From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail138.messagelabs.com (mail138.messagelabs.com [216.82.249.35]) by kanga.kvack.org (Postfix) with ESMTP id 1049D6B0074 for ; Wed, 16 Nov 2011 17:07:16 -0500 (EST) Received: by bke17 with SMTP id 17so1571758bke.14 for ; Wed, 16 Nov 2011 14:07:13 -0800 (PST) MIME-Version: 1.0 In-Reply-To: References: <20111114140421.GA27150@suse.de> <20111115173656.GJ27150@suse.de> <20111116002235.GA10958@barrios-laptop.redhat.com> <20111116004516.GA13028@barrios-laptop.redhat.com> Date: Thu, 17 Nov 2011 07:07:13 +0900 Message-ID: Subject: Re: [PATCH] mm: avoid livelock on !__GFP_FS allocations From: Minchan Kim Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Sender: owner-linux-mm@kvack.org List-ID: To: David Rientjes Cc: Pekka Enberg , Minchan Kim , Colin Cross , Mel Gorman , Andrew Morton , Pekka Enberg , KAMEZAWA Hiroyuki , Andrea Arcangeli , LKML , Linux-MM , "Rafael J. Wysocki" On Thu, Nov 17, 2011 at 6:44 AM, David Rientjes wrote= : > On Wed, 16 Nov 2011, Pekka Enberg wrote: > >> > diff --git a/kernel/power/suspend.c b/kernel/power/suspend.c >> > index fdd4263..01aa9b5 100644 >> > --- a/kernel/power/suspend.c >> > +++ b/kernel/power/suspend.c >> > @@ -297,9 +297,11 @@ int enter_state(suspend_state_t state) >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0goto Finish; >> > >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0pr_debug("PM: Entering %s sleep\n", pm_stat= es[state]); >> > + =C2=A0 =C2=A0 =C2=A0 oom_killer_disable(); >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0pm_restrict_gfp_mask(); >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0error =3D suspend_devices_and_enter(state); >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0pm_restore_gfp_mask(); >> > + =C2=A0 =C2=A0 =C2=A0 oom_killer_enable(); >> > >> > =C2=A0Finish: >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0pr_debug("PM: Finishing wakeup.\n"); >> > diff --git a/mm/page_alloc.c b/mm/page_alloc.c >> > index 6e8ecb6..d8c31b7 100644 >> > --- a/mm/page_alloc.c >> > +++ b/mm/page_alloc.c >> > @@ -2177,9 +2177,9 @@ rebalance: >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0 * running out of options and have to consi= der going OOM >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0 */ >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0if (!did_some_progress) { >> > - =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 if ((gfp_mask & __G= FP_FS) && !(gfp_mask & __GFP_NORETRY)) { >> > - =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2= =A0 =C2=A0 if (oom_killer_disabled) >> > + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 if (oom_killer_disa= bled) >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 = =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0goto nopage; > > You're allowing __GFP_NOFAIL allocations to fail. > >> > + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 if ((gfp_mask & __G= FP_FS) && !(gfp_mask & __GFP_NORETRY)) { >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 = =C2=A0 =C2=A0page =3D __alloc_pages_may_oom(gfp_mask, order, >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 = =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0zoneli= st, high_zoneidx, >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 = =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0nodema= sk, preferred_zone, >> > >> >> I'd prefer something like this. The whole 'gfp_allowed_flags' thing was >> designed to make GFP_KERNEL work during boot time where it's obviously s= afe to >> do that. I really don't think that's going to work suspend cleanly. >> > > Adding Rafael to the cc. > > This has been done since 2.6.34 and presumably has been working quite > well. =C2=A0I don't have a specific objection to gfp_allowed_flags to be = used > outside of boot since it seems plausible that there are system-level > contexts that would need different behavior in the page allocator and thi= s > does it effectively without major surgery or a slower fastpath. =C2=A0Sus= pend > is using it just like boot does before irqs are enabled, so I don't have > an objection to it. > My point isn't using gfp_allowed_flags(maybe it's Pekka's concern) but why adding new special case handling code like pm_suspended_storage. I think we can handle the issue with oom_killer_disabled(but the naming is = bad) --=20 Kind regards, Minchan Kim -- 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