**** BEGIN LOGGING AT Thu Jun 09 23:59:57 2005 Jun 10 00:00:04 then unsling to the disk, and reboot, and don't touch the web upgrade page (it has a button to exit maintenance mode when booted from a disk - don't push that button) Jun 10 00:00:19 check the df to make sure you are booting from the external disk and not the ramdisk Jun 10 00:18:23 03jeanfabrice * 10unslung/ (7 files in 3 dirs): Promoted wiley/poptop, a pptp server solution (courtesy of phedny from Chupa Forum) Jun 10 00:18:54 good morning all Jun 10 00:24:32 rwhitby-away: ping Jun 10 00:37:40 a question about the dhcp daemon, should it now be looking at /opt/etc/ for the configuration file instead of /etc/? Jun 10 00:38:21 there a two ways to change it, 1. change the Makefile.conf --> ETC = /etc Jun 10 00:38:43 2. include a -cf /opt/etc/dhcpd.conf into the S56dhcp startup script Jun 10 00:39:21 can anybody change the buildsystem please? Jun 10 01:47:25 03rwhitby * 10unslung/make/libtopfield.mk: Tagged at 0.5.0. Jun 10 01:48:02 03rwhitby * 10unslung/make/ftpd-topfield.mk: Updated to version 0.5.1 Jun 10 01:48:52 version 0.5.1 puts the ftpd-topfield.pid in the right place. Jun 10 01:53:14 rwhitby-away, your success with libusb is my inspiration Jun 10 01:53:29 I'm trying once again to get my camera working with the slug Jun 10 01:53:47 not having much luck but that could be bad libptp2 Jun 10 01:55:21 rwhitby-away: ftpd-topfield failed to build on nudi. any hint ? Jun 10 01:55:25 so now I'm trying latest gphoto2 Jun 10 01:56:29 jf-work: will check tonight Jun 10 01:56:39 does it get the cvs tag? Jun 10 01:58:02 what's this ? Jun 10 02:02:16 rwhitby, I'm still looking over what the differences are in the current zd1211 and the patch set that you forwarded to me. Jun 10 02:02:23 Its just taking a long time. Jun 10 02:10:14 jf-work: looks like sf.net public cvs hasn't updated yet - it will work in 6 hours or so Jun 10 02:11:29 dyoung: no rush Jun 10 02:12:21 ok Jun 10 02:35:13 03rwhitby * 10unslung/make/ftpd-topfield.mk: Revert back to previous tag until sf.net updates. Jun 10 02:35:47 gphoto2 segfaults just like ptpcam - I'm starting to suspect libusb again Jun 10 02:35:57 03rwhitby * 10unslung/make/ftpd-topfield.mk: Put the version back too. Jun 10 02:36:00 jf-work: try now - there will be another update tomorrow Jun 10 02:36:25 I'm building Jun 10 02:36:45 it will build 0.5.0 now, and 0.5.1 tomorrow Jun 10 02:37:22 hmm - Makefile.patch needs updating Jun 10 02:39:16 03rwhitby * 10unslung/sources/ftpd-topfield/Makefile.patch: Fixed the patch for 0.5.0 Jun 10 02:39:38 03rwhitby * 10unslung/sources/ftpd-topfield/Makefile.patch: Fixed the pid path. Jun 10 02:40:01 ok. that builds now Jun 10 02:41:30 ok Jun 10 02:41:34 jacques, which is the good shuttle mini barebones system to get these days? Jun 10 02:41:56 sn25p Jun 10 02:42:28 searching... Jun 10 02:44:32 I wonder if its loud... Jun 10 02:45:37 we discussed it quite a bit a few months ago Jun 10 02:46:01 along with me posting links to reviews which included dB charts Jun 10 02:46:24 jbot, status Jun 10 02:46:31 Since Wed Jun 8 17:08:31 2005, there have been 7 modifications, 123 questions, 0 dunnos, 0 morons and 118 commands. I have been awake for 1d 16h 37m 59s this session, and currently reference 108761 factoids. I'm using about 19500 kB of memory. With 0 active forks. Process time user/system 3298.75/145.1 child 214.58/16.73 Jun 10 02:46:54 there is a patch to store files in /opt instead of /, can anybody checkin the code please? http://hashphp.org/pastebin.php?pid=3609 Jun 10 02:48:05 Debating replacing my 3.06Ghz build box. Jun 10 02:49:19 dyoung, http://techreport.com/reviews/2005q1/shuttle-sn25p/index.x?pg=1 Jun 10 02:49:33 http://techreport.com/reviews/2005q1/shuttle-sn25p/index.x?pg=16 Jun 10 02:50:03 http://www.newegg.com/Product/Product.asp?Item=N82E16856101468 Jun 10 02:51:05 mr_claus: looks good to me Jun 10 02:51:19 rwhitby-away: i teste the package, it works Jun 10 02:51:30 who's listed as the maintainer? Jun 10 02:52:03 in the makefile? nobody Jun 10 02:52:19 jacques, cool. Jun 10 02:52:21 thanks Jun 10 02:52:38 mr_claus: got an sf.net id? Jun 10 02:52:39 np Jun 10 02:52:58 Just gotta find me a doo-doo video card. Jun 10 02:54:46 rwhitby-away: 430840 if you mean the userid Jun 10 02:55:17 heh, make that a pci-e card Jun 10 02:56:41 I'm so behind the times now. Does that replace AGP as the preferred video interface? Jun 10 02:57:24 mr_claus: no, the user name Jun 10 02:57:31 dyoung: yes Jun 10 02:57:35 dyoung: yeah, pci-express Jun 10 02:57:36 rwhitby-away: mr_claus Jun 10 02:58:19 dyoung, it replaces agp and pci Jun 10 02:59:04 most new motherboards only have pci-e slots, of varying widths Jun 10 02:59:11 1x 4x 8x 16x Jun 10 02:59:17 mr_claus: you now have CVS write access - use it wisely Jun 10 02:59:19 Wow. I wonder if a doo-doo pci-e card even exists... This machine doesnt need high end graphics. Must read.... Jun 10 03:00:13 rwhitby-away: hmm, i will try to leave all things intact Jun 10 03:00:14 And actually if the Xilinx ISE works under linux now, I can replace *two* machines with this one. Jun 10 03:00:24 rwhitby-away: there is a maitainer Jun 10 03:00:40 rwhitby-away: but i dont have the newest source there was a change two days ago Jun 10 03:01:19 rwhitby-away: i will send the patch to the maintainer, would that be the better way? Jun 10 03:01:41 mr_claus: who is the maintainer? Jun 10 03:02:11 Christopher. Jun 10 03:02:13 dyoung, there are some relatively inexpensive pci-e videocards, but I don't think I've seen any $30 ones or anything Jun 10 03:02:24 christopher blunck Jun 10 03:02:45 morning [g2] Jun 10 03:02:56 mr_claus: I think you're safe to update it. Christopher will be cool with it. Jun 10 03:03:12 just send him an email at his sf.net address to let him now you've done it Jun 10 03:03:13 rwhitby-away: ok Jun 10 03:03:21 (and that you discussed it in the channel first) Jun 10 03:03:38 It may be worthwhile to update the .mk file to the new template too, while youre in there. Jun 10 03:03:51 <[g2]> hey jacques Jun 10 03:04:02 <[g2]> morning rwhitby-away and dyoung Jun 10 03:04:08 Hi [g2] Jun 10 03:04:33 i'm sad, my two 160GB WD Drives sound like chainsaws now. Jun 10 03:04:45 now that... ? Jun 10 03:05:04 <[g2]> DOH! Jun 10 03:05:29 I havent had much luck with hardware recently. Jun 10 03:05:54 My slugs are punishing me for ignoring them I guess. Jun 10 03:06:45 The data is intact; just real noisy. Jun 10 03:07:59 Time to rebuild MediaSlug, better, stronger, faster..... Jun 10 03:14:38 <[g2]> Do you guys know how long we've been doing this stuff ? Last night I rebuilt the latest crosstool when I was talking to a guy in #elinux.. Jun 10 03:14:54 <[g2]> My last version was .28-rc37 Jun 10 03:15:08 <[g2]> I downloaded and built .35 Jun 10 03:15:34 <[g2]> man it's been quite some time since we played with that! Jun 10 03:15:58 almost a year. Jun 10 03:16:12 10 months or so? Jun 10 03:18:05 <[g2]> and what a great 10 months Jun 10 03:25:15 hmmm, pb is messing with binutils version in oe Jun 10 03:41:50 mr_claus: you need to check it out with :ext: to be able to check back in again Jun 10 03:42:09 and we just use basic CVS commands for checkin Jun 10 03:45:11 can someone flash 3.18 and find that upgrade.cgi script which makes the 8MB ramdisk and frees it to make room for maintenance mode upgrade? Jun 10 04:15:31 do i have to be root to checkout the sources? Jun 10 05:05:35 mr_claus: no, we recommend that you do *not* build as root. we do require you to be able to run sudo if you want to build crosstool-native Jun 10 05:06:21 dyoung: I finally found that upgrade.cgi code. It was in switchbox, and didn't make it into 4.x - and that's why maintenance mode won't work in 4.x Jun 10 05:07:29 well, we did kinda rush that switchbox removal into service..... Jun 10 05:08:39 rwhitby: hmm, i can checkout as user root but not with another user Jun 10 05:09:01 cvs.sourceforge.net: Connection timed out Jun 10 05:09:08 mr_claus: nobody else checks out as root Jun 10 05:09:10 if i'm not user root i get that error message Jun 10 05:09:15 glc: hi Jun 10 05:11:49 argll, my fault, sry Jun 10 05:22:53 I'm totally slugless now. Jun 10 05:23:12 I cant find my devslug, and MediaSlug's drive was sounding like a chainsaw. Jun 10 05:23:18 so its all off now. Jun 10 05:23:28 can't find .... ? Jun 10 05:23:37 yeah, its lost. Jun 10 05:23:43 you should put a picture of your development room sometime ... Jun 10 05:23:49 probably buried under a pile of '244's Jun 10 05:24:03 You would cringe. Jun 10 05:24:04 I know they're small but... Jun 10 05:24:19 my dev room is... um... Messy. Jun 10 05:24:29 go on - put up a piccie Jun 10 05:24:33 you've seen my shelf ... Jun 10 05:24:38 Its been lost for weeks. Jun 10 05:25:01 last time I saw it I was testing something for someone, I dont even remember what it was for. Jun 10 05:26:06 okay i'll post up some "dyoung secret laboratory" pics later. When I find my camera. :-) Jun 10 05:33:12 oh no - new binutils downloading when building unslung .... Jun 10 05:33:50 jacques: do I let it go, or do I pin the version back? Jun 10 05:36:47 * rwhitby-away decides to pin the version for now and let someone else be the bunny Jun 10 05:38:00 rwhitby-away, I intend to test the new version, but pinning it back is the wiser course IMHO Jun 10 05:38:23 I'll start a build now Jun 10 05:39:19 cool - I'm trying to get the maintenance mode working again, so I'm pinning back Jun 10 05:40:17 yeah, nothing can screw up a toolchain like an untested binutils Jun 10 05:41:24 I'm clearing my .ccache now, then I'll start a frech openslug-image build Jun 10 05:48:56 is the binutils lock pushed? Jun 10 05:50:12 pushing now Jun 10 05:50:24 03rwhitby 07 * r1.3468 10openembedded/ (4 files in 3 dirs): Unslung 5.2-alpha: added upgrade-real.cgi back, and locked the binutils version Jun 10 05:50:37 just about to flash that .... Jun 10 05:55:31 Firmware Version: V2.3R29-uNSLUng-5.2-alpha Jun 10 05:57:20 hopefully 5.x will have 'x' < 5 this time around .... Jun 10 06:02:45 exit maintenance mode works ... Jun 10 06:03:07 woo Jun 10 06:10:05 do you reckon we should just release this as a 4.x ? Jun 10 06:10:59 it is kind of a repair to 4.x Jun 10 06:11:22 yeah, that's what I'm thinking, since we're not removing maintenance mode Jun 10 06:12:00 or we could just let the masses wait. Jun 10 06:12:23 and release it wit the R29 base. Jun 10 06:13:27 it does have the R29 base now Jun 10 06:13:41 (by "this" I mean a tested version of what we have now) Jun 10 06:13:56 has Tiersten added those usb devices patch yet? Jun 10 06:14:06 Right, what I'm saying is that 5.x will be the R29 base. Its kind of a jump to release a 4.x with a R29 base eh? Jun 10 06:16:01 not really - it's an R29, not a V2.4 Jun 10 06:18:40 now, any README suggestions on the wiki to add ..... Jun 10 06:18:54 We should ask the new Unslung-manager-elect what he thinks. Jun 10 06:19:05 Note much at http://www.nslu2-linux.org/wiki/Unslung/SuggestedChangesToTheReadme Jun 10 06:19:41 s/Note/Not/ Jun 10 06:19:59 So much bellyaching about the doco, but only one suggestion on that page .... Jun 10 06:24:11 Hows the "for dummies" initiative going? Jun 10 06:24:59 dunno, haven't checked the yahoo group Jun 10 06:25:30 So should we just go and change the root password to always be "uNSLUng" all the time, in all relevant files? Jun 10 06:25:49 more additions to the SlugArmy. Jun 10 06:26:16 thats basically the first thing I've been doing myself as a post-install task. Jun 10 06:28:35 Wow now I'm all depressed that the 4.x initiative ended up being a bust. Jun 10 06:29:12 what do you mean? Jun 10 06:30:27 4.x works fine, unless to push buttons randomly, or want to upgrade to a new version after it :-) Jun 10 06:30:36 s/to push/you push/ Jun 10 06:42:02 03rwhitby 07 * r1.3469 10openembedded/ (5 files in 3 dirs): Unslung 4.21-alpha: updated README and NOTES, and fixed /mnt in maintenance mode. Jun 10 06:46:09 it looks like the binutils 2.16 openslug-image is actually going to complete Jun 10 06:48:32 NOTE: build 200506100542: completed Jun 10 06:48:51 I'll flash it to my unslung slug in a little while Jun 10 06:55:06 Hmm - maintenance mode upgrade still doesn't have enough ram Jun 10 06:55:22 ah - /dev and /var don't need to be ramdisks in maintenance mode .... Jun 10 07:10:52 and /initrd is stopping it from working Jun 10 07:16:19 hmm... What app can I use to talk to the slug over serial in windows Jun 10 07:16:22 ? Jun 10 07:16:45 HyperTerm? Jun 10 07:17:07 aha Jun 10 07:17:21 of course, hyperterm isn't installed... Bah :P Jun 10 07:18:39 ok, so it looks like we need to solve the initrd unmount before we can get maintenance mode working Jun 10 07:24:00 03rwhitby 07 * r1.3471 10openembedded/packages/nslu2-binary-only/unslung-rootfs/create-ramdisks.patch: Don't create /dev and /var ramdisks in Unslung maintenance mode Jun 10 07:41:10 Hmm - /initrd has nothing to do with it. Jun 10 07:41:21 web upgrade just doesn't want to work in 4.x Jun 10 07:47:21 ok, that's it from me tonight. unless anyone else has any ideas why maintenance mode web upgrade doesn't work in 4.x, then I'm all ears. Jun 10 07:48:07 if there are no other ideas, then we'll just call it a casualty of freeing up the extra 1MB of RAM from switchbox, and change it to use Upgrade Mode instead. Jun 10 07:51:58 no ideas here, unfortunately Jun 10 07:52:47 it just silently fails somewhere inside the Linksys upgrade.cgi binary Jun 10 07:53:28 there is plenty of RAM available, and the replacement upgrade.cgi script which creates and frees the ramdisk makes sure that the RAM is listed in the right category in /proc/meminfo Jun 10 07:54:29 and I even went as far as deleting /dev/console from the jffs2, and booting without a console (which then allowed me to unmount /initrd) and then killall'd lots of processes to free up even more RAM, and it still didn't work Jun 10 07:55:36 I'm going to flash the binutils 2.16 openslug Jun 10 07:55:59 openslug should be fine - I'm worried about a binutils 2.16 unslung ... Jun 10 07:56:05 <[g2]> go jacques go jacques flash that bad boy flash it Jun 10 07:56:16 (due to the linksys binaries) Jun 10 07:58:41 flashing Jun 10 07:59:01 toolchains can be very fragile Jun 10 07:59:12 tho it looks good for openslug because it all built Jun 10 07:59:54 I guess I can try an unslung too Jun 10 08:00:33 booting Jun 10 08:02:03 booted Jun 10 08:03:10 starting fresh clean unslung-image build Jun 10 08:03:39 (but with binutils 2.16) Jun 10 08:06:17 I guess while I'm waiting for unslung to build I can look at my other usb drive - I want to repartition that one too for turnuping :-) Jun 10 08:13:32 <[g2]> jacques, THX for the info and testing Jun 10 08:14:25 [g2], np :-) Jun 10 08:16:24 hmm I just thought of something - I haven't tested the armeb-native binutils 2.16 yet Jun 10 08:18:10 morning jbowler-away Jun 10 08:38:28 03rwhitby 07 * r1.3471.1.1 10openembedded/ (5 files in 4 dirs): Unslung 5.2-alpha: Removed the ipkg symlink on the jffs2 partition, and made unsling put it back again on the external disk Jun 10 08:42:24 Hello all Jun 10 08:47:43 <[g2]> Hi rodo37 Jun 10 08:48:22 So maintenance mode has to go .... Jun 10 08:48:32 <[g2]> nod. Nice post Jun 10 08:49:08 <[g2]> rwhitby, do you know what the memory footprint of thttpd is by chance ? Jun 10 08:49:17 It would be nice to know why upgrade.cgi isn't working, but it's not worth the effort IMHO. Jun 10 08:49:37 [g2]: it was around a MB I think. Jun 10 08:50:26 as soon as my flashing of 5.2-alpha completes I will be able to tell you Jun 10 08:50:38 <[g2]> ok thx Jun 10 08:54:40 444 root 2108 S /usr/sbin/thttpd -C /etc/thttpd.conf Jun 10 08:55:10 <[g2]> 2 MB ? Jun 10 08:55:15 <[g2]> 2.1 Jun 10 08:55:26 1054kb? Jun 10 08:57:07 Heh - an interesting side effect of removing /usr/bin/ipkg is that /etc is still in the path, so typing "ipkg" returns "ipkg: Permission denied" due to /etc/ipkg being a directory. Jun 10 08:57:25 unslung-image with binutils 2.16 completed Jun 10 08:57:36 I think I'll leave it like that, cause "Permission denied" is a good message for people who try to run ipkg on the jffs2. Jun 10 08:58:07 could just make it a tiny script that gives an informative message Jun 10 08:58:26 guess I'll flash unslung Jun 10 09:00:58 I'll wait for 12 hours before starting removal of maintenance mode. Just in case someone finds the answer to make it work. Jun 10 09:01:32 flashing Jun 10 09:01:41 * [g2] starts clean build of openslug with binutils 2.16... Won't be me Jun 10 09:02:21 I expect it would only be glc who could nut it out. He is a linuxrc guru now. Jun 10 09:03:12 booting Jun 10 09:04:44 beep Jun 10 09:05:35 well so far binutils 2.16 appears to be a nonissue Jun 10 09:05:52 for a change :-) Jun 10 09:06:43 so jacques it looks like we're keeping the same hours again .... Jun 10 09:07:28 rwhitby, yep seems that way Jun 10 09:07:48 I'd like to go to sleep soon but I might have a lunch appointment :-\ Jun 10 09:13:35 before i commit my changes i want to ask if i'm doing all things right, i change into my unslung directory and run cvs commit -m "my change description" Jun 10 09:16:46 as long as you have only changed that one file, then yes Jun 10 09:16:55 you can check this with "cvs update" Jun 10 09:17:10 it should remove "M" for your single modified file only Jun 10 09:17:15 s/remove/report/ Jun 10 09:17:46 alternately, you can give the single filename as the last argument to cvs commit Jun 10 09:19:12 night all Jun 10 09:19:35 `night rwhitby-asleep Jun 10 09:21:28 another benefit of removing maintenance mode is that we can fill up the rootfs on jffs2 with more stuff in the base firmware image. Jun 10 09:21:44 (like dropbear) Jun 10 09:21:55 yay, I'm for that Jun 10 09:22:40 ok, the ipkg thingo works Jun 10 09:52:18 03mr_claus * 10unslung/make/dhcp.mk: Look for dhcpd.leases, dhcpd.conf and dhcpd.pid at /opt instead of / Jun 10 13:16:38 I just finished the unslung build (from the repository sources) and the kernel fails to recognize the initrd. any experience? Jun 10 13:21:05 I investigated and the only difference between my kernel and the 4.20 kernel was the gcc version 3.4.4 instead of 3.4.3, otherwise diff-ing the two directories showed no diffs. Jun 10 15:44:39 ilg: I flashed unslung-5.2-alpha built from the repo about 8 hours ago Jun 10 15:44:49 will rebuild from scratch today to confirm Jun 10 15:48:06 right now I'm switching back to gcc-cross-3.4.3 to check if that version works Jun 10 15:49:15 I also checked the unslung-4.20-source.tar.gz archive and the highest gcc-cross is also 3.4.3 Jun 10 15:49:50 yeah, the change to 3.4.4 was recent, for unslung 5.x Jun 10 15:50:28 for corectness, the version I'm using is 4.21 (got it from repos on June 5th) Jun 10 15:50:48 ah, you'll want to update from the repo then Jun 10 15:52:20 what you say is that 4.21 was patched for 3.4.4 recently? Jun 10 15:53:41 which repo? bk or svn? Jun 10 15:53:52 bk Jun 10 15:54:11 we haven't changed unslung with regard to gcc, but there was a window when gcc was broken in OE Jun 10 15:55:05 jacques has built unslung 5.2 from scratch last night, and he was using 3.4.4 Jun 10 15:56:06 I haven't been building from scratch, but have just kicked one off now Jun 10 15:56:20 (mine takes a couple of hours) Jun 10 15:57:00 Linux version 2.4.22-xfs (unslung@unslung.org) (gcc version 3.4.4) #1 Mon Jun 6 Jun 10 15:57:08 it failed with: Jun 10 15:57:40 NetWinder Floating Point Emulator V0.97 (double precision) Jun 10 15:57:40 RAMDISK: Couldn't find valid RAM disk image starting at 0. Jun 10 15:57:40 Freeing initrd memory: 10240K Jun 10 15:57:40 FAT: bogus logical sector size 27209 Jun 10 15:57:40 Kernel panic: VFS: Unable to mount root fs on 01:00 Jun 10 15:58:05 what's the kernel command line printed earlier than that? Jun 10 15:58:13 replacing it with the kernel unpacked form the binary 4.20 worked Jun 10 15:58:42 Kernel command line: console=ttyS0,115200n8 root=/dev/ram0 initrd=0x01000000,10M Jun 10 15:58:43 mem=32M@0x00000000 Jun 10 15:59:21 I diff-ed it with the 4.20 kernel and except compiler version there were no differences Jun 10 15:59:39 bad flash? Jun 10 16:00:03 possible but unlikely Jun 10 16:01:29 bad programming rulled out, I did several programmings and all behaved the same Jun 10 16:02:13 I doubt a bad flash sector would allow the kernel to uncompress Jun 10 16:03:30 ah: the dyoung bug bit him Jun 10 16:04:04 ilg, you built the kernel at a time where dyoung's cmdline ?= change caused it to get the linksys kernel cmdline instead of the unslung kernel cmdline. Jun 10 16:04:12 yeah, sorry. Jun 10 16:04:17 rebuild the kernel from latest bk repo and you will be fine Jun 10 16:04:33 it's completely unrelated to gcc Jun 10 16:04:36 or if you wish to build what you have, change the ?= to = or bk undo it. Jun 10 16:05:50 ok, thanks, I'll give it a try Jun 10 16:06:06 so there doesn't seem to have been any revelations in the last 8 hours, so removal of maintenance mode will proceed today Jun 10 16:06:38 there's still a window of 4 hours or so before I can start on it, so any new revelations will have to be quick ones :-) Jun 10 16:07:00 Yank it. Jun 10 16:07:45 adios Maintainence mode. Jun 10 16:08:07 It would be nice to know its busted though. Jun 10 16:08:13 why its busted.... Jun 10 16:13:41 Anyone interested in taking a job in Shanghai? Jun 10 16:14:07 since I consider the web upgrade mode to be inferior anyway, I'm not all that driven to solve that problem Jun 10 16:14:21 beewoolie: does it pay a squillion dollars? Jun 10 16:14:28 Probably. Jun 10 16:14:44 It's for Intel and their xscale group. Jun 10 16:15:04 I was asked if I knew anyone willing to relocate there for a couple of years to help with the Linux stuff. Jun 10 16:15:20 I'd do it, but relo isn't really in my stars. Jun 10 16:16:36 I would take a job in Shanghai. Jun 10 16:16:56 dyoung: I'll give you a reference ... Jun 10 16:17:07 Excellent. Jun 10 16:17:34 They're looking for someone to do driver writing and BSP stuff. Jun 10 16:19:04 do they like UBE ? Jun 10 16:20:04 Somehow I dont think I'd fit into Intel Culture. Jun 10 16:20:16 Lemme think about it for a bit. Jun 10 16:20:16 Don't like bunny suits? Jun 10 16:22:25 UBE...wait... Jun 10 16:22:42 Ugly but effective... Jun 10 16:22:44 Right. Jun 10 16:22:56 I don't think there's a hurry. Jun 10 16:37:57 Maybe kergoth would like that job ;-) Jun 10 16:38:32 back later ... Jun 10 16:39:21 regarding the cmdline ?= bug, I just updated from the bk, can you point to a file where to check if the ? went away? I would like to avoid waiting several compiling hours to discover the same problem... Jun 10 16:43:39 linux/unslung-.....bb Jun 10 16:44:40 CMDLINE_ROOT = "root=/dev/mtdblock4 rootfstype=jffs2 rw init=/linuxrc mem=32M@0x00000000" Jun 10 16:49:03 yeah thats the one. Jun 10 17:12:38 There's a new way of reflashing in OpenSlug ChangeSet@1.3478. Run the command 'reflash help' to find out more. Jun 10 17:16:19 cool. Just did a pull. Building now. Jun 10 17:17:34 BTW if it goes wrong you can always recover by the standard RedBoot upgrade (if it doesn't go wrong all your config will be preserved.) Jun 10 17:22:19 nice Jun 10 17:22:53 So with this it should be possible to flash through a web interface? Jun 10 17:27:21 No: it's command line, 'reflash -i ', download the image from the build tree. Jun 10 18:35:19 rwhitby-away: What is the maint mode problem? Inability to enter? Jun 10 18:50:56 No. Fixed that. I see now. Jun 10 18:57:34 There is not enough free space to download the image. Jun 10 20:09:13 glc: even freeing up lots and lots of space still doesn't fix it Jun 10 20:09:43 did the upgrade-real.cgi trick. and killed lots of processes, and removed the /dev/console to umount initrd, and all that together still didn' work. Jun 10 21:11:14 Yeah. I can't get it to work either. Jun 10 21:12:15 it's unexplainable, and not debuggable Jun 10 21:13:00 so we get rid of it - it's too fragile. Jun 10 21:16:52 BTW, I think we should leave initrd mounted, so our scripts can reference it. Jun 10 21:17:04 since we will use upgrade mode, then we can leave it mounted Jun 10 21:19:35 Funny thing is that while testing before the release, I upgraded using the web interface both with and without the drive connected. Jun 10 21:19:51 yeah, it's very strange. Jun 10 21:20:05 it's never worked for me since we got rid of switchbox Jun 10 21:20:36 but the extra MB is worth it. we must prioritise run-time gains over once in a blue moon inconvenience Jun 10 21:22:35 I'll play with it a bit more tomorrow, but I would not feel bad just hacking it off. Jun 10 21:23:44 can anyone recommend a MTA for openslug? Jun 10 21:23:47 It must be checking for something in the image that isn't there any more. Jun 10 21:26:00 glc: I can't see anything in the strings. It's a mystery. Jun 10 21:26:33 glc: now that I have your input, I'm starting it's removal. Jun 10 21:27:19 OK. Jun 10 21:28:48 We will have to emphasize the need to verify Redboot or Sercomm access before upgrading. Jun 10 21:28:58 Like we don't now. Jun 10 21:44:19 heh, yeah :-) Jun 10 21:44:46 anyone upgrading from 4.x to 5.x will be forced to do that anyway ... Jun 10 21:45:02 hi guys Jun 10 21:45:29 hi Jun 10 21:46:40 I'm off to bed. Good night all. Jun 10 21:46:55 `night glc Jun 10 21:48:11 night glc - thanks for your input Jun 10 21:48:18 hey jacques Jun 10 22:25:22 jbowler-away: what is the devio incantation to erase /dev/mtdblock5 (or fill it with 0xff) ? Jun 10 22:34:30 fb#t- 255 Jun 10 22:35:08 Which, translated, means fill-bytes - with the value 255 Jun 10 22:36:23 Or you can use 0xff, or 0377 - it's just strtoul(value,0,) Jun 10 22:38:52 So the full command (assuming you just want to erase it) is devio "< Sorry, that should be devio ">>/dev/mtdblock5" "fb#255" Jun 10 22:40:02 (I've been writing too many things today which read from that partition ;-) **** ENDING LOGGING AT Fri Jun 10 23:59:57 2005