| Thaodan | i intend to host the t4 bot I just didn't came to set it up | 04:20 |
|---|---|---|
| liux | video recorder frozes when press to stop recoding | 04:43 |
| liux | Jun 25 03:08:38 Sailfish pulseaudio[1719]: failed to set source 'source.primary' port to 'input-back_mic' | 04:43 |
| liux | Jun 25 03:08:38 Sailfish pulseaudio[1719]: can't set profiles/ports to source backmicrophone | 04:43 |
| liux | Jun 25 03:09:02 Sailfish pulseaudio[1719]: Failed to read from stream. (err -22) | 04:43 |
| liux | D/AudioStreamInALSA( 1719): pcm_read() returned n < 0 | 04:44 |
| liux | abranson, please take a look at this issue if you have time. thanks in advance | 04:45 |
| abranson | liux: sounds like the sound routing thinks there's a microphone present that isn't there. aren't they specified somewhere in the hadk | 06:25 |
| jusa | liux: pactl list sources , look for input-foo ports, if you have only one mic (input-back_mic) as it would seem, add to /etc/pulse/xpolicy.conf.d file xvars.conf with this: (adjust per your device) https://paste.ubuntu.com/p/cy4SrtDPKW/ | 07:15 |
| jusa | ah, I read the whole thing only after writing that.. hmm.. issue probably isn't the one I wrote, but more likely something to do how the hal wants to handle input streams, try adding quirks=-close_input to pactl load-module module-droid-card line | 07:19 |
| Mister_Magister | jusa: maybe you know! i have no sms notification sound do you maybe know what could be a problem? | 09:00 |
| Mister_Magister | mal: ping is BINDER_SET_CONTEXT_MGR bad uid 1000 != 0 while dhi starts servicemanager bad? | 11:58 |
| liux | abranson, jusa, thank you for your help. unfortunately patches are not work for my issue | 13:09 |
| liux | paste complete logcat log here https://paste.ubuntu.com/p/J3t4bMxgwK/ | 13:10 |
| jusa | liux: could you take pulseaudio logs as well? as nemo systemctl --user stop pulseaudio ; pulseaudio -vvvv -n --file=/etc/pulse/arm_droid_default.pa | 13:21 |
| liux | jusa, ok, let me try | 13:22 |
| Mister_Magister | ooh i'm so close to gui | 13:33 |
| liux | jusa, pulseaudio log pasted here: https://paste.ubuntu.com/p/FyySJ3htNx/ | 13:41 |
| liux | It's strange video recorder does not freeze after pulseaudio restarted, just no sound | 13:41 |
| mal | Mister_Magister: try this https://pastebin.com/7sDP5NZC | 13:43 |
| mal | to system/core | 13:44 |
| jusa | liux: paste was cut after 871 lines? just before the interesting part :) | 13:45 |
| Mister_Magister | mal: i just removed the if | 13:45 |
| Mister_Magister | but will keep in mind | 13:45 |
| Mister_Magister | mal: failed to create composer client | 13:45 |
| Mister_Magister | does it ring the bell? | 13:45 |
| mal | well you shouldn't just randomly remove some checks | 13:46 |
| liux | jusa, terminated by itself...seem like segment fault there: I: [pulseaudio] droid-source.c: FIXME: Setting AUDIO_DEVICE_IN_BUILTIN_MIC as initial device. | 13:46 |
| liux | Segmentation fault | 13:46 |
| mal | Mister_Magister: probably you have some selinux issues if you are missing the patch I gave | 13:46 |
| Mister_Magister | kay i will restore that if | 13:46 |
| mal | liux: try using gdb to get backtrace of the crash | 13:46 |
| jusa | liux: and also just checking, could you add /etc/pulse/arm_droid_card_custom.pa with following line: load-module module-droid-card quirks=-close_input | 13:48 |
| jusa | liux: and retry | 13:48 |
| Mister_Magister | mal: patch fails heh | 13:51 |
| Mister_Magister | i will do it manually then | 13:52 |
| liux | jusa, systemctl --user stop pulseaudio ; pulseaudio -vvvv -n --file=/etc/pulse/arm_droid_card_custom.pa? | 13:52 |
| liux | jusa, also terminated by segment fault | 13:54 |
| mal | Mister_Magister: could be some small differences in different revisions | 13:54 |
| Mister_Magister | mal: i have this http://paste.opensuse.org/view//29218508 | 13:56 |
| Mister_Magister | probably from different patch | 13:56 |
| Mister_Magister | changing "dummy" to "" will change anything? | 13:57 |
| mal | where did that patch come from? | 13:58 |
| Mister_Magister | from TheKit patching sources xd | 13:59 |
| mal | would be interesting to see what other changes there are | 13:59 |
| Mister_Magister | hmm might be complicated | 13:59 |
| mal | there shouldn't be need for other patches, well there are two other ones but those are related to different things | 14:00 |
| Mister_Magister | mal: "[system/core] (hybris) don't call getpeercon() in property_service as SELinux is disabled" | 14:01 |
| Mister_Magister | is the title of commit | 14:01 |
| Mister_Magister | only this file was changed | 14:01 |
| mal | Mister_Magister: ok, the patch I gave probably replaces that | 14:02 |
| Mister_Magister | mal: so just change "dummy" to ""? | 14:03 |
| mal | take a backup of the patch and drop it and add the patch I gave | 14:03 |
| liux | mal, is there any config for coredump on sfos? just set to unlimited but cannot generate core file | 14:04 |
| T42 | No chat_id set! Add me to a Telegram group and say hi so I can find your group's chat_id! | 14:04 |
| mal | liux: try installing crashreporter | 14:08 |
| T42 | No chat_id set! Add me to a Telegram group and say hi so I can find your group's chat_id! | 14:08 |
| mal | Thaodan: fix the bot | 14:08 |
| T42 | No chat_id set! Add me to a Telegram group and say hi so I can find your group's chat_id! | 14:08 |
| Thaodan | sry | 14:09 |
| T42 | No chat_id set! Add me to a Telegram group and say hi so I can find your group's chat_id! | 14:09 |
| mal | liux: I meant crash-reporter | 14:09 |
| liux | mal, took long time install it. I can't get it's output location like "CrashReporter". where does it store crash info? | 14:26 |
| Thaodan | who manages the telegram group? | 14:27 |
| mal | liux: /var/cache/core-dumps | 14:29 |
| Mister_Magister | mal: kay what i have to make | 14:36 |
| Mister_Magister | make init? | 14:36 |
| liux | mal, it's empty after pulseaudio crashed. am i use it in a wrong way? | 14:38 |
| mal | liux: why can't you just manually use gdb | 14:39 |
| Mister_Magister | mal: nope i'm still getting failed to create composer client | 14:57 |
| Mister_Magister | not anymore…ish? | 14:58 |
| Mister_Magister | yep it changed lmao | 14:59 |
| Mister_Magister | getting Adreno-GSL: <ioctl_kgsl_driver_entry:777>: open(/dev/kgsl-3d0) failed: errno 13. Permission denied | 14:59 |
| Mister_Magister | mal: https://paste.opensuse.org/view//2a5a2b6e | 15:01 |
| liux | cannot know where exactly crashed: #0 0xb6c0ab02 in ?? () from /lib/libpthread.so.0 | 15:09 |
| liux | lack of symbols | 15:09 |
| Mister_Magister | mal: permissions to ion and kgsl seem to be wrong | 15:09 |
| mal | Mister_Magister: have you made sure the uevent.rc is generated properly to out/ | 15:09 |
| Mister_Magister | they are different than on my x2 | 15:09 |
| Mister_Magister | uevent? | 15:09 |
| mal | *ueventd.rc | 15:09 |
| Mister_Magister | erm | 15:10 |
| Mister_Magister | i removed everything from /usr/libexec…/etc/init and placed .rc files from faq | 15:10 |
| mal | nothing to do with that | 15:10 |
| Mister_Magister | kk | 15:10 |
| mal | this about a common issue in hybris-hal build | 15:10 |
| mal | sometimes generating/copying that fails on 15.1 base or newer and some addition make command needs to be run or the file copied manually | 15:11 |
| Mister_Magister | out/target/blablabla/root/ueventd.rc | 15:11 |
| Mister_Magister | exists | 15:11 |
| mal | some devices also have another one in vendor | 15:12 |
| Mister_Magister | i have only root and recovery | 15:12 |
| mal | check your device repos and see if all have been copied | 15:12 |
| Mister_Magister | kk | 15:12 |
| Mister_Magister | in device/stuff? | 15:12 |
| mal | like I said | 15:13 |
| Mister_Magister | hmm | 15:14 |
| Mister_Magister | i can't find any ueventd.rc | 15:14 |
| mal | grep for kgsl-3d0 | 15:15 |
| mal | liux: so install debug symbols, it tells the commands when you run gdb, check the backtrace for which libs you might need those for | 15:16 |
| Mister_Magister | mal: found it, it's single file | 15:17 |
| mal | which one | 15:18 |
| Mister_Magister | root | 15:18 |
| Mister_Magister | and it's there /ueventd.rc | 15:18 |
| mal | I meant in device repo | 15:19 |
| Mister_Magister | mal: device/qcom/common/rootdir/etc/ueventd.qcom.rc | 15:20 |
| Mister_Magister | there is no file to set ion tho | 15:21 |
| Mister_Magister | mal…? | 15:24 |
| mal | so is that file in out/ | 15:25 |
| Mister_Magister | in root/ueventd.rc | 15:25 |
| mal | did you check the generated udev rules in droid-hal rpms | 15:26 |
| mal | if it sets the permissions there | 15:26 |
| Mister_Magister | wait no | 15:26 |
| Mister_Magister | they are different files | 15:27 |
| Mister_Magister | where does this file come from lo' | 15:27 |
| Mister_Magister | what udev rules in hal | 15:28 |
| mal | the ones generated during droid-hal packaging | 15:28 |
| mal | you can find those in one of the rpms in /lib/udev/rules.d folder or something like that | 15:28 |
| Mister_Magister | KERNEL=="kgsl", MODE="0666", GROUP="root", OWNER="root" | 15:29 |
| Mister_Magister | is there | 15:29 |
| Mister_Magister | can't find ion | 15:29 |
| mal | the ion errors could be unrelated, those were from other process | 15:30 |
| Mister_Magister | but… | 15:30 |
| mal | from adsprpc which probably would need fixing of symlinks in /etc | 15:30 |
| Mister_Magister | okay | 15:30 |
| Mister_Magister | what about udev rules then | 15:31 |
| mal | how many times do I have to say the same thing, you need to make sure the needed device rules are generated there, you said you don't have kgsl-3d0 there | 15:31 |
| Mister_Magister | i said " [17:29] <Mister_Magister> KERNEL=="kgsl", MODE="0666", GROUP="root", OWNER="root" [17:29] <Mister_Magister> is there" | 15:32 |
| Mister_Magister | oh | 15:32 |
| Mister_Magister | its only kgsl not kgsl-3d0 | 15:32 |
| Mister_Magister | why is it not kgsl-3d0 tho | 15:34 |
| mal | like I said, you need to make sure the correct .rc files are copied to out/ | 15:34 |
| mal | you said some file had kgsl-3d0 but not sure if that ended up in out/ | 15:34 |
| Mister_Magister | i'm wondering why device/qcom one differs from out | 15:35 |
| Mister_Magister | qcom one has kgsl kgsl-3d0 kgsl-2d0 | 15:36 |
| Mister_Magister | but in out there is only kgsl | 15:36 |
| mal | Mister_Magister: https://github.com/mer-hybris/android_system_core/blob/hybris-15.1/rootdir/ueventd.rc | 15:36 |
| mal | Mister_Magister: because those come from different repos | 15:36 |
| Mister_Magister | oh | 15:37 |
| mal | just do what I said | 15:37 |
| Mister_Magister | "you need to make sure the correct .rc files are copied to out/" then it should be in root/etc/ueventd.qcom.rc but it's not being copied there | 15:37 |
| Mister_Magister | or is it | 15:38 |
| Mister_Magister | no it's not | 15:38 |
| Mister_Magister | mal: what script copies them? | 15:38 |
| mal | some Android.mk or like that | 15:39 |
| mal | check the device repos like I said | 15:39 |
| Mister_Magister | not in hybris? | 15:39 |
| Mister_Magister | okay | 15:39 |
| * Mister_Magister doesn't know how to check | 15:40 | |
| Mister_Magister | oh wait | 15:40 |
| Mister_Magister | mal: we removed vendorimage from building | 15:41 |
| Mister_Magister | mal: that could cause the problem? | 15:41 |
| mal | yes | 15:41 |
| Mister_Magister | can i workaround it by copying it manually to device for the time being? | 15:42 |
| mal | there is something wrong with the way we handle vendorimage stuff | 15:42 |
| mal | not to device but to out/ to correct place (check makefiles where) | 15:42 |
| mal | it has to be in out/ before you package droid-hal | 15:42 |
| Mister_Magister | okay | 15:42 |
| mal | so you completely missed what I was talking about, i.e. that droid-hal packaging generates udev rules from .rc files | 15:43 |
| Mister_Magister | i think it's in root | 15:44 |
| Mister_Magister | mal: sorry | 15:44 |
| Mister_Magister | mal: LOCAL_MODULE_PATH := $(TARGET_OUT_VENDOR_ETC)/init/hw is that out folder? | 15:46 |
| mal | that doesn't like the place where the uevent*.rc should go | 15:47 |
| mal | read the source https://github.com/mer-hybris/droid-hal-device/blob/master/droid-hal-device.inc#L404 | 15:48 |
| Mister_Magister | i copied it to root lol | 15:51 |
| Mister_Magister | ok now i have proper udev rules | 15:57 |
| Mister_Magister | OH SHIT | 15:59 |
| Mister_Magister | i had 1ms of test_hwc2 there | 15:59 |
| Mister_Magister | ITS WORKING | 15:59 |
| mal | magic | 16:00 |
| Mister_Magister | mal: lol thank you | 16:00 |
| Mister_Magister | i'm debugging it today since 8 | 16:01 |
| Mister_Magister | i fixed dhi and servicemanager | 16:01 |
| mal | well it's quite obvious that permission issues like that are due to udev rules | 16:01 |
| Mister_Magister | yeah but these came out in late game | 16:01 |
| Mister_Magister | let's see if i get to gui | 16:02 |
| Mister_Magister | YAY | 16:02 |
| Mister_Magister | gui | 16:02 |
| Mister_Magister | how do i change pixelratio lol it's locked | 16:16 |
| Mister_Magister | mal: how do i update dconf | 16:39 |
| mal | dconf update | 16:42 |
| Mister_Magister | that does whole nothing | 16:42 |
| Mister_Magister | there was file to touch that made dconf update at reboot | 16:42 |
| mal | I usually touch the folder | 16:42 |
| mal | like the vendor.d | 16:42 |
| Mister_Magister | whet | 16:43 |
| mal | https://github.com/mlehtima/droid-config-fp2-sibon/blob/master/sparse/usr/bin/droid/droid-cameradetect.sh#L51 | 16:43 |
| Mister_Magister | i'm trying to change pixel ratio but it won't change | 16:44 |
| Mister_Magister | locks won't unlock | 16:44 |
| mal | did you remove the lock file? | 16:44 |
| Mister_Magister | yes | 16:44 |
| Mister_Magister | that didn't help either | 16:45 |
| Mister_Magister | error: The operation attempted to modify one or more non-writable keys | 16:46 |
| Mister_Magister | i even changed /etc/dconf/db/vendor.d/silica-configs.txt | 16:46 |
| Mister_Magister | and that did whole nothing | 16:46 |
| mal | and then the touch and dconf update? | 16:47 |
| Mister_Magister | after that | 16:47 |
| Mister_Magister | it does whole nothing | 16:47 |
| mal | no idea then | 16:48 |
| Mister_Magister | hmm | 16:49 |
| Mister_Magister | rild is not running and there is no rild binary | 16:49 |
| mal | are you sure it even should be, new android uses different way for ril stuff, you need a different ofono plugin and configs | 16:50 |
| Mister_Magister | oh | 16:50 |
| Mister_Magister | then i need those | 16:50 |
| Mister_Magister | but not today | 16:50 |
| Mister_Magister | i just wanna change pixel ratio | 16:50 |
| Mister_Magister | lol to modprobe wlan it symlinks to vendor | 16:54 |
| Mister_Magister | and insmod requires some key | 16:54 |
| vknecht | maybe need CONFIG_MODULE_SIG_FORCE=n | 16:55 |
| Mister_Magister | maybe | 16:55 |
| Mister_Magister | i need different pixel ratio tho | 16:56 |
| Mister_Magister | lol fuck curved screens. gestures aren't working | 16:56 |
| vknecht | or sign module https://unix.stackexchange.com/questions/74022/sign-a-module-after-kernel-compilation | 16:56 |
| Mister_Magister | secure boot is disabled so why lol | 16:57 |
| vknecht | don't think it has to do with secure boot ; hit that one but couldn't figure how to get the key, so went with signature disabling... | 16:58 |
| mal | Mister_Magister: you might want to adjust the touch areas in dconf | 16:58 |
| Mister_Magister | dconf won't update so what's the point | 16:58 |
| mal | well you are obviously doing something wrong | 16:59 |
| Mister_Magister | i'm not | 16:59 |
| Mister_Magister | i even installed themepacksupport to do it for me | 17:01 |
| Mister_Magister | and guess what | 17:01 |
| Mister_Magister | it's not working | 17:01 |
| Mister_Magister | mal: could you release your lineage16 hybris? | 17:34 |
| mal | I probably need to rebase it first | 17:37 |
| Mister_Magister | ;-; | 17:37 |
| T42 | SailfishFreenodeIRCBridgeBot was added by: abhishek_0 | 17:47 |
| Mr_Doge | Ping | 17:48 |
| T42 | <abhishek_0> Pong | 17:48 |
| Mister_Magister | @abhishek_0 that should be three way handshake | 18:00 |
| T42 | <DSstill01> Noo irc guys are back :( | 18:03 |
| T42 | <meierrom> :)) | 18:10 |
| anpep | Hello there, I'm having a bit of trouble porting Sailfish to my device, anyone up to help a noob? | 18:34 |
| kimmoli | ask, dont ask to ask | 18:38 |
| anpep | ok, so I've got the system past early boot and I'm able to telnet over RNDIS, but my system does not have vendor EGL libraries | 18:40 |
| anpep | I was wondering where am I supposed to put them, or if Sailfish should mount the /system partition at boot (which is not) | 18:40 |
| T42 | <:D (MISSING @USERNAME! telegram.org/faq#usernames-and-t-me)> is possible download sailfish | 18:41 |
| T42 | <:D (MISSING @USERNAME! telegram.org/faq#usernames-and-t-me)> for motorola g6+ | 18:42 |
| T42 | <:D (MISSING @USERNAME! telegram.org/faq#usernames-and-t-me)> ? | 18:42 |
| Mister_Magister | @abhishek_0 can u remove reply option from that bot? | 18:48 |
| T42 | <eugenio_g7> Also edit events | 19:11 |
| vknecht | and missing usernames :) | 19:11 |
| Thaodan | I disabled edit events already | 19:32 |
| Mister_Magister | u can make reply to look like on irc | 19:32 |
| Thaodan | How? Explain it and I'll change it | 19:33 |
| Mister_Magister | dunno how i don't have source | 19:34 |
| Thaodan | ah that looked more like an answer than a question | 19:35 |
| T42 | <eugenio_g7> Yeah I guess it would be nice if replying ---> prepends the username of the user you're replying to and then the message | 19:42 |
| T42 | <eugenio_g7> Surely it needs a bit of thought as here most replies would be against T4 | 19:42 |
| mal | anpep: that suggests you don't have correctly done fixup-mountpoints or something else is broken | 20:58 |
| anpep | mal: I don't know what could possibly be wrong in fixup-mountpoints, I correctly added partition mappings (I checked thoroughly) for my device | 21:21 |
| anpep | I followed the hardware adaptation guide but the built rootfs seems to be incomplete. For instance, vendor libraries are nowhere to be found on system root, nor many system libraries for that matter (libminisf.so, etc. and also hardware libraries like libgralloc.*.so) | 21:22 |
| anpep | Also, back to the /system partition, I don't really know why but mounting it does not work half of the time, even with ro,noload (disables journal loading) options | 21:24 |
| piggz | mal: http://ix.io/1Lfm in this file, near the top there is some kclock messages .... i dont think thats normal? | 21:35 |
| piggz | mal: and this is a full boot log, with simialar http://ix.io/1MJP | 21:37 |
| deathmist | anpep: which base? on 15.1 cheeseburger (OnePlus 5) I settled on adding /system & /vendor to the device fstab file and got to UI right after :p how is your defconfig? | 21:38 |
| anpep | 15.1 bardockpro, which actually has no /vendor partition | 21:38 |
| anpep | I guess it will be ok just adding /system to fstab and symlinking /vendor -> /system/vendor; however I think I'm not doing things right, I'll keep reading the wikis | 21:39 |
| deathmist | https://pastebin.com/MVQ2eQxA there's my current working fstab file, since you have a qcom device you (most likely) have a fstab.qcom somewhere as well; find it and try adding ln7 (modify src as needed) from that paste and attempt "ln -s /system/vendor /vendor && reboot" when you boot again | 21:47 |
| deathmist | I found the line in a "recovery.fstab" file. just adding it to the regular fstab made everything work for me, but I have no clue if this still is the proper way to go; will do more research once I get the basics working on the adaptation :) | 21:50 |
| anpep | deathmist: trying that out now | 21:53 |
| anpep | I found no fstab.qcom, by the way | 21:53 |
| mal | some file with fstab in the name in device repo or one of the device repos if you have shared repos | 21:55 |
| anpep | same strange problem as before, mounting /system manually works but putting an entry in fstab fails silenty and attempting to mount manually somehow hangs the system, doing some debugging | 21:56 |
| deathmist | check out what find $ANDROID_ROOT/device -name "*fstab*" on ANDROID_ROOT and (if needed) other wildcard variations return to get the proper file | 21:56 |
| mal | anpep: what do you adding it to fstab, you mean to device repo fstab and then building hybris-hal and repackaging droid-hal? | 21:57 |
| mal | *what do you mean by adding | 21:57 |
| anpep | I'm working directly on the device to try and make it work before I do any changes on the system, guess that's not the way to go | 21:58 |
| deathmist | anpep: no matches with find $ANDROID_ROOT/device -name "*fstab*" ? | 22:09 |
| anpep | yes, found fstab.qcom with all the correct entries | 22:11 |
| anpep | I must be doing something wrong because those entries did not end up anywhere in hybris img or /system | 22:12 |
| deathmist | maybe check the fstab.qcom file returned by $ANDROID_ROOT/out -name "*fstab*", it should be identical to the one you already found. you are booting with hybris-boot image right? | 22:14 |
| anpep | there is no fstab.qcom at out/ | 22:18 |
| deathmist | ok what about the other ones? btw I also checked the logs your provided and found "cannot run command `/sbin/droid-hal-init': No such file or directory", which is no good if you want any kind of UI really | 22:19 |
| deathmist | I'm not experience enough to know which package creates /sbin/droid-hal-init, but did you adapt droid-hal-device, droid-hal-configs and droid-hal-version for your phone? this should be defined by default: "jolla-startupwizard-pre-user-session[3444]: [W] unknown:0 - No pixel ratio defined. Check your device configuration!" | 22:22 |
| deathmist | it's all in hadk pdf chapter 7 | 22:25 |
| mal | anpep: sailfish doesn't use fstab on device | 22:48 |
| mal | anpep: the fstab is parsed and the sfos build generates systemd mount service from those using fixup-mountpoints as help | 22:49 |
| mal | *mount services | 22:49 |
| mal | assuming the android build properly copies the fstab to out/ | 22:50 |
| anpep | mal: thanks, I didn't know that. As for the missing files, I may have misconfigured device-specific repositories | 22:50 |
| anpep | but I'm not really sure, looking into that now | 22:51 |
| mal | some newer bases need some extra command like make fstab.qcom or something like that, check output of "make modules" in HABUILD_SDK | 22:51 |
| anpep | I think the problem is that there is no single tree for my device | 22:51 |
| mal | anpep: the is a known bug in hybris-hal build that sometimes fstab and uevent*.rc files are not properly copies in hybris-15.1 base | 22:51 |
| mal | *there is | 22:52 |
| anpep | I mean, I have multiple device/ dependencies on my manifest, could it be that only files from bq/bardockpro are copied, even when it depends also on bq/msm8953-common? | 22:53 |
| mal | that is usually not a problem | 22:54 |
| mal | just check "make modules" like I said | 22:54 |
| mal | for things like fstab and ueventd*.rc | 22:54 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!