Oops in kernel boot on a 386 with a 3c507 (el16) using 2.1.61

Greg Alexander (galexand@sietch.bloomington.in.us)
Mon, 3 Nov 1997 01:26:42 -0500 (EST)

I can provide .config, etc., on request.
Description of the problem:
I get this oops right after identifying the partitions on my single hd:
Unable to handle kernel paging request at virtual address 000dfff6
current->tss.cr3 = 00101000, %cr3 = 00101000
*pde 00000000
Oops: 0002
CPU: 0
EIP: 0010:[<c019bba1>]
EFLAGS: 00010207
eax: c0002620 ebx: 000dfff6 ecx: 00000002 edx: 000d0000
esi: c01c7ea0 edi: 000dfff6 ebp: 000d0000 esp: c0007ed0
ds: 0018 es: 0018 ss: 0018
Process swapper (pid: 1, process nr: 1, stackpage=c0007000)
Stack: c01c7c9c c01c7c9c 00000000 c01cfa6c 00000001 00000001 c00026e0 c0000306
c0000300 c019b67b c01c7c9c c01c7c9c c01551d6 c01c7c9c c01c7c9c c01b0bbe
c01a4bfa c01cb1c1 c01c7c9c ffffffff c0007fe4 c0201228 00009000 00000002
Call Trace: [<c019b67b>] [<c01551d6>] [<c01b0bbe>] [<c01a4bfa>] [<c01a4bf1>] [<c013e4c4>] [<c01a4bed>]
[<c01a4bed>] [<c0179b97>] [<c01a4bed>] [<c012dbd0>] [<c01092da>] [<c011fbfc>] [<c01a0018>] [<c01080d3>]
[<c0107fdc>] [<c0108084>]
Code: f3 a5 66 a5 8b 44 24 28 8b 58 10 b9 15 00 00 00 89 df be aa
--------------------------------------------------------------------------
ksymoops output:
Using `../System.map' to map addresses to symbols.

>>EIP: c019bba1 <init_82586_mem+65/15c>
Trace: c019b67b <el16_open+b/24>
Trace: c01551d6 <dev_open+16/70>
Trace: c01b0bbe <NR_TYPES+3a6/875>
Trace: c01a4bfa <tvecs+4e0e/772f>
Trace: c01a4bf1 <tvecs+4e05/772f>
Trace: c013e4c4 <create_proc_entry+fc/10c>
Trace: c01a4bed <tvecs+4e01/772f>
Trace: c01a4bed <tvecs+4e01/772f>
Trace: c0179b97 <rpc_proc_register+2f/40>
Trace: c01a4bed <tvecs+4e01/772f>
Trace: c012dbd0 <sys_setup+40/44>
Trace: c01092da <system_call+3a/40>
Trace: c011fbfc <kswapd>
Trace: c01a0018 <tvecs+22c/772f>
Trace: c01080d3 <init+4f/1ac>
Trace: c0107fdc <this_must_match_init_task+1fdc/2000>
Trace: c01080d3 <init+4f/1ac>
Code: c019bba1 <init_82586_mem+65/15c>
Code: c019bba1 <init_82586_mem+65/15c> f3 a5 repz movsl %ds:(%esi),%es:(%edi)
Code: c019bba3 <init_82586_mem+67/15c> 66 a5 movsw %ds:(%esi),%es:(%edi)
Code: c019bba5 <init_82586_mem+69/15c> 8b 44 24 28 movl 0x28(%esp,1),%eax
Code: c019bba9 <init_82586_mem+6d/15c> 8b 58 10 movl 0x10(%eax),%ebx
Code: c019bbb2 <init_82586_mem+76/15c> b9 15 00 00 00 movl $0x15,%ecx
Code: c019bbb7 <init_82586_mem+7b/15c> 89 df movl %ebx,%edi
Code: c019bbb9 <init_82586_mem+7d/15c> be aa 00 90 90 movl $0x909000aa,%esi
Code: c019bbc4 <init_82586_mem+88/15c> 90 nop
-------------------------------------------------------------------
Description of the kernel:
this is a kernel built for 386 with support for 3c507. It is set to boot
off of NFS root off of a 2.0.31 machine. Everything about this setup was
fine with 2.0.30, but I figured I'd upgrade the 386 because it was
completely non-critical and I might find something. This is a fairly
minimal kernel. It also has the NFS server enabled in make menuconfig (for
the heck of it).

Description of computer:
CompuAdd 386SX/25MHz, with 387. Booting with lilo off of a small HD. It
has a 3c507 that worked fine (and still works with 2.0.30), but has had some
various problems that went away for not completely understood reasons.

If no one else feels like chasing this down because it's specific to the
3c507 or whatever, I'll probably go ahead and look into it next weekend.
But I figured that since I'm very unfamiliar with 2.1.x, someone else might
immediately see, "d'oh, I just knew that would break that. I know how to
fix it." If anyone has any vague ideas as to where this problem might be
originating, please let me know.
Thanks.

Greg Alexander - also <gralexan@indiana.edu> - http://sietch.home.ml.org/
----
Trashed filesystems are child's play -- if the hardware ain't smoking,
IT'S RUNNING TOO SLOW!
-- John Kelly, linux-kernel@vger.rutgers.edu contributor