**** BEGIN LOGGING AT Tue Sep 19 02:59:57 2006 Sep 19 11:38:11 <[g2]> what GNU tool chain are you guys using for the ARM micros. Crosstool ? Sep 19 15:24:43 <[g2]> ka6sox-office ping Sep 19 15:25:14 <[g2]> wookey_ around ? Sep 19 15:27:07 [g2]: yes Sep 19 15:27:47 <[g2]> wookey_ it's been forever since we last chatted. What's up with ballon v3 ? Sep 19 15:28:05 I got my hands on one at the w/e :-) Sep 19 15:28:19 First batch is in use at Cambrdige Eng dept Sep 19 15:28:39 respin being worked on in time for xmas Sep 19 15:29:21 (to fix various probs in initial board) Sep 19 15:29:35 FPGA and CPLD both seem to work and support for FPGA is now in bootloader Sep 19 15:29:52 <[g2]> sounds great Sep 19 15:30:00 Right now I am trying to work out why bootldr modified for B3 seems to be a bit broken on B2 Sep 19 15:30:21 <[g2]> no fpga ? Sep 19 15:30:27 can program files into yaffs filsystem, but direct programming of whole flash partition sseems to do nothing Sep 19 15:30:40 FPGA code is compiled out for this variant Sep 19 15:30:48 <[g2]> I was just teasing :) Sep 19 15:31:01 <[g2]> how do you like yaffs ? Sep 19 15:31:05 sorry - staring at code too long Sep 19 15:31:10 <[g2]> np :) Sep 19 15:31:20 <[g2]> I'm the one who should be sorry Sep 19 15:31:24 It's very cool Sep 19 15:31:36 <[g2]> alot cooler than jffs2 ? Sep 19 15:31:40 but we've had hassles with MTD integration which makes it a bit of a pain on linux Sep 19 15:31:46 faster than jffs2 Sep 19 15:31:59 <[g2]> is jffs3 out yet ? Sep 19 15:32:03 (but no compression built in to fs - big part of why it's faster) Sep 19 15:32:13 I understand it's being used on OPLC project Sep 19 15:32:15 <[g2]> ahh.. that makes sense Sep 19 15:32:31 so nice simple garbage collection, deterministic behaviour Sep 19 15:32:43 but you need to compress the data yourself Sep 19 15:32:52 <[g2]> a fair trade-off Sep 19 15:33:10 some korean guy is doing a load of comparative testing at the moment Sep 19 15:33:11 <[g2]> lord knows NAND is getting cheaper by the minute Sep 19 15:33:14 I look forward to his results Sep 19 15:33:52 yaffs got checkpointing in June (so boot scanning only happens if shutdown was not normal) Sep 19 15:34:04 huge speed increase on big flash drives Sep 19 15:34:15 I gather JFFS2 has also acquired similar facility Sep 19 15:34:54 I am reasonably hopeful yit will get into mainline with next kernel iteration Sep 19 15:36:08 [g2]: are lofts actually available to buy? didn't seem entirely clear from your website Sep 19 15:36:34 <[g2]> wookey_ yes they've been available for 8-9 months Sep 19 15:36:56 <[g2]> I've got to run right now, phone call but would love to talk later Sep 19 18:44:07 vmaster: tip for the day,, don't leave a defunt copy of openocd running for more than 24 hours, hehe Sep 19 18:45:57 prpplague: oh, what happened? Sep 19 18:46:47 vmaster: i powered off my device yesterday with an openocd running, check the box just a little while ago, the load level was almost 50 Sep 19 18:47:05 killed off openocd and it dropped to normal Sep 19 18:53:53 ah, hehehe Sep 19 20:57:52 <[g2]> vmaster which tool chain are you using for the SAM7 stuff ? Sep 19 20:58:50 * [g2] just used T0mW scripts to build a 4.0.2/newlib config Sep 19 20:59:14 <[g2]> I guess insight too Sep 19 21:00:28 i prefer codesourcerys toolchain, but the libraries from their latest release are built for armv5 Sep 19 21:00:39 codesourcery has the best gdb, imho Sep 19 21:03:31 i've used gnuarm, too, but they only provide 64-bit linux binaries these days Sep 19 21:16:12 <[g2]> I guess you don't build from source then Sep 19 21:16:26 <[g2]> and openocd interfaces to gdb right ? Sep 19 21:20:32 yeah, i don't build from source - prebuilt binaries are fine for the small testcases that i usually compile Sep 19 21:22:10 the openocd interfaces to gdb, yes Sep 19 21:22:26 it implements a gdb remote protocol server Sep 19 21:22:33 * [g2] hugs vmaster **** ENDING LOGGING AT Wed Sep 20 02:59:56 2006