**** BEGIN LOGGING AT Tue May 24 23:59:57 2005
May 25 03:57:05 <ep1220> [g2] I have a question on the NSLU2
May 25 03:57:58 <[g2]> ep1220, I'm all ears
May 25 03:58:22 <ep1220> I now have one here, opened.
May 25 03:58:43 <ep1220> For the cache Dld I need to control the Reset of the Xscale
May 25 03:59:00 <ep1220> In the Wiki I saw the rest button is connected to a GPIO?
May 25 03:59:12 <ep1220> DO You know more details on this ?
May 25 04:00:34 <dyoung-zzzz> what sort of info are you looking for?
May 25 04:00:51 <[g2]> ep1220, do you know which Reset pin on the chip ?  I mean the actual letter/number pair from the BGA pin out ?
May 25 04:01:32 <dyoung-zzzz> [g2] we should probably point him to the pictures of the bare board.
May 25 04:02:22 <[g2]> dyoung-zzzz, nod.  I was also going to let him know we've got the bare board and Allan can do traces
May 25 04:02:32 <dyoung-zzzz> http://www.nslu2-linux.org/wiki/Info/PhotosOfTheInternals
May 25 04:03:01 <dyoung-zzzz> The ones labelled "bare" are the board with no components.
May 25 04:03:50 <ep1220> I will look at the pics. And the datasheet for the Reset Ball number
May 25 04:04:22 <dyoung-zzzz> we shoudl ask allan to slap the thing on a scanner.
May 25 04:04:43 <dyoung-zzzz> assuming its flat enough....
May 25 04:08:18 <ep1220> scanning would be very good
May 25 04:08:55 <dyoung-zzzz> I dont seem to have the full ds handy atm.
May 25 04:10:08 <[g2]> I was thinking of getting a board x-ray'ed
May 25 04:10:36 <ep1220> I need the RESET_IN_N: AC13
May 25 04:12:48 <[g2]> Trevor's got a singleton
May 25 04:13:13 <[g2]> google "RESET_IN_N nslu2" and you get only his page
May 25 04:13:17 <[g2]> Tiersten_
May 25 04:14:42 <[g2]> actaully you don't even need the nslu2 to get that page :)
May 25 04:15:49 <dyoung-zzzz> it might be impratical to trace it out using pictures only.
May 25 04:16:16 <[g2]> yes we should just send a message to Allan
May 25 04:16:39 <[g2]> He's been in #nslu2-linux the past couple days
May 25 04:16:53 <[g2]> I'll send him an e-mail now
May 25 04:17:08 <dyoung-zzzz> ep1220, so the idea is to hold the xscale device in reset while you load things into the cache?
May 25 04:17:14 <ep1220> [g2]: Thanks.
May 25 04:17:39 <ep1220> dyoung-zzzz: Yes. Must be done this way.
May 25 04:17:51 <ep1220> else the CPU might interfere.
May 25 04:18:00 <dyoung-zzzz> Cool.
May 25 04:18:19 <dyoung-zzzz> For some reason I thought holding the CPU in Reset would interfere with JTAG operations.
May 25 04:18:33 <ep1220> yes it does. let me explain.
May 25 04:18:33 <dyoung-zzzz> But I'm pretty much a complete amateur.
May 25 04:18:50 <ep1220> 1) assert RESET_IN.
May 25 04:19:09 <ep1220> 2) write to a special JTAG register and set a "keep-reset" bit
May 25 04:19:22 <ep1220> 3) release RESET_IN
May 25 04:19:43 <ep1220> 4) now the CPU core is kept in reset, but other parts work
May 25 04:20:07 <dyoung-zzzz> I see.
May 25 04:20:23 <[g2]> 5) make changes
May 25 04:20:30 <[g2]> 6) release "keep-reset"
May 25 04:20:41 <[g2]> ?
May 25 04:20:49 <ep1220> [g2]: I see You are ready to help here :-)
May 25 04:21:04 <[g2]> I'm guessing there ?
May 25 04:21:20 <ep1220> is just as You said.
May 25 04:21:27 <[g2]> I'd assume that this is the same or similar feature that JTAG debugging uses
May 25 04:21:39 <dyoung-zzzz> this benefits us because we can twiddle cache bits faster than we can twiddle Flash bits?
May 25 04:22:11 <ep1220> the good is: You can add code wihout modifing flash or RAM
May 25 04:22:17 <[g2]> I'd think virtually immediately
May 25 04:22:18 <dyoung-zzzz> Okay can someone remind me what TRST is?
May 25 04:22:33 <ep1220> TRST is used to reset the JTAG engine
May 25 04:22:44 <ep1220> Does not affect CPU
May 25 04:22:55 <dyoung-zzzz> Got it.
May 25 04:23:28 <dyoung-zzzz> so actually, the twiddling of the CPU RESET_IN_N is not directly related to a JTAG Function.
May 25 04:23:47 <dyoung-zzzz> so we'd need an additional I/O pin to twiddle it.
May 25 04:24:02 <ep1220> dyoung-zzzz: yes
May 25 04:24:05 <dyoung-zzzz> (I'm trying to understand the whole picture, sorry to be quizzing you)
May 25 04:24:15 <ep1220> dyoung-zzzz: np
May 25 04:24:27 <dyoung-zzzz> as I said I'm pretty much a complete amateur.
May 25 04:24:30 <dyoung-zzzz> :-)
May 25 04:27:39 <[g2]> ep1220, is the keep-reset documented somewhere ? Like and Intel app note ?
May 25 04:27:49 <[g2]> Like an Intel ...
May 25 04:29:24 <dyoung-zzzz> how does one twiddle the cache?  During the CPU being "keep-reset-ed" is the BSR no longer huge?  Or am I just confused?
May 25 04:30:41 <[g2]> I'd guess the BSR doesn't change
May 25 04:31:14 <[g2]> We're not changing the TAP controller... just hold the logic in the CPU from getting clocked
May 25 04:31:21 <[g2]> holding
May 25 04:31:31 <dyoung-zzzz> Right....
May 25 04:31:47 <dyoung-zzzz> so twiddling the cache is still a 498b operation?
May 25 04:31:51 <[g2]> But I'm a *TOTAL AMATEUR* and just guessing
May 25 04:31:59 <ep1220> g2: it is in the Xscale core manual/debugging section
May 25 04:32:13 <[g2]> ep1220, cool THX!
May 25 04:32:19 <ep1220> Twiddling the chace is a 33bit/Word operation
May 25 04:32:34 <ep1220> You do not go thru the JTAG scan chain.
May 25 04:32:48 <prpplague> dyoung-zzzz: lots of times if your flash is blank and the processor starts up and reads a code of 0xffffffff which is an invalid arm code, the system will lock up
May 25 04:32:59 <ep1220> You select special registers by using "private" JTAG instructions
May 25 04:33:14 <prpplague> dyoung-zzzz: to prevent this you just hold the reset while running jtag 
May 25 04:33:38 <dyoung-zzzz> Aha, I see!  I didnt read enough of TFM yet!
May 25 04:34:38 <dyoung-zzzz> It all makes sense now.
May 25 04:35:11 <dyoung-zzzz> I think thats what Teirsten was suggesting a while back, but I didnt know what he was saying.
May 25 04:36:32 <[g2]> ep1220, So writing to the cache lines is a much shorter operation AND you probably get an immediate response as opposed to waiting for the Flash to program correct ?
May 25 04:36:51 <ep1220> g2: yes
May 25 04:37:16 <[g2]> or we may be able to stream writes back-to-back to the cache
May 25 04:37:44 <ep1220> g2: You write Adress + 8 data words (1 cache line)
May 25 04:38:09 <ep1220> as fast as te JTAG PC IF can do. Back -to-back.
May 25 04:38:15 <dyoung-zzzz> ep1220 do you have a good link/manual for JTAG?  these private instructions sound interesting.
May 25 04:38:16 <[g2]> heh... That sounds logical :)  Just like a cache line fill 
May 25 04:38:51 <[g2]> Right the onboard cache can run at 266Mhz
May 25 04:39:05 <ep1220> dyoung-zzzz: the private instructions depend on the CPU
May 25 04:39:14 <ep1220> Xscale different than ARM than MIPS
May 25 04:39:40 <ep1220> Usually You find them in the manuals
May 25 04:39:43 <[g2]> that's part of the reason the JTAG debugger cost money
May 25 04:39:49 <dyoung-zzzz> I see it. LDIC, etc....
May 25 04:40:04 <ep1220> dyoung-zzzz: exactly
May 25 04:40:51 <[g2]> dyoung-zzzz, what's the pub number for that (I've got so many doc's around :)
May 25 04:41:06 <dyoung-zzzz> 25248005.pdf
May 25 04:41:11 <[g2]> THX
May 25 04:41:17 <dyoung-zzzz> its the developer guide/manual
May 25 04:44:27 <dyoung-zzzz> HIGH_Z sounds like it might be of interest when progrogramming a "personality prom" In Circuit.
May 25 04:46:16 <ep1220> dyoung-zzzz: afaik, HIGH-Z disables all drivers;
May 25 04:46:18 <ep1220> Normally used in connection testing
May 25 04:46:26 <ep1220> is a standard JTAG cmd.
May 25 04:46:51 <dyoung-zzzz> This method sounds totoally doable.
May 25 04:47:03 <[g2]> ep1220, so HIGH_Z is used for Checking for shorts ?
May 25 04:47:24 <dyoung-zzzz> unless I'm hallucinating theres a via from the reset pin AC13
May 25 04:47:45 <ep1220> g2: asume You have FLASH/RAM/CPU all on the databus
May 25 04:48:18 <ep1220> You HiGHZ the CPU: no check FLASH-RAM (if both are in the JTAG chain)
May 25 04:48:56 <ep1220> Or You have an additional bed-of-nails tester, or ...
May 25 04:49:29 <[g2]> ep1220, I'm hoping to build some IXP4xx hw in the next couple months
May 25 04:49:50 <[g2]> I'm know a little about all these things, but I'm going to need to know more
May 25 04:50:00 * dyoung-zzzz me toos
May 25 04:50:06 <[g2]> and where I can cut cost and save $$
May 25 04:50:14 <dyoung-zzzz> I dont know anything about anything.
May 25 04:50:23 <[g2]> but still keep quality HIGH_Z
May 25 04:50:28 <[g2]> :)
May 25 04:50:28 <prpplague> ep1220: there is a really good book i got from amazon about using jtag for hardware testing, its proved very useful
May 25 04:50:44 <dyoung-zzzz> prpplague, Got a link?
May 25 04:50:55 <[g2]> no he's teasing us :)
May 25 04:50:57 <prpplague> dyoung-zzzz: grabing it now
May 25 04:51:14 <dyoung-zzzz> someone recently gave me a amazon GC, and I might as well spend it on something good instead of Yet Another Toy.
May 25 04:52:10 <ep1220> prpplague: is it HW test only, or also includes debugging ?
May 25 04:54:06 <[g2]> So LDIC and DCSR are our friends :)
May 25 04:54:16 <prpplague> http://www.amazon.com/exec/obidos/tg/detail/-/0750677309/qid=1117021996/sr=1-1/ref=sr_1_1/002-0698007-7396822?v=glance&s=books
May 25 04:54:27 <prpplague> ep1220: both software and hardware, but its not jtag specific
May 25 04:54:46 <prpplague> ep1220: mainly about procedures in testing
May 25 04:54:54 <dyoung-zzzz> Wow, thats some coin!
May 25 04:55:09 <prpplague> dyoung-zzzz: ??
May 25 04:55:23 <dyoung-zzzz> Maybe not too bad.
May 25 04:55:35 <dyoung-zzzz> I thought it said $85 for a second there.
May 25 04:55:44 <prpplague> oh, yea, its only 50
May 25 05:02:23 <[g2]> hey Tiersten did you catch what ep1220 was saying about directly loading the Icache ?
May 25 05:02:35 <ep1220> Is also available as e-book. Funny, costs more than the paperback.
May 25 05:11:42 <ep1220> [g2]: Do You have "public" specs of Your planned boards ?
May 25 05:12:44 <[g2]> ep1220, It's only me. I'll be working with hw designers/team and outsourcing the contract manufacturing
May 25 05:13:06 <[g2]> Initially I'm starting with re-worked items
May 25 05:13:49 <[g2]> So the first couple dozen/hundred units will be aimed at developers
May 25 05:14:22 <ep1220> i C, for routers/small servers/.. (?)
May 25 05:14:46 <[g2]> I see several different markets
May 25 05:15:25 <[g2]> a) developers/hobbyists, b) fanless routers/servers, c) infrastructure
May 25 05:16:29 <[g2]> I think just servicing market a) would be enough
May 25 05:16:36 <dyoung-zzzz> Webcam glue device.
May 25 05:16:42 <[g2]> and b and c are bonus items
May 25 05:16:50 <ep1220> You do complete kit, including case ?
May 25 05:17:28 <[g2]> I'm talking to several mechanicals people now
May 25 05:17:47 <[g2]> the case will probably drive the board layout
May 25 05:18:09 <dyoung-zzzz> Made in USA?
May 25 05:18:23 <[g2]> dyoung-zzzz, to start
May 25 05:18:31 <dyoung-zzzz> Cool.
May 25 05:19:03 <[g2]> I could be wrong but the shipping/time to go to asia is more expensive than the boards :)
May 25 05:19:47 <dyoung-zzzz> I think youre wrong about that.
May 25 05:20:19 <dyoung-zzzz> Injection Moulding is ludicrously cheap in china/se asia.
May 25 05:20:39 <dyoung-zzzz> but I have fundemental problems with stuff made in Slave Labour Camps.
May 25 05:20:51 <[g2]> well maybe you have connections there. I don't so it's at least weeks or months to sort out the connections
May 25 05:21:58 <[g2]> well last I read the Chinese make $200 a year and save about 30%
May 25 05:22:24 <dyoung-zzzz> We can talk when youre ready.
May 25 05:23:05 <[g2]> Do you know IM moulding ppl over there ?
May 25 05:23:25 <dyoung-zzzz> Joe Manual labourer makes around $6-$8/day.
May 25 05:23:53 <dyoung-zzzz> I can ask around; but I seem to remember that one of my contacts does plastics as a contractor to HP.
May 25 05:25:05 <[g2]> A cheap case is important
May 25 05:25:37 <[g2]> it's actually the biggest roadblock right now
May 25 05:25:56 <dyoung-zzzz> mail me some prelim size specs and I'll go poking around.
May 25 05:25:58 <[g2]> But I'm looking into other alternatives
May 25 05:26:38 <[g2]> Right now I'm starting on working to put together a costed BOM
May 25 05:26:38 <ep1220> [g2]: Maybe use a Mini-ITX case (?)
May 25 05:26:44 <dyoung-zzzz> stamped aluminium will reduce EMI
May 25 05:26:55 <[g2]> ep1220, That's one option
May 25 05:27:04 <[g2]> but those cases are expensive
May 25 05:27:17 <[g2]> $60-70 US
May 25 05:28:31 <ep1220> Street-price in the US ?
May 25 05:28:56 <[g2]> I think for $70 Joe Laborer will bend custom cases
May 25 05:29:44 <ep1220> but likely You need to commit to a larger number.
May 25 05:29:50 <[g2]> Allan does welding for racing cars I'm sure he'd know some connections or even be able to do something like that himself
May 25 05:30:07 <[g2]> ep1220, that's the catch-22
May 25 05:30:29 <dyoung-zzzz> Oops, did I say day?  That should be $6/week.
May 25 05:30:53 <[g2]> cases cost $70 in low-quantity, but for a $15-25K stamp they can be $7
May 25 05:31:01 <dyoung-zzzz> anyways, yes.  overseas manufacturing is cheep.
May 25 05:31:03 <[g2]> so it's a case of volumne
May 25 05:31:28 <ep1220> g2: always :-( - and thus a question of money needed upfront
May 25 05:31:33 <[g2]> Right but I may be able to buy someone elses cases that are already in volume
May 25 05:31:49 <dyoung-zzzz> You ought to be able to fit the device in some commodity case.
May 25 05:32:02 <[g2]> ep1220, it's really not so bad.
May 25 05:32:02 <dyoung-zzzz> or use one of these extruded alumnium cases.
May 25 05:32:12 <dyoung-zzzz> and select a boardsize appropriately.
May 25 05:32:19 <dyoung-zzzz> then the only custom part is the end pieces.
May 25 05:32:40 <[g2]> dyoung-zzzz, right that's the plan right now
May 25 05:33:12 <dyoung-zzzz> and I know for sure you can get aluminum stampins in "small/medium" quanities for cheap.
May 25 05:33:16 <[g2]> I'm looking to make a really affordable device
May 25 05:33:53 <[g2]> I think initially the options will be a) No case and b) and expensive case
May 25 05:34:00 <[g2]> or relatively expensive
May 25 05:34:03 <[g2]> $40
May 25 05:34:15 <dyoung-zzzz> I think I would disagree....
May 25 05:34:31 * [g2] would love to be proved wrong :)
May 25 05:34:47 <[g2]> what's the disagree part ?
May 25 05:34:52 <dyoung-zzzz> But it kind of depends on the volume.
May 25 05:35:07 <[g2]> well I'm talking the first 50 units
May 25 05:35:12 <[g2]> or maybe the first 200 units
May 25 05:35:15 <dyoung-zzzz> theres a significant price break on the PCB manufacture / layout.
May 25 05:35:22 <[g2]> I really depends on the upfront costs
May 25 05:35:31 <dyoung-zzzz> so if youre gonna make 10 with expensive case vs 100 with cheap case.....
May 25 05:35:56 <[g2]> I don't think layout will be bad at all
May 25 05:36:03 <[g2]> you know the size of the slug
May 25 05:36:04 <dyoung-zzzz> you might as well sell the first 10 of the "cheap case" with no case.
May 25 05:36:14 <dyoung-zzzz> true.
May 25 05:36:38 <[g2]> dyoung-zzzz, that's exactly what I said: "a) no case .... b) expensive case"
May 25 05:36:50 <dyoung-zzzz> oh yeah.
May 25 05:36:56 * dyoung-zzzz duhs.
May 25 05:37:15 <dyoung-zzzz> in that case, (a) I think.
May 25 05:37:17 <[g2]> initially, I'll lose sales, but eventually it should work out
May 25 05:37:41 <[g2]> or I'll find c) a reasonable cheap case
May 25 05:38:00 <dyoung-zzzz> you  should be able to find a commodity case.
May 25 05:39:27 <dyoung-zzzz> Either way, I'm still gonna sign up for the  "Developers kit".
May 25 05:39:37 <[g2]> THX :)
May 25 05:41:02 <dyoung-zzzz> lemme find my extruded aluminum links and send um to you.  it might be a few days.  but its incredibly cheap for those.
May 25 05:41:03 <[g2]> ep1220, dyoung-zzzz I really want to find out about the RESET_IN_N line
May 25 05:41:22 <[g2]> dyoung-zzzz, that'd be helpful thx
May 25 05:41:31 <dyoung-zzzz> allan should be around soon.
May 25 05:41:42 <dyoung-zzzz> or we could just mail him....
May 25 05:41:50 <[g2]> nod. I sent him an e-mail this morning about 45 minutes ago
May 25 05:42:38 <[g2]> ok 75 minutes ago
May 25 05:42:45 <dyoung-zzzz> ~slugtime
May 25 05:42:47 <purl> hmm... slugtime is 40 hour days 10 hours in 4 timezones with overlap
May 25 05:43:22 <[g2]> I truly love working with such talented motivated people
May 25 05:43:47 <[g2]> I really hope to make a go of this hw thing and get a bigger company going 
May 25 05:44:40 * dyoung-zzzz me toos
May 25 05:45:09 <ep1220> I too like this idea :-)
May 25 05:45:50 <[g2]> My day-dream is to turn into the next Linksys except be all open
May 25 05:46:26 <dyoung-zzzz> Good Plan.
May 25 05:47:12 <dyoung-zzzz> I've started playing with my CoolRunner again.
May 25 05:47:45 <dyoung-zzzz> debatng how much work its gonna be to invent a "dos-like" system on it....
May 25 05:47:50 <dyoung-zzzz> is there a FOSS DOS?
May 25 05:47:58 <[g2]> think so
May 25 05:48:09 <[g2]> FreeDOS ??
May 25 05:48:19 <dyoung-zzzz> then slap a SD/MMC card on it....
May 25 05:48:25 <[g2]> one of the guys in the ML for the slug was responsible for it
May 25 05:48:29 <dyoung-zzzz> but thats a step for later.
May 25 05:48:32 <dyoung-zzzz> oh?
May 25 05:48:45 <[g2]> The guys that does all the 
May 25 05:48:53 <[g2]> distributed processing stuff at NIST
May 25 05:49:12 <dyoung-zzzz> That would be a interesting port.
May 25 05:49:13 <[g2]> He's been playing with slugs for last couple months
May 25 05:49:36 <dyoung-zzzz> maybe i'l; bug him when I get closer to actually doing it.
May 25 05:50:03 <[g2]> FREEBIOS maybe 
May 25 05:51:00 <dyoung-zzzz> I should google it first.  What I need is picoblaze/DOS.
May 25 05:51:29 <[g2]> I'd think you could just run QEMU on the slug
May 25 05:53:08 <dyoung-zzzz> I need the power characteristics of a CR2.
May 25 05:53:24 <dyoung-zzzz> althow low power, the slug still takes a few watts.
May 25 05:55:55 * jacques hears "QEMU"
May 25 05:56:59 <jacques> unfortunately it does not support system emulation on arm, only user :-(
May 25 05:57:20 <jacques> but one of the devs says it shouldn't take too much work to get it there
May 25 05:57:36 <jacques> it's missing the arm versions of certain functions
May 25 05:57:40 <[g2]> jacques, are you familiar with softgun at all ?
May 25 05:58:12 <jacques> [g2], not really, I've only looked at their site a few times to track their progress
May 25 05:58:40 <[g2]> one of the guys in gentoo-embedded had a howto on it or was writing one up
May 25 05:58:46 <[g2]> it's on the ml
May 25 05:58:54 <[g2]> gentoo-embedded ml
May 25 05:58:56 <[g2]> that is
May 25 05:59:14 <jacques> very interesting - no BE yet tho ?
May 25 06:00:03 <[g2]> dunno I haven't gotten around to that
May 25 06:00:34 <jacques> I started messing with qemu fairly recently when lennert added armeb support
May 25 06:05:32 <dyoung-zzzz> Has anyone seen lennert recently?
May 25 06:05:50 <dyoung-zzzz> I'm kinda wondering if that serial cable ever made its way into his hands.
May 25 06:05:55 <dyoung-zzzz> mailed him, but no reply.
May 25 06:07:39 <jacques> hmm I haven't seen him
May 25 06:25:14 <[g2]> kitno445 just arrived in nslu2-linux
May 25 06:30:55 <kitno455> hey, its all the same guys!
May 25 06:31:05 <[g2]> a bunch
May 25 06:31:09 <kitno455> someone here wants to talk to me?
May 25 06:31:11 <[g2]> ep1220, meet kitno455 
May 25 06:31:41 <[g2]> ep1220, is kick'n ass on JTAG programming with both the USB 2.0 part and the FTDI part
May 25 06:32:00 <kitno455> greets to you ep1220
May 25 06:32:08 <kitno455> explain these?
May 25 06:32:17 <dyoung-zzzz> "these" ?
May 25 06:32:30 <kitno455> 'USB 2.0 part and the FTDI part'
May 25 06:32:38 <kitno455> status, etc
May 25 06:32:42 <dyoung-zzzz> Oh.
May 25 06:32:42 <[g2]> nod.
May 25 06:32:54 <dyoung-zzzz> FTDI2232
May 25 06:33:07 <dyoung-zzzz> NEC controller chip found in nslu2
May 25 06:33:17 <kitno455> right, i know it
May 25 06:33:18 <[g2]> no.....
May 25 06:33:33 <kitno455> pic micro
May 25 06:33:35 <[g2]> the FTDI is a USB chip
May 25 06:33:51 <dyoung-zzzz> hehe
May 25 06:33:58 <dyoung-zzzz> this is too funny.
May 25 06:34:11 <[g2]> The CY68013 is the USB device chip used by the USB 2.0 adapter board for the Spartan 3 FPGA development board from digilent/xilinx
May 25 06:34:24 <dyoung-zzzz> it is true you can buy the FTDI2232 in small quanitity from PIC Micro.
May 25 06:34:37 <dyoung-zzzz> oh, THAT USB part.
May 25 06:35:02 <dyoung-zzzz> I guess we were talking about apples and oranges AND bananas.
May 25 06:35:14 <kitno455> which one you want to be?
May 25 06:35:14 <[g2]> So ep1220 has reprogrammed the USB 2.0 Cypress part on the S3 dev board AND is experimenting with the FTDI chip that can do JTAG
May 25 06:35:29 <dyoung-zzzz> I'm a banana.
May 25 06:35:32 <kitno455> hold on, the ftdi talks jtag?
May 25 06:35:45 <dyoung-zzzz> the 2232 talks jtag.
May 25 06:35:46 <ka6sox-zzzz> dyoung-zzzz, I bought my FTDI parts from parallax..the Basic Stamp people
May 25 06:35:50 <[g2]> The C rev of the part has a JTAG mode
May 25 06:36:00 <kitno455> is C rev in slug?
May 25 06:36:15 <[g2]> kitno455, time out
May 25 06:36:22 * dyoung-zzzz hides
May 25 06:36:36 <[g2]> We are talking about doing high-speed JTAG programming to the SLUG
May 25 06:36:52 <[g2]> none of these chips are ON the slug
May 25 06:36:53 <kitno455> ok, you building usb2.0 jtag programmer?
May 25 06:36:58 <[g2]> exactly
May 25 06:37:13 <kitno455> using cypress devel chip
May 25 06:37:20 <kitno455> and ftdi chip?
May 25 06:37:25 <dyoung-zzzz> OR FTDI 2232
May 25 06:37:30 <[g2]> two different approaches
May 25 06:37:44 <kitno455> ok, why the f*ck did you not say so to start with :)
May 25 06:37:49 <[g2]> 1) Spartan 3 FPGA and Cypress USB 2.0 device 
May 25 06:37:53 <[g2]> 2) FTDI chip
May 25 06:38:05 <kitno455> fpga for jtag
May 25 06:38:24 <ka6sox-zzzz> kinda circular eh?
May 25 06:38:45 <dyoung-zzzz> your typical FPGA BSR is small though.
May 25 06:39:04 <kitno455> is there enough code over at opencores to get a usb2.0 device side AND jtag into a single fpga?
May 25 06:39:23 <dyoung-zzzz> not quite, we'll need to write it.
May 25 06:39:26 <[g2]> The point of all this is that by toggleing that RESET pin we may be able to load the on-board instruction cache directly
May 25 06:39:54 <[g2]> ep1220, has it worked out except for getting  access to that pin
May 25 06:40:06 <[g2]> or something close to that
May 25 06:40:30 <[g2]> the something close to that referring to the "has it all worked out"
May 25 06:40:41 <kitno455> ok, well thats good cause i dont know jtag from lpc, but i can use a meter to follow traces
May 25 06:40:56 <[g2]> and that's the key help we need !
May 25 06:40:58 <dyoung-zzzz> ... but you can weld.
May 25 06:41:04 <ep1220> Hi kitno455,
May 25 06:41:05 <kitno455> yeah, i can weld
May 25 06:41:07 <[g2]> and bend sheet metal
May 25 06:41:12 <kitno455> salutations
May 25 06:41:22 <kitno455> and design hi-performance suspensions
May 25 06:41:31 <ep1220>  it would be great if You can help find a point on the NSLU2 were I can connect a wire to the RESET line
May 25 06:41:33 <ka6sox-zzzz> kitno455, yes there is plenty of room for both in a S3
May 25 06:41:55 <kitno455> ka6sox-zzzz, does this mean that there is a branch #3 of this project?
May 25 06:42:07 <kitno455> do it all in a single fpga?
May 25 06:42:11 <ka6sox-zzzz> 1)
May 25 06:42:16 <dyoung-zzzz> that was the original plan.
May 25 06:42:25 <dyoung-zzzz> but we got sidetracked.
May 25 06:42:42 <ka6sox-zzzz> ya
May 25 06:42:43 <kitno455> but #1 now uses cypress (which is fine, nice and easy)
May 25 06:43:03 <dyoung-zzzz> Except require $500 devkit
May 25 06:43:18 <kitno455> yeah, well, there is no free lunch
May 25 06:43:21 <ka6sox-zzzz> because the cypress chip is the USB2 device that allows programming.
May 25 06:43:33 <ka6sox-zzzz> Unintended side benefit.
May 25 06:44:00 <kitno455> ep1220, i intend to get you info about whatever traces you need tonight
May 25 06:44:07 <ka6sox-zzzz> the FTDI chip came up because it has drivers that we can play with.
May 25 06:44:08 <kitno455> dont have board now
May 25 06:44:17 <ep1220> great :-) :-)
May 25 06:44:25 <[g2]> kitno455, ep1220 THX 
May 25 06:44:40 <kitno455> i also owe tommy some info on sdram and flash traces
May 25 06:44:54 <ep1220> kitno455: I need to write some SW as well.
May 25 06:45:01 <ka6sox-zzzz> kitno455, I'm intersted in the same info here.
May 25 06:45:11 <ep1220> kitno455: So friday is early enough for me.
May 25 06:45:13 <kitno455> ok, i think there is a page on wiki
May 25 06:45:23 <kitno455> about pins i already found
May 25 06:45:35 <kitno455> i dont know if i can edit it, but if so, i will add to it
May 25 06:48:13 <kitno455> ok, have work to do will be back tonight or tomorrow morning.
May 25 06:48:22 <kitno455> adios
May 25 06:48:31 <[g2]> kitno455, I think dyoung or ka6sox was saying there might be a NC for the CS line
May 25 06:48:34 <[g2]> adius
May 25 06:48:36 <[g2]> adios
May 25 06:49:02 <ep1220> kitno455: bye
May 25 06:49:06 <dyoung-zzzz> night.
May 25 06:49:08 <kitno455> i will check, but i think CS lines on sdram were connected. but i will look agian
May 25 06:49:25 <ka6sox-zzzz> nite dyoung-zzzz 
May 25 07:05:08 <[g2]> Ok Nokia gets it and goes to end my HW world-domination plan :)
May 25 07:05:13 <[g2]> http://www.mobileburn.com/review.jsp?Id=1376
May 25 07:08:43 <prpplague> [g2]: hehe thats funny, i interviewed for a job at TI to do kdrive port to the omap1710
May 25 07:09:50 <prpplague> [g2]: by the time they offered me the job, i'd already accepted the one i am at now
May 25 07:10:28 <[g2]> neat
May 25 07:10:42 <[g2]> I'll bet you are happy where you're at !
May 25 07:11:58 <prpplague> [g2]: hehe, yea, more hardware experience
May 25 07:12:23 <prpplague> [g2]: i've always been more on the software side of things, so working directly with the hardware designers is nice
May 25 07:13:04 <[g2]> I've done that a couple jobs and I really thing the hw/sw connection is important
**** BEGIN LOGGING AT Wed May 25 09:05:45 2005
May 25 17:43:50 <[g2]> beewoolie-afk, HEY!
May 25 17:44:00 <beewoolie-afk> [g2]: greetings.
May 25 17:44:06 <beewoolie-afk> Been busy with moving a client.
May 25 17:44:19 <[g2]> moving a client ?
May 25 17:44:21 <beewoolie-afk> This is the first day when I wasn't in the machine room, burried beneath a pile of cat5.
May 25 17:44:28 <beewoolie-afk> Yeah.  They moved offices.
May 25 17:44:36 <[g2]> cool
May 25 17:44:38 <beewoolie-afk> How's the good fight?
May 25 17:45:01 <[g2]> it's not a good fight it's an ass-kicking on our side :)
May 25 17:45:14 * beewoolie-afk celebrates
May 25 17:45:59 <[g2]> you've been missing the big JTAG stuff
May 25 17:46:26 <[g2]> ep1220 has been doing lots
May 25 17:46:48 <beewoolie> What's the latest?  Do we have a piece of HW?
May 25 17:47:03 <[g2]> We should very soon
May 25 17:47:11 <beewoolie> Excellent.
May 25 17:47:13 <[g2]> The parts were supposed to be in yesterday
May 25 17:48:12 <[g2]> ep1220 has made good progress with both the USB 2.0 device chip for the S3 boards (he's reprogrammed the on-board eeprom)
May 25 17:48:19 <[g2]> and with the FTDI chip
May 25 17:48:40 <[g2]> We were discussing loading up the I cache directly from JTAG
May 25 17:48:50 <beewoolie> Interesting idea.
May 25 17:48:59 <[g2]> so we could put APEX in the i cache and then wake up the chip
May 25 17:49:06 <beewoolie> that means that we store a program in cache and then execute from there.
May 25 17:49:17 <beewoolie> Hey!  That has an awesome consequence!
May 25 17:49:20 <[g2]> kitno455 is checking if the RESET_IN_N line AC13 comes out
May 25 17:49:34 <beewoolie> One of the problems with writing to flash is that byte-by-byte programming is slow.
May 25 17:49:37 <[g2]> well should know by tomorrow morning
May 25 17:49:52 <[g2]> and it's only 33bits/word :)
May 25 17:50:09 <beewoolie> If we can write the data to the cache, then we can program flash without initializin SDRAM.
May 25 17:51:22 <[g2]> beewoolie, nod.
May 25 17:56:41 <[g2]> so beewoolie I've got a question for you
May 25 17:56:51 <beewoolie> otp...hold on
May 25 17:56:55 <[g2]> np
May 25 18:34:44 <beewoolie> ok.  back.
May 25 18:34:47 <beewoolie> Whatsup
May 25 18:34:51 <beewoolie> [g2]: ping.
May 25 18:35:05 <[g2]> pong
May 25 18:35:08 <beewoolie> BTW, I like this writing to cache idea.
May 25 18:35:22 <beewoolie> It's great for machines that have separate I and D cache.
May 25 18:35:34 <[g2]> well we'll find out tomorrow if the line comes out
May 25 18:35:38 <beewoolie> It does mean we have to setup the MMU on most systems.
May 25 18:35:56 <beewoolie> Oh.  I see.  This is an issue for the scan chain?
May 25 18:36:06 <beewoolie> BTW, what is 33bits/word?
May 25 18:36:29 <[g2]> instead of 498 I think
May 25 18:36:41 <beewoolie> It's a separate chain?
May 25 18:36:41 <[g2]> it doesn't use the BSR
May 25 18:37:01 <beewoolie> Hmm.  Interesting.
May 25 18:37:12 <[g2]> LDIC
May 25 18:37:25 <[g2]> and DCSR
May 25 18:37:45 <[g2]> Load Instruction cache and Debug Control Status Register
May 25 18:37:59 <beewoolie> I need to get back on the telephone here.  Is there something quick you want to ask?
May 25 18:38:10 <[g2]> we can talk later
May 25 18:38:13 <beewoolie> OK.
May 25 18:38:15 <[g2]> or I'll send e-mail
May 25 18:38:16 <[g2]> cheers
May 25 18:38:24 <beewoolie> ttfn
May 25 18:39:12 <[g2]> beewoolie-afk, FYI see chapter 20 in the Intel Developer's Manual Pub 25248005.pdf
May 25 18:39:25 <beewoolie-afk> thx.
May 25 18:39:39 <[g2]> go yak on the phone!
May 25 18:39:43 <beewoolie-afk> already am
May 25 20:57:19 <[g2]> beewoolie-afk, hey
May 25 21:24:13 <ka6sox-away> ping pang pong
May 25 21:25:52 <dyoung-web> gu go a
May 25 21:26:10 <ka6sox-away> he he
May 25 21:27:03 <ByronT> that should have been "gu go ga" or even more correct would have "gi ga go"
**** ENDING LOGGING AT Wed May 25 23:59:56 2005