**** BEGIN LOGGING AT Sun Apr 11 02:59:56 2021
Apr 11 07:31:29 <Herrie> Morning!
Apr 11 07:34:44 <Herrie> Tofe: https://github.com/Herrie82/android_kernel_htc_msm8960/commit/12e0ba6b41b08a7cbc83364f68db0aa1d54843e1
Apr 11 07:43:33 <Herrie> Tofe: Useful if I put it in the correct branch :P https://github.com/Herrie82/android_kernel_htc_msm8960-1/commits/halium-9.0-defconfig-test
Apr 11 07:43:39 <Herrie> https://github.com/Herrie82/android_kernel_htc_msm8960-1/commit/eb7dd706b7b93d861aeff0f783d821016fb75c61
Apr 11 08:52:39 <Tofe> Herrie: good :)
Apr 11 08:52:42 <Tofe> Morning!
Apr 11 09:04:52 <Herrie> Tofe: FWIW I was using an older SRCREV in the recipe still
Apr 11 09:05:02 <Herrie> Not sure if it would build & boot with current SRCREV
Apr 11 09:05:21 <Herrie> I am also getting the same boost issue as you had it seems
Apr 11 09:06:29 <Tofe> Herrie: ah, the initramfs issue ?
Apr 11 09:06:51 <Herrie> Tofe: No boost not building at my end
Apr 11 09:07:08 <Herrie> Will try to fix it when rest of build finishes
Apr 11 09:07:33 <Tofe> I did a lot of cleansstate in my rebuild when I updated to recent gatesgarth
Apr 11 09:08:14 <Tofe> If you see you're on the same road, just build from scratch, you can afford it :)
Apr 11 09:11:00 <Herrie> Tofe: Yeah I alreayd removed tmp-glibc but that didnt do the trick it seems
Apr 11 09:12:04 <Tofe> the sstate cache is the issue
Apr 11 09:13:03 <Herrie> OK
Apr 11 09:13:38 <Herrie> I'll have a look in a minute
Apr 11 09:13:58 <Herrie> Just for kernel, you migth want to try to build with https://github.com/Herrie82/android_kernel_htc_msm8960-1/commit/eb7dd706b7b93d861aeff0f783d821016fb75c61 and see if it still works
Apr 11 09:19:19 <Herrie> Because the recipe was still pointing to 69a7e9b0836c390850bdbcecfab7b9b56fce672f
Apr 11 09:19:30 <Herrie> I suspect it will be OK, but who knows
Apr 11 10:26:55 <Tofe> Herrie: it removes quite a lot of fixes
Apr 11 10:29:04 <Tofe> ah no sorry, was looking at wrong branch
Apr 11 10:30:03 <Tofe> Herrie: you might want to switch the names of these halium branches
Apr 11 10:49:05 <Herrie> Tofe: Better? https://github.com/Herrie82/android_kernel_htc_msm8960-1
Apr 11 10:49:38 <Herrie> Renamed halium-9.0 to halium-9.0-old and renamed the halium-9.0-deconfig-test to halium-9.0
Apr 11 11:47:24 <Tofe> yes :)
Apr 11 12:23:27 <Tofe> Herrie: the new kernel build works as well
Apr 11 12:27:15 <Herrie> Tofe: I expected it to work, but good you confirmed
Apr 11 12:27:48 <Tofe> I can PR the bump
Apr 11 12:31:25 <Tofe> https://github.com/shr-distribution/meta-smartphone/pull/126
Apr 11 12:31:52 <Tofe> I also fixed the /boot mount move, so that we can also replace the uImage from our adb easily
Apr 11 13:13:50 <Tofe> Herrie: looks like we're missing something that would generate hwcomposer.msm8660 ; something like a "hardware/qcom/display-caf/msm8660" or alike
Apr 11 13:27:31 <Herrie> Wouldn't that be provided by vendor normally?
Apr 11 13:27:49 <Tofe> by qcom more precisely
Apr 11 13:28:21 <Tofe> but iirc, 8660 get mapped to 8960 in terms of hwcomposer
Apr 11 13:31:40 <Tofe> Currently I'm trying to integrate vendor/extras from DirtyUnicorns
Apr 11 13:37:43 <Tofe> doesn't want to work :( I'm now trying to symlink msm8960 to msm8660 in display-caf
Apr 11 13:38:24 <Tofe> nope
Apr 11 13:48:20 <Herrie> Tofe: Is this vendor/extras even in some of the dependency files?
Apr 11 13:48:28 <Herrie> I would expect it to be there if it's really needed for Tenderloin
Apr 11 13:48:47 <Tofe> it's in their general manifest
Apr 11 13:48:59 <Herrie> Ah ok
Apr 11 13:55:44 <Tofe> overall it's mainly a question of setting the right variables
Apr 11 14:04:49 <Tofe> damn, lineage is overriding vendor/extras and putting a wrong value for QCOM_HARDWARE_VARIANT
Apr 11 14:15:19 <Tofe> I think I got it right, but I had to patch vendor/lineage
Apr 11 14:22:21 <Tofe> https://github.com/Halium/hybris-patches/pull/19
Apr 11 14:30:36 <Tofe> kernel oops for msm_v4l2_open
Apr 11 14:31:00 <Tofe> NULL pointer
Apr 11 14:38:57 <Tofe> looks like checking the input pointers is too much to ask
Apr 11 14:45:59 <Herrie> Tofe: Hmmz weird that worked before
Apr 11 14:47:30 <Tofe> Herrie: that's when starting android in lxc
Apr 11 14:49:56 <Tofe> and now it's the gstreamer plugin that crashes, great
Apr 11 14:50:11 <Tofe> though a bit better
Apr 11 14:52:25 <Tofe> ah, it's the rendering thread that actually crashes :(
Apr 11 15:07:29 <Herrie> Couldn't that flag in qt5-hwcomposer-plugin that I pointed out yesterday help there?
Apr 11 15:11:56 <Tofe> Herrie: https://bpa.st/VSGQ not sure, maybe I should try, but I don't think it's related
Apr 11 15:12:45 <Tofe> here eglGetDisplay is working it seems
Apr 11 15:14:05 <Tofe> in logcat, no obvious error is thrown
Apr 11 15:14:24 <Herrie> This "Thread 16 "QSGRenderThread" received signal SIGSEGV, Segmentation fault." somehow rings a bell
Apr 11 15:15:00 <Tofe> we had sevral crashes like that in the past
Apr 11 15:16:11 <Tofe> mostly due to TLS and such I think
Apr 11 15:17:55 <Herrie> I also remember that there was something special with the Tenderloin framebuffer
Apr 11 15:18:01 <Herrie> Like it had a 3rd frame or something
Apr 11 15:18:04 <Herrie> While others have 2
**** ENDING LOGGING AT Mon Apr 12 02:59:56 2021