[PATCH] Add support for Traverse Technologies Geos mainboard

[Nathan Williams]

This board is similar to the AMD Norwich mainboard.

Signed-off-by: Nathan Williams

Issues with Geos mainboard

[Nathan Williams]

On 20/05/2010 11:23 AM, Nathan Williams wrote:
> This board is similar to the AMD Norwich mainboard.
>
> Signed-off-by: Nathan Williams

Currently I'm having 4 issues with coreboot-v4 on this board:

1.  On initial power-on, SeaBIOS doesn't get past "Press F12 for boot menu."

If I hit reboot, SeaBIOS completes:

Press F12 for boot menu.

Mapping hd drive 0x000fdc20 to 0
finalize PMM
malloc finalize
...

Once Linux has booted, successive reboots work fine.  It's only the initial power-on
that I seem to have trouble with.

I have tried going back as far as Rev 4976 (and patching in Edwin's RAM fixes), but it seems
this issue is still there.

2. Minor issue with valid_area() in src/boot/selfboot.c

I have a temporary workaround for this one where I simply return 1 and skip the test that would otherwise fail.

3. Rev 5476 introduces a test for CS5536, which is failing

Peter Stuge has been helping me on IRC to investigate this issue.  We think the key is that for some reason all_devices is 0.

4. Rev 5543 crashes after "Calling VSA module..."

http://coreboot.pastebin.com/jHxZBYZF

Calling VSA module...
oprom: INT# 0xd
oprom: eax: 10ffdf00 ebx: 80007800 ecx: 00000000 edx: 25fffc02
oprom: ebp: 00010ff4 esp: 00000fec edi: 00000000 esi: 00002080
oprom:  ip: 0647      cs: 6000   flags: 00000013
Oops, exception 13 while executing option rom
oprom: INT# 0xd
oprom: eax: 10ffdf00 ebx: 80007800 ecx: 00000000 edx: 25fffc02
oprom: ebp: 00010ff4 esp: 00000fec edi: 00000000 esi: 00002080
oprom:  ip: 0647      cs: 6000   flags: 00000013
...

Ignoring issue 1., I can successfully boot Linux from coreboot r5564 if I reverse r5476 and r5543.
If I only reverse r5543, Linux fails to boot because of an issue with my IDE: http://coreboot.pastebin.com/SGsRxdka

Any ideas/suggestions would be greatly appreciated.

Regards,
Nathan

Issues with Geos mainboard

[Nathan Williams]

On 20/05/2010 3:04 PM, Nathan Williams wrote:
> Minor issue with valid_area() in src/boot/selfboot.c
>
> I have a temporary workaround for this one where I simply return 1 and skip the test that would otherwise fail.

I've just realised that this one has been fixed recently, so I don't need my workaround patch any more.

Issues with Geos mainboard

[Nathan Williams]

On 20/05/2010 3:04 PM, Nathan Williams wrote:
> On initial power-on, SeaBIOS doesn't get past "Press F12 for boot menu."

...
> Rev 5476 introduces a test for CS5536, which is failing
>
> Peter Stuge has been helping me on IRC to investigate this issue.  We think the key is that for some reason all_devices is 0.
> Rev 5543 crashes after "Calling VSA module..."

I've done some testing with an Alix.1D and can reproduce issues 3. and 4. but for some reason I can't reproduce 1. on the Alix.

Regards,
Nathan

Issues with Geos mainboard

[Kevin O'Connor]

On Thu, May 20, 2010 at 03:04:01PM +1000, Nathan Williams wrote:
> On 20/05/2010 11:23 AM, Nathan Williams wrote:
> > This board is similar to the AMD Norwich mainboard.
> >
> > Signed-off-by: Nathan Williams
>
> Currently I'm having 4 issues with coreboot-v4 on this board:
> On initial power-on, SeaBIOS doesn't get past "Press F12 for boot menu."

Can you post the full log?  An error in "Press F12" is likely a
problem with the RTC.

-Kevin

Issues with Geos mainboard

[Nathan Williams]

On 20/05/2010 10:56 PM, Kevin O'Connor wrote:
> On Thu, May 20, 2010 at 03:04:01PM +1000, Nathan Williams wrote:
> > On 20/05/2010 11:23 AM, Nathan Williams wrote:
> > > This board is similar to the AMD Norwich mainboard.
> > >
> > > Signed-off-by: Nathan Williams
> >
> > Currently I'm having 4 issues with coreboot-v4 on this board:
> > On initial power-on, SeaBIOS doesn't get past "Press F12 for boot menu."
>
> Can you post the full log?  An error in "Press F12" is likely a
> problem with the RTC.
>
> -Kevin

http://coreboot.pastebin.com/nGUhXN40

Nathan

Issues with Geos mainboard

[Nathan Williams]

On 20/05/2010 10:56 PM, Kevin O'Connor wrote:
> On Thu, May 20, 2010 at 03:04:01PM +1000, Nathan Williams wrote:
> > On 20/05/2010 11:23 AM, Nathan Williams wrote:
> > > This board is similar to the AMD Norwich mainboard.
> > >
> > > Signed-off-by: Nathan Williams
> >
> > Currently I'm having 4 issues with coreboot-v4 on this board:
> > On initial power-on, SeaBIOS doesn't get past "Press F12 for boot menu."
>
> Can you post the full log?  An error in "Press F12" is likely a
> problem with the RTC.
>
> -Kevin

Ah.. It works if I have the CMOS battery in.

Thanks Kevin.

Issues with Geos mainboard

[Kevin O'Connor]

On Fri, May 21, 2010 at 06:26:04PM +1000, Nathan Williams wrote:
> On 20/05/2010 10:56 PM, Kevin O'Connor wrote:
> > On Thu, May 20, 2010 at 03:04:01PM +1000, Nathan Williams wrote:
> > > Currently I'm having 4 issues with coreboot-v4 on this board:
> > > On initial power-on, SeaBIOS doesn't get past "Press F12 for boot menu."
> >
> > Can you post the full log?  An error in "Press F12" is likely a
> > problem with the RTC.
>
> Ah.. It works if I have the CMOS battery in.

That's odd.  I don't know why a reboot would fix the RTC if it didn't
work the first boot.  Interestingly, some time back Stefan reported a
board which could only get past "Press F12" on the first boot.

I put together a patch to remove the dependency on the RTC periodic
irq for the boot menu.  (See:
http://www.seabios.org/pipermail/seabios/2010-May/000541.html ).  I
hope to check this into the SeaBIOS git repo in the next few days.

This should make SeaBIOS more robust even with these hardware errors.
(The bootloader/OS may still fail, but it should get past the SeaBIOS
init.)

-Kevin

[ron minnich]

On Wed, May 19, 2010 at 6:23 PM, Nathan Williams  wrote:
> This board is similar to the AMD Norwich mainboard.
>
> Signed-off-by: Nathan Williams

Acked-by: Ronald G. Minnich