Added correct support for kmem cgroup interaction with SLUB bulk API. Kernel tested with CONFIG_MEMCG_KMEM=y, and memcg setup[4] provide by Vladimir Davydov. And with my network stack use-case patchset applied, to actually activate the API. Patch01: I've verified the loop in slab_post_alloc_hook() gets removed by the compiler (when no debug options defined). This was actually tricky due to kernel gcc compile options, and I wrote a small program to figure this out [1]. Also tested memory exhaustion[3] to verify error: (label) code path. Patch02: If CONFIG_MEMCG_KMEM is enabled, we no longer handle error cases like passing of NULL pointers in the array to free. The "try_crash" mode test of module slab_bulk_test03 [2] have been adjusted. Kernel config wise, ran with combinations of: CONFIG_DEBUG_KMEMLEAK, CONFIG_KASAN and CONFIG_MEMCG_KMEM (Explicitly disabled/avoided CONFIG_KMEMCHECK) [1] https://github.com/netoptimizer/network-testing/blob/master/src/compiler_test01.c [2] https://github.com/netoptimizer/prototype-kernel/blob/master/kernel/mm/slab_bulk_test03.c [3] https://github.com/netoptimizer/prototype-kernel/blob/master/kernel/mm/slab_bulk_test04_exhaust_mem.c [4] http://thread.gmane.org/gmane.linux.kernel.mm/140860/focus=140865 --- Jesper Dangaard Brouer (2): slub: fix kmem cgroup bug in kmem_cache_alloc_bulk slub: add missing kmem cgroup support to kmem_cache_free_bulk mm/slub.c | 46 +++++++++++++++++++++++++++------------------- 1 file changed, 27 insertions(+), 19 deletions(-) -- -- 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>