IRC log of #maemo-ssu for Monday, 2013-09-16

Palinedko: PA not using alsa userspace API00:01
Paliit using direct kernel API00:01
Pali(if lennart did not changed it)00:01
Pali~poettering00:02
Pali~lennart00:02
infobotLennart is ~Poettering00:02
Pali~Poettering00:02
Paliinfobot dead?00:03
infobotyes :(00:03
nedkoPali: this is clarification not argument against my statement i guess?00:04
Palijust info which I known00:05
PaliI hate PA too00:06
nedkoso you are telling me that PA talks to ALSA code in linux kernel not via libasound2?00:06
Palilast time (in past) when I looked into PA it does not used any alsa userspace dmix/filters00:08
*** obironbo has quit IRC00:08
Paliand grabbed kernel alsa card directly00:08
Paliexclusively00:09
Palimaybe it was changed... I do not want to look into PA sources00:09
nedkoyou can grab the card directly and exclusively via the normal user-space ALSA API00:10
nedkoPA and JACK do this00:10
PaliI already read full lennart's ifplugd daemon and this was horrible...00:10
PaliI think that this guy does not know anything about networking and sockets00:11
PaliUsing busy waiting...00:11
nedkoi know for sure he knows this stuff very well, can you show an evidence to prove otherwise?00:12
Palilook into ifplugd source code00:14
Palior if you are lazy run: $ strace ifplugd00:14
Paliwith strace you will see at least busy waiting00:14
*** NIN101 has quit IRC00:26
nedkoPali: are you talking about daemon_retval_wait() ?00:30
*** wmarone_ has quit IRC00:31
*** sunny_s has joined #maemo-ssu00:31
Palinedko: I do not remember, but daemon did this: wait 1s, ask kernel for network adapter status, wait 1s, ask ....00:31
*** wmarone has joined #maemo-ssu00:31
PaliI do not want to read that lennart's code again...00:32
*** Martix_ has quit IRC00:42
nedkoPali: i read the code. and i think you are wrong. the code uses select() and there is indeed 1 second. it is used for things that you cannot do otherwise, no?00:47
Paliget network link status?00:48
nedkoyup, seems like this00:48
Paliyou can that00:48
nedkohow you do it?00:48
Paliopen netlink socket00:48
Palisend some data00:49
Paliand kernel will send you info when link change00:49
Palido not know all needed flags for netlink00:49
Palibut it is possible00:49
nedkohttp://git.0pointer.de/?p=ifplugd.git;a=blob;f=src/interface.c;h=c7747889639e55e5de71f20811d00b4315d4855c;hb=HEAD#l4600:50
nedkoIIRC you cannot detect the carrier this way00:50
nedkoi.e. UP != LOWER_UP00:51
nedkobut i may be wrong00:51
Palifor sure you can00:51
Palibut via netlink socket00:51
nedkohmm, maybe i gave wrong link...00:51
Paliand not using SIOCGIFFLAGS00:51
Paliifplugd is badly written00:52
Paliand using that SIOCGIFFLAGS with busy waiting00:52
*** sailus has joined #maemo-ssu00:52
nedkoPali: was it possible when lennart wrote that code?00:52
PaliI do not know00:53
Palibut this is still present in last version00:53
nedkothe last version is quite old00:53
Paliwhich is widely used00:53
Palinedko: RTM_GETLINK should be that netlink packet type01:02
Paliand RTMGRP_LINK group01:04
nedkothe code already uses RTMGRP_LINK01:06
nedkoso i wonder why he still polls01:07
Palibecause he is lennart?01:08
nedkoyeah, ad hominem works :-/01:08
nedkolatest ifplugd release is from 2005, the kernel documentation networking/operstates.txt was added in 200601:11
Paliand why he using RTMGRP_LINK before?01:13
nedkoPali: IFF_LOWER_UP was introduced in 2006 as well, after the latest ifplugd release01:17
nedkoPali: 2006 as well: http://lkml.indiana.edu/hypermail/linux/kernel/0602.2/1883.html01:20
nedkoand he wrote that code in 200301:22
Palihm, ok01:22
Palinedko: read that thread01:25
Paliand last email: http://lkml.indiana.edu/hypermail/linux/kernel/0602.2/2287.html01:25
nedkoPali: still 2006, three years after lennart wrote the ifplugd code for NETLINK stuff01:26
Paliin above email is written that it was possible before above patch... but I do not want to look into history... maybe really there was no API01:27
Paliwhat is bad is that last version of ifplugd does not support it and it is used in 201301:28
nedkoi agree01:28
*** oooaaaooo1 has joined #maemo-ssu01:41
*** oooaaaooo has quit IRC01:43
*** Pali has quit IRC01:55
*** sunny_s has quit IRC01:56
*** nox- has quit IRC02:42
*** M4rtinK has joined #maemo-ssu02:58
*** MohammadAG has quit IRC03:01
*** dos1 has quit IRC03:22
*** jonwil has joined #maemo-ssu04:00
*** M4rtinK has quit IRC04:31
*** LauRoman has quit IRC04:40
*** amiconn has quit IRC05:02
*** amiconn_ has joined #maemo-ssu05:02
*** amiconn_ is now known as amiconn05:02
*** n900-dk_ has joined #maemo-ssu05:30
*** n900-dk has quit IRC05:30
*** mkaindl has quit IRC05:56
*** amiconn has quit IRC05:59
*** amiconn_ has joined #maemo-ssu05:59
*** amiconn_ is now known as amiconn05:59
*** wmarone has quit IRC06:12
*** wmarone has joined #maemo-ssu07:20
*** Vlad_on_the_road has joined #maemo-ssu07:26
*** bromide is now known as discopig08:26
*** Vlad_on_the_road has quit IRC09:09
*** LauRoman has joined #maemo-ssu09:25
*** jonwil has quit IRC09:51
*** Martix_ has joined #maemo-ssu09:55
*** Martix_ has quit IRC10:00
*** LauRoman has quit IRC10:03
*** sunny_s has joined #maemo-ssu10:18
*** hatake_kakashi has quit IRC10:23
*** psycho_oreos has joined #maemo-ssu10:29
*** Martix_ has joined #maemo-ssu10:41
*** MohammadAG has joined #maemo-ssu10:42
*** luf has joined #maemo-ssu10:44
*** Martix_ has quit IRC10:45
*** Martix_ has joined #maemo-ssu10:45
*** Pali has joined #maemo-ssu10:45
*** Martix_ has quit IRC10:53
Paligitorious is not working again10:54
Paliit reject me to push anything10:54
Pali"Access denied or wrong repository path"10:54
Paligitorious started working11:53
*** dos1 has joined #maemo-ssu12:06
*** DrCode has quit IRC12:16
*** jonwil has joined #maemo-ssu12:50
*** DrCode has joined #maemo-ssu12:50
*** DrCode has quit IRC12:56
*** LauRoman has joined #maemo-ssu12:59
PaliFatPhil: ping13:07
*** DrCode has joined #maemo-ssu13:21
*** oooaaaooo1 has left #maemo-ssu13:46
*** arcean has joined #maemo-ssu14:30
FatPhilPali: pong!14:41
*** freemangordon_ is now known as freemangordon14:49
freemangordonPali: maybe we shoud switch to github14:49
freemangordonPali: BTW maybe you should rebase linux to 3.11 or 3.1214:50
*** _LauRoman has joined #maemo-ssu14:51
nedkomaemo for workgroups :)14:52
*** LauRoman has quit IRC14:54
freemangordonFatPhil: hi! Do you have some 3.x rx51 tree by chance :)?14:54
freemangordonnedko: hmm?14:55
freemangordonwin311?14:55
nedkofreemangordon: yes, linux-3.11 codename is "Linux for Workgroups"14:56
freemangordonomg, really? heh14:56
nedkoyes, check in /Makefile line 5 :)14:56
FatPhilfreemangordon: I did find a rather strange 2.6.3x rx-51 tree. I have no idea why it existed.15:01
*** _LauRoman is now known as LauRoman15:05
*** MrPingu has joined #maemo-ssu15:26
*** ruskie has quit IRC15:34
*** ruskie has joined #maemo-ssu15:34
*** luf has left #maemo-ssu15:37
*** ruskie has quit IRC15:39
*** ruskie has joined #maemo-ssu15:39
*** mkaindl has joined #maemo-ssu15:49
*** ruskie has quit IRC15:59
*** ruskie has joined #maemo-ssu16:04
*** ruskie has quit IRC16:05
*** ruskie has joined #maemo-ssu16:05
*** jonwil_ has joined #maemo-ssu16:06
*** jonwil has quit IRC16:07
*** jonwil_ is now known as jonwil16:07
*** oooaaaooo has joined #maemo-ssu16:08
oooaaaooohi guys im trying to install the cssu by clicking on the link on http://wiki.maemo.org/Community_SSU but HAM says "community-ssu-enabler" already installed16:10
thedead1440oooaaaooo: it means you already have it; are you sure you don't have CSSU?16:13
oooaaaoooit had stalled the last time i tried to update it for some reason or another and when i restarted the phone the maemo "cube" wasnt available on my desktop16:14
oooaaaooofailed to download the 23mb package16:14
thedead1440you may want to do refresh catalogues (or is it called update?) in HAM to allow it to complete the download/installation16:17
*** mkaindl has left #maemo-ssu16:18
oooaaaooothedead1440: which catalogs are related to cssu?16:19
thedead1440no i mean HAM has an update option; click on it for it to update the catalogues and present to you any updates which would complete the installation16:20
oooaaaooothedead1440: yeah but i would like to disable the non related catalogs to speed up the process16:20
thedead1440why mess with the catalogues? You clicked on the .install file right? It set up the catalogues for you. All you need to do now is allow HAM to complete the job instead of finding some shortcut which won't help most probably16:21
oooaaaoooi had non-standard catalogs installed before16:21
oooaaaoooand before coming in here i disabled some of them16:22
thedead1440sigh... just let HAM handle it instead of speculating...16:22
*** arcean_ has joined #maemo-ssu16:25
oooaaaooothedead1440: i have both testing and normal catalogues installed16:27
oooaaaooothedead1440: lol16:27
oooaaaoooi see a maemo5 CSSU package(thumb) update16:28
oooaaaoooi guess i install that16:28
*** arcean has quit IRC16:28
*** MrPingu has quit IRC16:39
*** MrPingu has joined #maemo-ssu16:52
*** oooaaaooo has left #maemo-ssu17:01
*** ruskie has quit IRC17:02
PaliFatPhil: 2.6.3x rx-51 tree is on meego gitorious and was there was n900 meego support17:05
PaliFatPhil: what I wanted to ask: is there any (nokia) tool for memory leak statistics?17:06
PaliI want to fix some mem leaks bugs in some application, I already used valgrind - but it does not show statictic - instead for every memleak show message17:06
Paliand if there are 10000 leaks, it is not possible to read valgrind output...17:07
Palisailus: ping17:07
Palifreemangordon: ping17:07
discopighi17:14
FatPhilPali: we had some scripts for that, I thought they were available from meego context17:16
*** ruskie has joined #maemo-ssu17:16
PaliFatPhil: if these scripts/tools are open source, can you give me link or at least name of tools?17:16
Paliwhat I found is: https://wiki.maemo.org/Documentation/devtools/maemo5/sp-memusage17:17
Pali(and sources are in maemo SDK repo)17:17
FatPhilsp-* yup, that's what we used17:17
FatPhilWell, there was a separate dedicated team using that17:18
Paliwhat does *sp* means?17:18
FatPhilnamed after the system performance team, perhaps?17:18
*** Martix_ has joined #maemo-ssu17:18
FatPhilWe're gaining one of that team here next week, I can ask him if he has any insights (but I leave here in 2 weeks...)17:19
Paliok, I will try to compile that tool for x86-64 and try to use it :-)17:19
Palisailus: freemangordon trying to get both n900 cameras working with 3.11 kernel, but he only see green output in mplayer17:21
jonwilok, just updated http://wiki.maemo.org/N900_dbus with some more info about com.nokia.phone.SSC17:21
Palisailus: do you have any idea why?17:21
DocScrutinizer05LAST FAMOUS WORDS:17:22
DocScrutinizer05[2013-09-16 15:28:18] <oooaaaooo> i see a maemo5 CSSU package(thumb) update17:22
DocScrutinizer05[2013-09-16 15:28:27] <oooaaaooo> i guess i install that17:22
DocScrutinizer05[2013-09-16 16:01:34] <-- oooaaaooo has left this channel.17:22
jonwilWe now have what in theory should be correct prototypes for all the cellular DBUS interfaces that we need to care about for Neo90017:23
kerioDocScrutinizer05: lol17:23
DocScrutinizer05actually "famous last words"17:24
* DocScrutinizer05 needs more coffee17:24
DocScrutinizer05jonwil: \o/17:27
jonwilwe still need to figure out all the arguments etc17:30
DocScrutinizer05dammit, the http://www.cncmods.net/files/dbus/ssc.xml link on http://wiki.maemo.org/N900_dbus a) should rather point to another wiki page inside wmo, and b) of course doesn't render at all on my konqueror17:30
jonwilbut its a start17:30
nedkoDocScrutinizer05: i'd like to make that page asciidoc17:30
DocScrutinizer05http://wiki.maemo.org/User:Joerg_rw/dbus17:31
DocScrutinizer05sidenote ^^^17:31
jonwilif you want to improve the N900_dbus page, feel free :)17:32
nedkoi also try to document the daemons that register the dbus names17:32
nedkothere are two of them17:32
nedkogot csd17:33
nedkofor csd17:33
jonwileverything on that page under Cellular functionality comes from the csd daemon and its plugins except com.nokia.phone.SSC17:34
jonwilcom.nokia.phone.SSC comes from sscd17:34
nedkonot on my n900 :]17:34
nedkomeh17:35
nedkoshame17:35
jonwilmeh what?17:35
nedkosame on my n90017:35
nedkosorry17:35
jonwilok17:35
jonwillet me identify who is using each of the relavent dbus interfaces :)17:40
jonwilcellular functionality is definatly the hardest part of this whole jigsaw puzzle, audio and GPS are next hardest17:41
jonwilAudio is only hard because of the need to deal with all the various closed-source bits in the system17:41
jonwiland how those things will change as a result of the differences in the audio system on N900 vs Neo90017:42
DocScrutinizer05ohyeah :-S17:42
DocScrutinizer05I'm still undecided if we shouldn't rather try to get the AIC32xxxx mixer into Neo900 hw17:43
jonwilok, so on the current GTA04 hardware, are there any binary kernel bits required?17:43
DocScrutinizer05dang, I missed to "bookmark" the quote from FatPhil(?) about OMAP audio and the SiErr(?) it had in every single function17:44
jonwiloh btw for pulseaudio we should maybe see what the meego/mer/nemo/whatever its called these days/etc guys are doing with the N900 stuff and see what pulseaudio bits they are using17:53
DocScrutinizer05(([2013-09-12 13:59:59] <FatPhil> Peter Ujfalusi once gave...)) found it. :-)17:53
PaliFatPhil: sp-memusage does not have any memleak detector18:02
FatPhilsp-endurance18:02
*** bsdmaniak has joined #maemo-ssu18:02
FatPhilvery coarse grained, but can tell you if there's a leak or not18:03
*** dos1 has quit IRC18:03
FatPhilI'm pretty sure I was able to link with Electric Fence18:06
FatPhilHmm, maybe that doesn't help with leaks, I've forgotten.18:08
*** NIN101 has joined #maemo-ssu18:11
*** amiconn_ has joined #maemo-ssu18:12
*** amiconn has quit IRC18:12
*** amiconn_ is now known as amiconn18:12
*** MrPingu has quit IRC18:15
*** MrPingu has joined #maemo-ssu18:16
*** freemangordon_ has joined #maemo-ssu18:23
*** freemangordon has quit IRC18:23
*** MrPingu has quit IRC18:40
*** dos1 has joined #maemo-ssu18:41
jonwilI doubt anyone outside of Nokia will ever be able to figure out what com.nokia.phone.sim.get_service_provider_info actually does19:17
dos1what does it return?19:28
jonwil'som numbers whos purpose I have never figured out19:30
jonwilsome19:30
*** dos11 has joined #maemo-ssu19:42
*** dos1 has quit IRC19:43
sixwheeledbeastPali: If I wish to have swap on uSD do I just 'echo swap_sd="1" >> /etc/default/mount-opts-overwrite'. Which update in "cssu-devel" provides this?19:43
Palisixwheeledbeast: that option will automatically enable all swap partitions on SD card during boot + it adds them to /etc/fstab19:44
*** jonwil has quit IRC19:48
sixwheeledbeastPali: Yep, I have edited mount-ops and tested this but noticed the comments at the start of the file.  Will this edit to mount-opts-overwrite work on all CSSU versions?19:54
Palisixwheeledbeast: all cssu versions which support swap_sd support also overwrite file19:55
Palisixwheeledbeast: and you really do not edit mount-opts file!19:56
Paliit can cause problems when updating via HAM19:56
Palior problems if you do not answer Y when updating via apt-get19:56
sixwheeledbeastI realise that from reading comment at stop of /etc/fstab ;) You don't happen to know which versions support it?19:57
freemangordon_Pali: pong20:06
*** freemangordon_ is now known as freemangordon20:06
*** MrPingu has joined #maemo-ssu20:07
*** Vlad_on_the_road has joined #maemo-ssu20:20
*** wiredless has joined #maemo-ssu20:59
wiredlesshaven't been here since nokia 770 days21:02
DocScrutinizer05wow, wellcome back. I just wonder if this channel even existed at 770 times21:10
wiredlessthink it did i could be wrong21:13
wiredlessany idea if urho kontorri comes in here21:14
wiredlesshe made theme maker for maemo21:14
wiredlessi know he wound up going to work for nokia21:16
DocScrutinizer05haven't "seen" kontorri for quite a while21:18
wiredlesshope he still has a job21:18
DocScrutinizer05~seen kontorri21:18
infobotDocScrutinizer05: i haven't seen 'kontorri'21:18
wiredlessso whats the newest hot hardware for maemo?21:19
DocScrutinizer05err, existing or projected?21:19
wiredlesssomething you can buy now21:20
FatPhil2nd hand n900s21:20
wiredlessi see21:20
DocScrutinizer05depending on your definition of "maemo" a N900 or a N921:20
FatPhilwell, you want a keyboard, so N950, perhaps?21:21
DocScrutinizer05some say N9 runs maemo6, some say it runs meego, some say it doesn't run at all21:21
DocScrutinizer05hard to buy ;-)21:21
wiredlessheh21:21
FatPhiltrue, rare as rocking-horse shit21:22
DocScrutinizer05also see www.neo900.org21:22
freemangordonFatPhil: IMO n900 has better KBD than N95021:22
DocScrutinizer05born from a widespread frustration about that situation21:22
freemangordonnot to say that opening n950's kbd is a nightmare. and impossible to be done with one hand21:23
DocScrutinizer05ack, n950 kbd is *cough*21:23
freemangordonPali: hmm, could our ssi problem be related to mux configuration?21:23
freemangordongoing to try it21:23
Palimaybe21:25
FatPhilfreemangordon: agree21:29
freemangordonPali: DocScrutinizer05: FatPhil: please verify http://pastebin.com/aAULuuyc21:35
freemangordonthe last comment is /* ssi1_wake */21:36
*** MrPingu has quit IRC21:38
FatPhilfreemangordon: see http://www.spinics.net/lists/linux-omap/msg96880.html21:38
wiredlesskeep the faith21:39
*** wiredless has left #maemo-ssu21:39
freemangordonFatPhil: hmm? some type correction21:40
freemangordonFatPhil: elaborate please21:40
*** arcean_ has quit IRC21:40
DocScrutinizer05freemangordon: huh?21:40
freemangordonDocScrutinizer05: I am almost sure mux is set for UART, not SSI21:42
freemangordonthat is why modem does not work21:42
freemangordonFatPhil: s/type/typo/21:42
DocScrutinizer05freemangordon: sorry, can't tell from that snippet21:43
freemangordonDocScrutinizer05: look at schematic, names are there21:44
freemangordonomap to rapu bus21:44
freemangordonFatPhil: oh, got it, we have that typo :)21:44
DocScrutinizer05when the names match with schematics, then I guess it's correct. Yes21:44
freemangordonDocScrutinizer05: I wanted you to check if they mach, along with inpto/output flags21:45
freemangordonI just wrote that snippet and want someone else to verify if it is correct21:45
DocScrutinizer05sorry, a tad under timing constraints right now21:45
freemangordonsure, np21:45
*** xes has joined #maemo-ssu21:46
DocScrutinizer05I can only speculate what your code snippet does. Looks to me as if the mux is set to mode1 for all pinds listed there21:46
freemangordonyep21:47
freemangordonmode 0 is UART21:47
freemangordonmode 1 *should* be SSI21:47
*** bsdmaniak has quit IRC21:50
*** arcean_ has joined #maemo-ssu21:50
freemangordondata is missing from TRM, but is in the kernel code21:51
DocScrutinizer05freemangordon: yes, from a cursory check that code snippet looks fine21:53
freemangordonthanks21:53
freemangordonbooting it, lets see21:53
freemangordon:D21:53
*** ruskie has quit IRC21:57
*** M4rtinK has joined #maemo-ssu22:04
freemangordonhmm:22:12
freemangordon/* gpio_74 */22:12
freemangordonOMAP3_MUX(DSS_DATA4, OMAP_PIN_INPUT_PULLDOWN | OMAP_MUX_MODE4),22:12
freemangordonthis is supposed to be CMT_EN22:12
DocScrutinizer05yes, and output, according to schematics22:17
freemangordon:nod:22:17
freemangordonI think all gpios are messed :(22:17
freemangordonbut how is that possible, it is set as OMAP_GPIO_SWITCH_FLAG_OUTPUT in the board code :(22:21
*** ruskie has joined #maemo-ssu22:24
*** DrCode has quit IRC22:41
*** DrCode has joined #maemo-ssu22:45
DocScrutinizer05freemangordon: when you're done with the work and have a beer, you *might* ponder about putting ETM into purpose, so you could tell who's doing what when22:48
DocScrutinizer05OMAP's Embedded Trace Macrocell is a nifty thing, and a pity to not exploit it22:49
*** sunny_s has quit IRC22:52
FatPhilETM is available in the harmattan kernel tree22:55
*** xes has quit IRC23:04
*** xes has joined #maemo-ssu23:06
Palifreemangordon: gpios are exported via sysfs and more maemo daemons setting them23:24
Palialso omap gpio switch driver doing something...23:24
freemangordonPali: I think it is NOLO that sets lots of things23:25
Paliyes and uboot setting someting too23:25
freemangordonhowever, having SSI gpios set to UART mode doesn;t look correct to me23:25
*** arcean_ has quit IRC23:25
Palifreemangordon: look here: https://elektranox.org/n900/libisi/power.html23:27
freemangordonPali: yes, I have similar script23:27
freemangordonand this is what sscd does23:28
freemangordonbut if gpios are not se correctly... ;)23:28
Palifreemangordon: look at data for omap gpio switch driver23:29
freemangordonPali: it is not enough, mux have to be st too23:30
freemangordon*set23:30
Palibecause OMAP atag table is not in kernel anymore I copied gpio data from NOLO atags directly to rx51 board code for omap gpio driver23:30
Paliand maybe there is some copy paste error23:30
Palibtw same data should be in u-boot23:30
Palimux are set by uboot too23:30
Palimaybe kernel resetting them now?23:30
freemangordonPali: SSI is not a public interface23:31
freemangordonPali: :nod:23:31
freemangordonlook at serial.c23:31
freemangordonserial driver sets all 4 uart pins to uart mode if they are in default state23:31
freemangordons/pins/devices/23:32
infobotfreemangordon meant: serial driver sets all 4 uart devices to uart mode if they are in default state23:32
freemangordonbut ssi shares pins with uart123:32
freemangordonso board code should tell the mux to put the in ssi mode23:32
freemangordontha same for mcbsp423:32
freemangordonPali: do you have KP compiled with debugfs enabled?23:33
Palikp should have debugfs23:34
freemangordonwell, 3.x has more info :)23:34
freemangordonPali: where to look for your "copied" gpois?23:35
freemangordongpios*23:35
Paliboard-rx51-peripherals.c23:35
Paliat begining23:35
Palifreemangordon: look also here: https://gitorious.org/meego-device-adaptation/n900_nokia-n900-configs/source/372022b4f463771534625dde7d714cad071fcea7:nokia-n900-configs.sh23:36
freemangordonPali: rx51_gpio_switches?23:36
Paliyes23:36
freemangordonok, I'll try this script23:37
freemangordonbut I guess the board code should do the same23:37
*** Vlad_on_the_road has quit IRC23:42
*** NIN101 has quit IRC23:51

Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!