Commit 30133177 authored Apr 08, 2015 by Abhi Das Committed by Bob Peterson Apr 08, 2015

gfs2: fix quota refresh race in do_glock()



quotad periodically syncs in-memory quotas to the ondisk quota file
and sets the QDF_REFRESH flag so that a subsequent read of a synced
quota is re-read from disk.

gfs2_quota_lock() checks for this flag and sets a 'force' bit to
force re-read from disk if requested. However, there is a race
condition here. It is possible for gfs2_quota_lock() to find the
QDF_REFRESH flag unset (i.e force=0) and quotad comes in immediately
after and syncs the relevant quota and sets the QDF_REFRESH flag.
gfs2_quota_lock() resumes with force=0 and uses the stale in-memory
quota usage values that result in miscalculations.

This patch fixes this race by moving the check for the QDF_REFRESH
flag check further out into the gfs2_quota_lock() process, i.e, in
do_glock(), under the protection of the quota glock.

Signed-off-by: Abhi Das <adas@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Acked-by: Steven Whitehouse <swhiteho@redhat.com>

parent 7b4ddfa7

fs/gfs2/quota.c

+4 −4

Original line number	Diff line number	Diff line
		@@ -923,6 +923,9 @@ static int do_glock(struct gfs2_quota_data *qd, int force_refresh,
		if (error)
		return error;

		if (test_and_clear_bit(QDF_REFRESH, &qd->qd_flags))
		force_refresh = FORCE;

		qd->qd_qb = (struct gfs2_quota_lvb )qd->qd_gl->gl_lksb.sb_lvbptr;

		if (force_refresh \|\| qd->qd_qb.qb_magic != cpu_to_be32(GFS2_MAGIC)) {
		@@ -974,11 +977,8 @@ int gfs2_quota_lock(struct gfs2_inode *ip, kuid_t uid, kgid_t gid)
		sizeof(struct gfs2_quota_data *), sort_qd, NULL);

		for (x = 0; x < ip->i_res->rs_qa_qd_num; x++) {
		int force = NO_FORCE;
		qd = ip->i_res->rs_qa_qd[x];
		if (test_and_clear_bit(QDF_REFRESH, &qd->qd_flags))
		force = FORCE;
		error = do_glock(qd, force, &ip->i_res->rs_qa_qd_ghs[x]);
		error = do_glock(qd, NO_FORCE, &ip->i_res->rs_qa_qd_ghs[x]);
		if (error)
		break;
		}