#qi-hardware IRC log for Friday, 2010-04-30

wolfspraullarsc: can you tell me which registers on the giantplus side are undocumented?01:05
wolfspraulI can test their willingness to share more documentation with us :-)01:05
vegyraupemorning04:17
gnufshttp://aligunduz.org/img/nanonote/timetravelunix.jpg04:18
wolfspraulvegyraupe: believe it or not, the digital picture frame is on the way to you!04:23
wolfspraulDHL just picked it up...04:23
wolfspraulgnufs: nice picture :-)04:24
wolfspraulcan you upload it to the wiki, http://en.qi-hardware.com/wiki/Special:Upload04:24
larscwolfspraul: the display is super stable now. even without setting the undocumented registers04:24
vegyraupewolfspraul: wow! great! yippie!04:25
wolfspraullarsc: oh, why that?04:25
wolfspraullarsc: about more recent kernel versions, I read somewhere 2.6.33 introduced something called 'compressed memory'. could this be helpful on the ben?04:26
larscwolfspraul: i used their init sequence, just without the undocumented registers04:27
wolfspraulhmm, OK04:27
larscwolfspraul: yes, it can be used and i've already been looking into it.04:27
wolfspraulif you can tell me the numbers of undocumented registers I can still ask04:27
gnufswolfspraul: http://en.qi-hardware.com/wiki/File:Timetravelunix.jpg04:28
zearhey wolfspraul, what's with that CECT C3100 phone?04:28
zearare you planning to port linux to it?04:28
wolfspraulah yes04:28
larscwolfspraul: register numbers 0xac, 0x96, 0x9704:28
wolfspraulzear: another thing on my list first - you mentioned a jlime meeting in poland. Can you add it to http://en.qi-hardware.com/wiki/Qi:Current_events04:29
wolfspraulso the C310004:29
wolfspraulpuh, bit of a long story04:29
zearwolfspraul, i'll do it as fast as i get sure it's official and everyone is welcome04:29
wolfspraulQi is copyleft hardware, not just Ben NanoNote04:29
wolfspraulwe are interested in adding GPS, see the talk about a Garmin-like device for OSM04:29
wolfsprauland of course also interested in phones04:30
zearyeah, i realize04:30
wolfspraulso we are following osmocombb04:30
zearthat's why I ask - is it planned to be a linux phone?04:30
wolfspraulfor now osmocom uses ti calypso and maybe locosto chipsets to hack into04:30
wolfspraulbut harald has an eye on mtk as well04:30
wolfspraulso when he asked on the list I offered a bit of MTK research support, since I just happened to be in shenzhen04:30
zearif you guys ever manage to port linux to a similar phone and still keep it at the low price, i'll definitely buy one04:31
zearor even two04:31
zeari always wanted to have a linux phone that looks like a phone and works like a phone, unlike freerunner ;)04:31
wolfspraulhttp://lists.osmocom.org/pipermail/baseband-devel/2010-April/000364.html04:32
wolfspraulzear: you would have to buy more like 50K and become a big disti for Poland :-)04:33
wolfspraulbut yes, I share your sentiments and of course we work on this stuff04:33
zearhaha, not gonna happen ;D04:33
wolfspraulwho knows04:33
wolfspraulif I can sell you these phones for 15 USD a piece, and we can get really functional software on it, you might find people in Poland that are very interested in financing it all. like large retail chains etc.04:33
wolfspraulbut lots of pieces are missing still04:34
zearif it had features like encrypted calls and texts04:34
zearthen it would probably sell good04:34
wolfspraulso one by one04:34
wolfspraulosmocom is being developed04:34
wolfspraulit's a long long way04:34
wolfspraulI'm actually somewhat doubtful about the GSM stacks since they are huge and moving forward04:34
wolfspraulbut the project is great still, and if I can support it a bit from the MTK side I'm happy to do that04:35
wolfspraulthis is 'alpha' stage though04:35
zearso it's gonna take years, eh?04:35
wolfspraulyes04:35
wolfspraulrealistically04:35
wolfspraulactually the Ben is a step there04:35
zearstill, it's something really interesting04:35
wolfspraulyou just need to see all the pieces04:35
wolfspraulbut unfortunately there are lots of pieces04:36
wolfspraulthat's why the SAKC, even Milkymist One, Elphel cameras, etc. all fit into this04:36
wolfspraulwe'll see04:36
wolfspraulso for this wiki page, yes I bought some of those phones, and tried to see how cheap & fast I can dig into the PCB04:36
wolfspraulwas quite successful04:36
wolfspraulI basically found a shop on the street that can do what you see on that wiki page for 100 USD in 2 days or so04:36
zearis it a form of reverse engineering?04:37
wolfspraulI'm not going to distinguish between forward engineering and reverse engineering.04:37
wolfspraulI only do forward engineering :-)04:37
wolfspraulI buy phones, then I take apart my property.04:37
wolfspraulthose pictures are lined up in physical order, i.e. you can zoom through the product04:38
zeari can only imagine how many iphones that company gets for taking apart :D04:38
wolfspraulnot many04:38
wolfspraulno point04:38
wolfspraulthere is not that much design work in the PCB anyway04:39
zearto produce cheap ripoff, i guess ;)04:39
wolfspraulno no04:39
wolfspraulit's all different than you think :-)04:39
zearah04:39
wolfspraulso anyway, did this explain the C3100 page in the wiki?04:40
zearwell, yeah, the PCB is completely different in the ripoffs, they just use the same form factor and look like the originals04:40
zeari guess so :)04:40
wolfspraulit's a support effort for a neighboring project, osmocombb04:40
wolfspraulwe'd love to do just what you are asking for04:40
wolfspraula phone that looks like a phone, and runs 100% free Linux04:40
wolfspraulbelieve it or not, by buying a nanonote, and hacking on it, and helping us with kernel, apps, Debian, etc. you are contributing towards that direction04:41
wolfspraulat least I hope so, and we don't die before reaching the goal :-)04:41
zear:D04:42
zearjust don't follow the way of openpandora and you'll be fine ;)04:42
larscwhat did they do wrong?04:42
zeartaking preorders for unexistant hardware04:43
zearit's about 3 years since they started the preorders04:43
zearand they still don't have a finished product04:43
wejpyeah, that really sucks04:43
zearand in the meanwhile many people resigned04:43
wejpalso they are not building open source hardware, so you can't really compare that anyway04:43
zearand they had to return their preorder money back04:43
vegyraupezear, wejp: the problem is not the way they finance the project, but the fact that they had no idea what they were getting it to04:44
vegyraupeit = in04:44
wejpvegyraupe, yes you are right, that is the biggest problem there04:44
zeari agree04:44
vegyraupeif you don't wanna be dependent on some ppl with big money you have to find alternative ways04:45
vegyraupeand I like the approach of community supported development better and better04:45
zearbtw, when can we expect any public information about how are the bens selling?04:45
vegyraupeof course with modifications to OP's way, but the same general idea04:45
wejpvegyraupe, yes, but in case of "open pandora" except the money they took for the preorders from the community, there hasn't been ver much influence on the actual product by the community04:46
vegyraupewejp: as I said with modifications ;)04:46
vegyraupeI would see the community as the customer04:46
vegyraupeso the customer creates specs04:46
wejpnot that i don't like the hardware design, i do, but still, it was only a very small group of people building a handheld console04:46
vegyraupethe company makes an offer04:46
vegyraupeand then the customer looks over it04:47
vegyraupeand the two parties agree to a plan04:47
wejpyyeah04:47
vegyraupewhen modifications need to be made04:47
vegyraupeit goes back to the customer for a decission04:47
vegyraupeand of course the customer wouldn't pay the whole sum up front04:47
vegyraupebut only part of it to finance development04:48
wejpproblem with the pandora was, it turned out to be even more expensive than what they got through the preorders04:48
vegyraupei don't follow them that closely04:49
vegyraupebut I am sure if that happened than probably because they hadn't done anything like that before04:49
wejpyes04:50
vegyraupeand considering that, it is actually amazing what they built04:50
wejpof course it is04:50
vegyraupei thinkit is a very good proof of concept and commendable that they even tried04:51
wolfspraulzear: we sold about 500 Ben so far04:51
zearwolfspraul, of how many in total?04:51
wolfspraulwhich is a good start, roughly in line with my estimates. Problem will be to sustain it now, which depends on improving software.04:51
wolfspraulzear: first batch we made 100004:51
zeari guess that's a good result?04:51
wolfspraulwe are taking all this very carefully04:52
wejpoh 500 sold out of 1000 isn't bad :)04:52
wolfspraulwe are spending a lot of time to visiting vendors, digging deeper into the production process04:52
wolfspraulwell we need to sell all 100004:52
wolfsprauland then the second 100004:52
wolfsprauletc04:52
wolfspraul:-)04:52
wejpsure :)04:52
wolfspraulalso drive price further down04:52
wejpi'm sure a future nanonote version which comes with wireless included would sell much better as this seems to be what most people interested in the device complain about in the first place. i for myself would be fine with just having usb host port (preferably a type a connector on the device), but internal wireless would be nice too04:56
wolfspraulmy current Ya NanoNote looks like Ben, but 4760 CPU, USB On-The-Go, more RAM (maximum supported by one chip), hopeRF RF module04:59
wolfspraulbut who knows it's all moving04:59
wolfspraulthe seeds are growing05:00
wejphow much RAM does the CPU support?05:00
wolfspraulwhich one?05:00
wolfspraulthe goal now is to max out the Ben05:00
wejpthe one you've chosen for the Ya05:00
wolfsprauldon't know05:00
wolfspraulfor one it support mobile DDR05:00
wolfspraulfinaly05:00
wolfspraulBen has SDRAM05:00
mortysHi all05:01
wolfspraulwejp: 'chosen' is the wrong word, we haven't chosen anything05:01
wolfspraulfocus now is Ben05:01
wejpbtw, the cpu speed of the ben which i would least complain about, memory bus is a little thin but cpu speed is fine05:01
zearwhat is a rf module? some sort of a wireless radio?05:01
wolfspraulwe need to show that we can bring out the performance of the HW to the normal end user05:01
wejpwolfspraul: yeah, i understand, i meant just the configuration you spoke about :)05:01
wolfspraulzear: yes05:02
wolfspraulthere were some mails about it on the list. just being evaluated right now, together with many other things...05:02
wejpso you are going with USB OTG to be able to keep the same case?05:02
zearyeah, read the mails already05:02
zearbut i don't really understand the idea behind it ;)05:03
wolfspraulzear: what do you mean with "don't understand"?05:04
zearwell, what it could be used for05:04
zearis this a well know and popular standard, that many devices use?05:05
zear*known05:05
larscotg?05:06
zearon the go, can pair two otg devices without a need of usb host05:06
larscno, i was wondering if you were asking about otg05:06
wejpnot only that, you can also attach a usb slave device to a usb otg capable device05:07
zearno, i was asking about the RF05:07
wejpi was asking about otg, if it has been "chosen" to be able to keep the ben nanonote case05:07
zearwejp, so how is it different than a host then?05:07
wejpzear, it can also run as normal slave mode, and the connector is usually the small slave port, not the large type A port05:08
zearwejp, that means it requires a mini>normal converter05:08
larscotg is basically a host and a device behind a inteligent transiver which decides which one is to use05:08
zeari have one and it's like 2,5cm big05:08
wejpzear, yes05:09
wejpthat's why i would prefer a USB A type port on the device, but i understand if this will not be done, to keep the ben case as it is05:09
zearyep05:09
vegyraupewejp: do you have any plans on making gmu playing videos as well?05:34
zearha, that would be cool05:34
vegyraupezear: +1 :)05:35
zearvegyraupe, but you mean a video support written from scratch, or some kind of a gui for mplayer?05:35
vegyraupezear: no idea about the details :)05:36
wejpvegyraupe: writing a video plaer isn't a simple task, so it requires a lot of work to do that05:36
wejpi'm not sure yet, if i'll do that. maybe i'll put some work into videoplaying when it is mostly feature-complete from the musicplayer point of view.05:36
zearit is mostly feature-complete music player already :)05:37
wejpzear, it it mostly is, there are only a few things left, i would like to add :)05:37
zear:)05:37
wejpand if you have every looked at ffmpeg (which one could use for creating a video player), you probably know that even doing very simple things with that library are rather complicated to say at least05:39
wejps/every/ever05:39
zearyeah, i always have to study the manual before i do anything with it05:39
wejpand that does not even involve putting a video on the screen yet05:40
wejpwhich also needs to be done, and it needs to be done in a fast way05:40
Action: vegyraupe just wanted to spark the thought :)05:40
wejpthen, ou need to keep video and audio in sync...05:40
wejphehe, yeah, that's fine :)05:40
zearoh, i didn't even know there's a "ffplay"05:41
wejpi just wanted to point out that this is a rather long way to go, if i want to do that ^^05:41
zearshould it be faster than mplayer?05:41
vegyraupecoz then one could create a "mediaplayer" which comes with gmu as the graphical interface :)05:41
wejpzear, not necessarily05:41
vegyraupecreate here: build05:41
wejpvegyraupe that could be an option, but i don't like it that much becaue until now, Gmu is a stand-alone player that does not need an external program to play stuff05:42
wejpof course you need some librarries but it doesn't call external programs05:42
vegyraupeexternal programs?05:43
zearwrite your own video processing lib then :D05:43
wejpwell, i'll think about the whole video stuff a little  :)05:43
vegyraupewejp: by mediaplayer I meant a physical device05:43
wejpzear, hehe, yeah, if i find the time to do that :D05:43
wejpvegyraupe, oh, ok i see :)05:44
vegyraupemy sis bought one of these05:44
vegyraupesized as a cheap digi cam05:44
vegyraupewith a screen05:44
vegyraupeand earplugs05:44
wejpi agree that it would be nice to use the same interface for all media formats05:44
vegyraupeand all it does is play videos and music05:44
vegyraupeand the avt2 boards and even teh ben is not too far away from that05:44
wejpeah05:45
wejpy05:45
wejpargh, my y button works only every second time i hit it or so05:45
zearvegyraupe, weird thing is, that even with ingenics patches and with an ipu, mplayer on the dingoo is really, really slow05:47
zearyou can't play anything greater than 320x240 without lags05:47
wejpthe video player from the default firmware isn't much better though05:48
zearwell, mplayer has to scale everything to 320x240 anyway, but it's more convinient to just let it play a dvd quality movie than to fiddle with ffmpeg and convert it to 320x24005:48
vegyraupezear: that's not quite my department ;)05:48
zear:P05:48
zearnot to mention a nononote, which has less available ram05:48
zear*nanonote05:48
wejpdoes the nanonote cpu even have the ipu?05:49
zearhmm.. no idea05:49
wejpit does have those SIMD instructions, but the IPU is an entirely different thing05:49
wejpi think only the 4740 has it not the 4720, but i could be wrong05:49
larscboth have it05:53
wejpyeah, just checked the datasheet, you are right05:53
larscthe only real difference between 4720 and 4740 is the form factor and the pin which are available05:53
wejpstill, if the dingoo has problems with playing videos it will be a real challenge to get that working on the nanonote. the dingoo has twice the memory bandwith compared to the nanonote :|05:55
wejpd05:55
zearingenics mplayer compiled for the nanonote using dingux toolchain just segfaults for me :P05:55
wejpmh05:55
wejpdo you know where it segfaults?05:56
zearsame with the binary some other guy from the community compiled and claimed it worked for his ben05:56
tuxbrainvideo conversation yummy yummy05:56
wejp:D05:56
zearwejp, nope, didn't bother bugtracing it05:56
wejpok05:56
wejpneed valgrind on the ben :D05:56
zear*bug tracking05:56
zeari heard valgrind is good, it helped my coder to track one nasty bug with the Tower of Gowno for the pacc contest05:57
zearwe couldn't find it with either gdb or strace05:57
wejpyes, valgrind is a great tool05:57
wejpi like it much better than stupid gdb05:58
cain__good morning09:28
cain__I have a little question09:28
cain__how can I return to console from gmenu2x?09:28
xiangfucain_, CTRL + ALT + F1.09:30
cain__thanks xiangfu09:30
xiangfucain_,  you can edit /etc/inittab. remove the "tty5/6/7/8"09:33
xiangfucain_, then you can go back to gmenu2x by CTRL+ALT+F509:33
xiangfucain_, you are welcome.09:34
Snakerhey would it be possible to add a Pointing stick http://en.wikipedia.org/wiki/Pointing_stick to future models of the NanoNote, that would make it awesome!10:11
wejpi like that idea :)10:23
zearme too, though do we really have a place for it?10:24
wejpwell maybe not in the middle of the keyboard, but rather above it between speaker an microphone10:24
wejpi would like that better than a touchscreen10:25
zearyeah, unless there's something on the PCB in there10:25
zearsame10:25
wejpyeah, not sure if it is, but shouldn't be difficult to find out :)10:25
darfgarfif anyone's interested, i'll be doing a proper release of nupdf later tonight15:09
vegyraupedarfgarf: hui, cool15:11
tuxbraindarfgarf: Yay!!! I was waiting for this :)15:16
darfgarfsorry for taking so long, the competition judging dragged on15:16
darfgarf(came 9th)15:16
tuxbrainI have not understand a word about you have said :P15:17
darfgarftha's ok XD  nothing important15:18
tuxbrainnot pushing at all, but you will do a Makefile to proper include on next distro?15:19
zeartuxbrain, the pacc competiton darfgarf (and me) participeted in has finished the 31th march, but the results of the contest were announced today15:19
tuxbrainok :)15:20
zearbtw, is it "on 31th" or "at 31th" or yet somehow different? :P15:20
darfgarftuxbrain: maybe, but i don't know much/anything about packaging for openwrt, and it's 1 month until exams start   (zear, 31st :P)15:20
zearso "the" wasn't needed as well? :P15:21
darfgarfthe 31st yes15:21
darfgarfor you could ignore it15:21
zearso i was right after all, hooray for me! ;P15:21
tuxbraindarfgarf: Ok, just make it compile and tell us what are the dependencies, someone (maybe me) will do the work of doing the Makefile, I think this is a key piece of soft for nano15:22
zeari suck at prepositions :P15:22
zeardarfgarf, owrt toolchain is a nightmare to set up15:23
darfgarftuxbrain: i'll include the compile scripts i use, and also compile it statically, so it can just be dropped in and used15:23
zearit's not as simple as throwing it to /opt and exporthing the path15:23
zearbtw tuxbrain, is there any pdf reader for the wiz?15:24
darfgarfzear: so it can wait or be done by someone else, i shouldn't even be touching it but it's the end of the week and i haven't stopped work at all15:24
zeari could recompile nupdf for it15:24
darfgarfzear: think there is15:24
darfgarflook up Mipdf15:24
tuxbrainzear, really I dont understand why you have the need of move the tolchain , it's easier to do the Makefile really15:25
zeardarfgarf, ah, you're right15:25
zeartuxbrain, because i want to just export the path and run ./configure --host= --target=15:25
zearand with the owrt toolchain you have to manually edit the makefiles becuse libs are located in an unusual place15:25
zear(the whole toolchain is divided in two dirs)15:25
tuxbrainI supose is an advantage being an ignorant in this case, cause I 'm not used to  compile in anyway , anyway is fine for me :)15:27
zeari might be wrong after all15:28
zeari hope i am and there's an easy way to set up owrt toolchain :)15:29
darfgarf_meh, i'll put it as a todo at the end of a very very long list of things15:29
zear:P15:30
tuxbrainowrt is thinked to do full firmware and I think is good for that porpouse, I might recognize is not such ergonomic for individual ports, but in the long run (make things manteinable on time) I think is a better aproach15:31
zearyeah, i agree to that15:31
zearbut that also means you need to set up every single new port to work with it15:31
zearand that takes time15:31
darfgarf_and keep it compatible with the dingoo at the same time15:32
zearit's just a matter of a makefile15:32
darfgarf_though that's just doing things in a slightly different way really15:32
zearif you have one single makefile for your project, it should be compatible with both the dingoo and owrt15:32
tuxbrainI hope we should find a way to make two words compatible15:32
zearjust use different makefiles for dingoo and owrt15:33
zearwe could always do a standalone owrt toolchain15:33
zearfor easy porting15:33
darfgarf_just uploaded v0.2 of nupdf to the dingoo file archive, here's the link:   http://dl.openhandhelds.org/cgi-bin/dingoo.cgi?0,0,0,0,113,34516:13
tuxbraindarfgarf:is this binary also compatible wit BNN16:28
tuxbrain?16:28
darfgarf_bnn?16:28
darfgarf_or if you slipped and meant to put nn, then yes, nupdf is the nn version, nupdf.dge is for dingoo (keybindings are the main difference)16:29
darfgarf_should just drop in and work...hopefully16:29
tuxbrainbnn=Ben NanoNote16:38
darfgarf_ah right, i just use nn (but the answer is yes put simply)16:39
kristianpaulbnn = nano = ben = ... :)16:55
kristianpaulbien16:55
kristianpaultuxbrain: what's the maximun speed you got with arduino serial stuff?16:55
tuxbrainI had not played with it, I just stay in the bnn defaults 57600 bps17:00
tuxbrainbut the theorical maximum is 115200 bps17:02
qi-commitsCarlos Camargo: Adding a simple plasma  example read write char short int, adding simulations http://qi-hw.com/p/nn-usb-fpga/622f59823:23
--- Sat May 1 201000:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!