15:00:34 #startmeeting 15:00:34 Meeting started Thu Aug 18 15:00:34 2011 UTC. The chair is NCommander. Information about MeetBot at http://wiki.ubuntu.com/AlanBell/mootbot. 15:00:34 Useful Commands: #topic #action #link #idea #voters #vote #chair #action #agreed #help #info #endmeeting. 15:01:11 #link https://wiki.ubuntu.com/ARM/Meeting/2011/20110818 15:01:41 hrm, new boot, and it seems to be lagging :-( 15:01:57 #topic Standing Items 15:01:57 TOPIC: Standing Items 15:02:01 G'day all 15:02:07 #link http://people.canonical.com/~platform/workitems/oneiric/ubuntu-armel.html 15:02:14 #link http://people.canonical.com/~platform/workitems/oneiric/ubuntu-armel-ubuntu-11.10-beta-1.html 15:02:26 you can use ubuntu-arm now 15:02:38 (i fixed that last week) 15:02:48 I can't decide if my laptop is completely foobared, or if its theinternet here :-( 15:02:57 * NCommander is lagging something feirce 15:03:18 http://people.canonical.com/~platform/workitems/oneiric/ubuntu-arm.html 15:03:18 regardless the burndown chart doesn't look happy 15:03:24 http://people.canonical.com/~platform/workitems/oneiric/ubuntu-arm-ubuntu-11.10-beta-1.html 15:03:35 thank you ogra_ 15:03:36 the beta one looks fine 15:03:54 we're a little above trend but not earth shattering 15:03:55 ugh, yeah, I'm lagging something hidiously bad 15:04:07 so I'll take your word for it (the image won't even load properly on thenew link for me) 15:04:14 heh 15:04:27 move :) 15:04:29 (the trendline for ubuntu-arm shows '12' as the start point so .... yeah) 15:04:40 yeah, ignore th etrend lines 15:04:42 #topic ARM Server Status (NCommander, Daviey) 15:04:42 TOPIC: ARM Server Status (NCommander, Daviey) 15:05:03 LXC on ARM should work out of the box once a new meta package to ship 2101 out of the box goes 15:05:04 i dont even know if there is a way to adjust it on other than the ubuntu-arm.html page 15:05:07 *1201 15:05:16 I'm now looking at openstack on ARM 15:05:41 and have some concerns since it seems to want an EC2-style image which we don't build for OMAP. Further research required 15:05:43 NCommander, is there anything i should change regarding the ac100 kernel ? (the server team uses dynabooks) 15:05:55 [action] NCommander to review openstack image requirements and report back 15:05:55 * meetingology NCommander to review openstack image requirements and report back 15:06:02 oh good, the old alias work 15:06:03 wrt LXC i mean 15:06:28 ogra_: run lxc-checkconfig. Ignore the namespace error, and make sure you can mount a cgroup with -o clone_children 15:06:37 k, thx 15:06:39 Ideally, just kick off a LXC instance 15:06:40 will test that 15:06:59 lxc-create does most of the hard work, poke me on IRC 15:07:13 will do, if i dont get along 15:07:13 My ac100 is currently dead and running natty so it won't be much help 15:07:23 ogra_: You could just diff your config with the omap4 one, ignore hardware bits, and make the rest match. :P 15:07:26 there should be an oneiric installer soon 15:07:29 #topic Kernel Status (cooloney, ppisati) 15:07:29 TOPIC: Kernel Status (cooloney, ppisati) 15:07:36 * NCommander LAGGGGGGGGGGGGGGGGS 15:07:46 infinity, yeahm, thats what i usually try ... 15:08:00 omap4 meta has been updated to 1202.6 (latest TI BSP + lxc fixes) 15:08:17 on the other end, work continues on the "arm scheduler bug" 15:08:18 Sweet. Next daily should then have LXC out of the box 15:08:28 (more or less) 15:08:34 found (and fixed) an smp clock bug, but didn't help 15:08:35 if you get a daily :) 15:08:41 * NCommander smacks ogra_ 15:09:04 that's an unhappy thought, and unhappy thoughts are disallowed while I'm chairing a meeting 15:09:21 heh 15:09:40 profiling is undergoing, but so far i didn't find the bttleneck yet 15:09:46 ppisati: sounds like the SMP issue on panda is proving to be an extremely difficult nut to crack 15:10:00 s/on panda// 15:10:06 NCommander: definitely, but i've started looking at it just yesterday, so... 15:10:11 let's see 15:10:21 ppisati, I think we do not have latencytop enabled in our omap4 kernel 15:10:21 infinity: point taken 15:10:24 but yes, it's nasty 15:10:42 I tried to use that to debug the issues two weeks ago, and then found it is ok on nosmp 15:10:52 janimo: someone told me that latencytop was broken on arm, but don't remember who 15:11:06 ppisati, Linaro have it enabled, so can't hurt 15:11:10 ah ok 15:11:14 i'll try it ou 15:11:15 t 15:11:16 I hear dit was ok un UP, not too ok on SMP 15:11:19 using perf right now 15:11:52 ah yes, and the sgw fix from ricardo :) 15:12:01 not pulled yet, but will shortly 15:12:12 ppisati, great :) 15:12:27 that's all from my side 15:12:36 good to have you back :) 15:12:38 ppisati: bestof luck. If you manage to fry your brain on this bug, then you've officially made it as a member of the ARM developer community 15:12:47 :) 15:12:48 heh 15:12:51 i will 15:13:06 [topic] ARM Porting/FTBFS status (NCommander, janimo) 15:13:06 TOPIC: ARM Porting/FTBFS status (NCommander, janimo) 15:13:39 main is looking happier 15:13:44 Yup. 15:13:56 And I have a ton of uploads sitting here to re-transition ocaml today. 15:14:02 who fixed ocaml? 15:14:04 Ah 15:14:17 Debian maintainer fixed it, I uploaded said fix. 15:14:24 Was it an underlying bug with binutil's handling Cortex A8 errta? 15:14:32 Not really. 15:14:49 nuts 15:14:58 that was the train of thought I was working on before life got busy 15:15:05 More undefined behaviour that required us to jam in an extra section and make everything happy again. 15:15:21 omg ! 15:15:26 libx86 fails 15:15:27 ocaml using packages still segfault 15:15:31 someone fix that ! 15:15:32 I saw a few today 15:15:35 I mean, one could argue it's binutils' fault, but ocaml was hand-crafting bits in a "workaround the compiler" sort of way, so their own foot they shot. 15:15:41 ogra_: $#!@I NFU'ed that package last cycle 15:15:50 janimo: Yes, see above about re-transitioning. 15:15:58 infinity, oki 15:16:02 janimo: Anything build before my last ocaml upload is broken. 15:16:19 s/build/built/ 15:16:32 infinity, I gave back two packages today and yesterday and they segfaulted 15:16:50 maybe you uploaded a new ocaml yesterday? 15:17:10 or they need explicit reupload for transitioning? 15:17:50 janimo: Which 2? If they depend on other non-transitioned stuff, they'd still break. 15:17:55 janimo: There's an order here. :) 15:18:40 pfft, orders .... just throw the whole bunch at the buildd until nothing returns :P 15:18:41 order? In the builds? awesome news :) 15:18:47 I'll check to see which exactly 15:18:48 janimo: (But I'd appreciate data points anyway, if it's only half-fixed, I'd like to know... I suspect you're just bouncing packages with broken/un-transitioned build-deps, though) 15:18:51 * NCommander beats ogra_ with something heavier 15:19:01 *squeek* 15:19:08 See? Fun noises. 15:19:17 just for you :) 15:19:50 move ? 15:20:05 Anyhow. 15:20:06 #topic ARM Image Status (ogra, NCommander) 15:20:06 TOPIC: ARM Image Status (ogra, NCommander) 15:20:20 desktop is broken thanks to archive skew 15:20:38 It's broken thanks to archive breakage. 15:20:43 on monday sw-center, since tuesday mono joined 15:20:45 (ie: every arch is broken) 15:21:05 sw-center is fine again, that affected all arches 15:21:58 beyond that, ac100 image is progressing, my final livefs testbuild should finish any minute 15:22:17 server preinstalled are fine, netboot are fine again since the last d-i uplaod 15:22:33 janimo, anything about mx5 to report ? 15:23:20 ogra_, nope, waiting for the kernels to test live image 15:23:42 ah, great (i wondered if my mic was muted :) ) 15:23:52 NCommander, nothing to add from my side 15:24:09 ENOTOBIN so skipping QA status 15:24:13 #topic Any Other Business 15:24:13 TOPIC: Any Other Business 15:24:35 well, i noticed this week how much support tobin does all the time .... 15:24:49 ogra_, nope, I just found an inappropriate moment to switch to xubuntu from unity-2d so sometimes my windows disappear :) 15:24:51 true enough he is missed 15:25:06 On the flip side, the pull of gravity in the house is no longer slated towards his office 15:25:09 * NCommander runs away 15:25:16 ogra_, same can be said about you 15:25:27 You running from all the things that are now heading your way? :P 15:25:32 well, my vacation is done now :) 15:25:35 pfft, we should just all go on vacation at the same time 15:25:47 No one wouldbe missed if the ARM team disappeared in December 15:25:48 and I'm very happy it is for sure 15:26:06 davidm, which remids me about the missing paperwork *grin* 15:26:09 OK one item is folks we would like to see at UDS 15:26:24 * NCommander will find some names 15:26:25 David Copperfield? 15:26:36 infinity, huh, really ? 15:26:38 Anyone we really need to sponsor? 15:26:50 * ogra_ would prefer claudia schiffer then 15:26:51 ogra_: It was the first sarcastic answer to pop to mind. :P 15:26:58 ogra_: But yours is better. 15:27:04 I've seen David Copperfield he is quite good 15:27:21 davidm: I assume we have our usual Linaro suspects coming already? 15:27:27 maybe he can make all ARM bugs go away 15:27:32 we dont sponsor linaro :) 15:27:42 ogra_: I know/figured. 15:27:47 yes Linaro will be there as their Linaro connect will be at UDS 15:28:15 We should probably see if we can get LXC upstream/openstack upstream to attend again 15:28:40 * ogra_ has some ideas but will have to talk to people first 15:28:42 if nothing else, I'd like to close the meeting out 15:28:46 go 15:28:59 #endmeeting