On Wed, Nov 29, 2017 at 11:12:44AM -0500, Doug Ledford wrote:
> On 11/19/17 1:46 AM, Leon Romanovsky wrote:
> > On Fri, Nov 17, 2017 at 02:47:40PM -0500, Doug Ledford wrote:
> >> Hi Stephen,
> >>
> >> Jason Gunthorpe is now co-maintaining the RDMA subsystem with me.  We
> >> have a shared repo that we can both write into.  I would like you to re-
> >> aim the RDMA linux-next pull point to:
> >>
> >> git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma.git tags/for-
> >> next
> >
> > Doug,
> >
> > Does it make sense to get rid of "k.o" notation for for-next and for-rc branches?
> > It is already on kernel.org and seems to me as a redundant.
>
> Jason and I talked about this offline.  The reason it's there is
> actually more technical than anything else.  In order to have WIP
> branches that you want to quickly/easily get on github, but to not have
> them pollute k.o with WIP branches that haven't passed 0day yet, the
> easy solution is to add this to your .git/config for your kernel repo:
>
>         push = refs/heads/k.o/*:refs/heads/k.o/*
>         push = +refs/tags/for-linus*:refs/tags/for-linus*
>         push = +refs/tags/for-next*:refs/tags/for-next*
>         push = refs/heads/master:refs/heads/master
>
> That keeps only the stuff you really want going to k.o going there, and
> everything can then go to your WIP repository.  Going forward, whether
> Jason decides to use a github repo for 0day like I do or not, it now has
> a second benefit in that we can each have our own private namespace,
> which is basically anything not in k.o/, and treat the k.o/ namespace as
> a shared namespace where we worry about branch name conflicts, fast
> forward only updates, etc.  So, it's redundant to the people pulling,
> but I still think it has value for Jason and I.

AS long it works for both of you, we can live with that too.

My scheme is similar to you, but with one repo:
 * master - follow Linus's master
 * rdma-next/rdma-rc - submission queues equivalent to your k.o
 * topic/* - for on-going development
 * testing/* - different experiments

Thanks

>
> --
> Doug Ledford <dledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
>     GPG KeyID: B826A3330E572FDD
>     Key fingerprint = AE6B 1BDA 122B 23B4 265B  1274 B826 A333 0E57 2FDD
>