Open CharlesTerrell opened 1 year ago
Yes, that was the pull req for this.
Tonight it failed on an auto-rebuild after a minor kernel upgrade. From the error messages, I think it's still getting confused about the unwieldy version numbers pulled in from the headers. I'll have a look at it tomorrow. For now, doing "sudo make dkms_remove" and then "sudo make dkms_install" (followed by a modprobe) got it running again. Having to do that every time defeats the purpose of dkms though.
The good news is, I figured out how to force the auto-rebuild process, so I'll be able to repeat and test the circumstances leading to the error.
As I did not write the dkms stuff, and do not support it, the most I will accept are patches.
The version string used by dkms comes from include/drv_types.h, specifically the line that says:
That's quite a version string! Where does it come from? Does it have any relation to updates done here? If so, how does it get updated?
If it came from elsewhere, I'll keep that in mind while tracking the rebuild problem.
That comes frrom the original driver. It is not updated here. You might track the git commit version, but I do not know how to do that from a dkms script.
Building via dkms on Ubuntu 22.10 failed with weird compile errors. For example, looking for hal/phydm in the kernel headers. The person who was maintaining the PPA seems to have last updated it several months ago, so that didn't work either. I got it to build and install on my laptop ("evolve iii maestro-ebook-11") by changing dkms.conf and Makefile to look a little more like the rtl8812au driver. Now it builds, installs, and seems to be working. Haven't tested uninstalling, though, nor has it been through a kernel upgrade yet.
I'll do a pull request with my changes.