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 SMTP id 467A66B01A7 for ; Mon, 15 Mar 2010 02:28:17 -0400 (EDT) Received: by pwi4 with SMTP id 4so126193pwi.14 for ; Sun, 14 Mar 2010 23:28:15 -0700 (PDT) MIME-Version: 1.0 In-Reply-To: <20100315143420.6ec3bdf9.kamezawa.hiroyu@jp.fujitsu.com> References: <1268412087-13536-1-git-send-email-mel@csn.ul.ie> <1268412087-13536-3-git-send-email-mel@csn.ul.ie> <28c262361003141728g4aa40901hb040144c5a4aeeed@mail.gmail.com> <20100315143420.6ec3bdf9.kamezawa.hiroyu@jp.fujitsu.com> Date: Mon, 15 Mar 2010 15:28:15 +0900 Message-ID: <28c262361003142328w610f0478sbc17880ffa454fe8@mail.gmail.com> Subject: Re: [PATCH 02/11] mm,migration: Do not try to migrate unmapped anonymous pages From: Minchan Kim Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Sender: owner-linux-mm@kvack.org To: KAMEZAWA Hiroyuki Cc: Mel Gorman , Andrew Morton , Andrea Arcangeli , Christoph Lameter , Adam Litke , Avi Kivity , David Rientjes , KOSAKI Motohiro , Rik van Riel , linux-kernel@vger.kernel.org, linux-mm@kvack.org List-ID: On Mon, Mar 15, 2010 at 2:34 PM, KAMEZAWA Hiroyuki wrote: > On Mon, 15 Mar 2010 09:28:08 +0900 > Minchan Kim wrote: > >> Hi, Mel. >> On Sat, Mar 13, 2010 at 1:41 AM, Mel Gorman wrote: >> > rmap_walk_anon() was triggering errors in memory compaction that looks= like >> > use-after-free errors in anon_vma. The problem appears to be that betw= een >> > the page being isolated from the LRU and rcu_read_lock() being taken, = the >> > mapcount of the page dropped to 0 and the anon_vma was freed. This pat= ch >> > skips the migration of anon pages that are not mapped by anyone. >> > >> > Signed-off-by: Mel Gorman >> > Acked-by: Rik van Riel >> > --- >> > =C2=A0mm/migrate.c | =C2=A0 10 ++++++++++ >> > =C2=A01 files changed, 10 insertions(+), 0 deletions(-) >> > >> > diff --git a/mm/migrate.c b/mm/migrate.c >> > index 98eaaf2..3c491e3 100644 >> > --- a/mm/migrate.c >> > +++ b/mm/migrate.c >> > @@ -602,6 +602,16 @@ static int unmap_and_move(new_page_t get_new_page= , unsigned long private, >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0 * just care Anon page here. >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0 */ >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0if (PageAnon(page)) { >> > + =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* If the page= has no mappings any more, just bail. An >> > + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0* unmapped an= on page is likely to be freed soon but worse, >> > + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0* it's possib= le its anon_vma disappeared between when >> > + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0* the page wa= s isolated and when we reached here while >> > + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0* the RCU loc= k was not held >> > + =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 if (!page_mapcount(= page)) >> >> As looking code about mapcount of page, I got confused. >> I think mapcount of page is protected by pte lock. >> But I can't find pte lock in unmap_and_move. > There is no pte_lock. > >> If I am right, what protects race between this condition check and >> rcu_read_lock? >> This patch makes race window very small but It can't remove race totally= . >> >> I think I am missing something. >> Pz, point me out. :) >> > > Hmm. This is my understanding of old story. > > At migration. > =C2=A01. we increase page_count(). > =C2=A02. isolate it from LRU. > =C2=A03. call try_to_unmap() under rcu_read_lock(). Then, > =C2=A04. replace pte with swp_entry_t made by PFN. under pte_lock. > =C2=A05. do migarate > =C2=A06. remap new pages. under pte_lock()> > =C2=A07. release rcu_read_lock(). > > Here, we don't care whether page->mapping holds valid anon_vma or not. > > Assume a racy threads which calls zap_pte_range() (or some other) > > a) When the thread finds valid pte under pte_lock and successfully call > =C2=A0 page_remove_rmap(). > =C2=A0 In this case, migration thread finds try_to_unmap doesn't unmap an= y pte. > =C2=A0 Then, at 6, remap pte will not work. > b) When the thread finds migrateion PTE(as swap entry) in zap_page_range(= ). > =C2=A0 In this case, migration doesn't find migrateion PTE and remap fail= s. > > Why rcu_read_lock() is necessary.. > =C2=A0- When page_mapcount() goes to 0, we shouldn't trust page->mapping = is valid. > =C2=A0- Possible cases are > =C2=A0 =C2=A0 =C2=A0 =C2=A0i) anon_vma (=3D page->mapping) is freed and u= sed for other object. > =C2=A0 =C2=A0 =C2=A0 =C2=A0ii) anon_vma (=3D page->mapping) is freed > =C2=A0 =C2=A0 =C2=A0 =C2=A0iii) anon_vma (=3D page->mapping) is freed and= used as anon_vma again. > > Here, anon_vma_cachep is created =C2=A0by SLAB_DESTROY_BY_RCU. Then, poss= ible cases > are only ii) and iii). While anon_vma is anon_vma, try_to_unmap and remap= _page > can work well because of the list of vmas and address check. IOW, remap r= outine > just do nothing if anon_vma is freed. > > I'm not sure by what logic "use-after-free anon_vma" is caught. But yes, > there will be case, "anon_vma is touched after freed.", I think. > > Thanks, > -Kame > Thanks for detail explanation, Kame. But it can't understand me enough, Sorry. Mel said he met "use-after-free errors in anon_vma". So added the check in unmap_and_move. if (PageAnon(page)) { .... if (!page_mapcount(page)) goto uncharge; rcu_read_lock(); My concern what protects racy mapcount of the page? For example, CPU A CPU B unmap_and_move page_mapcount check pass zap_pte_range <-- some stall --> pte_lock <-- some stall --> page_remove_rmap(map_count is zero!) <-- some stall --> pte_unlock <-- some stall --> anon_vma_unlink <-- some stall --> anon_vma free !!!! rcu_read_lock anon_vma has gone!! I think above scenario make error "use-after-free", again. What prevent above scenario? --=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/ . Don't email: email@kvack.org