Hi Ivan, Jay,
On Sat, Dec 15, 2001 at 02:56:38PM +0300, Ivan Kokshaysky wrote:
> On Fri, Dec 14, 2001 at 02:25:00PM -0500, Jay Estabrook wrote:
> > The patch looks fine to me (does the prototype need changing?).
> Argh, thanks. Surely it does.
I reversed my own patches and tried yours.
It works. I can do IDE DMA with decent performance and without corruption on
writing to the disk on my DPWS 600a.
As your patch is easier, I think it's preferable. It's also more general, as
it does not only handle the IDE DMA thing, but also other devices on the
primary PCI bus that possibly will run into trouble doing Busmaster DMA
reads across a page boundary.
The only issue is that we limit those affected motherboards to 2GB memory.
Maybe we should output a warning?
(And I really don't know what happens if you have more mem and try to do BM
DMA to a non-accessible address. There are things like bounce buffers for
those highmem machines on iA32, but I don't think we have such kludges on
Please submit your patch to Marcelo & Linus.
-- Kurt Garloff <firstname.lastname@example.org> Eindhoven, NL GPG key: See mail header, key servers Linux kernel development SuSE GmbH, Nuernberg, DE SCSI, Security
This archive was generated by hypermail 2b29 : Sun Dec 23 2001 - 21:00:33 EST