Re: [PATCH 1/2] target/riscv: Tentatively remove Zhinx* from ISA extension string

[Tsukasa OI]

On 2022/04/28 11:39, Weiwei Li wrote:
> 
> 在 2022/4/28 上午7:58, Alistair Francis 写道:
>> On Sun, Apr 24, 2022 at 3:22 PM Tsukasa OI <research_trasio@irq.a4lg.com> 
>> wrote:
>>> This commit disables ISA string conversion for Zhinx and Zhinxmin
>>> extensions for now.  Because extension category ordering of "H" is not
>>> ratified, their ordering is likely invalid.
>>>
>>> Once "H"-extension ordering is determined, we can add Zhinx* again.
>>>
>>> Signed-off-by: Tsukasa OI <research_trasio@irq.a4lg.com>
>> Weiwei Li does this sound alright to you?
>>
>> Alistair
> 
> Even though the rule says: "The first letter following the 'Z' conventionally 
> indicates the most closely
> 
> related alphabetical extension category, IMAFDQLCBKJTPVH", zhinx* is not 
> related to 'H' extension actually.
> 
> I think the most closely related alphabetical extension is 'F' extension.
> 
> Regards,
> 
> Weiwei Li

I don't quite agree that.

Although Zhinx* extensions are **functionally** related to 'F' (rather than
'H'), that's not how canonical ordering of Z* extensions works, at least
this is not how canonical ordering is implemented in GNU/LLVM toolchains.

Both toolchains orders Z* multi-letter extensions by second "character"-
first manner.

My interpretation is, Z* multi-letter extensions should be ordered by
category character (the second one) first.  Yes, it should have reflected
the most closely related single-letter extension, but it didn't happend
on Zhinx* extensions.

Thanks,

Tsukasa

> 
>>> ---
>>>   target/riscv/cpu.c | 2 --
>>>   1 file changed, 2 deletions(-)
>>>
>>> diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c
>>> index 0c774056c5..c765f7ff00 100644
>>> --- a/target/riscv/cpu.c
>>> +++ b/target/riscv/cpu.c
>>> @@ -954,8 +954,6 @@ static void riscv_isa_string_ext(RISCVCPU *cpu, char 
>>> **isa_str, int max_str_len)
>>>           ISA_EDATA_ENTRY(zfh, ext_zfh),
>>>           ISA_EDATA_ENTRY(zfhmin, ext_zfhmin),
>>>           ISA_EDATA_ENTRY(zfinx, ext_zfinx),
>>> -        ISA_EDATA_ENTRY(zhinx, ext_zhinx),
>>> -        ISA_EDATA_ENTRY(zhinxmin, ext_zhinxmin),
>>>           ISA_EDATA_ENTRY(zdinx, ext_zdinx),
>>>           ISA_EDATA_ENTRY(zba, ext_zba),
>>>           ISA_EDATA_ENTRY(zbb, ext_zbb),
>>> -- 
>>> 2.32.0
>>>
>