memcg: cleanup kmem cache creation/destruction functions naming
authorVladimir Davydov <vdavydov@parallels.com>
Wed, 4 Jun 2014 23:10:02 +0000 (16:10 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Wed, 4 Jun 2014 23:54:08 +0000 (16:54 -0700)
Current names are rather inconsistent. Let's try to improve them.

Brief change log:

** old name **                          ** new name **

kmem_cache_create_memcg                 memcg_create_kmem_cache
memcg_kmem_create_cache                 memcg_regsiter_cache
memcg_kmem_destroy_cache                memcg_unregister_cache

kmem_cache_destroy_memcg_children       memcg_cleanup_cache_params
mem_cgroup_destroy_all_caches           memcg_unregister_all_caches

create_work                             memcg_register_cache_work
memcg_create_cache_work_func            memcg_register_cache_func
memcg_create_cache_enqueue              memcg_schedule_register_cache

Signed-off-by: Vladimir Davydov <vdavydov@parallels.com>
Acked-by: Michal Hocko <mhocko@suse.cz>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
include/linux/memcontrol.h
include/linux/slab.h
mm/memcontrol.c
mm/slab_common.c

index dfc2929..eb65d29 100644 (file)
@@ -505,7 +505,7 @@ __memcg_kmem_get_cache(struct kmem_cache *cachep, gfp_t gfp);
 int __memcg_charge_slab(struct kmem_cache *cachep, gfp_t gfp, int order);
 void __memcg_uncharge_slab(struct kmem_cache *cachep, int order);
 
-int __kmem_cache_destroy_memcg_children(struct kmem_cache *s);
+int __memcg_cleanup_cache_params(struct kmem_cache *s);
 
 /**
  * memcg_kmem_newpage_charge: verify if a new kmem allocation is allowed.
index 86e5b26..1d9abb7 100644 (file)
@@ -116,7 +116,7 @@ struct kmem_cache *kmem_cache_create(const char *, size_t, size_t,
                        unsigned long,
                        void (*)(void *));
 #ifdef CONFIG_MEMCG_KMEM
-struct kmem_cache *kmem_cache_create_memcg(struct mem_cgroup *,
+struct kmem_cache *memcg_create_kmem_cache(struct mem_cgroup *,
                                           struct kmem_cache *,
                                           const char *);
 #endif
index 5e2bfcc..d176edb 100644 (file)
@@ -3132,8 +3132,8 @@ void memcg_free_cache_params(struct kmem_cache *s)
        kfree(s->memcg_params);
 }
 
-static void memcg_kmem_create_cache(struct mem_cgroup *memcg,
-                                   struct kmem_cache *root_cache)
+static void memcg_register_cache(struct mem_cgroup *memcg,
+                                struct kmem_cache *root_cache)
 {
        static char memcg_name_buf[NAME_MAX + 1]; /* protected by
                                                     memcg_slab_mutex */
@@ -3153,7 +3153,7 @@ static void memcg_kmem_create_cache(struct mem_cgroup *memcg,
                return;
 
        cgroup_name(memcg->css.cgroup, memcg_name_buf, NAME_MAX + 1);
-       cachep = kmem_cache_create_memcg(memcg, root_cache, memcg_name_buf);
+       cachep = memcg_create_kmem_cache(memcg, root_cache, memcg_name_buf);
        /*
         * If we could not create a memcg cache, do not complain, because
         * that's not critical at all as we can always proceed with the root
@@ -3175,7 +3175,7 @@ static void memcg_kmem_create_cache(struct mem_cgroup *memcg,
        root_cache->memcg_params->memcg_caches[id] = cachep;
 }
 
-static void memcg_kmem_destroy_cache(struct kmem_cache *cachep)
+static void memcg_unregister_cache(struct kmem_cache *cachep)
 {
        struct kmem_cache *root_cache;
        struct mem_cgroup *memcg;
@@ -3228,7 +3228,7 @@ static inline void memcg_resume_kmem_account(void)
        current->memcg_kmem_skip_account--;
 }
 
-int __kmem_cache_destroy_memcg_children(struct kmem_cache *s)
+int __memcg_cleanup_cache_params(struct kmem_cache *s)
 {
        struct kmem_cache *c;
        int i, failed = 0;
@@ -3239,7 +3239,7 @@ int __kmem_cache_destroy_memcg_children(struct kmem_cache *s)
                if (!c)
                        continue;
 
-               memcg_kmem_destroy_cache(c);
+               memcg_unregister_cache(c);
 
                if (cache_from_memcg_idx(s, i))
                        failed++;
@@ -3248,7 +3248,7 @@ int __kmem_cache_destroy_memcg_children(struct kmem_cache *s)
        return failed;
 }
 
-static void mem_cgroup_destroy_all_caches(struct mem_cgroup *memcg)
+static void memcg_unregister_all_caches(struct mem_cgroup *memcg)
 {
        struct kmem_cache *cachep;
        struct memcg_cache_params *params, *tmp;
@@ -3261,25 +3261,26 @@ static void mem_cgroup_destroy_all_caches(struct mem_cgroup *memcg)
                cachep = memcg_params_to_cache(params);
                kmem_cache_shrink(cachep);
                if (atomic_read(&cachep->memcg_params->nr_pages) == 0)
-                       memcg_kmem_destroy_cache(cachep);
+                       memcg_unregister_cache(cachep);
        }
        mutex_unlock(&memcg_slab_mutex);
 }
 
-struct create_work {
+struct memcg_register_cache_work {
        struct mem_cgroup *memcg;
        struct kmem_cache *cachep;
        struct work_struct work;
 };
 
-static void memcg_create_cache_work_func(struct work_struct *w)
+static void memcg_register_cache_func(struct work_struct *w)
 {
-       struct create_work *cw = container_of(w, struct create_work, work);
+       struct memcg_register_cache_work *cw =
+               container_of(w, struct memcg_register_cache_work, work);
        struct mem_cgroup *memcg = cw->memcg;
        struct kmem_cache *cachep = cw->cachep;
 
        mutex_lock(&memcg_slab_mutex);
-       memcg_kmem_create_cache(memcg, cachep);
+       memcg_register_cache(memcg, cachep);
        mutex_unlock(&memcg_slab_mutex);
 
        css_put(&memcg->css);
@@ -3289,12 +3290,12 @@ static void memcg_create_cache_work_func(struct work_struct *w)
 /*
  * Enqueue the creation of a per-memcg kmem_cache.
  */
-static void __memcg_create_cache_enqueue(struct mem_cgroup *memcg,
-                                        struct kmem_cache *cachep)
+static void __memcg_schedule_register_cache(struct mem_cgroup *memcg,
+                                           struct kmem_cache *cachep)
 {
-       struct create_work *cw;
+       struct memcg_register_cache_work *cw;
 
-       cw = kmalloc(sizeof(struct create_work), GFP_NOWAIT);
+       cw = kmalloc(sizeof(*cw), GFP_NOWAIT);
        if (cw == NULL) {
                css_put(&memcg->css);
                return;
@@ -3303,17 +3304,17 @@ static void __memcg_create_cache_enqueue(struct mem_cgroup *memcg,
        cw->memcg = memcg;
        cw->cachep = cachep;
 
-       INIT_WORK(&cw->work, memcg_create_cache_work_func);
+       INIT_WORK(&cw->work, memcg_register_cache_func);
        schedule_work(&cw->work);
 }
 
-static void memcg_create_cache_enqueue(struct mem_cgroup *memcg,
-                                      struct kmem_cache *cachep)
+static void memcg_schedule_register_cache(struct mem_cgroup *memcg,
+                                         struct kmem_cache *cachep)
 {
        /*
         * We need to stop accounting when we kmalloc, because if the
         * corresponding kmalloc cache is not yet created, the first allocation
-        * in __memcg_create_cache_enqueue will recurse.
+        * in __memcg_schedule_register_cache will recurse.
         *
         * However, it is better to enclose the whole function. Depending on
         * the debugging options enabled, INIT_WORK(), for instance, can
@@ -3322,7 +3323,7 @@ static void memcg_create_cache_enqueue(struct mem_cgroup *memcg,
         * the safest choice is to do it like this, wrapping the whole function.
         */
        memcg_stop_kmem_account();
-       __memcg_create_cache_enqueue(memcg, cachep);
+       __memcg_schedule_register_cache(memcg, cachep);
        memcg_resume_kmem_account();
 }
 
@@ -3393,16 +3394,11 @@ struct kmem_cache *__memcg_kmem_get_cache(struct kmem_cache *cachep,
         *
         * However, there are some clashes that can arrive from locking.
         * For instance, because we acquire the slab_mutex while doing
-        * kmem_cache_dup, this means no further allocation could happen
-        * with the slab_mutex held.
-        *
-        * Also, because cache creation issue get_online_cpus(), this
-        * creates a lock chain: memcg_slab_mutex -> cpu_hotplug_mutex,
-        * that ends up reversed during cpu hotplug. (cpuset allocates
-        * a bunch of GFP_KERNEL memory during cpuup). Due to all that,
-        * better to defer everything.
+        * memcg_create_kmem_cache, this means no further allocation
+        * could happen with the slab_mutex held. So it's better to
+        * defer everything.
         */
-       memcg_create_cache_enqueue(memcg, cachep);
+       memcg_schedule_register_cache(memcg, cachep);
        return cachep;
 out:
        rcu_read_unlock();
@@ -3526,7 +3522,7 @@ void __memcg_kmem_uncharge_pages(struct page *page, int order)
        memcg_uncharge_kmem(memcg, PAGE_SIZE << order);
 }
 #else
-static inline void mem_cgroup_destroy_all_caches(struct mem_cgroup *memcg)
+static inline void memcg_unregister_all_caches(struct mem_cgroup *memcg)
 {
 }
 #endif /* CONFIG_MEMCG_KMEM */
@@ -6372,7 +6368,7 @@ static void mem_cgroup_css_offline(struct cgroup_subsys_state *css)
        css_for_each_descendant_post(iter, css)
                mem_cgroup_reparent_charges(mem_cgroup_from_css(iter));
 
-       mem_cgroup_destroy_all_caches(memcg);
+       memcg_unregister_all_caches(memcg);
        vmpressure_cleanup(&memcg->vmpressure);
 }
 
index 3217561..48fafb6 100644 (file)
@@ -261,7 +261,7 @@ EXPORT_SYMBOL(kmem_cache_create);
 
 #ifdef CONFIG_MEMCG_KMEM
 /*
- * kmem_cache_create_memcg - Create a cache for a memory cgroup.
+ * memcg_create_kmem_cache - Create a cache for a memory cgroup.
  * @memcg: The memory cgroup the new cache is for.
  * @root_cache: The parent of the new cache.
  * @memcg_name: The name of the memory cgroup (used for naming the new cache).
@@ -270,7 +270,7 @@ EXPORT_SYMBOL(kmem_cache_create);
  * requests going from @memcg to @root_cache. The new cache inherits properties
  * from its parent.
  */
-struct kmem_cache *kmem_cache_create_memcg(struct mem_cgroup *memcg,
+struct kmem_cache *memcg_create_kmem_cache(struct mem_cgroup *memcg,
                                           struct kmem_cache *root_cache,
                                           const char *memcg_name)
 {
@@ -305,7 +305,7 @@ out_unlock:
        return s;
 }
 
-static int kmem_cache_destroy_memcg_children(struct kmem_cache *s)
+static int memcg_cleanup_cache_params(struct kmem_cache *s)
 {
        int rc;
 
@@ -314,13 +314,13 @@ static int kmem_cache_destroy_memcg_children(struct kmem_cache *s)
                return 0;
 
        mutex_unlock(&slab_mutex);
-       rc = __kmem_cache_destroy_memcg_children(s);
+       rc = __memcg_cleanup_cache_params(s);
        mutex_lock(&slab_mutex);
 
        return rc;
 }
 #else
-static int kmem_cache_destroy_memcg_children(struct kmem_cache *s)
+static int memcg_cleanup_cache_params(struct kmem_cache *s)
 {
        return 0;
 }
@@ -343,7 +343,7 @@ void kmem_cache_destroy(struct kmem_cache *s)
        if (s->refcount)
                goto out_unlock;
 
-       if (kmem_cache_destroy_memcg_children(s) != 0)
+       if (memcg_cleanup_cache_params(s) != 0)
                goto out_unlock;
 
        list_del(&s->list);