From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S935696Ab3BOLoh (ORCPT ); Fri, 15 Feb 2013 06:44:37 -0500 Received: from mail-ea0-f182.google.com ([209.85.215.182]:35957 "EHLO mail-ea0-f182.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753827Ab3BOLof (ORCPT ); Fri, 15 Feb 2013 06:44:35 -0500 Date: Fri, 15 Feb 2013 12:44:26 +0100 From: Ingo Molnar To: Yinghai Lu , Greg KH Cc: Thomas Gleixner , Linus Torvalds , Linux Kernel Mailing List , Jens Axboe , Alexander Viro , "Theodore Ts'o" , "H. Peter Anvin" , Laura Abbott , Mel Gorman Subject: [-rc7 regression] Buggy commit: "mm: use aligned zone start for pfn_to_bitidx calculation" Message-ID: <20130215114425.GD26955@gmail.com> References: <20130213111007.GA11367@gmail.com> <20130214144510.GC25282@gmail.com> <20130214145424.GA26071@gmail.com> <20130214150810.GA26095@gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: 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 [ Greg: -stable was Cc:-ed for this potentially buggy commit. ] * Yinghai Lu wrote: > On Thu, Feb 14, 2013 at 7:08 AM, Ingo Molnar wrote: > > these three: > > > > 10d73e655cef mm: bootmem: fix free_all_bootmem_core() with odd bitmap alignment > > c060f943d092 mm: use aligned zone start for pfn_to_bitidx calculation > > c0232ae861df mm: memblock: fix wrong memmove size in memblock_merge_regions() > > > > might have highmem=512m boot parameter sensitivity. > > 10d73e655cef should not be related, as we are using nobootmem wrapper > of memblock. > c0232ae861df is copying less while using the blank slot in memblock > memory or reserved array > so it should be ok too. > > c060f943d092 may be related as you config does not have > CONFIG_SPARSEMEM defined. Right, that's the commit causing the x86 regression: c060f943d0929f3e429c5d9522290584f6281d6e is the first bad commit commit c060f943d0929f3e429c5d9522290584f6281d6e Date: Fri Jan 11 14:31:51 2013 -0800 mm: use aligned zone start for pfn_to_bitidx calculation I've given 6d92d4f (its parent commit) quite a bit of testing so I'm 99% sure in the validity of this bisection result. Involved parties Cc:-ed. Revert attached below - it reverts cleanly and I'll test this today, to see whether this solves the sporadic lockups. Thanks, Ingo -------------------> >>From 6f3d5a8f38eb23f792b408d9bc8fee53a7d158f9 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Fri, 15 Feb 2013 12:43:22 +0100 Subject: [PATCH] Revert "mm: use aligned zone start for pfn_to_bitidx calculation" This reverts commit c060f943d0929f3e429c5d9522290584f6281d6e. Causes lockups. Signed-off-by: Ingo Molnar --- mm/page_alloc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/mm/page_alloc.c b/mm/page_alloc.c index cf46618..3a120bb 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -5590,7 +5590,7 @@ static inline int pfn_to_bitidx(struct zone *zone, unsigned long pfn) pfn &= (PAGES_PER_SECTION-1); return (pfn >> pageblock_order) * NR_PAGEBLOCK_BITS; #else - pfn = pfn - round_down(zone->zone_start_pfn, pageblock_nr_pages); + pfn = pfn - zone->zone_start_pfn; return (pfn >> pageblock_order) * NR_PAGEBLOCK_BITS; #endif /* CONFIG_SPARSEMEM */ }