[Qemu-ppc] PPC hardfloat (was: Re: [Qemu-devel] [Bug 1841442] Re: floating point emulation can fail to set FE_INEXACT)

[BALATON Zoltan]

On Mon, 26 Aug 2019, Richard Henderson wrote:
> That said, qemu-system-ppc64 will *never* use hardfloat, because ppc always
> need the current and correct result of inexact for emulation of the FI bit,
> which requires that we use the softfloat path.

That's bad news. I hoped that hardfloat for PPC can be implemented and 
previously it was thought it could be done after some reorganisation to 
prevent always reseting flags by moving them to environment or somewhere 
else (but I don't remember the details and maybe I never fully understood 
it in the first place). Could you please explain why do you think it's not 
possible? I don't know much about IEEE floating point but couldn't the 
flag be set based on exceptions we get? If the operations did not generate 
inexact eception the why should we set the flag and why is it not enough 
to set it when exception is caught to emulate it? (After the above reorg 
of course to preserve flags between FP ops.)

Currently using softfloat is a big source of slowness in PPC emulation 
that hopefully could be improved by using hardfloat where possible.

Regards,
BALATON Zoltan