#qi-hardware IRC log for Monday, 2011-10-17

whitequarkwpwrak: I've found the cause; the USB just cannot provide the current required00:35
whitequarkit works just fine from 13V 300mA power supply00:35
wpwrakquite power-hungry then00:35
wpwrakor your USB is too weak00:35
whitequarkyep, CCFL backlight isn't very energy-efficient00:36
whitequarkalso: the LCD itself eats 120mA@5V, and backlight wants 300mA@12V00:36
whitequarkthat's at least 720mA@5V total00:37
wpwrakseems that you need your own power supply, too00:38
whitequarkah, and also the module is powered through a FET (which was not selected by me). looks like additional 0.5V are falling on it00:38
whitequarkI should have checked that in first place :/00:38
whitequarkthat FET thingy is quite mysterious I'd say. see: it's a P-channel FET and the INVTOFF signal is active low, so I've added two inverters on BC54700:39
whitequarkwhen I feed it low level, it turns off, when I set it high, it turns on00:40
whitequarkwell, it works just the same if I cut the wire from the inverters to FET gate in two00:40
whitequarkexactly.00:40
whitequarkso it's floating in the air.00:40
whitequarkhow the hell it manages to work?!00:40
wpwrak;-)00:41
whitequark(I've simply forgot to solder in the gate. And when I noticed it... that was a shock, to say at least.)00:42
whitequarkthe inverters do not alter supply voltage in any way00:43
whitequarkthey just do some black magic, and FET does too, and it works.00:43
wpwraka circuit designed by wizards ;-)00:47
whitequarkthen I'm a wizard00:48
whitequarkwhich is doubtful :D00:48
DocScrutinizerwpwrak: ( how do you like that ? ) Well, it's OK if needed. I just don't see the exact purpose - rather contact the whole metal sheet with only one metallic spacer and use 3 plastic spacers01:27
wpwraknot sure if mechanically compatible plastic spacers are available01:29
DocScrutinizerusually such RC is used for "floating" GND that may have DC offset, or 50Hz, and you still want to make sure it's effectively grounded for ESD and RF01:29
DocScrutinizerwell, just make those R and C NC, same effect :-D01:30
DocScrutinizerI gather you do all this to avoid GND loops?01:30
wpwrakNC ? then the feet/spacers wouldn't be grounded anymore01:30
wpwraki don't know why all this is there :)01:31
DocScrutinizeraaah01:31
wpwraki would just isolate them :)01:31
DocScrutinizerthought you're trying to sell it to me01:31
wpwraknaw, i01:31
wpwrak'm trying to figure it out ;-)01:31
DocScrutinizer3 of 4 are nonsense anyway, better make them NC01:32
DocScrutinizerand after you did, you can make the 4th 0R01:32
wpwrak the only use this would have would be to ground the conductive tape, right ?01:33
DocScrutinizerright, then you would need 2 times 0R01:34
DocScrutinizerwhen using that alu tape idea01:34
wpwrak2 times ?01:35
wpwrakah, for each leg of the X ?01:35
DocScrutinizersure, the 2 stripes are not connected, so each one needs a post to gnd it01:35
DocScrutinizerjust ONE post per stripe01:35
wolfspraulaw_: hi good morning!01:36
DocScrutinizermore like good night :-D01:36
wolfspraulnah01:36
wolfspraulI'm sitting with my Venti Coffee here in bright morning sunlight01:36
aw_wolfspraul, good morning hi01:36
wolfspraulaw_: since it's Monday morning, let's chat a bit about the plan going forward01:37
DocScrutinizerI'm afraid I'll do the same when I don't manage to get some sleep next 3 h01:37
wolfspraulslowly rc3 is settling down01:37
wolfspraullet me look at the wiki page a bit :-)01:37
wolfspraulwhat's your plan this week?01:37
DocScrutinizerhi aw_ 01:38
aw_since the remaining 'packed for sale' is only 7 sets now and 23 set 'Avail-fix2b-l3' ready already; so need to do packing/assemble first this week01:39
aw_DocScrutinizer, hi 01:39
wolfspraul41 units have been sent out01:40
wolfsprauljust counted01:40
kristianpaul:o01:40
kristianpaul:-)01:40
wolfspraulok, those 7 need to be reflashed/locked again01:40
wpwrakpretty good for no real launch :)01:40
aw_wolfspraul, hopefully you don't have a 10-Pack shipped in 2 ~ 3 days. :-)01:40
wolfspraulprobably not01:41
wolfspraulso first you re-lock the 7, then finish another 2301:41
aw_so I may need to fo a batch work for 23 set in one time.01:41
wolfspraulnice01:41
wolfspraulthat will bring the total that are 100% good to 7101:41
wolfspraulthen we look at the other 1901:42
wolfspraulaw_: how about that other rc2 board?01:42
aw_i can do re-lock later01:42
wolfspraulit sounds like you need another week or so before we start looking into the 19, and talk about next runs such as rc4, labsw, or some other power supply I am currently investigating01:42
aw_aw_, no further on that bad rc2 board which I'll back to see it after 23 assembles and 7 relock.01:42
wolfspraulok01:43
wolfspraulWerner continues with the NOR corruption, it sounds like he will get to the real bottom of that this week01:43
wolfspraulthe relay keeps clicking away :-)01:43
Action: DocScrutinizer joins in on a bit NOR curruption, it's fun01:44
wolfspraulaw_: how much does a cheap ESD gun cost in Taipei?01:44
DocScrutinizerphaser01:44
wpwrak"your life" ;-)01:44
wolfspraulwpwrak: actually we are shipping out nice rc3 units almost as fast as Adam can make them01:45
kristianpaul*g*01:45
aw_wpwrak, hi i just realized the dumpotp, and 'poke' peek' cmd in jtag, ignore my email on list, i should have tried script then asked. ;-)01:45
DocScrutinizer:-P01:45
wpwrakwolfspraul: i'm actually a little worried about that ;-)01:45
wolfspraulnah01:45
wolfspraulrc3 was work intensive, so be it01:45
wolfspraulwe learn from that, and then rc4 will be better, hopefully (ahem)01:45
aw_wolfspraul, a cheap ESD gun? mmm...need to ask or/ check here wen first.01:46
wolfspraulyes01:46
wpwraki think you need to find an assistant for adam, to help with those routine tasks. as runs get larger, the workload will even get worse01:46
wolfspraulif we can find a cheap one, maybe you can get one...01:46
wolfspraulone by one01:46
wolfspraulno customer was turned back, because we don't have more customers either01:46
wolfspraulit needs to scale in sync01:46
wpwrakwhat number do you have in mind for rc4 ?01:47
wolfspraulyeah well01:47
DocScrutinizerwolfspraul: I can send you some fractions of the lightning that hit my house a few weeks ago - I bottled it ;-D01:47
wolfspraulI was hoping nobody asks for a while ;-)01:47
wolfspraulI don't know01:47
wolfspraulthere's a lot of things I want to achieve with rc4, including maybe boom and digikey sourcing (for some parts, not all)01:48
wolfspraulfrom a manufacturing standpoing, it should be 150 or so01:48
wolfspraulstandpoint01:48
wolfspraulbut then I want to improve the quality & product as much as possible, the whole dvi-i thing standing out01:49
wolfspraulor DocScrutinizer's 'X'01:49
wolfspraulwhat's the speed of rc3 sales? can we find new customers? distributors?01:49
wolfspraulsome parts may have long lead times, such as power supplies (2 months)01:50
wolfspraulthe cases took 8 months ;_)01:50
wpwrakhow are rc3 sales composed ? mainly people who have been waiting already ?01:50
wolfspraul:-)01:50
wolfspraulno01:50
wolfspraulnew random people from here and there01:50
wpwrak(case) ouch !01:50
wolfspraulwell01:50
wolfspraulroh can probably do it faster, but for 150, it's at least a month or so01:50
wolfsprauland it's very expensive01:51
wpwraktry kristianpaul's shop ?01:51
wolfspraulmaybe we can improve the metal sheet, buttons01:51
DocScrutinizer*cough* I get you those in < 1 week01:51
wolfspraulnah01:51
wolfspraulyou guys are doing easy superficial talk :-)01:51
Action: kristianpaul lacks QA01:51
DocScrutinizernot really, I got my acrylic shop here, and they do awesome things01:51
DocScrutinizerincl arbitrary CNC iirc01:52
wolfspraulthe screws are still not right01:52
wolfsprauleven though roh already spent an ungodly amount of time on them01:52
wolfspraulthe feet cost a fortune01:52
wpwrakbuttons should probably be milled. not sure how smooth the surface would be, but there doesn't need to be any fancy gluing and such01:52
wolfspraulthe metal sheet and insulation layer is a pain01:52
wolfspraulgluing the buttons out of 3 pieces is a pain01:52
wolfspraulacrylic tolerances cause a lot of issues01:52
wolfspraulwidth tolerance01:53
kristianpaulindeed, noo way 01:53
wolfspraulok just trust me - a lot of work01:53
kristianpaulsuperficial*01:53
wpwrakthe width tolerances are a bit outlandish, yes01:53
DocScrutinizergluWHAT?01:53
wpwraki dont think i could find any acrylic with such wide tolerances around here ;-)01:53
wolfspraulso with rc4, I want to improve a lot of things01:53
wolfspraulcheaper, easier to manufacture, etc.01:53
wolfspraulwpwrak: when did you do your last acrylic project?01:53
wpwraki haven't worked acrylic yet01:54
wolfspraulah :-)01:54
wpwrakbut i've been looking around for sources. problem here is to get it in reasonable quantities01:54
wolfspraulthere is always some problem01:54
wolfspraulthe magic with rc3 is that we actually ship a good product01:54
wolfspraulnot some 'but' included01:54
wpwrakthe suppliers sell it in large sheets, often 1 x 2 m or such.01:54
wolfspraulso with rc4, I want to do my absolute best01:55
wolfspraulimprovements all over01:55
DocScrutinizerwpwrak: you need a better shop01:55
wolfspraulfrom a volume perspective, it should be 15001:55
wolfspraulbut I also need to be careful that I don't sit on a lot of unsellable junk later01:55
wolfspraulhardware inventory is nasty01:55
wpwrakDocScrutinizer: the small shops want to do the engraving, so they don't sell the raw material :)01:55
DocScrutinizeryes01:55
DocScrutinizerwell, no01:56
wolfspraulthe pixels of the 'M' logo should be surface scanned, that would look much nicer I think01:56
wolfspraulanyway, there are hundreds of little things01:56
DocScrutinizernot here, they do, but also sell raw01:56
DocScrutinizerany qty01:56
wolfspraulfirst step, let's make the electrical & layout as good as we can01:56
wolfspraulah, I will add Joerg's 'X' idea now...01:56
wolfspraulto here http://en.qi-hardware.com/wiki/Milkymist_One_RC3_Known_Issues01:56
wpwrakDocScrutinizer: hmm, there's one next door. haven't asked them yet. maybe i can get a little piece from them, to get started01:57
wolfspraulwe could even explore a metal case01:57
wolfspraulwhat are the priorities?01:57
DocScrutinizeralso add my concerns about a VGA connector with no real mech support from anything but PCB01:57
wolfspraul#1 find as many paying customers as fast as possible01:57
kristianpaulmay be metal will be friendly with tolerances?01:58
wpwraki wouldn't change the case a lot. it currently "works".01:58
wolfspraul#2 never go 'out of stock', whether rc3 or rc4 or ever, we must always be able to sell a 'video synthesizer'01:58
kristianpaulwpwrak: it changes by it self :)01:58
wolfspraul#3 improve rc4 as much as possible, make it the best possible product we can make01:58
wpwrakDocScrutinizer: isn't that pretty much "industry standard" ?01:58
wolfspraul#4 from a numbers perspective, I'd like to double the run size, i.e. from 80 to 16001:58
DocScrutinizerdon't think so01:58
DocScrutinizermaybe on 2mm PCB gfk01:59
wolfspraulwpwrak: who do you think should have a labsw?01:59
wolfspraulwhat is the future potential of the board?01:59
wolfspraulI'm wondering whether we should make a labsw before/during/after rc402:00
wolfspraulkristianpaul: do you want to manufacture some labsw? :-)02:00
wolfspraulI pay, you export from Colombia to the world! :-)02:00
kristianpaulhum, interesting :)02:00
wolfspraulyes, I think technically it's within reach for you and the shops you have easy access to in Colombia, no?02:01
kristianpaulyes i can do all remotelly from home :)02:01
kristianpaulalmost*02:01
wolfspraulmy problem with labsw right now is that I don't fully understand who should have one, ideally, or what the long term potential could be02:02
wolfsprauldefinitely, that's the way to go02:02
wolfsprauland use local shops, whenever possible02:02
wolfspraulpcb, smt/dip02:02
wpwrak(labsw) adam should have one. after that, i don't care ;-)02:02
wolfspraulcome on, be more visionaire02:02
kristianpaul:)02:02
wolfspraulno other usefulness?02:02
kristianpaulproduction side i see, and agree wpwrak 02:03
wolfspraulI see I need to do the homework myself02:03
wpwrakoh, sure. but "productizing" it would be tricky02:03
wolfspraulread more about what it is :-)02:03
wolfsprauloh no, I don't want to productize it02:03
wpwrake.g., you probably can't sell it as 220 V capable02:03
wolfspraulit works with 110v now?02:03
wolfsprauland only that?02:03
kristianpaulat least you want to sell a board to extress M102:04
wpwrakthere's also the problem with sourcing case parts02:04
wolfspraulno need02:04
wolfspraulmy current understanding is that it's just a hacker tool02:04
wolfspraulfor use in production environment, of course02:04
wolfspraulor stress testing02:04
wolfspraulor whatever yet unthinkable use02:04
wpwrak(110 V) currently it's not designed to be suitable for voltages unsafe to touch. the coming design will be designed with 220 V in mind (with the usual safety margins, etc.)02:04
wpwrakbut i have no clue what safety regulations would have to be observed to make it really marketable as 220 V-ready02:05
kristianpaulwpwrak: next version will allo to add temperature probes?02:06
kristianpaulthat will be very usefull, all in a box02:06
kristianpaulno need for extra hubs or hacked usb cat5 adapters :)02:06
wpwrakfor a usage point of view, i'd consider it "safe for ~30 VDC" and "can operate with 220 V, but treat the whole unit as if it was a live wire"02:06
wolfspraulcome on, no safety regulations02:06
wpwrakkristianpaul: naw, no temp probes ;-))02:07
wolfspraulit's so amazing to me02:07
kristianpaulwhy?!!02:07
wpwrakwolfspraul: right, china :)02:07
kristianpaul:)02:07
wolfspraulChinese stuff is flooding the world02:07
wolfspraulwell, but better02:07
wolfspraulits' flooding the world, right?02:07
wolfspraulit's everywhere02:07
wolfsprauland I know how it's made02:07
kristianpaulyeah02:07
wolfspraulI feel it's like a global conspiracy02:07
wolfspraulon one side you have people who just totally don't care, don't know, dont' want to know, how they are randomly throwing shit together that maybe works or maybe not02:08
wolfsprauland they sell02:08
wolfsprauland on the other side you have those fine, sophisticated, suit-wearing and PhD possessing intellectuals02:08
wolfspraulthat are debating about paragraph A.3 subsection 15.9, some tiny detail in their regulations02:08
wolfspraulbut at home, everywhere in their lifes, they are using the crap from the monkeys02:09
wolfspraulwhat is going on???02:09
wolfspraul:-)02:09
wolfspraulit's hilarious to watch02:09
wolfspraulso whatever02:09
wolfspraulif you care about "safety regulations", then do so :-)02:09
wpwraki think it's a question of exposure02:10
wolfspraulI know one thing for sure - whatever you make will be 100 times more thought-through and more secure than a lot of things that are surrounding me and that I use every day02:10
wolfsprauleven if you think you are not meeting "safety regulations"02:10
wpwrakif we do something bad, we're easily exposed to the consequences02:10
wolfspraulyeah yeah02:10
wolfspraul:-)02:10
wolfspraulI wouldn't touch anything metallic in CHina on the street02:10
wpwrak;-)02:10
wolfspraulafter you've watched the 'electricians' repairing stuff once, you won't02:10
zrafawolfspraul: I do not touch any wall in my building outside my apartment02:11
zrafawolfspraul: I saw the electricians here as well :)02:11
wolfspraulI don't step on any drain covers on the streets either02:11
wolfsprauljust as any chinese wouldn't02:11
zrafaand I am not in china02:12
wolfspraulbecause still so many people are badly injured and die from falls into crappy covers that are loose, breaking, flip over, etc.02:12
wolfspraulwpwrak: you need to show more leadership, for the good of the world. you are the 'safety regulation'02:12
wpwrak;-))02:12
wolfspraul100 times better than any 'regulator' (read: totally incompetent bureaucrat)02:13
wolfspraulseriously, the system is broken. the regulation is just paper, totally meaningless except for the salaries of the people writing them, which are still real.02:13
wolfspraulwe have to take care and make good stuff, and once we think it's good, it is.02:13
kristianpaulwhere is the plug to be pull? :)02:13
wolfspraulwpwrak: who do you think watches the consequences? :-)02:14
wolfspraulthe regulators pouring over paper regulations?02:14
wolfspraulthe factory workers?02:14
wolfspraulthe business people selling what their workers produced?02:14
wolfspraulI think it's either you or nobody... :-)02:14
wpwrakwell, one thing with high voltages is that i don't know that topic too well. i've been googling for the most obvious issues, but i may well be unaware of some others02:15
wolfspraul:-)02:15
wolfspraulyou never stop02:15
wolfspraulthe fact that you even think that already makes the difference!02:15
wolfspraul:-)02:15
wolfspraulanyway, just ranting, sorry02:16
wolfspraulreality is a funny game sometimes02:16
wolfspraulso... labsw only good for Adam?02:16
wpwrakplus, there are some compromises between safety and usability. e.g., the banana receptacles i use would be considered unsafe for the last 10-20 years02:16
wpwrakwell, that's how many i plan to make :)02:16
wpwrakwell, a 2nd one for myself, too :)02:16
wolfspraulwhat's the vision with this thing?02:16
wpwrakdunno. the electronics are be reproduced elsewhere. mechanics are a little harder. that is, front/rear panel and case02:17
wpwraks/are/can/02:17
wolfspraulI need to think about it more.02:17
wolfspraulwhere it fits in the Qi universe in the long run02:18
wpwrak(electronics) there are some electromechanical bits i'm not entirely happy about. e.g., the relays can be socketed. but the sockets are a sourcing problem.02:18
kristianpaulautomate for  sure02:18
wpwrakkristianpaul: it;s either labsw or throwing the power switch ~30'000 times ;-)02:19
kristianpauloh yes :)02:19
wolfspraulI had a power strip once with built-in http server and remote control etc.02:20
wolfspraulforgot the brand, some German thing02:20
kristianpauli was thinking too in remote control of a future M1 freedom box as rejon called02:21
kristianpaulah yes that above02:22
kristianpaulbut why not temp sensor wpwrak ? :)02:23
kristianpaulor i'm missing a i2c port somwhere ...02:23
wpwrakwhat would be the use of knowing the temperature inside the labsw ?02:23
kristianpaulno no, room temperature02:24
wolfspraulcould labsw control a temperature chamber?02:24
wpwrakwell, that should be similar. labsw shouldn't get very warm02:24
kristianpaulwhy? well check this freaking expensice thing http://www.nagios.org/products/environmental/esensors02:24
kristianpaulexpensice/expensive02:25
wpwrak(temp chamber) you would probably start with the chamber, then add a controller02:25
kristianpaulcost same as M1 i cant beleive it02:25
wpwrakpricy :) maybe you can make one with a button cell and wpan :)02:26
kristianpaulwpwrak: hey wait, what about temp control for reflow process or something in a obven perhaps? (dunno how is that really usefull in SMT)02:26
wolfspraulkristianpaul: you need to get over pricey being bad02:26
wolfspraula business creates value, so it can reinvest that money into more progress, more value02:27
wolfspraulthe more it can charge, the better, even for its customers :-)02:27
kristianpaulyeah, they feel better etc.. :)02:27
wolfspraulnah it's not that simple02:27
kristianpaulah :)02:28
wolfspraulif you don't trust the business on the other side, maybe you should not buy from them at all02:28
wolfsprauleverybody will be better off...02:28
kristianpaul(reinvest) yes thats very important02:33
Action: DocScrutinizer calls it a day02:45
wolfsprauln802:46
wolfspraulyou might call it 2 days as well :-)02:46
kyakxiangfu: (building on 32bit host as well) - thanks!07:19
kyakthough the only thing needed frmo 32bit host is the SDK and Toolchain tarballs :) i don't think it makes sense to provide two different images07:21
xiangfukyak, I am working on it. already build.07:21
xiangfukyak, yes. agree.07:21
xiangfukyak, checkout: build.openmobilefree.net07:21
kyakcool, it started building :)07:22
kyaknot without errors though07:23
xiangfukyak, it is all same with the fidelio.qi-hardware.com. only it is 32bit. 07:23
kyakyep, i see07:23
xiangfuI meet some errors. I will fix today.07:24
xiangfuI will buy two hard disks recently. my host don't have much space now. after build the openwrt full_system it maybe 90% used. :)07:25
kyakyou are blogging to much ;)07:28
xiangfubackup my laptop hard disk :)07:29
xiangfuand the problem is laptop 500GB, Server 160GB :)07:30
kyaki thought everybody is keeping their stuff on a Cloud already :)07:33
qi-bot[commit] Xiangfu Liu: config.full_system: add recently new packages (master) http://qi-hw.com/p/openwrt-packages/e75b38e08:12
qi-bot[commit] Xiangfu Liu: config.full_system: sort all sections (master) http://qi-hw.com/p/openwrt-packages/3d8675408:12
blogictuxbrain: my package just arrived :)09:31
blogichope i find the time to start playing with the dongles09:31
blogic... on the weekend09:31
wolfspraulblogic: thanks for buying stuff from Tuxbrain! :-)09:36
blogichaha09:36
wolfspraulyes!09:36
wolfspraul:-)09:36
blogicits nice to see that the usb part of the dongle is a meag32u209:37
blogicwpwrak: you made these dongles, right ?09:39
blogicas in the atusb 802.15.4 thing09:40
wpwrakah yes, guilty as charged ;-)09:48
blogicnice09:48
wpwrakglad you like them ! :)09:52
wpwrakblogic: you got a pair of atusbs ?10:00
wpwrakor atben+atusb ?10:01
blogica pair10:02
blogici have no nano note10:02
blogici want to pair 2 openwrt devices10:02
blogici have a spare tpl1043 and rspro10:02
blogicso i will probably use those 2 units10:03
wpwrakany specific application in mind ?10:05
blogicnot sure10:08
blogici was contacted by a few italians that wanna play with this stuff10:08
blogicand asked for me to port the drievrs and apps to openwrt 10:08
blogicso i said, sure send me hw pr0n10:09
wpwrakheh, nice :)10:09
jow_laptopkyak: you should maybe backport the latest opkg fix from trunk, the current version contains a double free which may lead to a seg fault with a corrupted status db on certain package removal operations10:27
kyakjow_laptop: thanks for notifying; that would be pretty easy, since we are on trunk already.. Just a matter of catching up with trunk svn10:29
qi-bot[commit] Xiangfu Liu: optimize for ben nanonote (master) http://qi-hw.com/p/openwrt-xburst/42cfc6110:37
qi-bot[commit] Xiangfu Liu: [xburst] Improve mounttime (master) http://qi-hw.com/p/openwrt-xburst/c38952710:37
qi-bot[commit] Xiangfu Liu: nanonote optimize (master) http://qi-hw.com/p/openwrt-xburst/03f997710:37
qi-bot[commit] kyak: add kernel patch for setfont2 (master) http://qi-hw.com/p/openwrt-xburst/b5a9c9a10:37
qi-bot[commit] Xiangfu Liu: optimize for ben nanonote (master) http://qi-hw.com/p/openwrt-xburst/b27472110:37
qi-bot[commit] kyak: config-2.6.37: enable battery, disable RNDIS (master) http://qi-hw.com/p/openwrt-xburst/948a9cc10:37
qi-bot[commit] kyak: patches-2.6.37: support for Ben NAND partitioning (master) http://qi-hw.com/p/openwrt-xburst/1fdde4710:37
qi-bot[commit] David Kühling: linux kernel: add CONFIG_PROC_PAGE_MONITOR=y to allow for clean user-space DMA (master) http://qi-hw.com/p/openwrt-xburst/9048c1910:37
qi-bot[commit] kyak: config-2.6.37: enable options needed for keymouse (master) http://qi-hw.com/p/openwrt-xburst/252345f10:37
qi-bot[commit] kyak: trunk: fix kernel keymap for VolUp/Down and Del (master) http://qi-hw.com/p/openwrt-xburst/75e7b5610:37
qi-bot[commit] kyak: trunk: build sound modules in kernel (master) http://qi-hw.com/p/openwrt-xburst/2d62caf10:37
qi-bot[commit] kyak: trunk: add ks7010 support patch (master) http://qi-hw.com/p/openwrt-xburst/cd6e23210:37
qi-bot[commit] Xiangfu Liu: base-files, move it to openwrt-package/nanonote-files (master) http://qi-hw.com/p/openwrt-xburst/0eb17f710:37
qi-bot[commit] kyak: Disable syslogd and klogd (master) http://qi-hw.com/p/openwrt-xburst/0e98cca10:37
qi-bot[commit] Xiangfu Liu: uboot-xburst: update to 2010.06 (master) http://qi-hw.com/p/openwrt-xburst/67e619e10:37
qi-bot[commit] kyak: busybox: backport reverse history search patch (master) http://qi-hw.com/p/openwrt-xburst/f71812010:37
qi-bot[commit] Xiangfu Liu: uboot-xburst: enable-silent-console.patch (master) http://qi-hw.com/p/openwrt-xburst/e956c5610:37
qi-bot[commit] Xiangfu Liu: uboot silent console (master) http://qi-hw.com/p/openwrt-xburst/e64b1c610:37
qi-bot[commit] Xiangfu Liu: ben nanonote: forward patches to linux-3.0 (master) http://qi-hw.com/p/openwrt-xburst/bfd462510:37
qi-bot[commit] kyak: kernel-3.0: add fbcon color fonts patch (master) http://qi-hw.com/p/openwrt-xburst/87387eb10:37
kyakcool, all patches rebased on top even without conflicts10:38
wolfspraulkyak: thanks for helping, as usual!10:39
kyakwolfspraul: it's my pleasure to do at least something useful :)11:03
kyaki haven't been very active recently after changing my job..11:03
kyakthis employer sure does know how to make me busy11:04
wpwrakkyak: you gave up a job that let you do whatever you want pretty much 100% of the time ??!11:12
wolfspraulkyak: he, understand :-) but I'm not so worried about slow, as long as we keep chugging away things will be fine11:28
wolfsprauland slowly but surely the software on the Ben actually becomes usable :-)11:28
Aylahi11:30
Aylahow can I copy a full directory with /usr/bin/install from a makefile?11:31
jow_laptop$(CP) ... ...  ?11:33
Aylais that a clean way to do it?11:35
Aylaeverything seems to use $(INSTALL) instead11:35
jow_laptopdepends11:35
jow_laptopmany openwrt packages use $(CP), especially in conjunction with PKG_INSTALL as the inner Makefile takes care of moving files to the right palces with correct permissions then11:36
AylaI'm using buildroot :)11:36
jow_laptopbuildroot follows the same principle11:36
Ayla$(CP) is not defined here11:36
jow_laptopok, then the answer is you can't11:37
jow_laptopafair you can use wildcards to install more than one file11:37
Aylathat would be surprising11:37
Aylaah, ok11:37
jow_laptope.g. install foo/* target/foo/11:37
Aylabut here the 'foo' directory has subdirs...11:38
jow_laptopwell, bad luck ;)11:39
jow_laptopafaik install is not meant for recursive operation11:40
jow_laptopeither name each dir explicitely or use some find | xargs construct or use cp11:40
Aylameh.11:40
AylaI'll just use cp then11:40
kyakwpwrak: yeah, basically i did what you said :)12:02
wpwrakgot tired of living in paradise ? :)12:02
kristianpaulbrbr 19°C and cold mornings again 13:45
kristianpaulbut is nice i want do something usefull today ;)13:46
kristianpaulcool 0,0041 error% now one more flip flop and i guess will be enought17:40
wpwrakit's kinda scary that you approach the correctness of register accesses with statistics ... :)17:43
kristianpauloops, how i should then?17:48
wpwrakit should just work 100% ;-)17:49
kristianpaulah sure !17:49
kristianpaulI'm on that, my concern still :)17:49
larscwpwrak: all modern memory is based on statistics17:51
larscprobability theory is right word i guess17:52
wpwraklarsc: flash is evil ;-)17:56
larscharddisk use similar algorithems17:57
larscand even ram does17:57
kristianpaulhe, well hardisk are always dying is it? :)17:58
wpwrakECCs are like safety belts. they're good to have, but if your actions depends on them being there, you're probably doing something very wrong :)18:00
larscuhm, all modern memory uses some kind of ecc18:03
larscotherwise you wouldn't be able to get such high densities18:04
wpwrakoh, they have ECC (well, most do). but few depend on it as extensively as flash. particularly NAND flash.18:08
wpwraknot sure about what disks do internally, though18:09
larscsure?18:11
wpwrakand in any case, there are parts of a system that are generally considered error-free. registers of local peripherals tend to be among them.18:12
larscyou can stop progress ;)18:13
wpwrakwell, look at the data sheets. NAND is only guaranteed after ECC. NOR without. DRAM is very often even used without any error correction.18:13
larsci was more thinking about hdds18:14
wpwrakah, dunno about HDDs. they do it the way i like it - present a nice and fairly reliable interface, keep the quantum physicals out of my view :)18:16
larschehe18:16
wpwraks/physical/physic/  # gah18:17
kristianpaulweee, i listened a noaa satellite, now be prepare to record, what was that sox command?18:21
kyakhm. i'm playing with Ben under Windows right now by chance. Since there was no luck using RNDIS, it was disabled at all. Now i'm trying with linux-cdc-acm.inf18:24
kyaknow i have the COM6 device named "Gadget Serial"18:24
kyakit seems running fine - i.e. Windows doesn't report any error codes18:24
kyakbut i try to connect to COM6 with putty, and get "Unable to open connection"18:25
kyaki wonder what is this driver even doing...18:25
kyakit seems like USB serial18:25
kyakbut what needs to be done from Ben's side to accomplish it?18:25
kyakkernel doc says that i need to have either RNDIS or CDC ECM. This is already accomplished since RNDIS configuration is disabled in linux kernel18:26
kyakand i can't have both on Windows18:27
kristianpaulah, rec :)18:27
kyaki would expect linux-cdc-acm.inf to create a network interface instead of some non-working serial//18:28
kristianpaulkyak: Great !! so just enable gadget serial when compiler kernel and done?18:29
kristianpauli guess the ethernet usb must be disabled?18:30
kristianpaulover usb*18:30
kyakthis IS ethernet over usb :)18:30
kristianpaulahh18:30
kristianpaul;)18:30
kyakat least when i plug Ben in on LInux18:30
kyakhold on, i'm reading http://www.kernel.org/doc/Documentation/usb/gadget_serial.txt18:30
kyakit seems that i should have a serial interface on Ben's side now18:31
C-Keenyou will get both, a host and a target serial device18:31
kristianpaulthasts why i was asking :)18:31
kristianpauloh18:31
kyakhow do i know the name on target?18:31
kyakit has a bunch of /dev/tty* devices18:31
kyakdmesg is not very usefull18:32
C-Keenkyak: some ttyU0 or something there? If not check the module options and the driver docs 18:32
kristianpaul/dev/ttyS2 or P2? perhaps18:32
kristianpaulah18:32
C-Keenmight as well be ttyUSBsomething18:33
C-Keenthis should be the name on the host. *or* if it has anything in common with g_printer, which I have worked on lately it will be g_serieal18:33
C-Keenserial18:33
C-Keenthis gadget stuff is still a bit a mess18:34
kyakthere are tty[0-9][0-9], ttyS[0-1], ttyp[0-1]18:34
kristianpaulsame here18:34
kyakso it wasn't created then18:34
C-Keenis there a udev running?18:34
kyaknope18:35
kyaki'm not sure if there is any18:35
kyakthere is udevtrigger and hotplug218:36
kyakthough hotplug is probably for network stuff18:36
kyakcould it be that i'm confusing ACM with ECM?18:37
kyakit should be ttyGS018:37
C-Keenget the major / minor from sysfs and mknod it yourself18:38
kyakit's not in sysfs either -\18:38
kyaki can try the major/minor from the doc though18:39
C-Keenthe driver is loaded and it is not in sysfs?18:39
kyakwell, i'm looking in /sys/class/tty/ - is that right?18:40
C-KeenI am not sure. I will check tomorrow at work, if your problem still persists ping me again in 13 hours or later ;)18:41
kyakok, if i have windows in 13 hours :_)18:42
kyakhttp://www.kernel.org/doc/Documentation/usb/gadget_multi.txt18:46
kyakalso a good reading18:46
kyakit says CDC ECM should work with CDC ACM inf file?18:47
kyakCONFIG_USB_G_MULTI is not set - probablt this explains..18:48
whitequarkscreen -rD19:43
whitequarkoops.19:43
qi-botThe build was successfull, see images here: http://fidelio.qi-hardware.com/~xiangfu/compile-log/openwrt-xburst.full_system-10162011-1650/21:31
--- Tue Oct 18 201100:00

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