Hi Michal, On Thu, Nov 06, 2014 at 10:08:45AM +0100, Michal Hocko wrote: > I have encountered the same error as well. We need to move the forward > declaration up outside of CONFIG_NUMA: Yes, that's my fault, I'm sorry. Thank you for fixing this. BTW what do you think about the whole patch set that introduced it - https://lkml.org/lkml/2014/11/3/781 - w/o diving deeply into details, just by looking at the general idea described in the cover letter? Does it look like acceptable to you that a cgroup can get a cache with some objects left from the previous user? Or do you think it's better to give each cgroup its own cache as it used to be before and introduce cache auto-destruction somehow (that would be tricky though, but possible)? Or perhaps it'd be better to get rid of per-memcg caches altogether and share the same kmem cache for all kmem allocations keeping a pointer to the owner memcg in each kmem object? I'd really appreciate if you or Johannes could share your thoughts on it, because I'm afraid I can do something everybody will regret about in the future... Thanks, Vladimir -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxx. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>