xfs: make quota metadata truncation behavior consistent to user space
authorJie Liu <jeff.liu@oracle.com>
Fri, 22 Nov 2013 06:04:00 +0000 (14:04 +0800)
committerBen Myers <bpm@sgi.com>
Fri, 6 Dec 2013 20:06:15 +0000 (14:06 -0600)
commitc61a9e39f637373929a110ad2a5922a2e8b00f4c
treed0ad6bcf0c924d523072863679e0e09cf696fe4b
parentef701600fd26cace9d513ee174688a2b83832126
xfs: make quota metadata truncation behavior consistent to user space

In xfs_qm_scall_trunc_qfiles(), we ignore the error if failed to remove
the users quota metadata and proceed to remove groups and projects if
they are being there.  However, in user space, the remove operation will
break and return if failed to remove any kind of quota.
Also for v5 super block, we can enabled both group and project quota at
the same time, in this case the current error handling will cover the
group error with projects but they might failed due to different reasons.

It seems we'd better the error handling consistent to the user space and
don't trying to remove another kind of quota metadata if the previous
operation is failed.

Signed-off-by: Jie Liu <jeff.liu@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Ben Myers <bpm@sgi.com>
fs/xfs/xfs_qm_syscalls.c