**** BEGIN LOGGING AT Sat Jan 26 02:59:57 2008 Jan 26 03:06:21 mwester: last but not least, procedure on 4.8 works as advertised... :)... I'll spend some more time on 4.9 and see what I can come up with , it has something to do with the enet module/function not being active when it goes to mount the net volume Jan 26 03:15:01 aadamson: note that a big change from 4.8 to 4.9 is in the ethernet driver. Jan 26 03:15:17 4.8 used the old open source driver, 4.9 kernel uses the new open source driver. Jan 26 03:15:27 completely different animals, with no common heritage. Jan 26 03:58:38 03rwhitby * r994 10kernel/trunk/patches/2.6.24/KERNEL: Moved to 2.6.24 released Jan 26 04:02:11 Well, now building images should be easier at least Jan 26 04:02:21 I've rebuilt my debian desktop Jan 26 04:11:28 03bzhou * r7656 10optware/trunk/make/svn.mk: svn: 1.4.5 -> 1.4.6 Jan 26 04:44:08 rwhitby: Thanks for the info, I may play around and see if I can get the nfs (root swivel) to work on 4.9... gonna take figuring out when/how the enet driver goes active tho Jan 26 05:03:23 03rwhitby * r995 10kernel/trunk/patches/2.6.25/ (7 files): Added upstreamed patches from git Jan 26 05:39:18 03rwhitby * r996 10kernel/trunk/patches/2.6.25/ (8 files): Checked in patches submitted upstream Jan 26 05:40:30 03rwhitby * r997 10kernel/trunk/patches/2.6.25/KERNEL: Start with 2.6.24 kernel Jan 26 09:40:39 03oleo * r7657 10optware/trunk/make/transmission.mk: transmission: 4684->4839 Jan 26 17:28:48 03bzhou * r7658 10optware/trunk/make/transmission.mk: transmission: --disable-gtk --disable-wx instead of without-* Jan 26 18:38:26 rwhitby: Hi. I built a 2.6.23.14 kernel with dma_bounce disabled. My understanding is that it is only required if the DMA controller can only transfer data within a window smaller than the available memory size. The ixp42x processor does not have this limitation as far as I can tell. It kernel works fine but I get similar dma buffer allocation errors when swapping gets going. I mention this only to suggest that dma_bounce is not th Jan 26 18:38:26 e source of the problem. Jan 26 18:41:39 The other thing is that it takes time for the system to fail implying that the failure is due to either having too many things going on at once or somebody is not releasing the buffers after using them. Jan 26 19:04:00 Right now I have 128M slug memtestering 110M while running 'make update' in the master Makefile. It is using monotone-0.26 and is swapping continuously with no errors on the console. Jan 26 20:22:10 I am having a lot of Jan 26 15:20:32 kernel: hub 1-2:1.0: Cannot enable port 1. Maybe the USB cable is bad? Jan 26 20:22:16 but all my disk are working fine. Jan 26 20:22:24 anyone may know why? Jan 26 20:47:53 Anyone using the ethernet driver made by Krzysztof on the nslu? Jan 26 20:51:25 Yep, many folks have been using it. Any 2.6.23 (and beyond) kernel uses it. Jan 26 20:53:04 mwester: Well, i've already compiled and booted 2.6.23.14, but i can't bring the interface up: "SIOCSIFFLAGS: No such file or directory" Do you have an idea what's wrong? Jan 26 20:55:28 Check dmesg to see if there are any reasons for it to go wrong. Where did you get your kernel and patches? Jan 26 20:57:23 There are no obvious error messages for it. I've used the http://svn.nslu2-linux.org/svnroot/kernel repository for getting the kernel/patches. Jan 26 20:57:34 Is there any sign of it? Jan 26 20:59:15 Well, i can see the interface by using ifconfig -a, it has the correct mac adress, i can configure the ip and netmask but i can't bring it up because of that error. The only message in dmesg referring to it is "eth0: MII PHY 1 on NPE-B" Jan 26 21:01:02 ah, and: "NSLU2: Using MAC address $mine:) for port 0" Jan 26 21:01:13 What's your rootfs? Jan 26 21:02:52 a small gentoo from devroot.de Jan 26 21:06:21 Did you do whatever you have to do with gentoo in order to load the firmware to the interface? Jan 26 21:11:01 Oh! I didn't knew that this was necessary for this driver. I didn't found a /dev/ixp4xx_ucode or IxNpe for it. Jan 26 21:15:01 yes, the microcode is proprietary Jan 26 21:15:17 it uses the standard userspace firmware loader Jan 26 21:15:32 put a file called NPE-B in /lib/firmware/ Jan 26 21:17:39 http://ipkg.nslu2-linux.org/feeds/angstrom/unstable/armv5teb/ixp4xx-npe_2.4-r1_armv5teb.ipk has the firmware Jan 26 21:29:09 Great! Thanks:} Jan 26 21:53:41 03rwhitby * r998 10kernel/trunk/patches/2.6.25/ (4 files in 2 dirs): Forward-porting from 2.6.24 Jan 26 21:58:11 03rwhitby * r999 10kernel/trunk/patches/2.6.25/ (16 files): Forward-porting from 2.6.24 Jan 26 21:59:43 03rwhitby * r1000 10kernel/trunk/patches/2.6.25/ (4 files): Removed erroneous additions Jan 26 22:01:16 03rwhitby * r1001 10kernel/trunk/patches/2.6.25/ (3 files): Removed erroneous additions Jan 26 22:02:12 03rwhitby * r1002 10kernel/trunk/patches/2.6.25/series: Updated series Jan 26 22:09:46 w00t, found ya! Jan 26 22:10:10 :D Jan 26 22:10:26 Got a little problem. I put my NSLU2 in upgrade-mode, yet my laptop still does not see it with upslug2. Any hints as to what may be wrong? Jan 26 22:10:54 are you directly connecting the slug to the laptop? Jan 26 22:11:11 Strange thing is, Windows did see it (when i upgraded it previously - 6 months ago), but Linux (various distro's) doesn't. Jan 26 22:11:47 ka6sox-laptop: No, the slug is connected to the router, and i am wireless. The ip's used are in the 192.168.2.* range Jan 26 22:12:08 iarwain1, the IP isn't important Jan 26 22:12:17 try hardwiring your laptop. Jan 26 22:12:31 to the switch. Jan 26 22:12:32 Connecting it directly to the laptop? Jan 26 22:12:43 ah, so laptop=>router, router=>slug ? Jan 26 22:12:45 no, connect to a hardwire. Jan 26 22:12:48 yep Jan 26 22:12:51 kk Jan 26 22:17:17 ka6sox-laptop: i see it now, thanks! Jan 26 22:19:43 np...p00f Jan 26 22:27:15 Another quick question regarding to putting linux on the slug. uclibc or glibc ? Jan 26 22:27:31 your choice Jan 26 22:27:32 your choice Jan 26 22:27:37 >.< Jan 26 22:27:38 snap Jan 26 22:27:57 What's the main difference? (i don't know much about libc stuff :( ) Jan 26 22:28:00 iarwain1, i use both Jan 26 22:28:06 ^^ Jan 26 22:28:23 iarwain1: http://article.gmane.org/gmane.comp.misc.nslu2.linux/20610 Jan 26 22:28:33 Well, i just plan on using the slug for MPD (music), NFS (perhaps samba) and very perhaps a webserver. Jan 26 22:28:41 that was what I was looking for. Jan 26 22:29:38 rwhitby: Well, but i'm putting Gentoo on it (sorry i didn't mention before), but i can choose between uclibc and glibc. Jan 26 22:29:54 ask Joshin Jan 26 22:30:40 joshin: Could you tell me what the main difference would be between uclibc and glibc on Gentoo for the slug? Jan 26 22:30:58 I only use glibc. uclibc just drives me nuts Jan 26 22:31:07 Your sanity. Jan 26 22:31:16 Cheers :) Jan 26 22:31:17 Uclibc stuff takes up a bit less space Jan 26 22:31:38 Well, i have a 2Gb stick for it, guess that should be enough. Jan 26 22:31:47 It probably also uses less memory. But some things are a pain in the @** to use with it. Jan 26 22:32:01 "some things" ? Jan 26 22:32:04 it'll be good until you try to stage something during a ebuild Jan 26 22:32:04 There's an embedded Gentoo channel here on Freenode. Jan 26 22:32:17 Ask them the advantages of uclibc. Jan 26 22:32:31 joshin: Yeah, i went there, but seemed quite dead :( Jan 26 22:32:38 iarwain1: patience is a virtue Jan 26 22:33:05 i know, but sometimes it's hard to be patient (no excuse though) Jan 26 22:33:08 Its been over a year since I last tried uclibc but I couldn't even build irssi Jan 26 22:33:21 Hmm, then that answers my question :) Jan 26 22:33:23 Cheers Jan 26 22:34:32 I wouldn't attempt to emerge much on a 2GB stick. Jan 26 22:35:52 Yup. I mount all the portage files via NFS. Jan 26 22:37:12 how do you handle swap? Jan 26 22:38:17 Normally, to a microdrive or a spare 48MB SD card I don't care about. Jan 26 22:39:34 ah...I was wondering because swapping over NFS is problematic. Jan 26 22:39:36 The slug I'm on right now (running irssi) has been using the same SD card for about 2 years without a problem. Jan 26 22:41:17 Which means that it will fail in about 5 minutes Jan 26 22:44:37 joshin, but are you swapping a lot? Jan 26 22:45:25 josh@slug ~ $ uptime && free && ps ax |wc Jan 26 22:45:25 22:45:17 up 55 days, 22:14, 1 user, load average: 0.15, 0.09, 0.03 Jan 26 22:45:25 total used free shared buffers cached Jan 26 22:45:25 Mem: 30352 22884 7468 0 252 11028 Jan 26 22:45:25 -/+ buffers/cache: 11604 18748 Jan 26 22:45:27 Swap: 29864 1508 28356 Jan 26 22:45:29 32 172 1261 Jan 26 22:46:30 not too bad..the real question isn't how much but how often. Jan 26 22:47:18 Not often. The machine doesn't do all that much. The most intensive job on it is denyhosts. Jan 26 22:47:43 i run spamassassin. that swaps :) Jan 26 22:48:02 ouch, I'll bet it does. Jan 26 22:48:02 I'd imagine so. Jan 26 22:49:16 03rwhitby * r1003 10kernel/trunk/patches/2.6.25/ixp4xx-nslu2-nas100d-ethernet-support.patch: Updated ethernet patch from my git repo Jan 26 22:55:14 03rwhitby * r1004 10kernel/trunk/patches/2.6.25/defconfig: Updated vanilla defconfig Jan 26 22:55:53 full mainline support for the nslu2 nudges closer .... Jan 26 23:04:25 dumfrac: http://article.gmane.org/gmane.comp.misc.nslu2.linux/20969 Jan 26 23:08:30 That is scary that it would end up there so fast. Like 10 minutes tops. Jan 26 23:09:13 I hope its not so bad its not even wrong... Jan 26 23:11:37 Oh... an archive. Learning things every day. Jan 26 23:20:57 03rwhitby * r1005 10kernel/trunk/patches/2.6.25/ixp4xx-nslu2-nas100d-ethernet-support.patch: Cleaned up the patch (now passes checkpatch) Jan 26 23:21:33 sdm485: it's been passed to blogic, who is working this issue in his real job Jan 26 23:22:47 I hope it helps. Thanks for letting me know. Jan 26 23:24:37 I wonder if it is showing up now because there are so few systems that have usb_storage for swap and rootfs but are fast enought to get into trouble. Jan 26 23:25:06 I suppose the test would be to slow the fatslug down to 133MHz and see if it is still a problem Jan 27 00:24:45 rwhitby: Your makefile for crosscompiling was broken by the new releases of Debian. I'm not done fixing it, but so far, here are the changes. Replace all instance of 'linux-kernel-headers' with 'linux-libc-dev'. Replace the command "wget ${DEBIAN_POOL}/main/l/linux-kernel-headers/linux-kernel-headers_${HEADERS_VERSION}_${ARCH}.deb" with "wget ${DEBIAN_POOL}/main/l/linux-2.6/linux-libc-dev_${HEAD Jan 27 00:24:47 ERS_VERSION}_${ARCH}.deb". Change "GCC41_SOURCE_VERSION := $(shell dpkg -s gcc-4.1 | sed -n 's/Source:.*(\(.*\))/\1/p')" to "GCC41_SOURCE_VERSION := $(shell dpkg -s gcc-4.1 | sed -n 's/Version:[ ]*\(.*\)/\1/p')" Jan 27 00:25:51 Annirak, do you have developer access to our trac? Jan 27 00:26:02 No. Jan 27 00:26:20 if not email this to our mailing list? Jan 27 00:26:28 groups.yahoo.com/nslu2-linux Jan 27 00:26:39 Sure. I'll finish fixing it first tho Jan 27 00:26:45 thanks Jan 27 00:27:02 Right now, I'm stopped at: Jan 27 00:27:06 dpkg-buildpackage: warning: Build dependencies/conflicts unsatisfied; aborting. Jan 27 00:27:08 dpkg-buildpackage: warning: (Use -d flag to override.) Jan 27 00:27:09 make: *** [gcc-4.1-arm-linux-gnu_4.1.2-18_amd64.deb] Error 3 Jan 27 00:27:25 hmmm Jan 27 00:27:30 whups, missed a line there Jan 27 00:27:31 dpkg-checkbuilddeps: Unmet build dependencies: libc6-dev-arm-cross (>= 2.5) Jan 27 00:27:53 this is for building the cross-compiling toolchain? Jan 27 00:28:25 It's the makefile posted here: http://svn.nslu2-linux.org/svnroot/debian/trunk/toolchain/Makefile Jan 27 00:29:59 ah, okay email your patches to the list and I'm pretty sure that dumfrac will look at them. Jan 27 00:30:20 ok, this is weird Jan 27 00:30:33 I reran make all after that error, and it's running through this time Jan 27 00:30:40 oh, never mind Jan 27 00:30:46 it hung up again Jan 27 00:31:08 okay when you get it sorted post it up and it'll get looked at. Jan 27 00:34:30 Annirak: feel free to send me the patch directly Jan 27 00:34:40 'kay Jan 27 00:34:51 any thoughts on the problem above? Jan 27 00:39:29 Ah, I think it was caused by me reordering the packages to install Jan 27 00:39:57 mwester: any change you are around and I can pick your brain abit around the nfs boot under 4.9? The 4.8 stuff doesn't work Jan 27 00:40:31 rwhitby: should I be concerned about this? Jan 27 00:40:35 debian/rules.unpack:189: warning: overriding commands for target `stamps/01-unpack-stamp-' Jan 27 00:40:37 debian/rules.unpack:165: warning: ignoring old commands for target `stamps/01-unpack-stamp-' Jan 27 00:53:57 Ok, I found the problem. The 'all' build target was missing 'libgcc1-${ARCH}-cross_${GCC41_VERSION}_all.deb' and the 'libgcc1-${ARCH}-cross_${GCC41_VERSION}_all.deb' build target isn't defined. Jan 27 02:04:16 03bzhou * r7659 10optware/trunk/ (Makefile make/sg3-utils.mk): sg3-utils: added & promoted **** ENDING LOGGING AT Sun Jan 27 02:59:57 2008