Re: [resend][PATCH -mm] split_lru: fix pagevec_move_tail() doesn't treat unevictable page

[MinChan Kim]

So sorry to make a noisy . :-<

Gmail is mangling my mail.
I hope this is last.

 CPU1                                                           CPU2

 shm_unlock
 scan_mapping_unevictable_pages
 check_move_unevictable_page
 ClearPageUnevictable
                                                      rotate_reclaimable_page
							PageUnevictable(page) return 0

 SetPageUnevictable
 list_move(LRU_UNEVICTABLE)

                                                      local_irq_save
                                                      pagevec_move_tail



On Wed, Jul 2, 2008 at 9:49 AM, MinChan Kim <minchan.kim@xxxxxxxxx> wrote:
> Sorry, it seems mail sender problem.
> Resend
>
>  CPU1                                                           CPU2
>
>  shm_unlock
>  scan_mapping_unevictable_pages
>  check_move_unevictable_page
>  ClearPageUnevictable
>                                                      rotate_reclaimable_page
>
> PageUnevictable(page) return 0
>  SetPageUnevictable
>  list_move(LRU_UNEVICTABLE)
>
>                                                      local_irq_save
>                                                      pagevec_move_tail
>
>
> On Wed, Jul 2, 2008 at 9:39 AM, MinChan Kim <minchan.kim@xxxxxxxxx> wrote:
>> Hi, Rik and Kosaki-san
>>
>> I want to know exact race situation for remaining git log.
>> As you know, git log is important for me who is newbie to understand source
>>
>> There are many possibility in this race problem.
>>
>> Did you use hugepage in this test ?
>> I think that If you used hugepage, it seems to happen following race.
>>
>> --------------
>>
>> CPU1                                                           CPU2
>>
>> shm_unlock
>> scan_mapping_unevictable_pages
>> check_move_unevictable_page
>> ClearPageUnevictable                                 rotate_reclaimable_page
>>
>> PageUnevictable(page) return 0
>> SetPageUnevictable
>> list_move(LRU_UNEVICTABLE)
>>
>> local_irq_save
>>
>> pagevec_move_tail
>>
>> Do you think it is possible ?
>>
>> On Tue, Jul 1, 2008 at 10:38 PM, Rik van Riel <riel@xxxxxxxxxx> wrote:
>>> On Tue, 01 Jul 2008 17:26:51 +0900
>>> KOSAKI Motohiro <kosaki.motohiro@xxxxxxxxxxxxxx> wrote:
>>>
>>>> Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@xxxxxxxxxxxxxx>
>>>
>>> Acked-by: Rik van Riel <riel@xxxxxxxxxx>
>>>
>>> Good catch!
>>>
>>>> @@ -116,7 +116,7 @@ static void pagevec_move_tail(struct pag
>>>>                       zone = pagezone;
>>>>                       spin_lock(&zone->lru_lock);
>>>>               }
>>>> -             if (PageLRU(page) && !PageActive(page)) {
>>>> +             if (PageLRU(page) && !PageActive(page) && !PageUnevictable(page)) {
>>>>                       int lru = page_is_file_cache(page);
>>>>                       list_move_tail(&page->lru, &zone->lru[lru].list);
>>>>                       pgmoved++;
>>>
>>> --
>>> All rights reversed.
>>>
>>
>>
>>
>> --
>> Kinds regards,
>> MinChan Kim
>>
>
>
>
> --
> Kinds regards,
> MinChan Kim
>



-- 
Kinds regards,
MinChan Kim
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/