Re: [PATCH 4/4] don't do __ClearPageSlab before freeing slab page.

[Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>]

(2012/06/21 17:13), Glauber Costa wrote:
> On 06/21/2012 12:04 PM, David Rientjes wrote:
>> On Thu, 21 Jun 2012, Glauber Costa wrote:
>>
>>> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
>>> index 6092f33..fdec73e 100644
>>> --- a/mm/page_alloc.c
>>> +++ b/mm/page_alloc.c
>>> @@ -698,8 +698,10 @@ static bool free_pages_prepare(struct page *page, unsigned int order)
>>>
>>> if (PageAnon(page))
>>> page->mapping = NULL;
>>> - for (i = 0; i < (1 << order); i++)
>>> + for (i = 0; i < (1 << order); i++) {
>>> + __ClearPageSlab(page + i);
>>> bad += free_pages_check(page + i);
>>> + }
>>> if (bad)
>>> return false;
>>>
>>> @@ -2561,6 +2563,7 @@ EXPORT_SYMBOL(get_zeroed_page);
>>> void __free_pages(struct page *page, unsigned int order)
>>> {
>>> if (put_page_testzero(page)) {
>>> + __ClearPageSlab(page);
>>> if (order == 0)
>>> free_hot_cold_page(page, 0);
>>> else
>>
>> These are called from a number of different places that has nothing to do
>> with slab so it's certainly out of place here. Is there really no
>> alternative way of doing this?
>
> Well, if the requirement is that we must handle this from the page allocator, how else should I know if I must call the corresponding free functions ?
>
> Also note that other bits are tested inside the page allocator as well, such as MLock.
>
> I saw no other way, but if you have suggestions, I'd be open to try them, of course.
>

I'm sorry I don't understand the logic enough well.

Why check in __free_pages() is better than check in callers of slab.c/slub.c ?

Thanks,
-Kame






--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>