**** BEGIN LOGGING AT Wed Sep 13 02:59:56 2006 Sep 13 18:11:36 <[g2]> anybody seen AD around ? Sep 13 18:20:54 ~seen AchiestDragon Sep 13 18:21:10 achiestdragon was last seen on IRC in channel #openjtag, 44d 20h 46m 52s ago, saying: 'for debug etc , as sample ram ?'. Sep 13 18:21:34 hm, i've definitely seen him since then Sep 13 18:41:46 vmaster: you get my email? Sep 13 18:49:02 hey prpplague, yeah, got it Sep 13 18:49:42 we talked about how the output enables your boss wanted didn't make a lot of sense, did we? Sep 13 18:49:56 at least i hope we did :) Sep 13 18:51:00 the design is fine as it is, but i'll tie these output enables permanently on in the openocd Sep 13 18:52:23 vmaster: seems like we discussed it Sep 13 18:52:55 vmaster: which ones are you refering to? Sep 13 18:53:04 vmaster: i have the schematic in front of me now Sep 13 18:53:45 vmaster: ok, yea, i remember now Sep 13 18:54:05 vmaster: we have specific piece of hardware we want to use with this module Sep 13 18:54:09 ah, ok Sep 13 18:54:28 vmaster: so for standard jtag they aren't really needed Sep 13 18:54:55 yeah, exactly Sep 13 18:55:51 vmaster: i've started work on a small c app that will jtag on some code, then start it, causing the uart to receive data from the serial port Sep 13 18:56:49 vmaster: we plan to have 4 to 8 of these devices connected to a single pc for gang loading of main boards Sep 13 18:57:08 why do you want to use the serial port? Sep 13 18:57:46 i mean, jtag with the ft2232 is going to faster than 115200 Sep 13 18:57:50 vmaster: no specific reason than its already on the ft2232 and its easy to code for Sep 13 18:58:08 ah, ok Sep 13 18:58:19 vmaster: yea, the key is that we want to use some error checking Sep 13 18:59:08 vmaster: we'll use xmodem Sep 13 19:01:41 vmaster: basically we are just doing a flash of about 4mb on some NOR flash Sep 13 19:02:16 vmaster: what kind of performance do you normally get using a ft2232 to do flashing? Sep 13 19:04:19 haven't used the CFI flash driver in a while, but RAM download is ~120kb/s, and flash was about half of that, iirc Sep 13 19:04:53 thats pretty darn good Sep 13 19:05:04 vmaster: thats using openocd? Sep 13 19:05:15 yeah Sep 13 19:05:49 but not with default settings - you'll have to use "arm7_9 fast_memory_access enable" and "arm7_9 dcc_downloads enable" Sep 13 19:06:14 these are disabled by default to prevent users from outrunning devices running from a slow-clock like 32kHz Sep 13 19:06:16 vmaster: at 60kp/s that would be just a little over a minute to upload all 4mb Sep 13 19:06:41 i can write str7 flash at that speed, and the algorithms are comparable Sep 13 19:06:53 of course, the str7 is done in 4 seconds :) Sep 13 19:07:21 vmaster: any verification of the data done? Sep 13 19:07:37 nope, but you can read at >50kb/s Sep 13 19:08:11 hmm Sep 13 19:08:47 * prpplague wonders if he could script something up to do the load using openocd Sep 13 19:09:14 that's what people are doing Sep 13 19:09:42 vmaster: question i have is can i run 4 copies at once Sep 13 19:10:08 you can, but i suspect performance is going to suffer Sep 13 19:10:17 vmaster: hmm Sep 13 19:10:29 vmaster: well i should be able to test with openocd for the performance Sep 13 19:10:46 vmaster: are the changes you made in cvs/svn ? Sep 13 19:11:41 changes for your ft2232 layout? not yet Sep 13 19:11:50 ahh ok Sep 13 19:12:01 * prpplague makes the changes to his local copy Sep 13 19:12:14 brb Sep 13 19:25:27 vmaster: hmm, openocd doesn't seem to be finding me device Sep 13 19:25:42 do you use your own vid/pid? Sep 13 19:26:43 no, right now we are just using the defaults for testing, they match Sep 13 19:26:56 i think i dont have it enable for ftdi2232, recompiling now Sep 13 19:27:30 ftdi2232.c:33:18: ftdi.h: No such file or directory Sep 13 19:27:43 mhh, ftdi2232.c is dead Sep 13 19:28:03 there was a lot of duplicate code in ftdi2232.c and ftd2xx.c, now everything is in ft2232.c Sep 13 19:28:24 ahh, guess i need to grab the latest copy then Sep 13 19:28:56 ah, my performance figures are for FTD2XX, not for libftdi Sep 13 19:29:04 libftdi is a lot less Sep 13 19:29:26 ahh Sep 13 19:29:43 FTD2xx is the FTDI official lib correct? Sep 13 19:29:47 yes Sep 13 19:30:52 hmm Sep 13 19:31:01 lots of errors after svn update :( Sep 13 19:32:25 mhh, the current code compiles cleanly on linux (ftd2xx and libftdi), windows (ftd2xx) and freebsd (not sure what they're using) Sep 13 19:32:58 vmaster: yea i think its just it wasn't clean Sep 13 19:33:12 ok Sep 13 19:40:23 hmm Sep 13 19:40:29 compilied properly Sep 13 19:40:35 still not seeing my device though Sep 13 19:40:48 Info: openocd.c:82 main(): Open On-Chip Debugger (2006-09-07 20:00 CEST) Sep 13 19:40:48 Info: configuration.c:50 configuration_output_handler(): Command ft2232_device_desc not found Sep 13 19:40:48 Info: configuration.c:50 configuration_output_handler(): Command ft2232_layout not found Sep 13 19:40:48 Info: configuration.c:50 configuration_output_handler(): Command ft2232_vid_pid not found Sep 13 19:41:51 do you have "interface ft2232" in the .cfg file? Sep 13 19:42:29 yep Sep 13 19:43:58 did you configure with --enable-ft2232_ftd2xx or --enable-ft2232_libftdi? Sep 13 19:45:00 vmaster: with libftdi, changing now, my bad Sep 13 19:45:49 * prpplague grabs the ft2xx source Sep 13 19:45:59 mhh, no source Sep 13 19:46:07 lib i mean Sep 13 19:46:07 that's the problem with ftdi's own driver Sep 13 19:46:09 ok Sep 13 19:47:08 hmm Sep 13 19:47:16 i'm not seeing a place to dl that off hand Sep 13 19:47:22 thats on the ftdichip.com site? Sep 13 19:47:36 for 32-bit linux? hold on Sep 13 19:47:53 http://www.ftdichip.com/Drivers/D2XX/Linux/libftd2xx0.4.10.tar.gz Sep 13 19:48:02 the driver page is here: http://www.ftdichip.com/Drivers/D2XX.htm Sep 13 19:48:13 ahh Sep 13 19:49:46 argh Sep 13 19:49:50 someone needs to kick them Sep 13 19:50:01 don't they know how to make a proper tarball Sep 13 19:50:10 nope Sep 13 19:54:03 where do you normally put the ft2xx.h file? Sep 13 19:56:04 i have it in /usr/local/include Sep 13 19:56:21 you could also put it into the OpenOCD tree in src/jtag/ Sep 13 19:56:22 yea thats where i put it as well Sep 13 19:58:17 vmaster: yea that did the trick Sep 13 19:59:29 vmaster: a little closer Sep 13 20:08:56 prpplague: this is what i'd like to include in the OpenOCD code to support your device: http://mmd.ath.cx/ft2232_m5960.diff Sep 13 20:09:19 i didn't include led handling, as that's not really usable as it is now Sep 13 20:11:47 * prpplague looks Sep 13 20:14:02 * prpplague tests Sep 13 20:15:34 Debug: ft2232.c:972 ft2232_init(): 'ft2232' interface using FTD2XX with 'm5960' layout Sep 13 20:15:34 Error: ft2232.c:997 ft2232_init(): unable to open ftdi device: 2 Sep 13 20:15:34 Error: ft2232.c:1012 ft2232_init(): ListDevices: 3 Sep 13 20:17:09 did you specify the right ft2232_device_desc? Sep 13 20:18:01 i guess not, i just used m5960 as the description Sep 13 20:19:54 lsusb -v as root, the iProduct Sep 13 20:20:09 don't remember what the FTDI uses by default Sep 13 20:23:29 vmaster: ahh didn't catch that Sep 13 20:23:53 ah, sorry, not iProduct Sep 13 20:24:40 iInterface - but the OpenOCD should list all the devices it found Sep 13 20:25:38 yea got it Sep 13 20:26:39 vmaster: we it recognized the device Sep 13 20:27:01 Info: ft2232.c:359 ft2232_send_and_recv(): inter: 0.1451, inter2: 0.1451 end: 5.7912 Sep 13 20:27:01 Error: ft2232.c:367 ft2232_send_and_recv(): ft2232_expect_read (2) != ft2232_buffer_size (0) (0 retries) Sep 13 20:27:39 vmaster: thanks for the help, i'm gonna have to get at it in the morning Sep 13 20:27:52 heh, funny, i inserted that timeout just for one windows user Sep 13 20:28:05 end: 5.7912 means it waited more than 5 seconds for a reply Sep 13 20:28:10 doh Sep 13 20:28:34 yeah, just let me know if you need anything Sep 13 20:28:49 thanks, i'll get to doing some serious work tomorrow Sep 13 20:28:54 vmaster: later Sep 13 20:28:57 cya **** ENDING LOGGING AT Thu Sep 14 02:59:56 2006