From: Mel Gorman <mel-wPRd99KPJ+uzQB+pC5nmwQ@public•gmane.org>
To: David Rientjes <rientjes-hpIqsD4AKlfQT0dZR+AlfA@public•gmane.org>
Cc: Frans Pop <elendil-EIBgga6/0yRmR6Xm/wNWPw@public•gmane.org>,
Jiri Kosina <jkosina-AlSwsSmVLrQ@public•gmane.org>,
Sven Geggus
<lists-+AJD3D7QEjjt/htJsj1pd9AswbaBtrod@public•gmane.org>,
Karol Lewandowski
<karol.k.lewandowski-Re5JQEeQqe8AvxtiuMwx3w@public•gmane.org>,
Tobias Oetiker <tobi-7K0TWYW2a3pyDzI6CaY1VQ@public•gmane.org>,
"Rafael J. Wysocki" <rjw-KKrjLPT3xs0@public•gmane.org>,
David Miller <davem-fT/PcQaiUtIeIZ0/mPfg9Q@public•gmane.org>,
Reinette Chatre
<reinette.chatre-ral2JQCrhuEAvxtiuMwx3w@public•gmane.org>,
Kalle Valo <kalle.valo-X3B1VOXEql0@public•gmane.org>,
KOSAKI Motohiro
<kosaki.motohiro-+CUm20s59erQFUHtdCDX3A@public•gmane.org>,
Mohamed Abbas
<mohamed.abbas-ral2JQCrhuEAvxtiuMwx3w@public•gmane.org>,
Jens Axboe <jens.axboe-QHcLZuEGTsvQT0dZR+AlfA@public•gmane.org>,
"John W. Linville"
<linville-2XuSBdqkA4R54TAoqtyWWQ@public•gmane.org>,
Pekka Enberg <penberg-bbCR+/B0CizivPeTLB3BmA@public•gmane.org>,
Bartlomiej Zolnierkiewicz
<bzolnier-Re5JQEeQqe8AvxtiuMwx3w@public•gmane.org>,
Greg Kroah-Hartman <gregkh-l3A5Bk7waGM@public•gmane.org>,
Stephan von Krawczynski
<skraw-DcQCyzbjH0jQT0dZR+AlfA@public•gmane.org>,
Kernel Testers List
<kernel-testers-u79uwXL29TY76Z2rM5mHXA@public•gmane.org>,
netdev-u79uwXL29TY76Z2rM5mHXA@public•gmane.org,
linux-kernel-u79uwXL29TY76Z2rM5mHXA@public•gmane.org,
"linux-mm-Bw31MaZKKs3YtjvyW6yDsg@public•gmane.org"
<linux-mm-Bw31MaZKKs3YtjvyW6yDsg@public•gmane.org>
Subject: Re: [PATCH 4/5] page allocator: Pre-emptively wake kswapd when high-order watermarks are hit
Date: Fri, 23 Oct 2009 10:13:34 +0100 [thread overview]
Message-ID: <20091023091334.GV11778@csn.ul.ie> (raw)
In-Reply-To: <alpine.DEB.2.00.0910221227010.21601-X6Q0R45D7oAcqpCFd4KODRPsWskHk0ljAL8bYrjMMd8@public.gmane.org>
On Thu, Oct 22, 2009 at 12:41:42PM -0700, David Rientjes wrote:
> On Thu, 22 Oct 2009, Mel Gorman wrote:
>
> > diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> > index 7f2aa3e..851df40 100644
> > --- a/mm/page_alloc.c
> > +++ b/mm/page_alloc.c
> > @@ -1596,6 +1596,17 @@ try_next_zone:
> > return page;
> > }
> >
> > +static inline
> > +void wake_all_kswapd(unsigned int order, struct zonelist *zonelist,
> > + enum zone_type high_zoneidx)
> > +{
> > + struct zoneref *z;
> > + struct zone *zone;
> > +
> > + for_each_zone_zonelist(zone, z, zonelist, high_zoneidx)
> > + wakeup_kswapd(zone, order);
> > +}
> > +
> > static inline int
> > should_alloc_retry(gfp_t gfp_mask, unsigned int order,
> > unsigned long pages_reclaimed)
> > @@ -1730,18 +1741,18 @@ __alloc_pages_high_priority(gfp_t gfp_mask, unsigned int order,
> > congestion_wait(BLK_RW_ASYNC, HZ/50);
> > } while (!page && (gfp_mask & __GFP_NOFAIL));
> >
> > - return page;
> > -}
> > -
> > -static inline
> > -void wake_all_kswapd(unsigned int order, struct zonelist *zonelist,
> > - enum zone_type high_zoneidx)
> > -{
> > - struct zoneref *z;
> > - struct zone *zone;
> > + /*
> > + * If after a high-order allocation we are now below watermarks,
> > + * pre-emptively kick kswapd rather than having the next allocation
> > + * fail and have to wake up kswapd, potentially failing GFP_ATOMIC
> > + * allocations or entering direct reclaim
> > + */
> > + if (unlikely(order) && page && !zone_watermark_ok(preferred_zone, order,
> > + preferred_zone->watermark[ALLOC_WMARK_LOW],
> > + zone_idx(preferred_zone), ALLOC_WMARK_LOW))
> > + wake_all_kswapd(order, zonelist, high_zoneidx);
> >
> > - for_each_zone_zonelist(zone, z, zonelist, high_zoneidx)
> > - wakeup_kswapd(zone, order);
> > + return page;
> > }
> >
> > static inline int
>
> Hmm, is this really supposed to be added to __alloc_pages_high_priority()?
> By the patch description I was expecting kswapd to be woken up
> preemptively whenever the preferred zone is below ALLOC_WMARK_LOW and
> we're known to have just allocated at a higher order, not just when
> current was oom killed (when we should already be freeing a _lot_ of
> memory soon) or is doing a higher order allocation during direct reclaim.
>
It was a somewhat arbitrary choice to have it trigger in the event high
priority allocations were happening frequently.
> For the best coverage, it would have to be add the branch to the fastpath.
Agreed - specifically at the end of __alloc_pages_nodemask()
> That seems fine for a debugging aid and to see if progress is being made
> on the GFP_ATOMIC allocation issues, but doesn't seem like it should make
> its way to mainline, the subsequent GFP_ATOMIC allocation could already be
> happening and in the page allocator's slowpath at this point that this
> wakeup becomes unnecessary.
>
> If this is moved to the fastpath, why is this wake_all_kswapd() and not
> wakeup_kswapd(preferred_zone, order)? Do we need to kick kswapd in all
> zones even though they may be free just because preferred_zone is now
> below the watermark?
>
It probably makes no difference as zones are checked for their watermarks
before any real work happens. However, even if this patch makes a difference,
I don't want to see it merged. At best, it is an extremely heavy-handed
hack which is why I asked for it to be tested in isolation. It shouldn't
be necessary at all because sort of pre-emptive waking of kswapd was never
necessary before.
> Wouldn't it be better to do this on page_zone(page) instead of
> preferred_zone anyway?
>
No. The preferred_zone is the zone we should be allocating from. If we
failed to allocate from it, it implies the watermarks are not being met
so we want to wake it.
--
Mel Gorman
Part-time Phd Student Linux Technology Center
University of Limerick IBM Dublin Software Lab
next prev parent reply other threads:[~2009-10-23 9:13 UTC|newest]
Thread overview: 60+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-10-22 14:22 [PATCH 0/5] Candidate fix for increased number of GFP_ATOMIC failures V2 Mel Gorman
2009-10-22 14:22 ` [PATCH 3/5] vmscan: Force kswapd to take notice faster when high-order watermarks are being hit Mel Gorman
2009-10-27 2:42 ` KOSAKI Motohiro
2009-10-22 14:22 ` [PATCH 4/5] page allocator: Pre-emptively wake kswapd when high-order watermarks are hit Mel Gorman
[not found] ` <1256221356-26049-5-git-send-email-mel-wPRd99KPJ+uzQB+pC5nmwQ@public.gmane.org>
2009-10-22 19:41 ` David Rientjes
[not found] ` <alpine.DEB.2.00.0910221227010.21601-X6Q0R45D7oAcqpCFd4KODRPsWskHk0ljAL8bYrjMMd8@public.gmane.org>
2009-10-23 9:13 ` Mel Gorman [this message]
2009-10-23 9:36 ` David Rientjes
[not found] ` <alpine.DEB.2.00.0910230229010.28109-X6Q0R45D7oAcqpCFd4KODRPsWskHk0ljAL8bYrjMMd8@public.gmane.org>
2009-10-23 11:25 ` Mel Gorman
2009-10-23 11:31 ` Tobias Oetiker
[not found] ` <alpine.DEB.2.00.0910231329550.26462-EjsAmf5DE5zIvOfxy3zmAzgUDZmNtoG9@public.gmane.org>
2009-10-23 13:39 ` Mel Gorman
2009-10-27 2:42 ` KOSAKI Motohiro
[not found] ` <20091026235032.2F78.A69D9226-+CUm20s59erQFUHtdCDX3A@public.gmane.org>
2009-10-27 15:26 ` Mel Gorman
2009-10-22 14:22 ` [PATCH 5/5] ONLY-APPLY-IF-STILL-FAILING Revert 373c0a7e, 8aa7e847: Fix congestion_wait() sync/async vs read/write confusion Mel Gorman
2009-10-22 14:25 ` Against 2.6.31.4 [PATCH 5/5] " Mel Gorman
2009-10-22 21:49 ` [PATCH 5/5] ONLY-APPLY-IF-STILL-FAILING " Jens Axboe
2009-10-27 2:42 ` KOSAKI Motohiro
2009-10-27 10:29 ` Frans Pop
2009-10-22 14:47 ` [PATCH 0/5] Candidate fix for increased number of GFP_ATOMIC failures V2 Pekka Enberg
[not found] ` <84144f020910220747nba30d8bkc83c2569da79bd7c-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2009-10-22 16:03 ` Mel Gorman
2009-10-24 1:52 ` Christoph Lameter
2009-10-24 6:48 ` Pekka Enberg
2009-10-22 15:43 ` reinette chatre
2009-10-27 10:40 ` Mel Gorman
2009-10-27 23:34 ` reinette chatre
2009-10-23 7:31 ` Sven Geggus
2009-10-23 16:58 ` Karol Lewandowski
2009-10-23 21:12 ` Karol Lewandowski
2009-10-24 13:46 ` Mel LKML
[not found] ` <9ec2d7290910240646p75b93c68v6ea1648d628a9660-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2009-10-28 11:42 ` Karol Lewandowski
[not found] ` <20091028114208.GA14476-nLtalAL5mPp2RxbNQum0x1nzlInOXLuq@public.gmane.org>
2009-10-28 11:59 ` Mel Gorman
[not found] ` <20091028115926.GW8900-wPRd99KPJ+uzQB+pC5nmwQ@public.gmane.org>
2009-10-30 14:23 ` Karol Lewandowski
[not found] ` <20091030142350.GA9343-nLtalAL5mPp2RxbNQum0x1nzlInOXLuq@public.gmane.org>
2009-11-02 20:30 ` Mel Gorman
[not found] ` <20091102203034.GC22046-wPRd99KPJ+uzQB+pC5nmwQ@public.gmane.org>
2009-11-04 2:03 ` Karol Lewandowski
2009-10-28 12:55 ` Tobi Oetiker
2009-10-24 14:02 ` Sven Geggus
2009-10-27 13:27 ` Mel Gorman
[not found] ` <1256221356-26049-1-git-send-email-mel-wPRd99KPJ+uzQB+pC5nmwQ@public.gmane.org>
2009-10-22 14:22 ` [PATCH 1/5] page allocator: Always wake kswapd when restarting an allocation attempt after direct reclaim failed Mel Gorman
2009-10-22 14:24 ` [PATCH 1/5 Against 2.6.31.4] " Mel Gorman
2009-10-22 14:41 ` [PATCH 1/5] " Pekka Enberg
2009-10-22 15:49 ` Mel Gorman
[not found] ` <1256221356-26049-2-git-send-email-mel-wPRd99KPJ+uzQB+pC5nmwQ@public.gmane.org>
2009-10-26 1:11 ` KOSAKI Motohiro
2009-10-26 7:10 ` David Rientjes
2009-10-27 2:42 ` KOSAKI Motohiro
2009-10-27 12:27 ` Mel Gorman
2009-10-22 14:22 ` [PATCH 2/5] page allocator: Do not allow interrupts to use ALLOC_HARDER Mel Gorman
2009-10-22 16:33 ` Stephan von Krawczynski
[not found] ` <20091022183303.2448942d.skraw-DcQCyzbjH0jQT0dZR+AlfA@public.gmane.org>
2009-10-22 16:37 ` Mel Gorman
[not found] ` <20091022163752.GU11778-wPRd99KPJ+uzQB+pC5nmwQ@public.gmane.org>
2009-10-23 9:57 ` Stephan von Krawczynski
2009-10-24 2:03 ` Christoph Lameter
2009-10-27 15:19 ` Mel Gorman
2009-10-25 12:57 ` Stephan von Krawczynski
2009-10-26 1:15 ` KOSAKI Motohiro
2009-10-24 13:51 ` [PATCH 0/5] Candidate fix for increased number of GFP_ATOMIC failures V2 Frans Pop
2009-10-26 17:37 ` Tobias Oetiker
2009-10-27 15:36 ` Mel Gorman
2009-10-26 22:17 ` Frans Pop
2009-10-26 23:45 ` Frans Pop
2009-11-06 6:03 ` Tobias Diedrich
2009-11-06 9:24 ` Mel Gorman
[not found] ` <20091106092447.GC25926@csn.ul.ie>
2009-11-06 11:15 ` Tobias Diedrich
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20091023091334.GV11778@csn.ul.ie \
--to=mel-wprd99kpj+uzqb+pc5nmwq@public$(echo .)gmane.org \
--cc=bzolnier-Re5JQEeQqe8AvxtiuMwx3w@public$(echo .)gmane.org \
--cc=davem-fT/PcQaiUtIeIZ0/mPfg9Q@public$(echo .)gmane.org \
--cc=elendil-EIBgga6/0yRmR6Xm/wNWPw@public$(echo .)gmane.org \
--cc=gregkh-l3A5Bk7waGM@public$(echo .)gmane.org \
--cc=jens.axboe-QHcLZuEGTsvQT0dZR+AlfA@public$(echo .)gmane.org \
--cc=jkosina-AlSwsSmVLrQ@public$(echo .)gmane.org \
--cc=kalle.valo-X3B1VOXEql0@public$(echo .)gmane.org \
--cc=karol.k.lewandowski-Re5JQEeQqe8AvxtiuMwx3w@public$(echo .)gmane.org \
--cc=kernel-testers-u79uwXL29TY76Z2rM5mHXA@public$(echo .)gmane.org \
--cc=kosaki.motohiro-+CUm20s59erQFUHtdCDX3A@public$(echo .)gmane.org \
--cc=linux-kernel-u79uwXL29TY76Z2rM5mHXA@public$(echo .)gmane.org \
--cc=linux-mm-Bw31MaZKKs3YtjvyW6yDsg@public$(echo .)gmane.org \
--cc=linville-2XuSBdqkA4R54TAoqtyWWQ@public$(echo .)gmane.org \
--cc=lists-+AJD3D7QEjjt/htJsj1pd9AswbaBtrod@public$(echo .)gmane.org \
--cc=mohamed.abbas-ral2JQCrhuEAvxtiuMwx3w@public$(echo .)gmane.org \
--cc=netdev-u79uwXL29TY76Z2rM5mHXA@public$(echo .)gmane.org \
--cc=penberg-bbCR+/B0CizivPeTLB3BmA@public$(echo .)gmane.org \
--cc=reinette.chatre-ral2JQCrhuEAvxtiuMwx3w@public$(echo .)gmane.org \
--cc=rientjes-hpIqsD4AKlfQT0dZR+AlfA@public$(echo .)gmane.org \
--cc=rjw-KKrjLPT3xs0@public$(echo .)gmane.org \
--cc=skraw-DcQCyzbjH0jQT0dZR+AlfA@public$(echo .)gmane.org \
--cc=tobi-7K0TWYW2a3pyDzI6CaY1VQ@public$(echo .)gmane.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox