**** BEGIN LOGGING AT Sat Mar 25 10:59:57 2006 Mar 25 11:03:22 NAiL: trying to find out where the modules should be loaded then Mar 25 11:03:30 don't see a noticable change Mar 25 11:03:46 (not flashed yet, jsut examining rootfs directory) Mar 25 11:10:53 typo in conf ;) Mar 25 11:11:29 haha Mar 25 11:22:30 03repvik * r292 10/releases/slugos-3.6-beta/openembedded/conf/distro/ (debianslug.conf openslug.conf ucslugc.conf): debian/open/ucslugc: Add correct usb/storage modules to rootfs Mar 25 11:23:33 03repvik * r293 10/releases/slugos-3.6-beta/openembedded/packages/meta/slugos-packages.bb: slugos-packages: add cups to feeds Mar 25 11:24:52 hmmm, just ran diff on old and new rootfs Mar 25 11:25:05 only /etc/version changed Mar 25 11:31:53 03repvik * r294 10/releases/slugos-3.6-beta/openembedded/packages/slugos-init/files/leds: slugos-init: New leds script. Initscripts have not been updated to use it! Mar 25 11:32:46 that was half my to-do... Mar 25 11:39:43 03repvik * r295 10/releases/slugos-3.6-beta/openembedded/packages/slugos-init/files/leds: slugos-init: Update leds script to check if sysfs is mounted Mar 25 11:44:01 NAiL: I don't even know how to compile the kernel with this weird tool Mar 25 11:44:44 tbm: I would use debian, but still I don't know why it failed Mar 25 11:45:20 NAiL: I get an error when trtying to compile a openslug-image Mar 25 11:45:28 within the first commands Mar 25 11:45:40 what errors? Mar 25 11:46:21 /bin/sh: /media/sda3/scratch/debian_work/debonaras/openslug/bitbake/bin/bitbake: /usr/bin/env: bad interpreter: Permission denied Mar 25 11:47:14 but the real question is, how can I compile just one module? Mar 25 11:47:37 I don't need to build the whole image just for on kernel module Mar 25 11:48:53 to build the kernel, you need to build the tools. This is automated with the master makefile. Once you've built an image, building a custom kernel is easy. Mar 25 11:50:24 what is the target that builds the tools? Mar 25 11:50:56 make -image Mar 25 11:51:51 NAiL: for some reason I don't believe you :-D Mar 25 11:52:14 feel free to not believe me. Won't help you though ;) Mar 25 11:52:30 no, really, isn't there a dedicated target just for the toolchain? Mar 25 11:52:35 no Mar 25 11:52:59 in that case, how do I get over the error above? Mar 25 11:56:21 is there a paste-bin like page for this channel? Mar 25 11:56:32 no, just use pastebin.com Mar 25 11:56:45 ok, will paste erros there Mar 25 11:57:36 http://pastebin.com/621497 Mar 25 11:58:54 ls -las /usr/bin/env? Mar 25 11:59:01 any symlinks in that path, btw? Mar 25 11:59:53 NAiL: no Mar 25 12:00:08 03repvik * r296 10/releases/slugos-3.6-beta/openembedded/packages/slugos-init/files/leds: leds: make sure sysfs is mounted, and unmount after use if tmpmounted Mar 25 12:00:12 there is a symlink that points to it, though Mar 25 12:00:19 ~/slug Mar 25 12:00:37 you can't build from ~/slug Mar 25 12:00:44 points to /media/sda3/scratch/debian_work/debonaras/ Mar 25 12:00:46 since it's a symlink Mar 25 12:01:08 yes, but I am building from the target, not the symlink Mar 25 12:01:13 ok Mar 25 12:01:24 what's the permissions of env? Mar 25 12:01:27 make kernel does not work either Mar 25 12:01:39 default on debian(?) Mar 25 12:02:08 -rwxr-xr-x 1 root root 13888 2006-03-02 22:08 /usr/bin/env Mar 25 12:02:19 and you can run it manually? Mar 25 12:02:27 yes Mar 25 12:04:13 eddyp: /j openslug Mar 25 12:04:26 tbm: NAiL: is the debianslug target related to LE debian, or BE debian? Mar 25 12:04:31 le Mar 25 12:04:41 be debian just runs off of openslug Mar 25 12:05:08 NAiL: at least I got that right :) Mar 25 12:05:16 ;) Mar 25 12:06:22 hi all. I am trying to make the image from the svn, but getting problems with the ports_sources.redhat.com__20050627.tar.gz download. Anyone seeing this? Mar 25 12:06:38 yes Mar 25 12:06:49 unpack the tarball, repack it Mar 25 12:06:55 OK Mar 25 12:06:59 you *might* have to make a new md5sum for it Mar 25 12:07:13 it's just got crap attached on the end :) Mar 25 12:09:18 thanks. it started compiling Mar 25 12:10:00 np Mar 25 12:11:19 NAiL: not seeing differences in rootfs with modules_autoload :( Mar 25 12:12:02 now cleaned out work/* cache/* and removed stamps/slugos-image* and stamps/debianslug* Mar 25 12:26:28 still no change Mar 25 12:41:27 odd Mar 25 12:41:36 I'll try to look at it later. Cleaning house now. Mar 25 13:35:33 Hi again! Mar 25 13:37:55 Is there - after flashing Unsung5.5 - any difference between the capatibilities of the ports? My FlashDrive (in Fact MP3-Player) is only recongnized on USB1 and not on USB2. But i CAN mount the /dev/sdb1 and /dev/sdb2 manually - seems strange to me. Mar 25 13:38:24 IIRC, yes. But I'm not familiar with unslung. Mar 25 13:43:06 Hmm. Unslung does not have its own channel, right? Mar 25 14:14:35 What is the difference in loaded modules between booting Unslung5.5 from the internal Flash and booting from the Disk i've unslung to? /dev/sdb1 on my external WD Drive is there when i boot from the internal Flash but not, when booting from a FlashDrive in USB1, where i have unslung to :/ Mar 25 14:48:16 NAiL: ref. module loading, now trying a rebuild of kernel (or anything that depens on kernel.bbclass) to see if the /etc/modules file is populated Mar 25 14:48:46 working on about four things simultaneously ;) Mar 25 15:09:23 http://www.nslu2-linux.org/wiki/HowTo/TestAnImageInRamUsingRedBoot updated for FIS directory based images Mar 25 15:09:49 cool Mar 25 15:10:37 should add a Mar 25 15:10:37 should add a SplitAnImageIntoPartsUsingSlugImage page though Mar 25 15:10:51 since SlugTool is obsolete Mar 25 15:20:45 NAiL: now /etc/modules *is* populated Mar 25 15:20:46 + some Mar 25 15:21:30 or better, some files are found in /etc/modutils Mar 25 15:23:16 still not quite convinced this will work Mar 25 15:23:18 but looks better Mar 25 16:56:12 pepijn: did it work? Mar 25 16:59:08 looking at the initscripts i would think not Mar 25 16:59:16 /etc/init.d/modutils reads /etc/modules Mar 25 16:59:23 but /etc/modules is not populated Mar 25 16:59:53 howerver /etc/modutils.d gets (for example) usb-storage file, contents: usb-storage Mar 25 17:00:14 and ipkg postinit is different than before Mar 25 17:00:24 for usb-storage module Mar 25 17:00:55 --- rootfs.nousb/usr/lib/ipkg/info/kernel-module-usb-storage.postinst 2006-03-23 23:49:57.000000000 +0100 Mar 25 17:00:55 +++ rootfs/usr/lib/ipkg/info/kernel-module-usb-storage.postinst 2006-03-25 17:51:36.000000000 +0100 Mar 25 17:00:55 @@ -6,3 +6,6 @@ Mar 25 17:00:55 update-modules || true Mar 25 17:00:55 fi Mar 25 17:00:58 Mar 25 17:01:00 +if [ x"$D" = "x" ]; then Mar 25 17:01:02 + modprobe usb-storage || true Mar 25 17:01:04 +fi Mar 25 17:01:43 found several references on the net about the module_autoload feature of OE Mar 25 17:02:15 but can't figure out who is responsible for the updating of /etc/modules Mar 25 17:02:29 but that does seem to be the trick, when using modutils, that is Mar 25 17:02:35 bit suprised by that though Mar 25 17:02:44 should it not be module-init-tools ??? Mar 25 17:02:51 hmm Mar 25 17:02:53 dunno Mar 25 17:03:07 that is after all the 2.6 replacement for modutils (from pre 2.6 kernels) Mar 25 17:03:14 ah Mar 25 17:03:49 btw, in one IRC log i found, you were participating too ;) Mar 25 17:03:57 on #oe Mar 25 17:04:18 in module-init stuff? Mar 25 17:04:43 found that after a search for 'module_autoload oe' Mar 25 17:04:58 I do participate in #oe once in a while Mar 25 17:06:08 I do way too much development to remember a fraction of it :-P Mar 25 17:06:41 oh, don't get me wrong, was not suggesting you should remember, just found it funny ;) Mar 25 17:07:17 If I saw the context, I'd probably remember though ;) Mar 25 17:07:24 yeah Mar 25 17:07:55 http://purl.rikers.org/%23oe/20050807.html.gz Mar 25 17:07:56 ;) Mar 25 17:09:04 interesting part there (in our current context): Mar 25 17:09:05 20:37.09 lamikr pb: Hi, koen suggested to ask whether you know how to get oe build to add new module to /etc/modules file? Mar 25 17:09:06 20:37.44 lamikr pb: We would need to get omapts added there when we make gpe-image for familiar-0.8.3 distro. Mar 25 17:09:06 20:39.02 pb_ lamikr: generally, the hotplug mechanism would take care of that. But, if that isn't happening, you can use the MODULE_AUTOLOAD mechanism instead. Mar 25 17:09:06 20:39.54 JustinP NAiL: I doubt it. You can always run a vncserver on a linux box Mar 25 17:09:09 20:40.08 lamikr Where is this MODULE_AUTOLOAD method done? Mar 25 17:09:10 20:40.15 pb_ In kernel.bbclass. Mar 25 17:09:48 vncserver to get a remote desktop so I could run monotone-viz (or something) Mar 25 17:09:56 not really anything about module_autoload Mar 25 17:10:17 anyway Mar 25 17:10:51 the thing i don't get is that it is actually neccessary Mar 25 17:11:06 hotplug should do the trick ref. usb modules Mar 25 17:11:13 hmm Mar 25 17:11:17 at least for the base (hardware support) modules Mar 25 17:11:40 do you have a slug where the modules aren't loaded? Mar 25 17:11:51 I'd like to try one thing... Mar 25 17:12:21 i have a single slug, and it has a kernel with modules built into kernel Mar 25 17:12:25 ah Mar 25 17:12:50 but i'd like to get the slugos-3.6 stuff loaded onto it Mar 25 17:12:52 There was something about "uevent" in /sys. Mar 25 17:13:34 ah, hmm Mar 25 17:13:39 that doesn't create the devices Mar 25 17:13:41 All device events happening during the boot process until the udev daemon is running are obviously lost, because the infrastructure to handle these events lives on the root filesystem and is not available at that time. To cover that hole, the kernel now provides an uevent file for every device in the sysfs filesystem. By writing add to that file, the kernel will just emit exactly the same event again, as the one lost during bootup. A simp Mar 25 17:15:59 i am certain that adding lines to /etc/modules solves the problem of critical modules not being loaded Mar 25 17:16:09 just wondering: Mar 25 17:16:27 1) why this is neccessary for the uhci/ohci/ehci modules Mar 25 17:16:51 2) who (in the OE build process) is responsible for adding modules to /etc/modules Mar 25 17:44:03 guess i'll just have to give the image a spin Mar 25 17:50:08 03bzhou * 10unslung/Makefile: php-fcgi ready for testing Mar 25 17:58:10 NAiL: koen has a point Mar 25 17:58:22 S98configure runs 'ipkg-cl configure' Mar 25 18:04:09 but i would not expect /etc/modules to get populated then Mar 25 18:04:17 just a modprobe being executed by it Mar 25 18:05:13 I guess that is something that should be fixed Mar 25 18:06:36 if it works on every boot, it should be ok Mar 25 18:07:32 i'll run a testboot a bit later, as soon as i can 'kick' my wife from here PC (need to log er off) and my son is to bed ;) Mar 25 18:59:27 NAiL: it *did* work ;) Mar 25 18:59:44 good, send a diff to the ml ;) Mar 25 18:59:47 (cc me) Mar 25 18:59:47 /etc/modules had lines added Mar 25 19:00:21 which ml ? Mar 25 19:00:25 dev ? Mar 25 19:00:29 yeah Mar 25 19:00:31 k Mar 25 19:06:09 cc to ml.org ? Mar 25 19:07:10 ml.org? cc to nail @ nslu2-linux.org atleast ;) Mar 25 19:07:41 k, reading ml thru NNTP so no email from there Mar 25 19:18:21 Chatzilla went ballistic and crashed... Mar 25 19:53:07 Do I need to worry about the following warning during the kernel build from NSLU CVS at SF.net? "*** Warning: "ixOsalIrqLock" [/home/leon/sandbox/ixp4xx/nslu/kernel/ixp400_eth/ixp400_eth.ko] undefined!" Mar 25 19:53:25 (and similar warnings, this is the last). Mar 25 19:53:41 did you download the intel code? Mar 25 19:53:45 yes Mar 25 19:55:12 hmmm Mar 25 19:56:20 but the build continues? Mar 25 19:56:26 just warnings? Mar 25 19:56:55 yes build completes, yes. Mar 25 19:57:32 (I am rebuilding as we speak, with a 20000 line scroll buffer this time...) Mar 25 19:58:24 best way to build (AFAIK) is from 'Mastermakefile' and not from sf.net, but then again i never built unslung Mar 25 19:59:50 see: http://www.nslu2-linux.org/wiki/Unslung/BuildingUnslungAlphaVersion Mar 25 20:14:48 Ah, could it be all the kernels are LE? Mar 25 20:14:54 I am using a BE toolchain. Mar 25 20:15:05 and want to build BE binaries. Mar 25 20:15:23 unlikely Mar 25 20:15:37 reverse is more problematic, actually Mar 25 20:19:45 what do you mean by unlikely? Mar 25 20:20:00 kernels default to BE Mar 25 20:20:14 you have to perform tricks to get LE working Mar 25 20:20:37 (because slug starts up in BE mode) Mar 25 20:20:58 So the NSLU kernels are BE as well? Some of them (all) switch to LE mode. Mar 25 20:21:33 one count of LE i know of is to be able to use the Debian arm port Mar 25 20:21:39 that is LE based Mar 25 20:23:05 Well, the CVS kernel build for NSLU seem LE: /home/leon/sandbox/ixp4xx/buildroot/build_armeb_nofpu/staging_dir/bin/armeb-linux-gcc -D__KERNEL__ -I/home/leon/sandbox/ixp4xx/nslu/kernel/linux-2.6.16/include -Wall -Wno-trigraphs -fno-common -pipe -msoft-float -DMODULE -Isrc/include -D__linux -DCPU=33 -DXSCALE=33 -D__LINUX_ARM_ARCH__=5 -mtune=xscale -DIX_UPTOPIAMODE=0 -DIX_MPHY=1 -DIX_MPHYSINGLEPORT - Mar 25 20:23:06 mlittle-endian Mar 25 20:23:34 (see last keyword) Mar 25 20:28:16 my remarks are based on building from openembedded/mastermakefile Mar 25 20:29:03 Pepijn: OK, so that one builds a BE kernel according to you? Mar 25 20:30:23 as long as you do not use debianslug as targe Mar 25 20:30:23 t Mar 25 20:30:52 but it would be nice to know about the current status of the stuff you downloaded Mar 25 20:31:22 looking at the kernel-version it is quite recent Mar 25 20:31:41 but i know for a fact that the OE stuff is also actively maintained Mar 25 20:33:42 Thanks for the info. Mar 25 20:34:08 I will try out to see if the OE is easier to base my platform on. Mar 25 20:38:17 03marceln * 10unslung/ (make/nmap.mk sources/nmap/control): Converted the makefile to the new template style. Mar 25 20:41:29 03marceln * 10unslung/make/nmap.mk: Added pcre to the required package (http://groups.yahoo.com/group/nslu2-linux/message/12063). Mar 25 21:06:20 two of the nslu2 kernels are BE, two are LE Mar 25 21:06:36 I *think* the warnings you got are harmless Mar 25 21:06:49 (ie, I seem to remember I got them and stuff still worked) Mar 25 21:06:56 NAiL: so what about the status of CVS @ sf.net ? Mar 25 21:07:08 cvs @ sf.net is the bleeding edge kernel dev Mar 25 21:07:14 it has its own makefile Mar 25 21:07:30 it's pushed to OE once in a while Mar 25 21:07:50 just kernel? just unslung? Mar 25 21:08:02 those two are separate Mar 25 21:08:15 the two commits above are in the unslung repo Mar 25 21:08:23 the kernel has its own repo Mar 25 21:08:31 2.6-kernel, that is Mar 25 21:09:02 but the kernel for openslug is a stock kernel with patches, or? Mar 25 21:09:25 yes, and it's developed in the cvs repo mainly Mar 25 21:09:32 I can't remember the reason for this Mar 25 21:09:59 the repoman Mar 25 21:10:08 kolla: ltns ;) Mar 25 21:10:28 been idling :) Mar 25 21:11:07 NAiL: last patch seems ok Mar 25 21:11:12 rootfs looks fine Mar 25 21:11:35 (scsi-mod is not there since both usb-storage and sd_mod depend on it) Mar 25 21:12:22 pepijn: good Mar 25 21:12:44 I'm not sure if it needs to be in nslu2.conf or the distro-specific .conf Mar 25 21:13:26 kolla: are you at elgeseter? Mar 25 21:13:33 * AwayNAiL is Mar 25 21:13:43 i would say that all nslu2's need it Mar 25 21:14:12 pepijn: nslu2's, yes. ds101, no. Mar 25 21:14:25 but does ds101 use nslu2.conf ? Mar 25 21:14:34 yes Mar 25 21:14:36 as well as the lude Mar 25 21:14:37 ah Mar 25 21:14:39 and dsm600g Mar 25 21:14:41 NAiL: 2 are BE, 2 are LE, but that conflicts with it being a unified kernel, doesn't it? Starting make builds the kernel only once here. Mar 25 21:15:17 uhm scrap that Mar 25 21:15:28 there's one BE and one LE kernel Mar 25 21:15:45 it's decided by the CONFIG_BIG_ENDIAN in 2.6.16/defconfig Mar 25 21:16:48 maybe it should build both a LE and a BE kernel, but we haven't needed that so far Mar 25 21:17:13 Yes, I can select the ENDIAN variable in the Makefile at the top level of kernel. Mar 25 21:17:32 that is only when you want to build all, i.e. for binary distribution Mar 25 21:17:46 not when you know what you want, for private use Mar 25 21:17:51 No biggy, I just wondered if BE is supported by the patches, i.e. if it was tested on devices. Mar 25 21:18:00 yes, it's supported and tested Mar 25 21:18:06 The answer is yes then. Mar 25 21:18:11 Check. Mar 25 21:18:37 A big-endian build gives the same warnings. Let's going for a test run then. Mar 25 21:18:45 go actually Mar 25 21:19:05 * likewise runs up the stairs searching the IXDP425 in some carton boxes... Mar 25 21:19:38 test run, haha Mar 25 21:19:53 it did include actual running... Mar 25 21:20:58 any more kernel-related questions before I go back to my beer? ;) Mar 25 21:21:37 not from me, enjoy Mar 25 21:21:46 NAiL: beer and kernels make a good pair. Mar 25 21:21:57 NAiL: but I'll leave you to it. Thanks. Mar 25 21:22:03 hehe Mar 25 21:22:14 Later people :) Mar 25 21:22:18 * likewise waves Mar 25 21:22:25 "Skål" Mar 25 21:22:48 * likewise runs up the stairs again (for some exotic serial cable)... Mar 25 21:22:49 NAiL: you dropped an aa (?) there Mar 25 21:23:19 if only everybody used UTF-8 Mar 25 21:25:09 ziet er aardig Noors uit bij mij. Mar 25 21:25:51 guess my UTF-8 based macosx system has problems with its home language ;) Mar 25 21:27:31 * pepijn thinks the norwegian government should seriously consider dropping the three extra letters from its alphabet Mar 25 21:28:13 would give an enormous econimical benifit to not have to deal with those problems Mar 25 21:28:16 where're you from pepijn? Mar 25 21:28:26 from? or living? hehe Mar 25 21:28:38 .nl -> .no? Mar 25 21:28:38 i am a dutchman living in norway ;) Mar 25 21:28:57 since how long now? Mar 25 21:29:05 (the living in Norway part :-)) Mar 25 21:29:35 since december 2000 Mar 25 21:30:32 in NL i never really experienced problems with not having an ASCII alphabet Mar 25 21:30:41 My girlfriend and I are considering moving. Canada, Norway and Ireland are candidates. So I am quite interested to hear your overall experience! Mar 25 21:34:07 pepijn: My private msgs seem not to come through on the other window as I did not register myself yet? Mar 25 21:34:25 but you get mine? Mar 25 21:34:53 but yes, registration with nickserv might be causing that Mar 25 21:35:01 they recently changed that Mar 25 21:35:35 rwhitby: i changed the 'test an image from ram' wiki page to include the limitations for the newer builds Mar 25 21:35:56 pepijn: thx Mar 25 21:58:02 likewise: I'm norwegian, and been living in Ireland ;) Mar 25 21:58:41 Norway is way more expensive. But then again, you get paid so you can afford it ;) Mar 25 21:58:46 * pepijn is looking around for 'the canadian connection' Mar 25 21:58:59 * AwayNAiL spent a month in Quebec Mar 25 21:59:16 yes, the wages here make going abroad fun Mar 25 21:59:27 .nl is actually one of the few countries in .eu I haven't visited yet Mar 25 22:00:29 well, the you should Mar 25 22:00:35 s/the/then/ Mar 25 22:00:35 pepijn meant: well, then you should Mar 25 22:00:39 * AwayNAiL used to have a salary of 42.260 USD ;) Mar 25 22:00:50 uh, 42,260 Mar 25 22:01:06 * AwayNAiL always forgets.. The , and . are reversed here in .no Mar 25 22:01:35 yeah, I know. But all the people I know that visit .nl go to amster to get high. And I really don't feel much like that ;) Mar 25 22:01:46 huh? reversed her in .eu? you mean reversed on the other site of the ocean ;) Mar 25 22:01:54 well, yeah ;) Mar 25 22:03:16 well, .nl is more than am*dam and getting high Mar 25 22:03:30 yes, and I'd like to go there Mar 25 22:03:59 cheap beer, for example ;) Mar 25 22:04:01 hehe Mar 25 22:04:44 * AwayNAiL likes Kirin Mar 25 22:05:02 http://www.kirin.com/ ;) Mar 25 22:06:38 * AwayNAiL used to live next to the Heineken brewery in Cork... For some strange reason, I drank a lot of heineken ;) Mar 25 22:06:54 "next to" as in ~4 meters away Mar 25 22:07:00 hehe Mar 25 22:07:11 AwayNAiL: I liked the Irish isles "near" Cork Mar 25 22:07:27 The view from my bedroom window was a gigantic "Heineken" banner Mar 25 22:07:33 *nothing* else ;) Mar 25 22:07:53 likewise: Cork isn't the place you move to. It's a place you move from. Mar 25 22:08:08 It's an industrial town Mar 25 22:08:46 AwayNAiL: haha. Well haven't been there. Only hiked along the coastline with a tent. Excellent! Mar 25 22:08:53 yes Mar 25 22:08:53 NAiL: just flashed image built with patch i send earlier, and /dev/sd devices are available :) Mar 25 22:08:59 likewise: Been to Galway? Mar 25 22:09:05 pepijn: grand! :D Mar 25 22:09:10 * AwayNAiL liked Galway Mar 25 22:09:40 or just the other side of .ie? :) Mar 25 22:10:33 likewise: Any specific places you consider in Norway? Mar 25 22:11:36 AwayNAiL: Isles near Kerry up to Galway (maybe I am mistaken about Cork). Then, straight to Dublin for our flight out. Only been in .ie for 3 weeks Mar 25 22:12:27 AwayNAiL: No, we both should probably go there first. We have not orientated ourselves very well yet. Mar 25 22:12:40 sorry, that was meant for pepijn: Mar 25 22:13:19 well, if you come here, feel free to visit me ;) Mar 25 22:13:26 * AwayNAiL gotta go again Mar 25 22:13:32 this time we're going out ;) Mar 25 22:13:35 later Mar 25 22:13:53 CYa Mar 25 22:14:22 msg nickserv identify 9reen Mar 25 22:14:44 oops Mar 25 22:14:55 eno: never mind :-) We all need a beer. Mar 25 22:17:35 NAiL: now i am very uncertain about what happens when i run 'turnup disk /dev/sda1' Mar 25 22:17:44 i'll rollback for now Mar 25 22:17:52 investigate tomorrow Mar 25 22:20:36 night all Mar 25 22:23:05 03mwester * r20 10/trunk/unslung/: Removed the extern references; these will be replaced by physical copies. MJW Mar 26 01:32:27 03mwester * r21 10/trunk/unslung/ (10358 files in 2684 dirs): Import of Unslung 6.8-beta sources Mar 26 03:04:19 03mwester * r22 10/releases/Unslung-6.8-beta/: Create Unslung 6.8-beta release Mar 26 03:19:52 03mwester * r23 10/trunk/unslung/Makefile: Remove nslu2-linux from tarball Mar 26 03:52:07 go mwester! Mar 26 04:09:12 03mwester * r24 10/releases/Unslung-6.8-beta/Makefile: Remove nslu2-linux from tarball Mar 26 04:22:31 rwhitby: svn repo is set, the Makefile can create a tarball -- now how do we file-release the tarball? Mar 26 04:22:53 mwester: chat to ByronT Mar 26 04:23:32 Ok, I'll go find him... Mar 26 05:11:57 jbot_, test Mar 26 06:55:32 i need some help with understanding boot, ref. beta 3.6 release of slugos Mar 26 06:55:55 i doubt that 'turnup disk' will work on a NSLU2 Mar 26 06:56:16 since usb stuff is modularized and never loaded Mar 26 06:56:36 but how should this be solved? Mar 26 07:25:44 rwhitby: ping? Mar 26 07:53:34 mornin Mar 26 07:54:31 goedemorgen Mar 26 07:57:17 is the updated slugos.conf already available in monotone? Mar 26 07:57:46 you are thinking about the one with changed RDEPENDS? Mar 26 07:57:54 yes Mar 26 07:58:10 i am not sure, it is available on svn.berlios.de Mar 26 07:58:15 but there are more problems Mar 26 07:58:24 oh, like? Mar 26 07:58:31 modules are not loaded (just available) Mar 26 07:58:59 see ml nslu2-devel Mar 26 07:59:44 i fixed that yesterday (not checked in, just confirmed that it works) Mar 26 08:00:09 will chech the ml Mar 26 08:00:16 but there is still another issue after running 'turnup disk' and running debian on the slug Mar 26 08:01:29 loading the modules is simple, just list them in /etc/modutils/somefile (I used /etc/modutils/modules) and then run modutils.conf or so and make sure these files end up on the root fs Mar 26 08:02:19 the fix on the ml uses the OE method for fixing that ;) Mar 26 08:02:37 loading is simple, doing it the 'proper' way was a bit more of a challenge Mar 26 08:03:14 key is that modules should be added to /etc/modules (which is loaded by /etc/init.d/modutils) Mar 26 08:03:49 do you have an article nr or title keyword? I browsed three days back but apparently I overloooked it Mar 26 08:05:06 update-modules does create /etc/modules based upon what is in /etc/modutils/*, thought that was the proper way Mar 26 08:05:13 http://groups.yahoo.com/group/nslu2-developers/ Mar 26 08:05:37 well, yes, but those files need to be placed there by the build process ;) Mar 26 08:06:02 (two most recent posts) Mar 26 08:06:06 ah misread, was looking into nslu2-linux Mar 26 08:07:13 issue now is that 'turnup disk' will fail to boot, because it never loads the modules and just tries to mount the disk Mar 26 08:07:24 which will not be there since the modules are not there Mar 26 08:07:46 (did not try, but based on code analyzing) Mar 26 08:09:08 a crude fix would be to add the modprobes in boot/disk Mar 26 08:09:25 i guess that will be the fix, yes Mar 26 08:09:31 is the turnup failing or the booting Mar 26 08:09:42 the booting will fail Mar 26 08:09:58 turnup check validity of device and with the fixes in the build, that will work Mar 26 08:10:17 /dev/sda1 can be mounted w/o problems when booting from flash Mar 26 08:10:23 or turnup should be callled later after these modules are loaded Mar 26 08:10:36 which runs the flashes /sbin/init and thus runs /etc/init.d/modutils Mar 26 08:10:54 yeah but boot/disk is probably called before the modules are loaded Mar 26 08:11:01 yes, exactly Mar 26 08:11:25 the kernel uses init=/linuxrc which after turnup disk points to boot/disk Mar 26 08:11:54 so none of the flashes init is executed, it just mounts the disk, pivots to it and calls the disk's init Mar 26 08:12:14 think same is true for BE debian Mar 26 08:13:21 and yes, only solution i see is to get the modules loaded before the mount is executed, which means adding the loading to boot/disk Mar 26 08:13:51 maybe copy the modutils code for loading from /etc/modules to the top of boot/disk, or to /etc/default/functions Mar 26 08:14:22 but i'd like to get that confirmed by someone else Mar 26 08:14:31 or approved ;) Mar 26 08:14:50 understood Mar 26 08:15:08 i mean, solving a problem is one thing, but staying within boundaries of the 'philosophy' is quite something else Mar 26 08:16:14 an alternative for debian would be to use an initrd, but then the code would branch from the slugos development Mar 26 08:17:16 i agree, Mar 26 08:17:24 but i guess it is difficult to get ppl to respond on sunday mornings ;) Mar 26 08:17:32 in your mail I saw 5 modules, thought there were actually 6 or 7 Mar 26 08:17:48 * pepijn is usually up early on sundays to take care of his little son Mar 26 08:18:05 there are 6, depending on hardware Mar 26 08:18:32 but both sd_mod and usb-storage depend on scsi-mod, so loading of that is automatic Mar 26 08:19:00 uhci, ehci and ohci depend on hardware (usb 1.0, 1.1 and 2.0, in some order or another) Mar 26 08:19:50 understood, knew abou the *hci things Mar 26 08:20:29 as I want to give a demo next sat and do not have that much time the upcoming week I and I do want the latest pvr driver I decided for a quick workaround Mar 26 08:20:39 made these for now a y in defconfig Mar 26 08:21:34 then go with the quick workaround, i.e. modprobe in top of boot/disk Mar 26 08:21:40 not the most elegant way but I cannot fix this Mar 26 08:22:25 yeah, but I am on monotone so I need to patch the slugos.conf, your ml change and hack the modprobe Mar 26 08:23:11 yes, if none of the changes propegated Mar 26 08:23:54 bta, NAiL did not patch slugos.conf but the different distro confs (ref. copying of modules to /lib/modules) Mar 26 08:24:00