IRC log of #maemo-ssu for Wednesday, 2013-03-27

*** xes has joined #maemo-ssu00:07
*** andre__ has quit IRC00:28
*** Skry has quit IRC00:29
*** Skry has joined #maemo-ssu00:31
*** dhbiker has quit IRC00:38
*** Vlad_on_the_road has quit IRC00:43
*** Woody14619 has quit IRC01:31
*** Woody14619 has joined #maemo-ssu01:31
*** Woody14619 has quit IRC01:31
*** Woody14619 has joined #maemo-ssu01:31
*** unclouded has joined #maemo-ssu01:58
*** arcean_ has joined #maemo-ssu02:00
*** arcean has quit IRC02:01
*** _LauRoman has joined #maemo-ssu02:10
*** LauRoman has quit IRC02:13
*** LaoLang_cool has joined #maemo-ssu02:16
*** LaoLang_cool has quit IRC02:29
*** M4rtinK has quit IRC02:31
*** arcean_ has quit IRC02:36
*** kolp has quit IRC02:37
*** xes has quit IRC03:32
*** _LauRoman has quit IRC04:05
*** discopig has quit IRC04:22
*** discopig has joined #maemo-ssu04:22
*** discopig has quit IRC04:22
*** discopig has joined #maemo-ssu04:22
*** Pali has quit IRC04:48
*** amiconn has quit IRC05:39
*** amiconn_ has joined #maemo-ssu05:39
*** amiconn_ is now known as amiconn05:39
*** DocScrutinizer05 has quit IRC06:03
*** DocScrutinizer06 has joined #maemo-ssu06:03
*** DocScrutinizer06 is now known as DocScrutinizer0506:03
*** chainsawbike has quit IRC06:20
*** chainsawbike has joined #maemo-ssu06:21
*** M13 has joined #maemo-ssu06:47
*** amiconn has quit IRC09:09
*** amiconn has joined #maemo-ssu09:10
*** M13 has quit IRC09:47
*** andre__ has joined #maemo-ssu10:11
*** dhbiker has joined #maemo-ssu10:19
Estel_DocScrutinizer05,  extremely usefull info here10:26
Estel_pity that you won't put it on wiki, as it will gest lost, and in 3 weeks no one will remember it, too10:26
Estel_(the things about bmi, shutdown graceful and not graceful levels, etc)10:26
Estel_kerio,  and what are your experiences with no compression? :P10:27
Estel_kerio,  what do you think would it be much code change to stop behavior of missing sysfs nodes when not calibrated?10:28
Estel_in module?10:28
Estel_maybe forking it and 99% of sensible users using fork would convince Pali to stop selling bulls about "send patch to mainstream"10:28
Estel_if it's small change, maintaining it over next versions of bq27x00_battery would be easy10:29
*** Martix has joined #maemo-ssu10:34
*** M4rtinK has joined #maemo-ssu10:54
*** LauRoman has joined #maemo-ssu11:16
*** kolp has joined #maemo-ssu11:22
*** M4rtinK has quit IRC11:31
*** M13 has joined #maemo-ssu11:52
*** lizardo has joined #maemo-ssu12:26
*** Pali has joined #maemo-ssu12:38
*** unclouded has quit IRC12:55
*** sunny_s has joined #maemo-ssu13:09
*** LauRoman has quit IRC13:24
*** futpib has joined #maemo-ssu13:43
*** futpib has quit IRC14:26
*** sunny_s has quit IRC14:59
Paliping Estel_15:19
Paliping kerio15:19
PaliI updated bme replacement wiki15:21
Palithere is one question (at wiki): what should battery applet show when battery is not calibrated and is configured to not use rx51_battery driver15:22
kerio"<charge>/<full charge> (CI)"15:22
kerioit's already written in my proposal afaik15:22
*** xmlich02_ has joined #maemo-ssu15:37
*** xmlich02_ has quit IRC15:37
discopighi17:46
Palikerio, small update which allow to do not use design capacity in status menu applet18:31
Palipackage is in cssu-devel18:31
kerioyay18:32
keriowhat does it do if bq27k isn't calibrated?18:32
kerio0/0?18:32
PaliEstel_, I fixed hald-addon-bme, now it does not power off device when drivers are unloaded18:32
Palikerio, it will show message that data are no available or battery is not calibrated18:33
kerioand does the usual guesstimate for bars and percentage? alright18:33
Paliyes18:33
Palikerio, changes are on gitorious18:34
keriogitorious where?18:34
* kerio doesn't grok gitorious' interface18:35
Palimerlin1991: I pushed new version of rtcom-messaging-ui-portrait to cssu-devel. I only changed version string to 1.0-1 and it fixed that infinity upgrade loop18:59
Palimerlin1991: add this version of package to cssu-testing list18:59
*** NIN101 has joined #maemo-ssu19:21
merlin1991Pali: I'll do19:24
*** BCMM has joined #maemo-ssu19:37
*** luf has joined #maemo-ssu19:42
Estel_Pali,  thanks20:00
Estel_Pali,  is battery status applet also on your page, for non cssu-devel users? of course, I can install it from cssu-d, too, if not20:00
Estel_Pali,  so there remains one problem - about missing sysnode. You're not right, that it should be adressed by mainstream20:01
Estel_look how busybox-power does it - it fixes things for maemo, *then* send patches to upstream20:01
Estel_then, they're accepted in upstream20:01
Estel_DocScrutinizer05,  thought you may be interested in it (guy have problems with maemo repos)20:02
Estel_http://talk.maemo.org/showpost.php?p=1331800&postcount=6320:02
PaliEstel_, please read that email thread about bq27x al lkml20:03
Estel_ok, any highlights to look for?20:03
Palihttps://lkml.org/lkml/2013/1/19/16220:03
Palilink is on wiki20:03
Estel_btw, I'm mainly interested about our maemo case, not what mainstream does. It's good practice to fix for maemo, and send upstream - if they accept it it's fine, if not, fine too20:03
Estel_ah20:03
Estel_thanks20:03
Palihttps://wiki.maemo.org/Bme_replacement20:04
Estel_most of the times, they're much likely to accept it mainstream, if practically used "live" on some devices20:04
Estel_Pali,  seen mail on lkml, can't see hpw it is related20:05
Palicorrect solution is to use debugfs20:06
Estel_he talk about i2cget for userland, we fat about bq27x00_battery stopping censoring info for no reason20:06
Estel_s/fat/talk/20:06
infobotEstel_ meant: he talk about i2cget for userland, we talk about bq27x00_battery stopping censoring info for no reason20:06
Paliafk20:06
Estel_correct solution is to comment out those damn 3 lines of code, that make it censor data for no practical or theoretical reason20:07
Estel_so it always export to sysfs node what hardware tells it, period20:07
Estel_someone who thought it's good idea to censor that data when not calibrated, was fucked in head20:07
Estel_I must go off-line now, sorry20:07
DocScrutinizer05yes20:08
*** freemangordon has quit IRC20:25
Estel_Pali,  as a suplement - you've talked with every user of your early stage bme replacement. Everyone think censoring that sysfs node is freakin bad idea20:40
Estel_that must mean something. BTW, if me and DocScrutinizer05 speaks the same voice, it means even more :P20:40
kerioeven a broken clock etc. etc.20:41
kerioPali: hmm, is the default to not use design now?20:51
Palikerio, when battery is not calibrated, see wiki20:51
kerioto *not* use design i said20:52
keriooh i see20:52
*** M13 has quit IRC20:53
*** arcean has joined #maemo-ssu21:29
*** BCMM has quit IRC21:50
*** futpib has joined #maemo-ssu22:04
*** XDS2010 has quit IRC22:05
*** luf has quit IRC22:15
Estel_Pali,  new hald-addon-bme is in your page?22:34
PaliEstel_, yes, also dsme plugin22:35
Estel_any more packages updated?22:35
Estel_thanks22:35
Estel_installing now22:35
Estel_~pali22:35
infobotfrom memory, pali is http://atrey.karlin.mff.cuni.cz/~pali/22:35
Palidsme plugin now fallback to bq27x temperature if rx51 is not available22:35
Estel_new batt applet from cssu-dev only?22:35
Estel_nice22:35
Estel_~bme-replacement22:36
infobotbme-replacement is probably http://atrey.karlin.mff.cuni.cz/~pali/rx51-bme-replacement  http://atrey.karlin.mff.cuni.cz/~pali/projects/maemo/bme-replacement.html  See also: http://wiki.maemo.org/Bme_replacement . Please, use wiki page to report bugs/problems and/or solutions to them!22:36
Paliyes, battery plugin is in cssu-devel22:36
Estel_Pali,  and what about shutdown, was it fixed?22:37
Estel_edv1 and co?22:37
Palinot implemented yet22:37
Estel_I want to update wiki if smth is solved22:37
Estel_OK22:37
Paliit shutdown at edv122:37
Estel_but one can unload modules22:37
Paliyes22:37
Palihttps://gitorious.org/community-ssu/status-area-applet-battery/commit/3d3599743a598a833c1ecfa886ebf53fb6d11a1222:37
Estel_all right, I want to maintain wiki up to date, as no one else seems to be interested in it (and you're, rightly so, busy with coding)22:38
Estel_please, Pali, include that gconf value to set up own shutdown limit, too :P22:40
Paliok, later22:40
Estel_best would be to make it inclusive or exclusive with edv122:40
Estel_so, 2 gconf values22:40
Estel_one for treeshold voltage22:41
Estel_and one that triggers behavior of watching for shutdown voltage only after edv1 flag22:41
Estel_so, for example, 3000 mV shutdown from first gconf value, but triggered only, id edv1 flag was set (if 2nd gconf value, for example, "wait_for_edv1" is 1)22:42
Estel_or doesn't care about edv1 flag, if set otherwise22:42
Estel_it allow to rise voltage for devices with gsm problem, or lower it for dual-cells, ignoring spikes (due to waiting for edv1)22:42
Estel_Pali ^22:42
PaliEstel_, next simple patch will be to replace #define VOLTAGE with value from gconf22:43
Estel_OK22:43
Estel_looks great22:43
Estel_but you said voltage will be significant only, if edv1 flag isn't exported by kernel22:43
Estel_it would be nice to have this gconf value working even, when kernel properly exports edv1 flag22:44
PaliI will add gconf value to disable edv122:44
Estel_ok22:44
Pali(I wrote that on wiki)22:44
Estel_so I'm talking about - it's good idea to make it inclusive, no exclusive22:44
Estel_so one could have both $voltage from gconf AND edv1 gconf22:45
Estel_= shutdown at $voltage, but only if edv1 flag is set22:45
Paliinclusive is simple, exclusive will be hard to write22:45
Estel_or, when edv1 gconf is 0, shutdown at $voltage (ignoring edv1 totally)22:45
Estel_I think most people that are willing to change shutdown voltage would like to have it inclusive with edv122:46
Estel_to ensure that low voltage spike won't make device shutdown before edv1 flag22:46
Estel_voltage can drop even to 3000 mV, without edv1 flag still in place, as it needs 15 seconds22:46
Estel_Pali,  I'm glad inclusive will be easy to write, but it will work for people that want RISING voltage, too?22:47
Estel_for example, shutting down at 3300 mV?22:47
Estel_(for people with gsm modem problems)22:47
Estel_edv1 will be never set, then, so 3300 mV won't work without exclusive22:48
Estel_flexibility is key here, to less rewrites later22:48
Estel_when everyone will be able to set whatever fancy scenario, no one will pester you about features/bugs, later :P22:49
Estel_Pali,  I've seen your info about i2c_slave22:51
Estel_problem is, no one know how to use it :P22:51
Estel_properly22:51
PaliEstel_, what?22:51
Estel_from wiki, you said that problems can be solved by using i2c_slave22:51
Estel_I2C_SLAVE_FORCE22:51
Paliyou can use i2cget without problem with force flag22:52
Estel_no idea how to do it, properly, as no one want to access i2c wrong way and fuck something up22:52
Estel_I see22:52
Paliuse *only* for bq2720022:52
Estel_so it's just a matter of proper i2cget command syntax?22:52
Estel_yes22:52
Palii2cget -f22:52
Pali(check if -f is really force, but I think yes)22:53
Estel_and it stop complaining about resource in use?22:53
Estel_ok22:53
Estel_well, lemme check it, if I dissappear, it mean something went wrong :P22:53
Paliand double check if you are sending command to bq27200 address22:53
Estel_you're right, it works22:54
Estel_how to determine that addres is bq27200?22:54
Estel_address*22:55
Estel_0x5522:55
Estel_is bq27200?22:55
Estel_it solves many problems, then22:56
Estel_I still think sysfsnode shouldn't censor data, but I can at least workaround it by using 82cget directly22:57
Estel_not ideal, but at least works22:57
Estel_I appreciate you work on this, Pali, and I'm glad I could contribute by at least wiki22:58
Paliok22:58
Estel_just don't, *ever*, write "send patch to upstream", when we report something valid :P It makes people go berseker22:58
*** lizardo has quit IRC23:02
*** xes has joined #maemo-ssu23:09
DocScrutinizer05>>dsme plugin now fallback to bq27x temperature if rx51 is not available<< WTF? bq27x temperature is as meaningless as my little toe's temperature23:14
DocScrutinizer05actually tempreature from bq27200 is the most useless and void-of-any-meaning value you can read out from that chip23:15
*** Vlad_on_the_road has joined #maemo-ssu23:16
Estel_DocScrutinizer05,  except for the fact, that, in reality, it's always same as rx51 temperature :P23:17
Estel_i.e. difference never exceeds 1C23:17
Estel_and rarely ever exceeds 0.5C23:17
Estel_also, nothing ever use that temperature for any real-life purpose, so we can safely ignore it? or I'm wrong?23:18
* ShadowJK would expect bq27 temp to be close to that of the pcb it sits on23:19
DocScrutinizer05yeah, and temperature difference between my living room and outside rarely ever is >2°C (in spring and autumn, when my windows are opened wide) - MEH, the one is battery tempreature and the other one some bogus bullshit23:19
Estel_I think that Pali followed our suggestion, that presenting not-accurate data is better than no data at all :P23:19
Estel_why bogus?23:19
DocScrutinizer05because that chip is not even near to battery cell23:20
DocScrutinizer05and no, that value isn't useless or unused, bme uses it for batterty safety monitoring I'd say23:20
Palidsme needs some value and what is better? some hardcoded? some reported by bq27200?23:20
ShadowJKI don't think either are very close to the batteri23:20
DocScrutinizer05dang, my ignore filter failed23:20
Estel_Pali,  definitelly reported by bq27200 is better than hardcoded23:21
ShadowJKBut the thermistor attached to madc is first source, right?23:21
Estel_DocScrutinizer05,  not very mature to use some bullshit ignore filter as argument in civilized discussion, really *shrug* feel free to leave conversation, though23:21
DocScrutinizer05for cell temperature? there's only ONE "better", the theristor labeled "battery temp" in schematics23:21
Estel_ShadowJK,  yes23:21
ShadowJKEstel_; right23:22
Estel_fallbacks to bq27200 only if rx51 one isn't available23:22
ShadowJKSo I'd agree that bq27 is better than nothing23:22
ShadowJKFor sure better than the one in omap :)23:22
Estel_DocScrutinizer05 fails to understand, that dsme need *some* temperature to function23:22
DocScrutinizer05this is a safety relevant design detail, you shouldn't mess with it23:22
Estel_so if not bq27200, we would use dummy hardcoded palceholder23:22
Estel_placeholder, even23:23
ShadowJKAs long as madc continues to function, and battery temp thermistor stays attached, dsme will use exactly the same data as before, right?23:23
Estel_ShadowJK,  yes23:23
ShadowJKright..23:24
DocScrutinizer05ShadowJK: if those guys can't get the ADC-4(?) readout, then for sure they should SHUT DOWN that crap, and not try to handle LiIon cell based on some guestimate about cell temperature23:24
Estel_can't agree, no temperature can be due to unloaded module or whatever23:24
Estel_I want to be able to unload some module for debugging or whatsnit, without device shutting down in flames, because someone thought it's good idea to hardcode it.23:25
*** XDS2010 has joined #maemo-ssu23:26
ShadowJKThe battery temp sensor is kinda useless anyway, with that big air gap (or was it air gap + shield?) between it and battery :/23:26
Estel_agreed.23:26
Estel_air+ shield, in fact23:27
DocScrutinizer05not at all, the battery temp sensor is in contact with shield which in turn is in *tight* contact with battery23:27
Estel_sure, with insulating layer between. and definitelly not "tight"23:27
Estel_only springs are tight, without them battery can fell of on it's own23:28
Estel_s/can/will always/23:28
infobotEstel_ meant: only springs are tight, without them battery will always fell of on it's own23:28
Estel_springs, otoh, touch even thicker plastic insul. layer23:28
ShadowJKOh I remembered it as the batt thermistor not touching anything at all except for the point it's soldered to the pcb23:28
Estel_Because it's that fckin way23:29
ShadowJKah lol23:29
* DocScrutinizer05 feels like watching those experts who go like "that fuse is useless anyway, so let's use a nail to short it"23:29
Estel_IDK why DocScrutinizer05 thinks it's attached to shield23:29
Estel_in contact = may accidentaly touch it, from time to time, or not at all23:30
Estel_before it would detect any temp. change, it would be dissicipated in whole ground plane already23:30
ShadowJKhm, if I used my 1000mR battery, I'd get some mean heat dissipation, wonder if that would provoke bigger temperature gradients23:30
Estel_and bq27200 temp sensor would know it already, too23:30
*** Vlad_on_the_road has quit IRC23:31
Estel_belive me, when I got some overheating batteries, that rx51 sensor was last thing that knew about it23:31
Estel_I first felt it through backcover, lol23:31
Estel_30 seconds later sensor started to notice slowly rising temp23:32
ShadowJKNever had a situation like that myself23:32
Estel_when backcover was 60C already23:32
ShadowJKExcept on N81023:32
Estel_I had some funny batteries around23:32
Estel_Pali,  you said unloading both bq27x00_battery and rx51_battery won't cause reboot now. So dsme will work despite no temp. readings?23:33
PaliEstel_, I tested it and no reboot, but I do not know what will dsme do if it will not receivce temperature for 1hour/1day...23:34
Estel_btw, keeping fuse analogy... we're talking about "if that fuse fails, let's redirect it through another fuse", instead of using hardcoded value, with would be exactly using a nail.23:34
DocScrutinizer05oh fine. some self-certified experts declare battery sensor useless and thus suggest to replace its readout by a hardoced value or even some unrelated bogus shit. FINE, will you next redefine cmt to be a LTS modem?23:35
Estel_Pali,  OK.23:35
Estel_DocScrutinizer05,  you're talkung bullshit, no one replaces anything.23:35
Estel_we use 2nd value, when 1st value is unavailable.23:35
ShadowJKNew behaviour identical to previous in all normal conditions23:36
Estel_1st - marginal chances of that ever happening 2nd - if that ever happens, it's better than use hardcoded thing not related to anything23:36
Estel_ShadowJK,  exactly.23:36
PaliDocScrutinizer05, write solution what to do when rx51_battery driver is not loaded23:36
DocScrutinizer05except that sensor is exactly NOT about _normal_ situations23:36
Paliand reboot/shutdown device is not solution23:36
Estel_DocScrutinizer05 preffer shutdown, then.23:36
Estel_I don't agree23:36
Estel_I would like that energy spent on improving bme replacement wiki, or convincing that cripplig sysfs nodes is bad idea :P23:38
Estel_instead of arguing about bullshit temp. readout, which, in practical life, is completely irrelevant (meaning anything only in theoretical situations, and even there, discussably)23:39
DocScrutinizer05btw wtf is adc-x via anything bme related?23:40
Estel_Pali,  version number of packages haven't changed on your page, but content changed?23:41
Paliyes23:42
Palilook at date23:42
Estel_hald-addon-bme and dsme-thermaobject, yes?23:42
Estel_ok23:42
Estel_if you would, in future, bump versions, it would be less asking you about it23:42
Estel_just a suggestion23:42
DocScrutinizer05and wtf would it ever be "not available"??23:43
Estel_DocScrutinizer05,  due to rx51_battery module being not loaded.23:43
* DocScrutinizer05 pukes a little while heading out23:44
Estel_take your time and have all pleasure ;)23:45
*** LauRoman has joined #maemo-ssu23:48
*** freemangordon has joined #maemo-ssu23:50
*** freemangordon_ has joined #maemo-ssu23:50
*** freemangordon_ has left #maemo-ssu23:50
DocScrutinizer05on a sidenote: if nokia had thought bq27200 temp was *any* useful for battery overtemp/undertemp(!) protection, don't you think they *happily* had saved that thermistor from the BOM?23:52

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