Messages in this thread | | | Date | Wed, 7 May 2014 14:45:33 +0200 | From | Michal Hocko <> | Subject | Re: [PATCH -mm 2/2] memcg: cleanup kmem cache creation/destruction functions naming |
| |
On Wed 07-05-14 12:15:30, Vladimir Davydov wrote: > Current names are rather inconsistent. Let's try to improve them.
Yes the old names are a giant mess. I am not sure the new ones are that much better however.
> Brief change log: > > ** old name ** ** new name ** > > kmem_cache_create_memcg kmem_cache_request_memcg_copy
Both are bad because the first suggests we are creating memcg and the second that we are requesting a copy of memcg.
memcg_alloc_kmem_cache?
_copy suffix is a bit confusing. E.g. copy_mm and others either to shallow or deep copy depending on the context. This one always creats a deep copy. Also why it is imporatant to treat the created caches as copies?
> memcg_kmem_create_cache memcg_copy_kmem_cache
memcg_register_kmem_cache? It also allocates so this name is a bit awkward as well.
> memcg_kmem_destroy_cache memcg_destroy_kmem_cache_copy
memcg_unregister_kmem_cache to match the above?
> __kmem_cache_destroy_memcg_children __kmem_cache_destroy_memcg_copies > kmem_cache_destroy_memcg_children kmem_cache_destroy_memcg_copies
_children suffix is really confusing because they have different meaning in memcg and refer to children groups.
memcg_cleanup_kmem_chache_memcg_params? It doesn't have to live in the kmem_cache namespace because it only does only memcg kmem specific stuff.
> mem_cgroup_destroy_all_caches memcg_destroy_kmem_cache_copies > > create_work memcg_kmem_cache_copy_work
memcg_register_cache_work?
> memcg_create_cache_work_func memcg_kmem_cache_copy_work_func
memcg_register_cache_func?
> memcg_create_cache_enqueue memcg_schedule_kmem_cache_copy
memcg_schedule_register_cache?
-- Michal Hocko SUSE Labs
| |