Re: nfs: possible sync issue between nfs_call_unlink <-> nfs_async_unlink_release

From: Rafael David Tinoco <rafael.tinoco@linaro.org>
To: linux-kernel@vger.kernel.org
Cc: rafael.tinoco@linaro.org, Waiman Long <longman@redhat.com>
Subject: Re: nfs: possible sync issue between nfs_call_unlink <-> nfs_async_unlink_release
Date: Tue, 13 Nov 2018 22:28:52 -0200	[thread overview]
Message-ID: <c351b04b-e83e-d090-a6b2-b8b08a688872@linaro.org> (raw)
In-Reply-To: <CADtBP2AaVMgmYGsvBjsVjw9M4hCiYOcMMbPu6nS3Fmo8i4wY3w@mail.gmail.com>

On 7/3/18 6:04 PM, Rafael Tinoco wrote:
> BUG: https://bugs.linaro.org/show_bug.cgi?id=3731
> 
> During Linaro's Kernel Functional tests, we have observed the
> following situation:
> 
> [   52.651490] DEBUG_LOCKS_WARN_ON(sem->owner != ((struct task_struct *)1UL))
> [   52.651506] WARNING: CPU: 2 PID: 1457 at
> ./kernel/locking/rwsem.c:217 up_read_non_owner+0x5d/0x70
> [   52.674398] Modules linked in: x86_pkg_temp_thermal fuse
> [   52.679719] CPU: 2 PID: 1457 Comm: kworker/2:2 Not tainted 4.16.0 #1
> [   52.687448] Hardware name: Supermicro SYS-5019S-ML/X11SSH-F, BIOS
> 2.0b 07/27/2017
> [   52.694922] Workqueue: nfsiod rpc_async_release
> [   52.699454] RIP: 0010:up_read_non_owner+0x5d/0x70
> [   52.704157] RSP: 0018:ffff9cbf81a23dd0 EFLAGS: 00010282
> [   52.709376] RAX: 0000000000000000 RBX: ffff8dc1983c76c0 RCX: 0000000000000000
> [   52.716500] RDX: ffffffffbd2d26c9 RSI: 0000000000000001 RDI: ffffffffbd2d2889
> [   52.723652] RBP: ffff9cbf81a23dd8 R08: 0000000000000000 R09: 0000000000000000
> [   52.730782] R10: ffff9cbf81a23dd0 R11: 0000000000000000 R12: ffff8dc19abf8600
> [   52.737906] R13: ffff8dc19b6c0000 R14: 0000000000000000 R15: ffff8dc19bacad80
> [   52.745029] FS:  0000000000000000(0000) GS:ffff8dc1afd00000(0000)
> knlGS:0000000000000000
> [   52.753108] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> [   52.758845] CR2: 00007f33794665d8 CR3: 000000016c41e006 CR4: 00000000003606e0
> [   52.765968] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> [   52.773091] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
> [   52.780215] Call Trace:
> [   52.782695]  nfs_async_unlink_release+0x32/0x80
> [   52.787220]  rpc_free_task+0x30/0x50
> [   52.790789]  rpc_async_release+0x12/0x20
> [   52.794707]  process_one_work+0x25e/0x660
> [   52.798713]  worker_thread+0x4b/0x410
> [   52.802377]  kthread+0x10d/0x140
> [   52.805600]  ? rescuer_thread+0x3a0/0x3a0
> [   52.809652]  ? kthread_create_worker_on_cpu+0x70/0x70
> [   52.814702]  ? do_syscall_64+0x69/0x1b0
> [   52.818540]  ret_from_fork+0x3a/0x50

Full explanation: https://bugs.linaro.org/show_bug.cgi?id=3731#c15

Appears that:

commit 925b9cd1b89a94b7124d128c80dfc48f78a63098
Author: Waiman Long <longman@redhat.com>
Date:   Thu Sep 6 17:18:34 2018

     locking/rwsem: Make owner store task pointer of last owning reader

seem to have fixed this warning:

DEBUG_RWSEMS_WARN_ON() started differentiating unknown owners for 
semaphores, like in this case, when RPC function forces the release the 
dentry's semaphore to avoid issues with NFS filesystem logic (nfs_rmdir, 
the real user for the semaphore in question).

Thanks Waiman!
-
Rafael D. Tinoco
Linaro Kernel Validation