On 07/06/2012 12:16 PM, Christoph Lameter wrote:

On Mon, 18 Jun 2012, Glauber Costa wrote:

quoted

In the slub allocator, when the last object of a page goes away, we
don't necessarily free it - there is not necessarily a test for empty
page in any slab_free path.

This is the same btw in SLAB which keeps objects in per cpu caches and
keeps empty slab pages on special queues.

quoted

This patch marks all memcg caches as dead. kmem_cache_shrink is called
for the ones who are not yet dead - this will force internal cache
reorganization, and then all references to empty pages will be removed.

You need to call this also for slab to drain the caches and free the pages
on the empty list.

Doesn't the SLAB have a time-based reaper for that?

That's why I was less concerned with the SLAB, but I can certainly call
it for both.