2.0.33 bug running data acquisition daemon

Camm Maguire (camm@enhanced.com)
Fri, 30 Jan 1998 10:55:17 -0500 (EST)

Greetings! We have been successfully running a data acquisition
daemon under 2.0.29 for some time. The daemon basically reads a stream
of datagrams over the ethernet interface and stores them in a circular
queue. With 2.0.33, we get the following kernel panic:

-----------------------------------------------------------------------------
Unable to handle kernel NULL pointer dereference at virtual address c0000008
current->tss.cr3 = 00727000, ^_r3 = 00727000
*pde = 00102067
*pte = 00000000
Oops: 0000
CPU: 0
EIP: 0010:[memcpy_toiovec+82/144]
EFLAGS: 00010212
eax: 00001004 ebx: 00000190 ecx: 00000064 edx: bfffea90
esi: 00000008 edi: bfffea90 ebp: 00aaaedc esp: 00aaae28
ds: 0018 es: 002b fs: 002b gs: 002b ss: 0018
Process pcqetix1 (pid: 1741, process nr: 38, stackpage=00aaa000)
Stack: 00150018 00000190 00aaaee4 00755810 00aaaf64 00000190 0013b08b 00aaaedc
00000008 00000190 0014e001 0028fddc 00000008 00aaaedc 00000190 00755810
00aaaed8 00000000 00000000 0028fddc 0012f5f6 00152726 00755810 00aaaf64
Call Trace: [arp_req_set+328/688] [skb_copy_datagram_iovec+27/32] [udp_recvmsg+113/192] [do_select+438/496] [inet_recvmsg+118/144] [sys_recvfrom+327/400] [udp_rcv+955/976]
[ip_rcv+927/1248] [<0101a8c0>] [sys_socketcall+643/848] [system_call+85/128]
Code: f3 a5 83 e3 03 89 d9 f3 a4 07 8b 5c 24 10 01 5c 24 1c 29 5c
-----------------------------------------------------------------------------

The code uses shared memory. We really need this daemon to be very
stable, so any pointers/patches would be most appreciated.

Thanks!

-- 
Camm Maguire			     			camm@enhanced.com
==========================================================================
"The earth is but one country, and mankind its citizens."  --  Baha'u'llah