lkml.org 
[lkml]   [1997]   [Feb]   [24]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
SubjectRe: Unable to handle kernel paging request
Date
> > > Unable to handle kernel paging request at virtual address 40007000
> > > current->tss.cr3 = 01f84000, %cr3 = 01f84000
> > > *pde = 00ec5067
> > > *pte = 0013bc00
> > > Oops: 0000
> > > CPU: 0
> > > EIP: 0010:[<00157b45>]
> > > EFLAGS: 00010216
> > > eax: 00986bb0 ebx: 00000400 ecx: 00000100 edx: 016bd400
> > > esi: 40007000 edi: 016bd400 ebp: 00000400 esp: 01384c08
> >
> > Flipped 14th bit in esi, most likely. In other words, memory error.
>
> 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:

Yes, this program still causes an oops the first time it is run in
2.0.29, the second time it is run it hangs and can't be killed (it
still causes the oops though). I don't have the knowledge to fix this
but one thing I did notice is that the program works as expected (ie
it just segfaults) in 2.1.26, in earlier 2.1.xx kernels (the really
early ones) it segfaulted and caused a kernel message saying
"Exception at XXXXXXXX (YYYYYYYY)" so apparently the fix for this
problem might even be in one of the 2.1.xx patches...
Hope this helps...

>
> #include <stdio.h>
> #include <unistd.h>
> #include <sys/mman.h>
> #include <sys/types.h>
> #include <sys/stat.h>
> #include <fcntl.h>
> #include <string.h>
> #include <errno.h>
>
> void main ()
> {
> int fd;
> void *p;
> char buf[1024];
> fd = open("test", O_RDONLY);
> p = mmap(NULL, 200, PROT_WRITE, MAP_PRIVATE, fd, 0);
> strncpy (buf, p, 100);
> printf ("%s\n", buf);
> munmap (p, 200);
> exit (0);
> }
>
> The "Oops" from the other post is mostly identical-looking.
> (I haven't tried this, so if it doesn't work, I apologize. I'll try today)

Be careful if you try this, the system is unable to umount all
partitons cleanly after this so you never know what might happend...

> Of course, you'll have to enable core dumps.
>
> Bernd


PS. Sorry Bernd if you recieve several messages about this, I sent you
an other one that I ment to cc to linux-kernel but I think I forgot
that so I'm sending another one now...

--
Johnny Stenback, programmer / University of Vaasa, Computer Centre
E-Mail jst@uwasa.fi, Phone +358 50 5575 094

\
 
 \ /
  Last update: 2005-03-22 13:39    [from the cache]
©2003-2011 Jasper Spaans