Sunday, 2020-12-13

T42<Heng %lastname%> my touchscreen appears as /dev/input/mouse000:07
T42<Heng %lastname%> will it still work in sailfish with the lipstick config00:08
T42<linusdan> guys, is it normal to type droid-hal-init in telnet and boot into TWRP?00:39
T42<linusdan> P.S .: months ago I started updating the P3110 to the base of android 6 but did not know how to continue after compiling rootfs and abandoned it. I decided to come back today and I have the same problem: what to do now?00:42
T42<elros34> did you start droid-hal-init same ways as in Also it can be only started once00:44
T42<linusdan> no, i just typed droid-hal-init after entering telnet00:45
T42<elros34> did you masked it first?00:45
T42<linusdan> no00:46
T42<elros34> so that is your issue, it can be started only once00:47
T42<linusdan> how do I do that?00:49
T42<elros34> systemctl mask droid-hal-init, reboot then your can try to start it manually same as in BTW what issue you are trying to debug so you need to start it manually?00:52
T42<elros34> maybe you would need to  mask user@100000 also but I assume it will be not started if droid-hal-init is masked.00:55
T42<linusdan> ok, i will try :)00:57
T42<linusdan> about issue, after installing and restarting the tablet, it goes black screen00:57
T42<linusdan> it does not appear the interface00:57
T42<elros34> then more useful would be journalctl and logcat00:57
T42<elros34> this is old device so you can try with EGL_PLATFORM=fbdev00:59
T42<elros34> instead hwcomposer00:59
T42<linusdan> can you change without having to compile everything again?01:01
T42<elros34> sure, that env variable is  defined in /var/lib/environment/compositor/droid-hal-device.conf01:02
T42<linusdan> thanks! I'll test.01:04
T42<linusdan> I'm having to install rootfs again, I'll send you news as soon as I can :)01:04
T42<Lucas Thomas %lastname%>
T42<Samuel Martin %lastname%>
T42<Heng %lastname%> Spam02:29
T42<Spidey_SPIDEY_Spidey> @Manuvaidya81 [Ofcourse there are apps in fdroid like WhatsAp …], In my usage I need all gapps03:12
T42<ankaos> hı. ı build hybris-16.0. but not open screen device. ı send journalctl13:32
T42<ankaos> mal: where is the problem?13:35
malyou seem to have selinux enabled13:36
T42<ankaos> mal: ı dont see kernel out config selinux13:38
malwho did you disable selinux? did you do it the wrong way?13:48
maland then in kernel commandline selinux=013:49
T42<ankaos> mal: yes ı read this CONFIG_AUDIT=y CONFIG_SECURITY_SELINUX_BOOTPARAM=y13:50
malso why did you say "ı dont see kernel out config selinux" ?14:16
T42<ankaos> ./hybris/mer-kernel-check/mer_verify_kernel_config out/target/product/oxygen/obj/KERNEL_OBJ/.config -> not read CONFIG_SECURITY_SELINUX_BOOTPARAM=y14:27
T42<linusdan> good morning @elros34 and everyone :)14:34
T42<linusdan> I managed to copy the journalctl log to the tablet storage today:14:34
T42<linusdan> droid-hal-init continues to have the same problem after masking14:36
T42<linusdan> I don't quite understand the log, but the most common error was the "[grilio-socket] ERROR! Can't connect to RILD: No such file or directory"14:37
T42<linusdan> What would that be?14:38
mal@ankaos you seem to have some issue with mounting, check using ls -l /firmware and ls -l /dsp and ls -l /persist14:47
T42<elros34> @linusdan I don't see droid-hal-init at all in your logs so it's hard to guess14:59
T42<elros34> unmask it, reboot, then grab dmesg, journalctl --no-page -b0 and logcat14:59
T42<linusdan> ok15:01
T42<elros34> you seems to build some old sfos release. Which one is that?15:02
T42<linusdan> I'm using version
T42<ankaos> sh-3.2# ls -l firmware15:10
T42<ankaos> lrwxrwxrwx    1 root     root            20 Dec 13  2020 firmware -> /vendor/firmware_mnt15:10
T42<ankaos> sh-3.2# ls -l dsp15:10
T42<ankaos> lrwxrwxrwx    1 root     root            11 Dec 13  2020 dsp -> /vendor/dsp15:10
T42<ankaos> mal: answer :)15:11
T42<ankaos> sh-3.2# ls -l persist15:15
T42<ankaos> lrwxrwx--x    1 root     system          19 Dec 13  2020 persist -> /mnt/vendor/persist15:15
malis /vendor mounted correctly?15:17
T42<ankaos> yes15:24
T42<linusdan> @elros34 journalctl:
T42<linusdan> dmesg:
T42<elros34> is that log with EGL_PLATFORM=fbdev? What about logcat?15:37
T42<linusdan> yes, with fbdev :)15:49
T42<linusdan> oh wait, i forget of logcat 😂15:49
T42<elros34> also check lsmod because I guess some init*rc file try to insmod  module from /system and fail. Check what modules you have in /lib/modules/`uname -r`/15:58
T42<linusdan> it says that the logcat command is not found16:00
T42<elros34> full path is /usr/libexec/droid-hybris/system/bin/logcat16:01
T42<linusdan> logcat:
T42<linusdan> lsmod:
T42<elros34> so you should modprobe both omap and pvr modules16:25
T42<elros34> you can create conf files in /etc/modules-load.d/ so it happen automatically at boot time16:30
T42<linusdan> OK16:35
T42<linusdan> and how do I create these files with the configuration?16:35
T42<Heng %lastname%> soo...I set the droid-hal-device lipstick options to use the touchscreen at /dev/input/event4 but it is still not functionign16:37
T42<Heng %lastname%> what should I try to get it working16:37
T42<Heng %lastname%> Lipstick Options: `LIPSTICK_OPTIONS="-plugin evdevkeyboard:keymap=/usr/share/qt5/keymaps/droid.qmap -plugin evdevtouch:/dev/input/event4"`16:43
T42<elros34> @linusdan just put name of module without .ko. You can find examples probably even in your host system16:56
T42<elros34> Heng Ye: is that really touchscreen device? evdev_trace (mce-tools package) will print all your evdev devices and also events they generates. QT_LOGGING_RULES="qt.qpa.input=true" could also be useful17:02
T42<Heng %lastname%> @elros34 ...which means I should get wifi working first?17:07
T42<elros34> no, telnet via usb which is not hard and it is very usefull17:07
T42<Heng %lastname%> yes i have a console17:07
T42<Heng %lastname%> so reverse USB tethering? @elros3417:08
T42<elros34> ah, then you can find in HADK pdf how to get internet working17:08
T42<elros34> or download package in sdk target and copy to device17:09
T42<Heng %lastname%> @elros34 [ah, then you can find in HADK pdf how to get i …], It seems that the /dev/input/event number changes each boot19:03
T42<Heng %lastname%> now it's /dev/input/event719:03
T42<Heng %lastname%> evdev_trace -t gives ABS_MT_POSITION, ABS_MT_TOOL, ABS_MT_TOUCH, BTM_TOUCH, ABS_X, ABS_Y19:04
T42<Heng %lastname%> does this mean that it detected as touchscreen? @elros3419:04
T42<Heng %lastname%> Need udev rules for consistent path ? @elros3419:07
T42<elros34> Heng Ye: yes you can use udev, you can reuse this: KERNEL=="event*", SUBSYSTEM=="input", ATTRS{name}=="atmxt-i2c", ACTION=="add", TEST!="20:02
T42<elros34> /dev/touchscreen", SYMLINK+="touchscreen", ENV{ID_INPUT_KEY}="0"20:02
T42<elros34> you don't have ABS_MT_PRESSURE but maybe BTN_TOUCH is enough for evdevtouch20:04
T42<Heng %lastname%> @elros34 [Heng Ye: yes you can use udev, you can reuse t …], What is the priority (filename) i should set?20:04
T42<Heng %lastname%> @elros3420:04
T42<elros34> not sure it is relevant, I use 61-touchscreen.rules20:05
T42<Heng %lastname%> @elros34 [Heng Ye: yes you can use udev, you can reuse t …], soo...that didn't work, theres no /dev/touchscreen20:08
T42<elros34> did you change ATTRS{name}?20:09
T42<Heng %lastname%> @elros34 [did you change ATTRS{name}?], so I did the attribute walk, there are 6 levels of parent devices20:50
T42<Heng %lastname%> which one should I use @elros3420:50
T42<Heng %lastname%> the device itself does not have a name @elros3420:54
T42<elros34> so use any other attrs which is unique to this device20:54
T42<Heng %lastname%> @elros34 [so use any other attrs which is unique to this …], there are none20:54
T42<Heng %lastname%> the first parent device shares with 4 other inputs20:55
T42<elros34> oh :)20:55
T42<Heng %lastname%> @elros34 [oh :)], so...maybe use the Persistent Block Naming?20:57
T42<Heng %lastname%> it's at /devices/pci0000:00/808622C1:05/i2c-5/i2c-ELAN2562:00/0018:04F3:2562:0001/input/input25/event2020:57
T42<Heng %lastname%> or should i keep trying to do something with the udev rules20:57
T42<Heng %lastname%> @elros3420:57
T42<Heng %lastname%> the device itself has no ATTRs20:57
T42<elros34> not sure, maybe first just rty whether it works at all, so put correct node in droid-hal-device.conf and restart user@100000.service or maybe just lipstick.service20:59
T42<Heng %lastname%> @elros34 [not sure, maybe first just rty whether it work …], well...then I have to have root21:00
T42<elros34> that shoudn't be a problem. Using telnet you are already root otherwise use devel-su to gain root21:02
T42<Heng %lastname%> yay21:06
T42<Heng %lastname%> it works21:08
T42<Heng %lastname%> so I should try the persistent block naming? @elros3421:09
T42<elros34> not sure what do you mean. If you want to use /devices/pci0000*1/input/input25/event20 then I don't think it will work. qevdevtouchmanager.cpp expects "/dev/" string21:11
T42<Heng %lastname%> @elros34 [not sure what do you mean. If you want to use …], hmm...what else should i try then>21:11
T42<Heng %lastname%> @elros3421:11
T42<Heng %lastname%> @elros34 this looks interesting
T42<Heng %lastname%> is this relevant?21:17
T42<elros34> never used it but looks like perfect solution for you21:18
T42<Heng %lastname%> @elros34 [never used it but looks like perfect solution …], yeah but it's going to get confused with the digitizer @elros3421:39
T42<Heng %lastname%> is there any other way to get the udev to work without having ATTRs for the device?21:39
T42<elros34> evcap use device capabilities. I am pretty sure you will find unique ones. Does evcap for you touchscreen and digitizer? returns same caps?21:41
T42<Heng %lastname%> @elros34 [evcap use device capabilities. I am pretty sur …], the evdev_trace shows the same capabilites when i use touchscreen and digitizer21:41
T42<elros34> oh so use whatever you can and write simple script, even that long path you provided earlier21:46
T42<Heng %lastname%> @elros34 [oh so use whatever you can and write simple sc …], well even that is dynamic21:46
T42<Heng %lastname%> with the input20 and event2021:46
T42<elros34> haha21:46
T42<elros34> if that is really true then amend kernel driver or built driver touchscreen/digitizer as module and load it in correct order...21:49
T42<Heng %lastname%> they do slightly have a difference in evcap bitmask21:49
T42<Heng %lastname%> perhalps that will help21:49
T42hypnz was added by: hypnz21:54
T42<Heng %lastname%> yay fixed touchscren22:15
T42<Heng %lastname%> but once booted, it shuts down saying battery is empty, while actually its not empty22:26
T42<Heng %lastname%> is there a service or config i can modify for this? ^23:31
T42<elros34> mce takes care of battery. You can stop mce and run it in verbose mode to figure out what is wrong: mce -Tq -l"modules/battery-udev.c:*"23:41

Generated by 2.17.1 by Marius Gedminas - find it at!