On Mon, Apr 18, 2011 at 10:05:57PM +0200, Samuel Thibault wrote:
> I don't think this patch can be applied. Building a list and then
> freeing it was done for some reason, at least because kmem_free, (thus
> vm_map_delete), may actually need to use zalloc in order to split
> some areas (which is ac
Maksym Planeta, le Mon 18 Apr 2011 00:13:14 +0300, a écrit :
> When zone_gc starts collecting free memory, first it form a list
> (zone_free_page_list) of free
> pages and then by the end of function it frees this list.
>
> Patch makes releasing of free pages on spot. So this list isn't
> needed a
Maksym Planeta, le Mon 18 Apr 2011 00:13:14 +0300, a écrit :
> When zone_gc starts collecting free memory, first it form a list
> (zone_free_page_list) of free
> pages and then by the end of function it frees this list.
>
> Patch makes releasing of free pages on spot. So this list isn't
> needed a
When zone_gc starts collecting free memory, first it form a list
(zone_free_page_list) of free
pages and then by the end of function it frees this list.
Patch makes releasing of free pages on spot. So this list isn't
needed and one for loop becomes less.
There was a risk of deadlocking when kmem_
