PPP oops on 2.1.26

aidas@ixsrs4.ix.netcom.com
Thu, 20 Feb 1997 07:12:46 -0600 (CST)


I've been running 2.1.26. Lately, however, my phone company has been
messing with the lines and my link is _really_ bad. The thing is filled
with errors, and drops the connection routinely. I have pppd's chat script
program send debug messages to the syslog, so I can sit there and watch it
and tell if the line noise is unusually bad upon login.

Last night after dialing and redialing about 5 times, I got the oops
below. The system ran fine afterwards and ran until a power failure from a
thunderstorm rebooted it this morning.

Unable to handle kernel NULL pointer dereference at virtual address 0000002c
current->tss.cr3 = 00101000, Lr3 = 00101000
*pde = 00000000
Oops: 0000
CPU: 0
EIP: 0010:[<c0146f05>]
EFLAGS: 00010202
eax: c01c5a9c ebx: c1273900 ecx: 000052b0 edx: c1440d98
esi: 00000004 edi: 000052b0 ebp: 00000002 esp: c01c5a68
ds: 0018 es: 0018 ss: 0018
Process swapper (pid: 0, process nr: 0, stackpage=c01c3ba8)
Stack: c12739b0 c18208c0 c18208c0 c18209fc 00000200 c015326a c01c5a9c 00000002
000052b0 00000000 00000004 c1273874 c12739b0 c1440d98 c0150931 c18208c0
c12739b0 c18208c0 c18209fc 00000001 c01c5af0 00000002 00000000 c01519d9
Call Trace: [<c015326a>] [<c0150931>] [<c01519d9>] [<c015195c>] [<c01128cd>] [<c01195fb>] [<c010a4cf>]
[<c0109487>] [<c010a4b8>] [<c010925c>] [<c0109250>]
Code: 8b 46 28 c1 e0 05 31 c1 89 ea 81 e2 f0 f0 f0 f0 c1 ea 04 89
Aiee, killing interrupt handler
kfree of non-kmalloced memory: c01c5bf0, next= 00000000, order=0
kfree of non-kmalloced memory: c01c5be0, next= 00000000, order=0
kfree of non-kmalloced memory: c01c6cf4, next= 00000000, order=0
idle task may not sleep
ixsrs4 last message repeated 4 times
ixsrs4 kernel: Bad pmd in pte_alloc: b0670000
Feb 19 23:29:36 ixsrs4 kernel:
Feb 19 23:29:36 ixsrs4 kernel: Out of memory for chat.
Feb 19 23:29:36 ixsrs4 kernel: zap_pte_range: bad pmd (b0670000)
Feb 19 23:29:36 ixsrs4 kernel: zap_pte_range: bad pmd (50670000)
Feb 19 23:29:36 ixsrs4 kernel: free_one_pmd: bad directory entry 00000a31
Feb 19 23:29:36 ixsrs4 kernel: free_one_pmd: bad directory entry 00000179
Feb 19 23:29:36 ixsrs4 kernel: free_one_pmd: bad directory entry 00000062

Looking through the logs, the exact same error occured on the 18th, and
the 19th.

Running this through ksymoops reveals:

Using `/System.map' to map addresses to symbols.

>>EIP: c0146f05 <ip_route_output+21/fc>
Trace: c015326a <tcp_v4_rebuild_header+3a/98>
Trace: c0150931 <tcp_do_retransmit+101/240>
Trace: c01519d9 <tcp_retransmit_timer+7d/b4>
Trace: c01519d9 <tcp_retransmit_timer+7d/b4>
Trace: c01128cd <timer_bh+2f9/344>
Trace: c01195fb <do_bottom_half+4f/7c>
Trace: c010a4cf <ret_from_sys_call+f/60>
Trace: c0109487 <sys_idle+5b/74>
Trace: c010a4b8 <system_call+38/40>
Trace: c010925c <init>
Trace: c0109250 <start_kernel+200/20c>

Code: c0146f05 <ip_route_output+21/fc> movl 0x28(%esi),%eax
Code: c0146f08 <ip_route_output+24/fc> shll $0x5,%eax
Code: c0146f0b <ip_route_output+27/fc> xorl %eax,%ecx
Code: c0146f0d <ip_route_output+29/fc> movl %ebp,%edx
Code: c0146f0f <ip_route_output+2b/fc> andl $0xf0f0f0f0,%edx
Code: c0146f15 <ip_route_output+31/fc> shrl $0x4,%edx
Code: c0146f18 <ip_route_output+34/fc> movl %eax,(%eax)
Code: c0146f1a <ip_route_output+36/fc> nop
Code: c0146f1b <ip_route_output+37/fc> nop
Code: c0146f1c <ip_route_output+38/fc> nop

--
"Things fall apart; the centre cannot hold. Mere anarchy is loosed upon the
world; The blood-dimmed tide is loosed, and everywhere the ceremony of
innocence is drowned; The best lack all conviction, while the worst are full
of passionate intensity." -- "Second Coming" by Yeats.