**** BEGIN LOGGING AT Wed Nov 18 02:59:59 2015 Nov 18 03:00:16 unlike BB5 our modem has a simple plain standard PCM/I2S audio interface, exactly like the codec/mixer audiochip in N900 already does. And our modem already has all the audio enhancement embedded that is done in closed PA plugins on N900 Nov 18 03:01:07 if everything else fails we could attach the modem audio (and control) over a RTP (and SIP) loopback as a virtual SIP account Nov 18 03:03:13 freemangordon: ping Nov 18 03:06:49 basically asterisk is all that's needed for such "local loopback" Nov 18 03:07:11 and iirc asterisk is already available for maemo, and working Nov 18 03:09:11 I wouldn't even feel surprised if asterisk already offers a USB modem connection plugin for circuit switched calls Nov 18 03:10:25 and if it doesn't, we patch a AVM ISDN card driver into shape Nov 18 03:15:51 btw did we end up finding all the necessary parts to make the keyboard work (i.e. dome sheet)? Nov 18 03:33:12 we have sources, but no physical domesheets yet. We have alternatives which would involve a lot of manual work to place the 39 switches, and it lacks kbd backlight. It's a worst case fallback Nov 18 03:33:52 building the mere domesheet is basically a nobrainer. The lightspreader part is tricky Nov 18 04:10:44 isn't the lights preader part of the dome sheet on a N900? Nov 18 05:30:59 The domesheet comes together with motherboard, it's not something that can be sourced with mechanical parts Nov 18 05:31:34 And yes, the lightspreader part of domesheet-being-built-for-Neo900-motherboards turns out to be tricky? Nov 18 08:52:21 moin :) Nov 18 09:22:32 so someone is selling a Ti 81 calculator for USD17.50, and all of us 'geeks' are looking at it Nov 18 09:24:07 then, i whip up my N900 & launch ATI85 ;) Nov 18 09:24:11 ;) Nov 18 09:24:36 and you can has n900 with broken cellular for that amount Nov 18 09:25:19 The only reason the Texas Instruments calculators continue to exist (and continue to cost what they do new) is because TI have convinced school authorities all over the world to declare TI calculators the "approved" calculators for exams and stuff Nov 18 09:26:06 they are right in a way there are smart devices that are similar size but with enormous abilities Nov 18 09:27:01 that can connect to the site that will solve just any problem Nov 18 09:29:06 merlin1991: ping Nov 18 09:29:10 freemangordon: ping Nov 18 09:37:51 tbh i've never really used ATI85 in the 4 years i've had it installed. now feel like getting the manual for Ti calculators Nov 18 09:38:18 jonwil: isn't it a massive antitrust violation, anyway? Nov 18 09:39:33 There are other companies making graphics calculators and the school is free to choose those if they want (Casio makes them for one thing, not sure if the HP ones still exist or not) Nov 18 09:49:18 iirc they do Nov 18 09:49:49 and Casio calculators are "approved" by schools as well (at least in Europe) Nov 18 09:50:08 but I honestly think they're far behind the TI ones Nov 18 09:50:30 in my times we had pen and paper Nov 18 09:50:35 (or used to be... that was 10y ago :) Nov 18 09:50:37 to solve math problems Nov 18 09:50:39 ;) Nov 18 09:51:03 KotCzarny: actually I've hardly used a calculator to solve a maths problem Nov 18 09:53:20 why do you even use calculators in school anyway Nov 18 09:53:32 because lazy Nov 18 09:53:42 yeah but... graphing ones? Nov 18 09:53:52 same Nov 18 09:53:55 I use the calculator app on my N900 to do calculations all the time Nov 18 09:54:43 and I use the Windows calculator a LOT for doing stuff with hex numbers Nov 18 09:57:05 use use bash and printf for hex fun Nov 18 18:08:28 i have a problem connecting with talk.maemo.org, message says that my ip is on antispam blacklist. the thing is i don't remember any registration or post written by me Nov 18 18:13:12 might be your ip belong to some more people (now or in the past) that made it to the internet bad ip lists Nov 18 18:15:30 Kocie, mam ip z UPC, oni tak często zmieniają ip, że mam je od jakichś 2 lat Nov 18 18:20:21 t4nk226: czyli dynamiczna ipka, masz pecha ze poprzedni 'wlasciciel' cos nabroil Nov 18 18:20:45 chyba ze wchodziles na tmo juz z ta ipka, wtedy moze wirusa jakiegos masz Nov 18 18:23:20 nie było możliwości, żeby ktoś inny miał to ip, bo mam je od 2 lat non stop i nie zanosi się na zmianę (nawet na infolinii mówią, że pula jest wyczerpana), na talk.maemo wchodzilem jakis czas temu i problemow nie bylo, wirusow tez nie ma Nov 18 18:23:48 jeszcze opcja ze cala klasa ipek trafila na listy Nov 18 18:23:50 chyba ze to przez losowego useragenta przegladarki Nov 18 18:24:58 http://www.stopforumspam.com/ipcheck/89.76.153.60 Nov 18 18:25:01 pierwszy link z googla Nov 18 18:25:11 jednak trafiles na liste rblowa Nov 18 18:25:25 albo masz wirusa albo dziurawe proxy masz Nov 18 18:26:06 swietnie ... Nov 18 18:26:49 no ja bym na twoim miejscu sprawdzil antyvirem ew. jak masz tam jakis serwer to czy wlamu nie miales Nov 18 18:27:21 sprawdze, raczej nie jest fajnie miec poblokowanego za darmo neta Nov 18 18:29:28 ew. opcja ze ktos do ciebie z laptopem przyszedl wpial sie i tak trafiles Nov 18 18:29:44 przedwczoraj miales gosci? Nov 18 18:30:43 nie, wlasnie prowadze sledztwo, w niedziele udostepnialem na zewnatrz tylko wordpressa przez jakies pol godziny, pozniej juz nic Nov 18 18:31:01 odpalilem jednego antywira, przeleci 2 dyski i pewnie nic nie znajdzie Nov 18 18:31:11 no to juz masz punkt zaczepienia Nov 18 18:31:46 zgloszenia sa z 16.11 a wp byl online 15.11 kolo 22.30 Nov 18 18:32:20 mhm, tak czy owak posprawdzaj, bo to raczej samo z siebie sie nie zrobilo Nov 18 18:33:24 t4nk226: sorry I can't follow. Please use english language Nov 18 18:33:43 Doc: i have a problem connecting with talk.maemo.org, message says that my ip is on antispam blacklist. the thing is i don't remember any registration or post written by me Nov 18 18:33:55 I've seen that post Nov 18 18:34:26 stopforumspam.com sent by KotCzarny says that my ip was reported Nov 18 18:35:37 then try changing your IP maybe Nov 18 18:35:41 thinking about someone got into my wifi (hard password to break, so not so easy), maybe virus but it won't happen by itself Nov 18 18:35:51 though odds are the complete provider subnet is blacklisted Nov 18 18:36:32 ISP is not providing anything like that, the pool is so tight that even having dynamic ip got me stuck with this ip for about 2 years Nov 18 18:36:42 t4nk226: no, you got dynamic IP and when you do a reset to your DSL connection, odds are your IP will change Nov 18 18:36:59 ooh Nov 18 18:37:14 t4nk226: changing ip is not the solution, because clearly no one else used it, and you got in the rbl, so it might happen again Nov 18 18:37:28 well, then contact chem|st Nov 18 18:37:37 well, unfortunately not, the modem+router machine i have is shitty, i had a battle with it some time ago and nothing has changed Nov 18 18:40:17 how long the ip stays in stopforumspam database ? Nov 18 18:40:34 contact chem|st, ask him for assisantce to register to tmo Nov 18 18:40:36 you might want to email that db admins Nov 18 18:40:48 assistance even Nov 18 18:41:12 alternatively warfare Nov 18 18:41:25 both are in here Nov 18 18:43:00 chem|st ? sorry, not familiar with thos Nov 18 18:43:02 this* Nov 18 18:43:10 do a /whois Nov 18 18:43:14 its a nick Nov 18 18:44:59 that's an problem of your unfamiliarity with IRC ;-) Both are users in this channel, just like you Nov 18 18:46:31 t4nk226: please send a mail to techstaff@m.o asking for your ip being removed from the blocklist (include the ip), please. Nov 18 18:47:01 mail as @maemo.org if that wasn't clear Nov 18 18:47:43 ok, thanks Nov 18 18:47:49 ohh there's a factoid for that... Nov 18 18:48:00 ~unbanip Nov 18 18:48:01 hmm... unbanip is please contact techstaff maemo org with your request Nov 18 18:48:09 sixwheeledbeast: lol Nov 18 18:48:15 :) Nov 18 18:48:30 it's easier Nov 18 18:48:34 swb: ah, thats' how its named, i was thinking something with the staff in the name ;) Nov 18 18:48:53 ~techstaff Nov 18 19:12:01 I just wrote two mails, one for techstaff and one for stopforumspam, hope it'll help Nov 18 19:44:39 ~#maemo techstaff is techstaff@maemo.org - the folks that keep your maemo infra running. Devotion to Duty http://xkcd.com/705/ Nov 18 19:44:46 DocScrutinizer05: okay Nov 18 19:45:41 ~unbanip is also see ~techstaff Nov 18 19:45:42 okay, DocScrutinizer05 Nov 18 19:45:54 ~unbanip Nov 18 19:45:54 i heard unbanip is please contact techstaff maemo org with your request Nov 18 19:46:04 ugh Nov 18 19:46:12 ~factinfo unbanip Nov 18 19:46:13 unbanip -- created by DocScrutinizer05 <~saturn@openmoko/engineers/joerg> 31s ago. Nov 18 19:46:22 ~forget unbanip Nov 18 19:46:22 i forgot unbanip, DocScrutinizer05 Nov 18 19:46:28 ~literal unbanip Nov 18 19:46:28 "#maemo unbanip" is "please contact techstaff maemo org with your request" Nov 18 19:46:39 ~#maemo unbanip is also see ~techstaff Nov 18 19:46:40 okay, DocScrutinizer05 Nov 18 19:46:45 ~unbanip Nov 18 19:46:45 well, unbanip is please contact techstaff maemo org with your request, or see ~techstaff Nov 18 19:51:40 infobot: no, #maemo techstaff is techstaff(AT)maemo.org - the folks that keep your maemo infra running. Devotion to Duty http://xkcd.com/705/ Nov 18 19:51:40 okay, DocScrutinizer05 Nov 18 19:55:11 (when you already know which xkcd it is before clicking ... bleh :) Nov 18 20:04:26 well... actually techstaff did such an awesome job the the mere technical infra runs almost unattended since years now. Of course this doesn't apply to highly problematic exposed services like vBulletin aka tmo, which I'm sure gives its administrators a constatnt stream of 'fun' Nov 18 20:05:20 uptime of blades aka xen hosts is in the 600 days Nov 18 20:06:18 err >> 21:06:01 up 867 days, 19:28, 0 users, load average: 4.40, 4.35, 3.77 << Nov 18 20:13:52 freemangordon: ping Nov 18 20:18:45 hi Nov 18 20:28:48 jonwil: about your post on TMO... why do you need to wrap it if you want to keep them "as is" ? Nov 18 20:28:55 (pulseaudio) Nov 18 20:29:09 No, I meant we use specific functions from the blobs Nov 18 20:29:14 and clone or rewrite others Nov 18 20:29:30 i.e. we rewrite the stuff for cmtspeech Nov 18 20:29:35 but we keep e.g. fir_iq stuff Nov 18 20:29:46 or other stuff with all that math Nov 18 20:30:55 do you follow? Nov 18 20:34:06 not sure Nov 18 20:34:15 why would we rewrite only half of it? Nov 18 20:34:28 (I mean, why bother if we cant/dont remove the whole blob?) Nov 18 20:35:08 for Neo900 Nov 18 20:35:16 because we need to replace the cellular bits Nov 18 20:35:43 but we dont need to replace some of the other stuff since Neo900 has the same audio system as N900 other than the cellular stuff Nov 18 20:36:00 also we could use this for testing, i.e. clone some parts and use others, see that it works then clone more Nov 18 20:38:16 are we still talking about the nokia voice pa module? Nov 18 20:38:34 yes we are Nov 18 20:38:35 isn't this one quite "far" from the cmtspeech stuff in the pipeline? Nov 18 20:38:40 no, its not Nov 18 20:38:43 ah Nov 18 20:38:48 I see Nov 18 20:39:03 module-nokia-voice on N900 statically links to a library called libcmtspeech Nov 18 20:39:06 that's a pity Nov 18 20:39:11 which is what talks to the cellular radio Nov 18 20:39:14 yes it is a pitty Nov 18 20:39:18 now I get your point Nov 18 20:39:19 but we have to do what we need to Nov 18 20:39:36 does that mean SIP stuff doesn't go through nokia-voice? Nov 18 20:39:47 or do they have another entrypoint for SIP/RTP? Nov 18 20:40:02 s#SIP/RTP#telepathy calls# Nov 18 20:41:01 there is a different entrypoint (pulseaudio sink) Nov 18 20:42:53 For a cellular call, the cellular services daemon sends a signal (com.nokia.csd.Call.Instance.AudioConnect) to tell module-nokia-voice (and libcmtspeech) to start doing audio Nov 18 20:43:34 then libcmtspeech takes audio from the cellular modem and feeds it through various algorithms and out to other parts of PA and then the speakers Nov 18 20:43:39 and vice versa for microphone audio Nov 18 20:46:15 module-nokia-voice creates 2 sinks (raw sink and voip sink) and 3 sink inputs (aep sink input, hw sink input and cs call sink input) Nov 18 20:46:49 for sources there is raw source and voip source plus a hw source output Nov 18 20:47:31 looks like a monolithic monster Nov 18 20:47:49 yes its a beast Nov 18 20:55:56 now do you see the wisdom of my idea? :) Nov 18 20:57:07 kinda, yeah :) Nov 18 21:05:23 i was wondering... hasn't anyone thought of implementing a smartphone based on the zynq7000 soc from xilinx? even the gsm/3g/4g interface could be open source, implemented on the fpga part of he zynq Nov 18 21:06:07 wait! one PA plugin (module-nokia-voice) handles both csd and RTP audio? this sounds terribly wrong particularly for RTP which should get done by some FOSS module - there are quite a few, libccrtp comes to mind Nov 18 21:06:39 if i'm not mistaken osmocom software and docs would greatly help with this, as they have implemented the bts, so the terminal itself should be probably easier and possibly a subset of that Nov 18 21:07:02 Telepathy or some other part of the system feeds VoIP audio into module-nokia-voice which does some algos on it (including xprot for speaker protection) and feeds it down the chain into the rest of PA Nov 18 21:07:05 ceene: zynq is pretty big Nov 18 21:07:43 and pretty cool :P Nov 18 21:08:33 in any case once I get hold of freemangordon, I plan to ask for commit access to https://notabug.org/freemangordon/pulseaudio-nokia/ and see what I can figure out and add to that repo :) Nov 18 21:08:36 two arms, a big integrated fpga in which to implement almost all hardware Nov 18 21:08:37 yes, but implementing gsm etc on vhdl? Sounds like a full time job more multiple devs, if you can get this even certified Nov 18 21:08:48 ceene: a) osmocom bts stack is mostly unrelated to the MT radio stack, it's working in quite a different way than a phone b) no matter how you implement the hardware and th radio stack, it's illegal to use it when user can replace or modify it Nov 18 21:09:19 i know for a fact that xilinx has demonstrated an LTE client Nov 18 21:09:43 so? Nov 18 21:10:05 implemented on just two chips, the zynq and one thing from TI, if i recall correctly, a dual transceiver Nov 18 21:10:17 well, it can be done technically :) Nov 18 21:10:37 ceene: uuh nice, probably one of the more expensive IPs I guess Nov 18 21:10:53 i don't know if they sell it, to be honest Nov 18 21:11:12 but if they have implemented it, i guess you can buy it Nov 18 21:11:49 there are simpler cores for gsm that i think are included in basic licensing schemes Nov 18 21:12:13 2g is probably feasible Nov 18 21:13:41 ceene: I see kintex IPs (thats the big ones) Nov 18 21:13:47 also I wonder how much of practrical use such a proof of concept implementation would have. I mean, I worked with modem manufacturers and I can tell for sure the radio stack is a system way more complex and larger than for example a bare bones linux, and it's in the 3 digit man years what companies like ST-Ericsson invest into development of such stack, not included the effort to get everything tested in field and then approved by 12 Nov 18 21:13:48 dozen international authorieties like FCC Nov 18 21:15:50 one man can dream... Nov 18 21:15:50 so uuh here is a zynq based https://www.mathworks.com/examples/lte-system/12592-lte-transmitter-using-zynq-based-software-defined-radio-sdr Nov 18 21:16:10 you'd still need to go from vhdl/fpga to asic Nov 18 21:16:14 no matter what, you for sure can publish the radio stack source even under GPL, but as a manufacturer you must assure that only FCC-cerified sw stacks get used on your hardware. And FCC won't certify any firmware that doesn't enforce this limitation Nov 18 21:16:25 because you dont want to keep a fpga running on some pocket embedded system Nov 18 21:16:44 I can haz moar battery plz? Nov 18 21:17:37 DocScrutinizer05: fcc is an US only thing, isn't it? Nov 18 21:18:01 bencoh: yep, that's just a more real-life aspect that usually excludes FPGA from the list of stuff you want in a battery powered embedded system Nov 18 21:18:43 ceene: please take FCC for a wildcard for any authority ruling radio band usage in your country Nov 18 21:19:02 i don't know if power consumption would be worse than our n900 though... Nov 18 21:19:11 huh? Nov 18 21:19:16 n900 isnt that power hungry Nov 18 21:19:24 especially when idling Nov 18 21:19:35 the modem has a standby energy consumption of only 12mW Nov 18 21:19:45 in my experience the screen is the worst offender Nov 18 21:20:05 s/screen/backlight/ Nov 18 21:20:17 so maybe the communication subsystem in the end is not the thing that requires the most optimization Nov 18 21:20:22 no, the backlight is, plus data transfer over cellular, plus WLAN when in scanning mode Nov 18 21:20:43 add gps to the list Nov 18 21:20:49 whatever, screen needs baklight, so it's the same for me Nov 18 21:20:58 plus CPU when you run a really poorly written app that keeps the CPU busy Nov 18 21:21:19 Humpelstilzchen: yes, GPS can eat some power. Camera as well Nov 18 21:22:07 pareto says 80% of my power goes out to the screen, even if it's only lighted on for just 20% of the timwe Nov 18 21:22:22 ceene: backlight can get dimmed. Just one tick down from brightness 5 to 4 already makes backlight consume only 20-30% Nov 18 21:22:46 pareto? Nov 18 21:23:05 "pareto effect" Nov 18 21:23:12 80/20 Nov 18 21:23:16 did somebod ? Nov 18 21:23:50 I dont think so Nov 18 21:25:27 i still think a zynq smartphone would be pretty cool Nov 18 21:26:34 with a car lead battery, definitely :° Nov 18 21:27:13 lol Nov 18 21:27:23 it wouldn't be so bad Nov 18 21:44:10 ~power Nov 18 21:44:11 extra, extra, read all about it, power is http://wiki.maemo.org/N900_Hardware_Power_Consumption Nov 18 21:44:34 idea of system monitor ... meaning what in N900 context? Nov 18 21:45:36 check all (known) subsystems for activity, estimate the power consumption that comes from this for each of them, compare to measured power consumption as seen by BQ27200 Nov 18 21:57:02 sth like >> total measured energy consumption: 1980mW (520mA@3.81V) \n estimation for cellular: 1100mW, based on 'data: 41,4kBytes/sec RX, 528Bytes/s TX, in 9 packets/s', 'signal: -74dB', 'Radio Access Technology: UMTS', 'no call active' \n Screen backlight: 0mW, based on 'screen blaked, backlight off', 'actual brightness: 0%', 'configured brightness: 80%, ALS dynamic' \n Camera: 0mW based on 'not active' \n estimation for WLAN: 30mW Nov 18 21:57:04 based on: 'idle', 'data: 0Bytes/s', 'beacons: 20/s', 'PSM: high' \n ... Nov 18 22:00:11 and so on, for CPU, PowerVR, DSP, Audio, Vibrator, USB, BT, younameit Nov 18 22:02:30 are any DSP/PowerVR stats availalbe in /sys maybe or some other location? Nov 18 22:02:33 have a set of thresholds for alarms for each subsystem, like "when CPU eats >1000mW for >5 seconds", and also have an alarm that triggers on measured consumption massively non-compliant with what the summed up total of all subsystems would suggest Nov 18 22:02:56 Sicelo: see zzztop Nov 18 22:04:18 I don't know how much is actually exposed anywhere in /sys or /proc or elsewhere Nov 18 22:07:52 i'll study the script Nov 18 22:07:58 actually my idea for system monitor is of boader scope than only power consumption, though that's for sure the major focus and largest chunk of info it collects and supervises and displays Nov 18 22:08:28 broader even Nov 18 22:10:44 but zzztop and [h]top (or actually what they are based on) would also be info sources, so for example an alarm of JoergsSystemMonitor would also kick in when a process suddenly is using way more CPU time or RAM than it usually does Nov 18 22:30:46 anyway such monitoring tool would be *the* answer of choice to any complaints about "N900 battery standby time too short" Nov 18 22:32:24 it would literally list whatt's wrong, either some software that keeps CPU busy, or constant data traffic over cellular/GPRS, or WLAN in constamt scan mode, or backlight on all the time, or even "your battery capacity seems way lower that standard! battery old and worn?" Nov 18 22:33:33 lower thaN* standard **** ENDING LOGGING AT Thu Nov 19 02:59:58 2015