[PATCH 0/2] block: fix NOWAIT handling and locking in buffered write fallback

Linux block layer
 help / color / mirror / Atom feed

* [PATCH 0/2] block: fix NOWAIT handling and locking in buffered write fallback
@ 2026-10-15  9:11 Julian Sun
  2026-10-15  9:11 ` [PATCH 1/2] block: reject NOWAIT " Julian Sun
                   ` (2 more replies)
  0 siblings, 3 replies; 7+ messages in thread
From: Julian Sun @ 2026-10-15  9:11 UTC (permalink / raw)
  To: linux-block; +Cc: jack, axboe, viro, mcgrof, djwong

Hi,

This series fixes two issues in buffered fallback for direct writes to
block devices.

1. NOWAIT requests can fall back to buffered I/O and wait for writeback.
2. The fallback path lacks shared inode locking against set_blocksize().

Julian Sun (2):
  block: reject NOWAIT buffered write fallback
  block: take the inode lock for buffered write fallback

 block/fops.c | 25 ++++++++++++++++---------
 1 file changed, 16 insertions(+), 9 deletions(-)

-- 
2.39.5

^ permalink raw reply	[flat|nested] 7+ messages in thread

* [PATCH 1/2] block: reject NOWAIT buffered write fallback
  2026-10-15  9:11 [PATCH 0/2] block: fix NOWAIT handling and locking in buffered write fallback Julian Sun
@ 2026-10-15  9:11 ` Julian Sun
  2026-10-15  9:11 ` [PATCH 2/2] block: take the inode lock for " Julian Sun
  2026-10-15 12:41 ` [PATCH 0/2] block: fix NOWAIT handling and locking in " Jens Axboe
  2 siblings, 0 replies; 7+ messages in thread
From: Julian Sun @ 2026-10-15  9:11 UTC (permalink / raw)
  To: linux-block; +Cc: jack, axboe, viro, mcgrof, djwong

blkdev_write_iter() rejects NOWAIT buffered writes, but does not
enforce this restriction when direct I/O falls back to buffered I/O.
Successful fallback waits for writeback, violating NOWAIT semantics.

Reject NOWAIT requests in blkdev_buffered_write() with -EAGAIN.

Fixes: c35fc7a5abae ("block_dev: support RFW_NOWAIT on block device nodes")
Signed-off-by: Julian Sun <sunjunchao@bytedance.com>
---
 block/fops.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/block/fops.c b/block/fops.c
index 2ce7c6c4714e..982a713b4c27 100644
--- a/block/fops.c
+++ b/block/fops.c
@@ -709,6 +709,10 @@ blkdev_direct_write(struct kiocb *iocb, struct iov_iter *from)
 
 static ssize_t blkdev_buffered_write(struct kiocb *iocb, struct iov_iter *from)
 {
+	/* Buffered fallback waits for writeback, which NOWAIT cannot allow. */
+	if (iocb->ki_flags & IOCB_NOWAIT)
+		return -EAGAIN;
+
 	return iomap_file_buffered_write(iocb, from, &blkdev_iomap_ops, NULL,
 			NULL);
 }
-- 
2.39.5


^ permalink raw reply related	[flat|nested] 7+ messages in thread

* [PATCH 2/2] block: take the inode lock for buffered write fallback
  2026-10-15  9:11 [PATCH 0/2] block: fix NOWAIT handling and locking in buffered write fallback Julian Sun
  2026-10-15  9:11 ` [PATCH 1/2] block: reject NOWAIT " Julian Sun
@ 2026-10-15  9:11 ` Julian Sun
  2026-10-27  4:24   ` Shin'ichiro Kawasaki
  2026-10-15 12:41 ` [PATCH 0/2] block: fix NOWAIT handling and locking in " Jens Axboe
  2 siblings, 1 reply; 7+ messages in thread
From: Julian Sun @ 2026-10-15  9:11 UTC (permalink / raw)
  To: linux-block; +Cc: jack, axboe, viro, mcgrof, djwong

Commit c0e473a0d226 ("block: fix race between set_blocksize and read
paths") added shared inode locking around ordinary buffered writes,
but missed the direct-I/O fallback path.

Move the lock into blkdev_buffered_write() so both paths serialize
with set_blocksize().

Also correct the comment claiming that this path takes invalidate_lock;
it only takes i_rwsem.

Fixes: c0e473a0d226 ("block: fix race between set_blocksize and read paths")
Signed-off-by: Julian Sun <sunjunchao@bytedance.com>
---
 block/fops.c | 21 ++++++++++++---------
 1 file changed, 12 insertions(+), 9 deletions(-)

diff --git a/block/fops.c b/block/fops.c
index 982a713b4c27..74e7a0713d1a 100644
--- a/block/fops.c
+++ b/block/fops.c
@@ -709,12 +709,22 @@ blkdev_direct_write(struct kiocb *iocb, struct iov_iter *from)
 
 static ssize_t blkdev_buffered_write(struct kiocb *iocb, struct iov_iter *from)
 {
+	struct inode *inode = bdev_file_inode(iocb->ki_filp);
+	ssize_t ret;
+
 	/* Buffered fallback waits for writeback, which NOWAIT cannot allow. */
 	if (iocb->ki_flags & IOCB_NOWAIT)
 		return -EAGAIN;
 
-	return iomap_file_buffered_write(iocb, from, &blkdev_iomap_ops, NULL,
-			NULL);
+	/*
+	 * Take i_rwsem to avoid racing with set_blocksize() changing
+	 * i_blkbits/folio order and truncating the page cache.
+	 */
+	inode_lock_shared(inode);
+	ret = iomap_file_buffered_write(iocb, from, &blkdev_iomap_ops, NULL,
+					NULL);
+	inode_unlock_shared(inode);
+	return ret;
 }
 
 /*
@@ -773,14 +783,7 @@ static ssize_t blkdev_write_iter(struct kiocb *iocb, struct iov_iter *from)
 			ret = direct_write_fallback(iocb, from, ret,
 					blkdev_buffered_write(iocb, from));
 	} else {
-		/*
-		 * Take i_rwsem and invalidate_lock to avoid racing with
-		 * set_blocksize changing i_blkbits/folio order and punching
-		 * out the pagecache.
-		 */
-		inode_lock_shared(bd_inode);
 		ret = blkdev_buffered_write(iocb, from);
-		inode_unlock_shared(bd_inode);
 	}
 
 	if (ret > 0)
-- 
2.39.5


^ permalink raw reply related	[flat|nested] 7+ messages in thread

* Re: [PATCH 2/2] block: take the inode lock for buffered write fallback
  2026-10-15  9:11 ` [PATCH 2/2] block: take the inode lock for " Julian Sun
@ 2026-10-27  4:24   ` Shin'ichiro Kawasaki
  2026-10-27  6:13     ` Julian Sun
  0 siblings, 1 reply; 7+ messages in thread
From: Shin'ichiro Kawasaki @ 2026-10-27  4:24 UTC (permalink / raw)
  To: Julian Sun; +Cc: linux-block, jack, axboe, viro, mcgrof, djwong, tz2294

CC+: Tal Zassman,

On Oct 15, 2026 / 17:11, Julian Sun wrote:
> Commit c0e473a0d226 ("block: fix race between set_blocksize and read
> paths") added shared inode locking around ordinary buffered writes,
> but missed the direct-I/O fallback path.
> 
> Move the lock into blkdev_buffered_write() so both paths serialize
> with set_blocksize().
> 
> Also correct the comment claiming that this path takes invalidate_lock;
> it only takes i_rwsem.
> 
> Fixes: c0e473a0d226 ("block: fix race between set_blocksize and read paths")
> Signed-off-by: Julian Sun <sunjunchao@bytedance.com>

When I checking the new blktests test case block/050 by Tal Zussman [1], I found
the test case hangs [2]. I bisected and this kernel change triggers the hang.

Julian, may I ask you to take a look in the hang?

[1] https://lore.kernel.org/linux-block/20261015-blkdev-fixes-tests-v3-1-b6ffa6410514@columbia.edu/

[2] dmesg during test case hang

[   93.055848] [   T1042] run blktests block/050 at 2026-10-27 13:15:20.266665152
[   93.132099] [   T1042] null_blk: disk nullb1 created

[   93.204476] [   T1088] ============================================
[   93.205320] [   T1088] WARNING: possible recursive locking detected
[   93.206137] [   T1088] 7.3.0-rc2+ #715 Not tainted
[   93.206784] [   T1088] --------------------------------------------
[   93.207569] [   T1088] dio-fallback-ra/1088 is trying to acquire lock:
[   93.208415] [   T1088] ffff888100452a78 (&sb->s_type->i_mutex_key#13){++++}-{4:4}, at: blkdev_write_iter+0x928/0xf40
[   93.209769] [   T1088] 
                          but task is already holding lock:
[   93.210736] [   T1088] ffff888100452a78 (&sb->s_type->i_mutex_key#13){++++}-{4:4}, at: blkdev_write_iter+0x882/0xf40
[   93.212048] [   T1088] 
                          other info that might help us debug this:
[   93.213080] [   T1088]  Possible unsafe locking scenario:

[   93.214051] [   T1088]        CPU0
[   93.214501] [   T1088]        ----
[   93.214934] [   T1088]   lock(&sb->s_type->i_mutex_key#13);
[   93.215671] [   T1088]   lock(&sb->s_type->i_mutex_key#13);
[   93.216370] [   T1088] 
                           *** DEADLOCK ***

[   93.217409] [   T1088]  May be due to missing lock nesting notation

[   93.218468] [   T1088] locks held by dio-fallback-ra/1088: 1, last CPU#1:
[   93.219343] [   T1088]  #0: ffff888100452a78 (&sb->s_type->i_mutex_key#13){++++}-{4:4}, at: blkdev_write_iter+0x882/0xf40
[   93.220736] [   T1088] 
                          stack backtrace:
[   93.221490] [   T1088] CPU: 1 UID: 0 PID: 1088 Comm: dio-fallback-ra Not tainted 7.3.0-rc2+ #715 PREEMPT(full) 
[   93.222766] [   T1088] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.17.0-10.fc44 06/10/2025
[   93.223972] [   T1088] Call Trace:
[   93.224415] [   T1088]  <TASK>
[   93.224801] [   T1088]  dump_stack_lvl+0x6a/0x90
[   93.225381] [   T1088]  print_deadlock_bug.cold+0xc0/0xcd
[   93.226033] [   T1088]  __lock_acquire+0x1536/0x2490
[   93.226705] [   T1088]  lock_acquire+0x1b1/0x340
[   93.227262] [   T1088]  ? blkdev_write_iter+0x928/0xf40
[   93.227880] [   T1088]  ? __pfx___might_resched+0x10/0x10
[   93.228538] [   T1088]  down_read+0xa6/0x540
[   93.229049] [   T1088]  ? blkdev_write_iter+0x928/0xf40
[   93.229754] [   T1088]  ? __pfx_down_read+0x10/0x10
[   93.230374] [   T1088]  ? filemap_check_errors+0x58/0x100
[   93.231040] [   T1088]  blkdev_write_iter+0x928/0xf40
[   93.231722] [   T1088]  do_iter_readv_writev+0x48a/0xb80
[   93.232382] [   T1088]  ? __pfx_folio_add_lru_vma+0x10/0x10
[   93.233070] [   T1088]  ? rcu_read_unlock+0x1c/0x60
[   93.233764] [   T1088]  ? __folio_mod_stat+0x4c/0x190
[   93.234399] [   T1088]  ? __pfx_do_iter_readv_writev+0x10/0x10
[   93.235108] [   T1088]  ? folio_add_new_anon_rmap+0x34f/0xd40
[   93.235860] [   T1088]  ? selinux_file_permission+0x3c0/0x530
[   93.236577] [   T1088]  ? security_file_permission+0x44/0xc0
[   93.237316] [   T1088]  ? rw_verify_area+0x77/0x650
[   93.237917] [   T1088]  vfs_writev+0x27d/0xbc0
[   93.238484] [   T1088]  ? __pfx_vfs_writev+0x10/0x10
[   93.239095] [   T1088]  ? __handle_mm_fault+0x1584/0x1fb0
[   93.239816] [   T1088]  ? lock_acquire+0x1b1/0x340
[   93.240425] [   T1088]  ? __fget_files+0x1d8/0x440
[   93.241021] [   T1088]  ? do_pwritev+0x18a/0x300
[   93.241653] [   T1088]  do_pwritev+0x18a/0x300
[   93.242183] [   T1088]  ? __pfx_do_pwritev+0x10/0x10
[   93.242837] [   T1088]  ? lock_release+0x1b5/0x340
[   93.243451] [   T1088]  ? rcu_is_watching+0x12/0xb0
[   93.244054] [   T1088]  do_syscall_64+0xf2/0x690
[   93.244688] [   T1088]  ? do_syscall_64+0x2d/0x690
[   93.245277] [   T1088]  ? trace_hardirqs_off_finish+0x139/0x180
[   93.246010] [   T1088]  ? lockdep_hardirqs_off+0xaf/0xf0
[   93.246724] [   T1088]  ? do_syscall_64+0x60/0x690
[   93.247325] [   T1088]  entry_SYSCALL_64_after_hwframe+0x76/0x7e
[   93.248054] [   T1088] RIP: 0033:0x7f1cae028e32
[   93.248676] [   T1088] Code: 08 0f 85 91 41 ff ff 49 89 fb 48 89 f0 48 89 d7 48 89 ce 4c 89 c2 4d 89 ca 4c 8b 44 24 08 4c 8b 4c 24 10 4c 89 5c 24 08 0f 05 <c3> 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 00 f3 0f 1e fa 55 bf 01 00
[   93.251120] [   T1088] RSP: 002b:00007f1cadf98d98 EFLAGS: 00000246 ORIG_RAX: 0000000000000128
[   93.252230] [   T1088] RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007f1cae028e32
[   93.253253] [   T1088] RDX: 0000000000000002 RSI: 00007f1cadf98e20 RDI: 0000000000000004
[   93.254275] [   T1088] RBP: 00007f1cadf98dc0 R08: 0000000000000000 R09: 0000000000000000
[   93.255295] [   T1088] R10: 0000000000000000 R11: 0000000000000246 R12: 00007f1ca8001000
[   93.256307] [   T1088] R13: 00007fff0eee6560 R14: 0000000000000000 R15: 00007fff0eee6667
[   93.257329] [   T1088]  </TASK>
[  247.593474] [     T48] INFO: task dio-fallback-ra:1088 blocked for more than 122 seconds.
[  247.595947] [     T48]       Not tainted 7.3.0-rc2+ #715
[  247.598101] [     T48] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[  247.599361] [     T48] task:dio-fallback-ra state:D stack:0     pid:1088  tgid:1087  ppid:1042   task_flags:0x400040 flags:0x00080000
[  247.599727] [     T48] Call Trace:
[  247.600141] [     T48]  <TASK>
[  247.600524] [     T48]  __schedule+0x1fcc/0x6ef0
[  247.600905] [     T48]  ? asm_sysvec_apic_timer_interrupt+0x16/0x20
[  247.601293] [     T48]  ? trace_hardirqs_on+0x14/0x190
[  247.601670] [     T48]  ? __pfx___schedule+0x10/0x10
[  247.602045] [     T48]  ? lock_acquire+0x2f5/0x340
[  247.602430] [     T48]  ? lock_acquire+0x2f5/0x340
[  247.602801] [     T48]  ? rcu_is_watching+0x12/0xb0
[  247.603182] [     T48]  ? lock_release+0x285/0x340
[  247.603549] [     T48]  schedule+0xe3/0x2e0
[  247.603917] [     T48]  ? rwsem_down_read_slowpath+0x31e/0x10d0
[  247.604296] [     T48]  schedule_preempt_disabled+0x14/0x30
[  247.604665] [     T48]  rwsem_down_read_slowpath+0x61a/0x10d0
[  247.605050] [     T48]  ? __pfx_rwsem_down_read_slowpath+0x10/0x10
[  247.605412] [     T48]  ? __pfx___might_resched+0x10/0x10
[  247.605765] [     T48]  down_read+0xfa/0x540
[  247.606110] [     T48]  ? __pfx_down_read+0x10/0x10
[  247.606889] [     T48]  ? filemap_check_errors+0x58/0x100
[  247.607286] [     T48]  blkdev_write_iter+0x928/0xf40
[  247.607634] [     T48]  do_iter_readv_writev+0x48a/0xb80
[  247.607987] [     T48]  ? __pfx_folio_add_lru_vma+0x10/0x10
[  247.608348] [     T48]  ? rcu_read_unlock+0x1c/0x60
[  247.608707] [     T48]  ? __folio_mod_stat+0x4c/0x190
[  247.609083] [     T48]  ? __pfx_do_iter_readv_writev+0x10/0x10
[  247.609445] [     T48]  ? folio_add_new_anon_rmap+0x34f/0xd40
[  247.609802] [     T48]  ? selinux_file_permission+0x3c0/0x530
[  247.610167] [     T48]  ? security_file_permission+0x44/0xc0
[  247.610521] [     T48]  ? rw_verify_area+0x77/0x650
[  247.610878] [     T48]  vfs_writev+0x27d/0xbc0
[  247.611244] [     T48]  ? __pfx_vfs_writev+0x10/0x10
[  247.611611] [     T48]  ? __handle_mm_fault+0x1584/0x1fb0
[  247.611971] [     T48]  ? lock_acquire+0x1b1/0x340
[  247.612375] [     T48]  ? __fget_files+0x1d8/0x440
[  247.612731] [     T48]  ? do_pwritev+0x18a/0x300
[  247.613115] [     T48]  do_pwritev+0x18a/0x300
[  247.613472] [     T48]  ? __pfx_do_pwritev+0x10/0x10
[  247.613849] [     T48]  ? lock_release+0x1b5/0x340
[  247.614160] [     T48]  ? rcu_is_watching+0x12/0xb0
[  247.614478] [     T48]  do_syscall_64+0xf2/0x690
[  247.614823] [     T48]  ? do_syscall_64+0x2d/0x690
[  247.615146] [     T48]  ? trace_hardirqs_off_finish+0x139/0x180
[  247.615457] [     T48]  ? lockdep_hardirqs_off+0xaf/0xf0
[  247.615759] [     T48]  ? do_syscall_64+0x60/0x690
[  247.616054] [     T48]  entry_SYSCALL_64_after_hwframe+0x76/0x7e
[  247.616360] [     T48] RIP: 0033:0x7f1cae028e32
[  247.616652] [     T48] RSP: 002b:00007f1cadf98d98 EFLAGS: 00000246 ORIG_RAX: 0000000000000128
[  247.616984] [     T48] RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007f1cae028e32
[  247.617287] [     T48] RDX: 0000000000000002 RSI: 00007f1cadf98e20 RDI: 0000000000000004
[  247.617571] [     T48] RBP: 00007f1cadf98dc0 R08: 0000000000000000 R09: 0000000000000000
[  247.617850] [     T48] R10: 0000000000000000 R11: 0000000000000246 R12: 00007f1ca8001000
[  247.618139] [     T48] R13: 00007fff0eee6560 R14: 0000000000000000 R15: 00007fff0eee6667
[  247.618423] [     T48]  </TASK>
[  247.618698] [     T48] INFO: task dio-fallback-ra:1088 is blocked on an rw-semaphore, but the owner is not found.
[  247.618995] [     T48] INFO: task dio-fallback-ra:1089 blocked for more than 122 seconds.
[  247.619302] [     T48]       Not tainted 7.3.0-rc2+ #715
[  247.619669] [     T48] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[  247.619997] [     T48] task:dio-fallback-ra state:D stack:0     pid:1089  tgid:1087  ppid:1042   task_flags:0x400040 flags:0x00080000
[  247.620340] [     T48] Call Trace:
[  247.620660] [     T48]  <TASK>
[  247.620982] [     T48]  __schedule+0x1fcc/0x6ef0
[  247.621349] [     T48]  ? __pfx_bio_wait_end_io+0x10/0x10
[  247.621671] [     T48]  ? __pfx___schedule+0x10/0x10
[  247.622021] [     T48]  ? lock_acquire+0x2f5/0x340
[  247.622349] [     T48]  ? lock_acquire+0x2f5/0x340
[  247.622659] [     T48]  ? rcu_is_watching+0x12/0xb0
[  247.622973] [     T48]  ? lock_release+0x285/0x340
[  247.623294] [     T48]  schedule+0xe3/0x2e0
[  247.623602] [     T48]  ? rwsem_down_read_slowpath+0x31e/0x10d0
[  247.623957] [     T48]  schedule_preempt_disabled+0x14/0x30
[  247.624285] [     T48]  rwsem_down_read_slowpath+0x61a/0x10d0
[  247.624637] [     T48]  ? __pfx_rwsem_down_read_slowpath+0x10/0x10
[  247.624955] [     T48]  ? blkdev_direct_IO+0x86f/0x1c50
[  247.625322] [     T48]  ? __pfx___might_resched+0x10/0x10
[  247.625638] [     T48]  down_read+0xfa/0x540
[  247.625985] [     T48]  ? __pfx_down_read+0x10/0x10
[  247.626309] [     T48]  ? filemap_check_errors+0x58/0x100
[  247.626622] [     T48]  blkdev_write_iter+0x882/0xf40
[  247.626979] [     T48]  do_iter_readv_writev+0x48a/0xb80
[  247.627306] [     T48]  ? vfs_writev+0x831/0xbc0
[  247.627618] [     T48]  ? __pfx_do_iter_readv_writev+0x10/0x10
[  247.627976] [     T48]  ? selinux_file_permission+0x3c0/0x530
[  247.628303] [     T48]  ? security_file_permission+0x44/0xc0
[  247.628621] [     T48]  ? rw_verify_area+0x77/0x650
[  247.628941] [     T48]  vfs_writev+0x27d/0xbc0
[  247.629307] [     T48]  ? __pfx_vfs_writev+0x10/0x10
[  247.629628] [     T48]  ? __fget_files+0x4f/0x440
[  247.629981] [     T48]  ? __fget_files+0x1d8/0x440
[  247.630309] [     T48]  ? do_pwritev+0x18a/0x300
[  247.630623] [     T48]  do_pwritev+0x18a/0x300
[  247.630941] [     T48]  ? __pfx_do_pwritev+0x10/0x10
[  247.631265] [     T48]  ? lock_release+0x1b5/0x340
[  247.631577] [     T48]  ? rcu_is_watching+0x12/0xb0
[  247.631918] [     T48]  do_syscall_64+0xf2/0x690
[  247.632239] [     T48]  ? rcu_read_unlock+0x17/0x60
[  247.632548] [     T48]  ? lock_release+0x1b5/0x340
[  247.633267] [     T48]  ? __handle_mm_fault+0x1584/0x1fb0
[  247.633573] [     T48]  ? __lock_acquire+0x54a/0x2490
[  247.633875] [     T48]  ? __pfx___css_rstat_updated+0x10/0x10
[  247.634190] [     T48]  ? lock_acquire+0x1b1/0x340
[  247.634485] [     T48]  ? rcu_is_watching+0x12/0xb0
[  247.634783] [     T48]  ? count_memcg_events+0x107/0x4c0
[  247.635099] [     T48]  ? find_held_lock+0x2b/0x80
[  247.635397] [     T48]  ? rcu_read_unlock+0x17/0x60
[  247.635674] [     T48]  ? lock_release+0x1b5/0x340
[  247.635991] [     T48]  ? find_held_lock+0x2b/0x80
[  247.636308] [     T48]  ? exc_page_fault+0x94/0x140
[  247.636563] [     T48]  ? lock_release+0x1b5/0x340
[  247.636853] [     T48]  ? rcu_is_watching+0x12/0xb0
[  247.637100] [     T48]  ? do_syscall_64+0x2d/0x690
[  247.637359] [     T48]  ? trace_hardirqs_off_finish+0x139/0x180
[  247.637610] [     T48]  ? lockdep_hardirqs_off+0xaf/0xf0
[  247.637864] [     T48]  ? do_syscall_64+0x60/0x690
[  247.638113] [     T48]  entry_SYSCALL_64_after_hwframe+0x76/0x7e
[  247.638382] [     T48] RIP: 0033:0x7f1cae028e32
[  247.638639] [     T48] RSP: 002b:00007f1cad797d98 EFLAGS: 00000246 ORIG_RAX: 0000000000000128
[  247.638977] [     T48] RAX: ffffffffffffffda RBX: 0000000000004000 RCX: 00007f1cae028e32
[  247.639260] [     T48] RDX: 0000000000000002 RSI: 00007f1cad797e20 RDI: 0000000000000005
[  247.639534] [     T48] RBP: 00007f1cad797dc0 R08: 0000000000000000 R09: 0000000000000000
[  247.639810] [     T48] R10: 0000000000004000 R11: 0000000000000246 R12: 00007f1c9c001000
[  247.640082] [     T48] R13: 00007fff0eee6560 R14: 0000000000000000 R15: 00007fff0eee6667
[  247.640368] [     T48]  </TASK>
[  247.640636] [     T48] INFO: task dio-fallback-ra:1089 is blocked on an rw-semaphore, but the owner is not found.
[  247.640926] [     T48] INFO: task dio-fallback-ra:1091 blocked for more than 122 seconds.
[  247.641225] [     T48]       Not tainted 7.3.0-rc2+ #715
[  247.641512] [     T48] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[  247.641813] [     T48] task:dio-fallback-ra state:D stack:0     pid:1091  tgid:1087  ppid:1042   task_flags:0x400040 flags:0x00080000
[  247.642139] [     T48] Call Trace:
[  247.642446] [     T48]  <TASK>
[  247.642753] [     T48]  __schedule+0x1fcc/0x6ef0
[  247.643060] [     T48]  ? blk_stat_add+0xbc/0x450
[  247.643380] [     T48]  ? rcu_is_watching+0x12/0xb0
[  247.643685] [     T48]  ? lock_release+0x285/0x340
[  247.644028] [     T48]  ? percpu_ref_put_many.constprop.0+0x39/0x1d0
[  247.644348] [     T48]  ? __pfx___schedule+0x10/0x10
[  247.644655] [     T48]  ? lock_acquire+0x2f5/0x340
[  247.644964] [     T48]  ? lock_acquire+0x2f5/0x340
[  247.645318] [     T48]  ? rcu_is_watching+0x12/0xb0
[  247.645624] [     T48]  ? lock_release+0x285/0x340
[  247.645973] [     T48]  schedule+0xe3/0x2e0
[  247.646357] [     T48]  ? rwsem_down_read_slowpath+0x31e/0x10d0
[  247.646687] [     T48]  schedule_preempt_disabled+0x14/0x30
[  247.647028] [     T48]  rwsem_down_read_slowpath+0x61a/0x10d0
[  247.647354] [     T48]  ? __pfx_rwsem_down_read_slowpath+0x10/0x10
[  247.647669] [     T48]  ? blk_mq_dispatch_queue_requests+0x51a/0x7e0
[  247.648017] [     T48]  ? blk_mq_dispatch_queue_requests+0x576/0x7e0
[  247.648342] [     T48]  ? __pfx_folio_batch_move_lru+0x10/0x10
[  247.648658] [     T48]  ? __pfx___might_resched+0x10/0x10
[  247.649023] [     T48]  down_read+0xfa/0x540
[  247.649345] [     T48]  ? __pfx_down_read+0x10/0x10
[  247.649657] [     T48]  ? __blk_flush_plug+0x29f/0x540
[  247.649971] [     T48]  ? selinux_file_permission+0x3c0/0x530
[  247.650296] [     T48]  ? security_file_permission+0x44/0xc0
[  247.650610] [     T48]  blkdev_read_iter+0x15a/0x4e0
[  247.650927] [     T48]  ? rw_verify_area+0x77/0x650
[  247.651254] [     T48]  vfs_read+0x6f4/0xb30
[  247.651569] [     T48]  ? blk_finish_plug+0x52/0xc0
[  247.651887] [     T48]  ? __pfx_vfs_read+0x10/0x10
[  247.652243] [     T48]  ? lock_release+0x285/0x340
[  247.652560] [     T48]  __x64_sys_pread64+0x18d/0x250
[  247.652910] [     T48]  ? __pfx___x64_sys_pread64+0x10/0x10
[  247.653237] [     T48]  ? rcu_is_watching+0x12/0xb0
[  247.653552] [     T48]  ? rcu_is_watching+0x12/0xb0
[  247.653864] [     T48]  ? do_syscall_64+0x7f/0x690
[  247.654182] [     T48]  do_syscall_64+0xf2/0x690
[  247.654485] [     T48]  ? force_page_cache_ra+0x21c/0x370
[  247.654857] [     T48]  ? generic_fadvise+0x35c/0x660
[  247.655180] [     T48]  ? __pfx_generic_fadvise+0x10/0x10
[  247.655488] [     T48]  ? rcu_is_watching+0x12/0xb0
[  247.655786] [     T48]  ? lock_release+0x285/0x340
[  247.656078] [     T48]  ? __fget_files+0x1d8/0x440
[  247.656381] [     T48]  ? fput+0x1b/0xa0
[  247.656658] [     T48]  ? fput+0x48/0xa0
[  247.656969] [     T48]  ? rcu_is_watching+0x12/0xb0
[  247.657247] [     T48]  ? do_syscall_64+0x1ff/0x690
[  247.657507] [     T48]  ? trace_hardirqs_on_prepare+0x139/0x180
[  247.657806] [     T48]  ? do_syscall_64+0x21d/0x690
[  247.658049] [     T48]  ? vfs_read+0x3b6/0xb30
[  247.658305] [     T48]  ? vfs_read+0x3b6/0xb30
[  247.658547] [     T48]  ? lock_acquire+0x1b1/0x340
[  247.658792] [     T48]  ? __pfx_vfs_read+0x10/0x10
[  247.659034] [     T48]  ? __fget_files+0x1ce/0x440
[  247.659288] [     T48]  ? lock_release+0x1b5/0x340
[  247.659877] [     T48]  ? fput+0x1b/0xa0
[  247.660117] [     T48]  ? fput+0x48/0xa0
[  247.660364] [     T48]  ? __x64_sys_pread64+0x1ad/0x250
[  247.660604] [     T48]  ? __pfx___x64_sys_pread64+0x10/0x10
[  247.660884] [     T48]  ? rcu_is_watching+0x12/0xb0
[  247.661123] [     T48]  ? do_syscall_64+0x1ff/0x690
[  247.661407] [     T48]  ? trace_hardirqs_on_prepare+0x139/0x180
[  247.661649] [     T48]  ? do_syscall_64+0x21d/0x690
[  247.661939] [     T48]  ? rcu_is_watching+0x12/0xb0
[  247.662188] [     T48]  ? do_syscall_64+0x2d/0x690
[  247.662423] [     T48]  ? trace_hardirqs_off_finish+0x139/0x180
[  247.662658] [     T48]  ? lockdep_hardirqs_off+0xaf/0xf0
[  247.662900] [     T48]  ? do_syscall_64+0x60/0x690
[  247.663147] [     T48]  entry_SYSCALL_64_after_hwframe+0x76/0x7e
[  247.663388] [     T48] RIP: 0033:0x7f1cae028e32
[  247.663626] [     T48] RSP: 002b:00007f1ca45fedc8 EFLAGS: 00000246 ORIG_RAX: 0000000000000011
[  247.663880] [     T48] RAX: ffffffffffffffda RBX: 0000000000002000 RCX: 00007f1cae028e32
[  247.664144] [     T48] RDX: 0000000000002000 RSI: 00007f1c94000880 RDI: 0000000000000006
[  247.664441] [     T48] RBP: 00007f1ca45fedf0 R08: 0000000000000000 R09: 0000000000000000
[  247.664695] [     T48] R10: 0000000000002000 R11: 0000000000000246 R12: 00007f1c94000880
[  247.664995] [     T48] R13: 00007fff0eee6540 R14: 0000000000000000 R15: 00007fff0eee6647
[  247.665298] [     T48]  </TASK>
[  247.665551] [     T48] INFO: task dio-fallback-ra:1091 is blocked on an rw-semaphore, but the owner is not found.
[  247.665869] [     T48] INFO: task dio-fallback-ra:1092 blocked for more than 122 seconds.
[  247.666143] [     T48]       Not tainted 7.3.0-rc2+ #715
[  247.666429] [     T48] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[  247.666719] [     T48] task:dio-fallback-ra state:D stack:0     pid:1092  tgid:1087  ppid:1042   task_flags:0x400140 flags:0x00080000
[  247.667021] [     T48] Call Trace:
[  247.667327] [     T48]  <TASK>
[  247.667661] [     T48]  __schedule+0x1fcc/0x6ef0
[  247.667966] [     T48]  ? rcu_is_watching+0x12/0xb0
[  247.668307] [     T48]  ? __pfx___schedule+0x10/0x10
[  247.668606] [     T48]  ? lock_acquire+0x2f5/0x340
[  247.668906] [     T48]  ? lock_acquire+0x2f5/0x340
[  247.669210] [     T48]  ? rcu_is_watching+0x12/0xb0
[  247.669503] [     T48]  ? lock_release+0x285/0x340
[  247.669800] [     T48]  schedule+0xe3/0x2e0
[  247.670092] [     T48]  ? _raw_spin_unlock_irq+0x34/0x50
[  247.670400] [     T48]  schedule_preempt_disabled+0x14/0x30
[  247.670700] [     T48]  rwsem_down_write_slowpath+0x4b3/0x1830
[  247.671006] [     T48]  ? __pfx_rwsem_down_write_slowpath+0x10/0x10
[  247.671323] [     T48]  ? rcu_is_watching+0x12/0xb0
[  247.671626] [     T48]  ? lock_release+0x285/0x340
[  247.671930] [     T48]  ? __pfx___mutex_lock+0x10/0x10
[  247.672245] [     T48]  ? __pfx___might_resched+0x10/0x10
[  247.672553] [     T48]  down_write+0x1be/0x1f0
[  247.672994] [     T48]  ? __pfx_down_write+0x10/0x10
[  247.673328] [     T48]  ? bdev_open+0x90c/0xc60
[  247.673644] [     T48]  set_blocksize+0x1ad/0x460
[  247.673956] [     T48]  ? bdev_file_open_by_dev+0xe6/0x170
[  247.674279] [     T48]  blkdev_bszset+0x142/0x1e0
[  247.674592] [     T48]  ? __pfx_blkdev_bszset+0x10/0x10
[  247.674937] [     T48]  ? lock_acquire+0x2f5/0x340
[  247.675263] [     T48]  ? rcu_is_watching+0x12/0xb0
[  247.675583] [     T48]  blkdev_ioctl+0x4e9/0x5e0
[  247.675932] [     T48]  ? __pfx_blkdev_ioctl+0x10/0x10
[  247.676257] [     T48]  __x64_sys_ioctl+0x138/0x1c0
[  247.676569] [     T48]  do_syscall_64+0xf2/0x690
[  247.676882] [     T48]  ? rcu_read_unlock+0x1c/0x60
[  247.677268] [     T48]  ? handle_mm_fault+0x35f/0x660
[  247.677608] [     T48]  ? rcu_is_watching+0x12/0xb0
[  247.677952] [     T48]  ? lock_release+0x285/0x340
[  247.678315] [     T48]  ? rcu_is_watching+0x12/0xb0
[  247.678667] [     T48]  ? do_user_addr_fault+0x52e/0xaf0
[  247.679008] [     T48]  ? rcu_is_watching+0x12/0xb0
[  247.679350] [     T48]  ? do_syscall_64+0x2d/0x690
[  247.679690] [     T48]  ? trace_hardirqs_off_finish+0x139/0x180
[  247.680034] [     T48]  ? trace_hardirqs_off+0x14f/0x1a0
[  247.680369] [     T48]  ? do_syscall_64+0x60/0x690
[  247.680690] [     T48]  entry_SYSCALL_64_after_hwframe+0x76/0x7e
[  247.681055] [     T48] RIP: 0033:0x7f1cae0a073d
[  247.681399] [     T48] RSP: 002b:00007f1ca7ffedd0 EFLAGS: 00000246 ORIG_RAX: 0000000000000010
[  247.681773] [     T48] RAX: ffffffffffffffda RBX: 0000000000000200 RCX: 00007f1cae0a073d
[  247.682088] [     T48] RDX: 00007f1ca7ffee3c RSI: 0000000040081271 RDI: 0000000000000003
[  247.682386] [     T48] RBP: 00007f1ca7ffee20 R08: 0000000000000000 R09: 00007f1ca7fff6c0
[  247.682710] [     T48] R10: 0000000000000008 R11: 0000000000000246 R12: 0000000000000021
[  247.683016] [     T48] R13: 00007fff0eee6540 R14: 0000000000000000 R15: 00007fff0eee6647
[  247.683330] [     T48]  </TASK>
[  247.683643] [     T48] INFO: task dio-fallback-ra:1092 is blocked on an rw-semaphore, but the owner is not found.
[  247.683985] [     T48] INFO: lockdep is turned off.
[  370.473396] [     T48] INFO: task dio-fallback-ra:1088 blocked for more than 245 seconds.
[  370.475679] [     T48]       Not tainted 7.3.0-rc2+ #715
[  370.477650] [     T48] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[  370.479255] [     T48] task:dio-fallback-ra state:D stack:0     pid:1088  tgid:1087  ppid:1042   task_flags:0x400040 flags:0x00080000
[  370.479690] [     T48] Call Trace:
[  370.480085] [     T48]  <TASK>
[  370.480454] [     T48]  __schedule+0x1fcc/0x6ef0
[  370.480892] [     T48]  ? asm_sysvec_apic_timer_interrupt+0x16/0x20
[  370.481241] [     T48]  ? trace_hardirqs_on+0x14/0x190
[  370.481575] [     T48]  ? __pfx___schedule+0x10/0x10
[  370.481956] [     T48]  ? lock_acquire+0x2f5/0x340
[  370.482329] [     T48]  ? lock_acquire+0x2f5/0x340
[  370.482679] [     T48]  ? rcu_is_watching+0x12/0xb0
[  370.483565] [     T48]  ? lock_release+0x285/0x340
[  370.483890] [     T48]  schedule+0xe3/0x2e0
[  370.484268] [     T48]  ? rwsem_down_read_slowpath+0x31e/0x10d0
[  370.484619] [     T48]  schedule_preempt_disabled+0x14/0x30
[  370.484939] [     T48]  rwsem_down_read_slowpath+0x61a/0x10d0
[  370.485317] [     T48]  ? __pfx_rwsem_down_read_slowpath+0x10/0x10
[  370.485683] [     T48]  ? __pfx___might_resched+0x10/0x10
[  370.486011] [     T48]  down_read+0xfa/0x540
[  370.486382] [     T48]  ? __pfx_down_read+0x10/0x10
[  370.486738] [     T48]  ? filemap_check_errors+0x58/0x100
[  370.487087] [     T48]  blkdev_write_iter+0x928/0xf40
[  370.487701] [     T48]  do_iter_readv_writev+0x48a/0xb80
[  370.488224] [     T48]  ? __pfx_folio_add_lru_vma+0x10/0x10
[  370.488586] [     T48]  ? rcu_read_unlock+0x1c/0x60
[  370.488912] [     T48]  ? __folio_mod_stat+0x4c/0x190
[  370.489301] [     T48]  ? __pfx_do_iter_readv_writev+0x10/0x10
[  370.489661] [     T48]  ? folio_add_new_anon_rmap+0x34f/0xd40
[  370.490161] [     T48]  ? selinux_file_permission+0x3c0/0x530
[  370.490543] [     T48]  ? security_file_permission+0x44/0xc0
[  370.490874] [     T48]  ? rw_verify_area+0x77/0x650
[  370.491273] [     T48]  vfs_writev+0x27d/0xbc0
[  370.491634] [     T48]  ? __pfx_vfs_writev+0x10/0x10
[  370.491964] [     T48]  ? __handle_mm_fault+0x1584/0x1fb0
[  370.492350] [     T48]  ? lock_acquire+0x1b1/0x340
[  370.492713] [     T48]  ? __fget_files+0x1d8/0x440
[  370.493068] [     T48]  ? do_pwritev+0x18a/0x300
[  370.493440] [     T48]  do_pwritev+0x18a/0x300
[  370.493812] [     T48]  ? __pfx_do_pwritev+0x10/0x10
[  370.494186] [     T48]  ? lock_release+0x1b5/0x340
[  370.494532] [     T48]  ? rcu_is_watching+0x12/0xb0
[  370.494844] [     T48]  do_syscall_64+0xf2/0x690
[  370.495208] [     T48]  ? do_syscall_64+0x2d/0x690
[  370.495547] [     T48]  ? trace_hardirqs_off_finish+0x139/0x180
[  370.495856] [     T48]  ? lockdep_hardirqs_off+0xaf/0xf0
[  370.496326] [     T48]  ? do_syscall_64+0x60/0x690
[  370.496639] [     T48]  entry_SYSCALL_64_after_hwframe+0x76/0x7e
[  370.496933] [     T48] RIP: 0033:0x7f1cae028e32
[  370.497283] [     T48] RSP: 002b:00007f1cadf98d98 EFLAGS: 00000246 ORIG_RAX: 0000000000000128
[  370.497644] [     T48] RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007f1cae028e32
[  370.497965] [     T48] RDX: 0000000000000002 RSI: 00007f1cadf98e20 RDI: 0000000000000004
[  370.498308] [     T48] RBP: 00007f1cadf98dc0 R08: 0000000000000000 R09: 0000000000000000
[  370.498614] [     T48] R10: 0000000000000000 R11: 0000000000000246 R12: 00007f1ca8001000
[  370.498891] [     T48] R13: 00007fff0eee6560 R14: 0000000000000000 R15: 00007fff0eee6667
[  370.499233] [     T48]  </TASK>
[  370.499539] [     T48] INFO: task dio-fallback-ra:1088 is blocked on an rw-semaphore, but the owner is not found.
[  370.499833] [     T48] INFO: task dio-fallback-ra:1089 blocked for more than 245 seconds.
[  370.500182] [     T48]       Not tainted 7.3.0-rc2+ #715
[  370.500541] [     T48] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[  370.500849] [     T48] task:dio-fallback-ra state:D stack:0     pid:1089  tgid:1087  ppid:1042   task_flags:0x400040 flags:0x00080000
[  370.501230] [     T48] Call Trace:
[  370.501622] [     T48]  <TASK>
[  370.501970] [     T48]  __schedule+0x1fcc/0x6ef0
[  370.502371] [     T48]  ? __pfx_bio_wait_end_io+0x10/0x10
[  370.502711] [     T48]  ? __pfx___schedule+0x10/0x10
[  370.503052] [     T48]  ? lock_acquire+0x2f5/0x340
[  370.503416] [     T48]  ? lock_acquire+0x2f5/0x340
[  370.503759] [     T48]  ? rcu_is_watching+0x12/0xb0
[  370.504097] [     T48]  ? lock_release+0x285/0x340
[  370.504437] [     T48]  schedule+0xe3/0x2e0
[  370.504776] [     T48]  ? rwsem_down_read_slowpath+0x31e/0x10d0
[  370.505113] [     T48]  schedule_preempt_disabled+0x14/0x30
[  370.505480] [     T48]  rwsem_down_read_slowpath+0x61a/0x10d0
[  370.505841] [     T48]  ? __pfx_rwsem_down_read_slowpath+0x10/0x10
[  370.506224] [     T48]  ? blkdev_direct_IO+0x86f/0x1c50
[  370.506576] [     T48]  ? __pfx___might_resched+0x10/0x10
[  370.506938] [     T48]  down_read+0xfa/0x540
[  370.507302] [     T48]  ? __pfx_down_read+0x10/0x10
[  370.507653] [     T48]  ? filemap_check_errors+0x58/0x100
[  370.508052] [     T48]  blkdev_write_iter+0x882/0xf40
[  370.508582] [     T48]  do_iter_readv_writev+0x48a/0xb80
[  370.509013] [     T48]  ? vfs_writev+0x831/0xbc0
[  370.509401] [     T48]  ? __pfx_do_iter_readv_writev+0x10/0x10
[  370.510181] [     T48]  ? selinux_file_permission+0x3c0/0x530
[  370.510538] [     T48]  ? security_file_permission+0x44/0xc0
[  370.510887] [     T48]  ? rw_verify_area+0x77/0x650
[  370.511266] [     T48]  vfs_writev+0x27d/0xbc0
[  370.511627] [     T48]  ? __pfx_vfs_writev+0x10/0x10
[  370.511948] [     T48]  ? __fget_files+0x4f/0x440
[  370.512325] [     T48]  ? __fget_files+0x1d8/0x440
[  370.512676] [     T48]  ? do_pwritev+0x18a/0x300
[  370.512990] [     T48]  do_pwritev+0x18a/0x300
[  370.513356] [     T48]  ? __pfx_do_pwritev+0x10/0x10
[  370.513701] [     T48]  ? lock_release+0x1b5/0x340
[  370.514036] [     T48]  ? rcu_is_watching+0x12/0xb0
[  370.514376] [     T48]  do_syscall_64+0xf2/0x690
[  370.514724] [     T48]  ? rcu_read_unlock+0x17/0x60
[  370.515060] [     T48]  ? lock_release+0x1b5/0x340
[  370.515392] [     T48]  ? __handle_mm_fault+0x1584/0x1fb0
[  370.515725] [     T48]  ? __lock_acquire+0x54a/0x2490
[  370.516047] [     T48]  ? __pfx___css_rstat_updated+0x10/0x10
[  370.516374] [     T48]  ? lock_acquire+0x1b1/0x340
[  370.516712] [     T48]  ? rcu_is_watching+0x12/0xb0
[  370.517009] [     T48]  ? count_memcg_events+0x107/0x4c0
[  370.517347] [     T48]  ? find_held_lock+0x2b/0x80
[  370.517661] [     T48]  ? rcu_read_unlock+0x17/0x60
[  370.517937] [     T48]  ? lock_release+0x1b5/0x340
[  370.518259] [     T48]  ? find_held_lock+0x2b/0x80
[  370.518555] [     T48]  ? exc_page_fault+0x94/0x140
[  370.518810] [     T48]  ? lock_release+0x1b5/0x340
[  370.519079] [     T48]  ? rcu_is_watching+0x12/0xb0
[  370.519384] [     T48]  ? do_syscall_64+0x2d/0x690
[  370.519662] [     T48]  ? trace_hardirqs_off_finish+0x139/0x180
[  370.519911] [     T48]  ? lockdep_hardirqs_off+0xaf/0xf0
[  370.520212] [     T48]  ? do_syscall_64+0x60/0x690
[  370.520493] [     T48]  entry_SYSCALL_64_after_hwframe+0x76/0x7e
[  370.520750] [     T48] RIP: 0033:0x7f1cae028e32
[  370.521009] [     T48] RSP: 002b:00007f1cad797d98 EFLAGS: 00000246 ORIG_RAX: 0000000000000128
[  370.521333] [     T48] RAX: ffffffffffffffda RBX: 0000000000004000 RCX: 00007f1cae028e32
[  370.521635] [     T48] RDX: 0000000000000002 RSI: 00007f1cad797e20 RDI: 0000000000000005
[  370.521908] [     T48] RBP: 00007f1cad797dc0 R08: 0000000000000000 R09: 0000000000000000
[  370.522230] [     T48] R10: 0000000000004000 R11: 0000000000000246 R12: 00007f1c9c001000
[  370.522540] [     T48] R13: 00007fff0eee6560 R14: 0000000000000000 R15: 00007fff0eee6667
[  370.522815] [     T48]  </TASK>
[  370.523114] [     T48] INFO: task dio-fallback-ra:1089 is blocked on an rw-semaphore, but the owner is not found.
[  370.523427] [     T48] INFO: task dio-fallback-ra:1091 blocked for more than 245 seconds.
[  370.523746] [     T48]       Not tainted 7.3.0-rc2+ #715
[  370.524053] [     T48] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[  370.524380] [     T48] task:dio-fallback-ra state:D stack:0     pid:1091  tgid:1087  ppid:1042   task_flags:0x400040 flags:0x00080000
[  370.524727] [     T48] Call Trace:
[  370.525060] [     T48]  <TASK>
[  370.525391] [     T48]  __schedule+0x1fcc/0x6ef0
[  370.525728] [     T48]  ? blk_stat_add+0xbc/0x450
[  370.526058] [     T48]  ? rcu_is_watching+0x12/0xb0
[  370.526390] [     T48]  ? lock_release+0x285/0x340
[  370.526725] [     T48]  ? percpu_ref_put_many.constprop.0+0x39/0x1d0
[  370.527058] [     T48]  ? __pfx___schedule+0x10/0x10
[  370.527393] [     T48]  ? lock_acquire+0x2f5/0x340
[  370.527731] [     T48]  ? lock_acquire+0x2f5/0x340
[  370.528063] [     T48]  ? rcu_is_watching+0x12/0xb0
[  370.528396] [     T48]  ? lock_release+0x285/0x340
[  370.528737] [     T48]  schedule+0xe3/0x2e0
[  370.529061] [     T48]  ? rwsem_down_read_slowpath+0x31e/0x10d0
[  370.529483] [     T48]  schedule_preempt_disabled+0x14/0x30
[  370.529840] [     T48]  rwsem_down_read_slowpath+0x61a/0x10d0
[  370.530206] [     T48]  ? __pfx_rwsem_down_read_slowpath+0x10/0x10
[  370.530590] [     T48]  ? blk_mq_dispatch_queue_requests+0x51a/0x7e0
[  370.530905] [     T48]  ? blk_mq_dispatch_queue_requests+0x576/0x7e0
[  370.531290] [     T48]  ? __pfx_folio_batch_move_lru+0x10/0x10
[  370.531649] [     T48]  ? __pfx___might_resched+0x10/0x10
[  370.531992] [     T48]  down_read+0xfa/0x540
[  370.532351] [     T48]  ? __pfx_down_read+0x10/0x10
[  370.532693] [     T48]  ? __blk_flush_plug+0x29f/0x540
[  370.533008] [     T48]  ? selinux_file_permission+0x3c0/0x530
[  370.533367] [     T48]  ? security_file_permission+0x44/0xc0
[  370.533712] [     T48]  blkdev_read_iter+0x15a/0x4e0
[  370.534047] [     T48]  ? rw_verify_area+0x77/0x650
[  370.534389] [     T48]  vfs_read+0x6f4/0xb30
[  370.534704] [     T48]  ? blk_finish_plug+0x52/0xc0
[  370.535042] [     T48]  ? __pfx_vfs_read+0x10/0x10
[  370.535417] [     T48]  ? lock_release+0x285/0x340
[  370.535765] [     T48]  __x64_sys_pread64+0x18d/0x250
[  370.536104] [     T48]  ? __pfx___x64_sys_pread64+0x10/0x10
[  370.536520] [     T48]  ? rcu_is_watching+0x12/0xb0
[  370.536857] [     T48]  ? rcu_is_watching+0x12/0xb0
[  370.537224] [     T48]  ? do_syscall_64+0x7f/0x690
[  370.537532] [     T48]  do_syscall_64+0xf2/0x690
[  370.537834] [     T48]  ? force_page_cache_ra+0x21c/0x370
[  370.538191] [     T48]  ? generic_fadvise+0x35c/0x660
[  370.538522] [     T48]  ? __pfx_generic_fadvise+0x10/0x10
[  370.538817] [     T48]  ? rcu_is_watching+0x12/0xb0
[  370.539139] [     T48]  ? lock_release+0x285/0x340
[  370.539459] [     T48]  ? __fget_files+0x1d8/0x440
[  370.539778] [     T48]  ? fput+0x1b/0xa0
[  370.540079] [     T48]  ? fput+0x48/0xa0
[  370.540375] [     T48]  ? rcu_is_watching+0x12/0xb0
[  370.540671] [     T48]  ? do_syscall_64+0x1ff/0x690
[  370.540931] [     T48]  ? trace_hardirqs_on_prepare+0x139/0x180
[  370.541234] [     T48]  ? do_syscall_64+0x21d/0x690
[  370.541509] [     T48]  ? vfs_read+0x3b6/0xb30
[  370.541753] [     T48]  ? vfs_read+0x3b6/0xb30
[  370.541994] [     T48]  ? lock_acquire+0x1b1/0x340
[  370.542289] [     T48]  ? __pfx_vfs_read+0x10/0x10
[  370.542564] [     T48]  ? __fget_files+0x1ce/0x440
[  370.542806] [     T48]  ? lock_release+0x1b5/0x340
[  370.543070] [     T48]  ? fput+0x1b/0xa0
[  370.543333] [     T48]  ? fput+0x48/0xa0
[  370.543607] [     T48]  ? __x64_sys_pread64+0x1ad/0x250
[  370.543878] [     T48]  ? __pfx___x64_sys_pread64+0x10/0x10
[  370.544144] [     T48]  ? rcu_is_watching+0x12/0xb0
[  370.544442] [     T48]  ? do_syscall_64+0x1ff/0x690
[  370.544681] [     T48]  ? trace_hardirqs_on_prepare+0x139/0x180
[  370.544922] [     T48]  ? do_syscall_64+0x21d/0x690
[  370.545210] [     T48]  ? rcu_is_watching+0x12/0xb0
[  370.545480] [     T48]  ? do_syscall_64+0x2d/0x690
[  370.545714] [     T48]  ? trace_hardirqs_off_finish+0x139/0x180
[  370.545949] [     T48]  ? lockdep_hardirqs_off+0xaf/0xf0
[  370.546249] [     T48]  ? do_syscall_64+0x60/0x690
[  370.546517] [     T48]  entry_SYSCALL_64_after_hwframe+0x76/0x7e
[  370.546757] [     T48] RIP: 0033:0x7f1cae028e32
[  370.546995] [     T48] RSP: 002b:00007f1ca45fedc8 EFLAGS: 00000246 ORIG_RAX: 0000000000000011
[  370.547300] [     T48] RAX: ffffffffffffffda RBX: 0000000000002000 RCX: 00007f1cae028e32
[  370.547627] [     T48] RDX: 0000000000002000 RSI: 00007f1c94000880 RDI: 0000000000000006
[  370.547912] [     T48] RBP: 00007f1ca45fedf0 R08: 0000000000000000 R09: 0000000000000000
[  370.548218] [     T48] R10: 0000000000002000 R11: 0000000000000246 R12: 00007f1c94000880
[  370.548506] [     T48] R13: 00007fff0eee6540 R14: 0000000000000000 R15: 00007fff0eee6647
[  370.548766] [     T48]  </TASK>
[  370.549041] [     T48] INFO: task dio-fallback-ra:1091 is blocked on an rw-semaphore, but the owner is not found.
[  370.549342] [     T48] INFO: task dio-fallback-ra:1092 blocked for more than 245 seconds.
[  370.549635] [     T48]       Not tainted 7.3.0-rc2+ #715
[  370.549985] [     T48] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[  370.550319] [     T48] task:dio-fallback-ra state:D stack:0     pid:1092  tgid:1087  ppid:1042   task_flags:0x400140 flags:0x00080000
[  370.550642] [     T48] Call Trace:
[  370.550994] [     T48]  <TASK>
[  370.551339] [     T48]  __schedule+0x1fcc/0x6ef0
[  370.551672] [     T48]  ? rcu_is_watching+0x12/0xb0
[  370.552022] [     T48]  ? __pfx___schedule+0x10/0x10
[  370.552349] [     T48]  ? lock_acquire+0x2f5/0x340
[  370.552708] [     T48]  ? lock_acquire+0x2f5/0x340
[  370.553001] [     T48]  ? rcu_is_watching+0x12/0xb0
[  370.553342] [     T48]  ? lock_release+0x285/0x340
[  370.553667] [     T48]  schedule+0xe3/0x2e0
[  370.553961] [     T48]  ? _raw_spin_unlock_irq+0x34/0x50
[  370.554330] [     T48]  schedule_preempt_disabled+0x14/0x30
[  370.554645] [     T48]  rwsem_down_write_slowpath+0x4b3/0x1830
[  370.554949] [     T48]  ? __pfx_rwsem_down_write_slowpath+0x10/0x10
[  370.555308] [     T48]  ? rcu_is_watching+0x12/0xb0
[  370.555643] [     T48]  ? lock_release+0x285/0x340
[  370.555943] [     T48]  ? __pfx___mutex_lock+0x10/0x10
[  370.556300] [     T48]  ? __pfx___might_resched+0x10/0x10
[  370.556639] [     T48]  down_write+0x1be/0x1f0
[  370.556974] [     T48]  ? __pfx_down_write+0x10/0x10
[  370.557328] [     T48]  ? bdev_open+0x90c/0xc60
[  370.557667] [     T48]  set_blocksize+0x1ad/0x460
[  370.557975] [     T48]  ? bdev_file_open_by_dev+0xe6/0x170
[  370.558335] [     T48]  blkdev_bszset+0x142/0x1e0
[  370.558678] [     T48]  ? __pfx_blkdev_bszset+0x10/0x10
[  370.558991] [     T48]  ? lock_acquire+0x2f5/0x340
[  370.559374] [     T48]  ? rcu_is_watching+0x12/0xb0
[  370.559701] [     T48]  blkdev_ioctl+0x4e9/0x5e0
[  370.560016] [     T48]  ? __pfx_blkdev_ioctl+0x10/0x10
[  370.560411] [     T48]  __x64_sys_ioctl+0x138/0x1c0
[  370.560754] [     T48]  do_syscall_64+0xf2/0x690
[  370.561086] [     T48]  ? rcu_read_unlock+0x1c/0x60
[  370.561424] [     T48]  ? handle_mm_fault+0x35f/0x660
[  370.561767] [     T48]  ? rcu_is_watching+0x12/0xb0
[  370.562100] [     T48]  ? lock_release+0x285/0x340
[  370.562451] [     T48]  ? rcu_is_watching+0x12/0xb0
[  370.562806] [     T48]  ? do_user_addr_fault+0x52e/0xaf0
[  370.563149] [     T48]  ? rcu_is_watching+0x12/0xb0
[  370.563471] [     T48]  ? do_syscall_64+0x2d/0x690
[  370.563765] [     T48]  ? trace_hardirqs_off_finish+0x139/0x180
[  370.564087] [     T48]  ? trace_hardirqs_off+0x14f/0x1a0
[  370.564406] [     T48]  ? do_syscall_64+0x60/0x690
[  370.564729] [     T48]  entry_SYSCALL_64_after_hwframe+0x76/0x7e
[  370.565044] [     T48] RIP: 0033:0x7f1cae0a073d
[  370.565358] [     T48] RSP: 002b:00007f1ca7ffedd0 EFLAGS: 00000246 ORIG_RAX: 0000000000000010
[  370.565679] [     T48] RAX: ffffffffffffffda RBX: 0000000000000200 RCX: 00007f1cae0a073d
[  370.565968] [     T48] RDX: 00007f1ca7ffee3c RSI: 0000000040081271 RDI: 0000000000000003
[  370.566311] [     T48] RBP: 00007f1ca7ffee20 R08: 0000000000000000 R09: 00007f1ca7fff6c0
[  370.566623] [     T48] R10: 0000000000000008 R11: 0000000000000246 R12: 0000000000000021
[  370.566897] [     T48] R13: 00007fff0eee6540 R14: 0000000000000000 R15: 00007fff0eee6647
[  370.567218] [     T48]  </TASK>
[  370.567514] [     T48] INFO: task dio-fallback-ra:1092 is blocked on an rw-semaphore, but the owner is not found.
[  370.567808] [     T48] INFO: lockdep is turned off.

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [PATCH 2/2] block: take the inode lock for buffered write fallback
  2026-10-27  4:24   ` Shin'ichiro Kawasaki
@ 2026-10-27  6:13     ` Julian Sun
  2026-10-27  7:29       ` Shin'ichiro Kawasaki
  0 siblings, 1 reply; 7+ messages in thread
From: Julian Sun @ 2026-10-27  6:13 UTC (permalink / raw)
  To: Shin'ichiro Kawasaki
  Cc: linux-block, jack, axboe, viro, mcgrof, djwong, tz2294

On 10/27/26 12:24 PM, Shin'ichiro Kawasaki wrote:
> CC+: Tal Zassman,
> 
> On Oct 15, 2026 / 17:11, Julian Sun wrote:
>> Commit c0e473a0d226 ("block: fix race between set_blocksize and read
>> paths") added shared inode locking around ordinary buffered writes,
>> but missed the direct-I/O fallback path.
>>
>> Move the lock into blkdev_buffered_write() so both paths serialize
>> with set_blocksize().
>>
>> Also correct the comment claiming that this path takes invalidate_lock;
>> it only takes i_rwsem.
>>
>> Fixes: c0e473a0d226 ("block: fix race between set_blocksize and read paths")
>> Signed-off-by: Julian Sun <sunjunchao@bytedance.com>
> 
> When I checking the new blktests test case block/050 by Tal Zussman [1], I found
> the test case hangs [2]. I bisected and this kernel change triggers the hang.
> 
> Julian, may I ask you to take a look in the hang?

Hi Shin'ichiro,

Thanks for the report. This issue should be fixed by:
https://lore.kernel.org/all/20261022062513.77372-1-mzx199711@gmail.com/

Could you please apply that patch and rerun block/050?

> 
> [1] https://lore.kernel.org/linux-block/20261015-blkdev-fixes-tests-v3-1-b6ffa6410514@columbia.edu/
> 

Thanks,
-- 
Julian Sun <sunjunchao@bytedance.com>

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [PATCH 2/2] block: take the inode lock for buffered write fallback
  2026-10-27  6:13     ` Julian Sun
@ 2026-10-27  7:29       ` Shin'ichiro Kawasaki
  0 siblings, 0 replies; 7+ messages in thread
From: Shin'ichiro Kawasaki @ 2026-10-27  7:29 UTC (permalink / raw)
  To: Julian Sun; +Cc: linux-block, jack, axboe, viro, mcgrof, djwong, tz2294

On Oct 27, 2026 / 14:13, Julian Sun wrote:
...
> > When I checking the new blktests test case block/050 by Tal Zussman [1], I found
> > the test case hangs [2]. I bisected and this kernel change triggers the hang.
> > 
> > Julian, may I ask you to take a look in the hang?
> 
> Hi Shin'ichiro,
> 
> Thanks for the report. This issue should be fixed by:
> https://lore.kernel.org/all/20261022062513.77372-1-mzx199711@gmail.com/
> 
> Could you please apply that patch and rerun block/050?

Thanks, the patch avoided the block/050 hang. Good.

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [PATCH 0/2] block: fix NOWAIT handling and locking in buffered write fallback
  2026-10-15  9:11 [PATCH 0/2] block: fix NOWAIT handling and locking in buffered write fallback Julian Sun
  2026-10-15  9:11 ` [PATCH 1/2] block: reject NOWAIT " Julian Sun
  2026-10-15  9:11 ` [PATCH 2/2] block: take the inode lock for " Julian Sun
@ 2026-10-15 12:41 ` Jens Axboe
  2 siblings, 0 replies; 7+ messages in thread
From: Jens Axboe @ 2026-10-15 12:41 UTC (permalink / raw)
  To: linux-block, Julian Sun; +Cc: jack, viro, mcgrof, djwong


On Thu, 15 Oct 2026 17:11:03 +0800, Julian Sun wrote:
> This series fixes two issues in buffered fallback for direct writes to
> block devices.
> 
> 1. NOWAIT requests can fall back to buffered I/O and wait for writeback.
> 2. The fallback path lacks shared inode locking against set_blocksize().
> 
> Julian Sun (2):
>   block: reject NOWAIT buffered write fallback
>   block: take the inode lock for buffered write fallback
> 
> [...]

Applied, thanks!

[1/2] block: reject NOWAIT buffered write fallback
      commit: a08bf4b9f87898628e70882430b8dd4260730dba
[2/2] block: take the inode lock for buffered write fallback
      commit: 1b1be9ea4e34ccede1e7cf9ff2b6317328a5b6ad

Best regards,
-- 
Jens Axboe




^ permalink raw reply	[flat|nested] 7+ messages in thread

end of thread, other threads:[~2026-10-27  7:29 UTC | newest]

Thread overview: 7+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-10-15  9:11 [PATCH 0/2] block: fix NOWAIT handling and locking in buffered write fallback Julian Sun
2026-10-15  9:11 ` [PATCH 1/2] block: reject NOWAIT " Julian Sun
2026-10-15  9:11 ` [PATCH 2/2] block: take the inode lock for " Julian Sun
2026-10-27  4:24   ` Shin'ichiro Kawasaki
2026-10-27  6:13     ` Julian Sun
2026-10-27  7:29       ` Shin'ichiro Kawasaki
2026-10-15 12:41 ` [PATCH 0/2] block: fix NOWAIT handling and locking in " Jens Axboe

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox