Re: [PATCH memcg RFC] net: set proper memcg for net_init hooks allocations

[Vasily Averin <vvs@xxxxxxxxxx>]

On 4/21/22 18:56, Shakeel Butt wrote:
> On Sat, Apr 16, 2022 at 11:39 PM Vasily Averin <vvs@xxxxxxxxxx> wrote:
>> @@ -1147,7 +1148,13 @@ static int __register_pernet_operations(struct list_head *list,
>>                  * setup_net() and cleanup_net() are not possible.
>>                  */
>>                 for_each_net(net) {
>> +                       struct mem_cgroup *old, *memcg = NULL;
>> +#ifdef CONFIG_MEMCG
>> +                       memcg = (net == &init_net) ? root_mem_cgroup : mem_cgroup_from_obj(net);
> 
> memcg from obj is unstable, so you need a reference on memcg. You can
> introduce get_mem_cgroup_from_kmem() which works for both
> MEMCG_DATA_OBJCGS and MEMCG_DATA_KMEM. For uncharged objects (like
> init_net) it should return NULL.

Could you please elaborate with more details?
It seems to me mem_cgroup_from_obj() does everything exactly as you say:
- for slab objects it returns memcg taken from according slab->memcg_data
- for ex-slab objects (i.e. page->memcg_data & MEMCG_DATA_OBJCGS)
    page_memcg_check() returns NULL
- for kmem objects (i.e. page->memcg_data & MEMCG_DATA_KMEM) 
    page_memcg_check() returns objcg->memcg
- in another cases
    page_memcg_check() returns page->memcg_data,
    so for uncharged objects like init_net NULL should be returned.

I can introduce exported get_mem_cgroup_from_kmem(), however it should only
call mem_cgroup_from_obj(), perhaps under read_rcu_lock/unlock.

Do you mean something like this?

--- a/include/linux/memcontrol.h
+++ b/include/linux/memcontrol.h
@@ -1768,4 +1768,14 @@ static inline struct mem_cgroup *mem_cgroup_from_obj(void *p)
 
 #endif /* CONFIG_MEMCG_KMEM */
 
+static inline struct mem_cgroup *get_mem_cgroup_from_kmem(void *p)
+{
+	struct mem_cgroup *memcg;
+
+	rcu_read_lock();
+	memcg = mem_cgroup_from_obj(p);
+	rcu_read_unlock();
+
+	return memcg;
+}
 #endif /* _LINUX_MEMCONTROL_H */
diff --git a/net/core/net_namespace.c b/net/core/net_namespace.c
index a5b5bb99c644..4003c47965c9 100644
--- a/net/core/net_namespace.c
+++ b/net/core/net_namespace.c
@@ -26,6 +26,7 @@
 #include <net/net_namespace.h>
 #include <net/netns/generic.h>
 
+#include <linux/sched/mm.h>
 /*
  *	Our network namespace constructor/destructor lists
  */
@@ -1147,7 +1148,14 @@ static int __register_pernet_operations(struct list_head *list,
 		 * setup_net() and cleanup_net() are not possible.
 		 */
 		for_each_net(net) {
+			struct mem_cgroup *old, *memcg;
+
+			memcg = get_mem_cgroup_from_kmem(net);
+			if (memcg == NULL)
+				memcg = root_mem_cgroup;
+			old = set_active_memcg(memcg);
 			error = ops_init(ops, net);
+			set_active_memcg(old);
 			if (error)
 				goto out_undo;
 			list_add_tail(&net->exit_list, &net_exit_list);