#mer log for Sunday, 2012-05-27

« Saturday, 2012-05-26 Index Monday, 2012-05-28 »
*** phdeswer has quit IRC00:10
*** dijenerate has quit IRC00:11
*** dijenerate has joined #mer00:25
*** vgrade2 has joined #mer00:26
*** vgrade has quit IRC00:29
*** dijenerate has quit IRC00:29
*** dijenerate has joined #mer00:30
*** swerden has joined #mer01:11
*** swer has quit IRC01:12
*** beford has quit IRC01:13
*** beford has joined #mer01:26
*** tsdedst has quit IRC01:46
*** npm has quit IRC01:47
*** npm has joined #mer01:51
*** otep has quit IRC02:02
*** tsdedst has joined #mer02:03
*** otep has joined #mer02:16
*** dionet has joined #mer02:21
*** dionet has quit IRC02:53
*** furikku has joined #mer03:37
*** Alison_Chaiken has joined #mer04:08
*** board is now known as boar04:47
*** wmarone_ has joined #mer04:59
*** wmarone has quit IRC04:59
*** sonach has joined #mer05:10
*** InformatiQ has quit IRC05:12
*** beford has quit IRC05:17
*** InformatiQ has joined #mer05:18
*** tilgovi has joined #mer05:27
*** Alison_Chaiken has quit IRC05:37
Stskeepsmorn06:18
phaeronStskeeps: morning06:18
Stskeepsmorn phaeron :)06:19
phaeronso it wasn't imager's fault :)06:19
Stskeepsyeah, binfmt_misc is always such a mess06:20
phaeronand opensuse is still shipping qemu-0.14 with opensuse 12.106:20
*** harbaum has joined #mer06:46
*** himamura has joined #mer06:48
*** niqt has joined #mer06:59
*** niqt has quit IRC07:16
*** NIN101 has joined #mer07:33
*** rcg1 has joined #mer08:06
*** CosmoHill has joined #mer08:15
rcg1mornin08:16
Stskeepsmorn rcg1 o/08:16
*** M4rtinK has joined #mer08:17
Stskeepshow are things?08:17
CosmoHillmorning08:18
rcg1Stskeeps: quite well, finally had some time to enjoy the beach :)08:27
rcg1and am messing with the n950 vs. mer vs. battery status issue08:27
Stskeepswoo:nod:08:30
rcg1but so far i feel like being in an entirely dark room with my eyes closed xD08:33
Stskeepsdidn't we have libbmeipc somewhere?08:33
rcg1hmm am currently trying to get the bq27... kernel module running08:43
sonachStskeeps: when I making rootfs, it prints '110 packages to be installed, 100 packages gotten from cache, 10 packages to be downloaded'. but the command is 'mic create fs sdk_test.ks --outdir=/home/zhanghui/Mer/test --arch=armv7l --pkgmgr=yum', this means I don't appoint cache. so, what is happening?08:43
rcg1https://bugs.nemomobile.org/show_bug.cgi?id=5308:43
Stskeepssonach: it uses /var/tmp/mic/cache or something08:43
rcg1Stskeeps: ^ that's what i am currently trying.. so far i am at the same state as in the bug report08:44
rcg1so the kernel module builds, loads, and reports values but these values are far from making any sense08:44
Stskeepsrcg1: i don't think bq27 is what sits in n950 tbh08:44
sonachok,08:45
rcg1Stskeeps: ah ic.. at least for the n9 DocScrutinizer said it has a bq27.. dunno how similar the n950 is, though08:48
rcg1on the other hand he also mentioned twl5031_bbc being supported08:48
rcg1and for twl5031 there is already twl5031-aci.ko in the mer kernel08:49
rcg1dunno if it helps with the battery issue, though08:49
rcg1but shouldn't the fact that with the patch as included in the bug report values are reported via the bq27 module indicate that a bq27 is there? or could this be just some other arbitrary chip that responds to the requests?08:52
Stskeepsor a similar-ish chip08:53
*** ogoz_idle is now known as ogoz_08:54
rcg1hmm ic08:57
rcg1well.. am playing around a little and will see what turns out...08:57
rcg1still not sure if this twl5031 could be any help08:57
rcg1because there already a kernel module is built for the stock mer kernel08:57
rcg1Stskeeps: but thanks alot for the head up that it could be infact a totally different chip09:01
*** ogoz_ is now known as ogoz_idle09:05
*** ogoz_idle is now known as ogoz_09:06
*** CosmoHill has quit IRC09:07
*** ogoz_ has quit IRC09:13
*** _moguriso has joined #mer09:13
*** _moguriso is now known as moguriso_09:13
*** sonach has left #mer09:15
*** trbs has joined #mer09:18
*** araujo has quit IRC09:29
*** araujo has joined #mer09:34
*** araujo has joined #mer09:34
*** NIN102 has joined #mer09:34
*** NIN101 has quit IRC09:34
*** harbaum_ has joined #mer09:38
*** harbaum has quit IRC09:40
*** trbs has quit IRC09:47
*** harbaum_ has quit IRC09:49
*** BeholdMyGlory has quit IRC10:11
*** BeholdMyGlory has joined #mer10:12
*** CosmoHill has joined #mer10:17
Sleepy_Coderw00t: pretty widgets :D10:23
*** gimli has joined #mer10:37
DocScrutinizerrcg1: I said charger chip is bq24153 in n950. the chip on addr which is used by bq27200 in n900, on n950 doesn't behave like bq27200, best fit fo the N9(!) schematics was this other battery gauge chip bq27450 or what it's been10:39
DocScrutinizerrcg1: and I guess twl5031_<batTery10:39
DocScrutinizer<battery> is a deadend10:40
DocScrutinizersince the twl4031 isn't used for battery charging in N9.*10:40
rcg1DocScrutinizer: yeah, sorry if i was a little imprecise10:42
rcg1for me it boiled down to bq27xxxx the kernel module actually distinguished between bq27000 and bq2750010:43
DocScrutinizermaybe10:43
DocScrutinizerI never looked into the upstream kernel10:44
DocScrutinizerI just know the fremantle bq27200.ko is fsckdup to the bones10:44
DocScrutinizerin openmoko we used another non-mainline bq27k.ko that was pretty nice10:45
Stskeeps-g w00t10:45
Stskeepserr..10:45
DocScrutinizerbut as well would conflict with bme10:45
* w00t hands Stskeeps a /10:46
rcg1DocScrutinizer: ic10:47
*** jbos__ has joined #mer10:49
*** jbos_ has quit IRC10:52
*** jbos__ is now known as jbos_10:52
*** jbos__ has joined #mer11:09
*** sirdancealot has quit IRC11:10
*** arcean_ has joined #mer11:10
DocScrutinizerrcg1: http://www.ti.com/lit/ds/symlink/bq27520-g2.pdf11:11
DocScrutinizeris quite a different critter than bq27x0011:12
*** jbos_ has quit IRC11:12
rcg1DocScrutinizer: ic11:12
*** jbos__ is now known as jbos_11:12
*** sirdancealot has joined #mer11:13
rcg1am currently trying to see what happens if i'm using the bq27500 specific setting in the code of the backported patch as in the bug report11:13
rcg1at the same time i am looking at the chip with i2cdump and try to figure out if the data sheet makes sense with what i see there11:14
rcg1i.e. try to see if the "addresses"/"command codes" stated in the data sheet correspond to something i see in the dumped output11:15
DocScrutinizeryup11:15
rcg1i hope this could at least verify or reject the hypothesis we are dealing with an bq27520 here11:15
DocScrutinizercorrect approach11:15
*** dionet has joined #mer11:16
rcg1but i am very new to this very low level stuff.. so i might need a little time to figure things out and might not always choose the shortest path to a solution ;)11:16
rcg1DocScrutinizer: but thanks again for the help.. also your post you linked yesterday showed me direction to use the i2c-tools which are apparently very helpful for this job :)11:19
*** damiano has joined #mer11:36
damianohello11:37
Stskeepshello damiano :)11:37
damianonice to meet you Stskeeps11:42
damianoI need help to install meego on HP Touchpad11:43
Stskeepsmm, but this is mer :)11:43
damianook... but It's not the problem, I think the problem with nemo or cordia or any other mer implementation11:45
damianoin fact I tried only meego 1.2 images11:45
damianocould you give a look at the tutorial I followed?11:46
damianoplease?11:47
damianohttp://wiki.merproject.org/wiki/Adaptation/Touchpad11:47
damianoStskeeps, are you there?11:48
*** CosmoHill has quit IRC11:49
rcg1damiano: i assume he will eventually answer.. we are just not always at the computer ;)11:49
*** DocScrutinizer has quit IRC11:51
*** DocScrutinizer has joined #mer11:52
damianook thanks, I have to leave pc in ten minutes, I'll go back in another day11:52
*** swerden has quit IRC11:53
*** decibyte has quit IRC11:53
*** AndrewX192 has quit IRC11:53
*** steff has quit IRC11:53
*** jarkko^ has quit IRC11:53
*** sampos has quit IRC11:53
*** alterego has quit IRC11:53
*** noopman has quit IRC11:53
*** jarkko^ has joined #mer11:53
*** AndrewX192 has joined #mer11:53
*** AndrewX192 has quit IRC11:53
*** AndrewX192 has joined #mer11:53
*** Facefox has quit IRC11:55
*** Facefox has joined #mer11:59
*** CosmoHill has joined #mer12:00
rcg1well... looking at the data i highly doubt it is an bq27xxx chip...12:02
rcg1e.g., at 0x16+0x17 (time to empty) or 0x18+0x19 (time to full) i cannot observe any change at all12:05
rcg1generally there are actually very few parts that do actually change12:06
rcg1http://pastebin.com/VvE51v29 this is some of the data i gathered.. fwiw12:07
rcg1using "i2cdump 2 0x55"12:08
*** swerden has joined #mer12:16
*** decibyte has joined #mer12:16
*** steff has joined #mer12:16
*** sampos has joined #mer12:16
*** alterego has joined #mer12:16
*** noopman has joined #mer12:16
*** tommis has joined #mer12:24
DocScrutinizerrcg1: the 2 most relevant parameters to check for an unknown/unidentified I2C chip: I2C chip address (it's usually hardcoded to the chip, like e.g. 0x6b for bq24150), and then #2 most chips have an ID register with a well defined constant to read out from there.12:27
DocScrutinizerrcg1: regarding bq27520 I'm still struggling to understand the interface basic operation12:28
DocScrutinizerAIUI you got a index register to point somewhere into the addr room of the chip, and a data register (area) where the (range of) data from chip's addr room is mapped to12:30
DocScrutinizerthe datasheet is *very* cryptic and fuzzy12:30
ali1234slus955a.pdf?12:33
DocScrutinizerali1234: well, I'm currently looking at SLUSAB7A12:38
rcg1DocScrutinizer: i thought what you see with i2cdump is already the data you get from the chip?12:38
DocScrutinizerbq27520-G212:38
rcg1at least the address range corresponds to the commands given12:39
ali1234this is same but for -g112:39
DocScrutinizerrcg1: yes12:39
ali1234i2c commands look identical12:39
DocScrutinizerali1234: yes, G3 is currently the most recent though, it seems12:39
DocScrutinizeranyway...12:40
DocScrutinizer>>4.1.1.1.2 DEVICE_TYPE: 0x000112:40
DocScrutinizer>>Instructs the fuel gauge to return the device type to addresses 0x00/0x01. The bq27520-G2 device type returned is 0x0520.12:40
DocScrutinizerI understand this like: "write 0x0001 to command register (whatever this may be), and read back the values (like 0x0520) from data register12:41
rcg1DocScrutinizer: hmm, any way i can test this with i2c-tools?12:41
DocScrutinizerit's basically a double dereference AIUI12:41
DocScrutinizersure, you do a i2cset foo bar --word 0x000112:42
DocScrutinizerand then immediartely a i2cget --word foo bar DATAREGISTER12:42
DocScrutinizerbut I'm still just handwaving12:43
DocScrutinizerstill trying to decode this abyssmally fuzzy and obscure datasheet12:43
DocScrutinizeractually replace DATAREGISTER by 0x00 in my post 2 up12:44
DocScrutinizer>>Instructs the fuel gauge to return the device type **to addresses 0x00/0x01.**12:44
rcg1i2cset 2 0x55 0x00 0x0001 w12:45
rcg1that's what i did12:45
DocScrutinizeryep, that's what I think12:45
DocScrutinizerthen read out same addr12:45
rcg10x052112:45
DocScrutinizer\o/12:45
rcg1using this command line: "i2cget 2 0x55 0x00 w"12:46
DocScrutinizerso this is a bq27521 I bet my ass12:46
rcg1ah, alright :)12:47
DocScrutinizeryou see the basic opeation now?12:47
DocScrutinizerwrite COMMAND to 0x0012:48
DocScrutinizerread result from 0x0012:48
DocScrutinizeror 0x00+<offset>12:48
DocScrutinizerAIUI12:48
rcg1i see12:48
DocScrutinizerNFC why it has to be THAT complex12:49
DocScrutinizeralso NFC why this very sentence doesn'T show up in BOLD in the datasheet12:50
DocScrutinizerrcg1: do the same with COMMAND=0x0002 please (FW-VERSION)12:51
DocScrutinizeri2cset 2 0x55 0x00 0x0002 w; i2cget 2 0x55 0x00 w12:51
rcg10x000612:52
rcg1btw this is all on an n95012:52
DocScrutinizer>>4.1.1.1.3 FW_VERSION: 0x000212:53
DocScrutinizer>>Instructs the fuel gauge to return the firmware version to addresses 0x00/0x01.The bq27520-G2 firmware version returned is 0x0311.12:53
rcg1and btw.. looking further at the data sheet i get the impression that INITCOMP should be set12:54
rcg1but in fact when i query the status i get plain 012:54
DocScrutinizerhttp://forum.meego.com/showthread.php?t=461012:55
DocScrutinizerdpkg -L bme12:56
DocScrutinizer *candidate: /usr/lib/hwi/ic/bq27521.so12:56
DocScrutinizer12:56
DocScrutinizer closest hit: http://www.ti.com/product/bq27520-g2 - pinout matches with N9 schematics12:56
rcg1yeah, that should be exactly the chip..12:57
rcg1but i dunno how this information can help me12:57
DocScrutinizereither try to find datasheet for bq27521, or hope for differences from 27520 to 27521 are as small as between all the bq2415x12:58
rcg1aye12:58
DocScrutinizerobviously at least COMMAND 0x0001 matches ;-)12:59
rcg1but is there any way the /usr/lib/hwi/ic/bq27521.so lib can be of any help?12:59
DocScrutinizerif you find source code for that lib, sure you hit the jackpot12:59
DocScrutinizerif you find API, almost as good13:00
rcg1alright13:00
jarkko^http://i.imgur.com/Mev10.jpg13:00
rcg1interesting enough this lib is also there for mer13:00
jarkko^those vandals!13:00
StskeepsBostik: how do you feel about not having qtwidgets?13:01
DocScrutinizerrcg1: if you find bq27521.so src somewhere in the internets, please tosss me a link13:02
rcg1hmmz.. seem to be binaries :/13:03
rcg1"Copyright (c) Nokia Corporation 2010 All Rights Reserved."13:03
rcg1so no source13:04
DocScrutinizer:nod:13:06
DocScrutinizercurse Nokia13:06
*** damiano has quit IRC13:06
DocScrutinizerit seems to me the datasheet is so obfuscated because of mixing basic I2C operation principles with chip specific register and data descriptions13:07
rcg1hmm..might be13:07
rcg1well, if for one got a slight headache of focusing on mystical hex dumps13:08
DocScrutinizer7 COMMUNICATIONS13:08
rcg1gonna go to the beach..13:08
DocScrutinizeris quite what I'd expect to see from a wikipedia article about I2C basics13:08
* DocScrutinizer plans for sth similar (alas no beach here)13:09
rcg1alright, then have fun as well :)13:15
rcg1might mess with this chip some time later.. will as well update the bug report with some of our findings :)13:17
rcg1thanks again for the help, DocScrutinizer13:17
*** sonach has joined #mer13:19
rcg1bbl13:22
*** rcg1 has quit IRC13:22
*** sonach has left #mer13:27
*** Alison_Chaiken has joined #mer14:10
*** harbaum has joined #mer15:16
BostikStskeeps: you mean qt5-qtwidgets ?15:24
BostikStskeeps: quite a number of packages depend on those, since qtgui was split into -gui and -widgets15:24
*** dod has joined #mer15:28
*** harbaum has quit IRC15:28
*** CosmoHill has quit IRC15:32
*** tsdedst has quit IRC15:34
*** boar has quit IRC15:37
*** wmarone_ has quit IRC15:41
*** wmarone_ has joined #mer15:42
*** tsdedst has joined #mer15:46
*** boar has joined #mer15:49
*** CosmoHill has joined #mer15:51
*** slaine has joined #mer15:54
CosmoHillhi slaine15:54
StskeepsBostik: reason i'm asking is because i ran into qtwidgets dep'ing on libx1115:56
*** boar has quit IRC15:57
StskeepsBostik: i btw have a custom resolution virtualbox image with mer, busybox, qt5 and framebuffer eglfs image with llvmpipe capable of doing 60fps in qml215:58
Stskeepseven on my non-VT machine15:59
*** wmarone has joined #mer16:08
*** wmarone_ has quit IRC16:08
*** rcg1 has joined #mer16:16
rcg1re16:17
Stskeepswb16:17
rcg1thx16:29
*** InformatiQ has quit IRC16:37
rcg1oh my... this is going to take some time16:44
rcg1at least i found two fields that seem to increase when usb is connected16:44
rcg1;)16:44
rcg1and two fields that keep decreasing16:50
rcg1this actually looks promising16:50
rcg1the trick seems to read words instead of bytes... the data read using words makes much more sense16:51
*** arcean_ has quit IRC16:54
*** arcean_ has joined #mer16:54
BostikStskeeps: hmh, that might be a problem ... I can have a look at the sources tomorrow - perhaps I can patch out the x11-specific widget parts17:05
rcg1oh my.. there are strange things going on in that chip17:10
StskeepsBostik: yeah, w00t traced it down to some issue with tray icons17:18
w00tsrc/widgets/util/util.pri is where I think it goes wrong17:20
w00tthe X systemtray implementation pulls in CONFIG += x1117:20
*** trbs has joined #mer17:24
* Stskeeps has heaps of fun with locales today.17:25
* CosmoHill eeps at the repair price for a bike17:26
*** arcean_ has quit IRC17:36
*** InformatiQ has joined #mer17:40
*** diorahman has joined #mer17:40
diorahmanmy n950 doesn't give any response to flasher -i, is there a way to re-flash the bootloader (NOLO?) via other ports except USB?17:41
Stskeepsif you overwrote NOLO you're pretty SOL17:42
Stskeepsanyway, #harmattan might know mnore17:42
Stskeepsmore17:42
*** furikku has quit IRC18:01
*** InformatiQ has quit IRC18:12
*** swerden has quit IRC18:17
*** swerden has joined #mer18:19
*** peavey has quit IRC18:19
*** harbaum has joined #mer18:25
*** diorahman has quit IRC18:28
*** dod has quit IRC18:40
*** slaine has quit IRC18:45
rcg1hmm.. well i think you can scratch what I said about words vs. bytes18:53
rcg1that was me being plain stupid18:53
Bostikw00t: thanks for the pointer, I'll look at that tomorrow18:58
Bostika simple patch might be all that's needed for mer18:58
w00tBostik: well.. it's easy to patch around, but not so simple to patch without breaking that class18:59
Bostikfigures18:59
*** Behold has joined #mer19:00
Bostikif all source code in the world were to put as a map, Qt would be dead center of one of the areas marked "Here be dragons"19:00
Stskeepsno, that'd be powervr SGX19:00
Stskeeps:P19:00
*** BeholdMyGlory has quit IRC19:03
*** BeholdMyGlory has joined #mer19:06
Bostikhaven't looked inside that one, I only know what their regular kernel driver has eaten19:06
*** Behold has quit IRC19:06
*** rcg1 has quit IRC19:08
w00tBostik: i know my way around, so i'm not that bothered ;)19:09
*** poka has joined #mer19:16
*** arcean has joined #mer19:18
*** slx has joined #mer19:20
*** peavey has joined #mer19:21
*** swerden has quit IRC19:22
*** jstaniek has joined #mer19:32
*** poka has quit IRC19:38
*** sirdancealot has quit IRC19:42
*** arcean has quit IRC19:49
*** arcean has joined #mer19:50
Bostikhmm... the actual rule in widgets/util/util.pri says that "if XCB is available, use x11; otherwise use QPA"20:01
Bostikmight be enough to just do a rebuild with reduced build-deps20:01
*** kthomas_vh_ has joined #mer20:02
*** onekenthomas has quit IRC20:02
w00tStskeeps: http://qtl.me/mer14.png20:04
w00tSleepy_Coder: ^ since you were wondering about scaling20:04
Stskeepsnot bad20:04
Sleepy_Coderpurrdy :D20:04
* w00t plans to throw a gradient in there to make it prettier20:05
Bostik....althought XCB is used for more than just that, so maybe just patch the x11-case away and use qpa unconditionally20:05
Sleepy_Coderhmm wait, can widgets advertise to what aspect ratios they can scale?  the 1x2 in the lower left looks like it's chopped off? or... :o20:05
Bostikah well, for tomorrow20:05
Sleepy_Coderalso the 2x2 in the upper center looks like it's centered vertically in the upper 2 cells and just kind of hanging down over the lower 2 :x20:06
Stskeepsyeah, nite from me too20:06
Sleepy_Coderoh guess not20:07
w00tSleepy_Coder: the aspect ratio is up to the author20:07
w00tI just hacked it to use multiple ratios20:07
w00tand (by luck) it chopped off the date ;)20:07
Sleepy_Coderthe clock widget seems to hang in the top of its cell, so I thought it was being improperly centered, I'm weird20:07
Sleepy_Coderthe 1x1 version of it in the upper left gave me the clue20:07
Sleepy_Coderso me leiky :D20:07
*** niqt has joined #mer20:09
*** harbaum has quit IRC20:25
*** drussell has quit IRC20:54
*** tommis has quit IRC21:03
*** phaeron has quit IRC21:08
*** gimli has quit IRC21:27
*** trbs has quit IRC21:29
*** lardman|home has joined #mer21:31
*** lardman has quit IRC21:31
*** sandroandrade has joined #mer21:47
*** wmarone has quit IRC22:00
*** CosmoHill has quit IRC22:10
*** erbo_ has quit IRC22:17
*** erbo has joined #mer22:18
*** tsdedst has quit IRC22:23
*** speakman_ has quit IRC22:28
*** bef0rd has joined #mer22:28
*** bef0rd has joined #mer22:28
*** wmarone has joined #mer22:30
*** rcg1 has joined #mer22:31
*** niqt has quit IRC22:33
*** speakman has joined #mer22:35
*** speakman has joined #mer22:35
*** NIN102 has quit IRC22:36
*** cxl000 has quit IRC22:38
*** jstaniek has quit IRC23:04
*** bugzy has quit IRC23:06
*** kurtul has joined #mer23:26
*** arcean has quit IRC23:29
*** kurtul has quit IRC23:33
*** M4rtinK has quit IRC23:35
*** bugzy has joined #mer23:38
*** selesnie_ has quit IRC23:46
*** selesnie has joined #mer23:46
*** Bostik has quit IRC23:46
*** bugzy has quit IRC23:53
*** bugzy has joined #mer23:54
*** ALoGeNo has quit IRC23:58
« Saturday, 2012-05-26 Index Monday, 2012-05-28 »

Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!