**** BEGIN LOGGING AT Fri Mar 08 02:59:56 2019 Mar 08 07:05:10 JaMa: thanks ! Mar 08 07:05:11 Morning! Mar 08 07:05:28 Herrie: mako kernel has now been merged Mar 08 07:05:40 I think I have ~1 hour left to bump PR :p Mar 08 07:14:20 Tofe: Ah that was quick from UBPorts Mar 08 07:14:23 Usually they're not that quick Mar 08 07:14:32 Well you need JaMa for meta-smartphone Mar 08 07:14:46 And for Hammerhead it wasn't merged yet. Mar 08 07:14:49 bshah: ^ Mar 08 07:14:55 The guy did just a quick visual check... I don't know if ubports still supports mako? Mar 08 07:15:39 Tofe: Also not sure, but well ;) Mar 08 07:16:16 I wanted to first build complete sumo, then check the Halium kernel build, but, as you, "well" :) Mar 08 07:16:54 Well biggest issue was really the mounts not working with systemd Mar 08 07:16:59 Rest isn't rocket science ;) Mar 08 07:17:12 Well the systemd not working wasn't either in the end LOL Mar 08 07:17:12 JaMa: https://github.com/shr-distribution/meta-smartphone/pull/92 Mar 08 07:17:38 bshah: https://github.com/Halium/android_kernel_lge_hammerhead/pull/3/files Mar 08 07:19:12 at least we can say that things are going well for the sumo switch Mar 08 07:19:19 Tofe: I guess I'll PR documentation change to https://docs.halium.org/en/latest/porting/build-sources.html#documented-errors to outline the patch Mar 08 07:19:32 And it's more visible v.s. stuck somewhere in a commit & issue tracker Mar 08 07:20:14 good idea Mar 08 07:21:27 In this debug for the mounts, I was fairly happy that we have all our bootimg integrated in the toolchain Mar 08 07:22:15 So I'm not going to switch to the full halium-boot anytime soon :p Mar 08 07:29:16 Tofe: Well they use old toolchain for that Mar 08 07:29:24 So I don't see much reason to switch Mar 08 07:34:13 OK PR for Halium documentation made: https://github.com/Halium/docs/pull/132 Mar 08 08:08:15 Good ! Mar 08 09:21:08 Morning all, sorry been offline a couple of days, can someone bring me up to speed on the questions please? Mar 08 09:21:36 believe it or not my log doesn't go far enough back! Mar 08 09:22:04 novaldex: There is nothing too much for you Mar 08 09:22:13 There were some host key issues JaMa saws Mar 08 09:22:16 -s Mar 08 09:22:21 Herrie|Laptop: the words I like to hear! :) Mar 08 09:23:35 those host keys on bonaire? what was the target server? Mar 08 09:38:13 JaMa, it's not something i've seen here before, but it would indicate an IO issue somewhere Mar 08 09:38:35 i'm hoping it's only a one off, but i'll keep my eye on things & see if there was anything else going on at the time Mar 08 09:40:54 novaldex: COuldn't it be related to the work you did for the PReware feeds? Mar 08 09:40:59 Seems it was something on 25th of January Mar 08 09:41:18 for the keys? it is possible, but easily fixed Mar 08 09:42:06 did do quite a bit of cleaning out that day, was hoping I hadn't removed any of the wrong keys, bit harder to easily tell at a glance now Mar 08 09:45:13 novaldex: See http://logs.nslu2-linux.org/livelogs/webos-ports/webos-ports.20190306.txt at the bottom Mar 08 09:55:18 ah, github.com? that's a funny one to be missing.. let me take a look Mar 08 10:05:59 that's weird, i'm not getting a host key issue when connecting to the same host (name & IP match) Mar 08 10:09:39 novaldex: Could be JaMa fixed it already Mar 08 10:12:16 that would make sense.. Mar 08 10:25:58 Anyway we're making solid progress on some of the software migration bits ;) Seems we can move from Yocto Pyro to Sumo directly and skip Rocko after getting our 3.4 kernel devices to behave with systemd 233 and higher ;) Mar 08 10:26:10 Thud still has some issues, but we'll get there as well :) Mar 08 10:49:53 novaldex: the ssh key issue was github.com Mar 08 10:50:35 and yes, I've added it at the same time as reporting it here Mar 08 10:51:25 Tofe: you forgot to remove the .patch files, I'll fix it when cherry-picking Mar 08 10:57:18 JaMa: oops, you're right! Thanks Mar 08 11:05:55 Tofe: Herrie|Laptop: I'll merge your pull in few Mar 08 11:06:07 bshah: Thnx Mar 08 11:06:22 Also sent one for documentation, so we have it properly documented :) Mar 08 11:08:12 sorry for being MIA today/yesterday... I was sick with food poisoning :/ Mar 08 11:09:23 bshah: No worries, things like that happen. Hope you're feeling better! Mar 08 11:09:48 Last time I was sick from food was long long time ago... I think I was 7 LOL ;) Mar 08 11:10:06 I guess all the vinegar I drank during childhood (would drink the vinegar from pickle jars LOL) helped somehow Mar 08 11:10:49 although I admit I am quite unhappy about the 3.4 kernel getting fixed :P I wanted to kill it!! #letmekillitnowplease :( Mar 08 11:11:21 bshah: Who doesn't but well.... Would be good to keep it alive until mainline is viable ;) Mar 08 11:11:31 At least for our Tenderloin ;) Mar 08 11:12:32 :< Mar 08 11:18:49 bshah: Where you'd like to have it in the documentation? Mar 08 11:19:17 Herrie|Laptop: during initial preparations, where one would do initial kernel config etc Mar 08 11:19:42 So here somewhere? https://docs.halium.org/en/latest/porting/build-sources.html#modify-the-kernel-configuration Mar 08 11:19:47 aye Mar 08 11:20:41 OK can change that. Is there any reason why the "mer-kernel-check" is not pulled in by the manifest? Mar 08 11:23:11 no specific reason Mar 08 11:29:45 bshah: I guess I'll add it at some point. Will reduce the commands needed :) Mar 08 11:30:39 bshah: https://github.com/Halium/docs/pull/132 Mar 08 11:30:51 Updated and passed Travis CI check ;) Mar 08 11:41:53 Herrie|Laptop: I've approved changes, let's wait for others to approve chages Mar 08 11:42:00 OK :) Mar 08 11:42:10 if no-one does it I'll merge it tomorro Mar 08 11:46:25 OK :) Mar 08 12:06:51 thanks JaMa Mar 08 12:38:30 bshah: I'll soon begin testing that hammerhead mainline, I'm quite curious about it Mar 08 13:13:01 Tofe: you can try building pmos image with plasma-mobile UI to try it out Mar 08 13:13:04 :) Mar 08 13:17:27 bshah: Doing mainline kernel from our build is pretty easy too ;) Mar 08 13:17:28 mido, rosy, tissot done on jenkins builder, new sumo based testing :) Mar 08 13:20:08 JaMa: great! Mar 08 13:22:30 and mido has that kernel fix as well Mar 08 13:24:51 JaMa: Mido has 3.18 shouldn't need kernel fix. Or you mean the GCC 7 one? Mar 08 13:29:57 yes, the meta-smartphone PR Mar 08 13:30:54 sorry, I meant mako not mido Mar 08 13:31:20 first and last letter being the same, I cannot distinguis between these 2 :) Mar 08 17:00:36 https://github.com/shr-distribution/meta-smartphone/pull/93 this should solve our little wifi issue. To be confirmed... Mar 08 18:14:32 JaMa: ping Mar 08 18:28:29 Tofe: pong Mar 08 18:28:51 PR ready? Mar 08 18:29:43 yes, it's ready, but I had a question about something else :) Mar 08 18:30:09 JaMa: I'm playing with building linux-yocto for hammerhead, and get "do_kernel_metadata: Could not locate BSP definition for hammerhead/standard and no defconfig was provided" ... I found out it works better if I name it "defconfig", but can't I choose my own name ? Mar 08 18:31:33 yes, the file needs to be called defconfig Mar 08 18:31:44 going to try 5.0 on hammerhead? Mar 08 18:31:51 oh well, why not :) Mar 08 18:31:56 notice some linux-yocto changes I've merged today Mar 08 18:32:14 I'm on sumo, does it still apply ? Mar 08 18:32:29 (I should probably experiment on thud, I know) Mar 08 18:32:31 I have one small set of image change (adding more stuff to dev and example image) Mar 08 18:32:47 yes, I've backported it to sumo as well Mar 08 18:33:00 ah ok, I'll have a look Mar 08 18:33:29 we were using linux-yocto-dev with 4.19 even in pyro (because that's where in-kernel vboxvideo works with new mesa) Mar 08 18:33:35 https://github.com/torvalds/linux/commit/5cf5332d529bf7fd1f66378bb9611387c3e02b05#diff-6edc52f8a82e14082432721f3a52b2ab Mar 08 18:34:09 then I've upgraded it to 5.0-rc6 in thud and newer and yesterday versioned (not -dev) recipes were merged to oe-core, so I've backported those instead Mar 08 18:34:40 will cause some issues for e.g. lttng, perf in old sumo versions, but we're not including those Mar 08 18:34:55 are you flashing -dev or normal images? Mar 08 18:34:59 ok Mar 08 18:35:12 usually -dev Mar 08 18:35:18 I was wondering if e.g. added weston to -dev image can be the source of some of thud issues you were seeing Mar 08 18:35:51 e.g. there was a race between luna-next and weston who will initialize gfx first Mar 08 18:36:07 oh, weston is started automatically ? via systemd ? Mar 08 18:36:21 probably shouldn't be the root cause for the crash in render thread you were seeing, but still might be my fault Mar 08 18:36:27 ys with systemd service Mar 08 18:36:39 ok, indeed we don't want weston at all Mar 08 18:36:58 usually it fails to start (because luna-next was faster), but I was testing this only on qemux86 and tissot might be a bit different Mar 08 18:37:17 yes, it was just to debug gfx in vbox Mar 08 18:37:56 and it's still included for thud and newer Mar 08 18:39:02 weston will rarely be wanted on anything else than x86 and x64 Mar 08 18:39:16 https://github.com/webOS-ports/meta-webos-ports/commit/296dd761640346be8c719e2cf04a10b136ed8e67 Mar 08 18:39:26 maybe it's not the best criteria... Mesa then ? Mar 08 18:39:41 this is currently in jansa/sumo, that's why it was included in thud as well Mar 08 18:40:42 I would move weston* to the MESA include Mar 08 18:42:56 if it conflicts with luna-next then I should drop it for all and include it only in my test images Mar 08 18:47:24 well it's not sure at all that it's the culprit Mar 08 18:47:40 however, starting two wayland compositor is bound to create issues Mar 08 18:50:10 oh, zImage-dtb isn't a universal target ? I'll have to learn my linux basics again :) Mar 08 18:52:49 ah ok, things like https://cateee.net/lkddb/web-lkddb/ARM_APPENDED_DTB.html might come in handy... Mar 08 18:57:45 at least, linux-yocto does build. I'll still have to figure out at which step to include the dtb Mar 08 19:02:08 you can check meta-raspberrypi for that, it's also using linux-yocto with own defconfig and own dtbs Mar 08 19:03:17 oh, ok, thanks ! Mar 08 19:13:22 ok, dtb built, not yet appended to zImage... that's quite easy ! Mar 08 19:20:36 ahah, oops, I was going to use Halium... silly me Mar 08 19:33:47 And now it begins to be interesting: we don't have enough flexibility for our initramfs image; it always will take the init.sh dedicated to the Halium init script Mar 08 19:35:25 JaMa: I guess that in this case the usual way it to do a virtual/initrd-boot-scripts and several providers ? One for Halium, one for mainline Mar 08 19:35:30 is* Mar 08 19:38:00 mmmh or I could override using bbappend Mar 08 19:38:09 virtual/* is good only for build-time dependencies Mar 08 19:38:39 for runtime you need VIRTUAL-RUNTIME_foo Mar 08 19:41:15 Currently our boot image for devices is: android bootimg + initramfs image, the latter including initramfs-boot-android. In meta-smartphone/meta-android, the latter is very generic and would be fine for a mainline kernel too. But we override it in meta-webos-ports with a bbappend, always, and replace scripts with a Halium init. Mar 08 19:41:24 Something looks wrong here. Mar 08 19:43:58 I'll look at VRITUAL-RUNTIME_foo, maybe it'll solve this elegantly Mar 08 19:44:54 Yes, I think it will Mar 08 19:45:08 at least you can set the default to the generic one in meta-smartphone and then change it in meta-webos-ports with just one variable Mar 08 19:45:33 but I haven't check them now to see what exactly is the difference in the bbappend Mar 08 19:47:01 yes it would be a good idea, something like VIRTUAL-RUNTIME_android-initrd-scripts = "android-initrd-scripts-mainline" in meta-smartphone, VIRTUAL-RUNTIME_android-initrd-scripts = "android-initrd-scripts-halium" in meta-webosports, and once again overriden for hammerhead machine specifically Mar 08 19:47:18 that way we keep upward compatibility Mar 08 19:49:54 yes Mar 08 19:50:44 gtg, should get new sata controller tomorrow, need to move some more data around to free some disks Mar 08 19:52:06 ok :) Mar 08 19:52:31 also son will be baptised tomorrow and we have quite a few people here for the weekend, so I might need to help wife a bit and will be less around during the weekend Mar 08 19:52:57 Busy week-end I see ! Mar 08 19:53:09 Thanks for your help, and enjoy ! Mar 08 19:53:54 Thanks for yours! I'm really glad that we moved to sumo in such short time Mar 08 19:54:03 me too :) Mar 08 19:54:34 next steps thud qt-5.12.2 to be ready for warrior Mar 08 19:55:17 hehe Mar 08 19:56:31 NOTE: Running task 7257 of 7459 Mar 08 19:56:47 ^ looking forward for new Herrie-level builder :) Mar 08 21:01:59 And... here is my first built mainline hammerhead fastboot image :) Mar 08 21:05:10 Well, it doesn't boot, that's not really a surprise ^^ Mar 08 21:32:39 I'll build the whole image tomorrow. Mar 08 21:48:03 For the curious ones: https://github.com/Tofee/meta-webos-ports/commits/tofe/sumo/mainline-hammerhead and https://github.com/Tofee/meta-smartphone-1/commits/tofe/sumo/mainline-hammerhead Mar 08 21:48:35 time to sleep ! **** ENDING LOGGING AT Sat Mar 09 03:00:34 2019