[PATCH] RISC-V: Clobber V registers on syscalls

[Björn Töpel]

Andy Chiu <andy.chiu at sifive.com> writes:

> On Fri, Jun 23, 2023 at 12:38 AM Björn Töpel <bjorn at kernel.org> wrote:
>> This flow does avoid some context switch costs, but I wonder if this is
>> some that can be added later, when we can more reliable measure the
>> overhead. Premature optimization, and all that. ;-)
>>
>
> Sure, do you suggest any kinds of measurement, experiment, or
> benchmarking that could give out a figure on how things are different?

My take was; If you have access to actual V 1.0 hardware, and just not
Qemu, then we could do some actual real tests, measuring context switch
costs etc!


Björn