#qi-hardware IRC log for Monday, 2010-06-21

neil_Hello, I just got my NanoNote this morning and I'm well happy.  About the red LED on the right-hand side.  The manual says it's a power LED but it seems to indicate that it's charging instead.  Is that what it's doing?00:04
neil_so when it goes off even though it's still plugged in, it means it's fully charged?00:06
neil_great. thanks00:06
neil_how can I tell roughly how much charge remains in the battery from the command line?00:09
urandom_neil_ cat /sys/class/power_supply/battery/voltage_now00:22
qi-commitsXiangfu Liu: [usbboot] copy the jz4760.h file from u-boot, fix the jz4760 nand addr, cmd, data, port define http://qi-hw.com/p/xburst-tools/eb2382002:47
qi-commitsXiangfu Liu: this is a workaround. we have plan to update the libusb-0.1 http://qi-hw.com/p/xburst-tools/43d775b02:47
tuxbrainwolfgang: so you finally have flashed your the HK stock, any tips so I can do the same at little scale here?03:34
kyakwolfspraul: can i please be added to downloads.qi-hardware.com/people/ ? i would like to have a place at qi to conveniently store and update my ports :)06:02
tuxbrainkyak: you better put also put the Makefiles once verify it runs in http://projects.qi-hardware.com/index.php/p/openwrt-packages/ so it can be directly integrated in next distro06:09
kyaktuxbrain: this is exactly what i needed, thanks!06:11
kyakcan i commit anonymously there?06:13
tuxbrainnop, you have to register in the project application here http://projects.qi-hardware.com/index.php/login/ and ask to one of the project admin to let you in06:14
kyakalready registered.. ok06:16
tuxbrainask mirko Vogt or wolfspraul to commit access :) I really want you in :P06:17
kyakok, i guess i already asked wolfspraul for some kind of access :)06:18
wolfspraulkyak: can you email me a password of your choice? (wolfgang@sharism.cc)06:28
wolfspraulthe downloads server has a DAV access where I can create an account for you then you can upload to downloads/people06:30
kyakwolfspraul: sent!06:43
wolfspraulkyak: added you to openwrt-xburst and openwrt-packages06:56
wolfspraulthanks for any tidbit you can contribute there, that's really cool!06:56
wolfspraulnext the DAV account06:57
kyakwolfspraul: will try to do my best, thanks :)06:58
wolfspraulkyak: OK I added a DAV account too. Can you try with a DAV client whether you have read/write access to downloads.qi-hardware.com ?07:02
wolfspraulI create a people/kyak folder already07:02
kyakone sec07:03
kyakgoing to find a DAV client for linux :) never used it before07:03
kyakok, gnome can do this..07:04
wolfspraulsometimes the UI and how you enter username/password are a bit spooky in some clients07:04
wolfspraulso they say 'read only' etc., some of them bring the user/password dialog at failed write attempts...07:05
wolfspraulkyak: if you upload binaries to /people/kyak, please do not upload any binaries that cause patent problems, like MP3, MPEG4, H.26407:05
wolfsprauldoes it work? (can you try?)07:08
kyakwolfspraul: yes, working very well, thank you!07:18
wolfspraulgreat, thanks! adding commit rights and accounts is the best thing for the project07:25
kyakMy build of jfbterm.. It requires some configuration to display UTF-8 properly, and I havent figured out yet how to do it :)08:57
kyakand it's MUCH faster displaying, than the default terminal; and unifont looks more slim08:59
kyakby the way, there is also an fbterm build in people/kyak/tmp/fbterm.. but i suppose there is more chance for displaying of UTF-8 in console with jfbterm than fbterm09:01
wolfspraulwhat is the difference between jfbterm and fbterm?09:06
wolfspraulI only heard some Japanese like fbterm.09:06
wolfspraulin general I would like an easier way to increase/decrease console font sizes09:07
kyakjfbterm stands for Japanese fbterm :)09:11
kyaki hope it has a better support for utf-809:11
kyakcurrently i don't have any luck with both of them09:11
wolfspraulah there you go09:11
wolfspraulmaybe it's just the bfbterm (better fbterm)09:12
wolfspraulit seems to be quite commonly used in Japan, or so I heard09:12
kyaki'ld like to find a jwfbterm (Just a Working fbterm )09:12
kyaklol.. this is amazing what just happened :) i've launched jfbterm from a ssh session, and it started in Ben09:16
kyaknot i'm typing in ssh screen, and everything goes to Ben's screen :)09:16
kyaki wonder if this is a bug or a feature09:16
kyakit feels like connecting external keyboard to Ben09:17
kyaki think jfbterm has attached my ssh virtual terminal to a real tty09:17
zearkyak, i don't know about the nanonote, but if you launch programs on dingux via telnet, they render on dingoo's screen09:17
kyakzear: that's interesting.. i thought for X programs this should depend on DISPLAY variable, and for SDL they just don't launch09:19
kyakzear: yes, the same for Ben.. at least gmu is rendered on Ben's screen09:21
zearvia telnet, or via ssh?09:22
kyakvia ssh09:26
kyakshould be the same09:26
Action: kyak is building openwrt with utf-8 and locales support11:29
qi-commitsXiangfu Liu: [new package] gawk, o3read, surfraw by Jirka http://qi-hw.com/p/openwrt-packages/2dad1d311:39
rafakyak: is jfbterm better than fbterm?12:13
kyakrafa: it has support for bitmap fonts and supposedly supports i18n better. At least, when it comes to utf-8 support in console, there are many references to jfbterm12:19
kyaktheoretically, fbterm is also capable of text encoding, but anyway I didn't succeed with any of them yet12:20
rafakyak: I ask because it looks like jfbterm is old (last version is 2003 no?).. and fbterm is current, but I do not know which is better, that is why I ask you :)12:33
kyaklast version of jfbterm is 200512:35
kyakbut it doesn't matter as long as both don't do what i want :)12:35
kyaki'm just trying to play with both12:36
lekernelwhere can I buy a RS232/3.3V shifter in Europe?14:20
lekernelFrance preferably14:20
prpplaguelekernel: virtually any electronics outlet is going to carry uart level shifters14:23
lekernelprpplague, you overestimate the french electronics shops14:24
lekernelin my city they don't even have standard 2.54mm connector pins14:25
lekernelhe10 etc.14:25
lekernelplus everything is expensive as hell14:25
lekernellike 10x the digikey prices14:25
prpplaguelekernel: yea but you can afford to buy it since you have such good health care14:27
lekernelok, let's say farnell or radiospares prices then - both companies have French divisions14:29
TextmodeMorning all14:32
dptechhello where can i found the best tuto for flash to debian ?14:40
nebajothdptech: I followed http://www.tuxbrain.com/en/content/%E6%9C%AC-ben-nanonote-editingcompilingflashing-arduino-sketch14:41
nebajoththe first part of it anyway14:41
dptechnebajoth: thanks, I must to see before to go... so I want to be sure to make very good!14:47
dptechI see first part for debian, perfect!14:48
xdpirateWhat do you call an LCD screen like the one the NN has? (In regards to every even row of pixels lines up, and every odd row of pixels lines up, but the odds and evens differ by half a pixel)15:06
prpplaguexdpirate: the lcd in the NN is an passive matrix display with an 8-bit data interface15:14
xdpirateprpplague, thanks15:19
prpplaguexdpirate: it is a pretty decent display and fairly easy to interface to15:21
xdpirateit's hard to read small fonts when the pixel matrix is lined up like that tbh15:21
xdpiratebut i'm guessing the displays are cheaper to manufacture so15:22
prpplagueyea, they are mainly manufacturer for dedicated devices that use small graphics15:22
prpplaguei.e. game handhelds and control devices like a programmable thermostat15:23
prpplaguei have the display from the NN working with some OMAP3 and OMAP4 devices15:24
calamarzsdschulze: did you succeeded in rebuilding debian kernel for the nn with the openwrt toolchain?15:52
calamarzsdschulze: i was trying to follow your notes but got no luck booting it :(15:54
neil_Hi.  I've read http://en.qi-hardware.com/wiki/How_to_reflash over and over but I'm still not confident of how to re-flash -just- the rootfs.  Do I still need to use nerase if I don't want to erase the whole NAND?  If I do, would nerase 2048 4096 0 0 be the right thing to do ( followed by nprog for the rootfs)?19:17
larsc`nerase 16 4080 0 0` earases the rootfs19:27
emebwhy not just use the flash script and let it do u-boot, kernel and rootfs?19:27
neil_larsc: do you mean "4096" rather than "4080"?19:29
neil_or do you really mean 408019:29
neil_emeb: I've got the reflash-ben.sh, thanks.  I'm trying to understand it.  The page on the wiki does say to avoid re-flashing u-boot unless necessary19:30
larscneil_: the second parameter is the number of blocks to erase19:31
larscsince there are 4096 and you start at 16 you want to erase 4080 blocks19:31
neil_larsc: the wiki page says it's the block length.  ah.  does that mean "the number of blocks" and not "the block size", which is how I interpreted it?19:32
emebneil_: larsc: careful! using 4080 will erase everything from 16 on to the end, not just the rootfs.19:36
emebif you look at the reflash_ben.sh script you can see the parameter they use is 512, not 4080. The rootfs of the OpenWRT build doesn't take the whole rest of NAND19:37
emebin the OpenWRT build they reserve the last portion of NAND for the data partition of ~ 1.7GB.19:38
neil_emeb: I read exactly what you say somewhere on the wiki but "df" shows 1.8GB free on my rootfs.  is this the ubifs compression?19:40
emebneil_: no. when I do a df on mine I only see about 200M in the rootfs partion.19:41
larscit depends in your image version19:41
emebWhere did you get this image?19:41
neil_mine's 2010-10-??  It came with the device19:42
emebOK - that's the original. Pretty old u-boot and kernel too. You may want to update the u-boot and kernel because they're much improved since then.19:42
neil_emeb: OK, I'll update u-boot as well as you say.  sounds like having a data partition so my music isn't trashed when I update the rootfs is a good idea19:43
emebneil_: yep - that was what I thought too. There are some good instructions on the wiki for setting up the data partition after you reflash.19:46
emebI just used the script and it worked fine.19:46
neil_why does reflash-ben.sh only issue an nerase command for the rootfs and not for the u-boot and kernel partitions?19:47
emebthe new u-boot allows booting into USB mode without fiddling around with the shorting pads under the battery.19:47
emebneil_: no idea about the nerase. I was wondering that too19:47
neil_emeb: but it worked for you anyway?19:47
emebneil_: yes it worked fine both times I ran it.19:48
larscneil_: for the rootfs leftovers from the previous on need to be cleaned up. for kernel and uboot it doesn't matter19:48
neil_thanks for all the help.  I'm learning a lot and it's getting clearer19:50
emebneil_: looking at the logfile from the reflash script it appears that the nprog command does an erase automatically.19:50
larscneil_: if you think there is misleading or missing information on the wiki, please update it.19:51
emebit may be redundant to do it in the script.19:51
neil_larsc: OK, I will do.  thanks20:06
nebajothdoes the openwrt firmware use dumpkeys/loadkeys?21:34
nebajothor is that just a debian thing?21:34
nebajothanyone using debian: ping21:35
xiangfuI tried loadkeys in openwrt before. but not success. the BUSYBOX's loadkeys use different file format.21:43
nebajothwhat about dumpkeys?21:54
qwebirc64441Hello.. just flashed the Nanonote with Debian... Now it won't turn on..21:55
nebajothwhat happens when you take out the battery and plug in the usb cable?21:56
nebajothdoes the red light come on for a brief second, then go off?21:56
qwebirc64441Hmm.. let me check.21:56
qwebirc64441Yes, it does.21:57
nebajothyour bootloader is corrupt21:57
nebajothyou need to short the usb boot pins beneath the battery21:57
nebajothforce it into usb boot mode21:57
nebajothand then reflash21:57
nebajothI use tin foil21:58
nebajotha small piece21:58
qwebirc64441Just the boot-loader or everything?21:58
nebajothI'd wipe it clean21:58
nebajothand then restart21:58
nebajothdo the erase function21:58
nebajothwhat tutorial are you following?21:58
nebajoththat didn't work for me21:59
nebajoththe rootfs in particular21:59
nebajothI used tuxbrain's tutorial21:59
nebajothwith a different rootfs21:59
qwebirc64441Ok. Thanks.21:59
nebajoththe first part is what you need22:00
nebajothignore the stuff about the arduino22:00
nebajoth(unless you have one)22:00
nebajothspecifically this line: wget http://mister-muffin.de/nanonote/debian-sid.ubi22:00
nebajothI used the pyneo commands22:00
nebajothand that .ubi22:00
nebajothand it works a treat22:00
nebajothxiangfu: I occasionallyy get a message about scrubbing PEBs and moving them to other PEBs22:01
nebajothis this a matter for concern?22:01
wolfspraulqwebirc64441: your Nano is already unbricked? (just reading up)22:03
wolfspraulif your u-boot bootloader is indeed corrupted, you need to short the USB boot pin but there should be a little rubbery button included in your box22:04
nebajothoh yeys22:04
nebajothI lost mine22:04
wolfsprauluse that, much easier than aluminum foil22:04
nebajothalso hello wolfspraul22:04
wolfspraulnebajoth: email me your shipping address and we'll mail you a few more in regular mail22:04
wolfspraulI think they cost 0.5 US cents or so a piece, forgot :-)22:05
nebajothhaha, well shipping might be a little more22:05
qwebirc64441wolfspraul: Not yet. I'm in the process of putting it into USB-boot mode.22:05
nebajothany word on backlit keyboards for the Ya? :D22:05
wolfspraulsure but we'll just throw them in a regular letter22:05
nebajothoh snap22:05
nebajothI fixed my keymap problem22:06
nebajothfinally, a working + key22:06
nebajothnow I can use remind again22:06
wolfspraulnebajoth: please keep your expectations for those things (backlit keyboad) very low22:06
nebajothhaha, I do22:06
wolfspraulwe have to 100% focus on what this project is about - copyleft hardware22:06
wolfspraulwe sold 800 Ben, which is great22:06
wolfspraulbut of course totally not sustainable22:06
nebajothnot unless you increase the price a great deal :O22:06
wolfspraulso we need to be very careful, 100% focus on what is unique about our product22:07
wolfspraulincrease price = death22:07
wolfspraulthis is a hardware business22:07
wolfspraulI want to decrease prices.22:07
nebajothI dunno, key visibility in low light situations seems like a pretty universal requirement22:07
nebajothfor a "portable" device22:07
wolfspraulso anyway, when we think about Ya, we have some amount X of cash22:07
nebajothwhatever the use case scenario22:07
wolfsprauland it will be very little22:07
wolfsprauland we will focus it only on the areas that matter to actually be able to ship the product22:08
wolfspraulbut now, on the good side, I agree about your observations about the keyboard22:08
neil_what impact on the price would USB host have?22:08
nebajothalmost none, I'd think22:08
wolfspraulthat's the advantage of all of us using this thing, and the reason why we launched the project the way we do22:08
wolfspraulget a real product out first22:08
nebajoththe jz4740 has usb host capability22:08
wolfspraulcollect real life feedback, grow a community22:08
nebajothif I read the spec sheets correctly22:08
nebajothit'd have to be broken out into the circuitboard for a connector22:09
wolfspraulonce you do a full little computer with 2-3 full-time people, some volunteers, and some donated cash, every tiny thing becomes big :-)22:09
nebajothyeah, I hear you wolfspraul22:09
nebajothany chance you can send me some blank keys?22:09
nebajothI'll experiment with glow in the dark paints :D22:09
wolfspraulblank keys?22:09
nebajothI want to replace mine22:10
wolfspraulI think the print is under the plastic22:10
wolfspraulI'm not sure how easy it is to take the keyboard apart22:10
wolfspraulit may all break22:10
nebajothnot very22:10
wolfspraulif I have some old prototype/sample keyboard I can throw it in, for sure22:10
nebajothwhat do you do for a day job, wolfspraul?22:10
wolfspraulthis is my 'day job'22:10
nebajothwhere does the money come from to pay you?22:11
wolfspraulit's 10.10 AM in China22:11
wolfspraulfrom my savings22:11
nebajothyou're betting big on this22:11
nebajothdo you live in China?22:11
wolfspraulI don't think like that.22:11
wolfspraulyes, China22:11
nebajothyou are German by birth?22:11
wolfspraulkeep costs of living low, plus manufacturing is here22:11
wolfspraulyes. first Germany, then >10yr US make some money and savings, then 3+ years China spend the savings22:12
nebajothwhat do you work on mostly?22:12
wolfspraulhe he22:12
wolfspraulyeah, seriously22:12
wolfspraulthat's the short story22:12
wolfspraulI work on manufacturing issues22:12
nebajothpretty interesting story actually22:12
nebajothcould make a good magazine article22:12
nebajothwhen its time to get real publicity out about this unit22:13
wolfspraulI'm a software engineer originally.22:13
wolfspraulfun stuff, file systems22:13
neil_how do test a file system ( other than by using it)?22:13
neil_s/do/do you/22:14
wolfspraulsure run automated scripts/programs22:14
wolfspraulfile systems are hardcore, especially on multi-core/SMT machines, with swapping going on, etc.22:14
wolfspraulit was a fun time back then22:14
wolfspraulbut I have to say manufacturing is also really really interesting. totally overlooked.22:14
wolfspraulnebajoth: so far I think our NanoNote is off to a good start, isn't it?22:15
wolfspraulafter the first year...22:15
qwebirc64441Damn. It's a real bitch trying to get it into USB-boot mode.22:15
wolfspraulnow we throw all that we've learnt into Ya, very much looking forward to it22:15
wolfspraulqwebirc64441: yeah, one thing we need to improve in Ya :-)22:15
wolfspraulqwebirc64441: do you use the carbonized rubber button?22:15
neil_I'd say so.  Incredible that it can be made for $9922:15
nebajothwolfspraul: NanoNote rules, dude.22:16
wolfspraulunplug the USB cable22:16
nebajothI effing love this thing22:16
nebajothit really grows on you22:16
nebajothand I think that's mostly about hardware that just works22:16
wolfspraulpush the carbonized rubber button against the USB boot pins22:16
neil_I'm re-flashing for the 1st time right now.  It says I have a Jz4740.  I thought it was Jz4720?22:16
wolfspraulplug the USB cable back in22:16
wolfspraulah, sorry, forgot something :-)22:16
wolfspraulon your host, run "watch lsusb"22:16
wolfspraulwhen you see the ID "0x601a:4740" show up, that's your Nano in USB boot mode22:17
wolfspraulso it's easy to watch it there22:17
wolfspraulso if you plug the USB cable in, you can also press the power-on button (hold for 2-3 seconds to be sure)22:17
nebajothI prefer 'watch -n 0.2 lsusb'22:17
nebajothbut I'm an impatient guy22:18
wolfspraulyou only need to short the USB boot pins at the time when the CPU gets power22:18
wolfspraulyou don't need to hold it22:18
wolfspraul(after the CPU is powered)22:18
nebajothis holding it a problem?22:18
nebajothI'm breaking my usb-mode pads out to the side of the unit22:18
wolfspraulbasically when the CPU gets powered, it checks those pins. If short, branch to USB boot mode. if not short, boot from NAND.22:18
nebajothand I was going to put in a switch22:18
wolfspraulholding shouldn't be a problem22:18
nebajothok good22:18
wolfspraulneil_: 4720/4725/4740 is all the same die, just different packaging (=way the CPU is wrapped in plastic and fixed to the PCB)22:19
nebajothhow much memory is going into the Ya, you think?22:19
wolfspraulA LOT! :-)22:19
wolfspraulhe he22:19
wolfsprauldon't know22:19
wolfspraulwe will drive this together, the specs and all22:19
wolfspraulit will be fun22:19
wolfspraulwe have so many good people on board now22:19
wolfsprauljust need to cleanup a bit22:20
wolfspraulmerge the discussion & developer mailing lists into one22:20
neil_are you concerned about overlap with the Pandora?22:20
wolfspraulnot at all22:20
wolfspraulPandora is not copyleft hardware22:20
wolfspraulof course they run Linux and all so that's cool22:20
nebajothI want to get a Lemote22:20
wolfspraulneil_: we are very serious about this copyleft hardware thing22:20
nebajothchips with chinese symbols on them are 133722:20
wolfspraulwe need to work more on that, like GPL licensed production testing software (for Ya definitely)22:21
wolfspraulalso case design22:21
wolfspraulneed to cleanup some trademark issues (like 'microSD')22:21
nebajothcase design isn't open?22:21
wolfspraulwe worked hard on the patent problem, and still ongoing22:21
wolfspraulnebajoth: first question is "what is open case design?"22:21
wolfspraulright now we have design files open, but not 3D files22:21
nebajothschematics with a CC license?22:21
wolfspraulall that stuff is there22:22
wolfspraulI am talking about the plastic parts.22:22
wolfspraulof course the design files are all CC licensed.22:22
nebajothwe need to be able to build one of these with a RepRap22:22
neil_one day it would be nice to use a RepRap or one of its descendants to fab a new case at home22:22
wolfspraulthe problem with mechanical is that all professional software isproprietary22:22
wolfspraulanyway we work on that22:22
wolfspraulone by one22:22
wolfspraulwe start with the most valuable parts22:22
wolfspraulhey guys, Milkymist One boots Linux, did you know?22:23
nebajothno I had no idea22:23
wolfspraulthat's our future free GPL licensed CPU!!!22:23
nebajothis it fast? :P22:23
wolfspraulOn the XYZ NanoNote in 2020 :-)22:23
wolfspraulneil_: let me show you something I really like (still wrt your Pandora question)22:24
wolfspraulthis is awesome!22:24
wolfsprauldocumentation about the SMT process for Milkymist One22:24
neil_I'm getting "Error - can't read bulk data from Ingenic device:1344" when running sudo ./reflash_ben.sh.  I have the latest reflash.sh and xburst-tools22:24
wolfspraulI think this is the best and most thoroughly documented SMT process ever.22:24
nebajothneil_: is the unit actually in usb boot mode?22:25
neil_think so.  it started the flash process22:25
nebajothuh oh22:25
wolfsprauldid you see whether it flashed u-boot and kernel already?22:25
wolfspraulwhen did it stop?22:25
nebajothwolfspraul: that milkymist link is awesome22:26
neil_I have 601a:4740 in lsusb.  can I paste multi-line here or is that rude?22:26
wolfspraulok for me, but probably not needed22:26
wolfspraulok if you see that ID, it means you are in USB boot mode. good.22:26
neil_ Checking 454656 bytes... no check! End at Page: 60542156922:27
neil_ Skip a old bad block !===22:27
neil_fatal error occured - ABORTED22:27
neil_error while flashing bootloader:22:27
neil_that's from the log.txt22:27
wolfspraulboot loader22:27
wolfspraullet's try manually :-)22:27
wolfspraulI don't like the reflash script anyway...22:27
wolfspraulturn off the Nano22:27
wolfspraulunplug USB cable22:27
wolfspraulshort the USB boot pins with the carbonized rubber button22:28
neil_ok.  doing that now22:28
nebajothhow do I get a milkymist?22:28
wolfspraulplug the USB cable in, press POWER ON, see the 0x601a:4740 showing up22:28
neil_do I have to wait a minute like it says on the wiki?22:28
neil_ok, got vend/prod22:29
wolfspraulif you see the 601a:4740, all is good22:29
wolfspraulnow run 'usbboot' (as root)22:29
wolfsprauljust 'usbboot'22:29
wolfspraulit will go to a command prompt22:29
wolfspraulfirst command: "boot"22:29
neil_ok.  doing22:29
wolfspraulsecond command: "nprog 0 u-boot.bin 0 0 -n" (sorry u-boot.bin needs to be the real long filename of u-boot)22:29
wolfspraulhow about the nprog command?22:30
neil_got prompt back.  last thing it said was "Skip a old bad block"22:31
nebajothneil_: reinstall openssh-server, and install console-setup22:31
nebajoththe first will force it to generate a key for the ssh server22:31
nebajothand the second will fix a keymap issue22:31
wolfspraulneil_: let's run an erase command over your entire NAND first22:31
nebajothat least, an issue I was having22:31
nebajothqwebirc64441: ^^22:31
nebajothqwebirc64441: any luck?22:32
neil_nebajoth: does "boot" install an SSH server/client then?22:32
wolfspraulneil_: is that OK with you?22:32
wolfspraulso try "nerase 0 4096 0 0"22:32
nebajothneil_: eh?22:32
wolfspraulthat will nuke the entire 2 GB22:32
nebajothneil_: openssh is installed by default, if that's what you're asking22:32
nebajothbut it doesn't work22:32
wolfspraulafter that the nprog command again22:32
nebajothyou will almost certainly need to designate a swap file22:33
nebajothor use the microsd as a swap partition22:33
wolfspraulneil_: no no, "boot" will only load a tiny snipped to the Nano so usbboot can community with the Nano22:33
nebajoth(I do the latter, but xiangfu has done the former)22:33
wolfspraulouch it's getting confusing now, we talk about entirely different things22:33
nebajothI'm going to bed soon22:33
nebajothI was just pouring out information :P22:33
nebajothbut I've just confused things22:33
neil_wolfspraul: ok.  I think I got crossed wires just then.  have done nerase now.  last thing it said was "Force erase, no bad block infomation!"22:33
wolfspraulnow the nprog command again22:34
wolfspraul"nprog 0 u-boot.bin 0 0 -n"22:34
qwebirc64441nebajoth: Not yet. :/22:34
qwebirc64441nebajoth: Cant' get it into USB-boot mode.22:34
xiangfuit's u-boot-nand.bin22:34
neil_same:  Finish! (len 454656 start_page 0 page_num 111)22:34
neil_Error - can't read bulk data from Ingenic device:6422:34
neil_ Checking 454656 bytes... no check! End at Page: 840297722:34
neil_ Skip a old bad block !22:34
nebajothqwebirc64441: are you using the plastic doohickey or tinfoil?22:34
qwebirc64441The plastic thing.22:35
nebajothqwebirc64441, neil_: remember to sign up for nanohacks.org if you haven't already22:35
neil_wolfspraul: USB cable is the one that came with it.  host can only do USB 1.1.  is that relevant?22:35
nebajoththere's only a few of us posting stuff, but its the official home for debian discussion22:35
wolfspraulneil_: thinking22:36
wolfspraulneil_: can you try on another Linux host machine?22:36
wolfspraulor if your computer has multiple USB plugs, try the other ones22:37
neil_wolfspraul: can do.  will take a few mins to install xburst and so on22:37
wolfsprauldo you have multiple USB connectors on this computer?22:37
neil_yes, although they've always behaved the same, unlike some other machines I've seen22:37
wolfspraulseveral people have seen problems behind USB hubs, and sometimes there are hub chips on the mainboard22:37
wolfspraulcan you still turn on your Nano?22:38
wolfspraulor screen just stays black?22:38
qwebirc64441Well.. trial and error.. trial and error.. i'll keep trying. Thanks for all of your help.22:39
neil_inserting in either port shows:Bus 001 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub22:39
neil_Bus 002 Device 034: ID 601a:474022:39
neil_Bus 002 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub22:39
neil_Bus 003 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub22:39
neil_wolfspraul: can still turn it on.  it says unusual stuff22:40
wolfspraulmaybe the bootloader is just fune22:40
wolfspraullet's try to flash the kernel22:40
wolfspraulrun 'usbboot'22:40
wolfspraulthen at the command prompt 'boot'22:41
wolfspraulthen 'nprog 1024 uImage.bin 0 0 -n'22:41
neil_wolfspraul: aha.  it says on LCD: Wrong Image Format for bootm command22:41
wolfspraulreplace uImage.bin with the name of your kernel image22:41
wolfspraulwhat are you flashing actually? openwrt 2010-06-15?22:41
neil_wolfspraul: whatever the reflash-ben script downloaded22:41
wolfspraulopenwrt 2010-06-1522:42
wolfspraullet's try nprog for the kernel, see whether more/other errors show up22:42
neil_wolfspraul: it actually has a timestamp of 2010-06-16 but that might just be 'cos I'm in NZ22:42
wolfspraulfrom your log, my guess is that after the bootloader has been flashed successfully, usbboot looses connection with the Nano22:43
wolfspraul(can't read bulk data)22:43
neil_wolfspraul: ok, will try the kernel now then22:43
wolfspraulall errors after that are just because usbboot lost connection22:43
wolfspraulyeah let's try that22:43
wolfspraulmaybe you get other errors22:44
wolfspraulalso kernel is bigger, so that's good22:44
wolfspraulneil_: I have another idea - my favorite :-) microSD booting22:46
wolfsprauldo you have a microSD card?22:46
neil_wolfspraul: it says scarier things like "Comparing 524288 bytes - FAIL at off 3008, wrote 0x37, read 0xb5"22:46
wolfspraulpreferably a smaller one, 1 GB or so there is a bug in u-boot with larger microSD cards right now22:47
wolfspraulneil_: yeah OK.22:47
wolfspraulI doubt this is your Nano, it's USB related.22:47
wolfsprauljust a guess22:47
neil_wolfspraul: yes, 2GB uSD just arrived this morning.  it says it's Kingston but I suspect it's not after reading Bunnie's report22:47
wolfsprauldo you have a microSD card?22:47
wolfspraulBunnie's report is awesome!22:47
neil_wolfspraul: too true22:48
wolfspraulok let's try this22:48
neil_wolfspraul: I could try on another laptop first if that's better22:48
wolfspraulpartition the microSD card, one big ext2 partition22:48
wolfspraulunpack the openwrt rootfs into the root of that partition22:48
wolfspraul(you have to download the .tar.gz file)22:48
wolfspraulput the kernel into /boot/uImage on the microSD filesystem22:49
wolfspraulthen insert the microSD into your Nano, boot the Nano and hold the 'm' key22:49
wolfspraulwith some luck it should boot openwrt from the microSD card22:49
wolfspraulyes, another laptop is also good, we collect more data22:50
wolfspraulsorry for the inconvenience, your type of problem is new to me, never seen this kind of thing22:50
neil_wolfspraul: sorry for delay.   had to open the impossible packaging around the uSD card22:51
wolfspraulthis is what you need to unpack into the ext2 partition on the microSD card22:51
wolfsprauland then put this file into /boot/uImage (on the microSD card)22:52
wolfspraulneil_: it's a long shot, but we should try22:52
wolfspraulmy assumption is that your u-boot reflashed correctly (even though you got errors), and that your microSD card will not exhibit the bug we have in u-boot right now that makes > 50% of microSD cards fail booting22:53
wolfspraullet's see :-)22:53
neil_wolfspraul: doing all that now22:54
urandom_seems like python-gtk ipkg doesnt work on openwrt ben, also ruby-ncurses might not work22:58
urandom_at least pygame seems to work22:59
neil_urandom_: pygame is ported!?  cool23:00
nebajothoh snap23:00
nebajothapt-get install pygame23:00
neil_wolfspraul: do I just rename this blahblah-uImage.bin file to uImage on the SD?23:01
wolfspraulmake sure to put it in /boot/uImage23:01
neil_wolfspraul: ok.  I was going to put it in /, so good catch23:02
neil_wolfspraul: ok, all that is done23:02
neil_nebajoth: do you mean pygame for Debian on NN or pygame on OpenWRT?23:03
wolfspraulput it into the Nano, press power_on and hold the 'm' key23:03
urandom_when someone made a package for python-gtk and put it in the repo wouldnt he test it?23:04
neil_wolfspraul: sorry I only just thought of this: the battery has been out the whole time, powered only by USB23:04
neil_wolfspraul: could that be a problem?  maybe my USB port power is drooping or something. anyway, should I leave it as-is or replace the battery before booting the SD?23:05
wolfspraulshouldn't matter23:05
urandom_neil_ we have pygame for openwrt and debian, apt-get is for debian23:05
urandom_and jlime i think23:06
wolfspraulunplug the cable once (without battery) to make sure you start from a clean zero state23:06
neil_urandom_: that is very cool.  no need for me to cross-compile to develop most stuff then23:06
wolfspraulI have seen USB hosts that indeed had power problems (good idea!)23:06
wolfspraulbut let's just try this microSD booting now23:06
urandom_i personaly dont like pygame, ulove will be much better23:07
wolfspraulurandom_: remember we are just a few people and essentially doing a whole distro for an embedded device here23:07
wolfspraultest plan is sorely missing, I agree23:07
wolfspraulbut I have a feeling I will have to focus on reflash & software update issues for quite a while, as you see there are still too many people who get stuck on that.23:08
wolfspraulwith all the people that are helping we have made tremendous progress, but I think there are 2000+ packages in openwrt, and of course most of them are untested23:08
wolfsprauljlime has 18,000 packages or so, 2 or 3 volunteers :-)23:08
wolfspraulneil_: how's it going?23:09
neil_wolfspraul: wow. booted.  has desktop.  not seen that before.  quite flickery23:09
wolfspraulyeah, flickering23:09
wolfspraulsoftware bug, no worries23:09
neil_wolfspraul: it worked though!23:09
wolfspraulnot a hardware issue, definitely not23:09
urandom_i am not sure if we even should keep openwrt as a main distro, i mean i like it but maybe jlime is better23:09
neil_wolfspraul: I'll try dClock, "x" later on23:10
wolfspraulone sec23:10
wolfspraulwe haven't reflashed your NAND properly yet23:10
wolfspraullet's make sure all is good there23:10
neil_wolfspraul: yes, sorry I didn't think we were done.  mislead you there23:10
nebajoththe strength of going with openwrt as the main supported distro23:10
nebajothis that it lends itself to targetting specific niches23:11
wolfspraulI need help23:11
nebajothit enables people to put together custom mini-distros23:11
nebajothwithout all the overheard of OE23:11
wolfspraulneil_ has a working u-boot in NAND, but kernel and rootfs not yet23:11
nebajothwhich is what is used to build jlime, I believe23:11
wolfspraulI know how to write the rootfs from inside Linux, but never done the kernel23:11
nebajothyou mean the commandline commands to flash it?23:12
wolfspraulneil_: for the rootfs, what you need to do is this: download http://downloads.qi-hardware.com/software/images/Ben_NanoNote_2GB_NAND/2010-06-15/openwrt-xburst-qi_lb60-root.ubi23:12
wolfspraul(maybe you can find it on your host, reflash_ben.sh must have downloaded it already)23:12
wolfspraulcopy that file onto your microSD card23:12
wolfspraul(first turn off your Nano)23:12
wolfspraulthen boot Nano again, with that -root.ubi file on the microSD card23:12
neil_wolfspraul: ok.  doing23:13
wolfspraulthen after you booted, you go into command line and run "ubiformat /dev/mtd2 -f -root.ubi"23:13
wolfsprauland "ubiformat /dev/mtd3" (that's the empty 1.5 GB data partition to speed up boot times for now)23:13
wolfspraulso basically we have u-boot in mtd0 (already working), rootfs in mtd2 (see whether ubiformat runs through), empty data partition in mtd3 (also let's see what ubiformat says)23:14
wolfspraulthen the only thing missing will be mtd1, we have to get your /boot/uImage copied/flashed into /dev/mtd123:14
wolfspraulmaybe just cat or cp ?23:14
wolfspraulnebajoth: yes, that's what I meant23:14
wolfspraulhow can neil_ get the /boot/uImage kernel into flashed into /dev/mtd1 ?23:15
neil_wolfspraul: not "dd"?23:15
nebajoththose are all in the pyneo tutorial23:15
nebajothusbboot -c "boot"23:15
nebajothusbboot -c "nprog 0 openwrt-xburst-u-boot.bin 0 0 -n"23:15
nebajothusbboot -c "nprog 1024 openwrt-xburst-qi_lb60-uImage.bin 0 0 -n"23:15
nebajothusbboot -c "nprog 2048 debian-lenny.ubi 0 0 -n"23:15
nebajoththe openwrt ubi23:15
wolfspraulno no23:15
wolfspraulthis is different here23:15
nebajothoh, from inside another booted distro?23:16
wolfspraulfor some reason neil_ did not have luck flashing via usbboot23:16
neil_nebajoth: we've booted from SD23:16
nebajothgood lord23:16
wolfspraulbut luckily at least the u-boot made it via usbboot23:16
wolfspraulso then we let that u-boot load the kernel & rootfs from microSD23:16
wolfspraulnow we need to flash kernel & rootfs into NAND, from inside a running Linux system23:16
wolfspraulit's much more stable and trustworthy anyway23:17
wolfspraulthis is how I always reflash the rootfs, never via usbboot23:17
wolfspraulI normally do u-boot & kernel via usbboot, then rootfs via microSD23:17
wolfspraulalso allows better parallelization (I can use 6 or 8 microSD cards)23:17
wolfspraulbut I've never done the kernel from inside, that's why I need help23:18
wolfspraulmaybe just dd/cp/cat, maybe some mtd-utils utility?23:18
wolfspraulneil_: let's try the rootfs with ubiformat first, see whether ubiformat says anything unusual23:19
neil_wolfspraul: I guess we can get rootfs going first though23:19
wolfspraulI still want to make sure all is good with your NanoNote and the NAND chip23:19
neil_wolfspraul: ninja23:19
wolfspraulyes totally23:19
wolfspraulthe ubiformat will do a lot of checks and treat the NAND chip properly23:19
wolfspraulmuch much better than the hacked-together stuff running inside usbboot23:19
neil_wolfspraul: ok, booted. how do I get a terminal?23:20
wolfspraul(ctrl-alt-f5 will get you back to graphical)23:20
neil_wolfspraul: smart.  just like desktop linux23:21
urandom_this flashing from microSD stuff should be addet to the wiki23:21
neil_urandom_: I'll keep the IRC log and have a go at that if you don't mind23:21
urandom_and you dont need to keep it we have archives23:22
wolfspraulhmm, mtd-utils has a 'flashcp' command, maybe use that to copy the uImage to /dev/mtd1 ?23:24
nebajothcat /proc/mtd23:25
nebajoththere's the three partitions23:25
wolfspraulshould be 4 now23:26
nebajoth4 on mine23:26
wolfspraulbut we need a command to do 'the right thing', i.e. to copy the uImage to /dev/mtd1, not assuming uImage contains any OOB data, and skipping bad blocks if necessary23:26
wolfspraulthere's another command 'nandwrite'23:27
nebajothits nandwrite23:27
nebajothlook at the first codeblock here: http://elinux.org/BeagleBoardNAND23:28
nebajothunder X-Loader23:28
nebajothnand unlock23:28
wolfspraulI googled a bit people seem to use flashcp23:28
nebajothbut this uses the exact kind of command line structure23:29
nebajothas usbboot23:29
nebajoththat seems like a good sign23:29
neil_wolfspraul: mtd2 done.  no apparent error.  lots of output23:29
nebajoththat does look cool though23:29
wolfspraulyeah, u are right they say 'nandwrite -p'23:30
wolfspraulmaybe the same?23:30
wolfspraulthey use flashcp23:30
neil_wolfspraul: when doing mtd3: 1 of 3568 eraseblocks contain non-ubifs data23:30
wolfspraulneil_: the good news is, you cannot break anything! :-)23:30
wolfspraulhmm, don't know what that means23:31
nebajoththat's a lot more intuitive23:31
neil_wolfspraul: :-)23:31
nebajothacts like cp23:31
wolfsprauldoesn't sound scary23:31
wolfspraullet's try to flash the kernel now23:31
wolfspraulneil_: what doyou want to try? flashcp or nandwrite? :-)23:31
neil_wolfspraul: I've got a coin here23:31
wolfspraulnandwrite -p /dev/mtd1 /boot/uImage23:32
neil_wolfspraul: shall I proceed with this mtd3 format despite the warning?23:32
wolfspraulthere may be some options in ubiformat to nuke old erase counters/OOB data23:33
neil_wolfspraul: only 0 of 3568 have have valid erase counter.  is that a worry?23:33
nebajothwolfspraul will mail you another if it breaks23:33
wolfspraulmaybe we should nuke everything23:34
wolfspraulno it's not a worry23:34
wolfspraulwe are formatting23:34
wolfspraulideally in NAND you always want to carry forward old data, like erase counters23:34
neil_wolfspraul: sorry, just not used to so much output from UN*X commands23:34
neil_wolfspraul: ah, makes sense now23:34
wolfspraulbut on the Nano, we are not in an ideal world yet23:34
wolfspraulbugs are fixed all over the place constantly23:35
wolfsprauleven if you run 'ubiformat -h' and look at the example, see the second one :-)23:35
neil_wolfspraul: no other warnings.  mtd3 done.23:35
wolfspraulubiformat -q -e : be quiet and force erase counter to 023:36
wolfspraulnebajoth: these commands cannot damage anything. impossible.23:36
neil_I'll need to put uImage on the SD before nandwrite won't I?23:36
wolfspraulassuming you have no valuable data in NAND23:36
wolfspraulneil_: it should be there already, we booted from it23:36
wolfspraultry: nandwrite -p /dev/mtd1 /boot/uImage23:36
neil_wolfspraul: of course.  silly me23:37
wolfspraulwe might even be able to reuse your rootfs and somehow uninize it into NAND, but I never tried that before23:37
nebajothgod I love this thing23:37
wolfspraulanyway, this is about getting your nano to work, not all sorts of experiments :-)23:37
neil_wolfspraul: done.23:38
wolfspraulpower on, don't press 'm', see what happens23:38
neil_oops.  lots of "Uncorrectable error occurred"23:39
neil_ND read from offset 400000 failed -7723:39
wolfsprauldid u-boot load the kernel properly?23:39
neil_ERROR: can't get kernel image!23:40
neil_I guess not23:40
wolfspraulboot from microSD again23:40
neil_ok.  doing23:40
wolfspraulI'm trying the same thing on my Nano now23:41
wolfspraulwrote the kernel with nandwrite, rebooting...23:41
neil_this is very cool.  I planned to boot Debian from SD but I didn't plan on learning so much so fast23:41
wolfspraulah yes23:42
wolfsprauluncorrectable error :-)23:42
wolfspraulscrolling down wildly for a while23:42
neil_that's it23:42
wolfsprauldon't worry this has to do with the nandwrite23:42
wolfspraullet's try flashcp first23:42
nebajothor offsets23:42
wolfspraultrial and error for a while, don't we love the Chinese ! :-)23:42
wolfspraulknowledge can come later :-)23:42
neil_ok, ready for flashcp command line23:43
wolfsprauljust try you are ahead of me23:43
wolfspraulonly watch that you focus on /dev/mtd1 :-)23:43
wolfspraulmtd0 is our precious u-boot, let's not loose it23:43
wolfspraulmtd2 is the rootfs23:44
wolfspraulsure why not23:44
wolfspraulgood idea23:44
nebajothflash_erase /dev/mtd3?23:44
wolfsprauljust please watch out - mtd1 only23:44
neil_I'm used to double checking which devices I'm formatting but good point23:44
nebajothflash_erase /dev/mtd123:44
neil_nebajoth: good idea to do that before flashcp?23:45
wolfspraulflashcp seems to assume page-aligned data23:45
nebajothneil_: I think so23:45
nebajothespecially since you've run nand write23:45
neil_nebajoth: ok, I'll try that first then23:46
nebajothflashcp /path/to/your/uImage /dev/mtd123:46
wolfspraulnebajoth: flashcp doesn't work for me23:46
wolfspraulcomplains about not page-aligned data23:46
nebajothperhaps because of the nand write you both just ran?23:47
neil_flash_erase only erased 512KB for me I think23:47
neil_is that the size of mtd1?23:47
wolfspraulyou can try flash_erase /dev/mtd1 0 823:48
neil_that's nailed it, thanks23:48
neil_flashcp says "Attempt to write not page aligned data" though23:49
wolfspraulI just did nandwrite -p again, and now it boots!23:50
wolfspraulmaybe the flash_erase helped23:50
neil_we didn't erase before nandwrite though did we?23:51
wolfspraultry the nandwrite -p again23:51
neil_I think that's nailed it.  got OpenWRT logo23:53
neil_got desktop now23:53
wolfspraulhe. puh. finally.23:53
nebajothnandwrite works?23:54
wolfspraulI think your Nano is all fine.23:54
wolfspraulsomething is up with USB, for now I just blame usbboot :-)23:54
wolfspraulnebajoth: nandwrite after flash_erase did it23:54
neil_cat /etc/VERSION  2010-06-1523:54
neil_it could also be my seven-year old laptop that only supports USB 1.1, although I've never had any other USB-related trouble with it23:55
wolfspraulyes, possible23:55
wolfspraulbecause of usbboot23:55
wolfspraulcan you try to ssh into the Nano?23:55
wolfspraullet's make sure USB works now23:55
wolfspraulIP is
neil_oops. I just reset the IP to
wolfspraulsure why not23:56
wolfsprauljust check whether you can ssh in23:56
neil_host key has changed so I guess it works23:56
wolfspraulok then23:56
nebajoththat needs to be wikified23:57
wolfspraulbe a bit patient in the first few days, there are always some quirks that need to be understood23:57
wolfspraulmeanwhile we are working hard everywhere to make everything smoother and more solid23:57
wolfspraulyour feedback helps a lot!23:57
nebajothfor me, the making things work23:57
nebajothis the fun part23:57
wolfsprauland I learnt how to do flash_erase/nandwrite for the Linux kernel now. another reason to use usbboot less :-)23:57
nebajothbuying a NN was really just like buying membership in a club of people who think the same things are fun that I do23:57
nebajothwhich is pretty epic23:58
neil_bizarre that we needed to use flash_erase even after doing a whole-device erase to begin with23:58
wolfspraulno that's OK23:58
neil_actually, we tried to use usbboot to write the kernel.  it must have written something23:58
wolfspraulNAND is tricky, all pieces of software need to be well synced with each other23:58
wolfsprauland they are not23:58
wolfspraulin terms of where/how they store bad block info, eec data (different eec algorithms), etc.23:59
neil_pretty cool that you were able to sort it all out over IRC though23:59
wolfspraulalso in general free software is not very strong on NAND related things23:59
neil_big thanks to you all.  I now have a working NN with the latest software23:59
wolfspraulI am glad we have mtd-utils and ubi-utils now, they are fixing lots of bugs too23:59
--- Tue Jun 22 201000:00

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