Wednesday, 2014-08-06

vgradeSep 26 00:53:17 Jolla DSME[688]: IPHB: rtc at 1970-09-25 23:53:16 GMT00:00
vgradeSep 26 00:53:16 Jolla DSME[688]: IPHB: set to 2014-08-05 23:47:29 GMT00:00
vgradelook better00:00
vgradeand time was set correctly from mobile web00:00
vgradeon reboot00:00
vgradewill do a PR now00:00
*** cxl000 has quit IRC00:01
*** icelaglace has joined #sailfishos-porters00:05
vgradesledges: situ
*** diverse has left #sailfishos-porters00:42
*** boilerkim has quit IRC00:44
*** icelaglace has quit IRC01:05
*** alin has quit IRC01:06
*** arcean has quit IRC01:08
*** alin has joined #sailfishos-porters01:24
*** alin has joined #sailfishos-porters01:25
*** beidl has joined #sailfishos-porters01:59
*** beidl_ has quit IRC02:00
*** chrisi has quit IRC02:01
*** chrisi has joined #sailfishos-porters02:28
*** boilerkim has joined #sailfishos-porters02:42
*** dmt_ has quit IRC03:29
*** furikku has joined #sailfishos-porters03:59
situvgrade: I have already tried disabling that service, can you show me full journalctl log after disabling it ?04:58
*** ryukafalz has quit IRC05:12
*** ryukafalz has joined #sailfishos-porters05:26
*** VDVsx has quit IRC05:31
*** mailyaseen has joined #sailfishos-porters05:44
*** boilerkim has quit IRC05:53
*** boilerkim has joined #sailfishos-porters05:54
*** VDVsx has joined #sailfishos-porters05:58
*** phdeswer has quit IRC06:29
*** lpotter has quit IRC06:34
*** lpotter has joined #sailfishos-porters06:34
*** gogeta_ has joined #sailfishos-porters06:43
*** dmt_ has joined #sailfishos-porters06:49
*** vakkov has quit IRC06:55
*** giucam has quit IRC07:03
*** giucam has joined #sailfishos-porters07:03
*** giucam has joined #sailfishos-porters07:03
*** iTune has joined #sailfishos-porters07:14
alinvgrade: I have that disabled07:20
alinsitu: what happened with the time?07:21
alinI have that service deisabled but dsme still fails07:21
spiiroinalin: I might have some time later today to check the dsme behavior07:23
*** iTune has quit IRC07:26
*** carepack has joined #sailfishos-porters07:30
alinspiiroin: cool... ping me if you need any testing07:36
alinI shall be online once in office...07:36
*** alin has quit IRC07:38
*** kimmoli_ has joined #sailfishos-porters07:55
*** VDVsx has quit IRC07:56
*** kimmoli has quit IRC07:58
*** kimmoli_ is now known as kimmoli07:59
*** boilerkim has quit IRC07:59
*** boilerkim has joined #sailfishos-porters07:59
vgradesitu: dmse still gives the same error but the time on the device is set correctly08:04
*** VDVsx has joined #sailfishos-porters08:11
*** cxl000 has joined #sailfishos-porters08:15
gogeta_morning guys08:16
gogeta_I hate be in forced vacation08:17
gogeta_this week i work just two days :-(08:17
situvgrade: How many reboots you tried ?08:18
vgradesitu: 308:20
situsledges: can you confirm it works ?08:20
situsledges: on your device08:21
situvgrade: I had tried commenting out time_daemon service several times and it didn't fix the issue for me.08:21
*** alin has joined #sailfishos-porters08:21
*** alin has joined #sailfishos-porters08:21
vgradesitu: it does not fix the setting of the hardware rtc as this seems to be an issue even in stock android,
situvgrade: I understand that.08:22
vgradebut I'm getting a correct time picked up from network (wifi or mobile web)08:23
vgradetime shows as 1970 initially then gets updated08:23
*** phdeswer has joined #sailfishos-porters08:26
vgradesitu: + 0408 nightly08:27
dmt_this is my experience with time too08:27
dmt_uptime of 40y08:27
vgradedmt_: :)08:27
situvgrade: please disable setting time from mobile web or wifi08:28
situvgrade: it works for me too08:28
situvgrade: do not connect to mobile or wifi while testing this.08:28
situmobile internet *08:29
situwe want to pickup time from phone network.08:29
vgradesitu: well it wont set time because RTC is broken on N5 it seems08:29
alinvgrade: yap... that I have seen too... if I set it to automatically...08:29
alinvgrade: what surprises is the lack of DSME failure08:29
vgradealin: there is still the same failure message08:30
alinvgrade: ok... then is the same situation as last week08:30
alinvgrade: I thought we documented this...08:30
situvgrade: do you get correct time if you disable mobile web and wifi ?08:30
vgradesitu: right I understand now, you want to have time set without wifi or mobile internet08:31
alinvgrade: only half...08:31
situvgrade: Yeah08:31
*** boilerkim has quit IRC08:31
alinvgrade: yes... we would like dsme not to fail08:31
*** boilerkim has joined #sailfishos-porters08:31
alinvgrade: I will add now the info about automatic setting08:32
vgradealin: well thats not a dsme issue its an issue with the hardware, which is why qcom guys was so helpfull08:32
alinvgrade: I do not know I missed the help from qcom guy...08:33
alinvgrade: you mean the comment on situ post08:33
vgradealin: I was being sarcastic08:33
alinvgrade: good one...08:35
vgradesitu: so if the phone network does not supply the time then I don't see how without a settable RTC we can progress08:35
alinvgrade: spiiroin promised he may look today into dsme08:36
alinvgrade: how do you know android has the same issue? I do not remember testing this08:36
vgradesitu: we saw from the ophono test that there was no time data from both yours and my operators08:36
situvgrade: but on Jolla setting automatic time was working fine.08:37
situvgrade: was it being picked up from mobile internet or wifi ?08:37
vgradesitu: I don't think we did that test08:37
alinsitu: sledges vgrade dmt_ there is a snapshot today on cyanogen for hammerhead08:38
vgradesitu: I only showed you the output from the ophone test08:38
alinI will test with it and maybe we shall update our works based on it... especially as snapshots seem to stay on the cyanogen website for long time08:38
situvgrade: we did the test, read the backscroll in PMs.08:38
vgradesitu: we also turned off auto and wifi, set clock back, then turned on auto again and time was restored08:44
vgradeI guess I had mobile web on08:44
vgradealin, 11:22 < vgrade> situ: it does not fix the setting of the hardware rtc as this seems to be an issue even in stock android,
situvgrade: I am aware it's an issue with stock android.08:45
vgradesitu: that was fro alin , 11:36 < alin> vgrade: how do you know android has the same issue? I do not remember testing this08:46
alinvgrade: situ cool... this starts to sound like the other bug we have seen on chromium... pasted the link some time.08:50
alini remember the conclusion was about a race at the shitdown... and normal failure as the clocks are too far apart08:51
situvgrade: 11-05 16:03:22.329 W/        (  776): Unable to set rtc to 1383667402: Invalid argument08:52
situvgrade: This error is seen when dt property to enable write access to rtc is not set.08:52
*** VDVsx has quit IRC08:53
*** dmt__ has joined #sailfishos-porters08:53
*** dmt_ has quit IRC08:53
*** morphis has quit IRC08:53
*** sledges has quit IRC08:53
*** sledges has joined #sailfishos-porters08:54
*** sledges has quit IRC08:54
*** sledges has joined #sailfishos-porters08:54
vgradesitu: so the same issue08:54
vgradeand qcom man says don't chenge the dt property08:55
*** vrutkovs has joined #sailfishos-porters08:55
situI believe qcom was pretty dumb to disable write access to rtc.08:55
vgradeso without a settable hardware RTC or time picked up from wifi, mobile web, or mobile network then we will have wrong time08:56
situvgrade: time_daemon utility in android stores the time on filesystem somewhere.08:57
jusa_what's the HAL code for hammerhead/android 4.4.4 ?08:57
situvgrade: but still I wonder how dsme picks up the right time to set to rtc.08:57
vgradejusa_: HAL code?08:58
situvgrade: I had mobile web and wifi disabled on my device, still dsme tried to set correct time to rtc and failed.08:58
*** morphis has joined #sailfishos-porters08:59
alinvgrade: situ
jusa_vgrade: I mean for audio... hardware/VENDOR/audio in source tree09:00
vgradejusa_: sex09:00
situvgrade: embarassing :P09:00
alinsitu:  why?09:01
jusa_why's this commit?
alinsitu: sex is as important as food... generates all that strange chemicals in the brain09:01
jusa_I mean, weren't these patches for some other device than i9305?09:02
vgradejusa_: yes they were from hammerhead09:02
vgradeexcept that one09:03
jusa_vgrade: yep. so why is that patch there now? :) i9305 works without that IIRC.. the problem was input route setting messing with the output port..09:04
*** VDVsx has joined #sailfishos-porters09:05
jusa_this one:
vgradejusa_: ok, I can't recall if this was needed as we did not start on hammerhead until april and that commit is from feb09:07
vgradeso we should have had it when we started work on hammerhead09:07
*** dmt__ has quit IRC09:07
vgradeI can test without09:08
vgradeoh your commit is from sledges repo09:08
*** iTune has joined #sailfishos-porters09:09
sledgesmy repo only adds a fix to sbj09:09
sledges(and renames commits)09:09
jusa_I just picked that from there .. it's in upstrem master as well
sledgesi guess vgrade means to test without
vgradesledges: yes09:10
vgradejusa_: what exactly do you need from the n5 source tree I have it here09:11
jusa_at least if there are some workarounds needed only for one device, comment for all devices and preferably ifdef09:11
jusa_vgrade: hardware/VENDOR/audio is enough09:11
jusa_vgrade: and ./system/core/include/system dir as well09:12
jusa_vgrade: ..and ./hardware/libhardware/include/hardware09:12
sledgesjusa_: if you're asking whether speaker_drc_enabled is qc specific - nope, it's HALv2 specific, checked in tree yesterday09:13
sledgescan give you exact src line09:13
jusa_sledges: yep, but I still would really like to see the whole source + headers :)09:13
*** iTune has quit IRC09:13
jusa_if possible09:13
*** SK_work has joined #sailfishos-porters09:14
*** alin has quit IRC09:17
sledgessecs ;)09:20
vgradeI've tarred09:20
sledgesok, i was after only this:
sledgesthanks vgrade09:21
*** alin has joined #sailfishos-porters09:24
*** alin has joined #sailfishos-porters09:24
jusa_vgrade: thank you09:28
vgradesitu: sledges alin so time where are we?09:31
vgradestill blocked for release?09:31
alinvgrade: yes and no I will say...09:32
alinvgrade: practically if we let it on automatic all services that check a certificate to login do not login09:33
alinas they try to conect before the time is set09:33
alinvgrade: anyhow there is little we can do about...09:33
alinvgrade: so all we need is to document09:33
sledgeswe can switch automatic off during install09:34
vgradealin: I don;t follow the logic aout services and certificates.09:35
sledgesvgrade: these days time has to be set to world's time, otherwise phones (e.g. https) stop working :D09:36
vgradeIf you are connecting to services then you need a net connect and so time will be set correctly09:36
sledgestime doesn't get set correctly immediately09:36
sledgeseven if you are connected09:36
sledgesdifferent strategies whether your have only modem enabled, or also wlan09:37
vgraderight so you can;t use the sevices until time is corrected09:37
vgradeok, if we are to block on this09:38
vgradethen what do we need to implement09:38
alinvgrade: set up a google account... if the time is wrong when you start the phone... the account will not come online09:38
alinvgrade: nothing09:39
vgradealin: what do you mean09:41
alinvgrade: there is little we can do about that...09:41
alinunless instruct users to connect by hand once the time is correct09:42
vgradeI'm assuning we can't fix the root cause of the HW RTC not working so we have to determine if te current sailfish components can be configured to work with a broken RTC09:43
vgradeif not we need to implement that functionality by storing time in filesystem09:43
sledgesand to not power-off for long periods :)09:44
sledgesi wonder what's google's grace for time skew09:45
sledgesvgrade: are you up to speed with this discussion: ?09:45
vgradelets discuss with spiiroin when he gets time09:45
sledgesand do we have whole braindump of situ's findings?09:45
situsledges: I was told that timed-qt5 should read time from network and show correct time even if dsme fails to set rtc.09:48
vgradesitu: which network09:49
situvgrade: phone network.09:49
sledgesdo we admit defeat in fixing rtc ourselves?09:50
vgradesitu: so how long could that take09:50
vgradesledges: qcom guy says "don't do that"09:50
situsledges: I believe setting rtc is disabled through TrustZone.09:51
vgradesledges: and situ tried to do that and it still did not work09:51
sledgesvgrade:p yet time_daemon manages09:51
situsledges: time_daemon never sets rtc09:51
dmtis time also broken in the same way on cm/android?09:51
vgradesledges: it stores it on disk09:51
sledges... speechless, really!?09:51
*** Nokius has joined #sailfishos-porters09:52
situdmt: if I reflash my Nexus 5 with stock, I have to set time manually on first boot.09:52
sledgesdmt: fastboot boot boot.img from cm to quick check :)09:52
situdmt: and default time is set to 70's on first boot even with stock android.09:52
situsledges: I have checked with stock android, it has same issues.09:53
dmtand they get away with it by storing in fs then?09:53
situdmt: right09:53
sledgesriiight thanks now that was refreshing09:54
dmtwho would have thought on a flaship phone there is no working rtc..09:54
dmtfine ingeneering09:54
*** Nokius1 has quit IRC09:54
sledgesno wake-up alarms from powered off phone?09:55
situsledges: I don't think N5 support it.09:55
alteregodmt: why don't you think there's a working RTC alarm in N5?09:55
dmti don't09:55
dmtalterego: i'm only very surprised at the reality09:56
alteregoHrm, anyone tried setting an alarm under android and turning phone off?09:57
sledgesalterego: backlog? :)09:57
sledgestime is not saved across reboots09:57
situalterego: I think I have tried that once, but let me try it again.09:57
sledgesthey store it on fs09:57
alteregoOh, now I see :)09:58
* sledges is looking for a door handle09:58
sledgesso we are just a store-on-fs-fix away for dsme ;)09:59
situsledges: but when the phone boots, dsme tries to setup correct time on rtc and fails. How does dsme know correct time ?09:59
sledgessitu: debug?10:00
dmthow would storing on fs work if say the phone wakes up without network?10:00
sledgesdmt: you would only be few hours away, even less if you only rebooted10:00
vgradeIIRC timed changes system time when nitz date is received from cellular network; possibly connman could play some role too (ntp?)10:00
vgradevgrade: that from spiiroin last chat10:00
dmtfew hours out are a big deal if you are traveling by public transport10:00
sledgessitu: at the beginning we were sure time_daemon gets values from hwclock on boot. how did you discover it is on filesystem?10:01
situsledges: I remember it storing files somewhere, let me check.10:01
sledges(also let me get this straight: hwclock = rtc; system_time = software ?)10:01
situalterego: I missed the alarm when phone was off.10:02
situalterego: When I boot, it's shown in missed alarms in notifications.10:02
*** lpotter has quit IRC10:03
*** lpotter has joined #sailfishos-porters10:03
alteregositu: seriously lame :/10:04
sledgesdmt: no network in public transport? just try not to reboot the phone whilst commuting ;)10:05
sledges(especially going underground)10:05
situOops I can't run adb as root in production builds.10:05
dmtsledges: undergrown true, but flight mode is bigger issue10:05
sledgesmost people switch off phone for the night, it fetches time in the morning, ..., profit10:05
situvgrade: Can you strace time_daemon and check what files it is reading ?10:06
sledgessitu: worth deleting those files for good measure :))10:06
dmtand say you are not on your luck abroad with no roaming for the available nets..10:06
sledgesdmt: emergency calls = time info10:06
sledgesdmt: otherwise nexus5 would be recalled ;)10:07
dmtsledges: ok this is workable then i guess10:07
dmtsledges: another hypothetical: out in the fields/mountains with no net for a good while10:08
dmtsledges: would time halt?10:08
sledgesif phone is powered off - ofc10:08
dmtsledges: or can sw alone keep it going10:08
sledgesif on - no noticeable problems10:08
dmti'll reboot without sim to try10:09
sledgesdmt: reboot = no sw = time halt ;)10:09
*** vakkov has joined #sailfishos-porters10:09
sledgesbut reboot's fast, so you won't notice a thing10:09
sledgessitu: if time_daemon does the fs storing for us, why can't we keep it?10:10
dmtreboot shall start from 1970 now and move from there, shan't it?10:10
*** vakkov has quit IRC10:10
sledgesdmt: time_daemon enabled?10:11
*** vakkov has joined #sailfishos-porters10:11
situsledges: because time_daemon will try to lock /dev/rtc0 which is locked by dsme's iphb module already.10:13
dmtseems to work, time ticks with wifi disabled and no sim10:13
situsledges: and it will keep failing as it can't lock /dev/rtc010:13
dmtstarts at 1.feb.1970 and goes on10:13
situsledges: if we disable iphb module then time_daemon gets lock over /dev/rtc0 and shows correct time.10:14
*** lpotter has quit IRC10:14
sledgessitu: ok thanks for recap10:14
situsledges: but in some cases it has failed. I don't know why.10:14
*** lpotter has joined #sailfishos-porters10:14
dmttimed-qt5 is running only10:15
vgradetime_demon strace,
sledgesdmt: flash cm and test there if you like, too hairy state in sailfish10:16
sledgesso plan: disable time_daemon, investigate where dsme knows correct time on startup and fails to set it; if that's false alarm (pun): patch dsme to save time to fs on hammerhead10:17
dmtsledges: situ said s/he tried this and it was also in the 70's10:17
junnuvisledges: I got audio working on grouper.. but do you have any idea why volume control is not working? It claims that volume is muted even it's actually at 100% full time10:18
sledgesalterego: ^10:18
situ /data/system/time/ats_110:18
sledgessitu: nice10:18
sledgesvgrade: rm? ;)10:18
situ /data/system/time/ats_210:18
situI don't see any other relevant files.10:19
sledgesvgrade: taking that back, unless you got iphb deleted and that works10:19
sledgesjunnuvi: good work! :)10:20
vgradesledges: I stopped dsme.service10:20
junnuvisledges: thanks :)10:20
alteregojunnuvi: probably an issue with pa configs.10:22
junnuvialterego: probably, any ideas where to start investigate? with pacmd I can adjust volume levels10:22
alteregojunnuvi: I didn't get volume working because we didn't have buttons working at that time.10:22
alteregojunnuvi: I'm guessing you need to make sure that the pulse configs state the correct names for the audio output devices.10:23
alteregoThis is where jusa_ would probably be helpful for you :)10:24
junnuviok, I will try to compare configs between real jolla device and nexus, maybe that will help10:26
*** furikku has quit IRC10:26
sledgesjunnuvi: alterego:
sledgesyour key # might differ10:27
*** furikku has joined #sailfishos-porters10:27
sledgesalthough then it wouldn't even say "Mute"10:27
junnuviI don't think bbecause I get volume level indicator when pressing volume buttons10:28
sledgesthat ;)10:28
sledgesvgrade: if we want to double-check: completely disable time_daemon; ensure ats_* contain non-1970 date; reboot; observe date being 1970; stop dsme.service; launch time_daemon manually; see if date changes. remove ats_* files and repeat the process10:30
sledges(hardmasking dsme is not an option, as i don't think it will boot afterwards :)10:31
jusa_junnuvi: hi.. you mean volume is constantly 100% no matter what is set as the volume?10:31
*** vakkov has quit IRC10:32
*** dmt has quit IRC10:32
vgradesledges: I only masked to be able to provide the strace to situ10:32
jusa_if the volume seems to change but there is no effect to the actual volume level, I'm suspecting that either 1) hw volume control of droid-sink is broken (highly possible, I haven't tested the volume control with device that would have hw volume in "normal" playback), or 2) API incorrectly reports it has hw volume when in fact it doesn't10:34
junnuvijusa_: hmm, it looks like I don't have xpolicy.conf on nexus7, could that be reason?10:38
junnuvijusa_: yes, volume buttons does not change volume levels, from pacmd I can change volume levels10:39
jusa_junnuvi: ah, nexus 7... actually, n7 has totally different adaptation from everything else10:39
jusa_junnuvi: if you want it to behave like the rest of the devices, you'll need to re-do the adaptation10:40
jusa_or then merge those to somehow..10:40
situvgrade: did you try removing those files ?10:41
situI tried removing those files on cm, but time is correct on next boot.10:43
sledgessitu: good point on cm! and...  odddd10:44
situioctl(1, SNDCTL_TMR_TIMEBASE or SNDRV_TIMER_IOCTL_NEXT_DEVICE or TCGETS, {B38400 opost isig icanon echo ...}) = 010:44
situwhat would that do ?10:45
junnuvijusa_: actually it's already working pretty fine with audio. Only major issue is non-working volume controls.10:45
jusa_junnuvi: yep.. the reason why the adaptation is different is n7 doesn't have voice call, so it's fairly simple to just use ALSA directly10:46
jusa_you'll need to run module-stream-restore-nemo, module-meego-mainvolume, at least, to get working volume control10:47
alinsitu: I think that file are created on the fly... I do not have them... the ones under ../time/10:47
alinsledges: why do we use dsme? is really needed... seems ubuntu people do not10:48
sledgesalin: dsme is one of the core parts for sailfish, it's responsible not only for time10:50
alinsledges: yes I know... looked through the code...10:51
*** dmt has joined #sailfishos-porters10:51
alinsledges: anyhow10:51
sledgesDevice State Management Entity which handles various pieces of hardware.10:51
alinsledges: great...10:53
sledgesalin: you're in an awesome mood I see :D10:55
alinsledges: why? just wanted to know some things...10:59
alinsledges: now I am testing the cm snapsot and our latest image11:00
sledgessitu: with cm, could you take out sim and disable wlan, power down for 5 mins and check the time?11:01
sledgesalin: thanks for spotting+testing the snapshot, what a relief11:01
situsledges: will check that.11:03
*** arcean has joined #sailfishos-porters11:28
vgradeok, 13:30 < sledges> vgrade: if we want to double-check: completely disable time_daemon; ensure ats_* contain non-1970 date; reboot; observe  date being 1970; stop dsme.service; launch time_daemon manually; see if date changes. remove ats_* files and repeat the  process11:32
vgradeok so first part on reboot time is 1970 after stopping and launching time_daemon_manually date changes to 201411:33
sledgesvgrade: no sim nor wlan?11:34
vgradesledges: no wlan, mobile web switched off, sim installed11:35
sledgesok, i wonder what will happen with files deleted (from situ's cm tests - they are not those..)11:36
sledgesand chuck sim out once you're at it :)11:36
sledgesif it still sets, means some other file or actual talk to hwclock via qc closed blobs11:36
vgradesledges: will need to do later, dayjob calls11:39
* sledges nods11:42
vgradefor the record,11:43
vgrade[root@Jolla time]# hexdump ats_111:43
vgrade0000000 50d5 4470 0142 000011:43
*** vakkov has joined #sailfishos-porters11:51
alinvgrade: sledges situ dmt so... with m9... todays clean build using sf-mer... phone boots11:52
alinwe have no systemd workaround11:52
vgradealin: sf-mer git has a couple of workarounds, have you edited them out11:54
alinvgrade: the systemd are commented out11:54
vgradereboot reliably?11:55
alinvgrade: by the way first boot always start with time 1 january 201311:55
alinvgrade: yes... after install I had no issues11:56
alinvgrade: practically wipe/factory reset... sideload cm11m9 sideload sf... reboot11:56
alingot the tutorial screen11:56
*** RobJanc has joined #sailfishos-porters11:57
alinsledges: this is for you... the tutorial screen (by default there is no internet activated or wifi.. but still is asking me to sign in in the jolla account)11:59
situalin: Yes, timed-qt5 is programmed to set that date on first boot.12:00
sledgesalin: it's called startup-wizard, tutorial comes later. does it offer connection dialog?12:01
alinsledges: no12:01
sledgesthat's the bug12:04
sledgesthere's option to skip account creation/signin12:04
alinsledges: yes..12:05
vgradealin: will the source repos be tagged with M912:07
alinvgrade: cyanogen ones?12:09
alinvgrade: I do not know12:09
alinvgrade: I may check...12:09
alinvgrade: you think of clonning only that tag?12:09
vgradejust looked an they are not12:10
vgradeat least hammerhead device12:10
vgradewas wondering if we should matcj the source code to the /system bins.12:11
sledgesvgrade: which mainly means rebase ;)12:12
vgradehow does CM know what repos are integrated into M9 for example12:13
vgradewell which commit12:13
vgradeI mean manifest says which repo12:14
vgrademaybe their build system tells them12:15
vgradeafer a build what commit ids are for each repo12:15
vgradewas just wondering how they could reproduce a buld that's all12:16
vgradesledges: ok ats files moved, rebooted, date is 1970, stop dsme.service, hwclock returns 1970, no new ats files created, run time_daemon , date is still 197012:39
vgradesledges: situ so time is being read from those files12:39
vgradesledges: ok, stop time_daemon, renamed files back, run time_deamon, date 201412:41
vgradeafk for a while12:41
spiiroinlooks like dsme indeed forces system time to unmodifiable rtc time; I'll start figuring out where to change what without breaking other stuff12:53
alinAug 06 13:55:42 abbaton pulseaudio[8195]: can't set volume limit: don't know group 'btnotify'12:57
alinspiiroin: what do you mean unmodifiable rtc time?12:58
spiirointhat we can't set neither rtc time or alarm12:58
spiiroindsme has some logic for not going backwards in time during bootup, but it all assumes functioning rtc12:59
spiiroinnow we need to make it take into account situations where rtc has no chance of having correct time too12:59
spiiroini.e. if something else manages to get system time in sync with reality, dsme should not rewind back to seventies13:00
*** vrutkovs has quit IRC13:01
*** alin_ has joined #sailfishos-porters13:04
*** alin has quit IRC13:05
spiiroindone: a) if rtc is not writable, dsme will not sync system time to it b) if system time on dsme startup < last known system time, dsme moves system time to last known13:15
vgradespiiroin: is lask known already stored somewhere13:17
vgradespiiroin: or is the assumption that it in a working RTC13:17
spiiroinyes, but as said it was used for detecting rtc resets; if rtc does not work as expected, system time is lost as collateral damage13:18
situvgrade: did you figure out how dsme knows correct time to set to rtc ?13:19
vgradesitu: assuming from network (wifi, mobileweb) but no I've not looked any further than the testing I've done with time_daemon13:20
situvgrade: on my device both wifi and mobile web were off.13:21
vgradesitu: and13:21
situvgrade: and IIRC it was trying to set correct time in rtc and failed.13:22
vgradesitu: any sim card installed13:24
situvgrade: yes.13:24
spiiroinvgrade: the last-known-time is mtime of /var/tmp/saved-time13:25
vgraderight ok, that says -rw-r--r-- 1 root root    0 1970-09-26 13:41 saved-time13:27
spiiroinmine says (after reboot) -rw-r--r-- 1 root root 0 2014-08-06 16:26 /var/tmp/saved-time13:30
vgradeafter renabling wifi and resarting dsme so does mine13:32
vgradewell nearly -rw-r--r-- 1 root root    0 2014-08-06 14:31 saved-time13:33
vgradeassuming tz plays a part here13:33
spiiroinyes, but the problem is that it gets ignored without working rtc13:33
vgradespiiroin: yes, understood13:33
vgradeso I think we just answered situ's question about where it gets good time from13:34
situspiiroin: so I come back to my original suggestion to call settimeofday if setting rtc fails.13:35
spiiroinsitu: yes, I did a bit too fast code read initially13:35
spiiroinhacked fix is here if someone wants to give it a go13:35
*** iTune has joined #sailfishos-porters13:36
spiiroinmore specifically this commit13:36
spiiroin... but of course that fixes only the rewind issue; we still need to get correct time from somewhere13:37
situspiiroin: It looks like my network operator supports network time but the tests to get network time were failing on my device.13:40
situspiiroin: Can you download ofono-tests package and run the test to get network time on your device ?13:41
spiiroinsitu: I do not have a sim in my nexus 5 atm ...13:41
situspiiroin: Try on jolla device, it was failing for vgrade there.13:41
spiiroinsitu: that I can try13:42
situspiiroin: does your network operator support getting network time ?13:45
spiiroin... updating & installing still13:45
alin_spiiroin: dsme gets the correct time from somewhere13:46
spiiroinalin_: dsme gets it either from the saved-time file or rtc; otherwise it just notes that timed has changed system time13:47
situI think I will just leave a rant against qcom guys on that mail thread :P13:48
vgradesitu: at least ask why he says "don't do that"13:49
situvgrade: I think he was assuming we were on android.13:54
situI had left a reply for him, haven't heard back yet.13:55
vgradesitu: you made it clear and even gave him a link to dsme13:55
situvgrade: Right.13:55
spiiroinhmmm... does the rtc reset all the way back to some fixed point, or is it just not adjustable?13:57
spiiroinif it is the latter, then we might store rct-system_time diff at shutdown and apply it on bootup13:58
situspiiroin: I think it's not adjustable.13:58
spiiroinsitu: but it still always goes forwards?13:58
situspiiroin: Yes, I think so ( I don't have sailfish running on my device right now).13:59
spiiroinsitu: ok, I'll check once I get the ofono tests installed and done13:59
situspiiroin: Great, thanks.14:00
vgradespiiroin: yes hwclock returns a time moving fowwards14:01
spiiroinvgrade: over reboot too?14:01
vgradeyes hwclock time increments forward even after a reboot14:06
spiiroinvgrade: great, then we just need to store how much the rtc is off from system time; that should get us valid system time over reboots too14:07
vgradebefore reboot14:07
vgradeSat 26 Sep 1970 03:05:45 PM BST  -1.709299 seconds14:07
vgrade[root@Jolla nemo]# hwclock14:07
vgradeSat 26 Sep 1970 03:05:49 PM BST  -1.420551 seconds14:07
vgradeafter reboot14:07
situspiiroin: but you can't trust that for every device and should not store diff.14:07
vgradeSat 26 Sep 1970 03:10:43 PM BST  -1.035796 seconds14:07
vgrade[root@Jolla nemo]# hwclock14:07
vgradeSat 26 Sep 1970 03:11:11 PM BST  -1.884637 seconds14:07
spiiroinsitu: I mean locally; for that device14:07
situspiiroin: but you can't trust that rtc for every device will go forward.14:08
situspiiroin: or you can?14:08
spiiroinit is better than nothing; and only to be used if rtc is not writable14:09
situspiiroin: hmm ok.14:09
spiirointhe rtc will probably deviate from reality anyway without support from cellular or ntp14:09
*** alin_ has quit IRC14:09
spiirointhe "store error" should not be any worse than being able to set the rtc14:09
situspiiroin: ok, let's fix it.14:10
spiiroinsitu: what is the test binary from ofono-test you wanted me to run?14:10
situspiiroin: can't recall it's name, it's the which tests getting network time.14:11
spiiroinpresumably /usr/lib/ofono/test/get-network-time?14:11
situspiiroin: Yes14:11
spiiroinseems to be working - at least I got a timestamp { ..., UTC = 1407333960, ... }14:12
situspiiroin: ok, it returns empty list for me on Nexus 5.14:13
vgradereturns blank for me14:13
vgrade[ /ril_0/operator/23430 ] Status = current Technologies = umts  MobileNetworkCode = 30 Name = EE MobileCountryCode = 23414:13
vgradefrom this operator14:13
vgradeand also O2 in the Ok returned {}14:14
vgradewell we can't fix that14:15
situspiiroin: I am sure my operator support getting network time as my Nokia N8 is able to fetch network time correctly from same operator.14:15
situCould be an ofono bug.14:20
*** alin_ has joined #sailfishos-porters14:26
*** alin__ has joined #sailfishos-porters14:31
*** alin_ has quit IRC14:31
*** vrutkovs_ has joined #sailfishos-porters14:34
alin__spiiroin: yes it goes forward14:39
alin__from some time in 197014:39
alin__begining of it14:39
*** stephg has joined #sailfishos-porters14:50
*** gogeta_ has quit IRC14:51
*** phdeswer has quit IRC14:52
*** RobJanc has quit IRC14:52
dmtis there a way to reboot into CM withouth flashing anything?14:56
alin__spiiroin: Aug 29 21:28:48 abbaton DSME[684]: IPHB: set to 2014-08-06 14:38:08 GMT14:57
alin__that is correct and it fails14:58
alin__Aug 29 21:28:48 abbaton DSME[684]: IPHB: rtc at 1970-08-29 20:28:48 GMT14:58
alin__Aug 29 21:28:48 abbaton DSME[684]: IPHB: set to 2014-08-06 14:38:08 GMT14:58
alin__Aug 29 21:28:48 abbaton DSME[684]: IPHB: /dev/rtc0: RTC_SET_TIME: Invalid argument14:58
alin__so from somewhere steals the correct time14:58
vgradedmt: fastboot boot boot.img14:59
vgradealin__: ls -l /var/lib/dsme/time-state14:59
alin__vgrade: ls: cannot access /var/lib/dsme/time-state: No such file or directory15:00
*** carepack has quit IRC15:01
alin__vgrade: -rw-r--r-- 1 root root 4 2014-08-06 15:38 /var/lib/dsme/timed_state15:01
alin__vgrade: and inside 0 015:02
vgradealin__: so that matches what its trying to set15:02
vgradealin__: the datestamp of the file15:02
vgradetaking the tz into account15:02
vgrade\o stephg15:03
alin__vgrade: yap15:04
dmtused droid/out/target/product/hammerhead/boot.img and it went in recovery15:07
*** JohnyZ has quit IRC15:08
dmtshould have downloaded something from cm15:08
*** stephg has quit IRC15:09
*** JohnyZ has joined #sailfishos-porters15:09
vgradedmt: can you try adding -c "selinux=1"15:10
dmtwith the same image?15:10
vgradeif that does not work, then try a stock boot.img15:12
spiiroinalin__: the minval dsme uses is: max(saved_time, system_time, release_time, ...)15:12
dmtrecovery again15:13
spiiroinyou can see them listed if running with -v6 verbosity15:13
spiiroin /usr/sbin/dsme -p /usr/lib/dsme/ -l stderr -v615:13
sledgesdmt: use boot from cm.zip15:13
vgradesledges: what breaks cm boot from our boot.omg15:14
sledgeslack of tests15:15
vgradesledges: ?15:15
dmtworked this time15:16
sledgesi never thought hybris-boot generates cm's boot.img15:16
vgradeyou mean its not been a use case15:16
sledgesbut if it does, it boots already mer-modified kernel - maybe that?15:16
vgradeyes it does create one , as you say one ofthe kernel option changes15:17
sledgesand initrd15:19
sledgesprobably it's just a hybris-recovery evil twin15:19
sledgeshadk advices to rebuild cm tree separately, if run into problems15:19
vgradetwin should boot to sf then15:20
vgradeah and it would if selinux=0 was set15:21
vgradewith selinux=1 sf reboots to recovery15:21
dmtmakes sense15:23
dmtusing android is such a pain now after not using it for a week15:24
dmtdoes anyone know how to convince the terminal (in cm) not to vibrate on touch?15:26
dmti did set the things right in the general settings..15:26
*** gogeta_ has joined #sailfishos-porters15:39
sledgesvgrade: jusa_: just tested on n5: works fine with revert:
sledgesbooted with headset plugged in , did several tests, all ok15:46
vgradesledges: great. I've just tested spiiroin's new dsme15:47
sledgesgreat, how is it?15:47
vgradehave not been tansported into the 70's yet15:48
vgradebut of still have the delay until time gets updated from network15:53
*** mailyaseen has quit IRC15:54
*** arcean has quit IRC15:55
vgradeseems mine and situs operators don't send network time, or there is somthing add going on15:55
vgradeso time gets updated via mobile or wifi web as an when they've connected15:56
sledgescross check with cm?16:00
spiiroinI got the saved-rtc-delta working to the point that I can: stop dsme; move date to bogus value; start dsme -> correct system time gets restored16:01
vgradespiiroin: cool16:02
*** zZz0n is now known as zon16:05
spiiroin1 sec off after reboot, but that is kind of expected when fiddling with multiple 1 sec accuracy values16:05
*** alin__ has quit IRC16:06
*** alin has joined #sailfishos-porters16:09
spiiroinstop dsme; set bogus date; reboot -> works too16:09
vgradespiiroin: much better than 44 years out16:09
*** vrutkovs_ has quit IRC16:38
locusfhmm my device porting days are over I guess16:52
locusfsu: /bin/bash: Resource temporarily unavailable16:52
locusfthis is coming every time I enter Mer SDK16:52
locusfas it did on my earlier rootfs too16:52
*** alin has quit IRC16:53
vgradesitu: do you stil have CM installed?17:04
dmtlocusf: did you follow the installation outlined in the pdf and wiki?17:08
locusfdmt: yes I did17:08
locusfdmt: this is something my host does every now and then17:08
locusfreason still unknown17:08
dmtlocusf: archaic system/software?17:09
locusfdmt: ubuntu 14.0417:10
vgradelocusf: spin up a new HADK environment when required with dmt's great scripts17:14
situvgrade: Yes17:14
locusfvgrade: won't work as I have tried even reinstalling17:15
vgradesitu: what is the delat in setting correct time on cm after boot17:16
locusfvgrade: I tried to debug this with lbt as well17:16
situvgrade: I didn't understand the question.17:17
vgradeif you leave the device off for an hour how long does it take to show the correct time?17:17
vgradeif you have wifi and mobile web switched off17:18
situvgrade: Need to test this.17:18
vgradeso loooking to see if it picks it up from the cell network17:18
vgradeor if it does something similar to spiiroin is doing and using the delta from the last RTC17:19
situvgrade: I think it uses delta, I can just remove the sim and test it.17:20
spiiroinsitu: vgrade: I pushed the rtc delta stuff to the wip branch in case you want to test it17:20
vgradespiiroin: I will, cheers17:20
situspiiroin: Thanks, will test it.17:20
spiiroinI'll go offline for some time, will check in a bit later17:21
vgradelocusf: sounds odd17:21
vgradesome interaction with somthing else installed , sailfish SDK?17:22
*** spiiroin has quit IRC17:26
locusfI do have it installed yes17:30
*** Nokius has quit IRC17:34
*** dmt__ has joined #sailfishos-porters17:37
*** zarubond has joined #sailfishos-porters17:39
*** dmt has quit IRC17:44
*** spiiroin has joined #sailfishos-porters17:53
*** arcean has joined #sailfishos-porters17:58
vgradespiiroin: looks good18:03
vgradeneeds wifi or mobile web enabled to automoticaly update a manually set time wrong time but thats issue with cell time18:05
spiiroinvgrade: good to know, still need to make sure that it will not break if rtc happens to be settable ;-)18:05
vgradespiiroin: only have my jolla here, maybe we can get sledges to test on one of his HADK devices18:06
jusa_sledges: ok, great18:10
*** iTune has quit IRC18:13
vgradejusa_: thanks for your help18:16
*** piggz has quit IRC18:26
*** Sequenced_ has joined #sailfishos-porters18:29
*** piggz has joined #sailfishos-porters18:32
*** phdeswer has joined #sailfishos-porters18:40
*** furikku has quit IRC19:04
*** zarubond has quit IRC19:06
*** alin has joined #sailfishos-porters19:26
*** alin has joined #sailfishos-porters19:26
alinsledges: hi19:26
alinsitu: vgrade dmt__ is ok with you to update the wiki and suggest m9 to be used?19:27
vgradeevening alin19:27
alinvgrade: evening19:27
vgradealin: sure, I saw you tested19:27
alinvakkov: hi did you hear back from beidl?19:27
alinvgrade: yap I did.. would be nice if we will get one more test at least19:28
vgradealin: I've prepared OBS,
vgradethe one failure will go once the pulse commit goes upstream19:28
alinvgrade: cool... when time comes we can make it directly the official one19:29
vgradegtg dinner19:29
alinvgrade: ok...19:29
vakkovalin: haven't heard of him. have i missed something?19:30
alinvakkov: beidl thought you were porting maguro together19:33
alinvakkov: he is in army now19:33
vakkovalin: yes, i know ;D and yes we were :d19:34
vakkovhaven't heard of him. I thought there is something new that i don't know  :D19:34
alinvakkov: i see19:35
alinhaven;t heard from him19:35
vgradealin: we can hope19:41
alinvgrade: what is the problem with upstreaming the pulse audio?19:43
alinby the way... what happened with the time in the end?19:44
vgrade spiiroin has made some modifications to dsme19:47
vgradealin: pulse audio should be good now19:48
alinvgrade: cool19:49
vgradealin: still an issue when wifi and mobile web are not enabled as time is not being brought from the carriers situ and I use19:49
alinvgrade: I shall delete my branch on time... I see he did not commit anything back19:49
alinvgrade: but this  is strange by default none are enbaled when one starts the phone19:50
alini still do not understand where dsme gets the correct time19:50
alinthe one is trying to set and fails19:51
vgradealin: you mean on first boot, you are asked what the time is.  That will be stored now with the changes spiiroin has made today19:51
alinthat is nice19:52
alinAug 29 21:28:48 abbaton DSME[684]: IPHB: set to 2014-08-06 14:38:08 GMT19:52
alinbut this one... is taken from somewhere and is right19:52
vgradealin: what conditions?19:53
alinvgrade: I mean not mobile internet and no wifi19:55
vgradeand not set by you19:55
vgradeand not set via carrier update19:55
vgradeand not set by time_daemon19:56
alinvgrade: was probably set by me19:58
alintime _daemon i disable after install19:59
alinautomatic update disabled19:59
*** piggz has quit IRC20:15
*** shallow is now known as potlesk20:20
alinsledges: what about gps you said something about it and the word success20:24
vgradealin: last I remember it needed a bit of mw glue20:26
*** piggz has joined #sailfishos-porters20:27
vgradegoing for an early one tonight, see you tomorrow20:28
*** piggz has quit IRC20:39
alinvgrade: night20:39
*** piggz has joined #sailfishos-porters20:39
*** Xtal_ has joined #sailfishos-porters20:42
*** zon is now known as zZz0n20:57
*** Xtal_ has quit IRC21:05
*** lbt has quit IRC21:14
*** lbt has joined #sailfishos-porters21:14
*** Xtal_ has joined #sailfishos-porters21:23
*** Xtal_ has quit IRC21:27
*** Xtal_ has joined #sailfishos-porters21:27
*** Xtal_ has quit IRC21:34
spiiroinsitu: vgrade: wip branch updated again; should now work in device that have settable rtc too21:45
*** gogeta_ has quit IRC21:57
*** netchip has quit IRC22:22
*** chrisi has quit IRC23:03
*** cxl000 has quit IRC23:06
alinspiiroin: ok... I will build it and test it23:19
*** chrisi has joined #sailfishos-porters23:24
sledgesalin: i've passed the GPS bug internally, and sailors from australia replied - that's why i said [the bugzilla infra] is working :p Now I'm waiting for details on how much effort is required to fix this bug, and can community help (bit harder here, as it revolves around some sailfish components)23:32
sledgesnightn :)23:33
alinsledges: night23:35
alinsome strange error on obs23:35
sledgesalin: set _service url to
sledgesand add <param name="revision">HEAD</param>23:37
sledges(you have it set without the ".git" bit)23:37
alinsledges: yap you are right... I hshall go to sleap...23:38
sledgessame here23:38
sledgessleep 2880023:38
alinsledges: so long? I will kill my two neurons23:41
alinmore like 1200023:41
alin[   53s] vs /home/abuild/rpmbuild/SOURCES/dsme.spec=0.62.13+master.928.253.g4f8d29623:44
alin[   53s] Conflicting package versions23:44
alin[   53s] error: Bad exit status from /var/tmp/rpm-tmp.4xhjS4 (%build)23:44

Generated by 2.17.1 by Marius Gedminas - find it at!