Re: [PATCH 2/5] target/riscv: Add SiFive custom int8 matmul extension implied rules

[Max Chou <max.chou@sifive.com>]

On 2026-08-06 18:07, Daniel Henrique Barboza wrote:
> 
> 
> On 7/21/2026 9:20 AM, Max Chou wrote:
> > From: Frank Chang <frank.chang@sifive.com>
> > 
> > Add SiFive custom int8 matmul extension implied rules to enable the
> > implied extensions of SiFive custom int8 matmul extension recursively.
> > 
> > Signed-off-by: Frank Chang <frank.chang@sifive.com>
> > Signed-off-by: Max Chou <max.chou@sifive.com>
> > ---
> >   target/riscv/cpu.c | 19 +++++++++++++++++++
> >   1 file changed, 19 insertions(+)
> > 
> > diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c
> > index e643d7c7af..70baa054f9 100644
> > --- a/target/riscv/cpu.c
> > +++ b/target/riscv/cpu.c
> > @@ -2811,6 +2811,24 @@ static RISCVCPUImpliedExtsRule ZVFBFA_IMPLIED = {
> >       },
> >   };
> > +static RISCVCPUImpliedExtsRule XSFVQMACCDOD_IMPLIED = {
> > +    .ext = CPU_CFG_OFFSET(ext_xsfvqmaccdod),
> > +    .implied_multi_exts = {
> > +        CPU_CFG_OFFSET(ext_zve32x),
> > +
> > +        RISCV_IMPLIED_EXTS_RULE_END
> > +    },
> > +};
> > +
> > +static RISCVCPUImpliedExtsRule XSFVQMACCQOQ_IMPLIED = {
> > +    .ext = CPU_CFG_OFFSET(ext_xsfvqmaccqoq),
> > +    .implied_multi_exts = {
> > +        CPU_CFG_OFFSET(ext_zve32x),
> > +
> > +        RISCV_IMPLIED_EXTS_RULE_END
> > +    },
> > +};
> > +
> >   RISCVCPUImpliedExtsRule *riscv_misa_ext_implied_rules[] = {
> >       &RVA_IMPLIED, &RVD_IMPLIED, &RVF_IMPLIED,
> >       &RVM_IMPLIED, &RVV_IMPLIED, &RVG_IMPLIED,
> > @@ -2832,6 +2850,7 @@ RISCVCPUImpliedExtsRule *riscv_multi_ext_implied_rules[] = {
> >       &ZVKS_IMPLIED,  &ZVKSC_IMPLIED, &ZVKSG_IMPLIED, &SHA_IMPLIED,
> >       &SSCFG_IMPLIED, &SUPM_IMPLIED, &SSPM_IMPLIED, &SMCTR_IMPLIED,
> >       &SSCTR_IMPLIED, &SSSTATEEN_IMPLIED,
> > +    &XSFVQMACCDOD_IMPLIED, &XSFVQMACCQOQ_IMPLIED,
> 
> Not related to this patch but I wonder if we should put one rule per line
> in this array.  Every once in a while we have to add a rule that is
> in a line that already has a lot of stuff, then we need to change multiple
> lines to accommodate the new rule.
> 
> 
Hi Daniel,

Actually, I share the same observation and thought, and I agree with
you. Let me update this part in v2.

Thanks for the suggestion!
rnax

> As for the patch:
> 
> 
> 
> Reviewed-by: Daniel Henrique Barboza <daniel.barboza@oss.qualcomm.com>
> 
> 
> 
> 
> 
> >       NULL
> >   };
>