From: Nikita Danilov <Nikita@Namesys.COM>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Message-ID: <16150.43444.734717.485697@laputa.namesys.com>
Date: Thu, 17 Jul 2003 17:50:44 +0400
To: Andrew Morton <akpm@osdl.org>
Cc: Linux Kernel Mailing List <Linux-Kernel@Vger.Kernel.ORG>
Subject: Re: [PATCH] 1/5 VM changes: zone-pressure.patch
In-Reply-To: <20030709034251.6902c488.akpm@osdl.org>
References: <16139.54887.932511.717315@laputa.namesys.com>
	<20030709031203.3971d9b4.akpm@osdl.org>
	<16139.60502.110693.175421@laputa.namesys.com>
	<20030709034251.6902c488.akpm@osdl.org>
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1182
Lines: 34

Andrew Morton writes:
 > Nikita Danilov <Nikita@Namesys.COM> wrote:
 > >
 > >  > OK, fixes a bug.
 > > 
 > >  What bug?
 > 
 > Failing to consider mapped pages on the active list until the scanning
 > priority gets large.
 > 
 > I ran up your five patches on a 256MB box, running `qsbench -m 350'.  It got
 > all slow then the machine seized up.   I'll poke at it some.
 > 

My understanding of this is following:

1. kswapd/balance_pgdat runs through all priorities, but fails to rise
zone->free_pages above zone->pages_min. So it exet. Which is strange.

2. nobody is going to wake it up, because all memory allocators are
looping indefinitely in that newly introduced do_retry loop inside
__alloc_pages.

I think either wakeup_kswapd() should be called in __alloc_pages()
before goto rebalance, or kswapd should loop until it restores all zones
to the balance (actually, comment before balance_pgdat() says it does),
or both.

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