On Tue, Jun 26, 2012 at 04:48:08PM +0400, Glauber Costa wrote:

On 06/25/2012 10:38 PM, Tejun Heo wrote:

quoted

On Mon, Jun 25, 2012 at 06:55:35PM +0200, Frederic Weisbecker wrote:

quoted

On 06/25/2012 04:15 PM, Glauber Costa wrote:

quoted

Because those architectures will draw their stacks directly from
the page allocator, rather than the slab cache, we can directly
pass __GFP_KMEMCG flag, and issue the corresponding free_pages.

This code path is taken when the architecture doesn't define
CONFIG_ARCH_THREAD_INFO_ALLOCATOR (only ia64 seems to), and has
THREAD_SIZE >= PAGE_SIZE. Luckily, most - if not all - of the
remaining architectures fall in this category.

This will guarantee that every stack page is accounted to the memcg
the process currently lives on, and will have the allocations to fail
if they go over limit.

For the time being, I am defining a new variant of THREADINFO_GFP, not
to mess with the other path. Once the slab is also tracked by memcg,
we can get rid of that flag.

Tested to successfully protect against :(){ :|:& };:

Signed-off-by: Glauber Costa <redacted>
CC: Christoph Lameter <redacted>
CC: Pekka Enberg <penberg-bbCR+/B0CizivPeTLB3BmA@public.gmane.org>
CC: Michal Hocko <redacted>
CC: Kamezawa Hiroyuki <redacted>
CC: Johannes Weiner <redacted>
CC: Suleiman Souhlal <redacted>

Acked-by: Frederic Weisbecker <redacted>

Frederic, does this (with proper slab accounting added later) achieve
what you wanted with the task counter?

A note: Frederic may confirm, but I think he doesn't even need
the slab accounting to follow to achieve that goal.

Limiting is enough. But that requires internal accounting.