#milkymist IRC log for Monday, 2012-04-23

wpwrakkyak: (utf8) rejoice in the unblemished purity of plain 7 bit ASCII ! :)00:16
kristianpaulxiangfu: hi !00:17
kristianpaulsorry i remenber a log from you about a milkyminer taking too long to miner?00:17
kristianpaulor just calculated hash once..00:17
wpwrakdvdk: that sounds like the yaffs rootfs bug00:18
dvdkwpwrak: already reformatted.00:18
dvdkyes, remembered youm mail on devel@m00:19
xiangfukristianpaul, yes. I use the payload.py. it need ~5m give a result.00:19
wpwrakdvdk: i'm a bit surprised you ran into that one, though. what build did you install ?00:20
dvdk1.2 was 1.1 before00:20
wolfspra1lgood morning everybody00:20
wolfspra1lnew week, new luck00:20
wolfspra1lkicad, boom! where are thou00:20
wpwrakbeing hacked on, all of them i suppose :)00:21
wpwrakdvdk: hmm. makes me wonder what's inside 1.2 then. that problem should have popped up only recently, yet might be gone in the daily builds already.00:22
dvdkwpwrak: if even you don't know that... :)00:22
wpwrakdvdk: i know very little about the actual builds, only about the things that may go into them ;-)00:23
xiangfudvdk, try reflash_m1.sh --qi --data00:24
Action: dvdk tries to not think about what may be in there :)00:24
wpwrakat daytime make saussages. at night, we hack code ...00:25
dvdkxiangfu: just trying.  but i though --qi should reflash everything, including --data!?00:25
xiangfudvdk, --qi will reflash everything except data partition.00:25
dvdkwell, then help output from reflash_ben.sh is broken.00:25
xiangfudvdk, --data will reflash data partition.00:26
xiangfudvdk, ? you mean this one: CAUTION: if '--data' enable, it will REFLASH DATA PARTITION ?00:26
wpwrakhmm, 1.2 ought to be first of march. very odd00:27
dvdkxiangfu: ah, this line.  skipped over it  I guess :)00:28
dvdkafter seeing '--qi [VERSION] [--data]  ' i gave up understanting command line :)00:28
wpwrakthey changes the permissions on march 2, though. maybe that just crept in at the last moment00:28
wpwrakthat would also mean that all the people who have upgraded to 1.2 have chosen to suffer in silence00:29
xiangfudvdk, there is a workaround method to exit rendering without keyboard.00:30
xiangfudvdk,  Hold left button. it will exit the rendering mode. and goes to webupdate.00:31
dvdkah cool.00:32
xiangfuthe next release will automatic save all configuration.00:36
wpwrakxiangfu: you may want to check if 1.2 works with a yaffs created with an earlier release (and not auto-upgraded by my chmod commit from 4 days ago)00:39
xiangfuwpwrak, yes.00:40
wpwrakxiangfu: it seems that 1.2 may have picked up the new permission check in rtems, but nobody noticed. at least that would explain the problem dvdk has encountered00:40
xiangfuwpwrak, yes.00:41
xiangfuwpwrak, since I always reflash with --data. I must miss this problem.00:41
wpwrakah. maybe that explains it00:41
xiangfuwpwrak, working on that in next hour. :)00:41
Action: xiangfu reflash back to 1.1 with data partition.00:56
wpwrakls -l /ssd   shows 666 ?00:57
xiangfuwpwrak, yes.: drw-rw-rw-  1 root  root   512 Jan  1 00:00 ssd00:59
xiangfudrwxr-xr-x  1 root  root  7504 Jan  1 00:00 dev00:59
xiangfudrwxr-xr-x  1 root  root   536 Jan  1 00:00 etc00:59
xiangfudrw-rw-rw-  1 root  root   512 Jan  1 00:00 ssd00:59
xiangfu^^^ ls -l /00:59
xiangfunow reflashing to latest qi image.01:00
xiangfuwithout touch data partition01:00
wpwrak666, the number of the beast. so far so good :)01:02
xiangfucannot reproduce the bug. :(  after reflash latest qi image. it still works fine.01:08
wpwrakafter reflashing 1.2 ? or the latest daily build ?01:09
xiangfuafter reflashing  1.2, it still works fine.01:10
wpwrakand /ssd is still 666 ?01:10
xiangfulet me do this again.01:12
xiangfu1. I run reflash_m1.sh --rc3 00 3a: it will download image from : http://milkymist.org/updates/2011-11-29/ include the data partition.01:12
wpwrakperfect. that's certainly old enough01:12
GitHub61[board-m1] adamwang pushed 1 new commit to master: https://github.com/milkymist/board-m1/commit/4a2b950a6fefed364f2035794ed1cbc375237aae01:14
GitHub61[board-m1/master] corrected J21 pin assignments - Adam Wang01:14
xiangfuls -l shows: drw-rw-rw-  1 root  root   512 Jan  1 00:00 ssd, I can open file under patch editor.01:16
xiangfu2. now running reflash_m1.sh --release. which will donwload image from http://milkymist.org/updates/2012-03-01/ reflash without touch data partition.01:16
wpwrakwait ...01:16
xiangfupressed 'Enter'01:17
wpwrakthat's odd. if you installed all the 2011-11-29 stuff, you should still be able to open files01:17
wpwraksince RTEMS and data partition will both agree on using 66601:18
xiangfuyes. I can open file under patch editor.01:18
wpwrakoh, alright01:18
wpwraksomehow, i read "can't" :)01:18
wpwraklet's see how 2. goes then :)01:18
xiangfuafter [reflash_m1.sh --release], ls -l shows: drw-rw-rw-  1 root  root   512 Jan  1 00:00 ssd01:19
wpwrakvery good01:20
xiangfupatch editor open file just fine.01:20
wpwrakexcellent. that's how it should be01:20
wpwrakso dvdk ran into something else01:20
wpwrakthanks for checking !01:21
xiangfumaybe after step_1. I should edit some file. and save etc..01:21
wpwraknaw, the results look quite conclusive to me. with the permission problem, nothing worked01:22
xiangfuok. now I am reflashing the qi image.01:25
xiangfubuild under "2012-04-02/" it include the rtems-yaffs2 change.01:26
wolfspra1lwhy did the frog pad and alpha grip keyboards not work for david?01:27
xiangfuls -l shows: drwxr-xr-x  1 root  root   512 Jan  1 00:00 ssd.01:27
xiangfu( dvdk ran into something else) yes.01:28
wolfspra1llooks like 2 interesting little keyboards01:28
xiangfuhave to dump the hid description for check more detail.01:32
kristianpaul"new firmware doesn't allow me to exit patch via mouse click." arghhh that explain why on the demo i had i tough mouse support was broken !01:53
kristianpaulbut indeed is  a good safe measure01:54
wolfspra1lxiangfu: can you explain to david how to easily dump the hid descriptor?01:56
xiangfuI have replied his email. and kyak's email.01:56
wolfspra1lah :-)01:56
kristianpaulbtw rii kbd also supports touchpad?01:57
wolfspra1lyou have one, no?01:58
xiangfuhere is the detail by Werner: http://lists.en.qi-hardware.com/pipermail/discussion/2011-November/009206.html01:58
kristianpaulbut havent tried touchpad..01:58
wolfspra1lthat should only take 1 second01:58
kristianpaulwill do .)01:58
wolfspra1lyou mean the touch area on it?01:58
wolfspra1ljust try :-)01:58
Action: kristianpaul brings a extra monitor from the closet01:59
Action: kristianpaul turns on m102:03
kristianpaulewww :)02:03
kristianpaulI love yopu guys ! :)02:04
kristianpaulwhat was the short cut to get a rtems shell.. Fallenou ?02:12
wolfspra1lhe's sleeping I'm sure02:12
xiangfuflterm --port /dev/ttyUSB002:13
xiangfukristianpaul, ^02:13
xiangfukristianpaul, oh. you mean the rmmod?02:13
kristianpaulnope, in FN02:14
xiangfukristianpaul, you cannot do that under FN. no such GUI for rtems shell.02:15
kristianpaulhmm, sure?02:15
xiangfufrom my understanding. have to connect through /dev/ttyUSB002:18
kristianpaulwhere is the no need for PC ;)02:19
kristianpaullol the ezcap IR control seems work ;)02:20
kristianpaulah, partially02:21
kristianpaulwell :)02:21
xiangfuit not RC5.02:23
GitHub102[board-m1] adamwang pushed 1 new commit to master: https://github.com/milkymist/board-m1/commit/745baa077b33162046dfdebd4ccbad403033587003:09
GitHub102[board-m1/master] added VIDEOIN_A3V3, VIDEOIN_A1V8 components to fix ErrType(3): Pin connected to some others pins but no pin to drive it. (i.e. a local lable can't drive to a type of Power input). - Adam Wang03:09
GitHub76[board-m1] adamwang pushed 1 new commit to master: https://github.com/milkymist/board-m1/commit/64c29af1f0f4c1fdb9e4007f2fd69d45a5e721cc03:14
GitHub76[board-m1/master] added ETH_A3V3, ETH_A1V8, ETH_PLL1V8 components to fix ErrType(3): Pin connected to some others pins but no pin to drive it. (i.e. a local lable can't drive to a type of Power input). - Adam Wang03:14
wpwrakah, --qi also means "snapshot". that could explain dvdk's trouble. and indeed, the latest version is from 2012-04-02, just when that bug was around.07:30
wpwrak--qi vs. --release does indeed seem a bit redundant. --qi is only a small superset of --release, and far from having the granularity of http://fidelio.qi-hardware.com/~xiangfu/build-milkymist/07:33
xiangfuwpwrak, --qi have latest standby.bit splash.raw etc. for factory reflash in future. before the --release.07:36
xiangfuif we already have those standby.bit etc. it's a bit redundant. :-)07:46
wpwrakmaybe just add the missing files to milkymist.org/updates/current/ ?07:52
wpwrakthey do exist under http://milkymist.org/updates/, just scattered over various directories07:53
wolfspra1lif --qi means snapshot it should be called --snapshot, no?08:12
wpwraki'd rather call the things under build-milkymist a "snapshot"08:14
wpwrak--qi is sort of a release, just sometimes slightly time-shifted. and sometimes with what looks like multiple tries08:15
wpwrakcompare: --release, http://milkymist.org/updates/08:16
wpwrak--qi, http://downloads.qi-hardware.com/software/images/Milkymist_One/08:16
wpwrak--snapshot: http://fidelio.qi-hardware.com/~xiangfu/build-milkymist/08:16
wpwrak--snapshot is undocumented in the online help of reflash_m1.sh08:17
wpwrak(to paraphrase some old text adventure games) you're in a maze of twisty little releases, all different ;-)08:18
cladamwwpwrak, Does this "bzr checkout -r 3494 lp:kicad kicad.bzr" is to copy 3494 version from https://code.launchpad.net/kicad/+branches?field.lifecycle=DEVELOPMENT&field.lifecycle-empty-marker=1&field.sort_by=newest+first&field.sort_by-empty-marker=1  ???08:25
wpwraki guess the things it takes are roughly what you can find via that page ...08:29
wpwrakbut i don't know that web interface. so maybe there are files you can't easily access.08:29
GitHub169[board-m1] adamwang pushed 1 new commit to master: https://github.com/milkymist/board-m1/commit/7e99faf754dc7d8ef5ca48c6a5bd97d3024064fa08:54
GitHub169[board-m1/master] latest updated files - Adam Wang08:54
kyakkristianpaul: sorry, i didn10:18
kyak't understand your reference to CSR bus10:18
kyakwhat did you try to say? :)10:19
kyakxiangfu: for the HID description - i will do that. But from what wpwrak said, there is only basic HID support and he was looking forward to Linux10:21
kyakso.. not so clear - do you still need the HID description?10:22
xiangfuI think so. I would like to checkout what is different on your keyboard.10:23
kyakok, then i'll send it as soon as i can10:24
kyakwhen Linux hits M1, perhaps atusb would be upstream already10:25
kyakthis would have been awesome10:25
wpwrakkyak: it's always good to know what weird descriptors people use :)10:25
kyakwpwrak: so it is just for your amusement, i see :)10:25
wpwrakkyak: amusement is what we can be sure of. all the rest is more difficult ;-)10:26
kyakxiangfu: what is the version of oscsend you are using? I'm using the one provided by liblo (if i remember correctly) package in arch linux.. It displays OSD fine (though no UTF-8), but the /patch command doesn't owkr10:27
xiangfufor /patch. you have to configure under Flickernoise.10:28
kyakok, so it's not enabled by default10:29
kyakis it a 'yes' no? :)10:29
xiangfufor simple mode. there are no keyboard/IR/osc/midi configured.10:30
xiangfuI have plan to try to automatic hook keyboard/IR/osc/midi configure under simple mode.10:30
kyakwhat is the simple mode?10:31
kyakis it the simple mode in Flickernoise interface when you start performance?10:31
xiangfuyes. by default it using simple mode.10:31
xiangfuand there is a configure mode.10:32
kyakright.. i wondered what the configure mode means and what i can configure10:32
kyakshould i provide some configuration file?10:32
GitHub179[board-m1] wpwrak pushed 1 new commit to master: https://github.com/milkymist/board-m1/commit/32e1f7479cedaea12dd9975444f0482557f06d3b10:33
GitHub179[board-m1/master] r4/Power.sch: replaced PWR_FLAG with new POWERED - Werner Almesberger10:33
xiangfukyak, I would advice you update to recently daily build. that support automatic save all configure10:34
xiangfukyak, without this feature. it's a little bit hard to play with configure mode. :-)10:34
xiangfubasically, it edit the 'Keyboard' 'MIDI' 'OSC' etc. and click the 'Save' under Control Panel.10:35
kyakok, that i can do.. though it's hard to upgrade without jtag properly working.. have to upload firmware to FTP and update from files10:35
xiangfuthen when start the performance. chose the 'Configure' not 'Simple mode'10:35
xiangfus/ it edit the 'Keyboard' 'MIDI' 'OSC' etc./you edit the 'Keyboard' 'MIDI' 'OSC' etc.10:36
kyaki hope i don't need a keyboard to perform these configuration :)10:37
xiangfumouse is enough :-)  but exit render mode have to use 'Ctrl + ESC' or HOLD left button. (this will exit render mode and goto webupdate)10:38
kyakxiangfu: do you know why holding middle button turns off M1 if it runs (as expected), but then it immediately turns on again (unexpected)?10:40
wpwrakkyak: for the jtag, can you try with a different usb power ? also, which version of libusb do you have ? 1.0 or 0.1 ?10:41
xiangfukyak, hold middle button is for reboot.10:41
kyakxiangfu: hm.. i thought pressign all three buttons is for reboot. How can i power off?10:41
xiangfukyak, there is no 'PowerOFF' any more.10:41
wpwrakkyak: (middle) we used to have an "off" mode (standby) but that's gone now. so if you find any references to that behaviour, they're obsolete :)10:41
xiangfuPowerOFF == unplug the power cable.10:42
kyakwpwrak: my libusb is 1.0; what do you mean "different usb power"?10:42
wpwrakerr, different usb PORT10:42
wpwraksorry, multitasking confusion :)10:42
kyakhm, ok.. it's just convenient to leave it plugged in and just power off.. but whatevrr :)10:42
kyakwpwrak: i tried various ports of course, and different cable. But as i said, the same setup works fine for flashing Ben \10:43
wpwrak(leave it) if you turn off the screen, you'll never notice :) power consumption in "standby" wasn't all that much below normal power consumption anyway10:44
xiangfu( if you turn off the screen) :-D10:45
kyakok, that's fine :)10:45
wpwrak(jtag) that would suggest a software problem. hmm. i'm on ubuntu, also libusb 1.0, and it works just fine.10:46
kyakwpwrak: if i install ubuntu in a virtual machine and jtag works there, will it prove that the problem is a software?11:05
kristianpaulkyak: are you interested adding custom cores to milkymist SoC?11:06
kyakkristianpaul: ahh, ok, i got you now :)11:06
Fallenoukyak: I think yes11:06
Fallenouit will prove it's just libusb/urjtag/linux related problem11:06
wpwrakkyak: it would be a strong hint towards the problem being software, yes. or very wickedly buggy hardware :)11:06
kyakok, then i'll try ubuntu :)11:07
kristianpaulkyak: i noticed you are interested on hdl coding, dont you?11:12
wolfspra1lkyak: maybe you can just boot from a live ubuntu usb stick, or try ubuntu in a virtual machine11:17
wolfspra1las I explained, that fixed the same problem for me, and now I have a debian virtual machine just for m1 reflashing over jtag11:17
kyakkristianpaul: "coding" - no :) hdl autocode generation - perhaps :)11:26
kyakwolfspra1l: yep, booting from a stick - good idea11:26
GitHub164[board-m1] adamwang pushed 1 new commit to master: https://github.com/milkymist/board-m1/commit/de852f58dca2bf23ab14feac63db607132b3097513:07
GitHub164[board-m1/master] added bigger text within each hierarchical rectangle by functional naming. - Adam Wang13:07
Fallenou5/win 3613:15
kristianpaulkyak: autocode, you mean by using migen or such?14:12
GitHub142[board-m1] xiangfu pushed 1 new commit to master: https://github.com/milkymist/board-m1/commit/52f8f7865c87b34f338dc03488f846e6340b040714:14
GitHub142[board-m1/master] FPGAdec.sch: remove 2 useless junctions - Xiangfu14:14
kyakkristianpaul: by using mathworks hdl coder14:21
kristianpaulwhat? !14:26
kyak"what" in like "what is that" or like in "are you crazy?"? :)14:53
kyakwell, you had 20 minutes to google.. so the first question should disappear :)14:54
GitHub25[board-m1] wpwrak pushed 1 new commit to master: https://github.com/milkymist/board-m1/commit/09871289563b0eaa79f8300a94235fbb4e86ecb115:44
GitHub25[board-m1/master] r4/FPGA.sch: redraw FLASH_A0 through _A4, magically removing the ERC issues - Werner Almesberger15:44
kristianpaulkyak: what of, what is that sounds scary :)16:23
qi-botThe firmware build was successful, see images here: http://fidelio.qi-hardware.com/~xiangfu/build-milkymist/milkymist-firmware-20120423-1648/16:32
mwallewpwrak: shouldnt the directory be  777 ?17:03
mwallethe /ssd17:03
wpwrakmwalle: yes. and nowadays it is :)17:12
qi-botThe firmware (using branch) build was successful, checkout the VERSIONS for detail, see images here: http://fidelio.qi-hardware.com/~xiangfu/build-milkymist/milkymist-firmware-20120423-1832/18:16
mwalle*sigh* i guess i need an handwritten isr for usb token reception and transmission :(21:19
wpwrakas long as it doesn't have to be written by hand AND in blood ...21:37
wpwrakbut i guess i should have answered "when i was young, we had to write ISRs with our own blood" :)21:53
mwallehow old are you? :)22:15
wpwrakyeah, you got me. we already had computers back then. didn't have to summon demons to do our bidding.22:19
mwallehehe no vampire then :)22:20
wpwrakmidnight .. good hunt then ! ;-)22:25
--- Tue Apr 24 201200:00

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