**** BEGIN LOGGING AT Tue Jul 04 02:59:56 2006
Jul 04 09:08:19 <rwhitby> Have you guys seen the OpenOCD project and www.amontec.com ?
Jul 04 10:32:27 <ka6sox> hey NAiL 
Jul 04 10:32:34 <NAiL> hi there :)
Jul 04 10:32:44 <ka6sox> got insomnia tonight
Jul 04 10:32:54 <ka6sox> almost 4am here and I'm still up.
Jul 04 10:34:25 <rwhitby> hey ka6sox 
Jul 04 10:34:35 <rwhitby> ka6sox: are you involved in OpenOCD?
Jul 04 10:35:11 <ka6sox> rwhitby, nope
Jul 04 10:35:37 <ka6sox> that would be vmaster_
Jul 04 10:35:50 <rwhitby> Looks like a cool project.  I might be using it for work (we need to hook up Eclipse to a DSP core).
Jul 04 10:36:12 <ka6sox> fun
Jul 04 10:36:18 <ka6sox> vmaster is your guy.
Jul 04 10:36:35 <rwhitby> Is vmaster_ a regular here?
Jul 04 10:36:43 <ka6sox> I think that [g2] also has played with it.
Jul 04 10:36:56 <ka6sox> ya...vmaster is here pretty regularly...he's in .de
Jul 04 10:37:09 * rwhitby adds #openjtag to his auto-connect list ...
Jul 04 10:37:30 <rwhitby> I sent amontec an email seeing if they wanted to donate some hardware ...
Jul 04 10:37:36 <rwhitby> www.amontec.com
Jul 04 10:37:43 <ka6sox> good plan.
Jul 04 11:00:59 <vmaster_> hey
Jul 04 11:01:12 <ka6sox> morning vmaster.
Jul 04 11:01:19 <rwhitby> hey vmaster 
Jul 04 11:01:24 <vmaster> morning ka6sox, hey rwhitby 
Jul 04 11:01:37 <rwhitby> I hear you're the OpenOCD expert around these parts ...
Jul 04 11:01:54 <rwhitby> Are you on the development team of it?
Jul 04 11:02:05 <vmaster> guess i am the development team ;)
Jul 04 11:02:23 <vmaster> a few contributions from others, mostly bug fixes, and support for flash chips i don't own
Jul 04 11:02:37 <rwhitby> ah, you're Dominic?
Jul 04 11:02:42 <vmaster> yeah
Jul 04 11:02:47 <rwhitby> Pleased to meet you.
Jul 04 11:03:08 <rwhitby> I was just reading through your code, seeing how difficult it would be to add support for some Freescale processors ...
Jul 04 11:03:59 <vmaster> that really depends on the debug hardware used by freescale
Jul 04 11:04:09 <vmaster> arm7/9 and xscale for example is almost completely different
Jul 04 11:04:19 <vmaster> only the jtag stuff and architecture is the same
Jul 04 11:04:26 <vmaster> armv4/5t
Jul 04 11:04:42 <vmaster> oh yeah, and the gdb stuff
Jul 04 11:06:59 <rwhitby> My company (www.astc-design.com) is looking to connect Eclipse to some Freescale chips, and OpenOCD looks (on the face of it) to be a good platform for that sort of work ...
Jul 04 11:07:16 <vmaster> is there public documentation available on these chips?
Jul 04 11:07:33 <rwhitby> Good question.
Jul 04 11:08:35 <rwhitby> I know they give out Debug Access information to companies like Green Hills Software, etc.
Jul 04 11:08:44 <rwhitby> Not sure whether it's under NDA or not.
Jul 04 11:09:11 <rwhitby> The chip data books should have debug access chapters in them ...
Jul 04 11:10:28 * rwhitby checks www.freescale.com ...
Jul 04 11:12:50 <ka6sox> boo hisss...green hills....they are here in town.
Jul 04 11:21:35 <rwhitby> vmaster: here's an offer from the nslu2-linux project - we'll buy you a slug if you want to get OpenOCD to work with the NSLU2.
Jul 04 11:23:34 <rwhitby> vmaster: the Freescale chips we're looking at have the "EOnCE" (Enhanced On-Chip Emulation) block for debug access through jtag, and it's documented in the chip user manual.
Jul 04 11:23:45 <rwhitby> (e.g. for the 56800E DSP core)
Jul 04 11:24:19 * ka6sox programmed the 56K a LONG time ago with DSP in Assembly.
Jul 04 11:27:20 <vmaster> heh, that would be really cool
Jul 04 11:27:50 <vmaster> i'm having exams once again the next two weeks, so i wont work on the xscale stuff for some time, but after that I'd be glad to t
Jul 04 11:27:56 <vmaster> talk to you about your offer
Jul 04 11:28:19 <vmaster> i'll have a look at the freescale EOnCE
Jul 04 11:28:26 <rwhitby> Do you have a way to buy an NSLU2 locally?  Or would you require one to be shipped from the US?
Jul 04 11:28:54 <rwhitby> (If you can get one locally, we can just PayPal you the cost)
Jul 04 11:29:04 <vmaster> i can get one locally
Jul 04 11:29:19 <rwhitby> what's the cost (in USD$)?
Jul 04 11:29:42 <wookey> has dave bisset turned up on this channel to talk to anyone yet?
Jul 04 11:29:49 <wookey> he's the balloon 3 JTAG guy
Jul 04 11:30:00 <wookey> I suspect not so I'll kick him again
Jul 04 11:30:03 <vmaster> i don't think so
Jul 04 11:30:17 <rwhitby> vmaster: is this the main place on IRC where you discuss OpenOCD?
Jul 04 11:30:30 <wookey> (hi rwhitby)
Jul 04 11:30:43 * rwhitby waves to wookey - long time no speak
Jul 04 11:30:43 <vmaster> if ever, than in #openjtag, i guess
Jul 04 11:30:58 <vmaster> i'll grab something for lunch, bbiab
Jul 04 11:31:13 <rwhitby> I'll be gone to bed by then.
Jul 04 11:31:21 <rwhitby> Catch you another day.  Good luck with the exams.
Jul 04 11:33:09 <wookey> vmaster: I see that xscale support is 'coming along nicely' on june 12th. Does that mean we can use it for uploading files already, or is more needed?
Jul 04 11:33:22 <wookey> (we are using pxa270)
Jul 04 11:34:14 <rwhitby> wookey: is your company considering using OpenOCD?
Jul 04 11:34:26 <wookey> yes
Jul 04 11:34:51 <wookey> we currently have our own jtag tools which are a re-wrtitten (and much better) multi-chain jflash
Jul 04 11:34:54 <wookey> called bflash
Jul 04 11:35:22 <rwhitby> wookey: From your knowledge of OpenOCD, would it be able to support multi-core debugging?
Jul 04 11:35:28 <wookey> but we need some high-speed hardware and tools to the sort of thing XJTAG lets us do, but without everyone having to pay $18,000
Jul 04 11:35:50 <wookey> My knowledge is negligible, except for a talk vmaster gave at FOSDEM this year
Jul 04 11:36:08 <rwhitby> oh, is there an online copy of that talk somewhere I can look at?
Jul 04 11:36:24 <wookey> if openOCD and similar tools are the right sort of thing then it makes sense for us to develop that rather than NIH it
Jul 04 11:36:35 <rwhitby> wookey: my feelings exactly
Jul 04 11:36:55 <ka6sox> I think I saw a copy of the talk at one time....
Jul 04 11:37:03 <ka6sox> but I may be misteaken
Jul 04 11:37:21 <wookey> it was actually about openjtag. I'm not sure if openjtag==openOCD
Jul 04 11:37:30 <wookey> probably
Jul 04 11:37:48 <wookey> rwhitby: fodem.org seems to be down right now
Jul 04 11:37:49 <rwhitby> aha: http://free-electrons.com/community/videos/conferences
Jul 04 11:37:54 <wookey> look for the embedded track
Jul 04 11:37:56 <rwhitby> (from http://lwn.net/Articles/173968/)
Jul 04 11:38:05 <wookey> ah yes - he videoed it all - very handy
Jul 04 11:39:05 <ka6sox> http://openocd.berlios.de/OpenOCD_Fosdem2006_talk.odp
Jul 04 11:39:47 <ka6sox> those are the slides
Jul 04 11:40:26 <rwhitby> ka6sox: thx
Jul 04 11:40:49 <ka6sox> np
Jul 04 11:42:29 <rwhitby> also ftp://ftp.belnet.be/pub/mirror/FOSDEM/
Jul 04 11:46:06 <ka6sox> excellent
Jul 04 11:46:34 <rwhitby> Now I need to find a windows ogg player so i can watch them at work :-)
Jul 04 11:46:46 <vmaster> heh
Jul 04 11:46:55 <vmaster> the talk is more about how openocd works, than how to use it
Jul 04 11:47:04 <rwhitby> that's good.
Jul 04 11:47:18 <ka6sox> in order to port it to another arch thats what is needed.
Jul 04 11:47:23 <rwhitby> (especially for someone considering contributing to it)
Jul 04 11:47:33 <vmaster> mhh, not sure, it's a bit arm7/9 centric
Jul 04 11:47:56 <wookey> vmaster: what is the strongarm suport like?
Jul 04 11:48:04 <vmaster> there is no strong
Jul 04 11:48:06 <vmaster> arm support
Jul 04 11:48:18 <vmaster> the slide just shows the arm cores
Jul 04 11:48:44 <vmaster> ah, and "coming along nicely" means i got the basics working, but all the details have to be filled in
Jul 04 11:49:11 <wookey> OK, thanx. I will try to find time to take a proper look
Jul 04 11:49:42 <vmaster> the svn branch is horribly outdated - i'll let you know once there's something usable
Jul 04 11:50:27 <vmaster> currently i'm trying to get the OpenOCD build a native-windows version, without too many hacks
Jul 04 11:50:50 <rwhitby> vmaster: you don't work in the svn repo?
Jul 04 11:51:11 <vmaster> well, yeah, but I lack the discipline
Jul 04 11:51:36 <vmaster> sure, i could check everything in the way it is now
Jul 04 11:52:22 <vmaster> i try to do only small, documented changes on the /trunk
Jul 04 11:53:34 <vmaster> the xscale branch gets huge updates once i have all the bits working together
Jul 04 11:54:12 <rwhitby> Has the underlying architecture changed?  i.e. could someone use the current trunk as the basis to port to a new type of processor?
Jul 04 11:54:57 <vmaster> some very minor changes to the JTAG support, as XScale is a bit picky on how you move through the state machine
Jul 04 11:56:41 <vmaster> multi-core debugging should be possible, at least Laurent from Amontec reported that he used it on a board with two arm cores
Jul 04 11:57:12 <vmaster> davidc__: was having problems though, with his IPod, but then I know nothing about how these cores are connected
Jul 04 11:57:19 <rwhitby> heh, I was just talking to Laurent from Amontec on email tonight, asking for hardware donations to nslu2-linux ;-)
Jul 04 11:57:52 <vmaster> Olimex is going to sell something similar
Jul 04 11:58:05 <vmaster> you might want to ask them, too
Jul 04 11:59:58 <ka6sox> OLIMEX==Sparkfun in US.
Jul 04 12:00:07 <ka6sox> OLIMEX is in .bg.
Jul 04 12:00:13 <vmaster> yeah
Jul 04 12:00:35 <vmaster> I met Tsvetan (olimex) at the Embedded World this year in nuremberg
Jul 04 12:01:31 <ka6sox> I love nuremberg...spend a couple of romantic nights there.
Jul 04 12:02:01 <vmaster> yeah, the inner-city is really nice
Jul 04 12:02:32 <rwhitby> vmaster: olimex's product is this one?: http://www.olimex.com/dev/arm-usb-ocd.html
Jul 04 12:02:40 <vmaster> yeah
Jul 04 12:03:05 <vmaster> it's supposed to sell for around 60$
Jul 04 12:03:31 <vmaster> 1/3 of the amontec device
Jul 04 12:04:04 <ka6sox> still uses the FT2232?
Jul 04 12:04:15 <vmaster> i think so, although i haven't seen it yet
Jul 04 12:04:28 <vmaster> but he asked me how they should do the pin-out to have it work with OpenOCD
Jul 04 12:04:44 <vmaster> Amontec is planning to do a cheaper version, too
Jul 04 12:04:54 <ka6sox> k
Jul 04 12:06:13 <wookey> does openOCD support multiple JTAG chains?
Jul 04 12:06:34 <vmaster> mhh, not multiple chains, but multiple devices in a chain
Jul 04 12:06:52 <wookey> I recall DB saying the reason he didn't use any of the existing projects was that none supported multiple chains
Jul 04 12:07:06 <wookey> right - we put each of our devices on it's own port, we don't chain them
Jul 04 12:07:23 <wookey> s/it's/its/
Jul 04 12:07:24 <purl> wookey meant: right - we put each of our devices on its own port, we don't chain them
Jul 04 12:07:33 <vmaster> hmm, when do you have to operate multiple chains at the same time?
Jul 04 12:07:41 <vmaster> ah, board testing?
Jul 04 12:08:07 <wookey> and board programming - we nee to fill nand via CPU and FPGA on second chain
Jul 04 12:08:26 <wookey> and yes if there is a problem we want to monitor/control both
Jul 04 12:09:32 <wookey> separate ports is much simpler for new board bringup - easier to tell which bit is working/broken
Jul 04 12:10:00 <ka6sox> we are doing that a lot with our boards these days.
Jul 04 12:10:07 <wookey> does this concept break openOCD architecture badly, or could it be fitted in without too much trouble?
Jul 04 12:10:47 <vmaster> i think that could be added quite easily
Jul 04 12:11:01 <wookey> OK - that would make us much more enthusiastic :-)
Jul 04 12:11:18 <wookey> I can send you some hardware to play with if it helps
Jul 04 12:11:37 <wookey> (after the exams :-)
Jul 04 12:11:46 * rwhitby thinks vmaster will not pass his exams with all this hardware coming his way ...
Jul 04 12:12:12 <wookey> he only needs to do well enough not to get kicked out...
Jul 04 12:12:13 <ka6sox> he'll be thinking of dancing hardware.
Jul 04 12:13:08 <vmaster> heh, these are my last three exams ever i guess
Jul 04 12:14:03 <rwhitby> gotta go - sick child.  Catch you guys later.
Jul 04 12:14:10 <vmaster> cya
Jul 04 12:14:32 <ka6sox> bye for now
Jul 04 12:14:39 <ka6sox> time for sleep at 5:30 am
Jul 04 12:14:45 <vmaster> tonight's worldcup semi-final, so i won't be online anymore in a few hours
Jul 04 12:14:49 <vmaster> good night
Jul 04 12:21:09 <vmaster> wookey: regarding the multiple chains - do you think you could get your JTAG guy in here?
Jul 04 12:21:43 <wookey> I did tell him to come here, but he's not much of an IRC person.
Jul 04 12:21:59 <wookey> he did appear in #balloon so he has got it working...
Jul 04 12:22:11 <wookey> I'll reprod
Jul 04 14:17:24 <kay2>  http://www.vizufonusa.com/?topic=news&resource=CIP_6000
Jul 04 14:17:31 <kay2> it says it uses an arm920T
Jul 04 14:17:40 <kay2> as long as I know EP9302 is 920t
Jul 04 15:11:24 <wookey> vmaster: I mailed the list and cc:ed you. Hopefully someone who knows what they are talking about will turn up here at some point
Jul 04 15:16:09 <vmaster> nice, thank you
Jul 04 15:31:04 <vmaster> wookey: the StrongARM doesn't have debug support at all
Jul 04 15:31:23 <vmaster> at least not the SA1100 i have on my lart
Jul 04 15:34:04 <p2-mate> none of the strongarms has debug support afaik
Jul 04 15:34:46 <vmaster> yeah, i think so, but only know about the sa1100 for sure
Jul 04 15:35:12 <p2-mate> SA1110 doesn't either :)
Jul 04 15:37:52 <vmaster> wookey: and OpenOCD is more about debugging the code than the board - reading your mail I'm not sure if this is what you want
Jul 04 16:37:04 <wookey> vmaster: I see - so this is about using the built-in debug functionality on the CPU, not about using JTAG for simple programming (or complicated testing)
Jul 04 16:37:37 <wookey> As you say strongarm does not have that functionality in hardware
Jul 04 16:38:05 <wookey> So which software does have the functionality I am looking for?
Jul 04 16:41:28 <vmaster> well, it is about programming, too, but currently not about testing
Jul 04 16:41:36 <vmaster> i don't think there's free software
Jul 04 16:41:45 <vmaster> we've discussed this here before, but we weren't even sure what's necessary
Jul 04 16:43:26 <wookey> OK, so maybe our bflash could be a start others can use?
Jul 04 16:44:24 <wookey> are you going to answer mail to point out that I have wrong end of stick, or shall I do it?
Jul 04 16:44:38 <wookey> I expect you are about to watch footie match :-)
Jul 04 16:50:28 <vmaster> is your bflash available online somewhere?
Jul 04 16:58:30 <wookey> Not yet - I am hassling dave bisset to send it to me, and it will go on balloonboard.org
Jul 04 17:02:57 <vmaster> okay, I'll reply to your mail with what OpenOCD is, and what it's not
Jul 04 17:44:36 <davidc__> vmaster: I'm thinking its either that the debug macrocells in my board are nonstandard, or I've got some bad electrical noise on the signal lines, or that the thing is resetting itself in the middle of debug [I've seen some fluctuations in the power supply chunk of the board that looks dodgy]
Jul 04 17:44:44 <davidc__> vmaster: so don't worry about my issue too much for now ;)
Jul 04 17:46:53 <vmaster> davidc__: i have a fix for the two-targets problem
Jul 04 17:46:59 <vmaster> i forgot an i++ ;)
Jul 04 17:47:16 <davidc__> vmaster: heh, I was just gonna look at that today [ just got back from vacation]
Jul 04 17:47:21 <davidc__> but yeah, I'll do an svn up and grab that
Jul 04 17:48:06 <vmaster> it's not yet in svn, as i'm fighting the mingw stuff
Jul 04 17:48:43 <vmaster> in gdb_server, gdb_init(), add "i++;" at the end of the while loop, just before or after the target = target->next
Jul 04 17:50:15 <davidc__> alrighty, sounds good
Jul 04 17:52:09 <vmaster> okay, i'm off, watching the worldcup semi-finals
Jul 04 17:55:29 <ka6sox> enjoy!
**** ENDING LOGGING AT Wed Jul 05 02:59:56 2006