[Plugin] Intel-GVT-g


Recommended Posts

13 minutes ago, alturismo said:

i see you still have a active vgpu running, are you running more then 1 VM with it ? which is currently active ?

 

There could be the problem ...

 

I do NOT have any VM running where i have a vgpu attached to it. I have just 2 linux servers running, but without a vgpu assignment.

 

When I first experienced this issue, i thought it might help to delete all assignments and recreate them one by one.

Currently I have only created the following assignment:

 

4.thumb.png.568d57cf494b4dec8014527c2f668941.png

 

And this VM is stopped because it wont start without the error.

So no other VMs with a vgpu running.

 

 

 

Quote

or may just try this and see what the respond is

 

5.thumb.png.1afccd35b8dbe14d18d13d3dac8bce6d.png

also not working

 

Edited by greenflash24
Link to comment
5 minutes ago, greenflash24 said:

also not working

 

 

ok, there we have the error ;)

 

i guess its your "older" test you made before ...

 

now lets remove the existing one, replace the UID (REPLACEHERE) with the one i showed you above from your pasting, this should free up the space ...

 

echo 1 > "/sys/devices/pci0000:00/0000:00:02.0/REPLACEHERE/remove"

 

  • Like 1
Link to comment
2 minutes ago, greenflash24 said:

Thank you a lot for your help!

 

 

you welcome and enjoy, please see my other note about BIOS and RAM assignment and watch the used resolution (specially on RDP), max should be 1920x1080 ... when u increase u can also get vgpu page errors ending in a total crash, as long as you stay inside the specs you should be fine

  • Like 1
Link to comment
8 minutes ago, alturismo said:

btw, on the next reboot, take a look at your bios settings for the igpu and assign the max RAM you can do there ... im sadly also limited by my bios ... others can increase up to 1 GB and really can run more then 1 in v4 mode ... ;)

 

3 minutes ago, alturismo said:

 

you welcome and enjoy, please see my other note about BIOS and RAM assignment and watch the used resolution (specially on RDP), max should be 1920x1080 ... when u increase u can also get vgpu page errors ending in a total crash, as long as you stay inside the specs you should be fine

 

 

I have already done this on my last reboot, as I read in this thread that this might help.

I could increase it to 1GB, which i have already done.

 

Sadly I can not run more then 1 in v4 mode only in v8 mode, even when 1GB vram in BIOS is set.

Link to comment

I am still having reliability issues with my Win 10 VM and Intel-GVT,

 

I am now receiving a different error to previous errors:

(Previous error was "Tower kernel: gvt: guest page write error, gpa XXXXXXXX").

 

As pre previous post:

My details:

 

Found VGA Adapter:
Intel Corporation CoffeeLake-S GT2 [UHD Graphics 630] (rev 02) using mode V5_4 with Win10 VM

 

I have GuC/HuC Firmware loading: set to default

 

Win 10 VM is also assigned a Nvidia GPU (1050 Ti)

There is no other VM assigned a vGPU 

 

I have set RDP max resolution to 1920x1080 via Group Policy Editor on the VM

 

I have increased the BIOS values to the max for the following options:

dvmt pre-allocated

dvmt total gfx mem

 

 

My VM will work happily away with BlueIris doing some Intel based transcoding with approx 15-20% GPU usage reported via Task Manager but it crashes at random and then makes the VM system inaccessible on unRAID.

 

Diags attached.

 

Syslog reports the below at time of VM crash:

Nov  6 15:00:1Nov  6 15:00:18 Tower kernel: gvt: vgpu 1: fail: shadow page 0000000000000000 guest entry 0xffffffffffffffff type 9
Nov  6 15:00:18 Tower kernel: gvt: vgpu 1: fail: spt 00000000a1a28f04 guest entry 0xffffffffffffffff type 9
Nov  6 15:00:18 Tower kernel: gvt: vgpu 1: fail: shadow page 00000000a1a28f04 guest entry 0xffffffffffffffff type 9.
Nov  6 15:00:18 Tower kernel: gvt: guest page write error, gpa 2c6daa000
Nov  6 15:00:18 Tower kernel: gvt: vgpu 1: fail: shadow page 0000000000000000 guest entry 0xffffffffffffffff type 9
Nov  6 15:00:18 Tower kernel: gvt: vgpu 1: fail: spt 00000000a1a28f04 guest entry 0xffffffffffffffff type 9
Nov  6 15:00:18 Tower kernel: gvt: vgpu 1: fail: shadow page 00000000a1a28f04 guest entry 0xffffffffffffffff type 9.
Nov  6 15:00:18 Tower kernel: gvt: guest page write error, gpa 2c6daa008
Nov  6 15:00:18 Tower kernel: gvt: vgpu 1: fail: shadow page 0000000000000000 guest entry 0xffffffffffffffff type 9
Nov  6 15:00:18 Tower kernel: gvt: vgpu 1: fail: spt 00000000a1a28f04 guest entry 0xffffffffffffffff type 9
Nov  6 15:00:18 Tower kernel: gvt: vgpu 1: fail: shadow page 00000000a1a28f04 guest entry 0xffffffffffffffff type 9.
Nov  6 15:00:18 Tower kernel: gvt: guest page write error, gpa 2c6daa010

 

 

Diags attached

 

As always thanks so much for your work and any help that can be given.

 

tower-diagnostics-20211106-1919.zip

Link to comment
35 minutes ago, Capt.Insano said:

As always thanks so much for your work and any help that can be given.

 

there is not much todo, your error occurs due vgpu has not enough RAM sometime ...

 

all you can do or try, lower the resolution even more so your blue iris app may have more ressources and may set the highest fix value in the BIOS instead which hopefully is at least 256mb

Link to comment

I have 5 cameras in my set up but 4 of them are fairly low res so seem to be consuming very little effort from the iGPU.

 

I dont know of any BlueIris setting to limit usage, however usage is never over 15-20% 

 

 

Current Task Manager status of VM:

 

image.thumb.png.6dbd11b42d4c3e448abf2541280f3d65.png

 

image.png.9d88a956e01faf468a0dd1b21605a854.png

 

GPU 1 = Intel iGPU

GPU0 = Nvidia Card

 

Current Bios Settings:

 

954293897_PXL_20211106_214933483(Medium).thumb.jpg.091a645e3974733941a62ec1fb064e45.jpg

 

 

 

Was prev on MAX but changed to 256M instead of MAX to see if stability improves.

 

 

968716824_PXL_20211106_214915544(Medium).thumb.jpg.4df64fc9cfcb2d7aa8d7865862947eac.jpg

 

Left this on 1024M

 

 

799923102_PXL_20211106_215215473(Medium).thumb.jpg.bc65bf0695b4a7788a4a3d1f0b8edc2a.jpg

 

I also just changed the "Aperture Size" for the first time, I am not sure if it is relevant.

I was default on 256M but I increased it to 1024M (I tried 2048M but it caused issues booting unRAID)

 

 

 

Link to comment
  • 3 weeks later...

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.