slangasekdoko_: hmm, which qemu-linaro release has static lib problems?  The latest upload seems to build successfully - are there features silently disabled that I should be worrying about?00:34
plarsliuyq: hi01:58
plarsmwhudson: hi01:58
liuyqplars, hi01:58
liuyqplars, is ok now?01:59
plarsliuyq: yep01:59
=== michaelh1 is now known as michaelh1|away
=== michaelh1|away is now known as michaelh1
mwhudsonplars: hi02:47
plarsmwhudson: packaged and installed lava-android-test today02:48
plarsmwhudson: trying to get it running actual jobs02:48
mwhudsonplars: \o/02:48
plarsliuyq: ok, new tarball released, working on the package now03:32
liuyqok, thank you very much. I will submit a job and try. Please to to sleep.03:41
plarsliuyq: just did03:45
plarsliuyq: it's running now03:45
liuyqit failed:(03:54
liuyqit's better you go to sleep first. I can reproduce that error on my machine. It's strange. I will investigate that.03:56
liuyqI am sorry for so many bugs:(03:56
plarsliuyq: np, I hit something else though03:58
plarsliuyq: http://validation.linaro.org/lava-server/scheduler/job/199803:58
liuyqFor 1998, the error is "AttributeError: 'module' object has no attribute 'testobj'04:00
liuyq". I don't know why this occurs, I am investigating it now04:00
plarsliuyq: that may be my mistake04:00
plarsliuyq: I assumed those directories were something they were not04:00
plarsliuyq: hmm04:00
liuyqmay be the monkey directory and the monkey.py conflict04:01
plarsliuyq: right04:02
plarsliuyq: __init__.py was the wrong thing, I just need to have it keep that directory I guess04:02
liuyqso how can we fix? or rename the monkey directory?04:03
plarsliuyq: sorry, was in another window04:18
plarsliuyq: I just added it to the manifest04:18
plarsliuyq: ok, I will respin it in the morning, I have it patched on the server now04:19
liuyqok, thank you.04:20
liuyqI will first try it on my local. 04:21
liuyqdid you install the latest on validation.linaro.org?04:21
=== michaelh1 is now known as michaelh1|away
plarsliuyq: http://validation.linaro.org/lava-server/scheduler/job/199904:46
plarsliuyq: still hitting some errors04:47
plarsliuyq: but I need to go to bed04:47
plarsliuyq: sorry, will look at it more in the morning04:47
plarsdeepti: hi04:47
deeptiplars: Hello04:47
plarsdeepti: hey, could you resubmit one of your jobs that is failing to panda24?04:47
plarsdeepti: I'm just curious if it makes any difference, might not04:48
plarsdeepti: just if you get a chance04:48
deeptiplars: Ah! I just came into office .. I will try and do it in sometime.. 04:48
deeptiplars: do u want to be done immediately04:48
plarsdeepti: sure, whenever... just let me know if you get a chance04:48
plarsdeepti: nah, I'm going to bed.  Just if you get a chance04:48
deeptiplars: sure I will http://validation.linaro.org/lava-server/scheduler/job/1999 is this the one ?04:48
plarsdeepti: I don't expect it to make a difference, but who knows?04:48
plarsdeepti: no04:48
deeptiplars: before u leave04:48
plarsdeepti: that's something else04:48
plarsdeepti: I mean one of the kernel ci builds04:49
deeptiplars: did u see the reboot problem still there 04:49
plarsdeepti: that's what I am trying to find out04:49
plarsdeepti: I only see that with your jobs04:49
deeptiplars: I tried debugging yesterday night .. but no luck.. 04:49
plarsdeepti: I don't expect this to help, but maybe worth a try04:49
deeptiplars: yes may be there is something fishy in my hwpacks.. 04:49
plarsdeepti: panda24 has an updated uboot04:49
deeptiplars: I had not seen such a thing previously04:49
deeptiplars: but now since last week when u informed me I observed it..04:50
deeptiplars: thanks for pointing though04:50
plarsdeepti: sure :)04:50
deeptiplars: I will submit the job in a while..04:50
plarsdeepti: thanks for looking into it04:50
* plars waves04:50
=== plars is now known as plars-zzz
zmaI would like to enable DVI-D output on Pandaboard. I read that I need modify some bootargs, but can I modify them directly onto boot.ini file on SD card?05:42
mwhudsoni don't really know much about this sort of thing05:43
mwhudsonbut i think the answer is "yes, but it's more complicated than just editing the file"05:43
zmayeah, I suppose it is, because it didn't work :)05:43
mwhudsonyou need to run some mkimage command to put the file into a format uboot will read05:43
zmayes, I suppose something like that needs to be done.. I try to research it further..05:44
zmait's just that if only modifying boot.ini would be enough, I'd know there is something else wrong..05:45
deeptimwhudson: regd the 860022 .. u need the git describe to be part of the test run json file or it should be part of the build status json as well ?05:49
mwhudsondeepti: just the build status json would be fine05:49
deeptimwhudson: ok.. fine thanks05:49
zmaIs it possible to set bootargs at this phase? sudo linaro-media-create --rootfs ext3 --mmc /dev/sdb --dev panda --hwpack hwpack_linaro-lt-panda-x11-base_20110824-1_armel_supported.tar.gz --binary linary-n-ubuntu-desktop-tar-20110824-1.tar.gz05:56
=== michaelh1|away is now known as michaelh1
deeptimwhudson: is this what you expect 05:58
deeptigit describe05:58
mwhudsondeepti: not really06:00
mwhudsondeepti: maybe jenkins is tagging the tree you build from or something06:00
deeptimwhudson: or something like this git describe06:01
mwhudsondeepti: that latter06:01
deeptimwhudson: u mean v3.0.3-756-gf392cbf06:01
mwhudsondeepti: yes06:01
deeptimwhudson: ok...06:02
mwhudsondeepti: seems you maybe need to change the jenkins config06:02
mwhudsondeepti: https://issues.jenkins-ci.org/browse/JENKINS-5676 is relevant06:02
deeptimwhudson: to get that describe information ?06:03
mwhudsondeepti: yeah, to prevent jenkins making its own tag and so affecting the git describe output06:03
deeptimwhudson: ah! no need I can get that info .. its already present in the hwpack_info file06:04
mwhudsondeepti: ah excellent06:04
deeptimwhudson: I will just need to use the info in the build status06:04
deeptimwhudson: done .06:08
mwhudsondeepti: thanks!06:08
=== vila is now known as babune
=== babune is now known as vila
asac_zma: try our panda 11.09 RC builds or latest dailies ... i heard rumors that dvi and hdmi are now working and even with EDID/autodetect06:26
asac_zma: see http://lists.linaro.org/pipermail/linaro-dev/2011-September/007817.html06:28
asac_zma: and dont hesitate to fill out our test forms :)06:29
zmaasac_: thanks, that's good news. I also got mkimage work for 11.08. But I think I want to jump to 11.09 nevertheless :)06:32
asac_i agree :)06:33
faboasac_: there's still some corner case for EDID, bug 84962506:35
ubot2Launchpad bug 849625 in linaro-landing-team-ti "EDID and HDMI Audio not correct on Sony Bravia KLV-32J400A" [Low,Confirmed] https://launchpad.net/bugs/84962506:35
asac_details :)06:36
hrw~curse dash06:39
zygadeepti: hi06:42
deeptizyga: Hello06:42
faborsalveti: anything changed on the lt-origen hwpack?06:51
hrwhej czajkowski 07:09
rsalvetifabo: not that I know07:09
rsalvetifabo: slt should know07:10
czajkowskihrw: hi there 07:10
mabacfabo, rsalveti Vexpress hwpack fix. will need re-release and re-packaging of l-i-t if we want v2 hwpacks for Vexpress https://code.launchpad.net/~mabac/linaro-image-tools/bug-860225/+merge/77104 07:16
=== mabac is now known as mabac|afk
rsalvetimabac|afk: fabo: guess v1 should be fine for the release07:17
rsalvetiat least for vexpress07:17
hrwrsalveti: you are nosleep man ;)07:18
rsalvetihrw: haha, just got home07:18
hrwrsalveti: if all goes properly we will have multilib capable cross compiler today07:20
rsalvetihrw: oh, that would be awesome07:21
hrwadded some hackery to arm(el|hf)-cross-toolchain-base to get replacement of libc6-dev-armel_armhf (and vice versa) + symlink07:23
fabomabac|afk: rsalveti: I'm updating known issues for l-i-t/vexpress hwpack v207:24
rsalvetihrw: hm, cool07:27
rsalvetineed to check the details later :-)07:27
aviksilrsalveti: where can i find the kernel dbgsym package?07:44
drkhello, i'm having some random crashes w/ linaro (both 10.08 and 10.09) on my panda board, under heavy load. Most of print a kernel oops at the console (http://pastie.org/private/ojo1wnzfgm8kdlchacojdw << an example)07:45
drkany ideas ?07:45
rsalvetiaviksil: hm, seems jcrigby tried to fix it but it's still broken for current packages :-(07:45
rsalvetiso if you really want to test now, would suggest you to rebuild the package locally 07:45
rsalvetiuntil jcrigby is up again07:46
aviksilrsalveti: ok, will do :)07:46
=== mabac|afk is now known as mabac
mabacrsalveti, fabo ok. then there's no rush to get the fix released08:03
zmaasac_: tried 11.09, filled the web form and reported that DVI-D not working.08:11
zmaasac_: It works on 11.08 after modifying boot.scr with mkimage tool. But fresh install of 11.09 doesn't.08:12
doko_slangasek, 2011.09, davidgiluk did recommend building with --disable-cardsupport for the use mode packages08:14
asac_zma: did you file a bug?08:16
asac_that would be awesome08:16
asac_zma: against launchpad.net/linaro-ubuntu -> report  a bug08:16
rsalvetifor panda, as I reported, the hwpack is not using the latest lt kernel...08:19
rsalvetibecause the kernel was integrated after the 0926 build08:20
rsalvetifabo: panda hwpacks 20110927-1 should have latest lt kernel and u-boot integrated08:22
rsalvetiwe'll still need a respin after the debug package fix is properly uploaded08:22
rsalvetibut nothing that affects the testing08:22
rsalvetiso a respin agains the current panda hwpacks would be a nice thing to do08:22
* rsalveti away for a few hours08:23
asac_rsalveti: aye08:25
asac_there are two more days left :)08:29
birisalf_: ping https://bugs.launchpad.net/bugs/85133408:42
ubot2Ubuntu bug 851334 in linaro-android "GLMark2 does not run on Panda/Snowball" [Medium,Confirmed]08:42
alf_biris: yes, I am looking into it08:42
birisalf_: great :-) was hoping to get an update ... 08:43
davidgilukmru: Is that pngbench your own creation or a standard benchmark somewhere? (googling it gets stuff on Papua New Guinea...)08:53
doko_slangasek, pm215: how much are the openbios-ppc packages needed for qemu?09:00
davepigottdmart: ping09:05
janimorsalveti, are there oneiric based 11.09 images too?09:07
dmartdavepigott: pong09:08
davepigottdmart: quick question: Paul has asked me to upgrade the images on the pandas, and says he did it with an apt-get diet-upgrade and then manually copying the MLO and u-boot.img to the first partition...09:09
davepigottdmart: 1) What's the MLO? 2) Where would the u-boot.img be and 3) how do I copy them to the right place while I'm on the running panda?09:10
dmartdavepigott: MLO is just some first-stage bootloader.  It goes in the FAT partition at the start of the SD card like the other images.09:11
janimodavepigott, both are in the u-boot package for omap409:11
dmartYou should just be able to mount that FAT partition on the board and overwrite the images there09:11
janimounder /usr/lib/linaro-u-boot-omap4 or something09:11
davepigottdmart: janimo: Excellent. Thanks both!09:12
dmartYep, after the dist-upgrade, the updated images should be there09:12
dmartI'm assuming that uImage, uInitrd and the dtb blob get upgraded automatically09:12
davepigottdmart: I would assume so09:12
janimoindeed, I don't think manual copying is needed unless upgrade fails or you want to copy different non-packaged binaries there09:12
davepigottdmart: Paul said it worked.09:12
=== michaelh1 is now known as michaelh1|away
mwhudsondavepigott: hey, would it be possible to get a symlink to the latest validation db dump?09:29
zygamwhudson: hi09:35
zygastill up09:35
zygamwhudson: do you hae a moment for a quick chat?09:35
mwhudsonzyga: not for long09:35
mwhudsonzyga: not really09:35
=== doko_ is now known as doko
zmaasac_: bug created: https://bugs.launchpad.net/linaro-ubuntu/+bug/86040209:50
ubot2Ubuntu bug 860402 in linaro-ubuntu "DVI-D display not working on Pandaboard" [Undecided,New]09:50
asac_zma: asked something in bu09:52
hrwbug 860432 - ideas?10:30
ubot2Launchpad bug 860432 in multistrap "tried to grab repository for wrong architecture " [Undecided,New] https://launchpad.net/bugs/86043210:30
davidgilukhrw: This on oneiric?10:31
hrwlooks like I need to go to natty or older chroot10:31
davidgilukhrw: Have you got the foreign-arch stuff setup to do 386 and 64bit?10:32
* davidgiluk bets it's picking that up10:32
hrwso bug in apt even10:33
davidgilukhrw: Maybe, I've not quite figured out how that stuff works - I thought it was a dpkg config but then apt picks it up10:33
zmabug 860402. DVI-D works with latest hwpack, but new problems came ahead: pulseaudio consumes 99% CPU and some graphical artifacts on screen.10:54
ubot2Launchpad bug 860402 in linaro-ubuntu "DVI-D display not working on Pandaboard" [High,New] https://launchpad.net/bugs/86040210:54
perter-jimI have a question11:09
perter-jim[hdmi/edid]HDMI timings->pixel_clock->: 7425011:09
perter-jim237 <4>[    8.257934] [hdmi/edid]HDMI timings->x_res->: 128011:09
perter-jim238 <4>[    8.257934] [hdmi/edid]HDMI timings->y_res->: 72011:09
perter-jim239 <4>[    8.257965] [hdmi/edid]HDMI timings->hfp->: 18411:09
perter-jim240 <4>[    8.257965] [hdmi/edid]HDMI timings->hsw->: 4011:09
perter-jim241 <4>[    8.257995] [hdmi/edid]HDMI timings->hbp->: 47611:09
perter-jim242 <4>[    8.257995] [hdmi/edid]HDMI timings->vfp->: 511:09
perter-jim243 <4>[    8.257995] [hdmi/edid]HDMI timings->vsw->: 511:09
perter-jim244 <4>[    8.258026] [hdmi/edid]HDMI timings->vbp->: 2011:09
perter-jimtimings 720P?11:10
mrudavidgiluk: pngbench is a simple tool I made11:21
mruall it does is decode an image in a loop11:21
davidgilukmru: Could you share a copy - I'd like to profile it to see the type of memcpy's that are being called11:21
davidgilukmru: Combination of size and alignments11:22
zygahrw: hi11:24
hrwzyga: ho11:45
=== Quintasan_ is now known as Quintasan
zmahmm.. I blaclisted wl12xx at /etc/modprobe.d/blacklist.conf, however wl12xx module still gets loaded..12:02
zmaI'm trying to disable Pandaboards internal WLAN12:03
asac_zyga: do you know why lt-panda has more pwermgmt tests run than panda plain?12:05
ogra_zma, did you rebuild the initrd ?12:05
asac_zyga: http://validation.linaro.org/lava-server/dashboard/image_status/ubuntu-desktop+lt-panda/12:05
asac_zyga: vs. http://validation.linaro.org/lava-server/dashboard/image_status/ubuntu-desktop+panda/12:05
zmaogra_: I didn't.. I didn't know it's needed12:05
ogra_zma, well, for modules shipped in there it is :) else your module would be loaded before the blacklist kicks in 12:06
zygaasac_: nope12:06
ogra_not sure if the wlan mod is in there, but it only costs a sudo update-initramfs -u and a reboot to find out ;)12:06
zygaasac_: I suspect it is due to the frequency of submitted tests12:06
zygaasac_: but as we don't have a periodic scheduler I cannot really say12:06
asac_zyga: there are different test names12:07
zmaogra_: is there easy way to tell Ubuntu that I want to use USB WLAN dongle instead of internal WLAN?12:07
asac_zyga: if you click on them12:07
asac_lt-panda seems to run different tests :)12:07
ogra_you can surely do that through network-manager12:07
zygaasac_: that is perfectly possible12:08
asac_zyga: how?12:08
zygaasac_: the tests you wish to run are listed in the job you submit12:08
asac_but why are the daily lt-panda vs. panda different set up?12:08
asac_that feels like an issue :)12:08
zygaasac_: I don't know12:08
asac_plars-zzz: ^^12:08
zygaasac_: I'm not involved in that12:08
zmaogra_: I'm not so sure, atleast the network-manager-gnome doesn't give me a choice. Dunno if it has command line utility with more power.12:11
ogra_it should show you both devices in the panel menu12:11
ogra_just pick the USB one12:11
ogra_works fine here on my tegra netbook with oneiric installed12:12
zmano panel menu for network manager. If I go to tab "Wireless", and try to edit or add network, I have no choice for USB or anythhing.. and then it just continues to make connection with the internal ..12:15
* ogra_ wonders what kind of desktop you run there 12:15
zmaUbuntu 11.09 RC12:16
ogra_well, no idea what linaro uses as a desktop on that one 12:16
zmaGnome with Unity on top12:16
ogra_though i would expect it not to differ to much from an ubuntu oneiric image12:16
ogra_(where i definitely have a menu for NM that shows all attached network devices (including the 3G modem my device has)12:17
zmaI think I need to learn how to set it up from command line.. I can track from dmesg that device is attached and also see it on lsusb.. now I just need to tell from command line that this is the default device and then try to connect to access point12:22
plars-zzzasac_: that's a question for dlezcano, he wrote the test.  If I had to guess, lt-panda has more of the things enabled that they need12:22
davidgilukzma: Well see if there is any driver for it12:22
=== plars-zzz is now known as plars
asac_plars: ok12:22
asac_rsalveti: dzin: can you follow follow up eventually why the powermanagement test cases in lava change based on what hwpack we use?12:23
asac_that makes it HARD to create the powermanagement matrix :)12:23
asac_dlezcano: ^^12:23
asac_also you ;012:23
zmadavidgiluk: good point, I didn't look modprobe -l careful enough. I thought that the driver was something that isn't even on wireless category. I know PC Ubuntu 11.04 detects this WLAN card automatically, but it doesn't mean the driver would be included on Linaro by default I suppose.12:25
plarsdavepigott: hi, is panda24 offline for a reason?12:29
faboogra_: I subscribed you to our blueprint for ljt review ;)12:50
pfefferzamitdk, are you coming to the meeting?13:08
kikosalgado, ping?13:09
salgadohi kiko 13:09
kikosalgado, you want permissions for..?13:10
salgadokiko, I do?13:10
mabackiko, do you mean me perhaps? for the Kanbanboard? :)13:10
asac_kiko: mabac wanted permissions for the kandbanntool13:10
salgadokiko, I think it's mabac, no? 13:10
ogra_fabo, btw, i would have uploaded long ago if tgall_foo would have answered my question :) but it seems we miss each other regulary13:11
asac_ogra_: did you ask on bug?13:11
ogra_asac_, no13:12
asac_ogra_: might work better if you have zero overlap timezones :)13:12
ogra_we usually have overlap13:13
mabacumm has anyone noticed that status.linaro.org is not updating since Saturday. It spits out "HTTP Error 301: Moved Permanently" for something that's likely a project or series.13:13
asac_mabac: nope. where do you see that issue?13:13
kikomabac, you.. need more access, is that it?13:13
mabacasac_, http://status.linaro.org/update.log.txt13:13
kikomabac, oh, true. fixed now13:14
mabacasac_, last succesful update was 24 sep according to http://status.linaro.org/11.11/ 13:14
mabackiko, cool thanks! 13:14
asac_mabac: my guess is that series was renamed and status.l.o code doesnt deal with that properly13:14
asac_mabac: iirc there was a config file maintained somewhere saying which series to track ... maybe we can change to point to the 2011 series name?13:16
mabacasac_, there is a config file involved. I'll find it13:16
asac_mabac: or fix the code to properly go to location: header on 301 or 30213:18
asac_err 30713:18
tgall_fooogra_, I'm sorry ... I didn't realize you had a question13:19
ogra_tgall_foo, well, i pinged on #ubuntu-arm... whats the reason for all the lintian overriding ? why do you use such an old debhelper version =13:20
tgall_foo8.1.3 deb-helper is the latest greatest .. 9 isn't out yet13:21
tgall_fooogra_, correct me if I'm wrong13:21
dpcHi. Simple question. On Linaro's Ubuntu there are two images of linux kernel binary. One is on first patition along with u-boot.bin etc. And second is in the /boot directory. Why? :)13:22
hrwdpc: cause uboot reads first one 13:22
ogra_tgall_foo, ogra@horus:~$ apt-cache show debhelper|grep -i version13:22
ogra_Version: 8.9.0ubuntu113:22
tgall_fooso overrides ... in source.lintian-overrides we have package-needs-versioned-debhelper-build-depends 9  .. since ideally we want debhelper 9 for multiarch13:23
dpchrw: So why to keep second copy if first one would do?13:24
ogra_tgall_foo, thanks, thats all i wanted to know ... will upload now13:24
dpcAnd copies are not the same (slightly different013:24
tgall_fooogra_, ok thanks!13:25
dpcIs one generated somehow from the other?13:25
hrwdpc: /boot/ has plain image. first partition has uImage format13:25
davidgilukdpc: On some platforms the stuff in the u-boot directory has to be turned into a uboot image13:25
hrwdpc: some of debian/ubuntu devices needs to have booting kernel stored in flash directly - we use kernel in /boot and then store it wherever it is needed to boot device13:26
hrwdpc: and convert it to proper format - uImage, zImage, uImage+dt etc13:27
dpcThat explains a lot. Thanks. And last question is: where sits the code that does the coversion + copying?13:27
hrwdpc: /usr/sbin/flash-kernel mostly13:28
dpcThank you.13:30
rsalvetijanimo: I should have a working oneiric image later today13:31
rsalvetilet you know about my progress13:31
janimorsalveti, even for beagle?13:33
rsalvetijanimo: bug 851886 is still not fixed13:34
ubot2Launchpad bug 851886 in linaro-ubuntu "login not available on Oneiric based image for Panda" [High,Confirmed] https://launchpad.net/bugs/85188613:34
rsalvetiand I still need to copy the packages for oneiric 13:34
rsalvetito be in sync with the ones we have for the natty overlay13:34
janimorsalveti, I only have a beagle now. panda is broken13:34
=== nytowl_away is now known as nytowl
hrwfabo: what for is lp:~linaro-pkg team?13:42
fabohrw: it's created to Debian/Ubuntu packaging related work. As I started to push l-i-t and ljt to Debian, we needed a LP project for that13:44
dpcWhile I'm here: I was running 11.08 image on Qemu emulated vexpress platform and I didn't got the userspace outputs on the serial console. After "Freeing init memory ... " nothing were appearing. I used official binaries and followed the instruction. Is this a known issue with 11.08?13:44
ogra_tgall_foo, btw, calling autoreconf from the clean target is (despite not forbidden) really annoying ... you always need the build-deps installed even if you only want to roll a source package 13:44
ogra_(for the next package )13:44
hrwfabo: so any packaging stuff should be put there (when it comes to bzr branches)?13:44
fabohrw: yes, it might be good to centralize this effort and track the differences we could have between Debian/Ubuntu13:47
jcrigbyrsalveti, the dbgsym deb is still missing?14:03
jcrigbyrsalveti, looking at logs14:04
rsalvetijcrigby: yes14:09
=== shadeslayer is now known as shocklateboy42
rsalvetijcrigby: and it seems the fix for bug 669641 also didn't work14:21
ubot2Launchpad bug 669641 in linaro-ubuntu "systemtap fails to discover installed debug modules" [High,In progress] https://launchpad.net/bugs/66964114:21
rsalvetiat least from what aviksil said, still need to check14:22
=== shocklateboy42 is now known as shadeslayer
anmarjstultz_vm: Hello John.15:21
anmarjstultz_vm: your Android tree. Does it contain SoC specific bits for Android or just the common Android patches?15:22
ndecrsalveti: ping15:31
rsalvetindec: pong15:32
ndecrsalveti: hey! finally we meet ...15:32
ndecrsalveti: want to check with you how to handle nicely TI and Linaro 'overlay' PPAs...15:33
ndeci think every package will end up duplicated on both sides, and i can imagine most panda users to use both overlay most of the time... 15:33
rsalvetindec: well, I duplicated the packages basically because I wanted the oficial TI PPA to be used just at Ubuntu15:36
ndecrsalveti: if i understand correctly you will need to have the packages in the linaro overlay PPA to support the LEBs. similarly i need the same thing to support our releases...15:36
rsalvetito avoid conflicts and such15:36
rsalvetihaving both enable means that the TI PPA should be used for both Ubuntu and Linaro LEB releases15:37
ndecbut i can easily see ubuntu users to add the linaro overlay on their setup. and this is going to be a source of problems unless we are in sync15:37
rsalvetindec: sure, that's fine15:37
rsalvetionly thing that is different atm is the kernel packages15:37
ndecwhat if you just copy from TI PPA into linaro so that you control what gets into your LEB?15:37
rsalveti*kernel modules15:37
ndecyou mean pvr?15:38
rsalvetithat's fine, and that's what I did initially15:38
ndecwe are okay to rename our packages if that makes our lives simpler.15:38
rsalvetibut then I had to change it with some extra fixes, to support latest kernel and such, and that's why I then updated them only at the Linaro Overlay PPA15:38
ndecwe can also push the gbp git tree somewhere in case someone want to contribute15:38
rsalvetindec: sure, that would work15:39
ndecare you planning to copy the gst and ducati firmware as well when they are available?15:39
rsalvetindec: yes, if we're allowed to do that (in the license pov)15:40
ndecok ianal but if i am able to put in LP why wouldn't you be able to copy? 15:40
ndecit's already published on a non TI site anyways15:40
rsalvetindec: because you could put at the license that people are only able to get it from an specific location and such15:40
rsalvetioh, ok, then it should be fine15:41
ndecok... i am still expecting to receive the new license for the ducati firmware... no idea what will be in the license15:41
ndecok. so we will rename our pvr-omap-dkms to make yours, and push into the TI PPA. 15:42
ndecdo you have any patch in the kernel module that we need to take as well?15:42
rsalvetindec: great, that will also help when pushing the package to the ubuntu archive 15:42
rsalvetias the name will match with the one we already have for omap 315:42
ndecwhich archive?15:42
rsalvetindec: universe/multiverse15:43
ndecoh, you are planning to push that in universe?15:43
rsalvetior are you planning to keep it at your overlay for the rest of your life? :-)15:43
rsalvetiquestion is, why not?15:43
rsalvetiyou could still release newer versions at your own PPA15:43
ndecmaybe because it's useless without the corresponding user space bits15:43
rsalvetiwell, current package is still useful 15:44
rsalvetiso unless xorg changes it's abi, it should be fine15:44
rsalvetias the kernel part is open15:44
ndecok. i am fine with that anyways... 15:44
rsalvetindec: I have a few kernel patches at that package, but don't know if they are required for your new version15:45
ndecbtw, you gave andy some patch in the kernel that change the pvr module name, and we have to revert those, because they don't work with latest user space15:45
rsalvetithat depends on how well supported that is with latest kernel15:45
tgall_fooon arm about how many instructions does it take to stack a frame, call into a function,  execute one if test and then return ?   15:45
rsalvetindec: sure, once user space is available, just revert it15:45
looldmart: Hey, I think we had discussed exposing information about the CPU caches in the kernel, but I can't google it; would you have any pointer about this work?15:46
lool(exposing it to userspace that is)15:46
ndeclool: what kind of exposition? was talking about exactly this with sebjan few minutes ago...15:46
loolndec: details about size of the caches, layout or things like that15:46
ndeci want that too!15:47
loolndec: ISTR it was mostly about oprofile, valgrind, perf kind of tools when we discussed it in Linaro, but others came up with more use cases to expose it to any userspace15:47
loolthe first use case was reporting (and emulating properly) cache hits/misses15:47
loolthe second use case is calibrating your code for the cache size15:48
lool(e.g. use this or that implementation depending on the size of the cache)15:48
ndecmy main use case is to easily get the full info on how the caches are configured15:48
loolndec: could you share what you would use it for?15:49
ndecfirst to make sure they are configured properly...15:50
loolndec: This could be just dmesg then15:50
ndecbut we have tools that are checking all kind of configurations (clock, power, ... ). e.g. like a runtime audit of the platform. and if there was a way to retrieve the cache config, we could have that audited as well15:52
loolmakes sense15:52
davidgilukcache line size could be useful to know for optimisation at run time as well15:54
dmartlool: At present, the kernel doesn't expose that info in any systematic way15:57
dmartI don't remember the mail thread(s) now15:57
dmartIf we wanted to expose this, it would probably have to be exported via sysfs or similar -- but it may get complex (I believe this is one reason why the simple cache size in /proc/cpuinfo was abandoned)15:58
dmartAlso, I'm not sure whether the kernel even systematically knows all the cache geometry on recent CPUs15:59
kikodmart, hmm, wasn't there work done by the pmwg to expose that?16:02
davidgilukdmart: Is there a reason it couldn't expose it in exactly the same way as x86 - I can see that my laptop has a /sys/devices/system/cpu/cpu0/cache with loads of details (not sure if that's the only place)16:03
hrwlool: dmesg is not a tool to check things. it is too easy to fill that buffer16:07
loolhrw: I agree16:08
loolhrw: it's good enough for humans though, which is how I understood ndec's need at first, until he clarified usage of specialized checking tools16:09
ndecglad to see that you believe i am a human... ;-016:09
looldmart: Thanks; ISTR VDSO had come up in the discussions, hmm16:09
loolkiko: PMWG you say?  hmm maybe16:10
jstultz_vmanmar:  just android common patches on top of the linaro tree.16:10
loolamitk: Would you remember this?  (exposing CPU cache sizes to userspace in some way)16:10
wookeywhat's the right way to pin a PPA with higher priority than normal repo?16:14
wookeyapt-prefs can use a= or o=, but neither of those are set for this PPA16:15
ndecwookey: Package: *16:15
ndecPin: release o=LP-PPA-linaro-maintainers-overlay16:15
ndecPin-Priority: 10016:15
ndece.g. https://help.ubuntu.com/community/PinningHowto#Pinning_the_ubuntu-mozilla-daily_PPA16:15
hrwhave a nice rest of day16:15
wookeyyes, I found that but wonder what that string actually referred to16:16
wookeyapt-cache policy says: 500 http://ppa.launchpad.net/linaro-foundations/cross-alip/ubuntu/ natty/main armel Packages origin ppa.launchpad.net16:16
wookey'real repos have:  500 http://ports.ubuntu.com/ natty/universe armel Packages16:16
wookey     release v=11.04,o=Ubuntu,a=natty,n=natty,l=Ubuntu,c=universe16:16
ndeci would say LP-PPA-<team name>-<ppa name>16:17
wookeyah, OK.16:17
ndecso LP-PPA-linaro-foundations-cross-alip in your case16:17
wookeythis leads on to a second question of why apt-cache policy foo only shows the version in main repo, whilst apt-cache showsrc shows both main and PPA versions. BUt I'll try some pinning and see what happens16:19
loolwookey: this is from teh Release file16:22
SteveMcIntyrelool: mail looks ok :-)16:31
loolSteveMcIntyre: thanks, I wanted to ping you and forgot, great that you saw the email  :)16:34
asac_tgall_foo: good work on live-build ... is there a nice blueprint with headline/acceptance for that work this cycle16:40
tgall_fooasac_, yes ... just a second I'll dig it up16:40
tgall_fooasac_, https://blueprints.launchpad.net/linaro-ubuntu/+spec/live-build-linaro-media-create-integration16:41
* tgall_foo looks at the headline and decides to revise16:41
asac_tgall_foo: for medium and low you can revise acceptance and headline as it suites you :)16:41
asac_even if you noticed you want to deliver a beta prototype ... feel free to rename the blueprint to change scope like that16:42
asac_but sync with your PM and TL obviously before :)16:42
asac_tgall_foo: did you get initial feedback from upstream on this yet?16:42
anmarjstultz_vm: thanks16:43
tgall_fooasac_, not with linaro-media-create ...  I've got the patch generated I'm going to put onto the list, might be a long shot but I'm sure it'll generate some interesting discussion16:44
wookeyso I've set Pin: release o=LP-PPA-linaro-foundations-cross-alip in /etc/apt/preferences.d/crossppa-priority17:00
wookeybut it's not taking any notice.17:01
wookeyI notice that apt doesn't download the Release file (or at least keep it) unless the key has been installed with apt-key - never noticed that before17:01
wookeyapt-cache policy still does list an o= or a= for the PPA. should it?17:02
wookeythere isn;t some magic about preferences.f filenames is there - like sources.d files have to end in .list? 17:03
wookeyhmm, the cponfig file is defaintely being read, because if I pin a specific package then it gets listed at the end of apt-cache policy17:06
birisAmaranthus: ping got a question for the 1109 release18:53
birisAmaranthus JesseBarker: so in the meeting last week we talked releasing unity rebased to a month old branch, in order to avoid the issues from the latest hailstorm of oneiric-side changes for Unity code. In the release notes https://launchpad.net/unity-gles/2011.11/2011.09 Amaranthus mentioned using the latest Oneiric code for Unity .. Just wanted to make sure I understand - so basically the issues reported last week were overcome in the end?18:57
czajkowskimwhudson: joey http://fosdem.org/2012/call_for_devrooms19:57
amitklool: no, we haven't done any work around exposing cache info to userspace20:02
=== michaelh1|away is now known as michaelh1
dokomichaelh1, there are two more gcc/gas issues from the armel rebuild, didn't look at these further20:23
michaelh1"Error: selected processor does not support Thumb mode `rsc lr,lr,lr'" - an error with a bad error message which actually means "lr is not a valid operand"20:25
dokohrm, and I can't find the second one20:26
dokobut it was something similiar20:26
michaelh1I remember a similar one involving lr from a month or two ago.  Something regarding constraints.20:26
michaelh1Do these have a LP bug?  I can verrify it now but does it normally go through the arm-porting-jam queue?20:27
dokomichaelh1, bug 86088220:31
ubot2Launchpad bug 860882 in flint "flint ftbfs on armel (assembler errors)" [Medium,Confirmed] https://launchpad.net/bugs/86088220:31
michaelh1doko: ta.  I'll triage it.20:32
dokomichaelh1, the good news is no new ICEs ;-)20:32
michaelh1doko: it's an inline assembly problem.20:38
michaelh1doko: longlong.h contains the rsc and probably doesn't have the right constraints.20:39
dokomichaelh1, so nothing urgent20:39
michaelh1doko: there are C fallbacks but no obvious configure way to enable them/disable the asm versions.20:40
dokomichaelh1, the other one is 86089320:46
dokobug 86089320:46
ubot2Launchpad bug 860893 in libgii "libgii ftbfs on armel (assembler errors)" [High,Confirmed] https://launchpad.net/bugs/86089320:46
michaelh1doko: that's a swp instruction which doesn't exist in Thumb-2 mode.  Smells of inline assembly that needs to be changed to ldrex or __sync_foo()20:47
dokomichaelh1, yes inline ...20:48
mru__sync_foo is preferable if a suitable exists in your toolchain20:49
tgall_foohmm  is there some magic for getting the asm/* headers for user land builds?   I have linux-libc-dev and linux-headers-2.6 in build deps ...  would have thought htat would cover it  21:12
tgall_foodpkg-query -S doesn't seem to disagree21:13
tgall_foobut building in pbuilder it's missing21:13
michaelh1tgall_foo: some packages like gcc don't understand multiarch and look in /usr/include/asm instead of the new /usr/include/arm-linux-gnueabi/asm directory.21:14
michaelh1x86 has gcc-multiarch that puts in a symlink.  ARM doesn't.  These packages are rare though.21:15
tgall_foommm good point21:15
michaelh1What's the package?21:15
tgall_foomichaelh1, one that I'm building!21:15
michaelh1Go on.... :)21:15
tgall_fooit's adb which is one of the tools for android21:15
* michaelh1 he says unhelpfully21:18
michaelh1You could try a pbuilder --login, put in the symlink, try building from there, and see if it fixes the problem.  It should probably be fixed in adb though.21:19
tgall_fooactually that was spot on michaelh1, thank you21:28
michaelh1tgall_foo: adb probably has /usr/include hard coded in it somewhere.21:30
michaelh1tgall_foo: ask doko or slangasek about how to extract this from the host gcc instead.21:31
dokomichaelh1, armel does have gcc-multiarch too21:32
michaelh1doko: does that have the /usr/include/asm symlink?21:33
tgall_fooyeah the android source would do silly things like that21:33
dokotgall_foo, so either b-d on that, or make sure that /usr/include/$(gcc --print-multiarch) is searched21:33
michaelh1doko: what's the 'correct' way of finding the path to .../asm at compile time21:33
dokomichaelh1, ^^^21:33
dokofor ubuntu builds only, upstream patch is pending21:33
dokobut maybe include it into linaro?21:34
michaelh1Hmm.  -print-multiarch is unrecognised in the natty GCC.21:34
michaelh1Pfft.  No need for speechmarks around 'correct'.  Just correct.21:35
dokoyes, currently in oneiric only21:35
dokomichaelh1, well, the correct way would be to let the compiler do the work, and not second guessing ;-)21:43
pfefferzjcrigby, about https://bugs.edge.launchpad.net/linaro-android/+bug/81731521:57
ubot2Ubuntu bug 817315 in u-boot-linaro "eth0 doesn't get a valid MAC address on startup" [High,Fix released]21:57
mwhudsonzyga: around?21:59
=== michaelh1 is now known as michaelh1|away
rsalvetipfefferz: which u-boot are you using for mx53?22:08
rsalvetiU-Boot 2011.06 (Sep 27 2011 - 16:41:29)22:08
rsalvetidoesn't seems to be using the latest u-boot-linaro one22:08
zygamwhudson: sure22:12
=== michaelh1|away is now known as michaelh1
jcrigbypfefferz, I added a comment to the bug.  I just checked the git tree and it has all 5 mx patches brought forward from previous version.22:12
mwhudsonzyga: ah, i think i mostly figured out what's going on22:12
zygamwhudson: with?22:12
mwhudsonzyga: basically, the way i've been running lately is to have postgres running a snapshot of the db in a vm and lava running on my machine talking to this postgres22:13
mwhudsonzyga: this is somewhat annoying to set up22:13
jcrigbyrsalveti, the 2011.06 display may be a bug in the u-boot version print code22:14
rsalvetijcrigby: just for imx53?22:14
jcrigbyrsalveti, I'm checking22:15
zygamwhudson: aahh, nice22:15
zygamwhudson: snapshots on vm22:15
robclarku-boot, print code, incorrect?22:16
robclark(well, maybe it has been fixed more recently, but my u-boot still likes to tell me I have just 512MB or RAM ...)22:17
robclark(and that I have a 4430 when I'm using a 4460)22:17
pfefferzrsalveti, jcrigby tip master22:19
pfefferzrsalveti, jcrigby what about: FECWarning: failed to set MAC address22:20
rsalvetirobclark: the last one is just because the id line is something like printf("4430\n");22:20
zygamwhudson: I on the other hand have optimized a lot of the views22:21
jcrigbyrsalveti, ok I don't know22:21
rsalvetibut SPL now should be recognizing 4460, even es1.122:21
mwhudsonzyga: cool22:21
jcrigbywhen I build from git I get:22:21
zygamwhudson: I need to do a schema+data migration22:21
robclarkrsalveti, I figured as much..  probably the 512MB of RAM is same?22:21
jcrigbyU-Boot 2011.09-rc2-00042-g892cd2a (Sep 23 2011 - 07:18:23)22:21
zygamwhudson: to include aggregate pass/fail counts in test run mdoel22:21
robclarkI guess I should upgrade22:21
zygamwhudson: any objections or hints or how to do that?22:21
rsalvetirobclark: if you grab u-boot-linaro tip, it should work just fine22:21
rsalvetithere's a package already at the overlay ppa if you want22:22
mwhudsonzyga: sounds like a good idea in general22:22
rsalvetijcrigby: yeah, that's my point22:22
jcrigbyrsalveti, sorry for the noise22:22
rsalvetiI think the id part should be common for all boards22:22
mwhudsonzyga: do we need to do the "add the field as nullable, run a background job to fill out the field, add not null constraint" sort of dance?22:22
jcrigbyrsalveti, one would think yes22:23
rsalvetibut pfefferz keeps saying it's the tip :-)22:23
robclarkI'll build u-boot-linaro tip next time I rebuild u-boot22:23
rsalvetibut I believe that's a lie 22:23
pfefferzhere's what I tested22:23
pfefferzgit.linaro.org/boot/u-boot-linaro-stable at 1b1a2bc82bf9b8bbc1eb204ffdf0e4283036c85722:24
jcrigbypfefferz that 2011.07.122:24
jcrigbys/that/that is22:25
pfefferzjcrigby, okay22:25
zygamwhudson: I guess so, I wish I could find a way to make south --auto work on column alterations22:25
rsalvetinot tip :-)22:25
jcrigbypfefferz, is my master broken again?22:25
jcrigbylet me look22:25
jcrigbycould still be my bad22:25
jcrigbylikely is22:25
mwhudsonzyga: doesn't it?22:26
jcrigbyno such luck22:26
mwhudsoni always forget which of django's field options correspond to "not null"22:26
pfefferzjcrigby, so that master branch isn't tip?22:26
mwhudsonor rather, the difference between null and blank always trips me up22:27
jcrigbypfefferz, I believe master is tip is Linaro-u-boot-2011.0922:28
mwhudsonzyga: playing around, it looks like the queries to just fill out the data fields will take 5-10s to run, maybe it's worth 30s more downtime to avoid the fiddling?22:29
zygamwhudson: ah, perhaps it does, I tried renaming a column and that did not work without manual tweaks22:29
mwhudsonoh renaming a column, yeah, that sounds like it would need help22:29
zygamwhudson: south has db.rename_column()22:29
zygamwhudson: performance optimizations aside, django sucks somewhere22:30
zygamwhudson: oh22:30
zygamwhudson: let me show you22:30
pfefferzI think I know whats up I think we're not using u-boot from source on iMX5322:30
pfefferzI see U-Boot 2011.06 (Sep 27 2011 - 16:41:29)  22:30
pfefferzjcrigby, going to have to check that22:31
pfefferzjcrigby, that should solve a few issues22:31
jcrigbypfefferz, ok22:31
mwhudsonzyga: which bit of django is sucking today?22:31
zygamwhudson: the ORM apparently22:32
zygamwhudson: I'll explain in a second22:32
pfefferzbero|2, did you run into issues when you did the initial iMX53 work, and needed to use a precompiled bootloader?22:32
mwhudsonzyga: ok, i'm on the page22:32
zygamwhudson: ok22:32
pfefferzbero|2, I'll bet this is the issue with 3.0 not working22:32
zygamwhudson: go to bundle streams22:32
zygamwhudson: to lava-daily22:32
zygamwhudson: then switch to test runs 22:32
zygamwhudson: that query takes 17 seconds to run22:33
mwhudsonzyga: i think we need to do server side filtering :-)22:33
zygamwhudson: but... only in django, the raw sql answers instantly22:33
jcrigbypfefferz, we had that weird issue with O= on mx for awhile22:33
zygamwhudson: mapping the responseto objects takes long22:33
jcrigbyin 07 timeframe22:33
mwhudsonorms suck22:33
mwhudsonalthough 17 seconds for what, 10k objects22:33
kikoif you're grabbing too many objects.. definitely22:33
mwhudsonis pretty bad22:33
zyga    [sql] SELECT ... FROM "dashboard_app_testrun" INNER JOIN "dashboard_app_bundle" ON ("dashboard_app_testrun"."bundle_id" = "dashboard_app_bundle"."id") INNER JOIN "dashboard_app_bundlestream" ON ("dashboard_app_bundle"."bundle_stream_id" = "dashboard_app_bundlestream"."id") INNER JOIN "dashboard_app_test" ON ("dashboard_app_testrun"."test_id" = "dashboard_app_test"."id") WHERE "dashboard_app_bundle"."bundle_stream_id" = 122:34
pfefferzjcrigby, yeah...but this are working now right?22:34
zygathis is the query that I run on that page22:34
zygamwhudson: I suspect there are a lot of dupes there, I use select_related a lot22:34
bero|2pfefferz: We're building uboot from source on iMX53... And it's from normal u-boot-linaro-stable22:34
zygamwhudson: and it seems that I need to join to everything in the parten tree to make django happy22:35
jcrigbypfefferz, everything is fine these days, I was just wondering if bero went static for a while so as not be be bitten by the other problem22:35
jcrigbyand never went back22:35
bero|2pfefferz: At some point, we had to force a specific revision though (because of a build problem)22:35
mwhudsonzyga: i guess one of us needs to look into how server side filtering works with data table22:35
zygamwhudson: otherwise generating links to any of the objects re-requests that object due to missing fields/relations22:35
zygamwhudson: I started looking at that22:35
bero|2and I don't remember if I ever switched back to head, let me check...22:35
zygamwhudson: but it's just pushing the problem that a lot of the views were slow22:35
zygamwhudson: they are much faster now22:35
pfefferzbero|2, okay, I think theres a high likly hood that this is causing a few issues22:35
zygamwhudson: mind you this is on slowish sqlite22:35
mwhudsonit's not sensible to render a 8000 row table to show the user 25 rows22:36
zygamwhudson: first I wish to move to 1.3 to stop writing views as functions22:36
mwhudsonzyga: fair enough22:36
zygamwhudson: that's also true22:36
bero|2pfefferz: Just checked, it's using rev. Linaro-u-boot-2011.07.1 -- leftover from the compile failure. Not extremely old though... Let me switch to master and see if it changes anything...22:36
zygamwhudson: but just working on some views I can see that a dumb/naive code goes to O(N) while a simple analysis does O(1)22:37
zygamwhudson: where O counts the number of queries made22:37
pfefferzbero|2, ethernet should come up22:37
zygamwhudson: the amount of data is roughly the same in both cases22:37
zygamwhudson: (as I'm not removing anything)22:37
bero|2But the boot problem is almost certainly simply the kernel attempting to initialize both MX51 and MX53 bits, otherwise the messages we get out of it don't make much sense22:38
bero|2ethernet could be related to the old u-boot, yes...22:38
* bero|2 builds a new iMX53 build w/ master u-boot22:38
zygamwhudson: what do you think about how the sidebar looks like? particularly in the test run detail view22:38
pfefferzbero|2, its the fact that things are happening right at the boundary22:38
mwhudsonzyga: oh sure, it's super easy to write view code that does one (or more!) query per row of a table22:38
mwhudsonzyga: is that what you mean?22:39
* zyga goes downstairs for some tea22:39
zygamwhudson: yes22:39
pfefferzbero|2, looks like the Multimedi test is failing due to no/sdcard/mmtest/test-MP3.mp3, etc...22:39
pfefferzbero|2, let me try out when I copy all the file over22:45
* zyga is back22:47
mwhudsonzyga: i guess one approach is to write queries that just return the data that you need for the view, rather than objects22:52
=== vade_ is now known as vade
zygamwhudson: I'm still suspicious about what is actually slow, I'd need to profile this more 22:53
zygamwhudson: with .only() you can easily select one field from any model22:53
zygamwhudson: and still get the nice OO design22:53
zygamwhudson: unless the object wrapper is being _that_ slow I don't want to abandon this mode yet22:53
mwhudsonzyga: yes, i mean using things like .only()22:54
zygamwhudson: I'm already using that22:54
mwhudsonah, no i don't mean only22:55
mwhudsoni mean more like .values()22:55
zygayeah, I understand that22:55
zygathat discards parts of the ORM layers22:55
rsalveticrap, there goes another reboot23:00
=== michaelh1 is now known as michaelh1|away
pfefferzbero|2, any luck with the tip u-boot?23:07
pfefferzjcrigby, what config option do I need fo iMX53 u-boot?23:12
pfefferzjcrigby, cool, thanks23:13
bero|2pfefferz: I tried to run a full build after changing the manifest - still running... https://android-build.linaro.org/jenkins/job/berolinux_imx53-linaro-gcc-4.6/74/consoleText23:21
bero|2(I also pushed a new kernel rev. in there)23:21
pfefferzcool, all in   :)23:21
bero|2(that's a new 3.0 rev -- unlikely to boot ;) )23:22
pfefferzhave you seen arm-eabi-ld: cannot find -lgcc23:23
pfefferzmake[1]: *** [hello_world] Error 123:23
pfefferzeven after setting CROSS_COMPILE?23:23
bero|2No... When do you get that? Building u-boot manually?23:24
pfefferzI usually have no trouble23:24
pfefferzmaybe my config is gunked up23:24
pfefferzoh wow, looks like rain after 100 days of drought and 110 degF23:26
robclarkno rain up here..  although we did get some a little ways back23:28
pfefferzAustin's got some ugly clouds23:29
pfefferzman we need the rain23:29
bero|2pfefferz: I just tried... It works here if I just "git reset --hard; make CROSS_COMPILE=/..../arm-eabi- mxloco53_config; make CROSS_COMPILE=/...../arm-eabi-23:29
bero|2We aren't expecting any clouds here for 2 more weeks... But once those weeks are over, we'll probably get loads of snow23:31
pfefferzbero|2, :)23:32
pfefferzlooks like tornado weather23:32
pfefferzcool...no more badness 1000023:39
doanachold on tight pfefferz!23:42
pfefferzbero|2, which boot components are needed for iMX5323:42
bero|2Nothing special -- just uboot (u-boot.imx), the kernel and the initrd23:43
pfefferzI've got u-boot.bin23:43
pfefferzdo I need to copy it to u-boot.imx?23:43
bero|2No, u-boot.imx is somewhat different, let me check the exact command to create it (I think it's just make u-boot.imx)23:44
pfefferzyeah I see it23:44
pfefferzin Android23:44
bero|2make CROSS_COMPILE=/..../arm-eabi- u-boot.imx23:45
bero|2https://android-build.linaro.org/builds/~berolinux/imx53-linaro-gcc-4.6/#build=74 <-- just finished, has the current u-boot and the current 3.0 branch23:45
* bero|2 downloads23:45
pfefferzdrum role23:46
pfefferzbero|2, do you just replace the u-boot on the second partition?23:51
pfefferzjcrigby, I just built master23:53
pfefferzjcrigby, and I see U-Boot 2011.06 \23:53
jcrigbywhats the build date that you see?23:54
pfefferzSep 27 2011 - 16:41:2923:55
jcrigbyok back to source23:55
jcrigbywhat does git show say?23:55
pfefferz$git remote -v23:55
pfefferzI checked out remotes/origin/Linaro-u-boot-2011.0923:56
jcrigbyand git show says?23:56
pfefferzcommit 7da712c61fbdeb0fdd1b1c8c3f9005128c00437823:57
jcrigbymeanwhile I'll go try23:57
jcrigbythats right23:57
jcrigbyomap4460: tps initialization changes23:58
pfefferzwhen I build remotes/origin/Linaro-u-boot-2011.0923:58
pfefferzI still see, U-Boot 2011.06 (Sep 27 2011 - 16:41:29)23:58
dzinrsalveti: ping23:58
rsalvetidzin: pong23:59
=== michaelh1|away is now known as michaelh1
jcrigbypfefferz, I'm going to try it here23:59
dzinfor the native gcc-linaro package is the headline: "Headline: Native Gcc-Linaro packages are now available at the Overlay PPA. This package is basically gcc-linaro vanilla with minor modifications to make it multi-arch compatible." correct?23:59

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