**** BEGIN LOGGING AT Tue Jan 16 03:00:03 2018
Jan 16 06:26:13 <Herrie> Morning!
Jan 16 06:26:19 <Herrie> Tofe: On my Mako: Seems OK? https://bpaste.net/show/4071fce94658
Jan 16 07:57:56 <Tofe> Morning!
Jan 16 07:59:24 <Tofe> Herrie: well, adb debug does work :)
Jan 16 07:59:41 <Tofe> Just make sure it isn't enabled by default ;)
Jan 16 08:20:21 <Herrie|Laptop> Tofe: Yeah, I need to change that. I already tested the fastboot myimage.fastboot -c enable_adb and that works as expected :P
Jan 16 08:20:27 <Herrie|Laptop> Let me propose a PR
Jan 16 08:20:46 <Herrie|Laptop> This is using FunctionFS for N4, N5 and Mido.
Jan 16 08:20:57 <Herrie|Laptop> Just seems for TP we don't do adb at all in init
Jan 16 08:25:26 <Herrie|Laptop> I guess it might be good if we could unify the init.sh somehow
Jan 16 08:25:37 <Herrie|Laptop> Because the TP one is quite different from the others due to TP specific bits
Jan 16 08:29:32 <Tofe> But is FunctionFS required at all for adb to work?
Jan 16 08:31:05 <Herrie|Laptop> Tofe: It seems it replaced the legacy usb bits used for ADB from Android 5
Jan 16 08:31:17 <Herrie|Laptop> Seems that in newer Android Google killed the legacy bits completely
Jan 16 08:31:24 <Herrie|Laptop> At least if I understood it correctly
Jan 16 08:31:32 <Herrie|Laptop> So ADB now runs over FunctionFS
Jan 16 08:32:15 <Herrie|Laptop> Tofe: See https://android-review.googlesource.com/c/platform/system/core/+/31640 for some more background
Jan 16 08:32:34 <Herrie|Laptop> Seems FunctionFS is something more generic while ADB previously used custom USB gadget in kernel
Jan 16 08:33:09 <Tofe> ah ok, I understand
Jan 16 08:33:50 <Herrie|Laptop> I'm in favor of something more generic then custom bits ;)
Jan 16 08:33:58 <Herrie|Laptop> And as long as it works it should be fine
Jan 16 08:34:19 <Herrie|Laptop> Since you said previously you had issues with getting ADB to work?
Jan 16 08:34:35 <Herrie|Laptop> Which might be due to outdated Android tools, this or both ;)
Jan 16 08:37:15 <Tofe> yes, maybe both
Jan 16 08:38:07 <Herrie|Laptop> The standard android-gadget-setup needed some work too
Jan 16 08:38:17 <Herrie|Laptop> The way AsteroidOS guys are doing it is a bit weird
Jan 16 08:39:30 <Herrie|Laptop> I.e. They first set the values, then call the android-gadget-setup which would overwrite most: https://github.com/AsteroidOS/meta-asteroid/blob/master/recipes-core/initrdscripts/initramfs-boot-android/init.sh#L47 and https://github.com/openembedded/meta-openembedded/blob/pyro/meta-oe/recipes-devtools/android-tools/android-tools-conf/android-gadget-setup
Jan 16 08:40:13 <Herrie|Laptop> I tested it and as long as we keep L44,L45 and L50, L51 stuff works
Jan 16 08:40:19 <Herrie|Laptop> Without 50 and 51 adb doesn't work
Jan 16 08:49:05 <Tofe> Herrie|Laptop: I'd let's just keep the necessary stuff
Jan 16 08:49:44 <Tofe> i.e. no call to android-gadget-setup if it's useless
Jan 16 08:52:05 <Herrie|Laptop> Tofe: It's not useless for us, just not complete to get it working for FunctionFS
Jan 16 08:52:21 <Herrie|Laptop> I.e. Asteroid puts too many lines imho since android-gadget-setup overwrites them
Jan 16 08:52:27 <Herrie|Laptop> So I'll reduce it to a minimum
Jan 16 08:52:32 <Herrie|Laptop> So things work ;)
Jan 16 09:25:42 <Herrie|Laptop> Tofe: Is there by the way any preference over having telnet instead of adb or they can in general do the same?
Jan 16 09:25:49 <Herrie|Laptop> Just wondering why Halium guys have telnet
Jan 16 09:26:43 <Herrie|Laptop> Seems that adbd requires a lot less config
Jan 16 09:54:51 <Tofe> well, telnet can be accessed over wifi, not sure if adbd can do that
Jan 16 09:55:55 <Tofe> and maybe telnet is better managed by linux (i.e. as a terminal) then adb
Jan 16 09:55:58 <Tofe> than*
Jan 16 09:56:25 <Tofe> I'd say let's keep adb for now
Jan 16 09:56:35 <Herrie|Laptop> Tofe: Ah OK
Jan 16 09:56:36 <Tofe> as you say, it just works
Jan 16 09:56:43 <Herrie|Laptop> Yeah via wifi would be nice
Jan 16 09:56:51 <Herrie|Laptop> But then again, if you're devving you're fine with cable :P
Jan 16 09:57:26 <Tofe> Yes... it's even better, because you often don't have wifi anyway in initramfs
Jan 16 09:57:53 <Herrie|Laptop> I guess it helps we have yocto and can easily get stuff like android-tools in
Jan 16 09:57:55 <Tofe> and if you have adb in initramfs, then why not use it for the rest
Jan 16 09:58:02 <Herrie|Laptop> Having busybox only I guess telnetd is easier ;)
Jan 16 09:58:10 <Tofe> yup
Jan 16 09:58:53 <Tofe> Though they build their initrd via the Android toolchain, so including adb should be easy
Jan 16 10:00:13 <Herrie|Laptop> Well there are quite some patches in the android-tools for adb ;)
Jan 16 10:00:19 <Herrie|Laptop> They would need to modify bits too I guess
Jan 16 10:00:53 <Herrie|Laptop> I count 9 ;) https://github.com/openembedded/meta-openembedded/tree/pyro/meta-oe/recipes-devtools/android-tools/android-tools
Jan 16 10:01:07 <Herrie|Laptop> Not sure they would be all strictly required but they must be there for a reason
Jan 16 10:12:19 <Herrie|Laptop> Tofe: Any clues for my mido? It seems to switch to rootfs but then get stuck in systemd somewhere?
Jan 16 10:19:46 <Tofe> Herrie|Laptop: no, it's a bit mysterious here; I quite don't know why systemd would stop like that
Jan 16 10:20:09 <Herrie|Laptop> udev rules?
Jan 16 10:20:12 <Herrie|Laptop> I didn't update those yet
Jan 16 10:20:16 <Tofe> Maybe
Jan 16 10:21:04 <Tofe> I suggest you create the /var/log/journal directory from a recovery adb shell, reboot, and fetch the journal after a failed boot
Jan 16 10:21:44 <Tofe> or, you can try to see which service was started but never ended
Jan 16 10:23:27 <Herrie|Laptop> Tofe with recovery adb shell you mean TWRP?
Jan 16 10:23:31 <Herrie|Laptop> Then just adb in?
Jan 16 10:23:35 <Herrie|Laptop> Seems I don't even have var there
Jan 16 10:24:11 <Herrie|Laptop> simple mkdir -p /var/log/journal right?
Jan 16 10:33:16 <Herrie|Laptop> Hmmz that doesn't seem to work
Jan 16 10:33:33 <Herrie|Laptop> After it fails and I reboot back into twrp the /var/log/jounral isn't there
Jan 16 10:43:05 <Tofe> Yes I mean twrp
Jan 16 10:43:49 <Tofe> Ah sorry, more details is: you'll need to mount /data, and actually create /data/luneos/var/log/journal
Jan 16 10:44:05 <Tofe> (I think that where luneos is flash, iirc)
Jan 16 10:44:09 <Tofe> that's*
Jan 16 10:49:11 <Herrie|Laptop> That doesn't work because log is symlinked to volatile/log it seems?
Jan 16 10:49:20 <Herrie|Laptop> "No such file or directory"
Jan 16 12:00:28 <Herrie|Laptop> Tofe: Any objections/reservations on including https://github.com/shr-distribution/meta-smartphone/blob/cf30493ac4fab9058815bf9b5a72ab528b55b4aa/meta-android/recipes-core/initrdscripts/initramfs-boot-android/init.sh#L102 to L105 in our init.sh?
Jan 16 12:17:44 <Tofe> Herrie|Laptop: you'll need to remove the symlink first
Jan 16 12:19:37 <Tofe> Herrie|Laptop: the fallback looks like a good idea
Jan 16 12:19:47 <Herrie|Laptop> Tofe: That's what I thought as well
Jan 16 12:20:03 <Tofe> not sure it'll boot Android with our kernel, but it could very well be the case
Jan 16 12:20:06 <Herrie|Laptop> It's in the meta-smartphone one, there are some other useful variables that will allow for non mmcblk to work too
Jan 16 12:20:19 <Herrie|Laptop> So I'll take those too since I'm anyway on it
Jan 16 12:24:17 <Herrie|Laptop> Tofe: journal should be file or folder?
Jan 16 12:24:29 <Herrie|Laptop> Because I created folder and nothing gets placed in there it seems
Jan 16 12:49:52 <Tofe> Should be a directory
Jan 16 12:52:27 <Herrie|Laptop> Well nothing gets put in there it seems :S
Jan 16 12:52:44 <Herrie|Laptop> I have /data/luneos/var/log/journal/
Jan 16 12:54:34 <Tofe> or maybe our init mount something on top of it
Jan 16 12:58:33 <Tofe> https://github.com/Tofee/meta-smartphone-1/blob/pyro/meta-lg/recipes-core/base-files/base-files/hammerhead/fstab#L6
Jan 16 12:58:50 <Tofe> ah no sorry, it's not that
Jan 16 13:05:05 <Herrie|Laptop> Tofe: My proposal :) https://github.com/webOS-ports/meta-webos-ports/compare/herrie/initfix?expand=1
Jan 16 13:12:56 <Tofe> It does look good!
Jan 16 13:13:24 <Herrie|Laptop> Tofe: Let me do a final check
Jan 16 13:13:44 <Herrie|Laptop> I didn't test this sdcard_device but should work I'd say
Jan 16 13:43:53 <Herrie|Laptop> Tofe: OK seems to work :)
Jan 16 13:44:12 <Tofe> grat
Jan 16 13:44:13 <Tofe> great
Jan 16 13:45:57 <Herrie|Laptop> Just adb closes sometimes but it worked fine yesterday and I didn't change anything
Jan 16 13:46:00 <Herrie|Laptop> So could be just my pc
Jan 16 13:46:04 <Herrie|Laptop> That needs a reboot
Jan 16 13:46:21 <Herrie|Laptop> I.e. it lists the device but tells me "error: closed"
Jan 16 14:40:30 <Tofe> merged !
Jan 16 14:40:58 <Herrie|Laptop> Tofe: Got a new mido image awaiting at home where I removed the systemd bits for wifi and bt
Jan 16 14:41:05 <Herrie|Laptop> Maybe it'll go a bit further (or not...)
Jan 16 14:43:22 <Herrie|Laptop> Tofe: Thnx, it shouldn't hurt anyway the PR
Jan 16 14:48:07 <Tofe> Maybe
Jan 16 14:48:27 <Tofe> The journal thingy is a bit weird, I still don't see what goes in the way
Jan 16 14:48:44 <Herrie|Laptop> Anything I can do to debug?
Jan 16 14:48:52 <Herrie|Laptop> Even if it's a nasty hack?
Jan 16 15:42:16 <Tofe> I don't know yet
Jan 16 16:33:14 <Herrie|Laptop> Tofe: Found this one: https://github.com/systemd/systemd/issues/4277
Jan 16 16:48:34 <Herrie|Laptop> Tofe: modified journald.conf to set it to permanent instead of auto
Jan 16 16:48:37 <Herrie|Laptop> Maybe that will help
Jan 16 16:50:08 <Herrie|Laptop> Nope
Jan 16 18:10:39 <Herrie|TP> Tofe: Curious thing is that some of the partitions that appear in the logs I didn't specify anywhere like p12. So I'm unpacking a mido image and will grep around a bit to see where these come from
Jan 16 18:12:58 <Herrie|TP> Another thing I'll do is compare Hammerhead and Mido log to see which service might be the culprit
Jan 16 19:08:55 <Herrie> Tofe: Hmmz grep for mmcblk0p12 and mmcblk0p48 don't give anything
Jan 16 19:09:04 <Herrie> Not even when I search for p12 or p48
Jan 16 19:24:04 <Herrie> Tofe: I can only think of https://github.com/systemd/systemd/commits/master/rules/60-persistent-storage.rules where it might come from
Jan 16 19:24:13 <Herrie> Could be I need to use the lastest rules file?
Jan 16 20:07:54 <Herrie> Nope rules file doesn't help
Jan 16 20:13:13 <Herrie> Seems I don't have memnotify kernel module
Jan 16 20:13:31 <Herrie> And no kernel modules
Jan 16 20:13:38 <Herrie> Not sure that should make a difference
Jan 16 20:17:25 <Herrie> Tofe: OK some clue
Jan 16 20:17:35 <Herrie> Seems I don't have the "android-system.service" in my image somehow
Jan 16 20:25:16 <Herrie> I.e. I have wait-for-android.sh in my Hammerhead log, which doesn't appear in my mido log
Jan 16 20:25:24 <Herrie> I don't have wait-for-android.sh in image either
Jan 16 20:30:43 <Herrie> Seems that somehow android-system doesn't get installed :S
Jan 16 20:35:56 <Herrie> I don't really see where I'm doing stuff wrong
Jan 16 21:07:49 <Herrie> Tofe: OK found it I guess
Jan 16 21:08:53 <Herrie> Tofe: https://github.com/webOS-ports/meta-webos-ports/blob/pyro/meta-luneos/recipes-core/packagegroups/packagegroup-luneos-extended.bb#L145
Jan 16 21:08:58 <Herrie> I guess I need to add mido here
Jan 16 21:09:07 <Herrie> I guess we have memnotify now for Tenderloin too?
Jan 16 21:30:59 <Herrie> Now ahve some issues with qt5 qpa
Jan 16 21:31:07 <Herrie> Trying latest 5.9 commits from NotKit
Jan 16 21:41:41 <Herrie> I'll need some help it seems:P https://bpaste.net/show/9033804da5d8
**** ENDING LOGGING AT Wed Jan 17 03:00:01 2018