On Sat, 10 Apr 1999, Doug Ledford wrote:
> Alan Cox wrote:
> >
> > > This machines performs fairly well, despite having only 2 PII350's. Here is
> > > the problem: Only one processor is enabled. I've called Dell, and they say no
> > > support, and to call RH. Called them, no support either. So, I'm turning to
> > > alternative resources: Anyone have any suggestions for where to find
> > > documentation pertaining to compiling a Kernel, and/or, compiling a kernel
> > > w/SMP support? Any help would be greatly appreciated.
> >
> > If you paid Dell to install Linux properly on a box and they missed little
> > details like SMP support on a multiprocessor machine you should return it as
> > faulty. Go to one of the Linux hardware vendors and ask them to do the job
> > properly for you.
>
> To be fair to Dell, they sell Red Hat Linux 5.2 pre-installed on these
> systems. If Red Hat Linux 5.2 doesn't support SMP, then neither do
> they. Expecting something different from Dell is no different than
> expecting Dell to ship an SMP version of Win98. To that extent, Dell
> not shipping an SMP capable kernel is because we haven't released one
> for Red Hat Linux 5.2 and is our failing. The fact that they will still
> sell you multiple CPUs is in anticipation of being able to use the extra
> CPUs if you want to do a manual kernel compile or the day that we
> release an SMP kernel RPM that can be applied to the machine.
I don't think you can 'be fair' to Dell here. This is a totally stupid
policy, and would be just a dishonest as selling a quad xeon box with
win98 installed. And then letting _you_ figure out how to use all four
cpus (hint: spelled win NT server).
If they sell a dual box, with a non-SMP linux kernel installed... that
is totally stupid... And it will cause a little dissatisfaction with
Linux when the problem is the morons at Dell that are shipping such
nonsense.
>
> We decided not to ship an SMP kernel with Red Hat Linux 5.2 because we
> knew that there were some problems with 2.0.36 SMP. There were lots of
> people that could do OK with it, but then there were other systems that
> simply wouldn't work at all. Then there were the occasional lock up
> problems. Then there was the occasional SCSI sub-system goes belly up
> problem with my driver that's in the stock 2.0.36 (which I fixed in the
> 5.1.12 driver version just recently).
>
> Originally, if we released a 2.0.36 SMP kernel RPM, people would be mad
> that we shipped something that wasn't 100% "up to snuff" so to speak. I
> would also feel like we fell down if we did that. People rely on us to
> not only package things up, but to try and reasonably make sure those
> packages work. Shipping a known busted package would violate that
> expectation. At this point in time, with all of the patches there are
> for 2.0.36 it *might* be possible to do an SMP kernel and finally feel
> good about it.
>
all good points... but it doesn't excuse dell shipping a SMP box that
won't do "SMP" stuff. that's insane and a good way to lose customers.
> --
> Doug Ledford <[EMAIL PROTECTED]>
> Opinions expressed are my own, but
> they should be everybody's.
> -
> Linux SMP list: FIRST see FAQ at http://www.irisa.fr/prive/mentre/smp-faq/
> To Unsubscribe: send "unsubscribe linux-smp" to [EMAIL PROTECTED]
>
-
Linux SMP list: FIRST see FAQ at http://www.irisa.fr/prive/mentre/smp-faq/
To Unsubscribe: send "unsubscribe linux-smp" to [EMAIL PROTECTED]
