| 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
| ||
|
Message-ID: <56862fc0-3e4b-8d1e-ae15-0df32bf5e4c0@virtuozzo.com>
Date: Mon, 18 Mar 2019 12:33:15 +0300
From: Kirill Tkhai <ktkhai@...tuozzo.com>
To: Pankaj Suryawanshi <pankaj.suryawanshi@...fochips.com>,
Vlastimil Babka <vbabka@...e.cz>,
Michal Hocko <mhocko@...nel.org>,
"aneesh.kumar@...ux.ibm.com" <aneesh.kumar@...ux.ibm.com>
Cc: "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"minchan@...nel.org" <minchan@...nel.org>,
"linux-mm@...ck.org" <linux-mm@...ck.org>,
"khandual@...ux.vnet.ibm.com" <khandual@...ux.vnet.ibm.com>,
"hillf.zj@...baba-inc.com" <hillf.zj@...baba-inc.com>
Subject: Re: [External] Re: vmscan: Reclaim unevictable pages
Hi, Pankaj,
On 18.03.2019 12:09, Pankaj Suryawanshi wrote:
>
> Hello
>
> shrink_page_list() returns , number of pages reclaimed, when pages is unevictable it returns VM_BUG_ON_PAGE(PageLRU(page) || PageUnevicatble(page),page);
the general idea is shrink_page_list() can't iterate PageUnevictable() pages.
PageUnevictable() pages are never being added to lists, which shrink_page_list()
uses for iteration. Also, a page can't be marked as PageUnevictable(), when
it's attached to a shrinkable list.
So, the problem should be somewhere outside shrink_page_list().
I won't suggest you something about CMA, since I haven't dived in that code.
> We can add the unevictable pages in reclaim list in shrink_page_list(), return total number of reclaim pages including unevictable pages, let the caller handle unevictable pages.
>
> I think the problem is shrink_page_list is awkard. If page is unevictable it goto activate_locked->keep_locked->keep lables, keep lable list_add the unevictable pages and throw the VM_BUG instead of passing it to caller while it relies on caller for non-reclaimed-non-unevictable page's putback.
> I think we can make it consistent so that shrink_page_list could return non-reclaimed pages via page_list and caller can handle it. As an advance, it could try to migrate mlocked pages without retrial.
>
>
> Below is the issue of CMA_ALLOC of large size buffer : (Kernel version - 4.14.65 (On Android pie [ARM])).
>
> [ 24.718792] page dumped because: VM_BUG_ON_PAGE(PageLRU(page) || PageUnevictable(page))
> [ 24.726949] page->mem_cgroup:bd008c00
> [ 24.730693] ------------[ cut here ]------------
> [ 24.735304] kernel BUG at mm/vmscan.c:1350!
> [ 24.739478] Internal error: Oops - BUG: 0 [#1] PREEMPT SMP ARM
>
>
> Below is the patch which solved this issue :
>
> diff --git a/mm/vmscan.c b/mm/vmscan.c
> index be56e2e..12ac353 100644
> --- a/mm/vmscan.c
> +++ b/mm/vmscan.c
> @@ -998,7 +998,7 @@ static unsigned long shrink_page_list(struct list_head *page_list,
> sc->nr_scanned++;
>
> if (unlikely(!page_evictable(page)))
> - goto activate_locked;
> + goto cull_mlocked;
>
> if (!sc->may_unmap && page_mapped(page))
> goto keep_locked;
> @@ -1331,7 +1331,12 @@ static unsigned long shrink_page_list(struct list_head *page_list,
> } else
> list_add(&page->lru, &free_pages);
> continue;
> -
> +cull_mlocked:
> + if (PageSwapCache(page))
> + try_to_free_swap(page);
> + unlock_page(page);
> + list_add(&page->lru, &ret_pages);
> + continue;
> activate_locked:
> /* Not a candidate for swapping, so reclaim swap space. */
> if (PageSwapCache(page) && (mem_cgroup_swap_full(page) ||
>
>
>
>
> It fixes the below issue.
>
> 1. Large size buffer allocation using cma_alloc successful with unevictable pages.
>
> cma_alloc of current kernel will fail due to unevictable page
>
> Please let me know if anything i am missing.
>
> Regards,
> Pankaj
>
> From: Vlastimil Babka <vbabka@...e.cz>
> Sent: 18 March 2019 14:12:50
> To: Pankaj Suryawanshi; Kirill Tkhai; Michal Hocko; aneesh.kumar@...ux.ibm.com
> Cc: linux-kernel@...r.kernel.org; minchan@...nel.org; linux-mm@...ck.org; khandual@...ux.vnet.ibm.com; hillf.zj@...baba-inc.com
> Subject: Re: [External] Re: vmscan: Reclaim unevictable pages
>
>
> On 3/15/19 11:11 AM, Pankaj Suryawanshi wrote:
>>
>> [ cc Aneesh kumar, Anshuman, Hillf, Vlastimil]
>
> Can you send a proper patch with changelog explaining the change? I
> don't know the context of this thread.
>
>> From: Pankaj Suryawanshi
>> Sent: 15 March 2019 11:35:05
>> To: Kirill Tkhai; Michal Hocko
>> Cc: linux-kernel@...r.kernel.org; minchan@...nel.org; linux-mm@...ck.org
>> Subject: Re: Re: [External] Re: vmscan: Reclaim unevictable pages
>>
>>
>>
>> [ cc linux-mm ]
>>
>>
>> From: Pankaj Suryawanshi
>> Sent: 14 March 2019 19:14:40
>> To: Kirill Tkhai; Michal Hocko
>> Cc: linux-kernel@...r.kernel.org; minchan@...nel.org
>> Subject: Re: Re: [External] Re: vmscan: Reclaim unevictable pages
>>
>>
>>
>> Hello ,
>>
>> Please ignore the curly braces, they are just for debugging.
>>
>> Below is the updated patch.
>>
>>
>> diff --git a/mm/vmscan.c b/mm/vmscan.c
>> index be56e2e..12ac353 100644
>> --- a/mm/vmscan.c
>> +++ b/mm/vmscan.c
>> @@ -998,7 +998,7 @@ static unsigned long shrink_page_list(struct list_head *page_list,
>> sc->nr_scanned++;
>>
>> if (unlikely(!page_evictable(page)))
>> - goto activate_locked;
>> + goto cull_mlocked;
>>
>> if (!sc->may_unmap && page_mapped(page))
>> goto keep_locked;
>> @@ -1331,7 +1331,12 @@ static unsigned long shrink_page_list(struct list_head *page_list,
>> } else
>> list_add(&page->lru, &free_pages);
>> continue;
>> -
>> +cull_mlocked:
>> + if (PageSwapCache(page))
>> + try_to_free_swap(page);
>> + unlock_page(page);
>> + list_add(&page->lru, &ret_pages);
>> + continue;
>> activate_locked:
>> /* Not a candidate for swapping, so reclaim swap space. */
>> if (PageSwapCache(page) && (mem_cgroup_swap_full(page) ||
>>
>>
>>
>> Regards,
>> Pankaj
>>
>>
>> From: Kirill Tkhai <ktkhai@...tuozzo.com>
>> Sent: 14 March 2019 14:55:34
>> To: Pankaj Suryawanshi; Michal Hocko
>> Cc: linux-kernel@...r.kernel.org; minchan@...nel.org
>> Subject: Re: Re: [External] Re: vmscan: Reclaim unevictable pages
>>
>>
>> On 14.03.2019 11:52, Pankaj Suryawanshi wrote:
>>>
>>> I am using kernel version 4.14.65 (on Android pie [ARM]).
>>>
>>> No additional patches applied on top of vanilla.(Core MM).
>>>
>>> If I change in the vmscan.c as below patch, it will work.
>>
>> Sorry, but 4.14.65 does not have braces around trylock_page(),
>> like in your patch below.
>>
>> See https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/tree/mm/vmscan.c?h=v4.14.65
>>
>> [...]
>>
>>>> diff --git a/mm/vmscan.c b/mm/vmscan.c
>>>> index be56e2e..2e51edc 100644
>>>> --- a/mm/vmscan.c
>>>> +++ b/mm/vmscan.c
>>>> @@ -990,15 +990,17 @@ static unsigned long shrink_page_list(struct list_head *page_list,
>>>> page = lru_to_page(page_list);
>>>> list_del(&page->lru);
>>>>
>>>> if (!trylock_page(page)) {
>>>> goto keep;
>>>> }
>>
>> ************************************************************************************************************************************************************* eInfochips Business Disclaimer: This e-mail message and all attachments transmitted with it are intended solely for the use of the addressee and may contain legally privileged and confidential information. If the reader of this message is not the intended recipient, or an employee or agent responsible for delivering this message to the intended recipient, you are hereby notified that any dissemination, distribution, copying, or other use of this message or its attachments is strictly prohibited. If you have received this message in error, please notify the sender immediately by replying to this message and please delete it from your computer. Any views expressed in this message are those of the individual sender unless otherwise stated. Company has taken enough precautions to prevent the spread of viruses. However the company accepts no liability for any damage caused by any virus transmitted by this email. *************************************************************************************************************************************************************
>>
>
>
>
Powered by blists - more mailing lists