#milkymist IRC log for Thursday, 2013-02-21

larschm tried to replace the maxim parts with adi parts, but the pin compatible adi lvds receiver doesn't have built-in termination resistores :/11:42
lekernelwhich is good, since you don't want them11:46
lekernelnote that I picked the Maxim part which doesn't have those resistors11:46
lekernelhow fast do the ADI parts go?11:47
lekernelgive use the ADI sales pitch :)11:48
larscyea, just noticed that you use the MAX9121 and not the MAX912211:50
lekernelMAX9121 is a bit faster than ADN4666 (250MHz vs. 200MHz) which is why I'm using it11:50
larscthe ADI part is only specified for 200MHz instead of 250Mhz... exactly11:50
larscADN4668 is the pin compatible version11:51
lekernelI find it a bit frustrating that all those discrete parts are under 1Gbps, while GPUs happily run DRAM at 5+Gbps/pin11:52
lekernelFPGA IOs are a little better, but not that much11:53
larscshouldn't the gpx on the newer xilinx fpgas support up to 5Gbps?12:02
lekernelwithout enough pins to drive DRAM, and using differential IO12:04
lekernelGDDR5 does 5Gbps/pin on single-ended signals12:04
lekernelIOSERDES does a pale 1.8Gbps/pin12:05
azonenberghttp://i.imgur.com/5CPFuxF.png :D18:21
azonenbergscrew you chipscope18:21
larsclooks nice18:30
larscbut the text is not exactly vertically centered18:30
azonenbergWhich text, the channel names?18:32
azonenbergor the text in the signals18:32
azonenbergthe text in the signals should be, i used pango calls to get the line height and then put it at (signal height/2 - line height/2)18:32
azonenbergi know that the channel-name list on the left is way off18:33
larscthere is one more pixel on top18:38
larscfor the text in the signals18:38
larschave I ever mentioned I suffer from ocd ;)18:39
azonenberghmm, it might just not divide evenly18:40
azonenbergif the font is an odd number of pixels high and the signal is an even18:40
lekernelazonenberg, looks nice, though the obvious question is: why not gtkwave?19:08
azonenberglekernel: I considered gtkwave19:09
azonenbergin fact my first version used it19:09
lekerneltbh this looks a bit NIH19:09
azonenbergBut its internal architecture was not amenable to arbitrary protocol decode19:09
lekernelor sigrok19:09
azonenbergthe UI is almost irrelevant19:10
azonenbergmost of the work is the back end on the capture core19:10
azonenbergThe other problem is that gtkwave is gpl'd19:12
larscyea, using pulseview would be another alternative http://sigrok.org/wiki/PulseView19:12
larscalthough it is also gpl19:12
azonenbergMy entire stack as far down as i can get is going to be BSD19:12
azonenbergSeveral times i've found very nice looking apps and been ready to start integrating them, then realized they were GPL19:14
azonenbergWhich would mean i'd have to GPL a sizable fraction of my code19:14
azonenbergand that's not an option19:14
lekernelyour code can be BSD19:15
azonenbergnot if i link it with a gpl'd application19:15
azonenbergeverything i add would have to be gpl'd19:15
larsci wish there was some nice signal rendering _library_ under a liberal license19:15
azonenberglarsc: That's what i'm doing here19:16
azonenbergall of the protocol decode and back-end stuff is a standalone library19:16
lekernel"It is a lax, permissive non-copyleft free software license, *compatible with the GNU GPL.*"19:16
azonenbergBut the combination of the two becomes GPL'd19:16
lekernelthis sounds a lesser evil than NIH19:17
azonenbergMy bigger issue is gpl'd libraries19:17
azonenbergforcing me to gpl the entire application19:17
azonenbergthat is not acceptable19:17
azonenbergmy goal is to deliver a useful product that can be used by the widest possible audience19:18
azonenbergnot to push the ideology of free software19:18
azonenbergand that means rejecting anything that's too overly politicized like the gpl19:18
azonenbergStallman himself tried to convince me otherwise but when he said "I have no sympathy for anyone who uses such a ridiculously permissive license" that was the last straw19:19
azonenbergi couldn't care less if some company forks my code and turns it into a commercial product19:19
azonenbergas long as my copy is still usable by me then i lose nothing19:19
azonenbergand the world gains something19:20
lekernelyou're preaching to the converted here, I'm not a rms fan either19:21
lekerneland my conversations with him have been similar to yours19:22
azonenbergAnd as a result, if left with no choice19:22
azonenbergand I need a library to do something in a BSD'd application19:22
azonenbergI will replace a GPL'd library with a BSD'd equivalent19:22
azonenbergeven if that means writing it from scratch19:22
azonenbergLike i said, i had been using gtkwave as a standalone executable for my first version of the LA app19:23
azonenbergbut it lacked a lot of capabilities i needed to do proper live updates and protocol decoding19:23
azonenbergand i didn't want to put all of that time into patching a GPL'd project19:23
azonenbergSince i want my viewer to be a librarized component that i can drop into a larger appliction19:25
azonenbergwithout having to gpl it19:25
mwallelarsc: your simpleImage isnt working for me at all21:33
mwalleno prints21:33
larscmwalle: which qemu?21:33
mwallebut there shouldnt be any changes21:34
larscmine is 6 month old21:34
mwallecould you do a diff? eg. git fetch origin (or sth like that), git diff master..origin/master target-lm3221:35
larsccommit 63e588521:35
mwalleok mom21:35
mwalleis that an official commit?21:36
mwalleah that is a commit from me21:36
larscthe changes in target-lm32 look trivial21:37
larscbut I'd rather expect the problem to be in hw/milkymist* anyway21:37
mwallewhats your commandline?21:40
larsc-M milkymist -nographic -kernel arch/lm32/boot/simpleImage.milkymist_one -initrd /opt/Projects/openwrt-git/bin/lm32/openwrt-lm32-root.ext4  -s -append "init=/bin/sh root=/dev/ram0 console=ttyS0"21:40
mwalleim now at 63e588521:41
mwallebut still no output21:41
mwalleqemu-system-lm32 -M milkymist -kernel /tmp/simpleImage.milkymist_one -nographic -append "console=/dev/ttyS0"21:42
larscmd5sum of the kernel?21:42
mwalle3dcf992cd157f218414e8d9d4d8f9b13  /tmp/simpleImage.milkymist_one21:42
larscah, wait I don't have the one I uploaded anymore21:43
larscuploaded a new one a3a8e3fe5d21af6aaaa4990a71efb08721:49
mwalleworks neither21:52
mwalleyou're sure you are using a binary compiled from commit 63e588521:53
larscnot 100% but 90%-ish21:54
larscgit diff shows nothing21:54
mwallelarsc: does "-serial stdio -nographic" work for you?21:54
larscqemu: could not open serial device 'stdio': Success21:55
larscwhat a great error messages...21:55
mwalleyeah ;)21:56
mwallebtw i have some vmlinux here (3.1.0) which is working21:57
mwalleat least until it turns off the early boot console21:58
mwallewhich might not be enabled in your kernel21:58
mwalledo we support earlyprintk?21:59
mwalleah forget what i said, with the right kernel command line, my kernel works22:00
mwalleah and yours too22:00
mwallelol "console=/dev/ttyS0".. i need more sleep22:01
GitHub143[qemu] mwalle pushed 1 new commit to mmu: https://github.com/mwalle/qemu/commit/99e6157e87848b1329d3c831066017415eed8d3822:08
GitHub143qemu/mmu 99e6157 Michael Walle: mmu wip #922:08
GitHub62[qemu] mwalle pushed 1000 new commits to master: https://github.com/mwalle/qemu/compare/611029910906...3f15f98dc13922:10
GitHub62qemu/master 722bfec Richard Henderson: target-s390: Convert MULTIPLY AND ADD, SUBTRACT...22:10
GitHub62qemu/master 31aa97d Richard Henderson: target-s390: Convert TEST DATA CLASS...22:10
GitHub62qemu/master 5d7fd04 Richard Henderson: target-s390: Convert FP LOAD COMPLIMENT, NEGATIVE, POSITIVE...22:10
mwallelarsc: btw your rootfs gives me "sh: can't execute 'uci_load': No such file or directory"22:13
larscthat's why I use /bin/sh for init22:16
mwallei see22:21
--- Fri Feb 22 201300:00

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