On Mon, 7 Nov 2022 at 16:42, ~axelheider <axelhei...@git.sr.ht> wrote: > > From: Axel Heider <axel.hei...@hensoldt.net> > > Signed-off-by: Axel Heider <axel.hei...@hensoldt.net> > --- > hw/timer/imx_epit.c | 4 ++++ > 1 file changed, 4 insertions(+) > > diff --git a/hw/timer/imx_epit.c b/hw/timer/imx_epit.c > index 2e9dae0bc8..5315d9633e 100644 > --- a/hw/timer/imx_epit.c > +++ b/hw/timer/imx_epit.c > @@ -94,6 +94,10 @@ static void imx_epit_reset(DeviceState *dev) > s->lr = EPIT_TIMER_MAX; > s->cmp = 0; > s->cnt = 0; > + > + /* clear the interrupt */ > + qemu_irq_lower(s->irq); > + > ptimer_transaction_begin(s->timer_cmp); > ptimer_transaction_begin(s->timer_reload); > /* stop both timers */ > --
It's not valid to call qemu_irq_set/qemu_irq_lower from a legacy reset function (because whether it has an effect or not depends on whether the device on the other end of the line gets reset before or after this one, and there is no guaranteed order for devices being reset). The convention is that if the post-reset state of the IRQ line is 0, then you don't do anything. The device on the other end will reset into a state corresponding to "the input line is 0". thanks -- PMM
