[Qemu-devel] Re: tcg conditional set/move, round 3

[Laurent Desnogues <laurent.desnogues@gmail.com>]

On Fri, Dec 18, 2009 at 10:38 PM, Richard Henderson <rth@twiddle.net> wrote:
> On 12/18/2009 03:37 AM, Laurent Desnogues wrote:
>>>
>>>  tcg: Generic support for conditional set and conditional move.
>>
>> Needs cosmetics changes.
>
> Fixed, attachment 1.
>
>>>  tcg-x86_64: Implement setcond and movcond.
>>
>> Some cosmetics and comments, but overall good.
>
> Fixed, attachment 2.
>
>>>  tcg-i386: Implement small forward branches.
>>
>> I think this contains a bug.
>
> Fixed, attachment 3.  I've added an abort to patch_reloc to verify that the
> relocation is in range.  I've propagated the "small" flag to all of the
> branch functions so that...
>
>>>  tcg-i386: Simplify brcond2.
>>
>> I don't like the rewrite of brcond2.
>
> ... this patch is dropped.
>
>>>  tcg-i386: Implement setcond, movcond, setcond2.
>>
>> Not yet reviewed.
>
> Fixed, attachment 4.  Similar changes to the amd64 patch.

Everything looks good to me, though for i386 I'd not bet
my life it's 100% correct.  OTOH I think this is good enough
to go into mainline, so that people can start adding support
in the front-ends.

BTW for compiling to 32-bit on a 64-bit x86, the configure
script is broken as it checks gcc before having set -m32,
so passing -march=i686 will make it fail (in the end it means
I could not convince QEMU to use cmov :-).

Thanks,

Laurent

Acked-by: Laurent Desnogues <laurent.desnogues@gmail.com>