Re: [PATCH] memcg: fix stale swap cache leak v5

[Balbir Singh]

* Daisuke Nishimura <d-nishimura@xxxxxxxxxxxxxxxxx> [2009-05-01 13:33:17]:

> On Thu, 30 Apr 2009 23:42:46 +0530
> Balbir Singh <balbir@xxxxxxxxxxxxxxxxxx> wrote:
> 
> > * KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx> [2009-04-30 18:47:38]:
> > 
> > > On Thu, 30 Apr 2009 15:12:52 +0530
> > > Balbir Singh <balbir@xxxxxxxxxxxxxxxxxx> wrote:
> > > 
> > > > * KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx> [2009-04-30 18:04:26]:
> > > > 
> > > > > On Thu, 30 Apr 2009 16:35:39 +0900
> > > > > KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx> wrote:
> > > > > 
> > > > > > On Thu, 30 Apr 2009 16:16:27 +0900
> > > > > > KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx> wrote:
> > > > > > 
> > > > > > > This is v5 but all codes are rewritten.
> > > > > > > 
> > > > > > > After this patch, when memcg is used,
> > > > > > >  1. page's swapcount is checked after I/O (without locks). If the page is
> > > > > > >     stale swap cache, freeing routine will be scheduled.
> > > > > > >  2. vmscan.c calls try_to_free_swap() when __remove_mapping() fails.
> > > > > > > 
> > > > > > > Works well for me. no extra resources and no races.
> > > > > > > 
> > > > > > > Because my office will be closed until May/7, I'll not be able to make a
> > > > > > > response. Posting this for showing what I think of now.
> > > > > > > 
> > > > > > I found a hole immediately after posted this...sorry. plz ignore this patch/
> > > > > > see you again in the next month.
> > > > > > 
> > > > > I'm now wondering to disable "swapin readahed" completely when memcg is used...
> > > > > Then, half of the problems will go away immediately.
> > > > > And it's not so bad to try to free swapcache if swap writeback ends. Then, another
> > > > > half will go away...
> > > > >
> > > > 
> > > > Could you clarify? Will memcg not account for swapin readahead pages?
> > > >  
> > > swapin-readahead pages are _not_ accounted now. (And I think _never_)
> > > But has race and leak swp_entry account until global LRU runs.
> > > 
> > > "Don't do swapin-readahead, at all" will remove following race completely.
> > > ==
> > >          CPU0                  CPU1
> > >  free_swap_and_cache()
> > >                         read_swapcache_async()
> > > ==
> > > swp_entry to be freed will not be read-in.
> > > 
> > > I think there will no performance regression in _usual_ case even if no readahead.
> > > But has no number yet.
> > >
> > 
> > Kamezawa, Daisuke,
> > 
> > Can't we just correct the accounting and leave the page on the global
> > LRU?
> > 
> > Daisuke in the race conditions mentioned is (2) significant? Since the
> > accounting is already fixed during mem_cgroup_uncharge_page()?
> > 
> Do you mean type-2 stale swap caches I described before ?
> 
> They doesn't pressure mem.usage nor memsw.usage as you say,
> but consumes swp_entry(of cource, type-1 has this problem too).
> As a result, all the swap space can be used up and causes OOM.
> 

Good point..

> I've verified it long ago by:
> 
> - make swap space small(50MB).
> - set mem.limit(32MB).
> - run some programs(allocate, touch sometimes, exit) enough to
>   exceed mem.limit repeatedly(I used page01 included in ltp and run
>   5 instances 8MB per each in cpuset with 4cpus.).
> - wait for a very long time :) (2,30 hours IIRC)
>   You can see the usage of swap cache(grep SwapCached /proc/meminfo)
>   increasing gradually.
> 
> 
> BTW, I'm now testing a attached patch to fix type-2 with setting page-cluster
> to 0 to aboid type-1, and seeing what happens in the usage of swap cache.
> (I can't test it in large box though, because my office is closed till May 06.)
> 
> Thanks,
> Daisuke Nishimura.
> ===
> From: Daisuke Nishimura <nishimura@xxxxxxxxxxxxxxxxx>
> 
> memcg: free unused swapcache on swapout path
> 
> memcg cannot handle !PageCgroupUsed swapcache the owner process of which
> has been exited.
> 
> This patch is for handling such swap caches created by a race like below:
> 
>     Assume processA is exiting and pte points to a page(!PageSwapCache).
>     And processB is trying reclaim the page.
> 
>               processA                   |           processB
>     -------------------------------------+-------------------------------------
>       (page_remove_rmap())               |  (shrink_page_list())
>          mem_cgroup_uncharge_page()      |
>             ->uncharged because it's not |
>               PageSwapCache yet.         |
>               So, both mem/memsw.usage   |
>               are decremented.           |
>                                          |    add_to_swap() -> added to swap cache.
> 
>     If this page goes thorough without being freed for some reason, this page
>     doesn't goes back to memcg's LRU because of !PageCgroupUsed.
> 
> These swap cache cannot be freed in memcg's LRU scanning, and swp_entry cannot
> be freed properly as a result.
> This patch adds a hook after add_to_swap() to check the page is mapped by a
> process or not, and frees it if it has been unmapped already.
> 
> If a page has been on swap cache already when the owner process calls
> page_remove_rmap() -> mem_cgroup_uncharge_page(), the page is not uncharged.
> It goes back to memcg's LRU even if it goes through shrink_page_list()
> without being freed, so this patch ignores these case.
> 
> Signed-off-by: Daisuke Nishimura <nishimura@xxxxxxxxxxxxxxxxx>
> ---
>  include/linux/swap.h |   12 ++++++++++++
>  mm/memcontrol.c      |   14 ++++++++++++++
>  mm/vmscan.c          |    8 ++++++++
>  3 files changed, 34 insertions(+), 0 deletions(-)
> 
> diff --git a/include/linux/swap.h b/include/linux/swap.h
> index caf0767..8e75d7a 100644
> --- a/include/linux/swap.h
> +++ b/include/linux/swap.h
> @@ -336,11 +336,17 @@ static inline void disable_swap_token(void)
> 
>  #ifdef CONFIG_CGROUP_MEM_RES_CTLR
>  extern void mem_cgroup_uncharge_swapcache(struct page *page, swp_entry_t ent);
> +extern int memcg_free_unused_swapcache(struct page *page);
>  #else
>  static inline void
>  mem_cgroup_uncharge_swapcache(struct page *page, swp_entry_t ent)
>  {
>  }
> +static inline int
> +memcg_free_unused_swapcache(struct page *page)
> +{
> +	return 0;
> +}
>  #endif
>  #ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP
>  extern void mem_cgroup_uncharge_swap(swp_entry_t ent);
> @@ -431,6 +437,12 @@ static inline swp_entry_t get_swap_page(void)
>  #define has_swap_token(x) 0
>  #define disable_swap_token() do { } while(0)
> 
> +static inline int
> +memcg_free_unused_swapcache(struct page *page)
> +{
> +	return 0;
> +}
> +
>  #endif /* CONFIG_SWAP */
>  #endif /* __KERNEL__*/
>  #endif /* _LINUX_SWAP_H */
> diff --git a/mm/memcontrol.c b/mm/memcontrol.c
> index 01c2d8f..4f7e5b6 100644
> --- a/mm/memcontrol.c
> +++ b/mm/memcontrol.c
> @@ -1488,6 +1488,7 @@ void mem_cgroup_uncharge_cache_page(struct page *page)
>  	__mem_cgroup_uncharge_common(page, MEM_CGROUP_CHARGE_TYPE_CACHE);
>  }
> 
> +#ifdef CONFIG_SWAP
>  /*
>   * called from __delete_from_swap_cache() and drop "page" account.
>   * memcg information is recorded to swap_cgroup of "ent"
> @@ -1507,6 +1508,19 @@ void mem_cgroup_uncharge_swapcache(struct page *page, swp_entry_t ent)
>  		css_put(&memcg->css);
>  }
> 
> +int memcg_free_unused_swapcache(struct page *page)
> +{
> +	VM_BUG_ON(!PageLocked(page));
> +	VM_BUG_ON(!PageSwapCache(page));
> +
> +	if (mem_cgroup_disabled())
> +		return 0;
> +	if (!PageAnon(page) || page_mapped(page))
> +		return 0;
> +	return try_to_free_swap(page);	/* checks page_swapcount */
> +}
> +#endif /* CONFIG_SWAP */
> +
>  #ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP
>  /*
>   * called from swap_entry_free(). remove record in swap_cgroup and
> diff --git a/mm/vmscan.c b/mm/vmscan.c
> index eac9577..c1a7a6f 100644
> --- a/mm/vmscan.c
> +++ b/mm/vmscan.c
> @@ -656,6 +656,14 @@ static unsigned long shrink_page_list(struct list_head *page_list,
>  				goto keep_locked;
>  			if (!add_to_swap(page))
>  				goto activate_locked;
> +			/*
> +			 * The owner process might have uncharged the page
> +			 * (by page_remove_rmap()) before it has been added
> +			 * to swap cache.
> +			 * Check it here to avoid making it stale.
> +			 */
> +			if (memcg_free_unused_swapcache(page))
> +				goto keep_locked;
>  			may_enter_fs = 1;
>  		}
> 
> 
Looking through the patch, I have my doubts

 shrink_page_list() will catch the page - how? It is not on memcg's
LRU, so if we have a small cgroup with lots of memory, when the cgroup
is running out of memory, will this page show up in
shrink_page_list()?

-- 
	Balbir
--
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/