Re: [PATCH v8 6/6] zswap: shrinks zswap pool based on memory pressure

[Chengming Zhou <chengming.zhou@linux.dev>]

On 2023/12/6 13:59, Yosry Ahmed wrote:
> [..]
>>> @@ -526,6 +582,102 @@ static struct zswap_entry *zswap_entry_find_get(struct rb_root *root,
>>>       return entry;
>>>  }
>>>
>>> +/*********************************
>>> +* shrinker functions
>>> +**********************************/
>>> +static enum lru_status shrink_memcg_cb(struct list_head *item, struct list_lru_one *l,
>>> +                                    spinlock_t *lock, void *arg);
>>> +
>>> +static unsigned long zswap_shrinker_scan(struct shrinker *shrinker,
>>> +             struct shrink_control *sc)
>>> +{
>>> +     struct lruvec *lruvec = mem_cgroup_lruvec(sc->memcg, NODE_DATA(sc->nid));
>>> +     unsigned long shrink_ret, nr_protected, lru_size;
>>> +     struct zswap_pool *pool = shrinker->private_data;
>>> +     bool encountered_page_in_swapcache = false;
>>> +
>>> +     nr_protected =
>>> +             atomic_long_read(&lruvec->zswap_lruvec_state.nr_zswap_protected);
>>> +     lru_size = list_lru_shrink_count(&pool->list_lru, sc);
>>> +
>>> +     /*
>>> +      * Abort if the shrinker is disabled or if we are shrinking into the
>>> +      * protected region.
>>> +      *
>>> +      * This short-circuiting is necessary because if we have too many multiple
>>> +      * concurrent reclaimers getting the freeable zswap object counts at the
>>> +      * same time (before any of them made reasonable progress), the total
>>> +      * number of reclaimed objects might be more than the number of unprotected
>>> +      * objects (i.e the reclaimers will reclaim into the protected area of the
>>> +      * zswap LRU).
>>> +      */
>>> +     if (!zswap_shrinker_enabled || nr_protected >= lru_size - sc->nr_to_scan) {
>>> +             sc->nr_scanned = 0;
>>> +             return SHRINK_STOP;
>>> +     }
>>> +
>>> +     shrink_ret = list_lru_shrink_walk(&pool->list_lru, sc, &shrink_memcg_cb,
>>> +             &encountered_page_in_swapcache);
>>> +
>>> +     if (encountered_page_in_swapcache)
>>> +             return SHRINK_STOP;
>>> +
>>> +     return shrink_ret ? shrink_ret : SHRINK_STOP;
>>> +}
>>> +
>>> +static unsigned long zswap_shrinker_count(struct shrinker *shrinker,
>>> +             struct shrink_control *sc)
>>> +{
>>> +     struct zswap_pool *pool = shrinker->private_data;
>>> +     struct mem_cgroup *memcg = sc->memcg;
>>> +     struct lruvec *lruvec = mem_cgroup_lruvec(memcg, NODE_DATA(sc->nid));
>>> +     unsigned long nr_backing, nr_stored, nr_freeable, nr_protected;
>>> +
>>> +#ifdef CONFIG_MEMCG_KMEM
>>> +     cgroup_rstat_flush(memcg->css.cgroup);
>>> +     nr_backing = memcg_page_state(memcg, MEMCG_ZSWAP_B) >> PAGE_SHIFT;
>>> +     nr_stored = memcg_page_state(memcg, MEMCG_ZSWAPPED);
>>> +#else
>>> +     /* use pool stats instead of memcg stats */
>>> +     nr_backing = get_zswap_pool_size(pool) >> PAGE_SHIFT;
>>> +     nr_stored = atomic_read(&pool->nr_stored);
>>> +#endif
>>> +
>>> +     if (!zswap_shrinker_enabled || !nr_stored)
>> When I tested with this series, with !zswap_shrinker_enabled in the default case,
>> I found the performance is much worse than that without this patch.
>>
>> Testcase: memory.max=2G, zswap enabled, kernel build -j32 in a tmpfs directory.
>>
>> The reason seems the above cgroup_rstat_flush(), caused much rstat lock contention
>> to the zswap_store() path. And if I put the "zswap_shrinker_enabled" check above
>> the cgroup_rstat_flush(), the performance become much better.
>>
>> Maybe we can put the "zswap_shrinker_enabled" check above cgroup_rstat_flush()?
> 
> Yes, we should do nothing if !zswap_shrinker_enabled. We should also
> use mem_cgroup_flush_stats() here like other places unless accuracy is
> crucial, which I doubt given that reclaim uses
> mem_cgroup_flush_stats().
> 

Yes. After changing to use mem_cgroup_flush_stats() here, the performance
become much better.

> mem_cgroup_flush_stats() has some thresholding to make sure we don't
> do flushes unnecessarily, and I have a pending series in mm-unstable
> that makes that thresholding per-memcg. Keep in mind that adding a
> call to mem_cgroup_flush_stats() will cause a conflict in mm-unstable,

My test branch is linux-next 20231205, and it's all good after changing
to use mem_cgroup_flush_stats(memcg).

> because the series there adds a memcg argument to
> mem_cgroup_flush_stats(). That should be easily amenable though, I can
> post a fixlet for my series to add the memcg argument there on top of
> users if needed.
> 

It's great. Thanks!