**** BEGIN LOGGING AT Thu Jan 07 03:00:55 2021 Jan 07 07:12:40 Morning! Jan 07 08:31:40 Morning! Jan 07 08:32:05 Tofe: I saw you were fighting with some missing bits on Android side? Got that sorted? Jan 07 08:46:36 If you have some WIP stuff you could push, I'm happy to try with Mido or Tissot at my end Jan 07 09:15:38 Herrie: yes, the missing bits are there, now Jan 07 09:16:31 I'll push it soon yes; it's just some additional packaged products (I've put it in vendor .mk, but could be in device as well) Jan 07 09:17:07 Though it didn't solve my issue with luna-next, which seems to hang at some waiting point Jan 07 09:17:17 but logcat doesn't give much clue about that... Jan 07 09:17:56 I've tried to add minisfservice as a last try, but that's when the bluetooth (or audio?) driver began to crash again Jan 07 09:21:59 It could also be device specific so therefore might be good to check on another target Jan 07 09:22:07 That's my thinking at least Jan 07 09:22:11 yes, you're right Jan 07 09:29:00 Or could be something generic of course, but good to exclude the device specific reason at least Jan 07 09:29:22 We didn't have miniaf/minisf before and luna-next was OK then Jan 07 09:29:44 I assume the libhybris tests are OK or you don't get to that? Jan 07 09:32:56 Stracing those might give some clues that luna-next doesn't? Jan 07 09:52:10 I haven't tried them yet, but I intended to too, yes Jan 07 09:56:17 Herrie: can I just push into your android_device_xiaomi_rosy-2 ? looks the easiest way Jan 07 09:58:00 Too late, already forked :p Jan 07 09:59:16 https://github.com/Herrie82/android_device_xiaomi_rosy-2/pull/1 Jan 07 10:04:37 Tofe: Sorry a bit slow Jan 07 10:04:49 Managed to break the LCD of my laptop while swapping keyboard |:( Jan 07 10:05:11 And don't have HDMI->DVI cable, so a bit limited in terms of access for a day or 2 Jan 07 10:05:32 Cable should've arrived yesterday, but now scheduled for today. New LCD should come tomorrow hopefully :P Jan 07 10:09:26 Herrie: ah, that's unfortunate Jan 07 10:09:40 well, delivery is quick it seems Jan 07 10:09:41 Yup, well sh*t happens Jan 07 10:09:48 Comes from Canada with UPS ;) Jan 07 10:09:50 2 days Jan 07 10:10:00 Price same as here in Europe and quicker delivery it seems so well Jan 07 10:10:14 LCD was cheaper, shipping more expensive, but in the end all OK Jan 07 10:10:17 What a crazy world we leave in Jan 07 10:10:24 EUR 115 for a QHD incl. delivery not too bad Jan 07 10:10:38 For my "old" X1 Carbon 4th gen Jan 07 10:10:56 Replaced battery last week as well, went from < 1 hr back to 6-7 hrs on full charge :P Jan 07 10:14:47 good as new! Jan 07 10:20:47 Yup got the one with a bit upgraded capacity as well. 56Wh vs 52Wh so not bad Jan 07 10:25:22 Tofe: Weird all those dependencies are missing btw Jan 07 10:25:36 I guess they pull them in their generic arm device on Android side? Jan 07 10:27:13 some of them yes, like libandroid and libril iirc Jan 07 10:27:34 but here they are probably pulled indirectly by the UI components that we excluded Jan 07 10:34:33 OK, I'll try and see if I can put something together for Tissot maybe, since we both have it Jan 07 10:34:50 I thought you also made some adjustments to other scripts? Jan 07 10:41:07 mmh I just build vendorimage too Jan 07 10:41:15 or you mean for luneos? Jan 07 10:41:51 There I did a couple of changes yes, to adapt initramfs and lxc config Jan 07 11:11:50 Let me organize these modifications a bit, then I'll push them on my GH Jan 07 11:21:52 Tofe: I build vendor.img with my script anyway Jan 07 11:21:54 And pack it Jan 07 11:22:00 Yeah the LuneOS side of things Jan 07 13:06:36 Herrie: you can have a look in testing/halium-9.0 for meta-wop and smartphone, on my GH Jan 07 13:06:43 I hope I didn't mess up Jan 07 13:11:56 Tofe: OK looks pretty OK so far Jan 07 13:12:07 Will see if I can get some Tissot building shortly Jan 07 13:12:16 I didn't include the android-system-image-rosy.bb change regarding picking the newly built images Jan 07 13:12:24 as it's a local file for m Jan 07 13:12:25 me Jan 07 13:13:02 the QPA change should be PR-able directly today Jan 07 13:20:25 Yes Jan 07 13:20:36 Let's do that already Jan 07 13:20:41 Did you delete thepatches as well? Jan 07 13:22:21 Seems you didn't, might be good to drop the patch files Jan 07 13:22:26 I noticed sometimes we forget Jan 07 13:24:10 I yes, I forgot, let met add that Jan 07 13:29:47 ok done Jan 07 14:23:53 Trying some initial Tissot build Jan 07 14:26:36 good luck ! Jan 07 14:27:07 you already have a lineage-16 kernel ? Jan 07 14:27:18 not sure it's needed at all though Jan 07 14:28:31 Yes Jan 07 14:28:38 Updated defconfig Jan 07 14:29:05 ... with F_FS ;) Jan 07 14:32:51 function fs? Jan 07 14:34:39 Which flags I need for that? Jan 07 14:34:50 So I can add them to my script to update defconfig ;) Jan 07 14:36:49 let me check Jan 07 14:37:36 Seems it builds successfully but no vendor.img for Tissot somehow Jan 07 14:38:20 Ah might be the same issue you had with vendorsetup.sh Jan 07 14:38:27 from https://github.com/webOS-ports/meta-pine64-luneos/blob/master/recipes-kernel/linux/linux-pinephone/extra.cfg#L29 you'll need CONFIG_USB_F_FS CONFIG_USB_CONFIGFS CONFIG_USB_CONFIGFS_F_FS CONFIG_USB_FUNCTIONFS to "y" Jan 07 14:38:34 ^ at least Jan 07 14:39:15 vendorsetup.sh is mainly useful to get a working lunch Jan 07 14:39:28 not sure if it influences the build at all Jan 07 14:41:42 Well it seems it refers to a previous one: herrie@herrie-desktop:/media/herrie/HaliumDisk/9.0$ mka vendorimage Jan 07 14:41:43 build/make/core/product_config.mk:234: error: Can not locate config makefile for product "lineage_karnak". Jan 07 14:41:53 Karnak was a device I was helping someone to build Jan 07 14:42:53 ah, ok, yes it's a bit strange, there's some cache involved here... Jan 07 14:44:46 Yeah Jan 07 14:44:53 Let me simply delete build folder Jan 07 14:44:56 And update defconfig Jan 07 14:55:31 OK another TIssot build running Jan 07 15:12:10 Same issue Jan 07 15:12:20 Let me just delete my 9.0 folder and start from scratch Jan 07 15:12:38 Goes quickly enough :P Jan 07 15:33:39 Got a DVI->HDMI cable but doesn't work :S Jan 07 15:33:48 I guess I'll just wait for LCD tomorrow Jan 07 15:33:56 UPS is pretty good with their delivery usually :) Jan 07 16:01:13 Hmmz still no vendor.img Jan 07 16:53:28 Herrie: maybe it's just not a Treble image ? It's not an issue, as long as we don't want to go all-GSI Jan 07 16:53:59 Having only android-rootfs.img is fine, as long as /vendor is actually populated Jan 07 16:55:50 Seems it's system-as-root Jan 07 16:55:58 So probably non-treble Jan 07 16:59:49 well, treble often mean system-as-root, it's only /vendor that is separate Jan 07 17:00:32 system-as-root is mandatory from Android 10, and only allow on Android 9 in a case of an upgrade from a non-treble Jan 07 17:01:11 s/and only allow/and non-system-as-root is only allowed/ Jan 07 17:02:22 I guess it's to separate the kernel (vendor specific) from the initramfs (can be GSI) Jan 07 17:24:26 OK anyway will push it to Jenkins soon Jan 07 17:24:33 Then we can both use the same Jan 07 18:09:46 Tofe: Tissot image on: http://build.webos-ports.org/halium-luneos-9.0/ Jan 07 18:10:07 Halium image that is Jan 07 18:10:16 Will see if I can mock together a complete build in a bit as well Jan 07 18:20:04 Ok quickly put something together: https://github.com/Tofee/meta-smartphone-1/pull/1 Jan 07 18:20:07 Now building Jan 07 18:20:08 Let's see Jan 07 18:20:26 Not tested at all of course Jan 07 18:21:24 OK seems we need a little tweak already in case we don't have vendor.img: https://bpa.st/PN2Q Jan 07 18:22:43 ah, yes, that's a stupid mistake Jan 07 18:24:48 if you add it to the PR, I'll merge it Jan 07 19:08:15 What to add? Jan 07 19:09:43 Should add "if [ -e ${WORKDIR}/vendor.img ]; then" and fi ? Jan 07 19:11:56 exactly yes Jan 07 19:13:36 OK this should work I guess Jan 07 19:14:13 Haven't seen this one before on kernel compilation: "arch/arm64/Makefile:67: *** CROSS_COMPILE_ARM32 not defined or empty, the compat vDSO will not be built. Stop." Jan 07 19:15:14 Backport this one? https://patchwork.kernel.org/project/linux-arm-kernel/patch/20190222122430.21180-19-vincenzo.frascino@arm.com/ Jan 07 19:20:07 the error doesn't ring any bell either, let me see the link Jan 07 19:20:56 what is vDSO ? :) Jan 07 19:21:00 Or can we simply disable it in defconfig? Jan 07 19:21:01 https://cateee.net/lkddb/web-lkddb/COMPAT_VDSO.html Jan 07 19:21:31 Herrie: looks like your patch is already there, it's what triggers the error Jan 07 19:22:05 It's mentioned here: https://irclogs.sailfishos.org/logs/%23sailfishos-porters/%23sailfishos-porters.2019-05-17.log.html Jan 07 19:22:12 ah, yes, it's for old glibc, we don't care Jan 07 19:24:20 Let's try Jan 07 19:26:42 Plenty compilation errors.... Jan 07 19:27:35 https://paste.ubuntu.com/p/5t5BYmZmS3/ Jan 07 19:27:43 4.9 kernel, so wouldn't GCC to complain too much Jan 07 19:27:45 But who knows Jan 07 19:35:29 Will try this: https://github.com/Herrie82/android_kernel_xiaomi_msm8953-1/commit/a97c40ebe4eaa90dd468abb7b1a9d4c79406985f Jan 07 19:36:20 Seems LCD already is in Koeln for 1 hr, so I guess delivery tomorrow ;) Jan 07 19:50:23 Maybe delivered with a Kölsch Jan 07 19:52:13 your patch looks adapted Jan 07 19:53:14 mmmh "__u8 eir[0];" that's quite a weird line here ! Jan 07 19:53:18 Kolsch would benice Jan 07 19:53:34 Well addded a branch to test some "fixes" https://github.com/Herrie82/android_kernel_xiaomi_msm8953-1/commits/halium-9.0-fixes Jan 07 19:54:42 https://www.spinics.net/lists/linux-bluetooth/msg84890.html damn stupid syntac, can't they just write that's it's a pointer... Jan 07 19:55:49 Hehe yes Jan 07 20:02:02 Seems this kernel is quite a mess with GCC 9+ Jan 07 20:02:12 I'll leave it for today, got some other things to do still ;) Jan 07 20:02:23 Seems needs this: https://lkml.org/lkml/2020/4/30/1471 Jan 07 20:02:26 And other things as well Jan 07 20:04:43 it's a pity that a newer kernel is harder to build Jan 07 20:05:44 Yeah well it's a lot of the garbage qcom code Jan 07 20:06:09 Saw similar things with another Xiaomi with recent kernel Jan 07 20:06:37 Not very surprising indeed Jan 07 20:07:05 anyway, I'll disable the android lxc and give test_hwc a try on rosy Jan 07 23:00:25 Tofe: Seems this one has latest 4.9 kernel version with some fixes for newer GCC; https://github.com/d4rk-lucif3r/LuciferKernel/tree/NonOC-reb Jan 07 23:00:30 Will try that one instead Jan 07 23:52:51 Tofe: NFS seems somewhat broken, we don't use it so disabled it for now. Also memnotify module seems to have issues again: https://paste.ubuntu.com/p/CMgfDDv8zg/ Jan 07 23:53:01 I removed it from packagegroup for now Jan 07 23:53:22 Then kernel builds with a few fixes it seems and I get an image at least **** ENDING LOGGING AT Fri Jan 08 02:59:57 2021