07:10.39 | *** join/#openjtag dyoung-away (n=dyoung@nslu2-linux/dyoung) |
08:19.14 | *** part/#openjtag ka6sox-away (n=marina@nslu2-linux/ka6sox) |
09:04.07 | *** join/#openjtag ka6sox (n=ka6sox@nslu2-linux/ka6sox) |
11:13.36 | *** join/#openjtag dyoung-away (n=dyoung@nslu2-linux/dyoung) |
13:35.21 | *** join/#openjtag prpplague (n=billybob@72.22.146.238) |
14:59.06 | *** join/#openjtag vmaster_ (i=vmaster@p549B5BD6.dip.t-dialin.net) |
15:14.41 | *** join/#openjtag p2-mate (n=p2@NAT.office.mind.be) |
15:34.19 | *** join/#openjtag bullet (n=bullet@57.28.78.83.cust.bluewin.ch) |
15:34.23 | *** join/#openjtag p2-mate (n=p2@NAT.office.mind.be) |
17:16.26 | [g2] | vmaster ping |
17:16.44 | *** join/#openjtag ka6sox (n=ka6sox@nslu2-linux/ka6sox) |
17:17.31 | *** join/#openjtag [g2-lap] (n=g2@cpe-066-057-008-035.nc.res.rr.com) |
17:20.20 | *** join/#openjtag ka6sox (n=ka6sox@nslu2-linux/ka6sox) |
17:48.54 | vmaster | [g2] [g2-lap]: pong |
18:29.41 | *** join/#openjtag p2-mate (n=p2@NAT.office.mind.be) |
18:58.33 | [g2-lap] | vmaster hey |
18:58.51 | [g2-lap] | feel like talking about openOCD ? |
18:59.39 | [g2-lap] | p2-mate howdy |
19:00.50 | *** mode/#openjtag [+o [g2]] by ChanServ |
19:06.20 | vmaster | [g2-lap]: yeah, sure, you got 20 minutes |
19:07.19 | [g2] | vm sure |
19:07.25 | [g2] | one sec |
19:07.29 | *** mode/#openjtag [-o [g2]] by [g2] |
19:07.46 | [g2-lap] | vmaster Ok that's better |
19:09.09 | [g2-lap] | vmaster so we're jamming on the 93xx stuff |
19:09.46 | [g2-lap] | and I'm looking at some stuff on sparkfun and there's mention of the openOCD stuff |
19:10.02 | [g2-lap] | and I'd already been thinking about JTAG stuff a litle |
19:10.05 | vmaster | oh, where? |
19:10.12 | vmaster | i mean at sparkfun's site |
19:10.21 | [g2-lap] | in the forum |
19:10.27 | [g2-lap] | you and the olimex guy |
19:10.34 | vmaster | ah, yeah |
19:11.22 | [g2-lap] | so we've got debian and an OE rootfs running on the glomation and ts72xx |
19:11.22 | *** join/#openjtag p2-mate (n=p2@NAT.office.mind.be) |
19:11.26 | [g2-lap] | wb |
19:12.16 | [g2-lap] | So I was thinking hel., we could just have a custom app that runs on the 72xx or the 9302 version from Glomation |
19:12.32 | [g2-lap] | I think it'd fit in 4MB or could be loaded from the net |
19:13.08 | [g2-lap] | we'd only need 256K for RedBoot to load the app from the net, but I thin 4MB is the smallest NOR flash these days |
19:14.12 | vmaster | yeah, at least you wouldn't save a lot if you used less NOR |
19:14.22 | [g2-lap] | so for <$100 probably <$80 we'd have a 200Mhz ARM with 8+ DIOs, Ethernet, USB, serial that could run openOCD. |
19:14.34 | [g2-lap] | (or any other JTAG app too) |
19:15.03 | [g2-lap] | this won't be as cool or as fast as the geep |
19:15.32 | [g2-lap] | but 95% of the stuff would probably xfer right over to that platform when it's ready |
19:15.46 | vmaster | mhh, have you already done any testing how fast bitbanging could go? |
19:16.14 | [g2-lap] | vmaster not yet, but it should be pretty easy to do |
19:16.53 | [g2-lap] | I think the TS already has a PC104 connector on it |
19:17.32 | [g2-lap] | so if the DIOs are too slow would might be able to make a simple board |
19:18.01 | [g2-lap] | or maybe just get a very geep that's got the processor and fpga |
19:18.09 | [g2-lap] | s/very/very early/ |
19:19.51 | [g2-lap] | vmaster give it a little though |
19:19.53 | [g2-lap] | vmaster give it a little thought |
19:19.59 | vmaster | yeah, sure, i like the idea |
19:20.12 | [g2-lap] | how big is openOCD |
19:20.32 | vmaster | well, it isn't optimized at all - currently it requires quite a lot of ram, but that can be tweaked easily |
19:20.38 | [g2-lap] | is there a cmdline interface ? |
19:20.46 | vmaster | a telnet interface and a gdb server |
19:21.04 | [g2-lap] | what's "a lot of ram" ? |
19:21.25 | vmaster | well, several mb during runtime |
19:21.44 | [g2-lap] | these boards come with 32MB RAM |
19:21.47 | vmaster | the binary is <500kb |
19:21.54 | vmaster | with debug info and not stripped |
19:22.11 | vmaster | stripped it's <200kb |
19:22.20 | [g2-lap] | 64MB probably isn't a big deal |
19:22.35 | vmaster | even 32mb should be enough |
19:22.38 | vmaster | more than enough+ |
19:23.18 | [g2-lap] | right I was thinkging we could reserve a few Meg for JTAG buffers |
19:23.24 | *** join/#openjtag p2-mate (n=p2@NAT.office.mind.be) |
19:23.33 | [g2-lap] | then we'd just copy the images into the high buffers |
19:23.47 | [g2-lap] | past the area Linux is using |
19:24.29 | [g2-lap] | I'm also wondering how realtime (or how processor locked) we can go |
19:25.03 | [g2-lap] | I think we could disable all interrupts except for maybe a timer |
19:25.13 | [g2-lap] | and get in a _tight_ loop |
19:26.49 | [g2-lap] | vmaster I'll ping you in here from time-to-time |
19:27.07 | [g2-lap] | right now I'll see how fast the DIOs are with the TS and Glomation boards |
19:27.13 | [g2-lap] | based on a loop |
19:27.39 | [g2-lap] | running from jffs2 |
19:27.50 | [g2-lap] | no ethernet on |
19:27.52 | [g2-lap] | no usb |
19:28.02 | [g2-lap] | I think just the timer and serial |
19:28.21 | [g2-lap] | and serial won't generate any ints cause I won't be typing :) |
19:28.32 | vmaster | mhh, is there a real datasheet or user manual for the ep93xx? |
19:28.43 | [g2-lap] | sure many |
19:28.55 | [g2-lap] | are you aware of the wiki we've got going ? |
19:29.16 | vmaster | no |
19:29.20 | [g2-lap] | http://ep93xx.giantshoulderinc.com:8001/ep93xx |
19:29.39 | [g2-lap] | Click on the Cirrus link and go to the 9302 versus the 9312 |
19:30.20 | [g2-lap] | The CirrusEp9312Info link |
19:30.47 | [g2-lap] | And Three-Cheers for lennert !! |
19:30.54 | [g2-lap] | dwery too! |
19:34.55 | dwery | ty! :) |
19:35.15 | vmaster | hmm... ep's ssp might work for JTAG |
19:35.30 | vmaster | that would allow really fast jtagging |
19:36.07 | vmaster | well, i gotta run - beer time |
19:40.29 | [g2-lap] | cheers |
19:43.14 | lennert | [g2-lap]: three cheers for all the ep93xx people :) |
19:43.45 | [g2-lap] | lennert yeah especially those crack hw guys! |
19:44.01 | lennert | hehe |
19:44.19 | lennert | and the hw guys that were on crack when they designed the chip |
19:44.31 | lennert | openocd on the ep93xx is interesting |
19:44.37 | lennert | want me to test the bitbanging? |
19:45.48 | [g2-lap] | lennert if you want to |
19:46.01 | dwery | [g2-lap]: see #nslu2-linux |
19:46.12 | [g2-lap] | dwery Ok there is a sec |
19:47.30 | [g2] | dwery congrats! awesome! |
19:47.43 | lennert | what, who, where? :) |
19:47.44 | dwery | :-D |
19:47.52 | lennert | dwery: you became father? :) |
19:47.58 | dwery | argh! |
19:47.59 | dwery | no! |
19:48.01 | lennert | haha |
19:48.04 | lennert | then what? :) |
19:48.16 | dwery | just compiled the madwifi driver under arm le :) |
19:48.20 | lennert | oh :-) |
19:48.20 | lennert | cool |
21:58.40 | *** join/#openjtag p2-mate (n=p2@NAT.office.mind.be) |
22:50.05 | *** join/#openjtag p2-mate (n=p2@NAT.office.mind.be) |
23:04.31 | *** join/#openjtag p2-mate_ (n=p2@NAT.office.mind.be) |
23:05.37 | lennert | so, [g2] prodded me to do some gpio timings on the ep93xx |
23:06.20 | lennert | the numbers are very good |
23:07.20 | lennert | we can do about 10.5-odd million gpio reads per second |
23:07.28 | lennert | and 15.8-odd million gpio writes |
23:07.44 | [g2-lap] | vmaster |
23:07.58 | lennert | (at 200 MHz core clock speed) |
23:08.52 | vmaster | oh, nice |
23:09.40 | [g2-lap] | and since the boards have usb, ethernet, serial on board we could have some options to play with |
23:10.11 | [g2-lap] | the boards are in the $65-$85 range from glomation |
23:10.40 | lennert | vmaster: does openocd run _on_ arm? :) |
23:11.39 | [g2-lap] | btw lennert thx for looking into that and getting the fluke out |
23:11.52 | [g2-lap] | that sounds funny |
23:11.58 | lennert | haha |
23:12.00 | [g2-lap] | lennert get the fluke out! |
23:12.11 | vmaster | lennert: never tried on anything but x86 pc linux |
23:12.17 | vmaster | in theory it should |
23:12.19 | lennert | my fluke does 100 kHz max according to the docs |
23:13.18 | lennert | if i divide the clock by 1000 i get 16.25 kHz, if i divide it by 100 it's off the scale |
23:13.54 | lennert | so in theory it should be able to do 8 MHz |