Re: [Xen-devel] dom0/pvh: Dom0 PVH with PCI passthrough support status

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

On Mon, Nov 05, 2018 at 04:17:56PM +0200, Alexandru Vasile wrote:
> On Mon, Nov 05, 2018 at 12:57 PM, Wei Liu wrote:
> > On Mon, Nov 05, 2018 at 11:58:09AM +0200, Alexandru Vasile wrote:
> > > Hello,
> > > (XEN) event_channel.c:319:d0v1 EVTCHNOP failure: domain 1, error -22
> > > (XEN) event_channel.c:319:d0v3 EVTCHNOP failure: domain 1, error -22
> > Do you perhaps have more than one xenstored / xenconsoled running?
> 
> The processes listed by 'ps -aux | grep xen' immediately after dom0 boots
> are oxenstored, xenconsoled and xenwatchdogd.
> 
> I did start 'xencommons ' from my install folder due to xl showing the name
> of Dom0 as '(null)' and also because of a difference in output error from xl
> when creating a HVM DomU with passthrough (the output is captured in files
> xl_output_clean[0] and xl_output_xencommons[1]).
> 
> > > (XEN) Assertion 'fdom != p2m->domain' failed at p2m.c:504
> > You can work around this with:
> > 
> > ---8<---
> >  From 9437054299c1d360eb4fedd065d51965e560fc0c Mon Sep 17 00:00:00 2001
> > From: Wei Liu<wei.liu2@xxxxxxxxxx>
> > Date: Fri, 2 Nov 2018 14:55:04 +0000
> > Subject: [PATCH] DROP DONT POST
> > 
> > ---
> >   xen/arch/x86/mm.c | 2 +-
> >   1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/xen/arch/x86/mm.c b/xen/arch/x86/mm.c
> > index f043e43ac7..9c154783e4 100644
> > --- a/xen/arch/x86/mm.c
> > +++ b/xen/arch/x86/mm.c
> > @@ -4443,7 +4443,7 @@ int arch_acquire_resource(struct domain *d, unsigned 
> > int type,
> >       switch ( type )
> >       {
> > -#ifdef CONFIG_HVM
> > +#if 0
> >       case XENMEM_resource_ioreq_server:
> >       {
> >           ioservid_t ioservid = id;
> > -- 2.11.0
> 
> Thank you, after applying this patch dom0 no longer freezes or reboots,
> therefore I was able to capture all the error messages from the 'xl create'
> command [0] [1].

So the error message is:

libxl: error: libxl_qmp.c:334:qmp_handle_error_response: Domain 2:received an 
error message from QMP server: Mapping machine irq 18 to pirq -1 failed: 
Function not implemented

I haven't looked into the code, but I think QEMU is likely trying to
use a PHYSDEVOP_ or some other hypercall that's not been made
available to PVH. If you want to debug this you will have to figure
out which hypercall is being blocked by looking into the QEMU code
and finding where that error message is coming from.

> Another problem that I discovered while investigating pci passthrough is
> that Dom0 randomly freezes. I encountered this problem multiple times, but
> on two occasions I was able to capture the output form the serial console:
> the problem will reset dom0 after 5 seconds [2] or will cause the system to
> never reset [3].
> 
> I later applied the patch from [4] as seeing this is a recurrent problem,
> but I could not reproduce the bug even when stressing the system.

This is currently being looked into by Intel, and seems to be some
kind of hardware bug/errata.

Roger.

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel