**** BEGIN LOGGING AT Mon Sep 17 02:59:57 2007
Sep 17 18:19:06 <likewise> drath: ping. The cfi.patch works for me (tm).
Sep 17 18:20:36 <drath> hey likewise 
Sep 17 18:20:40 <drath> cool, thanks for letting me know
Sep 17 18:21:21 <likewise> drath: well, thanks for helping review the code on this.
Sep 17 18:21:53 <likewise> drath: BTW, any experience on what the fastest USB dongle is?
Sep 17 18:22:16 <likewise> I'm getting 30 kBytes/second with the Amontec/JTAGkey to a IXP420@533MHz.
Sep 17 18:22:29 <likewise> while flashing using workspace
Sep 17 18:48:09 <drath> they're all based on the FT2232, so speeed is the same
Sep 17 18:49:25 <drath> likewise: are you using libftdi? or the binary-only FTD2XX?
Sep 17 18:49:29 <likewise> drath: ok, so only a parallel port (or Ethernet one) can do better
Sep 17 18:49:37 <likewise> drath: binary-only under Cygwin
Sep 17 18:50:06 <drath> likewise: the Amontec JTAG Accelerator configuration is the fastest interface
Sep 17 18:50:17 <likewise> drath: I needed to create a production programming tool without losing our BDI2000.
Sep 17 18:50:34 <likewise> drath: but of course we will now see to use OpenOCD for debugging as well.
Sep 17 18:51:04 <likewise> drath: a next processor is a PowerPC though, so we need the BDI anyway
Sep 17 18:53:48 <drath> i've never had a chance to evaluate xscale performance, because the only xscale hardware i've got has a 200khz low-pass on TCK
Sep 17 18:56:43 <drath> but the transfer should be optimized already
Sep 17 18:57:01 <drath> did you measure download to ram speed?
Sep 17 18:57:20 <likewise> drath: no, flash speed (as reported by openocd when finished, after 16 MB)
Sep 17 18:57:35 <likewise> drath: will try to measure the download to RAM speed tomorrow
Sep 17 18:58:41 <drath> likewise: yeah, would be nice to know where the time is being spent
Sep 17 18:59:02 <likewise> drath: probably in windows :-)
Sep 17 18:59:28 <drath> likewise: running with -d slows down a lot, too - -d -l <logfile> improves speed by outputting to a log file instead of console
Sep 17 19:00:03 <likewise> didn't do any debugging/logging during measurement
Sep 17 19:00:26 <drath> ok
Sep 17 19:01:36 <drath> if RAM download turns out to be a lot faster we could increase the maximum buffer size, to reduce the overhead of halting-resuming
Sep 17 19:03:48 <drath> currently only 32kb are used - if you're flashing images of several mb the target needs to be halted and resumed quite often
Sep 17 19:37:00 <likewise> drath: ok, yes I see.
Sep 17 20:28:07 <fluff> anyone having trouble with jtagkey-tiny and the latest openocd?
Sep 17 20:31:35 <fluff> the interface is proving 3V, but there's nothing happening on any of the lines, such as TCK?
Sep 17 20:33:22 <prpplague> fluff: latest build using the flyswatter board works, so i'd guess its either your hardware or something specific to that board layout in openocd
Sep 17 20:34:16 <prpplague> fluff: got me on /ignore over in #h1940 ?
Sep 17 20:35:22 <fluff> hmm, am i still on that channel? my screen session froze and i've relogged in most places
Sep 17 20:36:44 <prpplague> fluff: yep
Sep 17 20:38:20 <fluff> hmm, i know the jtag connection is good, the parallel port verison works
Sep 17 20:38:42 <fluff> i've pulled the unused nRST to VCCtarget by 4K7
Sep 17 20:38:48 <fluff> and still the thing does f'all
Sep 17 20:38:57 <prpplague> hmm
Sep 17 20:39:11 * prpplague looks up the jtag-key-tiny
Sep 17 20:39:58 <fluff> wonder if it is just too low for it work properly, Vtarget is just under 3V
Sep 17 20:40:49 <prpplague> thats possible, i don't see a schematic available
Sep 17 20:41:05 <prpplague> it does say it supports as low as 2v
Sep 17 20:43:22 <prpplague> fluff: so Vref is set to the correct voltage?
Sep 17 20:44:31 <prpplague> fluff: the tiny is 100% usb powered correct?
Sep 17 20:45:28 <fluff> dunno
Sep 17 20:45:30 <prpplague> fluff: might be pulling too much current
Sep 17 20:45:39 <fluff> my scope is a little inaccurate, Vref is 2.8V
Sep 17 20:46:35 <prpplague> fluff: the board you are working with, i assume you have a seperate power supply for it? and not using the tiny to power the board?
Sep 17 20:50:54 <drath> fluff: the jtagkey-tiny page says Wide target
Sep 17 20:50:54 <drath> voltage range
Sep 17 20:50:56 <drath> 2.8V to 5.0V
Sep 17 20:51:10 <drath> not sure how good it is at the lower end of supported voltages
Sep 17 20:51:41 <prpplague> 2.8 or 2.0 ?
Sep 17 20:51:49 <drath> 2.8
Sep 17 20:52:00 <prpplague> hmm
Sep 17 20:52:05 <drath> "All JTAG signals are 5V down to 2.8V tolerant"
Sep 17 20:52:05 <prpplague> sparkfun's page says 2.0
Sep 17 20:52:14 <fluff> grr
Sep 17 20:52:17 <drath> oh, sparkfun has the jtagkey-tiny?
Sep 17 20:52:31 <fluff> for what the schematics say is 3.3V, is only reading 2.8V at the board end
Sep 17 20:52:34 <prpplague> supports ARM targets working in voltage range 2.0 \x{2013} 5.0 V DC
Sep 17 20:52:45 * fluff feels the urge to find some hardware people and make some hurtings
Sep 17 20:53:08 <prpplague> hehe
Sep 17 20:53:12 <prpplague> fluff: join the crowd
Sep 17 20:53:37 <prpplague> fluff: btw, i really need to talk to you about the s3c2410 spi driver as soon as you are available
Sep 17 20:54:01 <fluff> any particular problem?
Sep 17 20:54:01 <drath> prpplague: afaik sparkfun only offers the ARM-USB-TINY - that's from olimex
Sep 17 20:54:14 <drath> prpplague: the jtagkey-tiny is from amontec
Sep 17 20:54:21 <prpplague> drath: ahhh, my mistake
Sep 17 20:54:29 <prpplague> looking at the wrong item then, 
Sep 17 20:54:55 * fluff does some calcs to see what this psu is done
Sep 17 20:55:06 <fluff> drath: must get around to submitting my s3c24xx nand support
Sep 17 20:55:34 <drath> fluff: did you use the nand.h interface?
Sep 17 20:55:44 <fluff> think I did
Sep 17 20:55:56 <fluff> i've been ill since writing it which is why it hasn't gotten out yet
Sep 17 20:56:05 <drath> great :)
Sep 17 20:56:25 <drath> well, not the being ill thing, but the nand thing :)
Sep 17 20:57:05 <prpplague> fluff: that would be nice, i'd already started working on the nand support
Sep 17 20:57:13 <prpplague> fluff: hadn't completed it yet though
Sep 17 20:57:21 <fluff> the only bit i couldn't get working was using the comms registers to accelerate
Sep 17 20:57:29 <fluff> but this was baded on r17x
Sep 17 20:57:39 <fluff> whatever the problem may be fixed now
Sep 17 20:57:43 <prpplague> fluff: ahh
Sep 17 20:57:51 <fluff> after the next kernel release i'll push it to the list
Sep 17 20:57:55 <prpplague> fluff: i'd like to have that patch if you post it somewhere
Sep 17 21:02:05 <fluff> hmm, their calcs for their DC-DC show it to be set to 2.82V
Sep 17 21:02:14 <fluff> now to not get angry with client
Sep 17 21:02:56 <fluff> hmm, the old values made 3.1V
Sep 17 21:03:46 <prpplague> doh
Sep 17 21:14:10 <drath> fluff: are you using the OpenOCD on linux or on windows?
Sep 17 22:15:47 <fluff> drath: Linux
Sep 17 22:16:21 <fluff> would have been really useful to have ~5years ago when we first did our 2410 boards
Sep 17 22:17:01 <fluff> right, yes, they acknowledge the bug in DC-DC config
Sep 17 22:17:05 <fluff> new board being sent
Sep 17 22:17:29 <fluff> temp solution will be tommorow get the second bench supply out and connect 3V to the probe
Sep 17 22:19:36 <fluff> so far it seems to work well
Sep 17 22:19:54 <fluff> probably going to recommend it to customers
Sep 18 00:11:58 <fluff> goodnight all
**** ENDING LOGGING AT Tue Sep 18 02:59:59 2007