Re: [GIT PULL] bcache changes for 3.17

[Jens Axboe <axboe@kernel.dk>]

On 09/05/2014 03:45 PM, Greg KH wrote:
> On Fri, Sep 05, 2014 at 09:31:06AM +0200, Francis Moreau wrote:
>> On 08/10/2014 09:54 AM, Peter Kieser wrote:
>>>
>>> On 2014-08-05 9:58 AM, Jens Axboe wrote:
>>>> On 08/04/2014 10:33 PM, Kent Overstreet wrote:
>>>>> Hey Jens, here's the pull request for 3.17 - typically late, but lots of tasty
>>>>> fixes in this one :)
>>>> Normally I'd say no, but since it's basically just fixes, I guess we can
>>>> pull it in. But generally, it has to be in my hands a week before this,
>>>> so it can simmer a bit in for-next before going in...
>>>>
>>> Are these fixes going to be backported to 3.10 or other stable releases?
>>>
>>
>> Could you please answer this question ?
>>
>> If you don't want to maintain bcache for stable kernels (I can
>> understand that), can you mark it at least as unstable/experimental
>> stuff since it really is ?
> 
> WTF?
> 
> Just because a maintainer/developer doesn't want to do anything for the
> stable kernel releases does _NOT_ mean the code is
> "unstable/expreimental" at all.

That's not what he is saying at all. The code IS unstable in 3.10. And
the fact that nothing goes to stable for bcache, the situation wasn't
likely to change for 3.10. Nobody is saying "Oh nothing goes to stable,
lets mark it experimental".

> That's not how stable kernel releases work.  _IF_ a maintainer wants to
> / has the time to, they can mark patches for inclusion in stable kernel
> releases.  Given the huge list of patches that Jens just posted, I doubt
> that those are really something I would ever take for a stable kernel
> release.

Actually, all of those are pretty much stable material, since they fix
actual bugs that people hit. Which is the definition of what should go
to stable.

-- 
Jens Axboe