diff options
Diffstat (limited to 'results/classifier/accel-gemma3:12b/kvm/1671677')
| -rw-r--r-- | results/classifier/accel-gemma3:12b/kvm/1671677 | 23 |
1 files changed, 23 insertions, 0 deletions
diff --git a/results/classifier/accel-gemma3:12b/kvm/1671677 b/results/classifier/accel-gemma3:12b/kvm/1671677 new file mode 100644 index 000000000..0144f6903 --- /dev/null +++ b/results/classifier/accel-gemma3:12b/kvm/1671677 @@ -0,0 +1,23 @@ + +vfio-pci passthrough issue after resume from suspend + + +I'm running into a weird issue with the vfio-pci driver through qemu + +I use it on a laptop and I passthrough an external GPU connected via PCI express. In general this works well, however if the laptop has *ever* suspended since its last boot, then the windows guest reports an error 43 on the card and I get no output on the monitor that it is connected to. This is really weird to me since it works fine if I boot the latop from power-off, and hotplug the card. It's only if the laptop has ever suspended since it's last boot that I see this issue. Even if it was suspended before the card was ever hotplugged. + +In other words: +laptop off -> boot -> hotplug GPU : works great +laptop off -> boot -> do stuff (GPU *NOT* connected) -> sleep -> resume -> hotplug GPU: problem +laptop off -> boot -> hotplug GPU -> unplug GPU -> hotplug GPU : works great +laptop off -> boot -> hotplug GPU -> unplug GPU -> sleep -> resume -> hotplug GPU: problem + +Weird stuff... + +I'm honestly not sure that vfio-pci/qemu is to blame here since there are so many moving parts, but im not really sure where else to report this to + +What I have tried is using the sysfs interface to remove/rescan/poweroff/etc the PCI devices in questions (graphics card and it's HDMI audio) and this also does help. + +QEMU version: 2.6.1 + +Please let me know what other information I can provide \ No newline at end of file |