HELP: BAD TRAP, kernel panic on Solaris 2.4

HELP: BAD TRAP, kernel panic on Solaris 2.4

Post by Jozsef Kadlecs » Wed, 04 Dec 1996 04:00:00


On November 20th I installed all the recommended pacthes for Solaris 2.4
on our machine and now the machine panics:

Dec  2 14:49:06 unix: BAD TRAP: type=9 rp=f0ca8d74 addr=20202074 mmu_fsr=126 rw=1
Dec  2 14:49:06 unix: aleph: Data fault
Dec  2 14:49:06 unix: kernel read fault at addr=0x20202074, pme=0x0
Dec  2 14:49:06 unix: MMU sfsr=126: Invalid Address on supv data fetch at level 1

Dec  3 09:30:58 unix: BAD TRAP: type=9 rp=f0c78d74 addr=94 mmu_fsr=326 rw=1
Dec  3 09:30:58 unix: rts32: Data fault
Dec  3 09:30:58 unix: kernel read fault at addr=0x94, pme=0x0
Dec  3 09:30:58 unix: MMU sfsr=326: Invalid Address on supv data fetch at level 3

Is there any problem with the listed patches? Should I backout the kernel
patch? Or we have a hardware problem?

102479-02   SunOS 5.4: DNS spoofing is possible per Cern ca-96.02
101945-43   SunOS 5.4: patch for kernel
101959-11   SunOS 5.4: lp jumbo patch
101973-24   SunOS 5.4: fixes for libnsl and ypbind
102042-05   SunOS 5.4: usr/bin/mail jumbo patch
102044-01   SunOS 5.4: bug in mouse code makes "break root" attack possible
102049-05   SunOS 5.4: linker fixes
102066-12   SunOS 5.4: sendmail bug fixes
102070-01   SunOS 5.4: Bugfix for rpcbind/portmapper
102165-03   SunOS 5.4: fixes
102216-07   SunOS 5.4: klmmod and rpcmod patch
102218-03   SunOS 5.4: libbsm fixes
102277-02   SunOS 5.4: fixes
102437-03   SunOS 5.4: /usr/ccs/bin/as has an internal error
102656-01   SunOS 5.4: /dev/qec should protect against being opened directly
102664-01   SunOS 5.4: data fault in scanc() due to bad "cp" argument
102680-04   SunOS 5.4: fixes for ufsdump and wall
102693-03   SunOS 5.4: at and atrm fixes
102704-02   SunOS 5.4: jumbo patch for NIS commands
102711-01   SunOS 5.4: Creation of /tmp/ps_data is security problem
102741-01   SunOS 5.4: libm can hit SEGV in multi-threaded mode
102756-01   SunOS 5.4: expreserve still has security problem
102769-03   SunOS 5.4: statd fixes
102788-03   SunOS 5.4: Patch for sccs
102922-03   SunOS 5.4: inetd fixes
102960-01   SunOS 5.4: vipw has security problem
103070-01   SunOS 5.4: tip will read and print any uucp owned file
103270-01   SunOS 5.4: nissetup default permissions not secure enough
103706-01   SunOS 5.4: rpc.nisd_resolv rebuild for BIND 4.9.3
103813-01   SunOS 5.4: rdist suffers from buffer overflow
101878-13   OpenWindows 3.4: Xview Patch
102292-02   OpenWindows 3.4: filemgr (ff.core) fixes

Thank you any help,
Jozsef Kadlecsik

Address : KFKI Research Institute for Particle and Nuclear Physics
          P.O.B 49 Budapest, 1525 Hungary


1. 2.3->2.4 install/upgrade panic (BAD TRAP)

We are attempting to install/upgrade (I could upgrade but I'd rather do an
initial install) a SS 1000 with 128Mb of RAM, 2 system boards and 2 CPUs
to Solaris 2.4.  The system is currently already running 2.3.

I am able to boot the 2.4 (11/94 hardware) CDROM, start openwindows and
answer questions.  I answer questions about the network, name services,
hostname, IP address and timezone.  I get to the window just before where
it normally asks you if you want to do an Upgrade or an Initial install
(this is the window that says you are about to "tailor" the install and
has a Continue button).  I click on continue and after about 5 seconds
the system panics with a "BAD TRAP" saying <Memory Address Alignment>.
It then shows something like a stack trace of installtool.  I believe that
last call was something like ufs_dirlook().

I've called Sun software support on this and we're still clueless.  Someone
on the net (sorry, forgot your name!) suggested that I pull all but the lowest
32Mb of memory out and try it again -- still failed.

I have also tried:

- pulling one system board
- swapping system boards
- testing all memory from the openboot prompt
- pulling all SCSI devices (external devices -- the disks internal to the
  SS 1000 were still there)
- moving the 3 3rd party kernel modules we have to another location and
  rebuilding the kernel prior to another upgrade attempt
- commenting out all but /,/usr,/var, /opt and swap in the /etc/vfstab file
- harsh language :-)

all (with the exception of the last one) have given me the BAD TRAP panic
in exactly the same way at exactly the same time.  It's 100% reproducible on
this system.

This system has no history of hardware problems.  We are not running FDDI.
I see no errors (other than 2.3 software ones, grrrr) while running Solaris

My Sun support guy says he's going to turn me over to the kernel people
on Monday.  Sun doesn't think it's hardware.

Has anyone seen this before?  We believe that the move to 2.4 will fix a lot
of problems we're having with 2.3 (NIS+ especially) so this has now become
a critical issue for us.


-mark frost

2. Changing Depth in X

3. Solaris 2.4 on Sparc 1 - BAD TRAP on boot

4. Killing ALL offspring

5. BAD TRAP installing 2.4_x86/DU10 (HELP!)

6. Where to put startup commands

7. Q: bad trap panic Solaris 2.6

8. help with iptables script

9. Bad TRAP after installation of 2.4

10. can anyone help: "BAD TRAP", trap type = 0x31

11. PANIC: bad trap - alignment error

12. solaris2.6 (x86) install panics: BAD TRAP

13. httpd BAD TRAP, panic: Data fault