| T42 | <A_T_R> WARNING: CONFIG_SECURITY_SELINUX_BOOTPARAM is invalid | 07:05 |
|---|---|---|
| T42 | <A_T_R> how to propely enable selinux enable for hybris-17 | 07:05 |
| T42 | <A_T_R> 17.1 | 07:05 |
| T42 | <A_T_R> how to propely enable selinuxfor hybris-17 (edited) | 07:05 |
| T42 | <A_T_R> # CONFIG_AUDIT is not set | 07:06 |
| T42 | <A_T_R> BoardConfig.mk | 07:07 |
| T42 | <A_T_R> BOARD_KERNEL_CMDLINE += androidboot.selinux=permissive | 07:07 |
| T42 | <A_T_R> BOARD_KERNEL_CMDLINE += selinux=0 | 07:07 |
| T42 | <A_T_R> .config | 07:07 |
| T42 | <A_T_R> # CONFIG_AUDIT is not set | 07:07 |
| T42 | <Verevka86> https://github.com/SailfishOS-sagit/android_kernel_xiaomi_msm8998/commit/c12b32026cc648008dadc80bf5febe7a4cfc0a2d | 07:33 |
| T42 | <Verevka86> Set 0 if u want to disable selinux (re @A_T_R: WARNING: CONFIG_SECU...) | 07:33 |
| T42 | <elros34> @A_T_R obviously to enable selinux you would need to revert your config_audit change and selinux=0 but that will not help anyway. You have already tried it. You need to figure out why kernel fail. For example try flashing boot.img to see if it change anything, try with minimal defconfig changes or different compiler/kernel. | 09:00 |
| T42 | <elros34> Maybe also provide host dmesg I asked from the first boot after flashing kernel | 09:01 |
| T42 | <A_T_R> cmdline | 09:02 |
| T42 | <A_T_R> androidboot.selinux=permissive | 09:02 |
| T42 | <A_T_R> selinux=0 | 09:02 |
| T42 | <A_T_R> now compiling with config_audit=y | 09:02 |
| T42 | <elros34> what you want to achieve now? disable selinux "selinux=0" or selinux mode set to "permissive"? | 09:04 |
| T42 | <A_T_R> enable selinux | 09:05 |
| T42 | <elros34> by adding selinux=0?:) | 09:05 |
| T42 | <A_T_R> oh ,it set selinux disable . | 09:10 |
| T42 | <elros34> also revert your CONFIG_DEFAULT_SECURITY_SELINUX=y removal | 09:10 |
| T42 | <A_T_R> .config | 09:12 |
| T42 | <A_T_R> CONFIG_DEFAULT_SECURITY_SELINUX=y | 09:12 |
| T42 | <elros34> ok then, I only see your defconfig which has it removed | 09:13 |
| T42 | <A_T_R> host dmesg https://pastebin.com/3Qrr5Dzp | 09:15 |
| T42 | <A_T_R> problem missing initrd in kernel | 09:16 |
| T42 | <A_T_R> * kernel size = 15620417 bytes (14.90 MB) | 09:16 |
| T42 | <A_T_R> ramdisk size = 20 bytes (0.00 MB) | 09:16 |
| T42 | <elros34> so where in this dmesg first boot after flashing starts? Did you export needed env variable before building kernel? | 09:17 |
| T42 | <A_T_R> Did you export needed env variable before building kernel? | 09:20 |
| T42 | <A_T_R> no i guess | 09:20 |
| T42 | <elros34> how about my first question, line 1, 8 or 16? | 09:20 |
| T42 | <A_T_R> line 16 | 09:21 |
| T42 | <A_T_R> line 1-14 while device in fasbtoot | 09:21 |
| T42 | <A_T_R> abooting -i hybris-boot.img | 09:22 |
| T42 | <A_T_R> https://pastebin.com/vQz4rJr4 | 09:22 |
| T42 | <elros34> yeah, set that env variable | 09:24 |
| T42 | <A_T_R> what is that env variable ? | 09:24 |
| *** ggabriel_ is now known as ggabriel | 09:24 | |
| T42 | <elros34> the same we were talking yesterday and it's in hadk-faq hybris-17 section. If it's missing then initramfs is empty | 09:25 |
| T42 | <A_T_R> my mistaek i did export TEMPORARY_DISABLE_PATH_RESTRICTIONS=true | 09:28 |
| T42 | <elros34> try making any change in kernel then export it just before make hybris-boot and show your output from that | 09:31 |
| T42 | <A_T_R> cpio missing in my machine | 09:31 |
| T42 | <elros34> it's not allowed to be used without that env | 09:32 |
| T42 | <A_T_R> [ 99% 941/946] Making initramfs : out/target/product/X00TD/obj/ROOT/hybris-boot_intermediates/boot-initramfs.gz | 09:37 |
| T42 | <A_T_R> /bin/bash: cpio: command not found | 09:37 |
| T42 | <elros34> "which cpio" can't find it also? | 09:41 |
| T42 | <A_T_R> from terminal https://pastebin.com/cGYGJmEd | 09:41 |
| T42 | <hmmrock> I am curious to know that what is sailfish OS | 09:49 |
| T42 | <hmmrock> Is it like Ubuntu touch??? | 09:49 |
| Thaodan | Whever audio is in use I get warnings like SM8150 Media1: ASoC: no backend DAIs enabled for SM8150 Media1 | 10:49 |
| Thaodan | I guess that's normal? | 10:49 |
| Thaodan | AOSP 10, Sony Xperia I | 10:49 |
| T42 | <A_T_R> can anyone check the initramfs file size -here```hadk/out/target/product/X00TD/obj/ROOT/hybris-boot_intermediates/boot-initramfs.gz``` | 11:36 |
| T42 | <elros34> ~0.5MB. So have you checked that "which cpio" returns something? | 11:38 |
| T42 | <A_T_R> ABUILD_SDK [x00td] atr@atr-G41D3C:~/hadk$ which cpio | 11:38 |
| T42 | <A_T_R> HABUILD_SDK [x00td] atr@atr-G41D3C:~/hadk$ | 11:38 |
| T42 | <A_T_R> "which cpio" returns nothing | 11:39 |
| T42 | <A_T_R> in habuild | 11:39 |
| T42 | <elros34> which ubuntu version do you use? On older it's installed. Try to install it | 11:40 |
| T42 | <A_T_R> my host : ubutnu version 20.04.2 | 11:40 |
| T42 | <A_T_R> habuild ubuntu- latest ubuntu-focal | 11:40 |
| T42 | <A_T_R> Ubuntu-chroot | 11:41 |
| T42 | <A_T_R> https://releases.sailfishos.org/ubu/ubuntu-focal-20210531-android-rootfs.tar.bz2 | 11:41 |
| T42 | <A_T_R> HABUILD_SDK [x00td] atr@atr-G41D3C:~/hadk$ apt install cpio | 11:43 |
| T42 | <A_T_R> E: Could not open lock file /var/lib/dpkg/lock-frontend - open (13: Permission denied) | 11:43 |
| T42 | <A_T_R> E: Unable to acquire the dpkg frontend lock (/var/lib/dpkg/lock-frontend), are you root? | 11:43 |
| T42 | <A_T_R> HABUILD_SDK [x00td] atr@atr-G41D3C:~/hadk$ sudo apt install cpio | 11:43 |
| T42 | <A_T_R> sudo: account validation failure, is your account locked? | 11:43 |
| T42 | <A_T_R> HABUILD_SDK [x00td] atr@atr-G41D3C:~/hadk$ | 11:43 |
| T42 | <elros34> it's known issue, you need to fix it. IN the mean time just chroot /path/to_ubuntu/ and install it | 11:50 |
| T42 | <A_T_R> in platform_sdk | 11:59 |
| T42 | <A_T_R> sudo chroot $PLATFORM_SDK_ROOT/sdks/ubuntu/ /bin/bash | 11:59 |
| T42 | <A_T_R> chrooted | 11:59 |
| T42 | <A_T_R> in chroot | 12:00 |
| T42 | <A_T_R> apt-get install cpio | 12:00 |
| T42 | <A_T_R> error - https://pastebin.com/VebNncGe | 12:00 |
| T42 | <A_T_R> finally install cpio in chroot... | 12:49 |
| T42 | <A_T_R> compiled new kernel - booted to ui | 12:57 |
| T42 | <birdzhang> @A_T_R https://t.me/sailfishos_porters/127035 | 14:18 |
| T42 | <A_T_R> how to install *.rpm file for test ? | 16:12 |
| T42 | <elros34> tap on it:) | 16:12 |
| T42 | <elros34> or rpm -U path/to/rpm | 16:15 |
| T42 | <A_T_R> i used to build rpm/dhd/helpers/build_packages.sh --mw=https://github.com/SailfishOS-sagit/mapplauncherd && -i image ...atleast i could setup initial language the same issue looping the loading icon | 16:40 |
| T42 | <A_T_R> then* | 16:40 |
| T42 | <elros34> we are debugging issues by getting logs so you should get journalctl instead | 17:35 |
| T42 | <A_T_R> https://pastebin.com/tZNS2Vpw | 17:36 |
| T42 | <A_T_R> journalctl | 17:36 |
| T42 | <A_T_R> library "libGLESv2_adreno.so" not found | 17:38 |
| T42 | <A_T_R> Sep 08 19:49:54 MAXPROM1 invoker[6796]: library "eglSubDriverAndroid.so" not found | 17:38 |
| T42 | <A_T_R> Sep 08 19:49:55 MAXPROM1 invoker[6799]: library "libGLESv2_adreno.so" not found | 17:38 |
| T42 | <A_T_R> Sep 08 19:49:55 MAXPROM1 invoker[6799]: library "eglSubDriverAndroid.so" not found | 17:38 |
| T42 | <A_T_R> /odm@SailfishFreenodeIRCBridgeBot | 17:40 |
| mal | those messages are quite normal | 17:42 |
| Thaodan | I get them too | 17:50 |
| Thaodan | annoying but normal | 17:50 |
| Thaodan | @b100dian: Testd your ngfd PR works better now. | 17:51 |
| mal | @A_T_R are you able to get any output from /usr/libexec/droid-hybris/system/bin/logcat ? | 17:53 |
| T42 | <b100dian> Thaodan: anything still broken? | 18:02 |
| Thaodan | see comments on pr | 18:02 |
| Thaodan | vibration on incoming notification is a bit intense or other notifications are delayed my phone is vibrating for minutes now | 18:05 |
| Thaodan | the vibration stops when another apps uses the daemon | 18:06 |
| Thaodan | ah the long vibration comes from unlocking my phone with the fingerprint sensor | 18:10 |
| rinigus | speaking of annoying messages - mal, have you had a chance to look for /dev/sailfish/home probing by encsfa-fpd? | 18:37 |
| T42 | <b100dian> Thaodan I have one vibration that won't stop, that is alarm. I have it set as constant for 4 seconds, since ngfd fails to stop it sometimes. is fingerprint unlocking a constant effect took? | 18:47 |
| Thaodan | b100dian: Sorry but I didn't change the values you had set for custom after I took them from your port. | 19:20 |
| mal | rinigus: no, sorry, forgot about it | 19:21 |
| T42 | <b100dian> ah, but I don't have fingerprint reader | 19:21 |
| Thaodan | b100dian: All vibrations are now set to 240s duration that's the issue | 19:22 |
| T42 | <b100dian> regardless of config, that is? | 19:23 |
| Thaodan | that vibration is the touch_strong I guess that is what any app can request. | 19:25 |
| Thaodan | I think it's regardless of config since every vibration is 240s now. | 19:25 |
| Thaodan | _CUSTOM also isn't really saying much. If possible it would be great do handle custom and sine waveform vibrators transparent. | 19:27 |
| T42 | <b100dian> yes, but we don't know beforehand how a driver is configured to treat custom ints. or maybe I misinterpret your suggestion | 19:32 |
| T42 | <b100dian> I'll be at the keyboard in some mins bte, cannot look at the config now | 19:32 |
| T42 | <b100dian> *30 mins | 19:33 |
| Thaodan | My suggestion is to do the conversion of those custom ints in ngfd is possible or at least keep the configuration as easy as possible/non-device specific. | 19:41 |
| Thaodan | and I noticed that you don't have your branch rebased against master, please do that while testing your changes. | 19:50 |
| T42 | <b100dian> I have that in my fork's master. was OK with the last changes on my device but definitely there's something more to it | 19:56 |
| Thaodan | I mean upstream master. | 19:57 |
| T42 | <b100dian> I have the last commit from 27 days ago merged into https://github.com/sailfishos-on-tucana/ngfd/commits/master (not rebased) if that helps | 19:59 |
| T42 | <b100dian> back at the keyboard: looking at the logs, the only effect in your config with 24000 duration is ringtone. Does changing that also change the fingerprint unlock duration? | 20:02 |
| T42 | <b100dian> I ask because, the reason I implemented the CACHE_EFFECT define was that my device was effectively using the last effect uploaded even when I said to it to play another one. Maybe this is the case here, the last effect uploaded is ringtone and its duration is re-used? | 20:03 |
| Thaodan | ah ok, I was only looking at that feature branch. I don't like to change master and prefer to be in sync with upstream. I try again now and provide new logs. | 20:04 |
| Thaodan | Looked again. The issue is that the default duration value is now 240ms. | 20:11 |
| Thaodan | You can test the long vibration type in the components app under haptics. | 20:11 |
| T42 | <b100dian> Sorry I cant test right, I broke my device today (in an unrelated way:). So NGF_DEFAULT_DURATION has been 240ms since long. How is this an issue, is it that it translates to 240s? if yes I need to look into that driver again | 20:17 |
| Thaodan | of I'm sorry. Ok than I maybe mixed something up since also those notications that had the warning are so long. | 20:20 |
| T42 | <b100dian> which warning? | 20:26 |
| T42 | <b100dian> > ffmemless: short_DURATION not defined, using 240ms ? | 20:26 |
| Thaodan | I checked again now with the last commit you made all notications duration's are really long. E.g. I pressed the stop watch button and the effect was still there after 8min | 20:38 |
| rinigus | mal: no problem, I forgot to remind as well :) | 20:47 |
| T42 | <b100dian> Thaodan: do you have "upload constant effect, length" message in dmesg like here https://github.com/mer-hybris/android_kernel_sony_msm/blob/hybris-sony-aosp/LA.UM.7.1.r1/drivers/input/misc/qti-haptics.c#L859 ? | 20:54 |
| T42 | <b100dian> For later: if you could boot into AOSP with magisk and as root strace vendor.qti.hardware.vibrator@1.2-service, can you see EVIOCSFF ioctls? | 20:56 |
| T42 | <b100dian> (one thing I did to simulate dual boot was to remove fileencryption=ice from /vendor/etc/fstab.qcom and then just flash boot.img and hybris-boot.img alternatively. The former needs magisk flashed on top if you need root) | 20:58 |
| Thaodan | sorry I don't have aosp running just SailfishOS | 21:16 |
| Thaodan | I don't have those upload constant effect, length in dmesg | 21:19 |
| Thaodan | @b100dian: I ran the service with strace under SFOS without ngfd running of course: https://dpaste.org/3NyN | 21:29 |
| T42 | <b100dian> Those would have been interesting while the actual vibrator is in use under android.. | 21:59 |
| T42 | <b100dian> don't remember where the components app comes from (store?). It might use qt5-feedback-haptics-droid-vibrator or qt-mobility-haptics-ffmemless which I didn't touch. | 22:01 |
| T42 | <b100dian> but fingerprint unlock is definitely a ngfd problem:-S | 22:01 |
| T42 | <b100dian> according to https://www.kernel.org/doc/html/v4.11/admin-guide/dynamic-debug-howto.html#controlling-dynamic-debug-behaviour you can try # echo 'file <qti-haptics.c> +p'>/sys/kernel/debug/dynamic_debug/control | 22:06 |
| Thaodan | I use ffmemless native vibrator all apps use qt5-feedback-haptics vibrator with native or droid backend depending on the point | 22:07 |
| T42 | <b100dian> here's what I use https://github.com/sailfishos-on-tucana/droid-hal-version-tucana/blob/master/rpm/droid-hal-version-tucana.spec#L13 https://github.com/sailfishos-on-tucana/droid-config-tucana/blob/efe825957b2fd5f3a20c5830cac2af6390571112/patterns/patterns-sailfish-device-adaptation-tucana.inc#L29 | 22:11 |
| T42 | <b100dian> basically ffmemless on Qt side to (which doesn't work. so you get no keyboard vibration for example). But at least it doesn't act up | 22:11 |
| T42 | <b100dian> (I was planning to add custom effect support for Qt but didn't got the time) | 22:14 |
| Thaodan | This should be the same if I'm not wrong. You don't need to add support on Qt side since Qt has a ngfd plugin. | 22:17 |
| T42 | <elros34> qt apps can use ngfd but they do not have to, components example use qt not ngfd | 22:22 |
| T42 | <elros34> In lipstick you can find ngf usage in qml. They are not same | 22:24 |
| Thaodan | Everything that uses Qt-Feedback uses ngfd at least. | 22:27 |
| T42 | <elros34> how? | 22:27 |
| T42 | <elros34> qtfeedback use qt plugin not ngfd | 22:27 |
| T42 | <elros34> here is memless plugin for qtfeedback: https://github.com/sailfishos/qt-mobility-haptics-ffmemless/blob/master/qfeedback.cpp As you can see it use device directly without ngfd | 22:31 |
| T42 | <b100dian> So either I get vibration for keyboard from this discussion if Qt indeed has a ngfd plugin, or your errors Thaodan also come from Qt native vibrator:) | 22:31 |
| T42 | <b100dian> yup, that was on my list of pathcing this., year | 22:32 |
| T42 | <elros34> qt feedback native/droid or ffmemless do not use ngfd | 22:32 |
| T42 | <b100dian> Yes, but they may trigger vibrations on Thaodan's device, would be another possible cause for >8mins vibra | 22:33 |
| T42 | <elros34> depends on how he test plugin, if using components galery example then yeah | 22:34 |
| Thaodan | that is what I use https://github.com/sailfishos/libngf-qt and ngfd-plugin-native-vibrator | 22:35 |
| Thaodan | test was done with components galery | 22:35 |
| T42 | <elros34> ah so there is yet another plugin:) so I was wrong | 22:37 |
| Thaodan | libngf-qt replaces qt5-feedback-haptics-native-vibrator | 22:38 |
| T42 | <b100dian> that is actually good news for me:p if you wait long enough.. | 22:38 |
| T42 | <b100dian> but I need to see how to actually test that integration | 22:39 |
| T42 | <b100dian> Thaodan other than building that do you have any config pointers how to make use? | 22:42 |
| Thaodan | How to make use? | 22:42 |
| T42 | <b100dian> how to use that plugin | 22:43 |
| Thaodan | instead of installing qt5-feedback-haptics-droid-vibrator install libngf-qt | 22:49 |
| Thaodan | that's it | 22:50 |
| T42 | <b100dian> ok, I'll try that when I get to boot | 23:17 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!