lkml.org 
[lkml]   [2009]   [Apr]   [27]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH] fix leak of swap accounting as stale swap cache under memcg
On Tue, 28 Apr 2009 10:09:30 +0900
nishimura@mxp.nes.nec.co.jp wrote:

> > On Mon, 27 Apr 2009 21:08:56 +0900
> > Daisuke Nishimura <d-nishimura@mtf.biglobe.ne.jp> wrote:
> >
> >> > Index: mmotm-2.6.30-Apr24/mm/vmscan.c
> >> > ===================================================================
> >> > --- mmotm-2.6.30-Apr24.orig/mm/vmscan.c
> >> > +++ mmotm-2.6.30-Apr24/mm/vmscan.c
> >> > @@ -661,6 +661,9 @@ static unsigned long shrink_page_list(st
> >> > if (PageAnon(page) && !PageSwapCache(page)) {
> >> > if (!(sc->gfp_mask & __GFP_IO))
> >> > goto keep_locked;
> >> > + /* avoid making more stale swap caches */
> >> > + if (memcg_stale_swap_congestion())
> >> > + goto keep_locked;
> >> > if (!add_to_swap(page))
> >> > goto activate_locked;
> >> > may_enter_fs = 1;
> >> >
> >> Well, as I mentioned before(http://marc.info/?l=linux-kernel&m=124066623510867&w=2),
> >> this cannot avoid type-2(set !PageCgroupUsed by the owner process via
> >> page_remove_rmap()->mem_cgroup_uncharge_page() before being added to swap cache).
> >> If these swap caches go through shrink_page_list() without beeing freed
> >> for some reason, these swap caches doesn't go back to memcg's LRU.
> >>
> >> Type-2 doesn't pressure memsw.usage, but you can see it by plotting
> >> "grep SwapCached /proc/meminfo".
> >>
> >> And I don't think it's a good idea to add memcg_stale_swap_congestion() here.
> >> This means less possibility to reclaim pages.
> >>
> > Hmm. maybe adding congestion_wait() ?
> >
> I don't think no hook before add_to_swap() is needed.
>
> >> Do you dislike the patch I attached in the above mail ?
> >>
> > I doubt whether the patch covers all type-2 case.
> >
> hmm, I didn't see any leak anymore when I tested the patch.
>

At first, your patch
==
if (PageAnon(page) && !PageSwapCache(page)) {
if (!(sc->gfp_mask & __GFP_IO))
goto keep_locked;
- /* avoid making more stale swap caches */
- if (memcg_stale_swap_congestion())
- 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;
}
==
Should be
==

if (PageAnon(page) && !PageSwapCache(page)) {
... // don't modify here
}
if (PageAnon(page) && PageSwapCache(page) && !page_mapped(page)) {
if (try_to_free_page(page)) // or memcg_free_unused_swapcache()
goto free_it;
}
==
I think.

And we need hook to free_swap_and_cache() for handling PageWriteback() case.


> But because of machine time limit, I could only test for about 3 hours.
> (I had seen some leak at that point before applying my patch)
> I'll test for longer time if possible.
>
Sigh, my work time is also limited for these months ;(

> > I'll consider again and post v3.
> > But I'll go into a series of holidays, so, may not come back until May/6.
> >
> It's the same for me :)
>
Enjoy good holidays :)

Thanks,
-Kame






\
 
 \ /
  Last update: 2009-04-28 03:23    [W:0.176 / U:1.112 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site