xfs
[Top] [All Lists]

Re: IDE write cache and journaling file systems

To: Steve Lord <lord@xxxxxxx>
Subject: Re: IDE write cache and journaling file systems
From: Jens Axboe <axboe@xxxxxxx>
Date: Thu, 18 Apr 2002 18:09:18 +0200
Cc: Andi Kleen <ak@xxxxxxx>, Jim Buzbee <James.Buzbee@xxxxxxxxxxxx>, XFS List <linux-xfs@xxxxxxxxxxx>
In-reply-to: <1019145843.10294.10.camel@xxxxxxxxxxxxxxxxxxxx>
References: <3CBEE618.B220A393@xxxxxxxxxxxx> <1019144692.10200.8.camel@xxxxxxxxxxxxxxxxxxxx> <20020418175314.A21976@xxxxxxxxxxxxx> <20020418160344.GM2492@xxxxxxx> <1019145843.10294.10.camel@xxxxxxxxxxxxxxxxxxxx>
Sender: owner-linux-xfs@xxxxxxxxxxx
On Thu, Apr 18 2002, Steve Lord wrote:
> On Thu, 2002-04-18 at 11:03, Jens Axboe wrote:
> > On Thu, Apr 18 2002, Andi Kleen wrote:
> > > > 
> > > > Andi Kleen was experimenting with the ide cache flushing code in the
> > > > Suse kernel and adding some flushing calls to XFS. We talked about
> > > > the right place to add them, I am not sure if he has tried it yet.
> > > 
> > > I've tried it and also got it to work in an experimental state, but 
> > > decided
> > > to rewrite it to use barriers instead. I didn't yet get around to do this 
> > > rewrite. The reason for the rewrite is that just doing the flush slows
> > > it down a lot.
> > 
> > Using barriers is surely the right approach, and lets the kernel use
> > flushes or tag barriers as provided by the hardware.
> > 
> > > It requires considerable infrastructure not in the standard kernel.
> > 
> > ? Both the SuSE kernel has the infrastructure, and the 2.5 kernels so as
> > well.
> 
> Unless I am mistaken, Jim is tied down to a fairly old kernel.

Not too much of a problem, the barrier infrastructure + ide flush
support really isn't that much code, didn't even take long to write.
That just leaves the XFS parts, which I don't know anything about.

-- 
Jens Axboe


<Prev in Thread] Current Thread [Next in Thread>