From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-lb0-f197.google.com (mail-lb0-f197.google.com [209.85.217.197]) by kanga.kvack.org (Postfix) with ESMTP id F34436B007E for ; Thu, 2 Jun 2016 14:56:35 -0400 (EDT) Received: by mail-lb0-f197.google.com with SMTP id q17so27770042lbn.3 for ; Thu, 02 Jun 2016 11:56:35 -0700 (PDT) Received: from e06smtp06.uk.ibm.com (e06smtp06.uk.ibm.com. [195.75.94.102]) by mx.google.com with ESMTPS id w8si2377554wjz.220.2016.06.02.11.56.34 for (version=TLS1_2 cipher=AES128-SHA bits=128/128); Thu, 02 Jun 2016 11:56:34 -0700 (PDT) Received: from localhost by e06smtp06.uk.ibm.com with IBM ESMTP SMTP Gateway: Authorized Use Only! Violators will be prosecuted for from ; Thu, 2 Jun 2016 19:56:33 +0100 Received: from b06cxnps4075.portsmouth.uk.ibm.com (d06relay12.portsmouth.uk.ibm.com [9.149.109.197]) by d06dlp03.portsmouth.uk.ibm.com (Postfix) with ESMTP id 3168B1B0804B for ; Thu, 2 Jun 2016 19:57:33 +0100 (BST) Received: from d06av08.portsmouth.uk.ibm.com (d06av08.portsmouth.uk.ibm.com [9.149.37.249]) by b06cxnps4075.portsmouth.uk.ibm.com (8.14.9/8.14.9/NCO v10.0) with ESMTP id u52IuTbE65536050 for ; Thu, 2 Jun 2016 18:56:29 GMT Received: from d06av08.portsmouth.uk.ibm.com (localhost [127.0.0.1]) by d06av08.portsmouth.uk.ibm.com (8.14.4/8.14.4/NCO v10.0 AVout) with ESMTP id u52IuTH8030128 for ; Thu, 2 Jun 2016 12:56:29 -0600 Subject: Re: [BUG/REGRESSION] THP: broken page count after commit aa88b68c References: <20160602172141.75c006a9@thinkpad> <20160602155149.GB8493@node.shutemov.name> <20160602114031.64b178c823901c171ec82745@linux-foundation.org> From: Christian Borntraeger Message-ID: <5750815B.1090302@de.ibm.com> Date: Thu, 2 Jun 2016 20:56:27 +0200 MIME-Version: 1.0 In-Reply-To: <20160602114031.64b178c823901c171ec82745@linux-foundation.org> Content-Type: text/plain; charset=windows-1252 Content-Transfer-Encoding: 7bit Sender: owner-linux-mm@kvack.org List-ID: To: Andrew Morton , "Kirill A. Shutemov" Cc: Gerald Schaefer , "Kirill A. Shutemov" , Andrea Arcangeli , "Aneesh Kumar K.V" , Mel Gorman , Hugh Dickins , Johannes Weiner , Dave Hansen , Vlastimil Babka , Linus Torvalds , linux-mm@kvack.org, linux-kernel@vger.kernel.org, Martin Schwidefsky , Heiko Carstens On 06/02/2016 08:40 PM, Andrew Morton wrote: > On Thu, 2 Jun 2016 18:51:50 +0300 "Kirill A. Shutemov" wrote: > >> On Thu, Jun 02, 2016 at 05:21:41PM +0200, Gerald Schaefer wrote: >>> Christian Borntraeger reported a kernel panic after corrupt page counts, >>> and it turned out to be a regression introduced with commit aa88b68c >>> "thp: keep huge zero page pinned until tlb flush", at least on s390. >>> >>> put_huge_zero_page() was moved over from zap_huge_pmd() to release_pages(), >>> and it was replaced by tlb_remove_page(). However, release_pages() might >>> not always be triggered by (the arch-specific) tlb_remove_page(). >>> >>> On s390 we call free_page_and_swap_cache() from tlb_remove_page(), and not >>> tlb_flush_mmu() -> free_pages_and_swap_cache() like the generic version, >>> because we don't use the MMU-gather logic. Although both functions have very >>> similar names, they are doing very unsimilar things, in particular >>> free_page_xxx is just doing a put_page(), while free_pages_xxx calls >>> release_pages(). >>> >>> This of course results in very harmful put_page()s on the huge zero page, >>> on architectures where tlb_remove_page() is implemented in this way. It >>> seems to affect only s390 and sh, but sh doesn't have THP support, so >>> the problem (currently) probably only exists on s390. >>> >>> The following quick hack fixed the issue: >>> >>> diff --git a/mm/swap_state.c b/mm/swap_state.c >>> index 0d457e7..c99463a 100644 >>> --- a/mm/swap_state.c >>> +++ b/mm/swap_state.c >>> @@ -252,7 +252,10 @@ static inline void free_swap_cache(struct page *page) >>> void free_page_and_swap_cache(struct page *page) >>> { >>> free_swap_cache(page); >>> - put_page(page); >>> + if (is_huge_zero_page(page)) >>> + put_huge_zero_page(); >>> + else >>> + put_page(page); >>> } >>> >>> /* >> >> The fix looks good to me. > > Yes. A bit regrettable, but that's what release_pages() does. > > Can we have a signed-off-by please? Please also add CC: stable for 4.6 -- 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