On Thu, Jan 19, 2012 at 3:52 PM, Thomas Bächler <thomas@archlinux.org> wrote:
Who cares? :)
You'd be surprised... ;-)
2012.1.3 would fix nilfs as well as make it possible to move udev-177 to core (due to removing UDEV_TIMEOUT support which is causing problems).
Read https://bugs.archlinux.org/task/20764. Back then, the patch was added to make people shut up about hangs at boot. In such cases, why not let people break their own systems?
There are a few ways people can still fix apparent hangs on boot: * grab a coffee and wait for 120 seconds * report the problem with the broken kernel driver upstream to get it fixed properly * submit a patch to udev upstream to support a "settle timeout" setting in /etc/udev/udev.conf (preferably in such a way that it works well with the other timeouts in udev, afaik there are different timeouts that kick in at 30, 60, 120 and 180). * edit /etc/rc.d/functions manually and add "--timeout=30" The issue is now that if we keep UDEV_TIMEOUT=30 as it is now it will break lots of wireless drivers. This could probably be fixed as easily as increasing the default to 40, but that will only work until some other assumption is changed in udev. This kind of thing belongs in udev, not in initscripts. -t