| Message ID | 161537170060.6654.9430112746749476215.stgit@pasha-ThinkPad-X280 |
|---|---|
| State | New |
| Headers | show |
| Series | hw/i8254: fix vmstate load | expand |
On 10/03/21 11:21, Pavel Dovgalyuk wrote: > QEMU timer of channel 0 in i8254 is used to raise irq > at the specified moment of time. This irq can be disabled > with irq_disabled flag. But when vmstate of the pit is > loaded, timer may be rearmed despite the disabled interrupts. > This patch adds irq_disabled flag check to fix that. > > Signed-off-by: Pavel Dovgalyuk <Pavel.Dovgalyuk@ispras.ru> > --- > hw/timer/i8254.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/hw/timer/i8254.c b/hw/timer/i8254.c > index c01ee2c72a..c8388ea432 100644 > --- a/hw/timer/i8254.c > +++ b/hw/timer/i8254.c > @@ -324,7 +324,7 @@ static void pit_post_load(PITCommonState *s) > { > PITChannelState *sc = &s->channels[0]; > > - if (sc->next_transition_time != -1) { > + if (sc->next_transition_time != -1 && !sc->irq_disabled) { > timer_mod(sc->irq_timer, sc->next_transition_time); > } else { > timer_del(sc->irq_timer); > Queued, thanks. Paolo
* Pavel Dovgalyuk (pavel.dovgalyuk@ispras.ru) wrote: > QEMU timer of channel 0 in i8254 is used to raise irq > at the specified moment of time. This irq can be disabled > with irq_disabled flag. But when vmstate of the pit is > loaded, timer may be rearmed despite the disabled interrupts. > This patch adds irq_disabled flag check to fix that. > > Signed-off-by: Pavel Dovgalyuk <Pavel.Dovgalyuk@ispras.ru> Hi Pavel, I'm curious, did you see this cause a problem on a particular guest OS? Dave > --- > hw/timer/i8254.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/hw/timer/i8254.c b/hw/timer/i8254.c > index c01ee2c72a..c8388ea432 100644 > --- a/hw/timer/i8254.c > +++ b/hw/timer/i8254.c > @@ -324,7 +324,7 @@ static void pit_post_load(PITCommonState *s) > { > PITChannelState *sc = &s->channels[0]; > > - if (sc->next_transition_time != -1) { > + if (sc->next_transition_time != -1 && !sc->irq_disabled) { > timer_mod(sc->irq_timer, sc->next_transition_time); > } else { > timer_del(sc->irq_timer); > >
On 15.03.2021 23:13, Dr. David Alan Gilbert wrote: > * Pavel Dovgalyuk (pavel.dovgalyuk@ispras.ru) wrote: >> QEMU timer of channel 0 in i8254 is used to raise irq >> at the specified moment of time. This irq can be disabled >> with irq_disabled flag. But when vmstate of the pit is >> loaded, timer may be rearmed despite the disabled interrupts. >> This patch adds irq_disabled flag check to fix that. >> >> Signed-off-by: Pavel Dovgalyuk <Pavel.Dovgalyuk@ispras.ru> > > Hi Pavel, > I'm curious, did you see this cause a problem on a particular guest > OS? That was Windows 7 on i386. I found this when tested reverse debugging. > > Dave > >> --- >> hw/timer/i8254.c | 2 +- >> 1 file changed, 1 insertion(+), 1 deletion(-) >> >> diff --git a/hw/timer/i8254.c b/hw/timer/i8254.c >> index c01ee2c72a..c8388ea432 100644 >> --- a/hw/timer/i8254.c >> +++ b/hw/timer/i8254.c >> @@ -324,7 +324,7 @@ static void pit_post_load(PITCommonState *s) >> { >> PITChannelState *sc = &s->channels[0]; >> >> - if (sc->next_transition_time != -1) { >> + if (sc->next_transition_time != -1 && !sc->irq_disabled) { >> timer_mod(sc->irq_timer, sc->next_transition_time); >> } else { >> timer_del(sc->irq_timer); >> >>
* Pavel Dovgalyuk (pavel.dovgalyuk@ispras.ru) wrote: > On 15.03.2021 23:13, Dr. David Alan Gilbert wrote: > > * Pavel Dovgalyuk (pavel.dovgalyuk@ispras.ru) wrote: > > > QEMU timer of channel 0 in i8254 is used to raise irq > > > at the specified moment of time. This irq can be disabled > > > with irq_disabled flag. But when vmstate of the pit is > > > loaded, timer may be rearmed despite the disabled interrupts. > > > This patch adds irq_disabled flag check to fix that. > > > > > > Signed-off-by: Pavel Dovgalyuk <Pavel.Dovgalyuk@ispras.ru> > > > > Hi Pavel, > > I'm curious, did you see this cause a problem on a particular guest > > OS? > > That was Windows 7 on i386. > I found this when tested reverse debugging. Thanks; I like to know of migration fixes for when someone comes asking why something obscure breaks under migration :-) Dave > > > > Dave > > > > > --- > > > hw/timer/i8254.c | 2 +- > > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > > > diff --git a/hw/timer/i8254.c b/hw/timer/i8254.c > > > index c01ee2c72a..c8388ea432 100644 > > > --- a/hw/timer/i8254.c > > > +++ b/hw/timer/i8254.c > > > @@ -324,7 +324,7 @@ static void pit_post_load(PITCommonState *s) > > > { > > > PITChannelState *sc = &s->channels[0]; > > > - if (sc->next_transition_time != -1) { > > > + if (sc->next_transition_time != -1 && !sc->irq_disabled) { > > > timer_mod(sc->irq_timer, sc->next_transition_time); > > > } else { > > > timer_del(sc->irq_timer); > > > > > > >
diff --git a/hw/timer/i8254.c b/hw/timer/i8254.c index c01ee2c72a..c8388ea432 100644 --- a/hw/timer/i8254.c +++ b/hw/timer/i8254.c @@ -324,7 +324,7 @@ static void pit_post_load(PITCommonState *s) { PITChannelState *sc = &s->channels[0]; - if (sc->next_transition_time != -1) { + if (sc->next_transition_time != -1 && !sc->irq_disabled) { timer_mod(sc->irq_timer, sc->next_transition_time); } else { timer_del(sc->irq_timer);
QEMU timer of channel 0 in i8254 is used to raise irq at the specified moment of time. This irq can be disabled with irq_disabled flag. But when vmstate of the pit is loaded, timer may be rearmed despite the disabled interrupts. This patch adds irq_disabled flag check to fix that. Signed-off-by: Pavel Dovgalyuk <Pavel.Dovgalyuk@ispras.ru> --- hw/timer/i8254.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)