[prev in list] [next in list] [prev in thread] [next in thread] 

List:       linux-arm-kernel
Subject:    Re: [CFT] Mark IO as bufferable
From:       Russell King - ARM Linux <linux () arm ! linux ! org ! uk>
Date:       2005-06-29 9:43:20
Message-ID: 20050629104320.D3105 () flint ! arm ! linux ! org ! uk
[Download RAW message or body]

On Wed, Jun 29, 2005 at 10:29:10AM +0100, Richard Purdie wrote:
> On Tue, 2005-06-28 at 17:16 +0100, Russell King - ARM Linux wrote:
> > It'll use the write buffer.  What this means is that individual writes
> > should end up in the WB without stalling the CPU pipeline, and should
> > appear on the processor bus shortly thereafter.
> 
> > However, at present I'm just vaguely interested in what the effect is
> > on these older CPUs, and whether it looks like being a big issue.
> 
> The pxa255 system I tested on contains an ATI w100 graphics chip. You
> can't drive that chip with buffered IO due to the nature of the accesses
> you need to make. If I had to pick a reason for the failure of that
> patch, I'd blame the w100.
> 
> w100fb does map the chip's registers with ioremap_nocache which appears
> to be just the same as ioremap on arm. Perhaps this needs to be fixed?

The only things which will have changed are the static MT_DEVICE mappings.
ioremap and user-space non-memory mappings will still be completely
uncached/unbuffered.

-------------------------------------------------------------------
List admin: http://lists.arm.linux.org.uk/mailman/listinfo/linux-arm-kernel
FAQ:        http://www.arm.linux.org.uk/mailinglists/faq.php
Etiquette:  http://www.arm.linux.org.uk/mailinglists/etiquette.php
[prev in list] [next in list] [prev in thread] [next in thread]