On Fri, May 04, 2001 at 12:11:56PM +0200, Ingolf Steinbach wrote:
> Manuel Bouyer wrote:
> > Yes, see the wd(4) man page.
> 
> Thanks a lot. I'll try that. Can the value 0 be used for each nibble of the
> flags independently from the other nibbles? The man pages only mentions
> the 0x0000 case (i.e. all nibbles =0 at the same time).

Yes, they can be set to 0 independantly

> 
> > But this won't help you in this case.
> > The kernel isn't aware that your motherboard can do UDMA5 at the moment
> > and will use UDMA2 (that's part of the problem)
> 
> Hm. I don't understand that. I thought that UDMA5 makes the problems and
> that I just have to use some slower UDMA mode to workaround the problems.
> Does the kernel have to be aware of the motherboard's UDMA5-capability in
> this case? (I want the kernel to be forced to not use UDMA5!)

I didn't look at this controller's datasheet yet, but I suspect there's
a bit to reset to 0 somewhere to make it use lower mode - a bit which
is "reserved" in older VIA controllers so the driver doesn't touch it

> 
> > however in the dmesg I notice the kernel doesn't use Ultra-DMA at all.
> > Did you put flags to disable Ultra-DMA in your kernel config file ?
> 
> See the "Note that I have turned off UDMA at the moment (see the
> "wd0(pciide0:0:0):" line)" above the dmesg output in my original post.

Ha, ok.

Well, I'm working on support for newer VIA controllers. Stay tunned,
and watch the current-users mailing list :)

--
Manuel Bouyer, LIP6, Universite Paris VI.           Manuel.Bouyer@lip6.fr
--