Latest FC kernel still have SMP bugs?

[Jay Daniels]

On Fri, Apr 09, 2004 at 02:03:06AM -0400, Jay Daniels wrote:
> On Thu, Apr 08, 2004 at 10:21:46AM -0500, Jason L Tibbitts III wrote:
> > >>>>> "DJ" == Dave Jones <davej at redhat.com> writes:
> > 
> > DJ> For now I've just turned off the lowlat patch completely in
> > DJ> CVS. Its been a while since there's been a FC1 kernel update, so
> > DJ> there's quite a bit of stuff pending, so I'll try and get out an
> > DJ> update in the next week or so.
> > 
> > Thanks; I had assumed that you folks were working exclusively on 2.6
> > now but it's good to know that at least one more FC1 kernel update
> > will be forthcoming.
> > 
> > Folks who need this fixed now but who want to keep everything in RPM
> > are welcome to pull the packages I built from
> > http://www.math.uh.edu/~tibbs/kernel-packages or follow the
> > instructions posted in bug 109962 and build your own.  It's not
> > difficult.
> > 
> >  - J<
> 
> Downloading, it's late - will try your kernel-smp in the morn.
> 
> thanks,
> 
> jay


Didn't work, hanged while loading modules just like the stock fedora
core 1 smp kernel!

Where do I need to look to find out what the problem is with this dual
xeon processor box that won't run smp kernels?

Stock single processor kernel works great.  All smp kernels I've tried
on this box crashes just after checking /tmp and I believe it's
happening when loading modules.

It's a supermicro superserver board with dual xeon 2.66 ghz and 2GB of
ram.  The board model is a MBD-X5DEI-GG-0

Has anyone tested the latest Fedora Core 2 test smp kernels on such a
system?   If that those smp kernels work, I see no need to compile
from source.

Also note that the Redhat 9.0 smp kernel worked fine.


jay