[SeaBIOS] varlow/extrastack vs code

Paolo Bonzini pbonzini at redhat.com
Mon Feb 20 10:43:35 CET 2017



On 18/02/2017 17:55, Kevin O'Connor wrote:
> On Sat, Feb 18, 2017 at 04:11:28PM +0100, Paolo Bonzini wrote:
>> On 18/02/2017 04:45, Kevin O'Connor wrote:
>>> On Fri, Feb 17, 2017 at 02:12:57PM +0100, Paolo Bonzini wrote:
>>>> On 15/02/2017 18:35, Dr. David Alan Gilbert wrote:
>>>>> Yes it seems to.
>>>>> One worry is that if we ever fix the qemu triple-fault so it really
>>>>> does what you're describing and only resets the CPU, then I'm not
>>>>> sure your int3 is the right choice.
>>>>>
>>>>> The other question is whether that protected-mode exit switch
>>>>> works in practice on qemu; it's going to come back with a lot of it's
>>>>> devices reset.
>>>>
>>>> Right, keyboard reset (and port 92h reset) should be an INIT and not a
>>>> full system reset.  But that's also something that should be fixed, not
>>>> a quirk to rely on...
>>>
>>> Could we rely on the pci reset mechanism (port 0xcf9)?
>>
>> That's only 3-4 years old, but 0xcf9 followed by keyboard reset should
>> always work.
> 
> Since this would be qemu specific, I think it would be okay if reset
> only worked on qemu versions from the last few years.  Or, are you
> saying only some machine types would support it?  I checked, and
> reboots on isapc don't currently work at all (looks like the bios
> isn't mapped to 0xfff00000 on isapc) so I don't think that's a
> stopper.

isapc should never be SMP, so the keyboard reset would be enough even if
it were fixed to be an INIT.  Hence my suggestion of using 0xcf9 for
recent QEMU with PCI machine types, followed by keyboard reset for old
QEMU + isapc.

Paolo



More information about the SeaBIOS mailing list