Re: ksmbd stuck open response

[Namjae Jeon <linkinjeon@kernel.org>]

2024년 5월 9일 (목) 오전 6:26, Steve French <smfrench@gmail.com>님이 작성:
>
> I saw an example in 6.9-rc6 where ksmbd was failing to send an open
> response repeatedly (it showed up running test cifs/102
Where is cifs/102 test ? I can not find it in xfstests.
>
> On the client I see the call one stuck SMB3.1.1 open request (never
> returns) and call stack of:
>
> [root@fedora29 ~]# cat /proc/5042/stack
> [<0>] wait_for_response+0xd1/0x130 [cifs]
> [<0>] compound_send_recv+0x68e/0x10b0 [cifs]
> [<0>] cifs_send_recv+0x23/0x30 [cifs]
> [<0>] SMB2_open+0x378/0xbd0 [cifs]
> [<0>] smb2_open_file+0x171/0x560 [cifs]
> [<0>] cifs_do_create.isra.0+0x471/0xd40 [cifs]
> [<0>] cifs_atomic_open+0x382/0x780 [cifs]
> [<0>] lookup_open.isra.0+0x6b0/0x930
> [<0>] path_openat+0x491/0x10d0
> [<0>] do_filp_open+0x144/0x250
> [<0>] do_sys_openat2+0xe0/0x110
> [<0>] __x64_sys_openat+0xc1/0x120
> [<0>] do_syscall_64+0x78/0x180
> [<0>] entry_SYSCALL_64_after_hwframe+0x76/0x7e
>
> on the server I see lots of failure to send a message on the socket
> every five seconds:
> ...
> [Wed May  8 21:13:59 2024] ksmbd: Failed to send message: -11
> [Wed May  8 21:14:04 2024] ksmbd: Failed to send message: -11
> [Wed May  8 21:14:09 2024] ksmbd: Failed to send message: -11
> [Wed May  8 21:14:14 2024] ksmbd: Failed to send message: -11
> [Wed May  8 21:14:20 2024] ksmbd: Failed to send message: -11
> [Wed May  8 21:14:25 2024] ksmbd: Failed to send message: -11
> [Wed May  8 21:14:30 2024] ksmbd: Failed to send message: -11
> [Wed May  8 21:14:35 2024] ksmbd: Failed to send message: -11
> [Wed May  8 21:14:40 2024] ksmbd: Failed to send message: -11
> [Wed May  8 21:14:45 2024] ksmbd: Failed to send message: -11
>
> doing "ksmbd.control -s" did free that request but hangs on the server
> (and subsequent requests on the client fail).  Nothing obvious in
> dmesg on client or server, except the following (xfs bug?) which was
> logged after the "ksmbd.control -s" on the server:
>
> [24324.546128] Call Trace:
> [24324.546130]  <TASK>
> [24324.546134]  dump_stack_lvl+0x76/0xa0
> [24324.546141]  dump_stack+0x10/0x20
> [24324.546145]  xfs_error_report+0x4a/0x70 [xfs]
> [24324.546298]  ? xfs_remove+0x175/0x300 [xfs]
> [24324.546440]  xfs_trans_cancel+0x14b/0x170 [xfs]
> [24324.546582]  xfs_remove+0x175/0x300 [xfs]
> [24324.546724]  xfs_vn_unlink+0x53/0xb0 [xfs]
> [24324.546866]  vfs_unlink+0x146/0x2e0
> [24324.546872]  ksmbd_vfs_unlink+0xa9/0x140 [ksmbd]
> [24324.546888]  ? __pfx_session_fd_check+0x10/0x10 [ksmbd]
> [24324.546902]  __ksmbd_close_fd+0x2ba/0x2d0 [ksmbd]
> [24324.546916]  ? _raw_spin_unlock+0xe/0x40
> [24324.546920]  ? __pfx_session_fd_check+0x10/0x10 [ksmbd]
> [24324.546936]  __close_file_table_ids+0x60/0xb0 [ksmbd]
> [24324.546950]  ksmbd_destroy_file_table+0x22/0x60 [ksmbd]
> [24324.546966]  ksmbd_session_destroy+0x5a/0x1b0 [ksmbd]
> [24324.546984]  ksmbd_sessions_deregister+0x24c/0x270 [ksmbd]
> [24324.547001]  ksmbd_server_terminate_conn+0x12/0x30 [ksmbd]
> [24324.547016]  ksmbd_conn_handler_loop+0x203/0x370 [ksmbd]
> [24324.547034]  ? __pfx_ksmbd_conn_handler_loop+0x10/0x10 [ksmbd]
> [24324.547050]  kthread+0xe4/0x110
> [24324.547054]  ? __pfx_kthread+0x10/0x10
> [24324.547058]  ret_from_fork+0x47/0x70
> [24324.547062]  ? __pfx_kthread+0x10/0x10
> [24324.547065]  ret_from_fork_asm+0x1a/0x30
> [24324.547070]  </TASK>
>
>
> Any ideas?
> --
> Thanks,
>
> Steve
>