Re: [PATCH 2/3] target/i386: Enable page walking from MMIO memory

[Richard Henderson]

On 2/15/24 09:21, Richard Henderson wrote:
> On 2/15/24 05:01, Jonathan Cameron wrote:
> > -static bool ptw_translate(PTETranslate *inout, hwaddr addr)
> > +static bool ptw_translate(PTETranslate *inout, hwaddr addr, uint64_t ra)
>
> You do not need to pass in 'ra' here...
>
> >       flags = probe_access_full(inout->env, addr, 0, MMU_DATA_STORE,
> > -                              inout->ptw_idx, true, &inout->haddr, &full, 0);
> > +                              inout->ptw_idx, true, &inout->haddr, &full, ra);
>
> ... because this                                 ^^^^
> indicates a non-fault probe.
> A return address is only required for a faulting probe to raise an exception.

Bah.  We can always recurse and hit the load case too.  So:

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>


r~

> > -static inline uint32_t ptw_ldl(const PTETranslate *in)
> > +static inline uint32_t ptw_ldl(const PTETranslate *in, uint64_t ra)
> >  {
> >      if (likely(in->haddr)) {
> >          return ldl_p(in->haddr);
> >      }
> > -    return cpu_ldl_mmuidx_ra(in->env, in->gaddr, in->ptw_idx, 0);
> > +    return cpu_ldl_mmuidx_ra(in->env, in->gaddr, in->ptw_idx, ra);
> >  }
>
> However the loads do require 'ra' for the io_recompile, as you saw from the 
> backtrace.
>
>
> r~