From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail202.messagelabs.com (mail202.messagelabs.com [216.82.254.227]) by kanga.kvack.org (Postfix) with SMTP id 01DBE90010B for ; Fri, 6 May 2011 12:00:12 -0400 (EDT) Message-ID: <4DC41AF2.6000804@redhat.com> Date: Fri, 06 May 2011 11:59:46 -0400 From: Rik van Riel MIME-Version: 1.0 Subject: Re: [PATCH 2/4] VM/RMAP: Batch anon vma chain root locking in fork References: <1304623972-9159-1-git-send-email-andi@firstfloor.org> <1304623972-9159-3-git-send-email-andi@firstfloor.org> In-Reply-To: <1304623972-9159-3-git-send-email-andi@firstfloor.org> Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit Sender: owner-linux-mm@kvack.org List-ID: To: Andi Kleen Cc: linux-mm@kvack.org, Andrea Arcangeli , tim.c.chen@linux.intel.com, torvalds@linux-foundation.org, lwoodman@redhat.com, mel@csn.ul.ie, Andi Kleen On 05/05/2011 03:32 PM, Andi Kleen wrote: > From: Andi Kleen > > We found that the changes to take anon vma root chain lock lead > to excessive lock contention on a fork intensive workload on a 4S > system. > > Use the new batch lock infrastructure to optimize the fork() > path, where it is very common to acquire always the same lock. > > This patch does not really lower the contention, but batches > the lock taking/freeing to lower the bouncing overhead when > multiple forks are working at the same time. Essentially each > user will get more work done inside a locking region. > > Reported-by: Tim Chen > Cc: Andrea Arcangeli > Cc: Rik van Riel > Signed-off-by: Andi Kleen Reviewed-by: Rik van Riel -- All rights reversed -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@kvack.org. For more info on Linux MM, see: http://www.linux-mm.org/ . Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/ Don't email: email@kvack.org