RE: [PATCH v4 3/9] Support APX GPR32 with extend evex prefix

["Cui, Lili" <lili.cui@intel.com>]

> -----Original Message-----
> From: Jan Beulich <jbeulich@suse.com>
> Sent: Thursday, January 4, 2024 5:08 PM
> To: Cui, Lili <lili.cui@intel.com>
> Cc: Lu, Hongjiu <hongjiu.lu@intel.com>; binutils@sourceware.org
> Subject: Re: [PATCH v4 3/9] Support APX GPR32 with extend evex prefix
> 
> On 25.12.2023 13:23, Cui, Lili wrote:
> >> On 19.12.2023 13:12, Cui, Lili wrote:
> >>>  #define Vsz256 Vsz=VSZ256
> >>>  #define Vsz512 Vsz=VSZ512
> >>>
> >>> +// The template supports VEX format for cpuid and EVEX format for
> >>> +cpuid &
> >> apx_f.
> >>> +#define APX_F(cpuid) cpuid&(cpuid|APX_F)
> >>
> >> I think the comment wants to go into further detail. Please can you
> >> go back to read what I said when I suggested this construct, in
> >> particular regarding the stripping then done? However, with you not
> >> having found a need to fiddle with cpu_flags_match(), I wonder if this
> construct is needed in the first place.
> >> The earlier suggestion was entirely based on the assumption that
> >> stripping similar to that for other combined VEX/EVEX templates would
> >> be needed here, t
> >>
> >
> > Seeing this, I realized the problem and checked opcodes/i386-tbl.h,
> > for the following entry, we want to set CpuAPX_F and CpuBMI to 1, but
> > gen.c doesn't seem to support the format "cpuid&(cpuid|APX_F)",
> 
> I suppose you mean "cpuid|(cpuid&APX_F)"?
> 

I tried cpuid|(cpuid&APX_F). Currently this format is not supported.

> > in fact bzhi sets CpuBMI to 1 and CpuAPX_F to 0 . I'm not familiar with the
> relevant logic in gen.c and don't know how to debug it. when you have time,
> could you help take a look ?
> 
> Well, now that the patch was committed I of course can easily take a look.
> But with such an outstanding issue, how could the patch have been
> committed in the first place? And isn't it a requirement for making changes to
> i386-gen.c that you sufficiently understand the logic there (which admittedly
> has grown non-trivial)? Anyway - I will need to find time to play with what has
> been committed, but that's not likely to happen in time for 2.42. Hence it is
> quite possible that 2.42 will turn out to have incomplete and/or broken APX_F
> support. But perhaps this was intended to be this way by H.J., when he
> approved the entire series without even a single comment.
> 

Although there is a problem in this place, now apx can handle all cases correctly. On the contrary, when this place is corrected, cpu_flags_match() needs to be adjusted, otherwise there will be testcase fail.

Thanks,
Lili.