| mal | try changing it | 00:00 |
|---|---|---|
| sharks_ | why is it looking in lib64? | 00:00 |
| sharks_ | That folder doesn't exist lol | 00:00 |
| mal | I think we might have assumed devices using hybris-17.1 are 64-bit most of the time | 00:00 |
| mal | is there /usr/libexec/droid-hybris/system/lib/libselinux_stubs.so ? | 00:00 |
| sharks_ | Just checked that, yeh there is | 00:01 |
| mal | check the other .rc files in /usr/libexec/droid-hybris/system/etc for stubs and fix the paths | 00:01 |
| sharks_ | Will do | 00:02 |
| sharks_ | Okay, the only other file looking at lib64 was vndservicemanager.rc | 00:05 |
| sharks_ | Changed that, now reboot and see what happens? | 00:05 |
| sharks_ | Still the same issue in logcat, though now journal says hwservicemanager exited 4 times in 4 minutes rather than 4 times before boot completed | 00:10 |
| sharks_ | And it's no longer spamming that it's waiting a second | 00:10 |
| sharks_ | logcat & journal -> https://paste.opensuse.org/pastes/60d02c6abe27, https://paste.opensuse.org/pastes/60d02c6abe27 | 00:10 |
| sharks_ | oh wait, wrong links sorry | 00:11 |
| sharks_ | that was logcat, this is journal: https://paste.opensuse.org/pastes/4aa0b8358149 | 00:11 |
| ecrn | Jun 14 01:23:21 GalaxyTabA9.7 droid-hal-init: /system/etc/init/hwservicemanager.rc: 1: ignored duplicate definition of service 'hwservicemanager' | 00:20 |
| mal | ecrn: that is normal | 00:20 |
| ecrn | ok | 00:20 |
| sharks_ | I think I worked out what at least part of the problem may be | 00:20 |
| mal | it happens because it had already loaded the custom .rc from usr/libexec | 00:20 |
| mal | sharks_: ? | 00:20 |
| sharks_ | The hybris-16 changes in the hadk faq | 00:21 |
| sharks_ | I don't have those lines in logcat | 00:21 |
| sharks_ | and droid-hal-init has failed | 00:21 |
| sharks_ | ergo I need to add the files and symlinks, rebuild config packages, etc. | 00:21 |
| mal | what files? | 00:22 |
| mal | those parts of 16.0 are not relevant | 00:22 |
| sharks_ | Oh bugger. | 00:23 |
| sharks_ | Nevermind then. I'd got all excited I'd found something to fix | 00:23 |
| mal | only part of 16.0 instructions you do for 17.1 is "Before building hybris-hal run the following commands:" | 00:23 |
| sharks_ | Yeh I did that | 00:23 |
| mal | how you have "Service 'hwservicemanager' (pid 893) received signal 6" | 00:24 |
| mal | so at least a different issue | 00:24 |
| sharks_ | Yeh, but I can't find any other logs relating to hwservicemanager, so I don't know why it's being killed | 00:25 |
| mal | I was trying to check where to stubs came to the .rc files, were those there all the time? | 00:25 |
| sharks_ | Yeah, actually I think they came from the hybris-16 faq too | 00:26 |
| sharks_ | Maybe I did that wrong | 00:27 |
| mal | yeah, that is wrong | 00:27 |
| mal | remove the custom stuff you added manually to /usr/libexec/droid-hybris/system/etc/init | 00:28 |
| sharks_ | Bugger, okay. So I need to delete those from sparse, re-run `build-packages.sh -c`, and make a new image? | 00:28 |
| mal | you can remove manually from device first | 00:28 |
| sharks_ | Oh okay, just delete those files on-device and reboot? | 00:28 |
| mal | yes, the two servicemanager files | 00:29 |
| mal | assuming you had copied the whole files there | 00:29 |
| sharks_ | But not ofono and disabled_services? | 00:29 |
| mal | ofono can go also | 00:29 |
| mal | there is supposed to be this https://github.com/mer-hybris/droid-hal-configs/blob/master/sparse-10/usr/libexec/droid-hybris/system/etc/init/disabled_services.rc | 00:29 |
| sharks_ | Right, hang on and I'll fix that. My bad. Thanks | 00:30 |
| mal | and ecclist.rc | 00:30 |
| mal | https://github.com/mer-hybris/droid-hal-configs/tree/master/sparse-10/usr/libexec/droid-hybris/system/etc/init | 00:30 |
| sharks_ | Still got defaultservicemanager() is null :( | 00:35 |
| ecrn | if they got signal 6 (abort), it should be possible to get coredump | 00:36 |
| sharks_ | logcat -> https://paste.opensuse.org/pastes/9380d46b2baa | 00:36 |
| sharks_ | journal -> https://paste.opensuse.org/pastes/ba4f30f12074 | 00:36 |
| mal | so you removed the /usr/libexec/droid-hybris/system/etc/init/hwservicemanager.rc file? | 00:39 |
| sharks_ | Sure did | 00:39 |
| mal | wondering if trying to start hwservicemanager manually maybe even with gdb would reveal something | 00:45 |
| mal | anyway let's continue tomorrow | 00:47 |
| sharks_ | Yeh, I really appreciate your help mal. Get some rest, I've already had too many late nights due to this! | 00:48 |
| sharks_ | Hey Presto! I've got Lipstick! https://i.imgur.com/Pbehu17.jpeg | 01:54 |
| sharks_ | Touchscreen is dead, can't do anything with it, but at least I have a picture! | 01:54 |
| sharks_ | So `$ cat /sys/class/input/event1/device/name` returns 'sec_touchscreen', presumably that means the touch module exists | 02:01 |
| sharks_ | but `cat /dev/input/event1` has no output | 02:02 |
| ecrn | what was the issue? | 02:02 |
| sharks_ | Issue is the touchscreen is not responding | 02:02 |
| ecrn | but what was the issue that caused the hwservicemanager to abort? | 02:03 |
| sharks_ | Oh - I think there must have been an issue in how I set up the sparse directory. Mal had noticed I included some files from hybris-16 and missed some from hybris-17, so I started again, built a new image and made sure to include all the correct files. After re-creating my udev rules, now hwservicemanager works! | 02:04 |
| ecrn | ok | 02:05 |
| ecrn | the touch screen, well, I have a samsung device too, and had issues with the touch screen too | 02:06 |
| ecrn | but it worked until the lcd panel was turned off and then back on | 02:06 |
| sharks_ | Hang on - I closed the flip cover and opened it again, not only did the screen turn off and on again as it should, but now touch works? | 02:06 |
| ecrn | so the opposite what I had | 02:07 |
| ecrn | :> | 02:07 |
| sharks_ | Lol. Thanks, I'll see if I can go through the setup... | 02:07 |
| sharks_ | Okay, I have the y-axis inverted. This is a common problem on this model | 02:08 |
| sharks_ | If I click the left half of the screen, it registers a button on the right side | 02:08 |
| sharks_ | The guy who ported Lineage to this device had the same issue, I'll see if I can fix it the same way he did | 02:09 |
| sharks_ | wlan seems dead too, not terribly surprising | 02:10 |
| ecrn | there is also touchscreen-invert-x/y option in the device tree | 02:10 |
| sharks_ | Oh? That might be easier, thanks. I'll see if I can find it... | 02:11 |
| ecrn | which kernel version is that? | 02:11 |
| sharks_ | 3.10.108 | 02:12 |
| sharks_ | Okay, device basically works, just need to fix the inverted touchscreen, wifi, audio, etc. https://i.imgur.com/BcpqdkJ.png | 02:15 |
| ecrn | so the device tree properties wasn't probably implemented yet in the 3.10.108, so you need something else | 02:17 |
| ecrn | nice | 02:17 |
| sharks_ | Yeah that's okay, I think I have a solution for the inverted touch. The other issues are probably related to my defconfig still having many warnings, etc. | 02:18 |
| sharks_ | Oh, I actually have audio working already. Great! | 02:23 |
| ecrn | https://github.com/Galaxy-MSM8916/android_kernel_samsung_msm8916/blob/e8d5b0cb00c091722e14660a5512187c5a297a43/drivers/input/touch_disabler.c | 02:24 |
| ecrn | there is this thing in the kernel | 02:24 |
| sharks_ | Do you think it is the reason that touch requires the lcd to turn off and on again? | 02:25 |
| ecrn | maybe, you could try echo 1 > /sys/class/touch_disabler/control ; echo 1 > /sys/class/touch_disabler/enabled | 02:26 |
| ecrn | after reboot | 02:26 |
| ecrn | if it works then yes | 02:26 |
| sharks_ | Okay, just in the process of fixing the inverted touch, give me a few minutes. | 02:27 |
| sharks_ | Awesome, inverted touch is fixed. | 02:31 |
| T42 | <birdzhang> Cheers, haven't seen a new tablet port 🎉 | 02:36 |
| T42 | <birdzhang> in a while | 02:36 |
| sharks_ | Hmm, not sure. The paths are actually /sys/class/touch_disabler/touch_screen/control and /sys/class/touch_disabler/touch_screen/enabled, but the device fell asleep before I could type the commands! Anyway, not too worried about that at the minute, got bigger fish to fry | 02:36 |
| sharks_ | @birdzhang I only wanted to do it because I had two of these tablets in a drawer lol. | 02:37 |
| T42 | <birdzhang> XD | 02:40 |
| ecrn | the module would enable the touch screen on boot but only when touchscreen driver loads and probes the device before the lcd panel driver turns the lcd panel on | 02:40 |
| ecrn | that seems rather fragile condition if not enforced somehow | 02:40 |
| sharks_ | Do you know how to set the UI to tablet mode? | 02:45 |
| T42 | <elros34> I am not aware any tablet mode, what do you want to achieve? | 07:01 |
| sharks_ | This sort of thing? | 07:03 |
| sharks_ | https://www.slashgear.com/img/gallery/turns-out-the-jolla-tablet-is-really-onto-something/jolla-tablet-hands-on-sg-6-600x349.jpg | 07:03 |
| T42 | <elros34> I wonder whether this still works. What pixel ratio do you have set? | 07:05 |
| sharks_ | Honestly I can't even remember where that was set, but it's definitely wrong lol | 07:06 |
| T42 | <elros34> fix that first and maybe ui will change | 07:08 |
| T42 | <elros34> IIRC this should be in droid-config | 07:09 |
| sharks_ | I'm sure I gave it a value at some point early on, but I can't find it in the HADK now. Thanks - I'll have a poke around in droid-config... | 07:09 |
| sharks_ | ah - droid-configs/rpm/droid-config-$DEVICE.spec | 07:10 |
| sharks_ | pixel_ratio 1.0 | 07:11 |
| T42 | <Walid> first i loved the new background image and colors for sfos 5.0.0.29, fingerprint work on sfos 4.6.0 but in 5.0.0.29 no | 12:36 |
| T42 | <fredldotme> mind if I propose a build-time flag to allow overriding the HYBRIS_LINKER_DIR? getauxval(AT_SECURE) seems to screw around with my initial libvulkan + Snaps enablement. | 15:32 |
| mal | @fredldotme you mean also when AT_SECURE is set? | 15:36 |
| T42 | <fredldotme> still investigating, but it seems getenv("HYBRIS_LINKER_DIR") is not taken into account precisely because AT_SECURE is triggering to point it to NULL, which might be related to some setuid debugging happenings with Snaps | 15:38 |
| T42 | <fredldotme> afaict AppArmor on Ubuntu Touch causes AT_SECURE to be set after environment scrubbing, which might be why getenv() is ignored. | 15:54 |
| T42 | <fredldotme> okay no, something is definitely clearing environment variables, even SDL_VIDEODRIVER gets misdetected. | 17:47 |
| T42 | <fredldotme> test scenario: Snapped up vkQuake running confined, with libhybris loaded as a separate snap | 17:47 |
| T42 | <fredldotme> is something mucking around with environ in libhybris? It's the only real difference between my laptop and my smartphone. | 17:47 |
| mal | I think it should not do that, not sure what is happening | 17:55 |
| T42 | <xo0on> @Thaodan thanks! Do you actually have the xperia 5 II? did you read about the folks getting rid of their echos on 10 III by installing over A12? | 20:26 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
