Re: [Qemu-devel] [PATCH 4/7] move CPU models from cpus-x86_64.conf to C

[Andreas Färber]

Am 10.09.2012 17:07, schrieb Eduardo Habkost:
> On Mon, Sep 10, 2012 at 04:58:39PM +0200, Andreas Färber wrote:
>> Am 10.09.2012 16:50, schrieb Don Slutz:
>>> On 09/10/12 09:04, Igor Mammedov wrote:
>>>> But question unrelated to this patch is still stand if ia64 is valid
>>>> bit for
>>>> 01.EDX[30]?
>>>>
>>>>
>>> Intel® Processor Identification
>>> and the CPUID Instruction
>>> Application Note 485
>>> January 2006
>>>
>>> Order Number: 241618-030
>>>
>>> ...
>>>
>>> Updated Table 3-5 to include the feature flag definition (EDX[30]) for
>>> IA64 capabilities.
>>> ...
>>> 30 IA64 IA64 Capabilities The processor is a member of the Intel®
>>> Itanium® processor family
>>>                            and currently operating in IA32 emulation mode.
>>>
>>> ---------------
>>>
>>> Says that it is. Along with http://en.wikipedia.org/wiki/CPUID and
>>> http://www.sandpile.org/x86/cpuid.htm#level_0000_0001h (IA-64)
>>
>> Don't those semantics contradict the use in qemu-system-x86_64 or
>> qemu-system-i386 rather than qemu-system-ia64 then? We don't model ia64
>> CPUs here (just like we don't model ppc64 CPUs in ppc) so the flag could
>> never become 1 IIUC.
> 
> Correct, and the bit is always filtered out on both TCG and KVM modes.
> The name is in the table because we know the name/meaning of that
> feature bit, but it's impossible to enable it.
> 
> That said, I don't mind removing it from the table just to avoid
> confusion, but I also wouldn't mind keeping it (as it's harmless).

No objections from my side against having it for informational purpose.

Andreas

-- 
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer; HRB 16746 AG Nürnberg