#milkymist IRC log for Monday, 2011-03-21

FallenouI really don't know anything about it sorry00:00
FallenouI know better the driver side than the "user land" side lol00:00
kristianpaulheeh ok00:00
kristianpaulFallenou: did i forgot add this driver to table entry?00:09
Fallenouyou tell me :p00:10
kristianpauli think i did00:10
kristianpaul*forgot*00:10
kristianpauloh yes..00:11
kristianpaulhttp://www.rtems.com/ml/rtems-users/2004/december/msg00038.html00:16
kristianpaulmorning aw_00:31
aw_morning kristianpaul00:32
Fallenoui'm off to bed00:32
Fallenougn800:32
Fallenoumorning aw_ :)00:32
kristianpaulhmm00:32
kristianpaulFallenou: console.c is very related to his name i dunno if it can help me00:33
Fallenouoh and thanks for the synthesis log kristianpaul I will have a look00:33
kristianpaulit i want cat to /dev/console1 ..00:33
kristianpauls/it/if00:33
kristianpaulok Fallenou00:33
FallenouI don't understand00:33
kristianpauli want my driver be /dev/console100:34
Fallenouthen just modify the line in console.c :)00:34
Fallenouthat gives the name00:34
kristianpaulsure00:34
kristianpaulbut00:34
kristianpaulnot sure if the functionallity i want00:34
kristianpauli need read more00:34
FallenouI think it's what you want00:35
Fallenoujust modify line 165 of console.c00:35
kristianpaulthats already done00:35
kristianpaulok00:35
kristianpauli'll try as it is right now00:35
Fallenouyes don't ask youself too much question before even trying ;)00:35
kristianpaulheh00:36
kristianpaulbad habit00:36
Fallenoutry it, and ask yourself if it does not work :p00:36
kristianpaullol00:36
Fallenouactually it's good to think before trying, but not to think too much sometimes :p00:36
kristianpaulmore logical and productive way indeed00:36
Fallenouyou try, it fails you think you try again it fails etc00:36
Fallenoui'm off ! :)00:37
Fallenougood luck00:37
kristianpaulnice sleep00:37
kristianpauln800:37
Fallenouthanks00:38
kristianpaulaw_: very interesting your mail  about DV7125 and ground pins00:55
aw_kristianpaul, yes, me too00:55
kristianpaulIf wonder if you are going to test with lekernel help the "just ground the video output signals" theory.00:56
aw_I'll try to grep milkymist sources. surely I will.00:56
aw_the way is that trying to let no high active frequency comes from fpga.  i think that currently we are not very sure that if those parallel digital signal will influence audio quality.00:58
aw_so if stop digital video sources then can monitor if playing GUI the loud noise will be taken less. :-)00:59
aw_but i don't know every piece of higher s/w even driver needs to modify somewhere or just like 'pull down' all video out sources.01:01
aw_kristianpaul, thanks that you did the SoC Diagram. :-)01:04
kristianpaulAcording to lekernel reply on the ML it is just the HDL/Verilog part wich need to be modified, well i'm aware of way of disable modules of Milkymist wich is using the ENABLE_VIDEOIN variable, may be xianfgu can give, actually i asked that on the ML but not get a clear answers if this the easy way to test from last lekernel reply01:04
kristianpaulaw_: ah well it was already made by sebaestien, i just updated it a bit..01:05
aw_kristianpaul, addresses are that you added?01:06
kristianpaulaw_: yes i added01:06
aw_kristianpaul, ;-)01:06
kristianpaulalso some bus clarification from *my* consideration01:07
kristianpaulsome extra labels and removed old modules..01:07
kristianpaulclean up :-)01:07
aw_ah..great.01:08
aw_kristianpaul, I was thought that to let 'VGA_PSAVE_N' pin low to power down ADV7125. But he tried to pull all R/G/B sources low which is better idea than mine to totally simulate no high frequency.01:26
aw_kristianpaul, the AC97 wires and RGB routes they are close together came from fgpa area even there's one inner gnd layer between both. so do this experiment i think we can know if video sources influences audio. ;-)01:29
Action: kristianpaul back01:39
kristianpaulsorry was at dinner01:39
kristianpaulwtf happened with tmplab :(03:04
rohhm?03:05
kristianpaulit is a shop now?03:06
kristianpaulirc channel is dead?03:06
kristianpaulah, well new webpage is not a shop but wow is looks like03:08
kristianpaulor is me that thinks every where i saw an arduino + wires + motors is because sellling03:09
lekernelaw_: hi08:54
aw_lekernel, good morning08:55
aw_copied milkymist sources already08:56
lekernelok, do you have ise 13.108:56
lekernel?08:56
aw_mine is 12.3 not 13.108:57
aw_is it okay?08:57
xiangfume too. I have the 12.3. is the version very important?08:57
xiangfuI am downloading the 13.1 now08:57
aw_so we need 13.1?08:57
lekernelyes08:58
lekernel12.x doesn't meet timing and is very slow08:58
aw_okay..second...it may quite long time...08:58
lekernelah, and have you installed urjtag already?09:00
aw_yes, i had installed urjtag.09:01
Action: xiangfu download 13.1, need ~5h 09:03
aw_i am downloading..not sure how long it will take on mine.09:04
aw_lekernel, i think that it will take 3 ~ 4 hours to finish download. i got 10% only.09:40
aw_do i have others needed to install or download?09:40
lekernelyou may want to check that urjtag works and you can load (not flash) a bitstream with it10:04
aw_do you mean 'pld load' and such 'jtag -n batchfile'?10:32
aw_if yes, i think that i can, just don't know how to build *. bin, *fbi, *.fpg.10:34
lekernelyes, pld load10:34
lekernelyou can try with http://www.milkymist.org/msd/fjmem.bit.bz210:34
lekernel(use bunzip2 before)10:35
aw_xiangfu taught me used (bunzip) not bunzip2...10:38
xiangfuaw_: bunzip210:39
aw_xiangfu, huh? but last time you used 'bunzip', they are different?10:40
xiangfuaw_: sorry. maybe typo.10:41
aw_bunzip2 is for *.bz2? bunzip for *.bz?10:42
lekernelafaik there's no "bunzip" command, but anyway10:44
lekerneljust decompress that file :)10:44
aw_um..okay.. i see this. :-) http://www.linuxheadquarters.com/howto/basic/bz2.shtml10:46
kristianpaulmorning10:57
lekernelaw_: so are you able to 'pld load' fjmem.bit?11:01
aw_yes, able before..I need to practice more times. :-)11:03
aw_dinner time. I'll back.11:16
CIA-43antares: Sebastien Bourdeauducq master * rb9c3131 / (4 files): place: better constraints system (wip) - http://bit.ly/dIwdvn12:00
kristianpaulFallenou: /opt/rtems-4.11/lm32-rtems4.11/milkymist//lib/include/rtems/confdefs.h:886:7: error: UART1_DRIVER_TABLE_ENTRY no se declaró aquí (no en una función)12:54
kristianpaulha, this was not copied to the build folder13:01
lekernelaw: how is it going?13:03
Fallenoukristianpaul: try putting the driver table entry in the .h of the driver13:07
Fallenoulike it's done for the milkymist_gpio driver13:07
kristianpaulFallenou: i already did13:07
kristianpaulwait a mi13:07
kristianpauln13:07
Fallenouthen don't forget to add the header :)13:07
Fallenouin Makefile.am13:08
kristianpaulahh!13:08
Fallenou"include_bsp_HEADERS"13:08
awlekernel, just flashed soc-1.0RC2.fpg, bios-1.0RC2.bin, flickernoise-0.2.fbi13:08
Action: kristianpaul forget fullness13:08
kristianpaulsecond13:08
lekernelaw: you didn't need to, but good.13:08
awbut ISE download is very slow though..13:08
lekernelno surprise here :)13:09
awso i think that tomorrow I'll follow you to modify source then build bitstream file...seems I can not done this tonight.13:10
aw60% done only13:10
lekernelok13:10
lekernelany news on the other rc3 board improvements?13:11
lekerneldid you test and select a fuse to protect against overvoltage?13:11
lekerneland zener?13:11
awthose final two are not yet though..13:12
lekernelalso I should be receiving the wolfson audio codec samples tomorrow13:12
awduring sourcing bat I'll order fuse again.13:13
lekernelaw: just order samples of fuse, right?13:13
lekernelwe need to test it before production13:13
awno..the rest parts for rc3 batch I'll order all except fuse and zener I need to fine tune again..13:14
awsurely test them before rc3. just meanwhile ordering the rest.13:15
awalso I've not decided the parts for video input protection.13:16
lekernelI'd say just use 4 diodes... there are dual diodes in SOT-23, so you'd need only 2 extra packages per video channel13:18
awfor this, you 'd like zener diode or normal diode? actually I surveyed for a while. and i read this http://www.littelfuse.com/application/lcd-projector/svideo.html13:25
awthere's 'Due to the analog nature of the video signals, uni-polar TVS diode arrays are not recommended as the may clip the negative part of the video signal. ' mentioned13:26
lekernelnormal diode13:26
lekernelsure, but with two normal diodes in series, it'd only clip at about -1.4V13:26
awso i was wrong on using TVS thoughts.13:26
lekernelwhich a normal video signal isn't supposed to reach afaik13:26
aweven used normal diodes, you got correctly.13:27
awso the bias for supplies I'd like to use our +3.3V bias to clamp. how do you think?13:27
lekernelor yeah use those littelfuse components13:27
awright. so the littlefuse is still good for us. I just want to know that we add with good reasons and surveys more. :-)13:28
lekernelmaybe just take the littelfuse thingy13:28
lekernelit's made for that :)13:29
awsurely I can directly use the littlefuse proposed.13:29
lekernelless problems13:29
lekernelyup. i'm for the littelfuse integrated protection13:29
lekernellet's keep things simple13:29
awum...that one is also for IEC 61000-4-2 standard though. :-)13:30
awwell..sounds you'd like this although i surveyed another...well..we done this, just pick this. :-)13:31
lekernelyup. just order some littelfuse samples and test them13:32
lekernelso you need to sample: 1. littelfuse protections 2. input fuse 3. ~5V zener13:33
lekernelkeep me tuned about those13:34
lekernelmeanwhile i'll try the other audio codec i'll receive tomorrow13:34
awyes, I'll fine test them again.13:34
lekernelbbl13:34
awgood13:34
awyup..I'll keep downloading...but offline irc.  bbl13:35
kristianpaulargg same error14:20
kristianpauland i already ran automake also after modify Makefil.a14:20
kristianpaulm14:20
Fallenoudid you run the bootstrap ? (don't know fi it's necessary but maybe ...)14:22
kristianpaulyeap...14:22
kristianpaulwith -c then -p the no parameters14:22
kristianpauli also erased all in the build dir and confgiure... again14:22
Fallenouoh14:23
Fallenoudunno then14:23
kristianpauli could do a _full_ bootstrap14:24
Fallenouno it's really not the problem14:25
kristianpaulhhm14:25
Fallenoucan you paste again the exact error ?14:25
kristianpaulsure14:25
kristianpaulhttp://paste.debian.net/111458/14:26
kristianpaulxiangfu: fyi i'm trying to add suport to the second uart port in rtems14:28
Fallenoudid you check what's in the file ?14:28
Fallenouin confdefs.h14:28
kristianpaulyes14:28
Fallenoucan you paste the file too ?14:28
kristianpaulsip14:29
kristianpaulhttp://paste.debian.net/111459/14:29
kristianpaulOkay, uart1.h seems to be where it should be http://paste.debian.net/111460/14:32
Fallenouhumm can you paste your application code ?14:32
Fallenouthat declares the driver table14:32
kristianpaulyeah14:33
kristianpaulhttp://paste.debian.net/111461/ line 25914:34
kristianpaulhmm14:34
kristianpauli think i missed an include file14:34
Fallenouyes maybe include the uart1.h14:35
Fallenousomething like #include <bsp/uart1.h>14:36
kristianpaulyeah already did but no, same error14:38
Fallenouhumm I don't really understand14:40
Fallenoumaybe try declaring it on several lines with , \14:41
Fallenoudunno14:41
Fallenouyou can try to compile with -E14:41
Fallenouto just run the pre processor14:41
kristianpaulokay14:41
Fallenouand this can allow you to check what's generated14:41
Fallenouand to spot an obious syntax error14:41
Fallenoudue to defines14:42
kristianpaulthats a CFLAG ?14:43
kristianpauli mean i edit Makefile and add it to CFLAGS i guess14:44
Fallenoujust copy paste the command that shows up when you tye "make"14:45
Fallenouand add -E14:45
Fallenoueasy14:46
Fallenoulm32-rtems4.11-gcc -O9 -Wall -mbarrel-shift-enabled -mmultiply-enabled -mdivide-enabled -msign-extend-enabled -fsingle-precision-constant -I/opt/rtems-4.11/lm32-rtems4.11/milkymist//lib/include -c -o obj/main.o main.c14:46
kristianpaulah ok14:46
Fallenoubut just put -E14:46
Fallenouand then vim obj/main.o14:47
kristianpaulther is no obj/main.o14:48
kristianpauli apend -E to the end14:48
kristianpaullm32-rtems4.11-gcc -O9 -Wall -mbarrel-shift-enabled -mmultiply-enabled -mdivide-enabled -msign-extend-enabled -fsingle-precision-constant -I/opt/rtems-4.11/lm32-rtems4.11/milkymist//lib/include -c -o obj/main.o main.c -E14:48
Fallenoutry to put it before14:49
Fallenoulike -E -c -o14:49
kristianpaulok14:49
kristianpaulnow works14:49
kristianpaulwhat i look for ad main.o?14:51
kristianpaulno uart found..14:51
kristianpaulha14:53
kristianpaulfor line 39 (#include <bsp/uart1.h> ) jsut a comment # 39 "main.c" 2 no more14:54
kristianpaulhttp://kristianpaul.org/~paul/tmp/main.o14:55
kristianpaulah wait14:59
kristianpaulhttp://paste.debian.net/111468/15:00
kristianpauldamn what stupid typo did i this time..15:01
lekernelpff, farnell asking me to complete a form on dead trees about my company status so I could place orders... the joys of germany...15:46
lekernelnot to mention that my German credit card wouldn't work in their system (I wonder why I took one, really) and that I had to use my french mastercard15:46
lekernelstupid15:47
Fallenoulol16:05
lekerneland it's a 25 euro order. i'm sure the salaries of the people they paid to bother me with those stupid inquiries already offsets their markup by a pretty wide margin16:08
kristianpaulhahah16:10
CIA-43milkymist: Sebastien Bourdeauducq master * r079aeab / (3 files in 2 dirs): inline irq helper functions (Michael Walle) - http://bit.ly/fGVd8816:32
CIA-43milkymist: Sebastien Bourdeauducq master * r492c9ac / (4 files): doc: update overview paper - http://bit.ly/eukE1i16:32
CIA-43milkymist: Sebastien Bourdeauducq master * r107baea / (3 files): Updated SoC block diagram (Cristian Paul Peñaranda Rojas) - http://bit.ly/hNP6My16:32
lekernelhttp://www.milkymist.org/socdoc/paper_overview.pdf16:41
Fallenounice :)16:46
kristianpaul:D16:54
CIA-43antares: Sebastien Bourdeauducq master * rb6228de / (3 files): place: fully unconstrained initial placer working - http://bit.ly/g4Fhsg19:06
Fallenouoh19:06
Fallenouso the placer works now ? :)19:07
lekernelnot completely19:07
mumptaihi19:59
lekernelhi20:17
CIA-43antares: Sebastien Bourdeauducq master * r8512e1b / CMakeLists.txt : Fix mtwist build with LLVM - http://bit.ly/fSJyrI22:04
CIA-43antares: Sebastien Bourdeauducq master * r742996d / (3 files): place: infer relative placement constraints - http://bit.ly/gpt0JU22:04
Fallenoulekernel: oh you are using llvm to build antares and llhdl22:23
kristianpaulFallenou: missing gcc? ;-)22:24
Fallenouno I don't care22:24
Fallenouas long as it works :)22:24
FallenouI'm not a gcc fanatic22:24
kristianpaulWell,  I do!,llvm support is broken on debian mips... :-(22:24
Fallenoubut I never used llvm22:24
Fallenouoh :/22:25
kristianpaulyeah :-(22:25
kristianpaulanyway..22:27
kristianpaulonce i have a good reason, like programing my mm1, i'll dig on the internals of this issue22:28
kristianpaulphew http://paste.debian.net/111535/22:39
kristianpaulnow multiply by 12 and add to CSR22:39
lekernelwhat's that?22:41
kristianpaulpart of namuru correlator, a free implementation of a GP2021 chip22:43
kristianpaulgps receiver from 90's were really interesting to hack :-)22:44
CIA-43antares: Sebastien Bourdeauducq master * ra6114ae / (antares-place/constraints.c antares-place/initial.c): place: bugfixes - https://github.com/lekernel/antares/commit/a6114ae26394338d5879d3d48a99c47ced04a97722:58
Fallenouoh a correlator :)23:06
Fallenoukristianpaul: do you know what they mean by "chips" in a gps correlator ?23:06
FallenouI was told "the more chips you have the better !"23:06
Fallenoubut my teacher couldn't tell me what it is23:06
kristianpaulFallenou: chip = bit23:31
kristianpaulsure 1023 chips are something meanfull in order to get navigation data23:32
Fallenoubit of what ?23:32
kristianpaul?23:34
kristianpaul1023 chips/bits = C/A code23:34
kristianpaulC/A = coarse acquisition code23:34
Fallenouhum hum ok23:36
kristianpaulbut there is actuallt a GPS navigation data sctructure made of chips23:36
kristianpaullet me find a pic to make it clear23:36
kristianpaulthat hum dind sound good to me :-)23:36
kristianpauldint*23:36
Fallenouwell it's not very clear to me sorry :p23:38
FallenouIIRC the correlator is to increase the SNR , right ?23:38
Fallenou(signal / noise ratio)23:40
Fallenouhttp://www.datasheetdir.com/GP2021+GPS23:41
Fallenouohoh23:41
kristianpaulFallenou: zarlink website have more accutate informacion23:41
Fallenouhum ok23:42
kristianpaulcorrelator basically from my understanding allo to track the PRN codes in the raw data coming from the RF front end23:42
Fallenouit's digital analysis, right ?23:43
kristianpaulalso it should allow anticipate changes in the signal so you can track it no matter if you're movin or the satellite (actually this always moves)23:43
kristianpaulthere is also some DSP-like stuff on the software side23:44
kristianpaulthere are some algorythm about that,23:44
Fallenouseems nice23:44
kristianpauli dint dig on that part yet23:44
kristianpaul_but_ one of the obstacles for GPS signaling is that you should adapt to different enviroments23:45
Fallenouwould it be possible to implement a correlator on fpga?23:45
kristianpaulit is already23:45
kristianpaulnamuru project is about that23:45
Fallenouoh :)23:45
kristianpaulit is proven to work with a zarlink frontend and a niosII/Altera SoC23:46
Fallenouvery nice !23:48
kristianpaulnow i hope with a SiGE front end and the Milkymsit SoC ;-)23:49
kristianpauli have some dubts with clock but i think if the verilog code is not using some special altera libraries it can be migrated23:50
kristianpaulbut all i said is bullshit for now :-)23:50
Fallenoujust try23:51
Fallenouand you will know23:51
Fallenouif it works ;)23:51
kristianpaulyeah23:52
kristianpauli'm following you recomendations _really_23:52
--- Tue Mar 22 201100:00

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