Re: [PATCH -mm 00/14] Per memcg slab shrinkers

[Vladimir Davydov <vdavydov@parallels.com>]

ping

On Sun, Sep 21, 2014 at 07:14:32PM +0400, Vladimir Davydov wrote:
> Hi,
> 
> Kmem accounting of memcg is unusable now, because it lacks slab shrinker
> support. That means when we hit the limit we will get ENOMEM w/o any
> chance to recover. What we should do then is to call shrink_slab, which
> would reclaim old inode/dentry caches from this cgroup. This is what
> this patch set is intended to do.
> 
> Basically, it does two things. First, it introduces the notion of
> per-memcg slab shrinker. A shrinker that wants to reclaim objects per
> cgroup should mark itself as SHRINKER_MEMCG_AWARE. Then it will be
> passed the memory cgroup to scan from in shrink_control->memcg. For such
> shrinkers shrink_slab iterates over the whole cgroup subtree under the
> target cgroup and calls the shrinker for each kmem-active memory cgroup.
> 
> Secondly, this patch set makes the list_lru structure per-memcg. It's
> done transparently to list_lru users - everything they have to do is to
> tell list_lru_init that they want memcg-aware list_lru. Then the
> list_lru will automatically distribute objects among per-memcg lists
> basing on which cgroup the object is accounted to. This way to make FS
> shrinkers (icache, dcache) memcg-aware we only need to make them use
> memcg-aware list_lru, and this is what this patch set does.
> 
> The main difference of this patch set from my previous attempts to push
> memcg aware shrinkers is in how it handles css offline. Now we don't let
> list_lrus corresponding to dead memory cgroups hang around till all
> objects are freed. Instead we move lru items to the parent cgroup's lru
> list. This is really important, because this allows us to release
> memcg_cache_id used for indexing in per-memcg arrays. If we don't do
> this, the arrays will grow uncontrollably, which is really bad. Note, in
> comparison to user memory reparenting, which Johannes is going to get
> rid of, it's not racy and much easier to implement although it does
> impose some limitations on how list_lru locking can be implemented.
> Another difference is that it doesn't reparent charges, only list_lru
> entries - the css will be dangling until the last kmem object is freed.
> 
> As before, this patch set only enables per-memcg kmem reclaim when the
> pressure goes from memory.limit, not from memory.kmem.limit. Handling
> memory.kmem.limit is going to be tricky due to GFP_NOFS allocations, it
> will probably require a sort of soft limit to work properly. I'm leaving
> this for future work.
> 
> The patch set basically consists of three main parts and organized as
> follows:
> 
>  - Patches 1-3 implement per-memcg shrinker core with patches 1 and 2
>    preparing list_lru users for upcoming changes and patch 3 tuning
>    shrink_slab.
> 
>  - Patches 4-10 make memcg core release cache ids on offline doing a bit
>    of cleanup in the meanwhile. This is easy, because kmem_caches don't
>    need the cache id after css offline since there can't be allocations
>    going from a dead memcg. Note that most of these patches (namely 4-6,
>    and 8) were once merged, but then I decided to drop them, because I
>    didn't know how to deal with list_lrus at that time (see
>    https://lkml.org/lkml/2014/7/23/218).
> 
>  - Finally patches 11-14 make list_lru per-memcg and mark FS shrinkers
>    as memcg-aware. This is the most difficult part of this patch set
>    with patch 13 (unlucky :-) doing the most important work.
> 
> Reviews are more than welcome.