**** BEGIN LOGGING AT Mon May 14 03:00:02 2018 May 14 06:57:43 Morning May 14 07:32:11 Morning! May 14 07:52:36 Tofe: Mido flashed, booting now.... May 14 07:52:39 Fingers crossed LOL May 14 07:55:33 Hmmz stuck on logo :S May 14 07:55:38 Mi logo that is May 14 07:58:08 https://bpaste.net/show/3a7fff1e0b9a May 14 07:58:13 I'll reboot, maybe just one time thing May 14 07:59:14 Seems it was May 14 07:59:16 LuneOS logo now May 14 07:59:31 Fast glowing so probably another issue but well May 14 08:08:10 Tofe: This android-tools doesn't work on my mido, i.e. the detection of the functionfs it seems May 14 08:08:32 If I just push the .sh without conditional functionfs check it works May 14 08:09:02 Could be that the patch where it checks != tenderloin doesn't work in general May 14 08:09:08 Need to double check that May 14 08:09:20 Noticed it before but had no chance to fix it yet May 14 09:05:22 Tofe: I checked my image, somehow my android-gadget-setup script doesn't get patched so that means that https://github.com/webOS-ports/meta-webos-ports/blob/pyro/meta-luneos/recipes-android/android-tools/android-tools-conf_1.0.bbappend#L5 doesn't seem to work? May 14 09:11:50 Herrie|Laptop: ah, yes, maybe May 14 09:12:31 do you have the install log of that step ? May 14 09:13:17 Tofe: You mean the do_patch for android-tools-conf? May 14 09:13:44 no, do_install May 14 09:14:17 https://bpaste.net/show/953b4a1924e0 May 14 09:14:23 Last line seems to give a clue May 14 09:14:39 [[: not found May 14 09:14:42 Yes May 14 09:14:46 ok I did a bash-ism May 14 09:15:40 Herrie|Laptop: can you try with just simple brackets ? May 14 09:17:59 Tofe: Anbox so far: https://bpaste.net/show/5a22697a26fb May 14 09:18:30 My LNC doesn't start, not sure what's up with that May 14 09:18:41 I might want to try non-OSE version May 14 09:18:52 Though not sure it's critical for this May 14 09:18:52 what gives systemctl status anbox-container-manager ? May 14 09:19:24 Herrie|Laptop: you'll need a wayland compositor a some point :p May 14 09:19:27 at* May 14 09:19:34 https://bpaste.net/show/b173dfb734e0 May 14 09:19:49 Tofe: Well the compositor is there... i.e. luna-next starts, just not the cardshell it seems May 14 09:20:24 for the container, it looks like you have to image: do you have /var/lib/anbox/android.img ? May 14 09:20:29 s/to/no/ May 14 09:21:14 you'll have to check the anbox-data recipe, it failed somewhere I guess May 14 09:21:46 but seeing the anbox system-info output is quite good :) May 14 09:22:07 Well I guess it's the aarch64 and arm64 issue May 14 09:22:12 I could just manually push the image for now? May 14 09:22:20 yes sure May 14 09:24:13 not having the cardshell will be a problem: the window won't be shown on screen May 14 09:38:40 Tofe: Ah OK May 14 09:40:03 Tofe: android-tools-conf seems to work with single [ and ]] instead of [[ and ]] May 14 09:40:31 do_install log: https://bpaste.net/show/f1ca59a66a88 May 14 09:41:02 And output ipk has the functionfs lines in there May 14 09:41:06 I'll change the recipe May 14 09:42:00 https://github.com/webOS-ports/meta-webos-ports/commit/80dc053ed14ad71da089b8bc0e343cbc7e122cad May 14 09:45:37 Herrie|Laptop: thanks, perfect May 14 09:45:58 For anbox-data any suggestions on how to debug? May 14 09:47:14 First, is your recipe like this: https://github.com/Tofee/meta-webos-ports/blob/8112b536f7fa769eaa89f61a0639f6fbc99fc6c9/meta-luneos/recipes-support/anbox/anbox-data.bb May 14 09:48:31 oops, lunch time May 14 09:48:46 Well I have arm64 instead of aarch64 May 14 10:06:14 morphis: ping May 14 10:08:11 Tofe: OK think I might have found it May 14 10:08:17 Your recipe is fine it seems May 14 10:08:23 Just me needing more coffee :P May 14 10:39:04 Tofe: This looks better: https://bpaste.net/show/fcfffebed5b8 May 14 10:39:11 Just I guess my LNC not running is now the issue May 14 10:39:17 I should build I guess without the OSE bits May 14 10:40:05 During the process I actually get the Android logo with "starting ..." on screen May 14 10:40:14 I guess that's more then you get on 3.4 Hammerhead? May 14 10:47:32 Herrie|Laptop: can you (re)start the session-manager separately ? Kill it if it's runnning, and start it using anbox session-manager May 14 10:52:16 Tofe: It tells met "[ 2018-03-22 16:00:53] [daemon.cpp:58@Run] HOME not set in environment" May 14 10:54:06 ah yes, simply do "export HOME=/tmp" that'll do :) May 14 11:07:23 Now doesn't show anything, also no error message May 14 11:07:29 So I guess that's expected due to compositor May 14 11:08:00 both anbox session-manager and the anbox launch seem to be running May 14 11:08:11 I.e. they don't return to my prompt May 14 11:08:25 So I guess I should test this on the regular image May 14 11:11:14 That's where I got too May 14 11:11:49 if you want a little bit of output, there's a variable "export ANBOX_LOG_LEVEL=debug" (approximately) May 14 11:12:58 yes, it's this one https://github.com/anbox/anbox/blob/master/src/anbox/daemon.cpp#L48 May 14 11:15:05 Herrie|Laptop: I also got the "starting..." white window, when the session-manager isn't started yet May 14 11:16:50 Tofe: where does the output go to? May 14 11:16:52 Journalctl? May 14 11:19:29 to the console May 14 11:19:47 but only session-manager output something May 14 11:51:38 Tofe: https://bpaste.net/show/9d4b22f797ec May 14 11:53:11 Herrie|Laptop: ok, that's slightly better than me, I don't have surfaceless context on hammerhead May 14 11:53:18 the rest is identical May 14 11:53:32 Tofe: OK May 14 11:54:02 I had plenty of EGL warnings in logcat, on my side May 14 11:54:21 "W/Adreno-GSL( 0): : fd 27: error 2 No such file or directory" for example May 14 12:00:35 Well I have quite some of these: Mar 22 16:25:51 mido kernel: tz_get_target_freq ADRENO jumping level = 6 last_level = 5 total=16645 busy=6436 original busy_time=6436 May 14 12:00:59 But also without Anbox May 14 12:48:28 that one is fine, it's just CPU gouvernor it seems May 14 13:23:40 Tofe: I guess I have no other choice then to start a new build with pyro layers:P May 14 13:23:57 Tofe: I guess it wouldn't hurt to merge this into pyro and just enable it for mido on our builds? May 14 13:24:06 It shouldn't break anything else right? May 14 13:24:12 Just the image size increases 250MB or so :P May 14 13:46:17 Tofe: I guess you could drop the squashfs bits? I mean the module? I would assume these are already in the kernel tree? May 14 13:49:18 I mean Hammerhead, Mako & Tenderloin seem to have it? https://github.com/Halium/android_kernel_lge_hammerhead/tree/halium-5.1/fs/squashfs, https://github.com/ubports/android_kernel_google_msm/tree/cm-12.1/fs/squashfs and https://github.com/webOS-ports/android_kernel_htc_msm8960/tree/tenderloin/3.4/cm-12.1/fs/squashfs May 14 13:49:36 Or is that a different version? We should be able to suffice with just the defconfig changes? May 14 13:49:49 That's for this commit: https://github.com/Tofee/meta-webos-ports/commit/958a61668a6281a085a3a5f4ccceb322f19b5147 May 14 14:08:20 Herrie|Laptop: yes, I'll drop it May 14 14:13:36 Tofe: You're going to PR the other 2 libhybris patches as well? May 14 14:30:54 Herrie|Laptop: not sure, I've put them here because it was just taking dust in the folder May 14 14:31:07 Tofe: Ah ok May 14 14:31:38 But the test_camera "fix" may be not wanted by some people; it's not really a fix but it changes the binary into a waylant client May 14 14:32:00 s/t/d/ May 14 14:32:30 the printf fix is very empyrical, and it may just stop fixing the crash at any moment May 14 14:33:04 so for us it's an improvement, but it's either specific or fragile May 14 14:33:22 OK May 14 16:33:37 Herrie: pong May 14 16:43:36 morphis: interested in a couple of Anbox questions ? :) May 14 17:18:52 morphis: Specifically: We now got anbox in a bitbake recipe, so dropped the whole snappy stuff May 14 17:19:08 Do you know what causes the 3.10+ kernel requirement? May 14 19:59:41 Tofe: New build based on pyro with Anbox for mido is running locally May 14 20:01:52 Let's see if I can do something more with that one May 14 20:59:56 * elvispre can now confirm that webOS OSE runs on a Raspberry Pi 3 Model B but not on a Raspberry Pi 3 Model B+ May 14 21:01:29 elvispre: Well that doesn't help :P May 14 21:01:37 I mean that it doesn't run on B+ May 14 21:02:23 Herrie: It didn't help much that the first time I ordered the B they sent me another B+ :P May 14 21:02:29 Hehe May 14 21:02:40 Well the technical differences between the 2 aren't big May 14 21:02:46 Shouldn't be too hard to get it working May 14 21:02:54 It probably just needs another kernel May 14 21:03:08 I guess not. The differences are probably interesting. **** ENDING LOGGING AT Tue May 15 03:00:02 2018