libera/#maemo-leste/ Tuesday, 2022-10-04

Wizzupfreemangordon: did you end up mailing nikolaus wrt sgx userspace builds?00:35
freemangordonWizzup: no, will do today07:09
buZzuvos__: could you add those two middle tabs on maseratiBatteryAdapter.tar.xz ? then we could maybe turn it into a product, so people could add a thermistor thats the right value, maybe some BMS for safety , and easily adopt new cells07:36
buZzwhen i say product, i just mean some howto send to to pcbway/something to order flex pcbs yourself :)07:37
uvos__buZz: not so easy08:55
uvos__from memory: the tabs have to be raised relative to the screw holes08:56
uvos__you cant really do that with a pcb, you would need to add pads to the pcb and then solder some copper blocks to those or something like that08:56
uvos__also the way the current design works08:57
uvos__you have to have the copper layer up towards the battery door08:57
uvos__so you can solder the batery wires to the pcb08:57
uvos__if you flipp it over it would short against the rf shield08:58
freemangordonuvos__: re ofono logs - did you set both variables I told you?08:58
uvos__so you would have to do a 2 layer design08:58
uvos__freemangordon: yes i added -d /etc/default/ofono*08:58
uvos__and the envvar to /etc/init.d/ofono08:58
uvos__did it not work?08:58
freemangordonas I see only qmy logs, but not at ones08:58
uvos__ok08:59
uvos__hmm08:59
uvos__maybe i did something wrong08:59
uvos__i dident know what your soupposed to see in the log08:59
freemangordonwhen you have both set, one can see both qmi and at traces in ofono.log08:59
freemangordonnow it is very hard (if possible at all) to match ofono vs n_gsm logs09:00
uvos__cat /etc/default/ofono09:00
uvos__DAEMON_OPTS=-d09:00
uvos__cat /etc/init.d/ofono09:01
uvos__...09:01
uvos__export OFONO_QMI_DEBUG=109:01
uvos__something missing?09:01
freemangordonyes, I said two vars :)09:01
uvos__oh ok09:01
uvos__whats the other one?09:01
freemangordonsec09:01
buZzuvos__: maybe one of those pcb houses that also does manufacturing can add something liek that?09:02
uvos__sure most can09:03
uvos__its called pcb assembly09:03
buZzright09:03
uvos__and is quite expensive09:03
freemangordonuvos__: https://libera.irclog.whitequark.org/maemo-leste/2022-09-27#33042805;09:04
buZzheh yes, 250 usd for 10x 40x10mm flexpcb with stiffener and 2 SMD parts (thermistor and whatever lump of copper?)09:07
buZzoh wait, without pcb assembly the 10x 40x10mm flexpcb are 210 usd09:08
buZzmaybe i have expensive options selected :P09:08
buZz~135 usd for 100x 40x10mm flexpcb, 2 layers, 0.23mm with a 0.1mm stiffener09:15
buZzand assembled with 2 unique parts -> 540 usd for 100 pieces09:15
buZz¯\_(ツ)_/¯09:15
buZzuvos__: what about just 'solder something of copper here' as instruction? :P09:16
uvos__sure09:18
uvos__very small though09:18
uvos__for agv people09:18
uvos__freemangordon: ok ill regrab the logs in ~2h09:18
freemangordonok09:19
uvos__freemangordon: i gues for the sms log its not relevant09:29
uvos__as dmseg dosent show any at commands during this time?09:29
buZzuvos__: i'm not against distributing such pcbs, and also feel able to solder blobs and thermistors to it12:08
buZzbut not sure if thats the best goal to target, but 100 pcbs are the same price as 1012:08
freemangordonuvos__: if possible, clean the old logs before capturing new ones12:24
uvos__freemangordon: i got the logs12:25
uvos__i deleted the ofono log right before12:26
uvos__they are in te same place on my host12:26
uvos__the *2.log ones12:26
uvos__the logs are from boot12:27
freemangordonthanks12:27
uvos__through sim unlock (yes i misstyped the pin once - this is not important)12:27
uvos__then it connects12:27
uvos__and then it looses signal12:27
uvos__(not pictured) restarting ofono at that point gets the signal back12:28
uvos__i presume i dont need to recapture the sms problem (incomeing sms are allawys stuck in pending)12:29
uvos__since this dosent seam to touch the at interface at all12:29
uvos__OUTGOING not incomeing sry12:29
freemangordonuvos__: so, it does not understand when the signal is recovered?12:31
freemangordonis this the major issue?12:32
uvos__it never reconnects unless a restart is issued12:32
uvos__so either 1. its not instructing the modem to try and search for a netwok again (but why dose qmictl show one?), 2. it dosent realise the modem has reconnected12:33
uvos__3. the modem never looses signal at all12:33
uvos__if its 1 or 2, why dose the modem allways loose network registration right after ofono starts?12:33
uvos__(the first time)12:34
freemangordondo you have some timestamp of the lost signal?12:35
uvos__hmm not sure how to do that12:35
uvos__maybe tirgger on some dbus singal?12:35
freemangordonok, no issue12:35
uvos__ok so looks like the network is indeed lost right after ofono starts12:36
uvos__if you watch qmicli12:36
uvos__it goes to state12:36
freemangordonwhat state?12:36
uvos__not-registerd-seaarching for a while12:36
freemangordonok12:36
uvos__and then it reconnects (as shown by qmicli)12:37
freemangordonI guess it is this "Oct  4 12:20:51 localhost ofonod[3308]: drivers/qmimodem/network-registration.c:event_notify() signal with -1%(-125 dBm) on 4"12:37
uvos__(possible also messing with qmicli here changes the result)12:37
uvos__since it wakes the modem12:37
freemangordonand then "drivers/qmimodem/network-registration.c:extract_ss_info() roaming 2 lac -1 cellid -1 tech -1"12:38
freemangordonsrc/gprs.c:netreg_status_changed() 2 (searching)12:38
freemangordonok, will have a look12:39
uvos__the same thing also happens later12:39
uvos__if i really do loose connection by leaving the range of a tower12:39
uvos__it dosent seam to ever find a new one after that12:40
freemangordonok12:40
uvos__i again have to restart ofono for it to get it back12:40
freemangordonI will try to re-create by putting it in the fridge12:40
freemangordonor wrapping with a foil12:41
uvos__microwave works well12:41
uvos__(just dont turn it on :D)12:41
freemangordonuvos__: oh, btw, is there aby way to have a stable usb connection?12:41
freemangordon*a way12:41
uvos__usb otg?12:41
freemangordonno12:41
uvos__or slave12:41
freemangordonslave12:41
freemangordonwith networking12:41
freemangordonit disconnects here every now and then12:42
uvos__there is some issue with it dissconnecting when cpcap-charger changes charge state12:42
freemangordonmaking it a nightmare to debug radio issues12:42
uvos__maybe interferance12:42
freemangordonmaye rmmod cpcap-charger?12:42
uvos__just dont use cpcap-charger and power the device some other way12:42
uvos__freemangordon: just disable charge current in sysfs12:42
freemangordonok12:43
freemangordonthanks12:43
uvos__is sufficant12:43
uvos__as a workaround12:43
freemangordonthis is /usb node, right12:43
freemangordonthe limit12:43
uvos__right12:43
freemangordoncharge_current_limit?12:43
freemangordonok12:43
freemangordonttyl12:43
uvos__pretty sure yes12:43
uvos__ttyl12:43
freemangordonhmm, cannot make it lose signal :(16:50
freemangordonwripping it in foild does not help16:50
freemangordon*wrapping16:50
freemangordonfridge does not help either16:56
freemangordonuvos: please unlock the device, stop mce and try to repro17:08
freemangordonthough I don't see SCRN command17:09
freemangordonhmm17:09
freemangordonstill, make sure mce has send SCRN at least once17:11
freemangordonand then unlock, stop mce ane see if it still the same17:12
freemangordonuvos: what "and then it looses signal" is supposed to mean? there are no bars in hildon-status menu? or the place there of gsm is empty?17:58
freemangordon*fir gsm17:58
freemangordonuvos: oh, ok, I can reproduce it by doing  AT+SCRN=0 and then switching the technology18:05
freemangordon/usr/share/ofono/scripts/set-tech-preference gsm18:05
freemangordonfor example18:05
Wizzupfreemangordon: well, what should set scrn=0 then?18:29
freemangordon"(19,16,41) freemangordon: I need to think a while how to integrate ofono with screen lock"18:30
freemangordon:)18:30
freemangordonIOW- no idea18:30
freemangordonbut kernel does not seem to copy the data to all the clients18:30
freemangordonhmm, wait18:31
freemangordonmce does not red18:31
freemangordon*read18:31
freemangordonso, scratch that18:31
uvosfreemangordon: "uvos: what "and then it looses signal" is supposed to mean? there are no bars in hildon-status menu? or the place there of gsm is empty?"18:46
uvosbars go away so dose 2g so dose the operator name18:46
uvosofono ./list-modems lists no connection18:46
uvossomeone sets SCRN=0 on android18:47
uvosidk who presumeably its moto_ril18:48
uvossince it contains the string18:48
uvosfreemangordon: it happens even when i never allow mce to turn off the display18:49
uvosso mce never sets scrn at all18:50
uvosi can also try with mce totaly disabled in a sec18:50
freemangordonuvos: make sure SCRN=1 when you disable mce19:46
freemangordonas I said earlier - unlock, stop mce and retry19:47
Wizzupuvos: btw, did you see spinals typo fixes pr?20:13
buZzhehe20:14
buZzon what now? :P20:14
Wizzupfrom a few days ago I think20:20
buZzgee, not a single page on the wiki mentions sfone21:00
WizzupbuZz: mhm, well, it's lacking quite some things (libnotify is one)21:04
Wizzup(as does conversations)21:04
WizzupI think we usually don't mention things too explicitly unless we agree they work well enough21:04
Wizzupbut maybe that's not the right approach21:04
buZzi'm planning to at least go through the extras and make pages for it21:04
buZzi think we should document what we have, as people want to use it21:04
Wizzupagreed @ extras21:07
Wizzuponce sphone hits non-devel I think we ought to document it21:07
Wizzup(well)21:07
buZzwe could already document it though21:17
Wizzupmhm21:18
Wizzupsure21:18
buZzis there a easy way to query the current CellId ?21:27
buZzi'm grepping it out of iad_conndlg.log now, but feels a bit dirty :P21:28
Wizzupyeah, I will disable that soon (the debug logs)21:29
Wizzupmdbus2 will tell you21:29
Wizzupmdbus2 -s org.ofono /motmdm_0 org.ofono.NetworkRegistration.GetProperties21:30
buZzah yezz21:33
WizzupI am not sure if it updates with SCRN=0, but I think it does21:34
buZzthats backlight disabled mode?21:34
Wizzupscreen off yes21:35
buZzah 'jq' , i keep forgetting about that tool :) (cmdline json parser)21:37
Wizzupnot sure if you can use it to parse mdbus2 output21:37
Wizzupit's not json21:37
buZzsmells alike :P21:37
Wizzupwon't work, pretty sure21:37
buZzhmmz, opening the internet connections in statusbar still disconnects gprs , but not wifi21:38
buZznot sure if thats a bug or not :P21:38
buZzhmm, org.ofono.NetworkMonitor.GetServingCellInformation actually gets less information?21:44
buZzand .GetNeighbouringCellInformation isnt implemented :( aww21:45
freemangordonWizzup: https://github.com/maemo-leste-upstream-forks/ofono/commit/cbffe1d6d3a1872db5454327a4f8c6f20419bd5f21:50
freemangordonI know you said this is a hack21:50
buZzoh man, these all look so nice to have; https://kernel.googlesource.com/pub/scm/network/ofono/ofono/+/1.22/doc/networkmonitor-api.txt21:51
freemangordonbut, I think we have another issue21:51
buZzbut we dont :( (on d4)21:51
freemangordonwhat I see returned is 2 some times21:51
buZz2 registrations?21:52
buZzoh, 2 radio_if_count ?21:52
freemangordonwhich is not defined https://github.com/maemo-leste-upstream-forks/ofono/blob/maemo-ofono/drivers/qmimodem/nas.c#L2621:53
buZzModem.GetProperties does -claim- the modem supports 'rat' :P (whatever that may be)21:53
freemangordonRadio Access Technology21:53
freemangordonWizzup: see https://chromium.googlesource.com/chromiumos/third_party/libqmi/+/factory-4455.228.B/libqmi-glib/qmi-enums-nas.h#5921:53
freemangordonso the correct patch would be to add those, no?21:54
buZzarent those https://github.com/maemo-leste-upstream-forks/ofono/blob/maemo-ofono/drivers/qmimodem/nas.h#L10221:55
buZzah yeah, i see what you mean, seems logical to add the others21:55
buZzfreemangordon: where are you receiving CDMA_1XEVDO ?21:56
freemangordonhere21:56
freemangordonwell, on ofono logs21:56
buZzhehe, what nation? :)21:56
freemangordonBG21:56
buZzoh, cool :)21:56
freemangordonthis is just for a moment when I switch from umts to gsm21:56
Wizzupfreemangordon: sorry, what do you want to add?21:57
buZzThe CDMA Development Group states that, as of April 2014, there are 314 operators in 118 countries offering CDMA2000 1X and/or 1xEV-DO service.[7][needs update]21:57
buZzmy god21:57
buZz:D21:57
buZzWizzup: extra types of QMI_NAS_NETWORK_RAT_*21:57
buZzmissing ones21:57
freemangordonthe missing values21:57
buZz(which is now hardcoded to -1 always? or something, i didnt understand that)21:57
freemangordonI think those missing is the reason you get -121:57
buZzhttps://github.com/maemo-leste-upstream-forks/ofono/commit/cbffe1d6d3a1872db5454327a4f8c6f20419bd5f21:58
Wizzupfreemangordon: it will still override the value from radio0 with the value from radio1 ,no?21:58
freemangordonyes, but IIUC radio1 is with priority to radio021:58
Wizzupif you say cdma is21:58
freemangordoncdma is before gsm/umts21:59
freemangordonso gsm/umts will override21:59
Wizzupthen my patch is not necessary21:59
Wizzupbut it is21:59
Wizzupiiuc21:59
freemangordonI don't see why it is needed21:59
Wizzupok, maybe this without and see22:00
buZzat the very least, ofono will improve from knowing the missing libqmi network types22:00
freemangordonmhm22:00
freemangordonbut I will have to map22:00
buZzwhich -might- remove the -1 output22:00
buZzyeah, in that function?22:01
freemangordonwhat is CDMA? GSM?22:01
buZzUMTS22:01
WizzupCDMA is not umts22:01
Wizzupthis is the us radio tech22:01
freemangordonit is not umts22:01
buZzCDMA2000 compares to UMTS, a competing set of 3G standards, which is developed by 3GPP and used in Europe, Japan, China, and Singapore.22:01
buZzright22:01
buZzhttps://en.wikipedia.org/wiki/CDMA200022:01
freemangordonhttps://www.electronics-notes.com/articles/connectivity/cdmaone-cdma2000/cdma2000-1xevdo-basics.php22:01
buZzthis amazed me > The CDMA Development Group states that, as of April 2014, there are 314 operators in 118 countries offering CDMA2000 1X and/or 1xEV-DO service.[7][needs update]22:01
buZz-118- countries, is almost all the countries that exist?22:02
buZz> The intended 4G successor to CDMA2000 was UMB (Ultra Mobile Broadband); however, in November 2008, Qualcomm announced it was ending development of the technology, favoring LTE instead.[8]22:02
freemangordonWizzup: do you remember *why* that patch was needed?22:03
Wizzupyes, otherwise we get no technology reported, since cdma is second in the loop22:04
Wizzupwithout it you get no 2g/3g indicator22:04
freemangordonhmm, not here22:05
freemangordonhere it is first22:05
freemangordonso, reverting the patch and adding missing defines should be the right fix, no?22:05
freemangordonok, but how do we know which one to show?22:06
freemangordonlike, why is cdma even reported here?22:06
Wizzupwell, we want to show the one that is active/online22:07
freemangordonoh, ok, I see what you mean now:22:08
freemangordonOct  4 19:22:06 localhost ofonod[5563]: drivers/qmimodem/gprs.c:extract_ss_info() radio in use 522:08
freemangordonOct  4 19:22:06 localhost ofonod[5563]: drivers/qmimodem/gprs.c:extract_ss_info() radio in use 222:08
freemangordonso, how do select to show 5 and not 2?22:08
Wizzupdon't know22:08
freemangordonumts vs cdma?22:08
freemangordonugh22:09
buZzoooo, does that mean we could have more than just '2G' or '3G' displayed22:09
buZzbut also EVDO etc22:09
buZzso fancy!22:09
freemangordonsure, but do we want to?22:09
freemangordonI don;t think so22:10
buZzi'd love it :)22:10
buZznot that i'll see anything beside GPRS here, i think22:10
freemangordonok, I need some rest22:10
freemangordonwill continue tomorrow22:10
freemangordonnight!22:10
Wizzupttyl :)22:10
freemangordonoh, a mail from andrew22:11
freemangordonobviously I won;t reply today22:11
freemangordonbye22:11
buZzsleep tight22:12
buZzWizzup: so, what iap_conndlg.log is outputting as 'cellid' , is actually the LocationAreaCode22:30
buZzlol, friended ham nerd thinks he can spoof a 700mhz lte network for the droid4 :P22:42
WizzupbuZz: yes, I also have hw to do this22:46
buZzcool :)22:46
Wizzupbut never didi t22:46
WizzupI forget what it was called22:46
buZzhackRF i guess?22:46
buZzor rad1o22:47
Wizzupbladerf22:47
Wizzupwas like 800 usd or so22:47
buZzah, nice22:47
Wizzupit's just eating dust at my house for years22:47
Wizzup:(22:47
Wizzup(2 years)22:47
buZzrad1o was the badge of a cccamp, hackrf 'compatible'22:47
buZzbut its very narrow, maybe 2mhz?22:47
buZzzarya is using some VoLTE hardware for it22:48
buZzi think ..22:48

Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!