From mboxrd@z Thu Jan 1 00:00:00 1970 From: Steven Whitehouse Date: Fri, 26 Sep 2008 13:00:48 +0100 Subject: [Cluster-devel] [PATCH 05/12] GFS2: Fix race relating to glock min-hold time In-Reply-To: <1222430455-4632-5-git-send-email-swhiteho@redhat.com> References: <1222430455-4632-1-git-send-email-swhiteho@redhat.com> <1222430455-4632-2-git-send-email-swhiteho@redhat.com> <1222430455-4632-3-git-send-email-swhiteho@redhat.com> <1222430455-4632-4-git-send-email-swhiteho@redhat.com> <1222430455-4632-5-git-send-email-swhiteho@redhat.com> Message-ID: <1222430455-4632-6-git-send-email-swhiteho@redhat.com> List-Id: To: cluster-devel.redhat.com MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit In the case that a request for a glock arrives right after the grant reply has arrived, it sometimes means that the gl_tstamp field hasn't been updated recently enough. The net result is that the min-hold time for the glock is ignored. If this happens often enough, it leads to poor performance. This patch adds an additional test, so that if the reply pending bit is set on a glock, then it will select the maximum length of time for the min-hold time, rather than looking at gl_tstamp. Signed-off-by: Steven Whitehouse diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c index 4cbb695..806e1eb 100644 --- a/fs/gfs2/glock.c +++ b/fs/gfs2/glock.c @@ -1265,6 +1265,8 @@ static void blocking_cb(struct gfs2_sbd *sdp, struct lm_lockname *name, holdtime = gl->gl_tchange + gl->gl_ops->go_min_hold_time; if (time_before(now, holdtime)) delay = holdtime - now; + if (test_bit(GLF_REPLY_PENDING, &gl->gl_flags)) + delay = gl->gl_ops->go_min_hold_time; spin_lock(&gl->gl_spin); handle_callback(gl, state, 1, delay); -- 1.5.5.1