#qi-hardware IRC log for Thursday, 2011-09-29

hiyuhanyone can help me to fix my broken ben? http://lists.en.qi-hardware.com/pipermail/discussion/2011-September/009024.html02:06
kristianpaulhello hiyuh 02:13
kristianpaulYou can try flashing from memory card02:14
hiyuhhi kristianpaul 02:14
hiyuhyou mean using microSD card?02:15
kristianpaulbut you need still boot linux at least02:15
hiyuheh, i have not microSD mem ATM.02:15
kristianpaulwhat is your dmesg out when having this Error - can't read bulk data from Ingenic device messages btw?02:16
kristianpaullet find how force low-speed02:17
kristianpaullet me*02:17
kristianpaulah, werner alredy pointed it :)02:18
kristianpaulif you dont boot now, no rush, 02:19
hiyuhhmm, I wrote my ben won't boot (usbboot and/or standalone). but it's not, I realized, just standalone boot fails.02:20
kristianpaulgood, so you still have usbbot and kernel related error just after booting?02:20
kristianpaulhiyuh: had you already tried flashing with all your usb ports?02:21
hiyuhmy ben can do usbboot only atm. when tring standalone boot, my ben displays keeps black. so all I have to do is now runnig reflesh_ben.sh again w/ low-speed?02:24
kristianpaulwell, i think werner pointed some devices may not work in low-speed http://lists.en.qi-hardware.com/pipermail/discussion/2011-September/009021.html02:26
kristianpaulbut of course you can give it a try02:26
kristianpaulhiyuh: can you pastebin a lsusb output please?02:26
hiyuhkristianpaul, http://pastebin.com/Kwh3WTFK02:30
hiyuhw/ -v -> http://pastebin.com/PTXarcei02:31
kristianpaulalso dmesg02:31
kristianpaulI'm trying to follow this http://lists.en.qi-hardware.com/pipermail/discussion/2011-August/008508.html02:32
kristianpaulSo you be aware of tring full-speed just in case02:32
hiyuhdmesg -> http://pastebin.com/Wic2WFCm02:33
hiyuh"usb 1-1: new full speed USB device number 3 using ohci_hcd" in dmesg, so "(if it already enumerates as Full-Speed or Low-Speed, you shouldn't be reading this anyway ;-)"?02:35
kristianpaulso... hum02:37
kristianpauldo you have a really shorter usb cable? :)02:37
kristianpaulbtw, please try flash again as suguested and try to share with us dmesg out when "error" happens02:38
hiyuhi have no cable shorter than one came w/ ben. but longer ones are.02:39
hiyuhI'll try reflesh_ben.sh again.02:39
Action: hiyuh fails again.02:42
kristianpauldmesg ?02:42
kristianpaulthis may sound crazy, but do you have a usb hub near?02:45
wpwrakor maybe a different PC ?02:46
kristianpaulcan you try another cable too, this looks like a emi or bad cable issue according to some googleing :)02:46
kristianpaulwpwrak: I do no want to said that at the end :)02:46
kristianpaulbut yeah..02:46
hiyuhmy ben pluged thinkpad's usb port (a left one) directly.02:47
kristianpaulyes sure, but something is not right with that thinkpad usb port according to the last 3 lines from your dmesg02:47
hiyuhhmm, this thinkpad has another usb port on rear. i'll try again w/ it.02:50
hiyuhlsusb -v after pluged to rear usb port -> http://pastebin.com/riLJKUKY, dmesg after reflesh_ben.sh failed again -> http://pastebin.com/zNdUEgGE02:55
hiyuhmaybe this TP has internal USB hub?02:56
wolfspraulI'd say if you can, try a different computer02:56
kristianpaulhiyuh: perhaps02:56
kristianpauland ideed, you should try another computer anyway02:56
hiyuhbtw, why not using usb hub when doing reflesh_ben.sh?02:58
wpwraksometimes, hubs cause trouble. so it's standard practice to try to remove hubs from the equation if anything goes wrong. (not only for the ben.)03:00
wpwrakbut yous lsusb doesn't show any hubs. the "root hub" doesn't really count as a hub03:00
wpwrak(it's part of the USB host controller)03:01
qi-bot[commit] Werner Almesberger: README: mention also fab/ and mlztx/ (master) http://qi-hw.com/p/eda-tools/eff2e8e03:50
qi-bot[commit] Werner Almesberger: swapcmp/: swap (rename) components in a layout (master) http://qi-hw.com/p/eda-tools/886de1503:51
erikkugelhappy Jewish new year folks! ;)04:37
wolfspraulis it Jewish new year? wow04:38
wolfspraulit took me a few years in China to learn about and appreciate the significance of the Chinese new year, but I went through the process, so now I'm ready for any other discoveries! such as jewish new year! Happy new year then, indeed!04:39
erikkugelyup :) it's year 5772! But, I think the Chinese calendar might be the only one to beat THAT.04:39
qwebirc15270it's hiyuh on another box.05:09
qwebirc15270echo 6 >/sys/bus/usb/drivers/usb/usb1/../companion05:10
qwebirc15270bash: /sys/bus/usb/drivers/usb/usb1/../companion: No such file or directory05:10
wolfspraulqwebirc15270: sorry I lost track of your endeavor somewhere05:12
wolfspraulyou are trying to reflash your nano?05:12
wolfspraulwhat's the status of it right now? does it boot? (the nano)05:12
wolfsprauldid it boot before? what was the last thing you did to it when it was still working?05:12
wolfsprauljust trying to help...05:12
qwebirc15270wolfspraul: yes, using reflesh_ben.sh on another x86 box. but it fails. http://pastebin.com/uc95ND8Q05:14
wolfspraulok but let's unwind a little, it seems you are stuck trying some specific thing, so I'm looking for a workaround05:15
wolfsprauldid the nano work before? what was the last you did on it? does it boot now?05:15
wolfspraulthere are many ways to update it, I'm just trying to understand its current status05:15
qwebirc152701) my ben works as out-of-box. but shipped image looks outdated, so i want to update. 2) I tried to update by reflesh_ben.sh on gentoo/~ppc, but it fails. (maybe b/c endian bug in xburst-tools) 3) I tried to update by reflesh_ben.sh again on gentoo/~x86, but it fails. (maybe b/c thinkpad usb port sucks) 4) now I'm on an ancient fedora10 box I found, and try to usb speed change W/A.05:21
wolfsprauldoes your ben boot right now?05:22
qwebirc15270but it fails :( <- current status05:22
wolfspraulmaybe somethign is wrong earlier05:22
wolfspraulcan we step back for a moment?05:22
wolfsprauldoes your ben boot right now?05:22
qwebirc15270you mean usb boot?05:23
wolfspraulno, normal boot05:25
wolfsprauloh, it won't I think :-)05:26
wolfspraulI just looked at your pastebin transcript05:26
wolfspraulthat looks like you actually did some flashing, but then it aborted mid-way05:27
wolfspraulbut your u-boot is probably still ok (it looks like from the transcript)05:27
qwebirc15270normal boot fails. the display keeps black.05:29
qwebirc15270usb boot looks ok to me. http://pastebin.com/fyGLccr905:29
wolfspraulwhen rebooting your ben after reflashing, disconnect the usb cable and take out the battery, and wait 10-20 seconds05:31
wolfsprauljust in case05:31
wolfspraulyour first transcript looks like u-boot was flashed just fine05:32
wolfspraulthen the screen should turn on for sure05:32
wolfsprauleven if it cannot find or load the kernel05:32
qwebirc15270i see this -> Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(0,0)05:35
wolfspraulah great05:35
wolfspraulan alternative way to flash your ben is to make a bootable memory card, and then flash it from there05:35
wolfspraulthe only thing you need in NAND for that is u-boot, which you have05:35
wolfsprauldo you want to try that? do you have the patience?05:36
qwebirc15270you mean the way using microSD card?05:37
qwebirc15270eh, i have no microSD card atm...05:38
wolfspraulhmm, ok. then we have to find a more stable USB host05:38
wolfspraulmaybe you are just unlucky with the 2 systems you tried05:39
wolfspraulit seems to be working in principle, but then aborts after some time05:39
wolfspraulso I'm back to where everybody else was already before :-) no silver bullet...05:39
qwebirc15270nope, i tried 3 boxes (gentoo/~ppc, gentoo/~x86 and this ancient f10). and it looks i should use microSD.05:42
wolfspraulpainful, sorry to hear you had such bad luck05:44
wolfspraulare you using the latest xburst-tools ?05:44
qwebirc15270yes, tried git and latest .deb.05:45
wolfspraulnormally the flashing is robust, you made it past all the critical points (which is how to get into usbboot mode)05:45
qwebirc15270wolfspraul: anyway thank you, i'll be off from this ancient box. :)05:50
hiyuh2GB is enough to create bootable microSD?06:43
dvdkhiyuh: still around?07:34
dvdkjust read your reflashing endevour in the  qi-hw irc logs07:35
dvdkabout your questions: 2 GB should be more than enough for a bootable micro-sd07:35
dvdkon the other hand: what's wrong with the fedora box?  you wrote it won't support usb speed change, but maybe that's not neccessary on the fedora box anyways?07:36
dvdkBTW I've been mostly flashing at high-speed through a hub.  never saw a single error.07:36
wolfspraulyes that's how it should be07:41
wolfsprauland mostly is07:41
wolfspraulwe had a few rare cases where people had stubborn problems on one computer, but switching to another one always solved the problem for them07:41
wolfspraulnever had a case with someone who had 3 different computers in a row like this07:41
wolfspraultoo bad07:41
dvdkwolfspraul: I think he didn't try to actually flash via the 3rd (fedora) box.  just gave up when he weren't able to force it to full-speed (?)07:44
dvdki'd say he just needs to do a few more tries, switching cable, inserting a hub, and trying the fedora box at high-speed.07:44
hiyuhdvdk, thanks07:45
dvdkwolfspraul: btw milkymist is featured at theregister.co.uk07:45
dvdkwolfspraul: http://www.theregister.co.uk/2011/09/28/milkymist/07:45
dvdkhiyhuh: saw my answers above?  (10 minutes ago) anything you didn't already try?07:46
dvdkhiyuh: BTW here's the documentation for flashing via microsd-card (didn't try myself though):07:48
hiyuhdvdk: on the fedora, reflesh_ben.sh w/ high speed fails at end of kernel flashing. it boots but no rootfs.07:49
hiyuhyes, now I bought microSD and it does boot here.07:49
dvdkhiyuh: you really seem to be extremely unlucky :)07:49
hiyuhbut there is no mtdinfo :(07:50
wolfsprauldvdk: yes, featured, but... the article is confusing and we need to learn a lot from it and improve our story telling07:50
wolfspraulwe don't have your headline making skills! (vga abused...)07:50
wpwrakhiyuh: wait a minute ... does the flashing fail ? or does flashing succeed but then it doesn't boot ?07:50
dvdkhiyuh: have you configured network?  you should be able to install mtdtools using opkg07:51
wpwrakhiyuh: because there's also a failure scenario where the kernel just doesn't find the root file system. a while ago, the partitioning has been changed, but that change was never properly deployed, so there's the possibility that you have a system made from incompatible parts07:52
dvdkwhich firmware image did you install on the microsd?07:52
hiyuhwpwrak: dunno, booting after doing reflesh_ben.sh results kernel panic said no rootfs07:52
wpwrakhiyuh: so the reflash itself completed without an error ?07:52
dvdkwpwrak: didn't he say the reflash *of the kernel* completed, ?07:52
wpwrakdvdk: ah yes, good point07:53
wpwrakhiyuh: if you haven't done so yet, you may also have to reflash u-boot and the root file system07:54
wpwrakhiyuh: particularly if you've upgraded from a kernel with the old partitioning to one with the new partitioning07:54
hiyuhwpwrak, on fedora, reflesh_ben.sh vomits error at end of flashing kernel.07:54
wpwrakhiyuh: did you try reflashing on any of the other two systems ? or did you just check enumeration there ?07:55
dvdkhiyuh: try 'opkg install mtd-utils-flash-info mtd-utils-flash-erase', and probably many more tools.07:56
dvdkthe list of pakcages is here:07:56
dvdkcan also just copy the .ipk packages to your ben, then run 'opkg install *.ipk' (as long as there are no missing dependencies)07:56
hiyuhwpwrak, anyway, i would not like to try usbboot anymore. it's waste of time to me. i've tried 3 systems...07:57
hiyuhdvdk, thanks i'll try.07:58
dvdkwpwrak: i guess error rate is so high, that it fails somewhere in the huge rootfs part, but bootloader+kernel flashed ok07:58
wpwrakhiyuh: so you got an error on all three systems ? okay, that's bad then07:58
dvdkif the opkg works, we should probably update the wiki07:58
wpwrakdvdk: possible, yes. that would suggest a hardware problem on the ben side then. clock instability or something wrong with the USB electronics07:59
dvdkhiyuh: ah, you may need to run 'opkg update' before you can retrieve packages from the webm07:59
dvdkwpwrak: broken cable? :)08:00
dvdkwpwrak: the first system doesn't really count.  that's a PPC and usb-boot seems to be broken on big-endian08:01
hiyuhwpwrak, 1) my ben works as out-of-box. but shipped image looks outdated, so i want to update. 2) I tried to update by reflesh_ben.sh on gentoo/~ppc, but it fails. (maybe b/c endian bug in xburst-tools) 3) I tried to update by reflesh_ben.sh again on gentoo/~x86, but it fails. (maybe b/c thinkpad usb port sucks) 4) I'm on an ancient fedora10 box I found, and try to usb speed change W/A but it fails. using reflesh_ben.sh w/ high-speed also fails.08:02
hiyuh 5) now i'm tring to update w/ micorSD.08:02
hiyuhhmm, "Ethernet over USB" needs a driver on linux host?08:06
dvdkhyuh: should be included by default in all "normal" kernels08:06
dvdkbut you need to 'ifconfig usb0' on the  host for usb to appear08:06
dvdkjust watch 'tail -f /var/log/messages'08:07
hiyuhmaybe I should recompile my kernel to use this -> "Netchip Technology, Inc. Linux-USB Ethernet Gadget"08:08
dvdkhiyuh: you're using a self-made kernel?08:08
hiyuhdvdk, yes08:08
dvdkok, so not a "normal" kernel08:08
dvdkalso, normally routing isn't configured, so usb0 won't connect to the internet08:09
dvdkmost people around here setup routing + NAT for the usb0 device on the host.08:09
dvdki prefer to just run a http proxy on my pc, then configure opkg to use it08:09
Action: hiyuh uses abnormal kernel on junky boxes08:10
wpwrakhiyuh: (usbboot) hmm, looks bad, yes08:10
hiyuhmaybe, CONFIG_USB_ETH is needed?08:14
dvdkhiyuh: probably08:14
dvdksure you need the "ethernet gadget" code?  isn't that maybe tho wrong side of the driver (implementing usb client, not host code?)08:15
dvdkall the usb-gadget stuff is AFAIK for device-side drivers, not host-drivers08:15
hiyuhethernet over usb needs CONFIG_USB_NET_CDCETHER on linux host.09:07
hiyuhtried "opkg install mtd-utils-flash-info" on beb which boots from microSD, but mtdinfo is missing. why?11:57
hiyuhmy ben is updated finally.12:17
wolfspraulhiyuh: you deserve the xburst-tools survivor badge12:20
wolfspraulhanded over by dvdk or whoever helped push you over the finish line :-)12:21
wolfspraulglad to hear that btw, I was a bit worried you would be stuck12:21
dvdkhiyuh: so which mtd-package was it finally?  somebody ought to add that info to the wiki page12:39
hiyuhwolfspraul: thanks12:41
dvdkwolfspraul: i think it was kristianpaul who realized (some 8 hours ago!) that boot-loader was still working and micro-SD boot would still work.  i think that was the most important bit in the reflashing saga.12:41
hiyuhdvdk: mtdinfo is still missing. but it's not so important to update, wiki just shows usage of mtdinfo on http://en.qi-hardware.com/wiki/Updating_Ben_with_an_SD12:43
wpwrak(visual diffs) ah, they actually have done a bit more. just the picture at adafruit isn't so great12:45
hiyuhdvdk: http://pastebin.com/mu9tPfc312:46
wpwrakrejon: comment posted :-)13:03
wpwrakwolfspraul: ah, how about adding labsw to the schematics diffs ?14:34
wpwrakwe made some news, too :) http://www.adafruit.com/blog/2011/09/29/visual-revision-history-for-schematics-for-kicad-and-git/14:56
wolfsprauloh nice! [news]15:06
wolfspraulyes sure, will add tomorrow [labsw]15:06
wpwrakkewl, thanks ! it's in wernermisc/labsw/15:06
larscnice captcha on that site: "Prove you are human by reading this resistor: ..."15:17
rjeffrieswonder what will happen with this:17:25
rjeffriesthe Wall Street Journal reported that Nokia is developing a Linux-based operating system code-named Meltem for its low-end phones. The report, citing unnamed sources, said that Mary McDowell, Nokia's executive vice president in charge of mobile phones, is leading the effort. The focus is to improve Nokia's competitiveness in emerging markets, the report said, which also noted that Nokia has begun releasing dual-SIM phones in 17:25
erikkugelNokia's insistence on ignoring Android (which can by now run its lower end devices without any major investment on their end) amazes me. What are the chances this new linux flavour Nokia will make for their phone will be more "open" than Android?17:36
kristianpaulerikkugel: if they release the gui, i a good start :)18:58
kristianpaulalso low end phones sound nice i think :)18:58
virickyak: any good luck with offrss? :)19:25
kyakviric: didn't yet have a single chance to get to it (no time) - still using google reader :)19:25
virichehe ok :)19:26
kristianpauli can try, i currentky use newsbeuter, but if there is a good excuse to move, why not19:28
virickristianpaul: you're welcome! But it's up to you. It's already in the "works for me"19:35
kristianpaulsupport for pictures, looks a  good reason to me :)19:44
kristianpaulviric: so are you using it with elinks?19:44
viricwell, I mostly use it with firefox in my laptop19:45
viricbut I rarely used elinks or lynx in the nanonote19:45
kristianpauli got you !19:51
kristianpauli tought you'were talking about nanonote :)19:52
viricwell, it builds and works in the nanonote too :)19:56
viricI'm not advising you to run eclipse. I advise you to run a program that *also* works on the nanonote19:56
kristianpaulhaha ;)20:00
kristianpauloffrss: offrss.c:431: open_urls: Assertion `space != ((void *)0)' failed.20:19
kristianpaulno OPML ? :)20:19
viriceh? :)20:19
viricno OPML20:19
viricompl is a too modern thing20:20
viricwhat did you write in your feedurls.txt?20:20
kristianpaulan opml export from newsbeuter 20:22
viricah no no20:22
virickyak also wanted something like that.20:22
viricthe feedurls.txt format is simple and offrss-only.20:22
viricdescribed in the offrss web page20:23
kristianpaulok i'll look the webpage and not asume fancy features ;)20:23
virickristianpaul: or 'doc/home.wiki' in the package20:23
wpwrakDocScrutinizer: hmm, a bit of mystery: http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=ED2679-ND23:21
wpwrakDocScrutinizer: note the "wash-away". the data sheet says that the carrier dissolves in water. for what on earth would one use such a feature ?23:23
DocScrutinizermouse camera, with lens downwards?23:26
DocScrutinizernah, I think I have *no* idea23:27
wpwrakgood. i'm not alone with my puzzlement then ;-)23:29
DocScrutinizerjust can imagine you may want to solder a set of 6 (8?) contacts with proper raster, then for whatever reason need to get rid of the carrier23:29
wpwrakyeah, i could vaguely imagine sending the board through SMT afterwards, to fill the space between the pins with components. seems odd, though. and i'm not sure if this would actually work23:31
wpwraki mean, first problem: how to apply the solder paste ?23:31
wpwrakthen, won't have whatever soldering process you used for the DIP contaminated the SMT pads as well ?23:32
wpwrakand so on. sounds like an interesting solution for a problem that can't exist ;-)23:32
DocScrutinizer51sealing with acryl or epoxy23:46
--- Fri Sep 30 201100:00

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