On Thu, 24 Feb 2022 at 23:56, Richard Henderson
<richard.hender...@linaro.org> wrote:
>
> On 2/24/22 03:48, Amir Gonnen wrote:
> > +static void nios2_cpu_set_eic_irq(Nios2CPU *cpu, int level)
> > +{
> > + CPUNios2State *env = &cpu->env;
> > + CPUState *cs = CPU(cpu);
> > +
> > + env->irq_pending = level;
> > +
> > + if (env->irq_pending && nios2_take_eic_irq(cpu)) {
> > + env->irq_pending = 0;
> > + cpu_interrupt(cs, CPU_INTERRUPT_HARD);
> > + } else if (!env->irq_pending) {
> > + cpu_reset_interrupt(cs, CPU_INTERRUPT_HARD);
> > + }
> > +}
>
> This looks wrong. Of course, so does nios2_cpu_set_irq, from which you've
> cribbed this.
Bit of a thread from 2020 on nios2's odd interrupt handling;
https://lore.kernel.org/qemu-devel/3260735e-05ab-2d42-f7e4-914ad804f...@linaro.org/
though it's mostly just you saying the same things you're saying now :-)
xtensa's also weird in a similar way.
-- PMM
