**** BEGIN LOGGING AT Sun Feb 25 02:59:58 2007 Feb 25 03:04:00 03bzhou * r5621 10optware/trunk/make/glib.mk: glib: use standalone libiconv if not part of libc Feb 25 03:12:45 drif: Is there a way to integrate that into the unofficial Debian NSLU2 installers? Feb 25 03:29:35 03nail 07ts101_support * r749 10kernel/patches/2.6.15/ (24 files): add working 2.6.15 patchset Feb 25 04:06:37 03nail 07ts101_support * r750 10kernel/patches/2.6.16/ (22 files): Add working kernel set for 2.6.16 Feb 25 04:12:17 03nail 07ts101_support * r751 10kernel/patches/2.6.16/defconfig: add defconfig Feb 25 04:13:06 ? Feb 25 04:13:47 Working on a more recent kernel for the QNAP TS-101 Feb 25 04:14:01 nothing to do with the nslu2 Feb 25 04:15:37 03nail 07ts101_support * r752 10kernel/patches/2.6.17/ (24 files): Add working 2.6.17 patchset Feb 25 04:21:47 03nail 07ts101_support * r753 10kernel/patches/ (5 files in 5 dirs): Add notes for each kernel revision Feb 25 04:22:49 03nail 07ts101_support * r754 10kernel/patches/2.6.20/: Remove 2.6.20 patchset, since it doesn't work anyway Feb 25 04:26:03 03nail 07ts101_support * r755 10kernel/patches/2.6.18/ (24 files): Add working 2.6.18 patchset (Weird irq-problem aside) Feb 25 04:27:35 * mwester wonders if NAiL ever sleeps... Feb 25 04:27:40 yes, I do Feb 25 04:27:49 once in a while Feb 25 04:28:01 but now I'm on a roll :-P Feb 25 04:28:03 You're up to .18 now. Feb 25 04:28:26 03nail 07ts101_support * r756 10kernel/patches/2.6.18/README: Add note about usb IRQ-problem Feb 25 04:28:27 yeah Feb 25 04:28:31 working on .19 Feb 25 04:28:56 I'll go to bed when I get 2.6.20.1 working :-P Feb 25 04:29:08 or, atleast *booting* Feb 25 04:29:28 the last two kernels have some irq-problems that I don't know how to solve yet. Feb 25 04:31:35 Symptoms? Feb 25 04:33:19 It gives a dump during boot Feb 25 04:33:30 I don't have a filesystem I can boot to, so I can't diagnose further Feb 25 04:33:47 IRQ handler type mismatch for IRQ 17 Feb 25 04:33:47 Call Trace: Feb 25 04:33:47 [C04B7D90] [C000BCBC] show_stack+0x48/0x194 (unreliable) Feb 25 04:33:47 [C04B7DC0] [C003810C] setup_irq+0x13c/0x214 Feb 25 04:33:47 [C04B7DE0] [C00383A0] request_irq+0xac/0xdc Feb 25 04:33:50 etc. Feb 25 04:34:04 So, really bad IRQ problem then. Feb 25 04:34:15 Happens the second the usb-drivers load Feb 25 04:34:38 It sounds like it's a matter of changing the irq handler type Feb 25 04:34:45 ;) Feb 25 04:35:04 What happens if no USB is in the kernel - is the problem isolated to USB? Feb 25 04:35:31 no idea, I'll have to debug that when I'm a bit more awake Feb 25 04:35:47 Ok Feb 25 04:35:50 right now I'm just trying to get 2.6.20.1 to boot (And any previous kernels) Feb 25 04:37:47 * mwester_ must have offended the network gods! Feb 25 04:39:02 heh Feb 25 04:39:06 03nail 07ts101_support * r757 10kernel/patches/2.6.19/ (25 files): Add working patchset for 2.6.19 (Still with IRQ17 oddness Feb 25 04:39:58 ) Feb 25 04:40:18 * NAiL hastily adds missing ")" to the commit message :-P Feb 25 04:41:36 ok, time to cross fingers Feb 25 04:41:39 Compiling 2.6.20.1 now Feb 25 04:42:53 03nail 07ts101_support * r758 10kernel/patches/2.6.18/README: Update notes Feb 25 04:45:00 morning jacques Feb 25 04:47:28 Linux version 2.6.20.1 (repvik@apollo) (gcc version 3.4.6) #1 Sun Feb 25 06:19:44 CET 2007 Feb 25 04:47:28 Motorola SPS Sandpoint Test Platform Feb 25 04:47:30 Port by MontaVista Software, Inc. (source@mvista.com) Feb 25 04:47:31 ooh yeah ;-) Feb 25 04:48:23 03nail 07ts101_support * r759 10kernel/patches/2.6.20.1/ (25 files): Add working patchset for 2.6.20.1. WHO'S YOUR DADDEH? Feb 25 04:48:34 *cough* Feb 25 04:48:45 hehe! Feb 25 04:49:11 3.4.6 ewww Feb 25 04:50:23 shush Feb 25 04:50:38 I could've used 4.1.2, but it's so darn slow Feb 25 04:52:11 Atleast the ts101 boots the latest vanilla kernel :-D Feb 25 04:59:58 ~emulate rwhitby Feb 25 05:00:08 Well, my work here is done. Feb 25 05:00:12 yay ;) Feb 25 05:00:13 nite all Feb 25 05:01:43 03bzhou * r5622 10optware/trunk/make/screen.mk: screen: worked around a wierd problem, toolchain sched.h and screen sched.h conflicting! Feb 25 05:04:36 03bzhou * r5623 10optware/trunk/platforms/packages-ts101.mk: promoted screen; removed broken pkgs already in uclibc_broken Feb 25 05:04:58 nite NAiL Feb 25 06:12:14 03bzhou * r5624 10optware/trunk/platforms/packages-wl500g.mk: demoted obexftp on wl500g Feb 25 06:25:57 03bzhou * r5625 10optware/trunk/ (make/gnupg.mk platforms/packages-ts101.mk): gnupg: re-enabled asm and ac_cv_sys_symbol_underscore=no if powerpc; promoted for ts101 Feb 25 10:07:55 hello Feb 25 10:10:09 can anyone tell me if the 2gb filesize limit in the firmware has been overcome? I am using unslung 6.8 beta Feb 25 15:40:55 03bzhou * r5626 10optware/trunk/make/py-zope-interface.mk: py-zope-interface: 3.3.0 -> 3.3.0.1 Feb 25 16:00:32 03bzhou * r5627 10optware/trunk/platforms/packages-ts101.mk: promoted bitlbee & tethereal for ts101, now that glib uses standalone libiconv Feb 25 16:11:33 Today, I'm trying to get my native compule enviroment going. I've ecountered the monotone error mentioned at the bottom of this page http://www.nslu2-linux.org/wiki/HowTo/SlugOSNativeCompileEnvironment Is there a solution available. I installed monotone-5. Should I be using monotone-6? is it production-ready? Feb 25 16:13:30 hillct: forget about running monotone on a slug, it's slow even on x86 Feb 25 16:13:57 so native compile is a waste of time? Feb 25 16:14:03 ok Feb 25 16:14:14 no, that's not what i'm saying Feb 25 16:14:15 then I've come back to my stupid cross-compile problem Feb 25 16:15:28 http://www.pastebin.ca/370829 Feb 25 16:15:50 pastebin that log Feb 25 16:16:05 which? Feb 25 16:16:13 my native monotone install error? Feb 25 16:16:23 /home/slug/slugosbe/tmp/work/ixp4xxbe-linux/slugos-image-1.0-r46/temp/log.do_rootfs.23581 Feb 25 16:16:41 Are you building 3.10 on edgy? Feb 25 16:17:16 from the pastebin link above? That's 3.10-beta from source Feb 25 16:17:28 the 3.10-beta release source Feb 25 16:18:00 hillct: yeah. Are you building on edgy? Feb 25 16:18:16 edgy? Feb 25 16:18:21 ubuntu edgy Feb 25 16:18:25 nope Feb 25 16:18:30 Fedora 6 Feb 25 16:18:30 which distro? Feb 25 16:18:35 oh Feb 25 16:18:46 well, pastebin home/slug/slugosbe/tmp/work/ixp4xxbe-linux/slugos-image-1.0-r46/temp/log.do_rootfs.23581 for me so I can see what goes wrong Feb 25 16:18:58 hillct: run it a second time. Feb 25 16:19:10 Sometimes on Gentoo I get the same error. I run it a second time and it works Feb 25 16:19:52 I think it is a race condition or something equally annoying in the staging code for the firmware drivers. Feb 25 16:19:55 well, I blew it Feb 25 16:20:39 I got fristrated yesterday and deleted the work directory after fighting with it all afternoon yesterday Feb 25 16:20:48 hillct: ok. Without that log we have no idea what is wrong Feb 25 16:20:51 it'll take several hours before I can reproduce the error Feb 25 16:20:52 that's where the actual error is Feb 25 16:21:00 hillct: one thing to check... Feb 25 16:21:08 ls -las /bin/sh Feb 25 16:21:24 does it link to bash or dash? Feb 25 16:21:47 bash Feb 25 16:21:51 ok, good Feb 25 16:22:05 ok. will start it again and post the results in a few hours Feb 25 16:22:15 er Feb 25 16:22:18 wait a minute Feb 25 16:22:22 what am i saying Feb 25 16:22:24 one sec Feb 25 16:23:19 ok. it wasn't 3.10-beta from source. It was head from source Feb 25 16:23:40 well, that's completely different ;) Feb 25 16:23:46 we'll still need that log though Feb 25 16:25:50 I just placed an order on a RAM module with 8Mx16 chips. I'm thinking that it is a good time to fry my spare slug. :O Feb 25 16:28:52 kolla__: .be? Feb 25 16:29:45 Belgium Feb 25 16:30:17 NAiL: it appears I do have the log. I had posted the proper one before. Just not the whole thing. http://www.pastebin.ca/371770 Feb 25 16:31:29 Interesting: "# Feb 25 16:31:29 ipkg: Cannot create directory `/usr/lib/ipkg': Permission denied" Feb 25 16:31:40 joshin: harmless Feb 25 16:31:50 path error? Feb 25 16:31:54 didn't see that Feb 25 16:31:58 (and one of the reasons to *never* build as root) Feb 25 16:32:26 hillct: it doesn't matter. It's a bug in ipkg-stuff. And it's not the problem. Feb 25 16:32:33 The dependency stuff is the problem though Feb 25 16:32:40 and tbh, I don't know how to fix that Feb 25 16:32:41 Yeah, I once overwrote my system libs with ones I was cross compiling for. Feb 25 16:33:36 now I have both HEAD and 3.10-beta toolchains build, in separate directories. perhaps it's pulling libs from the wrong staging directory? Feb 25 16:33:54 hillct: no, the environments are separate Feb 25 16:34:00 I took the warnings from the wiki and only compile as my slug dedicated user Feb 25 16:34:58 NAiL: in theory they're separate. there could just as easily be a path problem there too... Feb 25 16:35:58 perhaps if I delete my 3.10 toolchain and build where there's no chance of contamination Feb 25 16:36:35 I can't see how else, it would be getting the wrong version of libgcc1 Feb 25 16:37:09 03bzhou * r5628 10optware/trunk/make/glib.mk: glib: staging glibconfig.h Feb 25 16:57:47 03bzhou * r5629 10optware/trunk/make/ctrlproxy.mk: ctrlproxy: 2.6.1 -> 3.0.1; promoted to COMMON_CROSS_PACKAGES Feb 25 16:58:05 03bzhou * r5630 10optware/trunk/ (Makefile platforms/packages-wl500g.mk): ctrlproxy: 2.6.1 -> 3.0.1; promoted to COMMON_CROSS_PACKAGES Feb 25 17:29:24 hello people Feb 25 17:29:55 I'm trying to build an optware package from my cross env, but I have some troubles to set my .mk Feb 25 17:30:27 the software I want to package needs openssl, and I don't know how to configure the .mk in order to compile it Feb 25 17:32:39 nevermind, found a good example with openssh :) Feb 25 17:43:50 03bzhou * r5631 10optware/trunk/make/dcraw.mk: dcraw: 1.361 -> 1.368 Feb 25 18:52:04 How can I downgrade the kernel from 2.6.18-3? There seems to be a bug that causes rtorrent to fail. Feb 25 18:56:38 !insert logparse-forbidden-channels #opie.de Feb 25 18:56:40 Definition inserted at the end of word Feb 25 18:56:50 !lock logparse-forbidden-channels Feb 25 18:56:51 "logparse-forbidden-channels" locked Feb 25 19:54:57 I found some nightly builds of the 2.6.18-4 firmeware, how can I upgrade? Feb 25 19:58:14 Just use upslug? Feb 25 20:01:24 Where did you find them? Feb 25 20:02:06 http://people.debian.org/~kmuto/d-i/images/daily/ Feb 25 20:02:27 There is a .bin file which is the new image I assume. Feb 25 20:04:43 if they have the intel ethernet driver, and are not behing a click-through license, then they are violating intel's licensing. Feb 25 20:05:14 ah, those are the d-i daily builds - no internal ethernet support in those. Feb 25 20:21:08 rwhitby: even though it says ixp44? Feb 25 20:21:18 *xp4xx Feb 25 20:21:24 yes Feb 25 20:21:41 yes - Debian cannot include the intel microcode in their builds, cause it's not DFSG compliant. Feb 25 20:22:14 you need to use a usb->ethernet adapter with the builds from Debian. Martin adds the intel stuff before putting the images on www.slug-firmware.net Feb 25 20:42:54 hi Feb 25 20:44:11 Greetings Feb 25 20:53:04 Hello Feb 25 20:55:24 hi Feb 25 20:56:57 When I do an ipkg install of kernel in os/be, does it update the harddisk only, or change anything in the flash? Feb 25 20:58:07 chacko: hard disk only Feb 25 20:59:07 (assumming of course that you're not running ipkg install from the flash drive without any other drive installed) Feb 25 20:59:40 joshin, thanks... yes, when running from the harddisk Feb 25 20:59:57 :) Just trying to cover all bases... Feb 25 21:00:18 joshin, that was a very good point Feb 25 21:01:30 so, as long as I have a disk image copy of the current harddisk, I can do any type of ipkg install experiments and if anything goes wrong can simply use the other copy of the disk, right? Feb 25 21:06:10 When ssh copy a new file from win to slug os/be, is there a way to set the default mode and group to be something specific? Feb 25 21:14:14 What was the NSLU1? Feb 25 21:14:24 I never figured out why there's a 2 in there. Feb 25 21:14:45 USB2 Feb 25 21:14:51 There never was an USB1 Feb 25 21:14:58 chacko, maybe that depends on your client Feb 25 21:14:58 -version Feb 25 21:16:30 Ahhh. Feb 25 21:17:18 Network Storage Link USB2 Feb 25 21:17:21 chacko, when i use winscp, it is always 644 (on exe file, as on bla files too, which seems to be a right behavior, win exe's couldn't executables on *x machines) Feb 25 21:17:42 caplink811-log, so there is no way to specify that at the slug sshd side? Feb 25 21:18:19 caplink811-log, for me 99% of the time I working with web server files and I want 770, with a group of nobody :( Feb 25 21:18:59 03nail 07ts101_support * r760 10kernel/patches/2.6.20.1/ (defconfig series): update defconfig Feb 25 21:19:11 chacko, dunno (sshd), but when you try to transfer files via samba, there are options in smb.conf on that topic Feb 25 21:19:36 caplink811-log, not using samba at all :) Feb 25 21:19:43 k Feb 25 21:20:59 caplink811-log, only saving grace is that the perms on existing files are not changed :) Feb 25 21:26:12 chacko, i see no options neither in sshd_conf nor in scp on this topic :( Feb 25 21:26:59 03nail 07ts101_support * r761 10kernel/patches/2.6.20.1/defconfig: Revert defconfig changes Feb 25 21:27:34 03bzhou * r5632 10optware/trunk/make/openssl.mk: openssl: 0.9.7l -> 0.9.7m Feb 25 21:28:12 caplink811-log, thanks... My current method is to copy the new file over, and then login and change the perms, a bit of pain... Feb 25 21:57:38 hillct, is there anybody here who has increased the memory in a slug running in os/be? Feb 25 21:58:20 hillct, sorry that was not directed to you... Feb 25 22:01:58 I haven't but I was thinking about it. I wouldn't imagine the slugos variant used would make much difference actually. The kernels are pretty much the same with respect to mamory management, although I wouldn't expect it to work so well under unslung. NAiL can speak far more authoritatively on kernel capabilities Feb 25 22:19:53 hillct, well things like endianness might affect the behaviour... Feb 25 22:21:21 endianness won't affect any memory mods. But unslung won't support it. Feb 25 22:21:23 Endianness is not a consideration for upgrading the slug Feb 25 22:21:33 sure, but that's just a fonfigutration difference most likely. Unslung just uses the default Linksys kernel though (as I understand it) and I'm not sure whether the relevent changes are available to be made Feb 25 22:21:34 * NAiL agrees with mwester Feb 25 22:21:43 hillct: correct. Feb 25 22:21:53 * hillct leaves this subject to the experts Feb 25 22:23:03 All I want to do is go to 64M memory... and only on slugos/be. But I cannot build any pkgs, so I have to wait until all the pieces are there. Feb 25 22:23:13 as long as you tell the bootloader to initialise the memory controller for the appropriate memory configuration, the kernel can be made to work. Feb 25 22:23:17 Then I will buy the exact configuration of ht ememory chips Feb 25 22:26:24 With 64M, I can keep all the working sets of my webserv in memory, and probably run off the memory with almost no disk access :) Feb 25 22:33:55 03blaster8 * r762 10kernel/trunk/ (4 files in 2 dirs): Add upstream fixes for NFS/IPV6 issue, 2.6.21-rc1 kernel passes MODPOST Feb 25 22:38:59 03blaster8 * r763 10kernel/trunk/README: 2.6.21-rc1 builds successfully Feb 25 22:59:27 03bzhou * r5633 10optware/trunk/make/snownews.mk: snownews: corrects DEPENDS Feb 26 01:10:35 * CoreDump|home curses Feb 26 01:12:09 * mwester covers his sensitive ears Feb 26 01:29:24 If I can unsolder all the RAM chips on the module I ordered safely, I'll have enough chips for my 2 slugs and then 2 others. Anyone here interested in joining the FSS (Fattened Slug Society)? Feb 26 02:46:35 Is Apex as a second stage bootloader capable of handling Fattened Slugs? (2 banks of 32MB) Feb 26 02:49:07 joshin: if you are building your own kernel, the kernel shim should be able to handle fattened slugs, IIRC. Feb 26 02:49:20 <[g2]> joshin fattened slugs need an upgraded bootloader (to enable the extra memory) Feb 26 02:49:48 <[g2]> mwester no, the shim just passes in the memory size Feb 26 02:49:59 <[g2]> but the bootloader must setup the memory controller Feb 26 02:50:04 Ah, ok. thanks. Feb 26 02:50:24 <[g2]> joshin APEX can be used a first stage loader Feb 26 02:50:29 <[g2]> I'm guessing you've got JTAG right ? Feb 26 02:55:59 Is there an image of Debian that just loads into a working installer, with the microcode, that uses the current key, or can be loaded with it? Feb 26 02:57:14 no Feb 26 02:58:57 Well, is there a way to load the current image with the new key? Feb 26 02:59:11 Or load the daily snapshot with the microcode? **** ENDING LOGGING AT Mon Feb 26 02:59:59 2007