Updated new SRU policy spec

[Luis R. Rodriguez]

On Fri, Jun 12, 2009 at 9:02 AM, Tim Gardner<tim.gardner at canonical.com> wrote:
> Luis R. Rodriguez wrote:
>> On Tue, Jun 9, 2009 at 10:39 AM, Stefan Bader<stefan.bader at canonical.com> wrote:

>> * What are "OMG kitten killer bugs" ? :)
>>
>
> An expression I acquired from Matthew Garrett. Like pornography, you'll
> know an OMG kitten killer bug when you see it. Is that definite enough?

Hehe.. very clear now.

>> * Can a requirement be that patches must first have an equivalent
>> upstream SHA1 sum on the stable kernel and mention that in the commit
>> log entry? If the concerns is that the patches may take a while to
>> trickle down the stable kernels then how about volunteering someone to
>> join the stable kernel patch review cycle and help with that process?
>>
>
> My intent is that updates to a released non-LTS kernel be throttled down
> to a mere trickle, e.g., CVE and oops fixes only. I want the bulk of the
>  distro kernel dev efforts directed at upstream where I believe we can
> be more effective in the long run.

Great!

Hm, and how about for LTS kernels?

> Any patches that are applied to a released kernel _must_ (with few
> exceptions) originate from upstream, either Linus' tree or one of the
> stable trees (which is effectively the same place).
>
>> * If for whatever reason the above is not possible ensure that at
>> least the patch has been posted and refer to URL for the patch in the
>> commit log entry.
>>
>
> All non-administrative commits to a released kernel _must_ have the
> original upstream SHA1, e.g., 'git cherry-pick -x ...'

Awesome!

  Luis