While using xjed -- a X editor, which has done any
segfaults before -- I have got a kernel oops:
The part of the syslog-file says:
---8<--------------------------------------------------------------
Unable to handle kernel paging request at virtual address 1a657c2c
current->tss.cr3 = 0125e000, <r3 = 0125e000
*pde = 00000000
Oops: 0000
CPU: 0
EIP: 0010:[<c012f485>]
EFLAGS: 00010206
eax: 1a657c00 ebx: c0ee1cc0 ecx: c14ed500 edx: c10fe3ef
esi: c0ee1cc0 edi: c0f62180 ebp: 00000001 esp: c117df50
ds: 0018 es: 0018 ss: 0018
Process xjed.bin (pid: 2108, process nr: 20, stackpage=c117d000)
Stack: c0ee1cc0 c1ff401c c0f62180 c012f64b c0f62180 c0ee1cc0 c1ff4000 00000001
00000001 bffff56c c14ed500 00001001 c1ff4014 00000008 08533825 c012f6ac
c1ff4000 00000000 00000001 00000000 080bdc48 c012d267 080bdc48 00000001
Call Trace: [<c012f64b>] [<c012f6ac>] [<c012d267>] [<c01098fa>]
Code: 83 78 2c 00 74 57 89 e0 89 c3 81 e3 00 e0 ff ff 8b 83 cc 01
---8<--------------------------------------------------------------
The output of the ksymoops-tool produces:
---8<--------------------------------------------------------------
Using `/boot/System.map' to map addresses to symbols.
>>EIP: c012f485 <do_follow_link+19/88>
Trace: c012f64b <lookup_dentry+157/190>
Trace: c012f6ac <__namei+28/80>
Trace: c012d267 <sys_newstat+7f/f4>
Trace: c01098fa <system_call+3a/40>
Code: c012f485 <do_follow_link+19/88>
Code: c012f485 <do_follow_link+19/88> 83 78 2c 00 cmpl $0x0,0x2c(%eax)
Code: c012f489 <do_follow_link+1d/88> 74 57 je c012f4e2 <do_follo
w_link+76/88>
Code: c012f48b <do_follow_link+1f/88> 89 e0 movl %esp,%eax
Code: c012f48d <do_follow_link+21/88> 89 c3 movl %eax,%ebx
Code: c012f495 <do_follow_link+29/88> 81 e3 00 e0 ff andl $0xffffe000,%ebx
Code: c012f49a <do_follow_link+2e/88> ff
Code: c012f49b <do_follow_link+2f/88> 8b 83 cc 01 00 movl 0x900001cc(%ebx),%
eax
Code: c012f4a6 <do_follow_link+3a/88> 90
Code: c012f4a7 <do_follow_link+3b/88> 90 nop
Code: c012f4a8 <do_follow_link+3c/88> 90 nop
---8<--------------------------------------------------------------
Again, the sum of MemFree, MemShared, Buffers, Cached of /proc/meminfo
does not match MemTotal -- again a loss of some MBytes...
(if this has something do with the oops above -- I don't know. But until now,
I've got no answer to my question to this loss-of-memory effect...)
Hardware:
32MByte "normal" RAM (no EDO-magics)
P90 Intel on a Intel Plato-board
SCSI-ncr53c815-Adaptor
two harddiscs 1.5GigaByte sum
kernel is configured to use modules
KEEP HACKING!
meino
----------------------------------
E-Mail: Meino Christian Cramer <mccramer@solfire.ludwigsburg.netsurf.de>
Date: 17-Aug-97
Time: 08:11:29
This message was sent by XFMail
----------------------------------