Red Hat Bugzilla – Bug 48141
laptop eth0 device partially broken
Last modified: 2014-03-16 22:21:42 EDT
Description of Problem: Fairfax (beta 1); Dell Latitude CPx with a Xircom RealPort 10/100 + 56K modem (non cardbus) after the system initializes the eth0 device does not come up service network restart does not start the device, only lo ifup eth0 works, however How Reproducible: always Steps to Reproduce: 1. install on Latitude CPx with Xircom RealPort 10/100 + 56K modem 2. ifconfig shows eth0 not up; service network restart doesn't start it 3. ifup eth0 works Actual Results: see above Expected Results: eth0 should be started automatically Additional Information: system files: /etc/sysconfig/network : NETWORKING=yes HOSTNAME=localhost.localdomain /etc/sysconfig/network-scripts/ifcfg-eth0 : DEVICE=eth0 BOOTPROTO=dhcp ONBOOT=no note: I realize ONBOOT=no is set. However, the files are identical for a 7.1 install on the same laptop and the eth0 device does come up when the machine comes up and when the network is restarted. At first I thought 7.1 was buggy, but I think I remember reading somewhere the behavior in 7.1 was intentional; that hotplug should take care of brining the device up (or something like that).
This is fixed with the new hotplug (2001-04-24-5) packages.
reopening... still an issue in beta 2 # rpm -q hotplug hotplug-2001_04_24-6
aaaaaaaargh. The -6 release broke more than it fixed. Try 2001_04_24-7. Does that work for you? Also, which dhcp client are you using?
tested with hotplug-2001_04_24-7 ... still not working I must be using dhcpcd: $ rpm -q pump package pump is not installed $ rpm -q dhcpcd dhcpcd-1.3.18pl8-12 I've noticed a strange insmod error during the detecting hardware phase of the sysinit (kudzu?) -- could this be related?
No, that shouldn't be a problem. Just to confirm, if you set it up to use static IPs, does it work OK? Also, what initscripts do you have installed?
initscripts-6.02-1 I'll check on the static IP
using a static IP works. so maybe the problem is a hotplug interaction with dhcpcd?
hotplug doesn't interact with dhcpcd - it only invokes ifup when a new interface appears.
does it work in beta3 if the calls to check_link_down are commented out of /sbin/ifup?
this is no longer a problem in RC 1
argh, reopening... While it worked once (and so I assumed it would work again), it is having problems again. Here's what is happening now. The interface appears to come up, but it doesn't get an IP address from the DHCP server (note: this is on the internal Meridian network), e.g., # ifconfig eth0 Link encap:Ethernet HWaddr 00:10:A4:C7:47:4D UP BROADCAST NOTRAILERS RUNNING MTU:1500 Metric:1 RX packets:0 errors:0 dropped:0 overruns:0 frame:0 TX packets:0 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 RX bytes:0 (0.0 b) TX bytes:2360 (2.3 Kb) lo Link encap:Local Loopback inet addr:127.0.0.1 Mask:255.0.0.0 UP LOOPBACK RUNNING MTU:16436 Metric:1 RX packets:20 errors:0 dropped:0 overruns:0 frame:0 TX packets:20 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 RX bytes:1520 (1.4 Kb) TX bytes:1520 (1.4 Kb) after a couple of minutes it appears to give up and eth0 goes away: # ifconfig lo Link encap:Local Loopback inet addr:127.0.0.1 Mask:255.0.0.0 UP LOOPBACK RUNNING MTU:16436 Metric:1 RX packets:32 errors:0 dropped:0 overruns:0 frame:0 TX packets:32 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 RX bytes:2480 (2.4 Kb) TX bytes:2480 (2.4 Kb) but when I try ifup eth0, it comes up: [root@localhost root]# ifup eth0 Determining IP information for eth0... done. it is possible that this is an internal Meridian problem, I suppose I'm using the latest rawhide components on RC1, e.g., initscripts-6.26-1 kernel-pcmcia-cs-3.1.27-9 kernel-2.4.7-2.5
What does 'ip addr ls' show if you run it while it is bringing it 'up'? Does using pump instead of dhcpcd work? :)
in some repeated testing, sometimes dhcpcd works, sometimes it doesn't ip addr ls shows: 1: lo: <LOOPBACK,UP> mtu 16436 qdisc noqueue link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 inet 127.0.0.1/8 brd 127.255.255.255 scope host lo 2: eth0: <BROADCAST,NOTRAILERS,UP> mtu 1500 qdisc pfifo_fast qlen 100 link/ether 00:10:a4:c7:47:4d brd ff:ff:ff:ff:ff:ff I tried pump (removing dhcpcd); testing so far shows that pump works without a problem
I'm not seeing this issue with the 7.2 release . . . Chris, this working with 7.2?
OK, just a sanity check here. If ONBOOT is 'n' . . . the interface shouldn't start should it? Isn't that the point of the flag anyway?
The scripts aren't executed on boot... so far, so good. The problem is that they are started later in the bootup process, when pcmcia is started and detects them - hotplug like.
OK, here's more information. With the Skipjack tree (re0320.0) hotplug is definitely broken. I upgraded my 7.2 system which had 'onboot=yes' and now eth0 doesn't start on boot. Doesn't start on network restart either. The only way to get eth0 up is to run 'ifup eth0' Bill is aware of the problem, but just wanted to document that there is indeed a problem.
I have this problem on my laptop (IBM TP X20) with the Lucent Wavelan card. I traced the problem to the is_available function. $alias is getting set to "off" when hotplug calls it, but it works correctly when run by hand. Commenting out the call to is_available fixes it on my system.
*** This bug has been marked as a duplicate of 61705 ***