Whamcloud - gitweb
LU-11296 osc: speed up page cache cleanup during blocking ASTs 90/33090/4
authorAndrew Perepechko <c17827@cray.com>
Fri, 10 Aug 2018 13:18:48 +0000 (16:18 +0300)
committerOleg Drokin <green@whamcloud.com>
Fri, 4 Jan 2019 04:45:17 +0000 (04:45 +0000)
While we are cleaning a write lock, we don't need to check if
page cache pages under this lock are covered by another lock.

If a client needs to give up its lock, cleaning gigabytes of
page cache can take quite a long time.

Signed-off-by: Andrew Perepechko <c17827@cray.com>
Cray-bug-id: LUS-6352
Change-Id: I576130216ed4de4e352ea697bddb5ff83046443a
Reviewed-by: Patrick Farrell <paf@cray.com>
Reviewed-by: Alexander Zarochentsev <c17826@cray.com>
Reviewed-on: https://review.whamcloud.com/33090
Reviewed-by: Jinshan Xiong <jinshan.xiong@gmail.com>
Tested-by: Jenkins
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
lustre/osc/osc_lock.c

index b2042a4..dca41e4 100644 (file)
@@ -377,7 +377,12 @@ static int osc_lock_flush(struct osc_object *obj, pgoff_t start, pgoff_t end,
                        rc = 0;
        }
 
-       rc2 = osc_lock_discard_pages(env, obj, start, end, discard);
+       /*
+        * Do not try to match other locks with CLM_WRITE since we already
+        * know there're none
+        */
+       rc2 = osc_lock_discard_pages(env, obj, start, end,
+                                    mode == CLM_WRITE || discard);
        if (rc == 0 && rc2 < 0)
                rc = rc2;