Re: [PATCH 3/3] target/nios2: Use gen_io_start around wrctl instruction

[Peter Maydell <peter.maydell@linaro.org>]

On Sun, 5 Jul 2020 at 21:54, Max Filippov <jcmvbkbc@gmail.com> wrote:
>
> On Sun, Jul 5, 2020 at 11:16 AM Max Filippov <jcmvbkbc@gmail.com> wrote:
> > On Sun, Jul 5, 2020 at 10:09 AM Peter Maydell <peter.maydell@linaro.org> wrote:
> > > On Thu, 2 Jul 2020 at 19:53, Richard Henderson
> > > <richard.henderson@linaro.org> wrote:
> > > > This isn't right.  Not so much the gen_io_start portion, but the entire
> > > > existence of helper_check_interrupt.
> > > I agree that it looks bogus (xtensa has a similar helper as well, incidentally),
> > I think there was a reason for it.
>
> ...and the reason is that this helper calls cpu_[re]set_interrupt
> to update CPU_INTERRUPT_HARD, which makes exit to the
> main CPU loop do something to handle IRQ.
> Maybe 'check_interrupt' is not a good name for that, but the
> action taken there seems right to me.

Usually I would expect that CPU_INTERRUPT_HARD would be
set by whatever was setting the interrupt (often a handler
for an inbound qemu_irq line to the CPU). Then the
cpu_exec_interrupt hook only actually does something if
both the INTERRUPT_HARD bit is set and the CPU register
says "and interrupts are unmasked". If you have a design
like that then "unmasking interrupts just means going out
to the main loop" will work.

thanks
-- PMM