On Fri, Apr 30, 2021 at 04:32:39PM +0800, Muchun Song wrote:

Before start, we should know about the following rules of list lrus.

- Only objects allocated with __GFP_ACCOUNT need to allocate
  the struct list_lru_node.
- The caller of allocating memory must know which list_lru the
  object will insert.

So we can allocate struct list_lru_node when allocating the
object instead of allocating it when list_lru_add().  It is easy, because
we already know the list_lru and memcg which the object belongs
to. So we can introduce a new helper to allocate the object and
list_lru_node. Like below.

I feel like there may be a simpler solution, although I'm not really
familiar with the list_lru situation.  The three caches you mention:

I have looked at the code closely. There are 3 different kmem_caches that
need to use this new API to allocate memory. They are inode_cachep,
dentry_cache and radix_tree_node_cachep. I think that it is easy to migrate.

are all filesystem.  So if there's a way of knowing which filesystems
are exposed to each container, we can allocate the list_lru structures at
"mount" time rather than at first allocation for a given cache/lru/memcg
combination.