Friday, 2014-10-31

*** zalan has joined #nemomobile00:14
*** wazd has quit IRC00:23
*** plfiorini has quit IRC00:25
*** ajalkane has quit IRC00:54
*** xhaakon has quit IRC00:58
*** ajalkane has joined #nemomobile01:00
*** xhaakon has joined #nemomobile01:04
*** phaeron has quit IRC01:22
*** DrCode has quit IRC01:46
*** DrCode has joined #nemomobile01:49
*** vakkov has quit IRC02:09
*** DrCode has quit IRC02:22
*** zalan has quit IRC02:23
*** zalan has joined #nemomobile02:26
*** zalan has joined #nemomobile02:26
*** vakkov has joined #nemomobile02:47
*** Jef91 has joined #nemomobile02:51
*** Morpog_PC__ has joined #nemomobile03:44
*** KaIRC has quit IRC03:56
*** flash1 has joined #nemomobile04:09
*** flash1 has quit IRC04:17
*** himamura has joined #nemomobile04:37
*** Behold has quit IRC04:44
*** lpotter has quit IRC04:52
*** lpotter has joined #nemomobile04:53
*** martyone_ has joined #nemomobile04:58
*** himamura has quit IRC05:00
*** Oksana has quit IRC05:24
*** spiiroin has quit IRC05:47
*** VDVsx_ has quit IRC05:50
*** furikku has joined #nemomobile05:59
*** kostaja has quit IRC06:00
*** VDVsx_ has joined #nemomobile06:01
*** olesalscheider has quit IRC06:14
*** filippz has joined #nemomobile06:17
*** Jef91 has quit IRC06:20
*** olesalscheider has joined #nemomobile06:20
*** WWDrakey has joined #nemomobile06:22
*** spiiroin has joined #nemomobile06:29
*** filippz has quit IRC06:29
*** spiiroin has quit IRC06:33
*** kostaja has joined #nemomobile06:37
*** Pat_o has joined #nemomobile06:38
*** filippz has joined #nemomobile06:42
*** spiiroin has joined #nemomobile06:45
*** messerting has quit IRC06:50
*** Sfiet_Konstantin has joined #nemomobile06:51
*** hedayat has joined #nemomobile06:55
*** DrCode has joined #nemomobile06:57
hedayatfilippz: Before going to some good news, I have a question. Is there any reason for which the N9 kernel shut down the phone automatically?06:58
filippzhedayat: Stskeeps has a point - if DSME isn't running or started early enough07:00
hedayatStskeeps: yeah, something like that? Is the watchdog always waiting for some kind of signal or it should be enabled somehow?07:00
filippzalso low battery is one to watch for :)07:00
hedayatfilippz: No, something other than low battery. hmmmm so the watchdog is enabled automatically?07:01
Stskeepsyou'll want to kick watchdogs in preinit perhaps07:01
filippzhedayat: I think that DSME gives them a "heartbeat" signal of some sort07:02
spiiroinbtw, I'm working on fixes for dsme that should allow it to start earlier07:03
hedayatI've seen preinits do start watchdog kicker, but they stop it right before booting the OS. So I thought if they don't start it, it would be disabled. :P07:03
spiiroinby not needing usb moded to be up and running before dsme07:03
hedayatOK, let me describe the real problem to see if it is related to watchdog/dsme.07:03
hedayatAs you probably know, I boot Nemo using a loop mounted filesystem images, which is done through a small initramfs07:04
hedayatIt just mounts the image and passes the control to systemd07:04
hedayatsometimes, phone turns off soon after starting systemd, so that systemd creates a log file but it is empty.07:05
hedayatShould I start a watchdog kicker in my initramfs? And, should I stop it (like what preinits do) before switch_root?07:06
*** nodevel has joined #nemomobile07:06
filippzIf you do get a shutdown at boot, my bet would be the watchdogs. Try to disable mce service and try booting a couple of times07:08
filippzwe had some problems getting it to boot with mce, but as of recently that problems went away07:09
hedayatfilippz: Yes, but it is probably before any userland services run by systemd; so probably mce is not running when the phone shuts down07:09
spiiroinby guess: mce is just one the things that gets started before dsme07:10
spiiroinand anything that makes dsme start earlier -> can help07:10
filippzmaybe systemd doesn't dump journal to flash before shutdown07:10
spiiroinif it is hwwd reset -> no flushing can happen07:10
filippzhedayat: try the mce thing and see what happens - there is no harm in that07:11
*** cxl000 has joined #nemomobile07:14
hedayatfilippz: spiiroin Stskeeps thanks, I'll go through the suggestions to see what happens. I will start a watchdog kicker, and see if disabling mce does help.07:15
spiiroinhedayat: if you can arrange the wd kicker to start early and get stopped when dsme starts up -> might be what is needed anyway07:16
hedayatOK, now another question: has anybody used the latest ofono package (used in latest nemo image) on a real hardware?07:16
*** SfietKonstantin has joined #nemomobile07:17
hedayatspiiroin: I'm going to try this first. What if I don't stop it? Is it required?07:17
Stskeepsfwiw the ofono that's there might be a bit fine tuned towards android ril07:17
spiiroinhedayat: no harm for debugging phase07:17
stephgit's friiiiday!
*** Sfiet_Konstantin has quit IRC07:17
spiiroin(dsme just wont be doing the kicking)07:17
Stskeepsif it causes trouble, check out an earlier version without android ril07:17
hedayatStskeeps: ah... probably that's it. Because, ril is the only plugin which must be disabled to get ofono correctly recognize N9 modem,07:18
hedayatStskeeps: and even with that, it doesn't work at all, while I had much more success with ofono-1.12 in the latest official Nemo image for N907:19
filippzhedayat: by success - you mean modem is working (pin entry, sms...)?07:20
Stskeepshedayat: hrmm07:20
hedayatAnd, I don't remember who said that N9/N900 modem is isimodem, he was correct! both isimodem and n900 plugins should be enabled. Apparently, each one is responsible for a part of functionality.07:21
hedayatfilippz: I have not yet tried to do call/sms, but pin entry works, and it connects to network.07:21
hedayatfilippz: Now I'm struggling to boot it again for further tests :)07:21
filippzhedayat: great news - CMT was broken for years in 3.5 kernel!07:22
filippzhedayat: are you sure that battery isn't low (sorry for mentioning it again)?07:23
hedayatfilippz: also, please include these packages in your ks: statefs-provider-ofono,  statefs-provider-qt5, statefs-loader-qt5 (these 2 packages are required by statefs-provider-ofono, so that one is probably sufficient to pull in the other 2)07:23
hedayatfilippz: Yes :)07:24
hedayatfilippz: But I'll make sure again if it doesn't boot with a watchdog kicker07:24
hedayatit seems to be booting. Why isn't there any startup splash screens? (or why doesn't it appear)? I've seen it on shutdown07:25
hedayatOK, booted successfully07:26
filippzhedayat: will do, but locusf is the one who does "public" builds (maybe we should public that ks somewhere so we can work on it together)07:26
hedayatStskeeps: I also say that in Sailfish image, pin entry works but then Sailfish says that there are no simcards. It is probably the same problem I see, because latest ofono terminated when I run enter-pin.07:27
hedayatfilippz: OK, I'll tell locusf too! :P07:27
hedayatlocusf: glacier PIN entry also works. However, glacier ui fonts are too large for N9 screen.07:28
hedayatOK, I just entered my PIN in galcier UI, and it shows the operator name (truncated) and generation :)07:29
locusfexcellent :)07:30
filippzhedayat: \o/07:30
filippzlocusf: what about that boot screen - do we need plymouth or something else?07:31
hedayatlocusf: I just wonder why it doesn't show Signal strength, while the values are reported in statefs07:32
locusffilippz: plymouth is in nemo:devel:mw which is dead now, I'm not sure it works07:33
locusfhedayat: strange, can you check journal for errors?07:33
hedayatlocusf: filippz: I see the shutdown screen, isn't it being shown by plymouth too?07:34
filippzlocusf: do you know how it's done on jolla?07:34
locusfhedayat: no, thats lipstick07:34
hedayatlocusf: I'm looking through logs...07:34
locusffilippz: no07:34
hedayatfilippz: hmmm at list the SFOF image for N9 doesn't show anything on boot either. But it also shows the shutdown logo07:35
*** vakkov has quit IRC07:36
filippzhedayat: then we're missing some setting - I'll look at this later today or tomorrow07:37
hedayatfilippz: thanks. But it's not a priority, I just wanted to know the situation. There are probably more interesting areas to work on for now ;)07:38
hedayatfilippz: And about CMT, I'm certainly not the one who did the real job.07:39
filippzhedayat: ty, I was just fixing/completing the hard work of others07:40
hedayatOK, me too!07:41
hedayatlocusf: at least there are no visible errors among the huge amount of logs. but there are no complains about missing statefs variables which I saw before07:42
hedayatlocusf: is the signal bars being updated regularly? (just guessed maybe it is zero because I needed to enter pin when booted).07:43
hedayatlocusf: file:///usr/share/lipstick-glacier-home-qt5/qml/compositor/WindowWrapperBase.qml:30: TypeError: Cannot read property of null07:45
hedayatlocusf: a number of errors related to glacier:
*** vakkov has joined #nemomobile07:48
locusfhedayat: I'll take a look at then later, gotta go now07:52
hedayatlocusf: OK, thanks07:52
*** vakkov has quit IRC07:55
hedayatfilippz: Sent an sms (almost) successfully. (it was UTF-8 encoded, but what was delivered was junk. But anyway, I'm using ofono test-sms script).07:55
*** veskuh has joined #nemomobile07:56
filippzhedayat: we'll - I'm calling this success no matter the junk :)07:56
hedayatme too! :D07:57
filippzat some point this was looking like SF - faenil informed me that  a NON-kernel-hacker ... spent more than a month trying to get telephony working on sailfish on N9 using the old kernel07:58
filippzand you are texting now!07:59
Aardfilippz: the _old_ kernel? what's there to get working for telephony? :p08:00
hedayatfilippz: And I dialed a number. But I wonder if I can test talking. Considering that sound doesn't work on the image, it should be impossible. But since you fixed that too, we should probably be able to call too08:00
hedayatfilippz: hmm... I'm not sure. Actually the telephony works on the old Nemo image on the website which uses the 2.6.32.x kernel.08:01
*** SfietKonstantin has quit IRC08:02
hedayatNow, maybe the major problem with telephony in Nemo is that there is no working UI for it. :P08:03
filippzAard, hedayat: parts of old kernel were used in the new one (
filippzbut this was the key to make it work:
hedayatfilippz: hmm never looked at the contents before. But why irq_to_gpio returns a wrong number? (If can be explained briefly) :P08:06
hedayatah, I see it was defined inside the hsi code, so apparently it worked fine in the old kernel? which is more weird as it doesn't seem to do anything which might depend on the kernel (except the irq value)!08:08
filippzhedayat: IIRC in the first approach someone c/p that function from old kernel as it was dropped in the new one (it was dropped for a reason - that reason being: it can't work like that)08:08
hedayatfilippz: :) OK! so probably the method IRQ values are assigned is changed.08:08
filippzhedayat: probably - but that's behind us - we're texting!08:09
hedayatfilippz: and, as I said, dial works too. Just wondering if we can actually talk, but hopefully it should work too!08:12
filippzhedayat: pulseadio has some inputs reported - one of them should be the mic - I didn't try it myself08:13
filippzyou can use this and try:
*** vakkov has joined #nemomobile08:15
hedayatfilippz: I'll look into it. I wonder if ofono test scripts does actually allow talking. let's see what happens.08:18
hedayatLet me summarize the needed changes before leaving: 1. older ofonod binary. I'm currently using ofonod from ofono-1.12. (binary is enough). (I didn't overwrite the binary, but overriden the service file to run the older binary)08:23
*** kostaja has quit IRC08:25
hedayat2. remove isimodem plugin from excluded plugins. Actually, I'm now excluding a single plugin: ril! but excluding more plugins is probably good for memory usage and maybe batter life? anyway, nothing works with RIL enabled. (its actually weird, but n900 plugin gets Permission denied errors reading /dev/cmt/ stuff when RIL is active!)08:25
*** hurrian has quit IRC08:27
hedayat3. changed ofono dbus policy: changing '<policy user="radio">' into '<policy group="users">' to allow nemo user to use ofono dbus services. Adding a new group (e.g. radio) and using it here is also an option, if preferred.08:28
*** kostaja has joined #nemomobile08:30
hedayatIIRC, these are the only required changes. I'll create a pr for the last 2. But should do something about the first problem. I'll hopefully try with previous release of ofono to see how it works, as Stskeeps suggested.08:31
hedayatleaving for now. :P08:31
*** dr_gogeta86 has quit IRC08:37
*** arcean has joined #nemomobile08:40
*** hurrian has joined #nemomobile09:00
*** hurrian has quit IRC09:00
*** hurrian has joined #nemomobile09:00
*** wazd has joined #nemomobile09:07
*** notmart has joined #nemomobile09:11
hedayatlocusf: signal strength works now (i restarted ofono to test something, and after re-entering the PIN the strength is now shown).09:11
locusfhedayat: good :)09:11
locusfalso its absolutely fantastic to see progress on telephony :)09:11
hedayatlocusf: and it (the strength bars) is also being updated09:12
locusfgotta buy a new prepaid to test it09:12
locusfhedayat: nice09:12
hedayatlocusf: Thanks. Even I can see events like missed calls :) It feels good. Just if there were some UI for seeing the call and answering it... ;)09:13
locusfthe bars are connected to contextkit-statefs bridge which subscribes to value changes09:13
locusfhedayat: we just gotta fix the dialer to use glacier components and prestart it as it does on Jolla09:14
locusfhedayat: does wlan work? I have yet to test it09:15
hedayatlocusf: I don't think so. In Sailfish, it works (at least shows available networks, but I have not tried to connect to any), but in Nemo no networks are being shown in Settings or when I click the icon on the bottom bar09:16
locusfI'm tem├ąted to get a nexus 5 for nemo development09:16
locusfhedayat: ok, its probably just a matter of activating it somehow09:17
hedayatlocusf: yeah09:17
locusfconnman-tests package has some scripts09:18
hedayatlocusf: I'll try that. Also, probably we need a number of fixes in dbus policies. I get lots of access denied errors about them. As it is required for ofono.09:20
locusfhedayat: ok09:20
locusfwe are gonna need ofono for android adaptations as well09:21
bencohofono-ril ?09:21
hedayatlocusf: created a pr against nemo-configs-n950-n909:23
bencoh(hmm looks like the current ofono-ril is the other way around)09:23
*** phaeron has joined #nemomobile09:24
locusfhedayat: merged09:24
hedayatlocusf: thanks :)09:25
locusfnp :)09:25
hedayatlocusf: How am I supposed to enable WiFi in the first place? :P09:28
hedayatdon't see any UI for it, and apparently it's off09:29
*** SK_work has quit IRC09:30
*** ZogG_laptop has joined #nemomobile09:30
hedayatlocusf: I can see WiFi networks using 'ifconfig wlan0 up;iwlist scan', but UI doesn't show anything. apparently, due to dbus permissions. I wonder if the UI user is considered 'at-console'....09:32
*** dr_gogeta86 has joined #nemomobile09:33
*** teve has quit IRC09:36
*** teve has joined #nemomobile09:47
*** xmlich02 has joined #nemomobile09:48
*** ZogG_laptop has quit IRC09:57
*** Wikiwide has joined #nemomobile10:15
lbtPSA: Mer OBS is likely to be down during today. Sorry but I had problems which stopped me yesterday/last night10:16
*** phaeron has quit IRC10:19
*** piggz has quit IRC10:21
*** piggz has joined #nemomobile10:30
locusfhedayat: try those connman test scripts, see this page
*** morphis has quit IRC10:43
*** morphis has joined #nemomobile10:47
hedayatlocusf: I'm going to look what's wrong with current dbus policies. I guess that 'at_console' part is supposed to be applied to UI user, but apparently, it doesn't. But I'll check the page too.10:53
hedayatlocusf: I wonder if changing ofono dbus policy was a proper solution.10:54
locusfhedayat: we probably just need nemo in radio group instead10:56
hedayatlocusf: No, the policy says 'user=radio', so it expects radio user rather than group.10:57
hedayatlocusf: Current nemo doesn't radio user/group at all10:57
locusfhedayat: ah ok10:58
hedayatlocusf: And reading other policies (IIRC bluetooth), I found out that it is expected that ofono itself is running as radio user10:58
hedayatbut it is probably  being run as root now10:58
*** ZogG_laptop has joined #nemomobile10:59
WikiwideAn old dream of mine... That different programs are treated like different users, with different permissions/policies.10:59
hedayatso I guess nemo user is expected to be able to communicate with these services using other policy definitions (at_console and the other one)11:00
*** piggz has quit IRC11:05
hedayatWikiwide: :)11:07
*** piggz has joined #nemomobile11:07
WikiwideOne acquaintance of mine runs everything as root, with only selected processes (such as browser) running under limited user - because they are the most risky ones. I run all programs as limited user, unless I have to run something as root. Having something more fine-tuned than limited-vs-root would be brilliant.11:09
*** piggz has quit IRC11:10
*** nodevel has quit IRC11:10
*** piggz has joined #nemomobile11:10
*** nodevel has joined #nemomobile11:10
hedayatWikiwide: I guess you'd love SELinux and recent sandboxing/isolating ideas/implementations. :P11:10
WikiwideLet's look it up... Ordinary user-processes often run in the unconfined domain, not restricted by SELinux but still restricted by the classic Linux access rights.11:12
hedayatWikiwide: With SELinux, a process can be as limited as you like: e.g. only connecting to specific ports are allowed for process X11:12
WikiwideActually, I am interested in Minix :)11:12
hedayatWikiwide: Yes, this is the 'userfriendly' approach towards SELinux. SELinux by default is a 'deny unless allowed' system.11:13
hedayatWikiwide: :)11:13
WikiwideDoes SELinux sandbox the hw drivers, so that keyboard driver would not affect the mouse, and so on?11:13
*** shentey has joined #nemomobile11:14
*** shentey has quit IRC11:14
WikiwideBecause most operating systems are monolithic. Most obviously, failure of one driver can mean failure of whole system, since drivers are somehow within the kernel. In Minix (as far as I remember), driver is just restarted, and does not bring anything else down.11:15
hedayatWikiwide: hmmmm, I'm not sure. AFAIK, it is mainly designed to restrict user-land processes rather than kernel ones. Specially since in Linux, these are actually not separate from kernel.11:15
*** shentey has joined #nemomobile11:15
hedayatWikiwide: Yes, Linux is certainly not a micro kernel. If you like micro kernel ones, you should look at modern micro kernels. AFAIK, there are a number of them out there.11:16
hedayatWikiwide: So, you might like something like Hurd (which was supposed to be such a kernel, I don't know how much it actually is)11:17
WikiwideI am recently interested in Plan 9 and Inferno.11:18
WikiwideNot sure what is in their kernel, though.11:18
hedayatWikiwide: hmm ok! I don't know either, but I 'feel' that Plan9 had a micro kernel. I don't trust this feeling though, I might have read about it once.11:21
WikiwideWikipedia says hybrid and does not say anything else.11:22
*** Guest47705 has quit IRC11:24
hedayatWikiwide: OK.11:26
*** arcean has quit IRC11:29
WikiwideHow is it going with users and policies?11:29
hedayatWikiwide: It seems that UI user is not considered at_console, this is because UI is run as user services11:32
hedayatI'm now trying make user services considered to be at_console, so that dbus policies are allowed for them11:32
*** nodevel has quit IRC11:34
*** jonwil has quit IRC11:37
*** lizardo has joined #nemomobile11:51
*** ZogG_laptop has quit IRC12:04
*** shentey has quit IRC12:05
*** shentey has joined #nemomobile12:05
*** louisdk has joined #nemomobile12:09
hedayatWell, this is a problem! I added pam_console in /etc/pam.d/systemd-user with hope that it is enough to recognize user services "at console' by dbus, but it was not.12:09
hedayatAnd it seems that the whole at_console concept is deprecated in favor of PolicyKit policies.12:10
hedayatAre there any plans in Mer/Nemo in this regard?12:10
locusfI have no idea12:11
hedayatProbably a solution for now is to actually change many dbus policies to explicitly allow users group.12:11
hedayatsed -i 's/ at_console="true"/ group="users"/' /path/to/all/dbus/policies :P12:14
*** shentey_ has joined #nemomobile12:15
*** shentey_ has quit IRC12:15
*** shentey_ has joined #nemomobile12:15
*** shentey has quit IRC12:17
*** Adriano_ has joined #nemomobile12:17
*** shentey_ has quit IRC12:24
*** shentey_ has joined #nemomobile12:25
*** Wikiwide has quit IRC12:25
*** hedayat has left #nemomobile12:26
*** flash1 has joined #nemomobile12:48
*** shentey_ has quit IRC12:49
*** shentey_ has joined #nemomobile12:50
*** shentey_ has quit IRC12:51
*** shentey__ has joined #nemomobile12:51
*** flash1 has quit IRC12:51
*** shentey has joined #nemomobile12:52
*** shentey__ has quit IRC12:52
*** KaIRC has joined #nemomobile12:58
*** martyone_ has quit IRC13:04
*** shentey has quit IRC13:06
*** piggz has quit IRC13:13
*** piggz has joined #nemomobile13:13
*** Jef91 has joined #nemomobile13:32
*** Jef91 has quit IRC13:32
*** Jef91 has joined #nemomobile13:32
*** phaeron has joined #nemomobile13:37
*** ZogG_laptop has joined #nemomobile13:59
*** zalan has quit IRC14:14
*** zalan has joined #nemomobile14:14
*** spiiroin has quit IRC14:17
*** zalan has quit IRC14:19
*** phdeswer has quit IRC14:22
*** Behold has joined #nemomobile14:28
*** filippz has quit IRC14:29
*** piggz has quit IRC14:32
*** piggz has joined #nemomobile14:33
*** jpetrell has quit IRC14:33
*** WWDrakey has left #nemomobile14:36
*** zalan has joined #nemomobile14:40
*** ZogG_laptop has quit IRC14:56
*** nodevel has joined #nemomobile15:00
*** spiiroin has joined #nemomobile15:08
*** flash1 has joined #nemomobile15:14
*** VDVsx_ has quit IRC15:16
*** flash1 has quit IRC15:16
*** himamura has joined #nemomobile15:33
*** r0kk3rz has joined #nemomobile15:39
*** arcean has joined #nemomobile15:42
*** Dynamit has quit IRC15:53
*** hedayat has joined #nemomobile16:00
hedayatI run the sed command against /etc/dbus-1/system.d/*, and it fixes both ofono (instead of changing the user="radio" one), and wireless networks are now visible in UI. It probably fixes some things related to bluetooth etc too (services which used at_console="true").16:02
hedayatI just wonder why Nemo/Sailfish doesn't list Ad-Hoc networks, like what Android does. I really hate it! Hope that it is not a design decision for Nemo. Harmattan does it, and it's sensible and really useful.16:04
hedayatFinally, replacing ofonod binary in sailfish and removing --nobacktrace (not supported by ofonod 1.12), makes simcard recognized in it too.16:05
*** vakkov has quit IRC16:08
*** wazd has quit IRC16:11
*** xhaakon has quit IRC16:24
*** xhaakon has joined #nemomobile16:25
*** xhaakon has quit IRC16:25
*** xhaakon has joined #nemomobile16:26
*** xhaakon has quit IRC16:26
*** xhaakon has joined #nemomobile16:27
*** xhaakon has quit IRC16:30
*** xhaakon has joined #nemomobile16:31
*** vakkov has joined #nemomobile16:32
*** tbr has quit IRC16:33
*** xhaakon has quit IRC16:36
*** xhaakon has joined #nemomobile16:37
*** Venemo has joined #nemomobile16:37
*** filippz has joined #nemomobile16:41
*** tbr has joined #nemomobile16:43
*** phaeron has quit IRC16:44
*** xhaakon has quit IRC16:49
*** xhaakon has joined #nemomobile16:49
*** filippz has quit IRC16:53
*** xhaakon has quit IRC17:11
*** xhaakon has joined #nemomobile17:12
*** messerting has joined #nemomobile17:12
*** shentey has joined #nemomobile17:12
*** VDVsx_ has joined #nemomobile17:29
*** vakkov has quit IRC17:51
*** Dynamit has joined #nemomobile17:58
*** veskuh has quit IRC18:15
*** notmart has quit IRC18:16
*** vakkov has joined #nemomobile18:19
*** dpurgin has quit IRC18:21
*** nodevel has quit IRC18:25
*** filippz has joined #nemomobile18:39
*** olesalscheider has quit IRC18:53
*** shentey has quit IRC18:56
*** olesalscheider has joined #nemomobile18:59
*** furikku has quit IRC18:59
filippz"Shutdown logo is easy to replace, but you will be having hard time replacing boot logo, as its builtin inside bootloader, so there is no easy way to change it." (
filippzI guess we'll have to use plymouth-lite on N9 since ubiboot has no support for boot logo AFAIK19:04
Aardfilippz: plymouth-lite messes up with the 32bpp framebuffer, only happy with 16bpp19:05
*** filippz has quit IRC19:08
*** filippz has joined #nemomobile19:08
filippzAard: I see no other alternative - I'll try to rebuild plymouth-lite (as current binary it's depending of older libpng) as see how it behaves19:11
Aardfilippz: moment19:11
hedayatfilippz: So, Jolla boot logo is inside their phones boot loader? OK, what does Harmattan do?19:11
Aardhedayat: overengineered custom splash screen, capable of videoplayback as well. our stack comes up too fast for videoplayback making sense19:12
hedayatfilippz: ubiboot supports loading png, so we can load something there (or in an initramfs, like mine); but Harmattan uses an animation for it.19:12
filippzhedayat: seem so - maybe moslo has some support in it... and since ubiboot is moslo based...19:12
filippzhedayat: a hint regarding "supports loading png" would be nice :)19:13
Aardfilippz: there, small present from us.
hedayatAard: thanks19:13
filippzAard: will try it ASAP, ty!19:13
Aardfilippz: basically, bring gles up as early as possible (how we did the wayland on n9 does that, for that purpose), and then just use gles to draw an image on the screen19:14
hedayatfilippz: There is a showpng in it. Actually the base graphical menu itself is a png.19:14
Aardwithout having to worry about all the framebuffer specific mess19:14
*** martyone_ has joined #nemomobile19:15
filippzAard, hedayat: I'll try glesplash first. Mainly because of this:
Aardfilippz: we've initially did it for the n9, used it on some other hardware, but in the end dropped it as early gles context on qc is not really doable19:17
filippzIn essence kernel resets display instead of relying on bootloader to do it for him - thus image from bootloader will dissapera within second or so...19:17
*** ajalkane has quit IRC19:23
hedayatfilippz: :(19:23
filippzhedayat: I know - that was also a fun game - guessing which boot parameter breaks booting19:24
hedayatfilippz: Yes, it's really fun! :P I've had this fun with the auto-shut-down problem, and soon realized that is independent of kernel params :D19:27
hedayatwhich is hopefully fixed (2 successful boots till now, with 0 failed boots) with a watchdog kicker :)19:28
filippzhedayat: as long as it works - I've had no problems booting it myself19:30
hedayatfilippz: Have you any idea why graphics seems to be a little sluggish? kernel problem or userspace one?19:31
hedayatfilippz: Yes, because you don't use an initramfs booting method as I do.19:32
filippzhedayat: I've noticed a lot of "slow update" warnings in dmesg - but they seem to be related to orientation change19:32
Aardhedayat: the wayland drivers are rather hackish. also there's no vsync19:33
filippzAard: I would point to that direction also since it was sluggis before orientation change19:34
hedayatfilippz: I see such message on every glacier animation, like moving between pages.19:34
filippzbut to be honest, I would say that 2.6 was better in that regart19:34
filippzhedayat: it's a waring from nokia display (as in lcd itself) driver - that measures time needed for image to be displayed19:35
Aardalso we're running in 32bpp for wayland, while more or less everybody else using this chip seems to run in 16bpp. took quite some time to figure out how to kick it into 32bpp back then19:35
filippzit seems that with other orientations image gets refreshed half-by-half (I'm not really an expert in that regard)19:36
filippzIf I'm reading it correctly ( we nog get PND_UPD_MODE_NORMAL/PND_UPD_TE_EXT19:38
hedayatAard: thanks for the explanations.19:38
filippzbefore that we would hit one of those:
*** martyone_ has quit IRC19:40
*** ajalkane has joined #nemomobile19:41
hedayathmmm... did 2.6.32x kernels have the same rotation as your new kernel?19:50
hedayatfilippz: ^19:50
filippzhedayat: no it was landcape, we switched to portrait few days ago since both glacier and sailfish home are (in a way) hardcoded to portrait devices19:51
hedayatfilippz: So, it didn't use UPD_MODE_NORMAL? (was refreshed half by half)?19:52
filippzlocusf knows more details as to why old colorfull home was working OK19:52
filippzhedayat: yes, it seems so - that would explain the warnings19:53
hedayathmm weird!19:54
filippzhedayat: I guess the code was never really tested in portrait mode, it does work, but not as fast19:55
locusffwiw glacier works on landscape as well19:57
*** piggz has quit IRC19:57
*** piggz has joined #nemomobile19:58
locusfso no hardcoding to portrait but depending on devices framebuffer native orientation19:58
locusflike p3110 native orientation is landscape19:58
locusfso glacier uses that19:59
locusfold colorful home could react to orientation changes19:59
filippzlocusf: yes -  but colorfull was responding to rotation events. Could we "instruct" glacier to do that also?19:59
filippz^yes "orientation changes" is the right expression20:00
locusffilippz: yes if we used applicationwindo asroot item ISO page20:00
locusfbut that doesn't work on lipstick as applicationwindow is qquickwindow20:01
filippzlocusf: and, there was problem when using applicationwindow QML IIRC20:01
filippzso definition of the problem would be: kernel doesn't behave very well in portrait fb orientation, but qquickwindow doesn't work for orinetation changes20:03
filippzOK, I guess we'll have to live with the current setup for now20:05
filippzty Aard, hedayat, locusf - I'm done for tonight, and will try glesplash tomorrow - good night to all20:09
hedayatfilippz: good night :)20:09
*** filippz has quit IRC20:09
locusffilippz: good night :)20:09
*** piggz has quit IRC20:09
hedayatAard: locusf: is it possible to run Nemo in 16bpp mode?20:09
*** piggz has joined #nemomobile20:10
*** phaeron has joined #nemomobile20:10
locusfhedayat: I wouldn't know20:10
Aardhedayat: not sure about nemo, at least sailfish is not possible -- the alpha channel you get from 32bpp is required20:11
hedayatAard: thanks, I also liked it for sailfish, to make it more smooth. That's unfortunate. :(20:12
locusfglesplash could also work for hadk devices20:16
Stskeepsnot really20:17
Stskeepsit will blow into fiery flames20:17
locusfyeah noticed coupling to omap20:19
locusfshould have looked at the code before blurting nonsense :)20:20
Aardlocusf: see above, it works for anything where you can get a gles context up very early20:22
locusfAard: but that wold require some rewriting and I don't have gles skills to do that :/20:23
Stskeepsit's not difficult.20:23
Stskeepsi don't either20:23
locusfok hmm20:27
*** hedayat has left #nemomobile20:49
*** ced117 has joined #nemomobile21:12
*** sandy_locke has joined #nemomobile21:19
*** sandy_locke has left #nemomobile21:20
*** sandy_locke has joined #nemomobile21:20
*** sandy_locke has left #nemomobile21:21
*** sandy_locke has joined #nemomobile21:21
*** sandy_locke has left #nemomobile21:22
*** sandy_locke has joined #nemomobile21:23
*** sandy_locke has left #nemomobile21:24
*** sandy_locke has joined #nemomobile21:24
*** sandy_locke has left #nemomobile21:27
*** sandy_locke has joined #nemomobile21:27
*** sandy_locke has left #nemomobile21:27
*** sandy_locke has joined #nemomobile21:28
*** sandy_locke has left #nemomobile21:30
*** louisdk has quit IRC21:32
*** sandy_locke has joined #nemomobile21:32
*** sandy_locke has left #nemomobile21:32
*** Adriano_ has quit IRC21:33
*** piggz has quit IRC21:33
*** jonwil has joined #nemomobile21:33
*** lizardo has quit IRC21:36
*** M4rtinK has joined #nemomobile21:39
*** SfietKonstantin has joined #nemomobile21:39
*** sandy_locke has joined #nemomobile21:42
*** sandy_locke has left #nemomobile21:42
*** Venemo has quit IRC21:47
*** SfietKonstantin has quit IRC21:54
*** sandy_locke has joined #nemomobile22:04
*** sandy_locke has quit IRC22:18
*** gabriel9 has joined #nemomobile22:54
*** Pat_o has quit IRC23:03
*** Wikiwide has joined #nemomobile23:07
*** DocScrutinizer05 has quit IRC23:27
*** DocScrutinizer05 has joined #nemomobile23:28
*** himamura has quit IRC23:35
*** r0kk3rz has quit IRC23:37
*** gabriel9 has quit IRC23:53
*** arcean has quit IRC23:55

Generated by 2.11.0 by Marius Gedminas - find it at!