xfs: avoid false quotacheck after unclean shutdown
authorEric Sandeen <sandeen@sandeen.net>
Mon, 4 Aug 2014 01:35:44 +0000 (11:35 +1000)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Fri, 14 Nov 2014 17:00:10 +0000 (09:00 -0800)
commit905290c5f3143eb78683bb24ec3f434fc0abb48b
tree624574db3abf5d9060098979a1873258c62ca387
parent1c6b7ee12064d9bbe84cd7ea85fa8bf8911e0b59
xfs: avoid false quotacheck after unclean shutdown

commit 5ef828c4152726f56751c78ea844f08d2b2a4fa3 upstream.

The commit

83e782e xfs: Remove incore use of XFS_OQUOTA_ENFD and XFS_OQUOTA_CHKD

added a new function xfs_sb_quota_from_disk() which swaps
on-disk XFS_OQUOTA_* flags for in-core XFS_GQUOTA_* and XFS_PQUOTA_*
flags after the superblock is read.

However, if log recovery is required, the superblock is read again,
and the modified in-core flags are re-read from disk, so we have
XFS_OQUOTA_* flags in memory again.  This causes the
XFS_QM_NEED_QUOTACHECK() test to be true, because the XFS_OQUOTA_CHKD
is still set, and not XFS_GQUOTA_CHKD or XFS_PQUOTA_CHKD.

Change xfs_sb_from_disk to call xfs_sb_quota_from disk and always
convert the disk flags to in-memory flags.

Add a lower-level function which can be called with "false" to
not convert the flags, so that the sb verifier can verify
exactly what was on disk, per Brian Foster's suggestion.

Reported-by: Cyril B. <cbay@excellency.fr>
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Cc: Arkadiusz Miƛkiewicz <arekm@maven.pl>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
fs/xfs/xfs_mount.c
fs/xfs/xfs_sb.c