Sunday, 2021-12-12

*** Daanct12 is now known as Danct1206:39
T42<eriki73> Not directly related to porting sfos but I have a question about pulseaudio-moules-droid: Should wired headset always show as plugged in?12:50
*** Daanct12 is now known as Danct1213:10
malit should not13:34
malchecking the output of "evdev_trace -t" can be use to see if the events from the headset are happening correctly, in theory also ohmd might need configuring in some cases13:34
T42<eriki73> mal: I was not on sailfish -> no ohm -> required patching modules. I just didn't realize that ohm does that on sfos...14:33
T42<eriki73> That is why I said is not directly related to porting sfos14:34
T42<eriki73> I just packaged the modules for Manjaro not realizing that we don't use ohm14:36
T42<eriki73> Well, it is fine now...14:37
T42<AntonlX> After mal patches there is binder and mknod errors:
T42<AntonlX> chmod 666 /dev/binder No such file or directory17:41
T42<AntonlX> but /dev/binder is there17:42
T42<foxclore> A question: verify_config checks not only kernel config mentioned in official hadk guide, but also obj/DTB_OBJ/.config18:08
T42<foxclore> Any ideas which defconfig generates dtb config? because my kernel config is fine, but the dtb one fails18:08
mal@AntonlX but is /dev/binder a symlink?18:10
T42<foxclore> addition: not mer-verify-config, but when building droid hal it checks dtb config18:10
malwhat do you mean by dtb one fails?18:29
T42<elros34> here you can find why it check DTB_OBJ/.config: It shouldn't be hard to fix it and check only KERNL_OBJ/.config or kernel/.config instead /*/.config18:38
T42<foxclore> mal: mer_verify_kernel_config ./out/target/product/apollon/obj/DTB_OBJ/.config18:42
T42<foxclore> It gives bunch of errors, yet I can't find defconfig which influences DTB_OBJECT config18:42
malis that really where the .config is?18:44
T42<foxclore> yep, there is one18:45
T42<elros34> so why don't you just edit so it uses correct KERNEL_OBJ/.config?18:47
T42<foxclore> but it does so, it uses both18:48
T42<foxclore> here is the line:18:48
T42<AntonlX> yes (re @SailfishFreenodeIRCBridgeBot: <mal>@AntonlX but is...)18:48
T42<elros34> but it should use only one18:48
T42<foxclore> Verifying kernel config18:49
T42<foxclore> + hybris/mer-kernel-check/mer_verify_kernel_config ./out/target/product/apollon/obj/DTB_OBJ/.config ./out/target/product/apollon/obj/KERNEL_OBJ/.config18:49
T42<foxclore> Here is full log:18:49
T42<elros34> I do not need it18:49
T42<AntonlX> binder -> /dev/binderfs/binder18:50
T42<elros34> just edit that one line so it use only 1 .config from KERNEL_OBJ/.config18:50
T42<elros34> If you read comment I linked then you will understand why both files are checked: just to make it work both with aosp and lineage18:51
T42<foxclore> oh okay, thank you18:51
T42<elros34> probably for < hybris-18  there wasn't such an issue18:52
T42<foxclore> but I am building for hybris-17.118:53
T42<elros34> so maybe exception, doesn't really matter but make sure you use correct barnch for kernel and you removed out/ directory before switching from 18 to 17 base18:57
T42<foxclore> yep, thank you18:57
T42<foxclore> Had to remove everything and re-sync from scratch18:57
mal@AntonlX so what are the permissions fo the target of the symlink?19:14
T42<AntonlX> chmod 666 /dev/binderfs/binder: No such file or directory19:16
malso nothing created the dynamic binder devices19:19
malsomething like this in defconfig could help: CONFIG_ANDROID_BINDER_DEVICES="binder,vndbinder,hwbinder"19:21
T42<AntonlX> CONFIG_ANDROID_BINDER_DEVICES="binder,hwbinder,vndbinder,anbox-binder,anbox-hwbinder,anbox-vndbinder" (re @SailfishFreenodeIRCBridgeBot: <mal>something like ...)19:23
T42<foxclore> Hmm, interesting error:19:31
T42<foxclore> rpm/dhd/helpers/ --gg19:31
T42<foxclore> Please build droidmedia as per HADK instructions19:31
T42<foxclore> !! Failed to pack_source_droidmedia-localbuild.sh19:31
T42<foxclore> I did make droidmedia which finished successfully19:32
T42<foxclore> but still asks me to build as per HADK19:32
T42<foxclore> Any ideas?19:32
maldo you see in out/ probably in out/target/product/*/system/lib*/libdroidmedia.so19:34
T42<foxclore> hmm, nop19:34
T42<foxclore> hmm, nope (edited)19:34
T42<foxclore> mal: its doesn't exist in out directory (19:35
malso what went wrong in droidmedia build19:36
T42<foxclore> well i dont know, is there a way to check?19:37
T42<foxclore> "make droidmedia" says "#### build completed successfully (7 seconds) ####19:37
T42<foxclore> "19:37
malafter that the should appear in out/19:39
T42<elros34> change somthing in droidmedia source file, save and run make again. Then it should print full path at the end of build19:40
T42<foxclore> where do I find droidmedia sources?19:40
T42<elros34> external/droidmedia19:41
T42<foxclore> hmm, I changed some stuff in hybris.c, but build log is:19:42
T42<foxclore> "ninja: no work to do"19:42
T42<foxclore> "build completed successfully (7 seconds)"19:42
malafaik hybris.c is not relevant for that19:44
maledit for example droidmedia.c19:45
T42<foxclore> same happened when changing external/droidmedia/droidmedia.cpp19:45
T42<foxclore> no impact19:45
T42<elros34> so are you sure that you have changed all paths and env when changing hybris base? In worst case make clean will do the job19:46
T42<foxclore> Well, I just deleted $ANDROID_ROOT and then started from scratch19:50
T42<foxclore> So make clean it is then, we'll see19:50
malwhy would you remove the sources, just removing out/ will make a completely clean rebuild using the existing sources20:07
T42<foxclore> I had to use completely different kernel and device tree20:13
malstill removing everything was not needed, assuming you used same android base20:16
T42<foxclore> Mal: well kind of. Still feels safer to start from scratch20:17
T42<elros34> mal: I think removing $ANDROID_ROOT was related to changing android base20:20

Generated by 2.17.1 by Marius Gedminas - find it at!