On Tue, Feb 9, 2021 at 12:33 PM Roman Gushchin [off-list ref] wrote:

On Tue, Feb 09, 2021 at 09:46:37AM -0800, Yang Shi wrote:

quoted

Since memcg_shrinker_map_size just can be changed under holding shrinker_rwsem
exclusively, the read side can be protected by holding read lock, so it sounds
superfluous to have a dedicated mutex.

Kirill Tkhai suggested use write lock since:

  * We want the assignment to shrinker_maps is visible for shrink_slab_memcg().
  * The rcu_dereference_protected() dereferrencing in shrink_slab_memcg(), but
    in case of we use READ lock in alloc_shrinker_maps(), the dereferrencing
    is not actually protected.
  * READ lock makes alloc_shrinker_info() racy against memory allocation fail.
    alloc_shrinker_info()->free_shrinker_info() may free memory right after
    shrink_slab_memcg() dereferenced it. You may say
    shrink_slab_memcg()->mem_cgroup_online() protects us from it? Yes, sure,
    but this is not the thing we want to remember in the future, since this
    spreads modularity.

And a test with heavy paging workload didn't show write lock makes things worse.

Acked-by: Vlastimil Babka <redacted>
Acked-by: Kirill Tkhai <redacted>
Signed-off-by: Yang Shi <redacted>

Acked-by: Roman Gushchin <redacted>

with a small nit (below):

quoted

---
 mm/vmscan.c | 16 ++++++----------
 1 file changed, 6 insertions(+), 10 deletions(-)

diff --git a/mm/vmscan.c b/mm/vmscan.c
index 96b08c79f18d..e4ddaaaeffe2 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c

@@ -187,7 +187,6 @@ static DECLARE_RWSEM(shrinker_rwsem);
 #ifdef CONFIG_MEMCG

 static int memcg_shrinker_map_size;
-static DEFINE_MUTEX(memcg_shrinker_map_mutex);

 static void free_shrinker_map_rcu(struct rcu_head *head)
 {

@@ -200,8 +199,6 @@ static int expand_one_shrinker_map(struct mem_cgroup *memcg,
      struct memcg_shrinker_map *new, *old;
      int nid;

-     lockdep_assert_held(&memcg_shrinker_map_mutex);
-

Why not check that shrinker_rwsem is down here?

No special reason, just because we know it was acquired before. We
could add the check, but not here. I think it'd be better to have the
assert in expand_shrinker_maps() since the rwsem was acquired before
calling it.

quoted

      for_each_node(nid) {
              old = rcu_dereference_protected(
                      mem_cgroup_nodeinfo(memcg, nid)->shrinker_map, true);

@@ -249,7 +246,7 @@ int alloc_shrinker_maps(struct mem_cgroup *memcg)
      if (mem_cgroup_is_root(memcg))
              return 0;

-     mutex_lock(&memcg_shrinker_map_mutex);
+     down_write(&shrinker_rwsem);
      size = memcg_shrinker_map_size;
      for_each_node(nid) {
              map = kvzalloc_node(sizeof(*map) + size, GFP_KERNEL, nid);

@@ -260,7 +257,7 @@ int alloc_shrinker_maps(struct mem_cgroup *memcg)
              }
              rcu_assign_pointer(memcg->nodeinfo[nid]->shrinker_map, map);
      }
-     mutex_unlock(&memcg_shrinker_map_mutex);
+     up_write(&shrinker_rwsem);

      return ret;
 }

@@ -275,9 +272,8 @@ static int expand_shrinker_maps(int new_id)
      if (size <= old_size)
              return 0;

-     mutex_lock(&memcg_shrinker_map_mutex);

And here as well. It will make the locking model more obvious and will help
to avoid errors in the future.

quoted

      if (!root_mem_cgroup)
-             goto unlock;
+             goto out;

      memcg = mem_cgroup_iter(NULL, NULL, NULL);
      do {

@@ -286,13 +282,13 @@ static int int new_id)
              ret = expand_one_shrinker_map(memcg, size, old_size);
              if (ret) {
                      mem_cgroup_iter_break(NULL, memcg);
-                     goto unlock;
+                     goto out;
              }
      } while ((memcg = mem_cgroup_iter(NULL, memcg, NULL)) != NULL);
-unlock:
+out:
      if (!ret)
              memcg_shrinker_map_size = size;
-     mutex_unlock(&memcg_shrinker_map_mutex);
+
      return ret;
 }

--

2.26.2