Re: [kbuild-all] [PATCH] locking_selftest: Save/restore migrate_disable_atomic in locking selftest

[Fengguang Wu <lkp@intel.com>]

Hi Sebastian,

On Tue, Dec 01, 2015 at 06:59:44PM +0100, Sebastian Andrzej Siewior wrote:
> On 11/23/2015 04:03 PM, Fengguang Wu wrote:
> > Thanks! Yes, a stable branch name would be better than "linux-4.1.y-rt"
> > that's like to become stable over time.
> 
> finally. I got to it.
> I pushed two branches @ pub/scm/linux/kernel/git/rt/linux-rt-devel.git:
> 
> 	for-kbuild-bot/current-stable [0]
> 	for-kbuild-bot/prepare-release [1]
> 
> Branch [0] should contain the last -RT release. This could be used for
> testing patches against (if you find one with the RT marker in subject).
> The tree should start a stable-tree marker (currently it is v4.1.13)
> and have -RT tree applied on top. You should be able compile after each
> commit (between the stable tag and HEAD) and nothing should introduce
> warnings or fail to compile.

Got it, I've updated the RT => for-kbuild-bot/current-stable mapping
accordingly, thanks for the info!

> The second branch [1] would be similar to the first one except that I
> plan to push stuff there before I make a release it. Does this make
> sense or do I over think this?

It's fair enough.

> If you do compile tests, it would be nice if you could enable
> CONFIG_PREEMPT_RT_FULL. That is where most of the changes start to work.
> Nevertheless it should also work without it(i.e. no preemption or
> desktop).

OK, I'll increase testing for CONFIG_PREEMPT_RT_FULL when it's the RT tree.

> If you have slightly different naming scheme or suggestions just tell
> me I will adapt to it:)
> 
> Thank you for the service.

You are welcome!

Thanks,
Fengguang