[22:08] Md: around [22:08] ? [22:11] mbiebl: yes [22:11] Md: hi [22:11] could you take a look at http://debs.michaelbiebl.de/bootchart.png [22:11] especially the modprobe calls during udevadm phase [22:12] Is it normal, that they take 3 secs? [22:14] mbiebl: I think so, IIRC there is some kind of serialization in the kernel. the most recent kernel and m-i-t should have some performance improvements [22:14] if you're using 2.6.30 there's no in-kernel lock [22:14] which m-i-t is it? [22:14] 3.7-pre9 [22:15] hmm, that should be fine [22:16] or is bootchart lying to me ;-) [22:17] well, those will be all your combined modprobes [22:18] but it's odd it's taking taht long [22:18] udevadm is ~0.5s for me [22:20] Keybuk: fwiw, this is a bootchart of a up-to-date karmic system: http://debs.michaelbiebl.de/pluto-karmic-20090609-1.png [22:22] the IO looks pretty crappy there [22:22] HDD or SSD? [22:22] HDD [22:22] and you have the same death-by-devkit that I see ;p [22:22] yeah, I was already puzzled by that [22:42] You call that a death by devkit? http://launchpadlibrarian.net/27624460/virhe-karmic-20090607-1.png :-P [22:43] wow, that looks bad :-o [22:44] TBH, it’s not devkit’s fault. https://bugs.edge.launchpad.net/ubuntu/+source/linux/+bug/384579 [22:44] lol [22:46] ion_: do you use MODULES=dep in initramfs.conf (I'm wondering because your initramfs stage is much shorter than mine) [22:47] MODULES=most