**** BEGIN LOGGING AT Tue Jan 19 02:59:57 2010 Jan 19 09:28:27 ogra: /win_xp/5 Jan 19 09:28:51 ogra: sorry. typo Jan 19 11:16:44 cooloney, intrestingly i think last time i tried suspend worked for me, it just didnt resume Jan 19 11:17:05 i.e. there is no way to wake up the board Jan 19 11:58:10 asac: The logic to disable CONFIG_NEON in the kernel because chromium will pick it up if present seems bogus to me (LP #507416) Jan 19 11:58:11 lool: Bug 507416 on http://launchpad.net/bugs/507416 is private Jan 19 11:59:23 lool, as i understood the plan is to switch it in the kernel at runtime/boottime Jan 19 11:59:38 ogra: Read the bug Jan 19 12:00:16 i'm subscribed to it Jan 19 12:00:35 however it doesnt match what was discussed on IRC Jan 19 12:00:54 as i understood cooloney does have a patch that can switch it dynamically Jan 19 12:01:07 Yes, that's what he says in the bug Jan 19 12:01:16 right Jan 19 12:01:31 so apps can look at cpuinfo and switch accordingly Jan 19 12:01:46 It's not cpuinfo but whatever Jan 19 12:01:47 if apps cant do it we'll have a -neon package of selected apps Jan 19 12:11:07 lool: yes, the way he refers to chromium is odd ;) Jan 19 12:11:11 i think thats none-sense Jan 19 12:11:17 chromium will use either no NEON Jan 19 12:11:25 or runtime detection ... Jan 19 12:11:33 but upstream doesnt want to do that so the patch would need to come from us Jan 19 12:11:58 so plan is to make a special NEON build available so we can compare benchmark Jan 19 12:12:07 and if its worth it consider doing the real implementation Jan 19 12:26:20 hi guys, am i wrong with the NEON issue? Jan 19 12:26:39 i am quite confused indeed, heh Jan 19 12:30:22 thats the drugs :P Jan 19 14:40:37 plars: so Jan 19 14:40:47 asac: yes? Jan 19 14:40:52 plars: if you install dove alternative ... and use ssh to log in Jan 19 14:40:54 and start apps Jan 19 14:40:58 do you still get the lock-up? Jan 19 14:41:03 not sure if you already answered that Jan 19 14:41:42 asac: haven't tried since last week, but I have no reason to suggest that it changed. All I was doing was booting in single - works booting to full gui - doesn't Jan 19 14:42:13 plars: right. so what we should do in any case is figure if its really dependent for some comüponent Jan 19 14:42:20 asac: with alternate, it recycles through the GDM constantly many times before locking up. With live image, the login is bypassed and it goes right into a session, where it locks before coming up completely Jan 19 14:42:33 so imo logging in using ssh and seeing if X apps still work would be great to narrow down to X server/graphics driver Jan 19 14:42:37 Could be worth trying to bring Xorg by hand Jan 19 14:43:02 plars: yes, i mean: try to not start X ... check if X apps alone over ssh trigger this. if not try to start X with stripped down features etc. Jan 19 14:43:05 until we know what is causing this. Jan 19 14:44:04 asac: I'll install with latest alt and give it a try in a bit, have to run to a doctor appointment right now Jan 19 14:45:26 ok thanks Jan 19 14:45:33 plars: the gdm restarts ... are those SIGILLs? Jan 19 14:46:13 If we get SIGILL, we at least have some chance to debug... Jan 19 14:46:53 asac: didn't catch if they were or not, very possibly Jan 19 14:48:54 so gdm restarts ... gnome-panel restarts Jan 19 14:49:01 when does it lock-up completely? Jan 19 14:49:16 just random i guess Jan 19 14:53:03 NCommander: can you kick off a gcc-4.4 package build on your dove? we might get something out of that from the testsuites run at the end of the build Jan 19 14:58:42 ok out for some errand Jan 19 15:30:19 dmart, ping Jan 19 15:51:00 dmart, unping (solved the issue) Jan 19 15:51:24 unpong Jan 19 15:51:31 heh Jan 19 16:10:35 Now that is a true sign of nerddom. Jan 19 18:14:25 asac: btw, chromium works fine here with v8 disabled... Jan 19 18:14:33 iow, it fails to run on armv5te only Jan 19 18:14:58 with v8 enabled v8 fails to build, with it disabled it fails to run :D Jan 19 18:15:07 however on armv7 works both ways Jan 19 18:15:34 guess it fails on armin76: v8 disabled? Jan 19 19:06:07 so basically nothing? Jan 19 19:11:45 asac: you know, the snapshot=false thing Jan 19 19:11:48 javascript works anyway Jan 19 19:11:56 it just doesn't on armv5te Jan 19 19:20:30 kk Jan 19 19:20:46 snapshot=true builds and just crashes? Jan 19 19:20:49 armin76: ? Jan 19 19:32:32 asac: on armv5te, snapshot=true fails to build, false fails to render any js page Jan 19 19:35:34 armin76: fails to build or crashes during build? in mksnapshot? Jan 19 19:37:17 crashes during build, mksnapshot Jan 19 19:50:16 asac: http://dpaste.com/147596/ Jan 19 19:52:05 armin76: for us it crashed there with gcc 4.4 because of inlining bug Jan 19 19:52:16 -fno-tree-sink or -fstrict-aliasing worked Jan 19 19:52:22 but i guess you face a different issue Jan 19 20:00:21 asac: its gcc-4.3 and works on armv7, so... Jan 19 20:00:41 ok Jan 19 20:06:24 asac, re: building gcc-4.4 (sorry for the delay on this work item): my dove board keeps hanging or gcc segfaults Jan 19 20:07:56 you broke it Jan 19 20:14:55 NCommander: while building or in the testsuite? Jan 19 20:15:23 asac, while building Jan 19 20:17:20 asac, I'm going to try one more trick up my shelve before giving up completely, but its not pormising Jan 19 20:17:55 * NCommander sighs Jan 19 20:18:02 * NCommander just hung in apt-get update Jan 19 20:20:20 NCommander: are you running lucid? Jan 19 20:20:25 asac, yup Jan 19 20:40:52 NCommander: so trick didnt work? Jan 19 20:41:19 maybe this gives more when run inside gdb session? Jan 19 20:41:30 i mean... just the gcc compile Jan 19 22:24:39 wow, that's just crazy Jan 19 22:24:49 I think it's bootchart that's killing it Jan 20 00:03:50 plars: killing what? Jan 20 00:03:56 dove? Jan 20 00:04:18 asac: yep Jan 20 00:04:31 plars: wow Jan 20 00:04:41 plars: so without it, all works fine? Jan 20 00:04:44 asac: yes, not what I was expecting Jan 20 00:05:05 asac: sorta... gdm still continually restarts. Haven't had a chance to focus on that yet Jan 20 00:05:42 plars: hmm ... so skipping gdm brings you a usable desktop and its stable? Jan 20 00:06:09 asac: haven't tried bringing up X on it's own, just pulling pybootchartgui Jan 20 00:06:32 how can you log in if gdm restarts? Jan 20 00:06:41 asac: command line Jan 20 00:06:49 ok; so no X Jan 20 00:07:18 at least a small improve Jan 20 00:07:21 asac: I suspect that X may start ok on its own, just didn't get a chance to try yet Jan 20 00:07:32 asac: yes, we have a workaround to make progress at least :) Jan 20 00:10:38 ok. when can you try that? Jan 20 00:11:02 also, didnt someone say that gdm was fixed and now gnome-panel restarts? Jan 20 00:15:25 asac: I can mess with it more tonight, trying to get kids fed right now Jan 20 00:38:35 asac: I'm starting an alternate install now so I have an installation to work from (easier than modifying the live image). Jan 20 00:38:49 great Jan 20 00:50:53 asac, GrueMaster: I had several packages out of date, after the update gdm was stable, but it hung once I got into X again. Looks like there is more than one thing that can cause the hang. I did try running a few gui programs (gcalctool, oowriter, etc) and nothing hung when running over ssh Jan 20 01:04:22 plars: do you have a Xorg log from that run? Jan 20 01:05:11 plars: another thing to test would be to just start X with a single app Jan 20 01:14:41 I'm wondering it it may be in the exception handler. It would be nice to know what is different in X0. **** ENDING LOGGING AT Wed Jan 20 02:59:57 2010