[PATCH] xfs: test for shut down fs in xfs_dir_fsync()

[Eric Sandeen]

On 4/28/14, 3:54 PM, Dave Chinner wrote:
> On Mon, Apr 28, 2014 at 11:35:16AM -0500, Eric Sandeen wrote:
>> Similar to xfs_file_fsync(), I think xfs_dir_fsync() needs
>> to test for a shut down fs, lest we go down paths we'll
>> never be able to complete; Boris reported that during some
>> stress tests he had threads stuck in xlog_cil_force_lsn
>> via xfs_dir_fsync().
>>
>> [ 3663.361709] sfsuspend-par   D ffff88042f0b4540     0  3981   3947 0x00000080
>>
>> [ 3663.394472] Call Trace:
>> [ 3663.397199]  [<ffffffff815f1889>] schedule+0x29/0x70
>> [ 3663.402743]  [<ffffffffa01feda5>] xlog_cil_force_lsn+0x185/0x1a0 [xfs]
>> [ 3663.416249]  [<ffffffffa01fd3af>] _xfs_log_force_lsn+0x6f/0x2f0 [xfs]
>> [ 3663.429271]  [<ffffffffa01a339d>] xfs_dir_fsync+0x7d/0xe0 [xfs]
>> [ 3663.435873]  [<ffffffff811df8c5>] do_fsync+0x65/0xa0
>> [ 3663.441408]  [<ffffffff811dfbc0>] SyS_fsync+0x10/0x20
>> [ 3663.447043]  [<ffffffff815fc7d9>] system_call_fastpath+0x16/0x1b
> 
> Wow, I believe it's taken this long for us to notice that we can't
> break out of xlog_cil_force_lsn() if we fail on xlog_write()
> from a CIL push.
> 
> I'd say that xlog_cil_force_lsn() needs log shutdown checks before
> it goes to sleep in xlog_wait()....
> 
>> Reported-by: Boris Ranto <branto at redhat.com>
>> Signed-off-by: Eric Sandeen <sandeen at redhat.com>
>> ---
>>
>> NB: While I've not asked Boris to test this yet, it seems
>> clear (?) that dir_fsync should behave the same as
>> file_fsync() in the face of a shut-down fs.
>>
>> diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c
>> index 4c749ab..2b94362 100644
>> --- a/fs/xfs/xfs_file.c
>> +++ b/fs/xfs/xfs_file.c
>> @@ -146,6 +146,9 @@ xfs_dir_fsync(
>>  
>>  	trace_xfs_dir_fsync(ip);
>>  
>> +	if (XFS_FORCED_SHUTDOWN(mp))
>> +		return -XFS_ERROR(EIO);
>> +
> 
> That won't hurt, but it won't fix the problem.

So, you did solve the problem properly I guess, in 
commit ac983517ec5941da0c58cacdbad10a231dc4e001
Author: Dave Chinner <dchinner at redhat.com>
Date:   Wed May 7 08:05:50 2014 +1000

    xfs: don't sleep in xlog_cil_force_lsn on shutdown
    

so thanks.  :)  Should my patch still go in, to be consistent
with file_fsync() paths?

-Eric

> Cheers,
> 
> Dave.
>