**** BEGIN LOGGING AT Sat Aug 10 02:59:58 2013 Aug 10 18:06:27 Panto, ping Aug 10 18:07:32 Ka6sox, ping Aug 10 18:08:20 mluckham-mobile, pong Aug 10 18:08:39 Hi tom, how are the bugs? Aug 10 18:12:50 I've been reading about sysfs and drivers at kernel.org and I wonder if there's a mailing list for the work panto's been doing Aug 10 18:13:51 I would also like to discuss possible design of jtag device driver implemention Aug 10 18:14:14 Maybe soneone is already doing that? Aug 10 18:15:27 It would be needed to use the prurproc approach for our project Aug 10 18:16:33 Sorry for spelling faults, fat fingers Aug 10 18:18:04 Home internet not working, a passing truck pulled the overhead cable down Aug 10 18:47:12 uh oh Aug 10 18:47:29 sorry, I'm out fighting the Moose over the guy wires... Aug 10 18:47:40 think I've got things stabilized enough for winter. Aug 10 19:01:48 Did you have to climb it? Aug 10 19:02:14 nope, I can't do that anymore Aug 10 19:03:39 Bad knees Aug 10 19:03:50 Weak fingers Aug 10 19:04:24 bad knees Aug 10 19:04:30 Are you in the middle of nowhere, or in a town? Aug 10 19:05:25 10kms north of a town Aug 10 19:05:34 I can't climb up 60m anymore Aug 10 19:06:29 Better than 500m Aug 10 19:07:05 none of the ones I've ever worked on were taller than 421m Aug 10 19:08:28 Lol ... still impractical, spend hours climbing up and get a bug in your eye, have to get down Aug 10 19:09:28 I've only gone up 124m on a tower...but no more. Aug 10 19:09:35 I sent email to you and Panto with my ideas Aug 10 19:09:36 okay back to your original questions. Aug 10 19:09:42 okay let me look. Aug 10 19:10:46 we could do this on the beagleboard forums....99.9% of the folks would be bored...but its a good place for this. Aug 10 19:11:28 I was thinking that the PRU would handle the JTAG signals. Aug 10 19:11:40 maybe I'm pushing this in the wrong direction... Aug 10 19:11:53 please give me ideas. Aug 10 19:12:37 Yes, pru will do it, but have to communicate with it and have a model Aug 10 19:12:54 For user mode programs Aug 10 19:13:33 My problem is lack of knowledge about jtag Aug 10 19:13:46 we will need commands/responses and data Aug 10 19:14:05 so things like start, length, abort, etc will need to be defined Aug 10 19:14:30 But I can see how to implement what jj2baile did in his demo code, using my ideas Aug 10 19:14:42 okay good. Aug 10 19:15:14 I'm on an airplane in 5hrs for about 9hrs...then I'll be back Aug 10 19:15:24 and can work on getting more done then. Aug 10 19:15:57 Could you give me some specific examples of signalling, including real time response requirements if any? Aug 10 19:17:05 I think the primary thing to remember is that you push bits out and they *do* come back as well. Aug 10 19:17:11 For example is tck needed to be carefully tined, or just used as a strobe? Aug 10 19:17:19 some of them are responses...and others are what you send. Aug 10 19:17:41 tck is the sync clock...nothing gets shifted out or in wihtout that. Aug 10 19:17:46 Yes, would use vring for responses Aug 10 19:18:36 responses could be results from a command/status or data Aug 10 19:19:01 I think a sysfs entry for command responses/status would be better than using a vring for this. Aug 10 19:19:14 the vrings should be used for data only Aug 10 19:19:53 Raw bytes? Aug 10 19:20:17 its a bitstream... Aug 10 19:20:34 so we have to parallized it to/from the app but serialize it in the PRU. Aug 10 19:21:14 JTAG doesn't come in neat 8 bit byte increments. Aug 10 19:21:31 539 bits might be a full frame. :P Aug 10 19:22:12 Ok, this is good to know Aug 10 19:22:56 Use sysfs for read/write jtag registers? Aug 10 19:25:16 mostly pru control/status I think Aug 10 19:26:54 We'll need to work it all out as a model, then start coding Aug 10 19:28:11 I'll write something up on the elinux.org wiki and use it as the starting place for the models. Aug 10 19:28:33 until I saw the remoteproc/vring stuff I didn't know what was possible. Aug 10 19:28:33 Ok. BTW I did write some loopback code, testing comms from a user mode program and pru0 Aug 10 19:28:49 this whole thing its "experimental" at this stage...bleeding edge. Aug 10 19:29:04 good. Aug 10 19:29:30 This could be a general model for jtag device drivers Aug 10 19:29:45 So best to get it right :) Aug 10 19:30:27 Also I made experimental changes to pru code, which worked Aug 10 19:31:25 Gotta go now too, and soaring tomorrow ... Can work on this next week starting Monday Aug 10 19:32:22 me too...flying for a while...need to integrate back with the Family when I get home tomorrow then I'll be starting Monday. Aug 10 19:32:46 Talk to you then ... bye **** ENDING LOGGING AT Sun Aug 11 02:59:58 2013