linux-kernel - Re: [PATCH 5/6] mm/page_alloc.c: avoid accessing uninitialized pcp page migratetype

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <bc14adc2-e688-1a46-be08-7626c69e6eb1@huawei.com>
Date:   Wed, 1 Sep 2021 16:18:53 +0800
From:   Miaohe Lin <linmiaohe@...wei.com>
To:     David Hildenbrand <david@...hat.com>
CC:     <akpm@...ux-foundation.org>, <vbabka@...e.cz>,
        <sfr@...b.auug.org.au>, <peterz@...radead.org>,
        <linux-mm@...ck.org>, <linux-kernel@...r.kernel.org>,
        Mel Gorman <mgorman@...hsingularity.net>
Subject: Re: [PATCH 5/6] mm/page_alloc.c: avoid accessing uninitialized pcp
 page migratetype

On 2021/9/1 16:10, David Hildenbrand wrote:
> On 01.09.21 10:02, Miaohe Lin wrote:
>> On 2021/8/31 22:23, David Hildenbrand wrote:
>>> On 31.08.21 15:43, Mel Gorman wrote:
>>>> On Mon, Aug 30, 2021 at 10:10:50PM +0800, Miaohe Lin wrote:
>>>>> If it's not prepared to free unref page, the pcp page migratetype is
>>>>> unset. Thus We will get rubbish from get_pcppage_migratetype() and
>>>>> might list_del &page->lru again after it's already deleted from the
>>>>> list leading to grumble about data corruption.
>>>>>
>>>>> Fixes: 3dcbe270d8ec ("mm/page_alloc: avoid conflating IRQs disabled with zone->lock")
>>>>> Signed-off-by: Miaohe Lin <linmiaohe@...wei.com>
>>>>
>>>> Acked-by: Mel Gorman <mgorman@...hsingularity.net>
>>>>
>>>> This fix is fairly important. Take this patch out and send it on its own
>>>> so it gets picked up relatively quickly. It does not belong in a series
>>>> that is mostly cosmetic cleanups.
>>>
>>> I think the commit id is wrong. Shouldn't that be
>>>
>>> df1acc856923 ("mm/page_alloc: avoid conflating IRQs disabled with zone->lock")
>>>
>>> ?
>>>
>>
>> Many thanks for pointing this out.
>>
>> I used to save the git log in a file to make life easier. But it seems this leads
>> to the old commit id above.
>>
>> commit 3dcbe270d8ec57e534f5c605279cdc3ceb1f044a
>> Author: Mel Gorman <mgorman@...hsingularity.net>
>> Date:   Fri Jun 4 14:20:03 2021 +1000
>>
>>      mm/page_alloc: avoid conflating IRQs disabled with zone->lock
>>
>> git name-rev 3dcbe270d8ec
>> 3dcbe270d8ec tags/next-20210604~2^2~196
>>
>> vs
>>
>> commit df1acc856923c0a65c28b588585449106c316b71
>> Author: Mel Gorman <mgorman@...hsingularity.net>
>> Date:   Mon Jun 28 19:42:00 2021 -0700
>>
>>      mm/page_alloc: avoid conflating IRQs disabled with zone->lock
>>
>> git name-rev df1acc856923
>> df1acc856923 tags/next-20210630~2^2~278
>>
>> Their contents are same but with different commit id. The previous one
>> could have been git-rebased.
>>
> 
> -mm tree commit ids keep changing until patches are upstream. Therefore, you can observe changing commit ids in -next. Always use the ones from Linus' tree, they are stable.
> 
Many thanks for your advice, David. :)