| T42 | <HengYeDev> @elros34 mask user@100000 makes it not bootloop | 00:46 |
|---|---|---|
| T42 | <HengYeDev> i will try that | 00:46 |
| T42 | <HengYeDev> also: whats going on with the mount line in that faq section? what am i supposed to do with that? | 00:47 |
| T42 | <LSolrac> Hello. I come once more with another issue, unfortunately. Am I not setting up the variable correctly? I used export beforehand as well : https://irc.thaodan.de/.imgstore/d4aeab3c/file_3326.jpg | 05:13 |
| T42 | <LSolrac> Is mkbootimg actually required as opposed to a general tool? Because I've got this error; build/make/core/Makefile:2084: error: hardware/samsung/mkbootimg.mk: No such file or directory | 07:19 |
| T42 | <LSolrac> Is mkbootimg actually required as opposed to a general tool? Because I've got this error; build/make/core/Makefile:2084: error: hardware/samsung/mkbootimg.mk: No such file or directory | 07:20 |
| T42 | <LSolrac> Edit; nevermind, its referenced in the source (edited) | 07:20 |
| T42 | <LSolrac> so we can just reference the hadk instead? (re @SailfishFreenodeIRCBridgeBot: <piggz>we already pa...) | 07:36 |
| T42 | <elros34> @HengYeDev it's copy paste error from old wiki, surfaceflinger works now? | 09:36 |
| piggz | mal: how do you enable all debugging in ofono? | 10:22 |
| Thaodan | piggz: build with mb2? Is strip run? I you can try to set cflags to -O0 -ggdb | 12:48 |
| piggz | Thaodan: i meant debug logs .... other dev though that -d wasnt giving all the info he expected | 12:49 |
| Thaodan | Hey Porters PSA: | 12:50 |
| Thaodan | * New HADK released: https://sailfishos.org/develop/hadk | 12:50 |
| Thaodan | * Minimum version for Platform SDK chroot 4.3.0.15 | 12:52 |
| Thaodan | * Update your droid-hal-device submodule to the latest master | 12:52 |
| Thaodan | * Minimum version of SDK toolings and targets (collectively called SDK build tools) is 4.3.0.15 | 12:52 |
| Thaodan | * Update your droid-hal-img-boot-$DEVICE/hybris-initrd submodule to the latest master | 12:52 |
| Thaodan | * If you are using build tools older than 4.3.0.15, we strongly recommend to `sdk-assistant remove` your targets and toolings, and reinstall them | 12:52 |
| Thaodan | * `$ANDROID_ROOT/droid-local-repo/$DEVICE` structure changed to flat without subdirs. Best to remove the directory and re-run `build_packages.sh` | 12:53 |
| Thaodan | * Examine the changes in the new HADK: https://paste.opensuse.org/99654bb8 -- especially mic.rst explaining how mb2 now uses a working copy of your build tools | 12:53 |
| Thaodan | * Please start adopting the new vocabulary to align better with Sailfish SDK: | 12:53 |
| Thaodan | * Use the term "build tools" to describe the tooling+target pair or build target where you don't need to be more specific | 12:53 |
| Thaodan | * A "snapshot" is an mb2's working copy of your build target. Read about CLEAN BUILDS in `mb2 --help`. | 12:53 |
| Thaodan | * Prefer using `mb2 -t $VENDOR-$DEVICE-$PORT_ARCH build-shell` instead of `sb2`, so you can forget those pesky `-m sdk-install` switches :) It will also keep your `$VENDOR-$DEVICE-$PORT_ARCH` target free of changes mistakenly introduced during the edit-build-test cycle. Check `mb2 --help`. | 12:53 |
| Thaodan | * If you need root, add `--maintain`; but obtaining a return value will only be fixed in the next release after 4.3.0 | 12:53 |
| Thaodan | * Should you hit an error like "Fatal: Failed to init build target snapshot", simply remove the build targets snapshot(s) with `sdk-assistant remove --snapshots-of $VENDOR-$DEVICE-$PORT_ARCH`. There is a known issue with resetting existing snapshots after file type change, which will be fixed in the next release after 4.3.0 | 12:53 |
| T42 | <edp_17> Thanks Thaodan! | 13:34 |
| T42 | <edp_17> I immadiately have a problem when building droidmedia: https://paste.ubuntu.com/p/jpZ8BNp6PG/ | 13:34 |
| Thaodan | edp_17: Updated like instructed above? | 13:35 |
| T42 | <edp_17> Yes. | 13:35 |
| Thaodan | I can look after lunch. It looks like an invalid option is passed to mb2. | 13:36 |
| Thaodan | Try running the build_packages.sh with bash -xv and look for the options passed to mb2 | 13:37 |
| T42 | <edp_17> I hope done it correctly: https://paste.ubuntu.com/p/w7xX9Z3BXX/ | 13:40 |
| Thaodan | Is that really the latest platform sdk version? | 13:59 |
| T42 | <edp_17> Nope, I have not updated the platform sdk, only the rest. I'm doing now. | 14:07 |
| Thaodan | that's the issue then | 14:09 |
| T42 | <edp_17> It is interesting because I've built 4.3 for another device a few days ago and I didn't update platform sdk either. | 14:15 |
| T42 | <edp_17> Thodan: Updating platform sdk has solved the issue. Now droidmedia has built. Thanks for the advice. 👍 | 14:18 |
| T42 | <edp_17> *Thaodan | 14:19 |
| T42 | <HengYeDev> I guess it works, it doesn't crash (re @elros34: @HengYeDev it's copy...) | 14:28 |
| T42 | <edp_17> Should I remove gmp-droid and gts-droid packages from my OBS? (They both have failed but if I understand correctly don't need them for 4.3 anyway.) | 14:39 |
| T42 | <edp_17> The new platform sdk is only 800MB while the previous one was 1.7GB. Is this reduction normal? | 14:45 |
| T42 | <elros34> @HengYeDev it should show something on screen, do you see errors in logcat when you run it? | 14:47 |
| T42 | <elros34> @edp_17 if you will remove them they will be also missing for < 4.3 builds so instead disable them | 14:49 |
| T42 | <HengYeDev> note: didn't apply mali patches yet | 14:50 |
| T42 | <HengYeDev> Logs after surfaceflinger starts: https://gist.github.com/HengYeDev/305e63f8dd040c59912b6edada906552 | 14:50 |
| T42 | <HengYeDev> nothing on screen @elros34 | 14:50 |
| T42 | <edp_17> @elros34 : How can I disable a package? In Meta, I've disabled building older than 4.3 repos. | 14:51 |
| T42 | <elros34> in package page -> repositories tab | 14:53 |
| T42 | <elros34> @HengYeDev AFAIU patch is not relevant for surfaceflinger | 14:55 |
| T42 | <edp_17> @elros34 : Got it and done. Thanks. | 14:56 |
| T42 | <HengYeDev> @elros34 anything in logcat seem critical? nothing seems obvious to me... | 14:57 |
| T42 | <elros34> only ignoring vsync messages but maybe something else is missing to get it working so do not bother about it too much now and try mali quirk with lipstick | 14:59 |
| Thaodan | voidanix[m]: I merged most of the important changes on droid-hal-kumano done by you now. | 15:01 |
| Thaodan | I skipped all the commit messages regarding submodules since I updated them my self. | 15:01 |
| T42 | <AntonlX> Is it possible to route all sound on hybris devices through alsa instead pulseaudio modules and android hals? | 15:04 |
| T42 | <AntonlX> Like on DontBeEvil using alsa-ucm-conf | 15:04 |
| Thaodan | I don't think so unless you redo the routing modules without pa | 15:05 |
| Thaodan | Aren't alsa ucm rules created manually? | 15:05 |
| Thaodan | Android to Alsa UCM on the fly converting would have been nice. | 15:06 |
| T42 | <AntonlX> Alsa ucm conf repo has configs for different codecs like wcd934x. Configs has lines like in mixer_path files from android. If I set up config for my codec theoretically sound will go through alsa. | 15:14 |
| T42 | <AntonlX> https://github.com/alsa-project/alsa-ucm-conf/tree/8264072ec906b9ff496f81543f7684be17bfd180/ucm2/codecs/wcd934x | 15:14 |
| Thaodan | I get that but you have to create those manually | 15:16 |
| Thaodan | I think you can use those instead of mixer_path files from Android but you have to do all the work for the soc your self. | 15:17 |
| Thaodan | alsa-ucm-conf is also used by pa | 15:17 |
| piggz | Thaodan: i have that packaged for pinephone pro already | 16:04 |
| piggz | https://build.sailfishos.org/package/show/nemo:devel:hw:pine:dontbeevil/alsa-ucm-conf | 16:05 |
| Thaodan | piggz: With routing configs created by you? | 16:22 |
| piggz | Thaodan: no, that package is just the raw auc repo ... in the ppp config repo i have the device config https://github.com/sailfish-on-dontbeevil/droid-config-pinephone/tree/pro/sparse/usr/share/alsa/ucm2 | 16:33 |
| piggz | its the latest upstream compatible with our alsa ... mal is maybe updating it? | 16:35 |
| xmn | Hey piggz, I remember in the past you recommending editing /etc/sailfish-community-device-encrytion/*.ini. But I'm actually not sure what to change. I haven't yet tried it on my pure Sailfish install, and only on the megi mutliboot ... so I understand if you don't want to comment on that since it not yours port really. | 17:00 |
| xmn | But I've tried some of these ideas, https://github.com/sailfishos-open/sailfish-device-encryption-community/blob/main/Setup.md | 17:00 |
| xmn | I'm basically trying to turn off encryptions and use the home dir in the regular place on the emmc/sdcard. | 17:01 |
| xmn | So with the multiboot it end up in a endless rotating loading circle. Once I get past the first setup. I can get to the pull down quick settings. | 17:02 |
| xmn | Anyway just thought I would ask. Thanks regardless if you can help with this. Cheers! | 17:03 |
| xmn | Oh this was for the pinephone | 17:03 |
| piggz | pinephone is my port ;) | 17:07 |
| piggz | try the latest script ... it asks a question at the end if you are running from sd and makes the edit for you | 17:07 |
| xmn | yuppers :), but not the megi multiboot, which is the one I'm trying to get running. | 17:08 |
| xmn | Though I will redo my version of yours too :) | 17:08 |
| xmn | from this link, https://xnux.eu/p-boot-demo/. megi says .. "Sailfish had some trouble starting up in my quick test, and I didn't figure out why." | 17:14 |
| xmn | I figured it's likely having problems building or locating the home dir. | 17:15 |
| T42 | <HengYeDev> @elros34 no luck, mali quirk applied, still bootloops | 17:27 |
| T42 | <HengYeDev> the bootloop is definitely caused by something of user@100000 since when it is masked it is not bootloop | 17:27 |
| T42 | <elros34> so try test_hwcomposer/minimer instead with masked user. Does "grep mali-hist-dump libhybris-common.so.1.0.0" find this string? | 17:29 |
| piggz | xmn: well, /etc/sailfish-community-device-encryption/devices.ini needs to have the location of the home partition .. its quite simple | 17:34 |
| T42 | <HengYeDev> @elros34 Yes, the string is there | 17:34 |
| T42 | <HengYeDev> strace hwcomposer: https://gist.github.com/HengYeDev/93caf06a454b2cec4c068d40adaa6c88 | 17:34 |
| T42 | <HengYeDev> it crashes right after access of binder? | 17:34 |
| T42 | <HengYeDev> logcat: https://gist.github.com/HengYeDev/53ce9aec2c7efa2d315a5bcf78234e35 | 17:37 |
| T42 | <HengYeDev> seems like | 17:37 |
| T42 | <HengYeDev> Not support EXYNOS_GET_DISPLAY_MODE_NUM : Inappropriate ioctl for device | 17:37 |
| T42 | <HengYeDev> could be the error? | 17:37 |
| T42 | <AntonlX> Thaodan, does Aliendalvik works on your hybris-11 port? I tried to run aliendalvik exacted from xperia 10 II. Does Aliendalik depends on hybris version? | 17:38 |
| xmn | piggz ah, so really don't need to remove anything. Just need to put the full location. So instead of /dev/mmcblk2p3, I could use as an example, something like this for megi setup ... /run/media/manjaro/23785feb-000-lots of number-333/sailfish/home | 17:40 |
| xmn | I'll give that a try later thx!! | 17:40 |
| T42 | <elros34> @HengYeDev could be but it doesn't have to be an issue, maybe android also have some error. Is /sys/devices/platform/148c0000.decon_t/vsync readable? | 17:41 |
| T42 | <elros34> same error* | 17:41 |
| piggz | yes, it just needs to be a block device that can have an lvm created on it | 17:41 |
| piggz | @AntonlX not sure discussing AD here will got you anywhere | 17:41 |
| xmn | okay will keep that in mind piggz. Good info! | 17:42 |
| T42 | <HengYeDev> @elros34 It's no such directory, there is only one decon and it is 148b0000.decon_f | 17:45 |
| T42 | <HengYeDev> the vsync there is -r--r--r-- | 17:45 |
| T42 | <elros34> try minimer, test_hwcomposer may not work always | 17:46 |
| T42 | <elros34> do you have something interesting in /data/vendor/log/hwc/hwc_error_log.txt? | 17:57 |
| T42 | <HengYeDev> no, nothing except ExynosDevice is initialized in the log | 18:24 |
| T42 | <HengYeDev> minimer allows for reproduction of the bootloop issue | 18:24 |
| T42 | <elros34> so what cause reboot? | 18:25 |
| T42 | <HengYeDev> user@100000 masked, no reboot | 18:27 |
| T42 | <HengYeDev> i start minimer, after few seconds it reboots | 18:27 |
| T42 | <HengYeDev> I think i got something interesting in the logcat: https://gist.github.com/HengYeDev/d32524ab361d1c628137a86d4a1f5c71#file-gistfile1-txt-L417 | 18:27 |
| T42 | <elros34> so Exynos.HWCService doesn't crash like with test_hwcomposer. Did you check /sys/fs/pstore or dmesg -w to get reason for crash or any other method | 18:33 |
| T42 | <HengYeDev> @elros34 Yes | 18:35 |
| T42 | <HengYeDev> dmesg: https://gist.github.com/HengYeDev/736bfeabbb3342e66641b47976b8a448#file-dmesg-L538 | 18:35 |
| T42 | <HengYeDev> not sure what's with the big dump of registers at the end | 18:35 |
| T42 | <HengYeDev> seems like its a kernel oops | 18:35 |
| T42 | <elros34> so some timout | 18:40 |
| Mister_Magister | mal: ping help my gps is not reporting accuracy and rinigus is mad at me because of that :< | 18:46 |
| T42 | <BenKerry> is there a port for the redmi note 10 pro | 18:48 |
| T42 | <elros34> @HengYeDev do you even reach that stage when "== hwcomposer module ==" is printed by hwcomposer plugin? | 18:49 |
| T42 | <elros34> @BenKerry search in this channel logs for codename for that device or in github droid-config-<codename> | 18:50 |
| rinigus | I confirm being mad at Mister_Magister since he cannot get precision of GPS signal and feed it to pure maps. curiously, looks like GPS is actually locked and updated as it should. how can that be? | 18:51 |
| rinigus | :) | 18:51 |
| Mister_Magister | there hasn't been any significant update in geoclue-provider-hybris | 18:52 |
| rinigus | It is not clear to me whether it ever gave you precision on your port | 18:52 |
| Mister_Magister | it did not | 18:53 |
| piggz | lol, following discussion across 2 channels | 18:54 |
| Mister_Magister | piggz: :) | 18:54 |
| Mister_Magister | piggz: your input is welcome | 18:54 |
| rinigus | on other channel. .... :) | 18:55 |
| T42 | <elros34> no significant change but there was announce for old devices to pull geoclue-provider-hybris-hal instead geoclue-provider-hybris | 18:55 |
| Mister_Magister | @elros34 kwsk (gib more info) | 18:55 |
| T42 | <BenKerry> Where can I find the GitHub | 18:56 |
| rinigus | but piggz channel grep was supposed to be back, wasn't it? | 18:56 |
| piggz | https://piggz.co.uk/sailfishos-porters-archive/index.php?log=2021-10-20.txt#line36 | 18:56 |
| T42 | <elros34> @Mister_Magister https://collabedit.com/kr9xx 2021.09.28 | 18:56 |
| Mister_Magister | fascinating | 18:57 |
| piggz | https://piggz.co.uk/sailfishos-porters-archive/index.php?log=2021-09-17.txt#line21 | 18:57 |
| mal | Mister_Magister: probably adding some debug printing to geoclue plugin might be useful | 18:58 |
| Mister_Magister | mal: yyyyup | 18:58 |
| Mister_Magister | i do in fact have the -hal so thats irrelevant | 18:59 |
| Mister_Magister | mal: any tips on how to add some debug? or do i have to go into code and add the printing? | 19:00 |
| T42 | <Usman> How to start porting on window ...? | 19:01 |
| mal | Mister_Magister: you can check if there is any relevant printing already, I can't remember what it can print | 19:04 |
| Mister_Magister | okay | 19:04 |
| Mister_Magister | is geoclue a systemd service even? | 19:04 |
| Mister_Magister | its not | 19:05 |
| T42 | <HengYeDev> Yes (re @elros34: @HengYeDev do you ev...) | 19:07 |
| Mister_Magister | mal: what is even starting the geoclue? | 19:07 |
| T42 | <HengYeDev> @elros34 I get all the way to QEglWindow https://gist.github.com/HengYeDev/a44ec23333e20a0e62a6216adf8d4899 | 19:08 |
| Mister_Magister | ah nice to know sfos is using 9yo geoclue | 19:12 |
| Mister_Magister | don't take me wrong i'm not surprised | 19:12 |
| piggz | Mister_Magister: if it aint broke.... :D | 19:15 |
| Mister_Magister | xd | 19:16 |
| piggz | geoclue upgrade will come with the qt upgrade :D | 19:16 |
| Mister_Magister | oh yes | 19:17 |
| Mister_Magister | right after the silica will become opensource | 19:17 |
| rinigus | piggz: geoclue2 has its own bag of problems. qt plugin had such bugs which indicate it was not used by anyone before - at least for navigation software. | 19:23 |
| mal | Mister_Magister: geoclue-master process | 19:31 |
| Mister_Magister | mal: geoclue-hybris | 19:31 |
| Mister_Magister | QT_LOGGING_RULES="*=true" /usr/libexec/geoclue-hybris | 19:31 |
| mal | Mister_Magister: I meant geoclue-master starts geoclue-hybris when needed | 19:31 |
| Mister_Magister | yep | 19:31 |
| Mister_Magister | i jsut started it by hand | 19:31 |
| T42 | <HengYeDev> @elros34 There is also things in pstore | 23:01 |
| T42 | <HengYeDev> console-ramoops: https://gist.github.com/HengYeDev/391da9c3600de73ccf299c65e1fa708f | 23:01 |
| T42 | <HengYeDev> dmesg-ramoops: https://gist.github.com/HengYeDev/8e96f4e8ef922f7d7d6339afe902c80b | 23:01 |
| T42 | <HengYeDev> there is also pmsg-ramoops but it is difficult to read | 23:02 |
| T42 | <HengYeDev> seems to generally all be the same thing - oops due to decon | 23:03 |
| T42 | <HengYeDev> interesting that surfaceflinger doesnt crash tho | 23:05 |
| T42 | <elros34> surfaceflinger is on android side so it should work, Maybe on never android bases it requires /system/bin/bootanimation or something like else to display something. There is even old qt5-qpa-surfaceflinger-plugin | 23:29 |
| T42 | <elros34> you could see if starting surfaceflinger then minimer generate something more useful | 23:30 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
