Re: v2.4.14 kernel compile problem
On Tue, Nov 06, 2001 at 08:53:51AM -0600, John Patton wrote:
> On Tue, Nov 06, 2001 at 02:11:52PM +0800, Patrick Cheong Shu Yang wrote:
> > Another one of those release many and release quick again....
> >
> > Uuughh...
> >
> > I remember someone once said regarding the impending release of 2.4.0
> > and why it was taking so long...that Linus just has much higher
> > standards...oh oh...this is definitely not in the correct direction in
> > support of the previous comment...
>
> The thing is that they tried to accomplish ALOT in 2.4... everything
> that they did is good and will lead to the linux kernel being a
> genuinely excellent kernel that can rightly compete with commercial
> unices, but all of the changes has introduced some new bugs that need
> to be worked out. The thing is, do you want a super stable toy kernel
> that works well for hobbyists and low-end desktop machines (one that is
> simple but inefficient)? Or do you want a kernel that strives to be the
> best... one that runs efficiently and well even under high load? The
> 2.2 series is quite stable, but it cannot compete with most commercial
> unices or even freeBSD on high end machines, while the 2.4 kernel has
> gone a long way towards being able to do so. But high end power and
> efficiency comes at the cost of much greater complexity, which is that
> much harder to maintain flawlessly. Also keep in mind that 2.4 is being
> actively worked on (despite it's declaration of being stable), and so
> it cannot be counted on to run perfectly just yet. Give it some time...
> Linus' standards are both high and ambitious, and when the smoke clears
> the kernel should be really top notch.
>
Although I appreciate the enormous work that has been done during the
development of the 2.4 kernel to bring it nearer to enterprise level
(it has already reached this level in many aspects) I'm not sure if
this is the right way to accomplish this goal.
I had problems with kernel versions 2.4.9 to 2.4.12 (once there were
errors while compiling NTFS support, another time there were problems
with the parallel port, and finally a lot of VMM errors in version
2.4.10). Therefore I've just made the change from 2.4.8 to a running
kernel 2.4.13. But now it seems this story will continue with 2.4.14.
I never had these problems when changing to a new kernel version of
the 2.2 series.
So with all these problems in differnet 2.4 releases and keeping in
mind that it's been actively worked on kernel 2.4 (as you wrote
before) I wouldn't call it stable. Maybe it would be better to do this
work on an unstable tree 2.5. But I think the Debian mailing list is
not the right place for such discussions.
Finally I agree with your conclusion: In the end the kernel will be
top notch and everybody - including me - will be happy :-)
Bye
Andreas Schmidt
Reply to: