From mboxrd@z Thu Jan 1 00:00:00 1970 From: Peter Zijlstra Subject: Re: [PATCH 00/28] mm: preemptibility -v3 Date: Mon, 07 Jun 2010 18:39:31 +0200 Message-ID: <1275928771.1645.572.camel@laptop> References: <20100607110654.606530953@chello.nl> <87hble6brb.fsf@basil.nowhere.org> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <87hble6brb.fsf@basil.nowhere.org> Sender: linux-kernel-owner@vger.kernel.org To: Andi Kleen Cc: Andrea Arcangeli , Avi Kivity , Thomas Gleixner , Rik van Riel , Ingo Molnar , akpm@firstfloor.org, linux-kernel@vger.kernel.org, linux-arch@vger.kernel.org, Benjamin Herrenschmidt , David Miller , Hugh Dickins , Mel Gorman , Nick Piggin , "Paul E. McKenney" List-Id: linux-arch.vger.kernel.org On Mon, 2010-06-07 at 18:36 +0200, Andi Kleen wrote: > Peter Zijlstra writes: > > > This patch-set makes part of the mm a lot more preemptible. It converts > > i_mmap_lock and anon_vma->lock to mutexes and makes mmu_gather fully > > preemptible. > > How about performance measurements? mutexes still behave quite > differently from spinlocks, especially under contention. What's your favourite benchmark to stress i_mmap_mutex/anon_vma->lock? a cache-hot kernel build didn't really show a difference.. From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from bombadil.infradead.org ([18.85.46.34]:58608 "EHLO bombadil.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752692Ab0FGQjk (ORCPT ); Mon, 7 Jun 2010 12:39:40 -0400 Received: from f199130.upc-f.chello.nl ([80.56.199.130] helo=dyad.programming.kicks-ass.net) by bombadil.infradead.org with esmtpsa (Exim 4.72 #1 (Red Hat Linux)) id 1OLfMS-0001P2-2c for linux-arch@vger.kernel.org; Mon, 07 Jun 2010 16:39:40 +0000 Subject: Re: [PATCH 00/28] mm: preemptibility -v3 From: Peter Zijlstra In-Reply-To: <87hble6brb.fsf@basil.nowhere.org> References: <20100607110654.606530953@chello.nl> <87hble6brb.fsf@basil.nowhere.org> Content-Type: text/plain; charset="UTF-8" Date: Mon, 07 Jun 2010 18:39:31 +0200 Message-ID: <1275928771.1645.572.camel@laptop> Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Sender: linux-arch-owner@vger.kernel.org List-ID: To: Andi Kleen Cc: Andrea Arcangeli , Avi Kivity , Thomas Gleixner , Rik van Riel , Ingo Molnar , akpm@firstfloor.org, linux-kernel@vger.kernel.org, linux-arch@vger.kernel.org, Benjamin Herrenschmidt , David Miller , Hugh Dickins , Mel Gorman , Nick Piggin , "Paul E. McKenney" Message-ID: <20100607163931.oQclzeYm3l3nyH6JEHLBBQ9fT3KmlOq7iFRA0ZtAbFI@z> On Mon, 2010-06-07 at 18:36 +0200, Andi Kleen wrote: > Peter Zijlstra writes: > > > This patch-set makes part of the mm a lot more preemptible. It converts > > i_mmap_lock and anon_vma->lock to mutexes and makes mmu_gather fully > > preemptible. > > How about performance measurements? mutexes still behave quite > differently from spinlocks, especially under contention. What's your favourite benchmark to stress i_mmap_mutex/anon_vma->lock? a cache-hot kernel build didn't really show a difference..