| T42 | <edp_17> This invoker issue is very annoying. I need to start all app from command line otherwise they all crash. 😔 | 00:00 |
|---|---|---|
| T42 | <elros34> have you tried mali quirks? | 00:00 |
| T42 | <edp_17> Not yet. How can I add that on device? | 00:02 |
| T42 | <elros34> instruction is in faq | 00:03 |
| T42 | <edp_17> 😄 | 00:03 |
| T42 | <edp_17> Well, the hadk only says to add the mali quirks into droid-hal-i9100.spec and rebuild packages. Then which packages I should install on the device? | 00:20 |
| mal | at least libhybris packages | 00:22 |
| T42 | <edp_17> Something not quite right around the call/messages. | 00:22 |
| T42 | <edp_17> I can receive text message but cannot respond (Problem with sending message) | 00:23 |
| T42 | <elros34> you want libhybris. To make sure that droid-hal header with mali quirks is picked up by libhybris you can remove hybris/mw/libhybris before building. grep mali-hist-dump libhybris-common.so.1.0.0 should tell whether you rebuild it correctly or not | 00:23 |
| T42 | <edp_17> I can make a call but when receive a call the dialer doesn't start. | 00:23 |
| T42 | <edp_17> Okay, I will install them once the pc finished with the build. | 00:24 |
| T42 | <edp_17> @elros34 : I already started to build. 😔 | 00:25 |
| T42 | <edp_17> In what folder I need to run this command? (grep mali-hist-dump libhybris-common.so.1.0.0) | 00:26 |
| T42 | <elros34> you need to upack libhybris*rpm and found that library | 00:26 |
| T42 | <edp_17> Okay, I'll try. | 00:27 |
| T42 | <edp_17> From libhybris folder do I need to reinstall only the libhybris*.rpm or all the other rpms too which name starts with libhybris? | 00:29 |
| T42 | <elros34> at least this rpm which contains mentioned library | 00:31 |
| T42 | <elros34> but to be safe you can reinstall all which are already installed on device | 00:32 |
| T42 | <edp_17> I unpacked the libhybris*.rpm and ran the 'grep mali-hist-dump libhybris-common.so.1.0.0' command in the folder where the libhybris-common.so.1.0.0 file was. | 00:47 |
| T42 | <edp_17> The output is: Binary file libhybris-common.so.1.0.0 matches | 00:47 |
| T42 | <elros34> good, it means quirks are enabled | 00:48 |
| T42 | <edp_17> Oh, and I installed all packages (except the *-devel*.rpm ones) on the device. | 00:48 |
| T42 | <edp_17> I guess should reboot now. | 00:49 |
| T42 | <edp_17> Nope. The mali quirks didn't solve the app crash issue. After the reboot they crash just like before. | 00:52 |
| T42 | <edp_17> I wanted to upload a new logcat -b radio but pastebin didn't let me as the size was over 512k | 00:55 |
| T42 | <edp_17> Another question. The hadk-faq says if I want to route the audio to the headphone I should find out the jack-match and jack-device values with running 'evdev_trace' command. | 01:07 |
| T42 | <edp_17> It does return with a lots of stuff but nor 'SW_HEADPHONE_INSERT*' neither 'SW_MICROPHONE_INSERT*' among them. | 01:08 |
| T42 | <edp_17> output of evdev-trace is here: https://pastebin.com/RTUjvTT3 | 01:09 |
| T42 | <elros34> then use switch/h2w like explained in faq | 01:10 |
| T42 | <edp_17> So, which event should I use? | 01:10 |
| T42 | <elros34> neither, use /sys/class/switch/h2w instead | 01:11 |
| T42 | <edp_17> The 'ls /sys/class/switch/h2w/' returns with 'name power state subsystem uevent' | 01:14 |
| Oksanaa | Just a tiny question: why isn't Fxtec Pro1 https://github.com/sailfish-on-fxtecpro1 listed under http://bit.ly/port-status aka https://wiki.merproject.org/wiki/Adaptations/libhybris ? | 01:15 |
| T42 | <edp_17> So I should just add model = uevent and switch = /sys/class/switch/h2w/ into the accessories.ini | 01:15 |
| r0kk3rz | Oksanaa: because nobody has put it there yet | 05:01 |
| T42 | <adampigg> I prob should! | 07:35 |
| mal | @adampigg does latte use libdrm-intel package? | 13:08 |
| T42 | <adampigg> mal: i think so @eugenio_g7 ^^ | 13:50 |
| T42 | <eugenio_g7> Yup | 14:10 |
| T42 | <adampigg> mal: why? | 14:29 |
| mal | we considered removing libpciaccess but then noticed that libdrm-intel needs it | 14:37 |
| T42 | <adampigg> for info, mesa master builds fail becuase of missing kernel headers, which isnt ideal | 15:12 |
| mal | @adampigg do you have the build on OBS? | 15:18 |
| T42 | <adampigg> mal: yes, either iin the latte or dontbeevil project | 15:28 |
| T42 | <adampigg> dontbeevil iirc | 15:28 |
| mal | @adampigg strange, http://man7.org/linux/man-pages/man2/kcmp.2.html "The kcmp() system call first appeared in Linux 3.5" and our kernel headers are 3.18 | 15:32 |
| T42 | <adampigg> does the package get installed? | 15:36 |
| mal | @adampigg I checked our kernel header sources and those have file "include/linux/kcmp.h", need to check the rpm | 15:36 |
| mal | interesting, it's not in the rpm but it's in sources | 15:38 |
| T42 | <adampigg> cool, i found a bug | 15:39 |
| mal | well not sure if it's a bug or not, maybe not all headers are installed by default | 15:41 |
| T42 | <adampigg> maybe they are in later versions | 15:44 |
| mal | @adampigg lol, it was moved to public and installed by default in 3.19, we have 3.18 | 15:45 |
| mal | https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=3f4994cfc15f38a3159c6e3a4b3ab2e1481a6b02 | 15:46 |
| T42 | <adampigg> ah, thats a pain! | 15:54 |
| T42 | <adampigg> what should we do? | 15:54 |
| T42 | <adampigg> add that patch to our kernel-header packahe | 15:56 |
| T42 | <adampigg> ? | 15:56 |
| T42 | <adampigg> why is our so old still? | 15:57 |
| mal | @adampigg we don't really need new headers so it has been kept like that | 16:13 |
| T42 | <adampigg> Now we do ;) | 16:16 |
| *** OhYash1 is now known as ohyash | 17:51 | |
| piggz | mal: i have configfs mounted on /sys/kernel/config, it has an empty usb_gadget folder .... where do i start? | 17:54 |
| rinigus | piggz: you started already well - step 1: ask mal | 17:55 |
| piggz | well, i started with the porters archive search, but there isnt mush info | 17:55 |
| piggz | ive created a usb-moded-configfs.ini and set the base folder | 17:55 |
| piggz | but, i feel i need to populate the folder with something | 17:56 |
| rinigus | piggz: its a bit different for me, I have it in /config. is it android device? | 17:58 |
| piggz | rinigus: nope | 18:01 |
| piggz | so, i dont have an andoird init file showing how it should be created | 18:02 |
| T42 | <adampigg> ok, lookinh | 18:03 |
| T42 | <elros34> afaik | 18:03 |
| rinigus | piggz: I suppose you could take as it is done on some android device where it works. or check out how its done on sfos x recovery | 18:03 |
| mal | rinigus: :D | 18:09 |
| piggz | mal: rinigus: well, my pc has a rndis device now :D | 18:11 |
| rinigus | :) | 18:12 |
| rinigus | piggz: now document in this channel so your first step (grep) would work next time | 18:12 |
| piggz | rinigus: first debug ;) | 18:13 |
| piggz | what do these errors mean | 18:13 |
| mal | piggz: show some output | 18:14 |
| piggz | write command errors....... | 18:14 |
| mal | and what you have in /config | 18:14 |
| piggz | mal: https://paste.gnome.org/pv3os634r | 18:14 |
| piggz | this is my script https://paste.gnome.org/poaoua5n9 | 18:15 |
| piggz | im leaving it as /sys/kernel/config as its mounted there automatically | 18:15 |
| mal | ok | 18:15 |
| piggz | usb_moded hasnt figured that out yet, mut im wondering what those write errors are about | 18:16 |
| mal | piggz: maybe use b.1, that is default in usb-moded | 18:17 |
| mal | https://git.sailfishos.org/mer-core/usb-moded/blob/master/src/usb_moded-configfs.c#L46 | 18:17 |
| mal | piggz: also you need to configure the rndis name | 18:18 |
| mal | piggz: https://piggz.co.uk/sailfishos-porters-archive/index.php?log=2019-10-08.txt#line69 | 18:21 |
| mal | piggz: and https://piggz.co.uk/sailfishos-porters-archive/index.php?log=2019-10-08.txt#line82 | 18:21 |
| piggz | mal: getting somewhere | 18:39 |
| mal | piggz: did those help? | 18:39 |
| piggz | mal: added | 18:40 |
| piggz | gadget_base_directory = /sys/kernel/config/usb_gadget/g1 | 18:40 |
| piggz | now, logs say | 18:40 |
| piggz | https://paste.gnome.org/pqhz6seor | 18:40 |
| piggz | the setup script still complains about write command | 18:41 |
| piggz | whats that about | 18:41 |
| piggz | mal: and those file missing errros by usb_moded are because of https://paste.gnome.org/prhm4sjeh | 18:42 |
| piggz | looks like they can be overwridden in config | 18:43 |
| piggz | ethaddr -> host_addr ? | 18:43 |
| piggz | not sure what wceiss is | 18:43 |
| piggz | mal: argh, no config for them https://git.sailfishos.org/mer-core/usb-moded/blob/master/src/usb_moded-configfs.c#L248 | 18:44 |
| mal | hmm, that looks familiar, the wceis error | 18:46 |
| piggz | mal: current usb_moded doesnt expect they can be configured ... maybe ill fix that | 18:46 |
| mal | piggz: I think the solution could be something else, wait a moment | 18:47 |
| piggz | it sees to work well enough to bring up networking ... mtp looks broken | 18:49 |
| piggz | ill create a service to run the script after local-fs.tafget and brfore usb-moded | 18:49 |
| mal | piggz: can you give a link to your kernel configs | 18:51 |
| piggz | mal: https://gitlab.com/pine64-org/linux/blob/pine64-kernel-5.4.y/arch/arm64/configs/pine64_defconfig | 18:53 |
| piggz | mal: do you want a print of the full config | 18:55 |
| mal | piggz: can you check from device if CONFIG_USB_U_ETHER is enabled? | 18:56 |
| piggz | mal: | 18:56 |
| piggz | [root@PinePhone ~]# grep USB_U_ETHER * | 18:56 |
| piggz | CONFIG_USB_U_ETHER=y | 18:56 |
| mal | hmm | 18:58 |
| piggz | mal: i made a service to start it, this is what my boot looks like https://paste.gnome.org/pgkkdwszb | 18:59 |
| piggz | so, it kinda works | 18:59 |
| piggz | just some odditiies | 18:59 |
| piggz | mal: what pacakge creates nemo user ? | 19:16 |
| mal | hmm, not sure | 19:18 |
| piggz | mal: i need droid-config-pinephone to be installe after that package to ensure script is run to add nemo to groups....seems our rootfs has the groups, but nemo is not a member | 19:22 |
| T42 | <adampigg> sailfish-setup | 20:02 |
| piggz | mal: fixed write errors....write is a bash function, not the binary | 20:25 |
| piggz | only error self is regarding rndis_bam.rndis' | 20:25 |
| piggz | is that needed? | 20:25 |
| mal | piggz: did you add the config I asked? | 20:29 |
| mal | about rndis name | 20:30 |
| piggz | yes... | 20:30 |
| piggz | [root@PinePhone ~]# cat /etc/usb-moded/usb-moded-configfs.ini | 20:30 |
| piggz | [configfs] | 20:30 |
| piggz | gadget_base_directory = /sys/kernel/config/usb_gadget/g1 | 20:30 |
| piggz | function_rndis = rndis.usb0 | 20:30 |
| piggz | the rndis_ban name comes from the init-script in hybris-boot | 20:30 |
| piggz | i copied it verbatim | 20:30 |
| piggz | maybe not required that bit? | 20:30 |
| mal | piggz: oh, that is wrong for your kernel probably, I assume you need to write rndis.usb0 | 20:32 |
| mal | not rndis_bam | 20:32 |
| piggz | mal: script does both https://github.com/mer-hybris/hybris-boot/blob/master/init-script#L207 | 20:32 |
| piggz | so, i can remove second i guess | 20:32 |
| mal | afaik rndis_bam is qcom thing | 20:32 |
| piggz | ok | 20:32 |
| piggz | mal: this is what i ended with https://github.com/sailfish-on-dontbeevil/droid-config-pinephone/commit/cafe8723f3e5857248273fa0740797e480c5adb6 | 20:42 |
| mal | piggz: looks reasonable, I only checked it quickly | 21:02 |
| piggz | np, thx | 21:03 |
| piggz | it might nee tweaked, but its a start | 21:03 |
| T42 | <edp_17> Good evening! | 21:18 |
| T42 | <edp_17> I tried to catch the log for an app when it is crashing. The journalct doesn't show any clue, it only says: "Application "Settings" with PID 2922 is not responding" | 21:26 |
| T42 | <edp_17> Then: "User terminated the unresponsive application "Settings" with PID 2922" | 21:27 |
| T42 | <edp_17> How can I see more about the issue? | 21:27 |
| r0kk3rz | which app? | 22:30 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!