Describe the bug
My Index HMD is not detected. However, EDID information does exist enough to identify the HMD. Something in the scanning and detection of HMDs fails. I suspect this could be due to only scanning a single level in /sys/class/drm/<SOME CONNECTED DEVICES/edid. However, on my system due to a more complex configuration here from a KVM switch, some monitor sleeping then re-connection, and multiple GPUs. I'm seeing this (hopefully not in error) having the effect that some edid's are another level deeper in the directory structure. A simple change to a recursive search for "edid" files should be sufficient to fix this.
Scanning doesn't seem to use the /sys interface, so I need to dig in the strace and look into output detection in OpenGL and Vulkan to find out other potential detection methods used.
Describe the bug
My Index HMD is not detected. However, EDID information does exist enough to identify the HMD. Something in the scanning and detection of HMDs fails.
I suspect this could be due to only scanning a single level in/sys/class/drm/<SOME CONNECTED DEVICES/edid. However, on my system due to a more complex configuration here from a KVM switch, some monitor sleeping then re-connection, and multiple GPUs. I'm seeing this (hopefully not in error) having the effect that someedid's are another level deeper in the directory structure. A simple change to a recursive search for "edid" files should be sufficient to fix this.