[PATCH 0/4] memcg fix swap accounting (28/May)

[KAMEZAWA Hiroyuki]

Removed "RFC"

This patch series is restructured as following. some bugs are fixed.
Thank you for all your helps.

[1/4] ....change interface of swap_duplicate()/swap_free()
    Adds an function swapcache_prepare() and swapcache_free().

[2/4] ....add SWAP_HAS_CACHE flag and modify reference counting in swap_map
    Add SWAP_HAS_CACHE flag to swap_map array for knowing an information that
    "there is an only swap cache and swap has no reference" 
    without extra call of find_get_page().

[3/4] ....reclaim swap-cache-only swap_entry when get_swap_page() find it.
    Now, swap_map can tell "there is no reference other than cache", we
    can reclaim it if necessary.
    This code reclaim swap entries if
    - vm_swap_full()==ture
    && there is no free swap cluster
    && get_swap_page() finds unused swap entry.

[4/4].... fix memcg's swap accounting
    This is for fixing memcg's swap account leak. like this
==
            processA                   |           processB
  -------------------------------------+-------------------------------------
    (free_swap_and_cache())            |  (read_swap_cache_async())
                                       |    swap_duplicate()
                                       |    __set_page_locked()
                                       |    add_to_swap_cache()
      swap_entry_free() == 0           |
      find_get_page() -> found         |
      try_lock_page() -> fail & return |
                                       |    lru_cache_add_anon()
                                       |      doesn't link this page to memcg's
                                       |      LRU, because of !PageCgroupUsed.
==
    This patch tries to fix this by uncharging account when swap's refcnt goes
    to 0 even if there is an unused swap-cache.

    Works quite well in my test.

Thanks,
-Kame


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/