Re: [PATCH 2/3] vmscan: make mapped executable pages the firstclass citizen

From: Wu Fengguang
Date: Tue May 19 2009 - 21:45:21 EST

On Tue, May 19, 2009 at 08:28:28PM +0800, KOSAKI Motohiro wrote:
> Hi
>
> 2009/5/19 Wu Fengguang <fengguang.wu@xxxxxxxxx>:
> > On Tue, May 19, 2009 at 04:06:35PM +0800, KOSAKI Motohiro wrote:
> >> > > > Like the console mode, the absolute nr_mapped drops considerably - to 1/13 of
> >> > > > the original size - during the streaming IO.
> >> > > >
> >> > > > The delta of pgmajfault is 3 vs 107 during IO, or 236 vs 393 during the whole
> >> > > > process.
> >> > >
> >> > > hmmm.
> >> > >
> >> > > about 100 page fault don't match Elladan's problem, I think.
> >> > > perhaps We missed any addional reproduce condition?
> >> >
> >> > Elladan's case is not the point of this test.
> >> > Elladan's IO is use-once, so probably not a caching problem at all.
> >> >
> >> > This test case is specifically devised to confirm whether this patch
> >> > works as expected. Conclusion: it is.
> >>
> >> Dejection ;-)
> >>
> >> The number should address the patch is useful or not. confirming as expected
> >> is not so great.
> >
> > OK, let's make the conclusion in this way:
> >
> > The changelog analyzed the possible beneficial situation, and this
> > test backs that theory with real numbers, ie: it successfully stops
> > major faults when the active file list is slowly scanned when there
> > are partially cache hot streaming IO.
> >
> > Another (amazing) finding of the test is, only around 1/10 mapped pages
> > are actively referenced in the absence of user activities.
> >
> > Shall we protect the remaining 9/10 inactive ones? This is a question ;-)
>
> Unfortunately, I don't reproduce again.
> I don't apply your patch yet. but mapped ratio is reduced only very little.

mapped ratio or absolute numbers? The ratio wont change much because
nr_mapped is already small.

> I think smem can show which library evicted. Can you try it?
>
> download: http://www.selenic.com/smem/
> usage: ./smem -m -r --abbreviate

Sure, but I don't see much change in its output (see attachments).

smem-console-0 is collected after fresh boot,
smem-console-1 is collected after the big IO.

> We can't decide 9/10 is important or not. we need know actual evicted file list.

Right. But what I measured is the activeness. Almost zero major page
faults means the evicted 90% mapped pages are inactive during the
long 300 seconds of IO.

Thanks,
Fengguang

> > Or, shall we take the "protect active VM_EXEC mapped pages" approach,
> > or Christoph's "protect all mapped pages all time, unless they grow
> > too large" attitude? ÂI still prefer the best effort VM_EXEC heuristics.
> >
> > 1) the partially cache hot streaming IO is far more likely to happen
> > Â on (file) servers. For them, evicting the 9/10 inactive mapped
> > Â pages over night should be acceptable for sysadms.
> >
> > 2) for use-once IO on desktop, we have Rik's active file list
> > Â protection heuristics, so nothing to worry at all.
> >
> > 3) for big working set small memory desktop, the active list will
> > Â still be scanned, in this situation, why not evict some of the
> > Â inactive mapped pages? If they have not been accessed for 1 minute,
> > Â they are not likely be the user focus, and the tight memory
> > Â constraint can only afford to cache the user focused working set.
> >
> > Does that make sense?
Map PIDs AVGPSS PSS
[heap] 2 1.5M 2.9M
<anonymous> 2 538.0K 1.1M
/usr/bin/python2.5 1 1.0M 1.0M
/bin/zsh4 1 592.0K 592.0K
/usr/lib/zsh/4.3.9/zsh/zle.so 1 216.0K 216.0K
/lib/libncursesw.so.5.7 1 156.0K 156.0K
/lib/libc-2.9.so 2 65.0K 130.0K
[stack] 2 62.0K 124.0K
/usr/lib/zsh/4.3.9/zsh/complete.so 1 56.0K 56.0K
/usr/lib/locale/locale-archive 2 28.0K 56.0K
/usr/lib/python2.5/lib-dynload/operator. 1 32.0K 32.0K
/lib/libm-2.9.so 2 15.0K 30.0K
/usr/lib/zsh/4.3.9/zsh/complist.so 1 28.0K 28.0K
/lib/ld-2.9.so 2 13.0K 26.0K
/usr/lib/zsh/4.3.9/zsh/parameter.so 1 24.0K 24.0K
/usr/lib/python2.5/lib-dynload/_struct.s 1 24.0K 24.0K
/usr/lib/zsh/4.3.9/zsh/zutil.so 1 20.0K 20.0K
/usr/lib/python2.5/lib-dynload/time.so 1 20.0K 20.0K
/usr/lib/python2.5/lib-dynload/strop.so 1 20.0K 20.0K
/usr/lib/python2.5/lib-dynload/_locale.s 1 20.0K 20.0K
/lib/libpthread-2.9.so 1 18.0K 18.0K
/lib/libdl-2.9.so 2 9.0K 18.0K
/usr/lib/zsh/4.3.9/zsh/rlimits.so 1 16.0K 16.0K
/lib/libcap.so.2.11 1 16.0K 16.0K
/lib/libattr.so.1.1.0 1 16.0K 16.0K
/usr/lib/zsh/4.3.9/zsh/terminfo.so 1 12.0K 12.0K
/usr/lib/python2.5/lib-dynload/grp.so 1 12.0K 12.0K
/lib/libutil-2.9.so 1 11.0K 11.0K
/lib/libnss_nis-2.9.so 1 11.0K 11.0K
/lib/libnss_files-2.9.so 1 10.0K 10.0K
/lib/libnss_compat-2.9.so 1 10.0K 10.0K
/lib/libnsl-2.9.so 1 10.0K 10.0K
/usr/lib/gconv/gconv-modules.cache 1 5.0K 5.0K
[vsyscall] 2 0 0
[vdso] 2 0 0
Map PIDs AVGPSS PSS
[heap] 2 1.5M 2.9M
<anonymous> 2 538.0K 1.1M
/usr/bin/python2.5 1 1.0M 1.0M
/bin/zsh4 1 496.0K 496.0K
/lib/libc-2.9.so 2 180.0K 360.0K
[stack] 2 60.0K 120.0K
/lib/ld-2.9.so 2 58.0K 116.0K
/usr/lib/zsh/4.3.9/zsh/zle.so 1 112.0K 112.0K
/lib/libncursesw.so.5.7 1 56.0K 56.0K
/usr/lib/locale/locale-archive 2 26.0K 52.0K
/lib/libpthread-2.9.so 1 48.0K 48.0K
/lib/libm-2.9.so 2 21.0K 42.0K
/usr/lib/python2.5/lib-dynload/operator. 1 32.0K 32.0K
/usr/lib/zsh/4.3.9/zsh/complete.so 1 24.0K 24.0K
/usr/lib/python2.5/lib-dynload/_struct.s 1 24.0K 24.0K
/lib/libdl-2.9.so 2 12.0K 24.0K
/usr/lib/python2.5/lib-dynload/time.so 1 20.0K 20.0K
/usr/lib/python2.5/lib-dynload/strop.so 1 20.0K 20.0K
/usr/lib/python2.5/lib-dynload/_locale.s 1 20.0K 20.0K
/lib/libutil-2.9.so 1 16.0K 16.0K
/usr/lib/python2.5/lib-dynload/grp.so 1 12.0K 12.0K
/lib/libcap.so.2.11 1 12.0K 12.0K
/lib/libnss_compat-2.9.so 1 9.0K 9.0K
/usr/lib/zsh/4.3.9/zsh/zutil.so 1 8.0K 8.0K
/usr/lib/zsh/4.3.9/zsh/rlimits.so 1 8.0K 8.0K
/lib/libnss_nis-2.9.so 1 8.0K 8.0K
/lib/libnss_files-2.9.so 1 8.0K 8.0K
/lib/libnsl-2.9.so 1 8.0K 8.0K
/usr/lib/zsh/4.3.9/zsh/terminfo.so 1 4.0K 4.0K
/usr/lib/zsh/4.3.9/zsh/parameter.so 1 4.0K 4.0K
/usr/lib/zsh/4.3.9/zsh/complist.so 1 4.0K 4.0K
/lib/libattr.so.1.1.0 1 4.0K 4.0K
[vsyscall] 2 0 0
[vdso] 2 0 0
/usr/lib/gconv/gconv-modules.cache 1 0 0