On 05/11/2012 03:56 PM, Christoph Lameter wrote:

On Fri, 11 May 2012, Glauber Costa wrote:

quoted

So we don't mix pages from multiple memcgs in the same cache - we believe that
would be too confusing.

Well subsystem create caches and other things that are shared between
multiple processes. How can you track that?

Each process that belongs to a memcg triggers the creation of a new 
child kmem cache.

quoted

/proc/slabinfo reflects this information, by listing the memcg-specific slabs.

What about /sys/kernel/slab/*?

 From the PoV of the global system, what you'll see is something like:
dentry , dentry(2:memcg1), dentry(2:memcg2), etc.

No attempt is made to provide any view of those caches in a logically 
grouped way - the global system can view them independently.