fury | leon-anavi: snapshot of raspberry pi 3 that i downloaded from 5-29 boots. unfortunately i set my machine to download 5-30, 5-31, 6-01, and 6-02 and the files were deleted before i finished downloading them :P so checking out 6-03 now | 00:03 |
---|---|---|
fury | downloading now * | 00:04 |
*** chbae has joined #automotive | 00:54 | |
fury | leon-anavi: snapshot build from 06-03 does not boot. so the change that broke it was somewhere between 05-29 and 06-03 | 01:02 |
khem | which release of OE/yocto are you guys using | 01:04 |
fury | morty i think | 01:04 |
fury | i may be wrong | 01:05 |
smurray | yeah, it's morty | 01:10 |
fury | leon-anavi: the uImage changed somehow and caused it to break. I downloaded the uImage from here: https://download.automotivelinux.org/AGL/release/dab/3.99.1/raspberrypi3/deploy/images/raspberrypi3/ - and plopped it down in place of the uImage that came on the snapshot from 06-03 - it works | 01:31 |
fury | actually, wait, let me double check | 01:31 |
fury | i may have mixed up my cards >.> | 01:31 |
fury | yep, used the one that was already working lol | 01:33 |
fury | ._. | 01:33 |
fury | sorry, false alarm | 01:33 |
fury | copied the whole boot partitoin and it worked tho | 01:39 |
khem | OK there was a kernel update that went into to morty | 01:39 |
khem | https://github.com/kraj/meta-raspberrypi/commit/6baa6641eae89f4f4d10189a4b53eb7ede7acb4c | 01:39 |
fury | looks like the one that's broken (06-03) is missing device_tree_address and has a different kernel | 01:39 |
fury | that kernel update might be it | 01:39 |
khem | are you failing with kernel 4.9.27 ? | 01:40 |
khem | and ok with 4.9.23 ? | 01:40 |
fury | looking at it booting up now, it's 4.4.50 | 01:40 |
fury | it's using the 4.4 branch | 01:41 |
khem | ok then nothing changed there for 4.4 | 01:41 |
fury | even so, i think it tells it which specific commit to use, if i'm not mistaken | 01:41 |
khem | OK so it could be AGLs own doing then | 01:42 |
khem | I use 4.9 and its all well with morty/pyro/master | 01:43 |
khem | as of today | 01:43 |
fury | is there a way to "reverse" a particular gerrit change? | 01:51 |
fury | to see if i can reverse one and see if that fixes it? | 01:51 |
fury | i suspect https://gerrit.automotivelinux.org/gerrit/#/c/9593/ may have made it build the kernel image differently and that makes it not like it, so if there's a quick command to reverse it i'll try that, or i could just manually apply the change | 01:53 |
mranostay | fury: er git revert? | 02:15 |
fury | i manually edited the file. forgot about revert :D tho doesn't that create a new commit? | 02:18 |
fury | i am running a new bitbake with the original defaultune | 02:19 |
fury | pretty sure it has to rebuild half the stuff as a result, and i'm on my laptop not my beastly build machine, so i'm gonna go hunt down some dinner :D bbl | 02:20 |
mranostay | does that matter? | 02:20 |
fury | not sure | 02:20 |
mranostay | unless you are pushing it | 02:20 |
mranostay | this is git not svn | 02:20 |
fury | i need to learn git command lines more | 02:20 |
fury | cause then to discard my commits i have to reset --hard or something | 02:24 |
fury | next time i have a printer i'm going to grab a git cheat sheet :D | 02:24 |
*** akilaw has joined #automotive | 02:36 | |
smurray | you can revert to test, then drop it with a repo sync or a manual git reset | 02:55 |
leon-anavi | fury, no problem for the false alarm :) thanks for your help. | 03:54 |
fury | no worries. i am waiting for the bitbake to finish - hours and hours it seems :D but i did narrow it down to something in the boot partition | 03:55 |
fury | because when i replace the boot partition with the earlier one, while keeping the rootfs, it works fine | 03:55 |
fury | built master just before this and verified that | 03:56 |
*** akilaw has quit IRC | 03:56 | |
fury | so, checking to see if that defaulttune caused it. but shoulda SSH'd into my build machine, it seems...i prolly could download the image faster if i built it there and downloaded it :P | 03:56 |
fury | especially with this laptop crashing when building qt O_o | 03:57 |
leon-anavi | for Poky (master and morty) the trick to enable serial console output is to add: ENABLE_UART = "1" | 03:58 |
leon-anavi | It is a rpi3 specific setting. It works like a charm in Poky (morty and master) for rpi3. I didn't see it in AGL but when I added it didn't work so I am investigating further. | 03:59 |
fury | whoops, lightmediascanner.h not found | 03:59 |
leon-anavi | lightmediascanner.h? | 04:00 |
leon-anavi | I noticed some changes to rygel and lightmediascanner but I don't expect a header file to be removed. | 04:00 |
fury | bitbake failed at media player, missing lightmediascanner.h, guessing a missing depend | 04:00 |
leon-anavi | are you sure there is such a problem? | 04:00 |
mranostay | uh oh | 04:00 |
fury | bit baked lightmediascanner and things are moving along again | 04:01 |
leon-anavi | fury, hm, is lightmediascanner in its depends? | 04:01 |
mranostay | possibly not | 04:02 |
fury | dunno...im afraid to switch over to another console to check, I logged into a text terminal to build AGL to try to avoid problems where the desktop environment crashes and kills the bitbake process | 04:02 |
fury | whenever it’s building Qt it crashes the desktop, lol | 04:03 |
mranostay | nope it is isn't | 04:03 |
fury | whoops | 04:03 |
mranostay | leon-anavi: yeah issue is is i never ran into that race condition when building | 04:03 |
mranostay | fury: i'll fix it but do bitbake lightmediascanner first *g* | 04:03 |
fury | I wonder how many depends issues there are like that | 04:04 |
fury | Yeah, I kinda guessed that and the bitbake is running happily now | 04:04 |
leon-anavi | mranostay, lightmediascanner and rygel were my first contributions to AGL. Be careful with them. | 04:04 |
fury | tho I did bitbake -c build lightmediascanner not sure if there’s a diff | 04:04 |
mranostay | leon-anavi: well rygel is disabled *g* | 04:05 |
mranostay | it is in the build but you need to manually enable it now | 04:05 |
mranostay | with systemctl --user | 04:05 |
leon-anavi | Couple of years ago Manuel Bachmann was integrating media manager and invited me to contribute my recipes for rygel and lightmediascanner which I had already fixed in Tizen. | 04:05 |
mranostay | leon-anavi: lms is there but fury seems to ran into a build issue with mediaplayer app being built before lms is | 04:06 |
leon-anavi | mranostay, right, I noticed this rygel change in the git logs. | 04:06 |
fury | Rygel likes to send SSDP requests all over the network, that was making IT unhappy...what does it do other than that? | 04:06 |
mranostay | fury: who knows who is using it.. hence mdp telling be to disable it starting than removing outright | 04:06 |
leon-anavi | fury, can you report your issue in JIRA? I will be happy to have a look and fix it. | 04:07 |
mranostay | leon-anavi: the rygel has one | 04:07 |
leon-anavi | My guess is that you just need lightmediascanner in DEPENDS of mediaplayer. | 04:07 |
fury | leon-anavi: for the rygel thing is https://jira.automotivelinux.org/browse/SPEC-613 | 04:08 |
mranostay | problem is you could not what it | 04:08 |
mranostay | lms that is | 04:08 |
mranostay | which is weird since the the checks should detect that | 04:08 |
fury | hadn’t encountered the lms build issue before now, does that need a JIRA issue too or naw? | 04:09 |
mranostay | fury: can't hurt i'll close if not an issue | 04:09 |
mranostay | i'd need a JIRA to submit a patch to anyway | 04:10 |
fury | ah. all patches need a corresponding JIRA issue these days? | 04:10 |
mranostay | always have in theory | 04:11 |
fury | theories :P | 04:11 |
leon-anavi | fury, report anything :) | 04:11 |
leon-anavi | by reporting it is becoming a known bug. It will be assigned to someone and it easier to fix it (sooner or later) :) | 04:12 |
mranostay | leon-anavi: that is asking for [SPEC-700] was the moon landing faked? :P | 04:12 |
fury | I like the traceability of an issue for every corresponding patch, but also it means a lot of work just to submit the work. Then again, been bit by that kind of “eh, it doesn’t need an issue, just commit and ship it” thing at work occasionally. I need to take the time to finish migrating to gitlab | 04:13 |
fury | once this bitbake finishes or crashes again I’ll hop on and submit the log file | 04:14 |
mranostay | fury: hey at least you don't have to enter them into ClearCase :) | 04:15 |
leon-anavi | mranostay, in this case SPEC-700 should be assigned to politics. Is there an issue for chemtrail? The .bg president has already promised that he will investigate it ;) | 04:15 |
mranostay | leon-anavi: had to google that... always good to know the balkans is as or more insane :) | 04:16 |
fury | I installed Redmine a while back, with some Apache integration for git so that Redmine credentials could pull/push to git...but somehow, that system broke down and I can no longer make new repos. Gitlab got me drooling over their full chat/issue/code/CI/CD stack demos, but the demos make it look WAY easier than it actually is, I’m having a hard time just installing it all :P | 04:16 |
fury | I’m almost afraid to google clear case...is it better or worse than emailing bug logs and change logs around as excel spreadsheets because excel hates opening up files from the network? | 04:22 |
mdp | mranostay: IMHO, all the unused components should be removed. that includes the agl-audio-plugin. Get an in-tree user or go away. | 04:29 |
mranostay | well lms is used but with the binding now | 04:30 |
mranostay | which i think dl9pf merged today | 04:30 |
fury | agl-audio-plugin? would I need that if I’m planning on adding a custom audio thingamajiggybobber? | 04:30 |
mdp | No, because we only load and use module-router which talks to genivi audiomanager and the AGL specific generic xml controller. | 04:32 |
mdp | mranostay: I was referring to rygel only. | 04:32 |
*** akilaw has joined #automotive | 04:43 | |
fury | mranostay: https://jira.automotivelinux.org/browse/SPEC-655 | 04:48 |
fury | whalep, just flashed the SD card, it's not the defaulttune that broke the pi | 04:50 |
mranostay | noted | 04:50 |
fury | i suspect it's the missing device_tree_address | 04:50 |
mranostay | fury: will look at tomorrow... bit spent atm | 04:59 |
fury | mranostay: no worries, didn't mean go fix it right now :P | 05:01 |
fury | just wanted to log it before i forgot | 05:02 |
fury | and thinking out loud bout the pi thing | 05:02 |
fury | i think i may've whittled it down to 3 binary differences between the working boot partition and the not working one | 05:03 |
fury | pretty sure it's kernel7.img | 05:10 |
*** User3209 has quit IRC | 05:18 | |
fury | yep, all i gotta do is throw it an older kernel7.img and it works fine | 05:19 |
fury | leon-anavi: any idea what could possibly change kernel7.img and break it? | 05:19 |
*** jobol has joined #automotive | 05:30 | |
leon-anavi | not sure | 05:30 |
leon-anavi | I have to go out. It's running time for me. See you later. | 05:31 |
fury | have a good one :D | 05:31 |
*** leon-anavi has quit IRC | 05:34 | |
*** araujo has quit IRC | 05:34 | |
*** voltbit has joined #automotive | 05:52 | |
*** jobol has quit IRC | 05:52 | |
akilaw | hi | 06:01 |
akilaw | Can you please explain core2-64 package inside the genivi-dev-platform/gdp-src-build/tmp/deploy/ipk path | 06:01 |
*** chbae has quit IRC | 06:04 | |
*** egy has joined #automotive | 08:08 | |
*** egy has quit IRC | 08:17 | |
*** egy has joined #automotive | 08:17 | |
*** akilaw has quit IRC | 08:49 | |
*** jlrmagnus has joined #automotive | 09:03 | |
*** egy has quit IRC | 10:20 | |
*** jlrmagnus has quit IRC | 10:43 | |
*** voltbit has quit IRC | 10:58 | |
*** aw- has joined #automotive | 11:51 | |
*** john3voltas has joined #automotive | 14:28 | |
*** aw- has quit IRC | 14:46 | |
*** AlisonChaiken has joined #automotive | 14:50 | |
khem | albanc: those are x86_64 arch packages for minnow platform | 15:14 |
khem | err it was meant for akilaw | 15:14 |
*** voltbit has joined #automotive | 15:38 | |
*** zeenix has joined #automotive | 15:53 | |
*** kooltux__ has joined #automotive | 16:10 | |
*** zeenix has quit IRC | 16:23 | |
voltbit | hello again, this happens when I try to install chinook https://pastebin.com/wTHFNx8f | 19:21 |
*** mkelly has joined #automotive | 19:28 | |
*** _corrupt has quit IRC | 19:42 | |
voltbit | nvm | 19:50 |
*** praneeth has quit IRC | 20:23 | |
*** praneeth has joined #automotive | 20:23 | |
*** mkelly has quit IRC | 20:54 | |
*** _corrupt has joined #automotive | 22:31 | |
mranostay | fury: ok fix attached to the jira | 23:02 |
mranostay | forgot to do "git clean -xfd" and rebuild to be sure there wasn't any artifacts it seems | 23:03 |
*** john3voltas has quit IRC | 23:26 | |
*** User3209 has joined #automotive | 23:36 | |
*** User3209 has joined #automotive | 23:37 | |
*** kooltux__ has quit IRC | 23:40 | |
*** voltbit has quit IRC | 23:45 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!