i440LX Patch

[Maciej Pijanka]

Hello,

I try to make biostar m6tld running, and created patch for 440lx using 440bx code as template.
Patch in attachment

Best regards
Maciej

[ron minnich]

It looks fine to me, although I think you can drop the Config.lb.

Acked-by: Ronald G. Minnich

[Maciej Pijanka]

On Fri, 27 Nov 2009, ron minnich wrote:
> It looks fine to me, although I think you can drop the Config.lb.

I left it because many other targets have it too.
> Acked-by: Ronald G. Minnich

I noticed that i forgot to add

Signed-off by: Maciej Pijanka

ps. If noone else see serious problems with this patch/code and have commit
rights please commit.

Maciej

[Stefan Reinauer]

Maciej Pijanka wrote:
> Hello,
>
> I try to make biostar m6tld running, and created patch for 440lx using 440bx code as template.
> Patch in attachment
>
> Best regards
> Maciej

Awesome! Committed as r4967... Do you have a mainboard patch to go with
it, too?

When you say "try" what was / is the problem?

All the best

Stefan

[Maciej Pijanka]

On Sat, 28 Nov 2009, Stefan Reinauer wrote:
> Maciej Pijanka wrote:
> > Hello,
> >
> > I try to make biostar m6tld running, and created patch for 440lx using 440bx code as template.
> > Patch in attachment
> >
> > Best regards
> > Maciej
>
> Awesome! Committed as r4967... Do you have a mainboard patch to go with
> it, too?

For some quality of patch, yes, but board dies around loading payload, and i did not have enough
time so far to debug, last attempt stuck when coreboot at machine showed me gdb beacon.
> When you say "try" what was / is the problem?

I am not sure if all code is correct, and there are few missing parts, ie cas latency is hard
coded to 3, agp is disabled (and i failed to find why i cant change APBASE), also i have weird
problem with reading/writing on this machine with read_config16/write_config16 because they return
incorrect data. I mean register *is* 16bit accessible, (it was DRT iirc), but two 8bit reads glued
with shift and or give other result than 16bit read, but i don't have log with output of 16bit read
on hand right now.

Maciej

[ron minnich]

On Sat, Nov 28, 2009 at 2:30 AM, Maciej Pijanka
 wrote:
> I mean register *is* 16bit accessible, (it was DRT iirc), but two 8bit reads glued
> with shift and or give other result than 16bit read, but i don't have log with output of 16bit read
> on hand right now.

There are weird cases where 8 bit accesses go to "a different place"
than a wider access. The first one that comes to mind is using cf8 for
pci config accesses but cf9 to reset the machine.

ron