**** BEGIN LOGGING AT Sat Sep 29 02:59:57 2012 **** BEGIN LOGGING AT Sat Sep 29 06:04:57 2012 Sep 29 08:24:10 morphis, hi Sep 29 09:01:46 hi ao2 Sep 29 09:02:04 hi GNUtoo-desktop Sep 29 09:02:04 let's discuss on #flashrom Sep 29 09:02:22 OK Sep 29 09:49:01 morphis, ping me when you're awake Sep 29 13:19:30 JaMa, hi Sep 29 14:37:58 WOW.!!!!!!!!!!!! I've systemd booting in seconds on GTA02 on linux-next!!!!!!!!!!!!! I tried to bisect the boot slowness problem for hours and hours many times(like at least 5 times)....and it seeem a simple kenrel configuration issue..... Sep 29 14:39:46 correct config: http://paste.debian.net/194304/ Sep 29 14:40:07 basically between last time and this time I added stuff for making bootchart happy Sep 29 14:40:15 and boom it was fast.... Sep 29 14:40:38 however I don't seem to be able to login trough serial Sep 29 14:40:41 maybe bad cable Sep 29 15:12:18 great! Sep 29 15:15:02 indeed Sep 29 15:27:06 GNUtoo-desktop: does suspend work reliably? Sep 29 15:27:45 no idea yet Sep 29 15:27:47 I'll try Sep 29 15:27:58 but on 3.5 or 3.6 it should be easier to fix Sep 29 15:28:06 like tell more stuff in dmesg Sep 29 15:28:10 since that's the reason why I'm still stuck at 2.6.34 Sep 29 15:28:31 for now booting means cherry-picking 1 commit Sep 29 15:28:41 the IRQ level commit Sep 29 15:28:53 I'll port all the other commits Sep 29 15:28:55 as usual Sep 29 16:04:34 hmmm still no getty.... Sep 29 16:04:43 maybe I need to set a password Sep 29 16:04:46 or something like that Sep 29 16:05:05 shr 20120913 om-gta02 ttySAC2 <- I've that but nothing after.... Sep 29 16:08:22 hmmm Sep 29 16:08:35 now it's slow again.... Sep 29 16:08:40 I changed QI version Sep 29 16:10:49 let me try to narrow this.... Sep 29 16:11:14 because the QI version that works is with special args for bootchart Sep 29 16:38:06 JaMa, hi Sep 29 16:38:21 it seems that running it in a wrapper is the cause of the speedup.... Sep 29 16:38:45 init=/bin/systemd seem to slow it down.... Sep 29 16:41:33 bootchart as wrapper? Sep 29 16:41:51 init=/bin/systemd -> slow Sep 29 16:42:14 init=/usr/sbin/bootchartd -> fast.... Sep 29 16:44:29 interesting, does systemd-analyze say something interesting? Sep 29 16:44:37 what's that? Sep 29 16:45:59 tool to analyze systemd startup time and blame units which takes a lot of time Sep 29 16:46:37 ok Sep 29 16:46:42 I need a shell for that Sep 29 16:46:50 http://build.shr-project.org/tests/jama/bootchart/systemd-20120525/ see *blame.txt files Sep 29 16:46:53 to compare Sep 29 16:47:01 ok Sep 29 16:47:49 how many system that spawn getty are there? Sep 29 16:48:28 shr 20120913 om-gta02 ttySAC2 Sep 29 16:48:30 and no shell Sep 29 16:48:41 no login: Sep 29 16:48:43 nothing.... Sep 29 16:48:48 I need a shell..... Sep 29 16:48:54 maybe init=/bin/ash? Sep 29 16:49:10 how does systemd-analyze work? Sep 29 16:49:46 just type "systemd-analyze" for overall time and "systemd-analyze blame" for unit times Sep 29 16:50:21 yes but on target? Sep 29 16:50:28 does it need systemd started Sep 29 16:50:35 because with systemd I get no shell Sep 29 16:50:46 yes it needs shell Sep 29 16:50:49 afaik Sep 29 17:46:29 JaMa, the problem is that I tried at least 5 times to bisect the problem and always failed Sep 29 17:46:49 it was always close to infinity Sep 29 17:46:54 because of issues with git bisect Sep 29 17:47:00 it was in an unbisectable zone Sep 29 17:47:08 + I had to apply patches on top Sep 29 17:47:09 etc... Sep 29 17:49:19 I really don't know what to do Sep 29 17:49:23 that makes me very nervous Sep 29 17:49:31 and we need more recent kernel on gta02 Sep 29 17:49:43 else the device will be abandoned in the future I guess **** ENDING LOGGING AT Sun Sep 30 02:59:59 2012