[Ksummit-2012-discuss] [ATTEND] stable kernel stuff and grumpy maintainers [bisection/rebase/-next]

Jason Wessel jason.wessel at windriver.com
Thu Jun 21 13:59:24 UTC 2012


On 06/21/2012 08:45 AM, Steven Rostedt wrote:
> On Thu, 2012-06-21 at 21:24 +0800, Fengguang Wu wrote:
>> On Thu, Jun 21, 2012 at 08:12:18AM -0400, Steven Rostedt wrote:
>>> On Thu, 2012-06-21 at 11:43 +0800, Fengguang Wu wrote:
>>>
>>>> Strictly speaking, 32 randconfigs may only cover a small fraction of
>>>> the full 32 combination space. That could be something improvable. 
>>>>
>>>
>>> But if you have a large number of people testing randconfigs and
>>> different hardware, that would be much better wont it? A SETI project
>>> for kernel builds ;-)
>>
>> Right. And just count me as one of them ;-)
> 
> Note, this was the motivation for pushing ktest. To get more people an
> easy setup to automate their tests. It wasn't to replace a setup if you
> already had one. It was to get new people to have a drop in tool that
> they can easily run randconfig tests or other things on their kernels.
> 
> It was to try to get that SETI mentality wrt kernel testing.
> 


There are really two groups we need to break into for the mentality.

1) The arbitrary developer who submits something to a maintainer.

   * Updating the "process for contributing" with some more
     recommendations like "do this..." would be good.

2) The maintainers with non-reversible trees

  * Having a maintainer who can get a 24 hour period before etching
    something in stone in non-reversible branches would be a world
    better than what we have today.
    
  * It is not clear if maintainers would bite on this or not, just
    because I would doesn't mean anyone else will.

Jason.


More information about the Ksummit-2012-discuss mailing list