**** BEGIN LOGGING AT Fri Sep 02 02:59:56 2005 Sep 02 03:56:12 03xora 07org.openembedded.dev * r0a900d63... 10/packages/gnumeric/ (gnumeric-1.5.5/remove-docs.patch gnumeric_1.5.5.bb): gnumeric: update to latest dev build 1.5.5 Sep 02 04:00:42 03mcdmx * 10unslung/make/ipkg.mk: patch ipkg to look in /opt/etc Sep 02 04:11:17 03koen 07org.openembedded.dev * r517a459e... 10/packages/libmimedir/libmimedir_cvs.bb: libmimedir_cvs: fix staging Sep 02 04:17:52 03koen 07org.openembedded.dev * r21480fa4... 10/packages/ipkg/ipkg_0.99.153.bb: set default_preference=-1 since it has problems Sep 02 04:17:55 03koen 07org.openembedded.dev * r4bd9a63f... 10/conf/distro/preferred-gpe-versions-2.7.inc: conf/distro/preferred-gpe-versions-2.7.inc: update gpe-bluetooth to 0.44 Sep 02 04:23:28 03dyoung 07org.openembedded.dev * rb144329b... 10/packages/ipkg/files/libdir.patch: ipkg: Add the libdir.patch that keeps getting lost Sep 02 04:31:12 03rwhitby 07org.openembedded.dev * rcb9a7aa4... 10/conf/machine/nslu2.conf: Pin ipkg to 0.99.152 until it's fixed for depends Sep 02 04:37:01 03mcdmx * 10unslung/make/ds101-bootstrap.mk: replace hardcoded arch, add wget-ssl and openssl to bootstrap archive Sep 02 04:38:06 03mcdmx * 10unslung/sources/ds101-bootstrap/bootstrap.sh: install openssl and wget, create cross-feed.conf Sep 02 05:51:10 03mickeyl 07org.openembedded.dev * rc6803271... 10/packages/ (36 files in 12 dirs): Sep 02 05:51:10 bluez: update and put related .bb files into one directory Sep 02 05:51:10 - bluez-libs, bluez-utils, and bluez-hcidump are pretty tied, so I think it makes sense to put them into the same directory Sep 02 05:51:10 - update bluez-libs to 2.20 Sep 02 05:51:10 - update bluez-hcidump to 1.25 Sep 02 05:51:12 - enable default-manfid patch Sep 02 06:06:31 03dyoung 07org.openembedded.dev * rdca105f1... 10/ (3 files in 3 dirs): zd1211: patch makefile to make compatible with 2.4 kernels; add to unslung-packages Sep 02 06:26:30 03koen 07org.openembedded.dev * r445191ac... 10/packages/gpe-bluetooth/gpe-bluetooth_0.44.bb: gpe-bluetooth-0.44: add it to the repo Sep 02 06:41:15 03koen 07org.openembedded.dev * rbc003912... 10/packages/libmimedir/libmimedir_cvs.bb: libmimedir_cvs: stage some more headers, patch courtesy Logan Owen, closes #285 Sep 02 07:36:16 03hrw 07org.openembedded.dev * r057273e8... 10/packages/opie-console/ (opie-console.inc opie-console_1.2.0.bb opie-console_cvs.bb): opie-console: added 'opie-keytabs' to DEPENDS Sep 02 07:36:19 03hrw 07org.openembedded.dev * rbc5db216... 10/packages/opie-embeddedkonsole/ (3 files): opie-embeddedkonsole: added 'opie-keytabs' into DEPENDS Sep 02 07:45:29 Morning all. I'm trying to boot a little-endian kernel on my slug, and seeing a lot of crashing / hanging. Anyone have any advice? Sep 02 07:45:57 I'm using the kernel image and initrd from the original debian on slug page... Sep 02 07:47:03 hannesr: what is the toolchain that you use? Sep 02 07:47:31 hannesr: oops.. u are using direct from there.. Sep 02 07:47:41 I just grabbed the binary from http://peter.korsgaard.com/articles/debian-nslu2.php Sep 02 07:47:49 <[g2]> hannesr what do you want to do ? Sep 02 07:48:01 <[g2]> there's a debian BE port Sep 02 07:48:11 I'm playing with a newer version of the intel ethernet driver Sep 02 07:48:24 I have it running on be Sep 02 07:48:25 <[g2]> 1.5 or 2.0 ? Sep 02 07:48:27 hannesr: arh i see.. Sep 02 07:48:29 1.5 Sep 02 07:48:51 <[g2]> great! Sep 02 07:49:02 so far, so great ;) Sep 02 07:49:10 <[g2]> can we get then changes to make it avialable for Openslug and Debian BE ? Sep 02 07:49:25 I can send you a patch Sep 02 07:49:53 I have very little kernel experience, though, so it's a kind of a dumb "works for me" merge of the existing patches. Sep 02 07:49:59 <[g2]> is you're patch in addition to the OpenSlug patches ? Sep 02 07:50:02 I have not compared performance to 1.4 Sep 02 07:50:50 I took the OpenSlug patches for 1.4 and applied the "equivalent" as I understand it of the changes to 1.5 Sep 02 07:51:01 <[g2]> the beauty of open-source is that ppl can compare and build upon the work of others Sep 02 07:51:26 <[g2]> there are clearly lots of capable ppl out there and a community can really kick a** Sep 02 07:51:40 indeed - I also used a patch made for APEX by Marc Singer as a starting point Sep 02 07:51:52 without that I wouldn't have had a hope :) Sep 02 07:51:57 <[g2]> exactly Sep 02 07:52:16 <[g2]> so you build upon beewoolie's (Marc) 1.5 port Sep 02 07:52:53 yip. After that it was just a matter of chasing down some missing include files and such. Sep 02 07:52:56 <[g2]> I'm sure he'll be happy to hear someone is using that Sep 02 07:53:05 and the small matter of the ethernet driver Sep 02 07:53:20 s/ethernet driver/ixp425_eth/ Sep 02 07:53:33 <[g2]> one often never knows how close one is to some goal Sep 02 07:54:12 I was surprised how easily I got it working myself... just a matter of doing the dokeywork. Sep 02 07:54:35 And of making sure you're compiling against the right kernel configuration... but that's another story. Sep 02 07:54:38 <[g2]> well excellent work Sep 02 07:54:53 <[g2]> Are you running APEX ? Sep 02 07:55:10 no, plain old redboot Sep 02 07:55:27 <[g2]> Ok... Sep 02 07:55:42 <[g2]> any plans to run APEX ? Sep 02 07:56:09 03rpurdie 07org.openembedded.dev * r5678fb66... 10/packages/linux/linux-openzaurus_2.6.13-mm1.bb: linux-oz-2.6: Update sharpsl_pm patch due to compile errors. Sep 02 07:56:18 I don't see a real advantage for me in doing so. And I'm afraid I'll brick the slug. Sep 02 07:57:21 any uclibc "experts" lurking around? Sep 02 07:57:26 the byte-swapping trick seems to work ok for booting le images Sep 02 07:57:27 <[g2]> so are you playing with the LE for learning and growth or to use the normal ARM Debian which is LE ? Sep 02 07:57:41 <[g2]> mcdmx, try /#uclibc :) Sep 02 07:57:49 tnx Sep 02 07:58:11 a bit of both - I wanted to learn a bit about driver code, and having debian packages seemed like a worthwhile thing Sep 02 07:58:32 of course with the armeb packages it looks like there are easier ways to get that now Sep 02 07:58:50 <[g2]> right Sep 02 07:59:06 <[g2]> there's over 1800+ now and growing Sep 02 07:59:14 nice Sep 02 07:59:37 <[g2]> I'm running the Debian BE stuff and have built a couple dozen packages already Sep 02 07:59:38 and they're all building natively on a slug? Sep 02 07:59:42 <[g2]> yep Sep 02 07:59:51 hannesr: what problem are you having with the binary kernel image from my homepage? Sep 02 08:00:30 oopses Sep 02 08:00:34 [g2]: yes, but 1800 is still far from the 16K packages on arm LE ;) Sep 02 08:00:54 hannesr: without messing with kernel drivers/modules? Sep 02 08:00:54 <[g2]> Jacmet depends on what you want to run :) Sep 02 08:01:04 [g2]: true Sep 02 08:01:09 yes, just the plain old lart image Sep 02 08:01:26 the lart image seems to reliably crash while trying to detect network adapters Sep 02 08:01:28 hannesr: I have been running that kernel ever since I made the page without problems Sep 02 08:01:44 which I figured might be because I don't have a usb ethernet adapter connected Sep 02 08:01:45 hannesr: is your slug overclocked? Sep 02 08:01:58 Jacmet: yes. I was thinking about that. Sep 02 08:02:00 hannesr: ahh, perhaps - I never tried it without Sep 02 08:02:39 I stuck a "/bin/sh" into the debian install scripts in the initrd to give me a shell before that point Sep 02 08:02:42 <[g2]> hannesr how big is the additional patch ? Sep 02 08:03:22 g2: I'll check... it's the same order of magnitude as the 1.4 patches Sep 02 08:03:35 no major additions by me Sep 02 08:04:08 <[g2]> are you running in interrupt or polled mode ? Sep 02 08:04:13 with the modified initrd it usually boots as far as the shell, though sometimes hangs at various point during boot Sep 02 08:04:37 running commands from the shell produces "invalid instruction" or "Unable to handle kernel paging request" oopses pretty quickly Sep 02 08:04:50 g2: interrupt mode. I haven't tried polled mode. Sep 02 08:05:01 <[g2]> hannesr what part of the world to you hail from ? Sep 02 08:05:26 I'm in germany Sep 02 08:05:35 <[g2]> ok.. cool Sep 02 08:05:49 <[g2]> I'm guessing you don't do JTAG debugging yet Sep 02 08:06:33 no. If you know anyone with JTAG capability in Berlin who can rescue my slug if I brick it, I might be persuaded to try APEX :) Sep 02 08:07:25 have there been LE stability problems with overclocked slugs before? Sep 02 08:07:47 or is there something I could have goofed up to cause these crashes? Sep 02 08:10:19 <[g2]> hannesr are you older than college age ? Sep 02 08:10:40 hannesr: you are the first one I hear anything about stability problems from Sep 02 08:11:12 g2: slightly... I have ~3 years professional experience Sep 02 08:11:27 <[g2]> excellent Sep 02 08:11:37 <[g2]> I'm just concerned about < 18 Sep 02 08:11:42 <[g2]> or 21 even Sep 02 08:12:01 * hannesr didn't know there was alcohol in #nslu2-linux Sep 02 08:12:02 <[g2]> concerned becuase I'm building hw Sep 02 08:12:24 <[g2]> and I'm not really interested in sending stuff on minors Sep 02 08:12:51 Here's a log of the slug booting and crashing http://cron.pastebin.com/352855 , maybe someone can spot something unusual happening Sep 02 08:13:02 <[g2]> not because a 16 year-old couldn't hack it Sep 02 08:13:27 <[g2]> it's just no worth hassling with Sep 02 08:13:45 <[g2]> s/no/not Sep 02 08:14:14 g2: you're building a slug-derived board, right? Sep 02 08:14:58 <[g2]> proto's are getting stuffed at the CM today :) Sep 02 08:15:04 <[g2]> nod. Sep 02 08:15:18 <[g2]> I'll launch the web site in a day or two Sep 02 08:15:32 <[g2]> it's all ready to go except for a little content Sep 02 08:16:06 03rpurdie 07org.openembedded.dev * rca81ad46... 10/packages/maemo/nokia770-init_1.0.bb: nokia770-init: change to hotplug instead of hotplug-dbus and only need a runtime dependency Sep 02 08:16:27 <[g2]> it's not a slug-derived board but an upscale board Sep 02 08:17:17 <[g2]> the relevent point that it comes with external serial and a JTAG header inside Sep 02 08:17:22 <[g2]> CF slot too Sep 02 08:17:44 <[g2]> I'll initially most likely ship the Debian BE port with it Sep 02 08:18:24 <[g2]> and with the 1.5 LE drivers and a tweaked Reboot the Debian LE could be avialable too Sep 02 08:18:45 <[g2]> I'm starting with BE because that makes more sense to me Sep 02 08:18:54 <[g2]> 1) it's been running for the last year+ Sep 02 08:19:13 <[g2]> 2) it's much more obscure and thus less of a target Sep 02 08:19:58 target? you mean in terms of hw competition? Sep 02 08:20:08 <[g2]> so between AIDE and JTAG all the bits on the box are verifiable Sep 02 08:20:23 <[g2]> no I mean it probability of exploit from the network Sep 02 08:21:00 <[g2]> the "Loft" is planned to be sold at $249 US for everything but a CF card/disk Sep 02 08:21:39 <[g2]> it roughly competes with the soekris, wrap, and mini-itx space Sep 02 08:22:34 will it have an ide controller, or is the CF slot intended as the main storage? Sep 02 08:22:35 <[g2]> I think the arch difference between the soekris and mini-itx boards are important as they are both x86 Sep 02 08:23:23 <[g2]> it's got CF, plus 2 USB 2.0, plus miniPCI and dual ethers Sep 02 08:23:31 <[g2]> right now 64MB Sep 02 08:23:49 <[g2]> one of my proto's will have 128MB Sep 02 08:24:26 <[g2]> debian actually recommends 128MB as the minimum for a server Sep 02 08:26:42 <[g2]> IMHO we are in a cornucopia of possibilities Sep 02 08:26:59 <[g2]> it really comes down to choice/preference Sep 02 08:27:45 <[g2]> and it signifies a maturation of the technology Sep 02 08:36:00 <[g2]> hannesr, it looks like everything isn't coming up properly Sep 02 08:36:14 <[g2]> you've got serial right ? Sep 02 08:36:39 <[g2]> so I'd disable the network drivers and boot without them first Sep 02 08:37:49 g2: you mean build a kernel image without any network support? Sep 02 08:38:10 (I don't see any network modules being loaded at the point where the crash happens) Sep 02 08:38:11 <[g2]> I mean don't load it from the rootfs Sep 02 08:38:36 <[g2]> maybe it's not and you are still seeing issues with the init Sep 02 08:39:36 I'll go through the init scripts and make sure they're not trying to load anything Sep 02 08:41:32 <[g2]> actually, just put a set -x in the main rc.d default script Sep 02 08:41:40 <[g2]> you can see what's happening Sep 02 08:42:19 <[g2]> and who the "bad guy" (errant sw) that attempts to write/access the bogus location Sep 02 08:49:04 added set -x to /sbin/debian-installer-startup. here's the log: http://cron.pastebin.com/352899 Sep 02 08:49:25 this time no crash message, it just hangs after I try the "ls" Sep 02 08:52:52 booted it again, and got a crash after a few commands: http://cron.pastebin.com/352901 Sep 02 08:57:08 <[g2]> right there's a NULL pointer getting used somewhere Sep 02 08:58:29 this stuff doesn't seem to be repeatable - sometimes it's an invalid instruction, sometimes it's a null pointer, and sometimes an invalid address. Sep 02 08:59:00 nor does it seem to be a specific command that triggers the crashes Sep 02 08:59:11 <[g2]> it's seems like a memory corruption/bad pointer issue to me at first glance Sep 02 09:00:52 <[g2]> hannesr I wonder how your setup is different form Jacmet's Sep 02 09:01:15 <[g2]> I'd think the slugs would all pretty much act the same with just flash and serial Sep 02 09:01:20 <[g2]> except for timing differences Sep 02 09:01:37 <[g2]> I wouldn't think you've got a race condition Sep 02 09:02:08 I don't have a usb ethernet adapter connected, that's all I can think of Sep 02 09:02:31 I've tried with and without a memory stick connected, and that seems to make no difference Sep 02 09:03:06 Jacmet: is your slug overclocked / has this kernel been confirmed to work on an overclocked slug? Sep 02 09:19:05 hannesr: yes, mine is overclocked, and there's no reason why it shouldn't work (E.G. it is de-underclocking) Sep 02 09:28:22 03mcdmx * 10unslung/make/ds101-bootstrap.mk: unpack required files to a subdirectory Sep 02 09:51:04 03justinp 07org.openembedded.dev * r9aa9c46d... 10/packages/ipkg/ipkg_0.99.153.bb: ipkg: fix bb syntax Sep 02 10:56:18 03justinp 07org.openembedded.dev * red6f5e4d... 10/packages/ (4 files in 3 dirs): gpe-dm, gpe-session-scripts, entrance: x11-common => xserver-common Sep 02 10:56:21 03justinp 07org.openembedded.dev * r6b8389ae... 10/packages/musicpd/ (mpd/save-volume-state.patch mpd_0.11.5.bb): mpd: Add a patch to save/restore volume state in the state_file Sep 02 10:56:23 03justinp 07org.openembedded.dev * r44e79e78... 10/packages/ (4 files in 3 dirs): gpe-dm, gpe-session-scripts, entrance: bump PR Sep 02 11:01:27 03orepvik * 10unslung/sources/crosstool/powerpc-603e.dat: Remove CRLF encoding. Sep 02 11:11:18 03justinp 07org.openembedded.dev * r9ae1962d... 10/packages/synergy/synergy_1.2.4.bb: Add synergy Sep 02 11:11:20 03justinp 07org.openembedded.dev * r43cbbf0c... 10/packages/slutils/ (3 files in 2 dirs): slutils: add slfb, a command-line utility to turn on/off the FrameBuffer for 2.4 Zaurus kernels Sep 02 11:20:04 03orepvik * 10unslung/Makefile: Split DS101_PACKAGES into common, ds101-specific and ds101g-specific Sep 02 11:30:07 03orepvik * 10unslung/make/openssh.mk: Fix missing sources. Sep 02 11:40:21 03mcdmx * 10unslung/sources/ds101-bootstrap/postinst: Make sure that there is a /bin/killall - optware start scripts rely on it Sep 02 11:58:30 Is there a specific reason optware bash is 2.x instead of 3.x? Sep 02 12:04:56 03jbowler * 10upslug/upslug.c: Make the verify target work. Sep 02 12:11:49 03jbowler * 10upslug2/ (15 files): Initial version of upslug2 Sep 02 14:01:11 03koen 07org.openembedded.dev * r0aa9fa51... 10/packages/gpe-session-scripts/gpe-session-scripts_0.66.bb: Add gpe-sessionscripts 0.66 Sep 02 14:11:10 03koen 07org.openembedded.dev * rd83dde07... 10/conf/distro/preferred-gpe-versions-2.7.inc: conf/distro/preferred-gpe-versions-2.7.inc: use gpe-sessionscripts 0.66 Sep 02 16:46:23 03justinp 07org.openembedded.dev * r6a06d3e1... 10/packages/e17/ (4 files in 3 dirs): entrance: set DISPLAY before calling Xinit Sep 02 16:46:25 03justinp 07org.openembedded.dev * rdd4e3c7a... 10/packages/e17/ (3 files in 3 dirs): entrance: run entrance from entrance_wrapper instead of from Xinit to fix login issues Sep 02 16:56:10 03justinp 07org.openembedded.dev * rfed15f0f... 10/packages/e17/entrance_20050810.bb: entrance: remove unused directory from SRC_URI Sep 02 17:11:07 03justinp 07org.openembedded.dev * r6b98c67d... 10/packages/e17/entrance_20050810.bb: entrance: remove more Xinit.d references Sep 02 18:55:25 03jbowler * 10upslug2/ (5 files): Sep 02 18:55:25 Add support for libpcap in place of the Linux specific stuff. Use: Sep 02 18:55:25 configure --with-libpcap to turn this on. Behaviour of the libpcap version Sep 02 18:55:25 under Linux is sub-optimal, the code has not yet been verified on BSD. Sep 02 19:49:57 03rwhitby * 10unslung/Makefile: Demoted qemu and qemu-libc-i386 cause they break the autobuild. Sep 02 19:56:06 03rwhitby 07org.openembedded.dev * r8a85b8d7... 10/conf/distro/ (openslug-packages.conf ucslugc-packages.conf): openslug-packages.conf,ucslugc-packages.conf: Updated location for bluez. Sep 02 20:01:03 03rwhitby 07org.nslu2-linux.dev * r67a6f708... 10/Makefile: Enabled ds101 and ds101g+ autobuilds. Sep 02 23:05:41 03jbowler * 10upslug2/configure.ac: Correct the version number to '2' Sep 03 01:38:57 03mcdmx * 10unslung/make/libnsl.mk: handle the ds101g OPTWARE_TARGET **** ENDING LOGGING AT Sat Sep 03 02:59:57 2005