**** BEGIN LOGGING AT Mon Feb 16 02:59:58 2009 Feb 16 19:18:55 Hi dirk2 Feb 16 19:22:32 hi Feb 16 19:22:57 any svf progress? Feb 16 19:23:36 no, but I think the sequence was 100% oke anyhow Feb 16 19:24:06 but you can't get the ID register 0x4? Feb 16 19:24:48 I need to look into that. this info needs to be on the elinux wiki Feb 16 19:25:55 I think depending on free time, I will go on with some more hacking directly in urjtag. Maybe I can implement a 'init OMAP3 command' Feb 16 19:25:56 but I had fun today creating a "visualisation" for the logic analyzer using ruby Feb 16 19:26:33 btw: Thanks for the logic analyzer hint last week. A friend of mine is thinking about buying it Feb 16 19:26:53 $160 is quite cheap Feb 16 19:27:04 and they propose linux software for ~March Feb 16 19:27:42 s/propose/promise/ Feb 16 19:27:43 dirk2 meant: and they promise linux software for ~March Feb 16 19:27:45 http://box.mmapps.net/analyzer.zip and the UI looks like this http://paste-it.net/raw/public/a44f1e4/ Feb 16 19:28:08 dirk2: they promissed it end jarnuary also Feb 16 19:28:29 keesj: well, I think for $160 it's ok ;) Feb 16 19:28:52 the protocol can only be very simple anyway Feb 16 19:29:15 I looked on sourceforge and there are at lease a few OK GUI's Feb 16 19:30:32 did you try to run the logic analyzer data through something like a perl script to create a JTAG protocol analysis? Feb 16 19:30:40 I think that usbee is a concurent of the one I am using https://www.sparkfun.com/commerce/product_info.php?products_id=8938 there are some comments that all match my experiance Feb 16 19:33:01 not yet , that wold really be trivial, perhaps the next step , but clearly not required any more. to follow the jtag state machine you only need to look at the clock and tms Feb 16 19:34:10 back to OMAP3 jtag: Looking at urjtag's 'discovery' output it (correctly) detects DR #7 as 8 bit long. Feb 16 19:34:22 If I understand TI's config correctly, we have to send 0x89 to this register. Feb 16 19:34:40 Then we have to send 0xa3002108 as 32 bit value to DR #2 (?). Feb 16 19:34:52 But DR #2 is detected as 1 bit length only: Detecting DR length for IR 000010 ... 1 Feb 16 19:35:57 yes I saw that , but bdi does the same. therefore I think it must be the discovery that goes wrong Feb 16 19:36:53 hmm, or with 0x89 we switch to a 32 bit register? Feb 16 19:37:39 for instance, write 0x89 to DR #7 == enable 32 bit register DR #2 Feb 16 19:43:36 we need to try the dbi sequence Feb 16 19:43:47 https://accesovpn.terra.es/http/elinux.org/BeagleBoardJTAG Feb 16 19:44:48 sorry, dbi? Feb 16 19:45:04 bdi Feb 16 19:45:33 and about Parameter : The port address field is '0x0f000000'. Parameter : The port address value is '3'. Feb 16 19:46:12 I think what they are saying is that in the command that follow Feb 16 19:47:08 the address used is (3) the as the bytes in [23:27] Feb 16 19:47:22 so Parameter : The send data value is '0xa3002108'. Feb 16 19:47:55 is't the same 3 and probably also saying that the DAP is located and the 4th port of the JCR Feb 16 19:48:15 bdi: yes (I wonder about the URL https://accesovpn.terra.es/ though) Feb 16 19:48:36 as there can be up to 15 tap's connected (needs 4 bytes) Feb 16 19:48:39 Parameter : The send data value is '0xa3002108'. Feb 16 19:48:55 yes, this is my understanding, too Feb 16 19:50:33 * keesj is searching for the screen shot showing the ICEPICK and the other components somewhere else on the wiki Feb 16 19:51:08 * dirk2 think to remember that it is in the TI slides Feb 16 19:52:57 http://focus.ti.com/lit/ml/sprp603/sprp603.pdf page 6 ? Feb 16 19:54:19 or page 10 ? Feb 16 19:58:00 keesj: btw: Do you like to attend with OMAP3 OpenOCD project at Beagle contest? Feb 16 19:58:29 keesj: I can't do it my self, but I like to help :) Feb 16 19:59:05 I don't see what I can gain doing that. It might help getting more people in Feb 16 19:59:35 You could win a beagle rev C2 Feb 16 20:03:20 keesj: Assuming we get UrJTAG to configure OMAP3 properly (by svf or source code hack), how would we proceed with OpenOCD then? Feb 16 20:04:39 That would be step two Feb 16 20:05:17 that is where one need to read the cortex trm and implement software support Feb 16 20:05:22 Do you have already any plan/idea for this second step? Feb 16 20:06:05 not as such no. Feb 16 20:06:36 I would procceed much in the same way I have been doing until now Feb 16 20:06:58 I think the openocd guys will hop in as it is then a "software only" issue Feb 16 20:07:35 keesj: Hmm, before TRM step OpenOCD needs 'do JTAG init the same way we did with urjtag and which showed to work' step? I.e. OpenOCD JTAG config has to be adapted to do stuff needed for OMAP3 correctly, no? Feb 16 20:08:18 yes , perhaps as you sugested urjtag could be the first step. Feb 16 20:08:56 my personal feeling is that once it works for urjtag we must have enough understanding of it to make it work under openocd Feb 16 20:09:37 yes. Or show the OpenOCD people how it works in urjtag and ask them 'how can we do it the same way in OpenOCD' Feb 16 20:09:39 I would also certainly look if it is possible to implement it on urjtag. Feb 16 20:10:57 'to implement it on urjtag': implement what? Don't use OpenOCD and use e.g. gdb with urjtag? Feb 16 20:12:39 openocd also has svf support. but when we know what to do it can be as simple as Feb 16 20:13:02 "don't" reset the damm thing once initialized. Feb 16 20:13:58 My biggest fear is that we need adaptive clocking or something that we can simply not do Feb 16 20:14:24 adaptive clocking? RTCK? Feb 16 20:14:32 yes **** ENDING LOGGING AT Mon Feb 16 23:03:53 2009 **** BEGIN LOGGING AT Mon Feb 16 23:04:14 2009 **** ENDING LOGGING AT Mon Feb 16 23:05:49 2009 **** BEGIN LOGGING AT Mon Feb 16 23:06:11 2009 **** ENDING LOGGING AT Tue Feb 17 02:59:58 2009