[PATCH v10 kvmtool 5/8] riscv: Add PLIC device emulation

[Anup Patel]

On Thu, Nov 18, 2021 at 3:43 PM Marc Zyngier <maz at kernel.org> wrote:
>
> On Tue, 16 Nov 2021 05:21:27 +0000,
> Anup Patel <anup.patel at wdc.com> wrote:
> >
> > The PLIC (platform level interrupt controller) manages peripheral
> > interrupts in RISC-V world. The per-CPU interrupts are managed
> > using CPU CSRs hence virtualized in-kernel by KVM RISC-V.
> >
> > This patch adds PLIC device emulation for KVMTOOL RISC-V.
> >
> > Signed-off-by: Vincent Chen <vincent.chen at sifive.com>
> > [For PLIC context CLAIM register emulation]
> > Signed-off-by: Anup Patel <anup.patel at wdc.com>
> > ---
> >  Makefile                     |   1 +
> >  riscv/include/kvm/kvm-arch.h |   2 +
> >  riscv/irq.c                  |   4 +-
> >  riscv/plic.c                 | 518 +++++++++++++++++++++++++++++++++++
> >  4 files changed, 523 insertions(+), 2 deletions(-)
> >  create mode 100644 riscv/plic.c
> >
>
> [...]
>
> > +static void plic__context_write(struct plic_state *s,
> > +                             struct plic_context *c,
> > +                             u64 offset, void *data)
> > +{
> > +     u32 val;
> > +     bool irq_update = false;
> > +
> > +     mutex_lock(&c->irq_lock);
> > +
> > +     switch (offset) {
> > +     case CONTEXT_THRESHOLD:
> > +             val = ioport__read32(data);
> > +             val &= ((1 << PRIORITY_PER_ID) - 1);
> > +             if (val <= s->max_prio)
> > +                     c->irq_priority_threshold = val;
> > +             else
> > +                     irq_update = true;
> > +             break;
> > +     case CONTEXT_CLAIM:
> > +             val = ioport__read32(data);
> > +             if (val < plic.num_irq) {
> > +                     c->irq_claimed[val / 32] &= ~(1 << (val % 32));
> > +                     irq_update = true;
> > +             }
>
> This seems to ignore the nasty bit of the PLIC spec where a write to
> CLAIM is ignored if the interrupt is masked.

Sure, I will update the CLAIM write emulation to ignore
masked interrupts.

Thanks for catching.

Regards,
Anup

>
>         M.
>
> --
> Without deviation from the norm, progress is not possible.