Hi guys,
I have a Intel CoffeeLake system (i3-8300 on C246 chipset), with installed Proxmox VE 6.2 (latest update status) that I wanted to use with several containers and one VM with fully passed-through i915 graphics.
Found a setup, that is nearly fully working (vm conf excerpt):
Now, that setup is already working pretty well, besides the following issue:
I have a Debian testing VM setup, that gets its syslog flooded to no end with "pipe error 0x00000880" error messages from i915.
I did some google searching and found the following bug report with a very similar, but not exactly the same error message:
https://bugs.freedesktop.org/show_bug.cgi?id=109099
Went ahead with the described workaround and rebuilt the Debian VM's kernel, with the mentioned code change in the "i915_drv.h" file, forcing the "intel_vtd_active" function to return "true".
With that, the errors are gone and it seems to work just fine.
Which is great, however, I was wondering:
- Is someone else is also experiencing this?
- Is it a already known issue?
- Why do I have to force the vt-d detection?
Thanks in advance for your comments.
Best regards,
Ingmar
I have a Intel CoffeeLake system (i3-8300 on C246 chipset), with installed Proxmox VE 6.2 (latest update status) that I wanted to use with several containers and one VM with fully passed-through i915 graphics.
Found a setup, that is nearly fully working (vm conf excerpt):
Code:
...
hostpci0: 00:1f,pcie=1
hostpci1: 00:02,pcie=1,x-vga=1
machine: pc-q35-4.1 (<== if I set anything newer than '4.1', the display output will no longer work)
serial0: socket
vga: serial0
...
Now, that setup is already working pretty well, besides the following issue:
I have a Debian testing VM setup, that gets its syslog flooded to no end with "pipe error 0x00000880" error messages from i915.
I did some google searching and found the following bug report with a very similar, but not exactly the same error message:
https://bugs.freedesktop.org/show_bug.cgi?id=109099
Went ahead with the described workaround and rebuilt the Debian VM's kernel, with the mentioned code change in the "i915_drv.h" file, forcing the "intel_vtd_active" function to return "true".
With that, the errors are gone and it seems to work just fine.
Which is great, however, I was wondering:
- Is someone else is also experiencing this?
- Is it a already known issue?
- Why do I have to force the vt-d detection?
Thanks in advance for your comments.
Best regards,
Ingmar