> On Mon, 24 Feb 1997, Bernd Schmidt wrote:
>
> > I'd like to agree, but I've seen another bug report like this, and I noticed
> > that in both cases, the function elf_core_dump appeared in the stack trace.
> > I even still have the original mail. The author (Johnny Stenback) claims the
> > problem is reproducible with the following test program:
>
> I've run Bernd's program several times now and am fairly convinced this is
> a kernel problem...not bad RAM. I got almost identical ksymoops data from
> both the user's mud and Bernd's program. I've upgraded to 2.0.29, and
> still get very similar results from Bernd's program. First run segfaults,
> additional runs go to state D and none can be killed...so I had to reboot
> again.
Very weird.. I ran it, too, and it simply segfaults, no oops, no state D,
no problem whatsoever.
There are two possibilities here:
1. file "test" doesn't exist, so mmap returns -1 (??? it should return
NULL), the first attempt to dereference the pointer segfaults, end of
story.
2. file "test" does exist, mmap returns a valid pointer, the first attempt
to read from that memory region segfaults because of PROT_WRITE protection
(no reading allowed). End of story.
Kernel 2.0.29 with minimal modifications, libc 5.4.23.
Any other takers?
Ionut
-- It is better to keep your mouth shut and be thought a fool, than to open it and remove all doubt.