Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S933624AbdCJTqf (ORCPT ); Fri, 10 Mar 2017 14:46:35 -0500 Received: from mail-pg0-f50.google.com ([74.125.83.50]:34503 "EHLO mail-pg0-f50.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755276AbdCJTq1 (ORCPT ); Fri, 10 Mar 2017 14:46:27 -0500 From: Shakeel Butt To: Johannes Weiner , Mel Gorman , Michal Hocko , Vlastimil Babka , Andrew Morton Cc: Jia He , Hillf Danton , linux-mm@kvack.org, linux-kernel@vger.kernel.org, Shakeel Butt Subject: [PATCH] mm: fix condition for throttle_direct_reclaim Date: Fri, 10 Mar 2017 11:46:20 -0800 Message-Id: <20170310194620.5021-1-shakeelb@google.com> X-Mailer: git-send-email 2.12.0.246.ga2ecc84866-goog Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 2149 Lines: 60 Recently kswapd has been modified to give up after MAX_RECLAIM_RETRIES number of unsucessful iterations. Before going to sleep, kswapd thread will unconditionally wakeup all threads sleeping on pfmemalloc_wait. However the awoken threads will recheck the watermarks and wake the kswapd thread and sleep again on pfmemalloc_wait. There is a chance of continuous back and forth between kswapd and direct reclaiming threads if the kswapd keep failing and thus defeat the purpose of adding backoff mechanism to kswapd. So, add kswapd_failures check on the throttle_direct_reclaim condition. Signed-off-by: Shakeel Butt --- mm/vmscan.c | 12 +++++++++--- 1 file changed, 9 insertions(+), 3 deletions(-) diff --git a/mm/vmscan.c b/mm/vmscan.c index bae698484e8e..b2d24cc7a161 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -2819,6 +2819,12 @@ static bool pfmemalloc_watermark_ok(pg_data_t *pgdat) return wmark_ok; } +static bool should_throttle_direct_reclaim(pg_data_t *pgdat) +{ + return (pgdat->kswapd_failures < MAX_RECLAIM_RETRIES && + !pfmemalloc_watermark_ok(pgdat)); +} + /* * Throttle direct reclaimers if backing storage is backed by the network * and the PFMEMALLOC reserve for the preferred node is getting dangerously @@ -2873,7 +2879,7 @@ static bool throttle_direct_reclaim(gfp_t gfp_mask, struct zonelist *zonelist, /* Throttle based on the first usable node */ pgdat = zone->zone_pgdat; - if (pfmemalloc_watermark_ok(pgdat)) + if (!should_throttle_direct_reclaim(pgdat)) goto out; break; } @@ -2895,14 +2901,14 @@ static bool throttle_direct_reclaim(gfp_t gfp_mask, struct zonelist *zonelist, */ if (!(gfp_mask & __GFP_FS)) { wait_event_interruptible_timeout(pgdat->pfmemalloc_wait, - pfmemalloc_watermark_ok(pgdat), HZ); + !should_throttle_direct_reclaim(pgdat), HZ); goto check_pending; } /* Throttle until kswapd wakes the process */ wait_event_killable(zone->zone_pgdat->pfmemalloc_wait, - pfmemalloc_watermark_ok(pgdat)); + !should_throttle_direct_reclaim(pgdat)); check_pending: if (fatal_signal_pending(current)) -- 2.12.0.246.ga2ecc84866-goog