diff options
Diffstat (limited to '')
-rw-r--r-- | fs/ceph/Kconfig | 3 | ||||
-rw-r--r-- | fs/ceph/Makefile | 2 | ||||
-rw-r--r-- | fs/ceph/acl.c | 13 | ||||
-rw-r--r-- | fs/ceph/addr.c | 1208 | ||||
-rw-r--r-- | fs/ceph/cache.c | 336 | ||||
-rw-r--r-- | fs/ceph/cache.h | 134 | ||||
-rw-r--r-- | fs/ceph/caps.c | 1828 | ||||
-rw-r--r-- | fs/ceph/debugfs.c | 197 | ||||
-rw-r--r-- | fs/ceph/dir.c | 357 | ||||
-rw-r--r-- | fs/ceph/export.c | 47 | ||||
-rw-r--r-- | fs/ceph/file.c | 993 | ||||
-rw-r--r-- | fs/ceph/inode.c | 589 | ||||
-rw-r--r-- | fs/ceph/io.c | 2 | ||||
-rw-r--r-- | fs/ceph/ioctl.c | 2 | ||||
-rw-r--r-- | fs/ceph/locks.c | 61 | ||||
-rw-r--r-- | fs/ceph/mds_client.c | 1647 | ||||
-rw-r--r-- | fs/ceph/mds_client.h | 96 | ||||
-rw-r--r-- | fs/ceph/mdsmap.c | 75 | ||||
-rw-r--r-- | fs/ceph/metric.c | 353 | ||||
-rw-r--r-- | fs/ceph/metric.h | 244 | ||||
-rw-r--r-- | fs/ceph/quota.c | 69 | ||||
-rw-r--r-- | fs/ceph/snap.c | 413 | ||||
-rw-r--r-- | fs/ceph/strings.c | 2 | ||||
-rw-r--r-- | fs/ceph/super.c | 346 | ||||
-rw-r--r-- | fs/ceph/super.h | 415 | ||||
-rw-r--r-- | fs/ceph/xattr.c | 175 |
26 files changed, 6383 insertions, 3224 deletions
diff --git a/fs/ceph/Kconfig b/fs/ceph/Kconfig index cf235f6eacf9..94df854147d3 100644 --- a/fs/ceph/Kconfig +++ b/fs/ceph/Kconfig @@ -6,6 +6,7 @@ config CEPH_FS select LIBCRC32C select CRYPTO_AES select CRYPTO + select NETFS_SUPPORT default n help Choose Y or M here to include support for mounting the @@ -13,7 +14,7 @@ config CEPH_FS scalable file system designed to provide high performance, reliable access to petabytes of storage. - More information at http://ceph.newdream.net/. + More information at https://ceph.io/. If unsure, say N. diff --git a/fs/ceph/Makefile b/fs/ceph/Makefile index 0a0823d378db..50c635dc7f71 100644 --- a/fs/ceph/Makefile +++ b/fs/ceph/Makefile @@ -8,7 +8,7 @@ obj-$(CONFIG_CEPH_FS) += ceph.o ceph-y := super.o inode.o dir.o file.o locks.o addr.o ioctl.o \ export.o caps.o snap.o xattr.o quota.o io.o \ mds_client.o mdsmap.o strings.o ceph_frag.o \ - debugfs.o util.o + debugfs.o util.o metric.o ceph-$(CONFIG_CEPH_FSCACHE) += cache.o ceph-$(CONFIG_CEPH_FS_POSIX_ACL) += acl.o diff --git a/fs/ceph/acl.c b/fs/ceph/acl.c index 26be6520d3fb..f4fc8e0b847c 100644 --- a/fs/ceph/acl.c +++ b/fs/ceph/acl.c @@ -22,14 +22,14 @@ static inline void ceph_set_cached_acl(struct inode *inode, struct ceph_inode_info *ci = ceph_inode(inode); spin_lock(&ci->i_ceph_lock); - if (__ceph_caps_issued_mask(ci, CEPH_CAP_XATTR_SHARED, 0)) + if (__ceph_caps_issued_mask_metric(ci, CEPH_CAP_XATTR_SHARED, 0)) set_cached_acl(inode, type, acl); else forget_cached_acl(inode, type); spin_unlock(&ci->i_ceph_lock); } -struct posix_acl *ceph_get_acl(struct inode *inode, int type) +struct posix_acl *ceph_get_acl(struct inode *inode, int type, bool rcu) { int size; unsigned int retry_cnt = 0; @@ -37,6 +37,9 @@ struct posix_acl *ceph_get_acl(struct inode *inode, int type) char *value = NULL; struct posix_acl *acl; + if (rcu) + return ERR_PTR(-ECHILD); + switch (type) { case ACL_TYPE_ACCESS: name = XATTR_NAME_POSIX_ACL_ACCESS; @@ -82,7 +85,8 @@ retry: return acl; } -int ceph_set_acl(struct inode *inode, struct posix_acl *acl, int type) +int ceph_set_acl(struct user_namespace *mnt_userns, struct inode *inode, + struct posix_acl *acl, int type) { int ret = 0, size = 0; const char *name = NULL; @@ -100,7 +104,8 @@ int ceph_set_acl(struct inode *inode, struct posix_acl *acl, int type) case ACL_TYPE_ACCESS: name = XATTR_NAME_POSIX_ACL_ACCESS; if (acl) { - ret = posix_acl_update_mode(inode, &new_mode, &acl); + ret = posix_acl_update_mode(&init_user_ns, inode, + &new_mode, &acl); if (ret) goto out; } diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c index 7ab616601141..dcf701b05cc1 100644 --- a/fs/ceph/addr.c +++ b/fs/ceph/addr.c @@ -4,17 +4,20 @@ #include <linux/backing-dev.h> #include <linux/fs.h> #include <linux/mm.h> +#include <linux/swap.h> #include <linux/pagemap.h> -#include <linux/writeback.h> /* generic_writepages */ #include <linux/slab.h> #include <linux/pagevec.h> #include <linux/task_io_accounting_ops.h> #include <linux/signal.h> #include <linux/iversion.h> +#include <linux/ktime.h> +#include <linux/netfs.h> #include "super.h" #include "mds_client.h" #include "cache.h" +#include "metric.h" #include <linux/ceph/osd_client.h> #include <linux/ceph/striper.h> @@ -59,6 +62,9 @@ (CONGESTION_ON_THRESH(congestion_kb) - \ (CONGESTION_ON_THRESH(congestion_kb) >> 2)) +static int ceph_netfs_check_write_begin(struct file *file, loff_t pos, unsigned int len, + struct folio **foliop, void **_fsdata); + static inline struct ceph_snap_context *page_snap_context(struct page *page) { if (PagePrivate(page)) @@ -70,22 +76,17 @@ static inline struct ceph_snap_context *page_snap_context(struct page *page) * Dirty a page. Optimistically adjust accounting, on the assumption * that we won't race with invalidate. If we do, readjust. */ -static int ceph_set_page_dirty(struct page *page) +static bool ceph_dirty_folio(struct address_space *mapping, struct folio *folio) { - struct address_space *mapping = page->mapping; struct inode *inode; struct ceph_inode_info *ci; struct ceph_snap_context *snapc; - int ret; - - if (unlikely(!mapping)) - return !TestSetPageDirty(page); - if (PageDirty(page)) { - dout("%p set_page_dirty %p idx %lu -- already dirty\n", - mapping->host, page, page->index); - BUG_ON(!PagePrivate(page)); - return 0; + if (folio_test_dirty(folio)) { + dout("%p dirty_folio %p idx %lu -- already dirty\n", + mapping->host, folio, folio->index); + VM_BUG_ON_FOLIO(!folio_test_private(folio), folio); + return false; } inode = mapping->host; @@ -109,369 +110,338 @@ static int ceph_set_page_dirty(struct page *page) if (ci->i_wrbuffer_ref == 0) ihold(inode); ++ci->i_wrbuffer_ref; - dout("%p set_page_dirty %p idx %lu head %d/%d -> %d/%d " + dout("%p dirty_folio %p idx %lu head %d/%d -> %d/%d " "snapc %p seq %lld (%d snaps)\n", - mapping->host, page, page->index, + mapping->host, folio, folio->index, ci->i_wrbuffer_ref-1, ci->i_wrbuffer_ref_head-1, ci->i_wrbuffer_ref, ci->i_wrbuffer_ref_head, snapc, snapc->seq, snapc->num_snaps); spin_unlock(&ci->i_ceph_lock); /* - * Reference snap context in page->private. Also set - * PagePrivate so that we get invalidatepage callback. + * Reference snap context in folio->private. Also set + * PagePrivate so that we get invalidate_folio callback. */ - BUG_ON(PagePrivate(page)); - page->private = (unsigned long)snapc; - SetPagePrivate(page); + VM_WARN_ON_FOLIO(folio->private, folio); + folio_attach_private(folio, snapc); - ret = __set_page_dirty_nobuffers(page); - WARN_ON(!PageLocked(page)); - WARN_ON(!page->mapping); - - return ret; + return ceph_fscache_dirty_folio(mapping, folio); } /* - * If we are truncating the full page (i.e. offset == 0), adjust the - * dirty page counters appropriately. Only called if there is private - * data on the page. + * If we are truncating the full folio (i.e. offset == 0), adjust the + * dirty folio counters appropriately. Only called if there is private + * data on the folio. */ -static void ceph_invalidatepage(struct page *page, unsigned int offset, - unsigned int length) +static void ceph_invalidate_folio(struct folio *folio, size_t offset, + size_t length) { struct inode *inode; struct ceph_inode_info *ci; - struct ceph_snap_context *snapc = page_snap_context(page); + struct ceph_snap_context *snapc; - inode = page->mapping->host; + inode = folio->mapping->host; ci = ceph_inode(inode); - if (offset != 0 || length != PAGE_SIZE) { - dout("%p invalidatepage %p idx %lu partial dirty page %u~%u\n", - inode, page, page->index, offset, length); + if (offset != 0 || length != folio_size(folio)) { + dout("%p invalidate_folio idx %lu partial dirty page %zu~%zu\n", + inode, folio->index, offset, length); return; } - ceph_invalidate_fscache_page(inode, page); + WARN_ON(!folio_test_locked(folio)); + if (folio_test_private(folio)) { + dout("%p invalidate_folio idx %lu full dirty page\n", + inode, folio->index); - WARN_ON(!PageLocked(page)); - if (!PagePrivate(page)) - return; + snapc = folio_detach_private(folio); + ceph_put_wrbuffer_cap_refs(ci, 1, snapc); + ceph_put_snap_context(snapc); + } + + folio_wait_fscache(folio); +} - ClearPageChecked(page); +static bool ceph_release_folio(struct folio *folio, gfp_t gfp) +{ + struct inode *inode = folio->mapping->host; - dout("%p invalidatepage %p idx %lu full dirty page\n", - inode, page, page->index); + dout("%llx:%llx release_folio idx %lu (%sdirty)\n", + ceph_vinop(inode), + folio->index, folio_test_dirty(folio) ? "" : "not "); - ceph_put_wrbuffer_cap_refs(ci, 1, snapc); - ceph_put_snap_context(snapc); - page->private = 0; - ClearPagePrivate(page); + if (folio_test_private(folio)) + return false; + + if (folio_test_fscache(folio)) { + if (current_is_kswapd() || !(gfp & __GFP_FS)) + return false; + folio_wait_fscache(folio); + } + ceph_fscache_note_page_release(inode); + return true; } -static int ceph_releasepage(struct page *page, gfp_t g) +static void ceph_netfs_expand_readahead(struct netfs_io_request *rreq) { - dout("%p releasepage %p idx %lu (%sdirty)\n", page->mapping->host, - page, page->index, PageDirty(page) ? "" : "not "); + struct inode *inode = rreq->inode; + struct ceph_inode_info *ci = ceph_inode(inode); + struct ceph_file_layout *lo = &ci->i_layout; + u32 blockoff; + u64 blockno; - /* Can we release the page from the cache? */ - if (!ceph_release_fscache_page(page, g)) - return 0; + /* Expand the start downward */ + blockno = div_u64_rem(rreq->start, lo->stripe_unit, &blockoff); + rreq->start = blockno * lo->stripe_unit; + rreq->len += blockoff; - return !PagePrivate(page); + /* Now, round up the length to the next block */ + rreq->len = roundup(rreq->len, lo->stripe_unit); } -/* - * read a single page, without unlocking it. - */ -static int ceph_do_readpage(struct file *filp, struct page *page) +static bool ceph_netfs_clamp_length(struct netfs_io_subrequest *subreq) { - struct inode *inode = file_inode(filp); - struct ceph_inode_info *ci = ceph_inode(inode); + struct inode *inode = subreq->rreq->inode; struct ceph_fs_client *fsc = ceph_inode_to_client(inode); - int err = 0; - u64 off = page_offset(page); - u64 len = PAGE_SIZE; + struct ceph_inode_info *ci = ceph_inode(inode); + u64 objno, objoff; + u32 xlen; + + /* Truncate the extent at the end of the current block */ + ceph_calc_file_object_mapping(&ci->i_layout, subreq->start, subreq->len, + &objno, &objoff, &xlen); + subreq->len = min(xlen, fsc->mount_options->rsize); + return true; +} - if (off >= i_size_read(inode)) { - zero_user_segment(page, 0, PAGE_SIZE); - SetPageUptodate(page); - return 0; - } +static void finish_netfs_read(struct ceph_osd_request *req) +{ + struct ceph_fs_client *fsc = ceph_inode_to_client(req->r_inode); + struct ceph_osd_data *osd_data = osd_req_op_extent_osd_data(req, 0); + struct netfs_io_subrequest *subreq = req->r_priv; + int num_pages; + int err = req->r_result; - if (ci->i_inline_version != CEPH_INLINE_NONE) { - /* - * Uptodate inline data should have been added - * into page cache while getting Fcr caps. - */ - if (off == 0) - return -EINVAL; - zero_user_segment(page, 0, PAGE_SIZE); - SetPageUptodate(page); - return 0; - } + ceph_update_read_metrics(&fsc->mdsc->metric, req->r_start_latency, + req->r_end_latency, osd_data->length, err); - err = ceph_readpage_from_fscache(inode, page); - if (err == 0) - return -EINPROGRESS; - - dout("readpage inode %p file %p page %p index %lu\n", - inode, filp, page, page->index); - err = ceph_osdc_readpages(&fsc->client->osdc, ceph_vino(inode), - &ci->i_layout, off, &len, - ci->i_truncate_seq, ci->i_truncate_size, - &page, 1, 0); + dout("%s: result %d subreq->len=%zu i_size=%lld\n", __func__, req->r_result, + subreq->len, i_size_read(req->r_inode)); + + /* no object means success but no data */ if (err == -ENOENT) err = 0; - if (err < 0) { - SetPageError(page); - ceph_fscache_readpage_cancel(inode, page); - if (err == -EBLACKLISTED) - fsc->blacklisted = true; - goto out; - } - if (err < PAGE_SIZE) - /* zero fill remainder of page */ - zero_user_segment(page, err, PAGE_SIZE); - else - flush_dcache_page(page); + else if (err == -EBLOCKLISTED) + fsc->blocklisted = true; - SetPageUptodate(page); - ceph_readpage_to_fscache(inode, page); + if (err >= 0 && err < subreq->len) + __set_bit(NETFS_SREQ_CLEAR_TAIL, &subreq->flags); -out: - return err < 0 ? err : 0; -} + netfs_subreq_terminated(subreq, err, false); -static int ceph_readpage(struct file *filp, struct page *page) -{ - int r = ceph_do_readpage(filp, page); - if (r != -EINPROGRESS) - unlock_page(page); - else - r = 0; - return r; + num_pages = calc_pages_for(osd_data->alignment, osd_data->length); + ceph_put_page_vector(osd_data->pages, num_pages, false); + iput(req->r_inode); } -/* - * Finish an async read(ahead) op. - */ -static void finish_read(struct ceph_osd_request *req) +static bool ceph_netfs_issue_op_inline(struct netfs_io_subrequest *subreq) { - struct inode *inode = req->r_inode; - struct ceph_osd_data *osd_data; - 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; + struct netfs_io_request *rreq = subreq->rreq; + struct inode *inode = rreq->inode; + struct ceph_mds_reply_info_parsed *rinfo; + struct ceph_mds_reply_info_in *iinfo; + struct ceph_mds_request *req; + struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(inode->i_sb); + struct ceph_inode_info *ci = ceph_inode(inode); + struct iov_iter iter; + ssize_t err = 0; + size_t len; + int mode; - dout("finish_read %p req %p rc %d bytes %d\n", inode, req, rc, bytes); - if (rc == -EBLACKLISTED) - ceph_inode_to_client(inode)->blacklisted = true; + __set_bit(NETFS_SREQ_CLEAR_TAIL, &subreq->flags); + __clear_bit(NETFS_SREQ_COPY_TO_CACHE, &subreq->flags); - /* unlock all pages, zeroing any data we didn't read */ - osd_data = osd_req_op_extent_osd_data(req, 0); - BUG_ON(osd_data->type != CEPH_OSD_DATA_TYPE_PAGES); - num_pages = calc_pages_for((u64)osd_data->alignment, - (u64)osd_data->length); - for (i = 0; i < num_pages; i++) { - struct page *page = osd_data->pages[i]; - - 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; - zero_user_segment(page, s, PAGE_SIZE); - } - dout("finish_read %p uptodate %p idx %lu\n", inode, page, - page->index); - flush_dcache_page(page); - SetPageUptodate(page); - ceph_readpage_to_fscache(inode, page); -unlock: - unlock_page(page); - put_page(page); - bytes -= PAGE_SIZE; + if (subreq->start >= inode->i_size) + goto out; + + /* We need to fetch the inline data. */ + mode = ceph_try_to_choose_auth_mds(inode, CEPH_STAT_CAP_INLINE_DATA); + req = ceph_mdsc_create_request(mdsc, CEPH_MDS_OP_GETATTR, mode); + if (IS_ERR(req)) { + err = PTR_ERR(req); + goto out; + } + req->r_ino1 = ci->i_vino; + req->r_args.getattr.mask = cpu_to_le32(CEPH_STAT_CAP_INLINE_DATA); + req->r_num_caps = 2; + + err = ceph_mdsc_do_request(mdsc, NULL, req); + if (err < 0) + goto out; + + rinfo = &req->r_reply_info; + iinfo = &rinfo->targeti; + if (iinfo->inline_version == CEPH_INLINE_NONE) { + /* The data got uninlined */ + ceph_mdsc_put_request(req); + return false; } - kfree(osd_data->pages); + + len = min_t(size_t, iinfo->inline_len - subreq->start, subreq->len); + iov_iter_xarray(&iter, READ, &rreq->mapping->i_pages, subreq->start, len); + err = copy_to_iter(iinfo->inline_data + subreq->start, len, &iter); + if (err == 0) + err = -EFAULT; + + ceph_mdsc_put_request(req); +out: + netfs_subreq_terminated(subreq, err, false); + return true; } -/* - * start an async read(ahead) operation. return nr_pages we submitted - * a read for on success, or negative error code. - */ -static int start_read(struct inode *inode, struct ceph_rw_context *rw_ctx, - struct list_head *page_list, int max) +static void ceph_netfs_issue_read(struct netfs_io_subrequest *subreq) { - struct ceph_osd_client *osdc = - &ceph_inode_to_client(inode)->client->osdc; + struct netfs_io_request *rreq = subreq->rreq; + struct inode *inode = rreq->inode; struct ceph_inode_info *ci = ceph_inode(inode); - struct page *page = lru_to_page(page_list); - struct ceph_vino vino; + struct ceph_fs_client *fsc = ceph_inode_to_client(inode); struct ceph_osd_request *req; - u64 off; - u64 len; - int i; + struct ceph_vino vino = ceph_vino(inode); + struct iov_iter iter; struct page **pages; - pgoff_t next_index; - int nr_pages = 0; - int got = 0; - int ret = 0; - - if (!rw_ctx) { - /* caller of readpages does not hold buffer and read caps - * (fadvise, madvise and readahead cases) */ - int want = CEPH_CAP_FILE_CACHE; - ret = ceph_try_get_caps(inode, CEPH_CAP_FILE_RD, want, - true, &got); - if (ret < 0) { - dout("start_read %p, error getting cap\n", inode); - } else if (!(got & want)) { - dout("start_read %p, no cache cap\n", inode); - ret = 0; - } - if (ret <= 0) { - if (got) - ceph_put_cap_refs(ci, got); - while (!list_empty(page_list)) { - page = lru_to_page(page_list); - list_del(&page->lru); - put_page(page); - } - return ret; - } - } + size_t page_off; + int err = 0; + u64 len = subreq->len; - off = (u64) page_offset(page); + if (ceph_has_inline_data(ci) && ceph_netfs_issue_op_inline(subreq)) + return; - /* count pages */ - next_index = page->index; - list_for_each_entry_reverse(page, page_list, lru) { - if (page->index != next_index) - break; - nr_pages++; - next_index++; - if (max && nr_pages == max) - break; - } - len = nr_pages << PAGE_SHIFT; - dout("start_read %p nr_pages %d is %lld~%lld\n", inode, nr_pages, - off, len); - vino = ceph_vino(inode); - req = ceph_osdc_new_request(osdc, &ci->i_layout, vino, off, &len, - 0, 1, CEPH_OSD_OP_READ, - CEPH_OSD_FLAG_READ, NULL, - ci->i_truncate_seq, ci->i_truncate_size, - false); + req = ceph_osdc_new_request(&fsc->client->osdc, &ci->i_layout, vino, subreq->start, &len, + 0, 1, CEPH_OSD_OP_READ, + CEPH_OSD_FLAG_READ | fsc->client->osdc.client->options->read_from_replica, + NULL, ci->i_truncate_seq, ci->i_truncate_size, false); if (IS_ERR(req)) { - ret = PTR_ERR(req); + err = PTR_ERR(req); + req = NULL; goto out; } - /* build page vector */ - nr_pages = calc_pages_for(0, len); - pages = kmalloc_array(nr_pages, sizeof(*pages), GFP_KERNEL); - if (!pages) { - ret = -ENOMEM; - goto out_put; - } - for (i = 0; i < nr_pages; ++i) { - page = list_entry(page_list->prev, struct page, lru); - BUG_ON(PageLocked(page)); - list_del(&page->lru); - - dout("start_read %p adding %p idx %lu\n", inode, page, - page->index); - if (add_to_page_cache_lru(page, &inode->i_data, page->index, - GFP_KERNEL)) { - ceph_fscache_uncache_page(inode, page); - put_page(page); - dout("start_read %p add_to_page_cache failed %p\n", - inode, page); - nr_pages = i; - if (nr_pages > 0) { - len = nr_pages << PAGE_SHIFT; - osd_req_op_extent_update(req, 0, len); - break; - } - goto out_pages; - } - pages[i] = page; + dout("%s: pos=%llu orig_len=%zu len=%llu\n", __func__, subreq->start, subreq->len, len); + iov_iter_xarray(&iter, READ, &rreq->mapping->i_pages, subreq->start, len); + err = iov_iter_get_pages_alloc2(&iter, &pages, len, &page_off); + if (err < 0) { + dout("%s: iov_ter_get_pages_alloc returned %d\n", __func__, err); + goto out; } + + /* should always give us a page-aligned read */ + WARN_ON_ONCE(page_off); + len = err; + err = 0; + osd_req_op_extent_osd_data_pages(req, 0, pages, len, 0, false, false); - req->r_callback = finish_read; + req->r_callback = finish_netfs_read; + req->r_priv = subreq; req->r_inode = inode; + ihold(inode); - dout("start_read %p starting %p %lld~%lld\n", inode, req, off, len); - ret = ceph_osdc_start_request(osdc, req, false); - if (ret < 0) - goto out_pages; + ceph_osdc_start_request(req->r_osdc, req); +out: ceph_osdc_put_request(req); + if (err) + netfs_subreq_terminated(subreq, err, false); + dout("%s: result %d\n", __func__, err); +} - /* After adding locked pages to page cache, the inode holds cache cap. - * So we can drop our cap refs. */ - if (got) - ceph_put_cap_refs(ci, got); +static int ceph_init_request(struct netfs_io_request *rreq, struct file *file) +{ + struct inode *inode = rreq->inode; + int got = 0, want = CEPH_CAP_FILE_CACHE; + int ret = 0; - return nr_pages; + if (rreq->origin != NETFS_READAHEAD) + return 0; -out_pages: - for (i = 0; i < nr_pages; ++i) { - ceph_fscache_readpage_cancel(inode, pages[i]); - unlock_page(pages[i]); + if (file) { + struct ceph_rw_context *rw_ctx; + struct ceph_file_info *fi = file->private_data; + + rw_ctx = ceph_find_rw_context(fi); + if (rw_ctx) + return 0; } - ceph_put_page_vector(pages, nr_pages, false); -out_put: - ceph_osdc_put_request(req); -out: + + /* + * readahead callers do not necessarily hold Fcb caps + * (e.g. fadvise, madvise). + */ + ret = ceph_try_get_caps(inode, CEPH_CAP_FILE_RD, want, true, &got); + if (ret < 0) { + dout("start_read %p, error getting cap\n", inode); + return ret; + } + + if (!(got & want)) { + dout("start_read %p, no cache cap\n", inode); + return -EACCES; + } + if (ret == 0) + return -EACCES; + + rreq->netfs_priv = (void *)(uintptr_t)got; + return 0; +} + +static void ceph_netfs_free_request(struct netfs_io_request *rreq) +{ + struct ceph_inode_info *ci = ceph_inode(rreq->inode); + int got = (uintptr_t)rreq->netfs_priv; + if (got) ceph_put_cap_refs(ci, got); - return ret; } +const struct netfs_request_ops ceph_netfs_ops = { + .init_request = ceph_init_request, + .free_request = ceph_netfs_free_request, + .begin_cache_operation = ceph_begin_cache_operation, + .issue_read = ceph_netfs_issue_read, + .expand_readahead = ceph_netfs_expand_readahead, + .clamp_length = ceph_netfs_clamp_length, + .check_write_begin = ceph_netfs_check_write_begin, +}; -/* - * Read multiple pages. Leave pages we don't read + unlock in page_list; - * the caller (VM) cleans them up. - */ -static int ceph_readpages(struct file *file, struct address_space *mapping, - struct list_head *page_list, unsigned nr_pages) +#ifdef CONFIG_CEPH_FSCACHE +static void ceph_set_page_fscache(struct page *page) { - struct inode *inode = file_inode(file); - struct ceph_fs_client *fsc = ceph_inode_to_client(inode); - struct ceph_file_info *fi = file->private_data; - struct ceph_rw_context *rw_ctx; - int rc = 0; - int max = 0; + set_page_fscache(page); +} - if (ceph_inode(inode)->i_inline_version != CEPH_INLINE_NONE) - return -EINVAL; +static void ceph_fscache_write_terminated(void *priv, ssize_t error, bool was_async) +{ + struct inode *inode = priv; - rc = ceph_readpages_from_fscache(mapping->host, mapping, page_list, - &nr_pages); + if (IS_ERR_VALUE(error) && error != -ENOBUFS) + ceph_fscache_invalidate(inode, false); +} - if (rc == 0) - goto out; +static void ceph_fscache_write_to_cache(struct inode *inode, u64 off, u64 len, bool caching) +{ + struct ceph_inode_info *ci = ceph_inode(inode); + struct fscache_cookie *cookie = ceph_fscache_cookie(ci); - rw_ctx = ceph_find_rw_context(fi); - max = fsc->mount_options->rsize >> PAGE_SHIFT; - dout("readpages %p file %p ctx %p nr_pages %d max %d\n", - inode, file, rw_ctx, nr_pages, max); - while (!list_empty(page_list)) { - rc = start_read(inode, rw_ctx, page_list, max); - if (rc < 0) - goto out; - } -out: - ceph_fscache_readpages_cancel(inode, page_list); + fscache_write_to_cache(cookie, inode->i_mapping, off, len, i_size_read(inode), + ceph_fscache_write_terminated, inode, caching); +} +#else +static inline void ceph_set_page_fscache(struct page *page) +{ +} - dout("readpages %p file %p ret %d\n", inode, file, rc); - return rc; +static inline void ceph_fscache_write_to_cache(struct inode *inode, u64 off, u64 len, bool caching) +{ } +#endif /* CONFIG_CEPH_FSCACHE */ struct ceph_writeback_ctl { @@ -565,8 +535,8 @@ static u64 get_writepages_data_length(struct inode *inode, spin_unlock(&ci->i_ceph_lock); WARN_ON(!found); } - if (end > page_offset(page) + PAGE_SIZE) - end = page_offset(page) + PAGE_SIZE; + if (end > page_offset(page) + thp_size(page)) + end = page_offset(page) + thp_size(page); return end > start ? end - start : 0; } @@ -578,20 +548,21 @@ static u64 get_writepages_data_length(struct inode *inode, */ static int writepage_nounlock(struct page *page, struct writeback_control *wbc) { - struct inode *inode; - struct ceph_inode_info *ci; - struct ceph_fs_client *fsc; + struct folio *folio = page_folio(page); + struct inode *inode = page->mapping->host; + struct ceph_inode_info *ci = ceph_inode(inode); + struct ceph_fs_client *fsc = ceph_inode_to_client(inode); struct ceph_snap_context *snapc, *oldest; loff_t page_off = page_offset(page); - int err, len = PAGE_SIZE; + int err; + loff_t len = thp_size(page); struct ceph_writeback_ctl ceph_wbc; + struct ceph_osd_client *osdc = &fsc->client->osdc; + struct ceph_osd_request *req; + bool caching = ceph_is_cache_enabled(inode); dout("writepage %p idx %lu\n", page, page->index); - inode = page->mapping->host; - ci = ceph_inode(inode); - fsc = ceph_inode_to_client(inode); - /* verify this is a writeable snap context */ snapc = page_snap_context(page); if (!snapc) { @@ -612,27 +583,52 @@ static int writepage_nounlock(struct page *page, struct writeback_control *wbc) /* is this a partial page at end of file? */ if (page_off >= ceph_wbc.i_size) { - dout("%p page eof %llu\n", page, ceph_wbc.i_size); - page->mapping->a_ops->invalidatepage(page, 0, PAGE_SIZE); + dout("folio at %lu beyond eof %llu\n", folio->index, + ceph_wbc.i_size); + folio_invalidate(folio, 0, folio_size(folio)); return 0; } if (ceph_wbc.i_size < page_off + len) len = ceph_wbc.i_size - page_off; - dout("writepage %p page %p index %lu on %llu~%u snapc %p seq %lld\n", + dout("writepage %p page %p index %lu on %llu~%llu snapc %p seq %lld\n", inode, page, page->index, page_off, len, snapc, snapc->seq); if (atomic_long_inc_return(&fsc->writeback_count) > CONGESTION_ON_THRESH(fsc->mount_options->congestion_kb)) - set_bdi_congested(inode_to_bdi(inode), BLK_RW_ASYNC); + fsc->write_congested = true; + + req = ceph_osdc_new_request(osdc, &ci->i_layout, ceph_vino(inode), page_off, &len, 0, 1, + CEPH_OSD_OP_WRITE, CEPH_OSD_FLAG_WRITE, snapc, + ceph_wbc.truncate_seq, ceph_wbc.truncate_size, + true); + if (IS_ERR(req)) { + redirty_page_for_writepage(wbc, page); + return PTR_ERR(req); + } set_page_writeback(page); - err = ceph_osdc_writepages(&fsc->client->osdc, ceph_vino(inode), - &ci->i_layout, snapc, page_off, len, - ceph_wbc.truncate_seq, - ceph_wbc.truncate_size, - &inode->i_mtime, &page, 1); + if (caching) + ceph_set_page_fscache(page); + ceph_fscache_write_to_cache(inode, page_off, len, caching); + + /* it may be a short write due to an object boundary */ + WARN_ON_ONCE(len > thp_size(page)); + osd_req_op_extent_osd_data_pages(req, 0, &page, len, 0, false, false); + dout("writepage %llu~%llu (%llu bytes)\n", page_off, len, len); + + req->r_mtime = inode->i_mtime; + ceph_osdc_start_request(osdc, req); + err = ceph_osdc_wait_request(osdc, req); + + ceph_update_write_metrics(&fsc->mdsc->metric, req->r_start_latency, + req->r_end_latency, len, err); + + ceph_osdc_put_request(req); + if (err == 0) + err = len; + if (err < 0) { struct writeback_control tmp_wbc; if (!wbc) @@ -644,8 +640,8 @@ static int writepage_nounlock(struct page *page, struct writeback_control *wbc) end_page_writeback(page); return err; } - if (err == -EBLACKLISTED) - fsc->blacklisted = true; + if (err == -EBLOCKLISTED) + fsc->blocklisted = true; dout("writepage setting page/mapping error %d %p\n", err, page); mapping_set_error(&inode->i_data, err); @@ -654,15 +650,15 @@ static int writepage_nounlock(struct page *page, struct writeback_control *wbc) dout("writepage cleaned page %p\n", page); err = 0; /* vfs expects us to return 0 */ } - page->private = 0; - ClearPagePrivate(page); + oldest = detach_page_private(page); + WARN_ON_ONCE(oldest != snapc); end_page_writeback(page); ceph_put_wrbuffer_cap_refs(ci, 1, snapc); ceph_put_snap_context(snapc); /* page's reference */ if (atomic_long_dec_return(&fsc->writeback_count) < CONGESTION_OFF_THRESH(fsc->mount_options->congestion_kb)) - clear_bdi_congested(inode_to_bdi(inode), BLK_RW_ASYNC); + fsc->write_congested = false; return err; } @@ -673,6 +669,13 @@ static int ceph_writepage(struct page *page, struct writeback_control *wbc) struct inode *inode = page->mapping->host; BUG_ON(!inode); ihold(inode); + + if (wbc->sync_mode == WB_SYNC_NONE && + ceph_inode_to_client(inode)->write_congested) + return AOP_WRITEPAGE_ACTIVATE; + + wait_on_page_fscache(page); + err = writepage_nounlock(page, wbc); if (err == -ERESTARTSYS) { /* direct memory reclaimer was killed by SIGKILL. return 0 @@ -702,14 +705,15 @@ static void writepages_finish(struct ceph_osd_request *req) struct ceph_snap_context *snapc = req->r_snapc; struct address_space *mapping = inode->i_mapping; struct ceph_fs_client *fsc = ceph_inode_to_client(inode); + unsigned int len = 0; bool remove_page; dout("writepages_finish %p rc %d\n", inode, rc); if (rc < 0) { mapping_set_error(mapping, rc); ceph_set_error_write(ci); - if (rc == -EBLACKLISTED) - fsc->blacklisted = true; + if (rc == -EBLOCKLISTED) + fsc->blocklisted = true; } else { ceph_clear_error_write(ci); } @@ -725,11 +729,15 @@ static void writepages_finish(struct ceph_osd_request *req) /* clean all pages */ for (i = 0; i < req->r_num_ops; i++) { - if (req->r_ops[i].op != CEPH_OSD_OP_WRITE) + if (req->r_ops[i].op != CEPH_OSD_OP_WRITE) { + pr_warn("%s incorrect op %d req %p index %d tid %llu\n", + __func__, req->r_ops[i].op, req, i, req->r_tid); break; + } osd_data = osd_req_op_extent_osd_data(req, i); BUG_ON(osd_data->type != CEPH_OSD_DATA_TYPE_PAGES); + len += osd_data->length; num_pages = calc_pages_for((u64)osd_data->alignment, (u64)osd_data->length); total_pages += num_pages; @@ -741,14 +749,11 @@ static void writepages_finish(struct ceph_osd_request *req) if (atomic_long_dec_return(&fsc->writeback_count) < CONGESTION_OFF_THRESH( fsc->mount_options->congestion_kb)) - clear_bdi_congested(inode_to_bdi(inode), - BLK_RW_ASYNC); + fsc->write_congested = false; - ceph_put_snap_context(page_snap_context(page)); - page->private = 0; - ClearPagePrivate(page); - dout("unlocking %p\n", page); + ceph_put_snap_context(detach_page_private(page)); end_page_writeback(page); + dout("unlocking %p\n", page); if (remove_page) generic_error_remove_page(inode->i_mapping, @@ -762,12 +767,14 @@ static void writepages_finish(struct ceph_osd_request *req) release_pages(osd_data->pages, num_pages); } + ceph_update_write_metrics(&fsc->mdsc->metric, req->r_start_latency, + req->r_end_latency, len, rc); + ceph_put_wrbuffer_cap_refs(ci, total_pages, snapc); osd_data = osd_req_op_extent_osd_data(req, 0); if (osd_data->pages_from_pool) - mempool_free(osd_data->pages, - ceph_sb_to_client(inode->i_sb)->wb_pagevec_pool); + mempool_free(osd_data->pages, ceph_wb_pagevec_pool); else kfree(osd_data->pages); ceph_osdc_put_request(req); @@ -792,12 +799,17 @@ static int ceph_writepages_start(struct address_space *mapping, struct ceph_writeback_ctl ceph_wbc; bool should_loop, range_whole = false; bool done = false; + bool caching = ceph_is_cache_enabled(inode); + + if (wbc->sync_mode == WB_SYNC_NONE && + fsc->write_congested) + return 0; dout("writepages_start %p (mode=%s)\n", inode, wbc->sync_mode == WB_SYNC_NONE ? "NONE" : (wbc->sync_mode == WB_SYNC_ALL ? "ALL" : "HOLD")); - if (READ_ONCE(fsc->mount_state) == CEPH_MOUNT_SHUTDOWN) { + if (ceph_inode_is_shutdown(inode)) { if (ci->i_wrbuffer_ref > 0) { pr_warn_ratelimited( "writepage_start %p %lld forced umount\n", @@ -859,17 +871,16 @@ retry: int num_ops = 0, op_idx; unsigned i, pvec_pages, max_pages, locked_pages = 0; struct page **pages = NULL, **data_pages; - mempool_t *pool = NULL; /* Becomes non-null if mempool used */ struct page *page; pgoff_t strip_unit_end = 0; u64 offset = 0, len = 0; + bool from_pool = false; max_pages = wsize >> PAGE_SHIFT; get_more_pages: - pvec_pages = pagevec_lookup_range_nr_tag(&pvec, mapping, &index, - end, PAGECACHE_TAG_DIRTY, - max_pages - locked_pages); + pvec_pages = pagevec_lookup_range_tag(&pvec, mapping, &index, + end, PAGECACHE_TAG_DIRTY); dout("pagevec_lookup_range_tag got %d\n", pvec_pages); if (!pvec_pages && !locked_pages) break; @@ -901,14 +912,16 @@ get_more_pages: continue; } if (page_offset(page) >= ceph_wbc.i_size) { - dout("%p page eof %llu\n", - page, ceph_wbc.i_size); + struct folio *folio = page_folio(page); + + dout("folio at %lu beyond eof %llu\n", + folio->index, ceph_wbc.i_size); if ((ceph_wbc.size_stable || - page_offset(page) >= i_size_read(inode)) && - clear_page_dirty_for_io(page)) - mapping->a_ops->invalidatepage(page, - 0, PAGE_SIZE); - unlock_page(page); + folio_pos(folio) >= i_size_read(inode)) && + folio_clear_dirty_for_io(folio)) + folio_invalidate(folio, 0, + folio_size(folio)); + folio_unlock(folio); continue; } if (strip_unit_end && (page->index > strip_unit_end)) { @@ -916,7 +929,7 @@ get_more_pages: unlock_page(page); break; } - if (PageWriteback(page)) { + if (PageWriteback(page) || PageFsCache(page)) { if (wbc->sync_mode == WB_SYNC_NONE) { dout("%p under writeback\n", page); unlock_page(page); @@ -924,6 +937,7 @@ get_more_pages: } dout("waiting on writeback %p\n", page); wait_on_page_writeback(page); + wait_on_page_fscache(page); } if (!clear_page_dirty_for_io(page)) { @@ -961,16 +975,16 @@ get_more_pages: sizeof(*pages), GFP_NOFS); if (!pages) { - pool = fsc->wb_pagevec_pool; - pages = mempool_alloc(pool, GFP_NOFS); + from_pool = true; + pages = mempool_alloc(ceph_wb_pagevec_pool, GFP_NOFS); BUG_ON(!pages); } len = 0; } else if (page->index != (offset + len) >> PAGE_SHIFT) { - if (num_ops >= (pool ? CEPH_OSD_SLAB_OPS : - CEPH_OSD_MAX_OPS)) { + if (num_ops >= (from_pool ? CEPH_OSD_SLAB_OPS : + CEPH_OSD_MAX_OPS)) { redirty_page_for_writepage(wbc, page); unlock_page(page); break; @@ -987,16 +1001,13 @@ get_more_pages: if (atomic_long_inc_return(&fsc->writeback_count) > CONGESTION_ON_THRESH( - fsc->mount_options->congestion_kb)) { - set_bdi_congested(inode_to_bdi(inode), - BLK_RW_ASYNC); - } - + fsc->mount_options->congestion_kb)) + fsc->write_congested = true; pages[locked_pages++] = page; pvec.pages[i] = NULL; - len += PAGE_SIZE; + len += thp_size(page); } /* did we get anything? */ @@ -1045,7 +1056,7 @@ new_request: BUG_ON(IS_ERR(req)); } BUG_ON(len < page_offset(pages[locked_pages - 1]) + - PAGE_SIZE - offset); + thp_size(page) - offset); req->r_callback = writepages_finish; req->r_inode = inode; @@ -1056,27 +1067,40 @@ new_request: op_idx = 0; for (i = 0; i < locked_pages; i++) { u64 cur_offset = page_offset(pages[i]); + /* + * Discontinuity in page range? Ceph can handle that by just passing + * multiple extents in the write op. + */ if (offset + len != cur_offset) { + /* If it's full, stop here */ if (op_idx + 1 == req->r_num_ops) break; + + /* Kick off an fscache write with what we have so far. */ + ceph_fscache_write_to_cache(inode, offset, len, caching); + + /* Start a new extent */ osd_req_op_extent_dup_last(req, op_idx, cur_offset - offset); dout("writepages got pages at %llu~%llu\n", offset, len); osd_req_op_extent_osd_data_pages(req, op_idx, data_pages, len, 0, - !!pool, false); + from_pool, false); osd_req_op_extent_update(req, op_idx, len); len = 0; - offset = cur_offset; + offset = cur_offset; data_pages = pages + i; op_idx++; } set_page_writeback(pages[i]); - len += PAGE_SIZE; + if (caching) + ceph_set_page_fscache(pages[i]); + len += thp_size(page); } + ceph_fscache_write_to_cache(inode, offset, len, caching); if (ceph_wbc.size_stable) { len = min(len, ceph_wbc.i_size - offset); @@ -1084,7 +1108,7 @@ new_request: /* writepages_finish() clears writeback pages * according to the data length, so make sure * data length covers all locked pages */ - u64 min_len = len + 1 - PAGE_SIZE; + u64 min_len = len + 1 - thp_size(page); len = get_writepages_data_length(inode, pages[i - 1], offset); len = max(len, min_len); @@ -1092,12 +1116,12 @@ new_request: dout("writepages got pages at %llu~%llu\n", offset, len); osd_req_op_extent_osd_data_pages(req, op_idx, data_pages, len, - 0, !!pool, false); + 0, from_pool, false); osd_req_op_extent_update(req, op_idx, len); BUG_ON(op_idx + 1 != req->r_num_ops); - pool = NULL; + from_pool = false; if (i < locked_pages) { BUG_ON(num_ops <= req->r_num_ops); num_ops -= req->r_num_ops; @@ -1108,8 +1132,8 @@ new_request: pages = kmalloc_array(locked_pages, sizeof(*pages), GFP_NOFS); if (!pages) { - pool = fsc->wb_pagevec_pool; - pages = mempool_alloc(pool, GFP_NOFS); + from_pool = true; + pages = mempool_alloc(ceph_wb_pagevec_pool, GFP_NOFS); BUG_ON(!pages); } memcpy(pages, data_pages + i, @@ -1124,8 +1148,7 @@ new_request: } req->r_mtime = inode->i_mtime; - rc = ceph_osdc_start_request(&fsc->client->osdc, req, true); - BUG_ON(rc); + ceph_osdc_start_request(&fsc->client->osdc, req); req = NULL; wbc->nr_to_write -= i; @@ -1204,110 +1227,86 @@ static int context_is_writeable_or_written(struct inode *inode, return ret; } -/* - * We are only allowed to write into/dirty the page if the page is - * clean, or already dirty within the same snap context. +/** + * ceph_find_incompatible - find an incompatible context and return it + * @page: page being dirtied * - * called with page locked. - * return success with page locked, - * or any failure (incl -EAGAIN) with page unlocked. + * We are only allowed to write into/dirty a page if the page is + * clean, or already dirty within the same snap context. Returns a + * conflicting context if there is one, NULL if there isn't, or a + * negative error code on other errors. + * + * Must be called with page lock held. */ -static int ceph_update_writeable_page(struct file *file, - loff_t pos, unsigned len, - struct page *page) +static struct ceph_snap_context * +ceph_find_incompatible(struct page *page) { - struct inode *inode = file_inode(file); - struct ceph_fs_client *fsc = ceph_inode_to_client(inode); + struct inode *inode = page->mapping->host; struct ceph_inode_info *ci = ceph_inode(inode); - loff_t page_off = pos & PAGE_MASK; - int pos_in_page = pos & ~PAGE_MASK; - int end_in_page = pos_in_page + len; - loff_t i_size; - int r; - struct ceph_snap_context *snapc, *oldest; - if (READ_ONCE(fsc->mount_state) == CEPH_MOUNT_SHUTDOWN) { - dout(" page %p forced umount\n", page); - unlock_page(page); - return -EIO; + if (ceph_inode_is_shutdown(inode)) { + dout(" page %p %llx:%llx is shutdown\n", page, + ceph_vinop(inode)); + return ERR_PTR(-ESTALE); } -retry_locked: - /* writepages currently holds page lock, but if we change that later, */ - wait_on_page_writeback(page); + for (;;) { + struct ceph_snap_context *snapc, *oldest; + + wait_on_page_writeback(page); + + snapc = page_snap_context(page); + if (!snapc || snapc == ci->i_head_snapc) + break; - snapc = page_snap_context(page); - if (snapc && snapc != ci->i_head_snapc) { /* * this page is already dirty in another (older) snap * context! is it writeable now? */ oldest = get_oldest_context(inode, NULL, NULL); if (snapc->seq > oldest->seq) { + /* not writeable -- return it for the caller to deal with */ ceph_put_snap_context(oldest); - dout(" page %p snapc %p not current or oldest\n", - page, snapc); - /* - * queue for writeback, and wait for snapc to - * be writeable or written - */ - snapc = ceph_get_snap_context(snapc); - unlock_page(page); - ceph_queue_writeback(inode); - r = wait_event_killable(ci->i_cap_wq, - context_is_writeable_or_written(inode, snapc)); - ceph_put_snap_context(snapc); - if (r == -ERESTARTSYS) - return r; - return -EAGAIN; + dout(" page %p snapc %p not current or oldest\n", page, snapc); + return ceph_get_snap_context(snapc); } ceph_put_snap_context(oldest); /* yay, writeable, do it now (without dropping page lock) */ - dout(" page %p snapc %p not current, but oldest\n", - page, snapc); - if (!clear_page_dirty_for_io(page)) - goto retry_locked; - r = writepage_nounlock(page, NULL); - if (r < 0) - goto fail_unlock; - goto retry_locked; - } - - if (PageUptodate(page)) { - dout(" page %p already uptodate\n", page); - return 0; + dout(" page %p snapc %p not current, but oldest\n", page, snapc); + if (clear_page_dirty_for_io(page)) { + int r = writepage_nounlock(page, NULL); + if (r < 0) + return ERR_PTR(r); + } } + return NULL; +} - /* full page? */ - if (pos_in_page == 0 && len == PAGE_SIZE) - return 0; - - /* past end of file? */ - i_size = i_size_read(inode); - - if (page_off >= i_size || - (pos_in_page == 0 && (pos+len) >= i_size && - end_in_page - pos_in_page != PAGE_SIZE)) { - dout(" zeroing %p 0 - %d and %d - %d\n", - page, pos_in_page, end_in_page, (int)PAGE_SIZE); - zero_user_segments(page, - 0, pos_in_page, - end_in_page, PAGE_SIZE); - return 0; - } +static int ceph_netfs_check_write_begin(struct file *file, loff_t pos, unsigned int len, + struct folio **foliop, void **_fsdata) +{ + struct inode *inode = file_inode(file); + struct ceph_inode_info *ci = ceph_inode(inode); + struct ceph_snap_context *snapc; - /* we need to read it. */ - r = ceph_do_readpage(file, page); - if (r < 0) { - if (r == -EINPROGRESS) - return -EAGAIN; - goto fail_unlock; + snapc = ceph_find_incompatible(folio_page(*foliop, 0)); + if (snapc) { + int r; + + folio_unlock(*foliop); + folio_put(*foliop); + *foliop = NULL; + if (IS_ERR(snapc)) + return PTR_ERR(snapc); + + ceph_queue_writeback(inode); + r = wait_event_killable(ci->i_cap_wq, + context_is_writeable_or_written(inode, snapc)); + ceph_put_snap_context(snapc); + return r == 0 ? -EAGAIN : r; } - goto retry_locked; -fail_unlock: - unlock_page(page); - return r; + return 0; } /* @@ -1315,31 +1314,22 @@ fail_unlock: * clean, or already dirty within the same snap context. */ static int ceph_write_begin(struct file *file, struct address_space *mapping, - loff_t pos, unsigned len, unsigned flags, + loff_t pos, unsigned len, struct page **pagep, void **fsdata) { struct inode *inode = file_inode(file); - struct page *page; - pgoff_t index = pos >> PAGE_SHIFT; + struct ceph_inode_info *ci = ceph_inode(inode); + struct folio *folio = NULL; int r; - do { - /* get a page */ - page = grab_cache_page_write_begin(mapping, index, 0); - if (!page) - return -ENOMEM; - - dout("write_begin file %p inode %p page %p %d~%d\n", file, - inode, page, (int)pos, (int)len); + r = netfs_write_begin(&ci->netfs, file, inode->i_mapping, pos, len, &folio, NULL); + if (r < 0) + return r; - r = ceph_update_writeable_page(file, pos, len, page); - if (r < 0) - put_page(page); - else - *pagep = page; - } while (r == -EAGAIN); - - return r; + folio_wait_fscache(folio); + WARN_ON_ONCE(!folio_test_locked(folio)); + *pagep = &folio->page; + return 0; } /* @@ -1348,32 +1338,33 @@ static int ceph_write_begin(struct file *file, struct address_space *mapping, */ static int ceph_write_end(struct file *file, struct address_space *mapping, loff_t pos, unsigned len, unsigned copied, - struct page *page, void *fsdata) + struct page *subpage, void *fsdata) { + struct folio *folio = page_folio(subpage); struct inode *inode = file_inode(file); bool check_cap = false; - dout("write_end file %p inode %p page %p %d~%d (%d)\n", file, - inode, page, (int)pos, (int)copied, (int)len); + dout("write_end file %p inode %p folio %p %d~%d (%d)\n", file, + inode, folio, (int)pos, (int)copied, (int)len); - /* zero the stale part of the page if we did a short copy */ - if (!PageUptodate(page)) { + if (!folio_test_uptodate(folio)) { + /* just return that nothing was copied on a short copy */ if (copied < len) { copied = 0; goto out; } - SetPageUptodate(page); + folio_mark_uptodate(folio); } /* did file size increase? */ if (pos+copied > i_size_read(inode)) check_cap = ceph_inode_set_size(inode, pos+copied); - set_page_dirty(page); + folio_mark_dirty(folio); out: - unlock_page(page); - put_page(page); + folio_unlock(folio); + folio_put(folio); if (check_cap) ceph_check_caps(ceph_inode(inode), CHECK_CAPS_AUTHONLY, NULL); @@ -1381,28 +1372,17 @@ out: return copied; } -/* - * we set .direct_IO to indicate direct io is supported, but since we - * intercept O_DIRECT reads and writes early, this function should - * never get called. - */ -static ssize_t ceph_direct_io(struct kiocb *iocb, struct iov_iter *iter) -{ - WARN_ON(1); - return -EINVAL; -} - const struct address_space_operations ceph_aops = { - .readpage = ceph_readpage, - .readpages = ceph_readpages, + .read_folio = netfs_read_folio, + .readahead = netfs_readahead, .writepage = ceph_writepage, .writepages = ceph_writepages_start, .write_begin = ceph_write_begin, .write_end = ceph_write_end, - .set_page_dirty = ceph_set_page_dirty, - .invalidatepage = ceph_invalidatepage, - .releasepage = ceph_releasepage, - .direct_IO = ceph_direct_io, + .dirty_folio = ceph_dirty_folio, + .invalidate_folio = ceph_invalidate_folio, + .release_folio = ceph_release_folio, + .direct_IO = noop_direct_IO, }; static void ceph_block_sigs(sigset_t *oldset) @@ -1426,44 +1406,42 @@ static vm_fault_t ceph_filemap_fault(struct vm_fault *vmf) struct inode *inode = file_inode(vma->vm_file); struct ceph_inode_info *ci = ceph_inode(inode); struct ceph_file_info *fi = vma->vm_file->private_data; - struct page *pinned_page = NULL; - loff_t off = vmf->pgoff << PAGE_SHIFT; + loff_t off = (loff_t)vmf->pgoff << PAGE_SHIFT; int want, got, err; sigset_t oldset; vm_fault_t ret = VM_FAULT_SIGBUS; + if (ceph_inode_is_shutdown(inode)) + return ret; + 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); + dout("filemap_fault %p %llx.%llx %llu trying to get caps\n", + inode, ceph_vinop(inode), off); if (fi->fmode & CEPH_FILE_MODE_LAZY) want = CEPH_CAP_FILE_CACHE | CEPH_CAP_FILE_LAZYIO; else want = CEPH_CAP_FILE_CACHE; got = 0; - err = ceph_get_caps(vma->vm_file, CEPH_CAP_FILE_RD, want, -1, - &got, &pinned_page); + err = ceph_get_caps(vma->vm_file, CEPH_CAP_FILE_RD, want, -1, &got); if (err < 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)); + dout("filemap_fault %p %llu got cap refs on %s\n", + inode, off, ceph_cap_string(got)); if ((got & (CEPH_CAP_FILE_CACHE | CEPH_CAP_FILE_LAZYIO)) || - ci->i_inline_version == CEPH_INLINE_NONE) { + !ceph_has_inline_data(ci)) { CEPH_DEFINE_RW_CONTEXT(rw_ctx, got); ceph_add_rw_context(fi, &rw_ctx); ret = filemap_fault(vmf); ceph_del_rw_context(fi, &rw_ctx); - dout("filemap_fault %p %llu~%zd drop cap refs %s ret %x\n", - inode, off, (size_t)PAGE_SIZE, - ceph_cap_string(got), ret); + dout("filemap_fault %p %llu drop cap refs %s ret %x\n", + inode, off, ceph_cap_string(got), ret); } else err = -EAGAIN; - if (pinned_page) - put_page(pinned_page); ceph_put_cap_refs(ci, got); if (err != -EAGAIN) @@ -1475,9 +1453,11 @@ static vm_fault_t ceph_filemap_fault(struct vm_fault *vmf) ret = VM_FAULT_SIGBUS; } else { struct address_space *mapping = inode->i_mapping; - struct page *page = find_or_create_page(mapping, 0, - mapping_gfp_constraint(mapping, - ~__GFP_FS)); + struct page *page; + + filemap_invalidate_lock_shared(mapping); + page = find_or_create_page(mapping, 0, + mapping_gfp_constraint(mapping, ~__GFP_FS)); if (!page) { ret = VM_FAULT_OOM; goto out_inline; @@ -1498,8 +1478,9 @@ static vm_fault_t ceph_filemap_fault(struct vm_fault *vmf) vmf->page = page; ret = VM_FAULT_MAJOR | VM_FAULT_LOCKED; out_inline: - dout("filemap_fault %p %llu~%zd read inline data ret %x\n", - inode, off, (size_t)PAGE_SIZE, ret); + filemap_invalidate_unlock_shared(mapping); + dout("filemap_fault %p %llu read inline data ret %x\n", + inode, off, ret); } out_restore: ceph_restore_sigs(&oldset); @@ -1509,9 +1490,6 @@ out_restore: return ret; } -/* - * Reuse write_begin here for simplicity. - */ static vm_fault_t ceph_page_mkwrite(struct vm_fault *vmf) { struct vm_area_struct *vma = vmf->vma; @@ -1527,6 +1505,9 @@ static vm_fault_t ceph_page_mkwrite(struct vm_fault *vmf) sigset_t oldset; vm_fault_t ret = VM_FAULT_SIGBUS; + if (ceph_inode_is_shutdown(inode)) + return ret; + prealloc_cf = ceph_alloc_cap_flush(); if (!prealloc_cf) return VM_FAULT_OOM; @@ -1534,23 +1515,10 @@ static vm_fault_t ceph_page_mkwrite(struct vm_fault *vmf) sb_start_pagefault(inode->i_sb); ceph_block_sigs(&oldset); - if (ci->i_inline_version != CEPH_INLINE_NONE) { - struct page *locked_page = NULL; - if (off == 0) { - lock_page(page); - locked_page = page; - } - err = ceph_uninline_data(vma->vm_file, locked_page); - if (locked_page) - unlock_page(locked_page); - if (err < 0) - goto out_free; - } - - if (off + PAGE_SIZE <= size) - len = PAGE_SIZE; + if (off + thp_size(page) <= size) + len = thp_size(page); else - len = size & ~PAGE_MASK; + len = offset_in_thp(page, size); dout("page_mkwrite %p %llx.%llx %llu~%zd getting caps i_size %llu\n", inode, ceph_vinop(inode), off, len, size); @@ -1560,8 +1528,7 @@ static vm_fault_t ceph_page_mkwrite(struct vm_fault *vmf) want = CEPH_CAP_FILE_BUFFER; got = 0; - err = ceph_get_caps(vma->vm_file, CEPH_CAP_FILE_WR, want, off + len, - &got, NULL); + err = ceph_get_caps(vma->vm_file, CEPH_CAP_FILE_WR, want, off + len, &got); if (err < 0) goto out_free; @@ -1573,27 +1540,40 @@ static vm_fault_t ceph_page_mkwrite(struct vm_fault *vmf) inode_inc_iversion_raw(inode); do { + struct ceph_snap_context *snapc; + lock_page(page); - if ((off > size) || (page->mapping != inode->i_mapping)) { + if (page_mkwrite_check_truncate(page, inode) < 0) { unlock_page(page); ret = VM_FAULT_NOPAGE; break; } - err = ceph_update_writeable_page(vma->vm_file, off, len, page); - if (err >= 0) { + snapc = ceph_find_incompatible(page); + if (!snapc) { /* success. we'll keep the page locked. */ set_page_dirty(page); ret = VM_FAULT_LOCKED; + break; } - } while (err == -EAGAIN); - if (ret == VM_FAULT_LOCKED || - ci->i_inline_version != CEPH_INLINE_NONE) { + unlock_page(page); + + if (IS_ERR(snapc)) { + ret = VM_FAULT_SIGBUS; + break; + } + + ceph_queue_writeback(inode); + err = wait_event_killable(ci->i_cap_wq, + context_is_writeable_or_written(inode, snapc)); + ceph_put_snap_context(snapc); + } while (err == 0); + + if (ret == VM_FAULT_LOCKED) { int dirty; spin_lock(&ci->i_ceph_lock); - ci->i_inline_version = CEPH_INLINE_NONE; dirty = __ceph_mark_dirty_caps(ci, CEPH_CAP_FILE_WR, &prealloc_cf); spin_unlock(&ci->i_ceph_lock); @@ -1603,7 +1583,7 @@ static vm_fault_t ceph_page_mkwrite(struct vm_fault *vmf) dout("page_mkwrite %p %llu~%zd dropping cap refs on %s ret %x\n", inode, off, len, ceph_cap_string(got), ret); - ceph_put_cap_refs(ci, got); + ceph_put_cap_refs_async(ci, got); out_free: ceph_restore_sigs(&oldset); sb_end_pagefault(inode->i_sb); @@ -1657,16 +1637,18 @@ void ceph_fill_inline_data(struct inode *inode, struct page *locked_page, } } -int ceph_uninline_data(struct file *filp, struct page *locked_page) +int ceph_uninline_data(struct file *file) { - struct inode *inode = file_inode(filp); + struct inode *inode = file_inode(file); struct ceph_inode_info *ci = ceph_inode(inode); struct ceph_fs_client *fsc = ceph_inode_to_client(inode); - struct ceph_osd_request *req; - struct page *page = NULL; - u64 len, inline_version; + struct ceph_osd_request *req = NULL; + struct ceph_cap_flush *prealloc_cf; + struct folio *folio = NULL; + u64 inline_version = CEPH_INLINE_NONE; + struct page *pages[1]; int err = 0; - bool from_pagecache = false; + u64 len; spin_lock(&ci->i_ceph_lock); inline_version = ci->i_inline_version; @@ -1675,64 +1657,43 @@ int ceph_uninline_data(struct file *filp, struct page *locked_page) dout("uninline_data %p %llx.%llx inline_version %llu\n", inode, ceph_vinop(inode), inline_version); - if (inline_version == 1 || /* initial version, no data */ - inline_version == CEPH_INLINE_NONE) - goto out; + if (inline_version == CEPH_INLINE_NONE) + return 0; - if (locked_page) { - page = locked_page; - WARN_ON(!PageUptodate(page)); - } else if (ceph_caps_issued(ci) & - (CEPH_CAP_FILE_CACHE|CEPH_CAP_FILE_LAZYIO)) { - page = find_get_page(inode->i_mapping, 0); - if (page) { - if (PageUptodate(page)) { - from_pagecache = true; - lock_page(page); - } else { - put_page(page); - page = NULL; - } - } - } + prealloc_cf = ceph_alloc_cap_flush(); + if (!prealloc_cf) + return -ENOMEM; - if (page) { - len = i_size_read(inode); - if (len > PAGE_SIZE) - len = PAGE_SIZE; - } else { - page = __page_cache_alloc(GFP_NOFS); - if (!page) { - err = -ENOMEM; - goto out; - } - err = __ceph_do_getattr(inode, page, - CEPH_STAT_CAP_INLINE_DATA, true); - if (err < 0) { - /* no inline data */ - if (err == -ENODATA) - err = 0; - goto out; - } - len = err; + if (inline_version == 1) /* initial version, no data */ + goto out_uninline; + + folio = read_mapping_folio(inode->i_mapping, 0, file); + if (IS_ERR(folio)) { + err = PTR_ERR(folio); + goto out; } + folio_lock(folio); + + len = i_size_read(inode); + if (len > folio_size(folio)) + len = folio_size(folio); + req = ceph_osdc_new_request(&fsc->client->osdc, &ci->i_layout, ceph_vino(inode), 0, &len, 0, 1, CEPH_OSD_OP_CREATE, CEPH_OSD_FLAG_WRITE, NULL, 0, 0, false); if (IS_ERR(req)) { err = PTR_ERR(req); - goto out; + goto out_unlock; } 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); + ceph_osdc_start_request(&fsc->client->osdc, req); + err = ceph_osdc_wait_request(&fsc->client->osdc, req); ceph_osdc_put_request(req); if (err < 0) - goto out; + goto out_unlock; req = ceph_osdc_new_request(&fsc->client->osdc, &ci->i_layout, ceph_vino(inode), 0, &len, 1, 3, @@ -1741,10 +1702,11 @@ int ceph_uninline_data(struct file *filp, struct page *locked_page) ci->i_truncate_size, false); if (IS_ERR(req)) { err = PTR_ERR(req); - goto out; + goto out_unlock; } - osd_req_op_extent_osd_data_pages(req, 1, &page, len, 0, false, false); + pages[0] = folio_page(folio, 0); + osd_req_op_extent_osd_data_pages(req, 1, pages, len, 0, false, false); { __le64 xattr_buf = cpu_to_le64(inline_version); @@ -1754,7 +1716,7 @@ int ceph_uninline_data(struct file *filp, struct page *locked_page) CEPH_OSD_CMPXATTR_OP_GT, CEPH_OSD_CMPXATTR_MODE_U64); if (err) - goto out_put; + goto out_put_req; } { @@ -1765,26 +1727,41 @@ int ceph_uninline_data(struct file *filp, struct page *locked_page) "inline_version", xattr_buf, xattr_len, 0, 0); if (err) - goto out_put; + goto out_put_req; } 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); -out_put: + ceph_osdc_start_request(&fsc->client->osdc, req); + err = ceph_osdc_wait_request(&fsc->client->osdc, req); + + ceph_update_write_metrics(&fsc->mdsc->metric, req->r_start_latency, + req->r_end_latency, len, err); + +out_uninline: + if (!err) { + int dirty; + + /* Set to CAP_INLINE_NONE and dirty the caps */ + down_read(&fsc->mdsc->snap_rwsem); + spin_lock(&ci->i_ceph_lock); + ci->i_inline_version = CEPH_INLINE_NONE; + dirty = __ceph_mark_dirty_caps(ci, CEPH_CAP_FILE_WR, &prealloc_cf); + spin_unlock(&ci->i_ceph_lock); + up_read(&fsc->mdsc->snap_rwsem); + if (dirty) + __mark_inode_dirty(inode, dirty); + } +out_put_req: ceph_osdc_put_request(req); if (err == -ECANCELED) err = 0; -out: - if (page && page != locked_page) { - if (from_pagecache) { - unlock_page(page); - put_page(page); - } else - __free_pages(page, 0); +out_unlock: + if (folio) { + folio_unlock(folio); + folio_put(folio); } - +out: + ceph_free_cap_flush(prealloc_cf); dout("uninline_data %p %llx.%llx inline_version %llu = %d\n", inode, ceph_vinop(inode), inline_version, err); return err; @@ -1799,9 +1776,8 @@ int ceph_mmap(struct file *file, struct vm_area_struct *vma) { struct address_space *mapping = file->f_mapping; - if (!mapping->a_ops->readpage) + if (!mapping->a_ops->read_folio) return -ENOEXEC; - file_accessed(file); vma->vm_ops = &ceph_vmops; return 0; } @@ -1814,7 +1790,7 @@ enum { static int __ceph_pool_perm_get(struct ceph_inode_info *ci, s64 pool, struct ceph_string *pool_ns) { - struct ceph_fs_client *fsc = ceph_inode_to_client(&ci->vfs_inode); + struct ceph_fs_client *fsc = ceph_inode_to_client(&ci->netfs.inode); struct ceph_mds_client *mdsc = fsc->mdsc; struct ceph_osd_request *rd_req = NULL, *wr_req = NULL; struct rb_node **p, *parent; @@ -1927,29 +1903,27 @@ static int __ceph_pool_perm_get(struct ceph_inode_info *ci, osd_req_op_raw_data_in_pages(rd_req, 0, pages, PAGE_SIZE, 0, false, true); - err = ceph_osdc_start_request(&fsc->client->osdc, rd_req, false); + ceph_osdc_start_request(&fsc->client->osdc, rd_req); - wr_req->r_mtime = ci->vfs_inode.i_mtime; - err2 = ceph_osdc_start_request(&fsc->client->osdc, wr_req, false); + wr_req->r_mtime = ci->netfs.inode.i_mtime; + ceph_osdc_start_request(&fsc->client->osdc, wr_req); - if (!err) - err = ceph_osdc_wait_request(&fsc->client->osdc, rd_req); - if (!err2) - err2 = ceph_osdc_wait_request(&fsc->client->osdc, wr_req); + err = ceph_osdc_wait_request(&fsc->client->osdc, rd_req); + err2 = ceph_osdc_wait_request(&fsc->client->osdc, wr_req); if (err >= 0 || err == -ENOENT) have |= POOL_READ; else if (err != -EPERM) { - if (err == -EBLACKLISTED) - fsc->blacklisted = true; + if (err == -EBLOCKLISTED) + fsc->blocklisted = true; goto out_unlock; } if (err2 == 0 || err2 == -EEXIST) have |= POOL_WRITE; else if (err2 != -EPERM) { - if (err2 == -EBLACKLISTED) - fsc->blacklisted = true; + if (err2 == -EBLOCKLISTED) + fsc->blocklisted = true; err = err2; goto out_unlock; } @@ -1994,6 +1968,10 @@ int ceph_pool_perm_check(struct inode *inode, int need) s64 pool; int ret, flags; + /* Only need to do this for regular files */ + if (!S_ISREG(inode->i_mode)) + return 0; + if (ci->i_vino.snap != CEPH_NOSNAP) { /* * Pool permission check needs to write to the first object. diff --git a/fs/ceph/cache.c b/fs/ceph/cache.c index 270b769607a2..177d8e8d73fe 100644 --- a/fs/ceph/cache.c +++ b/fs/ceph/cache.c @@ -12,341 +12,99 @@ #include "super.h" #include "cache.h" -struct ceph_aux_inode { - u64 version; - u64 mtime_sec; - u64 mtime_nsec; -}; - -struct fscache_netfs ceph_cache_netfs = { - .name = "ceph", - .version = 0, -}; - -static DEFINE_MUTEX(ceph_fscache_lock); -static LIST_HEAD(ceph_fscache_list); - -struct ceph_fscache_entry { - struct list_head list; - struct fscache_cookie *fscache; - size_t uniq_len; - /* The following members must be last */ - struct ceph_fsid fsid; - char uniquifier[0]; -}; - -static const struct fscache_cookie_def ceph_fscache_fsid_object_def = { - .name = "CEPH.fsid", - .type = FSCACHE_COOKIE_TYPE_INDEX, -}; - -int __init ceph_fscache_register(void) -{ - return fscache_register_netfs(&ceph_cache_netfs); -} - -void ceph_fscache_unregister(void) -{ - fscache_unregister_netfs(&ceph_cache_netfs); -} - -int ceph_fscache_register_fs(struct ceph_fs_client* fsc, struct fs_context *fc) -{ - const struct ceph_fsid *fsid = &fsc->client->fsid; - const char *fscache_uniq = fsc->mount_options->fscache_uniq; - size_t uniq_len = fscache_uniq ? strlen(fscache_uniq) : 0; - struct ceph_fscache_entry *ent; - int err = 0; - - mutex_lock(&ceph_fscache_lock); - list_for_each_entry(ent, &ceph_fscache_list, list) { - if (memcmp(&ent->fsid, fsid, sizeof(*fsid))) - continue; - if (ent->uniq_len != uniq_len) - continue; - if (uniq_len && memcmp(ent->uniquifier, fscache_uniq, uniq_len)) - continue; - - errorfc(fc, "fscache cookie already registered for fsid %pU, use fsc=<uniquifier> option", - fsid); - err = -EBUSY; - goto out_unlock; - } - - ent = kzalloc(sizeof(*ent) + uniq_len, GFP_KERNEL); - if (!ent) { - err = -ENOMEM; - goto out_unlock; - } - - memcpy(&ent->fsid, fsid, sizeof(*fsid)); - if (uniq_len > 0) { - memcpy(&ent->uniquifier, fscache_uniq, uniq_len); - ent->uniq_len = uniq_len; - } - - fsc->fscache = fscache_acquire_cookie(ceph_cache_netfs.primary_index, - &ceph_fscache_fsid_object_def, - &ent->fsid, sizeof(ent->fsid) + uniq_len, - NULL, 0, - fsc, 0, true); - - if (fsc->fscache) { - ent->fscache = fsc->fscache; - list_add_tail(&ent->list, &ceph_fscache_list); - } else { - kfree(ent); - errorfc(fc, "unable to register fscache cookie for fsid %pU", - fsid); - /* all other fs ignore this error */ - } -out_unlock: - mutex_unlock(&ceph_fscache_lock); - return err; -} - -static enum fscache_checkaux ceph_fscache_inode_check_aux( - void *cookie_netfs_data, const void *data, uint16_t dlen, - loff_t object_size) -{ - struct ceph_aux_inode aux; - struct ceph_inode_info* ci = cookie_netfs_data; - struct inode* inode = &ci->vfs_inode; - - if (dlen != sizeof(aux) || - i_size_read(inode) != object_size) - return FSCACHE_CHECKAUX_OBSOLETE; - - memset(&aux, 0, sizeof(aux)); - aux.version = ci->i_version; - aux.mtime_sec = inode->i_mtime.tv_sec; - aux.mtime_nsec = inode->i_mtime.tv_nsec; - - if (memcmp(data, &aux, sizeof(aux)) != 0) - return FSCACHE_CHECKAUX_OBSOLETE; - - dout("ceph inode 0x%p cached okay\n", ci); - return FSCACHE_CHECKAUX_OKAY; -} - -static const struct fscache_cookie_def ceph_fscache_inode_object_def = { - .name = "CEPH.inode", - .type = FSCACHE_COOKIE_TYPE_DATAFILE, - .check_aux = ceph_fscache_inode_check_aux, -}; - void ceph_fscache_register_inode_cookie(struct inode *inode) { struct ceph_inode_info *ci = ceph_inode(inode); struct ceph_fs_client *fsc = ceph_inode_to_client(inode); - struct ceph_aux_inode aux; - /* No caching for filesystem */ + /* No caching for filesystem? */ if (!fsc->fscache) return; - /* Only cache for regular files that are read only */ + /* Regular files only */ if (!S_ISREG(inode->i_mode)) return; - inode_lock_nested(inode, I_MUTEX_CHILD); - if (!ci->fscache) { - memset(&aux, 0, sizeof(aux)); - aux.version = ci->i_version; - aux.mtime_sec = inode->i_mtime.tv_sec; - aux.mtime_nsec = inode->i_mtime.tv_nsec; - ci->fscache = fscache_acquire_cookie(fsc->fscache, - &ceph_fscache_inode_object_def, - &ci->i_vino, sizeof(ci->i_vino), - &aux, sizeof(aux), - ci, i_size_read(inode), false); - } - inode_unlock(inode); -} - -void ceph_fscache_unregister_inode_cookie(struct ceph_inode_info* ci) -{ - struct fscache_cookie* cookie; - - if ((cookie = ci->fscache) == NULL) + /* Only new inodes! */ + if (!(inode->i_state & I_NEW)) return; - ci->fscache = NULL; + WARN_ON_ONCE(ci->netfs.cache); - fscache_uncache_all_inode_pages(cookie, &ci->vfs_inode); - fscache_relinquish_cookie(cookie, &ci->i_vino, false); + ci->netfs.cache = + fscache_acquire_cookie(fsc->fscache, 0, + &ci->i_vino, sizeof(ci->i_vino), + &ci->i_version, sizeof(ci->i_version), + i_size_read(inode)); } -static bool ceph_fscache_can_enable(void *data) +void ceph_fscache_unregister_inode_cookie(struct ceph_inode_info *ci) { - struct inode *inode = data; - return !inode_is_open_for_write(inode); + fscache_relinquish_cookie(ceph_fscache_cookie(ci), false); } -void ceph_fscache_file_set_cookie(struct inode *inode, struct file *filp) +void ceph_fscache_use_cookie(struct inode *inode, bool will_modify) { 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, &ci->i_vino, false); - fscache_uncache_all_inode_pages(ci->fscache, inode); - } else { - fscache_enable_cookie(ci->fscache, &ci->i_vino, i_size_read(inode), - ceph_fscache_can_enable, inode); - if (fscache_cookie_enabled(ci->fscache)) { - dout("fscache_file_set_cookie %p %p enabling cache\n", - inode, filp); - } - } -} - -static void ceph_readpage_from_fscache_complete(struct page *page, void *data, int error) -{ - if (!error) - SetPageUptodate(page); - - unlock_page(page); -} - -static inline bool cache_valid(struct ceph_inode_info *ci) -{ - return ci->i_fscache_gen == ci->i_rdcache_gen; -} - - -/* Atempt to read from the fscache, - * - * This function is called from the readpage_nounlock context. DO NOT attempt to - * unlock the page here (or in the callback). - */ -int ceph_readpage_from_fscache(struct inode *inode, struct page *page) -{ - struct ceph_inode_info *ci = ceph_inode(inode); - int ret; - - if (!cache_valid(ci)) - return -ENOBUFS; - - ret = fscache_read_or_alloc_page(ci->fscache, page, - ceph_readpage_from_fscache_complete, NULL, - GFP_KERNEL); - - switch (ret) { - case 0: /* Page found */ - dout("page read submitted\n"); - return 0; - case -ENOBUFS: /* Pages were not found, and can't be */ - case -ENODATA: /* Pages were not found */ - dout("page/inode not in cache\n"); - return ret; - default: - dout("%s: unknown error ret = %i\n", __func__, ret); - return ret; - } + fscache_use_cookie(ceph_fscache_cookie(ci), will_modify); } -int ceph_readpages_from_fscache(struct inode *inode, - struct address_space *mapping, - struct list_head *pages, - unsigned *nr_pages) +void ceph_fscache_unuse_cookie(struct inode *inode, bool update) { struct ceph_inode_info *ci = ceph_inode(inode); - int ret; - if (!cache_valid(ci)) - return -ENOBUFS; + if (update) { + loff_t i_size = i_size_read(inode); - ret = fscache_read_or_alloc_pages(ci->fscache, mapping, pages, nr_pages, - ceph_readpage_from_fscache_complete, - NULL, mapping_gfp_mask(mapping)); - - switch (ret) { - case 0: /* All pages found */ - dout("all-page read submitted\n"); - return 0; - case -ENOBUFS: /* Some pages were not found, and can't be */ - case -ENODATA: /* some pages were not found */ - dout("page/inode not in cache\n"); - return ret; - default: - dout("%s: unknown error ret = %i\n", __func__, ret); - return ret; + fscache_unuse_cookie(ceph_fscache_cookie(ci), + &ci->i_version, &i_size); + } else { + fscache_unuse_cookie(ceph_fscache_cookie(ci), NULL, NULL); } } -void ceph_readpage_to_fscache(struct inode *inode, struct page *page) +void ceph_fscache_update(struct inode *inode) { struct ceph_inode_info *ci = ceph_inode(inode); - int ret; - - if (!PageFsCache(page)) - return; - - if (!cache_valid(ci)) - return; + loff_t i_size = i_size_read(inode); - ret = fscache_write_page(ci->fscache, page, i_size_read(inode), - GFP_KERNEL); - if (ret) - fscache_uncache_page(ci->fscache, page); + fscache_update_cookie(ceph_fscache_cookie(ci), &ci->i_version, &i_size); } -void ceph_invalidate_fscache_page(struct inode* inode, struct page *page) +void ceph_fscache_invalidate(struct inode *inode, bool dio_write) { struct ceph_inode_info *ci = ceph_inode(inode); - if (!PageFsCache(page)) - return; - - fscache_wait_on_page_write(ci->fscache, page); - fscache_uncache_page(ci->fscache, page); + fscache_invalidate(ceph_fscache_cookie(ci), + &ci->i_version, i_size_read(inode), + dio_write ? FSCACHE_INVAL_DIO_WRITE : 0); } -void ceph_fscache_unregister_fs(struct ceph_fs_client* fsc) +int ceph_fscache_register_fs(struct ceph_fs_client* fsc, struct fs_context *fc) { - if (fscache_cookie_valid(fsc->fscache)) { - struct ceph_fscache_entry *ent; - bool found = false; + const struct ceph_fsid *fsid = &fsc->client->fsid; + const char *fscache_uniq = fsc->mount_options->fscache_uniq; + size_t uniq_len = fscache_uniq ? strlen(fscache_uniq) : 0; + char *name; + int err = 0; - mutex_lock(&ceph_fscache_lock); - list_for_each_entry(ent, &ceph_fscache_list, list) { - if (ent->fscache == fsc->fscache) { - list_del(&ent->list); - kfree(ent); - found = true; - break; - } - } - WARN_ON_ONCE(!found); - mutex_unlock(&ceph_fscache_lock); + name = kasprintf(GFP_KERNEL, "ceph,%pU%s%s", fsid, uniq_len ? "," : "", + uniq_len ? fscache_uniq : ""); + if (!name) + return -ENOMEM; - __fscache_relinquish_cookie(fsc->fscache, NULL, false); + fsc->fscache = fscache_acquire_volume(name, NULL, NULL, 0); + if (IS_ERR_OR_NULL(fsc->fscache)) { + errorfc(fc, "Unable to register fscache cookie for %s", name); + err = fsc->fscache ? PTR_ERR(fsc->fscache) : -EOPNOTSUPP; + fsc->fscache = NULL; } - fsc->fscache = NULL; + kfree(name); + return err; } -/* - * caller should hold CEPH_CAP_FILE_{RD,CACHE} - */ -void ceph_fscache_revalidate_cookie(struct ceph_inode_info *ci) +void ceph_fscache_unregister_fs(struct ceph_fs_client* fsc) { - if (cache_valid(ci)) - return; - - /* 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, &ci->i_vino)) - fscache_invalidate(ci->fscache); - spin_lock(&ci->i_ceph_lock); - ci->i_fscache_gen = ci->i_rdcache_gen; - spin_unlock(&ci->i_ceph_lock); - } - mutex_unlock(&ci->i_truncate_mutex); + fscache_relinquish_volume(fsc->fscache, NULL, false); } diff --git a/fs/ceph/cache.h b/fs/ceph/cache.h index 89dbdd1eb14a..dc502daac49a 100644 --- a/fs/ceph/cache.h +++ b/fs/ceph/cache.h @@ -9,85 +9,73 @@ #ifndef _CEPH_CACHE_H #define _CEPH_CACHE_H -#ifdef CONFIG_CEPH_FSCACHE - -extern struct fscache_netfs ceph_cache_netfs; +#include <linux/netfs.h> -int ceph_fscache_register(void); -void ceph_fscache_unregister(void); +#ifdef CONFIG_CEPH_FSCACHE +#include <linux/fscache.h> int ceph_fscache_register_fs(struct ceph_fs_client* fsc, struct fs_context *fc); void ceph_fscache_unregister_fs(struct ceph_fs_client* fsc); 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, - struct address_space *mapping, - struct list_head *pages, - 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_fscache_use_cookie(struct inode *inode, bool will_modify); +void ceph_fscache_unuse_cookie(struct inode *inode, bool update); -static inline void ceph_fscache_inode_init(struct ceph_inode_info *ci) -{ - ci->fscache = NULL; - ci->i_fscache_gen = 0; -} +void ceph_fscache_update(struct inode *inode); +void ceph_fscache_invalidate(struct inode *inode, bool dio_write); -static inline void ceph_fscache_invalidate(struct inode *inode) +static inline struct fscache_cookie *ceph_fscache_cookie(struct ceph_inode_info *ci) { - fscache_invalidate(ceph_inode(inode)->fscache); + return netfs_i_cookie(&ci->netfs); } -static inline void ceph_fscache_uncache_page(struct inode *inode, - struct page *page) +static inline void ceph_fscache_resize(struct inode *inode, loff_t to) { struct ceph_inode_info *ci = ceph_inode(inode); - return fscache_uncache_page(ci->fscache, page); -} + struct fscache_cookie *cookie = ceph_fscache_cookie(ci); -static inline int ceph_release_fscache_page(struct page *page, gfp_t gfp) -{ - struct inode* inode = page->mapping->host; - struct ceph_inode_info *ci = ceph_inode(inode); - return fscache_maybe_release_page(ci->fscache, page, gfp); + if (cookie) { + ceph_fscache_use_cookie(inode, true); + fscache_resize_cookie(cookie, to); + ceph_fscache_unuse_cookie(inode, true); + } } -static inline void ceph_fscache_readpage_cancel(struct inode *inode, - struct page *page) +static inline void ceph_fscache_unpin_writeback(struct inode *inode, + struct writeback_control *wbc) { - struct ceph_inode_info *ci = ceph_inode(inode); - if (fscache_cookie_valid(ci->fscache) && PageFsCache(page)) - __fscache_uncache_page(ci->fscache, page); + fscache_unpin_writeback(wbc, ceph_fscache_cookie(ceph_inode(inode))); } -static inline void ceph_fscache_readpages_cancel(struct inode *inode, - struct list_head *pages) +static inline int ceph_fscache_dirty_folio(struct address_space *mapping, + struct folio *folio) { - struct ceph_inode_info *ci = ceph_inode(inode); - return fscache_readpages_cancel(ci->fscache, pages); + struct ceph_inode_info *ci = ceph_inode(mapping->host); + + return fscache_dirty_folio(mapping, folio, ceph_fscache_cookie(ci)); } -static inline void ceph_disable_fscache_readpage(struct ceph_inode_info *ci) +static inline int ceph_begin_cache_operation(struct netfs_io_request *rreq) { - ci->i_fscache_gen = ci->i_rdcache_gen - 1; -} + struct fscache_cookie *cookie = ceph_fscache_cookie(ceph_inode(rreq->inode)); -#else + return fscache_begin_read_operation(&rreq->cache_resources, cookie); +} -static inline int ceph_fscache_register(void) +static inline bool ceph_is_cache_enabled(struct inode *inode) { - return 0; + return fscache_cookie_enabled(ceph_fscache_cookie(ceph_inode(inode))); } -static inline void ceph_fscache_unregister(void) +static inline void ceph_fscache_note_page_release(struct inode *inode) { -} + struct ceph_inode_info *ci = ceph_inode(inode); + fscache_note_page_release(ceph_fscache_cookie(ci)); +} +#else /* CONFIG_CEPH_FSCACHE */ static inline int ceph_fscache_register_fs(struct ceph_fs_client* fsc, struct fs_context *fc) { @@ -98,10 +86,6 @@ static inline void ceph_fscache_unregister_fs(struct ceph_fs_client* fsc) { } -static inline void ceph_fscache_inode_init(struct ceph_inode_info *ci) -{ -} - static inline void ceph_fscache_register_inode_cookie(struct inode *inode) { } @@ -110,67 +94,55 @@ static inline void ceph_fscache_unregister_inode_cookie(struct ceph_inode_info* { } -static inline void ceph_fscache_file_set_cookie(struct inode *inode, - struct file *filp) -{ -} - -static inline void ceph_fscache_revalidate_cookie(struct ceph_inode_info *ci) +static inline void ceph_fscache_use_cookie(struct inode *inode, bool will_modify) { } -static inline void ceph_fscache_uncache_page(struct inode *inode, - struct page *pages) +static inline void ceph_fscache_unuse_cookie(struct inode *inode, bool update) { } -static inline int ceph_readpage_from_fscache(struct inode* inode, - struct page *page) +static inline void ceph_fscache_update(struct inode *inode) { - return -ENOBUFS; } -static inline int ceph_readpages_from_fscache(struct inode *inode, - struct address_space *mapping, - struct list_head *pages, - unsigned *nr_pages) +static inline void ceph_fscache_invalidate(struct inode *inode, bool dio_write) { - return -ENOBUFS; } -static inline void ceph_readpage_to_fscache(struct inode *inode, - struct page *page) +static inline struct fscache_cookie *ceph_fscache_cookie(struct ceph_inode_info *ci) { + return NULL; } -static inline void ceph_fscache_invalidate(struct inode *inode) +static inline void ceph_fscache_resize(struct inode *inode, loff_t to) { } -static inline void ceph_invalidate_fscache_page(struct inode *inode, - struct page *page) +static inline void ceph_fscache_unpin_writeback(struct inode *inode, + struct writeback_control *wbc) { } -static inline int ceph_release_fscache_page(struct page *page, gfp_t gfp) +static inline int ceph_fscache_dirty_folio(struct address_space *mapping, + struct folio *folio) { - return 1; + return filemap_dirty_folio(mapping, folio); } -static inline void ceph_fscache_readpage_cancel(struct inode *inode, - struct page *page) +static inline bool ceph_is_cache_enabled(struct inode *inode) { + return false; } -static inline void ceph_fscache_readpages_cancel(struct inode *inode, - struct list_head *pages) +static inline int ceph_begin_cache_operation(struct netfs_io_request *rreq) { + return -ENOBUFS; } -static inline void ceph_disable_fscache_readpage(struct ceph_inode_info *ci) +static inline void ceph_fscache_note_page_release(struct inode *inode) { } - -#endif +#endif /* CONFIG_CEPH_FSCACHE */ #endif diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index 28ae0c134700..fb023f9fafcb 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -490,13 +490,10 @@ static void __cap_set_timeouts(struct ceph_mds_client *mdsc, struct ceph_inode_info *ci) { struct ceph_mount_options *opt = mdsc->fsc->mount_options; - - ci->i_hold_caps_min = round_jiffies(jiffies + - opt->caps_wanted_delay_min * HZ); ci->i_hold_caps_max = round_jiffies(jiffies + opt->caps_wanted_delay_max * HZ); - dout("__cap_set_timeouts %p min %lu max %lu\n", &ci->vfs_inode, - ci->i_hold_caps_min - jiffies, ci->i_hold_caps_max - jiffies); + dout("__cap_set_timeouts %p %lu\n", &ci->netfs.inode, + ci->i_hold_caps_max - jiffies); } /* @@ -508,10 +505,9 @@ static void __cap_set_timeouts(struct ceph_mds_client *mdsc, * -> we take mdsc->cap_delay_lock */ static void __cap_delay_requeue(struct ceph_mds_client *mdsc, - struct ceph_inode_info *ci, - bool set_timeout) + struct ceph_inode_info *ci) { - dout("__cap_delay_requeue %p flags %d at %lu\n", &ci->vfs_inode, + dout("__cap_delay_requeue %p flags 0x%lx at %lu\n", &ci->netfs.inode, ci->i_ceph_flags, ci->i_hold_caps_max); if (!mdsc->stopping) { spin_lock(&mdsc->cap_delay_lock); @@ -520,8 +516,7 @@ static void __cap_delay_requeue(struct ceph_mds_client *mdsc, goto no_change; list_del_init(&ci->i_cap_delay_list); } - if (set_timeout) - __cap_set_timeouts(mdsc, ci); + __cap_set_timeouts(mdsc, ci); list_add_tail(&ci->i_cap_delay_list, &mdsc->cap_delay_list); no_change: spin_unlock(&mdsc->cap_delay_lock); @@ -536,7 +531,7 @@ no_change: static void __cap_delay_requeue_front(struct ceph_mds_client *mdsc, struct ceph_inode_info *ci) { - dout("__cap_delay_requeue_front %p\n", &ci->vfs_inode); + dout("__cap_delay_requeue_front %p\n", &ci->netfs.inode); spin_lock(&mdsc->cap_delay_lock); ci->i_ceph_flags |= CEPH_I_FLUSH; if (!list_empty(&ci->i_cap_delay_list)) @@ -553,7 +548,7 @@ static void __cap_delay_requeue_front(struct ceph_mds_client *mdsc, static void __cap_delay_cancel(struct ceph_mds_client *mdsc, struct ceph_inode_info *ci) { - dout("__cap_delay_cancel %p\n", &ci->vfs_inode); + dout("__cap_delay_cancel %p\n", &ci->netfs.inode); if (list_empty(&ci->i_cap_delay_list)) return; spin_lock(&mdsc->cap_delay_lock); @@ -561,19 +556,20 @@ static void __cap_delay_cancel(struct ceph_mds_client *mdsc, spin_unlock(&mdsc->cap_delay_lock); } -/* - * Common issue checks for add_cap, handle_cap_grant. - */ +/* Common issue checks for add_cap, handle_cap_grant. */ static void __check_cap_issue(struct ceph_inode_info *ci, struct ceph_cap *cap, unsigned issued) { unsigned had = __ceph_caps_issued(ci, NULL); + lockdep_assert_held(&ci->i_ceph_lock); + /* * Each time we receive FILE_CACHE anew, we increment * i_rdcache_gen. */ - if ((issued & (CEPH_CAP_FILE_CACHE|CEPH_CAP_FILE_LAZYIO)) && + if (S_ISREG(ci->netfs.inode.i_mode) && + (issued & (CEPH_CAP_FILE_CACHE|CEPH_CAP_FILE_LAZYIO)) && (had & (CEPH_CAP_FILE_CACHE|CEPH_CAP_FILE_LAZYIO)) == 0) { ci->i_rdcache_gen++; } @@ -587,11 +583,39 @@ static void __check_cap_issue(struct ceph_inode_info *ci, struct ceph_cap *cap, if ((issued & CEPH_CAP_FILE_SHARED) != (had & CEPH_CAP_FILE_SHARED)) { if (issued & CEPH_CAP_FILE_SHARED) atomic_inc(&ci->i_shared_gen); - if (S_ISDIR(ci->vfs_inode.i_mode)) { - dout(" marking %p NOT complete\n", &ci->vfs_inode); + if (S_ISDIR(ci->netfs.inode.i_mode)) { + dout(" marking %p NOT complete\n", &ci->netfs.inode); __ceph_dir_clear_complete(ci); } } + + /* Wipe saved layout if we're losing DIR_CREATE caps */ + if (S_ISDIR(ci->netfs.inode.i_mode) && (had & CEPH_CAP_DIR_CREATE) && + !(issued & CEPH_CAP_DIR_CREATE)) { + ceph_put_string(rcu_dereference_raw(ci->i_cached_layout.pool_ns)); + memset(&ci->i_cached_layout, 0, sizeof(ci->i_cached_layout)); + } +} + +/** + * change_auth_cap_ses - move inode to appropriate lists when auth caps change + * @ci: inode to be moved + * @session: new auth caps session + */ +void change_auth_cap_ses(struct ceph_inode_info *ci, + struct ceph_mds_session *session) +{ + lockdep_assert_held(&ci->i_ceph_lock); + + if (list_empty(&ci->i_dirty_item) && list_empty(&ci->i_flushing_item)) + return; + + spin_lock(&session->s_mdsc->cap_dirty_lock); + if (!list_empty(&ci->i_dirty_item)) + list_move(&ci->i_dirty_item, &session->s_cap_dirty); + if (!list_empty(&ci->i_flushing_item)) + list_move_tail(&ci->i_flushing_item, &session->s_cap_flushing); + spin_unlock(&session->s_mdsc->cap_dirty_lock); } /* @@ -605,7 +629,7 @@ static void __check_cap_issue(struct ceph_inode_info *ci, struct ceph_cap *cap, */ void ceph_add_cap(struct inode *inode, struct ceph_mds_session *session, u64 cap_id, - int fmode, unsigned issued, unsigned wanted, + unsigned issued, unsigned wanted, unsigned seq, unsigned mseq, u64 realmino, int flags, struct ceph_cap **new_cap) { @@ -621,16 +645,7 @@ void ceph_add_cap(struct inode *inode, dout("add_cap %p mds%d cap %llx %s seq %d\n", inode, session->s_mds, cap_id, ceph_cap_string(issued), seq); - /* - * If we are opening the file, include file mode wanted bits - * in wanted. - */ - if (fmode >= 0) - wanted |= ceph_caps_for_mode(fmode); - - spin_lock(&session->s_gen_ttl_lock); - gen = session->s_cap_gen; - spin_unlock(&session->s_gen_ttl_lock); + gen = atomic_read(&session->s_cap_gen); cap = __get_cap_for_mds(ci, mds); if (!cap) { @@ -651,6 +666,7 @@ void ceph_add_cap(struct inode *inode, spin_lock(&session->s_cap_lock); list_add_tail(&cap->session_caps, &session->s_caps); session->s_nr_caps++; + atomic64_inc(&mdsc->metric.total_caps); spin_unlock(&session->s_cap_lock); } else { spin_lock(&session->s_cap_lock); @@ -687,29 +703,12 @@ void ceph_add_cap(struct inode *inode, */ struct ceph_snap_realm *realm = ceph_lookup_snap_realm(mdsc, realmino); - if (realm) { - struct ceph_snap_realm *oldrealm = ci->i_snap_realm; - if (oldrealm) { - spin_lock(&oldrealm->inodes_with_caps_lock); - list_del_init(&ci->i_snap_realm_item); - spin_unlock(&oldrealm->inodes_with_caps_lock); - } - - spin_lock(&realm->inodes_with_caps_lock); - list_add(&ci->i_snap_realm_item, - &realm->inodes_with_caps); - ci->i_snap_realm = realm; - if (realm->ino == ci->i_vino.ino) - realm->inode = inode; - spin_unlock(&realm->inodes_with_caps_lock); - - if (oldrealm) - ceph_put_snap_realm(mdsc, oldrealm); - } else { - pr_err("ceph_add_cap: couldn't find snap realm %llx\n", - realmino); - WARN_ON(!realm); - } + if (realm) + ceph_change_snap_realm(inode, realm); + else + WARN(1, "%s: couldn't find snap realm 0x%llx (ino 0x%llx oldrealm 0x%llx)\n", + __func__, realmino, ci->i_vino.ino, + ci->i_snap_realm ? ci->i_snap_realm->ino : 0); } __check_cap_issue(ci, cap, issued); @@ -725,12 +724,15 @@ void ceph_add_cap(struct inode *inode, dout(" issued %s, mds wanted %s, actual %s, queueing\n", ceph_cap_string(issued), ceph_cap_string(wanted), ceph_cap_string(actual_wanted)); - __cap_delay_requeue(mdsc, ci, true); + __cap_delay_requeue(mdsc, ci); } if (flags & CEPH_CAP_FLAG_AUTH) { if (!ci->i_auth_cap || ceph_seq_cmp(ci->i_auth_cap->mseq, mseq) < 0) { + if (ci->i_auth_cap && + ci->i_auth_cap->session != cap->session) + change_auth_cap_ses(ci, cap->session); ci->i_auth_cap = cap; cap->mds_wanted = wanted; } @@ -752,9 +754,7 @@ void ceph_add_cap(struct inode *inode, cap->issue_seq = seq; cap->mseq = mseq; cap->cap_gen = gen; - - if (fmode >= 0) - __ceph_get_fmode(ci, fmode); + wake_up_all(&ci->i_cap_wq); } /* @@ -767,14 +767,12 @@ static int __cap_is_valid(struct ceph_cap *cap) unsigned long ttl; u32 gen; - spin_lock(&cap->session->s_gen_ttl_lock); - gen = cap->session->s_cap_gen; + gen = atomic_read(&cap->session->s_cap_gen); ttl = cap->session->s_cap_ttl; - spin_unlock(&cap->session->s_gen_ttl_lock); if (cap->cap_gen < gen || time_after_eq(jiffies, ttl)) { dout("__cap_is_valid %p cap %p issued %s " - "but STALE (gen %u vs %u)\n", &cap->ci->vfs_inode, + "but STALE (gen %u vs %u)\n", &cap->ci->netfs.inode, cap, ceph_cap_string(cap->issued), cap->cap_gen, gen); return 0; } @@ -800,7 +798,7 @@ int __ceph_caps_issued(struct ceph_inode_info *ci, int *implemented) if (!__cap_is_valid(cap)) continue; dout("__ceph_caps_issued %p cap %p issued %s\n", - &ci->vfs_inode, cap, ceph_cap_string(cap->issued)); + &ci->netfs.inode, cap, ceph_cap_string(cap->issued)); have |= cap->issued; if (implemented) *implemented |= cap->implemented; @@ -847,12 +845,12 @@ static void __touch_cap(struct ceph_cap *cap) spin_lock(&s->s_cap_lock); if (!s->s_cap_iterator) { - dout("__touch_cap %p cap %p mds%d\n", &cap->ci->vfs_inode, cap, + dout("__touch_cap %p cap %p mds%d\n", &cap->ci->netfs.inode, cap, s->s_mds); list_move_tail(&cap->session_caps, &s->s_caps); } else { dout("__touch_cap %p cap %p mds%d NOP, iterating over caps\n", - &cap->ci->vfs_inode, cap, s->s_mds); + &cap->ci->netfs.inode, cap, s->s_mds); } spin_unlock(&s->s_cap_lock); } @@ -869,8 +867,8 @@ int __ceph_caps_issued_mask(struct ceph_inode_info *ci, int mask, int touch) int have = ci->i_snap_caps; if ((have & mask) == mask) { - dout("__ceph_caps_issued_mask ino 0x%lx snap issued %s" - " (mask %s)\n", ci->vfs_inode.i_ino, + dout("__ceph_caps_issued_mask ino 0x%llx snap issued %s" + " (mask %s)\n", ceph_ino(&ci->netfs.inode), ceph_cap_string(have), ceph_cap_string(mask)); return 1; @@ -881,8 +879,8 @@ int __ceph_caps_issued_mask(struct ceph_inode_info *ci, int mask, int touch) if (!__cap_is_valid(cap)) continue; if ((cap->issued & mask) == mask) { - dout("__ceph_caps_issued_mask ino 0x%lx cap %p issued %s" - " (mask %s)\n", ci->vfs_inode.i_ino, cap, + dout("__ceph_caps_issued_mask ino 0x%llx cap %p issued %s" + " (mask %s)\n", ceph_ino(&ci->netfs.inode), cap, ceph_cap_string(cap->issued), ceph_cap_string(mask)); if (touch) @@ -893,8 +891,8 @@ int __ceph_caps_issued_mask(struct ceph_inode_info *ci, int mask, int touch) /* does a combination of caps satisfy mask? */ have |= cap->issued; if ((have & mask) == mask) { - dout("__ceph_caps_issued_mask ino 0x%lx combo issued %s" - " (mask %s)\n", ci->vfs_inode.i_ino, + dout("__ceph_caps_issued_mask ino 0x%llx combo issued %s" + " (mask %s)\n", ceph_ino(&ci->netfs.inode), ceph_cap_string(cap->issued), ceph_cap_string(mask)); if (touch) { @@ -919,6 +917,20 @@ int __ceph_caps_issued_mask(struct ceph_inode_info *ci, int mask, int touch) return 0; } +int __ceph_caps_issued_mask_metric(struct ceph_inode_info *ci, int mask, + int touch) +{ + struct ceph_fs_client *fsc = ceph_sb_to_client(ci->netfs.inode.i_sb); + int r; + + r = __ceph_caps_issued_mask(ci, mask, touch); + if (r) + ceph_update_cap_hit(&fsc->mdsc->metric); + else + ceph_update_cap_mis(&fsc->mdsc->metric); + return r; +} + /* * Return true if mask caps are currently being revoked by an MDS. */ @@ -939,7 +951,7 @@ int __ceph_caps_revoking_other(struct ceph_inode_info *ci, int ceph_caps_revoking(struct ceph_inode_info *ci, int mask) { - struct inode *inode = &ci->vfs_inode; + struct inode *inode = &ci->netfs.inode; int ret; spin_lock(&ci->i_ceph_lock); @@ -958,29 +970,97 @@ int __ceph_caps_used(struct ceph_inode_info *ci) if (ci->i_rd_ref) used |= CEPH_CAP_FILE_RD; if (ci->i_rdcache_ref || - (!S_ISDIR(ci->vfs_inode.i_mode) && /* ignore readdir cache */ - ci->vfs_inode.i_data.nrpages)) + (S_ISREG(ci->netfs.inode.i_mode) && + ci->netfs.inode.i_data.nrpages)) used |= CEPH_CAP_FILE_CACHE; if (ci->i_wr_ref) used |= CEPH_CAP_FILE_WR; if (ci->i_wb_ref || ci->i_wrbuffer_ref) used |= CEPH_CAP_FILE_BUFFER; + if (ci->i_fx_ref) + used |= CEPH_CAP_FILE_EXCL; return used; } +#define FMODE_WAIT_BIAS 1000 + /* * wanted, by virtue of open file modes */ int __ceph_caps_file_wanted(struct ceph_inode_info *ci) { - int i, bits = 0; - for (i = 0; i < CEPH_FILE_MODE_BITS; i++) { - if (ci->i_nr_by_mode[i]) - bits |= 1 << i; + const int PIN_SHIFT = ffs(CEPH_FILE_MODE_PIN); + const int RD_SHIFT = ffs(CEPH_FILE_MODE_RD); + const int WR_SHIFT = ffs(CEPH_FILE_MODE_WR); + const int LAZY_SHIFT = ffs(CEPH_FILE_MODE_LAZY); + struct ceph_mount_options *opt = + ceph_inode_to_client(&ci->netfs.inode)->mount_options; + unsigned long used_cutoff = jiffies - opt->caps_wanted_delay_max * HZ; + unsigned long idle_cutoff = jiffies - opt->caps_wanted_delay_min * HZ; + + if (S_ISDIR(ci->netfs.inode.i_mode)) { + int want = 0; + + /* use used_cutoff here, to keep dir's wanted caps longer */ + if (ci->i_nr_by_mode[RD_SHIFT] > 0 || + time_after(ci->i_last_rd, used_cutoff)) + want |= CEPH_CAP_ANY_SHARED; + + if (ci->i_nr_by_mode[WR_SHIFT] > 0 || + time_after(ci->i_last_wr, used_cutoff)) { + want |= CEPH_CAP_ANY_SHARED | CEPH_CAP_FILE_EXCL; + if (opt->flags & CEPH_MOUNT_OPT_ASYNC_DIROPS) + want |= CEPH_CAP_ANY_DIR_OPS; + } + + if (want || ci->i_nr_by_mode[PIN_SHIFT] > 0) + want |= CEPH_CAP_PIN; + + return want; + } else { + int bits = 0; + + if (ci->i_nr_by_mode[RD_SHIFT] > 0) { + if (ci->i_nr_by_mode[RD_SHIFT] >= FMODE_WAIT_BIAS || + time_after(ci->i_last_rd, used_cutoff)) + bits |= 1 << RD_SHIFT; + } else if (time_after(ci->i_last_rd, idle_cutoff)) { + bits |= 1 << RD_SHIFT; + } + + if (ci->i_nr_by_mode[WR_SHIFT] > 0) { + if (ci->i_nr_by_mode[WR_SHIFT] >= FMODE_WAIT_BIAS || + time_after(ci->i_last_wr, used_cutoff)) + bits |= 1 << WR_SHIFT; + } else if (time_after(ci->i_last_wr, idle_cutoff)) { + bits |= 1 << WR_SHIFT; + } + + /* check lazyio only when read/write is wanted */ + if ((bits & (CEPH_FILE_MODE_RDWR << 1)) && + ci->i_nr_by_mode[LAZY_SHIFT] > 0) + bits |= 1 << LAZY_SHIFT; + + return bits ? ceph_caps_for_mode(bits >> 1) : 0; } - if (bits == 0) - return 0; - return ceph_caps_for_mode(bits >> 1); +} + +/* + * wanted, by virtue of open file modes AND cap refs (buffered/cached data) + */ +int __ceph_caps_wanted(struct ceph_inode_info *ci) +{ + int w = __ceph_caps_file_wanted(ci) | __ceph_caps_used(ci); + if (S_ISDIR(ci->netfs.inode.i_mode)) { + /* we want EXCL if holding caps of dir ops */ + if (w & CEPH_CAP_ANY_DIR_OPS) + w |= CEPH_CAP_FILE_EXCL; + } else { + /* we want EXCL if dirty data */ + if (w & CEPH_CAP_FILE_BUFFER) + w |= CEPH_CAP_FILE_EXCL; + } + return w; } /* @@ -1004,14 +1084,6 @@ int __ceph_caps_mds_wanted(struct ceph_inode_info *ci, bool check) return mds_wanted; } -/* - * called under i_ceph_lock - */ -static int __ceph_is_single_caps(struct ceph_inode_info *ci) -{ - return rb_first(&ci->i_caps) == rb_last(&ci->i_caps); -} - int ceph_is_any_caps(struct inode *inode) { struct ceph_inode_info *ci = ceph_inode(inode); @@ -1024,20 +1096,6 @@ int ceph_is_any_caps(struct inode *inode) return ret; } -static void drop_inode_snap_realm(struct ceph_inode_info *ci) -{ - struct ceph_snap_realm *realm = ci->i_snap_realm; - spin_lock(&realm->inodes_with_caps_lock); - list_del_init(&ci->i_snap_realm_item); - ci->i_snap_realm_counter++; - ci->i_snap_realm = NULL; - if (realm->ino == ci->i_vino.ino) - realm->inode = NULL; - spin_unlock(&realm->inodes_with_caps_lock); - ceph_put_snap_realm(ceph_sb_to_client(ci->vfs_inode.i_sb)->mdsc, - realm); -} - /* * Remove a cap. Take steps to deal with a racing iterate_session_caps. * @@ -1048,11 +1106,20 @@ void __ceph_remove_cap(struct ceph_cap *cap, bool queue_release) { struct ceph_mds_session *session = cap->session; struct ceph_inode_info *ci = cap->ci; - struct ceph_mds_client *mdsc = - ceph_sb_to_client(ci->vfs_inode.i_sb)->mdsc; + struct ceph_mds_client *mdsc; int removed = 0; - dout("__ceph_remove_cap %p from %p\n", cap, &ci->vfs_inode); + /* 'ci' being NULL means the remove have already occurred */ + if (!ci) { + dout("%s: cap inode is NULL\n", __func__); + return; + } + + lockdep_assert_held(&ci->i_ceph_lock); + + dout("__ceph_remove_cap %p from %p\n", cap, &ci->netfs.inode); + + mdsc = ceph_inode_to_client(&ci->netfs.inode)->mdsc; /* remove from inode's cap rbtree, and clear auth cap */ rb_erase(&cap->ci_node, &ci->i_caps); @@ -1068,6 +1135,7 @@ void __ceph_remove_cap(struct ceph_cap *cap, bool queue_release) } else { list_del_init(&cap->session_caps); session->s_nr_caps--; + atomic64_dec(&mdsc->metric.total_caps); cap->session = NULL; removed = 1; } @@ -1079,7 +1147,8 @@ void __ceph_remove_cap(struct ceph_cap *cap, bool queue_release) * s_cap_gen while session is in the reconnect state. */ if (queue_release && - (!session->s_cap_reconnect || cap->cap_gen == session->s_cap_gen)) { + (!session->s_cap_reconnect || + cap->cap_gen == atomic_read(&session->s_cap_gen))) { cap->queue_release = 1; if (removed) { __ceph_queue_cap_release(session, cap); @@ -1101,12 +1170,34 @@ void __ceph_remove_cap(struct ceph_cap *cap, bool queue_release) * keep i_snap_realm. */ if (ci->i_wr_ref == 0 && ci->i_snap_realm) - drop_inode_snap_realm(ci); + ceph_change_snap_realm(&ci->netfs.inode, NULL); __cap_delay_cancel(mdsc, ci); } } +void ceph_remove_cap(struct ceph_cap *cap, bool queue_release) +{ + struct ceph_inode_info *ci = cap->ci; + struct ceph_fs_client *fsc; + + /* 'ci' being NULL means the remove have already occurred */ + if (!ci) { + dout("%s: cap inode is NULL\n", __func__); + return; + } + + lockdep_assert_held(&ci->i_ceph_lock); + + fsc = ceph_inode_to_client(&ci->netfs.inode); + WARN_ON_ONCE(ci->i_auth_cap == cap && + !list_empty(&ci->i_dirty_item) && + !fsc->blocklisted && + !ceph_inode_is_shutdown(&ci->netfs.inode)); + + __ceph_remove_cap(cap, queue_release); +} + struct cap_msg_args { struct ceph_mds_session *session; u64 ino, cid, follows; @@ -1114,6 +1205,7 @@ struct cap_msg_args { u64 xattr_version; u64 change_attr; struct ceph_buffer *xattr_buf; + struct ceph_buffer *old_xattr_buf; struct timespec64 atime, mtime, ctime, btime; int op, caps, wanted, dirty; u32 seq, issue_seq, mseq, time_warp_seq; @@ -1122,39 +1214,31 @@ struct cap_msg_args { kgid_t gid; umode_t mode; bool inline_data; + bool wake; }; /* - * Build and send a cap message to the given MDS. - * - * Caller should be holding s_mutex. + * cap struct size + flock buffer size + inline version + inline data size + + * osd_epoch_barrier + oldest_flush_tid */ -static int send_cap_msg(struct cap_msg_args *arg) +#define CAP_MSG_SIZE (sizeof(struct ceph_mds_caps) + \ + 4 + 8 + 4 + 4 + 8 + 4 + 4 + 4 + 8 + 8 + 4) + +/* Marshal up the cap msg to the MDS */ +static void encode_cap_msg(struct ceph_msg *msg, struct cap_msg_args *arg) { struct ceph_mds_caps *fc; - struct ceph_msg *msg; void *p; - size_t extra_len; struct ceph_osd_client *osdc = &arg->session->s_mdsc->fsc->client->osdc; - dout("send_cap_msg %s %llx %llx caps %s wanted %s dirty %s" - " seq %u/%u tid %llu/%llu mseq %u follows %lld size %llu/%llu" - " xattr_ver %llu xattr_len %d\n", ceph_cap_op_name(arg->op), - arg->cid, arg->ino, ceph_cap_string(arg->caps), - ceph_cap_string(arg->wanted), ceph_cap_string(arg->dirty), - arg->seq, arg->issue_seq, arg->flush_tid, arg->oldest_flush_tid, - arg->mseq, arg->follows, arg->size, arg->max_size, - arg->xattr_version, + dout("%s %s %llx %llx caps %s wanted %s dirty %s seq %u/%u tid %llu/%llu mseq %u follows %lld size %llu/%llu xattr_ver %llu xattr_len %d\n", + __func__, ceph_cap_op_name(arg->op), arg->cid, arg->ino, + ceph_cap_string(arg->caps), ceph_cap_string(arg->wanted), + ceph_cap_string(arg->dirty), arg->seq, arg->issue_seq, + arg->flush_tid, arg->oldest_flush_tid, arg->mseq, arg->follows, + arg->size, arg->max_size, arg->xattr_version, arg->xattr_buf ? (int)arg->xattr_buf->vec.iov_len : 0); - /* flock buffer size + inline version + inline data size + - * osd_epoch_barrier + oldest_flush_tid */ - extra_len = 4 + 8 + 4 + 4 + 8 + 4 + 4 + 4 + 8 + 8 + 4; - msg = ceph_msg_new(CEPH_MSG_CLIENT_CAPS, sizeof(*fc) + extra_len, - GFP_NOFS, false); - if (!msg) - return -ENOMEM; - msg->hdr.version = cpu_to_le16(10); msg->hdr.tid = cpu_to_le64(arg->flush_tid); @@ -1226,9 +1310,6 @@ static int send_cap_msg(struct cap_msg_args *arg) /* Advisory flags (version 10) */ ceph_encode_32(&p, arg->flags); - - ceph_con_send(&arg->session->s_con, msg); - return 0; } /* @@ -1245,124 +1326,98 @@ void __ceph_remove_caps(struct ceph_inode_info *ci) while (p) { struct ceph_cap *cap = rb_entry(p, struct ceph_cap, ci_node); p = rb_next(p); - __ceph_remove_cap(cap, true); + ceph_remove_cap(cap, true); } spin_unlock(&ci->i_ceph_lock); } /* - * Send a cap msg on the given inode. Update our caps state, then - * drop i_ceph_lock and send the message. + * Prepare to send a cap message to an MDS. Update the cap state, and populate + * the arg struct with the parameters that will need to be sent. This should + * be done under the i_ceph_lock to guard against changes to cap state. * * Make note of max_size reported/requested from mds, revoked caps * that have now been implemented. - * - * Return non-zero if delayed release, or we experienced an error - * such that the caller should requeue + retry later. - * - * called with i_ceph_lock, then drops it. - * caller should hold snap_rwsem (read), s_mutex. */ -static int __send_cap(struct ceph_mds_client *mdsc, struct ceph_cap *cap, - int op, int flags, int used, int want, int retain, - int flushing, u64 flush_tid, u64 oldest_flush_tid) - __releases(cap->ci->i_ceph_lock) +static void __prep_cap(struct cap_msg_args *arg, struct ceph_cap *cap, + int op, int flags, int used, int want, int retain, + int flushing, u64 flush_tid, u64 oldest_flush_tid) { struct ceph_inode_info *ci = cap->ci; - struct inode *inode = &ci->vfs_inode; - struct ceph_buffer *old_blob = NULL; - struct cap_msg_args arg; + struct inode *inode = &ci->netfs.inode; int held, revoking; - int wake = 0; - int delayed = 0; - int ret; + + lockdep_assert_held(&ci->i_ceph_lock); held = cap->issued | cap->implemented; revoking = cap->implemented & ~cap->issued; retain &= ~revoking; - dout("__send_cap %p cap %p session %p %s -> %s (revoking %s)\n", - inode, cap, cap->session, + dout("%s %p cap %p session %p %s -> %s (revoking %s)\n", + __func__, inode, cap, cap->session, ceph_cap_string(held), ceph_cap_string(held & retain), ceph_cap_string(revoking)); BUG_ON((retain & CEPH_CAP_PIN) == 0); - arg.session = cap->session; - - /* don't release wanted unless we've waited a bit. */ - if ((ci->i_ceph_flags & CEPH_I_NODELAY) == 0 && - time_before(jiffies, ci->i_hold_caps_min)) { - dout(" delaying issued %s -> %s, wanted %s -> %s on send\n", - ceph_cap_string(cap->issued), - ceph_cap_string(cap->issued & retain), - ceph_cap_string(cap->mds_wanted), - ceph_cap_string(want)); - want |= cap->mds_wanted; - retain |= cap->issued; - delayed = 1; - } - ci->i_ceph_flags &= ~(CEPH_I_NODELAY | CEPH_I_FLUSH); - if (want & ~cap->mds_wanted) { - /* user space may open/close single file frequently. - * This avoids droping mds_wanted immediately after - * requesting new mds_wanted. - */ - __cap_set_timeouts(mdsc, ci); - } + ci->i_ceph_flags &= ~CEPH_I_FLUSH; cap->issued &= retain; /* drop bits we don't want */ - if (cap->implemented & ~cap->issued) { - /* - * Wake up any waiters on wanted -> needed transition. - * This is due to the weird transition from buffered - * to sync IO... we need to flush dirty pages _before_ - * allowing sync writes to avoid reordering. - */ - wake = 1; - } + /* + * Wake up any waiters on wanted -> needed transition. This is due to + * the weird transition from buffered to sync IO... we need to flush + * dirty pages _before_ allowing sync writes to avoid reordering. + */ + arg->wake = cap->implemented & ~cap->issued; cap->implemented &= cap->issued | used; cap->mds_wanted = want; - arg.ino = ceph_vino(inode).ino; - arg.cid = cap->cap_id; - arg.follows = flushing ? ci->i_head_snapc->seq : 0; - arg.flush_tid = flush_tid; - arg.oldest_flush_tid = oldest_flush_tid; - - arg.size = inode->i_size; - ci->i_reported_size = arg.size; - arg.max_size = ci->i_wanted_max_size; - ci->i_requested_max_size = arg.max_size; + arg->session = cap->session; + arg->ino = ceph_vino(inode).ino; + arg->cid = cap->cap_id; + arg->follows = flushing ? ci->i_head_snapc->seq : 0; + arg->flush_tid = flush_tid; + arg->oldest_flush_tid = oldest_flush_tid; + + arg->size = i_size_read(inode); + ci->i_reported_size = arg->size; + arg->max_size = ci->i_wanted_max_size; + if (cap == ci->i_auth_cap) { + if (want & CEPH_CAP_ANY_FILE_WR) + ci->i_requested_max_size = arg->max_size; + else + ci->i_requested_max_size = 0; + } if (flushing & CEPH_CAP_XATTR_EXCL) { - old_blob = __ceph_build_xattrs_blob(ci); - arg.xattr_version = ci->i_xattrs.version; - arg.xattr_buf = ci->i_xattrs.blob; + arg->old_xattr_buf = __ceph_build_xattrs_blob(ci); + arg->xattr_version = ci->i_xattrs.version; + arg->xattr_buf = ci->i_xattrs.blob; } else { - arg.xattr_buf = NULL; + arg->xattr_buf = NULL; + arg->old_xattr_buf = NULL; } - arg.mtime = inode->i_mtime; - arg.atime = inode->i_atime; - arg.ctime = inode->i_ctime; - arg.btime = ci->i_btime; - arg.change_attr = inode_peek_iversion_raw(inode); + arg->mtime = inode->i_mtime; + arg->atime = inode->i_atime; + arg->ctime = inode->i_ctime; + arg->btime = ci->i_btime; + arg->change_attr = inode_peek_iversion_raw(inode); - arg.op = op; - arg.caps = cap->implemented; - arg.wanted = want; - arg.dirty = flushing; + arg->op = op; + arg->caps = cap->implemented; + arg->wanted = want; + arg->dirty = flushing; - arg.seq = cap->seq; - arg.issue_seq = cap->issue_seq; - arg.mseq = cap->mseq; - arg.time_warp_seq = ci->i_time_warp_seq; + arg->seq = cap->seq; + arg->issue_seq = cap->issue_seq; + arg->mseq = cap->mseq; + arg->time_warp_seq = ci->i_time_warp_seq; - arg.uid = inode->i_uid; - arg.gid = inode->i_gid; - arg.mode = inode->i_mode; + arg->uid = inode->i_uid; + arg->gid = inode->i_gid; + arg->mode = inode->i_mode; - arg.inline_data = ci->i_inline_version != CEPH_INLINE_NONE; + arg->inline_data = ci->i_inline_version != CEPH_INLINE_NONE; if (!(flags & CEPH_CLIENT_CAPS_PENDING_CAPSNAP) && !list_empty(&ci->i_cap_snaps)) { struct ceph_cap_snap *capsnap; @@ -1375,22 +1430,35 @@ static int __send_cap(struct ceph_mds_client *mdsc, struct ceph_cap *cap, } } } - arg.flags = flags; - - spin_unlock(&ci->i_ceph_lock); + arg->flags = flags; +} - ceph_buffer_put(old_blob); +/* + * Send a cap msg on the given inode. + * + * Caller should hold snap_rwsem (read), s_mutex. + */ +static void __send_cap(struct cap_msg_args *arg, struct ceph_inode_info *ci) +{ + struct ceph_msg *msg; + struct inode *inode = &ci->netfs.inode; - ret = send_cap_msg(&arg); - if (ret < 0) { - dout("error sending cap msg, must requeue %p\n", inode); - delayed = 1; + msg = ceph_msg_new(CEPH_MSG_CLIENT_CAPS, CAP_MSG_SIZE, GFP_NOFS, false); + if (!msg) { + pr_err("error allocating cap msg: ino (%llx.%llx) flushing %s tid %llu, requeuing cap.\n", + ceph_vinop(inode), ceph_cap_string(arg->dirty), + arg->flush_tid); + spin_lock(&ci->i_ceph_lock); + __cap_delay_requeue(arg->session->s_mdsc, ci); + spin_unlock(&ci->i_ceph_lock); + return; } - if (wake) + encode_cap_msg(msg, arg); + ceph_con_send(&arg->session->s_con, msg); + ceph_buffer_put(arg->old_xattr_buf); + if (arg->wake) wake_up_all(&ci->i_cap_wq); - - return delayed; } static inline int __send_flush_snap(struct inode *inode, @@ -1399,6 +1467,11 @@ static inline int __send_flush_snap(struct inode *inode, u32 mseq, u64 oldest_flush_tid) { struct cap_msg_args arg; + struct ceph_msg *msg; + + msg = ceph_msg_new(CEPH_MSG_CLIENT_CAPS, CAP_MSG_SIZE, GFP_NOFS, false); + if (!msg) + return -ENOMEM; arg.session = session; arg.ino = ceph_vino(inode).ino; @@ -1411,6 +1484,7 @@ static inline int __send_flush_snap(struct inode *inode, arg.max_size = 0; arg.xattr_version = capsnap->xattr_version; arg.xattr_buf = capsnap->xattr_blob; + arg.old_xattr_buf = NULL; arg.atime = capsnap->atime; arg.mtime = capsnap->mtime; @@ -1434,8 +1508,11 @@ static inline int __send_flush_snap(struct inode *inode, arg.inline_data = capsnap->inline_data; arg.flags = 0; + arg.wake = false; - return send_cap_msg(&arg); + encode_cap_msg(msg, &arg); + ceph_con_send(&arg.session->s_con, msg); + return 0; } /* @@ -1445,14 +1522,14 @@ static inline int __send_flush_snap(struct inode *inode, * asynchronously back to the MDS once sync writes complete and dirty * data is written out. * - * Called under i_ceph_lock. Takes s_mutex as needed. + * Called under i_ceph_lock. */ static void __ceph_flush_snaps(struct ceph_inode_info *ci, struct ceph_mds_session *session) __releases(ci->i_ceph_lock) __acquires(ci->i_ceph_lock) { - struct inode *inode = &ci->vfs_inode; + struct inode *inode = &ci->netfs.inode; struct ceph_mds_client *mdsc = session->s_mdsc; struct ceph_cap_snap *capsnap; u64 oldest_flush_tid = 0; @@ -1501,7 +1578,7 @@ static void __ceph_flush_snaps(struct ceph_inode_info *ci, while (first_tid <= last_tid) { struct ceph_cap *cap = ci->i_auth_cap; - struct ceph_cap_flush *cf; + struct ceph_cap_flush *cf = NULL, *iter; int ret; if (!(cap && cap->session == session)) { @@ -1511,8 +1588,9 @@ static void __ceph_flush_snaps(struct ceph_inode_info *ci, } ret = -ENOENT; - list_for_each_entry(cf, &ci->i_cap_flush_list, i_list) { - if (cf->tid >= first_tid) { + list_for_each_entry(iter, &ci->i_cap_flush_list, i_list) { + if (iter->tid >= first_tid) { + cf = iter; ret = 0; break; } @@ -1545,7 +1623,7 @@ static void __ceph_flush_snaps(struct ceph_inode_info *ci, void ceph_flush_snaps(struct ceph_inode_info *ci, struct ceph_mds_session **psession) { - struct inode *inode = &ci->vfs_inode; + struct inode *inode = &ci->netfs.inode; struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc; struct ceph_mds_session *session = NULL; int mds; @@ -1567,7 +1645,6 @@ retry: mds = ci->i_auth_cap->session->s_mds; if (session && session->s_mds != mds) { dout(" oops, wrong session %p mutex\n", session); - mutex_unlock(&session->s_mutex); ceph_put_mds_session(session); session = NULL; } @@ -1576,10 +1653,6 @@ retry: mutex_lock(&mdsc->mutex); session = __ceph_lookup_mds_session(mdsc, mds); mutex_unlock(&mdsc->mutex); - if (session) { - dout(" inverting session/ino locks on %p\n", session); - mutex_lock(&session->s_mutex); - } goto retry; } @@ -1591,12 +1664,10 @@ retry: out: spin_unlock(&ci->i_ceph_lock); - if (psession) { + if (psession) *psession = session; - } else if (session) { - mutex_unlock(&session->s_mutex); + else ceph_put_mds_session(session); - } /* we flushed them all; remove this inode from the queue */ spin_lock(&mdsc->snap_flush_lock); list_del_init(&ci->i_snap_flush_item); @@ -1612,11 +1683,13 @@ int __ceph_mark_dirty_caps(struct ceph_inode_info *ci, int mask, struct ceph_cap_flush **pcf) { struct ceph_mds_client *mdsc = - ceph_sb_to_client(ci->vfs_inode.i_sb)->mdsc; - struct inode *inode = &ci->vfs_inode; + ceph_sb_to_client(ci->netfs.inode.i_sb)->mdsc; + struct inode *inode = &ci->netfs.inode; int was = ci->i_dirty_caps; int dirty = 0; + lockdep_assert_held(&ci->i_ceph_lock); + if (!ci->i_auth_cap) { pr_warn("__mark_dirty_caps %p %llx mask %s, " "but no auth cap (session was closed?)\n", @@ -1624,11 +1697,13 @@ int __ceph_mark_dirty_caps(struct ceph_inode_info *ci, int mask, return 0; } - dout("__mark_dirty_caps %p %s dirty %s -> %s\n", &ci->vfs_inode, + dout("__mark_dirty_caps %p %s dirty %s -> %s\n", &ci->netfs.inode, ceph_cap_string(mask), ceph_cap_string(was), ceph_cap_string(was | mask)); ci->i_dirty_caps |= mask; if (was == 0) { + struct ceph_mds_session *session = ci->i_auth_cap->session; + WARN_ON_ONCE(ci->i_prealloc_cap_flush); swap(ci->i_prealloc_cap_flush, *pcf); @@ -1638,10 +1713,10 @@ int __ceph_mark_dirty_caps(struct ceph_inode_info *ci, int mask, ci->i_snap_realm->cached_context); } dout(" inode %p now dirty snapc %p auth cap %p\n", - &ci->vfs_inode, ci->i_head_snapc, ci->i_auth_cap); + &ci->netfs.inode, ci->i_head_snapc, ci->i_auth_cap); BUG_ON(!list_empty(&ci->i_dirty_item)); spin_lock(&mdsc->cap_dirty_lock); - list_add(&ci->i_dirty_item, &mdsc->cap_dirty); + list_add(&ci->i_dirty_item, &session->s_cap_dirty); spin_unlock(&mdsc->cap_dirty_lock); if (ci->i_flushing_caps == 0) { ihold(inode); @@ -1654,13 +1729,20 @@ int __ceph_mark_dirty_caps(struct ceph_inode_info *ci, int mask, if (((was | ci->i_flushing_caps) & CEPH_CAP_FILE_BUFFER) && (mask & CEPH_CAP_FILE_BUFFER)) dirty |= I_DIRTY_DATASYNC; - __cap_delay_requeue(mdsc, ci, true); + __cap_delay_requeue(mdsc, ci); return dirty; } struct ceph_cap_flush *ceph_alloc_cap_flush(void) { - return kmem_cache_alloc(ceph_cap_flush_cachep, GFP_KERNEL); + struct ceph_cap_flush *cf; + + cf = kmem_cache_alloc(ceph_cap_flush_cachep, GFP_KERNEL); + if (!cf) + return NULL; + + cf->is_capsnap = false; + return cf; } void ceph_free_cap_flush(struct ceph_cap_flush *cf) @@ -1684,30 +1766,33 @@ static u64 __get_oldest_flush_tid(struct ceph_mds_client *mdsc) * Remove cap_flush from the mdsc's or inode's flushing cap list. * Return true if caller needs to wake up flush waiters. */ -static bool __finish_cap_flush(struct ceph_mds_client *mdsc, - struct ceph_inode_info *ci, - struct ceph_cap_flush *cf) +static bool __detach_cap_flush_from_mdsc(struct ceph_mds_client *mdsc, + struct ceph_cap_flush *cf) { struct ceph_cap_flush *prev; bool wake = cf->wake; - if (mdsc) { - /* are there older pending cap flushes? */ - if (wake && cf->g_list.prev != &mdsc->cap_flush_list) { - prev = list_prev_entry(cf, g_list); - prev->wake = true; - wake = false; - } - list_del(&cf->g_list); - } else if (ci) { - if (wake && cf->i_list.prev != &ci->i_cap_flush_list) { - prev = list_prev_entry(cf, i_list); - prev->wake = true; - wake = false; - } - list_del(&cf->i_list); - } else { - BUG_ON(1); + + if (wake && cf->g_list.prev != &mdsc->cap_flush_list) { + prev = list_prev_entry(cf, g_list); + prev->wake = true; + wake = false; } + list_del_init(&cf->g_list); + return wake; +} + +static bool __detach_cap_flush_from_ci(struct ceph_inode_info *ci, + struct ceph_cap_flush *cf) +{ + struct ceph_cap_flush *prev; + bool wake = cf->wake; + + if (wake && cf->i_list.prev != &ci->i_cap_flush_list) { + prev = list_prev_entry(cf, i_list); + prev->wake = true; + wake = false; + } + list_del_init(&cf->i_list); return wake; } @@ -1726,6 +1811,7 @@ static u64 __mark_caps_flushing(struct inode *inode, struct ceph_cap_flush *cf = NULL; int flushing; + lockdep_assert_held(&ci->i_ceph_lock); BUG_ON(ci->i_dirty_caps == 0); BUG_ON(list_empty(&ci->i_dirty_item)); BUG_ON(!ci->i_prealloc_cap_flush); @@ -1765,11 +1851,14 @@ static u64 __mark_caps_flushing(struct inode *inode, * try to invalidate mapping pages without blocking. */ static int try_nonblocking_invalidate(struct inode *inode) + __releases(ci->i_ceph_lock) + __acquires(ci->i_ceph_lock) { struct ceph_inode_info *ci = ceph_inode(inode); u32 invalidating_gen = ci->i_rdcache_gen; spin_unlock(&ci->i_ceph_lock); + ceph_fscache_invalidate(inode, false); invalidate_mapping_pages(&inode->i_data, 0, -1); spin_lock(&ci->i_ceph_lock); @@ -1787,7 +1876,7 @@ static int try_nonblocking_invalidate(struct inode *inode) bool __ceph_should_report_size(struct ceph_inode_info *ci) { - loff_t size = ci->vfs_inode.i_size; + loff_t size = i_size_read(&ci->netfs.inode); /* mds will adjust max size according to the reported size */ if (ci->i_flushing_caps & CEPH_CAP_FILE_WR) return false; @@ -1805,8 +1894,6 @@ bool __ceph_should_report_size(struct ceph_inode_info *ci) * versus held caps. Release, flush, ack revoked caps to mds as * appropriate. * - * CHECK_CAPS_NODELAY - caller is delayed work and we should not delay - * cap release further. * CHECK_CAPS_AUTHONLY - we should only check the auth cap * CHECK_CAPS_FLUSH - we should flush any dirty caps immediately, without * further delay. @@ -1814,45 +1901,51 @@ bool __ceph_should_report_size(struct ceph_inode_info *ci) void ceph_check_caps(struct ceph_inode_info *ci, int flags, struct ceph_mds_session *session) { - struct ceph_fs_client *fsc = ceph_inode_to_client(&ci->vfs_inode); - struct ceph_mds_client *mdsc = fsc->mdsc; - struct inode *inode = &ci->vfs_inode; + struct inode *inode = &ci->netfs.inode; + struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(inode->i_sb); struct ceph_cap *cap; u64 flush_tid, oldest_flush_tid; int file_wanted, used, cap_used; - int took_snap_rwsem = 0; /* true if mdsc->snap_rwsem held */ int issued, implemented, want, retain, revoking, flushing = 0; int mds = -1; /* keep track of how far we've gone through i_caps list to avoid an infinite loop on retry */ struct rb_node *p; - int delayed = 0, sent = 0; - bool no_delay = flags & CHECK_CAPS_NODELAY; bool queue_invalidate = false; bool tried_invalidate = false; + bool queue_writeback = false; - /* if we are unmounting, flush any unused caps immediately. */ - if (mdsc->stopping) - no_delay = true; + if (session) + ceph_get_mds_session(session); spin_lock(&ci->i_ceph_lock); + if (ci->i_ceph_flags & CEPH_I_ASYNC_CREATE) { + /* Don't send messages until we get async create reply */ + spin_unlock(&ci->i_ceph_lock); + ceph_put_mds_session(session); + return; + } if (ci->i_ceph_flags & CEPH_I_FLUSH) flags |= CHECK_CAPS_FLUSH; - - if (!(flags & CHECK_CAPS_AUTHONLY) || - (ci->i_auth_cap && __ceph_is_single_caps(ci))) - __cap_delay_cancel(mdsc, ci); - - goto retry_locked; retry: - spin_lock(&ci->i_ceph_lock); -retry_locked: + /* Caps wanted by virtue of active open files. */ file_wanted = __ceph_caps_file_wanted(ci); + + /* Caps which have active references against them */ used = __ceph_caps_used(ci); + + /* + * "issued" represents the current caps that the MDS wants us to have. + * "implemented" is the set that we have been granted, and includes the + * ones that have not yet been returned to the MDS (the "revoking" set, + * usually because they have outstanding references). + */ issued = __ceph_caps_issued(ci, &implemented); revoking = implemented & ~issued; want = file_wanted; + + /* The ones we currently want to retain (may be adjusted below) */ retain = file_wanted | used | CEPH_CAP_PIN; if (!mdsc->stopping && inode->i_nlink > 0) { if (file_wanted) { @@ -1866,10 +1959,11 @@ retry_locked: * revoking the shared cap on every create/unlink * operation. */ - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode)) { want = CEPH_CAP_ANY_SHARED; - else - want = CEPH_CAP_ANY_SHARED | CEPH_CAP_FILE_EXCL; + } else { + want |= CEPH_CAP_ANY_SHARED | CEPH_CAP_FILE_EXCL; + } retain |= want; } else { @@ -1884,40 +1978,44 @@ retry_locked: } } - dout("check_caps %p file_want %s used %s dirty %s flushing %s" - " issued %s revoking %s retain %s %s%s%s\n", inode, + dout("check_caps %llx.%llx file_want %s used %s dirty %s flushing %s" + " issued %s revoking %s retain %s %s%s%s\n", ceph_vinop(inode), ceph_cap_string(file_wanted), ceph_cap_string(used), ceph_cap_string(ci->i_dirty_caps), ceph_cap_string(ci->i_flushing_caps), ceph_cap_string(issued), ceph_cap_string(revoking), ceph_cap_string(retain), (flags & CHECK_CAPS_AUTHONLY) ? " AUTHONLY" : "", - (flags & CHECK_CAPS_NODELAY) ? " NODELAY" : "", - (flags & CHECK_CAPS_FLUSH) ? " FLUSH" : ""); + (flags & CHECK_CAPS_FLUSH) ? " FLUSH" : "", + (flags & CHECK_CAPS_NOINVAL) ? " NOINVAL" : ""); /* * If we no longer need to hold onto old our caps, and we may * have cached pages, but don't want them, then try to invalidate. * If we fail, it's because pages are locked.... try again later. */ - if ((!no_delay || mdsc->stopping) && - !S_ISDIR(inode->i_mode) && /* ignore readdir cache */ + if ((!(flags & CHECK_CAPS_NOINVAL) || mdsc->stopping) && + S_ISREG(inode->i_mode) && !(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 */ !tried_invalidate) { - dout("check_caps trying to invalidate on %p\n", inode); + dout("check_caps trying to invalidate on %llx.%llx\n", + ceph_vinop(inode)); if (try_nonblocking_invalidate(inode) < 0) { dout("check_caps queuing invalidate\n"); queue_invalidate = true; ci->i_rdcache_revoking = ci->i_rdcache_gen; } tried_invalidate = true; - goto retry_locked; + goto retry; } for (p = rb_first(&ci->i_caps); p; p = rb_next(p)) { + int mflags = 0; + struct cap_msg_args arg; + cap = rb_entry(p, struct ceph_cap, ci_node); /* avoid looping forever */ @@ -1925,8 +2023,10 @@ retry_locked: ((flags & CHECK_CAPS_AUTHONLY) && cap != ci->i_auth_cap)) continue; - /* NOTE: no side-effects allowed, until we take s_mutex */ - + /* + * If we have an auth cap, we don't need to consider any + * overlapping caps as used. + */ cap_used = used; if (ci->i_auth_cap && cap != ci->i_auth_cap) cap_used &= ~ci->i_auth_cap->issued; @@ -1966,55 +2066,44 @@ retry_locked: } /* completed revocation? going down and there are no caps? */ - if (revoking && (revoking & cap_used) == 0) { - dout("completed revocation of %s\n", - ceph_cap_string(cap->implemented & ~cap->issued)); - goto ack; + if (revoking) { + if ((revoking & cap_used) == 0) { + dout("completed revocation of %s\n", + ceph_cap_string(cap->implemented & ~cap->issued)); + goto ack; + } + + /* + * If the "i_wrbuffer_ref" was increased by mmap or generic + * cache write just before the ceph_check_caps() is called, + * the Fb capability revoking will fail this time. Then we + * must wait for the BDI's delayed work to flush the dirty + * pages and to release the "i_wrbuffer_ref", which will cost + * at most 5 seconds. That means the MDS needs to wait at + * most 5 seconds to finished the Fb capability's revocation. + * + * Let's queue a writeback for it. + */ + if (S_ISREG(inode->i_mode) && ci->i_wrbuffer_ref && + (revoking & CEPH_CAP_FILE_BUFFER)) + queue_writeback = true; } /* want more caps from mds? */ - if (want & ~(cap->mds_wanted | cap->issued)) - goto ack; + if (want & ~cap->mds_wanted) { + if (want & ~(cap->mds_wanted | cap->issued)) + goto ack; + if (!__cap_is_valid(cap)) + goto ack; + } /* things we might delay */ if ((cap->issued & ~retain) == 0) continue; /* nope, all good */ - if (no_delay) - goto ack; - - /* delay? */ - if ((ci->i_ceph_flags & CEPH_I_NODELAY) == 0 && - time_before(jiffies, ci->i_hold_caps_max)) { - dout(" delaying issued %s -> %s, wanted %s -> %s\n", - ceph_cap_string(cap->issued), - ceph_cap_string(cap->issued & retain), - ceph_cap_string(cap->mds_wanted), - ceph_cap_string(want)); - delayed++; - continue; - } - ack: - if (session && session != cap->session) { - dout("oops, wrong session %p mutex\n", session); - mutex_unlock(&session->s_mutex); - session = NULL; - } - if (!session) { - session = cap->session; - if (mutex_trylock(&session->s_mutex) == 0) { - dout("inverting session/ino locks on %p\n", - session); - spin_unlock(&ci->i_ceph_lock); - if (took_snap_rwsem) { - up_read(&mdsc->snap_rwsem); - took_snap_rwsem = 0; - } - mutex_lock(&session->s_mutex); - goto retry; - } - } + ceph_put_mds_session(session); + session = ceph_get_mds_session(cap->session); /* kick flushing and flush snaps before sending normal * cap message */ @@ -2026,26 +2115,16 @@ ack: if (ci->i_ceph_flags & CEPH_I_FLUSH_SNAPS) __ceph_flush_snaps(ci, session); - goto retry_locked; - } - - /* take snap_rwsem after session mutex */ - if (!took_snap_rwsem) { - if (down_read_trylock(&mdsc->snap_rwsem) == 0) { - dout("inverting snap/in locks on %p\n", - inode); - spin_unlock(&ci->i_ceph_lock); - down_read(&mdsc->snap_rwsem); - took_snap_rwsem = 1; - goto retry; - } - took_snap_rwsem = 1; + goto retry; } if (cap == ci->i_auth_cap && ci->i_dirty_caps) { flushing = ci->i_dirty_caps; flush_tid = __mark_caps_flushing(inode, session, false, &oldest_flush_tid); + if (flags & CHECK_CAPS_FLUSH && + list_empty(&session->s_cap_dirty)) + mflags |= CEPH_CLIENT_CAPS_SYNC; } else { flushing = 0; flush_tid = 0; @@ -2055,28 +2134,32 @@ ack: } mds = cap->mds; /* remember mds, so we don't repeat */ - sent++; - /* __send_cap drops i_ceph_lock */ - delayed += __send_cap(mdsc, cap, CEPH_CAP_OP_UPDATE, 0, - cap_used, want, retain, flushing, - flush_tid, oldest_flush_tid); + __prep_cap(&arg, cap, CEPH_CAP_OP_UPDATE, mflags, cap_used, + want, retain, flushing, flush_tid, oldest_flush_tid); + + spin_unlock(&ci->i_ceph_lock); + __send_cap(&arg, ci); + spin_lock(&ci->i_ceph_lock); + goto retry; /* retake i_ceph_lock and restart our cap scan. */ } - /* Reschedule delayed caps release if we delayed anything */ - if (delayed) - __cap_delay_requeue(mdsc, ci, false); + /* periodically re-calculate caps wanted by open files */ + if (__ceph_is_any_real_caps(ci) && + list_empty(&ci->i_cap_delay_list) && + (file_wanted & ~CEPH_CAP_PIN) && + !(used & (CEPH_CAP_FILE_RD | CEPH_CAP_ANY_FILE_WR))) { + __cap_delay_requeue(mdsc, ci); + } spin_unlock(&ci->i_ceph_lock); + ceph_put_mds_session(session); + if (queue_writeback) + ceph_queue_writeback(inode); if (queue_invalidate) ceph_queue_invalidate(inode); - - if (session) - mutex_unlock(&session->s_mutex); - if (took_snap_rwsem) - up_read(&mdsc->snap_rwsem); } /* @@ -2086,26 +2169,17 @@ static int try_flush_caps(struct inode *inode, u64 *ptid) { struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; struct ceph_inode_info *ci = ceph_inode(inode); - struct ceph_mds_session *session = NULL; int flushing = 0; u64 flush_tid = 0, oldest_flush_tid = 0; -retry: spin_lock(&ci->i_ceph_lock); retry_locked: if (ci->i_dirty_caps && ci->i_auth_cap) { struct ceph_cap *cap = ci->i_auth_cap; - int delayed; + struct cap_msg_args arg; + struct ceph_mds_session *session = cap->session; - if (session != cap->session) { - spin_unlock(&ci->i_ceph_lock); - if (session) - mutex_unlock(&session->s_mutex); - session = cap->session; - mutex_lock(&session->s_mutex); - goto retry; - } - if (cap->session->s_state < CEPH_MDS_SESSION_OPEN) { + if (session->s_state < CEPH_MDS_SESSION_OPEN) { spin_unlock(&ci->i_ceph_lock); goto out; } @@ -2123,19 +2197,13 @@ retry_locked: flush_tid = __mark_caps_flushing(inode, session, true, &oldest_flush_tid); - /* __send_cap drops i_ceph_lock */ - delayed = __send_cap(mdsc, cap, CEPH_CAP_OP_FLUSH, - CEPH_CLIENT_CAPS_SYNC, - __ceph_caps_used(ci), - __ceph_caps_wanted(ci), - (cap->issued | cap->implemented), - flushing, flush_tid, oldest_flush_tid); + __prep_cap(&arg, cap, CEPH_CAP_OP_FLUSH, CEPH_CLIENT_CAPS_SYNC, + __ceph_caps_used(ci), __ceph_caps_wanted(ci), + (cap->issued | cap->implemented), + flushing, flush_tid, oldest_flush_tid); + spin_unlock(&ci->i_ceph_lock); - if (delayed) { - spin_lock(&ci->i_ceph_lock); - __cap_delay_requeue(mdsc, ci, true); - spin_unlock(&ci->i_ceph_lock); - } + __send_cap(&arg, ci); } else { if (!list_empty(&ci->i_cap_flush_list)) { struct ceph_cap_flush *cf = @@ -2148,9 +2216,6 @@ retry_locked: spin_unlock(&ci->i_ceph_lock); } out: - if (session) - mutex_unlock(&session->s_mutex); - *ptid = flush_tid; return flushing; } @@ -2176,12 +2241,14 @@ static int caps_are_flushed(struct inode *inode, u64 flush_tid) } /* - * wait for any unsafe requests to complete. + * flush the mdlog and wait for any unsafe requests to complete. */ -static int unsafe_request_wait(struct inode *inode) +static int flush_mdlog_and_wait_inode_unsafe_requests(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 *req1 = NULL, *req2 = NULL; + unsigned int max_sessions; int ret, err = 0; spin_lock(&ci->i_unsafe_lock); @@ -2199,28 +2266,124 @@ static int unsafe_request_wait(struct inode *inode) } spin_unlock(&ci->i_unsafe_lock); - dout("unsafe_request_wait %p wait on tid %llu %llu\n", + /* + * The mdsc->max_sessions is unlikely to be changed + * mostly, here we will retry it by reallocating the + * sessions array memory to get rid of the mdsc->mutex + * lock. + */ +retry: + max_sessions = mdsc->max_sessions; + + /* + * Trigger to flush the journal logs in all the relevant MDSes + * manually, or in the worst case we must wait at most 5 seconds + * to wait the journal logs to be flushed by the MDSes periodically. + */ + if ((req1 || req2) && likely(max_sessions)) { + struct ceph_mds_session **sessions = NULL; + struct ceph_mds_session *s; + struct ceph_mds_request *req; + int i; + + sessions = kcalloc(max_sessions, sizeof(s), GFP_KERNEL); + if (!sessions) { + err = -ENOMEM; + goto out; + } + + spin_lock(&ci->i_unsafe_lock); + if (req1) { + list_for_each_entry(req, &ci->i_unsafe_dirops, + r_unsafe_dir_item) { + s = req->r_session; + if (!s) + continue; + if (unlikely(s->s_mds >= max_sessions)) { + spin_unlock(&ci->i_unsafe_lock); + for (i = 0; i < max_sessions; i++) { + s = sessions[i]; + if (s) + ceph_put_mds_session(s); + } + kfree(sessions); + goto retry; + } + if (!sessions[s->s_mds]) { + s = ceph_get_mds_session(s); + sessions[s->s_mds] = s; + } + } + } + if (req2) { + list_for_each_entry(req, &ci->i_unsafe_iops, + r_unsafe_target_item) { + s = req->r_session; + if (!s) + continue; + if (unlikely(s->s_mds >= max_sessions)) { + spin_unlock(&ci->i_unsafe_lock); + for (i = 0; i < max_sessions; i++) { + s = sessions[i]; + if (s) + ceph_put_mds_session(s); + } + kfree(sessions); + goto retry; + } + if (!sessions[s->s_mds]) { + s = ceph_get_mds_session(s); + sessions[s->s_mds] = s; + } + } + } + spin_unlock(&ci->i_unsafe_lock); + + /* the auth MDS */ + spin_lock(&ci->i_ceph_lock); + if (ci->i_auth_cap) { + s = ci->i_auth_cap->session; + if (!sessions[s->s_mds]) + sessions[s->s_mds] = ceph_get_mds_session(s); + } + spin_unlock(&ci->i_ceph_lock); + + /* send flush mdlog request to MDSes */ + for (i = 0; i < max_sessions; i++) { + s = sessions[i]; + if (s) { + send_flush_mdlog(s); + ceph_put_mds_session(s); + } + } + kfree(sessions); + } + + dout("%s %p wait on tid %llu %llu\n", __func__, inode, req1 ? req1->r_tid : 0ULL, req2 ? req2->r_tid : 0ULL); if (req1) { ret = !wait_for_completion_timeout(&req1->r_safe_completion, ceph_timeout_jiffies(req1->r_timeout)); if (ret) err = -EIO; - ceph_mdsc_put_request(req1); } if (req2) { ret = !wait_for_completion_timeout(&req2->r_safe_completion, ceph_timeout_jiffies(req2->r_timeout)); if (ret) err = -EIO; - ceph_mdsc_put_request(req2); } + +out: + if (req1) + ceph_mdsc_put_request(req1); + if (req2) + ceph_mdsc_put_request(req2); return err; } int ceph_fsync(struct file *file, loff_t start, loff_t end, int datasync) { - struct ceph_file_info *fi = file->private_data; struct inode *inode = file->f_mapping->host; struct ceph_inode_info *ci = ceph_inode(inode); u64 flush_tid; @@ -2233,10 +2396,14 @@ int ceph_fsync(struct file *file, loff_t start, loff_t end, int datasync) if (datasync) goto out; + ret = ceph_wait_on_async_create(inode); + if (ret) + goto out; + dirty = try_flush_caps(inode, &flush_tid); dout("fsync dirty caps are %s\n", ceph_cap_string(dirty)); - err = unsafe_request_wait(inode); + err = flush_mdlog_and_wait_inode_unsafe_requests(inode); /* * only wait on non-file metadata writeback (the mds @@ -2251,14 +2418,9 @@ int ceph_fsync(struct file *file, loff_t start, loff_t end, int datasync) if (err < 0) ret = err; - if (errseq_check(&ci->i_meta_err, READ_ONCE(fi->meta_err))) { - spin_lock(&file->f_lock); - err = errseq_check_and_advance(&ci->i_meta_err, - &fi->meta_err); - spin_unlock(&file->f_lock); - if (err < 0) - ret = err; - } + err = file_check_and_advance_wb_err(file); + if (err < 0) + ret = err; out: dout("fsync %p%s result=%d\n", inode, datasync ? " datasync" : "", ret); return ret; @@ -2279,7 +2441,11 @@ int ceph_write_inode(struct inode *inode, struct writeback_control *wbc) int wait = (wbc->sync_mode == WB_SYNC_ALL && !wbc->for_sync); dout("write_inode %p wait=%d\n", inode, wait); + ceph_fscache_unpin_writeback(inode, wbc); if (wait) { + err = ceph_wait_on_async_create(inode); + if (err) + return err; dirty = try_flush_caps(inode, &flush_tid); if (dirty) err = wait_event_interruptible(ci->i_cap_wq, @@ -2303,17 +2469,21 @@ static void __kick_flushing_caps(struct ceph_mds_client *mdsc, __releases(ci->i_ceph_lock) __acquires(ci->i_ceph_lock) { - struct inode *inode = &ci->vfs_inode; + struct inode *inode = &ci->netfs.inode; struct ceph_cap *cap; struct ceph_cap_flush *cf; int ret; u64 first_tid = 0; u64 last_snap_flush = 0; + /* Don't do anything until create reply comes in */ + if (ci->i_ceph_flags & CEPH_I_ASYNC_CREATE) + return; + ci->i_ceph_flags &= ~CEPH_I_KICK_FLUSH; list_for_each_entry_reverse(cf, &ci->i_cap_flush_list, i_list) { - if (!cf->caps) { + if (cf->is_capsnap) { last_snap_flush = cf->tid; break; } @@ -2332,25 +2502,20 @@ static void __kick_flushing_caps(struct ceph_mds_client *mdsc, first_tid = cf->tid + 1; - if (cf->caps) { + if (!cf->is_capsnap) { + struct cap_msg_args arg; + dout("kick_flushing_caps %p cap %p tid %llu %s\n", inode, cap, cf->tid, ceph_cap_string(cf->caps)); - ci->i_ceph_flags |= CEPH_I_NODELAY; - - ret = __send_cap(mdsc, cap, CEPH_CAP_OP_FLUSH, + __prep_cap(&arg, cap, CEPH_CAP_OP_FLUSH, (cf->tid < last_snap_flush ? CEPH_CLIENT_CAPS_PENDING_CAPSNAP : 0), __ceph_caps_used(ci), __ceph_caps_wanted(ci), (cap->issued | cap->implemented), cf->caps, cf->tid, oldest_flush_tid); - if (ret) { - pr_err("kick_flushing_caps: error sending " - "cap flush, ino (%llx.%llx) " - "tid %llu flushing %s\n", - ceph_vinop(inode), cf->tid, - ceph_cap_string(cf->caps)); - } + spin_unlock(&ci->i_ceph_lock); + __send_cap(&arg, ci); } else { struct ceph_cap_snap *capsnap = container_of(cf, struct ceph_cap_snap, @@ -2397,7 +2562,7 @@ void ceph_early_kick_flushing_caps(struct ceph_mds_client *mdsc, cap = ci->i_auth_cap; if (!(cap && cap->session == session)) { pr_err("%p auth cap %p not mds%d ???\n", - &ci->vfs_inode, cap, session->s_mds); + &ci->netfs.inode, cap, session->s_mds); spin_unlock(&ci->i_ceph_lock); continue; } @@ -2434,6 +2599,8 @@ void ceph_kick_flushing_caps(struct ceph_mds_client *mdsc, struct ceph_cap *cap; u64 oldest_flush_tid; + lockdep_assert_held(&session->s_mutex); + dout("kick_flushing_caps mds%d\n", session->s_mds); spin_lock(&mdsc->cap_dirty_lock); @@ -2445,7 +2612,7 @@ void ceph_kick_flushing_caps(struct ceph_mds_client *mdsc, cap = ci->i_auth_cap; if (!(cap && cap->session == session)) { pr_err("%p auth cap %p not mds%d ???\n", - &ci->vfs_inode, cap, session->s_mds); + &ci->netfs.inode, cap, session->s_mds); spin_unlock(&ci->i_ceph_lock); continue; } @@ -2457,16 +2624,15 @@ void ceph_kick_flushing_caps(struct ceph_mds_client *mdsc, } } -static void kick_flushing_inode_caps(struct ceph_mds_client *mdsc, - struct ceph_mds_session *session, - struct inode *inode) - __releases(ci->i_ceph_lock) +void ceph_kick_flushing_inode_caps(struct ceph_mds_session *session, + struct ceph_inode_info *ci) { - struct ceph_inode_info *ci = ceph_inode(inode); - struct ceph_cap *cap; + struct ceph_mds_client *mdsc = session->s_mdsc; + struct ceph_cap *cap = ci->i_auth_cap; + + lockdep_assert_held(&ci->i_ceph_lock); - cap = ci->i_auth_cap; - dout("kick_flushing_inode_caps %p flushing %s\n", inode, + dout("%s %p flushing %s\n", __func__, &ci->netfs.inode, ceph_cap_string(ci->i_flushing_caps)); if (!list_empty(&ci->i_cap_flush_list)) { @@ -2478,9 +2644,6 @@ static void kick_flushing_inode_caps(struct ceph_mds_client *mdsc, spin_unlock(&mdsc->cap_dirty_lock); __kick_flushing_caps(mdsc, session, ci, oldest_flush_tid); - spin_unlock(&ci->i_ceph_lock); - } else { - spin_unlock(&ci->i_ceph_lock); } } @@ -2488,18 +2651,20 @@ static void kick_flushing_inode_caps(struct ceph_mds_client *mdsc, /* * Take references to capabilities we hold, so that we don't release * them to the MDS prematurely. - * - * Protected by i_ceph_lock. */ -static void __take_cap_refs(struct ceph_inode_info *ci, int got, +void ceph_take_cap_refs(struct ceph_inode_info *ci, int got, bool snap_rwsem_locked) { + lockdep_assert_held(&ci->i_ceph_lock); + if (got & CEPH_CAP_PIN) ci->i_pin_ref++; if (got & CEPH_CAP_FILE_RD) ci->i_rd_ref++; if (got & CEPH_CAP_FILE_CACHE) ci->i_rdcache_ref++; + if (got & CEPH_CAP_FILE_EXCL) + ci->i_fx_ref++; if (got & CEPH_CAP_FILE_WR) { if (ci->i_wr_ref == 0 && !ci->i_head_snapc) { BUG_ON(!snap_rwsem_locked); @@ -2510,10 +2675,10 @@ static void __take_cap_refs(struct ceph_inode_info *ci, int got, } if (got & CEPH_CAP_FILE_BUFFER) { if (ci->i_wb_ref == 0) - ihold(&ci->vfs_inode); + ihold(&ci->netfs.inode); ci->i_wb_ref++; - dout("__take_cap_refs %p wb %d -> %d (?)\n", - &ci->vfs_inode, ci->i_wb_ref-1, ci->i_wb_ref); + dout("%s %p wb %d -> %d (?)\n", __func__, + &ci->netfs.inode, ci->i_wb_ref-1, ci->i_wb_ref); } } @@ -2524,14 +2689,16 @@ static void __take_cap_refs(struct ceph_inode_info *ci, int got, * Note that caller is responsible for ensuring max_size increases are * requested from the MDS. * - * Returns 0 if caps were not able to be acquired (yet), a 1 if they were, - * or a negative error code. - * - * FIXME: how does a 0 return differ from -EAGAIN? + * Returns 0 if caps were not able to be acquired (yet), 1 if succeed, + * or a negative error code. There are 3 speical error codes: + * -EAGAIN: need to sleep but non-blocking is specified + * -EFBIG: ask caller to call check_max_size() and try again. + * -EUCLEAN: ask caller to call ceph_renew_caps() and try again. */ enum { - NON_BLOCKING = 1, - CHECK_FILELOCK = 2, + /* first 8 bits are reserved for CEPH_FILE_MODE_FOO */ + NON_BLOCKING = (1 << 8), + CHECK_FILELOCK = (1 << 9), }; static int try_get_cap_refs(struct inode *inode, int need, int want, @@ -2541,7 +2708,6 @@ static int try_get_cap_refs(struct inode *inode, int need, int want, struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc; int ret = 0; int have, implemented; - int file_wanted; bool snap_rwsem_locked = false; dout("get_cap_refs %p need %s want %s\n", inode, @@ -2557,15 +2723,6 @@ again: goto out_unlock; } - /* make sure file is actually open */ - file_wanted = __ceph_caps_file_wanted(ci); - 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)); - ret = -EBADF; - goto out_unlock; - } - /* finish pending truncate */ while (ci->i_truncate_pending) { spin_unlock(&ci->i_ceph_lock); @@ -2584,7 +2741,7 @@ again: dout("get_cap_refs %p endoff %llu > maxsize %llu\n", inode, endoff, ci->i_max_size); if (endoff > ci->i_requested_max_size) - ret = -EAGAIN; + ret = ci->i_auth_cap ? -EFBIG : -EUCLEAN; goto out_unlock; } /* @@ -2603,13 +2760,17 @@ again: * on transition from wanted -> needed caps. This is needed * for WRBUFFER|WR -> WR to avoid a new WR sync write from * going before a prior buffered writeback happens. + * + * For RDCACHE|RD -> RD, there is not need to wait and we can + * just exclude the revoking caps and force to sync read. */ int not = want & ~(have & need); int revoking = implemented & ~have; + int exclude = revoking & not; dout("get_cap_refs %p have %s but not %s (revoking %s)\n", inode, ceph_cap_string(have), ceph_cap_string(not), ceph_cap_string(revoking)); - if ((revoking & not) == 0) { + if (!exclude || !(exclude & CEPH_CAP_FILE_BUFFER)) { if (!snap_rwsem_locked && !ci->i_head_snapc && (need & CEPH_CAP_FILE_WR)) { @@ -2630,55 +2791,60 @@ 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); + if ((have & want) == want) + *got = need | (want & ~exclude); + else + *got = need; + ceph_take_cap_refs(ci, *got, true); ret = 1; } } else { int session_readonly = false; - if ((need & CEPH_CAP_FILE_WR) && ci->i_auth_cap) { + int mds_wanted; + if (ci->i_auth_cap && + (need & (CEPH_CAP_FILE_WR | CEPH_CAP_FILE_EXCL))) { struct ceph_mds_session *s = ci->i_auth_cap->session; spin_lock(&s->s_cap_lock); session_readonly = s->s_readonly; spin_unlock(&s->s_cap_lock); } if (session_readonly) { - dout("get_cap_refs %p needed %s but mds%d readonly\n", + dout("get_cap_refs %p need %s but mds%d readonly\n", inode, ceph_cap_string(need), ci->i_auth_cap->mds); ret = -EROFS; goto out_unlock; } - if (ci->i_ceph_flags & CEPH_I_CAP_DROPPED) { - int mds_wanted; - if (READ_ONCE(mdsc->fsc->mount_state) == - CEPH_MOUNT_SHUTDOWN) { - dout("get_cap_refs %p forced umount\n", inode); - ret = -EIO; - goto out_unlock; - } - mds_wanted = __ceph_caps_mds_wanted(ci, false); - if (need & ~(mds_wanted & need)) { - dout("get_cap_refs %p caps were dropped" - " (session killed?)\n", inode); - ret = -ESTALE; - goto out_unlock; - } - if (!(file_wanted & ~mds_wanted)) - ci->i_ceph_flags &= ~CEPH_I_CAP_DROPPED; + if (ceph_inode_is_shutdown(inode)) { + dout("get_cap_refs %p inode is shutdown\n", inode); + ret = -ESTALE; + goto out_unlock; + } + mds_wanted = __ceph_caps_mds_wanted(ci, false); + if (need & ~mds_wanted) { + dout("get_cap_refs %p need %s > mds_wanted %s\n", + inode, ceph_cap_string(need), + ceph_cap_string(mds_wanted)); + ret = -EUCLEAN; + goto out_unlock; } - dout("get_cap_refs %p have %s needed %s\n", inode, + dout("get_cap_refs %p have %s need %s\n", inode, ceph_cap_string(have), ceph_cap_string(need)); } out_unlock: + + __ceph_touch_fmode(ci, mdsc, flags); + spin_unlock(&ci->i_ceph_lock); if (snap_rwsem_locked) up_read(&mdsc->snap_rwsem); + if (!ret) + ceph_update_cap_mis(&mdsc->metric); + else if (ret == 1) + ceph_update_cap_hit(&mdsc->metric); + dout("get_cap_refs %p ret %d got %s\n", inode, ret, ceph_cap_string(*got)); return ret; @@ -2712,20 +2878,40 @@ static void check_max_size(struct inode *inode, loff_t endoff) ceph_check_caps(ci, CHECK_CAPS_AUTHONLY, NULL); } +static inline int get_used_fmode(int caps) +{ + int fmode = 0; + if (caps & CEPH_CAP_FILE_RD) + fmode |= CEPH_FILE_MODE_RD; + if (caps & CEPH_CAP_FILE_WR) + fmode |= CEPH_FILE_MODE_WR; + return fmode; +} + int ceph_try_get_caps(struct inode *inode, int need, int want, bool nonblock, int *got) { - int ret; + int ret, flags; BUG_ON(need & ~CEPH_CAP_FILE_RD); - BUG_ON(want & ~(CEPH_CAP_FILE_CACHE|CEPH_CAP_FILE_LAZYIO|CEPH_CAP_FILE_SHARED)); - ret = ceph_pool_perm_check(inode, need); - if (ret < 0) - return ret; + BUG_ON(want & ~(CEPH_CAP_FILE_CACHE | CEPH_CAP_FILE_LAZYIO | + CEPH_CAP_FILE_SHARED | CEPH_CAP_FILE_EXCL | + CEPH_CAP_ANY_DIR_OPS)); + if (need) { + ret = ceph_pool_perm_check(inode, need); + if (ret < 0) + return ret; + } - ret = try_get_cap_refs(inode, need, want, 0, - (nonblock ? NON_BLOCKING : 0), got); - return ret == -EAGAIN ? 0 : ret; + flags = get_used_fmode(need | want); + if (nonblock) + flags |= NON_BLOCKING; + + ret = try_get_cap_refs(inode, need, want, 0, flags, got); + /* three special error codes */ + if (ret == -EAGAIN || ret == -EFBIG || ret == -EUCLEAN) + ret = 0; + return ret; } /* @@ -2733,8 +2919,7 @@ int ceph_try_get_caps(struct inode *inode, int need, int want, * due to a small max_size, make sure we check_max_size (and possibly * ask the mds) so we don't get hung up indefinitely. */ -int ceph_get_caps(struct file *filp, int need, int want, - loff_t endoff, int *got, struct page **pinned_page) +int ceph_get_caps(struct file *filp, int need, int want, loff_t endoff, int *got) { struct ceph_file_info *fi = filp->private_data; struct inode *inode = file_inode(filp); @@ -2750,22 +2935,22 @@ int ceph_get_caps(struct file *filp, int need, int want, fi->filp_gen != READ_ONCE(fsc->filp_gen)) return -EBADF; - while (true) { - if (endoff > 0) - check_max_size(inode, endoff); + flags = get_used_fmode(need | want); - flags = atomic_read(&fi->num_locks) ? CHECK_FILELOCK : 0; + while (true) { + flags &= CEPH_FILE_MODE_MASK; + if (atomic_read(&fi->num_locks)) + flags |= CHECK_FILELOCK; _got = 0; ret = try_get_cap_refs(inode, need, want, endoff, flags, &_got); - if (ret == -EAGAIN) - continue; + WARN_ON_ONCE(ret == -EAGAIN); if (!ret) { struct ceph_mds_client *mdsc = fsc->mdsc; struct cap_wait cw; DEFINE_WAIT_FUNC(wait, woken_wake_function); - cw.ino = inode->i_ino; + cw.ino = ceph_ino(inode); cw.tgid = current->tgid; cw.need = need; cw.want = want; @@ -2774,6 +2959,8 @@ int ceph_get_caps(struct file *filp, int need, int want, list_add(&cw.list, &mdsc->cap_wait_list); spin_unlock(&mdsc->caps_list_lock); + /* make sure used fmode not timeout */ + ceph_get_fmode(ci, flags, FMODE_WAIT_BIAS); add_wait_queue(&ci->i_cap_wq, &wait); flags |= NON_BLOCKING; @@ -2787,6 +2974,7 @@ int ceph_get_caps(struct file *filp, int need, int want, } remove_wait_queue(&ci->i_cap_wq, &wait); + ceph_put_fmode(ci, flags, FMODE_WAIT_BIAS); spin_lock(&mdsc->caps_list_lock); list_del(&cw.list); @@ -2804,26 +2992,36 @@ int ceph_get_caps(struct file *filp, int need, int want, } if (ret < 0) { - if (ret == -ESTALE) { + if (ret == -EFBIG || ret == -EUCLEAN) { + int ret2 = ceph_wait_on_async_create(inode); + if (ret2 < 0) + return ret2; + } + if (ret == -EFBIG) { + check_max_size(inode, endoff); + continue; + } + if (ret == -EUCLEAN) { /* session was killed, try renew caps */ - ret = ceph_renew_caps(inode); + ret = ceph_renew_caps(inode, flags); if (ret == 0) continue; } return ret; } - if (ci->i_inline_version != CEPH_INLINE_NONE && + if (S_ISREG(ci->netfs.inode.i_mode) && + ceph_has_inline_data(ci) && (_got & (CEPH_CAP_FILE_CACHE|CEPH_CAP_FILE_LAZYIO)) && i_size_read(inode) > 0) { struct page *page = find_get_page(inode->i_mapping, 0); if (page) { - if (PageUptodate(page)) { - *pinned_page = page; - break; - } + bool uptodate = PageUptodate(page); + put_page(page); + if (uptodate) + break; } /* * drop cap refs first because getattr while @@ -2845,10 +3043,6 @@ int ceph_get_caps(struct file *filp, 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; } @@ -2860,7 +3054,7 @@ int ceph_get_caps(struct file *filp, int need, int want, void ceph_get_cap_refs(struct ceph_inode_info *ci, int caps) { spin_lock(&ci->i_ceph_lock); - __take_cap_refs(ci, caps, false); + ceph_take_cap_refs(ci, caps, false); spin_unlock(&ci->i_ceph_lock); } @@ -2888,6 +3082,12 @@ static int ceph_try_drop_cap_snap(struct ceph_inode_info *ci, return 0; } +enum put_cap_refs_mode { + PUT_CAP_REFS_SYNC = 0, + PUT_CAP_REFS_NO_CHECK, + PUT_CAP_REFS_ASYNC, +}; + /* * Release cap refs. * @@ -2897,10 +3097,12 @@ static int ceph_try_drop_cap_snap(struct ceph_inode_info *ci, * If we are releasing a WR cap (from a sync write), finalize any affected * cap_snap, and wake up any waiters. */ -void ceph_put_cap_refs(struct ceph_inode_info *ci, int had) +static void __ceph_put_cap_refs(struct ceph_inode_info *ci, int had, + enum put_cap_refs_mode mode) { - struct inode *inode = &ci->vfs_inode; + struct inode *inode = &ci->netfs.inode; int last = 0, put = 0, flushsnaps = 0, wake = 0; + bool check_flushsnaps = false; spin_lock(&ci->i_ceph_lock); if (had & CEPH_CAP_PIN) @@ -2911,29 +3113,23 @@ void ceph_put_cap_refs(struct ceph_inode_info *ci, int had) if (had & CEPH_CAP_FILE_CACHE) if (--ci->i_rdcache_ref == 0) last++; + if (had & CEPH_CAP_FILE_EXCL) + if (--ci->i_fx_ref == 0) + last++; if (had & CEPH_CAP_FILE_BUFFER) { if (--ci->i_wb_ref == 0) { last++; + /* put the ref held by ceph_take_cap_refs() */ put++; + check_flushsnaps = true; } dout("put_cap_refs %p wb %d -> %d (?)\n", inode, ci->i_wb_ref+1, ci->i_wb_ref); } - if (had & CEPH_CAP_FILE_WR) + if (had & CEPH_CAP_FILE_WR) { if (--ci->i_wr_ref == 0) { last++; - if (__ceph_have_pending_cap_snap(ci)) { - struct ceph_cap_snap *capsnap = - list_last_entry(&ci->i_cap_snaps, - struct ceph_cap_snap, - ci_item); - capsnap->writing = 0; - if (ceph_try_drop_cap_snap(ci, capsnap)) - put++; - else if (__ceph_finish_cap_snap(ci, capsnap)) - flushsnaps = 1; - wake = 1; - } + check_flushsnaps = true; if (ci->i_wrbuffer_ref_head == 0 && ci->i_dirty_caps == 0 && ci->i_flushing_caps == 0) { @@ -2943,23 +3139,65 @@ void ceph_put_cap_refs(struct ceph_inode_info *ci, int had) } /* see comment in __ceph_remove_cap() */ if (!__ceph_is_any_real_caps(ci) && ci->i_snap_realm) - drop_inode_snap_realm(ci); + ceph_change_snap_realm(inode, NULL); } + } + if (check_flushsnaps && __ceph_have_pending_cap_snap(ci)) { + struct ceph_cap_snap *capsnap = + list_last_entry(&ci->i_cap_snaps, + struct ceph_cap_snap, + ci_item); + + capsnap->writing = 0; + if (ceph_try_drop_cap_snap(ci, capsnap)) + /* put the ref held by ceph_queue_cap_snap() */ + put++; + else if (__ceph_finish_cap_snap(ci, capsnap)) + flushsnaps = 1; + wake = 1; + } spin_unlock(&ci->i_ceph_lock); dout("put_cap_refs %p had %s%s%s\n", inode, ceph_cap_string(had), last ? " last" : "", put ? " put" : ""); - if (last && !flushsnaps) - ceph_check_caps(ci, 0, NULL); - else if (flushsnaps) - ceph_flush_snaps(ci, NULL); + switch (mode) { + case PUT_CAP_REFS_SYNC: + if (last) + ceph_check_caps(ci, 0, NULL); + else if (flushsnaps) + ceph_flush_snaps(ci, NULL); + break; + case PUT_CAP_REFS_ASYNC: + if (last) + ceph_queue_check_caps(inode); + else if (flushsnaps) + ceph_queue_flush_snaps(inode); + break; + default: + break; + } if (wake) wake_up_all(&ci->i_cap_wq); while (put-- > 0) iput(inode); } +void ceph_put_cap_refs(struct ceph_inode_info *ci, int had) +{ + __ceph_put_cap_refs(ci, had, PUT_CAP_REFS_SYNC); +} + +void ceph_put_cap_refs_async(struct ceph_inode_info *ci, int had) +{ + __ceph_put_cap_refs(ci, had, PUT_CAP_REFS_ASYNC); +} + +void ceph_put_cap_refs_no_check_caps(struct ceph_inode_info *ci, int had) +{ + __ceph_put_cap_refs(ci, had, PUT_CAP_REFS_NO_CHECK); +} + /* * Release @nr WRBUFFER refs on dirty pages for the given @snapc snap * context. Adjust per-snap dirty page accounting as appropriate. @@ -2970,11 +3208,10 @@ void ceph_put_cap_refs(struct ceph_inode_info *ci, int had) void ceph_put_wrbuffer_cap_refs(struct ceph_inode_info *ci, int nr, struct ceph_snap_context *snapc) { - struct inode *inode = &ci->vfs_inode; - struct ceph_cap_snap *capsnap = NULL; + struct inode *inode = &ci->netfs.inode; + struct ceph_cap_snap *capsnap = NULL, *iter; int put = 0; bool last = false; - bool found = false; bool flush_snaps = false; bool complete_capsnap = false; @@ -3001,13 +3238,22 @@ void ceph_put_wrbuffer_cap_refs(struct ceph_inode_info *ci, int nr, ci->i_wrbuffer_ref, ci->i_wrbuffer_ref_head, last ? " LAST" : ""); } else { - list_for_each_entry(capsnap, &ci->i_cap_snaps, ci_item) { - if (capsnap->context == snapc) { - found = true; + list_for_each_entry(iter, &ci->i_cap_snaps, ci_item) { + if (iter->context == snapc) { + capsnap = iter; break; } } - BUG_ON(!found); + + if (!capsnap) { + /* + * The capsnap should already be removed when removing + * auth cap in the case of a forced unmount. + */ + WARN_ON_ONCE(ci->i_auth_cap); + goto unlock; + } + capsnap->dirty_pages -= nr; if (capsnap->dirty_pages == 0) { complete_capsnap = true; @@ -3029,18 +3275,18 @@ void ceph_put_wrbuffer_cap_refs(struct ceph_inode_info *ci, int nr, complete_capsnap ? " (complete capsnap)" : ""); } +unlock: spin_unlock(&ci->i_ceph_lock); if (last) { - ceph_check_caps(ci, CHECK_CAPS_AUTHONLY, NULL); + ceph_check_caps(ci, 0, NULL); } else if (flush_snaps) { ceph_flush_snaps(ci, NULL); } if (complete_capsnap) wake_up_all(&ci->i_cap_wq); while (put-- > 0) { - /* avoid calling iput_final() in osd dispatch threads */ - ceph_async_iput(inode); + iput(inode); } } @@ -3114,7 +3360,7 @@ static void handle_cap_grant(struct inode *inode, u64 size = le64_to_cpu(grant->size); u64 max_size = le64_to_cpu(grant->max_size); unsigned char check_caps = 0; - bool was_stale = cap->cap_gen < session->s_cap_gen; + bool was_stale = cap->cap_gen < atomic_read(&session->s_cap_gen); bool wake = false; bool writeback = false; bool queue_trunc = false; @@ -3125,7 +3371,7 @@ static void handle_cap_grant(struct inode *inode, dout("handle_cap_grant inode %p cap %p mds%d seq %d %s\n", inode, cap, session->s_mds, seq, ceph_cap_string(newcaps)); dout(" size %llu max_size %llu, i_size %llu\n", size, max_size, - inode->i_size); + i_size_read(inode)); /* @@ -3133,7 +3379,7 @@ static void handle_cap_grant(struct inode *inode, * try to invalidate (once). (If there are dirty buffers, we * will invalidate _after_ writeback.) */ - if (!S_ISDIR(inode->i_mode) && /* don't invalidate readdir cache */ + if (S_ISREG(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_wb_ref)) { @@ -3166,7 +3412,7 @@ static void handle_cap_grant(struct inode *inode, } /* side effects now are allowed */ - cap->cap_gen = session->s_cap_gen; + cap->cap_gen = atomic_read(&session->s_cap_gen); cap->seq = seq; __check_cap_issue(ci, cap, newcaps); @@ -3175,7 +3421,13 @@ static void handle_cap_grant(struct inode *inode, if ((newcaps & CEPH_CAP_AUTH_SHARED) && (extra_info->issued & CEPH_CAP_AUTH_EXCL) == 0) { - inode->i_mode = le32_to_cpu(grant->mode); + umode_t mode = le32_to_cpu(grant->mode); + + if (inode_wrong_type(inode, mode)) + pr_warn_once("inode type changed! (ino %llx.%llx is 0%o, mds says 0%o)\n", + ceph_vinop(inode), inode->i_mode, mode); + else + inode->i_mode = mode; inode->i_uid = make_kuid(&init_user_ns, le32_to_cpu(grant->uid)); inode->i_gid = make_kgid(&init_user_ns, le32_to_cpu(grant->gid)); ci->i_btime = extra_info->btime; @@ -3187,8 +3439,7 @@ static void handle_cap_grant(struct inode *inode, if ((newcaps & CEPH_CAP_LINK_SHARED) && (extra_info->issued & CEPH_CAP_LINK_EXCL) == 0) { set_nlink(inode, le32_to_cpu(grant->nlink)); - if (inode->i_nlink == 0 && - (newcaps & (CEPH_CAP_LINK_SHARED | CEPH_CAP_LINK_EXCL))) + if (inode->i_nlink == 0) deleted_inode = true; } @@ -3258,10 +3509,6 @@ static void handle_cap_grant(struct inode *inode, ci->i_requested_max_size = 0; } wake = true; - } else if (ci->i_wanted_max_size > ci->i_max_size && - ci->i_wanted_max_size > ci->i_requested_max_size) { - /* CEPH_CAP_OP_IMPORT */ - wake = true; } } @@ -3297,16 +3544,20 @@ static void handle_cap_grant(struct inode *inode, ceph_cap_string(cap->issued), ceph_cap_string(newcaps), ceph_cap_string(revoking)); - if (revoking & used & CEPH_CAP_FILE_BUFFER) + if (S_ISREG(inode->i_mode) && + (revoking & used & CEPH_CAP_FILE_BUFFER)) writeback = true; /* initiate writeback; will delay ack */ - else if (revoking == CEPH_CAP_FILE_CACHE && - (newcaps & CEPH_CAP_FILE_LAZYIO) == 0 && - queue_invalidate) + else if (queue_invalidate && + revoking == CEPH_CAP_FILE_CACHE && + (newcaps & CEPH_CAP_FILE_LAZYIO) == 0) ; /* do nothing yet, invalidation will be queued */ else if (cap == ci->i_auth_cap) check_caps = 1; /* check auth cap only */ else check_caps = 2; /* check all caps */ + /* If there is new caps, try to wake up the waiters */ + if (~cap->issued & newcaps) + wake = true; cap->issued = newcaps; cap->implemented |= newcaps; } else if (cap->issued == newcaps) { @@ -3337,13 +3588,22 @@ static void handle_cap_grant(struct inode *inode, } if (le32_to_cpu(grant->op) == CEPH_CAP_OP_IMPORT) { - if (newcaps & ~extra_info->issued) - wake = true; - kick_flushing_inode_caps(session->s_mdsc, session, inode); + if (ci->i_auth_cap == cap) { + if (newcaps & ~extra_info->issued) + wake = true; + + if (ci->i_requested_max_size > max_size || + !(le32_to_cpu(grant->wanted) & CEPH_CAP_ANY_FILE_WR)) { + /* re-request max_size if necessary */ + ci->i_requested_max_size = 0; + wake = true; + } + + ceph_kick_flushing_inode_caps(session, ci); + } up_read(&session->s_mdsc->snap_rwsem); - } else { - spin_unlock(&ci->i_ceph_lock); } + spin_unlock(&ci->i_ceph_lock); if (fill_inline) ceph_fill_inline_data(inode, NULL, extra_info->inline_data, @@ -3366,13 +3626,12 @@ static void handle_cap_grant(struct inode *inode, if (wake) wake_up_all(&ci->i_cap_wq); + mutex_unlock(&session->s_mutex); if (check_caps == 1) - ceph_check_caps(ci, CHECK_CAPS_NODELAY|CHECK_CAPS_AUTHONLY, + ceph_check_caps(ci, CHECK_CAPS_AUTHONLY | CHECK_CAPS_NOINVAL, session); else if (check_caps == 2) - ceph_check_caps(ci, CHECK_CAPS_NODELAY, session); - else - mutex_unlock(&session->s_mutex); + ceph_check_caps(ci, CHECK_CAPS_NOINVAL, session); } /* @@ -3397,15 +3656,26 @@ static void handle_cap_flush_ack(struct inode *inode, u64 flush_tid, bool wake_mdsc = false; list_for_each_entry_safe(cf, tmp_cf, &ci->i_cap_flush_list, i_list) { + /* Is this the one that was flushed? */ if (cf->tid == flush_tid) cleaned = cf->caps; - if (cf->caps == 0) /* capsnap */ + + /* Is this a capsnap? */ + if (cf->is_capsnap) continue; + if (cf->tid <= flush_tid) { - if (__finish_cap_flush(NULL, ci, cf)) - wake_ci = true; + /* + * An earlier or current tid. The FLUSH_ACK should + * represent a superset of this flush's caps. + */ + wake_ci |= __detach_cap_flush_from_ci(ci, cf); list_add_tail(&cf->i_list, &to_remove); } else { + /* + * This is a later one. Any caps in it are still dirty + * so don't count them as cleaned. + */ cleaned &= ~cf->caps; if (!cleaned) break; @@ -3425,10 +3695,8 @@ static void handle_cap_flush_ack(struct inode *inode, u64 flush_tid, spin_lock(&mdsc->cap_dirty_lock); - list_for_each_entry(cf, &to_remove, i_list) { - if (__finish_cap_flush(mdsc, NULL, cf)) - wake_mdsc = true; - } + list_for_each_entry(cf, &to_remove, i_list) + wake_mdsc |= __detach_cap_flush_from_mdsc(mdsc, cf); if (ci->i_flushing_caps == 0) { if (list_empty(&ci->i_cap_flush_list)) { @@ -3438,7 +3706,7 @@ static void handle_cap_flush_ack(struct inode *inode, u64 flush_tid, session->s_mds, &list_first_entry(&session->s_cap_flushing, struct ceph_inode_info, - i_flushing_item)->vfs_inode); + i_flushing_item)->netfs.inode); } } mdsc->num_cap_flushing--; @@ -3466,8 +3734,9 @@ out: while (!list_empty(&to_remove)) { cf = list_first_entry(&to_remove, struct ceph_cap_flush, i_list); - list_del(&cf->i_list); - ceph_free_cap_flush(cf); + list_del_init(&cf->i_list); + if (!cf->is_capsnap) + ceph_free_cap_flush(cf); } if (wake_ci) @@ -3478,6 +3747,43 @@ out: iput(inode); } +void __ceph_remove_capsnap(struct inode *inode, struct ceph_cap_snap *capsnap, + bool *wake_ci, bool *wake_mdsc) +{ + struct ceph_inode_info *ci = ceph_inode(inode); + struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; + bool ret; + + lockdep_assert_held(&ci->i_ceph_lock); + + dout("removing capsnap %p, inode %p ci %p\n", capsnap, inode, ci); + + list_del_init(&capsnap->ci_item); + ret = __detach_cap_flush_from_ci(ci, &capsnap->cap_flush); + if (wake_ci) + *wake_ci = ret; + + spin_lock(&mdsc->cap_dirty_lock); + if (list_empty(&ci->i_cap_flush_list)) + list_del_init(&ci->i_flushing_item); + + ret = __detach_cap_flush_from_mdsc(mdsc, &capsnap->cap_flush); + if (wake_mdsc) + *wake_mdsc = ret; + spin_unlock(&mdsc->cap_dirty_lock); +} + +void ceph_remove_capsnap(struct inode *inode, struct ceph_cap_snap *capsnap, + bool *wake_ci, bool *wake_mdsc) +{ + struct ceph_inode_info *ci = ceph_inode(inode); + + lockdep_assert_held(&ci->i_ceph_lock); + + WARN_ON_ONCE(capsnap->dirty_pages || capsnap->writing); + __ceph_remove_capsnap(inode, capsnap, wake_ci, wake_mdsc); +} + /* * Handle FLUSHSNAP_ACK. MDS has flushed snap data to disk and we can * throw away our cap_snap. @@ -3491,8 +3797,7 @@ static void handle_cap_flushsnap_ack(struct inode *inode, u64 flush_tid, struct ceph_inode_info *ci = ceph_inode(inode); struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; u64 follows = le64_to_cpu(m->snap_follows); - struct ceph_cap_snap *capsnap; - bool flushed = false; + struct ceph_cap_snap *capsnap = NULL, *iter; bool wake_ci = false; bool wake_mdsc = false; @@ -3500,41 +3805,26 @@ static void handle_cap_flushsnap_ack(struct inode *inode, u64 flush_tid, inode, ci, session->s_mds, follows); spin_lock(&ci->i_ceph_lock); - list_for_each_entry(capsnap, &ci->i_cap_snaps, ci_item) { - if (capsnap->follows == follows) { - if (capsnap->cap_flush.tid != flush_tid) { + list_for_each_entry(iter, &ci->i_cap_snaps, ci_item) { + if (iter->follows == follows) { + if (iter->cap_flush.tid != flush_tid) { dout(" cap_snap %p follows %lld tid %lld !=" - " %lld\n", capsnap, follows, - flush_tid, capsnap->cap_flush.tid); + " %lld\n", iter, follows, + flush_tid, iter->cap_flush.tid); break; } - flushed = true; + capsnap = iter; break; } else { dout(" skipping cap_snap %p follows %lld\n", - capsnap, capsnap->follows); + iter, iter->follows); } } - if (flushed) { - WARN_ON(capsnap->dirty_pages || capsnap->writing); - dout(" removing %p cap_snap %p follows %lld\n", - inode, capsnap, follows); - list_del(&capsnap->ci_item); - if (__finish_cap_flush(NULL, ci, &capsnap->cap_flush)) - wake_ci = true; - - spin_lock(&mdsc->cap_dirty_lock); - - if (list_empty(&ci->i_cap_flush_list)) - list_del_init(&ci->i_flushing_item); - - if (__finish_cap_flush(mdsc, NULL, &capsnap->cap_flush)) - wake_mdsc = true; - - spin_unlock(&mdsc->cap_dirty_lock); - } + if (capsnap) + ceph_remove_capsnap(inode, capsnap, &wake_ci, &wake_mdsc); spin_unlock(&ci->i_ceph_lock); - if (flushed) { + + if (capsnap) { ceph_put_snap_context(capsnap->context); ceph_put_cap_snap(capsnap); if (wake_ci) @@ -3550,10 +3840,9 @@ static void handle_cap_flushsnap_ack(struct inode *inode, u64 flush_tid, * * caller hold s_mutex. */ -static void handle_cap_trunc(struct inode *inode, +static bool handle_cap_trunc(struct inode *inode, struct ceph_mds_caps *trunc, struct ceph_mds_session *session) - __releases(ci->i_ceph_lock) { struct ceph_inode_info *ci = ceph_inode(inode); int mds = session->s_mds; @@ -3564,7 +3853,9 @@ static void handle_cap_trunc(struct inode *inode, int implemented = 0; int dirty = __ceph_caps_dirty(ci); int issued = __ceph_caps_issued(ceph_inode(inode), &implemented); - int queue_trunc = 0; + bool queue_trunc = false; + + lockdep_assert_held(&ci->i_ceph_lock); issued |= implemented | dirty; @@ -3572,10 +3863,7 @@ static void handle_cap_trunc(struct inode *inode, inode, mds, seq, truncate_size, truncate_seq); queue_trunc = ceph_fill_file_size(inode, issued, truncate_seq, truncate_size, size); - spin_unlock(&ci->i_ceph_lock); - - if (queue_trunc) - ceph_queue_vmtruncate(inode); + return queue_trunc; } /* @@ -3613,15 +3901,14 @@ static void handle_cap_export(struct inode *inode, struct ceph_mds_caps *ex, dout("handle_cap_export inode %p ci %p mds%d mseq %d target %d\n", inode, ci, mds, mseq, target); retry: + down_read(&mdsc->snap_rwsem); spin_lock(&ci->i_ceph_lock); cap = __get_cap_for_mds(ci, mds); if (!cap || cap->cap_id != le64_to_cpu(ex->cap_id)) goto out_unlock; if (target < 0) { - if (cap->mds_wanted | cap->issued) - ci->i_ceph_flags |= CEPH_I_CAP_DROPPED; - __ceph_remove_cap(cap, false); + ceph_remove_cap(cap, false); goto out_unlock; } @@ -3651,24 +3938,18 @@ retry: tcap->issue_seq = t_seq - 1; tcap->issued |= issued; tcap->implemented |= issued; - if (cap == ci->i_auth_cap) + if (cap == ci->i_auth_cap) { ci->i_auth_cap = tcap; - - if (!list_empty(&ci->i_cap_flush_list) && - ci->i_auth_cap == tcap) { - spin_lock(&mdsc->cap_dirty_lock); - list_move_tail(&ci->i_flushing_item, - &tcap->session->s_cap_flushing); - spin_unlock(&mdsc->cap_dirty_lock); + change_auth_cap_ses(ci, tcap->session); } } - __ceph_remove_cap(cap, false); + ceph_remove_cap(cap, false); goto out_unlock; } else if (tsession) { /* add placeholder for the export tagert */ int flag = (cap == ci->i_auth_cap) ? CEPH_CAP_FLAG_AUTH : 0; tcap = new_cap; - ceph_add_cap(inode, tsession, t_cap_id, -1, issued, 0, + ceph_add_cap(inode, tsession, t_cap_id, issued, 0, t_seq - 1, t_mseq, (u64)-1, flag, &new_cap); if (!list_empty(&ci->i_cap_flush_list) && @@ -3679,11 +3960,12 @@ retry: spin_unlock(&mdsc->cap_dirty_lock); } - __ceph_remove_cap(cap, false); + ceph_remove_cap(cap, false); goto out_unlock; } spin_unlock(&ci->i_ceph_lock); + up_read(&mdsc->snap_rwsem); mutex_unlock(&session->s_mutex); /* open target session */ @@ -3703,11 +3985,13 @@ retry: WARN_ON(1); tsession = NULL; target = -1; + mutex_lock(&session->s_mutex); } goto retry; out_unlock: spin_unlock(&ci->i_ceph_lock); + up_read(&mdsc->snap_rwsem); mutex_unlock(&session->s_mutex); if (tsession) { mutex_unlock(&tsession->s_mutex); @@ -3727,7 +4011,6 @@ static void handle_cap_import(struct ceph_mds_client *mdsc, struct ceph_mds_cap_peer *ph, struct ceph_mds_session *session, struct ceph_cap **target_cap, int *old_issued) - __acquires(ci->i_ceph_lock) { struct ceph_inode_info *ci = ceph_inode(inode); struct ceph_cap *cap, *ocap, *new_cap = NULL; @@ -3752,14 +4035,13 @@ static void handle_cap_import(struct ceph_mds_client *mdsc, dout("handle_cap_import inode %p ci %p mds%d mseq %d peer %d\n", inode, ci, mds, mseq, peer); - retry: - spin_lock(&ci->i_ceph_lock); cap = __get_cap_for_mds(ci, mds); if (!cap) { if (!new_cap) { spin_unlock(&ci->i_ceph_lock); new_cap = ceph_get_cap(mdsc, NULL); + spin_lock(&ci->i_ceph_lock); goto retry; } cap = new_cap; @@ -3773,7 +4055,7 @@ retry: __ceph_caps_issued(ci, &issued); issued |= __ceph_caps_dirty(ci); - ceph_add_cap(inode, session, cap_id, -1, caps, wanted, seq, mseq, + ceph_add_cap(inode, session, cap_id, caps, wanted, seq, mseq, realmino, CEPH_CAP_FLAG_AUTH, &new_cap); ocap = peer >= 0 ? __get_cap_for_mds(ci, peer) : NULL; @@ -3791,12 +4073,9 @@ retry: ocap->mseq, mds, le32_to_cpu(ph->seq), le32_to_cpu(ph->mseq)); } - __ceph_remove_cap(ocap, (ph->flags & CEPH_CAP_FLAG_RELEASE)); + ceph_remove_cap(ocap, (ph->flags & CEPH_CAP_FLAG_RELEASE)); } - /* make sure we re-request max_size, if necessary */ - ci->i_requested_max_size = 0; - *old_issued = issued; *target_cap = cap; } @@ -3825,6 +4104,7 @@ void ceph_handle_caps(struct ceph_mds_session *session, size_t snaptrace_len; void *p, *end; struct cap_extra_info extra_info = {}; + bool queue_trunc; dout("handle_caps from mds%d\n", session->s_mds); @@ -3881,15 +4161,13 @@ void ceph_handle_caps(struct ceph_mds_session *session, } if (msg_version >= 8) { - u64 flush_tid; - u32 caller_uid, caller_gid; u32 pool_ns_len; /* version >= 6 */ - ceph_decode_64_safe(&p, end, flush_tid, bad); + ceph_decode_skip_64(&p, end, bad); // flush_tid /* version >= 7 */ - ceph_decode_32_safe(&p, end, caller_uid, bad); - ceph_decode_32_safe(&p, end, caller_gid, bad); + ceph_decode_skip_32(&p, end, bad); // caller_uid + ceph_decode_skip_32(&p, end, bad); // caller_gid /* version >= 8 */ ceph_decode_32_safe(&p, end, pool_ns_len, bad); if (pool_ns_len > 0) { @@ -3912,9 +4190,8 @@ void ceph_handle_caps(struct ceph_mds_session *session, } if (msg_version >= 11) { - u32 flags; /* version >= 10 */ - ceph_decode_32_safe(&p, end, flags, bad); + ceph_decode_skip_32(&p, end, bad); // flags /* version >= 11 */ extra_info.dirstat_valid = true; ceph_decode_64_safe(&p, end, extra_info.nfiles, bad); @@ -3923,12 +4200,11 @@ void ceph_handle_caps(struct ceph_mds_session *session, /* lookup ino */ inode = ceph_find_inode(mdsc->fsc->sb, vino); - ci = ceph_inode(inode); dout(" op %s ino %llx.%llx inode %p\n", ceph_cap_op_name(op), vino.ino, vino.snap, inode); mutex_lock(&session->s_mutex); - session->s_seq++; + inc_session_sequence(session); dout(" mds%d seq %lld cap seq %u\n", session->s_mds, session->s_seq, (unsigned)seq); @@ -3947,8 +4223,9 @@ void ceph_handle_caps(struct ceph_mds_session *session, __ceph_queue_cap_release(session, cap); spin_unlock(&session->s_cap_lock); } - goto done; + goto flush_cap_releases; } + ci = ceph_inode(inode); /* these will work even if we don't have a cap yet */ switch (op) { @@ -3972,6 +4249,7 @@ void ceph_handle_caps(struct ceph_mds_session *session, } else { down_read(&mdsc->snap_rwsem); } + spin_lock(&ci->i_ceph_lock); handle_cap_import(mdsc, inode, h, peer, session, &cap, &extra_info.issued); handle_cap_grant(inode, session, cap, @@ -4008,7 +4286,10 @@ void ceph_handle_caps(struct ceph_mds_session *session, break; case CEPH_CAP_OP_TRUNC: - handle_cap_trunc(inode, h, session); + queue_trunc = handle_cap_trunc(inode, h, session); + spin_unlock(&ci->i_ceph_lock); + if (queue_trunc) + ceph_queue_vmtruncate(inode); break; default: @@ -4020,9 +4301,9 @@ void ceph_handle_caps(struct ceph_mds_session *session, done: mutex_unlock(&session->s_mutex); done_unlocked: + iput(inode); +out: ceph_put_string(extra_info.pool_ns); - /* avoid calling iput_final() in mds dispatch threads */ - ceph_async_iput(inode); return; flush_cap_releases: @@ -4037,62 +4318,75 @@ flush_cap_releases: bad: pr_err("ceph_handle_caps: corrupt message\n"); ceph_msg_dump(msg); - return; + goto out; } /* * Delayed work handler to process end of delayed cap release LRU list. + * + * If new caps are added to the list while processing it, these won't get + * processed in this run. In this case, the ci->i_hold_caps_max will be + * returned so that the work can be scheduled accordingly. */ -void ceph_check_delayed_caps(struct ceph_mds_client *mdsc) +unsigned long ceph_check_delayed_caps(struct ceph_mds_client *mdsc) { struct inode *inode; struct ceph_inode_info *ci; - int flags = CHECK_CAPS_NODELAY; + struct ceph_mount_options *opt = mdsc->fsc->mount_options; + unsigned long delay_max = opt->caps_wanted_delay_max * HZ; + unsigned long loop_start = jiffies; + unsigned long delay = 0; dout("check_delayed_caps\n"); - while (1) { - spin_lock(&mdsc->cap_delay_lock); - if (list_empty(&mdsc->cap_delay_list)) - break; + spin_lock(&mdsc->cap_delay_lock); + while (!list_empty(&mdsc->cap_delay_list)) { ci = list_first_entry(&mdsc->cap_delay_list, struct ceph_inode_info, i_cap_delay_list); + if (time_before(loop_start, ci->i_hold_caps_max - delay_max)) { + dout("%s caps added recently. Exiting loop", __func__); + delay = ci->i_hold_caps_max; + break; + } if ((ci->i_ceph_flags & CEPH_I_FLUSH) == 0 && time_before(jiffies, ci->i_hold_caps_max)) break; list_del_init(&ci->i_cap_delay_list); - inode = igrab(&ci->vfs_inode); - spin_unlock(&mdsc->cap_delay_lock); - + inode = igrab(&ci->netfs.inode); if (inode) { + spin_unlock(&mdsc->cap_delay_lock); dout("check_delayed_caps on %p\n", inode); - ceph_check_caps(ci, flags, NULL); - /* avoid calling iput_final() in tick thread */ - ceph_async_iput(inode); + ceph_check_caps(ci, 0, NULL); + iput(inode); + spin_lock(&mdsc->cap_delay_lock); } } spin_unlock(&mdsc->cap_delay_lock); + + return delay; } /* * Flush all dirty caps to the mds */ -void ceph_flush_dirty_caps(struct ceph_mds_client *mdsc) +static void flush_dirty_session_caps(struct ceph_mds_session *s) { + struct ceph_mds_client *mdsc = s->s_mdsc; struct ceph_inode_info *ci; struct inode *inode; dout("flush_dirty_caps\n"); spin_lock(&mdsc->cap_dirty_lock); - while (!list_empty(&mdsc->cap_dirty)) { - ci = list_first_entry(&mdsc->cap_dirty, struct ceph_inode_info, + while (!list_empty(&s->s_cap_dirty)) { + ci = list_first_entry(&s->s_cap_dirty, struct ceph_inode_info, i_dirty_item); - inode = &ci->vfs_inode; + inode = &ci->netfs.inode; ihold(inode); - dout("flush_dirty_caps %p\n", inode); + dout("flush_dirty_caps %llx.%llx\n", ceph_vinop(inode)); spin_unlock(&mdsc->cap_dirty_lock); - ceph_check_caps(ci, CHECK_CAPS_NODELAY|CHECK_CAPS_FLUSH, NULL); + ceph_wait_on_async_create(inode); + ceph_check_caps(ci, CHECK_CAPS_FLUSH, NULL); iput(inode); spin_lock(&mdsc->cap_dirty_lock); } @@ -4100,14 +4394,53 @@ void ceph_flush_dirty_caps(struct ceph_mds_client *mdsc) dout("flush_dirty_caps done\n"); } -void __ceph_get_fmode(struct ceph_inode_info *ci, int fmode) +void ceph_flush_dirty_caps(struct ceph_mds_client *mdsc) { - int i; + ceph_mdsc_iterate_sessions(mdsc, flush_dirty_session_caps, true); +} + +void __ceph_touch_fmode(struct ceph_inode_info *ci, + struct ceph_mds_client *mdsc, int fmode) +{ + unsigned long now = jiffies; + if (fmode & CEPH_FILE_MODE_RD) + ci->i_last_rd = now; + if (fmode & CEPH_FILE_MODE_WR) + ci->i_last_wr = now; + /* queue periodic check */ + if (fmode && + __ceph_is_any_real_caps(ci) && + list_empty(&ci->i_cap_delay_list)) + __cap_delay_requeue(mdsc, ci); +} + +void ceph_get_fmode(struct ceph_inode_info *ci, int fmode, int count) +{ + struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(ci->netfs.inode.i_sb); int bits = (fmode << 1) | 1; + bool already_opened = false; + int i; + + if (count == 1) + atomic64_inc(&mdsc->metric.opened_files); + + spin_lock(&ci->i_ceph_lock); for (i = 0; i < CEPH_FILE_MODE_BITS; i++) { + /* + * If any of the mode ref is larger than 0, + * that means it has been already opened by + * others. Just skip checking the PIN ref. + */ + if (i && ci->i_nr_by_mode[i]) + already_opened = true; + if (bits & (1 << i)) - ci->i_nr_by_mode[i]++; + ci->i_nr_by_mode[i] += count; } + + if (!already_opened) + percpu_counter_inc(&mdsc->metric.opened_inodes); + spin_unlock(&ci->i_ceph_lock); } /* @@ -4115,26 +4448,35 @@ void __ceph_get_fmode(struct ceph_inode_info *ci, int fmode) * we may need to release capabilities to the MDS (or schedule * their delayed release). */ -void ceph_put_fmode(struct ceph_inode_info *ci, int fmode) +void ceph_put_fmode(struct ceph_inode_info *ci, int fmode, int count) { - int i, last = 0; + struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(ci->netfs.inode.i_sb); int bits = (fmode << 1) | 1; + bool is_closed = true; + int i; + + if (count == 1) + atomic64_dec(&mdsc->metric.opened_files); + spin_lock(&ci->i_ceph_lock); for (i = 0; i < CEPH_FILE_MODE_BITS; i++) { if (bits & (1 << i)) { - BUG_ON(ci->i_nr_by_mode[i] == 0); - if (--ci->i_nr_by_mode[i] == 0) - last++; + BUG_ON(ci->i_nr_by_mode[i] < count); + ci->i_nr_by_mode[i] -= count; } + + /* + * If any of the mode ref is not 0 after + * decreased, that means it is still opened + * by others. Just skip checking the PIN ref. + */ + if (i && ci->i_nr_by_mode[i]) + is_closed = false; } - dout("put_fmode %p fmode %d {%d,%d,%d,%d}\n", - &ci->vfs_inode, fmode, - ci->i_nr_by_mode[0], ci->i_nr_by_mode[1], - ci->i_nr_by_mode[2], ci->i_nr_by_mode[3]); - spin_unlock(&ci->i_ceph_lock); - if (last && ci->i_vino.snap == CEPH_NOSNAP) - ceph_check_caps(ci, 0, NULL); + if (is_closed) + percpu_counter_dec(&mdsc->metric.opened_inodes); + spin_unlock(&ci->i_ceph_lock); } /* @@ -4152,7 +4494,6 @@ int ceph_drop_caps_for_unlink(struct inode *inode) if (inode->i_nlink == 1) { drop |= ~(__ceph_caps_wanted(ci) | CEPH_CAP_PIN); - ci->i_ceph_flags |= CEPH_I_NODELAY; if (__ceph_caps_dirty(ci)) { struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc; @@ -4208,8 +4549,6 @@ int ceph_encode_inode_release(void **p, struct inode *inode, if (force || (cap->issued & drop)) { if (cap->issued & drop) { int wanted = __ceph_caps_wanted(ci); - if ((ci->i_ceph_flags & CEPH_I_NODELAY) == 0) - wanted |= cap->mds_wanted; dout("encode_inode_release %p cap %p " "%s -> %s, wanted %s -> %s\n", inode, cap, ceph_cap_string(cap->issued), @@ -4220,6 +4559,9 @@ int ceph_encode_inode_release(void **p, struct inode *inode, cap->issued &= ~drop; cap->implemented &= ~drop; cap->mds_wanted = wanted; + if (cap == ci->i_auth_cap && + !(wanted & CEPH_CAP_ANY_FILE_WR)) + ci->i_requested_max_size = 0; } else { dout("encode_inode_release %p cap %p %s" " (force)\n", inode, cap, @@ -4287,3 +4629,119 @@ int ceph_encode_dentry_release(void **p, struct dentry *dentry, spin_unlock(&dentry->d_lock); return ret; } + +static int remove_capsnaps(struct ceph_mds_client *mdsc, struct inode *inode) +{ + struct ceph_inode_info *ci = ceph_inode(inode); + struct ceph_cap_snap *capsnap; + int capsnap_release = 0; + + lockdep_assert_held(&ci->i_ceph_lock); + + dout("removing capsnaps, ci is %p, inode is %p\n", ci, inode); + + while (!list_empty(&ci->i_cap_snaps)) { + capsnap = list_first_entry(&ci->i_cap_snaps, + struct ceph_cap_snap, ci_item); + __ceph_remove_capsnap(inode, capsnap, NULL, NULL); + ceph_put_snap_context(capsnap->context); + ceph_put_cap_snap(capsnap); + capsnap_release++; + } + wake_up_all(&ci->i_cap_wq); + wake_up_all(&mdsc->cap_flushing_wq); + return capsnap_release; +} + +int ceph_purge_inode_cap(struct inode *inode, struct ceph_cap *cap, bool *invalidate) +{ + struct ceph_fs_client *fsc = ceph_inode_to_client(inode); + struct ceph_mds_client *mdsc = fsc->mdsc; + struct ceph_inode_info *ci = ceph_inode(inode); + bool is_auth; + bool dirty_dropped = false; + int iputs = 0; + + lockdep_assert_held(&ci->i_ceph_lock); + + dout("removing cap %p, ci is %p, inode is %p\n", + cap, ci, &ci->netfs.inode); + + is_auth = (cap == ci->i_auth_cap); + __ceph_remove_cap(cap, false); + if (is_auth) { + struct ceph_cap_flush *cf; + + if (ceph_inode_is_shutdown(inode)) { + if (inode->i_data.nrpages > 0) + *invalidate = true; + if (ci->i_wrbuffer_ref > 0) + mapping_set_error(&inode->i_data, -EIO); + } + + spin_lock(&mdsc->cap_dirty_lock); + + /* trash all of the cap flushes for this inode */ + while (!list_empty(&ci->i_cap_flush_list)) { + cf = list_first_entry(&ci->i_cap_flush_list, + struct ceph_cap_flush, i_list); + list_del_init(&cf->g_list); + list_del_init(&cf->i_list); + if (!cf->is_capsnap) + ceph_free_cap_flush(cf); + } + + if (!list_empty(&ci->i_dirty_item)) { + pr_warn_ratelimited( + " dropping dirty %s state for %p %lld\n", + ceph_cap_string(ci->i_dirty_caps), + inode, ceph_ino(inode)); + ci->i_dirty_caps = 0; + list_del_init(&ci->i_dirty_item); + dirty_dropped = true; + } + if (!list_empty(&ci->i_flushing_item)) { + pr_warn_ratelimited( + " dropping dirty+flushing %s state for %p %lld\n", + ceph_cap_string(ci->i_flushing_caps), + inode, ceph_ino(inode)); + ci->i_flushing_caps = 0; + list_del_init(&ci->i_flushing_item); + mdsc->num_cap_flushing--; + dirty_dropped = true; + } + spin_unlock(&mdsc->cap_dirty_lock); + + if (dirty_dropped) { + mapping_set_error(inode->i_mapping, -EIO); + + if (ci->i_wrbuffer_ref_head == 0 && + ci->i_wr_ref == 0 && + ci->i_dirty_caps == 0 && + ci->i_flushing_caps == 0) { + ceph_put_snap_context(ci->i_head_snapc); + ci->i_head_snapc = NULL; + } + } + + if (atomic_read(&ci->i_filelock_ref) > 0) { + /* make further file lock syscall return -EIO */ + ci->i_ceph_flags |= CEPH_I_ERROR_FILELOCK; + pr_warn_ratelimited(" dropping file locks for %p %lld\n", + inode, ceph_ino(inode)); + } + + if (!ci->i_dirty_caps && ci->i_prealloc_cap_flush) { + cf = ci->i_prealloc_cap_flush; + ci->i_prealloc_cap_flush = NULL; + if (!cf->is_capsnap) + ceph_free_cap_flush(cf); + } + + if (!list_empty(&ci->i_cap_snaps)) + iputs = remove_capsnaps(mdsc, inode); + } + if (dirty_dropped) + ++iputs; + return iputs; +} diff --git a/fs/ceph/debugfs.c b/fs/ceph/debugfs.c index fb7cabd98e7b..bec3c4549c07 100644 --- a/fs/ceph/debugfs.c +++ b/fs/ceph/debugfs.c @@ -7,6 +7,8 @@ #include <linux/ctype.h> #include <linux/debugfs.h> #include <linux/seq_file.h> +#include <linux/math64.h> +#include <linux/ktime.h> #include <linux/ceph/libceph.h> #include <linux/ceph/mon_client.h> @@ -18,6 +20,7 @@ #ifdef CONFIG_DEBUG_FS #include "mds_client.h" +#include "metric.h" static int mdsmap_show(struct seq_file *s, void *p) { @@ -124,11 +127,133 @@ static int mdsc_show(struct seq_file *s, void *p) return 0; } +#define CEPH_LAT_METRIC_SHOW(name, total, avg, min, max, sq) { \ + s64 _total, _avg, _min, _max, _sq, _st; \ + _avg = ktime_to_us(avg); \ + _min = ktime_to_us(min == KTIME_MAX ? 0 : min); \ + _max = ktime_to_us(max); \ + _total = total - 1; \ + _sq = _total > 0 ? DIV64_U64_ROUND_CLOSEST(sq, _total) : 0; \ + _st = int_sqrt64(_sq); \ + _st = ktime_to_us(_st); \ + seq_printf(s, "%-14s%-12lld%-16lld%-16lld%-16lld%lld\n", \ + name, total, _avg, _min, _max, _st); \ +} + +#define CEPH_SZ_METRIC_SHOW(name, total, avg, min, max, sum) { \ + u64 _min = min == U64_MAX ? 0 : min; \ + seq_printf(s, "%-14s%-12lld%-16llu%-16llu%-16llu%llu\n", \ + name, total, avg, _min, max, sum); \ +} + +static int metrics_file_show(struct seq_file *s, void *p) +{ + struct ceph_fs_client *fsc = s->private; + struct ceph_client_metric *m = &fsc->mdsc->metric; + + seq_printf(s, "item total\n"); + seq_printf(s, "------------------------------------------\n"); + seq_printf(s, "%-35s%lld\n", "total inodes", + percpu_counter_sum(&m->total_inodes)); + seq_printf(s, "%-35s%lld\n", "opened files", + atomic64_read(&m->opened_files)); + seq_printf(s, "%-35s%lld\n", "pinned i_caps", + atomic64_read(&m->total_caps)); + seq_printf(s, "%-35s%lld\n", "opened inodes", + percpu_counter_sum(&m->opened_inodes)); + return 0; +} + +static const char * const metric_str[] = { + "read", + "write", + "metadata", + "copyfrom" +}; +static int metrics_latency_show(struct seq_file *s, void *p) +{ + struct ceph_fs_client *fsc = s->private; + struct ceph_client_metric *cm = &fsc->mdsc->metric; + struct ceph_metric *m; + s64 total, avg, min, max, sq; + int i; + + seq_printf(s, "item total avg_lat(us) min_lat(us) max_lat(us) stdev(us)\n"); + seq_printf(s, "-----------------------------------------------------------------------------------\n"); + + for (i = 0; i < METRIC_MAX; i++) { + m = &cm->metric[i]; + spin_lock(&m->lock); + total = m->total; + avg = m->latency_avg; + min = m->latency_min; + max = m->latency_max; + sq = m->latency_sq_sum; + spin_unlock(&m->lock); + CEPH_LAT_METRIC_SHOW(metric_str[i], total, avg, min, max, sq); + } + + return 0; +} + +static int metrics_size_show(struct seq_file *s, void *p) +{ + struct ceph_fs_client *fsc = s->private; + struct ceph_client_metric *cm = &fsc->mdsc->metric; + struct ceph_metric *m; + s64 total; + u64 sum, avg, min, max; + int i; + + seq_printf(s, "item total avg_sz(bytes) min_sz(bytes) max_sz(bytes) total_sz(bytes)\n"); + seq_printf(s, "----------------------------------------------------------------------------------------\n"); + + for (i = 0; i < METRIC_MAX; i++) { + /* skip 'metadata' as it doesn't use the size metric */ + if (i == METRIC_METADATA) + continue; + m = &cm->metric[i]; + spin_lock(&m->lock); + total = m->total; + sum = m->size_sum; + avg = total > 0 ? DIV64_U64_ROUND_CLOSEST(sum, total) : 0; + min = m->size_min; + max = m->size_max; + spin_unlock(&m->lock); + CEPH_SZ_METRIC_SHOW(metric_str[i], total, avg, min, max, sum); + } + + return 0; +} + +static int metrics_caps_show(struct seq_file *s, void *p) +{ + struct ceph_fs_client *fsc = s->private; + struct ceph_client_metric *m = &fsc->mdsc->metric; + int nr_caps = 0; + + seq_printf(s, "item total miss hit\n"); + seq_printf(s, "-------------------------------------------------\n"); + + seq_printf(s, "%-14s%-16lld%-16lld%lld\n", "d_lease", + atomic64_read(&m->total_dentries), + percpu_counter_sum(&m->d_lease_mis), + percpu_counter_sum(&m->d_lease_hit)); + + nr_caps = atomic64_read(&m->total_caps); + seq_printf(s, "%-14s%-16d%-16lld%lld\n", "caps", nr_caps, + percpu_counter_sum(&m->i_caps_mis), + percpu_counter_sum(&m->i_caps_hit)); + + return 0; +} + static int caps_show_cb(struct inode *inode, struct ceph_cap *cap, void *p) { struct seq_file *s = p; - seq_printf(s, "0x%-17lx%-17s%-17s\n", inode->i_ino, + seq_printf(s, "0x%-17llx%-3d%-17s%-17s\n", ceph_ino(inode), + cap->session->s_mds, ceph_cap_string(cap->issued), ceph_cap_string(cap->implemented)); return 0; @@ -148,8 +273,8 @@ static int caps_show(struct seq_file *s, void *p) "reserved\t%d\n" "min\t\t%d\n\n", total, avail, used, reserved, min); - seq_printf(s, "ino issued implemented\n"); - seq_printf(s, "-----------------------------------------------\n"); + seq_printf(s, "ino mds issued implemented\n"); + seq_printf(s, "--------------------------------------------------\n"); mutex_lock(&mdsc->mutex); for (i = 0; i < mdsc->max_sessions; i++) { @@ -173,7 +298,7 @@ static int caps_show(struct seq_file *s, void *p) spin_lock(&mdsc->caps_list_lock); list_for_each_entry(cw, &mdsc->cap_wait_list, list) { - seq_printf(s, "%-13d0x%-17lx%-17s%-17s\n", cw->tgid, cw->ino, + seq_printf(s, "%-13d0x%-17llx%-17s%-17s\n", cw->tgid, cw->ino, ceph_cap_string(cw->need), ceph_cap_string(cw->want)); } @@ -188,7 +313,7 @@ static int mds_sessions_show(struct seq_file *s, void *ptr) struct ceph_mds_client *mdsc = fsc->mdsc; struct ceph_auth_client *ac = fsc->client->monc.auth; struct ceph_options *opt = fsc->client->options; - int mds = -1; + int mds; mutex_lock(&mdsc->mutex); @@ -218,10 +343,28 @@ static int mds_sessions_show(struct seq_file *s, void *ptr) return 0; } -CEPH_DEFINE_SHOW_FUNC(mdsmap_show) -CEPH_DEFINE_SHOW_FUNC(mdsc_show) -CEPH_DEFINE_SHOW_FUNC(caps_show) -CEPH_DEFINE_SHOW_FUNC(mds_sessions_show) +static int status_show(struct seq_file *s, void *p) +{ + struct ceph_fs_client *fsc = s->private; + struct ceph_entity_inst *inst = &fsc->client->msgr.inst; + struct ceph_entity_addr *client_addr = ceph_client_addr(fsc->client); + + seq_printf(s, "instance: %s.%lld %s/%u\n", ENTITY_NAME(inst->name), + ceph_pr_addr(client_addr), le32_to_cpu(client_addr->nonce)); + seq_printf(s, "blocklisted: %s\n", fsc->blocklisted ? "true" : "false"); + + return 0; +} + +DEFINE_SHOW_ATTRIBUTE(mdsmap); +DEFINE_SHOW_ATTRIBUTE(mdsc); +DEFINE_SHOW_ATTRIBUTE(caps); +DEFINE_SHOW_ATTRIBUTE(mds_sessions); +DEFINE_SHOW_ATTRIBUTE(status); +DEFINE_SHOW_ATTRIBUTE(metrics_file); +DEFINE_SHOW_ATTRIBUTE(metrics_latency); +DEFINE_SHOW_ATTRIBUTE(metrics_size); +DEFINE_SHOW_ATTRIBUTE(metrics_caps); /* @@ -255,7 +398,9 @@ void ceph_fs_debugfs_cleanup(struct ceph_fs_client *fsc) debugfs_remove(fsc->debugfs_mdsmap); debugfs_remove(fsc->debugfs_mds_sessions); debugfs_remove(fsc->debugfs_caps); + debugfs_remove(fsc->debugfs_status); debugfs_remove(fsc->debugfs_mdsc); + debugfs_remove_recursive(fsc->debugfs_metrics_dir); } void ceph_fs_debugfs_init(struct ceph_fs_client *fsc) @@ -271,7 +416,7 @@ void ceph_fs_debugfs_init(struct ceph_fs_client *fsc) &congestion_kb_fops); snprintf(name, sizeof(name), "../../bdi/%s", - dev_name(fsc->sb->s_bdi->dev)); + bdi_dev_name(fsc->sb->s_bdi)); fsc->debugfs_bdi = debugfs_create_symlink("bdi", fsc->client->debugfs_dir, @@ -281,25 +426,43 @@ void ceph_fs_debugfs_init(struct ceph_fs_client *fsc) 0400, fsc->client->debugfs_dir, fsc, - &mdsmap_show_fops); + &mdsmap_fops); fsc->debugfs_mds_sessions = debugfs_create_file("mds_sessions", 0400, fsc->client->debugfs_dir, fsc, - &mds_sessions_show_fops); + &mds_sessions_fops); fsc->debugfs_mdsc = debugfs_create_file("mdsc", 0400, fsc->client->debugfs_dir, fsc, - &mdsc_show_fops); + &mdsc_fops); fsc->debugfs_caps = debugfs_create_file("caps", - 0400, - fsc->client->debugfs_dir, - fsc, - &caps_show_fops); + 0400, + fsc->client->debugfs_dir, + fsc, + &caps_fops); + + fsc->debugfs_status = debugfs_create_file("status", + 0400, + fsc->client->debugfs_dir, + fsc, + &status_fops); + + fsc->debugfs_metrics_dir = debugfs_create_dir("metrics", + fsc->client->debugfs_dir); + + debugfs_create_file("file", 0400, fsc->debugfs_metrics_dir, fsc, + &metrics_file_fops); + debugfs_create_file("latency", 0400, fsc->debugfs_metrics_dir, fsc, + &metrics_latency_fops); + debugfs_create_file("size", 0400, fsc->debugfs_metrics_dir, fsc, + &metrics_size_fops); + debugfs_create_file("caps", 0400, fsc->debugfs_metrics_dir, fsc, + &metrics_caps_fops); } diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c index d0cd0aba5843..e7e2ebac330d 100644 --- a/fs/ceph/dir.c +++ b/fs/ceph/dir.c @@ -38,6 +38,7 @@ static int __dir_lease_try_check(const struct dentry *dentry); static int ceph_d_init(struct dentry *dentry) { struct ceph_dentry_info *di; + struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(dentry->d_sb); di = kmem_cache_zalloc(ceph_dentry_cachep, GFP_KERNEL); if (!di) @@ -48,6 +49,9 @@ static int ceph_d_init(struct dentry *dentry) di->time = jiffies; dentry->d_fsdata = di; INIT_LIST_HEAD(&di->lease_list); + + atomic64_inc(&mdsc->metric.total_dentries); + return 0; } @@ -141,7 +145,7 @@ __dcache_find_get_entry(struct dentry *parent, u64 idx, return ERR_PTR(-EAGAIN); } /* reading/filling the cache are serialized by - i_mutex, no need to use page lock */ + i_rwsem, no need to use page lock */ unlock_page(cache_ctl->page); cache_ctl->dentries = kmap(cache_ctl->page); } @@ -151,7 +155,7 @@ __dcache_find_get_entry(struct dentry *parent, u64 idx, 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. */ + * marked as complete while not holding the i_rwsem. */ if (ceph_dir_is_complete_ordered(dir) && ptr_pos < i_size_read(dir)) dentry = cache_ctl->dentries[cache_ctl->index]; else @@ -254,9 +258,7 @@ static int __dcache_readdir(struct file *file, struct dir_context *ctx, dentry, dentry, d_inode(dentry)); ctx->pos = di->offset; if (!dir_emit(ctx, dentry->d_name.name, - dentry->d_name.len, - ceph_translate_ino(dentry->d_sb, - d_inode(dentry)->i_ino), + dentry->d_name.len, ceph_present_inode(d_inode(dentry)), d_inode(dentry)->i_mode >> 12)) { dput(dentry); err = 0; @@ -319,30 +321,37 @@ static int ceph_readdir(struct file *file, struct dir_context *ctx) /* always start with . and .. */ if (ctx->pos == 0) { dout("readdir off 0 -> '.'\n"); - if (!dir_emit(ctx, ".", 1, - ceph_translate_ino(inode->i_sb, inode->i_ino), + if (!dir_emit(ctx, ".", 1, ceph_present_inode(inode), inode->i_mode >> 12)) return 0; ctx->pos = 1; } if (ctx->pos == 1) { - ino_t ino = parent_ino(file->f_path.dentry); + u64 ino; + struct dentry *dentry = file->f_path.dentry; + + spin_lock(&dentry->d_lock); + ino = ceph_present_inode(dentry->d_parent->d_inode); + spin_unlock(&dentry->d_lock); + dout("readdir off 1 -> '..'\n"); - if (!dir_emit(ctx, "..", 2, - ceph_translate_ino(inode->i_sb, ino), - inode->i_mode >> 12)) + if (!dir_emit(ctx, "..", 2, ino, inode->i_mode >> 12)) return 0; ctx->pos = 2; } - /* can we use the dcache? */ spin_lock(&ci->i_ceph_lock); + /* request Fx cap. if have Fx, we don't need to release Fs cap + * for later create/unlink. */ + __ceph_touch_fmode(ci, mdsc, CEPH_FILE_MODE_WR); + /* can we use the dcache? */ if (ceph_test_mount_opt(fsc, DCACHE) && !ceph_test_mount_opt(fsc, NOASYNCREADDIR) && ceph_snap(inode) != CEPH_SNAPDIR && __ceph_dir_is_complete_ordered(ci) && - __ceph_caps_issued_mask(ci, CEPH_CAP_FILE_SHARED, 1)) { + __ceph_caps_issued_mask_metric(ci, CEPH_CAP_FILE_SHARED, 1)) { int shared_gen = atomic_read(&ci->i_shared_gen); + spin_unlock(&ci->i_ceph_lock); err = __dcache_readdir(file, ctx, shared_gen); if (err != -EAGAIN) @@ -469,8 +478,11 @@ more: 2 : (fpos_off(rde->offset) + 1); err = note_last_dentry(dfi, rde->name, rde->name_len, next_offset); - if (err) + if (err) { + ceph_mdsc_put_request(dfi->last_readdir); + dfi->last_readdir = NULL; return err; + } } else if (req->r_reply_info.dir_end) { dfi->next_offset = 2; /* keep last name */ @@ -498,9 +510,6 @@ more: } for (; i < rinfo->dir_nr; i++) { struct ceph_mds_reply_dir_entry *rde = rinfo->dir_entries + i; - struct ceph_vino vino; - ino_t ino; - u32 ftype; BUG_ON(rde->offset < ctx->pos); @@ -510,13 +519,16 @@ more: 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)) { + ceph_present_ino(inode->i_sb, le64_to_cpu(rde->inode.in->ino)), + le32_to_cpu(rde->inode.in->mode) >> 12)) { + /* + * NOTE: Here no need to put the 'dfi->last_readdir', + * because when dir_emit stops us it's most likely + * doesn't have enough memory, etc. So for next readdir + * it will continue. + */ dout("filldir stopping us...\n"); return 0; } @@ -628,10 +640,12 @@ static loff_t ceph_dir_llseek(struct file *file, loff_t offset, int whence) switch (whence) { case SEEK_CUR: offset += file->f_pos; + break; case SEEK_SET: break; case SEEK_END: retval = -EOPNOTSUPP; + goto out; default: goto out; } @@ -662,25 +676,25 @@ out: /* * Handle lookups for the hidden .snap directory. */ -int ceph_handle_snapdir(struct ceph_mds_request *req, - struct dentry *dentry, int err) +struct dentry *ceph_handle_snapdir(struct ceph_mds_request *req, + struct dentry *dentry) { struct ceph_fs_client *fsc = ceph_sb_to_client(dentry->d_sb); - struct inode *parent = d_inode(dentry->d_parent); /* we hold i_mutex */ + struct inode *parent = d_inode(dentry->d_parent); /* we hold i_rwsem */ /* .snap dir? */ - if (err == -ENOENT && - ceph_snap(parent) == CEPH_NOSNAP && - strcmp(dentry->d_name.name, - fsc->mount_options->snapdir_name) == 0) { + if (ceph_snap(parent) == CEPH_NOSNAP && + strcmp(dentry->d_name.name, fsc->mount_options->snapdir_name) == 0) { + struct dentry *res; struct inode *inode = ceph_get_snapdir(parent); - dout("ENOENT on snapdir %p '%pd', linking to snapdir %p\n", - dentry, dentry, inode); - BUG_ON(!d_unhashed(dentry)); - d_add(dentry, inode); - err = 0; + + res = d_splice_alias(inode, dentry); + dout("ENOENT on snapdir %p '%pd', linking to snapdir %p. Spliced dentry %p\n", + dentry, dentry, inode, res); + if (res) + dentry = res; } - return err; + return dentry; } /* @@ -734,7 +748,7 @@ static struct dentry *ceph_lookup(struct inode *dir, struct dentry *dentry, unsigned int flags) { struct ceph_fs_client *fsc = ceph_sb_to_client(dir->i_sb); - struct ceph_mds_client *mdsc = fsc->mdsc; + struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(dir->i_sb); struct ceph_mds_request *req; int op; int mask; @@ -752,14 +766,15 @@ static struct dentry *ceph_lookup(struct inode *dir, struct dentry *dentry, struct ceph_dentry_info *di = ceph_dentry(dentry); spin_lock(&ci->i_ceph_lock); - dout(" dir %p flags are %d\n", dir, ci->i_ceph_flags); + dout(" dir %p flags are 0x%lx\n", dir, ci->i_ceph_flags); if (strncmp(dentry->d_name.name, fsc->mount_options->snapdir_name, dentry->d_name.len) && !is_root_ceph_dentry(dir, dentry) && ceph_test_mount_opt(fsc, DCACHE) && __ceph_dir_is_complete(ci) && - (__ceph_caps_issued_mask(ci, CEPH_CAP_FILE_SHARED, 1))) { + __ceph_caps_issued_mask_metric(ci, CEPH_CAP_FILE_SHARED, 1)) { + __ceph_touch_fmode(ci, mdsc, CEPH_FILE_MODE_RD); spin_unlock(&ci->i_ceph_lock); dout(" dir %p complete, -ENOENT\n", dir); d_add(dentry, NULL); @@ -782,10 +797,21 @@ static struct dentry *ceph_lookup(struct inode *dir, struct dentry *dentry, mask |= CEPH_CAP_XATTR_SHARED; req->r_args.getattr.mask = cpu_to_le32(mask); + ihold(dir); req->r_parent = dir; set_bit(CEPH_MDS_R_PARENT_LOCKED, &req->r_req_flags); err = ceph_mdsc_do_request(mdsc, NULL, req); - err = ceph_handle_snapdir(req, dentry, err); + if (err == -ENOENT) { + struct dentry *res; + + res = ceph_handle_snapdir(req, dentry); + if (IS_ERR(res)) { + err = PTR_ERR(res); + } else { + dentry = res; + err = 0; + } + } dentry = ceph_finish_lookup(req, dentry, err); ceph_mdsc_put_request(req); /* will dput(dentry) */ dout("lookup result=%p\n", dentry); @@ -819,11 +845,10 @@ int ceph_handle_notrace_create(struct inode *dir, struct dentry *dentry) return PTR_ERR(result); } -static int ceph_mknod(struct inode *dir, struct dentry *dentry, - umode_t mode, dev_t rdev) +static int ceph_mknod(struct user_namespace *mnt_userns, struct inode *dir, + struct dentry *dentry, umode_t mode, dev_t rdev) { - struct ceph_fs_client *fsc = ceph_sb_to_client(dir->i_sb); - struct ceph_mds_client *mdsc = fsc->mdsc; + struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(dir->i_sb); struct ceph_mds_request *req; struct ceph_acl_sec_ctx as_ctx = {}; int err; @@ -831,6 +856,10 @@ static int ceph_mknod(struct inode *dir, struct dentry *dentry, if (ceph_snap(dir) != CEPH_NOSNAP) return -EROFS; + err = ceph_wait_on_conflict_unlink(dentry); + if (err) + return err; + if (ceph_quota_is_max_files_exceeded(dir)) { err = -EDQUOT; goto out; @@ -853,6 +882,7 @@ static int ceph_mknod(struct inode *dir, struct dentry *dentry, req->r_dentry = dget(dentry); req->r_num_caps = 2; req->r_parent = dir; + ihold(dir); set_bit(CEPH_MDS_R_PARENT_LOCKED, &req->r_req_flags); req->r_args.mknod.mode = cpu_to_le32(mode); req->r_args.mknod.rdev = cpu_to_le32(rdev); @@ -875,17 +905,16 @@ out: return err; } -static int ceph_create(struct inode *dir, struct dentry *dentry, umode_t mode, - bool excl) +static int ceph_create(struct user_namespace *mnt_userns, struct inode *dir, + struct dentry *dentry, umode_t mode, bool excl) { - return ceph_mknod(dir, dentry, mode, 0); + return ceph_mknod(mnt_userns, dir, dentry, mode, 0); } -static int ceph_symlink(struct inode *dir, struct dentry *dentry, - const char *dest) +static int ceph_symlink(struct user_namespace *mnt_userns, struct inode *dir, + struct dentry *dentry, const char *dest) { - struct ceph_fs_client *fsc = ceph_sb_to_client(dir->i_sb); - struct ceph_mds_client *mdsc = fsc->mdsc; + struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(dir->i_sb); struct ceph_mds_request *req; struct ceph_acl_sec_ctx as_ctx = {}; int err; @@ -893,6 +922,10 @@ static int ceph_symlink(struct inode *dir, struct dentry *dentry, if (ceph_snap(dir) != CEPH_NOSNAP) return -EROFS; + err = ceph_wait_on_conflict_unlink(dentry); + if (err) + return err; + if (ceph_quota_is_max_files_exceeded(dir)) { err = -EDQUOT; goto out; @@ -915,11 +948,17 @@ static int ceph_symlink(struct inode *dir, struct dentry *dentry, goto out; } req->r_parent = dir; + ihold(dir); + set_bit(CEPH_MDS_R_PARENT_LOCKED, &req->r_req_flags); req->r_dentry = dget(dentry); req->r_num_caps = 2; req->r_dentry_drop = CEPH_CAP_FILE_SHARED | CEPH_CAP_AUTH_EXCL; req->r_dentry_unless = CEPH_CAP_FILE_EXCL; + if (as_ctx.pagelist) { + req->r_pagelist = as_ctx.pagelist; + as_ctx.pagelist = NULL; + } err = ceph_mdsc_do_request(mdsc, dir, req); if (!err && !req->r_reply_info.head->is_dentry) err = ceph_handle_notrace_create(dir, dentry); @@ -931,15 +970,19 @@ out: return err; } -static int ceph_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode) +static int ceph_mkdir(struct user_namespace *mnt_userns, struct inode *dir, + struct dentry *dentry, umode_t mode) { - struct ceph_fs_client *fsc = ceph_sb_to_client(dir->i_sb); - struct ceph_mds_client *mdsc = fsc->mdsc; + struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(dir->i_sb); struct ceph_mds_request *req; struct ceph_acl_sec_ctx as_ctx = {}; - int err = -EROFS; + int err; int op; + err = ceph_wait_on_conflict_unlink(dentry); + if (err) + return err; + if (ceph_snap(dir) == CEPH_SNAPDIR) { /* mkdir .snap/foo is a MKSNAP */ op = CEPH_MDS_OP_MKSNAP; @@ -949,6 +992,7 @@ static int ceph_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode) dout("mkdir dir %p dn %p mode 0%ho\n", dir, dentry, mode); op = CEPH_MDS_OP_MKDIR; } else { + err = -EROFS; goto out; } @@ -975,6 +1019,7 @@ static int ceph_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode) req->r_dentry = dget(dentry); req->r_num_caps = 2; req->r_parent = dir; + ihold(dir); set_bit(CEPH_MDS_R_PARENT_LOCKED, &req->r_req_flags); req->r_args.mkdir.mode = cpu_to_le32(mode); req->r_dentry_drop = CEPH_CAP_FILE_SHARED | CEPH_CAP_AUTH_EXCL; @@ -1001,11 +1046,14 @@ out: static int ceph_link(struct dentry *old_dentry, struct inode *dir, struct dentry *dentry) { - struct ceph_fs_client *fsc = ceph_sb_to_client(dir->i_sb); - struct ceph_mds_client *mdsc = fsc->mdsc; + struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(dir->i_sb); struct ceph_mds_request *req; int err; + err = ceph_wait_on_conflict_unlink(dentry); + if (err) + return err; + if (ceph_snap(dir) != CEPH_NOSNAP) return -EROFS; @@ -1020,6 +1068,7 @@ static int ceph_link(struct dentry *old_dentry, struct inode *dir, req->r_num_caps = 2; req->r_old_dentry = dget(old_dentry); req->r_parent = dir; + ihold(dir); set_bit(CEPH_MDS_R_PARENT_LOCKED, &req->r_req_flags); req->r_dentry_drop = CEPH_CAP_FILE_SHARED; req->r_dentry_unless = CEPH_CAP_FILE_EXCL; @@ -1036,6 +1085,96 @@ static int ceph_link(struct dentry *old_dentry, struct inode *dir, return err; } +static void ceph_async_unlink_cb(struct ceph_mds_client *mdsc, + struct ceph_mds_request *req) +{ + struct dentry *dentry = req->r_dentry; + struct ceph_fs_client *fsc = ceph_sb_to_client(dentry->d_sb); + struct ceph_dentry_info *di = ceph_dentry(dentry); + int result = req->r_err ? req->r_err : + le32_to_cpu(req->r_reply_info.head->result); + + if (!test_bit(CEPH_DENTRY_ASYNC_UNLINK_BIT, &di->flags)) + pr_warn("%s dentry %p:%pd async unlink bit is not set\n", + __func__, dentry, dentry); + + spin_lock(&fsc->async_unlink_conflict_lock); + hash_del_rcu(&di->hnode); + spin_unlock(&fsc->async_unlink_conflict_lock); + + spin_lock(&dentry->d_lock); + di->flags &= ~CEPH_DENTRY_ASYNC_UNLINK; + wake_up_bit(&di->flags, CEPH_DENTRY_ASYNC_UNLINK_BIT); + spin_unlock(&dentry->d_lock); + + synchronize_rcu(); + + if (result == -EJUKEBOX) + goto out; + + /* If op failed, mark everyone involved for errors */ + if (result) { + int pathlen = 0; + u64 base = 0; + char *path = ceph_mdsc_build_path(dentry, &pathlen, + &base, 0); + + /* mark error on parent + clear complete */ + mapping_set_error(req->r_parent->i_mapping, result); + ceph_dir_clear_complete(req->r_parent); + + /* drop the dentry -- we don't know its status */ + if (!d_unhashed(dentry)) + d_drop(dentry); + + /* mark inode itself for an error (since metadata is bogus) */ + mapping_set_error(req->r_old_inode->i_mapping, result); + + pr_warn("async unlink failure path=(%llx)%s result=%d!\n", + base, IS_ERR(path) ? "<<bad>>" : path, result); + ceph_mdsc_free_path(path, pathlen); + } +out: + iput(req->r_old_inode); + ceph_mdsc_release_dir_caps(req); +} + +static int get_caps_for_async_unlink(struct inode *dir, struct dentry *dentry) +{ + struct ceph_inode_info *ci = ceph_inode(dir); + struct ceph_dentry_info *di; + int got = 0, want = CEPH_CAP_FILE_EXCL | CEPH_CAP_DIR_UNLINK; + + spin_lock(&ci->i_ceph_lock); + if ((__ceph_caps_issued(ci, NULL) & want) == want) { + ceph_take_cap_refs(ci, want, false); + got = want; + } + spin_unlock(&ci->i_ceph_lock); + + /* If we didn't get anything, return 0 */ + if (!got) + return 0; + + spin_lock(&dentry->d_lock); + di = ceph_dentry(dentry); + /* + * - We are holding Fx, which implies Fs caps. + * - Only support async unlink for primary linkage + */ + if (atomic_read(&ci->i_shared_gen) != di->lease_shared_gen || + !(di->flags & CEPH_DENTRY_PRIMARY_LINK)) + want = 0; + spin_unlock(&dentry->d_lock); + + /* Do we still want what we've got? */ + if (want == got) + return got; + + ceph_put_cap_refs(ci, got); + return 0; +} + /* * rmdir and unlink are differ only by the metadata op code */ @@ -1045,6 +1184,7 @@ static int ceph_unlink(struct inode *dir, struct dentry *dentry) struct ceph_mds_client *mdsc = fsc->mdsc; struct inode *inode = d_inode(dentry); struct ceph_mds_request *req; + bool try_async = ceph_test_mount_opt(fsc, ASYNC_DIROPS); int err = -EROFS; int op; @@ -1059,6 +1199,7 @@ static int ceph_unlink(struct inode *dir, struct dentry *dentry) CEPH_MDS_OP_RMDIR : CEPH_MDS_OP_UNLINK; } else goto out; +retry: req = ceph_mdsc_create_request(mdsc, op, USE_AUTH_MDS); if (IS_ERR(req)) { err = PTR_ERR(req); @@ -1067,24 +1208,72 @@ static int ceph_unlink(struct inode *dir, struct dentry *dentry) req->r_dentry = dget(dentry); req->r_num_caps = 2; req->r_parent = dir; - set_bit(CEPH_MDS_R_PARENT_LOCKED, &req->r_req_flags); + ihold(dir); req->r_dentry_drop = CEPH_CAP_FILE_SHARED; req->r_dentry_unless = CEPH_CAP_FILE_EXCL; req->r_inode_drop = ceph_drop_caps_for_unlink(inode); - err = ceph_mdsc_do_request(mdsc, dir, req); - if (!err && !req->r_reply_info.head->is_dentry) - d_delete(dentry); + + if (try_async && op == CEPH_MDS_OP_UNLINK && + (req->r_dir_caps = get_caps_for_async_unlink(dir, dentry))) { + struct ceph_dentry_info *di = ceph_dentry(dentry); + + dout("async unlink on %llu/%.*s caps=%s", ceph_ino(dir), + dentry->d_name.len, dentry->d_name.name, + ceph_cap_string(req->r_dir_caps)); + set_bit(CEPH_MDS_R_ASYNC, &req->r_req_flags); + req->r_callback = ceph_async_unlink_cb; + req->r_old_inode = d_inode(dentry); + ihold(req->r_old_inode); + + spin_lock(&dentry->d_lock); + di->flags |= CEPH_DENTRY_ASYNC_UNLINK; + spin_unlock(&dentry->d_lock); + + spin_lock(&fsc->async_unlink_conflict_lock); + hash_add_rcu(fsc->async_unlink_conflict, &di->hnode, + dentry->d_name.hash); + spin_unlock(&fsc->async_unlink_conflict_lock); + + err = ceph_mdsc_submit_request(mdsc, dir, req); + if (!err) { + /* + * We have enough caps, so we assume that the unlink + * will succeed. Fix up the target inode and dcache. + */ + drop_nlink(inode); + d_delete(dentry); + } else { + spin_lock(&fsc->async_unlink_conflict_lock); + hash_del_rcu(&di->hnode); + spin_unlock(&fsc->async_unlink_conflict_lock); + + spin_lock(&dentry->d_lock); + di->flags &= ~CEPH_DENTRY_ASYNC_UNLINK; + spin_unlock(&dentry->d_lock); + + if (err == -EJUKEBOX) { + try_async = false; + ceph_mdsc_put_request(req); + goto retry; + } + } + } else { + set_bit(CEPH_MDS_R_PARENT_LOCKED, &req->r_req_flags); + err = ceph_mdsc_do_request(mdsc, dir, req); + if (!err && !req->r_reply_info.head->is_dentry) + d_delete(dentry); + } + ceph_mdsc_put_request(req); out: return err; } -static int ceph_rename(struct inode *old_dir, struct dentry *old_dentry, - struct inode *new_dir, struct dentry *new_dentry, - unsigned int flags) +static int ceph_rename(struct user_namespace *mnt_userns, struct inode *old_dir, + struct dentry *old_dentry, struct inode *new_dir, + struct dentry *new_dentry, unsigned int flags) { - struct ceph_fs_client *fsc = ceph_sb_to_client(old_dir->i_sb); - struct ceph_mds_client *mdsc = fsc->mdsc; + struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(old_dir->i_sb); struct ceph_mds_request *req; int op = CEPH_MDS_OP_RENAME; int err; @@ -1105,6 +1294,10 @@ static int ceph_rename(struct inode *old_dir, struct dentry *old_dentry, (!ceph_quota_is_same_realm(old_dir, new_dir))) return -EXDEV; + err = ceph_wait_on_conflict_unlink(new_dentry); + if (err) + return err; + dout("rename dir %p dentry %p to dir %p dentry %p\n", old_dir, old_dentry, new_dir, new_dentry); req = ceph_mdsc_create_request(mdsc, op, USE_AUTH_MDS); @@ -1116,6 +1309,7 @@ static int ceph_rename(struct inode *old_dir, struct dentry *old_dentry, req->r_old_dentry = dget(old_dentry); req->r_old_dentry_dir = old_dir; req->r_parent = new_dir; + ihold(new_dir); set_bit(CEPH_MDS_R_PARENT_LOCKED, &req->r_req_flags); req->r_old_dentry_drop = CEPH_CAP_FILE_SHARED; req->r_old_dentry_unless = CEPH_CAP_FILE_EXCL; @@ -1411,6 +1605,7 @@ void ceph_invalidate_dentry_lease(struct dentry *dentry) spin_lock(&dentry->d_lock); di->time = jiffies; di->lease_shared_gen = 0; + di->flags &= ~CEPH_DENTRY_PRIMARY_LINK; __dentry_lease_unlist(di); spin_unlock(&dentry->d_lock); } @@ -1431,10 +1626,8 @@ static bool __dentry_lease_is_valid(struct ceph_dentry_info *di) u32 gen; unsigned long ttl; - spin_lock(&session->s_gen_ttl_lock); - gen = session->s_cap_gen; + gen = atomic_read(&session->s_cap_gen); ttl = session->s_cap_ttl; - spin_unlock(&session->s_gen_ttl_lock); if (di->lease_gen == gen && time_before(jiffies, ttl) && @@ -1520,7 +1713,8 @@ static int __dir_lease_try_check(const struct dentry *dentry) /* * Check if directory-wide content lease/cap is valid. */ -static int dir_lease_is_valid(struct inode *dir, struct dentry *dentry) +static int dir_lease_is_valid(struct inode *dir, struct dentry *dentry, + struct ceph_mds_client *mdsc) { struct ceph_inode_info *ci = ceph_inode(dir); int valid; @@ -1528,7 +1722,10 @@ static int dir_lease_is_valid(struct inode *dir, struct dentry *dentry) spin_lock(&ci->i_ceph_lock); valid = __ceph_caps_issued_mask(ci, CEPH_CAP_FILE_SHARED, 1); - shared_gen = atomic_read(&ci->i_shared_gen); + if (valid) { + __ceph_touch_fmode(ci, mdsc, CEPH_FILE_MODE_RD); + shared_gen = atomic_read(&ci->i_shared_gen); + } spin_unlock(&ci->i_ceph_lock); if (valid) { struct ceph_dentry_info *di; @@ -1554,6 +1751,7 @@ static int ceph_d_revalidate(struct dentry *dentry, unsigned int flags) int valid = 0; struct dentry *parent; struct inode *dir, *inode; + struct ceph_mds_client *mdsc; if (flags & LOOKUP_RCU) { parent = READ_ONCE(dentry->d_parent); @@ -1570,6 +1768,8 @@ static int ceph_d_revalidate(struct dentry *dentry, unsigned int flags) dout("d_revalidate %p '%pd' inode %p offset 0x%llx\n", dentry, dentry, inode, ceph_dentry(dentry)->offset); + mdsc = ceph_sb_to_client(dir->i_sb)->mdsc; + /* always trust cached snapped dentries, snapdir dentry */ if (ceph_snap(dir) != CEPH_NOSNAP) { dout("d_revalidate %p '%pd' inode %p is SNAPPED\n", dentry, @@ -1581,7 +1781,7 @@ static int ceph_d_revalidate(struct dentry *dentry, unsigned int flags) valid = dentry_lease_is_valid(dentry, flags); if (valid == -ECHILD) return valid; - if (valid || dir_lease_is_valid(dir, dentry)) { + if (valid || dir_lease_is_valid(dir, dentry, mdsc)) { if (inode) valid = ceph_is_any_caps(inode); else @@ -1590,8 +1790,6 @@ static int ceph_d_revalidate(struct dentry *dentry, unsigned int flags) } if (!valid) { - struct ceph_mds_client *mdsc = - ceph_sb_to_client(dir->i_sb)->mdsc; struct ceph_mds_request *req; int op, err; u32 mask; @@ -1599,6 +1797,8 @@ static int ceph_d_revalidate(struct dentry *dentry, unsigned int flags) if (flags & LOOKUP_RCU) return -ECHILD; + percpu_counter_inc(&mdsc->metric.d_lease_mis); + op = ceph_snap(dir) == CEPH_SNAPDIR ? CEPH_MDS_OP_LOOKUPSNAP : CEPH_MDS_OP_LOOKUP; req = ceph_mdsc_create_request(mdsc, op, USE_ANY_MDS); @@ -1606,6 +1806,7 @@ static int ceph_d_revalidate(struct dentry *dentry, unsigned int flags) req->r_dentry = dget(dentry); req->r_num_caps = 2; req->r_parent = dir; + ihold(dir); mask = CEPH_STAT_CAP_INODE | CEPH_CAP_AUTH_SHARED; if (ceph_security_xattr_wanted(dir)) @@ -1622,7 +1823,7 @@ static int ceph_d_revalidate(struct dentry *dentry, unsigned int flags) case -ENOENT: if (d_really_is_negative(dentry)) valid = 1; - /* Fallthrough */ + fallthrough; default: break; } @@ -1630,6 +1831,8 @@ static int ceph_d_revalidate(struct dentry *dentry, unsigned int flags) dout("d_revalidate %p lookup result=%d\n", dentry, err); } + } else { + percpu_counter_inc(&mdsc->metric.d_lease_hit); } dout("d_revalidate %p %s\n", dentry, valid ? "valid" : "invalid"); @@ -1672,16 +1875,18 @@ static int ceph_d_delete(const struct dentry *dentry) static void ceph_d_release(struct dentry *dentry) { struct ceph_dentry_info *di = ceph_dentry(dentry); + struct ceph_fs_client *fsc = ceph_sb_to_client(dentry->d_sb); dout("d_release %p\n", dentry); + atomic64_dec(&fsc->mdsc->metric.total_dentries); + spin_lock(&dentry->d_lock); __dentry_lease_unlist(di); dentry->d_fsdata = NULL; spin_unlock(&dentry->d_lock); - if (di->lease_session) - ceph_put_mds_session(di->lease_session); + ceph_put_mds_session(di->lease_session); kmem_cache_free(ceph_dentry_cachep, di); } diff --git a/fs/ceph/export.c b/fs/ceph/export.c index b6bfa94332c3..f780e4e0d062 100644 --- a/fs/ceph/export.c +++ b/fs/ceph/export.c @@ -129,6 +129,10 @@ static struct inode *__lookup_inode(struct super_block *sb, u64 ino) vino.ino = ino; vino.snap = CEPH_NOSNAP; + + if (ceph_vino_is_reserved(vino)) + return ERR_PTR(-ESTALE); + inode = ceph_find_inode(sb, vino); if (!inode) { struct ceph_mds_request *req; @@ -153,6 +157,11 @@ static struct inode *__lookup_inode(struct super_block *sb, u64 ino) ceph_mdsc_put_request(req); if (!inode) return err < 0 ? ERR_PTR(err) : ERR_PTR(-ESTALE); + } else { + if (ceph_inode_is_shutdown(inode)) { + iput(inode); + return ERR_PTR(-ESTALE); + } } return inode; } @@ -172,9 +181,19 @@ struct inode *ceph_lookup_inode(struct super_block *sb, u64 ino) static struct dentry *__fh_to_dentry(struct super_block *sb, u64 ino) { struct inode *inode = __lookup_inode(sb, ino); + struct ceph_inode_info *ci = ceph_inode(inode); + int err; + if (IS_ERR(inode)) return ERR_CAST(inode); - if (inode->i_nlink == 0) { + /* We need LINK caps to reliably check i_nlink */ + err = ceph_do_getattr(inode, CEPH_CAP_LINK_SHARED, false); + if (err) { + iput(inode); + return ERR_PTR(err); + } + /* -ESTALE if inode as been unlinked and no file is open */ + if ((inode->i_nlink == 0) && !__ceph_is_file_opened(ci)) { iput(inode); return ERR_PTR(-ESTALE); } @@ -205,9 +224,18 @@ static struct dentry *__snapfh_to_dentry(struct super_block *sb, vino.ino = sfh->ino; vino.snap = sfh->snapid; } + + if (ceph_vino_is_reserved(vino)) + return ERR_PTR(-ESTALE); + inode = ceph_find_inode(sb, vino); - if (inode) + if (inode) { + if (ceph_inode_is_shutdown(inode)) { + iput(inode); + return ERR_PTR(-ESTALE); + } return d_obtain_alias(inode); + } req = ceph_mdsc_create_request(mdsc, CEPH_MDS_OP_LOOKUPINO, USE_ANY_MDS); @@ -241,9 +269,10 @@ static struct dentry *__snapfh_to_dentry(struct super_block *sb, ihold(inode); } else { /* mds does not support lookup snapped inode */ - err = -EOPNOTSUPP; - inode = NULL; + inode = ERR_PTR(-EOPNOTSUPP); } + } else { + inode = ERR_PTR(-ESTALE); } ceph_mdsc_put_request(req); @@ -254,8 +283,8 @@ static struct dentry *__snapfh_to_dentry(struct super_block *sb, dout("snapfh_to_dentry %llx.%llx parent %llx hash %x err=%d", vino.ino, vino.snap, sfh->parent_ino, sfh->hash, err); } - if (!inode) - return ERR_PTR(-ESTALE); + if (IS_ERR(inode)) + return ERR_CAST(inode); /* see comments in ceph_get_parent() */ return unlinked ? d_obtain_root(inode) : d_obtain_alias(inode); } @@ -315,6 +344,11 @@ static struct dentry *__get_parent(struct super_block *sb, req->r_num_caps = 1; err = ceph_mdsc_do_request(mdsc, NULL, req); + if (err) { + ceph_mdsc_put_request(req); + return ERR_PTR(err); + } + inode = req->r_target_inode; if (inode) ihold(inode); @@ -519,6 +553,7 @@ static int ceph_get_name(struct dentry *parent, char *name, ihold(inode); req->r_ino2 = ceph_vino(d_inode(parent)); req->r_parent = d_inode(parent); + ihold(req->r_parent); set_bit(CEPH_MDS_R_PARENT_LOCKED, &req->r_req_flags); req->r_num_caps = 2; err = ceph_mdsc_do_request(mdsc, NULL, req); diff --git a/fs/ceph/file.c b/fs/ceph/file.c index 7e0190b1f821..04fd34557de8 100644 --- a/fs/ceph/file.c +++ b/fs/ceph/file.c @@ -11,11 +11,13 @@ #include <linux/writeback.h> #include <linux/falloc.h> #include <linux/iversion.h> +#include <linux/ktime.h> #include "super.h" #include "mds_client.h" #include "cache.h" #include "io.h" +#include "metric.h" static __le32 ceph_flags_sys2wire(u32 flags) { @@ -93,12 +95,11 @@ static ssize_t __iter_get_bvecs(struct iov_iter *iter, size_t maxsize, size_t start; int idx = 0; - bytes = iov_iter_get_pages(iter, pages, maxsize - size, + bytes = iov_iter_get_pages2(iter, pages, maxsize - size, ITER_GET_BVECS_PAGES, &start); if (bytes < 0) return size ?: bytes; - iov_iter_advance(iter, bytes); size += bytes; for ( ; bytes; idx++, bvec_idx++) { @@ -180,8 +181,7 @@ static void put_bvecs(struct bio_vec *bvecs, int num_bvecs, bool should_dirty) static struct ceph_mds_request * prepare_open_request(struct super_block *sb, int flags, int create_mode) { - struct ceph_fs_client *fsc = ceph_sb_to_client(sb); - struct ceph_mds_client *mdsc = fsc->mdsc; + struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(sb); struct ceph_mds_request *req; int want_auth = USE_ANY_MDS; int op = (flags & O_CREAT) ? CEPH_MDS_OP_CREATE : CEPH_MDS_OP_OPEN; @@ -203,7 +203,10 @@ static int ceph_init_file_info(struct inode *inode, struct file *file, int fmode, bool isdir) { struct ceph_inode_info *ci = ceph_inode(inode); + struct ceph_mount_options *opt = + ceph_inode_to_client(&ci->netfs.inode)->mount_options; struct ceph_file_info *fi; + int ret; dout("%s %p %p 0%o (%s)\n", __func__, inode, file, inode->i_mode, isdir ? "dir" : "regular"); @@ -212,10 +215,8 @@ static int ceph_init_file_info(struct inode *inode, struct file *file, if (isdir) { struct ceph_dir_file_info *dfi = kmem_cache_zalloc(ceph_dir_file_cachep, GFP_KERNEL); - if (!dfi) { - ceph_put_fmode(ci, fmode); /* clean up */ + if (!dfi) return -ENOMEM; - } file->private_data = dfi; fi = &dfi->file_info; @@ -223,21 +224,37 @@ static int ceph_init_file_info(struct inode *inode, struct file *file, dfi->readdir_cache_idx = -1; } else { fi = kmem_cache_zalloc(ceph_file_cachep, GFP_KERNEL); - if (!fi) { - ceph_put_fmode(ci, fmode); /* clean up */ + if (!fi) return -ENOMEM; - } + + if (opt->flags & CEPH_MOUNT_OPT_NOPAGECACHE) + fi->flags |= CEPH_F_SYNC; file->private_data = fi; } + ceph_get_fmode(ci, fmode, 1); fi->fmode = fmode; + spin_lock_init(&fi->rw_contexts_lock); INIT_LIST_HEAD(&fi->rw_contexts); - fi->meta_err = errseq_sample(&ci->i_meta_err); fi->filp_gen = READ_ONCE(ceph_inode_to_client(inode)->filp_gen); + if ((file->f_mode & FMODE_WRITE) && ceph_has_inline_data(ci)) { + ret = ceph_uninline_data(file); + if (ret < 0) + goto error; + } + return 0; + +error: + ceph_fscache_unuse_cookie(inode, file->f_mode & FMODE_WRITE); + ceph_put_fmode(ci, fi->fmode, 1); + kmem_cache_free(ceph_file_cachep, fi); + /* wake up anyone waiting for caps on this inode */ + wake_up_all(&ci->i_cap_wq); + return ret; } /* @@ -250,20 +267,16 @@ static int ceph_init_file(struct inode *inode, struct file *file, int fmode) switch (inode->i_mode & S_IFMT) { case S_IFREG: - ceph_fscache_register_inode_cookie(inode); - ceph_fscache_file_set_cookie(inode, file); - /* fall through */ + ceph_fscache_use_cookie(inode, file->f_mode & FMODE_WRITE); + fallthrough; case S_IFDIR: ret = ceph_init_file_info(inode, file, fmode, S_ISDIR(inode->i_mode)); - if (ret) - return ret; break; case S_IFLNK: dout("init_file %p %p 0%o (symlink)\n", inode, file, inode->i_mode); - ceph_put_fmode(ceph_inode(inode), fmode); /* clean up */ break; default: @@ -273,7 +286,6 @@ static int ceph_init_file(struct inode *inode, struct file *file, int fmode) * we need to drop the open ref now, since we don't * have .release set to ceph_release. */ - ceph_put_fmode(ceph_inode(inode), fmode); /* clean up */ BUG_ON(inode->i_fop->release == ceph_release); /* call the proper open fop */ @@ -285,14 +297,15 @@ static int ceph_init_file(struct inode *inode, struct file *file, int fmode) /* * try renew caps after session gets killed. */ -int ceph_renew_caps(struct inode *inode) +int ceph_renew_caps(struct inode *inode, int fmode) { - struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; + struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(inode->i_sb); struct ceph_inode_info *ci = ceph_inode(inode); struct ceph_mds_request *req; int err, flags, wanted; spin_lock(&ci->i_ceph_lock); + __ceph_touch_fmode(ci, mdsc, fmode); wanted = __ceph_caps_file_wanted(ci); if (__ceph_is_any_real_caps(ci) && (!(wanted & CEPH_CAP_ANY_WR) || ci->i_auth_cap)) { @@ -326,7 +339,6 @@ int ceph_renew_caps(struct inode *inode) 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); @@ -372,9 +384,6 @@ int ceph_open(struct inode *inode, struct file *file) /* trivially open snapdir */ if (ceph_snap(inode) == CEPH_SNAPDIR) { - spin_lock(&ci->i_ceph_lock); - __ceph_get_fmode(ci, fmode); - spin_unlock(&ci->i_ceph_lock); return ceph_init_file(inode, file, fmode); } @@ -392,7 +401,7 @@ int ceph_open(struct inode *inode, struct file *file) dout("open %p fmode %d want %s issued %s using existing\n", inode, fmode, ceph_cap_string(wanted), ceph_cap_string(issued)); - __ceph_get_fmode(ci, fmode); + __ceph_touch_fmode(ci, mdsc, fmode); spin_unlock(&ci->i_ceph_lock); /* adjust wanted? */ @@ -404,7 +413,7 @@ int ceph_open(struct inode *inode, struct file *file) return ceph_init_file(inode, file, fmode); } else if (ceph_snap(inode) != CEPH_NOSNAP && (ci->i_snap_caps & wanted) == wanted) { - __ceph_get_fmode(ci, fmode); + __ceph_touch_fmode(ci, mdsc, fmode); spin_unlock(&ci->i_ceph_lock); return ceph_init_file(inode, file, fmode); } @@ -430,6 +439,284 @@ out: return err; } +/* Clone the layout from a synchronous create, if the dir now has Dc caps */ +static void +cache_file_layout(struct inode *dst, struct inode *src) +{ + struct ceph_inode_info *cdst = ceph_inode(dst); + struct ceph_inode_info *csrc = ceph_inode(src); + + spin_lock(&cdst->i_ceph_lock); + if ((__ceph_caps_issued(cdst, NULL) & CEPH_CAP_DIR_CREATE) && + !ceph_file_layout_is_valid(&cdst->i_cached_layout)) { + memcpy(&cdst->i_cached_layout, &csrc->i_layout, + sizeof(cdst->i_cached_layout)); + rcu_assign_pointer(cdst->i_cached_layout.pool_ns, + ceph_try_get_string(csrc->i_layout.pool_ns)); + } + spin_unlock(&cdst->i_ceph_lock); +} + +/* + * Try to set up an async create. We need caps, a file layout, and inode number, + * and either a lease on the dentry or complete dir info. If any of those + * criteria are not satisfied, then return false and the caller can go + * synchronous. + */ +static int try_prep_async_create(struct inode *dir, struct dentry *dentry, + struct ceph_file_layout *lo, u64 *pino) +{ + struct ceph_inode_info *ci = ceph_inode(dir); + struct ceph_dentry_info *di = ceph_dentry(dentry); + int got = 0, want = CEPH_CAP_FILE_EXCL | CEPH_CAP_DIR_CREATE; + u64 ino; + + spin_lock(&ci->i_ceph_lock); + /* No auth cap means no chance for Dc caps */ + if (!ci->i_auth_cap) + goto no_async; + + /* Any delegated inos? */ + if (xa_empty(&ci->i_auth_cap->session->s_delegated_inos)) + goto no_async; + + if (!ceph_file_layout_is_valid(&ci->i_cached_layout)) + goto no_async; + + if ((__ceph_caps_issued(ci, NULL) & want) != want) + goto no_async; + + if (d_in_lookup(dentry)) { + if (!__ceph_dir_is_complete(ci)) + goto no_async; + spin_lock(&dentry->d_lock); + di->lease_shared_gen = atomic_read(&ci->i_shared_gen); + spin_unlock(&dentry->d_lock); + } else if (atomic_read(&ci->i_shared_gen) != + READ_ONCE(di->lease_shared_gen)) { + goto no_async; + } + + ino = ceph_get_deleg_ino(ci->i_auth_cap->session); + if (!ino) + goto no_async; + + *pino = ino; + ceph_take_cap_refs(ci, want, false); + memcpy(lo, &ci->i_cached_layout, sizeof(*lo)); + rcu_assign_pointer(lo->pool_ns, + ceph_try_get_string(ci->i_cached_layout.pool_ns)); + got = want; +no_async: + spin_unlock(&ci->i_ceph_lock); + return got; +} + +static void restore_deleg_ino(struct inode *dir, u64 ino) +{ + struct ceph_inode_info *ci = ceph_inode(dir); + struct ceph_mds_session *s = NULL; + + spin_lock(&ci->i_ceph_lock); + if (ci->i_auth_cap) + s = ceph_get_mds_session(ci->i_auth_cap->session); + spin_unlock(&ci->i_ceph_lock); + if (s) { + int err = ceph_restore_deleg_ino(s, ino); + if (err) + pr_warn("ceph: unable to restore delegated ino 0x%llx to session: %d\n", + ino, err); + ceph_put_mds_session(s); + } +} + +static void wake_async_create_waiters(struct inode *inode, + struct ceph_mds_session *session) +{ + struct ceph_inode_info *ci = ceph_inode(inode); + + spin_lock(&ci->i_ceph_lock); + if (ci->i_ceph_flags & CEPH_I_ASYNC_CREATE) { + ci->i_ceph_flags &= ~CEPH_I_ASYNC_CREATE; + wake_up_bit(&ci->i_ceph_flags, CEPH_ASYNC_CREATE_BIT); + } + ceph_kick_flushing_inode_caps(session, ci); + spin_unlock(&ci->i_ceph_lock); +} + +static void ceph_async_create_cb(struct ceph_mds_client *mdsc, + struct ceph_mds_request *req) +{ + struct dentry *dentry = req->r_dentry; + struct inode *dinode = d_inode(dentry); + struct inode *tinode = req->r_target_inode; + int result = req->r_err ? req->r_err : + le32_to_cpu(req->r_reply_info.head->result); + + WARN_ON_ONCE(dinode && tinode && dinode != tinode); + + /* MDS changed -- caller must resubmit */ + if (result == -EJUKEBOX) + goto out; + + mapping_set_error(req->r_parent->i_mapping, result); + + if (result) { + int pathlen = 0; + u64 base = 0; + char *path = ceph_mdsc_build_path(req->r_dentry, &pathlen, + &base, 0); + + pr_warn("async create failure path=(%llx)%s result=%d!\n", + base, IS_ERR(path) ? "<<bad>>" : path, result); + ceph_mdsc_free_path(path, pathlen); + + ceph_dir_clear_complete(req->r_parent); + if (!d_unhashed(dentry)) + d_drop(dentry); + + if (dinode) { + mapping_set_error(dinode->i_mapping, result); + ceph_inode_shutdown(dinode); + wake_async_create_waiters(dinode, req->r_session); + } + } + + if (tinode) { + u64 ino = ceph_vino(tinode).ino; + + if (req->r_deleg_ino != ino) + pr_warn("%s: inode number mismatch! err=%d deleg_ino=0x%llx target=0x%llx\n", + __func__, req->r_err, req->r_deleg_ino, ino); + + mapping_set_error(tinode->i_mapping, result); + wake_async_create_waiters(tinode, req->r_session); + } else if (!result) { + pr_warn("%s: no req->r_target_inode for 0x%llx\n", __func__, + req->r_deleg_ino); + } +out: + ceph_mdsc_release_dir_caps(req); +} + +static int ceph_finish_async_create(struct inode *dir, struct dentry *dentry, + struct file *file, umode_t mode, + struct ceph_mds_request *req, + struct ceph_acl_sec_ctx *as_ctx, + struct ceph_file_layout *lo) +{ + int ret; + char xattr_buf[4]; + struct ceph_mds_reply_inode in = { }; + struct ceph_mds_reply_info_in iinfo = { .in = &in }; + struct ceph_inode_info *ci = ceph_inode(dir); + struct ceph_dentry_info *di = ceph_dentry(dentry); + struct inode *inode; + struct timespec64 now; + struct ceph_string *pool_ns; + struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(dir->i_sb); + struct ceph_vino vino = { .ino = req->r_deleg_ino, + .snap = CEPH_NOSNAP }; + + ktime_get_real_ts64(&now); + + inode = ceph_get_inode(dentry->d_sb, vino); + if (IS_ERR(inode)) + return PTR_ERR(inode); + + iinfo.inline_version = CEPH_INLINE_NONE; + iinfo.change_attr = 1; + ceph_encode_timespec64(&iinfo.btime, &now); + + if (req->r_pagelist) { + iinfo.xattr_len = req->r_pagelist->length; + iinfo.xattr_data = req->r_pagelist->mapped_tail; + } else { + /* fake it */ + iinfo.xattr_len = ARRAY_SIZE(xattr_buf); + iinfo.xattr_data = xattr_buf; + memset(iinfo.xattr_data, 0, iinfo.xattr_len); + } + + in.ino = cpu_to_le64(vino.ino); + in.snapid = cpu_to_le64(CEPH_NOSNAP); + in.version = cpu_to_le64(1); // ??? + in.cap.caps = in.cap.wanted = cpu_to_le32(CEPH_CAP_ALL_FILE); + in.cap.cap_id = cpu_to_le64(1); + in.cap.realm = cpu_to_le64(ci->i_snap_realm->ino); + in.cap.flags = CEPH_CAP_FLAG_AUTH; + in.ctime = in.mtime = in.atime = iinfo.btime; + in.truncate_seq = cpu_to_le32(1); + in.truncate_size = cpu_to_le64(-1ULL); + in.xattr_version = cpu_to_le64(1); + in.uid = cpu_to_le32(from_kuid(&init_user_ns, current_fsuid())); + if (dir->i_mode & S_ISGID) { + in.gid = cpu_to_le32(from_kgid(&init_user_ns, dir->i_gid)); + + /* Directories always inherit the setgid bit. */ + if (S_ISDIR(mode)) + mode |= S_ISGID; + } else { + in.gid = cpu_to_le32(from_kgid(&init_user_ns, current_fsgid())); + } + in.mode = cpu_to_le32((u32)mode); + + in.nlink = cpu_to_le32(1); + in.max_size = cpu_to_le64(lo->stripe_unit); + + ceph_file_layout_to_legacy(lo, &in.layout); + /* lo is private, so pool_ns can't change */ + pool_ns = rcu_dereference_raw(lo->pool_ns); + if (pool_ns) { + iinfo.pool_ns_len = pool_ns->len; + iinfo.pool_ns_data = pool_ns->str; + } + + down_read(&mdsc->snap_rwsem); + ret = ceph_fill_inode(inode, NULL, &iinfo, NULL, req->r_session, + req->r_fmode, NULL); + up_read(&mdsc->snap_rwsem); + if (ret) { + dout("%s failed to fill inode: %d\n", __func__, ret); + ceph_dir_clear_complete(dir); + if (!d_unhashed(dentry)) + d_drop(dentry); + if (inode->i_state & I_NEW) + discard_new_inode(inode); + } else { + struct dentry *dn; + + dout("%s d_adding new inode 0x%llx to 0x%llx/%s\n", __func__, + vino.ino, ceph_ino(dir), dentry->d_name.name); + ceph_dir_clear_ordered(dir); + ceph_init_inode_acls(inode, as_ctx); + if (inode->i_state & I_NEW) { + /* + * If it's not I_NEW, then someone created this before + * we got here. Assume the server is aware of it at + * that point and don't worry about setting + * CEPH_I_ASYNC_CREATE. + */ + ceph_inode(inode)->i_ceph_flags = CEPH_I_ASYNC_CREATE; + unlock_new_inode(inode); + } + if (d_in_lookup(dentry) || d_really_is_negative(dentry)) { + if (!d_unhashed(dentry)) + d_drop(dentry); + dn = d_splice_alias(inode, dentry); + WARN_ON_ONCE(dn && dn != dentry); + } + file->f_mode |= FMODE_CREATED; + ret = finish_open(file, dentry, ceph_open); + } + + spin_lock(&dentry->d_lock); + di->flags &= ~CEPH_DENTRY_ASYNC_CREATE; + wake_up_bit(&di->flags, CEPH_DENTRY_ASYNC_CREATE_BIT); + spin_unlock(&dentry->d_lock); + + return ret; +} /* * Do a lookup + open with a single request. If we get a non-existent @@ -443,6 +730,7 @@ int ceph_atomic_open(struct inode *dir, struct dentry *dentry, struct ceph_mds_request *req; struct dentry *dn; struct ceph_acl_sec_ctx as_ctx = {}; + bool try_async = ceph_test_mount_opt(fsc, ASYNC_DIROPS); int mask; int err; @@ -453,6 +741,15 @@ int ceph_atomic_open(struct inode *dir, struct dentry *dentry, if (dentry->d_name.len > NAME_MAX) return -ENAMETOOLONG; + err = ceph_wait_on_conflict_unlink(dentry); + if (err) + return err; + /* + * Do not truncate the file, since atomic_open is called before the + * permission check. The caller will do the truncation afterward. + */ + flags &= ~O_TRUNC; + if (flags & O_CREAT) { if (ceph_quota_is_max_files_exceeded(dir)) return -EDQUOT; @@ -462,11 +759,15 @@ int ceph_atomic_open(struct inode *dir, struct dentry *dentry, err = ceph_security_init_secctx(dentry, mode, &as_ctx); if (err < 0) goto out_ctx; + /* Async create can't handle more than a page of xattrs */ + if (as_ctx.pagelist && + !list_is_singular(&as_ctx.pagelist->head)) + try_async = false; } else if (!d_in_lookup(dentry)) { /* If it's not being looked up, it's negative */ return -ENOENT; } - +retry: /* do the open */ req = prepare_open_request(dir->i_sb, flags, mode); if (IS_ERR(req)) { @@ -475,30 +776,65 @@ int ceph_atomic_open(struct inode *dir, struct dentry *dentry, } req->r_dentry = dget(dentry); req->r_num_caps = 2; + mask = CEPH_STAT_CAP_INODE | CEPH_CAP_AUTH_SHARED; + if (ceph_security_xattr_wanted(dir)) + mask |= CEPH_CAP_XATTR_SHARED; + req->r_args.open.mask = cpu_to_le32(mask); + req->r_parent = dir; + ihold(dir); + if (flags & O_CREAT) { + struct ceph_file_layout lo; + req->r_dentry_drop = CEPH_CAP_FILE_SHARED | CEPH_CAP_AUTH_EXCL; req->r_dentry_unless = CEPH_CAP_FILE_EXCL; if (as_ctx.pagelist) { req->r_pagelist = as_ctx.pagelist; as_ctx.pagelist = NULL; } + if (try_async && + (req->r_dir_caps = + try_prep_async_create(dir, dentry, &lo, + &req->r_deleg_ino))) { + struct ceph_dentry_info *di = ceph_dentry(dentry); + + set_bit(CEPH_MDS_R_ASYNC, &req->r_req_flags); + req->r_args.open.flags |= cpu_to_le32(CEPH_O_EXCL); + req->r_callback = ceph_async_create_cb; + + spin_lock(&dentry->d_lock); + di->flags |= CEPH_DENTRY_ASYNC_CREATE; + spin_unlock(&dentry->d_lock); + + err = ceph_mdsc_submit_request(mdsc, dir, req); + if (!err) { + err = ceph_finish_async_create(dir, dentry, + file, mode, req, + &as_ctx, &lo); + } else if (err == -EJUKEBOX) { + restore_deleg_ino(dir, req->r_deleg_ino); + ceph_mdsc_put_request(req); + try_async = false; + ceph_put_string(rcu_dereference_raw(lo.pool_ns)); + goto retry; + } + ceph_put_string(rcu_dereference_raw(lo.pool_ns)); + goto out_req; + } } - mask = CEPH_STAT_CAP_INODE | CEPH_CAP_AUTH_SHARED; - if (ceph_security_xattr_wanted(dir)) - mask |= CEPH_CAP_XATTR_SHARED; - req->r_args.open.mask = cpu_to_le32(mask); - - req->r_parent = dir; set_bit(CEPH_MDS_R_PARENT_LOCKED, &req->r_req_flags); - err = ceph_mdsc_do_request(mdsc, - (flags & (O_CREAT|O_TRUNC)) ? dir : NULL, - req); - err = ceph_handle_snapdir(req, dentry, err); - if (err) - goto out_req; + err = ceph_mdsc_do_request(mdsc, (flags & O_CREAT) ? dir : NULL, req); + if (err == -ENOENT) { + dentry = ceph_handle_snapdir(req, dentry); + if (IS_ERR(dentry)) { + err = PTR_ERR(dentry); + goto out_req; + } + err = 0; + } - if ((flags & O_CREAT) && !req->r_reply_info.head->is_dentry) + if (!err && (flags & O_CREAT) && !req->r_reply_info.head->is_dentry) err = ceph_handle_notrace_create(dir, dentry); if (d_in_lookup(dentry)) { @@ -518,14 +854,15 @@ int ceph_atomic_open(struct inode *dir, struct dentry *dentry, } else { dout("atomic_open finish_open on dn %p\n", dn); if (req->r_op == CEPH_MDS_OP_CREATE && req->r_reply_info.has_create_ino) { - ceph_init_inode_acls(d_inode(dentry), &as_ctx); + struct inode *newino = d_inode(dentry); + + cache_file_layout(dir, newino); + ceph_init_inode_acls(newino, &as_ctx); file->f_mode |= FMODE_CREATED; } err = finish_open(file, dentry, ceph_open); } out_req: - if (!req->r_err && req->r_target_inode) - ceph_put_fmode(ceph_inode(req->r_target_inode), req->r_fmode); ceph_mdsc_put_request(req); out_ctx: ceph_release_acl_sec_ctx(&as_ctx); @@ -542,7 +879,7 @@ int ceph_release(struct inode *inode, struct file *file) dout("release inode %p dir file %p\n", inode, file); WARN_ON(!list_empty(&dfi->file_info.rw_contexts)); - ceph_put_fmode(ci, dfi->file_info.fmode); + ceph_put_fmode(ci, dfi->file_info.fmode, 1); if (dfi->last_readdir) ceph_mdsc_put_request(dfi->last_readdir); @@ -554,7 +891,9 @@ int ceph_release(struct inode *inode, struct file *file) dout("release inode %p regular file %p\n", inode, file); WARN_ON(!list_empty(&fi->rw_contexts)); - ceph_put_fmode(ci, fi->fmode); + ceph_fscache_unuse_cookie(inode, file->f_mode & FMODE_WRITE); + ceph_put_fmode(ci, fi->fmode, 1); + kmem_cache_free(ceph_file_cachep, fi); } @@ -590,6 +929,7 @@ static ssize_t ceph_sync_read(struct kiocb *iocb, struct iov_iter *to, ssize_t ret; u64 off = iocb->ki_pos; u64 len = iov_iter_count(to); + u64 i_size = i_size_read(inode); dout("sync_read on file %p %llu~%u %s\n", file, off, (unsigned)len, (file->f_flags & O_DIRECT) ? "O_DIRECT" : ""); @@ -613,8 +953,9 @@ static ssize_t ceph_sync_read(struct kiocb *iocb, struct iov_iter *to, struct page **pages; int num_pages; size_t page_off; - u64 i_size; bool more; + int idx; + size_t left; req = ceph_osdc_new_request(osdc, &ci->i_layout, ci->i_vino, off, &len, 0, 1, @@ -628,36 +969,25 @@ static ssize_t ceph_sync_read(struct kiocb *iocb, struct iov_iter *to, more = len < iov_iter_count(to); - if (unlikely(iov_iter_is_pipe(to))) { - ret = iov_iter_get_pages_alloc(to, &pages, len, - &page_off); - if (ret <= 0) { - ceph_osdc_put_request(req); - ret = -ENOMEM; - break; - } - num_pages = DIV_ROUND_UP(ret + page_off, PAGE_SIZE); - if (ret < len) { - len = ret; - osd_req_op_extent_update(req, 0, len); - more = false; - } - } else { - num_pages = calc_pages_for(off, len); - page_off = off & ~PAGE_MASK; - pages = ceph_alloc_page_vector(num_pages, GFP_KERNEL); - if (IS_ERR(pages)) { - ceph_osdc_put_request(req); - ret = PTR_ERR(pages); - break; - } + num_pages = calc_pages_for(off, len); + page_off = off & ~PAGE_MASK; + pages = ceph_alloc_page_vector(num_pages, GFP_KERNEL); + if (IS_ERR(pages)) { + ceph_osdc_put_request(req); + ret = PTR_ERR(pages); + break; } osd_req_op_extent_osd_data_pages(req, 0, pages, len, page_off, false, false); - ret = ceph_osdc_start_request(osdc, req, false); - if (!ret) - ret = ceph_osdc_wait_request(osdc, req); + ceph_osdc_start_request(osdc, req); + ret = ceph_osdc_wait_request(osdc, req); + + ceph_update_read_metrics(&fsc->mdsc->metric, + req->r_start_latency, + req->r_end_latency, + len, ret); + ceph_osdc_put_request(req); i_size = i_size_read(inode); @@ -675,36 +1005,27 @@ static ssize_t ceph_sync_read(struct kiocb *iocb, struct iov_iter *to, ret += zlen; } - if (unlikely(iov_iter_is_pipe(to))) { - if (ret > 0) { - iov_iter_advance(to, ret); - off += ret; - } else { - iov_iter_advance(to, 0); - } - ceph_put_page_vector(pages, num_pages, false); - } else { - int idx = 0; - size_t left = ret > 0 ? ret : 0; - while (left > 0) { - size_t len, copied; - page_off = off & ~PAGE_MASK; - len = min_t(size_t, left, PAGE_SIZE - page_off); - copied = copy_page_to_iter(pages[idx++], - page_off, len, to); - off += copied; - left -= copied; - if (copied < len) { - ret = -EFAULT; - break; - } + idx = 0; + left = ret > 0 ? ret : 0; + while (left > 0) { + size_t len, copied; + page_off = off & ~PAGE_MASK; + len = min_t(size_t, left, PAGE_SIZE - page_off); + SetPageUptodate(pages[idx]); + copied = copy_page_to_iter(pages[idx++], + page_off, len, to); + off += copied; + left -= copied; + if (copied < len) { + ret = -EFAULT; + break; } - ceph_release_page_vector(pages, num_pages); } + ceph_release_page_vector(pages, num_pages); if (ret < 0) { - if (ret == -EBLACKLISTED) - fsc->blacklisted = true; + if (ret == -EBLOCKLISTED) + fsc->blocklisted = true; break; } @@ -713,11 +1034,14 @@ static ssize_t ceph_sync_read(struct kiocb *iocb, struct iov_iter *to, } if (off > iocb->ki_pos) { - if (ret >= 0 && - iov_iter_count(to) > 0 && off >= i_size_read(inode)) + if (off >= i_size) { *retry_op = CHECK_EOF; - ret = off - iocb->ki_pos; - iocb->ki_pos = off; + ret = i_size - iocb->ki_pos; + iocb->ki_pos = i_size; + } else { + ret = off - iocb->ki_pos; + iocb->ki_pos = off; + } } dout("sync_read result %zd retry_op %d\n", ret, *retry_op); @@ -772,7 +1096,6 @@ static void ceph_aio_complete(struct inode *inode, } spin_lock(&ci->i_ceph_lock); - ci->i_inline_version = CEPH_INLINE_NONE; dirty = __ceph_mark_dirty_caps(ci, CEPH_CAP_FILE_WR, &aio_req->prealloc_cf); spin_unlock(&ci->i_ceph_lock); @@ -784,7 +1107,7 @@ static void ceph_aio_complete(struct inode *inode, ceph_put_cap_refs(ci, (aio_req->write ? CEPH_CAP_FILE_WR : CEPH_CAP_FILE_RD)); - aio_req->iocb->ki_complete(aio_req->iocb, ret, 0); + aio_req->iocb->ki_complete(aio_req->iocb, ret); ceph_free_cap_flush(aio_req->prealloc_cf); kfree(aio_req); @@ -796,12 +1119,13 @@ static void ceph_aio_complete_req(struct ceph_osd_request *req) struct inode *inode = req->r_inode; struct ceph_aio_request *aio_req = req->r_priv; struct ceph_osd_data *osd_data = osd_req_op_extent_osd_data(req, 0); + struct ceph_client_metric *metric = &ceph_sb_to_mdsc(inode->i_sb)->metric; + unsigned int len = osd_data->bvec_pos.iter.bi_size; BUG_ON(osd_data->type != CEPH_OSD_DATA_TYPE_BVECS); BUG_ON(!osd_data->num_bvecs); - dout("ceph_aio_complete_req %p rc %d bytes %u\n", - inode, rc, osd_data->bvec_pos.iter.bi_size); + dout("ceph_aio_complete_req %p rc %d bytes %u\n", inode, rc, len); if (rc == -EOLDSNAPC) { struct ceph_aio_work *aio_work; @@ -819,9 +1143,9 @@ static void ceph_aio_complete_req(struct ceph_osd_request *req) } else if (!aio_req->write) { if (rc == -ENOENT) rc = 0; - if (rc >= 0 && osd_data->bvec_pos.iter.bi_size > rc) { + if (rc >= 0 && len > rc) { struct iov_iter i; - int zlen = osd_data->bvec_pos.iter.bi_size - rc; + int zlen = len - rc; /* * If read is satisfied by single OSD request, @@ -838,13 +1162,22 @@ static void ceph_aio_complete_req(struct ceph_osd_request *req) } iov_iter_bvec(&i, READ, osd_data->bvec_pos.bvecs, - osd_data->num_bvecs, - osd_data->bvec_pos.iter.bi_size); + osd_data->num_bvecs, len); iov_iter_advance(&i, rc); iov_iter_zero(zlen, &i); } } + /* r_start_latency == 0 means the request was not submitted */ + if (req->r_start_latency) { + if (aio_req->write) + ceph_update_write_metrics(metric, req->r_start_latency, + req->r_end_latency, len, rc); + else + ceph_update_read_metrics(metric, req->r_start_latency, + req->r_end_latency, len, rc); + } + put_bvecs(osd_data->bvec_pos.bvecs, osd_data->num_bvecs, aio_req->should_dirty); ceph_osdc_put_request(req); @@ -911,7 +1244,7 @@ static void ceph_aio_retry_work(struct work_struct *work) req->r_inode = inode; req->r_priv = aio_req; - ret = ceph_osdc_start_request(req->r_osdc, req, false); + ceph_osdc_start_request(req->r_osdc, req); out: if (ret < 0) { req->r_result = ret; @@ -931,6 +1264,7 @@ ceph_direct_read_write(struct kiocb *iocb, struct iov_iter *iter, struct inode *inode = file_inode(file); struct ceph_inode_info *ci = ceph_inode(inode); struct ceph_fs_client *fsc = ceph_inode_to_client(inode); + struct ceph_client_metric *metric = &fsc->mdsc->metric; struct ceph_vino vino; struct ceph_osd_request *req; struct bio_vec *bvecs; @@ -942,7 +1276,7 @@ ceph_direct_read_write(struct kiocb *iocb, struct iov_iter *iter, size_t count = iov_iter_count(iter); loff_t pos = iocb->ki_pos; bool write = iov_iter_rw(iter) == WRITE; - bool should_dirty = !write && iter_is_iovec(iter); + bool should_dirty = !write && user_backed_iter(iter); if (write && ceph_snap(file_inode(file)) != CEPH_NOSNAP) return -EROFS; @@ -952,7 +1286,11 @@ ceph_direct_read_write(struct kiocb *iocb, struct iov_iter *iter, snapc, snapc ? snapc->seq : 0); if (write) { - int ret2 = invalidate_inode_pages2_range(inode->i_mapping, + int ret2; + + ceph_fscache_invalidate(inode, true); + + ret2 = invalidate_inode_pages2_range(inode->i_mapping, pos >> PAGE_SHIFT, (pos + count - 1) >> PAGE_SHIFT); if (ret2 < 0) @@ -1043,9 +1381,15 @@ ceph_direct_read_write(struct kiocb *iocb, struct iov_iter *iter, continue; } - ret = ceph_osdc_start_request(req->r_osdc, req, false); - if (!ret) - ret = ceph_osdc_wait_request(&fsc->client->osdc, req); + ceph_osdc_start_request(req->r_osdc, req); + ret = ceph_osdc_wait_request(&fsc->client->osdc, req); + + if (write) + ceph_update_write_metrics(metric, req->r_start_latency, + req->r_end_latency, len, ret); + else + ceph_update_read_metrics(metric, req->r_start_latency, + req->r_end_latency, len, ret); size = i_size_read(inode); if (!write) { @@ -1101,8 +1445,7 @@ ceph_direct_read_write(struct kiocb *iocb, struct iov_iter *iter, r_private_item); list_del_init(&req->r_private_item); if (ret >= 0) - ret = ceph_osdc_start_request(req->r_osdc, - req, false); + ceph_osdc_start_request(req->r_osdc, req); if (ret < 0) { req->r_result = ret; ceph_aio_complete_req(req); @@ -1156,6 +1499,7 @@ ceph_sync_write(struct kiocb *iocb, struct iov_iter *from, loff_t pos, if (ret < 0) return ret; + ceph_fscache_invalidate(inode, false); ret = invalidate_inode_pages2_range(inode->i_mapping, pos >> PAGE_SHIFT, (pos + count - 1) >> PAGE_SHIFT); @@ -1214,10 +1558,11 @@ ceph_sync_write(struct kiocb *iocb, struct iov_iter *from, loff_t pos, false, true); 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); + ceph_osdc_start_request(&fsc->client->osdc, req); + ret = ceph_osdc_wait_request(&fsc->client->osdc, req); + ceph_update_write_metrics(&fsc->mdsc->metric, req->r_start_latency, + req->r_end_latency, len, ret); out: ceph_osdc_put_request(req); if (ret != 0) { @@ -1259,28 +1604,31 @@ static ssize_t ceph_read_iter(struct kiocb *iocb, struct iov_iter *to) size_t len = iov_iter_count(to); struct inode *inode = file_inode(filp); struct ceph_inode_info *ci = ceph_inode(inode); - struct page *pinned_page = NULL; + bool direct_lock = iocb->ki_flags & IOCB_DIRECT; ssize_t ret; - int want, got = 0; + int want = 0, got = 0; int retry_op = 0, read = 0; again: dout("aio_read %p %llx.%llx %llu~%u trying to get caps on %p\n", inode, ceph_vinop(inode), iocb->ki_pos, (unsigned)len, inode); - if (iocb->ki_flags & IOCB_DIRECT) + if (ceph_inode_is_shutdown(inode)) + return -ESTALE; + + if (direct_lock) ceph_start_io_direct(inode); else ceph_start_io_read(inode); + if (!(fi->flags & CEPH_F_SYNC) && !direct_lock) + want |= CEPH_CAP_FILE_CACHE; if (fi->fmode & CEPH_FILE_MODE_LAZY) - want = CEPH_CAP_FILE_CACHE | CEPH_CAP_FILE_LAZYIO; - else - want = CEPH_CAP_FILE_CACHE; - ret = ceph_get_caps(filp, CEPH_CAP_FILE_RD, want, -1, - &got, &pinned_page); + want |= CEPH_CAP_FILE_LAZYIO; + + ret = ceph_get_caps(filp, CEPH_CAP_FILE_RD, want, -1, &got); if (ret < 0) { - if (iocb->ki_flags & IOCB_DIRECT) + if (direct_lock) ceph_end_io_direct(inode); else ceph_end_io_read(inode); @@ -1295,7 +1643,7 @@ again: inode, ceph_vinop(inode), iocb->ki_pos, (unsigned)len, ceph_cap_string(got)); - if (ci->i_inline_version == CEPH_INLINE_NONE) { + if (!ceph_has_inline_data(ci)) { if (!retry_op && (iocb->ki_flags & IOCB_DIRECT)) { ret = ceph_direct_read_write(iocb, to, NULL, NULL); @@ -1319,13 +1667,9 @@ again: dout("aio_read %p %llx.%llx dropping cap refs on %s = %d\n", inode, ceph_vinop(inode), ceph_cap_string(got), (int)ret); - if (pinned_page) { - put_page(pinned_page); - pinned_page = NULL; - } ceph_put_cap_refs(ci, got); - if (iocb->ki_flags & IOCB_DIRECT) + if (direct_lock) ceph_end_io_direct(inode); else ceph_end_io_read(inode); @@ -1415,13 +1759,19 @@ static ssize_t ceph_write_iter(struct kiocb *iocb, struct iov_iter *from) struct inode *inode = file_inode(file); struct ceph_inode_info *ci = ceph_inode(inode); struct ceph_fs_client *fsc = ceph_inode_to_client(inode); + struct ceph_osd_client *osdc = &fsc->client->osdc; struct ceph_cap_flush *prealloc_cf; ssize_t count, written = 0; - int err, want, got; + int err, want = 0, got; bool direct_lock = false; + u32 map_flags; + u64 pool_flags; loff_t pos; loff_t limit = max(i_size_read(inode), fsc->max_file_size); + if (ceph_inode_is_shutdown(inode)) + return -ESTALE; + if (ceph_snap(inode) != CEPH_NOSNAP) return -EROFS; @@ -1465,40 +1815,37 @@ retry_snap: goto out; } - err = file_remove_privs(file); - if (err) - goto out; - - err = file_update_time(file); - if (err) + down_read(&osdc->lock); + map_flags = osdc->osdmap->flags; + pool_flags = ceph_pg_pool_flags(osdc->osdmap, ci->i_layout.pool_id); + up_read(&osdc->lock); + if ((map_flags & CEPH_OSDMAP_FULL) || + (pool_flags & CEPH_POOL_FLAG_FULL)) { + err = -ENOSPC; goto out; - - inode_inc_iversion_raw(inode); - - if (ci->i_inline_version != CEPH_INLINE_NONE) { - err = ceph_uninline_data(file, NULL); - if (err < 0) - goto out; } - /* FIXME: not complete since it doesn't account for being at quota */ - if (ceph_osdmap_flag(&fsc->client->osdc, CEPH_OSDMAP_FULL)) { - err = -ENOSPC; + err = file_remove_privs(file); + if (err) goto out; - } dout("aio_write %p %llx.%llx %llu~%zd getting caps. i_size %llu\n", inode, ceph_vinop(inode), pos, count, i_size_read(inode)); + if (!(fi->flags & CEPH_F_SYNC) && !direct_lock) + want |= CEPH_CAP_FILE_BUFFER; if (fi->fmode & CEPH_FILE_MODE_LAZY) - want = CEPH_CAP_FILE_BUFFER | CEPH_CAP_FILE_LAZYIO; - else - want = CEPH_CAP_FILE_BUFFER; + want |= CEPH_CAP_FILE_LAZYIO; got = 0; - err = ceph_get_caps(file, CEPH_CAP_FILE_WR, want, pos + count, - &got, NULL); + err = ceph_get_caps(file, CEPH_CAP_FILE_WR, want, pos + count, &got); if (err < 0) goto out; + err = file_update_time(file); + if (err) + goto out_caps; + + inode_inc_iversion_raw(inode); + dout("aio_write %p %llx.%llx %llu~%zd got cap refs on %s\n", inode, ceph_vinop(inode), pos, count, ceph_cap_string(got)); @@ -1543,7 +1890,7 @@ retry_snap: * are pending vmtruncate. So write and vmtruncate * can not run at the same time */ - written = generic_perform_write(file, from, pos); + written = generic_perform_write(iocb, from); if (likely(written >= 0)) iocb->ki_pos = pos + written; ceph_end_io_write(inode); @@ -1553,14 +1900,13 @@ retry_snap: int dirty; spin_lock(&ci->i_ceph_lock); - ci->i_inline_version = CEPH_INLINE_NONE; dirty = __ceph_mark_dirty_caps(ci, CEPH_CAP_FILE_WR, &prealloc_cf); spin_unlock(&ci->i_ceph_lock); if (dirty) __mark_inode_dirty(inode, dirty); if (ceph_quota_is_max_bytes_approaching(inode, iocb->ki_pos)) - ceph_check_caps(ci, CHECK_CAPS_NODELAY, NULL); + ceph_check_caps(ci, CHECK_CAPS_FLUSH, NULL); } dout("aio_write %p %llx.%llx %llu~%u dropping cap refs on %s\n", @@ -1575,12 +1921,15 @@ retry_snap: } if (written >= 0) { - if (ceph_osdmap_flag(&fsc->client->osdc, CEPH_OSDMAP_NEARFULL)) + if ((map_flags & CEPH_OSDMAP_NEARFULL) || + (pool_flags & CEPH_POOL_FLAG_NEARFULL)) iocb->ki_flags |= IOCB_DSYNC; written = generic_write_sync(iocb, written); } goto out_unlocked; +out_caps: + ceph_put_cap_refs(ci, got); out: if (direct_lock) ceph_end_io_direct(inode); @@ -1597,57 +1946,15 @@ out_unlocked: */ static loff_t ceph_llseek(struct file *file, loff_t offset, int whence) { - struct inode *inode = file->f_mapping->host; - struct ceph_fs_client *fsc = ceph_inode_to_client(inode); - loff_t i_size; - loff_t ret; - - inode_lock(inode); - if (whence == SEEK_END || whence == SEEK_DATA || whence == SEEK_HOLE) { + struct inode *inode = file_inode(file); + int ret; + ret = ceph_do_getattr(inode, CEPH_STAT_CAP_SIZE, false); if (ret < 0) - goto out; - } - - i_size = i_size_read(inode); - switch (whence) { - case SEEK_END: - offset += i_size; - break; - case SEEK_CUR: - /* - * Here we special-case the lseek(fd, 0, SEEK_CUR) - * position-querying operation. Avoid rewriting the "same" - * f_pos value back to the file because a concurrent read(), - * write() or lseek() might have altered it - */ - if (offset == 0) { - ret = file->f_pos; - goto out; - } - offset += file->f_pos; - break; - case SEEK_DATA: - if (offset < 0 || offset >= i_size) { - ret = -ENXIO; - goto out; - } - break; - case SEEK_HOLE: - if (offset < 0 || offset >= i_size) { - ret = -ENXIO; - goto out; - } - offset = i_size; - break; + return ret; } - - ret = vfs_setpos(file, offset, max(i_size, fsc->max_file_size)); - -out: - inode_unlock(inode); - return ret; + return generic_file_llseek(file, offset, whence); } static inline void ceph_zero_partial_page( @@ -1716,12 +2023,10 @@ static int ceph_zero_partial_object(struct inode *inode, } 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); - if (ret == -ENOENT) - ret = 0; - } + ceph_osdc_start_request(&fsc->client->osdc, req); + ret = ceph_osdc_wait_request(&fsc->client->osdc, req); + if (ret == -ENOENT) + ret = 0; ceph_osdc_put_request(req); out: @@ -1804,12 +2109,6 @@ static long ceph_fallocate(struct file *file, int mode, goto unlock; } - if (ci->i_inline_version != CEPH_INLINE_NONE) { - ret = ceph_uninline_data(file, NULL); - if (ret < 0) - goto unlock; - } - size = i_size_read(inode); /* Are we punching a hole beyond EOF? */ @@ -1823,22 +2122,24 @@ static long ceph_fallocate(struct file *file, int mode, else want = CEPH_CAP_FILE_BUFFER; - ret = ceph_get_caps(file, CEPH_CAP_FILE_WR, want, endoff, &got, NULL); + ret = ceph_get_caps(file, CEPH_CAP_FILE_WR, want, endoff, &got); if (ret < 0) goto unlock; + filemap_invalidate_lock(inode->i_mapping); + ceph_fscache_invalidate(inode, false); ceph_zero_pagecache_range(inode, offset, length); ret = ceph_zero_objects(inode, offset, length); if (!ret) { spin_lock(&ci->i_ceph_lock); - ci->i_inline_version = CEPH_INLINE_NONE; dirty = __ceph_mark_dirty_caps(ci, CEPH_CAP_FILE_WR, &prealloc_cf); spin_unlock(&ci->i_ceph_lock); if (dirty) __mark_inode_dirty(inode, dirty); } + filemap_invalidate_unlock(inode->i_mapping); ceph_put_cap_refs(ci, got); unlock: @@ -1861,7 +2162,7 @@ static int get_rd_wr_caps(struct file *src_filp, int *src_got, retry_caps: ret = ceph_get_caps(dst_filp, CEPH_CAP_FILE_WR, CEPH_CAP_FILE_BUFFER, - dst_endoff, dst_got, NULL); + dst_endoff, dst_got); if (ret < 0) return ret; @@ -1883,7 +2184,7 @@ retry_caps: return ret; } ret = ceph_get_caps(src_filp, CEPH_CAP_FILE_RD, - CEPH_CAP_FILE_SHARED, -1, src_got, NULL); + CEPH_CAP_FILE_SHARED, -1, src_got); if (ret < 0) return ret; /*... drop src_ci caps too, and retry */ @@ -1936,6 +2237,127 @@ static int is_file_size_ok(struct inode *src_inode, struct inode *dst_inode, return 0; } +static struct ceph_osd_request * +ceph_alloc_copyfrom_request(struct ceph_osd_client *osdc, + u64 src_snapid, + struct ceph_object_id *src_oid, + struct ceph_object_locator *src_oloc, + struct ceph_object_id *dst_oid, + struct ceph_object_locator *dst_oloc, + u32 truncate_seq, u64 truncate_size) +{ + struct ceph_osd_request *req; + int ret; + u32 src_fadvise_flags = + CEPH_OSD_OP_FLAG_FADVISE_SEQUENTIAL | + CEPH_OSD_OP_FLAG_FADVISE_NOCACHE; + u32 dst_fadvise_flags = + CEPH_OSD_OP_FLAG_FADVISE_SEQUENTIAL | + CEPH_OSD_OP_FLAG_FADVISE_DONTNEED; + + req = ceph_osdc_alloc_request(osdc, NULL, 1, false, GFP_KERNEL); + if (!req) + return ERR_PTR(-ENOMEM); + + req->r_flags = CEPH_OSD_FLAG_WRITE; + + ceph_oloc_copy(&req->r_t.base_oloc, dst_oloc); + ceph_oid_copy(&req->r_t.base_oid, dst_oid); + + ret = osd_req_op_copy_from_init(req, src_snapid, 0, + src_oid, src_oloc, + src_fadvise_flags, + dst_fadvise_flags, + truncate_seq, + truncate_size, + CEPH_OSD_COPY_FROM_FLAG_TRUNCATE_SEQ); + if (ret) + goto out; + + ret = ceph_osdc_alloc_messages(req, GFP_KERNEL); + if (ret) + goto out; + + return req; + +out: + ceph_osdc_put_request(req); + return ERR_PTR(ret); +} + +static ssize_t ceph_do_objects_copy(struct ceph_inode_info *src_ci, u64 *src_off, + struct ceph_inode_info *dst_ci, u64 *dst_off, + struct ceph_fs_client *fsc, + size_t len, unsigned int flags) +{ + struct ceph_object_locator src_oloc, dst_oloc; + struct ceph_object_id src_oid, dst_oid; + struct ceph_osd_client *osdc; + struct ceph_osd_request *req; + size_t bytes = 0; + u64 src_objnum, src_objoff, dst_objnum, dst_objoff; + u32 src_objlen, dst_objlen; + u32 object_size = src_ci->i_layout.object_size; + int ret; + + src_oloc.pool = src_ci->i_layout.pool_id; + src_oloc.pool_ns = ceph_try_get_string(src_ci->i_layout.pool_ns); + dst_oloc.pool = dst_ci->i_layout.pool_id; + dst_oloc.pool_ns = ceph_try_get_string(dst_ci->i_layout.pool_ns); + osdc = &fsc->client->osdc; + + while (len >= object_size) { + ceph_calc_file_object_mapping(&src_ci->i_layout, *src_off, + object_size, &src_objnum, + &src_objoff, &src_objlen); + ceph_calc_file_object_mapping(&dst_ci->i_layout, *dst_off, + object_size, &dst_objnum, + &dst_objoff, &dst_objlen); + ceph_oid_init(&src_oid); + ceph_oid_printf(&src_oid, "%llx.%08llx", + src_ci->i_vino.ino, src_objnum); + ceph_oid_init(&dst_oid); + ceph_oid_printf(&dst_oid, "%llx.%08llx", + dst_ci->i_vino.ino, dst_objnum); + /* Do an object remote copy */ + req = ceph_alloc_copyfrom_request(osdc, src_ci->i_vino.snap, + &src_oid, &src_oloc, + &dst_oid, &dst_oloc, + dst_ci->i_truncate_seq, + dst_ci->i_truncate_size); + if (IS_ERR(req)) + ret = PTR_ERR(req); + else { + ceph_osdc_start_request(osdc, req); + ret = ceph_osdc_wait_request(osdc, req); + ceph_update_copyfrom_metrics(&fsc->mdsc->metric, + req->r_start_latency, + req->r_end_latency, + object_size, ret); + ceph_osdc_put_request(req); + } + if (ret) { + if (ret == -EOPNOTSUPP) { + fsc->have_copy_from2 = false; + pr_notice("OSDs don't support copy-from2; disabling copy offload\n"); + } + dout("ceph_osdc_copy_from returned %d\n", ret); + if (!bytes) + bytes = ret; + goto out; + } + len -= object_size; + bytes += object_size; + *src_off += object_size; + *dst_off += object_size; + } + +out: + ceph_oloc_destroy(&src_oloc); + ceph_oloc_destroy(&dst_oloc); + return bytes; +} + static ssize_t __ceph_copy_file_range(struct file *src_file, loff_t src_off, struct file *dst_file, loff_t dst_off, size_t len, unsigned int flags) @@ -1946,14 +2368,11 @@ static ssize_t __ceph_copy_file_range(struct file *src_file, loff_t src_off, struct ceph_inode_info *dst_ci = ceph_inode(dst_inode); struct ceph_cap_flush *prealloc_cf; struct ceph_fs_client *src_fsc = ceph_inode_to_client(src_inode); - struct ceph_object_locator src_oloc, dst_oloc; - struct ceph_object_id src_oid, dst_oid; - loff_t endoff = 0, size; - ssize_t ret = -EIO; + loff_t size; + ssize_t ret = -EIO, bytes; u64 src_objnum, dst_objnum, src_objoff, dst_objoff; - u32 src_objlen, dst_objlen, object_size; + u32 src_objlen, dst_objlen; int src_got = 0, dst_got = 0, err, dirty; - bool do_final_copy = false; if (src_inode->i_sb != dst_inode->i_sb) { struct ceph_fs_client *dst_fsc = ceph_inode_to_client(dst_inode); @@ -2031,22 +2450,15 @@ static ssize_t __ceph_copy_file_range(struct file *src_file, loff_t src_off, if (ret < 0) goto out_caps; - size = i_size_read(dst_inode); - endoff = dst_off + len; - /* Drop dst file cached pages */ + ceph_fscache_invalidate(dst_inode, false); ret = invalidate_inode_pages2_range(dst_inode->i_mapping, dst_off >> PAGE_SHIFT, - endoff >> PAGE_SHIFT); + (dst_off + len) >> PAGE_SHIFT); if (ret < 0) { dout("Failed to invalidate inode pages (%zd)\n", ret); ret = 0; /* XXX */ } - src_oloc.pool = src_ci->i_layout.pool_id; - src_oloc.pool_ns = ceph_try_get_string(src_ci->i_layout.pool_ns); - dst_oloc.pool = dst_ci->i_layout.pool_id; - dst_oloc.pool_ns = ceph_try_get_string(dst_ci->i_layout.pool_ns); - ceph_calc_file_object_mapping(&src_ci->i_layout, src_off, src_ci->i_layout.object_size, &src_objnum, &src_objoff, &src_objlen); @@ -2065,6 +2477,8 @@ static ssize_t __ceph_copy_file_range(struct file *src_file, loff_t src_off, * starting at the src_off */ if (src_objoff) { + dout("Initial partial copy of %u bytes\n", src_objlen); + /* * we need to temporarily drop all caps as we'll be calling * {read,write}_iter, which will get caps again. @@ -2072,8 +2486,9 @@ static ssize_t __ceph_copy_file_range(struct file *src_file, loff_t src_off, put_rd_wr_caps(src_ci, src_got, dst_ci, dst_got); ret = do_splice_direct(src_file, &src_off, dst_file, &dst_off, src_objlen, flags); - if (ret < 0) { - dout("do_splice_direct returned %d\n", err); + /* Abort on short copies or on error */ + if (ret < src_objlen) { + dout("Failed partial copy (%zd)\n", ret); goto out; } len -= ret; @@ -2086,69 +2501,31 @@ static ssize_t __ceph_copy_file_range(struct file *src_file, loff_t src_off, if (err < 0) goto out_caps; } - object_size = src_ci->i_layout.object_size; - while (len >= object_size) { - ceph_calc_file_object_mapping(&src_ci->i_layout, src_off, - object_size, &src_objnum, - &src_objoff, &src_objlen); - ceph_calc_file_object_mapping(&dst_ci->i_layout, dst_off, - object_size, &dst_objnum, - &dst_objoff, &dst_objlen); - ceph_oid_init(&src_oid); - ceph_oid_printf(&src_oid, "%llx.%08llx", - src_ci->i_vino.ino, src_objnum); - ceph_oid_init(&dst_oid); - ceph_oid_printf(&dst_oid, "%llx.%08llx", - dst_ci->i_vino.ino, dst_objnum); - /* Do an object remote copy */ - err = ceph_osdc_copy_from( - &src_fsc->client->osdc, - src_ci->i_vino.snap, 0, - &src_oid, &src_oloc, - CEPH_OSD_OP_FLAG_FADVISE_SEQUENTIAL | - CEPH_OSD_OP_FLAG_FADVISE_NOCACHE, - &dst_oid, &dst_oloc, - CEPH_OSD_OP_FLAG_FADVISE_SEQUENTIAL | - CEPH_OSD_OP_FLAG_FADVISE_DONTNEED, - dst_ci->i_truncate_seq, dst_ci->i_truncate_size, - CEPH_OSD_COPY_FROM_FLAG_TRUNCATE_SEQ); - if (err) { - if (err == -EOPNOTSUPP) { - src_fsc->have_copy_from2 = false; - pr_notice("OSDs don't support copy-from2; disabling copy offload\n"); - } - dout("ceph_osdc_copy_from returned %d\n", err); - if (!ret) - ret = err; - goto out_caps; - } - len -= object_size; - src_off += object_size; - dst_off += object_size; - ret += object_size; - } - if (len) - /* We still need one final local copy */ - do_final_copy = true; + size = i_size_read(dst_inode); + bytes = ceph_do_objects_copy(src_ci, &src_off, dst_ci, &dst_off, + src_fsc, len, flags); + if (bytes <= 0) { + if (!ret) + ret = bytes; + goto out_caps; + } + dout("Copied %zu bytes out of %zu\n", bytes, len); + len -= bytes; + ret += bytes; file_update_time(dst_file); inode_inc_iversion_raw(dst_inode); - if (endoff > size) { - int caps_flags = 0; - + if (dst_off > size) { /* Let the MDS know about dst file size change */ - if (ceph_quota_is_max_bytes_approaching(dst_inode, endoff)) - caps_flags |= CHECK_CAPS_NODELAY; - if (ceph_inode_set_size(dst_inode, endoff)) - caps_flags |= CHECK_CAPS_AUTHONLY; - if (caps_flags) - ceph_check_caps(dst_ci, caps_flags, NULL); + if (ceph_inode_set_size(dst_inode, dst_off) || + ceph_quota_is_max_bytes_approaching(dst_inode, dst_off)) + ceph_check_caps(dst_ci, CHECK_CAPS_AUTHONLY | CHECK_CAPS_FLUSH, + NULL); } /* Mark Fw dirty */ spin_lock(&dst_ci->i_ceph_lock); - dst_ci->i_inline_version = CEPH_INLINE_NONE; dirty = __ceph_mark_dirty_caps(dst_ci, CEPH_CAP_FILE_WR, &prealloc_cf); spin_unlock(&dst_ci->i_ceph_lock); if (dirty) @@ -2157,15 +2534,18 @@ static ssize_t __ceph_copy_file_range(struct file *src_file, loff_t src_off, out_caps: put_rd_wr_caps(src_ci, src_got, dst_ci, dst_got); - if (do_final_copy) { - err = do_splice_direct(src_file, &src_off, dst_file, - &dst_off, len, flags); - if (err < 0) { - dout("do_splice_direct returned %d\n", err); - goto out; - } - len -= err; - ret += err; + /* + * Do the final manual copy if we still have some bytes left, unless + * there were errors in remote object copies (len >= object_size). + */ + if (len && (len < src_ci->i_layout.object_size)) { + dout("Final partial copy of %zu bytes\n", len); + bytes = do_splice_direct(src_file, &src_off, dst_file, + &dst_off, len, flags); + if (bytes > 0) + ret += bytes; + else + dout("Failed partial copy (%zd)\n", bytes); } out: @@ -2198,6 +2578,7 @@ const struct file_operations ceph_file_fops = { .mmap = ceph_mmap, .fsync = ceph_fsync, .lock = ceph_lock, + .setlease = simple_nosetlease, .flock = ceph_flock, .splice_read = generic_file_splice_read, .splice_write = iter_file_splice_write, diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c index d01710a16a4a..4af5e55abc15 100644 --- a/fs/ceph/inode.c +++ b/fs/ceph/inode.c @@ -41,26 +41,31 @@ static void ceph_inode_work(struct work_struct *work); */ static int ceph_set_ino_cb(struct inode *inode, void *data) { - ceph_inode(inode)->i_vino = *(struct ceph_vino *)data; - inode->i_ino = ceph_vino_to_ino(*(struct ceph_vino *)data); + struct ceph_inode_info *ci = ceph_inode(inode); + struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(inode->i_sb); + + ci->i_vino = *(struct ceph_vino *)data; + inode->i_ino = ceph_vino_to_ino_t(ci->i_vino); inode_set_iversion_raw(inode, 0); + percpu_counter_inc(&mdsc->metric.total_inodes); + return 0; } struct inode *ceph_get_inode(struct super_block *sb, struct ceph_vino vino) { struct inode *inode; - ino_t t = ceph_vino_to_ino(vino); - inode = iget5_locked(sb, t, ceph_ino_compare, ceph_set_ino_cb, &vino); + if (ceph_vino_is_reserved(vino)) + return ERR_PTR(-EREMOTEIO); + + inode = iget5_locked(sb, (unsigned long)vino.ino, ceph_ino_compare, + ceph_set_ino_cb, &vino); if (!inode) return ERR_PTR(-ENOMEM); - if (inode->i_state & I_NEW) - dout("get_inode created new inode %p %llx.%llx ino %llx\n", - inode, ceph_vinop(inode), (u64)inode->i_ino); - dout("get_inode on %lu=%llx.%llx got %p\n", inode->i_ino, vino.ino, - vino.snap, inode); + dout("get_inode on %llu=%llx.%llx got %p new %d\n", ceph_present_inode(inode), + ceph_vinop(inode), inode, !!(inode->i_state & I_NEW)); return inode; } @@ -76,21 +81,44 @@ struct inode *ceph_get_snapdir(struct inode *parent) struct inode *inode = ceph_get_inode(parent->i_sb, vino); struct ceph_inode_info *ci = ceph_inode(inode); - BUG_ON(!S_ISDIR(parent->i_mode)); if (IS_ERR(inode)) return inode; + + if (!S_ISDIR(parent->i_mode)) { + pr_warn_once("bad snapdir parent type (mode=0%o)\n", + parent->i_mode); + goto err; + } + + if (!(inode->i_state & I_NEW) && !S_ISDIR(inode->i_mode)) { + pr_warn_once("bad snapdir inode type (mode=0%o)\n", + inode->i_mode); + goto err; + } + inode->i_mode = parent->i_mode; inode->i_uid = parent->i_uid; inode->i_gid = parent->i_gid; - inode->i_op = &ceph_snapdir_iops; - inode->i_fop = &ceph_snapdir_fops; - ci->i_snap_caps = CEPH_CAP_PIN; /* so we can open */ + inode->i_mtime = parent->i_mtime; + inode->i_ctime = parent->i_ctime; + inode->i_atime = parent->i_atime; ci->i_rbytes = 0; + ci->i_btime = ceph_inode(parent)->i_btime; - if (inode->i_state & I_NEW) + if (inode->i_state & I_NEW) { + inode->i_op = &ceph_snapdir_iops; + inode->i_fop = &ceph_snapdir_fops; + ci->i_snap_caps = CEPH_CAP_PIN; /* so we can open */ unlock_new_inode(inode); + } return inode; +err: + if ((inode->i_state & I_NEW)) + discard_new_inode(inode); + else + iput(inode); + return ERR_PTR(-ENOTDIR); } const struct inode_operations ceph_file_iops = { @@ -148,7 +176,7 @@ static struct ceph_inode_frag *__get_or_create_frag(struct ceph_inode_info *ci, rb_insert_color(&frag->node, &ci->i_fragtree); dout("get_or_create_frag added %llx.%llx frag %x\n", - ceph_vinop(&ci->vfs_inode), f); + ceph_vinop(&ci->netfs.inode), f); return frag; } @@ -334,7 +362,7 @@ static int ceph_fill_fragtree(struct inode *inode, if (nsplits != ci->i_fragtree_nsplits) { update = true; } else if (nsplits) { - i = prandom_u32() % nsplits; + i = prandom_u32_max(nsplits); id = le32_to_cpu(fragtree->splits[i].frag); if (!__ceph_find_frag(ci, id)) update = true; @@ -425,11 +453,14 @@ struct inode *ceph_alloc_inode(struct super_block *sb) struct ceph_inode_info *ci; int i; - ci = kmem_cache_alloc(ceph_inode_cachep, GFP_NOFS); + ci = alloc_inode_sb(sb, ceph_inode_cachep, GFP_NOFS); if (!ci) return NULL; - dout("alloc_inode %p\n", &ci->vfs_inode); + dout("alloc_inode %p\n", &ci->netfs.inode); + + /* Set parameters for the netfs library */ + netfs_inode_init(&ci->netfs, &ceph_netfs_ops); spin_lock_init(&ci->i_ceph_lock); @@ -447,6 +478,7 @@ struct inode *ceph_alloc_inode(struct super_block *sb) ci->i_max_files = 0; memset(&ci->i_dir_layout, 0, sizeof(ci->i_dir_layout)); + memset(&ci->i_cached_layout, 0, sizeof(ci->i_cached_layout)); RCU_INIT_POINTER(ci->i_layout.pool_ns, NULL); ci->i_fragtree = RB_ROOT; @@ -471,13 +503,13 @@ struct inode *ceph_alloc_inode(struct super_block *sb) ci->i_prealloc_cap_flush = NULL; INIT_LIST_HEAD(&ci->i_cap_flush_list); init_waitqueue_head(&ci->i_cap_wq); - ci->i_hold_caps_min = 0; ci->i_hold_caps_max = 0; INIT_LIST_HEAD(&ci->i_cap_delay_list); INIT_LIST_HEAD(&ci->i_cap_snaps); ci->i_head_snapc = NULL; ci->i_snap_caps = 0; + ci->i_last_rd = ci->i_last_wr = jiffies - 3600 * HZ; for (i = 0; i < CEPH_FILE_MODE_BITS; i++) ci->i_nr_by_mode[i] = 0; @@ -496,6 +528,7 @@ struct inode *ceph_alloc_inode(struct super_block *sb) ci->i_rdcache_ref = 0; ci->i_wr_ref = 0; ci->i_wb_ref = 0; + ci->i_fx_ref = 0; ci->i_wrbuffer_ref = 0; ci->i_wrbuffer_ref_head = 0; atomic_set(&ci->i_filelock_ref, 0); @@ -514,12 +547,7 @@ struct inode *ceph_alloc_inode(struct super_block *sb) INIT_WORK(&ci->i_work, ceph_inode_work); ci->i_work_mask = 0; memset(&ci->i_btime, '\0', sizeof(ci->i_btime)); - - ceph_fscache_inode_init(ci); - - ci->i_meta_err = 0; - - return &ci->vfs_inode; + return &ci->netfs.inode; } void ceph_free_inode(struct inode *inode) @@ -533,19 +561,24 @@ void ceph_free_inode(struct inode *inode) void ceph_evict_inode(struct inode *inode) { struct ceph_inode_info *ci = ceph_inode(inode); + struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(inode->i_sb); struct ceph_inode_frag *frag; struct rb_node *n; dout("evict_inode %p ino %llx.%llx\n", inode, ceph_vinop(inode)); + percpu_counter_dec(&mdsc->metric.total_inodes); + truncate_inode_pages_final(&inode->i_data); + if (inode->i_state & I_PINNING_FSCACHE_WB) + ceph_fscache_unuse_cookie(inode, true); clear_inode(inode); ceph_fscache_unregister_inode_cookie(ci); __ceph_remove_caps(ci); - if (__ceph_has_any_quota(ci)) + if (__ceph_has_quota(ci, QUOTA_GET_ANY)) ceph_adjust_quota_realms_count(inode, false); /* @@ -553,19 +586,10 @@ void ceph_evict_inode(struct inode *inode) * caps in i_snap_caps. */ if (ci->i_snap_realm) { - struct ceph_mds_client *mdsc = - ceph_inode_to_client(inode)->mdsc; if (ceph_snap(inode) == CEPH_NOSNAP) { - struct ceph_snap_realm *realm = ci->i_snap_realm; dout(" dropping residual ref to snap realm %p\n", - realm); - spin_lock(&realm->inodes_with_caps_lock); - list_del_init(&ci->i_snap_realm_item); - ci->i_snap_realm = NULL; - if (realm->ino == ci->i_vino.ino) - realm->inode = NULL; - spin_unlock(&realm->inodes_with_caps_lock); - ceph_put_snap_realm(mdsc, realm); + ci->i_snap_realm); + ceph_change_snap_realm(inode, NULL); } else { ceph_put_snapid_map(mdsc, ci->i_snapid_map); ci->i_snap_realm = NULL; @@ -586,6 +610,7 @@ void ceph_evict_inode(struct inode *inode) ceph_buffer_put(ci->i_xattrs.prealloc_blob); ceph_put_string(rcu_dereference_raw(ci->i_layout.pool_ns)); + ceph_put_string(rcu_dereference_raw(ci->i_cached_layout.pool_ns)); } static inline blkcnt_t calc_inode_blocks(u64 size) @@ -606,16 +631,23 @@ int ceph_fill_file_size(struct inode *inode, int issued, { struct ceph_inode_info *ci = ceph_inode(inode); int queue_trunc = 0; + loff_t isize = i_size_read(inode); if (ceph_seq_cmp(truncate_seq, ci->i_truncate_seq) > 0 || - (truncate_seq == ci->i_truncate_seq && size > inode->i_size)) { - dout("size %lld -> %llu\n", inode->i_size, size); + (truncate_seq == ci->i_truncate_seq && size > isize)) { + dout("size %lld -> %llu\n", isize, size); if (size > 0 && S_ISDIR(inode->i_mode)) { pr_err("fill_file_size non-zero size for directory\n"); size = 0; } i_size_write(inode, size); inode->i_blocks = calc_inode_blocks(size); + /* + * If we're expanding, then we should be able to just update + * the existing cookie. + */ + if (size > isize) + ceph_fscache_update(inode); ci->i_reported_size = size; if (truncate_seq != ci->i_truncate_seq) { dout("truncate_seq %u -> %u\n", @@ -636,7 +668,7 @@ int ceph_fill_file_size(struct inode *inode, int issued, if ((issued & (CEPH_CAP_FILE_CACHE| CEPH_CAP_FILE_BUFFER)) || mapping_mapped(inode->i_mapping) || - __ceph_caps_file_wanted(ci)) { + __ceph_is_file_opened(ci)) { ci->i_truncate_pending++; queue_trunc = 1; } @@ -648,10 +680,6 @@ int ceph_fill_file_size(struct inode *inode, int issued, truncate_size); ci->i_truncate_size = truncate_size; } - - if (queue_trunc) - ceph_fscache_invalidate(inode); - return queue_trunc; } @@ -727,13 +755,13 @@ void ceph_fill_file_time(struct inode *inode, int issued, * Populate an inode based on info from mds. May be called on new or * existing inodes. */ -static int fill_inode(struct inode *inode, struct page *locked_page, - struct ceph_mds_reply_info_in *iinfo, - struct ceph_mds_reply_dirfrag *dirinfo, - struct ceph_mds_session *session, int cap_fmode, - struct ceph_cap_reservation *caps_reservation) +int ceph_fill_inode(struct inode *inode, struct page *locked_page, + struct ceph_mds_reply_info_in *iinfo, + struct ceph_mds_reply_dirfrag *dirinfo, + struct ceph_mds_session *session, int cap_fmode, + struct ceph_cap_reservation *caps_reservation) { - struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc; + struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(inode->i_sb); struct ceph_mds_reply_inode *info = iinfo->in; struct ceph_inode_info *ci = ceph_inode(inode); int issued, new_issued, info_caps; @@ -747,11 +775,34 @@ static int fill_inode(struct inode *inode, struct page *locked_page, bool queue_trunc = false; bool new_version = false; bool fill_inline = false; + umode_t mode = le32_to_cpu(info->mode); + dev_t rdev = le32_to_cpu(info->rdev); + + lockdep_assert_held(&mdsc->snap_rwsem); - dout("fill_inode %p ino %llx.%llx v %llu had %llu\n", + dout("%s %p ino %llx.%llx v %llu had %llu\n", __func__, inode, ceph_vinop(inode), le64_to_cpu(info->version), ci->i_version); + /* Once I_NEW is cleared, we can't change type or dev numbers */ + if (inode->i_state & I_NEW) { + inode->i_mode = mode; + } else { + if (inode_wrong_type(inode, mode)) { + pr_warn_once("inode type changed! (ino %llx.%llx is 0%o, mds says 0%o)\n", + ceph_vinop(inode), inode->i_mode, mode); + return -ESTALE; + } + + if ((S_ISCHR(mode) || S_ISBLK(mode)) && inode->i_rdev != rdev) { + pr_warn_once("dev inode rdev changed! (ino %llx.%llx is %u:%u, mds says %u:%u)\n", + ceph_vinop(inode), MAJOR(inode->i_rdev), + MINOR(inode->i_rdev), MAJOR(rdev), + MINOR(rdev)); + return -ESTALE; + } + } + info_caps = le32_to_cpu(info->cap.caps); /* prealloc new cap struct */ @@ -769,7 +820,7 @@ static int fill_inode(struct inode *inode, struct page *locked_page, if (iinfo->xattr_len > 4) { xattr_blob = ceph_buffer_new(iinfo->xattr_len, GFP_NOFS); if (!xattr_blob) - pr_err("fill_inode ENOMEM xattr blob %d bytes\n", + pr_err("%s ENOMEM xattr blob %d bytes\n", __func__, iinfo->xattr_len); } @@ -805,8 +856,6 @@ static int fill_inode(struct inode *inode, struct page *locked_page, issued |= __ceph_caps_dirty(ci); new_issued = ~issued & info_caps; - /* update inode */ - inode->i_rdev = le32_to_cpu(info->rdev); /* directories have fl_stripe_unit set to zero */ if (le32_to_cpu(info->layout.fl_stripe_unit)) inode->i_blkbits = @@ -818,7 +867,7 @@ static int fill_inode(struct inode *inode, struct page *locked_page, if ((new_version || (new_issued & CEPH_CAP_AUTH_SHARED)) && (issued & CEPH_CAP_AUTH_EXCL) == 0) { - inode->i_mode = le32_to_cpu(info->mode); + inode->i_mode = mode; inode->i_uid = make_kuid(&init_user_ns, le32_to_cpu(info->uid)); inode->i_gid = make_kgid(&init_user_ns, le32_to_cpu(info->gid)); dout("%p mode 0%o uid.gid %d.%d\n", inode, inode->i_mode, @@ -884,6 +933,7 @@ static int fill_inode(struct inode *inode, struct page *locked_page, ci->i_rfiles = le64_to_cpu(info->rfiles); ci->i_rsubdirs = le64_to_cpu(info->rsubdirs); ci->i_dir_pin = iinfo->dir_pin; + ci->i_rsnaps = iinfo->rsnaps; ceph_decode_timespec64(&ci->i_rctime, &info->rctime); } } @@ -916,7 +966,7 @@ static int fill_inode(struct inode *inode, struct page *locked_page, case S_IFCHR: case S_IFSOCK: inode->i_blkbits = PAGE_SHIFT; - init_special_inode(inode, inode->i_mode, inode->i_rdev); + init_special_inode(inode, inode->i_mode, rdev); inode->i_op = &ceph_file_iops; break; case S_IFREG: @@ -932,8 +982,9 @@ static int fill_inode(struct inode *inode, struct page *locked_page, spin_unlock(&ci->i_ceph_lock); if (symlen != i_size_read(inode)) { - pr_err("fill_inode %llx.%llx BAD symlink " - "size %lld\n", ceph_vinop(inode), + pr_err("%s %llx.%llx BAD symlink " + "size %lld\n", __func__, + ceph_vinop(inode), i_size_read(inode)); i_size_write(inode, symlen); inode->i_blocks = calc_inode_blocks(symlen); @@ -957,7 +1008,7 @@ static int fill_inode(struct inode *inode, struct page *locked_page, inode->i_fop = &ceph_dir_fops; break; default: - pr_err("fill_inode %llx.%llx BAD mode 0%o\n", + pr_err("%s %llx.%llx BAD mode 0%o\n", __func__, ceph_vinop(inode), inode->i_mode); } @@ -966,7 +1017,7 @@ static int fill_inode(struct inode *inode, struct page *locked_page, if (ceph_snap(inode) == CEPH_NOSNAP) { ceph_add_cap(inode, session, le64_to_cpu(info->cap.cap_id), - cap_fmode, info_caps, + info_caps, le32_to_cpu(info->cap.wanted), le32_to_cpu(info->cap.seq), le32_to_cpu(info->cap.mseq), @@ -991,26 +1042,29 @@ static int fill_inode(struct inode *inode, struct page *locked_page, dout(" %p got snap_caps %s\n", inode, ceph_cap_string(info_caps)); ci->i_snap_caps |= info_caps; - if (cap_fmode >= 0) - __ceph_get_fmode(ci, cap_fmode); } - } else if (cap_fmode >= 0) { - pr_warn("mds issued no caps on %llx.%llx\n", - ceph_vinop(inode)); - __ceph_get_fmode(ci, cap_fmode); } if (iinfo->inline_version > 0 && iinfo->inline_version >= ci->i_inline_version) { int cache_caps = CEPH_CAP_FILE_CACHE | CEPH_CAP_FILE_LAZYIO; ci->i_inline_version = iinfo->inline_version; - if (ci->i_inline_version != CEPH_INLINE_NONE && + if (ceph_has_inline_data(ci) && (locked_page || (info_caps & cache_caps))) fill_inline = true; } + if (cap_fmode >= 0) { + if (!info_caps) + pr_warn("mds issued no caps on %llx.%llx\n", + ceph_vinop(inode)); + __ceph_touch_fmode(ci, mdsc, cap_fmode); + } + spin_unlock(&ci->i_ceph_lock); + ceph_fscache_register_inode_cookie(inode); + if (fill_inline) ceph_fill_inline_data(inode, locked_page, iinfo->inline_data, iinfo->inline_len); @@ -1050,6 +1104,7 @@ static void __update_dentry_lease(struct inode *dir, struct dentry *dentry, struct ceph_mds_session **old_lease_session) { struct ceph_dentry_info *di = ceph_dentry(dentry); + unsigned mask = le16_to_cpu(lease->mask); long unsigned duration = le32_to_cpu(lease->duration_ms); long unsigned ttl = from_time + (duration * HZ) / 1000; long unsigned half_ttl = from_time + (duration * HZ / 2) / 1000; @@ -1061,13 +1116,18 @@ static void __update_dentry_lease(struct inode *dir, struct dentry *dentry, if (ceph_snap(dir) != CEPH_NOSNAP) return; + if (mask & CEPH_LEASE_PRIMARY_LINK) + di->flags |= CEPH_DENTRY_PRIMARY_LINK; + else + di->flags &= ~CEPH_DENTRY_PRIMARY_LINK; + di->lease_shared_gen = atomic_read(&ceph_inode(dir)->i_shared_gen); - if (duration == 0) { + if (!(mask & CEPH_LEASE_VALID)) { __ceph_dentry_dir_lease_touch(di); return; } - if (di->lease_gen == session->s_cap_gen && + if (di->lease_gen == atomic_read(&session->s_cap_gen) && time_before(ttl, di->time)) return; /* we already have a newer lease. */ @@ -1078,7 +1138,7 @@ static void __update_dentry_lease(struct inode *dir, struct dentry *dentry, if (!di->lease_session) di->lease_session = ceph_get_mds_session(session); - di->lease_gen = session->s_cap_gen; + di->lease_gen = atomic_read(&session->s_cap_gen); di->lease_seq = le32_to_cpu(lease->seq); di->lease_renew_after = half_ttl; di->lease_renew_from = 0; @@ -1097,8 +1157,7 @@ static inline void update_dentry_lease(struct inode *dir, struct dentry *dentry, __update_dentry_lease(dir, dentry, lease, session, from_time, &old_lease_session); spin_unlock(&dentry->d_lock); - if (old_lease_session) - ceph_put_mds_session(old_lease_session); + ceph_put_mds_session(old_lease_session); } /* @@ -1143,13 +1202,12 @@ static void update_dentry_lease_careful(struct dentry *dentry, from_time, &old_lease_session); out_unlock: spin_unlock(&dentry->d_lock); - if (old_lease_session) - ceph_put_mds_session(old_lease_session); + ceph_put_mds_session(old_lease_session); } /* * splice a dentry to an inode. - * caller must hold directory i_mutex for this to be safe. + * caller must hold directory i_rwsem for this to be safe. */ static int splice_dentry(struct dentry **pdn, struct inode *in) { @@ -1239,10 +1297,9 @@ int ceph_fill_trace(struct super_block *sb, struct ceph_mds_request *req) struct inode *dir = req->r_parent; if (dir) { - err = fill_inode(dir, NULL, - &rinfo->diri, rinfo->dirfrag, - session, -1, - &req->r_caps_reservation); + err = ceph_fill_inode(dir, NULL, &rinfo->diri, + rinfo->dirfrag, session, -1, + &req->r_caps_reservation); if (err < 0) goto done; } else { @@ -1298,28 +1355,26 @@ retry_lookup: } if (rinfo->head->is_target) { - tvino.ino = le64_to_cpu(rinfo->targeti.in->ino); - tvino.snap = le64_to_cpu(rinfo->targeti.in->snapid); + /* Should be filled in by handle_reply */ + BUG_ON(!req->r_target_inode); - in = ceph_get_inode(sb, tvino); - if (IS_ERR(in)) { - err = PTR_ERR(in); - goto done; - } - - err = fill_inode(in, req->r_locked_page, &rinfo->targeti, NULL, - session, + in = req->r_target_inode; + err = ceph_fill_inode(in, req->r_locked_page, &rinfo->targeti, + NULL, session, (!test_bit(CEPH_MDS_R_ABORTED, &req->r_req_flags) && + !test_bit(CEPH_MDS_R_ASYNC, &req->r_req_flags) && rinfo->head->result == 0) ? req->r_fmode : -1, &req->r_caps_reservation); if (err < 0) { - pr_err("fill_inode badness %p %llx.%llx\n", + pr_err("ceph_fill_inode badness %p %llx.%llx\n", in, ceph_vinop(in)); + req->r_target_inode = NULL; if (in->i_state & I_NEW) discard_new_inode(in); + else + iput(in); goto done; } - req->r_target_inode = in; if (in->i_state & I_NEW) unlock_new_inode(in); } @@ -1411,10 +1466,12 @@ retry_lookup: } else if (have_lease) { if (d_unhashed(dn)) d_add(dn, NULL); + } + + if (!d_unhashed(dn) && have_lease) update_dentry_lease(dir, dn, rinfo->dlease, session, req->r_request_started); - } goto done; } @@ -1500,10 +1557,11 @@ 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, &rde->inode, NULL, session, - -1, &req->r_caps_reservation); + rc = ceph_fill_inode(in, NULL, &rde->inode, NULL, session, + -1, &req->r_caps_reservation); if (rc < 0) { - pr_err("fill_inode badness on %p got %d\n", in, rc); + pr_err("ceph_fill_inode badness on %p got %d\n", + in, rc); err = rc; if (in->i_state & I_NEW) { ihold(in); @@ -1513,8 +1571,7 @@ static int readdir_prepopulate_inodes_only(struct ceph_mds_request *req, unlock_new_inode(in); } - /* avoid calling iput_final() in mds dispatch threads */ - ceph_async_iput(in); + iput(in); } return err; @@ -1549,7 +1606,7 @@ static int fill_readdir_cache(struct inode *dir, struct dentry *dn, return idx == 0 ? -ENOMEM : 0; } /* reading/filling the cache are serialized by - * i_mutex, no need to use page lock */ + * i_rwsem, no need to use page lock */ unlock_page(ctl->page); ctl->dentries = kmap(ctl->page); if (idx == 0) @@ -1578,8 +1635,7 @@ int ceph_readdir_prepopulate(struct ceph_mds_request *req, struct dentry *dn; struct inode *in; int err = 0, skipped = 0, ret, i; - struct ceph_mds_request_head *rhead = req->r_request->front.iov_base; - u32 frag = le32_to_cpu(rhead->args.readdir.frag); + u32 frag = le32_to_cpu(req->r_args.readdir.frag); u32 last_hash = 0; u32 fpos_offset; struct ceph_readdir_cache_control cache_ctl = {}; @@ -1596,7 +1652,7 @@ int ceph_readdir_prepopulate(struct ceph_mds_request *req, } else if (rinfo->offset_hash) { /* mds understands offset_hash */ WARN_ON_ONCE(req->r_readdir_offset != 2); - last_hash = le32_to_cpu(rhead->args.readdir.offset_hash); + last_hash = le32_to_cpu(req->r_args.readdir.offset_hash); } } @@ -1707,18 +1763,16 @@ retry_lookup: } } - ret = fill_inode(in, NULL, &rde->inode, NULL, session, - -1, &req->r_caps_reservation); + ret = ceph_fill_inode(in, NULL, &rde->inode, NULL, session, + -1, &req->r_caps_reservation); if (ret < 0) { - pr_err("fill_inode badness on %p\n", in); + pr_err("ceph_fill_inode badness on %p\n", in); if (d_really_is_negative(dn)) { - /* avoid calling iput_final() in mds - * dispatch threads */ if (in->i_state & I_NEW) { ihold(in); discard_new_inode(in); } - ceph_async_iput(in); + iput(in); } d_drop(dn); err = ret; @@ -1731,7 +1785,7 @@ retry_lookup: if (ceph_security_xattr_deadlock(in)) { dout(" skip splicing dn %p to inode %p" " (security xattr deadlock)\n", dn, in); - ceph_async_iput(in); + iput(in); skipped++; goto next_item; } @@ -1772,89 +1826,29 @@ bool ceph_inode_set_size(struct inode *inode, loff_t size) bool ret; spin_lock(&ci->i_ceph_lock); - dout("set_size %p %llu -> %llu\n", inode, inode->i_size, size); + dout("set_size %p %llu -> %llu\n", inode, i_size_read(inode), size); i_size_write(inode, size); + ceph_fscache_update(inode); inode->i_blocks = calc_inode_blocks(size); ret = __ceph_should_report_size(ci); spin_unlock(&ci->i_ceph_lock); - return ret; -} -/* - * Put reference to inode, but avoid calling iput_final() in current thread. - * iput_final() may wait for reahahead pages. The wait can cause deadlock in - * some contexts. - */ -void ceph_async_iput(struct inode *inode) -{ - if (!inode) - return; - for (;;) { - if (atomic_add_unless(&inode->i_count, -1, 1)) - break; - if (queue_work(ceph_inode_to_client(inode)->inode_wq, - &ceph_inode(inode)->i_work)) - break; - /* queue work failed, i_count must be at least 2 */ - } -} - -/* - * Write back inode data in a worker thread. (This can't be done - * in the message handler context.) - */ -void ceph_queue_writeback(struct inode *inode) -{ - struct ceph_inode_info *ci = ceph_inode(inode); - set_bit(CEPH_I_WORK_WRITEBACK, &ci->i_work_mask); - - ihold(inode); - if (queue_work(ceph_inode_to_client(inode)->inode_wq, - &ci->i_work)) { - dout("ceph_queue_writeback %p\n", inode); - } else { - dout("ceph_queue_writeback %p already queued, mask=%lx\n", - inode, ci->i_work_mask); - iput(inode); - } -} - -/* - * queue an async invalidation - */ -void ceph_queue_invalidate(struct inode *inode) -{ - struct ceph_inode_info *ci = ceph_inode(inode); - set_bit(CEPH_I_WORK_INVALIDATE_PAGES, &ci->i_work_mask); - - ihold(inode); - if (queue_work(ceph_inode_to_client(inode)->inode_wq, - &ceph_inode(inode)->i_work)) { - dout("ceph_queue_invalidate %p\n", inode); - } else { - dout("ceph_queue_invalidate %p already queued, mask=%lx\n", - inode, ci->i_work_mask); - iput(inode); - } + return ret; } -/* - * Queue an async vmtruncate. If we fail to queue work, we will handle - * the truncation the next time we call __ceph_do_pending_vmtruncate. - */ -void ceph_queue_vmtruncate(struct inode *inode) +void ceph_queue_inode_work(struct inode *inode, int work_bit) { + struct ceph_fs_client *fsc = ceph_inode_to_client(inode); struct ceph_inode_info *ci = ceph_inode(inode); - set_bit(CEPH_I_WORK_VMTRUNCATE, &ci->i_work_mask); + set_bit(work_bit, &ci->i_work_mask); ihold(inode); - if (queue_work(ceph_inode_to_client(inode)->inode_wq, - &ci->i_work)) { - dout("ceph_queue_vmtruncate %p\n", inode); + if (queue_work(fsc->inode_wq, &ci->i_work)) { + dout("queue_inode_work %p, mask=%lx\n", inode, ci->i_work_mask); } else { - dout("ceph_queue_vmtruncate %p already queued, mask=%lx\n", + dout("queue_inode_work %p already queued, mask=%lx\n", inode, ci->i_work_mask); iput(inode); } @@ -1863,15 +1857,16 @@ void ceph_queue_vmtruncate(struct inode *inode) static void ceph_do_invalidate_pages(struct inode *inode) { struct ceph_inode_info *ci = ceph_inode(inode); - struct ceph_fs_client *fsc = ceph_inode_to_client(inode); u32 orig_gen; int check = 0; + ceph_fscache_invalidate(inode, false); + mutex_lock(&ci->i_truncate_mutex); - if (READ_ONCE(fsc->mount_state) == CEPH_MOUNT_SHUTDOWN) { - pr_warn_ratelimited("invalidate_pages %p %lld forced umount\n", - inode, ceph_ino(inode)); + if (ceph_inode_is_shutdown(inode)) { + pr_warn_ratelimited("%s: inode %llx.%llx is shut down\n", + __func__, ceph_vinop(inode)); mapping_set_error(inode->i_mapping, -EIO); truncate_pagecache(inode, 0); mutex_unlock(&ci->i_truncate_mutex); @@ -1892,7 +1887,8 @@ static void ceph_do_invalidate_pages(struct inode *inode) spin_unlock(&ci->i_ceph_lock); if (invalidate_inode_pages2(inode->i_mapping) < 0) { - pr_err("invalidate_pages %p fails\n", inode); + pr_err("invalidate_inode_pages2 %llx.%llx failed\n", + ceph_vinop(inode)); } spin_lock(&ci->i_ceph_lock); @@ -1958,6 +1954,7 @@ retry: ci->i_truncate_pending, to); spin_unlock(&ci->i_ceph_lock); + ceph_fscache_resize(inode, to); truncate_pagecache(inode, to); spin_lock(&ci->i_ceph_lock); @@ -1972,7 +1969,7 @@ retry: mutex_unlock(&ci->i_truncate_mutex); if (wrbuffer_refs == 0) - ceph_check_caps(ci, CHECK_CAPS_AUTHONLY, NULL); + ceph_check_caps(ci, 0, NULL); wake_up_all(&ci->i_cap_wq); } @@ -1981,7 +1978,7 @@ static void ceph_inode_work(struct work_struct *work) { struct ceph_inode_info *ci = container_of(work, struct ceph_inode_info, i_work); - struct inode *inode = &ci->vfs_inode; + struct inode *inode = &ci->netfs.inode; if (test_and_clear_bit(CEPH_I_WORK_WRITEBACK, &ci->i_work_mask)) { dout("writeback %p\n", inode); @@ -1993,6 +1990,12 @@ static void ceph_inode_work(struct work_struct *work) if (test_and_clear_bit(CEPH_I_WORK_VMTRUNCATE, &ci->i_work_mask)) __ceph_do_pending_vmtruncate(inode); + if (test_and_clear_bit(CEPH_I_WORK_CHECK_CAPS, &ci->i_work_mask)) + ceph_check_caps(ci, 0, NULL); + + if (test_and_clear_bit(CEPH_I_WORK_FLUSH_SNAPS, &ci->i_work_mask)) + ceph_flush_snaps(ci, NULL); + iput(inode); } @@ -2115,20 +2118,21 @@ int __ceph_setattr(struct inode *inode, struct iattr *attr) } } if (ia_valid & ATTR_SIZE) { - dout("setattr %p size %lld -> %lld\n", inode, - inode->i_size, attr->ia_size); - if ((issued & CEPH_CAP_FILE_EXCL) && - attr->ia_size > inode->i_size) { - i_size_write(inode, attr->ia_size); - inode->i_blocks = calc_inode_blocks(attr->ia_size); - ci->i_reported_size = attr->ia_size; - dirtied |= CEPH_CAP_FILE_EXCL; - ia_valid |= ATTR_MTIME; + loff_t isize = i_size_read(inode); + + dout("setattr %p size %lld -> %lld\n", inode, isize, attr->ia_size); + if ((issued & CEPH_CAP_FILE_EXCL) && attr->ia_size >= isize) { + if (attr->ia_size > isize) { + i_size_write(inode, attr->ia_size); + inode->i_blocks = calc_inode_blocks(attr->ia_size); + ci->i_reported_size = attr->ia_size; + dirtied |= CEPH_CAP_FILE_EXCL; + ia_valid |= ATTR_MTIME; + } } else if ((issued & CEPH_CAP_FILE_SHARED) == 0 || - attr->ia_size != inode->i_size) { + attr->ia_size != isize) { req->r_args.setattr.size = cpu_to_le64(attr->ia_size); - req->r_args.setattr.old_size = - cpu_to_le64(inode->i_size); + req->r_args.setattr.old_size = cpu_to_le64(isize); mask |= CEPH_SETATTR_SIZE; release |= CEPH_CAP_FILE_SHARED | CEPH_CAP_FILE_EXCL | CEPH_CAP_FILE_RD | CEPH_CAP_FILE_WR; @@ -2188,6 +2192,7 @@ int __ceph_setattr(struct inode *inode, struct iattr *attr) inode_dirty_flags = __ceph_mark_dirty_caps(ci, dirtied, &prealloc_cf); inode->i_ctime = attr->ia_ctime; + inode_inc_iversion_raw(inode); } release &= issued; @@ -2198,7 +2203,6 @@ int __ceph_setattr(struct inode *inode, struct iattr *attr) if (inode_dirty_flags) __mark_inode_dirty(inode, inode_dirty_flags); - if (mask) { req->r_inode = inode; ihold(inode); @@ -2223,7 +2227,8 @@ int __ceph_setattr(struct inode *inode, struct iattr *attr) /* * setattr */ -int ceph_setattr(struct dentry *dentry, struct iattr *attr) +int ceph_setattr(struct user_namespace *mnt_userns, struct dentry *dentry, + struct iattr *attr) { struct inode *inode = d_inode(dentry); struct ceph_fs_client *fsc = ceph_inode_to_client(inode); @@ -2232,12 +2237,15 @@ int ceph_setattr(struct dentry *dentry, struct iattr *attr) if (ceph_snap(inode) != CEPH_NOSNAP) return -EROFS; - err = setattr_prepare(dentry, attr); + if (ceph_inode_is_shutdown(inode)) + return -ESTALE; + + err = setattr_prepare(&init_user_ns, dentry, attr); if (err != 0) return err; if ((attr->ia_valid & ATTR_SIZE) && - attr->ia_size > max(inode->i_size, fsc->max_file_size)) + attr->ia_size > max(i_size_read(inode), fsc->max_file_size)) return -EFBIG; if ((attr->ia_valid & ATTR_SIZE) && @@ -2247,11 +2255,41 @@ int ceph_setattr(struct dentry *dentry, struct iattr *attr) err = __ceph_setattr(inode, attr); if (err >= 0 && (attr->ia_valid & ATTR_MODE)) - err = posix_acl_chmod(inode, attr->ia_mode); + err = posix_acl_chmod(&init_user_ns, inode, attr->ia_mode); return err; } +int ceph_try_to_choose_auth_mds(struct inode *inode, int mask) +{ + int issued = ceph_caps_issued(ceph_inode(inode)); + + /* + * If any 'x' caps is issued we can just choose the auth MDS + * instead of the random replica MDSes. Because only when the + * Locker is in LOCK_EXEC state will the loner client could + * get the 'x' caps. And if we send the getattr requests to + * any replica MDS it must auth pin and tries to rdlock from + * the auth MDS, and then the auth MDS need to do the Locker + * state transition to LOCK_SYNC. And after that the lock state + * will change back. + * + * This cost much when doing the Locker state transition and + * usually will need to revoke caps from clients. + * + * And for the 'Xs' caps for getxattr we will also choose the + * auth MDS, because the MDS side code is buggy due to setxattr + * won't notify the replica MDSes when the values changed and + * the replica MDS will return the old values. Though we will + * fix it in MDS code, but this still makes sense for old ceph. + */ + if (((mask & CEPH_CAP_ANY_SHARED) && (issued & CEPH_CAP_ANY_EXCL)) + || (mask & (CEPH_STAT_RSTAT | CEPH_STAT_CAP_XATTR))) + return USE_AUTH_MDS; + else + return USE_ANY_MDS; +} + /* * Verify that we have a lease on the given mask. If not, * do a getattr against an mds. @@ -2272,10 +2310,10 @@ int __ceph_do_getattr(struct inode *inode, struct page *locked_page, dout("do_getattr inode %p mask %s mode 0%o\n", inode, ceph_cap_string(mask), inode->i_mode); - if (!force && ceph_caps_issued_mask(ceph_inode(inode), mask, 1)) - return 0; + if (!force && ceph_caps_issued_mask_metric(ceph_inode(inode), mask, 1)) + return 0; - mode = (mask & CEPH_STAT_RSTAT) ? USE_AUTH_MDS : USE_ANY_MDS; + mode = ceph_try_to_choose_auth_mds(inode, mask); req = ceph_mdsc_create_request(mdsc, CEPH_MDS_OP_GETATTR, mode); if (IS_ERR(req)) return PTR_ERR(req); @@ -2290,7 +2328,8 @@ int __ceph_do_getattr(struct inode *inode, struct page *locked_page, if (inline_version == 0) { /* the reply is supposed to contain inline data */ err = -EINVAL; - } else if (inline_version == CEPH_INLINE_NONE) { + } else if (inline_version == CEPH_INLINE_NONE || + inline_version == 1) { err = -ENODATA; } else { err = req->r_reply_info.targeti.inline_len; @@ -2301,12 +2340,65 @@ int __ceph_do_getattr(struct inode *inode, struct page *locked_page, return err; } +int ceph_do_getvxattr(struct inode *inode, const char *name, void *value, + size_t size) +{ + 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 mode = USE_AUTH_MDS; + int err; + char *xattr_value; + size_t xattr_value_len; + + req = ceph_mdsc_create_request(mdsc, CEPH_MDS_OP_GETVXATTR, mode); + if (IS_ERR(req)) { + err = -ENOMEM; + goto out; + } + + req->r_feature_needed = CEPHFS_FEATURE_OP_GETVXATTR; + req->r_path2 = kstrdup(name, GFP_NOFS); + if (!req->r_path2) { + err = -ENOMEM; + goto put; + } + + ihold(inode); + req->r_inode = inode; + err = ceph_mdsc_do_request(mdsc, NULL, req); + if (err < 0) + goto put; + + xattr_value = req->r_reply_info.xattr_info.xattr_value; + xattr_value_len = req->r_reply_info.xattr_info.xattr_value_len; + + dout("do_getvxattr xattr_value_len:%zu, size:%zu\n", xattr_value_len, size); + + err = (int)xattr_value_len; + if (size == 0) + goto put; + + if (xattr_value_len > size) { + err = -ERANGE; + goto put; + } + + memcpy(value, xattr_value, xattr_value_len); +put: + ceph_mdsc_put_request(req); +out: + dout("do_getvxattr result=%d\n", err); + return err; +} + /* * Check inode permissions. We verify we have a valid value for * the AUTH cap, then call the generic handler. */ -int ceph_permission(struct inode *inode, int mask) +int ceph_permission(struct user_namespace *mnt_userns, struct inode *inode, + int mask) { int err; @@ -2316,20 +2408,28 @@ int ceph_permission(struct inode *inode, int mask) err = ceph_do_getattr(inode, CEPH_CAP_AUTH_SHARED, false); if (!err) - err = generic_permission(inode, mask); + err = generic_permission(&init_user_ns, inode, mask); return err; } /* Craft a mask of needed caps given a set of requested statx attrs. */ -static int statx_to_caps(u32 want) +static int statx_to_caps(u32 want, umode_t mode) { int mask = 0; if (want & (STATX_MODE|STATX_UID|STATX_GID|STATX_CTIME|STATX_BTIME)) mask |= CEPH_CAP_AUTH_SHARED; - if (want & (STATX_NLINK|STATX_CTIME)) - mask |= CEPH_CAP_LINK_SHARED; + if (want & (STATX_NLINK|STATX_CTIME)) { + /* + * The link count for directories depends on inode->i_subdirs, + * and that is only updated when Fs caps are held. + */ + if (S_ISDIR(mode)) + mask |= CEPH_CAP_FILE_SHARED; + else + mask |= CEPH_CAP_LINK_SHARED; + } if (want & (STATX_ATIME|STATX_MTIME|STATX_CTIME|STATX_SIZE| STATX_BLOCKS)) @@ -2345,24 +2445,29 @@ static int statx_to_caps(u32 want) * Get all the attributes. If we have sufficient caps for the requested attrs, * then we can avoid talking to the MDS at all. */ -int ceph_getattr(const struct path *path, struct kstat *stat, - u32 request_mask, unsigned int flags) +int ceph_getattr(struct user_namespace *mnt_userns, const struct path *path, + struct kstat *stat, u32 request_mask, unsigned int flags) { struct inode *inode = d_inode(path->dentry); + struct super_block *sb = inode->i_sb; struct ceph_inode_info *ci = ceph_inode(inode); u32 valid_mask = STATX_BASIC_STATS; int err = 0; + if (ceph_inode_is_shutdown(inode)) + return -ESTALE; + /* Skip the getattr altogether if we're asked not to sync */ - if (!(flags & AT_STATX_DONT_SYNC)) { - err = ceph_do_getattr(inode, statx_to_caps(request_mask), - flags & AT_STATX_FORCE_SYNC); + if ((flags & AT_STATX_SYNC_TYPE) != AT_STATX_DONT_SYNC) { + err = ceph_do_getattr(inode, + statx_to_caps(request_mask, inode->i_mode), + flags & AT_STATX_FORCE_SYNC); if (err) return err; } - generic_fillattr(inode, stat); - stat->ino = ceph_translate_ino(inode->i_sb, inode->i_ino); + generic_fillattr(&init_user_ns, inode, stat); + stat->ino = ceph_present_inode(inode); /* * btime on newly-allocated inodes is 0, so if this is still set to @@ -2374,16 +2479,34 @@ int ceph_getattr(const struct path *path, struct kstat *stat, } if (ceph_snap(inode) == CEPH_NOSNAP) - stat->dev = inode->i_sb->s_dev; + stat->dev = sb->s_dev; else stat->dev = ci->i_snapid_map ? ci->i_snapid_map->dev : 0; if (S_ISDIR(inode->i_mode)) { - if (ceph_test_mount_opt(ceph_sb_to_client(inode->i_sb), - RBYTES)) + if (ceph_test_mount_opt(ceph_sb_to_client(sb), RBYTES)) { stat->size = ci->i_rbytes; - else + } else if (ceph_snap(inode) == CEPH_SNAPDIR) { + struct ceph_inode_info *pci; + struct ceph_snap_realm *realm; + struct inode *parent; + + parent = ceph_lookup_inode(sb, ceph_ino(inode)); + if (!parent) + return PTR_ERR(parent); + + pci = ceph_inode(parent); + spin_lock(&pci->i_ceph_lock); + realm = pci->i_snap_realm; + if (realm) + stat->size = realm->num_snaps; + else + stat->size = 0; + spin_unlock(&pci->i_ceph_lock); + iput(parent); + } else { stat->size = ci->i_files + ci->i_subdirs; + } stat->blocks = 0; stat->blksize = 65536; /* @@ -2399,3 +2522,27 @@ int ceph_getattr(const struct path *path, struct kstat *stat, stat->result_mask = request_mask & valid_mask; return err; } + +void ceph_inode_shutdown(struct inode *inode) +{ + struct ceph_inode_info *ci = ceph_inode(inode); + struct rb_node *p; + int iputs = 0; + bool invalidate = false; + + spin_lock(&ci->i_ceph_lock); + ci->i_ceph_flags |= CEPH_I_SHUTDOWN; + p = rb_first(&ci->i_caps); + while (p) { + struct ceph_cap *cap = rb_entry(p, struct ceph_cap, ci_node); + + p = rb_next(p); + iputs += ceph_purge_inode_cap(inode, cap, &invalidate); + } + spin_unlock(&ci->i_ceph_lock); + + if (invalidate) + ceph_queue_invalidate(inode); + while (iputs--) + iput(inode); +} diff --git a/fs/ceph/io.c b/fs/ceph/io.c index 97602ea92ff4..c456509b31c3 100644 --- a/fs/ceph/io.c +++ b/fs/ceph/io.c @@ -118,7 +118,7 @@ static void ceph_block_buffered(struct ceph_inode_info *ci, struct inode *inode) } /** - * ceph_end_io_direct - declare the file is being used for direct i/o + * ceph_start_io_direct - declare the file is being used for direct i/o * @inode: file inode * * Declare that a direct I/O operation is about to start, and ensure diff --git a/fs/ceph/ioctl.c b/fs/ceph/ioctl.c index c90f03beb15d..6e061bf62ad4 100644 --- a/fs/ceph/ioctl.c +++ b/fs/ceph/ioctl.c @@ -243,11 +243,13 @@ static long ceph_ioctl_lazyio(struct file *file) struct ceph_file_info *fi = file->private_data; struct inode *inode = file_inode(file); struct ceph_inode_info *ci = ceph_inode(inode); + struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc; if ((fi->fmode & CEPH_FILE_MODE_LAZY) == 0) { spin_lock(&ci->i_ceph_lock); fi->fmode |= CEPH_FILE_MODE_LAZY; ci->i_nr_by_mode[ffs(CEPH_FILE_MODE_LAZY)]++; + __ceph_touch_fmode(ci, mdsc, fi->fmode); spin_unlock(&ci->i_ceph_lock); dout("ioctl_layzio: file %p marked lazy\n", file); diff --git a/fs/ceph/locks.c b/fs/ceph/locks.c index 544e9e85b120..3e2843e86e27 100644 --- a/fs/ceph/locks.c +++ b/fs/ceph/locks.c @@ -57,13 +57,13 @@ static const struct file_lock_operations ceph_fl_lock_ops = { .fl_release_private = ceph_fl_release_lock, }; -/** +/* * Implement fcntl and flock locking functions. */ static int ceph_lock_message(u8 lock_type, u16 operation, struct inode *inode, int cmd, u8 wait, struct file_lock *fl) { - struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; + struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(inode->i_sb); struct ceph_mds_request *req; int err; u64 length = 0; @@ -111,10 +111,10 @@ static int ceph_lock_message(u8 lock_type, u16 operation, struct inode *inode, req->r_args.filelock_change.length = cpu_to_le64(length); req->r_args.filelock_change.wait = wait; - if (wait) - req->r_wait_for_completion = ceph_lock_wait_for_completion; - - err = ceph_mdsc_do_request(mdsc, inode, req); + err = ceph_mdsc_submit_request(mdsc, inode, req); + if (!err) + err = ceph_mdsc_wait_request(mdsc, req, wait ? + ceph_lock_wait_for_completion : NULL); if (!err && operation == CEPH_MDS_OP_GETFILELOCK) { fl->fl_pid = -le64_to_cpu(req->r_reply_info.filelock_reply->pid); if (CEPH_LOCK_SHARED == req->r_reply_info.filelock_reply->type) @@ -210,7 +210,22 @@ static int ceph_lock_wait_for_completion(struct ceph_mds_client *mdsc, return 0; } -/** +static int try_unlock_file(struct file *file, struct file_lock *fl) +{ + int err; + unsigned int orig_flags = fl->fl_flags; + fl->fl_flags |= FL_EXISTS; + err = locks_lock_file_wait(file, fl); + fl->fl_flags = orig_flags; + if (err == -ENOENT) { + if (!(orig_flags & FL_EXISTS)) + err = 0; + return err; + } + return 1; +} + +/* * Attempt to set an fcntl lock. * For now, this just goes away to the server. Later it may be more awesome. */ @@ -225,9 +240,9 @@ int ceph_lock(struct file *file, int cmd, struct file_lock *fl) if (!(fl->fl_flags & FL_POSIX)) return -ENOLCK; - /* No mandatory locks */ - if (__mandatory_lock(file->f_mapping->host) && fl->fl_type != F_UNLCK) - return -ENOLCK; + + if (ceph_inode_is_shutdown(inode)) + return -ESTALE; dout("ceph_lock, fl_owner: %p\n", fl->fl_owner); @@ -255,9 +270,15 @@ int ceph_lock(struct file *file, int cmd, struct file_lock *fl) else lock_cmd = CEPH_LOCK_UNLOCK; + if (op == CEPH_MDS_OP_SETFILELOCK && F_UNLCK == fl->fl_type) { + err = try_unlock_file(file, fl); + if (err <= 0) + return err; + } + err = ceph_lock_message(CEPH_LOCK_FCNTL, op, inode, lock_cmd, wait, fl); if (!err) { - if (op == CEPH_MDS_OP_SETFILELOCK) { + if (op == CEPH_MDS_OP_SETFILELOCK && F_UNLCK != fl->fl_type) { dout("mds locked, locking locally\n"); err = posix_lock_file(file, fl, NULL); if (err) { @@ -284,9 +305,9 @@ int ceph_flock(struct file *file, int cmd, struct file_lock *fl) if (!(fl->fl_flags & FL_FLOCK)) return -ENOLCK; - /* No mandatory locks */ - if (fl->fl_type & LOCK_MAND) - return -EOPNOTSUPP; + + if (ceph_inode_is_shutdown(inode)) + return -ESTALE; dout("ceph_flock, fl_file: %p\n", fl->fl_file); @@ -311,9 +332,15 @@ int ceph_flock(struct file *file, int cmd, struct file_lock *fl) else lock_cmd = CEPH_LOCK_UNLOCK; + if (F_UNLCK == fl->fl_type) { + err = try_unlock_file(file, fl); + if (err <= 0) + return err; + } + err = ceph_lock_message(CEPH_LOCK_FLOCK, CEPH_MDS_OP_SETFILELOCK, inode, lock_cmd, wait, fl); - if (!err) { + if (!err && F_UNLCK != fl->fl_type) { err = locks_lock_file_wait(file, fl); if (err) { ceph_lock_message(CEPH_LOCK_FLOCK, @@ -381,7 +408,7 @@ static int lock_to_ceph_filelock(struct file_lock *lock, return err; } -/** +/* * Encode the flock and fcntl locks for the given inode into the ceph_filelock * array. Must be called with inode->i_lock already held. * If we encounter more of a specific lock type than expected, return -ENOSPC. @@ -431,7 +458,7 @@ fail: return err; } -/** +/* * Copy the encoded flock and fcntl locks into the pagelist. * Format is: #fcntl locks, sequential fcntl locks, #flock locks, * sequential flock locks. diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c index bbbbddf71326..26a0a8b9975e 100644 --- a/fs/ceph/mds_client.c +++ b/fs/ceph/mds_client.c @@ -10,6 +10,8 @@ #include <linux/seq_file.h> #include <linux/ratelimit.h> #include <linux/bits.h> +#include <linux/ktime.h> +#include <linux/bitmap.h> #include "super.h" #include "mds_client.h" @@ -175,6 +177,13 @@ static int parse_reply_info_in(void **p, void *end, memset(&info->snap_btime, 0, sizeof(info->snap_btime)); } + /* snapshot count, remains zero for v<=3 */ + if (struct_v >= 4) { + ceph_decode_64_safe(p, end, info->rsnaps, bad); + } else { + info->rsnaps = 0; + } + *p = end; } else { if (features & CEPH_FEATURE_MDS_INLINE_DATA) { @@ -213,7 +222,7 @@ static int parse_reply_info_in(void **p, void *end, } info->dir_pin = -ENODATA; - /* info->snap_btime remains zero */ + /* info->snap_btime and info->rsnaps remain zero */ } return 0; bad: @@ -415,21 +424,124 @@ bad: return -EIO; } + +#if BITS_PER_LONG == 64 + +#define DELEGATED_INO_AVAILABLE xa_mk_value(1) + +static int ceph_parse_deleg_inos(void **p, void *end, + struct ceph_mds_session *s) +{ + u32 sets; + + ceph_decode_32_safe(p, end, sets, bad); + dout("got %u sets of delegated inodes\n", sets); + while (sets--) { + u64 start, len; + + ceph_decode_64_safe(p, end, start, bad); + ceph_decode_64_safe(p, end, len, bad); + + /* Don't accept a delegation of system inodes */ + if (start < CEPH_INO_SYSTEM_BASE) { + pr_warn_ratelimited("ceph: ignoring reserved inode range delegation (start=0x%llx len=0x%llx)\n", + start, len); + continue; + } + while (len--) { + int err = xa_insert(&s->s_delegated_inos, start++, + DELEGATED_INO_AVAILABLE, + GFP_KERNEL); + if (!err) { + dout("added delegated inode 0x%llx\n", + start - 1); + } else if (err == -EBUSY) { + pr_warn("MDS delegated inode 0x%llx more than once.\n", + start - 1); + } else { + return err; + } + } + } + return 0; +bad: + return -EIO; +} + +u64 ceph_get_deleg_ino(struct ceph_mds_session *s) +{ + unsigned long ino; + void *val; + + xa_for_each(&s->s_delegated_inos, ino, val) { + val = xa_erase(&s->s_delegated_inos, ino); + if (val == DELEGATED_INO_AVAILABLE) + return ino; + } + return 0; +} + +int ceph_restore_deleg_ino(struct ceph_mds_session *s, u64 ino) +{ + return xa_insert(&s->s_delegated_inos, ino, DELEGATED_INO_AVAILABLE, + GFP_KERNEL); +} +#else /* BITS_PER_LONG == 64 */ +/* + * FIXME: xarrays can't handle 64-bit indexes on a 32-bit arch. For now, just + * ignore delegated_inos on 32 bit arch. Maybe eventually add xarrays for top + * and bottom words? + */ +static int ceph_parse_deleg_inos(void **p, void *end, + struct ceph_mds_session *s) +{ + u32 sets; + + ceph_decode_32_safe(p, end, sets, bad); + if (sets) + ceph_decode_skip_n(p, end, sets * 2 * sizeof(__le64), bad); + return 0; +bad: + return -EIO; +} + +u64 ceph_get_deleg_ino(struct ceph_mds_session *s) +{ + return 0; +} + +int ceph_restore_deleg_ino(struct ceph_mds_session *s, u64 ino) +{ + return 0; +} +#endif /* BITS_PER_LONG == 64 */ + /* * parse create results */ static int parse_reply_info_create(void **p, void *end, struct ceph_mds_reply_info_parsed *info, - u64 features) + u64 features, struct ceph_mds_session *s) { + int ret; + if (features == (u64)-1 || (features & CEPH_FEATURE_REPLY_CREATE_INODE)) { - /* Malformed reply? */ if (*p == end) { + /* Malformed reply? */ info->has_create_ino = false; - } else { + } else if (test_bit(CEPHFS_FEATURE_DELEG_INO, &s->s_features)) { info->has_create_ino = true; + /* struct_v, struct_compat, and len */ + ceph_decode_skip_n(p, end, 2 + sizeof(u32), bad); ceph_decode_64_safe(p, end, info->ino, bad); + ret = ceph_parse_deleg_inos(p, end, s); + if (ret) + return ret; + } else { + /* legacy */ + ceph_decode_64_safe(p, end, info->ino, bad); + info->has_create_ino = true; } } else { if (*p != end) @@ -443,12 +555,34 @@ bad: return -EIO; } +static int parse_reply_info_getvxattr(void **p, void *end, + struct ceph_mds_reply_info_parsed *info, + u64 features) +{ + u32 value_len; + + ceph_decode_skip_8(p, end, bad); /* skip current version: 1 */ + ceph_decode_skip_8(p, end, bad); /* skip first version: 1 */ + ceph_decode_skip_32(p, end, bad); /* skip payload length */ + + ceph_decode_32_safe(p, end, value_len, bad); + + if (value_len == end - *p) { + info->xattr_info.xattr_value = *p; + info->xattr_info.xattr_value_len = value_len; + *p = end; + return value_len; + } +bad: + return -EIO; +} + /* * parse extra results */ static int parse_reply_info_extra(void **p, void *end, struct ceph_mds_reply_info_parsed *info, - u64 features) + u64 features, struct ceph_mds_session *s) { u32 op = le32_to_cpu(info->head->op); @@ -457,7 +591,9 @@ static int parse_reply_info_extra(void **p, void *end, else if (op == CEPH_MDS_OP_READDIR || op == CEPH_MDS_OP_LSSNAP) return parse_reply_info_readdir(p, end, info, features); else if (op == CEPH_MDS_OP_CREATE) - return parse_reply_info_create(p, end, info, features); + return parse_reply_info_create(p, end, info, features, s); + else if (op == CEPH_MDS_OP_GETVXATTR) + return parse_reply_info_getvxattr(p, end, info, features); else return -EIO; } @@ -465,7 +601,7 @@ static int parse_reply_info_extra(void **p, void *end, /* * parse entire mds reply */ -static int parse_reply_info(struct ceph_msg *msg, +static int parse_reply_info(struct ceph_mds_session *s, struct ceph_msg *msg, struct ceph_mds_reply_info_parsed *info, u64 features) { @@ -490,7 +626,7 @@ static int parse_reply_info(struct ceph_msg *msg, ceph_decode_32_safe(&p, end, len, bad); if (len > 0) { ceph_decode_need(&p, end, len, bad); - err = parse_reply_info_extra(&p, p+len, info, features); + err = parse_reply_info_extra(&p, p+len, info, features, s); if (err < 0) goto out_bad; } @@ -519,6 +655,79 @@ static void destroy_reply_info(struct ceph_mds_reply_info_parsed *info) free_pages((unsigned long)info->dir_entries, get_order(info->dir_buf_size)); } +/* + * In async unlink case the kclient won't wait for the first reply + * from MDS and just drop all the links and unhash the dentry and then + * succeeds immediately. + * + * For any new create/link/rename,etc requests followed by using the + * same file names we must wait for the first reply of the inflight + * unlink request, or the MDS possibly will fail these following + * requests with -EEXIST if the inflight async unlink request was + * delayed for some reasons. + * + * And the worst case is that for the none async openc request it will + * successfully open the file if the CDentry hasn't been unlinked yet, + * but later the previous delayed async unlink request will remove the + * CDenty. That means the just created file is possiblly deleted later + * by accident. + * + * We need to wait for the inflight async unlink requests to finish + * when creating new files/directories by using the same file names. + */ +int ceph_wait_on_conflict_unlink(struct dentry *dentry) +{ + struct ceph_fs_client *fsc = ceph_sb_to_client(dentry->d_sb); + struct dentry *pdentry = dentry->d_parent; + struct dentry *udentry, *found = NULL; + struct ceph_dentry_info *di; + struct qstr dname; + u32 hash = dentry->d_name.hash; + int err; + + dname.name = dentry->d_name.name; + dname.len = dentry->d_name.len; + + rcu_read_lock(); + hash_for_each_possible_rcu(fsc->async_unlink_conflict, di, + hnode, hash) { + udentry = di->dentry; + + spin_lock(&udentry->d_lock); + if (udentry->d_name.hash != hash) + goto next; + if (unlikely(udentry->d_parent != pdentry)) + goto next; + if (!hash_hashed(&di->hnode)) + goto next; + + if (!test_bit(CEPH_DENTRY_ASYNC_UNLINK_BIT, &di->flags)) + pr_warn("%s dentry %p:%pd async unlink bit is not set\n", + __func__, dentry, dentry); + + if (!d_same_name(udentry, pdentry, &dname)) + goto next; + + spin_unlock(&udentry->d_lock); + found = dget(udentry); + break; +next: + spin_unlock(&udentry->d_lock); + } + rcu_read_unlock(); + + if (likely(!found)) + return 0; + + dout("%s dentry %p:%pd conflict with old %p:%pd\n", __func__, + dentry, dentry, found, found); + + err = wait_on_bit(&di->flags, CEPH_DENTRY_ASYNC_UNLINK_BIT, + TASK_KILLABLE); + dput(found); + return err; +} + /* * sessions @@ -541,23 +750,21 @@ const char *ceph_session_state_name(int s) struct ceph_mds_session *ceph_get_mds_session(struct ceph_mds_session *s) { - if (refcount_inc_not_zero(&s->s_ref)) { - dout("mdsc get_session %p %d -> %d\n", s, - refcount_read(&s->s_ref)-1, refcount_read(&s->s_ref)); + if (refcount_inc_not_zero(&s->s_ref)) return s; - } else { - dout("mdsc get_session %p 0 -- FAIL\n", s); - return NULL; - } + return NULL; } void ceph_put_mds_session(struct ceph_mds_session *s) { - dout("mdsc put_session %p %d -> %d\n", s, - refcount_read(&s->s_ref), refcount_read(&s->s_ref)-1); + if (IS_ERR_OR_NULL(s)) + return; + if (refcount_dec_and_test(&s->s_ref)) { if (s->s_auth.authorizer) ceph_auth_destroy_authorizer(s->s_auth.authorizer); + WARN_ON(mutex_is_locked(&s->s_mutex)); + xa_destroy(&s->s_delegated_inos); kfree(s); } } @@ -627,30 +834,23 @@ static struct ceph_mds_session *register_session(struct ceph_mds_client *mdsc, s->s_mdsc = mdsc; s->s_mds = mds; s->s_state = CEPH_MDS_SESSION_NEW; - s->s_ttl = 0; - s->s_seq = 0; mutex_init(&s->s_mutex); ceph_con_init(&s->s_con, s, &mds_con_ops, &mdsc->fsc->client->msgr); - spin_lock_init(&s->s_gen_ttl_lock); - s->s_cap_gen = 1; + atomic_set(&s->s_cap_gen, 1); s->s_cap_ttl = jiffies - 1; spin_lock_init(&s->s_cap_lock); - s->s_renew_requested = 0; - s->s_renew_seq = 0; INIT_LIST_HEAD(&s->s_caps); - s->s_nr_caps = 0; refcount_set(&s->s_ref, 1); INIT_LIST_HEAD(&s->s_waiting); INIT_LIST_HEAD(&s->s_unsafe); - s->s_num_cap_releases = 0; - s->s_cap_reconnect = 0; - s->s_cap_iterator = NULL; + xa_init(&s->s_delegated_inos); INIT_LIST_HEAD(&s->s_cap_releases); INIT_WORK(&s->s_cap_release_work, ceph_cap_release_work); + INIT_LIST_HEAD(&s->s_cap_dirty); INIT_LIST_HEAD(&s->s_cap_flushing); mdsc->sessions[mds] = s; @@ -694,11 +894,39 @@ static void put_request_session(struct ceph_mds_request *req) } } +void ceph_mdsc_iterate_sessions(struct ceph_mds_client *mdsc, + void (*cb)(struct ceph_mds_session *), + bool check_state) +{ + int mds; + + mutex_lock(&mdsc->mutex); + for (mds = 0; mds < mdsc->max_sessions; ++mds) { + struct ceph_mds_session *s; + + s = __ceph_lookup_mds_session(mdsc, mds); + if (!s) + continue; + + if (check_state && !check_session_state(s)) { + ceph_put_mds_session(s); + continue; + } + + mutex_unlock(&mdsc->mutex); + cb(s); + ceph_put_mds_session(s); + mutex_lock(&mdsc->mutex); + } + mutex_unlock(&mdsc->mutex); +} + void ceph_mdsc_release_request(struct kref *kref) { struct ceph_mds_request *req = container_of(kref, struct ceph_mds_request, r_kref); + ceph_mdsc_release_dir_caps_no_check(req); destroy_reply_info(&req->r_reply_info); if (req->r_request) ceph_msg_put(req->r_request); @@ -706,14 +934,13 @@ void ceph_mdsc_release_request(struct kref *kref) ceph_msg_put(req->r_reply); if (req->r_inode) { ceph_put_cap_refs(ceph_inode(req->r_inode), CEPH_CAP_PIN); - /* avoid calling iput_final() in mds dispatch threads */ - ceph_async_iput(req->r_inode); + iput(req->r_inode); } if (req->r_parent) { ceph_put_cap_refs(ceph_inode(req->r_parent), CEPH_CAP_PIN); - ceph_async_iput(req->r_parent); + iput(req->r_parent); } - ceph_async_iput(req->r_target_inode); + iput(req->r_target_inode); if (req->r_dentry) dput(req->r_dentry); if (req->r_old_dentry) @@ -727,16 +954,17 @@ void ceph_mdsc_release_request(struct kref *kref) */ ceph_put_cap_refs(ceph_inode(req->r_old_dentry_dir), CEPH_CAP_PIN); - ceph_async_iput(req->r_old_dentry_dir); + iput(req->r_old_dentry_dir); } kfree(req->r_path1); kfree(req->r_path2); + put_cred(req->r_cred); if (req->r_pagelist) ceph_pagelist_release(req->r_pagelist); put_request_session(req); ceph_unreserve_caps(req->r_mdsc, &req->r_caps_reservation); WARN_ON_ONCE(!list_empty(&req->r_wait)); - kfree(req); + kmem_cache_free(ceph_mds_request_cachep, req); } DEFINE_RB_FUNCS(request, struct ceph_mds_request, r_tid, r_node) @@ -786,15 +1014,19 @@ static void __register_request(struct ceph_mds_client *mdsc, ceph_mdsc_get_request(req); insert_request(&mdsc->request_tree, req); - req->r_uid = current_fsuid(); - req->r_gid = current_fsgid(); + req->r_cred = get_current_cred(); if (mdsc->oldest_tid == 0 && req->r_op != CEPH_MDS_OP_SETFILELOCK) mdsc->oldest_tid = req->r_tid; if (dir) { + struct ceph_inode_info *ci = ceph_inode(dir); + ihold(dir); req->r_unsafe_dir = dir; + spin_lock(&ci->i_unsafe_lock); + list_add_tail(&req->r_unsafe_dir_item, &ci->i_unsafe_dirops); + spin_unlock(&ci->i_unsafe_lock); } } @@ -822,8 +1054,7 @@ static void __unregister_request(struct ceph_mds_client *mdsc, erase_request(&mdsc->request_tree, req); - if (req->r_unsafe_dir && - test_bit(CEPH_MDS_R_GOT_UNSAFE, &req->r_req_flags)) { + if (req->r_unsafe_dir) { struct ceph_inode_info *ci = ceph_inode(req->r_unsafe_dir); spin_lock(&ci->i_unsafe_lock); list_del_init(&req->r_unsafe_dir_item); @@ -838,8 +1069,7 @@ static void __unregister_request(struct ceph_mds_client *mdsc, } if (req->r_unsafe_dir) { - /* avoid calling iput_final() in mds dispatch threads */ - ceph_async_iput(req->r_unsafe_dir); + iput(req->r_unsafe_dir); req->r_unsafe_dir = NULL; } @@ -993,8 +1223,7 @@ static int __choose_mds(struct ceph_mds_client *mdsc, frag.frag, mds); if (ceph_mdsmap_get_state(mdsc->mdsmap, mds) >= CEPH_MDS_STATE_ACTIVE) { - if (mode == USE_ANY_MDS && - !ceph_mdsmap_is_laggy(mdsc->mdsmap, + if (!ceph_mdsmap_is_laggy(mdsc->mdsmap, mds)) goto out; } @@ -1011,7 +1240,7 @@ static int __choose_mds(struct ceph_mds_client *mdsc, cap = rb_entry(rb_first(&ci->i_caps), struct ceph_cap, ci_node); if (!cap) { spin_unlock(&ci->i_ceph_lock); - ceph_async_iput(inode); + iput(inode); goto random; } mds = cap->session->s_mds; @@ -1020,9 +1249,7 @@ static int __choose_mds(struct ceph_mds_client *mdsc, cap == ci->i_auth_cap ? "auth " : "", cap); spin_unlock(&ci->i_ceph_lock); out: - /* avoid calling iput_final() while holding mdsc->mutex or - * in mds dispatch threads */ - ceph_async_iput(inode); + iput(inode); return mds; random: @@ -1038,7 +1265,7 @@ random: /* * session messages */ -static struct ceph_msg *create_session_msg(u32 op, u64 seq) +struct ceph_msg *ceph_create_session_msg(u32 op, u64 seq) { struct ceph_msg *msg; struct ceph_mds_session_head *h; @@ -1046,7 +1273,8 @@ static struct ceph_msg *create_session_msg(u32 op, u64 seq) msg = ceph_msg_new(CEPH_MSG_CLIENT_SESSION, sizeof(*h), GFP_NOFS, false); if (!msg) { - pr_err("create_session_msg ENOMEM creating msg\n"); + pr_err("ENOMEM creating session %s msg\n", + ceph_session_op_name(op)); return NULL; } h = msg->front.iov_base; @@ -1058,24 +1286,75 @@ static struct ceph_msg *create_session_msg(u32 op, u64 seq) static const unsigned char feature_bits[] = CEPHFS_FEATURES_CLIENT_SUPPORTED; #define FEATURE_BYTES(c) (DIV_ROUND_UP((size_t)feature_bits[c - 1] + 1, 64) * 8) -static void encode_supported_features(void **p, void *end) +static int encode_supported_features(void **p, void *end) { static const size_t count = ARRAY_SIZE(feature_bits); if (count > 0) { size_t i; size_t size = FEATURE_BYTES(count); + unsigned long bit; + + if (WARN_ON_ONCE(*p + 4 + size > end)) + return -ERANGE; - BUG_ON(*p + 4 + size > end); + ceph_encode_32(p, size); + memset(*p, 0, size); + for (i = 0; i < count; i++) { + bit = feature_bits[i]; + ((unsigned char *)(*p))[bit / 8] |= BIT(bit % 8); + } + *p += size; + } else { + if (WARN_ON_ONCE(*p + 4 > end)) + return -ERANGE; + + ceph_encode_32(p, 0); + } + + return 0; +} + +static const unsigned char metric_bits[] = CEPHFS_METRIC_SPEC_CLIENT_SUPPORTED; +#define METRIC_BYTES(cnt) (DIV_ROUND_UP((size_t)metric_bits[cnt - 1] + 1, 64) * 8) +static int encode_metric_spec(void **p, void *end) +{ + static const size_t count = ARRAY_SIZE(metric_bits); + + /* header */ + if (WARN_ON_ONCE(*p + 2 > end)) + return -ERANGE; + + ceph_encode_8(p, 1); /* version */ + ceph_encode_8(p, 1); /* compat */ + + if (count > 0) { + size_t i; + size_t size = METRIC_BYTES(count); + + if (WARN_ON_ONCE(*p + 4 + 4 + size > end)) + return -ERANGE; + + /* metric spec info length */ + ceph_encode_32(p, 4 + size); + + /* metric spec */ ceph_encode_32(p, size); memset(*p, 0, size); for (i = 0; i < count; i++) - ((unsigned char*)(*p))[i / 8] |= BIT(feature_bits[i] % 8); + ((unsigned char *)(*p))[i / 8] |= BIT(metric_bits[i] % 8); *p += size; } else { - BUG_ON(*p + 4 > end); + if (WARN_ON_ONCE(*p + 4 + 4 > end)) + return -ERANGE; + + /* metric spec info length */ + ceph_encode_32(p, 4); + /* metric spec */ ceph_encode_32(p, 0); } + + return 0; } /* @@ -1086,13 +1365,14 @@ static struct ceph_msg *create_session_open_msg(struct ceph_mds_client *mdsc, u6 { struct ceph_msg *msg; struct ceph_mds_session_head *h; - int i = -1; + int i; int extra_bytes = 0; int metadata_key_count = 0; struct ceph_options *opt = mdsc->fsc->client->options; struct ceph_mount_options *fsopt = mdsc->fsc->mount_options; size_t size, count; void *p, *end; + int ret; const char* metadata[][2] = { {"hostname", mdsc->nodename}, @@ -1117,12 +1397,19 @@ static struct ceph_msg *create_session_open_msg(struct ceph_mds_client *mdsc, u6 size = FEATURE_BYTES(count); extra_bytes += 4 + size; + /* metric spec */ + size = 0; + count = ARRAY_SIZE(metric_bits); + if (count > 0) + size = METRIC_BYTES(count); + extra_bytes += 2 + 4 + 4 + size; + /* Allocate the message */ msg = ceph_msg_new(CEPH_MSG_CLIENT_SESSION, sizeof(*h) + extra_bytes, GFP_NOFS, false); if (!msg) { - pr_err("create_session_msg ENOMEM creating msg\n"); - return NULL; + pr_err("ENOMEM creating session open msg\n"); + return ERR_PTR(-ENOMEM); } p = msg->front.iov_base; end = p + msg->front.iov_len; @@ -1135,9 +1422,9 @@ static struct ceph_msg *create_session_open_msg(struct ceph_mds_client *mdsc, u6 * Serialize client metadata into waiting buffer space, using * the format that userspace expects for map<string, string> * - * ClientSession messages with metadata are v3 + * ClientSession messages with metadata are v4 */ - msg->hdr.version = cpu_to_le16(3); + msg->hdr.version = cpu_to_le16(4); msg->hdr.compat_version = cpu_to_le16(1); /* The write pointer, following the session_head structure */ @@ -1159,7 +1446,20 @@ static struct ceph_msg *create_session_open_msg(struct ceph_mds_client *mdsc, u6 p += val_len; } - encode_supported_features(&p, end); + ret = encode_supported_features(&p, end); + if (ret) { + pr_err("encode_supported_features failed!\n"); + ceph_msg_put(msg); + return ERR_PTR(ret); + } + + ret = encode_metric_spec(&p, end); + if (ret) { + pr_err("encode_metric_spec failed!\n"); + ceph_msg_put(msg); + return ERR_PTR(ret); + } + msg->front.iov_len = p - msg->front.iov_base; msg->hdr.front_len = cpu_to_le32(msg->front.iov_len); @@ -1187,8 +1487,8 @@ static int __open_session(struct ceph_mds_client *mdsc, /* send connect message */ msg = create_session_open_msg(mdsc, session->s_seq); - if (!msg) - return -ENOMEM; + if (IS_ERR(msg)) + return PTR_ERR(msg); ceph_con_send(&session->s_con, msg); return 0; } @@ -1202,6 +1502,7 @@ static struct ceph_mds_session * __open_export_target_session(struct ceph_mds_client *mdsc, int target) { struct ceph_mds_session *session; + int ret; session = __ceph_lookup_mds_session(mdsc, target); if (!session) { @@ -1210,8 +1511,11 @@ __open_export_target_session(struct ceph_mds_client *mdsc, int target) return session; } if (session->s_state == CEPH_MDS_SESSION_NEW || - session->s_state == CEPH_MDS_SESSION_CLOSING) - __open_session(mdsc, session); + session->s_state == CEPH_MDS_SESSION_CLOSING) { + ret = __open_session(mdsc, session); + if (ret) + return ERR_PTR(ret); + } return session; } @@ -1246,8 +1550,7 @@ static void __open_export_target_sessions(struct ceph_mds_client *mdsc, for (i = 0; i < mi->num_export_targets; i++) { ts = __open_export_target_session(mdsc, mi->export_targets[i]); - if (!IS_ERR(ts)) - ceph_put_mds_session(ts); + ceph_put_mds_session(ts); } } @@ -1290,7 +1593,6 @@ static void cleanup_session_requests(struct ceph_mds_client *mdsc, { struct ceph_mds_request *req; struct rb_node *p; - struct ceph_inode_info *ci; dout("cleanup_session_requests mds%d\n", session->s_mds); mutex_lock(&mdsc->mutex); @@ -1299,16 +1601,10 @@ static void cleanup_session_requests(struct ceph_mds_client *mdsc, struct ceph_mds_request, r_unsafe_item); pr_warn_ratelimited(" dropping unsafe request %llu\n", req->r_tid); - if (req->r_target_inode) { - /* dropping unsafe change of inode's attributes */ - ci = ceph_inode(req->r_target_inode); - errseq_set(&ci->i_meta_err, -EIO); - } - if (req->r_unsafe_dir) { - /* dropping unsafe directory operation */ - ci = ceph_inode(req->r_unsafe_dir); - errseq_set(&ci->i_meta_err, -EIO); - } + if (req->r_target_inode) + mapping_set_error(req->r_target_inode->i_mapping, -EIO); + if (req->r_unsafe_dir) + mapping_set_error(req->r_unsafe_dir->i_mapping, -EIO); __unregister_request(mdsc, req); } /* zero r_attempts, so kick_requests() will re-send requests */ @@ -1344,7 +1640,7 @@ int ceph_iterate_session_caps(struct ceph_mds_session *session, p = session->s_caps.next; while (p != &session->s_caps) { cap = list_entry(p, struct ceph_cap, session_caps); - inode = igrab(&cap->ci->vfs_inode); + inode = igrab(&cap->ci->netfs.inode); if (!inode) { p = p->next; continue; @@ -1353,9 +1649,7 @@ int ceph_iterate_session_caps(struct ceph_mds_session *session, spin_unlock(&session->s_cap_lock); if (last_inode) { - /* avoid calling iput_final() while holding - * s_mutex or in mds dispatch threads */ - ceph_async_iput(last_inode); + iput(last_inode); last_inode = NULL; } if (old_cap) { @@ -1375,6 +1669,7 @@ int ceph_iterate_session_caps(struct ceph_mds_session *session, cap->session = NULL; list_del_init(&cap->session_caps); session->s_nr_caps--; + atomic64_dec(&session->s_mdsc->metric.total_caps); if (cap->queue_release) __ceph_queue_cap_release(session, cap); else @@ -1388,7 +1683,7 @@ out: session->s_cap_iterator = NULL; spin_unlock(&session->s_cap_lock); - ceph_async_iput(last_inode); + iput(last_inode); if (old_cap) ceph_put_cap(session->s_mdsc, old_cap); @@ -1398,98 +1693,20 @@ 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); - bool dirty_dropped = false; bool invalidate = false; + int iputs; dout("removing cap %p, ci is %p, inode is %p\n", - cap, ci, &ci->vfs_inode); + cap, ci, &ci->netfs.inode); spin_lock(&ci->i_ceph_lock); - if (cap->mds_wanted | cap->issued) - ci->i_ceph_flags |= CEPH_I_CAP_DROPPED; - __ceph_remove_cap(cap, false); - if (!ci->i_auth_cap) { - struct ceph_cap_flush *cf; - struct ceph_mds_client *mdsc = fsc->mdsc; - - if (READ_ONCE(fsc->mount_state) == CEPH_MOUNT_SHUTDOWN) { - if (inode->i_data.nrpages > 0) - invalidate = true; - if (ci->i_wrbuffer_ref > 0) - mapping_set_error(&inode->i_data, -EIO); - } - - while (!list_empty(&ci->i_cap_flush_list)) { - cf = list_first_entry(&ci->i_cap_flush_list, - struct ceph_cap_flush, i_list); - list_move(&cf->i_list, &to_remove); - } - - spin_lock(&mdsc->cap_dirty_lock); - - list_for_each_entry(cf, &to_remove, i_list) - list_del(&cf->g_list); - - if (!list_empty(&ci->i_dirty_item)) { - pr_warn_ratelimited( - " dropping dirty %s state for %p %lld\n", - ceph_cap_string(ci->i_dirty_caps), - inode, ceph_ino(inode)); - ci->i_dirty_caps = 0; - list_del_init(&ci->i_dirty_item); - dirty_dropped = true; - } - if (!list_empty(&ci->i_flushing_item)) { - pr_warn_ratelimited( - " dropping dirty+flushing %s state for %p %lld\n", - ceph_cap_string(ci->i_flushing_caps), - inode, ceph_ino(inode)); - ci->i_flushing_caps = 0; - list_del_init(&ci->i_flushing_item); - mdsc->num_cap_flushing--; - dirty_dropped = true; - } - spin_unlock(&mdsc->cap_dirty_lock); - - if (dirty_dropped) { - errseq_set(&ci->i_meta_err, -EIO); - - if (ci->i_wrbuffer_ref_head == 0 && - ci->i_wr_ref == 0 && - ci->i_dirty_caps == 0 && - ci->i_flushing_caps == 0) { - ceph_put_snap_context(ci->i_head_snapc); - ci->i_head_snapc = NULL; - } - } - - if (atomic_read(&ci->i_filelock_ref) > 0) { - /* make further file lock syscall return -EIO */ - ci->i_ceph_flags |= CEPH_I_ERROR_FILELOCK; - pr_warn_ratelimited(" dropping file locks for %p %lld\n", - inode, ceph_ino(inode)); - } - - if (!ci->i_dirty_caps && ci->i_prealloc_cap_flush) { - list_add(&ci->i_prealloc_cap_flush->i_list, &to_remove); - ci->i_prealloc_cap_flush = NULL; - } - } + iputs = ceph_purge_inode_cap(inode, cap, &invalidate); spin_unlock(&ci->i_ceph_lock); - while (!list_empty(&to_remove)) { - struct ceph_cap_flush *cf; - cf = list_first_entry(&to_remove, - struct ceph_cap_flush, i_list); - list_del(&cf->i_list); - ceph_free_cap_flush(cf); - } wake_up_all(&ci->i_cap_wq); if (invalidate) ceph_queue_invalidate(inode); - if (dirty_dropped) + while (iputs--) iput(inode); return 0; } @@ -1530,8 +1747,7 @@ static void remove_session_caps(struct ceph_mds_session *session) spin_unlock(&session->s_cap_lock); inode = ceph_find_inode(sb, vino); - /* avoid calling iput_final() while holding s_mutex */ - ceph_async_iput(inode); + iput(inode); spin_lock(&session->s_cap_lock); } @@ -1570,13 +1786,10 @@ static int wake_up_session_cb(struct inode *inode, struct ceph_cap *cap, ci->i_requested_max_size = 0; spin_unlock(&ci->i_ceph_lock); } else if (ev == RENEWCAPS) { - if (cap->cap_gen < cap->session->s_cap_gen) { + if (cap->cap_gen < atomic_read(&cap->session->s_cap_gen)) { /* mds did not re-issue stale cap */ spin_lock(&ci->i_ceph_lock); cap->issued = cap->implemented = CEPH_CAP_PIN; - /* make sure mds knows what we want */ - if (__ceph_caps_file_wanted(ci) & ~cap->mds_wanted) - ci->i_ceph_flags |= CEPH_I_CAP_DROPPED; spin_unlock(&ci->i_ceph_lock); } } else if (ev == FORCE_RO) { @@ -1620,8 +1833,8 @@ static int send_renew_caps(struct ceph_mds_client *mdsc, dout("send_renew_caps to mds%d (%s)\n", session->s_mds, ceph_mds_state_name(state)); - msg = create_session_msg(CEPH_SESSION_REQUEST_RENEWCAPS, - ++session->s_renew_seq); + msg = ceph_create_session_msg(CEPH_SESSION_REQUEST_RENEWCAPS, + ++session->s_renew_seq); if (!msg) return -ENOMEM; ceph_con_send(&session->s_con, msg); @@ -1635,7 +1848,7 @@ static int send_flushmsg_ack(struct ceph_mds_client *mdsc, dout("send_flushmsg_ack to mds%d (%s)s seq %lld\n", session->s_mds, ceph_session_state_name(session->s_state), seq); - msg = create_session_msg(CEPH_SESSION_FLUSHMSG_ACK, seq); + msg = ceph_create_session_msg(CEPH_SESSION_FLUSHMSG_ACK, seq); if (!msg) return -ENOMEM; ceph_con_send(&session->s_con, msg); @@ -1680,15 +1893,15 @@ static void renewed_caps(struct ceph_mds_client *mdsc, /* * send a session close request */ -static int request_close_session(struct ceph_mds_client *mdsc, - struct ceph_mds_session *session) +static int request_close_session(struct ceph_mds_session *session) { struct ceph_msg *msg; dout("request_close_session mds%d state %s seq %lld\n", session->s_mds, ceph_session_state_name(session->s_state), session->s_seq); - msg = create_session_msg(CEPH_SESSION_REQUEST_CLOSE, session->s_seq); + msg = ceph_create_session_msg(CEPH_SESSION_REQUEST_CLOSE, + session->s_seq); if (!msg) return -ENOMEM; ceph_con_send(&session->s_con, msg); @@ -1704,7 +1917,7 @@ static int __close_session(struct ceph_mds_client *mdsc, if (session->s_state >= CEPH_MDS_SESSION_CLOSING) return 0; session->s_state = CEPH_MDS_SESSION_CLOSING; - return request_close_session(mdsc, session); + return request_close_session(session); } static bool drop_negative_children(struct dentry *dentry) @@ -1772,7 +1985,8 @@ static int trim_caps_cb(struct inode *inode, struct ceph_cap *cap, void *arg) } /* The inode has cached pages, but it's no longer used. * we can safely drop it */ - if (wanted == 0 && used == CEPH_CAP_FILE_CACHE && + if (S_ISREG(inode->i_mode) && + wanted == 0 && used == CEPH_CAP_FILE_CACHE && !(oissued & CEPH_CAP_FILE_CACHE)) { used = 0; oissued = 0; @@ -1782,7 +1996,7 @@ static int trim_caps_cb(struct inode *inode, struct ceph_cap *cap, void *arg) if (oissued) { /* we aren't the only cap.. just remove us */ - __ceph_remove_cap(cap, true); + ceph_remove_cap(cap, true); (*remaining)--; } else { struct dentry *dentry; @@ -2064,7 +2278,8 @@ int ceph_alloc_readdir_reply_buffer(struct ceph_mds_request *req, order = get_order(size * num_entries); while (order >= 0) { rinfo->dir_entries = (void*)__get_free_pages(GFP_KERNEL | - __GFP_NOWARN, + __GFP_NOWARN | + __GFP_ZERO, order); if (rinfo->dir_entries) break; @@ -2089,18 +2304,21 @@ int ceph_alloc_readdir_reply_buffer(struct ceph_mds_request *req, struct ceph_mds_request * ceph_mdsc_create_request(struct ceph_mds_client *mdsc, int op, int mode) { - struct ceph_mds_request *req = kzalloc(sizeof(*req), GFP_NOFS); + struct ceph_mds_request *req; + req = kmem_cache_zalloc(ceph_mds_request_cachep, GFP_NOFS); if (!req) return ERR_PTR(-ENOMEM); mutex_init(&req->r_fill_mutex); req->r_mdsc = mdsc; req->r_started = jiffies; + req->r_start_latency = ktime_get(); req->r_resend_mds = -1; INIT_LIST_HEAD(&req->r_unsafe_dir_item); INIT_LIST_HEAD(&req->r_unsafe_target_item); req->r_fmode = -1; + req->r_feature_needed = -1; kref_init(&req->r_kref); RB_CLEAR_NODE(&req->r_node); INIT_LIST_HEAD(&req->r_wait); @@ -2299,15 +2517,33 @@ static int set_request_path_attr(struct inode *rinode, struct dentry *rdentry, return r; } +static void encode_timestamp_and_gids(void **p, + const struct ceph_mds_request *req) +{ + struct ceph_timespec ts; + int i; + + ceph_encode_timespec64(&ts, &req->r_stamp); + ceph_encode_copy(p, &ts, sizeof(ts)); + + /* gid_list */ + ceph_encode_32(p, req->r_cred->group_info->ngroups); + for (i = 0; i < req->r_cred->group_info->ngroups; i++) + ceph_encode_64(p, from_kgid(&init_user_ns, + req->r_cred->group_info->gid[i])); +} + /* * called under mdsc->mutex */ -static struct ceph_msg *create_request_message(struct ceph_mds_client *mdsc, +static struct ceph_msg *create_request_message(struct ceph_mds_session *session, struct ceph_mds_request *req, - int mds, bool drop_cap_releases) + bool drop_cap_releases) { + int mds = session->s_mds; + struct ceph_mds_client *mdsc = session->s_mdsc; struct ceph_msg *msg; - struct ceph_mds_request_head *head; + struct ceph_mds_request_head_old *head; const char *path1 = NULL; const char *path2 = NULL; u64 ino1 = 0, ino2 = 0; @@ -2317,6 +2553,7 @@ static struct ceph_msg *create_request_message(struct ceph_mds_client *mdsc, u16 releases; void *p, *end; int ret; + bool legacy = !(session->s_con.peer_features & CEPH_FEATURE_FS_BTIME); ret = set_request_path_attr(req->r_inode, req->r_dentry, req->r_parent, req->r_path1, req->r_ino1.ino, @@ -2338,14 +2575,16 @@ static struct ceph_msg *create_request_message(struct ceph_mds_client *mdsc, goto out_free1; } - len = sizeof(*head) + - pathlen1 + pathlen2 + 2*(1 + sizeof(u32) + sizeof(u64)) + + len = legacy ? sizeof(*head) : sizeof(struct ceph_mds_request_head); + len += pathlen1 + pathlen2 + 2*(1 + sizeof(u32) + sizeof(u64)) + sizeof(struct ceph_timespec); + len += sizeof(u32) + (sizeof(u64) * req->r_cred->group_info->ngroups); /* calculate (max) length for cap releases */ len += sizeof(struct ceph_mds_request_release) * (!!req->r_inode_drop + !!req->r_dentry_drop + !!req->r_old_inode_drop + !!req->r_old_dentry_drop); + if (req->r_dentry_drop) len += pathlen1; if (req->r_old_dentry_drop) @@ -2357,18 +2596,34 @@ static struct ceph_msg *create_request_message(struct ceph_mds_client *mdsc, goto out_free2; } - msg->hdr.version = cpu_to_le16(2); msg->hdr.tid = cpu_to_le64(req->r_tid); - head = msg->front.iov_base; - p = msg->front.iov_base + sizeof(*head); + /* + * The old ceph_mds_request_head didn't contain a version field, and + * one was added when we moved the message version from 3->4. + */ + if (legacy) { + msg->hdr.version = cpu_to_le16(3); + head = msg->front.iov_base; + p = msg->front.iov_base + sizeof(*head); + } else { + struct ceph_mds_request_head *new_head = msg->front.iov_base; + + msg->hdr.version = cpu_to_le16(4); + new_head->version = cpu_to_le16(CEPH_MDS_REQUEST_HEAD_VERSION); + head = (struct ceph_mds_request_head_old *)&new_head->oldest_client_tid; + p = msg->front.iov_base + sizeof(*new_head); + } + end = msg->front.iov_base + msg->front.iov_len; head->mdsmap_epoch = cpu_to_le32(mdsc->mdsmap->m_epoch); head->op = cpu_to_le32(req->r_op); - head->caller_uid = cpu_to_le32(from_kuid(&init_user_ns, req->r_uid)); - head->caller_gid = cpu_to_le32(from_kgid(&init_user_ns, req->r_gid)); - head->ino = 0; + head->caller_uid = cpu_to_le32(from_kuid(&init_user_ns, + req->r_cred->fsuid)); + head->caller_gid = cpu_to_le32(from_kgid(&init_user_ns, + req->r_cred->fsgid)); + head->ino = cpu_to_le64(req->r_deleg_ino); head->args = req->r_args; ceph_encode_filepath(&p, end, ino1, path1); @@ -2382,7 +2637,8 @@ static struct ceph_msg *create_request_message(struct ceph_mds_client *mdsc, if (req->r_inode_drop) releases += ceph_encode_inode_release(&p, req->r_inode ? req->r_inode : d_inode(req->r_dentry), - mds, req->r_inode_drop, req->r_inode_unless, 0); + mds, req->r_inode_drop, req->r_inode_unless, + req->r_op == CEPH_MDS_OP_READDIR); if (req->r_dentry_drop) releases += ceph_encode_dentry_release(&p, req->r_dentry, req->r_parent, mds, req->r_dentry_drop, @@ -2404,14 +2660,14 @@ static struct ceph_msg *create_request_message(struct ceph_mds_client *mdsc, head->num_releases = cpu_to_le16(releases); - /* time stamp */ - { - struct ceph_timespec ts; - ceph_encode_timespec64(&ts, &req->r_stamp); - ceph_encode_copy(&p, &ts, sizeof(ts)); + encode_timestamp_and_gids(&p, req); + + if (WARN_ON_ONCE(p > end)) { + ceph_msg_put(msg); + msg = ERR_PTR(-ERANGE); + goto out_free2; } - BUG_ON(p > end); msg->front.iov_len = p - msg->front.iov_base; msg->hdr.front_len = cpu_to_le32(msg->front.iov_len); @@ -2442,21 +2698,58 @@ out: static void complete_request(struct ceph_mds_client *mdsc, struct ceph_mds_request *req) { + req->r_end_latency = ktime_get(); + if (req->r_callback) req->r_callback(mdsc, req); complete_all(&req->r_completion); } +static struct ceph_mds_request_head_old * +find_old_request_head(void *p, u64 features) +{ + bool legacy = !(features & CEPH_FEATURE_FS_BTIME); + struct ceph_mds_request_head *new_head; + + if (legacy) + return (struct ceph_mds_request_head_old *)p; + new_head = (struct ceph_mds_request_head *)p; + return (struct ceph_mds_request_head_old *)&new_head->oldest_client_tid; +} + /* * called under mdsc->mutex */ -static int __prepare_send_request(struct ceph_mds_client *mdsc, +static int __prepare_send_request(struct ceph_mds_session *session, struct ceph_mds_request *req, - int mds, bool drop_cap_releases) + bool drop_cap_releases) { - struct ceph_mds_request_head *rhead; + int mds = session->s_mds; + struct ceph_mds_client *mdsc = session->s_mdsc; + struct ceph_mds_request_head_old *rhead; struct ceph_msg *msg; - int flags = 0; + int flags = 0, max_retry; + + /* + * The type of 'r_attempts' in kernel 'ceph_mds_request' + * is 'int', while in 'ceph_mds_request_head' the type of + * 'num_retry' is '__u8'. So in case the request retries + * exceeding 256 times, the MDS will receive a incorrect + * retry seq. + * + * In this case it's ususally a bug in MDS and continue + * retrying the request makes no sense. + * + * In future this could be fixed in ceph code, so avoid + * using the hardcode here. + */ + max_retry = sizeof_field(struct ceph_mds_request_head, num_retry); + max_retry = 1 << (max_retry * BITS_PER_BYTE); + if (req->r_attempts >= max_retry) { + pr_warn_ratelimited("%s request tid %llu seq overflow\n", + __func__, req->r_tid); + return -EMULTIHOP; + } req->r_attempts++; if (req->r_inode) { @@ -2468,11 +2761,12 @@ static int __prepare_send_request(struct ceph_mds_client *mdsc, else req->r_sent_on_mseq = -1; } - dout("prepare_send_request %p tid %lld %s (attempt %d)\n", req, + dout("%s %p tid %lld %s (attempt %d)\n", __func__, req, req->r_tid, ceph_mds_op_name(req->r_op), req->r_attempts); if (test_bit(CEPH_MDS_R_GOT_UNSAFE, &req->r_req_flags)) { void *p; + /* * Replay. Do not regenerate message (and rebuild * paths, etc.); just use the original message. @@ -2480,7 +2774,8 @@ static int __prepare_send_request(struct ceph_mds_client *mdsc, * d_move mangles the src name. */ msg = req->r_request; - rhead = msg->front.iov_base; + rhead = find_old_request_head(msg->front.iov_base, + session->s_con.peer_features); flags = le32_to_cpu(rhead->flags); flags |= CEPH_MDS_FLAG_REPLAY; @@ -2494,13 +2789,8 @@ static int __prepare_send_request(struct ceph_mds_client *mdsc, /* remove cap/dentry releases from message */ rhead->num_releases = 0; - /* time stamp */ p = msg->front.iov_base + req->r_request_release_offset; - { - struct ceph_timespec ts; - ceph_encode_timespec64(&ts, &req->r_stamp); - ceph_encode_copy(&p, &ts, sizeof(ts)); - } + encode_timestamp_and_gids(&p, req); msg->front.iov_len = p - msg->front.iov_base; msg->hdr.front_len = cpu_to_le32(msg->front.iov_len); @@ -2511,23 +2801,25 @@ static int __prepare_send_request(struct ceph_mds_client *mdsc, ceph_msg_put(req->r_request); req->r_request = NULL; } - msg = create_request_message(mdsc, req, mds, drop_cap_releases); + msg = create_request_message(session, req, drop_cap_releases); if (IS_ERR(msg)) { req->r_err = PTR_ERR(msg); return PTR_ERR(msg); } req->r_request = msg; - rhead = msg->front.iov_base; + rhead = find_old_request_head(msg->front.iov_base, + session->s_con.peer_features); rhead->oldest_client_tid = cpu_to_le64(__get_oldest_tid(mdsc)); if (test_bit(CEPH_MDS_R_GOT_UNSAFE, &req->r_req_flags)) flags |= CEPH_MDS_FLAG_REPLAY; + if (test_bit(CEPH_MDS_R_ASYNC, &req->r_req_flags)) + flags |= CEPH_MDS_FLAG_ASYNC; if (req->r_parent) flags |= CEPH_MDS_FLAG_WANT_DENTRY; rhead->flags = cpu_to_le32(flags); rhead->num_fwd = req->r_num_fwd; rhead->num_retry = req->r_attempts - 1; - rhead->ino = 0; dout(" r_parent = %p\n", req->r_parent); return 0; @@ -2536,15 +2828,13 @@ static int __prepare_send_request(struct ceph_mds_client *mdsc, /* * called under mdsc->mutex */ -static int __send_request(struct ceph_mds_client *mdsc, - struct ceph_mds_session *session, +static int __send_request(struct ceph_mds_session *session, struct ceph_mds_request *req, bool drop_cap_releases) { int err; - err = __prepare_send_request(mdsc, req, session->s_mds, - drop_cap_releases); + err = __prepare_send_request(session, req, drop_cap_releases); if (!err) { ceph_msg_get(req->r_request); ceph_con_send(&session->s_con, req->r_request); @@ -2573,7 +2863,7 @@ static void __do_request(struct ceph_mds_client *mdsc, if (req->r_timeout && time_after_eq(jiffies, req->r_started + req->r_timeout)) { dout("do_request timed out\n"); - err = -EIO; + err = -ETIMEDOUT; goto finish; } if (READ_ONCE(mdsc->fsc->mount_state) == CEPH_MOUNT_SHUTDOWN) { @@ -2605,6 +2895,10 @@ static void __do_request(struct ceph_mds_client *mdsc, mds = __choose_mds(mdsc, req, &random); if (mds < 0 || ceph_mdsmap_get_state(mdsc->mdsmap, mds) < CEPH_MDS_STATE_ACTIVE) { + if (test_bit(CEPH_MDS_R_ASYNC, &req->r_req_flags)) { + err = -EJUKEBOX; + goto finish; + } dout("do_request no mds or not active, waiting for map\n"); list_add(&req->r_wait, &mdsc->waiting_for_map); return; @@ -2623,15 +2917,46 @@ static void __do_request(struct ceph_mds_client *mdsc, dout("do_request mds%d session %p state %s\n", mds, session, ceph_session_state_name(session->s_state)); + + /* + * The old ceph will crash the MDSs when see unknown OPs + */ + if (req->r_feature_needed > 0 && + !test_bit(req->r_feature_needed, &session->s_features)) { + err = -EOPNOTSUPP; + goto out_session; + } + if (session->s_state != CEPH_MDS_SESSION_OPEN && session->s_state != CEPH_MDS_SESSION_HUNG) { + /* + * We cannot queue async requests since the caps and delegated + * inodes are bound to the session. Just return -EJUKEBOX and + * let the caller retry a sync request in that case. + */ + if (test_bit(CEPH_MDS_R_ASYNC, &req->r_req_flags)) { + err = -EJUKEBOX; + goto out_session; + } + + /* + * If the session has been REJECTED, then return a hard error, + * unless it's a CLEANRECOVER mount, in which case we'll queue + * it to the mdsc queue. + */ if (session->s_state == CEPH_MDS_SESSION_REJECTED) { - err = -EACCES; + if (ceph_test_mount_opt(mdsc->fsc, CLEANRECOVER)) + list_add(&req->r_wait, &mdsc->waiting_for_map); + else + err = -EACCES; goto out_session; } + if (session->s_state == CEPH_MDS_SESSION_NEW || session->s_state == CEPH_MDS_SESSION_CLOSING) { - __open_session(mdsc, session); + err = __open_session(mdsc, session); + if (err) + goto out_session; /* retry the same mds later */ if (random) req->r_resend_mds = mds; @@ -2646,7 +2971,65 @@ static void __do_request(struct ceph_mds_client *mdsc, if (req->r_request_started == 0) /* note request start time */ req->r_request_started = jiffies; - err = __send_request(mdsc, session, req, false); + /* + * For async create we will choose the auth MDS of frag in parent + * directory to send the request and ususally this works fine, but + * if the migrated the dirtory to another MDS before it could handle + * it the request will be forwarded. + * + * And then the auth cap will be changed. + */ + if (test_bit(CEPH_MDS_R_ASYNC, &req->r_req_flags) && req->r_num_fwd) { + struct ceph_dentry_info *di = ceph_dentry(req->r_dentry); + struct ceph_inode_info *ci; + struct ceph_cap *cap; + + /* + * The request maybe handled very fast and the new inode + * hasn't been linked to the dentry yet. We need to wait + * for the ceph_finish_async_create(), which shouldn't be + * stuck too long or fail in thoery, to finish when forwarding + * the request. + */ + if (!d_inode(req->r_dentry)) { + err = wait_on_bit(&di->flags, CEPH_DENTRY_ASYNC_CREATE_BIT, + TASK_KILLABLE); + if (err) { + mutex_lock(&req->r_fill_mutex); + set_bit(CEPH_MDS_R_ABORTED, &req->r_req_flags); + mutex_unlock(&req->r_fill_mutex); + goto out_session; + } + } + + ci = ceph_inode(d_inode(req->r_dentry)); + + spin_lock(&ci->i_ceph_lock); + cap = ci->i_auth_cap; + if (ci->i_ceph_flags & CEPH_I_ASYNC_CREATE && mds != cap->mds) { + dout("do_request session changed for auth cap %d -> %d\n", + cap->session->s_mds, session->s_mds); + + /* Remove the auth cap from old session */ + spin_lock(&cap->session->s_cap_lock); + cap->session->s_nr_caps--; + list_del_init(&cap->session_caps); + spin_unlock(&cap->session->s_cap_lock); + + /* Add the auth cap to the new session */ + cap->mds = mds; + cap->session = session; + spin_lock(&session->s_cap_lock); + session->s_nr_caps++; + list_add_tail(&cap->session_caps, &session->s_caps); + spin_unlock(&session->s_cap_lock); + + change_auth_cap_ses(ci, session); + } + spin_unlock(&ci->i_ceph_lock); + } + + err = __send_request(session, req, false); out_session: ceph_put_mds_session(session); @@ -2709,19 +3092,42 @@ static void kick_requests(struct ceph_mds_client *mdsc, int mds) int ceph_mdsc_submit_request(struct ceph_mds_client *mdsc, struct inode *dir, struct ceph_mds_request *req) { - int err; + int err = 0; /* take CAP_PIN refs for r_inode, r_parent, r_old_dentry */ if (req->r_inode) ceph_get_cap_refs(ceph_inode(req->r_inode), CEPH_CAP_PIN); if (req->r_parent) { - ceph_get_cap_refs(ceph_inode(req->r_parent), CEPH_CAP_PIN); - ihold(req->r_parent); + struct ceph_inode_info *ci = ceph_inode(req->r_parent); + int fmode = (req->r_op & CEPH_MDS_OP_WRITE) ? + CEPH_FILE_MODE_WR : CEPH_FILE_MODE_RD; + spin_lock(&ci->i_ceph_lock); + ceph_take_cap_refs(ci, CEPH_CAP_PIN, false); + __ceph_touch_fmode(ci, mdsc, fmode); + spin_unlock(&ci->i_ceph_lock); } if (req->r_old_dentry_dir) ceph_get_cap_refs(ceph_inode(req->r_old_dentry_dir), CEPH_CAP_PIN); + if (req->r_inode) { + err = ceph_wait_on_async_create(req->r_inode); + if (err) { + dout("%s: wait for async create returned: %d\n", + __func__, err); + return err; + } + } + + if (!err && req->r_old_inode) { + err = ceph_wait_on_async_create(req->r_old_inode); + if (err) { + dout("%s: wait for async create returned: %d\n", + __func__, err); + return err; + } + } + dout("submit_request on %p for inode %p\n", req, dir); mutex_lock(&mdsc->mutex); __register_request(mdsc, req, dir); @@ -2731,15 +3137,16 @@ int ceph_mdsc_submit_request(struct ceph_mds_client *mdsc, struct inode *dir, return err; } -static int ceph_mdsc_wait_request(struct ceph_mds_client *mdsc, - struct ceph_mds_request *req) +int ceph_mdsc_wait_request(struct ceph_mds_client *mdsc, + struct ceph_mds_request *req, + ceph_mds_request_wait_callback_t wait_func) { int err; /* wait */ dout("do_request waiting\n"); - if (!req->r_timeout && req->r_wait_for_completion) { - err = req->r_wait_for_completion(mdsc, req); + if (wait_func) { + err = wait_func(mdsc, req); } else { long timeleft = wait_for_completion_killable_timeout( &req->r_completion, @@ -2747,7 +3154,7 @@ static int ceph_mdsc_wait_request(struct ceph_mds_client *mdsc, if (timeleft > 0) err = 0; else if (!timeleft) - err = -EIO; /* timed out */ + err = -ETIMEDOUT; /* timed out */ else err = timeleft; /* killed */ } @@ -2796,7 +3203,7 @@ int ceph_mdsc_do_request(struct ceph_mds_client *mdsc, /* issue */ err = ceph_mdsc_submit_request(mdsc, dir, req); if (!err) - err = ceph_mdsc_wait_request(mdsc, req); + err = ceph_mdsc_wait_request(mdsc, req, NULL); dout("do_request %p done, result %d\n", req, err); return err; } @@ -2882,35 +3289,6 @@ static void handle_reply(struct ceph_mds_session *session, struct ceph_msg *msg) result = le32_to_cpu(head->result); - /* - * Handle an ESTALE - * if we're not talking to the authority, send to them - * if the authority has changed while we weren't looking, - * send to new authority - * Otherwise we just have to return an ESTALE - */ - if (result == -ESTALE) { - dout("got ESTALE on request %llu\n", req->r_tid); - req->r_resend_mds = -1; - if (req->r_direct_mode != USE_AUTH_MDS) { - dout("not using auth, setting for that now\n"); - req->r_direct_mode = USE_AUTH_MDS; - __do_request(mdsc, req); - mutex_unlock(&mdsc->mutex); - goto out; - } else { - int mds = __choose_mds(mdsc, req, NULL); - if (mds >= 0 && mds != req->r_session->s_mds) { - dout("but auth changed, so resending\n"); - __do_request(mdsc, req); - mutex_unlock(&mdsc->mutex); - goto out; - } - } - dout("have to return ESTALE on request %llu\n", req->r_tid); - } - - if (head->safe) { set_bit(CEPH_MDS_R_GOT_SAFE, &req->r_req_flags); __unregister_request(mdsc, req); @@ -2935,24 +3313,33 @@ static void handle_reply(struct ceph_mds_session *session, struct ceph_msg *msg) } else { set_bit(CEPH_MDS_R_GOT_UNSAFE, &req->r_req_flags); list_add_tail(&req->r_unsafe_item, &req->r_session->s_unsafe); - if (req->r_unsafe_dir) { - struct ceph_inode_info *ci = - ceph_inode(req->r_unsafe_dir); - spin_lock(&ci->i_unsafe_lock); - list_add_tail(&req->r_unsafe_dir_item, - &ci->i_unsafe_dirops); - spin_unlock(&ci->i_unsafe_lock); - } } dout("handle_reply tid %lld result %d\n", tid, result); rinfo = &req->r_reply_info; if (test_bit(CEPHFS_FEATURE_REPLY_ENCODING, &session->s_features)) - err = parse_reply_info(msg, rinfo, (u64)-1); + err = parse_reply_info(session, msg, rinfo, (u64)-1); else - err = parse_reply_info(msg, rinfo, session->s_con.peer_features); + err = parse_reply_info(session, msg, rinfo, session->s_con.peer_features); mutex_unlock(&mdsc->mutex); + /* Must find target inode outside of mutexes to avoid deadlocks */ + if ((err >= 0) && rinfo->head->is_target) { + struct inode *in; + struct ceph_vino tvino = { + .ino = le64_to_cpu(rinfo->targeti.in->ino), + .snap = le64_to_cpu(rinfo->targeti.in->snapid) + }; + + in = ceph_get_inode(mdsc->fsc->sb, tvino); + if (IS_ERR(in)) { + err = PTR_ERR(in); + mutex_lock(&session->s_mutex); + goto out_err; + } + req->r_target_inode = in; + } + mutex_lock(&session->s_mutex); if (err < 0) { pr_err("mdsc_handle_reply got corrupt reply mds%d(tid:%lld)\n", mds, tid); @@ -3020,6 +3407,9 @@ out_err: /* kick calling process */ complete_request(mdsc, req); + + ceph_update_metadata_metrics(&mdsc->metric, req->r_start_latency, + req->r_end_latency, err); out: ceph_mdsc_put_request(req); return; @@ -3041,6 +3431,7 @@ static void handle_forward(struct ceph_mds_client *mdsc, int err = -EINVAL; void *p = msg->front.iov_base; void *end = p + msg->front.iov_len; + bool aborted = false; ceph_decode_need(&p, end, 2*sizeof(u32), bad); next_mds = ceph_decode_32(&p); @@ -3049,16 +3440,41 @@ static void handle_forward(struct ceph_mds_client *mdsc, mutex_lock(&mdsc->mutex); req = lookup_get_request(mdsc, tid); if (!req) { + mutex_unlock(&mdsc->mutex); dout("forward tid %llu to mds%d - req dne\n", tid, next_mds); - goto out; /* dup reply? */ + return; /* dup reply? */ } if (test_bit(CEPH_MDS_R_ABORTED, &req->r_req_flags)) { dout("forward tid %llu aborted, unregistering\n", tid); __unregister_request(mdsc, req); } else if (fwd_seq <= req->r_num_fwd) { - dout("forward tid %llu to mds%d - old seq %d <= %d\n", - tid, next_mds, req->r_num_fwd, fwd_seq); + /* + * The type of 'num_fwd' in ceph 'MClientRequestForward' + * is 'int32_t', while in 'ceph_mds_request_head' the + * type is '__u8'. So in case the request bounces between + * MDSes exceeding 256 times, the client will get stuck. + * + * In this case it's ususally a bug in MDS and continue + * bouncing the request makes no sense. + * + * In future this could be fixed in ceph code, so avoid + * using the hardcode here. + */ + int max = sizeof_field(struct ceph_mds_request_head, num_fwd); + max = 1 << (max * BITS_PER_BYTE); + if (req->r_num_fwd >= max) { + mutex_lock(&req->r_fill_mutex); + req->r_err = -EMULTIHOP; + set_bit(CEPH_MDS_R_ABORTED, &req->r_req_flags); + mutex_unlock(&req->r_fill_mutex); + aborted = true; + pr_warn_ratelimited("forward tid %llu seq overflow\n", + tid); + } else { + dout("forward tid %llu to mds%d - old seq %d <= %d\n", + tid, next_mds, req->r_num_fwd, fwd_seq); + } } else { /* resend. forward race not possible; mds would drop */ dout("forward tid %llu to mds%d (we resend)\n", tid, next_mds); @@ -3070,9 +3486,12 @@ static void handle_forward(struct ceph_mds_client *mdsc, put_request_session(req); __do_request(mdsc, req); } - ceph_mdsc_put_request(req); -out: mutex_unlock(&mdsc->mutex); + + /* kick calling process */ + if (aborted) + complete_request(mdsc, req); + ceph_mdsc_put_request(req); return; bad: @@ -3080,7 +3499,7 @@ bad: } static int __decode_session_metadata(void **p, void *end, - bool *blacklisted) + bool *blocklisted) { /* map<string,string> */ u32 n; @@ -3094,8 +3513,12 @@ static int __decode_session_metadata(void **p, void *end, *p += len; ceph_decode_32_safe(p, end, len, bad); ceph_decode_need(p, end, len, bad); + /* + * Match "blocklisted (blacklisted)" from newer MDSes, + * or "blacklisted" from older MDSes. + */ if (err_str && strnstr(*p, "blacklisted", len)) - *blacklisted = true; + *blocklisted = true; *p += len; } return 0; @@ -3116,10 +3539,9 @@ static void handle_session(struct ceph_mds_session *session, void *end = p + msg->front.iov_len; struct ceph_mds_session_head *h; u32 op; - u64 seq; - unsigned long features = 0; + u64 seq, features = 0; int wake = 0; - bool blacklisted = false; + bool blocklisted = false; /* decode */ ceph_decode_need(&p, end, sizeof(*h), bad); @@ -3131,14 +3553,36 @@ static void handle_session(struct ceph_mds_session *session, if (msg_version >= 3) { u32 len; - /* version >= 2, metadata */ - if (__decode_session_metadata(&p, end, &blacklisted) < 0) + /* version >= 2 and < 5, decode metadata, skip otherwise + * as it's handled via flags. + */ + if (msg_version >= 5) + ceph_decode_skip_map(&p, end, string, string, bad); + else if (__decode_session_metadata(&p, end, &blocklisted) < 0) goto bad; + /* version >= 3, feature bits */ ceph_decode_32_safe(&p, end, len, bad); - ceph_decode_need(&p, end, len, bad); - memcpy(&features, p, min_t(size_t, len, sizeof(features))); - p += len; + if (len) { + ceph_decode_64_safe(&p, end, features, bad); + p += len - sizeof(features); + } + } + + if (msg_version >= 5) { + u32 flags, len; + + /* version >= 4 */ + ceph_decode_skip_16(&p, end, bad); /* struct_v, struct_cv */ + ceph_decode_32_safe(&p, end, len, bad); /* len */ + ceph_decode_skip_n(&p, end, len, bad); /* metric_spec */ + + /* version >= 5, flags */ + ceph_decode_32_safe(&p, end, flags, bad); + if (flags & CEPH_SESSION_BLOCKLISTED) { + pr_warn("mds%d session blocklisted\n", session->s_mds); + blocklisted = true; + } } mutex_lock(&mdsc->mutex); @@ -3165,9 +3609,26 @@ static void handle_session(struct ceph_mds_session *session, case CEPH_SESSION_OPEN: if (session->s_state == CEPH_MDS_SESSION_RECONNECTING) pr_info("mds%d reconnect success\n", session->s_mds); - session->s_state = CEPH_MDS_SESSION_OPEN; - session->s_features = features; - renewed_caps(mdsc, session, 0); + + if (session->s_state == CEPH_MDS_SESSION_OPEN) { + pr_notice("mds%d is already opened\n", session->s_mds); + } else { + session->s_state = CEPH_MDS_SESSION_OPEN; + session->s_features = features; + renewed_caps(mdsc, session, 0); + if (test_bit(CEPHFS_FEATURE_METRIC_COLLECT, + &session->s_features)) + metric_schedule_delayed(&mdsc->metric); + } + + /* + * The connection maybe broken and the session in client + * side has been reinitialized, need to update the seq + * anyway. + */ + if (!session->s_seq && seq) + session->s_seq = seq; + wake = 1; if (mdsc->stopping) __close_session(mdsc, session); @@ -3191,10 +3652,8 @@ static void handle_session(struct ceph_mds_session *session, case CEPH_SESSION_STALE: pr_info("mds%d caps went stale, renewing\n", session->s_mds); - spin_lock(&session->s_gen_ttl_lock); - session->s_cap_gen++; + atomic_inc(&session->s_cap_gen); session->s_cap_ttl = jiffies - 1; - spin_unlock(&session->s_gen_ttl_lock); send_renew_caps(mdsc, session); break; @@ -3220,8 +3679,8 @@ static void handle_session(struct ceph_mds_session *session, session->s_state = CEPH_MDS_SESSION_REJECTED; cleanup_session_requests(mdsc, session); remove_session_caps(session); - if (blacklisted) - mdsc->fsc->blacklisted = true; + if (blocklisted) + mdsc->fsc->blocklisted = true; wake = 2; /* for good measure */ break; @@ -3249,6 +3708,29 @@ bad: return; } +void ceph_mdsc_release_dir_caps(struct ceph_mds_request *req) +{ + int dcaps; + + dcaps = xchg(&req->r_dir_caps, 0); + if (dcaps) { + dout("releasing r_dir_caps=%s\n", ceph_cap_string(dcaps)); + ceph_put_cap_refs(ceph_inode(req->r_parent), dcaps); + } +} + +void ceph_mdsc_release_dir_caps_no_check(struct ceph_mds_request *req) +{ + int dcaps; + + dcaps = xchg(&req->r_dir_caps, 0); + if (dcaps) { + dout("releasing r_dir_caps=%s\n", ceph_cap_string(dcaps)); + ceph_put_cap_refs_no_check_caps(ceph_inode(req->r_parent), + dcaps); + } +} + /* * called under session->mutex. */ @@ -3262,7 +3744,7 @@ static void replay_unsafe_requests(struct ceph_mds_client *mdsc, mutex_lock(&mdsc->mutex); list_for_each_entry_safe(req, nreq, &session->s_unsafe, r_unsafe_item) - __send_request(mdsc, session, req, true); + __send_request(session, req, true); /* * also re-send old requests when MDS enters reconnect stage. So that MDS @@ -3276,9 +3758,14 @@ static void replay_unsafe_requests(struct ceph_mds_client *mdsc, continue; if (req->r_attempts == 0) continue; /* only old requests */ - if (req->r_session && - req->r_session->s_mds == session->s_mds) - __send_request(mdsc, session, req, true); + if (!req->r_session) + continue; + if (req->r_session->s_mds != session->s_mds) + continue; + + ceph_mdsc_release_dir_caps_no_check(req); + + __send_request(session, req, true); } mutex_unlock(&mdsc->mutex); } @@ -3359,10 +3846,43 @@ fail_msg: return err; } +static struct dentry* d_find_primary(struct inode *inode) +{ + struct dentry *alias, *dn = NULL; + + if (hlist_empty(&inode->i_dentry)) + return NULL; + + spin_lock(&inode->i_lock); + if (hlist_empty(&inode->i_dentry)) + goto out_unlock; + + if (S_ISDIR(inode->i_mode)) { + alias = hlist_entry(inode->i_dentry.first, struct dentry, d_u.d_alias); + if (!IS_ROOT(alias)) + dn = dget(alias); + goto out_unlock; + } + + hlist_for_each_entry(alias, &inode->i_dentry, d_u.d_alias) { + spin_lock(&alias->d_lock); + if (!d_unhashed(alias) && + (ceph_dentry(alias)->flags & CEPH_DENTRY_PRIMARY_LINK)) { + dn = dget_dlock(alias); + } + spin_unlock(&alias->d_lock); + if (dn) + break; + } +out_unlock: + spin_unlock(&inode->i_lock); + return dn; +} + /* * Encode information about a cap for a reconnect with the MDS. */ -static int encode_caps_cb(struct inode *inode, struct ceph_cap *cap, +static int reconnect_caps_cb(struct inode *inode, struct ceph_cap *cap, void *arg) { union { @@ -3372,36 +3892,63 @@ static int encode_caps_cb(struct inode *inode, struct ceph_cap *cap, struct ceph_inode_info *ci = cap->ci; struct ceph_reconnect_state *recon_state = arg; struct ceph_pagelist *pagelist = recon_state->pagelist; - int err; + struct dentry *dentry; + char *path; + int pathlen = 0, err; + u64 pathbase; u64 snap_follows; dout(" adding %p ino %llx.%llx cap %p %lld %s\n", inode, ceph_vinop(inode), cap, cap->cap_id, ceph_cap_string(cap->issued)); + dentry = d_find_primary(inode); + if (dentry) { + /* set pathbase to parent dir when msg_version >= 2 */ + path = ceph_mdsc_build_path(dentry, &pathlen, &pathbase, + recon_state->msg_version >= 2); + dput(dentry); + if (IS_ERR(path)) { + err = PTR_ERR(path); + goto out_err; + } + } else { + path = NULL; + pathbase = 0; + } + spin_lock(&ci->i_ceph_lock); cap->seq = 0; /* reset cap seq */ cap->issue_seq = 0; /* and issue_seq */ cap->mseq = 0; /* and migrate_seq */ - cap->cap_gen = cap->session->s_cap_gen; + cap->cap_gen = atomic_read(&cap->session->s_cap_gen); + + /* These are lost when the session goes away */ + if (S_ISDIR(inode->i_mode)) { + if (cap->issued & CEPH_CAP_DIR_CREATE) { + ceph_put_string(rcu_dereference_raw(ci->i_cached_layout.pool_ns)); + memset(&ci->i_cached_layout, 0, sizeof(ci->i_cached_layout)); + } + cap->issued &= ~CEPH_CAP_ANY_DIR_OPS; + } if (recon_state->msg_version >= 2) { rec.v2.cap_id = cpu_to_le64(cap->cap_id); rec.v2.wanted = cpu_to_le32(__ceph_caps_wanted(ci)); rec.v2.issued = cpu_to_le32(cap->issued); rec.v2.snaprealm = cpu_to_le64(ci->i_snap_realm->ino); - rec.v2.pathbase = 0; + rec.v2.pathbase = cpu_to_le64(pathbase); rec.v2.flock_len = (__force __le32) ((ci->i_ceph_flags & CEPH_I_ERROR_FILELOCK) ? 0 : 1); } else { rec.v1.cap_id = cpu_to_le64(cap->cap_id); rec.v1.wanted = cpu_to_le32(__ceph_caps_wanted(ci)); rec.v1.issued = cpu_to_le32(cap->issued); - rec.v1.size = cpu_to_le64(inode->i_size); + rec.v1.size = cpu_to_le64(i_size_read(inode)); ceph_encode_timespec64(&rec.v1.mtime, &inode->i_mtime); ceph_encode_timespec64(&rec.v1.atime, &inode->i_atime); rec.v1.snaprealm = cpu_to_le64(ci->i_snap_realm->ino); - rec.v1.pathbase = 0; + rec.v1.pathbase = cpu_to_le64(pathbase); } if (list_empty(&ci->i_cap_snaps)) { @@ -3463,7 +4010,7 @@ encode_again: sizeof(struct ceph_filelock); rec.v2.flock_len = cpu_to_le32(struct_len); - struct_len += sizeof(u32) + sizeof(rec.v2); + struct_len += sizeof(u32) + pathlen + sizeof(rec.v2); if (struct_v >= 2) struct_len += sizeof(u64); /* snap_follows */ @@ -3487,7 +4034,7 @@ encode_again: ceph_pagelist_encode_8(pagelist, 1); ceph_pagelist_encode_32(pagelist, struct_len); } - ceph_pagelist_encode_string(pagelist, NULL, 0); + ceph_pagelist_encode_string(pagelist, path, pathlen); ceph_pagelist_append(pagelist, &rec, sizeof(rec.v2)); ceph_locks_to_pagelist(flocks, pagelist, num_fcntl_locks, num_flock_locks); @@ -3496,39 +4043,20 @@ encode_again: out_freeflocks: kfree(flocks); } else { - u64 pathbase = 0; - int pathlen = 0; - char *path = NULL; - struct dentry *dentry; - - dentry = d_find_alias(inode); - if (dentry) { - path = ceph_mdsc_build_path(dentry, - &pathlen, &pathbase, 0); - dput(dentry); - if (IS_ERR(path)) { - err = PTR_ERR(path); - goto out_err; - } - rec.v1.pathbase = cpu_to_le64(pathbase); - } - err = ceph_pagelist_reserve(pagelist, sizeof(u64) + sizeof(u32) + pathlen + sizeof(rec.v1)); - if (err) { - goto out_freepath; - } + if (err) + goto out_err; ceph_pagelist_encode_64(pagelist, ceph_ino(inode)); ceph_pagelist_encode_string(pagelist, path, pathlen); ceph_pagelist_append(pagelist, &rec, sizeof(rec.v1)); -out_freepath: - ceph_mdsc_free_path(path, pathlen); } out_err: - if (err >= 0) + ceph_mdsc_free_path(path, pathlen); + if (!err) recon_state->nr_caps++; return err; } @@ -3602,8 +4130,6 @@ fail: * recovering MDS might have. * * This is a relatively heavyweight operation, but it's rare. - * - * called with mdsc->mutex held. */ static void send_mds_reconnect(struct ceph_mds_client *mdsc, struct ceph_mds_session *session) @@ -3626,6 +4152,8 @@ static void send_mds_reconnect(struct ceph_mds_client *mdsc, if (!reply) goto fail_nomsg; + xa_destroy(&session->s_delegated_inos); + mutex_lock(&session->s_mutex); session->s_state = CEPH_MDS_SESSION_RECONNECTING; session->s_seq = 0; @@ -3633,9 +4161,7 @@ static void send_mds_reconnect(struct ceph_mds_client *mdsc, dout("session %p state %s\n", session, ceph_session_state_name(session->s_state)); - spin_lock(&session->s_gen_ttl_lock); - session->s_cap_gen++; - spin_unlock(&session->s_gen_ttl_lock); + atomic_inc(&session->s_cap_gen); spin_lock(&session->s_cap_lock); /* don't know if session is readonly */ @@ -3681,7 +4207,7 @@ static void send_mds_reconnect(struct ceph_mds_client *mdsc, recon_state.msg_version = 2; } /* trsaverse this session's caps */ - err = ceph_iterate_session_caps(session, encode_caps_cb, &recon_state); + err = ceph_iterate_session_caps(session, reconnect_caps_cb, &recon_state); spin_lock(&session->s_cap_lock); session->s_cap_reconnect = 0; @@ -3782,13 +4308,21 @@ static void check_new_map(struct ceph_mds_client *mdsc, struct ceph_mdsmap *newmap, struct ceph_mdsmap *oldmap) { - int i; + int i, j, err; int oldstate, newstate; struct ceph_mds_session *s; + unsigned long targets[DIV_ROUND_UP(CEPH_MAX_MDS, sizeof(unsigned long))] = {0}; dout("check_new_map new %u old %u\n", newmap->m_epoch, oldmap->m_epoch); + if (newmap->m_info) { + for (i = 0; i < newmap->possible_max_rank; i++) { + for (j = 0; j < newmap->m_info[i].num_export_targets; j++) + set_bit(newmap->m_info[i].export_targets[j], targets); + } + } + for (i = 0; i < oldmap->possible_max_rank && i < mdsc->max_sessions; i++) { if (!mdsc->sessions[i]) continue; @@ -3842,6 +4376,7 @@ static void check_new_map(struct ceph_mds_client *mdsc, if (s->s_state == CEPH_MDS_SESSION_RESTARTING && newstate >= CEPH_MDS_STATE_RECONNECT) { mutex_unlock(&mdsc->mutex); + clear_bit(i, targets); send_mds_reconnect(mdsc, s); mutex_lock(&mdsc->mutex); } @@ -3855,11 +4390,60 @@ static void check_new_map(struct ceph_mds_client *mdsc, oldstate != CEPH_MDS_STATE_STARTING) pr_info("mds%d recovery completed\n", s->s_mds); kick_requests(mdsc, i); + mutex_unlock(&mdsc->mutex); + mutex_lock(&s->s_mutex); + mutex_lock(&mdsc->mutex); ceph_kick_flushing_caps(mdsc, s); + mutex_unlock(&s->s_mutex); wake_up_session_caps(s, RECONNECT); } } + /* + * Only open and reconnect sessions that don't exist yet. + */ + for (i = 0; i < newmap->possible_max_rank; i++) { + /* + * In case the import MDS is crashed just after + * the EImportStart journal is flushed, so when + * a standby MDS takes over it and is replaying + * the EImportStart journal the new MDS daemon + * will wait the client to reconnect it, but the + * client may never register/open the session yet. + * + * Will try to reconnect that MDS daemon if the + * rank number is in the export targets array and + * is the up:reconnect state. + */ + newstate = ceph_mdsmap_get_state(newmap, i); + if (!test_bit(i, targets) || newstate != CEPH_MDS_STATE_RECONNECT) + continue; + + /* + * The session maybe registered and opened by some + * requests which were choosing random MDSes during + * the mdsc->mutex's unlock/lock gap below in rare + * case. But the related MDS daemon will just queue + * that requests and be still waiting for the client's + * reconnection request in up:reconnect state. + */ + s = __ceph_lookup_mds_session(mdsc, i); + if (likely(!s)) { + s = __open_export_target_session(mdsc, i); + if (IS_ERR(s)) { + err = PTR_ERR(s); + pr_err("failed to open export target session, err %d\n", + err); + continue; + } + } + dout("send reconnect to export target mds.%d\n", i); + mutex_unlock(&mdsc->mutex); + send_mds_reconnect(mdsc, s); + ceph_put_mds_session(s); + mutex_lock(&mdsc->mutex); + } + for (i = 0; i < newmap->possible_max_rank && i < mdsc->max_sessions; i++) { s = mdsc->sessions[i]; if (!s) @@ -3928,7 +4512,7 @@ static void handle_lease(struct ceph_mds_client *mdsc, dname.len, dname.name); mutex_lock(&session->s_mutex); - session->s_seq++; + inc_session_sequence(session); if (!inode) { dout("handle_lease no inode %llx\n", vino.ino); @@ -3962,7 +4546,7 @@ static void handle_lease(struct ceph_mds_client *mdsc, case CEPH_MDS_LEASE_RENEW: if (di->lease_session == session && - di->lease_gen == session->s_cap_gen && + di->lease_gen == atomic_read(&session->s_cap_gen) && di->lease_renew_from && di->lease_renew_after == 0) { unsigned long duration = @@ -3990,8 +4574,7 @@ release: out: mutex_unlock(&session->s_mutex); - /* avoid calling iput_final() in mds dispatch threads */ - ceph_async_iput(inode); + iput(inode); return; bad: @@ -4027,35 +4610,17 @@ void ceph_mdsc_lease_send_msg(struct ceph_mds_session *session, memcpy((void *)(lease + 1) + 4, dentry->d_name.name, dentry->d_name.len); spin_unlock(&dentry->d_lock); - /* - * if this is a preemptive lease RELEASE, no need to - * flush request stream, since the actual request will - * soon follow. - */ - msg->more_to_follow = (action == CEPH_MDS_LEASE_RELEASE); ceph_con_send(&session->s_con, msg); } /* - * lock unlock sessions, to wait ongoing session activities + * lock unlock the session, to wait ongoing session activities */ -static void lock_unlock_sessions(struct ceph_mds_client *mdsc) +static void lock_unlock_session(struct ceph_mds_session *s) { - int i; - - mutex_lock(&mdsc->mutex); - for (i = 0; i < mdsc->max_sessions; i++) { - struct ceph_mds_session *s = __ceph_lookup_mds_session(mdsc, i); - if (!s) - continue; - mutex_unlock(&mdsc->mutex); - mutex_lock(&s->s_mutex); - mutex_unlock(&s->s_mutex); - ceph_put_mds_session(s); - mutex_lock(&mdsc->mutex); - } - mutex_unlock(&mdsc->mutex); + mutex_lock(&s->s_mutex); + mutex_unlock(&s->s_mutex); } static void maybe_recover_session(struct ceph_mds_client *mdsc) @@ -4068,38 +4633,84 @@ static void maybe_recover_session(struct ceph_mds_client *mdsc) if (READ_ONCE(fsc->mount_state) != CEPH_MOUNT_MOUNTED) return; - if (!READ_ONCE(fsc->blacklisted)) - return; - - if (fsc->last_auto_reconnect && - time_before(jiffies, fsc->last_auto_reconnect + HZ * 60 * 30)) + if (!READ_ONCE(fsc->blocklisted)) return; - pr_info("auto reconnect after blacklisted\n"); - fsc->last_auto_reconnect = jiffies; + pr_info("auto reconnect after blocklisted\n"); ceph_force_reconnect(fsc->sb); } +bool check_session_state(struct ceph_mds_session *s) +{ + switch (s->s_state) { + case CEPH_MDS_SESSION_OPEN: + if (s->s_ttl && time_after(jiffies, s->s_ttl)) { + s->s_state = CEPH_MDS_SESSION_HUNG; + pr_info("mds%d hung\n", s->s_mds); + } + break; + case CEPH_MDS_SESSION_CLOSING: + case CEPH_MDS_SESSION_NEW: + case CEPH_MDS_SESSION_RESTARTING: + case CEPH_MDS_SESSION_CLOSED: + case CEPH_MDS_SESSION_REJECTED: + return false; + } + + return true; +} + +/* + * If the sequence is incremented while we're waiting on a REQUEST_CLOSE reply, + * then we need to retransmit that request. + */ +void inc_session_sequence(struct ceph_mds_session *s) +{ + lockdep_assert_held(&s->s_mutex); + + s->s_seq++; + + if (s->s_state == CEPH_MDS_SESSION_CLOSING) { + int ret; + + dout("resending session close request for mds%d\n", s->s_mds); + ret = request_close_session(s); + if (ret < 0) + pr_err("unable to close session to mds%d: %d\n", + s->s_mds, ret); + } +} + /* - * delayed work -- periodically trim expired leases, renew caps with mds + * delayed work -- periodically trim expired leases, renew caps with mds. If + * the @delay parameter is set to 0 or if it's more than 5 secs, the default + * workqueue delay value of 5 secs will be used. */ -static void schedule_delayed(struct ceph_mds_client *mdsc) +static void schedule_delayed(struct ceph_mds_client *mdsc, unsigned long delay) { - int delay = 5; - unsigned hz = round_jiffies_relative(HZ * delay); - schedule_delayed_work(&mdsc->delayed_work, hz); + unsigned long max_delay = HZ * 5; + + /* 5 secs default delay */ + if (!delay || (delay > max_delay)) + delay = max_delay; + schedule_delayed_work(&mdsc->delayed_work, + round_jiffies_relative(delay)); } static void delayed_work(struct work_struct *work) { - int i; struct ceph_mds_client *mdsc = container_of(work, struct ceph_mds_client, delayed_work.work); + unsigned long delay; int renew_interval; int renew_caps; + int i; dout("mdsc delayed_work\n"); + if (mdsc->stopping) + return; + mutex_lock(&mdsc->mutex); renew_interval = mdsc->mdsmap->m_session_timeout >> 2; renew_caps = time_after_eq(jiffies, HZ*renew_interval + @@ -4111,23 +4722,8 @@ static void delayed_work(struct work_struct *work) struct ceph_mds_session *s = __ceph_lookup_mds_session(mdsc, i); if (!s) continue; - if (s->s_state == CEPH_MDS_SESSION_CLOSING) { - dout("resending session close request for mds%d\n", - s->s_mds); - request_close_session(mdsc, s); - ceph_put_mds_session(s); - continue; - } - if (s->s_ttl && time_after(jiffies, s->s_ttl)) { - if (s->s_state == CEPH_MDS_SESSION_OPEN) { - s->s_state = CEPH_MDS_SESSION_HUNG; - pr_info("mds%d hung\n", s->s_mds); - } - } - if (s->s_state == CEPH_MDS_SESSION_NEW || - s->s_state == CEPH_MDS_SESSION_RESTARTING || - s->s_state == CEPH_MDS_SESSION_REJECTED) { - /* this mds is failed or recovering, just wait */ + + if (!check_session_state(s)) { ceph_put_mds_session(s); continue; } @@ -4148,7 +4744,7 @@ static void delayed_work(struct work_struct *work) } mutex_unlock(&mdsc->mutex); - ceph_check_delayed_caps(mdsc); + delay = ceph_check_delayed_caps(mdsc); ceph_queue_cap_reclaim_work(mdsc); @@ -4156,13 +4752,14 @@ static void delayed_work(struct work_struct *work) maybe_recover_session(mdsc); - schedule_delayed(mdsc); + schedule_delayed(mdsc, delay); } int ceph_mdsc_init(struct ceph_fs_client *fsc) { struct ceph_mds_client *mdsc; + int err; mdsc = kzalloc(sizeof(struct ceph_mds_client), GFP_NOFS); if (!mdsc) @@ -4171,29 +4768,19 @@ int ceph_mdsc_init(struct ceph_fs_client *fsc) mutex_init(&mdsc->mutex); mdsc->mdsmap = kzalloc(sizeof(*mdsc->mdsmap), GFP_NOFS); if (!mdsc->mdsmap) { - kfree(mdsc); - return -ENOMEM; + err = -ENOMEM; + goto err_mdsc; } - fsc->mdsc = mdsc; init_completion(&mdsc->safe_umount_waiters); init_waitqueue_head(&mdsc->session_close_wq); INIT_LIST_HEAD(&mdsc->waiting_for_map); - mdsc->sessions = NULL; - atomic_set(&mdsc->num_sessions, 0); - mdsc->max_sessions = 0; - mdsc->stopping = 0; - atomic64_set(&mdsc->quotarealms_count, 0); mdsc->quotarealms_inodes = RB_ROOT; mutex_init(&mdsc->quotarealms_inodes_mutex); - mdsc->last_snap_seq = 0; init_rwsem(&mdsc->snap_rwsem); mdsc->snap_realms = RB_ROOT; INIT_LIST_HEAD(&mdsc->snap_empty); - mdsc->num_snap_realms = 0; spin_lock_init(&mdsc->snap_empty_lock); - mdsc->last_tid = 0; - mdsc->oldest_tid = 0; mdsc->request_tree = RB_ROOT; INIT_DELAYED_WORK(&mdsc->delayed_work, delayed_work); mdsc->last_renew_caps = jiffies; @@ -4204,13 +4791,13 @@ int ceph_mdsc_init(struct ceph_fs_client *fsc) spin_lock_init(&mdsc->snap_flush_lock); mdsc->last_cap_flush_tid = 1; INIT_LIST_HEAD(&mdsc->cap_flush_list); - INIT_LIST_HEAD(&mdsc->cap_dirty); INIT_LIST_HEAD(&mdsc->cap_dirty_migrating); - mdsc->num_cap_flushing = 0; spin_lock_init(&mdsc->cap_dirty_lock); init_waitqueue_head(&mdsc->cap_flushing_wq); INIT_WORK(&mdsc->cap_reclaim_work, ceph_cap_reclaim_work); - atomic_set(&mdsc->cap_reclaim_pending, 0); + err = ceph_metric_init(&mdsc->metric); + if (err) + goto err_mdsmap; spin_lock_init(&mdsc->dentry_list_lock); INIT_LIST_HEAD(&mdsc->dentry_leases); @@ -4228,7 +4815,15 @@ int ceph_mdsc_init(struct ceph_fs_client *fsc) strscpy(mdsc->nodename, utsname()->nodename, sizeof(mdsc->nodename)); + + fsc->mdsc = mdsc; return 0; + +err_mdsmap: + kfree(mdsc->mdsmap); +err_mdsc: + kfree(mdsc); + return err; } /* @@ -4261,6 +4856,30 @@ static void wait_requests(struct ceph_mds_client *mdsc) dout("wait_requests done\n"); } +void send_flush_mdlog(struct ceph_mds_session *s) +{ + struct ceph_msg *msg; + + /* + * Pre-luminous MDS crashes when it sees an unknown session request + */ + if (!CEPH_HAVE_FEATURE(s->s_con.peer_features, SERVER_LUMINOUS)) + return; + + mutex_lock(&s->s_mutex); + dout("request mdlog flush to mds%d (%s)s seq %lld\n", s->s_mds, + ceph_session_state_name(s->s_state), s->s_seq); + msg = ceph_create_session_msg(CEPH_SESSION_REQUEST_FLUSH_MDLOG, + s->s_seq); + if (!msg) { + pr_err("failed to request mdlog flush to mds%d (%s) seq %lld\n", + s->s_mds, ceph_session_state_name(s->s_state), s->s_seq); + } else { + ceph_con_send(&s->s_con, msg); + } + mutex_unlock(&s->s_mutex); +} + /* * called before mount is ro, and before dentries are torn down. * (hmm, does this still race with new lookups?) @@ -4270,7 +4889,8 @@ void ceph_mdsc_pre_umount(struct ceph_mds_client *mdsc) dout("pre_umount\n"); mdsc->stopping = 1; - lock_unlock_sessions(mdsc); + ceph_mdsc_iterate_sessions(mdsc, send_flush_mdlog, true); + ceph_mdsc_iterate_sessions(mdsc, lock_unlock_session, false); ceph_flush_dirty_caps(mdsc); wait_requests(mdsc); @@ -4284,15 +4904,17 @@ void ceph_mdsc_pre_umount(struct ceph_mds_client *mdsc) } /* - * wait for all write mds requests to flush. + * flush the mdlog and wait for all write mds requests to flush. */ -static void wait_unsafe_requests(struct ceph_mds_client *mdsc, u64 want_tid) +static void flush_mdlog_and_wait_mdsc_unsafe_requests(struct ceph_mds_client *mdsc, + u64 want_tid) { struct ceph_mds_request *req = NULL, *nextreq; + struct ceph_mds_session *last_session = NULL; struct rb_node *n; mutex_lock(&mdsc->mutex); - dout("wait_unsafe_requests want %lld\n", want_tid); + dout("%s want %lld\n", __func__, want_tid); restart: req = __get_oldest_req(mdsc); while (req && req->r_tid <= want_tid) { @@ -4304,14 +4926,32 @@ restart: nextreq = NULL; if (req->r_op != CEPH_MDS_OP_SETFILELOCK && (req->r_op & CEPH_MDS_OP_WRITE)) { + struct ceph_mds_session *s = req->r_session; + + if (!s) { + req = nextreq; + continue; + } + /* write op */ ceph_mdsc_get_request(req); if (nextreq) ceph_mdsc_get_request(nextreq); + s = ceph_get_mds_session(s); mutex_unlock(&mdsc->mutex); - dout("wait_unsafe_requests wait on %llu (want %llu)\n", + + /* send flush mdlog request to MDS */ + if (last_session != s) { + send_flush_mdlog(s); + ceph_put_mds_session(last_session); + last_session = s; + } else { + ceph_put_mds_session(s); + } + dout("%s wait on %llu (want %llu)\n", __func__, req->r_tid, want_tid); wait_for_completion(&req->r_safe_completion); + mutex_lock(&mdsc->mutex); ceph_mdsc_put_request(req); if (!nextreq) @@ -4326,14 +4966,15 @@ restart: req = nextreq; } mutex_unlock(&mdsc->mutex); - dout("wait_unsafe_requests done\n"); + ceph_put_mds_session(last_session); + dout("%s done\n", __func__); } void ceph_mdsc_sync(struct ceph_mds_client *mdsc) { u64 want_tid, want_flush; - if (READ_ONCE(mdsc->fsc->mount_state) == CEPH_MOUNT_SHUTDOWN) + if (READ_ONCE(mdsc->fsc->mount_state) >= CEPH_MOUNT_SHUTDOWN) return; dout("sync\n"); @@ -4355,7 +4996,7 @@ void ceph_mdsc_sync(struct ceph_mds_client *mdsc) dout("sync want tid %lld flush_seq %lld\n", want_tid, want_flush); - wait_unsafe_requests(mdsc, want_tid); + flush_mdlog_and_wait_mdsc_unsafe_requests(mdsc, want_tid); wait_caps_flush(mdsc, want_flush); } @@ -4420,7 +5061,7 @@ void ceph_mdsc_close_sessions(struct ceph_mds_client *mdsc) mutex_unlock(&mdsc->mutex); ceph_cleanup_snapid_map(mdsc); - ceph_cleanup_empty_realms(mdsc); + ceph_cleanup_global_and_empty_realms(mdsc); cancel_work_sync(&mdsc->cap_reclaim_work); cancel_delayed_work_sync(&mdsc->delayed_work); /* cancel timer */ @@ -4465,7 +5106,16 @@ void ceph_mdsc_force_umount(struct ceph_mds_client *mdsc) static void ceph_mdsc_stop(struct ceph_mds_client *mdsc) { dout("stop\n"); - cancel_delayed_work_sync(&mdsc->delayed_work); /* cancel timer */ + /* + * Make sure the delayed work stopped before releasing + * the resources. + * + * Because the cancel_delayed_work_sync() will only + * guarantee that the work finishes executing. But the + * delayed work will re-arm itself again after that. + */ + flush_delayed_work(&mdsc->delayed_work); + if (mdsc->mdsmap) ceph_mdsmap_destroy(mdsc->mdsmap); kfree(mdsc->sessions); @@ -4486,6 +5136,8 @@ void ceph_mdsc_destroy(struct ceph_fs_client *fsc) ceph_mdsc_stop(mdsc); + ceph_metric_destroy(&mdsc->metric); + fsc->mdsc = NULL; kfree(mdsc); dout("mdsc_destroy %p done\n", mdsc); @@ -4498,10 +5150,8 @@ void ceph_mdsc_handle_fsmap(struct ceph_mds_client *mdsc, struct ceph_msg *msg) void *p = msg->front.iov_base; void *end = p + msg->front.iov_len; u32 epoch; - u32 map_len; u32 num_fs; u32 mount_fscid = (u32)-1; - u8 struct_v, struct_cv; int err = -EINVAL; ceph_decode_need(&p, end, sizeof(u32), bad); @@ -4509,24 +5159,17 @@ void ceph_mdsc_handle_fsmap(struct ceph_mds_client *mdsc, struct ceph_msg *msg) dout("handle_fsmap epoch %u\n", epoch); - ceph_decode_need(&p, end, 2 + sizeof(u32), bad); - struct_v = ceph_decode_8(&p); - struct_cv = ceph_decode_8(&p); - map_len = ceph_decode_32(&p); - - ceph_decode_need(&p, end, sizeof(u32) * 3, bad); - p += sizeof(u32) * 2; /* skip epoch and legacy_client_fscid */ + /* struct_v, struct_cv, map_len, epoch, legacy_client_fscid */ + ceph_decode_skip_n(&p, end, 2 + sizeof(u32) * 3, bad); - num_fs = ceph_decode_32(&p); + ceph_decode_32_safe(&p, end, num_fs, bad); while (num_fs-- > 0) { void *info_p, *info_end; u32 info_len; - u8 info_v, info_cv; u32 fscid, namelen; ceph_decode_need(&p, end, 2 + sizeof(u32), bad); - info_v = ceph_decode_8(&p); - info_cv = ceph_decode_8(&p); + p += 2; // info_v, info_cv info_len = ceph_decode_32(&p); ceph_decode_need(&p, end, info_len, bad); info_p = p; @@ -4559,7 +5202,8 @@ void ceph_mdsc_handle_fsmap(struct ceph_mds_client *mdsc, struct ceph_msg *msg) return; bad: - pr_err("error decoding fsmap\n"); + pr_err("error decoding fsmap %d. Shutting down mount.\n", err); + ceph_umount_begin(mdsc->fsc->sb); err_out: mutex_lock(&mdsc->mutex); mdsc->mdsmap_err = err; @@ -4597,7 +5241,7 @@ void ceph_mdsc_handle_mdsmap(struct ceph_mds_client *mdsc, struct ceph_msg *msg) return; } - newmap = ceph_mdsmap_decode(&p, end); + newmap = ceph_mdsmap_decode(&p, end, ceph_msgr2(mdsc->fsc->client)); if (IS_ERR(newmap)) { err = PTR_ERR(newmap); goto bad_unlock; @@ -4620,17 +5264,18 @@ void ceph_mdsc_handle_mdsmap(struct ceph_mds_client *mdsc, struct ceph_msg *msg) mdsc->mdsmap->m_epoch); mutex_unlock(&mdsc->mutex); - schedule_delayed(mdsc); + schedule_delayed(mdsc, 0); return; bad_unlock: mutex_unlock(&mdsc->mutex); bad: - pr_err("error decoding mdsmap %d\n", err); + pr_err("error decoding mdsmap %d. Shutting down mount.\n", err); + ceph_umount_begin(mdsc->fsc->sb); return; } -static struct ceph_connection *con_get(struct ceph_connection *con) +static struct ceph_connection *mds_get_con(struct ceph_connection *con) { struct ceph_mds_session *s = con->private; @@ -4639,7 +5284,7 @@ static struct ceph_connection *con_get(struct ceph_connection *con) return NULL; } -static void con_put(struct ceph_connection *con) +static void mds_put_con(struct ceph_connection *con) { struct ceph_mds_session *s = con->private; @@ -4650,7 +5295,7 @@ static void con_put(struct ceph_connection *con) * if the client is unresponsive for long enough, the mds will kill * the session entirely. */ -static void peer_reset(struct ceph_connection *con) +static void mds_peer_reset(struct ceph_connection *con) { struct ceph_mds_session *s = con->private; struct ceph_mds_client *mdsc = s->s_mdsc; @@ -4659,7 +5304,7 @@ static void peer_reset(struct ceph_connection *con) send_mds_reconnect(mdsc, s); } -static void dispatch(struct ceph_connection *con, struct ceph_msg *msg) +static void mds_dispatch(struct ceph_connection *con, struct ceph_msg *msg) { struct ceph_mds_session *s = con->private; struct ceph_mds_client *mdsc = s->s_mdsc; @@ -4717,35 +5362,24 @@ out: * Note: returned pointer is the address of a structure that's * managed separately. Caller must *not* attempt to free it. */ -static struct ceph_auth_handshake *get_authorizer(struct ceph_connection *con, - int *proto, int force_new) +static struct ceph_auth_handshake * +mds_get_authorizer(struct ceph_connection *con, int *proto, int force_new) { struct ceph_mds_session *s = con->private; struct ceph_mds_client *mdsc = s->s_mdsc; struct ceph_auth_client *ac = mdsc->fsc->client->monc.auth; struct ceph_auth_handshake *auth = &s->s_auth; + int ret; - if (force_new && auth->authorizer) { - ceph_auth_destroy_authorizer(auth->authorizer); - auth->authorizer = NULL; - } - if (!auth->authorizer) { - int ret = ceph_auth_create_authorizer(ac, CEPH_ENTITY_TYPE_MDS, - auth); - if (ret) - return ERR_PTR(ret); - } else { - int ret = ceph_auth_update_authorizer(ac, CEPH_ENTITY_TYPE_MDS, - auth); - if (ret) - return ERR_PTR(ret); - } - *proto = ac->protocol; + ret = __ceph_auth_get_authorizer(ac, auth, CEPH_ENTITY_TYPE_MDS, + force_new, proto, NULL, NULL); + if (ret) + return ERR_PTR(ret); return auth; } -static int add_authorizer_challenge(struct ceph_connection *con, +static int mds_add_authorizer_challenge(struct ceph_connection *con, void *challenge_buf, int challenge_buf_len) { struct ceph_mds_session *s = con->private; @@ -4756,16 +5390,19 @@ static int add_authorizer_challenge(struct ceph_connection *con, challenge_buf, challenge_buf_len); } -static int verify_authorizer_reply(struct ceph_connection *con) +static int mds_verify_authorizer_reply(struct ceph_connection *con) { struct ceph_mds_session *s = con->private; struct ceph_mds_client *mdsc = s->s_mdsc; struct ceph_auth_client *ac = mdsc->fsc->client->monc.auth; + struct ceph_auth_handshake *auth = &s->s_auth; - return ceph_auth_verify_authorizer_reply(ac, s->s_auth.authorizer); + return ceph_auth_verify_authorizer_reply(ac, auth->authorizer, + auth->authorizer_reply_buf, auth->authorizer_reply_buf_len, + NULL, NULL, NULL, NULL); } -static int invalidate_authorizer(struct ceph_connection *con) +static int mds_invalidate_authorizer(struct ceph_connection *con) { struct ceph_mds_session *s = con->private; struct ceph_mds_client *mdsc = s->s_mdsc; @@ -4776,6 +5413,80 @@ static int invalidate_authorizer(struct ceph_connection *con) return ceph_monc_validate_auth(&mdsc->fsc->client->monc); } +static int mds_get_auth_request(struct ceph_connection *con, + void *buf, int *buf_len, + void **authorizer, int *authorizer_len) +{ + struct ceph_mds_session *s = con->private; + struct ceph_auth_client *ac = s->s_mdsc->fsc->client->monc.auth; + struct ceph_auth_handshake *auth = &s->s_auth; + int ret; + + ret = ceph_auth_get_authorizer(ac, auth, CEPH_ENTITY_TYPE_MDS, + buf, buf_len); + if (ret) + return ret; + + *authorizer = auth->authorizer_buf; + *authorizer_len = auth->authorizer_buf_len; + return 0; +} + +static int mds_handle_auth_reply_more(struct ceph_connection *con, + void *reply, int reply_len, + void *buf, int *buf_len, + void **authorizer, int *authorizer_len) +{ + struct ceph_mds_session *s = con->private; + struct ceph_auth_client *ac = s->s_mdsc->fsc->client->monc.auth; + struct ceph_auth_handshake *auth = &s->s_auth; + int ret; + + ret = ceph_auth_handle_svc_reply_more(ac, auth, reply, reply_len, + buf, buf_len); + if (ret) + return ret; + + *authorizer = auth->authorizer_buf; + *authorizer_len = auth->authorizer_buf_len; + return 0; +} + +static int mds_handle_auth_done(struct ceph_connection *con, + u64 global_id, void *reply, int reply_len, + u8 *session_key, int *session_key_len, + u8 *con_secret, int *con_secret_len) +{ + struct ceph_mds_session *s = con->private; + struct ceph_auth_client *ac = s->s_mdsc->fsc->client->monc.auth; + struct ceph_auth_handshake *auth = &s->s_auth; + + return ceph_auth_handle_svc_reply_done(ac, auth, reply, reply_len, + session_key, session_key_len, + con_secret, con_secret_len); +} + +static int mds_handle_auth_bad_method(struct ceph_connection *con, + int used_proto, int result, + const int *allowed_protos, int proto_cnt, + const int *allowed_modes, int mode_cnt) +{ + struct ceph_mds_session *s = con->private; + struct ceph_mon_client *monc = &s->s_mdsc->fsc->client->monc; + int ret; + + if (ceph_auth_handle_bad_authorizer(monc->auth, CEPH_ENTITY_TYPE_MDS, + used_proto, result, + allowed_protos, proto_cnt, + allowed_modes, mode_cnt)) { + ret = ceph_monc_validate_auth(monc); + if (ret) + return ret; + } + + return -EACCES; +} + static struct ceph_msg *mds_alloc_msg(struct ceph_connection *con, struct ceph_msg_header *hdr, int *skip) { @@ -4814,17 +5525,21 @@ static int mds_check_message_signature(struct ceph_msg *msg) } static const struct ceph_connection_operations mds_con_ops = { - .get = con_get, - .put = con_put, - .dispatch = dispatch, - .get_authorizer = get_authorizer, - .add_authorizer_challenge = add_authorizer_challenge, - .verify_authorizer_reply = verify_authorizer_reply, - .invalidate_authorizer = invalidate_authorizer, - .peer_reset = peer_reset, + .get = mds_get_con, + .put = mds_put_con, .alloc_msg = mds_alloc_msg, + .dispatch = mds_dispatch, + .peer_reset = mds_peer_reset, + .get_authorizer = mds_get_authorizer, + .add_authorizer_challenge = mds_add_authorizer_challenge, + .verify_authorizer_reply = mds_verify_authorizer_reply, + .invalidate_authorizer = mds_invalidate_authorizer, .sign_message = mds_sign_message, .check_message_signature = mds_check_message_signature, + .get_auth_request = mds_get_auth_request, + .handle_auth_reply_more = mds_handle_auth_reply_more, + .handle_auth_done = mds_handle_auth_done, + .handle_auth_bad_method = mds_handle_auth_bad_method, }; /* eof */ diff --git a/fs/ceph/mds_client.h b/fs/ceph/mds_client.h index 27a7446e10d3..0598faa50e2e 100644 --- a/fs/ceph/mds_client.h +++ b/fs/ceph/mds_client.h @@ -10,12 +10,16 @@ #include <linux/spinlock.h> #include <linux/refcount.h> #include <linux/utsname.h> +#include <linux/ktime.h> #include <linux/ceph/types.h> #include <linux/ceph/messenger.h> #include <linux/ceph/mdsmap.h> #include <linux/ceph/auth.h> +#include "metric.h" +#include "super.h" + /* The first 8 bits are reserved for old ceph releases */ enum ceph_feature_type { CEPHFS_FEATURE_MIMIC = 8, @@ -23,24 +27,26 @@ enum ceph_feature_type { CEPHFS_FEATURE_RECLAIM_CLIENT, CEPHFS_FEATURE_LAZY_CAP_WANTED, CEPHFS_FEATURE_MULTI_RECONNECT, + CEPHFS_FEATURE_DELEG_INO, + CEPHFS_FEATURE_METRIC_COLLECT, + CEPHFS_FEATURE_ALTERNATE_NAME, + CEPHFS_FEATURE_NOTIFY_SESSION_STATE, + CEPHFS_FEATURE_OP_GETVXATTR, - CEPHFS_FEATURE_MAX = CEPHFS_FEATURE_MULTI_RECONNECT, + CEPHFS_FEATURE_MAX = CEPHFS_FEATURE_OP_GETVXATTR, }; -/* - * This will always have the highest feature bit value - * as the last element of the array. - */ #define CEPHFS_FEATURES_CLIENT_SUPPORTED { \ 0, 1, 2, 3, 4, 5, 6, 7, \ CEPHFS_FEATURE_MIMIC, \ CEPHFS_FEATURE_REPLY_ENCODING, \ CEPHFS_FEATURE_LAZY_CAP_WANTED, \ CEPHFS_FEATURE_MULTI_RECONNECT, \ - \ - CEPHFS_FEATURE_MAX, \ + CEPHFS_FEATURE_DELEG_INO, \ + CEPHFS_FEATURE_METRIC_COLLECT, \ + CEPHFS_FEATURE_NOTIFY_SESSION_STATE, \ + CEPHFS_FEATURE_OP_GETVXATTR, \ } -#define CEPHFS_FEATURES_CLIENT_REQUIRED {} /* * Some lock dependencies: @@ -80,6 +86,7 @@ struct ceph_mds_reply_info_in { s32 dir_pin; struct ceph_timespec btime; struct ceph_timespec snap_btime; + u64 rsnaps; u64 change_attr; }; @@ -91,6 +98,11 @@ struct ceph_mds_reply_dir_entry { loff_t offset; }; +struct ceph_mds_reply_xattr { + char *xattr_value; + size_t xattr_value_len; +}; + /* * parsed info about an mds reply, including information about * either: 1) the target inode and/or its parent directory and dentry, @@ -106,6 +118,7 @@ struct ceph_mds_reply_info_parsed { char *dname; u32 dname_len; struct ceph_mds_reply_lease *dlease; + struct ceph_mds_reply_xattr xattr_info; /* extra */ union { @@ -177,10 +190,8 @@ struct ceph_mds_session { struct ceph_auth_handshake s_auth; - /* protected by s_gen_ttl_lock */ - spinlock_t s_gen_ttl_lock; - u32 s_cap_gen; /* inc each time we get mds stale msg */ - unsigned long s_cap_ttl; /* when session caps expire */ + atomic_t s_cap_gen; /* inc each time we get mds stale msg */ + unsigned long s_cap_ttl; /* when session caps expire. protected by s_mutex */ /* protected by s_cap_lock */ spinlock_t s_cap_lock; @@ -194,13 +205,18 @@ struct ceph_mds_session { struct list_head s_cap_releases; /* waiting cap_release messages */ struct work_struct s_cap_release_work; - /* protected by mutex */ + /* See ceph_inode_info->i_dirty_item. */ + struct list_head s_cap_dirty; /* inodes w/ dirty caps */ + + /* See ceph_inode_info->i_flushing_item. */ struct list_head s_cap_flushing; /* inodes w/ flushing caps */ + unsigned long s_renew_requested; /* last time we sent a renew req */ u64 s_renew_seq; struct list_head s_waiting; /* waiting requests */ struct list_head s_unsafe; /* unsafe requests */ + struct xarray s_delegated_inos; }; /* @@ -255,14 +271,15 @@ struct ceph_mds_request { #define CEPH_MDS_R_GOT_RESULT (5) /* got a result */ #define CEPH_MDS_R_DID_PREPOPULATE (6) /* prepopulated readdir */ #define CEPH_MDS_R_PARENT_LOCKED (7) /* is r_parent->i_rwsem wlocked? */ +#define CEPH_MDS_R_ASYNC (8) /* async request */ unsigned long r_req_flags; struct mutex r_fill_mutex; union ceph_mds_request_args r_args; int r_fmode; /* file mode, if expecting cap */ - kuid_t r_uid; - kgid_t r_gid; + int r_request_release_offset; + const struct cred *r_cred; struct timespec64 r_stamp; /* for choosing which mds to send this request to */ @@ -280,14 +297,19 @@ struct ceph_mds_request { int r_old_inode_drop, r_old_inode_unless; struct ceph_msg *r_request; /* original request */ - int r_request_release_offset; struct ceph_msg *r_reply; struct ceph_mds_reply_info_parsed r_reply_info; - struct page *r_locked_page; int r_err; + u32 r_readdir_offset; + + struct page *r_locked_page; + int r_dir_caps; + int r_num_caps; unsigned long r_timeout; /* optional. jiffies, 0 is "wait forever" */ unsigned long r_started; /* start time to measure timeout against */ + unsigned long r_start_latency; /* start time to measure latency */ + unsigned long r_end_latency; /* finish time to measure latency */ unsigned long r_request_started; /* start time for mds request only, used to measure lease durations */ @@ -304,21 +326,21 @@ struct ceph_mds_request { int r_num_fwd; /* number of forward attempts */ int r_resend_mds; /* mds to resend to next, if any*/ u32 r_sent_on_mseq; /* cap mseq request was sent at*/ + u64 r_deleg_ino; struct list_head r_wait; struct completion r_completion; struct completion r_safe_completion; ceph_mds_request_callback_t r_callback; - ceph_mds_request_wait_callback_t r_wait_for_completion; struct list_head r_unsafe_item; /* per-session unsafe list item */ long long r_dir_release_cnt; long long r_dir_ordered_cnt; int r_readdir_cache_idx; - u32 r_readdir_offset; + + int r_feature_needed; struct ceph_cap_reservation r_caps_reservation; - int r_num_caps; }; struct ceph_pool_perm { @@ -352,7 +374,7 @@ struct ceph_quotarealm_inode { struct cap_wait { struct list_head list; - unsigned long ino; + u64 ino; pid_t tgid; int need; int want; @@ -373,7 +395,7 @@ struct ceph_mds_client { struct ceph_mds_session **sessions; /* NULL for mds if no session */ atomic_t num_sessions; - int max_sessions; /* len of s_mds_sessions */ + int max_sessions; /* len of sessions array */ int stopping; /* true if shutting down */ atomic64_t quotarealms_count; /* # realms with quota */ @@ -411,7 +433,6 @@ struct ceph_mds_client { u64 last_cap_flush_tid; struct list_head cap_flush_list; - struct list_head cap_dirty; /* inodes with dirty caps */ struct list_head cap_dirty_migrating; /* ...that are migration... */ int num_cap_flushing; /* # caps we are flushing */ spinlock_t cap_dirty_lock; /* protects above items */ @@ -446,6 +467,8 @@ struct ceph_mds_client { struct list_head dentry_leases; /* fifo list */ struct list_head dentry_dir_leases; /* lru list */ + struct ceph_client_metric metric; + spinlock_t snapid_map_lock; struct rb_root snapid_map_tree; struct list_head snapid_map_lru; @@ -458,6 +481,9 @@ struct ceph_mds_client { extern const char *ceph_mds_op_name(int op); +extern bool check_session_state(struct ceph_mds_session *s); +void inc_session_sequence(struct ceph_mds_session *s); + extern struct ceph_mds_session * __ceph_lookup_mds_session(struct ceph_mds_client *, int mds); @@ -485,9 +511,14 @@ ceph_mdsc_create_request(struct ceph_mds_client *mdsc, int op, int mode); extern int ceph_mdsc_submit_request(struct ceph_mds_client *mdsc, struct inode *dir, struct ceph_mds_request *req); +int ceph_mdsc_wait_request(struct ceph_mds_client *mdsc, + struct ceph_mds_request *req, + ceph_mds_request_wait_callback_t wait_func); extern int ceph_mdsc_do_request(struct ceph_mds_client *mdsc, struct inode *dir, struct ceph_mds_request *req); +extern void ceph_mdsc_release_dir_caps(struct ceph_mds_request *req); +extern void ceph_mdsc_release_dir_caps_no_check(struct ceph_mds_request *req); static inline void ceph_mdsc_get_request(struct ceph_mds_request *req) { kref_get(&req->r_kref); @@ -498,6 +529,11 @@ static inline void ceph_mdsc_put_request(struct ceph_mds_request *req) kref_put(&req->r_kref, ceph_mdsc_release_request); } +extern void send_flush_mdlog(struct ceph_mds_session *s); +extern void ceph_mdsc_iterate_sessions(struct ceph_mds_client *mdsc, + void (*cb)(struct ceph_mds_session *), + bool check_state); +extern struct ceph_msg *ceph_create_session_msg(u32 op, u64 seq); extern void __ceph_queue_cap_release(struct ceph_mds_session *session, struct ceph_cap *cap); extern void ceph_flush_cap_releases(struct ceph_mds_client *mdsc, @@ -512,7 +548,7 @@ extern void ceph_mdsc_pre_umount(struct ceph_mds_client *mdsc); static inline void ceph_mdsc_free_path(char *path, int len) { - if (path) + if (!IS_ERR_OR_NULL(path)) __putname(path - (PATH_MAX - 1 - len)); } @@ -537,4 +573,16 @@ extern void ceph_mdsc_open_export_target_sessions(struct ceph_mds_client *mdsc, extern int ceph_trim_caps(struct ceph_mds_client *mdsc, struct ceph_mds_session *session, int max_caps); + +static inline int ceph_wait_on_async_create(struct inode *inode) +{ + struct ceph_inode_info *ci = ceph_inode(inode); + + return wait_on_bit(&ci->i_ceph_flags, CEPH_ASYNC_CREATE_BIT, + TASK_KILLABLE); +} + +extern int ceph_wait_on_conflict_unlink(struct dentry *dentry); +extern u64 ceph_get_deleg_ino(struct ceph_mds_session *session); +extern int ceph_restore_deleg_ino(struct ceph_mds_session *session, u64 ino); #endif diff --git a/fs/ceph/mdsmap.c b/fs/ceph/mdsmap.c index 889627817e52..3fbabc98e1f7 100644 --- a/fs/ceph/mdsmap.c +++ b/fs/ceph/mdsmap.c @@ -29,7 +29,7 @@ static int __mdsmap_get_random_mds(struct ceph_mdsmap *m, bool ignore_laggy) return -1; /* pick */ - n = prandom_u32() % n; + n = prandom_u32_max(n); for (j = 0, i = 0; i < m->possible_max_rank; i++) { if (CEPH_MDS_IS_READY(i, ignore_laggy)) j++; @@ -114,14 +114,15 @@ bad: * Ignore any fields we don't care about (there are quite a few of * them). */ -struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end) +struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end, bool msgr2) { struct ceph_mdsmap *m; const void *start = *p; int i, j, n; int err; - u8 mdsmap_v, mdsmap_cv; + u8 mdsmap_v; u16 mdsmap_ev; + u32 target; m = kzalloc(sizeof(*m), GFP_NOFS); if (!m) @@ -129,7 +130,7 @@ struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end) ceph_decode_need(p, end, 1 + 1, bad); mdsmap_v = ceph_decode_8(p); - mdsmap_cv = ceph_decode_8(p); + *p += sizeof(u8); /* mdsmap_cv */ if (mdsmap_v >= 4) { u32 mdsmap_len; ceph_decode_32_safe(p, end, mdsmap_len, bad); @@ -174,7 +175,6 @@ struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end) u64 global_id; u32 namelen; s32 mds, inc, state; - u64 state_seq; u8 info_v; void *info_end = NULL; struct ceph_entity_addr addr; @@ -189,9 +189,8 @@ struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end) 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); + *p += sizeof(u8); /* info_cv */ info_len = ceph_decode_32(p); info_end = *p + info_len; if (info_end > end) @@ -203,18 +202,19 @@ struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end) namelen = ceph_decode_32(p); /* skip mds name */ *p += namelen; - ceph_decode_need(p, end, - 4*sizeof(u32) + sizeof(u64) + - sizeof(addr) + sizeof(struct ceph_timespec), - bad); - mds = ceph_decode_32(p); - inc = ceph_decode_32(p); - state = ceph_decode_32(p); - state_seq = ceph_decode_64(p); - err = ceph_decode_entity_addr(p, end, &addr); + ceph_decode_32_safe(p, end, mds, bad); + ceph_decode_32_safe(p, end, inc, bad); + ceph_decode_32_safe(p, end, state, bad); + *p += sizeof(u64); /* state_seq */ + if (info_v >= 8) + err = ceph_decode_entity_addrvec(p, end, msgr2, &addr); + else + err = ceph_decode_entity_addr(p, end, &addr); if (err) goto corrupt; - ceph_decode_copy(p, &laggy_since, sizeof(laggy_since)); + + ceph_decode_copy_safe(p, end, &laggy_since, sizeof(laggy_since), + bad); laggy = laggy_since.tv_sec != 0 || laggy_since.tv_nsec != 0; *p += sizeof(u32); ceph_decode_32_safe(p, end, namelen, bad); @@ -245,8 +245,8 @@ struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end) } if (state <= 0) { - pr_warn("mdsmap_decode got incorrect state(%s)\n", - ceph_mds_state_name(state)); + dout("mdsmap_decode got incorrect state(%s)\n", + ceph_mds_state_name(state)); continue; } @@ -261,9 +261,10 @@ struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end) sizeof(u32), GFP_NOFS); if (!info->export_targets) goto nomem; - for (j = 0; j < num_export_targets; j++) - info->export_targets[j] = - ceph_decode_32(&pexport_targets); + for (j = 0; j < num_export_targets; j++) { + target = ceph_decode_32(&pexport_targets); + info->export_targets[j] = target; + } } else { info->export_targets = NULL; } @@ -351,12 +352,10 @@ struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end) __decode_and_drop_type(p, end, u8, bad_ext); } if (mdsmap_ev >= 8) { - u32 name_len; /* enabled */ ceph_decode_8_safe(p, end, m->m_enabled, bad_ext); - ceph_decode_32_safe(p, end, name_len, bad_ext); - ceph_decode_need(p, end, name_len, bad_ext); - *p += name_len; + /* fs_name */ + ceph_decode_skip_string(p, end, bad_ext); } /* damaged */ if (mdsmap_ev >= 9) { @@ -369,6 +368,22 @@ struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end) } else { m->m_damaged = false; } + if (mdsmap_ev >= 17) { + /* balancer */ + ceph_decode_skip_string(p, end, bad_ext); + /* standby_count_wanted */ + ceph_decode_skip_32(p, end, bad_ext); + /* old_max_mds */ + ceph_decode_skip_32(p, end, bad_ext); + /* min_compat_client */ + ceph_decode_skip_8(p, end, bad_ext); + /* required_client_features */ + ceph_decode_skip_set(p, end, 64, bad_ext); + ceph_decode_64_safe(p, end, m->m_max_xattr_size, bad_ext); + } else { + /* This forces the usage of the (sync) SETXATTR Op */ + m->m_max_xattr_size = 0; + } bad_ext: dout("mdsmap_decode m_enabled: %d, m_damaged: %d, m_num_laggy: %d\n", !!m->m_enabled, !!m->m_damaged, m->m_num_laggy); @@ -395,9 +410,11 @@ void ceph_mdsmap_destroy(struct ceph_mdsmap *m) { int i; - for (i = 0; i < m->possible_max_rank; i++) - kfree(m->m_info[i].export_targets); - kfree(m->m_info); + if (m->m_info) { + for (i = 0; i < m->possible_max_rank; i++) + kfree(m->m_info[i].export_targets); + kfree(m->m_info); + } kfree(m->m_data_pg_pools); kfree(m); } diff --git a/fs/ceph/metric.c b/fs/ceph/metric.c new file mode 100644 index 000000000000..c47347d2e84e --- /dev/null +++ b/fs/ceph/metric.c @@ -0,0 +1,353 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#include <linux/ceph/ceph_debug.h> + +#include <linux/types.h> +#include <linux/percpu_counter.h> +#include <linux/math64.h> + +#include "metric.h" +#include "mds_client.h" + +static void ktime_to_ceph_timespec(struct ceph_timespec *ts, ktime_t val) +{ + struct timespec64 t = ktime_to_timespec64(val); + ceph_encode_timespec64(ts, &t); +} + +static bool ceph_mdsc_send_metrics(struct ceph_mds_client *mdsc, + struct ceph_mds_session *s) +{ + struct ceph_metric_head *head; + struct ceph_metric_cap *cap; + struct ceph_metric_read_latency *read; + struct ceph_metric_write_latency *write; + struct ceph_metric_metadata_latency *meta; + struct ceph_metric_dlease *dlease; + struct ceph_opened_files *files; + struct ceph_pinned_icaps *icaps; + struct ceph_opened_inodes *inodes; + struct ceph_read_io_size *rsize; + struct ceph_write_io_size *wsize; + struct ceph_client_metric *m = &mdsc->metric; + u64 nr_caps = atomic64_read(&m->total_caps); + u32 header_len = sizeof(struct ceph_metric_header); + struct ceph_msg *msg; + s64 sum; + s32 items = 0; + s32 len; + + len = sizeof(*head) + sizeof(*cap) + sizeof(*read) + sizeof(*write) + + sizeof(*meta) + sizeof(*dlease) + sizeof(*files) + + sizeof(*icaps) + sizeof(*inodes) + sizeof(*rsize) + + sizeof(*wsize); + + msg = ceph_msg_new(CEPH_MSG_CLIENT_METRICS, len, GFP_NOFS, true); + if (!msg) { + pr_err("send metrics to mds%d, failed to allocate message\n", + s->s_mds); + return false; + } + + head = msg->front.iov_base; + + /* encode the cap metric */ + cap = (struct ceph_metric_cap *)(head + 1); + cap->header.type = cpu_to_le32(CLIENT_METRIC_TYPE_CAP_INFO); + cap->header.ver = 1; + cap->header.compat = 1; + cap->header.data_len = cpu_to_le32(sizeof(*cap) - header_len); + cap->hit = cpu_to_le64(percpu_counter_sum(&m->i_caps_hit)); + cap->mis = cpu_to_le64(percpu_counter_sum(&m->i_caps_mis)); + cap->total = cpu_to_le64(nr_caps); + items++; + + /* encode the read latency metric */ + read = (struct ceph_metric_read_latency *)(cap + 1); + read->header.type = cpu_to_le32(CLIENT_METRIC_TYPE_READ_LATENCY); + read->header.ver = 2; + read->header.compat = 1; + read->header.data_len = cpu_to_le32(sizeof(*read) - header_len); + sum = m->metric[METRIC_READ].latency_sum; + ktime_to_ceph_timespec(&read->lat, sum); + ktime_to_ceph_timespec(&read->avg, m->metric[METRIC_READ].latency_avg); + read->sq_sum = cpu_to_le64(m->metric[METRIC_READ].latency_sq_sum); + read->count = cpu_to_le64(m->metric[METRIC_READ].total); + items++; + + /* encode the write latency metric */ + write = (struct ceph_metric_write_latency *)(read + 1); + write->header.type = cpu_to_le32(CLIENT_METRIC_TYPE_WRITE_LATENCY); + write->header.ver = 2; + write->header.compat = 1; + write->header.data_len = cpu_to_le32(sizeof(*write) - header_len); + sum = m->metric[METRIC_WRITE].latency_sum; + ktime_to_ceph_timespec(&write->lat, sum); + ktime_to_ceph_timespec(&write->avg, m->metric[METRIC_WRITE].latency_avg); + write->sq_sum = cpu_to_le64(m->metric[METRIC_WRITE].latency_sq_sum); + write->count = cpu_to_le64(m->metric[METRIC_WRITE].total); + items++; + + /* encode the metadata latency metric */ + meta = (struct ceph_metric_metadata_latency *)(write + 1); + meta->header.type = cpu_to_le32(CLIENT_METRIC_TYPE_METADATA_LATENCY); + meta->header.ver = 2; + meta->header.compat = 1; + meta->header.data_len = cpu_to_le32(sizeof(*meta) - header_len); + sum = m->metric[METRIC_METADATA].latency_sum; + ktime_to_ceph_timespec(&meta->lat, sum); + ktime_to_ceph_timespec(&meta->avg, m->metric[METRIC_METADATA].latency_avg); + meta->sq_sum = cpu_to_le64(m->metric[METRIC_METADATA].latency_sq_sum); + meta->count = cpu_to_le64(m->metric[METRIC_METADATA].total); + items++; + + /* encode the dentry lease metric */ + dlease = (struct ceph_metric_dlease *)(meta + 1); + dlease->header.type = cpu_to_le32(CLIENT_METRIC_TYPE_DENTRY_LEASE); + dlease->header.ver = 1; + dlease->header.compat = 1; + dlease->header.data_len = cpu_to_le32(sizeof(*dlease) - header_len); + dlease->hit = cpu_to_le64(percpu_counter_sum(&m->d_lease_hit)); + dlease->mis = cpu_to_le64(percpu_counter_sum(&m->d_lease_mis)); + dlease->total = cpu_to_le64(atomic64_read(&m->total_dentries)); + items++; + + sum = percpu_counter_sum(&m->total_inodes); + + /* encode the opened files metric */ + files = (struct ceph_opened_files *)(dlease + 1); + files->header.type = cpu_to_le32(CLIENT_METRIC_TYPE_OPENED_FILES); + files->header.ver = 1; + files->header.compat = 1; + files->header.data_len = cpu_to_le32(sizeof(*files) - header_len); + files->opened_files = cpu_to_le64(atomic64_read(&m->opened_files)); + files->total = cpu_to_le64(sum); + items++; + + /* encode the pinned icaps metric */ + icaps = (struct ceph_pinned_icaps *)(files + 1); + icaps->header.type = cpu_to_le32(CLIENT_METRIC_TYPE_PINNED_ICAPS); + icaps->header.ver = 1; + icaps->header.compat = 1; + icaps->header.data_len = cpu_to_le32(sizeof(*icaps) - header_len); + icaps->pinned_icaps = cpu_to_le64(nr_caps); + icaps->total = cpu_to_le64(sum); + items++; + + /* encode the opened inodes metric */ + inodes = (struct ceph_opened_inodes *)(icaps + 1); + inodes->header.type = cpu_to_le32(CLIENT_METRIC_TYPE_OPENED_INODES); + inodes->header.ver = 1; + inodes->header.compat = 1; + inodes->header.data_len = cpu_to_le32(sizeof(*inodes) - header_len); + inodes->opened_inodes = cpu_to_le64(percpu_counter_sum(&m->opened_inodes)); + inodes->total = cpu_to_le64(sum); + items++; + + /* encode the read io size metric */ + rsize = (struct ceph_read_io_size *)(inodes + 1); + rsize->header.type = cpu_to_le32(CLIENT_METRIC_TYPE_READ_IO_SIZES); + rsize->header.ver = 1; + rsize->header.compat = 1; + rsize->header.data_len = cpu_to_le32(sizeof(*rsize) - header_len); + rsize->total_ops = cpu_to_le64(m->metric[METRIC_READ].total); + rsize->total_size = cpu_to_le64(m->metric[METRIC_READ].size_sum); + items++; + + /* encode the write io size metric */ + wsize = (struct ceph_write_io_size *)(rsize + 1); + wsize->header.type = cpu_to_le32(CLIENT_METRIC_TYPE_WRITE_IO_SIZES); + wsize->header.ver = 1; + wsize->header.compat = 1; + wsize->header.data_len = cpu_to_le32(sizeof(*wsize) - header_len); + wsize->total_ops = cpu_to_le64(m->metric[METRIC_WRITE].total); + wsize->total_size = cpu_to_le64(m->metric[METRIC_WRITE].size_sum); + items++; + + put_unaligned_le32(items, &head->num); + msg->front.iov_len = len; + msg->hdr.version = cpu_to_le16(1); + msg->hdr.compat_version = cpu_to_le16(1); + msg->hdr.front_len = cpu_to_le32(msg->front.iov_len); + ceph_con_send(&s->s_con, msg); + + return true; +} + + +static void metric_get_session(struct ceph_mds_client *mdsc) +{ + struct ceph_mds_session *s; + int i; + + mutex_lock(&mdsc->mutex); + for (i = 0; i < mdsc->max_sessions; i++) { + s = __ceph_lookup_mds_session(mdsc, i); + if (!s) + continue; + + /* + * Skip it if MDS doesn't support the metric collection, + * or the MDS will close the session's socket connection + * directly when it get this message. + */ + if (check_session_state(s) && + test_bit(CEPHFS_FEATURE_METRIC_COLLECT, &s->s_features)) { + mdsc->metric.session = s; + break; + } + + ceph_put_mds_session(s); + } + mutex_unlock(&mdsc->mutex); +} + +static void metric_delayed_work(struct work_struct *work) +{ + struct ceph_client_metric *m = + container_of(work, struct ceph_client_metric, delayed_work.work); + struct ceph_mds_client *mdsc = + container_of(m, struct ceph_mds_client, metric); + + if (mdsc->stopping) + return; + + if (!m->session || !check_session_state(m->session)) { + if (m->session) { + ceph_put_mds_session(m->session); + m->session = NULL; + } + metric_get_session(mdsc); + } + if (m->session) { + ceph_mdsc_send_metrics(mdsc, m->session); + metric_schedule_delayed(m); + } +} + +int ceph_metric_init(struct ceph_client_metric *m) +{ + struct ceph_metric *metric; + int ret, i; + + if (!m) + return -EINVAL; + + atomic64_set(&m->total_dentries, 0); + ret = percpu_counter_init(&m->d_lease_hit, 0, GFP_KERNEL); + if (ret) + return ret; + + ret = percpu_counter_init(&m->d_lease_mis, 0, GFP_KERNEL); + if (ret) + goto err_d_lease_mis; + + atomic64_set(&m->total_caps, 0); + ret = percpu_counter_init(&m->i_caps_hit, 0, GFP_KERNEL); + if (ret) + goto err_i_caps_hit; + + ret = percpu_counter_init(&m->i_caps_mis, 0, GFP_KERNEL); + if (ret) + goto err_i_caps_mis; + + for (i = 0; i < METRIC_MAX; i++) { + metric = &m->metric[i]; + spin_lock_init(&metric->lock); + metric->size_sum = 0; + metric->size_min = U64_MAX; + metric->size_max = 0; + metric->total = 0; + metric->latency_sum = 0; + metric->latency_avg = 0; + metric->latency_sq_sum = 0; + metric->latency_min = KTIME_MAX; + metric->latency_max = 0; + } + + atomic64_set(&m->opened_files, 0); + ret = percpu_counter_init(&m->opened_inodes, 0, GFP_KERNEL); + if (ret) + goto err_opened_inodes; + ret = percpu_counter_init(&m->total_inodes, 0, GFP_KERNEL); + if (ret) + goto err_total_inodes; + + m->session = NULL; + INIT_DELAYED_WORK(&m->delayed_work, metric_delayed_work); + + return 0; + +err_total_inodes: + percpu_counter_destroy(&m->opened_inodes); +err_opened_inodes: + percpu_counter_destroy(&m->i_caps_mis); +err_i_caps_mis: + percpu_counter_destroy(&m->i_caps_hit); +err_i_caps_hit: + percpu_counter_destroy(&m->d_lease_mis); +err_d_lease_mis: + percpu_counter_destroy(&m->d_lease_hit); + + return ret; +} + +void ceph_metric_destroy(struct ceph_client_metric *m) +{ + if (!m) + return; + + cancel_delayed_work_sync(&m->delayed_work); + + percpu_counter_destroy(&m->total_inodes); + percpu_counter_destroy(&m->opened_inodes); + percpu_counter_destroy(&m->i_caps_mis); + percpu_counter_destroy(&m->i_caps_hit); + percpu_counter_destroy(&m->d_lease_mis); + percpu_counter_destroy(&m->d_lease_hit); + + ceph_put_mds_session(m->session); +} + +#define METRIC_UPDATE_MIN_MAX(min, max, new) \ +{ \ + if (unlikely(new < min)) \ + min = new; \ + if (unlikely(new > max)) \ + max = new; \ +} + +static inline void __update_mean_and_stdev(ktime_t total, ktime_t *lavg, + ktime_t *sq_sump, ktime_t lat) +{ + ktime_t avg; + + if (unlikely(total == 1)) { + *lavg = lat; + } else { + /* the sq is (lat - old_avg) * (lat - new_avg) */ + avg = *lavg + div64_s64(lat - *lavg, total); + *sq_sump += (lat - *lavg)*(lat - avg); + *lavg = avg; + } +} + +void ceph_update_metrics(struct ceph_metric *m, + ktime_t r_start, ktime_t r_end, + unsigned int size, int rc) +{ + ktime_t lat = ktime_sub(r_end, r_start); + ktime_t total; + + if (unlikely(rc < 0 && rc != -ENOENT && rc != -ETIMEDOUT)) + return; + + spin_lock(&m->lock); + total = ++m->total; + m->size_sum += size; + METRIC_UPDATE_MIN_MAX(m->size_min, m->size_max, size); + m->latency_sum += lat; + METRIC_UPDATE_MIN_MAX(m->latency_min, m->latency_max, lat); + __update_mean_and_stdev(total, &m->latency_avg, &m->latency_sq_sum, + lat); + spin_unlock(&m->lock); +} diff --git a/fs/ceph/metric.h b/fs/ceph/metric.h new file mode 100644 index 000000000000..0d0c44bd3332 --- /dev/null +++ b/fs/ceph/metric.h @@ -0,0 +1,244 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef _FS_CEPH_MDS_METRIC_H +#define _FS_CEPH_MDS_METRIC_H + +#include <linux/ceph/types.h> +#include <linux/percpu_counter.h> +#include <linux/ktime.h> + +extern bool disable_send_metrics; + +enum ceph_metric_type { + CLIENT_METRIC_TYPE_CAP_INFO, + CLIENT_METRIC_TYPE_READ_LATENCY, + CLIENT_METRIC_TYPE_WRITE_LATENCY, + CLIENT_METRIC_TYPE_METADATA_LATENCY, + CLIENT_METRIC_TYPE_DENTRY_LEASE, + CLIENT_METRIC_TYPE_OPENED_FILES, + CLIENT_METRIC_TYPE_PINNED_ICAPS, + CLIENT_METRIC_TYPE_OPENED_INODES, + CLIENT_METRIC_TYPE_READ_IO_SIZES, + CLIENT_METRIC_TYPE_WRITE_IO_SIZES, + CLIENT_METRIC_TYPE_AVG_READ_LATENCY, + CLIENT_METRIC_TYPE_STDEV_READ_LATENCY, + CLIENT_METRIC_TYPE_AVG_WRITE_LATENCY, + CLIENT_METRIC_TYPE_STDEV_WRITE_LATENCY, + CLIENT_METRIC_TYPE_AVG_METADATA_LATENCY, + CLIENT_METRIC_TYPE_STDEV_METADATA_LATENCY, + + CLIENT_METRIC_TYPE_MAX = CLIENT_METRIC_TYPE_STDEV_METADATA_LATENCY, +}; + +/* + * This will always have the highest metric bit value + * as the last element of the array. + */ +#define CEPHFS_METRIC_SPEC_CLIENT_SUPPORTED { \ + CLIENT_METRIC_TYPE_CAP_INFO, \ + CLIENT_METRIC_TYPE_READ_LATENCY, \ + CLIENT_METRIC_TYPE_WRITE_LATENCY, \ + CLIENT_METRIC_TYPE_METADATA_LATENCY, \ + CLIENT_METRIC_TYPE_DENTRY_LEASE, \ + CLIENT_METRIC_TYPE_OPENED_FILES, \ + CLIENT_METRIC_TYPE_PINNED_ICAPS, \ + CLIENT_METRIC_TYPE_OPENED_INODES, \ + CLIENT_METRIC_TYPE_READ_IO_SIZES, \ + CLIENT_METRIC_TYPE_WRITE_IO_SIZES, \ + CLIENT_METRIC_TYPE_AVG_READ_LATENCY, \ + CLIENT_METRIC_TYPE_STDEV_READ_LATENCY, \ + CLIENT_METRIC_TYPE_AVG_WRITE_LATENCY, \ + CLIENT_METRIC_TYPE_STDEV_WRITE_LATENCY, \ + CLIENT_METRIC_TYPE_AVG_METADATA_LATENCY, \ + CLIENT_METRIC_TYPE_STDEV_METADATA_LATENCY, \ + \ + CLIENT_METRIC_TYPE_MAX, \ +} + +struct ceph_metric_header { + __le32 type; /* ceph metric type */ + __u8 ver; + __u8 compat; + __le32 data_len; /* length of sizeof(hit + mis + total) */ +} __packed; + +/* metric caps header */ +struct ceph_metric_cap { + struct ceph_metric_header header; + __le64 hit; + __le64 mis; + __le64 total; +} __packed; + +/* metric read latency header */ +struct ceph_metric_read_latency { + struct ceph_metric_header header; + struct ceph_timespec lat; + struct ceph_timespec avg; + __le64 sq_sum; + __le64 count; +} __packed; + +/* metric write latency header */ +struct ceph_metric_write_latency { + struct ceph_metric_header header; + struct ceph_timespec lat; + struct ceph_timespec avg; + __le64 sq_sum; + __le64 count; +} __packed; + +/* metric metadata latency header */ +struct ceph_metric_metadata_latency { + struct ceph_metric_header header; + struct ceph_timespec lat; + struct ceph_timespec avg; + __le64 sq_sum; + __le64 count; +} __packed; + +/* metric dentry lease header */ +struct ceph_metric_dlease { + struct ceph_metric_header header; + __le64 hit; + __le64 mis; + __le64 total; +} __packed; + +/* metric opened files header */ +struct ceph_opened_files { + struct ceph_metric_header header; + __le64 opened_files; + __le64 total; +} __packed; + +/* metric pinned i_caps header */ +struct ceph_pinned_icaps { + struct ceph_metric_header header; + __le64 pinned_icaps; + __le64 total; +} __packed; + +/* metric opened inodes header */ +struct ceph_opened_inodes { + struct ceph_metric_header header; + __le64 opened_inodes; + __le64 total; +} __packed; + +/* metric read io size header */ +struct ceph_read_io_size { + struct ceph_metric_header header; + __le64 total_ops; + __le64 total_size; +} __packed; + +/* metric write io size header */ +struct ceph_write_io_size { + struct ceph_metric_header header; + __le64 total_ops; + __le64 total_size; +} __packed; + +struct ceph_metric_head { + __le32 num; /* the number of metrics that will be sent */ +} __packed; + +enum metric_type { + METRIC_READ, + METRIC_WRITE, + METRIC_METADATA, + METRIC_COPYFROM, + METRIC_MAX +}; + +struct ceph_metric { + spinlock_t lock; + u64 total; + u64 size_sum; + u64 size_min; + u64 size_max; + ktime_t latency_sum; + ktime_t latency_avg; + ktime_t latency_sq_sum; + ktime_t latency_min; + ktime_t latency_max; +}; + +/* This is the global metrics */ +struct ceph_client_metric { + atomic64_t total_dentries; + struct percpu_counter d_lease_hit; + struct percpu_counter d_lease_mis; + + atomic64_t total_caps; + struct percpu_counter i_caps_hit; + struct percpu_counter i_caps_mis; + + struct ceph_metric metric[METRIC_MAX]; + + /* The total number of directories and files that are opened */ + atomic64_t opened_files; + + /* The total number of inodes that have opened files or directories */ + struct percpu_counter opened_inodes; + struct percpu_counter total_inodes; + + struct ceph_mds_session *session; + struct delayed_work delayed_work; /* delayed work */ +}; + +static inline void metric_schedule_delayed(struct ceph_client_metric *m) +{ + if (disable_send_metrics) + return; + + /* per second */ + schedule_delayed_work(&m->delayed_work, round_jiffies_relative(HZ)); +} + +extern int ceph_metric_init(struct ceph_client_metric *m); +extern void ceph_metric_destroy(struct ceph_client_metric *m); + +static inline void ceph_update_cap_hit(struct ceph_client_metric *m) +{ + percpu_counter_inc(&m->i_caps_hit); +} + +static inline void ceph_update_cap_mis(struct ceph_client_metric *m) +{ + percpu_counter_inc(&m->i_caps_mis); +} + +extern void ceph_update_metrics(struct ceph_metric *m, + ktime_t r_start, ktime_t r_end, + unsigned int size, int rc); + +static inline void ceph_update_read_metrics(struct ceph_client_metric *m, + ktime_t r_start, ktime_t r_end, + unsigned int size, int rc) +{ + ceph_update_metrics(&m->metric[METRIC_READ], + r_start, r_end, size, rc); +} +static inline void ceph_update_write_metrics(struct ceph_client_metric *m, + ktime_t r_start, ktime_t r_end, + unsigned int size, int rc) +{ + ceph_update_metrics(&m->metric[METRIC_WRITE], + r_start, r_end, size, rc); +} +static inline void ceph_update_metadata_metrics(struct ceph_client_metric *m, + ktime_t r_start, ktime_t r_end, + int rc) +{ + ceph_update_metrics(&m->metric[METRIC_METADATA], + r_start, r_end, 0, rc); +} +static inline void ceph_update_copyfrom_metrics(struct ceph_client_metric *m, + ktime_t r_start, ktime_t r_end, + unsigned int size, int rc) +{ + ceph_update_metrics(&m->metric[METRIC_COPYFROM], + r_start, r_end, size, rc); +} +#endif /* _FS_CEPH_MDS_METRIC_H */ diff --git a/fs/ceph/quota.c b/fs/ceph/quota.c index de56dee60540..64592adfe48f 100644 --- a/fs/ceph/quota.c +++ b/fs/ceph/quota.c @@ -12,7 +12,7 @@ void ceph_adjust_quota_realms_count(struct inode *inode, bool inc) { - struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc; + struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(inode->i_sb); if (inc) atomic64_inc(&mdsc->quotarealms_count); else @@ -21,14 +21,17 @@ void ceph_adjust_quota_realms_count(struct inode *inode, bool inc) static inline bool ceph_has_realms_with_quotas(struct inode *inode) { - struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc; - struct super_block *sb = mdsc->fsc->sb; + struct super_block *sb = inode->i_sb; + struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(sb); + struct inode *root = d_inode(sb->s_root); if (atomic64_read(&mdsc->quotarealms_count) > 0) return true; /* if root is the real CephFS root, we don't have quota realms */ - if (sb->s_root->d_inode && - (sb->s_root->d_inode->i_ino == CEPH_INO_ROOT)) + if (root && ceph_ino(root) == CEPH_INO_ROOT) + return false; + /* MDS stray dirs have no quota realms */ + if (ceph_vino_is_reserved(ceph_inode(inode)->i_vino)) return false; /* otherwise, we can't know for sure */ return true; @@ -53,7 +56,7 @@ void ceph_handle_quota(struct ceph_mds_client *mdsc, /* increment msg sequence number */ mutex_lock(&session->s_mutex); - session->s_seq++; + inc_session_sequence(session); mutex_unlock(&session->s_mutex); /* lookup inode */ @@ -74,8 +77,7 @@ void ceph_handle_quota(struct ceph_mds_client *mdsc, le64_to_cpu(h->max_files)); spin_unlock(&ci->i_ceph_lock); - /* avoid calling iput_final() in dispatch thread */ - ceph_async_iput(inode); + iput(inode); } static struct ceph_quotarealm_inode * @@ -159,8 +161,8 @@ static struct inode *lookup_quotarealm_inode(struct ceph_mds_client *mdsc, } if (IS_ERR(in)) { - pr_warn("Can't lookup inode %llx (err: %ld)\n", - realm->ino, PTR_ERR(in)); + dout("Can't lookup inode %llx (err: %ld)\n", + realm->ino, PTR_ERR(in)); qri->timeout = jiffies + msecs_to_jiffies(60 * 1000); /* XXX */ } else { qri->timeout = 0; @@ -193,9 +195,9 @@ void ceph_cleanup_quotarealms_inodes(struct ceph_mds_client *mdsc) /* * This function walks through the snaprealm for an inode and returns the - * ceph_snap_realm for the first snaprealm that has quotas set (either max_files - * or max_bytes). If the root is reached, return the root ceph_snap_realm - * instead. + * ceph_snap_realm for the first snaprealm that has quotas set (max_files, + * max_bytes, or any, depending on the 'which_quota' argument). If the root is + * reached, return the root ceph_snap_realm instead. * * Note that the caller is responsible for calling ceph_put_snap_realm() on the * returned realm. @@ -207,7 +209,9 @@ void ceph_cleanup_quotarealms_inodes(struct ceph_mds_client *mdsc) * will be restarted. */ static struct ceph_snap_realm *get_quota_realm(struct ceph_mds_client *mdsc, - struct inode *inode, bool retry) + struct inode *inode, + enum quota_get_realm which_quota, + bool retry) { struct ceph_inode_info *ci = NULL; struct ceph_snap_realm *realm, *next; @@ -246,9 +250,8 @@ restart: } ci = ceph_inode(in); - has_quota = __ceph_has_any_quota(ci); - /* avoid calling iput_final() while holding mdsc->snap_rwsem */ - ceph_async_iput(in); + has_quota = __ceph_has_quota(ci, which_quota); + iput(in); next = realm->parent; if (has_quota || !next) @@ -266,7 +269,7 @@ restart: bool ceph_quota_is_same_realm(struct inode *old, struct inode *new) { - struct ceph_mds_client *mdsc = ceph_inode_to_client(old)->mdsc; + struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(old->i_sb); struct ceph_snap_realm *old_realm, *new_realm; bool is_same; @@ -278,8 +281,8 @@ restart: * dropped and we can then restart the whole operation. */ down_read(&mdsc->snap_rwsem); - old_realm = get_quota_realm(mdsc, old, true); - new_realm = get_quota_realm(mdsc, new, false); + old_realm = get_quota_realm(mdsc, old, QUOTA_GET_ANY, true); + new_realm = get_quota_realm(mdsc, new, QUOTA_GET_ANY, false); if (PTR_ERR(new_realm) == -EAGAIN) { up_read(&mdsc->snap_rwsem); if (old_realm) @@ -313,7 +316,7 @@ enum quota_check_op { static bool check_quota_exceeded(struct inode *inode, enum quota_check_op op, loff_t delta) { - struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc; + struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(inode->i_sb); struct ceph_inode_info *ci; struct ceph_snap_realm *realm, *next; struct inode *in; @@ -361,8 +364,6 @@ restart: spin_unlock(&ci->i_ceph_lock); switch (op) { case QUOTA_CHECK_MAX_FILES_OP: - exceeded = (max && (rvalue >= max)); - break; case QUOTA_CHECK_MAX_BYTES_OP: exceeded = (max && (rvalue + delta > max)); break; @@ -385,8 +386,7 @@ restart: pr_warn("Invalid quota check op (%d)\n", op); exceeded = true; /* Just break the loop */ } - /* avoid calling iput_final() while holding mdsc->snap_rwsem */ - ceph_async_iput(in); + iput(in); next = realm->parent; if (exceeded || !next) @@ -417,7 +417,7 @@ bool ceph_quota_is_max_files_exceeded(struct inode *inode) WARN_ON(!S_ISDIR(inode->i_mode)); - return check_quota_exceeded(inode, QUOTA_CHECK_MAX_FILES_OP, 0); + return check_quota_exceeded(inode, QUOTA_CHECK_MAX_FILES_OP, 1); } /* @@ -485,7 +485,8 @@ bool ceph_quota_update_statfs(struct ceph_fs_client *fsc, struct kstatfs *buf) bool is_updated = false; down_read(&mdsc->snap_rwsem); - realm = get_quota_realm(mdsc, d_inode(fsc->sb->s_root), true); + realm = get_quota_realm(mdsc, d_inode(fsc->sb->s_root), + QUOTA_GET_MAX_BYTES, true); up_read(&mdsc->snap_rwsem); if (!realm) return false; @@ -499,10 +500,24 @@ bool ceph_quota_update_statfs(struct ceph_fs_client *fsc, struct kstatfs *buf) if (ci->i_max_bytes) { total = ci->i_max_bytes >> CEPH_BLOCK_SHIFT; used = ci->i_rbytes >> CEPH_BLOCK_SHIFT; + /* For quota size less than 4MB, use 4KB block size */ + if (!total) { + total = ci->i_max_bytes >> CEPH_4K_BLOCK_SHIFT; + used = ci->i_rbytes >> CEPH_4K_BLOCK_SHIFT; + buf->f_frsize = 1 << CEPH_4K_BLOCK_SHIFT; + } /* It is possible for a quota to be exceeded. * Report 'zero' in that case */ free = total > used ? total - used : 0; + /* For quota size less than 4KB, report the + * total=used=4KB,free=0 when quota is full + * and total=free=4KB, used=0 otherwise */ + if (!total) { + total = 1; + free = ci->i_max_bytes > ci->i_rbytes ? 1 : 0; + buf->f_frsize = 1 << CEPH_4K_BLOCK_SHIFT; + } } spin_unlock(&ci->i_ceph_lock); if (total) { diff --git a/fs/ceph/snap.c b/fs/ceph/snap.c index ccfcc66aaf44..864cdaa0d2bd 100644 --- a/fs/ceph/snap.c +++ b/fs/ceph/snap.c @@ -60,24 +60,26 @@ /* * increase ref count for the realm * - * caller must hold snap_rwsem for write. + * caller must hold snap_rwsem. */ void ceph_get_snap_realm(struct ceph_mds_client *mdsc, struct ceph_snap_realm *realm) { - dout("get_realm %p %d -> %d\n", realm, - atomic_read(&realm->nref), atomic_read(&realm->nref)+1); + lockdep_assert_held(&mdsc->snap_rwsem); + /* - * since we _only_ increment realm refs or empty the empty - * list with snap_rwsem held, adjusting the empty list here is - * safe. we do need to protect against concurrent empty list - * additions, however. + * The 0->1 and 1->0 transitions must take the snap_empty_lock + * atomically with the refcount change. Go ahead and bump the + * nref here, unless it's 0, in which case we take the spinlock + * and then do the increment and remove it from the list. */ - if (atomic_inc_return(&realm->nref) == 1) { - spin_lock(&mdsc->snap_empty_lock); + if (atomic_inc_not_zero(&realm->nref)) + return; + + spin_lock(&mdsc->snap_empty_lock); + if (atomic_inc_return(&realm->nref) == 1) list_del_init(&realm->empty_item); - spin_unlock(&mdsc->snap_empty_lock); - } + spin_unlock(&mdsc->snap_empty_lock); } static void __insert_snap_realm(struct rb_root *root, @@ -113,29 +115,36 @@ static struct ceph_snap_realm *ceph_create_snap_realm( { struct ceph_snap_realm *realm; + lockdep_assert_held_write(&mdsc->snap_rwsem); + realm = kzalloc(sizeof(*realm), GFP_NOFS); if (!realm) return ERR_PTR(-ENOMEM); - atomic_set(&realm->nref, 1); /* for caller */ + /* Do not release the global dummy snaprealm until unmouting */ + if (ino == CEPH_INO_GLOBAL_SNAPREALM) + atomic_set(&realm->nref, 2); + else + atomic_set(&realm->nref, 1); realm->ino = ino; INIT_LIST_HEAD(&realm->children); INIT_LIST_HEAD(&realm->child_item); INIT_LIST_HEAD(&realm->empty_item); INIT_LIST_HEAD(&realm->dirty_item); + INIT_LIST_HEAD(&realm->rebuild_item); INIT_LIST_HEAD(&realm->inodes_with_caps); spin_lock_init(&realm->inodes_with_caps_lock); __insert_snap_realm(&mdsc->snap_realms, realm); mdsc->num_snap_realms++; - dout("create_snap_realm %llx %p\n", realm->ino, realm); + dout("%s %llx %p\n", __func__, realm->ino, realm); return realm; } /* * lookup the realm rooted at @ino. * - * caller must hold snap_rwsem for write. + * caller must hold snap_rwsem. */ static struct ceph_snap_realm *__lookup_snap_realm(struct ceph_mds_client *mdsc, u64 ino) @@ -143,6 +152,8 @@ static struct ceph_snap_realm *__lookup_snap_realm(struct ceph_mds_client *mdsc, struct rb_node *n = mdsc->snap_realms.rb_node; struct ceph_snap_realm *r; + lockdep_assert_held(&mdsc->snap_rwsem); + while (n) { r = rb_entry(n, struct ceph_snap_realm, node); if (ino < r->ino) @@ -150,7 +161,7 @@ static struct ceph_snap_realm *__lookup_snap_realm(struct ceph_mds_client *mdsc, else if (ino > r->ino) n = n->rb_right; else { - dout("lookup_snap_realm %llx %p\n", r->ino, r); + dout("%s %llx %p\n", __func__, r->ino, r); return r; } } @@ -176,7 +187,9 @@ static void __put_snap_realm(struct ceph_mds_client *mdsc, static void __destroy_snap_realm(struct ceph_mds_client *mdsc, struct ceph_snap_realm *realm) { - dout("__destroy_snap_realm %p %llx\n", realm, realm->ino); + lockdep_assert_held_write(&mdsc->snap_rwsem); + + dout("%s %p %llx\n", __func__, realm, realm->ino); rb_erase(&realm->node, &mdsc->snap_realms); mdsc->num_snap_realms--; @@ -198,28 +211,30 @@ static void __destroy_snap_realm(struct ceph_mds_client *mdsc, static void __put_snap_realm(struct ceph_mds_client *mdsc, struct ceph_snap_realm *realm) { - dout("__put_snap_realm %llx %p %d -> %d\n", realm->ino, realm, - atomic_read(&realm->nref), atomic_read(&realm->nref)-1); + lockdep_assert_held_write(&mdsc->snap_rwsem); + + /* + * We do not require the snap_empty_lock here, as any caller that + * increments the value must hold the snap_rwsem. + */ if (atomic_dec_and_test(&realm->nref)) __destroy_snap_realm(mdsc, realm); } /* - * caller needn't hold any locks + * See comments in ceph_get_snap_realm. Caller needn't hold any locks. */ void ceph_put_snap_realm(struct ceph_mds_client *mdsc, struct ceph_snap_realm *realm) { - dout("put_snap_realm %llx %p %d -> %d\n", realm->ino, realm, - atomic_read(&realm->nref), atomic_read(&realm->nref)-1); - if (!atomic_dec_and_test(&realm->nref)) + if (!atomic_dec_and_lock(&realm->nref, &mdsc->snap_empty_lock)) return; if (down_write_trylock(&mdsc->snap_rwsem)) { + spin_unlock(&mdsc->snap_empty_lock); __destroy_snap_realm(mdsc, realm); up_write(&mdsc->snap_rwsem); } else { - spin_lock(&mdsc->snap_empty_lock); list_add(&realm->empty_item, &mdsc->snap_empty); spin_unlock(&mdsc->snap_empty_lock); } @@ -236,6 +251,8 @@ static void __cleanup_empty_realms(struct ceph_mds_client *mdsc) { struct ceph_snap_realm *realm; + lockdep_assert_held_write(&mdsc->snap_rwsem); + spin_lock(&mdsc->snap_empty_lock); while (!list_empty(&mdsc->snap_empty)) { realm = list_first_entry(&mdsc->snap_empty, @@ -248,9 +265,14 @@ static void __cleanup_empty_realms(struct ceph_mds_client *mdsc) spin_unlock(&mdsc->snap_empty_lock); } -void ceph_cleanup_empty_realms(struct ceph_mds_client *mdsc) +void ceph_cleanup_global_and_empty_realms(struct ceph_mds_client *mdsc) { + struct ceph_snap_realm *global_realm; + down_write(&mdsc->snap_rwsem); + global_realm = __lookup_snap_realm(mdsc, CEPH_INO_GLOBAL_SNAPREALM); + if (global_realm) + ceph_put_snap_realm(mdsc, global_realm); __cleanup_empty_realms(mdsc); up_write(&mdsc->snap_rwsem); } @@ -269,6 +291,8 @@ static int adjust_snap_realm_parent(struct ceph_mds_client *mdsc, { struct ceph_snap_realm *parent; + lockdep_assert_held_write(&mdsc->snap_rwsem); + if (realm->parent_ino == parentino) return 0; @@ -278,9 +302,8 @@ static int adjust_snap_realm_parent(struct ceph_mds_client *mdsc, if (IS_ERR(parent)) return PTR_ERR(parent); } - dout("adjust_snap_realm_parent %llx %p: %llx %p -> %llx %p\n", - realm->ino, realm, realm->parent_ino, realm->parent, - parentino, parent); + dout("%s %llx %p: %llx %p -> %llx %p\n", __func__, realm->ino, + realm, realm->parent_ino, realm->parent, parentino, parent); if (realm->parent) { list_del_init(&realm->child_item); ceph_put_snap_realm(mdsc, realm->parent); @@ -306,7 +329,8 @@ static int cmpu64_rev(const void *a, const void *b) * build the snap context for a given realm. */ static int build_snap_context(struct ceph_snap_realm *realm, - struct list_head* dirty_realms) + struct list_head *realm_queue, + struct list_head *dirty_realms) { struct ceph_snap_realm *parent = realm->parent; struct ceph_snap_context *snapc; @@ -320,9 +344,9 @@ static int build_snap_context(struct ceph_snap_realm *realm, */ if (parent) { if (!parent->cached_context) { - err = build_snap_context(parent, dirty_realms); - if (err) - goto fail; + /* add to the queue head */ + list_add(&parent->rebuild_item, realm_queue); + return 1; } num += parent->cached_context->num_snaps; } @@ -335,9 +359,8 @@ static int build_snap_context(struct ceph_snap_realm *realm, realm->cached_context->seq == realm->seq && (!parent || realm->cached_context->seq >= parent->cached_context->seq)) { - dout("build_snap_context %llx %p: %p seq %lld (%u snaps)" - " (unchanged)\n", - realm->ino, realm, realm->cached_context, + dout("%s %llx %p: %p seq %lld (%u snaps) (unchanged)\n", + __func__, realm->ino, realm, realm->cached_context, realm->cached_context->seq, (unsigned int)realm->cached_context->num_snaps); return 0; @@ -376,9 +399,8 @@ static int build_snap_context(struct ceph_snap_realm *realm, sort(snapc->snaps, num, sizeof(u64), cmpu64_rev, NULL); snapc->num_snaps = num; - dout("build_snap_context %llx %p: %p seq %lld (%u snaps)\n", - realm->ino, realm, snapc, snapc->seq, - (unsigned int) snapc->num_snaps); + dout("%s %llx %p: %p seq %lld (%u snaps)\n", __func__, realm->ino, + realm, snapc, snapc->seq, (unsigned int) snapc->num_snaps); ceph_put_snap_context(realm->cached_context); realm->cached_context = snapc; @@ -395,8 +417,7 @@ fail: ceph_put_snap_context(realm->cached_context); realm->cached_context = NULL; } - pr_err("build_snap_context %llx %p fail %d\n", realm->ino, - realm, err); + pr_err("%s %llx %p fail %d\n", __func__, realm->ino, realm, err); return err; } @@ -406,13 +427,50 @@ fail: static void rebuild_snap_realms(struct ceph_snap_realm *realm, struct list_head *dirty_realms) { - struct ceph_snap_realm *child; + LIST_HEAD(realm_queue); + int last = 0; + bool skip = false; + + list_add_tail(&realm->rebuild_item, &realm_queue); + + while (!list_empty(&realm_queue)) { + struct ceph_snap_realm *_realm, *child; + + _realm = list_first_entry(&realm_queue, + struct ceph_snap_realm, + rebuild_item); + + /* + * If the last building failed dues to memory + * issue, just empty the realm_queue and return + * to avoid infinite loop. + */ + if (last < 0) { + list_del_init(&_realm->rebuild_item); + continue; + } + + last = build_snap_context(_realm, &realm_queue, dirty_realms); + dout("%s %llx %p, %s\n", __func__, _realm->ino, _realm, + last > 0 ? "is deferred" : !last ? "succeeded" : "failed"); - dout("rebuild_snap_realms %llx %p\n", realm->ino, realm); - build_snap_context(realm, dirty_realms); + /* is any child in the list ? */ + list_for_each_entry(child, &_realm->children, child_item) { + if (!list_empty(&child->rebuild_item)) { + skip = true; + break; + } + } - list_for_each_entry(child, &realm->children, child_item) - rebuild_snap_realms(child, dirty_realms); + if (!skip) { + list_for_each_entry(child, &_realm->children, child_item) + list_add_tail(&child->rebuild_item, &realm_queue); + } + + /* last == 1 means need to build parent first */ + if (last <= 0) + list_del_init(&_realm->rebuild_item); + } } @@ -460,20 +518,15 @@ static bool has_new_snaps(struct ceph_snap_context *o, * Caller must hold snap_rwsem for read (i.e., the realm topology won't * change). */ -void ceph_queue_cap_snap(struct ceph_inode_info *ci) +static void ceph_queue_cap_snap(struct ceph_inode_info *ci, + struct ceph_cap_snap **pcapsnap) { - struct inode *inode = &ci->vfs_inode; - struct ceph_cap_snap *capsnap; + struct inode *inode = &ci->netfs.inode; struct ceph_snap_context *old_snapc, *new_snapc; + struct ceph_cap_snap *capsnap = *pcapsnap; struct ceph_buffer *old_blob = NULL; int used, dirty; - capsnap = kzalloc(sizeof(*capsnap), GFP_NOFS); - if (!capsnap) { - pr_err("ENOMEM allocating ceph_cap_snap on %p\n", inode); - return; - } - spin_lock(&ci->i_ceph_lock); used = __ceph_caps_used(ci); dirty = __ceph_caps_dirty(ci); @@ -494,12 +547,14 @@ void ceph_queue_cap_snap(struct ceph_inode_info *ci) as no new writes are allowed to start when pending, so any writes in progress now were started before the previous cap_snap. lucky us. */ - dout("queue_cap_snap %p already pending\n", inode); + dout("%s %p %llx.%llx already pending\n", + __func__, inode, ceph_vinop(inode)); goto update_snapc; } if (ci->i_wrbuffer_ref_head == 0 && !(dirty & (CEPH_CAP_ANY_EXCL|CEPH_CAP_FILE_WR))) { - dout("queue_cap_snap %p nothing dirty|writing\n", inode); + dout("%s %p %llx.%llx nothing dirty|writing\n", + __func__, inode, ceph_vinop(inode)); goto update_snapc; } @@ -519,20 +574,17 @@ void ceph_queue_cap_snap(struct ceph_inode_info *ci) } else { if (!(used & CEPH_CAP_FILE_WR) && ci->i_wrbuffer_ref_head == 0) { - dout("queue_cap_snap %p " - "no new_snap|dirty_page|writing\n", inode); + dout("%s %p %llx.%llx no new_snap|dirty_page|writing\n", + __func__, inode, ceph_vinop(inode)); goto update_snapc; } } - dout("queue_cap_snap %p cap_snap %p queuing under %p %s %s\n", - inode, capsnap, old_snapc, ceph_cap_string(dirty), - capsnap->need_flush ? "" : "no_flush"); + dout("%s %p %llx.%llx cap_snap %p queuing under %p %s %s\n", + __func__, inode, ceph_vinop(inode), capsnap, old_snapc, + ceph_cap_string(dirty), capsnap->need_flush ? "" : "no_flush"); ihold(inode); - refcount_set(&capsnap->nref, 1); - INIT_LIST_HEAD(&capsnap->ci_item); - capsnap->follows = old_snapc->seq; capsnap->issued = __ceph_caps_issued(ci, NULL); capsnap->dirty = dirty; @@ -562,31 +614,30 @@ void ceph_queue_cap_snap(struct ceph_inode_info *ci) list_add_tail(&capsnap->ci_item, &ci->i_cap_snaps); if (used & CEPH_CAP_FILE_WR) { - dout("queue_cap_snap %p cap_snap %p snapc %p" - " seq %llu used WR, now pending\n", inode, + dout("%s %p %llx.%llx cap_snap %p snapc %p seq %llu used WR," + " now pending\n", __func__, inode, ceph_vinop(inode), capsnap, old_snapc, old_snapc->seq); capsnap->writing = 1; } else { /* note mtime, size NOW. */ __ceph_finish_cap_snap(ci, capsnap); } - capsnap = NULL; + *pcapsnap = NULL; old_snapc = NULL; update_snapc: - if (ci->i_wrbuffer_ref_head == 0 && - ci->i_wr_ref == 0 && - ci->i_dirty_caps == 0 && - ci->i_flushing_caps == 0) { - ci->i_head_snapc = NULL; - } else { + if (ci->i_wrbuffer_ref_head == 0 && + ci->i_wr_ref == 0 && + ci->i_dirty_caps == 0 && + ci->i_flushing_caps == 0) { + ci->i_head_snapc = NULL; + } else { ci->i_head_snapc = ceph_get_snap_context(new_snapc); dout(" new snapc is %p\n", new_snapc); } spin_unlock(&ci->i_ceph_lock); ceph_buffer_put(old_blob); - kfree(capsnap); ceph_put_snap_context(old_snapc); } @@ -601,11 +652,11 @@ update_snapc: int __ceph_finish_cap_snap(struct ceph_inode_info *ci, struct ceph_cap_snap *capsnap) { - struct inode *inode = &ci->vfs_inode; - struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; + struct inode *inode = &ci->netfs.inode; + struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(inode->i_sb); BUG_ON(capsnap->writing); - capsnap->size = inode->i_size; + capsnap->size = i_size_read(inode); capsnap->mtime = inode->i_mtime; capsnap->atime = inode->i_atime; capsnap->ctime = inode->i_ctime; @@ -615,17 +666,28 @@ int __ceph_finish_cap_snap(struct ceph_inode_info *ci, capsnap->truncate_size = ci->i_truncate_size; capsnap->truncate_seq = ci->i_truncate_seq; if (capsnap->dirty_pages) { - dout("finish_cap_snap %p cap_snap %p snapc %p %llu %s s=%llu " - "still has %d dirty pages\n", inode, capsnap, - capsnap->context, capsnap->context->seq, - ceph_cap_string(capsnap->dirty), capsnap->size, - capsnap->dirty_pages); + dout("%s %p %llx.%llx cap_snap %p snapc %p %llu %s s=%llu " + "still has %d dirty pages\n", __func__, inode, + ceph_vinop(inode), capsnap, capsnap->context, + capsnap->context->seq, ceph_cap_string(capsnap->dirty), + capsnap->size, capsnap->dirty_pages); + return 0; + } + + /* Fb cap still in use, delay it */ + if (ci->i_wb_ref) { + dout("%s %p %llx.%llx cap_snap %p snapc %p %llu %s s=%llu " + "used WRBUFFER, delaying\n", __func__, inode, + ceph_vinop(inode), capsnap, capsnap->context, + capsnap->context->seq, ceph_cap_string(capsnap->dirty), + capsnap->size); + capsnap->writing = 1; return 0; } ci->i_ceph_flags |= CEPH_I_FLUSH_SNAPS; - dout("finish_cap_snap %p cap_snap %p snapc %p %llu %s s=%llu\n", - inode, capsnap, capsnap->context, + dout("%s %p %llx.%llx cap_snap %p snapc %p %llu %s s=%llu\n", + __func__, inode, ceph_vinop(inode), capsnap, capsnap->context, capsnap->context->seq, ceph_cap_string(capsnap->dirty), capsnap->size); @@ -644,26 +706,47 @@ static void queue_realm_cap_snaps(struct ceph_snap_realm *realm) { struct ceph_inode_info *ci; struct inode *lastinode = NULL; + struct ceph_cap_snap *capsnap = NULL; - dout("queue_realm_cap_snaps %p %llx inodes\n", realm, realm->ino); + dout("%s %p %llx inode\n", __func__, realm, realm->ino); spin_lock(&realm->inodes_with_caps_lock); list_for_each_entry(ci, &realm->inodes_with_caps, i_snap_realm_item) { - struct inode *inode = igrab(&ci->vfs_inode); + struct inode *inode = igrab(&ci->netfs.inode); if (!inode) continue; spin_unlock(&realm->inodes_with_caps_lock); - /* avoid calling iput_final() while holding - * mdsc->snap_rwsem or in mds dispatch threads */ - ceph_async_iput(lastinode); + iput(lastinode); lastinode = inode; - ceph_queue_cap_snap(ci); + + /* + * Allocate the capsnap memory outside of ceph_queue_cap_snap() + * to reduce very possible but unnecessary frequently memory + * allocate/free in this loop. + */ + if (!capsnap) { + capsnap = kmem_cache_zalloc(ceph_cap_snap_cachep, GFP_NOFS); + if (!capsnap) { + pr_err("ENOMEM allocating ceph_cap_snap on %p\n", + inode); + return; + } + } + capsnap->cap_flush.is_capsnap = true; + refcount_set(&capsnap->nref, 1); + INIT_LIST_HEAD(&capsnap->cap_flush.i_list); + INIT_LIST_HEAD(&capsnap->cap_flush.g_list); + INIT_LIST_HEAD(&capsnap->ci_item); + + ceph_queue_cap_snap(ci, &capsnap); spin_lock(&realm->inodes_with_caps_lock); } spin_unlock(&realm->inodes_with_caps_lock); - ceph_async_iput(lastinode); + iput(lastinode); - dout("queue_realm_cap_snaps %p %llx done\n", realm, realm->ino); + if (capsnap) + kmem_cache_free(ceph_cap_snap_cachep, capsnap); + dout("%s %p %llx done\n", __func__, realm, realm->ino); } /* @@ -682,12 +765,16 @@ int ceph_update_snap_trace(struct ceph_mds_client *mdsc, __le64 *prior_parent_snaps; /* encoded */ struct ceph_snap_realm *realm = NULL; struct ceph_snap_realm *first_realm = NULL; - int invalidate = 0; + struct ceph_snap_realm *realm_to_rebuild = NULL; + int rebuild_snapcs; int err = -ENOMEM; LIST_HEAD(dirty_realms); - dout("update_snap_trace deletion=%d\n", deletion); + lockdep_assert_held_write(&mdsc->snap_rwsem); + + dout("%s deletion=%d\n", __func__, deletion); more: + rebuild_snapcs = 0; ceph_decode_need(&p, e, sizeof(*ri), bad); ri = p; p += sizeof(*ri); @@ -711,10 +798,10 @@ more: err = adjust_snap_realm_parent(mdsc, realm, le64_to_cpu(ri->parent)); if (err < 0) goto fail; - invalidate += err; + rebuild_snapcs += err; if (le64_to_cpu(ri->seq) > realm->seq) { - dout("update_snap_trace updating %llx %p %lld -> %lld\n", + dout("%s updating %llx %p %lld -> %lld\n", __func__, realm->ino, realm, realm->seq, le64_to_cpu(ri->seq)); /* update realm parameters, snap lists */ realm->seq = le64_to_cpu(ri->seq); @@ -736,22 +823,30 @@ more: if (realm->seq > mdsc->last_snap_seq) mdsc->last_snap_seq = realm->seq; - invalidate = 1; + rebuild_snapcs = 1; } else if (!realm->cached_context) { - dout("update_snap_trace %llx %p seq %lld new\n", + dout("%s %llx %p seq %lld new\n", __func__, realm->ino, realm, realm->seq); - invalidate = 1; + rebuild_snapcs = 1; } else { - dout("update_snap_trace %llx %p seq %lld unchanged\n", + dout("%s %llx %p seq %lld unchanged\n", __func__, realm->ino, realm, realm->seq); } - dout("done with %llx %p, invalidated=%d, %p %p\n", realm->ino, - realm, invalidate, p, e); + dout("done with %llx %p, rebuild_snapcs=%d, %p %p\n", realm->ino, + realm, rebuild_snapcs, p, e); + + /* + * this will always track the uppest parent realm from which + * we need to rebuild the snapshot contexts _downward_ in + * hierarchy. + */ + if (rebuild_snapcs) + realm_to_rebuild = realm; - /* invalidate when we reach the _end_ (root) of the trace */ - if (invalidate && p >= e) - rebuild_snap_realms(realm, &dirty_realms); + /* rebuild_snapcs when we reach the _end_ (root) of the trace */ + if (realm_to_rebuild && p >= e) + rebuild_snap_realms(realm_to_rebuild, &dirty_realms); if (!first_realm) first_realm = realm; @@ -781,13 +876,13 @@ more: return 0; bad: - err = -EINVAL; + err = -EIO; fail: if (realm && !IS_ERR(realm)) ceph_put_snap_realm(mdsc, realm); if (first_realm) ceph_put_snap_realm(mdsc, first_realm); - pr_err("update_snap_trace error %d\n", err); + pr_err("%s error %d\n", __func__, err); return err; } @@ -804,29 +899,61 @@ static void flush_snaps(struct ceph_mds_client *mdsc) struct inode *inode; struct ceph_mds_session *session = NULL; - dout("flush_snaps\n"); + dout("%s\n", __func__); spin_lock(&mdsc->snap_flush_lock); while (!list_empty(&mdsc->snap_flush_list)) { ci = list_first_entry(&mdsc->snap_flush_list, struct ceph_inode_info, i_snap_flush_item); - inode = &ci->vfs_inode; + inode = &ci->netfs.inode; ihold(inode); spin_unlock(&mdsc->snap_flush_lock); ceph_flush_snaps(ci, &session); - /* avoid calling iput_final() while holding - * session->s_mutex or in mds dispatch threads */ - ceph_async_iput(inode); + iput(inode); spin_lock(&mdsc->snap_flush_lock); } spin_unlock(&mdsc->snap_flush_lock); - if (session) { - mutex_unlock(&session->s_mutex); - ceph_put_mds_session(session); - } - dout("flush_snaps done\n"); + ceph_put_mds_session(session); + dout("%s done\n", __func__); } +/** + * ceph_change_snap_realm - change the snap_realm for an inode + * @inode: inode to move to new snap realm + * @realm: new realm to move inode into (may be NULL) + * + * Detach an inode from its old snaprealm (if any) and attach it to + * the new snaprealm (if any). The old snap realm reference held by + * the inode is put. If realm is non-NULL, then the caller's reference + * to it is taken over by the inode. + */ +void ceph_change_snap_realm(struct inode *inode, struct ceph_snap_realm *realm) +{ + struct ceph_inode_info *ci = ceph_inode(inode); + struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc; + struct ceph_snap_realm *oldrealm = ci->i_snap_realm; + + lockdep_assert_held(&ci->i_ceph_lock); + + if (oldrealm) { + spin_lock(&oldrealm->inodes_with_caps_lock); + list_del_init(&ci->i_snap_realm_item); + if (oldrealm->ino == ci->i_vino.ino) + oldrealm->inode = NULL; + spin_unlock(&oldrealm->inodes_with_caps_lock); + ceph_put_snap_realm(mdsc, oldrealm); + } + + ci->i_snap_realm = realm; + + if (realm) { + spin_lock(&realm->inodes_with_caps_lock); + list_add(&ci->i_snap_realm_item, &realm->inodes_with_caps); + if (realm->ino == ci->i_vino.ino) + realm->inode = inode; + spin_unlock(&realm->inodes_with_caps_lock); + } +} /* * Handle a snap notification from the MDS. @@ -869,11 +996,11 @@ void ceph_handle_snap(struct ceph_mds_client *mdsc, trace_len = le32_to_cpu(h->trace_len); p += sizeof(*h); - dout("handle_snap from mds%d op %s split %llx tracelen %d\n", mds, - ceph_snap_op_name(op), split, trace_len); + dout("%s from mds%d op %s split %llx tracelen %d\n", __func__, + mds, ceph_snap_op_name(op), split, trace_len); mutex_lock(&session->s_mutex); - session->s_seq++; + inc_session_sequence(session); mutex_unlock(&session->s_mutex); down_write(&mdsc->snap_rwsem); @@ -913,7 +1040,6 @@ void ceph_handle_snap(struct ceph_mds_client *mdsc, }; struct inode *inode = ceph_find_inode(sb, vino); struct ceph_inode_info *ci; - struct ceph_snap_realm *oldrealm; if (!inode) continue; @@ -931,42 +1057,23 @@ void ceph_handle_snap(struct ceph_mds_client *mdsc, */ if (ci->i_snap_realm->created > le64_to_cpu(ri->created)) { - dout(" leaving %p in newer realm %llx %p\n", - inode, ci->i_snap_realm->ino, + dout(" leaving %p %llx.%llx in newer realm %llx %p\n", + inode, ceph_vinop(inode), ci->i_snap_realm->ino, ci->i_snap_realm); goto skip_inode; } - dout(" will move %p to split realm %llx %p\n", - inode, realm->ino, realm); - /* - * Move the inode to the new realm - */ - oldrealm = ci->i_snap_realm; - spin_lock(&oldrealm->inodes_with_caps_lock); - list_del_init(&ci->i_snap_realm_item); - spin_unlock(&oldrealm->inodes_with_caps_lock); - - spin_lock(&realm->inodes_with_caps_lock); - list_add(&ci->i_snap_realm_item, - &realm->inodes_with_caps); - ci->i_snap_realm = realm; - if (realm->ino == ci->i_vino.ino) - realm->inode = inode; - spin_unlock(&realm->inodes_with_caps_lock); - - spin_unlock(&ci->i_ceph_lock); + dout(" will move %p %llx.%llx to split realm %llx %p\n", + inode, ceph_vinop(inode), realm->ino, realm); ceph_get_snap_realm(mdsc, realm); - ceph_put_snap_realm(mdsc, oldrealm); - - /* avoid calling iput_final() while holding - * mdsc->snap_rwsem or mds in dispatch threads */ - ceph_async_iput(inode); + ceph_change_snap_realm(inode, realm); + spin_unlock(&ci->i_ceph_lock); + iput(inode); continue; skip_inode: spin_unlock(&ci->i_ceph_lock); - ceph_async_iput(inode); + iput(inode); } /* we may have taken some of the old realm's children. */ @@ -999,7 +1106,7 @@ skip_inode: return; bad: - pr_err("corrupt snap message from mds%d\n", mds); + pr_err("%s corrupt snap message from mds%d\n", __func__, mds); ceph_msg_dump(msg); out: if (locked_rwsem) @@ -1032,7 +1139,8 @@ struct ceph_snapid_map* ceph_get_snapid_map(struct ceph_mds_client *mdsc, } spin_unlock(&mdsc->snapid_map_lock); if (exist) { - dout("found snapid map %llx -> %x\n", exist->snap, exist->dev); + dout("%s found snapid map %llx -> %x\n", __func__, + exist->snap, exist->dev); return exist; } @@ -1076,11 +1184,13 @@ struct ceph_snapid_map* ceph_get_snapid_map(struct ceph_mds_client *mdsc, if (exist) { free_anon_bdev(sm->dev); kfree(sm); - dout("found snapid map %llx -> %x\n", exist->snap, exist->dev); + dout("%s found snapid map %llx -> %x\n", __func__, + exist->snap, exist->dev); return exist; } - dout("create snapid map %llx -> %x\n", sm->snap, sm->dev); + dout("%s create snapid map %llx -> %x\n", __func__, + sm->snap, sm->dev); return sm; } @@ -1155,5 +1265,6 @@ void ceph_cleanup_snapid_map(struct ceph_mds_client *mdsc) pr_err("snapid map %llx -> %x still in use\n", sm->snap, sm->dev); } + kfree(sm); } } diff --git a/fs/ceph/strings.c b/fs/ceph/strings.c index 4a79f3632260..e36e8948e728 100644 --- a/fs/ceph/strings.c +++ b/fs/ceph/strings.c @@ -46,6 +46,7 @@ const char *ceph_session_op_name(int op) case CEPH_SESSION_FLUSHMSG_ACK: return "flushmsg_ack"; case CEPH_SESSION_FORCE_RO: return "force_ro"; case CEPH_SESSION_REJECT: return "reject"; + case CEPH_SESSION_REQUEST_FLUSH_MDLOG: return "flush_mdlog"; } return "???"; } @@ -59,6 +60,7 @@ const char *ceph_mds_op_name(int op) case CEPH_MDS_OP_LOOKUPINO: return "lookupino"; case CEPH_MDS_OP_LOOKUPNAME: return "lookupname"; case CEPH_MDS_OP_GETATTR: return "getattr"; + case CEPH_MDS_OP_GETVXATTR: return "getvxattr"; case CEPH_MDS_OP_SETXATTR: return "setxattr"; case CEPH_MDS_OP_SETATTR: return "setattr"; case CEPH_MDS_OP_RMXATTR: return "rmxattr"; diff --git a/fs/ceph/super.c b/fs/ceph/super.c index c7f150686a53..3fc48b43cab0 100644 --- a/fs/ceph/super.c +++ b/fs/ceph/super.c @@ -27,6 +27,11 @@ #include <linux/ceph/auth.h> #include <linux/ceph/debugfs.h> +#include <uapi/linux/magic.h> + +static DEFINE_SPINLOCK(ceph_fsc_lock); +static LIST_HEAD(ceph_fsc_list); + /* * Ceph superblock operations * @@ -49,8 +54,7 @@ static int ceph_statfs(struct dentry *dentry, struct kstatfs *buf) struct ceph_fs_client *fsc = ceph_inode_to_client(d_inode(dentry)); struct ceph_mon_client *monc = &fsc->client->monc; struct ceph_statfs st; - u64 fsid; - int err; + int i, err; u64 data_pool; if (fsc->mdsc->mdsmap->m_num_data_pg_pools == 1) { @@ -68,15 +72,9 @@ static int ceph_statfs(struct dentry *dentry, struct kstatfs *buf) buf->f_type = CEPH_SUPER_MAGIC; /* ?? */ /* - * express utilization in terms of large blocks to avoid + * Express utilization in terms of large blocks to avoid * overflow on 32-bit machines. - * - * NOTE: for the time being, we make bsize == frsize to humor - * not-yet-ancient versions of glibc that are broken. - * Someday, we will probably want to report a real block - * size... whatever that may mean for a network file system! */ - buf->f_bsize = 1 << CEPH_BLOCK_SHIFT; buf->f_frsize = 1 << CEPH_BLOCK_SHIFT; /* @@ -91,18 +89,27 @@ static int ceph_statfs(struct dentry *dentry, struct kstatfs *buf) buf->f_bavail = le64_to_cpu(st.kb_avail) >> (CEPH_BLOCK_SHIFT-10); } + /* + * NOTE: for the time being, we make bsize == frsize to humor + * not-yet-ancient versions of glibc that are broken. + * Someday, we will probably want to report a real block + * size... whatever that may mean for a network file system! + */ + buf->f_bsize = buf->f_frsize; + buf->f_files = le64_to_cpu(st.num_objects); buf->f_ffree = -1; buf->f_namelen = NAME_MAX; /* Must convert the fsid, for consistent values across arches */ + buf->f_fsid.val[0] = 0; mutex_lock(&monc->mutex); - fsid = le64_to_cpu(*(__le64 *)(&monc->monmap->fsid)) ^ - le64_to_cpu(*((__le64 *)&monc->monmap->fsid + 1)); + for (i = 0 ; i < sizeof(monc->monmap->fsid) / sizeof(__le32) ; ++i) + buf->f_fsid.val[0] ^= le32_to_cpu(((__le32 *)&monc->monmap->fsid)[i]); mutex_unlock(&monc->mutex); - buf->f_fsid.val[0] = fsid & 0xffffffff; - buf->f_fsid.val[1] = fsid >> 32; + /* fold the fs_cluster_id into the upper bits */ + buf->f_fsid.val[1] = monc->fs_cluster_id; return 0; } @@ -143,6 +150,7 @@ enum { Opt_mds_namespace, Opt_recover_session, Opt_source, + Opt_mon_addr, /* string args above */ Opt_dirstat, Opt_rbytes, @@ -155,6 +163,8 @@ enum { Opt_acl, Opt_quotadf, Opt_copyfrom, + Opt_wsync, + Opt_pagecache, }; enum ceph_recover_session_mode { @@ -193,7 +203,10 @@ static const struct fs_parameter_spec ceph_mount_parameters[] = { fsparam_u32 ("rsize", Opt_rsize), fsparam_string ("snapdirname", Opt_snapdirname), fsparam_string ("source", Opt_source), + fsparam_string ("mon_addr", Opt_mon_addr), fsparam_u32 ("wsize", Opt_wsize), + fsparam_flag_no ("wsync", Opt_wsync), + fsparam_flag_no ("pagecache", Opt_pagecache), {} }; @@ -223,9 +236,92 @@ static void canonicalize_path(char *path) } /* - * Parse the source parameter. Distinguish the server list from the path. + * Check if the mds namespace in ceph_mount_options matches + * the passed in namespace string. First time match (when + * ->mds_namespace is NULL) is treated specially, since + * ->mds_namespace needs to be initialized by the caller. + */ +static int namespace_equals(struct ceph_mount_options *fsopt, + const char *namespace, size_t len) +{ + return !(fsopt->mds_namespace && + (strlen(fsopt->mds_namespace) != len || + strncmp(fsopt->mds_namespace, namespace, len))); +} + +static int ceph_parse_old_source(const char *dev_name, const char *dev_name_end, + struct fs_context *fc) +{ + int r; + struct ceph_parse_opts_ctx *pctx = fc->fs_private; + struct ceph_mount_options *fsopt = pctx->opts; + + if (*dev_name_end != ':') + return invalfc(fc, "separator ':' missing in source"); + + r = ceph_parse_mon_ips(dev_name, dev_name_end - dev_name, + pctx->copts, fc->log.log, ','); + if (r) + return r; + + fsopt->new_dev_syntax = false; + return 0; +} + +static int ceph_parse_new_source(const char *dev_name, const char *dev_name_end, + struct fs_context *fc) +{ + size_t len; + struct ceph_fsid fsid; + struct ceph_parse_opts_ctx *pctx = fc->fs_private; + struct ceph_mount_options *fsopt = pctx->opts; + char *fsid_start, *fs_name_start; + + if (*dev_name_end != '=') { + dout("separator '=' missing in source"); + return -EINVAL; + } + + fsid_start = strchr(dev_name, '@'); + if (!fsid_start) + return invalfc(fc, "missing cluster fsid"); + ++fsid_start; /* start of cluster fsid */ + + fs_name_start = strchr(fsid_start, '.'); + if (!fs_name_start) + return invalfc(fc, "missing file system name"); + + if (ceph_parse_fsid(fsid_start, &fsid)) + return invalfc(fc, "Invalid FSID"); + + ++fs_name_start; /* start of file system name */ + len = dev_name_end - fs_name_start; + + if (!namespace_equals(fsopt, fs_name_start, len)) + return invalfc(fc, "Mismatching mds_namespace"); + kfree(fsopt->mds_namespace); + fsopt->mds_namespace = kstrndup(fs_name_start, len, GFP_KERNEL); + if (!fsopt->mds_namespace) + return -ENOMEM; + dout("file system (mds namespace) '%s'\n", fsopt->mds_namespace); + + fsopt->new_dev_syntax = true; + return 0; +} + +/* + * Parse the source parameter for new device format. Distinguish the device + * spec from the path. Try parsing new device format and fallback to old + * format if needed. * - * The source will look like: + * New device syntax will looks like: + * <device_spec>=/<path> + * where + * <device_spec> is name@fsid.fsname + * <path> is optional, but if present must begin with '/' + * (monitor addresses are passed via mount option) + * + * Old device syntax is: * <server_spec>[,<server_spec>...]:[<path>] * where * <server_spec> is <ip>[:<port>] @@ -258,24 +354,44 @@ static int ceph_parse_source(struct fs_parameter *param, struct fs_context *fc) dev_name_end = dev_name + strlen(dev_name); } - dev_name_end--; /* back up to ':' separator */ - if (dev_name_end < dev_name || *dev_name_end != ':') - return invalfc(fc, "No path or : separator in source"); + dev_name_end--; /* back up to separator */ + if (dev_name_end < dev_name) + return invalfc(fc, "Path missing in source"); dout("device name '%.*s'\n", (int)(dev_name_end - dev_name), dev_name); if (fsopt->server_path) dout("server path '%s'\n", fsopt->server_path); - ret = ceph_parse_mon_ips(param->string, dev_name_end - dev_name, - pctx->copts, fc->log.log); - if (ret) - return ret; + dout("trying new device syntax"); + ret = ceph_parse_new_source(dev_name, dev_name_end, fc); + if (ret) { + if (ret != -EINVAL) + return ret; + dout("trying old device syntax"); + ret = ceph_parse_old_source(dev_name, dev_name_end, fc); + if (ret) + return ret; + } fc->source = param->string; param->string = NULL; return 0; } +static int ceph_parse_mon_addr(struct fs_parameter *param, + struct fs_context *fc) +{ + struct ceph_parse_opts_ctx *pctx = fc->fs_private; + struct ceph_mount_options *fsopt = pctx->opts; + + kfree(fsopt->mon_addr); + fsopt->mon_addr = param->string; + param->string = NULL; + + return ceph_parse_mon_ips(fsopt->mon_addr, strlen(fsopt->mon_addr), + pctx->copts, fc->log.log, '/'); +} + static int ceph_parse_mount_param(struct fs_context *fc, struct fs_parameter *param) { @@ -301,6 +417,8 @@ static int ceph_parse_mount_param(struct fs_context *fc, param->string = NULL; break; case Opt_mds_namespace: + if (!namespace_equals(fsopt, param->string, strlen(param->string))) + return invalfc(fc, "Mismatching mds_namespace"); kfree(fsopt->mds_namespace); fsopt->mds_namespace = param->string; param->string = NULL; @@ -318,6 +436,8 @@ static int ceph_parse_mount_param(struct fs_context *fc, if (fc->source) return invalfc(fc, "Multiple sources specified"); return ceph_parse_source(param, fc); + case Opt_mon_addr: + return ceph_parse_mon_addr(param, fc); case Opt_wsize: if (result.uint_32 < PAGE_SIZE || result.uint_32 > CEPH_MAX_WRITE_SIZE) @@ -444,6 +564,18 @@ static int ceph_parse_mount_param(struct fs_context *fc, fc->sb_flags &= ~SB_POSIXACL; } break; + case Opt_wsync: + if (!result.negated) + fsopt->flags &= ~CEPH_MOUNT_OPT_ASYNC_DIROPS; + else + fsopt->flags |= CEPH_MOUNT_OPT_ASYNC_DIROPS; + break; + case Opt_pagecache: + if (result.negated) + fsopt->flags |= CEPH_MOUNT_OPT_NOPAGECACHE; + else + fsopt->flags &= ~CEPH_MOUNT_OPT_NOPAGECACHE; + break; default: BUG(); } @@ -463,6 +595,7 @@ static void destroy_mount_options(struct ceph_mount_options *args) kfree(args->mds_namespace); kfree(args->server_path); kfree(args->fscache_uniq); + kfree(args->mon_addr); kfree(args); } @@ -506,6 +639,10 @@ static int compare_mount_options(struct ceph_mount_options *new_fsopt, if (ret) return ret; + ret = strcmp_null(fsopt1->mon_addr, fsopt2->mon_addr); + if (ret) + return ret; + return ceph_compare_options(new_opt, fsc->client); } @@ -561,12 +698,22 @@ static int ceph_show_options(struct seq_file *m, struct dentry *root) if ((fsopt->flags & CEPH_MOUNT_OPT_NOCOPYFROM) == 0) seq_puts(m, ",copyfrom"); - if (fsopt->mds_namespace) + /* dump mds_namespace when old device syntax is in use */ + if (fsopt->mds_namespace && !fsopt->new_dev_syntax) seq_show_option(m, "mds_namespace", fsopt->mds_namespace); + if (fsopt->mon_addr) + seq_printf(m, ",mon_addr=%s", fsopt->mon_addr); + if (fsopt->flags & CEPH_MOUNT_OPT_CLEANRECOVER) seq_show_option(m, "recover_session", "clean"); + if (!(fsopt->flags & CEPH_MOUNT_OPT_ASYNC_DIROPS)) + seq_puts(m, ",wsync"); + + if (fsopt->flags & CEPH_MOUNT_OPT_NOPAGECACHE) + seq_puts(m, ",nopagecache"); + if (fsopt->wsize != CEPH_MAX_WRITE_SIZE) seq_printf(m, ",wsize=%u", fsopt->wsize); if (fsopt->rsize != CEPH_MAX_READ_SIZE) @@ -623,8 +770,6 @@ static struct ceph_fs_client *create_fs_client(struct ceph_mount_options *fsopt, struct ceph_options *opt) { struct ceph_fs_client *fsc; - int page_count; - size_t size; int err; fsc = kzalloc(sizeof(*fsc), GFP_KERNEL); @@ -659,6 +804,7 @@ static struct ceph_fs_client *create_fs_client(struct ceph_mount_options *fsopt, fsc->have_copy_from2 = true; atomic_long_set(&fsc->writeback_count, 0); + fsc->write_congested = false; err = -ENOMEM; /* @@ -672,18 +818,15 @@ static struct ceph_fs_client *create_fs_client(struct ceph_mount_options *fsopt, if (!fsc->cap_wq) goto fail_inode_wq; - /* set up mempools */ - err = -ENOMEM; - page_count = fsc->mount_options->wsize >> PAGE_SHIFT; - size = sizeof (struct page *) * (page_count ? page_count : 1); - fsc->wb_pagevec_pool = mempool_create_kmalloc_pool(10, size); - if (!fsc->wb_pagevec_pool) - goto fail_cap_wq; + hash_init(fsc->async_unlink_conflict); + spin_lock_init(&fsc->async_unlink_conflict_lock); + + spin_lock(&ceph_fsc_lock); + list_add_tail(&fsc->metric_wakeup, &ceph_fsc_list); + spin_unlock(&ceph_fsc_lock); return fsc; -fail_cap_wq: - destroy_workqueue(fsc->cap_wq); fail_inode_wq: destroy_workqueue(fsc->inode_wq); fail_client: @@ -706,12 +849,14 @@ static void destroy_fs_client(struct ceph_fs_client *fsc) { dout("destroy_fs_client %p\n", fsc); + spin_lock(&ceph_fsc_lock); + list_del(&fsc->metric_wakeup); + spin_unlock(&ceph_fsc_lock); + ceph_mdsc_destroy(fsc); destroy_workqueue(fsc->inode_wq); destroy_workqueue(fsc->cap_wq); - mempool_destroy(fsc->wb_pagevec_pool); - destroy_mount_options(fsc->mount_options); ceph_destroy_client(fsc->client); @@ -725,15 +870,18 @@ static void destroy_fs_client(struct ceph_fs_client *fsc) */ struct kmem_cache *ceph_inode_cachep; struct kmem_cache *ceph_cap_cachep; +struct kmem_cache *ceph_cap_snap_cachep; struct kmem_cache *ceph_cap_flush_cachep; struct kmem_cache *ceph_dentry_cachep; struct kmem_cache *ceph_file_cachep; struct kmem_cache *ceph_dir_file_cachep; +struct kmem_cache *ceph_mds_request_cachep; +mempool_t *ceph_wb_pagevec_pool; static void ceph_inode_init_once(void *foo) { struct ceph_inode_info *ci = foo; - inode_init_once(&ci->vfs_inode); + inode_init_once(&ci->netfs.inode); } static int __init init_caches(void) @@ -751,6 +899,9 @@ static int __init init_caches(void) ceph_cap_cachep = KMEM_CACHE(ceph_cap, SLAB_MEM_SPREAD); if (!ceph_cap_cachep) goto bad_cap; + ceph_cap_snap_cachep = KMEM_CACHE(ceph_cap_snap, SLAB_MEM_SPREAD); + if (!ceph_cap_snap_cachep) + goto bad_cap_snap; ceph_cap_flush_cachep = KMEM_CACHE(ceph_cap_flush, SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD); if (!ceph_cap_flush_cachep) @@ -769,13 +920,19 @@ static int __init init_caches(void) if (!ceph_dir_file_cachep) goto bad_dir_file; - error = ceph_fscache_register(); - if (error) - goto bad_fscache; + ceph_mds_request_cachep = KMEM_CACHE(ceph_mds_request, SLAB_MEM_SPREAD); + if (!ceph_mds_request_cachep) + goto bad_mds_req; + + ceph_wb_pagevec_pool = mempool_create_kmalloc_pool(10, CEPH_MAX_WRITE_SIZE >> PAGE_SHIFT); + if (!ceph_wb_pagevec_pool) + goto bad_pagevec_pool; return 0; -bad_fscache: +bad_pagevec_pool: + kmem_cache_destroy(ceph_mds_request_cachep); +bad_mds_req: kmem_cache_destroy(ceph_dir_file_cachep); bad_dir_file: kmem_cache_destroy(ceph_file_cachep); @@ -784,6 +941,8 @@ bad_file: bad_dentry: kmem_cache_destroy(ceph_cap_flush_cachep); bad_cap_flush: + kmem_cache_destroy(ceph_cap_snap_cachep); +bad_cap_snap: kmem_cache_destroy(ceph_cap_cachep); bad_cap: kmem_cache_destroy(ceph_inode_cachep); @@ -800,19 +959,27 @@ static void destroy_caches(void) kmem_cache_destroy(ceph_inode_cachep); kmem_cache_destroy(ceph_cap_cachep); + kmem_cache_destroy(ceph_cap_snap_cachep); kmem_cache_destroy(ceph_cap_flush_cachep); kmem_cache_destroy(ceph_dentry_cachep); kmem_cache_destroy(ceph_file_cachep); kmem_cache_destroy(ceph_dir_file_cachep); + kmem_cache_destroy(ceph_mds_request_cachep); + mempool_destroy(ceph_wb_pagevec_pool); +} - ceph_fscache_unregister(); +static void __ceph_umount_begin(struct ceph_fs_client *fsc) +{ + ceph_osdc_abort_requests(&fsc->client->osdc, -EIO); + ceph_mdsc_force_umount(fsc->mdsc); + fsc->filp_gen++; // invalidate open files } /* - * ceph_umount_begin - initiate forced umount. Tear down down the + * ceph_umount_begin - initiate forced umount. Tear down the * mount, skipping steps that may hang while waiting for server(s). */ -static void ceph_umount_begin(struct super_block *sb) +void ceph_umount_begin(struct super_block *sb) { struct ceph_fs_client *fsc = ceph_sb_to_client(sb); @@ -820,9 +987,7 @@ static void ceph_umount_begin(struct super_block *sb) if (!fsc) return; fsc->mount_state = CEPH_MOUNT_SHUTDOWN; - ceph_osdc_abort_requests(&fsc->client->osdc, -EIO); - ceph_mdsc_force_umount(fsc->mdsc); - fsc->filp_gen++; // invalidate open files + __ceph_umount_begin(fsc); } static const struct super_operations ceph_super_ops = { @@ -959,6 +1124,7 @@ static int ceph_set_super(struct super_block *s, struct fs_context *fc) s->s_time_gran = 1; s->s_time_min = 0; s->s_time_max = U32_MAX; + s->s_flags |= SB_NODIRATIME | SB_NOATIME; ret = set_anon_super_fc(s, fc); if (ret != 0) @@ -974,16 +1140,16 @@ static int ceph_compare_super(struct super_block *sb, struct fs_context *fc) struct ceph_fs_client *new = fc->s_fs_info; struct ceph_mount_options *fsopt = new->mount_options; struct ceph_options *opt = new->client->options; - struct ceph_fs_client *other = ceph_sb_to_client(sb); + struct ceph_fs_client *fsc = ceph_sb_to_client(sb); dout("ceph_compare_super %p\n", sb); - if (compare_mount_options(fsopt, opt, other)) { + if (compare_mount_options(fsopt, opt, fsc)) { dout("monitor(s)/mount options don't match\n"); return 0; } if ((opt->flags & CEPH_OPT_FSID) && - ceph_fsid_compare(&opt->fsid, &other->client->fsid)) { + ceph_fsid_compare(&opt->fsid, &fsc->client->fsid)) { dout("fsid doesn't match\n"); return 0; } @@ -991,6 +1157,17 @@ static int ceph_compare_super(struct super_block *sb, struct fs_context *fc) dout("flags differ\n"); return 0; } + + if (fsc->blocklisted && !ceph_test_mount_opt(fsc, CLEANRECOVER)) { + dout("client is blocklisted (and CLEANRECOVER is not set)\n"); + return 0; + } + + if (fsc->mount_state == CEPH_MOUNT_SHUTDOWN) { + dout("client has been forcibly unmounted\n"); + return 0; + } + return 1; } @@ -1020,6 +1197,7 @@ static int ceph_setup_bdi(struct super_block *sb, struct ceph_fs_client *fsc) static int ceph_get_tree(struct fs_context *fc) { struct ceph_parse_opts_ctx *pctx = fc->fs_private; + struct ceph_mount_options *fsopt = pctx->opts; struct super_block *sb; struct ceph_fs_client *fsc; struct dentry *res; @@ -1031,6 +1209,8 @@ static int ceph_get_tree(struct fs_context *fc) if (!fc->source) return invalfc(fc, "No source"); + if (fsopt->new_dev_syntax && !fsopt->mon_addr) + return invalfc(fc, "No monitor address"); /* create client (which we may/may not use) */ fsc = create_fs_client(pctx->opts, pctx->copts); @@ -1107,6 +1287,22 @@ static void ceph_free_fc(struct fs_context *fc) static int ceph_reconfigure_fc(struct fs_context *fc) { + struct ceph_parse_opts_ctx *pctx = fc->fs_private; + struct ceph_mount_options *fsopt = pctx->opts; + struct ceph_fs_client *fsc = ceph_sb_to_client(fc->root->d_sb); + + if (fsopt->flags & CEPH_MOUNT_OPT_ASYNC_DIROPS) + ceph_set_mount_opt(fsc, ASYNC_DIROPS); + else + ceph_clear_mount_opt(fsc, ASYNC_DIROPS); + + if (strcmp_null(fsc->mount_options->mon_addr, fsopt->mon_addr)) { + kfree(fsc->mount_options->mon_addr); + fsc->mount_options->mon_addr = fsopt->mon_addr; + fsopt->mon_addr = NULL; + pr_notice("ceph: monitor addresses recorded, but not used for reconnection"); + } + sync_filesystem(fc->root->d_sb); return 0; } @@ -1172,14 +1368,13 @@ nomem: static void ceph_kill_sb(struct super_block *s) { struct ceph_fs_client *fsc = ceph_sb_to_client(s); - dev_t dev = s->s_dev; dout("kill_sb %p\n", s); ceph_mdsc_pre_umount(fsc->mdsc); flush_fs_workqueues(fsc); - generic_shutdown_super(s); + kill_anon_super(s); fsc->client->extra_mon_dispatch = NULL; ceph_fs_debugfs_cleanup(fsc); @@ -1187,7 +1382,6 @@ static void ceph_kill_sb(struct super_block *s) ceph_fscache_unregister_fs(fsc); destroy_fs_client(fsc); - free_anon_bdev(dev); } static struct file_system_type ceph_fs_type = { @@ -1204,19 +1398,20 @@ int ceph_force_reconnect(struct super_block *sb) struct ceph_fs_client *fsc = ceph_sb_to_client(sb); int err = 0; - ceph_umount_begin(sb); + fsc->mount_state = CEPH_MOUNT_RECOVER; + __ceph_umount_begin(fsc); /* Make sure all page caches get invalidated. * see remove_session_caps_cb() */ flush_workqueue(fsc->inode_wq); - /* In case that we were blacklisted. This also reset + /* In case that we were blocklisted. This also reset * all mon/osd connections */ ceph_reset_client_addr(fsc->client); ceph_osdc_clear_abort_err(&fsc->client->osdc); - fsc->blacklisted = false; + fsc->blocklisted = false; fsc->mount_state = CEPH_MOUNT_MOUNTED; if (sb->s_root) { @@ -1254,6 +1449,45 @@ static void __exit exit_ceph(void) destroy_caches(); } +static int param_set_metrics(const char *val, const struct kernel_param *kp) +{ + struct ceph_fs_client *fsc; + int ret; + + ret = param_set_bool(val, kp); + if (ret) { + pr_err("Failed to parse sending metrics switch value '%s'\n", + val); + return ret; + } else if (!disable_send_metrics) { + // wake up all the mds clients + spin_lock(&ceph_fsc_lock); + list_for_each_entry(fsc, &ceph_fsc_list, metric_wakeup) { + metric_schedule_delayed(&fsc->mdsc->metric); + } + spin_unlock(&ceph_fsc_lock); + } + + return 0; +} + +static const struct kernel_param_ops param_ops_metrics = { + .set = param_set_metrics, + .get = param_get_bool, +}; + +bool disable_send_metrics = false; +module_param_cb(disable_send_metrics, ¶m_ops_metrics, &disable_send_metrics, 0644); +MODULE_PARM_DESC(disable_send_metrics, "Enable sending perf metrics to ceph cluster (default: on)"); + +/* for both v1 and v2 syntax */ +static bool mount_support = true; +static const struct kernel_param_ops param_ops_mount_syntax = { + .get = param_get_bool, +}; +module_param_cb(mount_syntax_v1, ¶m_ops_mount_syntax, &mount_support, 0444); +module_param_cb(mount_syntax_v2, ¶m_ops_mount_syntax, &mount_support, 0444); + module_init(init_ceph); module_exit(exit_ceph); diff --git a/fs/ceph/super.h b/fs/ceph/super.h index 037cdfb2ad4f..40630e6f691c 100644 --- a/fs/ceph/super.h +++ b/fs/ceph/super.h @@ -17,22 +17,19 @@ #include <linux/posix_acl.h> #include <linux/refcount.h> #include <linux/security.h> - -#include <linux/ceph/libceph.h> - -#ifdef CONFIG_CEPH_FSCACHE +#include <linux/netfs.h> #include <linux/fscache.h> -#endif +#include <linux/hashtable.h> -/* f_type in struct statfs */ -#define CEPH_SUPER_MAGIC 0x00c36400 +#include <linux/ceph/libceph.h> /* large granularity for statfs utilization stats to facilitate * large volume sizes on 32-bit machines. */ #define CEPH_BLOCK_SHIFT 22 /* 4 MB */ #define CEPH_BLOCK (1 << CEPH_BLOCK_SHIFT) +#define CEPH_4K_BLOCK_SHIFT 12 /* 4 KB */ -#define CEPH_MOUNT_OPT_CLEANRECOVER (1<<1) /* auto reonnect (clean mode) after blacklisted */ +#define CEPH_MOUNT_OPT_CLEANRECOVER (1<<1) /* auto reonnect (clean mode) after blocklisted */ #define CEPH_MOUNT_OPT_DIRSTAT (1<<4) /* `cat dirname` for stats */ #define CEPH_MOUNT_OPT_RBYTES (1<<5) /* dir st_bytes = rbytes */ #define CEPH_MOUNT_OPT_NOASYNCREADDIR (1<<7) /* no dcache readdir */ @@ -43,13 +40,18 @@ #define CEPH_MOUNT_OPT_MOUNTWAIT (1<<12) /* mount waits if no mds is up */ #define CEPH_MOUNT_OPT_NOQUOTADF (1<<13) /* no root dir quota in statfs */ #define CEPH_MOUNT_OPT_NOCOPYFROM (1<<14) /* don't use RADOS 'copy-from' op */ +#define CEPH_MOUNT_OPT_ASYNC_DIROPS (1<<15) /* allow async directory ops */ +#define CEPH_MOUNT_OPT_NOPAGECACHE (1<<16) /* bypass pagecache altogether */ #define CEPH_MOUNT_OPT_DEFAULT \ (CEPH_MOUNT_OPT_DCACHE | \ - CEPH_MOUNT_OPT_NOCOPYFROM) + CEPH_MOUNT_OPT_NOCOPYFROM | \ + CEPH_MOUNT_OPT_ASYNC_DIROPS) #define ceph_set_mount_opt(fsc, opt) \ - (fsc)->mount_options->flags |= CEPH_MOUNT_OPT_##opt; + (fsc)->mount_options->flags |= CEPH_MOUNT_OPT_##opt +#define ceph_clear_mount_opt(fsc, opt) \ + (fsc)->mount_options->flags &= ~CEPH_MOUNT_OPT_##opt #define ceph_test_mount_opt(fsc, opt) \ (!!((fsc)->mount_options->flags & CEPH_MOUNT_OPT_##opt)) @@ -84,6 +86,8 @@ struct ceph_mount_options { unsigned int max_readdir; /* max readdir result (entries) */ unsigned int max_readdir_bytes; /* max readdir result (bytes) */ + bool new_dev_syntax; + /* * everything above this point can be memcmp'd; everything below * is handled in compare_mount_options() @@ -93,18 +97,22 @@ struct ceph_mount_options { char *mds_namespace; /* default NULL */ char *server_path; /* default NULL (means "/") */ char *fscache_uniq; /* default NULL */ + char *mon_addr; }; +#define CEPH_ASYNC_CREATE_CONFLICT_BITS 8 + struct ceph_fs_client { struct super_block *sb; + struct list_head metric_wakeup; + struct ceph_mount_options *mount_options; struct ceph_client *client; - unsigned long mount_state; + int mount_state; - unsigned long last_auto_reconnect; - bool blacklisted; + bool blocklisted; bool have_copy_from2; @@ -113,23 +121,27 @@ struct ceph_fs_client { struct ceph_mds_client *mdsc; - /* writeback */ - mempool_t *wb_pagevec_pool; atomic_long_t writeback_count; + bool write_congested; struct workqueue_struct *inode_wq; struct workqueue_struct *cap_wq; + DECLARE_HASHTABLE(async_unlink_conflict, CEPH_ASYNC_CREATE_CONFLICT_BITS); + spinlock_t async_unlink_conflict_lock; + #ifdef CONFIG_DEBUG_FS struct dentry *debugfs_dentry_lru, *debugfs_caps; struct dentry *debugfs_congestion_kb; struct dentry *debugfs_bdi; struct dentry *debugfs_mdsc, *debugfs_mdsmap; + struct dentry *debugfs_status; struct dentry *debugfs_mds_sessions; + struct dentry *debugfs_metrics_dir; #endif #ifdef CONFIG_CEPH_FSCACHE - struct fscache_cookie *fscache; + struct fscache_volume *fscache; #endif }; @@ -156,7 +168,8 @@ struct ceph_cap { int issued; /* latest, from the mds */ int implemented; /* implemented superset of issued (for revocation) */ - int mds, mds_wanted; + int mds; /* mds index for this cap */ + int mds_wanted; /* caps wanted from this mds */ }; /* caps to release */ struct { @@ -170,14 +183,15 @@ struct ceph_cap { struct list_head caps_item; }; -#define CHECK_CAPS_NODELAY 1 /* do not delay any further */ -#define CHECK_CAPS_AUTHONLY 2 /* only check auth cap */ -#define CHECK_CAPS_FLUSH 4 /* flush any dirty caps */ +#define CHECK_CAPS_AUTHONLY 1 /* only check auth cap */ +#define CHECK_CAPS_FLUSH 2 /* flush any dirty caps */ +#define CHECK_CAPS_NOINVAL 4 /* don't invalidate pagecache */ struct ceph_cap_flush { u64 tid; - int caps; /* 0 means capsnap */ + int caps; bool wake; /* wake up flush waiters when finish ? */ + bool is_capsnap; /* true means capsnap */ struct list_head g_list; // global struct list_head i_list; // per inode }; @@ -221,7 +235,7 @@ static inline void ceph_put_cap_snap(struct ceph_cap_snap *capsnap) if (refcount_dec_and_test(&capsnap->nref)) { if (capsnap->xattr_blob) ceph_buffer_put(capsnap->xattr_blob); - kfree(capsnap); + kmem_cache_free(ceph_cap_snap_cachep, capsnap); } } @@ -272,7 +286,8 @@ struct ceph_dentry_info { struct dentry *dentry; struct ceph_mds_session *lease_session; struct list_head lease_list; - unsigned flags; + struct hlist_node hnode; + unsigned long flags; int lease_shared_gen; u32 lease_gen; u32 lease_seq; @@ -281,9 +296,14 @@ struct ceph_dentry_info { u64 offset; }; -#define CEPH_DENTRY_REFERENCED 1 -#define CEPH_DENTRY_LEASE_LIST 2 -#define CEPH_DENTRY_SHRINK_LIST 4 +#define CEPH_DENTRY_REFERENCED (1 << 0) +#define CEPH_DENTRY_LEASE_LIST (1 << 1) +#define CEPH_DENTRY_SHRINK_LIST (1 << 2) +#define CEPH_DENTRY_PRIMARY_LINK (1 << 3) +#define CEPH_DENTRY_ASYNC_UNLINK_BIT (4) +#define CEPH_DENTRY_ASYNC_UNLINK (1 << CEPH_DENTRY_ASYNC_UNLINK_BIT) +#define CEPH_DENTRY_ASYNC_CREATE_BIT (5) +#define CEPH_DENTRY_ASYNC_CREATE (1 << CEPH_DENTRY_ASYNC_CREATE_BIT) struct ceph_inode_xattrs_info { /* @@ -307,6 +327,7 @@ struct ceph_inode_xattrs_info { * Ceph inode. */ struct ceph_inode_info { + struct netfs_inode netfs; /* Netfslib context and vfs inode */ struct ceph_vino i_vino; /* ceph ino + snap */ spinlock_t i_ceph_lock; @@ -315,18 +336,19 @@ struct ceph_inode_info { u64 i_inline_version; u32 i_time_warp_seq; - unsigned i_ceph_flags; + unsigned long i_ceph_flags; atomic64_t i_release_count; atomic64_t i_ordered_count; atomic64_t i_complete_seq[2]; struct ceph_dir_layout i_dir_layout; struct ceph_file_layout i_layout; + struct ceph_file_layout i_cached_layout; // for async creates char *i_symlink; /* for dirs */ struct timespec64 i_rctime; - u64 i_rbytes, i_rfiles, i_rsubdirs; + u64 i_rbytes, i_rfiles, i_rsubdirs, i_rsnaps; u64 i_files, i_subdirs; /* quotas */ @@ -345,14 +367,31 @@ struct ceph_inode_info { struct rb_root i_caps; /* cap list */ struct ceph_cap *i_auth_cap; /* authoritative cap, if any */ unsigned i_dirty_caps, i_flushing_caps; /* mask of dirtied fields */ - struct list_head i_dirty_item, i_flushing_item; + + /* + * Link to the auth cap's session's s_cap_dirty list. s_cap_dirty + * is protected by the mdsc->cap_dirty_lock, but each individual item + * is also protected by the inode's i_ceph_lock. Walking s_cap_dirty + * requires the mdsc->cap_dirty_lock. List presence for an item can + * be tested under the i_ceph_lock. Changing anything requires both. + */ + struct list_head i_dirty_item; + + /* + * Link to session's s_cap_flushing list. Protected in a similar + * fashion to i_dirty_item, but also by the s_mutex for changes. The + * s_cap_flushing list can be walked while holding either the s_mutex + * or msdc->cap_dirty_lock. List presence can also be checked while + * holding the i_ceph_lock for this inode. + */ + struct list_head i_flushing_item; + /* we need to track cap writeback on a per-cap-bit basis, to allow * overlapping, pipelined cap flushes to the mds. we can probably * reduce the tid to 8 bits if we're concerned about inode size. */ struct ceph_cap_flush *i_prealloc_cap_flush; struct list_head i_cap_flush_list; wait_queue_head_t i_cap_wq; /* threads waiting on a capability */ - unsigned long i_hold_caps_min; /* jiffies */ unsigned long i_hold_caps_max; /* jiffies */ struct list_head i_cap_delay_list; /* for delayed cap release to mds */ struct ceph_cap_reservation i_cap_migration_resv; @@ -361,6 +400,8 @@ struct ceph_inode_info { dirty|flushing caps */ unsigned i_snap_caps; /* cap bits for snapped files */ + unsigned long i_last_rd; + unsigned long i_last_wr; int i_nr_by_mode[CEPH_FILE_MODE_BITS]; /* open file counts */ struct mutex i_truncate_mutex; @@ -375,7 +416,7 @@ struct ceph_inode_info { /* held references to caps */ int i_pin_ref; - int i_rd_ref, i_rdcache_ref, i_wr_ref, i_wb_ref; + int i_rd_ref, i_rdcache_ref, i_wr_ref, i_wb_ref, i_fx_ref; int i_wrbuffer_ref, i_wrbuffer_ref_head; atomic_t i_filelock_ref; atomic_t i_shared_gen; /* increment each time we get FILE_SHARED */ @@ -390,7 +431,6 @@ struct ceph_inode_info { struct ceph_snap_realm *i_snap_realm; /* snap realm (if caps) */ struct ceph_snapid_map *i_snapid_map; /* snapid -> dev_t */ }; - int i_snap_realm_counter; /* snap realm (if caps) */ struct list_head i_snap_realm_item; struct list_head i_snap_flush_item; struct timespec64 i_btime; @@ -398,20 +438,12 @@ struct ceph_inode_info { struct work_struct i_work; unsigned long i_work_mask; - -#ifdef CONFIG_CEPH_FSCACHE - struct fscache_cookie *fscache; - u32 i_fscache_gen; -#endif - errseq_t i_meta_err; - - struct inode vfs_inode; /* at end */ }; static inline struct ceph_inode_info * ceph_inode(const struct inode *inode) { - return container_of(inode, struct ceph_inode_info, vfs_inode); + return container_of(inode, struct ceph_inode_info, netfs.inode); } static inline struct ceph_fs_client * @@ -426,21 +458,19 @@ ceph_sb_to_client(const struct super_block *sb) return (struct ceph_fs_client *)sb->s_fs_info; } +static inline struct ceph_mds_client * +ceph_sb_to_mdsc(const struct super_block *sb) +{ + return (struct ceph_mds_client *)ceph_sb_to_client(sb)->mdsc; +} + static inline struct ceph_vino ceph_vino(const struct inode *inode) { return ceph_inode(inode)->i_vino; } -/* - * ino_t is <64 bits on many architectures, blech. - * - * i_ino (kernel inode) st_ino (userspace) - * i386 32 32 - * x86_64+ino32 64 32 - * x86_64 64 64 - */ -static inline u32 ceph_ino_to_ino32(__u64 vino) +static inline u32 ceph_ino_to_ino32(u64 vino) { u32 ino = vino & 0xffffffff; ino ^= vino >> 32; @@ -450,34 +480,17 @@ static inline u32 ceph_ino_to_ino32(__u64 vino) } /* - * kernel i_ino value + * Inode numbers in cephfs are 64 bits, but inode->i_ino is 32-bits on + * some arches. We generally do not use this value inside the ceph driver, but + * we do want to set it to something, so that generic vfs code has an + * appropriate value for tracepoints and the like. */ -static inline ino_t ceph_vino_to_ino(struct ceph_vino vino) +static inline ino_t ceph_vino_to_ino_t(struct ceph_vino vino) { -#if BITS_PER_LONG == 32 - return ceph_ino_to_ino32(vino.ino); -#else + if (sizeof(ino_t) == sizeof(u32)) + return ceph_ino_to_ino32(vino.ino); return (ino_t)vino.ino; -#endif -} - -/* - * user-visible ino (stat, filldir) - */ -#if BITS_PER_LONG == 32 -static inline ino_t ceph_translate_ino(struct super_block *sb, ino_t ino) -{ - return ino; } -#else -static inline ino_t ceph_translate_ino(struct super_block *sb, ino_t ino) -{ - if (ceph_test_mount_opt(ceph_sb_to_client(sb), INO32)) - ino = ceph_ino_to_ino32(ino); - return ino; -} -#endif - /* for printf-style formatting */ #define ceph_vinop(i) ceph_inode(i)->i_vino.ino, ceph_inode(i)->i_vino.snap @@ -486,11 +499,34 @@ static inline u64 ceph_ino(struct inode *inode) { return ceph_inode(inode)->i_vino.ino; } + static inline u64 ceph_snap(struct inode *inode) { return ceph_inode(inode)->i_vino.snap; } +/** + * ceph_present_ino - format an inode number for presentation to userland + * @sb: superblock where the inode lives + * @ino: inode number to (possibly) convert + * + * If the user mounted with the ino32 option, then the 64-bit value needs + * to be converted to something that can fit inside 32 bits. Note that + * internal kernel code never uses this value, so this is entirely for + * userland consumption. + */ +static inline u64 ceph_present_ino(struct super_block *sb, u64 ino) +{ + if (unlikely(ceph_test_mount_opt(ceph_sb_to_client(sb), INO32))) + return ceph_ino_to_ino32(ino); + return ino; +} + +static inline u64 ceph_present_inode(struct inode *inode) +{ + return ceph_present_ino(inode->i_sb, ceph_ino(inode)); +} + static inline int ceph_ino_compare(struct inode *inode, void *data) { struct ceph_vino *pvino = (struct ceph_vino *)data; @@ -499,11 +535,44 @@ static inline int ceph_ino_compare(struct inode *inode, void *data) ci->i_vino.snap == pvino->snap; } +/* + * The MDS reserves a set of inodes for its own usage. These should never + * be accessible by clients, and so the MDS has no reason to ever hand these + * out. The range is CEPH_MDS_INO_MDSDIR_OFFSET..CEPH_INO_SYSTEM_BASE. + * + * These come from src/mds/mdstypes.h in the ceph sources. + */ +#define CEPH_MAX_MDS 0x100 +#define CEPH_NUM_STRAY 10 +#define CEPH_MDS_INO_MDSDIR_OFFSET (1 * CEPH_MAX_MDS) +#define CEPH_MDS_INO_LOG_OFFSET (2 * CEPH_MAX_MDS) +#define CEPH_INO_SYSTEM_BASE ((6*CEPH_MAX_MDS) + (CEPH_MAX_MDS * CEPH_NUM_STRAY)) + +static inline bool ceph_vino_is_reserved(const struct ceph_vino vino) +{ + if (vino.ino >= CEPH_INO_SYSTEM_BASE || + vino.ino < CEPH_MDS_INO_MDSDIR_OFFSET) + return false; + + /* Don't warn on mdsdirs */ + WARN_RATELIMIT(vino.ino >= CEPH_MDS_INO_LOG_OFFSET, + "Attempt to access reserved inode number 0x%llx", + vino.ino); + return true; +} + static inline struct inode *ceph_find_inode(struct super_block *sb, struct ceph_vino vino) { - ino_t t = ceph_vino_to_ino(vino); - return ilookup5(sb, t, ceph_ino_compare, &vino); + if (ceph_vino_is_reserved(vino)) + return NULL; + + /* + * NB: The hashval will be run through the fs/inode.c hash function + * anyway, so there is no need to squash the inode number down to + * 32-bits first. Just use low-order bits on arches with 32-bit long. + */ + return ilookup5(sb, (unsigned long)vino.ino, ceph_ino_compare, &vino); } @@ -511,25 +580,28 @@ static inline struct inode *ceph_find_inode(struct super_block *sb, * Ceph inode. */ #define CEPH_I_DIR_ORDERED (1 << 0) /* dentries in dir are ordered */ -#define CEPH_I_NODELAY (1 << 1) /* do not delay cap release */ #define CEPH_I_FLUSH (1 << 2) /* do not delay flush of dirty metadata */ #define CEPH_I_POOL_PERM (1 << 3) /* pool rd/wr bits are valid */ #define CEPH_I_POOL_RD (1 << 4) /* can read from pool */ #define CEPH_I_POOL_WR (1 << 5) /* can write to pool */ #define CEPH_I_SEC_INITED (1 << 6) /* security initialized */ -#define CEPH_I_CAP_DROPPED (1 << 7) /* caps were forcibly dropped */ -#define CEPH_I_KICK_FLUSH (1 << 8) /* kick flushing caps */ -#define CEPH_I_FLUSH_SNAPS (1 << 9) /* need flush snapss */ -#define CEPH_I_ERROR_WRITE (1 << 10) /* have seen write errors */ -#define CEPH_I_ERROR_FILELOCK (1 << 11) /* have seen file lock errors */ -#define CEPH_I_ODIRECT (1 << 12) /* inode in direct I/O mode */ +#define CEPH_I_KICK_FLUSH (1 << 7) /* kick flushing caps */ +#define CEPH_I_FLUSH_SNAPS (1 << 8) /* need flush snapss */ +#define CEPH_I_ERROR_WRITE (1 << 9) /* have seen write errors */ +#define CEPH_I_ERROR_FILELOCK (1 << 10) /* have seen file lock errors */ +#define CEPH_I_ODIRECT (1 << 11) /* inode in direct I/O mode */ +#define CEPH_ASYNC_CREATE_BIT (12) /* async create in flight for this */ +#define CEPH_I_ASYNC_CREATE (1 << CEPH_ASYNC_CREATE_BIT) +#define CEPH_I_SHUTDOWN (1 << 13) /* inode is no longer usable */ /* * Masks of ceph inode work. */ -#define CEPH_I_WORK_WRITEBACK 0 /* writeback */ -#define CEPH_I_WORK_INVALIDATE_PAGES 1 /* invalidate pages */ -#define CEPH_I_WORK_VMTRUNCATE 2 /* vmtruncate */ +#define CEPH_I_WORK_WRITEBACK 0 +#define CEPH_I_WORK_INVALIDATE_PAGES 1 +#define CEPH_I_WORK_VMTRUNCATE 2 +#define CEPH_I_WORK_CHECK_CAPS 3 +#define CEPH_I_WORK_FLUSH_SNAPS 4 /* * We set the ERROR_WRITE bit when we start seeing write errors on an inode @@ -638,6 +710,8 @@ static inline bool __ceph_is_any_real_caps(struct ceph_inode_info *ci) extern int __ceph_caps_issued(struct ceph_inode_info *ci, int *implemented); extern int __ceph_caps_issued_mask(struct ceph_inode_info *ci, int mask, int t); +extern int __ceph_caps_issued_mask_metric(struct ceph_inode_info *ci, int mask, + int t); extern int __ceph_caps_issued_other(struct ceph_inode_info *ci, struct ceph_cap *cap); @@ -650,12 +724,12 @@ static inline int ceph_caps_issued(struct ceph_inode_info *ci) return issued; } -static inline int ceph_caps_issued_mask(struct ceph_inode_info *ci, int mask, - int touch) +static inline int ceph_caps_issued_mask_metric(struct ceph_inode_info *ci, + int mask, int touch) { int r; spin_lock(&ci->i_ceph_lock); - r = __ceph_caps_issued_mask(ci, mask, touch); + r = __ceph_caps_issued_mask_metric(ci, mask, touch); spin_unlock(&ci->i_ceph_lock); return r; } @@ -674,18 +748,12 @@ extern int __ceph_caps_revoking_other(struct ceph_inode_info *ci, extern int ceph_caps_revoking(struct ceph_inode_info *ci, int mask); extern int __ceph_caps_used(struct ceph_inode_info *ci); -extern int __ceph_caps_file_wanted(struct ceph_inode_info *ci); - -/* - * wanted, by virtue of open file modes AND cap refs (buffered/cached data) - */ -static inline int __ceph_caps_wanted(struct ceph_inode_info *ci) +static inline bool __ceph_is_file_opened(struct ceph_inode_info *ci) { - int w = __ceph_caps_file_wanted(ci) | __ceph_caps_used(ci); - if (w & CEPH_CAP_FILE_BUFFER) - w |= CEPH_CAP_FILE_EXCL; /* we want EXCL if dirty data */ - return w; + return ci->i_nr_by_mode[0]; } +extern int __ceph_caps_file_wanted(struct ceph_inode_info *ci); +extern int __ceph_caps_wanted(struct ceph_inode_info *ci); /* what the mds thinks we want */ extern int __ceph_caps_mds_wanted(struct ceph_inode_info *ci, bool check); @@ -701,6 +769,8 @@ extern void ceph_unreserve_caps(struct ceph_mds_client *mdsc, extern void ceph_reservation_status(struct ceph_fs_client *client, int *total, int *avail, int *used, int *reserved, int *min); +extern void change_auth_cap_ses(struct ceph_inode_info *ci, + struct ceph_mds_session *session); @@ -717,7 +787,6 @@ struct ceph_file_info { spinlock_t rw_contexts_lock; struct list_head rw_contexts; - errseq_t meta_err; u32 filp_gen; atomic_t num_locks; }; @@ -822,6 +891,8 @@ struct ceph_snap_realm { struct list_head dirty_item; /* if realm needs new context */ + struct list_head rebuild_item; /* rebuild snap realms _downward_ in hierarchy */ + /* the current set of snaps for this realm */ struct ceph_snap_context *cached_context; @@ -871,13 +942,13 @@ extern void ceph_put_snap_realm(struct ceph_mds_client *mdsc, extern int ceph_update_snap_trace(struct ceph_mds_client *m, void *p, void *e, bool deletion, struct ceph_snap_realm **realm_ret); +void ceph_change_snap_realm(struct inode *inode, struct ceph_snap_realm *realm); extern void ceph_handle_snap(struct ceph_mds_client *mdsc, struct ceph_mds_session *session, struct ceph_msg *msg); -extern void ceph_queue_cap_snap(struct ceph_inode_info *ci); extern int __ceph_finish_cap_snap(struct ceph_inode_info *ci, struct ceph_cap_snap *capsnap); -extern void ceph_cleanup_empty_realms(struct ceph_mds_client *mdsc); +extern void ceph_cleanup_global_and_empty_realms(struct ceph_mds_client *mdsc); extern struct ceph_snapid_map *ceph_get_snapid_map(struct ceph_mds_client *mdsc, u64 snap); @@ -885,6 +956,7 @@ extern void ceph_put_snapid_map(struct ceph_mds_client* mdsc, struct ceph_snapid_map *sm); extern void ceph_trim_snapid_map(struct ceph_mds_client *mdsc); extern void ceph_cleanup_snapid_map(struct ceph_mds_client *mdsc); +void ceph_umount_begin(struct super_block *sb); /* @@ -899,6 +971,9 @@ static inline bool __ceph_have_pending_cap_snap(struct ceph_inode_info *ci) } /* inode.c */ +struct ceph_mds_reply_info_in; +struct ceph_mds_reply_dirfrag; + extern const struct inode_operations ceph_file_iops; extern struct inode *ceph_alloc_inode(struct super_block *sb); @@ -914,6 +989,11 @@ extern void ceph_fill_file_time(struct inode *inode, int issued, u64 time_warp_seq, struct timespec64 *ctime, struct timespec64 *mtime, struct timespec64 *atime); +extern int ceph_fill_inode(struct inode *inode, struct page *locked_page, + struct ceph_mds_reply_info_in *iinfo, + struct ceph_mds_reply_dirfrag *dirinfo, + struct ceph_mds_session *session, int cap_fmode, + struct ceph_cap_reservation *caps_reservation); extern int ceph_fill_trace(struct super_block *sb, struct ceph_mds_request *req); extern int ceph_readdir_prepopulate(struct ceph_mds_request *req, @@ -923,25 +1003,63 @@ extern int ceph_inode_holds_cap(struct inode *inode, int mask); extern bool ceph_inode_set_size(struct inode *inode, loff_t size); extern void __ceph_do_pending_vmtruncate(struct inode *inode); -extern void ceph_queue_vmtruncate(struct inode *inode); -extern void ceph_queue_invalidate(struct inode *inode); -extern void ceph_queue_writeback(struct inode *inode); -extern void ceph_async_iput(struct inode *inode); +void ceph_queue_inode_work(struct inode *inode, int work_bit); + +static inline void ceph_queue_vmtruncate(struct inode *inode) +{ + ceph_queue_inode_work(inode, CEPH_I_WORK_VMTRUNCATE); +} + +static inline void ceph_queue_invalidate(struct inode *inode) +{ + ceph_queue_inode_work(inode, CEPH_I_WORK_INVALIDATE_PAGES); +} + +static inline void ceph_queue_writeback(struct inode *inode) +{ + ceph_queue_inode_work(inode, CEPH_I_WORK_WRITEBACK); +} + +static inline void ceph_queue_check_caps(struct inode *inode) +{ + ceph_queue_inode_work(inode, CEPH_I_WORK_CHECK_CAPS); +} + +static inline void ceph_queue_flush_snaps(struct inode *inode) +{ + ceph_queue_inode_work(inode, CEPH_I_WORK_FLUSH_SNAPS); +} + +extern int ceph_try_to_choose_auth_mds(struct inode *inode, int mask); extern int __ceph_do_getattr(struct inode *inode, struct page *locked_page, int mask, bool force); 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_permission(struct user_namespace *mnt_userns, + 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(const struct path *path, struct kstat *stat, +extern int ceph_setattr(struct user_namespace *mnt_userns, + struct dentry *dentry, struct iattr *attr); +extern int ceph_getattr(struct user_namespace *mnt_userns, + const struct path *path, struct kstat *stat, u32 request_mask, unsigned int flags); +void ceph_inode_shutdown(struct inode *inode); + +static inline bool ceph_inode_is_shutdown(struct inode *inode) +{ + unsigned long flags = READ_ONCE(ceph_inode(inode)->i_ceph_flags); + struct ceph_fs_client *fsc = ceph_inode_to_client(inode); + int state = READ_ONCE(fsc->mount_state); + + return (flags & CEPH_I_SHUTDOWN) || state >= CEPH_MOUNT_SHUTDOWN; +} /* xattr.c */ int __ceph_setxattr(struct inode *, const char *, const void *, size_t, int); +int ceph_do_getvxattr(struct inode *inode, const char *name, void *value, size_t size); ssize_t __ceph_getxattr(struct inode *, const char *, void *, size_t); extern ssize_t ceph_listxattr(struct dentry *, char *, size_t); extern struct ceph_buffer *__ceph_build_xattrs_blob(struct ceph_inode_info *ci); @@ -997,8 +1115,9 @@ void ceph_release_acl_sec_ctx(struct ceph_acl_sec_ctx *as_ctx); /* acl.c */ #ifdef CONFIG_CEPH_FS_POSIX_ACL -struct posix_acl *ceph_get_acl(struct inode *, int); -int ceph_set_acl(struct inode *inode, struct posix_acl *acl, int type); +struct posix_acl *ceph_get_acl(struct inode *, int, bool); +int ceph_set_acl(struct user_namespace *mnt_userns, + struct inode *inode, struct posix_acl *acl, int type); int ceph_pre_init_acls(struct inode *dir, umode_t *mode, struct ceph_acl_sec_ctx *as_ctx); void ceph_init_inode_acls(struct inode *inode, @@ -1042,10 +1161,11 @@ extern struct ceph_cap *ceph_get_cap(struct ceph_mds_client *mdsc, struct ceph_cap_reservation *ctx); extern void ceph_add_cap(struct inode *inode, struct ceph_mds_session *session, u64 cap_id, - int fmode, unsigned issued, unsigned wanted, + unsigned issued, unsigned wanted, unsigned cap, unsigned seq, u64 realmino, int flags, struct ceph_cap **new_cap); extern void __ceph_remove_cap(struct ceph_cap *cap, bool queue_release); +extern void ceph_remove_cap(struct ceph_cap *cap, bool queue_release); extern void __ceph_remove_caps(struct ceph_inode_info *ci); extern void ceph_put_cap(struct ceph_mds_client *mdsc, struct ceph_cap *cap); @@ -1058,18 +1178,31 @@ extern void ceph_early_kick_flushing_caps(struct ceph_mds_client *mdsc, struct ceph_mds_session *session); extern void ceph_kick_flushing_caps(struct ceph_mds_client *mdsc, struct ceph_mds_session *session); +void ceph_kick_flushing_inode_caps(struct ceph_mds_session *session, + struct ceph_inode_info *ci); extern struct ceph_cap *ceph_get_cap_for_mds(struct ceph_inode_info *ci, int mds); +extern void ceph_take_cap_refs(struct ceph_inode_info *ci, int caps, + bool snap_rwsem_locked); extern void ceph_get_cap_refs(struct ceph_inode_info *ci, int caps); extern void ceph_put_cap_refs(struct ceph_inode_info *ci, int had); +extern void ceph_put_cap_refs_async(struct ceph_inode_info *ci, int had); +extern void ceph_put_cap_refs_no_check_caps(struct ceph_inode_info *ci, + int had); extern void ceph_put_wrbuffer_cap_refs(struct ceph_inode_info *ci, int nr, struct ceph_snap_context *snapc); +extern void __ceph_remove_capsnap(struct inode *inode, + struct ceph_cap_snap *capsnap, + bool *wake_ci, bool *wake_mdsc); +extern void ceph_remove_capsnap(struct inode *inode, + struct ceph_cap_snap *capsnap, + bool *wake_ci, bool *wake_mdsc); extern void ceph_flush_snaps(struct ceph_inode_info *ci, struct ceph_mds_session **psession); extern bool __ceph_should_report_size(struct ceph_inode_info *ci); extern void ceph_check_caps(struct ceph_inode_info *ci, int flags, struct ceph_mds_session *session); -extern void ceph_check_delayed_caps(struct ceph_mds_client *mdsc); +extern unsigned long ceph_check_delayed_caps(struct ceph_mds_client *mdsc); extern void ceph_flush_dirty_caps(struct ceph_mds_client *mdsc); extern int ceph_drop_caps_for_unlink(struct inode *inode); extern int ceph_encode_inode_release(void **p, struct inode *inode, @@ -1079,25 +1212,37 @@ extern int ceph_encode_dentry_release(void **p, struct dentry *dn, int mds, int drop, int unless); extern int ceph_get_caps(struct file *filp, int need, int want, - loff_t endoff, int *got, struct page **pinned_page); + loff_t endoff, int *got); extern int ceph_try_get_caps(struct inode *inode, int need, int want, bool nonblock, int *got); /* for counting open files by mode */ -extern void __ceph_get_fmode(struct ceph_inode_info *ci, int mode); -extern void ceph_put_fmode(struct ceph_inode_info *ci, int mode); +extern void ceph_get_fmode(struct ceph_inode_info *ci, int mode, int count); +extern void ceph_put_fmode(struct ceph_inode_info *ci, int mode, int count); +extern void __ceph_touch_fmode(struct ceph_inode_info *ci, + struct ceph_mds_client *mdsc, int fmode); /* addr.c */ extern const struct address_space_operations ceph_aops; +extern const struct netfs_request_ops ceph_netfs_ops; extern int ceph_mmap(struct file *file, struct vm_area_struct *vma); -extern int ceph_uninline_data(struct file *filp, struct page *locked_page); +extern int ceph_uninline_data(struct file *file); extern int ceph_pool_perm_check(struct inode *inode, int need); extern void ceph_pool_perm_destroy(struct ceph_mds_client* mdsc); +int ceph_purge_inode_cap(struct inode *inode, struct ceph_cap *cap, bool *invalidate); + +static inline bool ceph_has_inline_data(struct ceph_inode_info *ci) +{ + if (ci->i_inline_version == CEPH_INLINE_NONE || + ci->i_inline_version == 1) /* initial version, no data */ + return false; + return true; +} /* file.c */ extern const struct file_operations ceph_file_fops; -extern int ceph_renew_caps(struct inode *inode); +extern int ceph_renew_caps(struct inode *inode, int fmode); 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); @@ -1114,8 +1259,8 @@ extern const struct dentry_operations ceph_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); +extern struct dentry *ceph_handle_snapdir(struct ceph_mds_request *req, + struct dentry *dentry); extern struct dentry *ceph_finish_lookup(struct ceph_mds_request *req, struct dentry *dentry, int err); @@ -1151,9 +1296,29 @@ extern void ceph_fs_debugfs_init(struct ceph_fs_client *client); extern void ceph_fs_debugfs_cleanup(struct ceph_fs_client *client); /* quota.c */ -static inline bool __ceph_has_any_quota(struct ceph_inode_info *ci) + +enum quota_get_realm { + QUOTA_GET_MAX_FILES, + QUOTA_GET_MAX_BYTES, + QUOTA_GET_ANY +}; + +static inline bool __ceph_has_quota(struct ceph_inode_info *ci, + enum quota_get_realm which) { - return ci->i_max_files || ci->i_max_bytes; + bool has_quota = false; + + switch (which) { + case QUOTA_GET_MAX_BYTES: + has_quota = !!ci->i_max_bytes; + break; + case QUOTA_GET_MAX_FILES: + has_quota = !!ci->i_max_files; + break; + default: + has_quota = !!(ci->i_max_files || ci->i_max_bytes); + } + return has_quota; } extern void ceph_adjust_quota_realms_count(struct inode *inode, bool inc); @@ -1162,13 +1327,13 @@ static inline void __ceph_update_quota(struct ceph_inode_info *ci, u64 max_bytes, u64 max_files) { bool had_quota, has_quota; - had_quota = __ceph_has_any_quota(ci); + had_quota = __ceph_has_quota(ci, QUOTA_GET_ANY); ci->i_max_bytes = max_bytes; ci->i_max_files = max_files; - has_quota = __ceph_has_any_quota(ci); + has_quota = __ceph_has_quota(ci, QUOTA_GET_ANY); if (had_quota != has_quota) - ceph_adjust_quota_realms_count(&ci->vfs_inode, has_quota); + ceph_adjust_quota_realms_count(&ci->netfs.inode, has_quota); } extern void ceph_handle_quota(struct ceph_mds_client *mdsc, diff --git a/fs/ceph/xattr.c b/fs/ceph/xattr.c index 7b8a070a782d..f31350cda960 100644 --- a/fs/ceph/xattr.c +++ b/fs/ceph/xattr.c @@ -42,6 +42,7 @@ struct ceph_vxattr { #define VXATTR_FLAG_READONLY (1<<0) #define VXATTR_FLAG_HIDDEN (1<<1) #define VXATTR_FLAG_RSTAT (1<<2) +#define VXATTR_FLAG_DIRSTAT (1<<3) /* layouts */ @@ -56,7 +57,7 @@ static bool ceph_vxattrcb_layout_exists(struct ceph_inode_info *ci) static ssize_t ceph_vxattrcb_layout(struct ceph_inode_info *ci, char *val, size_t size) { - struct ceph_fs_client *fsc = ceph_sb_to_client(ci->vfs_inode.i_sb); + struct ceph_fs_client *fsc = ceph_sb_to_client(ci->netfs.inode.i_sb); struct ceph_osd_client *osdc = &fsc->client->osdc; struct ceph_string *pool_ns; s64 pool = ci->i_layout.pool_id; @@ -68,7 +69,7 @@ static ssize_t ceph_vxattrcb_layout(struct ceph_inode_info *ci, char *val, pool_ns = ceph_try_get_string(ci->i_layout.pool_ns); - dout("ceph_vxattrcb_layout %p\n", &ci->vfs_inode); + dout("ceph_vxattrcb_layout %p\n", &ci->netfs.inode); down_read(&osdc->lock); pool_name = ceph_pg_pool_name_by_id(osdc->osdmap, pool); if (pool_name) { @@ -116,7 +117,8 @@ static ssize_t ceph_vxattrcb_layout(struct ceph_inode_info *ci, char *val, * NULL terminates however, so call it on a temporary buffer and then memcpy * the result into place. */ -static int ceph_fmt_xattr(char *val, size_t size, const char *fmt, ...) +static __printf(3, 4) +int ceph_fmt_xattr(char *val, size_t size, const char *fmt, ...) { int ret; va_list args; @@ -159,7 +161,7 @@ static ssize_t ceph_vxattrcb_layout_pool(struct ceph_inode_info *ci, char *val, size_t size) { ssize_t ret; - struct ceph_fs_client *fsc = ceph_sb_to_client(ci->vfs_inode.i_sb); + struct ceph_fs_client *fsc = ceph_sb_to_client(ci->netfs.inode.i_sb); struct ceph_osd_client *osdc = &fsc->client->osdc; s64 pool = ci->i_layout.pool_id; const char *pool_name; @@ -231,6 +233,12 @@ static ssize_t ceph_vxattrcb_dir_rsubdirs(struct ceph_inode_info *ci, char *val, return ceph_fmt_xattr(val, size, "%lld", ci->i_rsubdirs); } +static ssize_t ceph_vxattrcb_dir_rsnaps(struct ceph_inode_info *ci, char *val, + size_t size) +{ + return ceph_fmt_xattr(val, size, "%lld", ci->i_rsnaps); +} + static ssize_t ceph_vxattrcb_dir_rbytes(struct ceph_inode_info *ci, char *val, size_t size) { @@ -302,6 +310,48 @@ static ssize_t ceph_vxattrcb_snap_btime(struct ceph_inode_info *ci, char *val, ci->i_snap_btime.tv_nsec); } +static ssize_t ceph_vxattrcb_cluster_fsid(struct ceph_inode_info *ci, + char *val, size_t size) +{ + struct ceph_fs_client *fsc = ceph_sb_to_client(ci->netfs.inode.i_sb); + + return ceph_fmt_xattr(val, size, "%pU", &fsc->client->fsid); +} + +static ssize_t ceph_vxattrcb_client_id(struct ceph_inode_info *ci, + char *val, size_t size) +{ + struct ceph_fs_client *fsc = ceph_sb_to_client(ci->netfs.inode.i_sb); + + return ceph_fmt_xattr(val, size, "client%lld", + ceph_client_gid(fsc->client)); +} + +static ssize_t ceph_vxattrcb_caps(struct ceph_inode_info *ci, char *val, + size_t size) +{ + int issued; + + spin_lock(&ci->i_ceph_lock); + issued = __ceph_caps_issued(ci, NULL); + spin_unlock(&ci->i_ceph_lock); + + return ceph_fmt_xattr(val, size, "%s/0x%x", + ceph_cap_string(issued), issued); +} + +static ssize_t ceph_vxattrcb_auth_mds(struct ceph_inode_info *ci, + char *val, size_t size) +{ + int ret; + + spin_lock(&ci->i_ceph_lock); + ret = ceph_fmt_xattr(val, size, "%d", + ci->i_auth_cap ? ci->i_auth_cap->session->s_mds : -1); + spin_unlock(&ci->i_ceph_lock); + return ret; +} + #define CEPH_XATTR_NAME(_type, _name) XATTR_CEPH_PREFIX #_type "." #_name #define CEPH_XATTR_NAME2(_type, _name, _name2) \ XATTR_CEPH_PREFIX #_type "." #_name "." #_name2 @@ -316,6 +366,14 @@ static ssize_t ceph_vxattrcb_snap_btime(struct ceph_inode_info *ci, char *val, } #define XATTR_RSTAT_FIELD(_type, _name) \ XATTR_NAME_CEPH(_type, _name, VXATTR_FLAG_RSTAT) +#define XATTR_RSTAT_FIELD_UPDATABLE(_type, _name) \ + { \ + .name = CEPH_XATTR_NAME(_type, _name), \ + .name_size = sizeof (CEPH_XATTR_NAME(_type, _name)), \ + .getxattr_cb = ceph_vxattrcb_ ## _type ## _ ## _name, \ + .exists_cb = NULL, \ + .flags = VXATTR_FLAG_RSTAT, \ + } #define XATTR_LAYOUT_FIELD(_type, _name, _field) \ { \ .name = CEPH_XATTR_NAME2(_type, _name, _field), \ @@ -346,14 +404,15 @@ static struct ceph_vxattr ceph_dir_vxattrs[] = { XATTR_LAYOUT_FIELD(dir, layout, object_size), XATTR_LAYOUT_FIELD(dir, layout, pool), XATTR_LAYOUT_FIELD(dir, layout, pool_namespace), - XATTR_NAME_CEPH(dir, entries, 0), - XATTR_NAME_CEPH(dir, files, 0), - XATTR_NAME_CEPH(dir, subdirs, 0), + XATTR_NAME_CEPH(dir, entries, VXATTR_FLAG_DIRSTAT), + XATTR_NAME_CEPH(dir, files, VXATTR_FLAG_DIRSTAT), + XATTR_NAME_CEPH(dir, subdirs, VXATTR_FLAG_DIRSTAT), XATTR_RSTAT_FIELD(dir, rentries), XATTR_RSTAT_FIELD(dir, rfiles), XATTR_RSTAT_FIELD(dir, rsubdirs), + XATTR_RSTAT_FIELD(dir, rsnaps), XATTR_RSTAT_FIELD(dir, rbytes), - XATTR_RSTAT_FIELD(dir, rctime), + XATTR_RSTAT_FIELD_UPDATABLE(dir, rctime), { .name = "ceph.dir.pin", .name_size = sizeof("ceph.dir.pin"), @@ -377,6 +436,13 @@ static struct ceph_vxattr ceph_dir_vxattrs[] = { .exists_cb = ceph_vxattrcb_snap_btime_exists, .flags = VXATTR_FLAG_READONLY, }, + { + .name = "ceph.caps", + .name_size = sizeof("ceph.caps"), + .getxattr_cb = ceph_vxattrcb_caps, + .exists_cb = NULL, + .flags = VXATTR_FLAG_HIDDEN, + }, { .name = NULL, 0 } /* Required table terminator */ }; @@ -402,6 +468,38 @@ static struct ceph_vxattr ceph_file_vxattrs[] = { .exists_cb = ceph_vxattrcb_snap_btime_exists, .flags = VXATTR_FLAG_READONLY, }, + { + .name = "ceph.caps", + .name_size = sizeof("ceph.caps"), + .getxattr_cb = ceph_vxattrcb_caps, + .exists_cb = NULL, + .flags = VXATTR_FLAG_HIDDEN, + }, + { .name = NULL, 0 } /* Required table terminator */ +}; + +static struct ceph_vxattr ceph_common_vxattrs[] = { + { + .name = "ceph.cluster_fsid", + .name_size = sizeof("ceph.cluster_fsid"), + .getxattr_cb = ceph_vxattrcb_cluster_fsid, + .exists_cb = NULL, + .flags = VXATTR_FLAG_READONLY, + }, + { + .name = "ceph.client_id", + .name_size = sizeof("ceph.client_id"), + .getxattr_cb = ceph_vxattrcb_client_id, + .exists_cb = NULL, + .flags = VXATTR_FLAG_READONLY, + }, + { + .name = "ceph.auth_mds", + .name_size = sizeof("ceph.auth_mds"), + .getxattr_cb = ceph_vxattrcb_auth_mds, + .exists_cb = NULL, + .flags = VXATTR_FLAG_READONLY, + }, { .name = NULL, 0 } /* Required table terminator */ }; @@ -427,6 +525,13 @@ static struct ceph_vxattr *ceph_match_vxattr(struct inode *inode, } } + vxattr = ceph_common_vxattrs; + while (vxattr->name) { + if (!strcmp(vxattr->name, name)) + return vxattr; + vxattr++; + } + return NULL; } @@ -497,10 +602,10 @@ static int __set_xattr(struct ceph_inode_info *ci, kfree(*newxattr); *newxattr = NULL; if (xattr->should_free_val) - kfree((void *)xattr->val); + kfree(xattr->val); if (update_xattr) { - kfree((void *)name); + kfree(name); name = xattr->name; } ci->i_xattrs.names_size -= xattr->name_len; @@ -524,7 +629,7 @@ static int __set_xattr(struct ceph_inode_info *ci, } dout("__set_xattr_val added %llx.%llx xattr %p %.*s=%.*s\n", - ceph_vinop(&ci->vfs_inode), xattr, name_len, name, val_len, val); + ceph_vinop(&ci->netfs.inode), xattr, name_len, name, val_len, val); return 0; } @@ -566,9 +671,9 @@ static void __free_xattr(struct ceph_inode_xattr *xattr) BUG_ON(!xattr); if (xattr->should_free_name) - kfree((void *)xattr->name); + kfree(xattr->name); if (xattr->should_free_val) - kfree((void *)xattr->val); + kfree(xattr->val); kfree(xattr); } @@ -582,9 +687,9 @@ static int __remove_xattr(struct ceph_inode_info *ci, rb_erase(&xattr->node, &ci->i_xattrs.index); if (xattr->should_free_name) - kfree((void *)xattr->name); + kfree(xattr->name); if (xattr->should_free_val) - kfree((void *)xattr->val); + kfree(xattr->val); ci->i_xattrs.names_size -= xattr->name_len; ci->i_xattrs.vals_size -= xattr->val_len; @@ -766,7 +871,7 @@ struct ceph_buffer *__ceph_build_xattrs_blob(struct ceph_inode_info *ci) struct ceph_buffer *old_blob = NULL; void *dest; - dout("__build_xattrs_blob %p\n", &ci->vfs_inode); + dout("__build_xattrs_blob %p\n", &ci->netfs.inode); if (ci->i_xattrs.dirty) { int need = __get_required_blob_size(ci, 0, 0); @@ -826,16 +931,21 @@ ssize_t __ceph_getxattr(struct inode *inode, const char *name, void *value, { struct ceph_inode_info *ci = ceph_inode(inode); struct ceph_inode_xattr *xattr; - struct ceph_vxattr *vxattr = NULL; + struct ceph_vxattr *vxattr; int req_mask; ssize_t err; + if (strncmp(name, XATTR_CEPH_PREFIX, XATTR_CEPH_PREFIX_LEN)) + goto handle_non_vxattrs; + /* let's see if a virtual xattr was requested */ vxattr = ceph_match_vxattr(inode, name); if (vxattr) { int mask = 0; if (vxattr->flags & VXATTR_FLAG_RSTAT) mask |= CEPH_STAT_RSTAT; + if (vxattr->flags & VXATTR_FLAG_DIRSTAT) + mask |= CEPH_CAP_FILE_SHARED; err = ceph_do_getattr(inode, mask, true); if (err) return err; @@ -846,8 +956,14 @@ ssize_t __ceph_getxattr(struct inode *inode, const char *name, void *value, err = -ERANGE; } return err; + } else { + err = ceph_do_getvxattr(inode, name, value, size); + /* this would happen with a new client and old server combo */ + if (err == -EOPNOTSUPP) + err = -ENODATA; + return err; } - +handle_non_vxattrs: req_mask = __get_request_mask(inode); spin_lock(&ci->i_ceph_lock); @@ -856,7 +972,7 @@ ssize_t __ceph_getxattr(struct inode *inode, const char *name, void *value, if (ci->i_xattrs.version == 0 || !((req_mask & CEPH_CAP_XATTR_SHARED) || - __ceph_caps_issued_mask(ci, CEPH_CAP_XATTR_SHARED, 1))) { + __ceph_caps_issued_mask_metric(ci, CEPH_CAP_XATTR_SHARED, 1))) { spin_unlock(&ci->i_ceph_lock); /* security module gets xattr while filling trace */ @@ -914,7 +1030,7 @@ ssize_t ceph_listxattr(struct dentry *dentry, char *names, size_t size) ci->i_xattrs.version, ci->i_xattrs.index_version); if (ci->i_xattrs.version == 0 || - !__ceph_caps_issued_mask(ci, CEPH_CAP_XATTR_SHARED, 1)) { + !__ceph_caps_issued_mask_metric(ci, CEPH_CAP_XATTR_SHARED, 1)) { spin_unlock(&ci->i_ceph_lock); err = ceph_do_getattr(inode, CEPH_STAT_CAP_XATTR, true); if (err) @@ -949,6 +1065,7 @@ static int ceph_sync_setxattr(struct inode *inode, const char *name, struct ceph_inode_info *ci = ceph_inode(inode); struct ceph_mds_request *req; struct ceph_mds_client *mdsc = fsc->mdsc; + struct ceph_osd_client *osdc = &fsc->client->osdc; struct ceph_pagelist *pagelist = NULL; int op = CEPH_MDS_OP_SETXATTR; int err; @@ -969,7 +1086,7 @@ static int ceph_sync_setxattr(struct inode *inode, const char *name, flags |= CEPH_XATTR_REMOVE; } - dout("setxattr value=%.*s\n", (int)size, value); + dout("setxattr value size: %zu\n", size); /* do request */ req = ceph_mdsc_create_request(mdsc, op, USE_AUTH_MDS); @@ -987,6 +1104,8 @@ static int ceph_sync_setxattr(struct inode *inode, const char *name, if (op == CEPH_MDS_OP_SETXATTR) { req->r_args.setxattr.flags = cpu_to_le32(flags); + req->r_args.setxattr.osdmap_epoch = + cpu_to_le32(osdc->osdmap->epoch); req->r_pagelist = pagelist; pagelist = NULL; } @@ -1065,8 +1184,14 @@ int __ceph_setxattr(struct inode *inode, const char *name, spin_lock(&ci->i_ceph_lock); retry: issued = __ceph_caps_issued(ci, NULL); - if (ci->i_xattrs.version == 0 || !(issued & CEPH_CAP_XATTR_EXCL)) + required_blob_size = __get_required_blob_size(ci, name_len, val_len); + if ((ci->i_xattrs.version == 0) || !(issued & CEPH_CAP_XATTR_EXCL) || + (required_blob_size > mdsc->mdsmap->m_max_xattr_size)) { + dout("%s do sync setxattr: version: %llu size: %d max: %llu\n", + __func__, ci->i_xattrs.version, required_blob_size, + mdsc->mdsmap->m_max_xattr_size); goto do_sync; + } if (!lock_snap_rwsem && !ci->i_head_snapc) { lock_snap_rwsem = true; @@ -1082,8 +1207,6 @@ retry: ceph_cap_string(issued)); __build_xattrs(inode); - required_blob_size = __get_required_blob_size(ci, name_len, val_len); - if (!ci->i_xattrs.prealloc_blob || required_blob_size > ci->i_xattrs.prealloc_blob->alloc_len) { struct ceph_buffer *blob; @@ -1162,6 +1285,7 @@ static int ceph_get_xattr_handler(const struct xattr_handler *handler, } static int ceph_set_xattr_handler(const struct xattr_handler *handler, + struct user_namespace *mnt_userns, struct dentry *unused, struct inode *inode, const char *name, const void *value, size_t size, int flags) @@ -1208,7 +1332,7 @@ int ceph_security_init_secctx(struct dentry *dentry, umode_t mode, int err; err = security_dentry_init_security(dentry, mode, &dentry->d_name, - &as_ctx->sec_ctx, + &name, &as_ctx->sec_ctx, &as_ctx->sec_ctxlen); if (err < 0) { WARN_ON_ONCE(err != -EOPNOTSUPP); @@ -1232,7 +1356,6 @@ int ceph_security_init_secctx(struct dentry *dentry, umode_t mode, * It only supports single security module and only selinux has * dentry_init_security hook. */ - name = XATTR_NAME_SELINUX; name_len = strlen(name); err = ceph_pagelist_reserve(pagelist, 4 * 2 + name_len + as_ctx->sec_ctxlen); |