[Qemu-devel] [RFT] ppc64 OpenBIOS

[Qemu-devel] [RFT] ppc64 OpenBIOS

[Andreas Färber]

Hello,

Except for some fs/iso9660/ issues, ppc64 guests are booting for me  
with the last five OpenBIOS patches posted.

You can find them together with the config change to use fs/grubfs/  
here:

   git://repo.or.cz/openbios/afaerber.git ppc64-boot

Please test your favorite ppc64 kernels and images with obj-ppc64/ 
openbios-qemu.elf!

If I don't hear of regressions, I'll commit those patches. Once that  
is done and the iso9660 issues hopefully fixed, I would ask for Blue  
to commit openbios-ppc and openbios-ppc64 blobs to QEMU. I'll then  
prepare a QEMU patch to switch the default qemu-system-ppc64  
ppc_newworld.c BIOS from openbios-ppc to openbios-ppc64.

Thanks,
Andreas

[Qemu-devel] Re: [RFT] ppc64 OpenBIOS

[Blue Swirl]

On Sat, Dec 18, 2010 at 2:30 AM, Andreas Färber <andreas.faerber@web.de> wrote:
> Hello,
>
> Except for some fs/iso9660/ issues, ppc64 guests are booting for me with the
> last five OpenBIOS patches posted.

What are the issues?

>
> You can find them together with the config change to use fs/grubfs/ here:
>
>  git://repo.or.cz/openbios/afaerber.git ppc64-boot
>
> Please test your favorite ppc64 kernels and images with
> obj-ppc64/openbios-qemu.elf!

Passes my simple tests. The only problem is that building forthstrap
on a 64 bit host fails.

> If I don't hear of regressions, I'll commit those patches. Once that is done
> and the iso9660 issues hopefully fixed, I would ask for Blue to commit
> openbios-ppc and openbios-ppc64 blobs to QEMU. I'll then prepare a QEMU
> patch to switch the default qemu-system-ppc64 ppc_newworld.c BIOS from
> openbios-ppc to openbios-ppc64.

I think openbios-ppc should be renamed to openbios-ppc32 for
consistency with Sparc naming.

[Qemu-devel] Re: [OpenBIOS] [RFT] ppc64 OpenBIOS

[Andreas Färber]

Am 18.12.2010 um 17:33 schrieb Blue Swirl:

> On Sat, Dec 18, 2010 at 2:30 AM, Andreas Färber <andreas.faerber@web.de 
> > wrote:
>> Hello,
>>
>> Except for some fs/iso9660/ issues, ppc64 guests are booting for me  
>> with the
>> last five OpenBIOS patches posted.
>
> What are the issues?

It still hangs after the "Trying" line, as reported in the "ppc64 boot  
issues" thread.
The exact cause is as of yet unknown.

>> You can find them together with the config change to use fs/grubfs/  
>> here:
>>
>>  git://repo.or.cz/openbios/afaerber.git ppc64-boot
>>
>> Please test your favorite ppc64 kernels and images with
>> obj-ppc64/openbios-qemu.elf!
>
> Passes my simple tests. The only problem is that building forthstrap
> on a 64 bit host fails.

Right. Should we hack switch-arch as an interim solution to make it  
work?

>> If I don't hear of regressions, I'll commit those patches. Once  
>> that is done
>> and the iso9660 issues hopefully fixed, I would ask for Blue to  
>> commit
>> openbios-ppc and openbios-ppc64 blobs to QEMU. I'll then prepare a  
>> QEMU
>> patch to switch the default qemu-system-ppc64 ppc_newworld.c BIOS  
>> from
>> openbios-ppc to openbios-ppc64.
>
> I think openbios-ppc should be renamed to openbios-ppc32 for
> consistency with Sparc naming.

No objection from my side.

Andreas

[Qemu-devel] Re: [OpenBIOS] [RFT] ppc64 OpenBIOS

[Blue Swirl]

On Sat, Dec 18, 2010 at 4:54 PM, Andreas Färber <andreas.faerber@web.de> wrote:
> Am 18.12.2010 um 17:33 schrieb Blue Swirl:
>
>> On Sat, Dec 18, 2010 at 2:30 AM, Andreas Färber <andreas.faerber@web.de>
>> wrote:
>>>
>>> Hello,
>>>
>>> Except for some fs/iso9660/ issues, ppc64 guests are booting for me with
>>> the
>>> last five OpenBIOS patches posted.
>>
>> What are the issues?
>
> It still hangs after the "Trying" line, as reported in the "ppc64 boot
> issues" thread.
> The exact cause is as of yet unknown.
>
>>> You can find them together with the config change to use fs/grubfs/ here:
>>>
>>>  git://repo.or.cz/openbios/afaerber.git ppc64-boot
>>>
>>> Please test your favorite ppc64 kernels and images with
>>> obj-ppc64/openbios-qemu.elf!
>>
>> Passes my simple tests. The only problem is that building forthstrap
>> on a 64 bit host fails.
>
> Right. Should we hack switch-arch as an interim solution to make it work?

I'd prefer a real fix, unless the problem persists.