[Pkg-xen-devel] Bug#988333: : VGA Intel IGD Passthrough to Debian Xen HVM DomUs not working

[Chuck Zmudzinski]

On 10/23/2021 11:11 AM, Hans van Kranenburg wrote:
> Hi!
>
> On 10/19/21 5:44 AM, Chuck Zmudzinski wrote:
>
>> I will continue to investigate this issue and try to bisect the problem
>> as it recurs in Dom0 for some version of Xen > 4.4 and <= 4.14. It
>> will obviously take some time since there are so many differences
>> between Xen 4.4 and 4.14.
> If you can make progress on that, and find an actual commit that changes
> the behavior, then we're probably at 95% towards finding a cause and
> solution. :) That'd be great.

In more recent tests, I have identified the regression occurs in Dom0
for some version of Xen > 4.4 and < Xen 4.5.0. In these old versions
of Xen, only the qemu traditional device model supports IGD
passthrough, so in these tests I am investigating the regression
that causes the crash in the Intel i915 drm kernel module in the
HVM, rather than the regression that causes problems with
assigning IRQs correctly, which as discussed earlier occurs only
with the upstream device model [1]. I am continuing to narrow down
the problem with the Intel i915 kernel module crash since now I
have started to build and test the Xen 4.5-unstable packages for
jessie with the traditional qemu device model. When the upstream
Xen 4.5 branch was first created, the bug was not present yet since
that was essentially a 4.4 release candidate, but it was present by
the time Xen 4.5.0 was released. I will hopefully be able to find the
commit along the development of the Xen 4.5 branch that
introduced the crash in the i915 drm kernel driver in the bullseye
HVM DomU. It will take some time because there are hundreds
of commits from the start of the Xen 4.5 branch to the release of
Xen 4.5.0, and since these are unstable pre-release candidate builds
of upstream Xen 4.5, it will not always be easy or even possible to
build and test them all.

If we can identify the cause of the i915 crash in Xen 4.5, then we can
decide if it is possible to fix it in modern versions of Xen.

>
> A possible time-saver that I can recommend is to send a post to the
> upstream xen-users list [0] about this already. Like "Hi all, I'm
> starting a HVM Linux domU with Linux 5.10.70 on a Xen 4.14.3 system with
> also 5.10.70 dom0 kernel, with this and this domU config file. It fails
> to start, this is the xl -vvv create output, and this error (the irq
> stuff) appears in the dom0 kernel log.". Try to keep it simple and not
> too long initially, without the surrounding stories, to increase chance
> of it being fully read.

I did send a message to the list last October [2], but did not get a
helpful reply, so that is why I am bisecting this problem myself.
If we can fix this problem with the Intel i915 drm kernel module, then
it is likely another bug against src:qemu (instead of src:xen) to fix
the problem with handling IRQs properly with the bullseye upstream
qemu device model will need to be opened and investigated to
allow IGD passthrough to work using an unmodified bullseye Dom0,
but I will only do that if we can fix the bug here in src:xen first.

> Have fun,
> Hans

I am having fun! Thanks,

Chuck

>
> [0] https://lists.xenproject.org/mailman/listinfo/xen-users
>

[1] https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=988333#36
[2] 
https://lists.xenproject.org/archives/html/xen-users/2021-10/msg00041.html