Monday, 2018-05-28

UmeaboyWhy doesn't the HADK mention anything about building for aarch64 based devices? The mer chroot tarball should be different then.00:32
MeowDudeThere is a build for maguro???? I replied to the dev on xda but *QUE?* The only reason I don't grab a Galaxy nexus is because A: battery and B: Sailfish OS support is garbage03:06
*** Nokius_ is now known as Nokius05:24
*** mog- is now known as mog10:01
steeeevehey, how to implement audio routing from . It has a function that sets a parameter to route. How do i implement it for my device? Do i have to add a config with my parameter or what?10:35
r0kk3rzsteeeeve: it would be better to look at the current droid module code10:40
r0kk3rzrather than some ancient slides10:40
r0kk3rzwhat parameter are you trying to set? and on what?10:42
steeeeve . in ril class it's ez, in pulseaudio i dont know how to set it10:44
r0kk3rzat a guess you'd want to do a 'droid.parameter.realcall=true' in xpolicy.conf somewhere10:51
r0kk3rzjusa probably knows how to feed that part10:53
steeeeveoh yes10:56
jusar0kk3rz: steeeeve: I'll check that out a bit later10:58
steeeeveim at school, i have to go now :D11:00
r0kk3rzthanks jusa11:00
r0kk3rzah i see it needs to be toggled when theres an actual call11:07
*** lynxis_ is now known as lynxis11:27
steeever0kk3rz: yes, it's toggled12:09
jusasteeeve: this probably needs some changes to implementation, but as of now you can test with following:12:23
jusa(moment :))12:25
jusamake a call, then12:27
jusapacmd update-sink-proplist sink.primary droid.parameter.realcall=on12:27
jusapactl set-sink-port sink.primary output-parking12:28
jusapactl set-sink-port sink.primary output-earpiece12:28
steeeveDo i have to change output-parking and output-earpiece with mines?12:28
jusaoutput-parking should stay that, but output-earpiece can be output-speaker or whatever12:29
steeeveThe one in mixer_paths right?12:29
jusaand if once doesn't cut it, repeate output-parking & output-foo , as the java code for some reason calls the routing twice12:29
steeeveHm, let's try!12:30
jusaoutput-parking is there just because of how pulseaudio works, pulseaudio port change to already active port is no-op, so the parking port is there just to avoid changing upstream pulseaudio code12:30
jusayou can see the pulseaudio port names with pactl list sinks12:33
steeeveok, got it12:34
steeevejusa: i dont get audio on calls12:36
steeevei tried running the 3 commands 2 times12:36
jusaok.. you already have audioflingerglue running ?12:36
steeevecommand plz?12:37
steeeveps -aux|grep audio shows  nothing about that12:37
r0kk3rz[m]look for miniaf12:39
steeevethen yes, it's running:miniafservice12:39
r0kk3rz[m]And its something you need to build yourself, details are in the hadk12:40
steeevei built them12:40
steeeveit must be running12:40
jusasteeeve: could you take logcat output of making a call and calling those command line calls12:42
jusasteeeve: also output of pactl list12:42
steeevebtw i have an amplifier (tfa9895), too12:44
jusahm it actually might be that the realcall needs to be set to the hw module instead of output stream.. you are using some cm as base right? could you find the audio adaptation repo12:47
steeeveim using stock hal12:47
steeevebecause of amplifier12:48
steeevethe ril class is what i need to make audio on calls working12:48
jusaplease elaborate, what amplifier, how is it connected etc? :)12:50
steeevetfa9895, which is connected inside the device12:50
steeevethats all i know12:50
ghosalmartinit seems my kernel is just all bad :P12:51
steeeveand there's a lib called which has some stuff needed by HAL12:51
jusahrm.. doesn't tell much.. but we could try calling hw module set_parameters first, a moment, I'll provide a patch12:54
steeeveWoho :P12:56
jusaapply that, and after installed repeat the sequence I mentioned above12:58
jusaargh, sorry,12:58
steeeveOk, let's see12:58
jusait's a bit wrong, moment..12:58
jusasteeeve: this one:
steeeve(By the way that parameter is isnide audio hal (cbecked with strings), so, that might be the solution)12:59
jusaalso experiment with output-speaker & output-earpiece when switching back and forth, I have no idea if that has any signifigance but it just might13:01
steeeveIm compiling13:05
ghosalmartinif a kernel panics so fast that ramoops arent even there is there a few obvious things to check?13:19
steeevejusa: working woowowowow13:20
jusasteeeve: cool :)13:20
steeeveghosalmarting: ramoops gets all panic13:20
jusais the realcall something that is needed always, or for some specific use case?13:20
steeeveno, only during calls13:20
jusayep, but for calls always with this specific samsung device ?13:21
jusaor just specific calls?13:21
jusaif it is always it's a bit simpler to implement13:21
steeeveall calls13:21
jusaI'll figure out some way to enable that from configuration and I'll let you know when a PR is ready (might take a couple of days)13:22
steeeveyeah, ping me, as long as irccloud works i should be online :D13:23
steeeveand thanks :)13:23
ghosalmartinsteeeve, doesnt seem to :(13:24
steeeveYay, but i have 1 week free :P13:27
r0kk3rzwoo it works :)14:15
r0kk3rzsteeeve: do you know if other phones do things like this?14:16
steeever0kk3rz: all samsung msm8916 devices that use stock audio hal14:26
r0kk3rzyeah right14:27
ghosalmartinhmm maybe a rebase vs the latest upstream helped my issue, now I can atleast build a image that crashes due to sailfishos changes :P16:05
ghosalmartinnot sure what broke it though it was broken when the commit here was 7th of may16:05
malghosalmartin: you used my latest code in github, I rebased those last week16:12
ghosalmartinmal: nah not yet, still tryna get resemblance of a booting kernel16:13
ghosalmartinso far the minute I apply any of hybris-boots configs it just ends in a boot loops with no logs16:14
r0kk3rzghosalmartin: which are you enabling?16:17
ghosalmartinr0kk3rz, these
ghosalmartinoh and disabling CONFIG_DUMMY16:17
ghosalmartinam about to either prove or disprove that statement with this latest build16:18
ghosalmartinif this build works then its something in those configs thats causing the problem / a dirty build issue16:18
r0kk3rzyeah thats a lot of trial and error16:19
r0kk3rzand even if you found the breaking flag, it wouldnt tell you why16:19
ghosalmartinfor now it would be better than nothing :P16:20
ghosalmartinplus it wasnt all a wasted effort, I managed to learn some things16:20
r0kk3rzi wonder if unwinding the changes since 14.1 would be a better idea16:21
ghosalmartinyep tis defo the hybris changes am making16:21
ghosalmartinand now I have decent ramoops from hybris-boot16:23
ghosalmartintake it this is at a script  init[1]: unhandled level 2 translation fault (11) at 0x000002e0, esr 0x9200004616:24
ghosalmartinat the init-script16:24
ghosalmartinr0kk3rz, does the init script panic if configs aren't fulfilled?16:25
r0kk3rznot sure16:26
r0kk3rzgot the logs?16:26
ghosalmartinactually lemme check16:27
ghosalmartinwell atleast its a positive breakthrough today :P16:29
r0kk3rzenable proc fs16:33
r0kk3rzand devtmpfs16:33
ghosalmartinam about to16:33
ghosalmartini disabled CONFIG_DUMMY with no issues16:33
r0kk3rzwell mounting those things is the second thing it does16:34
r0kk3rzso maybe its a bit unhappy without them16:34
ghosalmartinr0kk3rz, will those memory addresses related to anything like a line number in the file?16:35
r0kk3rzs/it/the init script16:35
r0kk3rznah i dont think so16:35
ghosalmartinah okays16:36
ghosalmartinI know it does for kernel traces which is handy i guess16:36
r0kk3rzive not really looked into it tbh16:38
ghosalmartinr0kk3rz, still no dice with those two, but the kernel isnt breaking like before so I guess Ill keep enabling flags. this could be an issue with busybox16:44
r0kk3rz[m]Yeah tbh i think most of those listed are needed16:46
ghosalmartinwell took a while but it was CONFIG_VT that was breaking the device17:18
ghosalmartinmal: how important is CONFIG_VT?17:19
malghosalmartin: no idea17:21
ghosalmartinaccording to docs you need a keyboard and display for it so not very important I bet :P17:22
ghosalmartinYou need at least one virtual terminal device in order to make use of your keyboard and monitor. Therefore, only people configuring an embedded system would want to say N here in order to save some memory; the only way to log into such a system is then via a serial or network connection.17:22
ghosalmartinalthough network connection hmm missed that bit17:22
r0kk3rz[m]Really? I wonder why...17:28
ghosalmartinr0kk3rz[m], time to look through those 200ish commits and see if any changes have been made to vt.h17:28
*** minimec_ is now known as minimec17:32
steeeveTo make bluetooth working what i have to do except for adding mer-vheck-config configs?18:39
r0kk3rzthat depends greatly on what bt chip you have18:43
r0kk3rzthats not a bt chip18:45
r0kk3rzsteeeve: where are your repos?18:46*18:46
r0kk3rzyou didnt fork the device repos there?18:47
steeeveI did18:48
steeeveYou mean android repos like kernel and device tree?18:48
steeeveOr sailfish configs?18:48
r0kk3rzandroid repos18:48
steeeveToo many accounts :P18:49
ghosalmartinis it a ROME chipset?18:52
malsteeeve: also in the future please use full urls, much easier to open the links :)18:52
r0kk3rzi looks like it might be a ROME chip18:52
ghosalmartinaww man :( i spent many a long night on ROME chipsets without much luck18:53
r0kk3rzits hard to say, looks like it automagically configs it based upon the chipset...18:54
ghosalmartinahh bullhead never had that script, but the lovely chiron does18:55
ghosalmartinso hopefully that'll be an easier time with its bluetooth18:56
malsteeeve: run "getprop qcom.bluetooth.soc" on your device18:56
steeevesmd :D18:57
steeeveLike hcismd woh18:57
malsteeeve: have you tried the usual bt script already?18:57
r0kk3rzyeah the stuff in the faq should be all you need18:58
steeeveThis: ?18:58
malsteeeve: this has the generic instructions
malsteeeve: yes, that is the script, it also needs the service18:59
steeeveOk, i have to add it :/18:59
malsteeeve: try on device first, just copy the script and the service and add the symlink to start the service19:00
steeeveAh sure19:00
steeeveAnd thanks for the info ^^19:00
steeeveAnd got the bluetooth wowow19:10
malsteeeve: that's why you read the faq :)19:10
steeeveI read it eh19:10
steeeveI built hcismd stuff before with the backports ;)19:10
steeeveI just needed the service19:11
malok, but forgot the actual bluetooth fix19:11
steeeveJust a lil service meh19:11
malyes, but a very important one19:11
steeeveSmall but big yeees19:12
malwhat issues do you still have with your device?19:12
steeeveNone when i'll get audio routing config19:13
steeeveAfter that ican publish the zip D:19:13
maldoes you device have fm radio?19:14
steeeveAh yes19:14
steeeveI havent tested it, yet19:14
malsteeeve: does headphone connector work?19:16
steeeveIt needed some stuff like in audio calls, but at least it's in hardware/qcom/fmradio and no one cares of fm radio, so, if it doesnt work, yes, i can try to fix it but it has the lowest priority ever19:17
steeeveYes headphones work19:17
malsteeeve: usually fm radio works if it uses that one qcom driver19:18
steeeveBut i will do a mass csd test soon, just to see general status19:18
steeevemal: it's something related to audio, but i can fix it myself i think19:19
maljust try the normal middleware first, some devices need a bit patching of the iris fm radio driver19:19
r0kk3rzfully working samsung port, thats rare19:26
steeeveYeah, just thanks to you all that helped me. Without you just few stuff was working :D19:32
steeeveNow i will try to apply fm patch to see if fm audio works (fm itself works), but if it doesnt, amen19:34
ThaodanIs there a way to strip out prebuild binaries from syncing when downloading the android sources?20:14
malThaodan: which ones do you mean in this case?20:17
Thaodanfolders like prebuits which contains certain gb of software is not used20:21
Thaodansomething like this mal20:37
malsome of those can be removed, just edit the manifest20:45
Mister_Magistermal: how can i enable flash on video?21:01
malMister_Magister: it should work, there should be an option for that in video recording mode21:03
Mister_Magisterthere is no21:03
Mister_Magisteri think i need to add it to gst-droid conf21:03
Mister_Magisteror jolla-camera conf21:04
maljolla-camera-hw.txt, change the flashValues in primary camera video mode to flashValues=[2, 32]21:05
malit seems droid-camres doesn't add the flosh torch for video mode21:05
Mister_Magisterthats what i thought i just changed it to 1, 2, 4 like in camera, thanks21:05
Mister_Magistermal: it doesnt add front flash too21:06
Mister_Magisterit only can into back flash on photo21:06
malit should be like I gave, not like front picture camera21:06
Mister_Magistermal: yeah i know thanks21:06
malI need to fix droid-camres21:07
Mister_Magistertotally :P21:07
Mister_Magisterdont forget front flash :P21:07
Mister_Magisterdroid-camres doesnt detect front flash too21:08
maldo some devices have front flash also?21:08
malit doesn't detect flash values at all, those are hardcoded21:08
Mister_Magisteryeah moto x force that i have (i just told you it doesnt detect front flash 2 times above xD)21:08
Mister_Magistermal: ahh k21:08
malMister_Magister: I just thought you meant back, I never heard of device with front flash, not sure who needs such a thing21:09
Mister_Magistermal: S E L F I E S21:10
malwell, I don't take selfies :)21:11
Mister_Magisterme too :D21:19
Mister_Magistermal: just adding 1, 2, 4 to front camera flash fixed it so and thanks for video values21:20
malMister_Magister: yes, thought that should fix it21:20
r0kk3rzmal: these days the front camera is much more important than the back :P21:48
malr0kk3rz: for some people21:48
Mister_Magisteri dont use camera21:49
Mister_Magisterjust like for showing something to people and i dont need good quality for that21:51

Generated by 2.17.1 by Marius Gedminas - find it at!