Good day,
I am facing a weird issue where I can passthrough the Vega M gpu to a Windows 10 VM and reboot/shutdown without issues.
Passing the gpu through to an Ubuntu 18.04 vm works fine for the first time, but once I reboot the vm, the host freeze (unable to access web interface or SSH).
If I shut the vm down, everything works ok till I try launching the vm, then the host freezes.
Could that be an ESXi reset issue with the AMD Vega GH M gpu ?
I read about disabling the gpu on the windows device manager before shutting it down, so I went ahead and did the equivalent on Ubuntu:
1- unbind the gpu from the amdgpu driver.
2- then echo 1 > /sys/devices/pci********/remove
The host still froze on turning the vm back on.
I then tried all combinations of resetting both the gpu HDMI and audio inside the /etc/vmware/passthru.map (d3d0, link, bridge) but that did not help to reboot the VM without freezing the host.
I am at my wits end here, so any suggestion would be appreciated.
I am facing a weird issue where I can passthrough the Vega M gpu to a Windows 10 VM and reboot/shutdown without issues.
Passing the gpu through to an Ubuntu 18.04 vm works fine for the first time, but once I reboot the vm, the host freeze (unable to access web interface or SSH).
If I shut the vm down, everything works ok till I try launching the vm, then the host freezes.
Could that be an ESXi reset issue with the AMD Vega GH M gpu ?
I read about disabling the gpu on the windows device manager before shutting it down, so I went ahead and did the equivalent on Ubuntu:
1- unbind the gpu from the amdgpu driver.
2- then echo 1 > /sys/devices/pci********/remove
The host still froze on turning the vm back on.
I then tried all combinations of resetting both the gpu HDMI and audio inside the /etc/vmware/passthru.map (d3d0, link, bridge) but that did not help to reboot the VM without freezing the host.
I am at my wits end here, so any suggestion would be appreciated.