RE: Future Linux devel. Kernels

From: Ron Van Dam (rvandam@liwave.com)
Date: Sun May 07 2000 - 13:40:53 EST


>> From my understanding the main reason why the production releases of new
>> kernels is always delayed is because of the "just one more feature"
problem.
>> What about creating two new development kernels (2.5 and 2.7) at the same
>> time and set fixed development goals of the 2.5 kernel before beginning
>> work. Anything outside of the 2.5 goals must go into 2.7.

>I'm completely again this. The main target is a stable kernel, and not to
>make some release date. If that is the target, things are going the WRONG
>way.

 I agree completely with your concern on with release dates and stablity.
But stablity is also dependent on how much new functionally is added. The
more stuff added in at one time (especially when its outside the project
scope) increases the complexity. If goals are established before work is
begun, there is a more likely chance of achieving a better development
cycle. Release dates are totally unimportant, but release goals are.

>> For instance is someone wishes to begin support for a new processor which
>> will most likely take a long time to complete, they can work on the 2.7
>> kernel. Some one that wants to simply add new driver support for a
standard
>> device (such as a new sound card, Ethernet Card, etc) can include their
work
>> in the 2.5 kernel.

>You normale test agains a stable release to be somehow sure that there is
>not a undiscovered bug biting you.

Absolutely. I thought by having a second development kernel for testing
large new features you would have a longer time for testing and tweaking.
Some features will simply required a much longer time to develop and
troubleshoot.

>>...
>> (Why buy multiple devices when I can share!)
>> TCP intercept -- verifies TCP connections before passing on the
>> connection to userland. To prevent DoS and Spoofed attacks.
>> Security integrity checking ( log if the system was booted with a
>> different kernel, log when kernel modules are loaded)
>> Enable Kernel Module signatures so any foriegn kernel modules will be
>> refused. (to avoid Kernel Module hacking).

> If you're that parainoid disable them.

 I thought about that. But some features can only be installed as kernel
modules.

Ron

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Sun May 07 2000 - 21:00:21 EST