Re: [PATCH 3/3] target/ppc: Merge fsqrt and fsqrts helpers

[Víctor Colombo]

On 05/09/2022 14:20, Richard Henderson wrote:
> Well, there would of course be no separate call, but 

I didn't understand what you meant here with 'no separate call'...

> do_float_check_status would:
>
>      int status = get_float_exception_flags(&env->fp_status);
>
>      set_float_exception_flags(0, &env->fp_status);
>
> straight away.  No extra call overhead, and the steady-state of softfp 
> exception flags
> outside of an in-progress fp operation is 0.

Right, makes sense. And what about when an invalid operation occurs,
with the corresponding exception enabled bit set?
float_invalid_op_* would stop the execution and do_float_check_status
would not be called, right? So it would require to call
set_float_exception_flags there too, correct?
If that's all that's necessary, I might be able to take a look at it and
come with a possible patch.

> Anything that can raise a softfp exception and doesn't do something with 
> it, either
> immediately within the same helper, or just afterward with 
> helper_float_check_status, is
> buggy.  With those fixed, helper_reset_fpstatus may be removed entirely.

Oh, that makes sense. It's easier to implement the idea using this
assumption.


--
Víctor Cora Colombo
Instituto de Pesquisas ELDORADO
Aviso Legal - Disclaimer <https://www.eldorado.org.br/disclaimer.html>