linux-kernel - Re: [PATCH v6 1/4] mm,hwpoison,hugetlb,memory_hotplug: hotremove memory section with hwpoisoned hugepage

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: <8a89e602-f80c-f5b9-ca54-7082a429bd52@huawei.com>
Date:   Mon, 17 Oct 2022 21:29:44 +0800
From:   Miaohe Lin <linmiaohe@...wei.com>
To:     HORIGUCHI NAOYA(堀口 直也) 
        <naoya.horiguchi@....com>
CC:     Naoya Horiguchi <naoya.horiguchi@...ux.dev>,
        "linux-mm@...ck.org" <linux-mm@...ck.org>,
        Andrew Morton <akpm@...ux-foundation.org>,
        David Hildenbrand <david@...hat.com>,
        Mike Kravetz <mike.kravetz@...cle.com>,
        Yang Shi <shy828301@...il.com>,
        Oscar Salvador <osalvador@...e.de>,
        Muchun Song <songmuchun@...edance.com>,
        Jane Chu <jane.chu@...cle.com>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v6 1/4] mm,hwpoison,hugetlb,memory_hotplug: hotremove
 memory section with hwpoisoned hugepage

On 2022/10/17 15:24, HORIGUCHI NAOYA(堀口 直也) wrote:
> On Sat, Oct 15, 2022 at 09:58:09AM +0800, Miaohe Lin wrote:
>> On 2022/10/7 9:07, Naoya Horiguchi wrote:
> ...
>>> @@ -1785,7 +1785,8 @@ void hugetlb_clear_page_hwpoison(struct page *hpage)
>>>   *   -EBUSY        - the hugepage is busy (try to retry)
>>>   *   -EHWPOISON    - the hugepage is already hwpoisoned
>>>   */
>>> -int __get_huge_page_for_hwpoison(unsigned long pfn, int flags)
>>> +int __get_huge_page_for_hwpoison(unsigned long pfn, int flags,
>>> +				 bool *migratable_cleared)
>>>  {
>>>  	struct page *page = pfn_to_page(pfn);
>>>  	struct page *head = compound_head(page);
>>> @@ -1815,6 +1816,15 @@ int __get_huge_page_for_hwpoison(unsigned long pfn, int flags)
>>>  		goto out;
>>>  	}
>>>  
>>> +	/*
>>> +	 * Clearing HPageMigratable for hwpoisoned hugepages to prevent them
>>> +	 * from being migrated by memory hotremove.
>>> +	 */
>>> +	if (count_increased) {
>>> +		*migratable_cleared = true;
>>> +		ClearHPageMigratable(head);
>>
>> I think I might be nitpicking... But it seems ClearHPageMigratable is not enough here.
>>   1. In MF_COUNT_INCREASED case, we don't know whether HPageMigratable is set.
>>   2. Even if HPageMigratable is set, there might be a race window before we clear HPageMigratable?
> 
> Maybe this race is what I mentioned in
> https://lore.kernel.org/linux-mm/20220928012647.GA597297@u2004.lan/
> (the second scenario).  My stance to this case is that the hugepage is not
> hwpoisoned even if some hardware (not visible to kernel) issue is in it,
> so hwpoison handler can/may not cope with the race.
> I guess that this could be handled by applying memcpy_mcsafe() mechanism
> to page migration, but I'm not sure of the feasibility.

Thanks Naoya. memcpy_mcsafe() might be a good idea to handle hwpoison with the memory copy in
page migration path. And [1] is doing the similar thing. If this mechanism is applicable, then
we could handle more memory failure scene. ;)

[1] https://lore.kernel.org/linux-mm/20221010160142.1087120-1-jiaqiyan@google.com/

> 
>> So "*migratable_cleared = TestClearHPageMigratable" might be better? But I might be wrong.
> 
> Thanks, this seems work for 1 (I need testing to check it), so I'll do this
> in the next post.

Many thanks for your work.

Thanks,
Miaohe Lin

> 
>>
>> With above fixed (if it's really a problem), this patch looks good to me.
>>
>> Reviewed-by: Miaohe Lin <linmiaohe@...wei.com>
> 
> Thank you very much.
> 
> - Naoya Horiguchi
>