[syzbot ci] Re: loop: improve loop aio perf by IOCB_NOWAIT

[syzbot ci] Re: loop: improve loop aio perf by IOCB_NOWAIT

[syzbot ci]

syzbot ci has tested the following series

[v1] loop: improve loop aio perf by IOCB_NOWAIT
https://lore.kernel.org/all/20250928132927.3672537-1-ming.lei@redhat.com
* [PATCH V4 1/6] loop: add helper lo_cmd_nr_bvec()
* [PATCH V4 2/6] loop: add helper lo_rw_aio_prep()
* [PATCH V4 3/6] loop: add lo_submit_rw_aio()
* [PATCH V4 4/6] loop: move command blkcg/memcg initialization into loop_queue_work
* [PATCH V4 5/6] loop: try to handle loop aio command via NOWAIT IO first
* [PATCH V4 6/6] loop: add hint for handling aio via IOCB_NOWAIT

and found the following issue:
WARNING in lo_submit_rw_aio

Full report is available here:
https://ci.syzbot.org/series/0ffdb6b4-a5fe-48da-9473-d2a926e780bd

***

WARNING in lo_submit_rw_aio

tree:      torvalds
URL:       https://kernel.googlesource.com/pub/scm/linux/kernel/git/torvalds/linux
base:      07e27ad16399afcd693be20211b0dfae63e0615f
arch:      amd64
compiler:  Debian clang version 20.1.8 (++20250708063551+0c9f909b7976-1~exp1~20250708183702.136), Debian LLD 20.1.8
config:    https://ci.syzbot.org/builds/3aba003b-2400-4e88-9a31-c09ab4e41a84/config
C repro:   https://ci.syzbot.org/findings/dc97454c-d87b-41f5-a44a-7182e666cfd5/c_repro
syz repro: https://ci.syzbot.org/findings/dc97454c-d87b-41f5-a44a-7182e666cfd5/syz_repro

------------[ cut here ]------------
WARNING: CPU: 0 PID: 5958 at drivers/block/loop.c:907 loop_inc_blocking_writes drivers/block/loop.c:907 [inline]
WARNING: CPU: 0 PID: 5958 at drivers/block/loop.c:907 loop_queue_work+0xb3b/0xc30 drivers/block/loop.c:1005
Modules linked in:
CPU: 0 UID: 0 PID: 5958 Comm: udevd Not tainted syzkaller #0 PREEMPT(full) 
Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.2-debian-1.16.2-1 04/01/2014
RIP: 0010:loop_inc_blocking_writes drivers/block/loop.c:907 [inline]
RIP: 0010:loop_queue_work+0xb3b/0xc30 drivers/block/loop.c:1005
Code: 33 bf 08 00 00 00 4c 89 ea e8 c1 89 7e fb 4c 89 f7 48 83 c4 30 5b 41 5c 41 5d 41 5e 41 5f 5d e9 cb 87 71 05 e8 26 36 b4 fb 90 <0f> 0b 90 e9 4e fe ff ff e8 18 36 b4 fb 48 83 c5 18 48 89 e8 48 c1
RSP: 0018:ffffc9000340ef38 EFLAGS: 00010093
RAX: ffffffff860b776a RBX: ffff88802187c000 RCX: ffff88810cee3980
RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000000
RBP: 0000000000000000 R08: 0000000000000003 R09: 0000000000000004
R10: dffffc0000000000 R11: fffff52000681dc4 R12: ffff88802187c158
R13: ffff88802187c110 R14: ffff888021989460 R15: ffff888021989418
FS:  00007f649c5a4c80(0000) GS:ffff8880b8612000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 000056520d53f000 CR3: 0000000109e48000 CR4: 00000000000006f0
Call Trace:
 <TASK>
 lo_submit_rw_aio+0x493/0x620 drivers/block/loop.c:441
 lo_rw_aio_nowait drivers/block/loop.c:508 [inline]
 loop_queue_rq+0x64d/0x840 drivers/block/loop.c:2026
 __blk_mq_issue_directly block/blk-mq.c:2695 [inline]
 blk_mq_request_issue_directly+0x3c1/0x710 block/blk-mq.c:2782
 blk_mq_issue_direct+0x2a0/0x660 block/blk-mq.c:2803
 blk_mq_dispatch_queue_requests+0x621/0x800 block/blk-mq.c:2878
 blk_mq_flush_plug_list+0x432/0x550 block/blk-mq.c:2961
 __blk_flush_plug+0x3d3/0x4b0 block/blk-core.c:1220
 blk_finish_plug+0x5e/0x90 block/blk-core.c:1247
 read_pages+0x3b2/0x580 mm/readahead.c:173
 page_cache_ra_unbounded+0x6b0/0x7b0 mm/readahead.c:297
 do_page_cache_ra mm/readahead.c:327 [inline]
 force_page_cache_ra mm/readahead.c:356 [inline]
 page_cache_sync_ra+0x3b9/0xb10 mm/readahead.c:572
 filemap_get_pages+0x43c/0x1ea0 mm/filemap.c:2603
 filemap_read+0x3f6/0x11a0 mm/filemap.c:2712
 blkdev_read_iter+0x30a/0x440 block/fops.c:852
 new_sync_read fs/read_write.c:491 [inline]
 vfs_read+0x55a/0xa30 fs/read_write.c:572
 ksys_read+0x145/0x250 fs/read_write.c:715
 do_syscall_x64 arch/x86/entry/syscall_64.c:63 [inline]
 do_syscall_64+0xfa/0x3b0 arch/x86/entry/syscall_64.c:94
 entry_SYSCALL_64_after_hwframe+0x77/0x7f
RIP: 0033:0x7f649c116b6a
Code: 00 3d 00 00 41 00 75 0d 50 48 8d 3d 2d 08 0a 00 e8 ea 7d 01 00 31 c0 e9 07 ff ff ff 64 8b 04 25 18 00 00 00 85 c0 75 1b 0f 05 <48> 3d 00 f0 ff ff 76 6c 48 8b 15 8f a2 0d 00 f7 d8 64 89 02 48 83
RSP: 002b:00007ffd6597a888 EFLAGS: 00000246 ORIG_RAX: 0000000000000000
RAX: ffffffffffffffda RBX: 0000000000040000 RCX: 00007f649c116b6a
RDX: 0000000000040000 RSI: 000056520d500438 RDI: 0000000000000009
RBP: 0000000000040000 R08: 000056520d500410 R09: 0000000000000010
R10: 0000000000004011 R11: 0000000000000246 R12: 000056520d500410
R13: 000056520d500428 R14: 000056520d3cf9c8 R15: 000056520d3cf970
 </TASK>


***

If these findings have caused you to resend the series or submit a
separate fix, please add the following tag to your commit message:
  Tested-by: syzbot@syzkaller.appspotmail.com

---
This report is generated by a bot. It may contain errors.
syzbot ci engineers can be reached at syzkaller@googlegroups.com.

Re: [syzbot ci] Re: loop: improve loop aio perf by IOCB_NOWAIT

[Ming Lei]

On Sun, Sep 28, 2025 at 11:42:20AM -0700, syzbot ci wrote:
> syzbot ci has tested the following series
> 
> [v1] loop: improve loop aio perf by IOCB_NOWAIT
> https://lore.kernel.org/all/20250928132927.3672537-1-ming.lei@redhat.com
> * [PATCH V4 1/6] loop: add helper lo_cmd_nr_bvec()
> * [PATCH V4 2/6] loop: add helper lo_rw_aio_prep()
> * [PATCH V4 3/6] loop: add lo_submit_rw_aio()
> * [PATCH V4 4/6] loop: move command blkcg/memcg initialization into loop_queue_work
> * [PATCH V4 5/6] loop: try to handle loop aio command via NOWAIT IO first
> * [PATCH V4 6/6] loop: add hint for handling aio via IOCB_NOWAIT
> 
> and found the following issue:
> WARNING in lo_submit_rw_aio
> 
> Full report is available here:
> https://ci.syzbot.org/series/0ffdb6b4-a5fe-48da-9473-d2a926e780bd
> 
> ***
> 
> WARNING in lo_submit_rw_aio
> 
> tree:      torvalds
> URL:       https://kernel.googlesource.com/pub/scm/linux/kernel/git/torvalds/linux
> base:      07e27ad16399afcd693be20211b0dfae63e0615f
> arch:      amd64
> compiler:  Debian clang version 20.1.8 (++20250708063551+0c9f909b7976-1~exp1~20250708183702.136), Debian LLD 20.1.8
> config:    https://ci.syzbot.org/builds/3aba003b-2400-4e88-9a31-c09ab4e41a84/config
> C repro:   https://ci.syzbot.org/findings/dc97454c-d87b-41f5-a44a-7182e666cfd5/c_repro
> syz repro: https://ci.syzbot.org/findings/dc97454c-d87b-41f5-a44a-7182e666cfd5/syz_repro
> 
> ------------[ cut here ]------------
> WARNING: CPU: 0 PID: 5958 at drivers/block/loop.c:907 loop_inc_blocking_writes drivers/block/loop.c:907 [inline]

Thanks for your report!

Looks wrong lock is asserted, and the following change can fix it:

diff --git a/drivers/block/loop.c b/drivers/block/loop.c
index 911262b648ce..f3372bf35fd5 100644
--- a/drivers/block/loop.c
+++ b/drivers/block/loop.c
@@ -904,7 +904,7 @@ static inline int queue_on_root_worker(struct cgroup_subsys_state *css)
 static inline void loop_inc_blocking_writes(struct loop_device *lo,
                struct loop_cmd *cmd)
 {
-       lockdep_assert_held(&lo->lo_mutex);
+       lockdep_assert_held(&lo->lo_work_lock);

        if (req_op(blk_mq_rq_from_pdu(cmd)) == REQ_OP_WRITE)
                lo->lo_nr_blocking_writes += 1;
@@ -913,7 +913,7 @@ static inline void loop_inc_blocking_writes(struct loop_device *lo,
 static inline void loop_dec_blocking_writes(struct loop_device *lo,
                struct loop_cmd *cmd)
 {
-       lockdep_assert_held(&lo->lo_mutex);
+       lockdep_assert_held(&lo->lo_work_lock);

        if (req_op(blk_mq_rq_from_pdu(cmd)) == REQ_OP_WRITE)
                lo->lo_nr_blocking_writes -= 1;




Thanks,
Ming