pfefferzk jstultz_vm04:14
pfefferzhey pabhishek05:09
pfefferzhey jackh05:09
jackhhey, pfefferz05:09
pfefferzhey jackh05:09
pabhishekpfefferz, hi05:09
pfefferzso it looks like the insitu testing of the kernel big>LITTLE stuff won't be needed05:10
pfefferzat this point05:10
pfefferzis Tian around?05:10
pfefferzjackh, they want to get the system level tests done in the test plan05:10
jackhyes, he is starting irc05:10
pfefferzand the kernel guys are going to do the other tests05:10
jackhk, so where's the new plan05:11
pfefferzhey tianhong-wang05:12
pfefferzjackh, well I need to get something together05:12
pfefferztianhong-wang, try to stay on irc as much as you can05:12
pfefferzjackh, looks like Annamalai123 may have solved the  Origen boot problem05:13
jackhpfefferz, we got the lcd display with hdmi05:13
jackhwhich boot problem?05:13
pfefferzjackh, great!05:13
pfefferzjackh, the Samsung crashing problem I emailed you about yesterday05:13
pfefferzAnnamalai123, have you pushed a change to Gerrit to fix that?05:14
jackhi think i located the problem...it's about the keys05:14
jackhi mean the suspend one05:14
Annamalai123pfefferz, the fix is in samsung kernel, and vishal told me that samsung kernel is not a part of gerrit05:14
pfefferzjackh, the keys?05:14
pfefferzAnnamalai123, right05:14
Annamalai123so i had sent the patch to kernel maintainer05:15
jackhya, i think the keys are not mapped correctly05:15
pfefferzAnnamalai123, okay cool05:15
jackhstill locating it now05:15
pfefferzjackh, k jackh05:15
pfefferztianhong-wang, hows that Android test stuff going?05:15
tianhong-wangIs that monkey?05:17
pfefferztianhong-wang, the monkey runner tests05:17
pfefferzPlay a local audio clip against software decode05:17
pfefferzPlay a local audio clip against hardware decode05:17
pfefferztianhong-wang, you should subscribe to this: https://blueprints.launchpad.net/linaro-android/+spec/linaro-android-auto-use-test05:18
tianhong-wangis Monkeyrunner?05:18
pfefferzjackh, tianhong-wang here's the launchpad page with our 12.03 plans05:18
pfefferztianhong-wang, you'll need to build monkeyrunner scripts05:19
tianhong-wangI still need a hdmi cable, which is on way05:19
pfefferzto execute all those scenarios05:19
pfefferzactually you should be able to start coding the tests now05:19
=== liuyq is now known as liuyq-afk
pfefferzyou can call them from the command line05:20
jackhthey can cover all the cases?05:24
pfefferzwell we have to do them05:25
jackhya, it looks very important for automation05:26
pfefferzI was actually think about having you and tianhong-wang work on them05:26
pfefferzplars, up?05:27
plarspfefferz: yep05:27
pfefferzplars, did I miss something with the test plan?05:28
pfefferzplars, all the C level tests are called out at Ubuntu05:28
pfefferzplars, all the C level tests are called out just for Ubuntu05:28
pfefferzplars, so does Amit still need to put them together?05:28
plarspfefferz: it was agreed on the call we had yesterday that we would run the android kernel on ubuntu rootfs to make this testing easier05:28
plarsif we're talking about the same thing05:29
plarsI'm not sure what you mean by "C level"05:29
pfefferzbefore all the test payloads05:29
pfefferzwe we're going to run the payloads like stress05:29
pfefferzand run them against Ubuntu and Android05:29
pfefferzand AMit is developing them05:30
plarspfefferz: can you give me an example?05:30
pfefferzbut the test plan just says Ubuntu05:30
pfefferzYou have an Ubuntu tab05:30
pfefferza Android kernel + Ubuntu Rootfs05:30
pfefferzand Android05:30
pfefferzbut we're meant to run all the tests in all the tabs on Android05:31
plarsok, that's not what we discussed05:31
pfefferzI'm confused05:31
plarsbut if the tests exist, I'm happy for them to be executed there as well05:31
pfefferzsure, but it should be in the doc05:31
pfefferzbecause we're building them now05:31
plarsthe point of running android kernel + ubuntu rootfs was because there was this discussion about not wanting to build new tests from scratch, and tests that were difficult to port to android and stuff05:32
pfefferzplars, yeah05:32
jackhall the utils are complied in android?05:32
pfefferzplars, but we were going to do it05:32
plarsso rsalveti brought up the idea of just producing an image with the android kernel + ubuntu rootfs to run those tests05:32
pfefferzplars, because people wanted everything tested on Android05:32
jackhif so, i think it's not complex to do them05:32
plarspfefferz: that's not what I understood from the call05:32
pfefferzall that we said in the call05:32
plarspfefferz: so it's easy enough to add then05:33
pfefferzwas that ricardo was going to use and Androidized kernel05:33
plarspfefferz: what all is he writing?05:33
pfefferzbut the test loads are all going to run on Android05:33
pfefferzplars, ?05:33
pfefferzwe already have stress in android05:33
pfefferzand libjpeg turbo05:33
pfefferzthe only thing to put in is the coherancy pakage05:33
plarspfefferz: the libjpegturbo test doesn't exist yet05:33
plarssomeone needs to write it05:34
pfefferzplars, actually it does05:34
pfefferztgall_foo's got it05:34
plarsstressapptest is the coherancy test we are using on ubuntu, unless someone comes up with a better one05:34
jackhi guess they should be compled in external directory05:34
plarspfefferz: no, we said someone needed to write it05:34
pfefferzright we're going to use stressapptest on Android05:34
plarsit's not the 2 second test that just converts a jpeg05:34
pfefferzplars, some one needs to right the run this 10 times05:34
plarsthat's not going to be useful for running while switching05:34
pfefferzbut the underlying test is written05:34
plarsI got the impression it was going to be more than that05:35
plarsstress was out a long time ago05:35
plarswe aren't using that now05:35
plarsit's no longer in the test plan for a while now05:35
pfefferzI'm confused05:35
plarsmemtester for mem stress05:35
plarsand iozone for disk io stress05:35
pfefferzokay, I'm not sure what we're doing then05:35
plarspfefferz: all of this was agreed in the call05:36
plarspfefferz: you were there! :)05:36
plarsI remember your loud phone :)05:36
pfefferzthe test plan I saw on tuesday had it in it05:36
plarsand then we had this review meeting where we discussed all this on tuesday05:36
pfefferzand everything looked okay05:37
pfefferzbut I went to go look for the test loads05:37
bhojbero, Hi05:37
plarspfefferz: that's when all of this that we're talking about now happened05:37
pfefferzto send my how are we going to bundle up the test05:37
pfefferzand everything is gone05:37
plarspfefferz: it's not gone, it changed forms, as we discussed to a case-by-case with clear pass/fail criteria05:38
pfefferzI'm really confused05:38
plarsyou requested that even05:38
plarsso am I now05:38
plarsbecause you were the one asking for a lot of this05:38
pfefferzbut I didn't ask for things to just run on Ubuntu05:38
pfefferzso that's confusion 105:38
pfefferzand I didn't ask to ditch stress05:38
pfefferzthat's confusion 205:39
rsalvetiwhat we agreed is that the normal tests that already exists for ubuntu would be run at ubuntu05:39
rsalvetiwithout porting it to android05:39
plarsno, rsalveti suggested that because there was this discussion about not wanting to write new tests from scratch, but a lot of the tests were ubuntu and not on android for low-level stuff05:39
rsalvetibecause we're already going to run the android kernel on it05:39
pfefferzrsalveti, actually I don't remember agreeing to that at all05:39
pfefferzmaybe I was out of the room05:39
jackhif you want to run the audio tests, i guess the alsa-utils should be compiled under external for android05:39
plarspfefferz: if you have extra tests that you want run on android, I think we can certainly get them in05:39
rsalvetiand for android, instead of porting the test cases, you'd work on specific test cases for android05:39
rsalvetipfefferz: you were there, you said it was ok for you05:39
rsalvetiI even wrote it down at the notes05:39
plarsthere's nothing that says this plan has to be carved in stone and never adapted as we find/create better tests!05:39
pfefferzwhat I said was05:39
pfefferzwe'd get it to work in Ubuntu05:40
pfefferzthen we'd prot it to Android05:40
rsalvetithe thing is, don't waste time porting test cases to android!05:40
pfefferzthen we'd port it to Android05:40
pfefferzbut we have to05:40
rsalvetibecause once we run it on ubuntu with the android kernel, it's all the same05:40
plarsyeah, it's the same kernel05:40
rsalvetiand you should focus on the android specific test cases only05:40
pfefferzeverything has to run on the Android platform05:40
rsalvetibecause we're testing *the kernel*05:40
rsalvetiwhatever, that was not what we agreed05:40
plarswe're not testing android05:40
pfefferzall the test cases that we deliver have to run on ANdroid05:40
rsalvetiyou want to port? go for it, but for me it's just wating time and resources05:40
pfefferzI have the guys to port05:41
rsalvetiwhy not start focusing on writing the test cases for android?05:41
rsalvetiinstead of porting them?05:41
pfefferzwe're doing that in parallel05:41
jackhwaste time??05:41
plarspfefferz: I don't think we have to beat this to death... it's what we discussed in the call, but if you want to add tests, great05:41
pfefferzI have 3 guys working on this05:41
rsalvetiif you have infinity resources, I'm ok05:41
rsalvetibut it's not what we agreed at the call05:41
pfefferzI'm just very confused05:41
pfefferzwhat I thought we agreed on was05:42
jackhsalveti only want them on ubuntu?05:42
pfefferzwe'd get the test working in Ubuntu05:42
pfefferzthen port them to Android05:42
rsalvetijackh: waste time because once we got it running on ubuntu, with the android kernel, there's no sense on porting the test cases for android05:42
pfefferzrsalveti, that's actually not true05:42
rsalvetibecause it's already the android kernel05:42
rsalvetiand that's the thing we're testing05:42
rsalvetiI'm off05:42
jackhagree, if not work means some android errors, ya?05:42
pfefferzlet me send a mail05:43
pfefferzI'm confused05:43
jackhi guess he means if the tests are passed on ubuntu, then means no hardware error05:43
rsalveticheck https://docs.google.com/a/linaro.org/document/d/1_tx0bjmHPppLy0pXG5JPLrJbwSxyV0iqzENpTVD4XTs/edit05:44
jackhit's maybe no use to do them again on android05:44
rsalvetithat's exactly my point05:44
rsalvetiif we already have ltp working on an ubuntu rootfs with the android kernel05:44
rsalvetiwhy porting ltp to android?05:44
rsalvetimakes no sense05:44
pfefferzactually the action was for you to review it with me05:45
rsalvetireview the document, but you agreed on the call05:45
rsalvetibecause the document was still showing the duplicated test cases05:45
rsalvetilike running foobar on ubuntu; porting foobar on android05:45
pfefferzrsalveti, look, I'm not trying to argue, I'm trying to explain that I'm confused05:45
pfefferzbecause my preception of the call is not what you're saying05:46
pfefferzwhat my preception is05:46
pfefferzis that I said we should port the stuff to Ubuntu05:46
pfefferzthen port it to Android05:46
pfefferzso that we can run the entire test set in Android05:46
pfefferzsince we're going to ship the Android platofmr to people05:47
pfefferzand they're going to want to run all the tests05:47
pfefferzlike George is asking for05:47
pfefferzso they all have to run on ANdroid05:47
rsalvetiwe don't need to port to ubuntu05:47
rsalvetithey are already working and enabled at ubuntu05:48
rsalvetiwe just need to run them05:48
pfefferzthen we just need to pull them into Androids05:48
pfefferzand port them05:48
rsalvetiI don't see why05:48
jackhif for the rest of the world to test their hardware, i think the porting to android is essential05:48
rsalvetiif you want to do it, go ahead05:48
pfefferzso can you send me a list of things we have to port?05:48
rsalvetibut we don't want to tesst the hardware05:48
pfefferzso all the test cases can rtun on Android05:48
rsalvetiwe want to test the big.LITTLE implementation at the kernel side05:48
rsalvetiwill do that later, but they are already at the test plan iirc05:49
rsalvetiplars should also have the list05:49
plarsI think we are seriously bikeshedding this test plan to death05:49
pfefferzplars, rsalveti is the list of things we have to pull into an Android build05:49
plarswe'll be done with big.LITTLE before the test plan is agreed to at this rate05:49
pfefferzplars, and rsalveti05:50
plarswe already know there are going to be tests added05:50
pfefferzas far as I understqand things05:50
pfefferzwe have to deliver all the tests we want to run05:50
plarsit was mentioned before that grant wanted to look at adding some ltp tests, but they still need to find some ltp tests that are useful to pull in for testing here if that's going to happen05:50
pfefferzin an Android build05:50
plarsfor example05:50
plarsso we'll be adding that at some point I feel quite sure05:50
plarsalso, there are tests that need to be written still05:51
pfefferzso we'll need to add  memtester05:51
plarsso this is going to undergo revisions05:51
pfefferzand the other packages to Android05:51
plarsand full android stack testing wasn't slated to happen until the end05:51
pfefferzotherwise the tests won't run05:51
rsalvetiwe want to test 2 things05:51
rsalveti1 - kernel05:51
rsalveti2 - android userspace interaction with the kernel05:51
pfefferzlook, I'm just going to call a stand up meeting tomorrow05:51
rsalvetitesting the kernel can be done with the ubuntu rootfs, because most of the tests are available there already05:51
pfefferzand get some clarity on this05:52
plarspfefferz: the big thing I'm concerned about on your end, is that the tests you put forth are pretty big05:52
rsalvetiand for 2 it'd be something you'd be working since start05:52
plarsand as you noted already, fast models is slow05:52
plarsso I expect one of two things will need to happen there05:52
pfefferzright now, the plan I thought I was working on05:52
rsalvetithen you say you can run on tiny android, but that's not difference from ubuntu05:52
pfefferzis gone05:52
plarseither 1. we find a way to make them run in a reasonable amount of time05:52
pfefferzand the plan I need to work on05:52
pfefferzis up in the air05:52
plars2. that list will have to be made more realistic05:52
pfefferzso I'm going to call a meeting so we can hash this out05:52
plarsso I think there's plenty to work on for android already05:52
plarswe did already05:53
pfefferzobviously not05:53
rsalveticheck the meeting notes05:53
pfefferzotherwise I wouldn't have these questions05:53
rsalvetithat's exactly what we discussed there05:53
pfefferzI need to figure out what George and Rooger want05:53
rsalveti# Action Ricardo based on comment on section 7 - Dependencies - This will be simplified as we'll be using the Android kernel with the Ubuntu images as well, so we don't need to port the generic test cases to Android anymore.05:54
rsalvetiwe don't need to have another call05:54
rsalvetiwe agreed on that already, but if you still want to do it05:54
rsalvetido it05:54
rsalvetiit's just that it'll just duplicate things05:54
plarsyeah, I don't think we have to beat this thing to death05:54
plarsnobody is saying you *can't* do it zach05:54
rsalvetiif you have resources, and want to do it, go for it05:55
rsalvetiit's just that *I* don't think it's necessary05:55
rsalvetifor me it seems there are more important things to test05:55
rsalvetiwith the real android stack05:55
pfefferzhere's the point I'm confused on05:56
pfefferzDo all the tests that we need to run to test big.LITTLE need to run in Android?05:56
pfefferzI believe that Roger and George will say yes05:56
pfefferzso I'm going to ask them05:56
pfefferzif the answer is no05:57
rsalvetiif you say need to run on android, of course they will say yes05:57
rsalvetieven without understanding it05:57
pfefferzthen I'm fine05:57
pfefferzif the answer is yes I'm fine05:57
pfefferzI just need to know05:57
pfefferzbecause if the answer is yes05:57
rsalvetiyou should ask if you want to duplicate work porting the test cases already available at ubuntu to android05:57
pfefferzthen we need to port stuff to Android05:57
rsalvetibecause the kernel is the *same*05:58
rsalvetieven with same configs05:58
rsalvetitesting with ubuntu, android or simply with busybox will be the same thing05:58
rsalvetiI don't really care if you want to port the test cases05:59
rsalvetiit's just that it doesn't seems to make sense05:59
pfefferzrsalveti, the point is05:59
rsalvetiso, do whatever you think it's the best for you05:59
pfefferzwe're not going to give people an Ubuntu buiild to test test cases 12305:59
rsalvetiwe'll be running the same tests with the android kernel on ubuntu as well05:59
pfefferzand an Android build to test cases 34505:59
pfefferzwe're going to give them an Android build to test cases 1234506:00
rsalvetiwe're going to do the testing and validation06:00
plarspfefferz: I don't think you need george's permission to add tests, if you want to add them, add them06:00
pfefferzso test case 1234506:00
pfefferzhas to run in Android06:00
rsalvetias long as you have the resources to do it, go for it06:00
pfefferzplars, no I want to be clear on this06:00
plarsactually no06:00
plarswe're going to give them a patch set06:00
plarsand some test cases06:00
pfefferzbecuase the test plan says otherwise06:00
plarsthat was the BOM discussion we had today06:00
pfefferzand I said I would send out how we're going to deliver the test cases06:01
pfefferzwhich is what I'm doing now06:01
pfefferzwhich is why I'm asking questions06:01
pfefferzbecause in Android06:01
plarsI think that was just because you were opposing what grant was suggesting06:01
plarsso they put it on you to put forth an alternate proposal06:01
pfefferzlook I don't know about that06:01
pfefferzbut I think I know what George is asking06:02
pfefferzso I'm going to get clarification06:02
pfefferzso that I don't get surprised06:02
pfefferzthis is not a "should we" discussion06:03
pfefferzits a "what does the customer want" discussion06:03
pfefferzand if there's double work to do06:03
pfefferzthen so be it06:03
pfefferzemail sent06:04
pfefferzif they say yes06:04
pfefferzwe have some work to do06:04
pfefferzif they say no just deleiver them as seperate things06:04
pfefferzthen we don't have work to do06:04
pfefferzeither way we know06:05
pfefferzright now we don't know06:05
pfefferzjackh and tianhong-wang take a look at monkey runner06:05
pfefferzand the test scenarios I've laied out06:06
pfefferzand the test scenarios I've laid out06:06
pfefferzjackh, instead of the kernel level tests06:06
jackhtesting them on all the boards?06:06
pfefferzthe tests will apply to all the boards06:07
jackhk, we haven't imx06:07
pfefferzjackh, I know06:08
pfefferzjackh, can you guys get a Panda?06:08
pfefferzthey're available now06:08
jackhwe will take snowball and origen to start with06:08
jackhi have panda here06:09
pfefferzthats fine06:09
pfefferzcool you can use that too06:09
pfefferzcan you go out and get the HDMI cable for Tian06:09
pfefferzI don't think we should be blocked on a cable06:09
jackhtian ordered two06:09
jackhwhich will reach here at monday06:09
pfefferzplease see what you can do with monkey runner ASAP06:11
jackhpriority 1?06:11
pfefferzhey kejun-zhou06:22
pfefferzhow's your stuff going?06:22
pfefferzkejun-zhou, looking at the gerrit review now06:22
kejun-zhouI fixed the sd card gallery issue06:22
kejun-zhouThe patch is so simple/06:23
kejun-zhoubut it really hard to find the root cause.06:23
pfefferzgood job!06:23
kejun-zhouI checked from java to sd card driver06:23
pfefferzwhere did you finally find it?06:24
pfefferzI mean what led you to look at vold.fstab?06:24
kejun-zhouAs I said in bp this week06:24
kejun-zhouI check the gallery software arch06:24
kejun-zhouThe setting-> storage gives me a inspire06:25
pfefferzdid it cause a print to go off that you had put in?06:26
plarssomeone has been monkeying with the dates in the dependencies and not modifying the change log06:26
jackhhey, guys, i dont think people gonna tested the gpio-key.kcm on all the boards...06:28
pfefferzplars, ???06:28
pfefferzjackh, could you post a little more info?06:28
plarsI did not put the dates like this... mounir alerted me that one of the dates in the dependencies section was very very wrong06:28
pfefferzhi jkridner06:28
plarsand then I see things like the android image being needed by 3/1806:29
plarswhich is way way earlier than we discussed06:29
jackh1, gpio-keys.kcm is not added into device.mk to make it copied to the board06:29
plarsnote to self, future test plans need to be in git or bzr06:29
plarsI gotta go figure out the dates all over again now06:29
jackh2, if you copy them to the board, then there will a parse error06:29
pfefferzplars, have you ever used google docs history06:30
pfefferzjackh, on Origen06:30
plarspfefferz: no, I was just looking at that, but it looks like you have to tag revisions for it to work06:30
pfefferzplars, actually no06:30
jackhpfefferz, i think we always use the wrong key files...06:30
plarspfefferz: it came up with nothing for me06:30
plarswhen I tried it06:30
pfefferzjackh, could be would you do a little investigation06:30
plarscrap, now gdocs isn't responding06:30
pfefferzand file a bug06:30
pfefferzFile > See Revision history06:31
jackhi have to figure that out, that's essential to wake-up06:31
pfefferzand you can select fine grained06:31
pfefferzjackh, awesome!06:31
jackhi free crazy how it can be like this?? no use ever use the right key files...06:32
jackhmy gosh...06:32
pfefferzyou're a hero jackh06:32
jackhohhhhhhhh.....eight Hs this time06:33
pfefferzgwell great fix kejun-zhou06:33
jackhkejun-zhou knows about this?06:34
jackhi m not sure got what u said...06:35
jackh Expected keyword, got '[type=QWERTY]', anything know why this??06:49
pfefferzkejun-zhou, https://blueprints.launchpad.net/linaro-android/+spec/linaro-android-cpuidle-snowball07:00
pfefferzmorning ryanharkin07:00
ryanharkinpfefferz: morning!  you still awake?!07:00
pfefferzthanks for testing the vexpress07:01
pfefferzryanharkin, I'm a 24 hour party person07:01
jackhi got some progress! the keys are wrong...07:02
ryanharkinpfefferz: i didn't get through it all, still more to do07:03
pfefferzjackh, cool!07:04
pfefferzryanharkin, you don't need to test everything for the boot test07:04
pfefferzryanharkin, just does it boot07:05
pfefferzconsole work07:05
pfefferzjust a few tests07:05
pfefferzwhat you did is great07:05
pfefferzfor the RC07:05
ryanharkinpfefferz: well, it does all that :)07:05
pfefferzpabhishek, all the RC builds looking good?07:05
pfefferzryanharkin, :)07:05
pfefferzryanharkin, now for the final build of the month07:05
pfefferzwe test everything07:05
pfefferzoff to bed for a bit07:07
ryanharkinwhen I was testing the email app, it wouldn't let me setup my gmail account, but the app seemed to work07:07
ryanharkinis that normal?!07:07
pfefferznot really07:07
pfefferzit should let you07:07
pfefferzoff for a bit07:07
pfefferzsee people soon ish07:07
ryanharkini might have been doing something daft, but i figured "default" settings would have worked for gmail07:07
ryanharkinyeah, go get your beauty sleep07:08
kishore_Hello everyone... I am getting some trouble on building kernel ... I got the source pack from http://www.linaro.org/linaro-blog/2011/09/15/switching-out-a-pre-built-android-images-kernel/  link ... But after grabing the tool chain, I am not able to proceed further... Can Anyone help me on this issue... ??08:07
jackhwhat's your detailed problem?08:09
kishore_jackh: In the link I gave in previous line I am able to do the things till grabing the toolchain , but after that "find the deconfig  step is failing with 404 error....08:11
jackhorigen? then make ARCH=arm android_origen_defconfig08:12
kishore_jackh: Yes origen only... i tried the command which you told me but its gets failed with following error... .08:13
kishore_/bin/bash: prebuilt/linux-x86/toolchain/arm-linux-androideabi-4.4.x/bin/arm-linux-androideabi-gcc: No such file or directory08:13
kishore_/bin/bash: prebuilt/linux-x86/toolchain/arm-linux-androideabi-4.4.x/bin/arm-linux-androideabi-gcc: No such file or directory08:13
kishore_/bin/bash: prebuilt/linux-x86/toolchain/arm-linux-androideabi-4.4.x/bin/arm-linux-androideabi-gcc: No such file or directory08:13
kishore_/bin/bash: bison: command not found08:13
kishore_build/core/config.mk:268: *** Error: could not find jdk tools.jar, please install JDK6, which you can download from java.sun.com.  Stop.08:13
jackhif u want to build the whole android, u need to set the toolchain prefix08:14
jackhfor bison, do a sudo apt-get install08:14
kishore_How to set the toolchain prifix??08:15
jackhsudo make -j16 HOST_CC=gcc-4.5 HOST_CXX=g++-4.5 HOST_CPP=cpp-4.5 TARGET_PRODUCT=origen TARGET_SIMULATOR=false TARGET_TOOLS_PREFIX=/media/one/origen_staging/android-toolchain-eabi/bin/arm-linux-androideabi- systemtarball userdatatarball08:16
kishore_jackh: One more thing jackh, every time i am installing any package I am getting some error like.08:17
kishore_invoke-rc.d: initscript rabbitmq-server, action "start" failed.08:17
kishore_dpkg: error processing rabbitmq-server (--configure):08:17
kishore_ subprocess installed post-installation script returned error exit status 108:17
kishore_Setting up m4 (1.4.16-1) ...08:17
kishore_Setting up bison (1:2.4.1.dfsg-3) ...08:17
kishore_update-alternatives: using /usr/bin/bison.yacc to provide /usr/bin/yacc (yacc) in auto mode.08:17
kishore_Errors were encountered while processing:08:17
kishore_ rabbitmq-server08:17
jackhi dont know why08:17
kishore_jackh: Its ok...08:18
kishore_jackh: The command you gave in this TARGET_TOOLS_PREFIX=/media/one/origen_staging/android-toolchain-eabi/bin/arm-linux-androideabi- systemtarball userdatatarball ... What you are trying to do. The path should be diffrent for different machine .. right??08:19
kishore_jackh: I am not getting what you are trying to do in this line??08:21
jackhto set the toolchain prefix08:21
kishore_jackh: I don't have any one//origen_staging/android-toolchain-eabi/bin/arm-linux-androideabi  directory in my /media directory.. So how can we set the toolchain prifix which doesnot exist..08:22
jackhtry wget to get it from linaro08:23
kishore_jackh:Anyway when i run your command to set the toolchain prifix i got this error build/core/product_config.mk:205: *** No matches for product "origen".  Stop.08:23
kishore_To get what??08:24
jackhget the toolchain08:24
kishore_I already downloaded the toolchain...08:24
kishore_I have it..08:24
kishore_I just tar it...08:24
jackhthen set it with TARGET_TOOLS_PREFIX=08:24
kishore_jackh: I exported the tool chain....08:29
kishore_What next I have to do??08:29
jackhif u repoed all the source code, then compile ti08:29
kishore_jackh; How??08:30
jackhwith a command simillar with this08:30
jackhsudo make -j16 HOST_CC=gcc-4.5 HOST_CXX=g++-4.5 HOST_CPP=cpp-4.5 TARGET_PRODUCT=origen TARGET_SIMULATOR=false TARGET_TOOLS_PREFIX=/media/one/origen_staging/android-toolchain-eabi/bin/arm-linux-androideabi- systemtarball userdatatarball08:30
arunbangari_I am trying to enable the eject command in busybox and when I try to compile I am getting erros..any help is appreciated :)10:41
mpoirierbero: good day14:49
mpoirierpfefferz: you around ?15:04
asac_bero: hello15:06
asac_will come back15:07
asac_sudo make -j16 HOST_CC=gcc-4.5 HOST_CXX=g++-4.5 HOST_CPP=cpp-4.5...15:15
asac_does that need to be exactly 4.5? or would 4.4 host also work?15:15
asac_bero: ?15:15
asac_I assume 4.6 isnt good either?15:15
bero4.6 works just fine15:16
beroso is 4.715:16
beroand 4.815:16
beroJust 4.4 is way too old15:16
asac_bero: for the host?15:16
asac_bero: how is that? google uses 4.4 upstream, no?15:16
asac_oh wait15:16
asac_they use 4.4 for target but 4.5 for host?15:16
beroI'm not sure what they use15:16
beroIt's possible that we broke 4.4 hosts with the strict-aliasing patches or something15:17
beroAll I know for sure is host 4.4 breaks the build15:17
asac_bero: some complain about GLIBC_... ?15:17
beroI use host 4.7 here (used host 4.6 until a while ago), so I know those are ok15:17
beroAh, right, that was the problem15:17
asac_bero: is that maybe becayuse we build our cross toolchain with 4.5?15:18
beroSo the problem in some more detail is that binutils gets built with 4.515:18
asac_and we cant build that with 4.4?15:18
beroAnd gold requires libstdc++, which has versioned symbols15:18
beroI don't know why we build the toolchain with 4.5 TBH15:18
beroIt's one of the things we do because it's always been that way15:18
asac_i would prefer if we would support ubuntu 10.04 out of the box... which has 4.4 host15:19
asac_thats LTS15:19
asac_and my rule was always we want to support building our stuff on most recent available LTS15:19
asac_and newer15:19
asac_bero: i guess i can just make my own toolchain build and try?15:19
beroIMO that's a dinosaur best forgotten about, but I can see why it would be useful to some15:19
beroI don't think it will cause problems15:20
asac_its a matter of supporting stuff for companies... those prefer to roll LTS15:20
asac_similar to how googles main instructions are even for 8.04 :)15:20
asac_bero: what job should I clone to try that?15:20
asac_https://android-build.linaro.org/builds/~linaro-android/toolchain-4.6-bzr/ ?15:20
pfefferz asac_ one sec here15:20
pfefferzasac_, we need to make sure we can build on the distro that Google specifies15:21
beroDepending on what you want, toolchain-4.6-bzr or toolchain-4.6-2012.0215:21
asac_pfefferz: right. but we dont :)15:21
asac_pfefferz: google says 8.0415:21
asac_we currently dont even support 10.0415:22
asac_we support 10.10 upwards15:22
asac_i found this because hackbox is running 10.0415:22
asac_bero: i dont see a tweak in the build args to change what gcc to use for building. guess thats not exported as a config?15:22
pfefferzGoogle says we should be able to build that stack AOSP on 10.0415:23
asac_let me check the google instructions15:23
asac_pfefferz: yep.15:23
beroIt isn't... But it's quite likely that binutils/gcc/gdb configure scripts just follow standards and will do what you mean if you export CC=gcc-2.0 or something15:23
pfefferzand people have built Android on the hackbox15:23
asac_pfefferz: so for that we need to spin our toolchain with 4.415:23
asac_and not 4.515:23
beroWe can build on prehistoric distros - as long as people install an updated compiler15:24
asac_pfefferz: on hackbox.linaro.org?15:24
asac_pfefferz: i doubt it ... all build depends have not been there15:24
asac_i installed them15:24
pfefferzbero, asac_ it sounds like theres some work to do here15:24
asac_dont bother15:24
pfefferzand a wiki page to put together15:24
pfefferzso that we can document it15:25
pfefferzI thinks its okay to be different, we just have to document it15:25
asac_the only thing that currently diverges from googl instructions is that you need to get a 4.5 host compiler from somewhere15:25
asac_which is only introduced from what i know because we spin our androideabi toolchain with a 4.5 host compiler15:26
pfefferzlike building with the 4.6 and 4.7 toolchain15:26
asac_pfefferz: right. but as bero said its not really a technical reason15:26
asac_for the host compiler to need to be 4.5 or higher15:26
asac_its just that we didnt care or notice15:26
beroWe don't know for sure, we just never tried15:26
asac_i am happy to try and play around and report back :)15:26
beroIt's possible that gold will barf on 4.4's prehistoric libstdc++15:26
asac_could be :)15:26
beroBut rather unlikely, gold doesn't make use of that many C++ features15:26
pfefferzasac_, would you build with everything on defaukt15:26
asac_pfefferz: you mean plain google instructions?15:27
asac_and not our instructions?15:27
asac_i can do that15:27
beroActually you have to edit build/Makefile.in in the toolchain15:27
asac_probably have to repo a clean android branch15:27
pfefferzand then file bugs for everything that doesn't work15:27
asac_and not ours15:27
beroThat has a couple of instances of15:27
asac_happy to15:27
beroif which g++-4.5; then export CC=gcc-4.5; export CXX=g++-4.5; fi ;15:27
beroThat need to be adapted of course15:27
pfefferzthen we'll go grom there15:27
pfefferzbero, would you put a little explanation about the current situation together15:28
pfefferzas far as you know it15:28
beroThose exports are there because we wanted to prevent the build from picking up gcc 4.6 on android-build, to support prehistoric distros15:28
beroThat may actually be another reason, do we even have a 4.4 compiler in the android-build environment?15:28
beroIf we don't, we can't really tell the build to use it15:28
pfefferzpfalcon, ^^^^15:29
asac_bero: yeah. can you commit suppport for selecting the toolchain to our scripts? or is that going through pfalcon ?15:29
beroI can take care of that15:29
pfefferzor bero I guess  :)15:29
beroAs long as the toolchains are installed in the environment15:29
asac_bero: good point. but even what is installed is in the scripts15:29
beroBasically we can make the "if which g++-4.5; then export CC=gcc-4.5; export CXX=g++-4.5; fi" bits already present in Makefile.in configurable15:30
asac_bero: so if maverick (which is what we use) uses 4.5 by default you can probably just install gcc-4.4 as well15:30
asac_bero: i dont like which ... i want to set the stuff through a build config15:30
asac_so i can select what to use in case both are installed15:30
asac_on top you add which as a fail safe i guess :)15:30
beroSelecting what is installed is pfalcon's job -- I don't have access there15:30
asac_anyway ... let me build plain AOSP for now15:30
pfalconasac_: support for selecting a toolchain was available for long time (many months)15:30
asac_pfalcon: we talk about the toolchain build15:31
asac_pfalcon: not the android build15:31
beroI can just edit the scripts inside the toolchain build15:31
asac_pfalcon: is that avialable there as well?15:31
pfalconasac_: ah, ok15:31
asac_bero: yep15:31
asac_bero: do you have commit access?15:31
pfalconasac_: actually, if it's about what packages installed on build slave, it's static for all types of builds15:31
beroasac_: To the Makefile.in (with the export CC=..... bits), yes15:32
asac_pfalcon: right. but we have gcc-4.4 installed? can you check?15:33
asac_bero: ok. if that allows us to pass through a config fromt he build server then thats what we want i guess15:33
pfefferzI think at the end of this discussion, we should support the instructions listed at: http://source.android.com/source/building.html15:33
beroasac_: Yes - we just need to make sure the compilers being selected are actually present.15:34
asac_its a good exercise that we have 10.04 now on hackbox.linaro.org15:34
pfalconasac_: http://bazaar.launchpad.net/~linaro-infrastructure/linaro-android-build-tools/trunk/view/head:/node/setup-build-android15:34
asac_bero: sure. but i think its a user mistake if they are not15:34
asac_so we dont need to do much sophisitication15:34
asac_my take15:34
pfalconasac_: that's what we *install*. as for what *is installed*, it's a Natty machine15:34
asac_let me check whats the default there15:35
pfalconasac_: I can check specifically for gcc-4.4 if needed15:35
asac_otherwise we need to add gcc-4.415:35
beroAnd g++-4.4 and the corresponding libstdc++-dev15:35
asac_pfalcon: do you have a natty chroot somewhere to check?15:36
* asac_ logs into a porter box at canonical that has all chroots15:36
pfalconasac_: no, but I can startup a build slave15:36
asac_oh no... the porter box name doesnt exist anymore15:37
* asac_ is getting old15:37
pfalconasac_: btw, this is good usecase for having dkg -l output as a build artifact, as we discussed previously15:37
asac_ok default is 4.5 in natty15:38
asac_pfalcon: so yeah, add extra 4.415:38
pfefferzbero, asac_ pfalcon https://blueprints.launchpad.net/linaro-android/+spec/improve-build-environment15:39
pfalconasac_, bero, pfefferz: Care to submit a ticket to https://bugs.launchpad.net/linaro-android-infrastructure ?15:39
asac_pfalcon: we dont have all what is needed to build the trunk manifest on our android.g.l.o15:44
asac_is that known?15:44
asac_ * [new tag]         upstream/r30707 -> upstream/r3070715:44
asac_error: Cannot fetch platform/external/libselinux15:44
asac_error: Cannot fetch platform/external/sepolicy15:44
pfalconasac_: no.  Care to submit a ticket to https://bugs.launchpad.net/linaro-android-infrastructure ? ;-)15:45
asac_pfalcon: bug 95091315:46
ubot5`Launchpad bug 950913 in Linaro Android Infrastructure "android.git.linaro.org doesn't have all AOSP trunk manifest trees mirrored" [Undecided,New] https://launchpad.net/bugs/95091315:46
pfalconasac_: thanks!15:47
asac_pfalcon: on top of a one time effort, I guess this should end up with a process that will ensure we will always have what is on trunk15:48
asac_that might be wortha blueprint15:48
asac_or whatever your team uses15:48
pfefferzmansson, ping15:49
asac_it is sooo blazingly fast to get the git stuff on that hackbox :)15:49
manssonhi pfefferz.15:49
pfalconasac_: we have what's in trunk, we just don't fetch automatically new projects. So, yes. I'll add my thoughts to the ticket, and make sure it's discussed with Danilo.15:49
asac_pfalcon: rock on15:50
koudI have build android with linaro toolchain on 10.0415:51
koud4.6 2012.02 release toolchain15:51
asac_koud: linaro android? or aosp?15:51
koudbut currently buggy15:52
koudthis forexample, https://bugs.launchpad.net/linaro-android/+bug/90629715:52
ubot5`Ubuntu bug 906297 in Linaro Android "Live wallpaper crashing on ICS -panda" [High,New]15:52
asac_koud: yeah. the fixes for that we have in our tree15:52
asac_(i hope)15:53
beroActually that one isn't fixed yet15:53
beroIt's a rather obscure bug in the code generator that needs more work15:53
beroI may have a fix, but it's untested currently15:53
kouddefault aosp build does not use -O3 or any other optimization flags?15:54
beroaosp uses -O2 for arm code and -Os for thumb code15:55
pfefferzmansson, did you get all the cards showing up in status.linaro.org15:55
beroBut this is unrelated to compiler flags15:55
asac_pfefferz: had a sync with lool an hour ago and we said we add them all to the NEW lane NOW ... if nothing happens today ping me and i will just do it for him15:56
koudwhat are those flags for then?15:56
beroOf course the flags affect how stuff is built, and you typically get faster code with -O315:56
manssonpfefferz. The trunk fix was merged today, not sure when it starts to work Will check...15:56
beroBut bug 906297 isn't related to optimization15:56
ubot5`Launchpad bug 906297 in Linaro Android "Live wallpaper crashing on ICS -panda" [High,New] https://launchpad.net/bugs/90629715:57
beroThe problem with that bug is that libbcc seems to rely on nonstandard behavior of gcc 4.4 that was since "fixed"15:57
koudah no ofcourse, I was just curious :P15:57
pfefferzasac_, mansson cool15:57
koudreason I tried building with linaro-toolchain was to try different compiler optimizations15:57
pfefferzasac_, I'll get the rest of the content in them15:57
pfefferzand finish the BP layout15:58
pfefferzbased on the plan I gave you15:58
pfefferzasac_, one thing15:58
pfefferzasac_, if we have the content in the TSC cards15:58
koudbut looking at the toolchain benchmarks it seems there is not much benefit compared to default aosp toolchain?15:58
pfefferzshould we put the discreate steps in15:58
pfefferzor just kind of give an over view of the work?15:58
asac_pfefferz: right. have you checked what dzin and ricardo did? they put the steps in the cards. i am not asking for that, but all the steps having nice headlines and acceptance is a good one to take from there15:59
asac_let me forward you their cards15:59
asac_look in each cards for the Work blocks ... those are kind of what i would expect to be in one blueprint each at this stage16:00
pfefferzokay, so those will be in the BPs16:01
pfefferzand I'll follow the same general form16:01
asac_cool. check what is good ... what is bad and innovate :)16:02
pfefferzasac_, would you send me a card that has the correct visibility requirments16:15
asac_pfefferz: https://linaro.papyrs.com/ANDROID2012-TEST-API seems to be fine. did I complain about that?16:23
asac_i can view it without being logged in16:23
pfefferzasac_, no I just want to make sure they're all right16:24
asac_pfefferz: to be super save use ...16:26
pfefferzasac_, k16:26
asac_those is one of the very first I created from the original copy16:26
pfefferzhmm... that just says Shared with All, Public16:28
pfefferzthat's easy16:28
=== don is now known as Guest85526
=== zyga is now known as zyga-afk
pfefferzasac_, where don't the accesibility controls show up for all cards16:31
pfefferzlike: https://linaro.papyrs.com/ANDROID2012-SMP-IMP16:31
pfefferzasac_, nevermind16:32
pfefferzasac_, I figure it out16:32
pfefferzmansson, would you give me full access to:16:33
manssonHm. hold on...16:33
manssonpfefferz: for ANDROID2012-TEST-API you have cpmlete access16:35
pfefferzmansson, I must not16:36
pfefferzI can't edit visibility16:36
pfefferzis there an admin setting?16:36
manssonAll(View, Edit & Admin) is what you have on all three16:37
pfefferzasac_, I have complete permissions for16:38
manssonpfefferz, on https://linaro.papyrs.com/pages/alphabetical/ do you see a grey box with a pen to the far right?16:39
manssonIt seems without the pen you don't even get a button. With the pen you get a button but the list od persons is empty UNLESS you are the creator of the page.16:44
asac_pfefferz: not sure what you mean16:46
asac_pfefferz: if you start with copying the card i gave you tings hopefully are fine :/16:47
pfefferzwhen I click on "Page Settings" on https://linaro.papyrs.com/ANDROID2012-TOOLS16:47
pfefferzI don't see the "Who has access controls"16:47
manssonDo page/page settings. I only see the list of people to modify rights on if I created the page. If somebody else submitted the page the list is empty Even if I have the right to delete the page. (am Admin)16:49
manssonExactly because I created that page.16:49
pfefferzthat's weird16:49
manssonThe recverse is also tru. I see blank page for pages that you created.16:49
pfefferzmansson, I guess make sure https://linaro.papyrs.com/ANDROID2012-SMP-IMP16:50
manssonThats rsalveti's and it is blank to me. Probably to you too?16:51
manssonpfefferz by blank I mean page settings is empty.16:52
pfefferzthat's why you can never see the permissions16:52
manssonBug or feature?16:52
pfefferzallright well everything is marked public now16:53
manssonMaybe I should file a bugreport to infrastructure.16:55
manssonpfefferz, asac. It appears that status.linaro.org does not care about papyrs, it only cares about roadmap.linaro.org. Thats why our cards and blueprints does not show up on http://status.linaro.org/lane/2012Q1.html . I guess I didn't realize that.17:05
bhojrsalveti, hi17:08
rsalvetibhoj: hey17:13
bhojrsalveti,  i wanted to know the naming you  are planning to have fastmodel builds . would like to use the same in android .17:14
rsalvetibhoj: check that with suihkulokki (#linaro)17:14
rsalvetihe's working on changing our tools to support fastmodel17:14
bhojrsalveti,  ok .17:14
=== zyga-afk is now known as zyga
=== dzin is now known as dzin_lunch
asac_mansson: right. as i mentioned above, if nothing happens today I will put all the new cards into the NEW lane ... at least the cards show up then17:29
asac_(not in the right quarter though)17:29
jedixpfefferz: I'm trying to build the kernel as per your email from  yesterday, but I'm getting an error "scripts/kconfig/zconf.tab.c:206:24: fatal error: zconf.hash.c: No such file or directory17:30
pfefferzwould you send my a mail with the exact steps?17:34
jedixthis is in the 12.02 clone, not the pull of the git kernel tree17:40
jedixit wanted me to make mrproper, so I did..17:40
jedixoh, hehe.. acutally I think it was that I didn't have ARCH set when making mrproper17:42
pfefferzthat may cause a problem17:42
jedixyeah, that's what it was.17:43
jedixit always wants me to run mrproper though17:43
mpoirierpfefferz: http://review.android.git.linaro.org/#change,1569 pls.17:59
jedixpfefferz: any idea on my issue?18:06
jedixdrives me nuts when something tells me to clean out my config when I need that to do what I want18:13
pfefferzmpoirier, on it18:32
pfefferzjedix, :)18:32
pfefferzjedix, hows the Trustzone stuff coming18:32
pfefferzmpoirier, merged18:33
mpoirierpfefferz: cool - I'll start a build18:33
=== dzin_lunch is now known as dzin
panicking_mikepfefferz, Hi19:01
panicking_mikeI'm not in that mailing list with my email19:01
panicking_mikedo you know where is the link to subscribe19:02
panicking_mikeI need to add android-platform on the loop19:06
panicking_mikeneed to go, I will do later :)19:06
pfefferzthat guys great19:15
_Thomaspfefferz: Ok, got the new android-3.3 kernel running late today19:23
pfefferz_Thomas, nice19:23
_Thomaspfefferz: With HW rendering (mali400)19:23
pfefferzis it the one we pointed to you?19:23
pfefferzmind sending a main about it to the list?19:24
_ThomasSure, but is there any reason not to include platform stuff for mali in the origen tree?19:25
_ThomasWe are thinking seriously about submitting patches to get our stuff included in the linaro origen tree, if that's the way to move forward to get things upstreamed?19:26
pfefferzthe reson is19:26
pfefferzwe can't19:26
pfefferzbecuase of how it has to be distributed19:26
pfefferzactually it should be in the platform tree (at least Mali support)19:26
_ThomasHuh? You can't distribute the parts in the kernel even?19:26
pfefferzfor MM no19:27
_Thomasyou are still missing stuff in the 3.3 tree to get mali working19:27
pfefferz_Thomas, I know...19:27
pfefferzdo you have some patches?19:27
_ThomasYes, I'll prepare some later today19:27
_ThomasWe're running 1080p 32bpp on hdmi now19:28
pfefferzplease send them in and we'll get them on the baseline19:28
pfefferzhey cool19:28
_ThomasBut since the vb2fb stuff is only single buffering, we don't have double buffering, yet19:28
_ThomasI also suspect that fimc is wrong on 3.3 kernel, but we're going to look at that next week19:34
pfefferz_Thomas, cool20:03
=== zyga is now known as zyga-weekend
pfefferzbero, +1 active desktops looks great!20:34
GregBHi all. We are running latest linaro android on panda ES. SIP registration fails with error 'no data connection'22:08
GregBAnyone has any ideas what is wrong?22:08
GregBDo have wifi connected.22:08
doanacGregB: most android guys are probably gone for the weekend. You might try posting your question to the linaro-android mailing list.23:04
beroWho maintains origen these days?23:57
beroGregB: No idea... I don't think anyone of us has tried to use SIP so far. Can you ping the SIP server?23:57

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