Tesla xenctrlext.unix_error Device or resource is busy

[kuznetcoff777]

Hi! I have XCP-ng release 8.2.1 and tesla, only one videocard (becasue of x8da3 motherboard, we have now network card and tesla without any videocard for video-kvm access)

I checked all VMs, no one more have this gpu, only this VM and it does not start. Can you please tell how to solve this problem?

[olivierlambert]

The GPU is probably not hidden to the Dom0

[kuznetcoff777]

olivierlambert
Yep, added excusion /opt/xensource/libexec/xen-cmdline --set-dom0 "xen-pciback.hide=(0000:05:00.0)" (https://docs.xcp-ng.org/compute/)
worked llike a sharm, thanks.

[olivierlambert]

Note that PCI passthrough will be entirely exposed in XO with XCP-ng 8.3, no need for the CLI anymore

[kuznetcoff777]

olivierlambert
After restart VM it began to show the same error, just like hide list is empty. If any way to check this hide list?

May 22 11:10:57 xcp kernel: [86647.738094] pciback 0000:05:00.0: timed out waiting for pending transaction; performing function level reset anyway
May 22 11:10:58 xcp kernel: [86648.986119] pciback 0000:05:00.0: not ready 1023ms after FLR; waiting
May 22 11:10:59 xcp kernel: [86650.042116] pciback 0000:05:00.0: not ready 2047ms after FLR; waiting
May 22 11:11:02 xcp kernel: [86652.346147] pciback 0000:05:00.0: not ready 4095ms after FLR; waiting
May 22 11:11:06 xcp kernel: [86656.698127] pciback 0000:05:00.0: not ready 8191ms after FLR; waiting
May 22 11:11:14 xcp kernel: [86665.146149] pciback 0000:05:00.0: not ready 16383ms after FLR; waiting
...
May 22 11:12:08 xcp kernel: [86718.394184] pciback 0000:05:00.0: not ready 65535ms after FLR; giving up

[olivierlambert]

Maybe this will ring a bell to anyone around here?

[kuznetcoff777]

Can someone please delete Solved sign? Cause problem still persists...
If any way to check what is in hide list?

[olivierlambert]

Put back as unresolved

[kuznetcoff777]

[11:12 xcp ~]# dmesg | grep pciback
[    0.000000] Command line: root=LABEL=root-apezjy ro nolvm hpet=disable rd.auto console=hvc0 console=tty0 quiet vga=785 splash plymouth.ignore-serial-consoles xen-pciback.hide=(0000:05:00.0)
[    3.313418] Kernel command line: root=LABEL=root-apezjy ro nolvm hpet=disable rd.auto console=hvc0 console=tty0 quiet vga=785 splash plymouth.ignore-serial-consoles xen-pciback.hide=(0000:05:00.0)

[20:49 xcp ~]# lspci -s 0000:05:00.0 -v
05:00.0 3D controller: NVIDIA Corporation TU104GL [Tesla T4] (rev ff) (prog-if ff)
        !!! Unknown header type 7f
        Kernel driver in use: pciback

Seems to be that it was on boot in exception and now it is still in. But VM cannot get it...

[olivierlambert]

You are only hidding one PCI address. In your dom0, with a lspci command, double check if you have other Nvidia references

[kuznetcoff777]

Yep, only one, cause one tesla onboard, no any other video devices (maybe it is a problem)

[20:53 xcp ~]# lspci | grep -i nvidia
05:00.0 3D controller: NVIDIA Corporation TU104GL [Tesla T4] (rev ff)