On 2018-07-24 22:09, Axel Beckert wrote: > Can you check if the kernel headers for the running kernel were > present when you encountered the issue?
This was in a sid chroot with only 4.17 headers installed (on a stretch host, so running 4.9). > If so, I might need to patch upstream's build system to cope with that > situation, as I'd expect that while the currently running kernel > version should be the default, but not the only one possible. (IIRC I > was able to compile the module for kernel 4.16 while running 4.17 with > the kernel headers already being installed while the 4.16 package has > been installed again.) dkms should be the one choosing the kernel version to build for. If the source unconditionally builds for the running kernel only, that's a bug. More modules had this problem a few years ago, but we got them all fixed so I didn't consider that as a possible cause :-) Andreas
