lkml.org 
[lkml]   [2009]   [Oct]   [13]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH][BUGFIX] vmscan: limit VM_EXEC protection to file pages
On Tue, 13 Oct 2009, Wu Fengguang wrote:

> It is possible to have !Anon but SwapBacked pages, and some apps could
> create huge number of such pages with MAP_SHARED|MAP_ANONYMOUS. These
> pages go into the ANON lru list, and hence shall not be protected: we
> only care mapped executable files. Failing to do so may trigger OOM.
>
> Tested-by: Christian Borntraeger <borntraeger@de.ibm.com>
> Reviewed-by: Rik van Riel <riel@redhat.com>
> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>

I'm not going to stand against this patch. But I would like to point
out that it will "penalize" (well, no longer favour) executables being
run from a tmpfs. Probably not a big deal.

And I want to put on record that (like Andrea) I really loathe this
(vm_flags & VM_EXEC) test: it's a heuristic unlike any other in page
reclaim, and one that is open to any application writer to take unfair
advantage of.

I know that it's there to make some things work better, and that it
has been successful (though now inevitably it's found to require a
tweak - how long until its next tweak?). But I do hope that one
day you will come up with something much more satisfactory here.

Hugh

> ---
> mm/vmscan.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> --- linux.orig/mm/vmscan.c 2009-10-13 09:49:05.000000000 +0800
> +++ linux/mm/vmscan.c 2009-10-13 09:49:37.000000000 +0800
> @@ -1356,7 +1356,7 @@ static void shrink_active_list(unsigned
> * IO, plus JVM can create lots of anon VM_EXEC pages,
> * so we ignore them here.
> */
> - if ((vm_flags & VM_EXEC) && !PageAnon(page)) {
> + if ((vm_flags & VM_EXEC) && page_is_file_cache(page)) {
> list_add(&page->lru, &l_active);
> continue;
> }


\
 
 \ /
  Last update: 2009-10-13 13:47    [W:0.069 / U:0.492 seconds]
©2003-2011 Jasper Spaans. Advertise on this site