Red Hat Bugzilla – Bug 180021
0.71 update breaks orinoco/firmware based NICs
Last modified: 2007-11-30 17:11:23 EST
From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.8.0.1) Gecko/20060124 Firefox/1.5.0.1 Description of problem: wlan0 no longer works and it seems to be replaced with eth1, or possibly wlan0 and eth1 seem to have their wires crossed somewhere. The driver identifies the card as a Prism I card instead of a Prism 2.5 card. This all started with the kernel-2.6.15-1.1830_FC4 & udev-0.71 update. If I go back to booting the previous FC4 kernel, the problem remains, which leads me to believe its a udev problem. This is basically the same issue in bug 179982, except i'm reporting it against udev instead of the kernel. I've got a Linksys WPC11v3 PCMCIA wifi card that uses the orinoco driver. Since the udev update to 0.71 as soon as I try to bring up the interface for the wifi card, I start getting an un-ending stream of errors like this: cannot change name of wlan0 to devWXYZ: File exists where WXYZ seems to be a random number that changes each time the error appears (and it appears many times per second, but gradually slows down over time). This is really really bad. Version-Release number of selected component (if applicable): udev-0.71 How reproducible: Always Steps to Reproduce: 1. setup networking for an orinoco driver based wifi NIC 2. attempt to bring up the eth0 interface for the NIC Actual Results: tons of "cannot change name of wlan0 to devWXYZ: File exists" errors and the interface never comes up. In fact, if the NIC is set to come up on boot, then the boot process never finishes as the errors just continue indefinitely (I waited 20 minutes, and it was still erroring out) Expected Results: eth0 comes up, no errors Additional info:
please post your /etc/sysconfig/network-scripts/ifcfg-* files here.
Created attachment 124735 [details] ifcfg-* files requested files attached in tarball.
More generally: udev-0.71 doesn't seem to be able to rename network interfaces. Here's my /etc/sysconfig/network-scripts/ifcfg-ipw0 file: DEVICE=ipw0 TYPE=Wireless MODE=Managed BOOTPROTO=dhcp ONBOOT=no Here's my /etc/udev/rules.d/10-ipw2100.rules file: # Name my wireless interface ipw0 instead of the next available "eth" number. KERNEL=="eth*", BUS=="pci", ID=="0000:02:03.0", DRIVER=="ipw2100", NAME="ipw0", OPTIONS="last_rule" If I modprobe the ipw2100 module, and then use udevtest, all seems well: $ udevtest /sys/class/net/eth1 net main: looking at device '/class/net/eth1' from subsystem 'net' main: opened class_dev->name='eth1' udev_rules_get_name: rule applied, 'eth1' becomes 'ipw0' rename_net_if: changing net interface name from 'eth1' to 'ipw0' udev_add_device: renamed netif to 'ipw0' I've inspected the log messages from udev, and nothing seems amiss. (I'll attach them in my next comment.) However, the rename never happens: $ ip link show 1: lo: <LOOPBACK,UP> mtu 16436 qdisc noqueue link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 2: eth0: <BROADCAST,MULTICAST> mtu 1500 qdisc noop qlen 1000 link/ether 00:0b:db:a1:19:98 brd ff:ff:ff:ff:ff:ff 13: eth1: <BROADCAST,MULTICAST> mtu 1500 qdisc noop qlen 1000 link/ether 00:04:23:48:8e:e6 brd ff:ff:ff:ff:ff:ff I can rename the interface manually, no problem: $ ifrename -c /dev/null -i eth1 -n ipw0 ipw0 $ ip link show 1: lo: <LOOPBACK,UP> mtu 16436 qdisc noqueue link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 2: eth0: <BROADCAST,MULTICAST> mtu 1500 qdisc noop qlen 1000 link/ether 00:0b:db:a1:19:98 brd ff:ff:ff:ff:ff:ff 13: ipw0: <BROADCAST,MULTICAST> mtu 1500 qdisc noop qlen 1000 link/ether 00:04:23:48:8e:e6 brd ff:ff:ff:ff:ff:ff The fact that this worked perfectly using udev-058, combined with the fact that udevtest thinks the rename should happen, leads me to believe that this is actually a regression/bug in udev-0.71.
Created attachment 124781 [details] udev log messages These are the log messages that are generated when I execute "modprobe ipw2100".
Is there a reason why this bug is still in the NEEDINFO_REPORTER state? I attached the requested information a week ago.
Harald, any ideas when this might see some attention? I'm deathly afraid of upgrading to FC5 because I have no clue whether this bug is still there, and I *really* need wifi to work.
ok, next step... comment out all modprobe lines from: /etc/udev/rules.d/* try to boot
oops... forget the last comment... wrong bugzilla :)
FWIW, I don't experience this problem with FC5 ( udev-084-13). More specifically, this line in a file in /etc/udev/rules.d: # Name my wireless interface ipw0 instead of the next available "eth" number. KERNEL=="eth*", BUS=="pci", ID=="0000:02:03.0", DRIVER=="ipw2100", NAME="ipw0" ...does what one would expect. Lonni, I suspect that you might have to use a similar rule to ensure that your wireless interface is named exactly what you want it to be, but if this works for me, it should work for you, too.
This report targets the FC3 or FC4 products, which have now been EOL'd. Could you please check that it still applies to a current Fedora release, and either update the target product or close it ? Thanks.
I'm still seeing this problem in FC5.
Just FYI, I've had no issues with udev being able to rename the interface of my ipw2100 adapter from eth1 to ipw0 since FC5, so I don't think this is a general udev problem per se.