diff --git a/drivers/md/md.c b/drivers/md/md.c index ac85ec73a409..772cffe02ff5 100644 --- a/drivers/md/md.c +++ b/drivers/md/md.c @@ -339,6 +339,7 @@ static int start_readonly; * so all the races disappear. */ static bool create_on_open = true; +static bool legacy_async_del_gendisk = true; /* * We have a system wide 'event count' that is incremented @@ -877,15 +878,18 @@ void mddev_unlock(struct mddev *mddev) export_rdev(rdev, mddev); } - /* Call del_gendisk after release reconfig_mutex to avoid - * deadlock (e.g. call del_gendisk under the lock and an - * access to sysfs files waits the lock) - * And MD_DELETED is only used for md raid which is set in - * do_md_stop. dm raid only uses md_stop to stop. So dm raid - * doesn't need to check MD_DELETED when getting reconfig lock - */ - if (test_bit(MD_DELETED, &mddev->flags)) - del_gendisk(mddev->gendisk); + if (!legacy_async_del_gendisk) { + /* + * Call del_gendisk after release reconfig_mutex to avoid + * deadlock (e.g. call del_gendisk under the lock and an + * access to sysfs files waits the lock) + * And MD_DELETED is only used for md raid which is set in + * do_md_stop. dm raid only uses md_stop to stop. So dm raid + * doesn't need to check MD_DELETED when getting reconfig lock + */ + if (test_bit(MD_DELETED, &mddev->flags)) + del_gendisk(mddev->gendisk); + } } EXPORT_SYMBOL_GPL(mddev_unlock); @@ -5818,6 +5822,13 @@ static void md_kobj_release(struct kobject *ko) { struct mddev *mddev = container_of(ko, struct mddev, kobj); + if (legacy_async_del_gendisk) { + if (mddev->sysfs_state) + sysfs_put(mddev->sysfs_state); + if (mddev->sysfs_level) + sysfs_put(mddev->sysfs_level); + del_gendisk(mddev->gendisk); + } put_disk(mddev->gendisk); } @@ -6021,6 +6032,9 @@ static int md_alloc_and_put(dev_t dev, char *name) { struct mddev *mddev = md_alloc(dev, name); + if (legacy_async_del_gendisk) + pr_warn("md: async del_gendisk mode will be removed in future, please upgrade to mdadm-4.5+\n"); + if (IS_ERR(mddev)) return PTR_ERR(mddev); mddev_put(mddev); @@ -6431,10 +6445,22 @@ static void md_clean(struct mddev *mddev) mddev->persistent = 0; mddev->level = LEVEL_NONE; mddev->clevel[0] = 0; - /* if UNTIL_STOP is set, it's cleared here */ - mddev->hold_active = 0; - /* Don't clear MD_CLOSING, or mddev can be opened again. */ - mddev->flags &= BIT_ULL_MASK(MD_CLOSING); + + /* + * For legacy_async_del_gendisk mode, it can stop the array in the + * middle of assembling it, then it still can access the array. So + * it needs to clear MD_CLOSING. If not legacy_async_del_gendisk, + * it can't open the array again after stopping it. So it doesn't + * clear MD_CLOSING. + */ + if (legacy_async_del_gendisk && mddev->hold_active) { + clear_bit(MD_CLOSING, &mddev->flags); + } else { + /* if UNTIL_STOP is set, it's cleared here */ + mddev->hold_active = 0; + /* Don't clear MD_CLOSING, or mddev can be opened again. */ + mddev->flags &= BIT_ULL_MASK(MD_CLOSING); + } mddev->sb_flags = 0; mddev->ro = MD_RDWR; mddev->metadata_type[0] = 0; @@ -6658,7 +6684,8 @@ static int do_md_stop(struct mddev *mddev, int mode) export_array(mddev); md_clean(mddev); - set_bit(MD_DELETED, &mddev->flags); + if (!legacy_async_del_gendisk) + set_bit(MD_DELETED, &mddev->flags); } md_new_event(); sysfs_notify_dirent_safe(mddev->sysfs_state); @@ -10392,6 +10419,7 @@ module_param_call(start_ro, set_ro, get_ro, NULL, S_IRUSR|S_IWUSR); module_param(start_dirty_degraded, int, S_IRUGO|S_IWUSR); module_param_call(new_array, add_named_array, NULL, NULL, S_IWUSR); module_param(create_on_open, bool, S_IRUSR|S_IWUSR); +module_param(legacy_async_del_gendisk, bool, 0600); MODULE_LICENSE("GPL"); MODULE_DESCRIPTION("MD RAID framework"); -- knotmirror 0.0.1 From c27973211ffcdf0a092eec265d5993e64b89adaf Mon Sep 17 00:00:00 2001 From: Xiao Ni Date: Fri, 15 Aug 2025 12:00:28 +0800 Subject: [PATCH 02/14] md: keep recovery_cp in mdp_superblock_s commit 907a99c314a5 ("md: rename recovery_cp to resync_offset") replaces recovery_cp with resync_offset in mdp_superblock_s which is in md_p.h. md_p.h is used in userspace too. So mdadm building fails because of this. This patch revert this change. Fixes: 907a99c314a5 ("md: rename recovery_cp to resync_offset") Signed-off-by: Xiao Ni Link: https://lore.kernel.org/linux-raid/20250815040028.18085-1-xni@redhat.com Signed-off-by: Yu Kuai --- drivers/md/md.c | 6 +++--- include/uapi/linux/raid/md_p.h | 2 +- 2 files changed, 4 insertions(+), 4 deletions(-) diff --git a/drivers/md/md.c b/drivers/md/md.c index 772cffe02ff5..3836fc7eff67 100644 --- a/drivers/md/md.c +++ b/drivers/md/md.c @@ -1423,7 +1423,7 @@ static int super_90_validate(struct mddev *mddev, struct md_rdev *freshest, stru else { if (sb->events_hi == sb->cp_events_hi && sb->events_lo == sb->cp_events_lo) { - mddev->resync_offset = sb->resync_offset; + mddev->resync_offset = sb->recovery_cp; } else mddev->resync_offset = 0; } @@ -1551,13 +1551,13 @@ static void super_90_sync(struct mddev *mddev, struct md_rdev *rdev) mddev->minor_version = sb->minor_version; if (mddev->in_sync) { - sb->resync_offset = mddev->resync_offset; + sb->recovery_cp = mddev->resync_offset; sb->cp_events_hi = (mddev->events>>32); sb->cp_events_lo = (u32)mddev->events; if (mddev->resync_offset == MaxSector) sb->state = (1<< MD_SB_CLEAN); } else - sb->resync_offset = 0; + sb->recovery_cp = 0; sb->layout = mddev->layout; sb->chunk_size = mddev->chunk_sectors << 9; diff --git a/include/uapi/linux/raid/md_p.h b/include/uapi/linux/raid/md_p.h index b13946287277..ac74133a4768 100644 --- a/include/uapi/linux/raid/md_p.h +++ b/include/uapi/linux/raid/md_p.h @@ -173,7 +173,7 @@ typedef struct mdp_superblock_s { #else #error unspecified endianness #endif - __u32 resync_offset; /* 11 resync checkpoint sector count */ + __u32 recovery_cp; /* 11 resync checkpoint sector count */ /* There are only valid for minor_version > 90 */ __u64 reshape_position; /* 12,13 next address in array-space for reshape */ __u32 new_level; /* 14 new level we are reshaping to */ -- knotmirror 0.0.1 From cb0780ad4333040a98e10f014b593ef738a3f31e Mon Sep 17 00:00:00 2001 From: Zheng Qixing Date: Sat, 16 Aug 2025 08:25:33 +0800 Subject: [PATCH 03/14] md: add helper rdev_needs_recovery() Add a helper for checking if an rdev needs recovery. Signed-off-by: Zheng Qixing Link: https://lore.kernel.org/linux-raid/20250816002534.1754356-2-zhengqixing@huaweicloud.com Signed-off-by: Yu Kuai --- drivers/md/md.c | 23 ++++++++++++----------- 1 file changed, 12 insertions(+), 11 deletions(-) diff --git a/drivers/md/md.c b/drivers/md/md.c index 3836fc7eff67..abd327ade4bd 100644 --- a/drivers/md/md.c +++ b/drivers/md/md.c @@ -4839,6 +4839,15 @@ out_unlock: static struct md_sysfs_entry md_metadata = __ATTR_PREALLOC(metadata_version, S_IRUGO|S_IWUSR, metadata_show, metadata_store); +static bool rdev_needs_recovery(struct md_rdev *rdev, sector_t sectors) +{ + return rdev->raid_disk >= 0 && + !test_bit(Journal, &rdev->flags) && + !test_bit(Faulty, &rdev->flags) && + !test_bit(In_sync, &rdev->flags) && + rdev->recovery_offset < sectors; +} + enum sync_action md_sync_action(struct mddev *mddev) { unsigned long recovery = mddev->recovery; @@ -8995,11 +9004,7 @@ static sector_t md_sync_position(struct mddev *mddev, enum sync_action action) start = MaxSector; rcu_read_lock(); rdev_for_each_rcu(rdev, mddev) - if (rdev->raid_disk >= 0 && - !test_bit(Journal, &rdev->flags) && - !test_bit(Faulty, &rdev->flags) && - !test_bit(In_sync, &rdev->flags) && - rdev->recovery_offset < start) + if (rdev_needs_recovery(rdev, start)) start = rdev->recovery_offset; rcu_read_unlock(); @@ -9358,12 +9363,8 @@ void md_do_sync(struct md_thread *thread) test_bit(MD_RECOVERY_RECOVER, &mddev->recovery)) { rcu_read_lock(); rdev_for_each_rcu(rdev, mddev) - if (rdev->raid_disk >= 0 && - mddev->delta_disks >= 0 && - !test_bit(Journal, &rdev->flags) && - !test_bit(Faulty, &rdev->flags) && - !test_bit(In_sync, &rdev->flags) && - rdev->recovery_offset < mddev->curr_resync) + if (mddev->delta_disks >= 0 && + rdev_needs_recovery(rdev, mddev->curr_resync)) rdev->recovery_offset = mddev->curr_resync; rcu_read_unlock(); } -- knotmirror 0.0.1 From b7ee30f0efd12f42735ae233071015389407966c Mon Sep 17 00:00:00 2001 From: Zheng Qixing Date: Sat, 16 Aug 2025 08:25:34 +0800 Subject: [PATCH 04/14] md: fix sync_action incorrect display during resync During raid resync, if a disk becomes faulty, the operation is briefly interrupted. The MD_RECOVERY_RECOVER flag triggered by the disk failure causes sync_action to incorrectly show "recover" instead of "resync". The same issue affects reshape operations. Reproduction steps: mdadm -Cv /dev/md1 -l1 -n4 -e1.2 /dev/sd{a..d} // -> resync happened mdadm -f /dev/md1 /dev/sda // -> resync interrupted cat sync_action -> recover Add progress checks in md_sync_action() for resync/recover/reshape to ensure the interface correctly reports the actual operation type. Fixes: 4b10a3bc67c1 ("md: ensure resync is prioritized over recovery") Signed-off-by: Zheng Qixing Link: https://lore.kernel.org/linux-raid/20250816002534.1754356-3-zhengqixing@huaweicloud.com Signed-off-by: Yu Kuai --- drivers/md/md.c | 37 +++++++++++++++++++++++++++++++++++-- 1 file changed, 35 insertions(+), 2 deletions(-) diff --git a/drivers/md/md.c b/drivers/md/md.c index abd327ade4bd..1baaf52c603c 100644 --- a/drivers/md/md.c +++ b/drivers/md/md.c @@ -4848,9 +4848,33 @@ static bool rdev_needs_recovery(struct md_rdev *rdev, sector_t sectors) rdev->recovery_offset < sectors; } +static enum sync_action md_get_active_sync_action(struct mddev *mddev) +{ + struct md_rdev *rdev; + bool is_recover = false; + + if (mddev->resync_offset < MaxSector) + return ACTION_RESYNC; + + if (mddev->reshape_position != MaxSector) + return ACTION_RESHAPE; + + rcu_read_lock(); + rdev_for_each_rcu(rdev, mddev) { + if (rdev_needs_recovery(rdev, MaxSector)) { + is_recover = true; + break; + } + } + rcu_read_unlock(); + + return is_recover ? ACTION_RECOVER : ACTION_IDLE; +} + enum sync_action md_sync_action(struct mddev *mddev) { unsigned long recovery = mddev->recovery; + enum sync_action active_action; /* * frozen has the highest priority, means running sync_thread will be @@ -4874,8 +4898,17 @@ enum sync_action md_sync_action(struct mddev *mddev) !test_bit(MD_RECOVERY_NEEDED, &recovery)) return ACTION_IDLE; - if (test_bit(MD_RECOVERY_RESHAPE, &recovery) || - mddev->reshape_position != MaxSector) + /* + * Check if any sync operation (resync/recover/reshape) is + * currently active. This ensures that only one sync operation + * can run at a time. Returns the type of active operation, or + * ACTION_IDLE if none are active. + */ + active_action = md_get_active_sync_action(mddev); + if (active_action != ACTION_IDLE) + return active_action; + + if (test_bit(MD_RECOVERY_RESHAPE, &recovery)) return ACTION_RESHAPE; if (test_bit(MD_RECOVERY_RECOVER, &recovery)) -- knotmirror 0.0.1 From 0227af355b50c526bf83ca52d67aef5d102e9b07 Mon Sep 17 00:00:00 2001 From: Akhilesh Patil Date: Sun, 17 Aug 2025 15:06:05 +0530 Subject: [PATCH 05/14] selftests: ublk: Use ARRAY_SIZE() macro to improve code Use ARRAY_SIZE() macro while calculating size of an array to improve code readability and reduce potential sizing errors. Implement this suggestion given by spatch tool by running coccinelle script - scripts/coccinelle/misc/array_size.cocci Follow ARRAY_SIZE() macro usage pattern in ublk.c introduced by, commit ec120093180b9 ("selftests: ublk: fix ublk_find_tgt()") wherever appropriate to maintain consistency. Signed-off-by: Akhilesh Patil Reviewed-by: Ming Lei Link: https://lore.kernel.org/r/aKGihYui6/Pcijbk@bhairav-test.ee.iitb.ac.in Signed-off-by: Jens Axboe --- tools/testing/selftests/ublk/kublk.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/tools/testing/selftests/ublk/kublk.c b/tools/testing/selftests/ublk/kublk.c index 95188065b2e9..6512dfbdbce3 100644 --- a/tools/testing/selftests/ublk/kublk.c +++ b/tools/testing/selftests/ublk/kublk.c @@ -1400,7 +1400,7 @@ static int cmd_dev_get_features(void) if (!((1ULL << i) & features)) continue; - if (i < sizeof(feat_map) / sizeof(feat_map[0])) + if (i < ARRAY_SIZE(feat_map)) feat = feat_map[i]; else feat = "unknown"; @@ -1477,7 +1477,7 @@ static void __cmd_create_help(char *exe, bool recovery) printf("\tdefault: nr_queues=2(max 32), depth=128(max 1024), dev_id=-1(auto allocation)\n"); printf("\tdefault: nthreads=nr_queues"); - for (i = 0; i < sizeof(tgt_ops_list) / sizeof(tgt_ops_list[0]); i++) { + for (i = 0; i < ARRAY_SIZE(tgt_ops_list); i++) { const struct ublk_tgt_ops *ops = tgt_ops_list[i]; if (ops->usage) -- knotmirror 0.0.1 From 61ca3b891b4b9667334c1356a73f28954c92d43a Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 18 Aug 2025 06:54:50 +0200 Subject: [PATCH 06/14] block: handle pi_tuple_size in queue_limits_stack_integrity queue_limits_stack_integrity needs to handle the new pi_tuple_size field, otherwise stacking PI-capable devices will always fail. Fixes: 76e45252a4ce ("block: introduce pi_tuple_size field in blk_integrity") Signed-off-by: Christoph Hellwig Reviewed-by: Anuj Gupta Reviewed-by: Martin K. Petersen Link: https://lore.kernel.org/r/20250818045456.1482889-2-hch@lst.de Signed-off-by: Jens Axboe --- block/blk-settings.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/block/blk-settings.c b/block/blk-settings.c index 07874e9b609f..491c0c48d52b 100644 --- a/block/blk-settings.c +++ b/block/blk-settings.c @@ -972,6 +972,8 @@ bool queue_limits_stack_integrity(struct queue_limits *t, goto incompatible; if (ti->csum_type != bi->csum_type) goto incompatible; + if (ti->pi_tuple_size != bi->pi_tuple_size) + goto incompatible; if ((ti->flags & BLK_INTEGRITY_REF_TAG) != (bi->flags & BLK_INTEGRITY_REF_TAG)) goto incompatible; @@ -980,6 +982,7 @@ bool queue_limits_stack_integrity(struct queue_limits *t, ti->flags |= (bi->flags & BLK_INTEGRITY_DEVICE_CAPABLE) | (bi->flags & BLK_INTEGRITY_REF_TAG); ti->csum_type = bi->csum_type; + ti->pi_tuple_size = bi->pi_tuple_size; ti->metadata_size = bi->metadata_size; ti->pi_offset = bi->pi_offset; ti->interval_exp = bi->interval_exp; -- knotmirror 0.0.1 From f4ae1744033d54b63c31a3664a4fdf5cebec7f27 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 18 Aug 2025 06:54:51 +0200 Subject: [PATCH 07/14] block: remove newlines from the warnings in blk_validate_integrity_limits Otherwise they are very hard to read in the kernel log. Signed-off-by: Christoph Hellwig Reviewed-by: Anuj Gupta Reviewed-by: Martin K. Petersen Link: https://lore.kernel.org/r/20250818045456.1482889-3-hch@lst.de Signed-off-by: Jens Axboe --- block/blk-settings.c | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-) diff --git a/block/blk-settings.c b/block/blk-settings.c index 491c0c48d52b..d6438e6c276d 100644 --- a/block/blk-settings.c +++ b/block/blk-settings.c @@ -157,16 +157,14 @@ static int blk_validate_integrity_limits(struct queue_limits *lim) switch (bi->csum_type) { case BLK_INTEGRITY_CSUM_NONE: if (bi->pi_tuple_size) { - pr_warn("pi_tuple_size must be 0 when checksum type \ - is none\n"); + pr_warn("pi_tuple_size must be 0 when checksum type is none\n"); return -EINVAL; } break; case BLK_INTEGRITY_CSUM_CRC: case BLK_INTEGRITY_CSUM_IP: if (bi->pi_tuple_size != sizeof(struct t10_pi_tuple)) { - pr_warn("pi_tuple_size mismatch for T10 PI: expected \ - %zu, got %u\n", + pr_warn("pi_tuple_size mismatch for T10 PI: expected %zu, got %u\n", sizeof(struct t10_pi_tuple), bi->pi_tuple_size); return -EINVAL; @@ -174,8 +172,7 @@ static int blk_validate_integrity_limits(struct queue_limits *lim) break; case BLK_INTEGRITY_CSUM_CRC64: if (bi->pi_tuple_size != sizeof(struct crc64_pi_tuple)) { - pr_warn("pi_tuple_size mismatch for CRC64 PI: \ - expected %zu, got %u\n", + pr_warn("pi_tuple_size mismatch for CRC64 PI: expected %zu, got %u\n", sizeof(struct crc64_pi_tuple), bi->pi_tuple_size); return -EINVAL; -- knotmirror 0.0.1 From 8aa5a3b68ad144da49a3d17f165e6561255e3529 Mon Sep 17 00:00:00 2001 From: Rajeev Mishra Date: Mon, 18 Aug 2025 18:48:20 +0000 Subject: [PATCH 08/14] loop: Consolidate size calculation logic into lo_calculate_size() Renamed get_size to lo_calculate_size and merged the logic from get_size and get_loop_size into a single function. Update all callers to use lo_calculate_size. This is done in preparation for improving the size detection logic. Signed-off-by: Rajeev Mishra Reviewed-by: Yu Kuai Link: https://lore.kernel.org/r/20250818184821.115033-2-rajeevm@hpe.com [axboe: massage commit message] Signed-off-by: Jens Axboe --- drivers/block/loop.c | 26 +++++++++----------------- 1 file changed, 9 insertions(+), 17 deletions(-) diff --git a/drivers/block/loop.c b/drivers/block/loop.c index 1b6ee91f8eb9..0e1b9eb9db10 100644 --- a/drivers/block/loop.c +++ b/drivers/block/loop.c @@ -137,20 +137,18 @@ static void loop_global_unlock(struct loop_device *lo, bool global) static int max_part; static int part_shift; -static loff_t get_size(loff_t offset, loff_t sizelimit, struct file *file) +static loff_t lo_calculate_size(struct loop_device *lo, struct file *file) { loff_t loopsize; - /* Compute loopsize in bytes */ loopsize = i_size_read(file->f_mapping->host); - if (offset > 0) - loopsize -= offset; + if (lo->lo_offset > 0) + loopsize -= lo->lo_offset; /* offset is beyond i_size, weird but possible */ if (loopsize < 0) return 0; - - if (sizelimit > 0 && sizelimit < loopsize) - loopsize = sizelimit; + if (lo->lo_sizelimit > 0 && lo->lo_sizelimit < loopsize) + loopsize = lo->lo_sizelimit; /* * Unfortunately, if we want to do I/O on the device, * the number of 512-byte sectors has to fit into a sector_t. @@ -158,11 +156,6 @@ static loff_t get_size(loff_t offset, loff_t sizelimit, struct file *file) return loopsize >> 9; } -static loff_t get_loop_size(struct loop_device *lo, struct file *file) -{ - return get_size(lo->lo_offset, lo->lo_sizelimit, file); -} - /* * We support direct I/O only if lo_offset is aligned with the logical I/O size * of backing device, and the logical block size of loop is bigger than that of @@ -569,7 +562,7 @@ static int loop_change_fd(struct loop_device *lo, struct block_device *bdev, error = -EINVAL; /* size of the new backing store needs to be the same */ - if (get_loop_size(lo, file) != get_loop_size(lo, old_file)) + if (lo_calculate_size(lo, file) != lo_calculate_size(lo, old_file)) goto out_err; /* @@ -1063,7 +1056,7 @@ static int loop_configure(struct loop_device *lo, blk_mode_t mode, loop_update_dio(lo); loop_sysfs_init(lo); - size = get_loop_size(lo, file); + size = lo_calculate_size(lo, file); loop_set_size(lo, size); /* Order wrt reading lo_state in loop_validate_file(). */ @@ -1255,8 +1248,7 @@ out_unfreeze: if (partscan) clear_bit(GD_SUPPRESS_PART_SCAN, &lo->lo_disk->state); if (!err && size_changed) { - loff_t new_size = get_size(lo->lo_offset, lo->lo_sizelimit, - lo->lo_backing_file); + loff_t new_size = lo_calculate_size(lo, lo->lo_backing_file); loop_set_size(lo, new_size); } out_unlock: @@ -1399,7 +1391,7 @@ static int loop_set_capacity(struct loop_device *lo) if (unlikely(lo->lo_state != Lo_bound)) return -ENXIO; - size = get_loop_size(lo, lo->lo_backing_file); + size = lo_calculate_size(lo, lo->lo_backing_file); loop_set_size(lo, size); return 0; -- knotmirror 0.0.1 From 47b71abd58461a67cae71d2f2a9d44379e4e2fcf Mon Sep 17 00:00:00 2001 From: Rajeev Mishra Date: Mon, 18 Aug 2025 18:48:21 +0000 Subject: [PATCH 09/14] loop: use vfs_getattr_nosec for accurate file size Use vfs_getattr_nosec() in lo_calculate_size() for getting the file size, rather than just read the cached inode size via i_size_read(). This provides better results than cached inode data, particularly for network filesystems where metadata may be stale. Signed-off-by: Rajeev Mishra Reviewed-by: Yu Kuai Link: https://lore.kernel.org/r/20250818184821.115033-3-rajeevm@hpe.com [axboe: massage commit message] Signed-off-by: Jens Axboe --- drivers/block/loop.c | 15 +++++++++++++-- 1 file changed, 13 insertions(+), 2 deletions(-) diff --git a/drivers/block/loop.c b/drivers/block/loop.c index 0e1b9eb9db10..57263c273f0f 100644 --- a/drivers/block/loop.c +++ b/drivers/block/loop.c @@ -139,9 +139,20 @@ static int part_shift; static loff_t lo_calculate_size(struct loop_device *lo, struct file *file) { + struct kstat stat; loff_t loopsize; - /* Compute loopsize in bytes */ - loopsize = i_size_read(file->f_mapping->host); + int ret; + + /* + * Get the accurate file size. This provides better results than + * cached inode data, particularly for network filesystems where + * metadata may be stale. + */ + ret = vfs_getattr_nosec(&file->f_path, &stat, STATX_SIZE, 0); + if (ret) + return 0; + + loopsize = stat.size; if (lo->lo_offset > 0) loopsize -= lo->lo_offset; /* offset is beyond i_size, weird but possible */ -- knotmirror 0.0.1 From d0a2b527d8c32e46ccb8a34053468d4ff0c27e5c Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 18 Aug 2025 12:11:02 +0200 Subject: [PATCH 10/14] block: tone down bio_check_eod bdev_nr_sectors() == 0 is a pattern used for block devices that have been hot removed, don't spam the log about them. Signed-off-by: Christoph Hellwig Link: https://lore.kernel.org/r/20250818101102.1604551-1-hch@lst.de Signed-off-by: Jens Axboe --- block/blk-core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/block/blk-core.c b/block/blk-core.c index fdac48aec5ef..4201504158a1 100644 --- a/block/blk-core.c +++ b/block/blk-core.c @@ -557,7 +557,7 @@ static inline int bio_check_eod(struct bio *bio) sector_t maxsector = bdev_nr_sectors(bio->bi_bdev); unsigned int nr_sectors = bio_sectors(bio); - if (nr_sectors && + if (nr_sectors && maxsector && (nr_sectors > maxsector || bio->bi_iter.bi_sector > maxsector - nr_sectors)) { pr_info_ratelimited("%s: attempt to access beyond end of device\n" -- knotmirror 0.0.1 From 2d82f3bd8910eb65e30bb2a3c9b945bfb3b6d661 Mon Sep 17 00:00:00 2001 From: Ming Lei Date: Fri, 15 Aug 2025 21:17:37 +0800 Subject: [PATCH 11/14] blk-mq: fix lockdep warning in __blk_mq_update_nr_hw_queues Commit 5989bfe6ac6b ("block: restore two stage elevator switch while running nr_hw_queue update") reintroduced a lockdep warning by calling blk_mq_freeze_queue_nomemsave() before switching the I/O scheduler. The function blk_mq_elv_switch_none() calls elevator_change_done(). Running this while the queue is frozen causes a lockdep warning. Fix this by reordering the operations: first, switch the I/O scheduler to 'none', and then freeze the queue. This ensures that elevator_change_done() is not called on an already frozen queue. And this way is safe because elevator_set_none() does freeze queue before switching to none. Also we still have to rely on blk_mq_elv_switch_back() for switching back, and it has to cover unfrozen queue case. Cc: Nilay Shroff Cc: Yu Kuai Fixes: 5989bfe6ac6b ("block: restore two stage elevator switch while running nr_hw_queue update") Signed-off-by: Ming Lei Reviewed-by: Yu Kuai Reviewed-by: Nilay Shroff Link: https://lore.kernel.org/r/20250815131737.331692-1-ming.lei@redhat.com Signed-off-by: Jens Axboe --- block/blk-mq.c | 13 +++++++++---- 1 file changed, 9 insertions(+), 4 deletions(-) diff --git a/block/blk-mq.c b/block/blk-mq.c index b67d6c02eceb..ba3a4b77f578 100644 --- a/block/blk-mq.c +++ b/block/blk-mq.c @@ -5033,6 +5033,7 @@ static void __blk_mq_update_nr_hw_queues(struct blk_mq_tag_set *set, unsigned int memflags; int i; struct xarray elv_tbl, et_tbl; + bool queues_frozen = false; lockdep_assert_held(&set->tag_list_lock); @@ -5056,9 +5057,6 @@ static void __blk_mq_update_nr_hw_queues(struct blk_mq_tag_set *set, blk_mq_sysfs_unregister_hctxs(q); } - list_for_each_entry(q, &set->tag_list, tag_set_list) - blk_mq_freeze_queue_nomemsave(q); - /* * Switch IO scheduler to 'none', cleaning up the data associated * with the previous scheduler. We will switch back once we are done @@ -5068,6 +5066,9 @@ static void __blk_mq_update_nr_hw_queues(struct blk_mq_tag_set *set, if (blk_mq_elv_switch_none(q, &elv_tbl)) goto switch_back; + list_for_each_entry(q, &set->tag_list, tag_set_list) + blk_mq_freeze_queue_nomemsave(q); + queues_frozen = true; if (blk_mq_realloc_tag_set_tags(set, nr_hw_queues) < 0) goto switch_back; @@ -5091,8 +5092,12 @@ fallback: } switch_back: /* The blk_mq_elv_switch_back unfreezes queue for us. */ - list_for_each_entry(q, &set->tag_list, tag_set_list) + list_for_each_entry(q, &set->tag_list, tag_set_list) { + /* switch_back expects queue to be frozen */ + if (!queues_frozen) + blk_mq_freeze_queue_nomemsave(q); blk_mq_elv_switch_back(q, &elv_tbl, &et_tbl); + } list_for_each_entry(q, &set->tag_list, tag_set_list) { blk_mq_sysfs_register_hctxs(q); -- knotmirror 0.0.1 From 275332877e2fa9d6efa7402b1e897f6c6ee695bb Mon Sep 17 00:00:00 2001 From: Nilay Shroff Date: Thu, 14 Aug 2025 13:54:57 +0530 Subject: [PATCH 12/14] block: skip q->rq_qos check in rq_qos_done_bio() If a bio has BIO_QOS_THROTTLED or BIO_QOS_MERGED set, it implicitly guarantees that q->rq_qos is present. Avoid re-checking q->rq_qos in this case and call __rq_qos_done_bio() directly as a minor optimization. Suggested-by : Yu Kuai Signed-off-by: Nilay Shroff Reviewed-by: Ming Lei Reviewed-by: Yu Kuai Link: https://lore.kernel.org/r/20250814082612.500845-2-nilay@linux.ibm.com Signed-off-by: Jens Axboe --- block/blk-rq-qos.h | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-) diff --git a/block/blk-rq-qos.h b/block/blk-rq-qos.h index 39749f4066fb..28125fc49eff 100644 --- a/block/blk-rq-qos.h +++ b/block/blk-rq-qos.h @@ -142,8 +142,14 @@ static inline void rq_qos_done_bio(struct bio *bio) bio->bi_bdev && (bio_flagged(bio, BIO_QOS_THROTTLED) || bio_flagged(bio, BIO_QOS_MERGED))) { struct request_queue *q = bdev_get_queue(bio->bi_bdev); - if (q->rq_qos) - __rq_qos_done_bio(q->rq_qos, bio); + + /* + * If a bio has BIO_QOS_xxx set, it implicitly implies that + * q->rq_qos is present. So, we skip re-checking q->rq_qos + * here as an extra optimization and directly call + * __rq_qos_done_bio(). + */ + __rq_qos_done_bio(q->rq_qos, bio); } } -- knotmirror 0.0.1 From ade1beea1c27657712aa8f594226d461639382ff Mon Sep 17 00:00:00 2001 From: Nilay Shroff Date: Thu, 14 Aug 2025 13:54:58 +0530 Subject: [PATCH 13/14] block: decrement block_rq_qos static key in rq_qos_del() rq_qos_add() increments the block_rq_qos static key when a QoS policy is attached. When a QoS policy is removed via rq_qos_del(), we must symmetrically decrement the static key. If this removal drops the last QoS policy from the queue (q->rq_qos becomes NULL), the static branch can be disabled and the jump label patched to a NOP, avoiding overhead on the hot path. This change ensures rq_qos_add()/rq_qos_del() keep the block_rq_qos static key balanced and prevents leaving the branch permanently enabled after the last policy is removed. Fixes: 033b667a823e ("block: blk-rq-qos: guard rq-qos helpers by static key") Signed-off-by: Nilay Shroff Reviewed-by: Ming Lei Reviewed-by: Yu Kuai Link: https://lore.kernel.org/r/20250814082612.500845-3-nilay@linux.ibm.com Signed-off-by: Jens Axboe --- block/blk-rq-qos.c | 1 + 1 file changed, 1 insertion(+) diff --git a/block/blk-rq-qos.c b/block/blk-rq-qos.c index 848591fb3c57..b1e24bb85ad2 100644 --- a/block/blk-rq-qos.c +++ b/block/blk-rq-qos.c @@ -374,6 +374,7 @@ void rq_qos_del(struct rq_qos *rqos) for (cur = &q->rq_qos; *cur; cur = &(*cur)->next) { if (*cur == rqos) { *cur = rqos->next; + static_branch_dec(&block_rq_qos); break; } } -- knotmirror 0.0.1 From 370ac285f23aecae40600851fb4a1a9e75e50973 Mon Sep 17 00:00:00 2001 From: Nilay Shroff Date: Thu, 14 Aug 2025 13:54:59 +0530 Subject: [PATCH 14/14] block: avoid cpu_hotplug_lock depedency on freeze_lock A recent lockdep[1] splat observed while running blktest block/005 reveals a potential deadlock caused by the cpu_hotplug_lock dependency on ->freeze_lock. This dependency was introduced by commit 033b667a823e ("block: blk-rq-qos: guard rq-qos helpers by static key"). That change added a static key to avoid fetching q->rq_qos when neither blk-wbt nor blk-iolatency is configured. The static key dynamically patches kernel text to a NOP when disabled, eliminating overhead of fetching q->rq_qos in the I/O hot path. However, enabling a static key at runtime requires acquiring both cpu_hotplug_lock and jump_label_mutex. When this happens after the queue has already been frozen (i.e., while holding ->freeze_lock), it creates a locking dependency from cpu_hotplug_lock to ->freeze_lock, which leads to a potential deadlock reported by lockdep [1]. To resolve this, replace the static key mechanism with q->queue_flags: QUEUE_FLAG_QOS_ENABLED. This flag is evaluated in the fast path before accessing q->rq_qos. If the flag is set, we proceed to fetch q->rq_qos; otherwise, the access is skipped. Since q->queue_flags is commonly accessed in IO hotpath and resides in the first cacheline of struct request_queue, checking it imposes minimal overhead while eliminating the deadlock risk. This change avoids the lockdep splat without introducing performance regressions. [1] https://lore.kernel.org/linux-block/4fdm37so3o4xricdgfosgmohn63aa7wj3ua4e5vpihoamwg3ui@fq42f5q5t5ic/ Reported-by: Shinichiro Kawasaki Closes: https://lore.kernel.org/linux-block/4fdm37so3o4xricdgfosgmohn63aa7wj3ua4e5vpihoamwg3ui@fq42f5q5t5ic/ Fixes: 033b667a823e ("block: blk-rq-qos: guard rq-qos helpers by static key") Tested-by: Shin'ichiro Kawasaki Signed-off-by: Nilay Shroff Reviewed-by: Ming Lei Reviewed-by: Yu Kuai Link: https://lore.kernel.org/r/20250814082612.500845-4-nilay@linux.ibm.com Signed-off-by: Jens Axboe --- block/blk-mq-debugfs.c | 1 + block/blk-rq-qos.c | 9 ++++--- block/blk-rq-qos.h | 54 ++++++++++++++++++++++++------------------ include/linux/blkdev.h | 1 + 4 files changed, 37 insertions(+), 28 deletions(-) diff --git a/block/blk-mq-debugfs.c b/block/blk-mq-debugfs.c index 7ed3e71f2fc0..32c65efdda46 100644 --- a/block/blk-mq-debugfs.c +++ b/block/blk-mq-debugfs.c @@ -95,6 +95,7 @@ static const char *const blk_queue_flag_name[] = { QUEUE_FLAG_NAME(SQ_SCHED), QUEUE_FLAG_NAME(DISABLE_WBT_DEF), QUEUE_FLAG_NAME(NO_ELV_SWITCH), + QUEUE_FLAG_NAME(QOS_ENABLED), }; #undef QUEUE_FLAG_NAME diff --git a/block/blk-rq-qos.c b/block/blk-rq-qos.c index b1e24bb85ad2..654478dfbc20 100644 --- a/block/blk-rq-qos.c +++ b/block/blk-rq-qos.c @@ -2,8 +2,6 @@ #include "blk-rq-qos.h" -__read_mostly DEFINE_STATIC_KEY_FALSE(block_rq_qos); - /* * Increment 'v', if 'v' is below 'below'. Returns true if we succeeded, * false if 'v' + 1 would be bigger than 'below'. @@ -319,8 +317,8 @@ void rq_qos_exit(struct request_queue *q) struct rq_qos *rqos = q->rq_qos; q->rq_qos = rqos->next; rqos->ops->exit(rqos); - static_branch_dec(&block_rq_qos); } + blk_queue_flag_clear(QUEUE_FLAG_QOS_ENABLED, q); mutex_unlock(&q->rq_qos_mutex); } @@ -346,7 +344,7 @@ int rq_qos_add(struct rq_qos *rqos, struct gendisk *disk, enum rq_qos_id id, goto ebusy; rqos->next = q->rq_qos; q->rq_qos = rqos; - static_branch_inc(&block_rq_qos); + blk_queue_flag_set(QUEUE_FLAG_QOS_ENABLED, q); blk_mq_unfreeze_queue(q, memflags); @@ -374,10 +372,11 @@ void rq_qos_del(struct rq_qos *rqos) for (cur = &q->rq_qos; *cur; cur = &(*cur)->next) { if (*cur == rqos) { *cur = rqos->next; - static_branch_dec(&block_rq_qos); break; } } + if (!q->rq_qos) + blk_queue_flag_clear(QUEUE_FLAG_QOS_ENABLED, q); blk_mq_unfreeze_queue(q, memflags); mutex_lock(&q->debugfs_mutex); diff --git a/block/blk-rq-qos.h b/block/blk-rq-qos.h index 28125fc49eff..1fe22000a379 100644 --- a/block/blk-rq-qos.h +++ b/block/blk-rq-qos.h @@ -12,7 +12,6 @@ #include "blk-mq-debugfs.h" struct blk_mq_debugfs_attr; -extern struct static_key_false block_rq_qos; enum rq_qos_id { RQ_QOS_WBT, @@ -113,49 +112,55 @@ void __rq_qos_queue_depth_changed(struct rq_qos *rqos); static inline void rq_qos_cleanup(struct request_queue *q, struct bio *bio) { - if (static_branch_unlikely(&block_rq_qos) && q->rq_qos) + if (unlikely(test_bit(QUEUE_FLAG_QOS_ENABLED, &q->queue_flags)) && + q->rq_qos) __rq_qos_cleanup(q->rq_qos, bio); } static inline void rq_qos_done(struct request_queue *q, struct request *rq) { - if (static_branch_unlikely(&block_rq_qos) && q->rq_qos && - !blk_rq_is_passthrough(rq)) + if (unlikely(test_bit(QUEUE_FLAG_QOS_ENABLED, &q->queue_flags)) && + q->rq_qos && !blk_rq_is_passthrough(rq)) __rq_qos_done(q->rq_qos, rq); } static inline void rq_qos_issue(struct request_queue *q, struct request *rq) { - if (static_branch_unlikely(&block_rq_qos) && q->rq_qos) + if (unlikely(test_bit(QUEUE_FLAG_QOS_ENABLED, &q->queue_flags)) && + q->rq_qos) __rq_qos_issue(q->rq_qos, rq); } static inline void rq_qos_requeue(struct request_queue *q, struct request *rq) { - if (static_branch_unlikely(&block_rq_qos) && q->rq_qos) + if (unlikely(test_bit(QUEUE_FLAG_QOS_ENABLED, &q->queue_flags)) && + q->rq_qos) __rq_qos_requeue(q->rq_qos, rq); } static inline void rq_qos_done_bio(struct bio *bio) { - if (static_branch_unlikely(&block_rq_qos) && - bio->bi_bdev && (bio_flagged(bio, BIO_QOS_THROTTLED) || - bio_flagged(bio, BIO_QOS_MERGED))) { - struct request_queue *q = bdev_get_queue(bio->bi_bdev); - - /* - * If a bio has BIO_QOS_xxx set, it implicitly implies that - * q->rq_qos is present. So, we skip re-checking q->rq_qos - * here as an extra optimization and directly call - * __rq_qos_done_bio(). - */ - __rq_qos_done_bio(q->rq_qos, bio); - } + struct request_queue *q; + + if (!bio->bi_bdev || (!bio_flagged(bio, BIO_QOS_THROTTLED) && + !bio_flagged(bio, BIO_QOS_MERGED))) + return; + + q = bdev_get_queue(bio->bi_bdev); + + /* + * If a bio has BIO_QOS_xxx set, it implicitly implies that + * q->rq_qos is present. So, we skip re-checking q->rq_qos + * here as an extra optimization and directly call + * __rq_qos_done_bio(). + */ + __rq_qos_done_bio(q->rq_qos, bio); } static inline void rq_qos_throttle(struct request_queue *q, struct bio *bio) { - if (static_branch_unlikely(&block_rq_qos) && q->rq_qos) { + if (unlikely(test_bit(QUEUE_FLAG_QOS_ENABLED, &q->queue_flags)) && + q->rq_qos) { bio_set_flag(bio, BIO_QOS_THROTTLED); __rq_qos_throttle(q->rq_qos, bio); } @@ -164,14 +169,16 @@ static inline void rq_qos_throttle(struct request_queue *q, struct bio *bio) static inline void rq_qos_track(struct request_queue *q, struct request *rq, struct bio *bio) { - if (static_branch_unlikely(&block_rq_qos) && q->rq_qos) + if (unlikely(test_bit(QUEUE_FLAG_QOS_ENABLED, &q->queue_flags)) && + q->rq_qos) __rq_qos_track(q->rq_qos, rq, bio); } static inline void rq_qos_merge(struct request_queue *q, struct request *rq, struct bio *bio) { - if (static_branch_unlikely(&block_rq_qos) && q->rq_qos) { + if (unlikely(test_bit(QUEUE_FLAG_QOS_ENABLED, &q->queue_flags)) && + q->rq_qos) { bio_set_flag(bio, BIO_QOS_MERGED); __rq_qos_merge(q->rq_qos, rq, bio); } @@ -179,7 +186,8 @@ static inline void rq_qos_merge(struct request_queue *q, struct request *rq, static inline void rq_qos_queue_depth_changed(struct request_queue *q) { - if (static_branch_unlikely(&block_rq_qos) && q->rq_qos) + if (unlikely(test_bit(QUEUE_FLAG_QOS_ENABLED, &q->queue_flags)) && + q->rq_qos) __rq_qos_queue_depth_changed(q->rq_qos); } diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h index 95886b404b16..fe1797bbec42 100644 --- a/include/linux/blkdev.h +++ b/include/linux/blkdev.h @@ -656,6 +656,7 @@ enum { QUEUE_FLAG_SQ_SCHED, /* single queue style io dispatch */ QUEUE_FLAG_DISABLE_WBT_DEF, /* for sched to disable/enable wbt */ QUEUE_FLAG_NO_ELV_SWITCH, /* can't switch elevator any more */ + QUEUE_FLAG_QOS_ENABLED, /* qos is enabled */ QUEUE_FLAG_MAX };