Is RISC-V Static Call worth implementing ?

[Andrew Jones]

On Fri, Sep 13, 2024 at 06:50:37PM GMT, Samuel Holland wrote:
> On 2024-09-13 3:55 PM, Conor Dooley wrote:
> > On Fri, Sep 13, 2024 at 03:05:33AM -0500, Juhan Jin wrote:
> >> Hi folks,
> >>
> >> I’m interested in implementing Static Call for RISC-V, and I want to
> >> know whether it is worth the efforts to implement Static Call for
> >> RISC-V.
> >>
> >> Does the aforementioned benefits merit a RISC-V static call implementation 
> >> (especially inline)? Or are the benefits so negligible that it’s simply not 
> >> worth the effort to do a RISC-V implementation?
> > 
> > Pretty sure we've talked about wanting it before - Samuel, Alex, Drew
> > Jones or Palmer might remember best what exactly we wanted it for
> > however, as I do not.
> 
> If I remember correctly, we had discussed it in the context of non-coherent DMA
> operations, but we decided that the benefit was negligible for that use case.
> 
> There are other places that could likely benefit from static calls, such as
> Sv39/Sv48/Sv57 differences, some users of riscv_has_extension_*(), or maybe some
> hot indirect SBI functions like sbi_set_timer(). But it depends on the overhead.

When guests have steal time accounting enabled then update_rq_clock()
would also likely benefit.

Thanks,
drew