Hpux 11.23 server panic: post_hndlr(): Unresolved kernel interruption

[likid0]

Hi,

I have 2 servers ,both of them have given panics in different days

The panic in both of them are exactly the same, they have the same stack trace:

panic: post_hndlr(): Unresolved kernel interruption

Stack Trace:
  IP                  Function Name
  0xe00000000052f310  post_hndlr+0xc80
  0xe00000000052e220  vm_hndlr+0x210
  0xe000000000bc0a40  bubbledown
  0xe0000000005fa810  qenable+0x10
  0xe0000000014261a0  td_dumpdev_unconf+0x5970
  0xe000000000810110  sq_wrapper+0xf0
  0xe000000000b712b0  str_sched_mp_daemon+0x810
  0xe000000000b6f5a0  str_sched_daemon+0x3a0
  0xe000000001017cc0  im_mpnetstr+0x60
  0xe00000000144e410  DoCalllist+0x3a0
End of Stack Trace

Really looks like a patch problem, since they have a very old QPK, but I have been looking around in the kb I can't pinpoint exactly watch patch is needed to solve the problem,we first need to find the bug so we can justify a patch update

Both have the same same level of OE and patches installed:

 HPUX11i-OE-Ent	B.11.23.0609   HP-UX Enterprise Operating Environment Component 

  QPKAPPS		B.11.23.0609.053a Applications Quality Pack Bundle for HP-UX 11i v2, September 2006 
  QPKBASE		B.11.23.0609.053a Base Quality Pack Bundle for HP-UX 11i v2, September 2006 

[Greg Tinker]

are you loading a dynamic kernel module at time of panic?

I only ask, because your footprint matches PHKL_31500

[Laurent Menase]

Hi

you probably need to ask to "td_dumpdev_unconf"  vendor.

Looks like you probably have some 3rd party vendor streams driver install.

Best regards

LM

[likid0]

Hi,

Thanks for the answer. I can't find any third party software installed that loads modules, I don't see any module in kcmodule that stands out.

The only third party software installed is a backup client that doesn't load any drivers.

I have attached the kcmodule ouput, to se if you can help me trackdown which is the 3rd party soft.

[Dennis Handly]

>you probably need to ask to "td_dumpdev_unconf"  vendor.

I find this entry on both 11.23 and 11.31:

$ nm -pxAN /usr/conf/*/* | grep td_dumpdev_unconf

T  /usr/conf/lib/libtd-pdk.a[td_dumpdev_real.o]:td_dumpdev_unconf  (11.23)
t  /usr/conf/mod/td:td_dumpdev_unconf   (11.31)

>because your footprint matches PHKL_31500

I'm not sure that's related since ALL customers on 11.23 should have that omnibus patch in 2004.

While I see post_hndlr, I don't see that stack trace.

[likid0]

Yes, all 11.23 since 2004 have that patch, it's when pa-risc was added to 11.23.

I can't relate it with a third party stream module

[likid0]

Sorry, I posted, without adding info, I have the same output:

0x0000000000000010 T  /usr/conf/lib/libtd-pdk.a[td_dumpdev_real.o]:td_dumpdev_unconf

> swlist -l file | grep -i /usr/conf/lib/libtd-pdk.a
  FC-TACHYON-TL.FC-TL-KRN: /usr/conf/lib/libtd-pdk.a

[Laurent Menase]

indeed you are right, what make me feel there was a 3rd party is that I never saw

td driver using streams framework.

  0xe0000000014261a0  td_dumpdev_unconf+0x5970
  0xe000000000810110  sq_wrapper+0xf0

so probably there is a static function after td_dumpdev_unconf which caused the panic.

So if you have PHKL_31500 on your system this need to be analyzed by hp support.

disassembling around that adress will be necessary.

[Dennis Handly]

>so probably there is a static function after td_dumpdev_unconf which caused the panic.

Ah, yes, the large offset from td_dumpdev_unconf.  But I would have thought that the kernel stack trace function would handle static functions?

But likid0 never mentioned whether it was PA or Integrity.

Possibly gdb would work for either?

gdb /stand/vmunix

disas 0xe0000000014261a0 0xe0000000014261a0+16

q

[likid0]

Hi Dennis, sorry I forgot to mention, this is a Itanium server.

(gdb)  disas 0xe0000000014261a0 0xe0000000014261a0+16
Dump of assembler code from 0xe0000000014261a0:0 to 0xe0000000014261b0:0:
;;; DOC Line Information: [Line, Column Start, Column End] [Line, Column] [Line]
;;; File: /ux/core/kern/common/net/telnet/str_telnet.c
;;; Line: 2613
0xe0000000014261a0:0 <telnet_lwsrv+0x210>:      
          ld8              r50=[r48]
0xe0000000014261a0:1 <telnet_lwsrv+0x211>:      
          mov              r51=r36;;
0xe0000000014261a0:2 <telnet_lwsrv+0x212>:      
          cmp.eq.unc       p2,p6=r0,r50

I had also seen it in the crashinfo trap section,but couldn't find any info on it. I have also attached it