**** BEGIN LOGGING AT Sat Apr 07 02:59:57 2007 Apr 07 03:00:34 does it use 2x5 header or did you customize it to use something like here: http://www.nslu2-linux.org/gallery/jtag/LandslideJTag Apr 07 03:00:58 except I also provided 3.3V and GND as follows (same pinout as above I believe): Apr 07 03:01:48 I used a single 1x5 header, and mounted in on top of the ethernet port Apr 07 03:02:00 oops I am wrong; I used a different pinout on my JTAG. I did it as follows: Apr 07 03:02:11 TDI TMS TCK VSS VCC TDO TRSTn Apr 07 03:02:41 even though TRSTn isn't used, it could have been used if I had a real wiggler. But openwince stuff didn't need it Apr 07 03:02:56 I just had tdi, tms, tdo, tck, gnd and 3.3V - no tsrtn Apr 07 03:03:29 if you can do one the same as the digilent parallel cable, that would be sweet. Apr 07 03:03:30 correct, trstn isn't needed, if you power up the board after connecting to printer port (holds it in reset) Apr 07 03:04:59 what is pinout of digilent? Apr 07 03:06:19 this is the cable: http://www.nuhorizons.com/products/digilent/jtag-cable.html Apr 07 03:06:36 and you can even read the pinout on that picture :-) Apr 07 03:08:19 gotcha... my 1x7 header used gnd as center pin (mainly b/c it was closest to ground of thru-hole cap) for rigidity Apr 07 03:08:53 rwhitby: what's inside that JTAG cable ? Would a simple circuit around FT245R be enough for "usb jtaging" ? Apr 07 03:09:44 pierrelux: I didn't use this; I used the instructions here: http://www.nslu2-linux.org/wiki/HowTo/RecoverFromABadFlashUsingJTAG Apr 07 03:09:54 pierrelux: dunno - never opened it. it's a fairly standard cable. Apr 07 03:10:09 245 is just a bidir buffer, just set the direction one-way and it works just like a 244 Apr 07 03:10:37 oh wait USB JTAG forget it Apr 07 03:11:08 you need some serious work on that. some kind of microcontroller, USB driver, etc. Apr 07 03:13:05 and plus, timing is pretty important with JTAG so you'd need some way to buffer up stuff as well. I heard that usb1.1 devices would be slower than parallel-port because of the handshaking, etc. required. Not sure about USB2.0 but that's not an enthusiast's market yet Apr 07 03:13:48 rwhitby: if JTAG is there, shouldn't be any reason it can't be programmed via openwince tools Apr 07 03:14:41 RobNC: do you have a jtag cable with a compatible pinout? Apr 07 03:14:59 I can make a dongle from my pinout to yours if needed Apr 07 03:15:01 (i.e. so you can test the 16MB flash before shipping) Apr 07 03:15:38 rwhitby: yeah that's a no-brainer Apr 07 03:17:01 sweet Apr 07 03:17:16 monotone took forever to compile. doing "make check". then I should be able to continue with the build env. Apr 07 03:19:46 tomorrow might be the best day to do it too... provided no one is in the lab and will bug me! :-) Apr 07 03:20:36 let me know if you need any other info (for the board, or for the build env) Apr 07 03:21:29 yeah, you only want bare PCB? shipping to my house was only 2lb, for entire box. Not sure delta in $$ between 1lb and 2lb to aus. Apr 07 03:23:20 RobNC: http://www.usps.com/global/globalexpressmail.htm Apr 07 03:24:24 this one looks neat: http://www.digilentinc.com/Products/Detail.cfm?Prod=JTAG-USB&Nav1=Products&Nav2=Cables Apr 07 03:24:27 should be able to get the $16.25 rate with the bare NSLU2 (board & plastic case, no power supply, no cables) Apr 07 03:24:44 RobNC: I do want the case :-) Apr 07 03:25:38 ok np! :-) (had to superglue the bottom tab on the faceplate as it broke off but it should be better now. Apr 07 03:26:19 rwhitby: is usps cheaper than fedex and ups? Apr 07 03:26:36 Global Express Mail to Australia is the cheapest I believe. Apr 07 03:26:54 (cheapest with tracking that is) Apr 07 03:28:07 can oe use monotone > 0.29 yet ? Apr 07 03:28:30 gotcha global express guaranteed - non-document Apr 07 03:28:51 oh crap don't tell me that - I'm building monotone 0.34 - is it not gonna work?? Apr 07 03:29:19 didn't used to Apr 07 03:29:30 0.34 is fine Apr 07 03:29:49 I have built 0.34 and am "make check"ing right now. Apr 07 03:29:52 you just need to mtn db migrate from 0.33 Apr 07 03:30:08 RobNC: how heavy is PCB + case only in lbs? Apr 07 03:30:32 fedex 2 lbs is $70 Apr 07 03:30:47 rwhitby: the whole package, which padding, was 2lbs according to amazon Apr 07 03:30:58 Ouch! Apr 07 03:31:14 that's right can't go fedex ground to aus :-) Apr 07 03:33:36 rwhitby I don't know how much for pcb+case - but it has to be less than 1lb, perhaps 0.5lb? Apr 07 03:35:16 usps global express mail (the site I linked) is the best way to send it. Apr 07 03:35:41 I've had great success with that previously when ka6sox and dyoung have sent me stuff. Apr 07 03:35:49 rwhitby ok np... first gotta mod the slug :-) Apr 07 03:36:07 RobNC: ka6sox is coming here now to confirm Apr 07 03:37:09 ka6sox-laptop: http://www.usps.com/global/globalexpressmail.htm right? Apr 07 03:37:17 yes Apr 07 03:37:37 $16.25 ? Apr 07 03:37:42 thats the best way (some days they say where he lives is off their map. Apr 07 03:37:54 (for slug board in case, no PSU, no other stuff) Apr 07 03:37:54 I think so if it fits in the box Apr 07 03:38:15 rwhitby holy smokes ups is $82 wow Apr 07 03:38:30 ka6sox-laptop: it'll be going to Adelaide city now, instead of 15 kms out of the city. Apr 07 03:38:43 right Apr 07 03:38:59 ka6sox ok cool... sometimes it can be cheaper if to business versus residence Apr 07 03:39:09 ka6sox-laptop: RobNC is making me a 256MB/16MB fatslug :-) Apr 07 03:39:26 excellent Apr 07 03:40:31 just so I don't forget anything: 4-pin serial header, 16M flash, 4x32x16 SDRAM w/CS jumper, de-underclock, JTAG, ? Apr 07 03:41:02 RobNC, some of the new ones don't require underclocking Apr 07 03:41:11 ka6sox this one does !! :-( Apr 07 03:41:23 ah Apr 07 03:41:28 an old stock one :D Apr 07 03:41:39 I would have thought not. You'd think that the old ones would be all gone Apr 07 03:42:12 no other changes, right, in regards to BOM, or PCB, for new ones? Apr 07 03:42:36 iirc they removed the resistor so it goes at full speed Apr 07 03:43:04 that's it; that's no biggie :-) Apr 07 03:43:11 right Apr 07 03:44:00 RobNC: yep, that list of mods is perfect :-) Apr 07 03:44:15 ok, cool... will get back to you rwhitby when mods are complete. gotta run - honey-do list :-) Apr 07 03:44:21 and tested ? Apr 07 03:44:35 may come back later after make check is done (for monotone) Apr 07 03:45:06 yes of course... I'll test it here on-line. Prolly won't flash it in the programmer, just use JTAG Apr 07 03:45:15 sweet. Apr 07 03:45:18 have a good easter! Apr 07 03:45:24 thanks, you too! Apr 07 04:37:06 hello all :) Apr 07 04:38:09 is there a live person here? Apr 07 04:38:31 I have a question re: kernel on nslu2 in debian Apr 07 04:38:45 on debian in nslu2... Apr 07 04:39:02 i need to have a 1000Hz ticks per sec kernel Apr 07 04:39:08 is it doable? Apr 07 05:01:55 03mwester * r812 10kernel/trunk/patches/2.6.21/06-dsmg600-upstream-support.patch: Reviewed and signed-off Apr 07 07:46:19 03blaster8 * r813 10kernel/trunk/patches/2.6.21/06-dsmg600-upstream-support.patch: DSMG-600 support submitted upstream Apr 07 08:02:05 03blaster8 * r814 10kernel/trunk/patches/2.6.20/ (00-linux-2.6.20.4.patch 00-linux-2.6.20.6.patch series): Update to 2.6.20.6, fixes remote crasher bug Apr 07 08:06:28 03blaster8 * r815 10kernel/trunk/ (README STATUS): Update docs Apr 07 08:11:52 03blaster8 07lopatch * r816 10kernel/patches/2.6.20/ (00-linux-2.6.20.4.patch 00-linux-2.6.20.6.patch series): Sync lopatch branch with trunk Apr 07 08:32:13 \\ Apr 07 11:41:09 03blaster8 * r817 10kernel/trunk/ (5 files in 2 dirs): Add trivial mach-ixp4xx Kconfig cleanup patch Apr 07 15:39:42 RedBoot is BE right? Apr 07 15:40:04 Yes Apr 07 15:40:24 thanks... devising some way to test RAM from RedBoot Apr 07 17:34:18 hi Apr 07 18:33:56 03bzhou * r5900 10optware/trunk/make/fetchmail.mk: fetchmail: 6.3.7 -> 6.3.8 Apr 07 18:52:01 03bzhou * r5901 10optware/trunk/make/git-core.mk: git-core: 1.4.4.4 -> 1.5.0.7, use PERL_HOSTPERL if available Apr 07 19:50:47 NAiL: see my comments in #oe. I don't know if this is slug specific or not Apr 07 19:51:31 no idea. I stay far away from libtool Apr 07 19:51:44 I believe libtool is contagious and causes braindamage. Apr 07 20:16:46 it may Apr 07 20:16:59 but it definately causes damage to my builds Apr 07 20:17:12 since it doesn't properly populate staging Apr 07 20:18:18 NAiL: can you build libtool-native just to confirm that it's not populating staging/i686-linux/share/libtool/libltdl ? Apr 07 20:18:59 sorry, not right now Apr 07 20:19:20 np Apr 07 20:20:09 puzzled Apr 07 20:21:16 could nano (even when using -w switch) break linuxrc file while editing? Apr 07 20:24:19 it shouldn't Apr 07 20:31:32 NAiL: it appears the libtool problem is either specific to slugos or specific ot my build enviroment Apr 07 20:37:10 part of "make slugosle" - says "Psyco JIT Compiler not available". Going to http://psyco.sf.net shows several versions... 1.5.2-linux-i386-2.3 2.4 and 2.5.tar.gz? Apr 07 20:37:59 RobNC: only available on ix86 Apr 07 20:38:03 not x64 Apr 07 20:38:06 or other archs Apr 07 20:38:21 not usable by arm? Apr 07 20:38:37 what host are you building on? Apr 07 20:38:42 i386 Apr 07 20:38:48 (i686 actually) Apr 07 20:39:13 should be fine. Figure out which python version you have, and install the psyco version that matches Apr 07 20:39:31 Ah ok that's the python version (second number) Apr 07 20:39:41 yes Apr 07 20:41:27 well, does anyone have a clue why I'm able to modify everything else in miniroot.gz except that file (linuxrc)? if I'm making a slightest edit - like adding a comment line # - boot breaks Apr 07 20:42:01 sounds crazy Apr 07 20:42:33 Thanks Nail. "python -V" does that Apr 07 23:36:10 anyone know how to resolve issues with "make slugosle" - looks like problem with makefile again... get some 404 Not Found errors Apr 07 23:57:59 RobNC: sounds like a changed upstream URL Apr 07 23:58:14 RobNC: which tarball do you need? Apr 07 23:58:18 RobNC: if there are sources missing, you can download them from sources.nslu2-linux.org ... Apr 07 23:58:27 yep, gonna do make update-master again. Apr 07 23:58:36 our autobuilders upload all required tarballs there Apr 08 00:07:42 CoreDump|home: thanks - looks like there are still issues with files missing - looks like Intel changed their website? Apr 08 00:08:30 ohh, the microcode needs to be downloaded manually Apr 08 00:08:41 rwhitby: thanks, trying to find what is needed... Apr 08 00:09:46 http://www.intel.com/Please-Read-The-BB-File/IPL_ixp400NpeLibrary-2_3.zip Apr 08 00:10:00 heh right Apr 08 00:11:33 looks like there's something else - ftp.osuosl.org - another 404 Apr 08 00:11:44 that's the mirror Apr 08 00:11:54 but we're not allowed to mirror the microcode Apr 08 00:12:05 "Please-Read-The-BB-File/..." Apr 08 00:12:19 I read that - but gottta read the BB file :-) Apr 08 00:12:48 you have to download it yourself, because you personally have to read, understand, and accept the Intel license. We can't accept it on your behalf. Apr 08 00:13:02 yep, heaven knows I've heard that once or twice :-) Apr 08 00:16:32 I take it that BB is bitbake. Never saw that mentioned but there are some clues here: http://www.nslu2-linux.org/wiki/Development/DownloadingTheIntelSourceCode Apr 08 00:18:22 RobNC: let us know if there is something clearer we could write there - the only thing we have is the failing URL - we have no other way of printing instructions in the build. Apr 08 00:19:52 IMHO, I wouldn't even go to the URL. I would do like what is done during the "make setup-host-centos" - where it refers you to download the monotone from http://venge.net/monotone Apr 08 00:22:34 I.e., if it finds the src tarball, then fine, continue. Else, abort and echo the message to download it from that Intel site. According to the Wiki, we're still using 2.1 even though Intel has updated it to 2.3.1 (only need 2.3.1 for IXP42x rev B1) Apr 08 00:25:21 rwhitby: you said you have no way of printing instructions during the build. Is that b/c NSLU2 project doesn't "own" the master makefile or something else? Apr 08 00:26:03 cause the error occurs during do_fetch in OE, a really low-level function Apr 08 00:26:24 master makefile could print something, but it would be nowhere near the actual error Apr 08 00:27:45 ok, gotcha... it would be hard to track. And it's not possible to have it even attempt the download from Intel (which will fail anyway); I'm sure I don't understand enough about this do_fetch process. Apr 08 00:34:50 I have to ask b/c it's not in the Wiki but there are the following things on the Intel website: (for v2.1): 1) Intel Hardware Access Software, 2) NPE Microcode, 3) Linux Devel Environment, 4) Redboot Boot Loader (which has 4 separate downloads) Apr 08 00:37:59 microcode (no crypto) Apr 08 00:38:11 please update wiki :-) Apr 08 00:45:38 YES I hate reading instructions that are half-baked :-) Apr 08 00:47:03 Just a dumb question but since I'm here, might has well get redboot eh? Apr 08 00:47:17 (in case rebuilding it is required for FatSlug) Apr 08 00:47:45 RobNC: we'd go for Apex instead of trying to rebuild RedBoot. Apr 08 00:48:53 rwhitby: ok, gotcha... that presumes it'll work as 2nd stage (not yet for me at least) Apr 08 00:49:17 we have built RedBoot before for the NAS100d (http://nslu.cvs.sourceforge.net/nslu/ludeboot/) Apr 08 00:51:32 so if need be, we would use that for the nslu2 with more flash and ram. but it's never been tested as a first-stage redboot, only as a second stage for the nas100d (which doesn't have network support in the first stage). Apr 08 00:52:13 the kernel will see the flash due to the cfi_probe routine in the board startup code. Apr 08 00:52:29 and we can always override ram size in apex. Apr 08 00:52:45 (nas100d and dsmg600 have 16MB flash, so we know how to work with that) Apr 08 00:53:48 RobNC: [g2] also builds redboot for ixp4xx regularly for his Loft boards Apr 08 00:53:52 yes, that is true... apex should work, it may just be the handoff that is broken Apr 08 00:54:17 and I got Apex 2nd stage to work with 64MiB detect (forced) but not 256MiB Apr 08 00:54:33 did apex fail, or did the kernel fail? Apr 08 00:54:56 Wiki states for download you need two files: Intel Hardware Access Software (driver software) and the NPE Microcode. Is this still true? (updating Wiki as we speak) Apr 08 00:55:01 we already suspect a kernel boundary at 64MB due to a known (but not understood, at least by me) dmabounce problem Apr 08 00:55:21 no longer true, since we use the open source driver with the proprietary microcode Apr 08 00:55:22 During the 256MiB, loading of the kernel from 2nd stage Apex failed. It just hung. Apr 08 00:55:54 ok, will update the Wiki to remove Intel Hardware Access S/W (since it's now open-source) Apr 08 00:56:07 RobNC: just being able to reliably write and read back from various parts of the 256MB without aliasing will be enough testing for me to be satisfied Apr 08 00:56:28 (Intel SW is not open source, we now have an open source alternative written from scratch) Apr 08 00:56:54 yes you are right I meant to say that; we have an alternative now (that is OSS) Apr 08 00:56:58 (write and read back from second stage Apex, for example) Apr 08 00:57:59 back later (taking kids to the park) - just leave any messages or questions in the log for me. Apr 08 01:01:25 ok thanks rwhitby - updating wiki, yada yada Apr 08 01:09:08 holy bloody shit Apr 08 01:09:37 * CoreDump|home just stumbled upon the mother of all pir4te sites Apr 08 01:15:31 is there a reason I couldn't include the actual filename "IPL_ixp400NpeLibrary-2_1.zip" in the "DownloadingTheIntelSourceCode" Wiki? Apr 08 01:18:07 no reason at all Apr 08 01:18:15 I think we're using 2.3 though Apr 08 01:21:25 Ok, more wiki oldness. Updated to v2.3. Apr 08 01:27:39 This is getting old - looks like Makefile is looking for IPL_ixp400NpeLibrary-2_3.zip but only IPL_ixp400NpeLibrary-2_3_2.zip is avail on Intel website Apr 08 01:29:08 Wonder is it safe to rename IPL_ixp400NpeLibrary-2_3_2.zip to IPL_ixp400NpeLibrary-2_3.zip Apr 08 01:35:48 seems to be working (so far) Apr 08 01:47:08 connection timing out to http://matt.ucc.asn.au/ Apr 08 01:47:38 get it from sources.nslu2-linux.org Apr 08 01:48:01 heh - no changelog for the microcode, so you have no idea what changed. Apr 08 01:50:57 true! ok, will do. Apr 08 01:53:29 Thanks rwhitby.. at the risk of seeming obtuse, this is all I get: "This is the place for archives." and FTP is not supported *dumb looks* Apr 08 01:59:53 I see - seems like sources.nslu2-linux.org redirects to a mirror - http://ftp.osuosl.org/pub/nslu2/sources/ Apr 08 02:02:03 surely IPL_ixp400NpeLibrary-2.3.zip isn't gonna be there since there's no way to agree to licensing conditions from Intel. Apr 08 02:03:02 nm - this was for the failing connection timing out. After some time it went to another mirror. Apr 08 02:04:31 These look really nifty for adding a serial port to the slug. http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=110066550130 Apr 08 02:04:51 cool - it finished slugosle. now doing slugosbe Apr 08 02:05:39 joshin - hard to find through-hole parts these days!!! :-) Apr 08 02:06:21 I think the USB prolific driver idea with 4-pin header is probably the most stable. Anyone know what futuredial cable can be used for that? Apr 08 02:06:32 Yeah. :( Apr 08 02:09:23 even RatShack doesn't have a lot of resistors left. Good luck getting them to tell you what a capacitor is Apr 08 02:49:32 RobNC: seen this? http://www.cse.unsw.edu.au/~cs9242/06/project/slug.shtml Apr 08 02:51:35 rwhitby: no I hadn't - intriguing... looking at the schematic **** ENDING LOGGING AT Sun Apr 08 02:59:57 2007