Date:   Sat, 26 Mar 2022 12:48:26 +0100
From:   Christian Schoenebeck <linux_oss@...debyte.com>
To:     David Kahurani <k.kahurani@...il.com>
Cc:     davem@...emloft.net, ericvh@...il.com, kuba@...nel.org,
        linux-kernel@...r.kernel.org, lucho@...kov.net,
        netdev@...r.kernel.org, syzkaller-bugs@...glegroups.com,
        v9fs-developer@...ts.sourceforge.net,
        syzbot+5e28cdb7ebd0f2389ca4@...kaller.appspotmail.com,
        asmadeus@...ewreck.org
Subject: Re: [syzbot] WARNING in p9_client_destroy

On Donnerstag, 24. März 2022 13:13:25 CET David Kahurani wrote:
> On Monday, February 28, 2022 at 4:38:57 AM UTC+3 asmadeus@...ewreck.org
> 
> wrote:
> > syzbot wrote on Sun, Feb 27, 2022 at 04:53:29PM -0800:
> > > kmem_cache_destroy 9p-fcall-cache: Slab cache still has objects when
> > > called from p9_client_destroy+0x213/0x370 net/9p/client.c:1100
> > 
> > hmm, there is no previous "Packet with tag %d has still references"
> > (sic) message, so this is probably because p9_tag_cleanup only relies on
> > rcu read lock for consistency, so even if the connection has been closed
> > above (clnt->trans_mode->close) there could have been a request sent
> > (= tag added) just before that which isn't visible on the destroying
> > side?
> > 
> > I guess adding an rcu_barrier() is what makes most sense here to protect
> > this case?
> > I'll send a patch in the next few days unless it was a stupid idea.
> 
> Looking at this brought me to the same conclusion.
> 
> ---------------------
> 
> From cd5a11207a140004bf55005fac7f7e4cec2fd075 Mon Sep 17 00:00:00 2001
> From: David Kahurani <k.kahurani@...il.com>
> Date: Thu, 24 Mar 2022 15:00:23 +0300
> Subject: [PATCH] net/9p: Flush any delayed rce free
> 
> As is best practice
> 
> kmem_cache_destroy 9p-fcall-cache: Slab cache still has objects when called
> from p9_client_destroy+0x213/0x370 net/9p/client.c:1100
> WARNING: CPU: 1 PID: 3701 at mm/slab_common.c:502 kmem_cache_destroy
> mm/slab_common.c:502 [inline]
> WARNING: CPU: 1 PID: 3701 at mm/slab_common.c:502
> kmem_cache_destroy+0x13b/0x140 mm/slab_common.c:490
> Modules linked in:
> CPU: 1 PID: 3701 Comm: syz-executor.3 Not tainted
> 5.17.0-rc5-syzkaller-00021-g23d04328444a #0
> Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.14.0-2 04/01/2014
> RIP: 0010:kmem_cache_destroy mm/slab_common.c:502 [inline]
> RIP: 0010:kmem_cache_destroy+0x13b/0x140 mm/slab_common.c:490
> Code: da a8 0e 48 89 ee e8 44 6e 15 00 eb c1 c3 48 8b 55 58 48 c7 c6 60 cd
> b6 89 48 c7 c7 30 83 3a 8b 48 8b 4c 24 18 e8 9b 30 60 07 <0f> 0b eb a0 90
> 41 55 49 89 d5 41 54 49 89 f4 55 48 89 fd 53 48 83
> RSP: 0018:ffffc90002767cf0 EFLAGS: 00010282
> RAX: 0000000000000000 RBX: 1ffff920004ecfa5 RCX: 0000000000000000
> RDX: ffff88801e56a280 RSI: ffffffff815f4b38 RDI: fffff520004ecf90
> RBP: ffff888020ba8b00 R08: 0000000000000000 R09: 0000000000000000
> R10: ffffffff815ef1ce R11: 0000000000000000 R12: 0000000000000001
> R13: ffffc90002767d68 R14: dffffc0000000000 R15: 0000000000000000
> FS:  00005555561b0400(0000) GS:ffff88802ca00000(0000) knlGS:0000000000000000
> CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> CR2: 0000555556ead708 CR3: 0000000068b97000 CR4: 0000000000150ef0
> DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
> Call Trace:
>  <TASK>
>  p9_client_destroy+0x213/0x370 net/9p/client.c:1100
>  v9fs_session_close+0x45/0x2d0 fs/9p/v9fs.c:504
>  v9fs_kill_super+0x49/0x90 fs/9p/vfs_super.c:226
>  deactivate_locked_super+0x94/0x160 fs/super.c:332
>  deactivate_super+0xad/0xd0 fs/super.c:363
>  cleanup_mnt+0x3a2/0x540 fs/namespace.c:1173
>  task_work_run+0xdd/0x1a0 kernel/task_work.c:164
>  tracehook_notify_resume include/linux/tracehook.h:188 [inline]
>  exit_to_user_mode_loop kernel/entry/common.c:175 [inline]
>  exit_to_user_mode_prepare+0x27e/0x290 kernel/entry/common.c:207
>  __syscall_exit_to_user_mode_work kernel/entry/common.c:289 [inline]
>  syscall_exit_to_user_mode+0x19/0x60 kernel/entry/common.c:300
>  do_syscall_64+0x42/0xb0 arch/x86/entry/common.c:86
>  entry_SYSCALL_64_after_hwframe+0x44/0xae
> RIP: 0033:0x7f5ff63ed4c7
> Code: ff ff ff f7 d8 64 89 01 48 83 c8 ff c3 66 0f 1f 44 00 00 31 f6 e9 09
> 00 00 00 66 0f 1f 84 00 00 00 00 00 b8 a6 00 00 00 0f 05 <48> 3d 01 f0 ff
> ff 73 01 c3 48 c7 c1 b8 ff ff ff f7 d8 64 89 01 48
> RSP: 002b:00007fff01862e98 EFLAGS: 00000246 ORIG_RAX: 00000000000000a6
> RAX: 0000000000000000 RBX: 0000000000000000 RCX: 00007f5ff63ed4c7
> RDX: 00007fff01862f6c RSI: 000000000000000a RDI: 00007fff01862f60
> RBP: 00007fff01862f60 R08: 00000000ffffffff R09: 00007fff01862d30
> R10: 00005555561b18b3 R11: 0000000000000246 R12: 00007f5ff64451ea
> R13: 00007fff01864020 R14: 00005555561b1810 R15: 00007fff01864060
>  </TASK>
> 
> Signed-off-by: David Kahurani <k.kahurani@...il.com>
> Reported-by: syzbot+5e28cdb7ebd0f2389ca4@...kaller.appspotmail.com

I'm not absolutely sure that this will really fix this issue, but it seems to 
be a good idea to add a rcu_barrier() call here nevertheless.

Reviewed-by: Christian Schoenebeck <linux_oss@...debyte.com>

> ---
>  net/9p/client.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/net/9p/client.c b/net/9p/client.c
> index 8bba0d9cf..67c51913a 100644
> --- a/net/9p/client.c
> +++ b/net/9p/client.c
> @@ -1097,6 +1097,7 @@ void p9_client_destroy(struct p9_client *clnt)
> 
>   p9_tag_cleanup(clnt);
> 
> + rcu_barrier();
>   kmem_cache_destroy(clnt->fcall_cache);
>   kfree(clnt);
>  }

Powered by blists - more mailing lists