HP DL360 G6 GPU passthrough

[sysco]

Hello everyone,
so I have been playing around with a couple of DL360s I picked up at work, one of then works just fine as a normal plex, VPN, etc. The second one was meant for GPU passthrough. Now that's where the issues start. I want to use a RADEON HD 6870 as the guest GPU with the onboard graphics as host. Both GPUs are fine when booting to other OS (Ubuntu, Kali) but when I try to pass through the Radeon to the Win10 host I don't get any signal from the GPU whatsoever.
I am quite new to all this so any help would be appreciated. And I don't know if this is a ProLiant issue, global unRAID issue or me just messing up the settings.
I can provide any additional information.

[sysco]

Just tried with HD 7570, works like a charm.

Still no luck with 6870 tho.

[1812]

 

 

in regards to the 6870 not passing through, it may be too old or just have general AMD passthrough issues. 

[sysco]

14 hours ago, 1812 said:

 

 

in regards to the 6870 not passing through, it may be too old or just have general AMD passthrough issues. 

I tried everything in those posts and still came out not working.

I forgot to mention that it worked on another system, it was the same motherboard but more ram and better CPUs. Worked without any problems but not on this one. Maybe the BIOS is different... I will check later.

[sysco]

Which BIOS is recommended? I have HP, Version P64 Sun 16 Aug 2015 12:00:00 AM CEST. I guess that it is too new or something.

[1812]

Which BIOS is recommended? I have HP, Version P64 Sun 16 Aug 2015 12:00:00 AM CEST. I guess that it is too new or something.

Try to start the vm and immediately after all fails post the zip file from tools>diagnostics


Sent from my iPad using Tapatalk

[sysco]

2 hours ago, 1812 said:

 

Try to start the vm and immediately after all fails post the zip file from tools>diagnostics


Sent from my iPad using Tapatalk

Well, nothing fails. I just don't see any output on the screen.

I don't even know anymore, it just booted up...

Still, here are the diagnostics

unraidnas-diagnostics-20190726-1411.zip

[1812]

16 minutes ago, sysco said:

Well, nothing fails. I just don't see any output on the screen.

I don't even know anymore, it just booted up...

Still, here are the diagnostics

unraidnas-diagnostics-20190726-1411.zip 64.88 kB · 0 downloads

 

a few things:

Jul 26 16:06:32 unraidNAS kernel: ACPI Error: SMBus/IPMI/GenericSerialBus write requires Buffer of length 66, found length 32 (20180810/exfield-393)
Jul 26 16:06:32 unraidNAS kernel: ACPI Error: Method parse/execution failed \_SB.PMI0._PMM, AE_AML_BUFFER_LIMIT (20180810/psparse-516)
Jul 26 16:06:32 unraidNAS kernel: ACPI Error: AE_AML_BUFFER_LIMIT, Evaluating _PMM (20180810/power_meter-338)
Jul 26 16:06:40 unraidNAS ntpd[1898]: kernel reports TIME_ERROR: 0x41: Clock Unsynchronized
Jul 26 16:06:51 unraidNAS kernel: ACPI Error: SMBus/IPMI/GenericSerialBus write requires Buffer of length 66, found length 32 (20180810/exfield-393)
Jul 26 16:06:51 unraidNAS kernel: ACPI Error: Method parse/execution failed \_SB.PMI0._PMM, AE_AML_BUFFER_LIMIT (20180810/psparse-516)
Jul 26 16:06:51 unraidNAS kernel: ACPI Error: AE_AML_BUFFER_LIMIT, Evaluating _PMM (20180810/power_meter-338)

 

you can fix that via the directions in the thread I posted earlier.

 

 

Jul 26 16:08:58 unraidNAS kernel: NMI: PCI system error (SERR) for reason a1 on CPU 0.
Jul 26 16:08:58 unraidNAS kernel: Dazed and confused, but trying to continue

 

I've gotten these randomly on proliants. Sometimes its hardware config, sometimes its bios setting being wonky. I'd recommend downloading the latest available from HP (they made one for all the spectre issues) It may fix that.

 

 

for your gpu, try recreating the windows 10 vm with seabios  

 

 

[sysco]

2 minutes ago, 1812 said:

 

a few things:

Jul 26 16:06:32 unraidNAS kernel: ACPI Error: SMBus/IPMI/GenericSerialBus write requires Buffer of length 66, found length 32 (20180810/exfield-393)
Jul 26 16:06:32 unraidNAS kernel: ACPI Error: Method parse/execution failed \_SB.PMI0._PMM, AE_AML_BUFFER_LIMIT (20180810/psparse-516)
Jul 26 16:06:32 unraidNAS kernel: ACPI Error: AE_AML_BUFFER_LIMIT, Evaluating _PMM (20180810/power_meter-338)
Jul 26 16:06:40 unraidNAS ntpd[1898]: kernel reports TIME_ERROR: 0x41: Clock Unsynchronized
Jul 26 16:06:51 unraidNAS kernel: ACPI Error: SMBus/IPMI/GenericSerialBus write requires Buffer of length 66, found length 32 (20180810/exfield-393)
Jul 26 16:06:51 unraidNAS kernel: ACPI Error: Method parse/execution failed \_SB.PMI0._PMM, AE_AML_BUFFER_LIMIT (20180810/psparse-516)
Jul 26 16:06:51 unraidNAS kernel: ACPI Error: AE_AML_BUFFER_LIMIT, Evaluating _PMM (20180810/power_meter-338)

 

you can fix that via the directions in the thread I posted earlier.

 

 

Jul 26 16:08:58 unraidNAS kernel: NMI: PCI system error (SERR) for reason a1 on CPU 0.
Jul 26 16:08:58 unraidNAS kernel: Dazed and confused, but trying to continue

 

I've gotten these randomly on proliants. Sometimes its hardware config, sometimes its bios setting being wonky. I'd recommend downloading the latest available from HP (they made one for all the spectre issues) It may fix that.

 

 

for your gpu, try recreating the windows 10 vm with seabios  

 

 

You mean the general thread or the rmrr error patching?

[sysco]

It just started working, i didn't do anything just booted as usual and it worked.

[1812]

7 hours ago, sysco said:

It just started working, i didn't do anything just booted as usual and it worked.

then my work here is done.

[sysco]

Just now, 1812 said:

then my work here is done.

Thank you. Might need some more of that after I get the Pcie extender from one stop systems and a better gpu