mccreary | Are you guys building your own kernels for the F(x)tec Pro1? | 02:02 |
---|---|---|
mccreary | We don't have kernel source yet... | 02:03 |
r0kk3rz | they would be yeah | 02:50 |
r0kk3rz | cant make sailfish run on stock android kernel | 02:50 |
piggz_ | mccreary: yes, you dont have the kernel sources ;) | 07:26 |
piggz_ | mal: i'll check that when the device loads again...however... | 07:45 |
piggz_ | as per the dmesg output, this is displayed: | 07:47 |
piggz_ | [ 3179.729987] wlan: driver loaded | 07:47 |
piggz_ | that message comes from | 07:47 |
piggz_ | pr_info("%s: driver loaded\n", WLAN_MODULE_NAME); | 07:48 |
piggz_ | we know that the firmware loading is started off from the the probe function, static int wlan_hdd_probe( | 07:49 |
piggz_ | and, that function, near the top has | 07:49 |
piggz_ | pr_info("%s: %sprobing driver v%s\n", WLAN_MODULE_NAME, | 07:50 |
piggz_ | reinit ? "re-" : "", QWLAN_VERSIONSTR); | 07:50 |
piggz_ | which isnt printed in the kernel | 07:50 |
deathmist | still trying to fix call audio on 15.1 and fixing broken symlinks in /etc now: which one of these files in /vendor should I symlink back to /etc/mixer_paths.xml? https://pastebin.com/iKCtRQxL | 09:41 |
T42 | <adampigg> mal: and yes, /lib/firmware link still exists | 09:59 |
deathmist | finally tested camera recording, it doesn't break everything after recording with the audiopolicyservice patch (y) recording quality is still bad tho with very visible I guess compression artifacts or something while moving even the slightest bit + a video is always completely black/garbled for the first 2 seconds | 10:58 |
deathmist | there artifacts ofc aren't visible in the viewfinder which looks perfect | 11:00 |
Mister_Magister | mal: you can create now repos on obs right? | 11:44 |
T42 | <SALPHA8506> PlatformSDK s-8506@s8506-X556UR:~/Desktop/SAILFISH$ sudo zypper ref | 11:46 |
T42 | <SALPHA8506> error: db4 error(-30974) from dbenv->open: DB_RUNRECOVERY: Fatal error, run database recovery | 11:46 |
T42 | <SALPHA8506> error: cannot open Packages index using db4 - (-30974) | 11:46 |
T42 | <SALPHA8506> error: cannot open Packages database in /var/lib/rpm | 11:46 |
T42 | <SALPHA8506> Target initialization failed: | 11:46 |
T42 | <SALPHA8506> Rpm Exception | 11:46 |
T42 | <SALPHA8506> @SALPHA8506 [PlatformSDK s-8506@s8506-X556UR:~/Desktop/SAIL …], ?? | 11:46 |
mal | Mister_Magister: yes | 12:01 |
Mister_Magister | mal: can you make devel and testing repo for asus Z01R? | 12:02 |
mal | maybe later tonight, at summer cottage now | 12:02 |
Mister_Magister | ah sure no hurry | 12:02 |
Mister_Magister | just sayin' | 12:03 |
Mister_Magister | hmcp: cannot stat `./out/target/product/Z01R/hybris-updater-script': No such file or directory | 12:22 |
Mister_Magister | nani | 12:22 |
Mister_Magister | make hybris-updater-script fixes it | 12:33 |
T42 | <SALPHA8506> HABUILD_SDK [] s-8506@s8506-X556UR:~/Desktop/SAILFISH$ mkdir -p $ANDROID_ROOT | 12:58 |
T42 | <SALPHA8506> mkdir: missing operand | 12:58 |
T42 | <SALPHA8506> Try 'mkdir --help' for more information. | 12:58 |
Mister_Magister | you r $ANDROID_ROOT variable is empty | 12:59 |
Mister_Magister | you entered habuild wrongly | 12:59 |
T42 | <SALPHA8506> how to fix it ? | 13:00 |
Mister_Magister | whats in your .hadk | 13:00 |
T42 | <SALPHA8506> HABUILD_SDK [] s-8506@s8506-X556UR:~/Desktop/SAILFISH$ cat ~/.hadk.env | 13:01 |
T42 | <SALPHA8506> export PLATFORM_SDK_ROOT="/srv/mer" | 13:01 |
T42 | <SALPHA8506> export ANDROID_ROOT="$HOME/hadk" | 13:01 |
T42 | <SALPHA8506> export VENDOR="z2_plus" | 13:01 |
T42 | <SALPHA8506> export DEVICE="zuk" | 13:01 |
T42 | <SALPHA8506> export PORT_ARCH="armv7hl" | 13:01 |
Mister_Magister | use https://susepaste.org or similar thanks | 13:01 |
Mister_Magister | so when you enter sfossdk do you have ANDROID_ROOT set? echo $ANDROID_ROOT | 13:02 |
T42 | <SALPHA8506> @Mister_Magister [so when you enter sfossdk do you have ANDROID_ …], no | 13:05 |
T42 | <SALPHA8506> how to set it ? | 13:05 |
Mister_Magister | don't use reply option | 13:05 |
Mister_Magister | you should have it | 13:05 |
Mister_Magister | export ANDROID_ROOT="$HOME/hadk" clearly sets it | 13:05 |
Mister_Magister | just run echo $ANDROID_ROOT | 13:05 |
deathmist | SALPHA8506: I see you were in a directory "~/Desktop/SAILFISH", is this your build tree? | 13:06 |
T42 | <SALPHA8506> no | 13:06 |
mccreary | piggz_, I guess that means F(x)tec hasn't published kernel source? :-( | 13:07 |
deathmist | SALPHA8506: check hadk pdf section 4 again to make sure you didn't miss something if this is your first time building | 13:08 |
T42 | <SALPHA8506> ok | 13:09 |
Mister_Magister | mccreary: they will after release i think | 13:09 |
mccreary | OK, thanks for the info. We have been promised pre-release access, but don't have it yet | 13:10 |
mccreary | FWIW, TWRP isn't very useful with the prebuilt kernel | 13:10 |
Mister_Magister | who needs twrp xd | 13:10 |
mccreary | Since you can use it when it is in the ramdisk in a boot partiton | 13:11 |
mccreary | I think TWRP is useful ;-) | 13:11 |
mccreary | With kernel source I can make it work from 'flashboot boot boot.img' | 13:12 |
Mister_Magister | whoa what just happend | 13:12 |
Mister_Magister | mic build only tar.bz2 | 13:12 |
Mister_Magister | no zip | 13:12 |
Mister_Magister | weird | 13:13 |
mal | Mister_Magister: if I remember correctly the scripts don't like codename in capital letters, check fp2 repo how I did it | 13:15 |
Mister_Magister | ok | 13:15 |
mal | so rpm_device and device or something in spec | 13:16 |
T42 | <SALPHA8506> https://pastebin.com/t9J2azag | 13:19 |
T42 | <SALPHA8506> repo: command not found | 13:19 |
deathmist | if you are willing to port you need the patience to read through the document; it states clearly above the command that you need to set it up https://source.android.com/setup/build/downloading#installing-repo | 13:20 |
T42 | <SALPHA8506> got it | 13:21 |
Mister_Magister | mal: after that and building dcd i'm getting rpm2cpio: /home/foidbgen/5z/android//droid-local-repo/Z01R/droid-configs/droid-config-Z01R-ssu-kickstarts-1-1.armv7hl.rpm: No such file or directory | 13:25 |
mal | Mister_Magister: so you didn't change it in all repos and also in .hadk.env | 13:30 |
mal | also target name | 13:30 |
Mister_Magister | mal: change everything?! lol | 13:31 |
Mister_Magister | also in dvd you have rpm_device but no device is that okay? | 13:31 |
mal | check fp2 repos | 13:32 |
mal | I don't remember what I did | 13:32 |
Mister_Magister | i'm doing that | 13:32 |
Mister_Magister | also in .hadk.env you have export DEVICE="fp2-sibon" or somthing else | 13:33 |
Mister_Magister | ok i'm starting to follow | 13:34 |
Mister_Magister | dunno if that will work tho | 13:34 |
T42 | <SALPHA8506> how to add repo ? | 13:36 |
mal | Mister_Magister: yes, the point is that sfos side uses non-capital letters and device in droid-hal spec tells dhd submodule what the android side codename is | 13:37 |
Mister_Magister | i see small problem here | 13:37 |
Mister_Magister | oh nvm | 13:37 |
deathmist | SALPHA8506: what do you mean by that? the device speficic repos in local_manifests? scavenge the LineageOS GitHub repos for your device & manually go through each lineage.dependencies file and add them in the manner specified in the hadk pdf | 13:40 |
deathmist | or the repo command? | 13:40 |
deathmist | the command you can add e.g. by adding ~/bin to PATH in ~/.hadk.env at the end: "export PATH=$HOME/bin:$PATH" | 13:43 |
deathmist | and then re-entering ubu-chroot | 13:44 |
mal | Mister_Magister: remember to replace the codename also inside other files in config repo, like patterns, probably easier to recreate the repo | 13:46 |
Mister_Magister | yeah yeah i did that | 13:46 |
Mister_Magister | replace everything | 13:46 |
Mister_Magister | it seems to be building | 13:50 |
Mister_Magister | mal: do you remember where build is stored or something like this? when i'm trying to build dvd it gets old paths or something and is not building because it wants to crete log in / | 14:36 |
Mister_Magister | removing hybris/mw helped for mw | 14:38 |
Mister_Magister | ah sorry i'm actually retard | 14:44 |
Mister_Magister | seems like updater-unpack doesn't work with my twrp | 15:01 |
T42 | <SALPHA8506> HABUILD_SDK [] s-8506@s8506-X556UR:~/hadk$ repo sync --fetch-submodules | 15:09 |
T42 | <SALPHA8506> error: command 'sync' requires repo to be installed first. | 15:09 |
T42 | <SALPHA8506> i installed repo then why it showing like this | 15:10 |
T42 | <SALPHA8506> https://paste.ubuntu.com/p/SdX9b3hGH6/ | 15:25 |
T42 | <SALPHA8506> HABUILD_SDK [z2_plus] s-8506@s8506-X556UR:~/hadk$ repo --version | 15:29 |
T42 | <SALPHA8506> error: repo is not installed. Use "repo init" to install it here. | 15:29 |
deathmist | SALPHA8506: did you read the messages I sent before? | 15:33 |
deathmist | doing "which repo" right now should give you the path to the command | 15:34 |
deathmist | if it doesn't, you didn't install it properly | 15:34 |
mal | Mister_Magister: recent twrp seems to cause many issues | 15:40 |
Mister_Magister | mal: i don't have rndis in kernal which causes bootloop error: 'KBUILD_MODNAME' undeclared here (not in a function | 15:40 |
T42 | <SALPHA8506> @deathmist [SALPHA8506: did you read the messages I sent b …], i read and set all the things but not working i create my repo and paste into /Home....../local_manifest/ | 15:47 |
T42 | <SALPHA8506> @deathmist [if you are willing to port you need the patien …], and i also installed repo | 15:48 |
deathmist | please don't use the reply feature as it doesn't format nice on IRC chat, could we talk in telegram private: add me @Deathmist1337 | 15:49 |
deathmist | you aren't using repo right also, check hadk for the repo init command you need to run | 15:54 |
piggz | mal: how can we override/disable built in services, that are part of /vendor/etc/init/ and not in individual .rc files? | 16:33 |
piggz | eg time_daemon from qcom.hw.init.rc | 16:34 |
piggz | hw/qcom.init.rc even | 16:36 |
deathmist | there might be a symlink issue (at least on 15.1): /usr/sbin/udevadm points to ../bin/udevadm (broken) when it should point to ../../bin/udevadm where the binary actually exists | 16:46 |
deathmist | doesn't seem to be a big issue as /bin comes before /usr/sbin in PATH, but it's still broken :p | 16:49 |
mal | piggz: just add some .rc to libexec | 17:18 |
mal | piggz: the .rc filename is not important, just the service name inside it | 17:18 |
piggz | mal: cool, i was right then...just turns out i was careless and created it in the wrong place! | 17:56 |
piggz | mal: except it didnt work.... | 17:58 |
piggz | see https://paste.gnome.org/pds9uorz9 | 17:59 |
piggz | https://paste.gnome.org/pgvhxn3ie | 18:00 |
deathmist | mal: could my current ril_subscription.conf cause no audio during a call? I'm still using the Xperia XA2 one | 18:02 |
mal | deathmist: should not affect that, what device is that? | 18:50 |
deathmist | cheeseburger (OnePlus 5) on 15.1 | 18:51 |
mal | piggz: try removing disabled from last line | 18:52 |
mal | piggz: btw, typo in the fake binary name | 18:52 |
mal | deathmist: and miniaf is running? | 19:10 |
mal | I think it's miniafservice | 19:10 |
deathmist | yeah it's running, grepping for it in journalctl just returns 1 line: "droid-hal-init: starting service 'miniaf'..." | 19:11 |
deathmist | should I get new logs during call with my fixed up /etc symlinks for audio codecs etc in-place? | 19:12 |
piggz | mal: nope, it still tries to start time_daemon over and over | 19:43 |
mal | deathmist: maybe | 19:48 |
mal | piggz: check processing order from droid-hal-init startup in dmesg | 19:49 |
piggz | mal: yup, that got it.... | 19:51 |
piggz | [ 76.236484] droid-hal-init: Parsing file /usr/libexec/droid-hybris/system/etc/init/time_daemon.rc... | 19:51 |
piggz | [ 76.237067] droid-hal-init: /usr/libexec/droid-hybris/system/etc/init/time_daemon.rc: 1: ignored duplicate definition of service 'time_daemon' | 19:51 |
piggz | it finds init.qcom.rc first | 19:51 |
mal | hmm, I though libexec should be first in order, it seems not | 20:01 |
mal | piggz: is that init.qcom.rc in / on device? | 20:02 |
piggz | mal: /vendor/etc/init/hw/init.qcom.rc | 20:03 |
piggz | modify init to ensure libexec is first? | 20:03 |
mal | check how it processes those | 20:03 |
piggz | yup | 20:04 |
piggz | mal: ok, it is because it first parses /init.rc ... and that included the /vendor one | 20:06 |
piggz | parser.ParseConfig("/init.rc"); | 20:06 |
piggz | parser.ParseConfig("/usr/libexec/droid-hybris/system/etc/init"); | 20:06 |
piggz | init.rc contains | 20:07 |
piggz | import /vendor/etc/init/hw/init.${ro.hardware}.rc | 20:07 |
piggz | so, swap the order? we _always_ want libexec versions first | 20:08 |
mal | maybe | 20:10 |
piggz | mal: nope, another way | 20:19 |
usvi | ll | 20:19 |
usvi | whoops | 20:19 |
piggz | i bet the other services already do this, i just didnt look.... | 20:21 |
piggz | oh, no they dont, so I just learned a new thing :D | 20:22 |
piggz | mal: keyword override | 20:23 |
piggz | i would recommend adding that to the other services in libexec | 20:23 |
mal | nice, didn't know that | 20:24 |
piggz | mal: found it by looking at the 'ognored' error message.... | 20:25 |
piggz | if (!service_->is_override()) { | 20:25 |
piggz | return Error() << "ignored duplicate definition of service '" << service_->name() | 20:25 |
deathmist | nice, I managed to break audio routing outside calls by adding the media & codec symlinks from /vendor/etc to /etc; anyways mal here is journalctl https://pastebin.com/eBbcRgW6 and logcat https://pastebin.com/tw5wvCAd during a call | 21:53 |
deathmist | and here is before call journalctl https://pastebin.com/CZPgxQpA and logcat https://pastebin.com/fM4DWskY | 21:57 |
deathmist | r0kk3rz sorry for ping but could you check my PR in hybris-boot for cheeseburger, I'd appreciate it! | 22:07 |
T42 | <SALPHA8506> https://paste.ubuntu.com/p/nbbsTSXbFb/ | 22:13 |
deathmist | SALPHA8506 could you link me your local device manifest? | 22:14 |
T42 | <SALPHA8506> how | 22:16 |
deathmist | just put it in another paste, it is in $ANDROID_ROOT/.repo/local_manifests/ | 22:17 |
T42 | <SALPHA8506> (Document) https://irc.thaodan.de/.imgstore/7vwiQA48aT.xml | 22:18 |
deathmist | that's one weird manifest, it's ok tho since this is one of the hardest parts in the beginning to get right; lemme check this | 22:19 |
T42 | <SALPHA8506> ok | 22:20 |
T42 | <SALPHA8506> https://paste.ubuntu.com/p/kHtn7hSFn5/ | 22:23 |
deathmist | sudo apt update && sudo apt install -y imagemagick | 22:26 |
T42 | <SALPHA8506> @SALPHA8506 [https://paste.ubuntu.com/p/nbbsTSXbFb/], same after installing | 22:31 |
deathmist | move your kernel dir from "kernel/zuk/z2_plus" to "kernel/zuk/msm8996" because that's what is should be (also fix in manifest) | 22:32 |
deathmist | no need to resync, just rename the directory / do mv in terminal | 22:32 |
T42 | <SALPHA8506> ninja: error: 'libsf_compat_layer', needed by 'hybris-hal', missing and no kn | 22:33 |
deathmist | http://bit.ly/faq-hadk search for "15.1" and "3.0.3.9" a bit later, if you don't find answers to your issues from the FAQ or logs https://piggz.co.uk/sailfishos-porters-archive/index.php then ask here | 22:37 |
T42 | <SALPHA8506> got it | 22:37 |
deathmist | you should probably bookmark them both for easy access ;) be aware of the dates on logs tho as some of the stuff e.g. android-8-initial branches are very much deprecated now | 22:46 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!