**** BEGIN LOGGING AT Mon May 16 23:59:56 2005
May 17 00:02:22 <ka6sox-away> it'd be quick!
May 17 00:03:48 <jbowler> 3ns?  You could get in quite a lot of switches before it stopped working.
May 17 00:04:27 <ka6sox-away> parts only good to 10ns!
May 17 00:05:29 <jbowler> Then it would be fine - a nice sine wave, only averaging about 10ma.
May 17 00:06:03 <jbowler> Not exactly digital electronics of course ;-)
May 17 00:08:05 <ka6sox-away> these aren't mixed signal CPLD's
May 17 00:25:42 <dyoung-zzzz> jacques, It seems I'm gonna be SOL for this series2 modification project.
May 17 00:26:25 <dyoung-zzzz> There isnt series2 support for Tystudio, and tivoweb apparently is busted for series2.
May 17 00:27:02 <jacques> ah damn :-(
May 17 00:27:10 <jacques> I knew s2 were a PITA
May 17 00:27:37 <jacques> I have two spare s1 that I bought to hack
May 17 00:27:42 <dyoung-zzzz> Wow.
May 17 00:28:02 <jacques> surpluscomputers.com had them for like $120 a while back
May 17 00:28:10 <dyoung-zzzz> I gotta keep my eyes open i guess.
May 17 00:28:31 <dyoung-zzzz> I found a bunch of instructions that are supposed to work; but I cant find all the bits to make it go yet.
May 17 00:32:17 <jacques> dyoung-zzzz, do you know what you're missing?
May 17 00:32:38 <dyoung-zzzz> a directory of s2 kernels atm.
May 17 00:32:52 <dyoung-zzzz> that I can run this killhdinitrd stuff on.
May 17 00:33:27 <jacques> you fighting with some hd locking thing?
May 17 00:34:11 <dyoung-zzzz> Interesting.  I just read that smeone compiled tytools for mips and it sorta works.
May 17 00:34:24 <dyoung-zzzz> I dont think I have the locking problem.
May 17 00:34:24 <jacques> ah yeah, they changed proc archs
May 17 00:34:41 <jacques> one thing about the tivo community that always annoyed me
May 17 00:34:57 <dyoung-zzzz> if I ever get this worked out I am going to have to cram it into wiki.
May 17 00:34:58 <jacques> ppl hold onto their source like it's MS code or something
May 17 00:35:17 <jacques> everything's all binary releases
May 17 00:35:22 <dyoung-zzzz> yeah, its the "see how l33t I am!"
May 17 00:35:27 <dyoung-zzzz> syndrome.
May 17 00:35:31 <jacques> exactly - extremely annoying
May 17 00:35:58 <jacques> just like on pocketpc - every dumbass who can write two lines of code think it's worth $10-$40
May 17 00:36:22 <dyoung-zzzz> Yeah, when youre *really* l33t like us, we release source like ... hmm.
May 17 00:36:58 <dyoung-zzzz> cant think of a good one. sorry. :-)
May 17 00:38:24 <dyoung-zzzz> I guess that must mean we're not all l33t and stuff. :-)
May 17 01:21:06 <ka6sox-away> d00d u n0t s00 l33t
May 17 01:22:33 <dyoung-zzzz> Grr.
May 17 01:22:42 <dyoung-zzzz> It shouldnt be this difficult to hack your Tivo. :-)
May 17 01:31:55 <jacques> they intentionally made it harder for s2
May 17 02:21:40 <dyoung-zzzz> I might have to revert to brute force.
**** BEGIN LOGGING AT Tue May 17 02:40:15 2005
May 17 03:36:27 <dyoung-zzzz> Ahhh.... Okay.
May 17 03:36:57 <dyoung-zzzz> if I can figure out how to enable telnet or get a console I found enough information to make this work.
May 17 10:51:49 <[g2]> beewoolie, I dropped off the hw today
May 17 10:52:01 <beewoolie> For rework?
May 17 10:52:05 <[g2]> nod
May 17 10:52:09 <beewoolie> Awesome.
May 17 10:52:19 <[g2]> looks like they can do NSLU2's too :)
May 17 10:54:09 <[g2]> so are you interested in a pimped out NSLU2 @ 64M, serial and JTAG ?
May 17 10:55:24 <[g2]> ep1220, hey
May 17 10:55:41 <[g2]> are you having any luck with the spartan3  and cypress part ?
May 17 10:56:06 <ep1220> [g2] Hi 
May 17 10:56:33 <ep1220> I can talk to the Cypress Chip from the Cypress SDK
May 17 10:57:06 <ep1220> Currently I am learning to use the FPGA software
May 17 10:57:22 <[g2]> AWESOME
May 17 10:57:47 <[g2]> that's the USB2 from the digilent board ?
May 17 10:57:52 <beewoolie> [g2]: Not sure about the pimped nslu2.  do you think they can fix the jtag pads I broke?
May 17 10:58:19 <[g2]> beewoolie, depends on the pads, probably not worth it
May 17 10:58:23 <ep1220> [g2]: yes.
May 17 10:58:31 * [g2] hugs ep1220 
May 17 10:59:45 <ep1220> [g2] but this is only a small step in writing a frimware to talk to the S3
May 17 11:00:20 <[g2]> well it's the biggest step taken so far and that's important
May 17 11:00:35 <[g2]> sometimes a little movement attracts others :)
May 17 11:00:58 <ep1220> We will see :-)
May 17 11:01:25 <[g2]> so are you reprogramming the cypress ?
May 17 11:01:25 <ep1220> I also have a FTDI 2232 module here.
May 17 11:01:41 <[g2]> is it the 2232C by chances ?
May 17 11:02:00 <ep1220> [g2] Yes, I want to evaluate it.
May 17 11:02:26 <[g2]> that's got the straight through to JTAG mode as an app note right
May 17 11:02:26 <ep1220> Need a few circuits around it to connect to a JTAG
May 17 11:02:55 <ep1220> [g2]: yes, it has JTAG TAP inside
May 17 11:03:26 <[g2]> We could really use that now
May 17 11:04:15 <ep1220> Advantage over my 8051 solution would be: It is available in very low volume
May 17 11:04:52 <[g2]> what do you mean by low volume ?
May 17 11:05:24 <ep1220> one can buy down to single units. the 8051 is 160+
May 17 11:07:25 <[g2]> how hard would it be to layout the 2332C design ?
May 17 11:07:36 <ka6sox-office> 2232?
May 17 11:07:45 <ka6sox-office> the FT part?
May 17 11:07:54 <ep1220> ka6sox-away: yes
May 17 11:08:00 <[g2]> ka6sox-office, 6 minutes ago in the log
May 17 11:08:25 <ka6sox-office> I see it...I have it almost done.
May 17 11:08:32 <ka6sox-office> worked on it last night.
May 17 11:08:57 <ka6sox-office> 2 chips and a bunch of pins.
May 17 11:09:02 <ka6sox-office> (and a USB connector.
May 17 11:09:23 <ka6sox-office> need to have somebody verify it.
May 17 11:09:50 <[g2]> what software did you use for the design layout ?
May 17 11:10:29 <ka6sox-office> PCB
May 17 11:10:35 <ka6sox-office> and geda
May 17 11:10:58 <[g2]> I know some hw guys that can review the design
May 17 11:11:09 <[g2]> what are you plans with it ?
May 17 11:11:12 <ka6sox-office> I'll post it up on the wiki
May 17 11:11:17 <ka6sox-office> I plan on building it.
May 17 11:11:22 <ep1220> ka6sox-away: if You want I can take a look at it
May 17 11:11:25 <ka6sox-office> and building about 15 or so of them.
May 17 11:11:44 <ka6sox-office> ep1220, i'll take you up on that.
May 17 11:12:07 <ka6sox-office> when I finish it today I'll post it up on the wiki and tell people here in channel.
May 17 11:12:26 <ep1220> ka6sox-away: openjtag wiki ?
May 17 11:12:40 <[g2]> ka6sox-office, do you have a costed BOM ?
May 17 11:13:20 <ka6sox-office> that will come right after I finish the "design"
May 17 11:13:40 <ka6sox-office> <$35 is what I'm looking for.
May 17 11:14:19 <[g2]> It that fully assembled or just the cost of the BOM and PCB ?
May 17 11:15:08 <ka6sox-office> dunno...the cost of the chip (ft2232c) is $8 each in small quantities.
May 17 11:15:25 <[g2]> define small please
May 17 11:15:26 <ka6sox-office> and the CPLD is $3.16/each
May 17 11:15:31 <ka6sox-office> <50
May 17 11:16:35 <ka6sox-office> then connectors and Board are all that is left I need to look at cases.
May 17 11:17:01 <[g2]> any idea how fast the price drops at 100 or 200 units ?
May 17 11:18:04 <ka6sox-office> at 100 the FT part is $6.20
May 17 11:18:32 <ka6sox-office> so 25% for 100qty.
May 17 11:19:32 <ka6sox-office> the boards then essentially become cost of production and NO setup (very low amoratization of setup)
May 17 11:19:54 <ka6sox-office> the biggest issue for me is the connectorization.
May 17 11:20:22 <ka6sox-office> the Digilent connector is fine + using flying leads alternatively.
May 17 11:21:29 <[g2]> are you just bringing 8 pins out ?
May 17 11:21:52 <[g2]> or 14, 20 ?
May 17 11:21:56 <ka6sox-office> haven't decided exactly what to bring out...was going to ask the community/
May 17 11:22:08 <ka6sox-office> what do you "want/need:
May 17 11:22:18 <[g2]> how many I/Os does the CPLD support ?
May 17 11:22:23 <ka6sox-office> brb
May 17 11:22:40 <ep1220> ka6sox-office: i would use the 2nd channel for 3.3/5V serial to USB
May 17 11:23:17 <[g2]> ep1220, at one point we had talked about using 1 channel for control and 1 for data
May 17 11:23:56 <ep1220> [g2] do not understand what You mean ?
May 17 11:24:35 <[g2]> iirc the FT2232C is a dual device no ?
May 17 11:25:12 <ep1220> Yes, Chan A can be JTAG, bit bang, serial , parallel.
May 17 11:25:21 <ep1220> Chan B is serial/parallel only
May 17 11:25:53 <ep1220> When Chan A does JTAG, it can also do 8bit IO
May 17 11:25:55 <[g2]> ok.. So Chan B would be serial/parallel control channel to setup the CPLD
May 17 11:26:31 <[g2]> and Chan A would drive the JTAG data
May 17 11:26:40 <ep1220> Possible.
May 17 11:26:57 <ep1220> But Chan A GPIO could also bit-bang the CPLD.
May 17 11:27:03 <ep1220> Then Chan B can be serial
May 17 11:27:08 <[g2]> then we'd have an out-of-band communications channel for settnig up the CPLD
May 17 11:27:27 <[g2]> quite possible
May 17 11:28:04 <[g2]> I've heard of data rates near 20Mb with the FTDI part
May 17 11:28:19 <ep1220> Mb == Mbit ?
May 17 11:28:25 <[g2]> yes
May 17 11:28:44 <[g2]> for JTAG iirc
May 17 11:28:49 <ep1220> No, the eval module doc: It can sustan 5.4Mb.
May 17 11:29:06 <ep1220> It is USB1.1 so 12Mb theoretical maximum
May 17 11:29:22 <ep1220> JTAG clock is 6 Mbit max
May 17 11:29:40 <[g2]> from that part
May 17 11:30:30 <[g2]> but the Cypress is a USB 2.0 High-speed device correct ?
May 17 11:30:58 <ep1220> yes
May 17 11:31:16 <ep1220> the 8051 does USB house-keeping
May 17 11:31:29 <ep1220> a DMA engine passes data along to the S3
May 17 11:32:07 <[g2]> so what's the max frame size 512 or can it be bigger
May 17 11:32:25 <[g2]> and which mode would you run the chip in 
May 17 11:32:41 <[g2]> bulk xfer, interrupt, async ?
May 17 11:33:34 <ep1220> g2: bulk
May 17 11:35:04 <ep1220> on usb1.1 "on-the-wire" packet size is max 64, 512 for USB2
May 17 11:35:26 <ep1220> But one can transfer larger "logical" packets. 
May 17 11:35:40 <ep1220> The driver splits them
May 17 11:36:35 <[g2]> nod.  I was reading something about the on chip memory and how many frames could be allocated to its buffer
May 17 11:38:33 <ep1220> in USB terms Frame has a special meaning.
May 17 11:38:51 <ep1220> E.g. on USB1. frame length is 1microsecond
May 17 11:39:14 <ep1220> It is used e.g. for syncing Audio devices
May 17 11:40:11 <ep1220> I prefer not to use the word frame for plain data,as it can be confusing
May 17 11:42:02 <ka6sox-office> 5.4Mb is correct.
May 17 11:42:44 <ka6sox-office> the idea is to use the Chan B as OOB setup.
May 17 11:43:02 <ka6sox-office> either parallel or serial
May 17 11:43:10 <ka6sox-office> but probably parallel.
May 17 11:44:39 <ka6sox-office> I've asked Xilinx about using different Vccio's on different I/O blocks.
May 17 11:44:46 <ka6sox-office> (since there are several)
May 17 11:45:47 <ka6sox-office> such that we can have the interface to the FT chip be 3.3v (off the USB bus) and the DUT side as whatever voltage is necessary for making the other side work.
May 17 11:47:00 <ep1220> ka6sox-office: what type of Xilinx CPLD ?
May 17 11:47:17 <ka6sox-office> XC9536XL-10vq44
May 17 11:50:12 <ep1220> ka6sox-office: Once You have posted the schematic, I will check it.
May 17 11:54:24 <ka6sox-office> ep1220, look in about 6hrs on the wiki and I'll announce it here.
May 17 11:56:50 <ep1220> ka6sox-office, I will look at it tomorrow morning (my time) 
May 17 11:57:02 <ka6sox-office> europe?
May 17 11:57:12 <ka6sox-office> CET
May 17 11:58:55 <ep1220> yes,
May 17 12:03:29 <[g2]> ep1220, so back to the USB2 and Spartan3
May 17 12:03:46 <ka6sox-office> wait for the S3e board...
May 17 12:03:48 <[g2]> You've got a linux host talking to the usb2 ?
May 17 12:03:51 <ka6sox-office> it will be MUCH better.
May 17 12:05:26 <ka6sox-office> bbiab..lunch
May 17 12:05:57 <ep1220> [g2] all done in windowze till now
May 17 12:06:09 <ep1220> the cypress SDK is win only
May 17 12:06:40 <[g2]> ep1220, OK...
May 17 12:06:53 <[g2]> we should really just reverse engineer that :)
May 17 12:07:21 <ep1220> is a possibility as well.
May 17 12:07:44 <ep1220> however, it limits our capabilities on talking to the S3
May 17 12:09:36 <ep1220> what the Cypress SDK does is not a secret.
May 17 12:10:40 <ep1220> but i  am reluctant on to putting time into Linux
May 17 12:10:52 <ep1220> till i am sure all works as i expect.
May 17 12:11:17 <[g2]> nod.
May 17 12:11:23 <[g2]> that's understandable
May 17 12:13:56 <ep1220> performance wise there are some open questions.
May 17 12:14:27 <ep1220> Depends on the implementation of the USB engines in the chips
May 17 12:14:34 <ep1220> both cypress and FTDI
May 17 12:16:20 <ep1220>  2 hours ago you were talking about the NSLU2 ?
May 17 12:16:38 <ep1220> JTag is 3.3V on this model ?
May 17 12:18:22 <[g2]> http://www.nslu2-linux.org/wiki/Info/PinoutOfJTAGPort
May 17 12:18:29 <[g2]> think so on the JTAG
May 17 12:18:33 <[g2]> the 3.3v
May 17 12:18:50 <[g2]> and Yes that was for the NSLU2 and the RV042 and RV082
May 17 12:21:15 <ep1220> Do the RV042/082 have a JTAG connector on board ?
May 17 12:21:27 <[g2]> there's a header
May 17 12:21:55 <[g2]> I'll be trying to verify that the JTAG is all brought out to the header and working
May 17 12:22:58 <ep1220> good
May 17 12:24:31 <[g2]> Hey did you guys see the PMOD-RS232 at digilent ?
May 17 12:27:45 <ep1220> Looks good.
May 17 12:28:09 <ep1220> If shipping to here were not so expensive I would order one ;-)
May 17 13:32:38 <beewoolie> [g2]: Hey man.  I assume you've created jffs2 filesystems for openslug, right?
May 17 13:43:34 <jacques> beewoolie, maybe I can help
May 17 13:44:02 <beewoolie> jacques: OK.  I've been working on the JFFS2 reader code.  
May 17 13:44:18 <jacques> cool
May 17 13:44:27 <beewoolie> There's some sticky issues having to do with the way it logs.
May 17 13:44:48 <beewoolie> the first question I have is this.  Has the mkjffs2 program been fixed to properly handle hard links?
May 17 13:44:58 <beewoolie> do we use hard links in the jffs2 for openslug?
May 17 13:45:25 <jacques> I'm not aware of us using hard links in openslug.
May 17 13:45:37 <beewoolie> OK.  I suspect it may still be broken.
May 17 13:45:54 <beewoolie> If you really want to help, you could do me a big favor.
May 17 13:45:54 <jacques> you have a recent mtd version ?
May 17 13:46:09 <beewoolie> I'd like to have a jffs2 filesystem that has been used.
May 17 13:46:41 <beewoolie> I haven't looked at the code in a while, for mkjffs2.  the hard-links stuff was unimplemented as of 1.42, iirc.
May 17 13:47:43 <beewoolie> So, if you'd like to make this easy on me, you'd extract the jffs2 filesystem from memory and send it to me.
May 17 13:48:22 <[g2]> beewoolie, just dd the jffs2 partition
May 17 13:48:32 <[g2]> I can email you one
May 17 13:48:33 <beewoolie> jacques: BTW, the endian stuff isn't really a problem.  I discovered that it is always in native byte-order.  Reasonable choice considering the way things are used.
May 17 13:48:48 <beewoolie> [g2]: I don't have one right now.  I still havent built it.
May 17 13:48:59 <beewoolie> If you want to email one, that would be supercool.
May 17 13:49:10 <[g2]> there's a binary at openslug.nslu2-linux.org :)
May 17 13:49:21 <[g2]> sure I'll be happy to e-mail you one
May 17 13:49:22 <beewoolie> Right.  I want to use apex and I don't want to fuss.
May 17 13:49:37 <beewoolie> Having a built one will make it easier to tinker.
May 17 13:49:45 <[g2]> do you just want the whole image ?
May 17 13:50:06 <[g2]> meaning you'll have the kernel too
May 17 13:50:42 <[g2]> I think the whole shooting match squeezes down to like 4MB
May 17 13:50:48 <[g2]> bzipped2
May 17 13:50:49 <beewoolie> What I want is just the jffs2 part.  I want to see how changes to the filesystem come out in the logs.
May 17 13:51:03 <beewoolie> If the kernel is in there, then that's fine.
May 17 13:51:16 <[g2]> the kernel is in there
May 17 13:51:34 <beewoolie> so, you have two copies of the kernel right now?
May 17 13:51:39 <[g2]> nod
May 17 13:51:44 <beewoolie> hmm.
May 17 13:52:00 <[g2]> it's just the way OE builds it and we haven't taken it out
May 17 13:52:20 <[g2]> because I was hoping one day we could run it out of the rootfs :)
May 17 13:52:26 <beewoolie> Well, I have the basics down for reading jffs2.  I just want to make sure that a working jffs2 doesn't throw any curves.
May 17 13:53:02 <[g2]> I may have deleted the kernel from the jffs2 on the system I've got
May 17 13:53:11 <beewoolie> That's fine.
May 17 13:53:15 <beewoolie> I'd like to see that, too.
May 17 13:56:39 <jacques> beewoolie, when you say "used" you mean a jffs2 fs which has been mounted ?
May 17 13:57:06 <beewoolie> jacques: mounted.  Files written.  Files deleted.  Files renamed.  Whatever.
May 17 14:12:44 <[g2]> kas11, HEY !!
May 17 14:16:33 <kas11> hiya [g2] ;)
May 17 14:17:23 <kas11> waz up in toyland?
May 17 14:18:54 <[g2]> beewoolie, you've got e-mail
May 17 14:19:07 <beewoolie> Yes I do.
May 17 14:19:23 <kas11> I've been buzy breaking everything with Ubuntu...wondering if I really need to get a grip and just do debian instead...I found I was way too mentally lazy for gentoo
May 17 14:19:24 <[g2]> kas11, did you here the APEX news ?
May 17 14:19:25 <beewoolie> You're a saint.
May 17 14:19:52 <[g2]> yeah kas11 is a saint
May 17 14:20:04 * beewoolie hugs [g2] and kas11 
May 17 14:20:13 <kas11> I heard you had APEX booting the fat slag iirc, [g2]
May 17 14:20:40 <[g2]> kas11, it's boots both the 64MB avila and the 32MB slug
May 17 14:20:45 <[g2]> and all works fine
May 17 14:20:58 * kas11 rotflhao at saint moniker
May 17 14:21:08 <kas11> very kewl
May 17 14:21:26 * [g2] embraces beewoolie 
May 17 14:21:57 * kas11 knows beewoolie is one smart cookie
May 17 14:22:10 <[g2]> and a hard worker too!
May 17 14:22:22 <[g2]> kas11, two things
May 17 14:22:57 <[g2]> 1) There's an insight from jbowler that the RTC is causing NTP to hang on openslug
May 17 14:23:25 <[g2]> 2) For testing Gigabit NAS, I got an AMD64 that rocks da house
May 17 14:23:51 <[g2]> and I've sent some boards to the re-work fram
May 17 14:23:54 <[g2]> farm
May 17 14:25:28 <kas11> A64s do rock...I have a poor old 3200 athlon xp that doesn't even get powered up anymore
May 17 14:25:50 <[g2]> I wiped my 37G Raptor drive in 712 seconds (nearly 49MBs)
May 17 14:26:13 <kas11> darn....'tis fast
May 17 14:26:17 * [g2] is almost afraid to move to 64 bits
May 17 14:26:31 <[g2]> that's in 32bit compatibility mode
May 17 14:26:51 <[g2]> the other things is that the Rapter is not the fastest drive I've got
May 17 14:27:00 <kas11> I mostly run 32 bits...can't see all that much difference in 64 bit mode and there are hassles
May 17 14:27:36 <[g2]> I got a Maxtor 300G 16MB and it's got NCQ but was getting 60MB on the P4 where the Raptor was only getting 55
May 17 14:28:32 <kas11> the 2.6.11 386 ubuntu doesn't like me anymore so I am running it 64 bit tho...seems like Nautilus won't start but the logs are devoid of anything useful
May 17 14:28:38 <[g2]> So I'll be trying a pair of those Raid 0 it a little while and pulling / pushing across the GigE pipe with Jumbo frames
May 17 14:29:12 <[g2]> I'm just running a vanilla 2.6.11 Knoppix on it
May 17 14:29:25 <[g2]> I've looked in to building Gentoo on it
May 17 14:29:54 <[g2]> next time I talk to solar and spanky I find out if it's worth it
May 17 14:30:06 <[g2]> I've also toyed with building OpenSlug on it
May 17 14:31:20 <kas11> I am still prevaricating on GigE here...once I get a handle on debian/ubuntu I'll take the plunge...but since I am only using the 2 A64s, I may just do a short crossover cable and mess with my routing tables
May 17 14:32:55 <kas11> I got to a point where I had X on gentoo amd64...but doing all of gnome was more than I was up to
May 17 14:33:58 <kas11> all those cycles down the drain...5.5 hrs with the cpu pretty well pegged  *lol*
May 17 14:34:18 <[g2]> yeah... Gentoo can peg the box
May 17 14:34:26 <[g2]> it's really a good stress test
May 17 14:35:50 <kas11> yeppers...that was a good part of the reason why I did it...I am overclocking this box (/me hangs head) so I wanted to see what it did when I really pushed it
May 17 14:47:04 <kas11> [g2] I have no real idea about why x1205-rtc is hanging NTP...I'd have to move stuff over to ubuntu and build via the old makefile to even check...and ubuntu doesn't like to mount Fedora ext3 partitions so nothing has been moved over yet
May 17 14:48:26 <kas11> I had virtually no userland on my slug so never looked a ntp
May 17 15:03:23 <[g2]-dinner> np. It was just FYI
May 17 15:06:56 <kas11> is OpenSlug still using BK?
May 17 15:12:51 <[g2]> it is there, but there is also a tarball snapshot of the metadata
May 17 15:12:57 <[g2]> and a binary :)
May 17 15:20:17 <kas11> I was just wondering if they had moved to subversion or something ... I never did get the whole OE thing
May 17 15:25:00 <ka6sox-office> Hiya ka6sox-away 
May 17 15:25:03 <ka6sox-office> heh
May 17 15:25:08 <ka6sox-office> hiya kas11 
May 17 15:31:59 <kas11> hey ka6sox-office ;)
May 17 15:35:13 <ka6sox-office> how have you been?
May 17 15:41:11 <ka6sox-office> kas11, how did you make out with the Cypress chip in the Digilent board?
May 17 15:44:20 <[g2]> kas11, so how's the gumstix ?
May 17 15:50:53 <kas11> haven't been doing anything with either, ka6sox-office ... just been dinking around with the full sized boxes
May 17 15:51:13 <ka6sox-office> hopefuly your own and NOT the neighbors!
May 17 15:51:50 <kas11> well....
May 17 15:52:24 <ka6sox-office> that'll be $100 please!
May 17 15:53:06 <kas11> nothing where any money changes hands is fair to say
May 17 15:55:46 <ka6sox-office> bummer
May 17 16:01:25 <kas11> I've been thinking I need to figure out ebay so I can unload some of this stuff ... digilent and slug need new homes and prolly the gumstix too
May 17 16:02:58 <[g2]> Doh!
May 17 16:18:26 * dyoung-web might be interested.
May 17 16:43:25 <jbowler> beewoolie: jffs2 fully supports hard links - I'm pretty sure I would have noticed by now if there were any problems, there is something broken in the OE build process for jffs2 with respect to a file with >1 hard link (but not a directory) - that could be the mkfs, but I suspect something in the ipkg handling of a tar archive.
May 17 16:52:38 <jacques> jbowler-away, what about this (from mkfs.jffs2 man page)?
May 17 16:52:41 <jacques> mkfs.jffs2  does not properly handle hard links in the input directory structure.  Currently, hard linked
May 17 16:52:41 <jacques>        files will be expanded to multiple identical files in the output image.
May 17 16:54:07 <beewoolie> jacques: that's what I'm talking about.
May 17 16:54:20 <beewoolie> It seems like it would have been an easy thing to fix.
May 17 16:54:40 <jbowler-away> jacques: from memory when I looked at the rootfs after an ipkg install the hard links (after the first) weren't there at all, but I might be wrong - so I don't think I ever managed to find out what mkfs.jffs2 did ;-)
May 17 16:54:42 <[g2]> beewoolie, any luck reading the jffs2 info ?
May 17 16:56:04 <beewoolie> [g2]: I can read the data fine.  I've come up with a plan that should make the read process reasonably efficient.
May 17 16:56:30 <[g2]> excellent ... we only need to read to boot the kernel :)
May 17 16:56:36 <beewoolie> One of the sample pieces of code re-reads the data over and over again.
May 17 16:57:06 <beewoolie> Oh yeah.  There won't be a write function.  I've been worrying over the algorithm so that it doesn't take 30s to read the kernel.
May 17 16:57:20 <[g2]> nod.
May 17 16:57:33 <beewoolie> BTW, this is substantially simpler than the USB driver would be.
May 17 16:58:16 * [g2] has been reading about USB some doay
May 17 16:58:18 <[g2]> today
May 17 17:05:01 <dyoung-web> USB would be pretty handy.
May 17 17:05:55 <dyoung-web> you can put a stick in with a known good kernel and at the bootloader test for "OhCrapIBrokeMyOpenSlug" or something and if present use that one instead.
May 17 17:06:17 <dyoung-web> I'm sure thats one thing [g2] had in mind.
May 17 17:06:18 <[g2]> You know all told.... I think there are more developers/part-time of course on nslu2-linux than in the whole company at Linksys (156 headcount)
May 17 17:06:33 <dyoung-web> 156 at linksys or us?
May 17 17:06:43 <[g2]> that's linksys's number
May 17 17:06:48 <[g2]> I think we've got more :)
May 17 17:07:10 <dyoung-web> Could be.
May 17 17:09:51 <dyoung-web> we have 75 with CVS write access.
May 17 17:13:23 <[g2]> really that many ?
May 17 17:13:27 <[g2]> is that on sf ?
May 17 17:13:34 <dyoung-web> yes.
May 17 17:13:40 <[g2]> wow
May 17 17:13:45 <dyoung-web> Unless I grossly miscounted, its around there.
May 17 17:15:58 <[g2]> Developers: 70
May 17 17:16:21 <[g2]> simply amazing
May 17 17:19:24 <beewoolie> dyoung-web: The trouble with USB support is that it is relatively complex.  There's the OHCI driver, the EHCI driver, and then there are drivers for each of the device types.
May 17 17:19:31 <beewoolie> And, there's a PCI component.
May 17 17:19:50 <[g2]> nod.
May 17 17:20:02 <beewoolie> I'm not saying I don't want to do it.
May 17 17:20:04 <[g2]> Beaten Paths are for Beaten Men :)
May 17 17:20:17 <beewoolie> I am saying that I'd like to pick some of the lower hanging fruit first.
May 17 17:20:30 <[g2]> nod.
May 17 17:20:38 <beewoolie> uboot has usb.  Heh heh.
May 17 17:21:01 <[g2]> are you suggesting we graft it from there ?
May 17 17:21:30 <beewoolie> Having someone elses effort always helps.  I read the uboot and blob implementations of JFFS2 before I wrote something new.
May 17 17:21:54 <[g2]> the beauty of open source
May 17 17:21:59 <[g2]> let freedom ring
May 17 17:23:03 <[g2]> after I learn some more about USB I'll know whether BULK or INTERRUPT xfers will be easier to implement
May 17 17:23:05 <beewoolie> Perhaps I'm exxaggerating the complexity as the uboot code isn't very much.
May 17 17:23:22 <beewoolie> I think that that has to do with where it is scheduled.
May 17 17:23:45 <[g2]> they are different Xfer types from what I know
May 17 17:24:01 <dyoung-web> yeah, I fully realise the scope of the usb problem.
May 17 17:24:05 <beewoolie> uboot appears to use bulk.
May 17 17:24:18 <[g2]> We'll need bulk for booting off the usb devices directly
May 17 17:25:41 <[g2]-ldd3-usb-ch> back to reading the LDD3 USB chapter
May 17 17:28:35 <[g2]-ldd3-usb-ch> wow
May 17 17:29:28 <dyoung-web> Wow Lau Lau.
May 17 17:29:33 <dyoung-web> We go Kau Kau.
May 17 17:29:36 <[g2]-ldd3-usb-ch> T:  Bus=01 Lev=01 Prnt=01 Port=00 Cnt=01 Dev#=  2 Spd=480 MxCh= 0
May 17 17:29:36 <[g2]-ldd3-usb-ch> D:  Ver= 2.00 Cls=00(>ifc ) Sub=00 Prot=00 MxPS=64 #Cfgs=  1
May 17 17:29:36 <[g2]-ldd3-usb-ch> P:  Vendor=05dc ProdID=a410 Rev=30.00
May 17 17:29:36 <[g2]-ldd3-usb-ch> S:  Manufacturer=LEXAR MEDIA
May 17 17:29:36 <[g2]-ldd3-usb-ch> S:  Product=JUMPDRIVE
May 17 17:29:36 <dyoung-web> ??
May 17 17:29:37 <[g2]-ldd3-usb-ch> S:  SerialNumber=0A4E................
May 17 17:29:39 <[g2]-ldd3-usb-ch> C:* #Ifs= 1 Cfg#= 1 Atr=80 MxPwr=100mA
May 17 17:29:41 <[g2]-ldd3-usb-ch> I:  If#= 0 Alt= 0 #EPs= 2 Cls=08(stor.) Sub=06 Prot=50 Driver=usb-storage
May 17 17:29:43 <[g2]-ldd3-usb-ch> E:  Ad=81(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
May 17 17:29:45 <[g2]-ldd3-usb-ch> E:  Ad=02(O) Atr=02(Bulk) MxPS= 512 Ivl=31875us
May 17 17:30:07 <dyoung-web> What am I wowing at?
May 17 17:30:17 <[g2]-ldd3-usb-ch> mounting usbfs I can read the serial number of the compact flash stick
May 17 17:30:18 <dyoung-web> the power?
May 17 17:30:27 <dyoung-web> oh.
May 17 17:30:36 <dyoung-web> neato .
May 17 17:30:44 <[g2]-ldd3-usb-ch> can you say fingerprint ? :)
May 17 17:31:31 <[g2]-ldd3-usb-ch> Rev 30.00 looks funny :)
May 17 17:32:50 <[g2]-ldd3-usb-ch> Hey the Maxtor has a serial # too :)
May 17 20:10:42 <ka6sox-laptop> dyoung-web: ping
**** ENDING LOGGING AT Tue May 17 23:59:56 2005