**** BEGIN LOGGING AT Sun Apr 17 23:59:57 2005
**** BEGIN LOGGING AT Mon Apr 18 01:42:33 2005
Apr 18 09:54:55 <[g2]> beewoolie-away, hey!
Apr 18 09:55:13 <beewoolie-away> [g2]: what up dog?
Apr 18 09:55:40 <[g2]> I was thinking about kernels and bootloaders :)
Apr 18 09:56:14 <[g2]> I had a *REALLY* crazy idea :)
Apr 18 10:00:50 <beewoolie-away> ...and
Apr 18 10:02:56 <beewoolie-away> [g2]: I thought about the PCI problem a little.  It seems like we should survey the PCI registers and compare them.
Apr 18 10:04:20 <[g2]> I wondering whether a JTAG debugger or an on-board debugger that automatically steps most lines of assembler, or an emulator could
Apr 18 10:04:51 <[g2]> just capture all the writes to the various different register spaces (PCI, different areas on-chip, etc....)
Apr 18 10:05:36 <[g2]> Even if it runs at 1/100 the speed, the boot process is 20-30 seconds thats 1/2 hour
Apr 18 10:05:44 <[g2]> and you've have all the writes captured
Apr 18 10:05:53 <[g2]> probably even by process id/name
Apr 18 10:06:11 <beewoolie-away> The openjtag platform would be really good for this.
Apr 18 10:06:26 <[g2]> It *totally* would, but its not there yet
Apr 18 10:06:31 <[g2]> or even close
Apr 18 10:07:04 <beewoolie-away> I don't have anything cabable of this at the moment.  AFAIK, only jtag can make this kind of cycle-by-cycle inspection.
Apr 18 10:08:55 <[g2]> Ever here of these guys ? http://www.etoolsmiths.com/
Apr 18 10:28:32 <beewoolie-away> [g2]: What does it cost?
Apr 18 10:40:17 <JMunakra> g2, you want to capture all accesses a processor makes?
Apr 18 10:41:07 <[g2]> JMunakra, actually just all accesses to the configuration areas of on-chip and off-chip devices
Apr 18 10:42:15 <JMunakra> I have a library that supports JTAG operations (bit-shift and that stuff)
Apr 18 10:42:33 <JMunakra> it's not published currently, but any day now :)
Apr 18 10:42:56 <[g2]> what kind of devices does it drive ?
Apr 18 10:43:07 <JMunakra> Parport
Apr 18 10:43:26 <JMunakra> plugin interface for adding cable types
Apr 18 10:43:27 <[g2]> Does that include the Wiggler ?
Apr 18 10:43:39 <[g2]> or Wiggler like devices ?
Apr 18 10:43:42 <JMunakra> I guess so (didn't try)
Apr 18 10:43:53 <[g2]> All C ?
Apr 18 10:43:56 <JMunakra> but it should be easily adjustable if not
Apr 18 10:44:01 <JMunakra> yeah, C
Apr 18 10:44:27 <JMunakra> brb
Apr 18 10:45:13 <[g2]> beewoolie-away, ARM/Xscale has got just a handful of store opterations right ?
Apr 18 10:47:12 <[g2]> I guess DMA is an issue
Apr 18 10:47:19 <beewoolie-away> Off the top of my head, I think there are two.  Store, and store multiple.  Depends on how you define 'operations'.  Most opcodes have several bits controlling how the data is transferred: access width, source registers, and so on.
Apr 18 10:47:39 <[g2]> nod.
Apr 18 10:47:43 <beewoolie-away> IIRC, jtag give access to the bus which means that *all* accesses are available.
Apr 18 10:47:50 <beewoolie-away> Regardless of the source.
Apr 18 10:48:03 <prpplague> beewoolie-away: lo
Apr 18 10:48:07 <prpplague> beewoolie-away: how goes it?
Apr 18 10:48:10 <beewoolie-away> prpplague: Hey man.
Apr 18 10:48:16 <[g2]> So the processor runs in a slowed down mode ?
Apr 18 10:48:28 <[g2]> non-realtime
Apr 18 10:48:28 <beewoolie-away> prpplague: Not bad.  I'm learning to hate autotools more and more, day by day.
Apr 18 10:48:35 <prpplague> beewoolie-away: btdt
Apr 18 10:49:09 <beewoolie-away> [g2]: I suppose so.  I don't know details.  I've read some of the TI docs on JTAG.  
Apr 18 10:49:28 <beewoolie-away> If I ever get an openjtag device, I'll probably buy the spec.
Apr 18 10:51:23 <JMunakra> the 1149 spec?
Apr 18 10:51:47 <beewoolie-away> JMunakra: That's the one.  Seems worth owning if I'm going to hack on jtag.
Apr 18 10:52:17 <JMunakra> not really, as I was told. What you need is the 1149 state diagram
Apr 18 10:52:38 <JMunakra> I encourage reading the ARM7TDMI docs, very detailed on JTAG action.
Apr 18 10:52:48 <JMunakra> it also has said state diagram
Apr 18 10:52:53 <beewoolie-away> That's published.  One of the things I'd like to do is parse bsdl files directly.
Apr 18 10:53:01 <beewoolie-away> JMunakra: fair enough.
Apr 18 10:53:19 <beewoolie-away> I don't have a problem supporting IEEE.
Apr 18 10:53:22 <JMunakra> :)
Apr 18 10:53:48 <beewoolie-away> The openwince method isn't very convenient for adding new CPUs.
Apr 18 10:54:01 <beewoolie-away> And it's error prone.
Apr 18 10:54:35 <JMunakra> g2: you switch the processor clock through JTAG
Apr 18 10:54:42 <JMunakra> when in debug mode
Apr 18 10:55:58 <[g2]> JMunakra, ok so you can control the processor a clock at a time ?
Apr 18 10:56:46 <JMunakra> It just doesn't run by itself while in debug mode. Well, I guess it depends on the debug unit, but that's how it is on ARM.
Apr 18 10:57:09 <prpplague> JMunakra: hmm, the arm7tdmi docs i have are very vague on jtag
Apr 18 10:57:23 <prpplague> JMunakra: in that a printed doc or a pdf somewhere?
Apr 18 10:57:36 <JMunakra> Annex B, "Debug in Detail"
Apr 18 10:57:41 <JMunakra> did you read that?
Apr 18 10:57:50 <prpplague> JMunakra: from which document?
Apr 18 10:58:05 <prpplague> JMunakra: the armarm?
Apr 18 10:58:52 <JMunakra> No, the ARMARM has the architecture. The different implementations (ARM7, ARM9, etc) have their own manuals
Apr 18 10:59:15 * prpplague checks his arm ltd cd
Apr 18 10:59:40 <JMunakra> its called ARM7TDMI Technical Reference Manual, Rev. 4, ARM DDI 0210B
Apr 18 11:03:13 * prpplague downloads
Apr 18 11:03:25 <beewoolie-away> [g2]: I think you may be missing the point.
Apr 18 11:03:25 <[g2]> url ?
Apr 18 11:03:38 <[g2]> beewoolie-away, which point would that be
Apr 18 11:03:46 <prpplague> JMunakra: i've always just used the cpu reference manual, didn't think about getting a core manual
Apr 18 11:04:01 <beewoolie-away> [g2]: changing the JTAG clock (or whatever) doesn't mean you can read the data.
Apr 18 11:04:02 <prpplague> [g2]: http://www.arm.com/pdfs/DDI0210C_7tdmi_r4p1_trm.pdf
Apr 18 11:04:14 <beewoolie-away> We need to single step to read the jtag register.
Apr 18 11:05:03 <beewoolie-away> prpplague: I also believe that the JTAG state machine is always the same.
Apr 18 11:05:18 <JMunakra> beewoolie: it is
Apr 18 11:05:20 <beewoolie-away> Make that the 'clever' JTAG state machine.
Apr 18 11:05:45 <beewoolie-away> It's always refreshing to see a good piece of engineering.
Apr 18 11:07:20 <prpplague> lovely, just the docs i was looking for
Apr 18 11:08:06 <JMunakra> Ever wondered about the ARMARM saying "XXX is IMPLEMENTATION DEPENDENT" ? :P
Apr 18 11:09:55 <prpplague> JMunakra: yea yea, i always thought that it would be described it the manufacturer specific docs
Apr 18 11:15:51 <[g2]> JMunakra, thx for the reference and prpplague thx for the URL
Apr 18 11:17:47 <prpplague> [g2]: np, glad to share
Apr 18 11:18:09 <prpplague> [g2]: i've been collecting jtag info at http://www.elinux.org/wiki/JTAG
Apr 18 11:29:16 <JMunakra> prpplague, beewoolie: what are you guys doing on the openjtag project?
Apr 18 11:32:54 <prpplague> JMunakra: hehe, i'm a lurker here, i do bringups for ARM boards
Apr 18 11:33:07 <prpplague> JMunakra: i just like to listen in on the jtag conversation
Apr 18 11:33:08 <[g2]> http://www.openjtag.net/wiki/
Apr 18 11:33:27 <[g2]> The page needs some updating
Apr 18 11:34:34 <JMunakra> g2, what do you do on the project?
Apr 18 11:35:22 <[g2]> I was the guy that had the original idea
Apr 18 11:35:32 <[g2]> ~[g2]
Apr 18 11:35:32 <qbot> i haven't a clue, [g2]
Apr 18 11:35:33 <purl> i heard [g2] is really cool, and an an all around nice guy!  Also he is a real funny guy, the master of NSLU2 Buildroot and a known Slug Sacrificer.  And editor of SlugNews.
Apr 18 11:36:08 <[g2]> "Slug Sacrificer" I gave an NSLU2 unit to a guy to desolder all the chips and map out the traces
Apr 18 11:36:32 <[g2]> to find the JTAG connections if they came out from the BGA processor somewhere
Apr 18 11:36:57 <JMunakra> and the guy died in the process?
Apr 18 11:37:17 <[g2]> jacques, and a couple others have acutally soldered on to the JTAG points and reflashed 
Apr 18 11:37:42 <[g2]> The NSLU2 == Slug
Apr 18 11:37:52 <JMunakra> IC
Apr 18 11:38:02 <[g2]> are you familiar with the www.nslu2-linux.org ?
Apr 18 11:38:53 <[g2]> http://www.nslu2-linux.org/wiki/Info/PhotosOfTheInternals
Apr 18 11:39:15 <JMunakra> not really, but I attended the CCC Easterhegg in Hamburg, Germany, and there was literally a room full of people hacking those, and two sessions about it too
Apr 18 11:39:44 <[g2]> I've heard we've got a strong German branch :)
Apr 18 11:39:47 <JMunakra> also about something called gentoo-linux
Apr 18 11:39:53 <JMunakra> :-)
Apr 18 11:40:18 <[g2]> well gentoo linux and gentoo-embedded are other topics :)
Apr 18 11:40:36 <[g2]> So I'm guessing you are from Germany :)
Apr 18 11:40:59 <JMunakra> yeah
Apr 18 11:41:04 <JMunakra> so what did you / do you currently do on Openjtag, besides establishing a need for it?
Apr 18 11:41:58 <[g2]> I'm a founder of nslu2-linux and the release manager of the OpenSlug version which is a firmware replacement for the NSLU2
Apr 18 11:42:27 <[g2]> I've been working with several of the guys in here on bootloader replacment
Apr 18 11:43:21 <[g2]> beewoolie-away, is the primary author of the APEX bootloarder which is an ARM bootloader that currently boots the NSLU2, but the Ethernet and USB are initialized properly yet
Apr 18 11:43:32 <JMunakra> I'm working on an allpurpose jtag lib, that's why I try to find out who really works with it
Apr 18 11:43:36 <[g2]> we've been debugging those issues
Apr 18 11:43:50 <[g2]> sounds like a good match.
Apr 18 11:44:12 <[g2]> I've got a couple other ARM targets that have JTAG headers and could possibly be used for testing
Apr 18 11:44:45 <[g2]> Ultimately, I'm interested in high-speed devices debugging
Apr 18 11:44:50 <JMunakra> I have a working ARM7 debug stub for gdb
Apr 18 11:45:06 <JMunakra> no thumb code tho
Apr 18 11:45:07 <[g2]> how much different is an ARM9 ?
Apr 18 11:45:36 <JMunakra> dunno :)
Apr 18 11:45:53 <[g2]> I've seen some rally cheap ARM7's
Apr 18 11:46:28 <JMunakra> Absolutely
Apr 18 11:48:43 <JMunakra> So who's actively working on the JTAG code? ka6sox only?
Apr 18 11:50:12 <prpplague> JMunakra: i think dev is kinda at a standstill at the momment
Apr 18 11:50:20 <JMunakra> I see
Apr 18 11:50:26 <prpplague> JMunakra: beewoolie-away is working on some stuff for sharp
Apr 18 11:50:41 <prpplague> JMunakra: and i've been swamped with a new handheld
Apr 18 11:51:11 <beewoolie-away> JMunakra: the gating factor has been hardware.
Apr 18 11:51:20 <JMunakra> well, there is http://lap.epfl.ch/dev/arm/jelie/
Apr 18 11:51:24 <beewoolie-away> JMunakra: We arrived at a design that would minimize the HW phase.
Apr 18 11:51:38 <prpplague> JMunakra: yea, i've got jelie, but haven't had time to test
Apr 18 11:51:52 <prpplague> JMunakra: we have a custom HG dongle that i was gonna use
Apr 18 11:52:04 <prpplague> JMunakra: along with adding support for the sharp processors
Apr 18 11:52:24 <JMunakra> HG dongle?
Apr 18 11:52:33 <beewoolie-away> We'e been interested in producing hw as well as sw so that the system can work on any platform.
Apr 18 11:52:47 <prpplague> JMunakra: holly-gates
Apr 18 11:52:54 <JMunakra> ?? 
Apr 18 11:52:55 <prpplague> JMunakra: one sec, let me get you a url
Apr 18 11:52:58 <beewoolie-away> Most of the dongles don't support volatage changes.
Apr 18 11:53:36 <JMunakra> oh, what are you aiming for? 1.8/3.3/5 ?
Apr 18 11:54:08 <prpplague> JMunakra: http://www.lart.tudelft.nl/projects/jtag/
Apr 18 11:54:58 <prpplague> JMunakra: our dongle is based on the same design, however, we've add things like nTRST, and RST
Apr 18 11:56:06 <JMunakra> i see
Apr 18 11:56:38 <JMunakra> i think ka6sox was working on something FPGA based?
Apr 18 11:57:06 <prpplague> JMunakra: yea, sorry, when i say "we" i mean the company i work for
Apr 18 11:57:22 <JMunakra> who do you work for?
Apr 18 11:57:36 <prpplague> http://www.amltd.com
Apr 18 11:58:34 <prpplague> JMunakra: we have a full range of uclinux and linux devices
Apr 18 11:58:34 <JMunakra> oh, you made your own design tools? nice
Apr 18 11:59:36 <prpplague> JMunakra: we have full design team, plastics, hardware, electronics, and in house pcb manufacturer for small runs
Apr 18 11:59:51 <JMunakra> cool
Apr 18 12:00:50 <JMunakra> haha
Apr 18 12:00:53 <JMunakra> "We require all new employees to submit to a pre-employment drug screen"
Apr 18 12:03:27 <beewoolie-away> prpplague: speaking of your jtag dongle.  Any sign of one heading out your door to mine?
Apr 18 12:03:56 <prpplague> beewoolie-away: hell no, i just got my vonage phone working last week, i'll call this afternoon and find out
Apr 18 12:04:03 <JMunakra> So what do you think about the USB solution (Anchor) from jelie?
Apr 18 12:04:11 <prpplague> beewoolie-away: i don't think the shiping ppl ever read their email
Apr 18 12:30:52 <JMunakra> later guys
Apr 18 17:41:23 <[g2]> beewoolie-away, got a slug handy ?
Apr 18 17:41:52 <beewoolie-away> [g2]: Why?
Apr 18 17:42:18 <[g2]> can you compute a checksum for me ?
Apr 18 17:42:27 <beewoolie-away> [g2]: Do you know if I can download unslung 4.x?
Apr 18 17:42:29 <[g2]> I'm comparing to make sure lots of ppl have the same one
Apr 18 17:42:35 <beewoolie-away> Checksum on what?
Apr 18 17:42:38 <[g2]> I'm checking right now
Apr 18 17:43:08 <[g2]> I don't see the 4.x beta
Apr 18 17:43:21 <[g2]> I didn't think there were 20 Alpha users yet
Apr 18 17:43:30 <[g2]> http://sourceforge.net/project/showfiles.php?group_id=116564
Apr 18 17:43:44 <beewoolie-away> rwhitby as looking for testers.  I thought it was for unslung, but I'm not usre.
Apr 18 17:44:04 <beewoolie-away> I looked there already.
Apr 18 17:44:05 <[g2]> yes, standard release policy is 20 Alpha users building from source
Apr 18 17:44:24 <beewoolie-away> Maybe he was talking about openslug.
Apr 18 17:44:47 <[g2]> NO, OpenSlug has had 20+ for over a month
Apr 18 17:45:12 <[g2]> there's 28 OpenSluggers now
Apr 18 17:45:20 <[g2]> Only 15 4.x
Apr 18 17:45:28 <[g2]> http://groups.yahoo.com/group/nslu2-linux/database
Apr 18 17:45:49 <[g2]> 2 alternate boot loaders
Apr 18 17:46:54 <beewoolie-away> Maybe that's it, then.
Apr 18 17:47:11 <[g2]> nod. I won't get "released" until there's 20
Apr 18 17:47:27 <[g2]> Maybe if you asked rwhitby he'd send you a binary
Apr 18 17:47:33 <beewoolie-away> I loathe yahoo groups.
Apr 18 17:47:47 <beewoolie-away> can I build 4.x from OE?
Apr 18 17:47:53 <[g2]> most probably
Apr 18 17:48:34 <[g2]> I'd imagine you'd just mv tmp to tmp_openslug and change the distro from openslug to unslung and rebuild the whole shooting match
Apr 18 17:48:57 <[g2]> although It's been many months since I did that
Apr 18 17:49:02 <[g2]> build unslung
Apr 18 17:49:25 <beewoolie-away> I'm building unslung right now.
Apr 18 17:49:40 <[g2]> cool
Apr 18 17:49:56 <[g2]> so you got a NSLU2 around that you've got console to ?
Apr 18 17:50:02 <beewoolie-away> I've got the big disk ready to go.  all I need is a working system and I'll be putting one into a test rotation before delopying it as a server on  my network.
Apr 18 17:50:24 <beewoolie-away> what checksum do you need?
Apr 18 17:55:50 <jacques> anyone built openslug lately?
Apr 18 17:55:59 <[g2]> nope
Apr 18 17:56:10 <[g2]> I've been bootloadering and native compileing
Apr 18 17:56:13 <beewoolie-away> We were just talking about that.
Apr 18 17:56:13 <[g2]> I've been bootloadering and native compiling
Apr 18 17:56:21 <[g2]> bbiab
Apr 18 17:56:31 <jacques> I tried a few days ago and it didn't complete
Apr 18 17:57:28 <jacques> might be something simple - seemed like some disconnect between kernel config and some .bb
Apr 18 17:57:28 <[g2]> I've been thinking about taking a snapshot of the OE files
Apr 18 17:59:07 <beewoolie-away> [g2]: I'm going to go out for a bit.  I can look at some slug things in a couple of hours.
Apr 18 17:59:26 <[g2]> ok
Apr 18 17:59:51 <[g2]> when you get back can you type dd if=/dev/mtdblock0 bs=1000 count=262 | md5sum 
Apr 18 18:00:00 <[g2]> jacques, can you do that too ?
Apr 18 18:03:48 <jacques> 83deb02a8df713e638afc13bc2b4f46c
Apr 18 18:04:04 <[g2]> thx, that looks familiar :)
Apr 18 18:04:18 <[g2]> ~logs
Apr 18 18:04:21 <purl> apt/ibot/jbot/purl all log to http://ibot.rikers.org/<channelname>/ where channelname is html encoded ie: %23debian | lines that start with a space are not shown | some channels have stats at http://ibot.rikers.org/stats/<channelname>.html.gz
Apr 18 18:04:21 <qbot> [g2]: are you smoking crack?
Apr 18 18:04:51 <[g2]> eno	83deb02a8df713e638afc13bc2b4f46c  -
Apr 18 18:05:06 <[g2]> I got the same think on my two slugs
Apr 18 18:05:27 <[g2]> I makes for a fancy compression algorithm
Apr 18 18:15:33 <jacques> anyone know how to create pdf files under linux?
Apr 18 18:15:58 <[g2]> doesn't Latex create pdfs ?
Apr 18 18:16:16 <[g2]> I know there are conversion utilities from like ps2pdf
Apr 18 18:16:41 <[g2]> And I'm sure Adobe under WINE would do a nice job, but I don't think that's what you are thinking
Apr 18 18:17:04 <jacques> no, that latter is not what I want
Apr 18 18:17:09 <[g2]> :)
Apr 18 18:17:16 <jacques> i dont understand why abiword doesn't just do it
Apr 18 18:18:43 <jacques> ah ok, I found this; http://www.abisource.com/twiki/bin/view/Abiword/FaqCreatingPDFs
Apr 18 18:20:10 <[g2]> jacques, did you see that thing (on /. irrc) about GPL fonts ? 
Apr 18 18:20:54 <jacques> saw the blurb, disn't read it
Apr 18 18:21:51 <[g2]> they basic thing was that from a recent court ruling document writtin using GPL fonts my be have to be GPL
Apr 18 18:22:27 <[g2]> I think ppl are looking in to getting an exception, (kinda like lgpl) so the documents could still abe copyrighted as normal doc
Apr 18 18:22:46 <jacques> bah
Apr 18 18:23:05 <jacques> there is a bug requesting builtin pdf exporting capabilities for abiword - from 2001
Apr 18 18:23:11 <jacques> still open
Apr 18 18:23:19 <[g2]> DOH!
Apr 18 18:28:06 <jacques> looks like if I re-build abiword with gnome support, and build gnome-print (neither of which I want to do) I can get pdf files
Apr 18 18:28:19 <jacques> maybe I'll try pdflatex
Apr 18 18:31:42 <jacques> or ps2pdf
**** ENDING LOGGING AT Mon Apr 18 23:59:56 2005