| *** Danct12 is now known as Guest8981 | 01:30 | |
| T42 | <k1gen> mal: https://vomitb.in/raw/QQzMg3nM2y | 08:05 |
|---|---|---|
| T42 | <k1gen> mal: you said earlier "not sure yet why that one variable is null which then causes the crash". does this mean I should just wait for you to fix the sound issue upstream? sorry if I'm being cheeky, I just think you and elros34 noticed that without you I have zero chance of getting the sound to work. should I hope for a magical sound fix or forget about it for some time? | 09:04 |
| T42 | <elros34> @b100dian based on fact there is nothing in device dmesg about connected usb I wonder if this is the case for your device: https://review.lineageos.org/c/LineageOS/android_device_xiaomi_sm6150-common/+/291494/4 | 09:20 |
| T42 | <b100dian> Wow, that's really a must try! Thanks @elros34 . I can confirm I have that line in my init recovery. I'll test as soon as I am able to and report back. | 09:26 |
| T42 | <edp_17> @elros34 : Hi. On OBS, I've added kfbluezqt package and built revision: 430824e3316d322a75a6f923003253c117f05861. It's built but I only got kfbluezqt-bluez5*.rpm packages in the repo however, I'd need bluez4. I guess this is not the way you suggested to bring support back to kfbluezqt. :) Can you help me out please? | 09:35 |
| T42 | <edp_17> I've tried to set back to revision 9e70d50ced79979c54c2b97da966666b5af41ad7 which is the latest it has bluez4, but got error in OBS. | 09:41 |
| T42 | <b100dian> @elros34 I got telnet! | 09:47 |
| Mister_Magister | LETSGOOOOOOOOOOO | 09:48 |
| Mister_Magister | i mean congrats | 09:48 |
| T42 | <b100dian> note to self: also search recovery init.rc s not only vendor/system ones | 09:51 |
| Mister_Magister | grep asd $(find / -name *.rc) xd | 09:51 |
| T42 | <b100dian> but then I may have found that line and ignored it. Elros found it with the nice, unignorable, explanation | 09:52 |
| T42 | <elros34> awesome, I figure out it after reading your post on forum: dmesg from lineage recovery, usb was initialized when ueventd started | 09:59 |
| T42 | <edp_17> This was the right revision but I had to remove the bluez5 .spec file. Now I have kfbluezqt-bluez4*.rpm packages in my repo. :D Hopefully they will work. (re @edp_17: I've tried to set ba...) | 10:02 |
| T42 | <elros34> @edp_17 if it would only be so simple like building older revision. If you don't want backported bluetooth in kernel then you must apply removed bluez4 patches from older kfbkuezqt version on top of update kfbbluez. I expected its not trivial and needs understanding the code that is why I didn't even try it | 10:03 |
| T42 | <elros34> that would work only if you could build closed parts from jolla which depends on newer kfbluezqt | 10:04 |
| T42 | <edp_17> Oh, okay. | 10:05 |
| T42 | <edp_17> So, you are saying the easiest way is getting bluetooth driver backported into kernel? | 10:08 |
| T42 | <elros34> not sure which is easier but I already spened some time on driver and get it working partially | 10:09 |
| T42 | <edp_17> @elros34 : You are right. On OBS building "pulseaudio-modules-bluez4" is unresolvable because "have choice for pkgconfig(bluez) >= 4.101: bluez-libs-devel bluez5-libs-devel" | 10:19 |
| T42 | <edp_17> What does it mean? | 10:19 |
| T42 | <Mister_Magister> don't use bluez4 it's nonexistent anymore | 10:22 |
| T42 | <Mister_Magister> use bluez5 | 10:22 |
| mal | you need to set prefer in project config https://build.sailfishos.org/project/prjconf/nemo:devel:hw:fairphone:fp2-sibon if there is conflict, that has some example for other packages | 10:22 |
| T42 | <b100dian> @elros34 It's a proof that there's no such thing as oversharing:P. I now cannot unsee that line saying "Manually set USB mode peripheral". Thanks again! | 10:25 |
| T42 | <elros34> @edp_17 IIRC real issue is lipstick-jolla-home-qt5 other open source packages you can rebuild | 10:28 |
| T42 | <edp_17> I think the problem is that I have droid-config-$DEVICE-bluez5 instead of droid-config-$DEVICE-bluez4 | 10:38 |
| T42 | <edp_17> Hmm, no, it should be something else. Because on older sfos version I also only have droid-config-$DEVICE-bluez5 | 10:40 |
| T42 | <edp_17> mal : I've set Prefer: droid-config-$DEVICE-bluez4 in Project Config but didn't help. | 10:43 |
| T42 | <elros34> problem is libKF5BluezQt.so.5 on 4.4.0 -> libKF5BluezQt.so.6 on 4.5.0 required by lipstick so even if you make obs build working this will not help | 10:49 |
| T42 | <edp_17> Great. :( | 10:50 |
| mal | well there is the option of porting the bluez4 changes to kf5bluez | 10:51 |
| T42 | <elros34> yeas that is what I trying to explain earlier either kf5bluez or kernel backporting | 10:52 |
| *** amccarthy is now known as Guest9018 | 10:52 | |
| *** amccarthy_ is now known as amccarthy | 10:52 | |
| T42 | <edp_17> Now it seems kernel backporting is the only futureeproof way. Otherwise, might need to do this battle after every sfos update. | 10:54 |
| mal | @k1gen pulseaudio-modules-droid crash you had, the one which needed changes to the xml, has been fixed if you build latest version | 11:29 |
| T42 | <k1gen> mal: how do I pull changes so I can rebuild it? | 11:39 |
| mal | build_packages.sh --mw=pulseaudio-modules-droid | 11:41 |
| mal | I think that should pull in updates, it should mention it in the output | 11:42 |
| T42 | <k1gen> yeah, it does | 11:42 |
| T42 | <elros34> have you tried using pulseaudio-modules-droid-jb2q instead? No idea if that has any chance to work | 11:43 |
| T42 | <k1gen> no, I haven't yet. let me try the new pulseaudio-modules-droid version first | 11:44 |
| T42 | <elros34> don't forget to remove modified xml | 11:45 |
| T42 | <k1gen> mal: https://vomitb.in/raw/mczVK8dBGT | 11:56 |
| T42 | <k1gen> without gdb: https://vomitb.in/raw/G1PywqxHVZ | 11:57 |
| T42 | <k1gen> I deleted /etc/audio_policy_configuration.xml | 11:57 |
| mal | are you sure you used HYBRIS_USE_VENDOR_NAMESPACE=yes ? | 11:59 |
| mal | when using gdb | 12:00 |
| mal | at least in the latter link you probably didn't use it | 12:00 |
| T42 | <k1gen> mal: yes, sorry, should've exported it: https://vomitb.in/raw/XdASuLUY4v | 12:01 |
| mal | I assume the backtrace is the same as before? | 12:02 |
| mal | hmm, but it's doesn't show as much there now | 12:02 |
| T42 | <k1gen> yeah | 12:05 |
| T42 | <elros34> did you install new debug packages? | 12:05 |
| T42 | <k1gen> sure | 12:05 |
| T42 | <k1gen> pulseaudio-modules-droid, -common, -devel, -hidl | 12:05 |
| T42 | <elros34> debug not devel | 12:06 |
| T42 | <k1gen> used zypper in -f --replacefiles pulseaudio-modules-droid-* | 12:06 |
| T42 | <k1gen> there is no -debug package | 12:07 |
| mal | can you get output of this: HYBRIS_USE_VENDOR_NAMESPACE=yes strace -o outfile.txt -f -s2048 pulseaudio --daemonize=no -n --file=/etc/pulse/arm_droid_default.pa -vvvvvvvvvv | 12:07 |
| T42 | <k1gen> sure, a minute | 12:07 |
| mal | @elros34 afaik packages built with build_packages.sh don't have separate debug packages, symbols are included always | 12:08 |
| T42 | <elros34> you are right, I have it changed since always | 12:08 |
| mal | that command I have should write the output to outfile.txt if I did it correctly | 12:11 |
| T42 | <k1gen> mal: I'm sorry, ix.io doesn't return the link | 12:11 |
| T42 | <k1gen> any other way to send a file so big contents can't be copied separately? | 12:12 |
| T42 | <elros34> about that gdb you can also try: thread apply all bt full | 12:13 |
| T42 | <k1gen> @elros34: https://vomitb.in/raw/TYMMjGRS46 | 12:14 |
| mal | that failure seems different than the one yesterday, the backtrace is not very useful though in this case, not sure why | 12:20 |
| T42 | <elros34> do you watch also logcat during tests? | 12:21 |
| mal | that could be useful indeed | 12:21 |
| T42 | <k1gen> @elros34 no, I didn't. what should I repeat with logcat in background? | 12:22 |
| T42 | <k1gen> let me reboot so it can be readable | 12:22 |
| T42 | <elros34> you should watch what happens during pulseaudi start and crash aslo mask pulseaudio.service | 12:23 |
| T42 | <edp_17> When building droid-hal-version on OBS, I am getting unresolved: "have choice for hybris-libsensorfw-qt5: hybris-libsensorfw-qt5-binder hybris-libsensorfw-qt5-hal" | 12:25 |
| T42 | <edp_17> I am using "hybris-libsensorfw-qt5" (not the hal version) in patterns and a previous version was built successfully. | 12:26 |
| mal | did you accidentally build the hal package? | 12:28 |
| mal | you should not build it if you don't need it | 12:28 |
| T42 | <elros34> same for yamui, do not build something which is in common repo | 12:31 |
| T42 | <edp_17> Hmm, yes. there is hybris-libsensorfw-qt5-hal*.rpm package in the repo, but how? | 12:39 |
| mal | which android base are you using? | 12:39 |
| T42 | <edp_17> Lineage17.1 | 12:39 |
| T42 | <elros34> because you are building it https://build.merproject.org/package/show/nemo:devel:hw:motorola:ocean/sensorfw-qt5-hybris | 12:39 |
| mal | yeah, that should not be there if you use binder backend | 12:41 |
| mal | I removed that | 12:41 |
| T42 | <edp_17> Okay, I've removed it. I just wonder how did it work before for 4.4? 🤔 | 12:42 |
| T42 | <edp_17> droid-hal-version is buuilding now. Thanks! | 12:43 |
| T42 | <elros34> remove yamui.. | 12:43 |
| mal | also libglibutil is probably not something that should be there | 12:44 |
| mal | @edp_17 are you intentionally building that newew version of yamui? | 12:45 |
| T42 | <edp_17> I excluded both yamui* from 4.5. | 12:45 |
| T42 | <edp_17> mal : No, it wasn't intentional. I thought should use the latest version. | 12:48 |
| T42 | <edp_17> So, should I remove all these: libglibutil, yamui, yamuisplash ? | 12:48 |
| mal | well depends if you want to have the new version, it has some benefits | 12:52 |
| T42 | <k1gen> mal: logcat: https://vomitb.in/raw/chgckgd6G5, stdout is same | 12:52 |
| T42 | <edp_17> mal : I don't know the benefit of newer version of yamui. What about libglibutil? | 12:57 |
| T42 | <TheVancedGamer> @elros34 i'm back at where I was, now I get the following errors: | 12:57 |
| T42 | <TheVancedGamer> https://hastebin.com/share/ovoxexoxol.bash | 12:57 |
| T42 | <TheVancedGamer> doing `make allmod` states there is no such target | 12:58 |
| T42 | <elros34> allmod not make allmod | 12:59 |
| T42 | <TheVancedGamer> oh | 12:59 |
| T42 | <TheVancedGamer> my bad, sorry | 12:59 |
| T42 | <TheVancedGamer> what am I looking for there? | 13:00 |
| T42 | <elros34> and provide full logs, few lines is just waste of time because I could already answer partial of your question | 13:00 |
| T42 | <TheVancedGamer> this was the entire droid-hal-angelica.log file | 13:00 |
| T42 | <elros34> hmm so search for same: fstab ueventd | 13:02 |
| T42 | <TheVancedGamer> ``` | 13:02 |
| T42 | <TheVancedGamer> build/make/core/board_config.mk:TARGET_COPY_OUT_RAMDISK := $(TARGET_COPY_OUT_ROOT)``` | 13:02 |
| T42 | <TheVancedGamer> | 13:02 |
| T42 | <TheVancedGamer> there was this, but still confuses me if TARGET_COPY_OUT_ROOT's value is root, then why does fstab end up in ramdisk folder | 13:02 |
| T42 | <TheVancedGamer> ``` | 13:02 |
| T42 | <TheVancedGamer> HABUILD_SDK [angelica] [thevancedgamer@ProDesk-arch hadk]$ allmod | grep fstab | 13:02 |
| T42 | <TheVancedGamer> fstab.mt6762 | 13:02 |
| T42 | <TheVancedGamer> fstab.mt6765 | 13:02 |
| T42 | <TheVancedGamer> fstab.postinstall | 13:02 |
| T42 | <TheVancedGamer> HABUILD_SDK [angelica] [thevancedgamer@ProDesk-arch hadk]$ allmod | grep ueventd | 13:03 |
| T42 | <TheVancedGamer> ueventd.rc | 13:03 |
| T42 | <TheVancedGamer> ueventd.rc.recovery``` | 13:03 |
| T42 | <elros34> choose correct one and build | 13:03 |
| T42 | <TheVancedGamer> but, I am? | 13:03 |
| T42 | <TheVancedGamer> as I said, fstab ends up in ramdisk/ and etc/vendor_overlay | 13:03 |
| T42 | <elros34> ueventd too? | 13:04 |
| T42 | <TheVancedGamer> ueventd is in system/etc | 13:04 |
| T42 | <elros34> maybe just copy fstab and that will be enough for script | 13:05 |
| T42 | <TheVancedGamer> https://hastebin.com/share/uqereqekir.bash | 13:06 |
| T42 | <TheVancedGamer> I feel like maybe root/ is unnecessary now | 13:06 |
| T42 | <TheVancedGamer> ramdisk serves the same purpose anyhow | 13:06 |
| T42 | <k1gen> mal: any thoughts on pulseaudio logcat? also wanted to ask how can I fix failing metadata.mount and mnt-vendor-persist.mount | 13:08 |
| mal | @elros34 @TheVancedGamer ueventd.rc is needed in out also to make udev rules | 13:10 |
| T42 | <TheVancedGamer> realized that with hybris-17.1 | 13:10 |
| T42 | <TheVancedGamer> ``` | 13:11 |
| T42 | <TheVancedGamer> prebuilt_etc { | 13:12 |
| T42 | <TheVancedGamer> name: "ueventd.rc", | 13:12 |
| T42 | <TheVancedGamer> src: "ueventd.rc", | 13:12 |
| T42 | <TheVancedGamer> recovery_available: true, | 13:12 |
| T42 | <TheVancedGamer> }``` | 13:12 |
| T42 | <TheVancedGamer> because i've disabled vendor partition build this would install it in /system/etc which it does | 13:12 |
| T42 | <elros34> but AFAIK ueventd doesn't need to be in root anyway you will se in droid-hal-device.log | 13:17 |
| T42 | <TheVancedGamer> but for now, what do I put in root? | 13:17 |
| T42 | <elros34> what I sad fstab | 13:17 |
| T42 | <elros34> said* | 13:17 |
| T42 | <TheVancedGamer> ok | 13:18 |
| T42 | <elros34> later you may check if your ueventd on /vendor or somwhere on device is same as then one in out/ | 13:19 |
| T42 | <TheVancedGamer> @elros34 https://hastebin.com/share/opirajurok.bash | 13:25 |
| T42 | <TheVancedGamer> should I put all device-specific init there as well? | 13:25 |
| T42 | <elros34> not sure if you should, either add '|| true' to that sed command or copy some rc files to be sure all of them are scanned and any unusual mountpoint is extracted | 13:29 |
| T42 | <TheVancedGamer> so now I have https://hastebin.com/share/ihovumeboh.bash | 13:44 |
| T42 | <TheVancedGamer> init.zygote* aren't in modules, they aren't anywhere in the out dir, yet they are defined in build/make | 13:45 |
| T42 | <elros34> you don't need zygote so I wouldn't worry about that too much | 13:45 |
| T42 | <TheVancedGamer> but it's failing | 13:45 |
| T42 | <elros34> || true | 13:46 |
| T42 | <TheVancedGamer> ok | 13:46 |
| *** attah_ is now known as attah | 13:46 | |
| T42 | <TheVancedGamer> where do I add it though? | 13:49 |
| T42 | <TheVancedGamer> sorry for the questions, first time getting this far | 13:49 |
| T42 | <elros34> you see where it fails so search for "system/etc/init/hw/init.zygote*.rc" in droid-hal-device.inc | 13:50 |
| T42 | <elros34> btw do you use correct device repo for lineage 18? | 13:51 |
| T42 | <TheVancedGamer> oh | 13:51 |
| T42 | <TheVancedGamer> so that's how it works | 13:51 |
| T42 | <TheVancedGamer> github.com/redmi-mt6765/android_device_xiaomi_dandelion | 13:51 |
| T42 | <elros34> thats the fstab not such tiny few lines one: https://github.com/Redmi-MT6765/android_device_xiaomi_dandelion/blob/lineage-18.1/rootdir/etc/fstab.mt6762 | 13:56 |
| T42 | <TheVancedGamer> I found it already, yes | 13:56 |
| T42 | <TheVancedGamer> 6762 one is for 9A | 13:57 |
| T42 | <TheVancedGamer> (I made half the tree if that helps :P) | 13:57 |
| T42 | <elros34> there is one mount bind in init.mt6765.rc | 13:58 |
| T42 | <TheVancedGamer> now it wants default.prop | 13:59 |
| T42 | <TheVancedGamer> which I also know is named prop.default, none of which are located in the out dir, or in allmod | 13:59 |
| T42 | <elros34> picky droid hal:) | 13:59 |
| T42 | <elros34> ignore it:P | 13:59 |
| T42 | <TheVancedGamer> || true as well? | 13:59 |
| T42 | <elros34> no its not in script section | 14:00 |
| T42 | <TheVancedGamer> but it's fatal | 14:00 |
| T42 | <TheVancedGamer> how do I ignore it now? | 14:00 |
| T42 | <elros34> remove it | 14:00 |
| T42 | <TheVancedGamer> worked, thanks | 14:04 |
| T42 | <TheVancedGamer> (it was in droid-hal-device.inc) | 14:04 |
| T42 | <TheVancedGamer> getting ccache not found while building middleware, hadk-hot isn't opening so I can't find the command | 14:16 |
| T42 | <elros34> channel logs is also great source of knowledge | 14:16 |
| T42 | <k1gen> yeah, I had the same error | 14:21 |
| T42 | <TheVancedGamer> I remember mal doing something with dhd module to fix kf5bluezqt, did I miss something? | 14:23 |
| mal | @TheVancedGamer I haven't yet made the PR to remove that useless build from it and also some other fixes | 14:27 |
| T42 | <k1gen> mal: I haven't moved on with broken sound, waiting for your answer on logcat I sent | 14:28 |
| mal | @k1gen could you add file /etc/pulse/arm_droid_card_custom.pa with this content: load-module module-droid-card hw_volume=false | 14:28 |
| T42 | <k1gen> sure | 14:29 |
| T42 | <k1gen> give me couple of minutes | 14:29 |
| T42 | <TheVancedGamer> @elros34 https://www.toptal.com/developers/hastebin | 14:35 |
| T42 | <TheVancedGamer> I did build mw, no errors there, gg was fine as well | 14:35 |
| T42 | <TheVancedGamer> (building droid-hal-version) | 14:35 |
| T42 | <k1gen> mal: segfault on the same place, logcat: https://vomitb.in/raw/fMFCojxfv6 | 14:35 |
| T42 | <k1gen> stdout of pa: https://vomitb.in/raw/hHSj6ghHUf | 14:36 |
| mal | please use gdb so we can be sure there is no difference | 14:37 |
| T42 | <k1gen> of course: https://vomitb.in/raw/4fkXCyiCVM | 14:40 |
| T42 | <elros34> whats that "Airbrush Manager"? | 14:40 |
| T42 | <k1gen> I have no idea | 14:41 |
| T42 | <elros34> what are the permission of ls -al /dev/ab_sm | 14:43 |
| T42 | <k1gen> [defaultuser@flame ~]$ ls -al /dev/ab_sm | 14:43 |
| T42 | <k1gen> crw------- 1 root root 10, 27 Mar 27 15:48 /dev/ab_sm | 14:43 |
| T42 | <elros34> do you have ueventd* somwhere in /vendor or /system? | 14:44 |
| T42 | <k1gen> [root@flame defaultuser]# find 2>/dev/null /vendor -iname ueventd* | 14:45 |
| T42 | <k1gen> /vendor/ueventd.rc | 14:45 |
| T42 | <k1gen> [root@flame defaultuser]# find 2>/dev/null /system -iname ueventd* | 14:45 |
| T42 | <k1gen> /system/bin/ueventd | 14:45 |
| T42 | <k1gen> /system/etc/ueventd.rc | 14:45 |
| T42 | <elros34> does any of it have ab_sm entry? | 14:45 |
| T42 | <k1gen> [root@flame defaultuser]# grep 'ab_sm' /vendor/ueventd.rc | 14:46 |
| T42 | <k1gen> /dev/ab_sm 0600 vendor_airbrush vendor_airbrush | 14:46 |
| T42 | <k1gen> [root@flame defaultuser]# grep 'ab_sm' /system/etc/ueventd.rc | 14:46 |
| T42 | <k1gen> [root@flame defaultuser]# | 14:46 |
| T42 | <elros34> so you need to copy that file to out/** according to hadk-hot and build droid-hal. Can you show whole file maybe there something audio related there too | 14:47 |
| T42 | <k1gen> /vendor/ueventd.rc? | 14:48 |
| T42 | <elros34> yeah | 14:48 |
| T42 | <elros34> @TheVancedGamer probably same for your device | 14:49 |
| T42 | <k1gen> @elros34 http://ix.io/4rYi | 14:49 |
| T42 | <elros34> there is something, you have to compare it with /dev on device or build new droid-hal with updated rules | 14:52 |
| T42 | <k1gen> I'm sorry, I don't understand what should I do exactly | 14:53 |
| mal | @k1gen could you try HYBRIS_ENABLE_LINKER_DEBUG_MAP=1 HYBRIS_USE_VENDOR_NAMESPACE=yes gdb --args pulseaudio --daemonize=no -n --file=/etc/pulse/arm_droid_default.pa -vvvvvvvvvv | 14:53 |
| T42 | <k1gen> mal: https://vomitb.in/raw/6FiScDXUBk | 14:54 |
| mal | hmm, still not useful backtrace | 15:06 |
| T42 | <k1gen> :C | 15:07 |
| T42 | <k1gen> very sad. I can live without sensors and vibro, even without camera if I really want to, but without sound... | 15:12 |
| T42 | <k1gen> especially after we got everything else working flawlessly | 15:12 |
| mal | you never gave the things I asked for vibrator | 15:19 |
| T42 | <Mister_Magister> Audio is pain, sometimes unfixable | 15:26 |
| T42 | <Mister_Magister> Is this motorola (re @k1gen: especially after we ...) | 15:26 |
| T42 | <k1gen> google, close one | 15:26 |
| T42 | <Mister_Magister> Ah yes | 15:26 |
| T42 | <Mister_Magister> Should've expected that from them | 15:27 |
| T42 | <k1gen> mal: actually, I did: today at 00:37 and 11:05. https://vomitb.in/raw/BRPlfDvnJf | 15:28 |
| T42 | <elros34> you are lucky anything works without proper permissions for /dev/* | 15:29 |
| T42 | <k1gen> @elros34 you didn't explain what you meant by "there is something, you have to compare it with /dev on device or build new droid-hal with updated rules" | 15:30 |
| T42 | <elros34> I told you to search in hadk-hot, its there ueventd. You think I wrote all this lines to repeat again and again know solutions here | 15:30 |
| T42 | <k1gen> got it, thank you | 15:31 |
| T42 | <k1gen> @elros34 diff out/target/product/flame/vendor/ueventd.rc out/target/product/flame/vendor/ueventd.flame.rc returns nothing | 15:58 |
| T42 | <k1gen> ueventd.flame.rc is the one I scp'd from device | 15:59 |
| T42 | <elros34> what about grep ab_sm /lib/udev/rules.d/99?-* | 16:05 |
| T42 | <k1gen> [root@flame defaultuser]# grep ab_sm /lib/udev/rules.d/99?-* | 16:06 |
| T42 | <k1gen> /lib/udev/rules.d/999-android-system.rules:#:/dev/ab_sm 0600 vendor_airbrush vendor_airbrush | 16:06 |
| T42 | <k1gen> /lib/udev/rules.d/999-android-system.rules:KERNEL=="ab_sm", MODE="0600", GROUP="vendor_airbrush", OWNER="vendor_airbrush" | 16:06 |
| T42 | <edp_17> @k1gen : I just realized last time I gave you wrong command for running the mic manually. You might figured out already. If you use tar.gz instead of tar.bz2, you'll need to use this: sudo mic create fs --arch=$PORT_ARCH --tokenmap=ARCH:$PORT_ARCH,RELEASE:$RELEASE,EXTRA_NAME:$EXTRA_NAME --record-pkgs=name,url --outdir=sfe-$DEVICE-$RELEASE$EXTRA_NAME --pack-to=sfe-$DEVICE-$RELEASE$EXTRA_NAME.tar.gz $ANDROID_ROOT/Jolla-@RELEASE@-$DEVICE-@ARCH@.k | 16:06 |
| T42 | <k1gen> @edp_17 yeah, I realized this on my own, it wasn't hard. you helped a lot with initial command syntax | 16:06 |
| T42 | <elros34> ok so this looks correct except permission is probably overwritten by some init*rc file | 16:07 |
| mal | @k1gen the other patch I asked is not in that paste | 16:10 |
| mal | *path | 16:10 |
| T42 | <k1gen> which one then? | 16:10 |
| mal | something like /sys/class/leds/vibrator* | 16:11 |
| T42 | <k1gen> mal: [root@flame defaultuser]# ls -l /sys/class/leds/vibrator | 16:12 |
| T42 | <k1gen> lrwxrwxrwx 1 root root 0 Mar 27 15:48 /sys/class/leds/vibrator -> ../../devices/platform/soc/c94000.i2c/i2c-4/4-0043/leds/vibrator | 16:12 |
| mal | what are the permissions of that target? | 16:13 |
| T42 | <k1gen> @elros34 https://vomitb.in/raw/OJrkv2yskI one of these? | 16:13 |
| T42 | <elros34> I guess | 16:13 |
| T42 | <k1gen> mal: https://vomitb.in/raw/aMB7BLFPMF | 16:14 |
| T42 | <k1gen> @elros34 how can I prevent this? delete ab_sm line in those files, reboot and check? | 16:15 |
| T42 | <elros34> more or less | 16:15 |
| mal | @k1gen looks like a permissions issue, those activate and duration paths should have write permission also for group | 16:17 |
| T42 | <k1gen> @elros34 no ab_sm in init*rc on device: https://vomitb.in/raw/TpUaxB9nB0 | 16:18 |
| T42 | <k1gen> mal: chmod g+rw? | 16:21 |
| T42 | <elros34> do you even have vendor_airbrush? what happen when you chown file | 16:23 |
| T42 | <k1gen> [root@flame defaultuser]# find 2>/dev/null / -iname airbrush | 16:24 |
| T42 | <k1gen> /sys/kernel/debug/airbrush | 16:24 |
| T42 | <k1gen> /sys/kernel/debug/tracing/events/airbrush | 16:24 |
| T42 | <k1gen> /sys/kernel/debug/tracing/instances/pixel-trace/events/airbrush | 16:24 |
| T42 | <k1gen> /sys/kernel/debug/tracing/instances/wifi/events/airbrush | 16:24 |
| T42 | <k1gen> /sys/firmware/devicetree/base/soc/thermal-zones/pm8150b-vbat-lvl0/cooling-maps/airbrush | 16:24 |
| T42 | <k1gen> /data/vendor/hardware/airbrush | 16:24 |
| T42 | <k1gen> chown what file exactly? | 16:24 |
| T42 | <elros34> the same /dev/ab_sm why are talking about | 16:25 |
| T42 | <k1gen> [root@flame defaultuser]# chown vendor_airbrush:vendor_airbrush /dev/ab_sm | 16:26 |
| T42 | <k1gen> chown: unknown user/group vendor_airbrush:vendor_airbrush | 16:26 |
| mal | @k1gen yes, that chmod should fix it, but proper fix that remains after reboot probably should be in .rc files or udev, depends on how the device does things | 16:28 |
| mal | that is for vibrator | 16:28 |
| T42 | <k1gen> mal: https://vomitb.in/raw/EAg7Ge1bxF, no vibration in csd | 16:31 |
| mal | did you restart ngfd? | 16:36 |
| T42 | <k1gen> I did now, it works. how can I make this persistent? | 16:38 |
| mal | check files in /vendor/etc/init/* and /system/etc/init/* and /usr/libexec/droid-hybris/system/etc/init/* for things changing permission of that path, so grep for chmod.*vibrator | 16:40 |
| mal | I used this kind of file https://github.com/mlehtima/droid-config-fp4/blob/devel/sparse/usr/libexec/droid-hybris/system/etc/init/vibrator.rc | 16:43 |
| mal | note that the paths are different than the ones you have | 16:44 |
| mal | and the last part might not be needed, but if you have some vibrator service running then disable it like that based on the service name you have | 16:44 |
| T42 | <k1gen> I don't think in any of directories you listed there's chmod*vibrator line | 16:45 |
| T42 | <TheVancedGamer> mal: question about the FP4 port, where did you get A13 vendor from? was it just la.um blobs or from fp themselves? | 16:45 |
| mal | those are probably just the blobs from fairphone | 16:45 |
| T42 | <TheVancedGamer> I hear you got SFOS on android 13 | 16:46 |
| mal | from themuppets repo | 16:46 |
| mal | yes | 16:46 |
| T42 | <TheVancedGamer> from what I know, 13 wasn't released by fp themselves | 16:46 |
| T42 | <TheVancedGamer> yet anyway | 16:46 |
| mal | yeah, that is based on lineage | 16:46 |
| T42 | <TheVancedGamer> was it released as an image or I need to build it myself? | 16:46 |
| mal | also fairphone hasn't released aosp sources for fp4, at least not yet | 16:47 |
| mal | what was released as image? | 16:47 |
| T42 | <TheVancedGamer> the sfos build | 16:48 |
| mal | I haven't released any builds for fp4 yet | 16:48 |
| mal | because there are couple of blocker bugs | 16:49 |
| mal | most critical is call audio, second is mobile data and third a bit less crtitical is video recording, everything else works | 16:49 |
| T42 | <k1gen> mal: no chmod*vibrator in init*rc files in /vendor, /system and droid-hybris/system | 16:57 |
| T42 | <elros34> so search in udev rules or in other places, you just did the same search for other file. You can just add similar to fp4 file and hope it will work | 16:59 |
| mal | only two ways those permissions can happen, either .rc or .sh files in android init or udev | 17:00 |
| T42 | <k1gen> grep -r 'chmod*vibrator' /lib/udev/rules.d also returns nothing | 17:16 |
| T42 | <k1gen> as well as grep 2>/dev/null -R 'chmod*vibrator' /vendor and grep 2>/dev/null -R 'chmod*vibrator' /system, weird. where are the permissions then? | 17:23 |
| T42 | <elros34> in /system_root? or in some .sh file | 17:26 |
| mal | maybe try grepping just for vibrator | 17:50 |
| T42 | <k1gen> already, no mentions in /system, one file in /vendor | 17:54 |
| T42 | <k1gen> grep /system_root still running | 17:54 |
| T42 | <elros34> but why you care so much, add vibrator.rc overwrite and see if it work | 17:55 |
| T42 | <Arunkumarneelam> Can anyone send me guidence notes? | 17:59 |
| T42 | <k1gen> mal: what to replace FP4_HYBRIS_DISABLED with? | 18:00 |
| T42 | <k1gen> FLAME_HYBRIS_DISABLED? | 18:00 |
| T42 | <k1gen> I have this service: service vendor.vibrator.cs40l25 /vendor/bin/hw/android.hardware.vibrator-service.cs40l25 | 18:04 |
| T42 | <elros34> can be even FLAME_HYBRIS_ENABLED | 18:05 |
| T42 | <k1gen> that's what I have now: https://vomitb.in/raw/vihLNS526k | 18:06 |
| T42 | <k1gen> should I enable this somehow or just reboot? | 18:06 |
| T42 | <elros34> reboot, couldn't you use just /sys/class/leds/vibrator path instead i2c address path? | 18:07 |
| T42 | <TheVancedGamer> @elros34 https://hastebin.com/share/ejezinabud.vbnet | 18:12 |
| T42 | <TheVancedGamer> mw, gg and configs, everything built fine | 18:12 |
| T42 | <elros34> nothing in this paste but if its fine then fine:) | 18:13 |
| T42 | <TheVancedGamer> https://hastebin.com/share/unupeceyoq.vbnet | 18:14 |
| T42 | <TheVancedGamer> entire log | 18:14 |
| T42 | <elros34> search for ofono-configs-binder and follow instruction. Don't you guys read it? | 18:15 |
| T42 | <TheVancedGamer> oh | 18:15 |
| T42 | <TheVancedGamer> I didn't actually see it, sorry | 18:15 |
| T42 | <TheVancedGamer> ``` | 18:22 |
| T42 | <TheVancedGamer> File /etc/bluez5/bluetooth/main.conf | 18:22 |
| T42 | <TheVancedGamer> from install of | 18:22 |
| T42 | <TheVancedGamer> droid-config-angelica-bluez5-1-202303271821.armv7hl (dir:/home/thevancedgamer/hadk/droid-local-repo/angelica) | 18:22 |
| T42 | <TheVancedGamer> conflicts with file | 18:22 |
| T42 | <TheVancedGamer> /etc/bluetooth/main.conf | 18:22 |
| T42 | <TheVancedGamer> from package | 18:22 |
| T42 | <TheVancedGamer> bluez5-configs-mer-5.65+git1-1.9.1.jolla.armv7hl (@System)``` | 18:23 |
| T42 | <TheVancedGamer> | 18:23 |
| T42 | <TheVancedGamer> I did add | 18:23 |
| T42 | <TheVancedGamer> Provides: bluez5-configs | 18:23 |
| T42 | <TheVancedGamer> Obsoletes: bluez5-configs-mer | 18:23 |
| T42 | <TheVancedGamer> nvm, running the sb2 command seems to have fixed it | 18:25 |
| T42 | <k1gen> @elros34 vibro doesn't work after reboot, permissions are still being rewritten | 18:55 |
| T42 | <k1gen> overridden by what though... | 18:55 |
| T42 | <elros34> do you still see disabled service? Is file even parsed? | 18:58 |
| mal | @k1gen do you see anything logs about vibrator? | 18:59 |
| T42 | <k1gen> https://vomitb.in/raw/j3XNlhPbPg | 18:59 |
| T42 | <elros34> wrong permission | 19:00 |
| T42 | <k1gen> what it's supposed to be? | 19:00 |
| T42 | <k1gen> I have rw-rw-r-- | 19:01 |
| T42 | <elros34> I do not know you have plenty of them in device, check one | 19:01 |
| T42 | <k1gen> got it | 19:01 |
| T42 | <k1gen> it's constantly vibrating from lockscreen on highest power, had to turn off the device | 19:05 |
| T42 | <k1gen> can I fix it from twrp chroot? | 19:06 |
| mal | hmm, that is odd | 19:08 |
| mal | if you have persistent logs then you could check is logs tell something what is happening | 19:08 |
| T42 | <k1gen> same journalctl command? | 19:09 |
| T42 | <k1gen> I'm in chroot | 19:10 |
| T42 | <k1gen> https://vomitb.in/raw/jvf5AaGPiv | 19:11 |
| T42 | <k1gen> seems like everything is fine | 19:16 |
| T42 | <k1gen> mal: here's full journalctl: https://vomitb.in/raw/67ny7QDxrr | 19:26 |
| T42 | <k1gen> @elros34 any ideas why vibrator works on max setting? | 19:33 |
| T42 | <elros34> not really, mask ngfd and start it manually, then watch sysfs entries and see if it vibration will start/stop, what values are there. I wonder if ngfd can works correctly without audio. Try disabling quick access to camera in lockscreen | 19:36 |
| T42 | <elros34> and try to make that airbrush working for now by just adding some more than needed permissions | 19:37 |
| T42 | <elros34> disable vendor.time_daemon in disabled_services.rc | 19:38 |
| T42 | <k1gen> what's the exact line? service vendor.time_daemon /vendor/bin/hw/time_daemon.service_HYBRIS_DISABLED? | 19:44 |
| T42 | <elros34> probably, you can confirm in time_daemon.rc file you have | 19:45 |
| T42 | <elros34> I wonder you should enable vold if servicemanager waits for it | 19:46 |
| T42 | <k1gen> sorry, where is this time_daemon.rc supposed to be located? | 19:46 |
| T42 | <elros34> path is in your logs | 19:47 |
| T42 | <k1gen> I see, thanks | 19:47 |
| T42 | <k1gen> what else can I do before rebooting from chroot? | 19:48 |
| T42 | <k1gen> should I remove the service vold /system/bin/vold_HYBRIS_DISABLED line? | 19:49 |
| T42 | <elros34> you can try but mal will probably know better | 19:50 |
| T42 | <elros34> hope it will not encrypt enything:) | 19:55 |
| T42 | <k1gen> :| | 19:56 |
| T42 | <XAP2P> xD (re @elros34: hope it will not enc...) | 19:56 |
| T42 | <k1gen> mal: what do you think, should I enable vold? if /data gets corrupted or anything, this will set me back for a week, if not more. and it will be demoralizing as hell | 19:59 |
| T42 | <elros34> I wonder if this waiting spam is from some init*rc file with "wait.*vold" line | 20:04 |
| T42 | <elros34> but there is only wait_for_prop and wait <path> so probably no. If you scare so much then just ignore it for now | 20:05 |
| T42 | <k1gen> reboot then. with masked ngfd vibro shouldn't work, right? | 20:06 |
| T42 | <elros34> it shouldn't unless issue is in some other service | 20:07 |
| T42 | <k1gen> alright, I disabled time_daemon, reboot and then what? vibro doesn't work | 20:14 |
| T42 | <elros34> and how time daemon is related to vibrator? | 20:16 |
| T42 | <elros34> like I said start ngfd manually | 20:17 |
| T42 | <k1gen> "adding some more than needed permissions" like chmod 777 /dev/ab_sm? | 20:17 |
| T42 | <elros34> sure | 20:17 |
| T42 | <k1gen> @elros34 after I launch ngfd manually, phone starts to vibrate, and even after I send SIGKILL to ngfd doesn't stop until shutdown: https://vomitb.in/raw/QlZBSwxATZ | 20:28 |
| T42 | <elros34> you can probably write to sysfs to stop it. ngfd also have support for -v argument | 20:30 |
| T42 | <k1gen> by the way, if I chmod vibrator files manually, vibration is low-power and works fine, but this... I'm afraid it might damage the phone if turned on for a while | 20:30 |
| T42 | <elros34> chmod change anything? | 20:32 |
| T42 | <k1gen> not anymore. if I chmod manually and then start ngfd, device vibrates with max power again | 20:35 |
| T42 | <k1gen> could you please provide a command to stop it so I could paste it fast into ssh? | 20:35 |
| mal | @k1gen vold should not be enabled afaik | 20:37 |
| T42 | <k1gen> then it's good I haven't | 20:37 |
| mal | what are you trying to fix now? | 20:38 |
| T42 | <k1gen> strong vibration on ngfd launch | 20:39 |
| mal | can you see what it writes to sysfs? | 20:39 |
| T42 | <k1gen> I'll go to sleep in 15 minutes though | 20:39 |
| T42 | <k1gen> every time that vibration starts I immediately turn off the device, because I don't want to damage it | 20:40 |
| mal | and does vibrator work normally if you enable it manually by writing to the duration and activate paths | 20:40 |
| T42 | <k1gen> not anymore. if I chmod manually and then start ngfd, device vibrates with max power again | 20:40 |
| mal | I mean write directly to those paths from terminal | 20:40 |
| T42 | <k1gen> ah, like that, no. what should I echo? | 20:41 |
| mal | afaik duration is in milliseconds and so some reasonable value and then start by writing 1 to activate and stop by writing 0 to activate | 20:42 |
| T42 | <k1gen> I wrote 10 to duration, 1 to activate, it started vibrating like crazy, then I wrote 0 to activate - and it didn't stop | 20:46 |
| mal | very odd | 20:46 |
| mal | do you see anything in dmesg when that happens | 20:49 |
| T42 | <k1gen> mal, I have to go rest, we can continue tomorrow. we're done with sound until you or elros figure out a fix, right? it's in higher priority to me than vibration and anything else | 20:50 |
| T42 | <k1gen> I'll answer that tomorrow in the morning, if you don't mind | 20:50 |
| mal | can you show what files there are in /sys/devices/platform/soc/c94000.i2c/i2c-4/4-0043/leds/vibrator/ | 20:52 |
| T42 | <k1gen> mal: it's from today, there's a link in chat history: https://vomitb.in/raw/aMB7BLFPMF | 20:54 |
| mal | @k1gen one thing to try, use that brightness path instead of activate, write 1 to start and 0 to stop | 21:00 |
| T42 | <k1gen> sure, I'll do that in the morning. did you ignore my question or just didn't see it? :) | 21:01 |
| mal | so much text that I probably have missed something | 21:02 |
| T42 | <k1gen> 23:50 one | 21:02 |
| mal | ok | 21:02 |
| T42 | <elros34> mal: I think this is the code so yes probably brightness file : https://github.com/LineageOS/android_kernel_google_coral/blob/lineage-18.1/drivers/misc/cs40l2x.c | 21:03 |
| mal | @elros34 yes, that is where I found it, quite odd why that is different | 21:03 |
| T42 | <k1gen> mal: let me repeat the question: I can forget about sound until you or elros or someone smart enough figure out a fix for that weird segfault, right? | 21:11 |
| mal | well we need your help since you have the device, it probably require doing some more testing in certain ways | 21:12 |
| T42 | <k1gen> well as long as you guys are helping I hope we'll get there | 21:14 |
| T42 | <elros34> mal canjb2q plugin can work on hybris 18? If yes I would try that | 21:14 |
| mal | I have been trying to get call audio working on my device for longer than you have been doing that, it me more time to get even basic audio working but this needed some special things | 21:15 |
| mal | @elros34 not sure really, it could be tested | 21:17 |
| T42 | <k1gen> just tell me what to do™️ | 21:20 |
| T42 | <k1gen> goodnight | 21:20 |
| T42 | <b100dian> So if you "chmod the files manually" the vibration works fine? Otherwise, it just doesn't stop and it's maximum ? | 23:10 |
| T42 | <b100dian> Btw, I have vold on my previous port. But maybe I didn't search enough who was waiting for it. Since you have your screen on just ignore it, I didn't (and no, that doesn't encrypt by itself). But of course you should make your changes on the host sooner rather than later, or you lose context and forget some changes. | 23:10 |
| mal | @b100dian I understood that it never stops, we suspect that the activate is not working as it should | 23:21 |
| mal | and usually it should be possible to see from logcat what process is waiting for vold or any other service | 23:22 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!