[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-users] Intel HD4000 IGD pass through appears to work, but monitor complains about 'no signal'



Update:
I'm now able to install debian on domU.
The issue I mentioned before is due to a silly mistake of incorrect nfs permission setting.
Basically this time I can see linux desktop on TV panel. However, it is still not fully working.
Even 2D looks perfect, 3D acceleration is not working at all
For 3D, glxinfo report looks good, but glxigears does not render anything.
Only once among many times I can see a set of static gears instead of running ones.
I also tried xbmc, a gpu accelerated program and basically it doesn't work.
But frankly I could still switch to console (KMS) even when the fullscreen 3D app appears to be hanging.

And this time I can't have the audio working (even it did work for the openelec os img).

There seems to be some problems about interrupt routing and lead to missing interrupts according to logs:

Linux debvm 3.2.0-4-amd64 #1 SMP Debian 3.2.32-1 x86_64 GNU/Linux
Dec 2 21:52:44 debvm kernel: [ 1085.604071] [drm:i915_hangcheck_ring_idle] *ERROR* Hangcheck timer elapsed... blt ring idle [waiting on 3354, at 3354], missed IRQ?
Dec 2 21:56:50 debvm kernel: [ 1332.076071] [drm:i915_hangcheck_ring_idle] *ERROR* Hangcheck timer elapsed... render ring idle [waiting on 11297, at 11297], missed IRQ?
Dec 2 22:32:48 debvm kernel: [ 7.220073] hda-intel: azx_get_response timeout, switching to polling mode: last cmd=0x000f0000
Dec 2 22:45:32 debvm kernel: [ 776.392084] hda-intel: No response from codec, disabling MSI: last cmd=0x002f0600
Dec 2 22:45:33 debvm kernel: [ 777.400075] hda_intel: azx_get_response timeout, switching to single_cmd mode: last cmd=0x002f0600

Log from xen dmesg:
(XEN) Intel VT-d supported page sizes: 4kB.
(XEN) Intel VT-d supported page sizes: 4kB.
(XEN) Intel VT-d Snoop Control not enabled.
(XEN) Intel VT-d Dom0 DMA Passthrough not enabled.
(XEN) Intel VT-d Queued Invalidation enabled.
(XEN) Intel VT-d Interrupt Remapping enabled.
(XEN) Intel VT-d Shared EPT tables not enabled.
(XEN) I/O virtualisation enabled
(XEN)Â - Dom0 mode: Relaxed
(XEN) Enabled directed EOI with ioapic_ack_old on!
(XEN) ENABLING IO-APIC IRQs
(XEN)Â -> Using old ACK method
(XEN) Platform timer appears to have unexpectedly wrapped 1 times.
(XEN) Platform timer is 14.318MHz HPET
(XEN) Allocated console ring of 16 KiB.
(XEN) VMX: Supported advanced features:
(XEN)Â - APIC MMIO access virtualisation
(XEN)Â - APIC TPR shadow
(XEN)Â - Extended Page Tables (EPT)
(XEN)Â - Virtual-Processor Identifiers (VPID)
(XEN)Â - Virtual NMI
(XEN)Â - MSR direct-access bitmap
(XEN)Â - Unrestricted Guest
(XEN) HVM: ASIDs enabled.
(XEN) HVM: VMX enabled
(XEN) HVM: Hardware Assisted Paging (HAP) detected
(XEN) HVM: HAP page sizes: 4kB, 2MB
(XEN) printk: 1 messages suppressed.
(XEN) vmsi.c:122:d32767 Unsupported delivery mode 3
(XEN) vmsi.c:122:d32767 Unsupported delivery mode 3
(XEN) vmsi.c:122:d32767 Unsupported delivery mode 3
(XEN) physdev.c:186: dom17: 16:-1 already mapped to 16
(XEN) vmsi.c:122:d32767 Unsupported delivery mode 3
(XEN) vmsi.c:122:d17 Unsupported delivery mode 3

qemu-dm log:

pt_iomem_map: e_phys=ffffffff maddr=f7c18000 type=0 len=4096 index=0 first_map=0
pt_pci_write_config: Warning: Guest attempt to set address to unused Base Address Register. [00:06.0][Offset:30h][Length:4]
pt_iomem_map: e_phys=f1424000 maddr=f7c18000 type=0 len=4096 index=0 first_map=0
igd_pci_read: pci_config_read: 0:0.0: addr=52 len=2 val=ffff0000
pt_msgctrl_reg_write: guest enabling MSI, disable MSI-INTx translation
pci_intx: intx=1
pt_msi_update: Update msi with pirq 36 gvec 0 gflags 3034
pt_msgctrl_reg_write: guest enabling MSI, disable MSI-INTx translation
pci_intx: intx=1
pt_msi_update: Update msi with pirq 37 gvec 0 gflags 3033
igd_pci_read: pci_config_read: 0:0.0: addr=52 len=2 val=ffff0000
igd_pci_read: pci_config_read: 0:0.0: addr=0 len=4 val=1508086



On Sun, Dec 2, 2012 at 12:46 PM, G.R. <firemeteor@xxxxxxxxxxxxxxxxxxxxx> wrote:
I had been trying to install win7 && debian as domU these days and the result is frustrating.
For win7, I was able to install and got video output on TV.
But it simply BSOD once intel driver is installed.
The BSOD says something about SYSTEM_SERVICE_EXCEPTION. not sure what does it means.

Also, the TV screen starts to flashing with random colors once I get the win7 VM reboot.
I need to assign the VGA controller back to host once and then to pciback again to have it back to work.
Quite annoying.


For debian, I tried both debian 6 / Wheezy installer. Both of them hang when trying to partition the disk xvda.
I tried to manually use fdisk && dd to access the device and it seems that the access never got response.



On Thu, Nov 29, 2012 at 11:26 PM, Ricardo Jesus <ricardo.meb.jesus@xxxxxxxxx> wrote:
Oh... DQ77MK has DVI-I, DVI-D and HDMI so it's a little different.

The fact that Win7 installer outputs makes me further suggest that you try another Linux domU.

When I get home I'll try to assign the USB 3.0 controller to a domU and let you know.


On Thu, Nov 29, 2012 at 3:11 PM, G.R. <firemeteor@xxxxxxxxxxxxxxxxxxxxx> wrote:
Update: DVI output does not work either for the openelec 2.0 domU.
The intel HD4000 seems does not have internal DVI output.
The xrandr shows that the on-board DVI port is converted from an internal hdmi output (hdmi1 this time).
So I won't surprise that it does not make any difference.

However, another experiment brings some positive result -- the win7 installer actually showed some output on TV.
I'll try to finish the install and see what will happen with an Intel driver.



On Thu, Nov 29, 2012 at 10:07 PM, Ricardo Jesus <ricardo.meb.jesus@xxxxxxxxx> wrote:
Haven't tried to passthrough the USB 3.0 controller both the 2 USB 2.0 controller pass just fine.

Try to setup another Linux domU for example Wheezy or Ubuntu 12.04 LTS. Those are guaranteed to worked.


On Thu, Nov 29, 2012 at 1:09 PM, G.R. <firemeteor@xxxxxxxxxxxxxxxxxxxxx> wrote:
>Btw did you try doing IGD passthrough to another Linux domU?
No, this is the first VM so far.

I tried to upgrade the bios and it does not help. Haven't try out DVI port yet. Will do it later.

PS: I can now confirm that the usb ctlr also works -- I can get data by open the device file directly.
However, it only works for USB2 ctlrs (at least before bios upgrade) -- system complaining that it cannot reset the xhci controller.
What about yours?


Here comes the logs, compressed to save traffic:


On Thu, Nov 29, 2012 at 6:52 PM, Ricardo Jesus <ricardo.meb.jesus@xxxxxxxxx> wrote:
/var/log/xen/qemu-dm-<vm-name> and <vm-name>.cfg would be used.

Btw did you try doing IGD passthrough to another Linux domU?

On Thu, Nov 29, 2012 at 10:27 AM, G.R. <firemeteor@xxxxxxxxxxxxxxxxxxxxx> wrote:
> Intel site is kind of confusing about whether H77 is Vt-d capable.
> This page claims that it is not supported by the chipset:
> http://ark.intel.com/products/64018/Intel-BD82H77-PCH
> And this page lists a series of board that support Vt-d, including some H77
> based products.
> http://www.intel.com/support/motherboards/desktop/sb/CS-030922.htm
>
> AsRock has vt-d config in both the bios and their manual.
> And I have dumped the ACPI table from bios and confirmed that it looks fine
> by checking the content against the spec.
>
> So I tend to believe this is actually working, which is partially proved by
> the audio playback experiment I've done.
>
> I'll do more experiments to root cause the issue to night, including the DVI
> port you mentioned.
> And thanks to your advice about bios update. Once I believe I had the newest
> version, but as I just checked on the web site,
> they seem to released a new version for win 8 about two weeks ago :-).
>
> Of course I'm willing to post logs. What kind of logs do you believe to be
> interesting?
> And any specific debug option do I need to enable before dumping the log?
> Instructions needed, especially for XEN related switch -- I'm really a
> newbie here.
>
>
>
>
>
> On Thu, Nov 29, 2012 at 5:59 PM, Ricardo Jesus <ricardo.meb.jesus@xxxxxxxxx>
> wrote:
>>
>> Hi,
>>
>> I am the guy that wrote the HowTo.
>>
>> Are you sure the H77 supports VT-d? At least when I chose the DQ77MK I
>> only recall Q77 as the desktop chipset that supported VT-d though I
>> read somewhere that Intel has since updated the list of VT-d able
>> desktop boards.
>>
>> Also try to use your monitor's DVI input. I say this because I've a
>> BIOS back experienced a weird bug where the motherboard didn't ouput
>> from DVI so I had to use VGA output at the time.
>>
>> You might also consider updating your motherboard's BIOS.
>>
>> I must that say that I've also managed to passthrough the IGD to
>> various Linux domUs (Debian and Ubuntu for instance).
>>
>> Any logs you can send? Post your domU's config so that the list users
>> can have a look.
>>
>> On Thu, Nov 29, 2012 at 9:24 AM, G.R. <firemeteor@xxxxxxxxxxxxxxxxxxxxx>
>> wrote:
>> > I've seen a recent report claiming that it works:
>> >
>> > http://linux-bsd-sharing.blogspot.pt/2012/10/howto-xen-413-windows-8-hvm-domu-with.html
>> > However, in my case, with a slightly different HW configuration the
>> > story is
>> > totally different. :-ï
>> >
>> > My config:
>> > HW: I7-3770 + AsRock H77M-ITX board
>> > Dom0: Debian Wheezy stock xen (v4.1.3) + Debian stock kernel (3.2.x /
>> > 3.6.x)
>> > 64 bit version.
>> >
>> > Note that both of us have only one gpu in the system.
>> > They difference is that the lucky guy has an DQ77MK while mine is an
>> > AsRock.
>> > Another different is that he picks win 8 as domU, while my target is
>> > openelec 2.0, a linux 3.2.x based HTPC distro.
>> >
>> > So basically what I can see is a functional domU, almost, I mean.
>> > It does not freeze / lock up at all.
>> > All the assigned devices (IGD, audio, USB) are recognized and appears to
>> > work.
>> > I can hear sound from aplay, so audio is working.
>> > I can see the mouse I attached through lsusb so it appears to work too
>> > --
>> > but I can't 100% confident since I haven't seen the pointer moving on
>> > screen...
>> > The IGD is shown in lspci properly and the vainfo tool shows that the HW
>> > codec is fine.
>> > The Xorg.0.log shows that intel driver is properly loaded, EDID from
>> > panel
>> > is correctly obtained.
>> > xrandr tool can show the connected output (I've tried both VGA1 /
>> > hdmi3).
>> > It appears that I can switch resolution or turn-off the display through
>> > xrandr without any error reported.
>> > The only *little* :-) issue is that I can't see any output on monitor /
>> > TV.
>> > Both of them complain about no signal and go back to power saving.
>> >
>> > Well, I'm totally confused about the current status I'm at and need your
>> > advice.
>> > Is there any thing I should do to diagnosis and root cause the problem?
>> > I'm new to xen so am not familiar about the xen style of debugging.
>> > Would like to provide any log as you believe beneficial.
>> >
>> > Thanks,
>> > Timothy
>> >
>> > _______________________________________________
>> > Xen-users mailing list
>> > Xen-users@xxxxxxxxxxxxx
>> > http://lists.xen.org/xen-users
>
>






_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxx
http://lists.xen.org/xen-users

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.