Technical article 105866

Token Ring network system panics showing "dlpi_unlink" function in crash trace.

OpenServer 5.0.2

Written · Last revised

Problem

I have several SCO OpenServer 5.0.2 systems running TCP/IP over a Token Ring network. The systems run several different manufacturers' Token Ring cards including Madge, IBM, and 3COM. On several occasions, at various times of the day, and seemingly unconnected to network loading, the systems have panicked. Running crash(ADM) on these panic dumps shows the following:

> panic
System Messages:
...
...
...
TDC 3800
Unexpected trap in kernel mode:
cr0 0x8001001B     cr2  0x00040150     cr3 0x00002000     tlb
0x00000000
ss  0x00000B9C     uesp 0x00000000     efl 0x00010202     ipl
0x00000001
cs  0x00000158     eip  0xF00BB8F1     err 0x00000003     trap
0x0000000E
eax 0xF0264438     ecx  0x0004014C     edx 0x3381C6F5     ebx
0xFD2D3F8C
esp 0xE0000B64     ebp  0xE0000B7C     esi 0xF0264418     edi
0x00000000
ds  0x00000160     es   0x00000160     fs  0x00000000     gs
0x00000000
cpu 0x00000001
PANIC: k_trap - Kernel mode trap type 0x0000000E
Trying to dump 20383 pages to dumpdev hd (1/41), 255 pages
per '.'
..............

Panic String: k_trap - Kernel mode trap type 0x%x

Kernel Trap.  Kernel Registers saved at 0xe0000b34
ERR=3, TRAPNO=14
cs:eip=0158:f00bb8f1 Flags=10202
ds = 0160   es = 0160   fs = 0000   gs = 0000
esi= f0264418   edi= 00000000   ebp= e0000b7c   esp= e0000b64
eax= f0264438   ebx= fd2d3f8c   ecx= 0004014c   edx= 3381c6f5
Kernel Stack before Trap:
STKADDR   FRAMEPTR  FUNCTION   POSSIBLE ARGUMENTS
e0000b64  e0000b7c  dlpi_unlin (net1route_table,0,
net1route_table+0x20,net1route_table+0x24)
e0000b84  e0000b9c  dlpi_get_n (net1route_table,net1route_
table+0xad8)
e0000ba4  e0000bac  dlpi_reloc (net1route_table,net1route_
table+0xad8,0,net1route_table)
e0000bb4  e0000bc4  dlpi_get_n (net1route_table,0x21,
0xfd2d3f30,net1cardinfo)
e0000bcc  e0000bec  dlpiSR_aut (net1route_table,u+0xc44,0x3,
u+0xc44)
e0000bf4  e0000c08  dlpi_send_ (net1sapinfo+0xd0,u+0xc44,0,
net1sapinfo+0xd0)
e0000c10  e0000c28  dlpi_llc_r (net1sapinfo+0xd0,u+0xc44,
0xfcee4ea0,0xfd2d3f30)
e0000c30  e0000c7c  mdi_primit (0xfcee4ea0,0xfd2d3f30,
0xfcee4ea0,0xfd2d3f30)
e0000c84  e0000c94  mdi_messag (0xfcee4ea0,0xfd2d3f30,0x5,0)
e0000c9c  e0000cac  dlpilrsrv  (0xfcee4ea0,0x1,0,0x7)
e0000cb4  e0000cd0  queuerun   (curpri,0,0x1)
e0000cd8  e0000ce4  runqueues  (0x7,0,proc+0x87b8)
e0000cec  e0000cf8  pswtch     (sleep+0x20d,0x3e8,0,u+0x1148)
e0000d00  e0000d1c  swtch      (selwait,0x11a,0,sysent+0x2b8)
e0000d24  e0000e00  poll       (0,0x40219c,0,u+0xe28)
e0000e08  e0000e28  systrap    (u+0xe34)

The systems are running the latest version of the respective network card drivers and Support Level Supplement (SLS) OSS449F, the Network Maintenance Supplement for SCO OpenServer.

Cause

The dlpi_unlink function is part of the SCO DLPI (Data Link Provider Interface). This provides a link between the physical network card and its driver and the OpenServer 5 network protocol stack. For example:

=================
OpenServer 5.0.2
+++++++++++++++++
STREAM Head
=================
TCP
-----------------
IP
                        <-----+
-----------------             | DLPI and MDI modules
MAC                     <-----|
~~~~~~~~~~~~~~~~~
Network Card
~~~~~~|~~~~~~~~~~
      |
      ---> LAN

In this case, the DLPI module is unable to handle the amount of incoming route information from the network. Source routing information in a Token Ring environment is provided as a function of the network card driver.

Solution

The problem has been resolved in SCO OpenServer Release 5.0.4.

See also

scohelp(XC) for details on DLPI and Token Ring networking.

From the Xinuos technical article database, record 105866. Reproduced with permission. The original is at www.sco.com/ta/105866, which is reachable but blocked from search engines. Xinuos marks this article against a retired product line. That refers to the product’s lifecycle, not to whether the advice still works — plenty of it still applies to running systems.

Next step

Need this fixed rather than explained?

Tell us the release, what the system runs and what it is doing. We will reply by email with what we would do next.

  • You deal with the people who do the work
  • No obligation, and no upgrade required before we will help
  • We tell you plainly if something is not worth doing

Send an enquiry

We reply by email. If a server is down, say so in the first line. You can also write to help@scoexpert.com.