[PATCH 4/5] lib: sbi: Extend hart protection abstraction to allow ID configuration

[Anup Patel]

On Wed, Aug 5, 2026 at 3:08 PM Yu-Chien Peter Lin <peter.lin at sifive.com> wrote:
>
> Hi Anup,
>
> On Mon, Jul 20, 2026 at 04:19:10PM +0530, Anup Patel wrote:
> > On Mon, Jul 20, 2026 at 3:40 PM Yu-Chien Peter Lin <peter.lin at sifive.com> wrote:
> > >
> > > Hi Anup and Raymond,
> > >
> > > On Fri, Jul 17, 2026 at 01:14:12PM +0530, Anup Patel wrote:
> > > > There are various ID configuration ISA extensions such as RISC-V Worlds,
> > > > Supervisor domain ID, QoS ID, etc which need to be re-configured upon
> > > > domain context switch on a hart. Extend hart protection abstraction to
> > > > support both memory protection and ID configuration ISA extensions.
> > >
> > > I will rework my "[RFC,00/12] Add RISC-V Worlds ISA support to OpenSBI"
> > > series based on this series.
> >
> > Thanks Peter!
> >
> > The RISC-V world DT property parsing can be done at the
> > time of parsing domains from FDT (like already done in this
> > series). The hart platform abstraction will be only for configuring
> > the RISC-V world CSRs.
> >
> > May be you can add sbi_hart_worlds.c (just like sbi_hart_pmp.c) ?
> >
> > For RISC-V world checkers, we can have separate FDT based driver
> > under lib/utils which will register sbi_domain_data instance upon probe
> > of each RISC-V world checker instance. The domain_setup() and
> > domain_cleanup() callbacks of the RISC-V worlds checker driver will
> > program the RISC-V world checker MMIO registers.
>
> I expect the wgChecker driver will be implemented as a global FDT
> driver, not bound to domain state. A checker slot table contains
> all WIDs' permissions simultaneously and remains unchanged during
> runtime (might be locked). So I think we should not add wgChecker
> domain_setup/domain_cleanup callbacks.
>

I agree with your suggestion. The wgChecker is quite simple and
does not need any in-memory data structures to be allocated (like
IO-MPT) so a normal FDT driver probed via platform final init should
be fine.

Regards,
Anup