#milkymist IRC log for Thursday, 2011-08-04

awxiangfu, do you know what it wants to do after showed up 'Bitstream length: 1484404' while flashing m1?01:33
xiangfuaw, let me check.01:36
xiangfuaw, from the reflash script file:pld load ${FJMEM}01:39
xiangfuinitbus fjmem opcode=00001001:39
xiangfufrequency 600000001:39
xiangfuBitstream length: 1484404 means load fjmem done.01:39
xiangfunext should be 'frequency 6000000'01:39
xiangfushould be output "Setting TCK frequency to 6000000 Hz"01:40
awxiangfu, http://pastebin.com/nGUxmFA601:40
awso 1484404 is done means load fjmem finished? so then what's the next step script trys to do? will 'frequency 6000000' mean UrJtag starting to set somethings into fpga?01:44
xiangfusorry. next should be "initbus fjmem opcode=000010"01:45
awwhat does this meaning? fpga trys to do somethings you guess?01:48
xiangfuaw, I would advice change on Jtag board. try reflash again.01:49
xiangfuusing another jtag board, try reflash again.01:49
xiangfuinitbus mean try to initializes the external memory bus, then we can access the flash chip01:51
xiangfuaw, each m1 you using different jtag board, right? then we can try to using one working jtag board on this m1. to detec if it's jtag board problem.01:52
awxiangfu, yes, each m1 with its won jtag board. mmm...good idea01:54
kristianpaulalso i wondered the dmesg output after connect that board, but yes better try another01:55
awxiangfu, your this idea makes me happy now. ;-)01:56
kristianpaul:)01:57
awalthough I don't change another jtag board to 0x32, but now it's flashing... I just tried to plug out and replug again though. ;-)01:57
awox32: hmm...d2/d3 dimly lit after flashed. with BEN shorter cable. :(01:59
awlet's change a new jtag board.01:59
awafter changed to other jtag rc1 board...still stops at 'Bitstream length: 1484404'02:03
kristianpaulhum, but jtag serial cabled passed testing all okay it seems, so dunno if can be a jtag-serial board  issue02:03
wolfspraulsounds like the problem is with the flash chip?02:03
xiangfumaybe02:04
awused a new jtag rc2 still stops at there.02:04
awokay, so you guys think it's probably a flash chip problem related?02:04
awnot bad. at least narrow the area to find possible cause.02:05
awtks02:05
wolfspraulwell02:06
wolfspraulwe are guessing02:06
wolfspraulbut at least 0x32 never successfully flashed and booted before02:06
wolfspraulunlike for example 0x3402:07
wolfspraulso for 0x32, yeah, maybe it's just a flash soldering problem?02:07
awyes, never successfully flashed, but at least today became only was flashing... once02:07
wolfspraulyes, but then it didn't boot (configure)02:08
kristianpaulyeah, probably flash something around or flash chip it self02:08
awpossible . I'll see it. just asked here to know more steps about scripts. ;-)02:08
kristianpaulbut from the pastebin looks like jtag lost connection with fjmem...02:09
kristianpaulxiangfu: hi,02:16
xiangfukristianpaul, hi02:16
kristianpaulxiangfu: you did some dd from /dev/zero to the first blocks of the 512Mb flash you sent me, to make it recognice by mm bios?02:17
kristianpaulor formated? (mkfs.fvat)02:18
kristianpauli made the mistake of format the card now it is recogniced but i hit a bug with the fat library implemented in bios...02:18
kristianpaulin wich it dint recognice files on the memory..02:19
xiangfukristianpaul, ? I lost. what file I sent to you?02:19
kristianpaulcard02:19
kristianpaul512mb and 1Gb remenber02:19
kristianpaul?02:19
xiangfuoh. memcard. 512MB.02:20
kristianpaulsorry my english is not very good today..02:20
kristianpaulyeah02:20
xiangfuthe BIOS only know fat16.02:20
kristianpaulit was working well, actually i  put boot.bin cmdline.txt and initrd.bin on it02:20
kristianpaulyeah, but happen that i formated the card, in fat16, but the error is not about fat32 not supported.02:21
xiangfucheck out this file: https://github.com/milkymist/scripts/blob/master/format_memcard.sh02:21
kristianpaulahh ;)02:21
xiangfuwhich is used on RC3 memcards.02:21
kristianpaulbut those are brand new cards isnt=02:21
kristianpaul?02:21
kristianpaulcause the 1Gb you sent me was almost all zero, expect the 512Mb that seems to have a remaning metadata, but anyway02:22
kristianpaulthe point is you already gave me a good link to see :)02:22
kristianpaulmkdosfs """"02:23
kristianpaulahhh02:23
kristianpauli see now02:23
kristianpaulthanks xiangfu02:23
xiangfunope. :)02:23
kristianpaulhum this is script is new, i see02:24
kristianpaulha very hackisght the tr :)02:26
kristianpaulbut kind of you and lekernel ;)02:26
kristianpauli can try linux again this weekend02:26
kristianpaulok..02:26
xiangfukristianpaul, which linux image you using?02:27
xiangfumaybe you can try image here: http://fidelio.qi-hardware.com/~xiangfu/compile-log/openwrt-milkymist.minimal-08032011-0953/02:27
kristianpauloh, ok02:28
kristianpauli'll go for some tea and try :)02:28
kristianpaulfrom commit 5a05105aee18d4abb6cd8597c4e0db254b1dde4602:28
xiangfucan we build the 'initrd.bin' inside the 'boot.bin'?02:29
xiangfulars said we don't need cmdline.txt anymore. if we build 'initrd.bin' inside the 'boot.bin' then we only needs one file02:29
xiangfualso we need change the 'openwrt-lm32-root.ext4' to 4M02:30
kristianpauli like this openwrt-milkymit, is so small :)02:32
kristianpaulmay be i can try to package osgps ;)02:33
xiangfuyes. I have enable the all package as modules when build openwrt-milkymist. but it's not working now. maybe miss something.02:35
kristianpaulso it can void ipkg stuff and build the package directly to rootfs?02:39
xiangfuthat needs IMAGE BUILDER in openwrt. I only tried once. it not works well.02:45
xiangfuit's like you already build all packages. but want select 10 or 20 of them in rootfs, then using IMAGE BUILDER. that can fast build rootfs.02:45
xiangfuit's just unpack the package then create a rootfs. no compile I remember.02:46
kristianpaulstill formating, i guess take a while?..02:47
xiangfuno. should be very fast. make the "DEV=sdb" correct02:50
xiangfudont' format your harddisk.02:50
xiangfumake the sure the 'DEV=sdb' is your memcard02:50
xiangfuin your computer02:50
kristianpaulsure, i fixed DEV02:52
kristianpaulcause sdb is a _important_ parition on my computer ;)02:52
xiangfuyes. :)02:54
kristianpaulactually get stucj in fdisk02:58
kristianpaul13971 pts/35   R+     0:15 fdisk -b 512 /dev/sdc02:58
kristianpaulhum that FDISK_CMD_FILE..03:07
kristianpaulfollwing manually the fdisk automation but something not right03:07
xiangfuhmm..03:09
kristianpaulignore, if work for you okay, now i should find the way to make it work in my system :)03:14
kristianpaulbut imho expect command is a better way to automate input to others commands03:16
xiangfusorry. what you mean?03:19
kristianpaulbut a bit complex tought.. https://secure.wikimedia.org/wikipedia/en/wiki/Expect03:19
xiangfukristianpaul, oh. command 'expect' , I got it.03:28
xiangfuwill try that next time.03:35
aw0x72: actually is both Y1 Xtal no signal and L1's pad not soldering well. Replaced a Y1 then audio codec circuit work well. ;-)07:10
wolfspraulnice. another one fixed :-)07:10
AlarmI have installed and unpacked Xilinx_ISE_DS_Lin_13.2_O.61xd.0.0. Is there a command to run?10:54
Alarmto complete the installation?10:55
larscunpacking should be enough11:12
kristianpaulAlarm: source /opt/Xilinx/13.2/ISE_DS/settings32.sh11:15
kristianpaulAlarm: or whatver you install dir11:15
kristianpaulyou can run "ise" command to make sure11:16
kristianpaulbut dont bother to run that thign as from milkymist/boards/milkymist-one/flas a make will be enought i think11:17
kristianpaulalmost forgot, http://lekernel.net/blog/2010/05/xilinx-ise-12-phoning-home/11:25
GitHub117[extras-m1] yizhangsh pushed 3 new commits to master: https://github.com/milkymist/extras-m1/compare/751a2ff...236db4714:55
GitHub117[extras-m1/master] added bleeds, adjust M1 logo color - Yi Zhang14:55
GitHub117[extras-m1/master] Merge branch 'master' of https://github.com/milkymist/extras-m1 - Yi Zhang14:55
GitHub117[extras-m1/master] adjust color to colse to Pantone 2935 C - Yi Zhang14:55
AlarmI installed the Webpack ise in paltest but make does not work because missing xst16:21
kristianpaulAlarm: do you ran source command ober the right settings.sh file?16:33
kristianpaulalso can you please use pastebin.com or such so we can see clearly error you got?16:34
Alarmkristianpaul: http://pastebin.com/YLZWBmT216:36
kristianpaulAlarm: where did you installed Xilinx ISE?16:37
kristianpaulfull dir please16:38
kristianpaulsource /opt/Xilinx/13.2/ISE_DS/settings32.sh16:38
kristianpaulis the command in my system i ran in order to set right env variables for xst to run16:38
kristianpaulso you should run something similar in your system,16:39
Alarmkristianpaul: /opt/Xilinx/13.2/ISE_DS/16:39
kristianpaulsame16:39
kristianpaulrun that then16:39
kristianpaulnot of my fully care, but you should try to avoid xst as run, we never know what we can get from it ;)16:40
kristianpaulbut ignore my last comment for now :)16:40
kristianpaulrun xst as root i mean, sorry16:45
GitHub141[linux-milkymist] mwalle pushed 1 new commit to master: http://bit.ly/plnR5A21:56
GitHub141[linux-milkymist/master] lm32: unify asm-offsets.[ch] - Michael Walle21:56
--- Fri Aug 5 201100:00

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