Hi Kirill Tkhai,
Please see mm/vmscan.c in which it first added to list and than throw the error
:
--------------------------------------------------------------------------------------------------
keep:
list_add(&page->lru, &ret_pages);
VM_BUG_ON_PAGE(PageLRU(page) || PageUnevictable(page), page);
---------------------------------------------------------------------------------------------------
Before throwing error, pages are added to list, this is under iteration of
shrink_page_list().
From: Kirill Tkhai <[email protected]>
Sent: 18 March 2019 15:03:15
To: Pankaj Suryawanshi; Vlastimil Babka; Michal Hocko;
[email protected]
Cc: [email protected]; [email protected]; [email protected];
[email protected]; [email protected]
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 <[email protected]>
> Sent: 18 March 2019 14:12:50
> To: Pankaj Suryawanshi; Kirill Tkhai; Michal Hocko; [email protected]
> Cc: [email protected]; [email protected]; [email protected];
> [email protected]; [email protected]
> 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: [email protected]; [email protected]; [email protected]
>> 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: [email protected]; [email protected]
>> 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 <[email protected]>
>> Sent: 14 March 2019 14:55:34
>> To: Pankaj Suryawanshi; Michal Hocko
>> Cc: [email protected]; [email protected]
>> 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.
>> *************************************************************************************************************************************************************
>>
>
>
>
