[00:24] mwhudson, ok [00:30] zyga: wb [00:49] mwhudson, ok so now I can run lava commands via celery [00:49] mwhudson, I guess the next step is to lava-toolify the dispatcher [00:49] mwhudson, any objections? [00:53] zyga: nope [00:53] zyga: feel free to fix assorted cruddiness on the way past :) [00:53] mwhudson, would you mid if I moved all board support to extensions? [00:53] mwhudson, and keep the dispatcher a shell for them [00:53] mwhudson, maybe moving code back to core as we see fit for it to be there (and be supported) [00:54] zyga: um, probably not [00:54] mwhudson, initially we'd have one big lava-dispatcher-legacy package with the current code [00:54] zyga: it still should be reasonably simple to install though [00:54] mwhudson, then we could see how to clean that up (maybe redesign on top of the agent code) [00:54] mwhudson, look up celery-and-redis on pypi [00:54] mwhudson, bundles :-) [00:55] pip install lava-stack -> everything -> we could bundle that in .pybundle [00:55] zyga: celery-and-redis? [00:55] ah right [00:55] mwhudson, it's a pypi package that bundles both celery and redis dependencies [00:55] yeah [00:55] something like that [00:55] mwhudson, pip install lava-and-panda ;-) [00:55] mwhudson, pip install lava-tests-for-ubuntu [00:55] or something like that [00:55] could be cool [00:55] anyway [00:55] food for thought [00:56] I spent far too much time playing with celery [00:56] let's use it [00:56] everything ends up looking a bit like debian packaging i guess [00:56] zyga: +1 [00:56] mwhudson, oh, I think I have a semi-elegant way to fix settings [00:56] mwhudson, lava.celery.settings will do magic [00:56] mwhudson, if lava-server is around then we pull configuration from lava-server [00:57] mwhudson, then extensions can contribute_to_settings() to add celery tasks [00:57] mwhudson, broker is configured in settings.conf [00:58] mwhudson, result backend, I think, should be fixed to amqp for the moment, perhaps lava-server might change that to django but I'm not really sure if that is a good choice [00:58] mwhudson, then if lava-server is not available [00:58] mwhudson, right now I hardcode settings for local development [00:58] mwhudson, we only need to support three kinds of users for this module [00:59] mwhudson, celeryd --config=lava.celery.settings # I now see the need to rename that to lava.celery.config ;) [00:59] aaah [00:59] my scheduler changes depend on a new django-restricted-resource [00:59] i think [01:00] mwhudson, lava cloudify ... # I can see it either talking to lava-server if inside one venv or perhaps having some command line switches [01:00] mwhudson, new? [01:00] mwhudson, did you patch it? [01:00] zyga: no, but the most recent release has is_owned_by(user) for group owned resources dtrt [01:00] and i had a too old one installed locally [01:00] mwhudson, ah [01:00] mwhudson, I remember that [01:00] i wonder if staging has the old version [01:01] hm no [01:06] huh [01:06] oh [01:06] zyga: the change i need for django-restricted-resource hasn't been released yet [01:06] mwhudson, oh [01:06] mwhudson, well that's a shame :) [01:07] mwhudson, go ahead and release it please :) [01:09] yeah will do [01:14] Submitting dist/django-restricted-resource-0.2.7.tar.gz to http://pypi.python.org/pypi [01:14] Upload failed (502): Bad Gateway [01:14] COME ON [01:17] I guess pycon is a bad day for any releases [01:18] it's back up now [01:19] .. and released [01:23] Hi , i just download http://releases.linaro.org/12.02/ubuntu/oneiric-images/alip/hwpack_linaro-lt-vexpress-a9_20120221-1_armel_supported.tar.gz and http://releases.linaro.org/12.02/ubuntu/oneiric-images/alip/vexpress-a9-alip.img.gz .. try with linaro-media-create but it report bad checksum. [01:27] zyga: can you look at http://staging.validation.linaro.org/scheduler/ please? [01:27] looking [01:27] if you log in (you may need to log out and in again) you should be able to view job 11979 but not 11978 [01:28] mwhudson, is that openid? [01:28] mwhudson, I still see both now [01:28] mwhudson, I'll log out and back [01:28] zyga: do you have links for the job id though? [01:28] mwhudson, no wait [01:28] mwhudson, I see both as a guest [01:28] right [01:29] no [01:29] just job numbers [01:29] but no details? [01:29] right [01:29] cool [01:29] let me log in now [01:29] mwhudson, works as expected [01:29] mwhudson, is that via openid group membership? [01:30] zyga: yes [01:30] the linaro group on launchpad in this case [01:31] mwhudson, how does it look like in python> [01:31] zyga: thanks for checking [01:31] zyga: what do you mean? [01:31] it looks like OPENID_ENABLE_AUTO_TEAM_MAPPING=True [01:31] mwhudson, does that create "linaro" team in django? [01:31] yes [01:31] mwhudson, and makes me a member of that team? [01:31] no [01:31] wow, cool :D [01:31] :D [01:31] if you create a team in django [01:31] ah [01:32] sane and cool [01:32] it will ask launchpad if you are a member of that team when you log in [01:32] does it check back to kick me out? [01:32] I see [01:32] so we could create a group called ubuntu-core-dev, if we felt like it [01:32] mwhudson, I have a proof-of-concept that captures stdout and sends that back [01:32] nice, really nice [01:32] so [01:32] maybe [01:32] the only issue is that it all happens at log in time [01:33] it's time for lava-lab-job-submitters on launchpad? [01:33] so if someone leaves a team on lp, that isn't reflected until they log in again [01:33] zyga: yes, i think so [01:33] mwhudson, right, that is expected [01:33] mwhudson, good work :) [01:36] zyga: btw [01:36] zyga: anonymous bundle streams are a terrible terrible hack [01:36] i guess you knew this :) [01:36] D:D:D [01:37] I have only myself to blame [01:37] and OAuth [01:37] heh yes, always blame oauth [01:37] I think they have value but are gravely misused [01:37] well [01:38] i was more thinking of the implementation [01:38] user = User.objects.get_or_create(username="anonymous-owner")[0] === davidrusling_ is now known as davidrusling [01:39] robclark kicks it :) [01:39] aaaa [01:39] that is indeed super ugly [01:39] I forgot about that [01:39] I wrote that after d-r-r came to be [01:39] mwhudson, and required ownership of anything [01:40] (django-drm ;-) [01:40] yeah, i can see where it comes from [01:40] mwhudson, it would be nice to cure this one day [01:42] zyga: in d-r-r? [01:42] django-restricted-resource [01:44] mwhudson, OMG [01:44] mwhudson, :D [01:44] mwhudson, OMG [01:44] mwhudson, lol [01:44] zyga: ? [01:44] mwhudson, so have I told you that by accident I can run lava-test remotely via celery [01:44] mwhudson, so now I can run celeryd on a board and run tests there [01:44] :D [01:44] haha [01:45] plars, ^^^^ [01:45] well, that's nice [01:45] so where are we with that agent idea? [01:45] what are the minimal deps to run celeryd? [01:45] lava agent -> run celeryd setup for lava-lab with config in preseeded directory [01:46] mwhudson, kombu, pure python stuff for rabbitmq, celery [01:46] mwhudson, w8 [01:46] mwhudson, anyjson, ampqlib, kombu [01:46] mwhudson, and python-dateutil [01:46] mwhudson, I did not think about this [01:47] mwhudson, this is so cool, we could even _wait_ for the board to be "up" enough [01:47] mwhudson, I see some downsides [01:47] mwhudson, we don't have any yet [01:47] mwhudson, but any network test might disrupt this [01:48] mwhudson, unless this is designed correctly or we run ppp on the serial line [01:48] still [01:48] food for thought :-) [01:48] plars, I told you celery is _the_ thing ;-) [01:48] Dear all : Where to find the document that describe the definition of linaro 'm', 'n'. 'o' ? [01:49] wowon, where did you hear about 'm' 'n' 'o' [01:50] wowon: I'm guessing you are talking about the series of release? This would be for the ubuntu releases based on ubuntu {m=maverick, n=natty, o=oneiric, p=precise} [01:50] ah [01:51] zyga: you accidentally made lava totally distributed? (there's a funny parody of a meme there somewhere I think) [01:51] plars, well all lava-tool [01:51] kiko, I wanted to also include the corresponding drm-prime patch, but I'm still in the middle of rebasing and refactoring/simplifying some of that, but since folks are waiting I decided I couldn't wait for sending out the dma-buf patch/rfc.. [01:51] hopefully will have the drm-prime part of it tomorrow [01:51] plars, lava-test would make a fine agent to run on the board [01:51] plars, we could simply start it after booting [01:51] plars, then just talk to it directly via the API [01:53] plars, well not via API but still close enough [01:53] plars, single codebase is still an annoying factor [01:53] zyga: interesting [01:54] plars : thankyou for your enlightment . so .. is it true that linaro-o-server-tar-20120221-0.tar.gz means --> linaro oneiric server ? [01:55] wowon: yes, linaro-server image, based on ubuntu oneiric, and it's the daily build from February 21, 2012 [01:57] plars : thankyou ... can I later install x11 on it ? ... is it just like (for pc) ubuntu-server version ? [01:59] zyga: eyeballs please? https://code.launchpad.net/~mwhudson/lava-scheduler/private-job-tweaks/+merge/97556 [01:59] (pretty simple) [01:59] applying eyeballs [02:06] mwhudson, done [02:06] mwhudson, nothing wrong with the code but I'd like a change there please [02:06] zyga: ok [02:06] zyga: i don't think d-r-r and the django permissions system intersect at all do they? [02:06] mwhudson, by design, they don't [02:07] mwhudson, there's a description why [02:07] mwhudson, in short: permissions are complicated [02:07] mwhudson, and that was not the point of d-r-r [02:07] mwhudson, user code should implement permissions [02:07] ok [02:07] sounds sensible [02:07] mwhudson, and use the ownership and sharing enums to aid in that [02:08] mwhudson, I was hoping that django 1.3 would bring a better permission system [02:08] mwhudson, and I think it did [02:10] All ..... is there any support for RockChip RK-2918 Cortex-A9 ? [02:12] wowon, linaro focuses on member hardware, that SoC would likely get community support only (whatever is in the current upstream kernel) [02:12] wowon, I'm not familiar with that chip though [02:13] mwhudson, I just looked, I think django has a three-argument has_perm(with the action _and_ the target) but I cannot find any [02:13] mwhudson, I know there are 3rd party extensions that provide that [02:13] mwhudson, perhaps they somehow plug into django but it seems my desire to have a generic permission API is unfulfilled [02:15] zyga: so um [02:15] zyga: i think one can upload using put to any bundle you can see [02:15] mwhudson, :D [02:15] mwhudson, bug? [02:15] looks like it [02:15] mwhudson, ouch [02:15] lava-dashboard-tool put --dashboard-url http://127.0.0.1:8000/RPC2/ 4beb2b83f425a4be77107582f37ab4ddf65735f5 /public/personal/zkrynicki/testing/ [02:15] mwhudson, there goes security [02:15] Stored as bundle 4beb2b83f425a4be77107582f37ab4ddf65735f5 [02:16] so yeah [02:16] * zyga looks at the code [02:16] Zyga : thankyou for your info ... so the only chance for me is to try it my self. anyway ... that RK2918 is commonly used by chinese manufacturer .... their android setop box is priced low. [02:17] wowon, they are free to become linaro members, partners or to work as a community supported variant [02:19] mwhudson, I'm fixing that [02:19] zyga: ok [02:19] mwhudson, I'll add a test case too [02:19] zyga: please make it an api that the scheduler can call :_) [02:20] zyga : kindly please explain me more about 'work as community supported variant' .... well i'm not the manufacturer .. i just in progress at developing such fleet management application, and I found that this kind of appliance will economically match my requirement [02:20] mwhudson, heh, fair enough :) [02:20] wowon, anyone can step up and work as community inside linaro to support non-member hardware [02:21] wowon, I'm not sure anyone has but it's not like we're going to prevent that [02:26] Zyga : Thankyou. Argh .. i'll learn more using qemu ..... and be back soon I have more infos. Thanks [02:38] mwhudson, sorry, on it now [02:46] zyga: let me know when you have something you want me to review [02:46] mwhudson, almost done [02:47] mwhudson, I've realized how much testing with lava-server manage -d test sucks [02:47] mwhudson, plust how many errors there are in the dashboard test suite now [02:48] mwhudson, aww [02:48] mwhudson, now I want your factory [02:48] mwhudson, I have my own but yours is cleaner [02:50] zyga: at least manage -d test completes fairly quickly (if that's using sqlite) [03:00] mwhudson, is there a bug on that? [03:00] mwhudson, I've got the core ready [03:03] mwhudson, do you have the bug number? [03:03] zyga: no, i can file one [03:04] thanks [03:05] * zyga builds docs [03:05] zyga: https://bugs.launchpad.net/lava-dashboard/+bug/955669 [03:05] Launchpad bug 955669 in lava-dashboard "anyone who can see a bundle stream can place a bundle into it" [Undecided,New] [03:05] mwhudson, thanks [03:05] * mwhudson wonders who will get the precious bug 1000000 [03:07] zyga: i need to leave in 10 mins btw [03:07] * zyga hurries up [03:07] zyga: i can check in later, but probably (hopefully?) after you're asleep === DrKmobs[0] is now known as DrKmobs [03:09] ok [03:09] mwhudson, the branch is up [03:10] https://code.launchpad.net/~zkrynicki/lava-dashboard/fix-955669/+merge/97563 [03:11] mwhudson, I guess I'll wait for your review [03:11] mwhudson, and call it a day [03:11] zyga: +1 [03:11] plars, tomorrow I'll release alpha cloudified dispatcher [03:12] zyga: clicked around on lp too [03:13] mwhudson, merged [03:13] zyga: thanks! [03:13] zyga: i assume if i update my branch to call the new method, you'll be happy with it? [03:13] mwhudson, yes :) [03:13] cool [03:13] ttyl [03:13] are you going back today [03:14] or is 10min your daily EOD [03:15] mwhudson, remember you need lava-dashboard >= 0.13.dev [03:15] question about auto root login , how to disable it ? [03:16] zyga: no, i'm stopping early today [03:16] (started early too) [03:18] I see [03:18] see you tomorrow than [03:19] good night === michaelh1 is now known as michaelh1|away [05:08] hi ... kindly please help me .... i really really need to disable the auto root login ... how to do it ?? [06:45] springz: liuyq: do you know where I can find a description/doc of metadata in lava json file? [06:47] fabo, in job definition file or result? [06:47] springz: job definition [06:48] fabo, e.g. http://validation.linaro.org/lava-server/scheduler/job/15586/definition [06:49] fabo, "metadata": { [06:49] fabo, others can be retrieved when test image running [06:50] springz: I don't get it. for example, I'm looking for all key/value pairs that I can put in metadata section. [06:51] springz: the job definition of a running job will give me only the one that have been specified by the user, if I understand correctly [06:52] fabo, right, I don't know where all pairs defined, need to look for [06:55] fabo, springz, maybe there is no definite definition, we can set any key and value pair in the metadata section === e-ndy|afk is now known as e-ndy [06:58] fabo, liuyq yes, if you mean the hidden ones which will collect by lava itself, from lava-dispatcher code, there are like target.hostname, device_type [07:00] springz: liuyq: ok, sounds good. thanks! [07:28] bhoj: ping [07:33] hrw: I've created oneiric-armel-alip-dev successfully [07:34] hrw: I started from my precise-armhf-alip configuration [07:34] it doesn't contain your clean up script yet, I'm testing now === michaelh1|away is now known as michaelh1 [08:05] ndec: a colleague tells me he got null WS working on beagle and panda boards [08:18] HI, I'm using arm-linux-gnueabi from ubuntu, i found that it can only produce code for armv7, i appended the -mcpu paramter according the linaro ppa's instruction, but it doesn't work.please help [08:23] wenyi: it can enlight you -> http://comments.gmane.org/gmane.linux.linaro.toolchain/2185 [08:27] hrw: it worked, including your scripts. config available on http://people.linaro.org/~fboudra/live-build/oneiric-armel-alip-dev.tar.bz2 [08:28] Hi fabo .... know I have linaro image with plain openbox and slim ... build from nano. Currently i'm trying on compiling Navit with it. [08:29] wowon: good progress :) [08:29] fabo: Thanks.I've read the link, and it's helpful. [08:54] aviksil, hi [08:55] bhoj: in my local fast models setup, it takes around 20 s to complete "sleep 1". is it the usual behavior? [08:56] aviksil, yes. its correct. the simulated clock itself is not accurate. [08:56] bhoj: hmm, ok [08:57] bhoj: and do you know how much it takes to run "memtester 64M 1"? [08:57] bhoj: for "memtester 4K 1" it takes few minutes for me [08:58] aviksil, I have never run it [08:58] bhoj: oh ok [08:58] aviksil, may be pundiramit would have tried it . [08:59] bhoj: ok, thanks. will ask him [09:19] i am probably missing something: i crosscompiled my own kernel for the imx53 following some wiki link, then installed the deb into my linaro-ubuntu-natty install but the system isn't picking up the new kernel automagically as it does in my workstation. since there's no grub, is there some extra step i am missing? [09:19] hi [09:19] fabo: fetched [09:20] c10ud: boot board, run flash-kernel [09:21] hrw, thanks, trying asap [09:38] fabo: same problem with your files. will check in oneiric chroot [09:41] heeen: ok... that's certainly possible. that must be with a different DDK than the one we currently maintain... do you know where he got his? there are indeed several different flavours of it.. [09:42] "kernel whatever2.6.38-linaro-mx51 does not match your subarchitecture mx5, therefore not writing it to flash" ..any clues? [09:42] hrw: what's your host? I'm using precise/amd64 [09:43] fabo: precise/amd64 [09:44] hmm I can't reproduce here... [09:50] * michaelh1 waves goodnight === michaelh1 is now known as michaelh1|away [09:50] ndec: romaxa.info/qws-embed/sgx/ [09:55] isn't there some flash-kernel "by hand"? i tried getting an updated version but no luck [09:56] it still complains mx51 != mx5 [09:56] but i cloned from linaro/ubuntu/linaro-natty so i don't know what could be wrong (and used the provided debian folder) [10:00] ohh flash-kernel is a script, now let's see [10:04] c10ud: which mx51 boards are you using? i have a babbage and a sharp pcz1 that i do plan to try get opengl es running on when i get some time [10:05] imx53 but i'm starting to think this kernel is no good for [10:05] it [10:08] the ubuntu flash-kernel has mx5 support since ages, make sure your kernel package is named properly [10:08] mx51 is dead in ubuntu [10:09] ogra_, i found that out the hard way, because it isn't booting up ;) [10:09] heh [10:10] anyway, i cloned from http://git.linaro.org/git/ubuntu/linux-linaro-natty.git i thought it had the right kernels for each board (and maybe mx51 was something like mx5x) [10:10] looks like i was wrong [10:11] mx51 is the old freescale babbage board thats long out of production [10:11] last support for it in ubuntu was 10.04 [10:12] i see. then i just need to find the upstream git for the kernel i was running on the board [10:12] oh man, why don't you build with some more drivers [10:13] eheh [10:13] note there is an ubuntu image for the quickstart board that should work out of the box [10:13] is your HW something different ? [10:14] nope, everything was working well (i created the os with the linaro tools quite some time ago, hence the need for natty kernel..should be 11.04 iirc) but i needed drivers for ftdi and some wifi dongles [10:15] ah [10:19] wow [10:19] good === danilo_ is now known as danilos [10:19] hi [10:20] what will be minimum size for LPDDR for ICS [10:20] fabo: same in clean oneiric/amd64 chroot [10:21] jk_: I would say 512MB [10:21] is it possible to cut down still... by removing things realted to camera, gps and BT etc [10:22] jk_: my nexus s has 512MB ram which means ~380MB for system and it gives feel 'I want more ram' [10:23] really... does this means..very slow when you use camera [10:24] hard to tell but I feel that ICS requires dualcore cpus - it is slow by itself [10:24] yes, i agree with you Dual core is a good option [10:25] but nexus s does not have dual core i suppose [10:25] ics on archos tablet (omap4430, 512mb ram) works nicer then on nexus s (onecore a8/512mb ram) [10:26] that is good... [10:27] omap 4430 is dua core and also has DSP on board I supose [10:28] and more [10:28] yes some h/w acce units for vedio [10:29] do u think archos tablet PC has open source version of linux 3.0 and ICS on it [10:29] nope [10:30] but you will be able to install it soon [10:30] hi av500, thanks ...join command worked well [10:30] hrw: I'm trying on ci as I don't have any clue on what's going on and the differences in our precise/amd64 build host [10:30] fabo: neither do I - checking clean precise chroot now [10:30] fabo, still issues with qemu ? [10:30] av500, soon ? Do you know how it could be done ? I could add any packages since they have a squashfs image on that thing. [10:30] fabo: anyway I will create configs for my images and then push to ci [10:31] bhoj: av500 is archos guy ;) [10:31] hrw, aahhh :) . [10:31] bhoj: and you can run most of archos tablets with ext4 rootfs [10:31] good [10:31] av500 [10:31] i have acer tablet [10:31] ogra_: no, I have a work around for qemu and pm215 has fixed qemu-linaro. 2012.03 should land today. [10:31] jk_: heresy! [10:32] but it is very old model c200 [10:32] fabo, well, i dont see the issue anyway on x86 :) [10:32] hrw, I was expecting archos g9 tab to have standard android partitions but it did not :) . [10:32] bhoj: true [10:32] since we made tablets longer than aapl and google, we had our boot scheme already :) [10:32] bhoj: I have read that you can hack a bit and get it running from mmcblk0p3 [10:33] really av500 [10:33] do you folks take tablet design contract wok [10:33] depends [10:33] means what [10:33] hrw, haven't explored it much . need to try it out :) . [10:33] do you have money? :) [10:33] money...yes VCs have it [10:33] bhoj: basically go to xda develoeprs [10:34] great, send some [10:34] if you give good cost, i can aske VC to pay u [10:34] i hope av500 is not from china [10:35] 怎么回事呢？ [10:35] my worry all on LPDDR and fall out elpida [10:35] wowwwwwwwww nooo yea.. i can not read [10:35] neither can I [10:35] what does this means [10:35] then it is ok [10:35] same boat [10:36] slow boat to china... [10:36] do u think 512 MB with OMAP 4460 will be good tablet oc [10:37] oc? [10:37] oc means... free [10:38] how come u did not know this one [10:38] when we wre stiudnets, if some ask some thing ..we call it oc [10:38] students... [10:39]