AMD RX 6800 XT reset bug happens every time


Recommended Posts

Did some testing and found that I do not need the plugin that @ich777 suggested. The problem seemed to be a VM template throwing errors. 

When I tried passing through the USB function of my GPU, the template gave me an error telling me that a device could not be found. I could not fix that error and just kept unstubbing the GPU parts to make it work again. Turns out that the solution was to just create a new template with all the parts of the GPU passed through. 

  • Like 1
  • Thanks 1
Link to comment
  • 2 weeks later...

having somewhat similiar problems with my Radeon rx 6800xt.

Unraid 6.9.2

All  4 components for the gpu are passed trough to the Win10 Vm.

Win 10 VM boots up, running some games and performance tests works. 

but then randomly from 1 day to 1 hrs running the VM the amd software reports driver crash, the gpu fans ramps up to maximum the monitor switches to windows basic adapter and the Win 10 VM freezes only possible to stop the VM in unraid gui.

 

tested drivers: Adrenalin 21.8.2 and Adrenalin 21.6.1

machine type: q35 -5.0 and q35 -5.1

VM logs show no warnings or error but the system logs does.

1296382253_vfiorx6800xt.thumb.png.40c151842bfabb7f9b12a04620e0af1b.png

 

is there something i should include in the XML as per this post setting a vendor id?

or as @ich777 mention in this post upgrading to 6.10 rc1 and install amd vendor reset?

the big Navi cars should not be affected by the reset bug so i am strugling at this moment to pin point the cause.

 

Edited by TIE Fighter
Link to comment
11 minutes ago, TIE Fighter said:

having somewhat similiar problems with my Radeon rx 6800xt.

Unraid 6.9.2

All  4 components for the gpu are passed trough to the Win10 Vm.

Win 10 VM boots up, running some games and performance tests works. 

but then randomly from 1 day to 1 hrs running the VM the amd software reports driver crash, the gpu fans ramps up to maximum the monitor switches to windows basic adapter and the Win 10 VM freezes only possible to stop the VM in unraid gui.

 

tested drivers: Adrenalin 21.8.2 and Adrenalin 21.6.1

machine type: q35 -5.0 and q35 -5.1

VM logs show no warnings or error but the system logs does.

1296382253_vfiorx6800xt.thumb.png.40c151842bfabb7f9b12a04620e0af1b.png

 

is there something i should include in the XML as per this post setting a vendor id?

or as @ich777 mention in this post upgrading to 6.10 rc1 and install amd vendor reset?

the big Navi cars should not be affected by the reset bug so i am strugling at this moment to pin point the cause.

 

Hi, 

 

I do not have the knowledge to help you out. Sorry. But in my case it was just an issue of not creating a good VM template. 

You mention that you can boot the system just fine and even game, run benchmarks and stresstest it. That is great. But it crashes. The crash looks like it is caused by the AMD driver. I am also running 21.8.2 and do not have that issue. 

 

When my system did crash, the card did not ramp up it's fans but stopped them. 

 

You mention that you have 4 pieces of the GPU passed through. I thought that there only were 3. 

Here is a screenshot of my VM. I hope that it helps. 

image.thumb.png.eedc8c6f1a250c62a5c1fd552dc03abc.png

 

image.thumb.png.a2746ed62f63186105b9ef393ff75570.png

 

I do not have any vendor reset things going on. It seems like my card does not have the AMD reset bug. 

Link to comment
21 hours ago, workermaster said:

You mention that you have 4 pieces of the GPU passed through. I thought that there only were 3. 

Here is a screenshot of my VM. I hope that it helps. 

there are 4 on my Radeon RX 6800xt on the bus 10.x.x all passed1204086414_Namnls.thumb.png.5f9da1b1c2012ee02a93712629879ce3.png

21 hours ago, workermaster said:

by the AMD driver. I am also running 21.8.2 and do not have that issue. 

atleast it is was not a driver issue.

infact i found the issue was caused by the S3 sleep plugin. when the plugin didnt pick up any disk activity a sleep command was issued trying to put the server asleep this made the VM crash and sleep was put on hault. my second vm was restarted so that made me think it could also be related to the host, i thought the S3 Sleep plugin would know if any Vms were infact running but it looks like it didnt.

 

to find the ip adress for the VMs i Open unraid terminal and type:

 

arp -a | grep br0

 

the Vms ip adress to be put to "wait for host inactivity"

436169458_Namnls2.thumb.png.371f48e9c0e87db29056b1b267d2d2f3.png

this made all the issues dissapear 🙂

 

cheers

 

 

Edited by TIE Fighter
Link to comment
2 hours ago, TIE Fighter said:

there are 4 on my Radeon RX 6800xt on the bus 10.x.x all passed1204086414_Namnls.thumb.png.5f9da1b1c2012ee02a93712629879ce3.png

atleast it is was not a driver issue.

infact i found the issue was caused by the S3 sleep plugin. when the plugin didnt pick up any disk activity a sleep command was issued trying to put the server asleep this made the VM crash and sleep was put on hault. my second vm was restarted so that made me think it could also be related to the host, i thought the S3 Sleep plugin would know if any Vms were infact running but it looks like it didnt.

 

to find the ip adress for the VMs i Open unraid terminal and type:

 

arp -a | grep br0

 

the Vms ip adress to be put to "wait for host inactivity"

436169458_Namnls2.thumb.png.371f48e9c0e87db29056b1b267d2d2f3.png

this made all the issues dissapear 🙂

 

cheers

 

 

Good to hear that you found the issue. 

 

I also see that clearly my screenshot shows that my GPU also has 4 parts passed through. My mistake there. 

Link to comment
  • 1 month later...

@ich777 @glennv

I was hoping you guys see something in here that I am not able to. The issue is the same: I am able to pass through the RX 6800XT, boot into Windows, download the latest AMD Radeon drivers for installation and when the VM restarts during the graphic driver installation, Unraid crashes! Not just the VM! 

 

I am on  6.10.0-rc1 and have installed the AMD Vendor Reset patch. Unraid boots in UEFI mode

 

Any help much appreciated. Spent a fruitless 7 hours on this now...trying various permutations and combinations.

diagnostics-20211019-2000.zip WinVMXML.md

Link to comment
@ich777 @glennv
I was hoping you guys see something in here that I am not able to. The issue is the same: I am able to pass through the RX 6800XT, boot into Windows, download the latest AMD Radeon drivers for installation and when the VM restarts during the graphic driver installation, Unraid crashes! Not just the VM! 
 
I am on  6.10.0-rc1 and have installed the AMD Vendor Reset patch. Unraid boots in UEFI mode
 
Any help much appreciated. Spent a fruitless 7 hours on this now...trying various permutations and combinations.
diagnostics-20211019-2000.zip WinVMXML.md
As far as I know the 6xxx series should not be affected from the AMD Reset bug.

Make sure you've created a OVMF VM and also start unRAID with UEFI not with Legacy (CSM).

Sent from my C64

Link to comment
52 minutes ago, ich777 said:

As far as I know the 6xxx series should not be affected from the AMD Reset bug.

Make sure you've created a OVMF VM and also start unRAID with UEFI not with Legacy (CSM).

Sent from my C64
 

Thanks for getting back.

I got it working. I uninstalled the AMD Vendor Reset plugin from UnRaid. I disabled `Resizable BAR` and `Above 4G memory/Crypto Currency mining` in my MSI x570 Tomahawk Wifi and everything fell into place. Wow..took me 8 hours 😞 

Link to comment
Thanks for getting back.
I got it working. I uninstalled the AMD Vendor Reset plugin from UnRaid. I disabled `Resizable BAR` and `Above 4G memory/Crypto Currency mining` in my MSI x570 Tomahawk Wifi and everything fell into place. Wow..took me 8 hours  
Glad to hear that everyrhing is working now.

AMD is a little behind when it comes to virtualisation from my perspective.

Sent from my C64

Link to comment
2 minutes ago, ich777 said:

Glad to hear that everyrhing is working now.

AMD is a little behind when it comes to virtualisation from my perspective.

Sent from my C64
 

Yeah. Wish things were easier.

For posterity, in case someone else is facing similar issues, I should also mention that:

- I just used a vBIOS from techpowerup and did not dump a bios from my GPU (was having an issue with spaceinvaderone's script anyway). Apart from user scripts and CA, I had no plugins installed

- I had to use the param `video=efifb:off` in Unraid boot options

- Unraid was booting in UEFI mode

- VM type was a Q35 6.0

  • Like 2
Link to comment

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.