Kernel Oops! (info attached) 2.4.x+EtherTalk+LPRng

Disclaimer: IANAKH (I am not a kernel hacker.)

Oops! SYMPTOM:

Printing to an EtherTalk printer using netatalk-1.5pre6 and kernel 2.4.[2,5,6,7] generates an Oops! (Oops! and related details attached at end of message.)

Configuration (software):

2.4 kernels (2.4.2 and 2.4.5-2.4.7)
netatalk-1.5pre6
LPRng 3.7.4
glibc 2.2
gcc 2.96

Configuration (hardware):

Athlon-C 1.0GHz (not overclocked)
Asus A7V133 + 768MB premium brand RAM
AMD PCNet32 PCI Ethernet*
(*device is behind a DEC 21152 PCI bridge)
Apple LaserWriter IIg on EtherTalk

CPU0

  0:     351613          XT-PIC  timer
  1:       6651          XT-PIC  keyboard
  2:          0          XT-PIC  cascade
  5:      28495          XT-PIC  sym53c8xx, PCnet/FAST 79C971
  9:        105          XT-PIC  bttv, sym53c8xx, usb-uhci, usb-uhci
 10:         61          XT-PIC  sym53c8xx
 11:     333887          XT-PIC  es1371, nvidia
 12:     102148          XT-PIC  PS/2 Mouse
 14:        132          XT-PIC  ide0
 15:         79          XT-PIC  ide1
NMI:          0
ERR:          0

OTHER INFO:

I didn't experience this problem under 2.2 kernels, though of course this proves nothing... System is under moderate load and may enjoy uptimes of several weeks under heavy memory pressure without problems -- until I try to print via EtherTalk.

Usually the first spooled job prints okay; it is on the second job that the OOPS usually seems to occur and the system is then completely wedged... Magic SysRq doesn't work, have to reset hard.

Thanks in advance for help, anyone! I'm not subscribed to the list, but I watch the archives daily.

-------------------------------------------

OOPS (already ksymoops'ed) DETAILS:

Unable to handle kernel paging request at virtual address fffffffc
c0110633
*pde = 00001063
Oops: 0000
CPU:    0
EIP:    0010:[<c0110633>]

Using defaults from ksymoops -t elf32-i386 -a i386
EFLAGS: 00010017
eax: e8afec28 ebx: 00000000 ecx: 00000001 edx: e8afec2c
esi: e9c22200 edi: fffffff8 ebp: c02e1f1c esp: c02e1f00
ds: 0018 es: 0018 ss: 0018
Process swapper (pid: 0, stackpage=c02e1000)

Stack: e8afec2c 00000001 00000286 00000001 e8afec28 e9c22200 fffffffd 00000046
       c821b106 e8ba0000 c02dcd00 ee978d40 00000000 e9c22200 ee978140 c021a7f1
       e9c22200 ee978440 ee97849c c021b60b 00000287 00000000 c021b6d5 ee978140
Call Trace: [<c021b106>] [<c021a7f1>] [<c021b60b>] [<c021b6d5>] [<c021ec82>] [<c011651b>] [<c01080ec>]
       [<c01051a0>] [<c01051a0>] p<c0106d54>] [<c01051a0>] [<c01051a0>] [<c01051c3>] [<c0105242>] [<c0105000>]
Code: 8b 4f 04 8b 1b 8b 01 85 45 f0 74 4f 31 c0 9c 5e fa c7 01 00

Trace; c021b106 <sock_def_write_space+36/90>
Trace; c021a7f1 <sock_wfree+21/40>
Trace; c021b60b <kfree_skbmem+b/60>
Trace; c021b6d5 <__kfree_skb+75/120>
Trace; c021ec82 <net_tx_action+52/c0>
Trace; c011651b <do_softirq+4b/90>
Trace; c01080ec <do_IRQ+9c/b0>
Trace; c01051a0 <default_idle+0/30>
Trace; c01051a0 <default_idle+0/30>
Code; c0110633 <__wake_up+33/a0>
00000000 <_EIP>:
Code; c0110633 <__wake_up+33/a0> <=====
0: 8b 4f 04 mov 0x4(%edi),%ecx <=====
Code; c0110636 <__wake_up+36/a0>
3: 8b 1b mov (%ebx),%ebx
Code; c0110638 <__wake_up+38/a0>
5: 8b 01 mov (%ecx),%eax
Code; c011063a <__wake_up+3a/a0>
7: 85 45 f0 test %eax,0xfffffff0(%ebp)
Code; c011063d <__wake_up+3d/a0>
a: 74 4f je 5b <_EIP+0x5b> c011068e <__wake_up+8e/a0>
Code; c011063f <__wake_up+3f/a0>
c: 31 c0 xor %eax,%eax
Code; c0110641 <__wake_up+41/a0>
e: 9c pushf
Code; c0110642 <__wake_up+42/a0>
f: 5e pop %esi
Code; c0110643 <__wake_up+43/a0>
10: fa cli
Code; c0110644 <__wake_up+44/a0>
11: c7 01 00 00 00 00 movl $0x0,(%ecx)

Kernel panic: Aiee, killing interrupt handler!