Re: how to proceed with a driver problem
On Wed, 2009-04-08 at 14:31 -0600, dann frazier wrote:
> On Wed, Apr 08, 2009 at 01:10:16PM -0700, Ross Boylan wrote:
> > It appears that the ath5k wireless driver is crashing my system. It
> > leaves almost no info in the logs when it does so (and thus has taken
> > several weeks to find).
> >
> > Should I report that as a bug against the appropriate kernel image (the
> > standard Lenny amd64), or take it directly upstream?
> >
> > Also, if anyone has any advice on how to proceed, even just how to get
> > more info about the failures, would be great.
> >
> > The system is new hardware (2 CPU x 4 core each Xeon) running a fresh
> > install of Lenny.
> > Linux version 2.6.26-1-amd64 (Debian 2.6.26-13)
> > it does not (yet) have the latest security fixes.
> > I'm not actually using the wireless--it doesn't have the password for
> > the local network. I may need the wireless later.
> >
> > The system stayed up over a week with ath5k unloaded and blacklisted;
> > before it was crashing c. daily. I just loaded the driver, and the
> > system crashed about 20 hours later.
> >
> > I'd appreciate if you cc'd me on the response.
>
> I'd suggest trying with the latest available trunk snapshot kernel
> (see http://wiki.debian.org/) and, if its still causing a problem,
> report upstream at bugzilla.kernel.org.
I've done that, but aptitude said it was holding linux-libc-dev at the
current level. Is that a problem? Will there be an issue with any of
my other libs working with the new kernel?
Reading the kernel handbook makes me wonder why no low level libraries
are not necessary for each ABI version of the kernel. I thought some,
like libc, did make kernel calls internally and so would be sensitive to
changes.
--
Ross Boylan wk: (415) 514-8146
185 Berry St #5700 ross@biostat.ucsf.edu
Dept of Epidemiology and Biostatistics fax: (415) 514-8150
University of California, San Francisco
San Francisco, CA 94107-1739 hm: (415) 550-1062
Reply to: