**** BEGIN LOGGING AT Wed Jun 27 03:00:02 2018
Jun 27 05:33:21 <bshah> Tofe: morning
Jun 27 05:35:46 <bshah> Tofe: so two things...
Jun 27 05:35:55 <bshah> halium 5 does have libdsyscall patch
Jun 27 05:36:02 <bshah> and halium 7 doesn't in raw mode
Jun 27 05:36:11 <bshah> but my branch does have it
Jun 27 05:36:24 <bshah> but neither does work.. so I guess we need to talk to krnlyng
Jun 27 05:37:11 <bshah> would be good if for example Herrie can test it on halium 7 device along with two of my branches
Jun 27 05:37:21 <bshah> and with TLS patch disabled
Jun 27 05:41:31 <Herrie> Morning!
Jun 27 05:41:49 <Herrie> bshah: Tofe got a 7 device as well now he's working on ;)
Jun 27 05:42:27 <bshah> it's still WIP though :p
Jun 27 05:50:04 <Herrie> True
Jun 27 05:50:21 <Herrie> Just we'll be moving soon so lots of things to sort, don't have much free time at hand currently
Jun 27 05:50:32 <Herrie> But please paste the links and will see what I can do
Jun 27 05:59:44 <bshah> Herrie: for starters, I think you can try upgrading to 5.11.1 and see if it rightfuly breaks the device (assuming it have TLS patch disabled)
Jun 27 06:00:02 <bshah> s/device/compositor ;)
Jun 27 06:36:28 <Tofe> Morning
Jun 27 06:38:44 <Tofe> bshah: for me it's not necessarily the compositor that break; in my case here, it was a qml app
Jun 27 06:39:07 <bshah> oh
Jun 27 06:39:29 <bshah> wait, so you mean compositor worked fine even earlier?
Jun 27 06:39:36 <Tofe> The crash always occurs in QSGRenderThread though
Jun 27 06:40:16 <Tofe> bshah: I...will have to check our logs; that's possible
Jun 27 06:44:57 <Tofe> It was also an app, it seems
Jun 27 06:45:40 <Tofe> "Jun 13 18:29:19 <Tofe> ok, some progress: I could start some Qt apps by disabling TLS in Qt"
Jun 27 06:46:33 <bshah> hm
Jun 27 06:47:22 <Tofe> However I guess it really depends on what the app/compositor does
Jun 27 06:54:49 <Tofe> bshah: I do think your sigbus also comes from the same __set_errno issue: 0x1d is an address not aligned with 4-bytes, and that could trigger sigbus
Jun 27 06:56:08 <bshah> .. but.. as krnlyng mentiond, __set_errno is probably irrelevent in newer target
Jun 27 06:56:14 <bshah> and my device is halium 7
Jun 27 06:56:32 <Tofe> ah, yes, right, I forgot
Jun 27 06:56:49 <bshah> I guess I best get backtrace
Jun 27 06:56:56 <Tofe> yup
Jun 27 06:57:14 <Tofe> if you have halium's build symbols, it's not that long
Jun 27 06:58:52 <bshah> yeah.. although, since I am at it.. let me finish qt 5.11.1 migration first
Jun 27 06:59:15 <Tofe> bshah: put a breakpoint on https://github.com/qt/qtbase/blob/5.11/src/corelib/thread/qthread_unix.cpp#L191 , then when in the problematic thread number add a watch on currentThreadData's address, and that's it, pretty much
Jun 27 07:00:37 <bshah> okay
Jun 27 07:20:28 <Herrie|Laptop> Morning
Jun 27 07:22:18 <Tofe> Herrie|Laptop: o/
Jun 27 07:59:31 <JaMa> morning
Jun 27 08:07:33 <Herrie|Laptop> JaMa: Morning!
Jun 27 08:13:43 <Herrie|Laptop> bshah: Redid this one so it can merge clean: https://github.com/Halium/android/pull/26
Jun 27 08:13:49 <JaMa> Herrie|Laptop: how is new house going?
Jun 27 08:13:53 <Herrie|Laptop> The 5.1 one was merged a while ago
Jun 27 08:14:02 <Herrie|Laptop> JaMa: Mortgage approved, keys next Thursday :D
Jun 27 08:14:17 <Herrie|Laptop> Now sorting kitchen.... Delivery time 12 weeks :S
Jun 27 08:14:44 <Herrie|Laptop> Those "efficient" Germans :P
Jun 27 08:31:14 <JaMa> nice :)
Jun 27 08:46:28 <Herrie|Laptop> Tofe: I'm testing something now so we could use the halium Android repo directly with local_manifest for our changes.
Jun 27 08:46:47 <Herrie|Laptop> We would need to add something to Halium documentation as well, but that's not a whole lot of work
Jun 27 08:46:55 <Herrie|Laptop> Will PR once I have it tested & working
Jun 27 08:48:22 <Tofe> Herrie|Laptop: why would we need to add something on Halium's documentation ?
Jun 27 08:49:41 <Herrie|Laptop> Tofe: Because you'd need to run the extra setup script to create the symlink which is just for LuneOS
Jun 27 08:49:51 <Herrie|Laptop> In general it doesn't hurt to write some documentation in Halium with our specifics
Jun 27 08:50:10 <Tofe> Herrie|Laptop: I think we can avoid that link thing
Jun 27 08:51:42 <Tofe> Let's imagine our luneos manifest becomes the main manifest; and halium's manifest is checked out in a subfolder of the root repository. Then we would only need, at the beginning of our manifest, to say 'include "halium-manifest/android.xml" '
Jun 27 08:52:36 <Herrie|Laptop> Tofe: Ah that would work too I guess
Jun 27 08:52:42 <Herrie|Laptop> I assumed Halium as main manifest
Jun 27 08:52:50 <Herrie|Laptop> But I guess result would be the same in the end
Jun 27 08:53:14 <Tofe> However that would mean "repo init" with our manifest + "git clone" of Halium's manifest somewhere else
Jun 27 08:55:07 <Tofe> I don't really see how this could be made automatic; maybe with git submodules, but I never really understood that part of git
Jun 27 08:56:19 <Herrie|Laptop> Let me see if my approach works
Jun 27 08:56:27 <Herrie|Laptop> We can always check with bshah if he has a preference too
Jun 27 08:57:08 <Tofe> He'll probably say he prefers that Halium can be reused as-is ;)
Jun 27 08:57:59 <Herrie|Laptop> Tofe: Well we don't need/want halium/hybris-boot ;)
Jun 27 08:59:00 <Herrie|Laptop> Tofe: Well in general we could add some documentation to https://docs.halium.org/en/latest/porting/install-build/index.html
Jun 27 08:59:07 <Herrie|Laptop> Doesn't hurt to have some things documented there
Jun 27 08:59:39 <Tofe> Sure
Jun 27 10:44:55 <saidinesh5> Tofe: so after so much usage any complaints , any problems with your redmi note 5 pro?
Jun 27 10:57:23 <Tofe> saidinesh5: well, it's a bit too big for my hands
Jun 27 10:58:01 <Tofe> otherwise, I'm quite happy: battery is great, everything works, regular updates...
Jun 27 10:59:03 <saidinesh5> Ahh cool.. it is pro right? the one with dual cameras?
Jun 27 11:01:30 <Tofe> yup; to be precise, its name is just "Redmi Note 5" in France, but it's dual camera and same specs as Pro
Jun 27 11:03:05 <saidinesh5> Ahh
Jun 27 11:18:57 <Tofe> saidinesh5: but if I would have to buy a Xiaomi device, I'd probably hesitate Mi 8 SE ... when that'll be out :)
Jun 27 11:19:05 <Tofe> +with
Jun 27 11:19:44 <Tofe> (Amoled screen, nice size, nice price, battery a bit less generous)
Jun 27 11:20:03 <saidinesh5> Ahh but no headphone jack!
Jun 27 11:20:09 <saidinesh5> Tofe: this is for a family member
Jun 27 11:20:23 <saidinesh5> so was only looking for good battery life and screen
Jun 27 11:21:00 <saidinesh5> i currently have a Lenovo P2 (my old Mi3 recently kicked the bucket).. it is quite nice if Lenovo didn't ditch it software wise
Jun 27 11:23:00 <saidinesh5> Amoled, 5000 mAh battery .. snapdragon 625
Jun 27 11:56:38 <Herrie|Laptop> Tofe: My "solution": https://github.com/Herrie82/android/commit/e394c5a53cef2ea9f58bf5acb67e2ca97faac90c and https://github.com/Herrie82/halium-luneos
Jun 27 11:56:59 <Herrie|Laptop> Which means that you'd have to run ./halium/halium-luneos/setup before ./halium/devices/setup DEVICENAME
Jun 27 11:57:01 <Herrie|Laptop> Then it'll work
Jun 27 11:58:45 <Tofe> Well that's one approach yes
Jun 27 11:59:03 <Tofe> I'll sketch mine tonight and we'll compare :)
Jun 27 11:59:35 <Herrie|Laptop> Anyway this way we can get rid of our custom bits :)
Jun 27 12:00:24 <Tofe> exactly
Jun 27 12:07:37 <Herrie|Laptop> Especially those patches eventually
Jun 27 12:07:50 <Herrie|Laptop> Should be OK for most targets already, but need to fix the Onyx and some others in Halium still
Jun 27 13:44:09 <Herrie|Laptop> Tofe: Small cosmetic PR: https://github.com/webOS-ports/meta-webos-ports/pull/291/files
Jun 27 13:47:06 <Tofe> merged (even rebased by GH, good tool :) )
Jun 27 13:51:00 <Herrie|Laptop> Tofe: And some Tissot work for which it'll do rebasing as well: https://github.com/webOS-ports/meta-webos-ports/pull/292/files
Jun 27 13:51:15 <Herrie|Laptop> You might want to add the si_swapinfo to your kernel for Rosy
Jun 27 13:51:19 <Herrie|Laptop> Otherwise it'll break :P
Jun 27 13:51:26 <Herrie|Laptop> I.e.: https://github.com/Herrie82/android_kernel_xiaomi_msm8953/commit/154357bba583889378648ebd096d7973c8ae96d3
Jun 27 13:52:14 <Tofe> yup
Jun 27 17:55:15 <Tofe> Herrie: actually my rosy does boot, after adding logs into the initrd... it's just that I forgot to add the ${LIBHYBRIS_RDEPENDS}  that you've added today for rosy... so I didn't have the Halium image in the rootfs...
Jun 27 17:58:15 <Herrie> Tofe: ok
Jun 27 17:58:20 <Herrie> Good
Jun 27 18:00:09 <Tofe> yup
Jun 27 18:00:15 <Herrie> That's progress :D
Jun 27 18:00:39 <Tofe> yes, now I hope it'll just go smoothly
Jun 27 18:00:43 <Herrie> See it was something small, silly but crucial :P
Jun 27 18:01:02 <Herrie> Like you expected
Jun 27 18:01:06 <Tofe> It's still annoying that my pstore or last_kmsg doesn't work, but well
Jun 27 18:01:12 <Herrie> I had that too on mine
Jun 27 18:01:34 <Herrie> Tofe: That's defconfig I think
Jun 27 18:01:59 <Tofe> yes, probably
Jun 27 18:02:13 <Herrie> Something like https://github.com/Herrie82/android_kernel_xiaomi_msm8953/commit/f974d591e9553d614fe4cc56b190f9aa598c8197
Jun 27 18:02:13 <Tofe> not worth it though, now that I identified the issue
Jun 27 18:03:27 <Tofe> https://github.com/Tofee/android_kernel_xiaomi_msm8953/blob/tofe/work/arch/arm64/configs/lineageos_rosy_defconfig#L4060 nope, it'll be something else (I kind of merged your defconfig :) )
Jun 27 18:25:21 <Tofe> Herrie: I... think I see a LuneOS logo
Jun 27 18:25:35 <Herrie> Ooh'
Jun 27 18:25:44 <Herrie> Exciting!
Jun 27 18:25:49 <Tofe> but no adb yet
Jun 27 18:25:56 <Tofe> so, mixed feelings
Jun 27 18:26:07 <Tofe> functionfs maybe?
Jun 27 18:26:18 <Herrie> Tofe: For adb you need to fix functionfs
Jun 27 18:26:24 <Tofe> ...and firstuse is there
Jun 27 18:27:07 <Tofe> mmh it hanged, somehow
Jun 27 18:27:14 <Herrie> Use this in android-gadget-setup: https://bpaste.net/show/3e7230791323
Jun 27 18:27:28 <Herrie> And push via recovery
Jun 27 18:27:40 <Tofe> ok thanks
Jun 27 18:28:24 <Herrie> This one doesn't work properly: https://github.com/webOS-ports/meta-webos-ports/blob/pyro/meta-luneos/recipes-android/android-tools/android-tools-conf/0001_Enable_FunctionFS.install_patch
Jun 27 18:28:35 <Herrie> I have that on mido too
Jun 27 18:28:42 <Herrie> Seems timing issue
Jun 27 18:28:57 <Herrie> Didn't have much chance to debug
Jun 27 18:29:13 <Tofe> ok
Jun 27 18:29:21 <Tofe> so it's just removing the "if"
Jun 27 18:30:40 <Herrie> Yup
Jun 27 18:30:53 <Herrie> Seems so
Jun 27 18:31:05 <Tofe> ok, edited with vi on recovery
Jun 27 18:31:27 <Herrie> That works for me on mido
Jun 27 18:33:38 <Tofe> nope, didn't work
Jun 27 18:55:12 <Tofe> Herrie: that's the only change you did to get adb working ? because it doesn't seem to work for me, unfortunately
Jun 27 19:14:37 <Herrie> Tofe: I think so
Jun 27 19:25:55 <Herrie|TP> Tofe: I would just adb push the file via recovery and chmod it if I recall correctly
Jun 27 19:26:38 <Herrie|TP> You could try to see if you could set the ran first use flag in recovery, then start fingerterm and output log to file
Jun 27 19:26:46 <Herrie|TP> Bit nasty but well
Jun 27 19:31:30 <Tofe> oh, I could finish firstuse :)
Jun 27 19:31:31 <Herrie|TP> It's a bit weird that both pstore and functionfs don't seem to work.
Jun 27 19:31:38 <Herrie|TP> Tofe: Ah ok
Jun 27 19:31:53 <Tofe> functionfs is present in /proc/filesystems
Jun 27 19:32:18 <Tofe> so it's either a timing issue, or something that hasn't been properly configured by the lxc container
Jun 27 19:32:19 <Herrie|TP> That's what I also had but the if didn't work for me somehow
Jun 27 19:32:46 <Herrie|TP> Not sure why your add wouldn't work then
Jun 27 19:33:00 <Tofe> I'll try starting it by hand from fingerterm
Jun 27 19:33:14 <Herrie|TP> Could be permissions for the android-gadget-setup but vi shouldn't change that
Jun 27 19:33:18 <Tofe> I have the same screen timeout as on mido, it's quite annoying
Jun 27 19:33:53 <Herrie|TP> Tofe: Well that's probably faulty nyx conf
Jun 27 19:34:14 <Herrie|TP> It's annoying but when fixed on 1 probably on both
Jun 27 19:34:23 <Herrie|TP> I guess audio works too?
Jun 27 19:34:56 <Herrie|TP> Worked for me out of box, so did for elvispre on onyx it seems
Jun 27 19:35:45 <elvispre> It did
Jun 27 19:36:01 <elvispre> (And does)
Jun 27 19:36:47 <Tofe> no audio here...
Jun 27 19:37:10 <Tofe> but the diagnostic shouldn't be hard, once I get this damned adb to work
Jun 27 19:38:14 <Herrie|TP> Tofe: Yeah Fingerterm is no fun to use for anything more then really necessary :-P
Jun 27 20:15:46 <elvispre> I'm getting that '=== LuneOS/Halium ===' message (and so on) in my journalctl log endlessly at pretty much one second intervals
Jun 27 20:16:06 <elvispre> Seems a bit suspicious to me
Jun 27 20:16:14 <Tofe> elvispre: huh ? that message comes from the initrd boot
Jun 27 20:16:36 <elvispre> Tofe: I get it again and again and again.
Jun 27 20:16:56 <Tofe> https://github.com/webOS-ports/meta-webos-ports/blob/master/meta-luneos/recipes-core/initrdscripts/initramfs-boot-android/init.sh#L103
Jun 27 20:17:05 <elvispre> Pretty much the whole boot log again and again and again, actually.
Jun 27 20:17:07 <Tofe> I'm really not sure how this can happen
Jun 27 20:17:23 <Tofe> ah, mmh, dmesg in a loop ? weird
Jun 27 20:17:52 <elvispre> Can the android get into a boot loop inside the container?
Jun 27 20:19:45 <elvispre> My onyx log file is ~6 MB on about the same timescale as my mako one is ~500 KB
Jun 27 20:20:21 <Tofe> does journalctl give a culprit as to who is outputting this ?
Jun 27 20:21:09 <elvispre> "onyx kernel" mostly
Jun 27 20:21:42 <elvispre> "onyx kernel" always, for that particular line
Jun 27 20:22:22 <Tofe> mmh ok
Jun 27 20:22:59 <elvispre> So not normal then :)
Jun 27 20:23:29 <Herrie> elvispre: Could be something like selinux or audit
Jun 27 20:24:00 <Herrie> elvispre: You also might want to try to debug android side of things with logcat
Jun 27 20:24:19 <Herrie> I.e.: http://docs.halium.org/en/latest/porting/debug-build/debug-android-userspace.html
Jun 27 20:24:26 <Herrie> There shouldn't be a whole lot in there
Jun 27 20:24:35 <Herrie> If there is, it's necessaryt o address
Jun 27 20:24:41 <Herrie> I had quite some when I started with mido
Jun 27 20:37:17 <Herrie> Once my logcat was cleaner the dmesg also got a lot cleaner
Jun 27 21:15:55 <Tofe> boot time on rosy is quite long... I'll have to see why
**** ENDING LOGGING AT Thu Jun 28 03:00:02 2018