| 00:06.26 | ali_as | Nite all. |
| 04:15.21 | *** join/#openjtag jkilb_ (n=jkilb@p5B2093CD.dip0.t-ipconnect.de) |
| 05:59.46 | *** join/#openjtag drath_ (i=vmaster@p5B07DF4E.dip.t-dialin.net) |
| 06:24.45 | *** join/#openjtag toi (n=pleemans@d5152D3B4.access.telenet.be) |
| 09:57.35 | *** join/#openjtag ali_as (n=as@ambix.plus.com) |
| 11:09.27 | *** join/#openjtag key2 (n=key2@193.33.36.20) |
| 12:27.23 | ali_as | "TAP_SD & TAP_SI are illegal end states," |
| 12:27.28 | ali_as | Hmm. |
| 13:20.55 | *** join/#openjtag prpplague (n=dave@mail.americanmicrosystems.com) |
| 15:12.30 | drath | ali_as: how did you trigger that message? |
| 15:15.10 | ali_as | I didn't, I've been reading the source. |
| 15:15.23 | ali_as | It seems a slightly shortsighted requirement. |
| 15:17.16 | ali_as | I'm not so sure about reading TDO when TCK is high, either. For some reason this has given me errors in my own dabblings. |
| 15:19.13 | ali_as | lifts his hat. |
| 15:19.18 | ali_as | I'm Alex btw. |
| 15:33.16 | drath | i'm about to get home from the office, i'll be back in ~15 minutes |
| 15:54.20 | drath | ali_as: Well, declaring SD and SI as illegal is fine with the current code, I suppose, but I didn't like the change either |
| 15:55.07 | drath | ali_as: Especially since there's nothing wrong with allowing them as endstates |
| 15:57.54 | ali_as | ST20 debug protocol involves moving a value into IR, then moving into DR and staying there, uses it like a clocked full duplex debug interface. |
| 16:01.57 | drath | ali_as: ah, nice |
| 16:02.04 | drath | ali_as: do you have any pointer to some docs? |
| 16:03.42 | drath | ali_as: hmm, i suppose ST20 is somewhat deprecated? |
| 16:05.38 | ali_as | (sorry phone) |
| 16:09.57 | ali_as | ST20 became the STi55xx series which is used in a huge number of products, mainly set top boxes, DVD players. |
| 16:10.57 | ali_as | I don't know of a document with all the info in but I've written debug programs for it for a few years. |
| 16:12.07 | ali_as | I'm currently working on EJTAG, and specifically the idea of designing a hardware interface to speed it up. I was looking for 'GEEP-B' but I'm told that project died. |
| 16:13.27 | drath | mhh, I looked at EJTAG very briefly a long time ago, but it seemed overly complicated, and difficult to implement for reasonable performance |
| 16:13.42 | drath | something like the EJTAG telling the emulator which address to serve |
| 16:14.30 | ali_as | It's nasty. If you want to read so much as one byte, you have to stall the CPU and feed it the opcodes for the instructions that tell it to read the byte and hand it over. |
| 16:15.31 | drath | well, you have to do the same with ARM JTAG, but at least you can scan in a large number of instructions in one go |
| 16:15.48 | drath | using a load multiple it's possible to read 14 registers (56 byte) in one go |
| 16:15.59 | ali_as | ARM JTAG is completely new to me. |
| 16:17.35 | ali_as | I'll be sticking around but if you want a simple st20 source to read I can find one of the original "Tomas Vlad" programs. |
| 16:20.41 | ali_as | DCC OK? |
| 16:21.33 | drath | yeah, sure |
| 16:22.56 | ali_as | I might not have my ports setup correctly for dcc. |
| 16:26.40 | ali_as | http://pastebin.com/pastebin.php?dl=m70fede06 |
| 16:29.47 | drath | thx, i'll have a look later today, lunch is calling |
| 16:29.50 | ali_as | And english explanation here http://pastebin.com/m5d6c16c6 |
| 16:29.54 | drath | mhh, dinner |
| 16:30.03 | ali_as | Ok, I have to afk anyway. |
| 17:27.39 | *** join/#openjtag sn9 (n=danielg4@gimpelevich.san-francisco.ca.us) [NETSPLIT VICTIM] |
| 17:53.54 | *** join/#openjtag dwer1 (n=dwery@81-208-60-204.ip.fastwebnet.it) |
| 18:47.49 | *** join/#openjtag dwery (n=dwery@nslu2-linux/dwery) |
| 18:52.30 | *** join/#openjtag rwhitby` (n=rwhitby@ppp240-74.static.internode.on.net) |
| 19:15.45 | *** join/#openjtag dwer1 (n=dwery@81-208-60-204.ip.fastwebnet.it) |
| 19:51.04 | gandhijee | anyone know any magic or where i can start to find magic to make openOCD work with an IXP4xx device? |
| 20:04.28 | drath | gandhijee: it should work as it is |
| 20:04.49 | drath | gandhijee: which board do you use, and which jtag interface? |
| 20:07.29 | gandhijee | i use a wiggler interface |
| 20:07.41 | gandhijee | and the design is based of the azusa design, |
| 20:07.55 | gandhijee | but its using a different flash part |
| 20:10.05 | drath | gandhijee: i had a look at what you pasted on pastebin yesterday - looks like the JTAG interface isn't working at all |
| 20:10.24 | drath | gandhijee: you'll have to verify the pinout |
| 20:10.46 | drath | gandhijee: you need TMS, TCK, TDI, TDO, TRST and SRST (system reset) |
| 20:10.50 | gandhijee | i know the iface is working cause i can program it via windows |
| 20:11.09 | gandhijee | its a wiggler device, but i think i got it to see the device on the chain now |
| 20:11.27 | drath | gandhijee: what tool did you use on windows? |
| 20:11.53 | gandhijee | OCDemon Flash Memorty Programmer |
| 20:12.26 | gandhijee | i got it to see the device now, but it said the number of dis devices in JTAG chain doesn't match config |
| 20:12.46 | drath | can you paste the output? |
| 20:13.08 | gandhijee | sure, just one sec |
| 20:13.34 | gandhijee | http://pastebin.ca/966531 |
| 20:14.03 | gandhijee | thats strange |
| 20:14.18 | drath | gandhijee: what does your .cfg look like? |
| 20:14.32 | gandhijee | http://pastebin.ca/966532 |
| 20:14.36 | gandhijee | look at that one |
| 20:14.49 | gandhijee | the first time i ran it, it didn't find the part, the 2nd time it did |
| 20:15.16 | gandhijee | http://pastebin.ca/966534 |
| 20:16.21 | drath | ok, that .cfg can't work - you lack a jtag_device line, and the target line needs to know which device in the chain (you have only one, so it's 0) it should talk to |
| 20:16.52 | gandhijee | o |
| 20:17.09 | gandhijee | i kinda pieced that cfg together from random sources |
| 20:17.14 | drath | you could have a look at doc/configs/xscale_ixp42x.cfg |
| 20:17.34 | gandhijee | i'll go do just that now |
| 20:17.36 | gandhijee | thanks man |
| 20:17.49 | drath | regarding the detection problem: could be all kinds of problems - parallel ports are often flaky |
| 20:17.56 | gandhijee | fair enough |
| 20:18.12 | drath | you're running linux? |
| 20:18.17 | gandhijee | you know how i would use openOCD to program the flash on the board? are there some type of commands i can issue |
| 20:18.18 | gandhijee | ya |
| 20:18.33 | drath | did you use direct access, or do you use ppdev? |
| 20:18.41 | gandhijee | direct |
| 20:18.54 | drath | you could try ppdev, and see if it's more stable |
| 20:19.15 | gandhijee | ok |
| 20:19.25 | drath | it's an additional configure option, and you have to change the parport_port to 0 |
| 20:19.27 | gandhijee | for whatever reason udev doesn't create the parport0 device |
| 20:19.42 | gandhijee | and if i try to do it by hand with mknod it doesn't seem to werk either |
| 20:19.51 | drath | did you load ppdev? |
| 20:20.10 | gandhijee | brb |
| 20:41.36 | gandhijee | ppdev should be a module, right? |
| 20:48.50 | drath | yeah |
| 20:51.49 | gandhijee | and is there a place i can read more about the .cfg file? i don't understand all the parts in it |
| 21:03.58 | ali_as | gandhijee, have you been able to get a scan chain list? |
| 21:04.51 | ali_as | Have you tried debugging the chain manually? |
| 21:05.47 | gandhijee | i am heading out |
| 21:05.58 | gandhijee | i will get back on when i am at my house |
| 21:06.23 | ali_as | I'll be off to sleep in an hour or so. |
| 21:06.31 | gandhijee | o |
| 21:06.46 | gandhijee | how would i start to debug the chain manually? |
| 21:07.24 | ali_as | Hang on. |
| 21:08.23 | ali_as | http://www.macraigor.com/ I thought had a brain dead one for free. |
| 21:08.54 | gandhijee | i have a wiggler |
| 21:08.57 | gandhijee | already |
| 21:09.06 | ali_as | I've done this with a Xilinx one using 'Impact' and I've done it with a Vlad type one using 'Jkeys'. |
| 21:09.16 | ali_as | Yes, software I'm thinking of. |
| 21:09.25 | gandhijee | o |
| 21:09.39 | ali_as | Just something that will let you toggle TDI/TCK/TMS manually. |
| 21:09.46 | ali_as | And let you watch TDO. |
| 21:10.11 | ali_as | An oscilloscope is helpful but not required. |
| 21:13.41 | ali_as | All I can find are pay ones on macraigor. Thats not helpful. |
| 21:16.31 | ali_as | What OS are you in? |
| 21:39.26 | ali_as | OCDCommander installs with 'JTAG Scan Chain analyser" but it might not help much. |
| 23:05.12 | gandhijee | linux |