NetBSD-Bugs archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

install/48093: sysinst cores with floating point exception on arm/BEAGLEBONE



>Number:         48093
>Category:       install
>Synopsis:       sysinst cores with floating point exception on arm/BEAGLEBONE
>Confidential:   no
>Severity:       non-critical
>Priority:       low
>Responsible:    install-manager
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Sun Jul 28 22:35:00 +0000 2013
>Originator:     Dennis Ferguson
>Release:        6.99.23
>Organization:
>Environment:
No uname with an install kernel.  sysctl says this:

kern.ostype = NetBSD
kern.osrelease = 6.99.23
kern.osrevision = 699002300
kern.version = NetBSD 6.99.23 (BEAGLEBONE_INSTALL) #0: Sun Jul 28 14:21:04 EDT 
2013 
dennis%ts1.to.mistimed.ca@localhost:/usr/obj/sys/arch/evbarm/compile/BEAGLEBONE_INSTALL
>Description:
sysinst starts on a BeagleBone Black but cores just after it says it is writing 
the MBR, like this:

 Setting up the DOS partition table ... 
 pid 8 (sysinst), uid 0: exited on signal 8 (core dumped)
[1]   Floating point exception (core dumped) sysinst

I tried this with install kernels compiled for all three arm abi's, i.e. arm, 
earm and earmhf, just to make sure it had nothing to do with that. All did the 
same thing at the same spot.

I have very limited ability to debug this further right now.  I did manage to 
run gdb on the core with a binary without debugging symbols and got:

# gdb sysinst ../sysinst.core
GNU gdb (GDB) 7.3.1
Copyright (C) 2011 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
and "show warranty" for details.
This GDB was configured as "arm--netbsdelf".
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>...
Reading symbols from /root/evbarm/sysinst...(no debugging symbols found)...done.
[New process 1]
Core was generated by `sysinst'.
Program terminated with signal 8, Arithmetic exception.
#0  0x00183518 in ?? ()
(gdb) where
#0  0x00183518 in ?? ()
#1  0x00183414 in ?? ()
#2  0x00183414 in ?? ()
Backtrace stopped: previous frame identical to this frame (corrupt stack?)
(gdb) quit

I have no idea whether any of that is valid or not, but if the addresses are 
correct than the exception occurred in a library function rather than the 
sysinst text itself, since the latter only extends to 0x00036e4c.

I wish I had more to offer.  NetBSD does run on this board, I have a by-hand 
installation on the flash card which seems to run fine but it is hard to get 
that far without sysinst to fill in some of the missing files.
>How-To-Repeat:
Boot a BEAGLEBONE_INSTALL kernel on a BB Black.  Getting one may require this 
patch:


Index: distrib/evbarm/instkernel/instkernel/Makefile
===================================================================
RCS file: /cvsroot/src/distrib/evbarm/instkernel/instkernel/Makefile,v
retrieving revision 1.20
diff -u -r1.20 Makefile
--- distrib/evbarm/instkernel/instkernel/Makefile       6 May 2013 17:42:30 
-0000       1.20
+++ distrib/evbarm/instkernel/instkernel/Makefile       28 Jul 2013 22:16:24 
-0000
@@ -34,6 +34,7 @@
     ${MACHINE_ARCH} == "earmhf"
 # Little endian (any ABI) platforms.
 MDSETTARGETS+=         RPI_INSTALL             ${SSHRAMDISK}   -
+MDSETTARGETS+=         BEAGLEBONE_INSTALL      ${RAMDISK}      -
 .endif
 
 MDSET_RELEASEDIR=      installation/instkernel

Select English as the language and the board to be an `Other' (it isn't a 
Raspberry Pi), fill out the forms for the MBR and BSD disk labels and then 
watch it crash.

I know this isn't much to go on, but I wanted to note this somewhere.  I can 
investigate further if I know what to look at.
>Fix:



Home | Main Index | Thread Index | Old Index