Re: 2.5 kernel hangs system

Murray J. Root (murrayr@brain.org)
Tue, 22 Apr 2003 22:31:28 -0400


On Tue, Apr 22, 2003 at 06:55:11PM -0700, Randy.Dunlap wrote:
> > On Wed, Apr 23, 2003 at 01:11:46AM +0200, Felipe Alfaro Solana wrote:
> >> On Tue, 2003-04-22 at 23:00, Richard B. Johnson wrote:
> >> > First, I don't understand how as you say, "suggestions are
> >> > desperately needed" on a developmental kernel. These things are
> >> > not known to work on all configurations and some information like
> >> > "It gives me hex codes..." is worthless. Please write down
> >> > these "hex-codes" and, after booting a version the works, run them
> >> > through ksymoops. If you don't know what that is:
> >>
> >> ksymoops? I thought 2.5 kernels didn't need ksymoops anymore and that
> >> function names were automatically "guessed" in call stack traces.
> >>
> >
> > IFF you use "include symbols" when building you shouldn't need ksymoops.
> > IMO, if you're using 2.5.x you really should include the symbols - chances
> > are you'll need em.
>
> Maybe we are reading this differently, but it sounded to me like the
> original system hang never reached the kernel | system log and that
> some hex codes were the only clues. In that case, pushing them thru
> ksymoops does still make some sense, doesn't it?
> How else would you determine where the hang occurred?
>
I read it that he was getting an oops with just hex values cause
he didn't have the symbols included (note the reference to "it
just gives hex codes"). I could be wrong, since I'm only looking
at the quote of a quote of someone's interpretation. :)
With symbols included an oops DOES print the actual call trace in
readable form. (I've seen enough oops to be certain of that :)

-- 
Murray J. Root

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