On Tue, May 10, 2011 at 17:34, Koen Kooi <k...@dominion.thruhere.net> wrote:
> So after rebuilding eglibc against 2.6.37 headers and udev from git: > > root@beagleboard-systemd:~# time ( udevadm trigger ; udevadm settle ) > > real 0m10.475s Ok, at least the timeout issue seems gone. How many device do you have? find /sys -name uevent | wc -l How many devices have a modalias: find /sys -name modalias | wc -l > And the matching systemd-analyze plot: > http://dominion.thruhere.net/koen/angstrom/systemd/git-with-accept4.svg What does: time udevadm test /sys/class/tty/ttyO0 show? If you run it only for certain subsystems, what does it give you: time ( udevadm trigger -s block; udevadm settle ) time ( udevadm trigger -s tty; udevadm settle ) > I'll do some more digging after linuxtag. I'll be around Fri+Sat. Maybe we'll walk into each other ... Kay _______________________________________________ systemd-devel mailing list systemd-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/systemd-devel
