00:51.41 | *** join/#neo900 mordac_ (~mordac_@c-24-6-219-139.hsd1.ca.comcast.net) |
01:18.30 | *** join/#neo900 dos1 (~dos1@pinkiedb.com) |
01:18.38 | *** join/#neo900 dos11 (~dos@unaffiliated/dos1) |
01:19.08 | *** join/#neo900 dos11 (~dos@unaffiliated/dos1) |
01:26.21 | *** join/#neo900 dos1 (~dos1@pinkiedb.com) |
01:28.43 | *** join/#neo900 dos1 (~dos1@pinkiedb.com) |
01:29.30 | *** join/#neo900 dos1 (~dos1@pinkiedb.com) |
01:33.29 | *** join/#neo900 dos1 (~dos1@unaffiliated/dos1) |
01:59.16 | *** join/#neo900 ddark_ (~ddark@188.120.138.60) |
02:14.23 | *** join/#neo900 nicksydney (~quassel@140.137-50-210.dynamic.dsl.syd.iprimus.net.au) |
02:20.19 | *** join/#neo900 chainsawbike (~chainsawb@unaffiliated/chainsawbike) |
05:31.08 | *** join/#neo900 nicksydney (~quassel@140.137-50-210.dynamic.dsl.syd.iprimus.net.au) |
05:37.14 | *** part/#neo900 freemangordon (~ivo@46.249.74.23) |
05:37.34 | *** join/#neo900 freemangordon (~ivo@46.249.74.23) |
05:55.57 | *** join/#neo900 Fortuona (~Fortuona@2001:638:208:fd5f:31c5:dfd3:a572:f240) |
06:12.27 | *** join/#neo900 sixwheeledbeast (~paul@2a01:348:6:860a:39da:8964:d21f:2c26) |
07:41.19 | *** join/#neo900 Kabouik (~quassel@147.99.218.243) |
08:17.01 | *** join/#neo900 parmegv (~parmegv@ma.sdf.org) |
08:22.42 | parmegv | hey people! I cannot use duckduckgo nor google to search a string in the irc logger: https://encrypted.google.com/search?hl=en&q=site%3Ahttp%3A%2F%2Finfobot.rikers.org%2F%2523neo900%2F%20phoronix yields no result! |
08:23.35 | parmegv | I wanted to know if there has been any comparison between new900 and fairphone, in terms of "fair trade" ideas |
08:24.06 | *** join/#neo900 remarc (~sequantz@194.11.24.44) |
08:24.47 | parmegv | I've seen that Golden Delicious Computers is from Germany, but I don't know if they're targetting an ethical audience or not |
08:28.03 | parmegv | ha, reading http://projects.goldelico.com/p/openphoenux/downloads/get/OpenPhoenux.pdf, on page 6 it says "hardware production near to |
08:28.06 | parmegv | users (Europe) under fair labour |
08:28.08 | parmegv | conditions" |
08:28.32 | parmegv | that's good news |
08:38.19 | *** join/#neo900 che1 (~che@p5B07380F.dip0.t-ipconnect.de) |
08:48.13 | *** join/#neo900 xes (~xes@unaffiliated/xes) |
08:52.36 | *** join/#neo900 edgar2 (~edgar2@84-230-8-29.elisa-mobile.fi) |
09:09.46 | *** join/#neo900 Fortuona (~Fortuona@2001:638:208:fd5f:b0b9:908f:809e:671) |
10:15.54 | DocScrutinizer05 | that's about as "fair" as it gets, even for fairphone |
10:17.04 | DocScrutinizer05 | I consider fairphone more of a propaganda concept and for that it's fine. Their environmental dedication however is futile |
10:45.21 | *** join/#neo900 mvaenskae (~mvaenskae@unaffiliated/mvaenskae) |
11:31.46 | parmegv | DocScrutinizer05: they have published how they've chosen their supplier, and also their "requirements" |
11:32.21 | parmegv | I don't know how could I prove they're being dishonest, although I also think a bit like you |
11:36.27 | DocScrutinizer05 | they are not dishonest, they simply emphasize their efforts to do "green" electronics, but there's no such thing like "green electronics", you can't build your own chips with coltan you dug up in your bavarian garden |
11:37.40 | DocScrutinizer05 | it's a bit like claiming you're doing something for environment by using plastic bags that are 100mm shorter than usual |
11:37.53 | DocScrutinizer05 | 10mm |
11:38.36 | DocScrutinizer05 | by re-using the N900 cases we probably do more for environment than greenphone with their "fair sourcing" |
11:41.23 | DocScrutinizer05 | in my book greenphone is a great PR concept to create more awareness for the problems in compulsive hightech planned obsolescence and dumping prices, but it doesn't change a single aspect of what's really happening, their percentage of really "green" components is maybe 2% in their phone |
11:42.31 | DocScrutinizer05 | similar to the green mouse, which has a case of wood instead of PVC or whatever, but the whole electronics inside is same toxic crap as with any arbitrary other mouse |
11:44.04 | parmegv | what I care the most is about fair incomes for workers, do you have an opinion on that? |
11:44.08 | parmegv | btw, what's your book? |
11:44.36 | DocScrutinizer05 | true, the case is "green" (not evaluating how much of non-green resources been used up to produce that wooden case), but does that realy make a difference for the environment, given the case is probably ~3% of the total amount of resource usage footprint of such mouse |
11:45.12 | DocScrutinizer05 | "my book" == my opinion, my point of view |
11:45.27 | parmegv | ah, ok |
11:45.40 | DocScrutinizer05 | I have a pretty clear idea about fair income: mine is waaaay too low right now |
11:45.54 | DocScrutinizer05 | and that's honest, not sarcastic |
11:47.16 | parmegv | I think that fairphone is doing a god job there |
11:47.33 | DocScrutinizer05 | we're producing at Munich/Bavaria/Germany and the fab charges normal prices for that. We can't do anything about chips and components, we need to buy them where we can get them. And we all (4 guys of Neo900 team) don't see any fair income right now |
11:48.19 | parmegv | oh, you're not being paid? I thought donations are for you too |
11:49.13 | DocScrutinizer05 | yes, that's the plan, so from next month on I _must_ pay me a monthly fee at the level of a trainee or somesuch |
11:49.28 | parmegv | good news for you then :) |
11:50.11 | DocScrutinizer05 | not really, when I would do a regular job as coder monkey at an arbitrary company around here, I'd earn more for sure |
11:50.58 | parmegv | oh yes, for sure, but you're enjoying much more with this, instead of being a code monkey ;) |
11:51.01 | DocScrutinizer05 | this project is born from love and enthusiasm for the N900 and GTA02/04 |
11:51.14 | DocScrutinizer05 | and *free* linux on our phones |
11:52.13 | parmegv | and regarding chips and components... perhaps you could talk to fairphone and other companies trying to leverage the ethical standards, and join forces |
11:53.13 | DocScrutinizer05 | for now my ballance is pretty much a negatibe one: expenses in the 1000 to 3000 EUR, plus a 9 months I wasn't really able to do any other works thanks to managing Neo900 and thus my work didn't pay me bagels during that time |
11:53.54 | parmegv | DocScrutinizer05: you could freelance while managing this |
11:54.11 | DocScrutinizer05 | no, we won't talk to fairphone, we have no evangelist attitude, and they don't have any resources we could exploit for Neo900 |
11:54.32 | DocScrutinizer05 | parmegv: sorry? |
11:55.35 | parmegv | I didn't want to look agressive, perhaps my sentence was too short |
11:57.12 | parmegv | it was just a suggestion, to enjoy your work without suffering from low income |
11:57.25 | DocScrutinizer05 | I don't understand |
11:57.30 | parmegv | but nvm, I shouldn't have said anything like that, it's your personal life |
11:57.45 | DocScrutinizer05 | I'm freelancer basically for whole my professional lifetime |
11:58.39 | parmegv | then great, you can spend time on this! |
11:59.11 | DocScrutinizer05 | and when I say "I wasn't really able to do any other works" then that's because e.g. I've not found the time to leave my flat since last weekend, sitting in front of PC 20h per day |
11:59.35 | DocScrutinizer05 | I think there's some sort of misconception at your end |
11:59.59 | parmegv | yes, but I already understood |
12:02.09 | DocScrutinizer05 | re GreenPhone: we easily could do mega advertisement "We are even MORE green than GreenPhone" but we're not interested in that |
12:03.26 | DocScrutinizer05 | as already stated GreenPhone is marketing an idea and an ideology, but they don't really sell any new product, their hardware is not any "greener" than ours |
12:04.02 | DocScrutinizer05 | they just "take care" about that aspect. Well, we would do as well, when we'd see a way this would result in *any* change |
12:05.05 | DocScrutinizer05 | we're using exactly the same chips and resistors and capacitors they use too. And our manufacturing is exclusively in Munich/Bavaria |
12:05.55 | parmegv | it's your choice, but I think that as well as Libre Software is a good choice for society and people make efforts to make them know, this is also a good choice and people clearly need to hear about this |
12:06.07 | DocScrutinizer05 | but they tried to find the "greenest" chips, and pestered chip manufacturers about it, which is a good thing but sth we don't want to do |
12:06.14 | parmegv | but again, you'll have your reasons not to "evangelize" |
12:07.09 | DocScrutinizer05 | well, we also won't fund a political party despite it would be "the right thing to do" |
12:07.32 | DocScrutinizer05 | it's just not my profession |
12:07.53 | parmegv | aha, yes, I understand that last statement |
12:08.22 | parmegv | thanks for your convo, I have to leave, I'll stay here to read and/or write! |
12:08.31 | parmegv | bye! |
12:08.36 | DocScrutinizer05 | I'm an EE, not a marketing and spindoctor crack that could run a second "GreenPhone campaign" |
12:09.14 | DocScrutinizer05 | bye, thanks for your interest and suggestions! :-) |
12:12.45 | DocScrutinizer05 | my "political" statement is by building hardware that can last 10 years, is re-using existing components wherever reasonable, and clearly blocks the whole big brother evil tendencies as much as possible, and particularly on the latter we're best you can get *anywhere* |
12:13.59 | DocScrutinizer05 | I also think I'm fostering community by what I'm doing, which also has its political aspects |
12:17.30 | DocScrutinizer05 | then there are less conceptual aspects like me helping at least a 50 N900 users to fix their broken USB, so a 50 new whyPhones or whatever didn't get produced and discarded after 2 years |
12:20.18 | DocScrutinizer05 | but regarding what GreenPhone aims at, I only can support that idea by looking into their results and buying any green parts they found or managed manufacturers to build. Of course we're doing this, but so far their results regarding that are close to zilch |
12:33.52 | *** join/#neo900 mvaenskae (~mvaenskae@unaffiliated/mvaenskae) |
12:35.08 | *** join/#neo900 freemangordon_ (~ivo@85-118-92-26.mtel.net) |
12:35.35 | *** join/#neo900 freemangordon_ (~ivo@85-118-92-26.mtel.net) |
13:05.31 | FIQ | dc |
13:06.13 | FIQ | DocScrutinizer05: isn't fairphone also a lot about social sustainable development as much as environmental? |
13:06.31 | FIQ | from what I've seen, they did quite well on that regard, but you're probably right about the environmental issues though |
13:07.07 | mvaenskae | fairphone? i didn't know phones grew on trees |
13:18.17 | mvaenskae | the idea about a fair phone is very nice _but_ that is neither economical nor possible |
13:18.32 | mvaenskae | creating chips is a very dirty procedure |
13:28.31 | DocScrutinizer05 | exactly |
13:29.20 | mvaenskae | also the pcbs are done using acids iirc and usually the products created by that are not very healthy |
13:29.26 | DocScrutinizer05 | and no tiny manufacturer like fairphone or Neo900 is going to force the chip makers into tidying up that dirty process |
13:30.28 | DocScrutinizer05 | well, since our PCBs are built in Germany, I'm confident that those acids are properly recycled. The copper is too valuable to, to discard the acid with dissolved copper in it |
13:30.35 | mvaenskae | DocScrutinizer05: even if wanted it cannot be cleaned up by much as the silicon wafer creation is so complex these days that a particle of dust can destroy thousands of possible chips |
13:30.43 | *** join/#neo900 freemangordon_ (~ivo@85-118-92-26.mtel.net) |
13:30.54 | DocScrutinizer05 | yes, I know |
13:31.26 | mvaenskae | i was extremely impressed hearing upon chip manufacturing as a short sidenote in a chemistry lecture |
13:31.46 | mvaenskae | also, great news getting another person aboard the neo900 :) |
13:32.03 | DocScrutinizer05 | and nothing fairphone can do about it. Also the cltan sources and other rare earth are where they are and the west is happy when china sells enough of that stuff to us so we *can* build chips at all |
13:32.12 | DocScrutinizer05 | coltan* |
13:32.14 | mvaenskae | and i liked seeing the presentation of the current development on the neo900, much appreciated |
13:32.55 | DocScrutinizer05 | not like we could force china to not make a mess out of the rare earth digging in their own country |
13:33.18 | mvaenskae | i assume the big chip on the video where dos called the phone back and forth was the modem? |
13:33.34 | DocScrutinizer05 | yep |
13:33.53 | DocScrutinizer05 | PHS8 |
13:34.50 | mvaenskae | the chip is really large, would there be even enough room for 2 modems? |
13:35.12 | DocScrutinizer05 | it's a module, but anyway. Hardly enough room for two of them |
13:35.41 | DocScrutinizer05 | well, they would "just fit" but honestly, that gets nasty |
13:36.51 | DocScrutinizer05 | I also can't see anybody willing to pay the additional 200+ bucks for such a device with 2 modem modules |
13:37.24 | mvaenskae | well, would allow people to get regional lte and world wide umts :) |
13:37.33 | DocScrutinizer05 | let alone trouble with RF design |
13:38.08 | mvaenskae | true, if i was smart enough i would try solving it but nah... i can hardly work with any electrical circuits |
13:38.48 | DocScrutinizer05 | I'd be *really* intersted in internal circuitry of the Cinterion P*S8, I *guess* it's using tunable wide range filters, so the supported bands are just a question of firmware |
13:39.28 | mvaenskae | mosst likely firmware, companies usually pay for software licenses these days not hardware |
13:39.55 | DocScrutinizer05 | I mean, when such module is supporting e.g. 1800/1900 GSM, then why wouldn't it be able to support UMTS on same frequencies |
13:40.58 | mvaenskae | exactly, the hardware for 2g 1800/1900 is there but not for 3g? you gotta be kidding |
13:41.12 | DocScrutinizer05 | I even suspect they have *one* filter (or two, for UL and DL) and those are tunable to any frequency between at least 1800 and 2500 |
13:41.38 | mvaenskae | well, time for you guys to RE it ;) |
13:41.52 | DocScrutinizer05 | they might use a separate circuitry for the 800MHz range maybe |
13:42.04 | mvaenskae | if i had the knowledge i would most definitly give it a try |
13:42.09 | DocScrutinizer05 | yep, I'm planning to eventually disassemble one of those modules |
13:42.09 | mvaenskae | makes a nice hack |
13:42.27 | DocScrutinizer05 | alas we can't hack the module hw nor firmware |
13:42.40 | mvaenskae | one cannot update the firmware? |
13:43.01 | DocScrutinizer05 | yes, you can, but you can't change it. Much like a tivoized smartphone |
13:43.20 | DocScrutinizer05 | signed by cryptokeys |
13:43.34 | DocScrutinizer05 | chenge one byte and it doesn't flash anymore |
13:43.38 | DocScrutinizer05 | change |
13:44.22 | DocScrutinizer05 | in that regard it's exactly like the arbitrary average android phone, just for some reason |
13:44.47 | DocScrutinizer05 | and unlike those phones you cannot root it |
13:45.02 | DocScrutinizer05 | at least no exploit known so far |
13:45.17 | DocScrutinizer05 | doesn't mean there's none, but just not known (yet) |
13:45.43 | mvaenskae | int getRandomNumber() { return 4; } |
13:45.45 | DocScrutinizer05 | and industry meanwhile learned to protect their stuff and do signature and crypto right |
13:45.50 | mvaenskae | that can always happen ;) |
13:46.49 | DocScrutinizer05 | sure, you always *might* find unsealed development images that allow you to flash arbitrary other stuff then since you can hack those dev imgs |
13:47.05 | DocScrutinizer05 | eventually, and for *some* types of modem |
13:47.38 | DocScrutinizer05 | or manufacturer even forgot or omitted sealing of the productive firmware, like happened for calypso |
13:48.01 | mvaenskae | frmware sealing? |
13:48.34 | DocScrutinizer05 | but honestly, all that is mega botring stuff as long as you don't have the docs needed to understand what those 50MB of firmware are *doing* |
13:49.13 | DocScrutinizer05 | so you could apply reasonable meaningful hacks, beyond "I changed "Cinterion" string to "You're pwnd" |
13:49.59 | mvaenskae | DocScrutinizer05: so what most android devs do? |
13:50.20 | mvaenskae | not to discredit them but very few of them seem to turn out to be true hackers |
13:51.44 | *** join/#neo900 mvaenskae (~mvaenskae@unaffiliated/mvaenskae) |
13:55.11 | DocScrutinizer05 | I never looked too deep into android. AIUI most of what they do is userland changes |
13:56.45 | DocScrutinizer05 | and even while android maybe is FOSS allegedly, many phones are locked so you can't simply update to an arbitrary new system version |
13:57.10 | mvaenskae | and if you do, you void your warranty |
13:57.13 | mvaenskae | great concept |
13:57.29 | DocScrutinizer05 | dunno what exactly cyanogen mod is doing, and on which hw platforms it (doesn't) work |
13:58.19 | DocScrutinizer05 | see e.g fruitPhone, you need to exploit a vulnerability in the core system to "root" it |
13:59.22 | DocScrutinizer05 | well, kinda OT for this channel. Luckily Neo900 has none of all that shite |
13:59.35 | DocScrutinizer05 | except contained inside modem module |
13:59.51 | DocScrutinizer05 | and believe me, you don't want to hack the modem |
14:00.36 | mvaenskae | DocScrutinizer05: i was just about to propose that; if you do, you have a slogan "a phone like only the nsa has" :D |
14:00.44 | DocScrutinizer05 | I worked a year for ST-Ericsson, seen modem firmware for LTE from inside. You do _not_ want to hack that |
14:01.01 | mvaenskae | is it full of goto? |
14:01.27 | DocScrutinizer05 | not really, though it depends on which modem and which firmware |
14:01.42 | DocScrutinizer05 | but is has a complecity that exceeds that of e.g. maemo |
14:01.49 | DocScrutinizer05 | complexity |
14:02.09 | mvaenskae | of the whole OS? |
14:02.28 | DocScrutinizer05 | there are several *hundreds* of manyears development going just into the firmware |
14:02.49 | DocScrutinizer05 | not talking about the chip/silicon |
14:03.53 | mvaenskae | wow, i didn't expect it to be that much :o |
14:04.20 | DocScrutinizer05 | firmwares are several dozen or even hundreds of MB size nowadays, and that's without any fancy GUI stuff that eats lots of storage |
14:05.26 | DocScrutinizer05 | I think the ST-E firmware for NovaThor was several GB of source |
14:06.28 | mvaenskae | just for radio communications?! |
14:08.05 | DocScrutinizer05 | for radio stack, data, handover, debugging, servicing the interfaces (UART, HSI, USB, whatnot else), battery management, you name it |
14:08.30 | DocScrutinizer05 | SMS, MMS, |
14:08.38 | DocScrutinizer05 | GPS |
14:08.53 | DocScrutinizer05 | GSM, UMTS, LTE |
14:09.25 | DocScrutinizer05 | AT interpreter (that's a joke somebody could write in a week's work) |
14:09.56 | DocScrutinizer05 | MIMO, 911, SIM handling |
14:10.02 | DocScrutinizer05 | addrbook |
14:10.08 | DocScrutinizer05 | SIMlock |
14:10.17 | mvaenskae | :( |
14:10.18 | DocScrutinizer05 | protection |
14:10.24 | mvaenskae | that's extreme |
14:10.44 | mvaenskae | i will _never_ work with modems D: |
14:10.52 | DocScrutinizer05 | you do *not+ want to hack that unless you got the sources |
14:11.40 | DocScrutinizer05 | and then all that stuff has to run realtime, so it guarantees data rates of up to 100Mbps |
14:12.15 | che1 | So I guess there isn't any known source leak? |
14:12.30 | mvaenskae | and they still managed to have that thing work as a tracking device?! holy moly |
14:12.45 | mvaenskae | che1: not without either NSA involvement or a big bucket of cash |
14:12.57 | DocScrutinizer05 | there's about 20 or more known firmwares that are completely independant of each other, using their own proprietary OS and whatnot |
14:12.58 | che1 | :( |
14:13.38 | mvaenskae | that's a major case of reinventing the wheel |
14:14.30 | DocScrutinizer05 | calypso firmware leaked, but that's really stupid one and 10 years old, and it's already some dozen of MB source, nevertheless can do only GSM and not even EDGE |
14:15.44 | DocScrutinizer05 | but calypso chipset isn't sealed |
14:15.48 | DocScrutinizer05 | usually |
14:16.15 | DocScrutinizer05 | so osmocom still using this chipset, and they haven't found an alternative yet |
14:16.51 | mvaenskae | i hope there is somewhere some line of gplv3 code in all modems |
14:17.03 | DocScrutinizer05 | LOL, no way |
14:17.20 | mvaenskae | guess who would jump in a triangle on the roof :D |
14:17.28 | DocScrutinizer05 | this isn't linux, this is a commercial OS you prolly never heard about |
14:17.48 | mvaenskae | windows special edition of millenium edition? |
14:18.51 | DocScrutinizer05 | those are typically realtime OS which not even use a MMU and separated addr space |
14:20.42 | DocScrutinizer05 | see http://en.wikipedia.org/wiki/Operating_System_Embedded |
14:21.23 | DocScrutinizer05 | >>who at the time was a consultant from ENEA with an assignment at Ericsson<< guess what NovaThor firmware is using |
14:22.07 | *** part/#neo900 freemangordon_ (~ivo@85-118-92-26.mtel.net) |
14:24.02 | DocScrutinizer05 | well, I have to correct myself, of course there's MMU, but many "processes" are just threads of the kernel itself |
14:24.09 | DocScrutinizer05 | and share one addr space |
14:25.59 | DocScrutinizer05 | note that all modern chips have a asymmetrical multicore design |
14:26.41 | DocScrutinizer05 | NovaThor is said to have two A9 and one M4, plus a number of auxiliary processors |
14:28.12 | DocScrutinizer05 | the complexity of the modem can easily outperform the complexity of ate application environment of a smartphone |
14:28.37 | DocScrutinizer05 | s/ ate/ hte/ |
14:28.41 | DocScrutinizer05 | dang |
14:31.01 | DocScrutinizer05 | actually the usual featurephone tasks (gui and addr book and maybe SMS etc) is easily vanishing in between the gaps of a modem OS |
14:32.32 | DocScrutinizer05 | and basically that's how all modern android phones work. they don't have a dedicated APE anymore, the whole APE is running on the CPUs of the modem |
14:32.42 | DocScrutinizer05 | see Jolla |
14:33.37 | DocScrutinizer05 | and since APE and modem (firmware) share same CPU, they of course also share same RAM, and that creates a vulnerability you CANNOT fix |
14:33.50 | DocScrutinizer05 | Neo900 doe NOT use this approach |
14:34.57 | DocScrutinizer05 | when Linux is just a process running in a "VM" as one of many processes inside the modem OS, then how could it ever be safe? |
14:36.14 | DocScrutinizer05 | and "Linus" includes android here, unlike my usual way to distinguish between the two |
14:36.20 | DocScrutinizer05 | Linux* |
14:37.28 | DocScrutinizer05 | guess why Jolla can't publish firware to reflash their phone, not even for the linux part of it |
14:44.50 | DocScrutinizer05 | an OS like ENEA OSE has nothing in common with usual linux or other desktop OS as you know them |
14:46.28 | DocScrutinizer05 | e.g. almost all memory is "static", since you don't want dynamic memory allocation during runtime, since that causes all sorts of problems in realtime environments |
14:46.53 | DocScrutinizer05 | almost all processes are "preloaded" during boot time |
14:48.43 | DocScrutinizer05 | these two points alone are sufficient to change stuff dramatically from what you may be used to, into something that looks rather bizarre to the common developer used to "app" or even kernel development on linux etc |
14:50.39 | *** join/#neo900 mvaenskae_ (~mvaenskae@unaffiliated/mvaenskae) |
14:50.45 | mvaenskae_ | sorry, just got distracted a bit by some colleagues D: |
14:56.20 | mvaenskae | i guess i can understand why modem chips are so expensive though (with the newfound deep insight of how complex they are) |
15:11.09 | DocScrutinizer05 | I guess the average LTE modem chip has some 2000 to 30000 man years gone to its development before it starts selling and earning money for the company |
15:17.33 | mvaenskae | i wonder how that compares to the linux kernel |
15:18.04 | DocScrutinizer05 | the linux kernel is kind of simple compared to a complete modem firmware |
15:19.49 | mvaenskae | i meant in terms of coding time :) |
15:24.14 | DocScrutinizer05 | analog, probably. Though a lot of 2evolution" went into linux kernel |
15:24.21 | DocScrutinizer05 | s/2/"/ |
15:25.48 | DocScrutinizer05 | so while writing a linux kernel from scratch nowadays should be significantly less effort than writing a LTE firmware, the kernel nevertheless might have seen more man years of work going into it during the life span of linux |
15:26.57 | DocScrutinizer05 | in a mere comparison of LOC the kernel loses by magnitudes |
15:27.44 | mvaenskae | oh, wow |
15:27.49 | DocScrutinizer05 | when you consider LOC proportional to coding time to write those lines, then it's clear |
15:28.19 | DocScrutinizer05 | kernel is a few MB, while, as mentioned above, NovaThor firmware is several GB |
15:29.20 | DocScrutinizer05 | no surprise, the firmware is *using* a kernel etc of something like OSE, and then building on top of that |
15:30.05 | mvaenskae | oh, that indeed makes it much larger |
15:30.45 | DocScrutinizer05 | and I didn't even count the man years in OSE in my above considerations |
15:31.19 | DocScrutinizer05 | OSE it relatively "simple" compared to an average linux kernel |
15:31.27 | mvaenskae | is "our" lte chip that immense? |
15:31.48 | DocScrutinizer05 | for sure, I don't see why and how it could be any simpler |
15:31.54 | DocScrutinizer05 | or smaler |
15:38.59 | mvaenskae | i meant is it multiple gigabyte ;) |
15:39.30 | che1 | If this question is stupid I'm sorry, but as you put it, firmwares conceptually sound more simple to me. The 'few' protocols to handle, you mentioned above, fixed address space, processes loaded at boot, ... (Probably I'm greatly mistaking as I'm less experienced) |
15:39.30 | che1 | So I wonder if it's the high optimisation that makes them so complex? Or what (else) have I overseen/underestimated? |
15:44.38 | mvaenskae | che1: i assume a large part of the code is debugging related |
15:48.35 | che1 | In that order of size? And wouldn't it be removed before it's copied to the devices? |
15:49.41 | mvaenskae | i assume the gigabyte sizes are sources, not compiled binaries |
15:50.01 | che1 | Ah, right... |
15:51.25 | mvaenskae | if it was binaries then holy moly would tht chip spy 24/7/365 |
15:51.34 | mvaenskae | the nsa would love it |
15:57.15 | che1 | When you believe some people they already do... |
15:59.56 | mvaenskae | they do it some extend but they are not magic |
16:00.32 | che1 | Was referring especially to the 24/7/365, but yeah... |
16:01.13 | mvaenskae | give them enough manpower and money and they'll do it |
16:01.24 | mvaenskae | but also they can't beat the maths/physics |
16:01.54 | che1 | Have you heard about the latest upcoming FB app? Mic on all the time? They don't need more manpower... |
16:03.14 | mvaenskae | well, for me they most certainly need it, not using FB :) |
16:03.22 | mvaenskae | and with the neo900 i can ditch android |
16:03.35 | che1 | True story... |
16:03.42 | che1 | :) |
16:06.01 | mvaenskae | and as i am beginning to harden my systems and trying to use more alternative solutions which i can verify by source code checking i can get even more off of the grid |
16:06.13 | *** join/#neo900 mordac__ (~mordac@c-24-6-219-139.hsd1.ca.comcast.net) |
16:06.26 | *** join/#neo900 mordac___ (~mordac_@c-24-6-219-139.hsd1.ca.comcast.net) |
16:06.41 | che1 | The problem here is: there will still be people running around having their mic on... |
16:07.08 | che1 | You don't (sadly?) controll their devices... |
16:09.21 | mvaenskae | that is true |
16:09.39 | mvaenskae | that is also what bruce said in his talk which i was able to attend |
16:10.20 | mvaenskae | the problem is, more and more people around you are using these solutions and those are then doing the fieldwork |
16:10.54 | mvaenskae | at least i can give myself and others the peace of mind "if i turn off cellular noone knows where we truly are" |
16:12.42 | che1 | Nothing to add here... :) |
16:13.31 | mvaenskae | there is; this might be the last chance to get such a device, chips are turning more and more into a single non-seperable die |
16:13.51 | mvaenskae | and that gives us no guarantee on what is on and what not |
16:53.59 | DocScrutinizer05 | ~ping |
16:53.59 | infobot | ~pong |
16:57.43 | DocScrutinizer05 | mvaenskae: exactly |
16:59.05 | DocScrutinizer05 | and about FB: ROTFL! I suggested that for Siri, to do all sorts of *really* funny stuff, like listening to people next table that one single microphone couldn't even understand what they say. 8 mics spread across the room *can* do that |
17:00.03 | DocScrutinizer05 | exact acoustic location of arbitrary events, be it bomb explosions, cars driving by, or airplanes (incl altitude!) |
17:00.42 | DocScrutinizer05 | there's a DAMN LOT you can do with a few millions of always listening microphones all across the country |
17:01.58 | DocScrutinizer05 | shoot a gun, 5 s later the alarm not only located your position to the meter but also the direction in which you shoot, and probably even the type of rifle you used and if you hit your target or not |
17:02.48 | DocScrutinizer05 | they can't _really_ plan to do that! |
17:03.51 | dos1 | amazing |
17:04.06 | dos1 | and so scary at the same time... |
17:04.48 | DocScrutinizer05 | if they do, I'd have to finally build my microwave pulse gun, and spread a small extension to all FB apps that not only enables mic but also _the speaker_ and shouts "aim at here!" when I do my special whistle |
17:05.30 | DocScrutinizer05 | ;-) |
17:08.00 | parmegv | DocScrutinizer05: I love this conversation. I'm a Computer Engineer with little knowledge about real Electronics: how could I start learning about all these things? just reading? reading and coding? what to read? what to code? |
17:08.01 | *** join/#neo900 wpwrak (~werner@154-164-231-201.fibertel.com.ar) |
17:08.47 | DocScrutinizer05 | parmegv: I have no advice how to learn having weird ideas like I do. Maybe you're born with that ;-) |
17:09.02 | parmegv | hahaha |
17:09.07 | mvaenskae | parmegv: one important asset is to try and break stuff |
17:09.25 | mvaenskae | and try to think out of the box, think of what stuff cannot do |
17:10.07 | mvaenskae | also reading up on them, surrounding yourself with security experts/following their blogs/websites can help |
17:10.12 | DocScrutinizer05 | actually long times in my life my "nickname" (though there wasn't such stuff back when) been (translated) "expert for incorrect handling" |
17:10.38 | mvaenskae | one very famous person whom i just met on wednesday and inspired me was bruce schneier |
17:10.56 | mvaenskae | he had a talk here at a local university and i HAD to attend ;) |
17:11.06 | parmegv | :) |
17:11.28 | DocScrutinizer05 | I bet on that (attending) |
17:11.42 | mvaenskae | DocScrutinizer05: i have his talk recorded |
17:11.47 | DocScrutinizer05 | :-) |
17:11.47 | mvaenskae | on mp3 192kbit/s |
17:12.10 | DocScrutinizer05 | don't you know this is illegal? ;-) |
17:12.13 | parmegv | Currently I'm freelancong for LEAP, so I'm learning more and more about security :) |
17:12.20 | wpwrak | not even lossless ? that's cheap ! |
17:12.26 | parmegv | *freelancing |
17:12.37 | DocScrutinizer05 | what's LEAP |
17:12.39 | DocScrutinizer05 | ? |
17:12.39 | mvaenskae | DocScrutinizer05: he used a microphone, it could have been recording all along |
17:12.46 | parmegv | leap.se |
17:12.54 | mvaenskae | wpwrak: either that or wav and wav uses way too much storage |
17:12.58 | mvaenskae | also makes bad for online sharing |
17:13.07 | mvaenskae | it was roughly 70 minutes |
17:13.09 | parmegv | encrypted email, easy vpn |
17:13.25 | DocScrutinizer05 | btw, may I introduce wpwrak: Werner Almesberger. Hi! And welcome! |
17:13.38 | parmegv | In fact I joined this channel in part because we have similar interests |
17:14.03 | mvaenskae | is Werner the newest member of neo900? |
17:14.06 | DocScrutinizer05 | yes |
17:14.09 | parmegv | wpwrak: welcome yo this beautiful channel |
17:14.13 | wpwrak | that's me, yes |
17:14.15 | parmegv | *to |
17:14.17 | wpwrak | thanks ! :) |
17:14.34 | parmegv | what are you going to do in neo900? |
17:14.49 | mvaenskae | wpwrak: "welcome aboard" <-- i would want to write that but i am just an enthusiast, not someon working with neo900 ug ;) |
17:15.21 | wpwrak | first i'll poke around a little, see if i can spot anything immediately suspicious in the design. an extra pair of eyes. |
17:15.46 | wpwrak | later on, i may make the production testing process |
17:15.49 | DocScrutinizer05 | which is always extremely appreciated |
17:15.58 | mvaenskae | well, i gotta go now home but i'll be back in a moment, wlan-ap :) |
17:16.59 | DocScrutinizer05 | damn, I again missed to go shopping |
17:18.13 | DocScrutinizer05 | waves |
17:18.16 | DocScrutinizer05 | bbl |
17:18.30 | DocScrutinizer05 | wpwrak: a bit of backscroll |
17:18.47 | DocScrutinizer05 | [2014-05-23 Fri 18:01:54] <che1> Have you heard about the latest upcoming FB app? Mic on all the time? They don't need more manpower... |
17:19.24 | DocScrutinizer05 | [2014-05-23 Fri 18:59:04] <DocScrutinizer05> and about FB: ROTFL! I suggested that for Siri, to do all sorts of *really* funny stuff, like listening to people next table that one single microphone couldn't even understand what they say. 8 mics spread across the room *can* do that. shoot a gun, 5 s later the alarm not only located your position to the meter but also the direction in which you shoot, and probably even the type of rifle you |
17:19.25 | DocScrutinizer05 | used and if you hit your target or not |
17:19.58 | DocScrutinizer05 | I guess I even chatted with you about that topic, no? |
17:20.00 | wpwrak | (mic on) add a FIR-capable low-power camera and total control is yours :) |
17:20.05 | dos1 | DocScrutinizer05 is posting after "bbl" - he'll be there for next few hours until he decides to go "afk" again and repeats the story ;) |
17:20.28 | DocScrutinizer05 | dos1: ssssh! that's secret! |
17:20.35 | DocScrutinizer05 | o/ |
17:25.52 | *** join/#neo900 mvaenskae (~mvaenskae@unaffiliated/mvaenskae) |
17:26.00 | mvaenskae | and back :) |
17:26.22 | mvaenskae | wpwrak: you better make videos of the testing process =p |
17:27.44 | wpwrak | you mean, per device ? :) |
17:27.50 | mvaenskae | DocScrutinizer05: also, how would it be illegal to record a public presentation? |
17:28.52 | mvaenskae | wpwrak: per function that works and is accompanied by a cool flashy outcome, like firecrackers! |
17:29.58 | wpwrak | there's sparks and smoke -> "it works". hmm :) |
17:30.37 | mvaenskae | yes! |
17:30.56 | mvaenskae | actually, could you guys implement a killswitch for the memory system? |
17:31.18 | mvaenskae | like me pressing a button and pooof, crispy chips |
17:32.08 | DocScrutinizer05 | we're trying hard to NOT inmplement any such killswitches |
17:32.16 | mvaenskae | :( |
17:32.56 | mvaenskae | it would make a great show and might work in emergencies as a flash/smoke grenade though D: |
17:32.57 | DocScrutinizer05 | risk of those getting triggered by accident or by a rogueware is way too high |
17:33.43 | mvaenskae | wait, you're trying not to... is there something i should know of? =p |
17:33.49 | wpwrak | there's an easier approach. all you need is a screwdriver and a cat. 1) open device to get access to the keyboard. 2) rub cat. 3) discharge into exposed contacts in device. 4) repeat a few times. |
17:34.04 | DocScrutinizer05 | and at least where I live, it's even illegal to sell or possess such stuff without special permission |
17:34.14 | mvaenskae | wpwrak: my cat is quarter a globe away D: |
17:34.17 | wpwrak | this should produce some cat-a-strophic memory loss |
17:34.27 | mvaenskae | DocScrutinizer05: wät?! |
17:34.50 | wpwrak | the germans are serious about not allowing pets in apartments :) |
17:35.04 | dos1 | :D |
17:35.08 | mvaenskae | how does that need special permission? Ö.ö |
17:35.42 | DocScrutinizer05 | any explosives of any kind not allowed to sell without special type approval and permission |
17:35.54 | mvaenskae | dos1: btw, nice videos on the prototype! |
17:36.06 | mvaenskae | DocScrutinizer05: no need for explosives, just some extra volts |
17:37.05 | dos1 | mvaenskae: thanks! not the last ones for sure :) |
17:37.12 | DocScrutinizer05 | extra volts are a poor "killswitch" |
17:38.17 | DocScrutinizer05 | when you really want to phyically destroy the data, you better do it right or it's a big scum |
17:38.34 | mvaenskae | DocScrutinizer05: why wouldn;t the volts work? |
17:38.46 | mvaenskae | dos1: that's the spirit :) |
17:39.01 | DocScrutinizer05 | because the only burn out a very limited area of the chip, while leaving the storage cells intact |
17:39.52 | DocScrutinizer05 | that's like deleting the partition table on a HDD, though on a more hw level |
17:40.11 | DocScrutinizer05 | it can get fixed "easily" when somebody really wants your data |
17:40.41 | mvaenskae | :( |
17:40.42 | DocScrutinizer05 | a simple "erase all" command would work better than "some volts extra" |
17:40.52 | mvaenskae | oh... interesting |
17:41.19 | mvaenskae | i didn't realize that |
17:43.29 | mvaenskae | well, it was worth a shot ;) |
17:43.31 | DocScrutinizer05 | get a properly encrypted filesystem, which has a linked list of mutually depending keys for encryption, so when deleting one sector/block you can't decode any of the whole content |
17:46.11 | mvaenskae | luks? |
17:46.25 | mvaenskae | speaking of, i met the dev as well on the same day the bruce schneier talk was :D |
17:46.30 | mvaenskae | the dev of luks |
17:46.32 | DocScrutinizer05 | or even better: use really stron encryption 2k bits or more, and store the (8 chars-)encrypted full length key on a fixed block on NAND |
17:47.01 | mvaenskae | best day ever, the icing on the cake was the signature of bruce schneier on my lemote yeeloong :D |
17:47.07 | DocScrutinizer05 | then your kill switch would just be erasure of that single NAND block |
17:47.11 | mvaenskae | anyways, bus arriving at station, brb |
17:50.19 | DocScrutinizer05 | btw this whole encryption thing is part of what I'm planning to offer as a special high-security OS for Neo900 - eventually |
17:51.24 | DocScrutinizer05 | along with proper ZRTP based SIP and VPN, sandboxed browser and a few other thing |
17:51.32 | DocScrutinizer05 | s |
17:52.01 | DocScrutinizer05 | such OS obviously doesn't come for free |
17:53.14 | DocScrutinizer05 | I will offer it as a service, incl the internet side infra supporting it |
17:54.27 | DocScrutinizer05 | secure net based storage. Update service and repo that doesn't offer any apps that might have security threats, etc |
17:54.33 | *** join/#neo900 mvaenskae (~mvaenskae@unaffiliated/mvaenskae) |
17:54.56 | mvaenskae | and /home :) |
17:57.09 | mvaenskae | DocScrutinizer05: so we are not using plausible deniability but destructive methods from the looks |
17:59.05 | DocScrutinizer05 | where's plausible deniability in fried chips? |
18:00.29 | mvaenskae | true, very true, but sometimes it might just get "stolen" |
18:00.44 | DocScrutinizer05 | plausible deniability is a property of encrypted fs, when you can't tell from looking at the encrypted fs if there's something stored inside or not |
18:00.44 | mvaenskae | and analysis will surely reveal such a sector, wouldn't it? |
18:05.01 | DocScrutinizer05 | what sector? |
18:06.05 | mvaenskae | well, there needs to be some space reserved for such a sector and method for managing it (deletion, usage) |
18:06.21 | mvaenskae | and that in an unencrypted state |
18:08.13 | DocScrutinizer05 | the NAND borne encryption key? you could have 10 of then, on 10 different NAND locations, and each single one of them decrypts an actually existing cryptfs embedded and coexisting on the one encrypted partition, So depending on first digit of your password you enter, you see one of 10 alternative filesystems coexisting on device. Who's going to tell which of them never had any data, which is fake, and which is the true one? |
18:08.49 | mvaenskae | interesting proposal |
18:09.02 | mvaenskae | i like that idea |
18:10.38 | DocScrutinizer05 | and of course 5 of the 10 possible digits might point to a NAND sector that doesn't store *any* valid encryption key. Who's going to tell if it ever had one and if that been the one for the true cryptfs with the protected content? |
18:11.28 | DocScrutinizer05 | perfect deniability accomplished |
18:13.01 | mvaenskae | how are the cryptFS' managed? are they overlapping or of given size in terms of the partition? |
18:13.38 | DocScrutinizer05 | to go even further, you can set up your 2killswitch" in a way so it would start with the true key sector, but decrement pointer when triggered, so next time you trigger it, it erases one of the bogus encryption key sectors. Nobody can tell if you already triggered it or not |
18:14.58 | mvaenskae | this is getting complicated |
18:15.56 | DocScrutinizer05 | obviously the true one needs to have marked as used the blocks inside crypt partition that are used by the other fake filesystems. The fake filesystems however (when existing) can mark as free / used whatever they like |
18:17.08 | DocScrutinizer05 | mind you: from outside you can't tell if a block is used or not, that's a mandatory property of a proper cryptfs |
18:19.37 | DocScrutinizer05 | and there's nothing complicated in an app that when started erased block 10800 in NAND and then changes own pointwer to 10700. Next invocation it erases 10700 and changes own pointer to 10600. ASO |
18:20.39 | mvaenskae | and what happens if the pointer is 00000? |
18:22.20 | DocScrutinizer05 | create files with random names and one block size on 10100, 10200 ... 10900, and place your true encryption key (of course itself being encrypted by your password) into 10800. Place similar keys but pointing to fake cryptfs into the other blocks |
18:22.51 | DocScrutinizer05 | huh? the pointer isn't supposed to go down to anything like 00000 |
18:24.17 | DocScrutinizer05 | of course the pointer in the killswitch is a threat then, since it +might* point to the right key |
18:25.14 | DocScrutinizer05 | but then the killswitch itself most likely living inside your encrypted filesystem, so nobody could reach it after it triggered |
18:26.01 | DocScrutinizer05 | in the fake filesystems you would have other killswitches that point to their according NAND sector. So no need the killswitch decrements own pointer |
18:26.09 | DocScrutinizer05 | sorry, I'm afk now |
18:27.35 | DocScrutinizer05 | and of course the killswitch doesn't simply erase the NAND based key, but replaces it with one that points to a fake fs |
18:30.00 | DocScrutinizer05 | so you have 10 fake and one true fs on your crypt partition, and you have 9 fake and one true key in 10 NAND sectors. after triggering the killswitch you got 10 fake and no true key in NAND |
18:31.12 | DocScrutinizer05 | nobody can tell if you have 2 or 4 or 7 or 10 fs on cryptpart. And nobody can tell if or not you triggered killswitch |
18:32.13 | DocScrutinizer05 | you can memorize passwords for 1 or 2 or all of the fake fs, it doesn't matter |
18:32.57 | DocScrutinizer05 | plausible deniability, nobody can know if you hd one more password than the one you tell them when forced to |
18:33.04 | DocScrutinizer05 | had* |
18:34.07 | DocScrutinizer05 | you need to know ONE fake password at least, so you can state it's the one and only and _true_ password |
18:35.05 | DocScrutinizer05 | of course you even can have multiple _true_ filesystems in use, each with its own password |
18:35.11 | DocScrutinizer05 | cya |
19:19.56 | *** join/#neo900 mordac_ (~mordac_@68-27-229-89.pools.spcsdns.net) |
19:47.29 | *** join/#neo900 mordac_ (~mordac_@216.253.193.158) |
20:51.05 | *** join/#neo900 archtimmy (~archtimmy@unaffiliated/archtimmy) |
21:03.25 | *** join/#neo900 xes (~xes@unaffiliated/xes) |