Note: This bug is displayed in read-only format because the product is no longer active in Red Hat Bugzilla.

Bug 114071

Summary: kudzu/system-config-network crashes when new ADSL modem and eth device is present
Product: [Retired] Red Hat Raw Hide Reporter: Moritz Barsnick <moritz>
Component: kudzuAssignee: Bill Nottingham <notting>
Status: CLOSED DUPLICATE QA Contact: David Lawrence <dkl>
Severity: medium Docs Contact:
Priority: high    
Version: 1.0CC: moritz, rvokal, ssj4goko4
Target Milestone: ---   
Target Release: ---   
Hardware: i686   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2006-02-21 19:00:53 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Attachments:
Description Flags
typical output of "ifconfig -a"
none
typical output of "lsmod"
none
typical output of "lsusb"
none
typical content of /proc/bus/usb/devices
none
patch against kudzu-1.1.40 to fix segfault none

Description Moritz Barsnick 2004-01-22 09:18:37 UTC
Description of problem:
I have a new USB ADSL modem, with a non-kernel driver. The driver 
works, and so does the modem. When modem and driver are active (eth1 
is present), system-config-network crashes (segfaults) immediately.

Version-Release number of selected component (if applicable):
1.3.13-2

How reproducible:
always

Steps to Reproduce:
1. Install driver (module + hotplug userdata).
2. Activate modem (hotplug or boot)
3. Call system-config-network (from cmdline, menu, or from
   system-control-network)
  
Actual results:
Segmentation fault (gnome_segv pops up)

Expected results:
Program runs.

Additional info:
This is hindering me from properly configuring my ADSL modem, because 
I'm too stupid to use "route"/"ip". ;-)

It is an Allied Telesyn AT-AR215 USB modem (an "ADImodem", i.e. an 
Analog Devices modem). I use the German driver from http://eagle-usb.
eisfront.de/, which I built as an RPM (myself). The driver works 
fine: hotplug recognizes the device (lsmod shows eagle-usb), 
initializes it with firmware, the device syncs with the DSL network, 
and the driver presents the system with an eth1 (eth0 is present as 
well, my Eth card). If I use rp-pppoe and pppd manually (and some 
guessing with "route add ..."), it operates fine. Yet starting 
system-config-network in any possible way gives a segmentation fault 
very quickly. I think system-config-network is still parsing 
files/output at this point. GPlease give me instructions for an 
"strace [-f -ff]" call, and I will deliver any info required.

If I remove the driver, or reboot without the modem connected, 
system-config-network can start fine. Yet I can't configure the 
connection without the device. ;-) And I need this nice little config 
app, because I'm too stupid for all the mechanisms behind it, right 
now.

BTW, kudzu also crashes when this device is present. ;-)

I will attach some files to show what the system looks like when the 
driver is loaded. I'm open for any other requests.

Thanks a bunch,
Moritz

Comment 1 Moritz Barsnick 2004-01-22 09:20:54 UTC
Created attachment 97156 [details]
typical output of "ifconfig -a"

This is typical output of "ifconfig -a" when the eagle-usb driver is loaded.
The new interface eth1 is not activated yet though.

Comment 2 Moritz Barsnick 2004-01-22 09:22:24 UTC
Created attachment 97157 [details]
typical output of "lsmod"

This is typical output of "lsmod" when the eagle-usb driver is loaded.

Comment 3 Moritz Barsnick 2004-01-22 09:23:16 UTC
Created attachment 97158 [details]
typical output of "lsusb"

This is typical output of "lsusb" when the eagle-usb driver is loaded.

Comment 4 Moritz Barsnick 2004-01-22 09:24:30 UTC
Created attachment 97159 [details]
typical content of /proc/bus/usb/devices

This is typical content of /proc/bus/usb/devices when the eagle-usb driver is
loaded.

Comment 5 Harald Hoyer 2004-01-22 09:47:54 UTC
system-config-network uses kudzu, which segfaults.
reassigning

Comment 6 Moritz Barsnick 2004-01-22 13:38:53 UTC
Thanks to this pointer to kudzu (I thought it was a separate problem) 
and the Fedora-list archive, specifically
http://www.redhat.com/archives/fedora-list/2004-January/msg04701.html
, I think this bug may be a duplicate of bugzilla Bug #106332. 
Especially the contents of attachment 94965 [details] and attachment 94966 [details] very 
much resemble what I saw when stracing and gdb'ing either 
system-config-network or kudzu (can't remember which ;->); i.e. the 
functions which the backtrace lists, and the 
/etc/sysconfig/network-scripts/ifcfg-lo at the and of the strace!

Nevertheless, I hope my info (especially tracing it down to a certain 
device/interface) helps track this down.

Has attachment 96307 [details] been integrated into kudzu yet?

Thanks

Comment 7 Moritz Barsnick 2004-01-23 08:20:57 UTC
Indeed, the patch in attachment 96307 [details] does the job for me. I rebuilt 
the kudzu RPM with an updated patch applied, and kudzu no longer 
segfaults when the new device is present. system-config-network also 
starts up fine now. I still can't configure the connection with scn, 
but that's a different, separate bug report.

I'm attaching an updated patch which applies properly against 
kudzu-1.1.40 and post a reference in the report for Bug #106332.

Comment 8 Moritz Barsnick 2004-01-23 08:25:08 UTC
Created attachment 97205 [details]
patch against kudzu-1.1.40 to fix segfault

This patch is to fix the described segfault of kudzu. It is an update of the
one included in attachment 96307 [details]. Thanks to Norm Brake for the latter!

Comment 9 Bill Nottingham 2004-02-04 06:11:59 UTC
*** Bug 111979 has been marked as a duplicate of this bug. ***

Comment 10 Bill Nottingham 2004-02-04 06:13:15 UTC
*** This bug has been marked as a duplicate of 106332 ***

*** This bug has been marked as a duplicate of 106332 ***

Comment 11 Red Hat Bugzilla 2006-02-21 19:00:53 UTC
Changed to 'CLOSED' state since 'RESOLVED' has been deprecated.