On Sat, Mar 4, 2023 at 3:21 AM Karel Gardas <karel@functional.vision> wrote:
> > One remark as an random observer here. > "branch 6-freebsd-12" has caught my eyes. Let me ask shouldn't > development patches go into the master branch from which they may be > later cherry-picked if needed and pushed into 6-freebsd-12 branch? > > Just few weeks ago guys were having a discussion how to sync branches > since they diverged due to patches pushed into both directions without a > proper syncing... > Mea culpa, In general, you're absolutely correct. The CGEM driver in particular is one of the places where the current state between 6-freebsd-12 does not match. When I got the CGEM driver working initially for ZynqMP, it was on 6-freebsd-12 and at the time there was a lot of complicated work in progress (some of the other things that still do not match) targeted at master from others in the community that I didn't want to disturb by causing more churn. I was waiting for those to resolve before I moved forward on master, but that never actually happened. Looking back, any patches I had to master would have been unlikely to cause them any real headache, but that's hindsight for you. Another complication is that the FreeBSD 13 CGEM driver has a lot of the functionality that has been backported to 6-freebsd-12 or changed from it, so some patches become near-irrelevant, some need a complete rework, and some apply cleanly. This may be one of the cases where it applies relatively cleanly, but there would be no way to test it without bringing up the driver to full functionality and I don't have a good idea of the level of effort that would take. All that said, getting CGEM on master up to a functioning level is on my todo list so that patches can be applied to both and tested properly. Kinsey
_______________________________________________ devel mailing list devel@rtems.org http://lists.rtems.org/mailman/listinfo/devel
