Re: [PATCH 2/2] tcg/riscv: Use BEXTI for single-bit extractions

[Richard Henderson <richard.henderson@linaro.org>]

On 1/7/25 02:29, Philippe Mathieu-Daudé wrote:
> Hi Richard,
> 
> On 2/1/25 19:16, Richard Henderson wrote:
>> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
>> ---
>>   tcg/riscv/tcg-target-has.h |  8 +++++++-
>>   tcg/riscv/tcg-target.c.inc | 13 +++++++++++--
>>   2 files changed, 18 insertions(+), 3 deletions(-)
>>
>> diff --git a/tcg/riscv/tcg-target-has.h b/tcg/riscv/tcg-target-has.h
>> index 10e61edc45..ea38ee5cbb 100644
>> --- a/tcg/riscv/tcg-target-has.h
>> +++ b/tcg/riscv/tcg-target-has.h
>> @@ -64,7 +64,13 @@ tcg_target_extract_valid(TCGType type, unsigned ofs, unsigned len)
>>           /* ofs > 0 uses SRLIW; ofs == 0 uses add.uw. */
>>           return ofs || (cpuinfo & CPUINFO_ZBA);
>>       }
>> -    return (cpuinfo & CPUINFO_ZBB) && ofs == 0 && len == 16;
>> +    switch (len) {
>> +    case 1:
>> +        return (cpuinfo & CPUINFO_ZBS) && ofs != 0;
> 
> Why can't we have ofs=0?

To prefer ANDI.


r~