On Mon, Mar 22, 1999 at 04:47:06PM -0500, Ulf Carlsson wrote:
> I'm having some problems with gdb for mips. I can't examine core files
> since gdb itself dumps core files when I try start gdb (gdb nsgsml core).
> Are there other ways to check where a program crashes (except for printf()s)?
> It isn't possible to run nsgsml in gdb either, or at least it doesn't
> provide the information I want.
hmm, if memory serves right, I had reading core working at some point.
But I just tried it again and it fails like you described. I debugged
gdb some time ago and it's was a real nightmare. IMHO fixing gdb isn't
a short term task and I'm right now trying to plug all SCSI holes
(two bugs smashed, but still too much known bugs unfixed). And after
the SCSI stuff I'll port the RedHat Rawhide installer, so I'm not
available to fix gdb for some time.
For a short term solution edit arch/mips/mm/fault.c and look for the
second #if 0 (after the label bad_area) and change that to a #if 1.
This will give you a kernel message for every segfault, but it won't
help you for bus errors (but it's also possible to add printks there, too).
This device has completely bogus header. Compaq scores again :-|
It's a host bridge, but it should be called ghost bridge instead ;^)
[Martin `MJ' Mares on linux-kernel]