**** BEGIN LOGGING AT Sat Jan 17 02:59:56 2009 Jan 17 07:29:12 * flyback is tired of these MOTHERCANUCKING SNAKES in his mothercanucking dream Jan 17 08:13:14 * flyback just got a null pointer deference error on the server he brought home from work to test the drives on but the cdrom for the live linux cd hit a read error so that might just be the cause Jan 17 08:14:58 * flyback I always like to explain null pointer deference errors as the scene in indiana jones III where has to spell out and walk on the stones that say "jehova" and he thinks it starts with a "j" Jan 17 09:14:28 * flyback was in shock how well this old pII 300 64 meg ram played video via vnc last n ight and if only he had sound :P Jan 17 16:44:59 * flyback ponders what flash drive to carry with him to the party and some other stuff, bbl Jan 17 17:52:05 * flyback bbl has to get some medical supplies so he can try to enjoy a company party with as little pain as possible Jan 17 21:28:27 Hi ! Jan 17 21:29:02 How do I extract and appropriate tap-definition from my hardware ? Jan 17 21:48:52 You'll need to know something about the hardware... How long is the instruction register, what the manufacturer ID is, if any... Jan 17 21:49:09 What's the board? Jan 17 22:34:59 noz: it is a hitex lpc3250-comstick Jan 17 22:36:59 http://www.ehitex.de/p_info.php?products_id=373&xID=8b4e13ddb743ee9bea9237adc1693aca Jan 17 22:44:29 Nice board. https://lists.berlios.de/pipermail/openocd-development/2008-November/003666.html (first Google hit on the CPU part) suggests that the IR is 4 bits long Jan 17 22:46:31 So something like 'jtag newtap arm926 cpu -irlen 4 -ircapture 0x01 -irmask 0xf -expected-id (something - start with 0xFFFFFFFF and work it out from debug) Jan 17 22:48:49 Possibly 0x07926f0f Jan 17 22:56:22 noz: ok, it then gives me this: Error: Could not validate JTAG scan chain, IR mismatch, scan returned 0x3f. tap=arm926.cpu pos=0 expected 0x1 got 3 Jan 17 22:57:12 noz: I really don't know what to do with it mostly because I don't fully understand how it all works. Jan 17 22:57:55 Hmm. One thing a lot of the tools lack is auto-detection. It's a shame, because it should be perfectly possible with the JTAG spec... Jan 17 22:58:54 noz: ... basically I know what type of tage I have, so fitting the cfg-file to the actual read of the IR would be possible right ? Jan 17 22:59:00 I found an excellent tutorial on JTAG - search for "Boundary-Scan Tutorial" from Asset Jan 17 23:00:12 I would have thought so Jan 17 23:00:24 noz: thanks ..... Jan 17 23:01:19 You could try with an older version, and the older style configuration line of "jtag_device 4 0x1 0xf 0xe", in case I got the translation to new style config wrong Jan 17 23:06:47 noz: I allready tried that yesterday. Actually the 0.1.0 release iss better since specifying somthin completely wrong as ft2232_device_desc gives me an error saying "LPC3250-Stick A" "LPC3250-Stick B" which is comforting for me becaus that informatin I guess is read from the device. Jan 17 23:08:14 Sadly this channel is a little quiet (unless you're interested in flyback's life ;) Jan 17 23:09:23 "flyback's life" ? Jan 17 23:09:27 Have you tried with -d3? Do you want to pastebin a little more of the context around the error? Jan 17 23:09:42 noz: sure ... 2 sec Jan 17 23:10:27 (Sorry - just a little dig - there's been nothing here for days except occasional posts from him telling us all what he's up to!) Jan 17 23:11:15 noz: config-file : http://pastebin.ca/1311201 Jan 17 23:11:58 k Jan 17 23:12:15 noz: -d 3 output: http://pastebin.ca/1311203 Jan 17 23:15:03 Can you compile openocd? Try again with 'make CFLAGS="-g -O2 -D_DEBUG_JTAG_IO_"' - gives more info about what's going on at a very low level Jan 17 23:15:39 noz: sure Jan 17 23:15:57 noz: version 0.1.0 is okay? Jan 17 23:16:05 The other thign you could try is slowing everything down with a jtag_speed 20 or similar Jan 17 23:16:26 noz: will that give different output? Jan 17 23:16:46 It may give the device a better chance, if it's a timing issue Jan 17 23:17:07 (not very scientific I know, but hey) Jan 17 23:17:47 noz: as low as 1 gives no difference... but it was worth the try Jan 17 23:18:13 Hang on a mo - you've got "JTAG communication failure, check connection, JTAG interface, targetpower etc." in the log. Jan 17 23:18:55 Nothing's going to work until you can get beyond that Jan 17 23:19:12 (speed is inverse - the higher the number, the slower it goes) Jan 17 23:19:51 Presumably it's all powered from the USB? Jan 17 23:25:24 noz: yes Jan 17 23:26:35 Then I'm stuck. Sorry. The basic comms to the JTAG adapter part aren't working. Jan 17 23:26:43 ... let alone the CPU Jan 17 23:27:00 noz: -d 3 with jtag_io_debug_flag: http://pastebin.ca/1311218 lesser informatin that before. Jan 17 23:27:08 noz: .... hmmm ok. Jan 17 23:27:37 noz: but how am I able to read out "LPC3250-stick" ? Jan 17 23:28:03 See. Line 52 - it's reading in all 1's Jan 17 23:28:49 Sorry - I don't have a ft2232-based interface, so I don't really know the things you might tweak. Jan 17 23:29:30 noz: hold on the jtag_speed in inversed ... that is 1 is faster that 60 ? Jan 17 23:29:46 Yes. It's a divisor Jan 17 23:30:03 Try 20 to start with Jan 17 23:30:18 hmm ... then 1 is problably a bad value ;D I had is set to 60. **** ENDING LOGGING AT Sun Jan 18 02:59:57 2009