Re: [Qemu-devel] kqemu version 1.3.0pre5

[Fabrice Bellard <fabrice@bellard.org>]

Ed Swierk wrote:
> On 3/28/06, Jens Axboe <qemu@kernel.dk> wrote:
> 
>>>monitor/mwait feature present.
>>>using mwait in idle threads.
>>
>>[snip]
>>
>>
>>>invalid operand: 0000 [#1]
>>>Modules linked in:
>>>CPU:    0
>>>EIP:    0060:[<c0101147>]    Not tainted VLI
>>>EFLAGS: 00010246   (2.6.14-1.1656_FC4)
>>>EIP is at mwait_idle+0x2f/0x41
>>
>>I don't think qemu supports PNI, which includes the monitor/mwait
>>additions. I wonder why Linux detects that. You can probably get around
>>it for now by either passing idle=poll as a boot parameter, or compile
>>your kernel for plain i586 for instance.
> 
> 
> It seems that with -kernel-kqemu, the guest kernel is seeing the CPUID
> of the host machine rather than the one normally generated by qemu.
> 
> The workarounds you suggest do work--thanks for your help. However,
> ideally kqemu would trap the CPUID instruction and mask the feature
> bits for unsupported CPU features.

The problem is that it is not possible to trap on the CPUID instruction 
:-( So the only possible patch is to support PNI in QEMU. For 
monitor/mwait for example, doing nops should suffice...

Fabrice.