From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755522Ab3H1Vwz (ORCPT ); Wed, 28 Aug 2013 17:52:55 -0400 Received: from mga02.intel.com ([134.134.136.20]:18127 "EHLO mga02.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753468Ab3H1Vwx (ORCPT ); Wed, 28 Aug 2013 17:52:53 -0400 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="4.89,978,1367996400"; d="scan'208";a="394965642" Subject: [PATCH] Avoid useless inodes and dentries reclamation From: Tim Chen To: Alexander Viro , Jan Kara , Dave Chinner Cc: Dave Hansen , Andi Kleen , Matthew Wilcox , linux-fsdevel , linux-kernel Content-Type: text/plain; charset="UTF-8" Date: Wed, 28 Aug 2013 14:52:12 -0700 Message-ID: <1377726732.3625.31.camel@schen9-DESK> Mime-Version: 1.0 X-Mailer: Evolution 2.32.3 (2.32.3-1.fc14) Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org This patch detects that when free inodes and dentries are really low, their reclamation is skipped so we do not have to contend on the global sb_lock uselessly under memory pressure. Otherwise we create a log jam trying to acquire the sb_lock in prune_super(), with little or no freed memory to show for the effort. The profile below shows a multi-threaded large file read exerting pressure on memory with page cache usage. It is dominated by the sb_lock contention in the cpu cycles profile. The patch eliminates the sb_lock contention almost entirely for prune_super(). 43.94% usemem [kernel.kallsyms] [k] _raw_spin_lock | --- _raw_spin_lock | |--32.44%-- grab_super_passive | prune_super | shrink_slab | do_try_to_free_pages | try_to_free_pages | __alloc_pages_nodemask | alloc_pages_current | |--32.18%-- put_super | drop_super | prune_super | shrink_slab | do_try_to_free_pages | try_to_free_pages | __alloc_pages_nodemask | alloc_pages_current Signed-off-by: Tim Chen --- fs/super.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/fs/super.c b/fs/super.c index 68307c0..70fa26c 100644 --- a/fs/super.c +++ b/fs/super.c @@ -53,6 +53,7 @@ static char *sb_writers_name[SB_FREEZE_LEVELS] = { * shrinker path and that leads to deadlock on the shrinker_rwsem. Hence we * take a passive reference to the superblock to avoid this from occurring. */ +#define SB_CACHE_LOW 5 static int prune_super(struct shrinker *shrink, struct shrink_control *sc) { struct super_block *sb; @@ -68,6 +69,13 @@ static int prune_super(struct shrinker *shrink, struct shrink_control *sc) if (sc->nr_to_scan && !(sc->gfp_mask & __GFP_FS)) return -1; + /* + * Don't prune if we have few cached objects to reclaim to + * avoid useless sb_lock contention + */ + if ((sb->s_nr_dentry_unused + sb->s_nr_inodes_unused) <= SB_CACHE_LOW) + return -1; + if (!grab_super_passive(sb)) return -1; -- 1.7.11.7