xfs: preserve default grace interval during quotacheck
authorDarrick J. Wong <darrick.wong@oracle.com>
Wed, 25 Mar 2020 03:12:53 +0000 (20:12 -0700)
committerDarrick J. Wong <darrick.wong@oracle.com>
Thu, 26 Mar 2020 15:19:24 +0000 (08:19 -0700)
When quotacheck runs, it zeroes all the timer fields in every dquot.
Unfortunately, it also does this to the root dquot, which erases any
preconfigured grace intervals and warning limits that the administrator
may have set.  Worse yet, the incore copies of those variables remain
set.  This cache coherence problem manifests itself as the grace
interval mysteriously being reset back to the defaults at the /next/
mount.

Fix it by not resetting the root disk dquot's timer and warning fields.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
fs/xfs/xfs_qm.c

index de1d2c606c14e6309037924badf37d381242af42..cabdb755adaec6e023107c9cffb27d61ae494926 100644 (file)
@@ -871,12 +871,20 @@ xfs_qm_reset_dqcounts(
                ddq->d_bcount = 0;
                ddq->d_icount = 0;
                ddq->d_rtbcount = 0;
-               ddq->d_btimer = 0;
-               ddq->d_itimer = 0;
-               ddq->d_rtbtimer = 0;
-               ddq->d_bwarns = 0;
-               ddq->d_iwarns = 0;
-               ddq->d_rtbwarns = 0;
+
+               /*
+                * dquot id 0 stores the default grace period and the maximum
+                * warning limit that were set by the administrator, so we
+                * should not reset them.
+                */
+               if (ddq->d_id != 0) {
+                       ddq->d_btimer = 0;
+                       ddq->d_itimer = 0;
+                       ddq->d_rtbtimer = 0;
+                       ddq->d_bwarns = 0;
+                       ddq->d_iwarns = 0;
+                       ddq->d_rtbwarns = 0;
+               }
 
                if (xfs_sb_version_hascrc(&mp->m_sb)) {
                        xfs_update_cksum((char *)&dqb[j],