>> The description in the spec is"Each bit of Op1 is inverted and logically 
>> ANDed with the corresponding bits in vs2",
>> so I think the "and" should be placed  outside.

Ah. Yes.


juzhe.zh...@rivai.ai
 
From: Feng Wang
Date: 2023-12-20 16:09
To: juzhe.zh...@rivai.ai; gcc-patches
CC: kito.cheng; Jeff Law
Subject: Re: Re: [PATCH v4 2/3] RISC-V: Add crypto machine descriptions
2023-12-20 15:12 juzhe.zhong <juzhe.zh...@rivai.ai> wrote:
 
>+       (and:VI
>+         (match_operand:VI 3 "register_operand"         "vr, vr, vr, vr")
>+         (not:VI (match_operand:VI 4 "register_operand" "vr, vr, vr, vr")))
 
>Swap the order:
 
>
 
>(not:VI (match_operand:VI 4 "register_operand" "vr, vr, vr, vr")
 
>       (and:VI
 
>         (match_operand:VI 3 "register_operand"         "vr, vr, vr, vr")
 
The description in the spec is"Each bit of Op1 is inverted and logically ANDed 
with the corresponding bits in vs2",
so I think the "and" should be placed  outside.
 
>
 
>+       (and:VI_QHS
 
>+          (match_operand:VI_QHS 3 "register_operand"   "vr, vr,vr, vr")
 
>+          (not:<VEL>
 
>+            (match_operand:<VEL> 4 "register_operand"  " r,  r, r,  r")))
 
>
 
>->
 
>
 
>+       (and:VI_QHS+          (not:VI_QHS+            (vec_duplicate:VI_QHS
 
>
 
>+            (match_operand:<VEL> 4 "register_operand"  " r,  r, r,  r")))
 
 
 
>+          (match_operand:VI_QHS 3 "register_operand"   "vr, vr,vr, vr"))
 
OK. Will add vec_duplicate.
 
>
 
>Can you add EEW64 vx test on RV32 ?
 
>
 
>
 
>
 
>
 
>juzhe.zh...@rivai.ai
 
> 
 
The test cases are added at [PATCH v4 3/3] RISC-V: Add crypto vector 
api-testing cases.
gcc.target/riscv/zvbb_vandn_vx_constraint.c
gcc.target/riscv/zvbc_vx_constraint-2.c: New test.
gcc.target/riscv/zvbc_vx_constraint-1.c: New test.
 
Thanks.
 

Reply via email to