pci passthrough of snd_hda_intel or via usb to oneiric or precise hvm breaks xen pvops drivers

[Mark Hurenkamp]

When starting oneiric as a xen hvm domain (from oneiric), passing through vga, usb & audio, the audio device seems to cause xen to fail pvops initialization.
The xen domu is running oneiric with kernel 3.0.0-12-server, symptoms are the same for 3.0.0-11-server.
My old ubuntu maverick installation works properly on the same host.

The xen config file for the hvm looks like this:
kernel = "/usr/lib/xen/boot/hvmloader"
builder='hvm'
memory = 900
device_model='/usr/lib64/xen/bin/qemu-dm'

name = "yavin"
vcpus=8
vif = ['mac=aa:00:00:00:01:09,bridge=loc']
disk = [
        'phy:/dev/vm/yavin-oneiric,ioemu:hda,w',
        'file:/images/ubuntu-11.10-beta2-server-amd64.iso,hdc:cdrom,r',
]
boot="cd"
sdl=0
vnc=1
vnclisten="0.0.0.0"
vncdisplay=9
vncunused=1
stdvga=0
serial='pty'
monitor=1
pci = [
    '0000:06:00.0', # 06:00.0 usb
    '0000:02:00.0', # 02:00.0 fglrx (tried radeon as well)
    '0000:02:00.1' # 02:00.1 snd_hda_intel
]

Relevant part of the boot messages:
Loading, please wait...
[ 2.401864] udevd[106]: starting version 173
Begin: Loading essential drivers ... [ 2.456569] xen map irq failed -22
[ 2.475372] xen-platform-pci 0000:00:03.0: PCI INT A: failed to register GSI
[ 2.509994] xen-platform-pci: probe of 0000:00:03.0 failed with error -1

The above error results in the xdv* devices not being created, and thus i'm stuck at the initramfs prompt.
Commenting out the audio device from the above configuration makes things work (albeit without sound).

[Mark Hurenkamp (mark-hurenkamp)]

I noticed a similar problem when upgrading another ubuntu-maverick hvm to oneiric.
This time it seems to be due to passing through the uhci & ehci devices.
(i tried to determine wether it is due to a single device, but any one of these seems
to cause problems for me)

Configuration:
===========
kernel = '/usr/lib/xen/boot/hvmloader'
builder = 'hvm'
memory = '1000'
device_model='/usr/lib/xen/bin/qemu-dm'
serial='pty'
disk = [ 'phy:/dev/vm/cularin-oneiric,hda,w' ]
name = 'cularin'
vif = [ 'mac=aa:20:00:00:02:17, bridge=loc' ]
boot='c'
vfb = [ 'vnc=1,vnclisten=0.0.0.0,vncdisplay=7' ]
usb=1
usbdevice='tablet'
monitor=1
pci = [
    '0000:08:01.0', # uhci_hcd: hauppauge pvr 500 #1
    '0000:08:01.2', # ehci_hcd: hauppauge pvr 500 #1
    '0000:08:02.0', # uhci_hcd: hauppauge pvr 500 #2
    '0000:08:02.2', # ehci_hcd: hauppauge pvr 500 #2
]

[Brad Figg (brad-figg)]

This bug is missing log files that will aid in diagnosing the problem. From a terminal window please run:

apport-collect 860175

and then change the status of the bug to 'Confirmed'.

If, due to the nature of the issue you have encountered, you are unable to run this command, please add a comment stating that fact and change the bug status to 'Confirmed'.

This change has been made by an automated script, maintained by the Ubuntu Kernel Team.

[Mark Hurenkamp (mark-hurenkamp)]

I did some more testing, and noticed that adding "acpi=0" to both configurations makes things work.

Running the command from the (guest) system in the 'fault' configuration is not possible due to the nature of the bug, since it prevents the boot from completing. Running the command in a different configuration, will yield unreliable information, therefore i will not do so for now.
If you would like me to run it from the host, or from a guest with a different configuration, then i can still do that, but do let me know which you find most interesting.

For now i'm setting this to 'confirmed' in line with the above comment.

[Brad Figg (brad-figg)]

Thank you for taking the time to file a bug report on this issue.

However, given the number of bugs that the Kernel Team receives during any development cycle it is impossible for us to review them all. Therefore, we occasionally resort to using automated bots to request further testing. This is such a request.

We have noted that there is a newer version of the development kernel than the one you last tested when this issue was found. Please test again with the newer kernel and indicate in the bug if this issue still exists or not.

If the bug still exists, change the bug status from Incomplete to Confirmed. If the bug no longer exists, change the bug status from Incomplete to Fix Released.

Thank you for your help, we really do appreciate it.

[Mark Hurenkamp (mark-hurenkamp)]

Thanks for the heads up. I do understand that you are probably very busy with bug reports and appreciate your efforts in trying to push out a stable linux release free for us all to use. I don't mind trying out any updates that might solve my issue, however, i do not currently see any updates in the kernel available.

My current kernel is:
root@cularin:/etc# uname -a
Linux cularin 3.0.0-12-server #19-Ubuntu SMP Fri Sep 23 23:06:20 UTC 2011 x86_64 x86_64 x86_64 GNU/Linux

If this is not the latest kernel, then do let me know how i can find a more recent one.

[Mark Hurenkamp (mark-hurenkamp)]

update: with kernel 2.6.35-30-server from maverick things work fine, so i'm using that for now.

[Brad Figg (brad-figg)]

Thank you for taking the time to file a bug report on this issue.

However, given the number of bugs that the Kernel Team receives during any development cycle it is impossible for us to review them all. Therefore, we occasionally resort to using automated bots to request further testing. This is such a request.

We have noted that there is a newer version of the development kernel than the one you last tested when this issue was found. Please test again with the newer kernel and indicate in the bug if this issue still exists or not.

If the bug still exists, change the bug status from Incomplete to Confirmed. If the bug no longer exists, change the bug status from Incomplete to Fix Released.

Thank you for your help, we really do appreciate it.

[Mark Hurenkamp (mark-hurenkamp)]

Tested against latest kernel (3.0.0-12.20). Problem still exists.

[Joseph Salisbury (jsalisbury)]

Would it be possible for you to test the latest upstream kernel? It will allow additional upstream developers to examine the issue. Refer to https://wiki.ubuntu.com/KernelMainlineBuilds . If possible, please test the latest release candidate kernel (Not a kernel in the daily directory). Once you've tested the upstream kernel, please remove the 'needs-upstream-testing' tag(Only that one tag, please leave the others). This can be done by clicking on the yellow pencil icon next to the tag located at the bottom of the bug description and deleting the 'needs-upstream-testing' text.

If this bug is fixed by the mainline kernel, please add the following tag 'kernel-fixed-upstream-KERNEL-VERSION'. For example, if kernel version 3.1-rc9 fixed and issue, the tag would be: 'kernel-fixed-upstream-v3.1-rc9'.

If the mainline kernel does not fix this bug, please add the tag: 'kernel-bug-exists-upstream'.

Thanks in advance.

[Mark Hurenkamp (mark-hurenkamp)]

Tested against linux-image-3.1.0-030100rc10-generic_3.1.0-030100rc10.201110200610_amd64.deb as provided by the ppa mentioned on the referred page.
Problem exists in this mainline kernel as well, i've updated the tags to reflect this.

[Joseph Salisbury (jsalisbury)]

This issue appears to be an upstream bug, since you tested the latest upstream kernel. Would it be possible for you to open an upstream bug report at bugzilla.kernel.org [1]? That will allow the upstream Developers to examine the issue, and may provide a quicker resolution to the bug.

If you are comfortable with opening a bug upstream, It would be great if you can report back the upstream bug number in this bug report. That will allow us to link this bug to the upstream report.

[1] https://wiki.ubuntu.com/Bugs/Upstream/kernel

[Mark Hurenkamp (mark-hurenkamp)]

I just went over to bugzilla.kernel.org, but it doesn't seem to be up.
(perhaps it is due to the recent security breach?)
So i can't submit, nor read some of the information referred to by the page in your ref [1].
Will keep an eye on kernel.org for now.

[Mark Hurenkamp (mark-hurenkamp)]

I tested some more kernels from the ppa mentioned on the page referred to by comment #12

The last one to work for me is a 2.6.35 maverick kernel.
Problems start from 2.6.36 onward, i guess that's where the pvops block driver has been included.
(i tested latest versions of 2.6.36, 2.6.37, 2.6.38 and 2.6.39 available in the ppa)

My guess is that pci passthrough with pvops never really worked at all in mainline kernels (for some of my hw that is).

2.6.37 shows something new:
[ 1.490097] ACPI: bus type pci registered
[ 1.492766] PCI: Using configuration type 1 for base access
[ 1.496987] bio: create slab <bio-0> at 0
[ 1.506986] ACPI: Interpreter enabled
[ 1.510008] ACPI: (supports S0 S3 S4 S5)
[ 1.514761] ACPI: Using IOAPIC for interrupt routing
[ 1.540064] ACPI Exception: AE_NOT_FOUND, Evaluating _PRW (20101013/scan-723)

[Mark Hurenkamp (mark-hurenkamp)]

Meanwhile i did post a message on linux-kernel and xen-devel (although i did not see it appear on the latter), refering to this problem report. So far i've seen no responses, and bugzilla.kernel.org still seems to be down.
(see http://article.gmane.org/gmane.linux.kernel/1207316 )

[Mark Hurenkamp (mark-hurenkamp)]

Did some tests with precise pangolin alpha-1, result is the same as above.
Used kernel:

# uname -a
Linux hoth 3.2.0-2-generic #5-Ubuntu SMP Mon Nov 28 18:10:23 UTC 2011 x86_64 x86_64 x86_64 GNU/Linux

[Mark Hurenkamp (mark-hurenkamp)]

I upgraded the precise kernel on one of my virtual machines recently (to 3.2.0-17-generic) and it seems to solve this issue at least for the snd_hda_intel case on that vm. Still need to check out if it fixes the hauppauge t500 problems.

[Mark Hurenkamp (mark-hurenkamp)]

I did a complete update of the vm, so it might be due to an updated xen hypervisor or xen utils as well.

[Joseph Salisbury (jsalisbury)]

@Mark

Thanks for the update. Does the latest precise kernel resolve this bug? If so, can it be marked as "Fix Released"?

[Mark Hurenkamp (mark-hurenkamp)]

I upgraded my other VM with the two Hauppauge T500 cards to precise, and that too seems to work fine with the current precise kernel (no missing devices, pvops works fine).
So i can confirm that at in the configuration precise host with precise vm, this issue is solved now.

As far as I am concerned you can mark this as "Fix Released".

[Joseph Salisbury (jsalisbury)]

Thanks, Mark.