At the July 2004 kernel summit, it was decided that there was no need to fork a 2.7 kernel [forum] to introduce new functionality into the Linux kernel. Instead, the decision was made that it was possible for Andrew Morton [interview] and Linus Torvalds to continue working together to first merge things into Andrew's -mm tree, and then after testing the changes to merge them into Linus' mainline tree [story]. This of course led to discussion, with some confusion as to how the 2.6 kernel [forum] could be considered stable while new features were still being merged in [story]. During another short discussion nine months after this decision, Rik van Riel [interview] offered some insight into why the new development model works:

"Things get merged one change at a time, and stabilised one change at a time. This is a big change from the even/odd numbered kernel series, where sometimes a bug crops up without anybody knowing exactly what change introduced it. The current development model seems to go much smoother than anything I've seen before."

From: Chuck Wolber [email blocked]
To:  linux-kernel
Subject: Development Model
Date: 	Mon, 18 Apr 2005 22:31:43 -0700 (PDT)


Greetings,

For months I have been reading as much as I can about the current 
stable/unstable development model, but still have a question.

Has the Linux Kernel reached a point where the majority of developers feel 
that (at least for now) no *MAJOR* "rip it out, stomp on it, burn it and 
start over" parts of the kernel exist any longer? In other words, do you 
feel that we're in a good place at this point and that incremental 
improvements will rule the day until $COOL_IDEA comes along and requires a 
refactoring of 2.x (where x is an odd number) porportions?


Thank you,

..Chuck..


P.S. Please CC me on replies, as I am not subscribed at this time.

-- 
http://www.quantumlinux.com 
 Quantum Linux Laboratories, LLC.
 ACCELERATING Business with Open Technology

 "The measure of the restoration lies in the extent to which we apply 
  social values more noble than mere monetary profit." - FDR



From: Arjan van de Ven [email blocked]
Subject: Re: Development Model
Date: 	Tue, 19 Apr 2005 16:50:48 +0200

On Mon, 2005-04-18 at 22:31 -0700, Chuck Wolber wrote:
> Greetings,
> 
> For months I have been reading as much as I can about the current 
> stable/unstable development model, but still have a question.
> 
> Has the Linux Kernel reached a point where the majority of developers feel 
> that (at least for now) no *MAJOR* "rip it out, stomp on it, burn it and 
> start over" parts of the kernel exist any longer? In other words, do you 
> feel that we're in a good place at this point and that incremental 
> improvements will rule the day until $COOL_IDEA comes along and requires a 
> refactoring of 2.x (where x is an odd number) porportions?


actually we have shown (and I like the model very much, it's a great way
to get many features production ready and in the hand of users/customers
really fast) that it doesn't take an odd number release branch to get
major changes in. Instead it takes careful design and sufficient testing
and review and most of the changes can go in anyway.

A good example is the 4 level page tables... we entirely changed how
pagetables worked (eg added another level) midway the 2.6 series and we
could because of the factors I wrote above. It even didn't majorly
destabilize the 2.6 tree to become unusable (although the change was
obviously not pain free for kernel architecture maintainers). 

The VM got replaced as well on the fly. The tty layer got half fixed and
seems to be in progress to be fixed even more. The scsi layer has gone
through an overhaul with the creation of transport classes etc etc. All
on the fly by doing careful design (eg stepwise), testing and review.

If we can do all that... what WOULD warrant a odd numbered release which
would be out of the hands of users for a long time (resulting in
features not being available for the users for a likewise long time,
which in turn means very little testing etc etc)



From: Rik van Riel [email blocked]
Subject: Re: Development Model
Date: 	Tue, 19 Apr 2005 15:24:00 -0400 (EDT)

On Tue, 19 Apr 2005, Arjan van de Ven wrote:

> actually we have shown (and I like the model very much, it's a great way
> to get many features production ready and in the hand of users/customers
> really fast) that it doesn't take an odd number release branch to get
> major changes in. Instead it takes careful design and sufficient testing
> and review and most of the changes can go in anyway.

Perhaps even more importantly, things get merged one change
at a time, and stabilised one change at a time.  This is a
big change from the even/odd numbered kernel series, where
sometimes a bug crops up without anybody knowing exactly
what change introduced it.

The current development model seems to go much smoother than
anything I've seen before.

-- 
"Debugging is twice as hard as writing the code in the first place.
Therefore, if you write the code as cleverly as possible, you are,
by definition, not smart enough to debug it." - Brian W. Kernighan

• Archive of above thread
• KernelTrap interview with Rik van Riel