**** BEGIN LOGGING AT Fri Sep 30 02:59:57 2011 Sep 30 09:22:07 ogra_: infinity: you know that the alsa-lib updates from last night is wrong? Sep 30 09:22:30 grep SDP4430 /usr/share/alsa/ucm/Panda/* Sep 30 09:22:34 is wrong... Sep 30 09:22:55 aww, thats alsa-lib instead of alsa, right ? Sep 30 09:23:19 hmm, no Sep 30 09:23:27 whats wrong about it = Sep 30 09:23:28 ? Sep 30 09:24:12 no. the ucm files are in share/alsa/ucm Sep 30 09:24:23 do you have audio working? Sep 30 09:24:32 the .conf is in the subdir here Sep 30 09:24:46 * ogra_ isnt near a panda and actually on his way out atm Sep 30 09:24:57 i'll test when i come back Sep 30 09:25:16 (some RL stuff) Sep 30 09:26:56 if you run alsaucm it use the UCM config to config the driver, and the config params are taken in the script. in the Panda scripts we have thinkgs like hw:SDP4430 which will fail on Panda... Sep 30 09:27:18 GrueMaster: ^^ Sep 30 09:27:37 right Sep 30 09:27:45 so that might need a fix then Sep 30 09:28:03 infinity, ^^^ Sep 30 09:29:51 It was fixed. Sep 30 09:29:58 * ogra_ -> really out now ... will be back for the call Sep 30 09:30:44 infinity: nope. you didn't change the params in the config file. you still have references to hw:SDP4430 *inside* the files. Sep 30 09:30:59 on my panda: Sep 30 09:30:59 alsaucm -c Panda set _verb HiFi Sep 30 09:30:59 Im setting defaults Sep 30 09:30:59 ALSA lib main.c:260:(execute_sequence) unable to open ctl device 'hw:SDP4430' Sep 30 09:30:59 ALSA lib main.c:1287:(set_verb_user) error: failed to initialize new use case: HiFi Sep 30 09:31:04 Oh, feh. I forgot the sed on the version I uploaded. Sep 30 09:31:05 after the upgrade Sep 30 09:31:12 Balls. Sep 30 09:31:19 It was right in the local test version. :/ Sep 30 09:35:06 ndec1: Fixed. Sep 30 09:35:21 great. Sep 30 09:35:34 ndec1: Thanks for the catch. Annoyance of testing on one machine and uploading on another. :/ Sep 30 09:35:38 while(1) apt-get update && apt-get upgrade ;-) Sep 30 10:02:16 infinity, are idle arm builders assigned build jobs randomly or is there a prioritization? Sep 30 10:07:23 janimo: It's either alphabetical or based on a serial in the table (ie: it's just a table lookup), but you could talk to wgrant about seeing if they can be reordered. ;) Sep 30 10:07:45 infinity, would make sense to reoder and put pandas first no? Sep 30 10:08:05 janimo: I suppose, except that armel isn't keeping those anyway. Sep 30 10:08:08 It's pretty much random and difficult to change. Sep 30 10:08:28 wgrant: Shouldn't be random, surely. I'm sure it's deterministic. Sep 30 10:08:28 wgrant, so not even alphabetical as infinity said? Sep 30 10:08:41 infinity: Deterministic, but not predictable. Sep 30 10:08:41 wgrant: I'm assuming based on the order they get spit out of the table. Sep 30 10:08:46 infinity: Not any more. Sep 30 10:08:49 Oh. Sep 30 10:09:04 Because SQL is hard, and we had to do it even worse? :) Sep 30 10:09:19 Each builder has its own state machine in the Twisted daemon. Sep 30 10:09:28 Oh, right. Sep 30 10:09:29 It's entirely asynchronous. Sep 30 10:09:34 the name is fitting Sep 30 10:09:39 I forgot about the parallel rewrit. Sep 30 10:09:40 e Sep 30 10:09:45 Once a builder is done, we schedule the next scan for 15s from whenever it finished. Sep 30 10:09:55 And dispatching is done on a per-builder basis. Sep 30 10:10:09 So whenever a builder is not building anything, the next scan will query the DB to see if there are any jobs. Sep 30 10:10:35 We could possibly rework it to queue a bit differently, but at present no statement can be made about the ordering. Sep 30 10:10:54 Yeah, like I said, I forgot entirely about that rewrite. Sep 30 10:11:00 No point trying to weight them. Sep 30 10:11:05 Speaking of armhf... when's it going to happen? Sep 30 10:11:38 wgrant: Well, we may have missed our cutoff for getting a gcc upload in. :/ Sep 30 10:11:48 wgrant: So, it might just have to open with P. Sep 30 10:12:18 Yeah. Sep 30 10:12:35 Not happy about all the delays, but whatever. Life goes on. Sep 30 10:12:43 We'll just have to get an army of Pandas to make up for it. Sep 30 10:13:03 or imx53s as they are easier to get Sep 30 10:14:22 Maybe we'll have A15s of some variety eventually... Sep 30 10:21:51 infinity: bzr commit, git log -p|less. you are not alone Sep 30 11:53:59 janimo, the panda cluster took a year to happen from when i made the first proposal to today Sep 30 11:54:19 janimo, better bet for mx56 or so :) Sep 30 11:54:53 or mx58 or however freescale will call the overnext generation Sep 30 12:04:55 ogra_, a difference is pandas were scarce, whereas imx53 seems to be in good supply Sep 30 12:05:18 janimo, still you need to desing the whole crap Sep 30 12:05:30 and the needs more pressing now and also the communication better Sep 30 12:05:42 davidm, said they'd fit in the same case he designed Sep 30 13:00:27 imx53 eh? Sep 30 13:25:29 ogra, i.MX615 probably is the next big chip Sep 30 13:27:05 with a D or a Q at the end for how many cores it has Sep 30 14:12:06 actually if i disable SUSPEND*, panda doesn't hang anymore Sep 30 14:12:18 i really don't understand how they reproduced it without SUSPEND support Sep 30 14:12:20 doh Sep 30 14:12:56 yeah, me neither Sep 30 14:13:04 but they also use the pvr driver Sep 30 14:15:08 but that is still triggered only from suspend Sep 30 14:35:43 ppisati: you disabled SUSPEND support in the kernel, right? Sep 30 14:36:08 sebjan: as a test? Sep 30 14:36:15 sebjan: yes, and it doesn' hang anymore Sep 30 14:36:23 ok Sep 30 14:36:29 the kernel in the archive has Sep 30 14:36:33 (wait) Sep 30 14:37:09 http://paste.ubuntu.com/699814/ Sep 30 14:37:15 this is what i reverted Sep 30 14:37:19 this way, no more hangs Sep 30 14:37:22 well, as i said in the call, its mostly a gnome-power-manager bug .... Sep 30 14:37:30 which is being worked on Sep 30 14:37:40 so, we keep it on? Sep 30 14:37:59 well, it doesnt work on the kernel side either ... we might want to switch it off :) Sep 30 14:38:15 asd :) Sep 30 14:38:25 that suspend doesnt work and that g-p-m triggers suspend are surely two separate bugs Sep 30 14:38:41 well, actually i think the only bug is kernel side Sep 30 14:38:53 g-p-m is a bit too aggressive IMO Sep 30 14:38:58 to suspend the thing after 30mins Sep 30 14:39:02 well, the system forcefully goes to suspend after 30min idle time Sep 30 14:39:07 yep Sep 30 14:39:14 thats the g-p-m bug Sep 30 14:39:19 yep Sep 30 14:39:23 anyway Sep 30 14:39:32 if we can't work it out, i switch it off Sep 30 14:39:33 ok? Sep 30 14:39:43 yep Sep 30 14:39:52 doesnt do any good to have it on :) Sep 30 14:40:01 if it just locks the system at least Sep 30 14:40:22 righty right my friend :) Sep 30 14:40:53 yes, it's safer/easier to deactivate it for now Sep 30 14:45:27 sebjan: does it happen with Sep 30 14:45:33 sebjan: git://git.linaro.org/ubuntu/linux-linaro-oneiric.git too? Sep 30 14:48:39 ppisati: I don't know. Not sure if this tree has the TI patches Sep 30 14:49:30 ppisati: I use the tilt-linux-linaro-3.0 branch from Andy's tree Sep 30 14:50:04 sebjan: ok, and it does exhibit the problem even with no SUSPEND, right? Sep 30 14:51:18 ppisati: well, we can see a similar problem with no suspend but are unsure of its origin. It could be related to the sgx driver activation. That will be interesting to test Ubuntu kernel + sgx Sep 30 14:51:38 (in this case we don't have any suspend traces in the console) Sep 30 14:52:06 sebjan: and how do you reproduce it? Sep 30 14:52:20 sebjan: just leave the board idle? Sep 30 14:52:25 ppisati: well, just wait some time :) Sep 30 14:52:27 anyone build mongodb lately? Sep 30 14:52:31 the 2.0 release builds Sep 30 14:57:04 sebjan: half an hour is enought? Sep 30 14:59:30 ppisati: not sure. I know it happens quickly sometimes Sep 30 15:03:19 ogra_: i remember there was an installer bug yesterday. what was this problem? Sep 30 15:10:10 bug 806751 Sep 30 15:10:11 ndec1: you have a ti branch you are using to test 4460 support? Sep 30 15:10:11 Launchpad bug 806751 in debian-installer "Boot partition on SD is too small on omap/omap4" [Medium,New] https://launchpad.net/bugs/806751 Sep 30 15:10:23 bug 820621 Sep 30 15:10:25 Launchpad bug 820621 in flash-kernel "netinstall fails to make omap system bootable during install" [High,Fix released] https://launchpad.net/bugs/820621 Sep 30 15:10:33 bug 709245 Sep 30 15:10:35 Launchpad bug 709245 in linux-ti-omap4 "ARM SMP scheduler performance bug" [High,Fix committed] https://launchpad.net/bugs/709245 Sep 30 15:10:42 bug 779410 Sep 30 15:10:44 Launchpad bug 779410 in flash-kernel "package initramfs-tools 0.98.8ubuntu3 failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Medium,Fix released] https://launchpad.net/bugs/779410 Sep 30 15:11:08 all fix are in the archives? Sep 30 15:11:11 since when? Sep 30 15:13:48 ndec1, sorry, that wasnt for you, i'm preparing the release report for #ubuntu-meeting Sep 30 15:14:04 ah! Sep 30 15:16:02 ogra_: we are getting 'installer crashed' with one of our recent image. could we be missing a pkg? Sep 30 15:16:06 that you've fixed? Sep 30 15:18:52 hmm, we fixed plenty things Sep 30 15:21:30 ogra_: does the very latest image work fine ? e.g. for the installer? Sep 30 15:21:52 for me the last image worked, yeah Sep 30 15:22:03 i used it yesterday to hunt down the icon bug Sep 30 15:34:08 ndec1: There are still issues during oem-config that I see, but they are very timing specific. I can reliably reproduce them here on some SD cards, but not others. Sep 30 15:34:21 This is the 20110928 image. Sep 30 15:48:33 prpplague: there's one from the ti landing team Sep 30 15:48:56 prpplague: http://git.linaro.org/gitweb?p=people/andygreen/kernel-tilt.git;a=shortlog;h=refs/heads/tilt-linux-linaro-3.0 Sep 30 15:49:04 I know this branch works fine at 4460 Sep 30 15:49:08 rsalveti: yea ndec1 pointed me to that one Sep 30 15:49:23 there's also the tracking branch, based on 3.1 Sep 30 15:49:32 but not working properly I believe Sep 30 15:49:41 missing some hwmod changes Sep 30 15:49:47 correct Sep 30 15:49:59 prpplague: what was the question? Sep 30 15:51:49 ndec1: just need a build for some 4460 testing Sep 30 15:52:27 prpplague: funny you ask... because if I needed one i would generally ask you ;-) Sep 30 15:53:00 ndec1: hehe, well, i;'ve been testing the head of all the TI trees and everyone has something wrong with them Sep 30 15:53:41 prpplague: there isn't just 1 tree? Sep 30 15:54:04 ndec1: no Sep 30 15:54:12 ;-) Sep 30 15:58:14 haha :-) Sep 30 16:11:07 sebjan: i left the board idle for a while but no hang Sep 30 16:11:18 sebjan: titl-linux-linaro + ubuntu userspace Sep 30 16:11:32 dist-upgrade an hour ago or so Sep 30 16:21:29 sebjan: ok, if i force the suspend it hangs Sep 30 16:21:35 as expected Sep 30 18:37:48 ndec1: core 4460 support so far looks good on that one you referenced Sep 30 19:16:24 I am going to try my cruzer key (not the kingston) Sep 30 22:28:44 Hey look, the slideshow works. Sep 30 22:28:51 That's much less ugly. Sep 30 23:08:59 GrueMaster / ogra_ any of you guys know if userspace spi dev is enabled in the panda supported kernels? Sep 30 23:09:08 (for ubuntu releases that is ) **** ENDING LOGGING AT Sat Oct 01 02:59:56 2011