linux-kernel - Re: [PATCH 8/8] vm: Add an tuning knob for vm.max_writeback

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20090908193427.GP22901@mit.edu>
Date:	Tue, 8 Sep 2009 15:34:27 -0400
From:	Theodore Tso <tytso@....edu>
To:	Christoph Hellwig <hch@...radead.org>
Cc:	Chris Mason <chris.mason@...cle.com>,
	Peter Zijlstra <peterz@...radead.org>,
	Artem Bityutskiy <dedekind1@...il.com>,
	Jens Axboe <jens.axboe@...cle.com>,
	linux-kernel@...r.kernel.org, linux-fsdevel@...r.kernel.org,
	david@...morbit.com, akpm@...ux-foundation.org, jack@...e.cz
Subject: Re: [PATCH 8/8] vm: Add an tuning knob for vm.max_writeback_mb

On Tue, Sep 08, 2009 at 02:19:37PM -0400, Christoph Hellwig wrote:
> > (This wasn't a proposed change; it's in Linux 2.6 mainline already;
> > see fs/xfs/linux-2.6/xfs_aops.c, in xfs_vm_writepage).  The fact that
> > filesystems are playing games like this should be a clear indication
> > that things are badly broken above....
> 
> Note that we did not put in this hack behind anyones back.  The first
> version from Chris was posted on fsdevel, lkml, the ext3 list and so on:
> 
> An when we finally decided that we absolute need it it also made another
> roundtrip to linux-mm to hope that we'd get something better from the VM
> people:

Sorry, I didn't want to imply that this was done behind anyone's back
--- rather that the fact that we need to do this sort of thing is an
indication that something is badly broken in the page writeback
functions.

I was reacting to Peter's argument that we shouldn't have a
knob/tunable to adjust this limit.  If we can figure out something
which is auto-tuning, that's all very well and good, but the fact that
people have been pointing a problem here for at least a full year,
maybe we should have the tunable now, and then later on, if the VM
crowd can figure out something more clever, we can retire the tunable
at some point in the future.

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