subject: exec: make exec path waiting for mmap_sem killable setup_arg_pages requires mmap_sem for write. If the waiting task gets killed by the oom killer it would block oom_reaper from asynchronous address space reclaim and reduce the chances of timely OOM resolving. Wait for the lock in the killable mode and return with EINTR if the task got killed while waiting. All the callers are already handling error path and the fatal signal doesn't need any additional treatment. The same applies to __bprm_mm_init. Signed-off-by: Michal Hocko Acked-by: Oleg Nesterov Acked-by: Vlastimil Babka Cc: Alexander Viro Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/exec.c | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/exec.c b/fs/exec.c index 4f079bd083af..887c1c955df8 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -263,7 +263,10 @@ static int __bprm_mm_init(struct linux_binprm *bprm) if (!vma) return -ENOMEM; - down_write(&mm->mmap_sem); + if (down_write_killable(&mm->mmap_sem)) { + err = -EINTR; + goto err_free; + } vma->vm_mm = mm; /* @@ -290,6 +293,7 @@ static int __bprm_mm_init(struct linux_binprm *bprm) return 0; err: up_write(&mm->mmap_sem); +err_free: bprm->vma = NULL; kmem_cache_free(vm_area_cachep, vma); return err; @@ -696,7 +700,9 @@ int setup_arg_pages(struct linux_binprm *bprm, bprm->loader -= stack_shift; bprm->exec -= stack_shift; - down_write(&mm->mmap_sem); + if (down_write_killable(&mm->mmap_sem)) + return -EINTR; + vm_flags = VM_STACK_FLAGS; /* -- cgit v1.3-8-gc7d7 From 1112018cefc586cba924770a2b28bff6d2f0aa5c Mon Sep 17 00:00:00 2001 From: Andreas Gruenbacher Date: Mon, 16 May 2016 21:46:30 +0200 Subject: ubifs: ubifs_dump_inode: Fix dumping field bulk_read The wrong field (xattr) is dumped here due to a copy-and-paste error. Signed-off-by: Andreas Gruenbacher Signed-off-by: Richard Weinberger --- fs/ubifs/debug.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ubifs/debug.c b/fs/ubifs/debug.c index 595ca0debe11..69e287e20732 100644 --- a/fs/ubifs/debug.c +++ b/fs/ubifs/debug.c @@ -260,7 +260,7 @@ void ubifs_dump_inode(struct ubifs_info *c, const struct inode *inode) pr_err("\txattr_names %u\n", ui->xattr_names); pr_err("\tdirty %u\n", ui->dirty); pr_err("\txattr %u\n", ui->xattr); - pr_err("\tbulk_read %u\n", ui->xattr); + pr_err("\tbulk_read %u\n", ui->bulk_read); pr_err("\tsynced_i_size %llu\n", (unsigned long long)ui->synced_i_size); pr_err("\tui_size %llu\n", -- cgit v1.3-8-gc7d7 From b9d8905e4a751e2cdc0fb474856b7183c594dcc6 Mon Sep 17 00:00:00 2001 From: Mikulas Patocka Date: Tue, 24 May 2016 22:51:27 +0200 Subject: reiserfs: check kstrdup failure Check out-of-memory failure of the kstrdup option. Note that the argument "arg" may be NULL (in that case kstrup returns NULL), so out of memory condition happened if arg was non-NULL and kstrdup returned NULL. The patch also changes the call to replace_mount_options - if we didn't pass any filesystem-specific options, we don't call replace_mount_options (thus we don't erase existing reported options). Note that to properly report options after remount, the reiserfs filesystem should implement the show_options method. Without the show_options method, options changed with remount replace existing options. Signed-off-by: Mikulas Patocka Signed-off-by: Jan Kara --- fs/reiserfs/super.c | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/reiserfs/super.c b/fs/reiserfs/super.c index b8f2d1e8c645..c72c16c5a60f 100644 --- a/fs/reiserfs/super.c +++ b/fs/reiserfs/super.c @@ -1393,7 +1393,7 @@ static int reiserfs_remount(struct super_block *s, int *mount_flags, char *arg) unsigned long safe_mask = 0; unsigned int commit_max_age = (unsigned int)-1; struct reiserfs_journal *journal = SB_JOURNAL(s); - char *new_opts = kstrdup(arg, GFP_KERNEL); + char *new_opts; int err; char *qf_names[REISERFS_MAXQUOTAS]; unsigned int qfmt = 0; @@ -1401,6 +1401,10 @@ static int reiserfs_remount(struct super_block *s, int *mount_flags, char *arg) int i; #endif + new_opts = kstrdup(arg, GFP_KERNEL); + if (arg && !new_opts) + return -ENOMEM; + sync_filesystem(s); reiserfs_write_lock(s); @@ -1546,7 +1550,8 @@ static int reiserfs_remount(struct super_block *s, int *mount_flags, char *arg) } out_ok_unlocked: - replace_mount_options(s, new_opts); + if (new_opts) + replace_mount_options(s, new_opts); return 0; out_err_unlock: -- cgit v1.3-8-gc7d7 From 2997bfd0492d33d230b1475909a0211157a734bc Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Mon, 23 May 2016 13:21:01 +0300 Subject: NFS: checking for NULL instead of IS_ERR() in nfs_commit_file() nfs_create_request() doesn't return NULL, it returns error pointers. Fixes: 67911c8f18b5 ('NFS: Add nfs_commit_file()') Signed-off-by: Dan Carpenter Signed-off-by: Anna Schumaker --- fs/nfs/write.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/nfs/write.c b/fs/nfs/write.c index 4dac51ba1f7e..54a4eb6bc462 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -1735,8 +1735,8 @@ int nfs_commit_file(struct file *file, struct nfs_write_verifier *verf) open = get_nfs_open_context(nfs_file_open_context(file)); req = nfs_create_request(open, NULL, NULL, 0, i_size_read(inode)); - if (!req) { - ret = -ENOMEM; + if (IS_ERR(req)) { + ret = PTR_ERR(req); goto out_put; } -- cgit v1.3-8-gc7d7 From ade8febde0271513360bac44883dbebad44276c3 Mon Sep 17 00:00:00 2001 From: Weston Andros Adamson Date: Wed, 25 May 2016 10:07:23 -0400 Subject: nfs: avoid race that crashes nfs_init_commit Since the patch "NFS: Allow multiple commit requests in flight per file" we can run multiple simultaneous commits on the same inode. This introduced a race over collecting pages to commit that made it possible to call nfs_init_commit() with an empty list - which causes crashes like the one below. The fix is to catch this race and avoid calling nfs_init_commit and initiate_commit when there is no work to do. Here is the crash: [600522.076832] BUG: unable to handle kernel NULL pointer dereference at 0000000000000040 [600522.078475] IP: [] nfs_init_commit+0x22/0x130 [nfs] [600522.078745] PGD 4272b1067 PUD 4272cb067 PMD 0 [600522.078972] Oops: 0000 [#1] SMP [600522.079204] Modules linked in: nfsv3 nfs_layout_flexfiles rpcsec_gss_krb5 nfsv4 dns_resolver nfs fscache dcdbas ip6t_rpfilter ip6t_REJECT nf_reject_ipv6 xt_conntrack ebtable_nat ebtable_broute bridge stp llc ebtable_filter ebtables ip6table_nat nf_conntrack_ipv6 nf_defrag_ipv6 nf_nat_ipv6 ip6table_mangle ip6table_security ip6table_raw ip6table_filter ip6_tables iptable_nat nf_conntrack_ipv4 nf_defrag_ipv4 nf_nat_ipv4 nf_nat nf_conntrack iptable_mangle iptable_security iptable_raw vmw_vsock_vmci_transport vsock bonding ipmi_devintf ipmi_msghandler coretemp crct10dif_pclmul crc32_pclmul ghash_clmulni_intel ppdev vmw_balloon parport_pc parport acpi_cpufreq vmw_vmci i2c_piix4 shpchp nfsd auth_rpcgss nfs_acl lockd grace sunrpc xfs libcrc32c vmwgfx drm_kms_helper ttm drm crc32c_intel serio_raw vmxnet3 [600522.081380] vmw_pvscsi ata_generic pata_acpi [600522.081809] CPU: 3 PID: 15667 Comm: /usr/bin/python Not tainted 4.1.9-100.pd.88.el7.x86_64 #1 [600522.082281] Hardware name: VMware, Inc. VMware Virtual Platform/440BX Desktop Reference Platform, BIOS 6.00 09/30/2014 [600522.082814] task: ffff8800bbbfa780 ti: ffff88042ae84000 task.ti: ffff88042ae84000 [600522.083378] RIP: 0010:[] [] nfs_init_commit+0x22/0x130 [nfs] [600522.083973] RSP: 0018:ffff88042ae87438 EFLAGS: 00010246 [600522.084571] RAX: 0000000000000000 RBX: ffff880003485e40 RCX: ffff88042ae87588 [600522.085188] RDX: 0000000000000000 RSI: ffff88042ae874b0 RDI: ffff880003485e40 [600522.085756] RBP: ffff88042ae87448 R08: ffff880003486010 R09: ffff88042ae874b0 [600522.086332] R10: 0000000000000000 R11: 0000000000000005 R12: ffff88042ae872d0 [600522.086905] R13: ffff88042ae874b0 R14: ffff880003485e40 R15: ffff88042704c840 [600522.087484] FS: 00007f4728ff2740(0000) GS:ffff88043fd80000(0000) knlGS:0000000000000000 [600522.088070] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b [600522.088663] CR2: 0000000000000040 CR3: 000000042b6aa000 CR4: 00000000001406e0 [600522.089327] Stack: [600522.089926] 0000000000000001 ffff88042ae87588 ffff88042ae874f8 ffffffffa04f09fa [600522.090549] 0000000000017840 0000000000017840 ffff88042ae87588 ffff8803258d9930 [600522.091169] ffff88042ae87578 ffffffffa0563d80 0000000000000000 ffff88042704c840 [600522.091789] Call Trace: [600522.092420] [] pnfs_generic_commit_pagelist+0x1da/0x320 [nfsv4] [600522.093052] [] ? ff_layout_commit_prepare_v3+0x30/0x30 [nfs_layout_flexfiles] [600522.093696] [] ff_layout_commit_pagelist+0x15/0x20 [nfs_layout_flexfiles] [600522.094359] [] nfs_generic_commit_list+0xe8/0x120 [nfs] [600522.095032] [] nfs_commit_inode+0xba/0x110 [nfs] [600522.095719] [] nfs_release_page+0x44/0xd0 [nfs] [600522.096410] [] try_to_release_page+0x32/0x50 [600522.097109] [] shrink_page_list+0x961/0xb30 [600522.097812] [] shrink_inactive_list+0x1cd/0x550 [600522.098530] [] shrink_lruvec+0x635/0x840 [600522.099250] [] shrink_zone+0xf0/0x2f0 [600522.099974] [] do_try_to_free_pages+0x192/0x470 [600522.100709] [] try_to_free_pages+0xda/0x170 [600522.101464] [] __alloc_pages_nodemask+0x588/0x970 [600522.102235] [] alloc_pages_vma+0xb5/0x230 [600522.103000] [] ? cpumask_any_but+0x39/0x50 [600522.103774] [] wp_page_copy.isra.55+0x95/0x490 [600522.104558] [] ? __wake_up+0x48/0x60 [600522.105357] [] do_wp_page+0xab/0x4f0 [600522.106137] [] ? release_task+0x36b/0x470 [600522.106902] [] ? eventfd_ctx_read+0x67/0x1c0 [600522.107659] [] handle_mm_fault+0xc78/0x1900 [600522.108431] [] __do_page_fault+0x181/0x420 [600522.109173] [] ? __audit_syscall_exit+0x1e6/0x280 [600522.109893] [] do_page_fault+0x30/0x80 [600522.110594] [] ? syscall_trace_leave+0xc6/0x120 [600522.111288] [] page_fault+0x28/0x30 [600522.111947] Code: 5d c3 0f 1f 80 00 00 00 00 0f 1f 44 00 00 55 4c 8d 87 d0 01 00 00 48 89 e5 53 48 89 fb 48 83 ec 08 4c 8b 0e 49 8b 41 18 4c 39 ce <48> 8b 40 40 4c 8b 50 30 74 24 48 8b 87 d0 01 00 00 48 8b 7e 08 [600522.113343] RIP [] nfs_init_commit+0x22/0x130 [nfs] [600522.114003] RSP [600522.114636] CR2: 0000000000000040 Fixes: af7cf057 (NFS: Allow multiple commit requests in flight per file) CC: stable@vger.kernel.org Signed-off-by: Weston Andros Adamson Signed-off-by: Anna Schumaker --- fs/nfs/pnfs_nfs.c | 28 ++++++++++++++++++++++++++++ fs/nfs/write.c | 4 ++++ 2 files changed, 32 insertions(+) (limited to 'fs') diff --git a/fs/nfs/pnfs_nfs.c b/fs/nfs/pnfs_nfs.c index d2a7c9f7aa94..0dfc476da3e1 100644 --- a/fs/nfs/pnfs_nfs.c +++ b/fs/nfs/pnfs_nfs.c @@ -246,6 +246,23 @@ void pnfs_fetch_commit_bucket_list(struct list_head *pages, } +/* Helper function for pnfs_generic_commit_pagelist to catch an empty + * page list. This can happen when two commits race. */ +static bool +pnfs_generic_commit_cancel_empty_pagelist(struct list_head *pages, + struct nfs_commit_data *data, + struct nfs_commit_info *cinfo) +{ + if (list_empty(pages)) { + if (atomic_dec_and_test(&cinfo->mds->rpcs_out)) + wake_up_atomic_t(&cinfo->mds->rpcs_out); + nfs_commitdata_release(data); + return true; + } + + return false; +} + /* This follows nfs_commit_list pretty closely */ int pnfs_generic_commit_pagelist(struct inode *inode, struct list_head *mds_pages, @@ -280,6 +297,11 @@ pnfs_generic_commit_pagelist(struct inode *inode, struct list_head *mds_pages, list_for_each_entry_safe(data, tmp, &list, pages) { list_del_init(&data->pages); if (data->ds_commit_index < 0) { + /* another commit raced with us */ + if (pnfs_generic_commit_cancel_empty_pagelist(mds_pages, + data, cinfo)) + continue; + nfs_init_commit(data, mds_pages, NULL, cinfo); nfs_initiate_commit(NFS_CLIENT(inode), data, NFS_PROTO(data->inode), @@ -288,6 +310,12 @@ pnfs_generic_commit_pagelist(struct inode *inode, struct list_head *mds_pages, LIST_HEAD(pages); pnfs_fetch_commit_bucket_list(&pages, data, cinfo); + + /* another commit raced with us */ + if (pnfs_generic_commit_cancel_empty_pagelist(&pages, + data, cinfo)) + continue; + nfs_init_commit(data, &pages, data->lseg, cinfo); initiate_commit(data, how); } diff --git a/fs/nfs/write.c b/fs/nfs/write.c index 54a4eb6bc462..e1c74d3db64d 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -1710,6 +1710,10 @@ nfs_commit_list(struct inode *inode, struct list_head *head, int how, { struct nfs_commit_data *data; + /* another commit raced with us */ + if (list_empty(head)) + return 0; + data = nfs_commitdata_alloc(); if (!data) -- cgit v1.3-8-gc7d7 From fb1084e3320555eca0fb86c9eaad5429ee761282 Mon Sep 17 00:00:00 2001 From: Tom Haynes Date: Wed, 25 May 2016 07:31:12 -0700 Subject: nfs/flexfiles: Helper function to detect FF_FLAGS_NO_READ_IO The mds can inform the client not to use the IOMODE_RW layout segment for doing READs. I.e., it is basically a IOMODE_WRITE layout segment. It would do this to not interfere with the WRITEs. Signed-off-by: Tom Haynes Signed-off-by: Anna Schumaker --- fs/nfs/flexfilelayout/flexfilelayout.h | 11 ++++++++++- fs/nfs/flexfilelayout/flexfilelayoutdev.c | 6 ++++++ 2 files changed, 16 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/flexfilelayout/flexfilelayout.h b/fs/nfs/flexfilelayout/flexfilelayout.h index b54058122647..1bcdb15d0c41 100644 --- a/fs/nfs/flexfilelayout/flexfilelayout.h +++ b/fs/nfs/flexfilelayout/flexfilelayout.h @@ -10,7 +10,8 @@ #define FS_NFS_NFS4FLEXFILELAYOUT_H #define FF_FLAGS_NO_LAYOUTCOMMIT 1 -#define FF_FLAGS_NO_IO_THRU_MDS 2 +#define FF_FLAGS_NO_IO_THRU_MDS 2 +#define FF_FLAGS_NO_READ_IO 4 #include "../pnfs.h" @@ -152,6 +153,12 @@ ff_layout_no_fallback_to_mds(struct pnfs_layout_segment *lseg) return FF_LAYOUT_LSEG(lseg)->flags & FF_FLAGS_NO_IO_THRU_MDS; } +static inline bool +ff_layout_no_read_on_rw(struct pnfs_layout_segment *lseg) +{ + return FF_LAYOUT_LSEG(lseg)->flags & FF_FLAGS_NO_READ_IO; +} + static inline bool ff_layout_test_devid_unavailable(struct nfs4_deviceid_node *node) { @@ -192,4 +199,6 @@ struct rpc_cred *ff_layout_get_ds_cred(struct pnfs_layout_segment *lseg, u32 ds_idx, struct rpc_cred *mdscred); bool ff_layout_has_available_ds(struct pnfs_layout_segment *lseg); bool ff_layout_avoid_mds_available_ds(struct pnfs_layout_segment *lseg); +bool ff_layout_avoid_read_on_rw(struct pnfs_layout_segment *lseg); + #endif /* FS_NFS_NFS4FLEXFILELAYOUT_H */ diff --git a/fs/nfs/flexfilelayout/flexfilelayoutdev.c b/fs/nfs/flexfilelayout/flexfilelayoutdev.c index 35d84d0c0f21..0aa36be71fce 100644 --- a/fs/nfs/flexfilelayout/flexfilelayoutdev.c +++ b/fs/nfs/flexfilelayout/flexfilelayoutdev.c @@ -557,6 +557,12 @@ bool ff_layout_avoid_mds_available_ds(struct pnfs_layout_segment *lseg) ff_layout_has_available_ds(lseg); } +bool ff_layout_avoid_read_on_rw(struct pnfs_layout_segment *lseg) +{ + return lseg->pls_range.iomode == IOMODE_RW && + ff_layout_no_read_on_rw(lseg); +} + module_param(dataserver_retrans, uint, 0644); MODULE_PARM_DESC(dataserver_retrans, "The number of times the NFSv4.1 client " "retries a request before it attempts further " -- cgit v1.3-8-gc7d7 From 1c8b5b6e8b570a8038fa42cf9e7c23782bd4882c Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Fri, 13 May 2016 17:06:59 -0700 Subject: Btrfs: free sys_array eb as soon as possible While reading sys_chunk_array in superblock, btrfs creates a temporary extent buffer. Since we don't use it after finishing reading sys_chunk_array, we don't need to keep it in memory. Signed-off-by: Liu Bo Signed-off-by: David Sterba --- fs/btrfs/volumes.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index e0290303bb35..fd3e8280b549 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -6630,13 +6630,13 @@ int btrfs_read_sys_array(struct btrfs_root *root) sb_array_offset += len; cur_offset += len; } - free_extent_buffer(sb); + free_extent_buffer_stale(sb); return ret; out_short_read: printk(KERN_ERR "BTRFS: sys_array too short to read %u bytes at offset %u\n", len, cur_offset); - free_extent_buffer(sb); + free_extent_buffer_stale(sb); return -EIO; } -- cgit v1.3-8-gc7d7 From 2d324f59f343967a03eeb2690f0ff178304d0687 Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Tue, 17 May 2016 17:21:48 -0700 Subject: Btrfs: fix unexpected return value of fiemap btrfs's fiemap is supposed to return 0 on success and return < 0 on error. however, ret becomes 1 after looking up the last file extent: btrfs_lookup_file_extent -> btrfs_search_slot(..., ins_len=0, cow=0) and if the offset is beyond EOF, we'll get 'path' pointed to the place of potentail insertion, and ret == 1. This may confuse applications using ioctl(FIEL_IOC_FIEMAP). Signed-off-by: Liu Bo Signed-off-by: David Sterba --- fs/btrfs/extent_io.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 2f83448d34fe..f661688c2102 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -4389,8 +4389,12 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, if (ret < 0) { btrfs_free_path(path); return ret; + } else { + WARN_ON(!ret); + if (ret == 1) + ret = 0; } - WARN_ON(!ret); + path->slots[0]--; btrfs_item_key_to_cpu(path->nodes[0], &found_key, path->slots[0]); found_type = found_key.type; -- cgit v1.3-8-gc7d7 From f1fee6534dcfbf468a159789aa202db2bce8c200 Mon Sep 17 00:00:00 2001 From: Zhao Lei Date: Tue, 17 May 2016 17:37:38 +0800 Subject: btrfs: scrub: Set bbio to NULL before calling btrfs_map_block We usually call btrfs_put_bbio() when btrfs_map_block() failed, btrfs_put_bbio() works right whether bbio is a valid value, or NULL. But there is a exception, in some case, btrfs_map_block() will return fail without touching *bbio(keeping its original value), and if bbio was not initialized yet, invalid memory accessing will happened. Above case is in scrub_missing_raid56_pages(), and similar case in scrub_raid56_parity(). Signed-off-by: Zhao Lei Signed-off-by: David Sterba --- fs/btrfs/scrub.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c index fa35cdc46494..ab3b9d604919 100644 --- a/fs/btrfs/scrub.c +++ b/fs/btrfs/scrub.c @@ -2181,7 +2181,7 @@ static void scrub_missing_raid56_pages(struct scrub_block *sblock) struct btrfs_fs_info *fs_info = sctx->dev_root->fs_info; u64 length = sblock->page_count * PAGE_SIZE; u64 logical = sblock->pagev[0]->logical; - struct btrfs_bio *bbio; + struct btrfs_bio *bbio = NULL; struct bio *bio; struct btrfs_raid_bio *rbio; int ret; @@ -2982,6 +2982,7 @@ again: extent_len); mapped_length = extent_len; + bbio = NULL; ret = btrfs_map_block(fs_info, READ, extent_logical, &mapped_length, &bbio, 0); if (!ret) { -- cgit v1.3-8-gc7d7 From 0132761017e012ab4dc8584d679503f2ba26ca86 Mon Sep 17 00:00:00 2001 From: Nicholas D Steeves Date: Thu, 19 May 2016 21:18:45 -0400 Subject: btrfs: fix string and comment grammatical issues and typos Signed-off-by: Nicholas D Steeves Signed-off-by: David Sterba --- fs/btrfs/backref.c | 2 +- fs/btrfs/btrfs_inode.h | 2 +- fs/btrfs/check-integrity.c | 2 +- fs/btrfs/ctree.c | 14 +++++++------- fs/btrfs/ctree.h | 6 +++--- fs/btrfs/delayed-ref.h | 2 +- fs/btrfs/dev-replace.c | 2 +- fs/btrfs/disk-io.c | 10 +++++----- fs/btrfs/extent-tree.c | 32 ++++++++++++++++---------------- fs/btrfs/extent_io.c | 4 ++-- fs/btrfs/extent_map.c | 2 +- fs/btrfs/file.c | 4 ++-- fs/btrfs/free-space-cache.c | 2 +- fs/btrfs/free-space-cache.h | 2 +- fs/btrfs/inode.c | 22 +++++++++++----------- fs/btrfs/ioctl.c | 10 +++++----- fs/btrfs/ordered-data.h | 2 +- fs/btrfs/qgroup.c | 16 ++++++++-------- fs/btrfs/raid56.c | 6 +++--- fs/btrfs/relocation.c | 12 ++++++------ fs/btrfs/root-tree.c | 4 ++-- fs/btrfs/scrub.c | 4 ++-- fs/btrfs/send.c | 6 +++--- fs/btrfs/struct-funcs.c | 2 +- fs/btrfs/super.c | 8 ++++---- fs/btrfs/tests/extent-io-tests.c | 2 +- fs/btrfs/tests/free-space-tests.c | 7 ++++--- fs/btrfs/tests/inode-tests.c | 2 +- fs/btrfs/tests/qgroup-tests.c | 2 +- fs/btrfs/transaction.h | 2 +- fs/btrfs/tree-log.c | 8 ++++---- fs/btrfs/ulist.c | 2 +- fs/btrfs/volumes.c | 8 ++++---- 33 files changed, 106 insertions(+), 105 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/backref.c b/fs/btrfs/backref.c index 80e8472d618b..b8b598740bbd 100644 --- a/fs/btrfs/backref.c +++ b/fs/btrfs/backref.c @@ -1939,7 +1939,7 @@ static int inode_to_path(u64 inum, u32 name_len, unsigned long name_off, * from ipath->fspath->val[i]. * when it returns, there are ipath->fspath->elem_cnt number of paths available * in ipath->fspath->val[]. when the allocated space wasn't sufficient, the - * number of missed paths in recored in ipath->fspath->elem_missed, otherwise, + * number of missed paths is recorded in ipath->fspath->elem_missed, otherwise, * it's zero. ipath->fspath->bytes_missing holds the number of bytes that would * have been needed to return all paths. */ diff --git a/fs/btrfs/btrfs_inode.h b/fs/btrfs/btrfs_inode.h index 61205e3bbefa..c0a2018761f9 100644 --- a/fs/btrfs/btrfs_inode.h +++ b/fs/btrfs/btrfs_inode.h @@ -303,7 +303,7 @@ struct btrfs_dio_private { struct bio *dio_bio; /* - * The original bio may be splited to several sub-bios, this is + * The original bio may be split to several sub-bios, this is * done during endio of sub-bios */ int (*subio_endio)(struct inode *, struct btrfs_io_bio *, int); diff --git a/fs/btrfs/check-integrity.c b/fs/btrfs/check-integrity.c index 516e19d1d202..b677a6ea6001 100644 --- a/fs/btrfs/check-integrity.c +++ b/fs/btrfs/check-integrity.c @@ -1939,7 +1939,7 @@ again: /* * Clear all references of this block. Do not free * the block itself even if is not referenced anymore - * because it still carries valueable information + * because it still carries valuable information * like whether it was ever written and IO completed. */ list_for_each_entry_safe(l, tmp, &block->ref_to_list, diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c index ec7928a27aaa..4997f175b446 100644 --- a/fs/btrfs/ctree.c +++ b/fs/btrfs/ctree.c @@ -156,7 +156,7 @@ struct extent_buffer *btrfs_root_node(struct btrfs_root *root) /* * RCU really hurts here, we could free up the root node because - * it was cow'ed but we may not get the new root node yet so do + * it was COWed but we may not get the new root node yet so do * the inc_not_zero dance and if it doesn't work then * synchronize_rcu and try again. */ @@ -955,7 +955,7 @@ int btrfs_block_can_be_shared(struct btrfs_root *root, struct extent_buffer *buf) { /* - * Tree blocks not in refernece counted trees and tree roots + * Tree blocks not in reference counted trees and tree roots * are never shared. If a block was allocated after the last * snapshot and the block was not allocated by tree relocation, * we know the block is not shared. @@ -1270,7 +1270,7 @@ __tree_mod_log_oldest_root(struct btrfs_fs_info *fs_info, /* * tm is a pointer to the first operation to rewind within eb. then, all - * previous operations will be rewinded (until we reach something older than + * previous operations will be rewound (until we reach something older than * time_seq). */ static void @@ -1345,7 +1345,7 @@ __tree_mod_log_rewind(struct btrfs_fs_info *fs_info, struct extent_buffer *eb, } /* - * Called with eb read locked. If the buffer cannot be rewinded, the same buffer + * Called with eb read locked. If the buffer cannot be rewound, the same buffer * is returned. If rewind operations happen, a fresh buffer is returned. The * returned buffer is always read-locked. If the returned buffer is not the * input buffer, the lock on the input buffer is released and the input buffer @@ -1516,7 +1516,7 @@ static inline int should_cow_block(struct btrfs_trans_handle *trans, * 3) the root is not forced COW. * * What is forced COW: - * when we create snapshot during commiting the transaction, + * when we create snapshot during committing the transaction, * after we've finished coping src root, we must COW the shared * block to ensure the metadata consistency. */ @@ -1531,7 +1531,7 @@ static inline int should_cow_block(struct btrfs_trans_handle *trans, /* * cows a single block, see __btrfs_cow_block for the real work. - * This version of it has extra checks so that a block isn't cow'd more than + * This version of it has extra checks so that a block isn't COWed more than * once per transaction, as long as it hasn't been written yet */ noinline int btrfs_cow_block(struct btrfs_trans_handle *trans, @@ -2986,7 +2986,7 @@ again: btrfs_unlock_up_safe(p, level + 1); /* - * Since we can unwind eb's we want to do a real search every + * Since we can unwind ebs we want to do a real search every * time. */ prev_cmp = -1; diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 84a6a5b3384a..65e7a9fa881f 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -186,7 +186,7 @@ static const int btrfs_csum_sizes[] = { 4 }; /* four bytes for CRC32 */ #define BTRFS_EMPTY_DIR_SIZE 0 -/* spefic to btrfs_map_block(), therefore not in include/linux/blk_types.h */ +/* specific to btrfs_map_block(), therefore not in include/linux/blk_types.h */ #define REQ_GET_READ_MIRRORS (1 << 30) #define BTRFS_FT_UNKNOWN 0 @@ -1221,7 +1221,7 @@ struct btrfs_space_info { * bytes_pinned does not reflect the bytes that will be pinned once the * delayed refs are flushed, so this counter is inc'ed every time we * call btrfs_free_extent so it is a realtime count of what will be - * freed once the transaction is committed. It will be zero'ed every + * freed once the transaction is committed. It will be zeroed every * time the transaction commits. */ struct percpu_counter total_bytes_pinned; @@ -2392,7 +2392,7 @@ static inline void btrfs_init_map_token (struct btrfs_map_token *token) token->kaddr = NULL; } -/* some macros to generate set/get funcs for the struct fields. This +/* some macros to generate set/get functions for the struct fields. This * assumes there is a lefoo_to_cpu for every type, so lets make a simple * one for u8: */ diff --git a/fs/btrfs/delayed-ref.h b/fs/btrfs/delayed-ref.h index c24b653c7343..5fca9534a271 100644 --- a/fs/btrfs/delayed-ref.h +++ b/fs/btrfs/delayed-ref.h @@ -188,7 +188,7 @@ struct btrfs_delayed_ref_root { /* * To make qgroup to skip given root. - * This is for snapshot, as btrfs_qgroup_inherit() will manully + * This is for snapshot, as btrfs_qgroup_inherit() will manually * modify counters for snapshot and its source, so we should skip * the snapshot in new_root/old_roots or it will get calculated twice */ diff --git a/fs/btrfs/dev-replace.c b/fs/btrfs/dev-replace.c index 26bcb487f958..0584b6c2df57 100644 --- a/fs/btrfs/dev-replace.c +++ b/fs/btrfs/dev-replace.c @@ -441,7 +441,7 @@ leave: } /* - * blocked until all flighting bios are finished. + * blocked until all in-flight bios operations are finished. */ static void btrfs_rm_dev_replace_blocked(struct btrfs_fs_info *fs_info) { diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 263823f28ba4..aacc707ef952 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -384,7 +384,7 @@ static int verify_parent_transid(struct extent_io_tree *io_tree, /* * Things reading via commit roots that don't have normal protection, * like send, can have a really old block in cache that may point at a - * block that has been free'd and re-allocated. So don't clear uptodate + * block that has been freed and re-allocated. So don't clear uptodate * if we find an eb that is under IO (dirty/writeback) because we could * end up reading in the stale data and then writing it back out and * making everybody very sad. @@ -418,7 +418,7 @@ static int btrfs_check_super_csum(char *raw_disk_sb) /* * The super_block structure does not span the whole * BTRFS_SUPER_INFO_SIZE range, we expect that the unused space - * is filled with zeros and is included in the checkum. + * is filled with zeros and is included in the checksum. */ crc = btrfs_csum_data(raw_disk_sb + BTRFS_CSUM_SIZE, crc, BTRFS_SUPER_INFO_SIZE - BTRFS_CSUM_SIZE); @@ -600,7 +600,7 @@ static noinline int check_leaf(struct btrfs_root *root, /* * Check to make sure that we don't point outside of the leaf, - * just incase all the items are consistent to eachother, but + * just in case all the items are consistent to each other, but * all point outside of the leaf. */ if (btrfs_item_end_nr(leaf, slot) > @@ -3017,7 +3017,7 @@ retry_root_backup: } /* - * Mount does not set all options immediatelly, we can do it now and do + * Mount does not set all options immediately, we can do it now and do * not have to wait for transaction commit */ btrfs_apply_pending_changes(fs_info); @@ -3245,7 +3245,7 @@ static void btrfs_end_buffer_write_sync(struct buffer_head *bh, int uptodate) btrfs_warn_rl_in_rcu(device->dev_root->fs_info, "lost page write due to IO error on %s", rcu_str_deref(device->name)); - /* note, we dont' set_buffer_write_io_error because we have + /* note, we don't set_buffer_write_io_error because we have * our own ways of dealing with the IO errors */ clear_buffer_uptodate(bh); diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 0ddc996a40e1..939a043eabfa 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -980,7 +980,7 @@ out_free: * event that tree block loses its owner tree's reference and do the * back refs conversion. * - * When a tree block is COW'd through a tree, there are four cases: + * When a tree block is COWed through a tree, there are four cases: * * The reference count of the block is one and the tree is the block's * owner tree. Nothing to do in this case. @@ -2595,7 +2595,7 @@ static noinline int __btrfs_run_delayed_refs(struct btrfs_trans_handle *trans, } /* - * Need to drop our head ref lock and re-aqcuire the + * Need to drop our head ref lock and re-acquire the * delayed ref lock and then re-check to make sure * nobody got added. */ @@ -2747,7 +2747,7 @@ static inline u64 heads_to_leaves(struct btrfs_root *root, u64 heads) /* * We don't ever fill up leaves all the way so multiply by 2 just to be - * closer to what we're really going to want to ouse. + * closer to what we're really going to want to use. */ return div_u64(num_bytes, BTRFS_LEAF_DATA_SIZE(root)); } @@ -2851,7 +2851,7 @@ static void delayed_ref_async_start(struct btrfs_work *work) } /* - * trans->sync means that when we call end_transaciton, we won't + * trans->sync means that when we call end_transaction, we won't * wait on delayed refs */ trans->sync = true; @@ -4243,7 +4243,7 @@ void btrfs_free_reserved_data_space_noquota(struct inode *inode, u64 start, * Called if we need to clear a data reservation for this inode * Normally in a error case. * - * This one will handle the per-indoe data rsv map for accurate reserved + * This one will handle the per-inode data rsv map for accurate reserved * space framework. */ void btrfs_free_reserved_data_space(struct inode *inode, u64 start, u64 len) @@ -4911,7 +4911,7 @@ void btrfs_init_async_reclaim_work(struct work_struct *work) * @orig_bytes - the number of bytes we want * @flush - whether or not we can flush to make our reservation * - * This will reserve orgi_bytes number of bytes from the space info associated + * This will reserve orig_bytes number of bytes from the space info associated * with the block_rsv. If there is not enough space it will make an attempt to * flush out space to make room. It will do this by flushing delalloc if * possible or committing the transaction. If flush is 0 then no attempts to @@ -5516,7 +5516,7 @@ void btrfs_orphan_release_metadata(struct inode *inode) * common file/directory operations, they change two fs/file trees * and root tree, the number of items that the qgroup reserves is * different with the free space reservation. So we can not use - * the space reseravtion mechanism in start_transaction(). + * the space reservation mechanism in start_transaction(). */ int btrfs_subvolume_reserve_metadata(struct btrfs_root *root, struct btrfs_block_rsv *rsv, @@ -5565,7 +5565,7 @@ void btrfs_subvolume_release_metadata(struct btrfs_root *root, /** * drop_outstanding_extent - drop an outstanding extent * @inode: the inode we're dropping the extent for - * @num_bytes: the number of bytes we're relaseing. + * @num_bytes: the number of bytes we're releasing. * * This is called when we are freeing up an outstanding extent, either called * after an error or after an extent is written. This will return the number of @@ -5591,7 +5591,7 @@ static unsigned drop_outstanding_extent(struct inode *inode, u64 num_bytes) drop_inode_space = 1; /* - * If we have more or the same amount of outsanding extents than we have + * If we have more or the same amount of outstanding extents than we have * reserved then we need to leave the reserved extents count alone. */ if (BTRFS_I(inode)->outstanding_extents >= @@ -5605,8 +5605,8 @@ static unsigned drop_outstanding_extent(struct inode *inode, u64 num_bytes) } /** - * calc_csum_metadata_size - return the amount of metada space that must be - * reserved/free'd for the given bytes. + * calc_csum_metadata_size - return the amount of metadata space that must be + * reserved/freed for the given bytes. * @inode: the inode we're manipulating * @num_bytes: the number of bytes in question * @reserve: 1 if we are reserving space, 0 if we are freeing space @@ -5758,7 +5758,7 @@ out_fail: /* * This is tricky, but first we need to figure out how much we - * free'd from any free-ers that occurred during this + * freed from any free-ers that occurred during this * reservation, so we reset ->csum_bytes to the csum_bytes * before we dropped our lock, and then call the free for the * number of bytes that were freed while we were trying our @@ -5780,7 +5780,7 @@ out_fail: /* * Now reset ->csum_bytes to what it should be. If bytes is - * more than to_free then we would have free'd more space had we + * more than to_free then we would have freed more space had we * not had an artificially high ->csum_bytes, so we need to free * the remainder. If bytes is the same or less then we don't * need to do anything, the other free-ers did the correct @@ -7471,7 +7471,7 @@ loop: if (loop == LOOP_CACHING_NOWAIT) { /* * We want to skip the LOOP_CACHING_WAIT step if we - * don't have any unached bgs and we've alrelady done a + * don't have any uncached bgs and we've already done a * full search through. */ if (orig_have_caching_bg || !full_search) @@ -7873,7 +7873,7 @@ int btrfs_alloc_logged_file_extent(struct btrfs_trans_handle *trans, /* * Mixed block groups will exclude before processing the log so we only - * need to do the exlude dance if this fs isn't mixed. + * need to do the exclude dance if this fs isn't mixed. */ if (!btrfs_fs_incompat(root->fs_info, MIXED_GROUPS)) { ret = __exclude_logged_extent(root, ins->objectid, ins->offset); @@ -9317,7 +9317,7 @@ u64 btrfs_account_ro_block_groups_free_space(struct btrfs_space_info *sinfo) u64 free_bytes = 0; int factor; - /* It's df, we don't care if it's racey */ + /* It's df, we don't care if it's racy */ if (list_empty(&sinfo->ro_bgs)) return 0; diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index c1139bcf8870..5fdf1fcb03c9 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -4591,7 +4591,7 @@ static void btrfs_release_extent_buffer_page(struct extent_buffer *eb) if (mapped) spin_unlock(&page->mapping->private_lock); - /* One for when we alloced the page */ + /* One for when we allocated the page */ put_page(page); } while (index != 0); } @@ -5751,7 +5751,7 @@ int try_release_extent_buffer(struct page *page) struct extent_buffer *eb; /* - * We need to make sure noboody is attaching this page to an eb right + * We need to make sure nobody is attaching this page to an eb right * now. */ spin_lock(&page->mapping->private_lock); diff --git a/fs/btrfs/extent_map.c b/fs/btrfs/extent_map.c index 318b048eb254..e0715fcfb11e 100644 --- a/fs/btrfs/extent_map.c +++ b/fs/btrfs/extent_map.c @@ -62,7 +62,7 @@ struct extent_map *alloc_extent_map(void) /** * free_extent_map - drop reference count of an extent_map - * @em: extent map being releasead + * @em: extent map being released * * Drops the reference out on @em by one and free the structure * if the reference count hits zero. diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index 8d7b5a45c005..50dac40d9561 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -2024,7 +2024,7 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync) BTRFS_I(inode)->last_trans <= root->fs_info->last_trans_committed)) { /* - * We'v had everything committed since the last time we were + * We've had everything committed since the last time we were * modified so clear this flag in case it was set for whatever * reason, it's no longer relevant. */ @@ -2372,7 +2372,7 @@ static int btrfs_punch_hole(struct inode *inode, loff_t offset, loff_t len) /* Check the aligned pages after the first unaligned page, * if offset != orig_start, which means the first unaligned page - * including serveral following pages are already in holes, + * including several following pages are already in holes, * the extra check can be skipped */ if (offset == orig_start) { /* after truncate page, check hole again */ diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c index 5e6062c26129..c6dc1183f542 100644 --- a/fs/btrfs/free-space-cache.c +++ b/fs/btrfs/free-space-cache.c @@ -1983,7 +1983,7 @@ static bool use_bitmap(struct btrfs_free_space_ctl *ctl, /* * If this block group has some small extents we don't want to * use up all of our free slots in the cache with them, we want - * to reserve them to larger extents, however if we have plent + * to reserve them to larger extents, however if we have plenty * of cache left then go ahead an dadd them, no sense in adding * the overhead of a bitmap if we don't have to. */ diff --git a/fs/btrfs/free-space-cache.h b/fs/btrfs/free-space-cache.h index 33178c490ace..3af651c2bbc7 100644 --- a/fs/btrfs/free-space-cache.h +++ b/fs/btrfs/free-space-cache.h @@ -123,7 +123,7 @@ int btrfs_return_cluster_to_free_space( int btrfs_trim_block_group(struct btrfs_block_group_cache *block_group, u64 *trimmed, u64 start, u64 end, u64 minlen); -/* Support functions for runnint our sanity tests */ +/* Support functions for running our sanity tests */ #ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS int test_add_free_space_entry(struct btrfs_block_group_cache *cache, u64 offset, u64 bytes, bool bitmap); diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 1a4bd193962c..2027bbd4b05e 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -455,7 +455,7 @@ again: /* * skip compression for a small file range(<=blocksize) that - * isn't an inline extent, since it dosen't save disk space at all. + * isn't an inline extent, since it doesn't save disk space at all. */ if (total_compressed <= blocksize && (start > 0 || end + 1 < BTRFS_I(inode)->disk_i_size)) @@ -3705,7 +3705,7 @@ cache_index: * and doesn't have an inode ref with the name "bar" anymore. * * Setting last_unlink_trans to last_trans is a pessimistic approach, - * but it guarantees correctness at the expense of ocassional full + * but it guarantees correctness at the expense of occasional full * transaction commits on fsync if our inode is a directory, or if our * inode is not a directory, logging its parent unnecessarily. */ @@ -4961,7 +4961,7 @@ static int btrfs_setsize(struct inode *inode, struct iattr *attr) * be instantly completed which will give us extents that need * to be truncated. If we fail to get an orphan inode down we * could have left over extents that were never meant to live, - * so we need to garuntee from this point on that everything + * so we need to guarantee from this point on that everything * will be consistent. */ ret = btrfs_orphan_add(trans, inode); @@ -5231,7 +5231,7 @@ void btrfs_evict_inode(struct inode *inode) } /* - * We can't just steal from the global reserve, we need tomake + * We can't just steal from the global reserve, we need to make * sure there is room to do it, if not we need to commit and try * again. */ @@ -7407,7 +7407,7 @@ static int lock_extent_direct(struct inode *inode, u64 lockstart, u64 lockend, cached_state); /* * We're concerned with the entire range that we're going to be - * doing DIO to, so we need to make sure theres no ordered + * doing DIO to, so we need to make sure there's no ordered * extents in this range. */ ordered = btrfs_lookup_ordered_range(inode, lockstart, @@ -7569,7 +7569,7 @@ static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock, if (current->journal_info) { /* * Need to pull our outstanding extents and set journal_info to NULL so - * that anything that needs to check if there's a transction doesn't get + * that anything that needs to check if there's a transaction doesn't get * confused. */ dio_data = current->journal_info; @@ -7602,7 +7602,7 @@ static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock, * decompress it, so there will be buffering required no matter what we * do, so go ahead and fallback to buffered. * - * We return -ENOTBLK because thats what makes DIO go ahead and go back + * We return -ENOTBLK because that's what makes DIO go ahead and go back * to buffered IO. Don't blame me, this is the price we pay for using * the generic code. */ @@ -9018,7 +9018,7 @@ static int btrfs_truncate(struct inode *inode) return ret; /* - * Yes ladies and gentelment, this is indeed ugly. The fact is we have + * Yes ladies and gentlemen, this is indeed ugly. The fact is we have * 3 things going on here * * 1) We need to reserve space for our orphan item and the space to @@ -9032,15 +9032,15 @@ static int btrfs_truncate(struct inode *inode) * space reserved in case it uses space during the truncate (thank you * very much snapshotting). * - * And we need these to all be seperate. The fact is we can use alot of + * And we need these to all be separate. The fact is we can use a lot of * space doing the truncate, and we have no earthly idea how much space - * we will use, so we need the truncate reservation to be seperate so it + * we will use, so we need the truncate reservation to be separate so it * doesn't end up using space reserved for updating the inode or * removing the orphan item. We also need to be able to stop the * transaction and start a new one, which means we need to be able to * update the inode several times, and we have no idea of knowing how * many times that will be, so we can't just reserve 1 item for the - * entirety of the opration, so that has to be done seperately as well. + * entirety of the operation, so that has to be done separately as well. * Then there is the orphan item, which does indeed need to be held on * to for the whole operation, and we need nobody to touch this reserved * space except the orphan code. diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index 72f48252a55d..972cd0b00b21 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -296,7 +296,7 @@ static int btrfs_ioctl_setflags(struct file *file, void __user *arg) } } else { /* - * Revert back under same assuptions as above + * Revert back under same assumptions as above */ if (S_ISREG(mode)) { if (inode->i_size == 0) @@ -461,7 +461,7 @@ static noinline int create_subvol(struct inode *dir, /* * Don't create subvolume whose level is not zero. Or qgroup will be - * screwed up since it assume subvolme qgroup's level to be 0. + * screwed up since it assumes subvolume qgroup's level to be 0. */ if (btrfs_qgroup_level(objectid)) return -ENOSPC; @@ -771,7 +771,7 @@ free_pending: * a. be owner of dir, or * b. be owner of victim, or * c. have CAP_FOWNER capability - * 6. If the victim is append-only or immutable we can't do antyhing with + * 6. If the victim is append-only or immutable we can't do anything with * links pointing to it. * 7. If we were asked to remove a directory and victim isn't one - ENOTDIR. * 8. If we were asked to remove a non-directory and victim isn't one - EISDIR. @@ -4585,7 +4585,7 @@ again: } /* - * mut. excl. ops lock is locked. Three possibilites: + * mut. excl. ops lock is locked. Three possibilities: * (1) some other op is running * (2) balance is running * (3) balance is paused -- special case (think resume) @@ -5490,7 +5490,7 @@ long btrfs_ioctl(struct file *file, unsigned int ret = btrfs_sync_fs(file_inode(file)->i_sb, 1); /* * The transaction thread may want to do more work, - * namely it pokes the cleaner ktread that will start + * namely it pokes the cleaner kthread that will start * processing uncleaned subvols. */ wake_up_process(root->fs_info->transaction_kthread); diff --git a/fs/btrfs/ordered-data.h b/fs/btrfs/ordered-data.h index 23c96059cef2..ae068bbddc0d 100644 --- a/fs/btrfs/ordered-data.h +++ b/fs/btrfs/ordered-data.h @@ -58,7 +58,7 @@ struct btrfs_ordered_sum { #define BTRFS_ORDERED_COMPRESSED 3 /* writing a zlib compressed extent */ -#define BTRFS_ORDERED_PREALLOC 4 /* set when writing to prealloced extent */ +#define BTRFS_ORDERED_PREALLOC 4 /* set when writing to preallocated extent */ #define BTRFS_ORDERED_DIRECT 5 /* set when we're doing DIO with this extent */ diff --git a/fs/btrfs/qgroup.c b/fs/btrfs/qgroup.c index 34af959b4e0f..9d4c05b14f6e 100644 --- a/fs/btrfs/qgroup.c +++ b/fs/btrfs/qgroup.c @@ -85,7 +85,7 @@ struct btrfs_qgroup { /* * temp variables for accounting operations - * Refer to qgroup_shared_accouting() for details. + * Refer to qgroup_shared_accounting() for details. */ u64 old_refcnt; u64 new_refcnt; @@ -499,7 +499,7 @@ void btrfs_free_qgroup_config(struct btrfs_fs_info *fs_info) } /* * we call btrfs_free_qgroup_config() when umounting - * filesystem and disabling quota, so we set qgroup_ulit + * filesystem and disabling quota, so we set qgroup_ulist * to be null here to avoid double free. */ ulist_free(fs_info->qgroup_ulist); @@ -1036,7 +1036,7 @@ static void qgroup_dirty(struct btrfs_fs_info *fs_info, /* * The easy accounting, if we are adding/removing the only ref for an extent - * then this qgroup and all of the parent qgroups get their refrence and + * then this qgroup and all of the parent qgroups get their reference and * exclusive counts adjusted. * * Caller should hold fs_info->qgroup_lock. @@ -1436,7 +1436,7 @@ int btrfs_qgroup_prepare_account_extents(struct btrfs_trans_handle *trans, /* * No need to do lock, since this function will only be called in - * btrfs_commmit_transaction(). + * btrfs_commit_transaction(). */ node = rb_first(&delayed_refs->dirty_extent_root); while (node) { @@ -1557,7 +1557,7 @@ static int qgroup_update_refcnt(struct btrfs_fs_info *fs_info, * A: cur_old_roots < nr_old_roots (not exclusive before) * !A: cur_old_roots == nr_old_roots (possible exclusive before) * B: cur_new_roots < nr_new_roots (not exclusive now) - * !B: cur_new_roots == nr_new_roots (possible exclsuive now) + * !B: cur_new_roots == nr_new_roots (possible exclusive now) * * Results: * +: Possible sharing -> exclusive -: Possible exclusive -> sharing @@ -1851,7 +1851,7 @@ out: } /* - * Copy the acounting information between qgroups. This is necessary + * Copy the accounting information between qgroups. This is necessary * when a snapshot or a subvolume is created. Throwing an error will * cause a transaction abort so we take extra care here to only error * when a readonly fs is a reasonable outcome. @@ -2340,7 +2340,7 @@ out: mutex_unlock(&fs_info->qgroup_rescan_lock); /* - * only update status, since the previous part has alreay updated the + * only update status, since the previous part has already updated the * qgroup info. */ trans = btrfs_start_transaction(fs_info->quota_root, 1); @@ -2670,7 +2670,7 @@ void btrfs_qgroup_free_meta(struct btrfs_root *root, int num_bytes) } /* - * Check qgroup reserved space leaking, normally at destory inode + * Check qgroup reserved space leaking, normally at destroy inode * time */ void btrfs_qgroup_check_reserved_leak(struct inode *inode) diff --git a/fs/btrfs/raid56.c b/fs/btrfs/raid56.c index 0b7792e02dd5..f8b6d411a034 100644 --- a/fs/btrfs/raid56.c +++ b/fs/btrfs/raid56.c @@ -576,7 +576,7 @@ static int rbio_can_merge(struct btrfs_raid_bio *last, * we can't merge with cached rbios, since the * idea is that when we merge the destination * rbio is going to run our IO for us. We can - * steal from cached rbio's though, other functions + * steal from cached rbios though, other functions * handle that. */ if (test_bit(RBIO_CACHE_BIT, &last->flags) || @@ -2368,7 +2368,7 @@ static noinline void finish_parity_scrub(struct btrfs_raid_bio *rbio, run_xor(pointers + 1, nr_data - 1, PAGE_SIZE); } - /* Check scrubbing pairty and repair it */ + /* Check scrubbing parity and repair it */ p = rbio_stripe_page(rbio, rbio->scrubp, pagenr); parity = kmap(p); if (memcmp(parity, pointers[rbio->scrubp], PAGE_SIZE)) @@ -2493,7 +2493,7 @@ static void validate_rbio_for_parity_scrub(struct btrfs_raid_bio *rbio) /* * Here means we got one corrupted data stripe and one * corrupted parity on RAID6, if the corrupted parity - * is scrubbing parity, luckly, use the other one to repair + * is scrubbing parity, luckily, use the other one to repair * the data, or we can not repair the data stripe. */ if (failp != rbio->scrubp) diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c index bd5ea1a8a9f1..852be779a834 100644 --- a/fs/btrfs/relocation.c +++ b/fs/btrfs/relocation.c @@ -668,8 +668,8 @@ int find_inline_backref(struct extent_buffer *leaf, int slot, * roots of b-trees that reference the tree block. * * the basic idea of this function is check backrefs of a given block - * to find upper level blocks that refernece the block, and then check - * bakcrefs of these upper level blocks recursively. the recursion stop + * to find upper level blocks that reference the block, and then check + * backrefs of these upper level blocks recursively. the recursion stop * when tree root is reached or backrefs for the block is cached. * * NOTE: if we find backrefs for a block are cached, we know backrefs @@ -1160,7 +1160,7 @@ out: if (!RB_EMPTY_NODE(&upper->rb_node)) continue; - /* Add this guy's upper edges to the list to proces */ + /* Add this guy's upper edges to the list to process */ list_for_each_entry(edge, &upper->upper, list[LOWER]) list_add_tail(&edge->list[UPPER], &list); if (list_empty(&upper->upper)) @@ -2396,7 +2396,7 @@ again: } /* - * we keep the old last snapshod transid in rtranid when we + * we keep the old last snapshot transid in rtranid when we * created the relocation tree. */ last_snap = btrfs_root_rtransid(&reloc_root->root_item); @@ -2616,7 +2616,7 @@ static int reserve_metadata_space(struct btrfs_trans_handle *trans, * only one thread can access block_rsv at this point, * so we don't need hold lock to protect block_rsv. * we expand more reservation size here to allow enough - * space for relocation and we will return eailer in + * space for relocation and we will return earlier in * enospc case. */ rc->block_rsv->size = tmp + rc->extent_root->nodesize * @@ -4591,7 +4591,7 @@ int btrfs_reloc_cow_block(struct btrfs_trans_handle *trans, /* * called before creating snapshot. it calculates metadata reservation - * requried for relocating tree blocks in the snapshot + * required for relocating tree blocks in the snapshot */ void btrfs_reloc_pre_snapshot(struct btrfs_pending_snapshot *pending, u64 *bytes_to_reserve) diff --git a/fs/btrfs/root-tree.c b/fs/btrfs/root-tree.c index 9fcd6dfc3266..5ddf14e1b2d2 100644 --- a/fs/btrfs/root-tree.c +++ b/fs/btrfs/root-tree.c @@ -71,9 +71,9 @@ static void btrfs_read_root_item(struct extent_buffer *eb, int slot, * search_key: the key to search * path: the path we search * root_item: the root item of the tree we look for - * root_key: the reak key of the tree we look for + * root_key: the root key of the tree we look for * - * If ->offset of 'seach_key' is -1ULL, it means we are not sure the offset + * If ->offset of 'search_key' is -1ULL, it means we are not sure the offset * of the search key, just lookup the root with the highest offset for a * given objectid. * diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c index 841dae874a21..6838b2ca3cab 100644 --- a/fs/btrfs/scrub.c +++ b/fs/btrfs/scrub.c @@ -1044,7 +1044,7 @@ nodatasum_case: /* * !is_metadata and !have_csum, this means that the data - * might not be COW'ed, that it might be modified + * might not be COWed, that it might be modified * concurrently. The general strategy to work on the * commit root does not help in the case when COW is not * used. @@ -1125,7 +1125,7 @@ nodatasum_case: * the 2nd page of mirror #1 faces I/O errors, and the 2nd page * of mirror #2 is readable but the final checksum test fails, * then the 2nd page of mirror #3 could be tried, whether now - * the final checksum succeedes. But this would be a rare + * the final checksum succeeds. But this would be a rare * exception and is therefore not implemented. At least it is * avoided that the good copy is overwritten. * A more useful improvement would be to pick the sectors diff --git a/fs/btrfs/send.c b/fs/btrfs/send.c index 8d358c547c59..e9710f461008 100644 --- a/fs/btrfs/send.c +++ b/fs/btrfs/send.c @@ -1831,7 +1831,7 @@ static int will_overwrite_ref(struct send_ctx *sctx, u64 dir, u64 dir_gen, /* * If we have a parent root we need to verify that the parent dir was - * not delted and then re-created, if it was then we have no overwrite + * not deleted and then re-created, if it was then we have no overwrite * and we can just unlink this entry. */ if (sctx->parent_root) { @@ -4192,9 +4192,9 @@ static int __process_new_xattr(int num, struct btrfs_key *di_key, return -ENOMEM; /* - * This hack is needed because empty acl's are stored as zero byte + * This hack is needed because empty acls are stored as zero byte * data in xattrs. Problem with that is, that receiving these zero byte - * acl's will fail later. To fix this, we send a dummy acl list that + * acls will fail later. To fix this, we send a dummy acl list that * only contains the version number and no entries. */ if (!strncmp(name, XATTR_NAME_POSIX_ACL_ACCESS, name_len) || diff --git a/fs/btrfs/struct-funcs.c b/fs/btrfs/struct-funcs.c index e05619f241be..875c757e73e2 100644 --- a/fs/btrfs/struct-funcs.c +++ b/fs/btrfs/struct-funcs.c @@ -36,7 +36,7 @@ static inline void put_unaligned_le8(u8 val, void *p) * * The end result is that anyone who #includes ctree.h gets a * declaration for the btrfs_set_foo functions and btrfs_foo functions, - * which are wappers of btrfs_set_token_#bits functions and + * which are wrappers of btrfs_set_token_#bits functions and * btrfs_get_token_#bits functions, which are defined in this file. * * These setget functions do all the extent_buffer related mapping diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index 00b8f37cc306..01ef1865db63 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -121,7 +121,7 @@ static void btrfs_handle_error(struct btrfs_fs_info *fs_info) * Note that a running device replace operation is not * canceled here although there is no way to update * the progress. It would add the risk of a deadlock, - * therefore the canceling is ommited. The only penalty + * therefore the canceling is omitted. The only penalty * is that some I/O remains active until the procedure * completes. The next time when the filesystem is * mounted writeable again, the device replace @@ -1881,7 +1881,7 @@ static int btrfs_calc_avail_data_space(struct btrfs_root *root, u64 *free_bytes) int ret; /* - * We aren't under the device list lock, so this is racey-ish, but good + * We aren't under the device list lock, so this is racy-ish, but good * enough for our purposes. */ nr_devices = fs_info->fs_devices->open_devices; @@ -1900,7 +1900,7 @@ static int btrfs_calc_avail_data_space(struct btrfs_root *root, u64 *free_bytes) if (!devices_info) return -ENOMEM; - /* calc min stripe number for data space alloction */ + /* calc min stripe number for data space allocation */ type = btrfs_get_alloc_profile(root, 1); if (type & BTRFS_BLOCK_GROUP_RAID0) { min_stripes = 2; @@ -1936,7 +1936,7 @@ static int btrfs_calc_avail_data_space(struct btrfs_root *root, u64 *free_bytes) avail_space *= BTRFS_STRIPE_LEN; /* - * In order to avoid overwritting the superblock on the drive, + * In order to avoid overwriting the superblock on the drive, * btrfs starts at an offset of at least 1MB when doing chunk * allocation. */ diff --git a/fs/btrfs/tests/extent-io-tests.c b/fs/btrfs/tests/extent-io-tests.c index c190d2a520d1..55724607f79b 100644 --- a/fs/btrfs/tests/extent-io-tests.c +++ b/fs/btrfs/tests/extent-io-tests.c @@ -176,7 +176,7 @@ static int test_find_delalloc(void) locked_page = find_lock_page(inode->i_mapping, test_start >> PAGE_SHIFT); if (!locked_page) { - test_msg("Could'nt find the locked page\n"); + test_msg("Couldn't find the locked page\n"); goto out_bits; } start = test_start; diff --git a/fs/btrfs/tests/free-space-tests.c b/fs/btrfs/tests/free-space-tests.c index 514247515312..0eeb8f3d6b67 100644 --- a/fs/btrfs/tests/free-space-tests.c +++ b/fs/btrfs/tests/free-space-tests.c @@ -25,7 +25,7 @@ #define BITS_PER_BITMAP (PAGE_SIZE * 8) /* - * This test just does basic sanity checking, making sure we can add an exten + * This test just does basic sanity checking, making sure we can add an extent * entry and remove space from either end and the middle, and make sure we can * remove space that covers adjacent extent entries. */ @@ -396,8 +396,9 @@ static int check_cache_empty(struct btrfs_block_group_cache *cache) * wasn't optimal as they could be spread all over the block group while under * concurrency (extra overhead and fragmentation). * - * This stealing approach is benefical, since we always prefer to allocate from - * extent entries, both for clustered and non-clustered allocation requests. + * This stealing approach is beneficial, since we always prefer to allocate + * from extent entries, both for clustered and non-clustered allocation + * requests. */ static int test_steal_space_from_bitmap_to_extent(struct btrfs_block_group_cache *cache) diff --git a/fs/btrfs/tests/inode-tests.c b/fs/btrfs/tests/inode-tests.c index 863a6a3af1f8..8a25fe8b7c45 100644 --- a/fs/btrfs/tests/inode-tests.c +++ b/fs/btrfs/tests/inode-tests.c @@ -264,7 +264,7 @@ static noinline int test_btrfs_get_extent(void) /* * We will just free a dummy node if it's ref count is 2 so we need an - * extra ref so our searches don't accidently release our page. + * extra ref so our searches don't accidentally release our page. */ extent_buffer_get(root->node); btrfs_set_header_nritems(root->node, 0); diff --git a/fs/btrfs/tests/qgroup-tests.c b/fs/btrfs/tests/qgroup-tests.c index 8ea5d34bc5a2..8aa4ded31326 100644 --- a/fs/btrfs/tests/qgroup-tests.c +++ b/fs/btrfs/tests/qgroup-tests.c @@ -234,7 +234,7 @@ static int test_no_shared_qgroup(struct btrfs_root *root) } /* - * Since the test trans doesn't havee the complicated delayed refs, + * Since the test trans doesn't have the complicated delayed refs, * we can only call btrfs_qgroup_account_extent() directly to test * quota. */ diff --git a/fs/btrfs/transaction.h b/fs/btrfs/transaction.h index 72be51f7ca2f..9fe0ec2bf0fe 100644 --- a/fs/btrfs/transaction.h +++ b/fs/btrfs/transaction.h @@ -144,7 +144,7 @@ struct btrfs_pending_snapshot { /* block reservation for the operation */ struct btrfs_block_rsv block_rsv; u64 qgroup_reserved; - /* extra metadata reseration for relocation */ + /* extra metadata reservation for relocation */ int error; bool readonly; struct list_head list; diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c index a111e275ab9e..f3a8939d63fb 100644 --- a/fs/btrfs/tree-log.c +++ b/fs/btrfs/tree-log.c @@ -2330,7 +2330,7 @@ static int replay_one_buffer(struct btrfs_root *log, struct extent_buffer *eb, break; /* for regular files, make sure corresponding - * orhpan item exist. extents past the new EOF + * orphan item exist. extents past the new EOF * will be truncated later by orphan cleanup. */ if (S_ISREG(mode)) { @@ -4937,7 +4937,7 @@ out_unlock: * the actual unlink operation, so if we do this check before a concurrent task * sets last_unlink_trans it means we've logged a consistent version/state of * all the inode items, otherwise we are not sure and must do a transaction - * commit (the concurrent task migth have only updated last_unlink_trans before + * commit (the concurrent task might have only updated last_unlink_trans before * we logged the inode or it might have also done the unlink). */ static bool btrfs_must_commit_transaction(struct btrfs_trans_handle *trans, @@ -4996,7 +4996,7 @@ static noinline int check_parent_dirs_for_sync(struct btrfs_trans_handle *trans, while (1) { /* * If we are logging a directory then we start with our inode, - * not our parents inode, so we need to skipp setting the + * not our parent's inode, so we need to skip setting the * logged_trans so that further down in the log code we don't * think this inode has already been logged. */ @@ -5375,7 +5375,7 @@ static int btrfs_log_inode_parent(struct btrfs_trans_handle *trans, log_dentries = true; /* - * On unlink we must make sure all our current and old parent directores + * On unlink we must make sure all our current and old parent directory * inodes are fully logged. This is to prevent leaving dangling * directory index entries in directories that were our parents but are * not anymore. Not doing this results in old parent directory being diff --git a/fs/btrfs/ulist.c b/fs/btrfs/ulist.c index 91feb2bdefee..b1434bb57e36 100644 --- a/fs/btrfs/ulist.c +++ b/fs/btrfs/ulist.c @@ -28,7 +28,7 @@ * } * ulist_free(ulist); * - * This assumes the graph nodes are adressable by u64. This stems from the + * This assumes the graph nodes are addressable by u64. This stems from the * usage for tree enumeration in btrfs, where the logical addresses are * 64 bit. * diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index bd0f45fb38c4..8fe7fff5ffb7 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -2165,7 +2165,7 @@ static int btrfs_prepare_sprout(struct btrfs_root *root) } /* - * strore the expected generation for seed devices in device items. + * Store the expected generation for seed devices in device items. */ static int btrfs_finish_sprout(struct btrfs_trans_handle *trans, struct btrfs_root *root) @@ -3362,7 +3362,7 @@ static int should_balance_chunk(struct btrfs_root *root, } else if ((bargs->flags & BTRFS_BALANCE_ARGS_LIMIT_RANGE)) { /* * Same logic as the 'limit' filter; the minimum cannot be - * determined here because we do not have the global informatoin + * determined here because we do not have the global information * about the count of all chunks that satisfy the filters. */ if (bargs->limit_max == 0) @@ -6032,7 +6032,7 @@ static void bbio_error(struct btrfs_bio *bbio, struct bio *bio, u64 logical) { atomic_inc(&bbio->error); if (atomic_dec_and_test(&bbio->stripes_pending)) { - /* Shoud be the original bio. */ + /* Should be the original bio. */ WARN_ON(bio != bbio->orig_bio); btrfs_io_bio(bio)->mirror_num = bbio->mirror_num; @@ -6516,7 +6516,7 @@ int btrfs_read_sys_array(struct btrfs_root *root) set_extent_buffer_uptodate(sb); btrfs_set_buffer_lockdep_class(root->root_key.objectid, sb, 0); /* - * The sb extent buffer is artifical and just used to read the system array. + * The sb extent buffer is artificial and just used to read the system array. * set_extent_buffer_uptodate() call does not properly mark all it's * pages up-to-date when the page is larger: extent does not cover the * whole page and consequently check_page_uptodate does not find all -- cgit v1.3-8-gc7d7 From aaf431b4f92152d46ab54079692633aa422262b1 Mon Sep 17 00:00:00 2001 From: Andreas Gruenbacher Date: Mon, 9 May 2016 13:28:49 +0200 Subject: xattr: Fail with -EINVAL for NULL attribute names Commit 98e9cb57 improved the xattr name checks in xattr_resolve_name but didn't update the NULL attribute name check appropriately, so NULL attribute names lead to NULL pointer dereferences. Turn that into -EINVAL results instead. Signed-off-by: Andreas Gruenbacher fs/xattr.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) Signed-off-by: Al Viro --- fs/xattr.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/xattr.c b/fs/xattr.c index b11945e15fde..4231488f2122 100644 --- a/fs/xattr.c +++ b/fs/xattr.c @@ -668,7 +668,7 @@ xattr_resolve_name(const struct xattr_handler **handlers, const char **name) const struct xattr_handler *handler; if (!*name) - return NULL; + return ERR_PTR(-EINVAL); for_each_xattr_handler(handlers, handler) { const char *n; -- cgit v1.3-8-gc7d7 From 0040773bff7b585fc948565a0558e5a6a4680e96 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Wed, 25 May 2016 17:34:41 -0400 Subject: make xattr_resolve_handlers() safe to use with NULL ->s_xattr Signed-off-by: Al Viro --- fs/xattr.c | 1 + 1 file changed, 1 insertion(+) (limited to 'fs') diff --git a/fs/xattr.c b/fs/xattr.c index 4231488f2122..fc81e771488a 100644 --- a/fs/xattr.c +++ b/fs/xattr.c @@ -655,6 +655,7 @@ strcmp_prefix(const char *a, const char *a_prefix) * operations to the correct xattr_handler. */ #define for_each_xattr_handler(handlers, handler) \ + if (handlers) \ for ((handler) = *(handlers)++; \ (handler) != NULL; \ (handler) = *(handlers)++) -- cgit v1.3-8-gc7d7 From 3ed97d6345a36a0a61e6af62ad8a66ca40f1aa2e Mon Sep 17 00:00:00 2001 From: Ilya Dryomov Date: Tue, 26 Apr 2016 15:05:29 +0200 Subject: libceph: make ceph_osdc_put_request() accept NULL Signed-off-by: Ilya Dryomov --- fs/ceph/addr.c | 9 +++------ net/ceph/osd_client.c | 8 +++++--- 2 files changed, 8 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c index 4801571f51cb..3e61fc8bb371 100644 --- a/fs/ceph/addr.c +++ b/fs/ceph/addr.c @@ -1099,8 +1099,7 @@ release_pvec_pages: mapping->writeback_index = index; out: - if (req) - ceph_osdc_put_request(req); + ceph_osdc_put_request(req); ceph_put_snap_context(snapc); dout("writepages done, rc = %d\n", rc); return rc; @@ -1824,10 +1823,8 @@ static int __ceph_pool_perm_get(struct ceph_inode_info *ci, u32 pool) out_unlock: up_write(&mdsc->pool_perm_rwsem); - if (rd_req) - ceph_osdc_put_request(rd_req); - if (wr_req) - ceph_osdc_put_request(wr_req); + ceph_osdc_put_request(rd_req); + ceph_osdc_put_request(wr_req); out: if (!err) err = have; diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c index 40a53a70efdf..cacce9e35f08 100644 --- a/net/ceph/osd_client.c +++ b/net/ceph/osd_client.c @@ -354,9 +354,11 @@ EXPORT_SYMBOL(ceph_osdc_get_request); void ceph_osdc_put_request(struct ceph_osd_request *req) { - dout("%s %p (was %d)\n", __func__, req, - atomic_read(&req->r_kref.refcount)); - kref_put(&req->r_kref, ceph_osdc_release_request); + if (req) { + dout("%s %p (was %d)\n", __func__, req, + atomic_read(&req->r_kref.refcount)); + kref_put(&req->r_kref, ceph_osdc_release_request); + } } EXPORT_SYMBOL(ceph_osdc_put_request); -- cgit v1.3-8-gc7d7 From 13d1ad16d05eebb4db977eb955716b9da2c19fbd Mon Sep 17 00:00:00 2001 From: Ilya Dryomov Date: Wed, 27 Apr 2016 14:15:51 +0200 Subject: libceph: move message allocation out of ceph_osdc_alloc_request() The size of ->r_request and ->r_reply messages depends on the size of the object name (ceph_object_id), while the size of ceph_osd_request is fixed. Move message allocation into a separate function that would have to be called after ceph_object_id and ceph_object_locator (which is also going to become variable in size with RADOS namespaces) have been filled in: req = ceph_osdc_alloc_request(...); r_base_oid> r_base_oloc> ceph_osdc_alloc_messages(req); Signed-off-by: Ilya Dryomov --- drivers/block/rbd.c | 18 ++++++++- fs/ceph/addr.c | 8 ++++ fs/ceph/file.c | 7 ++++ include/linux/ceph/osd_client.h | 1 + net/ceph/osd_client.c | 88 +++++++++++++++++++++++------------------ 5 files changed, 82 insertions(+), 40 deletions(-) (limited to 'fs') diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c index c3089f32a392..bda4deade82e 100644 --- a/drivers/block/rbd.c +++ b/drivers/block/rbd.c @@ -1954,7 +1954,7 @@ static struct ceph_osd_request *rbd_osd_req_create( osd_req = ceph_osdc_alloc_request(osdc, snapc, num_ops, false, GFP_NOIO); if (!osd_req) - return NULL; /* ENOMEM */ + goto fail; if (op_type == OBJ_OP_WRITE || op_type == OBJ_OP_DISCARD) osd_req->r_flags = CEPH_OSD_FLAG_WRITE | CEPH_OSD_FLAG_ONDISK; @@ -1967,7 +1967,14 @@ static struct ceph_osd_request *rbd_osd_req_create( osd_req->r_base_oloc.pool = ceph_file_layout_pg_pool(rbd_dev->layout); ceph_oid_set_name(&osd_req->r_base_oid, obj_request->object_name); + if (ceph_osdc_alloc_messages(osd_req, GFP_NOIO)) + goto fail; + return osd_req; + +fail: + ceph_osdc_put_request(osd_req); + return NULL; } /* @@ -2003,7 +2010,7 @@ rbd_osd_req_create_copyup(struct rbd_obj_request *obj_request) osd_req = ceph_osdc_alloc_request(osdc, snapc, num_osd_ops, false, GFP_NOIO); if (!osd_req) - return NULL; /* ENOMEM */ + goto fail; osd_req->r_flags = CEPH_OSD_FLAG_WRITE | CEPH_OSD_FLAG_ONDISK; osd_req->r_callback = rbd_osd_req_callback; @@ -2012,7 +2019,14 @@ rbd_osd_req_create_copyup(struct rbd_obj_request *obj_request) osd_req->r_base_oloc.pool = ceph_file_layout_pg_pool(rbd_dev->layout); ceph_oid_set_name(&osd_req->r_base_oid, obj_request->object_name); + if (ceph_osdc_alloc_messages(osd_req, GFP_NOIO)) + goto fail; + return osd_req; + +fail: + ceph_osdc_put_request(osd_req); + return NULL; } diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c index 3e61fc8bb371..6fee7e0b8931 100644 --- a/fs/ceph/addr.c +++ b/fs/ceph/addr.c @@ -1762,6 +1762,10 @@ static int __ceph_pool_perm_get(struct ceph_inode_info *ci, u32 pool) "%llx.00000000", ci->i_vino.ino); rd_req->r_base_oid.name_len = strlen(rd_req->r_base_oid.name); + err = ceph_osdc_alloc_messages(rd_req, GFP_NOFS); + if (err) + goto out_unlock; + wr_req = ceph_osdc_alloc_request(&fsc->client->osdc, NULL, 1, false, GFP_NOFS); if (!wr_req) { @@ -1775,6 +1779,10 @@ static int __ceph_pool_perm_get(struct ceph_inode_info *ci, u32 pool) wr_req->r_base_oloc.pool = pool; wr_req->r_base_oid = rd_req->r_base_oid; + err = ceph_osdc_alloc_messages(wr_req, GFP_NOFS); + if (err) + goto out_unlock; + /* one page should be large enough for STAT data */ pages = ceph_alloc_page_vector(1, GFP_KERNEL); if (IS_ERR(pages)) { diff --git a/fs/ceph/file.c b/fs/ceph/file.c index a79f9269831e..5d46d106bbb7 100644 --- a/fs/ceph/file.c +++ b/fs/ceph/file.c @@ -717,6 +717,13 @@ static void ceph_aio_retry_work(struct work_struct *work) req->r_base_oloc = orig_req->r_base_oloc; req->r_base_oid = orig_req->r_base_oid; + ret = ceph_osdc_alloc_messages(req, GFP_NOFS); + if (ret) { + ceph_osdc_put_request(req); + req = orig_req; + goto out; + } + req->r_ops[0] = orig_req->r_ops[0]; osd_req_op_init(req, 1, CEPH_OSD_OP_STARTSYNC, 0); diff --git a/include/linux/ceph/osd_client.h b/include/linux/ceph/osd_client.h index cbf460927c42..66a1fcd5bff7 100644 --- a/include/linux/ceph/osd_client.h +++ b/include/linux/ceph/osd_client.h @@ -322,6 +322,7 @@ extern struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client * unsigned int num_ops, bool use_mempool, gfp_t gfp_flags); +int ceph_osdc_alloc_messages(struct ceph_osd_request *req, gfp_t gfp); extern void ceph_osdc_build_request(struct ceph_osd_request *req, u64 off, struct ceph_snap_context *snapc, diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c index ccb9539dc780..d66dacc9d0d4 100644 --- a/net/ceph/osd_client.c +++ b/net/ceph/osd_client.c @@ -369,8 +369,6 @@ struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc, gfp_t gfp_flags) { struct ceph_osd_request *req; - struct ceph_msg *msg; - size_t msg_size; if (use_mempool) { BUG_ON(num_ops > CEPH_OSD_SLAB_OPS); @@ -407,53 +405,59 @@ struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc, req->r_base_oloc.pool = -1; req->r_target_oloc.pool = -1; - msg_size = OSD_OPREPLY_FRONT_LEN; - if (num_ops > CEPH_OSD_SLAB_OPS) { - /* ceph_osd_op and rval */ - msg_size += (num_ops - CEPH_OSD_SLAB_OPS) * - (sizeof(struct ceph_osd_op) + 4); - } + dout("%s req %p\n", __func__, req); + return req; +} +EXPORT_SYMBOL(ceph_osdc_alloc_request); - /* create reply message */ - if (use_mempool) - msg = ceph_msgpool_get(&osdc->msgpool_op_reply, 0); - else - msg = ceph_msg_new(CEPH_MSG_OSD_OPREPLY, msg_size, - gfp_flags, true); - if (!msg) { - ceph_osdc_put_request(req); - return NULL; - } - req->r_reply = msg; +int ceph_osdc_alloc_messages(struct ceph_osd_request *req, gfp_t gfp) +{ + struct ceph_osd_client *osdc = req->r_osdc; + struct ceph_msg *msg; + int msg_size; + /* create request message */ msg_size = 4 + 4 + 4; /* client_inc, osdmap_epoch, flags */ msg_size += 4 + 4 + 4 + 8; /* mtime, reassert_version */ msg_size += 2 + 4 + 8 + 4 + 4; /* oloc */ msg_size += 1 + 8 + 4 + 4; /* pgid */ - msg_size += 4 + CEPH_MAX_OID_NAME_LEN; /* oid */ - msg_size += 2 + num_ops * sizeof(struct ceph_osd_op); + msg_size += 4 + req->r_base_oid.name_len; /* oid */ + msg_size += 2 + req->r_num_ops * sizeof(struct ceph_osd_op); msg_size += 8; /* snapid */ msg_size += 8; /* snap_seq */ - msg_size += 4 + 8 * (snapc ? snapc->num_snaps : 0); /* snaps */ + msg_size += 4 + 8 * (req->r_snapc ? req->r_snapc->num_snaps : 0); msg_size += 4; /* retry_attempt */ - /* create request message; allow space for oid */ - if (use_mempool) + if (req->r_mempool) msg = ceph_msgpool_get(&osdc->msgpool_op, 0); else - msg = ceph_msg_new(CEPH_MSG_OSD_OP, msg_size, gfp_flags, true); - if (!msg) { - ceph_osdc_put_request(req); - return NULL; - } + msg = ceph_msg_new(CEPH_MSG_OSD_OP, msg_size, gfp, true); + if (!msg) + return -ENOMEM; memset(msg->front.iov_base, 0, msg->front.iov_len); - req->r_request = msg; - return req; + /* create reply message */ + msg_size = OSD_OPREPLY_FRONT_LEN; + if (req->r_num_ops > CEPH_OSD_SLAB_OPS) { + /* ceph_osd_op and rval */ + msg_size += (req->r_num_ops - CEPH_OSD_SLAB_OPS) * + (sizeof(struct ceph_osd_op) + 4); + } + + if (req->r_mempool) + msg = ceph_msgpool_get(&osdc->msgpool_op_reply, 0); + else + msg = ceph_msg_new(CEPH_MSG_OSD_OPREPLY, msg_size, gfp, true); + if (!msg) + return -ENOMEM; + + req->r_reply = msg; + + return 0; } -EXPORT_SYMBOL(ceph_osdc_alloc_request); +EXPORT_SYMBOL(ceph_osdc_alloc_messages); static bool osd_req_opcode_valid(u16 opcode) { @@ -828,17 +832,17 @@ struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc, req = ceph_osdc_alloc_request(osdc, snapc, num_ops, use_mempool, GFP_NOFS); - if (!req) - return ERR_PTR(-ENOMEM); + if (!req) { + r = -ENOMEM; + goto fail; + } req->r_flags = flags; /* calculate max write size */ r = calc_layout(layout, off, plen, &objnum, &objoff, &objlen); - if (r < 0) { - ceph_osdc_put_request(req); - return ERR_PTR(r); - } + if (r) + goto fail; if (opcode == CEPH_OSD_OP_CREATE || opcode == CEPH_OSD_OP_DELETE) { osd_req_op_init(req, which, opcode, 0); @@ -864,7 +868,15 @@ struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc, "%llx.%08llx", vino.ino, objnum); req->r_base_oid.name_len = strlen(req->r_base_oid.name); + r = ceph_osdc_alloc_messages(req, GFP_NOFS); + if (r) + goto fail; + return req; + +fail: + ceph_osdc_put_request(req); + return ERR_PTR(r); } EXPORT_SYMBOL(ceph_osdc_new_request); -- cgit v1.3-8-gc7d7 From d30291b985d1854565d7f2c82a4457869d5265e8 Mon Sep 17 00:00:00 2001 From: Ilya Dryomov Date: Fri, 29 Apr 2016 19:54:20 +0200 Subject: libceph: variable-sized ceph_object_id Currently ceph_object_id can hold object names of up to 100 (CEPH_MAX_OID_NAME_LEN) characters. This is enough for all use cases, expect one - long rbd image names: - a format 1 header is named ".rbd" - an object that points to a format 2 header is named "rbd_id." We operate on these potentially long-named objects during rbd map, and, for format 1 images, during header refresh. (A format 2 header name is a small system-generated string.) Lift this 100 character limit by making ceph_object_id be able to point to an externally-allocated string. Apart from being able to work with almost arbitrarily-long named objects, this allows us to reduce the size of ceph_object_id from >100 bytes to 64 bytes. Signed-off-by: Ilya Dryomov --- drivers/block/rbd.c | 8 +++- fs/ceph/addr.c | 6 +-- fs/ceph/file.c | 2 +- fs/ceph/ioctl.c | 2 +- include/linux/ceph/osdmap.h | 62 ++++++++++++++++++------------ net/ceph/debugfs.c | 2 +- net/ceph/osd_client.c | 16 +++++--- net/ceph/osdmap.c | 93 ++++++++++++++++++++++++++++++++++++++++++++- 8 files changed, 150 insertions(+), 41 deletions(-) (limited to 'fs') diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c index bda4deade82e..3bf93a2a20f0 100644 --- a/drivers/block/rbd.c +++ b/drivers/block/rbd.c @@ -1965,7 +1965,9 @@ static struct ceph_osd_request *rbd_osd_req_create( osd_req->r_priv = obj_request; osd_req->r_base_oloc.pool = ceph_file_layout_pg_pool(rbd_dev->layout); - ceph_oid_set_name(&osd_req->r_base_oid, obj_request->object_name); + if (ceph_oid_aprintf(&osd_req->r_base_oid, GFP_NOIO, "%s", + obj_request->object_name)) + goto fail; if (ceph_osdc_alloc_messages(osd_req, GFP_NOIO)) goto fail; @@ -2017,7 +2019,9 @@ rbd_osd_req_create_copyup(struct rbd_obj_request *obj_request) osd_req->r_priv = obj_request; osd_req->r_base_oloc.pool = ceph_file_layout_pg_pool(rbd_dev->layout); - ceph_oid_set_name(&osd_req->r_base_oid, obj_request->object_name); + if (ceph_oid_aprintf(&osd_req->r_base_oid, GFP_NOIO, "%s", + obj_request->object_name)) + goto fail; if (ceph_osdc_alloc_messages(osd_req, GFP_NOIO)) goto fail; diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c index 6fee7e0b8931..6f28dd9bacb2 100644 --- a/fs/ceph/addr.c +++ b/fs/ceph/addr.c @@ -1758,9 +1758,7 @@ static int __ceph_pool_perm_get(struct ceph_inode_info *ci, u32 pool) rd_req->r_flags = CEPH_OSD_FLAG_READ; osd_req_op_init(rd_req, 0, CEPH_OSD_OP_STAT, 0); rd_req->r_base_oloc.pool = pool; - snprintf(rd_req->r_base_oid.name, sizeof(rd_req->r_base_oid.name), - "%llx.00000000", ci->i_vino.ino); - rd_req->r_base_oid.name_len = strlen(rd_req->r_base_oid.name); + ceph_oid_printf(&rd_req->r_base_oid, "%llx.00000000", ci->i_vino.ino); err = ceph_osdc_alloc_messages(rd_req, GFP_NOFS); if (err) @@ -1777,7 +1775,7 @@ static int __ceph_pool_perm_get(struct ceph_inode_info *ci, u32 pool) CEPH_OSD_FLAG_ACK | CEPH_OSD_FLAG_ONDISK; osd_req_op_init(wr_req, 0, CEPH_OSD_OP_CREATE, CEPH_OSD_OP_FLAG_EXCL); wr_req->r_base_oloc.pool = pool; - wr_req->r_base_oid = rd_req->r_base_oid; + ceph_oid_copy(&wr_req->r_base_oid, &rd_req->r_base_oid); err = ceph_osdc_alloc_messages(wr_req, GFP_NOFS); if (err) diff --git a/fs/ceph/file.c b/fs/ceph/file.c index 5d46d106bbb7..9d470397e249 100644 --- a/fs/ceph/file.c +++ b/fs/ceph/file.c @@ -715,7 +715,7 @@ static void ceph_aio_retry_work(struct work_struct *work) CEPH_OSD_FLAG_ONDISK | CEPH_OSD_FLAG_WRITE; req->r_base_oloc = orig_req->r_base_oloc; - req->r_base_oid = orig_req->r_base_oid; + ceph_oid_copy(&req->r_base_oid, &orig_req->r_base_oid); ret = ceph_osdc_alloc_messages(req, GFP_NOFS); if (ret) { diff --git a/fs/ceph/ioctl.c b/fs/ceph/ioctl.c index f851d8d70158..db296709784a 100644 --- a/fs/ceph/ioctl.c +++ b/fs/ceph/ioctl.c @@ -213,7 +213,7 @@ static long ceph_ioctl_get_dataloc(struct file *file, void __user *arg) ceph_ino(inode), dl.object_no); oloc.pool = ceph_file_layout_pg_pool(ci->i_layout); - ceph_oid_set_name(&oid, dl.object_name); + ceph_oid_printf(&oid, "%s", dl.object_name); r = ceph_oloc_oid_to_pg(osdc->osdmap, &oloc, &oid, &pgid); if (r < 0) { diff --git a/include/linux/ceph/osdmap.h b/include/linux/ceph/osdmap.h index e55c08bc3a96..777a29412706 100644 --- a/include/linux/ceph/osdmap.h +++ b/include/linux/ceph/osdmap.h @@ -64,11 +64,47 @@ struct ceph_object_locator { */ #define CEPH_MAX_OID_NAME_LEN 100 +/* + * 51-char inline_name is long enough for all cephfs and all but one + * rbd requests: in ".rbd"/"rbd_id." can be + * arbitrarily long (~PAGE_SIZE). It's done once during rbd map; all + * other rbd requests fit into inline_name. + * + * Makes ceph_object_id 64 bytes on 64-bit. + */ +#define CEPH_OID_INLINE_LEN 52 + +/* + * Both inline and external buffers have space for a NUL-terminator, + * which is carried around. It's not required though - RADOS object + * names don't have to be NUL-terminated and may contain NULs. + */ struct ceph_object_id { - char name[CEPH_MAX_OID_NAME_LEN]; + char *name; + char inline_name[CEPH_OID_INLINE_LEN]; int name_len; }; +static inline void ceph_oid_init(struct ceph_object_id *oid) +{ + oid->name = oid->inline_name; + oid->name_len = 0; +} + +static inline bool ceph_oid_empty(const struct ceph_object_id *oid) +{ + return oid->name == oid->inline_name && !oid->name_len; +} + +void ceph_oid_copy(struct ceph_object_id *dest, + const struct ceph_object_id *src); +__printf(2, 3) +void ceph_oid_printf(struct ceph_object_id *oid, const char *fmt, ...); +__printf(3, 4) +int ceph_oid_aprintf(struct ceph_object_id *oid, gfp_t gfp, + const char *fmt, ...); +void ceph_oid_destroy(struct ceph_object_id *oid); + struct ceph_pg_mapping { struct rb_node node; struct ceph_pg pgid; @@ -113,30 +149,6 @@ struct ceph_osdmap { int crush_scratch_ary[CEPH_PG_MAX_SIZE * 3]; }; -static inline void ceph_oid_set_name(struct ceph_object_id *oid, - const char *name) -{ - int len; - - len = strlen(name); - if (len > sizeof(oid->name)) { - WARN(1, "ceph_oid_set_name '%s' len %d vs %zu, truncating\n", - name, len, sizeof(oid->name)); - len = sizeof(oid->name); - } - - memcpy(oid->name, name, len); - oid->name_len = len; -} - -static inline void ceph_oid_copy(struct ceph_object_id *dest, - struct ceph_object_id *src) -{ - BUG_ON(src->name_len > sizeof(dest->name)); - memcpy(dest->name, src->name, src->name_len); - dest->name_len = src->name_len; -} - static inline int ceph_osd_exists(struct ceph_osdmap *map, int osd) { return osd >= 0 && osd < map->max_osd && diff --git a/net/ceph/debugfs.c b/net/ceph/debugfs.c index b902fbc7863e..6f8413293d15 100644 --- a/net/ceph/debugfs.c +++ b/net/ceph/debugfs.c @@ -161,7 +161,7 @@ static int osdc_show(struct seq_file *s, void *pp) req->r_osd ? req->r_osd->o_osd : -1, req->r_pgid.pool, req->r_pgid.seed); - seq_printf(s, "%.*s", req->r_base_oid.name_len, + seq_printf(s, "%*pE", req->r_base_oid.name_len, req->r_base_oid.name); if (req->r_reassert_version.epoch) diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c index 75e27bd3d372..95910aed8e2e 100644 --- a/net/ceph/osd_client.c +++ b/net/ceph/osd_client.c @@ -334,7 +334,10 @@ static void ceph_osdc_release_request(struct kref *kref) for (which = 0; which < req->r_num_ops; which++) osd_req_op_data_release(req, which); + ceph_oid_destroy(&req->r_base_oid); + ceph_oid_destroy(&req->r_target_oid); ceph_put_snap_context(req->r_snapc); + if (req->r_mempool) mempool_free(req, req->r_osdc->req_mempool); else if (req->r_num_ops <= CEPH_OSD_SLAB_OPS) @@ -401,7 +404,9 @@ struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc, INIT_LIST_HEAD(&req->r_req_lru_item); INIT_LIST_HEAD(&req->r_osd_item); + ceph_oid_init(&req->r_base_oid); req->r_base_oloc.pool = -1; + ceph_oid_init(&req->r_target_oid); req->r_target_oloc.pool = -1; dout("%s req %p\n", __func__, req); @@ -415,6 +420,8 @@ int ceph_osdc_alloc_messages(struct ceph_osd_request *req, gfp_t gfp) struct ceph_msg *msg; int msg_size; + WARN_ON(ceph_oid_empty(&req->r_base_oid)); + /* create request message */ msg_size = 4 + 4 + 4; /* client_inc, osdmap_epoch, flags */ msg_size += 4 + 4 + 4 + 8; /* mtime, reassert_version */ @@ -859,10 +866,7 @@ struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc, } req->r_base_oloc.pool = ceph_file_layout_pg_pool(*layout); - - snprintf(req->r_base_oid.name, sizeof(req->r_base_oid.name), - "%llx.%08llx", vino.ino, objnum); - req->r_base_oid.name_len = strlen(req->r_base_oid.name); + ceph_oid_printf(&req->r_base_oid, "%llx.%08llx", vino.ino, objnum); r = ceph_osdc_alloc_messages(req, GFP_NOFS); if (r) @@ -1410,7 +1414,7 @@ static int __calc_request_pg(struct ceph_osdmap *osdmap, req->r_target_oloc = req->r_base_oloc; /* struct */ need_check_tiering = true; } - if (req->r_target_oid.name_len == 0) { + if (ceph_oid_empty(&req->r_target_oid)) { ceph_oid_copy(&req->r_target_oid, &req->r_base_oid); need_check_tiering = true; } @@ -2501,7 +2505,7 @@ void ceph_osdc_build_request(struct ceph_osd_request *req, u64 off, /* oid */ ceph_encode_32(&p, req->r_base_oid.name_len); memcpy(p, req->r_base_oid.name, req->r_base_oid.name_len); - dout("oid '%.*s' len %d\n", req->r_base_oid.name_len, + dout("oid %*pE len %d\n", req->r_base_oid.name_len, req->r_base_oid.name, req->r_base_oid.name_len); p += req->r_base_oid.name_len; diff --git a/net/ceph/osdmap.c b/net/ceph/osdmap.c index 243574c8cf33..4668b871ca47 100644 --- a/net/ceph/osdmap.c +++ b/net/ceph/osdmap.c @@ -1381,8 +1381,99 @@ bad: return ERR_PTR(err); } +void ceph_oid_copy(struct ceph_object_id *dest, + const struct ceph_object_id *src) +{ + WARN_ON(!ceph_oid_empty(dest)); + + if (src->name != src->inline_name) { + /* very rare, see ceph_object_id definition */ + dest->name = kmalloc(src->name_len + 1, + GFP_NOIO | __GFP_NOFAIL); + } + memcpy(dest->name, src->name, src->name_len + 1); + dest->name_len = src->name_len; +} +EXPORT_SYMBOL(ceph_oid_copy); +static __printf(2, 0) +int oid_printf_vargs(struct ceph_object_id *oid, const char *fmt, va_list ap) +{ + int len; + + WARN_ON(!ceph_oid_empty(oid)); + + len = vsnprintf(oid->inline_name, sizeof(oid->inline_name), fmt, ap); + if (len >= sizeof(oid->inline_name)) + return len; + + oid->name_len = len; + return 0; +} + +/* + * If oid doesn't fit into inline buffer, BUG. + */ +void ceph_oid_printf(struct ceph_object_id *oid, const char *fmt, ...) +{ + va_list ap; + + va_start(ap, fmt); + BUG_ON(oid_printf_vargs(oid, fmt, ap)); + va_end(ap); +} +EXPORT_SYMBOL(ceph_oid_printf); + +static __printf(3, 0) +int oid_aprintf_vargs(struct ceph_object_id *oid, gfp_t gfp, + const char *fmt, va_list ap) +{ + va_list aq; + int len; + + va_copy(aq, ap); + len = oid_printf_vargs(oid, fmt, aq); + va_end(aq); + + if (len) { + char *external_name; + + external_name = kmalloc(len + 1, gfp); + if (!external_name) + return -ENOMEM; + + oid->name = external_name; + WARN_ON(vsnprintf(oid->name, len + 1, fmt, ap) != len); + oid->name_len = len; + } + + return 0; +} + +/* + * If oid doesn't fit into inline buffer, allocate. + */ +int ceph_oid_aprintf(struct ceph_object_id *oid, gfp_t gfp, + const char *fmt, ...) +{ + va_list ap; + int ret; + + va_start(ap, fmt); + ret = oid_aprintf_vargs(oid, gfp, fmt, ap); + va_end(ap); + + return ret; +} +EXPORT_SYMBOL(ceph_oid_aprintf); + +void ceph_oid_destroy(struct ceph_object_id *oid) +{ + if (oid->name != oid->inline_name) + kfree(oid->name); +} +EXPORT_SYMBOL(ceph_oid_destroy); /* * calculate file layout from given offset, length. @@ -1474,7 +1565,7 @@ int ceph_oloc_oid_to_pg(struct ceph_osdmap *osdmap, pg_out->seed = ceph_str_hash(pi->object_hash, oid->name, oid->name_len); - dout("%s '%.*s' pgid %llu.%x\n", __func__, oid->name_len, oid->name, + dout("%s %*pE pgid %llu.%x\n", __func__, oid->name_len, oid->name, pg_out->pool, pg_out->seed); return 0; } -- cgit v1.3-8-gc7d7 From fcd00b68bbe2bf5606cb45c2cd4a250a390bcc1f Mon Sep 17 00:00:00 2001 From: Ilya Dryomov Date: Thu, 28 Apr 2016 16:07:22 +0200 Subject: libceph: DEFINE_RB_FUNCS macro Given struct foo { u64 id; struct rb_node bar_node; }; generate insert_bar(), erase_bar() and lookup_bar() functions with DEFINE_RB_FUNCS(bar, struct foo, id, bar_node) The key is assumed to be an integer (u64, int, etc), compared with < and >. nodefld has to be initialized with RB_CLEAR_NODE(). Start using it for MDS, MON and OSD requests and OSD sessions. Signed-off-by: Ilya Dryomov --- fs/ceph/mds_client.c | 54 ++++++------------------ include/linux/ceph/libceph.h | 57 ++++++++++++++++++++++++++ net/ceph/mon_client.c | 52 ++++-------------------- net/ceph/osd_client.c | 97 +++++--------------------------------------- 4 files changed, 88 insertions(+), 172 deletions(-) (limited to 'fs') diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c index 85b8517f17a0..cff85af425d4 100644 --- a/fs/ceph/mds_client.c +++ b/fs/ceph/mds_client.c @@ -567,51 +567,23 @@ void ceph_mdsc_release_request(struct kref *kref) kfree(req); } +DEFINE_RB_FUNCS(request, struct ceph_mds_request, r_tid, r_node) + /* * lookup session, bump ref if found. * * called under mdsc->mutex. */ -static struct ceph_mds_request *__lookup_request(struct ceph_mds_client *mdsc, - u64 tid) +static struct ceph_mds_request * +lookup_get_request(struct ceph_mds_client *mdsc, u64 tid) { struct ceph_mds_request *req; - struct rb_node *n = mdsc->request_tree.rb_node; - - while (n) { - req = rb_entry(n, struct ceph_mds_request, r_node); - if (tid < req->r_tid) - n = n->rb_left; - else if (tid > req->r_tid) - n = n->rb_right; - else { - ceph_mdsc_get_request(req); - return req; - } - } - return NULL; -} -static void __insert_request(struct ceph_mds_client *mdsc, - struct ceph_mds_request *new) -{ - struct rb_node **p = &mdsc->request_tree.rb_node; - struct rb_node *parent = NULL; - struct ceph_mds_request *req = NULL; + req = lookup_request(&mdsc->request_tree, tid); + if (req) + ceph_mdsc_get_request(req); - while (*p) { - parent = *p; - req = rb_entry(parent, struct ceph_mds_request, r_node); - if (new->r_tid < req->r_tid) - p = &(*p)->rb_left; - else if (new->r_tid > req->r_tid) - p = &(*p)->rb_right; - else - BUG(); - } - - rb_link_node(&new->r_node, parent, p); - rb_insert_color(&new->r_node, &mdsc->request_tree); + return req; } /* @@ -630,7 +602,7 @@ static void __register_request(struct ceph_mds_client *mdsc, req->r_num_caps); dout("__register_request %p tid %lld\n", req, req->r_tid); ceph_mdsc_get_request(req); - __insert_request(mdsc, req); + insert_request(&mdsc->request_tree, req); req->r_uid = current_fsuid(); req->r_gid = current_fsgid(); @@ -663,8 +635,7 @@ static void __unregister_request(struct ceph_mds_client *mdsc, } } - rb_erase(&req->r_node, &mdsc->request_tree); - RB_CLEAR_NODE(&req->r_node); + erase_request(&mdsc->request_tree, req); if (req->r_unsafe_dir && req->r_got_unsafe) { struct ceph_inode_info *ci = ceph_inode(req->r_unsafe_dir); @@ -1722,6 +1693,7 @@ ceph_mdsc_create_request(struct ceph_mds_client *mdsc, int op, int mode) INIT_LIST_HEAD(&req->r_unsafe_target_item); req->r_fmode = -1; kref_init(&req->r_kref); + RB_CLEAR_NODE(&req->r_node); INIT_LIST_HEAD(&req->r_wait); init_completion(&req->r_completion); init_completion(&req->r_safe_completion); @@ -2414,7 +2386,7 @@ static void handle_reply(struct ceph_mds_session *session, struct ceph_msg *msg) /* get request, session */ tid = le64_to_cpu(msg->hdr.tid); mutex_lock(&mdsc->mutex); - req = __lookup_request(mdsc, tid); + req = lookup_get_request(mdsc, tid); if (!req) { dout("handle_reply on unknown tid %llu\n", tid); mutex_unlock(&mdsc->mutex); @@ -2604,7 +2576,7 @@ static void handle_forward(struct ceph_mds_client *mdsc, fwd_seq = ceph_decode_32(&p); mutex_lock(&mdsc->mutex); - req = __lookup_request(mdsc, tid); + req = lookup_get_request(mdsc, tid); if (!req) { dout("forward tid %llu to mds%d - req dne\n", tid, next_mds); goto out; /* dup reply? */ diff --git a/include/linux/ceph/libceph.h b/include/linux/ceph/libceph.h index db92a8d4926e..690985daad1c 100644 --- a/include/linux/ceph/libceph.h +++ b/include/linux/ceph/libceph.h @@ -180,6 +180,63 @@ static inline int calc_pages_for(u64 off, u64 len) (off >> PAGE_SHIFT); } +/* + * These are not meant to be generic - an integer key is assumed. + */ +#define DEFINE_RB_INSDEL_FUNCS(name, type, keyfld, nodefld) \ +static void insert_##name(struct rb_root *root, type *t) \ +{ \ + struct rb_node **n = &root->rb_node; \ + struct rb_node *parent = NULL; \ + \ + BUG_ON(!RB_EMPTY_NODE(&t->nodefld)); \ + \ + while (*n) { \ + type *cur = rb_entry(*n, type, nodefld); \ + \ + parent = *n; \ + if (t->keyfld < cur->keyfld) \ + n = &(*n)->rb_left; \ + else if (t->keyfld > cur->keyfld) \ + n = &(*n)->rb_right; \ + else \ + BUG(); \ + } \ + \ + rb_link_node(&t->nodefld, parent, n); \ + rb_insert_color(&t->nodefld, root); \ +} \ +static void erase_##name(struct rb_root *root, type *t) \ +{ \ + BUG_ON(RB_EMPTY_NODE(&t->nodefld)); \ + rb_erase(&t->nodefld, root); \ + RB_CLEAR_NODE(&t->nodefld); \ +} + +#define DEFINE_RB_LOOKUP_FUNC(name, type, keyfld, nodefld) \ +static type *lookup_##name(struct rb_root *root, \ + typeof(((type *)0)->keyfld) key) \ +{ \ + struct rb_node *n = root->rb_node; \ + \ + while (n) { \ + type *cur = rb_entry(n, type, nodefld); \ + \ + if (key < cur->keyfld) \ + n = n->rb_left; \ + else if (key > cur->keyfld) \ + n = n->rb_right; \ + else \ + return cur; \ + } \ + \ + return NULL; \ +} + +#define DEFINE_RB_FUNCS(name, type, keyfld, nodefld) \ +DEFINE_RB_INSDEL_FUNCS(name, type, keyfld, nodefld) \ +DEFINE_RB_LOOKUP_FUNC(name, type, keyfld, nodefld) + extern struct kmem_cache *ceph_inode_cachep; extern struct kmem_cache *ceph_cap_cachep; extern struct kmem_cache *ceph_cap_flush_cachep; diff --git a/net/ceph/mon_client.c b/net/ceph/mon_client.c index 3dfafdad92aa..a426a4b03e75 100644 --- a/net/ceph/mon_client.c +++ b/net/ceph/mon_client.c @@ -478,45 +478,7 @@ out: /* * generic requests (currently statfs, mon_get_version) */ -static struct ceph_mon_generic_request *__lookup_generic_req( - struct ceph_mon_client *monc, u64 tid) -{ - struct ceph_mon_generic_request *req; - struct rb_node *n = monc->generic_request_tree.rb_node; - - while (n) { - req = rb_entry(n, struct ceph_mon_generic_request, node); - if (tid < req->tid) - n = n->rb_left; - else if (tid > req->tid) - n = n->rb_right; - else - return req; - } - return NULL; -} - -static void __insert_generic_request(struct ceph_mon_client *monc, - struct ceph_mon_generic_request *new) -{ - struct rb_node **p = &monc->generic_request_tree.rb_node; - struct rb_node *parent = NULL; - struct ceph_mon_generic_request *req = NULL; - - while (*p) { - parent = *p; - req = rb_entry(parent, struct ceph_mon_generic_request, node); - if (new->tid < req->tid) - p = &(*p)->rb_left; - else if (new->tid > req->tid) - p = &(*p)->rb_right; - else - BUG(); - } - - rb_link_node(&new->node, parent, p); - rb_insert_color(&new->node, &monc->generic_request_tree); -} +DEFINE_RB_FUNCS(generic_request, struct ceph_mon_generic_request, tid, node) static void release_generic_request(struct kref *kref) { @@ -551,7 +513,7 @@ static struct ceph_msg *get_generic_reply(struct ceph_connection *con, struct ceph_msg *m; mutex_lock(&monc->mutex); - req = __lookup_generic_req(monc, tid); + req = lookup_generic_request(&monc->generic_request_tree, tid); if (!req) { dout("get_generic_reply %lld dne\n", tid); *skip = 1; @@ -578,14 +540,14 @@ static int __do_generic_request(struct ceph_mon_client *monc, u64 tid, /* register request */ req->tid = tid != 0 ? tid : ++monc->last_tid; req->request->hdr.tid = cpu_to_le64(req->tid); - __insert_generic_request(monc, req); + insert_generic_request(&monc->generic_request_tree, req); ceph_con_send(&monc->con, ceph_msg_get(req->request)); mutex_unlock(&monc->mutex); err = wait_for_completion_interruptible(&req->completion); mutex_lock(&monc->mutex); - rb_erase(&req->node, &monc->generic_request_tree); + erase_generic_request(&monc->generic_request_tree, req); if (!err) err = req->result; @@ -619,7 +581,7 @@ static void handle_statfs_reply(struct ceph_mon_client *monc, dout("handle_statfs_reply %p tid %llu\n", msg, tid); mutex_lock(&monc->mutex); - req = __lookup_generic_req(monc, tid); + req = lookup_generic_request(&monc->generic_request_tree, tid); if (req) { *(struct ceph_statfs *)req->buf = reply->st; req->result = 0; @@ -651,6 +613,7 @@ int ceph_monc_do_statfs(struct ceph_mon_client *monc, struct ceph_statfs *buf) return -ENOMEM; kref_init(&req->kref); + RB_CLEAR_NODE(&req->node); req->buf = buf; init_completion(&req->completion); @@ -696,7 +659,7 @@ static void handle_get_version_reply(struct ceph_mon_client *monc, goto bad; mutex_lock(&monc->mutex); - req = __lookup_generic_req(monc, handle); + req = lookup_generic_request(&monc->generic_request_tree, handle); if (req) { *(u64 *)req->buf = ceph_decode_64(&p); req->result = 0; @@ -732,6 +695,7 @@ int ceph_monc_do_get_version(struct ceph_mon_client *monc, const char *what, return -ENOMEM; kref_init(&req->kref); + RB_CLEAR_NODE(&req->node); req->buf = newest; init_completion(&req->completion); diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c index c423e11d6857..8256051ed88f 100644 --- a/net/ceph/osd_client.c +++ b/net/ceph/osd_client.c @@ -875,45 +875,7 @@ EXPORT_SYMBOL(ceph_osdc_new_request); /* * We keep osd requests in an rbtree, sorted by ->r_tid. */ -static void __insert_request(struct ceph_osd_client *osdc, - struct ceph_osd_request *new) -{ - struct rb_node **p = &osdc->requests.rb_node; - struct rb_node *parent = NULL; - struct ceph_osd_request *req = NULL; - - while (*p) { - parent = *p; - req = rb_entry(parent, struct ceph_osd_request, r_node); - if (new->r_tid < req->r_tid) - p = &(*p)->rb_left; - else if (new->r_tid > req->r_tid) - p = &(*p)->rb_right; - else - BUG(); - } - - rb_link_node(&new->r_node, parent, p); - rb_insert_color(&new->r_node, &osdc->requests); -} - -static struct ceph_osd_request *__lookup_request(struct ceph_osd_client *osdc, - u64 tid) -{ - struct ceph_osd_request *req; - struct rb_node *n = osdc->requests.rb_node; - - while (n) { - req = rb_entry(n, struct ceph_osd_request, r_node); - if (tid < req->r_tid) - n = n->rb_left; - else if (tid > req->r_tid) - n = n->rb_right; - else - return req; - } - return NULL; -} +DEFINE_RB_FUNCS(request, struct ceph_osd_request, r_tid, r_node) static struct ceph_osd_request * __lookup_request_ge(struct ceph_osd_client *osdc, @@ -1101,6 +1063,8 @@ static void put_osd(struct ceph_osd *osd) } } +DEFINE_RB_FUNCS(osd, struct ceph_osd, o_osd, o_node) + /* * remove an osd from our map */ @@ -1111,8 +1075,7 @@ static void __remove_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd) WARN_ON(!list_empty(&osd->o_linger_requests)); list_del_init(&osd->o_osd_lru); - rb_erase(&osd->o_node, &osdc->osds); - RB_CLEAR_NODE(&osd->o_node); + erase_osd(&osdc->osds, osd); } static void remove_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd) @@ -1188,45 +1151,6 @@ static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd) return 0; } -static void __insert_osd(struct ceph_osd_client *osdc, struct ceph_osd *new) -{ - struct rb_node **p = &osdc->osds.rb_node; - struct rb_node *parent = NULL; - struct ceph_osd *osd = NULL; - - dout("__insert_osd %p osd%d\n", new, new->o_osd); - while (*p) { - parent = *p; - osd = rb_entry(parent, struct ceph_osd, o_node); - if (new->o_osd < osd->o_osd) - p = &(*p)->rb_left; - else if (new->o_osd > osd->o_osd) - p = &(*p)->rb_right; - else - BUG(); - } - - rb_link_node(&new->o_node, parent, p); - rb_insert_color(&new->o_node, &osdc->osds); -} - -static struct ceph_osd *__lookup_osd(struct ceph_osd_client *osdc, int o) -{ - struct ceph_osd *osd; - struct rb_node *n = osdc->osds.rb_node; - - while (n) { - osd = rb_entry(n, struct ceph_osd, o_node); - if (o < osd->o_osd) - n = n->rb_left; - else if (o > osd->o_osd) - n = n->rb_right; - else - return osd; - } - return NULL; -} - static void __schedule_osd_timeout(struct ceph_osd_client *osdc) { schedule_delayed_work(&osdc->timeout_work, @@ -1248,7 +1172,7 @@ static void __register_request(struct ceph_osd_client *osdc, req->r_tid = ++osdc->last_tid; req->r_request->hdr.tid = cpu_to_le64(req->r_tid); dout("__register_request %p tid %lld\n", req, req->r_tid); - __insert_request(osdc, req); + insert_request(&osdc->requests, req); ceph_osdc_get_request(req); osdc->num_requests++; if (osdc->num_requests == 1) { @@ -1270,8 +1194,7 @@ static void __unregister_request(struct ceph_osd_client *osdc, } dout("__unregister_request %p tid %lld\n", req, req->r_tid); - rb_erase(&req->r_node, &osdc->requests); - RB_CLEAR_NODE(&req->r_node); + erase_request(&osdc->requests, req); osdc->num_requests--; if (req->r_osd) { @@ -1482,7 +1405,7 @@ static int __map_request(struct ceph_osd_client *osdc, req->r_osd = NULL; } - req->r_osd = __lookup_osd(osdc, o); + req->r_osd = lookup_osd(&osdc->osds, o); if (!req->r_osd && o >= 0) { err = -ENOMEM; req->r_osd = create_osd(osdc, o); @@ -1492,7 +1415,7 @@ static int __map_request(struct ceph_osd_client *osdc, } dout("map_request osd %p is osd%d\n", req->r_osd, o); - __insert_osd(osdc, req->r_osd); + insert_osd(&osdc->osds, req->r_osd); ceph_con_open(&req->r_osd->o_con, CEPH_ENTITY_TYPE_OSD, o, @@ -1822,7 +1745,7 @@ static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg) /* lookup */ down_read(&osdc->map_sem); mutex_lock(&osdc->request_mutex); - req = __lookup_request(osdc, tid); + req = lookup_request(&osdc->requests, tid); if (req == NULL) { dout("handle_reply tid %llu dne\n", tid); goto bad_mutex; @@ -2880,7 +2803,7 @@ static struct ceph_msg *get_reply(struct ceph_connection *con, tid = le64_to_cpu(hdr->tid); mutex_lock(&osdc->request_mutex); - req = __lookup_request(osdc, tid); + req = lookup_request(&osdc->requests, tid); if (!req) { dout("%s osd%d tid %llu unknown, skipping\n", __func__, osd->o_osd, tid); -- cgit v1.3-8-gc7d7 From d9591f5e28686277d9312d3c7422faf1368b305e Mon Sep 17 00:00:00 2001 From: Ilya Dryomov Date: Thu, 28 Apr 2016 16:07:22 +0200 Subject: libceph: rename ceph_oloc_oid_to_pg() Rename ceph_oloc_oid_to_pg() to ceph_object_locator_to_pg(). Emphasise that returned is raw PG and return -ENOENT instead of -EIO if the pool doesn't exist. Signed-off-by: Ilya Dryomov --- fs/ceph/ioctl.c | 2 +- include/linux/ceph/osdmap.h | 9 ++++----- net/ceph/osd_client.c | 4 ++-- net/ceph/osdmap.c | 31 ++++++++++++++++--------------- 4 files changed, 23 insertions(+), 23 deletions(-) (limited to 'fs') diff --git a/fs/ceph/ioctl.c b/fs/ceph/ioctl.c index db296709784a..cca7fff22725 100644 --- a/fs/ceph/ioctl.c +++ b/fs/ceph/ioctl.c @@ -215,7 +215,7 @@ static long ceph_ioctl_get_dataloc(struct file *file, void __user *arg) oloc.pool = ceph_file_layout_pg_pool(ci->i_layout); ceph_oid_printf(&oid, "%s", dl.object_name); - r = ceph_oloc_oid_to_pg(osdc->osdmap, &oloc, &oid, &pgid); + r = ceph_object_locator_to_pg(osdc->osdmap, &oid, &oloc, &pgid); if (r < 0) { up_read(&osdc->map_sem); return r; diff --git a/include/linux/ceph/osdmap.h b/include/linux/ceph/osdmap.h index ce7a41a182d4..b70440c05b49 100644 --- a/include/linux/ceph/osdmap.h +++ b/include/linux/ceph/osdmap.h @@ -213,11 +213,10 @@ extern int ceph_calc_file_object_mapping(struct ceph_file_layout *layout, u64 off, u64 len, u64 *bno, u64 *oxoff, u64 *oxlen); -/* calculate mapping of object to a placement group */ -extern int ceph_oloc_oid_to_pg(struct ceph_osdmap *osdmap, - struct ceph_object_locator *oloc, - struct ceph_object_id *oid, - struct ceph_pg *pg_out); +int ceph_object_locator_to_pg(struct ceph_osdmap *osdmap, + struct ceph_object_id *oid, + struct ceph_object_locator *oloc, + struct ceph_pg *raw_pgid); extern int ceph_calc_pg_acting(struct ceph_osdmap *osdmap, struct ceph_pg pgid, diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c index 8256051ed88f..cb9f1953f5fb 100644 --- a/net/ceph/osd_client.c +++ b/net/ceph/osd_client.c @@ -1324,8 +1324,8 @@ static int __calc_request_pg(struct ceph_osdmap *osdmap, /* !pi is caught in ceph_oloc_oid_to_pg() */ } - return ceph_oloc_oid_to_pg(osdmap, &req->r_target_oloc, - &req->r_target_oid, pg_out); + return ceph_object_locator_to_pg(osdmap, &req->r_target_oid, + &req->r_target_oloc, pg_out); } static void __enqueue_request(struct ceph_osd_request *req) diff --git a/net/ceph/osdmap.c b/net/ceph/osdmap.c index 9a0cc072a909..6267839cb246 100644 --- a/net/ceph/osdmap.c +++ b/net/ceph/osdmap.c @@ -1545,30 +1545,31 @@ invalid: EXPORT_SYMBOL(ceph_calc_file_object_mapping); /* - * Calculate mapping of a (oloc, oid) pair to a PG. Should only be - * called with target's (oloc, oid), since tiering isn't taken into - * account. + * Map an object into a PG. + * + * Should only be called with target_oid and target_oloc (as opposed to + * base_oid and base_oloc), since tiering isn't taken into account. */ -int ceph_oloc_oid_to_pg(struct ceph_osdmap *osdmap, - struct ceph_object_locator *oloc, - struct ceph_object_id *oid, - struct ceph_pg *pg_out) +int ceph_object_locator_to_pg(struct ceph_osdmap *osdmap, + struct ceph_object_id *oid, + struct ceph_object_locator *oloc, + struct ceph_pg *raw_pgid) { struct ceph_pg_pool_info *pi; - pi = __lookup_pg_pool(&osdmap->pg_pools, oloc->pool); + pi = ceph_pg_pool_by_id(osdmap, oloc->pool); if (!pi) - return -EIO; + return -ENOENT; - pg_out->pool = oloc->pool; - pg_out->seed = ceph_str_hash(pi->object_hash, oid->name, - oid->name_len); + raw_pgid->pool = oloc->pool; + raw_pgid->seed = ceph_str_hash(pi->object_hash, oid->name, + oid->name_len); - dout("%s %*pE pgid %llu.%x\n", __func__, oid->name_len, oid->name, - pg_out->pool, pg_out->seed); + dout("%s %*pE -> raw_pgid %llu.%x\n", __func__, oid->name_len, + oid->name, raw_pgid->pool, raw_pgid->seed); return 0; } -EXPORT_SYMBOL(ceph_oloc_oid_to_pg); +EXPORT_SYMBOL(ceph_object_locator_to_pg); static int do_crush(struct ceph_osdmap *map, int ruleno, int x, int *result, int result_max, -- cgit v1.3-8-gc7d7 From f81f16339a05775df600b2ff75a79be1864975c1 Mon Sep 17 00:00:00 2001 From: Ilya Dryomov Date: Thu, 28 Apr 2016 16:07:23 +0200 Subject: libceph: rename ceph_calc_pg_primary() Rename ceph_calc_pg_primary() to ceph_pg_to_acting_primary() to emphasise that it returns acting primary. Signed-off-by: Ilya Dryomov --- fs/ceph/ioctl.c | 2 +- include/linux/ceph/osdmap.h | 4 ++-- net/ceph/osdmap.c | 9 +++++---- 3 files changed, 8 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/ceph/ioctl.c b/fs/ceph/ioctl.c index cca7fff22725..1831ad6cf066 100644 --- a/fs/ceph/ioctl.c +++ b/fs/ceph/ioctl.c @@ -221,7 +221,7 @@ static long ceph_ioctl_get_dataloc(struct file *file, void __user *arg) return r; } - dl.osd = ceph_calc_pg_primary(osdc->osdmap, pgid); + dl.osd = ceph_pg_to_acting_primary(osdc->osdmap, &pgid); if (dl.osd >= 0) { struct ceph_entity_addr *a = ceph_osd_addr(osdc->osdmap, dl.osd); diff --git a/include/linux/ceph/osdmap.h b/include/linux/ceph/osdmap.h index 942189d311e0..3fd978a1639b 100644 --- a/include/linux/ceph/osdmap.h +++ b/include/linux/ceph/osdmap.h @@ -236,8 +236,8 @@ void ceph_pg_to_up_acting_osds(struct ceph_osdmap *osdmap, const struct ceph_pg *raw_pgid, struct ceph_osds *up, struct ceph_osds *acting); -extern int ceph_calc_pg_primary(struct ceph_osdmap *osdmap, - struct ceph_pg pgid); +int ceph_pg_to_acting_primary(struct ceph_osdmap *osdmap, + const struct ceph_pg *raw_pgid); extern struct ceph_pg_pool_info *ceph_pg_pool_by_id(struct ceph_osdmap *map, u64 id); diff --git a/net/ceph/osdmap.c b/net/ceph/osdmap.c index f5fc8fc63879..656384a8fd1e 100644 --- a/net/ceph/osdmap.c +++ b/net/ceph/osdmap.c @@ -1896,13 +1896,14 @@ out: } /* - * Return primary osd for given pgid, or -1 if none. + * Return acting primary for given PG, or -1 if none. */ -int ceph_calc_pg_primary(struct ceph_osdmap *osdmap, struct ceph_pg pgid) +int ceph_pg_to_acting_primary(struct ceph_osdmap *osdmap, + const struct ceph_pg *raw_pgid) { struct ceph_osds up, acting; - ceph_pg_to_up_acting_osds(osdmap, &pgid, &up, &acting); + ceph_pg_to_up_acting_osds(osdmap, raw_pgid, &up, &acting); return acting.primary; } -EXPORT_SYMBOL(ceph_calc_pg_primary); +EXPORT_SYMBOL(ceph_pg_to_acting_primary); -- cgit v1.3-8-gc7d7 From 63244fa123a755e4bbaee03022b68613c71d1332 Mon Sep 17 00:00:00 2001 From: Ilya Dryomov Date: Thu, 28 Apr 2016 16:07:23 +0200 Subject: libceph: introduce ceph_osd_request_target, calc_target() Introduce ceph_osd_request_target, containing all mapping-related fields of ceph_osd_request and calc_target() for calculating mappings and populating it. Signed-off-by: Ilya Dryomov --- fs/ceph/addr.c | 2 +- fs/ceph/file.c | 2 +- include/linux/ceph/osd_client.h | 23 ++++++ include/linux/ceph/osdmap.h | 34 +++++++++ include/linux/ceph/rados.h | 5 ++ net/ceph/osd_client.c | 157 +++++++++++++++++++++++++++++++++++++++- net/ceph/osdmap.c | 121 +++++++++++++++++++++++++++++++ 7 files changed, 340 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c index 6f28dd9bacb2..c5d75486823b 100644 --- a/fs/ceph/addr.c +++ b/fs/ceph/addr.c @@ -1774,7 +1774,7 @@ static int __ceph_pool_perm_get(struct ceph_inode_info *ci, u32 pool) wr_req->r_flags = CEPH_OSD_FLAG_WRITE | CEPH_OSD_FLAG_ACK | CEPH_OSD_FLAG_ONDISK; osd_req_op_init(wr_req, 0, CEPH_OSD_OP_CREATE, CEPH_OSD_OP_FLAG_EXCL); - wr_req->r_base_oloc.pool = pool; + ceph_oloc_copy(&wr_req->r_base_oloc, &rd_req->r_base_oloc); ceph_oid_copy(&wr_req->r_base_oid, &rd_req->r_base_oid); err = ceph_osdc_alloc_messages(wr_req, GFP_NOFS); diff --git a/fs/ceph/file.c b/fs/ceph/file.c index 9d470397e249..36b4a41dfa67 100644 --- a/fs/ceph/file.c +++ b/fs/ceph/file.c @@ -714,7 +714,7 @@ static void ceph_aio_retry_work(struct work_struct *work) req->r_flags = CEPH_OSD_FLAG_ORDERSNAP | CEPH_OSD_FLAG_ONDISK | CEPH_OSD_FLAG_WRITE; - req->r_base_oloc = orig_req->r_base_oloc; + ceph_oloc_copy(&req->r_base_oloc, &orig_req->r_base_oloc); ceph_oid_copy(&req->r_base_oid, &orig_req->r_base_oid); ret = ceph_osdc_alloc_messages(req, GFP_NOFS); diff --git a/include/linux/ceph/osd_client.h b/include/linux/ceph/osd_client.h index 63854a8df183..48806ee4488d 100644 --- a/include/linux/ceph/osd_client.h +++ b/include/linux/ceph/osd_client.h @@ -24,6 +24,8 @@ typedef void (*ceph_osdc_callback_t)(struct ceph_osd_request *, struct ceph_msg *); typedef void (*ceph_osdc_unsafe_callback_t)(struct ceph_osd_request *, bool); +#define CEPH_HOMELESS_OSD -1 + /* a given osd we're communicating with */ struct ceph_osd { atomic_t o_ref; @@ -118,6 +120,27 @@ struct ceph_osd_req_op { }; }; +struct ceph_osd_request_target { + struct ceph_object_id base_oid; + struct ceph_object_locator base_oloc; + struct ceph_object_id target_oid; + struct ceph_object_locator target_oloc; + + struct ceph_pg pgid; + u32 pg_num; + u32 pg_num_mask; + struct ceph_osds acting; + struct ceph_osds up; + int size; + int min_size; + bool sort_bitwise; + + unsigned int flags; /* CEPH_OSD_FLAG_* */ + bool paused; + + int osd; +}; + /* an in-flight request */ struct ceph_osd_request { u64 r_tid; /* unique for this client */ diff --git a/include/linux/ceph/osdmap.h b/include/linux/ceph/osdmap.h index 989294d0b8d2..420bb7968b25 100644 --- a/include/linux/ceph/osdmap.h +++ b/include/linux/ceph/osdmap.h @@ -28,6 +28,7 @@ int ceph_pg_compare(const struct ceph_pg *lhs, const struct ceph_pg *rhs); #define CEPH_POOL_FLAG_HASHPSPOOL (1ULL << 0) /* hash pg seed and pool id together */ +#define CEPH_POOL_FLAG_FULL (1ULL << 1) /* pool is full */ struct ceph_pg_pool_info { struct rb_node node; @@ -62,6 +63,22 @@ struct ceph_object_locator { s64 pool; }; +static inline void ceph_oloc_init(struct ceph_object_locator *oloc) +{ + oloc->pool = -1; +} + +static inline bool ceph_oloc_empty(const struct ceph_object_locator *oloc) +{ + return oloc->pool == -1; +} + +static inline void ceph_oloc_copy(struct ceph_object_locator *dest, + const struct ceph_object_locator *src) +{ + dest->pool = src->pool; +} + /* * Maximum supported by kernel client object name length * @@ -227,6 +244,23 @@ static inline void ceph_osds_init(struct ceph_osds *set) void ceph_osds_copy(struct ceph_osds *dest, const struct ceph_osds *src); +bool ceph_is_new_interval(const struct ceph_osds *old_acting, + const struct ceph_osds *new_acting, + const struct ceph_osds *old_up, + const struct ceph_osds *new_up, + int old_size, + int new_size, + int old_min_size, + int new_min_size, + u32 old_pg_num, + u32 new_pg_num, + bool old_sort_bitwise, + bool new_sort_bitwise, + const struct ceph_pg *pgid); +bool ceph_osds_changed(const struct ceph_osds *old_acting, + const struct ceph_osds *new_acting, + bool any_change); + /* calculate mapping of a file extent to an object */ extern int ceph_calc_file_object_mapping(struct ceph_file_layout *layout, u64 off, u64 len, diff --git a/include/linux/ceph/rados.h b/include/linux/ceph/rados.h index 913c87c26d33..f28ed864e682 100644 --- a/include/linux/ceph/rados.h +++ b/include/linux/ceph/rados.h @@ -153,6 +153,11 @@ extern const char *ceph_osd_state_name(int s); #define CEPH_OSDMAP_NOIN (1<<8) /* block osd auto mark-in */ #define CEPH_OSDMAP_NOBACKFILL (1<<9) /* block osd backfill */ #define CEPH_OSDMAP_NORECOVER (1<<10) /* block osd recovery and backfill */ +#define CEPH_OSDMAP_NOSCRUB (1<<11) /* block periodic scrub */ +#define CEPH_OSDMAP_NODEEP_SCRUB (1<<12) /* block periodic deep-scrub */ +#define CEPH_OSDMAP_NOTIERAGENT (1<<13) /* disable tiering agent */ +#define CEPH_OSDMAP_NOREBALANCE (1<<14) /* block osd backfill unless pg is degraded */ +#define CEPH_OSDMAP_SORTBITWISE (1<<15) /* use bitwise hobject_t sort */ /* * The error code to return when an OSD can't handle a write diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c index 0ff400a56cd6..cff3a7e29233 100644 --- a/net/ceph/osd_client.c +++ b/net/ceph/osd_client.c @@ -298,6 +298,30 @@ static void osd_req_op_data_release(struct ceph_osd_request *osd_req, } } +/* + * Assumes @t is zero-initialized. + */ +static void target_init(struct ceph_osd_request_target *t) +{ + ceph_oid_init(&t->base_oid); + ceph_oloc_init(&t->base_oloc); + ceph_oid_init(&t->target_oid); + ceph_oloc_init(&t->target_oloc); + + ceph_osds_init(&t->acting); + ceph_osds_init(&t->up); + t->size = -1; + t->min_size = -1; + + t->osd = CEPH_HOMELESS_OSD; +} + +static void target_destroy(struct ceph_osd_request_target *t) +{ + ceph_oid_destroy(&t->base_oid); + ceph_oid_destroy(&t->target_oid); +} + /* * requests */ @@ -1273,6 +1297,11 @@ void ceph_osdc_set_request_linger(struct ceph_osd_client *osdc, } EXPORT_SYMBOL(ceph_osdc_set_request_linger); +static bool __pool_full(struct ceph_pg_pool_info *pi) +{ + return pi->flags & CEPH_POOL_FLAG_FULL; +} + /* * Returns whether a request should be blocked from being sent * based on the current osdmap and osd_client settings. @@ -1289,6 +1318,20 @@ static bool __req_should_be_paused(struct ceph_osd_client *osdc, (req->r_flags & CEPH_OSD_FLAG_WRITE && pausewr); } +static bool target_should_be_paused(struct ceph_osd_client *osdc, + const struct ceph_osd_request_target *t, + struct ceph_pg_pool_info *pi) +{ + bool pauserd = ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_PAUSERD); + bool pausewr = ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_PAUSEWR) || + ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_FULL) || + __pool_full(pi); + + WARN_ON(pi->id != t->base_oloc.pool); + return (t->flags & CEPH_OSD_FLAG_READ && pauserd) || + (t->flags & CEPH_OSD_FLAG_WRITE && pausewr); +} + /* * Calculate mapping of a request to a PG. Takes tiering into account. */ @@ -1328,6 +1371,116 @@ static int __calc_request_pg(struct ceph_osdmap *osdmap, &req->r_target_oloc, pg_out); } +enum calc_target_result { + CALC_TARGET_NO_ACTION = 0, + CALC_TARGET_NEED_RESEND, + CALC_TARGET_POOL_DNE, +}; + +static enum calc_target_result calc_target(struct ceph_osd_client *osdc, + struct ceph_osd_request_target *t, + u32 *last_force_resend, + bool any_change) +{ + struct ceph_pg_pool_info *pi; + struct ceph_pg pgid, last_pgid; + struct ceph_osds up, acting; + bool force_resend = false; + bool need_check_tiering = false; + bool need_resend = false; + bool sort_bitwise = ceph_osdmap_flag(osdc->osdmap, + CEPH_OSDMAP_SORTBITWISE); + enum calc_target_result ct_res; + int ret; + + pi = ceph_pg_pool_by_id(osdc->osdmap, t->base_oloc.pool); + if (!pi) { + t->osd = CEPH_HOMELESS_OSD; + ct_res = CALC_TARGET_POOL_DNE; + goto out; + } + + if (osdc->osdmap->epoch == pi->last_force_request_resend) { + if (last_force_resend && + *last_force_resend < pi->last_force_request_resend) { + *last_force_resend = pi->last_force_request_resend; + force_resend = true; + } else if (!last_force_resend) { + force_resend = true; + } + } + if (ceph_oid_empty(&t->target_oid) || force_resend) { + ceph_oid_copy(&t->target_oid, &t->base_oid); + need_check_tiering = true; + } + if (ceph_oloc_empty(&t->target_oloc) || force_resend) { + ceph_oloc_copy(&t->target_oloc, &t->base_oloc); + need_check_tiering = true; + } + + if (need_check_tiering && + (t->flags & CEPH_OSD_FLAG_IGNORE_OVERLAY) == 0) { + if (t->flags & CEPH_OSD_FLAG_READ && pi->read_tier >= 0) + t->target_oloc.pool = pi->read_tier; + if (t->flags & CEPH_OSD_FLAG_WRITE && pi->write_tier >= 0) + t->target_oloc.pool = pi->write_tier; + } + + ret = ceph_object_locator_to_pg(osdc->osdmap, &t->target_oid, + &t->target_oloc, &pgid); + if (ret) { + WARN_ON(ret != -ENOENT); + t->osd = CEPH_HOMELESS_OSD; + ct_res = CALC_TARGET_POOL_DNE; + goto out; + } + last_pgid.pool = pgid.pool; + last_pgid.seed = ceph_stable_mod(pgid.seed, t->pg_num, t->pg_num_mask); + + ceph_pg_to_up_acting_osds(osdc->osdmap, &pgid, &up, &acting); + if (any_change && + ceph_is_new_interval(&t->acting, + &acting, + &t->up, + &up, + t->size, + pi->size, + t->min_size, + pi->min_size, + t->pg_num, + pi->pg_num, + t->sort_bitwise, + sort_bitwise, + &last_pgid)) + force_resend = true; + + if (t->paused && !target_should_be_paused(osdc, t, pi)) { + t->paused = false; + need_resend = true; + } + + if (ceph_pg_compare(&t->pgid, &pgid) || + ceph_osds_changed(&t->acting, &acting, any_change) || + force_resend) { + t->pgid = pgid; /* struct */ + ceph_osds_copy(&t->acting, &acting); + ceph_osds_copy(&t->up, &up); + t->size = pi->size; + t->min_size = pi->min_size; + t->pg_num = pi->pg_num; + t->pg_num_mask = pi->pg_num_mask; + t->sort_bitwise = sort_bitwise; + + t->osd = acting.primary; + need_resend = true; + } + + ct_res = need_resend ? CALC_TARGET_NEED_RESEND : CALC_TARGET_NO_ACTION; +out: + dout("%s t %p -> ct_res %d osd %d\n", __func__, t, ct_res, t->osd); + return ct_res; +} + static void __enqueue_request(struct ceph_osd_request *req) { struct ceph_osd_client *osdc = req->r_osdc; @@ -1805,12 +1958,12 @@ static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg) redir.oloc.pool = -1; } - if (redir.oloc.pool != -1) { + if (!ceph_oloc_empty(&redir.oloc)) { dout("redirect pool %lld\n", redir.oloc.pool); __unregister_request(osdc, req); - req->r_target_oloc = redir.oloc; /* struct */ + ceph_oloc_copy(&req->r_target_oloc, &redir.oloc); /* * Start redirect requests with nofail=true. If diff --git a/net/ceph/osdmap.c b/net/ceph/osdmap.c index 66c3ebead92f..7d4a5b43085e 100644 --- a/net/ceph/osdmap.c +++ b/net/ceph/osdmap.c @@ -1521,6 +1521,32 @@ void ceph_oid_destroy(struct ceph_object_id *oid) } EXPORT_SYMBOL(ceph_oid_destroy); +/* + * osds only + */ +static bool __osds_equal(const struct ceph_osds *lhs, + const struct ceph_osds *rhs) +{ + if (lhs->size == rhs->size && + !memcmp(lhs->osds, rhs->osds, rhs->size * sizeof(rhs->osds[0]))) + return true; + + return false; +} + +/* + * osds + primary + */ +static bool osds_equal(const struct ceph_osds *lhs, + const struct ceph_osds *rhs) +{ + if (__osds_equal(lhs, rhs) && + lhs->primary == rhs->primary) + return true; + + return false; +} + static bool osds_valid(const struct ceph_osds *set) { /* non-empty set */ @@ -1553,6 +1579,101 @@ void ceph_osds_copy(struct ceph_osds *dest, const struct ceph_osds *src) dest->primary = src->primary; } +static bool is_split(const struct ceph_pg *pgid, + u32 old_pg_num, + u32 new_pg_num) +{ + int old_bits = calc_bits_of(old_pg_num); + int old_mask = (1 << old_bits) - 1; + int n; + + WARN_ON(pgid->seed >= old_pg_num); + if (new_pg_num <= old_pg_num) + return false; + + for (n = 1; ; n++) { + int next_bit = n << (old_bits - 1); + u32 s = next_bit | pgid->seed; + + if (s < old_pg_num || s == pgid->seed) + continue; + if (s >= new_pg_num) + break; + + s = ceph_stable_mod(s, old_pg_num, old_mask); + if (s == pgid->seed) + return true; + } + + return false; +} + +bool ceph_is_new_interval(const struct ceph_osds *old_acting, + const struct ceph_osds *new_acting, + const struct ceph_osds *old_up, + const struct ceph_osds *new_up, + int old_size, + int new_size, + int old_min_size, + int new_min_size, + u32 old_pg_num, + u32 new_pg_num, + bool old_sort_bitwise, + bool new_sort_bitwise, + const struct ceph_pg *pgid) +{ + return !osds_equal(old_acting, new_acting) || + !osds_equal(old_up, new_up) || + old_size != new_size || + old_min_size != new_min_size || + is_split(pgid, old_pg_num, new_pg_num) || + old_sort_bitwise != new_sort_bitwise; +} + +static int calc_pg_rank(int osd, const struct ceph_osds *acting) +{ + int i; + + for (i = 0; i < acting->size; i++) { + if (acting->osds[i] == osd) + return i; + } + + return -1; +} + +static bool primary_changed(const struct ceph_osds *old_acting, + const struct ceph_osds *new_acting) +{ + if (!old_acting->size && !new_acting->size) + return false; /* both still empty */ + + if (!old_acting->size ^ !new_acting->size) + return true; /* was empty, now not, or vice versa */ + + if (old_acting->primary != new_acting->primary) + return true; /* primary changed */ + + if (calc_pg_rank(old_acting->primary, old_acting) != + calc_pg_rank(new_acting->primary, new_acting)) + return true; + + return false; /* same primary (tho replicas may have changed) */ +} + +bool ceph_osds_changed(const struct ceph_osds *old_acting, + const struct ceph_osds *new_acting, + bool any_change) +{ + if (primary_changed(old_acting, new_acting)) + return true; + + if (any_change && !__osds_equal(old_acting, new_acting)) + return true; + + return false; +} + /* * calculate file layout from given offset, length. * fill in correct oid, logical length, and object extent -- cgit v1.3-8-gc7d7 From bb873b539154ab51893430b4ad6ba4051775276a Mon Sep 17 00:00:00 2001 From: Ilya Dryomov Date: Thu, 26 May 2016 00:29:52 +0200 Subject: libceph: switch to calc_target(), part 2 The crux of this is getting rid of ceph_osdc_build_request(), so that MOSDOp can be encoded not before but after calc_target() calculates the actual target. Encoding now happens within ceph_osdc_start_request(). Also nuked is the accompanying bunch of pointers into the encoded buffer that was used to update fields on each send - instead, the entire front is re-encoded. If we want to support target->name_len != base->name_len in the future, there is no other way, because oid is surrounded by other fields in the encoded buffer. Encoding OSD ops and adding data items to the request message were mixed together in osd_req_encode_op(). While we want to re-encode OSD ops, we don't want to add duplicate data items to the message when resending, so all call to ceph_osdc_msg_data_add() are factored out into a new setup_request_data(). Signed-off-by: Ilya Dryomov --- drivers/block/rbd.c | 18 +- fs/ceph/addr.c | 16 +- fs/ceph/file.c | 16 +- include/linux/ceph/osd_client.h | 29 ++-- include/linux/ceph/rados.h | 7 + net/ceph/debugfs.c | 61 ++++--- net/ceph/osd_client.c | 355 ++++++++++++++++++++-------------------- 7 files changed, 247 insertions(+), 255 deletions(-) (limited to 'fs') diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c index f3ea927f93de..0e598916e048 100644 --- a/drivers/block/rbd.c +++ b/drivers/block/rbd.c @@ -1896,27 +1896,17 @@ static void rbd_osd_req_format_read(struct rbd_obj_request *obj_request) { struct rbd_img_request *img_request = obj_request->img_request; struct ceph_osd_request *osd_req = obj_request->osd_req; - u64 snap_id; - - rbd_assert(osd_req != NULL); - snap_id = img_request ? img_request->snap_id : CEPH_NOSNAP; - ceph_osdc_build_request(osd_req, obj_request->offset, - NULL, snap_id, NULL); + if (img_request) + osd_req->r_snapid = img_request->snap_id; } static void rbd_osd_req_format_write(struct rbd_obj_request *obj_request) { - struct rbd_img_request *img_request = obj_request->img_request; struct ceph_osd_request *osd_req = obj_request->osd_req; - struct ceph_snap_context *snapc; - struct timespec mtime = CURRENT_TIME; - - rbd_assert(osd_req != NULL); - snapc = img_request ? img_request->snapc : NULL; - ceph_osdc_build_request(osd_req, obj_request->offset, - snapc, CEPH_NOSNAP, &mtime); + osd_req->r_mtime = CURRENT_TIME; + osd_req->r_data_offset = obj_request->offset; } /* diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c index c5d75486823b..59b3c3fbd3bd 100644 --- a/fs/ceph/addr.c +++ b/fs/ceph/addr.c @@ -376,8 +376,6 @@ static int start_read(struct inode *inode, struct list_head *page_list, int max) req->r_callback = finish_read; req->r_inode = inode; - ceph_osdc_build_request(req, off, NULL, vino.snap, NULL); - dout("start_read %p starting %p %lld~%lld\n", inode, req, off, len); ret = ceph_osdc_start_request(osdc, req, false); if (ret < 0) @@ -1063,10 +1061,7 @@ new_request: pages = NULL; } - vino = ceph_vino(inode); - ceph_osdc_build_request(req, offset, snapc, vino.snap, - &inode->i_mtime); - + req->r_mtime = inode->i_mtime; rc = ceph_osdc_start_request(&fsc->client->osdc, req, true); BUG_ON(rc); req = NULL; @@ -1614,7 +1609,7 @@ int ceph_uninline_data(struct file *filp, struct page *locked_page) goto out; } - ceph_osdc_build_request(req, 0, NULL, CEPH_NOSNAP, &inode->i_mtime); + req->r_mtime = inode->i_mtime; err = ceph_osdc_start_request(&fsc->client->osdc, req, false); if (!err) err = ceph_osdc_wait_request(&fsc->client->osdc, req); @@ -1657,7 +1652,7 @@ int ceph_uninline_data(struct file *filp, struct page *locked_page) goto out_put; } - ceph_osdc_build_request(req, 0, NULL, CEPH_NOSNAP, &inode->i_mtime); + req->r_mtime = inode->i_mtime; err = ceph_osdc_start_request(&fsc->client->osdc, req, false); if (!err) err = ceph_osdc_wait_request(&fsc->client->osdc, req); @@ -1790,12 +1785,9 @@ static int __ceph_pool_perm_get(struct ceph_inode_info *ci, u32 pool) osd_req_op_raw_data_in_pages(rd_req, 0, pages, PAGE_SIZE, 0, false, true); - ceph_osdc_build_request(rd_req, 0, NULL, CEPH_NOSNAP, - &ci->vfs_inode.i_mtime); err = ceph_osdc_start_request(&fsc->client->osdc, rd_req, false); - ceph_osdc_build_request(wr_req, 0, NULL, CEPH_NOSNAP, - &ci->vfs_inode.i_mtime); + wr_req->r_mtime = ci->vfs_inode.i_mtime; err2 = ceph_osdc_start_request(&fsc->client->osdc, wr_req, false); if (!err) diff --git a/fs/ceph/file.c b/fs/ceph/file.c index 36b4a41dfa67..52e4b72dd5de 100644 --- a/fs/ceph/file.c +++ b/fs/ceph/file.c @@ -727,8 +727,8 @@ static void ceph_aio_retry_work(struct work_struct *work) req->r_ops[0] = orig_req->r_ops[0]; osd_req_op_init(req, 1, CEPH_OSD_OP_STARTSYNC, 0); - ceph_osdc_build_request(req, req->r_ops[0].extent.offset, - snapc, CEPH_NOSNAP, &aio_req->mtime); + req->r_mtime = aio_req->mtime; + req->r_data_offset = req->r_ops[0].extent.offset; ceph_osdc_put_request(orig_req); @@ -882,14 +882,12 @@ ceph_direct_read_write(struct kiocb *iocb, struct iov_iter *iter, (pos+len) | (PAGE_SIZE - 1)); osd_req_op_init(req, 1, CEPH_OSD_OP_STARTSYNC, 0); + req->r_mtime = mtime; } - osd_req_op_extent_osd_data_pages(req, 0, pages, len, start, false, false); - ceph_osdc_build_request(req, pos, snapc, vino.snap, &mtime); - if (aio_req) { aio_req->total_len += len; aio_req->num_reqs++; @@ -1074,9 +1072,7 @@ ceph_sync_write(struct kiocb *iocb, struct iov_iter *from, loff_t pos, osd_req_op_extent_osd_data_pages(req, 0, pages, len, 0, false, true); - /* BUG_ON(vino.snap != CEPH_NOSNAP); */ - ceph_osdc_build_request(req, pos, snapc, vino.snap, &mtime); - + req->r_mtime = mtime; ret = ceph_osdc_start_request(&fsc->client->osdc, req, false); if (!ret) ret = ceph_osdc_wait_request(&fsc->client->osdc, req); @@ -1532,9 +1528,7 @@ static int ceph_zero_partial_object(struct inode *inode, goto out; } - ceph_osdc_build_request(req, offset, NULL, ceph_vino(inode).snap, - &inode->i_mtime); - + req->r_mtime = inode->i_mtime; ret = ceph_osdc_start_request(&fsc->client->osdc, req, false); if (!ret) { ret = ceph_osdc_wait_request(&fsc->client->osdc, req); diff --git a/include/linux/ceph/osd_client.h b/include/linux/ceph/osd_client.h index 03bf9d9e1517..67a37d98e0ca 100644 --- a/include/linux/ceph/osd_client.h +++ b/include/linux/ceph/osd_client.h @@ -104,7 +104,7 @@ struct ceph_osd_req_op { struct ceph_osd_data response_data; __u8 class_len; __u8 method_len; - __u8 argc; + u32 indata_len; } cls; struct { u64 cookie; @@ -162,14 +162,6 @@ struct ceph_osd_request { /* request osd ops array */ unsigned int r_num_ops; - /* these are updated on each send */ - __le32 *r_request_osdmap_epoch; - __le32 *r_request_flags; - __le64 *r_request_pool; - void *r_request_pgid; - __le32 *r_request_attempts; - struct ceph_eversion *r_request_reassert_version; - int r_result; int r_got_reply; int r_linger; @@ -180,16 +172,22 @@ struct ceph_osd_request { struct completion r_completion, r_safe_completion; ceph_osdc_callback_t r_callback; ceph_osdc_unsafe_callback_t r_unsafe_callback; - struct ceph_eversion r_reassert_version; struct list_head r_unsafe_item; struct inode *r_inode; /* for use by callbacks */ void *r_priv; /* ditto */ - u64 r_snapid; - unsigned long r_stamp; /* send OR check time */ + /* set by submitter */ + u64 r_snapid; /* for reads, CEPH_NOSNAP o/w */ + struct ceph_snap_context *r_snapc; /* for writes */ + struct timespec r_mtime; /* ditto */ + u64 r_data_offset; /* ditto */ - struct ceph_snap_context *r_snapc; /* snap context for writes */ + /* internal */ + unsigned long r_stamp; /* jiffies, send or check time */ + int r_attempts; + struct ceph_eversion r_replay_version; /* aka reassert_version */ + u32 r_last_force_resend; struct ceph_osd_req_op r_ops[]; }; @@ -334,11 +332,6 @@ extern struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client * gfp_t gfp_flags); int ceph_osdc_alloc_messages(struct ceph_osd_request *req, gfp_t gfp); -extern void ceph_osdc_build_request(struct ceph_osd_request *req, u64 off, - struct ceph_snap_context *snapc, - u64 snap_id, - struct timespec *mtime); - extern struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *, struct ceph_file_layout *layout, struct ceph_vino vino, diff --git a/include/linux/ceph/rados.h b/include/linux/ceph/rados.h index f28ed864e682..28740a58f32c 100644 --- a/include/linux/ceph/rados.h +++ b/include/linux/ceph/rados.h @@ -394,6 +394,13 @@ enum { CEPH_OSD_FLAG_SKIPRWLOCKS = 0x10000, /* skip rw locks */ CEPH_OSD_FLAG_IGNORE_OVERLAY = 0x20000, /* ignore pool overlay */ CEPH_OSD_FLAG_FLUSH = 0x40000, /* this is part of flush */ + CEPH_OSD_FLAG_MAP_SNAP_CLONE = 0x80000, /* map snap direct to clone id */ + CEPH_OSD_FLAG_ENFORCE_SNAPC = 0x100000, /* use snapc provided even if + pool uses pool snaps */ + CEPH_OSD_FLAG_REDIRECTED = 0x200000, /* op has been redirected */ + CEPH_OSD_FLAG_KNOWN_REDIR = 0x400000, /* redirect bit is authoritative */ + CEPH_OSD_FLAG_FULL_TRY = 0x800000, /* try op despite full flag */ + CEPH_OSD_FLAG_FULL_FORCE = 0x1000000, /* force op despite full flag */ }; enum { diff --git a/net/ceph/debugfs.c b/net/ceph/debugfs.c index 0c11ab5f8c30..6d3ff713edeb 100644 --- a/net/ceph/debugfs.c +++ b/net/ceph/debugfs.c @@ -145,6 +145,43 @@ static int monc_show(struct seq_file *s, void *p) return 0; } +static void dump_target(struct seq_file *s, struct ceph_osd_request_target *t) +{ + int i; + + seq_printf(s, "osd%d\t%llu.%x\t[", t->osd, t->pgid.pool, t->pgid.seed); + for (i = 0; i < t->up.size; i++) + seq_printf(s, "%s%d", (!i ? "" : ","), t->up.osds[i]); + seq_printf(s, "]/%d\t[", t->up.primary); + for (i = 0; i < t->acting.size; i++) + seq_printf(s, "%s%d", (!i ? "" : ","), t->acting.osds[i]); + seq_printf(s, "]/%d\t%*pE\t0x%x", t->acting.primary, + t->target_oid.name_len, t->target_oid.name, t->flags); + if (t->paused) + seq_puts(s, "\tP"); +} + +static void dump_request(struct seq_file *s, struct ceph_osd_request *req) +{ + int i; + + seq_printf(s, "%llu\t", req->r_tid); + dump_target(s, &req->r_t); + + seq_printf(s, "\t%d\t%u'%llu", req->r_attempts, + le32_to_cpu(req->r_replay_version.epoch), + le64_to_cpu(req->r_replay_version.version)); + + for (i = 0; i < req->r_num_ops; i++) { + struct ceph_osd_req_op *op = &req->r_ops[i]; + + seq_printf(s, "%s%s", (i == 0 ? "\t" : ","), + ceph_osd_op_name(op->op)); + } + + seq_putc(s, '\n'); +} + static int osdc_show(struct seq_file *s, void *pp) { struct ceph_client *client = s->private; @@ -154,32 +191,10 @@ static int osdc_show(struct seq_file *s, void *pp) mutex_lock(&osdc->request_mutex); for (p = rb_first(&osdc->requests); p; p = rb_next(p)) { struct ceph_osd_request *req; - unsigned int i; - int opcode; req = rb_entry(p, struct ceph_osd_request, r_node); - seq_printf(s, "%lld\tosd%d\t%lld.%x\t", req->r_tid, - req->r_osd ? req->r_osd->o_osd : -1, - req->r_t.pgid.pool, req->r_t.pgid.seed); - - seq_printf(s, "%*pE", req->r_base_oid.name_len, - req->r_base_oid.name); - - if (req->r_reassert_version.epoch) - seq_printf(s, "\t%u'%llu", - (unsigned int)le32_to_cpu(req->r_reassert_version.epoch), - le64_to_cpu(req->r_reassert_version.version)); - else - seq_printf(s, "\t"); - - for (i = 0; i < req->r_num_ops; i++) { - opcode = req->r_ops[i].op; - seq_printf(s, "%s%s", (i == 0 ? "\t" : ","), - ceph_osd_op_name(opcode)); - } - - seq_printf(s, "\n"); + dump_request(s, req); } mutex_unlock(&osdc->request_mutex); return 0; diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c index 013101598c41..8a008f083283 100644 --- a/net/ceph/osd_client.c +++ b/net/ceph/osd_client.c @@ -34,8 +34,6 @@ static void __unregister_request(struct ceph_osd_client *osdc, static void __unregister_linger_request(struct ceph_osd_client *osdc, struct ceph_osd_request *req); static void __enqueue_request(struct ceph_osd_request *req); -static void __send_request(struct ceph_osd_client *osdc, - struct ceph_osd_request *req); /* * Implement client access to distributed object storage cluster. @@ -209,6 +207,8 @@ void osd_req_op_cls_request_data_pagelist( osd_data = osd_req_op_data(osd_req, which, cls, request_data); ceph_osd_data_pagelist_init(osd_data, pagelist); + osd_req->r_ops[which].cls.indata_len += pagelist->length; + osd_req->r_ops[which].indata_len += pagelist->length; } EXPORT_SYMBOL(osd_req_op_cls_request_data_pagelist); @@ -221,6 +221,8 @@ void osd_req_op_cls_request_data_pages(struct ceph_osd_request *osd_req, osd_data = osd_req_op_data(osd_req, which, cls, request_data); ceph_osd_data_pages_init(osd_data, pages, length, alignment, pages_from_pool, own_pages); + osd_req->r_ops[which].cls.indata_len += length; + osd_req->r_ops[which].indata_len += length; } EXPORT_SYMBOL(osd_req_op_cls_request_data_pages); @@ -610,8 +612,6 @@ void osd_req_op_cls_init(struct ceph_osd_request *osd_req, unsigned int which, osd_req_op_cls_request_info_pagelist(osd_req, which, pagelist); - op->cls.argc = 0; /* currently unused */ - op->indata_len = payload_len; } EXPORT_SYMBOL(osd_req_op_cls_init); @@ -709,16 +709,9 @@ static void ceph_osdc_msg_data_add(struct ceph_msg *msg, } } -static u64 osd_req_encode_op(struct ceph_osd_request *req, - struct ceph_osd_op *dst, unsigned int which) +static u32 osd_req_encode_op(struct ceph_osd_op *dst, + const struct ceph_osd_req_op *src) { - struct ceph_osd_req_op *src; - struct ceph_osd_data *osd_data; - u64 request_data_len = 0; - u64 data_length; - - BUG_ON(which >= req->r_num_ops); - src = &req->r_ops[which]; if (WARN_ON(!osd_req_opcode_valid(src->op))) { pr_err("unrecognized osd opcode %d\n", src->op); @@ -727,49 +720,23 @@ static u64 osd_req_encode_op(struct ceph_osd_request *req, switch (src->op) { case CEPH_OSD_OP_STAT: - osd_data = &src->raw_data_in; - ceph_osdc_msg_data_add(req->r_reply, osd_data); break; case CEPH_OSD_OP_READ: case CEPH_OSD_OP_WRITE: case CEPH_OSD_OP_WRITEFULL: case CEPH_OSD_OP_ZERO: case CEPH_OSD_OP_TRUNCATE: - if (src->op == CEPH_OSD_OP_WRITE || - src->op == CEPH_OSD_OP_WRITEFULL) - request_data_len = src->extent.length; dst->extent.offset = cpu_to_le64(src->extent.offset); dst->extent.length = cpu_to_le64(src->extent.length); dst->extent.truncate_size = cpu_to_le64(src->extent.truncate_size); dst->extent.truncate_seq = cpu_to_le32(src->extent.truncate_seq); - osd_data = &src->extent.osd_data; - if (src->op == CEPH_OSD_OP_WRITE || - src->op == CEPH_OSD_OP_WRITEFULL) - ceph_osdc_msg_data_add(req->r_request, osd_data); - else - ceph_osdc_msg_data_add(req->r_reply, osd_data); break; case CEPH_OSD_OP_CALL: dst->cls.class_len = src->cls.class_len; dst->cls.method_len = src->cls.method_len; - osd_data = &src->cls.request_info; - ceph_osdc_msg_data_add(req->r_request, osd_data); - BUG_ON(osd_data->type != CEPH_OSD_DATA_TYPE_PAGELIST); - request_data_len = osd_data->pagelist->length; - - osd_data = &src->cls.request_data; - data_length = ceph_osd_data_length(osd_data); - if (data_length) { - BUG_ON(osd_data->type == CEPH_OSD_DATA_TYPE_NONE); - dst->cls.indata_len = cpu_to_le32(data_length); - ceph_osdc_msg_data_add(req->r_request, osd_data); - src->indata_len += data_length; - request_data_len += data_length; - } - osd_data = &src->cls.response_data; - ceph_osdc_msg_data_add(req->r_reply, osd_data); + dst->cls.indata_len = cpu_to_le32(src->cls.indata_len); break; case CEPH_OSD_OP_STARTSYNC: break; @@ -791,9 +758,6 @@ static u64 osd_req_encode_op(struct ceph_osd_request *req, dst->xattr.value_len = cpu_to_le32(src->xattr.value_len); dst->xattr.cmp_op = src->xattr.cmp_op; dst->xattr.cmp_mode = src->xattr.cmp_mode; - osd_data = &src->xattr.osd_data; - ceph_osdc_msg_data_add(req->r_request, osd_data); - request_data_len = osd_data->pagelist->length; break; case CEPH_OSD_OP_CREATE: case CEPH_OSD_OP_DELETE: @@ -810,7 +774,7 @@ static u64 osd_req_encode_op(struct ceph_osd_request *req, dst->flags = cpu_to_le32(src->flags); dst->payload_len = cpu_to_le32(src->indata_len); - return request_data_len; + return src->indata_len; } /* @@ -852,8 +816,6 @@ struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc, goto fail; } - req->r_flags = flags; - /* calculate max write size */ r = calc_layout(layout, off, plen, &objnum, &objoff, &objlen); if (r) @@ -877,9 +839,14 @@ struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc, truncate_size, truncate_seq); } + req->r_flags = flags; req->r_base_oloc.pool = ceph_file_layout_pg_pool(*layout); ceph_oid_printf(&req->r_base_oid, "%llx.%08llx", vino.ino, objnum); + req->r_snapid = vino.snap; + if (flags & CEPH_OSD_FLAG_WRITE) + req->r_data_offset = off; + r = ceph_osdc_alloc_messages(req, GFP_NOFS); if (r) goto fail; @@ -1509,37 +1476,173 @@ out: return err; } -/* - * caller should hold map_sem (for read) and request_mutex - */ -static void __send_request(struct ceph_osd_client *osdc, - struct ceph_osd_request *req) +static void setup_request_data(struct ceph_osd_request *req, + struct ceph_msg *msg) { - void *p; + u32 data_len = 0; + int i; + + if (!list_empty(&msg->data)) + return; - dout("send_request %p tid %llu to osd%d flags %d pg %lld.%x\n", - req, req->r_tid, req->r_osd->o_osd, req->r_flags, - req->r_t.pgid.pool, req->r_t.pgid.seed); + WARN_ON(msg->data_length); + for (i = 0; i < req->r_num_ops; i++) { + struct ceph_osd_req_op *op = &req->r_ops[i]; + + switch (op->op) { + /* request */ + case CEPH_OSD_OP_WRITE: + case CEPH_OSD_OP_WRITEFULL: + WARN_ON(op->indata_len != op->extent.length); + ceph_osdc_msg_data_add(msg, &op->extent.osd_data); + break; + case CEPH_OSD_OP_SETXATTR: + case CEPH_OSD_OP_CMPXATTR: + WARN_ON(op->indata_len != op->xattr.name_len + + op->xattr.value_len); + ceph_osdc_msg_data_add(msg, &op->xattr.osd_data); + break; + + /* reply */ + case CEPH_OSD_OP_STAT: + ceph_osdc_msg_data_add(req->r_reply, + &op->raw_data_in); + break; + case CEPH_OSD_OP_READ: + ceph_osdc_msg_data_add(req->r_reply, + &op->extent.osd_data); + break; + + /* both */ + case CEPH_OSD_OP_CALL: + WARN_ON(op->indata_len != op->cls.class_len + + op->cls.method_len + + op->cls.indata_len); + ceph_osdc_msg_data_add(msg, &op->cls.request_info); + /* optional, can be NONE */ + ceph_osdc_msg_data_add(msg, &op->cls.request_data); + /* optional, can be NONE */ + ceph_osdc_msg_data_add(req->r_reply, + &op->cls.response_data); + break; + } + + data_len += op->indata_len; + } - /* fill in message content that changes each time we send it */ - put_unaligned_le32(osdc->osdmap->epoch, req->r_request_osdmap_epoch); - put_unaligned_le32(req->r_flags, req->r_request_flags); - put_unaligned_le64(req->r_t.target_oloc.pool, req->r_request_pool); - p = req->r_request_pgid; + WARN_ON(data_len != msg->data_length); +} + +static void encode_request(struct ceph_osd_request *req, struct ceph_msg *msg) +{ + void *p = msg->front.iov_base; + void *const end = p + msg->front_alloc_len; + u32 data_len = 0; + int i; + + if (req->r_flags & CEPH_OSD_FLAG_WRITE) { + /* snapshots aren't writeable */ + WARN_ON(req->r_snapid != CEPH_NOSNAP); + } else { + WARN_ON(req->r_mtime.tv_sec || req->r_mtime.tv_nsec || + req->r_data_offset || req->r_snapc); + } + + setup_request_data(req, msg); + + ceph_encode_32(&p, 1); /* client_inc, always 1 */ + ceph_encode_32(&p, req->r_osdc->osdmap->epoch); + ceph_encode_32(&p, req->r_flags); + ceph_encode_timespec(p, &req->r_mtime); + p += sizeof(struct ceph_timespec); + /* aka reassert_version */ + memcpy(p, &req->r_replay_version, sizeof(req->r_replay_version)); + p += sizeof(req->r_replay_version); + + /* oloc */ + ceph_encode_8(&p, 4); + ceph_encode_8(&p, 4); + ceph_encode_32(&p, 8 + 4 + 4); + ceph_encode_64(&p, req->r_t.target_oloc.pool); + ceph_encode_32(&p, -1); /* preferred */ + ceph_encode_32(&p, 0); /* key len */ + + /* pgid */ + ceph_encode_8(&p, 1); ceph_encode_64(&p, req->r_t.pgid.pool); ceph_encode_32(&p, req->r_t.pgid.seed); - put_unaligned_le64(1, req->r_request_attempts); /* FIXME */ - memcpy(req->r_request_reassert_version, &req->r_reassert_version, - sizeof(req->r_reassert_version)); + ceph_encode_32(&p, -1); /* preferred */ - req->r_stamp = jiffies; - list_move_tail(&req->r_req_lru_item, &osdc->req_lru); + /* oid */ + ceph_encode_32(&p, req->r_t.target_oid.name_len); + memcpy(p, req->r_t.target_oid.name, req->r_t.target_oid.name_len); + p += req->r_t.target_oid.name_len; - ceph_msg_get(req->r_request); /* send consumes a ref */ + /* ops, can imply data */ + ceph_encode_16(&p, req->r_num_ops); + for (i = 0; i < req->r_num_ops; i++) { + data_len += osd_req_encode_op(p, &req->r_ops[i]); + p += sizeof(struct ceph_osd_op); + } - req->r_sent = req->r_osd->o_incarnation; + ceph_encode_64(&p, req->r_snapid); /* snapid */ + if (req->r_snapc) { + ceph_encode_64(&p, req->r_snapc->seq); + ceph_encode_32(&p, req->r_snapc->num_snaps); + for (i = 0; i < req->r_snapc->num_snaps; i++) + ceph_encode_64(&p, req->r_snapc->snaps[i]); + } else { + ceph_encode_64(&p, 0); /* snap_seq */ + ceph_encode_32(&p, 0); /* snaps len */ + } + + ceph_encode_32(&p, req->r_attempts); /* retry_attempt */ + + BUG_ON(p > end); + msg->front.iov_len = p - msg->front.iov_base; + msg->hdr.version = cpu_to_le16(4); /* MOSDOp v4 */ + msg->hdr.front_len = cpu_to_le32(msg->front.iov_len); + msg->hdr.data_len = cpu_to_le32(data_len); + /* + * The header "data_off" is a hint to the receiver allowing it + * to align received data into its buffers such that there's no + * need to re-copy it before writing it to disk (direct I/O). + */ + msg->hdr.data_off = cpu_to_le16(req->r_data_offset); - ceph_con_send(&req->r_osd->o_con, req->r_request); + dout("%s req %p oid %*pE oid_len %d front %zu data %u\n", __func__, + req, req->r_t.target_oid.name_len, req->r_t.target_oid.name, + req->r_t.target_oid.name_len, msg->front.iov_len, data_len); +} + +/* + * @req has to be assigned a tid and registered. + */ +static void send_request(struct ceph_osd_request *req) +{ + struct ceph_osd *osd = req->r_osd; + + WARN_ON(osd->o_osd != req->r_t.osd); + + req->r_flags |= CEPH_OSD_FLAG_KNOWN_REDIR; + if (req->r_attempts) + req->r_flags |= CEPH_OSD_FLAG_RETRY; + else + WARN_ON(req->r_flags & CEPH_OSD_FLAG_RETRY); + + encode_request(req, req->r_request); + + dout("%s req %p tid %llu to pg %llu.%x osd%d flags 0x%x attempt %d\n", + __func__, req, req->r_tid, req->r_t.pgid.pool, req->r_t.pgid.seed, + req->r_t.osd, req->r_flags, req->r_attempts); + + req->r_t.paused = false; + req->r_stamp = jiffies; + req->r_attempts++; + + req->r_sent = osd->o_incarnation; + req->r_request->hdr.tid = cpu_to_le64(req->r_tid); + ceph_con_send(&osd->o_con, ceph_msg_get(req->r_request)); } /* @@ -1550,8 +1653,10 @@ static void __send_queued(struct ceph_osd_client *osdc) struct ceph_osd_request *req, *tmp; dout("__send_queued\n"); - list_for_each_entry_safe(req, tmp, &osdc->req_unsent, r_req_lru_item) - __send_request(osdc, req); + list_for_each_entry_safe(req, tmp, &osdc->req_unsent, r_req_lru_item) { + list_move_tail(&req->r_req_lru_item, &osdc->req_lru); + send_request(req); + } } /* @@ -1915,8 +2020,8 @@ static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg) req->r_result = bytes; /* in case this is a write and we need to replay, */ - req->r_reassert_version.epoch = cpu_to_le32(reassert_epoch); - req->r_reassert_version.version = cpu_to_le64(reassert_version); + req->r_replay_version.epoch = cpu_to_le32(reassert_epoch); + req->r_replay_version.version = cpu_to_le64(reassert_version); req->r_got_reply = 1; } else if ((flags & CEPH_OSD_FLAG_ONDISK) == 0) { @@ -2432,105 +2537,6 @@ bad: pr_err("osdc handle_watch_notify corrupt msg\n"); } -/* - * build new request AND message - * - */ -void ceph_osdc_build_request(struct ceph_osd_request *req, u64 off, - struct ceph_snap_context *snapc, u64 snap_id, - struct timespec *mtime) -{ - struct ceph_msg *msg = req->r_request; - void *p; - size_t msg_size; - int flags = req->r_flags; - u64 data_len; - unsigned int i; - - req->r_snapid = snap_id; - WARN_ON(snapc != req->r_snapc); - - /* encode request */ - msg->hdr.version = cpu_to_le16(4); - - p = msg->front.iov_base; - ceph_encode_32(&p, 1); /* client_inc is always 1 */ - req->r_request_osdmap_epoch = p; - p += 4; - req->r_request_flags = p; - p += 4; - if (req->r_flags & CEPH_OSD_FLAG_WRITE) - ceph_encode_timespec(p, mtime); - p += sizeof(struct ceph_timespec); - req->r_request_reassert_version = p; - p += sizeof(struct ceph_eversion); /* will get filled in */ - - /* oloc */ - ceph_encode_8(&p, 4); - ceph_encode_8(&p, 4); - ceph_encode_32(&p, 8 + 4 + 4); - req->r_request_pool = p; - p += 8; - ceph_encode_32(&p, -1); /* preferred */ - ceph_encode_32(&p, 0); /* key len */ - - ceph_encode_8(&p, 1); - req->r_request_pgid = p; - p += 8 + 4; - ceph_encode_32(&p, -1); /* preferred */ - - /* oid */ - ceph_encode_32(&p, req->r_base_oid.name_len); - memcpy(p, req->r_base_oid.name, req->r_base_oid.name_len); - dout("oid %*pE len %d\n", req->r_base_oid.name_len, - req->r_base_oid.name, req->r_base_oid.name_len); - p += req->r_base_oid.name_len; - - /* ops--can imply data */ - ceph_encode_16(&p, (u16)req->r_num_ops); - data_len = 0; - for (i = 0; i < req->r_num_ops; i++) { - data_len += osd_req_encode_op(req, p, i); - p += sizeof(struct ceph_osd_op); - } - - /* snaps */ - ceph_encode_64(&p, req->r_snapid); - ceph_encode_64(&p, req->r_snapc ? req->r_snapc->seq : 0); - ceph_encode_32(&p, req->r_snapc ? req->r_snapc->num_snaps : 0); - if (req->r_snapc) { - for (i = 0; i < req->r_snapc->num_snaps; i++) { - ceph_encode_64(&p, req->r_snapc->snaps[i]); - } - } - - req->r_request_attempts = p; - p += 4; - - /* data */ - if (flags & CEPH_OSD_FLAG_WRITE) { - u16 data_off; - - /* - * The header "data_off" is a hint to the receiver - * allowing it to align received data into its - * buffers such that there's no need to re-copy - * it before writing it to disk (direct I/O). - */ - data_off = (u16) (off & 0xffff); - req->r_request->hdr.data_off = cpu_to_le16(data_off); - } - req->r_request->hdr.data_len = cpu_to_le32(data_len); - - BUG_ON(p > msg->front.iov_base + msg->front.iov_len); - msg_size = p - msg->front.iov_base; - msg->front.iov_len = msg_size; - msg->hdr.front_len = cpu_to_le32(msg_size); - - dout("build_request msg_size was %d\n", (int)msg_size); -} -EXPORT_SYMBOL(ceph_osdc_build_request); - /* * Register request, send initial attempt. */ @@ -2749,15 +2755,12 @@ int ceph_osdc_readpages(struct ceph_osd_client *osdc, return PTR_ERR(req); /* it may be a short read due to an object boundary */ - osd_req_op_extent_osd_data_pages(req, 0, pages, *plen, page_align, false, false); dout("readpages final extent is %llu~%llu (%llu bytes align %d)\n", off, *plen, *plen, page_align); - ceph_osdc_build_request(req, off, NULL, vino.snap, NULL); - rc = ceph_osdc_start_request(osdc, req, false); if (!rc) rc = ceph_osdc_wait_request(osdc, req); @@ -2783,7 +2786,6 @@ int ceph_osdc_writepages(struct ceph_osd_client *osdc, struct ceph_vino vino, int rc = 0; int page_align = off & ~PAGE_MASK; - BUG_ON(vino.snap != CEPH_NOSNAP); /* snapshots aren't writeable */ req = ceph_osdc_new_request(osdc, layout, vino, off, &len, 0, 1, CEPH_OSD_OP_WRITE, CEPH_OSD_FLAG_ONDISK | CEPH_OSD_FLAG_WRITE, @@ -2797,8 +2799,7 @@ int ceph_osdc_writepages(struct ceph_osd_client *osdc, struct ceph_vino vino, false, false); dout("writepages %llu~%llu (%llu bytes)\n", off, len, len); - ceph_osdc_build_request(req, off, snapc, CEPH_NOSNAP, mtime); - + req->r_mtime = *mtime; rc = ceph_osdc_start_request(osdc, req, true); if (!rc) rc = ceph_osdc_wait_request(osdc, req); -- cgit v1.3-8-gc7d7 From 85e084feb47349d62989efe1713a8723af95f4ea Mon Sep 17 00:00:00 2001 From: Ilya Dryomov Date: Thu, 28 Apr 2016 16:07:24 +0200 Subject: libceph: drop msg argument from ceph_osdc_callback_t finish_read(), its only user, uses it to get to hdr.data_len, which is what ->r_result is set to on success. This gains us the ability to safely call callbacks from contexts other than reply, e.g. map check. Signed-off-by: Ilya Dryomov --- drivers/block/rbd.c | 5 ++--- fs/ceph/addr.c | 9 ++++----- fs/ceph/file.c | 7 +++---- include/linux/ceph/osd_client.h | 3 +-- net/ceph/osd_client.c | 4 ++-- 5 files changed, 12 insertions(+), 16 deletions(-) (limited to 'fs') diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c index 0e598916e048..82b03aa509e6 100644 --- a/drivers/block/rbd.c +++ b/drivers/block/rbd.c @@ -1828,13 +1828,12 @@ static void rbd_osd_call_callback(struct rbd_obj_request *obj_request) obj_request_done_set(obj_request); } -static void rbd_osd_req_callback(struct ceph_osd_request *osd_req, - struct ceph_msg *msg) +static void rbd_osd_req_callback(struct ceph_osd_request *osd_req) { struct rbd_obj_request *obj_request = osd_req->r_priv; u16 opcode; - dout("%s: osd_req %p msg %p\n", __func__, osd_req, msg); + dout("%s: osd_req %p\n", __func__, osd_req); rbd_assert(osd_req == obj_request->osd_req); if (obj_request_img_data_test(obj_request)) { rbd_assert(obj_request->img_request); diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c index 59b3c3fbd3bd..a11756a39471 100644 --- a/fs/ceph/addr.c +++ b/fs/ceph/addr.c @@ -257,12 +257,12 @@ static int ceph_readpage(struct file *filp, struct page *page) /* * Finish an async read(ahead) op. */ -static void finish_read(struct ceph_osd_request *req, struct ceph_msg *msg) +static void finish_read(struct ceph_osd_request *req) { struct inode *inode = req->r_inode; struct ceph_osd_data *osd_data; - int rc = req->r_result; - int bytes = le32_to_cpu(msg->hdr.data_len); + int rc = req->r_result <= 0 ? req->r_result : 0; + int bytes = req->r_result >= 0 ? req->r_result : 0; int num_pages; int i; @@ -598,8 +598,7 @@ static void ceph_release_pages(struct page **pages, int num) * If we get an error, set the mapping error bit, but not the individual * page error bits. */ -static void writepages_finish(struct ceph_osd_request *req, - struct ceph_msg *msg) +static void writepages_finish(struct ceph_osd_request *req) { struct inode *inode = req->r_inode; struct ceph_inode_info *ci = ceph_inode(inode); diff --git a/fs/ceph/file.c b/fs/ceph/file.c index 52e4b72dd5de..e75fd0b028e9 100644 --- a/fs/ceph/file.c +++ b/fs/ceph/file.c @@ -616,8 +616,7 @@ static void ceph_aio_complete(struct inode *inode, kfree(aio_req); } -static void ceph_aio_complete_req(struct ceph_osd_request *req, - struct ceph_msg *msg) +static void ceph_aio_complete_req(struct ceph_osd_request *req) { int rc = req->r_result; struct inode *inode = req->r_inode; @@ -740,7 +739,7 @@ static void ceph_aio_retry_work(struct work_struct *work) out: if (ret < 0) { req->r_result = ret; - ceph_aio_complete_req(req, NULL); + ceph_aio_complete_req(req); } ceph_put_snap_context(snapc); @@ -961,7 +960,7 @@ ceph_direct_read_write(struct kiocb *iocb, struct iov_iter *iter, req, false); if (ret < 0) { req->r_result = ret; - ceph_aio_complete_req(req, NULL); + ceph_aio_complete_req(req); } } return -EIOCBQUEUED; diff --git a/include/linux/ceph/osd_client.h b/include/linux/ceph/osd_client.h index 67a37d98e0ca..3bebd60e7f9f 100644 --- a/include/linux/ceph/osd_client.h +++ b/include/linux/ceph/osd_client.h @@ -20,8 +20,7 @@ struct ceph_osd_client; /* * completion callback for async writepages */ -typedef void (*ceph_osdc_callback_t)(struct ceph_osd_request *, - struct ceph_msg *); +typedef void (*ceph_osdc_callback_t)(struct ceph_osd_request *); typedef void (*ceph_osdc_unsafe_callback_t)(struct ceph_osd_request *, bool); #define CEPH_HOMELESS_OSD -1 diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c index 8a008f083283..2a30c0bb3045 100644 --- a/net/ceph/osd_client.c +++ b/net/ceph/osd_client.c @@ -2048,7 +2048,7 @@ static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg) result >= 0 && !(flags & CEPH_OSD_FLAG_ONDISK)) req->r_unsafe_callback(req, true); if (req->r_callback) - req->r_callback(req, msg); + req->r_callback(req); else complete_all(&req->r_completion); } @@ -2072,7 +2072,7 @@ bad_put: req->r_result = -EIO; __unregister_request(osdc, req); if (req->r_callback) - req->r_callback(req, msg); + req->r_callback(req); else complete_all(&req->r_completion); complete_request(req); -- cgit v1.3-8-gc7d7 From fe5da05e979830b43b115d8a18ead521d507c783 Mon Sep 17 00:00:00 2001 From: Ilya Dryomov Date: Thu, 28 Apr 2016 16:07:24 +0200 Subject: libceph: redo callbacks and factor out MOSDOpReply decoding If you specify ACK | ONDISK and set ->r_unsafe_callback, both ->r_callback and ->r_unsafe_callback(true) are called on ack. This is very confusing. Redo this so that only one of them is called: ->r_unsafe_callback(true), on ack ->r_unsafe_callback(false), on commit or ->r_callback, on ack|commit Decode everything in decode_MOSDOpReply() to reduce clutter. Signed-off-by: Ilya Dryomov --- fs/ceph/addr.c | 3 +- fs/ceph/file.c | 2 + include/linux/ceph/osd_client.h | 5 +- net/ceph/osd_client.c | 362 +++++++++++++++++++++++----------------- 4 files changed, 215 insertions(+), 157 deletions(-) (limited to 'fs') diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c index a11756a39471..f47418477629 100644 --- a/fs/ceph/addr.c +++ b/fs/ceph/addr.c @@ -1765,8 +1765,7 @@ static int __ceph_pool_perm_get(struct ceph_inode_info *ci, u32 pool) goto out_unlock; } - wr_req->r_flags = CEPH_OSD_FLAG_WRITE | - CEPH_OSD_FLAG_ACK | CEPH_OSD_FLAG_ONDISK; + wr_req->r_flags = CEPH_OSD_FLAG_WRITE | CEPH_OSD_FLAG_ACK; osd_req_op_init(wr_req, 0, CEPH_OSD_OP_CREATE, CEPH_OSD_OP_FLAG_EXCL); ceph_oloc_copy(&wr_req->r_base_oloc, &rd_req->r_base_oloc); ceph_oid_copy(&wr_req->r_base_oid, &rd_req->r_base_oid); diff --git a/fs/ceph/file.c b/fs/ceph/file.c index e75fd0b028e9..30fd49eb25b4 100644 --- a/fs/ceph/file.c +++ b/fs/ceph/file.c @@ -770,6 +770,8 @@ static void ceph_sync_write_unsafe(struct ceph_osd_request *req, bool unsafe) list_add_tail(&req->r_unsafe_item, &ci->i_unsafe_writes); spin_unlock(&ci->i_unsafe_lock); + + complete_all(&req->r_completion); } else { spin_lock(&ci->i_unsafe_lock); list_del_init(&req->r_unsafe_item); diff --git a/include/linux/ceph/osd_client.h b/include/linux/ceph/osd_client.h index 3bebd60e7f9f..2415dc0cb008 100644 --- a/include/linux/ceph/osd_client.h +++ b/include/linux/ceph/osd_client.h @@ -162,13 +162,14 @@ struct ceph_osd_request { unsigned int r_num_ops; int r_result; - int r_got_reply; + bool r_got_reply; int r_linger; struct ceph_osd_client *r_osdc; struct kref r_kref; bool r_mempool; - struct completion r_completion, r_safe_completion; + struct completion r_completion; + struct completion r_safe_completion; /* fsync waiter */ ceph_osdc_callback_t r_callback; ceph_osdc_unsafe_callback_t r_unsafe_callback; struct list_head r_unsafe_item; diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c index 2a30c0bb3045..baf2844b00d6 100644 --- a/net/ceph/osd_client.c +++ b/net/ceph/osd_client.c @@ -1693,6 +1693,14 @@ static int __ceph_osdc_start_request(struct ceph_osd_client *osdc, return 0; } +static void __complete_request(struct ceph_osd_request *req) +{ + if (req->r_callback) + req->r_callback(req); + else + complete_all(&req->r_completion); +} + /* * Timeout callback, called every N seconds when 1 or more osd * requests has been active for more than N seconds. When this @@ -1875,107 +1883,76 @@ e_inval: goto out; } -static void complete_request(struct ceph_osd_request *req) -{ - complete_all(&req->r_safe_completion); /* fsync waiter */ -} +struct MOSDOpReply { + struct ceph_pg pgid; + u64 flags; + int result; + u32 epoch; + int num_ops; + u32 outdata_len[CEPH_OSD_MAX_OPS]; + s32 rval[CEPH_OSD_MAX_OPS]; + int retry_attempt; + struct ceph_eversion replay_version; + u64 user_version; + struct ceph_request_redirect redirect; +}; -/* - * handle osd op reply. either call the callback if it is specified, - * or do the completion to wake up the waiting thread. - */ -static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg) +static int decode_MOSDOpReply(const struct ceph_msg *msg, struct MOSDOpReply *m) { - void *p, *end; - struct ceph_osd_request *req; - struct ceph_request_redirect redir; - u64 tid; - int object_len; - unsigned int numops; - int payload_len, flags; - s32 result; - s32 retry_attempt; - struct ceph_pg pg; - int err; - u32 reassert_epoch; - u64 reassert_version; - u32 osdmap_epoch; - int already_completed; - u32 bytes; + void *p = msg->front.iov_base; + void *const end = p + msg->front.iov_len; + u16 version = le16_to_cpu(msg->hdr.version); + struct ceph_eversion bad_replay_version; u8 decode_redir; - unsigned int i; - - tid = le64_to_cpu(msg->hdr.tid); - dout("handle_reply %p tid %llu\n", msg, tid); - - p = msg->front.iov_base; - end = p + msg->front.iov_len; + u32 len; + int ret; + int i; - ceph_decode_need(&p, end, 4, bad); - object_len = ceph_decode_32(&p); - ceph_decode_need(&p, end, object_len, bad); - p += object_len; + ceph_decode_32_safe(&p, end, len, e_inval); + ceph_decode_need(&p, end, len, e_inval); + p += len; /* skip oid */ - err = ceph_decode_pgid(&p, end, &pg); - if (err) - goto bad; + ret = ceph_decode_pgid(&p, end, &m->pgid); + if (ret) + return ret; - ceph_decode_need(&p, end, 8 + 4 + 4 + 8 + 4, bad); - flags = ceph_decode_64(&p); - result = ceph_decode_32(&p); - reassert_epoch = ceph_decode_32(&p); - reassert_version = ceph_decode_64(&p); - osdmap_epoch = ceph_decode_32(&p); + ceph_decode_64_safe(&p, end, m->flags, e_inval); + ceph_decode_32_safe(&p, end, m->result, e_inval); + ceph_decode_need(&p, end, sizeof(bad_replay_version), e_inval); + memcpy(&bad_replay_version, p, sizeof(bad_replay_version)); + p += sizeof(bad_replay_version); + ceph_decode_32_safe(&p, end, m->epoch, e_inval); - /* lookup */ - down_read(&osdc->map_sem); - mutex_lock(&osdc->request_mutex); - req = lookup_request(&osdc->requests, tid); - if (req == NULL) { - dout("handle_reply tid %llu dne\n", tid); - goto bad_mutex; - } - ceph_osdc_get_request(req); + ceph_decode_32_safe(&p, end, m->num_ops, e_inval); + if (m->num_ops > ARRAY_SIZE(m->outdata_len)) + goto e_inval; - dout("handle_reply %p tid %llu req %p result %d\n", msg, tid, - req, result); - - ceph_decode_need(&p, end, 4, bad_put); - numops = ceph_decode_32(&p); - if (numops > CEPH_OSD_MAX_OPS) - goto bad_put; - if (numops != req->r_num_ops) - goto bad_put; - payload_len = 0; - ceph_decode_need(&p, end, numops * sizeof(struct ceph_osd_op), bad_put); - for (i = 0; i < numops; i++) { + ceph_decode_need(&p, end, m->num_ops * sizeof(struct ceph_osd_op), + e_inval); + for (i = 0; i < m->num_ops; i++) { struct ceph_osd_op *op = p; - int len; - len = le32_to_cpu(op->payload_len); - req->r_ops[i].outdata_len = len; - dout(" op %d has %d bytes\n", i, len); - payload_len += len; + m->outdata_len[i] = le32_to_cpu(op->payload_len); p += sizeof(*op); } - bytes = le32_to_cpu(msg->hdr.data_len); - if (payload_len != bytes) { - pr_warn("sum of op payload lens %d != data_len %d\n", - payload_len, bytes); - goto bad_put; - } - ceph_decode_need(&p, end, 4 + numops * 4, bad_put); - retry_attempt = ceph_decode_32(&p); - for (i = 0; i < numops; i++) - req->r_ops[i].rval = ceph_decode_32(&p); + ceph_decode_32_safe(&p, end, m->retry_attempt, e_inval); + for (i = 0; i < m->num_ops; i++) + ceph_decode_32_safe(&p, end, m->rval[i], e_inval); - if (le16_to_cpu(msg->hdr.version) >= 6) { - p += 8 + 4; /* skip replay_version */ - p += 8; /* skip user_version */ + if (version >= 5) { + ceph_decode_need(&p, end, sizeof(m->replay_version), e_inval); + memcpy(&m->replay_version, p, sizeof(m->replay_version)); + p += sizeof(m->replay_version); + ceph_decode_64_safe(&p, end, m->user_version, e_inval); + } else { + m->replay_version = bad_replay_version; /* struct */ + m->user_version = le64_to_cpu(m->replay_version.version); + } - if (le16_to_cpu(msg->hdr.version) >= 7) - ceph_decode_8_safe(&p, end, decode_redir, bad_put); + if (version >= 6) { + if (version >= 7) + ceph_decode_8_safe(&p, end, decode_redir, e_inval); else decode_redir = 1; } else { @@ -1983,19 +1960,96 @@ static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg) } if (decode_redir) { - err = ceph_redirect_decode(&p, end, &redir); - if (err) - goto bad_put; + ret = ceph_redirect_decode(&p, end, &m->redirect); + if (ret) + return ret; } else { - redir.oloc.pool = -1; + ceph_oloc_init(&m->redirect.oloc); } - if (!ceph_oloc_empty(&redir.oloc)) { - dout("redirect pool %lld\n", redir.oloc.pool); + return 0; + +e_inval: + return -EINVAL; +} + +/* + * We are done with @req if + * - @m is a safe reply, or + * - @m is an unsafe reply and we didn't want a safe one + */ +static bool done_request(const struct ceph_osd_request *req, + const struct MOSDOpReply *m) +{ + return (m->result < 0 || + (m->flags & CEPH_OSD_FLAG_ONDISK) || + !(req->r_flags & CEPH_OSD_FLAG_ONDISK)); +} +/* + * handle osd op reply. either call the callback if it is specified, + * or do the completion to wake up the waiting thread. + * + * ->r_unsafe_callback is set? yes no + * + * first reply is OK (needed r_cb/r_completion, r_cb/r_completion, + * any or needed/got safe) r_safe_completion r_safe_completion + * + * first reply is unsafe r_unsafe_cb(true) (nothing) + * + * when we get the safe reply r_unsafe_cb(false), r_cb/r_completion, + * r_safe_completion r_safe_completion + */ +static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg) +{ + struct ceph_osd_request *req; + struct MOSDOpReply m; + u64 tid = le64_to_cpu(msg->hdr.tid); + u32 data_len = 0; + bool already_acked; + int ret; + int i; + + dout("%s msg %p tid %llu\n", __func__, msg, tid); + + down_read(&osdc->map_sem); + mutex_lock(&osdc->request_mutex); + req = lookup_request(&osdc->requests, tid); + if (!req) { + dout("%s no tid %llu\n", __func__, tid); + goto out_unlock; + } + ceph_osdc_get_request(req); + + ret = decode_MOSDOpReply(msg, &m); + if (ret) { + pr_err("failed to decode MOSDOpReply for tid %llu: %d\n", + req->r_tid, ret); + ceph_msg_dump(msg); + goto fail_request; + } + dout("%s req %p tid %llu flags 0x%llx pgid %llu.%x epoch %u attempt %d v %u'%llu uv %llu\n", + __func__, req, req->r_tid, m.flags, m.pgid.pool, m.pgid.seed, + m.epoch, m.retry_attempt, le32_to_cpu(m.replay_version.epoch), + le64_to_cpu(m.replay_version.version), m.user_version); + + if (m.retry_attempt >= 0) { + if (m.retry_attempt != req->r_attempts - 1) { + dout("req %p tid %llu retry_attempt %d != %d, ignoring\n", + req, req->r_tid, m.retry_attempt, + req->r_attempts - 1); + goto out_put; + } + } else { + WARN_ON(1); /* MOSDOpReply v4 is assumed */ + } + + if (!ceph_oloc_empty(&m.redirect.oloc)) { + dout("req %p tid %llu redirect pool %lld\n", req, req->r_tid, + m.redirect.oloc.pool); __unregister_request(osdc, req); - ceph_oloc_copy(&req->r_t.target_oloc, &redir.oloc); + ceph_oloc_copy(&req->r_t.target_oloc, &m.redirect.oloc); /* * Start redirect requests with nofail=true. If @@ -2005,85 +2059,85 @@ static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg) * successfully. In the future we might want to follow * original request's nofail setting here. */ - err = __ceph_osdc_start_request(osdc, req, true); - BUG_ON(err); + ret = __ceph_osdc_start_request(osdc, req, true); + BUG_ON(ret); - goto out_unlock; + goto out_put; } - already_completed = req->r_got_reply; - if (!req->r_got_reply) { - req->r_result = result; - dout("handle_reply result %d bytes %d\n", req->r_result, - bytes); - if (req->r_result == 0) - req->r_result = bytes; - - /* in case this is a write and we need to replay, */ - req->r_replay_version.epoch = cpu_to_le32(reassert_epoch); - req->r_replay_version.version = cpu_to_le64(reassert_version); - - req->r_got_reply = 1; - } else if ((flags & CEPH_OSD_FLAG_ONDISK) == 0) { - dout("handle_reply tid %llu dup ack\n", tid); - goto out_unlock; + if (m.num_ops != req->r_num_ops) { + pr_err("num_ops %d != %d for tid %llu\n", m.num_ops, + req->r_num_ops, req->r_tid); + goto fail_request; } - - dout("handle_reply tid %llu flags %d\n", tid, flags); - - if (req->r_linger && (flags & CEPH_OSD_FLAG_ONDISK)) - __register_linger_request(osdc, req); - - /* either this is a read, or we got the safe response */ - if (result < 0 || - (flags & CEPH_OSD_FLAG_ONDISK) || - ((flags & CEPH_OSD_FLAG_WRITE) == 0)) + for (i = 0; i < req->r_num_ops; i++) { + dout(" req %p tid %llu op %d rval %d len %u\n", req, + req->r_tid, i, m.rval[i], m.outdata_len[i]); + req->r_ops[i].rval = m.rval[i]; + req->r_ops[i].outdata_len = m.outdata_len[i]; + data_len += m.outdata_len[i]; + } + if (data_len != le32_to_cpu(msg->hdr.data_len)) { + pr_err("sum of lens %u != %u for tid %llu\n", data_len, + le32_to_cpu(msg->hdr.data_len), req->r_tid); + goto fail_request; + } + dout("%s req %p tid %llu acked %d result %d data_len %u\n", __func__, + req, req->r_tid, req->r_got_reply, m.result, data_len); + + already_acked = req->r_got_reply; + if (!already_acked) { + req->r_result = m.result ?: data_len; + req->r_replay_version = m.replay_version; /* struct */ + req->r_got_reply = true; + } else if (!(m.flags & CEPH_OSD_FLAG_ONDISK)) { + dout("req %p tid %llu dup ack\n", req, req->r_tid); + goto out_put; + } + + if (done_request(req, &m)) { __unregister_request(osdc, req); + if (req->r_linger) { + WARN_ON(req->r_unsafe_callback); + __register_linger_request(osdc, req); + } + } mutex_unlock(&osdc->request_mutex); up_read(&osdc->map_sem); - if (!already_completed) { - if (req->r_unsafe_callback && - result >= 0 && !(flags & CEPH_OSD_FLAG_ONDISK)) - req->r_unsafe_callback(req, true); - if (req->r_callback) - req->r_callback(req); - else - complete_all(&req->r_completion); - } - - if (flags & CEPH_OSD_FLAG_ONDISK) { - if (req->r_unsafe_callback && already_completed) + if (done_request(req, &m)) { + if (already_acked && req->r_unsafe_callback) { + dout("req %p tid %llu safe-cb\n", req, req->r_tid); req->r_unsafe_callback(req, false); - complete_request(req); + } else { + dout("req %p tid %llu cb\n", req, req->r_tid); + __complete_request(req); + } + } else { + if (req->r_unsafe_callback) { + dout("req %p tid %llu unsafe-cb\n", req, req->r_tid); + req->r_unsafe_callback(req, true); + } else { + WARN_ON(1); + } } + if (m.flags & CEPH_OSD_FLAG_ONDISK) + complete_all(&req->r_safe_completion); -out: - dout("req=%p req->r_linger=%d\n", req, req->r_linger); ceph_osdc_put_request(req); return; -out_unlock: - mutex_unlock(&osdc->request_mutex); - up_read(&osdc->map_sem); - goto out; -bad_put: +fail_request: req->r_result = -EIO; __unregister_request(osdc, req); - if (req->r_callback) - req->r_callback(req); - else - complete_all(&req->r_completion); - complete_request(req); + __complete_request(req); + complete_all(&req->r_safe_completion); +out_put: ceph_osdc_put_request(req); -bad_mutex: +out_unlock: mutex_unlock(&osdc->request_mutex); up_read(&osdc->map_sem); -bad: - pr_err("corrupt osd_op_reply got %d %d\n", - (int)msg->front.iov_len, le32_to_cpu(msg->hdr.front_len)); - ceph_msg_dump(msg); } static void reset_changed_osds(struct ceph_osd_client *osdc) @@ -2591,7 +2645,9 @@ int ceph_osdc_wait_request(struct ceph_osd_client *osdc, if (rc < 0) { dout("%s %p tid %llu interrupted\n", __func__, req, req->r_tid); ceph_osdc_cancel_request(req); - complete_request(req); + + /* kludge - need to to wake ceph_osdc_sync() */ + complete_all(&req->r_safe_completion); return rc; } -- cgit v1.3-8-gc7d7 From 5aea3dcd50215fa9563270251ad7323e2f2490ee Mon Sep 17 00:00:00 2001 From: Ilya Dryomov Date: Thu, 28 Apr 2016 16:07:26 +0200 Subject: libceph: a major OSD client update This is a major sync up, up to ~Jewel. The highlights are: - per-session request trees (vs a global per-client tree) - per-session locking (vs a global per-client rwlock) - homeless OSD session - no ad-hoc global per-client lists - support for pool quotas - foundation for watch/notify v2 support - foundation for map check (pool deletion detection) support The switchover is incomplete: lingering requests can be setup and teared down but aren't ever reestablished. This functionality is restored with the introduction of the new lingering infrastructure (ceph_osd_linger_request, linger_work, etc) in a later commit. Signed-off-by: Ilya Dryomov --- fs/ceph/ioctl.c | 8 +- fs/ceph/xattr.c | 8 +- include/linux/ceph/osd_client.h | 18 +- net/ceph/debugfs.c | 34 +- net/ceph/osd_client.c | 1164 +++++++++++++++++++-------------------- 5 files changed, 602 insertions(+), 630 deletions(-) (limited to 'fs') diff --git a/fs/ceph/ioctl.c b/fs/ceph/ioctl.c index 1831ad6cf066..be6b1657b1af 100644 --- a/fs/ceph/ioctl.c +++ b/fs/ceph/ioctl.c @@ -193,12 +193,12 @@ static long ceph_ioctl_get_dataloc(struct file *file, void __user *arg) if (copy_from_user(&dl, arg, sizeof(dl))) return -EFAULT; - down_read(&osdc->map_sem); + down_read(&osdc->lock); r = ceph_calc_file_object_mapping(&ci->i_layout, dl.file_offset, len, &dl.object_no, &dl.object_offset, &olen); if (r < 0) { - up_read(&osdc->map_sem); + up_read(&osdc->lock); return -EIO; } dl.file_offset -= dl.object_offset; @@ -217,7 +217,7 @@ static long ceph_ioctl_get_dataloc(struct file *file, void __user *arg) r = ceph_object_locator_to_pg(osdc->osdmap, &oid, &oloc, &pgid); if (r < 0) { - up_read(&osdc->map_sem); + up_read(&osdc->lock); return r; } @@ -230,7 +230,7 @@ static long ceph_ioctl_get_dataloc(struct file *file, void __user *arg) } else { memset(&dl.osd_addr, 0, sizeof(dl.osd_addr)); } - up_read(&osdc->map_sem); + up_read(&osdc->lock); /* send result back to user */ if (copy_to_user(arg, &dl, sizeof(dl))) diff --git a/fs/ceph/xattr.c b/fs/ceph/xattr.c index 9410abdef3ce..5afabc4bf4c7 100644 --- a/fs/ceph/xattr.c +++ b/fs/ceph/xattr.c @@ -75,7 +75,7 @@ static size_t ceph_vxattrcb_layout(struct ceph_inode_info *ci, char *val, char buf[128]; dout("ceph_vxattrcb_layout %p\n", &ci->vfs_inode); - down_read(&osdc->map_sem); + down_read(&osdc->lock); pool_name = ceph_pg_pool_name_by_id(osdc->osdmap, pool); if (pool_name) { size_t len = strlen(pool_name); @@ -107,7 +107,7 @@ static size_t ceph_vxattrcb_layout(struct ceph_inode_info *ci, char *val, ret = -ERANGE; } } - up_read(&osdc->map_sem); + up_read(&osdc->lock); return ret; } @@ -141,13 +141,13 @@ static size_t ceph_vxattrcb_layout_pool(struct ceph_inode_info *ci, s64 pool = ceph_file_layout_pg_pool(ci->i_layout); const char *pool_name; - down_read(&osdc->map_sem); + down_read(&osdc->lock); pool_name = ceph_pg_pool_name_by_id(osdc->osdmap, pool); if (pool_name) ret = snprintf(val, size, "%s", pool_name); else ret = snprintf(val, size, "%lld", (unsigned long long)pool); - up_read(&osdc->map_sem); + up_read(&osdc->lock); return ret; } diff --git a/include/linux/ceph/osd_client.h b/include/linux/ceph/osd_client.h index 486d681694c4..342f22f1f040 100644 --- a/include/linux/ceph/osd_client.h +++ b/include/linux/ceph/osd_client.h @@ -33,12 +33,13 @@ struct ceph_osd { int o_incarnation; struct rb_node o_node; struct ceph_connection o_con; - struct list_head o_requests; + struct rb_root o_requests; struct list_head o_linger_requests; struct list_head o_osd_lru; struct ceph_auth_handshake o_auth; unsigned long lru_ttl; struct list_head o_keepalive_item; + struct mutex lock; }; #define CEPH_OSD_SLAB_OPS 2 @@ -144,8 +145,6 @@ struct ceph_osd_request_target { struct ceph_osd_request { u64 r_tid; /* unique for this client */ struct rb_node r_node; - struct list_head r_req_lru_item; - struct list_head r_osd_item; struct list_head r_linger_item; struct list_head r_linger_osd_item; struct ceph_osd *r_osd; @@ -219,19 +218,16 @@ struct ceph_osd_client { struct ceph_client *client; struct ceph_osdmap *osdmap; /* current map */ - struct rw_semaphore map_sem; + struct rw_semaphore lock; - struct mutex request_mutex; struct rb_root osds; /* osds */ struct list_head osd_lru; /* idle osds */ spinlock_t osd_lru_lock; - u64 last_tid; /* tid of last request */ - struct rb_root requests; /* pending requests */ - struct list_head req_lru; /* in-flight lru */ - struct list_head req_unsent; /* unsent/need-resend queue */ - struct list_head req_notarget; /* map to no osd */ struct list_head req_linger; /* lingering requests */ - int num_requests; + struct ceph_osd homeless_osd; + atomic64_t last_tid; /* tid of last request */ + atomic_t num_requests; + atomic_t num_homeless; struct delayed_work timeout_work; struct delayed_work osds_timeout_work; #ifdef CONFIG_DEBUG_FS diff --git a/net/ceph/debugfs.c b/net/ceph/debugfs.c index 6d3ff713edeb..61dbd9de4650 100644 --- a/net/ceph/debugfs.c +++ b/net/ceph/debugfs.c @@ -182,21 +182,39 @@ static void dump_request(struct seq_file *s, struct ceph_osd_request *req) seq_putc(s, '\n'); } +static void dump_requests(struct seq_file *s, struct ceph_osd *osd) +{ + struct rb_node *n; + + mutex_lock(&osd->lock); + for (n = rb_first(&osd->o_requests); n; n = rb_next(n)) { + struct ceph_osd_request *req = + rb_entry(n, struct ceph_osd_request, r_node); + + dump_request(s, req); + } + + mutex_unlock(&osd->lock); +} + static int osdc_show(struct seq_file *s, void *pp) { struct ceph_client *client = s->private; struct ceph_osd_client *osdc = &client->osdc; - struct rb_node *p; - - mutex_lock(&osdc->request_mutex); - for (p = rb_first(&osdc->requests); p; p = rb_next(p)) { - struct ceph_osd_request *req; + struct rb_node *n; - req = rb_entry(p, struct ceph_osd_request, r_node); + down_read(&osdc->lock); + seq_printf(s, "REQUESTS %d homeless %d\n", + atomic_read(&osdc->num_requests), + atomic_read(&osdc->num_homeless)); + for (n = rb_first(&osdc->osds); n; n = rb_next(n)) { + struct ceph_osd *osd = rb_entry(n, struct ceph_osd, o_node); - dump_request(s, req); + dump_requests(s, osd); } - mutex_unlock(&osdc->request_mutex); + dump_requests(s, &osdc->homeless_osd); + + up_read(&osdc->lock); return 0; } diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c index d1c8e06f1261..4c856c87b1a9 100644 --- a/net/ceph/osd_client.c +++ b/net/ceph/osd_client.c @@ -25,16 +25,6 @@ static struct kmem_cache *ceph_osd_request_cache; static const struct ceph_connection_operations osd_con_ops; -static void __send_queued(struct ceph_osd_client *osdc); -static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd); -static void __register_request(struct ceph_osd_client *osdc, - struct ceph_osd_request *req); -static void __unregister_request(struct ceph_osd_client *osdc, - struct ceph_osd_request *req); -static void __unregister_linger_request(struct ceph_osd_client *osdc, - struct ceph_osd_request *req); -static void __enqueue_request(struct ceph_osd_request *req); - /* * Implement client access to distributed object storage cluster. * @@ -53,6 +43,43 @@ static void __enqueue_request(struct ceph_osd_request *req); * channel with an OSD is reset. */ +static void link_request(struct ceph_osd *osd, struct ceph_osd_request *req); +static void unlink_request(struct ceph_osd *osd, struct ceph_osd_request *req); + +#if 1 +static inline bool rwsem_is_wrlocked(struct rw_semaphore *sem) +{ + bool wrlocked = true; + + if (unlikely(down_read_trylock(sem))) { + wrlocked = false; + up_read(sem); + } + + return wrlocked; +} +static inline void verify_osdc_locked(struct ceph_osd_client *osdc) +{ + WARN_ON(!rwsem_is_locked(&osdc->lock)); +} +static inline void verify_osdc_wrlocked(struct ceph_osd_client *osdc) +{ + WARN_ON(!rwsem_is_wrlocked(&osdc->lock)); +} +static inline void verify_osd_locked(struct ceph_osd *osd) +{ + struct ceph_osd_client *osdc = osd->o_osdc; + + WARN_ON(!(mutex_is_locked(&osd->lock) && + rwsem_is_locked(&osdc->lock)) && + !rwsem_is_wrlocked(&osdc->lock)); +} +#else +static inline void verify_osdc_locked(struct ceph_osd_client *osdc) { } +static inline void verify_osdc_wrlocked(struct ceph_osd_client *osdc) { } +static inline void verify_osd_locked(struct ceph_osd *osd) { } +#endif + /* * calculate the mapping of a file extent onto an object, and fill out the * request accordingly. shorten extent as necessary if it crosses an @@ -336,18 +363,14 @@ static void ceph_osdc_release_request(struct kref *kref) dout("%s %p (r_request %p r_reply %p)\n", __func__, req, req->r_request, req->r_reply); WARN_ON(!RB_EMPTY_NODE(&req->r_node)); - WARN_ON(!list_empty(&req->r_req_lru_item)); - WARN_ON(!list_empty(&req->r_osd_item)); WARN_ON(!list_empty(&req->r_linger_item)); WARN_ON(!list_empty(&req->r_linger_osd_item)); WARN_ON(req->r_osd); if (req->r_request) ceph_msg_put(req->r_request); - if (req->r_reply) { - ceph_msg_revoke_incoming(req->r_reply); + if (req->r_reply) ceph_msg_put(req->r_reply); - } for (which = 0; which < req->r_num_ops; which++) osd_req_op_data_release(req, which); @@ -418,8 +441,6 @@ struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc, INIT_LIST_HEAD(&req->r_unsafe_item); INIT_LIST_HEAD(&req->r_linger_item); INIT_LIST_HEAD(&req->r_linger_osd_item); - INIT_LIST_HEAD(&req->r_req_lru_item); - INIT_LIST_HEAD(&req->r_osd_item); target_init(&req->r_t); @@ -869,141 +890,11 @@ static bool osd_homeless(struct ceph_osd *osd) return osd->o_osd == CEPH_HOMELESS_OSD; } -static struct ceph_osd_request * -__lookup_request_ge(struct ceph_osd_client *osdc, - u64 tid) -{ - struct ceph_osd_request *req; - struct rb_node *n = osdc->requests.rb_node; - - while (n) { - req = rb_entry(n, struct ceph_osd_request, r_node); - if (tid < req->r_tid) { - if (!n->rb_left) - return req; - n = n->rb_left; - } else if (tid > req->r_tid) { - n = n->rb_right; - } else { - return req; - } - } - return NULL; -} - -static void __kick_linger_request(struct ceph_osd_request *req) -{ - struct ceph_osd_client *osdc = req->r_osdc; - struct ceph_osd *osd = req->r_osd; - - /* - * Linger requests need to be resent with a new tid to avoid - * the dup op detection logic on the OSDs. Achieve this with - * a re-register dance instead of open-coding. - */ - ceph_osdc_get_request(req); - if (!list_empty(&req->r_linger_item)) - __unregister_linger_request(osdc, req); - else - __unregister_request(osdc, req); - __register_request(osdc, req); - ceph_osdc_put_request(req); - - /* - * Unless request has been registered as both normal and - * lingering, __unregister{,_linger}_request clears r_osd. - * However, here we need to preserve r_osd to make sure we - * requeue on the same OSD. - */ - WARN_ON(req->r_osd || !osd); - req->r_osd = osd; - - dout("%s requeueing %p tid %llu\n", __func__, req, req->r_tid); - __enqueue_request(req); -} - -/* - * Resubmit requests pending on the given osd. - */ -static void __kick_osd_requests(struct ceph_osd_client *osdc, - struct ceph_osd *osd) -{ - struct ceph_osd_request *req, *nreq; - LIST_HEAD(resend); - LIST_HEAD(resend_linger); - int err; - - dout("%s osd%d\n", __func__, osd->o_osd); - err = __reset_osd(osdc, osd); - if (err) - return; - - /* - * Build up a list of requests to resend by traversing the - * osd's list of requests. Requests for a given object are - * sent in tid order, and that is also the order they're - * kept on this list. Therefore all requests that are in - * flight will be found first, followed by all requests that - * have not yet been sent. And to resend requests while - * preserving this order we will want to put any sent - * requests back on the front of the osd client's unsent - * list. - * - * So we build a separate ordered list of already-sent - * requests for the affected osd and splice it onto the - * front of the osd client's unsent list. Once we've seen a - * request that has not yet been sent we're done. Those - * requests are already sitting right where they belong. - */ - list_for_each_entry(req, &osd->o_requests, r_osd_item) { - if (!req->r_sent) - break; - - if (!req->r_linger) { - dout("%s requeueing %p tid %llu\n", __func__, req, - req->r_tid); - list_move_tail(&req->r_req_lru_item, &resend); - req->r_flags |= CEPH_OSD_FLAG_RETRY; - } else { - list_move_tail(&req->r_req_lru_item, &resend_linger); - } - } - list_splice(&resend, &osdc->req_unsent); - - /* - * Both registered and not yet registered linger requests are - * enqueued with a new tid on the same OSD. We add/move them - * to req_unsent/o_requests at the end to keep things in tid - * order. - */ - list_for_each_entry_safe(req, nreq, &osd->o_linger_requests, - r_linger_osd_item) { - WARN_ON(!list_empty(&req->r_req_lru_item)); - __kick_linger_request(req); - } - - list_for_each_entry_safe(req, nreq, &resend_linger, r_req_lru_item) - __kick_linger_request(req); -} - -/* - * If the osd connection drops, we need to resubmit all requests. - */ -static void osd_reset(struct ceph_connection *con) +static bool osd_registered(struct ceph_osd *osd) { - struct ceph_osd *osd = con->private; - struct ceph_osd_client *osdc; + verify_osdc_locked(osd->o_osdc); - if (!osd) - return; - dout("osd_reset osd%d\n", osd->o_osd); - osdc = osd->o_osdc; - down_read(&osdc->map_sem); - mutex_lock(&osdc->request_mutex); - __kick_osd_requests(osdc, osd); - __send_queued(osdc); - mutex_unlock(&osdc->request_mutex); - up_read(&osdc->map_sem); + return !RB_EMPTY_NODE(&osd->o_node); } /* @@ -1013,17 +904,18 @@ static void osd_init(struct ceph_osd *osd) { atomic_set(&osd->o_ref, 1); RB_CLEAR_NODE(&osd->o_node); - INIT_LIST_HEAD(&osd->o_requests); + osd->o_requests = RB_ROOT; INIT_LIST_HEAD(&osd->o_linger_requests); INIT_LIST_HEAD(&osd->o_osd_lru); INIT_LIST_HEAD(&osd->o_keepalive_item); osd->o_incarnation = 1; + mutex_init(&osd->lock); } static void osd_cleanup(struct ceph_osd *osd) { WARN_ON(!RB_EMPTY_NODE(&osd->o_node)); - WARN_ON(!list_empty(&osd->o_requests)); + WARN_ON(!RB_EMPTY_ROOT(&osd->o_requests)); WARN_ON(!list_empty(&osd->o_linger_requests)); WARN_ON(!list_empty(&osd->o_osd_lru)); WARN_ON(!list_empty(&osd->o_keepalive_item)); @@ -1077,30 +969,6 @@ static void put_osd(struct ceph_osd *osd) DEFINE_RB_FUNCS(osd, struct ceph_osd, o_osd, o_node) -/* - * remove an osd from our map - */ -static void __remove_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd) -{ - dout("%s %p osd%d\n", __func__, osd, osd->o_osd); - WARN_ON(!list_empty(&osd->o_requests)); - WARN_ON(!list_empty(&osd->o_linger_requests)); - - list_del_init(&osd->o_osd_lru); - erase_osd(&osdc->osds, osd); -} - -static void remove_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd) -{ - dout("%s %p osd%d\n", __func__, osd, osd->o_osd); - - if (!RB_EMPTY_NODE(&osd->o_node)) { - ceph_con_close(&osd->o_con); - __remove_osd(osdc, osd); - put_osd(osd); - } -} - static void __move_osd_to_lru(struct ceph_osd *osd) { struct ceph_osd_client *osdc = osd->o_osdc; @@ -1117,7 +985,7 @@ static void __move_osd_to_lru(struct ceph_osd *osd) static void maybe_move_osd_to_lru(struct ceph_osd *osd) { - if (list_empty(&osd->o_requests) && + if (RB_EMPTY_ROOT(&osd->o_requests) && list_empty(&osd->o_linger_requests)) __move_osd_to_lru(osd); } @@ -1134,30 +1002,64 @@ static void __remove_osd_from_lru(struct ceph_osd *osd) spin_unlock(&osdc->osd_lru_lock); } +/* + * Close the connection and assign any leftover requests to the + * homeless session. + */ +static void close_osd(struct ceph_osd *osd) +{ + struct ceph_osd_client *osdc = osd->o_osdc; + struct rb_node *n; + + verify_osdc_wrlocked(osdc); + dout("%s osd %p osd%d\n", __func__, osd, osd->o_osd); + + ceph_con_close(&osd->o_con); + + for (n = rb_first(&osd->o_requests); n; ) { + struct ceph_osd_request *req = + rb_entry(n, struct ceph_osd_request, r_node); + + n = rb_next(n); /* unlink_request() */ + + dout(" reassigning req %p tid %llu\n", req, req->r_tid); + unlink_request(osd, req); + link_request(&osdc->homeless_osd, req); + } + + __remove_osd_from_lru(osd); + erase_osd(&osdc->osds, osd); + put_osd(osd); +} + /* * reset osd connect */ -static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd) +static int reopen_osd(struct ceph_osd *osd) { struct ceph_entity_addr *peer_addr; - dout("__reset_osd %p osd%d\n", osd, osd->o_osd); - if (list_empty(&osd->o_requests) && + dout("%s osd %p osd%d\n", __func__, osd, osd->o_osd); + + if (RB_EMPTY_ROOT(&osd->o_requests) && list_empty(&osd->o_linger_requests)) { - remove_osd(osdc, osd); + close_osd(osd); return -ENODEV; } - peer_addr = &osdc->osdmap->osd_addr[osd->o_osd]; + peer_addr = &osd->o_osdc->osdmap->osd_addr[osd->o_osd]; if (!memcmp(peer_addr, &osd->o_con.peer_addr, sizeof (*peer_addr)) && !ceph_con_opened(&osd->o_con)) { - struct ceph_osd_request *req; + struct rb_node *n; dout("osd addr hasn't changed and connection never opened, " "letting msgr retry\n"); /* touch each r_stamp for handle_timeout()'s benfit */ - list_for_each_entry(req, &osd->o_requests, r_osd_item) + for (n = rb_first(&osd->o_requests); n; n = rb_next(n)) { + struct ceph_osd_request *req = + rb_entry(n, struct ceph_osd_request, r_node); req->r_stamp = jiffies; + } return -EAGAIN; } @@ -1169,73 +1071,84 @@ static int __reset_osd(struct ceph_osd_client *osdc, struct ceph_osd *osd) return 0; } -/* - * Register request, assign tid. If this is the first request, set up - * the timeout event. - */ -static void __register_request(struct ceph_osd_client *osdc, - struct ceph_osd_request *req) -{ - req->r_tid = ++osdc->last_tid; - req->r_request->hdr.tid = cpu_to_le64(req->r_tid); - dout("__register_request %p tid %lld\n", req, req->r_tid); - insert_request(&osdc->requests, req); - ceph_osdc_get_request(req); - osdc->num_requests++; -} - -/* - * called under osdc->request_mutex - */ -static void __unregister_request(struct ceph_osd_client *osdc, - struct ceph_osd_request *req) +static struct ceph_osd *lookup_create_osd(struct ceph_osd_client *osdc, int o, + bool wrlocked) { - if (RB_EMPTY_NODE(&req->r_node)) { - dout("__unregister_request %p tid %lld not registered\n", - req, req->r_tid); - return; - } + struct ceph_osd *osd; - dout("__unregister_request %p tid %lld\n", req, req->r_tid); - erase_request(&osdc->requests, req); - osdc->num_requests--; + if (wrlocked) + verify_osdc_wrlocked(osdc); + else + verify_osdc_locked(osdc); - if (req->r_osd) { - /* make sure the original request isn't in flight. */ - ceph_msg_revoke(req->r_request); + if (o != CEPH_HOMELESS_OSD) + osd = lookup_osd(&osdc->osds, o); + else + osd = &osdc->homeless_osd; + if (!osd) { + if (!wrlocked) + return ERR_PTR(-EAGAIN); - list_del_init(&req->r_osd_item); - maybe_move_osd_to_lru(req->r_osd); - if (list_empty(&req->r_linger_osd_item)) - req->r_osd = NULL; + osd = create_osd(osdc, o); + insert_osd(&osdc->osds, osd); + ceph_con_open(&osd->o_con, CEPH_ENTITY_TYPE_OSD, osd->o_osd, + &osdc->osdmap->osd_addr[osd->o_osd]); } - list_del_init(&req->r_req_lru_item); - ceph_osdc_put_request(req); + dout("%s osdc %p osd%d -> osd %p\n", __func__, osdc, o, osd); + return osd; } /* - * Cancel a previously queued request message + * Create request <-> OSD session relation. + * + * @req has to be assigned a tid, @osd may be homeless. */ -static void __cancel_request(struct ceph_osd_request *req) +static void link_request(struct ceph_osd *osd, struct ceph_osd_request *req) { - if (req->r_sent && req->r_osd) { - ceph_msg_revoke(req->r_request); - req->r_sent = 0; - } + verify_osd_locked(osd); + WARN_ON(!req->r_tid || req->r_osd); + dout("%s osd %p osd%d req %p tid %llu\n", __func__, osd, osd->o_osd, + req, req->r_tid); + + if (!osd_homeless(osd)) + __remove_osd_from_lru(osd); + else + atomic_inc(&osd->o_osdc->num_homeless); + + get_osd(osd); + insert_request(&osd->o_requests, req); + req->r_osd = osd; } -static void __register_linger_request(struct ceph_osd_client *osdc, +static void unlink_request(struct ceph_osd *osd, struct ceph_osd_request *req) +{ + verify_osd_locked(osd); + WARN_ON(req->r_osd != osd); + dout("%s osd %p osd%d req %p tid %llu\n", __func__, osd, osd->o_osd, + req, req->r_tid); + + req->r_osd = NULL; + erase_request(&osd->o_requests, req); + put_osd(osd); + + if (!osd_homeless(osd)) + maybe_move_osd_to_lru(osd); + else + atomic_dec(&osd->o_osdc->num_homeless); +} + +static void __register_linger_request(struct ceph_osd *osd, struct ceph_osd_request *req) { dout("%s %p tid %llu\n", __func__, req, req->r_tid); WARN_ON(!req->r_linger); ceph_osdc_get_request(req); - list_add_tail(&req->r_linger_item, &osdc->req_linger); - if (req->r_osd) - list_add_tail(&req->r_linger_osd_item, - &req->r_osd->o_linger_requests); + list_add_tail(&req->r_linger_item, &osd->o_osdc->req_linger); + list_add_tail(&req->r_linger_osd_item, &osd->o_linger_requests); + __remove_osd_from_lru(osd); + req->r_osd = osd; } static void __unregister_linger_request(struct ceph_osd_client *osdc, @@ -1255,7 +1168,7 @@ static void __unregister_linger_request(struct ceph_osd_client *osdc, if (req->r_osd) { list_del_init(&req->r_linger_osd_item); maybe_move_osd_to_lru(req->r_osd); - if (list_empty(&req->r_osd_item)) + if (RB_EMPTY_ROOT(&req->r_osd->o_requests)) req->r_osd = NULL; } ceph_osdc_put_request(req); @@ -1291,11 +1204,20 @@ static bool have_pool_full(struct ceph_osd_client *osdc) return false; } +static bool pool_full(struct ceph_osd_client *osdc, s64 pool_id) +{ + struct ceph_pg_pool_info *pi; + + pi = ceph_pg_pool_by_id(osdc->osdmap, pool_id); + if (!pi) + return false; + + return __pool_full(pi); +} + /* * Returns whether a request should be blocked from being sent * based on the current osdmap and osd_client settings. - * - * Caller should hold map_sem for read. */ static bool target_should_be_paused(struct ceph_osd_client *osdc, const struct ceph_osd_request_target *t, @@ -1421,87 +1343,6 @@ out: return ct_res; } -static void __enqueue_request(struct ceph_osd_request *req) -{ - struct ceph_osd_client *osdc = req->r_osdc; - - dout("%s %p tid %llu to osd%d\n", __func__, req, req->r_tid, - req->r_osd ? req->r_osd->o_osd : -1); - - if (req->r_osd) { - __remove_osd_from_lru(req->r_osd); - list_add_tail(&req->r_osd_item, &req->r_osd->o_requests); - list_move_tail(&req->r_req_lru_item, &osdc->req_unsent); - } else { - list_move_tail(&req->r_req_lru_item, &osdc->req_notarget); - } -} - -/* - * Pick an osd (the first 'up' osd in the pg), allocate the osd struct - * (as needed), and set the request r_osd appropriately. If there is - * no up osd, set r_osd to NULL. Move the request to the appropriate list - * (unsent, homeless) or leave on in-flight lru. - * - * Return 0 if unchanged, 1 if changed, or negative on error. - * - * Caller should hold map_sem for read and request_mutex. - */ -static int __map_request(struct ceph_osd_client *osdc, - struct ceph_osd_request *req, int force_resend) -{ - enum calc_target_result ct_res; - int err; - - dout("map_request %p tid %lld\n", req, req->r_tid); - - ct_res = calc_target(osdc, &req->r_t, NULL, force_resend); - switch (ct_res) { - case CALC_TARGET_POOL_DNE: - list_move(&req->r_req_lru_item, &osdc->req_notarget); - return -EIO; - case CALC_TARGET_NO_ACTION: - return 0; /* no change */ - default: - BUG_ON(ct_res != CALC_TARGET_NEED_RESEND); - } - - dout("map_request tid %llu pgid %lld.%x osd%d (was osd%d)\n", - req->r_tid, req->r_t.pgid.pool, req->r_t.pgid.seed, req->r_t.osd, - req->r_osd ? req->r_osd->o_osd : -1); - - if (req->r_osd) { - __cancel_request(req); - list_del_init(&req->r_osd_item); - list_del_init(&req->r_linger_osd_item); - req->r_osd = NULL; - } - - req->r_osd = lookup_osd(&osdc->osds, req->r_t.osd); - if (!req->r_osd && req->r_t.osd >= 0) { - err = -ENOMEM; - req->r_osd = create_osd(osdc, req->r_t.osd); - if (!req->r_osd) { - list_move(&req->r_req_lru_item, &osdc->req_notarget); - goto out; - } - - dout("map_request osd %p is osd%d\n", req->r_osd, - req->r_osd->o_osd); - insert_osd(&osdc->osds, req->r_osd); - - ceph_con_open(&req->r_osd->o_con, - CEPH_ENTITY_TYPE_OSD, req->r_osd->o_osd, - &osdc->osdmap->osd_addr[req->r_osd->o_osd]); - } - - __enqueue_request(req); - err = 1; /* osd or pg changed */ - -out: - return err; -} - static void setup_request_data(struct ceph_osd_request *req, struct ceph_msg *msg) { @@ -1648,8 +1489,16 @@ static void send_request(struct ceph_osd_request *req) { struct ceph_osd *osd = req->r_osd; + verify_osd_locked(osd); WARN_ON(osd->o_osd != req->r_t.osd); + /* + * We may have a previously queued request message hanging + * around. Cancel it to avoid corrupting the msgr. + */ + if (req->r_sent) + ceph_msg_revoke(req->r_request); + req->r_flags |= CEPH_OSD_FLAG_KNOWN_REDIR; if (req->r_attempts) req->r_flags |= CEPH_OSD_FLAG_RETRY; @@ -1671,24 +1520,11 @@ static void send_request(struct ceph_osd_request *req) ceph_con_send(&osd->o_con, ceph_msg_get(req->r_request)); } -/* - * Send any requests in the queue (req_unsent). - */ -static void __send_queued(struct ceph_osd_client *osdc) -{ - struct ceph_osd_request *req, *tmp; - - dout("__send_queued\n"); - list_for_each_entry_safe(req, tmp, &osdc->req_unsent, r_req_lru_item) { - list_move_tail(&req->r_req_lru_item, &osdc->req_lru); - send_request(req); - } -} - static void maybe_request_map(struct ceph_osd_client *osdc) { bool continuous = false; + verify_osdc_locked(osdc); WARN_ON(!osdc->osdmap->epoch); if (ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_FULL) || @@ -1705,38 +1541,121 @@ static void maybe_request_map(struct ceph_osd_client *osdc) ceph_monc_renew_subs(&osdc->client->monc); } -/* - * Caller should hold map_sem for read and request_mutex. - */ -static int __ceph_osdc_start_request(struct ceph_osd_client *osdc, - struct ceph_osd_request *req, - bool nofail) +static void __submit_request(struct ceph_osd_request *req, bool wrlocked) { - int rc; + struct ceph_osd_client *osdc = req->r_osdc; + struct ceph_osd *osd; + bool need_send = false; + bool promoted = false; - __register_request(osdc, req); - req->r_sent = 0; - req->r_got_reply = 0; - rc = __map_request(osdc, req, 0); - if (rc < 0) { - if (nofail) { - dout("osdc_start_request failed map, " - " will retry %lld\n", req->r_tid); - rc = 0; - } else { - __unregister_request(osdc, req); - } - return rc; + WARN_ON(req->r_tid || req->r_got_reply); + dout("%s req %p wrlocked %d\n", __func__, req, wrlocked); + +again: + calc_target(osdc, &req->r_t, &req->r_last_force_resend, false); + osd = lookup_create_osd(osdc, req->r_t.osd, wrlocked); + if (IS_ERR(osd)) { + WARN_ON(PTR_ERR(osd) != -EAGAIN || wrlocked); + goto promote; } - if (req->r_osd == NULL) { - dout("send_request %p no up osds in pg\n", req); - ceph_monc_request_next_osdmap(&osdc->client->monc); + if ((req->r_flags & CEPH_OSD_FLAG_WRITE) && + ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_PAUSEWR)) { + dout("req %p pausewr\n", req); + req->r_t.paused = true; + maybe_request_map(osdc); + } else if ((req->r_flags & CEPH_OSD_FLAG_READ) && + ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_PAUSERD)) { + dout("req %p pauserd\n", req); + req->r_t.paused = true; + maybe_request_map(osdc); + } else if ((req->r_flags & CEPH_OSD_FLAG_WRITE) && + !(req->r_flags & (CEPH_OSD_FLAG_FULL_TRY | + CEPH_OSD_FLAG_FULL_FORCE)) && + (ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_FULL) || + pool_full(osdc, req->r_t.base_oloc.pool))) { + dout("req %p full/pool_full\n", req); + pr_warn_ratelimited("FULL or reached pool quota\n"); + req->r_t.paused = true; + maybe_request_map(osdc); + } else if (!osd_homeless(osd)) { + need_send = true; } else { - __send_queued(osdc); + maybe_request_map(osdc); } - return 0; + mutex_lock(&osd->lock); + /* + * Assign the tid atomically with send_request() to protect + * multiple writes to the same object from racing with each + * other, resulting in out of order ops on the OSDs. + */ + req->r_tid = atomic64_inc_return(&osdc->last_tid); + link_request(osd, req); + if (need_send) + send_request(req); + mutex_unlock(&osd->lock); + + if (promoted) + downgrade_write(&osdc->lock); + return; + +promote: + up_read(&osdc->lock); + down_write(&osdc->lock); + wrlocked = true; + promoted = true; + goto again; +} + +static void account_request(struct ceph_osd_request *req) +{ + unsigned int mask = CEPH_OSD_FLAG_ACK | CEPH_OSD_FLAG_ONDISK; + + if (req->r_flags & CEPH_OSD_FLAG_READ) { + WARN_ON(req->r_flags & mask); + req->r_flags |= CEPH_OSD_FLAG_ACK; + } else if (req->r_flags & CEPH_OSD_FLAG_WRITE) + WARN_ON(!(req->r_flags & mask)); + else + WARN_ON(1); + + WARN_ON(req->r_unsafe_callback && (req->r_flags & mask) != mask); + atomic_inc(&req->r_osdc->num_requests); +} + +static void submit_request(struct ceph_osd_request *req, bool wrlocked) +{ + ceph_osdc_get_request(req); + account_request(req); + __submit_request(req, wrlocked); +} + +static void __finish_request(struct ceph_osd_request *req) +{ + struct ceph_osd_client *osdc = req->r_osdc; + struct ceph_osd *osd = req->r_osd; + + verify_osd_locked(osd); + dout("%s req %p tid %llu\n", __func__, req, req->r_tid); + + unlink_request(osd, req); + atomic_dec(&osdc->num_requests); + + /* + * If an OSD has failed or returned and a request has been sent + * twice, it's possible to get a reply and end up here while the + * request message is queued for delivery. We will ignore the + * reply, so not a big deal, but better to try and catch it. + */ + ceph_msg_revoke(req->r_request); + ceph_msg_revoke_incoming(req->r_reply); +} + +static void finish_request(struct ceph_osd_request *req) +{ + __finish_request(req); + ceph_osdc_put_request(req); } static void __complete_request(struct ceph_osd_request *req) @@ -1747,6 +1666,13 @@ static void __complete_request(struct ceph_osd_request *req) complete_all(&req->r_completion); } +static void cancel_request(struct ceph_osd_request *req) +{ + dout("%s req %p tid %llu\n", __func__, req, req->r_tid); + + finish_request(req); +} + /* * Timeout callback, called every N seconds. When 1 or more OSD * requests has been active for more than N seconds, we send a keepalive @@ -1758,44 +1684,49 @@ static void handle_timeout(struct work_struct *work) struct ceph_osd_client *osdc = container_of(work, struct ceph_osd_client, timeout_work.work); struct ceph_options *opts = osdc->client->options; - struct ceph_osd_request *req; - struct ceph_osd *osd; - struct list_head slow_osds; - dout("timeout\n"); - down_read(&osdc->map_sem); - - ceph_monc_request_next_osdmap(&osdc->client->monc); + unsigned long cutoff = jiffies - opts->osd_keepalive_timeout; + LIST_HEAD(slow_osds); + struct rb_node *n, *p; - mutex_lock(&osdc->request_mutex); + dout("%s osdc %p\n", __func__, osdc); + down_write(&osdc->lock); /* * ping osds that are a bit slow. this ensures that if there * is a break in the TCP connection we will notice, and reopen * a connection with that osd (from the fault callback). */ - INIT_LIST_HEAD(&slow_osds); - list_for_each_entry(req, &osdc->req_lru, r_req_lru_item) { - if (time_before(jiffies, - req->r_stamp + opts->osd_keepalive_timeout)) - break; + for (n = rb_first(&osdc->osds); n; n = rb_next(n)) { + struct ceph_osd *osd = rb_entry(n, struct ceph_osd, o_node); + bool found = false; + + for (p = rb_first(&osd->o_requests); p; p = rb_next(p)) { + struct ceph_osd_request *req = + rb_entry(p, struct ceph_osd_request, r_node); + + if (time_before(req->r_stamp, cutoff)) { + dout(" req %p tid %llu on osd%d is laggy\n", + req, req->r_tid, osd->o_osd); + found = true; + } + } - osd = req->r_osd; - BUG_ON(!osd); - dout(" tid %llu is slow, will send keepalive on osd%d\n", - req->r_tid, osd->o_osd); - list_move_tail(&osd->o_keepalive_item, &slow_osds); + if (found) + list_move_tail(&osd->o_keepalive_item, &slow_osds); } + + if (atomic_read(&osdc->num_homeless) || !list_empty(&slow_osds)) + maybe_request_map(osdc); + while (!list_empty(&slow_osds)) { - osd = list_entry(slow_osds.next, struct ceph_osd, - o_keepalive_item); + struct ceph_osd *osd = list_first_entry(&slow_osds, + struct ceph_osd, + o_keepalive_item); list_del_init(&osd->o_keepalive_item); ceph_con_keepalive(&osd->o_con); } - __send_queued(osdc); - mutex_unlock(&osdc->request_mutex); - up_read(&osdc->map_sem); - + up_write(&osdc->lock); schedule_delayed_work(&osdc->timeout_work, osdc->client->options->osd_keepalive_timeout); } @@ -1809,18 +1740,17 @@ static void handle_osds_timeout(struct work_struct *work) struct ceph_osd *osd, *nosd; dout("%s osdc %p\n", __func__, osdc); - down_read(&osdc->map_sem); - mutex_lock(&osdc->request_mutex); - + down_write(&osdc->lock); list_for_each_entry_safe(osd, nosd, &osdc->osd_lru, o_osd_lru) { if (time_before(jiffies, osd->lru_ttl)) break; - remove_osd(osdc, osd); + WARN_ON(!RB_EMPTY_ROOT(&osd->o_requests)); + WARN_ON(!list_empty(&osd->o_linger_requests)); + close_osd(osd); } - mutex_unlock(&osdc->request_mutex); - up_read(&osdc->map_sem); + up_write(&osdc->lock); schedule_delayed_work(&osdc->osds_timeout_work, round_jiffies_relative(delay)); } @@ -2045,8 +1975,9 @@ static bool done_request(const struct ceph_osd_request *req, * when we get the safe reply r_unsafe_cb(false), r_cb/r_completion, * r_safe_completion r_safe_completion */ -static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg) +static void handle_reply(struct ceph_osd *osd, struct ceph_msg *msg) { + struct ceph_osd_client *osdc = osd->o_osdc; struct ceph_osd_request *req; struct MOSDOpReply m; u64 tid = le64_to_cpu(msg->hdr.tid); @@ -2057,14 +1988,19 @@ static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg) dout("%s msg %p tid %llu\n", __func__, msg, tid); - down_read(&osdc->map_sem); - mutex_lock(&osdc->request_mutex); - req = lookup_request(&osdc->requests, tid); + down_read(&osdc->lock); + if (!osd_registered(osd)) { + dout("%s osd%d unknown\n", __func__, osd->o_osd); + goto out_unlock_osdc; + } + WARN_ON(osd->o_osd != le64_to_cpu(msg->hdr.src.num)); + + mutex_lock(&osd->lock); + req = lookup_request(&osd->o_requests, tid); if (!req) { - dout("%s no tid %llu\n", __func__, tid); - goto out_unlock; + dout("%s osd%d tid %llu unknown\n", __func__, osd->o_osd, tid); + goto out_unlock_session; } - ceph_osdc_get_request(req); ret = decode_MOSDOpReply(msg, &m); if (ret) { @@ -2083,7 +2019,7 @@ static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg) dout("req %p tid %llu retry_attempt %d != %d, ignoring\n", req, req->r_tid, m.retry_attempt, req->r_attempts - 1); - goto out_put; + goto out_unlock_session; } } else { WARN_ON(1); /* MOSDOpReply v4 is assumed */ @@ -2092,22 +2028,14 @@ static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg) if (!ceph_oloc_empty(&m.redirect.oloc)) { dout("req %p tid %llu redirect pool %lld\n", req, req->r_tid, m.redirect.oloc.pool); - __unregister_request(osdc, req); + unlink_request(osd, req); + mutex_unlock(&osd->lock); ceph_oloc_copy(&req->r_t.target_oloc, &m.redirect.oloc); - - /* - * Start redirect requests with nofail=true. If - * mapping fails, request will end up on the notarget - * list, waiting for the new osdmap (which can take - * a while), even though the original request mapped - * successfully. In the future we might want to follow - * original request's nofail setting here. - */ - ret = __ceph_osdc_start_request(osdc, req, true); - BUG_ON(ret); - - goto out_put; + req->r_flags |= CEPH_OSD_FLAG_REDIRECTED; + req->r_tid = 0; + __submit_request(req, false); + goto out_unlock_osdc; } if (m.num_ops != req->r_num_ops) { @@ -2137,19 +2065,19 @@ static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg) req->r_got_reply = true; } else if (!(m.flags & CEPH_OSD_FLAG_ONDISK)) { dout("req %p tid %llu dup ack\n", req, req->r_tid); - goto out_put; + goto out_unlock_session; } if (done_request(req, &m)) { - __unregister_request(osdc, req); + __finish_request(req); if (req->r_linger) { WARN_ON(req->r_unsafe_callback); - __register_linger_request(osdc, req); + __register_linger_request(osd, req); } } - mutex_unlock(&osdc->request_mutex); - up_read(&osdc->map_sem); + mutex_unlock(&osd->lock); + up_read(&osdc->lock); if (done_request(req, &m)) { if (already_acked && req->r_unsafe_callback) { @@ -2175,14 +2103,13 @@ static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg) fail_request: req->r_result = -EIO; - __unregister_request(osdc, req); + __finish_request(req); __complete_request(req); complete_all(&req->r_safe_completion); -out_put: - ceph_osdc_put_request(req); -out_unlock: - mutex_unlock(&osdc->request_mutex); - up_read(&osdc->map_sem); +out_unlock_session: + mutex_unlock(&osd->lock); +out_unlock_osdc: + up_read(&osdc->lock); } static void set_pool_was_full(struct ceph_osd_client *osdc) @@ -2197,126 +2124,66 @@ static void set_pool_was_full(struct ceph_osd_client *osdc) } } -static void reset_changed_osds(struct ceph_osd_client *osdc) +static bool pool_cleared_full(struct ceph_osd_client *osdc, s64 pool_id) { - struct rb_node *p, *n; + struct ceph_pg_pool_info *pi; - dout("%s %p\n", __func__, osdc); - for (p = rb_first(&osdc->osds); p; p = n) { - struct ceph_osd *osd = rb_entry(p, struct ceph_osd, o_node); + pi = ceph_pg_pool_by_id(osdc->osdmap, pool_id); + if (!pi) + return false; - n = rb_next(p); - if (!ceph_osd_is_up(osdc->osdmap, osd->o_osd) || - memcmp(&osd->o_con.peer_addr, - ceph_osd_addr(osdc->osdmap, - osd->o_osd), - sizeof(struct ceph_entity_addr)) != 0) - __reset_osd(osdc, osd); - } + return pi->was_full && !__pool_full(pi); } /* - * Requeue requests whose mapping to an OSD has changed. If requests map to - * no osd, request a new map. - * - * Caller should hold map_sem for read. + * Requeue requests whose mapping to an OSD has changed. */ -static void kick_requests(struct ceph_osd_client *osdc, bool force_resend, - bool force_resend_writes) +static void scan_requests(struct ceph_osd *osd, + bool force_resend, + bool cleared_full, + bool check_pool_cleared_full, + struct rb_root *need_resend, + struct list_head *need_resend_linger) { - struct ceph_osd_request *req, *nreq; - struct rb_node *p; - int needmap = 0; - int err; - bool force_resend_req; - - dout("kick_requests %s %s\n", force_resend ? " (force resend)" : "", - force_resend_writes ? " (force resend writes)" : ""); - mutex_lock(&osdc->request_mutex); - for (p = rb_first(&osdc->requests); p; ) { - req = rb_entry(p, struct ceph_osd_request, r_node); - p = rb_next(p); - - /* - * For linger requests that have not yet been - * registered, move them to the linger list; they'll - * be sent to the osd in the loop below. Unregister - * the request before re-registering it as a linger - * request to ensure the __map_request() below - * will decide it needs to be sent. - */ - if (req->r_linger && list_empty(&req->r_linger_item)) { - dout("%p tid %llu restart on osd%d\n", - req, req->r_tid, - req->r_osd ? req->r_osd->o_osd : -1); - ceph_osdc_get_request(req); - __unregister_request(osdc, req); - __register_linger_request(osdc, req); - ceph_osdc_put_request(req); - continue; - } - - force_resend_req = force_resend || - (force_resend_writes && - req->r_flags & CEPH_OSD_FLAG_WRITE); - err = __map_request(osdc, req, force_resend_req); - if (err < 0) - continue; /* error */ - if (req->r_osd == NULL) { - dout("%p tid %llu maps to no osd\n", req, req->r_tid); - needmap++; /* request a newer map */ - } else if (err > 0) { - if (!req->r_linger) { - dout("%p tid %llu requeued on osd%d\n", req, - req->r_tid, - req->r_osd ? req->r_osd->o_osd : -1); - req->r_flags |= CEPH_OSD_FLAG_RETRY; - } - } - } - - list_for_each_entry_safe(req, nreq, &osdc->req_linger, - r_linger_item) { - dout("linger req=%p req->r_osd=%p\n", req, req->r_osd); - - err = __map_request(osdc, req, - force_resend || force_resend_writes); - dout("__map_request returned %d\n", err); - if (err < 0) - continue; /* hrm! */ - if (req->r_osd == NULL || err > 0) { - if (req->r_osd == NULL) { - dout("lingering %p tid %llu maps to no osd\n", - req, req->r_tid); - /* - * A homeless lingering request makes - * no sense, as it's job is to keep - * a particular OSD connection open. - * Request a newer map and kick the - * request, knowing that it won't be - * resent until we actually get a map - * that can tell us where to send it. - */ - needmap++; - } - - dout("kicking lingering %p tid %llu osd%d\n", req, - req->r_tid, req->r_osd ? req->r_osd->o_osd : -1); - __register_request(osdc, req); - __unregister_linger_request(osdc, req); + struct ceph_osd_client *osdc = osd->o_osdc; + struct rb_node *n; + bool force_resend_writes; + + for (n = rb_first(&osd->o_requests); n; ) { + struct ceph_osd_request *req = + rb_entry(n, struct ceph_osd_request, r_node); + enum calc_target_result ct_res; + + n = rb_next(n); /* unlink_request() */ + + dout("%s req %p tid %llu\n", __func__, req, req->r_tid); + ct_res = calc_target(osdc, &req->r_t, + &req->r_last_force_resend, false); + switch (ct_res) { + case CALC_TARGET_NO_ACTION: + force_resend_writes = cleared_full || + (check_pool_cleared_full && + pool_cleared_full(osdc, req->r_t.base_oloc.pool)); + if (!force_resend && + (!(req->r_flags & CEPH_OSD_FLAG_WRITE) || + !force_resend_writes)) + break; + + /* fall through */ + case CALC_TARGET_NEED_RESEND: + unlink_request(osd, req); + insert_request(need_resend, req); + break; + case CALC_TARGET_POOL_DNE: + break; } } - reset_changed_osds(osdc); - mutex_unlock(&osdc->request_mutex); - - if (needmap) { - dout("%d requests for down osds, need new map\n", needmap); - ceph_monc_request_next_osdmap(&osdc->client->monc); - } } static int handle_one_map(struct ceph_osd_client *osdc, - void *p, void *end, bool incremental) + void *p, void *end, bool incremental, + struct rb_root *need_resend, + struct list_head *need_resend_linger) { struct ceph_osdmap *newmap; struct rb_node *n; @@ -2362,11 +2229,51 @@ static int handle_one_map(struct ceph_osd_client *osdc, } was_full &= !ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_FULL); - kick_requests(osdc, skipped_map, was_full); + scan_requests(&osdc->homeless_osd, skipped_map, was_full, true, + need_resend, need_resend_linger); + + for (n = rb_first(&osdc->osds); n; ) { + struct ceph_osd *osd = rb_entry(n, struct ceph_osd, o_node); + + n = rb_next(n); /* close_osd() */ + + scan_requests(osd, skipped_map, was_full, true, need_resend, + need_resend_linger); + if (!ceph_osd_is_up(osdc->osdmap, osd->o_osd) || + memcmp(&osd->o_con.peer_addr, + ceph_osd_addr(osdc->osdmap, osd->o_osd), + sizeof(struct ceph_entity_addr))) + close_osd(osd); + } return 0; } +static void kick_requests(struct ceph_osd_client *osdc, + struct rb_root *need_resend, + struct list_head *need_resend_linger) +{ + struct rb_node *n; + + for (n = rb_first(need_resend); n; ) { + struct ceph_osd_request *req = + rb_entry(n, struct ceph_osd_request, r_node); + struct ceph_osd *osd; + + n = rb_next(n); + erase_request(need_resend, req); /* before link_request() */ + + WARN_ON(req->r_osd); + calc_target(osdc, &req->r_t, NULL, false); + osd = lookup_create_osd(osdc, req->r_t.osd, true); + link_request(osd, req); + if (!req->r_linger) { + if (!osd_homeless(osd) && !req->r_t.paused) + send_request(req); + } + } +} + /* * Process updated osd map. * @@ -2381,13 +2288,15 @@ void ceph_osdc_handle_map(struct ceph_osd_client *osdc, struct ceph_msg *msg) u32 nr_maps, maplen; u32 epoch; struct ceph_fsid fsid; + struct rb_root need_resend = RB_ROOT; + LIST_HEAD(need_resend_linger); bool handled_incremental = false; bool was_pauserd, was_pausewr; bool pauserd, pausewr; int err; dout("%s have %u\n", __func__, osdc->osdmap->epoch); - down_write(&osdc->map_sem); + down_write(&osdc->lock); /* verify fsid */ ceph_decode_need(&p, end, sizeof(fsid), bad); @@ -2412,7 +2321,8 @@ void ceph_osdc_handle_map(struct ceph_osd_client *osdc, struct ceph_msg *msg) osdc->osdmap->epoch + 1 == epoch) { dout("applying incremental map %u len %d\n", epoch, maplen); - err = handle_one_map(osdc, p, p + maplen, true); + err = handle_one_map(osdc, p, p + maplen, true, + &need_resend, &need_resend_linger); if (err) goto bad; handled_incremental = true; @@ -2443,7 +2353,8 @@ void ceph_osdc_handle_map(struct ceph_osd_client *osdc, struct ceph_msg *msg) osdc->osdmap->epoch); } else { dout("taking full map %u len %d\n", epoch, maplen); - err = handle_one_map(osdc, p, p + maplen, false); + err = handle_one_map(osdc, p, p + maplen, false, + &need_resend, &need_resend_linger); if (err) goto bad; } @@ -2464,20 +2375,60 @@ done: if (was_pauserd || was_pausewr || pauserd || pausewr) maybe_request_map(osdc); - mutex_lock(&osdc->request_mutex); - __send_queued(osdc); - mutex_unlock(&osdc->request_mutex); + kick_requests(osdc, &need_resend, &need_resend_linger); ceph_monc_got_map(&osdc->client->monc, CEPH_SUB_OSDMAP, osdc->osdmap->epoch); - up_write(&osdc->map_sem); + up_write(&osdc->lock); wake_up_all(&osdc->client->auth_wq); return; bad: pr_err("osdc handle_map corrupt msg\n"); ceph_msg_dump(msg); - up_write(&osdc->map_sem); + up_write(&osdc->lock); +} + +/* + * Resubmit requests pending on the given osd. + */ +static void kick_osd_requests(struct ceph_osd *osd) +{ + struct rb_node *n; + + for (n = rb_first(&osd->o_requests); n; n = rb_next(n)) { + struct ceph_osd_request *req = + rb_entry(n, struct ceph_osd_request, r_node); + + if (!req->r_linger) { + if (!req->r_t.paused) + send_request(req); + } + } +} + +/* + * If the osd connection drops, we need to resubmit all requests. + */ +static void osd_fault(struct ceph_connection *con) +{ + struct ceph_osd *osd = con->private; + struct ceph_osd_client *osdc = osd->o_osdc; + + dout("%s osd %p osd%d\n", __func__, osd, osd->o_osd); + + down_write(&osdc->lock); + if (!osd_registered(osd)) { + dout("%s osd%d unknown\n", __func__, osd->o_osd); + goto out_unlock; + } + + if (!reopen_osd(osd)) + kick_osd_requests(osd); + maybe_request_map(osdc); + +out_unlock: + up_write(&osdc->lock); } /* @@ -2680,17 +2631,11 @@ int ceph_osdc_start_request(struct ceph_osd_client *osdc, struct ceph_osd_request *req, bool nofail) { - int rc; - - down_read(&osdc->map_sem); - mutex_lock(&osdc->request_mutex); - - rc = __ceph_osdc_start_request(osdc, req, nofail); + down_read(&osdc->lock); + submit_request(req, false); + up_read(&osdc->lock); - mutex_unlock(&osdc->request_mutex); - up_read(&osdc->map_sem); - - return rc; + return 0; } EXPORT_SYMBOL(ceph_osdc_start_request); @@ -2703,13 +2648,12 @@ void ceph_osdc_cancel_request(struct ceph_osd_request *req) { struct ceph_osd_client *osdc = req->r_osdc; - mutex_lock(&osdc->request_mutex); + down_write(&osdc->lock); if (req->r_linger) __unregister_linger_request(osdc, req); - __unregister_request(osdc, req); - mutex_unlock(&osdc->request_mutex); - - dout("%s %p tid %llu canceled\n", __func__, req, req->r_tid); + if (req->r_osd) + cancel_request(req); + up_write(&osdc->lock); } EXPORT_SYMBOL(ceph_osdc_cancel_request); @@ -2744,32 +2688,40 @@ EXPORT_SYMBOL(ceph_osdc_wait_request); */ void ceph_osdc_sync(struct ceph_osd_client *osdc) { - struct ceph_osd_request *req; - u64 last_tid, next_tid = 0; + struct rb_node *n, *p; + u64 last_tid = atomic64_read(&osdc->last_tid); - mutex_lock(&osdc->request_mutex); - last_tid = osdc->last_tid; - while (1) { - req = __lookup_request_ge(osdc, next_tid); - if (!req) - break; - if (req->r_tid > last_tid) - break; +again: + down_read(&osdc->lock); + for (n = rb_first(&osdc->osds); n; n = rb_next(n)) { + struct ceph_osd *osd = rb_entry(n, struct ceph_osd, o_node); + + mutex_lock(&osd->lock); + for (p = rb_first(&osd->o_requests); p; p = rb_next(p)) { + struct ceph_osd_request *req = + rb_entry(p, struct ceph_osd_request, r_node); + + if (req->r_tid > last_tid) + break; + + if (!(req->r_flags & CEPH_OSD_FLAG_WRITE)) + continue; - next_tid = req->r_tid + 1; - if ((req->r_flags & CEPH_OSD_FLAG_WRITE) == 0) - continue; + ceph_osdc_get_request(req); + mutex_unlock(&osd->lock); + up_read(&osdc->lock); + dout("%s waiting on req %p tid %llu last_tid %llu\n", + __func__, req, req->r_tid, last_tid); + wait_for_completion(&req->r_safe_completion); + ceph_osdc_put_request(req); + goto again; + } - ceph_osdc_get_request(req); - mutex_unlock(&osdc->request_mutex); - dout("sync waiting on tid %llu (last is %llu)\n", - req->r_tid, last_tid); - wait_for_completion(&req->r_safe_completion); - mutex_lock(&osdc->request_mutex); - ceph_osdc_put_request(req); + mutex_unlock(&osd->lock); } - mutex_unlock(&osdc->request_mutex); - dout("sync done (thru tid %llu)\n", last_tid); + + up_read(&osdc->lock); + dout("%s done last_tid %llu\n", __func__, last_tid); } EXPORT_SYMBOL(ceph_osdc_sync); @@ -2793,18 +2745,14 @@ int ceph_osdc_init(struct ceph_osd_client *osdc, struct ceph_client *client) dout("init\n"); osdc->client = client; - init_rwsem(&osdc->map_sem); - mutex_init(&osdc->request_mutex); - osdc->last_tid = 0; + init_rwsem(&osdc->lock); osdc->osds = RB_ROOT; INIT_LIST_HEAD(&osdc->osd_lru); spin_lock_init(&osdc->osd_lru_lock); - osdc->requests = RB_ROOT; - INIT_LIST_HEAD(&osdc->req_lru); - INIT_LIST_HEAD(&osdc->req_unsent); - INIT_LIST_HEAD(&osdc->req_notarget); INIT_LIST_HEAD(&osdc->req_linger); - osdc->num_requests = 0; + osd_init(&osdc->homeless_osd); + osdc->homeless_osd.o_osdc = osdc; + osdc->homeless_osd.o_osd = CEPH_HOMELESS_OSD; INIT_DELAYED_WORK(&osdc->timeout_work, handle_timeout); INIT_DELAYED_WORK(&osdc->osds_timeout_work, handle_osds_timeout); spin_lock_init(&osdc->event_lock); @@ -2861,13 +2809,19 @@ void ceph_osdc_stop(struct ceph_osd_client *osdc) cancel_delayed_work_sync(&osdc->timeout_work); cancel_delayed_work_sync(&osdc->osds_timeout_work); - mutex_lock(&osdc->request_mutex); + down_write(&osdc->lock); while (!RB_EMPTY_ROOT(&osdc->osds)) { struct ceph_osd *osd = rb_entry(rb_first(&osdc->osds), struct ceph_osd, o_node); - remove_osd(osdc, osd); + close_osd(osd); } - mutex_unlock(&osdc->request_mutex); + up_write(&osdc->lock); + WARN_ON(atomic_read(&osdc->homeless_osd.o_ref) != 1); + osd_cleanup(&osdc->homeless_osd); + + WARN_ON(!list_empty(&osdc->osd_lru)); + WARN_ON(atomic_read(&osdc->num_requests)); + WARN_ON(atomic_read(&osdc->num_homeless)); ceph_osdmap_destroy(osdc->osdmap); mempool_destroy(osdc->req_mempool); @@ -2982,19 +2936,15 @@ EXPORT_SYMBOL(ceph_osdc_cleanup); static void dispatch(struct ceph_connection *con, struct ceph_msg *msg) { struct ceph_osd *osd = con->private; - struct ceph_osd_client *osdc; + struct ceph_osd_client *osdc = osd->o_osdc; int type = le16_to_cpu(msg->hdr.type); - if (!osd) - goto out; - osdc = osd->o_osdc; - switch (type) { case CEPH_MSG_OSD_MAP: ceph_osdc_handle_map(osdc, msg); break; case CEPH_MSG_OSD_OPREPLY: - handle_reply(osdc, msg); + handle_reply(osd, msg); break; case CEPH_MSG_WATCH_NOTIFY: handle_watch_notify(osdc, msg); @@ -3004,7 +2954,7 @@ static void dispatch(struct ceph_connection *con, struct ceph_msg *msg) pr_err("received unknown message type %d %s\n", type, ceph_msg_type_name(type)); } -out: + ceph_msg_put(msg); } @@ -3019,21 +2969,27 @@ static struct ceph_msg *get_reply(struct ceph_connection *con, { struct ceph_osd *osd = con->private; struct ceph_osd_client *osdc = osd->o_osdc; - struct ceph_msg *m; + struct ceph_msg *m = NULL; struct ceph_osd_request *req; int front_len = le32_to_cpu(hdr->front_len); int data_len = le32_to_cpu(hdr->data_len); - u64 tid; + u64 tid = le64_to_cpu(hdr->tid); - tid = le64_to_cpu(hdr->tid); - mutex_lock(&osdc->request_mutex); - req = lookup_request(&osdc->requests, tid); + down_read(&osdc->lock); + if (!osd_registered(osd)) { + dout("%s osd%d unknown, skipping\n", __func__, osd->o_osd); + *skip = 1; + goto out_unlock_osdc; + } + WARN_ON(osd->o_osd != le64_to_cpu(hdr->src.num)); + + mutex_lock(&osd->lock); + req = lookup_request(&osd->o_requests, tid); if (!req) { dout("%s osd%d tid %llu unknown, skipping\n", __func__, osd->o_osd, tid); - m = NULL; *skip = 1; - goto out; + goto out_unlock_session; } ceph_msg_revoke_incoming(req->r_reply); @@ -3045,7 +3001,7 @@ static struct ceph_msg *get_reply(struct ceph_connection *con, m = ceph_msg_new(CEPH_MSG_OSD_OPREPLY, front_len, GFP_NOFS, false); if (!m) - goto out; + goto out_unlock_session; ceph_msg_put(req->r_reply); req->r_reply = m; } @@ -3056,14 +3012,16 @@ static struct ceph_msg *get_reply(struct ceph_connection *con, req->r_reply->data_length); m = NULL; *skip = 1; - goto out; + goto out_unlock_session; } m = ceph_msg_get(req->r_reply); dout("get_reply tid %lld %p\n", tid, m); -out: - mutex_unlock(&osdc->request_mutex); +out_unlock_session: + mutex_unlock(&osd->lock); +out_unlock_osdc: + up_read(&osdc->lock); return m; } @@ -3083,8 +3041,8 @@ static struct ceph_msg *alloc_msg(struct ceph_connection *con, case CEPH_MSG_OSD_OPREPLY: return get_reply(con, hdr, skip); default: - pr_info("alloc_msg unexpected msg type %d from osd%d\n", type, - osd->o_osd); + pr_warn("%s osd%d unknown msg type %d, skipping\n", __func__, + osd->o_osd, type); *skip = 1; return NULL; } @@ -3188,5 +3146,5 @@ static const struct ceph_connection_operations osd_con_ops = { .alloc_msg = alloc_msg, .sign_message = osd_sign_message, .check_message_signature = osd_check_message_signature, - .fault = osd_reset, + .fault = osd_fault, }; -- cgit v1.3-8-gc7d7 From 235a09821c2bc71d9d07f12217ce2ac00db99eba Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Wed, 30 Mar 2016 17:18:34 +0800 Subject: ceph: multiple filesystem support To access non-default filesystem, we just need to subscribe to mdsmap. and add a new mount option for mds namespace id. Signed-off-by: Yan, Zheng [idryomov@gmail.com: switch to a new libceph API] Signed-off-by: Ilya Dryomov --- fs/ceph/super.c | 9 +++++++++ fs/ceph/super.h | 1 + 2 files changed, 10 insertions(+) (limited to 'fs') diff --git a/fs/ceph/super.c b/fs/ceph/super.c index f12d5e2955c2..452770dc9fc1 100644 --- a/fs/ceph/super.c +++ b/fs/ceph/super.c @@ -108,6 +108,7 @@ static int ceph_sync_fs(struct super_block *sb, int wait) * mount options */ enum { + Opt_mds_namespace, Opt_wsize, Opt_rsize, Opt_rasize, @@ -143,6 +144,7 @@ enum { }; static match_table_t fsopt_tokens = { + {Opt_mds_namespace, "mds_namespace=%d"}, {Opt_wsize, "wsize=%d"}, {Opt_rsize, "rsize=%d"}, {Opt_rasize, "rasize=%d"}, @@ -212,6 +214,9 @@ static int parse_fsopt_token(char *c, void *private) break; /* misc */ + case Opt_mds_namespace: + fsopt->mds_namespace = intval; + break; case Opt_wsize: fsopt->wsize = intval; break; @@ -367,6 +372,7 @@ static int parse_mount_options(struct ceph_mount_options **pfsopt, fsopt->max_readdir = CEPH_MAX_READDIR_DEFAULT; fsopt->max_readdir_bytes = CEPH_MAX_READDIR_BYTES_DEFAULT; fsopt->congestion_kb = default_congestion_kb(); + fsopt->mds_namespace = CEPH_FS_CLUSTER_ID_NONE; /* * Distinguish the server list from the path in "dev_name". @@ -457,6 +463,8 @@ static int ceph_show_options(struct seq_file *m, struct dentry *root) seq_puts(m, ",noacl"); #endif + if (fsopt->mds_namespace != CEPH_FS_CLUSTER_ID_NONE) + seq_printf(m, ",mds_namespace=%d", fsopt->mds_namespace); if (fsopt->wsize) seq_printf(m, ",wsize=%d", fsopt->wsize); if (fsopt->rsize != CEPH_RSIZE_DEFAULT) @@ -530,6 +538,7 @@ static struct ceph_fs_client *create_fs_client(struct ceph_mount_options *fsopt, goto fail; } fsc->client->extra_mon_dispatch = extra_mon_dispatch; + fsc->client->monc.fs_cluster_id = fsopt->mds_namespace; ceph_monc_want_map(&fsc->client->monc, CEPH_SUB_MDSMAP, 0, true); fsc->mount_options = fsopt; diff --git a/fs/ceph/super.h b/fs/ceph/super.h index e705c4d612d7..db2200f5ba80 100644 --- a/fs/ceph/super.h +++ b/fs/ceph/super.h @@ -62,6 +62,7 @@ struct ceph_mount_options { int cap_release_safety; int max_readdir; /* max readdir result (entires) */ int max_readdir_bytes; /* max readdir result (bytes) */ + int mds_namespace; /* * everything above this point can be memcmp'd; everything below -- cgit v1.3-8-gc7d7 From d463a43d69f4af85887671d76182437775fd1631 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Thu, 31 Mar 2016 15:53:01 +0800 Subject: ceph: CEPH_FEATURE_MDSENC support Signed-off-by: Yan, Zheng --- fs/ceph/mdsmap.c | 43 ++++++++++++++++++++++++++++++++++--------- fs/ceph/super.c | 5 ++--- 2 files changed, 36 insertions(+), 12 deletions(-) (limited to 'fs') diff --git a/fs/ceph/mdsmap.c b/fs/ceph/mdsmap.c index 261531e55e9d..8c3591a7fbae 100644 --- a/fs/ceph/mdsmap.c +++ b/fs/ceph/mdsmap.c @@ -54,16 +54,21 @@ struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end) const void *start = *p; int i, j, n; int err = -EINVAL; - u16 version; + u8 mdsmap_v, mdsmap_cv; m = kzalloc(sizeof(*m), GFP_NOFS); if (m == NULL) return ERR_PTR(-ENOMEM); - ceph_decode_16_safe(p, end, version, bad); - if (version > 3) { - pr_warn("got mdsmap version %d > 3, failing", version); - goto bad; + ceph_decode_need(p, end, 1 + 1, bad); + mdsmap_v = ceph_decode_8(p); + mdsmap_cv = ceph_decode_8(p); + if (mdsmap_v >= 4) { + u32 mdsmap_len; + ceph_decode_32_safe(p, end, mdsmap_len, bad); + if (end < *p + mdsmap_len) + goto bad; + end = *p + mdsmap_len; } ceph_decode_need(p, end, 8*sizeof(u32) + sizeof(u64), bad); @@ -87,16 +92,29 @@ struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end) u32 namelen; s32 mds, inc, state; u64 state_seq; - u8 infoversion; + u8 info_v; + void *info_end = NULL; struct ceph_entity_addr addr; u32 num_export_targets; void *pexport_targets = NULL; struct ceph_timespec laggy_since; struct ceph_mds_info *info; - ceph_decode_need(p, end, sizeof(u64)*2 + 1 + sizeof(u32), bad); + ceph_decode_need(p, end, sizeof(u64) + 1, bad); global_id = ceph_decode_64(p); - infoversion = ceph_decode_8(p); + info_v= ceph_decode_8(p); + if (info_v >= 4) { + u32 info_len; + u8 info_cv; + ceph_decode_need(p, end, 1 + sizeof(u32), bad); + info_cv = ceph_decode_8(p); + info_len = ceph_decode_32(p); + info_end = *p + info_len; + if (info_end > end) + goto bad; + } + + ceph_decode_need(p, end, sizeof(u64) + sizeof(u32), bad); *p += sizeof(u64); namelen = ceph_decode_32(p); /* skip mds name */ *p += namelen; @@ -115,7 +133,7 @@ struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end) *p += sizeof(u32); ceph_decode_32_safe(p, end, namelen, bad); *p += namelen; - if (infoversion >= 2) { + if (info_v >= 2) { ceph_decode_32_safe(p, end, num_export_targets, bad); pexport_targets = *p; *p += num_export_targets * sizeof(u32); @@ -123,6 +141,12 @@ struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end) num_export_targets = 0; } + if (info_end && *p != info_end) { + if (*p > info_end) + goto bad; + *p = info_end; + } + dout("mdsmap_decode %d/%d %lld mds%d.%d %s %s\n", i+1, n, global_id, mds, inc, ceph_pr_addr(&addr.in_addr), @@ -163,6 +187,7 @@ struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end) m->m_cas_pg_pool = ceph_decode_64(p); /* ok, we don't care about the rest. */ + *p = end; dout("mdsmap_decode success epoch %u\n", m->m_epoch); return m; diff --git a/fs/ceph/super.c b/fs/ceph/super.c index 452770dc9fc1..d714ab20ad24 100644 --- a/fs/ceph/super.c +++ b/fs/ceph/super.c @@ -519,9 +519,8 @@ static struct ceph_fs_client *create_fs_client(struct ceph_mount_options *fsopt, { struct ceph_fs_client *fsc; const u64 supported_features = - CEPH_FEATURE_FLOCK | - CEPH_FEATURE_DIRLAYOUTHASH | - CEPH_FEATURE_MDS_INLINE_DATA; + CEPH_FEATURE_FLOCK | CEPH_FEATURE_DIRLAYOUTHASH | + CEPH_FEATURE_MDSENC | CEPH_FEATURE_MDS_INLINE_DATA; const u64 required_features = 0; int page_count; size_t size; -- cgit v1.3-8-gc7d7 From 77310320c299b0dc050037ff8fc29fd1861fb005 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Fri, 8 Apr 2016 15:27:16 +0800 Subject: ceph: renew caps for read/write if mds session got killed. When mds session gets killed, read/write operation may hang. Client waits for Frw caps, but mds does not know what caps client wants. To recover this, client sends an open request to mds. The request will tell mds what caps client wants. Signed-off-by: Yan, Zheng --- fs/ceph/caps.c | 43 ++++++++++++++++++++++++++++++++---------- fs/ceph/file.c | 53 ++++++++++++++++++++++++++++++++++++++++++++++++++++ fs/ceph/mds_client.c | 6 +++++- fs/ceph/super.h | 2 ++ 4 files changed, 93 insertions(+), 11 deletions(-) (limited to 'fs') diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index cfaeef18cbca..fab93c66d879 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -2317,7 +2317,7 @@ again: /* make sure file is actually open */ file_wanted = __ceph_caps_file_wanted(ci); - if ((file_wanted & need) == 0) { + if ((file_wanted & need) != need) { dout("try_get_cap_refs need %s file_wanted %s, EBADF\n", ceph_cap_string(need), ceph_cap_string(file_wanted)); *err = -EBADF; @@ -2412,12 +2412,26 @@ again: goto out_unlock; } - if (!__ceph_is_any_caps(ci) && - ACCESS_ONCE(mdsc->fsc->mount_state) == CEPH_MOUNT_SHUTDOWN) { - dout("get_cap_refs %p forced umount\n", inode); - *err = -EIO; - ret = 1; - goto out_unlock; + if (ci->i_ceph_flags & CEPH_I_CAP_DROPPED) { + int mds_wanted; + if (ACCESS_ONCE(mdsc->fsc->mount_state) == + CEPH_MOUNT_SHUTDOWN) { + dout("get_cap_refs %p forced umount\n", inode); + *err = -EIO; + ret = 1; + goto out_unlock; + } + mds_wanted = __ceph_caps_mds_wanted(ci); + if ((mds_wanted & need) != need) { + dout("get_cap_refs %p caps were dropped" + " (session killed?)\n", inode); + *err = -ESTALE; + ret = 1; + goto out_unlock; + } + if ((mds_wanted & file_wanted) == + (file_wanted & (CEPH_CAP_FILE_RD|CEPH_CAP_FILE_WR))) + ci->i_ceph_flags &= ~CEPH_I_CAP_DROPPED; } dout("get_cap_refs %p have %s needed %s\n", inode, @@ -2487,7 +2501,7 @@ int ceph_get_caps(struct ceph_inode_info *ci, int need, int want, if (err == -EAGAIN) continue; if (err < 0) - return err; + ret = err; } else { ret = wait_event_interruptible(ci->i_cap_wq, try_get_cap_refs(ci, need, want, endoff, @@ -2496,8 +2510,15 @@ int ceph_get_caps(struct ceph_inode_info *ci, int need, int want, continue; if (err < 0) ret = err; - if (ret < 0) - return ret; + } + if (ret < 0) { + if (err == -ESTALE) { + /* session was killed, try renew caps */ + ret = ceph_renew_caps(&ci->vfs_inode); + if (ret == 0) + continue; + } + return ret; } if (ci->i_inline_version != CEPH_INLINE_NONE && @@ -3226,6 +3247,8 @@ retry: if (target < 0) { __ceph_remove_cap(cap, false); + if (!ci->i_auth_cap) + ci->i_ceph_flags |= CEPH_I_CAP_DROPPED; goto out_unlock; } diff --git a/fs/ceph/file.c b/fs/ceph/file.c index 30fd49eb25b4..996e9ec81e4e 100644 --- a/fs/ceph/file.c +++ b/fs/ceph/file.c @@ -191,6 +191,59 @@ static int ceph_init_file(struct inode *inode, struct file *file, int fmode) return ret; } +/* + * try renew caps after session gets killed. + */ +int ceph_renew_caps(struct inode *inode) +{ + struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; + struct ceph_inode_info *ci = ceph_inode(inode); + struct ceph_mds_request *req; + int err, flags, wanted; + + spin_lock(&ci->i_ceph_lock); + wanted = __ceph_caps_file_wanted(ci); + if (__ceph_is_any_real_caps(ci) && + (!(wanted & CEPH_CAP_ANY_WR) == 0 || ci->i_auth_cap)) { + int issued = __ceph_caps_issued(ci, NULL); + spin_unlock(&ci->i_ceph_lock); + dout("renew caps %p want %s issued %s updating mds_wanted\n", + inode, ceph_cap_string(wanted), ceph_cap_string(issued)); + ceph_check_caps(ci, 0, NULL); + return 0; + } + spin_unlock(&ci->i_ceph_lock); + + flags = 0; + if ((wanted & CEPH_CAP_FILE_RD) && (wanted & CEPH_CAP_FILE_WR)) + flags = O_RDWR; + else if (wanted & CEPH_CAP_FILE_RD) + flags = O_RDONLY; + else if (wanted & CEPH_CAP_FILE_WR) + flags = O_WRONLY; +#ifdef O_LAZY + if (wanted & CEPH_CAP_FILE_LAZYIO) + flags |= O_LAZY; +#endif + + req = prepare_open_request(inode->i_sb, flags, 0); + if (IS_ERR(req)) { + err = PTR_ERR(req); + goto out; + } + + req->r_inode = inode; + ihold(inode); + req->r_num_caps = 1; + req->r_fmode = -1; + + err = ceph_mdsc_do_request(mdsc, NULL, req); + ceph_mdsc_put_request(req); +out: + dout("renew caps %p open result=%d\n", inode, err); + return err < 0 ? err : 0; +} + /* * If we already have the requisite capabilities, we can satisfy * the open request locally (no need to request new caps from the diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c index cff85af425d4..1e5965d17cb1 100644 --- a/fs/ceph/mds_client.c +++ b/fs/ceph/mds_client.c @@ -1133,6 +1133,8 @@ static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap, struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; + ci->i_ceph_flags |= CEPH_I_CAP_DROPPED; + while (true) { struct rb_node *n = rb_first(&ci->i_cap_flush_tree); if (!n) @@ -1181,7 +1183,9 @@ static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap, list_del(&cf->list); ceph_free_cap_flush(cf); } - while (drop--) + + wake_up_all(&ci->i_cap_wq); + if (drop) iput(inode); return 0; } diff --git a/fs/ceph/super.h b/fs/ceph/super.h index db2200f5ba80..5fef3a6397db 100644 --- a/fs/ceph/super.h +++ b/fs/ceph/super.h @@ -470,6 +470,7 @@ static inline struct inode *ceph_find_inode(struct super_block *sb, #define CEPH_I_POOL_RD (1 << 5) /* can read from pool */ #define CEPH_I_POOL_WR (1 << 6) /* can write to pool */ #define CEPH_I_SEC_INITED (1 << 7) /* security initialized */ +#define CEPH_I_CAP_DROPPED (1 << 8) /* caps were forcibly dropped */ static inline void __ceph_dir_set_complete(struct ceph_inode_info *ci, long long release_count, @@ -932,6 +933,7 @@ extern void ceph_pool_perm_destroy(struct ceph_mds_client* mdsc); /* file.c */ extern const struct file_operations ceph_file_fops; +extern int ceph_renew_caps(struct inode *inode); extern int ceph_open(struct inode *inode, struct file *file); extern int ceph_atomic_open(struct inode *dir, struct dentry *dentry, struct file *file, unsigned flags, umode_t mode, -- cgit v1.3-8-gc7d7 From 6c93df5db628e710697c43bc1bd78a786549a548 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Fri, 15 Apr 2016 13:56:12 +0800 Subject: ceph: don't call truncate_pagecache in ceph_writepages_start truncate_pagecache() may decrease inode's reference. This can cause deadlock if inode's last reference is dropped and iput_final() wants to evict the inode. (evict() calls inode_wait_for_writeback(), which waits for ceph_writepages_start() to return). The fix is use work thead to truncate dirty pages. Also add 'forced umount' check to ceph_update_writeable_page(), which prevents new pages getting dirty. Signed-off-by: Yan, Zheng --- fs/ceph/addr.c | 14 ++++++++++++-- fs/ceph/inode.c | 11 +++++++++++ fs/ceph/mds_client.c | 22 +++++++++++++++------- 3 files changed, 38 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c index f47418477629..d52e3bcfda7c 100644 --- a/fs/ceph/addr.c +++ b/fs/ceph/addr.c @@ -715,8 +715,11 @@ static int ceph_writepages_start(struct address_space *mapping, (wbc->sync_mode == WB_SYNC_ALL ? "ALL" : "HOLD")); if (ACCESS_ONCE(fsc->mount_state) == CEPH_MOUNT_SHUTDOWN) { - pr_warn("writepage_start %p on forced umount\n", inode); - truncate_pagecache(inode, 0); + if (ci->i_wrbuffer_ref > 0) { + pr_warn_ratelimited( + "writepage_start %p %lld forced umount\n", + inode, ceph_ino(inode)); + } mapping_set_error(mapping, -EIO); return -EIO; /* we're in a forced umount, don't write! */ } @@ -1127,6 +1130,7 @@ static int ceph_update_writeable_page(struct file *file, struct page *page) { struct inode *inode = file_inode(file); + struct ceph_fs_client *fsc = ceph_inode_to_client(inode); struct ceph_inode_info *ci = ceph_inode(inode); loff_t page_off = pos & PAGE_MASK; int pos_in_page = pos & ~PAGE_MASK; @@ -1135,6 +1139,12 @@ static int ceph_update_writeable_page(struct file *file, int r; struct ceph_snap_context *snapc, *oldest; + if (ACCESS_ONCE(fsc->mount_state) == CEPH_MOUNT_SHUTDOWN) { + dout(" page %p forced umount\n", page); + unlock_page(page); + return -EIO; + } + retry_locked: /* writepages currently holds page lock, but if we change that later, */ wait_on_page_writeback(page); diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c index edfade037738..b906e02cddad 100644 --- a/fs/ceph/inode.c +++ b/fs/ceph/inode.c @@ -1623,10 +1623,21 @@ static void ceph_invalidate_work(struct work_struct *work) struct ceph_inode_info *ci = container_of(work, struct ceph_inode_info, i_pg_inv_work); struct inode *inode = &ci->vfs_inode; + struct ceph_fs_client *fsc = ceph_inode_to_client(inode); u32 orig_gen; int check = 0; mutex_lock(&ci->i_truncate_mutex); + + if (ACCESS_ONCE(fsc->mount_state) == CEPH_MOUNT_SHUTDOWN) { + pr_warn_ratelimited("invalidate_pages %p %lld forced umount\n", + inode, ceph_ino(inode)); + mapping_set_error(inode->i_mapping, -EIO); + truncate_pagecache(inode, 0); + mutex_unlock(&ci->i_truncate_mutex); + goto out; + } + spin_lock(&ci->i_ceph_lock); dout("invalidate_pages %p gen %d revoking %d\n", inode, ci->i_rdcache_gen, ci->i_rdcache_revoking); diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c index 1e5965d17cb1..cbe6c0afdadc 100644 --- a/fs/ceph/mds_client.c +++ b/fs/ceph/mds_client.c @@ -1120,9 +1120,11 @@ out: static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap, void *arg) { + struct ceph_fs_client *fsc = (struct ceph_fs_client *)arg; struct ceph_inode_info *ci = ceph_inode(inode); LIST_HEAD(to_remove); - int drop = 0; + bool drop = false; + bool invalidate = false; dout("removing cap %p, ci is %p, inode is %p\n", cap, ci, &ci->vfs_inode); @@ -1130,11 +1132,14 @@ static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap, __ceph_remove_cap(cap, false); if (!ci->i_auth_cap) { struct ceph_cap_flush *cf; - struct ceph_mds_client *mdsc = - ceph_sb_to_client(inode->i_sb)->mdsc; + struct ceph_mds_client *mdsc = fsc->mdsc; ci->i_ceph_flags |= CEPH_I_CAP_DROPPED; + if (ci->i_wrbuffer_ref > 0 && + ACCESS_ONCE(fsc->mount_state) == CEPH_MOUNT_SHUTDOWN) + invalidate = true; + while (true) { struct rb_node *n = rb_first(&ci->i_cap_flush_tree); if (!n) @@ -1156,7 +1161,7 @@ static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap, inode, ceph_ino(inode)); ci->i_dirty_caps = 0; list_del_init(&ci->i_dirty_item); - drop = 1; + drop = true; } if (!list_empty(&ci->i_flushing_item)) { pr_warn_ratelimited( @@ -1166,7 +1171,7 @@ static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap, ci->i_flushing_caps = 0; list_del_init(&ci->i_flushing_item); mdsc->num_cap_flushing--; - drop = 1; + drop = true; } spin_unlock(&mdsc->cap_dirty_lock); @@ -1185,6 +1190,8 @@ static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap, } wake_up_all(&ci->i_cap_wq); + if (invalidate) + ceph_queue_invalidate(inode); if (drop) iput(inode); return 0; @@ -1195,12 +1202,13 @@ static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap, */ static void remove_session_caps(struct ceph_mds_session *session) { + struct ceph_fs_client *fsc = session->s_mdsc->fsc; + struct super_block *sb = fsc->sb; dout("remove_session_caps on %p\n", session); - iterate_session_caps(session, remove_session_caps_cb, NULL); + iterate_session_caps(session, remove_session_caps_cb, fsc); spin_lock(&session->s_cap_lock); if (session->s_nr_caps > 0) { - struct super_block *sb = session->s_mdsc->fsc->sb; struct inode *inode; struct ceph_cap *cap, *prev = NULL; struct ceph_vino vino; -- cgit v1.3-8-gc7d7 From aeda081c5eb3d7668b64e2dc9bfc0bf30ef53632 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Mon, 18 Apr 2016 16:51:37 +0800 Subject: ceph: don't show symlink target in debugfs/mdsc symlink target is useless for debug and can be very long. It's annoying to show it in debugfs/mdsc. Signed-off-by: Yan, Zheng --- fs/ceph/debugfs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ceph/debugfs.c b/fs/ceph/debugfs.c index 31f831471ed2..39ff678e567f 100644 --- a/fs/ceph/debugfs.c +++ b/fs/ceph/debugfs.c @@ -109,7 +109,7 @@ static int mdsc_show(struct seq_file *s, void *p) path ? path : ""); spin_unlock(&req->r_old_dentry->d_lock); kfree(path); - } else if (req->r_path2) { + } else if (req->r_path2 && req->r_op != CEPH_MDS_OP_SYMLINK) { if (req->r_ino2.ino) seq_printf(s, " #%llx/%s", req->r_ino2.ino, req->r_path2); -- cgit v1.3-8-gc7d7 From 3f38495409b613071021fca86629df7ae81820ad Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Thu, 21 Apr 2016 11:09:55 +0800 Subject: ceph: report mount root in session metadata Signed-off-by: Yan, Zheng --- fs/ceph/mds_client.c | 4 +++- fs/ceph/super.c | 33 +++++++++++++++++++-------------- fs/ceph/super.h | 1 + 3 files changed, 23 insertions(+), 15 deletions(-) (limited to 'fs') diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c index cbe6c0afdadc..047f723bdbe6 100644 --- a/fs/ceph/mds_client.c +++ b/fs/ceph/mds_client.c @@ -839,12 +839,14 @@ static struct ceph_msg *create_session_open_msg(struct ceph_mds_client *mdsc, u6 int metadata_bytes = 0; int metadata_key_count = 0; struct ceph_options *opt = mdsc->fsc->client->options; + struct ceph_mount_options *fsopt = mdsc->fsc->mount_options; void *p; const char* metadata[][2] = { {"hostname", utsname()->nodename}, {"kernel_version", utsname()->release}, - {"entity_id", opt->name ? opt->name : ""}, + {"entity_id", opt->name ? : ""}, + {"root", fsopt->server_path ? : "/"}, {NULL, NULL} }; diff --git a/fs/ceph/super.c b/fs/ceph/super.c index d714ab20ad24..91e02481ce06 100644 --- a/fs/ceph/super.c +++ b/fs/ceph/super.c @@ -302,6 +302,7 @@ static void destroy_mount_options(struct ceph_mount_options *args) { dout("destroy_mount_options %p\n", args); kfree(args->snapdir_name); + kfree(args->server_path); kfree(args); } @@ -333,14 +334,17 @@ static int compare_mount_options(struct ceph_mount_options *new_fsopt, if (ret) return ret; + ret = strcmp_null(fsopt1->server_path, fsopt2->server_path); + if (ret) + return ret; + return ceph_compare_options(new_opt, fsc->client); } static int parse_mount_options(struct ceph_mount_options **pfsopt, struct ceph_options **popt, int flags, char *options, - const char *dev_name, - const char **path) + const char *dev_name) { struct ceph_mount_options *fsopt; const char *dev_name_end; @@ -386,12 +390,13 @@ static int parse_mount_options(struct ceph_mount_options **pfsopt, */ dev_name_end = strchr(dev_name, '/'); if (dev_name_end) { - /* skip over leading '/' for path */ - *path = dev_name_end + 1; + fsopt->server_path = kstrdup(dev_name_end, GFP_KERNEL); + if (!fsopt->server_path) { + err = -ENOMEM; + goto out; + } } else { - /* path is empty */ dev_name_end = dev_name + strlen(dev_name); - *path = dev_name_end; } err = -EINVAL; dev_name_end--; /* back up to ':' separator */ @@ -401,7 +406,8 @@ static int parse_mount_options(struct ceph_mount_options **pfsopt, goto out; } dout("device name '%.*s'\n", (int)(dev_name_end - dev_name), dev_name); - dout("server path '%s'\n", *path); + if (fsopt->server_path) + dout("server path '%s'\n", fsopt->server_path); *popt = ceph_parse_options(options, dev_name, dev_name_end, parse_fsopt_token, (void *)fsopt); @@ -793,8 +799,7 @@ out: /* * mount: join the ceph cluster, and open root directory. */ -static struct dentry *ceph_real_mount(struct ceph_fs_client *fsc, - const char *path) +static struct dentry *ceph_real_mount(struct ceph_fs_client *fsc) { int err; unsigned long started = jiffies; /* note the start time */ @@ -823,11 +828,12 @@ static struct dentry *ceph_real_mount(struct ceph_fs_client *fsc, goto fail; } - if (path[0] == 0) { + if (!fsc->mount_options->server_path) { root = fsc->sb->s_root; dget(root); } else { - dout("mount opening base mountpoint\n"); + const char *path = fsc->mount_options->server_path + 1; + dout("mount opening path %s\n", path); root = open_root_dentry(fsc, path, started); if (IS_ERR(root)) { err = PTR_ERR(root); @@ -943,7 +949,6 @@ static struct dentry *ceph_mount(struct file_system_type *fs_type, struct dentry *res; int err; int (*compare_super)(struct super_block *, void *) = ceph_compare_super; - const char *path = NULL; struct ceph_mount_options *fsopt = NULL; struct ceph_options *opt = NULL; @@ -952,7 +957,7 @@ static struct dentry *ceph_mount(struct file_system_type *fs_type, #ifdef CONFIG_CEPH_FS_POSIX_ACL flags |= MS_POSIXACL; #endif - err = parse_mount_options(&fsopt, &opt, flags, data, dev_name, &path); + err = parse_mount_options(&fsopt, &opt, flags, data, dev_name); if (err < 0) { res = ERR_PTR(err); goto out_final; @@ -995,7 +1000,7 @@ static struct dentry *ceph_mount(struct file_system_type *fs_type, } } - res = ceph_real_mount(fsc, path); + res = ceph_real_mount(fsc); if (IS_ERR(res)) goto out_splat; dout("root %p inode %p ino %llx.%llx\n", res, diff --git a/fs/ceph/super.h b/fs/ceph/super.h index 5fef3a6397db..0ea86406f463 100644 --- a/fs/ceph/super.h +++ b/fs/ceph/super.h @@ -70,6 +70,7 @@ struct ceph_mount_options { */ char *snapdir_name; /* default ".snap" */ + char *server_path; /* default "/" */ }; struct ceph_fs_client { -- cgit v1.3-8-gc7d7 From 04303d8ad06985bed4ea4f46018fc8f55a0962a8 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Thu, 21 Apr 2016 12:11:54 +0800 Subject: ceph: use CEPH_MDS_OP_RMXATTR request to remove xattr Setxattr with NULL value and XATTR_REPLACE flag should be equivalent to removexattr. But current MDS does not support deleting vxattrs through MDS_OP_SETXATTR request. The workaround is sending MDS_OP_RMXATTR request if setxattr actually removs xattr. Signed-off-by: Yan, Zheng --- fs/ceph/xattr.c | 17 +++++++++++------ 1 file changed, 11 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/ceph/xattr.c b/fs/ceph/xattr.c index 5afabc4bf4c7..426c83cc683a 100644 --- a/fs/ceph/xattr.c +++ b/fs/ceph/xattr.c @@ -886,6 +886,7 @@ static int ceph_sync_setxattr(struct dentry *dentry, const char *name, struct ceph_mds_request *req; struct ceph_mds_client *mdsc = fsc->mdsc; struct ceph_pagelist *pagelist = NULL; + int op = CEPH_MDS_OP_SETXATTR; int err; if (size > 0) { @@ -899,20 +900,21 @@ static int ceph_sync_setxattr(struct dentry *dentry, const char *name, if (err) goto out; } else if (!value) { - flags |= CEPH_XATTR_REMOVE; + if (flags & CEPH_XATTR_REPLACE) + op = CEPH_MDS_OP_RMXATTR; + else + flags |= CEPH_XATTR_REMOVE; } dout("setxattr value=%.*s\n", (int)size, value); /* do request */ - req = ceph_mdsc_create_request(mdsc, CEPH_MDS_OP_SETXATTR, - USE_AUTH_MDS); + req = ceph_mdsc_create_request(mdsc, op, USE_AUTH_MDS); if (IS_ERR(req)) { err = PTR_ERR(req); goto out; } - req->r_args.setxattr.flags = cpu_to_le32(flags); req->r_path2 = kstrdup(name, GFP_NOFS); if (!req->r_path2) { ceph_mdsc_put_request(req); @@ -920,8 +922,11 @@ static int ceph_sync_setxattr(struct dentry *dentry, const char *name, goto out; } - req->r_pagelist = pagelist; - pagelist = NULL; + if (op == CEPH_MDS_OP_SETXATTR) { + req->r_args.setxattr.flags = cpu_to_le32(flags); + req->r_pagelist = pagelist; + pagelist = NULL; + } req->r_inode = inode; ihold(inode); -- cgit v1.3-8-gc7d7 From c530cd24c23818af8da35fe556f9a7754211e50a Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Thu, 28 Apr 2016 17:43:35 +0800 Subject: ceph: search cache postion for dcache readdir use binary search to find cache index that corresponds to readdir postion. Signed-off-by: Yan, Zheng --- fs/ceph/dir.c | 129 +++++++++++++++++++++++++++++++++++++--------------------- 1 file changed, 83 insertions(+), 46 deletions(-) (limited to 'fs') diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c index 4fb2bbc2a272..cdea450c5902 100644 --- a/fs/ceph/dir.c +++ b/fs/ceph/dir.c @@ -110,6 +110,50 @@ static int note_last_dentry(struct ceph_file_info *fi, const char *name, return 0; } + +static struct dentry * +__dcache_find_get_entry(struct dentry *parent, u64 idx, + struct ceph_readdir_cache_control *cache_ctl) +{ + struct inode *dir = d_inode(parent); + struct dentry *dentry; + unsigned idx_mask = (PAGE_SIZE / sizeof(struct dentry *)) - 1; + loff_t ptr_pos = idx * sizeof(struct dentry *); + pgoff_t ptr_pgoff = ptr_pos >> PAGE_SHIFT; + + if (ptr_pos >= i_size_read(dir)) + return NULL; + + if (!cache_ctl->page || ptr_pgoff != page_index(cache_ctl->page)) { + ceph_readdir_cache_release(cache_ctl); + cache_ctl->page = find_lock_page(&dir->i_data, ptr_pgoff); + if (!cache_ctl->page) { + dout(" page %lu not found\n", ptr_pgoff); + return ERR_PTR(-EAGAIN); + } + /* reading/filling the cache are serialized by + i_mutex, no need to use page lock */ + unlock_page(cache_ctl->page); + cache_ctl->dentries = kmap(cache_ctl->page); + } + + cache_ctl->index = idx & idx_mask; + + rcu_read_lock(); + spin_lock(&parent->d_lock); + /* check i_size again here, because empty directory can be + * marked as complete while not holding the i_mutex. */ + if (ceph_dir_is_complete_ordered(dir) && ptr_pos < i_size_read(dir)) + dentry = cache_ctl->dentries[cache_ctl->index]; + else + dentry = NULL; + spin_unlock(&parent->d_lock); + if (dentry && !lockref_get_not_dead(&dentry->d_lockref)) + dentry = NULL; + rcu_read_unlock(); + return dentry ? : ERR_PTR(-EAGAIN); +} + /* * When possible, we try to satisfy a readdir by peeking at the * dcache. We make this work by carefully ordering dentries on @@ -129,62 +173,57 @@ static int __dcache_readdir(struct file *file, struct dir_context *ctx, struct inode *dir = d_inode(parent); struct dentry *dentry, *last = NULL; struct ceph_dentry_info *di; - unsigned nsize = PAGE_SIZE / sizeof(struct dentry *); - int err = 0; - loff_t ptr_pos = 0; struct ceph_readdir_cache_control cache_ctl = {}; + u64 idx = 0; + int err = 0; dout("__dcache_readdir %p v%u at %llu\n", dir, shared_gen, ctx->pos); - /* we can calculate cache index for the first dirfrag */ - if (ceph_frag_is_leftmost(fpos_frag(ctx->pos))) { - cache_ctl.index = fpos_off(ctx->pos) - 2; - BUG_ON(cache_ctl.index < 0); - ptr_pos = cache_ctl.index * sizeof(struct dentry *); + /* search start position */ + if (ctx->pos > 2) { + u64 count = div_u64(i_size_read(dir), sizeof(struct dentry *)); + while (count > 0) { + u64 step = count >> 1; + dentry = __dcache_find_get_entry(parent, idx + step, + &cache_ctl); + if (!dentry) { + /* use linar search */ + idx = 0; + break; + } + if (IS_ERR(dentry)) { + err = PTR_ERR(dentry); + goto out; + } + di = ceph_dentry(dentry); + spin_lock(&dentry->d_lock); + if (fpos_cmp(di->offset, ctx->pos) < 0) { + idx += step + 1; + count -= step + 1; + } else { + count = step; + } + spin_unlock(&dentry->d_lock); + dput(dentry); + } + + dout("__dcache_readdir %p cache idx %llu\n", dir, idx); } - while (true) { - pgoff_t pgoff; - bool emit_dentry; - if (ptr_pos >= i_size_read(dir)) { + for (;;) { + bool emit_dentry = false; + dentry = __dcache_find_get_entry(parent, idx++, &cache_ctl); + if (!dentry) { fi->flags |= CEPH_F_ATEND; err = 0; break; } - - err = -EAGAIN; - pgoff = ptr_pos >> PAGE_SHIFT; - if (!cache_ctl.page || pgoff != page_index(cache_ctl.page)) { - ceph_readdir_cache_release(&cache_ctl); - cache_ctl.page = find_lock_page(&dir->i_data, pgoff); - if (!cache_ctl.page) { - dout(" page %lu not found\n", pgoff); - break; - } - /* reading/filling the cache are serialized by - * i_mutex, no need to use page lock */ - unlock_page(cache_ctl.page); - cache_ctl.dentries = kmap(cache_ctl.page); + if (IS_ERR(dentry)) { + err = PTR_ERR(dentry); + goto out; } - rcu_read_lock(); - spin_lock(&parent->d_lock); - /* check i_size again here, because empty directory can be - * marked as complete while not holding the i_mutex. */ - if (ceph_dir_is_complete_ordered(dir) && - ptr_pos < i_size_read(dir)) - dentry = cache_ctl.dentries[cache_ctl.index % nsize]; - else - dentry = NULL; - spin_unlock(&parent->d_lock); - if (dentry && !lockref_get_not_dead(&dentry->d_lockref)) - dentry = NULL; - rcu_read_unlock(); - if (!dentry) - break; - - emit_dentry = false; di = ceph_dentry(dentry); spin_lock(&dentry->d_lock); if (di->lease_shared_gen == shared_gen && @@ -217,10 +256,8 @@ static int __dcache_readdir(struct file *file, struct dir_context *ctx, } else { dput(dentry); } - - cache_ctl.index++; - ptr_pos += sizeof(struct dentry *); } +out: ceph_readdir_cache_release(&cache_ctl); if (last) { int ret; -- cgit v1.3-8-gc7d7 From 1cd42a429174689c0df1c37b642654a1ab4d1506 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Fri, 29 Apr 2016 15:58:32 +0800 Subject: ceph: remove unnecessary checks in __dcache_readdir we never add snapdir and the hidden .ceph dir into readdir cache Signed-off-by: Yan, Zheng --- fs/ceph/dir.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'fs') diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c index cdea450c5902..734508be1526 100644 --- a/fs/ceph/dir.c +++ b/fs/ceph/dir.c @@ -228,8 +228,6 @@ static int __dcache_readdir(struct file *file, struct dir_context *ctx, spin_lock(&dentry->d_lock); if (di->lease_shared_gen == shared_gen && d_really_is_positive(dentry) && - ceph_snap(d_inode(dentry)) != CEPH_SNAPDIR && - ceph_ino(d_inode(dentry)) != CEPH_INO_CEPH && fpos_cmp(ctx->pos, di->offset) <= 0) { emit_dentry = true; } -- cgit v1.3-8-gc7d7 From a78600e7c4fb47fb5ef34265456b731fde27a9c3 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Wed, 27 Apr 2016 17:32:34 +0800 Subject: ceph: simplify 'offset in frag' don't distinguish leftmost frag from other frags. always use 2 as first entry's offset. Signed-off-by: Yan, Zheng --- fs/ceph/dir.c | 12 +++--------- fs/ceph/inode.c | 5 +---- 2 files changed, 4 insertions(+), 13 deletions(-) (limited to 'fs') diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c index 734508be1526..31a2cdf39210 100644 --- a/fs/ceph/dir.c +++ b/fs/ceph/dir.c @@ -414,10 +414,7 @@ more: if (req->r_reply_info.dir_end) { kfree(fi->last_name); fi->last_name = NULL; - if (ceph_frag_is_rightmost(frag)) - fi->next_offset = 2; - else - fi->next_offset = 0; + fi->next_offset = 2; } else { err = note_last_dentry(fi, rinfo->dir_dname[rinfo->dir_nr-1], @@ -468,7 +465,7 @@ more: /* more frags? */ if (!ceph_frag_is_rightmost(frag)) { frag = ceph_frag_next(frag); - off = 0; + off = 2; ctx->pos = ceph_make_fpos(frag, off); dout("readdir next frag is %x\n", frag); goto more; @@ -511,10 +508,7 @@ static void reset_readdir(struct ceph_file_info *fi, unsigned frag) fi->last_name = NULL; fi->dir_release_count = 0; fi->readdir_cache_idx = -1; - if (ceph_frag_is_leftmost(frag)) - fi->next_offset = 2; /* compensate for . and .. */ - else - fi->next_offset = 0; + fi->next_offset = 2; /* compensate for . and .. */ fi->flags &= ~CEPH_F_ATEND; } diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c index b906e02cddad..51ce5ce59f04 100644 --- a/fs/ceph/inode.c +++ b/fs/ceph/inode.c @@ -1405,10 +1405,7 @@ int ceph_readdir_prepopulate(struct ceph_mds_request *req, dout("readdir_prepopulate got new frag %x -> %x\n", frag, le32_to_cpu(rinfo->dir_dir->frag)); frag = le32_to_cpu(rinfo->dir_dir->frag); - if (ceph_frag_is_leftmost(frag)) - req->r_readdir_offset = 2; - else - req->r_readdir_offset = 0; + req->r_readdir_offset = 2; } if (le32_to_cpu(rinfo->head->op) == CEPH_MDS_OP_LSSNAP) { -- cgit v1.3-8-gc7d7 From 2a5beea3f1b6544d6c72ea220e860a2eda2f9104 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Thu, 28 Apr 2016 09:37:39 +0800 Subject: ceph: define struct for dir entry in readdir reply This avoids defining multiple arrays for entries in readdir reply Signed-off-by: Yan, Zheng --- fs/ceph/dir.c | 27 ++++++++++++--------------- fs/ceph/inode.c | 21 +++++++++++---------- fs/ceph/mds_client.c | 42 +++++++++++++++++++----------------------- fs/ceph/mds_client.h | 12 ++++++++---- 4 files changed, 50 insertions(+), 52 deletions(-) (limited to 'fs') diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c index 31a2cdf39210..68530acea2c8 100644 --- a/fs/ceph/dir.c +++ b/fs/ceph/dir.c @@ -416,9 +416,9 @@ more: fi->last_name = NULL; fi->next_offset = 2; } else { - err = note_last_dentry(fi, - rinfo->dir_dname[rinfo->dir_nr-1], - rinfo->dir_dname_len[rinfo->dir_nr-1], + struct ceph_mds_reply_dir_entry *rde = + rinfo->dir_entries + (rinfo->dir_nr-1); + err = note_last_dentry(fi, rde->name, rde->name_len, fi->next_offset + rinfo->dir_nr); if (err) return err; @@ -431,24 +431,21 @@ more: ctx->pos = ceph_make_fpos(frag, off); while (off >= fi->offset && off - fi->offset < rinfo->dir_nr) { - struct ceph_mds_reply_inode *in = - rinfo->dir_in[off - fi->offset].in; + struct ceph_mds_reply_dir_entry *rde = + rinfo->dir_entries + (off - fi->offset); struct ceph_vino vino; ino_t ino; dout("readdir off %d (%d/%d) -> %lld '%.*s' %p\n", off, off - fi->offset, rinfo->dir_nr, ctx->pos, - rinfo->dir_dname_len[off - fi->offset], - rinfo->dir_dname[off - fi->offset], in); - BUG_ON(!in); - ftype = le32_to_cpu(in->mode) >> 12; - vino.ino = le64_to_cpu(in->ino); - vino.snap = le64_to_cpu(in->snapid); + rde->name_len, rde->name, &rde->inode.in); + BUG_ON(!rde->inode.in); + ftype = le32_to_cpu(rde->inode.in->mode) >> 12; + vino.ino = le64_to_cpu(rde->inode.in->ino); + vino.snap = le64_to_cpu(rde->inode.in->snapid); ino = ceph_vino_to_ino(vino); - if (!dir_emit(ctx, - rinfo->dir_dname[off - fi->offset], - rinfo->dir_dname_len[off - fi->offset], - ceph_translate_ino(inode->i_sb, ino), ftype)) { + if (!dir_emit(ctx, rde->name, rde->name_len, + ceph_translate_ino(inode->i_sb, ino), ftype)) { dout("filldir stopping us...\n"); return 0; } diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c index 51ce5ce59f04..40d081d7028e 100644 --- a/fs/ceph/inode.c +++ b/fs/ceph/inode.c @@ -1308,12 +1308,13 @@ static int readdir_prepopulate_inodes_only(struct ceph_mds_request *req, int i, err = 0; for (i = 0; i < rinfo->dir_nr; i++) { + struct ceph_mds_reply_dir_entry *rde = rinfo->dir_entries + i; struct ceph_vino vino; struct inode *in; int rc; - vino.ino = le64_to_cpu(rinfo->dir_in[i].in->ino); - vino.snap = le64_to_cpu(rinfo->dir_in[i].in->snapid); + vino.ino = le64_to_cpu(rde->inode.in->ino); + vino.snap = le64_to_cpu(rde->inode.in->snapid); in = ceph_get_inode(req->r_dentry->d_sb, vino); if (IS_ERR(in)) { @@ -1321,7 +1322,7 @@ static int readdir_prepopulate_inodes_only(struct ceph_mds_request *req, dout("new_inode badness got %d\n", err); continue; } - rc = fill_inode(in, NULL, &rinfo->dir_in[i], NULL, session, + rc = fill_inode(in, NULL, &rde->inode, NULL, session, req->r_request_started, -1, &req->r_caps_reservation); if (rc < 0) { @@ -1433,14 +1434,15 @@ int ceph_readdir_prepopulate(struct ceph_mds_request *req, /* FIXME: release caps/leases if error occurs */ for (i = 0; i < rinfo->dir_nr; i++) { + struct ceph_mds_reply_dir_entry *rde = rinfo->dir_entries + i; struct ceph_vino vino; - dname.name = rinfo->dir_dname[i]; - dname.len = rinfo->dir_dname_len[i]; + dname.name = rde->name; + dname.len = rde->name_len; dname.hash = full_name_hash(dname.name, dname.len); - vino.ino = le64_to_cpu(rinfo->dir_in[i].in->ino); - vino.snap = le64_to_cpu(rinfo->dir_in[i].in->snapid); + vino.ino = le64_to_cpu(rde->inode.in->ino); + vino.snap = le64_to_cpu(rde->inode.in->snapid); retry_lookup: dn = d_lookup(parent, &dname); @@ -1486,7 +1488,7 @@ retry_lookup: } } - ret = fill_inode(in, NULL, &rinfo->dir_in[i], NULL, session, + ret = fill_inode(in, NULL, &rde->inode, NULL, session, req->r_request_started, -1, &req->r_caps_reservation); if (ret < 0) { @@ -1522,8 +1524,7 @@ retry_lookup: di = dn->d_fsdata; di->offset = ceph_make_fpos(frag, i + req->r_readdir_offset); - update_dentry_lease(dn, rinfo->dir_dlease[i], - req->r_session, + update_dentry_lease(dn, rde->lease, req->r_session, req->r_request_started); if (err == 0 && skipped == 0 && cache_ctl.index >= 0) { diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c index 047f723bdbe6..6220d3caf7ab 100644 --- a/fs/ceph/mds_client.c +++ b/fs/ceph/mds_client.c @@ -186,12 +186,9 @@ static int parse_reply_info_dir(void **p, void *end, if (num == 0) goto done; - BUG_ON(!info->dir_in); - info->dir_dname = (void *)(info->dir_in + num); - info->dir_dname_len = (void *)(info->dir_dname + num); - info->dir_dlease = (void *)(info->dir_dname_len + num); - if ((unsigned long)(info->dir_dlease + num) > - (unsigned long)info->dir_in + info->dir_buf_size) { + BUG_ON(!info->dir_entries); + if ((unsigned long)(info->dir_entries + num) > + (unsigned long)info->dir_entries + info->dir_buf_size) { pr_err("dir contents are larger than expected\n"); WARN_ON(1); goto bad; @@ -199,19 +196,19 @@ static int parse_reply_info_dir(void **p, void *end, info->dir_nr = num; while (num) { + struct ceph_mds_reply_dir_entry *rde = info->dir_entries + i; /* dentry */ ceph_decode_need(p, end, sizeof(u32)*2, bad); - info->dir_dname_len[i] = ceph_decode_32(p); - ceph_decode_need(p, end, info->dir_dname_len[i], bad); - info->dir_dname[i] = *p; - *p += info->dir_dname_len[i]; - dout("parsed dir dname '%.*s'\n", info->dir_dname_len[i], - info->dir_dname[i]); - info->dir_dlease[i] = *p; + rde->name_len = ceph_decode_32(p); + ceph_decode_need(p, end, rde->name_len, bad); + rde->name = *p; + *p += rde->name_len; + dout("parsed dir dname '%.*s'\n", rde->name_len, rde->name); + rde->lease = *p; *p += sizeof(struct ceph_mds_reply_lease); /* inode */ - err = parse_reply_info_in(p, end, &info->dir_in[i], features); + err = parse_reply_info_in(p, end, &rde->inode, features); if (err < 0) goto out_bad; i++; @@ -345,9 +342,9 @@ out_bad: static void destroy_reply_info(struct ceph_mds_reply_info_parsed *info) { - if (!info->dir_in) + if (!info->dir_entries) return; - free_pages((unsigned long)info->dir_in, get_order(info->dir_buf_size)); + free_pages((unsigned long)info->dir_entries, get_order(info->dir_buf_size)); } @@ -1656,8 +1653,7 @@ int ceph_alloc_readdir_reply_buffer(struct ceph_mds_request *req, struct ceph_inode_info *ci = ceph_inode(dir); struct ceph_mds_reply_info_parsed *rinfo = &req->r_reply_info; struct ceph_mount_options *opt = req->r_mdsc->fsc->mount_options; - size_t size = sizeof(*rinfo->dir_in) + sizeof(*rinfo->dir_dname_len) + - sizeof(*rinfo->dir_dname) + sizeof(*rinfo->dir_dlease); + size_t size = sizeof(struct ceph_mds_reply_dir_entry); int order, num_entries; spin_lock(&ci->i_ceph_lock); @@ -1668,14 +1664,14 @@ int ceph_alloc_readdir_reply_buffer(struct ceph_mds_request *req, order = get_order(size * num_entries); while (order >= 0) { - rinfo->dir_in = (void*)__get_free_pages(GFP_KERNEL | - __GFP_NOWARN, - order); - if (rinfo->dir_in) + rinfo->dir_entries = (void*)__get_free_pages(GFP_KERNEL | + __GFP_NOWARN, + order); + if (rinfo->dir_entries) break; order--; } - if (!rinfo->dir_in) + if (!rinfo->dir_entries) return -ENOMEM; num_entries = (PAGE_SIZE << order) / size; diff --git a/fs/ceph/mds_client.h b/fs/ceph/mds_client.h index ee69a537dba5..0b84f9c0afa3 100644 --- a/fs/ceph/mds_client.h +++ b/fs/ceph/mds_client.h @@ -47,6 +47,13 @@ struct ceph_mds_reply_info_in { u32 pool_ns_len; }; +struct ceph_mds_reply_dir_entry { + char *name; + u32 name_len; + struct ceph_mds_reply_lease *lease; + struct ceph_mds_reply_info_in inode; +}; + /* * parsed info about an mds reply, including information about * either: 1) the target inode and/or its parent directory and dentry, @@ -73,11 +80,8 @@ struct ceph_mds_reply_info_parsed { struct ceph_mds_reply_dirfrag *dir_dir; size_t dir_buf_size; int dir_nr; - char **dir_dname; - u32 *dir_dname_len; - struct ceph_mds_reply_lease **dir_dlease; - struct ceph_mds_reply_info_in *dir_in; u8 dir_complete, dir_end; + struct ceph_mds_reply_dir_entry *dir_entries; }; /* for create results */ -- cgit v1.3-8-gc7d7 From 956d39d631dbcf7b57854873a24e309047f2a7f5 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Wed, 27 Apr 2016 17:48:30 +0800 Subject: ceph: define 'end/complete' in readdir reply as bit flags Set a flag in readdir request, which indicates that client interprets 'end/complete' as bit flags. So that mds can reply additional flags in readdir reply. Signed-off-by: Yan, Zheng --- fs/ceph/dir.c | 2 ++ fs/ceph/mds_client.c | 7 +++++-- fs/ceph/mds_client.h | 2 +- include/linux/ceph/ceph_fs.h | 12 ++++++++++++ 4 files changed, 20 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c index 68530acea2c8..ebcbd1c946b4 100644 --- a/fs/ceph/dir.c +++ b/fs/ceph/dir.c @@ -365,6 +365,8 @@ more: req->r_readdir_cache_idx = fi->readdir_cache_idx; req->r_readdir_offset = fi->next_offset; req->r_args.readdir.frag = cpu_to_le32(frag); + req->r_args.readdir.flags = + cpu_to_le16(CEPH_READDIR_REPLY_BITFLAGS); req->r_inode = inode; ihold(inode); diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c index 6220d3caf7ab..1c2befcd24fb 100644 --- a/fs/ceph/mds_client.c +++ b/fs/ceph/mds_client.c @@ -181,8 +181,11 @@ static int parse_reply_info_dir(void **p, void *end, ceph_decode_need(p, end, sizeof(num) + 2, bad); num = ceph_decode_32(p); - info->dir_end = ceph_decode_8(p); - info->dir_complete = ceph_decode_8(p); + { + u16 flags = ceph_decode_16(p); + info->dir_end = !!(flags & CEPH_READDIR_FRAG_END); + info->dir_complete = !!(flags & CEPH_READDIR_FRAG_COMPLETE); + } if (num == 0) goto done; diff --git a/fs/ceph/mds_client.h b/fs/ceph/mds_client.h index 0b84f9c0afa3..2a865812a41b 100644 --- a/fs/ceph/mds_client.h +++ b/fs/ceph/mds_client.h @@ -80,7 +80,7 @@ struct ceph_mds_reply_info_parsed { struct ceph_mds_reply_dirfrag *dir_dir; size_t dir_buf_size; int dir_nr; - u8 dir_complete, dir_end; + bool dir_complete, dir_end; struct ceph_mds_reply_dir_entry *dir_entries; }; diff --git a/include/linux/ceph/ceph_fs.h b/include/linux/ceph/ceph_fs.h index bae833d0d055..a811c5e98bfa 100644 --- a/include/linux/ceph/ceph_fs.h +++ b/include/linux/ceph/ceph_fs.h @@ -347,6 +347,17 @@ extern const char *ceph_mds_op_name(int op); #define CEPH_XATTR_REPLACE (1 << 1) #define CEPH_XATTR_REMOVE (1 << 31) +/* + * readdir request flags; + */ +#define CEPH_READDIR_REPLY_BITFLAGS (1<<0) + +/* + * readdir reply flags. + */ +#define CEPH_READDIR_FRAG_END (1<<0) +#define CEPH_READDIR_FRAG_COMPLETE (1<<8) + union ceph_mds_request_args { struct { __le32 mask; /* CEPH_CAP_* */ @@ -364,6 +375,7 @@ union ceph_mds_request_args { __le32 frag; /* which dir fragment */ __le32 max_entries; /* how many dentries to grab */ __le32 max_bytes; + __le16 flags; } __attribute__ ((packed)) readdir; struct { __le32 mode; -- cgit v1.3-8-gc7d7 From 8974eebd38737c9534d81c4131c5fdb1fe24d3e9 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Thu, 28 Apr 2016 15:17:40 +0800 Subject: ceph: record 'offset' for each entry of readdir result This is preparation for using hash value as dentry 'offset' Signed-off-by: Yan, Zheng --- fs/ceph/dir.c | 83 ++++++++++++++++++++++++++++++++++------------------ fs/ceph/inode.c | 1 + fs/ceph/mds_client.c | 2 ++ fs/ceph/mds_client.h | 1 + fs/ceph/super.h | 1 - 5 files changed, 59 insertions(+), 29 deletions(-) (limited to 'fs') diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c index ebcbd1c946b4..6ae635605be5 100644 --- a/fs/ceph/dir.c +++ b/fs/ceph/dir.c @@ -277,12 +277,12 @@ static int ceph_readdir(struct file *file, struct dir_context *ctx) struct ceph_fs_client *fsc = ceph_inode_to_client(inode); struct ceph_mds_client *mdsc = fsc->mdsc; unsigned frag = fpos_frag(ctx->pos); - int off = fpos_off(ctx->pos); + int i; int err; u32 ftype; struct ceph_mds_reply_info_parsed *rinfo; - dout("readdir %p file %p frag %u off %u\n", inode, file, frag, off); + dout("readdir %p file %p pos %llx\n", inode, file, ctx->pos); if (fi->flags & CEPH_F_ATEND) return 0; @@ -294,7 +294,6 @@ static int ceph_readdir(struct file *file, struct dir_context *ctx) inode->i_mode >> 12)) return 0; ctx->pos = 1; - off = 1; } if (ctx->pos == 1) { ino_t ino = parent_ino(file->f_path.dentry); @@ -304,7 +303,6 @@ static int ceph_readdir(struct file *file, struct dir_context *ctx) inode->i_mode >> 12)) return 0; ctx->pos = 2; - off = 2; } /* can we use the dcache? */ @@ -320,7 +318,6 @@ static int ceph_readdir(struct file *file, struct dir_context *ctx) if (err != -EAGAIN) return err; frag = fpos_frag(ctx->pos); - off = fpos_off(ctx->pos); } else { spin_unlock(&ci->i_ceph_lock); } @@ -386,12 +383,12 @@ more: rinfo = &req->r_reply_info; if (le32_to_cpu(rinfo->dir_dir->frag) != frag) { frag = le32_to_cpu(rinfo->dir_dir->frag); - off = req->r_readdir_offset; - fi->next_offset = off; + fi->next_offset = req->r_readdir_offset; + /* adjust ctx->pos to beginning of frag */ + ctx->pos = ceph_make_fpos(frag, fi->next_offset); } fi->frag = frag; - fi->offset = fi->next_offset; fi->last_readdir = req; if (req->r_did_prepopulate) { @@ -399,7 +396,8 @@ more: if (fi->readdir_cache_idx < 0) { /* preclude from marking dir ordered */ fi->dir_ordered_count = 0; - } else if (ceph_frag_is_leftmost(frag) && off == 2) { + } else if (ceph_frag_is_leftmost(frag) && + fi->next_offset == 2) { /* note dir version at start of readdir so * we can tell if any dentries get dropped */ fi->dir_release_count = req->r_dir_release_cnt; @@ -421,37 +419,54 @@ more: struct ceph_mds_reply_dir_entry *rde = rinfo->dir_entries + (rinfo->dir_nr-1); err = note_last_dentry(fi, rde->name, rde->name_len, - fi->next_offset + rinfo->dir_nr); + fpos_off(rde->offset) + 1); if (err) return err; } } rinfo = &fi->last_readdir->r_reply_info; - dout("readdir frag %x num %d off %d chunkoff %d\n", frag, - rinfo->dir_nr, off, fi->offset); + dout("readdir frag %x num %d pos %llx chunk first %llx\n", + frag, rinfo->dir_nr, ctx->pos, + rinfo->dir_nr ? rinfo->dir_entries[0].offset : 0LL); - ctx->pos = ceph_make_fpos(frag, off); - while (off >= fi->offset && off - fi->offset < rinfo->dir_nr) { - struct ceph_mds_reply_dir_entry *rde = - rinfo->dir_entries + (off - fi->offset); + i = 0; + /* search start position */ + if (rinfo->dir_nr > 0) { + int step, nr = rinfo->dir_nr; + while (nr > 0) { + step = nr >> 1; + if (rinfo->dir_entries[i + step].offset < ctx->pos) { + i += step + 1; + nr -= step + 1; + } else { + nr = step; + } + } + } + for (; i < rinfo->dir_nr; i++) { + struct ceph_mds_reply_dir_entry *rde = rinfo->dir_entries + i; struct ceph_vino vino; ino_t ino; - dout("readdir off %d (%d/%d) -> %lld '%.*s' %p\n", - off, off - fi->offset, rinfo->dir_nr, ctx->pos, + BUG_ON(rde->offset < ctx->pos); + + ctx->pos = rde->offset; + dout("readdir (%d/%d) -> %llx '%.*s' %p\n", + i, rinfo->dir_nr, ctx->pos, rde->name_len, rde->name, &rde->inode.in); + BUG_ON(!rde->inode.in); ftype = le32_to_cpu(rde->inode.in->mode) >> 12; vino.ino = le64_to_cpu(rde->inode.in->ino); vino.snap = le64_to_cpu(rde->inode.in->snapid); ino = ceph_vino_to_ino(vino); + if (!dir_emit(ctx, rde->name, rde->name_len, ceph_translate_ino(inode->i_sb, ino), ftype)) { dout("filldir stopping us...\n"); return 0; } - off++; ctx->pos++; } @@ -464,8 +479,7 @@ more: /* more frags? */ if (!ceph_frag_is_rightmost(frag)) { frag = ceph_frag_next(frag); - off = 2; - ctx->pos = ceph_make_fpos(frag, off); + ctx->pos = ceph_make_fpos(frag, 2); dout("readdir next frag is %x\n", frag); goto more; } @@ -497,7 +511,7 @@ more: return 0; } -static void reset_readdir(struct ceph_file_info *fi, unsigned frag) +static void reset_readdir(struct ceph_file_info *fi) { if (fi->last_readdir) { ceph_mdsc_put_request(fi->last_readdir); @@ -511,6 +525,23 @@ static void reset_readdir(struct ceph_file_info *fi, unsigned frag) fi->flags &= ~CEPH_F_ATEND; } +/* + * discard buffered readdir content on seekdir(0), or seek to new frag, + * or seek prior to current chunk + */ +static bool need_reset_readdir(struct ceph_file_info *fi, loff_t new_pos) +{ + struct ceph_mds_reply_info_parsed *rinfo; + if (new_pos == 0) + return true; + if (fpos_frag(new_pos) != fi->frag) + return true; + rinfo = fi->last_readdir ? &fi->last_readdir->r_reply_info : NULL; + if (!rinfo || !rinfo->dir_nr) + return true; + return new_pos < rinfo->dir_entries[0].offset;; +} + static loff_t ceph_dir_llseek(struct file *file, loff_t offset, int whence) { struct ceph_file_info *fi = file->private_data; @@ -539,13 +570,9 @@ static loff_t ceph_dir_llseek(struct file *file, loff_t offset, int whence) } retval = offset; - if (offset == 0 || - fpos_frag(offset) != fi->frag || - fpos_off(offset) < fi->offset) { - /* discard buffered readdir content on seekdir(0), or - * seek to new frag, or seek prior to current chunk */ + if (need_reset_readdir(fi, offset)) { dout("dir_llseek dropping %p content\n", file); - reset_readdir(fi, fpos_frag(offset)); + reset_readdir(fi); } else if (fpos_cmp(offset, old_offset) > 0) { /* reset dir_release_count if we did a forward seek */ fi->dir_release_count = 0; diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c index 40d081d7028e..b53c95903aeb 100644 --- a/fs/ceph/inode.c +++ b/fs/ceph/inode.c @@ -1523,6 +1523,7 @@ retry_lookup: di = dn->d_fsdata; di->offset = ceph_make_fpos(frag, i + req->r_readdir_offset); + rde->offset = di->offset; update_dentry_lease(dn, rde->lease, req->r_session, req->r_request_started); diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c index 1c2befcd24fb..48def22fc7b9 100644 --- a/fs/ceph/mds_client.c +++ b/fs/ceph/mds_client.c @@ -214,6 +214,8 @@ static int parse_reply_info_dir(void **p, void *end, err = parse_reply_info_in(p, end, &rde->inode, features); if (err < 0) goto out_bad; + /* ceph_readdir_prepopulate() will update it */ + rde->offset = 0; i++; num--; } diff --git a/fs/ceph/mds_client.h b/fs/ceph/mds_client.h index 2a865812a41b..4ce19d852657 100644 --- a/fs/ceph/mds_client.h +++ b/fs/ceph/mds_client.h @@ -52,6 +52,7 @@ struct ceph_mds_reply_dir_entry { u32 name_len; struct ceph_mds_reply_lease *lease; struct ceph_mds_reply_info_in inode; + loff_t offset; }; /* diff --git a/fs/ceph/super.h b/fs/ceph/super.h index 0ea86406f463..0628099ba1f2 100644 --- a/fs/ceph/super.h +++ b/fs/ceph/super.h @@ -635,7 +635,6 @@ struct ceph_file_info { struct ceph_mds_request *last_readdir; /* readdir: position within a frag */ - unsigned offset; /* offset of last chunk, adjusted for . and .. */ unsigned next_offset; /* offset of next chunk (last_name's + 1) */ char *last_name; /* last entry in previous chunk */ long long dir_release_count; -- cgit v1.3-8-gc7d7 From 076c40f18d10489e29c515bf5936952830df5e16 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Thu, 28 Apr 2016 22:56:44 +0800 Subject: ceph: don't forbid marking directory complete after forward seek Forward seek within same frag does not update fi->last_name, it will not affect contents of later readdir reply. So there is no need to forbid marking directory complete Signed-off-by: Yan, Zheng --- fs/ceph/dir.c | 5 ----- 1 file changed, 5 deletions(-) (limited to 'fs') diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c index 6ae635605be5..e954ea2fb710 100644 --- a/fs/ceph/dir.c +++ b/fs/ceph/dir.c @@ -546,7 +546,6 @@ static loff_t ceph_dir_llseek(struct file *file, loff_t offset, int whence) { struct ceph_file_info *fi = file->private_data; struct inode *inode = file->f_mapping->host; - loff_t old_offset = ceph_make_fpos(fi->frag, fi->next_offset); loff_t retval; inode_lock(inode); @@ -573,10 +572,6 @@ static loff_t ceph_dir_llseek(struct file *file, loff_t offset, int whence) if (need_reset_readdir(fi, offset)) { dout("dir_llseek dropping %p content\n", file); reset_readdir(fi); - } else if (fpos_cmp(offset, old_offset) > 0) { - /* reset dir_release_count if we did a forward seek */ - fi->dir_release_count = 0; - fi->readdir_cache_idx = -1; } } out: -- cgit v1.3-8-gc7d7 From f3c4ebe65ea149ec892f94474233cfebe9cbe299 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Fri, 29 Apr 2016 11:27:30 +0800 Subject: ceph: using hash value to compose dentry offset If MDS sorts dentries in dirfrag in hash order, we use hash value to compose dentry offset. dentry offset is: (0xff << 52) | ((24 bits hash) << 28) | (the nth entry hash hash collision) This offset is stable across directory fragmentation. This alos means there is no need to reset readdir offset if directory get fragmented in the middle of readdir. Signed-off-by: Yan, Zheng --- fs/ceph/dir.c | 140 ++++++++++++++++++++++++++++++++----------- fs/ceph/inode.c | 31 ++++++++-- fs/ceph/mds_client.c | 1 + fs/ceph/mds_client.h | 4 +- fs/ceph/super.h | 6 +- include/linux/ceph/ceph_fs.h | 1 + 6 files changed, 136 insertions(+), 47 deletions(-) (limited to 'fs') diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c index e954ea2fb710..4850c3624a87 100644 --- a/fs/ceph/dir.c +++ b/fs/ceph/dir.c @@ -69,16 +69,42 @@ out_unlock: } /* - * for readdir, we encode the directory frag and offset within that - * frag into f_pos. + * for f_pos for readdir: + * - hash order: + * (0xff << 52) | ((24 bits hash) << 28) | + * (the nth entry has hash collision); + * - frag+name order; + * ((frag value) << 28) | (the nth entry in frag); */ +#define OFFSET_BITS 28 +#define OFFSET_MASK ((1 << OFFSET_BITS) - 1) +#define HASH_ORDER (0xffull << (OFFSET_BITS + 24)) +loff_t ceph_make_fpos(unsigned high, unsigned off, bool hash_order) +{ + loff_t fpos = ((loff_t)high << 28) | (loff_t)off; + if (hash_order) + fpos |= HASH_ORDER; + return fpos; +} + +static bool is_hash_order(loff_t p) +{ + return (p & HASH_ORDER) == HASH_ORDER; +} + static unsigned fpos_frag(loff_t p) { - return p >> 32; + return p >> OFFSET_BITS; } + +static unsigned fpos_hash(loff_t p) +{ + return ceph_frag_value(fpos_frag(p)); +} + static unsigned fpos_off(loff_t p) { - return p & 0xffffffff; + return p & OFFSET_MASK; } static int fpos_cmp(loff_t l, loff_t r) @@ -177,7 +203,7 @@ static int __dcache_readdir(struct file *file, struct dir_context *ctx, u64 idx = 0; int err = 0; - dout("__dcache_readdir %p v%u at %llu\n", dir, shared_gen, ctx->pos); + dout("__dcache_readdir %p v%u at %llx\n", dir, shared_gen, ctx->pos); /* search start position */ if (ctx->pos > 2) { @@ -234,7 +260,7 @@ static int __dcache_readdir(struct file *file, struct dir_context *ctx, spin_unlock(&dentry->d_lock); if (emit_dentry) { - dout(" %llu (%llu) dentry %p %pd %p\n", di->offset, ctx->pos, + dout(" %llx dentry %p %pd %p\n", di->offset, dentry, dentry, d_inode(dentry)); ctx->pos = di->offset; if (!dir_emit(ctx, dentry->d_name.name, @@ -269,6 +295,16 @@ out: return err; } +static bool need_send_readdir(struct ceph_file_info *fi, loff_t pos) +{ + if (!fi->last_readdir) + return true; + if (is_hash_order(pos)) + return !ceph_frag_contains_value(fi->frag, fpos_hash(pos)); + else + return fi->frag != fpos_frag(pos); +} + static int ceph_readdir(struct file *file, struct dir_context *ctx) { struct ceph_file_info *fi = file->private_data; @@ -276,7 +312,6 @@ static int ceph_readdir(struct file *file, struct dir_context *ctx) struct ceph_inode_info *ci = ceph_inode(inode); struct ceph_fs_client *fsc = ceph_inode_to_client(inode); struct ceph_mds_client *mdsc = fsc->mdsc; - unsigned frag = fpos_frag(ctx->pos); int i; int err; u32 ftype; @@ -317,7 +352,6 @@ static int ceph_readdir(struct file *file, struct dir_context *ctx) err = __dcache_readdir(file, ctx, shared_gen); if (err != -EAGAIN) return err; - frag = fpos_frag(ctx->pos); } else { spin_unlock(&ci->i_ceph_lock); } @@ -325,8 +359,9 @@ static int ceph_readdir(struct file *file, struct dir_context *ctx) /* proceed with a normal readdir */ more: /* do we have the correct frag content buffered? */ - if (fi->frag != frag || fi->last_readdir == NULL) { + if (need_send_readdir(fi, ctx->pos)) { struct ceph_mds_request *req; + unsigned frag; int op = ceph_snap(inode) == CEPH_SNAPDIR ? CEPH_MDS_OP_LSSNAP : CEPH_MDS_OP_READDIR; @@ -336,6 +371,13 @@ more: fi->last_readdir = NULL; } + if (is_hash_order(ctx->pos)) { + frag = ceph_choose_frag(ci, fpos_hash(ctx->pos), + NULL, NULL); + } else { + frag = fpos_frag(ctx->pos); + } + dout("readdir fetching %llx.%llx frag %x offset '%s'\n", ceph_vinop(inode), frag, fi->last_name); req = ceph_mdsc_create_request(mdsc, op, USE_AUTH_MDS); @@ -373,19 +415,23 @@ more: ceph_mdsc_put_request(req); return err; } - dout("readdir got and parsed readdir result=%d" - " on frag %x, end=%d, complete=%d\n", err, frag, + dout("readdir got and parsed readdir result=%d on " + "frag %x, end=%d, complete=%d, hash_order=%d\n", + err, frag, (int)req->r_reply_info.dir_end, - (int)req->r_reply_info.dir_complete); - + (int)req->r_reply_info.dir_complete, + (int)req->r_reply_info.hash_order); - /* note next offset and last dentry name */ rinfo = &req->r_reply_info; if (le32_to_cpu(rinfo->dir_dir->frag) != frag) { frag = le32_to_cpu(rinfo->dir_dir->frag); - fi->next_offset = req->r_readdir_offset; - /* adjust ctx->pos to beginning of frag */ - ctx->pos = ceph_make_fpos(frag, fi->next_offset); + if (!rinfo->hash_order) { + fi->next_offset = req->r_readdir_offset; + /* adjust ctx->pos to beginning of frag */ + ctx->pos = ceph_make_fpos(frag, + fi->next_offset, + false); + } } fi->frag = frag; @@ -411,23 +457,25 @@ more: fi->dir_release_count = 0; } - if (req->r_reply_info.dir_end) { - kfree(fi->last_name); - fi->last_name = NULL; - fi->next_offset = 2; - } else { + /* note next offset and last dentry name */ + if (rinfo->dir_nr > 0) { struct ceph_mds_reply_dir_entry *rde = rinfo->dir_entries + (rinfo->dir_nr-1); + unsigned next_offset = req->r_reply_info.dir_end ? + 2 : (fpos_off(rde->offset) + 1); err = note_last_dentry(fi, rde->name, rde->name_len, - fpos_off(rde->offset) + 1); + next_offset); if (err) return err; + } else if (req->r_reply_info.dir_end) { + fi->next_offset = 2; + /* keep last name */ } } rinfo = &fi->last_readdir->r_reply_info; dout("readdir frag %x num %d pos %llx chunk first %llx\n", - frag, rinfo->dir_nr, ctx->pos, + fi->frag, rinfo->dir_nr, ctx->pos, rinfo->dir_nr ? rinfo->dir_entries[0].offset : 0LL); i = 0; @@ -470,16 +518,26 @@ more: ctx->pos++; } - if (fi->last_name) { + if (fi->next_offset > 2) { ceph_mdsc_put_request(fi->last_readdir); fi->last_readdir = NULL; goto more; } /* more frags? */ - if (!ceph_frag_is_rightmost(frag)) { - frag = ceph_frag_next(frag); - ctx->pos = ceph_make_fpos(frag, 2); + if (!ceph_frag_is_rightmost(fi->frag)) { + unsigned frag = ceph_frag_next(fi->frag); + if (is_hash_order(ctx->pos)) { + loff_t new_pos = ceph_make_fpos(ceph_frag_value(frag), + fi->next_offset, true); + if (new_pos > ctx->pos) + ctx->pos = new_pos; + /* keep last_name */ + } else { + ctx->pos = ceph_make_fpos(frag, fi->next_offset, false); + kfree(fi->last_name); + fi->last_name = NULL; + } dout("readdir next frag is %x\n", frag); goto more; } @@ -532,14 +590,21 @@ static void reset_readdir(struct ceph_file_info *fi) static bool need_reset_readdir(struct ceph_file_info *fi, loff_t new_pos) { struct ceph_mds_reply_info_parsed *rinfo; + loff_t chunk_offset; if (new_pos == 0) return true; - if (fpos_frag(new_pos) != fi->frag) + if (is_hash_order(new_pos)) { + /* no need to reset last_name for a forward seek when + * dentries are sotred in hash order */ + } else if (fi->frag |= fpos_frag(new_pos)) { return true; + } rinfo = fi->last_readdir ? &fi->last_readdir->r_reply_info : NULL; if (!rinfo || !rinfo->dir_nr) return true; - return new_pos < rinfo->dir_entries[0].offset;; + chunk_offset = rinfo->dir_entries[0].offset; + return new_pos < chunk_offset || + is_hash_order(new_pos) != is_hash_order(chunk_offset); } static loff_t ceph_dir_llseek(struct file *file, loff_t offset, int whence) @@ -562,17 +627,22 @@ static loff_t ceph_dir_llseek(struct file *file, loff_t offset, int whence) } if (offset >= 0) { + if (need_reset_readdir(fi, offset)) { + dout("dir_llseek dropping %p content\n", file); + reset_readdir(fi); + } else if (is_hash_order(offset) && offset > file->f_pos) { + /* for hash offset, we don't know if a forward seek + * is within same frag */ + fi->dir_release_count = 0; + fi->readdir_cache_idx = -1; + } + if (offset != file->f_pos) { file->f_pos = offset; file->f_version = 0; fi->flags &= ~CEPH_F_ATEND; } retval = offset; - - if (need_reset_readdir(fi, offset)) { - dout("dir_llseek dropping %p content\n", file); - reset_readdir(fi); - } } out: inode_unlock(inode); diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c index b53c95903aeb..f51b6fd5f570 100644 --- a/fs/ceph/inode.c +++ b/fs/ceph/inode.c @@ -1387,6 +1387,7 @@ int ceph_readdir_prepopulate(struct ceph_mds_request *req, struct ceph_mds_session *session) { struct dentry *parent = req->r_dentry; + struct ceph_inode_info *ci = ceph_inode(d_inode(parent)); struct ceph_mds_reply_info_parsed *rinfo = &req->r_reply_info; struct qstr dname; struct dentry *dn; @@ -1394,19 +1395,27 @@ int ceph_readdir_prepopulate(struct ceph_mds_request *req, int err = 0, skipped = 0, ret, i; struct inode *snapdir = NULL; struct ceph_mds_request_head *rhead = req->r_request->front.iov_base; - struct ceph_dentry_info *di; u32 frag = le32_to_cpu(rhead->args.readdir.frag); + u32 last_hash = 0; + u32 fpos_offset; struct ceph_readdir_cache_control cache_ctl = {}; if (req->r_aborted) return readdir_prepopulate_inodes_only(req, session); + if (rinfo->hash_order && req->r_path2) { + last_hash = ceph_str_hash(ci->i_dir_layout.dl_dir_hash, + req->r_path2, strlen(req->r_path2)); + last_hash = ceph_frag_value(last_hash); + } + if (rinfo->dir_dir && le32_to_cpu(rinfo->dir_dir->frag) != frag) { dout("readdir_prepopulate got new frag %x -> %x\n", frag, le32_to_cpu(rinfo->dir_dir->frag)); frag = le32_to_cpu(rinfo->dir_dir->frag); - req->r_readdir_offset = 2; + if (!rinfo->hash_order) + req->r_readdir_offset = 2; } if (le32_to_cpu(rinfo->head->op) == CEPH_MDS_OP_LSSNAP) { @@ -1424,13 +1433,13 @@ int ceph_readdir_prepopulate(struct ceph_mds_request *req, if (ceph_frag_is_leftmost(frag) && req->r_readdir_offset == 2) { /* note dir version at start of readdir so we can tell * if any dentries get dropped */ - struct ceph_inode_info *ci = ceph_inode(d_inode(parent)); req->r_dir_release_cnt = atomic64_read(&ci->i_release_count); req->r_dir_ordered_cnt = atomic64_read(&ci->i_ordered_count); req->r_readdir_cache_idx = 0; } cache_ctl.index = req->r_readdir_cache_idx; + fpos_offset = req->r_readdir_offset; /* FIXME: release caps/leases if error occurs */ for (i = 0; i < rinfo->dir_nr; i++) { @@ -1444,6 +1453,18 @@ int ceph_readdir_prepopulate(struct ceph_mds_request *req, vino.ino = le64_to_cpu(rde->inode.in->ino); vino.snap = le64_to_cpu(rde->inode.in->snapid); + if (rinfo->hash_order) { + u32 hash = ceph_str_hash(ci->i_dir_layout.dl_dir_hash, + rde->name, rde->name_len); + hash = ceph_frag_value(hash); + if (hash != last_hash) + fpos_offset = 2; + last_hash = hash; + rde->offset = ceph_make_fpos(hash, fpos_offset++, true); + } else { + rde->offset = ceph_make_fpos(frag, fpos_offset++, false); + } + retry_lookup: dn = d_lookup(parent, &dname); dout("d_lookup on parent=%p name=%.*s got %p\n", @@ -1521,9 +1542,7 @@ retry_lookup: dn = realdn; } - di = dn->d_fsdata; - di->offset = ceph_make_fpos(frag, i + req->r_readdir_offset); - rde->offset = di->offset; + ceph_dentry(dn)->offset = rde->offset; update_dentry_lease(dn, rde->lease, req->r_session, req->r_request_started); diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c index 48def22fc7b9..7ad31283d510 100644 --- a/fs/ceph/mds_client.c +++ b/fs/ceph/mds_client.c @@ -185,6 +185,7 @@ static int parse_reply_info_dir(void **p, void *end, u16 flags = ceph_decode_16(p); info->dir_end = !!(flags & CEPH_READDIR_FRAG_END); info->dir_complete = !!(flags & CEPH_READDIR_FRAG_COMPLETE); + info->hash_order = !!(flags & CEPH_READDIR_HASH_ORDER); } if (num == 0) goto done; diff --git a/fs/ceph/mds_client.h b/fs/ceph/mds_client.h index 4ce19d852657..e7d38aac7109 100644 --- a/fs/ceph/mds_client.h +++ b/fs/ceph/mds_client.h @@ -81,7 +81,9 @@ struct ceph_mds_reply_info_parsed { struct ceph_mds_reply_dirfrag *dir_dir; size_t dir_buf_size; int dir_nr; - bool dir_complete, dir_end; + bool dir_complete; + bool dir_end; + bool hash_order; struct ceph_mds_reply_dir_entry *dir_entries; }; diff --git a/fs/ceph/super.h b/fs/ceph/super.h index 0628099ba1f2..c9b671dfff81 100644 --- a/fs/ceph/super.h +++ b/fs/ceph/super.h @@ -540,11 +540,6 @@ static inline struct ceph_dentry_info *ceph_dentry(struct dentry *dentry) return (struct ceph_dentry_info *)dentry->d_fsdata; } -static inline loff_t ceph_make_fpos(unsigned frag, unsigned off) -{ - return ((loff_t)frag << 32) | (loff_t)off; -} - /* * caps helpers */ @@ -949,6 +944,7 @@ extern const struct inode_operations ceph_snapdir_iops; extern const struct dentry_operations ceph_dentry_ops, ceph_snap_dentry_ops, ceph_snapdir_dentry_ops; +extern loff_t ceph_make_fpos(unsigned high, unsigned off, bool hash_order); extern int ceph_handle_notrace_create(struct inode *dir, struct dentry *dentry); extern int ceph_handle_snapdir(struct ceph_mds_request *req, struct dentry *dentry, int err); diff --git a/include/linux/ceph/ceph_fs.h b/include/linux/ceph/ceph_fs.h index a811c5e98bfa..dfce616002ad 100644 --- a/include/linux/ceph/ceph_fs.h +++ b/include/linux/ceph/ceph_fs.h @@ -357,6 +357,7 @@ extern const char *ceph_mds_op_name(int op); */ #define CEPH_READDIR_FRAG_END (1<<0) #define CEPH_READDIR_FRAG_COMPLETE (1<<8) +#define CEPH_READDIR_HASH_ORDER (1<<9) union ceph_mds_request_args { struct { -- cgit v1.3-8-gc7d7 From 209ae762a64397557be2bc48490fd968912e9a71 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Fri, 29 Apr 2016 23:40:23 +0800 Subject: ceph: fix inode reference leak Signed-off-by: Yan, Zheng --- fs/ceph/inode.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c index f51b6fd5f570..4889aaa72acc 100644 --- a/fs/ceph/inode.c +++ b/fs/ceph/inode.c @@ -1328,8 +1328,8 @@ static int readdir_prepopulate_inodes_only(struct ceph_mds_request *req, if (rc < 0) { pr_err("fill_inode badness on %p got %d\n", in, rc); err = rc; - continue; } + iput(in); } return err; -- cgit v1.3-8-gc7d7 From a407846ef7c6e7905ab0d1ebc0c86fbebb065b87 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Tue, 3 May 2016 20:55:50 +0800 Subject: ceph: don't assume frag tree splits in mds reply are sorted The algorithm that updates i_fragtree relies on that the frag tree splits in mds reply are of the same order of i_fragtree. This is not true because current MDS encodes frag tree splits in ascending order of (unsigned)frag_t. But nodes in i_fragtree are sorted according to ceph_frag_compare(). The fix is sort the frag tree splits first, then updates i_fragtree. Signed-off-by: Yan, Zheng --- fs/ceph/inode.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) (limited to 'fs') diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c index 4889aaa72acc..d91eb6b49259 100644 --- a/fs/ceph/inode.c +++ b/fs/ceph/inode.c @@ -10,6 +10,7 @@ #include #include #include +#include #include "super.h" #include "mds_client.h" @@ -299,6 +300,13 @@ out: return err; } +static int frag_tree_split_cmp(const void *l, const void *r) +{ + struct ceph_frag_tree_split *ls = (struct ceph_frag_tree_split*)l; + struct ceph_frag_tree_split *rs = (struct ceph_frag_tree_split*)r; + return ceph_frag_compare(ls->frag, rs->frag); +} + static int ceph_fill_fragtree(struct inode *inode, struct ceph_frag_tree_head *fragtree, struct ceph_mds_reply_dirfrag *dirinfo) @@ -331,6 +339,11 @@ static int ceph_fill_fragtree(struct inode *inode, if (!update) goto out_unlock; + if (nsplits > 1) { + sort(fragtree->splits, nsplits, sizeof(fragtree->splits[0]), + frag_tree_split_cmp, NULL); + } + dout("fill_fragtree %llx.%llx\n", ceph_vinop(inode)); rb_node = rb_first(&ci->i_fragtree); for (i = 0; i < nsplits; i++) { -- cgit v1.3-8-gc7d7 From 421721195a68f46a8218c664154076b5b06f5f51 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Tue, 3 May 2016 22:33:20 +0800 Subject: ceph: fix dir_auth check in ceph_fill_dirfrag() -1 is CDIR_AUTH_PARENT, it means dir's auth mds is the same as inode's auth mds Signed-off-by: Yan, Zheng --- fs/ceph/inode.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'fs') diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c index d91eb6b49259..562f57d15f46 100644 --- a/fs/ceph/inode.c +++ b/fs/ceph/inode.c @@ -254,6 +254,9 @@ static int ceph_fill_dirfrag(struct inode *inode, diri_auth = ci->i_auth_cap->mds; spin_unlock(&ci->i_ceph_lock); + if (mds == -1) /* CDIR_AUTH_PARENT */ + mds = diri_auth; + mutex_lock(&ci->i_fragtree_mutex); if (ndist == 0 && mds == diri_auth) { /* no delegation info needed. */ -- cgit v1.3-8-gc7d7 From a4b7431f39438f415cef1fa28502ec58016686f7 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Wed, 4 May 2016 11:05:10 +0800 Subject: ceph: keep leaf frag when updating fragtree Nodes in i_fragtree are sorted according to ceph_compare_frag(). It means frag node in i_fragtree always follow its direct parent node. To check if a leaf node is valid, we just need to check if it's child of previous split node. Signed-off-by: Yan, Zheng --- fs/ceph/inode.c | 28 +++++++++++++++++++++++----- 1 file changed, 23 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c index 562f57d15f46..09713f32ea88 100644 --- a/fs/ceph/inode.c +++ b/fs/ceph/inode.c @@ -310,12 +310,21 @@ static int frag_tree_split_cmp(const void *l, const void *r) return ceph_frag_compare(ls->frag, rs->frag); } +static bool is_frag_child(u32 f, struct ceph_inode_frag *frag) +{ + if (!frag) + return f == ceph_frag_make(0, 0); + if (ceph_frag_bits(f) != ceph_frag_bits(frag->frag) + frag->split_by) + return false; + return ceph_frag_contains_value(frag->frag, ceph_frag_value(f)); +} + static int ceph_fill_fragtree(struct inode *inode, struct ceph_frag_tree_head *fragtree, struct ceph_mds_reply_dirfrag *dirinfo) { struct ceph_inode_info *ci = ceph_inode(inode); - struct ceph_inode_frag *frag; + struct ceph_inode_frag *frag, *prev_frag = NULL; struct rb_node *rb_node; int i; u32 id, nsplits; @@ -362,8 +371,12 @@ static int ceph_fill_fragtree(struct inode *inode, break; } rb_node = rb_next(rb_node); - rb_erase(&frag->node, &ci->i_fragtree); - kfree(frag); + /* delete stale split/leaf node */ + if (frag->split_by > 0 || + !is_frag_child(frag->frag, prev_frag)) { + rb_erase(&frag->node, &ci->i_fragtree); + kfree(frag); + } frag = NULL; } if (!frag) { @@ -373,12 +386,17 @@ static int ceph_fill_fragtree(struct inode *inode, } frag->split_by = le32_to_cpu(fragtree->splits[i].by); dout(" frag %x split by %d\n", frag->frag, frag->split_by); + prev_frag = frag; } while (rb_node) { frag = rb_entry(rb_node, struct ceph_inode_frag, node); rb_node = rb_next(rb_node); - rb_erase(&frag->node, &ci->i_fragtree); - kfree(frag); + /* delete stale split/leaf node */ + if (frag->split_by > 0 || + !is_frag_child(frag->frag, prev_frag)) { + rb_erase(&frag->node, &ci->i_fragtree); + kfree(frag); + } } out_unlock: mutex_unlock(&ci->i_fragtree_mutex); -- cgit v1.3-8-gc7d7 From 1b1bc16d66a7c7af3b4f30d1cf5a363168b217f4 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Wed, 4 May 2016 11:40:30 +0800 Subject: ceph: improve fragtree change detection check if number of splits in i_fragtree is equal to number of splits in mds reply Signed-off-by: Yan, Zheng --- fs/ceph/inode.c | 24 ++++++++++++++++++++---- fs/ceph/super.h | 1 + 2 files changed, 21 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c index 09713f32ea88..7ba8e1c2557b 100644 --- a/fs/ceph/inode.c +++ b/fs/ceph/inode.c @@ -326,13 +326,15 @@ static int ceph_fill_fragtree(struct inode *inode, struct ceph_inode_info *ci = ceph_inode(inode); struct ceph_inode_frag *frag, *prev_frag = NULL; struct rb_node *rb_node; - int i; - u32 id, nsplits; + unsigned i, split_by, nsplits; + u32 id; bool update = false; mutex_lock(&ci->i_fragtree_mutex); nsplits = le32_to_cpu(fragtree->nsplits); - if (nsplits) { + if (nsplits != ci->i_fragtree_nsplits) { + update = true; + } else if (nsplits) { i = prandom_u32() % nsplits; id = le32_to_cpu(fragtree->splits[i].frag); if (!__ceph_find_frag(ci, id)) @@ -360,6 +362,13 @@ static int ceph_fill_fragtree(struct inode *inode, rb_node = rb_first(&ci->i_fragtree); for (i = 0; i < nsplits; i++) { id = le32_to_cpu(fragtree->splits[i].frag); + split_by = le32_to_cpu(fragtree->splits[i].by); + if (split_by == 0 || ceph_frag_bits(id) + split_by > 24) { + pr_err("fill_fragtree %llx.%llx invalid split %d/%u, " + "frag %x split by %d\n", ceph_vinop(inode), + i, nsplits, id, split_by); + continue; + } frag = NULL; while (rb_node) { frag = rb_entry(rb_node, struct ceph_inode_frag, node); @@ -375,6 +384,8 @@ static int ceph_fill_fragtree(struct inode *inode, if (frag->split_by > 0 || !is_frag_child(frag->frag, prev_frag)) { rb_erase(&frag->node, &ci->i_fragtree); + if (frag->split_by > 0) + ci->i_fragtree_nsplits--; kfree(frag); } frag = NULL; @@ -384,7 +395,9 @@ static int ceph_fill_fragtree(struct inode *inode, if (IS_ERR(frag)) continue; } - frag->split_by = le32_to_cpu(fragtree->splits[i].by); + if (frag->split_by == 0) + ci->i_fragtree_nsplits++; + frag->split_by = split_by; dout(" frag %x split by %d\n", frag->frag, frag->split_by); prev_frag = frag; } @@ -395,6 +408,8 @@ static int ceph_fill_fragtree(struct inode *inode, if (frag->split_by > 0 || !is_frag_child(frag->frag, prev_frag)) { rb_erase(&frag->node, &ci->i_fragtree); + if (frag->split_by > 0) + ci->i_fragtree_nsplits--; kfree(frag); } } @@ -546,6 +561,7 @@ void ceph_destroy_inode(struct inode *inode) rb_erase(n, &ci->i_fragtree); kfree(frag); } + ci->i_fragtree_nsplits = 0; __ceph_destroy_xattrs(ci); if (ci->i_xattrs.blob) diff --git a/fs/ceph/super.h b/fs/ceph/super.h index c9b671dfff81..a268f18d2119 100644 --- a/fs/ceph/super.h +++ b/fs/ceph/super.h @@ -297,6 +297,7 @@ struct ceph_inode_info { u64 i_files, i_subdirs; struct rb_root i_fragtree; + int i_fragtree_nsplits; struct mutex i_fragtree_mutex; struct ceph_inode_xattrs_info i_xattrs; -- cgit v1.3-8-gc7d7 From 224a7542b8fdde3cc7c600f8b0870c5541a9f678 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Thu, 5 May 2016 16:40:17 +0800 Subject: ceph: tolerate bad i_size for symlink inode A mds bug can cause symlink's size to be truncated to zero. Signed-off-by: Yan, Zheng --- fs/ceph/inode.c | 22 +++++++++++++++------- 1 file changed, 15 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c index 7ba8e1c2557b..89d08155986d 100644 --- a/fs/ceph/inode.c +++ b/fs/ceph/inode.c @@ -582,6 +582,11 @@ int ceph_drop_inode(struct inode *inode) return 1; } +static inline blkcnt_t calc_inode_blocks(u64 size) +{ + return (size + (1<<9) - 1) >> 9; +} + /* * Helpers to fill in size, ctime, mtime, and atime. We have to be * careful because either the client or MDS may have more up to date @@ -604,7 +609,7 @@ int ceph_fill_file_size(struct inode *inode, int issued, size = 0; } i_size_write(inode, size); - inode->i_blocks = (size + (1<<9) - 1) >> 9; + inode->i_blocks = calc_inode_blocks(size); ci->i_reported_size = size; if (truncate_seq != ci->i_truncate_seq) { dout("truncate_seq %u -> %u\n", @@ -863,9 +868,13 @@ static int fill_inode(struct inode *inode, struct page *locked_page, spin_unlock(&ci->i_ceph_lock); - err = -EINVAL; - if (WARN_ON(symlen != i_size_read(inode))) - goto out; + if (symlen != i_size_read(inode)) { + pr_err("fill_inode %llx.%llx BAD symlink " + "size %lld\n", ceph_vinop(inode), + i_size_read(inode)); + i_size_write(inode, symlen); + inode->i_blocks = calc_inode_blocks(symlen); + } err = -ENOMEM; sym = kstrndup(iinfo->symlink, symlen, GFP_NOFS); @@ -1629,7 +1638,7 @@ int ceph_inode_set_size(struct inode *inode, loff_t size) spin_lock(&ci->i_ceph_lock); dout("set_size %p %llu -> %llu\n", inode, inode->i_size, size); i_size_write(inode, size); - inode->i_blocks = (size + (1 << 9) - 1) >> 9; + inode->i_blocks = calc_inode_blocks(size); /* tell the MDS if we are approaching max_size */ if ((size << 1) >= ci->i_max_size && @@ -2002,8 +2011,7 @@ int ceph_setattr(struct dentry *dentry, struct iattr *attr) if ((issued & CEPH_CAP_FILE_EXCL) && attr->ia_size > inode->i_size) { i_size_write(inode, attr->ia_size); - inode->i_blocks = - (attr->ia_size + (1 << 9) - 1) >> 9; + inode->i_blocks = calc_inode_blocks(attr->ia_size); inode->i_ctime = attr->ia_ctime; ci->i_reported_size = attr->ia_size; dirtied |= CEPH_CAP_FILE_EXCL; -- cgit v1.3-8-gc7d7 From 3b33f692c84c28cc8178aaeeb9264d82b48787f1 Mon Sep 17 00:00:00 2001 From: Zhang Zhuoyu Date: Fri, 25 Mar 2016 05:18:39 -0400 Subject: ceph: make logical calculation functions return bool This patch makes serverl logical caculation functions return bool to improve readability due to these particular functions only using 0/1 as their return value. No functional change. Signed-off-by: Zhang Zhuoyu --- fs/ceph/cache.c | 2 +- fs/ceph/dir.c | 2 +- include/linux/ceph/ceph_frag.h | 4 ++-- include/linux/ceph/decode.h | 2 +- include/linux/ceph/osdmap.h | 6 +++--- net/ceph/ceph_common.c | 2 +- 6 files changed, 9 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/ceph/cache.c b/fs/ceph/cache.c index a351480dbabc..c052b5bf219b 100644 --- a/fs/ceph/cache.c +++ b/fs/ceph/cache.c @@ -236,7 +236,7 @@ static void ceph_vfs_readpage_complete_unlock(struct page *page, void *data, int unlock_page(page); } -static inline int cache_valid(struct ceph_inode_info *ci) +static inline bool cache_valid(struct ceph_inode_info *ci) { return ((ceph_caps_issued(ci) & CEPH_CAP_FILE_CACHE) && (ci->i_fscache_gen == ci->i_rdcache_gen)); diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c index 4850c3624a87..f6279a1bd6ec 100644 --- a/fs/ceph/dir.c +++ b/fs/ceph/dir.c @@ -710,7 +710,7 @@ struct dentry *ceph_finish_lookup(struct ceph_mds_request *req, return dentry; } -static int is_root_ceph_dentry(struct inode *inode, struct dentry *dentry) +static bool is_root_ceph_dentry(struct inode *inode, struct dentry *dentry) { return ceph_ino(inode) == CEPH_INO_ROOT && strncmp(dentry->d_name.name, ".ceph", 5) == 0; diff --git a/include/linux/ceph/ceph_frag.h b/include/linux/ceph/ceph_frag.h index b827e066e55a..146507df8650 100644 --- a/include/linux/ceph/ceph_frag.h +++ b/include/linux/ceph/ceph_frag.h @@ -51,11 +51,11 @@ static inline __u32 ceph_frag_make_child(__u32 f, int by, int i) return ceph_frag_make(newbits, ceph_frag_value(f) | (i << (24 - newbits))); } -static inline int ceph_frag_is_leftmost(__u32 f) +static inline bool ceph_frag_is_leftmost(__u32 f) { return ceph_frag_value(f) == 0; } -static inline int ceph_frag_is_rightmost(__u32 f) +static inline bool ceph_frag_is_rightmost(__u32 f) { return ceph_frag_value(f) == ceph_frag_mask(f); } diff --git a/include/linux/ceph/decode.h b/include/linux/ceph/decode.h index a6ef9cc267ec..19e9932f3e77 100644 --- a/include/linux/ceph/decode.h +++ b/include/linux/ceph/decode.h @@ -47,7 +47,7 @@ static inline void ceph_decode_copy(void **p, void *pv, size_t n) /* * bounds check input. */ -static inline int ceph_has_room(void **p, void *end, size_t n) +static inline bool ceph_has_room(void **p, void *end, size_t n) { return end >= *p && n <= end - *p; } diff --git a/include/linux/ceph/osdmap.h b/include/linux/ceph/osdmap.h index 821e16fff39a..ddc426b22d81 100644 --- a/include/linux/ceph/osdmap.h +++ b/include/linux/ceph/osdmap.h @@ -172,19 +172,19 @@ struct ceph_osdmap { int crush_scratch_ary[CEPH_PG_MAX_SIZE * 3]; }; -static inline int ceph_osd_exists(struct ceph_osdmap *map, int osd) +static inline bool ceph_osd_exists(struct ceph_osdmap *map, int osd) { return osd >= 0 && osd < map->max_osd && (map->osd_state[osd] & CEPH_OSD_EXISTS); } -static inline int ceph_osd_is_up(struct ceph_osdmap *map, int osd) +static inline bool ceph_osd_is_up(struct ceph_osdmap *map, int osd) { return ceph_osd_exists(map, osd) && (map->osd_state[osd] & CEPH_OSD_UP); } -static inline int ceph_osd_is_down(struct ceph_osdmap *map, int osd) +static inline bool ceph_osd_is_down(struct ceph_osdmap *map, int osd) { return !ceph_osd_is_up(map, osd); } diff --git a/net/ceph/ceph_common.c b/net/ceph/ceph_common.c index dcc18c6f7cf9..55d2bfee16d7 100644 --- a/net/ceph/ceph_common.c +++ b/net/ceph/ceph_common.c @@ -651,7 +651,7 @@ EXPORT_SYMBOL(ceph_destroy_client); /* * true if we have the mon map (and have thus joined the cluster) */ -static int have_mon_and_osd_map(struct ceph_client *client) +static bool have_mon_and_osd_map(struct ceph_client *client) { return client->monc.monmap && client->monc.monmap->epoch && client->osdc.osdmap && client->osdc.osdmap->epoch; -- cgit v1.3-8-gc7d7 From 4f7e89f6ace0f6cd2f20110efd2d405e26bcbf31 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Tue, 10 May 2016 18:40:28 +0800 Subject: ceph: block non-fatal signals for fault/page_mkwrite Fault and page_mkwrite are supposed to be uninterruptable. But they call ceph functions that are interruptible. So they should block signals before calling functions that are interruptible Signed-off-by: Yan, Zheng --- fs/ceph/addr.c | 66 ++++++++++++++++++++++++++++++++++------------------------ 1 file changed, 39 insertions(+), 27 deletions(-) (limited to 'fs') diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c index d52e3bcfda7c..3e204b9ff9f7 100644 --- a/fs/ceph/addr.c +++ b/fs/ceph/addr.c @@ -1315,6 +1315,17 @@ const struct address_space_operations ceph_aops = { .direct_IO = ceph_direct_io, }; +static void ceph_block_sigs(sigset_t *oldset) +{ + sigset_t mask; + siginitsetinv(&mask, sigmask(SIGKILL)); + sigprocmask(SIG_BLOCK, &mask, oldset); +} + +static void ceph_restore_sigs(sigset_t *oldset) +{ + sigprocmask(SIG_SETMASK, oldset, NULL); +} /* * vm ops @@ -1327,6 +1338,9 @@ static int ceph_filemap_fault(struct vm_area_struct *vma, struct vm_fault *vmf) struct page *pinned_page = NULL; loff_t off = vmf->pgoff << PAGE_SHIFT; int want, got, ret; + sigset_t oldset; + + ceph_block_sigs(&oldset); dout("filemap_fault %p %llx.%llx %llu~%zd trying to get caps\n", inode, ceph_vinop(inode), off, (size_t)PAGE_SIZE); @@ -1334,16 +1348,12 @@ static int ceph_filemap_fault(struct vm_area_struct *vma, struct vm_fault *vmf) want = CEPH_CAP_FILE_CACHE | CEPH_CAP_FILE_LAZYIO; else want = CEPH_CAP_FILE_CACHE; - while (1) { - got = 0; - ret = ceph_get_caps(ci, CEPH_CAP_FILE_RD, want, - -1, &got, &pinned_page); - if (ret == 0) - break; - if (ret != -ERESTARTSYS) { - WARN_ON(1); - return VM_FAULT_SIGBUS; - } + + got = 0; + ret = ceph_get_caps(ci, CEPH_CAP_FILE_RD, want, -1, &got, &pinned_page); + if (ret < 0) { + ret = VM_FAULT_SIGBUS; + goto out_restore; } dout("filemap_fault %p %llu~%zd got cap refs on %s\n", inode, off, (size_t)PAGE_SIZE, ceph_cap_string(got)); @@ -1361,7 +1371,7 @@ static int ceph_filemap_fault(struct vm_area_struct *vma, struct vm_fault *vmf) ceph_put_cap_refs(ci, got); if (ret != -EAGAIN) - return ret; + goto out_restore; /* read inline data */ if (off >= PAGE_SIZE) { @@ -1375,7 +1385,7 @@ static int ceph_filemap_fault(struct vm_area_struct *vma, struct vm_fault *vmf) ~__GFP_FS)); if (!page) { ret = VM_FAULT_OOM; - goto out; + goto out_inline; } ret1 = __ceph_do_getattr(inode, page, CEPH_STAT_CAP_INLINE_DATA, true); @@ -1383,7 +1393,7 @@ static int ceph_filemap_fault(struct vm_area_struct *vma, struct vm_fault *vmf) unlock_page(page); put_page(page); ret = VM_FAULT_SIGBUS; - goto out; + goto out_inline; } if (ret1 < PAGE_SIZE) zero_user_segment(page, ret1, PAGE_SIZE); @@ -1392,10 +1402,12 @@ static int ceph_filemap_fault(struct vm_area_struct *vma, struct vm_fault *vmf) SetPageUptodate(page); vmf->page = page; ret = VM_FAULT_MAJOR | VM_FAULT_LOCKED; +out_inline: + dout("filemap_fault %p %llu~%zd read inline data ret %d\n", + inode, off, (size_t)PAGE_SIZE, ret); } -out: - dout("filemap_fault %p %llu~%zd read inline data ret %d\n", - inode, off, (size_t)PAGE_SIZE, ret); +out_restore: + ceph_restore_sigs(&oldset); return ret; } @@ -1413,11 +1425,14 @@ static int ceph_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) loff_t size = i_size_read(inode); size_t len; int want, got, ret; + sigset_t oldset; prealloc_cf = ceph_alloc_cap_flush(); if (!prealloc_cf) return VM_FAULT_SIGBUS; + ceph_block_sigs(&oldset); + if (ci->i_inline_version != CEPH_INLINE_NONE) { struct page *locked_page = NULL; if (off == 0) { @@ -1444,17 +1459,13 @@ static int ceph_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) want = CEPH_CAP_FILE_BUFFER | CEPH_CAP_FILE_LAZYIO; else want = CEPH_CAP_FILE_BUFFER; - while (1) { - got = 0; - ret = ceph_get_caps(ci, CEPH_CAP_FILE_WR, want, off + len, - &got, NULL); - if (ret == 0) - break; - if (ret != -ERESTARTSYS) { - WARN_ON(1); - ret = VM_FAULT_SIGBUS; - goto out_free; - } + + got = 0; + ret = ceph_get_caps(ci, CEPH_CAP_FILE_WR, want, off + len, + &got, NULL); + if (ret < 0) { + ret = VM_FAULT_SIGBUS; + goto out_free; } dout("page_mkwrite %p %llu~%zd got cap refs on %s\n", inode, off, len, ceph_cap_string(got)); @@ -1499,6 +1510,7 @@ out: inode, off, len, ceph_cap_string(got), ret); ceph_put_cap_refs(ci, got); out_free: + ceph_restore_sigs(&oldset); ceph_free_cap_flush(prealloc_cf); return ret; -- cgit v1.3-8-gc7d7 From 6ce026e411c4f36c9e51189d28a5dd9d08095b9d Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Tue, 10 May 2016 18:59:13 +0800 Subject: ceph: make fault/page_mkwrite return VM_FAULT_OOM for -ENOMEM Signed-off-by: Yan, Zheng --- fs/ceph/addr.c | 37 +++++++++++++++++-------------------- 1 file changed, 17 insertions(+), 20 deletions(-) (limited to 'fs') diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c index 3e204b9ff9f7..d128bb65746d 100644 --- a/fs/ceph/addr.c +++ b/fs/ceph/addr.c @@ -1351,10 +1351,9 @@ static int ceph_filemap_fault(struct vm_area_struct *vma, struct vm_fault *vmf) got = 0; ret = ceph_get_caps(ci, CEPH_CAP_FILE_RD, want, -1, &got, &pinned_page); - if (ret < 0) { - ret = VM_FAULT_SIGBUS; + if (ret < 0) goto out_restore; - } + dout("filemap_fault %p %llu~%zd got cap refs on %s\n", inode, off, (size_t)PAGE_SIZE, ceph_cap_string(got)); @@ -1392,7 +1391,10 @@ static int ceph_filemap_fault(struct vm_area_struct *vma, struct vm_fault *vmf) if (ret1 < 0 || off >= i_size_read(inode)) { unlock_page(page); put_page(page); - ret = VM_FAULT_SIGBUS; + if (ret1 < 0) + ret = ret1; + else + ret = VM_FAULT_SIGBUS; goto out_inline; } if (ret1 < PAGE_SIZE) @@ -1408,6 +1410,9 @@ out_inline: } out_restore: ceph_restore_sigs(&oldset); + if (ret < 0) + ret = (ret == -ENOMEM) ? VM_FAULT_OOM : VM_FAULT_SIGBUS; + return ret; } @@ -1429,7 +1434,7 @@ static int ceph_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) prealloc_cf = ceph_alloc_cap_flush(); if (!prealloc_cf) - return VM_FAULT_SIGBUS; + return VM_FAULT_OOM; ceph_block_sigs(&oldset); @@ -1442,10 +1447,8 @@ static int ceph_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) ret = ceph_uninline_data(vma->vm_file, locked_page); if (locked_page) unlock_page(locked_page); - if (ret < 0) { - ret = VM_FAULT_SIGBUS; + if (ret < 0) goto out_free; - } } if (off + PAGE_SIZE <= size) @@ -1463,10 +1466,9 @@ static int ceph_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) got = 0; ret = ceph_get_caps(ci, CEPH_CAP_FILE_WR, want, off + len, &got, NULL); - if (ret < 0) { - ret = VM_FAULT_SIGBUS; + if (ret < 0) goto out_free; - } + dout("page_mkwrite %p %llu~%zd got cap refs on %s\n", inode, off, len, ceph_cap_string(got)); @@ -1475,10 +1477,9 @@ static int ceph_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) lock_page(page); - ret = VM_FAULT_NOPAGE; - if ((off > size) || - (page->mapping != inode->i_mapping)) { + if ((off > size) || (page->mapping != inode->i_mapping)) { unlock_page(page); + ret = VM_FAULT_NOPAGE; goto out; } @@ -1487,11 +1488,6 @@ static int ceph_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) /* success. we'll keep the page locked. */ set_page_dirty(page); ret = VM_FAULT_LOCKED; - } else { - if (ret == -ENOMEM) - ret = VM_FAULT_OOM; - else - ret = VM_FAULT_SIGBUS; } out: if (ret == VM_FAULT_LOCKED || @@ -1512,7 +1508,8 @@ out: out_free: ceph_restore_sigs(&oldset); ceph_free_cap_flush(prealloc_cf); - + if (ret < 0) + ret = (ret == -ENOMEM) ? VM_FAULT_OOM : VM_FAULT_SIGBUS; return ret; } -- cgit v1.3-8-gc7d7 From f0b33df57a5f03c637f75ead7cb4d978c59cc63d Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Tue, 10 May 2016 19:09:06 +0800 Subject: ceph: handle -EAGAIN returned by ceph_update_writeable_page() when ceph_update_writeable_page() return -EAGAIN, caller should lock the page and call ceph_update_writeable_page() again. Signed-off-by: Yan, Zheng --- fs/ceph/addr.c | 28 +++++++++++++++------------- 1 file changed, 15 insertions(+), 13 deletions(-) (limited to 'fs') diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c index d128bb65746d..97ee5d1fbb61 100644 --- a/fs/ceph/addr.c +++ b/fs/ceph/addr.c @@ -1475,21 +1475,23 @@ static int ceph_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) /* Update time before taking page lock */ file_update_time(vma->vm_file); - lock_page(page); + do { + lock_page(page); - if ((off > size) || (page->mapping != inode->i_mapping)) { - unlock_page(page); - ret = VM_FAULT_NOPAGE; - goto out; - } + if ((off > size) || (page->mapping != inode->i_mapping)) { + unlock_page(page); + ret = VM_FAULT_NOPAGE; + break; + } + + ret = ceph_update_writeable_page(vma->vm_file, off, len, page); + if (ret >= 0) { + /* success. we'll keep the page locked. */ + set_page_dirty(page); + ret = VM_FAULT_LOCKED; + } + } while (ret == -EAGAIN); - ret = ceph_update_writeable_page(vma->vm_file, off, len, page); - if (ret >= 0) { - /* success. we'll keep the page locked. */ - set_page_dirty(page); - ret = VM_FAULT_LOCKED; - } -out: if (ret == VM_FAULT_LOCKED || ci->i_inline_version != CEPH_INLINE_NONE) { int dirty; -- cgit v1.3-8-gc7d7 From a78bbd4b29c29784f0addb5e3b35790c7ed178ae Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Fri, 13 May 2016 11:30:24 +0800 Subject: ceph: make ceph_update_writeable_page() uninterruptible ceph_update_writeable_page() is used by ceph_write_begin(). It beaks atomicity of write operation if it's interruptible. Signed-off-by: Yan, Zheng --- fs/ceph/addr.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c index 97ee5d1fbb61..4aa8e375e648 100644 --- a/fs/ceph/addr.c +++ b/fs/ceph/addr.c @@ -1168,7 +1168,7 @@ retry_locked: snapc = ceph_get_snap_context(snapc); unlock_page(page); ceph_queue_writeback(inode); - r = wait_event_interruptible(ci->i_cap_wq, + r = wait_event_killable(ci->i_cap_wq, context_is_writeable_or_written(inode, snapc)); ceph_put_snap_context(snapc); if (r == -ERESTARTSYS) -- cgit v1.3-8-gc7d7 From ad15ec06e51b6eb73981428109e32c75cbad7d3d Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Fri, 13 May 2016 17:29:51 +0800 Subject: ceph: handle interrupted ceph_writepage() writepage() can be interrupted when it's called by direct memory reclaimer (the direct memory relaimer is killed). To avoid lossing data, we redirty the page. Signed-off-by: Yan, Zheng --- fs/ceph/addr.c | 22 ++++++++++++++++++---- 1 file changed, 18 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c index 4aa8e375e648..080a9cab3ee1 100644 --- a/fs/ceph/addr.c +++ b/fs/ceph/addr.c @@ -544,11 +544,21 @@ static int writepage_nounlock(struct page *page, struct writeback_control *wbc) truncate_seq, truncate_size, &inode->i_mtime, &page, 1); if (err < 0) { - dout("writepage setting page/mapping error %d %p\n", err, page); + struct writeback_control tmp_wbc; + if (!wbc) + wbc = &tmp_wbc; + if (err == -ERESTARTSYS) { + /* killed by SIGKILL */ + dout("writepage interrupted page %p\n", page); + redirty_page_for_writepage(wbc, page); + end_page_writeback(page); + goto out; + } + dout("writepage setting page/mapping error %d %p\n", + err, page); SetPageError(page); mapping_set_error(&inode->i_data, err); - if (wbc) - wbc->pages_skipped++; + wbc->pages_skipped++; } else { dout("writepage cleaned page %p\n", page); err = 0; /* vfs expects us to return 0 */ @@ -569,12 +579,16 @@ static int ceph_writepage(struct page *page, struct writeback_control *wbc) BUG_ON(!inode); ihold(inode); err = writepage_nounlock(page, wbc); + if (err == -ERESTARTSYS) { + /* direct memory reclaimer was killed by SIGKILL. return 0 + * to prevent caller from setting mapping/page error */ + err = 0; + } unlock_page(page); iput(inode); return err; } - /* * lame release_pages helper. release_pages() isn't exported to * modules. -- cgit v1.3-8-gc7d7 From b109eec6f4332bd517e2f41e207037c4b9065094 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Fri, 13 May 2016 17:54:17 +0800 Subject: ceph: SetPageError() for writeback pages if writepages fails Signed-off-by: Yan, Zheng --- fs/ceph/addr.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c index 080a9cab3ee1..8fa1f91cbf16 100644 --- a/fs/ceph/addr.c +++ b/fs/ceph/addr.c @@ -626,7 +626,6 @@ static void writepages_finish(struct ceph_osd_request *req) struct ceph_fs_client *fsc = ceph_inode_to_client(inode); bool remove_page; - dout("writepages_finish %p rc %d\n", inode, rc); if (rc < 0) mapping_set_error(mapping, rc); @@ -661,6 +660,9 @@ static void writepages_finish(struct ceph_osd_request *req) clear_bdi_congested(&fsc->backing_dev_info, BLK_RW_ASYNC); + if (rc < 0) + SetPageError(page); + ceph_put_snap_context(page_snap_context(page)); page->private = 0; ClearPagePrivate(page); -- cgit v1.3-8-gc7d7 From 9abd4db713704aac146395e079224ddd716e9b95 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Wed, 18 May 2016 20:58:26 +0800 Subject: ceph: don't use truncate_pagecache() to invalidate read cache truncate_pagecache() drops dirty pages, it's dangerous to use it to invalidate read cache. Besides, we shouldn't start invalidating read cache while there are buffer writers. Because buffer writers may add dirty pages later. Signed-off-by: Yan, Zheng --- fs/ceph/caps.c | 8 ++++---- fs/ceph/inode.c | 4 +++- 2 files changed, 7 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index fab93c66d879..c17b5d76d75e 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -1656,7 +1656,7 @@ retry_locked: */ if ((!is_delayed || mdsc->stopping) && !S_ISDIR(inode->i_mode) && /* ignore readdir cache */ - ci->i_wrbuffer_ref == 0 && /* no dirty pages... */ + !(ci->i_wb_ref || ci->i_wrbuffer_ref) && /* no dirty pages... */ inode->i_data.nrpages && /* have cached pages */ (revoking & (CEPH_CAP_FILE_CACHE| CEPH_CAP_FILE_LAZYIO)) && /* or revoking cache */ @@ -1698,8 +1698,8 @@ retry_locked: revoking = cap->implemented & ~cap->issued; dout(" mds%d cap %p used %s issued %s implemented %s revoking %s\n", - cap->mds, cap, ceph_cap_string(cap->issued), - ceph_cap_string(cap_used), + cap->mds, cap, ceph_cap_string(cap_used), + ceph_cap_string(cap->issued), ceph_cap_string(cap->implemented), ceph_cap_string(revoking)); @@ -2828,7 +2828,7 @@ static void handle_cap_grant(struct ceph_mds_client *mdsc, if (!S_ISDIR(inode->i_mode) && /* don't invalidate readdir cache */ ((cap->issued & ~newcaps) & CEPH_CAP_FILE_CACHE) && (newcaps & CEPH_CAP_FILE_LAZYIO) == 0 && - !ci->i_wrbuffer_ref) { + !(ci->i_wrbuffer_ref || ci->i_wb_ref)) { if (try_nonblocking_invalidate(inode)) { /* there were locked pages.. invalidate later in a separate thread. */ diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c index 89d08155986d..07495ba61fe3 100644 --- a/fs/ceph/inode.c +++ b/fs/ceph/inode.c @@ -1728,7 +1728,9 @@ static void ceph_invalidate_work(struct work_struct *work) orig_gen = ci->i_rdcache_gen; spin_unlock(&ci->i_ceph_lock); - truncate_pagecache(inode, 0); + if (invalidate_inode_pages2(inode->i_mapping) < 0) { + pr_err("invalidate_pages %p fails\n", inode); + } spin_lock(&ci->i_ceph_lock); if (orig_gen == ci->i_rdcache_gen && -- cgit v1.3-8-gc7d7 From e536030934aebf049fe6aaebc58dd37aeee21840 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Thu, 19 May 2016 19:15:19 +0800 Subject: ceph: fix wake_up_session_cb() We should reset i_requested_max_size before waking the waiters. (zero i_requested_max_size make waiter re-request the max size) Signed-off-by: Yan, Zheng --- fs/ceph/mds_client.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c index 7ad31283d510..2103b823bec0 100644 --- a/fs/ceph/mds_client.c +++ b/fs/ceph/mds_client.c @@ -1258,13 +1258,13 @@ static int wake_up_session_cb(struct inode *inode, struct ceph_cap *cap, { struct ceph_inode_info *ci = ceph_inode(inode); - wake_up_all(&ci->i_cap_wq); if (arg) { spin_lock(&ci->i_ceph_lock); ci->i_wanted_max_size = 0; ci->i_requested_max_size = 0; spin_unlock(&ci->i_ceph_lock); } + wake_up_all(&ci->i_cap_wq); return 0; } -- cgit v1.3-8-gc7d7 From 002354112f1e3cc7400ef48b853aefb90e801588 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Thu, 26 May 2016 00:05:12 -0400 Subject: restore killability of old mutex_lock_killable(&inode->i_mutex) users The ones that are taking it exclusive, that is... Signed-off-by: Al Viro --- fs/btrfs/ioctl.c | 18 +++++++----------- fs/overlayfs/readdir.c | 4 +--- fs/readdir.c | 12 ++++++------ 3 files changed, 14 insertions(+), 20 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index 4e700694b741..c24473fac2fb 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -846,11 +846,9 @@ static noinline int btrfs_mksubvol(struct path *parent, struct dentry *dentry; int error; - inode_lock_nested(dir, I_MUTEX_PARENT); - // XXX: should've been - // mutex_lock_killable_nested(&dir->i_mutex, I_MUTEX_PARENT); - // if (error == -EINTR) - // return error; + error = down_write_killable_nested(&dir->i_rwsem, I_MUTEX_PARENT); + if (error == -EINTR) + return error; dentry = lookup_one_len(name, parent->dentry, namelen); error = PTR_ERR(dentry); @@ -2377,11 +2375,9 @@ static noinline int btrfs_ioctl_snap_destroy(struct file *file, goto out; - inode_lock_nested(dir, I_MUTEX_PARENT); - // XXX: should've been - // err = mutex_lock_killable_nested(&dir->i_mutex, I_MUTEX_PARENT); - // if (err == -EINTR) - // goto out_drop_write; + err = down_write_killable_nested(&dir->i_rwsem, I_MUTEX_PARENT); + if (err == -EINTR) + goto out_drop_write; dentry = lookup_one_len(vol_args->name, parent, namelen); if (IS_ERR(dentry)) { err = PTR_ERR(dentry); @@ -2571,7 +2567,7 @@ out_dput: dput(dentry); out_unlock_dir: inode_unlock(dir); -//out_drop_write: +out_drop_write: mnt_drop_write_file(file); out: kfree(vol_args); diff --git a/fs/overlayfs/readdir.c b/fs/overlayfs/readdir.c index da186ee4f846..9e79f6458dfb 100644 --- a/fs/overlayfs/readdir.c +++ b/fs/overlayfs/readdir.c @@ -218,9 +218,7 @@ static int ovl_check_whiteouts(struct dentry *dir, struct ovl_readdir_data *rdd) cap_raise(override_cred->cap_effective, CAP_DAC_OVERRIDE); old_cred = override_creds(override_cred); - inode_lock(dir->d_inode); - err = 0; - // XXX: err = mutex_lock_killable(&dir->d_inode->i_mutex); + err = down_write_killable(&dir->d_inode->i_rwsem); if (!err) { while (rdd->first_maybe_whiteout) { p = rdd->first_maybe_whiteout; diff --git a/fs/readdir.c b/fs/readdir.c index 68ef06efe6bc..9d0212c374d6 100644 --- a/fs/readdir.c +++ b/fs/readdir.c @@ -35,13 +35,13 @@ int iterate_dir(struct file *file, struct dir_context *ctx) if (res) goto out; - if (shared) + if (shared) { inode_lock_shared(inode); - else - inode_lock(inode); - // res = mutex_lock_killable(&inode->i_mutex); - // if (res) - // goto out; + } else { + res = down_write_killable(&inode->i_rwsem); + if (res) + goto out; + } res = -ENOENT; if (!IS_DEADDIR(inode)) { -- cgit v1.3-8-gc7d7 From 602c4cd452d93355166daf8a662c84a5f85d2856 Mon Sep 17 00:00:00 2001 From: Tom Haynes Date: Wed, 25 May 2016 07:31:13 -0700 Subject: nfs/flexfiles: Use the layout segment for reading unless it a IOMODE_RW and reading is disabled Signed-off-by: Tom Haynes Signed-off-by: Anna Schumaker --- fs/nfs/flexfilelayout/flexfilelayout.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/nfs/flexfilelayout/flexfilelayout.c b/fs/nfs/flexfilelayout/flexfilelayout.c index 40bccb2c06d1..a7aeb7407252 100644 --- a/fs/nfs/flexfilelayout/flexfilelayout.c +++ b/fs/nfs/flexfilelayout/flexfilelayout.c @@ -830,7 +830,8 @@ ff_layout_pg_init_read(struct nfs_pageio_descriptor *pgio, int ds_idx; /* Use full layout for now */ - if (!pgio->pg_lseg) { + if (!pgio->pg_lseg || ff_layout_avoid_read_on_rw(pgio->pg_lseg)) { + pnfs_put_lseg(pgio->pg_lseg); pgio->pg_lseg = pnfs_update_layout(pgio->pg_inode, req->wb_context, 0, @@ -840,9 +841,9 @@ ff_layout_pg_init_read(struct nfs_pageio_descriptor *pgio, if (IS_ERR(pgio->pg_lseg)) { pgio->pg_error = PTR_ERR(pgio->pg_lseg); pgio->pg_lseg = NULL; - return; } } + /* If no lseg, fall back to read through mds */ if (pgio->pg_lseg == NULL) goto out_mds; -- cgit v1.3-8-gc7d7 From c7d73af2d249f0323f5cdb171a59497ce80011fb Mon Sep 17 00:00:00 2001 From: Tom Haynes Date: Wed, 25 May 2016 07:31:14 -0700 Subject: pnfs: pnfs_update_layout needs to consider if strict iomode checking is on As flexfiles has FF_FLAGS_NO_READ_IO, there is a need to generically support enforcing that a IOMODE_RW segment will not allow READ I/O. Signed-off-by: Tom Haynes Signed-off-by: Anna Schumaker --- fs/nfs/filelayout/filelayout.c | 2 ++ fs/nfs/flexfilelayout/flexfilelayout.c | 49 +++++++++++++++++++++++++--------- fs/nfs/pnfs.c | 34 ++++++++++++++--------- fs/nfs/pnfs.h | 1 + 4 files changed, 61 insertions(+), 25 deletions(-) (limited to 'fs') diff --git a/fs/nfs/filelayout/filelayout.c b/fs/nfs/filelayout/filelayout.c index 3e50057eeadf..aa59757389dc 100644 --- a/fs/nfs/filelayout/filelayout.c +++ b/fs/nfs/filelayout/filelayout.c @@ -890,6 +890,7 @@ filelayout_pg_init_read(struct nfs_pageio_descriptor *pgio, 0, NFS4_MAX_UINT64, IOMODE_READ, + false, GFP_KERNEL); if (IS_ERR(pgio->pg_lseg)) { pgio->pg_error = PTR_ERR(pgio->pg_lseg); @@ -915,6 +916,7 @@ filelayout_pg_init_write(struct nfs_pageio_descriptor *pgio, 0, NFS4_MAX_UINT64, IOMODE_RW, + false, GFP_NOFS); if (IS_ERR(pgio->pg_lseg)) { pgio->pg_error = PTR_ERR(pgio->pg_lseg); diff --git a/fs/nfs/flexfilelayout/flexfilelayout.c b/fs/nfs/flexfilelayout/flexfilelayout.c index a7aeb7407252..0e8018bc9880 100644 --- a/fs/nfs/flexfilelayout/flexfilelayout.c +++ b/fs/nfs/flexfilelayout/flexfilelayout.c @@ -820,6 +820,36 @@ ff_layout_choose_best_ds_for_read(struct pnfs_layout_segment *lseg, return NULL; } +static void +ff_layout_pg_get_read(struct nfs_pageio_descriptor *pgio, + struct nfs_page *req, + bool strict_iomode) +{ +retry_strict: + pnfs_put_lseg(pgio->pg_lseg); + pgio->pg_lseg = pnfs_update_layout(pgio->pg_inode, + req->wb_context, + 0, + NFS4_MAX_UINT64, + IOMODE_READ, + strict_iomode, + GFP_KERNEL); + if (IS_ERR(pgio->pg_lseg)) { + pgio->pg_error = PTR_ERR(pgio->pg_lseg); + pgio->pg_lseg = NULL; + } + + /* If we don't have checking, do get a IOMODE_RW + * segment, and the server wants to avoid READs + * there, then retry! + */ + if (pgio->pg_lseg && !strict_iomode && + ff_layout_avoid_read_on_rw(pgio->pg_lseg)) { + strict_iomode = true; + goto retry_strict; + } +} + static void ff_layout_pg_init_read(struct nfs_pageio_descriptor *pgio, struct nfs_page *req) @@ -830,19 +860,10 @@ ff_layout_pg_init_read(struct nfs_pageio_descriptor *pgio, int ds_idx; /* Use full layout for now */ - if (!pgio->pg_lseg || ff_layout_avoid_read_on_rw(pgio->pg_lseg)) { - pnfs_put_lseg(pgio->pg_lseg); - pgio->pg_lseg = pnfs_update_layout(pgio->pg_inode, - req->wb_context, - 0, - NFS4_MAX_UINT64, - IOMODE_READ, - GFP_KERNEL); - if (IS_ERR(pgio->pg_lseg)) { - pgio->pg_error = PTR_ERR(pgio->pg_lseg); - pgio->pg_lseg = NULL; - } - } + if (!pgio->pg_lseg) + ff_layout_pg_get_read(pgio, req, false); + else if (ff_layout_avoid_read_on_rw(pgio->pg_lseg)) + ff_layout_pg_get_read(pgio, req, true); /* If no lseg, fall back to read through mds */ if (pgio->pg_lseg == NULL) @@ -894,6 +915,7 @@ ff_layout_pg_init_write(struct nfs_pageio_descriptor *pgio, 0, NFS4_MAX_UINT64, IOMODE_RW, + false, GFP_NOFS); if (IS_ERR(pgio->pg_lseg)) { pgio->pg_error = PTR_ERR(pgio->pg_lseg); @@ -952,6 +974,7 @@ ff_layout_pg_get_mirror_count_write(struct nfs_pageio_descriptor *pgio, 0, NFS4_MAX_UINT64, IOMODE_RW, + false, GFP_NOFS); if (IS_ERR(pgio->pg_lseg)) { pgio->pg_error = PTR_ERR(pgio->pg_lseg); diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index 79ae3049608d..0c7e0d45a4de 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -1321,23 +1321,28 @@ out_existing: /* * iomode matching rules: - * iomode lseg match - * ----- ----- ----- - * ANY READ true - * ANY RW true - * RW READ false - * RW RW true - * READ READ true - * READ RW true + * iomode lseg strict match + * iomode + * ----- ----- ------ ----- + * ANY READ N/A true + * ANY RW N/A true + * RW READ N/A false + * RW RW N/A true + * READ READ N/A true + * READ RW true false + * READ RW false true */ static bool pnfs_lseg_range_match(const struct pnfs_layout_range *ls_range, - const struct pnfs_layout_range *range) + const struct pnfs_layout_range *range, + bool strict_iomode) { struct pnfs_layout_range range1; if ((range->iomode == IOMODE_RW && ls_range->iomode != IOMODE_RW) || + (range->iomode != ls_range->iomode && + strict_iomode == true) || !pnfs_lseg_range_intersecting(ls_range, range)) return 0; @@ -1352,7 +1357,8 @@ pnfs_lseg_range_match(const struct pnfs_layout_range *ls_range, */ static struct pnfs_layout_segment * pnfs_find_lseg(struct pnfs_layout_hdr *lo, - struct pnfs_layout_range *range) + struct pnfs_layout_range *range, + bool strict_iomode) { struct pnfs_layout_segment *lseg, *ret = NULL; @@ -1361,7 +1367,8 @@ pnfs_find_lseg(struct pnfs_layout_hdr *lo, list_for_each_entry(lseg, &lo->plh_segs, pls_list) { if (test_bit(NFS_LSEG_VALID, &lseg->pls_flags) && !test_bit(NFS_LSEG_LAYOUTRETURN, &lseg->pls_flags) && - pnfs_lseg_range_match(&lseg->pls_range, range)) { + pnfs_lseg_range_match(&lseg->pls_range, range, + strict_iomode)) { ret = pnfs_get_lseg(lseg); break; } @@ -1478,6 +1485,7 @@ pnfs_update_layout(struct inode *ino, loff_t pos, u64 count, enum pnfs_iomode iomode, + bool strict_iomode, gfp_t gfp_flags) { struct pnfs_layout_range arg = { @@ -1539,7 +1547,7 @@ lookup_again: goto out_unlock; } - lseg = pnfs_find_lseg(lo, &arg); + lseg = pnfs_find_lseg(lo, &arg, strict_iomode); if (lseg) { trace_pnfs_update_layout(ino, pos, count, iomode, lo, lseg, PNFS_UPDATE_LAYOUT_FOUND_CACHED); @@ -1883,6 +1891,7 @@ pnfs_generic_pg_init_read(struct nfs_pageio_descriptor *pgio, struct nfs_page *r req_offset(req), rd_size, IOMODE_READ, + false, GFP_KERNEL); if (IS_ERR(pgio->pg_lseg)) { pgio->pg_error = PTR_ERR(pgio->pg_lseg); @@ -1907,6 +1916,7 @@ pnfs_generic_pg_init_write(struct nfs_pageio_descriptor *pgio, req_offset(req), wb_size, IOMODE_RW, + false, GFP_NOFS); if (IS_ERR(pgio->pg_lseg)) { pgio->pg_error = PTR_ERR(pgio->pg_lseg); diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h index f9f3331bef49..b21bd0bee784 100644 --- a/fs/nfs/pnfs.h +++ b/fs/nfs/pnfs.h @@ -288,6 +288,7 @@ struct pnfs_layout_segment *pnfs_update_layout(struct inode *ino, loff_t pos, u64 count, enum pnfs_iomode iomode, + bool strict_iomode, gfp_t gfp_flags); void pnfs_clear_layoutreturn_waitbit(struct pnfs_layout_hdr *lo); -- cgit v1.3-8-gc7d7 From 56244ef151c3cd11f505020ab0b3f45454363bcc Mon Sep 17 00:00:00 2001 From: Chris Mason Date: Mon, 16 May 2016 09:21:01 -0700 Subject: Btrfs: fix handling of faults from btrfs_copy_from_user When btrfs_copy_from_user isn't able to copy all of the pages, we need to adjust our accounting to reflect the work that was actually done. Commit 2e78c927d79 changed around the decisions a little and we ended up skipping the accounting adjustments some of the time. This commit makes sure that when we don't copy anything at all, we still hop into the adjustments, and switches to release_bytes instead of write_bytes, since write_bytes isn't aligned. The accounting errors led to warnings during btrfs_destroy_inode: [ 70.847532] WARNING: CPU: 10 PID: 514 at fs/btrfs/inode.c:9350 btrfs_destroy_inode+0x2b3/0x2c0 [ 70.847536] Modules linked in: i2c_piix4 virtio_net i2c_core input_leds button led_class serio_raw acpi_cpufreq sch_fq_codel autofs4 virtio_blk [ 70.847538] CPU: 10 PID: 514 Comm: umount Tainted: G W 4.6.0-rc6_00062_g2997da1-dirty #23 [ 70.847539] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.9.0-1.fc24 04/01/2014 [ 70.847542] 0000000000000000 ffff880ff5cafab8 ffffffff8149d5e9 0000000000000202 [ 70.847543] 0000000000000000 0000000000000000 0000000000000000 ffff880ff5cafb08 [ 70.847547] ffffffff8107bdfd ffff880ff5cafaf8 000024868120013d ffff880ff5cafb28 [ 70.847547] Call Trace: [ 70.847550] [] dump_stack+0x51/0x78 [ 70.847551] [] __warn+0xfd/0x120 [ 70.847553] [] warn_slowpath_null+0x1d/0x20 [ 70.847555] [] btrfs_destroy_inode+0x2b3/0x2c0 [ 70.847556] [] ? __destroy_inode+0x71/0x140 [ 70.847558] [] destroy_inode+0x43/0x70 [ 70.847559] [] ? wake_up_bit+0x2f/0x40 [ 70.847560] [] evict+0x148/0x1d0 [ 70.847562] [] ? start_transaction+0x3de/0x460 [ 70.847564] [] dispose_list+0x59/0x80 [ 70.847565] [] evict_inodes+0x180/0x190 [ 70.847566] [] ? __sync_filesystem+0x3f/0x50 [ 70.847568] [] generic_shutdown_super+0x48/0x100 [ 70.847569] [] ? woken_wake_function+0x20/0x20 [ 70.847571] [] kill_anon_super+0x16/0x30 [ 70.847573] [] btrfs_kill_super+0x1e/0x130 [ 70.847574] [] deactivate_locked_super+0x4e/0x90 [ 70.847576] [] deactivate_super+0x51/0x70 [ 70.847577] [] cleanup_mnt+0x3f/0x80 [ 70.847579] [] __cleanup_mnt+0x12/0x20 [ 70.847581] [] task_work_run+0x68/0xa0 [ 70.847582] [] exit_to_usermode_loop+0xd6/0xe0 [ 70.847583] [] do_syscall_64+0xbd/0x170 [ 70.847586] [] entry_SYSCALL64_slow_path+0x25/0x25 This is the test program I used to force short returns from btrfs_copy_from_user void *dontneed(void *arg) { char *p = arg; int ret; while(1) { ret = madvise(p, BUFSIZE/4, MADV_DONTNEED); if (ret) { perror("madvise"); exit(1); } } } int main(int ac, char **av) { int ret; int fd; char *filename; unsigned long offset; char *buf; int i; pthread_t tid; if (ac != 2) { fprintf(stderr, "usage: dammitdave filename\n"); exit(1); } buf = mmap(NULL, BUFSIZE, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0); if (buf == MAP_FAILED) { perror("mmap"); exit(1); } memset(buf, 'a', BUFSIZE); filename = av[1]; ret = pthread_create(&tid, NULL, dontneed, buf); if (ret) { fprintf(stderr, "error %d from pthread_create\n", ret); exit(1); } ret = pthread_detach(tid); if (ret) { fprintf(stderr, "pthread detach failed %d\n", ret); exit(1); } while (1) { fd = open(filename, O_RDWR | O_CREAT, 0600); if (fd < 0) { perror("open"); exit(1); } for (i = 0; i < ROUNDS; i++) { int this_write = BUFSIZE; offset = rand() % MAXSIZE; ret = pwrite(fd, buf, this_write, offset); if (ret < 0) { perror("pwrite"); exit(1); } else if (ret != this_write) { fprintf(stderr, "short write to %s offset %lu ret %d\n", filename, offset, ret); exit(1); } if (i == ROUNDS - 1) { ret = sync_file_range(fd, offset, 4096, SYNC_FILE_RANGE_WRITE); if (ret < 0) { perror("sync_file_range"); exit(1); } } } ret = ftruncate(fd, 0); if (ret < 0) { perror("ftruncate"); exit(1); } ret = close(fd); if (ret) { perror("close"); exit(1); } ret = unlink(filename); if (ret) { perror("unlink"); exit(1); } } return 0; } Signed-off-by: Chris Mason Reported-by: Dave Jones Fixes: 2e78c927d79333f299a8ac81c2fd2952caeef335 cc: stable@vger.kernel.org # v4.6 Signed-off-by: Chris Mason --- fs/btrfs/file.c | 27 +++++++++++++++++---------- 1 file changed, 17 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index 1933e8b07ed7..159a93450e26 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -1596,6 +1596,13 @@ again: copied = btrfs_copy_from_user(pos, write_bytes, pages, i); + num_sectors = BTRFS_BYTES_TO_BLKS(root->fs_info, + reserve_bytes); + dirty_sectors = round_up(copied + sector_offset, + root->sectorsize); + dirty_sectors = BTRFS_BYTES_TO_BLKS(root->fs_info, + dirty_sectors); + /* * if we have trouble faulting in the pages, fall * back to one page at a time @@ -1605,6 +1612,7 @@ again: if (copied == 0) { force_page_uptodate = true; + dirty_sectors = 0; dirty_pages = 0; } else { force_page_uptodate = false; @@ -1615,20 +1623,19 @@ again: /* * If we had a short copy we need to release the excess delaloc * bytes we reserved. We need to increment outstanding_extents - * because btrfs_delalloc_release_space will decrement it, but + * because btrfs_delalloc_release_space and + * btrfs_delalloc_release_metadata will decrement it, but * we still have an outstanding extent for the chunk we actually * managed to copy. */ - num_sectors = BTRFS_BYTES_TO_BLKS(root->fs_info, - reserve_bytes); - dirty_sectors = round_up(copied + sector_offset, - root->sectorsize); - dirty_sectors = BTRFS_BYTES_TO_BLKS(root->fs_info, - dirty_sectors); - if (num_sectors > dirty_sectors) { - release_bytes = (write_bytes - copied) - & ~((u64)root->sectorsize - 1); + /* + * we round down because we don't want to count + * any partial blocks actually sent through the + * IO machines + */ + release_bytes = round_down(release_bytes - copied, + root->sectorsize); if (copied > 0) { spin_lock(&BTRFS_I(inode)->lock); BTRFS_I(inode)->outstanding_extents++; -- cgit v1.3-8-gc7d7 From 1f3a437fa027c75b26a64e2e71d7c374ff70f866 Mon Sep 17 00:00:00 2001 From: Eric Ren Date: Thu, 26 May 2016 15:16:16 -0700 Subject: ocfs2: fix improper handling of return errno Previously, if a bad inode was found in ocfs2_iget(), -ESTALE was returned back to the caller anyway. Since commit d2b9d71a2da7 ("ocfs2: check/fix inode block for online file check") can handle with return value from ocfs2_read_locked_inode() now, we know the exact errno returned for us. Link: http://lkml.kernel.org/r/1463970656-18413-1-git-send-email-zren@suse.com Signed-off-by: Eric Ren Reviewed-by: Joseph Qi Cc: Mark Fasheh Cc: Joel Becker Cc: Junxiao Bi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ocfs2/inode.c | 7 +------ 1 file changed, 1 insertion(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/inode.c b/fs/ocfs2/inode.c index 0748777f2e2a..c56a7679df93 100644 --- a/fs/ocfs2/inode.c +++ b/fs/ocfs2/inode.c @@ -176,12 +176,7 @@ struct inode *ocfs2_iget(struct ocfs2_super *osb, u64 blkno, unsigned flags, } if (is_bad_inode(inode)) { iput(inode); - if ((flags & OCFS2_FI_FLAG_FILECHECK_CHK) || - (flags & OCFS2_FI_FLAG_FILECHECK_FIX)) - /* Return OCFS2_FILECHECK_ERR_XXX related errno */ - inode = ERR_PTR(rc); - else - inode = ERR_PTR(-ESTALE); + inode = ERR_PTR(rc); goto bail; } -- cgit v1.3-8-gc7d7 From 3fe6e52f062643676eb4518d68cee3bc1272091b Mon Sep 17 00:00:00 2001 From: Antonio Murdaca Date: Thu, 7 Apr 2016 15:48:25 +0200 Subject: ovl: override creds with the ones from the superblock mounter In user namespace the whiteout creation fails with -EPERM because the current process isn't capable(CAP_SYS_ADMIN) when setting xattr. A simple reproducer: $ mkdir upper lower work merged lower/dir $ sudo mount -t overlay overlay -olowerdir=lower,upperdir=upper,workdir=work merged $ unshare -m -p -f -U -r bash Now as root in the user namespace: \# touch merged/dir/{1,2,3} # this will force a copy up of lower/dir \# rm -fR merged/* This ends up failing with -EPERM after the files in dir has been correctly deleted: unlinkat(4, "2", 0) = 0 unlinkat(4, "1", 0) = 0 unlinkat(4, "3", 0) = 0 close(4) = 0 unlinkat(AT_FDCWD, "merged/dir", AT_REMOVEDIR) = -1 EPERM (Operation not permitted) Interestingly, if you don't place files in merged/dir you can remove it, meaning if upper/dir does not exist, creating the char device file works properly in that same location. This patch uses ovl_sb_creator_cred() to get the cred struct from the superblock mounter and override the old cred with these new ones so that the whiteout creation is possible because overlay is wrong in assuming that the creds it will get with prepare_creds will be in the initial user namespace. The old cap_raise game is removed in favor of just overriding the old cred struct. This patch also drops from ovl_copy_up_one() the following two lines: override_cred->fsuid = stat->uid; override_cred->fsgid = stat->gid; This is because the correct uid and gid are taken directly with the stat struct and correctly set with ovl_set_attr(). Signed-off-by: Antonio Murdaca Signed-off-by: Miklos Szeredi --- fs/overlayfs/copy_up.c | 26 +------------------ fs/overlayfs/dir.c | 67 ++++-------------------------------------------- fs/overlayfs/overlayfs.h | 1 + fs/overlayfs/readdir.c | 14 +++------- fs/overlayfs/super.c | 18 ++++++++++++- 5 files changed, 27 insertions(+), 99 deletions(-) (limited to 'fs') diff --git a/fs/overlayfs/copy_up.c b/fs/overlayfs/copy_up.c index cc514da6f3e7..80aa6f1eb336 100644 --- a/fs/overlayfs/copy_up.c +++ b/fs/overlayfs/copy_up.c @@ -336,7 +336,6 @@ int ovl_copy_up_one(struct dentry *parent, struct dentry *dentry, struct dentry *upperdir; struct dentry *upperdentry; const struct cred *old_cred; - struct cred *override_cred; char *link = NULL; if (WARN_ON(!workdir)) @@ -357,28 +356,7 @@ int ovl_copy_up_one(struct dentry *parent, struct dentry *dentry, return PTR_ERR(link); } - err = -ENOMEM; - override_cred = prepare_creds(); - if (!override_cred) - goto out_free_link; - - override_cred->fsuid = stat->uid; - override_cred->fsgid = stat->gid; - /* - * CAP_SYS_ADMIN for copying up extended attributes - * CAP_DAC_OVERRIDE for create - * CAP_FOWNER for chmod, timestamp update - * CAP_FSETID for chmod - * CAP_CHOWN for chown - * CAP_MKNOD for mknod - */ - cap_raise(override_cred->cap_effective, CAP_SYS_ADMIN); - cap_raise(override_cred->cap_effective, CAP_DAC_OVERRIDE); - cap_raise(override_cred->cap_effective, CAP_FOWNER); - cap_raise(override_cred->cap_effective, CAP_FSETID); - cap_raise(override_cred->cap_effective, CAP_CHOWN); - cap_raise(override_cred->cap_effective, CAP_MKNOD); - old_cred = override_creds(override_cred); + old_cred = ovl_override_creds(dentry->d_sb); err = -EIO; if (lock_rename(workdir, upperdir) != NULL) { @@ -401,9 +379,7 @@ int ovl_copy_up_one(struct dentry *parent, struct dentry *dentry, out_unlock: unlock_rename(workdir, upperdir); revert_creds(old_cred); - put_cred(override_cred); -out_free_link: if (link) free_page((unsigned long) link); diff --git a/fs/overlayfs/dir.c b/fs/overlayfs/dir.c index b3fc0a35bf62..22f0253a3567 100644 --- a/fs/overlayfs/dir.c +++ b/fs/overlayfs/dir.c @@ -405,28 +405,13 @@ static int ovl_create_or_link(struct dentry *dentry, int mode, dev_t rdev, err = ovl_create_upper(dentry, inode, &stat, link, hardlink); } else { const struct cred *old_cred; - struct cred *override_cred; - err = -ENOMEM; - override_cred = prepare_creds(); - if (!override_cred) - goto out_iput; - - /* - * CAP_SYS_ADMIN for setting opaque xattr - * CAP_DAC_OVERRIDE for create in workdir, rename - * CAP_FOWNER for removing whiteout from sticky dir - */ - cap_raise(override_cred->cap_effective, CAP_SYS_ADMIN); - cap_raise(override_cred->cap_effective, CAP_DAC_OVERRIDE); - cap_raise(override_cred->cap_effective, CAP_FOWNER); - old_cred = override_creds(override_cred); + old_cred = ovl_override_creds(dentry->d_sb); err = ovl_create_over_whiteout(dentry, inode, &stat, link, hardlink); revert_creds(old_cred); - put_cred(override_cred); } if (!err) @@ -662,32 +647,11 @@ static int ovl_do_remove(struct dentry *dentry, bool is_dir) if (OVL_TYPE_PURE_UPPER(type)) { err = ovl_remove_upper(dentry, is_dir); } else { - const struct cred *old_cred; - struct cred *override_cred; - - err = -ENOMEM; - override_cred = prepare_creds(); - if (!override_cred) - goto out_drop_write; - - /* - * CAP_SYS_ADMIN for setting xattr on whiteout, opaque dir - * CAP_DAC_OVERRIDE for create in workdir, rename - * CAP_FOWNER for removing whiteout from sticky dir - * CAP_FSETID for chmod of opaque dir - * CAP_CHOWN for chown of opaque dir - */ - cap_raise(override_cred->cap_effective, CAP_SYS_ADMIN); - cap_raise(override_cred->cap_effective, CAP_DAC_OVERRIDE); - cap_raise(override_cred->cap_effective, CAP_FOWNER); - cap_raise(override_cred->cap_effective, CAP_FSETID); - cap_raise(override_cred->cap_effective, CAP_CHOWN); - old_cred = override_creds(override_cred); + const struct cred *old_cred = ovl_override_creds(dentry->d_sb); err = ovl_remove_and_whiteout(dentry, is_dir); revert_creds(old_cred); - put_cred(override_cred); } out_drop_write: ovl_drop_write(dentry); @@ -725,7 +689,6 @@ static int ovl_rename2(struct inode *olddir, struct dentry *old, bool new_is_dir = false; struct dentry *opaquedir = NULL; const struct cred *old_cred = NULL; - struct cred *override_cred = NULL; err = -EINVAL; if (flags & ~(RENAME_EXCHANGE | RENAME_NOREPLACE)) @@ -794,26 +757,8 @@ static int ovl_rename2(struct inode *olddir, struct dentry *old, old_opaque = !OVL_TYPE_PURE_UPPER(old_type); new_opaque = !OVL_TYPE_PURE_UPPER(new_type); - if (old_opaque || new_opaque) { - err = -ENOMEM; - override_cred = prepare_creds(); - if (!override_cred) - goto out_drop_write; - - /* - * CAP_SYS_ADMIN for setting xattr on whiteout, opaque dir - * CAP_DAC_OVERRIDE for create in workdir - * CAP_FOWNER for removing whiteout from sticky dir - * CAP_FSETID for chmod of opaque dir - * CAP_CHOWN for chown of opaque dir - */ - cap_raise(override_cred->cap_effective, CAP_SYS_ADMIN); - cap_raise(override_cred->cap_effective, CAP_DAC_OVERRIDE); - cap_raise(override_cred->cap_effective, CAP_FOWNER); - cap_raise(override_cred->cap_effective, CAP_FSETID); - cap_raise(override_cred->cap_effective, CAP_CHOWN); - old_cred = override_creds(override_cred); - } + if (old_opaque || new_opaque) + old_cred = ovl_override_creds(old->d_sb); if (overwrite && OVL_TYPE_MERGE_OR_LOWER(new_type) && new_is_dir) { opaquedir = ovl_check_empty_and_clear(new); @@ -943,10 +888,8 @@ out_dput_old: out_unlock: unlock_rename(new_upperdir, old_upperdir); out_revert_creds: - if (old_opaque || new_opaque) { + if (old_opaque || new_opaque) revert_creds(old_cred); - put_cred(override_cred); - } out_drop_write: ovl_drop_write(old); out: diff --git a/fs/overlayfs/overlayfs.h b/fs/overlayfs/overlayfs.h index 6a7090f4a441..4cebeb24c08d 100644 --- a/fs/overlayfs/overlayfs.h +++ b/fs/overlayfs/overlayfs.h @@ -153,6 +153,7 @@ void ovl_drop_write(struct dentry *dentry); bool ovl_dentry_is_opaque(struct dentry *dentry); void ovl_dentry_set_opaque(struct dentry *dentry, bool opaque); bool ovl_is_whiteout(struct dentry *dentry); +const struct cred *ovl_override_creds(struct super_block *sb); void ovl_dentry_update(struct dentry *dentry, struct dentry *upperdentry); struct dentry *ovl_lookup(struct inode *dir, struct dentry *dentry, unsigned int flags); diff --git a/fs/overlayfs/readdir.c b/fs/overlayfs/readdir.c index 6ec1e43a9a54..e9206bc8598f 100644 --- a/fs/overlayfs/readdir.c +++ b/fs/overlayfs/readdir.c @@ -36,6 +36,7 @@ struct ovl_dir_cache { struct ovl_readdir_data { struct dir_context ctx; + struct dentry *dentry; bool is_lowest; struct rb_root root; struct list_head *list; @@ -206,17 +207,8 @@ static int ovl_check_whiteouts(struct dentry *dir, struct ovl_readdir_data *rdd) struct ovl_cache_entry *p; struct dentry *dentry; const struct cred *old_cred; - struct cred *override_cred; - - override_cred = prepare_creds(); - if (!override_cred) - return -ENOMEM; - /* - * CAP_DAC_OVERRIDE for lookup - */ - cap_raise(override_cred->cap_effective, CAP_DAC_OVERRIDE); - old_cred = override_creds(override_cred); + old_cred = ovl_override_creds(rdd->dentry->d_sb); err = mutex_lock_killable(&dir->d_inode->i_mutex); if (!err) { @@ -232,7 +224,6 @@ static int ovl_check_whiteouts(struct dentry *dir, struct ovl_readdir_data *rdd) inode_unlock(dir->d_inode); } revert_creds(old_cred); - put_cred(override_cred); return err; } @@ -288,6 +279,7 @@ static int ovl_dir_read_merged(struct dentry *dentry, struct list_head *list) struct path realpath; struct ovl_readdir_data rdd = { .ctx.actor = ovl_fill_merge, + .dentry = dentry, .list = list, .root = RB_ROOT, .is_lowest = false, diff --git a/fs/overlayfs/super.c b/fs/overlayfs/super.c index 791235e03d17..d659f766ceff 100644 --- a/fs/overlayfs/super.c +++ b/fs/overlayfs/super.c @@ -42,6 +42,8 @@ struct ovl_fs { long lower_namelen; /* pathnames of lower and upper dirs, for show_options */ struct ovl_config config; + /* creds of process who forced instantiation of super block */ + const struct cred *creator_cred; }; struct ovl_dir_cache; @@ -265,6 +267,13 @@ bool ovl_is_whiteout(struct dentry *dentry) return inode && IS_WHITEOUT(inode); } +const struct cred *ovl_override_creds(struct super_block *sb) +{ + struct ovl_fs *ofs = sb->s_fs_info; + + return override_creds(ofs->creator_cred); +} + static bool ovl_is_opaquedir(struct dentry *dentry) { int res; @@ -603,6 +612,7 @@ static void ovl_put_super(struct super_block *sb) kfree(ufs->config.lowerdir); kfree(ufs->config.upperdir); kfree(ufs->config.workdir); + put_cred(ufs->creator_cred); kfree(ufs); } @@ -1108,10 +1118,14 @@ static int ovl_fill_super(struct super_block *sb, void *data, int silent) else sb->s_d_op = &ovl_dentry_operations; + ufs->creator_cred = prepare_creds(); + if (!ufs->creator_cred) + goto out_put_lower_mnt; + err = -ENOMEM; oe = ovl_alloc_entry(numlower); if (!oe) - goto out_put_lower_mnt; + goto out_put_cred; root_dentry = d_make_root(ovl_new_inode(sb, S_IFDIR, oe)); if (!root_dentry) @@ -1144,6 +1158,8 @@ static int ovl_fill_super(struct super_block *sb, void *data, int silent) out_free_oe: kfree(oe); +out_put_cred: + put_cred(ufs->creator_cred); out_put_lower_mnt: for (i = 0; i < ufs->numlower; i++) mntput(ufs->lower_mnt[i]); -- cgit v1.3-8-gc7d7 From 21765194cecf2e4514ad75244df459f188140a0f Mon Sep 17 00:00:00 2001 From: Vivek Goyal Date: Fri, 20 May 2016 09:04:26 -0400 Subject: ovl: Do d_type check only if work dir creation was successful d_type check requires successful creation of workdir as iterates through work dir and expects work dir to be present in it. If that's not the case, this check will always return d_type not supported even if underlying filesystem might be supporting it. So don't do this check if work dir creation failed in previous step. Signed-off-by: Vivek Goyal Signed-off-by: Miklos Szeredi --- fs/overlayfs/super.c | 19 +++++++++++-------- 1 file changed, 11 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/overlayfs/super.c b/fs/overlayfs/super.c index d659f766ceff..781e1d86d2ce 100644 --- a/fs/overlayfs/super.c +++ b/fs/overlayfs/super.c @@ -1074,16 +1074,19 @@ static int ovl_fill_super(struct super_block *sb, void *data, int silent) /* * Upper should support d_type, else whiteouts are visible. * Given workdir and upper are on same fs, we can do - * iterate_dir() on workdir. + * iterate_dir() on workdir. This check requires successful + * creation of workdir in previous step. */ - err = ovl_check_d_type_supported(&workpath); - if (err < 0) - goto out_put_workdir; + if (ufs->workdir) { + err = ovl_check_d_type_supported(&workpath); + if (err < 0) + goto out_put_workdir; - if (!err) { - pr_err("overlayfs: upper fs needs to support d_type.\n"); - err = -EINVAL; - goto out_put_workdir; + if (!err) { + pr_err("overlayfs: upper fs needs to support d_type.\n"); + err = -EINVAL; + goto out_put_workdir; + } } } -- cgit v1.3-8-gc7d7 From 5930122683dff58f0846b0f0405b4bd598a3ba6a Mon Sep 17 00:00:00 2001 From: Al Viro Date: Fri, 27 May 2016 10:19:30 -0400 Subject: switch xattr_handler->set() to passing dentry and inode separately preparation for similar switch in ->setxattr() (see the next commit for rationale). Signed-off-by: Al Viro --- fs/9p/acl.c | 6 +++--- fs/9p/xattr.c | 5 +++-- fs/btrfs/xattr.c | 12 +++++------- fs/ceph/xattr.c | 7 ++++--- fs/cifs/xattr.c | 9 +++++---- fs/ext2/xattr_security.c | 7 ++++--- fs/ext2/xattr_trusted.c | 7 ++++--- fs/ext2/xattr_user.c | 9 +++++---- fs/ext4/xattr_security.c | 7 ++++--- fs/ext4/xattr_trusted.c | 7 ++++--- fs/ext4/xattr_user.c | 9 +++++---- fs/f2fs/xattr.c | 12 ++++++------ fs/gfs2/xattr.c | 6 +++--- fs/hfsplus/xattr.c | 12 ++++++------ fs/hfsplus/xattr.h | 2 +- fs/hfsplus/xattr_security.c | 7 ++++--- fs/hfsplus/xattr_trusted.c | 7 ++++--- fs/hfsplus/xattr_user.c | 7 ++++--- fs/jffs2/security.c | 7 ++++--- fs/jffs2/xattr_trusted.c | 7 ++++--- fs/jffs2/xattr_user.c | 7 ++++--- fs/jfs/xattr.c | 14 ++++++-------- fs/nfs/nfs4proc.c | 19 +++++++++---------- fs/ocfs2/xattr.c | 23 +++++++++++++---------- fs/orangefs/xattr.c | 10 ++++++---- fs/posix_acl.c | 6 +++--- fs/reiserfs/xattr_security.c | 9 +++++---- fs/reiserfs/xattr_trusted.c | 9 +++++---- fs/reiserfs/xattr_user.c | 9 +++++---- fs/ubifs/xattr.c | 7 +++---- fs/xattr.c | 6 ++++-- fs/xfs/xfs_xattr.c | 9 +++++---- include/linux/xattr.h | 4 ++-- mm/shmem.c | 7 ++++--- 34 files changed, 156 insertions(+), 135 deletions(-) (limited to 'fs') diff --git a/fs/9p/acl.c b/fs/9p/acl.c index eb3589edf485..0576eaeb60b9 100644 --- a/fs/9p/acl.c +++ b/fs/9p/acl.c @@ -239,13 +239,13 @@ static int v9fs_xattr_get_acl(const struct xattr_handler *handler, } static int v9fs_xattr_set_acl(const struct xattr_handler *handler, - struct dentry *dentry, const char *name, - const void *value, size_t size, int flags) + struct dentry *dentry, struct inode *inode, + const char *name, const void *value, + size_t size, int flags) { int retval; struct posix_acl *acl; struct v9fs_session_info *v9ses; - struct inode *inode = d_inode(dentry); v9ses = v9fs_dentry2v9ses(dentry); /* diff --git a/fs/9p/xattr.c b/fs/9p/xattr.c index 18c62bae9591..a6bd349bab23 100644 --- a/fs/9p/xattr.c +++ b/fs/9p/xattr.c @@ -147,8 +147,9 @@ static int v9fs_xattr_handler_get(const struct xattr_handler *handler, } static int v9fs_xattr_handler_set(const struct xattr_handler *handler, - struct dentry *dentry, const char *name, - const void *value, size_t size, int flags) + struct dentry *dentry, struct inode *inode, + const char *name, const void *value, + size_t size, int flags) { const char *full_name = xattr_full_name(handler, name); diff --git a/fs/btrfs/xattr.c b/fs/btrfs/xattr.c index 3bfb252206c7..d1a177a3dbe8 100644 --- a/fs/btrfs/xattr.c +++ b/fs/btrfs/xattr.c @@ -380,23 +380,21 @@ static int btrfs_xattr_handler_get(const struct xattr_handler *handler, } static int btrfs_xattr_handler_set(const struct xattr_handler *handler, - struct dentry *dentry, const char *name, - const void *buffer, size_t size, - int flags) + struct dentry *unused, struct inode *inode, + const char *name, const void *buffer, + size_t size, int flags) { - struct inode *inode = d_inode(dentry); - name = xattr_full_name(handler, name); return __btrfs_setxattr(NULL, inode, name, buffer, size, flags); } static int btrfs_xattr_handler_set_prop(const struct xattr_handler *handler, - struct dentry *dentry, + struct dentry *unused, struct inode *inode, const char *name, const void *value, size_t size, int flags) { name = xattr_full_name(handler, name); - return btrfs_set_prop(d_inode(dentry), name, value, size, flags); + return btrfs_set_prop(inode, name, value, size, flags); } static const struct xattr_handler btrfs_security_xattr_handler = { diff --git a/fs/ceph/xattr.c b/fs/ceph/xattr.c index 0d66722c6a52..2baa6939dfe6 100644 --- a/fs/ceph/xattr.c +++ b/fs/ceph/xattr.c @@ -1051,12 +1051,13 @@ static int ceph_get_xattr_handler(const struct xattr_handler *handler, } static int ceph_set_xattr_handler(const struct xattr_handler *handler, - struct dentry *dentry, const char *name, - const void *value, size_t size, int flags) + struct dentry *unused, struct inode *inode, + const char *name, const void *value, + size_t size, int flags) { if (!ceph_is_valid_xattr(name)) return -EOPNOTSUPP; - return __ceph_setxattr(d_inode(dentry), name, value, size, flags); + return __ceph_setxattr(inode, name, value, size, flags); } const struct xattr_handler ceph_other_xattr_handler = { diff --git a/fs/cifs/xattr.c b/fs/cifs/xattr.c index c8b77aa24a1d..5e23f64c0804 100644 --- a/fs/cifs/xattr.c +++ b/fs/cifs/xattr.c @@ -39,8 +39,9 @@ enum { XATTR_USER, XATTR_CIFS_ACL, XATTR_ACL_ACCESS, XATTR_ACL_DEFAULT }; static int cifs_xattr_set(const struct xattr_handler *handler, - struct dentry *dentry, const char *name, - const void *value, size_t size, int flags) + struct dentry *dentry, struct inode *inode, + const char *name, const void *value, + size_t size, int flags) { int rc = -EOPNOTSUPP; unsigned int xid; @@ -99,12 +100,12 @@ static int cifs_xattr_set(const struct xattr_handler *handler, if (value && pTcon->ses->server->ops->set_acl) rc = pTcon->ses->server->ops->set_acl(pacl, - size, d_inode(dentry), + size, inode, full_path, CIFS_ACL_DACL); else rc = -EOPNOTSUPP; if (rc == 0) /* force revalidate of the inode */ - CIFS_I(d_inode(dentry))->time = 0; + CIFS_I(inode)->time = 0; kfree(pacl); } #endif /* CONFIG_CIFS_ACL */ diff --git a/fs/ext2/xattr_security.c b/fs/ext2/xattr_security.c index 7fd3b867ce65..7b9e9c1842d5 100644 --- a/fs/ext2/xattr_security.c +++ b/fs/ext2/xattr_security.c @@ -18,10 +18,11 @@ ext2_xattr_security_get(const struct xattr_handler *handler, static int ext2_xattr_security_set(const struct xattr_handler *handler, - struct dentry *dentry, const char *name, - const void *value, size_t size, int flags) + struct dentry *unused, struct inode *inode, + const char *name, const void *value, + size_t size, int flags) { - return ext2_xattr_set(d_inode(dentry), EXT2_XATTR_INDEX_SECURITY, name, + return ext2_xattr_set(inode, EXT2_XATTR_INDEX_SECURITY, name, value, size, flags); } diff --git a/fs/ext2/xattr_trusted.c b/fs/ext2/xattr_trusted.c index 0f85705ff519..65049b71af13 100644 --- a/fs/ext2/xattr_trusted.c +++ b/fs/ext2/xattr_trusted.c @@ -25,10 +25,11 @@ ext2_xattr_trusted_get(const struct xattr_handler *handler, static int ext2_xattr_trusted_set(const struct xattr_handler *handler, - struct dentry *dentry, const char *name, - const void *value, size_t size, int flags) + struct dentry *unused, struct inode *inode, + const char *name, const void *value, + size_t size, int flags) { - return ext2_xattr_set(d_inode(dentry), EXT2_XATTR_INDEX_TRUSTED, name, + return ext2_xattr_set(inode, EXT2_XATTR_INDEX_TRUSTED, name, value, size, flags); } diff --git a/fs/ext2/xattr_user.c b/fs/ext2/xattr_user.c index 1fafd27037cc..fb2f992ae763 100644 --- a/fs/ext2/xattr_user.c +++ b/fs/ext2/xattr_user.c @@ -29,13 +29,14 @@ ext2_xattr_user_get(const struct xattr_handler *handler, static int ext2_xattr_user_set(const struct xattr_handler *handler, - struct dentry *dentry, const char *name, - const void *value, size_t size, int flags) + struct dentry *unused, struct inode *inode, + const char *name, const void *value, + size_t size, int flags) { - if (!test_opt(dentry->d_sb, XATTR_USER)) + if (!test_opt(inode->i_sb, XATTR_USER)) return -EOPNOTSUPP; - return ext2_xattr_set(d_inode(dentry), EXT2_XATTR_INDEX_USER, + return ext2_xattr_set(inode, EXT2_XATTR_INDEX_USER, name, value, size, flags); } diff --git a/fs/ext4/xattr_security.c b/fs/ext4/xattr_security.c index 123a7d010efe..a8921112030d 100644 --- a/fs/ext4/xattr_security.c +++ b/fs/ext4/xattr_security.c @@ -22,10 +22,11 @@ ext4_xattr_security_get(const struct xattr_handler *handler, static int ext4_xattr_security_set(const struct xattr_handler *handler, - struct dentry *dentry, const char *name, - const void *value, size_t size, int flags) + struct dentry *unused, struct inode *inode, + const char *name, const void *value, + size_t size, int flags) { - return ext4_xattr_set(d_inode(dentry), EXT4_XATTR_INDEX_SECURITY, + return ext4_xattr_set(inode, EXT4_XATTR_INDEX_SECURITY, name, value, size, flags); } diff --git a/fs/ext4/xattr_trusted.c b/fs/ext4/xattr_trusted.c index 60652fa24cbc..c7765c735714 100644 --- a/fs/ext4/xattr_trusted.c +++ b/fs/ext4/xattr_trusted.c @@ -29,10 +29,11 @@ ext4_xattr_trusted_get(const struct xattr_handler *handler, static int ext4_xattr_trusted_set(const struct xattr_handler *handler, - struct dentry *dentry, const char *name, - const void *value, size_t size, int flags) + struct dentry *unused, struct inode *inode, + const char *name, const void *value, + size_t size, int flags) { - return ext4_xattr_set(d_inode(dentry), EXT4_XATTR_INDEX_TRUSTED, + return ext4_xattr_set(inode, EXT4_XATTR_INDEX_TRUSTED, name, value, size, flags); } diff --git a/fs/ext4/xattr_user.c b/fs/ext4/xattr_user.c index 17a446ffecd3..ca20e423034b 100644 --- a/fs/ext4/xattr_user.c +++ b/fs/ext4/xattr_user.c @@ -30,12 +30,13 @@ ext4_xattr_user_get(const struct xattr_handler *handler, static int ext4_xattr_user_set(const struct xattr_handler *handler, - struct dentry *dentry, const char *name, - const void *value, size_t size, int flags) + struct dentry *unused, struct inode *inode, + const char *name, const void *value, + size_t size, int flags) { - if (!test_opt(dentry->d_sb, XATTR_USER)) + if (!test_opt(inode->i_sb, XATTR_USER)) return -EOPNOTSUPP; - return ext4_xattr_set(d_inode(dentry), EXT4_XATTR_INDEX_USER, + return ext4_xattr_set(inode, EXT4_XATTR_INDEX_USER, name, value, size, flags); } diff --git a/fs/f2fs/xattr.c b/fs/f2fs/xattr.c index 00ea56797258..e3decae3acfb 100644 --- a/fs/f2fs/xattr.c +++ b/fs/f2fs/xattr.c @@ -50,10 +50,11 @@ static int f2fs_xattr_generic_get(const struct xattr_handler *handler, } static int f2fs_xattr_generic_set(const struct xattr_handler *handler, - struct dentry *dentry, const char *name, const void *value, + struct dentry *unused, struct inode *inode, + const char *name, const void *value, size_t size, int flags) { - struct f2fs_sb_info *sbi = F2FS_SB(dentry->d_sb); + struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb); switch (handler->flags) { case F2FS_XATTR_INDEX_USER: @@ -69,7 +70,7 @@ static int f2fs_xattr_generic_set(const struct xattr_handler *handler, default: return -EINVAL; } - return f2fs_setxattr(d_inode(dentry), handler->flags, name, + return f2fs_setxattr(inode, handler->flags, name, value, size, NULL, flags); } @@ -95,11 +96,10 @@ static int f2fs_xattr_advise_get(const struct xattr_handler *handler, } static int f2fs_xattr_advise_set(const struct xattr_handler *handler, - struct dentry *dentry, const char *name, const void *value, + struct dentry *unused, struct inode *inode, + const char *name, const void *value, size_t size, int flags) { - struct inode *inode = d_inode(dentry); - if (!inode_owner_or_capable(inode)) return -EPERM; if (value == NULL) diff --git a/fs/gfs2/xattr.c b/fs/gfs2/xattr.c index f42ab53bd30d..3a2853504084 100644 --- a/fs/gfs2/xattr.c +++ b/fs/gfs2/xattr.c @@ -1251,10 +1251,10 @@ int __gfs2_xattr_set(struct inode *inode, const char *name, } static int gfs2_xattr_set(const struct xattr_handler *handler, - struct dentry *dentry, const char *name, - const void *value, size_t size, int flags) + struct dentry *unused, struct inode *inode, + const char *name, const void *value, + size_t size, int flags) { - struct inode *inode = d_inode(dentry); struct gfs2_inode *ip = GFS2_I(inode); struct gfs2_holder gh; int ret; diff --git a/fs/hfsplus/xattr.c b/fs/hfsplus/xattr.c index 4f118d282a7a..d37bb88dc746 100644 --- a/fs/hfsplus/xattr.c +++ b/fs/hfsplus/xattr.c @@ -424,7 +424,7 @@ static int copy_name(char *buffer, const char *xattr_name, int name_len) return len; } -int hfsplus_setxattr(struct dentry *dentry, const char *name, +int hfsplus_setxattr(struct inode *inode, const char *name, const void *value, size_t size, int flags, const char *prefix, size_t prefixlen) { @@ -437,8 +437,7 @@ int hfsplus_setxattr(struct dentry *dentry, const char *name, return -ENOMEM; strcpy(xattr_name, prefix); strcpy(xattr_name + prefixlen, name); - res = __hfsplus_setxattr(d_inode(dentry), xattr_name, value, size, - flags); + res = __hfsplus_setxattr(inode, xattr_name, value, size, flags); kfree(xattr_name); return res; } @@ -864,8 +863,9 @@ static int hfsplus_osx_getxattr(const struct xattr_handler *handler, } static int hfsplus_osx_setxattr(const struct xattr_handler *handler, - struct dentry *dentry, const char *name, - const void *buffer, size_t size, int flags) + struct dentry *unused, struct inode *inode, + const char *name, const void *buffer, + size_t size, int flags) { /* * Don't allow setting properly prefixed attributes @@ -880,7 +880,7 @@ static int hfsplus_osx_setxattr(const struct xattr_handler *handler, * creates), so we pass the name through unmodified (after * ensuring it doesn't conflict with another namespace). */ - return __hfsplus_setxattr(d_inode(dentry), name, buffer, size, flags); + return __hfsplus_setxattr(inode, name, buffer, size, flags); } const struct xattr_handler hfsplus_xattr_osx_handler = { diff --git a/fs/hfsplus/xattr.h b/fs/hfsplus/xattr.h index d04ba6f58df2..68f6b539371f 100644 --- a/fs/hfsplus/xattr.h +++ b/fs/hfsplus/xattr.h @@ -21,7 +21,7 @@ extern const struct xattr_handler *hfsplus_xattr_handlers[]; int __hfsplus_setxattr(struct inode *inode, const char *name, const void *value, size_t size, int flags); -int hfsplus_setxattr(struct dentry *dentry, const char *name, +int hfsplus_setxattr(struct inode *inode, const char *name, const void *value, size_t size, int flags, const char *prefix, size_t prefixlen); diff --git a/fs/hfsplus/xattr_security.c b/fs/hfsplus/xattr_security.c index ae2ca8c2e335..37b3efa733ef 100644 --- a/fs/hfsplus/xattr_security.c +++ b/fs/hfsplus/xattr_security.c @@ -23,10 +23,11 @@ static int hfsplus_security_getxattr(const struct xattr_handler *handler, } static int hfsplus_security_setxattr(const struct xattr_handler *handler, - struct dentry *dentry, const char *name, - const void *buffer, size_t size, int flags) + struct dentry *unused, struct inode *inode, + const char *name, const void *buffer, + size_t size, int flags) { - return hfsplus_setxattr(dentry, name, buffer, size, flags, + return hfsplus_setxattr(inode, name, buffer, size, flags, XATTR_SECURITY_PREFIX, XATTR_SECURITY_PREFIX_LEN); } diff --git a/fs/hfsplus/xattr_trusted.c b/fs/hfsplus/xattr_trusted.c index eae2947060aa..94519d6c627d 100644 --- a/fs/hfsplus/xattr_trusted.c +++ b/fs/hfsplus/xattr_trusted.c @@ -21,10 +21,11 @@ static int hfsplus_trusted_getxattr(const struct xattr_handler *handler, } static int hfsplus_trusted_setxattr(const struct xattr_handler *handler, - struct dentry *dentry, const char *name, - const void *buffer, size_t size, int flags) + struct dentry *unused, struct inode *inode, + const char *name, const void *buffer, + size_t size, int flags) { - return hfsplus_setxattr(dentry, name, buffer, size, flags, + return hfsplus_setxattr(inode, name, buffer, size, flags, XATTR_TRUSTED_PREFIX, XATTR_TRUSTED_PREFIX_LEN); } diff --git a/fs/hfsplus/xattr_user.c b/fs/hfsplus/xattr_user.c index 3c9eec3e4c7b..fae6c0ea0030 100644 --- a/fs/hfsplus/xattr_user.c +++ b/fs/hfsplus/xattr_user.c @@ -21,10 +21,11 @@ static int hfsplus_user_getxattr(const struct xattr_handler *handler, } static int hfsplus_user_setxattr(const struct xattr_handler *handler, - struct dentry *dentry, const char *name, - const void *buffer, size_t size, int flags) + struct dentry *unused, struct inode *inode, + const char *name, const void *buffer, + size_t size, int flags) { - return hfsplus_setxattr(dentry, name, buffer, size, flags, + return hfsplus_setxattr(inode, name, buffer, size, flags, XATTR_USER_PREFIX, XATTR_USER_PREFIX_LEN); } diff --git a/fs/jffs2/security.c b/fs/jffs2/security.c index 3ed9a4b49778..c2332e30f218 100644 --- a/fs/jffs2/security.c +++ b/fs/jffs2/security.c @@ -57,10 +57,11 @@ static int jffs2_security_getxattr(const struct xattr_handler *handler, } static int jffs2_security_setxattr(const struct xattr_handler *handler, - struct dentry *dentry, const char *name, - const void *buffer, size_t size, int flags) + struct dentry *unused, struct inode *inode, + const char *name, const void *buffer, + size_t size, int flags) { - return do_jffs2_setxattr(d_inode(dentry), JFFS2_XPREFIX_SECURITY, + return do_jffs2_setxattr(inode, JFFS2_XPREFIX_SECURITY, name, buffer, size, flags); } diff --git a/fs/jffs2/xattr_trusted.c b/fs/jffs2/xattr_trusted.c index 4ebecff1d922..5d6030826c52 100644 --- a/fs/jffs2/xattr_trusted.c +++ b/fs/jffs2/xattr_trusted.c @@ -25,10 +25,11 @@ static int jffs2_trusted_getxattr(const struct xattr_handler *handler, } static int jffs2_trusted_setxattr(const struct xattr_handler *handler, - struct dentry *dentry, const char *name, - const void *buffer, size_t size, int flags) + struct dentry *unused, struct inode *inode, + const char *name, const void *buffer, + size_t size, int flags) { - return do_jffs2_setxattr(d_inode(dentry), JFFS2_XPREFIX_TRUSTED, + return do_jffs2_setxattr(inode, JFFS2_XPREFIX_TRUSTED, name, buffer, size, flags); } diff --git a/fs/jffs2/xattr_user.c b/fs/jffs2/xattr_user.c index bce249e1b277..9d027b4abcf9 100644 --- a/fs/jffs2/xattr_user.c +++ b/fs/jffs2/xattr_user.c @@ -25,10 +25,11 @@ static int jffs2_user_getxattr(const struct xattr_handler *handler, } static int jffs2_user_setxattr(const struct xattr_handler *handler, - struct dentry *dentry, const char *name, - const void *buffer, size_t size, int flags) + struct dentry *unused, struct inode *inode, + const char *name, const void *buffer, + size_t size, int flags) { - return do_jffs2_setxattr(d_inode(dentry), JFFS2_XPREFIX_USER, + return do_jffs2_setxattr(inode, JFFS2_XPREFIX_USER, name, buffer, size, flags); } diff --git a/fs/jfs/xattr.c b/fs/jfs/xattr.c index beb182b503b3..0bf3c33aedff 100644 --- a/fs/jfs/xattr.c +++ b/fs/jfs/xattr.c @@ -943,11 +943,10 @@ static int jfs_xattr_get(const struct xattr_handler *handler, } static int jfs_xattr_set(const struct xattr_handler *handler, - struct dentry *dentry, const char *name, - const void *value, size_t size, int flags) + struct dentry *unused, struct inode *inode, + const char *name, const void *value, + size_t size, int flags) { - struct inode *inode = d_inode(dentry); - name = xattr_full_name(handler, name); return __jfs_xattr_set(inode, name, value, size, flags); } @@ -962,11 +961,10 @@ static int jfs_xattr_get_os2(const struct xattr_handler *handler, } static int jfs_xattr_set_os2(const struct xattr_handler *handler, - struct dentry *dentry, const char *name, - const void *value, size_t size, int flags) + struct dentry *unused, struct inode *inode, + const char *name, const void *value, + size_t size, int flags) { - struct inode *inode = d_inode(dentry); - if (is_known_namespace(name)) return -EOPNOTSUPP; return __jfs_xattr_set(inode, name, value, size, flags); diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 084e8570da18..2e802ec47b8a 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -4993,12 +4993,11 @@ static int nfs4_do_set_security_label(struct inode *inode, } static int -nfs4_set_security_label(struct dentry *dentry, const void *buf, size_t buflen) +nfs4_set_security_label(struct inode *inode, const void *buf, size_t buflen) { struct nfs4_label ilabel, *olabel = NULL; struct nfs_fattr fattr; struct rpc_cred *cred; - struct inode *inode = d_inode(dentry); int status; if (!nfs_server_capable(inode, NFS_CAP_SECURITY_LABEL)) @@ -6255,11 +6254,11 @@ nfs4_release_lockowner(struct nfs_server *server, struct nfs4_lock_state *lsp) #define XATTR_NAME_NFSV4_ACL "system.nfs4_acl" static int nfs4_xattr_set_nfs4_acl(const struct xattr_handler *handler, - struct dentry *dentry, const char *key, - const void *buf, size_t buflen, - int flags) + struct dentry *unused, struct inode *inode, + const char *key, const void *buf, + size_t buflen, int flags) { - return nfs4_proc_set_acl(d_inode(dentry), buf, buflen); + return nfs4_proc_set_acl(inode, buf, buflen); } static int nfs4_xattr_get_nfs4_acl(const struct xattr_handler *handler, @@ -6277,12 +6276,12 @@ static bool nfs4_xattr_list_nfs4_acl(struct dentry *dentry) #ifdef CONFIG_NFS_V4_SECURITY_LABEL static int nfs4_xattr_set_nfs4_label(const struct xattr_handler *handler, - struct dentry *dentry, const char *key, - const void *buf, size_t buflen, - int flags) + struct dentry *unused, struct inode *inode, + const char *key, const void *buf, + size_t buflen, int flags) { if (security_ismaclabel(key)) - return nfs4_set_security_label(dentry, buf, buflen); + return nfs4_set_security_label(inode, buf, buflen); return -EOPNOTSUPP; } diff --git a/fs/ocfs2/xattr.c b/fs/ocfs2/xattr.c index ad16995c9e7a..d2053853951e 100644 --- a/fs/ocfs2/xattr.c +++ b/fs/ocfs2/xattr.c @@ -7254,10 +7254,11 @@ static int ocfs2_xattr_security_get(const struct xattr_handler *handler, } static int ocfs2_xattr_security_set(const struct xattr_handler *handler, - struct dentry *dentry, const char *name, - const void *value, size_t size, int flags) + struct dentry *unused, struct inode *inode, + const char *name, const void *value, + size_t size, int flags) { - return ocfs2_xattr_set(d_inode(dentry), OCFS2_XATTR_INDEX_SECURITY, + return ocfs2_xattr_set(inode, OCFS2_XATTR_INDEX_SECURITY, name, value, size, flags); } @@ -7325,10 +7326,11 @@ static int ocfs2_xattr_trusted_get(const struct xattr_handler *handler, } static int ocfs2_xattr_trusted_set(const struct xattr_handler *handler, - struct dentry *dentry, const char *name, - const void *value, size_t size, int flags) + struct dentry *unused, struct inode *inode, + const char *name, const void *value, + size_t size, int flags) { - return ocfs2_xattr_set(d_inode(dentry), OCFS2_XATTR_INDEX_TRUSTED, + return ocfs2_xattr_set(inode, OCFS2_XATTR_INDEX_TRUSTED, name, value, size, flags); } @@ -7354,15 +7356,16 @@ static int ocfs2_xattr_user_get(const struct xattr_handler *handler, } static int ocfs2_xattr_user_set(const struct xattr_handler *handler, - struct dentry *dentry, const char *name, - const void *value, size_t size, int flags) + struct dentry *unused, struct inode *inode, + const char *name, const void *value, + size_t size, int flags) { - struct ocfs2_super *osb = OCFS2_SB(dentry->d_sb); + struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); if (osb->s_mount_opt & OCFS2_MOUNT_NOUSERXATTR) return -EOPNOTSUPP; - return ocfs2_xattr_set(d_inode(dentry), OCFS2_XATTR_INDEX_USER, + return ocfs2_xattr_set(inode, OCFS2_XATTR_INDEX_USER, name, value, size, flags); } diff --git a/fs/orangefs/xattr.c b/fs/orangefs/xattr.c index 99c19545752c..5893ddde0e4b 100644 --- a/fs/orangefs/xattr.c +++ b/fs/orangefs/xattr.c @@ -448,13 +448,14 @@ out_unlock: } static int orangefs_xattr_set_default(const struct xattr_handler *handler, - struct dentry *dentry, + struct dentry *unused, + struct inode *inode, const char *name, const void *buffer, size_t size, int flags) { - return orangefs_inode_setxattr(dentry->d_inode, + return orangefs_inode_setxattr(inode, ORANGEFS_XATTR_NAME_DEFAULT_PREFIX, name, buffer, @@ -478,13 +479,14 @@ static int orangefs_xattr_get_default(const struct xattr_handler *handler, } static int orangefs_xattr_set_trusted(const struct xattr_handler *handler, - struct dentry *dentry, + struct dentry *unused, + struct inode *inode, const char *name, const void *buffer, size_t size, int flags) { - return orangefs_inode_setxattr(dentry->d_inode, + return orangefs_inode_setxattr(inode, ORANGEFS_XATTR_NAME_TRUSTED_PREFIX, name, buffer, diff --git a/fs/posix_acl.c b/fs/posix_acl.c index 2c60f17e7d92..8a4a266beff3 100644 --- a/fs/posix_acl.c +++ b/fs/posix_acl.c @@ -822,10 +822,10 @@ posix_acl_xattr_get(const struct xattr_handler *handler, static int posix_acl_xattr_set(const struct xattr_handler *handler, - struct dentry *dentry, const char *name, - const void *value, size_t size, int flags) + struct dentry *unused, struct inode *inode, + const char *name, const void *value, + size_t size, int flags) { - struct inode *inode = d_backing_inode(dentry); struct posix_acl *acl = NULL; int ret; diff --git a/fs/reiserfs/xattr_security.c b/fs/reiserfs/xattr_security.c index 86aeb9dd805a..e4cbb7719906 100644 --- a/fs/reiserfs/xattr_security.c +++ b/fs/reiserfs/xattr_security.c @@ -20,13 +20,14 @@ security_get(const struct xattr_handler *handler, struct dentry *unused, } static int -security_set(const struct xattr_handler *handler, struct dentry *dentry, - const char *name, const void *buffer, size_t size, int flags) +security_set(const struct xattr_handler *handler, struct dentry *unused, + struct inode *inode, const char *name, const void *buffer, + size_t size, int flags) { - if (IS_PRIVATE(d_inode(dentry))) + if (IS_PRIVATE(inode)) return -EPERM; - return reiserfs_xattr_set(d_inode(dentry), + return reiserfs_xattr_set(inode, xattr_full_name(handler, name), buffer, size, flags); } diff --git a/fs/reiserfs/xattr_trusted.c b/fs/reiserfs/xattr_trusted.c index 31837f031f59..f15a5f9e84ce 100644 --- a/fs/reiserfs/xattr_trusted.c +++ b/fs/reiserfs/xattr_trusted.c @@ -19,13 +19,14 @@ trusted_get(const struct xattr_handler *handler, struct dentry *unused, } static int -trusted_set(const struct xattr_handler *handler, struct dentry *dentry, - const char *name, const void *buffer, size_t size, int flags) +trusted_set(const struct xattr_handler *handler, struct dentry *unused, + struct inode *inode, const char *name, const void *buffer, + size_t size, int flags) { - if (!capable(CAP_SYS_ADMIN) || IS_PRIVATE(d_inode(dentry))) + if (!capable(CAP_SYS_ADMIN) || IS_PRIVATE(inode)) return -EPERM; - return reiserfs_xattr_set(d_inode(dentry), + return reiserfs_xattr_set(inode, xattr_full_name(handler, name), buffer, size, flags); } diff --git a/fs/reiserfs/xattr_user.c b/fs/reiserfs/xattr_user.c index f7c39731684b..dc59df43b2db 100644 --- a/fs/reiserfs/xattr_user.c +++ b/fs/reiserfs/xattr_user.c @@ -17,12 +17,13 @@ user_get(const struct xattr_handler *handler, struct dentry *unused, } static int -user_set(const struct xattr_handler *handler, struct dentry *dentry, - const char *name, const void *buffer, size_t size, int flags) +user_set(const struct xattr_handler *handler, struct dentry *unused, + struct inode *inode, const char *name, const void *buffer, + size_t size, int flags) { - if (!reiserfs_xattrs_user(dentry->d_sb)) + if (!reiserfs_xattrs_user(inode->i_sb)) return -EOPNOTSUPP; - return reiserfs_xattr_set(d_inode(dentry), + return reiserfs_xattr_set(inode, xattr_full_name(handler, name), buffer, size, flags); } diff --git a/fs/ubifs/xattr.c b/fs/ubifs/xattr.c index 6c277eb6aef9..b5fc27969e9d 100644 --- a/fs/ubifs/xattr.c +++ b/fs/ubifs/xattr.c @@ -579,11 +579,10 @@ static int ubifs_xattr_get(const struct xattr_handler *handler, } static int ubifs_xattr_set(const struct xattr_handler *handler, - struct dentry *dentry, const char *name, - const void *value, size_t size, int flags) + struct dentry *dentry, struct inode *inode, + const char *name, const void *value, + size_t size, int flags) { - struct inode *inode = d_inode(dentry); - dbg_gen("xattr '%s', host ino %lu ('%pd'), size %zd", name, inode->i_ino, dentry, size); diff --git a/fs/xattr.c b/fs/xattr.c index fc81e771488a..b16d07889700 100644 --- a/fs/xattr.c +++ b/fs/xattr.c @@ -754,7 +754,8 @@ generic_setxattr(struct dentry *dentry, const char *name, const void *value, siz handler = xattr_resolve_name(dentry->d_sb->s_xattr, &name); if (IS_ERR(handler)) return PTR_ERR(handler); - return handler->set(handler, dentry, name, value, size, flags); + return handler->set(handler, dentry, d_inode(dentry), name, value, + size, flags); } /* @@ -769,7 +770,8 @@ generic_removexattr(struct dentry *dentry, const char *name) handler = xattr_resolve_name(dentry->d_sb->s_xattr, &name); if (IS_ERR(handler)) return PTR_ERR(handler); - return handler->set(handler, dentry, name, NULL, 0, XATTR_REPLACE); + return handler->set(handler, dentry, d_inode(dentry), name, NULL, + 0, XATTR_REPLACE); } EXPORT_SYMBOL(generic_getxattr); diff --git a/fs/xfs/xfs_xattr.c b/fs/xfs/xfs_xattr.c index d111f691f313..2773b155cb56 100644 --- a/fs/xfs/xfs_xattr.c +++ b/fs/xfs/xfs_xattr.c @@ -74,11 +74,12 @@ xfs_forget_acl( } static int -xfs_xattr_set(const struct xattr_handler *handler, struct dentry *dentry, - const char *name, const void *value, size_t size, int flags) +xfs_xattr_set(const struct xattr_handler *handler, struct dentry *unused, + struct inode *inode, const char *name, const void *value, + size_t size, int flags) { int xflags = handler->flags; - struct xfs_inode *ip = XFS_I(d_inode(dentry)); + struct xfs_inode *ip = XFS_I(inode); int error; /* Convert Linux syscall to XFS internal ATTR flags */ @@ -92,7 +93,7 @@ xfs_xattr_set(const struct xattr_handler *handler, struct dentry *dentry, error = xfs_attr_set(ip, (unsigned char *)name, (void *)value, size, xflags); if (!error) - xfs_forget_acl(d_inode(dentry), name, xflags); + xfs_forget_acl(inode, name, xflags); return error; } diff --git a/include/linux/xattr.h b/include/linux/xattr.h index 1cc4c578deb9..76beb206741a 100644 --- a/include/linux/xattr.h +++ b/include/linux/xattr.h @@ -33,8 +33,8 @@ struct xattr_handler { struct inode *inode, const char *name, void *buffer, size_t size); int (*set)(const struct xattr_handler *, struct dentry *dentry, - const char *name, const void *buffer, size_t size, - int flags); + struct inode *inode, const char *name, const void *buffer, + size_t size, int flags); }; const char *xattr_full_name(const struct xattr_handler *, const char *); diff --git a/mm/shmem.c b/mm/shmem.c index e418a995427d..a36144909b28 100644 --- a/mm/shmem.c +++ b/mm/shmem.c @@ -2645,10 +2645,11 @@ static int shmem_xattr_handler_get(const struct xattr_handler *handler, } static int shmem_xattr_handler_set(const struct xattr_handler *handler, - struct dentry *dentry, const char *name, - const void *value, size_t size, int flags) + struct dentry *unused, struct inode *inode, + const char *name, const void *value, + size_t size, int flags) { - struct shmem_inode_info *info = SHMEM_I(d_inode(dentry)); + struct shmem_inode_info *info = SHMEM_I(inode); name = xattr_full_name(handler, name); return simple_xattr_set(&info->xattrs, name, value, size, flags); -- cgit v1.3-8-gc7d7 From e0cbb79805083b8862182341ebf72266d58f6d12 Mon Sep 17 00:00:00 2001 From: Junxiao Bi Date: Fri, 27 May 2016 14:26:58 -0700 Subject: ocfs2: o2hb: add negotiate timer This series of patches is to fix the issue that when storage down, all nodes will fence self due to write timeout. With this patch set, all nodes will keep going until storage back online, except if the following issue happens, then all nodes will do as before to fence self. 1. io error got 2. network between nodes down 3. nodes panic This patch (of 6): When storage down, all nodes will fence self due to write timeout. The negotiate timer is designed to avoid this, with it node will wait until storage up again. Negotiate timer working in the following way: 1. The timer expires before write timeout timer, its timeout is half of write timeout now. It is re-queued along with write timeout timer. If expires, it will send NEGO_TIMEOUT message to master node(node with lowest node number). This message does nothing but marks a bit in a bitmap recording which nodes are negotiating timeout on master node. 2. If storage down, nodes will send this message to master node, then when master node finds its bitmap including all online nodes, it sends NEGO_APPROVL message to all nodes one by one, this message will re-queue write timeout timer and negotiate timer. For any node doesn't receive this message or meets some issue when handling this message, it will be fenced. If storage up at any time, o2hb_thread will run and re-queue all the timer, nothing will be affected by these two steps. Signed-off-by: Junxiao Bi Reviewed-by: Ryan Ding Reviewed-by: Mark Fasheh Cc: Gang He Cc: rwxybh Cc: Joel Becker Cc: Joseph Qi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ocfs2/cluster/heartbeat.c | 51 ++++++++++++++++++++++++++++++++++++++++---- 1 file changed, 47 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/cluster/heartbeat.c b/fs/ocfs2/cluster/heartbeat.c index a8d15beee5cb..750c950f4e1f 100644 --- a/fs/ocfs2/cluster/heartbeat.c +++ b/fs/ocfs2/cluster/heartbeat.c @@ -272,6 +272,10 @@ struct o2hb_region { struct delayed_work hr_write_timeout_work; unsigned long hr_last_timeout_start; + /* negotiate timer, used to negotiate extending hb timeout. */ + struct delayed_work hr_nego_timeout_work; + unsigned long hr_nego_node_bitmap[BITS_TO_LONGS(O2NM_MAX_NODES)]; + /* Used during o2hb_check_slot to hold a copy of the block * being checked because we temporarily have to zero out the * crc field. */ @@ -319,7 +323,7 @@ static void o2hb_write_timeout(struct work_struct *work) o2quo_disk_timeout(); } -static void o2hb_arm_write_timeout(struct o2hb_region *reg) +static void o2hb_arm_timeout(struct o2hb_region *reg) { /* Arm writeout only after thread reaches steady state */ if (atomic_read(®->hr_steady_iterations) != 0) @@ -337,11 +341,49 @@ static void o2hb_arm_write_timeout(struct o2hb_region *reg) reg->hr_last_timeout_start = jiffies; schedule_delayed_work(®->hr_write_timeout_work, msecs_to_jiffies(O2HB_MAX_WRITE_TIMEOUT_MS)); + + cancel_delayed_work(®->hr_nego_timeout_work); + /* negotiate timeout must be less than write timeout. */ + schedule_delayed_work(®->hr_nego_timeout_work, + msecs_to_jiffies(O2HB_MAX_WRITE_TIMEOUT_MS)/2); + memset(reg->hr_nego_node_bitmap, 0, sizeof(reg->hr_nego_node_bitmap)); } -static void o2hb_disarm_write_timeout(struct o2hb_region *reg) +static void o2hb_disarm_timeout(struct o2hb_region *reg) { cancel_delayed_work_sync(®->hr_write_timeout_work); + cancel_delayed_work_sync(®->hr_nego_timeout_work); +} + +static void o2hb_nego_timeout(struct work_struct *work) +{ + unsigned long live_node_bitmap[BITS_TO_LONGS(O2NM_MAX_NODES)]; + int master_node; + struct o2hb_region *reg; + + reg = container_of(work, struct o2hb_region, hr_nego_timeout_work.work); + o2hb_fill_node_map(live_node_bitmap, sizeof(live_node_bitmap)); + /* lowest node as master node to make negotiate decision. */ + master_node = find_next_bit(live_node_bitmap, O2NM_MAX_NODES, 0); + + if (master_node == o2nm_this_node()) { + set_bit(master_node, reg->hr_nego_node_bitmap); + if (memcmp(reg->hr_nego_node_bitmap, live_node_bitmap, + sizeof(reg->hr_nego_node_bitmap))) { + /* check negotiate bitmap every second to do timeout + * approve decision. + */ + schedule_delayed_work(®->hr_nego_timeout_work, + msecs_to_jiffies(1000)); + + return; + } + + /* approve negotiate timeout request. */ + } else { + /* negotiate timeout with master node. */ + } + } static inline void o2hb_bio_wait_init(struct o2hb_bio_wait_ctxt *wc) @@ -1032,7 +1074,7 @@ static int o2hb_do_disk_heartbeat(struct o2hb_region *reg) /* Skip disarming the timeout if own slot has stale/bad data */ if (own_slot_ok) { o2hb_set_quorum_device(reg); - o2hb_arm_write_timeout(reg); + o2hb_arm_timeout(reg); } bail: @@ -1114,7 +1156,7 @@ static int o2hb_thread(void *data) } } - o2hb_disarm_write_timeout(reg); + o2hb_disarm_timeout(reg); /* unclean stop is only used in very bad situation */ for(i = 0; !reg->hr_unclean_stop && i < reg->hr_blocks; i++) @@ -1762,6 +1804,7 @@ static ssize_t o2hb_region_dev_store(struct config_item *item, } INIT_DELAYED_WORK(®->hr_write_timeout_work, o2hb_write_timeout); + INIT_DELAYED_WORK(®->hr_nego_timeout_work, o2hb_nego_timeout); /* * A node is considered live after it has beat LIVE_THRESHOLD -- cgit v1.3-8-gc7d7 From 34069b886f95356d68bf8315fa648c4ab3193cdd Mon Sep 17 00:00:00 2001 From: Junxiao Bi Date: Fri, 27 May 2016 14:27:01 -0700 Subject: ocfs2: o2hb: add NEGO_TIMEOUT message This message is sent to master node when non-master nodes's negotiate timer expired. Master node records these nodes in a bitmap which is used to do write timeout timer re-queue decision. Signed-off-by: Junxiao Bi Reviewed-by: Ryan Ding Reviewed-by: Mark Fasheh Cc: Gang He Cc: rwxybh Cc: Joel Becker Cc: Joseph Qi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ocfs2/cluster/heartbeat.c | 66 +++++++++++++++++++++++++++++++++++++++++++- 1 file changed, 65 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ocfs2/cluster/heartbeat.c b/fs/ocfs2/cluster/heartbeat.c index 750c950f4e1f..454c89076833 100644 --- a/fs/ocfs2/cluster/heartbeat.c +++ b/fs/ocfs2/cluster/heartbeat.c @@ -280,6 +280,10 @@ struct o2hb_region { * being checked because we temporarily have to zero out the * crc field. */ struct o2hb_disk_heartbeat_block *hr_tmp_block; + + /* Message key for negotiate timeout message. */ + unsigned int hr_key; + struct list_head hr_handler_list; }; struct o2hb_bio_wait_ctxt { @@ -288,6 +292,14 @@ struct o2hb_bio_wait_ctxt { int wc_error; }; +enum { + O2HB_NEGO_TIMEOUT_MSG = 1, +}; + +struct o2hb_nego_msg { + u8 node_num; +}; + static void o2hb_write_timeout(struct work_struct *work) { int failed, quorum; @@ -355,6 +367,24 @@ static void o2hb_disarm_timeout(struct o2hb_region *reg) cancel_delayed_work_sync(®->hr_nego_timeout_work); } +static int o2hb_send_nego_msg(int key, int type, u8 target) +{ + struct o2hb_nego_msg msg; + int status, ret; + + msg.node_num = o2nm_this_node(); +again: + ret = o2net_send_message(type, key, &msg, sizeof(msg), + target, &status); + + if (ret == -EAGAIN || ret == -ENOMEM) { + msleep(100); + goto again; + } + + return ret; +} + static void o2hb_nego_timeout(struct work_struct *work) { unsigned long live_node_bitmap[BITS_TO_LONGS(O2NM_MAX_NODES)]; @@ -382,8 +412,24 @@ static void o2hb_nego_timeout(struct work_struct *work) /* approve negotiate timeout request. */ } else { /* negotiate timeout with master node. */ + o2hb_send_nego_msg(reg->hr_key, O2HB_NEGO_TIMEOUT_MSG, + master_node); } +} + +static int o2hb_nego_timeout_handler(struct o2net_msg *msg, u32 len, void *data, + void **ret_data) +{ + struct o2hb_region *reg = data; + struct o2hb_nego_msg *nego_msg; + nego_msg = (struct o2hb_nego_msg *)msg->buf; + if (nego_msg->node_num < O2NM_MAX_NODES) + set_bit(nego_msg->node_num, reg->hr_nego_node_bitmap); + else + mlog(ML_ERROR, "got nego timeout message from bad node.\n"); + + return 0; } static inline void o2hb_bio_wait_init(struct o2hb_bio_wait_ctxt *wc) @@ -1493,6 +1539,7 @@ static void o2hb_region_release(struct config_item *item) list_del(®->hr_all_item); spin_unlock(&o2hb_live_lock); + o2net_unregister_handler_list(®->hr_handler_list); kfree(reg); } @@ -2038,13 +2085,30 @@ static struct config_item *o2hb_heartbeat_group_make_item(struct config_group *g config_item_init_type_name(®->hr_item, name, &o2hb_region_type); + /* this is the same way to generate msg key as dlm, for local heartbeat, + * name is also the same, so make initial crc value different to avoid + * message key conflict. + */ + reg->hr_key = crc32_le(reg->hr_region_num + O2NM_MAX_REGIONS, + name, strlen(name)); + INIT_LIST_HEAD(®->hr_handler_list); + ret = o2net_register_handler(O2HB_NEGO_TIMEOUT_MSG, reg->hr_key, + sizeof(struct o2hb_nego_msg), + o2hb_nego_timeout_handler, + reg, NULL, ®->hr_handler_list); + if (ret) + goto free; + ret = o2hb_debug_region_init(reg, o2hb_debug_dir); if (ret) { config_item_put(®->hr_item); - goto free; + goto unregister_handler; } return ®->hr_item; + +unregister_handler: + o2net_unregister_handler_list(®->hr_handler_list); free: kfree(reg); return ERR_PTR(ret); -- cgit v1.3-8-gc7d7 From e76f8237a2f7b7220980c0fb3c6d0b1d48ba79ad Mon Sep 17 00:00:00 2001 From: Junxiao Bi Date: Fri, 27 May 2016 14:27:04 -0700 Subject: ocfs2: o2hb: add NEGOTIATE_APPROVE message This message is used to re-queue write timeout timer and negotiate timer when all nodes suffer a write hung to storage, this makes node not fence self if storage down. Signed-off-by: Junxiao Bi Reviewed-by: Ryan Ding Reviewed-by: Mark Fasheh Cc: Gang He Cc: rwxybh Cc: Joel Becker Cc: Joseph Qi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ocfs2/cluster/heartbeat.c | 28 +++++++++++++++++++++++++++- 1 file changed, 27 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ocfs2/cluster/heartbeat.c b/fs/ocfs2/cluster/heartbeat.c index 454c89076833..e929b15d6162 100644 --- a/fs/ocfs2/cluster/heartbeat.c +++ b/fs/ocfs2/cluster/heartbeat.c @@ -294,6 +294,7 @@ struct o2hb_bio_wait_ctxt { enum { O2HB_NEGO_TIMEOUT_MSG = 1, + O2HB_NEGO_APPROVE_MSG = 2, }; struct o2hb_nego_msg { @@ -388,7 +389,7 @@ again: static void o2hb_nego_timeout(struct work_struct *work) { unsigned long live_node_bitmap[BITS_TO_LONGS(O2NM_MAX_NODES)]; - int master_node; + int master_node, i; struct o2hb_region *reg; reg = container_of(work, struct o2hb_region, hr_nego_timeout_work.work); @@ -410,6 +411,17 @@ static void o2hb_nego_timeout(struct work_struct *work) } /* approve negotiate timeout request. */ + o2hb_arm_timeout(reg); + + i = -1; + while ((i = find_next_bit(live_node_bitmap, + O2NM_MAX_NODES, i + 1)) < O2NM_MAX_NODES) { + if (i == master_node) + continue; + + o2hb_send_nego_msg(reg->hr_key, + O2HB_NEGO_APPROVE_MSG, i); + } } else { /* negotiate timeout with master node. */ o2hb_send_nego_msg(reg->hr_key, O2HB_NEGO_TIMEOUT_MSG, @@ -432,6 +444,13 @@ static int o2hb_nego_timeout_handler(struct o2net_msg *msg, u32 len, void *data, return 0; } +static int o2hb_nego_approve_handler(struct o2net_msg *msg, u32 len, void *data, + void **ret_data) +{ + o2hb_arm_timeout(data); + return 0; +} + static inline void o2hb_bio_wait_init(struct o2hb_bio_wait_ctxt *wc) { atomic_set(&wc->wc_num_reqs, 1); @@ -2099,6 +2118,13 @@ static struct config_item *o2hb_heartbeat_group_make_item(struct config_group *g if (ret) goto free; + ret = o2net_register_handler(O2HB_NEGO_APPROVE_MSG, reg->hr_key, + sizeof(struct o2hb_nego_msg), + o2hb_nego_approve_handler, + reg, NULL, ®->hr_handler_list); + if (ret) + goto unregister_handler; + ret = o2hb_debug_region_init(reg, o2hb_debug_dir); if (ret) { config_item_put(®->hr_item); -- cgit v1.3-8-gc7d7 From 1bd1290283d7939478062e80bdd9719d3a21522f Mon Sep 17 00:00:00 2001 From: Junxiao Bi Date: Fri, 27 May 2016 14:27:07 -0700 Subject: ocfs2: o2hb: add some user/debug log Signed-off-by: Junxiao Bi Reviewed-by: Ryan Ding Reviewed-by: Mark Fasheh Cc: Gang He Cc: rwxybh Cc: Joel Becker Cc: Joseph Qi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ocfs2/cluster/heartbeat.c | 39 ++++++++++++++++++++++++++++++++------- 1 file changed, 32 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/cluster/heartbeat.c b/fs/ocfs2/cluster/heartbeat.c index e929b15d6162..84ebeb5678c6 100644 --- a/fs/ocfs2/cluster/heartbeat.c +++ b/fs/ocfs2/cluster/heartbeat.c @@ -292,6 +292,8 @@ struct o2hb_bio_wait_ctxt { int wc_error; }; +#define O2HB_NEGO_TIMEOUT_MS (O2HB_MAX_WRITE_TIMEOUT_MS/2) + enum { O2HB_NEGO_TIMEOUT_MSG = 1, O2HB_NEGO_APPROVE_MSG = 2, @@ -358,7 +360,7 @@ static void o2hb_arm_timeout(struct o2hb_region *reg) cancel_delayed_work(®->hr_nego_timeout_work); /* negotiate timeout must be less than write timeout. */ schedule_delayed_work(®->hr_nego_timeout_work, - msecs_to_jiffies(O2HB_MAX_WRITE_TIMEOUT_MS)/2); + msecs_to_jiffies(O2HB_NEGO_TIMEOUT_MS)); memset(reg->hr_nego_node_bitmap, 0, sizeof(reg->hr_nego_node_bitmap)); } @@ -389,7 +391,7 @@ again: static void o2hb_nego_timeout(struct work_struct *work) { unsigned long live_node_bitmap[BITS_TO_LONGS(O2NM_MAX_NODES)]; - int master_node, i; + int master_node, i, ret; struct o2hb_region *reg; reg = container_of(work, struct o2hb_region, hr_nego_timeout_work.work); @@ -398,7 +400,12 @@ static void o2hb_nego_timeout(struct work_struct *work) master_node = find_next_bit(live_node_bitmap, O2NM_MAX_NODES, 0); if (master_node == o2nm_this_node()) { - set_bit(master_node, reg->hr_nego_node_bitmap); + if (!test_bit(master_node, reg->hr_nego_node_bitmap)) { + printk(KERN_NOTICE "o2hb: node %d hb write hung for %ds on region %s (%s).\n", + o2nm_this_node(), O2HB_NEGO_TIMEOUT_MS/1000, + config_item_name(®->hr_item), reg->hr_dev_name); + set_bit(master_node, reg->hr_nego_node_bitmap); + } if (memcmp(reg->hr_nego_node_bitmap, live_node_bitmap, sizeof(reg->hr_nego_node_bitmap))) { /* check negotiate bitmap every second to do timeout @@ -410,6 +417,8 @@ static void o2hb_nego_timeout(struct work_struct *work) return; } + printk(KERN_NOTICE "o2hb: all nodes hb write hung, maybe region %s (%s) is down.\n", + config_item_name(®->hr_item), reg->hr_dev_name); /* approve negotiate timeout request. */ o2hb_arm_timeout(reg); @@ -419,13 +428,23 @@ static void o2hb_nego_timeout(struct work_struct *work) if (i == master_node) continue; - o2hb_send_nego_msg(reg->hr_key, + mlog(ML_HEARTBEAT, "send NEGO_APPROVE msg to node %d\n", i); + ret = o2hb_send_nego_msg(reg->hr_key, O2HB_NEGO_APPROVE_MSG, i); + if (ret) + mlog(ML_ERROR, "send NEGO_APPROVE msg to node %d fail %d\n", + i, ret); } } else { /* negotiate timeout with master node. */ - o2hb_send_nego_msg(reg->hr_key, O2HB_NEGO_TIMEOUT_MSG, - master_node); + printk(KERN_NOTICE "o2hb: node %d hb write hung for %ds on region %s (%s), negotiate timeout with node %d.\n", + o2nm_this_node(), O2HB_NEGO_TIMEOUT_MS/1000, config_item_name(®->hr_item), + reg->hr_dev_name, master_node); + ret = o2hb_send_nego_msg(reg->hr_key, O2HB_NEGO_TIMEOUT_MSG, + master_node); + if (ret) + mlog(ML_ERROR, "send NEGO_TIMEOUT msg to node %d fail %d\n", + master_node, ret); } } @@ -436,6 +455,8 @@ static int o2hb_nego_timeout_handler(struct o2net_msg *msg, u32 len, void *data, struct o2hb_nego_msg *nego_msg; nego_msg = (struct o2hb_nego_msg *)msg->buf; + printk(KERN_NOTICE "o2hb: receive negotiate timeout message from node %d on region %s (%s).\n", + nego_msg->node_num, config_item_name(®->hr_item), reg->hr_dev_name); if (nego_msg->node_num < O2NM_MAX_NODES) set_bit(nego_msg->node_num, reg->hr_nego_node_bitmap); else @@ -447,7 +468,11 @@ static int o2hb_nego_timeout_handler(struct o2net_msg *msg, u32 len, void *data, static int o2hb_nego_approve_handler(struct o2net_msg *msg, u32 len, void *data, void **ret_data) { - o2hb_arm_timeout(data); + struct o2hb_region *reg = data; + + printk(KERN_NOTICE "o2hb: negotiate timeout approved by master node on region %s (%s).\n", + config_item_name(®->hr_item), reg->hr_dev_name); + o2hb_arm_timeout(reg); return 0; } -- cgit v1.3-8-gc7d7 From 88dbe98dc72fa0c2b778210e2079cd69d2477b36 Mon Sep 17 00:00:00 2001 From: Junxiao Bi Date: Fri, 27 May 2016 14:27:10 -0700 Subject: ocfs2: o2hb: don't negotiate if last hb fail Sometimes io error is returned when storage is down for a while. Like for iscsi device, stroage is made offline when session timeout, and this will make all io return -EIO. For this case, nodes shouldn't do negotiate timeout but should fence self. So let nodes fence self when o2hb_do_disk_heartbeat return an error, this is the same behavior with o2hb without negotiate timer. Signed-off-by: Junxiao Bi Reviewed-by: Ryan Ding Reviewed-by: Mark Fasheh Cc: Gang He Cc: rwxybh Cc: Joel Becker Cc: Joseph Qi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ocfs2/cluster/heartbeat.c | 10 ++++++++++ 1 file changed, 10 insertions(+) (limited to 'fs') diff --git a/fs/ocfs2/cluster/heartbeat.c b/fs/ocfs2/cluster/heartbeat.c index 84ebeb5678c6..31a5c0803e0a 100644 --- a/fs/ocfs2/cluster/heartbeat.c +++ b/fs/ocfs2/cluster/heartbeat.c @@ -284,6 +284,9 @@ struct o2hb_region { /* Message key for negotiate timeout message. */ unsigned int hr_key; struct list_head hr_handler_list; + + /* last hb status, 0 for success, other value for error. */ + int hr_last_hb_status; }; struct o2hb_bio_wait_ctxt { @@ -395,6 +398,12 @@ static void o2hb_nego_timeout(struct work_struct *work) struct o2hb_region *reg; reg = container_of(work, struct o2hb_region, hr_nego_timeout_work.work); + /* don't negotiate timeout if last hb failed since it is very + * possible io failed. Should let write timeout fence self. + */ + if (reg->hr_last_hb_status) + return; + o2hb_fill_node_map(live_node_bitmap, sizeof(live_node_bitmap)); /* lowest node as master node to make negotiate decision. */ master_node = find_next_bit(live_node_bitmap, O2NM_MAX_NODES, 0); @@ -1228,6 +1237,7 @@ static int o2hb_thread(void *data) before_hb = ktime_get_real(); ret = o2hb_do_disk_heartbeat(reg); + reg->hr_last_hb_status = ret; after_hb = ktime_get_real(); -- cgit v1.3-8-gc7d7 From 6633ca573165e000867e50caf94f60a75399b68b Mon Sep 17 00:00:00 2001 From: Junxiao Bi Date: Fri, 27 May 2016 14:27:13 -0700 Subject: ocfs2: o2hb: fix hb hung time hr_last_timeout_start should be set as the last time where hb is still OK. When hb write timeout, hung time will be (jiffies - hr_last_timeout_start). Signed-off-by: Junxiao Bi Reviewed-by: Ryan Ding Reviewed-by: Mark Fasheh Cc: Gang He Cc: rwxybh Cc: Joel Becker Cc: Joseph Qi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ocfs2/cluster/heartbeat.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ocfs2/cluster/heartbeat.c b/fs/ocfs2/cluster/heartbeat.c index 31a5c0803e0a..6aaf3e351391 100644 --- a/fs/ocfs2/cluster/heartbeat.c +++ b/fs/ocfs2/cluster/heartbeat.c @@ -356,7 +356,6 @@ static void o2hb_arm_timeout(struct o2hb_region *reg) spin_unlock(&o2hb_live_lock); } cancel_delayed_work(®->hr_write_timeout_work); - reg->hr_last_timeout_start = jiffies; schedule_delayed_work(®->hr_write_timeout_work, msecs_to_jiffies(O2HB_MAX_WRITE_TIMEOUT_MS)); @@ -1174,6 +1173,7 @@ static int o2hb_do_disk_heartbeat(struct o2hb_region *reg) if (own_slot_ok) { o2hb_set_quorum_device(reg); o2hb_arm_timeout(reg); + reg->hr_last_timeout_start = jiffies; } bail: -- cgit v1.3-8-gc7d7 From 38b52efd218bf2a11a5b4a8f56052cee6684cfec Mon Sep 17 00:00:00 2001 From: Junxiao Bi Date: Fri, 27 May 2016 14:27:16 -0700 Subject: ocfs2: bump up o2cb network protocol version Two new messages are added to support negotiating hb timeout. Stop nodes frmo talking an old version to mount as they will cause the negotiation to fail. Link: http://lkml.kernel.org/r/1464231615-27939-1-git-send-email-junxiao.bi@oracle.com Signed-off-by: Junxiao Bi Reviewed-by: Mark Fasheh Cc: Joel Becker Cc: Joseph Qi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ocfs2/cluster/tcp_internal.h | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ocfs2/cluster/tcp_internal.h b/fs/ocfs2/cluster/tcp_internal.h index b95e7df5b76a..94b18369b1cc 100644 --- a/fs/ocfs2/cluster/tcp_internal.h +++ b/fs/ocfs2/cluster/tcp_internal.h @@ -44,6 +44,9 @@ * version here in tcp_internal.h should not need to be bumped for * filesystem locking changes. * + * New in version 12 + * - Negotiate hb timeout when storage is down. + * * New in version 11 * - Negotiation of filesystem locking in the dlm join. * @@ -75,7 +78,7 @@ * - full 64 bit i_size in the metadata lock lvbs * - introduction of "rw" lock and pushing meta/data locking down */ -#define O2NET_PROTOCOL_VERSION 11ULL +#define O2NET_PROTOCOL_VERSION 12ULL struct o2net_handshake { __be64 protocol_version; __be64 connector_id; -- cgit v1.3-8-gc7d7 From 9ecd10b7a0270803fd5f36ab93173e6d5b41b895 Mon Sep 17 00:00:00 2001 From: Eryu Guan Date: Fri, 27 May 2016 14:27:18 -0700 Subject: direct-io: fix direct write stale data exposure from concurrent buffered read Currently direct writes inside i_size on a DIO_SKIP_HOLES filesystem are not allowed to allocate blocks(get_more_blocks() sets 'create' to 0 before calling get_block() callback), if it's a sparse file, direct writes fall back to buffered writes to avoid stale data exposure from concurrent buffered read. But there're two cases that can result in stale data exposure are not correctly detected. 1. The detection for "writing inside i_size" is not sufficient, writes can be treated as "extending writes" wrongly. For example, direct write 1FSB (file system block) to a 1FSB sparse file on ext2/3/4, starting from offset 0, in this case it's writing inside i_size, but 'create' is non-zero, because 'block_in_file' and '(i_size_read(inode) >> blkbits' are both zero. 2. Direct writes starting from or beyong i_size (not inside i_size) also could trigger block allocation and expose stale data. For example, consider a sparse file with i_size of 2k, and a write to offset 2k or 3k into the file, with a filesystem block size of 4k. (Thanks to Jeff Moyer for pointing this case out in his review.) The first problem can be demostrated by running ltp-aiodio test ADSP045 many times. When testing on extN filesystems, I see test failures occasionally, buffered read could read non-zero (stale) data. ADSP045: dio_sparse -a 4k -w 4k -s 2k -n 1 dio_sparse 0 TINFO : Dirtying free blocks dio_sparse 0 TINFO : Starting I/O tests non zero buffer at buf[0] => 0xffffffaa,ffffffaa,ffffffaa,ffffffaa non-zero read at offset 0 dio_sparse 0 TINFO : Killing childrens(s) dio_sparse 1 TFAIL : dio_sparse.c:191: 1 children(s) exited abnormally The second problem can also be reproduced easily by a hacked dio_sparse program, which accepts an option to specify the write offset. What we should really do is to disable block allocation for writes that could result in filling holes inside i_size. Link: http://lkml.kernel.org/r/1463156728-13357-1-git-send-email-guaneryu@gmail.com Reviewed-by: Jan Kara Signed-off-by: Eryu Guan Cc: Al Viro Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/direct-io.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/direct-io.c b/fs/direct-io.c index 3bf3f20f8ecc..f3b4408be590 100644 --- a/fs/direct-io.c +++ b/fs/direct-io.c @@ -628,11 +628,11 @@ static int get_more_blocks(struct dio *dio, struct dio_submit *sdio, map_bh->b_size = fs_count << i_blkbits; /* - * For writes inside i_size on a DIO_SKIP_HOLES filesystem we - * forbid block creations: only overwrites are permitted. - * We will return early to the caller once we see an - * unmapped buffer head returned, and the caller will fall - * back to buffered I/O. + * For writes that could fill holes inside i_size on a + * DIO_SKIP_HOLES filesystem we forbid block creations: only + * overwrites are permitted. We will return early to the caller + * once we see an unmapped buffer head returned, and the caller + * will fall back to buffered I/O. * * Otherwise the decision is left to the get_blocks method, * which may decide to handle it or also return an unmapped @@ -640,8 +640,8 @@ static int get_more_blocks(struct dio *dio, struct dio_submit *sdio, */ create = dio->rw & WRITE; if (dio->flags & DIO_SKIP_HOLES) { - if (sdio->block_in_file < (i_size_read(dio->inode) >> - sdio->blkbits)) + if (fs_startblk <= ((i_size_read(dio->inode) - 1) >> + i_blkbits)) create = 0; } -- cgit v1.3-8-gc7d7 From 287980e49ffc0f6d911601e7e352a812ed27768e Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Fri, 27 May 2016 23:23:25 +0200 Subject: remove lots of IS_ERR_VALUE abuses Most users of IS_ERR_VALUE() in the kernel are wrong, as they pass an 'int' into a function that takes an 'unsigned long' argument. This happens to work because the type is sign-extended on 64-bit architectures before it gets converted into an unsigned type. However, anything that passes an 'unsigned short' or 'unsigned int' argument into IS_ERR_VALUE() is guaranteed to be broken, as are 8-bit integers and types that are wider than 'unsigned long'. Andrzej Hajda has already fixed a lot of the worst abusers that were causing actual bugs, but it would be nice to prevent any users that are not passing 'unsigned long' arguments. This patch changes all users of IS_ERR_VALUE() that I could find on 32-bit ARM randconfig builds and x86 allmodconfig. For the moment, this doesn't change the definition of IS_ERR_VALUE() because there are probably still architecture specific users elsewhere. Almost all the warnings I got are for files that are better off using 'if (err)' or 'if (err < 0)'. The only legitimate user I could find that we get a warning for is the (32-bit only) freescale fman driver, so I did not remove the IS_ERR_VALUE() there but changed the type to 'unsigned long'. For 9pfs, I just worked around one user whose calling conventions are so obscure that I did not dare change the behavior. I was using this definition for testing: #define IS_ERR_VALUE(x) ((unsigned long*)NULL == (typeof (x)*)NULL && \ unlikely((unsigned long long)(x) >= (unsigned long long)(typeof(x))-MAX_ERRNO)) which ends up making all 16-bit or wider types work correctly with the most plausible interpretation of what IS_ERR_VALUE() was supposed to return according to its users, but also causes a compile-time warning for any users that do not pass an 'unsigned long' argument. I suggested this approach earlier this year, but back then we ended up deciding to just fix the users that are obviously broken. After the initial warning that caused me to get involved in the discussion (fs/gfs2/dir.c) showed up again in the mainline kernel, Linus asked me to send the whole thing again. [ Updated the 9p parts as per Al Viro - Linus ] Signed-off-by: Arnd Bergmann Cc: Andrzej Hajda Cc: Andrew Morton Link: https://lkml.org/lkml/2016/1/7/363 Link: https://lkml.org/lkml/2016/5/27/486 Acked-by: Srinivas Kandagatla # For nvmem part Signed-off-by: Linus Torvalds --- drivers/acpi/acpi_dbg.c | 22 +++++++++++----------- drivers/ata/sata_highbank.c | 2 +- drivers/clk/tegra/clk-tegra210.c | 2 +- drivers/cpufreq/omap-cpufreq.c | 2 +- drivers/crypto/caam/ctrl.c | 2 +- drivers/dma/sun4i-dma.c | 16 ++++++++-------- drivers/gpio/gpio-xlp.c | 2 +- drivers/gpu/drm/sti/sti_vtg.c | 4 ++-- drivers/gpu/drm/tilcdc/tilcdc_tfp410.c | 2 +- drivers/gpu/host1x/hw/intr_hw.c | 2 +- drivers/iommu/arm-smmu-v3.c | 18 +++++++++--------- drivers/iommu/arm-smmu.c | 8 ++++---- drivers/irqchip/irq-clps711x.c | 2 +- drivers/irqchip/irq-gic.c | 2 +- drivers/irqchip/irq-hip04.c | 2 +- drivers/irqchip/spear-shirq.c | 2 +- drivers/media/i2c/adp1653.c | 10 +++++----- drivers/media/platform/s5p-tv/mixer_drv.c | 2 +- drivers/mfd/twl4030-irq.c | 2 +- drivers/mmc/core/mmc.c | 4 ++-- drivers/mmc/host/dw_mmc.c | 6 +++--- drivers/mmc/host/sdhci-esdhc-imx.c | 2 +- drivers/mmc/host/sdhci-of-at91.c | 2 +- drivers/mmc/host/sdhci.c | 4 ++-- drivers/net/ethernet/freescale/fman/fman.c | 2 +- drivers/net/ethernet/freescale/fman/fman_muram.c | 4 ++-- drivers/net/ethernet/freescale/fman/fman_muram.h | 4 ++-- drivers/net/wireless/ti/wlcore/spi.c | 4 ++-- drivers/nvmem/core.c | 22 +++++++++++----------- drivers/tty/serial/amba-pl011.c | 2 +- drivers/tty/serial/sprd_serial.c | 2 +- drivers/video/fbdev/da8xx-fb.c | 4 ++-- fs/afs/write.c | 4 ---- fs/binfmt_flat.c | 6 +++--- fs/gfs2/dir.c | 15 +++++++++------ kernel/pid.c | 2 +- net/9p/client.c | 8 ++++---- sound/soc/qcom/lpass-platform.c | 4 ++-- 38 files changed, 102 insertions(+), 103 deletions(-) (limited to 'fs') diff --git a/drivers/acpi/acpi_dbg.c b/drivers/acpi/acpi_dbg.c index 15e4604efba7..1f4128487dd4 100644 --- a/drivers/acpi/acpi_dbg.c +++ b/drivers/acpi/acpi_dbg.c @@ -265,7 +265,7 @@ static int acpi_aml_write_kern(const char *buf, int len) char *p; ret = acpi_aml_lock_write(crc, ACPI_AML_OUT_KERN); - if (IS_ERR_VALUE(ret)) + if (ret < 0) return ret; /* sync tail before inserting logs */ smp_mb(); @@ -286,7 +286,7 @@ static int acpi_aml_readb_kern(void) char *p; ret = acpi_aml_lock_read(crc, ACPI_AML_IN_KERN); - if (IS_ERR_VALUE(ret)) + if (ret < 0) return ret; /* sync head before removing cmds */ smp_rmb(); @@ -330,7 +330,7 @@ again: goto again; break; } - if (IS_ERR_VALUE(ret)) + if (ret < 0) break; size += ret; count -= ret; @@ -373,7 +373,7 @@ again: if (ret == 0) goto again; } - if (IS_ERR_VALUE(ret)) + if (ret < 0) break; *(msg + size) = (char)ret; size++; @@ -526,7 +526,7 @@ static int acpi_aml_open(struct inode *inode, struct file *file) } acpi_aml_io.users++; err_lock: - if (IS_ERR_VALUE(ret)) { + if (ret < 0) { if (acpi_aml_active_reader == file) acpi_aml_active_reader = NULL; } @@ -587,7 +587,7 @@ static int acpi_aml_read_user(char __user *buf, int len) char *p; ret = acpi_aml_lock_read(crc, ACPI_AML_OUT_USER); - if (IS_ERR_VALUE(ret)) + if (ret < 0) return ret; /* sync head before removing logs */ smp_rmb(); @@ -602,7 +602,7 @@ static int acpi_aml_read_user(char __user *buf, int len) crc->tail = (crc->tail + n) & (ACPI_AML_BUF_SIZE - 1); ret = n; out: - acpi_aml_unlock_fifo(ACPI_AML_OUT_USER, !IS_ERR_VALUE(ret)); + acpi_aml_unlock_fifo(ACPI_AML_OUT_USER, !ret); return ret; } @@ -634,7 +634,7 @@ again: goto again; } } - if (IS_ERR_VALUE(ret)) { + if (ret < 0) { if (!acpi_aml_running()) ret = 0; break; @@ -657,7 +657,7 @@ static int acpi_aml_write_user(const char __user *buf, int len) char *p; ret = acpi_aml_lock_write(crc, ACPI_AML_IN_USER); - if (IS_ERR_VALUE(ret)) + if (ret < 0) return ret; /* sync tail before inserting cmds */ smp_mb(); @@ -672,7 +672,7 @@ static int acpi_aml_write_user(const char __user *buf, int len) crc->head = (crc->head + n) & (ACPI_AML_BUF_SIZE - 1); ret = n; out: - acpi_aml_unlock_fifo(ACPI_AML_IN_USER, !IS_ERR_VALUE(ret)); + acpi_aml_unlock_fifo(ACPI_AML_IN_USER, !ret); return n; } @@ -704,7 +704,7 @@ again: goto again; } } - if (IS_ERR_VALUE(ret)) { + if (ret < 0) { if (!acpi_aml_running()) ret = 0; break; diff --git a/drivers/ata/sata_highbank.c b/drivers/ata/sata_highbank.c index 8638d575b2b9..aafb8cc03523 100644 --- a/drivers/ata/sata_highbank.c +++ b/drivers/ata/sata_highbank.c @@ -197,7 +197,7 @@ static void highbank_set_em_messages(struct device *dev, for (i = 0; i < SGPIO_PINS; i++) { err = of_get_named_gpio(np, "calxeda,sgpio-gpio", i); - if (IS_ERR_VALUE(err)) + if (err < 0) return; pdata->sgpio_gpio[i] = err; diff --git a/drivers/clk/tegra/clk-tegra210.c b/drivers/clk/tegra/clk-tegra210.c index b8551813ec43..456cf586d2c2 100644 --- a/drivers/clk/tegra/clk-tegra210.c +++ b/drivers/clk/tegra/clk-tegra210.c @@ -1221,7 +1221,7 @@ static int tegra210_pll_fixed_mdiv_cfg(struct clk_hw *hw, p = rate >= params->vco_min ? 1 : -EINVAL; } - if (IS_ERR_VALUE(p)) + if (p < 0) return -EINVAL; cfg->m = tegra_pll_get_fixed_mdiv(hw, input_rate); diff --git a/drivers/cpufreq/omap-cpufreq.c b/drivers/cpufreq/omap-cpufreq.c index cead9bec4843..376e63ca94e8 100644 --- a/drivers/cpufreq/omap-cpufreq.c +++ b/drivers/cpufreq/omap-cpufreq.c @@ -54,7 +54,7 @@ static int omap_target(struct cpufreq_policy *policy, unsigned int index) freq = new_freq * 1000; ret = clk_round_rate(policy->clk, freq); - if (IS_ERR_VALUE(ret)) { + if (ret < 0) { dev_warn(mpu_dev, "CPUfreq: Cannot find matching frequency for %lu\n", freq); diff --git a/drivers/crypto/caam/ctrl.c b/drivers/crypto/caam/ctrl.c index 44d30b45f3cc..5ad5f3009ae0 100644 --- a/drivers/crypto/caam/ctrl.c +++ b/drivers/crypto/caam/ctrl.c @@ -402,7 +402,7 @@ int caam_get_era(void) ret = of_property_read_u32(caam_node, "fsl,sec-era", &prop); of_node_put(caam_node); - return IS_ERR_VALUE(ret) ? -ENOTSUPP : prop; + return ret ? -ENOTSUPP : prop; } EXPORT_SYMBOL(caam_get_era); diff --git a/drivers/dma/sun4i-dma.c b/drivers/dma/sun4i-dma.c index e0df233dde92..57aa227bfadb 100644 --- a/drivers/dma/sun4i-dma.c +++ b/drivers/dma/sun4i-dma.c @@ -461,25 +461,25 @@ generate_ndma_promise(struct dma_chan *chan, dma_addr_t src, dma_addr_t dest, /* Source burst */ ret = convert_burst(sconfig->src_maxburst); - if (IS_ERR_VALUE(ret)) + if (ret < 0) goto fail; promise->cfg |= SUN4I_DMA_CFG_SRC_BURST_LENGTH(ret); /* Destination burst */ ret = convert_burst(sconfig->dst_maxburst); - if (IS_ERR_VALUE(ret)) + if (ret < 0) goto fail; promise->cfg |= SUN4I_DMA_CFG_DST_BURST_LENGTH(ret); /* Source bus width */ ret = convert_buswidth(sconfig->src_addr_width); - if (IS_ERR_VALUE(ret)) + if (ret < 0) goto fail; promise->cfg |= SUN4I_DMA_CFG_SRC_DATA_WIDTH(ret); /* Destination bus width */ ret = convert_buswidth(sconfig->dst_addr_width); - if (IS_ERR_VALUE(ret)) + if (ret < 0) goto fail; promise->cfg |= SUN4I_DMA_CFG_DST_DATA_WIDTH(ret); @@ -518,25 +518,25 @@ generate_ddma_promise(struct dma_chan *chan, dma_addr_t src, dma_addr_t dest, /* Source burst */ ret = convert_burst(sconfig->src_maxburst); - if (IS_ERR_VALUE(ret)) + if (ret < 0) goto fail; promise->cfg |= SUN4I_DMA_CFG_SRC_BURST_LENGTH(ret); /* Destination burst */ ret = convert_burst(sconfig->dst_maxburst); - if (IS_ERR_VALUE(ret)) + if (ret < 0) goto fail; promise->cfg |= SUN4I_DMA_CFG_DST_BURST_LENGTH(ret); /* Source bus width */ ret = convert_buswidth(sconfig->src_addr_width); - if (IS_ERR_VALUE(ret)) + if (ret < 0) goto fail; promise->cfg |= SUN4I_DMA_CFG_SRC_DATA_WIDTH(ret); /* Destination bus width */ ret = convert_buswidth(sconfig->dst_addr_width); - if (IS_ERR_VALUE(ret)) + if (ret < 0) goto fail; promise->cfg |= SUN4I_DMA_CFG_DST_DATA_WIDTH(ret); diff --git a/drivers/gpio/gpio-xlp.c b/drivers/gpio/gpio-xlp.c index 08897dc11915..1a33a19d95b9 100644 --- a/drivers/gpio/gpio-xlp.c +++ b/drivers/gpio/gpio-xlp.c @@ -393,7 +393,7 @@ static int xlp_gpio_probe(struct platform_device *pdev) irq_base = irq_alloc_descs(-1, 0, gc->ngpio, 0); else irq_base = irq_alloc_descs(-1, XLP_GPIO_IRQ_BASE, gc->ngpio, 0); - if (IS_ERR_VALUE(irq_base)) { + if (irq_base < 0) { dev_err(&pdev->dev, "Failed to allocate IRQ numbers\n"); return irq_base; } diff --git a/drivers/gpu/drm/sti/sti_vtg.c b/drivers/gpu/drm/sti/sti_vtg.c index 32c7986b63ab..6bf4ce466d20 100644 --- a/drivers/gpu/drm/sti/sti_vtg.c +++ b/drivers/gpu/drm/sti/sti_vtg.c @@ -437,7 +437,7 @@ static int vtg_probe(struct platform_device *pdev) return -EPROBE_DEFER; } else { vtg->irq = platform_get_irq(pdev, 0); - if (IS_ERR_VALUE(vtg->irq)) { + if (vtg->irq < 0) { DRM_ERROR("Failed to get VTG interrupt\n"); return vtg->irq; } @@ -447,7 +447,7 @@ static int vtg_probe(struct platform_device *pdev) ret = devm_request_threaded_irq(dev, vtg->irq, vtg_irq, vtg_irq_thread, IRQF_ONESHOT, dev_name(dev), vtg); - if (IS_ERR_VALUE(ret)) { + if (ret < 0) { DRM_ERROR("Failed to register VTG interrupt\n"); return ret; } diff --git a/drivers/gpu/drm/tilcdc/tilcdc_tfp410.c b/drivers/gpu/drm/tilcdc/tilcdc_tfp410.c index 7716f42f8aab..6b8c5b3bf588 100644 --- a/drivers/gpu/drm/tilcdc/tilcdc_tfp410.c +++ b/drivers/gpu/drm/tilcdc/tilcdc_tfp410.c @@ -342,7 +342,7 @@ static int tfp410_probe(struct platform_device *pdev) tfp410_mod->gpio = of_get_named_gpio_flags(node, "powerdn-gpio", 0, NULL); - if (IS_ERR_VALUE(tfp410_mod->gpio)) { + if (tfp410_mod->gpio < 0) { dev_warn(&pdev->dev, "No power down GPIO\n"); } else { ret = gpio_request(tfp410_mod->gpio, "DVI_PDn"); diff --git a/drivers/gpu/host1x/hw/intr_hw.c b/drivers/gpu/host1x/hw/intr_hw.c index 498b37e39058..e1e31e9e67cd 100644 --- a/drivers/gpu/host1x/hw/intr_hw.c +++ b/drivers/gpu/host1x/hw/intr_hw.c @@ -85,7 +85,7 @@ static int _host1x_intr_init_host_sync(struct host1x *host, u32 cpm, err = devm_request_irq(host->dev, host->intr_syncpt_irq, syncpt_thresh_isr, IRQF_SHARED, "host1x_syncpt", host); - if (IS_ERR_VALUE(err)) { + if (err < 0) { WARN_ON(1); return err; } diff --git a/drivers/iommu/arm-smmu-v3.c b/drivers/iommu/arm-smmu-v3.c index ebab33e77d67..94b68213c50d 100644 --- a/drivers/iommu/arm-smmu-v3.c +++ b/drivers/iommu/arm-smmu-v3.c @@ -1477,7 +1477,7 @@ static int arm_smmu_domain_finalise_s1(struct arm_smmu_domain *smmu_domain, struct arm_smmu_s1_cfg *cfg = &smmu_domain->s1_cfg; asid = arm_smmu_bitmap_alloc(smmu->asid_map, smmu->asid_bits); - if (IS_ERR_VALUE(asid)) + if (asid < 0) return asid; cfg->cdptr = dmam_alloc_coherent(smmu->dev, CTXDESC_CD_DWORDS << 3, @@ -1508,7 +1508,7 @@ static int arm_smmu_domain_finalise_s2(struct arm_smmu_domain *smmu_domain, struct arm_smmu_s2_cfg *cfg = &smmu_domain->s2_cfg; vmid = arm_smmu_bitmap_alloc(smmu->vmid_map, smmu->vmid_bits); - if (IS_ERR_VALUE(vmid)) + if (vmid < 0) return vmid; cfg->vmid = (u16)vmid; @@ -1569,7 +1569,7 @@ static int arm_smmu_domain_finalise(struct iommu_domain *domain) smmu_domain->pgtbl_ops = pgtbl_ops; ret = finalise_stage_fn(smmu_domain, &pgtbl_cfg); - if (IS_ERR_VALUE(ret)) + if (ret < 0) free_io_pgtable_ops(pgtbl_ops); return ret; @@ -1642,7 +1642,7 @@ static void arm_smmu_detach_dev(struct device *dev) struct arm_smmu_group *smmu_group = arm_smmu_group_get(dev); smmu_group->ste.bypass = true; - if (IS_ERR_VALUE(arm_smmu_install_ste_for_group(smmu_group))) + if (arm_smmu_install_ste_for_group(smmu_group) < 0) dev_warn(dev, "failed to install bypass STE\n"); smmu_group->domain = NULL; @@ -1694,7 +1694,7 @@ static int arm_smmu_attach_dev(struct iommu_domain *domain, struct device *dev) smmu_group->ste.bypass = domain->type == IOMMU_DOMAIN_DMA; ret = arm_smmu_install_ste_for_group(smmu_group); - if (IS_ERR_VALUE(ret)) + if (ret < 0) smmu_group->domain = NULL; out_unlock: @@ -2235,7 +2235,7 @@ static int arm_smmu_setup_irqs(struct arm_smmu_device *smmu) arm_smmu_evtq_handler, arm_smmu_evtq_thread, 0, "arm-smmu-v3-evtq", smmu); - if (IS_ERR_VALUE(ret)) + if (ret < 0) dev_warn(smmu->dev, "failed to enable evtq irq\n"); } @@ -2244,7 +2244,7 @@ static int arm_smmu_setup_irqs(struct arm_smmu_device *smmu) ret = devm_request_irq(smmu->dev, irq, arm_smmu_cmdq_sync_handler, 0, "arm-smmu-v3-cmdq-sync", smmu); - if (IS_ERR_VALUE(ret)) + if (ret < 0) dev_warn(smmu->dev, "failed to enable cmdq-sync irq\n"); } @@ -2252,7 +2252,7 @@ static int arm_smmu_setup_irqs(struct arm_smmu_device *smmu) if (irq) { ret = devm_request_irq(smmu->dev, irq, arm_smmu_gerror_handler, 0, "arm-smmu-v3-gerror", smmu); - if (IS_ERR_VALUE(ret)) + if (ret < 0) dev_warn(smmu->dev, "failed to enable gerror irq\n"); } @@ -2264,7 +2264,7 @@ static int arm_smmu_setup_irqs(struct arm_smmu_device *smmu) arm_smmu_priq_thread, 0, "arm-smmu-v3-priq", smmu); - if (IS_ERR_VALUE(ret)) + if (ret < 0) dev_warn(smmu->dev, "failed to enable priq irq\n"); else diff --git a/drivers/iommu/arm-smmu.c b/drivers/iommu/arm-smmu.c index e206ce7a4e4b..9345a3fcb706 100644 --- a/drivers/iommu/arm-smmu.c +++ b/drivers/iommu/arm-smmu.c @@ -950,7 +950,7 @@ static int arm_smmu_init_domain_context(struct iommu_domain *domain, ret = __arm_smmu_alloc_bitmap(smmu->context_map, start, smmu->num_context_banks); - if (IS_ERR_VALUE(ret)) + if (ret < 0) goto out_unlock; cfg->cbndx = ret; @@ -989,7 +989,7 @@ static int arm_smmu_init_domain_context(struct iommu_domain *domain, irq = smmu->irqs[smmu->num_global_irqs + cfg->irptndx]; ret = request_irq(irq, arm_smmu_context_fault, IRQF_SHARED, "arm-smmu-context-fault", domain); - if (IS_ERR_VALUE(ret)) { + if (ret < 0) { dev_err(smmu->dev, "failed to request context IRQ %d (%u)\n", cfg->irptndx, irq); cfg->irptndx = INVALID_IRPTNDX; @@ -1099,7 +1099,7 @@ static int arm_smmu_master_configure_smrs(struct arm_smmu_device *smmu, for (i = 0; i < cfg->num_streamids; ++i) { int idx = __arm_smmu_alloc_bitmap(smmu->smr_map, 0, smmu->num_mapping_groups); - if (IS_ERR_VALUE(idx)) { + if (idx < 0) { dev_err(smmu->dev, "failed to allocate free SMR\n"); goto err_free_smrs; } @@ -1233,7 +1233,7 @@ static int arm_smmu_attach_dev(struct iommu_domain *domain, struct device *dev) /* Ensure that the domain is finalised */ ret = arm_smmu_init_domain_context(domain, smmu); - if (IS_ERR_VALUE(ret)) + if (ret < 0) return ret; /* diff --git a/drivers/irqchip/irq-clps711x.c b/drivers/irqchip/irq-clps711x.c index eb5eb0cd414d..2223b3f15d68 100644 --- a/drivers/irqchip/irq-clps711x.c +++ b/drivers/irqchip/irq-clps711x.c @@ -182,7 +182,7 @@ static int __init _clps711x_intc_init(struct device_node *np, writel_relaxed(0, clps711x_intc->intmr[2]); err = irq_alloc_descs(-1, 0, ARRAY_SIZE(clps711x_irqs), numa_node_id()); - if (IS_ERR_VALUE(err)) + if (err < 0) goto out_iounmap; clps711x_intc->ops.map = clps711x_intc_irq_map; diff --git a/drivers/irqchip/irq-gic.c b/drivers/irqchip/irq-gic.c index b4e647179346..fbc4ae2afd29 100644 --- a/drivers/irqchip/irq-gic.c +++ b/drivers/irqchip/irq-gic.c @@ -1123,7 +1123,7 @@ static int __init __gic_init_bases(struct gic_chip_data *gic, int irq_start, irq_base = irq_alloc_descs(irq_start, 16, gic_irqs, numa_node_id()); - if (IS_ERR_VALUE(irq_base)) { + if (irq_base < 0) { WARN(1, "Cannot allocate irq_descs @ IRQ%d, assuming pre-allocated\n", irq_start); irq_base = irq_start; diff --git a/drivers/irqchip/irq-hip04.c b/drivers/irqchip/irq-hip04.c index 9688d2e2a636..9e25d8ce08e5 100644 --- a/drivers/irqchip/irq-hip04.c +++ b/drivers/irqchip/irq-hip04.c @@ -402,7 +402,7 @@ hip04_of_init(struct device_node *node, struct device_node *parent) nr_irqs -= hwirq_base; /* calculate # of irqs to allocate */ irq_base = irq_alloc_descs(-1, hwirq_base, nr_irqs, numa_node_id()); - if (IS_ERR_VALUE(irq_base)) { + if (irq_base < 0) { pr_err("failed to allocate IRQ numbers\n"); return -EINVAL; } diff --git a/drivers/irqchip/spear-shirq.c b/drivers/irqchip/spear-shirq.c index 1ccd2abed65f..1518ba31a80c 100644 --- a/drivers/irqchip/spear-shirq.c +++ b/drivers/irqchip/spear-shirq.c @@ -232,7 +232,7 @@ static int __init shirq_init(struct spear_shirq **shirq_blocks, int block_nr, nr_irqs += shirq_blocks[i]->nr_irqs; virq_base = irq_alloc_descs(-1, 0, nr_irqs, 0); - if (IS_ERR_VALUE(virq_base)) { + if (virq_base < 0) { pr_err("%s: irq desc alloc failed\n", __func__); goto err_unmap; } diff --git a/drivers/media/i2c/adp1653.c b/drivers/media/i2c/adp1653.c index 9e1731c565e7..e191e295c951 100644 --- a/drivers/media/i2c/adp1653.c +++ b/drivers/media/i2c/adp1653.c @@ -95,7 +95,7 @@ static int adp1653_get_fault(struct adp1653_flash *flash) int rval; fault = i2c_smbus_read_byte_data(client, ADP1653_REG_FAULT); - if (IS_ERR_VALUE(fault)) + if (fault < 0) return fault; flash->fault |= fault; @@ -105,13 +105,13 @@ static int adp1653_get_fault(struct adp1653_flash *flash) /* Clear faults. */ rval = i2c_smbus_write_byte_data(client, ADP1653_REG_OUT_SEL, 0); - if (IS_ERR_VALUE(rval)) + if (rval < 0) return rval; flash->led_mode->val = V4L2_FLASH_LED_MODE_NONE; rval = adp1653_update_hw(flash); - if (IS_ERR_VALUE(rval)) + if (rval) return rval; return flash->fault; @@ -158,7 +158,7 @@ static int adp1653_get_ctrl(struct v4l2_ctrl *ctrl) int rval; rval = adp1653_get_fault(flash); - if (IS_ERR_VALUE(rval)) + if (rval) return rval; ctrl->cur.val = 0; @@ -184,7 +184,7 @@ static int adp1653_set_ctrl(struct v4l2_ctrl *ctrl) int rval; rval = adp1653_get_fault(flash); - if (IS_ERR_VALUE(rval)) + if (rval) return rval; if ((rval & (ADP1653_REG_FAULT_FLT_SCP | ADP1653_REG_FAULT_FLT_OT | diff --git a/drivers/media/platform/s5p-tv/mixer_drv.c b/drivers/media/platform/s5p-tv/mixer_drv.c index 5ef67774971d..8a5d19469ddc 100644 --- a/drivers/media/platform/s5p-tv/mixer_drv.c +++ b/drivers/media/platform/s5p-tv/mixer_drv.c @@ -146,7 +146,7 @@ int mxr_power_get(struct mxr_device *mdev) /* returning 1 means that power is already enabled, * so zero success be returned */ - if (IS_ERR_VALUE(ret)) + if (ret < 0) return ret; return 0; } diff --git a/drivers/mfd/twl4030-irq.c b/drivers/mfd/twl4030-irq.c index 40e51b0baa46..b46c0cfc27d9 100644 --- a/drivers/mfd/twl4030-irq.c +++ b/drivers/mfd/twl4030-irq.c @@ -696,7 +696,7 @@ int twl4030_init_irq(struct device *dev, int irq_num) nr_irqs = TWL4030_PWR_NR_IRQS + TWL4030_CORE_NR_IRQS; irq_base = irq_alloc_descs(-1, 0, nr_irqs, 0); - if (IS_ERR_VALUE(irq_base)) { + if (irq_base < 0) { dev_err(dev, "Fail to allocate IRQ descs\n"); return irq_base; } diff --git a/drivers/mmc/core/mmc.c b/drivers/mmc/core/mmc.c index b81b08f81325..c984321d1881 100644 --- a/drivers/mmc/core/mmc.c +++ b/drivers/mmc/core/mmc.c @@ -1276,7 +1276,7 @@ static int mmc_select_hs200(struct mmc_card *card) * switch to HS200 mode if bus width is set successfully. */ err = mmc_select_bus_width(card); - if (!IS_ERR_VALUE(err)) { + if (!err) { val = EXT_CSD_TIMING_HS200 | card->drive_strength << EXT_CSD_DRV_STR_SHIFT; err = __mmc_switch(card, EXT_CSD_CMD_SET_NORMAL, @@ -1583,7 +1583,7 @@ static int mmc_init_card(struct mmc_host *host, u32 ocr, } else if (mmc_card_hs(card)) { /* Select the desired bus width optionally */ err = mmc_select_bus_width(card); - if (!IS_ERR_VALUE(err)) { + if (!err) { err = mmc_select_hs_ddr(card); if (err) goto free_card; diff --git a/drivers/mmc/host/dw_mmc.c b/drivers/mmc/host/dw_mmc.c index 829a6eebcdce..2cc6123b1df9 100644 --- a/drivers/mmc/host/dw_mmc.c +++ b/drivers/mmc/host/dw_mmc.c @@ -1431,7 +1431,7 @@ static int dw_mci_get_ro(struct mmc_host *mmc) int gpio_ro = mmc_gpio_get_ro(mmc); /* Use platform get_ro function, else try on board write protect */ - if (!IS_ERR_VALUE(gpio_ro)) + if (gpio_ro >= 0) read_only = gpio_ro; else read_only = @@ -1454,7 +1454,7 @@ static int dw_mci_get_cd(struct mmc_host *mmc) if ((mmc->caps & MMC_CAP_NEEDS_POLL) || (mmc->caps & MMC_CAP_NONREMOVABLE)) present = 1; - else if (!IS_ERR_VALUE(gpio_cd)) + else if (gpio_cd >= 0) present = gpio_cd; else present = (mci_readl(slot->host, CDETECT) & (1 << slot->id)) @@ -2927,7 +2927,7 @@ static void dw_mci_enable_cd(struct dw_mci *host) if (slot->mmc->caps & MMC_CAP_NEEDS_POLL) return; - if (IS_ERR_VALUE(mmc_gpio_get_cd(slot->mmc))) + if (mmc_gpio_get_cd(slot->mmc) < 0) break; } if (i == host->num_slots) diff --git a/drivers/mmc/host/sdhci-esdhc-imx.c b/drivers/mmc/host/sdhci-esdhc-imx.c index 2d300d87cda8..9d3ae1f4bd3c 100644 --- a/drivers/mmc/host/sdhci-esdhc-imx.c +++ b/drivers/mmc/host/sdhci-esdhc-imx.c @@ -1011,7 +1011,7 @@ sdhci_esdhc_imx_probe_dt(struct platform_device *pdev, if (ret) return ret; - if (!IS_ERR_VALUE(mmc_gpio_get_cd(host->mmc))) + if (mmc_gpio_get_cd(host->mmc) >= 0) host->quirks &= ~SDHCI_QUIRK_BROKEN_CARD_DETECTION; return 0; diff --git a/drivers/mmc/host/sdhci-of-at91.c b/drivers/mmc/host/sdhci-of-at91.c index 25f779e09d8e..d4cef713d246 100644 --- a/drivers/mmc/host/sdhci-of-at91.c +++ b/drivers/mmc/host/sdhci-of-at91.c @@ -289,7 +289,7 @@ static int sdhci_at91_probe(struct platform_device *pdev) * to enable polling via device tree with broken-cd property. */ if (!(host->mmc->caps & MMC_CAP_NONREMOVABLE) && - IS_ERR_VALUE(mmc_gpio_get_cd(host->mmc))) { + mmc_gpio_get_cd(host->mmc) < 0) { host->mmc->caps |= MMC_CAP_NEEDS_POLL; host->quirks &= ~SDHCI_QUIRK_BROKEN_CARD_DETECTION; } diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c index e010ea4eb6f5..0e3d7c056cb1 100644 --- a/drivers/mmc/host/sdhci.c +++ b/drivers/mmc/host/sdhci.c @@ -1624,7 +1624,7 @@ static int sdhci_get_cd(struct mmc_host *mmc) * Try slot gpio detect, if defined it take precedence * over build in controller functionality */ - if (!IS_ERR_VALUE(gpio_cd)) + if (gpio_cd >= 0) return !!gpio_cd; /* If polling, assume that the card is always present. */ @@ -3077,7 +3077,7 @@ int sdhci_add_host(struct sdhci_host *host) if ((host->quirks & SDHCI_QUIRK_BROKEN_CARD_DETECTION) && !(mmc->caps & MMC_CAP_NONREMOVABLE) && - IS_ERR_VALUE(mmc_gpio_get_cd(host->mmc))) + mmc_gpio_get_cd(host->mmc) < 0) mmc->caps |= MMC_CAP_NEEDS_POLL; /* If there are external regulators, get them */ diff --git a/drivers/net/ethernet/freescale/fman/fman.c b/drivers/net/ethernet/freescale/fman/fman.c index bcb9dccada4d..1de2e1e51c2b 100644 --- a/drivers/net/ethernet/freescale/fman/fman.c +++ b/drivers/net/ethernet/freescale/fman/fman.c @@ -615,7 +615,7 @@ struct fman { struct fman_cfg *cfg; struct muram_info *muram; /* cam section in muram */ - int cam_offset; + unsigned long cam_offset; size_t cam_size; /* Fifo in MURAM */ int fifo_offset; diff --git a/drivers/net/ethernet/freescale/fman/fman_muram.c b/drivers/net/ethernet/freescale/fman/fman_muram.c index 4eb0e9ac7182..47394c45b6e8 100644 --- a/drivers/net/ethernet/freescale/fman/fman_muram.c +++ b/drivers/net/ethernet/freescale/fman/fman_muram.c @@ -129,7 +129,7 @@ unsigned long fman_muram_offset_to_vbase(struct muram_info *muram, * * Return: address of the allocated memory; NULL otherwise. */ -int fman_muram_alloc(struct muram_info *muram, size_t size) +unsigned long fman_muram_alloc(struct muram_info *muram, size_t size) { unsigned long vaddr; @@ -150,7 +150,7 @@ int fman_muram_alloc(struct muram_info *muram, size_t size) * * Free an allocated memory from FM-MURAM partition. */ -void fman_muram_free_mem(struct muram_info *muram, u32 offset, size_t size) +void fman_muram_free_mem(struct muram_info *muram, unsigned long offset, size_t size) { unsigned long addr = fman_muram_offset_to_vbase(muram, offset); diff --git a/drivers/net/ethernet/freescale/fman/fman_muram.h b/drivers/net/ethernet/freescale/fman/fman_muram.h index dbf0af9e5bb5..889649ad8931 100644 --- a/drivers/net/ethernet/freescale/fman/fman_muram.h +++ b/drivers/net/ethernet/freescale/fman/fman_muram.h @@ -44,8 +44,8 @@ struct muram_info *fman_muram_init(phys_addr_t base, size_t size); unsigned long fman_muram_offset_to_vbase(struct muram_info *muram, unsigned long offset); -int fman_muram_alloc(struct muram_info *muram, size_t size); +unsigned long fman_muram_alloc(struct muram_info *muram, size_t size); -void fman_muram_free_mem(struct muram_info *muram, u32 offset, size_t size); +void fman_muram_free_mem(struct muram_info *muram, unsigned long offset, size_t size); #endif /* __FM_MURAM_EXT */ diff --git a/drivers/net/wireless/ti/wlcore/spi.c b/drivers/net/wireless/ti/wlcore/spi.c index 020ac1a4b408..cea9443c22a6 100644 --- a/drivers/net/wireless/ti/wlcore/spi.c +++ b/drivers/net/wireless/ti/wlcore/spi.c @@ -382,7 +382,7 @@ static int wlcore_probe_of(struct spi_device *spi, struct wl12xx_spi_glue *glue, ret = of_property_read_u32(dt_node, "ref-clock-frequency", &pdev_data->ref_clock_freq); - if (IS_ERR_VALUE(ret)) { + if (ret) { dev_err(glue->dev, "can't get reference clock frequency (%d)\n", ret); return ret; @@ -425,7 +425,7 @@ static int wl1271_probe(struct spi_device *spi) } ret = wlcore_probe_of(spi, glue, &pdev_data); - if (IS_ERR_VALUE(ret)) { + if (ret) { dev_err(glue->dev, "can't get device tree parameters (%d)\n", ret); return ret; diff --git a/drivers/nvmem/core.c b/drivers/nvmem/core.c index bb4ea123547f..965911d9b36a 100644 --- a/drivers/nvmem/core.c +++ b/drivers/nvmem/core.c @@ -113,7 +113,7 @@ static ssize_t bin_attr_nvmem_read(struct file *filp, struct kobject *kobj, rc = nvmem_reg_read(nvmem, pos, buf, count); - if (IS_ERR_VALUE(rc)) + if (rc) return rc; return count; @@ -147,7 +147,7 @@ static ssize_t bin_attr_nvmem_write(struct file *filp, struct kobject *kobj, rc = nvmem_reg_write(nvmem, pos, buf, count); - if (IS_ERR_VALUE(rc)) + if (rc) return rc; return count; @@ -366,7 +366,7 @@ static int nvmem_add_cells(struct nvmem_device *nvmem, } rval = nvmem_cell_info_to_nvmem_cell(nvmem, &info[i], cells[i]); - if (IS_ERR_VALUE(rval)) { + if (rval) { kfree(cells[i]); goto err; } @@ -963,7 +963,7 @@ static int __nvmem_cell_read(struct nvmem_device *nvmem, rc = nvmem_reg_read(nvmem, cell->offset, buf, cell->bytes); - if (IS_ERR_VALUE(rc)) + if (rc) return rc; /* shift bits in-place */ @@ -998,7 +998,7 @@ void *nvmem_cell_read(struct nvmem_cell *cell, size_t *len) return ERR_PTR(-ENOMEM); rc = __nvmem_cell_read(nvmem, cell, buf, len); - if (IS_ERR_VALUE(rc)) { + if (rc) { kfree(buf); return ERR_PTR(rc); } @@ -1083,7 +1083,7 @@ int nvmem_cell_write(struct nvmem_cell *cell, void *buf, size_t len) if (cell->bit_offset || cell->nbits) kfree(buf); - if (IS_ERR_VALUE(rc)) + if (rc) return rc; return len; @@ -1111,11 +1111,11 @@ ssize_t nvmem_device_cell_read(struct nvmem_device *nvmem, return -EINVAL; rc = nvmem_cell_info_to_nvmem_cell(nvmem, info, &cell); - if (IS_ERR_VALUE(rc)) + if (rc) return rc; rc = __nvmem_cell_read(nvmem, &cell, buf, &len); - if (IS_ERR_VALUE(rc)) + if (rc) return rc; return len; @@ -1141,7 +1141,7 @@ int nvmem_device_cell_write(struct nvmem_device *nvmem, return -EINVAL; rc = nvmem_cell_info_to_nvmem_cell(nvmem, info, &cell); - if (IS_ERR_VALUE(rc)) + if (rc) return rc; return nvmem_cell_write(&cell, buf, cell.bytes); @@ -1170,7 +1170,7 @@ int nvmem_device_read(struct nvmem_device *nvmem, rc = nvmem_reg_read(nvmem, offset, buf, bytes); - if (IS_ERR_VALUE(rc)) + if (rc) return rc; return bytes; @@ -1198,7 +1198,7 @@ int nvmem_device_write(struct nvmem_device *nvmem, rc = nvmem_reg_write(nvmem, offset, buf, bytes); - if (IS_ERR_VALUE(rc)) + if (rc) return rc; diff --git a/drivers/tty/serial/amba-pl011.c b/drivers/tty/serial/amba-pl011.c index a2aa655f56c4..1b7331e40d79 100644 --- a/drivers/tty/serial/amba-pl011.c +++ b/drivers/tty/serial/amba-pl011.c @@ -2360,7 +2360,7 @@ static int pl011_probe_dt_alias(int index, struct device *dev) return ret; ret = of_alias_get_id(np, "serial"); - if (IS_ERR_VALUE(ret)) { + if (ret < 0) { seen_dev_without_alias = true; ret = index; } else { diff --git a/drivers/tty/serial/sprd_serial.c b/drivers/tty/serial/sprd_serial.c index 18971063f95f..699447aa8b43 100644 --- a/drivers/tty/serial/sprd_serial.c +++ b/drivers/tty/serial/sprd_serial.c @@ -654,7 +654,7 @@ static int sprd_probe_dt_alias(int index, struct device *dev) return ret; ret = of_alias_get_id(np, "serial"); - if (IS_ERR_VALUE(ret)) + if (ret < 0) ret = index; else if (ret >= ARRAY_SIZE(sprd_port) || sprd_port[ret] != NULL) { dev_warn(dev, "requested serial port %d not available.\n", ret); diff --git a/drivers/video/fbdev/da8xx-fb.c b/drivers/video/fbdev/da8xx-fb.c index d8d583d32a37..c229b1a0d13b 100644 --- a/drivers/video/fbdev/da8xx-fb.c +++ b/drivers/video/fbdev/da8xx-fb.c @@ -713,7 +713,7 @@ static int da8xx_fb_config_clk_divider(struct da8xx_fb_par *par, if (par->lcdc_clk_rate != lcdc_clk_rate) { ret = clk_set_rate(par->lcdc_clk, lcdc_clk_rate); - if (IS_ERR_VALUE(ret)) { + if (ret) { dev_err(par->dev, "unable to set clock rate at %u\n", lcdc_clk_rate); @@ -784,7 +784,7 @@ static int lcd_init(struct da8xx_fb_par *par, const struct lcd_ctrl_config *cfg, int ret = 0; ret = da8xx_fb_calc_config_clk_divider(par, panel); - if (IS_ERR_VALUE(ret)) { + if (ret) { dev_err(par->dev, "unable to configure clock\n"); return ret; } diff --git a/fs/afs/write.c b/fs/afs/write.c index 65de439bdc4f..14d506efd1aa 100644 --- a/fs/afs/write.c +++ b/fs/afs/write.c @@ -643,10 +643,6 @@ ssize_t afs_file_write(struct kiocb *iocb, struct iov_iter *from) return 0; result = generic_file_write_iter(iocb, from); - if (IS_ERR_VALUE(result)) { - _leave(" = %zd", result); - return result; - } _leave(" = %zd", result); return result; diff --git a/fs/binfmt_flat.c b/fs/binfmt_flat.c index f723cd3a455c..caf9e39bb82b 100644 --- a/fs/binfmt_flat.c +++ b/fs/binfmt_flat.c @@ -337,7 +337,7 @@ calc_reloc(unsigned long r, struct lib_info *p, int curid, int internalp) "(%d != %d)", (unsigned) r, curid, id); goto failed; } else if ( ! p->lib_list[id].loaded && - IS_ERR_VALUE(load_flat_shared_library(id, p))) { + load_flat_shared_library(id, p) < 0) { printk("BINFMT_FLAT: failed to load library %d", id); goto failed; } @@ -837,7 +837,7 @@ static int load_flat_shared_library(int id, struct lib_info *libs) res = prepare_binprm(&bprm); - if (!IS_ERR_VALUE(res)) + if (!res) res = load_flat_file(&bprm, libs, id, NULL); abort_creds(bprm.cred); @@ -883,7 +883,7 @@ static int load_flat_binary(struct linux_binprm * bprm) stack_len += FLAT_STACK_ALIGN - 1; /* reserve for upcoming alignment */ res = load_flat_file(bprm, &libinfo, 0, &stack_len); - if (IS_ERR_VALUE(res)) + if (res < 0) return res; /* Update data segment pointers for all libraries */ diff --git a/fs/gfs2/dir.c b/fs/gfs2/dir.c index 4a01f30e9995..271d93905bac 100644 --- a/fs/gfs2/dir.c +++ b/fs/gfs2/dir.c @@ -783,12 +783,15 @@ static int get_leaf_nr(struct gfs2_inode *dip, u32 index, u64 *leaf_out) { __be64 *hash; + int error; hash = gfs2_dir_get_hash_table(dip); - if (IS_ERR(hash)) - return PTR_ERR(hash); - *leaf_out = be64_to_cpu(*(hash + index)); - return 0; + error = PTR_ERR_OR_ZERO(hash); + + if (!error) + *leaf_out = be64_to_cpu(*(hash + index)); + + return error; } static int get_first_leaf(struct gfs2_inode *dip, u32 index, @@ -798,7 +801,7 @@ static int get_first_leaf(struct gfs2_inode *dip, u32 index, int error; error = get_leaf_nr(dip, index, &leaf_no); - if (!IS_ERR_VALUE(error)) + if (!error) error = get_leaf(dip, leaf_no, bh_out); return error; @@ -1014,7 +1017,7 @@ static int dir_split_leaf(struct inode *inode, const struct qstr *name) index = name->hash >> (32 - dip->i_depth); error = get_leaf_nr(dip, index, &leaf_no); - if (IS_ERR_VALUE(error)) + if (error) return error; /* Get the old leaf block */ diff --git a/kernel/pid.c b/kernel/pid.c index 4d73a834c7e6..f66162f2359b 100644 --- a/kernel/pid.c +++ b/kernel/pid.c @@ -311,7 +311,7 @@ struct pid *alloc_pid(struct pid_namespace *ns) pid->level = ns->level; for (i = ns->level; i >= 0; i--) { nr = alloc_pidmap(tmp); - if (IS_ERR_VALUE(nr)) { + if (nr < 0) { retval = nr; goto out_free; } diff --git a/net/9p/client.c b/net/9p/client.c index ea79ee9a7348..3fc94a49ccd5 100644 --- a/net/9p/client.c +++ b/net/9p/client.c @@ -518,10 +518,10 @@ static int p9_check_errors(struct p9_client *c, struct p9_req_t *req) if (err) goto out_err; - if (p9_is_proto_dotu(c)) + if (p9_is_proto_dotu(c) && ecode < 512) err = -ecode; - if (!err || !IS_ERR_VALUE(err)) { + if (!err) { err = p9_errstr2errno(ename, strlen(ename)); p9_debug(P9_DEBUG_9P, "<<< RERROR (%d) %s\n", @@ -605,10 +605,10 @@ static int p9_check_zc_errors(struct p9_client *c, struct p9_req_t *req, if (err) goto out_err; - if (p9_is_proto_dotu(c)) + if (p9_is_proto_dotu(c) && ecode < 512) err = -ecode; - if (!err || !IS_ERR_VALUE(err)) { + if (!err) { err = p9_errstr2errno(ename, strlen(ename)); p9_debug(P9_DEBUG_9P, "<<< RERROR (%d) %s\n", diff --git a/sound/soc/qcom/lpass-platform.c b/sound/soc/qcom/lpass-platform.c index 6e8665430bd5..ddfe34434765 100644 --- a/sound/soc/qcom/lpass-platform.c +++ b/sound/soc/qcom/lpass-platform.c @@ -491,7 +491,7 @@ static int lpass_platform_pcm_new(struct snd_soc_pcm_runtime *soc_runtime) data->rdma_ch = v->alloc_dma_channel(drvdata, SNDRV_PCM_STREAM_PLAYBACK); - if (IS_ERR_VALUE(data->rdma_ch)) + if (data->rdma_ch < 0) return data->rdma_ch; drvdata->substream[data->rdma_ch] = psubstream; @@ -518,7 +518,7 @@ static int lpass_platform_pcm_new(struct snd_soc_pcm_runtime *soc_runtime) data->wrdma_ch = v->alloc_dma_channel(drvdata, SNDRV_PCM_STREAM_CAPTURE); - if (IS_ERR_VALUE(data->wrdma_ch)) + if (data->wrdma_ch < 0) goto capture_alloc_err; drvdata->substream[data->wrdma_ch] = csubstream; -- cgit v1.3-8-gc7d7 From 5d22fc25d4fc8096d2d7df27ea1893d4e055e764 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Fri, 27 May 2016 15:57:31 -0700 Subject: mm: remove more IS_ERR_VALUE abuses The do_brk() and vm_brk() return value was "unsigned long" and returned the starting address on success, and an error value on failure. The reasons are entirely historical, and go back to it basically behaving like the mmap() interface does. However, nobody actually wanted that interface, and it causes totally pointless IS_ERR_VALUE() confusion. What every single caller actually wants is just the simpler integer return of zero for success and negative error number on failure. So just convert to that much clearer and more common calling convention, and get rid of all the IS_ERR_VALUE() uses wrt vm_brk(). Signed-off-by: Linus Torvalds --- arch/x86/ia32/ia32_aout.c | 17 ++++++++--------- fs/binfmt_aout.c | 18 +++++++----------- fs/binfmt_elf.c | 11 +++++------ include/linux/mm.h | 2 +- mm/mmap.c | 16 ++++++++-------- mm/nommu.c | 2 +- 6 files changed, 30 insertions(+), 36 deletions(-) (limited to 'fs') diff --git a/arch/x86/ia32/ia32_aout.c b/arch/x86/ia32/ia32_aout.c index f5e737ff0022..cb26f18d43af 100644 --- a/arch/x86/ia32/ia32_aout.c +++ b/arch/x86/ia32/ia32_aout.c @@ -116,12 +116,12 @@ static struct linux_binfmt aout_format = { .min_coredump = PAGE_SIZE }; -static unsigned long set_brk(unsigned long start, unsigned long end) +static int set_brk(unsigned long start, unsigned long end) { start = PAGE_ALIGN(start); end = PAGE_ALIGN(end); if (end <= start) - return start; + return 0; return vm_brk(start, end - start); } @@ -321,7 +321,7 @@ static int load_aout_binary(struct linux_binprm *bprm) error = vm_brk(text_addr & PAGE_MASK, map_size); - if (error != (text_addr & PAGE_MASK)) + if (error) return error; error = read_code(bprm->file, text_addr, 32, @@ -350,7 +350,7 @@ static int load_aout_binary(struct linux_binprm *bprm) if (!bprm->file->f_op->mmap || (fd_offset & ~PAGE_MASK) != 0) { error = vm_brk(N_TXTADDR(ex), ex.a_text+ex.a_data); - if (IS_ERR_VALUE(error)) + if (error) return error; read_code(bprm->file, N_TXTADDR(ex), fd_offset, @@ -378,7 +378,7 @@ static int load_aout_binary(struct linux_binprm *bprm) beyond_if: error = set_brk(current->mm->start_brk, current->mm->brk); - if (IS_ERR_VALUE(error)) + if (error) return error; set_binfmt(&aout_format); @@ -441,7 +441,7 @@ static int load_aout_library(struct file *file) } #endif retval = vm_brk(start_addr, ex.a_text + ex.a_data + ex.a_bss); - if (IS_ERR_VALUE(retval)) + if (retval) goto out; read_code(file, start_addr, N_TXTOFF(ex), @@ -461,9 +461,8 @@ static int load_aout_library(struct file *file) len = PAGE_ALIGN(ex.a_text + ex.a_data); bss = ex.a_text + ex.a_data + ex.a_bss; if (bss > len) { - error = vm_brk(start_addr + len, bss - len); - retval = error; - if (error != start_addr + len) + retval = vm_brk(start_addr + len, bss - len); + if (retval) goto out; } retval = 0; diff --git a/fs/binfmt_aout.c b/fs/binfmt_aout.c index 2fab9f130e51..64b331ae3428 100644 --- a/fs/binfmt_aout.c +++ b/fs/binfmt_aout.c @@ -127,11 +127,8 @@ static int set_brk(unsigned long start, unsigned long end) { start = PAGE_ALIGN(start); end = PAGE_ALIGN(end); - if (end > start) { - unsigned long addr; - addr = vm_brk(start, end - start); - if (BAD_ADDR(addr)) - return addr; + if (end > start) + return vm_brk(start, end - start); } return 0; } @@ -275,7 +272,7 @@ static int load_aout_binary(struct linux_binprm * bprm) map_size = ex.a_text+ex.a_data; #endif error = vm_brk(text_addr & PAGE_MASK, map_size); - if (error != (text_addr & PAGE_MASK)) + if (error) return error; error = read_code(bprm->file, text_addr, pos, @@ -298,7 +295,7 @@ static int load_aout_binary(struct linux_binprm * bprm) if (!bprm->file->f_op->mmap||((fd_offset & ~PAGE_MASK) != 0)) { error = vm_brk(N_TXTADDR(ex), ex.a_text+ex.a_data); - if (IS_ERR_VALUE(error)) + if (error) return error; read_code(bprm->file, N_TXTADDR(ex), fd_offset, @@ -382,7 +379,7 @@ static int load_aout_library(struct file *file) file); } retval = vm_brk(start_addr, ex.a_text + ex.a_data + ex.a_bss); - if (IS_ERR_VALUE(retval)) + if (retval) goto out; read_code(file, start_addr, N_TXTOFF(ex), @@ -402,9 +399,8 @@ static int load_aout_library(struct file *file) len = PAGE_ALIGN(ex.a_text + ex.a_data); bss = ex.a_text + ex.a_data + ex.a_bss; if (bss > len) { - error = vm_brk(start_addr + len, bss - len); - retval = error; - if (error != start_addr + len) + retval = vm_brk(start_addr + len, bss - len); + if (retval) goto out; } retval = 0; diff --git a/fs/binfmt_elf.c b/fs/binfmt_elf.c index 938fc4ede764..e158b22ef32f 100644 --- a/fs/binfmt_elf.c +++ b/fs/binfmt_elf.c @@ -96,10 +96,9 @@ static int set_brk(unsigned long start, unsigned long end) start = ELF_PAGEALIGN(start); end = ELF_PAGEALIGN(end); if (end > start) { - unsigned long addr; - addr = vm_brk(start, end - start); - if (BAD_ADDR(addr)) - return addr; + int error = vm_brk(start, end - start); + if (error) + return error; } current->mm->start_brk = current->mm->brk = end; return 0; @@ -629,7 +628,7 @@ static unsigned long load_elf_interp(struct elfhdr *interp_elf_ex, /* Map the last of the bss segment */ error = vm_brk(elf_bss, last_bss - elf_bss); - if (BAD_ADDR(error)) + if (error) goto out; } @@ -1178,7 +1177,7 @@ static int load_elf_library(struct file *file) bss = eppnt->p_memsz + eppnt->p_vaddr; if (bss > len) { error = vm_brk(len, bss - len); - if (BAD_ADDR(error)) + if (error) goto out_free_ph; } error = 0; diff --git a/include/linux/mm.h b/include/linux/mm.h index a00ec816233a..5df5feb49575 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -2018,7 +2018,7 @@ static inline void mm_populate(unsigned long addr, unsigned long len) {} #endif /* These take the mm semaphore themselves */ -extern unsigned long __must_check vm_brk(unsigned long, unsigned long); +extern int __must_check vm_brk(unsigned long, unsigned long); extern int vm_munmap(unsigned long, size_t); extern unsigned long __must_check vm_mmap(struct file *, unsigned long, unsigned long, unsigned long, diff --git a/mm/mmap.c b/mm/mmap.c index d3d9a94ca031..de2c1769cc68 100644 --- a/mm/mmap.c +++ b/mm/mmap.c @@ -168,7 +168,7 @@ static struct vm_area_struct *remove_vma(struct vm_area_struct *vma) return next; } -static unsigned long do_brk(unsigned long addr, unsigned long len); +static int do_brk(unsigned long addr, unsigned long len); SYSCALL_DEFINE1(brk, unsigned long, brk) { @@ -224,7 +224,7 @@ SYSCALL_DEFINE1(brk, unsigned long, brk) goto out; /* Ok, looks good - let it rip. */ - if (do_brk(oldbrk, newbrk-oldbrk) != oldbrk) + if (do_brk(oldbrk, newbrk-oldbrk) < 0) goto out; set_brk: @@ -2625,7 +2625,7 @@ static inline void verify_mm_writelocked(struct mm_struct *mm) * anonymous maps. eventually we may be able to do some * brk-specific accounting here. */ -static unsigned long do_brk(unsigned long addr, unsigned long len) +static int do_brk(unsigned long addr, unsigned long len) { struct mm_struct *mm = current->mm; struct vm_area_struct *vma, *prev; @@ -2636,7 +2636,7 @@ static unsigned long do_brk(unsigned long addr, unsigned long len) len = PAGE_ALIGN(len); if (!len) - return addr; + return 0; flags = VM_DATA_DEFAULT_FLAGS | VM_ACCOUNT | mm->def_flags; @@ -2703,13 +2703,13 @@ out: if (flags & VM_LOCKED) mm->locked_vm += (len >> PAGE_SHIFT); vma->vm_flags |= VM_SOFTDIRTY; - return addr; + return 0; } -unsigned long vm_brk(unsigned long addr, unsigned long len) +int vm_brk(unsigned long addr, unsigned long len) { struct mm_struct *mm = current->mm; - unsigned long ret; + int ret; bool populate; if (down_write_killable(&mm->mmap_sem)) @@ -2718,7 +2718,7 @@ unsigned long vm_brk(unsigned long addr, unsigned long len) ret = do_brk(addr, len); populate = ((mm->def_flags & VM_LOCKED) != 0); up_write(&mm->mmap_sem); - if (populate) + if (populate && !ret) mm_populate(addr, len); return ret; } diff --git a/mm/nommu.c b/mm/nommu.c index c8bd59a03c71..c2e58880207f 100644 --- a/mm/nommu.c +++ b/mm/nommu.c @@ -1682,7 +1682,7 @@ void exit_mmap(struct mm_struct *mm) } } -unsigned long vm_brk(unsigned long addr, unsigned long len) +int vm_brk(unsigned long addr, unsigned long len) { return -ENOMEM; } -- cgit v1.3-8-gc7d7 From 3767e255b390d72f9a33c08d9e86c5f21f25860f Mon Sep 17 00:00:00 2001 From: Al Viro Date: Fri, 27 May 2016 11:06:05 -0400 Subject: switch ->setxattr() to passing dentry and inode separately smack ->d_instantiate() uses ->setxattr(), so to be able to call it before we'd hashed the new dentry and attached it to inode, we need ->setxattr() instances getting the inode as an explicit argument rather than obtaining it from dentry. Similar change for ->getxattr() had been done in commit ce23e64. Unlike ->getxattr() (which is used by both selinux and smack instances of ->d_instantiate()) ->setxattr() is used only by smack one and unfortunately it got missed back then. Reported-by: Seung-Woo Kim Tested-by: Casey Schaufler Signed-off-by: Al Viro --- Documentation/filesystems/porting | 7 +++++++ drivers/staging/lustre/lustre/llite/llite_internal.h | 4 ++-- drivers/staging/lustre/lustre/llite/xattr.c | 6 ++---- fs/bad_inode.c | 4 ++-- fs/ecryptfs/crypto.c | 9 +++++---- fs/ecryptfs/ecryptfs_kernel.h | 4 ++-- fs/ecryptfs/inode.c | 7 ++++--- fs/ecryptfs/mmap.c | 3 ++- fs/fuse/dir.c | 6 +++--- fs/hfs/attr.c | 6 +++--- fs/hfs/hfs_fs.h | 2 +- fs/kernfs/inode.c | 11 ++++++----- fs/kernfs/kernfs-internal.h | 3 ++- fs/libfs.c | 5 +++-- fs/overlayfs/inode.c | 5 +++-- fs/overlayfs/overlayfs.h | 5 +++-- fs/xattr.c | 8 ++++---- include/linux/fs.h | 3 ++- include/linux/xattr.h | 3 ++- security/smack/smack_lsm.c | 2 +- 20 files changed, 59 insertions(+), 44 deletions(-) (limited to 'fs') diff --git a/Documentation/filesystems/porting b/Documentation/filesystems/porting index 46f3bb7a02f5..a5fb89cac615 100644 --- a/Documentation/filesystems/porting +++ b/Documentation/filesystems/porting @@ -578,3 +578,10 @@ in your dentry operations instead. -- [mandatory] ->atomic_open() calls without O_CREAT may happen in parallel. +-- +[mandatory] + ->setxattr() and xattr_handler.set() get dentry and inode passed separately. + dentry might be yet to be attached to inode, so do _not_ use its ->d_inode + in the instances. Rationale: !@#!@# security_d_instantiate() needs to be + called before we attach dentry to inode and !@#!@##!@$!$#!@#$!@$!@$ smack + ->d_instantiate() uses not just ->getxattr() but ->setxattr() as well. diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h index ce1f949430f1..3f2f30b6542c 100644 --- a/drivers/staging/lustre/lustre/llite/llite_internal.h +++ b/drivers/staging/lustre/lustre/llite/llite_internal.h @@ -976,8 +976,8 @@ static inline __u64 ll_file_maxbytes(struct inode *inode) } /* llite/xattr.c */ -int ll_setxattr(struct dentry *dentry, const char *name, - const void *value, size_t size, int flags); +int ll_setxattr(struct dentry *dentry, struct inode *inode, + const char *name, const void *value, size_t size, int flags); ssize_t ll_getxattr(struct dentry *dentry, struct inode *inode, const char *name, void *buffer, size_t size); ssize_t ll_listxattr(struct dentry *dentry, char *buffer, size_t size); diff --git a/drivers/staging/lustre/lustre/llite/xattr.c b/drivers/staging/lustre/lustre/llite/xattr.c index ed4de04381c3..608014b0dbcd 100644 --- a/drivers/staging/lustre/lustre/llite/xattr.c +++ b/drivers/staging/lustre/lustre/llite/xattr.c @@ -211,11 +211,9 @@ int ll_setxattr_common(struct inode *inode, const char *name, return 0; } -int ll_setxattr(struct dentry *dentry, const char *name, - const void *value, size_t size, int flags) +int ll_setxattr(struct dentry *dentry, struct inode *inode, + const char *name, const void *value, size_t size, int flags) { - struct inode *inode = d_inode(dentry); - LASSERT(inode); LASSERT(name); diff --git a/fs/bad_inode.c b/fs/bad_inode.c index 72e35b721608..3ba385eaa26e 100644 --- a/fs/bad_inode.c +++ b/fs/bad_inode.c @@ -100,8 +100,8 @@ static int bad_inode_setattr(struct dentry *direntry, struct iattr *attrs) return -EIO; } -static int bad_inode_setxattr(struct dentry *dentry, const char *name, - const void *value, size_t size, int flags) +static int bad_inode_setxattr(struct dentry *dentry, struct inode *inode, + const char *name, const void *value, size_t size, int flags) { return -EIO; } diff --git a/fs/ecryptfs/crypto.c b/fs/ecryptfs/crypto.c index ebd40f46ed4c..0d8eb3455b34 100644 --- a/fs/ecryptfs/crypto.c +++ b/fs/ecryptfs/crypto.c @@ -1141,12 +1141,13 @@ ecryptfs_write_metadata_to_contents(struct inode *ecryptfs_inode, static int ecryptfs_write_metadata_to_xattr(struct dentry *ecryptfs_dentry, + struct inode *ecryptfs_inode, char *page_virt, size_t size) { int rc; - rc = ecryptfs_setxattr(ecryptfs_dentry, ECRYPTFS_XATTR_NAME, page_virt, - size, 0); + rc = ecryptfs_setxattr(ecryptfs_dentry, ecryptfs_inode, + ECRYPTFS_XATTR_NAME, page_virt, size, 0); return rc; } @@ -1215,8 +1216,8 @@ int ecryptfs_write_metadata(struct dentry *ecryptfs_dentry, goto out_free; } if (crypt_stat->flags & ECRYPTFS_METADATA_IN_XATTR) - rc = ecryptfs_write_metadata_to_xattr(ecryptfs_dentry, virt, - size); + rc = ecryptfs_write_metadata_to_xattr(ecryptfs_dentry, ecryptfs_inode, + virt, size); else rc = ecryptfs_write_metadata_to_contents(ecryptfs_inode, virt, virt_len); diff --git a/fs/ecryptfs/ecryptfs_kernel.h b/fs/ecryptfs/ecryptfs_kernel.h index 3ec495db7e82..4ba1547bb9ad 100644 --- a/fs/ecryptfs/ecryptfs_kernel.h +++ b/fs/ecryptfs/ecryptfs_kernel.h @@ -609,8 +609,8 @@ ssize_t ecryptfs_getxattr_lower(struct dentry *lower_dentry, struct inode *lower_inode, const char *name, void *value, size_t size); int -ecryptfs_setxattr(struct dentry *dentry, const char *name, const void *value, - size_t size, int flags); +ecryptfs_setxattr(struct dentry *dentry, struct inode *inode, const char *name, + const void *value, size_t size, int flags); int ecryptfs_read_xattr_region(char *page_virt, struct inode *ecryptfs_inode); #ifdef CONFIG_ECRYPT_FS_MESSAGING int ecryptfs_process_response(struct ecryptfs_daemon *daemon, diff --git a/fs/ecryptfs/inode.c b/fs/ecryptfs/inode.c index 318b04689d76..9d153b6a1d72 100644 --- a/fs/ecryptfs/inode.c +++ b/fs/ecryptfs/inode.c @@ -1001,7 +1001,8 @@ static int ecryptfs_getattr(struct vfsmount *mnt, struct dentry *dentry, } int -ecryptfs_setxattr(struct dentry *dentry, const char *name, const void *value, +ecryptfs_setxattr(struct dentry *dentry, struct inode *inode, + const char *name, const void *value, size_t size, int flags) { int rc = 0; @@ -1014,8 +1015,8 @@ ecryptfs_setxattr(struct dentry *dentry, const char *name, const void *value, } rc = vfs_setxattr(lower_dentry, name, value, size, flags); - if (!rc && d_really_is_positive(dentry)) - fsstack_copy_attr_all(d_inode(dentry), d_inode(lower_dentry)); + if (!rc && inode) + fsstack_copy_attr_all(inode, d_inode(lower_dentry)); out: return rc; } diff --git a/fs/ecryptfs/mmap.c b/fs/ecryptfs/mmap.c index 148d11b514fb..9c3437c8a5b1 100644 --- a/fs/ecryptfs/mmap.c +++ b/fs/ecryptfs/mmap.c @@ -442,7 +442,8 @@ static int ecryptfs_write_inode_size_to_xattr(struct inode *ecryptfs_inode) if (size < 0) size = 8; put_unaligned_be64(i_size_read(ecryptfs_inode), xattr_virt); - rc = lower_inode->i_op->setxattr(lower_dentry, ECRYPTFS_XATTR_NAME, + rc = lower_inode->i_op->setxattr(lower_dentry, lower_inode, + ECRYPTFS_XATTR_NAME, xattr_virt, size, 0); inode_unlock(lower_inode); if (rc) diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c index b9419058108f..ccd4971cc6c1 100644 --- a/fs/fuse/dir.c +++ b/fs/fuse/dir.c @@ -1719,10 +1719,10 @@ static int fuse_getattr(struct vfsmount *mnt, struct dentry *entry, return fuse_update_attributes(inode, stat, NULL, NULL); } -static int fuse_setxattr(struct dentry *entry, const char *name, - const void *value, size_t size, int flags) +static int fuse_setxattr(struct dentry *unused, struct inode *inode, + const char *name, const void *value, + size_t size, int flags) { - struct inode *inode = d_inode(entry); struct fuse_conn *fc = get_fuse_conn(inode); FUSE_ARGS(args); struct fuse_setxattr_in inarg; diff --git a/fs/hfs/attr.c b/fs/hfs/attr.c index 064f92f17efc..d9a86919fdf6 100644 --- a/fs/hfs/attr.c +++ b/fs/hfs/attr.c @@ -13,10 +13,10 @@ #include "hfs_fs.h" #include "btree.h" -int hfs_setxattr(struct dentry *dentry, const char *name, - const void *value, size_t size, int flags) +int hfs_setxattr(struct dentry *unused, struct inode *inode, + const char *name, const void *value, + size_t size, int flags) { - struct inode *inode = d_inode(dentry); struct hfs_find_data fd; hfs_cat_rec rec; struct hfs_cat_file *file; diff --git a/fs/hfs/hfs_fs.h b/fs/hfs/hfs_fs.h index fa3eed86837c..ee2f385811c8 100644 --- a/fs/hfs/hfs_fs.h +++ b/fs/hfs/hfs_fs.h @@ -212,7 +212,7 @@ extern void hfs_evict_inode(struct inode *); extern void hfs_delete_inode(struct inode *); /* attr.c */ -extern int hfs_setxattr(struct dentry *dentry, const char *name, +extern int hfs_setxattr(struct dentry *dentry, struct inode *inode, const char *name, const void *value, size_t size, int flags); extern ssize_t hfs_getxattr(struct dentry *dentry, struct inode *inode, const char *name, void *value, size_t size); diff --git a/fs/kernfs/inode.c b/fs/kernfs/inode.c index 1719649d7ad7..63b925d5ba1e 100644 --- a/fs/kernfs/inode.c +++ b/fs/kernfs/inode.c @@ -160,10 +160,11 @@ static int kernfs_node_setsecdata(struct kernfs_node *kn, void **secdata, return 0; } -int kernfs_iop_setxattr(struct dentry *dentry, const char *name, - const void *value, size_t size, int flags) +int kernfs_iop_setxattr(struct dentry *unused, struct inode *inode, + const char *name, const void *value, + size_t size, int flags) { - struct kernfs_node *kn = dentry->d_fsdata; + struct kernfs_node *kn = inode->i_private; struct kernfs_iattrs *attrs; void *secdata; int error; @@ -175,11 +176,11 @@ int kernfs_iop_setxattr(struct dentry *dentry, const char *name, if (!strncmp(name, XATTR_SECURITY_PREFIX, XATTR_SECURITY_PREFIX_LEN)) { const char *suffix = name + XATTR_SECURITY_PREFIX_LEN; - error = security_inode_setsecurity(d_inode(dentry), suffix, + error = security_inode_setsecurity(inode, suffix, value, size, flags); if (error) return error; - error = security_inode_getsecctx(d_inode(dentry), + error = security_inode_getsecctx(inode, &secdata, &secdata_len); if (error) return error; diff --git a/fs/kernfs/kernfs-internal.h b/fs/kernfs/kernfs-internal.h index 45c9192c276e..37159235ac10 100644 --- a/fs/kernfs/kernfs-internal.h +++ b/fs/kernfs/kernfs-internal.h @@ -81,7 +81,8 @@ int kernfs_iop_permission(struct inode *inode, int mask); int kernfs_iop_setattr(struct dentry *dentry, struct iattr *iattr); int kernfs_iop_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat); -int kernfs_iop_setxattr(struct dentry *dentry, const char *name, const void *value, +int kernfs_iop_setxattr(struct dentry *dentry, struct inode *inode, + const char *name, const void *value, size_t size, int flags); int kernfs_iop_removexattr(struct dentry *dentry, const char *name); ssize_t kernfs_iop_getxattr(struct dentry *dentry, struct inode *inode, diff --git a/fs/libfs.c b/fs/libfs.c index 8765ff1adc07..3db2721144c2 100644 --- a/fs/libfs.c +++ b/fs/libfs.c @@ -1118,8 +1118,9 @@ static int empty_dir_setattr(struct dentry *dentry, struct iattr *attr) return -EPERM; } -static int empty_dir_setxattr(struct dentry *dentry, const char *name, - const void *value, size_t size, int flags) +static int empty_dir_setxattr(struct dentry *dentry, struct inode *inode, + const char *name, const void *value, + size_t size, int flags) { return -EOPNOTSUPP; } diff --git a/fs/overlayfs/inode.c b/fs/overlayfs/inode.c index c7b31a03dc9c..0ed7c4012437 100644 --- a/fs/overlayfs/inode.c +++ b/fs/overlayfs/inode.c @@ -210,8 +210,9 @@ static bool ovl_is_private_xattr(const char *name) return strncmp(name, OVL_XATTR_PRE_NAME, OVL_XATTR_PRE_LEN) == 0; } -int ovl_setxattr(struct dentry *dentry, const char *name, - const void *value, size_t size, int flags) +int ovl_setxattr(struct dentry *dentry, struct inode *inode, + const char *name, const void *value, + size_t size, int flags) { int err; struct dentry *upperdentry; diff --git a/fs/overlayfs/overlayfs.h b/fs/overlayfs/overlayfs.h index 99ec4b035237..d79577eb3937 100644 --- a/fs/overlayfs/overlayfs.h +++ b/fs/overlayfs/overlayfs.h @@ -171,8 +171,9 @@ int ovl_check_d_type_supported(struct path *realpath); /* inode.c */ int ovl_setattr(struct dentry *dentry, struct iattr *attr); int ovl_permission(struct inode *inode, int mask); -int ovl_setxattr(struct dentry *dentry, const char *name, - const void *value, size_t size, int flags); +int ovl_setxattr(struct dentry *dentry, struct inode *inode, + const char *name, const void *value, + size_t size, int flags); ssize_t ovl_getxattr(struct dentry *dentry, struct inode *inode, const char *name, void *value, size_t size); ssize_t ovl_listxattr(struct dentry *dentry, char *list, size_t size); diff --git a/fs/xattr.c b/fs/xattr.c index b16d07889700..4beafc43daa5 100644 --- a/fs/xattr.c +++ b/fs/xattr.c @@ -100,7 +100,7 @@ int __vfs_setxattr_noperm(struct dentry *dentry, const char *name, if (issec) inode->i_flags &= ~S_NOSEC; if (inode->i_op->setxattr) { - error = inode->i_op->setxattr(dentry, name, value, size, flags); + error = inode->i_op->setxattr(dentry, inode, name, value, size, flags); if (!error) { fsnotify_xattr(dentry); security_inode_post_setxattr(dentry, name, value, @@ -745,7 +745,8 @@ generic_listxattr(struct dentry *dentry, char *buffer, size_t buffer_size) * Find the handler for the prefix and dispatch its set() operation. */ int -generic_setxattr(struct dentry *dentry, const char *name, const void *value, size_t size, int flags) +generic_setxattr(struct dentry *dentry, struct inode *inode, const char *name, + const void *value, size_t size, int flags) { const struct xattr_handler *handler; @@ -754,8 +755,7 @@ generic_setxattr(struct dentry *dentry, const char *name, const void *value, siz handler = xattr_resolve_name(dentry->d_sb->s_xattr, &name); if (IS_ERR(handler)) return PTR_ERR(handler); - return handler->set(handler, dentry, d_inode(dentry), name, value, - size, flags); + return handler->set(handler, dentry, inode, name, value, size, flags); } /* diff --git a/include/linux/fs.h b/include/linux/fs.h index 5f61431d8673..62bdb0a6cf2d 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1730,7 +1730,8 @@ struct inode_operations { struct inode *, struct dentry *, unsigned int); int (*setattr) (struct dentry *, struct iattr *); int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *); - int (*setxattr) (struct dentry *, const char *,const void *,size_t,int); + int (*setxattr) (struct dentry *, struct inode *, + const char *, const void *, size_t, int); ssize_t (*getxattr) (struct dentry *, struct inode *, const char *, void *, size_t); ssize_t (*listxattr) (struct dentry *, char *, size_t); diff --git a/include/linux/xattr.h b/include/linux/xattr.h index 76beb206741a..94079bab9243 100644 --- a/include/linux/xattr.h +++ b/include/linux/xattr.h @@ -54,7 +54,8 @@ int vfs_removexattr(struct dentry *, const char *); ssize_t generic_getxattr(struct dentry *dentry, struct inode *inode, const char *name, void *buffer, size_t size); ssize_t generic_listxattr(struct dentry *dentry, char *buffer, size_t buffer_size); -int generic_setxattr(struct dentry *dentry, const char *name, const void *value, size_t size, int flags); +int generic_setxattr(struct dentry *dentry, struct inode *inode, + const char *name, const void *value, size_t size, int flags); int generic_removexattr(struct dentry *dentry, const char *name); ssize_t vfs_getxattr_alloc(struct dentry *dentry, const char *name, char **xattr_value, size_t size, gfp_t flags); diff --git a/security/smack/smack_lsm.c b/security/smack/smack_lsm.c index ff2b8c3cf7a9..6777295f4b2b 100644 --- a/security/smack/smack_lsm.c +++ b/security/smack/smack_lsm.c @@ -3514,7 +3514,7 @@ static void smack_d_instantiate(struct dentry *opt_dentry, struct inode *inode) */ if (isp->smk_flags & SMK_INODE_CHANGED) { isp->smk_flags &= ~SMK_INODE_CHANGED; - rc = inode->i_op->setxattr(dp, + rc = inode->i_op->setxattr(dp, inode, XATTR_NAME_SMACKTRANSMUTE, TRANS_TRUE, TRANS_TRUE_SIZE, 0); -- cgit v1.3-8-gc7d7 From e0714ec4f9efe7b86828b0dcc077fd8f5d8e5e91 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Fri, 27 May 2016 17:20:27 -0700 Subject: nfs: fix anonymous member initializer build failure with older compilers MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Older versions of gcc don't understand named initializers inside a anonymous structure or union member. It can be worked around by adding the bracin gin the initializer for the anonymous member. Without this, gcc 4.4.4 will fail the build with CC fs/nfs/nfs4state.o fs/nfs/nfs4state.c:69: error: unknown field ‘data’ specified in initializer fs/nfs/nfs4state.c:69: warning: missing braces around initializer fs/nfs/nfs4state.c:69: warning: (near initialization for ‘zero_stateid..data’) make[2]: *** [fs/nfs/nfs4state.o] Error 1 introduced in commit 93b717fd81bf ("NFSv4: Label stateids with the type") Reported-and-tested-by: Boris Ostrovsky Cc: Anna Schumaker Cc: Trond Myklebust Signed-off-by: Linus Torvalds --- fs/nfs/nfs4state.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c index 5075592df145..9679f4749364 100644 --- a/fs/nfs/nfs4state.c +++ b/fs/nfs/nfs4state.c @@ -66,7 +66,7 @@ #define OPENOWNER_POOL_SIZE 8 const nfs4_stateid zero_stateid = { - .data = { 0 }, + { .data = { 0 } }, .type = NFS4_SPECIAL_STATEID_TYPE, }; static DEFINE_MUTEX(nfs_clid_init_mutex); -- cgit v1.3-8-gc7d7 From fcfd2fbf22d2587196890103d41e3d554c47da0e Mon Sep 17 00:00:00 2001 From: George Spelvin Date: Fri, 20 May 2016 08:41:37 -0400 Subject: fs/namei.c: Add hashlen_string() function We'd like to make more use of the highly-optimized dcache hash functions throughout the kernel, rather than have every subsystem create its own, and a function that hashes basic null-terminated strings is required for that. (The name is to emphasize that it returns both hash and length.) It's actually useful in the dcache itself, specifically d_alloc_name(). Other uses in the next patch. full_name_hash() is also tweaked to make it more generally useful: 1) Take a "char *" rather than "unsigned char *" argument, to be consistent with hash_name(). 2) Handle zero-length inputs. If we want more callers, we don't want to make them worry about corner cases. Signed-off-by: George Spelvin --- fs/dcache.c | 3 +-- fs/namei.c | 51 +++++++++++++++++++++++++++++++++++++++++----- include/linux/stringhash.h | 8 ++++++-- 3 files changed, 53 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/dcache.c b/fs/dcache.c index d5ecc6e477da..19b751806789 100644 --- a/fs/dcache.c +++ b/fs/dcache.c @@ -1653,8 +1653,7 @@ struct dentry *d_alloc_name(struct dentry *parent, const char *name) struct qstr q; q.name = name; - q.len = strlen(name); - q.hash = full_name_hash(q.name, q.len); + q.hash_len = hashlen_string(name); return d_alloc(parent, &q); } EXPORT_SYMBOL(d_alloc_name); diff --git a/fs/namei.c b/fs/namei.c index 42f8ca038254..dd98d43a54f8 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -1822,19 +1822,20 @@ static inline unsigned long mix_hash(unsigned long hash) #endif -unsigned int full_name_hash(const unsigned char *name, unsigned int len) +/* Return the hash of a string of known length */ +unsigned int full_name_hash(const char *name, unsigned int len) { unsigned long a, hash = 0; for (;;) { + if (!len) + goto done; a = load_unaligned_zeropad(name); if (len < sizeof(unsigned long)) break; hash = mix_hash(hash + a); name += sizeof(unsigned long); len -= sizeof(unsigned long); - if (!len) - goto done; } hash += a & bytemask_from_count(len); done: @@ -1842,6 +1843,29 @@ done: } EXPORT_SYMBOL(full_name_hash); +/* Return the "hash_len" (hash and length) of a null-terminated string */ +u64 hashlen_string(const char *name) +{ + unsigned long a, adata, mask, hash, len; + const struct word_at_a_time constants = WORD_AT_A_TIME_CONSTANTS; + + hash = a = 0; + len = -sizeof(unsigned long); + do { + hash = mix_hash(hash + a); + len += sizeof(unsigned long); + a = load_unaligned_zeropad(name+len); + } while (!has_zero(a, &adata, &constants)); + + adata = prep_zero_mask(a, adata, &constants); + mask = create_zero_mask(adata); + hash += a & zero_bytemask(mask); + len += find_zero(mask); + + return hashlen_create(fold_hash(hash), len); +} +EXPORT_SYMBOL(hashlen_string); + /* * Calculate the length and hash of the path component, and * return the "hash_len" as the result. @@ -1872,15 +1896,32 @@ static inline u64 hash_name(const char *name) #else -unsigned int full_name_hash(const unsigned char *name, unsigned int len) +/* Return the hash of a string of known length */ +unsigned int full_name_hash(const char *name, unsigned int len) { unsigned long hash = init_name_hash(); while (len--) - hash = partial_name_hash(*name++, hash); + hash = partial_name_hash((unsigned char)*name++, hash); return end_name_hash(hash); } EXPORT_SYMBOL(full_name_hash); +/* Return the "hash_len" (hash and length) of a null-terminated string */ +u64 hash_string(const char *name) +{ + unsigned long hash = init_name_hash(); + unsigned long len = 0, c; + + c = (unsigned char)*name; + do { + len++; + hash = partial_name_hash(c, hash); + c = (unsigned char)name[len]; + } while (c); + return hashlen_create(end_name_hash(hash), len); +} +EXPORT_SYMBOL(hash_string); + /* * We know there's a real path component here of at least * one character. diff --git a/include/linux/stringhash.h b/include/linux/stringhash.h index 2eaaaf6d2776..451771d9b9c0 100644 --- a/include/linux/stringhash.h +++ b/include/linux/stringhash.h @@ -1,7 +1,8 @@ #ifndef __LINUX_STRINGHASH_H #define __LINUX_STRINGHASH_H -#include +#include /* For __pure */ +#include /* For u32, u64 */ /* * Routines for hashing strings of bytes to a 32-bit hash value. @@ -59,7 +60,7 @@ static inline unsigned long end_name_hash(unsigned long hash) * * If not set, this falls back to a wrapper around the preceding. */ -extern unsigned int full_name_hash(const unsigned char *, unsigned int); +extern unsigned int __pure full_name_hash(const char *, unsigned int); /* * A hash_len is a u64 with the hash of a string in the low @@ -69,4 +70,7 @@ extern unsigned int full_name_hash(const unsigned char *, unsigned int); #define hashlen_len(hashlen) ((u32)((hashlen) >> 32)) #define hashlen_create(hash, len) ((u64)(len)<<32 | (u32)(hash)) +/* Return the "hash_len" (hash and length) of a null-terminated string */ +extern u64 __pure hashlen_string(const char *name); + #endif /* __LINUX_STRINGHASH_H */ -- cgit v1.3-8-gc7d7 From 2a18da7a9c7886f1c7307f8d3f23f24318583f03 Mon Sep 17 00:00:00 2001 From: George Spelvin Date: Mon, 23 May 2016 07:43:58 -0400 Subject: fs/namei.c: Improve dcache hash function Patch 0fed3ac866 improved the hash mixing, but the function is slower than necessary; there's a 7-instruction dependency chain (10 on x86) each loop iteration. Word-at-a-time access is a very tight loop (which is good, because link_path_walk() is one of the hottest code paths in the entire kernel), and the hash mixing function must not have a longer latency to avoid slowing it down. There do not appear to be any published fast hash functions that: 1) Operate on the input a word at a time, and 2) Don't need to know the length of the input beforehand, and 3) Have a single iterated mixing function, not needing conditional branches or unrolling to distinguish different loop iterations. One of the algorithms which comes closest is Yann Collet's xxHash, but that's two dependent multiplies per word, which is too much. The key insights in this design are: 1) Barring expensive ops like multiplies, to diffuse one input bit across 64 bits of hash state takes at least log2(64) = 6 sequentially dependent instructions. That is more cycles than we'd like. 2) An operation like "hash ^= hash << 13" requires a second temporary register anyway, and on a 2-operand machine like x86, it's three instructions. 3) A better use of a second register is to hold a two-word hash state. With careful design, no temporaries are needed at all, so it doesn't increase register pressure. And this gets rid of register copying on 2-operand machines, so the code is smaller and faster. 4) Using two words of state weakens the requirement for one-round mixing; we now have two rounds of mixing before cancellation is possible. 5) A two-word hash state also allows operations on both halves to be done in parallel, so on a superscalar processor we get more mixing in fewer cycles. I ended up using a mixing function inspired by the ChaCha and Speck round functions. It is 6 simple instructions and 3 cycles per iteration (assuming multiply by 9 can be done by an "lea" instruction): x ^= *input++; y ^= x; x = ROL(x, K1); x += y; y = ROL(y, K2); y *= 9; Not only is this reversible, two consecutive rounds are reversible: if you are given the initial and final states, but not the intermediate state, it is possible to compute both input words. This means that at least 3 words of input are required to create a collision. (It also has the property, used by hash_name() to avoid a branch, that it hashes all-zero to all-zero.) The rotate constants K1 and K2 were found by experiment. The search took a sample of random initial states (I used 1023) and considered the effect of flipping each of the 64 input bits on each of the 128 output bits two rounds later. Each of the 8192 pairs can be considered a biased coin, and adding up the Shannon entropy of all of them produces a score. The best-scoring shifts also did well in other tests (flipping bits in y, trying 3 or 4 rounds of mixing, flipping all 64*63/2 pairs of input bits), so the choice was made with the additional constraint that the sum of the shifts is odd and not too close to the word size. The final state is then folded into a 32-bit hash value by a less carefully optimized multiply-based scheme. This also has to be fast, as pathname components tend to be short (the most common case is one iteration!), but there's some room for latency, as there is a fair bit of intervening logic before the hash value is used for anything. (Performance verified with "bonnie++ -s 0 -n 1536:-2" on tmpfs. I need a better benchmark; the numbers seem to show a slight dip in performance between 4.6.0 and this patch, but they're too noisy to quote.) Special thanks to Bruce fields for diligent testing which uncovered a nasty fencepost error in an earlier version of this patch. [checkpatch.pl formatting complaints noted and respectfully disagreed with.] Signed-off-by: George Spelvin Tested-by: J. Bruce Fields --- fs/namei.c | 121 +++++++++++++++++++++++++++++++++++++++++-------------------- 1 file changed, 81 insertions(+), 40 deletions(-) (limited to 'fs') diff --git a/fs/namei.c b/fs/namei.c index dd98d43a54f8..a49cbd7efcaa 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -35,6 +35,7 @@ #include #include #include +#include #include #include "internal.h" @@ -1788,44 +1789,89 @@ static int walk_component(struct nameidata *nd, int flags) #include #ifdef CONFIG_64BIT - -static inline unsigned int fold_hash(unsigned long hash) -{ - return hash_64(hash, 32); -} +/* + * Register pressure in the mixing function is an issue, particularly + * on 32-bit x86, but almost any function requires one state value and + * one temporary. Instead, use a function designed for two state values + * and no temporaries. + * + * This function cannot create a collision in only two iterations, so + * we have two iterations to achieve avalanche. In those two iterations, + * we have six layers of mixing, which is enough to spread one bit's + * influence out to 2^6 = 64 state bits. + * + * Rotate constants are scored by considering either 64 one-bit input + * deltas or 64*63/2 = 2016 two-bit input deltas, and finding the + * probability of that delta causing a change to each of the 128 output + * bits, using a sample of random initial states. + * + * The Shannon entropy of the computed probabilities is then summed + * to produce a score. Ideally, any input change has a 50% chance of + * toggling any given output bit. + * + * Mixing scores (in bits) for (12,45): + * Input delta: 1-bit 2-bit + * 1 round: 713.3 42542.6 + * 2 rounds: 2753.7 140389.8 + * 3 rounds: 5954.1 233458.2 + * 4 rounds: 7862.6 256672.2 + * Perfect: 8192 258048 + * (64*128) (64*63/2 * 128) + */ +#define HASH_MIX(x, y, a) \ + ( x ^= (a), \ + y ^= x, x = rol64(x,12),\ + x += y, y = rol64(y,45),\ + y *= 9 ) /* - * This is George Marsaglia's XORSHIFT generator. - * It implements a maximum-period LFSR in only a few - * instructions. It also has the property (required - * by hash_name()) that mix_hash(0) = 0. + * Fold two longs into one 32-bit hash value. This must be fast, but + * latency isn't quite as critical, as there is a fair bit of additional + * work done before the hash value is used. */ -static inline unsigned long mix_hash(unsigned long hash) +static inline unsigned int fold_hash(unsigned long x, unsigned long y) { - hash ^= hash << 13; - hash ^= hash >> 7; - hash ^= hash << 17; - return hash; + y ^= x * GOLDEN_RATIO_64; + y *= GOLDEN_RATIO_64; + return y >> 32; } #else /* 32-bit case */ -#define fold_hash(x) (x) +/* + * Mixing scores (in bits) for (7,20): + * Input delta: 1-bit 2-bit + * 1 round: 330.3 9201.6 + * 2 rounds: 1246.4 25475.4 + * 3 rounds: 1907.1 31295.1 + * 4 rounds: 2042.3 31718.6 + * Perfect: 2048 31744 + * (32*64) (32*31/2 * 64) + */ +#define HASH_MIX(x, y, a) \ + ( x ^= (a), \ + y ^= x, x = rol32(x, 7),\ + x += y, y = rol32(y,20),\ + y *= 9 ) -static inline unsigned long mix_hash(unsigned long hash) +static inline unsigned int fold_hash(unsigned long x, unsigned long y) { - hash ^= hash << 13; - hash ^= hash >> 17; - hash ^= hash << 5; - return hash; + /* Use arch-optimized multiply if one exists */ + return __hash_32(y ^ __hash_32(x)); } #endif -/* Return the hash of a string of known length */ +/* + * Return the hash of a string of known length. This is carfully + * designed to match hash_name(), which is the more critical function. + * In particular, we must end by hashing a final word containing 0..7 + * payload bytes, to match the way that hash_name() iterates until it + * finds the delimiter after the name. + */ unsigned int full_name_hash(const char *name, unsigned int len) { - unsigned long a, hash = 0; + unsigned long a, x = 0, y = 0; for (;;) { if (!len) @@ -1833,36 +1879,34 @@ unsigned int full_name_hash(const char *name, unsigned int len) a = load_unaligned_zeropad(name); if (len < sizeof(unsigned long)) break; - hash = mix_hash(hash + a); + HASH_MIX(x, y, a); name += sizeof(unsigned long); len -= sizeof(unsigned long); } - hash += a & bytemask_from_count(len); + x ^= a & bytemask_from_count(len); done: - return fold_hash(hash); + return fold_hash(x, y); } EXPORT_SYMBOL(full_name_hash); /* Return the "hash_len" (hash and length) of a null-terminated string */ u64 hashlen_string(const char *name) { - unsigned long a, adata, mask, hash, len; + unsigned long a = 0, x = 0, y = 0, adata, mask, len; const struct word_at_a_time constants = WORD_AT_A_TIME_CONSTANTS; - hash = a = 0; len = -sizeof(unsigned long); do { - hash = mix_hash(hash + a); + HASH_MIX(x, y, a); len += sizeof(unsigned long); a = load_unaligned_zeropad(name+len); } while (!has_zero(a, &adata, &constants)); adata = prep_zero_mask(a, adata, &constants); mask = create_zero_mask(adata); - hash += a & zero_bytemask(mask); - len += find_zero(mask); + x ^= a & zero_bytemask(mask); - return hashlen_create(fold_hash(hash), len); + return hashlen_create(fold_hash(x, y), len + find_zero(mask)); } EXPORT_SYMBOL(hashlen_string); @@ -1872,13 +1916,12 @@ EXPORT_SYMBOL(hashlen_string); */ static inline u64 hash_name(const char *name) { - unsigned long a, b, adata, bdata, mask, hash, len; + unsigned long a = 0, b, x = 0, y = 0, adata, bdata, mask, len; const struct word_at_a_time constants = WORD_AT_A_TIME_CONSTANTS; - hash = a = 0; len = -sizeof(unsigned long); do { - hash = mix_hash(hash + a); + HASH_MIX(x, y, a); len += sizeof(unsigned long); a = load_unaligned_zeropad(name+len); b = a ^ REPEAT_BYTE('/'); @@ -1886,15 +1929,13 @@ static inline u64 hash_name(const char *name) adata = prep_zero_mask(a, adata, &constants); bdata = prep_zero_mask(b, bdata, &constants); - mask = create_zero_mask(adata | bdata); + x ^= a & zero_bytemask(mask); - hash += a & zero_bytemask(mask); - len += find_zero(mask); - return hashlen_create(fold_hash(hash), len); + return hashlen_create(fold_hash(x, y), len + find_zero(mask)); } -#else +#else /* !CONFIG_DCACHE_WORD_ACCESS: Slow, byte-at-a-time version */ /* Return the hash of a string of known length */ unsigned int full_name_hash(const char *name, unsigned int len) @@ -1965,7 +2006,7 @@ static int link_path_walk(const char *name, struct nameidata *nd) int type; err = may_lookup(nd); - if (err) + if (err) return err; hash_len = hash_name(name); -- cgit v1.3-8-gc7d7 From 468a9428521e7d00fb21250af363eb94dc1d6861 Mon Sep 17 00:00:00 2001 From: George Spelvin Date: Thu, 26 May 2016 22:11:51 -0400 Subject: : Add support for architecture-specific functions This is just the infrastructure; there are no users yet. This is modelled on CONFIG_ARCH_RANDOM; a CONFIG_ symbol declares the existence of . That file may define its own versions of various functions, and define HAVE_* symbols (no CONFIG_ prefix!) to suppress the generic ones. Included is a self-test (in lib/test_hash.c) that verifies the basics. It is NOT in general required that the arch-specific functions compute the same thing as the generic, but if a HAVE_* symbol is defined with the value 1, then equality is tested. Signed-off-by: George Spelvin Cc: Geert Uytterhoeven Cc: Greg Ungerer Cc: Andreas Schwab Cc: Philippe De Muyter Cc: linux-m68k@lists.linux-m68k.org Cc: Alistair Francis Cc: Michal Simek Cc: Yoshinori Sato Cc: uclinux-h8-devel@lists.sourceforge.jp --- arch/Kconfig | 8 ++ fs/namei.c | 6 +- include/linux/hash.h | 27 +++++- lib/Kconfig.debug | 11 +++ lib/Makefile | 1 + lib/test_hash.c | 250 +++++++++++++++++++++++++++++++++++++++++++++++++++ 6 files changed, 299 insertions(+), 4 deletions(-) create mode 100644 lib/test_hash.c (limited to 'fs') diff --git a/arch/Kconfig b/arch/Kconfig index 81869a5e7e17..96406e4db995 100644 --- a/arch/Kconfig +++ b/arch/Kconfig @@ -589,6 +589,14 @@ config HAVE_STACK_VALIDATION Architecture supports the 'objtool check' host tool command, which performs compile-time stack metadata validation. +config HAVE_ARCH_HASH + bool + default n + help + If this is set, the architecture provides an + file which provides platform-specific implementations of some + functions in or fs/namei.c. + # # ABI hall of shame # diff --git a/fs/namei.c b/fs/namei.c index a49cbd7efcaa..968dae025230 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -1788,7 +1788,11 @@ static int walk_component(struct nameidata *nd, int flags) #include -#ifdef CONFIG_64BIT +#ifdef HASH_MIX + +/* Architecture provides HASH_MIX and fold_hash() in */ + +#elif defined(CONFIG_64BIT) /* * Register pressure in the mixing function is an issue, particularly * on 32-bit x86, but almost any function requires one state value and diff --git a/include/linux/hash.h b/include/linux/hash.h index 613cfde3a1e0..ad6fa21d977b 100644 --- a/include/linux/hash.h +++ b/include/linux/hash.h @@ -41,19 +41,40 @@ #define GOLDEN_RATIO_32 0x61C88647 #define GOLDEN_RATIO_64 0x61C8864680B583EBull +#ifdef CONFIG_HAVE_ARCH_HASH +/* This header may use the GOLDEN_RATIO_xx constants */ +#include +#endif -static inline u32 __hash_32(u32 val) +/* + * The _generic versions exist only so lib/test_hash.c can compare + * the arch-optimized versions with the generic. + * + * Note that if you change these, any that aren't updated + * to match need to have their HAVE_ARCH_* define values updated so the + * self-test will not false-positive. + */ +#ifndef HAVE_ARCH__HASH_32 +#define __hash_32 __hash_32_generic +#endif +static inline u32 __hash_32_generic(u32 val) { return val * GOLDEN_RATIO_32; } -static inline u32 hash_32(u32 val, unsigned int bits) +#ifndef HAVE_ARCH_HASH_32 +#define hash_32 hash_32_generic +#endif +static inline u32 hash_32_generic(u32 val, unsigned int bits) { /* High bits are more random, so use them. */ return __hash_32(val) >> (32 - bits); } -static __always_inline u32 hash_64(u64 val, unsigned int bits) +#ifndef HAVE_ARCH_HASH_64 +#define hash_64 hash_64_generic +#endif +static __always_inline u32 hash_64_generic(u64 val, unsigned int bits) { #if BITS_PER_LONG == 64 /* 64x64-bit multiply is efficient on all 64-bit processors */ diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index 1e9a607534ca..18ec69ba8eb6 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -1815,6 +1815,17 @@ config TEST_RHASHTABLE If unsure, say N. +config TEST_HASH + tristate "Perform selftest on hash functions" + default n + help + Enable this option to test the kernel's integer () + and string () hash functions on boot + (or module load). + + This is intended to help people writing architecture-specific + optimized versions. If unsure, say N. + endmenu # runtime tests config PROVIDE_OHCI1394_DMA_INIT diff --git a/lib/Makefile b/lib/Makefile index 7bd6fd436c97..f80b1a1b3afd 100644 --- a/lib/Makefile +++ b/lib/Makefile @@ -48,6 +48,7 @@ obj-$(CONFIG_TEST_HEXDUMP) += test_hexdump.o obj-y += kstrtox.o obj-$(CONFIG_TEST_BPF) += test_bpf.o obj-$(CONFIG_TEST_FIRMWARE) += test_firmware.o +obj-$(CONFIG_TEST_HASH) += test_hash.o obj-$(CONFIG_TEST_KASAN) += test_kasan.o obj-$(CONFIG_TEST_KSTRTOX) += test-kstrtox.o obj-$(CONFIG_TEST_LKM) += test_module.o diff --git a/lib/test_hash.c b/lib/test_hash.c new file mode 100644 index 000000000000..c9549c8b4909 --- /dev/null +++ b/lib/test_hash.c @@ -0,0 +1,250 @@ +/* + * Test cases for and + * This just verifies that various ways of computing a hash + * produce the same thing and, for cases where a k-bit hash + * value is requested, is of the requested size. + * + * We fill a buffer with a 255-byte null-terminated string, + * and use both full_name_hash() and hashlen_string() to hash the + * substrings from i to j, where 0 <= i < j < 256. + * + * The returned values are used to check that __hash_32() and + * __hash_32_generic() compute the same thing. Likewise hash_32() + * and hash_64(). + */ + +#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt "\n" + +#include +#include +#include +#include +#include +#include + +/* 32-bit XORSHIFT generator. Seed must not be zero. */ +static u32 __init __attribute_const__ +xorshift(u32 seed) +{ + seed ^= seed << 13; + seed ^= seed >> 17; + seed ^= seed << 5; + return seed; +} + +/* Given a non-zero x, returns a non-zero byte. */ +static u8 __init __attribute_const__ +mod255(u32 x) +{ + x = (x & 0xffff) + (x >> 16); /* 1 <= x <= 0x1fffe */ + x = (x & 0xff) + (x >> 8); /* 1 <= x <= 0x2fd */ + x = (x & 0xff) + (x >> 8); /* 1 <= x <= 0x100 */ + x = (x & 0xff) + (x >> 8); /* 1 <= x <= 0xff */ + return x; +} + +/* Fill the buffer with non-zero bytes. */ +static void __init +fill_buf(char *buf, size_t len, u32 seed) +{ + size_t i; + + for (i = 0; i < len; i++) { + seed = xorshift(seed); + buf[i] = mod255(seed); + } +} + +/* + * Test the various integer hash functions. h64 (or its low-order bits) + * is the integer to hash. hash_or accumulates the OR of the hash values, + * which are later checked to see that they cover all the requested bits. + * + * Because these functions (as opposed to the string hashes) are all + * inline, the code being tested is actually in the module, and you can + * recompile and re-test the module without rebooting. + */ +static bool __init +test_int_hash(unsigned long long h64, u32 hash_or[2][33]) +{ + int k; + u32 h0 = (u32)h64, h1, h2; + + /* Test __hash32 */ + hash_or[0][0] |= h1 = __hash_32(h0); +#ifdef HAVE_ARCH__HASH_32 + hash_or[1][0] |= h2 = __hash_32_generic(h0); +#if HAVE_ARCH__HASH_32 == 1 + if (h1 != h2) { + pr_err("__hash_32(%#x) = %#x != __hash_32_generic() = %#x", + h0, h1, h2); + return false; + } +#endif +#endif + + /* Test k = 1..32 bits */ + for (k = 1; k <= 32; k++) { + u32 const m = ((u32)2 << (k-1)) - 1; /* Low k bits set */ + + /* Test hash_32 */ + hash_or[0][k] |= h1 = hash_32(h0, k); + if (h1 > m) { + pr_err("hash_32(%#x, %d) = %#x > %#x", h0, k, h1, m); + return false; + } +#ifdef HAVE_ARCH_HASH_32 + h2 = hash_32_generic(h0, k); +#if HAVE_ARCH_HASH_32 == 1 + if (h1 != h2) { + pr_err("hash_32(%#x, %d) = %#x != hash_32_generic() " + " = %#x", h0, k, h1, h2); + return false; + } +#else + if (h2 > m) { + pr_err("hash_32_generic(%#x, %d) = %#x > %#x", + h0, k, h1, m); + return false; + } +#endif +#endif + /* Test hash_64 */ + hash_or[1][k] |= h1 = hash_64(h64, k); + if (h1 > m) { + pr_err("hash_64(%#llx, %d) = %#x > %#x", h64, k, h1, m); + return false; + } +#ifdef HAVE_ARCH_HASH_64 + h2 = hash_64_generic(h64, k); +#if HAVE_ARCH_HASH_64 == 1 + if (h1 != h2) { + pr_err("hash_64(%#llx, %d) = %#x != hash_64_generic() " + "= %#x", h64, k, h1, h2); + return false; + } +#else + if (h2 > m) { + pr_err("hash_64_generic(%#llx, %d) = %#x > %#x", + h64, k, h1, m); + return false; + } +#endif +#endif + } + + (void)h2; /* Suppress unused variable warning */ + return true; +} + +#define SIZE 256 /* Run time is cubic in SIZE */ + +static int __init +test_hash_init(void) +{ + char buf[SIZE+1]; + u32 string_or = 0, hash_or[2][33] = { 0 }; + unsigned tests = 0; + unsigned long long h64 = 0; + int i, j; + + fill_buf(buf, SIZE, 1); + + /* Test every possible non-empty substring in the buffer. */ + for (j = SIZE; j > 0; --j) { + buf[j] = '\0'; + + for (i = 0; i <= j; i++) { + u64 hashlen = hashlen_string(buf+i); + u32 h0 = full_name_hash(buf+i, j-i); + + /* Check that hashlen_string gets the length right */ + if (hashlen_len(hashlen) != j-i) { + pr_err("hashlen_string(%d..%d) returned length" + " %u, expected %d", + i, j, hashlen_len(hashlen), j-i); + return -EINVAL; + } + /* Check that the hashes match */ + if (hashlen_hash(hashlen) != h0) { + pr_err("hashlen_string(%d..%d) = %08x != " + "full_name_hash() = %08x", + i, j, hashlen_hash(hashlen), h0); + return -EINVAL; + } + + string_or |= h0; + h64 = h64 << 32 | h0; /* For use with hash_64 */ + if (!test_int_hash(h64, hash_or)) + return -EINVAL; + tests++; + } /* i */ + } /* j */ + + /* The OR of all the hash values should cover all the bits */ + if (~string_or) { + pr_err("OR of all string hash results = %#x != %#x", + string_or, -1u); + return -EINVAL; + } + if (~hash_or[0][0]) { + pr_err("OR of all __hash_32 results = %#x != %#x", + hash_or[0][0], -1u); + return -EINVAL; + } +#ifdef HAVE_ARCH__HASH_32 +#if HAVE_ARCH__HASH_32 != 1 /* Test is pointless if results match */ + if (~hash_or[1][0]) { + pr_err("OR of all __hash_32_generic results = %#x != %#x", + hash_or[1][0], -1u); + return -EINVAL; + } +#endif +#endif + + /* Likewise for all the i-bit hash values */ + for (i = 1; i <= 32; i++) { + u32 const m = ((u32)2 << (i-1)) - 1; /* Low i bits set */ + + if (hash_or[0][i] != m) { + pr_err("OR of all hash_32(%d) results = %#x " + "(%#x expected)", i, hash_or[0][i], m); + return -EINVAL; + } + if (hash_or[1][i] != m) { + pr_err("OR of all hash_64(%d) results = %#x " + "(%#x expected)", i, hash_or[1][i], m); + return -EINVAL; + } + } + + /* Issue notices about skipped tests. */ +#ifndef HAVE_ARCH__HASH_32 + pr_info("__hash_32() has no arch implementation to test."); +#elif HAVE_ARCH__HASH_32 != 1 + pr_info("__hash_32() is arch-specific; not compared to generic."); +#endif +#ifndef HAVE_ARCH_HASH_32 + pr_info("hash_32() has no arch implementation to test."); +#elif HAVE_ARCH_HASH_32 != 1 + pr_info("hash_32() is arch-specific; not compared to generic."); +#endif +#ifndef HAVE_ARCH_HASH_64 + pr_info("hash_64() has no arch implementation to test."); +#elif HAVE_ARCH_HASH_64 != 1 + pr_info("hash_64() is arch-specific; not compared to generic."); +#endif + + pr_notice("%u tests passed.", tests); + + return 0; +} + +static void __exit test_hash_exit(void) +{ +} + +module_init(test_hash_init); /* Does everything */ +module_exit(test_hash_exit); /* Does nothing */ + +MODULE_LICENSE("GPL"); -- cgit v1.3-8-gc7d7 From d66492bce151ac4c477cf1de97171777c0d62d20 Mon Sep 17 00:00:00 2001 From: Guenter Roeck Date: Sat, 28 May 2016 15:26:02 -0700 Subject: fs: fix binfmt_aout.c build error Various builds (such as i386:allmodconfig) fail with fs/binfmt_aout.c:133:2: error: expected identifier or '(' before 'return' fs/binfmt_aout.c:134:1: error: expected identifier or '(' before '}' token [ Oops. My bad, I had stupidly thought that "allmodconfig" covered this on x86-64 too, but it obviously doesn't. Egg on my face. - Linus ] Fixes: 5d22fc25d4fc ("mm: remove more IS_ERR_VALUE abuses") Signed-off-by: Guenter Roeck Signed-off-by: Linus Torvalds --- fs/binfmt_aout.c | 1 - 1 file changed, 1 deletion(-) (limited to 'fs') diff --git a/fs/binfmt_aout.c b/fs/binfmt_aout.c index 64b331ae3428..ae1b5404fced 100644 --- a/fs/binfmt_aout.c +++ b/fs/binfmt_aout.c @@ -129,7 +129,6 @@ static int set_brk(unsigned long start, unsigned long end) end = PAGE_ALIGN(end); if (end > start) return vm_brk(start, end - start); - } return 0; } -- cgit v1.3-8-gc7d7 From 44d51706b4685f965cd32acde3fe0fcc1e6198e8 Mon Sep 17 00:00:00 2001 From: Mikulas Patocka Date: Tue, 24 May 2016 22:47:00 +0200 Subject: hpfs: fix remount failure when there are no options changed Commit ce657611baf9 ("hpfs: kstrdup() out of memory handling") checks if the kstrdup function returns NULL due to out-of-memory condition. However, if we are remounting a filesystem with no change to filesystem-specific options, the parameter data is NULL. In this case, kstrdup returns NULL (because it was passed NULL parameter), although no out of memory condition exists. The mount syscall then fails with ENOMEM. This patch fixes the bug. We fail with ENOMEM only if data is non-NULL. The patch also changes the call to replace_mount_options - if we didn't pass any filesystem-specific options, we don't call replace_mount_options (thus we don't erase existing reported options). Fixes: ce657611baf9 ("hpfs: kstrdup() out of memory handling") Signed-off-by: Mikulas Patocka Cc: stable@vger.kernel.org Signed-off-by: Linus Torvalds --- fs/hpfs/super.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/hpfs/super.c b/fs/hpfs/super.c index 458cf463047b..46ad0ef8b4cc 100644 --- a/fs/hpfs/super.c +++ b/fs/hpfs/super.c @@ -455,7 +455,7 @@ static int hpfs_remount_fs(struct super_block *s, int *flags, char *data) struct hpfs_sb_info *sbi = hpfs_sb(s); char *new_opts = kstrdup(data, GFP_KERNEL); - if (!new_opts) + if (data && !new_opts) return -ENOMEM; sync_filesystem(s); @@ -493,7 +493,8 @@ static int hpfs_remount_fs(struct super_block *s, int *flags, char *data) if (!(*flags & MS_RDONLY)) mark_dirty(s, 1); - replace_mount_options(s, new_opts); + if (new_opts) + replace_mount_options(s, new_opts); hpfs_unlock(s); return 0; -- cgit v1.3-8-gc7d7 From 01d6e08711bf90bc4d7ead14a93a0cbd73b1896a Mon Sep 17 00:00:00 2001 From: Mikulas Patocka Date: Tue, 24 May 2016 22:48:33 +0200 Subject: affs: fix remount failure when there are no options changed Commit c8f33d0bec99 ("affs: kstrdup() memory handling") checks if the kstrdup function returns NULL due to out-of-memory condition. However, if we are remounting a filesystem with no change to filesystem-specific options, the parameter data is NULL. In this case, kstrdup returns NULL (because it was passed NULL parameter), although no out of memory condition exists. The mount syscall then fails with ENOMEM. This patch fixes the bug. We fail with ENOMEM only if data is non-NULL. The patch also changes the call to replace_mount_options - if we didn't pass any filesystem-specific options, we don't call replace_mount_options (thus we don't erase existing reported options). Fixes: c8f33d0bec99 ("affs: kstrdup() memory handling") Signed-off-by: Mikulas Patocka Cc: stable@vger.kernel.org # v4.1+ Signed-off-by: Linus Torvalds --- fs/affs/super.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/affs/super.c b/fs/affs/super.c index 2a6713b6b9f4..d6384863192c 100644 --- a/fs/affs/super.c +++ b/fs/affs/super.c @@ -528,7 +528,7 @@ affs_remount(struct super_block *sb, int *flags, char *data) char *prefix = NULL; new_opts = kstrdup(data, GFP_KERNEL); - if (!new_opts) + if (data && !new_opts) return -ENOMEM; pr_debug("%s(flags=0x%x,opts=\"%s\")\n", __func__, *flags, data); @@ -546,7 +546,8 @@ affs_remount(struct super_block *sb, int *flags, char *data) } flush_delayed_work(&sbi->sb_work); - replace_mount_options(sb, new_opts); + if (new_opts) + replace_mount_options(sb, new_opts); sbi->s_flags = mount_flags; sbi->s_mode = mode; -- cgit v1.3-8-gc7d7 From 037369b872940cd923835a0a589763180c4a36bc Mon Sep 17 00:00:00 2001 From: Mikulas Patocka Date: Tue, 24 May 2016 22:49:18 +0200 Subject: hpfs: implement the show_options method The HPFS filesystem used generic_show_options to produce string that is displayed in /proc/mounts. However, there is a problem that the options may disappear after remount. If we mount the filesystem with option1 and then remount it with option2, /proc/mounts should show both option1 and option2, however it only shows option2 because the whole option string is replaced with replace_mount_options in hpfs_remount_fs. To fix this bug, implement the hpfs_show_options function that prints options that are currently selected. Signed-off-by: Mikulas Patocka Cc: stable@vger.kernel.org Signed-off-by: Linus Torvalds --- fs/hpfs/super.c | 43 ++++++++++++++++++++++++++++++++----------- 1 file changed, 32 insertions(+), 11 deletions(-) (limited to 'fs') diff --git a/fs/hpfs/super.c b/fs/hpfs/super.c index 46ad0ef8b4cc..82067ca22f2b 100644 --- a/fs/hpfs/super.c +++ b/fs/hpfs/super.c @@ -15,6 +15,7 @@ #include #include #include +#include /* Mark the filesystem dirty, so that chkdsk checks it when os/2 booted */ @@ -453,10 +454,6 @@ static int hpfs_remount_fs(struct super_block *s, int *flags, char *data) int lowercase, eas, chk, errs, chkdsk, timeshift; int o; struct hpfs_sb_info *sbi = hpfs_sb(s); - char *new_opts = kstrdup(data, GFP_KERNEL); - - if (data && !new_opts) - return -ENOMEM; sync_filesystem(s); @@ -493,18 +490,44 @@ static int hpfs_remount_fs(struct super_block *s, int *flags, char *data) if (!(*flags & MS_RDONLY)) mark_dirty(s, 1); - if (new_opts) - replace_mount_options(s, new_opts); - hpfs_unlock(s); return 0; out_err: hpfs_unlock(s); - kfree(new_opts); return -EINVAL; } +static int hpfs_show_options(struct seq_file *seq, struct dentry *root) +{ + struct hpfs_sb_info *sbi = hpfs_sb(root->d_sb); + + seq_printf(seq, ",uid=%u", from_kuid_munged(&init_user_ns, sbi->sb_uid)); + seq_printf(seq, ",gid=%u", from_kgid_munged(&init_user_ns, sbi->sb_gid)); + seq_printf(seq, ",umask=%03o", (~sbi->sb_mode & 0777)); + if (sbi->sb_lowercase) + seq_printf(seq, ",case=lower"); + if (!sbi->sb_chk) + seq_printf(seq, ",check=none"); + if (sbi->sb_chk == 2) + seq_printf(seq, ",check=strict"); + if (!sbi->sb_err) + seq_printf(seq, ",errors=continue"); + if (sbi->sb_err == 2) + seq_printf(seq, ",errors=panic"); + if (!sbi->sb_chkdsk) + seq_printf(seq, ",chkdsk=no"); + if (sbi->sb_chkdsk == 2) + seq_printf(seq, ",chkdsk=always"); + if (!sbi->sb_eas) + seq_printf(seq, ",eas=no"); + if (sbi->sb_eas == 1) + seq_printf(seq, ",eas=ro"); + if (sbi->sb_timeshift) + seq_printf(seq, ",timeshift=%d", sbi->sb_timeshift); + return 0; +} + /* Super operations */ static const struct super_operations hpfs_sops = @@ -515,7 +538,7 @@ static const struct super_operations hpfs_sops = .put_super = hpfs_put_super, .statfs = hpfs_statfs, .remount_fs = hpfs_remount_fs, - .show_options = generic_show_options, + .show_options = hpfs_show_options, }; static int hpfs_fill_super(struct super_block *s, void *options, int silent) @@ -538,8 +561,6 @@ static int hpfs_fill_super(struct super_block *s, void *options, int silent) int o; - save_mount_options(s, options); - sbi = kzalloc(sizeof(*sbi), GFP_KERNEL); if (!sbi) { return -ENOMEM; -- cgit v1.3-8-gc7d7 From f2a031b66ec6da1512f9db228e4174ad2c7d2688 Mon Sep 17 00:00:00 2001 From: George Spelvin Date: Sun, 29 May 2016 01:26:41 -0400 Subject: Rename other copy of hash_string to hashlen_string The original name was simply hash_string(), but that conflicted with a function with that name in drivers/base/power/trace.c, and I decided that calling it "hashlen_" was better anyway. But you have to do it in two places. [ This caused build errors for architectures that don't define CONFIG_DCACHE_WORD_ACCESS - Linus ] Signed-off-by: George Spelvin Reported-by: Guenter Roeck Fixes: fcfd2fbf22d2 ("fs/namei.c: Add hashlen_string() function") Signed-off-by: Linus Torvalds --- fs/namei.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/namei.c b/fs/namei.c index e7bf99d387d0..47edc9981298 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -1962,7 +1962,7 @@ unsigned int full_name_hash(const char *name, unsigned int len) EXPORT_SYMBOL(full_name_hash); /* Return the "hash_len" (hash and length) of a null-terminated string */ -u64 hash_string(const char *name) +u64 hashlen_string(const char *name) { unsigned long hash = init_name_hash(); unsigned long len = 0, c; @@ -1975,7 +1975,7 @@ u64 hash_string(const char *name) } while (c); return hashlen_create(end_name_hash(hash), len); } -EXPORT_SYMBOL(hash_string); +EXPORT_SYMBOL(hashlen_string); /* * We know there's a real path component here of at least -- cgit v1.3-8-gc7d7 From e0ab7af9bd1c25e40b9a1dbeb2a634b88267b9d7 Mon Sep 17 00:00:00 2001 From: George Spelvin Date: Sun, 29 May 2016 08:05:56 -0400 Subject: hash_string: Fix zero-length case for !DCACHE_WORD_ACCESS The self-test was updated to cover zero-length strings; the function needs to be updated, too. Reported-by: Geert Uytterhoeven Signed-off-by: George Spelvin Fixes: fcfd2fbf22d2 ("fs/namei.c: Add hashlen_string() function") Signed-off-by: Linus Torvalds --- fs/namei.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/namei.c b/fs/namei.c index 47edc9981298..4c4f95ac8aa5 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -1968,11 +1968,11 @@ u64 hashlen_string(const char *name) unsigned long len = 0, c; c = (unsigned char)*name; - do { + while (c) { len++; hash = partial_name_hash(c, hash); c = (unsigned char)name[len]; - } while (c); + } return hashlen_create(end_name_hash(hash), len); } EXPORT_SYMBOL(hashlen_string); -- cgit v1.3-8-gc7d7 From ce7791ffee1e1ee9f97193b817c7dd1fa6746aad Mon Sep 17 00:00:00 2001 From: Filipe Manana Date: Fri, 20 May 2016 01:57:20 +0100 Subject: Btrfs: fix race between readahead and device replace/removal The list of devices is protected by the device_list_mutex and the device replace code, in its finishing phase correctly takes that mutex before removing the source device from that list. However the readahead code was iterating that list without acquiring the respective mutex leading to crashes later on due to invalid memory accesses: [125671.831036] general protection fault: 0000 [#1] PREEMPT SMP [125671.832129] Modules linked in: btrfs dm_flakey dm_mod crc32c_generic xor raid6_pq acpi_cpufreq tpm_tis tpm ppdev evdev parport_pc psmouse sg parport processor ser [125671.834973] CPU: 10 PID: 19603 Comm: kworker/u32:19 Tainted: G W 4.6.0-rc7-btrfs-next-29+ #1 [125671.834973] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS by qemu-project.org 04/01/2014 [125671.834973] Workqueue: btrfs-readahead btrfs_readahead_helper [btrfs] [125671.834973] task: ffff8801ac520540 ti: ffff8801ac918000 task.ti: ffff8801ac918000 [125671.834973] RIP: 0010:[] [] __radix_tree_lookup+0x6a/0x105 [125671.834973] RSP: 0018:ffff8801ac91bc28 EFLAGS: 00010206 [125671.834973] RAX: 0000000000000000 RBX: 6b6b6b6b6b6b6b6a RCX: 0000000000000000 [125671.834973] RDX: 0000000000000000 RSI: 00000000000c1bff RDI: ffff88002ebd62a8 [125671.834973] RBP: ffff8801ac91bc70 R08: 0000000000000001 R09: 0000000000000000 [125671.834973] R10: ffff8801ac91bc70 R11: 0000000000000000 R12: ffff88002ebd62a8 [125671.834973] R13: 0000000000000000 R14: 0000000000000000 R15: 00000000000c1bff [125671.834973] FS: 0000000000000000(0000) GS:ffff88023fd40000(0000) knlGS:0000000000000000 [125671.834973] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [125671.834973] CR2: 000000000073cae4 CR3: 00000000b7723000 CR4: 00000000000006e0 [125671.834973] Stack: [125671.834973] 0000000000000000 ffff8801422d5600 ffff8802286bbc00 0000000000000000 [125671.834973] 0000000000000001 ffff8802286bbc00 00000000000c1bff 0000000000000000 [125671.834973] ffff88002e639eb8 ffff8801ac91bc80 ffffffff81270541 ffff8801ac91bcb0 [125671.834973] Call Trace: [125671.834973] [] radix_tree_lookup+0xd/0xf [125671.834973] [] reada_peer_zones_set_lock+0x3e/0x60 [btrfs] [125671.834973] [] reada_pick_zone+0x29/0x103 [btrfs] [125671.834973] [] reada_start_machine_worker+0x129/0x2d3 [btrfs] [125671.834973] [] btrfs_scrubparity_helper+0x185/0x3aa [btrfs] [125671.834973] [] btrfs_readahead_helper+0xe/0x10 [btrfs] [125671.834973] [] process_one_work+0x271/0x4e9 [125671.834973] [] worker_thread+0x1eb/0x2c9 [125671.834973] [] ? rescuer_thread+0x2b3/0x2b3 [125671.834973] [] kthread+0xd4/0xdc [125671.834973] [] ret_from_fork+0x22/0x40 [125671.834973] [] ? kthread_stop+0x286/0x286 So fix this by taking the device_list_mutex in the readahead code. We can't use here the lighter approach of using a rcu_read_lock() and rcu_read_unlock() pair together with a list_for_each_entry_rcu() call because we end up doing calls to sleeping functions (kzalloc()) in the respective code path. Signed-off-by: Filipe Manana Reviewed-by: Josef Bacik --- fs/btrfs/reada.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/reada.c b/fs/btrfs/reada.c index 298631eaee78..8428db7cd88f 100644 --- a/fs/btrfs/reada.c +++ b/fs/btrfs/reada.c @@ -761,12 +761,14 @@ static void __reada_start_machine(struct btrfs_fs_info *fs_info) do { enqueued = 0; + mutex_lock(&fs_devices->device_list_mutex); list_for_each_entry(device, &fs_devices->devices, dev_list) { if (atomic_read(&device->reada_in_flight) < MAX_IN_FLIGHT) enqueued += reada_start_machine_dev(fs_info, device); } + mutex_unlock(&fs_devices->device_list_mutex); total += enqueued; } while (enqueued && total < 10000); -- cgit v1.3-8-gc7d7 From 57ba4cb85bffc0c7c6567c89d23713721fea9655 Mon Sep 17 00:00:00 2001 From: Filipe Manana Date: Fri, 20 May 2016 04:34:23 +0100 Subject: Btrfs: fix race between device replace and block group removal When it's finishing, the device replace code iterates all extent maps representing block group and for each one that has a stripe that refers to the source device, it replaces its device with the target device. However when it replaces the source device with the target device it, the target device still has an ID of 0ULL (BTRFS_DEV_REPLACE_DEVID), only after its ID is changed to match the one from the source device. This leads to races with the chunk removal code that can temporarly see a device with an ID of 0ULL and then attempt to use that ID to remove items from the device tree and fail, causing a transaction abort: [ 9238.594364] BTRFS info (device sdf): dev_replace from /dev/sdf (devid 3) to /dev/sde finished [ 9238.594377] ------------[ cut here ]------------ [ 9238.594402] WARNING: CPU: 14 PID: 21566 at fs/btrfs/volumes.c:2771 btrfs_remove_chunk+0x2e5/0x793 [btrfs] [ 9238.594403] BTRFS: Transaction aborted (error 1) [ 9238.594416] Modules linked in: btrfs crc32c_generic acpi_cpufreq xor tpm_tis tpm raid6_pq ppdev parport_pc processor psmouse parport i2c_piix4 evdev sg i2c_core se rio_raw pcspkr button loop autofs4 ext4 crc16 jbd2 mbcache sr_mod cdrom sd_mod ata_generic virtio_scsi ata_piix virtio_pci libata virtio_ring virtio e1000 scsi_mod fl oppy [last unloaded: btrfs] [ 9238.594418] CPU: 14 PID: 21566 Comm: btrfs-cleaner Not tainted 4.6.0-rc7-btrfs-next-29+ #1 [ 9238.594419] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS by qemu-project.org 04/01/2014 [ 9238.594421] 0000000000000000 ffff88017f1dbc60 ffffffff8126b42c ffff88017f1dbcb0 [ 9238.594422] 0000000000000000 ffff88017f1dbca0 ffffffff81052b14 00000ad37f1dbd18 [ 9238.594423] 0000000000000001 ffff88018068a558 ffff88005c4b9c00 ffff880233f60db0 [ 9238.594424] Call Trace: [ 9238.594428] [] dump_stack+0x67/0x90 [ 9238.594430] [] __warn+0xc2/0xdd [ 9238.594432] [] warn_slowpath_fmt+0x4b/0x53 [ 9238.594434] [] ? kmem_cache_free+0x128/0x188 [ 9238.594450] [] btrfs_remove_chunk+0x2e5/0x793 [btrfs] [ 9238.594452] [] ? arch_local_irq_save+0x9/0xc [ 9238.594464] [] btrfs_delete_unused_bgs+0x317/0x382 [btrfs] [ 9238.594476] [] cleaner_kthread+0x1ad/0x1c7 [btrfs] [ 9238.594489] [] ? btree_invalidatepage+0x8e/0x8e [btrfs] [ 9238.594490] [] kthread+0xd4/0xdc [ 9238.594494] [] ret_from_fork+0x22/0x40 [ 9238.594495] [] ? kthread_stop+0x286/0x286 [ 9238.594496] ---[ end trace 183efbe50275f059 ]--- The sequence of steps leading to this is like the following: CPU 1 CPU 2 btrfs_dev_replace_finishing() at this point dev_replace->tgtdev->devid == BTRFS_DEV_REPLACE_DEVID (0ULL) ... btrfs_start_transaction() btrfs_commit_transaction() btrfs_delete_unused_bgs() btrfs_remove_chunk() looks up for the extent map corresponding to the chunk lock_chunks() (chunk_mutex) check_system_chunk() unlock_chunks() (chunk_mutex) locks fs_info->chunk_mutex btrfs_dev_replace_update_device_in_mapping_tree() --> iterates fs_info->mapping_tree and replaces the device in every extent map's map->stripes[] with dev_replace->tgtdev, which still has an id of 0ULL (BTRFS_DEV_REPLACE_DEVID) iterates over all stripes from the extent map --> calls btrfs_free_dev_extent() passing it the target device that still has an ID of 0ULL --> btrfs_free_dev_extent() fails --> aborts current transaction finishes setting up the target device, namely it sets tgtdev->devid to the value of srcdev->devid (which is necessarily > 0) frees the srcdev unlocks fs_info->chunk_mutex So fix this by taking the device list mutex while processing the stripes for the chunk's extent map. This is similar to the race between device replace and block group creation that was fixed by commit 50460e37186a ("Btrfs: fix race when finishing dev replace leading to transaction abort"). Signed-off-by: Filipe Manana Reviewed-by: Josef Bacik --- fs/btrfs/volumes.c | 11 +++++++++++ 1 file changed, 11 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 9c01824eef08..04ca48362ef1 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -2761,6 +2761,7 @@ int btrfs_remove_chunk(struct btrfs_trans_handle *trans, u64 dev_extent_len = 0; u64 chunk_objectid = BTRFS_FIRST_CHUNK_TREE_OBJECTID; int i, ret = 0; + struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices; /* Just in case */ root = root->fs_info->chunk_root; @@ -2787,12 +2788,19 @@ int btrfs_remove_chunk(struct btrfs_trans_handle *trans, check_system_chunk(trans, extent_root, map->type); unlock_chunks(root->fs_info->chunk_root); + /* + * Take the device list mutex to prevent races with the final phase of + * a device replace operation that replaces the device object associated + * with map stripes (dev-replace.c:btrfs_dev_replace_finishing()). + */ + mutex_lock(&fs_devices->device_list_mutex); for (i = 0; i < map->num_stripes; i++) { struct btrfs_device *device = map->stripes[i].dev; ret = btrfs_free_dev_extent(trans, device, map->stripes[i].physical, &dev_extent_len); if (ret) { + mutex_unlock(&fs_devices->device_list_mutex); btrfs_abort_transaction(trans, root, ret); goto out; } @@ -2811,11 +2819,14 @@ int btrfs_remove_chunk(struct btrfs_trans_handle *trans, if (map->stripes[i].dev) { ret = btrfs_update_device(trans, map->stripes[i].dev); if (ret) { + mutex_unlock(&fs_devices->device_list_mutex); btrfs_abort_transaction(trans, root, ret); goto out; } } } + mutex_unlock(&fs_devices->device_list_mutex); + ret = btrfs_free_chunk(trans, root, chunk_objectid, chunk_offset); if (ret) { btrfs_abort_transaction(trans, root, ret); -- cgit v1.3-8-gc7d7 From f0e9b7d6401959816599191d1d9db90b6fd750db Mon Sep 17 00:00:00 2001 From: Filipe Manana Date: Sat, 14 May 2016 09:12:53 +0100 Subject: Btrfs: fix race setting block group readonly during device replace When we do a device replace, for each device extent we find from the source device, we set the corresponding block group to readonly mode to prevent writes into it from happening while we are copying the device extent from the source to the target device. However just before we set the block group to readonly mode some concurrent task might have already allocated an extent from it or decided it could perform a nocow write into one of its extents, which can make the device replace process to miss copying an extent since it uses the extent tree's commit root to search for extents and only once it finishes searching for all extents belonging to the block group it does set the left cursor to the logical end address of the block group - this is a problem if the respective ordered extents finish while we are searching for extents using the extent tree's commit root and no transaction commit happens while we are iterating the tree, since it's the delayed references created by the ordered extents (when they complete) that insert the extent items into the extent tree (using the non-commit root of course). Example: CPU 1 CPU 2 btrfs_dev_replace_start() btrfs_scrub_dev() scrub_enumerate_chunks() --> finds device extent belonging to block group X starts buffered write against some inode writepages is run against that inode forcing dellaloc to run btrfs_writepages() extent_writepages() extent_write_cache_pages() __extent_writepage() writepage_delalloc() run_delalloc_range() cow_file_range() btrfs_reserve_extent() --> allocates an extent from block group X (which is not yet in RO mode) btrfs_add_ordered_extent() --> creates ordered extent Y flush_epd_write_bio() --> bio against the extent from block group X is submitted btrfs_inc_block_group_ro(bg X) --> sets block group X to readonly scrub_chunk(bg X) scrub_stripe(device extent from srcdev) --> keeps searching for extent items belonging to the block group using the extent tree's commit root --> it never blocks due to fs_info->scrub_pause_req as no one tries to commit transaction N --> copies all extents found from the source device into the target device --> finishes search loop bio completes ordered extent Y completes and creates delayed data reference which will add an extent item to the extent tree when run (typically at transaction commit time) --> so the task doing the scrub/device replace at CPU 1 misses this and does not copy this extent into the new/target device btrfs_dec_block_group_ro(bg X) --> turns block group X back to RW mode dev_replace->cursor_left is set to the logical end offset of block group X So fix this by waiting for all cow and nocow writes after setting a block group to readonly mode. Signed-off-by: Filipe Manana Reviewed-by: Josef Bacik --- fs/btrfs/ordered-data.c | 6 +++++- fs/btrfs/ordered-data.h | 2 +- fs/btrfs/scrub.c | 40 ++++++++++++++++++++++++++++++++++++++++ 3 files changed, 46 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ordered-data.c b/fs/btrfs/ordered-data.c index 559170464d7c..e96634a725c3 100644 --- a/fs/btrfs/ordered-data.c +++ b/fs/btrfs/ordered-data.c @@ -718,12 +718,13 @@ int btrfs_wait_ordered_extents(struct btrfs_root *root, int nr, return count; } -void btrfs_wait_ordered_roots(struct btrfs_fs_info *fs_info, int nr, +int btrfs_wait_ordered_roots(struct btrfs_fs_info *fs_info, int nr, const u64 range_start, const u64 range_len) { struct btrfs_root *root; struct list_head splice; int done; + int total_done = 0; INIT_LIST_HEAD(&splice); @@ -742,6 +743,7 @@ void btrfs_wait_ordered_roots(struct btrfs_fs_info *fs_info, int nr, done = btrfs_wait_ordered_extents(root, nr, range_start, range_len); btrfs_put_fs_root(root); + total_done += done; spin_lock(&fs_info->ordered_root_lock); if (nr != -1) { @@ -752,6 +754,8 @@ void btrfs_wait_ordered_roots(struct btrfs_fs_info *fs_info, int nr, list_splice_tail(&splice, &fs_info->ordered_roots); spin_unlock(&fs_info->ordered_root_lock); mutex_unlock(&fs_info->ordered_operations_mutex); + + return total_done; } /* diff --git a/fs/btrfs/ordered-data.h b/fs/btrfs/ordered-data.h index 2049c9be85ee..451507776ff5 100644 --- a/fs/btrfs/ordered-data.h +++ b/fs/btrfs/ordered-data.h @@ -199,7 +199,7 @@ int btrfs_find_ordered_sum(struct inode *inode, u64 offset, u64 disk_bytenr, u32 *sum, int len); int btrfs_wait_ordered_extents(struct btrfs_root *root, int nr, const u64 range_start, const u64 range_len); -void btrfs_wait_ordered_roots(struct btrfs_fs_info *fs_info, int nr, +int btrfs_wait_ordered_roots(struct btrfs_fs_info *fs_info, int nr, const u64 range_start, const u64 range_len); void btrfs_get_logged_extents(struct inode *inode, struct list_head *logged_list, diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c index 46d847f66e4b..1611572d47bd 100644 --- a/fs/btrfs/scrub.c +++ b/fs/btrfs/scrub.c @@ -3582,6 +3582,46 @@ int scrub_enumerate_chunks(struct scrub_ctx *sctx, */ scrub_pause_on(fs_info); ret = btrfs_inc_block_group_ro(root, cache); + if (!ret && is_dev_replace) { + /* + * If we are doing a device replace wait for any tasks + * that started dellaloc right before we set the block + * group to RO mode, as they might have just allocated + * an extent from it or decided they could do a nocow + * write. And if any such tasks did that, wait for their + * ordered extents to complete and then commit the + * current transaction, so that we can later see the new + * extent items in the extent tree - the ordered extents + * create delayed data references (for cow writes) when + * they complete, which will be run and insert the + * corresponding extent items into the extent tree when + * we commit the transaction they used when running + * inode.c:btrfs_finish_ordered_io(). We later use + * the commit root of the extent tree to find extents + * to copy from the srcdev into the tgtdev, and we don't + * want to miss any new extents. + */ + btrfs_wait_block_group_reservations(cache); + btrfs_wait_nocow_writers(cache); + ret = btrfs_wait_ordered_roots(fs_info, -1, + cache->key.objectid, + cache->key.offset); + if (ret > 0) { + struct btrfs_trans_handle *trans; + + trans = btrfs_join_transaction(root); + if (IS_ERR(trans)) + ret = PTR_ERR(trans); + else + ret = btrfs_commit_transaction(trans, + root); + if (ret) { + scrub_pause_off(fs_info); + btrfs_put_block_group(cache); + break; + } + } + } scrub_pause_off(fs_info); if (ret == 0) { -- cgit v1.3-8-gc7d7 From 81e87a736c5581e83f52dc054b368993ec7f16d7 Mon Sep 17 00:00:00 2001 From: Filipe Manana Date: Sat, 14 May 2016 16:32:35 +0100 Subject: Btrfs: fix unprotected assignment of the left cursor for device replace We were assigning new values to fields of the device replace object without holding the respective lock after processing each device extent. This is important for the left cursor field which can be accessed by a concurrent task running __btrfs_map_block (which, correctly, takes the device replace lock). So change these fields while holding the device replace lock. Signed-off-by: Filipe Manana Reviewed-by: Josef Bacik --- fs/btrfs/scrub.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c index 1611572d47bd..19a1eda70361 100644 --- a/fs/btrfs/scrub.c +++ b/fs/btrfs/scrub.c @@ -3642,9 +3642,11 @@ int scrub_enumerate_chunks(struct scrub_ctx *sctx, break; } + btrfs_dev_replace_lock(&fs_info->dev_replace, 1); dev_replace->cursor_right = found_key.offset + length; dev_replace->cursor_left = found_key.offset; dev_replace->item_needs_writeback = 1; + btrfs_dev_replace_unlock(&fs_info->dev_replace, 1); ret = scrub_chunk(sctx, scrub_dev, chunk_offset, length, found_key.offset, cache, is_dev_replace); @@ -3718,8 +3720,10 @@ int scrub_enumerate_chunks(struct scrub_ctx *sctx, break; } + btrfs_dev_replace_lock(&fs_info->dev_replace, 1); dev_replace->cursor_left = dev_replace->cursor_right; dev_replace->item_needs_writeback = 1; + btrfs_dev_replace_unlock(&fs_info->dev_replace, 1); skip: key.offset = found_key.offset + length; btrfs_release_path(path); -- cgit v1.3-8-gc7d7 From 1a1a8b732c7e958e6eba0680439e814efde2362d Mon Sep 17 00:00:00 2001 From: Filipe Manana Date: Sat, 14 May 2016 19:44:40 +0100 Subject: Btrfs: fix race setting block group back to RW mode during device replace After it finishes processing a device extent, the device replace code sets back the block group to RW mode and then after that it sets the left cursor to match the logical end address of the block group, so that future writes into extents belonging to the block group go both the source (old) and target (new) devices. However from the moment we turn the block group back to RW mode we have a short time window, that lasts until we update the left cursor's value, where extents can be allocated from the block group and written to, in which case they will not be copied/written to the target (new) device. Fix this by updating the left cursor's value before turning the block group back to RW mode. Signed-off-by: Filipe Manana Reviewed-by: Josef Bacik --- fs/btrfs/scrub.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c index 19a1eda70361..70427ef66b04 100644 --- a/fs/btrfs/scrub.c +++ b/fs/btrfs/scrub.c @@ -3682,6 +3682,11 @@ int scrub_enumerate_chunks(struct scrub_ctx *sctx, scrub_pause_off(fs_info); + btrfs_dev_replace_lock(&fs_info->dev_replace, 1); + dev_replace->cursor_left = dev_replace->cursor_right; + dev_replace->item_needs_writeback = 1; + btrfs_dev_replace_unlock(&fs_info->dev_replace, 1); + if (ro_set) btrfs_dec_block_group_ro(root, cache); @@ -3719,11 +3724,6 @@ int scrub_enumerate_chunks(struct scrub_ctx *sctx, ret = -ENOMEM; break; } - - btrfs_dev_replace_lock(&fs_info->dev_replace, 1); - dev_replace->cursor_left = dev_replace->cursor_right; - dev_replace->item_needs_writeback = 1; - btrfs_dev_replace_unlock(&fs_info->dev_replace, 1); skip: key.offset = found_key.offset + length; btrfs_release_path(path); -- cgit v1.3-8-gc7d7 From 22ab04e814f4fe2ce72a13d291491f98ef6ac757 Mon Sep 17 00:00:00 2001 From: Filipe Manana Date: Wed, 18 May 2016 20:29:44 +0100 Subject: Btrfs: fix race between device replace and chunk allocation While iterating and copying extents from the source device, the device replace code keeps adjusting a left cursor that is used to make sure that once we finish processing a device extent, any future writes to extents from the corresponding block group will get into both the source and target devices. This left cursor is also used for resuming the device replace operation at mount time. However using this left cursor to decide whether writes go into both devices or only the source device is not enough to guarantee we don't miss copying extents into the target device. There are two cases where the current approach fails. The first one is related to when there are holes in the device and they get allocated for new block groups while the device replace operation is iterating the device extents (more on this explained below). The second one is that when that loop over the device extents finishes, we start dellaloc, wait for all ordered extents and then commit the current transaction, we might have got new block groups allocated that are now using a device extent that has an offset greater then or equals to the value of the left cursor, in which case writes to extents belonging to these new block groups will get issued only to the source device. For the first case where the current approach of using a left cursor fails, consider the source device currently has the following layout: [ extent bg A ] [ hole, unallocated space ] [extent bg B ] 3Gb 4Gb 5Gb While we are iterating the device extents from the source device using the commit root of the device tree, the following happens: CPU 1 CPU 2 scrub_enumerate_chunks() --> searches the device tree for extents belonging to the source device using the device tree's commit root --> 1st iteration finds extent belonging to block group A --> sets block group A to RO mode (btrfs_inc_block_group_ro) --> sets cursor left to found_key.offset which is 3Gb --> scrub_chunk() starts copies all allocated extents from block group's A stripe at source device into target device btrfs_alloc_chunk() --> allocates device extent in the range [4Gb, 5Gb[ from the source device for a new block group C extent allocated from block group C for a direct IO, buffered write or btree node/leaf extent is written to, perhaps in response to a writepages() call from the VM or directly through direct IO the write is made only against the source device and not against the target device because the extent's offset is in the interval [4Gb, 5Gb[ which is larger then the value of cursor_left (3Gb) --> scrub_chunks() finishes --> updates left cursor from 3Gb to 4Gb --> btrfs_dec_block_group_ro() sets block group A back to RW mode --> 2nd iteration finds extent belonging to block group B - it did not find the new extent in the range [4Gb, 5Gb[ for block group C because we are using the device tree's commit root or even because the block group's items are not all yet inserted in the respective btrees, that is, the block group is still attached to some transaction handle's new_bgs list and btrfs_create_pending_block_groups() was not called yet against that transaction handle, so the device extent items were not yet inserted into the devices tree --> so we end not copying anything from the newly allocated device extent from the source device to the target device So fix this by making __btrfs_map_block() always redirect writes to the target device as well, independently of the left cursor's value. With this change the left cursor is now used only for the purpose of tracking progress and allow a mount operation to resume a device replace. Signed-off-by: Filipe Manana Reviewed-by: Josef Bacik --- fs/btrfs/volumes.c | 21 +++++++++------------ 1 file changed, 9 insertions(+), 12 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 04ca48362ef1..765aabd9145f 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -5773,20 +5773,17 @@ static int __btrfs_map_block(struct btrfs_fs_info *fs_info, int rw, } } if (found) { - if (physical_of_found + map->stripe_len <= - dev_replace->cursor_left) { - struct btrfs_bio_stripe *tgtdev_stripe = - bbio->stripes + num_stripes; + struct btrfs_bio_stripe *tgtdev_stripe = + bbio->stripes + num_stripes; - tgtdev_stripe->physical = physical_of_found; - tgtdev_stripe->length = - bbio->stripes[index_srcdev].length; - tgtdev_stripe->dev = dev_replace->tgtdev; - bbio->tgtdev_map[index_srcdev] = num_stripes; + tgtdev_stripe->physical = physical_of_found; + tgtdev_stripe->length = + bbio->stripes[index_srcdev].length; + tgtdev_stripe->dev = dev_replace->tgtdev; + bbio->tgtdev_map[index_srcdev] = num_stripes; - tgtdev_indexes++; - num_stripes++; - } + tgtdev_indexes++; + num_stripes++; } } -- cgit v1.3-8-gc7d7 From b7ec35b304b64af2830027350cc99d31e6e537c2 Mon Sep 17 00:00:00 2001 From: Ilya Dryomov Date: Thu, 28 Apr 2016 16:07:25 +0200 Subject: libceph: change ceph_osdmap_flag() to take osdc For the benefit of every single caller, take osdc instead of map. Also, now that osdc->osdmap can't ever be NULL, drop the check. Signed-off-by: Ilya Dryomov --- fs/ceph/file.c | 8 ++++---- include/linux/ceph/osd_client.h | 5 +++++ include/linux/ceph/osdmap.h | 5 ----- net/ceph/osd_client.c | 39 +++++++++++++++++++-------------------- 4 files changed, 28 insertions(+), 29 deletions(-) (limited to 'fs') diff --git a/fs/ceph/file.c b/fs/ceph/file.c index a888df6f2d71..8eeb9f579db5 100644 --- a/fs/ceph/file.c +++ b/fs/ceph/file.c @@ -1349,7 +1349,7 @@ static ssize_t ceph_write_iter(struct kiocb *iocb, struct iov_iter *from) } retry_snap: - if (ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_FULL)) { + if (ceph_osdmap_flag(osdc, CEPH_OSDMAP_FULL)) { err = -ENOSPC; goto out; } @@ -1440,7 +1440,7 @@ retry_snap: ceph_put_cap_refs(ci, got); if (written >= 0) { - if (ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_NEARFULL)) + if (ceph_osdmap_flag(osdc, CEPH_OSDMAP_NEARFULL)) iocb->ki_flags |= IOCB_DSYNC; written = generic_write_sync(iocb, written); @@ -1672,8 +1672,8 @@ static long ceph_fallocate(struct file *file, int mode, goto unlock; } - if (ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_FULL) && - !(mode & FALLOC_FL_PUNCH_HOLE)) { + if (ceph_osdmap_flag(osdc, CEPH_OSDMAP_FULL) && + !(mode & FALLOC_FL_PUNCH_HOLE)) { ret = -ENOSPC; goto unlock; } diff --git a/include/linux/ceph/osd_client.h b/include/linux/ceph/osd_client.h index 19b14862d3e0..1b3b6e155392 100644 --- a/include/linux/ceph/osd_client.h +++ b/include/linux/ceph/osd_client.h @@ -279,6 +279,11 @@ struct ceph_osd_client { struct workqueue_struct *notify_wq; }; +static inline bool ceph_osdmap_flag(struct ceph_osd_client *osdc, int flag) +{ + return osdc->osdmap->flags & flag; +} + extern int ceph_osdc_setup(void); extern void ceph_osdc_cleanup(void); diff --git a/include/linux/ceph/osdmap.h b/include/linux/ceph/osdmap.h index ddc426b22d81..9ccf4dbe55f8 100644 --- a/include/linux/ceph/osdmap.h +++ b/include/linux/ceph/osdmap.h @@ -189,11 +189,6 @@ static inline bool ceph_osd_is_down(struct ceph_osdmap *map, int osd) return !ceph_osd_is_up(map, osd); } -static inline bool ceph_osdmap_flag(struct ceph_osdmap *map, int flag) -{ - return map && (map->flags & flag); -} - extern char *ceph_osdmap_state_str(char *str, int len, int state); extern u32 ceph_get_primary_affinity(struct ceph_osdmap *map, int osd); diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c index 0160d7d09a1e..79c3bad87e62 100644 --- a/net/ceph/osd_client.c +++ b/net/ceph/osd_client.c @@ -1276,9 +1276,9 @@ static bool target_should_be_paused(struct ceph_osd_client *osdc, const struct ceph_osd_request_target *t, struct ceph_pg_pool_info *pi) { - bool pauserd = ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_PAUSERD); - bool pausewr = ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_PAUSEWR) || - ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_FULL) || + bool pauserd = ceph_osdmap_flag(osdc, CEPH_OSDMAP_PAUSERD); + bool pausewr = ceph_osdmap_flag(osdc, CEPH_OSDMAP_PAUSEWR) || + ceph_osdmap_flag(osdc, CEPH_OSDMAP_FULL) || __pool_full(pi); WARN_ON(pi->id != t->base_oloc.pool); @@ -1303,8 +1303,7 @@ static enum calc_target_result calc_target(struct ceph_osd_client *osdc, bool force_resend = false; bool need_check_tiering = false; bool need_resend = false; - bool sort_bitwise = ceph_osdmap_flag(osdc->osdmap, - CEPH_OSDMAP_SORTBITWISE); + bool sort_bitwise = ceph_osdmap_flag(osdc, CEPH_OSDMAP_SORTBITWISE); enum calc_target_result ct_res; int ret; @@ -1590,9 +1589,9 @@ static void maybe_request_map(struct ceph_osd_client *osdc) verify_osdc_locked(osdc); WARN_ON(!osdc->osdmap->epoch); - if (ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_FULL) || - ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_PAUSERD) || - ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_PAUSEWR)) { + if (ceph_osdmap_flag(osdc, CEPH_OSDMAP_FULL) || + ceph_osdmap_flag(osdc, CEPH_OSDMAP_PAUSERD) || + ceph_osdmap_flag(osdc, CEPH_OSDMAP_PAUSEWR)) { dout("%s osdc %p continuous\n", __func__, osdc); continuous = true; } else { @@ -1629,19 +1628,19 @@ again: } if ((req->r_flags & CEPH_OSD_FLAG_WRITE) && - ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_PAUSEWR)) { + ceph_osdmap_flag(osdc, CEPH_OSDMAP_PAUSEWR)) { dout("req %p pausewr\n", req); req->r_t.paused = true; maybe_request_map(osdc); } else if ((req->r_flags & CEPH_OSD_FLAG_READ) && - ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_PAUSERD)) { + ceph_osdmap_flag(osdc, CEPH_OSDMAP_PAUSERD)) { dout("req %p pauserd\n", req); req->r_t.paused = true; maybe_request_map(osdc); } else if ((req->r_flags & CEPH_OSD_FLAG_WRITE) && !(req->r_flags & (CEPH_OSD_FLAG_FULL_TRY | CEPH_OSD_FLAG_FULL_FORCE)) && - (ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_FULL) || + (ceph_osdmap_flag(osdc, CEPH_OSDMAP_FULL) || pool_full(osdc, req->r_t.base_oloc.pool))) { dout("req %p full/pool_full\n", req); pr_warn_ratelimited("FULL or reached pool quota\n"); @@ -2280,7 +2279,7 @@ static void send_linger_ping(struct ceph_osd_linger_request *lreq) struct ceph_osd_request *req = lreq->ping_req; struct ceph_osd_req_op *op = &req->r_ops[0]; - if (ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_PAUSERD)) { + if (ceph_osdmap_flag(osdc, CEPH_OSDMAP_PAUSERD)) { dout("%s PAUSERD\n", __func__); return; } @@ -3050,7 +3049,7 @@ static int handle_one_map(struct ceph_osd_client *osdc, bool skipped_map = false; bool was_full; - was_full = ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_FULL); + was_full = ceph_osdmap_flag(osdc, CEPH_OSDMAP_FULL); set_pool_was_full(osdc); if (incremental) @@ -3088,7 +3087,7 @@ static int handle_one_map(struct ceph_osd_client *osdc, osdc->osdmap = newmap; } - was_full &= !ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_FULL); + was_full &= !ceph_osdmap_flag(osdc, CEPH_OSDMAP_FULL); scan_requests(&osdc->homeless_osd, skipped_map, was_full, true, need_resend, need_resend_linger); @@ -3174,9 +3173,9 @@ void ceph_osdc_handle_map(struct ceph_osd_client *osdc, struct ceph_msg *msg) if (ceph_check_fsid(osdc->client, &fsid) < 0) goto bad; - was_pauserd = ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_PAUSERD); - was_pausewr = ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_PAUSEWR) || - ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_FULL) || + was_pauserd = ceph_osdmap_flag(osdc, CEPH_OSDMAP_PAUSERD); + was_pausewr = ceph_osdmap_flag(osdc, CEPH_OSDMAP_PAUSEWR) || + ceph_osdmap_flag(osdc, CEPH_OSDMAP_FULL) || have_pool_full(osdc); /* incremental maps */ @@ -3238,9 +3237,9 @@ done: * we find out when we are no longer full and stop returning * ENOSPC. */ - pauserd = ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_PAUSERD); - pausewr = ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_PAUSEWR) || - ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_FULL) || + pauserd = ceph_osdmap_flag(osdc, CEPH_OSDMAP_PAUSERD); + pausewr = ceph_osdmap_flag(osdc, CEPH_OSDMAP_PAUSEWR) || + ceph_osdmap_flag(osdc, CEPH_OSDMAP_FULL) || have_pool_full(osdc); if (was_pauserd || was_pausewr || pauserd || pausewr) maybe_request_map(osdc); -- cgit v1.3-8-gc7d7 From 2999241daa8d77947f108dfbde35c268cd7bd709 Mon Sep 17 00:00:00 2001 From: Filipe Manana Date: Fri, 27 May 2016 17:42:05 +0100 Subject: Btrfs: fix race between device replace and discard While we are finishing a device replace operation, we can make a discard operation (fs mounted with -o discard) do an invalid memory access like the one reported by the following trace: [ 3206.384654] general protection fault: 0000 [#1] PREEMPT SMP [ 3206.387520] Modules linked in: dm_mod btrfs crc32c_generic xor raid6_pq acpi_cpufreq tpm_tis psmouse tpm ppdev sg parport_pc evdev i2c_piix4 parport processor serio_raw i2c_core pcspkr button loop autofs4 ext4 crc16 jbd2 mbcache sr_mod cdrom ata_generic sd_mod virtio_scsi ata_piix libata virtio_pci virtio_ring scsi_mod e1000 virtio floppy [last unloaded: btrfs] [ 3206.388595] CPU: 14 PID: 29194 Comm: fsstress Not tainted 4.6.0-rc7-btrfs-next-29+ #1 [ 3206.388595] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS by qemu-project.org 04/01/2014 [ 3206.388595] task: ffff88017ace0100 ti: ffff880171b98000 task.ti: ffff880171b98000 [ 3206.388595] RIP: 0010:[] [] blkdev_issue_discard+0x5c/0x2a7 [ 3206.388595] RSP: 0018:ffff880171b9bb80 EFLAGS: 00010246 [ 3206.388595] RAX: ffff880171b9bc28 RBX: 000000000090d000 RCX: 0000000000000000 [ 3206.388595] RDX: ffffffff82fa1b48 RSI: ffffffff8179f46c RDI: ffffffff82fa1b48 [ 3206.388595] RBP: ffff880171b9bcc0 R08: 0000000000000000 R09: 0000000000000001 [ 3206.388595] R10: ffff880171b9bce0 R11: 000000000090f000 R12: ffff880171b9bbe8 [ 3206.388595] R13: 0000000000000010 R14: 0000000000004868 R15: 6b6b6b6b6b6b6b6b [ 3206.388595] FS: 00007f6182e4e700(0000) GS:ffff88023fdc0000(0000) knlGS:0000000000000000 [ 3206.388595] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 3206.388595] CR2: 00007f617c2bbb18 CR3: 000000017ad9c000 CR4: 00000000000006e0 [ 3206.388595] Stack: [ 3206.388595] 0000000000004878 0000000000000000 0000000002400040 0000000000000000 [ 3206.388595] 0000000000000000 ffff880171b9bbe8 ffff880171b9bbb0 ffff880171b9bbb0 [ 3206.388595] ffff880171b9bbc0 ffff880171b9bbc0 ffff880171b9bbd0 ffff880171b9bbd0 [ 3206.388595] Call Trace: [ 3206.388595] [] btrfs_issue_discard+0x12f/0x143 [btrfs] [ 3206.388595] [] ? btrfs_issue_discard+0x12f/0x143 [btrfs] [ 3206.388595] [] btrfs_discard_extent+0x87/0xde [btrfs] [ 3206.388595] [] btrfs_finish_extent_commit+0xb2/0x1df [btrfs] [ 3206.388595] [] ? __mutex_unlock_slowpath+0x150/0x15b [ 3206.388595] [] btrfs_commit_transaction+0x7fc/0x980 [btrfs] [ 3206.388595] [] ? __mutex_unlock_slowpath+0x150/0x15b [ 3206.388595] [] btrfs_sync_file+0x38f/0x428 [btrfs] [ 3206.388595] [] vfs_fsync_range+0x8c/0x9e [ 3206.388595] [] vfs_fsync+0x1c/0x1e [ 3206.388595] [] do_fsync+0x31/0x4a [ 3206.388595] [] SyS_fsync+0x10/0x14 [ 3206.388595] [] entry_SYSCALL_64_fastpath+0x18/0xa8 [ 3206.388595] [] ? time_hardirqs_off+0x9/0x14 [ 3206.388595] [] ? trace_hardirqs_off_caller+0x1f/0xaa This happens because when we call btrfs_map_block() from btrfs_discard_extent() to get a btrfs_bio structure, the device replace operation has not finished yet, but before we use the device of one of the stripes from the returned btrfs_bio structure, the device object is freed. This is illustrated by the following diagram. CPU 1 CPU 2 btrfs_dev_replace_start() (...) btrfs_dev_replace_finishing() btrfs_start_transaction() btrfs_commit_transaction() (...) btrfs_sync_file() btrfs_start_transaction() (...) btrfs_commit_transaction() btrfs_finish_extent_commit() btrfs_discard_extent() btrfs_map_block() --> returns a struct btrfs_bio with a stripe that has a device field pointing to source device of the replace operation (the device that is being replaced) mutex_lock(&uuid_mutex) mutex_lock(&fs_info->fs_devices->device_list_mutex) mutex_lock(&fs_info->chunk_mutex) btrfs_dev_replace_update_device_in_mapping_tree() --> iterates the mapping tree and for each extent map that has a stripe pointing to the source device, it updates the stripe to point to the target device instead btrfs_rm_dev_replace_blocked() --> waits for fs_info->bio_counter to go down to 0 btrfs_rm_dev_replace_remove_srcdev() --> removes source device from the list of devices mutex_unlock(&fs_info->chunk_mutex) mutex_unlock(&fs_info->fs_devices->device_list_mutex) mutex_unlock(&uuid_mutex) btrfs_rm_dev_replace_free_srcdev() --> frees the source device --> iterates over all stripes of the returned struct btrfs_bio --> for each stripe it dereferences its device pointer --> it ends up finding a pointer to the device used as the source device for the replace operation and that was already freed So fix this by surrounding the call to btrfs_map_block(), and the code that uses the returned struct btrfs_bio, with calls to btrfs_bio_counter_inc_blocked() and btrfs_bio_counter_dec(), so that the finishing phase of the device replace operation blocks until the the bio counter decreases to zero before it frees the source device. This is the same approach we do at btrfs_map_bio() for example. Signed-off-by: Filipe Manana Reviewed-by: Josef Bacik --- fs/btrfs/extent-tree.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index a400951e8678..689d25ac6a68 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -2042,6 +2042,11 @@ int btrfs_discard_extent(struct btrfs_root *root, u64 bytenr, struct btrfs_bio *bbio = NULL; + /* + * Avoid races with device replace and make sure our bbio has devices + * associated to its stripes that don't go away while we are discarding. + */ + btrfs_bio_counter_inc_blocked(root->fs_info); /* Tell the block device(s) that the sectors can be discarded */ ret = btrfs_map_block(root->fs_info, REQ_DISCARD, bytenr, &num_bytes, &bbio, 0); @@ -2074,6 +2079,7 @@ int btrfs_discard_extent(struct btrfs_root *root, u64 bytenr, } btrfs_put_bbio(bbio); } + btrfs_bio_counter_dec(root->fs_info); if (actual_bytes) *actual_bytes = discarded_bytes; -- cgit v1.3-8-gc7d7 From b5de8d0df80fa87f1f97fbcc4bbc8cad0a018802 Mon Sep 17 00:00:00 2001 From: Filipe Manana Date: Fri, 27 May 2016 22:21:27 +0100 Subject: Btrfs: fix race between device replace and read repair While we are finishing a device replace operation we can have a concurrent task trying to do a read repair operation, in which case it will call btrfs_map_block() to get a struct btrfs_bio which can have a stripe that points to the source device of the device replace operation. This allows for the read repair task to dereference the stripe's device pointer after the device replace operation has freed the source device, resulting in an invalid memory access. This is similar to the problem solved by my previous patch in the same series and named "Btrfs: fix race between device replace and discard". So fix this by surrounding the call to btrfs_map_block() and the code that uses the returned struct btrfs_bio with calls to btrfs_bio_counter_inc_blocked() and btrfs_bio_counter_dec(), giving the proper serialization with the finishing phase of the device replace operation. Signed-off-by: Filipe Manana Reviewed-by: Josef Bacik --- fs/btrfs/extent_io.c | 10 ++++++++++ 1 file changed, 10 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 3cd57825c75f..6e953de83f08 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -2025,9 +2025,16 @@ int repair_io_failure(struct inode *inode, u64 start, u64 length, u64 logical, bio->bi_iter.bi_size = 0; map_length = length; + /* + * Avoid races with device replace and make sure our bbio has devices + * associated to its stripes that don't go away while we are doing the + * read repair operation. + */ + btrfs_bio_counter_inc_blocked(fs_info); ret = btrfs_map_block(fs_info, WRITE, logical, &map_length, &bbio, mirror_num); if (ret) { + btrfs_bio_counter_dec(fs_info); bio_put(bio); return -EIO; } @@ -2037,6 +2044,7 @@ int repair_io_failure(struct inode *inode, u64 start, u64 length, u64 logical, dev = bbio->stripes[mirror_num-1].dev; btrfs_put_bbio(bbio); if (!dev || !dev->bdev || !dev->writeable) { + btrfs_bio_counter_dec(fs_info); bio_put(bio); return -EIO; } @@ -2045,6 +2053,7 @@ int repair_io_failure(struct inode *inode, u64 start, u64 length, u64 logical, if (btrfsic_submit_bio_wait(WRITE_SYNC, bio)) { /* try to remap that extent elsewhere? */ + btrfs_bio_counter_dec(fs_info); bio_put(bio); btrfs_dev_stat_inc_and_print(dev, BTRFS_DEV_STAT_WRITE_ERRS); return -EIO; @@ -2054,6 +2063,7 @@ int repair_io_failure(struct inode *inode, u64 start, u64 length, u64 logical, "read error corrected: ino %llu off %llu (dev %s sector %llu)", btrfs_ino(inode), start, rcu_str_deref(dev->name), sector); + btrfs_bio_counter_dec(fs_info); bio_put(bio); return 0; } -- cgit v1.3-8-gc7d7 From 65d4f4c151a5fa7b2dacaaf70def3f95001766d7 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Wed, 23 Sep 2015 15:00:37 -0400 Subject: Btrfs: end transaction if we abort when creating uuid root We still need to call btrfs_end_transaction if we call btrfs_abort_transaction, otherwise we hang and make me super grumpy. Thanks, Signed-off-by: Josef Bacik Signed-off-by: David Sterba --- fs/btrfs/volumes.c | 1 + 1 file changed, 1 insertion(+) (limited to 'fs') diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 9c01824eef08..673c72ab4fbe 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -4230,6 +4230,7 @@ int btrfs_create_uuid_tree(struct btrfs_fs_info *fs_info) if (IS_ERR(uuid_root)) { ret = PTR_ERR(uuid_root); btrfs_abort_transaction(trans, tree_root, ret); + btrfs_end_transaction(trans, tree_root); return ret; } -- cgit v1.3-8-gc7d7 From d2138455286f9db8fbff9b50b53b51766f6c1f92 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Tue, 17 May 2016 11:52:48 +0800 Subject: FS-Cache: wake write waiter after invalidating writes Signed-off-by: Yan, Zheng Acked-by: David Howells --- fs/fscache/page.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'fs') diff --git a/fs/fscache/page.c b/fs/fscache/page.c index 3078b679fcd1..c8c4f79c7ce1 100644 --- a/fs/fscache/page.c +++ b/fs/fscache/page.c @@ -887,6 +887,8 @@ void fscache_invalidate_writes(struct fscache_cookie *cookie) put_page(results[i]); } + wake_up_bit(&cookie->flags, 0); + _leave(""); } -- cgit v1.3-8-gc7d7 From 480ce08a70e4179f34808a3bdbfe6627f624cf54 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Fri, 20 May 2016 18:32:31 +0800 Subject: FS-Cache: make check_consistency callback return int __fscache_check_consistency() calls check_consistency() callback and return the callback's return value. But the return type of check_consistency() is bool. So __fscache_check_consistency() return 1 if the cache is inconsistent. This is inconsistent with the document. Signed-off-by: Yan, Zheng Acked-by: David Howells --- fs/cachefiles/interface.c | 2 +- include/linux/fscache-cache.h | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/cachefiles/interface.c b/fs/cachefiles/interface.c index 861d611b8c05..ce5f345d70f5 100644 --- a/fs/cachefiles/interface.c +++ b/fs/cachefiles/interface.c @@ -380,7 +380,7 @@ static void cachefiles_sync_cache(struct fscache_cache *_cache) * check if the backing cache is updated to FS-Cache * - called by FS-Cache when evaluates if need to invalidate the cache */ -static bool cachefiles_check_consistency(struct fscache_operation *op) +static int cachefiles_check_consistency(struct fscache_operation *op) { struct cachefiles_object *object; struct cachefiles_cache *cache; diff --git a/include/linux/fscache-cache.h b/include/linux/fscache-cache.h index 604e1526cd00..13ba552e6c09 100644 --- a/include/linux/fscache-cache.h +++ b/include/linux/fscache-cache.h @@ -241,7 +241,7 @@ struct fscache_cache_ops { /* check the consistency between the backing cache and the FS-Cache * cookie */ - bool (*check_consistency)(struct fscache_operation *op); + int (*check_consistency)(struct fscache_operation *op); /* store the updated auxiliary data on an object */ void (*update_object)(struct fscache_object *object); -- cgit v1.3-8-gc7d7 From 368e35857dfab264f512d040c4486c9b13297988 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Tue, 17 May 2016 11:58:02 +0800 Subject: ceph: call __fscache_uncache_page() if readpages fails If readpages fails, fscache needs to cleanup its internal state. Signed-off-by: Yan, Zheng --- fs/ceph/addr.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c index eeb71e5de27a..4ff62fc98fb5 100644 --- a/fs/ceph/addr.c +++ b/fs/ceph/addr.c @@ -276,8 +276,10 @@ static void finish_read(struct ceph_osd_request *req) for (i = 0; i < num_pages; i++) { struct page *page = osd_data->pages[i]; - if (rc < 0 && rc != -ENOENT) + if (rc < 0 && rc != -ENOENT) { + ceph_fscache_readpage_cancel(inode, page); goto unlock; + } if (bytes < (int)PAGE_SIZE) { /* zero (remainder of) page */ int s = bytes < 0 ? 0 : bytes; -- cgit v1.3-8-gc7d7 From 1464975816c79a7cd28dc314384f060a122a9d55 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Fri, 20 May 2016 15:41:20 +0800 Subject: ceph: avoid unnecessary fscache invalidation/revlidation ceph_fill_file_size() has already called ceph_fscache_invalidate() if it return true. Signed-off-by: Yan, Zheng --- fs/ceph/caps.c | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index c17b5d76d75e..7bdf7d59a36d 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -2993,10 +2993,9 @@ static void handle_cap_grant(struct ceph_mds_client *mdsc, if (fill_inline) ceph_fill_inline_data(inode, NULL, inline_data, inline_len); - if (queue_trunc) { + if (queue_trunc) ceph_queue_vmtruncate(inode); - ceph_queue_revalidate(inode); - } else if (queue_revalidate) + else if (queue_revalidate) ceph_queue_revalidate(inode); if (writeback) @@ -3199,10 +3198,8 @@ static void handle_cap_trunc(struct inode *inode, truncate_seq, truncate_size, size); spin_unlock(&ci->i_ceph_lock); - if (queue_trunc) { + if (queue_trunc) ceph_queue_vmtruncate(inode); - ceph_fscache_invalidate(inode); - } } /* -- cgit v1.3-8-gc7d7 From 46b59b2be05a71d80d76883d2f495f182d768f47 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Wed, 18 May 2016 15:25:03 +0800 Subject: ceph: disable fscache when inode is opened for write All other filesystems do not add dirty pages to fscache. They all disable fscache when inode is opened for write. Only ceph adds dirty pages to fscache, but the code is buggy. Signed-off-by: Yan, Zheng --- fs/ceph/addr.c | 2 -- fs/ceph/cache.c | 54 ++++++++++++++++++++++++++++++++++++++---------------- fs/ceph/cache.h | 30 ++++++++++++------------------ fs/ceph/file.c | 19 ++----------------- 4 files changed, 52 insertions(+), 53 deletions(-) (limited to 'fs') diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c index 4ff62fc98fb5..26a9d10d75e9 100644 --- a/fs/ceph/addr.c +++ b/fs/ceph/addr.c @@ -537,8 +537,6 @@ static int writepage_nounlock(struct page *page, struct writeback_control *wbc) CONGESTION_ON_THRESH(fsc->mount_options->congestion_kb)) set_bdi_congested(&fsc->backing_dev_info, BLK_RW_ASYNC); - ceph_readpage_to_fscache(inode, page); - set_page_writeback(page); err = ceph_osdc_writepages(osdc, ceph_vino(inode), &ci->i_layout, snapc, diff --git a/fs/ceph/cache.c b/fs/ceph/cache.c index c052b5bf219b..c19db6afd0cc 100644 --- a/fs/ceph/cache.c +++ b/fs/ceph/cache.c @@ -181,32 +181,26 @@ static const struct fscache_cookie_def ceph_fscache_inode_object_def = { .now_uncached = ceph_fscache_inode_now_uncached, }; -void ceph_fscache_register_inode_cookie(struct ceph_fs_client* fsc, - struct ceph_inode_info* ci) +void ceph_fscache_register_inode_cookie(struct inode *inode) { - struct inode* inode = &ci->vfs_inode; + struct ceph_inode_info *ci = ceph_inode(inode); + struct ceph_fs_client *fsc = ceph_inode_to_client(inode); /* No caching for filesystem */ if (fsc->fscache == NULL) return; /* Only cache for regular files that are read only */ - if ((ci->vfs_inode.i_mode & S_IFREG) == 0) + if (!S_ISREG(inode->i_mode)) return; - /* Avoid multiple racing open requests */ - inode_lock(inode); - - if (ci->fscache) - goto done; - - ci->fscache = fscache_acquire_cookie(fsc->fscache, - &ceph_fscache_inode_object_def, - ci, true); - fscache_check_consistency(ci->fscache); -done: + inode_lock_nested(inode, I_MUTEX_CHILD); + if (!ci->fscache) { + ci->fscache = fscache_acquire_cookie(fsc->fscache, + &ceph_fscache_inode_object_def, + ci, false); + } inode_unlock(inode); - } void ceph_fscache_unregister_inode_cookie(struct ceph_inode_info* ci) @@ -222,6 +216,34 @@ void ceph_fscache_unregister_inode_cookie(struct ceph_inode_info* ci) fscache_relinquish_cookie(cookie, 0); } +static bool ceph_fscache_can_enable(void *data) +{ + struct inode *inode = data; + return !inode_is_open_for_write(inode); +} + +void ceph_fscache_file_set_cookie(struct inode *inode, struct file *filp) +{ + struct ceph_inode_info *ci = ceph_inode(inode); + + if (!fscache_cookie_valid(ci->fscache)) + return; + + if (inode_is_open_for_write(inode)) { + dout("fscache_file_set_cookie %p %p disabling cache\n", + inode, filp); + fscache_disable_cookie(ci->fscache, false); + fscache_uncache_all_inode_pages(ci->fscache, inode); + } else { + fscache_enable_cookie(ci->fscache, ceph_fscache_can_enable, + inode); + if (fscache_cookie_enabled(ci->fscache)) { + dout("fscache_file_set_cookie %p %p enabing cache\n", + inode, filp); + } + } +} + static void ceph_vfs_readpage_complete(struct page *page, void *data, int error) { if (!error) diff --git a/fs/ceph/cache.h b/fs/ceph/cache.h index 5ac591bd012b..dfe9b5823bd3 100644 --- a/fs/ceph/cache.h +++ b/fs/ceph/cache.h @@ -35,9 +35,9 @@ int ceph_fscache_register_fs(struct ceph_fs_client* fsc); void ceph_fscache_unregister_fs(struct ceph_fs_client* fsc); void ceph_fscache_inode_init(struct ceph_inode_info *ci); -void ceph_fscache_register_inode_cookie(struct ceph_fs_client* fsc, - struct ceph_inode_info* ci); +void ceph_fscache_register_inode_cookie(struct inode *inode); void ceph_fscache_unregister_inode_cookie(struct ceph_inode_info* ci); +void ceph_fscache_file_set_cookie(struct inode *inode, struct file *filp); int ceph_readpage_from_fscache(struct inode *inode, struct page *page); int ceph_readpages_from_fscache(struct inode *inode, @@ -48,12 +48,6 @@ void ceph_readpage_to_fscache(struct inode *inode, struct page *page); void ceph_invalidate_fscache_page(struct inode* inode, struct page *page); void ceph_queue_revalidate(struct inode *inode); -static inline void ceph_fscache_update_objectsize(struct inode *inode) -{ - struct ceph_inode_info *ci = ceph_inode(inode); - fscache_attr_changed(ci->fscache); -} - static inline void ceph_fscache_invalidate(struct inode *inode) { fscache_invalidate(ceph_inode(inode)->fscache); @@ -112,8 +106,16 @@ static inline void ceph_fscache_inode_init(struct ceph_inode_info *ci) { } -static inline void ceph_fscache_register_inode_cookie(struct ceph_fs_client* parent_fsc, - struct ceph_inode_info* ci) +static inline void ceph_fscache_register_inode_cookie(struct inode *inode) +{ +} + +static inline void ceph_fscache_unregister_inode_cookie(struct ceph_inode_info* ci) +{ +} + +static inline void ceph_fscache_file_set_cookie(struct inode *inode, + struct file *filp) { } @@ -141,10 +143,6 @@ static inline void ceph_readpage_to_fscache(struct inode *inode, { } -static inline void ceph_fscache_update_objectsize(struct inode *inode) -{ -} - static inline void ceph_fscache_invalidate(struct inode *inode) { } @@ -154,10 +152,6 @@ static inline void ceph_invalidate_fscache_page(struct inode *inode, { } -static inline void ceph_fscache_unregister_inode_cookie(struct ceph_inode_info* ci) -{ -} - static inline int ceph_release_fscache_page(struct page *page, gfp_t gfp) { return 1; diff --git a/fs/ceph/file.c b/fs/ceph/file.c index 8eeb9f579db5..ce2f5795e44b 100644 --- a/fs/ceph/file.c +++ b/fs/ceph/file.c @@ -137,23 +137,11 @@ static int ceph_init_file(struct inode *inode, struct file *file, int fmode) { struct ceph_file_info *cf; int ret = 0; - struct ceph_inode_info *ci = ceph_inode(inode); - struct ceph_fs_client *fsc = ceph_sb_to_client(inode->i_sb); - struct ceph_mds_client *mdsc = fsc->mdsc; switch (inode->i_mode & S_IFMT) { case S_IFREG: - /* First file open request creates the cookie, we want to keep - * this cookie around for the filetime of the inode as not to - * have to worry about fscache register / revoke / operation - * races. - * - * Also, if we know the operation is going to invalidate data - * (non readonly) just nuke the cache right away. - */ - ceph_fscache_register_inode_cookie(mdsc->fsc, ci); - if ((fmode & CEPH_FILE_MODE_WR)) - ceph_fscache_invalidate(inode); + ceph_fscache_register_inode_cookie(inode); + ceph_fscache_file_set_cookie(inode, file); case S_IFDIR: dout("init_file %p %p 0%o (regular)\n", inode, file, inode->i_mode); @@ -1407,7 +1395,6 @@ retry_snap: iov_iter_advance(from, written); ceph_put_snap_context(snapc); } else { - loff_t old_size = i_size_read(inode); /* * No need to acquire the i_truncate_mutex. Because * the MDS revokes Fwb caps before sending truncate @@ -1418,8 +1405,6 @@ retry_snap: written = generic_perform_write(file, from, pos); if (likely(written >= 0)) iocb->ki_pos = pos + written; - if (i_size_read(inode) > old_size) - ceph_fscache_update_objectsize(inode); inode_unlock(inode); } -- cgit v1.3-8-gc7d7 From f7f7e7a0635dedd5064fba255cb3facfa87b06d6 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Wed, 18 May 2016 20:31:55 +0800 Subject: ceph: improve fscache revalidation There are several issues in fscache revalidation code. - In ceph_revalidate_work(), fscache_invalidate() is called when fscache_check_consistency() return 0. This is complete wrong because 0 means cache is valid. - Handle_cap_grant() calls ceph_queue_revalidate() if client already has CAP_FILE_CACHE. This code is confusing. Client should revalidate the cache each time it got CAP_FILE_CACHE anew. - In Handle_cap_grant(), fscache_invalidate() is called if MDS revokes CAP_FILE_CACHE. This is inconsistency with the case that inode get evicted. In the later case, the cache is not discarded. Client may use the cache when inode is reloaded. This patch moves the fscache revalidation into ceph_get_caps(). Client revalidates the cache after it gets CAP_FILE_CACHE. i_rdcache_gen should keep constance while CAP_FILE_CACHE is used. If i_fscache_gen is not equal to i_rdcache_gen, client needs to check cache's consistency. Signed-off-by: Yan, Zheng --- fs/ceph/cache.c | 84 ++++++++++++--------------------------------------------- fs/ceph/cache.h | 20 +++++++++++--- fs/ceph/caps.c | 16 +++++------ fs/ceph/super.h | 4 +-- 4 files changed, 41 insertions(+), 83 deletions(-) (limited to 'fs') diff --git a/fs/ceph/cache.c b/fs/ceph/cache.c index c19db6afd0cc..5b3f4828f214 100644 --- a/fs/ceph/cache.c +++ b/fs/ceph/cache.c @@ -69,15 +69,8 @@ int ceph_fscache_register_fs(struct ceph_fs_client* fsc) fsc->fscache = fscache_acquire_cookie(ceph_cache_netfs.primary_index, &ceph_fscache_fsid_object_def, fsc, true); - - if (fsc->fscache == NULL) { + if (!fsc->fscache) pr_err("Unable to resgister fsid: %p fscache cookie", fsc); - return 0; - } - - fsc->revalidate_wq = alloc_workqueue("ceph-revalidate", 0, 1); - if (fsc->revalidate_wq == NULL) - return -ENOMEM; return 0; } @@ -260,8 +253,7 @@ static void ceph_vfs_readpage_complete_unlock(struct page *page, void *data, int static inline bool cache_valid(struct ceph_inode_info *ci) { - return ((ceph_caps_issued(ci) & CEPH_CAP_FILE_CACHE) && - (ci->i_fscache_gen == ci->i_rdcache_gen)); + return ci->i_fscache_gen == ci->i_rdcache_gen; } @@ -354,69 +346,27 @@ void ceph_invalidate_fscache_page(struct inode* inode, struct page *page) void ceph_fscache_unregister_fs(struct ceph_fs_client* fsc) { - if (fsc->revalidate_wq) - destroy_workqueue(fsc->revalidate_wq); - fscache_relinquish_cookie(fsc->fscache, 0); fsc->fscache = NULL; } -static void ceph_revalidate_work(struct work_struct *work) -{ - int issued; - u32 orig_gen; - struct ceph_inode_info *ci = container_of(work, struct ceph_inode_info, - i_revalidate_work); - struct inode *inode = &ci->vfs_inode; - - spin_lock(&ci->i_ceph_lock); - issued = __ceph_caps_issued(ci, NULL); - orig_gen = ci->i_rdcache_gen; - spin_unlock(&ci->i_ceph_lock); - - if (!(issued & CEPH_CAP_FILE_CACHE)) { - dout("revalidate_work lost cache before validation %p\n", - inode); - goto out; - } - - if (!fscache_check_consistency(ci->fscache)) - fscache_invalidate(ci->fscache); - - spin_lock(&ci->i_ceph_lock); - /* Update the new valid generation (backwards sanity check too) */ - if (orig_gen > ci->i_fscache_gen) { - ci->i_fscache_gen = orig_gen; - } - spin_unlock(&ci->i_ceph_lock); - -out: - iput(&ci->vfs_inode); -} - -void ceph_queue_revalidate(struct inode *inode) +/* + * caller should hold CEPH_CAP_FILE_{RD,CACHE} + */ +void ceph_fscache_revalidate_cookie(struct ceph_inode_info *ci) { - struct ceph_fs_client *fsc = ceph_sb_to_client(inode->i_sb); - struct ceph_inode_info *ci = ceph_inode(inode); - - if (fsc->revalidate_wq == NULL || ci->fscache == NULL) + if (cache_valid(ci)) return; - ihold(inode); - - if (queue_work(ceph_sb_to_client(inode->i_sb)->revalidate_wq, - &ci->i_revalidate_work)) { - dout("ceph_queue_revalidate %p\n", inode); - } else { - dout("ceph_queue_revalidate %p failed\n)", inode); - iput(inode); + /* resue i_truncate_mutex. There should be no pending + * truncate while the caller holds CEPH_CAP_FILE_RD */ + mutex_lock(&ci->i_truncate_mutex); + if (!cache_valid(ci)) { + if (fscache_check_consistency(ci->fscache)) + fscache_invalidate(ci->fscache); + spin_lock(&ci->i_ceph_lock); + ci->i_fscache_gen = ci->i_rdcache_gen; + spin_unlock(&ci->i_ceph_lock); } -} - -void ceph_fscache_inode_init(struct ceph_inode_info *ci) -{ - ci->fscache = NULL; - /* The first load is verifed cookie open time */ - ci->i_fscache_gen = 1; - INIT_WORK(&ci->i_revalidate_work, ceph_revalidate_work); + mutex_unlock(&ci->i_truncate_mutex); } diff --git a/fs/ceph/cache.h b/fs/ceph/cache.h index dfe9b5823bd3..7e72c7594f0c 100644 --- a/fs/ceph/cache.h +++ b/fs/ceph/cache.h @@ -34,10 +34,10 @@ void ceph_fscache_unregister(void); int ceph_fscache_register_fs(struct ceph_fs_client* fsc); void ceph_fscache_unregister_fs(struct ceph_fs_client* fsc); -void ceph_fscache_inode_init(struct ceph_inode_info *ci); void ceph_fscache_register_inode_cookie(struct inode *inode); void ceph_fscache_unregister_inode_cookie(struct ceph_inode_info* ci); void ceph_fscache_file_set_cookie(struct inode *inode, struct file *filp); +void ceph_fscache_revalidate_cookie(struct ceph_inode_info *ci); int ceph_readpage_from_fscache(struct inode *inode, struct page *page); int ceph_readpages_from_fscache(struct inode *inode, @@ -46,7 +46,12 @@ int ceph_readpages_from_fscache(struct inode *inode, unsigned *nr_pages); void ceph_readpage_to_fscache(struct inode *inode, struct page *page); void ceph_invalidate_fscache_page(struct inode* inode, struct page *page); -void ceph_queue_revalidate(struct inode *inode); + +static inline void ceph_fscache_inode_init(struct ceph_inode_info *ci) +{ + ci->fscache = NULL; + ci->i_fscache_gen = 0; +} static inline void ceph_fscache_invalidate(struct inode *inode) { @@ -82,6 +87,11 @@ static inline void ceph_fscache_readpages_cancel(struct inode *inode, return fscache_readpages_cancel(ci->fscache, pages); } +static inline void ceph_disable_fscache_readpage(struct ceph_inode_info *ci) +{ + ci->i_fscache_gen = ci->i_rdcache_gen - 1; +} + #else static inline int ceph_fscache_register(void) @@ -119,6 +129,10 @@ static inline void ceph_fscache_file_set_cookie(struct inode *inode, { } +static inline void ceph_fscache_revalidate_cookie(struct ceph_inode_info *ci) +{ +} + static inline void ceph_fscache_uncache_page(struct inode *inode, struct page *pages) { @@ -167,7 +181,7 @@ static inline void ceph_fscache_readpages_cancel(struct inode *inode, { } -static inline void ceph_queue_revalidate(struct inode *inode) +static inline void ceph_disable_fscache_readpage(struct ceph_inode_info *ci) { } diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index 7bdf7d59a36d..6f60d0a3d0f9 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -2393,6 +2393,9 @@ again: snap_rwsem_locked = true; } *got = need | (have & want); + if ((need & CEPH_CAP_FILE_RD) && + !(*got & CEPH_CAP_FILE_CACHE)) + ceph_disable_fscache_readpage(ci); __take_cap_refs(ci, *got, true); ret = 1; } @@ -2554,6 +2557,9 @@ int ceph_get_caps(struct ceph_inode_info *ci, int need, int want, break; } + if ((_got & CEPH_CAP_FILE_RD) && (_got & CEPH_CAP_FILE_CACHE)) + ceph_fscache_revalidate_cookie(ci); + *got = _got; return 0; } @@ -2795,7 +2801,6 @@ static void handle_cap_grant(struct ceph_mds_client *mdsc, bool writeback = false; bool queue_trunc = false; bool queue_invalidate = false; - bool queue_revalidate = false; bool deleted_inode = false; bool fill_inline = false; @@ -2837,8 +2842,6 @@ static void handle_cap_grant(struct ceph_mds_client *mdsc, ci->i_rdcache_revoking = ci->i_rdcache_gen; } } - - ceph_fscache_invalidate(inode); } /* side effects now are allowed */ @@ -2880,11 +2883,6 @@ static void handle_cap_grant(struct ceph_mds_client *mdsc, } } - /* Do we need to revalidate our fscache cookie. Don't bother on the - * first cache cap as we already validate at cookie creation time. */ - if ((issued & CEPH_CAP_FILE_CACHE) && ci->i_rdcache_gen > 1) - queue_revalidate = true; - if (newcaps & CEPH_CAP_ANY_RD) { /* ctime/mtime/atime? */ ceph_decode_timespec(&mtime, &grant->mtime); @@ -2995,8 +2993,6 @@ static void handle_cap_grant(struct ceph_mds_client *mdsc, if (queue_trunc) ceph_queue_vmtruncate(inode); - else if (queue_revalidate) - ceph_queue_revalidate(inode); if (writeback) /* diff --git a/fs/ceph/super.h b/fs/ceph/super.h index 0130a8592191..0168b49fb6ad 100644 --- a/fs/ceph/super.h +++ b/fs/ceph/super.h @@ -103,7 +103,6 @@ struct ceph_fs_client { #ifdef CONFIG_CEPH_FSCACHE struct fscache_cookie *fscache; - struct workqueue_struct *revalidate_wq; #endif }; @@ -360,8 +359,7 @@ struct ceph_inode_info { #ifdef CONFIG_CEPH_FSCACHE struct fscache_cookie *fscache; - u32 i_fscache_gen; /* sequence, for delayed fscache validate */ - struct work_struct i_revalidate_work; + u32 i_fscache_gen; #endif struct inode vfs_inode; /* at end */ }; -- cgit v1.3-8-gc7d7 From f6973c09490c919398fc87d9c6bec9c0b670c4c4 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Fri, 20 May 2016 16:57:29 +0800 Subject: ceph: use i_version to check validity of fscache Signed-off-by: Yan, Zheng --- fs/ceph/cache.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'fs') diff --git a/fs/ceph/cache.c b/fs/ceph/cache.c index 5b3f4828f214..238c55b01723 100644 --- a/fs/ceph/cache.c +++ b/fs/ceph/cache.c @@ -25,6 +25,7 @@ #include "cache.h" struct ceph_aux_inode { + u64 version; struct timespec mtime; loff_t size; }; @@ -98,6 +99,7 @@ static uint16_t ceph_fscache_inode_get_aux(const void *cookie_netfs_data, const struct inode* inode = &ci->vfs_inode; memset(&aux, 0, sizeof(aux)); + aux.version = ci->i_version; aux.mtime = inode->i_mtime; aux.size = i_size_read(inode); @@ -124,6 +126,7 @@ static enum fscache_checkaux ceph_fscache_inode_check_aux( return FSCACHE_CHECKAUX_OBSOLETE; memset(&aux, 0, sizeof(aux)); + aux.version = ci->i_version; aux.mtime = inode->i_mtime; aux.size = i_size_read(inode); -- cgit v1.3-8-gc7d7 From 5473e0c426ffaeaa19734987b153c2a7f33b8706 Mon Sep 17 00:00:00 2001 From: Feifei Xu Date: Wed, 1 Jun 2016 19:18:23 +0800 Subject: Btrfs: test_check_exists: Fix infinite loop when searching for free space entries On a ppc64 machine using 64K as the block size, assume that the RB tree at btrfs_free_space_ctl->free_space_offset contains following two entries: 1. A bitmap entry having an offset value of 0 and having the bits corresponding to the address range [128M+512K, 128M+768K] set. 2. An extent entry corresponding to the address range [128M-256K, 128M-128K] In such a scenario, test_check_exists() invoked for checking the existence of address range [128M+768K, 256M] can lead to an infinite loop as explained below: - Checking for the extent entry fails. - Checking for a bitmap entry results in the free space info in range [128M+512K, 128M+768K] beng returned. - rb_prev(info) returns NULL because the bitmap entry starting from offset 0 comes first in the RB tree. - current_node = bitmap node. - while (current_node) tmp = rb_next(bitmap_node);/*tmp is extent based free space entry*/ Since extent based free space entry's last address is smaller than the address being searched for (i.e. 128M+768K) we incorrectly again obtain the extent node as the "next right node" of the RB tree and thus end up looping infinitely. This patch fixes the issue by checking the "tmp" variable which point to the most recently searched free space node. Reviewed-by: Josef Bacik Reviewed-by: Chandan Rajendra Signed-off-by: Feifei Xu Signed-off-by: David Sterba --- fs/btrfs/free-space-cache.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c index c6dc1183f542..fa623359d5b8 100644 --- a/fs/btrfs/free-space-cache.c +++ b/fs/btrfs/free-space-cache.c @@ -3662,7 +3662,7 @@ have_info: if (tmp->offset + tmp->bytes < offset) break; if (offset + bytes < tmp->offset) { - n = rb_prev(&info->offset_index); + n = rb_prev(&tmp->offset_index); continue; } info = tmp; @@ -3676,7 +3676,7 @@ have_info: if (offset + bytes < tmp->offset) break; if (tmp->offset + tmp->bytes < offset) { - n = rb_next(&info->offset_index); + n = rb_next(&tmp->offset_index); continue; } info = tmp; -- cgit v1.3-8-gc7d7 From 0ef6447a3d2f014e49069c4da33f905ed803aa2a Mon Sep 17 00:00:00 2001 From: Feifei Xu Date: Wed, 1 Jun 2016 19:18:24 +0800 Subject: Btrfs: Fix integer overflow when calculating bytes_per_bitmap On ppc64, bytes_per_bitmap will be (65536*8*65536). Hence append UL to fix integer overflow. Reviewed-by: Josef Bacik Reviewed-by: Chandan Rajendra Signed-off-by: Feifei Xu Signed-off-by: David Sterba --- fs/btrfs/free-space-cache.c | 12 ++++++------ fs/btrfs/tests/free-space-tests.c | 2 +- 2 files changed, 7 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c index fa623359d5b8..2813ef0718a2 100644 --- a/fs/btrfs/free-space-cache.c +++ b/fs/btrfs/free-space-cache.c @@ -29,7 +29,7 @@ #include "inode-map.h" #include "volumes.h" -#define BITS_PER_BITMAP (PAGE_SIZE * 8) +#define BITS_PER_BITMAP (PAGE_SIZE * 8UL) #define MAX_CACHE_BYTES_PER_GIG SZ_32K struct btrfs_trim_range { @@ -1415,11 +1415,11 @@ static inline u64 offset_to_bitmap(struct btrfs_free_space_ctl *ctl, u64 offset) { u64 bitmap_start; - u32 bytes_per_bitmap; + u64 bytes_per_bitmap; bytes_per_bitmap = BITS_PER_BITMAP * ctl->unit; bitmap_start = offset - ctl->start; - bitmap_start = div_u64(bitmap_start, bytes_per_bitmap); + bitmap_start = div64_u64(bitmap_start, bytes_per_bitmap); bitmap_start *= bytes_per_bitmap; bitmap_start += ctl->start; @@ -1638,10 +1638,10 @@ static void recalculate_thresholds(struct btrfs_free_space_ctl *ctl) u64 bitmap_bytes; u64 extent_bytes; u64 size = block_group->key.offset; - u32 bytes_per_bg = BITS_PER_BITMAP * ctl->unit; - u32 max_bitmaps = div_u64(size + bytes_per_bg - 1, bytes_per_bg); + u64 bytes_per_bg = BITS_PER_BITMAP * ctl->unit; + u64 max_bitmaps = div64_u64(size + bytes_per_bg - 1, bytes_per_bg); - max_bitmaps = max_t(u32, max_bitmaps, 1); + max_bitmaps = max_t(u64, max_bitmaps, 1); ASSERT(ctl->total_bitmaps <= max_bitmaps); diff --git a/fs/btrfs/tests/free-space-tests.c b/fs/btrfs/tests/free-space-tests.c index 0eeb8f3d6b67..f3756d6b9ba2 100644 --- a/fs/btrfs/tests/free-space-tests.c +++ b/fs/btrfs/tests/free-space-tests.c @@ -22,7 +22,7 @@ #include "../disk-io.h" #include "../free-space-cache.h" -#define BITS_PER_BITMAP (PAGE_SIZE * 8) +#define BITS_PER_BITMAP (PAGE_SIZE * 8UL) /* * This test just does basic sanity checking, making sure we can add an extent -- cgit v1.3-8-gc7d7 From b9ef22dedde08ab1b4ccd5f53344984c4dcb89f4 Mon Sep 17 00:00:00 2001 From: Feifei Xu Date: Wed, 1 Jun 2016 19:18:25 +0800 Subject: Btrfs: self-tests: Support non-4k page size self-tests code assumes 4k as the sectorsize and nodesize. This commit fix hardcoded 4K. Enables the self-tests code to be executed on non-4k page sized systems (e.g. ppc64). Reviewed-by: Josef Bacik Signed-off-by: Feifei Xu Signed-off-by: Chandan Rajendra Signed-off-by: David Sterba --- fs/btrfs/ctree.c | 6 +- fs/btrfs/disk-io.c | 8 +- fs/btrfs/disk-io.h | 2 +- fs/btrfs/extent_io.c | 10 +- fs/btrfs/extent_io.h | 4 +- fs/btrfs/free-space-cache.c | 2 +- fs/btrfs/super.c | 16 +- fs/btrfs/tests/btrfs-tests.c | 6 +- fs/btrfs/tests/btrfs-tests.h | 27 +-- fs/btrfs/tests/extent-buffer-tests.c | 11 +- fs/btrfs/tests/extent-io-tests.c | 27 +-- fs/btrfs/tests/free-space-tests.c | 68 ++++--- fs/btrfs/tests/free-space-tree-tests.c | 30 +-- fs/btrfs/tests/inode-tests.c | 344 ++++++++++++++++++--------------- fs/btrfs/tests/qgroup-tests.c | 85 ++++---- 15 files changed, 355 insertions(+), 291 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c index 427c36b430a6..46025688f1d0 100644 --- a/fs/btrfs/ctree.c +++ b/fs/btrfs/ctree.c @@ -1373,7 +1373,8 @@ tree_mod_log_rewind(struct btrfs_fs_info *fs_info, struct btrfs_path *path, if (tm->op == MOD_LOG_KEY_REMOVE_WHILE_FREEING) { BUG_ON(tm->slot != 0); - eb_rewin = alloc_dummy_extent_buffer(fs_info, eb->start); + eb_rewin = alloc_dummy_extent_buffer(fs_info, eb->start, + eb->len); if (!eb_rewin) { btrfs_tree_read_unlock_blocking(eb); free_extent_buffer(eb); @@ -1454,7 +1455,8 @@ get_old_root(struct btrfs_root *root, u64 time_seq) } else if (old_root) { btrfs_tree_read_unlock(eb_root); free_extent_buffer(eb_root); - eb = alloc_dummy_extent_buffer(root->fs_info, logical); + eb = alloc_dummy_extent_buffer(root->fs_info, logical, + root->nodesize); } else { btrfs_set_lock_blocking_rw(eb_root, BTRFS_READ_LOCK); eb = btrfs_clone_extent_buffer(eb_root); diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 6628fca9f4ed..06ef433552b3 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1147,7 +1147,8 @@ struct extent_buffer *btrfs_find_create_tree_block(struct btrfs_root *root, u64 bytenr) { if (btrfs_test_is_dummy_root(root)) - return alloc_test_extent_buffer(root->fs_info, bytenr); + return alloc_test_extent_buffer(root->fs_info, bytenr, + root->nodesize); return alloc_extent_buffer(root->fs_info, bytenr); } @@ -1314,14 +1315,15 @@ static struct btrfs_root *btrfs_alloc_root(struct btrfs_fs_info *fs_info, #ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS /* Should only be used by the testing infrastructure */ -struct btrfs_root *btrfs_alloc_dummy_root(void) +struct btrfs_root *btrfs_alloc_dummy_root(u32 sectorsize, u32 nodesize) { struct btrfs_root *root; root = btrfs_alloc_root(NULL, GFP_KERNEL); if (!root) return ERR_PTR(-ENOMEM); - __setup_root(4096, 4096, 4096, root, NULL, 1); + /* We don't use the stripesize in selftest, set it as sectorsize */ + __setup_root(nodesize, sectorsize, sectorsize, root, NULL, 1); set_bit(BTRFS_ROOT_DUMMY_ROOT, &root->state); root->alloc_bytenr = 0; diff --git a/fs/btrfs/disk-io.h b/fs/btrfs/disk-io.h index 8e79d0070bcf..acba821499a9 100644 --- a/fs/btrfs/disk-io.h +++ b/fs/btrfs/disk-io.h @@ -90,7 +90,7 @@ void btrfs_drop_and_free_fs_root(struct btrfs_fs_info *fs_info, void btrfs_free_fs_root(struct btrfs_root *root); #ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS -struct btrfs_root *btrfs_alloc_dummy_root(void); +struct btrfs_root *btrfs_alloc_dummy_root(u32 sectorsize, u32 nodesize); #endif /* diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 3cd57825c75f..b0a554d4204a 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -4718,16 +4718,16 @@ err: } struct extent_buffer *alloc_dummy_extent_buffer(struct btrfs_fs_info *fs_info, - u64 start) + u64 start, u32 nodesize) { unsigned long len; if (!fs_info) { /* * Called only from tests that don't always have a fs_info - * available, but we know that nodesize is 4096 + * available */ - len = 4096; + len = nodesize; } else { len = fs_info->tree_root->nodesize; } @@ -4823,7 +4823,7 @@ struct extent_buffer *find_extent_buffer(struct btrfs_fs_info *fs_info, #ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS struct extent_buffer *alloc_test_extent_buffer(struct btrfs_fs_info *fs_info, - u64 start) + u64 start, u32 nodesize) { struct extent_buffer *eb, *exists = NULL; int ret; @@ -4831,7 +4831,7 @@ struct extent_buffer *alloc_test_extent_buffer(struct btrfs_fs_info *fs_info, eb = find_extent_buffer(fs_info, start); if (eb) return eb; - eb = alloc_dummy_extent_buffer(fs_info, start); + eb = alloc_dummy_extent_buffer(fs_info, start, nodesize); if (!eb) return NULL; eb->fs_info = fs_info; diff --git a/fs/btrfs/extent_io.h b/fs/btrfs/extent_io.h index 1baf19c9b79d..c0c1c4fef6ce 100644 --- a/fs/btrfs/extent_io.h +++ b/fs/btrfs/extent_io.h @@ -348,7 +348,7 @@ struct extent_buffer *alloc_extent_buffer(struct btrfs_fs_info *fs_info, struct extent_buffer *__alloc_dummy_extent_buffer(struct btrfs_fs_info *fs_info, u64 start, unsigned long len); struct extent_buffer *alloc_dummy_extent_buffer(struct btrfs_fs_info *fs_info, - u64 start); + u64 start, u32 nodesize); struct extent_buffer *btrfs_clone_extent_buffer(struct extent_buffer *src); struct extent_buffer *find_extent_buffer(struct btrfs_fs_info *fs_info, u64 start); @@ -468,5 +468,5 @@ noinline u64 find_lock_delalloc_range(struct inode *inode, u64 *end, u64 max_bytes); #endif struct extent_buffer *alloc_test_extent_buffer(struct btrfs_fs_info *fs_info, - u64 start); + u64 start, u32 nodesize); #endif diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c index 2813ef0718a2..69d270f6602c 100644 --- a/fs/btrfs/free-space-cache.c +++ b/fs/btrfs/free-space-cache.c @@ -1660,7 +1660,7 @@ static void recalculate_thresholds(struct btrfs_free_space_ctl *ctl) * sure we don't go over our overall goal of MAX_CACHE_BYTES_PER_GIG as * we add more bitmaps. */ - bitmap_bytes = (ctl->total_bitmaps + 1) * PAGE_SIZE; + bitmap_bytes = (ctl->total_bitmaps + 1) * ctl->unit; if (bitmap_bytes >= max_bytes) { ctl->extents_thresh = 0; diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index 4e59a91a11e0..5b0b354ca594 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -2319,27 +2319,31 @@ static void btrfs_print_mod_info(void) static int btrfs_run_sanity_tests(void) { int ret; + u32 sectorsize, nodesize; + sectorsize = PAGE_SIZE; + nodesize = PAGE_SIZE; ret = btrfs_init_test_fs(); if (ret) return ret; - ret = btrfs_test_free_space_cache(); + ret = btrfs_test_free_space_cache(sectorsize, nodesize); if (ret) goto out; - ret = btrfs_test_extent_buffer_operations(); + ret = btrfs_test_extent_buffer_operations(sectorsize, + nodesize); if (ret) goto out; - ret = btrfs_test_extent_io(); + ret = btrfs_test_extent_io(sectorsize, nodesize); if (ret) goto out; - ret = btrfs_test_inodes(); + ret = btrfs_test_inodes(sectorsize, nodesize); if (ret) goto out; - ret = btrfs_test_qgroups(); + ret = btrfs_test_qgroups(sectorsize, nodesize); if (ret) goto out; - ret = btrfs_test_free_space_tree(); + ret = btrfs_test_free_space_tree(sectorsize, nodesize); out: btrfs_destroy_test_fs(); return ret; diff --git a/fs/btrfs/tests/btrfs-tests.c b/fs/btrfs/tests/btrfs-tests.c index f54bf450bad3..10eb249ef891 100644 --- a/fs/btrfs/tests/btrfs-tests.c +++ b/fs/btrfs/tests/btrfs-tests.c @@ -175,7 +175,7 @@ void btrfs_free_dummy_root(struct btrfs_root *root) } struct btrfs_block_group_cache * -btrfs_alloc_dummy_block_group(unsigned long length) +btrfs_alloc_dummy_block_group(unsigned long length, u32 sectorsize) { struct btrfs_block_group_cache *cache; @@ -192,8 +192,8 @@ btrfs_alloc_dummy_block_group(unsigned long length) cache->key.objectid = 0; cache->key.offset = length; cache->key.type = BTRFS_BLOCK_GROUP_ITEM_KEY; - cache->sectorsize = 4096; - cache->full_stripe_len = 4096; + cache->sectorsize = sectorsize; + cache->full_stripe_len = sectorsize; INIT_LIST_HEAD(&cache->list); INIT_LIST_HEAD(&cache->cluster_list); diff --git a/fs/btrfs/tests/btrfs-tests.h b/fs/btrfs/tests/btrfs-tests.h index 054b8c73c951..66fb6b701eb7 100644 --- a/fs/btrfs/tests/btrfs-tests.h +++ b/fs/btrfs/tests/btrfs-tests.h @@ -26,27 +26,28 @@ struct btrfs_root; struct btrfs_trans_handle; -int btrfs_test_free_space_cache(void); -int btrfs_test_extent_buffer_operations(void); -int btrfs_test_extent_io(void); -int btrfs_test_inodes(void); -int btrfs_test_qgroups(void); -int btrfs_test_free_space_tree(void); +int btrfs_test_free_space_cache(u32 sectorsize, u32 nodesize); +int btrfs_test_extent_buffer_operations(u32 sectorsize, u32 nodesize); +int btrfs_test_extent_io(u32 sectorsize, u32 nodesize); +int btrfs_test_inodes(u32 sectorsize, u32 nodesize); +int btrfs_test_qgroups(u32 sectorsize, u32 nodesize); +int btrfs_test_free_space_tree(u32 sectorsize, u32 nodesize); int btrfs_init_test_fs(void); void btrfs_destroy_test_fs(void); struct inode *btrfs_new_test_inode(void); struct btrfs_fs_info *btrfs_alloc_dummy_fs_info(void); void btrfs_free_dummy_root(struct btrfs_root *root); struct btrfs_block_group_cache * -btrfs_alloc_dummy_block_group(unsigned long length); +btrfs_alloc_dummy_block_group(unsigned long length, u32 sectorsize); void btrfs_free_dummy_block_group(struct btrfs_block_group_cache *cache); void btrfs_init_dummy_trans(struct btrfs_trans_handle *trans); #else -static inline int btrfs_test_free_space_cache(void) +static inline int btrfs_test_free_space_cache(u32 sectorsize, u32 nodesize) { return 0; } -static inline int btrfs_test_extent_buffer_operations(void) +static inline int btrfs_test_extent_buffer_operations(u32 sectorsize, + u32 nodesize) { return 0; } @@ -57,19 +58,19 @@ static inline int btrfs_init_test_fs(void) static inline void btrfs_destroy_test_fs(void) { } -static inline int btrfs_test_extent_io(void) +static inline int btrfs_test_extent_io(u32 sectorsize, u32 nodesize) { return 0; } -static inline int btrfs_test_inodes(void) +static inline int btrfs_test_inodes(u32 sectorsize, u32 nodesize) { return 0; } -static inline int btrfs_test_qgroups(void) +static inline int btrfs_test_qgroups(u32 sectorsize, u32 nodesize) { return 0; } -static inline int btrfs_test_free_space_tree(void) +static inline int btrfs_test_free_space_tree(u32 sectorsize, u32 nodesize) { return 0; } diff --git a/fs/btrfs/tests/extent-buffer-tests.c b/fs/btrfs/tests/extent-buffer-tests.c index f51963a8f929..17b110f39f06 100644 --- a/fs/btrfs/tests/extent-buffer-tests.c +++ b/fs/btrfs/tests/extent-buffer-tests.c @@ -22,7 +22,7 @@ #include "../extent_io.h" #include "../disk-io.h" -static int test_btrfs_split_item(void) +static int test_btrfs_split_item(u32 sectorsize, u32 nodesize) { struct btrfs_path *path; struct btrfs_root *root; @@ -40,7 +40,7 @@ static int test_btrfs_split_item(void) test_msg("Running btrfs_split_item tests\n"); - root = btrfs_alloc_dummy_root(); + root = btrfs_alloc_dummy_root(sectorsize, nodesize); if (IS_ERR(root)) { test_msg("Could not allocate root\n"); return PTR_ERR(root); @@ -53,7 +53,8 @@ static int test_btrfs_split_item(void) return -ENOMEM; } - path->nodes[0] = eb = alloc_dummy_extent_buffer(NULL, 4096); + path->nodes[0] = eb = alloc_dummy_extent_buffer(NULL, nodesize, + nodesize); if (!eb) { test_msg("Could not allocate dummy buffer\n"); ret = -ENOMEM; @@ -222,8 +223,8 @@ out: return ret; } -int btrfs_test_extent_buffer_operations(void) +int btrfs_test_extent_buffer_operations(u32 sectorsize, u32 nodesize) { test_msg("Running extent buffer operation tests"); - return test_btrfs_split_item(); + return test_btrfs_split_item(sectorsize, nodesize); } diff --git a/fs/btrfs/tests/extent-io-tests.c b/fs/btrfs/tests/extent-io-tests.c index 55724607f79b..755fceb2e5f5 100644 --- a/fs/btrfs/tests/extent-io-tests.c +++ b/fs/btrfs/tests/extent-io-tests.c @@ -65,7 +65,7 @@ static noinline int process_page_range(struct inode *inode, u64 start, u64 end, return count; } -static int test_find_delalloc(void) +static int test_find_delalloc(u32 sectorsize) { struct inode *inode; struct extent_io_tree tmp; @@ -113,7 +113,7 @@ static int test_find_delalloc(void) * |--- delalloc ---| * |--- search ---| */ - set_extent_delalloc(&tmp, 0, 4095, NULL); + set_extent_delalloc(&tmp, 0, sectorsize - 1, NULL); start = 0; end = 0; found = find_lock_delalloc_range(inode, &tmp, locked_page, &start, @@ -122,9 +122,9 @@ static int test_find_delalloc(void) test_msg("Should have found at least one delalloc\n"); goto out_bits; } - if (start != 0 || end != 4095) { - test_msg("Expected start 0 end 4095, got start %Lu end %Lu\n", - start, end); + if (start != 0 || end != (sectorsize - 1)) { + test_msg("Expected start 0 end %u, got start %llu end %llu\n", + sectorsize - 1, start, end); goto out_bits; } unlock_extent(&tmp, start, end); @@ -144,7 +144,7 @@ static int test_find_delalloc(void) test_msg("Couldn't find the locked page\n"); goto out_bits; } - set_extent_delalloc(&tmp, 4096, max_bytes - 1, NULL); + set_extent_delalloc(&tmp, sectorsize, max_bytes - 1, NULL); start = test_start; end = 0; found = find_lock_delalloc_range(inode, &tmp, locked_page, &start, @@ -172,7 +172,7 @@ static int test_find_delalloc(void) * |--- delalloc ---| * |--- search ---| */ - test_start = max_bytes + 4096; + test_start = max_bytes + sectorsize; locked_page = find_lock_page(inode->i_mapping, test_start >> PAGE_SHIFT); if (!locked_page) { @@ -351,14 +351,15 @@ static int __test_eb_bitmaps(unsigned long *bitmap, struct extent_buffer *eb, return 0; } -static int test_eb_bitmaps(void) +static int test_eb_bitmaps(u32 sectorsize, u32 nodesize) { - unsigned long len = PAGE_SIZE * 4; + unsigned long len; unsigned long *bitmap; struct extent_buffer *eb; int ret; test_msg("Running extent buffer bitmap tests\n"); + len = sectorsize * 4; bitmap = kmalloc(len, GFP_KERNEL); if (!bitmap) { @@ -379,7 +380,7 @@ static int test_eb_bitmaps(void) /* Do it over again with an extent buffer which isn't page-aligned. */ free_extent_buffer(eb); - eb = __alloc_dummy_extent_buffer(NULL, PAGE_SIZE / 2, len); + eb = __alloc_dummy_extent_buffer(NULL, nodesize / 2, len); if (!eb) { test_msg("Couldn't allocate test extent buffer\n"); kfree(bitmap); @@ -393,17 +394,17 @@ out: return ret; } -int btrfs_test_extent_io(void) +int btrfs_test_extent_io(u32 sectorsize, u32 nodesize) { int ret; test_msg("Running extent I/O tests\n"); - ret = test_find_delalloc(); + ret = test_find_delalloc(sectorsize); if (ret) goto out; - ret = test_eb_bitmaps(); + ret = test_eb_bitmaps(sectorsize, nodesize); out: test_msg("Extent I/O tests finished\n"); return ret; diff --git a/fs/btrfs/tests/free-space-tests.c b/fs/btrfs/tests/free-space-tests.c index f3756d6b9ba2..029343b3e854 100644 --- a/fs/btrfs/tests/free-space-tests.c +++ b/fs/btrfs/tests/free-space-tests.c @@ -99,7 +99,8 @@ static int test_extents(struct btrfs_block_group_cache *cache) return 0; } -static int test_bitmaps(struct btrfs_block_group_cache *cache) +static int test_bitmaps(struct btrfs_block_group_cache *cache, + u32 sectorsize) { u64 next_bitmap_offset; int ret; @@ -139,7 +140,7 @@ static int test_bitmaps(struct btrfs_block_group_cache *cache) * The first bitmap we have starts at offset 0 so the next one is just * at the end of the first bitmap. */ - next_bitmap_offset = (u64)(BITS_PER_BITMAP * 4096); + next_bitmap_offset = (u64)(BITS_PER_BITMAP * sectorsize); /* Test a bit straddling two bitmaps */ ret = test_add_free_space_entry(cache, next_bitmap_offset - SZ_2M, @@ -167,9 +168,10 @@ static int test_bitmaps(struct btrfs_block_group_cache *cache) } /* This is the high grade jackassery */ -static int test_bitmaps_and_extents(struct btrfs_block_group_cache *cache) +static int test_bitmaps_and_extents(struct btrfs_block_group_cache *cache, + u32 sectorsize) { - u64 bitmap_offset = (u64)(BITS_PER_BITMAP * 4096); + u64 bitmap_offset = (u64)(BITS_PER_BITMAP * sectorsize); int ret; test_msg("Running bitmap and extent tests\n"); @@ -401,7 +403,8 @@ static int check_cache_empty(struct btrfs_block_group_cache *cache) * requests. */ static int -test_steal_space_from_bitmap_to_extent(struct btrfs_block_group_cache *cache) +test_steal_space_from_bitmap_to_extent(struct btrfs_block_group_cache *cache, + u32 sectorsize) { int ret; u64 offset; @@ -539,7 +542,7 @@ test_steal_space_from_bitmap_to_extent(struct btrfs_block_group_cache *cache) * The goal is to test that the bitmap entry space stealing doesn't * steal this space region. */ - ret = btrfs_add_free_space(cache, SZ_128M + SZ_16M, 4096); + ret = btrfs_add_free_space(cache, SZ_128M + SZ_16M, sectorsize); if (ret) { test_msg("Error adding free space: %d\n", ret); return ret; @@ -597,8 +600,8 @@ test_steal_space_from_bitmap_to_extent(struct btrfs_block_group_cache *cache) return -ENOENT; } - if (cache->free_space_ctl->free_space != (SZ_1M + 4096)) { - test_msg("Cache free space is not 1Mb + 4Kb\n"); + if (cache->free_space_ctl->free_space != (SZ_1M + sectorsize)) { + test_msg("Cache free space is not 1Mb + %u\n", sectorsize); return -EINVAL; } @@ -611,22 +614,25 @@ test_steal_space_from_bitmap_to_extent(struct btrfs_block_group_cache *cache) return -EINVAL; } - /* All that remains is a 4Kb free space region in a bitmap. Confirm. */ + /* + * All that remains is a sectorsize free space region in a bitmap. + * Confirm. + */ ret = check_num_extents_and_bitmaps(cache, 1, 1); if (ret) return ret; - if (cache->free_space_ctl->free_space != 4096) { - test_msg("Cache free space is not 4Kb\n"); + if (cache->free_space_ctl->free_space != sectorsize) { + test_msg("Cache free space is not %u\n", sectorsize); return -EINVAL; } offset = btrfs_find_space_for_alloc(cache, - 0, 4096, 0, + 0, sectorsize, 0, &max_extent_size); if (offset != (SZ_128M + SZ_16M)) { - test_msg("Failed to allocate 4Kb from space cache, returned offset is: %llu\n", - offset); + test_msg("Failed to allocate %u, returned offset : %llu\n", + sectorsize, offset); return -EINVAL; } @@ -733,7 +739,7 @@ test_steal_space_from_bitmap_to_extent(struct btrfs_block_group_cache *cache) * The goal is to test that the bitmap entry space stealing doesn't * steal this space region. */ - ret = btrfs_add_free_space(cache, SZ_32M, 8192); + ret = btrfs_add_free_space(cache, SZ_32M, 2 * sectorsize); if (ret) { test_msg("Error adding free space: %d\n", ret); return ret; @@ -757,7 +763,7 @@ test_steal_space_from_bitmap_to_extent(struct btrfs_block_group_cache *cache) /* * Confirm that our extent entry didn't stole all free space from the - * bitmap, because of the small 8Kb free space region. + * bitmap, because of the small 2 * sectorsize free space region. */ ret = check_num_extents_and_bitmaps(cache, 2, 1); if (ret) @@ -783,8 +789,8 @@ test_steal_space_from_bitmap_to_extent(struct btrfs_block_group_cache *cache) return -ENOENT; } - if (cache->free_space_ctl->free_space != (SZ_1M + 8192)) { - test_msg("Cache free space is not 1Mb + 8Kb\n"); + if (cache->free_space_ctl->free_space != (SZ_1M + 2 * sectorsize)) { + test_msg("Cache free space is not 1Mb + %u\n", 2 * sectorsize); return -EINVAL; } @@ -796,21 +802,25 @@ test_steal_space_from_bitmap_to_extent(struct btrfs_block_group_cache *cache) return -EINVAL; } - /* All that remains is a 8Kb free space region in a bitmap. Confirm. */ + /* + * All that remains is 2 * sectorsize free space region + * in a bitmap. Confirm. + */ ret = check_num_extents_and_bitmaps(cache, 1, 1); if (ret) return ret; - if (cache->free_space_ctl->free_space != 8192) { - test_msg("Cache free space is not 8Kb\n"); + if (cache->free_space_ctl->free_space != 2 * sectorsize) { + test_msg("Cache free space is not %u\n", 2 * sectorsize); return -EINVAL; } offset = btrfs_find_space_for_alloc(cache, - 0, 8192, 0, + 0, 2 * sectorsize, 0, &max_extent_size); if (offset != SZ_32M) { - test_msg("Failed to allocate 8Kb from space cache, returned offset is: %llu\n", + test_msg("Failed to allocate %u, offset: %llu\n", + 2 * sectorsize, offset); return -EINVAL; } @@ -825,7 +835,7 @@ test_steal_space_from_bitmap_to_extent(struct btrfs_block_group_cache *cache) return 0; } -int btrfs_test_free_space_cache(void) +int btrfs_test_free_space_cache(u32 sectorsize, u32 nodesize) { struct btrfs_block_group_cache *cache; struct btrfs_root *root = NULL; @@ -833,13 +843,13 @@ int btrfs_test_free_space_cache(void) test_msg("Running btrfs free space cache tests\n"); - cache = btrfs_alloc_dummy_block_group(1024 * 1024 * 1024); + cache = btrfs_alloc_dummy_block_group(1024 * 1024 * 1024, sectorsize); if (!cache) { test_msg("Couldn't run the tests\n"); return 0; } - root = btrfs_alloc_dummy_root(); + root = btrfs_alloc_dummy_root(sectorsize, nodesize); if (IS_ERR(root)) { ret = PTR_ERR(root); goto out; @@ -855,14 +865,14 @@ int btrfs_test_free_space_cache(void) ret = test_extents(cache); if (ret) goto out; - ret = test_bitmaps(cache); + ret = test_bitmaps(cache, sectorsize); if (ret) goto out; - ret = test_bitmaps_and_extents(cache); + ret = test_bitmaps_and_extents(cache, sectorsize); if (ret) goto out; - ret = test_steal_space_from_bitmap_to_extent(cache); + ret = test_steal_space_from_bitmap_to_extent(cache, sectorsize); out: btrfs_free_dummy_block_group(cache); btrfs_free_dummy_root(root); diff --git a/fs/btrfs/tests/free-space-tree-tests.c b/fs/btrfs/tests/free-space-tree-tests.c index 7cea4462acd5..aac507085ab0 100644 --- a/fs/btrfs/tests/free-space-tree-tests.c +++ b/fs/btrfs/tests/free-space-tree-tests.c @@ -16,6 +16,7 @@ * Boston, MA 021110-1307, USA. */ +#include #include "btrfs-tests.h" #include "../ctree.h" #include "../disk-io.h" @@ -30,7 +31,7 @@ struct free_space_extent { * The test cases align their operations to this in order to hit some of the * edge cases in the bitmap code. */ -#define BITMAP_RANGE (BTRFS_FREE_SPACE_BITMAP_BITS * 4096) +#define BITMAP_RANGE (BTRFS_FREE_SPACE_BITMAP_BITS * PAGE_SIZE) static int __check_free_space_extents(struct btrfs_trans_handle *trans, struct btrfs_fs_info *fs_info, @@ -439,7 +440,8 @@ typedef int (*test_func_t)(struct btrfs_trans_handle *, struct btrfs_block_group_cache *, struct btrfs_path *); -static int run_test(test_func_t test_func, int bitmaps) +static int run_test(test_func_t test_func, int bitmaps, + u32 sectorsize, u32 nodesize) { struct btrfs_root *root = NULL; struct btrfs_block_group_cache *cache = NULL; @@ -447,7 +449,7 @@ static int run_test(test_func_t test_func, int bitmaps) struct btrfs_path *path = NULL; int ret; - root = btrfs_alloc_dummy_root(); + root = btrfs_alloc_dummy_root(sectorsize, nodesize); if (IS_ERR(root)) { test_msg("Couldn't allocate dummy root\n"); ret = PTR_ERR(root); @@ -466,7 +468,8 @@ static int run_test(test_func_t test_func, int bitmaps) root->fs_info->free_space_root = root; root->fs_info->tree_root = root; - root->node = alloc_test_extent_buffer(root->fs_info, 4096); + root->node = alloc_test_extent_buffer(root->fs_info, + nodesize, nodesize); if (!root->node) { test_msg("Couldn't allocate dummy buffer\n"); ret = -ENOMEM; @@ -474,9 +477,9 @@ static int run_test(test_func_t test_func, int bitmaps) } btrfs_set_header_level(root->node, 0); btrfs_set_header_nritems(root->node, 0); - root->alloc_bytenr += 8192; + root->alloc_bytenr += 2 * nodesize; - cache = btrfs_alloc_dummy_block_group(8 * BITMAP_RANGE); + cache = btrfs_alloc_dummy_block_group(8 * BITMAP_RANGE, sectorsize); if (!cache) { test_msg("Couldn't allocate dummy block group cache\n"); ret = -ENOMEM; @@ -534,17 +537,18 @@ out: return ret; } -static int run_test_both_formats(test_func_t test_func) +static int run_test_both_formats(test_func_t test_func, + u32 sectorsize, u32 nodesize) { int ret; - ret = run_test(test_func, 0); + ret = run_test(test_func, 0, sectorsize, nodesize); if (ret) return ret; - return run_test(test_func, 1); + return run_test(test_func, 1, sectorsize, nodesize); } -int btrfs_test_free_space_tree(void) +int btrfs_test_free_space_tree(u32 sectorsize, u32 nodesize) { test_func_t tests[] = { test_empty_block_group, @@ -561,9 +565,11 @@ int btrfs_test_free_space_tree(void) test_msg("Running free space tree tests\n"); for (i = 0; i < ARRAY_SIZE(tests); i++) { - int ret = run_test_both_formats(tests[i]); + int ret = run_test_both_formats(tests[i], sectorsize, + nodesize); if (ret) { - test_msg("%pf failed\n", tests[i]); + test_msg("%pf : sectorsize %u failed\n", + tests[i], sectorsize); return ret; } } diff --git a/fs/btrfs/tests/inode-tests.c b/fs/btrfs/tests/inode-tests.c index 8a25fe8b7c45..29648c0a39f1 100644 --- a/fs/btrfs/tests/inode-tests.c +++ b/fs/btrfs/tests/inode-tests.c @@ -16,6 +16,7 @@ * Boston, MA 021110-1307, USA. */ +#include #include "btrfs-tests.h" #include "../ctree.h" #include "../btrfs_inode.h" @@ -86,19 +87,19 @@ static void insert_inode_item_key(struct btrfs_root *root) * diagram of how the extents will look though this may not be possible we still * want to make sure everything acts normally (the last number is not inclusive) * - * [0 - 5][5 - 6][6 - 10][10 - 4096][ 4096 - 8192 ][8192 - 12288] - * [hole ][inline][ hole ][ regular ][regular1 split][ hole ] + * [0 - 5][5 - 6][ 6 - 4096 ][ 4096 - 4100][4100 - 8195][8195 - 12291] + * [hole ][inline][hole but no extent][ hole ][ regular ][regular1 split] * - * [ 12288 - 20480][20480 - 24576][ 24576 - 28672 ][28672 - 36864][36864 - 45056] - * [regular1 split][ prealloc1 ][prealloc1 written][ prealloc1 ][ compressed ] + * [12291 - 16387][16387 - 24579][24579 - 28675][ 28675 - 32771][32771 - 36867 ] + * [ hole ][regular1 split][ prealloc ][ prealloc1 ][prealloc1 written] * - * [45056 - 49152][49152-53248][53248-61440][61440-65536][ 65536+81920 ] - * [ compressed1 ][ regular ][compressed1][ regular ][ hole but no extent] + * [36867 - 45059][45059 - 53251][53251 - 57347][57347 - 61443][61443- 69635] + * [ prealloc1 ][ compressed ][ compressed1 ][ regular ][ compressed1] * - * [81920-86016] - * [ regular ] + * [69635-73731][ 73731 - 86019 ][86019-90115] + * [ regular ][ hole but no extent][ regular ] */ -static void setup_file_extents(struct btrfs_root *root) +static void setup_file_extents(struct btrfs_root *root, u32 sectorsize) { int slot = 0; u64 disk_bytenr = SZ_1M; @@ -119,7 +120,7 @@ static void setup_file_extents(struct btrfs_root *root) insert_extent(root, offset, 1, 1, 0, 0, 0, BTRFS_FILE_EXTENT_INLINE, 0, slot); slot++; - offset = 4096; + offset = sectorsize; /* Now another hole */ insert_extent(root, offset, 4, 4, 0, 0, 0, BTRFS_FILE_EXTENT_REG, 0, @@ -128,99 +129,106 @@ static void setup_file_extents(struct btrfs_root *root) offset += 4; /* Now for a regular extent */ - insert_extent(root, offset, 4095, 4095, 0, disk_bytenr, 4096, - BTRFS_FILE_EXTENT_REG, 0, slot); + insert_extent(root, offset, sectorsize - 1, sectorsize - 1, 0, + disk_bytenr, sectorsize, BTRFS_FILE_EXTENT_REG, 0, slot); slot++; - disk_bytenr += 4096; - offset += 4095; + disk_bytenr += sectorsize; + offset += sectorsize - 1; /* * Now for 3 extents that were split from a hole punch so we test * offsets properly. */ - insert_extent(root, offset, 4096, 16384, 0, disk_bytenr, 16384, - BTRFS_FILE_EXTENT_REG, 0, slot); + insert_extent(root, offset, sectorsize, 4 * sectorsize, 0, disk_bytenr, + 4 * sectorsize, BTRFS_FILE_EXTENT_REG, 0, slot); slot++; - offset += 4096; - insert_extent(root, offset, 4096, 4096, 0, 0, 0, BTRFS_FILE_EXTENT_REG, - 0, slot); + offset += sectorsize; + insert_extent(root, offset, sectorsize, sectorsize, 0, 0, 0, + BTRFS_FILE_EXTENT_REG, 0, slot); slot++; - offset += 4096; - insert_extent(root, offset, 8192, 16384, 8192, disk_bytenr, 16384, + offset += sectorsize; + insert_extent(root, offset, 2 * sectorsize, 4 * sectorsize, + 2 * sectorsize, disk_bytenr, 4 * sectorsize, BTRFS_FILE_EXTENT_REG, 0, slot); slot++; - offset += 8192; - disk_bytenr += 16384; + offset += 2 * sectorsize; + disk_bytenr += 4 * sectorsize; /* Now for a unwritten prealloc extent */ - insert_extent(root, offset, 4096, 4096, 0, disk_bytenr, 4096, - BTRFS_FILE_EXTENT_PREALLOC, 0, slot); + insert_extent(root, offset, sectorsize, sectorsize, 0, disk_bytenr, + sectorsize, BTRFS_FILE_EXTENT_PREALLOC, 0, slot); slot++; - offset += 4096; + offset += sectorsize; /* * We want to jack up disk_bytenr a little more so the em stuff doesn't * merge our records. */ - disk_bytenr += 8192; + disk_bytenr += 2 * sectorsize; /* * Now for a partially written prealloc extent, basically the same as * the hole punch example above. Ram_bytes never changes when you mark * extents written btw. */ - insert_extent(root, offset, 4096, 16384, 0, disk_bytenr, 16384, - BTRFS_FILE_EXTENT_PREALLOC, 0, slot); + insert_extent(root, offset, sectorsize, 4 * sectorsize, 0, disk_bytenr, + 4 * sectorsize, BTRFS_FILE_EXTENT_PREALLOC, 0, slot); slot++; - offset += 4096; - insert_extent(root, offset, 4096, 16384, 4096, disk_bytenr, 16384, - BTRFS_FILE_EXTENT_REG, 0, slot); + offset += sectorsize; + insert_extent(root, offset, sectorsize, 4 * sectorsize, sectorsize, + disk_bytenr, 4 * sectorsize, BTRFS_FILE_EXTENT_REG, 0, + slot); slot++; - offset += 4096; - insert_extent(root, offset, 8192, 16384, 8192, disk_bytenr, 16384, + offset += sectorsize; + insert_extent(root, offset, 2 * sectorsize, 4 * sectorsize, + 2 * sectorsize, disk_bytenr, 4 * sectorsize, BTRFS_FILE_EXTENT_PREALLOC, 0, slot); slot++; - offset += 8192; - disk_bytenr += 16384; + offset += 2 * sectorsize; + disk_bytenr += 4 * sectorsize; /* Now a normal compressed extent */ - insert_extent(root, offset, 8192, 8192, 0, disk_bytenr, 4096, - BTRFS_FILE_EXTENT_REG, BTRFS_COMPRESS_ZLIB, slot); + insert_extent(root, offset, 2 * sectorsize, 2 * sectorsize, 0, + disk_bytenr, sectorsize, BTRFS_FILE_EXTENT_REG, + BTRFS_COMPRESS_ZLIB, slot); slot++; - offset += 8192; + offset += 2 * sectorsize; /* No merges */ - disk_bytenr += 8192; + disk_bytenr += 2 * sectorsize; /* Now a split compressed extent */ - insert_extent(root, offset, 4096, 16384, 0, disk_bytenr, 4096, - BTRFS_FILE_EXTENT_REG, BTRFS_COMPRESS_ZLIB, slot); + insert_extent(root, offset, sectorsize, 4 * sectorsize, 0, disk_bytenr, + sectorsize, BTRFS_FILE_EXTENT_REG, + BTRFS_COMPRESS_ZLIB, slot); slot++; - offset += 4096; - insert_extent(root, offset, 4096, 4096, 0, disk_bytenr + 4096, 4096, + offset += sectorsize; + insert_extent(root, offset, sectorsize, sectorsize, 0, + disk_bytenr + sectorsize, sectorsize, BTRFS_FILE_EXTENT_REG, 0, slot); slot++; - offset += 4096; - insert_extent(root, offset, 8192, 16384, 8192, disk_bytenr, 4096, + offset += sectorsize; + insert_extent(root, offset, 2 * sectorsize, 4 * sectorsize, + 2 * sectorsize, disk_bytenr, sectorsize, BTRFS_FILE_EXTENT_REG, BTRFS_COMPRESS_ZLIB, slot); slot++; - offset += 8192; - disk_bytenr += 8192; + offset += 2 * sectorsize; + disk_bytenr += 2 * sectorsize; /* Now extents that have a hole but no hole extent */ - insert_extent(root, offset, 4096, 4096, 0, disk_bytenr, 4096, - BTRFS_FILE_EXTENT_REG, 0, slot); + insert_extent(root, offset, sectorsize, sectorsize, 0, disk_bytenr, + sectorsize, BTRFS_FILE_EXTENT_REG, 0, slot); slot++; - offset += 16384; - disk_bytenr += 4096; - insert_extent(root, offset, 4096, 4096, 0, disk_bytenr, 4096, - BTRFS_FILE_EXTENT_REG, 0, slot); + offset += 4 * sectorsize; + disk_bytenr += sectorsize; + insert_extent(root, offset, sectorsize, sectorsize, 0, disk_bytenr, + sectorsize, BTRFS_FILE_EXTENT_REG, 0, slot); } static unsigned long prealloc_only = 0; static unsigned long compressed_only = 0; static unsigned long vacancy_only = 0; -static noinline int test_btrfs_get_extent(void) +static noinline int test_btrfs_get_extent(u32 sectorsize, u32 nodesize) { struct inode *inode = NULL; struct btrfs_root *root = NULL; @@ -240,7 +248,7 @@ static noinline int test_btrfs_get_extent(void) BTRFS_I(inode)->location.objectid = BTRFS_FIRST_FREE_OBJECTID; BTRFS_I(inode)->location.offset = 0; - root = btrfs_alloc_dummy_root(); + root = btrfs_alloc_dummy_root(sectorsize, nodesize); if (IS_ERR(root)) { test_msg("Couldn't allocate root\n"); goto out; @@ -256,7 +264,7 @@ static noinline int test_btrfs_get_extent(void) goto out; } - root->node = alloc_dummy_extent_buffer(NULL, 4096); + root->node = alloc_dummy_extent_buffer(NULL, nodesize, nodesize); if (!root->node) { test_msg("Couldn't allocate dummy buffer\n"); goto out; @@ -273,7 +281,7 @@ static noinline int test_btrfs_get_extent(void) /* First with no extents */ BTRFS_I(inode)->root = root; - em = btrfs_get_extent(inode, NULL, 0, 0, 4096, 0); + em = btrfs_get_extent(inode, NULL, 0, 0, sectorsize, 0); if (IS_ERR(em)) { em = NULL; test_msg("Got an error when we shouldn't have\n"); @@ -295,7 +303,7 @@ static noinline int test_btrfs_get_extent(void) * setup_file_extents, so if you change anything there you need to * update the comment and update the expected values below. */ - setup_file_extents(root); + setup_file_extents(root, sectorsize); em = btrfs_get_extent(inode, NULL, 0, 0, (u64)-1, 0); if (IS_ERR(em)) { @@ -318,7 +326,7 @@ static noinline int test_btrfs_get_extent(void) offset = em->start + em->len; free_extent_map(em); - em = btrfs_get_extent(inode, NULL, 0, offset, 4096, 0); + em = btrfs_get_extent(inode, NULL, 0, offset, sectorsize, 0); if (IS_ERR(em)) { test_msg("Got an error when we shouldn't have\n"); goto out; @@ -327,7 +335,8 @@ static noinline int test_btrfs_get_extent(void) test_msg("Expected an inline, got %llu\n", em->block_start); goto out; } - if (em->start != offset || em->len != 4091) { + + if (em->start != offset || em->len != (sectorsize - 5)) { test_msg("Unexpected extent wanted start %llu len 1, got start " "%llu len %llu\n", offset, em->start, em->len); goto out; @@ -344,7 +353,7 @@ static noinline int test_btrfs_get_extent(void) offset = em->start + em->len; free_extent_map(em); - em = btrfs_get_extent(inode, NULL, 0, offset, 4096, 0); + em = btrfs_get_extent(inode, NULL, 0, offset, sectorsize, 0); if (IS_ERR(em)) { test_msg("Got an error when we shouldn't have\n"); goto out; @@ -366,7 +375,7 @@ static noinline int test_btrfs_get_extent(void) free_extent_map(em); /* Regular extent */ - em = btrfs_get_extent(inode, NULL, 0, offset, 4096, 0); + em = btrfs_get_extent(inode, NULL, 0, offset, sectorsize, 0); if (IS_ERR(em)) { test_msg("Got an error when we shouldn't have\n"); goto out; @@ -375,7 +384,7 @@ static noinline int test_btrfs_get_extent(void) test_msg("Expected a real extent, got %llu\n", em->block_start); goto out; } - if (em->start != offset || em->len != 4095) { + if (em->start != offset || em->len != sectorsize - 1) { test_msg("Unexpected extent wanted start %llu len 4095, got " "start %llu len %llu\n", offset, em->start, em->len); goto out; @@ -393,7 +402,7 @@ static noinline int test_btrfs_get_extent(void) free_extent_map(em); /* The next 3 are split extents */ - em = btrfs_get_extent(inode, NULL, 0, offset, 4096, 0); + em = btrfs_get_extent(inode, NULL, 0, offset, sectorsize, 0); if (IS_ERR(em)) { test_msg("Got an error when we shouldn't have\n"); goto out; @@ -402,9 +411,10 @@ static noinline int test_btrfs_get_extent(void) test_msg("Expected a real extent, got %llu\n", em->block_start); goto out; } - if (em->start != offset || em->len != 4096) { - test_msg("Unexpected extent wanted start %llu len 4096, got " - "start %llu len %llu\n", offset, em->start, em->len); + if (em->start != offset || em->len != sectorsize) { + test_msg("Unexpected extent start %llu len %u, " + "got start %llu len %llu\n", + offset, sectorsize, em->start, em->len); goto out; } if (em->flags != 0) { @@ -421,7 +431,7 @@ static noinline int test_btrfs_get_extent(void) offset = em->start + em->len; free_extent_map(em); - em = btrfs_get_extent(inode, NULL, 0, offset, 4096, 0); + em = btrfs_get_extent(inode, NULL, 0, offset, sectorsize, 0); if (IS_ERR(em)) { test_msg("Got an error when we shouldn't have\n"); goto out; @@ -430,9 +440,10 @@ static noinline int test_btrfs_get_extent(void) test_msg("Expected a hole, got %llu\n", em->block_start); goto out; } - if (em->start != offset || em->len != 4096) { - test_msg("Unexpected extent wanted start %llu len 4096, got " - "start %llu len %llu\n", offset, em->start, em->len); + if (em->start != offset || em->len != sectorsize) { + test_msg("Unexpected extent wanted start %llu len %u, " + "got start %llu len %llu\n", + offset, sectorsize, em->start, em->len); goto out; } if (em->flags != 0) { @@ -442,7 +453,7 @@ static noinline int test_btrfs_get_extent(void) offset = em->start + em->len; free_extent_map(em); - em = btrfs_get_extent(inode, NULL, 0, offset, 4096, 0); + em = btrfs_get_extent(inode, NULL, 0, offset, sectorsize, 0); if (IS_ERR(em)) { test_msg("Got an error when we shouldn't have\n"); goto out; @@ -451,9 +462,10 @@ static noinline int test_btrfs_get_extent(void) test_msg("Expected a real extent, got %llu\n", em->block_start); goto out; } - if (em->start != offset || em->len != 8192) { - test_msg("Unexpected extent wanted start %llu len 8192, got " - "start %llu len %llu\n", offset, em->start, em->len); + if (em->start != offset || em->len != 2 * sectorsize) { + test_msg("Unexpected extent wanted start %llu len %u, " + "got start %llu len %llu\n", + offset, 2 * sectorsize, em->start, em->len); goto out; } if (em->flags != 0) { @@ -475,7 +487,7 @@ static noinline int test_btrfs_get_extent(void) free_extent_map(em); /* Prealloc extent */ - em = btrfs_get_extent(inode, NULL, 0, offset, 4096, 0); + em = btrfs_get_extent(inode, NULL, 0, offset, sectorsize, 0); if (IS_ERR(em)) { test_msg("Got an error when we shouldn't have\n"); goto out; @@ -484,9 +496,10 @@ static noinline int test_btrfs_get_extent(void) test_msg("Expected a real extent, got %llu\n", em->block_start); goto out; } - if (em->start != offset || em->len != 4096) { - test_msg("Unexpected extent wanted start %llu len 4096, got " - "start %llu len %llu\n", offset, em->start, em->len); + if (em->start != offset || em->len != sectorsize) { + test_msg("Unexpected extent wanted start %llu len %u, " + "got start %llu len %llu\n", + offset, sectorsize, em->start, em->len); goto out; } if (em->flags != prealloc_only) { @@ -503,7 +516,7 @@ static noinline int test_btrfs_get_extent(void) free_extent_map(em); /* The next 3 are a half written prealloc extent */ - em = btrfs_get_extent(inode, NULL, 0, offset, 4096, 0); + em = btrfs_get_extent(inode, NULL, 0, offset, sectorsize, 0); if (IS_ERR(em)) { test_msg("Got an error when we shouldn't have\n"); goto out; @@ -512,9 +525,10 @@ static noinline int test_btrfs_get_extent(void) test_msg("Expected a real extent, got %llu\n", em->block_start); goto out; } - if (em->start != offset || em->len != 4096) { - test_msg("Unexpected extent wanted start %llu len 4096, got " - "start %llu len %llu\n", offset, em->start, em->len); + if (em->start != offset || em->len != sectorsize) { + test_msg("Unexpected extent wanted start %llu len %u, " + "got start %llu len %llu\n", + offset, sectorsize, em->start, em->len); goto out; } if (em->flags != prealloc_only) { @@ -532,7 +546,7 @@ static noinline int test_btrfs_get_extent(void) offset = em->start + em->len; free_extent_map(em); - em = btrfs_get_extent(inode, NULL, 0, offset, 4096, 0); + em = btrfs_get_extent(inode, NULL, 0, offset, sectorsize, 0); if (IS_ERR(em)) { test_msg("Got an error when we shouldn't have\n"); goto out; @@ -541,9 +555,10 @@ static noinline int test_btrfs_get_extent(void) test_msg("Expected a real extent, got %llu\n", em->block_start); goto out; } - if (em->start != offset || em->len != 4096) { - test_msg("Unexpected extent wanted start %llu len 4096, got " - "start %llu len %llu\n", offset, em->start, em->len); + if (em->start != offset || em->len != sectorsize) { + test_msg("Unexpected extent wanted start %llu len %u, " + "got start %llu len %llu\n", + offset, sectorsize, em->start, em->len); goto out; } if (em->flags != 0) { @@ -564,7 +579,7 @@ static noinline int test_btrfs_get_extent(void) offset = em->start + em->len; free_extent_map(em); - em = btrfs_get_extent(inode, NULL, 0, offset, 4096, 0); + em = btrfs_get_extent(inode, NULL, 0, offset, sectorsize, 0); if (IS_ERR(em)) { test_msg("Got an error when we shouldn't have\n"); goto out; @@ -573,9 +588,10 @@ static noinline int test_btrfs_get_extent(void) test_msg("Expected a real extent, got %llu\n", em->block_start); goto out; } - if (em->start != offset || em->len != 8192) { - test_msg("Unexpected extent wanted start %llu len 8192, got " - "start %llu len %llu\n", offset, em->start, em->len); + if (em->start != offset || em->len != 2 * sectorsize) { + test_msg("Unexpected extent wanted start %llu len %u, " + "got start %llu len %llu\n", + offset, 2 * sectorsize, em->start, em->len); goto out; } if (em->flags != prealloc_only) { @@ -598,7 +614,7 @@ static noinline int test_btrfs_get_extent(void) free_extent_map(em); /* Now for the compressed extent */ - em = btrfs_get_extent(inode, NULL, 0, offset, 4096, 0); + em = btrfs_get_extent(inode, NULL, 0, offset, sectorsize, 0); if (IS_ERR(em)) { test_msg("Got an error when we shouldn't have\n"); goto out; @@ -607,9 +623,10 @@ static noinline int test_btrfs_get_extent(void) test_msg("Expected a real extent, got %llu\n", em->block_start); goto out; } - if (em->start != offset || em->len != 8192) { - test_msg("Unexpected extent wanted start %llu len 8192, got " - "start %llu len %llu\n", offset, em->start, em->len); + if (em->start != offset || em->len != 2 * sectorsize) { + test_msg("Unexpected extent wanted start %llu len %u," + "got start %llu len %llu\n", + offset, 2 * sectorsize, em->start, em->len); goto out; } if (em->flags != compressed_only) { @@ -631,7 +648,7 @@ static noinline int test_btrfs_get_extent(void) free_extent_map(em); /* Split compressed extent */ - em = btrfs_get_extent(inode, NULL, 0, offset, 4096, 0); + em = btrfs_get_extent(inode, NULL, 0, offset, sectorsize, 0); if (IS_ERR(em)) { test_msg("Got an error when we shouldn't have\n"); goto out; @@ -640,9 +657,10 @@ static noinline int test_btrfs_get_extent(void) test_msg("Expected a real extent, got %llu\n", em->block_start); goto out; } - if (em->start != offset || em->len != 4096) { - test_msg("Unexpected extent wanted start %llu len 4096, got " - "start %llu len %llu\n", offset, em->start, em->len); + if (em->start != offset || em->len != sectorsize) { + test_msg("Unexpected extent wanted start %llu len %u," + "got start %llu len %llu\n", + offset, sectorsize, em->start, em->len); goto out; } if (em->flags != compressed_only) { @@ -665,7 +683,7 @@ static noinline int test_btrfs_get_extent(void) offset = em->start + em->len; free_extent_map(em); - em = btrfs_get_extent(inode, NULL, 0, offset, 4096, 0); + em = btrfs_get_extent(inode, NULL, 0, offset, sectorsize, 0); if (IS_ERR(em)) { test_msg("Got an error when we shouldn't have\n"); goto out; @@ -674,9 +692,10 @@ static noinline int test_btrfs_get_extent(void) test_msg("Expected a real extent, got %llu\n", em->block_start); goto out; } - if (em->start != offset || em->len != 4096) { - test_msg("Unexpected extent wanted start %llu len 4096, got " - "start %llu len %llu\n", offset, em->start, em->len); + if (em->start != offset || em->len != sectorsize) { + test_msg("Unexpected extent wanted start %llu len %u, " + "got start %llu len %llu\n", + offset, sectorsize, em->start, em->len); goto out; } if (em->flags != 0) { @@ -691,7 +710,7 @@ static noinline int test_btrfs_get_extent(void) offset = em->start + em->len; free_extent_map(em); - em = btrfs_get_extent(inode, NULL, 0, offset, 4096, 0); + em = btrfs_get_extent(inode, NULL, 0, offset, sectorsize, 0); if (IS_ERR(em)) { test_msg("Got an error when we shouldn't have\n"); goto out; @@ -701,9 +720,10 @@ static noinline int test_btrfs_get_extent(void) disk_bytenr, em->block_start); goto out; } - if (em->start != offset || em->len != 8192) { - test_msg("Unexpected extent wanted start %llu len 8192, got " - "start %llu len %llu\n", offset, em->start, em->len); + if (em->start != offset || em->len != 2 * sectorsize) { + test_msg("Unexpected extent wanted start %llu len %u, " + "got start %llu len %llu\n", + offset, 2 * sectorsize, em->start, em->len); goto out; } if (em->flags != compressed_only) { @@ -725,7 +745,7 @@ static noinline int test_btrfs_get_extent(void) free_extent_map(em); /* A hole between regular extents but no hole extent */ - em = btrfs_get_extent(inode, NULL, 0, offset + 6, 4096, 0); + em = btrfs_get_extent(inode, NULL, 0, offset + 6, sectorsize, 0); if (IS_ERR(em)) { test_msg("Got an error when we shouldn't have\n"); goto out; @@ -734,9 +754,10 @@ static noinline int test_btrfs_get_extent(void) test_msg("Expected a real extent, got %llu\n", em->block_start); goto out; } - if (em->start != offset || em->len != 4096) { - test_msg("Unexpected extent wanted start %llu len 4096, got " - "start %llu len %llu\n", offset, em->start, em->len); + if (em->start != offset || em->len != sectorsize) { + test_msg("Unexpected extent wanted start %llu len %u, " + "got start %llu len %llu\n", + offset, sectorsize, em->start, em->len); goto out; } if (em->flags != 0) { @@ -765,9 +786,10 @@ static noinline int test_btrfs_get_extent(void) * length of the actual hole, if this changes we'll have to change this * test. */ - if (em->start != offset || em->len != 12288) { - test_msg("Unexpected extent wanted start %llu len 12288, got " - "start %llu len %llu\n", offset, em->start, em->len); + if (em->start != offset || em->len != 3 * sectorsize) { + test_msg("Unexpected extent wanted start %llu len %u, " + "got start %llu len %llu\n", + offset, 3 * sectorsize, em->start, em->len); goto out; } if (em->flags != vacancy_only) { @@ -783,7 +805,7 @@ static noinline int test_btrfs_get_extent(void) offset = em->start + em->len; free_extent_map(em); - em = btrfs_get_extent(inode, NULL, 0, offset, 4096, 0); + em = btrfs_get_extent(inode, NULL, 0, offset, sectorsize, 0); if (IS_ERR(em)) { test_msg("Got an error when we shouldn't have\n"); goto out; @@ -792,9 +814,10 @@ static noinline int test_btrfs_get_extent(void) test_msg("Expected a real extent, got %llu\n", em->block_start); goto out; } - if (em->start != offset || em->len != 4096) { - test_msg("Unexpected extent wanted start %llu len 4096, got " - "start %llu len %llu\n", offset, em->start, em->len); + if (em->start != offset || em->len != sectorsize) { + test_msg("Unexpected extent wanted start %llu len %u," + "got start %llu len %llu\n", + offset, sectorsize, em->start, em->len); goto out; } if (em->flags != 0) { @@ -815,7 +838,7 @@ out: return ret; } -static int test_hole_first(void) +static int test_hole_first(u32 sectorsize, u32 nodesize) { struct inode *inode = NULL; struct btrfs_root *root = NULL; @@ -832,7 +855,7 @@ static int test_hole_first(void) BTRFS_I(inode)->location.objectid = BTRFS_FIRST_FREE_OBJECTID; BTRFS_I(inode)->location.offset = 0; - root = btrfs_alloc_dummy_root(); + root = btrfs_alloc_dummy_root(sectorsize, nodesize); if (IS_ERR(root)) { test_msg("Couldn't allocate root\n"); goto out; @@ -844,7 +867,7 @@ static int test_hole_first(void) goto out; } - root->node = alloc_dummy_extent_buffer(NULL, 4096); + root->node = alloc_dummy_extent_buffer(NULL, nodesize, nodesize); if (!root->node) { test_msg("Couldn't allocate dummy buffer\n"); goto out; @@ -861,9 +884,9 @@ static int test_hole_first(void) * btrfs_get_extent. */ insert_inode_item_key(root); - insert_extent(root, 4096, 4096, 4096, 0, 4096, 4096, - BTRFS_FILE_EXTENT_REG, 0, 1); - em = btrfs_get_extent(inode, NULL, 0, 0, 8192, 0); + insert_extent(root, sectorsize, sectorsize, sectorsize, 0, sectorsize, + sectorsize, BTRFS_FILE_EXTENT_REG, 0, 1); + em = btrfs_get_extent(inode, NULL, 0, 0, 2 * sectorsize, 0); if (IS_ERR(em)) { test_msg("Got an error when we shouldn't have\n"); goto out; @@ -872,9 +895,10 @@ static int test_hole_first(void) test_msg("Expected a hole, got %llu\n", em->block_start); goto out; } - if (em->start != 0 || em->len != 4096) { - test_msg("Unexpected extent wanted start 0 len 4096, got start " - "%llu len %llu\n", em->start, em->len); + if (em->start != 0 || em->len != sectorsize) { + test_msg("Unexpected extent wanted start 0 len %u, " + "got start %llu len %llu\n", + sectorsize, em->start, em->len); goto out; } if (em->flags != vacancy_only) { @@ -884,18 +908,19 @@ static int test_hole_first(void) } free_extent_map(em); - em = btrfs_get_extent(inode, NULL, 0, 4096, 8192, 0); + em = btrfs_get_extent(inode, NULL, 0, sectorsize, 2 * sectorsize, 0); if (IS_ERR(em)) { test_msg("Got an error when we shouldn't have\n"); goto out; } - if (em->block_start != 4096) { + if (em->block_start != sectorsize) { test_msg("Expected a real extent, got %llu\n", em->block_start); goto out; } - if (em->start != 4096 || em->len != 4096) { - test_msg("Unexpected extent wanted start 4096 len 4096, got " - "start %llu len %llu\n", em->start, em->len); + if (em->start != sectorsize || em->len != sectorsize) { + test_msg("Unexpected extent wanted start %u len %u, " + "got start %llu len %llu\n", + sectorsize, sectorsize, em->start, em->len); goto out; } if (em->flags != 0) { @@ -912,7 +937,7 @@ out: return ret; } -static int test_extent_accounting(void) +static int test_extent_accounting(u32 sectorsize, u32 nodesize) { struct inode *inode = NULL; struct btrfs_root *root = NULL; @@ -924,7 +949,7 @@ static int test_extent_accounting(void) return ret; } - root = btrfs_alloc_dummy_root(); + root = btrfs_alloc_dummy_root(sectorsize, nodesize); if (IS_ERR(root)) { test_msg("Couldn't allocate root\n"); goto out; @@ -954,10 +979,11 @@ static int test_extent_accounting(void) goto out; } - /* [BTRFS_MAX_EXTENT_SIZE][4k] */ + /* [BTRFS_MAX_EXTENT_SIZE][sectorsize] */ BTRFS_I(inode)->outstanding_extents++; ret = btrfs_set_extent_delalloc(inode, BTRFS_MAX_EXTENT_SIZE, - BTRFS_MAX_EXTENT_SIZE + 4095, NULL); + BTRFS_MAX_EXTENT_SIZE + sectorsize - 1, + NULL); if (ret) { test_msg("btrfs_set_extent_delalloc returned %d\n", ret); goto out; @@ -969,10 +995,10 @@ static int test_extent_accounting(void) goto out; } - /* [BTRFS_MAX_EXTENT_SIZE/2][4K HOLE][the rest] */ + /* [BTRFS_MAX_EXTENT_SIZE/2][sectorsize HOLE][the rest] */ ret = clear_extent_bit(&BTRFS_I(inode)->io_tree, BTRFS_MAX_EXTENT_SIZE >> 1, - (BTRFS_MAX_EXTENT_SIZE >> 1) + 4095, + (BTRFS_MAX_EXTENT_SIZE >> 1) + sectorsize - 1, EXTENT_DELALLOC | EXTENT_DIRTY | EXTENT_UPTODATE | EXTENT_DO_ACCOUNTING, 0, 0, NULL, GFP_KERNEL); @@ -987,10 +1013,11 @@ static int test_extent_accounting(void) goto out; } - /* [BTRFS_MAX_EXTENT_SIZE][4K] */ + /* [BTRFS_MAX_EXTENT_SIZE][sectorsize] */ BTRFS_I(inode)->outstanding_extents++; ret = btrfs_set_extent_delalloc(inode, BTRFS_MAX_EXTENT_SIZE >> 1, - (BTRFS_MAX_EXTENT_SIZE >> 1) + 4095, + (BTRFS_MAX_EXTENT_SIZE >> 1) + + sectorsize - 1, NULL); if (ret) { test_msg("btrfs_set_extent_delalloc returned %d\n", ret); @@ -1004,16 +1031,17 @@ static int test_extent_accounting(void) } /* - * [BTRFS_MAX_EXTENT_SIZE+4K][4K HOLE][BTRFS_MAX_EXTENT_SIZE+4K] + * [BTRFS_MAX_EXTENT_SIZE+sectorsize][sectorsize HOLE][BTRFS_MAX_EXTENT_SIZE+sectorsize] * * I'm artificially adding 2 to outstanding_extents because in the * buffered IO case we'd add things up as we go, but I don't feel like * doing that here, this isn't the interesting case we want to test. */ BTRFS_I(inode)->outstanding_extents += 2; - ret = btrfs_set_extent_delalloc(inode, BTRFS_MAX_EXTENT_SIZE + 8192, - (BTRFS_MAX_EXTENT_SIZE << 1) + 12287, - NULL); + ret = btrfs_set_extent_delalloc(inode, + BTRFS_MAX_EXTENT_SIZE + 2 * sectorsize, + (BTRFS_MAX_EXTENT_SIZE << 1) + 3 * sectorsize - 1, + NULL); if (ret) { test_msg("btrfs_set_extent_delalloc returned %d\n", ret); goto out; @@ -1025,10 +1053,13 @@ static int test_extent_accounting(void) goto out; } - /* [BTRFS_MAX_EXTENT_SIZE+4k][4k][BTRFS_MAX_EXTENT_SIZE+4k] */ + /* + * [BTRFS_MAX_EXTENT_SIZE+sectorsize][sectorsize][BTRFS_MAX_EXTENT_SIZE+sectorsize] + */ BTRFS_I(inode)->outstanding_extents++; - ret = btrfs_set_extent_delalloc(inode, BTRFS_MAX_EXTENT_SIZE+4096, - BTRFS_MAX_EXTENT_SIZE+8191, NULL); + ret = btrfs_set_extent_delalloc(inode, + BTRFS_MAX_EXTENT_SIZE + sectorsize, + BTRFS_MAX_EXTENT_SIZE + 2 * sectorsize - 1, NULL); if (ret) { test_msg("btrfs_set_extent_delalloc returned %d\n", ret); goto out; @@ -1042,8 +1073,8 @@ static int test_extent_accounting(void) /* [BTRFS_MAX_EXTENT_SIZE+4k][4K HOLE][BTRFS_MAX_EXTENT_SIZE+4k] */ ret = clear_extent_bit(&BTRFS_I(inode)->io_tree, - BTRFS_MAX_EXTENT_SIZE+4096, - BTRFS_MAX_EXTENT_SIZE+8191, + BTRFS_MAX_EXTENT_SIZE + sectorsize, + BTRFS_MAX_EXTENT_SIZE + 2 * sectorsize - 1, EXTENT_DIRTY | EXTENT_DELALLOC | EXTENT_DO_ACCOUNTING | EXTENT_UPTODATE, 0, 0, NULL, GFP_KERNEL); @@ -1063,8 +1094,9 @@ static int test_extent_accounting(void) * might fail and I'd rather satisfy my paranoia at this point. */ BTRFS_I(inode)->outstanding_extents++; - ret = btrfs_set_extent_delalloc(inode, BTRFS_MAX_EXTENT_SIZE+4096, - BTRFS_MAX_EXTENT_SIZE+8191, NULL); + ret = btrfs_set_extent_delalloc(inode, + BTRFS_MAX_EXTENT_SIZE + sectorsize, + BTRFS_MAX_EXTENT_SIZE + 2 * sectorsize - 1, NULL); if (ret) { test_msg("btrfs_set_extent_delalloc returned %d\n", ret); goto out; @@ -1103,7 +1135,7 @@ out: return ret; } -int btrfs_test_inodes(void) +int btrfs_test_inodes(u32 sectorsize, u32 nodesize) { int ret; @@ -1112,13 +1144,13 @@ int btrfs_test_inodes(void) set_bit(EXTENT_FLAG_PREALLOC, &prealloc_only); test_msg("Running btrfs_get_extent tests\n"); - ret = test_btrfs_get_extent(); + ret = test_btrfs_get_extent(sectorsize, nodesize); if (ret) return ret; test_msg("Running hole first btrfs_get_extent test\n"); - ret = test_hole_first(); + ret = test_hole_first(sectorsize, nodesize); if (ret) return ret; test_msg("Running outstanding_extents tests\n"); - return test_extent_accounting(); + return test_extent_accounting(sectorsize, nodesize); } diff --git a/fs/btrfs/tests/qgroup-tests.c b/fs/btrfs/tests/qgroup-tests.c index 8aa4ded31326..9c1d27352f19 100644 --- a/fs/btrfs/tests/qgroup-tests.c +++ b/fs/btrfs/tests/qgroup-tests.c @@ -16,6 +16,7 @@ * Boston, MA 021110-1307, USA. */ +#include #include "btrfs-tests.h" #include "../ctree.h" #include "../transaction.h" @@ -216,7 +217,8 @@ static int remove_extent_ref(struct btrfs_root *root, u64 bytenr, return ret; } -static int test_no_shared_qgroup(struct btrfs_root *root) +static int test_no_shared_qgroup(struct btrfs_root *root, + u32 sectorsize, u32 nodesize) { struct btrfs_trans_handle trans; struct btrfs_fs_info *fs_info = root->fs_info; @@ -238,18 +240,18 @@ static int test_no_shared_qgroup(struct btrfs_root *root) * we can only call btrfs_qgroup_account_extent() directly to test * quota. */ - ret = btrfs_find_all_roots(&trans, fs_info, 4096, 0, &old_roots); + ret = btrfs_find_all_roots(&trans, fs_info, nodesize, 0, &old_roots); if (ret) { ulist_free(old_roots); test_msg("Couldn't find old roots: %d\n", ret); return ret; } - ret = insert_normal_tree_ref(root, 4096, 4096, 0, 5); + ret = insert_normal_tree_ref(root, nodesize, nodesize, 0, 5); if (ret) return ret; - ret = btrfs_find_all_roots(&trans, fs_info, 4096, 0, &new_roots); + ret = btrfs_find_all_roots(&trans, fs_info, nodesize, 0, &new_roots); if (ret) { ulist_free(old_roots); ulist_free(new_roots); @@ -257,32 +259,32 @@ static int test_no_shared_qgroup(struct btrfs_root *root) return ret; } - ret = btrfs_qgroup_account_extent(&trans, fs_info, 4096, 4096, - old_roots, new_roots); + ret = btrfs_qgroup_account_extent(&trans, fs_info, nodesize, + nodesize, old_roots, new_roots); if (ret) { test_msg("Couldn't account space for a qgroup %d\n", ret); return ret; } - if (btrfs_verify_qgroup_counts(fs_info, 5, 4096, 4096)) { + if (btrfs_verify_qgroup_counts(fs_info, 5, nodesize, nodesize)) { test_msg("Qgroup counts didn't match expected values\n"); return -EINVAL; } old_roots = NULL; new_roots = NULL; - ret = btrfs_find_all_roots(&trans, fs_info, 4096, 0, &old_roots); + ret = btrfs_find_all_roots(&trans, fs_info, nodesize, 0, &old_roots); if (ret) { ulist_free(old_roots); test_msg("Couldn't find old roots: %d\n", ret); return ret; } - ret = remove_extent_item(root, 4096, 4096); + ret = remove_extent_item(root, nodesize, nodesize); if (ret) return -EINVAL; - ret = btrfs_find_all_roots(&trans, fs_info, 4096, 0, &new_roots); + ret = btrfs_find_all_roots(&trans, fs_info, nodesize, 0, &new_roots); if (ret) { ulist_free(old_roots); ulist_free(new_roots); @@ -290,8 +292,8 @@ static int test_no_shared_qgroup(struct btrfs_root *root) return ret; } - ret = btrfs_qgroup_account_extent(&trans, fs_info, 4096, 4096, - old_roots, new_roots); + ret = btrfs_qgroup_account_extent(&trans, fs_info, nodesize, + nodesize, old_roots, new_roots); if (ret) { test_msg("Couldn't account space for a qgroup %d\n", ret); return -EINVAL; @@ -310,7 +312,8 @@ static int test_no_shared_qgroup(struct btrfs_root *root) * right, also remove one of the roots and make sure the exclusive count is * adjusted properly. */ -static int test_multiple_refs(struct btrfs_root *root) +static int test_multiple_refs(struct btrfs_root *root, + u32 sectorsize, u32 nodesize) { struct btrfs_trans_handle trans; struct btrfs_fs_info *fs_info = root->fs_info; @@ -329,18 +332,18 @@ static int test_multiple_refs(struct btrfs_root *root) return ret; } - ret = btrfs_find_all_roots(&trans, fs_info, 4096, 0, &old_roots); + ret = btrfs_find_all_roots(&trans, fs_info, nodesize, 0, &old_roots); if (ret) { ulist_free(old_roots); test_msg("Couldn't find old roots: %d\n", ret); return ret; } - ret = insert_normal_tree_ref(root, 4096, 4096, 0, 5); + ret = insert_normal_tree_ref(root, nodesize, nodesize, 0, 5); if (ret) return ret; - ret = btrfs_find_all_roots(&trans, fs_info, 4096, 0, &new_roots); + ret = btrfs_find_all_roots(&trans, fs_info, nodesize, 0, &new_roots); if (ret) { ulist_free(old_roots); ulist_free(new_roots); @@ -348,30 +351,31 @@ static int test_multiple_refs(struct btrfs_root *root) return ret; } - ret = btrfs_qgroup_account_extent(&trans, fs_info, 4096, 4096, - old_roots, new_roots); + ret = btrfs_qgroup_account_extent(&trans, fs_info, nodesize, + nodesize, old_roots, new_roots); if (ret) { test_msg("Couldn't account space for a qgroup %d\n", ret); return ret; } - if (btrfs_verify_qgroup_counts(fs_info, 5, 4096, 4096)) { + if (btrfs_verify_qgroup_counts(fs_info, 5, + nodesize, nodesize)) { test_msg("Qgroup counts didn't match expected values\n"); return -EINVAL; } - ret = btrfs_find_all_roots(&trans, fs_info, 4096, 0, &old_roots); + ret = btrfs_find_all_roots(&trans, fs_info, nodesize, 0, &old_roots); if (ret) { ulist_free(old_roots); test_msg("Couldn't find old roots: %d\n", ret); return ret; } - ret = add_tree_ref(root, 4096, 4096, 0, 256); + ret = add_tree_ref(root, nodesize, nodesize, 0, 256); if (ret) return ret; - ret = btrfs_find_all_roots(&trans, fs_info, 4096, 0, &new_roots); + ret = btrfs_find_all_roots(&trans, fs_info, nodesize, 0, &new_roots); if (ret) { ulist_free(old_roots); ulist_free(new_roots); @@ -379,35 +383,35 @@ static int test_multiple_refs(struct btrfs_root *root) return ret; } - ret = btrfs_qgroup_account_extent(&trans, fs_info, 4096, 4096, - old_roots, new_roots); + ret = btrfs_qgroup_account_extent(&trans, fs_info, nodesize, + nodesize, old_roots, new_roots); if (ret) { test_msg("Couldn't account space for a qgroup %d\n", ret); return ret; } - if (btrfs_verify_qgroup_counts(fs_info, 5, 4096, 0)) { + if (btrfs_verify_qgroup_counts(fs_info, 5, nodesize, 0)) { test_msg("Qgroup counts didn't match expected values\n"); return -EINVAL; } - if (btrfs_verify_qgroup_counts(fs_info, 256, 4096, 0)) { + if (btrfs_verify_qgroup_counts(fs_info, 256, nodesize, 0)) { test_msg("Qgroup counts didn't match expected values\n"); return -EINVAL; } - ret = btrfs_find_all_roots(&trans, fs_info, 4096, 0, &old_roots); + ret = btrfs_find_all_roots(&trans, fs_info, nodesize, 0, &old_roots); if (ret) { ulist_free(old_roots); test_msg("Couldn't find old roots: %d\n", ret); return ret; } - ret = remove_extent_ref(root, 4096, 4096, 0, 256); + ret = remove_extent_ref(root, nodesize, nodesize, 0, 256); if (ret) return ret; - ret = btrfs_find_all_roots(&trans, fs_info, 4096, 0, &new_roots); + ret = btrfs_find_all_roots(&trans, fs_info, nodesize, 0, &new_roots); if (ret) { ulist_free(old_roots); ulist_free(new_roots); @@ -415,8 +419,8 @@ static int test_multiple_refs(struct btrfs_root *root) return ret; } - ret = btrfs_qgroup_account_extent(&trans, fs_info, 4096, 4096, - old_roots, new_roots); + ret = btrfs_qgroup_account_extent(&trans, fs_info, nodesize, + nodesize, old_roots, new_roots); if (ret) { test_msg("Couldn't account space for a qgroup %d\n", ret); return ret; @@ -427,7 +431,7 @@ static int test_multiple_refs(struct btrfs_root *root) return -EINVAL; } - if (btrfs_verify_qgroup_counts(fs_info, 5, 4096, 4096)) { + if (btrfs_verify_qgroup_counts(fs_info, 5, nodesize, nodesize)) { test_msg("Qgroup counts didn't match expected values\n"); return -EINVAL; } @@ -435,13 +439,13 @@ static int test_multiple_refs(struct btrfs_root *root) return 0; } -int btrfs_test_qgroups(void) +int btrfs_test_qgroups(u32 sectorsize, u32 nodesize) { struct btrfs_root *root; struct btrfs_root *tmp_root; int ret = 0; - root = btrfs_alloc_dummy_root(); + root = btrfs_alloc_dummy_root(sectorsize, nodesize); if (IS_ERR(root)) { test_msg("Couldn't allocate root\n"); return PTR_ERR(root); @@ -468,7 +472,8 @@ int btrfs_test_qgroups(void) * Can't use bytenr 0, some things freak out * *cough*backref walking code*cough* */ - root->node = alloc_test_extent_buffer(root->fs_info, 4096); + root->node = alloc_test_extent_buffer(root->fs_info, nodesize, + nodesize); if (!root->node) { test_msg("Couldn't allocate dummy buffer\n"); ret = -ENOMEM; @@ -476,9 +481,9 @@ int btrfs_test_qgroups(void) } btrfs_set_header_level(root->node, 0); btrfs_set_header_nritems(root->node, 0); - root->alloc_bytenr += 8192; + root->alloc_bytenr += 2 * nodesize; - tmp_root = btrfs_alloc_dummy_root(); + tmp_root = btrfs_alloc_dummy_root(sectorsize, nodesize); if (IS_ERR(tmp_root)) { test_msg("Couldn't allocate a fs root\n"); ret = PTR_ERR(tmp_root); @@ -493,7 +498,7 @@ int btrfs_test_qgroups(void) goto out; } - tmp_root = btrfs_alloc_dummy_root(); + tmp_root = btrfs_alloc_dummy_root(sectorsize, nodesize); if (IS_ERR(tmp_root)) { test_msg("Couldn't allocate a fs root\n"); ret = PTR_ERR(tmp_root); @@ -508,10 +513,10 @@ int btrfs_test_qgroups(void) } test_msg("Running qgroup tests\n"); - ret = test_no_shared_qgroup(root); + ret = test_no_shared_qgroup(root, sectorsize, nodesize); if (ret) goto out; - ret = test_multiple_refs(root); + ret = test_multiple_refs(root, sectorsize, nodesize); out: btrfs_free_dummy_root(root); return ret; -- cgit v1.3-8-gc7d7 From 8dff9c85341032767d7b519217a79ea04cd676b0 Mon Sep 17 00:00:00 2001 From: Chris Mason Date: Sat, 19 Sep 2015 11:28:25 -0700 Subject: Btrfs: deal with duplciates during extent_map insertion in btrfs_get_extent When dealing with inline extents, btrfs_get_extent will incorrectly try to insert a duplicate extent_map. The dup hits -EEXIST from add_extent_map, but then we try to merge with the existing one and end up trying to insert a zero length extent_map. This actually works most of the time, except when there are extent maps past the end of the inline extent. rocksdb will trigger this sometimes because it preallocates an extent and then truncates down. Josef made a script to trigger with xfs_io: #!/bin/bash xfs_io -f -c "pwrite 0 1000" inline xfs_io -c "falloc -k 4k 1M" inline xfs_io -c "pread 0 1000" -c "fadvise -d 0 1000" -c "pread 0 1000" inline xfs_io -c "fadvise -d 0 1000" inline cat inline You'll get EIOs trying to read inline after this because add_extent_map is returning EEXIST Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 13 ++++++++++++- 1 file changed, 12 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 2debd42c4660..e5558d9e396e 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -6979,7 +6979,18 @@ insert: * existing will always be non-NULL, since there must be * extent causing the -EEXIST. */ - if (start >= extent_map_end(existing) || + if (existing->start == em->start && + extent_map_end(existing) == extent_map_end(em) && + em->block_start == existing->block_start) { + /* + * these two extents are the same, it happens + * with inlines especially + */ + free_extent_map(em); + em = existing; + err = 0; + + } else if (start >= extent_map_end(existing) || start <= existing->start) { /* * The existing extent map is the one nearest to -- cgit v1.3-8-gc7d7 From fac7d1917dfddfa53e98524f0abfbe60252740fe Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 4 Jun 2016 11:41:49 -0400 Subject: fix EOPENSTALE bug in do_last() EOPENSTALE occuring at the last component of a trailing symlink ends up with do_last() retrying its lookup. After the symlink body has been discarded. The thing is, all this retry_lookup logics in there is not needed at all - the upper layers will do the right thing if we simply return that -EOPENSTALE as we would with any other error. Trying to microoptimize in do_last() is a lot of headache for no good reason. Cc: stable@vger.kernel.org # v4.2+ Tested-by: Oleg Drokin Reviewed-and-Tested-by: Jeff Layton Signed-off-by: Al Viro --- fs/namei.c | 43 ++++--------------------------------------- 1 file changed, 4 insertions(+), 39 deletions(-) (limited to 'fs') diff --git a/fs/namei.c b/fs/namei.c index 4c4f95ac8aa5..3d9511e656ab 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -3166,9 +3166,7 @@ static int do_last(struct nameidata *nd, int acc_mode = op->acc_mode; unsigned seq; struct inode *inode; - struct path save_parent = { .dentry = NULL, .mnt = NULL }; struct path path; - bool retried = false; int error; nd->flags &= ~LOOKUP_PARENT; @@ -3211,7 +3209,6 @@ static int do_last(struct nameidata *nd, return -EISDIR; } -retry_lookup: if (open_flag & (O_CREAT | O_TRUNC | O_WRONLY | O_RDWR)) { error = mnt_want_write(nd->path.mnt); if (!error) @@ -3292,23 +3289,14 @@ finish_lookup: if (unlikely(error)) return error; - if ((nd->flags & LOOKUP_RCU) || nd->path.mnt != path.mnt) { - path_to_nameidata(&path, nd); - } else { - save_parent.dentry = nd->path.dentry; - save_parent.mnt = mntget(path.mnt); - nd->path.dentry = path.dentry; - - } + path_to_nameidata(&path, nd); nd->inode = inode; nd->seq = seq; /* Why this, you ask? _Now_ we might have grown LOOKUP_JUMPED... */ finish_open: error = complete_walk(nd); - if (error) { - path_put(&save_parent); + if (error) return error; - } audit_inode(nd->name, nd->path.dentry, 0); error = -EISDIR; if ((open_flag & O_CREAT) && d_is_dir(nd->path.dentry)) @@ -3331,13 +3319,9 @@ finish_open_created: goto out; BUG_ON(*opened & FILE_OPENED); /* once it's opened, it's opened */ error = vfs_open(&nd->path, file, current_cred()); - if (!error) { - *opened |= FILE_OPENED; - } else { - if (error == -EOPENSTALE) - goto stale_open; + if (error) goto out; - } + *opened |= FILE_OPENED; opened: error = open_check_o_direct(file); if (!error) @@ -3353,26 +3337,7 @@ out: } if (got_write) mnt_drop_write(nd->path.mnt); - path_put(&save_parent); return error; - -stale_open: - /* If no saved parent or already retried then can't retry */ - if (!save_parent.dentry || retried) - goto out; - - BUG_ON(save_parent.dentry != dir); - path_put(&nd->path); - nd->path = save_parent; - nd->inode = dir->d_inode; - save_parent.mnt = NULL; - save_parent.dentry = NULL; - if (got_write) { - mnt_drop_write(nd->path.mnt); - got_write = false; - } - retried = true; - goto retry_lookup; } static int do_tmpfile(struct nameidata *nd, unsigned flags, -- cgit v1.3-8-gc7d7 From e6ec03a25f12b312b7e0c037fe4a6471c4ee5665 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sun, 5 Jun 2016 00:23:09 -0400 Subject: autofs braino fix for do_last() It's an analogue of commit 7500c38a (fix the braino in "namei: massage lookup_slow() to be usable by lookup_one_len_unlocked()"). The same problem (->lookup()-returned unhashed negative dentry just might be an autofs one with ->d_manage() that would wait until the daemon makes it positive) applies in do_last() - we need to do follow_managed() first. Fortunately, remaining callers of follow_managed() are OK - only autofs has that weirdness (negative dentry that does not mean an instant -ENOENT)) and autofs never has its negative dentries hashed, so we can't pick one from a dcache lookup. ->d_manage() is a bloody mess ;-/ Cc: stable@vger.kernel.org # v4.6 Spotted-by: Ian Kent Signed-off-by: Al Viro --- fs/namei.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/namei.c b/fs/namei.c index 3d9511e656ab..d7c0cac56d89 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -3260,6 +3260,10 @@ static int do_last(struct nameidata *nd, got_write = false; } + error = follow_managed(&path, nd); + if (unlikely(error < 0)) + return error; + if (unlikely(d_is_negative(path.dentry))) { path_to_nameidata(&path, nd); return -ENOENT; @@ -3275,10 +3279,6 @@ static int do_last(struct nameidata *nd, return -EEXIST; } - error = follow_managed(&path, nd); - if (unlikely(error < 0)) - return error; - seq = 0; /* out of RCU mode, so the value doesn't matter */ inode = d_backing_inode(path.dentry); finish_lookup: -- cgit v1.3-8-gc7d7 From eedf265aa003b4781de24cfed40a655a664457e6 Mon Sep 17 00:00:00 2001 From: "Eric W. Biederman" Date: Thu, 2 Jun 2016 10:29:47 -0500 Subject: devpts: Make each mount of devpts an independent filesystem. The /dev/ptmx device node is changed to lookup the directory entry "pts" in the same directory as the /dev/ptmx device node was opened in. If there is a "pts" entry and that entry is a devpts filesystem /dev/ptmx uses that filesystem. Otherwise the open of /dev/ptmx fails. The DEVPTS_MULTIPLE_INSTANCES configuration option is removed, so that userspace can now safely depend on each mount of devpts creating a new instance of the filesystem. Each mount of devpts is now a separate and equal filesystem. Reserved ttys are now available to all instances of devpts where the mounter is in the initial mount namespace. A new vfs helper path_pts is introduced that finds a directory entry named "pts" in the directory of the passed in path, and changes the passed in path to point to it. The helper path_pts uses a function path_parent_directory that was factored out of follow_dotdot. In the implementation of devpts: - devpts_mnt is killed as it is no longer meaningful if all mounts of devpts are equal. - pts_sb_from_inode is replaced by just inode->i_sb as all cached inodes in the tty layer are now from the devpts filesystem. - devpts_add_ref is rolled into the new function devpts_ptmx. And the unnecessary inode hold is removed. - devpts_del_ref is renamed devpts_release and reduced to just a deacrivate_super. - The newinstance mount option continues to be accepted but is now ignored. In devpts_fs.h definitions for when !CONFIG_UNIX98_PTYS are removed as they are never used. Documentation/filesystems/devices.txt is updated to describe the current situation. This has been verified to work properly on openwrt-15.05, centos5, centos6, centos7, debian-6.0.2, debian-7.9, debian-8.2, ubuntu-14.04.3, ubuntu-15.10, fedora23, magia-5, mint-17.3, opensuse-42.1, slackware-14.1, gentoo-20151225 (13.0?), archlinux-2015-12-01. With the caveat that on centos6 and on slackware-14.1 that there wind up being two instances of the devpts filesystem mounted on /dev/pts, the lower copy does not end up getting used. Signed-off-by: "Eric W. Biederman" Cc: Greg KH Cc: Peter Hurley Cc: Peter Anvin Cc: Andy Lutomirski Cc: Al Viro Cc: Serge Hallyn Cc: Willy Tarreau Cc: Aurelien Jarno Cc: One Thousand Gnomes Cc: Jann Horn Cc: Jiri Slaby Cc: Florian Weimer Cc: Konstantin Khlebnikov Signed-off-by: Linus Torvalds --- Documentation/filesystems/devpts.txt | 145 +++----------------------- drivers/tty/Kconfig | 11 -- drivers/tty/pty.c | 15 +-- fs/devpts/inode.c | 191 ++++++++++------------------------- fs/namei.c | 49 +++++++-- include/linux/devpts_fs.h | 9 +- include/linux/namei.h | 2 + 7 files changed, 126 insertions(+), 296 deletions(-) (limited to 'fs') diff --git a/Documentation/filesystems/devpts.txt b/Documentation/filesystems/devpts.txt index 30d2fcb32f72..9f94fe276dea 100644 --- a/Documentation/filesystems/devpts.txt +++ b/Documentation/filesystems/devpts.txt @@ -1,141 +1,26 @@ +Each mount of the devpts filesystem is now distinct such that ptys +and their indicies allocated in one mount are independent from ptys +and their indicies in all other mounts. -To support containers, we now allow multiple instances of devpts filesystem, -such that indices of ptys allocated in one instance are independent of indices -allocated in other instances of devpts. +All mounts of the devpts filesystem now create a /dev/pts/ptmx node +with permissions 0000. -To preserve backward compatibility, this support for multiple instances is -enabled only if: +To retain backwards compatibility the a ptmx device node (aka any node +created with "mknod name c 5 2") when opened will look for an instance +of devpts under the name "pts" in the same directory as the ptmx device +node. - - CONFIG_DEVPTS_MULTIPLE_INSTANCES=y, and - - '-o newinstance' mount option is specified while mounting devpts - -IOW, devpts now supports both single-instance and multi-instance semantics. - -If CONFIG_DEVPTS_MULTIPLE_INSTANCES=n, there is no change in behavior and -this referred to as the "legacy" mode. In this mode, the new mount options -(-o newinstance and -o ptmxmode) will be ignored with a 'bogus option' message -on console. - -If CONFIG_DEVPTS_MULTIPLE_INSTANCES=y and devpts is mounted without the -'newinstance' option (as in current start-up scripts) the new mount binds -to the initial kernel mount of devpts. This mode is referred to as the -'single-instance' mode and the current, single-instance semantics are -preserved, i.e PTYs are common across the system. - -The only difference between this single-instance mode and the legacy mode -is the presence of new, '/dev/pts/ptmx' node with permissions 0000, which -can safely be ignored. - -If CONFIG_DEVPTS_MULTIPLE_INSTANCES=y and 'newinstance' option is specified, -the mount is considered to be in the multi-instance mode and a new instance -of the devpts fs is created. Any ptys created in this instance are independent -of ptys in other instances of devpts. Like in the single-instance mode, the -/dev/pts/ptmx node is present. To effectively use the multi-instance mode, -open of /dev/ptmx must be a redirected to '/dev/pts/ptmx' using a symlink or -bind-mount. - -Eg: A container startup script could do the following: - - $ chmod 0666 /dev/pts/ptmx - $ rm /dev/ptmx - $ ln -s pts/ptmx /dev/ptmx - $ ns_exec -cm /bin/bash - - # We are now in new container - - $ umount /dev/pts - $ mount -t devpts -o newinstance lxcpts /dev/pts - $ sshd -p 1234 - -where 'ns_exec -cm /bin/bash' calls clone() with CLONE_NEWNS flag and execs -/bin/bash in the child process. A pty created by the sshd is not visible in -the original mount of /dev/pts. +As an option instead of placing a /dev/ptmx device node at /dev/ptmx +it is possible to place a symlink to /dev/pts/ptmx at /dev/ptmx or +to bind mount /dev/ptx/ptmx to /dev/ptmx. If you opt for using +the devpts filesystem in this manner devpts should be mounted with +the ptmxmode=0666, or chmod 0666 /dev/pts/ptmx should be called. Total count of pty pairs in all instances is limited by sysctls: kernel.pty.max = 4096 - global limit -kernel.pty.reserve = 1024 - reserve for initial instance +kernel.pty.reserve = 1024 - reserved for filesystems mounted from the initial mount namespace kernel.pty.nr - current count of ptys Per-instance limit could be set by adding mount option "max=". This feature was added in kernel 3.4 together with sysctl kernel.pty.reserve. In kernels older than 3.4 sysctl kernel.pty.max works as per-instance limit. - -User-space changes ------------------- - -In multi-instance mode (i.e '-o newinstance' mount option is specified at least -once), following user-space issues should be noted. - -1. If -o newinstance mount option is never used, /dev/pts/ptmx can be ignored - and no change is needed to system-startup scripts. - -2. To effectively use multi-instance mode (i.e -o newinstance is specified) - administrators or startup scripts should "redirect" open of /dev/ptmx to - /dev/pts/ptmx using either a bind mount or symlink. - - $ mount -t devpts -o newinstance devpts /dev/pts - - followed by either - - $ rm /dev/ptmx - $ ln -s pts/ptmx /dev/ptmx - $ chmod 666 /dev/pts/ptmx - or - $ mount -o bind /dev/pts/ptmx /dev/ptmx - -3. The '/dev/ptmx -> pts/ptmx' symlink is the preferred method since it - enables better error-reporting and treats both single-instance and - multi-instance mounts similarly. - - But this method requires that system-startup scripts set the mode of - /dev/pts/ptmx correctly (default mode is 0000). The scripts can set the - mode by, either - - - adding ptmxmode mount option to devpts entry in /etc/fstab, or - - using 'chmod 0666 /dev/pts/ptmx' - -4. If multi-instance mode mount is needed for containers, but the system - startup scripts have not yet been updated, container-startup scripts - should bind mount /dev/ptmx to /dev/pts/ptmx to avoid breaking single- - instance mounts. - - Or, in general, container-startup scripts should use: - - mount -t devpts -o newinstance -o ptmxmode=0666 devpts /dev/pts - if [ ! -L /dev/ptmx ]; then - mount -o bind /dev/pts/ptmx /dev/ptmx - fi - - When all devpts mounts are multi-instance, /dev/ptmx can permanently be - a symlink to pts/ptmx and the bind mount can be ignored. - -5. A multi-instance mount that is not accompanied by the /dev/ptmx to - /dev/pts/ptmx redirection would result in an unusable/unreachable pty. - - mount -t devpts -o newinstance lxcpts /dev/pts - - immediately followed by: - - open("/dev/ptmx") - - would create a pty, say /dev/pts/7, in the initial kernel mount. - But /dev/pts/7 would be invisible in the new mount. - -6. The permissions for /dev/pts/ptmx node should be specified when mounting - /dev/pts, using the '-o ptmxmode=%o' mount option (default is 0000). - - mount -t devpts -o newinstance -o ptmxmode=0644 devpts /dev/pts - - The permissions can be later be changed as usual with 'chmod'. - - chmod 666 /dev/pts/ptmx - -7. A mount of devpts without the 'newinstance' option results in binding to - initial kernel mount. This behavior while preserving legacy semantics, - does not provide strict isolation in a container environment. i.e by - mounting devpts without the 'newinstance' option, a container could - get visibility into the 'host' or root container's devpts. - - To workaround this and have strict isolation, all mounts of devpts, - including the mount in the root container, should use the newinstance - option. diff --git a/drivers/tty/Kconfig b/drivers/tty/Kconfig index 82c4d2e45319..95103054c0e4 100644 --- a/drivers/tty/Kconfig +++ b/drivers/tty/Kconfig @@ -120,17 +120,6 @@ config UNIX98_PTYS All modern Linux systems use the Unix98 ptys. Say Y unless you're on an embedded system and want to conserve memory. -config DEVPTS_MULTIPLE_INSTANCES - bool "Support multiple instances of devpts" - depends on UNIX98_PTYS - default n - ---help--- - Enable support for multiple instances of devpts filesystem. - If you want to have isolated PTY namespaces (eg: in containers), - say Y here. Otherwise, say N. If enabled, each mount of devpts - filesystem with the '-o newinstance' option will create an - independent PTY namespace. - config LEGACY_PTYS bool "Legacy (BSD) PTY support" default y diff --git a/drivers/tty/pty.c b/drivers/tty/pty.c index dd4b8417e7f4..f856c4544eea 100644 --- a/drivers/tty/pty.c +++ b/drivers/tty/pty.c @@ -668,7 +668,7 @@ static void pty_unix98_remove(struct tty_driver *driver, struct tty_struct *tty) else fsi = tty->link->driver_data; devpts_kill_index(fsi, tty->index); - devpts_put_ref(fsi); + devpts_release(fsi); } static const struct tty_operations ptm_unix98_ops = { @@ -733,10 +733,11 @@ static int ptmx_open(struct inode *inode, struct file *filp) if (retval) return retval; - fsi = devpts_get_ref(inode, filp); - retval = -ENODEV; - if (!fsi) + fsi = devpts_acquire(filp); + if (IS_ERR(fsi)) { + retval = PTR_ERR(fsi); goto out_free_file; + } /* find a device that is not in use. */ mutex_lock(&devpts_mutex); @@ -745,7 +746,7 @@ static int ptmx_open(struct inode *inode, struct file *filp) retval = index; if (index < 0) - goto out_put_ref; + goto out_put_fsi; mutex_lock(&tty_mutex); @@ -789,8 +790,8 @@ err_release: return retval; out: devpts_kill_index(fsi, index); -out_put_ref: - devpts_put_ref(fsi); +out_put_fsi: + devpts_release(fsi); out_free_file: tty_free_file(filp); return retval; diff --git a/fs/devpts/inode.c b/fs/devpts/inode.c index 0b2954d7172d..37c134a132c7 100644 --- a/fs/devpts/inode.c +++ b/fs/devpts/inode.c @@ -95,8 +95,6 @@ static struct ctl_table pty_root_table[] = { static DEFINE_MUTEX(allocated_ptys_lock); -static struct vfsmount *devpts_mnt; - struct pts_mount_opts { int setuid; int setgid; @@ -104,7 +102,7 @@ struct pts_mount_opts { kgid_t gid; umode_t mode; umode_t ptmxmode; - int newinstance; + int reserve; int max; }; @@ -117,11 +115,9 @@ static const match_table_t tokens = { {Opt_uid, "uid=%u"}, {Opt_gid, "gid=%u"}, {Opt_mode, "mode=%o"}, -#ifdef CONFIG_DEVPTS_MULTIPLE_INSTANCES {Opt_ptmxmode, "ptmxmode=%o"}, {Opt_newinstance, "newinstance"}, {Opt_max, "max=%d"}, -#endif {Opt_err, NULL} }; @@ -137,15 +133,48 @@ static inline struct pts_fs_info *DEVPTS_SB(struct super_block *sb) return sb->s_fs_info; } -static inline struct super_block *pts_sb_from_inode(struct inode *inode) +struct pts_fs_info *devpts_acquire(struct file *filp) { -#ifdef CONFIG_DEVPTS_MULTIPLE_INSTANCES - if (inode->i_sb->s_magic == DEVPTS_SUPER_MAGIC) - return inode->i_sb; -#endif - if (!devpts_mnt) - return NULL; - return devpts_mnt->mnt_sb; + struct pts_fs_info *result; + struct path path; + struct super_block *sb; + int err; + + path = filp->f_path; + path_get(&path); + + /* Has the devpts filesystem already been found? */ + sb = path.mnt->mnt_sb; + if (sb->s_magic != DEVPTS_SUPER_MAGIC) { + /* Is a devpts filesystem at "pts" in the same directory? */ + err = path_pts(&path); + if (err) { + result = ERR_PTR(err); + goto out; + } + + /* Is the path the root of a devpts filesystem? */ + result = ERR_PTR(-ENODEV); + sb = path.mnt->mnt_sb; + if ((sb->s_magic != DEVPTS_SUPER_MAGIC) || + (path.mnt->mnt_root != sb->s_root)) + goto out; + } + + /* + * pty code needs to hold extra references in case of last /dev/tty close + */ + atomic_inc(&sb->s_active); + result = DEVPTS_SB(sb); + +out: + path_put(&path); + return result; +} + +void devpts_release(struct pts_fs_info *fsi) +{ + deactivate_super(fsi->sb); } #define PARSE_MOUNT 0 @@ -154,9 +183,7 @@ static inline struct super_block *pts_sb_from_inode(struct inode *inode) /* * parse_mount_options(): * Set @opts to mount options specified in @data. If an option is not - * specified in @data, set it to its default value. The exception is - * 'newinstance' option which can only be set/cleared on a mount (i.e. - * cannot be changed during remount). + * specified in @data, set it to its default value. * * Note: @data may be NULL (in which case all options are set to default). */ @@ -174,9 +201,12 @@ static int parse_mount_options(char *data, int op, struct pts_mount_opts *opts) opts->ptmxmode = DEVPTS_DEFAULT_PTMX_MODE; opts->max = NR_UNIX98_PTY_MAX; - /* newinstance makes sense only on initial mount */ + /* Only allow instances mounted from the initial mount + * namespace to tap the reserve pool of ptys. + */ if (op == PARSE_MOUNT) - opts->newinstance = 0; + opts->reserve = + (current->nsproxy->mnt_ns == init_task.nsproxy->mnt_ns); while ((p = strsep(&data, ",")) != NULL) { substring_t args[MAX_OPT_ARGS]; @@ -211,16 +241,12 @@ static int parse_mount_options(char *data, int op, struct pts_mount_opts *opts) return -EINVAL; opts->mode = option & S_IALLUGO; break; -#ifdef CONFIG_DEVPTS_MULTIPLE_INSTANCES case Opt_ptmxmode: if (match_octal(&args[0], &option)) return -EINVAL; opts->ptmxmode = option & S_IALLUGO; break; case Opt_newinstance: - /* newinstance makes sense only on initial mount */ - if (op == PARSE_MOUNT) - opts->newinstance = 1; break; case Opt_max: if (match_int(&args[0], &option) || @@ -228,7 +254,6 @@ static int parse_mount_options(char *data, int op, struct pts_mount_opts *opts) return -EINVAL; opts->max = option; break; -#endif default: pr_err("called with bogus options\n"); return -EINVAL; @@ -238,7 +263,6 @@ static int parse_mount_options(char *data, int op, struct pts_mount_opts *opts) return 0; } -#ifdef CONFIG_DEVPTS_MULTIPLE_INSTANCES static int mknod_ptmx(struct super_block *sb) { int mode; @@ -305,12 +329,6 @@ static void update_ptmx_mode(struct pts_fs_info *fsi) inode->i_mode = S_IFCHR|fsi->mount_opts.ptmxmode; } } -#else -static inline void update_ptmx_mode(struct pts_fs_info *fsi) -{ - return; -} -#endif static int devpts_remount(struct super_block *sb, int *flags, char *data) { @@ -344,11 +362,9 @@ static int devpts_show_options(struct seq_file *seq, struct dentry *root) seq_printf(seq, ",gid=%u", from_kgid_munged(&init_user_ns, opts->gid)); seq_printf(seq, ",mode=%03o", opts->mode); -#ifdef CONFIG_DEVPTS_MULTIPLE_INSTANCES seq_printf(seq, ",ptmxmode=%03o", opts->ptmxmode); if (opts->max < NR_UNIX98_PTY_MAX) seq_printf(seq, ",max=%d", opts->max); -#endif return 0; } @@ -410,40 +426,11 @@ fail: return -ENOMEM; } -#ifdef CONFIG_DEVPTS_MULTIPLE_INSTANCES -static int compare_init_pts_sb(struct super_block *s, void *p) -{ - if (devpts_mnt) - return devpts_mnt->mnt_sb == s; - return 0; -} - /* * devpts_mount() * - * If the '-o newinstance' mount option was specified, mount a new - * (private) instance of devpts. PTYs created in this instance are - * independent of the PTYs in other devpts instances. - * - * If the '-o newinstance' option was not specified, mount/remount the - * initial kernel mount of devpts. This type of mount gives the - * legacy, single-instance semantics. - * - * The 'newinstance' option is needed to support multiple namespace - * semantics in devpts while preserving backward compatibility of the - * current 'single-namespace' semantics. i.e all mounts of devpts - * without the 'newinstance' mount option should bind to the initial - * kernel mount, like mount_single(). - * - * Mounts with 'newinstance' option create a new, private namespace. - * - * NOTE: - * - * For single-mount semantics, devpts cannot use mount_single(), - * because mount_single()/sget() find and use the super-block from - * the most recent mount of devpts. But that recent mount may be a - * 'newinstance' mount and mount_single() would pick the newinstance - * super-block instead of the initial super-block. + * Mount a new (private) instance of devpts. PTYs created in this + * instance are independent of the PTYs in other devpts instances. */ static struct dentry *devpts_mount(struct file_system_type *fs_type, int flags, const char *dev_name, void *data) @@ -456,18 +443,7 @@ static struct dentry *devpts_mount(struct file_system_type *fs_type, if (error) return ERR_PTR(error); - /* Require newinstance for all user namespace mounts to ensure - * the mount options are not changed. - */ - if ((current_user_ns() != &init_user_ns) && !opts.newinstance) - return ERR_PTR(-EINVAL); - - if (opts.newinstance) - s = sget(fs_type, NULL, set_anon_super, flags, NULL); - else - s = sget(fs_type, compare_init_pts_sb, set_anon_super, flags, - NULL); - + s = sget(fs_type, NULL, set_anon_super, flags, NULL); if (IS_ERR(s)) return ERR_CAST(s); @@ -491,18 +467,6 @@ out_undo_sget: return ERR_PTR(error); } -#else -/* - * This supports only the legacy single-instance semantics (no - * multiple-instance semantics) - */ -static struct dentry *devpts_mount(struct file_system_type *fs_type, int flags, - const char *dev_name, void *data) -{ - return mount_single(fs_type, flags, data, devpts_fill_super); -} -#endif - static void devpts_kill_sb(struct super_block *sb) { struct pts_fs_info *fsi = DEVPTS_SB(sb); @@ -516,9 +480,7 @@ static struct file_system_type devpts_fs_type = { .name = "devpts", .mount = devpts_mount, .kill_sb = devpts_kill_sb, -#ifdef CONFIG_DEVPTS_MULTIPLE_INSTANCES .fs_flags = FS_USERNS_MOUNT | FS_USERNS_DEV_MOUNT, -#endif }; /* @@ -531,16 +493,13 @@ int devpts_new_index(struct pts_fs_info *fsi) int index; int ida_ret; - if (!fsi) - return -ENODEV; - retry: if (!ida_pre_get(&fsi->allocated_ptys, GFP_KERNEL)) return -ENOMEM; mutex_lock(&allocated_ptys_lock); - if (pty_count >= pty_limit - - (fsi->mount_opts.newinstance ? pty_reserve : 0)) { + if (pty_count >= (pty_limit - + (fsi->mount_opts.reserve ? 0 : pty_reserve))) { mutex_unlock(&allocated_ptys_lock); return -ENOSPC; } @@ -571,30 +530,6 @@ void devpts_kill_index(struct pts_fs_info *fsi, int idx) mutex_unlock(&allocated_ptys_lock); } -/* - * pty code needs to hold extra references in case of last /dev/tty close - */ -struct pts_fs_info *devpts_get_ref(struct inode *ptmx_inode, struct file *file) -{ - struct super_block *sb; - struct pts_fs_info *fsi; - - sb = pts_sb_from_inode(ptmx_inode); - if (!sb) - return NULL; - fsi = DEVPTS_SB(sb); - if (!fsi) - return NULL; - - atomic_inc(&sb->s_active); - return fsi; -} - -void devpts_put_ref(struct pts_fs_info *fsi) -{ - deactivate_super(fsi->sb); -} - /** * devpts_pty_new -- create a new inode in /dev/pts/ * @ptmx_inode: inode of the master @@ -607,16 +542,12 @@ void devpts_put_ref(struct pts_fs_info *fsi) struct dentry *devpts_pty_new(struct pts_fs_info *fsi, int index, void *priv) { struct dentry *dentry; - struct super_block *sb; + struct super_block *sb = fsi->sb; struct inode *inode; struct dentry *root; struct pts_mount_opts *opts; char s[12]; - if (!fsi) - return ERR_PTR(-ENODEV); - - sb = fsi->sb; root = sb->s_root; opts = &fsi->mount_opts; @@ -676,20 +607,8 @@ void devpts_pty_kill(struct dentry *dentry) static int __init init_devpts_fs(void) { int err = register_filesystem(&devpts_fs_type); - struct ctl_table_header *table; - if (!err) { - struct vfsmount *mnt; - - table = register_sysctl_table(pty_root_table); - mnt = kern_mount(&devpts_fs_type); - if (IS_ERR(mnt)) { - err = PTR_ERR(mnt); - unregister_filesystem(&devpts_fs_type); - unregister_sysctl_table(table); - } else { - devpts_mnt = mnt; - } + register_sysctl_table(pty_root_table); } return err; } diff --git a/fs/namei.c b/fs/namei.c index 4c4f95ac8aa5..6a82fb7e2127 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -1416,21 +1416,28 @@ static void follow_mount(struct path *path) } } +static int path_parent_directory(struct path *path) +{ + struct dentry *old = path->dentry; + /* rare case of legitimate dget_parent()... */ + path->dentry = dget_parent(path->dentry); + dput(old); + if (unlikely(!path_connected(path))) + return -ENOENT; + return 0; +} + static int follow_dotdot(struct nameidata *nd) { while(1) { - struct dentry *old = nd->path.dentry; - if (nd->path.dentry == nd->root.dentry && nd->path.mnt == nd->root.mnt) { break; } if (nd->path.dentry != nd->path.mnt->mnt_root) { - /* rare case of legitimate dget_parent()... */ - nd->path.dentry = dget_parent(nd->path.dentry); - dput(old); - if (unlikely(!path_connected(&nd->path))) - return -ENOENT; + int ret = path_parent_directory(&nd->path); + if (ret) + return ret; break; } if (!follow_up(&nd->path)) @@ -2514,6 +2521,34 @@ struct dentry *lookup_one_len_unlocked(const char *name, } EXPORT_SYMBOL(lookup_one_len_unlocked); +#ifdef CONFIG_UNIX98_PTYS +int path_pts(struct path *path) +{ + /* Find something mounted on "pts" in the same directory as + * the input path. + */ + struct dentry *child, *parent; + struct qstr this; + int ret; + + ret = path_parent_directory(path); + if (ret) + return ret; + + parent = path->dentry; + this.name = "pts"; + this.len = 3; + child = d_hash_and_lookup(parent, &this); + if (!child) + return -ENOENT; + + path->dentry = child; + dput(parent); + follow_mount(path); + return 0; +} +#endif + int user_path_at_empty(int dfd, const char __user *name, unsigned flags, struct path *path, int *empty) { diff --git a/include/linux/devpts_fs.h b/include/linux/devpts_fs.h index 5871f292b596..277ab9af9ac2 100644 --- a/include/linux/devpts_fs.h +++ b/include/linux/devpts_fs.h @@ -15,13 +15,12 @@ #include -struct pts_fs_info; - #ifdef CONFIG_UNIX98_PTYS -/* Look up a pts fs info and get a ref to it */ -struct pts_fs_info *devpts_get_ref(struct inode *, struct file *); -void devpts_put_ref(struct pts_fs_info *); +struct pts_fs_info; + +struct pts_fs_info *devpts_acquire(struct file *); +void devpts_release(struct pts_fs_info *); int devpts_new_index(struct pts_fs_info *); void devpts_kill_index(struct pts_fs_info *, int); diff --git a/include/linux/namei.h b/include/linux/namei.h index ec5ec2818a28..d3d0398f2a1b 100644 --- a/include/linux/namei.h +++ b/include/linux/namei.h @@ -45,6 +45,8 @@ enum {LAST_NORM, LAST_ROOT, LAST_DOT, LAST_DOTDOT, LAST_BIND}; #define LOOKUP_ROOT 0x2000 #define LOOKUP_EMPTY 0x4000 +extern int path_pts(struct path *path); + extern int user_path_at_empty(int, const char __user *, unsigned, struct path *, int *empty); static inline int user_path_at(int dfd, const char __user *name, unsigned flags, -- cgit v1.3-8-gc7d7 From d865177a5e749827f248f6363f5100d3a2f66b0f Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Fri, 3 Jun 2016 17:41:42 -0700 Subject: Btrfs: clear uptodate flags of pages in sys_array eb We set uptodate flag to pages in the temporary sys_array eb, but do not clear the flag after free eb. As the special btree inode may still hold a reference on those pages, the uptodate flag can remain alive in them. If btrfs_super_chunk_root has been intentionally changed to the offset of this sys_array eb, reading chunk_root will read content of sys_array and it will skip our beautiful checks in btree_readpage_end_io_hook() because of "pages of eb are uptodate => eb is uptodate" This adds the 'clear uptodate' part to force it to read from disk. Reviewed-by: Josef Bacik Signed-off-by: Liu Bo Signed-off-by: David Sterba --- fs/btrfs/volumes.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 673c72ab4fbe..42ccde43053b 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -6631,12 +6631,14 @@ int btrfs_read_sys_array(struct btrfs_root *root) sb_array_offset += len; cur_offset += len; } + clear_extent_buffer_uptodate(sb); free_extent_buffer_stale(sb); return ret; out_short_read: printk(KERN_ERR "BTRFS: sys_array too short to read %u bytes at offset %u\n", len, cur_offset); + clear_extent_buffer_uptodate(sb); free_extent_buffer_stale(sb); return -EIO; } -- cgit v1.3-8-gc7d7 From 99e3ecfcb9f4ca35192d20a5bea158b81f600062 Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Fri, 3 Jun 2016 12:05:14 -0700 Subject: Btrfs: add more validation checks for superblock This adds validation checks for super_total_bytes, super_bytes_used and super_stripesize, super_num_devices. Reported-by: Vegard Nossum Reported-by: Quentin Casasnovas Reviewed-by: David Sterba Signed-off-by: Liu Bo Signed-off-by: David Sterba --- fs/btrfs/disk-io.c | 11 +++++++++++ fs/btrfs/volumes.c | 24 ++++++++++++++++++++++++ 2 files changed, 35 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 6628fca9f4ed..2bd5f5e975cc 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -4130,6 +4130,17 @@ static int btrfs_check_super_valid(struct btrfs_fs_info *fs_info, * Hint to catch really bogus numbers, bitflips or so, more exact checks are * done later */ + if (btrfs_super_bytes_used(sb) < 6 * btrfs_super_nodesize(sb)) { + btrfs_err(fs_info, "bytes_used is too small %llu", + btrfs_super_bytes_used(sb)); + ret = -EINVAL; + } + if (!is_power_of_2(btrfs_super_stripesize(sb)) || + btrfs_super_stripesize(sb) != sectorsize) { + btrfs_err(fs_info, "invalid stripesize %u", + btrfs_super_stripesize(sb)); + ret = -EINVAL; + } if (btrfs_super_num_devices(sb) > (1UL << 31)) printk(KERN_WARNING "BTRFS: suspicious number of devices: %llu\n", btrfs_super_num_devices(sb)); diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 42ccde43053b..fd5c9e69894a 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -6651,6 +6651,7 @@ int btrfs_read_chunk_tree(struct btrfs_root *root) struct btrfs_key found_key; int ret; int slot; + u64 total_dev = 0; root = root->fs_info->chunk_root; @@ -6692,6 +6693,7 @@ int btrfs_read_chunk_tree(struct btrfs_root *root) ret = read_one_dev(root, leaf, dev_item); if (ret) goto error; + total_dev++; } else if (found_key.type == BTRFS_CHUNK_ITEM_KEY) { struct btrfs_chunk *chunk; chunk = btrfs_item_ptr(leaf, slot, struct btrfs_chunk); @@ -6701,6 +6703,28 @@ int btrfs_read_chunk_tree(struct btrfs_root *root) } path->slots[0]++; } + + /* + * After loading chunk tree, we've got all device information, + * do another round of validation checks. + */ + if (total_dev != root->fs_info->fs_devices->total_devices) { + btrfs_err(root->fs_info, + "super_num_devices %llu mismatch with num_devices %llu found here", + btrfs_super_num_devices(root->fs_info->super_copy), + total_dev); + ret = -EINVAL; + goto error; + } + if (btrfs_super_total_bytes(root->fs_info->super_copy) < + root->fs_info->fs_devices->total_rw_bytes) { + btrfs_err(root->fs_info, + "super_total_bytes %llu mismatch with fs_devices total_rw_bytes %llu", + btrfs_super_total_bytes(root->fs_info->super_copy), + root->fs_info->fs_devices->total_rw_bytes); + ret = -EINVAL; + goto error; + } ret = 0; error: unlock_chunks(root); -- cgit v1.3-8-gc7d7 From e06cd3dd7cea50e87663a88acdfdb7ac1c53a5ca Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Fri, 3 Jun 2016 12:05:15 -0700 Subject: Btrfs: add validadtion checks for chunk loading To prevent fuzzed filesystem images from panic the whole system, we need various validation checks to refuse to mount such an image if btrfs finds any invalid value during loading chunks, including both sys_array and regular chunks. Note that these checks may not be sufficient to cover all corner cases, feel free to add more checks. Reported-by: Vegard Nossum Reported-by: Quentin Casasnovas Reviewed-by: David Sterba Signed-off-by: Liu Bo Signed-off-by: David Sterba --- fs/btrfs/volumes.c | 82 ++++++++++++++++++++++++++++++++++++++++++++---------- 1 file changed, 67 insertions(+), 15 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index fd5c9e69894a..74507b05061b 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -6251,27 +6251,23 @@ struct btrfs_device *btrfs_alloc_device(struct btrfs_fs_info *fs_info, return dev; } -static int read_one_chunk(struct btrfs_root *root, struct btrfs_key *key, - struct extent_buffer *leaf, - struct btrfs_chunk *chunk) +/* Return -EIO if any error, otherwise return 0. */ +static int btrfs_check_chunk_valid(struct btrfs_root *root, + struct extent_buffer *leaf, + struct btrfs_chunk *chunk, u64 logical) { - struct btrfs_mapping_tree *map_tree = &root->fs_info->mapping_tree; - struct map_lookup *map; - struct extent_map *em; - u64 logical; u64 length; u64 stripe_len; - u64 devid; - u8 uuid[BTRFS_UUID_SIZE]; - int num_stripes; - int ret; - int i; + u16 num_stripes; + u16 sub_stripes; + u64 type; - logical = key->offset; length = btrfs_chunk_length(leaf, chunk); stripe_len = btrfs_chunk_stripe_len(leaf, chunk); num_stripes = btrfs_chunk_num_stripes(leaf, chunk); - /* Validation check */ + sub_stripes = btrfs_chunk_sub_stripes(leaf, chunk); + type = btrfs_chunk_type(leaf, chunk); + if (!num_stripes) { btrfs_err(root->fs_info, "invalid chunk num_stripes: %u", num_stripes); @@ -6282,6 +6278,11 @@ static int read_one_chunk(struct btrfs_root *root, struct btrfs_key *key, "invalid chunk logical %llu", logical); return -EIO; } + if (btrfs_chunk_sector_size(leaf, chunk) != root->sectorsize) { + btrfs_err(root->fs_info, "invalid chunk sectorsize %u", + btrfs_chunk_sector_size(leaf, chunk)); + return -EIO; + } if (!length || !IS_ALIGNED(length, root->sectorsize)) { btrfs_err(root->fs_info, "invalid chunk length %llu", length); @@ -6293,13 +6294,54 @@ static int read_one_chunk(struct btrfs_root *root, struct btrfs_key *key, return -EIO; } if (~(BTRFS_BLOCK_GROUP_TYPE_MASK | BTRFS_BLOCK_GROUP_PROFILE_MASK) & - btrfs_chunk_type(leaf, chunk)) { + type) { btrfs_err(root->fs_info, "unrecognized chunk type: %llu", ~(BTRFS_BLOCK_GROUP_TYPE_MASK | BTRFS_BLOCK_GROUP_PROFILE_MASK) & btrfs_chunk_type(leaf, chunk)); return -EIO; } + if ((type & BTRFS_BLOCK_GROUP_RAID10 && sub_stripes != 2) || + (type & BTRFS_BLOCK_GROUP_RAID1 && num_stripes < 1) || + (type & BTRFS_BLOCK_GROUP_RAID5 && num_stripes < 2) || + (type & BTRFS_BLOCK_GROUP_RAID6 && num_stripes < 3) || + (type & BTRFS_BLOCK_GROUP_DUP && num_stripes > 2) || + ((type & BTRFS_BLOCK_GROUP_PROFILE_MASK) == 0 && + num_stripes != 1)) { + btrfs_err(root->fs_info, + "invalid num_stripes:sub_stripes %u:%u for profile %llu", + num_stripes, sub_stripes, + type & BTRFS_BLOCK_GROUP_PROFILE_MASK); + return -EIO; + } + + return 0; +} + +static int read_one_chunk(struct btrfs_root *root, struct btrfs_key *key, + struct extent_buffer *leaf, + struct btrfs_chunk *chunk) +{ + struct btrfs_mapping_tree *map_tree = &root->fs_info->mapping_tree; + struct map_lookup *map; + struct extent_map *em; + u64 logical; + u64 length; + u64 stripe_len; + u64 devid; + u8 uuid[BTRFS_UUID_SIZE]; + int num_stripes; + int ret; + int i; + + logical = key->offset; + length = btrfs_chunk_length(leaf, chunk); + stripe_len = btrfs_chunk_stripe_len(leaf, chunk); + num_stripes = btrfs_chunk_num_stripes(leaf, chunk); + + ret = btrfs_check_chunk_valid(root, leaf, chunk, logical); + if (ret) + return ret; read_lock(&map_tree->map_tree.lock); em = lookup_extent_mapping(&map_tree->map_tree, logical, 1); @@ -6547,6 +6589,7 @@ int btrfs_read_sys_array(struct btrfs_root *root) u32 array_size; u32 len = 0; u32 cur_offset; + u64 type; struct btrfs_key key; ASSERT(BTRFS_SUPER_INFO_SIZE <= root->nodesize); @@ -6613,6 +6656,15 @@ int btrfs_read_sys_array(struct btrfs_root *root) break; } + type = btrfs_chunk_type(sb, chunk); + if ((type & BTRFS_BLOCK_GROUP_SYSTEM) == 0) { + btrfs_err(root->fs_info, + "invalid chunk type %llu in sys_array at offset %u", + type, cur_offset); + ret = -EIO; + break; + } + len = btrfs_chunk_item_size(num_stripes); if (cur_offset + len > array_size) goto out_short_read; -- cgit v1.3-8-gc7d7 From 5f9e1059d9347191b271bf7d13bd83db57594d2a Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Wed, 16 Sep 2015 15:34:53 +0200 Subject: btrfs: advertise which crc32c implementation is being used at module load Since several architectures support hardware-accelerated crc32c calculation, it would be nice to confirm that btrfs is actually using it. We can see an elevated use count for the module, but it doesn't actually show who the users are. This patch simply prints the name of the driver after successfully initializing the shash. Signed-off-by: Jeff Mahoney [ added a helper and used in module load-time message ] Signed-off-by: David Sterba --- fs/btrfs/hash.c | 5 +++++ fs/btrfs/hash.h | 1 + fs/btrfs/super.c | 5 +++-- 3 files changed, 9 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/hash.c b/fs/btrfs/hash.c index aae520b2aee5..a97fdc156a03 100644 --- a/fs/btrfs/hash.c +++ b/fs/btrfs/hash.c @@ -24,6 +24,11 @@ int __init btrfs_hash_init(void) return PTR_ERR_OR_ZERO(tfm); } +const char* btrfs_crc32c_impl(void) +{ + return crypto_tfm_alg_driver_name(crypto_shash_tfm(tfm)); +} + void btrfs_hash_exit(void) { crypto_free_shash(tfm); diff --git a/fs/btrfs/hash.h b/fs/btrfs/hash.h index 118a2316e5d3..c3a2ec554361 100644 --- a/fs/btrfs/hash.h +++ b/fs/btrfs/hash.h @@ -22,6 +22,7 @@ int __init btrfs_hash_init(void); void btrfs_hash_exit(void); +const char* btrfs_crc32c_impl(void); u32 btrfs_crc32c(u32 crc, const void *address, unsigned int length); diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index 4e59a91a11e0..4397a303fc6c 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -2303,7 +2303,7 @@ static void btrfs_interface_exit(void) static void btrfs_print_mod_info(void) { - printk(KERN_INFO "Btrfs loaded" + printk(KERN_INFO "Btrfs loaded, crc32c=%s" #ifdef CONFIG_BTRFS_DEBUG ", debug=on" #endif @@ -2313,7 +2313,8 @@ static void btrfs_print_mod_info(void) #ifdef CONFIG_BTRFS_FS_CHECK_INTEGRITY ", integrity-checker=on" #endif - "\n"); + "\n", + btrfs_crc32c_impl()); } static int btrfs_run_sanity_tests(void) -- cgit v1.3-8-gc7d7 From b581755b1c565391c72d03b157ba2dd0b18e9d15 Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Mon, 6 Jun 2016 16:21:37 +0200 Subject: ovl: xattr filter fix a) ovl_need_xattr_filter() is wrong, we can have multiple lower layers overlaid, all of which (except the lowest one) honouring the "trusted.overlay.opaque" xattr. So need to filter everything except the bottom and the pure-upper layer. b) we no longer can assume that inode is attached to dentry in get/setxattr. This patch unconditionally filters private xattrs to fix both of the above. Performance impact for get/removexattrs is likely in the noise. For listxattrs it might be measurable in pathological cases, but I very much hope nobody cares. If they do, we'll fix it then. Reported-by: Vivek Goyal Signed-off-by: Miklos Szeredi Fixes: b96809173e94 ("security_d_instantiate(): move to the point prior to attaching dentry to inode") --- fs/overlayfs/inode.c | 26 ++++++-------------------- 1 file changed, 6 insertions(+), 20 deletions(-) (limited to 'fs') diff --git a/fs/overlayfs/inode.c b/fs/overlayfs/inode.c index 0ed7c4012437..1dbeab6cf96e 100644 --- a/fs/overlayfs/inode.c +++ b/fs/overlayfs/inode.c @@ -238,41 +238,27 @@ out: return err; } -static bool ovl_need_xattr_filter(struct dentry *dentry, - enum ovl_path_type type) -{ - if ((type & (__OVL_PATH_PURE | __OVL_PATH_UPPER)) == __OVL_PATH_UPPER) - return S_ISDIR(dentry->d_inode->i_mode); - else - return false; -} - ssize_t ovl_getxattr(struct dentry *dentry, struct inode *inode, const char *name, void *value, size_t size) { - struct path realpath; - enum ovl_path_type type = ovl_path_real(dentry, &realpath); + struct dentry *realdentry = ovl_dentry_real(dentry); - if (ovl_need_xattr_filter(dentry, type) && ovl_is_private_xattr(name)) + if (ovl_is_private_xattr(name)) return -ENODATA; - return vfs_getxattr(realpath.dentry, name, value, size); + return vfs_getxattr(realdentry, name, value, size); } ssize_t ovl_listxattr(struct dentry *dentry, char *list, size_t size) { - struct path realpath; - enum ovl_path_type type = ovl_path_real(dentry, &realpath); + struct dentry *realdentry = ovl_dentry_real(dentry); ssize_t res; int off; - res = vfs_listxattr(realpath.dentry, list, size); + res = vfs_listxattr(realdentry, list, size); if (res <= 0 || size == 0) return res; - if (!ovl_need_xattr_filter(dentry, type)) - return res; - /* filter out private xattrs */ for (off = 0; off < res;) { char *s = list + off; @@ -302,7 +288,7 @@ int ovl_removexattr(struct dentry *dentry, const char *name) goto out; err = -ENODATA; - if (ovl_need_xattr_filter(dentry, type) && ovl_is_private_xattr(name)) + if (ovl_is_private_xattr(name)) goto out_drop_write; if (!OVL_TYPE_UPPER(type)) { -- cgit v1.3-8-gc7d7 From ed9e4afdb0551e3ef4ee8433fe664433a20ef73a Mon Sep 17 00:00:00 2001 From: Feifei Xu Date: Wed, 1 Jun 2016 19:18:26 +0800 Subject: Btrfs: self-tests: Execute page straddling test only when nodesize < PAGE_SIZE On ppc64, PAGE_SIZE is 64k which is same as BTRFS_MAX_METADATA_BLOCKSIZE. In such a scenario, we will never be able to have an extent buffer containing more than one page. Hence in such cases this commit does not execute the page straddling tests. Reviewed-by: Josef Bacik Signed-off-by: Feifei Xu Signed-off-by: Chandan Rajendra Signed-off-by: David Sterba --- fs/btrfs/tests/extent-io-tests.c | 49 ++++++++++++++++++++++++---------------- 1 file changed, 30 insertions(+), 19 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/tests/extent-io-tests.c b/fs/btrfs/tests/extent-io-tests.c index 755fceb2e5f5..2794fed71fa4 100644 --- a/fs/btrfs/tests/extent-io-tests.c +++ b/fs/btrfs/tests/extent-io-tests.c @@ -21,6 +21,7 @@ #include #include #include "btrfs-tests.h" +#include "../ctree.h" #include "../extent_io.h" #define PROCESS_UNLOCK (1 << 0) @@ -298,25 +299,29 @@ static int __test_eb_bitmaps(unsigned long *bitmap, struct extent_buffer *eb, return -EINVAL; } - bitmap_set(bitmap, (PAGE_SIZE - sizeof(long) / 2) * BITS_PER_BYTE, - sizeof(long) * BITS_PER_BYTE); - extent_buffer_bitmap_set(eb, PAGE_SIZE - sizeof(long) / 2, 0, - sizeof(long) * BITS_PER_BYTE); - if (memcmp_extent_buffer(eb, bitmap, 0, len) != 0) { - test_msg("Setting straddling pages failed\n"); - return -EINVAL; - } + /* Straddling pages test */ + if (len > PAGE_SIZE) { + bitmap_set(bitmap, + (PAGE_SIZE - sizeof(long) / 2) * BITS_PER_BYTE, + sizeof(long) * BITS_PER_BYTE); + extent_buffer_bitmap_set(eb, PAGE_SIZE - sizeof(long) / 2, 0, + sizeof(long) * BITS_PER_BYTE); + if (memcmp_extent_buffer(eb, bitmap, 0, len) != 0) { + test_msg("Setting straddling pages failed\n"); + return -EINVAL; + } - bitmap_set(bitmap, 0, len * BITS_PER_BYTE); - bitmap_clear(bitmap, - (PAGE_SIZE - sizeof(long) / 2) * BITS_PER_BYTE, - sizeof(long) * BITS_PER_BYTE); - extent_buffer_bitmap_set(eb, 0, 0, len * BITS_PER_BYTE); - extent_buffer_bitmap_clear(eb, PAGE_SIZE - sizeof(long) / 2, 0, - sizeof(long) * BITS_PER_BYTE); - if (memcmp_extent_buffer(eb, bitmap, 0, len) != 0) { - test_msg("Clearing straddling pages failed\n"); - return -EINVAL; + bitmap_set(bitmap, 0, len * BITS_PER_BYTE); + bitmap_clear(bitmap, + (PAGE_SIZE - sizeof(long) / 2) * BITS_PER_BYTE, + sizeof(long) * BITS_PER_BYTE); + extent_buffer_bitmap_set(eb, 0, 0, len * BITS_PER_BYTE); + extent_buffer_bitmap_clear(eb, PAGE_SIZE - sizeof(long) / 2, 0, + sizeof(long) * BITS_PER_BYTE); + if (memcmp_extent_buffer(eb, bitmap, 0, len) != 0) { + test_msg("Clearing straddling pages failed\n"); + return -EINVAL; + } } /* @@ -359,7 +364,13 @@ static int test_eb_bitmaps(u32 sectorsize, u32 nodesize) int ret; test_msg("Running extent buffer bitmap tests\n"); - len = sectorsize * 4; + + /* + * In ppc64, sectorsize can be 64K, thus 4 * 64K will be larger than + * BTRFS_MAX_METADATA_BLOCKSIZE. + */ + len = (sectorsize < BTRFS_MAX_METADATA_BLOCKSIZE) + ? sectorsize * 4 : sectorsize; bitmap = kmalloc(len, GFP_KERNEL); if (!bitmap) { -- cgit v1.3-8-gc7d7 From d94f43b4c6b88b9393422868d8dfde14c5923858 Mon Sep 17 00:00:00 2001 From: Feifei Xu Date: Wed, 1 Jun 2016 19:18:27 +0800 Subject: Btrfs: self-tests: Support testing all possible sectorsizes and nodesizes To test all possible sectorsizes, this commit adds a sectorsize array. This commit executes the tests for all possible sectorsizes and nodesizes. Reviewed-by: Josef Bacik Signed-off-by: Chandan Rajendra Signed-off-by: Feifei Xu Signed-off-by: David Sterba --- fs/btrfs/super.c | 54 ++++++++++++++++++++++++++++++++---------------------- 1 file changed, 32 insertions(+), 22 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index 5b0b354ca594..c49d7ae69617 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -2318,32 +2318,42 @@ static void btrfs_print_mod_info(void) static int btrfs_run_sanity_tests(void) { - int ret; + int ret, i; u32 sectorsize, nodesize; - - sectorsize = PAGE_SIZE; - nodesize = PAGE_SIZE; + u32 test_sectorsize[] = { + PAGE_SIZE, + }; ret = btrfs_init_test_fs(); if (ret) return ret; - - ret = btrfs_test_free_space_cache(sectorsize, nodesize); - if (ret) - goto out; - ret = btrfs_test_extent_buffer_operations(sectorsize, - nodesize); - if (ret) - goto out; - ret = btrfs_test_extent_io(sectorsize, nodesize); - if (ret) - goto out; - ret = btrfs_test_inodes(sectorsize, nodesize); - if (ret) - goto out; - ret = btrfs_test_qgroups(sectorsize, nodesize); - if (ret) - goto out; - ret = btrfs_test_free_space_tree(sectorsize, nodesize); + for (i = 0; i < ARRAY_SIZE(test_sectorsize); i++) { + sectorsize = test_sectorsize[i]; + for (nodesize = sectorsize; + nodesize <= BTRFS_MAX_METADATA_BLOCKSIZE; + nodesize <<= 1) { + pr_info("BTRFS: selftest: sectorsize: %u nodesize: %u\n", + sectorsize, nodesize); + ret = btrfs_test_free_space_cache(sectorsize, nodesize); + if (ret) + goto out; + ret = btrfs_test_extent_buffer_operations(sectorsize, + nodesize); + if (ret) + goto out; + ret = btrfs_test_extent_io(sectorsize, nodesize); + if (ret) + goto out; + ret = btrfs_test_inodes(sectorsize, nodesize); + if (ret) + goto out; + ret = btrfs_test_qgroups(sectorsize, nodesize); + if (ret) + goto out; + ret = btrfs_test_free_space_tree(sectorsize, nodesize); + if (ret) + goto out; + } + } out: btrfs_destroy_test_fs(); return ret; -- cgit v1.3-8-gc7d7 From ef9f2db365c31433e52b0c5863793273bb632666 Mon Sep 17 00:00:00 2001 From: Feifei Xu Date: Wed, 1 Jun 2016 19:18:28 +0800 Subject: Btrfs: self-tests: Use macros instead of constants and add missing newline This commit replaces numerical constants with appropriate preprocessor macros. Reviewed-by: Josef Bacik Signed-off-by: Chandan Rajendra Signed-off-by: Feifei Xu Signed-off-by: David Sterba --- fs/btrfs/disk-io.c | 3 ++- fs/btrfs/tests/extent-buffer-tests.c | 2 +- fs/btrfs/tests/qgroup-tests.c | 44 +++++++++++++++++++++++------------- 3 files changed, 31 insertions(+), 18 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 06ef433552b3..c3764ddb2a46 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1323,7 +1323,8 @@ struct btrfs_root *btrfs_alloc_dummy_root(u32 sectorsize, u32 nodesize) if (!root) return ERR_PTR(-ENOMEM); /* We don't use the stripesize in selftest, set it as sectorsize */ - __setup_root(nodesize, sectorsize, sectorsize, root, NULL, 1); + __setup_root(nodesize, sectorsize, sectorsize, root, NULL, + BTRFS_ROOT_TREE_OBJECTID); set_bit(BTRFS_ROOT_DUMMY_ROOT, &root->state); root->alloc_bytenr = 0; diff --git a/fs/btrfs/tests/extent-buffer-tests.c b/fs/btrfs/tests/extent-buffer-tests.c index 17b110f39f06..4f8cbd1ec5ee 100644 --- a/fs/btrfs/tests/extent-buffer-tests.c +++ b/fs/btrfs/tests/extent-buffer-tests.c @@ -225,6 +225,6 @@ out: int btrfs_test_extent_buffer_operations(u32 sectorsize, u32 nodesize) { - test_msg("Running extent buffer operation tests"); + test_msg("Running extent buffer operation tests\n"); return test_btrfs_split_item(sectorsize, nodesize); } diff --git a/fs/btrfs/tests/qgroup-tests.c b/fs/btrfs/tests/qgroup-tests.c index 9c1d27352f19..57a12c0d680b 100644 --- a/fs/btrfs/tests/qgroup-tests.c +++ b/fs/btrfs/tests/qgroup-tests.c @@ -229,7 +229,7 @@ static int test_no_shared_qgroup(struct btrfs_root *root, btrfs_init_dummy_trans(&trans); test_msg("Qgroup basic add\n"); - ret = btrfs_create_qgroup(NULL, fs_info, 5); + ret = btrfs_create_qgroup(NULL, fs_info, BTRFS_FS_TREE_OBJECTID); if (ret) { test_msg("Couldn't create a qgroup %d\n", ret); return ret; @@ -247,7 +247,8 @@ static int test_no_shared_qgroup(struct btrfs_root *root, return ret; } - ret = insert_normal_tree_ref(root, nodesize, nodesize, 0, 5); + ret = insert_normal_tree_ref(root, nodesize, nodesize, 0, + BTRFS_FS_TREE_OBJECTID); if (ret) return ret; @@ -266,7 +267,8 @@ static int test_no_shared_qgroup(struct btrfs_root *root, return ret; } - if (btrfs_verify_qgroup_counts(fs_info, 5, nodesize, nodesize)) { + if (btrfs_verify_qgroup_counts(fs_info, BTRFS_FS_TREE_OBJECTID, + nodesize, nodesize)) { test_msg("Qgroup counts didn't match expected values\n"); return -EINVAL; } @@ -299,7 +301,7 @@ static int test_no_shared_qgroup(struct btrfs_root *root, return -EINVAL; } - if (btrfs_verify_qgroup_counts(fs_info, 5, 0, 0)) { + if (btrfs_verify_qgroup_counts(fs_info, BTRFS_FS_TREE_OBJECTID, 0, 0)) { test_msg("Qgroup counts didn't match expected values\n"); return -EINVAL; } @@ -325,8 +327,11 @@ static int test_multiple_refs(struct btrfs_root *root, test_msg("Qgroup multiple refs test\n"); - /* We have 5 created already from the previous test */ - ret = btrfs_create_qgroup(NULL, fs_info, 256); + /* + * We have BTRFS_FS_TREE_OBJECTID created already from the + * previous test. + */ + ret = btrfs_create_qgroup(NULL, fs_info, BTRFS_FIRST_FREE_OBJECTID); if (ret) { test_msg("Couldn't create a qgroup %d\n", ret); return ret; @@ -339,7 +344,8 @@ static int test_multiple_refs(struct btrfs_root *root, return ret; } - ret = insert_normal_tree_ref(root, nodesize, nodesize, 0, 5); + ret = insert_normal_tree_ref(root, nodesize, nodesize, 0, + BTRFS_FS_TREE_OBJECTID); if (ret) return ret; @@ -358,7 +364,7 @@ static int test_multiple_refs(struct btrfs_root *root, return ret; } - if (btrfs_verify_qgroup_counts(fs_info, 5, + if (btrfs_verify_qgroup_counts(fs_info, BTRFS_FS_TREE_OBJECTID, nodesize, nodesize)) { test_msg("Qgroup counts didn't match expected values\n"); return -EINVAL; @@ -371,7 +377,8 @@ static int test_multiple_refs(struct btrfs_root *root, return ret; } - ret = add_tree_ref(root, nodesize, nodesize, 0, 256); + ret = add_tree_ref(root, nodesize, nodesize, 0, + BTRFS_FIRST_FREE_OBJECTID); if (ret) return ret; @@ -390,12 +397,14 @@ static int test_multiple_refs(struct btrfs_root *root, return ret; } - if (btrfs_verify_qgroup_counts(fs_info, 5, nodesize, 0)) { + if (btrfs_verify_qgroup_counts(fs_info, BTRFS_FS_TREE_OBJECTID, + nodesize, 0)) { test_msg("Qgroup counts didn't match expected values\n"); return -EINVAL; } - if (btrfs_verify_qgroup_counts(fs_info, 256, nodesize, 0)) { + if (btrfs_verify_qgroup_counts(fs_info, BTRFS_FIRST_FREE_OBJECTID, + nodesize, 0)) { test_msg("Qgroup counts didn't match expected values\n"); return -EINVAL; } @@ -407,7 +416,8 @@ static int test_multiple_refs(struct btrfs_root *root, return ret; } - ret = remove_extent_ref(root, nodesize, nodesize, 0, 256); + ret = remove_extent_ref(root, nodesize, nodesize, 0, + BTRFS_FIRST_FREE_OBJECTID); if (ret) return ret; @@ -426,12 +436,14 @@ static int test_multiple_refs(struct btrfs_root *root, return ret; } - if (btrfs_verify_qgroup_counts(fs_info, 256, 0, 0)) { + if (btrfs_verify_qgroup_counts(fs_info, BTRFS_FIRST_FREE_OBJECTID, + 0, 0)) { test_msg("Qgroup counts didn't match expected values\n"); return -EINVAL; } - if (btrfs_verify_qgroup_counts(fs_info, 5, nodesize, nodesize)) { + if (btrfs_verify_qgroup_counts(fs_info, BTRFS_FS_TREE_OBJECTID, + nodesize, nodesize)) { test_msg("Qgroup counts didn't match expected values\n"); return -EINVAL; } @@ -490,7 +502,7 @@ int btrfs_test_qgroups(u32 sectorsize, u32 nodesize) goto out; } - tmp_root->root_key.objectid = 5; + tmp_root->root_key.objectid = BTRFS_FS_TREE_OBJECTID; root->fs_info->fs_root = tmp_root; ret = btrfs_insert_fs_root(root->fs_info, tmp_root); if (ret) { @@ -505,7 +517,7 @@ int btrfs_test_qgroups(u32 sectorsize, u32 nodesize) goto out; } - tmp_root->root_key.objectid = 256; + tmp_root->root_key.objectid = BTRFS_FIRST_FREE_OBJECTID; ret = btrfs_insert_fs_root(root->fs_info, tmp_root); if (ret) { test_msg("Couldn't insert fs root %d\n", ret); -- cgit v1.3-8-gc7d7 From 36b3dc05b4650e81eca7d60d548a92b014595eb1 Mon Sep 17 00:00:00 2001 From: Feifei Xu Date: Wed, 1 Jun 2016 19:18:29 +0800 Subject: Btrfs: self-tests: Fix test_bitmaps fail on 64k sectorsize With 64K sectorsize, 1G sized block group cannot span across bitmaps. To execute test_bitmaps() function, this commit allocates "BITS_PER_BITMAP * sectorsize + PAGE_SIZE" sized block group. Reviewed-by: Josef Bacik Reviewed-by: Chandan Rajendra Signed-off-by: Feifei Xu Signed-off-by: David Sterba --- fs/btrfs/tests/free-space-tests.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/tests/free-space-tests.c b/fs/btrfs/tests/free-space-tests.c index 029343b3e854..3956bb2ff84c 100644 --- a/fs/btrfs/tests/free-space-tests.c +++ b/fs/btrfs/tests/free-space-tests.c @@ -843,7 +843,13 @@ int btrfs_test_free_space_cache(u32 sectorsize, u32 nodesize) test_msg("Running btrfs free space cache tests\n"); - cache = btrfs_alloc_dummy_block_group(1024 * 1024 * 1024, sectorsize); + /* + * For ppc64 (with 64k page size), bytes per bitmap might be + * larger than 1G. To make bitmap test available in ppc64, + * alloc dummy block group whose size cross bitmaps. + */ + cache = btrfs_alloc_dummy_block_group(BITS_PER_BITMAP * sectorsize + + PAGE_SIZE, sectorsize); if (!cache) { test_msg("Couldn't run the tests\n"); return 0; -- cgit v1.3-8-gc7d7 From 34b3e6c92af1fa3f7067e4fa05ffa9d8bd41c96c Mon Sep 17 00:00:00 2001 From: Feifei Xu Date: Wed, 1 Jun 2016 19:18:30 +0800 Subject: Btrfs: self-tests: Fix extent buffer bitmap test fail on BE system In __test_eb_bitmaps(), we write random data to a bitmap. Then copy the bitmap to another bitmap that resides inside an extent buffer. Later we verify the values of corresponding bits in the bitmap and the bitmap inside the extent buffer. However, extent_buffer_test_bit() reads in byte granularity while test_bit() reads in unsigned long granularity. Hence we end up comparing wrong bits on big-endian systems such as ppc64. This commit fixes the issue by reading the bitmap in byte granularity. Reviewed-by: Josef Bacik Reviewed-by: Chandan Rajendra Signed-off-by: Feifei Xu Signed-off-by: David Sterba --- fs/btrfs/tests/extent-io-tests.c | 12 +++++++++++- 1 file changed, 11 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/tests/extent-io-tests.c b/fs/btrfs/tests/extent-io-tests.c index 2794fed71fa4..d19ab0317283 100644 --- a/fs/btrfs/tests/extent-io-tests.c +++ b/fs/btrfs/tests/extent-io-tests.c @@ -273,6 +273,16 @@ out: return ret; } +/** + * test_bit_in_byte - Determine whether a bit is set in a byte + * @nr: bit number to test + * @addr: Address to start counting from + */ +static inline int test_bit_in_byte(int nr, const u8 *addr) +{ + return 1UL & (addr[nr / BITS_PER_BYTE] >> (nr & (BITS_PER_BYTE - 1))); +} + static int __test_eb_bitmaps(unsigned long *bitmap, struct extent_buffer *eb, unsigned long len) { @@ -338,7 +348,7 @@ static int __test_eb_bitmaps(unsigned long *bitmap, struct extent_buffer *eb, for (i = 0; i < len * BITS_PER_BYTE; i++) { int bit, bit1; - bit = !!test_bit(i, bitmap); + bit = !!test_bit_in_byte(i, (u8 *)bitmap); bit1 = !!extent_buffer_test_bit(eb, 0, i); if (bit1 != bit) { test_msg("Testing bit pattern failed\n"); -- cgit v1.3-8-gc7d7 From 97c1df3e54e811aed484a036a798b4b25d002ecf Mon Sep 17 00:00:00 2001 From: "Eric W. Biederman" Date: Mon, 6 Jun 2016 15:36:07 -0500 Subject: mnt: If fs_fully_visible fails call put_filesystem. Add this trivial missing error handling. Cc: stable@vger.kernel.org Fixes: 1b852bceb0d1 ("mnt: Refactor the logic for mounting sysfs and proc in a user namespace") Signed-off-by: "Eric W. Biederman" --- fs/namespace.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/namespace.c b/fs/namespace.c index 4fb1691b4355..9d45c8a3414f 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -2409,8 +2409,10 @@ static int do_new_mount(struct path *path, const char *fstype, int flags, mnt_flags |= MNT_NODEV | MNT_LOCK_NODEV; } if (type->fs_flags & FS_USERNS_VISIBLE) { - if (!fs_fully_visible(type, &mnt_flags)) + if (!fs_fully_visible(type, &mnt_flags)) { + put_filesystem(type); return -EPERM; + } } } -- cgit v1.3-8-gc7d7 From d71ed6c930ac7d8f88f3cef6624a7e826392d61f Mon Sep 17 00:00:00 2001 From: "Eric W. Biederman" Date: Fri, 27 May 2016 14:50:05 -0500 Subject: mnt: fs_fully_visible test the proper mount for MNT_LOCKED MNT_LOCKED implies on a child mount implies the child is locked to the parent. So while looping through the children the children should be tested (not their parent). Typically an unshare of a mount namespace locks all mounts together making both the parent and the slave as locked but there are a few corner cases where other things work. Cc: stable@vger.kernel.org Fixes: ceeb0e5d39fc ("vfs: Ignore unlocked mounts in fs_fully_visible") Reported-by: Seth Forshee Signed-off-by: "Eric W. Biederman" --- fs/namespace.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/namespace.c b/fs/namespace.c index 9d45c8a3414f..a7ec92c051f5 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -3273,7 +3273,7 @@ static bool fs_fully_visible(struct file_system_type *type, int *new_mnt_flags) list_for_each_entry(child, &mnt->mnt_mounts, mnt_child) { struct inode *inode = child->mnt_mountpoint->d_inode; /* Only worry about locked mounts */ - if (!(mnt_flags & MNT_LOCKED)) + if (!(child->mnt.mnt_flags & MNT_LOCKED)) continue; /* Is the directory permanetly empty? */ if (!is_empty_dir_inode(inode)) -- cgit v1.3-8-gc7d7 From 3d56c25e3bb0726a5c5e16fc2d9e38f8ed763085 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 7 Jun 2016 21:26:55 -0400 Subject: fix d_walk()/non-delayed __d_free() race Ascend-to-parent logics in d_walk() depends on all encountered child dentries not getting freed without an RCU delay. Unfortunately, in quite a few cases it is not true, with hard-to-hit oopsable race as the result. Fortunately, the fix is simiple; right now the rule is "if it ever been hashed, freeing must be delayed" and changing it to "if it ever had a parent, freeing must be delayed" closes that hole and covers all cases the old rule used to cover. Moreover, pipes and sockets remain _not_ covered, so we do not introduce RCU delay in the cases which are the reason for having that delay conditional in the first place. Cc: stable@vger.kernel.org # v3.2+ (and watch out for __d_materialise_dentry()) Signed-off-by: Al Viro --- fs/dcache.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/dcache.c b/fs/dcache.c index ad4a542e9bab..817c243c1ff1 100644 --- a/fs/dcache.c +++ b/fs/dcache.c @@ -1636,7 +1636,7 @@ struct dentry *d_alloc(struct dentry * parent, const struct qstr *name) struct dentry *dentry = __d_alloc(parent->d_sb, name); if (!dentry) return NULL; - + dentry->d_flags |= DCACHE_RCUACCESS; spin_lock(&parent->d_lock); /* * don't need child lock because it is not subject @@ -2358,7 +2358,6 @@ static void __d_rehash(struct dentry * entry, struct hlist_bl_head *b) { BUG_ON(!d_unhashed(entry)); hlist_bl_lock(b); - entry->d_flags |= DCACHE_RCUACCESS; hlist_bl_add_head_rcu(&entry->d_hash, b); hlist_bl_unlock(b); } @@ -2843,6 +2842,7 @@ static void __d_move(struct dentry *dentry, struct dentry *target, /* ... and switch them in the tree */ if (IS_ROOT(dentry)) { /* splicing a tree */ + dentry->d_flags |= DCACHE_RCUACCESS; dentry->d_parent = target->d_parent; target->d_parent = target; list_del_init(&target->d_child); -- cgit v1.3-8-gc7d7 From a01e718f7241c53f564402f7acff373eed5bd166 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 7 Jun 2016 21:53:51 -0400 Subject: fix a regression in atomic_open() open("/foo/no_such_file", O_RDONLY | O_CREAT) on should fail with EACCES when /foo is not writable; failing with ENOENT is obviously wrong. That got broken by a braino introduced when moving the creat_error logics from atomic_open() to lookup_open(). Easy to fix, fortunately. Spotted-by: "Yan, Zheng" Tested-by: "Yan, Zheng" Signed-off-by: Al Viro --- fs/namei.c | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/namei.c b/fs/namei.c index d7c0cac56d89..28cb1cd8507c 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -2995,9 +2995,13 @@ static int atomic_open(struct nameidata *nd, struct dentry *dentry, } if (*opened & FILE_CREATED) fsnotify_create(dir, dentry); - path->dentry = dentry; - path->mnt = nd->path.mnt; - return 1; + if (unlikely(d_is_negative(dentry))) { + error = -ENOENT; + } else { + path->dentry = dentry; + path->mnt = nd->path.mnt; + return 1; + } } } dput(dentry); -- cgit v1.3-8-gc7d7 From 1607f09c226d1378439c411baaaa020042750338 Mon Sep 17 00:00:00 2001 From: Mateusz Guzik Date: Sun, 5 Jun 2016 23:14:14 +0200 Subject: coredump: fix dumping through pipes MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The offset in the core file used to be tracked with ->written field of the coredump_params structure. The field was retired in favour of file->f_pos. However, ->f_pos is not maintained for pipes which leads to breakage. Restore explicit tracking of the offset in coredump_params. Introduce ->pos field for this purpose since ->written was already reused. Fixes: a00839395103 ("get rid of coredump_params->written"). Reported-by: Zbigniew Jędrzejewski-Szmek Signed-off-by: Mateusz Guzik Reviewed-by: Omar Sandoval Signed-off-by: Al Viro --- arch/powerpc/platforms/cell/spufs/coredump.c | 2 +- fs/binfmt_elf.c | 2 +- fs/binfmt_elf_fdpic.c | 2 +- fs/coredump.c | 4 +++- include/linux/binfmts.h | 1 + 5 files changed, 7 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/arch/powerpc/platforms/cell/spufs/coredump.c b/arch/powerpc/platforms/cell/spufs/coredump.c index 84fb984f29c1..85c85eb3e245 100644 --- a/arch/powerpc/platforms/cell/spufs/coredump.c +++ b/arch/powerpc/platforms/cell/spufs/coredump.c @@ -172,7 +172,7 @@ static int spufs_arch_write_note(struct spu_context *ctx, int i, if (rc < 0) goto out; - skip = roundup(cprm->file->f_pos - total + sz, 4) - cprm->file->f_pos; + skip = roundup(cprm->pos - total + sz, 4) - cprm->pos; if (!dump_skip(cprm, skip)) goto Eio; out: diff --git a/fs/binfmt_elf.c b/fs/binfmt_elf.c index e158b22ef32f..a7a28110dc80 100644 --- a/fs/binfmt_elf.c +++ b/fs/binfmt_elf.c @@ -2275,7 +2275,7 @@ static int elf_core_dump(struct coredump_params *cprm) goto end_coredump; /* Align to page */ - if (!dump_skip(cprm, dataoff - cprm->file->f_pos)) + if (!dump_skip(cprm, dataoff - cprm->pos)) goto end_coredump; for (i = 0, vma = first_vma(current, gate_vma); vma != NULL; diff --git a/fs/binfmt_elf_fdpic.c b/fs/binfmt_elf_fdpic.c index 71ade0e556b7..203589311bf8 100644 --- a/fs/binfmt_elf_fdpic.c +++ b/fs/binfmt_elf_fdpic.c @@ -1787,7 +1787,7 @@ static int elf_fdpic_core_dump(struct coredump_params *cprm) goto end_coredump; } - if (!dump_skip(cprm, dataoff - cprm->file->f_pos)) + if (!dump_skip(cprm, dataoff - cprm->pos)) goto end_coredump; if (!elf_fdpic_dump_segments(cprm)) diff --git a/fs/coredump.c b/fs/coredump.c index 38a7ab87e10a..281b768000e6 100644 --- a/fs/coredump.c +++ b/fs/coredump.c @@ -794,6 +794,7 @@ int dump_emit(struct coredump_params *cprm, const void *addr, int nr) return 0; file->f_pos = pos; cprm->written += n; + cprm->pos += n; nr -= n; } return 1; @@ -808,6 +809,7 @@ int dump_skip(struct coredump_params *cprm, size_t nr) if (dump_interrupted() || file->f_op->llseek(file, nr, SEEK_CUR) < 0) return 0; + cprm->pos += nr; return 1; } else { while (nr > PAGE_SIZE) { @@ -822,7 +824,7 @@ EXPORT_SYMBOL(dump_skip); int dump_align(struct coredump_params *cprm, int align) { - unsigned mod = cprm->file->f_pos & (align - 1); + unsigned mod = cprm->pos & (align - 1); if (align & (align - 1)) return 0; return mod ? dump_skip(cprm, align - mod) : 1; diff --git a/include/linux/binfmts.h b/include/linux/binfmts.h index 576e4639ca60..314b3caa701c 100644 --- a/include/linux/binfmts.h +++ b/include/linux/binfmts.h @@ -65,6 +65,7 @@ struct coredump_params { unsigned long limit; unsigned long mm_flags; loff_t written; + loff_t pos; }; /* -- cgit v1.3-8-gc7d7 From ba65dc5ef16f82fba77869cecf7a7d515f61446b Mon Sep 17 00:00:00 2001 From: Al Viro Date: Fri, 10 Jun 2016 11:32:47 -0400 Subject: much milder d_walk() race d_walk() relies upon the tree not getting rearranged under it without rename_lock being touched. And we do grab rename_lock around the places that change the tree topology. Unfortunately, branch reordering is just as bad from d_walk() POV and we have two places that do it without touching rename_lock - one in handling of cursors (for ramfs-style directories) and another in autofs. autofs one is a separate story; this commit deals with the cursors. * mark cursor dentries explicitly at allocation time * make __dentry_kill() leave ->d_child.next pointing to the next non-cursor sibling, making sure that it won't be moved around unnoticed before the parent is relocked on ascend-to-parent path in d_walk(). * make d_walk() skip cursors explicitly; strictly speaking it's not necessary (all callbacks we pass to d_walk() are no-ops on cursors), but it makes analysis easier. Signed-off-by: Al Viro --- fs/dcache.c | 58 ++++++++++++++++++++++++++++++++++++++++++++------ fs/internal.h | 1 + fs/libfs.c | 4 +--- include/linux/dcache.h | 1 + 4 files changed, 55 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/dcache.c b/fs/dcache.c index 817c243c1ff1..b7eddfd35aa5 100644 --- a/fs/dcache.c +++ b/fs/dcache.c @@ -507,6 +507,44 @@ void d_drop(struct dentry *dentry) } EXPORT_SYMBOL(d_drop); +static inline void dentry_unlist(struct dentry *dentry, struct dentry *parent) +{ + struct dentry *next; + /* + * Inform d_walk() and shrink_dentry_list() that we are no longer + * attached to the dentry tree + */ + dentry->d_flags |= DCACHE_DENTRY_KILLED; + if (unlikely(list_empty(&dentry->d_child))) + return; + __list_del_entry(&dentry->d_child); + /* + * Cursors can move around the list of children. While we'd been + * a normal list member, it didn't matter - ->d_child.next would've + * been updated. However, from now on it won't be and for the + * things like d_walk() it might end up with a nasty surprise. + * Normally d_walk() doesn't care about cursors moving around - + * ->d_lock on parent prevents that and since a cursor has no children + * of its own, we get through it without ever unlocking the parent. + * There is one exception, though - if we ascend from a child that + * gets killed as soon as we unlock it, the next sibling is found + * using the value left in its ->d_child.next. And if _that_ + * pointed to a cursor, and cursor got moved (e.g. by lseek()) + * before d_walk() regains parent->d_lock, we'll end up skipping + * everything the cursor had been moved past. + * + * Solution: make sure that the pointer left behind in ->d_child.next + * points to something that won't be moving around. I.e. skip the + * cursors. + */ + while (dentry->d_child.next != &parent->d_subdirs) { + next = list_entry(dentry->d_child.next, struct dentry, d_child); + if (likely(!(next->d_flags & DCACHE_DENTRY_CURSOR))) + break; + dentry->d_child.next = next->d_child.next; + } +} + static void __dentry_kill(struct dentry *dentry) { struct dentry *parent = NULL; @@ -532,12 +570,7 @@ static void __dentry_kill(struct dentry *dentry) } /* if it was on the hash then remove it */ __d_drop(dentry); - __list_del_entry(&dentry->d_child); - /* - * Inform d_walk() that we are no longer attached to the - * dentry tree - */ - dentry->d_flags |= DCACHE_DENTRY_KILLED; + dentry_unlist(dentry, parent); if (parent) spin_unlock(&parent->d_lock); dentry_iput(dentry); @@ -1203,6 +1236,9 @@ resume: struct dentry *dentry = list_entry(tmp, struct dentry, d_child); next = tmp->next; + if (unlikely(dentry->d_flags & DCACHE_DENTRY_CURSOR)) + continue; + spin_lock_nested(&dentry->d_lock, DENTRY_D_LOCK_NESTED); ret = enter(data, dentry); @@ -1651,6 +1687,16 @@ struct dentry *d_alloc(struct dentry * parent, const struct qstr *name) } EXPORT_SYMBOL(d_alloc); +struct dentry *d_alloc_cursor(struct dentry * parent) +{ + struct dentry *dentry = __d_alloc(parent->d_sb, NULL); + if (dentry) { + dentry->d_flags |= DCACHE_RCUACCESS | DCACHE_DENTRY_CURSOR; + dentry->d_parent = dget(parent); + } + return dentry; +} + /** * d_alloc_pseudo - allocate a dentry (for lookup-less filesystems) * @sb: the superblock diff --git a/fs/internal.h b/fs/internal.h index b71deeecea17..f57ced528cde 100644 --- a/fs/internal.h +++ b/fs/internal.h @@ -130,6 +130,7 @@ extern int invalidate_inodes(struct super_block *, bool); extern struct dentry *__d_alloc(struct super_block *, const struct qstr *); extern int d_set_mounted(struct dentry *dentry); extern long prune_dcache_sb(struct super_block *sb, struct shrink_control *sc); +extern struct dentry *d_alloc_cursor(struct dentry *); /* * read_write.c diff --git a/fs/libfs.c b/fs/libfs.c index 3db2721144c2..cedeacbae303 100644 --- a/fs/libfs.c +++ b/fs/libfs.c @@ -71,9 +71,7 @@ EXPORT_SYMBOL(simple_lookup); int dcache_dir_open(struct inode *inode, struct file *file) { - static struct qstr cursor_name = QSTR_INIT(".", 1); - - file->private_data = d_alloc(file->f_path.dentry, &cursor_name); + file->private_data = d_alloc_cursor(file->f_path.dentry); return file->private_data ? 0 : -ENOMEM; } diff --git a/include/linux/dcache.h b/include/linux/dcache.h index 484c8792da82..bcd0c64e3ed8 100644 --- a/include/linux/dcache.h +++ b/include/linux/dcache.h @@ -212,6 +212,7 @@ struct dentry_operations { #define DCACHE_OP_REAL 0x08000000 #define DCACHE_PAR_LOOKUP 0x10000000 /* being looked up (with parent locked shared) */ +#define DCACHE_DENTRY_CURSOR 0x20000000 extern seqlock_t rename_lock; -- cgit v1.3-8-gc7d7 From e54ad7f1ee263ffa5a2de9c609d58dfa27b21cd9 Mon Sep 17 00:00:00 2001 From: Jann Horn Date: Wed, 1 Jun 2016 11:55:05 +0200 Subject: proc: prevent stacking filesystems on top This prevents stacking filesystems (ecryptfs and overlayfs) from using procfs as lower filesystem. There is too much magic going on inside procfs, and there is no good reason to stack stuff on top of procfs. (For example, procfs does access checks in VFS open handlers, and ecryptfs by design calls open handlers from a kernel thread that doesn't drop privileges or so.) Signed-off-by: Jann Horn Cc: stable@vger.kernel.org Signed-off-by: Linus Torvalds --- fs/proc/root.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'fs') diff --git a/fs/proc/root.c b/fs/proc/root.c index 361ab4ee42fc..ec649c92d270 100644 --- a/fs/proc/root.c +++ b/fs/proc/root.c @@ -121,6 +121,13 @@ static struct dentry *proc_mount(struct file_system_type *fs_type, if (IS_ERR(sb)) return ERR_CAST(sb); + /* + * procfs isn't actually a stacking filesystem; however, there is + * too much magic going on inside it to permit stacking things on + * top of it + */ + sb->s_stack_depth = FILESYSTEM_MAX_STACK_DEPTH; + if (!proc_parse_options(options, ns)) { deactivate_locked_super(sb); return ERR_PTR(-EINVAL); -- cgit v1.3-8-gc7d7 From 2f36db71009304b3f0b95afacd8eba1f9f046b87 Mon Sep 17 00:00:00 2001 From: Jann Horn Date: Wed, 1 Jun 2016 11:55:06 +0200 Subject: ecryptfs: forbid opening files without mmap handler This prevents users from triggering a stack overflow through a recursive invocation of pagefault handling that involves mapping procfs files into virtual memory. Signed-off-by: Jann Horn Acked-by: Tyler Hicks Cc: stable@vger.kernel.org Signed-off-by: Linus Torvalds --- fs/ecryptfs/kthread.c | 13 +++++++++++-- 1 file changed, 11 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/ecryptfs/kthread.c b/fs/ecryptfs/kthread.c index 866bb18efefe..e818f5ac7a26 100644 --- a/fs/ecryptfs/kthread.c +++ b/fs/ecryptfs/kthread.c @@ -25,6 +25,7 @@ #include #include #include +#include #include "ecryptfs_kernel.h" struct ecryptfs_open_req { @@ -147,7 +148,7 @@ int ecryptfs_privileged_open(struct file **lower_file, flags |= IS_RDONLY(d_inode(lower_dentry)) ? O_RDONLY : O_RDWR; (*lower_file) = dentry_open(&req.path, flags, cred); if (!IS_ERR(*lower_file)) - goto out; + goto have_file; if ((flags & O_ACCMODE) == O_RDONLY) { rc = PTR_ERR((*lower_file)); goto out; @@ -165,8 +166,16 @@ int ecryptfs_privileged_open(struct file **lower_file, mutex_unlock(&ecryptfs_kthread_ctl.mux); wake_up(&ecryptfs_kthread_ctl.wait); wait_for_completion(&req.done); - if (IS_ERR(*lower_file)) + if (IS_ERR(*lower_file)) { rc = PTR_ERR(*lower_file); + goto out; + } +have_file: + if ((*lower_file)->f_op->mmap == NULL) { + fput(*lower_file); + *lower_file = NULL; + rc = -EMEDIUMTYPE; + } out: return rc; } -- cgit v1.3-8-gc7d7 From ea01a18494b3d7a91b2f1f2a6a5aaef4741bc294 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sun, 12 Jun 2016 11:24:46 -0400 Subject: autofs races * make autofs4_expire_indirect() skip the dentries being in process of expiry * do *not* mess with list_move(); making sure that dentry with AUTOFS_INF_EXPIRING are not picked for expiry is enough. * do not remove NO_RCU when we set EXPIRING, don't bother with smp_mb() there. Clear it at the same time we clear EXPIRING. Makes a bunch of tests simpler. * rename NO_RCU to WANT_EXPIRE, which is what it really is. Signed-off-by: Al Viro --- fs/autofs4/autofs_i.h | 8 ++++++-- fs/autofs4/expire.c | 27 ++++++++------------------- fs/autofs4/root.c | 2 +- 3 files changed, 15 insertions(+), 22 deletions(-) (limited to 'fs') diff --git a/fs/autofs4/autofs_i.h b/fs/autofs4/autofs_i.h index f0d268b97d19..a439548de785 100644 --- a/fs/autofs4/autofs_i.h +++ b/fs/autofs4/autofs_i.h @@ -70,9 +70,13 @@ struct autofs_info { }; #define AUTOFS_INF_EXPIRING (1<<0) /* dentry in the process of expiring */ -#define AUTOFS_INF_NO_RCU (1<<1) /* the dentry is being considered +#define AUTOFS_INF_WANT_EXPIRE (1<<1) /* the dentry is being considered * for expiry, so RCU_walk is - * not permitted + * not permitted. If it progresses to + * actual expiry attempt, the flag is + * not cleared when EXPIRING is set - + * in that case it gets cleared only + * when it comes to clearing EXPIRING. */ #define AUTOFS_INF_PENDING (1<<2) /* dentry pending mount */ diff --git a/fs/autofs4/expire.c b/fs/autofs4/expire.c index 9510d8d2e9cd..b493909e7492 100644 --- a/fs/autofs4/expire.c +++ b/fs/autofs4/expire.c @@ -316,19 +316,17 @@ struct dentry *autofs4_expire_direct(struct super_block *sb, if (ino->flags & AUTOFS_INF_PENDING) goto out; if (!autofs4_direct_busy(mnt, root, timeout, do_now)) { - ino->flags |= AUTOFS_INF_NO_RCU; + ino->flags |= AUTOFS_INF_WANT_EXPIRE; spin_unlock(&sbi->fs_lock); synchronize_rcu(); spin_lock(&sbi->fs_lock); if (!autofs4_direct_busy(mnt, root, timeout, do_now)) { ino->flags |= AUTOFS_INF_EXPIRING; - smp_mb(); - ino->flags &= ~AUTOFS_INF_NO_RCU; init_completion(&ino->expire_complete); spin_unlock(&sbi->fs_lock); return root; } - ino->flags &= ~AUTOFS_INF_NO_RCU; + ino->flags &= ~AUTOFS_INF_WANT_EXPIRE; } out: spin_unlock(&sbi->fs_lock); @@ -446,7 +444,7 @@ struct dentry *autofs4_expire_indirect(struct super_block *sb, while ((dentry = get_next_positive_subdir(dentry, root))) { spin_lock(&sbi->fs_lock); ino = autofs4_dentry_ino(dentry); - if (ino->flags & AUTOFS_INF_NO_RCU) + if (ino->flags & AUTOFS_INF_WANT_EXPIRE) expired = NULL; else expired = should_expire(dentry, mnt, timeout, how); @@ -455,7 +453,7 @@ struct dentry *autofs4_expire_indirect(struct super_block *sb, continue; } ino = autofs4_dentry_ino(expired); - ino->flags |= AUTOFS_INF_NO_RCU; + ino->flags |= AUTOFS_INF_WANT_EXPIRE; spin_unlock(&sbi->fs_lock); synchronize_rcu(); spin_lock(&sbi->fs_lock); @@ -465,7 +463,7 @@ struct dentry *autofs4_expire_indirect(struct super_block *sb, goto found; } - ino->flags &= ~AUTOFS_INF_NO_RCU; + ino->flags &= ~AUTOFS_INF_WANT_EXPIRE; if (expired != dentry) dput(expired); spin_unlock(&sbi->fs_lock); @@ -475,17 +473,8 @@ struct dentry *autofs4_expire_indirect(struct super_block *sb, found: pr_debug("returning %p %pd\n", expired, expired); ino->flags |= AUTOFS_INF_EXPIRING; - smp_mb(); - ino->flags &= ~AUTOFS_INF_NO_RCU; init_completion(&ino->expire_complete); spin_unlock(&sbi->fs_lock); - spin_lock(&sbi->lookup_lock); - spin_lock(&expired->d_parent->d_lock); - spin_lock_nested(&expired->d_lock, DENTRY_D_LOCK_NESTED); - list_move(&expired->d_parent->d_subdirs, &expired->d_child); - spin_unlock(&expired->d_lock); - spin_unlock(&expired->d_parent->d_lock); - spin_unlock(&sbi->lookup_lock); return expired; } @@ -496,7 +485,7 @@ int autofs4_expire_wait(struct dentry *dentry, int rcu_walk) int status; /* Block on any pending expire */ - if (!(ino->flags & (AUTOFS_INF_EXPIRING | AUTOFS_INF_NO_RCU))) + if (!(ino->flags & AUTOFS_INF_WANT_EXPIRE)) return 0; if (rcu_walk) return -ECHILD; @@ -554,7 +543,7 @@ int autofs4_expire_run(struct super_block *sb, ino = autofs4_dentry_ino(dentry); /* avoid rapid-fire expire attempts if expiry fails */ ino->last_used = now; - ino->flags &= ~AUTOFS_INF_EXPIRING; + ino->flags &= ~(AUTOFS_INF_EXPIRING|AUTOFS_INF_WANT_EXPIRE); complete_all(&ino->expire_complete); spin_unlock(&sbi->fs_lock); @@ -583,7 +572,7 @@ int autofs4_do_expire_multi(struct super_block *sb, struct vfsmount *mnt, spin_lock(&sbi->fs_lock); /* avoid rapid-fire expire attempts if expiry fails */ ino->last_used = now; - ino->flags &= ~AUTOFS_INF_EXPIRING; + ino->flags &= ~(AUTOFS_INF_EXPIRING|AUTOFS_INF_WANT_EXPIRE); complete_all(&ino->expire_complete); spin_unlock(&sbi->fs_lock); dput(dentry); diff --git a/fs/autofs4/root.c b/fs/autofs4/root.c index 78bd80298528..3767f6641af1 100644 --- a/fs/autofs4/root.c +++ b/fs/autofs4/root.c @@ -458,7 +458,7 @@ static int autofs4_d_manage(struct dentry *dentry, bool rcu_walk) */ struct inode *inode; - if (ino->flags & (AUTOFS_INF_EXPIRING | AUTOFS_INF_NO_RCU)) + if (ino->flags & AUTOFS_INF_WANT_EXPIRE) return 0; if (d_mountpoint(dentry)) return 0; -- cgit v1.3-8-gc7d7 From eee930163c6a2507aa3b41edf8f7e3b7a2049769 Mon Sep 17 00:00:00 2001 From: Geert Uytterhoeven Date: Fri, 25 Mar 2016 11:47:11 +0100 Subject: nfsd: Fix NFSD_MDS_PR_KEY on 32-bit by adding ULL postfix MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit On 32-bit: fs/nfsd/blocklayout.c: In function ‘nfsd4_block_get_device_info_scsi’: fs/nfsd/blocklayout.c:337: warning: integer constant is too large for ‘long’ type fs/nfsd/blocklayout.c:344: warning: integer constant is too large for ‘long’ type fs/nfsd/blocklayout.c: In function ‘nfsd4_scsi_fence_client’: fs/nfsd/blocklayout.c:385: warning: integer constant is too large for ‘long’ type Add the missing "ULL" postfix to 64-bit constant NFSD_MDS_PR_KEY to fix this. Fixes: f99d4fbdae6765d0 ("nfsd: add SCSI layout support") Signed-off-by: Geert Uytterhoeven Reviewed-by: Christoph Hellwig Signed-off-by: J. Bruce Fields --- fs/nfsd/blocklayout.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfsd/blocklayout.c b/fs/nfsd/blocklayout.c index e55b5242614d..31f3df193bdb 100644 --- a/fs/nfsd/blocklayout.c +++ b/fs/nfsd/blocklayout.c @@ -290,7 +290,7 @@ out_free_buf: return error; } -#define NFSD_MDS_PR_KEY 0x0100000000000000 +#define NFSD_MDS_PR_KEY 0x0100000000000000ULL /* * We use the client ID as a unique key for the reservations. -- cgit v1.3-8-gc7d7 From 695e9df010e40f407f4830dc11d53dce957710ba Mon Sep 17 00:00:00 2001 From: "Eric W. Biederman" Date: Fri, 10 Jun 2016 12:21:40 -0500 Subject: mnt: Account for MS_RDONLY in fs_fully_visible In rare cases it is possible for s_flags & MS_RDONLY to be set but MNT_READONLY to be clear. This starting combination can cause fs_fully_visible to fail to ensure that the new mount is readonly. Therefore force MNT_LOCK_READONLY in the new mount if MS_RDONLY is set on the source filesystem of the mount. In general both MS_RDONLY and MNT_READONLY are set at the same for mounts so I don't expect any programs to care. Nor do I expect MS_RDONLY to be set on proc or sysfs in the initial user namespace, which further decreases the likelyhood of problems. Which means this change should only affect system configurations by paranoid sysadmins who should welcome the additional protection as it keeps people from wriggling out of their policies. Cc: stable@vger.kernel.org Fixes: 8c6cf9cc829f ("mnt: Modify fs_fully_visible to deal with locked ro nodev and atime") Signed-off-by: "Eric W. Biederman" --- fs/namespace.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'fs') diff --git a/fs/namespace.c b/fs/namespace.c index a7ec92c051f5..783004af5707 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -3247,6 +3247,10 @@ static bool fs_fully_visible(struct file_system_type *type, int *new_mnt_flags) if (mnt->mnt.mnt_sb->s_iflags & SB_I_NOEXEC) mnt_flags &= ~(MNT_LOCK_NOSUID | MNT_LOCK_NOEXEC); + /* Don't miss readonly hidden in the superblock flags */ + if (mnt->mnt.mnt_sb->s_flags & MS_RDONLY) + mnt_flags |= MNT_LOCK_READONLY; + /* Verify the mount flags are equal to or more permissive * than the proposed new mount. */ -- cgit v1.3-8-gc7d7 From b10e3e90485e32e4cea9e35d2295ee7bffaeff73 Mon Sep 17 00:00:00 2001 From: Nicolai Stange Date: Tue, 24 May 2016 13:08:53 +0200 Subject: debugfs: full_proxy_open(): free proxy on ->open() failure Debugfs' full_proxy_open(), the ->open() installed at all inodes created through debugfs_create_file(), - grabs a reference to the original struct file_operations instance passed to debugfs_create_file(), - dynamically allocates a proxy struct file_operations instance wrapping the original - and installs this at the file's ->f_op. Afterwards, it calls the original ->open() and passes its return value back to the VFS layer. Now, if that return value indicates failure, the VFS layer won't ever call ->release() and thus, neither the reference to the original file_operations nor the memory for the proxy file_operations will get released, i.e. both are leaked. Upon failure of the original fops' ->open(), undo the proxy installation. That is: - Set the struct file ->f_op to what it had been when full_proxy_open() was entered. - Drop the reference to the original file_operations. - Free the memory holding the proxy file_operations. Fixes: 49d200deaa68 ("debugfs: prevent access to removed files' private data") Signed-off-by: Nicolai Stange Signed-off-by: Greg Kroah-Hartman --- fs/debugfs/file.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/debugfs/file.c b/fs/debugfs/file.c index 9c1c9a01b7e5..d1ec80331414 100644 --- a/fs/debugfs/file.c +++ b/fs/debugfs/file.c @@ -262,8 +262,10 @@ static int full_proxy_open(struct inode *inode, struct file *filp) if (real_fops->open) { r = real_fops->open(inode, filp); - - if (filp->f_op != proxy_fops) { + if (r) { + replace_fops(filp, d_inode(dentry)->i_fop); + goto free_proxy; + } else if (filp->f_op != proxy_fops) { /* No protection against file removal anymore. */ WARN(1, "debugfs file owner replaced proxy fops: %pd", dentry); -- cgit v1.3-8-gc7d7 From 75f0b68b75dabb3ff551440163fd67b3fc62901a Mon Sep 17 00:00:00 2001 From: Nicolai Stange Date: Tue, 24 May 2016 13:08:54 +0200 Subject: debugfs: open_proxy_open(): avoid double fops release Debugfs' open_proxy_open(), the ->open() installed at all inodes created through debugfs_create_file_unsafe(), - grabs a reference to the original file_operations instance passed to debugfs_create_file_unsafe() via fops_get(), - installs it at the file's ->f_op by means of replace_fops() - and calls fops_put() on it. Since the semantics of replace_fops() are such that the reference's ownership is transferred, the subsequent fops_put() will result in a double release when the file is eventually closed. Currently, this is not an issue since fops_put() basically does a module_put() on the file_operations' ->owner only and there don't exist any modules calling debugfs_create_file_unsafe() yet. This is expected to change in the future though, c.f. commit c64688081490 ("debugfs: add support for self-protecting attribute file fops"). Remove the call to fops_put() from open_proxy_open(). Fixes: 9fd4dcece43a ("debugfs: prevent access to possibly dead file_operations at file open") Signed-off-by: Nicolai Stange Signed-off-by: Greg Kroah-Hartman --- fs/debugfs/file.c | 1 - 1 file changed, 1 deletion(-) (limited to 'fs') diff --git a/fs/debugfs/file.c b/fs/debugfs/file.c index d1ec80331414..592059f88e04 100644 --- a/fs/debugfs/file.c +++ b/fs/debugfs/file.c @@ -127,7 +127,6 @@ static int open_proxy_open(struct inode *inode, struct file *filp) r = real_fops->open(inode, filp); out: - fops_put(real_fops); debugfs_use_file_finish(srcu_idx); return r; } -- cgit v1.3-8-gc7d7 From d0e13f5bbe4be7c8f27736fc40503dcec04b7de0 Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Wed, 15 Jun 2016 14:18:59 +0200 Subject: ovl: fix uid/gid when creating over whiteout Fix a regression when creating a file over a whiteout. The new file/directory needs to use the current fsuid/fsgid, not the ones from the mounter's credentials. The refcounting is a bit tricky: prepare_creds() sets an original refcount, override_creds() gets one more, which revert_cred() drops. So 1) we need to expicitly put the mounter's credentials when overriding with the updated one 2) we need to put the original ref to the updated creds (and this can safely be done before revert_creds(), since we'll still have the ref from override_creds()). Reported-by: Stephen Smalley Fixes: 3fe6e52f0626 ("ovl: override creds with the ones from the superblock mounter") Signed-off-by: Miklos Szeredi --- fs/overlayfs/dir.c | 13 +++++++++++-- 1 file changed, 11 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/overlayfs/dir.c b/fs/overlayfs/dir.c index 22f0253a3567..c2a6b0894022 100644 --- a/fs/overlayfs/dir.c +++ b/fs/overlayfs/dir.c @@ -405,12 +405,21 @@ static int ovl_create_or_link(struct dentry *dentry, int mode, dev_t rdev, err = ovl_create_upper(dentry, inode, &stat, link, hardlink); } else { const struct cred *old_cred; + struct cred *override_cred; old_cred = ovl_override_creds(dentry->d_sb); - err = ovl_create_over_whiteout(dentry, inode, &stat, link, - hardlink); + err = -ENOMEM; + override_cred = prepare_creds(); + if (override_cred) { + override_cred->fsuid = old_cred->fsuid; + override_cred->fsgid = old_cred->fsgid; + put_cred(override_creds(override_cred)); + put_cred(override_cred); + err = ovl_create_over_whiteout(dentry, inode, &stat, + link, hardlink); + } revert_creds(old_cred); } -- cgit v1.3-8-gc7d7 From d50039ea5ee63c589b0434baa5ecf6e5075bb6f9 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Mon, 16 May 2016 17:03:42 -0400 Subject: nfsd4/rpc: move backchannel create logic into rpc code Also simplify the logic a bit. Cc: stable@vger.kernel.org Signed-off-by: J. Bruce Fields Acked-by: Trond Myklebust --- fs/nfsd/nfs4callback.c | 18 +----------------- include/linux/sunrpc/clnt.h | 2 -- net/sunrpc/clnt.c | 12 ++++++++++-- 3 files changed, 11 insertions(+), 21 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c index 7389cb1d7409..04c68d900324 100644 --- a/fs/nfsd/nfs4callback.c +++ b/fs/nfsd/nfs4callback.c @@ -710,22 +710,6 @@ static struct rpc_cred *get_backchannel_cred(struct nfs4_client *clp, struct rpc } } -static struct rpc_clnt *create_backchannel_client(struct rpc_create_args *args) -{ - struct rpc_xprt *xprt; - - if (args->protocol != XPRT_TRANSPORT_BC_TCP) - return rpc_create(args); - - xprt = args->bc_xprt->xpt_bc_xprt; - if (xprt) { - xprt_get(xprt); - return rpc_create_xprt(args, xprt); - } - - return rpc_create(args); -} - static int setup_callback_client(struct nfs4_client *clp, struct nfs4_cb_conn *conn, struct nfsd4_session *ses) { int maxtime = max_cb_time(clp->net); @@ -768,7 +752,7 @@ static int setup_callback_client(struct nfs4_client *clp, struct nfs4_cb_conn *c args.authflavor = ses->se_cb_sec.flavor; } /* Create RPC client */ - client = create_backchannel_client(&args); + client = rpc_create(&args); if (IS_ERR(client)) { dprintk("NFSD: couldn't create callback client: %ld\n", PTR_ERR(client)); diff --git a/include/linux/sunrpc/clnt.h b/include/linux/sunrpc/clnt.h index 19c659d1c0f8..b6810c92b8bb 100644 --- a/include/linux/sunrpc/clnt.h +++ b/include/linux/sunrpc/clnt.h @@ -137,8 +137,6 @@ struct rpc_create_args { #define RPC_CLNT_CREATE_NO_RETRANS_TIMEOUT (1UL << 9) struct rpc_clnt *rpc_create(struct rpc_create_args *args); -struct rpc_clnt *rpc_create_xprt(struct rpc_create_args *args, - struct rpc_xprt *xprt); struct rpc_clnt *rpc_bind_new_program(struct rpc_clnt *, const struct rpc_program *, u32); struct rpc_clnt *rpc_clone_client(struct rpc_clnt *); diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c index 173c5dd2d751..b33721d41adc 100644 --- a/net/sunrpc/clnt.c +++ b/net/sunrpc/clnt.c @@ -446,7 +446,7 @@ out_no_rpciod: return ERR_PTR(err); } -struct rpc_clnt *rpc_create_xprt(struct rpc_create_args *args, +static struct rpc_clnt *rpc_create_xprt(struct rpc_create_args *args, struct rpc_xprt *xprt) { struct rpc_clnt *clnt = NULL; @@ -484,7 +484,6 @@ struct rpc_clnt *rpc_create_xprt(struct rpc_create_args *args, return clnt; } -EXPORT_SYMBOL_GPL(rpc_create_xprt); /** * rpc_create - create an RPC client and transport with one call @@ -510,6 +509,15 @@ struct rpc_clnt *rpc_create(struct rpc_create_args *args) }; char servername[48]; + if (args->bc_xprt) { + WARN_ON(args->protocol != XPRT_TRANSPORT_BC_TCP); + xprt = args->bc_xprt->xpt_bc_xprt; + if (xprt) { + xprt_get(xprt); + return rpc_create_xprt(args, xprt); + } + } + if (args->flags & RPC_CLNT_CREATE_INFINITE_SLOTS) xprtargs.flags |= XPRT_CREATE_INFINITE_SLOTS; if (args->flags & RPC_CLNT_CREATE_NO_IDLE_TIMEOUT) -- cgit v1.3-8-gc7d7 From feb9dad5209280085d5b0c094fa67e7a8d75c81a Mon Sep 17 00:00:00 2001 From: Oleg Drokin Date: Tue, 14 Jun 2016 23:28:04 -0400 Subject: nfsd: Always lock state exclusively. It used to be the case that state had an rwlock that was locked for write by downgrades, but for read for upgrades (opens). Well, the problem is if there are two competing opens for the same state, they step on each other toes potentially leading to leaking file descriptors from the state structure, since access mode is a bitmap only set once. Signed-off-by: Oleg Drokin Cc: stable@vger.kernel.org Signed-off-by: J. Bruce Fields --- fs/nfsd/nfs4state.c | 40 ++++++++++++++++++++-------------------- fs/nfsd/state.h | 2 +- 2 files changed, 21 insertions(+), 21 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index f5f82e145018..c927d36d8ac6 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -3502,7 +3502,7 @@ init_open_stateid(struct nfs4_ol_stateid *stp, struct nfs4_file *fp, stp->st_access_bmap = 0; stp->st_deny_bmap = 0; stp->st_openstp = NULL; - init_rwsem(&stp->st_rwsem); + mutex_init(&stp->st_mutex); list_add(&stp->st_perstateowner, &oo->oo_owner.so_stateids); list_add(&stp->st_perfile, &fp->fi_stateids); @@ -4335,10 +4335,10 @@ nfsd4_process_open2(struct svc_rqst *rqstp, struct svc_fh *current_fh, struct nf */ if (stp) { /* Stateid was found, this is an OPEN upgrade */ - down_read(&stp->st_rwsem); + mutex_lock(&stp->st_mutex); status = nfs4_upgrade_open(rqstp, fp, current_fh, stp, open); if (status) { - up_read(&stp->st_rwsem); + mutex_unlock(&stp->st_mutex); goto out; } } else { @@ -4348,19 +4348,19 @@ nfsd4_process_open2(struct svc_rqst *rqstp, struct svc_fh *current_fh, struct nf if (swapstp) { nfs4_put_stid(&stp->st_stid); stp = swapstp; - down_read(&stp->st_rwsem); + mutex_lock(&stp->st_mutex); status = nfs4_upgrade_open(rqstp, fp, current_fh, stp, open); if (status) { - up_read(&stp->st_rwsem); + mutex_unlock(&stp->st_mutex); goto out; } goto upgrade_out; } - down_read(&stp->st_rwsem); + mutex_lock(&stp->st_mutex); status = nfs4_get_vfs_file(rqstp, fp, current_fh, stp, open); if (status) { - up_read(&stp->st_rwsem); + mutex_unlock(&stp->st_mutex); release_open_stateid(stp); goto out; } @@ -4372,7 +4372,7 @@ nfsd4_process_open2(struct svc_rqst *rqstp, struct svc_fh *current_fh, struct nf } upgrade_out: nfs4_inc_and_copy_stateid(&open->op_stateid, &stp->st_stid); - up_read(&stp->st_rwsem); + mutex_unlock(&stp->st_mutex); if (nfsd4_has_session(&resp->cstate)) { if (open->op_deleg_want & NFS4_SHARE_WANT_NO_DELEG) { @@ -4977,12 +4977,12 @@ static __be32 nfs4_seqid_op_checks(struct nfsd4_compound_state *cstate, stateid_ * revoked delegations are kept only for free_stateid. */ return nfserr_bad_stateid; - down_write(&stp->st_rwsem); + mutex_lock(&stp->st_mutex); status = check_stateid_generation(stateid, &stp->st_stid.sc_stateid, nfsd4_has_session(cstate)); if (status == nfs_ok) status = nfs4_check_fh(current_fh, &stp->st_stid); if (status != nfs_ok) - up_write(&stp->st_rwsem); + mutex_unlock(&stp->st_mutex); return status; } @@ -5030,7 +5030,7 @@ static __be32 nfs4_preprocess_confirmed_seqid_op(struct nfsd4_compound_state *cs return status; oo = openowner(stp->st_stateowner); if (!(oo->oo_flags & NFS4_OO_CONFIRMED)) { - up_write(&stp->st_rwsem); + mutex_unlock(&stp->st_mutex); nfs4_put_stid(&stp->st_stid); return nfserr_bad_stateid; } @@ -5062,12 +5062,12 @@ nfsd4_open_confirm(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, oo = openowner(stp->st_stateowner); status = nfserr_bad_stateid; if (oo->oo_flags & NFS4_OO_CONFIRMED) { - up_write(&stp->st_rwsem); + mutex_unlock(&stp->st_mutex); goto put_stateid; } oo->oo_flags |= NFS4_OO_CONFIRMED; nfs4_inc_and_copy_stateid(&oc->oc_resp_stateid, &stp->st_stid); - up_write(&stp->st_rwsem); + mutex_unlock(&stp->st_mutex); dprintk("NFSD: %s: success, seqid=%d stateid=" STATEID_FMT "\n", __func__, oc->oc_seqid, STATEID_VAL(&stp->st_stid.sc_stateid)); @@ -5143,7 +5143,7 @@ nfsd4_open_downgrade(struct svc_rqst *rqstp, nfs4_inc_and_copy_stateid(&od->od_stateid, &stp->st_stid); status = nfs_ok; put_stateid: - up_write(&stp->st_rwsem); + mutex_unlock(&stp->st_mutex); nfs4_put_stid(&stp->st_stid); out: nfsd4_bump_seqid(cstate, status); @@ -5196,7 +5196,7 @@ nfsd4_close(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, if (status) goto out; nfs4_inc_and_copy_stateid(&close->cl_stateid, &stp->st_stid); - up_write(&stp->st_rwsem); + mutex_unlock(&stp->st_mutex); nfsd4_close_open_stateid(stp); @@ -5422,7 +5422,7 @@ init_lock_stateid(struct nfs4_ol_stateid *stp, struct nfs4_lockowner *lo, stp->st_access_bmap = 0; stp->st_deny_bmap = open_stp->st_deny_bmap; stp->st_openstp = open_stp; - init_rwsem(&stp->st_rwsem); + mutex_init(&stp->st_mutex); list_add(&stp->st_locks, &open_stp->st_locks); list_add(&stp->st_perstateowner, &lo->lo_owner.so_stateids); spin_lock(&fp->fi_lock); @@ -5591,7 +5591,7 @@ nfsd4_lock(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, &open_stp, nn); if (status) goto out; - up_write(&open_stp->st_rwsem); + mutex_unlock(&open_stp->st_mutex); open_sop = openowner(open_stp->st_stateowner); status = nfserr_bad_stateid; if (!same_clid(&open_sop->oo_owner.so_client->cl_clientid, @@ -5600,7 +5600,7 @@ nfsd4_lock(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, status = lookup_or_create_lock_state(cstate, open_stp, lock, &lock_stp, &new); if (status == nfs_ok) - down_write(&lock_stp->st_rwsem); + mutex_lock(&lock_stp->st_mutex); } else { status = nfs4_preprocess_seqid_op(cstate, lock->lk_old_lock_seqid, @@ -5704,7 +5704,7 @@ out: seqid_mutating_err(ntohl(status))) lock_sop->lo_owner.so_seqid++; - up_write(&lock_stp->st_rwsem); + mutex_unlock(&lock_stp->st_mutex); /* * If this is a new, never-before-used stateid, and we are @@ -5874,7 +5874,7 @@ nfsd4_locku(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, fput: fput(filp); put_stateid: - up_write(&stp->st_rwsem); + mutex_unlock(&stp->st_mutex); nfs4_put_stid(&stp->st_stid); out: nfsd4_bump_seqid(cstate, status); diff --git a/fs/nfsd/state.h b/fs/nfsd/state.h index 986e51e5ceac..64053eadeb81 100644 --- a/fs/nfsd/state.h +++ b/fs/nfsd/state.h @@ -535,7 +535,7 @@ struct nfs4_ol_stateid { unsigned char st_access_bmap; unsigned char st_deny_bmap; struct nfs4_ol_stateid *st_openstp; - struct rw_semaphore st_rwsem; + struct mutex st_mutex; }; static inline struct nfs4_ol_stateid *openlockstateid(struct nfs4_stid *s) -- cgit v1.3-8-gc7d7 From 5cc1fb2a093e254b656c64ff24b0b76bed1d34d9 Mon Sep 17 00:00:00 2001 From: Oleg Drokin Date: Tue, 14 Jun 2016 23:28:05 -0400 Subject: nfsd: Extend the mutex holding region around in nfsd4_process_open2() To avoid racing entry into nfs4_get_vfs_file(). Make init_open_stateid() return with locked stateid to be unlocked by the caller. Signed-off-by: Oleg Drokin Cc: stable@vger.kernel.org Signed-off-by: J. Bruce Fields --- fs/nfsd/nfs4state.c | 16 +++++++++++++--- 1 file changed, 13 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index c927d36d8ac6..94854a06c9ad 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -3487,6 +3487,10 @@ init_open_stateid(struct nfs4_ol_stateid *stp, struct nfs4_file *fp, struct nfs4_openowner *oo = open->op_openowner; struct nfs4_ol_stateid *retstp = NULL; + /* We are moving these outside of the spinlocks to avoid the warnings */ + mutex_init(&stp->st_mutex); + mutex_lock(&stp->st_mutex); + spin_lock(&oo->oo_owner.so_client->cl_lock); spin_lock(&fp->fi_lock); @@ -3502,13 +3506,17 @@ init_open_stateid(struct nfs4_ol_stateid *stp, struct nfs4_file *fp, stp->st_access_bmap = 0; stp->st_deny_bmap = 0; stp->st_openstp = NULL; - mutex_init(&stp->st_mutex); list_add(&stp->st_perstateowner, &oo->oo_owner.so_stateids); list_add(&stp->st_perfile, &fp->fi_stateids); out_unlock: spin_unlock(&fp->fi_lock); spin_unlock(&oo->oo_owner.so_client->cl_lock); + if (retstp) { + mutex_lock(&retstp->st_mutex); + /* Not that we need to, just for neatness */ + mutex_unlock(&stp->st_mutex); + } return retstp; } @@ -4344,11 +4352,14 @@ nfsd4_process_open2(struct svc_rqst *rqstp, struct svc_fh *current_fh, struct nf } else { stp = open->op_stp; open->op_stp = NULL; + /* + * init_open_stateid() either returns a locked stateid + * it found, or initializes and locks the new one we passed in + */ swapstp = init_open_stateid(stp, fp, open); if (swapstp) { nfs4_put_stid(&stp->st_stid); stp = swapstp; - mutex_lock(&stp->st_mutex); status = nfs4_upgrade_open(rqstp, fp, current_fh, stp, open); if (status) { @@ -4357,7 +4368,6 @@ nfsd4_process_open2(struct svc_rqst *rqstp, struct svc_fh *current_fh, struct nf } goto upgrade_out; } - mutex_lock(&stp->st_mutex); status = nfs4_get_vfs_file(rqstp, fp, current_fh, stp, open); if (status) { mutex_unlock(&stp->st_mutex); -- cgit v1.3-8-gc7d7 From 8c7245abda877d4689b3371db8ae2a4400d7d9ce Mon Sep 17 00:00:00 2001 From: Oleg Drokin Date: Tue, 14 Jun 2016 23:28:06 -0400 Subject: nfsd: Make init_open_stateid() a bit more whole Move the state selection logic inside from the caller, always making it return correct stp to use. Signed-off-by: J . Bruce Fields Signed-off-by: Oleg Drokin Signed-off-by: J. Bruce Fields --- fs/nfsd/nfs4state.c | 27 ++++++++++++--------------- 1 file changed, 12 insertions(+), 15 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index 94854a06c9ad..70d0b9b33031 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -3480,13 +3480,14 @@ alloc_init_open_stateowner(unsigned int strhashval, struct nfsd4_open *open, } static struct nfs4_ol_stateid * -init_open_stateid(struct nfs4_ol_stateid *stp, struct nfs4_file *fp, - struct nfsd4_open *open) +init_open_stateid(struct nfs4_file *fp, struct nfsd4_open *open) { struct nfs4_openowner *oo = open->op_openowner; struct nfs4_ol_stateid *retstp = NULL; + struct nfs4_ol_stateid *stp; + stp = open->op_stp; /* We are moving these outside of the spinlocks to avoid the warnings */ mutex_init(&stp->st_mutex); mutex_lock(&stp->st_mutex); @@ -3497,6 +3498,8 @@ init_open_stateid(struct nfs4_ol_stateid *stp, struct nfs4_file *fp, retstp = nfsd4_find_existing_open(fp, open); if (retstp) goto out_unlock; + + open->op_stp = NULL; atomic_inc(&stp->st_stid.sc_count); stp->st_stid.sc_type = NFS4_OPEN_STID; INIT_LIST_HEAD(&stp->st_locks); @@ -3514,10 +3517,11 @@ out_unlock: spin_unlock(&oo->oo_owner.so_client->cl_lock); if (retstp) { mutex_lock(&retstp->st_mutex); - /* Not that we need to, just for neatness */ + /* To keep mutex tracking happy */ mutex_unlock(&stp->st_mutex); + stp = retstp; } - return retstp; + return stp; } /* @@ -4313,7 +4317,6 @@ nfsd4_process_open2(struct svc_rqst *rqstp, struct svc_fh *current_fh, struct nf struct nfs4_client *cl = open->op_openowner->oo_owner.so_client; struct nfs4_file *fp = NULL; struct nfs4_ol_stateid *stp = NULL; - struct nfs4_ol_stateid *swapstp = NULL; struct nfs4_delegation *dp = NULL; __be32 status; @@ -4350,16 +4353,10 @@ nfsd4_process_open2(struct svc_rqst *rqstp, struct svc_fh *current_fh, struct nf goto out; } } else { - stp = open->op_stp; - open->op_stp = NULL; - /* - * init_open_stateid() either returns a locked stateid - * it found, or initializes and locks the new one we passed in - */ - swapstp = init_open_stateid(stp, fp, open); - if (swapstp) { - nfs4_put_stid(&stp->st_stid); - stp = swapstp; + /* stp is returned locked. */ + stp = init_open_stateid(fp, open); + /* See if we lost the race to some other thread */ + if (stp->st_access_bmap != 0) { status = nfs4_upgrade_open(rqstp, fp, current_fh, stp, open); if (status) { -- cgit v1.3-8-gc7d7 From 16ff4b454f1b56e8d89a9075feed0dd6ac510c3d Mon Sep 17 00:00:00 2001 From: Heinrich Schuchardt Date: Sat, 11 Jun 2016 18:11:10 +0200 Subject: btrfs: Use correct format specifier Component mirror_num of struct btrfsic_block is defined as unsigned int. Use %u as format specifier. Signed-off-by: Heinrich Schuchardt Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/check-integrity.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/check-integrity.c b/fs/btrfs/check-integrity.c index b677a6ea6001..7706c8dc5fa6 100644 --- a/fs/btrfs/check-integrity.c +++ b/fs/btrfs/check-integrity.c @@ -2645,7 +2645,7 @@ static void btrfsic_dump_tree_sub(const struct btrfsic_state *state, * This algorithm is recursive because the amount of used stack space * is very small and the max recursion depth is limited. */ - indent_add = sprintf(buf, "%c-%llu(%s/%llu/%d)", + indent_add = sprintf(buf, "%c-%llu(%s/%llu/%u)", btrfsic_get_block_type(state, block), block->logical_bytenr, block->dev_state->name, block->dev_bytenr, block->mirror_num); -- cgit v1.3-8-gc7d7 From c871b0f2fd27e7f9097d507f47de5270f88003b9 Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Mon, 6 Jun 2016 12:01:23 -0700 Subject: Btrfs: check if extent buffer is aligned to sectorsize Thanks to fuzz testing, we can pass an invalid bytenr to extent buffer via alloc_extent_buffer(). An unaligned eb can have more pages than it should have, which ends up extent buffer's leak or some corrupted content in extent buffer. This adds a warning to let us quickly know what was happening. Now that alloc_extent_buffer() no more returns NULL, this changes its caller and callers of its caller to match with the new error handling. Signed-off-by: Liu Bo Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/ctree.c | 2 ++ fs/btrfs/disk-io.c | 8 ++++---- fs/btrfs/extent-tree.c | 10 ++++++---- fs/btrfs/extent_io.c | 15 ++++++++++++--- fs/btrfs/tree-log.c | 4 ++-- fs/btrfs/volumes.c | 4 ++-- 6 files changed, 28 insertions(+), 15 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c index 46025688f1d0..827c949fa4bc 100644 --- a/fs/btrfs/ctree.c +++ b/fs/btrfs/ctree.c @@ -2512,6 +2512,8 @@ read_block_for_search(struct btrfs_trans_handle *trans, if (!btrfs_buffer_uptodate(tmp, 0, 0)) ret = -EIO; free_extent_buffer(tmp); + } else { + ret = PTR_ERR(tmp); } return ret; } diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 1142127f6e5e..7b5d5e8efde6 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1098,7 +1098,7 @@ void readahead_tree_block(struct btrfs_root *root, u64 bytenr) struct inode *btree_inode = root->fs_info->btree_inode; buf = btrfs_find_create_tree_block(root, bytenr); - if (!buf) + if (IS_ERR(buf)) return; read_extent_buffer_pages(&BTRFS_I(btree_inode)->io_tree, buf, 0, WAIT_NONE, btree_get_extent, 0); @@ -1114,7 +1114,7 @@ int reada_tree_block_flagged(struct btrfs_root *root, u64 bytenr, int ret; buf = btrfs_find_create_tree_block(root, bytenr); - if (!buf) + if (IS_ERR(buf)) return 0; set_bit(EXTENT_BUFFER_READAHEAD, &buf->bflags); @@ -1172,8 +1172,8 @@ struct extent_buffer *read_tree_block(struct btrfs_root *root, u64 bytenr, int ret; buf = btrfs_find_create_tree_block(root, bytenr); - if (!buf) - return ERR_PTR(-ENOMEM); + if (IS_ERR(buf)) + return buf; ret = btree_read_extent_buffer_pages(root, buf, 0, parent_transid); if (ret) { diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 689d25ac6a68..5439e85c4813 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -8016,8 +8016,9 @@ btrfs_init_new_buffer(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct extent_buffer *buf; buf = btrfs_find_create_tree_block(root, bytenr); - if (!buf) - return ERR_PTR(-ENOMEM); + if (IS_ERR(buf)) + return buf; + btrfs_set_header_generation(buf, trans->transid); btrfs_set_buffer_lockdep_class(root->root_key.objectid, buf, level); btrfs_tree_lock(buf); @@ -8659,8 +8660,9 @@ static noinline int do_walk_down(struct btrfs_trans_handle *trans, next = btrfs_find_tree_block(root->fs_info, bytenr); if (!next) { next = btrfs_find_create_tree_block(root, bytenr); - if (!next) - return -ENOMEM; + if (IS_ERR(next)) + return PTR_ERR(next); + btrfs_set_buffer_lockdep_class(root->root_key.objectid, next, level - 1); reada = 1; diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index a3412d68ad37..aaee3ef55ed8 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -4892,18 +4892,25 @@ struct extent_buffer *alloc_extent_buffer(struct btrfs_fs_info *fs_info, int uptodate = 1; int ret; + if (!IS_ALIGNED(start, fs_info->tree_root->sectorsize)) { + btrfs_err(fs_info, "bad tree block start %llu", start); + return ERR_PTR(-EINVAL); + } + eb = find_extent_buffer(fs_info, start); if (eb) return eb; eb = __alloc_extent_buffer(fs_info, start, len); if (!eb) - return NULL; + return ERR_PTR(-ENOMEM); for (i = 0; i < num_pages; i++, index++) { p = find_or_create_page(mapping, index, GFP_NOFS|__GFP_NOFAIL); - if (!p) + if (!p) { + exists = ERR_PTR(-ENOMEM); goto free_eb; + } spin_lock(&mapping->private_lock); if (PagePrivate(p)) { @@ -4948,8 +4955,10 @@ struct extent_buffer *alloc_extent_buffer(struct btrfs_fs_info *fs_info, set_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags); again: ret = radix_tree_preload(GFP_NOFS); - if (ret) + if (ret) { + exists = ERR_PTR(ret); goto free_eb; + } spin_lock(&fs_info->buffer_lock); ret = radix_tree_insert(&fs_info->buffer_radix, diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c index aa8fed11f749..8ab1dc64cbba 100644 --- a/fs/btrfs/tree-log.c +++ b/fs/btrfs/tree-log.c @@ -2422,8 +2422,8 @@ static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans, root_owner = btrfs_header_owner(parent); next = btrfs_find_create_tree_block(root, bytenr); - if (!next) - return -ENOMEM; + if (IS_ERR(next)) + return PTR_ERR(next); if (*level == 1) { ret = wc->process_func(root, next, wc, ptr_gen); diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index fcbda4341f7d..c3a2900c6030 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -6607,8 +6607,8 @@ int btrfs_read_sys_array(struct btrfs_root *root) * overallocate but we can keep it as-is, only the first page is used. */ sb = btrfs_find_create_tree_block(root, BTRFS_SUPER_INFO_OFFSET); - if (!sb) - return -ENOMEM; + if (IS_ERR(sb)) + return PTR_ERR(sb); set_extent_buffer_uptodate(sb); btrfs_set_buffer_lockdep_class(root->root_key.objectid, sb, 0); /* -- cgit v1.3-8-gc7d7 From 64c12921e11b3a0c10d088606e328c58e29274d8 Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Wed, 8 Jun 2016 00:36:38 -0400 Subject: btrfs: account for non-CoW'd blocks in btrfs_abort_transaction The test for !trans->blocks_used in btrfs_abort_transaction is insufficient to determine whether it's safe to drop the transaction handle on the floor. btrfs_cow_block, informed by should_cow_block, can return blocks that have already been CoW'd in the current transaction. trans->blocks_used is only incremented for new block allocations. If an operation overlaps the blocks in the current transaction entirely and must abort the transaction, we'll happily let it clean up the trans handle even though it may have modified the blocks and will commit an incomplete operation. In the long-term, I'd like to do closer tracking of when the fs is actually modified so we can still recover as gracefully as possible, but that approach will need some discussion. In the short term, since this is the only code using trans->blocks_used, let's just switch it to a bool indicating whether any blocks were used and set it when should_cow_block returns false. Cc: stable@vger.kernel.org # 3.4+ Signed-off-by: Jeff Mahoney Reviewed-by: Filipe Manana Signed-off-by: David Sterba --- fs/btrfs/ctree.c | 5 ++++- fs/btrfs/extent-tree.c | 2 +- fs/btrfs/super.c | 2 +- fs/btrfs/transaction.h | 2 +- 4 files changed, 7 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c index 827c949fa4bc..6276add8538a 100644 --- a/fs/btrfs/ctree.c +++ b/fs/btrfs/ctree.c @@ -1554,6 +1554,7 @@ noinline int btrfs_cow_block(struct btrfs_trans_handle *trans, trans->transid, root->fs_info->generation); if (!should_cow_block(trans, root, buf)) { + trans->dirty = true; *cow_ret = buf; return 0; } @@ -2777,8 +2778,10 @@ again: * then we don't want to set the path blocking, * so we test it here */ - if (!should_cow_block(trans, root, b)) + if (!should_cow_block(trans, root, b)) { + trans->dirty = true; goto cow_done; + } /* * must have write locks on this node and the diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 5439e85c4813..29e5d000bbee 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -8045,7 +8045,7 @@ btrfs_init_new_buffer(struct btrfs_trans_handle *trans, struct btrfs_root *root, set_extent_dirty(&trans->transaction->dirty_pages, buf->start, buf->start + buf->len - 1, GFP_NOFS); } - trans->blocks_used++; + trans->dirty = true; /* this returns a buffer locked for blocking */ return buf; } diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index 4339b6613f19..bf70d33b5791 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -235,7 +235,7 @@ void __btrfs_abort_transaction(struct btrfs_trans_handle *trans, trans->aborted = errno; /* Nothing used. The other threads that have joined this * transaction may be able to continue. */ - if (!trans->blocks_used && list_empty(&trans->new_bgs)) { + if (!trans->dirty && list_empty(&trans->new_bgs)) { const char *errstr; errstr = btrfs_decode_error(errno); diff --git a/fs/btrfs/transaction.h b/fs/btrfs/transaction.h index 9fe0ec2bf0fe..c5abee4f01ad 100644 --- a/fs/btrfs/transaction.h +++ b/fs/btrfs/transaction.h @@ -110,7 +110,6 @@ struct btrfs_trans_handle { u64 chunk_bytes_reserved; unsigned long use_count; unsigned long blocks_reserved; - unsigned long blocks_used; unsigned long delayed_ref_updates; struct btrfs_transaction *transaction; struct btrfs_block_rsv *block_rsv; @@ -121,6 +120,7 @@ struct btrfs_trans_handle { bool can_flush_pending_bgs; bool reloc_reserved; bool sync; + bool dirty; unsigned int type; /* * this root is only needed to validate that the root passed to -- cgit v1.3-8-gc7d7 From 3b6571c180da85e43550c608e954ab7b2a31d954 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 27 May 2016 13:03:04 -0400 Subject: Btrfs: don't BUG_ON() in btrfs_orphan_add This is just a screwup for developers, so change it to an ASSERT() so developers notice when things go wrong and deal with the error appropriately if ASSERT() isn't enabled. Thanks, Signed-off-by: Josef Bacik Reviewed-by: Mark Fasheh Signed-off-by: David Sterba --- fs/btrfs/inode.c | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index e5558d9e396e..bb62418b8023 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -3271,7 +3271,16 @@ int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode) /* grab metadata reservation from transaction handle */ if (reserve) { ret = btrfs_orphan_reserve_metadata(trans, inode); - BUG_ON(ret); /* -ENOSPC in reservation; Logic error? JDM */ + ASSERT(!ret); + if (ret) { + atomic_dec(&root->orphan_inodes); + clear_bit(BTRFS_INODE_ORPHAN_META_RESERVED, + &BTRFS_I(inode)->runtime_flags); + if (insert) + clear_bit(BTRFS_INODE_HAS_ORPHAN_ITEM, + &BTRFS_I(inode)->runtime_flags); + return ret; + } } /* insert an orphan item to track this unlinked/truncated file */ -- cgit v1.3-8-gc7d7 From 90c711ab380d633bf85249bd5d819edb601feda7 Mon Sep 17 00:00:00 2001 From: Zygo Blaxell Date: Sun, 12 Jun 2016 23:39:58 -0400 Subject: btrfs: avoid blocking open_ctree from cleaner_kthread This fixes a problem introduced in commit 2f3165ecf103599f82bf0ea254039db335fb5005 "btrfs: don't force mounts to wait for cleaner_kthread to delete one or more subvolumes". open_ctree eventually calls btrfs_replay_log which in turn calls btrfs_commit_super which tries to lock the cleaner_mutex, causing a recursive mutex deadlock during mount. Instead of playing whack-a-mole trying to keep up with all the functions that may want to lock cleaner_mutex, put all the cleaner_mutex lockers back where they were, and attack the problem more directly: keep cleaner_kthread asleep until the filesystem is mounted. When filesystems are mounted read-only and later remounted read-write, open_ctree did not set fs_info->open and neither does anything else. Set this flag in btrfs_remount so that neither btrfs_delete_unused_bgs nor cleaner_kthread get confused by the common case of "/" filesystem read-only mount followed by read-write remount. Signed-off-by: Zygo Blaxell Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/disk-io.c | 25 ++++++++++--------------- fs/btrfs/super.c | 2 ++ 2 files changed, 12 insertions(+), 15 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 7b5d5e8efde6..789f5f233940 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1806,6 +1806,13 @@ static int cleaner_kthread(void *arg) if (btrfs_need_cleaner_sleep(root)) goto sleep; + /* + * Do not do anything if we might cause open_ctree() to block + * before we have finished mounting the filesystem. + */ + if (!root->fs_info->open) + goto sleep; + if (!mutex_trylock(&root->fs_info->cleaner_mutex)) goto sleep; @@ -2520,7 +2527,6 @@ int open_ctree(struct super_block *sb, int num_backups_tried = 0; int backup_index = 0; int max_active; - bool cleaner_mutex_locked = false; tree_root = fs_info->tree_root = btrfs_alloc_root(fs_info, GFP_KERNEL); chunk_root = fs_info->chunk_root = btrfs_alloc_root(fs_info, GFP_KERNEL); @@ -2999,13 +3005,6 @@ retry_root_backup: goto fail_sysfs; } - /* - * Hold the cleaner_mutex thread here so that we don't block - * for a long time on btrfs_recover_relocation. cleaner_kthread - * will wait for us to finish mounting the filesystem. - */ - mutex_lock(&fs_info->cleaner_mutex); - cleaner_mutex_locked = true; fs_info->cleaner_kthread = kthread_run(cleaner_kthread, tree_root, "btrfs-cleaner"); if (IS_ERR(fs_info->cleaner_kthread)) @@ -3065,8 +3064,10 @@ retry_root_backup: ret = btrfs_cleanup_fs_roots(fs_info); if (ret) goto fail_qgroup; - /* We locked cleaner_mutex before creating cleaner_kthread. */ + + mutex_lock(&fs_info->cleaner_mutex); ret = btrfs_recover_relocation(tree_root); + mutex_unlock(&fs_info->cleaner_mutex); if (ret < 0) { btrfs_warn(fs_info, "failed to recover relocation: %d", ret); @@ -3074,8 +3075,6 @@ retry_root_backup: goto fail_qgroup; } } - mutex_unlock(&fs_info->cleaner_mutex); - cleaner_mutex_locked = false; location.objectid = BTRFS_FS_TREE_OBJECTID; location.type = BTRFS_ROOT_ITEM_KEY; @@ -3189,10 +3188,6 @@ fail_cleaner: filemap_write_and_wait(fs_info->btree_inode->i_mapping); fail_sysfs: - if (cleaner_mutex_locked) { - mutex_unlock(&fs_info->cleaner_mutex); - cleaner_mutex_locked = false; - } btrfs_sysfs_remove_mounted(fs_info); fail_fsdev_sysfs: diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index bf70d33b5791..60e7179ed4b7 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -1807,6 +1807,8 @@ static int btrfs_remount(struct super_block *sb, int *flags, char *data) } } sb->s_flags &= ~MS_RDONLY; + + fs_info->open = 1; } out: wake_up_process(fs_info->transaction_kthread); -- cgit v1.3-8-gc7d7 From f7af3934c2bccba261972261ac8ebcbf92a346b2 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Fri, 17 Jun 2016 18:15:25 +0200 Subject: btrfs: use new error message helper in qgroup_account_snapshot We've renamed btrfs_std_error, this one is left from last merge. Signed-off-by: David Sterba --- fs/btrfs/transaction.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c index f6e24cb423ae..4e74b5733030 100644 --- a/fs/btrfs/transaction.c +++ b/fs/btrfs/transaction.c @@ -1385,7 +1385,7 @@ static int qgroup_account_snapshot(struct btrfs_trans_handle *trans, switch_commit_roots(trans->transaction, fs_info); ret = btrfs_write_and_wait_transaction(trans, src); if (ret) - btrfs_std_error(fs_info, ret, + btrfs_handle_fs_error(fs_info, ret, "Error while writing out transaction for qgroup"); out: -- cgit v1.3-8-gc7d7 From 89c5a5441d703ba068699524ae59f7806e9b173d Mon Sep 17 00:00:00 2001 From: David Sterba Date: Thu, 16 Jun 2016 17:34:28 +0200 Subject: btrfs: remove build fixup for qgroup_account_snapshot Introduced in 2c1984f244838477aab ("btrfs: build fixup for qgroup_account_snapshot") as temporary bisectability build fixup. Signed-off-by: David Sterba --- fs/btrfs/transaction.c | 5 ----- 1 file changed, 5 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c index 4e74b5733030..765845742fde 100644 --- a/fs/btrfs/transaction.c +++ b/fs/btrfs/transaction.c @@ -1311,11 +1311,6 @@ int btrfs_defrag_root(struct btrfs_root *root) return ret; } -/* Bisesctability fixup, remove in 4.8 */ -#ifndef btrfs_std_error -#define btrfs_std_error btrfs_handle_fs_error -#endif - /* * Do all special snapshot related qgroup dirty hack. * -- cgit v1.3-8-gc7d7 From dd5c93111dc9d26e038ac437f7a403d617e82c62 Mon Sep 17 00:00:00 2001 From: Chandan Rajendra Date: Thu, 16 Jun 2016 22:07:58 +0530 Subject: Btrfs: btrfs_check_super_valid: Allow 4096 as stripesize Older btrfs-progs/mkfs.btrfs sets 4096 as the stripesize. Hence restricting stripesize to be equal to sectorsize would cause super block validation to return an error on architectures where PAGE_SIZE is not equal to 4096. Hence as a workaround, this commit allows stripesize to be set to 4096 bytes. Signed-off-by: Chandan Rajendra Signed-off-by: David Sterba --- fs/btrfs/disk-io.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 789f5f233940..54cca7a1572b 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -4134,7 +4134,8 @@ static int btrfs_check_super_valid(struct btrfs_fs_info *fs_info, ret = -EINVAL; } if (!is_power_of_2(btrfs_super_stripesize(sb)) || - btrfs_super_stripesize(sb) != sectorsize) { + ((btrfs_super_stripesize(sb) != sectorsize) && + (btrfs_super_stripesize(sb) != 4096))) { btrfs_err(fs_info, "invalid stripesize %u", btrfs_super_stripesize(sb)); ret = -EINVAL; -- cgit v1.3-8-gc7d7 From e7d6ef9790bc281f5c29d0132b68031248523fe8 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Mon, 20 Jun 2016 01:35:59 -0400 Subject: fix idiotic braino in d_alloc_parallel() Check for d_unhashed() while searching in in-lookup hash was absolutely wrong. Worse, it masked a deadlock on dget() done under bitlock that nests inside ->d_lock. Thanks to J. R. Okajima for spotting it. Spotted-by: "J. R. Okajima" Wearing-brown-paperbag: Al Viro Signed-off-by: Al Viro --- fs/dcache.c | 17 ++++++++++++----- 1 file changed, 12 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/dcache.c b/fs/dcache.c index b7eddfd35aa5..d6847d7b123d 100644 --- a/fs/dcache.c +++ b/fs/dcache.c @@ -2503,7 +2503,6 @@ retry: rcu_read_unlock(); goto retry; } - rcu_read_unlock(); /* * No changes for the parent since the beginning of d_lookup(). * Since all removals from the chain happen with hlist_bl_lock(), @@ -2516,8 +2515,6 @@ retry: continue; if (dentry->d_parent != parent) continue; - if (d_unhashed(dentry)) - continue; if (parent->d_flags & DCACHE_OP_COMPARE) { int tlen = dentry->d_name.len; const char *tname = dentry->d_name.name; @@ -2529,9 +2526,18 @@ retry: if (dentry_cmp(dentry, str, len)) continue; } - dget(dentry); hlist_bl_unlock(b); - /* somebody is doing lookup for it right now; wait for it */ + /* now we can try to grab a reference */ + if (!lockref_get_not_dead(&dentry->d_lockref)) { + rcu_read_unlock(); + goto retry; + } + + rcu_read_unlock(); + /* + * somebody is likely to be still doing lookup for it; + * wait for them to finish + */ spin_lock(&dentry->d_lock); d_wait_lookup(dentry); /* @@ -2562,6 +2568,7 @@ retry: dput(new); return dentry; } + rcu_read_unlock(); /* we can't take ->d_lock here; it's OK, though. */ new->d_flags |= DCACHE_PAR_LOOKUP; new->d_wait = wq; -- cgit v1.3-8-gc7d7 From 4ac1c17b2044a1b4b2fbed74451947e905fc2992 Mon Sep 17 00:00:00 2001 From: "Kirill A. Shutemov" Date: Thu, 16 Jun 2016 23:26:15 +0200 Subject: UBIFS: Implement ->migratepage() During page migrations UBIFS might get confused and the following assert triggers: [ 213.480000] UBIFS assert failed in ubifs_set_page_dirty at 1451 (pid 436) [ 213.490000] CPU: 0 PID: 436 Comm: drm-stress-test Not tainted 4.4.4-00176-geaa802524636-dirty #1008 [ 213.490000] Hardware name: Allwinner sun4i/sun5i Families [ 213.490000] [] (unwind_backtrace) from [] (show_stack+0x10/0x14) [ 213.490000] [] (show_stack) from [] (dump_stack+0x8c/0xa0) [ 213.490000] [] (dump_stack) from [] (ubifs_set_page_dirty+0x44/0x50) [ 213.490000] [] (ubifs_set_page_dirty) from [] (try_to_unmap_one+0x10c/0x3a8) [ 213.490000] [] (try_to_unmap_one) from [] (rmap_walk+0xb4/0x290) [ 213.490000] [] (rmap_walk) from [] (try_to_unmap+0x64/0x80) [ 213.490000] [] (try_to_unmap) from [] (migrate_pages+0x328/0x7a0) [ 213.490000] [] (migrate_pages) from [] (alloc_contig_range+0x168/0x2f4) [ 213.490000] [] (alloc_contig_range) from [] (cma_alloc+0x170/0x2c0) [ 213.490000] [] (cma_alloc) from [] (__alloc_from_contiguous+0x38/0xd8) [ 213.490000] [] (__alloc_from_contiguous) from [] (__dma_alloc+0x23c/0x274) [ 213.490000] [] (__dma_alloc) from [] (arm_dma_alloc+0x54/0x5c) [ 213.490000] [] (arm_dma_alloc) from [] (drm_gem_cma_create+0xb8/0xf0) [ 213.490000] [] (drm_gem_cma_create) from [] (drm_gem_cma_create_with_handle+0x1c/0xe8) [ 213.490000] [] (drm_gem_cma_create_with_handle) from [] (drm_gem_cma_dumb_create+0x3c/0x48) [ 213.490000] [] (drm_gem_cma_dumb_create) from [] (drm_ioctl+0x12c/0x444) [ 213.490000] [] (drm_ioctl) from [] (do_vfs_ioctl+0x3f4/0x614) [ 213.490000] [] (do_vfs_ioctl) from [] (SyS_ioctl+0x34/0x5c) [ 213.490000] [] (SyS_ioctl) from [] (ret_fast_syscall+0x0/0x34) UBIFS is using PagePrivate() which can have different meanings across filesystems. Therefore the generic page migration code cannot handle this case correctly. We have to implement our own migration function which basically does a plain copy but also duplicates the page private flag. UBIFS is not a block device filesystem and cannot use buffer_migrate_page(). Cc: stable@vger.kernel.org Signed-off-by: Kirill A. Shutemov [rw: Massaged changelog, build fixes, etc...] Signed-off-by: Richard Weinberger Acked-by: Christoph Hellwig --- fs/ubifs/file.c | 24 ++++++++++++++++++++++++ 1 file changed, 24 insertions(+) (limited to 'fs') diff --git a/fs/ubifs/file.c b/fs/ubifs/file.c index 08316972ff93..7bbf420d1289 100644 --- a/fs/ubifs/file.c +++ b/fs/ubifs/file.c @@ -52,6 +52,7 @@ #include "ubifs.h" #include #include +#include static int read_block(struct inode *inode, void *addr, unsigned int block, struct ubifs_data_node *dn) @@ -1452,6 +1453,26 @@ static int ubifs_set_page_dirty(struct page *page) return ret; } +#ifdef CONFIG_MIGRATION +static int ubifs_migrate_page(struct address_space *mapping, + struct page *newpage, struct page *page, enum migrate_mode mode) +{ + int rc; + + rc = migrate_page_move_mapping(mapping, newpage, page, NULL, mode, 0); + if (rc != MIGRATEPAGE_SUCCESS) + return rc; + + if (PagePrivate(page)) { + ClearPagePrivate(page); + SetPagePrivate(newpage); + } + + migrate_page_copy(newpage, page); + return MIGRATEPAGE_SUCCESS; +} +#endif + static int ubifs_releasepage(struct page *page, gfp_t unused_gfp_flags) { /* @@ -1591,6 +1612,9 @@ const struct address_space_operations ubifs_file_address_operations = { .write_end = ubifs_write_end, .invalidatepage = ubifs_invalidatepage, .set_page_dirty = ubifs_set_page_dirty, +#ifdef CONFIG_MIGRATION + .migratepage = ubifs_migrate_page, +#endif .releasepage = ubifs_releasepage, }; -- cgit v1.3-8-gc7d7 From 31b9655f439a26856edca0f3f8daa368a61f16d5 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Mon, 11 Apr 2016 17:37:40 -0400 Subject: Btrfs: track transid for delayed ref flushing Using the offwakecputime bpf script I noticed most of our time was spent waiting on the delayed ref throttling. This is what is supposed to happen, but sometimes the transaction can commit and then we're waiting for throttling that doesn't matter anymore. So change this stuff to be a little smarter by tracking the transid we were in when we initiated the throttling. If the transaction we get is different then we can just bail out. This resulted in a 50% speedup in my fs_mark test, and reduced the amount of time spent throttling by 60 seconds over the entire run (which is about 30 minutes). Thanks, Signed-off-by: Josef Bacik Signed-off-by: Chris Mason --- fs/btrfs/ctree.h | 2 +- fs/btrfs/extent-tree.c | 15 ++++++++++++--- fs/btrfs/inode.c | 1 + fs/btrfs/transaction.c | 3 ++- 4 files changed, 16 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 101c3cfd3f7c..4274a7bfdaed 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -2518,7 +2518,7 @@ void btrfs_put_block_group(struct btrfs_block_group_cache *cache); int btrfs_run_delayed_refs(struct btrfs_trans_handle *trans, struct btrfs_root *root, unsigned long count); int btrfs_async_run_delayed_refs(struct btrfs_root *root, - unsigned long count, int wait); + unsigned long count, u64 transid, int wait); int btrfs_lookup_data_extent(struct btrfs_root *root, u64 start, u64 len); int btrfs_lookup_extent_info(struct btrfs_trans_handle *trans, struct btrfs_root *root, u64 bytenr, diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 29e5d000bbee..ecfa52002363 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -2835,6 +2835,7 @@ int btrfs_should_throttle_delayed_refs(struct btrfs_trans_handle *trans, struct async_delayed_refs { struct btrfs_root *root; + u64 transid; int count; int error; int sync; @@ -2850,9 +2851,16 @@ static void delayed_ref_async_start(struct btrfs_work *work) async = container_of(work, struct async_delayed_refs, work); - trans = btrfs_join_transaction(async->root); + trans = btrfs_attach_transaction(async->root); if (IS_ERR(trans)) { - async->error = PTR_ERR(trans); + if (PTR_ERR(trans) != -ENOENT) + async->error = PTR_ERR(trans); + goto done; + } + + /* Don't bother flushing if we got into a different transaction */ + if (trans->transid != async->transid) { + btrfs_end_transaction(trans, async->root); goto done; } @@ -2876,7 +2884,7 @@ done: } int btrfs_async_run_delayed_refs(struct btrfs_root *root, - unsigned long count, int wait) + unsigned long count, u64 transid, int wait) { struct async_delayed_refs *async; int ret; @@ -2888,6 +2896,7 @@ int btrfs_async_run_delayed_refs(struct btrfs_root *root, async->root = root->fs_info->tree_root; async->count = count; async->error = 0; + async->transid = transid; if (wait) async->sync = 1; else diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index bb62418b8023..78582e339f33 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -4558,6 +4558,7 @@ delete: BUG_ON(ret); if (btrfs_should_throttle_delayed_refs(trans, root)) btrfs_async_run_delayed_refs(root, + trans->transid, trans->delayed_ref_updates * 2, 0); if (be_nice) { if (truncate_space_check(trans, root, diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c index 765845742fde..948aa186b353 100644 --- a/fs/btrfs/transaction.c +++ b/fs/btrfs/transaction.c @@ -818,6 +818,7 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans, { struct btrfs_transaction *cur_trans = trans->transaction; struct btrfs_fs_info *info = root->fs_info; + u64 transid = trans->transid; unsigned long cur = trans->delayed_ref_updates; int lock = (trans->type != TRANS_JOIN_NOLOCK); int err = 0; @@ -905,7 +906,7 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans, kmem_cache_free(btrfs_trans_handle_cachep, trans); if (must_run_delayed_refs) { - btrfs_async_run_delayed_refs(root, cur, + btrfs_async_run_delayed_refs(root, cur, transid, must_run_delayed_refs == 1); } return err; -- cgit v1.3-8-gc7d7 From 0f873eca82a0bee45f38862e0ea2ac7b1c2a31bd Mon Sep 17 00:00:00 2001 From: Chris Mason Date: Wed, 27 Apr 2016 09:59:38 -0400 Subject: btrfs: fix deadlock in delayed_ref_async_start "Btrfs: track transid for delayed ref flushing" was deadlocking on btrfs_attach_transaction because its not safe to call from the async delayed ref start code. This commit brings back btrfs_join_transaction instead and checks for a blocked commit. Signed-off-by: Josef Bacik Signed-off-by: Chris Mason --- fs/btrfs/extent-tree.c | 20 +++++++++++--------- 1 file changed, 11 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index ecfa52002363..82b912a293ab 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -2851,16 +2851,13 @@ static void delayed_ref_async_start(struct btrfs_work *work) async = container_of(work, struct async_delayed_refs, work); - trans = btrfs_attach_transaction(async->root); - if (IS_ERR(trans)) { - if (PTR_ERR(trans) != -ENOENT) - async->error = PTR_ERR(trans); + /* if the commit is already started, we don't need to wait here */ + if (btrfs_transaction_blocked(async->root->fs_info)) goto done; - } - /* Don't bother flushing if we got into a different transaction */ - if (trans->transid != async->transid) { - btrfs_end_transaction(trans, async->root); + trans = btrfs_join_transaction(async->root); + if (IS_ERR(trans)) { + async->error = PTR_ERR(trans); goto done; } @@ -2869,10 +2866,15 @@ static void delayed_ref_async_start(struct btrfs_work *work) * wait on delayed refs */ trans->sync = true; + + /* Don't bother flushing if we got into a different transaction */ + if (trans->transid > async->transid) + goto end; + ret = btrfs_run_delayed_refs(trans, async->root, async->count); if (ret) async->error = ret; - +end: ret = btrfs_end_transaction(trans, async->root); if (ret && !async->error) async->error = ret; -- cgit v1.3-8-gc7d7 From c6887cd11149d7325328749f06719071e6c725c6 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 25 Mar 2016 13:26:00 -0400 Subject: Btrfs: don't do nocow check unless we have to Before we write into prealloc/nocow space we have to make sure that there are no references to the extents we are writing into, which means checking the extent tree and csum tree in the case of nocow. So we don't want to do the nocow dance unless we can't reserve data space, since it's a serious drag on performance. With the following sequence fallocate -l10737418240 /mnt/btrfs-test/file cp --reflink /mnt/btrfs-test/file /mnt/btrfs-test/link fio --name=randwrite --rw=randwrite --bs=4k --filename=/mnt/btrfs-test/file \ --end_fsync=1 we get the worst case scenario where we have to fall back on to doing the check anyway. Without this patch lat (usec): min=5, max=111598, avg=27.65, stdev=124.51 write: io=10240MB, bw=126876KB/s, iops=31718, runt= 82646msec With this patch lat (usec): min=3, max=91210, avg=14.09, stdev=110.62 write: io=10240MB, bw=212753KB/s, iops=53188, runt= 49286msec We get twice the throughput, half of the runtime, and half of the average latency. Thanks, Signed-off-by: Josef Bacik [ PAGE_CACHE_ removal related fixups ] Signed-off-by: David Sterba Signed-off-by: Chris Mason --- fs/btrfs/file.c | 44 ++++++++++++++++++++++---------------------- 1 file changed, 22 insertions(+), 22 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index 159a93450e26..8a538abb597a 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -1534,30 +1534,30 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file, reserve_bytes = round_up(write_bytes + sector_offset, root->sectorsize); - if ((BTRFS_I(inode)->flags & (BTRFS_INODE_NODATACOW | - BTRFS_INODE_PREALLOC)) && - check_can_nocow(inode, pos, &write_bytes) > 0) { - /* - * For nodata cow case, no need to reserve - * data space. - */ - only_release_metadata = true; - /* - * our prealloc extent may be smaller than - * write_bytes, so scale down. - */ - num_pages = DIV_ROUND_UP(write_bytes + offset, - PAGE_SIZE); - reserve_bytes = round_up(write_bytes + sector_offset, - root->sectorsize); - goto reserve_metadata; - } - ret = btrfs_check_data_free_space(inode, pos, write_bytes); - if (ret < 0) - break; + if (ret < 0) { + if ((BTRFS_I(inode)->flags & (BTRFS_INODE_NODATACOW | + BTRFS_INODE_PREALLOC)) && + check_can_nocow(inode, pos, &write_bytes) > 0) { + /* + * For nodata cow case, no need to reserve + * data space. + */ + only_release_metadata = true; + /* + * our prealloc extent may be smaller than + * write_bytes, so scale down. + */ + num_pages = DIV_ROUND_UP(write_bytes + offset, + PAGE_SIZE); + reserve_bytes = round_up(write_bytes + + sector_offset, + root->sectorsize); + } else { + break; + } + } -reserve_metadata: ret = btrfs_delalloc_reserve_metadata(inode, reserve_bytes); if (ret) { if (!only_release_metadata) -- cgit v1.3-8-gc7d7 From 04e1b65af2085d4102b2b5d2fd1e050f8ee63092 Mon Sep 17 00:00:00 2001 From: Wei Yongjun Date: Fri, 17 Jun 2016 17:20:40 +0000 Subject: Btrfs: fix error return code in btrfs_init_test_fs() Fix to return a negative error code from the kern_mount() error handling case instead of 0(ret is set to 0 by register_filesystem), as done elsewhere in this function. Signed-off-by: Wei Yongjun Reviewed-by: Omar Sandoval Signed-off-by: David Sterba Signed-off-by: Chris Mason --- fs/btrfs/tests/btrfs-tests.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/tests/btrfs-tests.c b/fs/btrfs/tests/btrfs-tests.c index 10eb249ef891..02223f3f78f4 100644 --- a/fs/btrfs/tests/btrfs-tests.c +++ b/fs/btrfs/tests/btrfs-tests.c @@ -68,7 +68,7 @@ int btrfs_init_test_fs(void) if (IS_ERR(test_mnt)) { printk(KERN_ERR "btrfs: cannot mount test file system\n"); unregister_filesystem(&test_type); - return ret; + return PTR_ERR(test_mnt); } return 0; } -- cgit v1.3-8-gc7d7 From 415b35a55b57a701afe7391d32a6bb0193b7d3da Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Fri, 17 Jun 2016 19:16:21 -0700 Subject: Btrfs: fix error handling in map_private_extent_buffer map_private_extent_buffer() can return -EINVAL in two different cases, 1. when the requested contents span two pages if nodesize is larger than pagesize, 2. when it detects something insane. The 2nd one used to be only a WARN_ON(1), and we decided to return a error to callers, but we didn't fix up all its callers, which will be addressed by this patch. Without this, btrfs may end up with 'general protection', ie. reading invalid memory. Reported-by: Vegard Nossum Signed-off-by: Liu Bo Signed-off-by: David Sterba Signed-off-by: Chris Mason --- fs/btrfs/ctree.c | 6 +++++- fs/btrfs/extent_io.c | 7 ++++++- 2 files changed, 11 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c index 6276add8538a..a85cf7d23309 100644 --- a/fs/btrfs/ctree.c +++ b/fs/btrfs/ctree.c @@ -1786,10 +1786,12 @@ static noinline int generic_bin_search(struct extent_buffer *eb, if (!err) { tmp = (struct btrfs_disk_key *)(kaddr + offset - map_start); - } else { + } else if (err == 1) { read_extent_buffer(eb, &unaligned, offset, sizeof(unaligned)); tmp = &unaligned; + } else { + return err; } } else { @@ -2830,6 +2832,8 @@ cow_done: } ret = key_search(b, key, level, &prev_cmp, &slot); + if (ret < 0) + goto done; if (level != 0) { int dec = 0; diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index aaee3ef55ed8..75533adef998 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -5342,6 +5342,11 @@ int read_extent_buffer_to_user(struct extent_buffer *eb, void __user *dstv, return ret; } +/* + * return 0 if the item is found within a page. + * return 1 if the item spans two pages. + * return -EINVAL otherwise. + */ int map_private_extent_buffer(struct extent_buffer *eb, unsigned long start, unsigned long min_len, char **map, unsigned long *map_start, @@ -5356,7 +5361,7 @@ int map_private_extent_buffer(struct extent_buffer *eb, unsigned long start, PAGE_SHIFT; if (i != end_i) - return -EINVAL; + return 1; if (i == 0) { offset = start_offset; -- cgit v1.3-8-gc7d7 From c0d2f6104e8ab2eb75e58e72494ad4b69c5227f8 Mon Sep 17 00:00:00 2001 From: Wang Xiaoguang Date: Wed, 22 Jun 2016 09:57:01 +0800 Subject: btrfs: fix disk_i_size update bug when fallocate() fails When doing truncate operation, btrfs_setsize() will first call truncate_setsize() to set new inode->i_size, but if later btrfs_truncate() fails, btrfs_setsize() will call "i_size_write(inode, BTRFS_I(inode)->disk_i_size)" to reset the inmemory inode size, now bug occurs. It's because for truncate case btrfs_ordered_update_i_size() directly uses inode->i_size to update BTRFS_I(inode)->disk_i_size, indeed we should use the "offset" argument to update disk_i_size. Here is the call graph: ==>btrfs_truncate() ====>btrfs_truncate_inode_items() ======>btrfs_ordered_update_i_size(inode, last_size, NULL); Here btrfs_ordered_update_i_size()'s offset argument is last_size. And below test case can reveal this bug: dd if=/dev/zero of=fs.img bs=$((1024*1024)) count=100 dev=$(losetup --show -f fs.img) mkdir -p /mnt/mntpoint mkfs.btrfs -f $dev mount $dev /mnt/mntpoint cd /mnt/mntpoint echo "workdir is: /mnt/mntpoint" blocksize=$((128 * 1024)) dd if=/dev/zero of=testfile bs=$blocksize count=1 sync count=$((17*1024*1024*1024/blocksize)) echo "file size is:" $((count*blocksize)) for ((i = 1; i <= $count; i++)); do i=$((i + 1)) dst_offset=$((blocksize * i)) xfs_io -f -c "reflink testfile 0 $dst_offset $blocksize"\ testfile > /dev/null done sync truncate --size 0 testfile ls -l testfile du -sh testfile exit In this case, truncate operation will fail for enospc reason and "du -sh testfile" returns value greater than 0, but testfile's size is 0, we need to reflect correct inode->i_size. Signed-off-by: Wang Xiaoguang Signed-off-by: David Sterba Signed-off-by: Chris Mason --- fs/btrfs/ordered-data.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/ordered-data.c b/fs/btrfs/ordered-data.c index e96634a725c3..aca8264f4a49 100644 --- a/fs/btrfs/ordered-data.c +++ b/fs/btrfs/ordered-data.c @@ -968,6 +968,7 @@ int btrfs_ordered_update_i_size(struct inode *inode, u64 offset, struct rb_node *prev = NULL; struct btrfs_ordered_extent *test; int ret = 1; + u64 orig_offset = offset; spin_lock_irq(&tree->lock); if (ordered) { @@ -983,7 +984,7 @@ int btrfs_ordered_update_i_size(struct inode *inode, u64 offset, /* truncate file */ if (disk_i_size > i_size) { - BTRFS_I(inode)->disk_i_size = i_size; + BTRFS_I(inode)->disk_i_size = orig_offset; ret = 0; goto out; } -- cgit v1.3-8-gc7d7 From b7f67055d2df9b8f68f02e49d256ee3973999bd2 Mon Sep 17 00:00:00 2001 From: Chandan Rajendra Date: Thu, 23 Jun 2016 15:16:44 +0530 Subject: Btrfs: Force stripesize to the value of sectorsize Btrfs code currently assumes stripesize to be same as sectorsize. However Btrfs-progs (until commit df05c7ed455f519e6e15e46196392e4757257305) has been setting btrfs_super_block->stripesize to a value of 4096. This commit makes sure that the value of btrfs_super_block->stripesize is a power of 2. Later, it unconditionally sets btrfs_root->stripesize to sectorsize. Signed-off-by: Chandan Rajendra Reviewed-by: David Sterba Signed-off-by: David Sterba Signed-off-by: Chris Mason --- fs/btrfs/disk-io.c | 6 ++---- fs/btrfs/volumes.c | 4 ++-- 2 files changed, 4 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 54cca7a1572b..60ce1190307b 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -2806,7 +2806,7 @@ int open_ctree(struct super_block *sb, nodesize = btrfs_super_nodesize(disk_super); sectorsize = btrfs_super_sectorsize(disk_super); - stripesize = btrfs_super_stripesize(disk_super); + stripesize = sectorsize; fs_info->dirty_metadata_batch = nodesize * (1 + ilog2(nr_cpu_ids)); fs_info->delalloc_batch = sectorsize * 512 * (1 + ilog2(nr_cpu_ids)); @@ -4133,9 +4133,7 @@ static int btrfs_check_super_valid(struct btrfs_fs_info *fs_info, btrfs_super_bytes_used(sb)); ret = -EINVAL; } - if (!is_power_of_2(btrfs_super_stripesize(sb)) || - ((btrfs_super_stripesize(sb) != sectorsize) && - (btrfs_super_stripesize(sb) != 4096))) { + if (!is_power_of_2(btrfs_super_stripesize(sb))) { btrfs_err(fs_info, "invalid stripesize %u", btrfs_super_stripesize(sb)); ret = -EINVAL; diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index c3a2900c6030..64eec2cec1d3 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -4694,12 +4694,12 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans, if (type & BTRFS_BLOCK_GROUP_RAID5) { raid_stripe_len = find_raid56_stripe_len(ndevs - 1, - btrfs_super_stripesize(info->super_copy)); + extent_root->stripesize); data_stripes = num_stripes - 1; } if (type & BTRFS_BLOCK_GROUP_RAID6) { raid_stripe_len = find_raid56_stripe_len(ndevs - 2, - btrfs_super_stripesize(info->super_copy)); + extent_root->stripesize); data_stripes = num_stripes - 2; } -- cgit v1.3-8-gc7d7 From 3d22462ae915743f3be5bf1ab3d4a6b72c2bb6c9 Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Tue, 24 May 2016 06:27:44 -0400 Subject: cifs: stuff the fl_owner into "pid" field in the lock request Right now, we send the tgid cross the wire. What we really want to send though is a hashed fl_owner_t since samba treats this field as a generic lockowner. It turns out that because we enforce and release locks locally before they are ever sent to the server, this patch makes no difference in behavior. Still, setting OFD locks on the server using the process pid seems wrong, so I think this patch still makes sense. Signed-off-by: Jeff Layton Signed-off-by: Steve French Acked-by: Pavel Shilovsky Acked-by: Sachin Prabhu --- fs/cifs/cifsfs.c | 3 +++ fs/cifs/cifsglob.h | 1 + fs/cifs/file.c | 14 +++++++++++--- 3 files changed, 15 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c index 5d8b7edf8a8f..5d841f39c4b7 100644 --- a/fs/cifs/cifsfs.c +++ b/fs/cifs/cifsfs.c @@ -87,6 +87,7 @@ extern mempool_t *cifs_req_poolp; extern mempool_t *cifs_mid_poolp; struct workqueue_struct *cifsiod_wq; +__u32 cifs_lock_secret; /* * Bumps refcount for cifs super block. @@ -1266,6 +1267,8 @@ init_cifs(void) spin_lock_init(&cifs_file_list_lock); spin_lock_init(&GlobalMid_Lock); + get_random_bytes(&cifs_lock_secret, sizeof(cifs_lock_secret)); + if (cifs_max_pending < 2) { cifs_max_pending = 2; cifs_dbg(FYI, "cifs_max_pending set to min of 2\n"); diff --git a/fs/cifs/cifsglob.h b/fs/cifs/cifsglob.h index bba106cdc43c..8f1d8c1e72be 100644 --- a/fs/cifs/cifsglob.h +++ b/fs/cifs/cifsglob.h @@ -1619,6 +1619,7 @@ void cifs_oplock_break(struct work_struct *work); extern const struct slow_work_ops cifs_oplock_break_ops; extern struct workqueue_struct *cifsiod_wq; +extern __u32 cifs_lock_secret; extern mempool_t *cifs_mid_poolp; diff --git a/fs/cifs/file.c b/fs/cifs/file.c index 9793ae0bcaa2..d4890b6dc22d 100644 --- a/fs/cifs/file.c +++ b/fs/cifs/file.c @@ -1112,6 +1112,12 @@ cifs_push_mandatory_locks(struct cifsFileInfo *cfile) return rc; } +static __u32 +hash_lockowner(fl_owner_t owner) +{ + return cifs_lock_secret ^ hash32_ptr((const void *)owner); +} + struct lock_to_push { struct list_head llist; __u64 offset; @@ -1178,7 +1184,7 @@ cifs_push_posix_locks(struct cifsFileInfo *cfile) else type = CIFS_WRLCK; lck = list_entry(el, struct lock_to_push, llist); - lck->pid = flock->fl_pid; + lck->pid = hash_lockowner(flock->fl_owner); lck->netfid = cfile->fid.netfid; lck->length = length; lck->type = type; @@ -1305,7 +1311,8 @@ cifs_getlk(struct file *file, struct file_lock *flock, __u32 type, posix_lock_type = CIFS_RDLCK; else posix_lock_type = CIFS_WRLCK; - rc = CIFSSMBPosixLock(xid, tcon, netfid, current->tgid, + rc = CIFSSMBPosixLock(xid, tcon, netfid, + hash_lockowner(flock->fl_owner), flock->fl_start, length, flock, posix_lock_type, wait_flag); return rc; @@ -1505,7 +1512,8 @@ cifs_setlk(struct file *file, struct file_lock *flock, __u32 type, posix_lock_type = CIFS_UNLCK; rc = CIFSSMBPosixLock(xid, tcon, cfile->fid.netfid, - current->tgid, flock->fl_start, length, + hash_lockowner(flock->fl_owner), + flock->fl_start, length, NULL, posix_lock_type, wait_flag); goto out; } -- cgit v1.3-8-gc7d7 From 202d772ba02b1deb8835a631cd8255943d1906a0 Mon Sep 17 00:00:00 2001 From: Jerome Marchand Date: Thu, 26 May 2016 11:52:24 +0200 Subject: cifs: use CIFS_MAX_DOMAINNAME_LEN when converting the domain name Currently in build_ntlmssp_auth_blob(), when converting the domain name to UTF16, CIFS_MAX_USERNAME_LEN limit is used. It should be CIFS_MAX_DOMAINNAME_LEN. This patch fixes this. Signed-off-by: Jerome Marchand Signed-off-by: Steve French --- fs/cifs/sess.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/cifs/sess.c b/fs/cifs/sess.c index af0ec2d5ad0e..c3d086e2bd56 100644 --- a/fs/cifs/sess.c +++ b/fs/cifs/sess.c @@ -430,7 +430,7 @@ int build_ntlmssp_auth_blob(unsigned char *pbuffer, } else { int len; len = cifs_strtoUTF16((__le16 *)tmp, ses->domainName, - CIFS_MAX_USERNAME_LEN, nls_cp); + CIFS_MAX_DOMAINNAME_LEN, nls_cp); len *= 2; /* unicode is 2 bytes each */ sec_blob->DomainName.BufferOffset = cpu_to_le32(tmp - pbuffer); sec_blob->DomainName.Length = cpu_to_le16(len); -- cgit v1.3-8-gc7d7 From b8da344b74c822e966c6d19d6b2321efe82c5d97 Mon Sep 17 00:00:00 2001 From: Jerome Marchand Date: Thu, 26 May 2016 11:52:25 +0200 Subject: cifs: dynamic allocation of ntlmssp blob In sess_auth_rawntlmssp_authenticate(), the ntlmssp blob is allocated statically and its size is an "empirical" 5*sizeof(struct _AUTHENTICATE_MESSAGE) (320B on x86_64). I don't know where this value comes from or if it was ever appropriate, but it is currently insufficient: the user and domain name in UTF16 could take 1kB by themselves. Because of that, build_ntlmssp_auth_blob() might corrupt memory (out-of-bounds write). The size of ntlmssp_blob in SMB2_sess_setup() is too small too (sizeof(struct _NEGOTIATE_MESSAGE) + 500). This patch allocates the blob dynamically in build_ntlmssp_auth_blob(). Signed-off-by: Jerome Marchand Signed-off-by: Steve French CC: Stable --- fs/cifs/ntlmssp.h | 2 +- fs/cifs/sess.c | 76 ++++++++++++++++++++++++++++++------------------------- fs/cifs/smb2pdu.c | 10 ++------ 3 files changed, 45 insertions(+), 43 deletions(-) (limited to 'fs') diff --git a/fs/cifs/ntlmssp.h b/fs/cifs/ntlmssp.h index 848249fa120f..3079b38f0afb 100644 --- a/fs/cifs/ntlmssp.h +++ b/fs/cifs/ntlmssp.h @@ -133,6 +133,6 @@ typedef struct _AUTHENTICATE_MESSAGE { int decode_ntlmssp_challenge(char *bcc_ptr, int blob_len, struct cifs_ses *ses); void build_ntlmssp_negotiate_blob(unsigned char *pbuffer, struct cifs_ses *ses); -int build_ntlmssp_auth_blob(unsigned char *pbuffer, u16 *buflen, +int build_ntlmssp_auth_blob(unsigned char **pbuffer, u16 *buflen, struct cifs_ses *ses, const struct nls_table *nls_cp); diff --git a/fs/cifs/sess.c b/fs/cifs/sess.c index c3d086e2bd56..a42e99c8e00e 100644 --- a/fs/cifs/sess.c +++ b/fs/cifs/sess.c @@ -364,19 +364,43 @@ void build_ntlmssp_negotiate_blob(unsigned char *pbuffer, sec_blob->DomainName.MaximumLength = 0; } -/* We do not malloc the blob, it is passed in pbuffer, because its - maximum possible size is fixed and small, making this approach cleaner. - This function returns the length of the data in the blob */ -int build_ntlmssp_auth_blob(unsigned char *pbuffer, +static int size_of_ntlmssp_blob(struct cifs_ses *ses) +{ + int sz = sizeof(AUTHENTICATE_MESSAGE) + ses->auth_key.len + - CIFS_SESS_KEY_SIZE + CIFS_CPHTXT_SIZE + 2; + + if (ses->domainName) + sz += 2 * strnlen(ses->domainName, CIFS_MAX_DOMAINNAME_LEN); + else + sz += 2; + + if (ses->user_name) + sz += 2 * strnlen(ses->user_name, CIFS_MAX_USERNAME_LEN); + else + sz += 2; + + return sz; +} + +int build_ntlmssp_auth_blob(unsigned char **pbuffer, u16 *buflen, struct cifs_ses *ses, const struct nls_table *nls_cp) { int rc; - AUTHENTICATE_MESSAGE *sec_blob = (AUTHENTICATE_MESSAGE *)pbuffer; + AUTHENTICATE_MESSAGE *sec_blob; __u32 flags; unsigned char *tmp; + rc = setup_ntlmv2_rsp(ses, nls_cp); + if (rc) { + cifs_dbg(VFS, "Error %d during NTLMSSP authentication\n", rc); + *buflen = 0; + goto setup_ntlmv2_ret; + } + *pbuffer = kmalloc(size_of_ntlmssp_blob(ses), GFP_KERNEL); + sec_blob = (AUTHENTICATE_MESSAGE *)*pbuffer; + memcpy(sec_blob->Signature, NTLMSSP_SIGNATURE, 8); sec_blob->MessageType = NtLmAuthenticate; @@ -391,7 +415,7 @@ int build_ntlmssp_auth_blob(unsigned char *pbuffer, flags |= NTLMSSP_NEGOTIATE_KEY_XCH; } - tmp = pbuffer + sizeof(AUTHENTICATE_MESSAGE); + tmp = *pbuffer + sizeof(AUTHENTICATE_MESSAGE); sec_blob->NegotiateFlags = cpu_to_le32(flags); sec_blob->LmChallengeResponse.BufferOffset = @@ -399,13 +423,9 @@ int build_ntlmssp_auth_blob(unsigned char *pbuffer, sec_blob->LmChallengeResponse.Length = 0; sec_blob->LmChallengeResponse.MaximumLength = 0; - sec_blob->NtChallengeResponse.BufferOffset = cpu_to_le32(tmp - pbuffer); + sec_blob->NtChallengeResponse.BufferOffset = + cpu_to_le32(tmp - *pbuffer); if (ses->user_name != NULL) { - rc = setup_ntlmv2_rsp(ses, nls_cp); - if (rc) { - cifs_dbg(VFS, "Error %d during NTLMSSP authentication\n", rc); - goto setup_ntlmv2_ret; - } memcpy(tmp, ses->auth_key.response + CIFS_SESS_KEY_SIZE, ses->auth_key.len - CIFS_SESS_KEY_SIZE); tmp += ses->auth_key.len - CIFS_SESS_KEY_SIZE; @@ -423,7 +443,7 @@ int build_ntlmssp_auth_blob(unsigned char *pbuffer, } if (ses->domainName == NULL) { - sec_blob->DomainName.BufferOffset = cpu_to_le32(tmp - pbuffer); + sec_blob->DomainName.BufferOffset = cpu_to_le32(tmp - *pbuffer); sec_blob->DomainName.Length = 0; sec_blob->DomainName.MaximumLength = 0; tmp += 2; @@ -432,14 +452,14 @@ int build_ntlmssp_auth_blob(unsigned char *pbuffer, len = cifs_strtoUTF16((__le16 *)tmp, ses->domainName, CIFS_MAX_DOMAINNAME_LEN, nls_cp); len *= 2; /* unicode is 2 bytes each */ - sec_blob->DomainName.BufferOffset = cpu_to_le32(tmp - pbuffer); + sec_blob->DomainName.BufferOffset = cpu_to_le32(tmp - *pbuffer); sec_blob->DomainName.Length = cpu_to_le16(len); sec_blob->DomainName.MaximumLength = cpu_to_le16(len); tmp += len; } if (ses->user_name == NULL) { - sec_blob->UserName.BufferOffset = cpu_to_le32(tmp - pbuffer); + sec_blob->UserName.BufferOffset = cpu_to_le32(tmp - *pbuffer); sec_blob->UserName.Length = 0; sec_blob->UserName.MaximumLength = 0; tmp += 2; @@ -448,13 +468,13 @@ int build_ntlmssp_auth_blob(unsigned char *pbuffer, len = cifs_strtoUTF16((__le16 *)tmp, ses->user_name, CIFS_MAX_USERNAME_LEN, nls_cp); len *= 2; /* unicode is 2 bytes each */ - sec_blob->UserName.BufferOffset = cpu_to_le32(tmp - pbuffer); + sec_blob->UserName.BufferOffset = cpu_to_le32(tmp - *pbuffer); sec_blob->UserName.Length = cpu_to_le16(len); sec_blob->UserName.MaximumLength = cpu_to_le16(len); tmp += len; } - sec_blob->WorkstationName.BufferOffset = cpu_to_le32(tmp - pbuffer); + sec_blob->WorkstationName.BufferOffset = cpu_to_le32(tmp - *pbuffer); sec_blob->WorkstationName.Length = 0; sec_blob->WorkstationName.MaximumLength = 0; tmp += 2; @@ -463,19 +483,19 @@ int build_ntlmssp_auth_blob(unsigned char *pbuffer, (ses->ntlmssp->server_flags & NTLMSSP_NEGOTIATE_EXTENDED_SEC)) && !calc_seckey(ses)) { memcpy(tmp, ses->ntlmssp->ciphertext, CIFS_CPHTXT_SIZE); - sec_blob->SessionKey.BufferOffset = cpu_to_le32(tmp - pbuffer); + sec_blob->SessionKey.BufferOffset = cpu_to_le32(tmp - *pbuffer); sec_blob->SessionKey.Length = cpu_to_le16(CIFS_CPHTXT_SIZE); sec_blob->SessionKey.MaximumLength = cpu_to_le16(CIFS_CPHTXT_SIZE); tmp += CIFS_CPHTXT_SIZE; } else { - sec_blob->SessionKey.BufferOffset = cpu_to_le32(tmp - pbuffer); + sec_blob->SessionKey.BufferOffset = cpu_to_le32(tmp - *pbuffer); sec_blob->SessionKey.Length = 0; sec_blob->SessionKey.MaximumLength = 0; } + *buflen = tmp - *pbuffer; setup_ntlmv2_ret: - *buflen = tmp - pbuffer; return rc; } @@ -1266,7 +1286,7 @@ sess_auth_rawntlmssp_authenticate(struct sess_data *sess_data) struct cifs_ses *ses = sess_data->ses; __u16 bytes_remaining; char *bcc_ptr; - char *ntlmsspblob = NULL; + unsigned char *ntlmsspblob = NULL; u16 blob_len; cifs_dbg(FYI, "rawntlmssp session setup authenticate phase\n"); @@ -1279,19 +1299,7 @@ sess_auth_rawntlmssp_authenticate(struct sess_data *sess_data) /* Build security blob before we assemble the request */ pSMB = (SESSION_SETUP_ANDX *)sess_data->iov[0].iov_base; smb_buf = (struct smb_hdr *)pSMB; - /* - * 5 is an empirical value, large enough to hold - * authenticate message plus max 10 of av paris, - * domain, user, workstation names, flags, etc. - */ - ntlmsspblob = kzalloc(5*sizeof(struct _AUTHENTICATE_MESSAGE), - GFP_KERNEL); - if (!ntlmsspblob) { - rc = -ENOMEM; - goto out; - } - - rc = build_ntlmssp_auth_blob(ntlmsspblob, + rc = build_ntlmssp_auth_blob(&ntlmsspblob, &blob_len, ses, sess_data->nls_cp); if (rc) goto out_free_ntlmsspblob; diff --git a/fs/cifs/smb2pdu.c b/fs/cifs/smb2pdu.c index 8f38e33d365b..c3e61a7a7c7c 100644 --- a/fs/cifs/smb2pdu.c +++ b/fs/cifs/smb2pdu.c @@ -588,7 +588,7 @@ SMB2_sess_setup(const unsigned int xid, struct cifs_ses *ses, u16 blob_length = 0; struct key *spnego_key = NULL; char *security_blob = NULL; - char *ntlmssp_blob = NULL; + unsigned char *ntlmssp_blob = NULL; bool use_spnego = false; /* else use raw ntlmssp */ cifs_dbg(FYI, "Session Setup\n"); @@ -713,13 +713,7 @@ ssetup_ntlmssp_authenticate: iov[1].iov_len = blob_length; } else if (phase == NtLmAuthenticate) { req->hdr.SessionId = ses->Suid; - ntlmssp_blob = kzalloc(sizeof(struct _NEGOTIATE_MESSAGE) + 500, - GFP_KERNEL); - if (ntlmssp_blob == NULL) { - rc = -ENOMEM; - goto ssetup_exit; - } - rc = build_ntlmssp_auth_blob(ntlmssp_blob, &blob_length, ses, + rc = build_ntlmssp_auth_blob(&ntlmssp_blob, &blob_length, ses, nls_cp); if (rc) { cifs_dbg(FYI, "build_ntlmssp_auth_blob failed %d\n", -- cgit v1.3-8-gc7d7 From a6b6befbb2806697461962edb044e3376a771ebb Mon Sep 17 00:00:00 2001 From: Luis de Bethencourt Date: Wed, 8 Jun 2016 17:02:32 +0100 Subject: cifs: check hash calculating succeeded calc_lanman_hash() could return -ENOMEM or other errors, we should check that everything went fine before using the calculated key. Signed-off-by: Luis de Bethencourt Signed-off-by: Steve French --- fs/cifs/sess.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'fs') diff --git a/fs/cifs/sess.c b/fs/cifs/sess.c index a42e99c8e00e..538d9b55699a 100644 --- a/fs/cifs/sess.c +++ b/fs/cifs/sess.c @@ -710,6 +710,8 @@ sess_auth_lanman(struct sess_data *sess_data) rc = calc_lanman_hash(ses->password, ses->server->cryptkey, ses->server->sec_mode & SECMODE_PW_ENCRYPT ? true : false, lnm_session_key); + if (rc) + goto out; memcpy(bcc_ptr, (char *)lnm_session_key, CIFS_AUTH_RESP_SIZE); bcc_ptr += CIFS_AUTH_RESP_SIZE; -- cgit v1.3-8-gc7d7 From 485e71e8fb6356c08c7fc6bcce4bf02c9a9a663f Mon Sep 17 00:00:00 2001 From: Andreas Gruenbacher Date: Wed, 22 Jun 2016 23:57:25 +0200 Subject: posix_acl: Add set_posix_acl Factor out part of posix_acl_xattr_set into a common function that takes a posix_acl, which nfsd can also call. The prototype already exists in include/linux/posix_acl.h. Signed-off-by: Andreas Gruenbacher Cc: stable@vger.kernel.org Cc: Christoph Hellwig Cc: Al Viro Signed-off-by: J. Bruce Fields --- fs/posix_acl.c | 42 +++++++++++++++++++++++------------------- 1 file changed, 23 insertions(+), 19 deletions(-) (limited to 'fs') diff --git a/fs/posix_acl.c b/fs/posix_acl.c index 8a4a266beff3..edc452c2a563 100644 --- a/fs/posix_acl.c +++ b/fs/posix_acl.c @@ -820,39 +820,43 @@ posix_acl_xattr_get(const struct xattr_handler *handler, return error; } -static int -posix_acl_xattr_set(const struct xattr_handler *handler, - struct dentry *unused, struct inode *inode, - const char *name, const void *value, - size_t size, int flags) +int +set_posix_acl(struct inode *inode, int type, struct posix_acl *acl) { - struct posix_acl *acl = NULL; - int ret; - if (!IS_POSIXACL(inode)) return -EOPNOTSUPP; if (!inode->i_op->set_acl) return -EOPNOTSUPP; - if (handler->flags == ACL_TYPE_DEFAULT && !S_ISDIR(inode->i_mode)) - return value ? -EACCES : 0; + if (type == ACL_TYPE_DEFAULT && !S_ISDIR(inode->i_mode)) + return acl ? -EACCES : 0; if (!inode_owner_or_capable(inode)) return -EPERM; + if (acl) { + int ret = posix_acl_valid(acl); + if (ret) + return ret; + } + return inode->i_op->set_acl(inode, acl, type); +} +EXPORT_SYMBOL(set_posix_acl); + +static int +posix_acl_xattr_set(const struct xattr_handler *handler, + struct dentry *unused, struct inode *inode, + const char *name, const void *value, + size_t size, int flags) +{ + struct posix_acl *acl = NULL; + int ret; + if (value) { acl = posix_acl_from_xattr(&init_user_ns, value, size); if (IS_ERR(acl)) return PTR_ERR(acl); - - if (acl) { - ret = posix_acl_valid(acl); - if (ret) - goto out; - } } - - ret = inode->i_op->set_acl(inode, acl, handler->flags); -out: + ret = set_posix_acl(inode, handler->flags, acl); posix_acl_release(acl); return ret; } -- cgit v1.3-8-gc7d7 From 999653786df6954a31044528ac3f7a5dadca08f4 Mon Sep 17 00:00:00 2001 From: Ben Hutchings Date: Wed, 22 Jun 2016 19:43:35 +0100 Subject: nfsd: check permissions when setting ACLs Use set_posix_acl, which includes proper permission checks, instead of calling ->set_acl directly. Without this anyone may be able to grant themselves permissions to a file by setting the ACL. Lock the inode to make the new checks atomic with respect to set_acl. (Also, nfsd was the only caller of set_acl not locking the inode, so I suspect this may fix other races.) This also simplifies the code, and ensures our ACLs are checked by posix_acl_valid. The permission checks and the inode locking were lost with commit 4ac7249e, which changed nfsd to use the set_acl inode operation directly instead of going through xattr handlers. Reported-by: David Sinquin [agreunba@redhat.com: use set_posix_acl] Fixes: 4ac7249e Cc: Christoph Hellwig Cc: Al Viro Cc: stable@vger.kernel.org Signed-off-by: J. Bruce Fields --- fs/nfsd/nfs2acl.c | 20 ++++++++++---------- fs/nfsd/nfs3acl.c | 16 +++++++--------- fs/nfsd/nfs4acl.c | 16 ++++++++-------- 3 files changed, 25 insertions(+), 27 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/nfs2acl.c b/fs/nfsd/nfs2acl.c index 1580ea6fd64d..d08cd88155c7 100644 --- a/fs/nfsd/nfs2acl.c +++ b/fs/nfsd/nfs2acl.c @@ -104,22 +104,21 @@ static __be32 nfsacld_proc_setacl(struct svc_rqst * rqstp, goto out; inode = d_inode(fh->fh_dentry); - if (!IS_POSIXACL(inode) || !inode->i_op->set_acl) { - error = -EOPNOTSUPP; - goto out_errno; - } error = fh_want_write(fh); if (error) goto out_errno; - error = inode->i_op->set_acl(inode, argp->acl_access, ACL_TYPE_ACCESS); + fh_lock(fh); + + error = set_posix_acl(inode, ACL_TYPE_ACCESS, argp->acl_access); if (error) - goto out_drop_write; - error = inode->i_op->set_acl(inode, argp->acl_default, - ACL_TYPE_DEFAULT); + goto out_drop_lock; + error = set_posix_acl(inode, ACL_TYPE_DEFAULT, argp->acl_default); if (error) - goto out_drop_write; + goto out_drop_lock; + + fh_unlock(fh); fh_drop_write(fh); @@ -131,7 +130,8 @@ out: posix_acl_release(argp->acl_access); posix_acl_release(argp->acl_default); return nfserr; -out_drop_write: +out_drop_lock: + fh_unlock(fh); fh_drop_write(fh); out_errno: nfserr = nfserrno(error); diff --git a/fs/nfsd/nfs3acl.c b/fs/nfsd/nfs3acl.c index 01df4cd7c753..0c890347cde3 100644 --- a/fs/nfsd/nfs3acl.c +++ b/fs/nfsd/nfs3acl.c @@ -95,22 +95,20 @@ static __be32 nfsd3_proc_setacl(struct svc_rqst * rqstp, goto out; inode = d_inode(fh->fh_dentry); - if (!IS_POSIXACL(inode) || !inode->i_op->set_acl) { - error = -EOPNOTSUPP; - goto out_errno; - } error = fh_want_write(fh); if (error) goto out_errno; - error = inode->i_op->set_acl(inode, argp->acl_access, ACL_TYPE_ACCESS); + fh_lock(fh); + + error = set_posix_acl(inode, ACL_TYPE_ACCESS, argp->acl_access); if (error) - goto out_drop_write; - error = inode->i_op->set_acl(inode, argp->acl_default, - ACL_TYPE_DEFAULT); + goto out_drop_lock; + error = set_posix_acl(inode, ACL_TYPE_DEFAULT, argp->acl_default); -out_drop_write: +out_drop_lock: + fh_unlock(fh); fh_drop_write(fh); out_errno: nfserr = nfserrno(error); diff --git a/fs/nfsd/nfs4acl.c b/fs/nfsd/nfs4acl.c index 6adabd6049b7..71292a0d6f09 100644 --- a/fs/nfsd/nfs4acl.c +++ b/fs/nfsd/nfs4acl.c @@ -770,9 +770,6 @@ nfsd4_set_nfs4_acl(struct svc_rqst *rqstp, struct svc_fh *fhp, dentry = fhp->fh_dentry; inode = d_inode(dentry); - if (!inode->i_op->set_acl || !IS_POSIXACL(inode)) - return nfserr_attrnotsupp; - if (S_ISDIR(inode->i_mode)) flags = NFS4_ACL_DIR; @@ -782,16 +779,19 @@ nfsd4_set_nfs4_acl(struct svc_rqst *rqstp, struct svc_fh *fhp, if (host_error < 0) goto out_nfserr; - host_error = inode->i_op->set_acl(inode, pacl, ACL_TYPE_ACCESS); + fh_lock(fhp); + + host_error = set_posix_acl(inode, ACL_TYPE_ACCESS, pacl); if (host_error < 0) - goto out_release; + goto out_drop_lock; if (S_ISDIR(inode->i_mode)) { - host_error = inode->i_op->set_acl(inode, dpacl, - ACL_TYPE_DEFAULT); + host_error = set_posix_acl(inode, ACL_TYPE_DEFAULT, dpacl); } -out_release: +out_drop_lock: + fh_unlock(fhp); + posix_acl_release(pacl); posix_acl_release(dpacl); out_nfserr: -- cgit v1.3-8-gc7d7 From 4fcd1813e6404dd4420c7d12fb483f9320f0bf93 Mon Sep 17 00:00:00 2001 From: Steve French Date: Wed, 22 Jun 2016 20:12:05 -0500 Subject: Fix reconnect to not defer smb3 session reconnect long after socket reconnect Azure server blocks clients that open a socket and don't do anything on it. In our reconnect scenarios, we can reconnect the tcp session and detect the socket is available but we defer the negprot and SMB3 session setup and tree connect reconnection until the next i/o is requested, but this looks suspicous to some servers who expect SMB3 negprog and session setup soon after a socket is created. In the echo thread, reconnect SMB3 sessions and tree connections that are disconnected. A later patch will replay persistent (and resilient) handle opens. CC: Stable Signed-off-by: Steve French Acked-by: Pavel Shilovsky --- fs/cifs/connect.c | 4 +++- fs/cifs/smb2pdu.c | 27 +++++++++++++++++++++++++++ 2 files changed, 30 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c index 66736f57b5ab..7d2b15c06090 100644 --- a/fs/cifs/connect.c +++ b/fs/cifs/connect.c @@ -428,7 +428,9 @@ cifs_echo_request(struct work_struct *work) * server->ops->need_neg() == true. Also, no need to ping if * we got a response recently. */ - if (!server->ops->need_neg || server->ops->need_neg(server) || + + if (server->tcpStatus == CifsNeedReconnect || + server->tcpStatus == CifsExiting || server->tcpStatus == CifsNew || (server->ops->can_echo && !server->ops->can_echo(server)) || time_before(jiffies, server->lstrp + echo_interval - HZ)) goto requeue_echo; diff --git a/fs/cifs/smb2pdu.c b/fs/cifs/smb2pdu.c index c3e61a7a7c7c..29e06db5f187 100644 --- a/fs/cifs/smb2pdu.c +++ b/fs/cifs/smb2pdu.c @@ -1812,6 +1812,33 @@ SMB2_echo(struct TCP_Server_Info *server) cifs_dbg(FYI, "In echo request\n"); + if (server->tcpStatus == CifsNeedNegotiate) { + struct list_head *tmp, *tmp2; + struct cifs_ses *ses; + struct cifs_tcon *tcon; + + cifs_dbg(FYI, "Need negotiate, reconnecting tcons\n"); + spin_lock(&cifs_tcp_ses_lock); + list_for_each(tmp, &server->smb_ses_list) { + ses = list_entry(tmp, struct cifs_ses, smb_ses_list); + list_for_each(tmp2, &ses->tcon_list) { + tcon = list_entry(tmp2, struct cifs_tcon, + tcon_list); + /* add check for persistent handle reconnect */ + if (tcon && tcon->need_reconnect) { + spin_unlock(&cifs_tcp_ses_lock); + rc = smb2_reconnect(SMB2_ECHO, tcon); + spin_lock(&cifs_tcp_ses_lock); + } + } + } + spin_unlock(&cifs_tcp_ses_lock); + } + + /* if no session, renegotiate failed above */ + if (server->tcpStatus == CifsNeedNegotiate) + return -EIO; + rc = small_smb2_init(SMB2_ECHO, NULL, (void **)&req); if (rc) return rc; -- cgit v1.3-8-gc7d7 From 45e8a2583d97ca758a55c608f78c4cef562644d1 Mon Sep 17 00:00:00 2001 From: Steve French Date: Wed, 22 Jun 2016 21:07:32 -0500 Subject: File names with trailing period or space need special case conversion POSIX allows files with trailing spaces or a trailing period but SMB3 does not, so convert these using the normal Services For Mac mapping as we do for other reserved characters such as : < > | ? * This is similar to what Macs do for the same problem over SMB3. CC: Stable Signed-off-by: Steve French Acked-by: Pavel Shilovsky --- fs/cifs/cifs_unicode.c | 33 +++++++++++++++++++++++++++++---- fs/cifs/cifs_unicode.h | 2 ++ 2 files changed, 31 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/cifs/cifs_unicode.c b/fs/cifs/cifs_unicode.c index 5a53ac6b1e02..02b071bf3732 100644 --- a/fs/cifs/cifs_unicode.c +++ b/fs/cifs/cifs_unicode.c @@ -101,6 +101,12 @@ convert_sfm_char(const __u16 src_char, char *target) case SFM_SLASH: *target = '\\'; break; + case SFM_SPACE: + *target = ' '; + break; + case SFM_PERIOD: + *target = '.'; + break; default: return false; } @@ -404,7 +410,7 @@ static __le16 convert_to_sfu_char(char src_char) return dest_char; } -static __le16 convert_to_sfm_char(char src_char) +static __le16 convert_to_sfm_char(char src_char, bool end_of_string) { __le16 dest_char; @@ -427,6 +433,18 @@ static __le16 convert_to_sfm_char(char src_char) case '|': dest_char = cpu_to_le16(SFM_PIPE); break; + case '.': + if (end_of_string) + dest_char = cpu_to_le16(SFM_PERIOD); + else + dest_char = 0; + break; + case ' ': + if (end_of_string) + dest_char = cpu_to_le16(SFM_SPACE); + else + dest_char = 0; + break; default: dest_char = 0; } @@ -469,9 +487,16 @@ cifsConvertToUTF16(__le16 *target, const char *source, int srclen, /* see if we must remap this char */ if (map_chars == SFU_MAP_UNI_RSVD) dst_char = convert_to_sfu_char(src_char); - else if (map_chars == SFM_MAP_UNI_RSVD) - dst_char = convert_to_sfm_char(src_char); - else + else if (map_chars == SFM_MAP_UNI_RSVD) { + bool end_of_string; + + if (i == srclen - 1) + end_of_string = true; + else + end_of_string = false; + + dst_char = convert_to_sfm_char(src_char, end_of_string); + } else dst_char = 0; /* * FIXME: We can not handle remapping backslash (UNI_SLASH) diff --git a/fs/cifs/cifs_unicode.h b/fs/cifs/cifs_unicode.h index bdc52cb9a676..479bc0a941f3 100644 --- a/fs/cifs/cifs_unicode.h +++ b/fs/cifs/cifs_unicode.h @@ -64,6 +64,8 @@ #define SFM_LESSTHAN ((__u16) 0xF023) #define SFM_PIPE ((__u16) 0xF027) #define SFM_SLASH ((__u16) 0xF026) +#define SFM_PERIOD ((__u16) 0xF028) +#define SFM_SPACE ((__u16) 0xF029) /* * Mapping mechanism to use when one of the seven reserved characters is -- cgit v1.3-8-gc7d7 From f2db19719a4e789a58ac024b43f12eeb9e458074 Mon Sep 17 00:00:00 2001 From: Michal Hocko Date: Fri, 24 Jun 2016 14:49:28 -0700 Subject: jbd2: get rid of superfluous __GFP_REPEAT jbd2_alloc is explicit about its allocation preferences wrt. the allocation size. Sub page allocations go to the slab allocator and larger are using either the page allocator or vmalloc. This is all good but the logic is unnecessarily complex. 1) as per Ted, the vmalloc fallback is a left-over: : jbd2_alloc is only passed in the bh->b_size, which can't be PAGE_SIZE, so : the code path that calls vmalloc() should never get called. When we : conveted jbd2_alloc() to suppor sub-page size allocations in commit : d2eecb039368, there was an assumption that it could be called with a size : greater than PAGE_SIZE, but that's certaily not true today. Moreover vmalloc allocation might even lead to a deadlock because the callers expect GFP_NOFS context while vmalloc is GFP_KERNEL. 2) __GFP_REPEAT for requests <= PAGE_ALLOC_COSTLY_ORDER is ignored since the flag was introduced. Let's simplify the code flow and use the slab allocator for sub-page requests and the page allocator for others. Even though order > 0 is not currently used as per above leave that option open. Link: http://lkml.kernel.org/r/1464599699-30131-18-git-send-email-mhocko@kernel.org Signed-off-by: Michal Hocko Reviewed-by: Jan Kara Cc: "Theodore Ts'o" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/jbd2/journal.c | 32 +++++++------------------------- 1 file changed, 7 insertions(+), 25 deletions(-) (limited to 'fs') diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c index b31852f76f46..e3ca4b4cac84 100644 --- a/fs/jbd2/journal.c +++ b/fs/jbd2/journal.c @@ -2329,18 +2329,10 @@ void *jbd2_alloc(size_t size, gfp_t flags) BUG_ON(size & (size-1)); /* Must be a power of 2 */ - flags |= __GFP_REPEAT; - if (size == PAGE_SIZE) - ptr = (void *)__get_free_pages(flags, 0); - else if (size > PAGE_SIZE) { - int order = get_order(size); - - if (order < 3) - ptr = (void *)__get_free_pages(flags, order); - else - ptr = vmalloc(size); - } else + if (size < PAGE_SIZE) ptr = kmem_cache_alloc(get_slab(size), flags); + else + ptr = (void *)__get_free_pages(flags, get_order(size)); /* Check alignment; SLUB has gotten this wrong in the past, * and this can lead to user data corruption! */ @@ -2351,20 +2343,10 @@ void *jbd2_alloc(size_t size, gfp_t flags) void jbd2_free(void *ptr, size_t size) { - if (size == PAGE_SIZE) { - free_pages((unsigned long)ptr, 0); - return; - } - if (size > PAGE_SIZE) { - int order = get_order(size); - - if (order < 3) - free_pages((unsigned long)ptr, order); - else - vfree(ptr); - return; - } - kmem_cache_free(get_slab(size), ptr); + if (size < PAGE_SIZE) + kmem_cache_free(get_slab(size), ptr); + else + free_pages((unsigned long)ptr, get_order(size)); }; /* -- cgit v1.3-8-gc7d7 From 7186ee06b66313dae0d34ec5241fda7c4a537cb9 Mon Sep 17 00:00:00 2001 From: Gang He Date: Fri, 24 Jun 2016 14:50:13 -0700 Subject: ocfs2: disable BUG assertions in reading blocks According to some high-load testing, these two BUG assertions were encountered, this led system panic. Actually, there were some discussions about removing these two BUG() assertions, it would not bring any side effect. Then, I did the the following changes, 1) use the existing macro CATCH_BH_JBD_RACES to wrap BUG() in the ocfs2_read_blocks_sync function like before. 2) disable the macro CATCH_BH_JBD_RACES in Makefile by default. Link: http://lkml.kernel.org/r/1466574294-26863-1-git-send-email-ghe@suse.com Signed-off-by: Gang He Cc: Mark Fasheh Cc: Joel Becker Cc: Junxiao Bi Cc: Joseph Qi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ocfs2/Makefile | 2 -- fs/ocfs2/buffer_head_io.c | 5 +++++ 2 files changed, 5 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/Makefile b/fs/ocfs2/Makefile index e27e6527912b..4342c7ee7d20 100644 --- a/fs/ocfs2/Makefile +++ b/fs/ocfs2/Makefile @@ -1,7 +1,5 @@ ccflags-y := -Ifs/ocfs2 -ccflags-y += -DCATCH_BH_JBD_RACES - obj-$(CONFIG_OCFS2_FS) += \ ocfs2.o \ ocfs2_stackglue.o diff --git a/fs/ocfs2/buffer_head_io.c b/fs/ocfs2/buffer_head_io.c index fe50ded1b4ce..498641eed2db 100644 --- a/fs/ocfs2/buffer_head_io.c +++ b/fs/ocfs2/buffer_head_io.c @@ -139,11 +139,16 @@ int ocfs2_read_blocks_sync(struct ocfs2_super *osb, u64 block, lock_buffer(bh); if (buffer_jbd(bh)) { +#ifdef CATCH_BH_JBD_RACES mlog(ML_ERROR, "block %llu had the JBD bit set " "while I was in lock_buffer!", (unsigned long long)bh->b_blocknr); BUG(); +#else + unlock_buffer(bh); + continue; +#endif } clear_buffer_uptodate(bh); -- cgit v1.3-8-gc7d7 From 63d2f95d63396059200c391ca87161897b99e74a Mon Sep 17 00:00:00 2001 From: Torsten Hilbrich Date: Fri, 24 Jun 2016 14:50:18 -0700 Subject: fs/nilfs2: fix potential underflow in call to crc32_le The value `bytes' comes from the filesystem which is about to be mounted. We cannot trust that the value is always in the range we expect it to be. Check its value before using it to calculate the length for the crc32_le call. It value must be larger (or equal) sumoff + 4. This fixes a kernel bug when accidentially mounting an image file which had the nilfs2 magic value 0x3434 at the right offset 0x406 by chance. The bytes 0x01 0x00 were stored at 0x408 and were interpreted as a s_bytes value of 1. This caused an underflow when substracting sumoff + 4 (20) in the call to crc32_le. BUG: unable to handle kernel paging request at ffff88021e600000 IP: crc32_le+0x36/0x100 ... Call Trace: nilfs_valid_sb.part.5+0x52/0x60 [nilfs2] nilfs_load_super_block+0x142/0x300 [nilfs2] init_nilfs+0x60/0x390 [nilfs2] nilfs_mount+0x302/0x520 [nilfs2] mount_fs+0x38/0x160 vfs_kern_mount+0x67/0x110 do_mount+0x269/0xe00 SyS_mount+0x9f/0x100 entry_SYSCALL_64_fastpath+0x16/0x71 Link: http://lkml.kernel.org/r/1466778587-5184-2-git-send-email-konishi.ryusuke@lab.ntt.co.jp Signed-off-by: Torsten Hilbrich Tested-by: Torsten Hilbrich Signed-off-by: Ryusuke Konishi Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nilfs2/the_nilfs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nilfs2/the_nilfs.c b/fs/nilfs2/the_nilfs.c index 809bd2de7ad0..e9fd241b9a0a 100644 --- a/fs/nilfs2/the_nilfs.c +++ b/fs/nilfs2/the_nilfs.c @@ -439,7 +439,7 @@ static int nilfs_valid_sb(struct nilfs_super_block *sbp) if (!sbp || le16_to_cpu(sbp->s_magic) != NILFS_SUPER_MAGIC) return 0; bytes = le16_to_cpu(sbp->s_bytes); - if (bytes > BLOCK_SIZE) + if (bytes < sumoff + 4 || bytes > BLOCK_SIZE) return 0; crc = crc32_le(le32_to_cpu(sbp->s_crc_seed), (unsigned char *)sbp, sumoff); -- cgit v1.3-8-gc7d7 From 5a9294e5c535deab69831076af15cd35e1c95f8b Mon Sep 17 00:00:00 2001 From: Andrey Vagin Date: Fri, 24 Jun 2016 14:50:27 -0700 Subject: autofs: don't get stuck in a loop if vfs_write() returns an error __vfs_write() returns a negative value in a error case. Link: http://lkml.kernel.org/r/20160616083108.6278.65815.stgit@pluto.themaw.net Signed-off-by: Andrey Vagin Signed-off-by: Ian Kent Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/autofs4/waitq.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/autofs4/waitq.c b/fs/autofs4/waitq.c index 0146d911f468..631f1554c87b 100644 --- a/fs/autofs4/waitq.c +++ b/fs/autofs4/waitq.c @@ -66,11 +66,12 @@ static int autofs4_write(struct autofs_sb_info *sbi, set_fs(KERNEL_DS); mutex_lock(&sbi->pipe_mutex); - wr = __vfs_write(file, data, bytes, &file->f_pos); - while (bytes && wr) { + while (bytes) { + wr = __vfs_write(file, data, bytes, &file->f_pos); + if (wr <= 0) + break; data += wr; bytes -= wr; - wr = __vfs_write(file, data, bytes, &file->f_pos); } mutex_unlock(&sbi->pipe_mutex); -- cgit v1.3-8-gc7d7 From 02dbfc99b424dde3cf0a492ed3bec4f222441754 Mon Sep 17 00:00:00 2001 From: Omar Sandoval Date: Fri, 20 May 2016 13:50:33 -0700 Subject: Btrfs: fix ->iterate_shared() by upgrading i_rwsem for delayed nodes Commit fe742fd4f90f ("Revert "btrfs: switch to ->iterate_shared()"") backed out the conversion to ->iterate_shared() for Btrfs because the delayed inode handling in btrfs_real_readdir() is racy. However, we can still do readdir in parallel if there are no delayed nodes. This is a temporary fix which upgrades the shared inode lock to an exclusive lock only when we have delayed items until we come up with a more complete solution. While we're here, rename the btrfs_{get,put}_delayed_items functions to make it very clear that they're just for readdir. Tested with xfstests and by doing a parallel kernel build: while make tinyconfig && make -j4 && git clean dqfx; do : done along with a bunch of parallel finds in another shell: while true; do for ((i=0; i<4; i++)); do find . >/dev/null & done wait done Signed-off-by: Omar Sandoval Signed-off-by: David Sterba Signed-off-by: Chris Mason --- fs/btrfs/delayed-inode.c | 27 ++++++++++++++++++++++----- fs/btrfs/delayed-inode.h | 10 ++++++---- fs/btrfs/inode.c | 10 ++++++---- 3 files changed, 34 insertions(+), 13 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/delayed-inode.c b/fs/btrfs/delayed-inode.c index 61561c2a3f96..d3aaabbfada0 100644 --- a/fs/btrfs/delayed-inode.c +++ b/fs/btrfs/delayed-inode.c @@ -1606,15 +1606,23 @@ int btrfs_inode_delayed_dir_index_count(struct inode *inode) return 0; } -void btrfs_get_delayed_items(struct inode *inode, struct list_head *ins_list, - struct list_head *del_list) +bool btrfs_readdir_get_delayed_items(struct inode *inode, + struct list_head *ins_list, + struct list_head *del_list) { struct btrfs_delayed_node *delayed_node; struct btrfs_delayed_item *item; delayed_node = btrfs_get_delayed_node(inode); if (!delayed_node) - return; + return false; + + /* + * We can only do one readdir with delayed items at a time because of + * item->readdir_list. + */ + inode_unlock_shared(inode); + inode_lock(inode); mutex_lock(&delayed_node->mutex); item = __btrfs_first_delayed_insertion_item(delayed_node); @@ -1641,10 +1649,13 @@ void btrfs_get_delayed_items(struct inode *inode, struct list_head *ins_list, * requeue or dequeue this delayed node. */ atomic_dec(&delayed_node->refs); + + return true; } -void btrfs_put_delayed_items(struct list_head *ins_list, - struct list_head *del_list) +void btrfs_readdir_put_delayed_items(struct inode *inode, + struct list_head *ins_list, + struct list_head *del_list) { struct btrfs_delayed_item *curr, *next; @@ -1659,6 +1670,12 @@ void btrfs_put_delayed_items(struct list_head *ins_list, if (atomic_dec_and_test(&curr->refs)) kfree(curr); } + + /* + * The VFS is going to do up_read(), so we need to downgrade back to a + * read lock. + */ + downgrade_write(&inode->i_rwsem); } int btrfs_should_delete_dir_index(struct list_head *del_list, diff --git a/fs/btrfs/delayed-inode.h b/fs/btrfs/delayed-inode.h index 0167853c84ae..2495b3d4075f 100644 --- a/fs/btrfs/delayed-inode.h +++ b/fs/btrfs/delayed-inode.h @@ -137,10 +137,12 @@ void btrfs_kill_all_delayed_nodes(struct btrfs_root *root); void btrfs_destroy_delayed_inodes(struct btrfs_root *root); /* Used for readdir() */ -void btrfs_get_delayed_items(struct inode *inode, struct list_head *ins_list, - struct list_head *del_list); -void btrfs_put_delayed_items(struct list_head *ins_list, - struct list_head *del_list); +bool btrfs_readdir_get_delayed_items(struct inode *inode, + struct list_head *ins_list, + struct list_head *del_list); +void btrfs_readdir_put_delayed_items(struct inode *inode, + struct list_head *ins_list, + struct list_head *del_list); int btrfs_should_delete_dir_index(struct list_head *del_list, u64 index); int btrfs_readdir_delayed_dir_index(struct dir_context *ctx, diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index d2be95cfb6d1..969a25c5abcb 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -5757,6 +5757,7 @@ static int btrfs_real_readdir(struct file *file, struct dir_context *ctx) int name_len; int is_curr = 0; /* ctx->pos points to the current index? */ bool emitted; + bool put = false; /* FIXME, use a real flag for deciding about the key type */ if (root->fs_info->tree_root == root) @@ -5774,7 +5775,8 @@ static int btrfs_real_readdir(struct file *file, struct dir_context *ctx) if (key_type == BTRFS_DIR_INDEX_KEY) { INIT_LIST_HEAD(&ins_list); INIT_LIST_HEAD(&del_list); - btrfs_get_delayed_items(inode, &ins_list, &del_list); + put = btrfs_readdir_get_delayed_items(inode, &ins_list, + &del_list); } key.type = key_type; @@ -5921,8 +5923,8 @@ next: nopos: ret = 0; err: - if (key_type == BTRFS_DIR_INDEX_KEY) - btrfs_put_delayed_items(&ins_list, &del_list); + if (put) + btrfs_readdir_put_delayed_items(inode, &ins_list, &del_list); btrfs_free_path(path); return ret; } @@ -10534,7 +10536,7 @@ static const struct inode_operations btrfs_dir_ro_inode_operations = { static const struct file_operations btrfs_dir_file_operations = { .llseek = generic_file_llseek, .read = generic_read_dir, - .iterate = btrfs_real_readdir, + .iterate_shared = btrfs_real_readdir, .unlocked_ioctl = btrfs_ioctl, #ifdef CONFIG_COMPAT .compat_ioctl = btrfs_compat_ioctl, -- cgit v1.3-8-gc7d7