On 17.12.21 08:55, C Smith wrote: > There are development scenarios where different versions of a driver > are substituted. I can tell you some long stories when one day we meet > over a few beers...
I can imagine those cases - but only when you patch the driver locally anyway. In that case, a local instrumentation patch is not an issue to carry along that. If you are using Xenomai upstream, the driver version is defined by the Xenomai version. If you like to make that more precise, submit a patch that also forwards (prepare-kernel.sh -> XENO_VERSION_SRCREV) and dumps (kernel/cobalt/init.c, sys_init) the git revision. That may even solve your local modification scenario as well if that uses git to manage the changes. There are reasons why Linux no longer prints version information for drivers as well (except for historic cases). The kernel version / revision is sufficient, and the boot log would explode if every loaded driver did that. Jan -- Siemens AG, Technology Competence Center Embedded Linux