Re: [Qemu-devel] [RFC PATCH v2 10/49] rtl8139: adding new fields to vmstate

[Pavel Dovgaluk]

> From: Paolo Bonzini [mailto:address@hidden On Behalf Of Paolo Bonzini
> Il 17/07/2014 13:02, Pavel Dovgalyuk ha scritto:
> > This patch adds virtual clock-dependent timers to VMState to allow correct
> > saving and restoring the state of RTL8139 network controller.
> >
> > Signed-off-by: Pavel Dovgalyuk <address@hidden>
> > ---
> >  hw/net/rtl8139.c |    5 ++++-
> >  1 files changed, 4 insertions(+), 1 deletions(-)
> >
> > diff --git a/hw/net/rtl8139.c b/hw/net/rtl8139.c
> > index 90bc5ec..992caf0 100644
> > --- a/hw/net/rtl8139.c
> > +++ b/hw/net/rtl8139.c
> > @@ -3289,7 +3289,7 @@ static void rtl8139_pre_save(void *opaque)
> >
> >  static const VMStateDescription vmstate_rtl8139 = {
> >      .name = "rtl8139",
> > -    .version_id = 4,
> > +    .version_id = 5,
> >      .minimum_version_id = 3,
> >      .post_load = rtl8139_post_load,
> >      .pre_save  = rtl8139_pre_save,
> > @@ -3363,6 +3363,9 @@ static const VMStateDescription vmstate_rtl8139 = {
> >          VMSTATE_STRUCT(tally_counters, RTL8139State, 0,
> >                         vmstate_tally_counters, RTL8139TallyCounters),
> >
> > +        VMSTATE_TIMER_V(timer, RTL8139State, 5),
> 
> timer need not be migrated, because it is reinstated by rtl8139_post_load.
> 

  That's true for normal execution.
  In replay execution mode post_load can be called before cached virtual clock
values are loaded. This may cause invalid setting of the timer and raising
an IRQ which didn't happen in record mode.
  I will update this patch to fix post_load function and avoid this 
non-deterministic behavior.

Pavel Dovgalyuk