RE: [Qemu-devel] qemu/hw piix_pci.c

[Igor Lvovsky]

> -----Original Message-----
> From: Thiemo Seufer [mailto:address@hidden
> Sent: Monday, October 22, 2007 12:16 AM
> To: Avi Kivity
> Cc: address@hidden; Igor Lvovsky
> Subject: Re: [Qemu-devel] qemu/hw piix_pci.c
> 
> Avi Kivity wrote:
> >> Modified files:
> >>    hw             : piix_pci.c
> >> Log message:
> >>    Enable ACPI interrupts, by Igor Lvovsky.
> >>
> >> CVSWeb URLs:
> >>
> http://cvs.savannah.gnu.org/viewcvs/qemu/hw/piix_pci.c?cvsroot=qemu&r1=1.1
> 1&r2=1.12
> >>
> >>
> >
> > IMO this is a bug.  The patch enables the ACPI PCI interrupt every time
> it
> > is asserted.  If this is really needed, this implies that the guest OS
> > detects some sort of inconsistency and keeps turning the ACPI interrupt
> > off.
> >
> > The bug should be addressed instead of worked-around in this way.
> 
> Igor, could you clarify what happens there?

[Igor Lvovsky] Yes sure. I rechecked it and it seems like reasonable conclusion.
As I mentioned before, enable bit of PIRQRCA should be 0, but somehow
Windows set it all the time. One of the reasons can be some sort 
of inconsistency.
My last patch can be temporary workaround and now we can get the ACPI 
interrupts without disrupt anything  else, but I'll try to find full solution 
for this issue.
        Igor
 



> 
> 
> Thiemo