Sunday, 2019-06-30

T42<SALPHA8506> s-8506@s8506-X556UR:~$ sb2 -t $VENDOR-$DEVICE-$PORT_ARCH gcc main.c -o test01:04
T42<SALPHA8506> sb2: Error: Invalid target specified, aborting.01:04
T42<SALPHA8506> not getting any connection (not showing in dmesg) after install sailfish with lineage 15.106:12
T42<SALPHA8506> how to check log ?06:12
T42<SALPHA8506> only stuck at device logo06:13
T42<LittleBoy8506> @LittleBoy8506 [not getting any connection (not showing in dme …], not even telnet work06:20
deathmistSALPHA8506: you should have an init.log in /data/.stowaways/sailfishos/init.log accessible from recovery06:52
deathmistfor me it was the RNDIS kernel options, could you show defconfig?06:53
deathmistdoes the device reboot after 65 sec if you just let it sit there? also ch06:55
deathmist*check dmesg on host while it's booting and sudo lsusb -v06:55
deathmist| grep iSerial06:56
T42<LittleBoy8506> @LittleBoy8506 [], defconfig06:57
T42<LittleBoy8506> device not rebooting06:57
T42<LittleBoy8506> it stuck only06:57
deathmistnothing on host dmesg and lsusb -v | grep iSerial when it's connected? what about the init.log?06:59
T42<LittleBoy8506> it's connectd now07:00
T42<LittleBoy8506> but stuck07:00
deathmistthat doesn't answer much, enable CONFIG_AUDIT and CONFIG_FW_LOADER_USER_HELPER in your defconfig and go to device/zuk/msm8XXX_common/ and find BOARD_KERNEL_CMDLINE then append "audit=0 selinux=0" at the end => make -j`nproc` hybris-boot07:03
deathmistand flash the new boot image to check if it changed anything07:04
deathmistactually make hybris-recovery instead07:04
deathmistI also see CONFIG_DM_VERITY is enabled, I don't remember if that caused issues (should probably unset it)07:05
deathmistCONFIG_USB_CONFIGFS_RNDIS should be enabled too07:07
deathmistthat probably actually does nothing because if your device doesn't reboot then RNDIS should be up, just the matter of getting the options right. remove CONFIG_CGROUP_MEM_RES_CTLR* options too as you have 4.4 kernel (right?) and change CONFIG_NETPRIO_CGROUP => CONFIG_CGROUP_NET_PRIO (it got changed in newer kernel versions)07:18
T42<LittleBoy8506> 3.18 kernel not 4.407:27
deathmistok then no need for those in last message07:27
T42<LittleBoy8506> i think selinux should disable right ?07:28
T42<LittleBoy8506> be^07:28
deathmistsome devices need it BEFORE botting, so disabling it via cmdline is safer if you don't want to check yourself :p07:29
deathmistalso CONFIG_CGROUP_MEM_RES_CTLRy,!# systemd (optional):, ignore if kernel version >= 3.10 so CONFIG_CGROUP_MEM_RES_CTLR* should be removed anyways07:29
T42<LittleBoy8506> what about this CONFIG_NETPRIO_CGROUP07:30
deathmistit's not important, if it's set in .config then no need to change07:31
T42<LittleBoy8506> ok07:31
Mister_Magistersomebody any idea how to debug kernel without any logs other than destroying phone to get serial?09:06
deathmistI only remember last_kmsg / console_ramoops but if you cannot boot to recovery without fully restarting or the file isn't there I have no clue, which device and base are you working on? you sure about selinux options and kernel cmdline since I had a bunch of issues with those before and also forced encryption of /data09:09
Mister_Magisterdeathmist: little boy little you know selinux has nothing to do with booting kernel09:23
Mister_Magisterand even with encrypted /data it should boot09:23
deathmistsorry, I have way too little experience then to help at all09:23
r0kk3rzMister_Magister: what device?11:33
Mister_Magisterzenfone 5z11:34
r0kk3rzboots with stock config?11:36
Mister_Magisterstock doesnt have rndis11:37
r0kk3rzyoud need to screw with the init script11:38
Mister_Magisteri even cleared it out it desnt matter11:44
r0kk3rzcleared what out?11:45
Mister_Magisterinit script11:45
r0kk3rzyeah you need it to signal something obviously11:46
r0kk3rzbut if stock no boot, and no logs, then good luck!11:47
Mister_Magisteryeah like i tried putting installing busybox at the beginning and the running sleep 30 but that did nothing11:50
Mister_Magisteractual content of the init script doesn't matter11:50
T42<DSstill01> Guys im finally tryin to fix fm radio. Channel scanning works, however audio doesnt get routed in fm path. I dont have input-fm_rx but only output-fm and input-fm_tuner. I tried to enable the second one but it says 'shared memfd open() failed: function not implemented. How to get audio routed to fm, then oO?15:07
malhow did you try to enable the input?15:11
T42<DSstill01> Yes, it's in what i wrote15:12
T42<DSstill01> I get that error15:12
malso files like these but xvars with correct device
T42<DSstill01> I tried that15:15
T42<DSstill01> With input-fm_tuner in xvars15:15
maldid you check what pactl says about routing15:16
T42<DSstill01> Nop15:16
T42<DSstill01> What's the command? :p15:16
malnot sure check help of that15:17
T42<DSstill01> Ok15:17
T42<DSstill01> I dont see anything about thay15:18
T42<DSstill01> I can try with pulseaudio verbose output15:20
malso pactl and then list or something doesn't the inputs and outputs15:23
T42<DSstill01> What?15:26
T42<DSstill01> Also, on android i dont see anything with 'fm' grepping, that's why im sure it doesnt get routed15:27
malpactl can show what inputs and outputs are available, and which ones are used15:29
malon android?15:30
T42<DSstill01> in android it's play-fm if im not wrong15:30
T42<DSstill01> Yeah, play-fm and play-fm speaker-amp for speaker15:32
malI have no idea what you are talking about15:33
T42<DSstill01> This is the stock rom's mixer_paths which is the one i use15:33
T42<DSstill01> You can search for 'play-fm'15:34
maljust to be sure, the other "file" in the link I gave earlier is symlink15:35
T42<DSstill01> Ah lol15:35
T42<DSstill01> I opened it in github and though it was a text file15:36
malthe fmradio.conf, github shows symlinks in a bit confusing way15:37
T42<DSstill01> Yay15:37
T42<DSstill01> I'll reboot and see15:38
Mister_Magistermal: my sfos kernel is booting lineage15:41
T42<DSstill01> Now something works, but it's like a mic, i mean, it plays what the mic hears15:44
Mister_Magisterohh u r close15:44
T42<DSstill01> You think? :/15:45
Mister_Magisterlike there is single step from there15:45
T42<DSstill01> @Mister_Magister [mal: my sfos kernel is booting lineage], Lol15:45
Mister_MagisterDO NOT USE REPLY OPTION15:45
Mister_Magisternot on this chat15:45
T42<DSstill01> @Mister_Magister [like there is single step from there], I had the same issue in lineage 15.1 when playing with fmapp service15:46
T42<DSstill01> So?15:46
Mister_Magisterstop using it15:46
T42<DSstill01> What should i reply?15:46
T42<DSstill01> To what you said15:46
Mister_Magister just type my nick at the beginning15:47
T42<DSstill01> Then you wont know whats the message im replying to15:47
Mister_Magisterit comes from context15:48
Mister_Magisterthat's how irc works15:48
Mister_Magisterpeople lived like that for at least 20 years15:48
T42<DSstill01> Sometimes it doesnt15:48
Mister_Magisterthen you clarify what you are talking about15:48
Mister_Magisterreply option in telegram formates badly in irc15:49
malif irc nick i mentioned usually people get a highlight15:49
Mister_Magistermal: any idea for my problem? did i forgot some patch? init script in ramdisk doesn't seem to be running15:50
T42<DSstill01> Yes but if the user im replying to is like mister magister who spams 50 consecutive message, he wont know what's the one im replying to15:51
Mister_Magisteryes i will15:52
T42<DSstill01> IF you say so..15:54
T42<meierrom> @DSstill01: I do agree with you, but you simply can't use the reply feature here because of the many IRC users here.16:11
T42<DSstill01> Ok16:12
T42<meierrom> @DSstill01: also don't edit your posts. You will be just fine :)16:12
mal@DSstill01 any luck with fm radio now? either you have incorrect xvars file or you might need one hack16:16
maland in this case telegram group (or what those are called) is an extension of the original IRC channel so rules of that should be followed16:24
T42<DSstill01> Nope, i can only hear myself when fm is on16:45
T42<DSstill01> In xvars i tried both input-fm_tuner and output-fm with no luck16:46
T42<DSstill01> And after a quick search on github i saw that many users use input-fm_tuner in xvars, so it should be correct16:50
maloutput is of course not right, that is for fm transmission afaik17:05
T42<DSstill01> (Photo, 720x1280) In logcat i found this line17:13
T42<DSstill01> Don't know if it's referred to input-fm_tuner17:14
malI told you already to check pactl, using "pactl list | grep fm" you should see he input device17:18
T42<DSstill01> That was with input-fm_tuner17:19
T42<DSstill01> That command shows only input-fm_tuner and output-fm, and excluding the output one, input-fm_tuner should be the right one, but it gives me that error in ligcat17:20
maltry replacing /etc/pulse/ with
T42<DSstill01> Ok, in the meantime i found that it uses rec-main-mic device17:25
T42<DSstill01> Now i have to go to eat outside, i will continue tonight17:31
T42<OhYash> Is sailfish ported on one of those bezel-less phones?18:19
T42<OhYash> yet18:19
malwhy do you ask?18:29
piggzmal: which ofono plugin do i need tof new android?18:41
piggzfound it18:41
malpiggz: yes, that one, and check the config from nile repo18:52
piggzyup, my search facility wins again ;)18:53
piggzmal: so far i gave myself a busy spinner ... typical ofono issue19:02
T42<abhishek_0> Any news about hybris16?19:07
piggz@abhishek_0 im building using it atm on the fxtec .... patches ore not in the repos yet afaik19:09
piggzthink wip19:09
T42<abhishek_0> Okay so work is going on, future seems bright I think19:10
malI have just been a bit lazy19:10
T42<eugenio_g7> Also there are Sony-aosp9 trees in the mer-hybris org19:10
T42<OhYash> mal: asked cuz I was planning to buy a new phone and would prefer some with sailfish support19:11
piggzmal: ofono seems to connect to android ok, then timeouts....
piggzand PA / mce problem19:13
malpiggz: check what ofono version you have19:25
piggzmal: 1.21+git5119:29
malwhy so new? downgrade to usual version19:29
T42<DSstill01> im back. So, the current issue is thats even if i set input-fm_tuner in xarg conf, it routes audio to rec-mic-main which is probably why i hear myself instead of fm channels19:30
maldo what I said19:30
maltry the file I linked19:30
malthe .pa file19:30
T42<DSstill01> ok19:31
T42<DSstill01> im sceptical tho :D19:31
piggzmal: this is why19:31
piggzi+ | ofono                      | package    | 1.21+git51-1.4.1.jolla  | armv7hl | adaptation-community-common19:31
piggzv  | ofono                      | package    | 1.21+git44-1.19.1.jolla | armv7hl | jolla19:31
malpiggz: damn webhooks recreated the package there19:32
malpiggz: I removed it from devel common19:32
T42<DSstill01> mal: it still uses rec-main-mic:
malyou should be looking at pactl like I said19:38
malto see if it selects the correct one19:38
malthe route selection happens in sfos side19:39
T42<DSstill01> i dont know what's the option to see that. in —help there's nothing good19:39
piggzmal: same behaviour19:39
T42<DSstill01> However if i hear the noises around the phone it means it's really routing to mic19:40
malwhat, it gives nice list of commands to use19:42
mal"pactl help"19:42
malpiggz: exactly the same?19:43
T42<DSstill01> To manage stuff but not to debug what i need to debug19:43
malwell that tell what source pa is using19:44
T42<DSstill01> Which one? The first 4 dont and the last ones are to set them19:46
piggzmal: pretty much i think ... there is a meesage from statefs in there
piggzwhen not filtering for ofono19:50
malpiggz: so ofono fails after that or what? does it still show the timeouts?19:53
malyou should know not to show partial logs19:53
piggzmal: yes, timeouts eventually....19:53
piggzin logcat, i have some modem related messages19:53
piggz05-25 09:06:54.420  4095  4286 D PerMgrLib: QCRIL voting for modem19:55
piggz05-25 09:06:54.422  3809  3982 D PerMgrSrv: QCRIL voting for modem19:55
piggz05-25 09:06:54.422  3809  3982 D PerMgrSrv: modem client QCRIL is voter already19:56
T42<DSstill01> (Photo, 720x1280) Ok, so the problem starts here. I can try to hack it to use fm instead of mic19:58
mal@DSstill01 come on, yyou are focusing on completely wrong part19:59
piggzmal: maybe its a problme for another night, logcat is very busy atm with sensor errors every 25ms, and still havnt for wlan to work19:59
mal@DSstill01 I told the issue is most likely in sfos side, ignore the whole android side for now and focus on sfos side audio routing20:00
T42<DSstill01> Mal: no, that's a mixer_paths that has nothing to do with lineage msmxxxx ones. Maybe pa is confused with it :P20:00
malare you sure you copied the file I said correctly, i.e. replaced the existing file20:00
T42<DSstill01> I copied it20:00
malwhat? so you didn't replace file with different name with the new file?20:01
T42<DSstill01> I repaced it as you wanted20:01
T42<meierrom> @OhYash: the fx1tec is probably the phone with the smallest bezels currently but it costs a little fortune and the port is still wip.20:01
T42<DSstill01> I,overwrote the normal one, i didnt copy it with its name20:02
malI reall have no idea if you have done everything I said correctly20:03
T42<OhYash> @meierrom True tho. Too expensive for me.20:04
T42<OhYash> Can think of porting SF to one of the new notch-less ones but I'll have to find one with source available20:04
T42<DSstill01> I did it, trust me20:05
T42<OhYash> Though if I ignore source availability and port support, Meizu's upcoming 16Xs looks like the best shot to minimize bezels amongst all these trying. I'm eyeing on that one.20:06
mal@DSstill01 during the years of porting I have learn not trust anything20:14
T42<meierrom> @OhYash: the xa2 plus is a real beauty with smaller bezels and it's Jolla supported.20:14
mal*not to trust20:14
T42<DSstill01> Trust me, trust me ;)20:14
T42<OhYash> I'm with mal on this one. We never know what we miss out20:15
T42<DSstill01> It was a file to copy in a directory, come on lol20:16
T42<OhYash> lol20:18
T42<OhYash> @meierrom XA2 has always been there but I never really considered it until now. Looks beautiful indeed.20:19
T42<DSstill01> I found the way to debug with pactl and when i enable fm it says that active port is input-fm_tuner, so, in sailfishos side it should be correct,right?20:50
malso it seems21:00
malhopefully the .pa is correct for that21:00
malthere were some routing issues on some devices, routing was done at wrong time so it wasn't really used21:01
T42<DSstill01> In my opinion, it's routing to the correct path found in mixer_paths21:01
malI assume you have tested using fm radio in android and it works there?21:02
T42<DSstill01> Yes, with some different audiomanager parameters set to start it respect the ones uded in lineage21:03
T42<DSstill01> But pa doesnt seem to use any of them, included the ones used in linesge FMapp221:04
T42<DSstill01> Or whatever it's called21:04
deathmist could there be something wrong in my audio, media & codec symlinks? still no audio/mic during call, just very low volume random crackling on the other side. I have no. even with proper mixer_paths.xml and media_profiles.xml files from OOS /vendor same issue persists :/21:35
deathmisthere's pactl list device is cheeseburger (OnePlus 5) and on 15.1 base21:37
piggzmal: apart from grepping init and .sh files, any suggestions for kicking the wlan into action?21:39
piggzso far ive search for insmod, /sys/, echo, write, wlan, wifi and probably some more21:40
piggzpoll, polling21:40
piggznothing has jumped out so far21:40
malpiggz: did you make sure the mounts are correct and the symlink I said is there?21:46
T42<adampigg> Mal, yes there is a symlink for /lib/firmware....21:47
malnot that21:48
T42<adampigg> Which?21:48
maldid you have /vendor/firmware_mnt? or something like that21:48
maland you symlinked it to /firmware?21:49
T42<adampigg> Firmware_mnt was empty, and not touched it21:50
malbut afaik based on fstab there should be something mounted there?21:51
malare all of the things in fstab mounted (well the reasonable ones)21:52
piggzmal: probably not...... /mist, /vendor/firmware_mnt /vendor/bt_firmware and /mnt/vendor/persist are missing...21:57
malpiggz: what did I say to you about fixing those :D21:57
maltrust me, I know what I'm talking about21:58
piggzaye aye, looking21:58
piggzi got lost looking at kernel code21:58
malpiggz: I think I said to fix those first and if it won't help then look into kernel21:58
piggzmal: they mount manually....21:59
piggzi wonder if its an order thing21:59
piggzlike i think Mister_Magister said, after=vendor22:00
malpiggz: so try wlan after mounting those and then checking that the symlink I said works22:00
malalso add symlink /persist22:00
malto obvious place22:00
Mister_Magisterpiggz: i think systemd should manage order but JUST IN CASE it doesnt22:00
malwell you can get a log from systemd to see the order in which services ran22:01
piggzmal: /firmware is already a link to vendor/firmware, not /vendor/firmware_mnt22:04
piggzthese are my mounts now
piggzcant link to / as ro, and cant make rw22:07
Mister_Magistermal: when we can expect fix btw22:07
piggzMister_Magister: get in line22:07
Mister_Magisterpiggz: >:-(22:07
Mister_Magisterfix would be useful and it sounded like it won't take much time22:08
Mister_Magisteri like how piggz is doing is part22:08
piggzmal: the kernel messages i see come from calls to pr_info .... there is a pr_info i would exepct to see if the driver started to probe, and load a firmware22:09
piggzthink im calling it a night, late ... Mister_Magister, your time is now :D22:10
malpiggz: why is / readonly?22:11
malMister_Magister: forgot about the fix already, I'll try to remember tomorrow22:11
malpiggz: did you add the /firmware symlink yourself earlier22:12
T42<adampigg> Mal, i dont think so...i added it /lib/firmware as that was in the search path22:13
T42<adampigg> In the paste you provided22:14
malwhy is / readonly?22:15
malI doubt sfos likes that22:16
T42<adampigg> I did wonder that22:16
malanything about that in log, maybe you have gotten root a bit corrupted22:17
T42<adampigg> Mal, it isnt...22:19
T42<adampigg> I can remove a file in /, but cant create a link. ..22:20
T42<adampigg> Mal,.user fail22:23
T42<adampigg> Works if delete existing link first,.it wanted to creat link inside /firmware, which was.ro22:23
malany difference in behavior after correct mounts and symlinks22:25
T42<adampigg> Mal nope, messages end at 'driver loaded'22:25
T42<adampigg> No probe message22:26
maldid you also add the /persist symlink22:28
T42<adampigg> Mal, i mounted /persist22:29
Mister_Magister / could be ro due to missing patch or new android22:30
T42<adampigg> No, its not ro, my bad22:30
*** ahjolinna_ is now known as ahjolinna23:48

Generated by 2.17.1 by Marius Gedminas - find it at!