Re: Shark memory problems
Erik Mouw wrote:
> On Tue, 16 May 2000 22:45:22 +0200 (CES), Alexander Schulz wrote:
> > I am trying to get 2.3.99-pre8 to work, but it crashes rather soon after
> > booting:
> >
> > Unable to handle kernel paging request at virtual address c5fff000
> > -> there is no memory at c5fff000
> > pgd = c0004000
> > *pgd = 00000000, *pmd = 00000000
> > Internal error: Oops: 0
> > CPU: 0
> > pc : [<c0128500>] lr : [<00000100>]
> > ...
If you can copy that info to a file and run ksymoops (with some options to point
it to the right kernel image and System.map, see the manpage), that provides a
lot more debugging info- kind of like a backtrace. But I don't think the stock
tree is the relevant active development tree you want, there seem to be two
separate efforts, the RMK kernels Erik writes about and the Netwinder cvs
repository (which is probably more netwinder-specific). Each of these sends
patches to Linus for inclusion in the official tree, but its state at any given
time is unpredictable at best, and depends on which patches Linus liked.
> > While booting, it says the following:
> >
> > Linux version 2.3.99-pre8 (un23@silvermine) (gcc version egcs-2.91.66
> ^^^^^^^^^^^
> Did you apply Russel King's patch-2.3.99-pre8-rmk1.gz? The ARM patches
> are not yet integrated into the main kernel tree. Get his patch from
>
> ftp://ftp.arm.linux.org.uk/pub/armlinux/source/kernel-patches/v2.3/
Just a touch of feedback FWIW: the cyber2000fb in pre8-rmk1 is telling Xfb that
the display is 15-bit, to which X replies, "Can't do 15-bit packed pixel"; this
wasn't a problem for 16-bit fb in earlier releases (e.g. pre6-rmk2). Either
way, X crashes the system with various problems, such as "frame pointer
underflow".
I'm going to try the netwinder.org cvs sources if I get some time.
Zeen,
-Adam P.
Reply to: