Re: [PATCH] RISC-V: Fix regressions due to 86de9b66480b710202a2898cf513db105d8c432f

2024-01-22 Thread Robin Dapp 
> No, we didn't undo the optimization.
> 
> We just disallow move pattern for (set (reg) (VL_REGNUM)).

Ah, what I referred to was the opposite direction.  We allow
(subreg:V8QI (reg:DI ...)) which is not touched by this patch.

Then it is OK.

Regards
 Robin

Re: Re: [PATCH] RISC-V: Fix regressions due to 86de9b66480b710202a2898cf513db105d8c432f

2024-01-22 Thread juzhe.zh...@rivai.ai 
No, we didn't undo the optimization.

We just disallow move pattern for (set (reg) (VL_REGNUM)).




juzhe.zh...@rivai.ai
 
From: Robin Dapp
Date: 2024-01-22 19:25
To: Juzhe-Zhong; gcc-patches
CC: rdapp.gcc; kito.cheng; kito.cheng; jeffreyalaw
Subject: Re: [PATCH] RISC-V: Fix regressions due to 
86de9b66480b710202a2898cf513db105d8c432f
 
Hi Juzhe,
 
in principle this seems ok to me but I wonder about:
 
> We shouldn't worry about subreg:...VL_REGNUM since it's impossible
> that we can have such situation,
 
I think we allow this in legitimize_move for situations like
(subreg:SI (reg:V4QI)).  That was not added for correctness but
optimization - are we sure we don't undo this optimization with
that change?
 
Regards
Robin

Re: [PATCH] RISC-V: Fix regressions due to 86de9b66480b710202a2898cf513db105d8c432f

2024-01-22 Thread Robin Dapp 


Hi Juzhe,

in principle this seems ok to me but I wonder about:

> We shouldn't worry about subreg:...VL_REGNUM since it's impossible
> that we can have such situation,

I think we allow this in legitimize_move for situations like
(subreg:SI (reg:V4QI)).  That was not added for correctness but
optimization - are we sure we don't undo this optimization with
that change?

Regards
 Robin