From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Message-ID: <5425378A.4090702@suse.cz> Date: Fri, 26 Sep 2014 11:53:14 +0200 From: Jiri Slaby MIME-Version: 1.0 To: Mel Gorman CC: Linux-Stable , LKML Subject: Re: [PATCH 08/97] mm, thp: do not allow thp faults to avoid cpuset restrictions References: <1409250945-30874-1-git-send-email-mgorman@suse.de> <1409250945-30874-9-git-send-email-mgorman@suse.de> In-Reply-To: <1409250945-30874-9-git-send-email-mgorman@suse.de> Content-Type: text/plain; charset=iso-8859-2 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: On 08/28/2014, 08:34 PM, Mel Gorman wrote: > From: David Rientjes > > commit b104a35d32025ca740539db2808aa3385d0f30eb upstream. > > The page allocator relies on __GFP_WAIT to determine if ALLOC_CPUSET > should be set in allocflags. ALLOC_CPUSET controls if a page allocation > should be restricted only to the set of allowed cpuset mems. > > Transparent hugepages clears __GFP_WAIT when defrag is disabled to prevent > the fault path from using memory compaction or direct reclaim. Thus, it > is unfairly able to allocate outside of its cpuset mems restriction as a > side-effect. > > This patch ensures that ALLOC_CPUSET is only cleared when the gfp mask is > truly GFP_ATOMIC by verifying it is also not a thp allocation. Hi, the time has come to apply this series. Thanks a lot for doing that. It all applies cleanly except this one, as it is already in 2.6.27. So skipping makes it work nicely. -- js suse labs