Summary: | ipw2200 Firmware fails to load during 2005.0 boot up | ||
---|---|---|---|
Product: | Gentoo Release Media | Reporter: | Bryan <gtg627p> |
Component: | Everything | Assignee: | Gentoo Release Team <releng> |
Status: | RESOLVED DUPLICATE | ||
Severity: | normal | CC: | henrik, nefercheprure |
Priority: | High | ||
Version: | unspecified | ||
Hardware: | x86 | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- |
Description
Bryan
2005-04-25 15:22:17 UTC
I successfully helped the reporter on irc: 00:06:51 <@brix> frostedeg: try 'echo -n "60" > /sys/class/firmware/timeout' 00:07:10 <@brix> frostedeg: then 'modprobe -r ipw2200 && modprobe ipw2200' 00:08:13 < frostedeg> bingo 00:08:16 < frostedeg> got eth0 now This issue is tracked upstream as http://bughost.org/bugzilla/show_bug.cgi?id=601 This does not works for me, even trying echo 100 > /sys/class/firmware/timeout modprobe -r ipw2200 && modprobe ipw2200 I was running ipw2200 v1.0.4 and firmware v2.3 successfully (latest at this time) and sudenly the firmware stopped loading as described here. The only thing I remember doing was a change somewhere in /etc/modules.* to change the alias from 'eth1' to 'wlan0', but it didnt work as expected and I reverted to the previous configuration. After that the driver stopped working correctly. Actually the last time I actually successfully used the wireless interface was a week ago, under 2.6.10 but in the meantime upgraded to 2.6.11-suspend2 (thans Brix for the convenient ebuild :-) I didn't touch hotplug or anything else that seems to be related and confirmed that the directory firmware is correct. I tried downgrade the firmware to v2.2 with no luck, and then tried ipw2200 v.1.0.3 and it worked (even with firmware 2.3). So I am leaving it working like this as I don't have time right now to dig further... A similar problem occured after "emerge -u system" on an installed gentoo. After some investigations I have found the echo 100 > /sys/class/firmware/timeout solution elsewhere. It works only partially, however. (It shows a nondeterministic behavior: sometimes it works, sometimes it does not.) After further investigation I figured out that /proc/sys/kernel/hotplug contains /sbin/udevsend. Setting it to /sbin/hotplug eliminated the problem altogether. The problem started with the upgrade of udev from 045 to 058, because /sbin/rc sets the /proc/sys/kernel/hotplug to /sbin/udevsend based on the version of udev. Therefore, this could possibly be construed a bug in udev and/or /sbin/rc. As a temporary solution, the line: echo "/sbin/udevsend" > /proc/sys/kernel/hotplug in /sbin/rc can be commented out. |