Re: NetBSD dom0 PVH: hardware interrupts stalls

[Roger Pau Monné <roger.pau@xxxxxxxxxx>]

On Thu, Nov 26, 2020 at 03:16:08PM +0100, Manuel Bouyer wrote:
> On Thu, Nov 26, 2020 at 02:34:44PM +0100, Roger Pau Monné wrote:
> > On Tue, Nov 24, 2020 at 05:09:14PM +0100, Manuel Bouyer wrote:
> > > On Tue, Nov 24, 2020 at 04:49:17PM +0100, Roger Pau Monné wrote:
> > > > Could you also give a try with ioapic_ack=new on the Xen command line?
> > > 
> > > With this I still have the interrupt issue, but Xen doesn't panic on 'i'.
> > > http://www-soc.lip6.fr/~bouyer/xen-log8.txt
> > 
> > Sorry for the delay, I have yet another debug patch for you to try.
> > Can you remove the ioapic_ack=new from the command line and rebuild
> > the hypervisor with the provided patch applied and debug trace
> > enabled? (`gmake -C xen menuconfig` and go into Debugging Options to
> > find it).
> 
> menuconfig doens't build on NetBSD, I set CONFIG_DEBUG_TRACE=y in
> .config. I guess it is enough ?
> 
> For the record, my boot commad line is now
> menu=Boot Xen PVH:load /test console=com0 root=dk0 -vx; multiboot 
> /xen-test.gz dom0_mem=1024M console=com2 com2=57600,8n1,,0 loglvl=all 
> guest_loglvl=all gnttab_max_nr_frames=64 dom0=pvh iommu=debug dom0_vcpus_pin 
> sync_console dom0_max_vcpus=1 watchdog=force iommu=no-intremap
> 
> 
> > 
> > Then once the system stalls use the 'T' debug key to dump the buffer.
> 
> Here it is. It seems to be stuck in an infinite loop, I hit the 'R' key
> after several minutes
> http://www-soc.lip6.fr/~bouyer/xen-log9.txt

Oh, that's actually very useful. The interrupt is being constantly
injected from the hardware and received by Xen, it's just not then
injected into dom0 - that's the bit we are missing. Let me look into
adding some more debug to that path, hopefully it will tell us where
things are getting blocked.

Roger.