[PATCH 5/6] x86/APX: optimize certain {nf}-form insns to BMI2 ones

[Jiang, Haochen]

> -----Original Message-----
> From: Jan Beulich <jbeulich@suse.com>
> Sent: Friday, June 14, 2024 8:14 PM
> To: Binutils <binutils@sourceware.org>
> Cc: H.J. Lu <hjl.tools@gmail.com>; Cui, Lili <lili.cui@intel.com>
> Subject: [PATCH 5/6] x86/APX: optimize certain {nf}-form insns to BMI2 ones
> 
> ..., as those leave EFLAGS untouched anyway. That's a shorter encoding,
> available as long as no eGPR is in use anywhere.
> ---
> RFC: Especially because of the need to explicitly enable BMI2 this may
>      be deemed not worth it; seeking views.
> 

If we need explicitly enable BMI2, maybe doing that in compiler is a better choice
since we always enable everything by default in assembler in the real world usage.

However, since actually the change only benefits codesize, not latency (Throughput
might be benefited but not for all cases). Doing that in assembler seems also
reasonable.

Going either way seems ok to me.

Thx,
Haochen