/srv/irclogs.linaro.org/2012/03/09/#linaro.txt

plarszyga: sure, it would help to do that, and also the rootfs extraction in tmpfs I'm sure, but we'd need to restrict it somehow, else we could *easily* overrun the available space, even with the ram we have on that box00:08
avinashhmHi friends , i am trying to boot linaro minimal image on origen using prebuilt binaries .. but i am getting below error - http://paste.ubuntu.com/875417/ .. when i try to unpack .. any pointers on what may be wrong ? 00:47
mwhudsonzyga: still movieing?00:56
avinashhmavinashhm, nevermind .. i think the hwpack was not downloaded properly .. on downloading by a weblink instead of wget, it worked clean .. thanks guys00:58
=== agreen-away is now known as agreen
=== michaelh1 is now known as michaelh1|away
=== michaelh1|away is now known as michaelh1
MustafaFHi, I have a lava question01:44
MustafaFI just deployed a test lava instance on my machine, and I am trying to add a job to it with lava-tool01:44
MustafaFbut I cant get auth-add to work01:45
MustafaFit gives me a 404 when I try to run this: lava-tool auth-add http://mustafax@localhost/lava-server/RPC2/ 01:45
MustafaFAm I doing something wrong?01:46
mwhudsonMustafaF: you probably don't want /lava-server in there01:49
MustafaFmwhudson: Thanks! I am now reaching the server01:52
MustafaFmwhudson: but the auth token is being rejected01:52
mwhudsonMustafaF: well, i guess the token or user name is wrong then?01:54
araujohi mwhudson o/01:55
mwhudsonaraujo: hello01:56
araujomwhudson, got the qatracker working before ... probably will add few tweaks here and there, specially for deployment01:57
araujo:P01:57
araujoit is very handy too :)01:57
mwhudsonaraujo: yay01:57
mwhudsonaraujo: please submit merge proposals of any improvements you make01:57
araujomwhudson, heheok :D01:57
araujomwhudson, I wanted to ask you about the LavaClient class too .... the entry point for clients ..... I can see two parameters, 'config' and 'context' .... what do they specify?01:58
mwhudsonaraujo: the config is the device config01:59
mwhudsonthe context sort of represents the whole run of the dispatcher01:59
araujommm... I see01:59
araujomwhudson, is that "run of the dispatcher" available in some constant or something?02:00
mwhudsonaraujo: no02:01
mwhudsonaraujo: what do you want it for?02:01
araujomwhudson, well, I am checking right now the LavaClient class, just trying to understand how it works , specially for the qemu support02:01
mwhudsonaraujo: so something ... the context probably ... decides which client subclass to create02:02
araujomwhudson, I was wondering if I could try some tests from the lava-server shell02:02
mwhudsonaraujo: not so easily i tihnk02:04
araujoI see02:05
araujowell.. the code looks straight forward,... except for the context thing :P02:05
mwhudsonaraujo: the context class is in lava_dispatcher/context.py02:05
mwhudsoni don't think its especially complicated...02:06
araujomwhudson, I see it there :D02:07
araujook, checking now02:07
=== michaelh1 is now known as michaelh1|away
tsnyderhi. Does anyone know if all of the 12.02 oneric images for the Origen are supposed to login root to the serial console by default on boot?  I'm trying to get the first Origen boards booting for the first time at Samsung Austin.02:15
mwhudsonargh02:33
mwhudsondjango-restricted-resource release pls02:33
* mwhudson runs out of battery02:44
mwhudsonbiab02:45
avinashhmHi friends, is there any linaro snapshots, which i can use to boot s5pv310 .. i am trying to boot SMDK v310 with latest .. but seems not booting .. and i read that s5pv310 is not supported in latest snapshots .. pls let me know any pointers02:46
plarsspringz: Hi!03:00
plarsspringz: time for sync?03:00
springzplars, hello, yes!03:00
plarsliuyq: Hi, welcome back!03:33
plarsliuyq: give me just a few more minutes to finish talking to spring and we'll talk ok?03:33
liuyqhimamura, 03:33
liuyqplars, ok03:33
mwhudsonplars: excitement! https://code.launchpad.net/~mwhudson/lava-scheduler/create-private-job/+merge/9670304:16
plarsmwhudson: cool!04:17
* mwhudson EOWs04:46
=== liuyq is now known as liuyq-afk
=== vade_ is now known as vade
fabospringz: when you remove a feature from a blueprint, please discuss it before.06:31
fabospringz: I re-added the graphical view for health status, but postponed.06:32
springzfabo, ok, it may be replaced by another implementing way, javascript06:34
fabospringz: good to know. do you think we can have it in staging at some point so I can take a look?06:38
fabonot necessarily this cycle06:38
springzfabo, no, I just start to investigate on it, and I need to find some examples06:39
faboI'll suggest to take a look to kernel ci view...06:42
plarsfabo: I always forget, do we release 3rd thursday of each month, or 2nd to last thursday?07:13
plarsfabo: yeah, spring and I talked a bit about that bp at our sync07:21
plarsfabo: I don't think the kernel ci view is right for the health check jobs, but it would be useful to have some indicator of trending07:21
faboplars: 3rd -> March 23th07:31
faboplars: you need to dig into the table, filter incomplete/complete, it requires effort07:32
plarsfabo: if it's the 3rd Thursday, then that's next week, 3/15, not 3/2307:32
faboplars: the kernel ci view is the only one graphical and you see if a test failed or not07:33
plarsfabo: for health jobs? no, not anymore07:33
plarsfabo: just click on the health status link and sort by the status :)07:33
faboplars: sorry one week before last thursday07:33
* fabo looks07:34
faboplars: you talk about http://validation.linaro.org/lava-server/scheduler/labhealth/ ?07:36
plarsfabo: yep07:36
plarsfabo: it's sorted by board by default07:36
fabowell pass/unknown/fail :)07:36
plarsfabo: but if you click the health status column twice, you'll get a reverse sort which will bubble the fails up to the top07:37
plarsfabo, springz: What I'd also like to see in this view are the comment fields (so we can see why it was offlined) and the trend07:38
fabosorry I still prefer the kernel ci view, with green/red :) in one shot I see the failed one07:38
faboinfo is there but I don't like the visual ;)07:39
plarsfabo: but think about it this way, if you mapped this onto a view like that, it would be 50 columns wide07:39
plarsfabo: I'm happy to consider other options for visualization though07:39
plarsmuch like cat-skinning, there are lots of ways we could visualize this07:40
plarsfabo: but what we have now is already a huge improvement over what we did have07:40
fabosure, one default of the kernel ci view: it's wide/big07:40
plarsmakes it *way* easier to see where there are problems, and it forces action already07:40
faboagree, -> info is there but I don't like the visual ;)07:41
* fabo dreams of personal dashboard07:41
plars+107:41
faboit's details07:41
plarshow's your javascript? :)07:41
fabohehe07:42
plarspersonal dashboard is something that we've discussed before, and would be very cool indeed07:42
faboI used one with bugzilla, it's very cool and convenient07:42
faboplars: http://bayoteers.org/blog/extensions/bud/07:44
fabothat's very cool ^^^07:44
plarsfabo: yes, but it doesn't actually let you define your own views, just aggregate them all on one page07:46
plarsbut yes, that's pretty cool too07:46
faboplars: it depends on the widget available and how powerful they are.07:47
faboI'll poke ericlr07:48
plarsfabo: with the data policy stuff, I think we also need to add some public-ok views07:48
plarsfabo: and I think the kernel-ci-view is a good one to try to apply more generally, for instance, to ubuntu-ci jobs07:48
plarsmaybe even android07:48
=== e-ndy|afk is now known as e-ndy
faboplars: btw, go to BED :)07:49
faboyou're going to miss the weekly meeting07:49
plarsfabo: oh crap07:49
plarsfabo: yeah, that's in a few hours07:49
plarsfabo: thanks, now I can't go to bed07:49
plarsnow I'm worried I won't wake up for it07:50
faboin 6h07:50
fabobetter sleep 6h than nothing ;)07:50
zygahi08:33
fabomorning zyga 08:36
zygafabo, good morning08:41
zygafabo, I'll be back after breakfast08:41
amitkvingu: perhaps you can help answer Jonas, I'm feeling like a stuck record now :)08:54
vinguamitk: dlezcano: hi, ok let me check what I can do. daniel shoudl also have arguments08:56
zygare09:04
amitkvingu: I'll let you and dlezcano handle it.09:07
davepigott_ryanharkin: ping09:27
ryanharkindavepigott: i'm not here ;)09:28
davepigott_ryanharkin: Me neither09:28
ryanharkindavepigott_: no, you have an extra underscore today09:28
ryanharkindavepigott_: so what's the story?09:29
davepigott_ryanharkin: :) OK. My plan is to take the 01/2012 lt release and use that as a master image. Then manually walk through the process to see if that works. If it does, I'll run an automated lava test and see what happens. 09:29
zygadavepigott_, hi09:29
zygadavepigott_, is your wife okay?09:29
davepigott_zyga: Well, she still hobbles around but her foot isn't broken as we feared.09:30
davepigott_zyga: It's some sort of muscular sprain or a torn ligament in her foot09:30
davepigott_zyga: I'm in the office this morning, but will be heading home for the time when my son gets home from school to help her out09:30
ryanharkindavepigott_: good plan, wget works for me, as per email09:30
davepigott_ryanharkin: That's why I'm taking that route. Eliminate the unknowns09:31
ryanharkindavepigott_: grand job09:31
=== agreen is now known as agreen-away
zygadavepigott_, I hope she recovers okay, that kind of thing can be very painful09:32
ryanharkindavepigott_: zyga: aye, it sounds awful!09:32
davepigott_zyga: Yeah, and my son is still recovering from his foot operation. When the two of them walk round the house it's like I live on the set of "Night of the Living Dead". :)09:32
ryanharkinlol :)09:33
zygaheh09:33
ryanharkinmy cat could join in, he's just has a foot operation too :)09:33
=== amitk is now known as amitk-afk
amitdkamitk-afk, hi amit10:02
pantoamitk-afk, ping if !afk10:07
=== e-ndy is now known as e-ndy|afk
aviksilfabo: ping10:36
zygacode reviews done10:40
zygafeel free to bash me for questions10:40
zyganow onto merging10:40
faboaviksil: pong10:44
aviksilfabo: the glew FTBFS for armhf in overlay may be due to older libc6-dev in ftpmaster.internal10:45
aviksiltrying to build it locally10:45
faboaviksil: ok. I fixed libpfm4 2 days ago.10:52
aviksilfabo: thanks. saw it10:54
ChiThudavepigott_, ping11:08
davepigott_ChiThu: pong11:09
ChiThudavepigott_, how is your wife ?11:09
davepigott_ChiThu: We went and had her foot x-rayed yesterday, and it's not broken. She's pulled some ligaments and bruised her muscles, but not as bad as we feared. Thanks for asking!11:10
zygabrb11:11
ChiThudavepigott_, I got a question from you about lp:~le-chi-thu/lava-dispatcher/lava-test-deployment-fixes11:12
davepigott_ChiThu: ?11:13
ChiThudavepigott_, "It's probably because I don't fully understand, but why have you removed the deployment check?" what do you mean ?11:13
davepigott_ChiThu: Ah. OK. There was a check previously to see that lava-test was installed. It seems that that check was removed in your change11:13
ChiThudavepigott_, It does the check with "which lava-test" instead.11:14
ChiThudavepigott_, and "lava-test reset" direct after.11:14
davepigott_ChiThu: But it doesn't seem to take any action if that fails. Or am I missing something?11:15
ChiThudavepigott_, the run command will check the return code and raise exception if not null.11:15
davepigott_ChiThu: Ah! OK. That explains it. Just ignore me then. :)11:16
ChiThudavepigott_, the failing panda devices turned of by scheduler . Have you check them out ? 11:18
davepigott_ChiThu: Not yet. I'm busy with vexpress at the moment11:18
ChiThudavepigott_, it is not urgent since we have enough panda devices to serve the job queue.11:19
ChiThuzyga, hi11:20
davepigott_ChiThu: :) OK. Thanks11:20
zygaChiThu, hi11:20
ChiThuzyga, the dispatcher trunk need to test before release.11:21
zygaChiThu, I read11:21
zygaChiThu, I was thinking about how to do that easily11:21
zygaChiThu, it's the hardest component to test today11:21
ChiThuzyga, if I want to do it , lava staging env.11:21
ChiThuzyga, is it possible and how ?11:22
zygaChiThu, if you want to do it in the lab it requires some skill and careful changes11:22
zygaChiThu, start with taking a few boards away 11:22
zygaChiThu, offline them in production11:22
zygaChiThu, then deploy staging from scratch11:22
zygaChiThu, add those devices 11:22
zygaChiThu, copy dispatcher configs from production11:22
zygaChiThu, then you can run a few jobs to see what's going on11:23
hrwcool. mx53loco lost contact with microsd card11:24
ChiThuzyga, we should have at least one panda reserved for staging env.11:24
zygaChiThu, we can move them back and forth manually11:24
zygaChiThu, I plan on introducing a new component11:24
zygaChiThu, that will make this painless11:25
zygaChiThu, but I need to rewrite most of the dispatcher first as that's more important to get right11:25
ChiThuzyga, it is easy to make misstake.11:25
zygaChiThu, so for now I'd just do it manually11:25
zygaChiThu, yeah, I know11:25
zygathe bad part11:25
zygais that we cannot even test this easily11:25
zyga(the test procedure cannot be tested)11:25
zygaas lab setup is impossible to replicate at home11:25
zygafor now11:26
zygaI recommend:11:26
zyga1) following devops policy11:26
zygaget a buddy, do it with someone, double-check with your partner11:26
ChiThuzyga, why the lava-dispatcher need to rewrite for this purpose ?11:26
fabopm215: I'm rebuilding qemu package with your patch. you'll get the testing result soon11:26
zygaChiThu, not for this purpose11:26
zygaplan how the test procedure looks like, write down the scenario, include copy-paste shell code if possible11:27
zygareview the scenario with your peer buddy11:27
zygause a shared screen session11:27
zygahave one person follow the scenario11:27
zygaand other acking each command before execution11:27
zygait requires discipline but people do harder things daily11:28
zygaChiThu, the dispatcher needs to be rewritten to become stable and testable11:28
zygaChiThu, and to become a nice place for extensions and not an endless place for patching and fixing 11:28
zygaanyway, that's not related11:28
zygathe new component would be instance-less device manager 11:29
zygathat actually governs stuff like serial access, physical/virtual device allocation, device history, etc11:29
zygaaccess control, billing, power management ;)11:29
ChiThuzyga, I want to deploy the latest lava-dispatcher. can you help me out ? Do what you use to do when release lava-dispatcher. I do not dear to play with staging env.11:33
zygaI don't release it so I don't know 11:33
zygahistorically I guess we probably neglected testing rigorously 11:33
zygacould you review changes since last release?11:35
zygalook at what should be tested11:35
zygaChiThu, frankly current situation is hopeless, it's so impractical to test the dispatcher I don't blame anyone11:38
zygaChiThu, but that has to change and that's my priority11:39
zygaChiThu, I want to make sure anyone can test the dispatcher in less than ten seconds without any hardware after any single patch 11:39
ChiThuzyga, what are we testing without hardware ? It is not acceptant tests before deployment anyway.11:41
zygaChiThu, no but the dispatcher is _really_ not that hardware tied11:41
zygaChiThu, the dispatcher does not run on the hardware11:41
zygaChiThu, everything else is testable and can be replaced with dummies that replicate what the hardware did 11:41
ChiThuzyga, the purpose with staging is have a mini real env to do some regression tests before deployment.11:42
zygaChiThu, yes, but that's separate11:42
zygaChiThu, it's a new level of quality11:42
zygaChiThu, staging and end-to-end testing is not what I'm after right now11:42
zygaChiThu, if we get to a point where we can test the dispatcher and make sure there are no regressions without needing the hardware but still with very high confidence levels11:43
ChiThuzyga, I can easy test the dispatch locally with my board.11:43
zygaChiThu, then overall quality will be better overnight11:43
zygaChiThu, that's the fallacy11:43
zygaChiThu, dispatcher requires hours to test11:43
zygaChiThu, it should require seconds11:43
zygaChiThu, and your board is not perfect11:43
zygaChiThu, it changes over time11:44
zygaChiThu, such tests are less different11:44
zygaer11:44
zygaless valuable, different11:44
zygaChiThu, the dispatcher is really a big makefile if you look at it, it does stuff depending on rules and output11:44
zygaChiThu, that _is_ testable11:44
zygaChiThu, sure, the output may change (new hardware fault, new software fault  - both outside the dispatcher)11:44
zygaChiThu, but that's a different class of problems11:45
zygaChiThu, and those problems can and should be solved in a different way11:45
ChiThuzyga, can you guide me through the steps need to test the lava-dispatcher on staging env today ? I need to learn it.11:50
zygayes11:50
zygado you have root access there?11:50
ChiThuzyga, I don't think so. Can use give me access ?11:50
zygadavepigott_, around?11:51
zygadavepigott_, would you mind assisting ChiThu with the whole process? I will help with high-level topics but I'd like to devote most of my time for reviews, merges and coding11:51
salgadopfalcon, you joining us?11:51
pfalconi'm in11:52
davepigott_zyga: You mean giving him root access? Sure.11:52
zygadavepigott_, no, I'm on that11:52
davepigott_zyga: You mean testing lava-dispatcher?11:52
zygadavepigott_, I meant both of you following the devops protocol to do dispatcher tests in staging11:52
davepigott_zyga: I'm testing in isolation still, on a cloud instance. Haven't used staging.11:53
zygadavepigott_, cloud would be fine as long as it can reach conmux11:53
zygadavepigott_, and we can reach the 80 port on the cloud device 11:53
davepigott_zyga: Oh. OK. Sure.11:53
zygadavepigott_, do we have any spare public IPs?11:54
davepigott_zyga: Which you can now. :)11:54
zygadavepigott_, otherwise it's control11:54
davepigott_zyga: Yep. I'll help ChiThu set up and test his cloud instance. 11:54
zygadavepigott_, will you forward 80 to something on control?11:54
zygadavepigott_, and open the router11:54
ChiThudavepigott_, so the dispatcher running cloud instance will able access the devices conmux  serial port ?11:55
davepigott_ChiThu: Which device?11:55
ChiThudavepigott_, example if I borrow a panda board from production.11:56
davepigott_ChiThu: Yeah. That's not a problem, once you're logged into control you can ssh onto the cloud instance and I can guide you from there11:58
ChiThudavepigott_, I am going for lunch now. When will you be available ?12:00
davepigott_ChiThu: ABout an hour. But only for an hour. I have to go home to help my wife when our son gets home from school12:00
ChiThudavepigott_, ok. now it is 12 PM in UK. I ping you around 1 PM UK time. ok ?12:02
davepigott_ChiThu: Perfect12:02
asac_hrw: do we have native x86 gcc 4.5 backports in a ppa as well?12:09
hrwasac_: i386/amd64 ones12:10
hrwasac_: no, cross only12:10
hrwasac_: decision was "cross only"12:10
asac_doko: is there a place to get a cool 4.5 gcc from a ppa for lucid?12:12
=== e-ndy|afk is now known as e-ndy
zygadavepigott_, quick question12:25
davepigott_zyga: sure12:25
zygadavepigott_, how do you construct udev rules for USB serial12:25
zygadavepigott_, my adapters don't have anything unique12:25
davepigott_zyga: Really? You read my guide on the wiki?12:26
zygadavepigott_, apart from the full path to the USB hub/port chain I've plugged them to12:26
zygadavepigott_, I did read it a while ago but I cannot remember and from a quick experiment it seems my hardware is not unique in any way12:26
davepigott_zyga: My snowballs all have a unique ID on the board. What version of snowball?12:26
zygaah12:26
zygano snowballs12:26
zygano12:26
zygarandom USB/serial12:27
davepigott_zyga: AH. :)12:27
zygaI thought we also use them12:27
zygasnowballs are easy :)12:27
davepigott_Well, the other way to do it, that I've used in the past, is the USB/serial adaptors I use have a unique ID12:27
=== amitk-afk is now known as amitk
zygadavepigott_, could you tell me the brand you have and the udevadm monitor --property dump?12:27
davepigott_zyga: Give me a few minutes12:28
zygathanks, it's not urgent12:29
amitkpanto: hi12:32
pantohi amitk 12:32
davepigott_zyga: It's a Belkin one. http://pastebin.ubuntu.com/875972/12:34
zygadavepigott_, thanks12:34
zygadavepigott_, so you're looking at PRODUCT?12:35
davepigott_zyga: Can't remember how I set it up. It's been months and I don't have the config any more (re-install since then) :(12:36
zygadavepigott_, I have identical output12:37
zygadavepigott_, I don't think there is anything unique there12:37
davepigott_zyga: Ah. OK. :(12:37
zygadavepigott_, apart from DEVPATH12:37
zygawhich is essentially the socket and wiring for my USB hubs12:37
davepigott_zyga: Could have sworn it was that deivce.12:37
zygadavepigott_, no worries, I'm curious if there is a brand with real serial number 12:37
zygadavepigott_, or was this a mistake early on somewhere12:37
zygaI think my beagle has died12:39
davepigott_zyga: :(12:39
zygahttps://pastebin.linaro.org/424/12:39
zygathat's what I get on "boot"12:39
fabopm215: bug updated, works for me with nano. I'm testing ubuntu-desktop now.12:42
pm215interesting. we don't want to drop the 'reserve 4G' patch though, that is fixing genuine problems in other cases12:43
fabopm215: with 'reserve 4G' patch, it fails.12:44
pm215fabo: can I get this thing to actually tell me what commands it's running in the chroot?12:47
fabopm215: I need to check. there's debug options but I'm not sure it will give you what you want12:50
pm215fabo: tracked it down, it's a gpg --import that is failing12:58
davepigott_zyga: Yep. That looks like a dead board. You're sure your sd card is ok?13:05
zygadavepigott_, the card was purchased and programmed once13:06
zygaand works on my desktop13:06
davepigott_zyga: So it's ok. Yep. Looks like a dead board. :(13:06
pm215zyga: I assume the s/w on the card is known-good on that board from before?13:10
zygapm215, yes13:10
zygapm215, I could try DDing it again or cmp'ing the card if I can find the template13:10
zygapm215, but the bootloader did load13:10
zygaotherwise there's not output13:10
pm215zyga: it's just I know that at least some recent U-Boot SPL revs don't program the SDRC correctly, so if you'd just updated the s/w then it could be you were unlucky and getting marginal RAM timings13:11
pm215but if this s/w load has worked fine up til now then I agree with davepigott_ that h/w failure is more likely13:12
geshamabac, I've changed license text in linaro-license-protection and created MP13:12
zygapm215, this device is offline, it could not update itself13:12
zygapm215, so what do you recon, hardware failure in the SoC? the MMC adapter? the card?13:12
zygastatic damage?13:12
pm215zyga: no idea, I'm afraid13:13
zygathanks for the input anyway13:13
zygawhat is SDRC? and SPL?13:13
mabacgesha, thanks. I'll take a look13:13
geshamabac, next step is to replace the ste.html file on mombin in /srv/releases.linaro.org/www/licenses/ with new one13:13
geshamabac, but I'm not sure if this could be done with branch deploy13:14
pm215SDRC is the omap3 memory controller13:14
ChiThudavepigott_, ping13:14
pm215SPL is the variant of u-boot that can act as a first-stage bootloader (replacing X-loader)13:14
davepigott_ChiThu: pong13:15
davepigott_ChiThu: OK. Can you ssh into control?13:15
ChiThudavepigott_, I have ssh  into cloud instance13:15
davepigott_ChiThu: You created one already? 13:15
ChiThudavepigott_, yes13:16
mabacgesha, approved. I'll log on to mombin and look around.13:16
davepigott_ChiThu: ok. That's good. Now, we need to install a few things first. Let me just get the list...13:16
geshamabac, thnx13:16
davepigott_ChiThu: First apt-get install gcc make13:17
zygadavepigott_, wait13:17
zygadavepigott_, why are you doing this?13:17
ChiThudavepigott_, doney13:17
zygadavepigott_, if you're using lava-deployment-tool this is _not_ needed13:18
zygadavepigott_, right?13:18
zygadavepigott_, I'd rather know our deps are bad 13:18
davepigott_zyga: Ah! OK. I wasn't using l-d-t. :)13:18
zygadavepigott_, then practice black magic "now I shall teach thu the secret art of deployment"13:18
ChiThudavepigott_, yes, I suppose I will use lava-deployment-tool13:18
zygadavepigott_, we're deploying lava now, aren't we?13:18
davepigott_zyga: I was just going to run a stand-alone dispatcher, which is what I've been doing for my tests13:19
zygaah13:19
zygain that case ignore me13:19
zyga:)13:19
davepigott_plars: ping13:19
geshamabac, I've merged it into trunk13:19
davepigott_zyga: Also, when you install conmux, there's a perl library missing that needs to be installed. Don't know if l-d-t knows about that dependency.13:20
mabacgesha, thanks. I can't find anything to suggest that the files on releases.linaro.org are taken from the branch. 13:20
ChiThudavepigott_, ok, let continue with stand-alone dispatcher. no lava-deployment-tool13:20
mabacgesha, I'm not sure yet but we might have to just replace the file manually13:20
davepigott_ChiThu: sudo add-apt-repository ppa:linaro-maintainers/tools13:21
davepigott_ChiThu: sudo apt-get update13:21
davepigott_ChiThu: sudo apt-get install conmux13:21
zygaah conmux13:21
ChiThudavepigott_, done13:24
geshamabac, if it will work than do it :) all the magic was done by Danilo13:24
geshamabac, at least we can try13:24
mabacgesha, will check in a few minutes. going to take a call right now13:25
davepigott_ChiThu: OK. Now sudo perl -MCPAN -e shell13:26
davepigott_ChiThu: In perl type: install URI::Escape13:27
davepigott_ChiThu: Then quite from perl13:27
davepigott_ChiThu: Did that succeed?13:27
davepigott_^*quit13:28
ChiThudavepigott_, waiting  "searching for the best CSPAN mirrors"13:28
davepigott_ChiThu: AH yes. OK.13:29
davepigott_ChiThu: When that's done, you create a conmux config file for the board you're going to test. The other thing you need to do is stop the conmux daemon for that device on control. Can't have two conmux daemons trying to talk to the same device13:30
zygadavepigott_, can we backup the config once you are done13:31
anmarChiThu: Hello there. around?13:31
zygadavepigott_, this seems like something worth keeping13:31
zygadavepigott_, along with the instance 13:31
ChiThuanmar, hi13:31
zygadavepigott_, (although we can turn it off, just not terminate it)13:31
zygaChiThu, thanks for doing this in the cloud, very good!13:31
davepigott_zyga: Yes. It is. I'm looking into how I can turn the instance into a new launchable image13:31
anmarChiThu: I am looking here: https://ci.linaro.org/jenkins/view/Linux%20Linaro%20Snowball%20builds/job/linux-igloo-kernel-integration-ubuntu-ux500/13:31
zygadavepigott_, I'd rather script creating from plain server image13:31
anmarChiThu: which branch are you testing in ci?13:31
ChiThuquit13:32
zygadavepigott_, we need to build juju + cheff/puppet skills13:32
davepigott_zyga: Yeah. OK. Good idea.13:32
davepigott_zyga: From scratch! :)13:32
zygadavepigott_, yes13:32
ChiThuanmar, two branches. let me check13:32
anmarChiThu: please. 13:32
anmarChiThu: it says 3.2 kernel.. but I think that is a mistake.13:32
davepigott_zyga: I'll dump a bare bones script out and then we can work on it afterwards. I'll stick it up as a google doc13:33
zygaIZAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA13:33
davepigott_zyga: ???13:33
zygaIZA13:33
zygaTATA13:34
zygaMAMA13:34
alfzyga: Hello zyga's kid :)13:35
plarsheh13:35
zygaJAS13:35
davepigott_plars: On vexpress I'm getting a timeout on read_nonblocking() during the mkfs.ext313:35
zygaalf, bingo13:35
zygaalf, she's typing that helself13:35
plarsdavepigott_: sounds like it's just really slow13:36
plarsdavepigott_: that's not something that should corrupt the whole sd though13:36
davepigott_plars: Which it is, but it means it fails the deploy_linaro13:36
ChiThuanmar, stable-ubuntu-ux500-3.2 & integration-ubuntu-ux500 of git://igloocommunity.org/git/kernel/igloo-kernel.git13:36
davepigott_plars: It doesn't with the new build I'm using.13:36
zygaYASEMINAAAAA13:36
plarsdavepigott_: did you reproduce manually?13:36
plarszyga: I guess she's feeling better? :)13:36
zygaYANIA13:37
davepigott_plars: I'm using the Jan/12 release and it doesn't corrupt the sd any more (as per Ryan's recommendation) - just that it times out and fails13:37
zygaMARIA13:38
zygare13:38
zygashe's better yeah13:38
zygabut still at home13:38
zygashe went upstairs to see what I'm doing and give me a hug13:38
anmarChiThu: ok.. so the branches are correct.13:38
anmarChiThu: but the contents for the "Tree:" is wrong in that page.13:39
anmarChiThu: it is 3.3 kernel .. not 3.213:39
davepigott_salgado: daverusling said you might know who to ask about getting my touch screen working in oneiric on a Fujitsu P1610. 13:39
davepigott_zyga: Awwww. 13:39
zygaIZA is her name13:39
davepigott_Hi Iza13:40
zygashe can write that both on paper and keyboard13:40
zyga:)13:40
* zyga is so proud13:40
zygaanyway13:40
zygaback to work13:40
faboheh :)13:40
davepigott_Very smart13:40
ChiThuanmar, I updated the page. now it says Tree: and Branch:13:41
davepigott_plars: How do I increase the read_nonblocking timeout?13:41
anmarChiThu: perfect. 13:41
anmarChiThu: that was the easy part.13:41
anmarChiThu: Now I have to figure out why is it failing the boot tests on this branch where other people are booting it just fine. 13:42
anmarplars: when you get a minute, please ping me . I have an issue with CI I want to sort out. 13:42
zygaI have a cool idea :D13:43
zygawe should absolutely measure one thing today13:43
zygathe speed of writing the test image to the card13:43
zygawe should have this for _every_ test13:43
zygaplars, can we retroactively scan this from deployment logs?13:43
ChiThuanmar, which one is failing ? 13:44
anmarChiThu: the itegration one (https://ci.linaro.org/jenkins/view/Linux%20Linaro%20Snowball%20builds/job/linux-igloo-kernel-integration-ubuntu-ux500/)13:44
anmarChiThu: look at #35 for example.13:45
anmarChiThu: I am trying to figure out why it failed. 13:45
anmarChiThu: I am still learning how the Jenkins works.13:45
anmarChiThu: hmm... looks like the CI Loop is failing not the kernel.13:47
anmarChiThu: look here https://ci.linaro.org/jenkins/job/linux-igloo-kernel-integration-ubuntu-ux500/35/consoleText13:47
ChiThuanmar, on the console log, it is something about ilegal email address..13:47
anmarChiThu: BINGO :)13:47
anmarChiThu: can you fix that? or do I have to file a bug with sombody else?13:48
ChiThuanmar, I try to fix it and make a new build13:48
anmarChiThu: can you please add my e-mail address (anmar.oueja@linaro.org) to the list of recipients as well?13:48
ChiThuanmar, it will cost you some $ :-)13:49
faboarf13:50
faboI can make it cheaper ;)13:50
anmarChiThu: I can pay you in Beer ;-)13:52
plarsanmar: sure, I have a meeting in 5 min. but in +1 hour we can talk13:54
faboChiThu zyga davepigott_ plars springz - weekly meeting in 5min13:54
faboplars: it's sorted out13:54
plarsfabo: yes13:55
plarsfabo: I know we had actions at the last meeting13:55
plarsfabo: is there any place I can easily recover those?13:55
faboplars: being added to the notified people list13:55
ChiThuanmar, something wrong with the make file. http://paste.ubuntu.com/87604513:56
faboChiThu is the maintainer but he isn't notified...13:56
plarsdavepigott_: -> #linaro-meeting?13:59
ChiThufabo, I am maintainer of the CI job itself. not the git14:00
faboChiThu: when a build fails you should be notified as well. in current case, the jenkins error about mailing is yours, not LT14:00
anmarplars: no worries. ChiThu is doing a great job taking care of me 14:02
plarsgreat, thanks ChiThu :)14:03
ChiThuanmar, I will try to build locally on my laptop to reproduce the problem later today.14:14
JesseBarkeralf, checkout lp:~glcompbench-dev/glcompbench/blur14:24
anmarChiThu: thanks 14:30
anmarlag: is this erro due to code or package of ubuntu http://paste.ubuntu.com/876045/14:31
anmars/erro/error14:31
laganmar: It looks like he's trying to do a "clean mrproper" (can you even do that?) in 'Documentation/DocBook/media'14:33
alfJesseBarker: Nice! I still get bad window texture messages, though. Also the windows are not scaled to fit to a square, which is not necessarily a problem, depends on what we want to display14:40
faborsalveti: around?14:47
JesseBarkeralf, getting pulled into a meeting, but want to follow up on that.  i didn't get errors from either the nvidia proprietary (glx/opengl) or the mesa software (egl/opengles) renderers, and i was resizing both the "canvas" and the xterms being composited.14:48
alfJesseBarker: ok, we can talk later14:49
zuhlag: anmar: FWIW, http://igloocommunity.org/irc/logs/igloo/%23igloo.2012-02-09.log.html#t2012-02-09T14:37:5514:52
zuhThat's with integration branch back then, not sure if it was 3.2 or 3.3 already14:53
zuhI guess 3.314:53
anmarzuh: thanks.... 14:57
zuhAlso for the record, mrproper target doesn't work in current integration branch either14:57
anmarlag: so what should ChiThu do. You are the kernel expert?14:58
anmarzuh: can you please file a bug against that and assign it to philippe.14:58
zuhanmar: sure15:02
dokoasac_, ubuntu-toolchain-r/{ppa,test}15:02
asac_doko: which one is the "stable backport" ?15:06
asac_ppa?15:06
asac_or test?15:06
* asac_ guesses ppa, but wants to be sure15:06
dokoasac_, https://wiki.ubuntu.com/ToolChain15:06
asac_doko: hmm. no 4.5 for lucid?15:08
asac_guess not :)15:08
dokoasac_, open your eyes15:09
dokohttps://launchpad.net/~ubuntu-toolchain-r/+archive/test/+packages15:09
asac_doko: ++15:12
asac_doko: what does the -r stand for?15:12
asac_doko: nevermind15:12
asac_“PPA for Ubuntu Toolchain Uploads (restricted)” team15:12
zuhanmar: https://bugs.launchpad.net/igloocommunity/+bug/95088315:12
ubot2`Launchpad bug 950883 in igloocommunity ""mrproper" target broken on IK 3.3" [Wishlist,New]15:12
ogra_its the right handside ... if you want a stereo toolchain you also need -l15:13
anmarzuh: thanks15:13
asac_ogra_: :)15:13
dokoasac_, the non-virtualized builds can't have non-canonical members in the group15:13
anmarzuh: i am going to up the priority since this is blocking us from getting CI loop results in place.15:13
anmarlag: got a minute to chat?15:14
lagSure15:14
asac_doko: except for linaro ppas :)15:14
anmarlag: gtalk?15:14
laganmar: Mumble or here?15:14
asac_hehe15:14
laganmar: Okay15:14
asac_doko: is it safe to install 4.5 from test and 4.6 from ppa on the same machine?15:18
dokoasac_, yes, should be, just note that the shared libs then come from 4.615:29
asac_doko: kk15:29
asac_we discussed and want to keep the 4.4 only state15:29
asac_for now15:29
asac_to see if we can get stuff work with that15:29
tgall_foolool, Hi loic, wanna discuss over irc / mumble / g+ hangout ?15:30
looltgall_foo: oh hey, whereever you prefer15:31
loolvoice sounds like a nice plus; video up to you15:31
loolskype and viber ok too15:31
loolI'm in https://talkgadget.google.com/hangouts/extras/linaro.org/lool right now15:32
tgall_foolool, well I'm already connected via mumble 15:32
tgall_foook .. I can set up for g+ .. half a moment15:32
asac_doko: natty has gcc-4.4 as default still? do you remember?15:35
* ogra_ thought natty was 4.515:35
dokoasac_, ogra is right15:38
asac_yep confirmed15:38
hrwhave a nice weekend15:49
pvrAh right, weekend already..15:50
pvrDoes anyone here have some experience working with the Freescale libEGL & OpenGL ES APIs? I can't find any documentation on their EGL extensions, and there's *1* use case in google results.15:51
pvrIt's making me crazy...15:52
=== e-ndy is now known as e-ndy|afk
pfalconmabac: Have you seen in error maillist:16:01
pfalcon  File "/srv/status.linaro.org//work-items-config/current.cfg", line 12016:01
pfalcon    'linaro-android': 2012q1,16:01
pfalcon                           ^16:01
pfalconSyntaxError: invalid syntax16:01
pfalconmabac: I don't see it being backed by any bzr code changes in either prod or staging branch, so not sure how to proceed with it16:02
pfalconsalgado: maybe you have an idea ^^^16:02
salgadopfalcon, 2012q1 should be a string, I guess?16:02
pfalconsalgado: yep. the q is where from that change went, I don't see it in bzr16:03
salgadopfalcon, where are you looking at?16:03
pfalconsalgado: I would guess it may be a "fix" for that config issue, but again, no trace in prod/staging branch16:03
pfalconsalgado: https://code.launchpad.net/~linaro-infrastructure/launchpad-work-items-tracker/linaro-staging16:04
pfalconsalgado: https://code.launchpad.net/~linaro-infrastructure/launchpad-work-items-tracker/linaro16:04
salgadopfalcon, the config lives in a separate branch16:04
pfalcondamn ;-)16:04
salgadolinaro-config, IIRC16:04
pfalconsalgado: ok, was fixed by mabac already, false alarm16:05
pfalconsalgado: and thanks for hint ;-)16:05
salgadonp :)16:07
springzzyga, hi16:09
mabacpfalcon, I was a little trigger happy earlier today and approved that change. sorry about the noise16:10
=== salgado is now known as salgado-lunch
pfalconmabac: np, thanks for fixing it. and I'm just try to do that maintenance team thing and watch the stuff ;-)16:16
zygaspringz, hi16:26
* zyga is about to eat dinner16:26
=== zyga is now known as zyga-afk
springzzyga-afk, in your comment of mp: https://code.launchpad.net/~qzhang/lava-dispatcher/add-health_check-field/+merge/96715, you mentioned to add a default value of new field, how can I add one?16:36
kikorobher! long time no talk16:36
=== salgado-lunch is now known as salgado
robherkiko: what's up?16:49
zyga-afkspringz, "default": value16:50
kikorobher, not much, basically emailed to death. what about over in austin?16:51
springzzyga-afk, got it16:51
robherkiko: just general annoyance with u-boot code and maintainer responses or lack of...16:54
nekohayorsalveti, I shall try to help you with the development of the LAVA QA Tracker.16:54
zyga-afknekohayo, great!16:55
nekohayohowever, I do not know django at all and the codebase feels alien to me currently, so I guess I need to start learning django first16:55
nekohayoI'll probably pester you folks with questions eventually16:55
zyga-afknekohayo, if you need help just ask16:56
nekohayoI'm a bit unsure how to get started/what I should try to work on initially/how to synchronise with your development efforts16:56
rsalvetinekohayo: django is quite easy to learn 16:56
nekohayoI presume so... it would be unpythonic for them to make it significantly more complex than straight python16:57
zyga-afknekohayo, templates are not python16:58
zyga-afknekohayo, I'd recommend looking at our flavour of django first16:58
zyga-afknekohayo, then trying to write a toy app16:58
nekohayowait, there are "flavours" of that dialect? ;)16:58
zyga-afknekohayo, we're doing some things differently (lava-server for example)16:58
zyga-afknekohayo, no, we're just not a run-off-the-mill django app16:59
zyga-afkwe're developing our own local standards16:59
zyga-afklike lava-server extensions16:59
zyga-afkweb api16:59
zyga-afktable display api16:59
zyga-afkso that's unique to lava16:59
zyga-afkbut most of the code is just python with django16:59
davepigottplars: TImed it. It took 21 minutes, so I made the timeout 1800s in the code and it passed that point. :)17:00
ryanharkin \o/ yay!17:01
zyga-afknekohayo, if you really want your hands dirty we need to do some extra steps17:02
zyga-afknekohayo, it'd be nice to re-think what we want QA tracker to be17:02
zyga-afknekohayo, and write a blueprint17:02
nekohayoI know "straight" python, but you pretty much have to assume I'm starting from zero knowledge on LAVA and django :)17:03
zyga-afknekohayo, django will be bigger hit17:04
nekohayoI expect it might take me a significant amount of time to learn and get somewhat comfortable17:04
zyga-afknekohayo, start with the django tutorial17:04
nekohayosure17:04
zyga-afknekohayo, don't use ubuntu packages17:04
zyga-afknekohayo, see how virtualenv works if you're not accustomed to that17:04
zyga-afknekohayo, deploy postgresql17:04
zyga-afknekohayo, install rabbitmq 17:05
zyga-afknekohayo, our toolkit is python + django + postgresql via django ORM + rabbitmq (incoming) via celery 17:05
zyga-afknekohayo, you won't need all 17:05
zyga-afknekohayo, but you'll come across all 17:05
GrueMasterrsalveti: I'm being asked on #beagle when the sgx drivers for omap3 will pe ported to armhf.  Any news I can pass along?17:08
nekohayothe fun thing too is that I'm running Fedora17:08
zyga-afknekohayo, interesting17:08
nekohayoto test lava I have to run it in an ubuntu virtualbox instance17:08
zyga-afknekohayo, I had some fedora patches17:08
zyga-afknekohayo, no no17:08
zyga-afknekohayo, you can run on fedora17:08
zyga-afknekohayo, but you'll have problems with linaro images17:08
nekohayozyga-afk, well there's no such thing as a ppa for it no?17:08
zyga-afknekohayo, lava should be good17:08
zyga-afknekohayo, for lava?17:08
zyga-afknekohayo, no, we aborted that17:09
zyga-afknekohayo, I had lava running on F15 (or 16 cannot remember) last year17:09
rsalvetiGrueMaster: not from TI, but I'll try the hack to transform the current armel on to be compatible with armhf 17:09
zyga-afknekohayo, we need systemd scripts 17:09
zyga-afknekohayo, that's all17:09
rsalvetiGrueMaster: something I was planning to check this weekend17:09
zyga-afknekohayo, I don't run fedora daily, it's too annoying 17:09
GrueMasterrsalveti: Excellent, thanks.17:09
zyga-afknekohayo, but if you want that should be all fine17:09
zyga-afknekohayo, curious, what is your interest with lava?17:13
bhojsuihkulokki, Hi17:15
=== zyga-afk is now known as zyga
zyganekohayo, which fedora are you running now?17:17
bhojsuihkulokki,  i wanted to know the naming you  are planning to have fastmodel builds . would like to use the same in android .  17:18
=== dzin is now known as dzin_lunch
zygasuihkulokki, hi17:18
zygasuihkulokki, I'd like to sync with yuo17:18
alfrsalveti: you've got mail17:18
zygasuihkulokki, I'm responsible for implementing lava support for fast models17:19
rsalvetialf: let me check17:19
zygasuihkulokki, when would you have the time to talk (and timezone if you can)17:19
nekohayozyga, F1617:20
rsalvetialf: I think I can easily generate an ubuntu-desktop rootfs with latest packages included17:20
zyganekohayo, cool17:20
rsalvetialf: let me copy the packages to the overlay ppa17:21
nekohayolooking forward to flying a F22 or F3517:21
zyganekohayo, could you check coulple of things for me17:21
zyganekohayo, I need to know if all of our non-python deps are available17:21
zyganekohayo, I'm not fluent in fedora so you'd be my interpreter here17:21
nekohayosure17:21
zyganekohayo, postgresql, apache2, rabbitmq17:21
zyganekohayo, uwsgi17:21
nekohayoI used to run ubuntu, so I have a decent enough package name translation skill now17:21
alfrsalveti: we don't want the current packages, I need to update the linaro-gles2 branches first17:22
zyganekohayo, and apache module for uwsgi17:22
zyganekohayo, note: we build uwsgi from source17:22
zyganekohayo, we may switch to another app server but that's it for now17:22
rsalvetialf: oh, ok, that's the updated code you said 17:22
rsalvetialf: I'm fine with it 17:22
* ogra_ hugs alf 17:22
rsalvetialf: go for it :-)17:23
ogra_awesome progress !17:23
rsalvetialf: seems unity failed to build again17:23
alfrsalveti: If we are referring to the same failure, it's because unity depends on a newer compiz than the one we currently have. The updated one is OK.17:24
rsalvetialf: oh, ok then17:24
alfrsalveti: I will push in a minute. These will get built with the gcc version currently in the archive, right?17:25
rsalvetialf: yup17:26
nekohayozyga, "httpd postgresql-server rabbitmq-server"17:32
zyganekohayo, what versions do you have?17:32
nekohayothere is no uwsgi package17:32
zygapostgres and rabbit17:32
zyganekohayo, that's okay17:32
zyganekohayo, could you get the equivalent of python-dev and run pip install uwsgi in a venv?17:32
nekohayopostgresql-server 9.1.317:32
nekohayorabbitmq-server 2.6.117:33
nekohayothe equivalent of *-dev (from ubuntu/debian) is *-devel in fedora17:33
=== rsalveti changed the topic of #linaro to: Linaro 12.02: http://bit.ly/y7Z1v4 | Logs at: http://irclogs.linaro.org/ | Ask Linaro: http://ask.linaro.org/ | Porting Jam: http://bit.ly/arm-porting-queue | ARM Porting Jam today!
zyganekohayo, it should be all good 17:34
zyganekohayo, we're using upstream rabbitmq package17:34
zyganekohayo, as the one in precise is still older and the new one has nice features but that's not important to QA tracker17:34
nekohayoI would have been surprised if ubuntu had newer package versions than fedora's ;)17:35
zyganekohayo, it has updated unity I've heard ;-)17:35
zygaanyway17:35
zyganekohayo, I'm sure you'll be able to work fine17:35
zyganekohayo, what would you like the lava-qatracker to do?17:36
nekohayoI have to say that this bit is unclear to me for the time being :)17:36
zyganekohayo, what is the motivation then?17:36
anmarrsalveti: Hello Ricardo. Are you around?17:36
rsalvetianmar: yup, what can I do for you sr?17:36
anmarrsalveti: I need to know if and if not, then how can we get snowball HWPacks with the latest 3.3 integration branch17:37
anmarrsalveti: we can mumble if it is faster for you.17:37
nekohayoit seems to be a little bit hard to deploy at times, and there's stuff like hardcoded paths in the code17:37
rsalvetianmar: I just need the tree, branch and config you want to be used by our CI17:37
rsalvetianmar: then jcrigby can set it up at ci.linaro.org17:38
rsalvetiand the package will be published at the kernel PPA, already for oneiric and precise 17:38
anmarrsalveti: so we have this https://ci.linaro.org/jenkins/view/Linux%20Linaro%20Snowball%20builds/job/linux-igloo-kernel-integration-ubuntu-ux500/17:38
zyganekohayo, qatracker was never finished17:38
anmarrsalveti: in ci loop17:38
zyganekohayo, it was a quick demo done during a sprint17:38
rsalvetithen we can simply generate a hwpack from that17:38
nekohayoI saw that :)17:38
anmarrsalveti: ah.. perfect. 17:38
anmarrsalveti: there is a bug mpoirier is fixing to dislodge our CI loop.17:38
nekohayobut I was told that development on it would pick up again in the near future17:39
rsalvetianmar: https://ci.linaro.org/jenkins/view/Ubuntu%20CI%20PKG/17:39
anmarrsalveti: but once fixed, you can build the images. 17:39
rsalvetianmar: this is where we'll be adding your tree 17:39
zyganekohayo, I saw some interest but nobody from the people I work with daily is working on it now17:39
anmarrsalveti: perfect.17:39
rsalvetithat way we can do what we plan/decide at connect, that is to track and package your kernel automatically with ci17:39
zyganekohayo, it'd be nice to send a message to linaro-validation@lists.linaro.org17:39
zyganekohayo, and check if that is true17:39
zyganekohayo, I'd love to see that part finsihed17:39
zyganekohayo, although in a different shape17:40
jcrigbyrsalveti, did we ever use the 3.2 origen output?17:40
anmarrsalveti: that is exactly what I want :)17:40
zyganekohayo, I really don't like the current "look at bzr tree to find test cases" approach17:40
rsalvetianmar: I believe jcrigby will be enabling a bunch of new trees next week17:40
anmarrsalveti: cool.17:40
jcrigbyanmar, yes17:40
rsalvetijcrigby: I used it, but it's the same as the current one available at the samsung ppa17:40
anmarrsalveti: these are HWPacks or just kernel packages?17:40
rsalvetianmar: just the kernel packages, but having the hwpack is something quite easy to do17:41
rsalvetionce we have the kernel packages available, I'll set it up for you17:41
anmarrsalveti: can we get that as well please.17:41
rsalvetisure17:41
zyganekohayo, do you have a goal, are you interested in using that locally? just hacking on it? 17:41
rsalvetiI want to test and make sure it's working well before announce it back to you ;-)17:41
anmarrsalveti: perfect. I will make sure you get the information you need to make the .debs and packages.17:41
rsalvetiand that will include having the hwpack17:41
rsalvetianmar: great17:41
ramanaams_cs, hi17:41
anmarrsalveti: pleasure as always Ricardo. Thank you .17:42
rsalvetijcrigby: but as the samsung lt asked to track 3.3 now, we can simply drop the 3.2 and move to their 3.3 tree17:42
rsalvetianmar: np :-)17:42
ams_cshi17:42
ams_csramana: hi17:42
jcrigbyrsalveti, yes that is what I was asking, I would like to retire unused kernels17:42
rsalvetijcrigby: sounds good :-)17:42
rsalvetianmar: https://blueprints.launchpad.net/linaro-ubuntu/+spec/kernel-ci-fixes-for-12.0317:43
davepigottzyga: I set up an ssh config so I can connect to my cloud instance nice and neatly from home. Very cool. :)17:47
zygadavepigott, would you mind sharing? :)17:47
davepigottzyga: Not at all. One moment...17:47
davepigottzyga: https://pastebin.linaro.org/425/17:49
zygadavepigott, I wish we had something that allows you to connect to any instance we have17:53
zygaI have that for canonical private cloud17:53
davepigottzyga: I'm working on that. 17:53
kenwszyga: fabo: Hi, As discussed the other I've create a small shell script that automates building oe-core+meta-linaro: http://people.linaro.org/~kwerner/oe-core/script/build-oe-core-linaro.sh17:54
kenwsThe part that runs qemu and validates the image is missing (there are a few thing still unclear)17:54
zygakenws, cool17:55
zygakenws, let me have a look17:55
kenwsMaybe we can talk next week about how to proceed : )17:55
zygakenws, gladly17:55
zygakenws, wait17:55
zygakenws, is the interesting "product" / artefact anything that is in ${OUT_DIR}17:56
kenwsyep17:56
zygaok17:56
zygaI'll play with this 17:56
kenwsthis is where the kernel and roots for qemu will be placed17:56
zygaand talk to you next week17:56
zygathis is very good though17:57
kenwscool, thanks! 17:57
kenwshave nice weekend17:57
zygathanks for writing this17:57
zygalikewise17:57
kenwss/the other/the other day/17:57
alfrsalveti: Branch updating done. We will probably need to schedule a unity build when the new compiz* daily builds are done.17:57
rsalvetialf: yeah, cool17:58
rsalvetialf: thanks17:58
alfrsalveti: Also we will probably get a broken build of compiz-plugins-main at first because it depends on the new compiz-core, so will probably need to reschedule that as well.17:59
alfrsalveti: Just in case you are around when all this happens :)17:59
rsalvetialf: yeah, will look over it during the weekend, as it taks quite a while to have all the packages built 18:00
nekohayozyga, generally speaking I don't have a very clear goal other than "improve the thing" yet. If you folks have a list of "easy starter tasks/bugs" (à la "gnome-love") that would help. In the meantime I have to get familiar with django.18:01
avinashhmgood morning, after flashing the linaro snapshots to the sd card, is there an option to update the u-boot ?18:09
avinashhmthe kernel and initrd are in /media/boot .. could update that .. but don't find the u-boot .. hence the question .. any poiters will be helpful .. thanks18:10
zyganekohayo, we don't have that yet, although if you want to get involved we could easily show you a load of easy tasks18:12
zyganekohayo, not all are django though18:12
zyganekohayo, we have a big non-web part too18:12
nekohayooh?18:12
zyganekohayo, yeah18:13
zyganekohayo, lava-tool* is all pure python18:13
zyganekohayo, lava-dispatcher18:13
zyganekohayo, lots of smaller libraries18:13
zyganekohayo, one thing that will probably come out of this18:13
zyganekohayo, that you may not have thought of before18:14
zyganekohayo, is review of our test result spec/format18:14
zyganekohayo, as I believe manual tests change that 18:14
zyganekohayo, one of the things we've identified as needed are bug references18:14
zyga(someone tests, finds a bug and notes that)18:14
zygaand free form comments (we have something but it's not clearly for that)18:15
zygathat's separate from supporting comments as a feature in lava (any user commenting on anything in the system)18:15
zyganekohayo, per-test-result attachments are also probably needed18:15
zyganekohayo, right now attachments are only associated with the whole test run as that's what we needed for "I run a test and here's the output"18:16
zyganekohayo, but for manual tests people have asked for ability to add screenshots, crash dumps, etc18:16
zyganekohayo, you should get familiar with our test result spec: linaro-dashboard-bundle.rtfd.org18:16
zyganekohayo, that document is not perfect but should give you plenty of context18:17
zyganekohayo, it does not cover everything, look at the raw schema to get that, it's somewhat hard to read on the first go18:17
fabokenws: I'll setup a job on ci.linaro.org to run your script. there's mostly nothing to do except get your script and running it.18:35
davepigottryanharkin: ping18:37
ryanharkindavepigott: pong18:38
davepigottryanharkin: DId I ever comment on how slow the vexpress is? ;)18:38
ryanharkin:O18:38
davepigottryanharkin: I started a lava test at about 4:30. It's still running. But at least it's running. :)18:39
zygadavepigott, wow, what are you running?18:39
zygadavepigott, is that te FPGA variant or the one with silicon?18:39
davepigottzyga: This is just deployment! :)18:39
ryanharkindavepigott: seems quite quick to me... I have to use these boards all day, you know ;)18:40
davepigottryanharkin: :D18:40
zygadavepigott, hmm18:40
zygadavepigott, what is it doing now?18:40
zygadavepigott, are you home?18:40
ryanharkini run my builds on them and everything, much better than my old 486 box.  Just about ;)18:41
davepigottzyga: It's installing a load of packages under chroot18:41
davepigottzyga: Yes. Home18:41
zygadavepigott, I think io is slow18:41
zygadavepigott, does it have USB?18:41
davepigottzyga: According to ryanharkin USB is worse than sd18:41
ryanharkinzyga: don't say "USB" to it!!! it's crash!18:41
ryanharkinit'll crash18:42
zygadavepigott, is it USB 1.1 ?18:42
davepigottryanharkin: ^18:42
zygadavepigott, SD is so crappy I run _everything_ off USB18:42
davepigottzyga: It's not about the USB spec. It's the support implementation on board18:43
zygaah18:43
ryanharkinzyga: don't know actually, it's not the USB protocol that's slowing it down, ... wot dave said18:43
ryanharkinzyga: no DMA18:43
zygaryanharkin, USB has no DMA18:43
davepigottryanharkin: You stole the acronym right out of my mouth18:44
zygaryanharkin, AFAIK18:44
zygaryanharkin, does SD?18:44
davepigottzyga: ryanharkin is talking about sd18:44
ryanharkinzyga: the board has no DMA controller at all18:44
zygaryanharkin, from what I know USB has no support for any kind of DMA anywhere but I could be mis-interpreting stuff18:44
zyga(and perhaps on x86 something else helps with DMA for DMA-less USB)18:45
ryanharkinzyga: i think you are.  USB devices usually use DMA to transfer large amounts of data from the controller into RAM18:45
ryanharkinzyga, but that's implementation specific18:46
zygaryanharkin, maybe, from what I remember the biggest difference between USB and firewire was that (apart from hub architecture) firewire had DMA suport18:46
zyga(and stream reservation but AFAIR USB got those too)18:46
zygaand the CPU has to poke the controller all the time18:47
zygaperhaps nowadays the controller hides some of that crap 18:47
zygaand that's why we have [AOE]HCI18:47
zygaancient, obsolete and enchanced ;-)18:47
zyga(just making those up though)18:47
ryanharkinzyga, I thought that the firewire protocol was handled on chip, the usb protocol by the CPU, so the CPU has to inspect all the USB packets.  I didn't think it was a DMA issue18:48
zygaI'm not an expert on this18:48
ryanharkinzyga, but I'm no USB expert :O so I could be wrong18:48
ryanharkinsnap!18:48
zygaheh18:48
zygaI guess I'm wrong18:48
zygahttp://www.mjmwired.net/kernel/Documentation/usb/dma.txt18:48
zygasorry for the random mirror18:48
=== dzin_lunch is now known as dzin
davepigottryanharkin: zyga: What I'm concerned about here is that vexpress is going to take 3+ hours to run a lava health check test, let alone anything that actually does something18:49
zygadavepigott, yeah18:50
davepigottzyga: ryanharkin Actually, worse, because we deploy both android and ubuntu in health check18:50
zygaI think health checks are somewhat bogus ... somewhat18:50
zygathree things can break:18:50
ryanharkindavepigott: i feel your pain, but I have to go and drink myself silly.  sorry to run off on you!18:50
zyga1) SD cards go bad18:50
zyga2) master images may be corrupted, we can fix that with software18:50
zyga3) cabling/configuration can break/be setup incorrectly 18:51
davepigottryanharkin: It's Friday. I don't blame you. Starting soon myself. Have to go and get the Indian in a minute18:51
zygaI think our current pandora box of problems is that masters "degrade"18:51
zygaand that timeouts we have are even more bogus than the "I tell the future" advert on the tv shop channel with premium numbers to call18:52
zygaso18:52
zygagiven that perspective18:52
zyga1) we need to test that18:52
zygaand that's not going away ever18:52
zyga2) we will eventually never ever need to test that18:52
zyga3) that's something I'd like to know more about18:52
zygadavepigott, I think we should pick the right set of tests18:53
davepigottzyga: We'll have to cherry pick very carefully18:53
zygadavepigott, still I cannot shake the feeling that even if we ran everything all the time to "check" jobs18:53
zygadavepigott, we'd still hit timeouts all the time18:53
zygadavepigott, and break our master images faster18:53
zygadavepigott, and not be a bit smarter18:54
zygadavepigott, perhaps reprogramming the SD via a very special test would be a good alternative?18:54
davepigottzyga: Yeah. I know what you mean18:54
zygaor perhaps boards really die18:54
davepigottzyga: If we get these dual ported sd interfaces it will make a huge difference18:54
zyga(like often die)18:54
zygadavepigott, initially18:54
zygadavepigott, but I think that's still a crack solution18:54
zygadavepigott, they will allow us to test bootloaders18:55
zygadavepigott, but will actually lessen the incentive to fix our master images18:55
zygadavepigott, by pretending we can recover them 18:55
davepigottzyga: Back in 30 minutes. 18:55
zyga:)18:55
zygaah18:58
zygakenws, OE, finally something to keep my 4 cores busy18:59
shadeslayerwookey: ping19:25
shadeslayerwookey: would it be possible for you to get a new multistrap package into the ubuntu archives? The current one is *old*19:26
wookeyshadeslayer: erm, yes I guess so. 19:34
wookeythere was some difficulty, but I foget what it was19:35
wookeyI was trying to get rid of the diff so it would update automatically, and ran into somethig that made that awkward IIRC (upstream mostly :-)19:35
shadeslayerwookey: I'm looking at the diff right now19:36
shadeslayerwookey: there's not alot of diff in the packaging, most of it is ... well, we can drop all of it I think19:36
wookeyyeah it was mostl pdebuild-cross support I thnk , and probably no-one cares much about that19:37
shadeslayerwell, pdebuild-cross isn't installable right now19:37
shadeslayerwhich is why I noticed the issue19:37
wookeyOK. I can;t actually do ubuntu uploads anyway so I can look at the diff and ask someone nicely, but it sounds like you're already doing that?19:38
shadeslayeryeah19:39
shadeslayerI'm test building the package right now ;)19:39
wookeyIf you want to send me something to say 'yep looks good to me' that's fine. Or if you just upload the current debian version with a couple of configs tweaked to have ubuntu archive paths, that's fine too.19:40
shadeslayerwookey: http://paste.kde.org/43623219:40
shadeslayerwookey: I don't see anything that needs doing really, there are no ubuntu specific patches that I see19:41
shadeslayerthere is stuff like : -examples/grip-squeeze.conf ./usr/share/multistrap/19:42
shadeslayerugh19:44
shadeslayerhttps://launchpadlibrarian.net/56394851/multistrap-2.1.6ubuntu3.patch isn't upstream19:44
wookeyyou need the ubuntu configs19:44
wookey-cross/ubuntu/*.conf ./usr/share/multistrap/ubuntu/19:44
wookeybut yes, that's about it19:45
* wookey has to go home - more later19:46
shadeslayerwookey: this package is seriously messed up :(19:46
shadeslayerwookey: just one last thing19:46
shadeslayerwookey: all these changes should really be in a seprate patch19:46
wookeyOK. I was entirely clueless about ubuntu dev when I did that.19:47
shadeslayer:D19:48
shadeslayerI've added it to my TODO list, will pursue on Monday19:48
wookeyso basically we should take upstream and just have a debian/patches patch for all the changes (which is basically archive/packagename changes config)19:49
* wookey really is gone now 19:49
shadeslayeryep19:50
=== cody-somerville_ is now known as cody-somerville
=== zyga is now known as zyga-weekend
plarsrsalveti: ping?22:01
plarssuihkulokki: ping?22:01
=== e-ndy|afk is now known as e-ndy
=== e-ndy is now known as e-ndy|afk
ChiThuplars, ping22:27
=== salgado is now known as salgado-afk
janrinzehi, i have tried just about anything for a couple of weeks now but there seems to be no way of getting hdmi to work on the origen board with the linaro kernels.. anyone here who feels let down like me about this result?23:37

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!