Re: [PATCH v2 1/3] RISC-V Vector Extension Support

["Maciej W. Rozycki" <macro@orcam.me.uk>]

On Fri, 25 Sep 2026, Kirill Radkin wrote:

> diff --git a/gdb/arch/riscv.h b/gdb/arch/riscv.h
> index 1443bc20a7d..0261d4d0f9e 100644
> --- a/gdb/arch/riscv.h
> +++ b/gdb/arch/riscv.h
> @@ -51,7 +51,7 @@ struct riscv_gdbarch_features
>       target should be 16 and 4 for an embedded subset compliant target (with
>       'Zve32*' extension), but GDB doesn't currently mind, and will accept any
>       vector size.  */
> -  int vlen = 0;
> +  int vlenb = 0;

 What's the rationale behind the member rename?  If this is legitimate, 
then can you please submit a preparatory change that does only the rename 
without mixing it with new code, which only causes obfuscation?

> @@ -68,9 +68,8 @@ struct riscv_gdbarch_features
>    /* Equality operator.  */
>    bool operator== (const struct riscv_gdbarch_features &rhs) const
>    {
> -    return (xlen == rhs.xlen && flen == rhs.flen
> -	    && embedded == rhs.embedded && vlen == rhs.vlen
> -	    && has_fflags_reg == rhs.has_fflags_reg
> +    return (xlen == rhs.xlen && flen == rhs.flen && embedded == rhs.embedded
> +	    && vlenb == rhs.vlenb && has_fflags_reg == rhs.has_fflags_reg
>  	    && has_frm_reg == rhs.has_frm_reg
>  	    && has_fcsr_reg == rhs.has_fcsr_reg);
>    }
> @@ -84,13 +83,10 @@ struct riscv_gdbarch_features
>    /* Used by std::unordered_map to hash feature sets.  */
>    std::size_t hash () const noexcept
>    {
> -    std::size_t val = ((embedded ? 1 : 0) << 10
> -		       | (has_fflags_reg ? 1 : 0) << 11
> -		       | (has_frm_reg ? 1 : 0) << 12
> -		       | (has_fcsr_reg ? 1 : 0) << 13
> -		       | (xlen & 0x1f) << 5
> -		       | (flen & 0x1f) << 0
> -		       | (vlen & 0x3fff) << 14);
> +    std::size_t val
> +      = ((embedded ? 1 : 0) << 10 | (has_fflags_reg ? 1 : 0) << 11
> +	 | (has_frm_reg ? 1 : 0) << 12 | (has_fcsr_reg ? 1 : 0) << 13
> +	 | (xlen & 0x1f) << 5 | (flen & 0x1f) << 0 | (vlenb & 0x3fff) << 14);

 The only functional change here is AFAICT s/vlen/vlenb/ and the rest is 
reformatting, which obfuscates the update.  Would you then please remove 
the formatting changes and if you indeed think they're needed, then submit 
them as a follow-up cleanup?

 I'll yet see if I can look through the rest of your changes, but this 
just stood out prominently for me.

  Maciej