From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S965458AbbDXPRo (ORCPT ); Fri, 24 Apr 2015 11:17:44 -0400 Received: from cantor2.suse.de ([195.135.220.15]:44030 "EHLO mx2.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S933814AbbDXPQ4 (ORCPT ); Fri, 24 Apr 2015 11:16:56 -0400 Date: Fri, 24 Apr 2015 16:16:52 +0100 From: Mel Gorman To: Vlastimil Babka Cc: Linux-MM , Rik van Riel , Hugh Dickins , Minchan Kim , Dave Hansen , Andi Kleen , LKML Subject: Re: [RFC PATCH 0/6] TLB flush multiple pages with a single IPI v3 Message-ID: <20150424151652.GC2449@suse.de> References: <1429612880-21415-1-git-send-email-mgorman@suse.de> <553A573E.2000608@suse.cz> MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-15 Content-Disposition: inline In-Reply-To: <553A573E.2000608@suse.cz> User-Agent: Mutt/1.5.21 (2010-09-15) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Fri, Apr 24, 2015 at 04:46:22PM +0200, Vlastimil Babka wrote: > On 04/21/2015 12:41 PM, Mel Gorman wrote: > >Changelog since V2 > >o Ensure TLBs are flushed before pages are freed (mel) > > I admit not reading all the patches thoroughly, but doesn't this > change of ordering mean that you no longer need the architectural > guarantee discussed in patch 2? No. If we unmap a page to write it to disk then we cannot allow a CPU to write to the physical page being written through a cached entry. -- Mel Gorman SUSE Labs