**** BEGIN LOGGING AT Sun Oct 23 02:59:57 2005
Oct 23 03:48:33 <CIA-8> 03rwhitby * 10debian/configs/nslu2_defconfig: Removed the obsolete defconfig from 2.6.12
Oct 23 04:09:54 <CIA-8> 03rwhitby * 10debian/configs/nslu2_defconfig-2.6.14-rc4: Added some non-functional configuration settings for things which are in the Debian kernel source (and therefore need to turned off) but are not in the OpenEmbedded kernel source.
Oct 23 04:10:20 <CIA-8> 03rwhitby * 10kernel/2.6.14/defconfig: Added some non-functional configuration settings for things which are in the Debian kernel source (and therefore need to turned off) but are not in the OpenEmbedded kernel source.
Oct 23 04:13:15 <CIA-8> 03rwhitby * 10kernel/2.6.14/defconfig: Added another missing item to be turned off.
Oct 23 04:13:41 <CIA-8> 03rwhitby * 10debian/configs/nslu2_defconfig-2.6.14-rc4: Added another missing item to be turned off.
Oct 23 04:26:41 <CIA-8> 03koen 07org.openembedded.dev * r5c2d88f0... 10/packages/gpe-mini-browser/ (gpe-mini-browser-hildon_0.17.bb gpe-mini-browser_0.17.bb): gpe-mini-browser: add normal and hildon version of 0.17
Oct 23 06:02:37 <ferrix> what? hildon?
Oct 23 06:03:03 <rwhitby> OE runs on the 770
Oct 23 06:03:17 <ferrix> I do know... that's my job
Oct 23 06:03:44 <rwhitby> sorry, I didn't understand your question then.
Oct 23 06:04:20 <ferrix> well... I thought that was one of the non-disclosable words
Oct 23 06:04:26 <ferrix> I was wrong :)
Oct 23 06:04:49 <rwhitby> "hildon" is plastered all over the SDK site
Oct 23 06:05:00 <ferrix> googled that much
Oct 23 06:06:11 <ferrix> well... I bet nothing about it will be secret as soon as it will be released in Q3 this year ;)
Oct 23 06:28:27 <dwery> rwhitby: kaweth.o: I think it has been patched in/after 2.6.12
Oct 23 06:28:42 <dwery> bttv.o, not required imho
Oct 23 06:29:07 <dwery> video-media.o, not required too
Oct 23 06:29:38 <dwery> mm.. no, the kawet bug is stil lthere it seems..
Oct 23 06:29:45 <rwhitby> there's a whole lot of video modules enabled.  I wonder how many of them are actually possible to hook up to a slug.
Oct 23 06:30:30 <rwhitby> (I know that there is at least one video solution for the slug, so some will be required.  But probably all those that are enabled)
Oct 23 06:30:45 <dwery> wasn't the media/ subdir for video4linux et al?
Oct 23 06:31:23 <rwhitby> dwery: we now have a 2.6.14-rc4 kernel which works unchanged for debian and openslug and ucslugc.
Oct 23 06:31:31 <dwery> great!
Oct 23 06:31:34 <rwhitby> (all big-endian)
Oct 23 06:31:47 <NAiL> rwhitby: btw, __bad_udelay seems to be a bug in the module.. it sleeps to long
Oct 23 06:32:05 <dwery> NAiL: yes.. 
Oct 23 06:32:13 <rwhitby> NAiL: yeah, I know.  I also know that someone posted about fixing it, but the fix never made it to the repo.
Oct 23 06:32:26 <dwery> I would simply split the delay
Oct 23 06:33:11 <dwery> rwhitby: I have some spare time. Need me to test something?
Oct 23 06:34:47 <rwhitby> dwery: I need someone with a running armel system and a serial port, to build a little-endian 2.6.14-rc4 kernel from the debian cvs repo, and copy it's modules to their rootfs, and then boot with the 2.6.14-rc4 kernel out of redboot ram
Oct 23 06:35:12 <rwhitby> to confirm that this kernel defconfig also works with debian armel.
Oct 23 06:36:36 <dwery> ok
Oct 23 06:37:40 <rwhitby> you should just be able to change the ENDIAN=b to ENDIAN=l in the makefile, and make kernel
Oct 23 06:40:57 <dwery> debian cvs checked out
Oct 23 06:41:04 <dwery> preparing plain kernel dir
Oct 23 06:41:37 <dwery> or you want me to download specific kernel sources/
Oct 23 06:41:38 <dwery> ?
Oct 23 06:42:13 <rwhitby> oh, you need a particular one, hang on
Oct 23 06:42:24 <rwhitby> http://packages.vergenet.net/testing/linux-2.6-2.6.13+2.6.14-rc4/linux-source-2.6.14_2.6.13+2.6.14-rc4-0experimental.1_all.deb
Oct 23 06:42:25 <dwery> ok
Oct 23 06:42:36 <rwhitby> put that in the downloads dir, and the makefile will unpack it for you
Oct 23 06:43:05 <dwery> ack. do I also need IAL?
Oct 23 06:43:26 <rwhitby> I don't have IAL for LE
Oct 23 06:43:36 <dwery> ok
Oct 23 06:44:57 <rwhitby> night
Oct 23 06:45:21 <dwery> bye
Oct 23 06:47:47 <dwery> rwhitby-asleep: errors. still there?
Oct 23 06:49:08 <dwery> http://rafb.net/paste/results/LGurTl93.html
Oct 23 06:51:21 <dwery> tried also with armel. Either I need a more recent make-kpkg or more command line flags..
Oct 23 06:55:51 <dwery> using --arch-arm
Oct 23 06:55:58 <dwery> --arch=arm
Oct 23 07:06:54 <CIA-8> 03rwhitby * 10debian/Makefile: Fixed the ARCH=armeb for LE builds.
Oct 23 07:08:56 <rwhitby-asleep> dwery: that should fix it
Oct 23 07:09:20 <jelle> anybody here that can check my dmesg output to see what's wrong?
Oct 23 07:10:39 <dwery> jelle: paste it on rafb.net and i'll check
Oct 23 07:10:44 <dwery> rwhitby-asleep: kernel compiled ok
Oct 23 07:10:59 <dwery> I need to byteswap it, right?
Oct 23 07:11:23 <rwhitby-asleep> are you using the latest devio which should have done it automatically ?
Oct 23 07:11:41 <dwery> i downloaded it from sf.net, october 20 release
Oct 23 07:12:14 <dwery> given that the kernel needs to be byteswapped only when tftping
Oct 23 07:12:24 <dwery> i thought I had to to it manually
Oct 23 07:13:55 <rwhitby-asleep> the "xb $,4" does the byte-swapping for you.
Oct 23 07:14:40 <dwery> ok. but then this kernel is not suitable for flashing, right?
Oct 23 07:15:57 <rwhitby-asleep> um.  I thought everything was BE byte streams now
Oct 23 07:18:37 <dwery> do I need a specific initrd?
Oct 23 07:19:01 <rwhitby-asleep> dwery: it loads the rootfs in jffs2
Oct 23 07:19:22 <dwery> which I don't have...
Oct 23 07:19:55 <rwhitby-asleep> ok, you'll need to compile in ext support and change the cmdline.
Oct 23 07:20:20 <dwery> maybe it's easier to use upslug to write the required jffs2
Oct 23 07:20:32 <rwhitby-asleep> we're thinking of always having the kernel boot to a ucslugc or ucslugcle jffs2 rootfs, and then from there you can turnup to debian disk partition
Oct 23 07:21:07 <rwhitby-asleep> cause the jffs2 partition needs to have all the modules you need in it
Oct 23 07:21:29 <dwery> ok. how can I build/download a jffs2 rootfs?
Oct 23 07:21:31 <rwhitby-asleep> and then you can just symlink /lib/modules to /initrd/lib/modules in your external rootfs and never worry about kernel modules again :-)
Oct 23 07:21:49 <rwhitby-asleep> Hmm - you'll want an LE jffs2 rootfs, right?
Oct 23 07:21:58 <dwery> I think so
Oct 23 07:22:01 <rwhitby-asleep> dunno.
Oct 23 07:22:27 <rwhitby-asleep> alternatively, you can change the cmdline to load an initrd.gz and just test it with the LE debian installer image ....
Oct 23 07:22:40 <dwery> how do you build the BE jffs2?
Oct 23 07:22:56 <rwhitby-asleep> I just use an openslug or ucslugc image built in OE
Oct 23 07:23:51 <dwery> can you create a new one using the LE flag?
Oct 23 07:24:10 <rwhitby-asleep> we're also thinking of putting the ixp modules in the unused 1 block ramdisk partition in flash (with the overflow in the last block) and having the /linuxrc load those modules before pivoting
Oct 23 07:24:46 <rwhitby-asleep> dwery: I can't but jbowler can create a ucslugcle - I don't think he's checked in the configs for ucslugcle yet
Oct 23 07:25:06 <jelle> it's there...
Oct 23 07:25:19 <dwery> or, If you can send me the BE jffs2 i'll mount and recreate
Oct 23 07:25:25 <dwery> jelle: post the link
Oct 23 07:25:38 <jelle> http://www.rafb.net/paste/results/NR2Qlq13.html
Oct 23 07:26:21 <rwhitby-asleep> the idea is that you flash the kernel, ixp modules, and a uslugc-derived jffs2, and then you run ucslugc, mount your disk, install the debian unconfigured rootfs there, and then turnup disk to it.
Oct 23 07:26:44 <jelle> it complains a little about mounting vfat, but that shouldn't stop it i hope...
Oct 23 07:26:49 <rwhitby-asleep> if you stuff up your debian installation, you just unplug the disk, reboot, and start again without having to reflash.
Oct 23 07:26:50 <dwery> jelle: what's your problem, exactly.
Oct 23 07:27:23 <jelle> It doesn't boot from sda1. won't mount sda1
Oct 23 07:27:23 <dwery> rwhitby-asleep: ack
Oct 23 07:28:03 <rwhitby-asleep> dwery: that scheme also means that the ixp modules never appear in debian or the installer image, just in the ucslugc flash image (which we have behind the click-through license)
Oct 23 07:28:19 <dwery> jelle: sda1 seems moujnted correctly
Oct 23 07:28:41 <dwery> jelle: you rootfs is on /dev/mtdblock4
Oct 23 07:30:03 <rwhitby-asleep> dwery: it also means that we never have to deal with debian kernel and modules in the repository - they're always part of the downloadable flash images for the slug.
Oct 23 07:30:09 <jelle> so there shouldn't be a problem? I thought I'd see more "sda" references in the result of the "mount" command...
Oct 23 07:30:23 <jelle> I'l do more testing, thankx.
Oct 23 07:30:56 <rwhitby-asleep> dwery: have fun.
Oct 23 07:30:59 * rwhitby-asleep goes to bed
Oct 23 07:32:03 <dwery> rwhitby-asleep: the BE jffs2 please :)
Oct 23 07:35:27 <rwhitby-asleep> rwhitby@sarge:/home/slug/ucslugc/tmp/deploy/images$ ls -l /mnt/hgfs/Desktop/ucslugc.jffs2
Oct 23 07:35:27 <rwhitby-asleep> -rwx------  1 root root 1572864 2005-10-24 00:04 /mnt/hgfs/Desktop/ucslugc.jffs2
Oct 23 07:35:27 <rwhitby-asleep> rwhitby@sarge:/home/slug/ucslugc/tmp/deploy/images$ md5sum /mnt/hgfs/Desktop/ucslugc.jffs2
Oct 23 07:35:27 <rwhitby-asleep> 9364de38b46e7ff5bc96eef8560690e5  /mnt/hgfs/Desktop/ucslugc.jffs2
Oct 23 07:35:40 <rwhitby-asleep> sending now
Oct 23 07:35:56 <dwery> thanks
Oct 23 07:36:48 <rwhitby-asleep> I think it would be easier for you to just test against the initrd-debian.gz-swap after changing the command line to load an initrd
Oct 23 07:37:19 <dwery> you're right
Oct 23 07:37:21 <rwhitby-asleep> ok, the mail is gone, and so am I :-)
Oct 23 07:37:27 <dwery> good night
Oct 23 12:39:10 <caplink811_log> ~seen inge_ba
Oct 23 12:39:14 <jbot> i haven't seen 'inge_ba', caplink811_log
Oct 23 12:39:19 <caplink811_log> ~seen inge_b
Oct 23 12:39:21 <jbot> caplink811_log: i haven't seen 'inge_b'
Oct 23 13:21:22 <NAiL> ~seen ingeba
Oct 23 13:21:26 <jbot> ingeba <n=ingeba@129.240.168.42> was last seen on IRC in channel #nslu2-linux, 59d 6h 7m 3s ago, saying: 'I'll stay and filter spam'.
Oct 23 13:22:05 <NAiL> ~seen ingeba-toandfrom
Oct 23 13:22:08 <jbot> ingeba-toandfrom <~ingeba@251.80-203-45.nextgentel.com> was last seen on IRC in channel #nslu2-linux, 76d 22h 51m 33s ago, saying: 'dyoung: ping?'.
Oct 23 13:26:21 <caplink811_log> thx NAiL 
Oct 23 13:27:25 <NAiL> np ;)
Oct 23 17:23:07 <rwhitby-web> rwhitby-treo: test
Oct 23 20:43:53 <rwhitby-treo> re
Oct 23 22:48:19 <CIA-8> 03rpedde * 10unslung/ (make/mt-daapd.mk sources/mt-daapd/itunes5.patch): Update mt-daapd to 0.2.3
Oct 24 00:51:02 <CIA-8> 03jbowler 07org.openembedded.dev * r00415d9b... 10/packages/ixp425-eth/ (ixp400-eth-1.4/ixp400-le-be.patch ixp400-eth_1.4.bb): 
Oct 24 00:51:02 <CIA-8> ixp400-eth: fix reboot (module unload) problem in LE patch for 1.4
Oct 24 00:51:02 <CIA-8>  - The LE patch removed code to initialise a semaphore, this caused
Oct 24 00:51:02 <CIA-8>  - a crash when the semaphore was used during the device close
Oct 24 00:51:02 <CIA-8>  - Also change config to use the 'only one NPE' option in the build,
Oct 24 00:51:03 <CIA-8>  - however by examination this doesn't work very well.
Oct 24 01:01:35 <CIA-8> 03jbowler 07org.openembedded.dev * r0cc93a00... 10/packages/ixp4xx/ixp4xx-csr_1.4.bb: 
Oct 24 01:01:35 <CIA-8> ixp4xx-csr: improve Intel license handling in 1.4
Oct 24 01:01:35 <CIA-8>  - version 1.4 of ixp4xx-csr now behaves like the other versions and
Oct 24 01:01:35 <CIA-8>  - requires users of the package to download the source code themselves
Oct 24 01:01:35 <CIA-8>  - from Intel. This helps ensure than no one accidentally violates the
Oct 24 01:01:36 <CIA-8>  - license agreement (by forcing all builders of the software to read it)
Oct 24 01:01:38 <CIA-8> 03jbowler 07org.openembedded.dev * r1a06c99a... 10/packages/meta/openslug-image.bb: 
Oct 24 01:01:40 <CIA-8> openslug-image: add udev to RDEPENDS within openslug-image itself
Oct 24 01:01:43 <CIA-8>  - This change effectively forces all users of openslug-image to use
Oct 24 01:01:44 <CIA-8>  - udev. Before this change the DISTRO file had to set udev to be
Oct 24 01:01:46 <CIA-8>  - installed. This ensures consistency and removes the requirement to
Oct 24 01:01:48 <CIA-8>  - maintain a fully and correctly populated static dev directory (it is
Oct 24 01:01:50 <CIA-8>  - currently broken and will probably remain so!)
Oct 24 01:01:52 <CIA-8> 03jbowler 07org.openembedded.dev * r95824471... 10/conf/ (9 files in 2 dirs): (log message trimmed)
Oct 24 01:01:55 <CIA-8> ucslugc, nslu2: formalise different build configs, move to IAL 2.0 and nslu2-kernel
Oct 24 01:01:57 <CIA-8>  - ucslugc is substantially as before, however the configuration is now
Oct 24 01:02:00 <CIA-8>  - expressed as a specific instance of nslu2-btu.conf (big-endian, thumb,
Oct 24 01:02:01 <CIA-8>  - uclibc). The other six buildable variants of ucslugc have been added
Oct 24 01:02:03 <CIA-8>  - as separate configuration files, all with names of the form 'nslu2-???'
Oct 24 01:02:05 <CIA-8>  - for big vs little, arm vs thumb and glibc vs uclibc.
Oct 24 01:02:07 <CIA-8> 03jbowler 07org.openembedded.dev * rabfd9f00... 10/packages/uclibc/uclibc-0.9.28/ (ucslugc/uClibc.distro nslu2/uClibc.distro): 
Oct 24 01:02:10 <CIA-8> uclibc: rename ucslugc config file to nslu2 in 0.9.28
Oct 24 01:02:12 <CIA-8>  - the file is actually used in all NSLU2 uclibc builds, even though
Oct 24 01:02:14 <CIA-8>  - it contains distro settings, therefore to avoid replicating it it
Oct 24 01:02:17 <CIA-8>  - is made the default for all NSLU2
Oct 24 01:08:23 <CIA-8> 03jbowler * 10kernel/2.6.14/10-mtdpart-redboot-fis-byteswap.patch: Remove compiler warning messages.
Oct 24 01:13:47 <CIA-8> 03jbowler * 10kernel/2.6.14/10-ixp4xx-gpio.patch: This patch is in 2.6.14, so no longer needed.
Oct 24 01:20:53 <CIA-8> 03jbowler 07org.openembedded.dev * r2849dc70... 10/packages/linux/ (2 files in 2 dirs): nslu2-kernel: fix redboot.c compiler warning messages in 2.6.14(all)
Oct 24 01:27:42 <rwhitby> http://embedded.hk/modules/newbb/viewtopic.php?topic_id=115&forum=7 <- unslung interest from hong kong
Oct 24 01:55:23 <rwhitby> jbowler-zzz: Lots of different configurations for SlugOS !  Nice work.
Oct 24 01:55:51 <jelle> root
Oct 24 01:56:07 * jelle ignores himself
**** ENDING LOGGING AT Mon Oct 24 02:59:57 2005