mal | not sure, maybe some firmware issues or other problems during the boot | 00:00 |
---|---|---|
mal | logs should tell | 00:00 |
T42 | <edp_17> mal: I'll check. | 00:02 |
T42 | <edp_17> I've sorted the other issue. Now the mobile data works. 😃 | 00:02 |
T42 | <edp_17> Although, none of the camera works. Both shows black image only. (there is no viewfinder in csd) | 00:05 |
T42 | <edp_17> @elros34 : Can you help me with the MTP, please? I've wanted to apply this patch https://github.com/elros34/android_kernel_motorola_msm8960-common/commit/35eca6b56d13260ab5997aaeba12184b55636282 but my drivers/usb/gadget/android.c is very different and don't have drivers/usb/gadget/f_fs.c. | 01:18 |
T42 | <edp_17> When select MTP on device I got this error: https://paste.ubuntu.com/p/tvGMNTvg7b/ (as you see I tried to issue the mount command manually as well) | 01:18 |
T42 | <edp_17> Do you have idea how to fix it? | 01:18 |
T42 | <edp_17> mal: the lightsensor should be ok because the value changes in /sys/class/lightsensor/switch_cmd/lightsensor_file_state (It has high value (746) when put under light, otherwise lower (119) | 01:31 |
T42 | <elros34> @edp_17 that commit is wrong (I need to amend it) as I am not the author of those changes. Anyway you should be able to add these bunch of functions manually. The hard part is only functions_store() method which you probably need to rewrite for your kernel | 02:05 |
T42 | <Verevka86> hybris17.1 works for me only with audit turned off in defconfig, what are the consequences? Until I have it working - ril, wifi, camera ... | 05:45 |
T42 | <Braillynn> Finally got Sailfish OS installed. It doesn't seem to work with Mint Mobile. | 05:47 |
T42 | <Ljroc> You probably need to set an apn | 06:02 |
rinigus | mal : good morning! I see. Anyway I can fix the issue with building it? | 06:32 |
T42 | <edp_17> @elros34 : Thanks. What's wrong with that commit? I don't remember now, as it was long time ago, but I might use that in my S2 port where mtp works. | 07:28 |
rinigus | mal: found it - missing include in Android.mk of droidmedia... | 07:42 |
mal | rinigus: yes, it needs a conditional include there | 07:52 |
rinigus | mal: do you want me to submit PR to droidmedia or you would make it yourself? | 08:06 |
*** Thaodan_ is now known as Thaodan | 08:07 | |
mal | rinigus: did you test the change worked and fixed the audio policy issue on your device | 08:14 |
rinigus | mal: going on with it. it will take some time, will probably manage towards the evening | 08:15 |
T42 | guan %lastname% was added by: guan %lastname% | 08:16 |
mal | rinigus: https://github.com/sailfishos/droidmedia/pull/75 | 08:23 |
mal | oops | 08:26 |
mal | hmm, maybe messed up something with the env.mk file | 08:26 |
rinigus | mal: you just dropped it. it is an empty file | 08:27 |
mal | yes, I was only supposed to add it to .gitignore not drop it | 08:28 |
rinigus | mal: could be a side effect. otherwise we don't ignore | 08:28 |
mal | rinigus: it seem gitignore is not working as I thought it would | 08:30 |
deathmist_ | @Verevka86 so you got hybris 17.1 into GUI? was it an aarch64 build? | 08:38 |
T42 | <Verevka86> @deathmist_ [@Verevka86 so you got hybris 17.1 into GUI? wa …], Yes | 08:39 |
deathmist_ | could you upload repos? I'm struggling to understand why I can't even run logcat on my fresh 17.1 port | 08:40 |
mal | deathmist_: so have you fixed selinux? | 08:46 |
deathmist_ | I don't know how, @Verevka86 did you have to do anything about SELinux on 17.1? | 08:49 |
deathmist_ | I tried to naively put it to permissive mode but that broke things even more and still didn't let me run logcat | 08:51 |
mal | deathmist_: you have the symlinks somewhere under /etc/selinux, replace those with the actual files, at least if you didn't build using packaged system and vendor | 09:01 |
T42 | <elros34> @edp_17 author not the code itself | 09:06 |
T42 | <Verevka86> @mal [deathmist_: you have the symlinks somewhere un …], https://github.com/SailfishOS-sagit/android_kernel_xiaomi_msm8998/commit/7b844427909f742c1cac5923cb0fb690e52cc226 | 09:07 |
T42 | <Verevka86> https://github.com/SailfishOS-sagit/android_device_xiaomi_msm8998-common/commit/2a0ba6e60bd95b14f054a599c344afea9e60f7e6 | 09:07 |
swanux | Could you help me with HAL packaging? (chapter 7.2, but the PDF seems outdated and incomplete, especially at this part). Here's the error: https://pastebin.com/c74uYyZV | 09:41 |
T42 | <Verevka86> @swanux [Could you help me with HAL packaging? (chapter …], https://github.com/SailfishOS-sagit/droid-hal-sagit/blob/master/droid-hal-sagit.spec#L6 | 09:44 |
deathmist_ | anux: you simply cannot change some of them on all kernels; some are armv7 specific while others only appear on x86 kernels. tbh the script could really use a rewrite ;) | 10:20 |
deathmist_ | and ofc some of the options can only be set on newer/older kernels | 10:20 |
rinigus | mal: minimediaservice fails to start now with library "libaudiopolicyservice.so" not found | 10:25 |
rinigus | mal: that lib is in 32 bit section as /usr/libexec/droid-hybris/system/lib/libaudiopolicyservice.so | 10:26 |
rinigus | as probably expected for 32 bit minimediaservice | 10:27 |
*** monich_ is now known as monich | 10:51 | |
swanux | deathmist_ I see, thank you. Regarding rewrite a lots of tools need them related to linux on phones I guess. | 11:36 |
swanux | In this case, how can I get around this one? https://pastebin.com/MB7pTWwk | 11:36 |
T42 | <edp_17> mal: No wonder I couldn't find /sys/class/leds. On this device there is no leds. 😃 | 11:49 |
T42 | <edp_17> I'll try to use the two softkeys for that (just like on the Galaxy S2). | 11:49 |
T42 | <edp_17> I've tested all sensors on Android (the same what I use as base on sfos) and all works. I don't understand why light sensor goes amok. | 11:51 |
T42 | <edp_17> On the device I also have barometer. How can I utilise that in sfos? piggz, do you remember that? | 11:52 |
mal | there is no app for barometer but it's supported by sensorfw and qtsensors | 12:22 |
T42 | <edp_17> mal: thanks. | 13:21 |
mal | I have patched messwerk app with barometer support, I should make PR to upstream | 13:28 |
T42 | <elros34> @edp_17 just for the record, I have updated commit author here: https://github.com/elros34/android_kernel_motorola_msm8960-common/commit/b9827c091fd99c3d8929ca7e4bf10c8bcde6c7ad | 13:32 |
T42 | <edp_17> mal: That sounds amazing. Do you have an rpm to share? 😉 (Just until it goes live.) | 13:32 |
T42 | <edp_17> @elros34 : Thanks. I was wrong about MTP. As the n7000 has the same (very similar) kernel to the i9100, I could apply the mtp fix from that work. I am about to try it on device. Fingers crossed. 🤞 | 13:33 |
T42 | <elros34> ok | 13:34 |
rinigus | @edp_17: I have used that messwerk; it may be somewhere at OBS, in mal's repos. | 13:37 |
mal | @edp_17 https://build.sailfishos.org/package/show/home:mal:apps/Messwerk | 13:37 |
T42 | <edp_17> Since I turned the light sensore off, the device is much snappier and all apps do work with firejail correctly. (Well, there are some error message in the logs but still work ok so far.) | 13:37 |
T42 | <edp_17> mal, rinigus: Thanks both. | 13:38 |
T42 | <edp_17> @elros34 : MTP works! Whoohoo. 😆 | 13:41 |
T42 | <edp_17> Now, I am trying to utilise the two softkeys as "leds". Just like on the i9100. | 13:42 |
T42 | <elros34> @edp_17 I wonder whether you have same issue as me on these old kernel devices that sometimes copying big files over mtp from ubuntu to device fail. I think on windows it's better but I prefer mass storage mode so I am not sure anymore. | 13:45 |
T42 | <edp_17> @elros34 : I have not tried that but happy to test it for you. What do you mean under "big files"? | 13:47 |
T42 | <elros34> >= 300MB | 13:49 |
T42 | <edp_17> @elros34 : One by one, I've copied over two files (340MB and 958MB) There was no issue with any of them. | 13:53 |
T42 | <elros34> ok thank | 13:54 |
T42 | <elros34> sometimes it also works correctly for my so I am not sure what triggers that issue | 13:55 |
T42 | <edp_17> Btw in messwerk: Rotation, Accelerometer, Magnetometer and Pressure sensor do not work. Where should I switch them on? | 13:56 |
T42 | <elros34> At least accelerometer should work if your sensorfw works and you have hw-settings | 13:57 |
T42 | <edp_17> What is interesting, all sensors (except light and accelerometer) do work in csd. So, I am not sure what am missing. 🤔 | 13:58 |
T42 | <edp_17> In csd Accelerometer doesn't move at all. (Just like in messwerk) | 13:59 |
deathmist_ | mal: could https://github.com/mer-hybris/droid-hal-device/pull/295 get merged so local building for SFOS 4.0 will work without needing manual steps during build_packages? | 14:03 |
T42 | <elros34> @edp_17 running sensorfwd -l=test might give you some ideas what is wrong | 14:16 |
T42 | <edp_17> @elros34 : Thanks! I got this: https://paste.ubuntu.com/p/JTntvCkcXJ/ | 14:17 |
T42 | <elros34> sensorfwd.service is system service so stop it first and then start manually sensorfw as root | 14:19 |
T42 | <edp_17> Now I got this: https://paste.ubuntu.com/p/Sdv4876nYJ/ | 14:21 |
T42 | <edp_17> Ops. I accidentally renamed a file on device /etc/mce and I don't know what was the name of the file. By the content, can you advise me the file name? | 14:28 |
T42 | <elros34> what file? /etc/mce is a directory. about sensorfw, maybe adding is also necessary -c=/etc/sensorfw/primaryuse.conf | 14:30 |
T42 | <edp_17> I wanted to create a file 60-binary.ini and accidentally renamed an existing one. This is the content: https://paste.ubuntu.com/p/qY3FWmgtkp/ | 14:31 |
T42 | <elros34> 60-powerkey-lipstick-jolla-home.conf | 14:33 |
T42 | <edp_17> Thanks! | 14:33 |
T42 | <edp_17> Here is the sensorfwd log with primaryuse.conf: https://paste.ubuntu.com/p/82mnqj5R6X/ | 14:34 |
T42 | <edp_17> Oh, and I have this file in /etc/mce but I am not sure I need it: https://paste.ubuntu.com/p/T4mbvxSrw6/ | 14:36 |
T42 | <elros34> HYBRIS CTL setActive0=ACCELEROMETER, true) -> -19=Unknown error -19. Don't you have anything in other logs? I guess you should have some process responsible for sensors started by /init*rc | 14:49 |
T42 | <edp_17> I have only found this one where sensor was mentioned: https://paste.ubuntu.com/p/2w8P8ns8sg/ | 15:02 |
T42 | <elros34> according to device repo logcat should have something like AkmSensor | 15:19 |
T42 | <edp_17> @elros34 : Yes, there is "AkmSensor: unable to load AKM Library, dlopen failed": https://paste.ubuntu.com/p/MYZmDkRYgQ/ | 15:31 |
T42 | <edp_17> What is that NetlinkListener? It is spamming the log. | 15:32 |
T42 | <elros34> that explains -19 error ENODEV. I think I saw that netlinklistenter error many times in logs so probably is not critical | 15:41 |
T42 | <elros34> do you have libakm.so in system or vendor? | 15:45 |
T42 | <elros34> so this might be related: /system/lib/libakm.so: has text relocations | 15:50 |
rinigus | mal: any idea on how to make minimediaservice (32bit) aware that libaudiopolicyservice.so is in /usr/libexec/droid-hybris/system/lib/libaudiopolicyservice.so? right now it just doesn't start | 15:51 |
mal | what does strace tell? | 15:51 |
rinigus | mal: will have to check | 15:55 |
T42 | <edp_17> @elros34 [do you have libakm.so in system or vendor?], I have: /system/lib/libakm.so | 15:58 |
T42 | <edp_17> Should I symlink it to /system or /vendor? | 15:59 |
T42 | <elros34> I don't think it will help | 16:00 |
swanux | Any idea on this (Error <creator>[02/25 15:59:02] : URLGrabber error: http://repo.merproject.org/obs/nemo:/devel:/hw:/common/sailfish_latest_aarch64/repodata/repomd.xml - [Errno 14] HTTP Error 404 - Not Found) error when running rpm/dhd/helpers/build_packages.sh --mic ? Why is it looking for a non-existent page? | 16:01 |
T42 | <elros34> @edp_17 https://github.com/libhybris/libhybris/blob/506718cc54ceb954e68287d8370cbe30225185bf/hybris/common/n/linker.cpp#L4218. Either comment out that false return or maybe analyze that code to see whether TARGET_NEEDS_PLATFORM_TEXT_RELOCATIONS could help | 16:06 |
rinigus | mal: strace - it doesn't look in droid dir https://dpaste.org/ORJ9#L348 | 16:11 |
mal | rinigus: ok, so the known issue I saw somewhere else also, something more needed in ld.config.29.txt, for now just add preload there | 16:16 |
rinigus | mal: haven't used preload so far - how do I do that? | 16:18 |
mal | to .rc file: setenv LD_PRELOAD /usr/libexec/droid-hybris/system/lib/libaudiopolicyservice.so | 16:21 |
mal | to the correct service | 16:21 |
rinigus | mal: thanks! | 16:22 |
mal | rinigus: does camera now work well? | 16:25 |
rinigus | mal: for some reason it didn't work. still cannot find the lib | 16:27 |
rinigus | let me check if I can do it manually | 16:27 |
rinigus | corresponding .rc file https://dpaste.org/fj9r | 16:27 |
rinigus | mal: when running strace, I am getting error: ERROR: ld.so: object '/usr/libexec/droid-hybris/system/lib/libaudiopolicyservice.so' from LD_PRELOAD cannot be preloaded (wrong ELF class: ELFCLASS32): ignored. | 16:29 |
mal | rinigus: is minimediaservice really 32-bit? | 16:31 |
rinigus | mal: /usr/libexec/droid-hybris/system/bin/minimediaservice: ELF 32-bit LSB shared object, ARM, EABI5 version 1 (SYSV), dynamically linked, interpreter /system/bin/linker | 16:32 |
rinigus | :( | 16:32 |
rinigus | mal: if I add missing libs as symlinks at /system/lib to /usr/libexec/droid-hybris/system/lib, I can get it run. | 16:52 |
rinigus | but now I am missing /usr/libexec/droid-hybris/system/bin/miniafservice - don't have that on device | 16:53 |
mal | rinigus: miniaf is not used anymore in new ports | 16:53 |
mal | this is used instead https://github.com/mer-hybris/audiosystem-passthrough | 16:53 |
rinigus | mal: damn, asked about it yesterday. sorry. | 16:54 |
mal | you might need this is patterns Requires: audiosystem-passthrough-dummy-af | 16:54 |
rinigus | but why do we have service miniaf /usr/libexec/droid-hybris/system/bin/miniafservice in servicemanager.rc | 16:54 |
mal | rinigus: because we have been too lazy to remove it from new bases | 16:54 |
mal | we just rebase the patches but never through of removing that | 16:55 |
rinigus | mal: I have that in the patterns (audiosystem-passthrough-dummy-af) | 16:55 |
rinigus | and logcat full of "ServiceManager: Waiting for service 'media.audio_flinger' on '/dev/binder'" | 16:56 |
rinigus | ServiceManager: Service media.audio_flinger didn't start. Returning NULL | 16:56 |
mal | hmm | 16:56 |
mal | is the process running? | 16:57 |
rinigus | mal: /usr/libexec/audiosystem-passthrough/audiosystem-passthrough is running; but nothing called media.audio_flinger or something androidy named audio (pgrep -l audio) | 16:59 |
mal | install libgbinder-tools and check binder-list | 17:00 |
mal | did the service start successfully? | 17:00 |
rinigus | mal: will have to take a break. will do , thank you for help | 17:01 |
mal | I will build a new image today for my device and test also | 17:02 |
*** jrayhawk_ is now known as jrayhawk | 17:20 | |
T42 | <edp_17> @elros34 [@edp_17 https://github.com/libhybris/libhybris …], Thanks @elros34 . Is this not a problem on your device with that old kernel? | 17:27 |
T42 | <elros34> it's not about kernel but libraries. If I am right then your libakm.so is from older android version | 17:31 |
swanux | When I try to flash sailfish zip I get error 7, failed to extract. I tried manually, and it gives to following error: | 18:04 |
swanux | tar: exec bunzip2: No such file or directory | 18:04 |
swanux | tar: read error | 18:04 |
swanux | Any ideas? (the android base is correct and I tried with twrp adb shell) | 18:04 |
rinigus | mal: not sure what I am expected to get from binder-list command: https://dpaste.org/VK1z | 18:10 |
mal | rinigus: maybe you need to check device /dev/binder not hwbinder, assuming it works with that | 18:22 |
rinigus | mal: running `binder-list -d /dev/binder` returns nothing. but at the same time, "ServiceManager: Waiting for service 'media.audio_flinger' on '/dev/binder'..." in logcat | 18:27 |
rinigus | maybe I should try to rebuild image without that MINIMEDIA_AUDIOPOLICYSERVICE_ENABLE:=1 to see whether it worked there. in older logcats I don't see that message... | 18:28 |
rinigus | ... just pain to do it all without OBS, I must say | 18:29 |
piggz | mal: armed with a clean source tree, im trying _just_ the egl init change to see how that affects browser | 18:30 |
mal | piggz: cool | 18:35 |
mal | rinigus: do you see anything other related to audio flinger in logcat? | 18:36 |
rinigus | let me grab a fresh one | 18:36 |
mal | rinigus: also try stopping the audiosystem-passthrough and start the process manually with strace | 18:36 |
mal | to see if it fails somehow | 18:36 |
rinigus | will do | 18:37 |
mal | rinigus: remember the parameters when starting it | 18:37 |
swanux | Same flash problem with lineage recovery... | 18:48 |
mal | one way do ugly installation is to flash the kernel only and then after booting extract the rootfs via telnet | 19:09 |
swanux | mal In the meantime I worked around the problem (thanks God android is linux based and at least I understand that part). | 19:27 |
swanux | After the manual install (extract rootfs then flashing hybris-boot) I have a good old bootloop... Do you know some typical / most common problem that causes it? | 19:27 |
mal | how soon does it reboot? can you telnet in? | 19:29 |
mal | I'll be away for a while, maybe someone else can help | 19:29 |
swanux | Vendor logo for normal time, then when the OS boots black screen for ~5s, reboot, vendor logo etc... So no telnet during this time window I assume. | 19:33 |
rinigus | mal: running AUDIOSYSTEM_PASSTHROUGH_TYPE=af /usr/libexec/audiosystem-passthrough/audiosystem-passthrough --address dummy | 19:45 |
rinigus | [audiosystem-passthrough] ERROR: Failed to add media.audio_flinger (-1) | 19:46 |
rinigus | strace: https://dpaste.org/jQRd | 19:46 |
rinigus | that error was in journal as well | 19:46 |
rinigus | mal: ^ | 19:46 |
T42 | <elros34> @swanux don't assume, check in host dmesg according to hadk pdf | 19:53 |
rinigus | mal: when running verbose https://dpaste.org/kZqf | 19:57 |
mal | rinigus: do you have /etc/gbinder.conf | 20:00 |
swanux | T42 checked dmesg & lsusb, no telnet | 20:00 |
mal | swanux: which android base? | 20:00 |
rinigus | mal: yes, with ApiLevel = 29 under [General] | 20:03 |
rinigus | mal: comes via sparse-10 | 20:03 |
mal | rinigus: try moving it away | 20:03 |
mal | and try again | 20:03 |
rinigus | ok | 20:03 |
swanux | mal It's the 17.1, so I'll try with 16.0 next before digging deep inside the possible problems with this one. I'll see how it goes | 20:08 |
rinigus | mal: no difference. audiosystem-passthrough fails as service, when running using the same command as above, same verbose output | 20:08 |
mal | swanux: 17.1 has some special things which have not been documented yet, did you use "export TEMPORARY_DISABLE_PATH_RESTRICTIONS=true" before building hybris-hal | 20:18 |
mal | rinigus: try apilevel 28 | 20:20 |
rinigus | mal: don't see any difference, unfortunately | 20:24 |
rinigus | (rebooted after changing) | 20:25 |
rinigus | mal: I am on libgbinder 1.0.47, as the latest sfos release | 20:27 |
mal | rinigus: too old | 20:32 |
mal | rinigus: update to at least 1.1.4 | 20:35 |
piggz | mal: seems i have a minimal patch to fix browser | 21:21 |
piggz | doesnt fix email | 21:21 |
piggz | but neither did the big patch | 21:22 |
piggz | mal: https://git.sailfishos.org/piggz/gecko-dev/tree/mesa | 22:03 |
piggz | https://git.sailfishos.org/mer-core/gecko-dev/merge_requests/241 | 22:34 |
mal | piggz: nitpick, fix the useless whitespace change near the end of the diff | 22:39 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!