[coreboot] Coreboot doesn't boot on Arima HDAMA rev.G mainboard

[Ward Vandewege]

On Tue, May 04, 2010 at 11:59:26PM +0200, Joop Boonen wrote:
> On Tue, May 4, 2010 8:16 pm, Myles Watson wrote:
> > On Tue, May 4, 2010 at 11:54 AM, Myles Watson <mylesgw at gmail.com> wrote:
> >>
> >>> rev 4920 works without a problem, also with the extracted VGA. I get
> >>> the
> >>> FILO screen.
> >>
> >> Great.  You could check the values from Kconfig (build/config.h) and
> >> compare
> >> them to the values from 4920.  That would be the easiest thing to fix.
> > I think the most likely culprit is SB_HT_CHAIN_ON_BUS0.  Could you
> > change it to 1 in src/mainboard/arima/hdama/Kconfig, make oldconfig,
> > and test?
> 
> I've tested it. I get the filo screen now.
> 
> When I do a probe I don't see any IDE device yet. Neither IDE nor SATA
> SIL3114 drive. This also didn't work for version 4920.

Yeah, the Sil3114 needs special handling for its SATA ports. The Tyan S2881
board has the same problem. I'm not sure about the IDE ports.

In the past I've used a linux kernel payload to get around that. Rudolf has
had success with SeaBIOS (and the SiL3114 option rom, I believe)

  http://www.coreboot.org/pipermail/coreboot/2009-February/044781.htm

but I have not been able to replicate that yet on my Tyan s2881.

I wonder if we could make Coreboot do the necessary to initialize that
controller, so that we don't need that binary blob or a full blown linux
kernel anymore.

There appear to be public datasheets for Sil3114, referenced here

  https://ata.wiki.kernel.org/index.php/Sata_sil 

And the kernel driver also knows how to bring it up, since using a linux
kernel as a payload has worked for me in the past.

Thanks,
Ward.

-- 
Ward Vandewege <ward at gnu.org>