From e7508d48565060af5d89f10cb83c9359c8ae1310 Mon Sep 17 00:00:00 2001 From: Md Haris Iqbal Date: Thu, 10 Dec 2020 11:18:20 +0100 Subject: block/rnbd-clt: Get rid of warning regarding size argument in strlcpy The kernel test robot triggerred the following warning, >> drivers/block/rnbd/rnbd-clt.c:1397:42: warning: size argument in 'strlcpy' call appears to be size of the source; expected the size of the destination [-Wstrlcpy-strlcat-size] strlcpy(dev->pathname, pathname, strlen(pathname) + 1); ~~~~~~~^~~~~~~~~~~~~ To get rid of the above warning, use a kstrdup as Bart suggested. Fixes: 64e8a6ece1a5 ("block/rnbd-clt: Dynamically alloc buffer for pathname & blk_symlink_name") Reported-by: kernel test robot Signed-off-by: Md Haris Iqbal Signed-off-by: Jack Wang Signed-off-by: Jens Axboe --- drivers/block/rnbd/rnbd-clt.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'drivers') diff --git a/drivers/block/rnbd/rnbd-clt.c b/drivers/block/rnbd/rnbd-clt.c index a199b190c73d..d63f0974bd04 100644 --- a/drivers/block/rnbd/rnbd-clt.c +++ b/drivers/block/rnbd/rnbd-clt.c @@ -1388,12 +1388,11 @@ static struct rnbd_clt_dev *init_dev(struct rnbd_clt_session *sess, goto out_queues; } - dev->pathname = kzalloc(strlen(pathname) + 1, GFP_KERNEL); + dev->pathname = kstrdup(pathname, GFP_KERNEL); if (!dev->pathname) { ret = -ENOMEM; goto out_queues; } - strlcpy(dev->pathname, pathname, strlen(pathname) + 1); dev->clt_device_id = ret; dev->sess = sess; -- cgit v1.2.3-59-g8ed1b From 46067844efdb8275ade705923120fc5391543b53 Mon Sep 17 00:00:00 2001 From: Jack Wang Date: Thu, 10 Dec 2020 11:18:21 +0100 Subject: block/rnbd-clt: Fix possible memleak In error case, we do not free the memory for blk_symlink_name. Do it by free the memory in error case, and set to NULL afterwards. Also fix the condition in rnbd_clt_remove_dev_symlink. Fixes: 64e8a6ece1a5 ("block/rnbd-clt: Dynamically alloc buffer for pathname & blk_symlink_name") Signed-off-by: Jack Wang Reviewed-by: Md Haris Iqbal Signed-off-by: Jens Axboe --- drivers/block/rnbd/rnbd-clt-sysfs.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'drivers') diff --git a/drivers/block/rnbd/rnbd-clt-sysfs.c b/drivers/block/rnbd/rnbd-clt-sysfs.c index a7caeedeb198..d4aa6bfc9555 100644 --- a/drivers/block/rnbd/rnbd-clt-sysfs.c +++ b/drivers/block/rnbd/rnbd-clt-sysfs.c @@ -432,7 +432,7 @@ void rnbd_clt_remove_dev_symlink(struct rnbd_clt_dev *dev) * i.e. rnbd_clt_unmap_dev_store() leading to a sysfs warning because * of sysfs link already was removed already. */ - if (strlen(dev->blk_symlink_name) && try_module_get(THIS_MODULE)) { + if (dev->blk_symlink_name && try_module_get(THIS_MODULE)) { sysfs_remove_link(rnbd_devs_kobj, dev->blk_symlink_name); kfree(dev->blk_symlink_name); module_put(THIS_MODULE); @@ -521,7 +521,8 @@ static int rnbd_clt_add_dev_symlink(struct rnbd_clt_dev *dev) return 0; out_err: - dev->blk_symlink_name[0] = '\0'; + kfree(dev->blk_symlink_name); + dev->blk_symlink_name = NULL ; return ret; } -- cgit v1.2.3-59-g8ed1b From 87019e7d99d707e60e20ea3245a561419d5de5ce Mon Sep 17 00:00:00 2001 From: Md Haris Iqbal Date: Thu, 10 Dec 2020 11:18:22 +0100 Subject: block/rnbd-srv: Protect dev session sysfs removal Since the removal of the session sysfs can also be called from the function destroy_sess, there is a need to protect the call from the function rnbd_srv_sess_dev_force_close Fixes: 786998050cbc ("block/rnbd-srv: close a mapped device from server side.") Signed-off-by: Md Haris Iqbal Reviewed-by: Guoqing Jiang Signed-off-by: Jack Wang Signed-off-by: Jens Axboe --- drivers/block/rnbd/rnbd-srv.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'drivers') diff --git a/drivers/block/rnbd/rnbd-srv.c b/drivers/block/rnbd/rnbd-srv.c index d1ee72ed8384..066411cce5e2 100644 --- a/drivers/block/rnbd/rnbd-srv.c +++ b/drivers/block/rnbd/rnbd-srv.c @@ -338,9 +338,10 @@ static int rnbd_srv_link_ev(struct rtrs_srv *rtrs, void rnbd_srv_sess_dev_force_close(struct rnbd_srv_sess_dev *sess_dev) { + mutex_lock(&sess_dev->sess->lock); rnbd_srv_destroy_dev_session_sysfs(sess_dev); + mutex_unlock(&sess_dev->sess->lock); sess_dev->keep_id = true; - } static int process_msg_close(struct rtrs_srv *rtrs, -- cgit v1.2.3-59-g8ed1b From 3877ece01e46f01fae0fbc00df93d0e5f23196b0 Mon Sep 17 00:00:00 2001 From: Jack Wang Date: Thu, 10 Dec 2020 11:18:23 +0100 Subject: block/rnbd: Fix typos Signed-off-by: Jack Wang Signed-off-by: Jens Axboe --- drivers/block/rnbd/rnbd-clt.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'drivers') diff --git a/drivers/block/rnbd/rnbd-clt.c b/drivers/block/rnbd/rnbd-clt.c index d63f0974bd04..3a2e6e8ed6b1 100644 --- a/drivers/block/rnbd/rnbd-clt.c +++ b/drivers/block/rnbd/rnbd-clt.c @@ -359,7 +359,7 @@ static struct rnbd_iu *rnbd_get_iu(struct rnbd_clt_session *sess, * 2nd reference is dropped after confirmation with the response is * returned. * 1st and 2nd can happen in any order, so the rnbd_iu should be - * released (rtrs_permit returned to ibbtrs) only leased after both + * released (rtrs_permit returned to rtrs) only after both * are finished. */ atomic_set(&iu->refcount, 2); @@ -803,7 +803,7 @@ static struct rnbd_clt_session *alloc_sess(const char *sessname) rnbd_init_cpu_qlists(sess->cpu_queues); /* - * That is simple percpu variable which stores cpu indeces, which are + * That is simple percpu variable which stores cpu indices, which are * incremented on each access. We need that for the sake of fairness * to wake up queues in a round-robin manner. */ @@ -1666,7 +1666,7 @@ static void rnbd_destroy_sessions(void) /* * Here at this point there is no any concurrent access to sessions * list and devices list: - * 1. New session or device can'be be created - session sysfs files + * 1. New session or device can't be created - session sysfs files * are removed. * 2. Device or session can't be removed - module reference is taken * into account in unmap device sysfs callback. -- cgit v1.2.3-59-g8ed1b From 512c781fd28cb401ee9f2843e32bf4640732c671 Mon Sep 17 00:00:00 2001 From: Gioh Kim Date: Thu, 10 Dec 2020 11:18:24 +0100 Subject: block/rnbd: Set write-back cache and fua same to the target device The rnbd-client always sets the write-back cache and fua attributes of the rnbd device queue regardless of the target device on the server. That generates IO hang issue when the target device does not support both of write-back cacne and fua. This patch adds more fields for the cache policy and fua into the device opening message. The rnbd-server sends the information if the target device supports the write-back cache and fua and rnbd-client recevives it and set the device queue accordingly. Signed-off-by: Gioh Kim [jwang: some minor change, rename a few varables, remove unrelated comments.] Signed-off-by: Jack Wang Signed-off-by: Jens Axboe --- drivers/block/rnbd/rnbd-clt.c | 8 +++++--- drivers/block/rnbd/rnbd-clt.h | 2 ++ drivers/block/rnbd/rnbd-proto.h | 9 ++++++++- drivers/block/rnbd/rnbd-srv.c | 9 +++++++-- 4 files changed, 22 insertions(+), 6 deletions(-) (limited to 'drivers') diff --git a/drivers/block/rnbd/rnbd-clt.c b/drivers/block/rnbd/rnbd-clt.c index 3a2e6e8ed6b1..b5fffbdeb263 100644 --- a/drivers/block/rnbd/rnbd-clt.c +++ b/drivers/block/rnbd/rnbd-clt.c @@ -88,6 +88,8 @@ static int rnbd_clt_set_dev_attr(struct rnbd_clt_dev *dev, dev->discard_alignment = le32_to_cpu(rsp->discard_alignment); dev->secure_discard = le16_to_cpu(rsp->secure_discard); dev->rotational = rsp->rotational; + dev->wc = !!(rsp->cache_policy & RNBD_WRITEBACK); + dev->fua = !!(rsp->cache_policy & RNBD_FUA); dev->max_hw_sectors = sess->max_io_size / SECTOR_SIZE; dev->max_segments = BMAX_SEGMENTS; @@ -1305,7 +1307,7 @@ static void setup_request_queue(struct rnbd_clt_dev *dev) blk_queue_max_segments(dev->queue, dev->max_segments); blk_queue_io_opt(dev->queue, dev->sess->max_io_size); blk_queue_virt_boundary(dev->queue, SZ_4K - 1); - blk_queue_write_cache(dev->queue, true, true); + blk_queue_write_cache(dev->queue, dev->wc, dev->fua); dev->queue->queuedata = dev; } @@ -1528,13 +1530,13 @@ struct rnbd_clt_dev *rnbd_clt_map_device(const char *sessname, } rnbd_clt_info(dev, - "map_device: Device mapped as %s (nsectors: %zu, logical_block_size: %d, physical_block_size: %d, max_write_same_sectors: %d, max_discard_sectors: %d, discard_granularity: %d, discard_alignment: %d, secure_discard: %d, max_segments: %d, max_hw_sectors: %d, rotational: %d)\n", + "map_device: Device mapped as %s (nsectors: %zu, logical_block_size: %d, physical_block_size: %d, max_write_same_sectors: %d, max_discard_sectors: %d, discard_granularity: %d, discard_alignment: %d, secure_discard: %d, max_segments: %d, max_hw_sectors: %d, rotational: %d, wc: %d, fua: %d)\n", dev->gd->disk_name, dev->nsectors, dev->logical_block_size, dev->physical_block_size, dev->max_write_same_sectors, dev->max_discard_sectors, dev->discard_granularity, dev->discard_alignment, dev->secure_discard, dev->max_segments, - dev->max_hw_sectors, dev->rotational); + dev->max_hw_sectors, dev->rotational, dev->wc, dev->fua); mutex_unlock(&dev->lock); diff --git a/drivers/block/rnbd/rnbd-clt.h b/drivers/block/rnbd/rnbd-clt.h index b193d5904050..efd67ae286ca 100644 --- a/drivers/block/rnbd/rnbd-clt.h +++ b/drivers/block/rnbd/rnbd-clt.h @@ -112,6 +112,8 @@ struct rnbd_clt_dev { enum rnbd_access_mode access_mode; bool read_only; bool rotational; + bool wc; + bool fua; u32 max_hw_sectors; u32 max_write_same_sectors; u32 max_discard_sectors; diff --git a/drivers/block/rnbd/rnbd-proto.h b/drivers/block/rnbd/rnbd-proto.h index ca166241452c..c1bc5c0fef71 100644 --- a/drivers/block/rnbd/rnbd-proto.h +++ b/drivers/block/rnbd/rnbd-proto.h @@ -108,6 +108,11 @@ struct rnbd_msg_close { __le32 device_id; }; +enum rnbd_cache_policy { + RNBD_FUA = 1 << 0, + RNBD_WRITEBACK = 1 << 1, +}; + /** * struct rnbd_msg_open_rsp - response message to RNBD_MSG_OPEN * @hdr: message header @@ -124,6 +129,7 @@ struct rnbd_msg_close { * @max_segments: max segments hardware support in one transfer * @secure_discard: supports secure discard * @rotation: is a rotational disc? + * @cache_policy: support write-back caching or FUA? */ struct rnbd_msg_open_rsp { struct rnbd_msg_hdr hdr; @@ -139,7 +145,8 @@ struct rnbd_msg_open_rsp { __le16 max_segments; __le16 secure_discard; u8 rotational; - u8 reserved[11]; + u8 cache_policy; + u8 reserved[10]; }; /** diff --git a/drivers/block/rnbd/rnbd-srv.c b/drivers/block/rnbd/rnbd-srv.c index 066411cce5e2..b8e44331e494 100644 --- a/drivers/block/rnbd/rnbd-srv.c +++ b/drivers/block/rnbd/rnbd-srv.c @@ -550,6 +550,7 @@ static void rnbd_srv_fill_msg_open_rsp(struct rnbd_msg_open_rsp *rsp, struct rnbd_srv_sess_dev *sess_dev) { struct rnbd_dev *rnbd_dev = sess_dev->rnbd_dev; + struct request_queue *q = bdev_get_queue(rnbd_dev->bdev); rsp->hdr.type = cpu_to_le16(RNBD_MSG_OPEN_RSP); rsp->device_id = @@ -574,8 +575,12 @@ static void rnbd_srv_fill_msg_open_rsp(struct rnbd_msg_open_rsp *rsp, cpu_to_le32(rnbd_dev_get_discard_alignment(rnbd_dev)); rsp->secure_discard = cpu_to_le16(rnbd_dev_get_secure_discard(rnbd_dev)); - rsp->rotational = - !blk_queue_nonrot(bdev_get_queue(rnbd_dev->bdev)); + rsp->rotational = !blk_queue_nonrot(q); + rsp->cache_policy = 0; + if (test_bit(QUEUE_FLAG_WC, &q->queue_flags)) + rsp->cache_policy |= RNBD_WRITEBACK; + if (blk_queue_fua(q)) + rsp->cache_policy |= RNBD_FUA; } static struct rnbd_srv_sess_dev * -- cgit v1.2.3-59-g8ed1b From 5a1328d0c3a757cdd8c65f4dfe0a02502a5810bc Mon Sep 17 00:00:00 2001 From: Gioh Kim Date: Thu, 10 Dec 2020 11:18:25 +0100 Subject: block/rnbd-clt: Dynamically allocate sglist for rnbd_iu The BMAX_SEGMENT static array for scatterlist is embedded in rnbd_iu structure to avoid memory allocation in hot IO path. In many cases, we do need only several sg entries because many IOs have only several segments. This patch change rnbd_iu to check the number of segments in the request and allocate sglist dynamically. For io path, use sg_alloc_table_chained to allocate sg list faster. First it makes two sg entries after pdu of request. The sg_alloc_table_chained uses the pre-allocated sg entries if the number of segments of the request is less than two. So it reduces the number of memory allocation. Signed-off-by: Gioh Kim Signed-off-by: Jack Wang Signed-off-by: Jens Axboe --- drivers/block/rnbd/rnbd-clt.c | 60 ++++++++++++++++++++++++------------------- drivers/block/rnbd/rnbd-clt.h | 10 +++++++- 2 files changed, 43 insertions(+), 27 deletions(-) (limited to 'drivers') diff --git a/drivers/block/rnbd/rnbd-clt.c b/drivers/block/rnbd/rnbd-clt.c index b5fffbdeb263..5941ff7c83a8 100644 --- a/drivers/block/rnbd/rnbd-clt.c +++ b/drivers/block/rnbd/rnbd-clt.c @@ -384,6 +384,7 @@ static void rnbd_softirq_done_fn(struct request *rq) struct rnbd_iu *iu; iu = blk_mq_rq_to_pdu(rq); + sg_free_table_chained(&iu->sgt, RNBD_INLINE_SG_CNT); rnbd_put_permit(sess, iu->permit); blk_mq_end_request(rq, errno_to_blk_status(iu->errno)); } @@ -477,7 +478,7 @@ static int send_msg_close(struct rnbd_clt_dev *dev, u32 device_id, bool wait) iu->buf = NULL; iu->dev = dev; - sg_mark_end(&iu->sglist[0]); + sg_alloc_table(&iu->sgt, 1, GFP_KERNEL); msg.hdr.type = cpu_to_le16(RNBD_MSG_CLOSE); msg.device_id = cpu_to_le32(device_id); @@ -492,6 +493,7 @@ static int send_msg_close(struct rnbd_clt_dev *dev, u32 device_id, bool wait) err = errno; } + sg_free_table(&iu->sgt); rnbd_put_iu(sess, iu); return err; } @@ -564,7 +566,8 @@ static int send_msg_open(struct rnbd_clt_dev *dev, bool wait) iu->buf = rsp; iu->dev = dev; - sg_init_one(iu->sglist, rsp, sizeof(*rsp)); + sg_alloc_table(&iu->sgt, 1, GFP_KERNEL); + sg_init_one(iu->sgt.sgl, rsp, sizeof(*rsp)); msg.hdr.type = cpu_to_le16(RNBD_MSG_OPEN); msg.access_mode = dev->access_mode; @@ -572,7 +575,7 @@ static int send_msg_open(struct rnbd_clt_dev *dev, bool wait) WARN_ON(!rnbd_clt_get_dev(dev)); err = send_usr_msg(sess->rtrs, READ, iu, - &vec, sizeof(*rsp), iu->sglist, 1, + &vec, sizeof(*rsp), iu->sgt.sgl, 1, msg_open_conf, &errno, wait); if (err) { rnbd_clt_put_dev(dev); @@ -582,6 +585,7 @@ static int send_msg_open(struct rnbd_clt_dev *dev, bool wait) err = errno; } + sg_free_table(&iu->sgt); rnbd_put_iu(sess, iu); return err; } @@ -610,7 +614,8 @@ static int send_msg_sess_info(struct rnbd_clt_session *sess, bool wait) iu->buf = rsp; iu->sess = sess; - sg_init_one(iu->sglist, rsp, sizeof(*rsp)); + sg_alloc_table(&iu->sgt, 1, GFP_KERNEL); + sg_init_one(iu->sgt.sgl, rsp, sizeof(*rsp)); msg.hdr.type = cpu_to_le16(RNBD_MSG_SESS_INFO); msg.ver = RNBD_PROTO_VER_MAJOR; @@ -626,7 +631,7 @@ static int send_msg_sess_info(struct rnbd_clt_session *sess, bool wait) goto put_iu; } err = send_usr_msg(sess->rtrs, READ, iu, - &vec, sizeof(*rsp), iu->sglist, 1, + &vec, sizeof(*rsp), iu->sgt.sgl, 1, msg_sess_info_conf, &errno, wait); if (err) { rnbd_clt_put_sess(sess); @@ -636,7 +641,7 @@ put_iu: } else { err = errno; } - + sg_free_table(&iu->sgt); rnbd_put_iu(sess, iu); return err; } @@ -1016,11 +1021,10 @@ static int rnbd_client_xfer_request(struct rnbd_clt_dev *dev, * See queue limits. */ if (req_op(rq) != REQ_OP_DISCARD) - sg_cnt = blk_rq_map_sg(dev->queue, rq, iu->sglist); + sg_cnt = blk_rq_map_sg(dev->queue, rq, iu->sgt.sgl); if (sg_cnt == 0) - /* Do not forget to mark the end */ - sg_mark_end(&iu->sglist[0]); + sg_mark_end(&iu->sgt.sgl[0]); msg.hdr.type = cpu_to_le16(RNBD_MSG_IO); msg.device_id = cpu_to_le32(dev->device_id); @@ -1029,13 +1033,13 @@ static int rnbd_client_xfer_request(struct rnbd_clt_dev *dev, .iov_base = &msg, .iov_len = sizeof(msg) }; - size = rnbd_clt_get_sg_size(iu->sglist, sg_cnt); + size = rnbd_clt_get_sg_size(iu->sgt.sgl, sg_cnt); req_ops = (struct rtrs_clt_req_ops) { .priv = iu, .conf_fn = msg_io_conf, }; err = rtrs_clt_request(rq_data_dir(rq), &req_ops, rtrs, permit, - &vec, 1, size, iu->sglist, sg_cnt); + &vec, 1, size, iu->sgt.sgl, sg_cnt); if (unlikely(err)) { rnbd_clt_err_rl(dev, "RTRS failed to transfer IO, err: %d\n", err); @@ -1122,6 +1126,7 @@ static blk_status_t rnbd_queue_rq(struct blk_mq_hw_ctx *hctx, struct rnbd_clt_dev *dev = rq->rq_disk->private_data; struct rnbd_iu *iu = blk_mq_rq_to_pdu(rq); int err; + blk_status_t ret = BLK_STS_IOERR; if (unlikely(dev->dev_state != DEV_STATE_MAPPED)) return BLK_STS_IOERR; @@ -1133,32 +1138,35 @@ static blk_status_t rnbd_queue_rq(struct blk_mq_hw_ctx *hctx, return BLK_STS_RESOURCE; } + iu->sgt.sgl = iu->first_sgl; + err = sg_alloc_table_chained(&iu->sgt, + /* Even-if the request has no segment, + * sglist must have one entry at least */ + blk_rq_nr_phys_segments(rq) ? : 1, + iu->sgt.sgl, + RNBD_INLINE_SG_CNT); + if (err) { + rnbd_clt_err_rl(dev, "sg_alloc_table_chained ret=%d\n", err); + rnbd_clt_dev_kick_mq_queue(dev, hctx, 10/*ms*/); + rnbd_put_permit(dev->sess, iu->permit); + return BLK_STS_RESOURCE; + } + blk_mq_start_request(rq); err = rnbd_client_xfer_request(dev, rq, iu); if (likely(err == 0)) return BLK_STS_OK; if (unlikely(err == -EAGAIN || err == -ENOMEM)) { rnbd_clt_dev_kick_mq_queue(dev, hctx, 10/*ms*/); - rnbd_put_permit(dev->sess, iu->permit); - return BLK_STS_RESOURCE; + ret = BLK_STS_RESOURCE; } - + sg_free_table_chained(&iu->sgt, RNBD_INLINE_SG_CNT); rnbd_put_permit(dev->sess, iu->permit); - return BLK_STS_IOERR; -} - -static int rnbd_init_request(struct blk_mq_tag_set *set, struct request *rq, - unsigned int hctx_idx, unsigned int numa_node) -{ - struct rnbd_iu *iu = blk_mq_rq_to_pdu(rq); - - sg_init_table(iu->sglist, BMAX_SEGMENTS); - return 0; + return ret; } static struct blk_mq_ops rnbd_mq_ops = { .queue_rq = rnbd_queue_rq, - .init_request = rnbd_init_request, .complete = rnbd_softirq_done_fn, }; @@ -1172,7 +1180,7 @@ static int setup_mq_tags(struct rnbd_clt_session *sess) tag_set->numa_node = NUMA_NO_NODE; tag_set->flags = BLK_MQ_F_SHOULD_MERGE | BLK_MQ_F_TAG_QUEUE_SHARED; - tag_set->cmd_size = sizeof(struct rnbd_iu); + tag_set->cmd_size = sizeof(struct rnbd_iu) + RNBD_RDMA_SGL_SIZE; tag_set->nr_hw_queues = num_online_cpus(); return blk_mq_alloc_tag_set(tag_set); diff --git a/drivers/block/rnbd/rnbd-clt.h b/drivers/block/rnbd/rnbd-clt.h index efd67ae286ca..537d499dad3b 100644 --- a/drivers/block/rnbd/rnbd-clt.h +++ b/drivers/block/rnbd/rnbd-clt.h @@ -44,6 +44,13 @@ struct rnbd_iu_comp { int errno; }; +#ifdef CONFIG_ARCH_NO_SG_CHAIN +#define RNBD_INLINE_SG_CNT 0 +#else +#define RNBD_INLINE_SG_CNT 2 +#endif +#define RNBD_RDMA_SGL_SIZE (sizeof(struct scatterlist) * RNBD_INLINE_SG_CNT) + struct rnbd_iu { union { struct request *rq; /* for block io */ @@ -56,11 +63,12 @@ struct rnbd_iu { /* use to send msg associated with a sess */ struct rnbd_clt_session *sess; }; - struct scatterlist sglist[BMAX_SEGMENTS]; + struct sg_table sgt; struct work_struct work; int errno; struct rnbd_iu_comp comp; atomic_t refcount; + struct scatterlist first_sgl[]; /* must be the last one */ }; struct rnbd_cpu_qlist { -- cgit v1.2.3-59-g8ed1b From 9aaf9a2aba0c2b5f0fc6dfeb011f0b4c8e224a73 Mon Sep 17 00:00:00 2001 From: Gioh Kim Date: Thu, 10 Dec 2020 11:18:26 +0100 Subject: block/rnbd-clt: Does not request pdu to rtrs-clt Previously the rnbd client requested the rtrs to allocate rnbd_iu just after the rtrs_iu. So the rnbd client passes the size of rnbd_iu for rtrs_clt_open() and rtrs creates an array of rnbd_iu and rtrs_iu. For IO handling, rnbd_iu exists after the request because we pass the size of rnbd_iu when setting the tag-set. Therefore we do not use the rnbd_iu allocated by rtrs for IO handling. We only use the rnbd_iu allocated by rtrs when doing session initialization. Almost all rnbd_iu allocated by rtrs are wasted. By this patch the rnbd client does not request rnbd_iu allocation to rtrs but allocate it for itself when doing session initialization. Also remove unused rtrs_permit_to_pdu from rtrs. Signed-off-by: Gioh Kim Signed-off-by: Jack Wang Signed-off-by: Jens Axboe --- drivers/block/rnbd/rnbd-clt.c | 17 +++++++++++++---- drivers/infiniband/ulp/rtrs/rtrs-clt.c | 6 ------ drivers/infiniband/ulp/rtrs/rtrs.h | 7 ------- 3 files changed, 13 insertions(+), 17 deletions(-) (limited to 'drivers') diff --git a/drivers/block/rnbd/rnbd-clt.c b/drivers/block/rnbd/rnbd-clt.c index 5941ff7c83a8..96e3f9fe8241 100644 --- a/drivers/block/rnbd/rnbd-clt.c +++ b/drivers/block/rnbd/rnbd-clt.c @@ -349,12 +349,19 @@ static struct rnbd_iu *rnbd_get_iu(struct rnbd_clt_session *sess, struct rnbd_iu *iu; struct rtrs_permit *permit; + iu = kzalloc(sizeof(*iu), GFP_KERNEL); + if (!iu) { + return NULL; + } + permit = rnbd_get_permit(sess, con_type, wait ? RTRS_PERMIT_WAIT : RTRS_PERMIT_NOWAIT); - if (unlikely(!permit)) + if (unlikely(!permit)) { + kfree(iu); return NULL; - iu = rtrs_permit_to_pdu(permit); + } + iu->permit = permit; /* * 1st reference is dropped after finishing sending a "user" message, @@ -373,8 +380,10 @@ static struct rnbd_iu *rnbd_get_iu(struct rnbd_clt_session *sess, static void rnbd_put_iu(struct rnbd_clt_session *sess, struct rnbd_iu *iu) { - if (atomic_dec_and_test(&iu->refcount)) + if (atomic_dec_and_test(&iu->refcount)) { rnbd_put_permit(sess, iu->permit); + kfree(iu); + } } static void rnbd_softirq_done_fn(struct request *rq) @@ -1218,7 +1227,7 @@ find_and_get_or_create_sess(const char *sessname, */ sess->rtrs = rtrs_clt_open(&rtrs_ops, sessname, paths, path_cnt, port_nr, - sizeof(struct rnbd_iu), + 0, /* Do not use pdu of rtrs */ RECONNECT_DELAY, BMAX_SEGMENTS, BLK_MAX_SEGMENT_SIZE, MAX_RECONNECTS); diff --git a/drivers/infiniband/ulp/rtrs/rtrs-clt.c b/drivers/infiniband/ulp/rtrs/rtrs-clt.c index 560865f65dc4..67f86c405a26 100644 --- a/drivers/infiniband/ulp/rtrs/rtrs-clt.c +++ b/drivers/infiniband/ulp/rtrs/rtrs-clt.c @@ -157,12 +157,6 @@ void rtrs_clt_put_permit(struct rtrs_clt *clt, struct rtrs_permit *permit) } EXPORT_SYMBOL(rtrs_clt_put_permit); -void *rtrs_permit_to_pdu(struct rtrs_permit *permit) -{ - return permit + 1; -} -EXPORT_SYMBOL(rtrs_permit_to_pdu); - /** * rtrs_permit_to_clt_con() - returns RDMA connection pointer by the permit * @sess: client session pointer diff --git a/drivers/infiniband/ulp/rtrs/rtrs.h b/drivers/infiniband/ulp/rtrs/rtrs.h index 9af750f4d783..8738e90e715a 100644 --- a/drivers/infiniband/ulp/rtrs/rtrs.h +++ b/drivers/infiniband/ulp/rtrs/rtrs.h @@ -63,13 +63,6 @@ struct rtrs_clt *rtrs_clt_open(struct rtrs_clt_ops *ops, void rtrs_clt_close(struct rtrs_clt *sess); -/** - * rtrs_permit_to_pdu() - converts rtrs_permit to opaque pdu pointer - * @permit: RTRS permit pointer, it associates the memory allocation for future - * RDMA operation. - */ -void *rtrs_permit_to_pdu(struct rtrs_permit *permit); - enum { RTRS_PERMIT_NOWAIT = 0, RTRS_PERMIT_WAIT = 1, -- cgit v1.2.3-59-g8ed1b From 1aba169e770911fb2afa63eb859883c4de2191e3 Mon Sep 17 00:00:00 2001 From: Josh Triplett Date: Thu, 17 Dec 2020 00:58:47 -0800 Subject: nbd: Respect max_part for all partition scans The creation path of the NBD device respects max_part and only scans for partitions if max_part is not 0. However, some other code paths ignore max_part, and unconditionally scan for partitions. Add a check for max_part on each partition scan. Signed-off-by: Josh Triplett Reviewed-by: Josef Bacik Signed-off-by: Jens Axboe --- drivers/block/nbd.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) (limited to 'drivers') diff --git a/drivers/block/nbd.c b/drivers/block/nbd.c index 92f84ed0ba9e..6727358e147d 100644 --- a/drivers/block/nbd.c +++ b/drivers/block/nbd.c @@ -318,7 +318,8 @@ static int nbd_set_size(struct nbd_device *nbd, loff_t bytesize, blk_queue_logical_block_size(nbd->disk->queue, blksize); blk_queue_physical_block_size(nbd->disk->queue, blksize); - set_bit(GD_NEED_PART_SCAN, &nbd->disk->state); + if (max_part) + set_bit(GD_NEED_PART_SCAN, &nbd->disk->state); if (!set_capacity_and_notify(nbd->disk, bytesize >> 9)) kobject_uevent(&nbd_to_dev(nbd)->kobj, KOBJ_CHANGE); return 0; @@ -1476,9 +1477,11 @@ static int nbd_open(struct block_device *bdev, fmode_t mode) refcount_set(&nbd->config_refs, 1); refcount_inc(&nbd->refs); mutex_unlock(&nbd->config_lock); - set_bit(GD_NEED_PART_SCAN, &bdev->bd_disk->state); + if (max_part) + set_bit(GD_NEED_PART_SCAN, &bdev->bd_disk->state); } else if (nbd_disconnected(nbd->config)) { - set_bit(GD_NEED_PART_SCAN, &bdev->bd_disk->state); + if (max_part) + set_bit(GD_NEED_PART_SCAN, &bdev->bd_disk->state); } out: mutex_unlock(&nbd_index_mutex); -- cgit v1.2.3-59-g8ed1b From 658a337a606f48b7ebe451591f7681d383fa115e Mon Sep 17 00:00:00 2001 From: Stefan Haberland Date: Thu, 17 Dec 2020 16:59:04 +0100 Subject: s390/dasd: fix hanging device offline processing For an LCU update a read unit address configuration IO is required. This is started using sleep_on(), which has early exit paths in case the device is not usable for IO. For example when it is in offline processing. In those cases the LCU update should fail and not be retried. Therefore lcu_update_work checks if EOPNOTSUPP is returned or not. Commit 41995342b40c ("s390/dasd: fix endless loop after read unit address configuration") accidentally removed the EOPNOTSUPP return code from read_unit_address_configuration(), which in turn might lead to an endless loop of the LCU update in offline processing. Fix by returning EOPNOTSUPP again if the device is not able to perform the request. Fixes: 41995342b40c ("s390/dasd: fix endless loop after read unit address configuration") Cc: stable@vger.kernel.org #5.3 Signed-off-by: Stefan Haberland Reviewed-by: Jan Hoeppner Signed-off-by: Jens Axboe --- drivers/s390/block/dasd_alias.c | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) (limited to 'drivers') diff --git a/drivers/s390/block/dasd_alias.c b/drivers/s390/block/dasd_alias.c index 99f86612f775..31e8b5d48e86 100644 --- a/drivers/s390/block/dasd_alias.c +++ b/drivers/s390/block/dasd_alias.c @@ -462,11 +462,19 @@ static int read_unit_address_configuration(struct dasd_device *device, spin_unlock_irqrestore(&lcu->lock, flags); rc = dasd_sleep_on(cqr); - if (rc && !suborder_not_supported(cqr)) { + if (!rc) + goto out; + + if (suborder_not_supported(cqr)) { + /* suborder not supported or device unusable for IO */ + rc = -EOPNOTSUPP; + } else { + /* IO failed but should be retried */ spin_lock_irqsave(&lcu->lock, flags); lcu->flags |= NEED_UAC_UPDATE; spin_unlock_irqrestore(&lcu->lock, flags); } +out: dasd_sfree_request(cqr, cqr->memdev); return rc; } -- cgit v1.2.3-59-g8ed1b From a29ea01653493b94ea12bb2b89d1564a265081b6 Mon Sep 17 00:00:00 2001 From: Stefan Haberland Date: Thu, 17 Dec 2020 16:59:05 +0100 Subject: s390/dasd: prevent inconsistent LCU device data Prevent _lcu_update from adding a device to a pavgroup if the LCU still requires an update. The data is not reliable any longer and in parallel devices might have been moved on the lists already. This might lead to list corruptions or invalid PAV grouping. Only add devices to a pavgroup if the LCU is up to date. Additional steps are taken by the scheduled lcu update. Fixes: 8e09f21574ea ("[S390] dasd: add hyper PAV support to DASD device driver, part 1") Cc: stable@vger.kernel.org Signed-off-by: Stefan Haberland Reviewed-by: Jan Hoeppner Signed-off-by: Jens Axboe --- drivers/s390/block/dasd_alias.c | 9 +++++++++ 1 file changed, 9 insertions(+) (limited to 'drivers') diff --git a/drivers/s390/block/dasd_alias.c b/drivers/s390/block/dasd_alias.c index 31e8b5d48e86..f841518de6c5 100644 --- a/drivers/s390/block/dasd_alias.c +++ b/drivers/s390/block/dasd_alias.c @@ -511,6 +511,14 @@ static int _lcu_update(struct dasd_device *refdev, struct alias_lcu *lcu) return rc; spin_lock_irqsave(&lcu->lock, flags); + /* + * there is another update needed skip the remaining handling + * the data might already be outdated + * but especially do not add the device to an LCU with pending + * update + */ + if (lcu->flags & NEED_UAC_UPDATE) + goto out; lcu->pav = NO_PAV; for (i = 0; i < MAX_DEVICES_PER_LCU; ++i) { switch (lcu->uac->unit[i].ua_type) { @@ -529,6 +537,7 @@ static int _lcu_update(struct dasd_device *refdev, struct alias_lcu *lcu) alias_list) { _add_device_to_lcu(lcu, device, refdev); } +out: spin_unlock_irqrestore(&lcu->lock, flags); return 0; } -- cgit v1.2.3-59-g8ed1b From 0ede91f83aa335da1c3ec68eb0f9e228f269f6d8 Mon Sep 17 00:00:00 2001 From: Stefan Haberland Date: Thu, 17 Dec 2020 16:59:06 +0100 Subject: s390/dasd: fix list corruption of pavgroup group list dasd_alias_add_device() moves devices to the active_devices list in case of a scheduled LCU update regardless if they have previously been in a pavgroup or not. Example: device A and B are in the same pavgroup. Device A has already been in a pavgroup and the private->pavgroup pointer is set and points to a valid pavgroup. While going through dasd_add_device it is moved from the pavgroup to the active_devices list. In parallel device B might be removed from the same pavgroup in remove_device_from_lcu() which in turn checks if the group is empty and deletes it accordingly because device A has already been removed from there. When now device A enters remove_device_from_lcu() it is tried to remove it from the pavgroup again because the pavgroup pointer is still set and again the empty group will be cleaned up which leads to a list corruption. Fix by setting private->pavgroup to NULL in dasd_add_device. If the device has been the last device on the pavgroup an empty pavgroup remains but this will be cleaned up by the scheduled lcu_update which iterates over all existing pavgroups. Fixes: 8e09f21574ea ("[S390] dasd: add hyper PAV support to DASD device driver, part 1") Cc: stable@vger.kernel.org Signed-off-by: Stefan Haberland Reviewed-by: Jan Hoeppner Signed-off-by: Jens Axboe --- drivers/s390/block/dasd_alias.c | 1 + 1 file changed, 1 insertion(+) (limited to 'drivers') diff --git a/drivers/s390/block/dasd_alias.c b/drivers/s390/block/dasd_alias.c index f841518de6c5..60344e13e87b 100644 --- a/drivers/s390/block/dasd_alias.c +++ b/drivers/s390/block/dasd_alias.c @@ -642,6 +642,7 @@ int dasd_alias_add_device(struct dasd_device *device) } if (lcu->flags & UPDATE_PENDING) { list_move(&device->alias_list, &lcu->active_devices); + private->pavgroup = NULL; _schedule_lcu_update(lcu, device); } spin_unlock_irqrestore(&lcu->lock, flags); -- cgit v1.2.3-59-g8ed1b From 53a7f655834c7c335bf683f248208d4fbe4b47bc Mon Sep 17 00:00:00 2001 From: Stefan Haberland Date: Thu, 17 Dec 2020 16:59:07 +0100 Subject: s390/dasd: fix list corruption of lcu list In dasd_alias_disconnect_device_from_lcu the device is removed from any list on the LCU. Afterwards the LCU is removed from the lcu list if it does not contain devices any longer. The lcu->lock protects the lcu from parallel updates. But to cancel all workers and wait for completion the lcu->lock has to be unlocked. If two devices are removed in parallel and both are removed from the LCU the first device that takes the lcu->lock again will delete the LCU because it is already empty but the second device also tries to free the LCU which leads to a list corruption of the lcu list. Fix by removing the device right before the lcu is checked without unlocking the lcu->lock in between. Fixes: 8e09f21574ea ("[S390] dasd: add hyper PAV support to DASD device driver, part 1") Cc: stable@vger.kernel.org Signed-off-by: Stefan Haberland Reviewed-by: Jan Hoeppner Signed-off-by: Jens Axboe --- drivers/s390/block/dasd_alias.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers') diff --git a/drivers/s390/block/dasd_alias.c b/drivers/s390/block/dasd_alias.c index 60344e13e87b..dc78a523a69f 100644 --- a/drivers/s390/block/dasd_alias.c +++ b/drivers/s390/block/dasd_alias.c @@ -256,7 +256,6 @@ void dasd_alias_disconnect_device_from_lcu(struct dasd_device *device) return; device->discipline->get_uid(device, &uid); spin_lock_irqsave(&lcu->lock, flags); - list_del_init(&device->alias_list); /* make sure that the workers don't use this device */ if (device == lcu->suc_data.device) { spin_unlock_irqrestore(&lcu->lock, flags); @@ -283,6 +282,7 @@ void dasd_alias_disconnect_device_from_lcu(struct dasd_device *device) spin_lock_irqsave(&aliastree.lock, flags); spin_lock(&lcu->lock); + list_del_init(&device->alias_list); if (list_empty(&lcu->grouplist) && list_empty(&lcu->active_devices) && list_empty(&lcu->inactive_devices)) { -- cgit v1.2.3-59-g8ed1b From 117ae250cfa3718f21bd07df0650dfbe3bc3a823 Mon Sep 17 00:00:00 2001 From: Yi Li Date: Wed, 23 Dec 2020 23:04:21 +0800 Subject: bcache:remove a superfluous check in register_bcache There have no reassign the bdev after check It is IS_ERR. the double check !IS_ERR(bdev) is superfluous. After commit 4e7b5671c6a8 ("block: remove i_bdev"), "Switch the block device lookup interfaces to directly work with a dev_t so that struct block_device references are only acquired by the blkdev_get variants (and the blk-cgroup special case). This means that we now don't need an extra reference in the inode and can generally simplify handling of struct block_device to keep the lookups contained in the core block layer code." so after lookup_bdev call, there no need to do bdput. remove a superfluous check the bdev & don't call bdput after lookup_bdev. Fixes: 4e7b5671c6a8("block: remove i_bdev") Signed-off-by: Yi Li Reviewed-by: Christoph Hellwig Signed-off-by: Coly Li Signed-off-by: Jens Axboe --- drivers/md/bcache/super.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'drivers') diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c index 0e06d721cd8e..a4752ac410dc 100644 --- a/drivers/md/bcache/super.c +++ b/drivers/md/bcache/super.c @@ -2535,8 +2535,6 @@ static ssize_t register_bcache(struct kobject *k, struct kobj_attribute *attr, else err = "device busy"; mutex_unlock(&bch_register_lock); - if (!IS_ERR(bdev)) - bdput(bdev); if (attr == &ksysfs_register_quiet) goto done; } -- cgit v1.2.3-59-g8ed1b From 46926127d76359b46659c556df7b4aa1b6325d90 Mon Sep 17 00:00:00 2001 From: Zheng Yongjun Date: Wed, 23 Dec 2020 23:04:22 +0800 Subject: md/bcache: convert comma to semicolon Replace a comma between expression statements by a semicolon. Signed-off-by: Zheng Yongjun Signed-off-by: Coly Li Signed-off-by: Jens Axboe --- drivers/md/bcache/sysfs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers') diff --git a/drivers/md/bcache/sysfs.c b/drivers/md/bcache/sysfs.c index 554e3afc9b68..00a520c03f41 100644 --- a/drivers/md/bcache/sysfs.c +++ b/drivers/md/bcache/sysfs.c @@ -404,7 +404,7 @@ STORE(__cached_dev) if (!env) return -ENOMEM; add_uevent_var(env, "DRIVER=bcache"); - add_uevent_var(env, "CACHED_UUID=%pU", dc->sb.uuid), + add_uevent_var(env, "CACHED_UUID=%pU", dc->sb.uuid); add_uevent_var(env, "CACHED_LABEL=%s", buf); kobject_uevent_env(&disk_to_dev(dc->disk.disk)->kobj, KOBJ_CHANGE, -- cgit v1.2.3-59-g8ed1b