Date: Wed, 24 Jun 2015 11:16:50 +0200
From: Michal Hocko <mhocko@suse.cz>
To: Mark Hills <mark@xwax.org>
Cc: Vlastimil Babka <vbabka@suse.cz>, linux-mm@kvack.org,
        Mel Gorman <mgorman@suse.de>, Johannes Weiner <hannes@cmpxchg.org>,
        LKML <linux-kernel@vger.kernel.org>
Subject: Re: Write throughput impaired by touching dirty_ratio
Message-ID: <20150624091650.GC32756@dhcp22.suse.cz>
References: <1506191513210.2879@stax.localdomain>
 <558A69F8.2080304@suse.cz>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <558A69F8.2080304@suse.cz>
User-Agent: Mutt/1.5.23 (2014-03-12)
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 967
Lines: 26

On Wed 24-06-15 10:27:36, Vlastimil Babka wrote:
> [add some CC's]
> 
> On 06/19/2015 05:16 PM, Mark Hills wrote:
[...]
> > The system is an HP xw6600, running i686 kernel. This happens whether 

How many CPUs does the machine have?

> > internal SATA HDD, SSD or external USB drive is used. I first saw this on 
> > kernel 4.0.4, and 4.0.5 is also affected.

OK so this is 32b kernel which might be the most important part. What is
the value of /proc/sys/vm/highmem_is_dirtyable? Also how does your low
mem vs higmem look when you are setting the ratio (cat /proc/zoneinfo)?

It seems Vlastimil is right and a bogus ratelimit_pages is calculated
and your writers are throttled every few pages.
-- 
Michal Hocko
SUSE Labs
--
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/