Bug 5915 - if ip_always_defrag is missing then kudzu fails
Summary: if ip_always_defrag is missing then kudzu fails
Alias: None
Product: Red Hat Linux
Classification: Retired
Component: kudzu
Version: 6.1
Hardware: All
OS: Linux
Target Milestone: ---
Assignee: Bill Nottingham
QA Contact:
Depends On:
TreeView+ depends on / blocked
Reported: 1999-10-13 19:22 UTC by shishz
Modified: 2014-03-17 02:10 UTC (History)
1 user (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Clone Of:
Last Closed: 2000-02-10 16:13:10 UTC

Attachments (Terms of Use)

Description shishz 1999-10-13 19:22:00 UTC

After an upgrade from 6.0 -> 6.1, I compiled my own
2.2.12 kernel without ip_always_defrag.

I received a /proc/sys/net/ipv4/ip_always_defrag is missing
cannot control ip defragmentation and kudzu was marked
off as [failed]...

I made the following change to /etc/rc.d/init.d/network:
not sure if this is the correct way to do this but it
worked out ok... (initscripts-4.48-1)

There may be something more going on, since I didn't
expect /etc/rc.d/init.d/network to cause
/etc/rc.d/init.d/kudzu to fail.

--- network	Wed Oct 13 14:35:29 1999
+++ network.new	Wed Oct 13 14:35:29 1999
@@ -77,8 +77,10 @@
 	# Good for masquerading. Not so good otherwise.
 	if [ -d /proc/sys/net/ipv4 ]; then
 	    if [ ! -f /proc/sys/net/ipv4/ip_always_defrag ]; then
-	       echo "/proc/sys/net/ipv4/ip_always_defrag is
missing --" \
-	            "cannot control IP defragmentation" >&2
+	       if [ "$DEFRAG_IPV4" = "yes" -o "$DEFRAG_IPV4" =
"true" ]; then
+	          echo "/proc/sys/net/ipv4/ip_always_defrag is
missing --" \
+	               "cannot control IP defragmentation" >&2
+	       fi
 	       if [ "$DEFRAG_IPV4" = "yes" -o "$DEFRAG_IPV4" =
"true" ]; then


Comment 1 Bill Nottingham 1999-10-13 19:37:59 UTC
The message is OK as it is, I would think; it's just stating
a fact.

Most likely kudzu is failing because of something else.

What happens if you run it at the command line?

Do you have any ide-scsi devices?

Comment 2 shishz 1999-10-14 16:28:59 UTC
Ah, your right... I backstepped what I had done and found that kudzu
seg faults if CONFIG_FIREWALL=y on my machine.

I don't use the firewall or have any masquerading on this machine
so I don't know if this is a bug in kudzu or not.

Two 3c509b nics, pentium, no ide-scsi devices, no scsi.

Everything is fine when I recompiled with config_firewall off.

Here is a bt if your interested:

This GDB was configured as "i386-redhat-linux"...
(no debugging symbols found)...
Core was generated by `kudzu'.
Program terminated with signal 11, Segmentation fault.
Reading symbols from /usr/lib/libnewt.so.0.50...(no debugging symbols
Reading symbols from /usr/lib/libslang.so.1...(no debugging symbols
Reading symbols from /lib/libc.so.6...done.
Reading symbols from /lib/libm.so.6...done.
Reading symbols from /lib/ld-linux.so.2...done.
#0  0x8050e18 in strcpy () at ../sysdeps/generic/strcpy.c:30
30	../sysdeps/generic/strcpy.c: No such file or directory.
(gdb) bt
#0  0x8050e18 in strcpy () at ../sysdeps/generic/strcpy.c:30
#1  0x8088810 in ?? ()
#2  0x805124c in strcpy () at ../sysdeps/generic/strcpy.c:30
#3  0x804d179 in strcpy () at ../sysdeps/generic/strcpy.c:30
#4  0x804c2de in strcpy () at ../sysdeps/generic/strcpy.c:30
#5  0x400841eb in __libc_start_main (main=0x804c020 <strcpy+8928>,
    argv=0xbffffd14, init=0x8049650 <_init>, fini=0x80568cc <_fini>,
    rtld_fini=0x4000a610 <_dl_fini>, stack_end=0xbffffd0c)
    at ../sysdeps/generic/libc-start.c:90

Comment 3 Bill Nottingham 1999-10-18 16:28:59 UTC
CONFIG_FIREWALL?  There's nothing in kudzu that should
depend on that one way or another.

Unfortunately, the binary you have doesn't have debugging symbols,
so I can't figure out where it died at.

Comment 4 Bill Nottingham 2000-02-08 22:48:59 UTC
(Apologies for the extreme delay)

Did you compile your kernel without CONFIG_PNP_PARPORT? This
*will* cause a crash with the kudzu that shipped in 6.1,
and it's since been fixed.

Comment 5 shishz 2000-02-10 16:13:59 UTC
Thanks, that was it...

Note You need to log in before you can comment on or make changes to this bug.