[SeaBIOS] [PATCH 0/6] pci patches

[Alexey Korolev]

On 22/05/12 18:38, Gerd Hoffmann wrote:
> On 05/22/12 08:21, Alexey Korolev wrote:
>> Hi,
>>
>>> On 05/18/12 03:50, Alexey Korolev wrote:
>>>> Hi,
>>>>
>>>> It looks Win2008 R2 is not working with this patches if 64bit devices are present and going to occupy a 64bit range.
>>>> I've got the same BSOD as before.
>>>> Is there any information I could collect to help you debug this issue?
>>> Is win2k8 fine with the fixed entry (the reverted patch)?
>> If patches are reverted no BSOD occurs, but no 64bit resources available too :(.
> No, I mean your patch which adds a 64bit entry unconditionally (which
> was reverted due to winxp issues), does that one work fine with w2k8?
> i.e. commit 74f96123e7e37c219403b50e39dabc8e8c450948
Ah I see. Yes the seabios with the static 64bit entry works without problems on Win2008r2. 
>
>>> You can try add DBUG("...") calls into the _CRS Method to figure the
>>> problematic call.  That adds debug output to the seabios log.  Doesn't
>>> work with winxp, works with win7 (and linux).  Lets hope that w2k8 is
>>> closer to win7 ...
>> I tried but log doesn't have any DBUG messages inside.
>> I've just outputed the seabios log to serial port. Is it suitable?
> Hmm, no.  The magic words are:
>
> qemu $usualargsforyourguest \
>    -chardev file,id=bioslog,path=/path/to/seabios.log \
>    -device isa-debugcon,iobase=0x402,chardev=bioslog
>
> That adds the seabios debug port and logfile windup.  The DBUG function
> in the dsdt uses the debug port too.  Redirecting seabios output to
> serial will *not* show the acpi debug stuff as this goes directly to the
> debug port, seabios isn't involved here at all, they just share the port
> so all output conveniently shows up in the same logfile.
Here is the log.

pci0 _crs: enter
pci0 _crs: 64bit
pci0 _crs: done

I've also tried the latest qemu version and ivshmem with 64bit support. Result is the same.

Thanks,
Alexey