Thursday, 2019-01-03

*** ChanServ sets mode: +v T402:45
entilmy fp2 is acting up again :( it's always been prone to reboot itself, esp. when I use the browser, but now it doesn't find its sim cards again at all... and to top that off, there are unsynced calendar events07:50
*** ChanServ sets mode: +v T408:22
T4<srmane4444> Hi everyone. I also port with @MarcoDS_bit and @karan5chaos SFOS for wyred. Build Hybris-hal. Will post the log on pastebin: Why so many warning s and how to fix them?09:09
T4<It_sMike> Hey sailors :)09:43
T4<It_sMike> Tried and checked everything that I could, but the sensor problem is still present...09:45
T4<It_sMike> Could this problem relates to 50-ffmemless.ini? because I make a bit change to those....09:47
T4<It_sMike> Just deleted 50-ffmemless.ini and done something like this : …
r0kk3rzdo you use ffmemless driver?10:01
r0kk3rzthats also vibra stuff, not really anything to do with sensors10:01
T4<It_sMike> Yes..., So... Any idea?! About this weird problem?10:05
r0kk3rzwhats the actual problem?10:06
T4<It_sMike> Sensors are not working....10:06
r0kk3rzin detail?10:06
T4<It_sMike> In action ... Actually10:07
T4<srmane4444> @srmane4444 [Hi everyone. I also port with @MarcoDS_bit and …], No answer?13:50
T4<srmane4444> @srmane4444 [Hi everyone. I also port with @MarcoDS_bit and …], [Edit] Quote: No answer?13:50
T4<srmane4444> @srmane4444 [Hi everyone. I also port with @MarcoDS_bit and …], [Edit] @M1kaela_freenode_relaybot Quote: No answer?13:51
malfirst of all do use reply in telegram, it looks terrible in IRC side13:54
mal*do not use13:54
T4<It_sMike> :)13:54
mal@srmane4444 did the other porters build hybris-hal successfully yet? do they get the same error?14:01
r0kk3rzedit also spams us14:01
malyes, that too14:01
*** liar is now known as krnlyng14:20
T4<It_sMike> mal? Just find out this: on sailfish 2.2, sensors working fine with the current config, but on 3.0 non of them works... Any idea!?14:22
malthe question is whether the issue is in android side or sailfish side14:22
malI did ask for logs14:23
T4<It_sMike> I'll send them to you once I got home :)14:24
*** ChanServ sets mode: +v T414:41
portagerConnection check. Previously, it did not work register.14:44
T4<srmane4444> @mal [first of all do use reply in telegram, it look …], mal, I did it like you said.14:47
T4<MarcoDS_bit> I got a new error!
portagerSo, I port SFOS for whyred, like MarcoDS_bit, it went a little further, but we can’t properly organize the work to work together. What should I do and what should he do.14:51
T4<srmane4444> I'm just not used to talking in a group where bot sends all this garbage to another resource. Excuse me. I'll keep it on mind.15:03
T4<It_sMike> @srmane4444 So use IRC then :)15:15
T4<It_sMike> mal which logs should I send for you ?15:16
T4<It_sMike> only journald.conf?15:17
mal@It_sMike output of commands journalctl and /usr/libexec/droid-hybris/system/bin/logcat15:30
T4<It_sMike> on the way ...15:31
T4<MarcoDS_bit> mhh any help?15:41
malthat says error is because compatibility_matrix.xml fails15:45
T4<It_sMike> Output of journalctl: …
T4<MarcoDS_bit> mal, I don't actually have a  … compatibility_matrix.xml in the said directory, wasn't it supposed to be created automatically? :P15:47
mal@It_sMike could you please stop skipping the http:// from beginning of links, it's annoying that I have to copy the link and not open it directly15:49
T4<It_sMike> Output of /usr/libexec/droid-hybris/system/bin/logcat : …
mal@It_sMike ah you have the usual problem, too new sensorfw hybris build15:50
T4<It_sMike> Oh sorry mal..... didn't know that.... so sorry...15:50
T4<It_sMike> … Here you are :)15:51
T4<It_sMike> ah you have the usual problem, too new sensorfw hybris build  … Hmmmm.... What's this problem then?15:51
T4<It_sMike> Could you explain more about it?15:53
malgo to hybris/mw/sensorfw, run git reset --hard 7af22d7d77a8c9aca2bd8029a5337e4428ab5239 and then back to $ANDROID_ROOT and then rpm/dhd/helpers/ -b hybris/mw/sensorfw -s rpm/sensorfw-qt5-hybris.spec15:53
T4<It_sMike> on the way :) THANKS a lot mal :)15:54
T4<MarcoDS_bit> @MarcoDS_bit [], I get this, "No android-base.cfg is found in these paths:' '", do anyone know in which file I have to put the right directory gor the android-base.cfg file?15:58
T4<It_sMike> Everything's working fine but flash light :/16:23
T4<It_sMike> But let me build a new zip with mic....16:24
malmaybe you need something like this
T4<It_sMike> Should try it! Thanks again mal :)16:24
malassuming you meant the flashlight shortcut16:24
T4<It_sMike> Yes16:25
T4<It_sMike> Stucked on " on " toggle :/ and nothing happens16:25
entilmal: 01-03 16:44:55.872  9497 12155 E qdhwcomposer: isValidDimension: layer handle is NULL16:45
entilmal: is that to be worried about?16:45
entilI did `grep -v` on that to see if there's anything about the modem not being found, but there isn't16:49
entilbut I need my phone to work as a phone :(16:51
entilI still have that japanese sim card, I removed it and now csd told me it found one modem out of two17:00
entilgonna see what I get with no sim cards17:00
entillooking at settings, there's this brief view of the sim card selection that told me no sim1 was inserted but then replaced by a text saying no sim cards are17:00
entiland with no sim cards, no modems found either17:02
malentil: "isValidDimension: layer handle is NULL" is normal17:07
entilmal: ok, and the weird thing is, now that I inserted only the one sim into slot1 (last time was slot2) it worked... and I'm not excluding hardware problems because all hardware seems to be shit, but it would be nice to know how to verify...17:09
entilI have all this other weirdness, like usually when I power down the phone, it boots back up immediately, like 4/5 times, and these *constant* reboots when reading the web17:10
malentil: browser is quite bad, some websites cause out of memory issues17:12
entilmal: like apparently :P17:13
entilanyway, I guessed it might be oom, because apart from hardware, all websites are shit too17:14
malentil: need to check why oom is killing the device and not just browser17:14
entilmal: is there a persistent log across reboots?17:14
entilnow I'm just worried I'll 1) read the web 2) have the phone reboot 3) lose the modems again17:16
entiljournalctl does not mention oom17:17
malentil: yes, if you enable Storage=persistent in /etc/systemd/journald.conf17:17
entilmal: ok, thanks, I did that and restarted journald :) let's see what happens next time it dies...17:19
malI need to dig out one website that caused reproducable oom17:19
entilJan 03 19:19:33 Sailfish [18515]: [D] unknown:0 - connection-manager.cpp 107 setupSocketConnection p2p error: QDBusError("org.freedesktop.DBus.Error.FileNotFound", "Failed to connect to socket /run/user/100000/signond/socket: No such fil17:20
entilmal: pretty much anything :P but I tend to read and its linked articles (many of which on on my commutes17:21
entilJan 03 19:19:33 Sailfish [18519]: signondaemon.cpp 388 init Failed to SUID root. Secure storage will not be available.17:21
entilwhat's with all of this now17:21
entilJan 03 19:19:56 Sailfish kernel: [17:19:56.986458] [00000003F0E9AD81] [wpa_s]  wlan: [E :HDD] wlan_hdd_get_frame_logs: Frame Logging not init!17:22
T4<MarcoDS_bit> Is that normal that my hybris-boot.img is only 25.8MB?17:28
T4<It_sMike> mal, I still have the problem with flash light, even clean install ( creating new system with mic ) and adding 999-torch.rules under the rules.d directory not fixed that... So what should I do now :) ?17:33
T4<It_sMike> Is the problem refers to sensorfw again or it's from somewhere else?17:35
T4<It_sMike> Oh and forgot to say, I've already defined flashlight in hw-settings.ini17:36
mal@MarcoDS_bit hybris-boot.img is the kernel, not the rootfs17:39
T4<MarcoDS_bit> oh right17:40
mal@It_sMike logs, journalctl at least, you can also try to enable flashlight manually via /sys, you just need to find the correct path17:40
mal@It_sMike did you make sure the path in 999-torch.rules is correct for your device?17:40
entilmal: are those errors acceptable? I usually freak out when I see errors I can't understand but sound scary ;P17:59
T4<eugenio_g7> \o/ got audio somewhat working, a bit distorted but at least something comes out from the speakers!18:00
malentil: probably fine18:06
T4<It_sMike> did you make sure the path in 999-torch.rules is correct for your device? … Oh yes :) didn't attention to path, should take a look at it again. Sorry for being noob :D18:06
entilmal: wlan works, but what's that suid one?18:07
malentil: not sure, need to check if I get that also18:31
malentil: so what really is broken in your device now?18:31
entilmal: calendar syncing as well18:36
entilmal: I have the msyncd debug magics somewhere in my irc logs, but probably won't get to debug tonight18:37
entilmal: and then I'll look at journalctl after the next uncontrolled reboot18:37
entilI did have time to find this:18:47
entil[W] unknown:0 - Profile file not found: "/home/nemo/.cache/msyncd/sync/caldav-sync-28.xml"18:47
entil[W] unknown:0 - Failed to load profile: "caldav-sync-28"18:47
entil"something's broken"18:47
malentil: ok, I have never used calendar sync19:11
piggz@eugenio_g7 im starting to understand how this driver works wrt suspend/PM20:26
piggzi can now get the system into a suspend state, and have the  work20:26
piggzim just trying some more configurations to see what options are available20:27
piggz@eugenio_g7 how does this look?
piggzmal, r0kk3rz: ^^ finally20:40
piggzlost days on a fscking wlan droiver!20:41
r0kk3rzhow did you fix it?20:41
piggzthe driver has a config option for supporting runtime power management20:41
malpiggz: ah, so a bit like the mess in xperia x bt20:42
piggzthe runtime PM rarely (never?) goes into a sleep state becuase of a wakelock20:42
piggzthe wakelock also prevents mem sleep20:42
piggzthere is a also a config option for wakelocks20:42
piggzso, turning off wakelocks and runtime PM allows sleeping20:42
piggzwhich all seems a little backwards!20:43
r0kk3rzslightly counter intuitive20:43
piggzwith runtime PM and wakelocks enabled, I commented out the one wakelock (event lock) that prevented sleeping generally .....20:45
piggzthis resulted in a system that couldnt connect to networks becuase the runtime PM kicks in after 100ms....20:45
malpiggz: maybe you should write some notes to faq20:45
piggzi extended the runtime PM timer to 5000, which worked, but i sense this was a bit hacky!20:45
piggzand probably only worked becuase it associated before the timer expored!20:47
T4<eugenio_g7> piggz: awesome!20:50
*** ChanServ sets mode: +v T420:54
piggz@eugenio_g7 feel free to confirm ... just comment those 2 lines in the Makefile20:58
piggzjust tring now with and without the deep_sleep config option21:03
piggzi dont think it doesn anything on linux anyway, its an android thing, and seems can only be triggered from an android command interface21:03
T4<eugenio_g7> Will try21:09
piggzso, if you could get audio working.......!21:12
piggz@eugenio_g7, with the device unplugged, wifi on, and screen off ... im getting approx 50% suspend state21:14
T4<eugenio_g7> working on it! :) Made some progress, audio now plays although it's pretty distorted21:14
T4<eugenio_g7> but ok21:14
piggzwhat is involved with that21:14
piggzalsa or PA configs?21:14
T4<eugenio_g7> alsa, it sounds bad even while using aplay21:15
T4<eugenio_g7> but I guess it's a matter of getting the correct controls right21:15
piggz@eugenio_g7 not sure how to automate fixing BT ... we need to change the filename of a file on /system so that it is loaded ... i hackyish systemd script ought to do it, but not ideal, and only needsd done the once21:21
piggzmal^ thoughts?21:21
piggzor, modify bluez5-tools to loaded it, but that seems OTT!21:21
T4<eugenio_g7> hmm, could we ship the file directly? I don't think we need the android rom anymore21:22
malpiggz: which file?21:23
piggzmal: that BT firmware blob21:23
malpiggz: can't you make a symlink in /usr/libexec ...21:24
maldepends on how it's loaded21:24
piggzloaded by hciattach21:24
malpiggz: which exact folder?21:24
piggzhciattach loads from /etc/firmeware21:25
piggzwhich, I have symlinked to /system/vendor/firmeware21:25
piggzin there is bt/21:25
piggzthe system ships with BCM4356A2_001.003.015.0084.0268_ORC.hcd21:25
malpiggz: instead of symlinking the folder symlink needed files and make sure the symlink has the correct name?21:25
piggzbut needs to be called BCM4354A2.hcd21:25
piggzyes, maybe could do that21:26
malpiggz: my way doesn't require renaming anything in /system21:26
malpiggz: sometimes the solution is simple :)21:26
piggzi need to check if hciattah will load from a symlink ... i think i tried, but will try again21:27
T4<eugenio_g7> piggz: we could ship firmwares in a separate package and use whatever name we want21:27
malthat is not preferred21:27
T4<eugenio_g7> I'd like to avoid requiring the base LOS ROM21:28
piggzfull on port!21:28
piggzwell .. we have options21:28
T4<It_sMike> mal the problem was from droidmedia :)21:29
T4<eugenio_g7> we don't use the android bits anymore after all (excluding the kernel)21:29
T4<It_sMike> should fix droidmedia....21:29
mal@It_sMike what was the problem there?21:30
T4<It_sMike> flashlight, camera and video playback21:30
T4<It_sMike> all of them gone when I installed the old droidmedia package21:31
piggz@eugenio_g7 wonder if we can use v4l for the camera?21:31
mal@It_sMike not sure if I understand, so those work with old one but not with new?21:32
T4<It_sMike> this one : …
T4<It_sMike> yeah!21:32
malso what is the difference in that, that looks like it has some patch21:33
T4<It_sMike> but when I build the droidmedia using its old source face with those problems ag ain.... any idea!?21:33
malbased on the version21:33
T4<eugenio_g7> piggz: not sure, but getting the cameras to work would be pretty nice, given that they work only on miui and windows21:33
T4<eugenio_g7> there is also the funky battery reads to fix21:34
T4<It_sMike> maybe !?21:34
mal@It_sMike anything in logcat?21:34
T4<It_sMike> didn't checked it yet....21:34
T4<eugenio_g7> I don't think it's Sailfish related though, I flashed Lineage the other day to tinker with a working audio stack and it experienced the same issue21:35
T4<eugenio_g7> so there is something pretty wrong on the kernel side21:35
T4<eugenio_g7> that, or both our batteries aren't healthy21:35
T4<It_sMike> mal, did other porters apply patchs on droidmedia to make it working on them devices?21:36
piggzah yes, it works on windows, i wonder what interface it used21:37
piggzyeah, my battery is all over the place!21:38
mal@It_sMike I use one patch myself21:38
T4<It_sMike> Hmmm.... is it too hard to be done !?21:38
malcheck logcat like I said21:39
malor pastebin the output21:39
T4<It_sMike> with broken droidmedia? or with working droidmedia?21:39
malbroken of course21:41
malwhy would I want look at logs from working one?21:41
T4<It_sMike> :) sorry for being noob one again21:41
T4<It_sMike> once*21:41
T4<It_sMike> Here's the logcat dear mal :) …
mal@It_sMike as I suspected you need to apply this to your for droidmedia
T4<It_sMike> Let me try it mal22:07
T4<It_sMike> And THANKS again :) … You helped me a lot :)22:07
malI assume you know how to get the raw patch from github? or add git remote and cherry-pick the commit22:08
T4<It_sMike> I know it... it's ok mal :)22:10
maldo you remember what you need to rebuild after that?22:12
malyou don't have to rebuild everything22:12
T4<It_sMike> gst-droid ?22:12
malyou do the chapter 13.2 but you can skip gst-droid, just do the droidmedia parts22:13
T4<It_sMike> :) OK then22:13
T4<It_sMike> it would be take a while for me....22:13
piggzcancaro is a xiaomi isnt it?22:15
piggzi too use that patch22:15
T4<It_sMike> Yes :) Xiaomi Mi 4 / Mi 322:15
UmeaboyGuys, I still can't build hybris-hal with the r47 branch.22:42
UmeaboyI know I've followed every instruction correctly from the HADK PDF.22:43
malwhat is the error now?22:43
UmeaboyStill the same:
malshow your local manifest22:51
T4<linusdan> People, I was able to compile the zip of the sailfish for the i9100! Installed via TWRP but not normal out of the samsung logo when restarting :(22:57
UmeaboyThere is no local_manifests in $HOME/hadk/.repo/ as the HADK doesn't tell you to create one manually. The only manifest is the one for the hybris aosp branch that I initiated.22:59
malUmeaboy: how about chapter 5.2?23:02
malUmeaboy: how in the hell did you even get it to build without local manifest?23:02
UmeaboyI don't know to be honest.23:03
malgrep .repo for the device repo23:03
Umeaboyps aux | grep maple ?23:05
malps aux, wtf, wy don't you think a little23:06
Umeaboy /home/kristoffer/hadk/device/sony/maple23:06
malyou are missing completely what I'm asking23:07
UmeaboyOK. What ARE you asking for?23:07
malWHAT DAMN MANIFEST IS THAT REPO DEFINED???????????????????????23:07
UmeaboyI haven't got the slightest idea of which manifest it uses.23:16
malUmeaboy: so you don't understand at all how android build works?23:17
r0kk3rzwell its pulled in by something, it doesnt get there by magic23:17
Umeaboymal: Not down to the very core.23:17
malthis is like the simplest part of android build23:17
r0kk3rzthis isnt the very core, this is the basics23:17
r0kk3rzlooks like the default.xml has device items23:21
UmeaboyYeah, but shouldn't that create a manifest from sonyxperiadev after you've done the lunch command?23:23
malUmeaboy: since you don't even try to solve anything I solved it for you, check that device/qomc/common repo history and you will see that you need older revision, you'll see the commit that brought in json-c23:24
malit was quite obvious that if there is conflict with two repo then git log should be checked23:25
malthe reason is probably that we use r47 where as sony already uses r5223:25
Umeaboymal: That would be good to add to the HADK.23:25
malhadk is not that specific23:27
maland sony manifest is special23:27
UmeaboySo I just switch branch to a lower revision and repo --fetch-submodules and then try again?23:28
malsony common doesn't have older revision, it has just o-mr1 branch23:29
maljust go to the damn folder, check git log for suitable commit and then git reset --hard some_hash23:29
malsorry for being a bit rude, I'm way too tired and get annoyed easily when I don't sleep enough23:36
UmeaboyNo problem. Understandable.23:49
UmeaboyI think the main difference between this build of hybris-hal is that the build guide for Xperia X (that DID build the image) and following the equivalent instructions in the HADK PDF is that it uses an older AOSP version: repo init -u git:// -b hybris-sony-aosp-6.0.1_r80-20170902 -m tagged-manifest.xml23:52

Generated by 2.17.1 by Marius Gedminas - find it at!