UHD 750 Passthrough Android TV VM

[BeardElk]

1 minute ago, SimonF said:

That was just my path when i ran the command. Can you post diags?

 

do you intel gpu top loaded in plugins?

I had it installed but removed it and rebooted before blacklisting. 

So not anymore!

belk-diagnostics-20220131-2106.zip

[SimonF]

I cannot see the igpu in the vfio bindings have you checked the check box in system devices.

 

Loading config from /boot/config/vfio-pci.cfg
BIND=0000:01:00.0|10de:2182 0000:01:00.1|10de:1aeb 0000:01:00.2|10de:1aec 0000:01:00.3|10de:1aed

[BeardElk]

37 minutes ago, SimonF said:

I cannot see the igpu in the vfio bindings have you checked the check box in system devices.

 

Loading config from /boot/config/vfio-pci.cfg
BIND=0000:01:00.0|10de:2182 0000:01:00.1|10de:1aeb 0000:01:00.2|10de:1aec 0000:01:00.3|10de:1aed

Sorry, I removed it after i got this: 

root@BelK:~#  lspci -Dvs 00:02.0
0000:00:02.0 VGA compatible controller: Intel Corporation RocketLake-S GT1 [UHD Graphics 750] (rev 04) (prog-if 00 [VGA controller])
        DeviceName: Onboard - Video
        Subsystem: Gigabyte Technology Co., Ltd Device d000
        Flags: fast devsel, IRQ 16, IOMMU group 2
        Memory at 82000000 (64-bit, non-prefetchable) [size=16M]
        Memory at 40000000 (64-bit, prefetchable) [size=256M]
        I/O ports at 6000 [size=64]
        Capabilities: [40] Vendor Specific Information: Len=0c <?>
        Capabilities: [70] Express Root Complex Integrated Endpoint, MSI 00
        Capabilities: [ac] MSI: Enable- Count=1/1 Maskable+ 64bit-
        Capabilities: [d0] Power Management version 2
        Capabilities: [100] Process Address Space ID (PASID)
        Capabilities: [200] Address Translation Service (ATS)
        Capabilities: [300] Page Request Interface (PRI)
        Kernel driver in use: vfio-pci
        Kernel modules: i915

 

But now i´ve bound them again.

belk-diagnostics-20220201-0906.zip

[SimonF]

10 hours ago, BeardElk said:

Sorry, I removed it after i got this: 

root@BelK:~#  lspci -Dvs 00:02.0
0000:00:02.0 VGA compatible controller: Intel Corporation RocketLake-S GT1 [UHD Graphics 750] (rev 04) (prog-if 00 [VGA controller])
        DeviceName: Onboard - Video
        Subsystem: Gigabyte Technology Co., Ltd Device d000
        Flags: fast devsel, IRQ 16, IOMMU group 2
        Memory at 82000000 (64-bit, non-prefetchable) [size=16M]
        Memory at 40000000 (64-bit, prefetchable) [size=256M]
        I/O ports at 6000 [size=64]
        Capabilities: [40] Vendor Specific Information: Len=0c <?>
        Capabilities: [70] Express Root Complex Integrated Endpoint, MSI 00
        Capabilities: [ac] MSI: Enable- Count=1/1 Maskable+ 64bit-
        Capabilities: [d0] Power Management version 2
        Capabilities: [100] Process Address Space ID (PASID)
        Capabilities: [200] Address Translation Service (ATS)
        Capabilities: [300] Page Request Interface (PRI)
        Kernel driver in use: vfio-pci
        Kernel modules: i915

 

But now i´ve bound them again.

belk-diagnostics-20220201-0906.zip 129.4 kB · 1 download

Which vm? 

[BeardElk]

33 minutes ago, SimonF said:

Which vm? 

 

Any VM. Linux/android tv initially but any VM at this point. 

[SimonF]

40 minutes ago, BeardElk said:

 

Any VM. Linux/android tv initially but any VM at this point. 

When does the error appear when you try to save with igpu selected?

[BeardElk]

Just now, SimonF said:

When does the error appear when you try to save with igpu selected?

Both windows and Linux/android tv. Every VM I've tried i get "attempt at dubble bla bla".

[SimonF]

Have you tried i440fx as the machine type rather than Q35?

[BeardElk]

20 hours ago, SimonF said:

Have you tried i440fx as the machine type rather than Q35?

Tried it now, no change. 
If I edit any existing vm I get "XML error: Attempted double use of PCI Address 0000:00:02.0" but if I take the same vdisk and create a "new" vm with exactly the same everything, except from setting Intel uhd 750 as vga and audio, I´m able to create the vm with the gpu "connected", but as soon as I try and start it the screen just goes blank and I get "no input". 

 

I´ve tried creating a new vm with manjaro (had luck with manjaro previously) but same thing there. If I edit the vm "attempted double" and if I remove it and create a new one with uhd 750 as vga/audio from start, i´m able to start it but no output. 

Edited February 2, 2022 by BeardElk

[SimonF]

17 hours ago, BeardElk said:

Tried it now, no change. 
If I edit any existing vm I get "XML error: Attempted double use of PCI Address 0000:00:02.0" but if I take the same vdisk and create a "new" vm with exactly the same everything, except from setting Intel uhd 750 as vga and audio, I´m able to create the vm with the gpu "connected", but as soon as I try and start it the screen just goes blank and I get "no input". 

 

I´ve tried creating a new vm with manjaro (had luck with manjaro previously) but same thing there. If I edit the vm "attempted double" and if I remove it and create a new one with uhd 750 as vga/audio from start, i´m able to start it but no output. 

I have been reading another thread, It may need seabios but I cannot get my Alderlake system to work I get 

 

Feb  2 17:05:08 computenode usb_manager: Info: rc.usb_manager Autoconnect Start SIGMACHIP_USB_Keyboard vm: Windows 10 Device in Use action ignored. 
Feb  2 17:05:09 computenode kernel: vfio-pci 0000:00:02.0: vfio_ecap_init: hiding ecap 0x1b@0x100
Feb  2 17:05:09 computenode kernel: vfio-pci 0000:00:02.0: Invalid PCI ROM header signature: expecting 0xaa55, got 0xffb6

 

other thread for info.

 

 

[BeardElk]

7 hours ago, SimonF said:

I have been reading another thread, It may need seabios but I cannot get my Alderlake system to work I get 

 

Feb  2 17:05:08 computenode usb_manager: Info: rc.usb_manager Autoconnect Start SIGMACHIP_USB_Keyboard vm: Windows 10 Device in Use action ignored. 
Feb  2 17:05:09 computenode kernel: vfio-pci 0000:00:02.0: vfio_ecap_init: hiding ecap 0x1b@0x100
Feb  2 17:05:09 computenode kernel: vfio-pci 0000:00:02.0: Invalid PCI ROM header signature: expecting 0xaa55, got 0xffb6

 

other thread for info.

 

 

I´ll give it a try, but when I first setup my windows vm I had hella problem just getting it to work period, so I tried every combination of ovfm/seabios/i440fx/q35 that is possible until I landed on the setup I´ve got working now with the 1660ti, everything else gave errors or blackscreens. 

Thats why i tried with a linux vm, (in the past they have been more forgiving about the specific setup, working with just about any config i´ve tried), but even that wont work! 
But I´ll give seabios a go on both windows and linux and report back! 

[BeardElk]

On 2/2/2022 at 6:24 PM, SimonF said:

I have been reading another thread, It may need seabios but I cannot get my Alderlake system to work I get 

 

Feb  2 17:05:08 computenode usb_manager: Info: rc.usb_manager Autoconnect Start SIGMACHIP_USB_Keyboard vm: Windows 10 Device in Use action ignored. 
Feb  2 17:05:09 computenode kernel: vfio-pci 0000:00:02.0: vfio_ecap_init: hiding ecap 0x1b@0x100
Feb  2 17:05:09 computenode kernel: vfio-pci 0000:00:02.0: Invalid PCI ROM header signature: expecting 0xaa55, got 0xffb6

 

other thread for info.

 

 

After I´ve created the vm, I still cannot change from vnc to uhd750, and if I delete the vm (not the image) and create a new one with identical specs except uhd750 as vga and sound, I get this: 
"device 0000:00:02.0 cannot support legacy mode due to existing devices at address 1f.0
2022-02-08T23:29:10.336064Z qemu-system-x86_64: vfio: Cannot reset device 0000:00:1f.5, no available reset mechanism.
2022-02-08T23:29:10.341984Z qemu-system-x86_64: vfio: Cannot reset device 0000:00:1f.3, no available reset mechanism.
2022-02-08T23:29:10.452120Z qemu-system-x86_64: vfio: Cannot reset device 0000:00:1f.5, no available reset mechanism.
2022-02-08T23:29:10.455971Z qemu-system-x86_64: vfio: Cannot reset device 0000:00:1f.3, no available reset mechanism."

 

I´ve never gotten a seabios vm working on the hardware I´m running now.

Still get no input on the display as soon as I try to start the vm with igpu connected. 
I´ve rebooted the server so I can see when it goes black. 

Starting to give up hope on this, and regretting not going ryzen.......

[BeardElk]

I can change between my 1660ti and 1050ti like I´m supposed to, but Nothing is working for my igpu. 
For shits and giggles I swapped my transcode gpu (1050ti) with my working win11 gpu (1660ti) and it was plug and play. 
De-select for unbind my 1660ti, bind the 1050ti, reboot, and swap in plex docker/win11 vm. 
Probably gonna stick with this setup since I rarely game on that vm, and IF I want to, its an easy swap. 
Just upset cuz I planned on using igpu for win11 vm and something different for my 1050ti, but you cant have it all I guess.

Edited February 9, 2022 by BeardElk