03:39:40 | Russ | wb tim |
03:39:50 | Russ | trying to get my jtag stuff working |
04:23:39 | BZFlag | how far have you gotten? |
04:23:56 | BZFlag | got the jflash from cvs checked out and built? |
04:24:17 | Russ | ah |
04:24:28 | Russ | yah, looking at jflash |
04:24:34 | Russ | not getting anything on TDO |
04:24:42 | BZFlag | tdo? |
04:25:06 | Russ | its the jtag output pin |
04:25:19 | BZFlag | ah. |
04:25:27 | BZFlag | you've got leds there right? |
04:25:40 | Russ | been testing that stuff |
04:25:53 | Russ | going to hook up the LCD, and see if reset works |
04:26:26 | BZFlag | you soldered on a header? and built a cable? |
04:26:36 | Russ | yup, yup |
04:27:30 | BZFlag | using a holly dongle? |
04:27:58 | Russ | no, built something |
04:28:07 | BZFlag | ahh. |
04:28:28 | BZFlag | jflash has only been tested on holly's dongle with bi-di par ports. |
04:28:42 | BZFlag | will not work with par port in any other mode. |
04:29:39 | Russ | ah, I have an SPP, and an EPP port |
04:29:49 | Russ | the dongle is the same pinout |
04:30:10 | BZFlag | can you set either to just bidi? |
04:30:26 | Russ | I think SPP and EPP are bidir |
04:30:36 | Russ | my JTAG hardware at work is an ISA card |
04:30:42 | BZFlag | I know mine did not work in spp or epp mode. just bidi |
04:30:50 | Russ | ah |
04:30:57 | Russ | then I suppose I'll reboot |
04:31:06 | BZFlag | k |
04:40:15 | Russ | no go...parport1: PC-style at 0x378, irq 7 [PCSPP,TRISTATE] |
04:42:35 | Russ | TDO is reading .460V |
04:43:52 | Russ | which is going through the AHC as a 1 |
04:45:21 | Russ | I put .1uF caps on TMS, TDI, and TCLK from the parport |
04:45:31 | Russ | and then a .1uF cap on the IC |
04:45:37 | BZFlag | try --port=0x378 |
04:46:46 | BZFlag | so like ./flash_locks --port=0x378 where flash_locks is a symlink to jflash |
04:47:28 | Russ | nope |
04:47:35 | BZFlag | or can you disable parport0 ? |
04:47:50 | Russ | going to use parapin to do some testing |
04:47:58 | Russ | the .4V on TDO is the weird thing |
04:48:30 | BZFlag | hmmm... I don't know what the state should be. |
04:49:22 | Russ | should be 0V or 3.3V |
04:49:57 | BZFlag | so a 1v margin for error? ;-/ |
04:50:22 | Russ | well, with the AHC logic, I think .4V is in the undefined area |
04:51:10 | BZFlag | don't know what to tell ya. I know a few folks have it working, but I have not measured the lines. |
04:51:40 | BZFlag | I only have one phone here setup for jtag, but I just got more headers so I will build a few more. |
04:52:34 | BZFlag | I know people have seen cabling trouble before, I have some single cable (not ribbon) cables here and I can't get any of them to work well. |
04:52:45 | BZFlag | the ribbon cable in the pics on the site works well. |
04:53:01 | BZFlag | could be things are not quite spec. |
04:53:06 | Russ | I just used a cable from a MB->parallel port |
04:53:34 | Russ | cut off the end, and soldered it to some perfboard |
04:53:48 | Russ | seems one of my pins isn't clearing |
04:54:34 | BZFlag | hw problem? |
04:54:48 | Russ | seems to be |
04:54:50 | BZFlag | gotta reboot to run some silly windows app. back in a few |
04:54:53 | Russ | k |
05:06:11 | Russ | was it worth it? |
05:06:31 | Russ | even when the tuxscreen isn't connected, its stuck |
05:06:51 | Russ | interesingly enough, when I think it should be high, it 4.5V |
05:06:59 | Russ | when I think it should be low, it 5V |
05:08:31 | BZFlag | I hated it, but it worked. |
05:09:00 | BZFlag | that sounds strange. |
05:13:09 | Russ | what program was it? |
05:13:09 | Russ | maybe its a bad chip |
05:13:10 | BZFlag | I've got a lucent access point here (ap1000PING :175180124 |
05:15:13 | Russ | be interested to see how jtagmap.o works |
05:16:43 | BZFlag | switched them back now? what's next? |
05:17:15 | Russ | waiting for the iron to heat up |
05:17:35 | BZFlag | cool. bbiab |
05:29:22 | BZFlag | back. |
05:29:40 | BZFlag | firmware upgrade solved the dhcp client issue. ;-) |
05:29:51 | BZFlag | how goes the hacking? |
05:32:17 | Russ | hmm...no help yet |
05:53:22 | BZFlag | lag... |
05:53:38 | BZFlag | got a camera? |
05:53:48 | BZFlag | want to see my setup? |
06:23:15 | Russ | ah, your back |
06:23:15 | Russ | I don't have a camera here, but I'll see any pics you got |
06:24:59 | BZFlag | nice network tonight huh? |
06:25:20 | Russ | yah |
06:26:48 | BZFlag | hang on... |
06:28:52 | BZFlag | http://guest:guest@rikers.org:8000/ |
06:29:19 | BZFlag | k, what do you want to see? ;-) |
06:30:00 | Russ | a holly gates dongle |
06:30:15 | Russ | and a screwdriver that probably has a red end |
06:30:22 | Russ | some blue laplink cables |
06:30:59 | BZFlag | close |
06:31:19 | Russ | it stopped... |
06:31:28 | Russ | er, there we go |
06:32:12 | BZFlag | on sec... restarting... |
06:32:32 | Russ | what is the loopback for anyway? |
06:33:35 | BZFlag | better? |
06:33:52 | BZFlag | no idea. the cable I got had it, so I've always wired it. |
06:34:00 | Russ | galeon keeps stopping |
06:34:07 | Russ | lemme try ns4.7 |
06:34:24 | BZFlag | I think It just stops when I connect here. |
06:36:25 | Russ | expected 0xbeef, returned 0x4110 |
06:36:27 | Russ | getting better |
06:37:58 | Russ | what do you notice about that? |
06:38:10 | BZFlag | xored |
06:38:36 | BZFlag | well, with ffff that is. |
06:38:52 | BZFlag | bitswapped would be better explaination I suppose. |
06:39:03 | Russ | not is the term |
06:39:26 | BZFlag | ah, even better. |
06:39:51 | Russ | I think my caps are messed |
06:42:48 | Russ | I need to sleep a bit between accesses |
06:42:53 | BZFlag | ThaT wOulD lOok sTrAnGe. |
06:43:04 | BZFlag | is punchy |
06:43:25 | Russ | is there any GPIO's I can see? (input or output? |
06:43:58 | Russ | all the GPIO's but GPIO0 read as on |
06:44:36 | Russ | and batt fault is on (no batteries) |
06:44:39 | BZFlag | have not traced any if that's what you mean. |
06:44:55 | BZFlag | there are batts in the box... ;-) |
06:45:11 | Russ | I know, the board is not in the case right now (thanks btw) |
06:46:01 | BZFlag | just need a 6v power supply. got plenty right? |
06:46:14 | Russ | heh, course not |
06:46:26 | BZFlag | oh, and that bats were in all the boxes. |
06:46:55 | BZFlag | can you reach the clips on the case? that seems like a good toggle test point to me. |
06:47:30 | Russ | It seems like stuff is working to me |
06:47:38 | Russ | any clue which caps are important |
06:48:17 | Russ | I think I need to remove my input caps |
06:51:19 | BZFlag | no idea. trial an error? |
06:51:33 | Russ | argh...its perf board |
06:51:39 | BZFlag | got a scope? watch for spikes? |
06:51:56 | BZFlag | you have used jflash before? |
06:52:45 | BZFlag | the interface is a bit strange, if output is redirected, it dumps, if input is redirected it loads. both can be there, the dump happens first. |
06:53:22 | Russ | I have a few jtag apps |
06:53:30 | Russ | using jtagedit right now |
06:53:53 | Russ | the first thing it does is write and read from the loopback register |
06:54:02 | Russ | not a chance I would have a scope |
06:54:09 | Russ | don't even have one at work |
06:54:50 | BZFlag | me neither. someday I'll ebay one. |
06:56:43 | BZFlag | so anything you want to see on the cam? |
06:58:16 | Russ | I dunno |
06:58:30 | Russ | removing solder with a wick atm |
07:01:20 | Russ | much better |
07:01:44 | Russ | have you looked at jtagedit? |
07:01:55 | BZFlag | nope. does it do amd? |
07:02:06 | BZFlag | that's what's on your board too right? |
07:02:23 | Russ | jtagedit is GTK lowlevel |
07:02:34 | Russ | you act like you are the CPU |
07:02:43 | BZFlag | cool. |
07:02:49 | Russ | type in an address, some data, read or write |
07:03:01 | Russ | it sets the addressing line appropriotely |
07:03:07 | Russ | watch stuff in realtime, etc |
07:03:36 | Russ | I think something is still wrong with my dongle tho |
07:05:33 | Russ | and it just stopped working |
07:06:02 | Russ | the other module I have is a kernel module for MTD, so any flash MTD supports, you can use |
07:06:59 | Russ | is the loopback jtag reset |
07:07:07 | Russ | (nTRST) |
07:08:18 | BZFlag | ever done gdb over jtag? |
07:08:29 | BZFlag | that would be very useful to me. |
07:08:39 | BZFlag | debugging blob itself. ;-) |
07:08:52 | Russ | gdb over jtag wouldn't work on SA-1100 I don't think |
07:09:06 | BZFlag | why? |
07:09:37 | Russ | because on the SA-1100 it only has a boundry scan interface (read/modify pins) |
07:10:23 | BZFlag | hmm... so the remote would have to duplicate all the cpu actions... |
07:10:40 | BZFlag | and then step everything... so performance would be dog slow. |
07:10:51 | Russ | which would mean it would make more sense to have an emulator |
07:10:56 | Russ | (as you would need one) |
07:11:02 | BZFlag | right. |
07:11:04 | Russ | also, you couldn't use DRAM |
07:11:13 | Russ | because you would miss all your refresheh |
07:11:33 | BZFlag | eek. forgot that minor point. ;-/ |
07:11:54 | Russ | not sure where I need caps.... |
07:12:10 | Russ | or if .1uF is too much |
07:12:11 | BZFlag | things still flakey? |
07:12:27 | Russ | yah |
07:12:35 | Russ | expected 0xbeef, returned 0x3c0f |
07:12:42 | Russ | expected 0xbeef, returned 0x000f |
07:12:46 | Russ | expected 0xbeef, returned 000000 |
07:12:50 | Russ | expected 0xbeef, returned 0x200f |
07:12:54 | Russ | expected 0xbeef, returned 0x380f |
07:12:58 | Russ | expected 0xbeef, returned 0xbecf |
07:13:28 | Russ | its schmitt triggered logic |
07:14:39 | Russ | it seems to only be missing bits (1->0) |
07:15:24 | Russ | so I think the inputs are ok |
07:16:05 | BZFlag | does the screen keep resetting? |
07:16:22 | Russ | I don't have the screen plugged in |
07:16:46 | BZFlag | I find the device resets after I do most jtag accesses so I need to delay between accesses. |
07:17:15 | BZFlag | punch a hole in the back of the case and put it all back together. |
07:17:17 | Russ | if you circle the reset state three times, it will reset |
07:17:32 | Russ | the cable can go through the ram slot |
07:17:45 | Russ | hey look, its a tuxscreen |
07:18:10 | Russ | I think I'll put a cap on TDO |
07:18:41 | Russ | the device shouldn't reset while waiting in jtag |
07:18:57 | Russ | unless its something strange your jtag software is doing |
07:24:16 | Russ | nice plug hack |
07:25:30 | Russ | argh...its much worse with that cap |
07:25:48 | BZFlag | k, in focus now and everything. |
07:26:54 | Russ | yah, I hope I won't have to use jtag for more than a few weeks |
07:27:07 | Russ | is the loopback nTRST? |
07:27:15 | BZFlag | no idea. |
07:27:35 | Russ | then why is it there? |
07:27:37 | BZFlag | I got all this stuff with almost no docs. |
07:27:58 | BZFlag | almost everything I know is up on the wiki. |
07:28:21 | BZFlag | I'm still digging for more whereever I can. |
07:28:25 | Russ | right, but why did you put the loop in there |
07:28:34 | BZFlag | I got the cable this way. |
07:28:45 | Russ | who made the cable? |
07:28:52 | BZFlag | I had one phone that had working jtag, I hacked more of them to be the same. |
07:29:06 | BZFlag | moberg, but I think he jus copied another... |
07:29:16 | BZFlag | don't know where the first one came from. |
07:29:29 | Russ | maybe phillips |
07:29:35 | BZFlag | perhaps. |
07:29:47 | BZFlag | but shaggy never had one. |
07:29:54 | BZFlag | I'm sending one to him. |
07:34:43 | Russ | expected 0xbeef, returned 0x7dd |
07:34:45 | Russ | er |
07:34:50 | Russ | expected 0xbeef, returned 0x7dde |
07:34:55 | Russ | everytime now.... |
07:39:31 | Russ | 7dde == beef << 1 |
07:39:54 | BZFlag | hmmm.... timing issue? |
07:40:26 | BZFlag | are your cables all exactly the same length? |
07:40:43 | Russ | jtag doesn't really have a timing issue |
07:41:21 | Russ | check my manual... |
07:42:48 | Russ | I took the jtag part out and put it somewhere....oops |
07:44:28 | Russ | so why not use the serial port for blob debugging? |
07:47:35 | BZFlag | ok for debugging I guess, have to put something like gdbstubs in. |
07:47:58 | Russ | I mean for the current run from ram problems |
07:48:29 | BZFlag | yeah. That's what I'm going to do next. |
07:49:05 | Russ | probably will give good clues as to why a kernel isn't booting |
07:49:05 | BZFlag | in #TuxScreen? |
07:49:12 | Russ | not atm |
07:49:35 | BZFlag | Tangent and I are having the same conversation. ;-) |