[PATCH] x86: SETcc doesn't permit W suffix

[Cui, Lili]

> -----Original Message-----
> From: Jan Beulich <jbeulich@suse.com>
> Sent: Monday, November 18, 2024 7:04 PM
> To: H.J. Lu <hjl.tools@gmail.com>; Cui, Lili <lili.cui@intel.com>
> Cc: Binutils <binutils@sourceware.org>
> Subject: Re: [PATCH] x86: SETcc doesn't permit W suffix
> 
> On 18.11.2024 11:50, Jan Beulich wrote:
> > Accidentally I had removed No_wSuf when cloning the extra template.
> >
> > --- a/opcodes/i386-opc.tbl
> > +++ b/opcodes/i386-opc.tbl
> > @@ -562,7 +562,7 @@ set<cc>, 0xf24<cc:opc>/0, APX_F, Modrm|N
> > setzu<cc>, 0xf24<cc:opc>/0, APX_F,
> > Modrm|No_wSuf|No_lSuf|No_sSuf|No_qSuf|VexWIG|EVexMap4|ZU,
> { Reg8 }  //
> > As an extension, allow SETcc without the ZU infix, instead taking the full  //
> register as operand.
> > -set<cc>, 0xf24<cc:opc>/0, APX_F, Modrm|No_bSuf|No_sSuf|EVexMap4|ZU,
> {
> > Reg32|Reg64 }
> > +set<cc>, 0xf24<cc:opc>/0, APX_F,
> > +Modrm|No_bSuf|No_wSuf|No_sSuf|EVexMap4|ZU, { Reg32|Reg64 }
> 
> While I noticed the above issue in the context of "x86/Solaris: support Sun
> form of CMOVcc", I'll arbitrarily reply here: That same documentation (as
> before, the only AT&T spec I'm aware of) does specifically not name 'b' suffixes
> for SETcc. Still we allow them, and we even have one single test for it. It feels
> wrong though. Probably we shouldn't remove support for the pre-APX form.
> Yet I'm heavily inclined to disallow it for the APX ones that we gained only
> relatively recently. Thoughts? Either both prohibit b, or both allow b.

It looks like SDM doesn't disable the b suffix for SETcc, but I see your point, our default output is b now, I think it's better to keep them the same, either disabling b in both or allowing b in both.

Thanks,
Lili.