Re: NetBSD dom0 PVH: hardware interrupts stalls

[Manuel Bouyer <bouyer@xxxxxxxxxxxxxxx>]

On Thu, Nov 19, 2020 at 04:57:18PM +0100, Manuel Bouyer wrote:
> On Thu, Nov 19, 2020 at 03:19:15PM +0100, Roger Pau Monné wrote:
> > I've got two different debug patches for you to try. I'm attaching both
> > to this email but I think we should start with Jan's suggestion
> > (conditional_print.patch). That patch will only print extra messages
> > between the ioregsel 3 ... ioregsel f existing debug messages, you
> > will have to trigger this from NetBSD by using ioapic_dump_raw AFAICT.
> 
> thanks. I didn't see any change in behavior or XEN output with this
> patch (especially the vioapic_deliver string doesn't show up in the
> logs).

I tried forcing print to 1, and I still don't see "vioapic_deliver" on the
console. I changed the patch to:
#define vioapic_deliver(vioapic, irq) ({\ 
    if ( /* print && irq == 34 */ 1 ) \
        printk("%s:%d:%s: vioapic_deliver %d\n", __FILE__, __LINE__, __func__, i
rq); \
    _vioapic_deliver(vioapic, irq); })

and got:
[  13.8853432] ioapic2: pin2 0x0000a067 0x00000000^M
[  13.8853432] mfii0: cmd timeout ccb 0xffff9780023b7d60 status 0x40000008^M
(XEN) *** Serial input to Xen (type 'CTRL-a' three times to switch input)
(XEN) vioapic.c:511:vioapic_irq_positive_edge: vioapic_deliver 2
[  17.0001093] mfii0: cmd aborted ccb 0xffff9780023b7d60^M
(XEN) vioapic.c:511:vioapic_irq_positive_edge: vioapic_deliver 2
[  17.0217772] config_pending_decr: scsibus0 0^M
(XEN) vioapic.c:511:vioapic_irq_positive_edge: vioapic_deliver 2
[  17.(XEN) vioapic.c:511:vioapic_irq_positive_edge: vioapic_deliver 2
0417095] config_finalize 2185^M

So I guess that the interrupt delivery on XEN output is
vioapic.c:511

-- 
Manuel Bouyer <bouyer@xxxxxxxxxxxxxxx>
     NetBSD: 26 ans d'experience feront toujours la difference
--