#milkymist IRC log for Thursday, 2012-02-23

cladamwmorning !01:03
cladamwwpwrak, if still reserve un-used fpga pin, you meant to connect limiter switch's fault pin ? right ?01:03
wpwrakyeah. and maybe the enable pin as well.01:09
wpwrakenable would have three functions: 1) turning off the DCC5V if we want to do this for some reason. 2) re-enable DCC5V after a voltage inversion (output > input). 3) in case we have troubles sourcing the right chip, increase the choice for alternatives.01:12
wpwrak1) and 2) seem unlikely or exotic. but 3) could be interesting.01:13
cladamwGood, Ven can be worked with 3.3V directly. Vfault don't know...checking...01:27
cladamwFault is open-drain, also good news for 3.3V.01:31
davidc__Hey; migen looks pretty interesting - is it in a state where non-milkymist-people should start writing code against it, or is a lot of change in the interfaces expected?01:32
GitHub73[milkymist] xiangfu pushed 1 new commit to master: https://github.com/milkymist/milkymist/commit/8cbdb53f9c2bf6d7de456d2d850faac338720adf03:56
GitHub73[milkymist/master] Makefile: add sdk-install, tools-install - Xiangfu Liu03:56
GitHub127[scripts] xiangfu pushed 1 new commit to master: http://git.io/sJp1cg03:56
GitHub127[scripts/master] update with milkymist Makefile change - Xiangfu Liu03:56
wpwrakxiangfu: while you're at it, how about s/successfull/successful/ ? :)03:58
xiangfuwpwrak, you mean the daily build?03:59
wpwrakthe message in the Makefile04:00
wpwrakthe word "successfull" doesn't exist. it's "successful". a rather common mistake :)04:01
xiangfuwpwrak, thanks. fixed04:02
GitHub65[scripts] xiangfu pushed 1 new commit to master: http://git.io/3Tmq2w04:03
GitHub65[scripts/master] fix typo - Xiangfu Liu04:03
wpwrakthanks ! :)04:03
xiangfuwpwrak, when we can merge the 'imaginarium' branch to master?04:03
wpwrakit's already there, isn't it ?04:03
wpwraki just keep it around in case i want to do more experiments there.04:04
xiangfuwpwrak, oh. sorry. yes. merged at 02-18.04:05
xiangfuwpwrak, I have to find out what wrong with my usb-midi(ICON) .04:09
wpwrakso it's still not working. hmm ....04:11
wpwrakmaybe try some of my binaries. they're a bit older but at least the soc shouldn't have changed: http://downloads.qi-hardware.com/people/werner/tmp/m1-20120126/04:12
xiangfuI am reflashing with 'http://fidelio.qi-hardware.com/~xiangfu/build-milkymist/milkymist-firmware-20120222-1455/' now. for cleanup and sync all with latest code.04:12
wpwrakand http://downloads.qi-hardware.com/people/werner/tmp/m1-20120125/04:12
wpwrakicon should work with these04:12
xiangfuwpwrak, works with 'm1-20120125'04:18
wpwrakvictory !!! ;-)04:19
xiangfuok. it's not soc.fpg problem.04:20
xiangfuso must be flickernoise again.04:20
wpwrakif something usb fails, that would point to a broken milkmist.git (softusb) -> rtems -> flickernoise path04:21
wpwrak(and it's good that it isn't a soc problem. because i wouldn't have known where to look in this case :)04:22
wpwrak(path) actually, it's milkymist.git (softusb) -> milkymist.git (libhal) -> rtems -> flickernoise04:23
wpwrakway too complicated :)04:23
xiangfuwpwrak, we have some regression on display by using 13.4 Xilinx ISE. too bad.04:26
xiangfuwpwrak, the screenshot look ok.04:26
xiangfubut my monitor shows something wrong. :(04:27
xiangfulet's start with milkymist.git(softusb) :)04:34
wpwrakdisplay regression sounds nasty04:40
xiangfuit's like the 'GREEN' shift to left.04:49
wpwrakhmm. that sounds a bit like the return of the gradients issue i reported a while ago. it eventually just disappeared for some reason.05:14
xiangfuyes. there is a gradient line of green points on button and blank. I will take a picture later.05:21
xiangfuwpwrak, found the problem.06:12
xiangfuwpwrak, I miss your patch on rtems.06:12
xiangfumust be something wrong with the build script file06:18
GitHub199[scripts] xiangfu pushed 1 new commit to master: http://git.io/1b5-5Q06:40
GitHub199[scripts/master] fix the quilt apply patch problem. pop -a -f before push - Xiangfu Liu06:40
xiangfulet's wait next build. try to find out the 13.4 display regression. :)06:47
Action: xiangfu have the usb-midi working great. try to write a demo on ICON :)06:48
Fallenou00:19 < wolfspraul> Fallenou: R4, we removed the 'C' < Ah yes I forgot that :)08:42
Fallenoumorning08:42
xiangfumorning08:50
GitHub106[milkymist] xiangfu force-pushed master from 8cbdb53 to 12368fb: https://github.com/milkymist/milkymist/commits/master08:56
GitHub106[milkymist/master] Makefile: add sdk-install, tools-install - Xiangfu Liu08:56
xiangfufix my last commit. that wrong with 'make tools' since there is a folder named 'tools'08:57
GitHub63[milkymist] xiangfu force-pushed hid from b05e7e7 to c3566b0: https://github.com/milkymist/milkymist/commits/hid09:02
GitHub63[milkymist/hid] pass ep_status to ep process - Xiangfu Liu09:02
GitHub63[milkymist/hid] softusb-input: add GET_HID_REPORT_DESCRIPTOR support - Xiangfu Liu09:02
GitHub63[milkymist/hid] cleanup code style, remove space after if - Xiangfu Liu09:02
xiangfurebase 'hid' on 'master' for easy review the patches on hid.09:03
wpwrakxiangfu: (quilt issue found) great !10:01
xiangfuwpwrak, yes. I am play the wheel.fnp now. try to write another demo for ICON :)10:03
wpwrakgreat ! the more demos, the better :)10:04
lekerneldavidc__: feel free to write code against it. no experimental software = no fun.10:36
davidc__lekernel: heh, that was the general assumption. I'm quite happy to deal with bleeding-edgeness / bughunting; I just don't want to write against something that is definitely going to be rewritten :)10:38
qi-botThe firmware build was successful, see images here: http://fidelio.qi-hardware.com/~xiangfu/build-milkymist/milkymist-firmware-20120223-0959/10:39
Action: xiangfu added a new rtems diff after build: http://fidelio.qi-hardware.com/~xiangfu/build-milkymist/milkymist-firmware-20120223-0959/rtems.on.f80b3a3.diff10:42
lekerneldavidc__: there are no major rewrites planned10:51
lekernel(yet *g*)10:52
davidc__lekernel: cool. One area I'd like to develop is automatic clock-crossing-logic generation; since thats a tedious PITA10:59
lekernelhmm... everything in migen is designed for synchronous circuits11:00
lekernelalso this stuff is often architecture-dependent11:01
davidc__lekernel: yeah to both. I haven't figured out if the first can be worked around11:01
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-20120223-1139/12:20
wpwrakfunny. i had "wheel" running for about a day and the two differ by about 4 frames. that twice the exactly some calculations, although with a slight difference in initial angle.13:27
lekernelhaha! I found rejon's famous conference. http://www.ttivanguard.com/conference/2011/nextgens11.html13:40
rejongood work13:50
lekernelgood work getting in there :) but I wonder what sort of outcome can be expected from such events13:58
rejonIts true, but its so easy just to stay in free software world too, so I wonder what happens there?13:58
rejonIts always hard work getting to know new people and scene13:58
rejonreminds me I need to followup with them13:58
rejonI'm working on some more events, but will refrain talking about until more real ;)13:59
lekernelyes sure, always good to go and see what other people are doing14:00
GitHub44[scripts] xiangfu pushed 1 new commit to master: http://git.io/DmlEnA14:25
GitHub44[scripts/master] checkout to master branch since we have branch build now - Xiangfu Liu14:25
Action: xiangfu will work on the 'green gradient line' issue tomorrow. 14:30
qi-botThe firmware build was successful, see images here: http://fidelio.qi-hardware.com/~xiangfu/build-milkymist/milkymist-firmware-20120223-1511/15:51
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-20120223-1651/17:32
qi-botThe firmware build was successful, see images here: http://fidelio.qi-hardware.com/~xiangfu/build-milkymist/milkymist-firmware-20120223-1842/19:22
GitHub61[milkymist-ng] sbourdeauducq pushed 1 new commit to master: http://git.io/6T5_ng20:27
GitHub61[milkymist-ng/master] dfii: new design - Sebastien Bourdeauducq20:27
cozyAnyone trying to do any 3D projection (mapping) stuff with Milkymist Ones?20:29
Action: lekernel is successfully operating the DDR command signals at 166MHz20:48
lekernelseems the read path is working too, but there's some data alignment issue20:50
lekernel(at 333Mt/s/pin)20:50
GitHub186[milkymist-ng] sbourdeauducq pushed 1 new commit to master: http://git.io/c1j21w21:06
GitHub186[milkymist-ng/master] dfii: set data mask - Sebastien Bourdeauducq21:06
lekernelyay! almost works. and data doesn't get corrupted.21:10
kristianpaulbravo ! :)21:11
lekernelit reads/writes the same data, only at the wrong places, and with inserted zero's in some cycles...21:12
lekernelsounds like relatively easy bugs ...21:12
lekernelhm probably the serdes are acting up21:29
lekernelor the bufpll21:29
lekernelinstead of writing 0001020304050607f0f1f2f3f4f5f6f721:32
lekernelI get 04050607f0f1f2f3f4f5f6f7f4f5f6f7 and a similar bug in the read path21:33
lekernel(I can deduce what got actually written by reordering the bursts)21:34
--- Fri Feb 24 201200:00

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