**** BEGIN LOGGING AT Thu Aug 11 23:59:56 2005 Aug 12 00:00:22 hmmm Aug 12 00:00:34 I have installed 'perl-template-toolkit' using ipkg Aug 12 00:00:52 however, i cannot find any of the files listed here: http://ipkgfind.nslu2-linux.org/details.php?package=perl-template-toolkit&id=481709&official=&format= Aug 12 00:01:01 can anyone confirm? Aug 12 00:01:38 apologies if this isn't a developer question...hoping some perl programmers would know Aug 12 00:02:57 example: ./opt/lib/perl5/site_perl/5.8.6/armv5b-linux/Template/Plugin/XML/DOM.pm is not found Aug 12 00:05:29 argh i'm an idiot **** BEGIN LOGGING AT Fri Aug 12 07:18:11 2005 Aug 12 08:25:42 03jbowler 07org.openembedded.nslu2-linux * r09d7f197... 10/conf/ (distro/ucslugc.conf machine/nslu2.conf): Aug 12 08:25:42 nslu2.conf: change to the IPKG_ARCHS to include the correct architectures Aug 12 08:25:42 (armv*b for all armv* suppported by the xscale processor). The old Aug 12 08:25:42 'generic' armeb is still suppported (of course) and, by default, this is Aug 12 08:25:42 the ARCH of the generated packages - this is technically wrong, the Aug 12 08:25:44 packages are actually armv5teb - but changing this would make the feeds Aug 12 08:25:46 non-functional for existing systems. Aug 12 08:53:35 Hi, where can I find the default OpenSlug kernel config? Aug 12 08:54:36 under openembedded/packages/linux/openslug-kernel-2.6.12.2/ I think Aug 12 08:56:11 ok thanks :) Aug 12 08:57:21 is recompiling the kernel on openslug pretty much like on any linux box? Aug 12 08:57:31 untar under /usr/src/linux, make etc? Aug 12 09:01:29 matt_: you shoud try to use the benefits of bitbake Aug 12 09:02:01 do you have the toolchain installed? Aug 12 09:02:13 on a desktop pc? Aug 12 09:02:35 no... how would bitbake help me for kernel recompiling? Aug 12 09:03:46 well, you can just do your modifications, invaludate the stamps and use bb openslug-kernel to do the rest Aug 12 09:03:55 invalidate Aug 12 09:06:10 invalidate the stamps? Aug 12 09:06:27 you mean use bitbake to create a custom firmware on my desktop then upslug it to my slug? Aug 12 09:10:36 you can upslug the whole image, or you can reflash just the kernel if you have a recent openslug on the slug with reflash -k after installing the new kernel ipk. Aug 12 09:12:52 I see - so it'll only reflash the affected part of the image? Aug 12 09:13:25 what does invalidating the stamps mean? Aug 12 09:14:19 when using the bitbake build system, a stamp is created at major points during the build process Aug 12 09:14:39 usualy in tmp/stamps Aug 12 09:16:18 what does the stamp represent? like a file hash? Aug 12 09:16:33 and how do i invalidate the stamps? Aug 12 09:17:15 No, they are simple files. Just delete them Aug 12 09:18:37 ah Aug 12 09:18:44 but what do they do? Aug 12 09:19:49 For example, after making changes to work/openslug-kernel-2.6.12.2-r8/ you can delete openslug-kernel-2.6.12.2-r8.do_compile and the rest. Then you issue bb openslug-kernel and it will repeate the phases Aug 12 09:20:22 openslug-kernel-2.6.12.2-r8.do_* are in tmp/stamps Aug 12 09:20:28 I see Aug 12 09:20:45 so after that will I need to use make again, or will the flash image be ready? Aug 12 09:21:53 If you properly remove the stamps from the modification point till the end, bb will package the kernel and the module files Aug 12 09:22:06 package them into the flash file? Aug 12 09:22:12 or do I still need to rerun make? Aug 12 09:22:23 what make? Aug 12 09:22:37 the kernel will be compiled, packaged, etc. Aug 12 09:22:45 what you get is a .ipk of the kernel Aug 12 09:23:01 if you like a complete new image Aug 12 09:23:09 you have to repeat bb openslug-image Aug 12 09:23:36 so you need to delete the stamps for that also, and than use upslug to flash the new image. Aug 12 09:24:15 but if you have a recent openslug on the slug, it is much more easy to use the reflash script after installing the newly created kernel package with ipkg Aug 12 09:39:16 oh I seeeeee Aug 12 09:39:30 so it's a case of installing the new kernel package with ipkg - that's nifty :) Aug 12 09:39:34 thanks hanjo :) Aug 12 09:43:37 yes, as long as the modules did not change, you just need to install the new kernel package, and then you need to reflash -k /boot/zImage Aug 12 09:47:44 ok, where is the reflash tool? Aug 12 09:48:02 ...and what happens if the modules do change? Aug 12 09:51:10 then you need to install the module packages also. reflash is part of the openslug-init package i think Aug 12 10:22:48 the perl package in OE ... does that install a full perl or just miniperl ? Aug 12 10:23:02 got a hardware question in nslu2-general, can someone have a look? Aug 12 10:29:07 SpanKY: full perl, no perl modules. Aug 12 10:29:22 thanks Aug 12 11:35:44 03jbowler 07org.openembedded.nslu2-linux * r3d3d7753... 10/packages/monotone/ (monotone-5_0.22.bb monotone.inc): Aug 12 11:35:44 monotone-5 0.22 - the LUA hook for anonymous_read must be updated (moved Aug 12 11:35:44 into the standard read hook). Aug 12 11:35:44 test 247 (lua function existsonpath) fails - apparently there is a problem Aug 12 11:35:44 with search the path for executables. This is the only remaining Aug 12 11:35:45 unexpected failure (all unexpected failures in previous versions except the Aug 12 11:35:47 netsync 'largish-file' test now function correctly on an NSLU2). Aug 12 11:35:49 03jbowler 07org.openembedded.nslu2-linux * r23548cd4... 10/ (16 files in 13 dirs): Aug 12 11:35:51 propagate from branch 'org.openembedded.dev' (head 167e15fd041bdf8e86fff3513e13934a284b3d06) Aug 12 11:35:53 to branch 'org.openembedded.nslu2-linux' (head 3d3d7753981cab1f26a692e87498137a0d06cdc9) Aug 12 11:40:35 03jbowler 07org.openembedded.nslu2-linux * r96b19529... 10/packages/meta/ucslugc-packages.bb: UcSlugC: add dev packages Aug 12 11:58:49 Has anyone done any work on supporting a USB WiFi NIC? Aug 12 11:59:19 Going to Fry's anyway, and I thought I'd pick up one to play with. Aug 12 11:59:35 If there are giants' shoulders to stand upon, I'm not too proud. :) Aug 12 12:00:19 Not yet but two of the developers have a ZD1211 based USB dongle which has open source drivers Aug 12 12:00:42 Tiersten: Mine hasn't arrived yet, so it's only one :( Aug 12 12:00:52 Ah, excellent. Aug 12 12:01:04 ZD1211 = Cheap Airlink stuff. ;) Aug 12 12:02:05 Thanks! Aug 12 12:05:46 Okay. 1 dev has it :) Aug 12 12:05:49 One more is pending Aug 12 12:05:55 yeah Aug 12 12:09:25 Can someone point me towards a good writeup on how do deal with lib dependencies in cross compiliation? Aug 12 12:09:38 I'm still beating my head against the wall trying to get libptp to configure. Aug 12 12:11:55 It builds just fine on its own, but I'm after ptpcam, which depends on libusb, and I don't seem to be able to tell configure where to find libusb (even though I've staged it) Aug 12 13:50:11 Just got an AirLink WiFi dongle, and it is indeed a ZyDAS 1211 Aug 12 14:01:44 I have a error on installing libusb into openslug 2.5 Aug 12 14:01:53 It can't find ldconfig! Aug 12 14:02:08 Should i log this as a bug? Aug 12 14:02:17 ls /sbin Aug 12 14:02:24 its in there. Aug 12 14:02:45 Yes it is! Aug 12 14:05:07 The cause was "sudo". Because i used sudo "/sbin" wasn't in my path. Aug 12 14:12:27 dyoung-zzzz, i'd like to take a stab at sound input this weekend. i have external SoundBlaster 24bit USB to try. i've unslung, and gather i'll need to move to openslug, but could use a pointer for how to proceed. Aug 12 14:39:33 anyone know a cheaper version of something similar to a slug? Aug 12 14:40:10 Cheaper than the slug? You can get a slug for like... $80 Aug 12 14:40:42 NAiL: $80 US? Aug 12 14:40:52 Think it was $80 Aug 12 14:40:56 hang on a sec Aug 12 14:41:40 oh, you're in .so Aug 12 14:41:41 .se Aug 12 14:41:45 yeah Aug 12 14:41:53 Then it's a bit more (well, way more) expensive Aug 12 14:42:03 http://www.newegg.com/Product/Product.asp?Item=N82E16833124036 <-- $79.99 at newegg Aug 12 14:42:08 but they won't ship over here Aug 12 14:42:13 (.no) Aug 12 14:42:26 i really wish the slug was in the standard linksys stackable formfactor as well Aug 12 14:44:09 where is .no? Aug 12 14:44:11 norway? Aug 12 14:44:22 jepp Aug 12 14:46:39 you wouldn't happen to know the cheapest way to get some usb ports :P Aug 12 14:47:45 some usb ports on what? a PC? Aug 12 14:48:47 i have a wrt54g currently Aug 12 14:49:00 i'm thinking i should have got the asus one with the usb port Aug 12 15:08:54 mithro: yep, you should have got the asus wl500gx :-) Aug 12 15:12:42 that is the delux model right? Aug 12 15:17:24 yep, make sure you get the deluxe model for the usb 2.0 ports Aug 12 15:19:04 i can't find them anywhere here :/ Aug 12 15:19:58 Appears to not be available in .no either **** BEGIN LOGGING AT Fri Aug 12 15:27:41 2005 Aug 12 15:28:18 http://komplett.no/k/ki.asp?sku=308518&cks=SER Aug 12 15:29:41 thats the deluxe one Aug 12 15:29:59 wonder if they are avaliable in sweden too... Aug 12 15:30:18 DaKa: you in Sweden? Aug 12 15:30:21 yep Aug 12 15:30:29 they are found in denmark Aug 12 15:30:45 http://www.edbpriser.dk/Listprices.asp?ID=134800# Aug 12 15:32:22 DaKa: where in Sweden are you? Aug 12 15:32:33 mithro: Stockholm Aug 12 15:33:46 hm.. the WL-500G deluxe is avaliable on komplett.se too Aug 12 15:35:11 1095 KR Aug 12 15:35:47 mithro: where are you? Aug 12 15:38:58 Stockholm Aug 12 15:39:09 in Lapis :P studying at KTH Aug 12 15:47:02 :) Aug 12 15:50:11 DaKa: where in stockholm are you located? Aug 12 15:50:35 right now at work on kungsholmen, thorildsplan Aug 12 15:51:01 (yes, I know its sad beeing at work at 00:55.. but :-) Aug 12 15:57:21 anyone know what these print servers run? Aug 12 15:57:40 make autobuild on openslug-native only took 8 hours (most of that was 140 monotone revisions in) Aug 12 15:57:47 mithro: model no? Aug 12 15:58:03 you could use the asus as a print server too. Aug 12 15:59:44 rwhitby: the print servers are really cheap, so if they did run a really cut back version of linux it shouldn't be hard to use them as a file server instead Aug 12 16:10:07 mithro: are you talking about the linksys print servers? Aug 12 16:12:13 any of the usb print servers :P Aug 12 16:12:42 need a model number, so I can look at the firmware for it. Aug 12 16:14:21 PSUS4-EU Aug 12 16:14:24 linksys Aug 12 16:14:41 Belkin F1UP0001 Aug 12 16:16:30 from the firmware image, it looks like the PSUS4-EU uses an ADM5120 processor Aug 12 16:16:56 and is also done by SerComm (like the NSLU2) Aug 12 16:17:50 which means? Aug 12 16:18:32 http://www.linux-mips.org/wiki/ADMtek Aug 12 16:18:33 it runs linux or ecos or something? Aug 12 16:19:43 4Mb flash, 32Mb ram, 175MHz, not that bad.. Aug 12 16:23:41 so far that kinda of makes it the cheapest option Aug 12 16:24:13 only usb 1.1 however Aug 12 16:24:56 what about the Belkin one? Aug 12 16:36:10 Hey, I was on earlier and speaking to Hanjo (I think) about compiling a custom kernel for openslug Aug 12 16:36:25 he was very helpful, but I need a couple more hints Aug 12 16:37:19 in order to use bitbake to recompile my custom kernel, can I just edit defconfig in the directory /slug/openembedded/packages/linux/openslug-kernel-2.6.12.2 and then use bitbake to make a kernel ipkg? Aug 12 16:38:35 I've found this: http://www.nslu2-linux.org/wiki/HowTo/TweakKernelOrPackagesInOE Aug 12 16:38:59 but I where does the bb command come from? by installing bitbake like this: http://www.nslu2-linux.org/wiki/OpenEmbedded/GettingStarted ? Aug 12 16:42:52 mithro: can't find any firmware for the belkin one Aug 12 16:44:43 mithro: apart from the nslu2, most things you'll find will not have much 2.6 kernel support. Aug 12 16:48:27 * mithro ponders Aug 12 16:52:36 bah, why does international shipping have to cost so much? Aug 12 18:18:53 so rwhitby, do you know where to get a cheap slug or wl-500g deluxe? Aug 12 18:20:01 get someone in US here to buy it and ship it. Aug 12 18:20:40 a slug or wl500gx fits in the USD$14 international fixed price small box. Aug 12 18:24:43 so rwhitby :P Aug 12 18:24:58 rwhitby isn't in the US Aug 12 18:25:03 dam Aug 12 18:25:49 n Aug 12 18:26:13 I am, but don't the import taxes negate the savings? Aug 12 18:39:41 what price can you get a wl500gx in the US? Aug 12 18:41:11 checking. I know the slug is $89US (72.61 EU) Aug 12 18:42:01 slug costs $79.99US at newegg Aug 12 18:42:16 NAiL: I gotta stop forgetting to check them Aug 12 18:42:35 You're looking at $93.99 inc shipping to EU then Aug 12 18:42:51 wl: http://www.newegg.com/Product/Product.asp?Item=N82E16839112005 Aug 12 18:43:07 A NSLU2 in the UK and most of Europe appears to be roughly $125 Aug 12 18:47:52 Tiersten: so you would save roughly $35? Aug 12 18:48:06 Yes Aug 12 18:48:33 Probably more like $40-$45 since that doesn't include shipping from the UK onwards Aug 12 18:48:51 Tiersten: ? Aug 12 18:49:16 $125 doesn't include the price of the extra shipping needed Aug 12 18:49:21 $93.99 does Aug 12 18:49:51 At a rough guess, a NSLU2 bought in the UK and then shipped to somewhere else would roughly cost $135 Aug 12 18:50:19 A total saving therefore of about $40 if you get it from the US and don't incur any import or customs taxes Aug 12 20:34:55 [12:36] rwhitby-away: # df Aug 12 20:34:55 [12:36] rwhitby-away: Filesystem 1k-blocks Used Available Use% Mounted on Aug 12 20:34:55 [12:36] rwhitby-away: "/dev/ram0 10647 10111 426 96% /" Aug 12 20:34:55 [12:36] rwhitby-away: "/dev/sdb1 248716 0 248716 0% /share/hdd/data/FLASH_1_1_1" Aug 12 20:34:57 [12:40] rwhitby-away: From web interface, with one flash key in each port: Aug 12 20:34:59 [12:40] rwhitby-away: USB Port 1: Ready (FAT/NTFS), 242MB (0% Free) Aug 12 20:35:01 [12:40] rwhitby-away: USB Port 2: Ready (FAT/NTFS), 117MB (100% Free) Aug 12 20:35:03 [12:41] rwhitby-away: (the % free on port1 is wrong) Aug 12 20:35:05 [12:41] rwhitby-away: and df: Aug 12 20:35:07 [12:41] rwhitby-away: "/dev/sdb1 248716 0 248716 0% /share/hdd/data/FLASH_1_1_1" Aug 12 20:35:10 [12:41] rwhitby-away: "/dev/sda1 120204 0 120204 0% /share/flash/data/FLASH_2_1_1" Aug 12 20:35:13 [12:41] rwhitby-away: I guess the _1_1_1 support multiple partitions Aug 12 20:35:15 [12:43] rwhitby-away: now to try a 1.5Gb NTFS cornice Aug 12 20:35:17 [12:45] rwhitby-away (n=rwhitby@nslu2-linux/rwhitby) is now known as rwhitby Aug 12 20:35:19 [12:49] rwhitby: Attached scsi removable disk sdc at scsi1, channel 0, id 0, lun 0 Aug 12 20:35:21 [12:49] rwhitby: SCSI device sdc: 2858688 512-byte hdwr sectors (1464 MB) Aug 12 20:35:23 [12:49] rwhitby: is removable disk Aug 12 20:35:25 [12:49] rwhitby: sdc: test WP failed, assume Write Enabled Aug 12 20:35:27 [12:49] rwhitby: sdc: sdc1 Aug 12 20:35:29 [12:49] rwhitby: >>>>>>>>>>>>>>>>>>>>Device /dev/sdc connected into PORT 1 Aug 12 20:35:31 [12:49] rwhitby: Try to mount FAT/NTFS volumes Aug 12 20:35:33 [12:49] rwhitby: >>>>>>>>>>>>>>>>>>>>mount cmd:/bin/mount -tufsd /dev/sdc1 /share/hdd/data/FLASH_1_2_1 -o umask=0007,fmask=0760,dmask=0770,uid=501,gid=501,iocharset=cp437 2>/dev/null Aug 12 20:35:36 [12:52] rwhitby: can definitely do NTFS R/W and serve it up through web interface and samba Aug 12 20:35:38 [12:53] rwhitby: "/dev/sdc1 on /share/hdd/data/FLASH_1_2_1 type ufsd (rw)" Aug 12 20:35:40 [12:55] rwhitby: ftp server support in the web interface Aug 12 20:35:42 [12:59] rwhitby: more than two disks supported Aug 12 20:35:44 [12:59] rwhitby: USB Port 1: Multiple Drives Ready (FAT/NTFS) Aug 12 20:35:46 [12:59] rwhitby: CORNICE Inc. Storage Ele : 1388MB (99% Free) Aug 12 20:35:48 [12:59] rwhitby: FLASH: 242MB (0% Free) Aug 12 20:35:50 [12:59] rwhitby: USB Port 2: Ready (FAT/NTFS), 117MB (100% Free) Aug 12 20:35:52 experiments with Linksys 2.3R63 Aug 12 20:35:54 Note the multiple disk and NTFS R/W support Aug 12 20:36:35 Three disks on one port on a hub, so far ... Aug 12 20:39:04 "/dev/sde1 on /share/hdd/data/HDD_1_4_1 type ufsd (rw)" Aug 12 20:44:53 USB Port 1: Multiple Drives Ready (FAT/NTFS) Aug 12 20:44:54 CORNICE Inc. Storage Ele : 1388MB (99% Free) Aug 12 20:44:54 UrDisk USB FLASH DISK : 117MB (100% Free) Aug 12 20:44:54 SAMSUNG SV4084D : 38923MB (99% Free) Aug 12 20:44:54 FLASH: 242MB (0% Free) Aug 12 20:44:55 USB Port 2: Not Installed Aug 12 20:45:10 looks like the free space on the last disk on Port 1 is always wrong Aug 12 20:47:52 # ls /proc/hdd* Aug 12 20:47:52 "/proc/hdd_sde" Aug 12 20:47:52 # ls /proc/flash* Aug 12 20:47:52 "/proc/flash_sda /proc/flash_sdb /proc/flash_sdc" Aug 12 20:51:16 handles hot-plug and hot-disconnect of FAT and NTFS drives Aug 12 20:52:52 NTFS hard disk accepted on Port2, but not NTFS cornice Aug 12 20:53:11 NTFS cornice accepted on Port1 Aug 12 20:53:47 (Cornice is detected as FLASH, as per stock linksys) Aug 12 20:55:32 Hmm - maybe NTFS hard disk is not accepted on Port 2 ... Aug 12 21:00:15 OK, NTFS is not accepted at all on Port 2 Aug 12 21:02:54 NTFS support is only on Port 1 Aug 12 21:03:48 that sounds like a bug... Aug 12 21:08:08 no, I expect that might be intentional Aug 12 21:12:23 I would have thought the two storage devices was intentional too, but that limitation is removed in R63? Aug 12 21:12:45 Where did you read that? Aug 12 21:13:27 [22:31] Three disks on one port on a hub, so far ... Aug 12 21:13:30 scroll up Aug 12 21:13:39 yep, you seem to only be limited by /dev/sdXNN devices (which is 5 drives with 15 partitions each) Aug 12 21:14:08 Great.... They've fiddled with the USB code yet again... Aug 12 21:14:11 like I said... if more than 1 NTFS can be seen on port 1, then it sounds like a bug.... Aug 12 21:14:43 looks like more than 1 NTFS on Port 1, but no NTFS on Port 2. Aug 13 01:02:08 jp30-work:ping Aug 13 01:06:01 rwhitby: how do I know when appweb gets moved from TESTING into the main feed ? Aug 13 01:07:37 copperbeech: you'll see it in the CVS Makefile, and then it will automatically be in the feed about 15 minutes or so after that. Aug 13 01:09:12 ok who is responsible for moving it up into the main feed? Should I be doing that or is it jp30s perogative? Aug 13 01:12:14 in the makefile? or the actual feed? Aug 13 01:15:45 in the unslung/Makefile ? Aug 13 01:15:59 jp30 is the package manager and will take cae of it. Aug 13 01:16:01 care. Aug 13 01:16:09 Is it an emergency? Aug 13 01:18:12 :-) great - nope, but I was going to write a mail to the list when it came onto the feeds... Aug 13 01:18:22 hehe Aug 13 01:18:34 Patience. :-) Aug 13 01:23:58 he he :-) Aug 13 08:06:49 exit Aug 13 08:06:50 exit Aug 13 08:06:51 exit Aug 13 08:43:02 Daka: ping Aug 13 08:49:40 think I found the problem - the permissions of the / (i.e. root) directory were set to 340 for some unknown reason Aug 13 08:54:48 zin: btw, updated your dropbear to openssh guide: http://www.nslu2-linux.org/wiki/HowTo/SwapFromDropbearToOpenSSH Aug 13 09:03:45 btw, if anyone tries doing make build-openslug-2.5-beta, I think they will find it pretty badly broken Aug 13 09:05:29 blaster8: oh? Why? Aug 13 09:05:46 it just fails really quickly Aug 13 09:05:53 what error? Aug 13 09:06:06 you probably have to do make setup-openslug-2.5-beta first Aug 13 09:06:14 probably, yes Aug 13 09:07:08 make[1]: Entering directory `/home/slugdev/slug/releases/OpenSlug-2.5-beta' Aug 13 09:07:09 make[1]: Makefile: No such file or directory Aug 13 09:07:14 seems to be a favourite Aug 13 09:07:23 hehe Aug 13 09:07:24 yeah Aug 13 09:07:32 then you'll need to do setup first Aug 13 09:07:50 The 2.5-beta targets aren't exactly set up for common users :P Aug 13 09:08:07 I like the way, when gentoo sees links where there is no file attached (if you see what I mean) it puts them in flashing white on red background Aug 13 09:08:13 was wondering what that was about Aug 13 09:08:52 you can probably copy the dircolors file from gentoo and put it on the slug Aug 13 09:09:25 But you'll probably need updated coreutils/util-linux/bash for it to work thataway Aug 13 09:10:26 I've got bash running nicely on unslung Aug 13 09:10:33 but Openslug is the way forward Aug 13 09:35:03 hy Aug 13 09:40:48 heya Aug 13 09:50:40 blaster8: ah! I knew it! Aug 13 09:50:52 NAiL: Still no luck with Aug 13 09:51:07 Openslug-2.5 beta build Aug 13 09:51:40 DaKa: worryingly this was a default gentoo install, so someone else may have this problem Aug 13 09:52:19 hm. if that really is so.. that's really really strange Aug 13 09:52:39 because that is just wrong.. Aug 13 09:53:08 indeed so (2005.1 too) Aug 13 09:55:55 that means any perl script using fastcwd() can fail.. Aug 13 09:56:27 the boost build (using jam, remember) had the same problem Aug 13 10:07:35 i'm pretty sure that it belongs to nslu2-general but isn't upslug supposed to quit when the firmware upgrade is over (ie when the slug restarts) ? Aug 13 10:08:42 MooZ: yes Aug 13 10:09:09 What's the last line you get from upslug? Aug 13 10:10:35 Writing Flash...0123456789abcdef0123456789abc Aug 13 10:10:46 it hangs on c? Aug 13 10:10:54 it quits Aug 13 10:11:03 quits? Aug 13 10:11:05 hmm Aug 13 10:11:08 that's weird Aug 13 10:11:19 and the slug remains in maintenance mode Aug 13 10:11:37 did you try again? Aug 13 10:12:01 * NAiL notes that the reset button now powers off the slug when it's hung :D Aug 13 10:12:06 (on OpenSlug) Aug 13 10:12:35 eek Aug 13 10:13:10 MooZ: what? Aug 13 10:13:24 http://rafb.net/paste/results/ccrfsN99.html Aug 13 10:13:47 hmm Aug 13 10:14:02 Should get jbowler to look at that Aug 13 10:14:40 No idea what that means Aug 13 10:27:08 ok Aug 13 10:30:12 bbl Aug 13 10:58:12 no joe package for openslug? Aug 13 10:58:39 hmm? Aug 13 10:58:42 there should be Aug 13 10:58:57 I am pretty sure I put that in the feed Aug 13 10:59:01 2 sec Aug 13 10:59:17 stable or unstable? Aug 13 11:00:24 It's in the stable feed, lemme put it in unstable as well Aug 13 11:00:42 I think I'm running stable (2.5-beta) Aug 13 11:00:52 _openslug_, not unslung Aug 13 11:00:56 ;) Aug 13 11:01:20 hmm Aug 13 11:01:27 how'd that disappear from the feed then? Aug 13 11:01:34 I've got it installed on my stable slug Aug 13 11:01:38 2.5-beta Aug 13 11:01:43 weird Aug 13 11:02:07 # ipkg install joe Aug 13 11:02:07 Nothing to be done Aug 13 11:02:07 An error ocurred, return value: 4. Aug 13 11:02:07 Collected errors: Aug 13 11:02:07 Cannot find package joe. Aug 13 11:02:09 I'm on the case ;) Aug 13 11:02:13 that's what I get Aug 13 11:02:20 root@maimedslug:/home# ipkg install joe Aug 13 11:02:20 Package joe (3.1-r0) installed in root is up to date. Aug 13 11:02:20 Nothing to be done Aug 13 11:02:21 k, thx :) Aug 13 11:05:48 do you guys know of anyone trying to port netatalk to NSLU2? Aug 13 11:05:53 nope Aug 13 11:06:09 it's just that NFS is actually quite bad Aug 13 11:06:41 it really isn't designed as a simple file storage system - so no reliable read/write soft mounts or anything I actually need Aug 13 11:07:54 Matta: It'll be in the feed as soon as I clean up my repo Aug 13 11:08:47 hmm Aug 13 11:08:53 Where's CIA? Aug 13 11:09:26 check to see if #commits is working maybe Aug 13 11:09:54 no cia there either. Guess it's down Aug 13 11:10:58 k, thx NAiL Aug 13 11:11:35 np Aug 13 11:11:45 should've been there in the first place, dunno why it's gone Aug 13 11:12:22 give it some... 10-15 minutes and try again Aug 13 11:15:14 k Aug 13 11:17:06 It's building a few packages, and it'll push to the feed afterwards Aug 13 11:17:11 at coreutils right now Aug 13 11:19:04 util-linux Aug 13 11:22:25 pushing stuff now Aug 13 11:22:36 where should I specify swap space? /etc/fstab doesn't seem to work... and is it recommended to run a ext2 or 3 fs? default is apparently ext2 Aug 13 11:23:38 ext3 is better on disks Aug 13 11:23:48 since it has a journal Aug 13 11:24:12 specify it in fstab, that should work nicely... Aug 13 11:24:25 /dev/ swap swap defaults 0 0 Aug 13 11:29:16 k, swap is apparently on, top doesn't show it though Aug 13 11:29:36 free should show it Aug 13 11:29:40 what top do you use? Aug 13 11:29:47 coreutils has a better top ;) Aug 13 11:29:58 ipkg update && ipkg install joe Aug 13 11:30:29 Installing joe (3.1-r0) to root... Aug 13 11:30:32 thx! ;) Aug 13 11:30:35 np Aug 13 11:43:08 you might want to install coreutils as well Aug 13 11:47:51 yup, did that Aug 13 11:47:58 EXT2-fs warning (device sda1): ext2_fill_super: mounting ext3 filesystem as ext2 Aug 13 11:48:23 for some reason it mounts my drive as ext2 even though I have ext3 specified in /etc/fstab Aug 13 11:59:51 yeah Aug 13 11:59:57 edit /initrd/linuxrc Aug 13 12:00:09 you have to remount initrd rw first though Aug 13 12:21:51 ah, thx Aug 13 13:39:47 rwhitby: can you link a wiki page for me? Aug 13 13:40:44 oh - asleep :-) anyone else here have wiki admin priviledges? Aug 13 13:43:34 yeah Aug 13 13:43:48 what's the url and where to link it to? Aug 13 13:46:01 http://www.nslu2-linux.org/wiki/Info/ComparingWebServers Aug 13 13:46:13 Probably needs linkng off the info page ? Aug 13 13:47:34 <[g2]> copperbeech, page looks nice Aug 13 13:48:25 thanks... don't really know enough about thttpd to do it justice rally, but when I did play around with it it was a but limited for stuff like SSL and virtual sites etc. Aug 13 13:49:09 <[g2]> do you use AppWeb much ? Aug 13 13:49:38 well I just built the latest version integrated with PHP Aug 13 13:49:44 it's in TESTING now. Aug 13 13:50:15 yes I use it as my webserver - the developers are really responsive, and I think it is simple to configure Aug 13 13:50:20 <[g2]> I'm guessing you are running on Unslung Aug 13 13:50:22 it's neat Aug 13 13:50:39 yep - but it should compile up pretty easily on openslug Aug 13 13:50:47 <[g2]> isn't it perl based ? Aug 13 13:50:54 nope Aug 13 13:51:00 <[g2]> php? Aug 13 13:51:03 copperbeech, done Aug 13 13:51:12 it's an embedded webserver - compiles from C, C++ Aug 13 13:51:21 <[g2]> hmmm Aug 13 13:51:43 it eve has some utilities to compile web pages into code modules for fully compiled embedded systems... Aug 13 13:51:58 the link is down under "Other" on the Information page Aug 13 13:52:10 <[g2]> I'm really interested in the memory/performance footprint numbers Aug 13 13:52:26 I think it's a great blend of the efficiencies of thttpd vs the functionality of apache. Aug 13 13:53:20 are there any commands I can run from BASH to give you those figures? Aug 13 13:53:50 <[g2]> cat /proc/cpu/meminfo helps Aug 13 13:54:06 <[g2]> there are some other tools depending on which firmware version you are running Aug 13 13:55:15 hmmm dont have a /proc/cpu/meminfo Aug 13 13:58:29 [g2] I spooled some infor in the appweb channel for you :-) Aug 13 13:58:33 ByronT: Thanks Aug 13 13:58:34 <[g2]> nod Aug 13 13:58:48 <[g2]> it looks like you are using nearly 64MB memory Aug 13 13:59:01 <[g2]> the mem+swap used Aug 13 13:59:06 [g2] yeah, but I think that may be more to do with the media server Aug 13 13:59:32 <[g2]> you can shut it down an see how it changes :) Aug 13 13:59:40 I was wondering if there is a way of getting the resource usage for each processes Aug 13 13:59:51 yeah - ok I'll try that Aug 13 13:59:52 <[g2]> lsof Aug 13 14:00:16 <[g2]> I think lsof -p (pid number) and there's a memory option Aug 13 14:00:33 <[g2]> I've run across some memory tools like that Aug 13 14:00:47 <[g2]> I haven't put them together or written up things yet Aug 13 14:00:58 <[g2]> I'll be doing more of that real soon now :) Aug 13 14:04:24 <[g2]> on a quick glance that looks like 52MB vers 48MB for no AppWeb Aug 13 14:05:04 The documentation says that the memory footprint is 110k - whatever that means http://www.appwebserver.org/products/appWeb/embedded-web-server.html Aug 13 14:05:36 <[g2]> that's probably 110K before you serve up your first web page :) Aug 13 14:05:52 "Over 3500 requests per second on a PC class device. Memory footprint from 110K. Code and web pages are fully ROMable." Aug 13 14:06:39 yeah - these things probably cache requests etc... I had 4 threads running on my box - not sure if they were spawned or just not cleaned up from my testing... Aug 13 14:07:13 <[g2]> that's why I like to "Trust -- be verify" Aug 13 14:07:16 <[g2]> be/but Aug 13 14:08:52 yes - we should probably run the tests from clean to be fair... there should be an appweb package on the feeds soon to do some comparative tests. Aug 13 14:09:21 the make file is in the CVS if you want build it yourself... Aug 13 14:09:25 <[g2]> you probably want to install lsof if you don't already have it installed Aug 13 14:09:34 yes Aug 13 14:09:49 <[g2]> I think the +M dumps memory and file info Aug 13 14:10:08 [g2] if you get any stats we can record them on the wiki page http://www.nslu2-linux.org/wiki/Info/ComparingWebServers Aug 13 14:10:34 <[g2]> will do. Aug 13 14:10:59 <[g2]> I'll also be launching a couple of my one sites soon also Aug 13 14:11:08 <[g2]> heh... my own Aug 13 14:15:53 [g2] seems to me that much of the memory footprint of appweb is taken by the php libraries Aug 13 14:16:34 <[g2]> nod... looking now Aug 13 14:16:45 <[g2]> I think their .so is 300K Aug 13 14:16:49 <[g2]> plus some other stuff Aug 13 14:25:45 apache is working really well on the slug.. Aug 13 14:26:20 <[g2]> DaKa, great Aug 13 14:26:29 <[g2]> we can compare and contrast Aug 13 14:27:00 <[g2]> I had 2.0.54 and php5 running with openssl and it seemed pretty zippy on the FatSlug Aug 13 14:27:20 <[g2]> I'm planning on serving up my web site on one of my units Aug 13 14:27:22 60% cpu idle with 25 concurrent siege keep-alive users on the default 1,4Kb html Aug 13 14:27:34 should probably test siege on something bigger Aug 13 14:27:43 and on php.. Aug 13 14:28:02 <[g2]> yeah I think it'll take some testing to figure out Aug 13 14:28:17 <[g2]> the whole apache/php/mysql/perl/.... Aug 13 14:28:21 <[g2]> etc... Aug 13 14:28:27 yeah... Aug 13 14:29:20 [DaKa] If we can do some tests we should record the results in the wiki Aug 13 14:29:49 with 50 users I'm down to 15% free cpu.. but it runs along really smooth Aug 13 14:29:53 [g2] if you come up with a good way of testing these we should agree on some sort of standards to provide a basis for the comparison. Aug 13 14:30:05 <[g2]> nod. Aug 13 14:30:21 should probably make some plans on how to test things.. Aug 13 14:30:23 <[g2]> One of the web site's I'll be launching will focus on such things Aug 13 14:30:41 :-) Aug 13 14:30:45 <[g2]> We do a lot of great things on the wiki, but I'll be looking to provide much more info Aug 13 14:31:06 yeah - the wiki is not the be all and end all Aug 13 14:31:20 ... almost Aug 13 14:31:26 <[g2]> it's great for coordinating and posting lots of stuff Aug 13 14:33:02 hm.. 100 keep-alive users was not that good.. it didn't miss any hits, but some took 20 secs Aug 13 14:33:32 most were under 0,1 secs Aug 13 14:44:34 <[g2]> DaKa, is that a fat turbo ? Aug 13 14:44:40 nope Aug 13 14:44:43 just turbo Aug 13 14:45:12 Transaction rate: 96.99 trans/sec Aug 13 14:45:28 for a 34KB text file Aug 13 14:45:51 and that uses ~70% cpu Aug 13 14:47:12 so, if you make a small static page you should survive a slashdotting easily Aug 13 14:47:28 <[g2]> excellent Aug 13 14:47:43 Transaction rate: 94.31 trans/sec Aug 13 14:47:44 Throughput: 3285636.51 bytes/sec Aug 13 14:48:24 <[g2]> that's 24Mbs/sec Aug 13 14:48:43 yep.. for a single small text-file Aug 13 14:48:55 should test with a 500KB random garbage file.. Aug 13 14:52:29 ow.. 100% cpu and a load of 20+.. Aug 13 14:52:33 Transaction rate: 22.21 trans/sec Aug 13 14:52:33 Throughput: 11370109.09 bytes/sec Aug 13 14:52:37 but.. Aug 13 14:52:39 thats... Aug 13 14:52:47 <[g2]> near line rate Aug 13 14:52:50 yep Aug 13 14:52:54 <[g2]> 88Mbs Aug 13 14:52:59 <[g2]> + Aug 13 14:53:17 <[g2]> I think the trick is to limit the incoming requests with iptables Aug 13 14:53:24 <[g2]> and the tc Aug 13 14:53:24 but the slug was really sluggish, and some transactions took lots of time Aug 13 14:53:36 but apache is not that bad Aug 13 14:53:44 <[g2]> you are probably memory limited Aug 13 14:53:51 <[g2]> and you start swapping Aug 13 14:53:55 lets see... Aug 13 14:53:57 swapoff -a Aug 13 14:54:07 * [g2] hugs DaKa Aug 13 14:54:17 [DaKa] would be interesting to see the equivalent for thttpd and appweb ? Aug 13 14:54:19 22 transactions per second? Aug 13 14:54:29 what are you using for these tests? Aug 13 14:54:40 on a 500KB file Aug 13 14:54:43 <[g2]> copperbeech, agreed Aug 13 14:54:49 I'm using siege Aug 13 14:54:52 * [g2] hugs dyoung Aug 13 14:54:53 siege? Aug 13 14:54:55 <[g2]> hey Aug 13 14:54:57 a really nice http test program Aug 13 14:55:02 ok need to find that/ Aug 13 14:55:06 <[g2]> url ? Aug 13 14:55:08 apt-get install siege Aug 13 14:55:10 :) Aug 13 14:55:34 mornign [g2] Aug 13 14:55:42 <[g2]> aloha Aug 13 14:55:47 Im getting the same values with swap off, just some requests takes longer Aug 13 14:56:04 should probably lower sieges simulated number of users Aug 13 14:56:43 <[g2]> http://www.joedog.org/siege/ Aug 13 14:57:26 <[g2]> duress.... Aug 13 14:57:59 <[g2]> "measure the performance of their code under duress" http://www.joedog.org/ Aug 13 14:58:10 <[g2]> yummy Aug 13 14:58:36 with 15 keep-alive users on that file I have 20% cpu free, and things gets serverd in ~0,1 secs Aug 13 14:58:58 Transaction rate: 20.86 trans/sec Aug 13 14:58:58 Throughput: 10678034.73 bytes/sec Aug 13 14:59:27 "that file" is 'dd if=/dev/urandom of=test count=500 bs=1024' Aug 13 14:59:28 siege segfults for me. Aug 13 14:59:42 right, the config-file gets empty Aug 13 14:59:53 oh. do I need to twiddle it? Aug 13 14:59:55 heh Aug 13 14:59:57 * dyoung lazy Aug 13 15:00:08 something is wrong with siege.config Aug 13 15:00:15 <[g2]> dyoung, testing the robustness of the distro :) Aug 13 15:00:29 <[g2]> stuff shouldn't segfault Aug 13 15:00:45 <[g2]> uh-oh.... dinner time Aug 13 15:00:53 <[g2]> 2 minute warning Aug 13 15:00:57 food good. Aug 13 15:00:58 [g2]: so when's the site going up ? Aug 13 15:01:05 <[g2]> rwhitby, hey... Aug 13 15:01:08 <[g2]> probably this week Aug 13 15:01:11 and do you have the prototypes yet? Aug 13 15:01:30 <[g2]> the T1 line was installed and we should "turnup" this week Aug 13 15:01:41 dyoung: if you havn't fixed it already: wget http://david.thg.se/saker/.siegerc Aug 13 15:01:54 <[g2]> rwhitby, should be this weeks Aug 13 15:01:55 <[g2]> rwhitby, should be this week Aug 13 15:02:06 <[g2]> I'll get an update Monday or Tuesday Aug 13 15:02:41 <[g2]> I've started working on the web sites Aug 13 15:03:06 <[g2]-dinner> bbl Aug 13 15:03:19 DaKa: I'm going to try mini_httpd combined with PHP asap - It's extremely small, supports virtual hosting and has been security audited (security patchset over at http://m0n0.ch/wall/docbook-current/image-guide-packages.html#id2615985) Aug 13 15:03:49 First get it working on Openslug then move over to UcSlugC Aug 13 15:05:07 I dunno, but I think siege just broke thttpd Aug 13 15:05:13 hm.. looks good.. Aug 13 15:05:16 :-) Aug 13 15:05:56 dyoung: heh.. Aug 13 15:06:09 heh... Aug 13 15:06:18 heh indeed Aug 13 15:06:26 I think siege broke his internet Aug 13 15:07:00 dyoung: it broke your internet too? :-) Aug 13 15:07:06 hehe yp., it broke Aug 13 15:07:18 actually it did. Aug 13 15:07:27 someone did manage to crash the internet once Aug 13 15:07:29 it broke the slug on which the proxy is running. Aug 13 15:07:50 might want to lower the number of users and use connection = close Aug 13 15:08:15 dunno how well thttpd handles keep-alive Aug 13 15:08:42 probably quite well Aug 13 15:08:53 ppl use it for download/file servers mostly Aug 13 15:09:18 I'd just run vsftpd Aug 13 15:13:21 hm.. apache seems to use 2,8MB ram after some abuse Aug 13 15:13:36 that's good Aug 13 15:13:44 I had to back set it to 5 transactions. Aug 13 15:14:01 I get a whopping 0.33 transactions/s Aug 13 15:14:22 initially 1,7MB memory Aug 13 15:15:02 after the first 10 requests it's at 2,7MB Aug 13 15:15:07 so.. apache is not that bad Aug 13 15:15:12 but.. have to test with php Aug 13 15:15:38 dyoung: on what file are you testing that? Aug 13 15:15:45 rwhitby:ping Aug 13 15:15:54 pong Aug 13 15:16:19 thats thttpd with a 500k file "dd if=/dev/zero of=500k bs=1024 count=512" Aug 13 15:16:33 uh.. that's like.. really bad.. Aug 13 15:16:40 really bad Aug 13 15:16:41 can't be right Aug 13 15:17:26 rwhitby: did we agree whether appweb package should be plain appweb or php-appweb ? Aug 13 15:18:35 plain Aug 13 15:18:51 was jp30's call IIRC Aug 13 15:19:22 I'm probably not licking it right. Aug 13 15:20:16 ok - last thing I did was to try and remove the php-appweb and appweb-php.mk files from the cvs and left a single appweb-mk file, but when I do ipkg update I see that we have a php-appweb package listed and no appweb.... Aug 13 15:20:27 Do you think I made a mistake somewhere ? Aug 13 15:24:13 well i got a slug today Aug 13 15:24:20 copperbeech: depends what the top Makefile says Aug 13 15:24:52 I'll have to ask jp30 to have a look at it. Aug 13 15:24:55 CROSS_PACKAGES determines what is in the cross feed Aug 13 15:24:56 what should be the first thing i do :P Aug 13 15:25:04 mithro: unslung or openslug? Aug 13 15:25:12 oh, remove R83 first. Aug 13 15:25:20 and put serial on it Aug 13 15:25:43 (since you're in #oe, I assume you're gonna be hacking the kernel on it, so you'll want serial) Aug 13 15:26:04 mithro: Put on 2.3R29 Aug 13 15:26:16 mithro: format the disk as ext3 Aug 13 15:26:18 why R29 ? Aug 13 15:26:37 well, my drive didn't format in R63 Aug 13 15:27:01 yeah, since mithro is asking here, then we can assume he's a developer, and would go straight to unslung or openslug ... Aug 13 15:27:02 Oh, I was htinking more why not R25. Aug 13 15:27:03 heh Aug 13 15:27:03 i'm thinking openslug Aug 13 15:27:11 openslug is your friend. Aug 13 15:27:14 since i have done a lot of OE work already :P Aug 13 15:27:29 unslug is still based on the original firmware right? Aug 13 15:27:31 mithro: you want to build you're own image? Aug 13 15:27:32 mithro: openslug Aug 13 15:27:54 rwhitby: maybe, most probably just my own packages Aug 13 15:28:27 mithro: I'd suggest installing the openslug 2.5 beta binary, and setting up your build machine with the master makefile. Aug 13 15:28:52 but bear in mind that the make openslug-2.5-beta target is a little flaky atm Aug 13 15:29:01 then rebuild openslug latest and reflash -i that image - then you will be able to contribute kernel defconfig changes too Aug 13 15:29:37 Hey, is [g2] stlil around? Aug 13 15:30:15 <[g2]-dinner> dyoung, back Aug 13 15:30:36 [g2]: do you s till have 1400 revs in? if so, I have a solution for you. Aug 13 15:31:06 (if he doesn't have local loft changes) Aug 13 15:31:13 <[g2]> I'm at 784 of 1219 Aug 13 15:31:23 <[g2]> for the latest update Aug 13 15:31:37 <[g2]> the pull that started a couple hours ago Aug 13 15:31:54 [g2]: are you using the master makefile? Aug 13 15:32:12 <[g2]> I did a make update with an old makefile Aug 13 15:33:16 if you don't have local changes, then starting from scratch with the master makefile will be *much* faster, as it downloads a prebuilt monotone db Aug 13 15:33:34 make sure you save your keys Aug 13 15:34:24 openslug is OE right? Aug 13 15:34:38 yes, totally Aug 13 15:34:40 <[g2]> mithro it's built with OE Aug 13 15:34:46 as is UcSlugC Aug 13 15:34:47 cool Aug 13 15:35:11 just checking incase we had gone of on the wrong tangent Aug 13 15:35:21 i guess i should first test if the slug actually works Aug 13 15:35:49 very good start Aug 13 15:38:05 uh.. using keep-alive against thttpd was really really bad.. Aug 13 15:39:10 hehe.. at 160 transactions/sec syslogd uses 17% cpu Aug 13 15:39:27 can that be turned off?? Aug 13 15:39:54 Ohyeah, I discovered that. Aug 13 15:40:22 that got me to only 7 trans/s Aug 13 15:40:24 is there a config-file for thttpd somehwere? Aug 13 15:40:39 or how does this work? Aug 13 15:40:42 nail said "no, but I should probablyu put ome someplace" Aug 13 15:40:52 without the typos. Aug 13 15:40:56 ok.. Aug 13 15:41:12 so it just uses some default right now? Aug 13 15:41:20 dyoung, i have unslung done, but i assume that to get audio input (for frog logger) i'll need to set up my own openslug build environment, or no? Aug 13 15:41:50 geojeff: It could be made to work in unslung maybe too. Aug 13 15:42:17 I think the only thing holding you back in libao which allegedly has runtime issues. I say allegedly, beucase I have no way to test that. Aug 13 15:43:13 I never found it in the ipkg feeds (back when I was playing with audio on the slug) Aug 13 15:43:19 libao that is Aug 13 15:43:58 i will try using an external USB soundblaster. Aug 13 15:44:24 (it was $50) Aug 13 15:45:15 got a sample thttpd.conf Aug 13 15:46:36 -l /dev/null will kill logging Aug 13 15:46:38 ByronT, its not in the feed because its commented in the top-level make file as "having runtime issues" Aug 13 15:47:03 its unclear to me if it works any better in openslug. Aug 13 15:47:13 I have a imic in my Amazon Cart; but I didnt check out yet. Aug 13 15:47:26 or logfile=/dev/null in /etc/thttpd.conf (see here for details http://www.die.net/doc/linux/man/man8/thttpd.8.html} Aug 13 15:49:26 dyoung, iMic or other solution might be cheaper for me, but soundblaster is something for me to try with. i'll get up to speed on doing basic unslung ipkg updates tomorrow. Aug 13 15:50:35 If you do decide to go the unslung route; be aware that significant patching to libao will probably be required. Aug 13 15:50:50 If thats the case, you might be able to steal some patches from OE. Aug 13 15:51:01 "borrow" rather. Aug 13 15:51:23 ok, looking at libao now Aug 13 15:51:30 <[g2]> "reuse" Aug 13 16:01:15 why does linksys say to put certain things in certain usb ports? Aug 13 16:02:03 hm.. I can make apache work as good as thttpd on small static files, but then apache works as bad as thttpd on keep-alive connections.. Aug 13 16:02:15 and whats with the thing you clamp on the cable? Aug 13 16:03:14 <[g2]> EMI supression Aug 13 16:03:32 why does it need EMI supression? Aug 13 16:04:14 <[g2]> well chips are ethernets are often noisy (EMI wise) Aug 13 16:05:26 mithro: stock linksys firmware has funky scsi assignments and stuff in the kernel. ignore it for openslug. Aug 13 16:06:39 The ferrite ring is so it complies with the standards. Chances are that you'll be fine without it anyway Aug 13 16:07:54 now the slug is connected if should pick up a DHCP address right? Aug 13 16:08:19 no, it starts out being 192.168.1.77 Aug 13 16:08:57 Server Name: LKG7F3A15 Aug 13 16:08:58 Version: V2.3R24 Aug 13 16:09:58 mithro: if you ever loose connectivity with the slug while on the default firmware, it may have fallen to its static address, so check it out before starting to worry Aug 13 16:13:07 okay i gave it a better address now Aug 13 16:13:15 so how to i get openslug onto this thing :P Aug 13 16:13:45 openslug supports usb hubs? Aug 13 16:13:45 from linux or windows? Aug 13 16:13:49 linux Aug 13 16:14:08 mithro: openslug is vanilla 2.6.latestish kernel with OE rootfs. Aug 13 16:14:45 downloading 2.5 beta now Aug 13 16:16:31 downloaded reading instructions Aug 13 16:23:08 upslugging Aug 13 16:25:06 oh well, should get some sleep, native apache bb is working great, and just needs a bit polishing, might get some time tomorrow Aug 13 16:25:12 bbl Aug 13 16:26:50 okay just ran upslug Aug 13 16:26:58 and the slug just beeped at me Aug 13 16:27:27 <[g2]> DaKa, sweet dreams Aug 13 16:34:11 <[g2]> dyoung, running siege ? Aug 13 16:34:21 Yeah, I was. Aug 13 16:34:28 <[g2]> cool Aug 13 16:34:35 <[g2]> I'm reading the website now Aug 13 16:34:48 <[g2]> can you pastebin some output ? Aug 13 16:34:57 sure what do you want of? Aug 13 16:35:00 httpd? Aug 13 16:35:03 h Aug 13 16:35:03 t Aug 13 16:35:08 thttpd Aug 13 16:35:14 <[g2]> sure Aug 13 16:35:30 <[g2]> is this on a turbo/fat/stock unit ? Aug 13 16:35:40 I had to turn off keep-alve, so its "close" now Aug 13 16:35:46 turbo only Aug 13 16:35:59 I lost the Fat arms race. Aug 13 16:36:13 <[g2]> to DaKa or ka6sox ? Aug 13 16:36:15 EXT2-fs warning (device sda2): ext2_fill_super: mounting ext3 filesystem as ext2 Aug 13 16:36:15 ? Aug 13 16:36:30 Transaction rate: 8.52 trans/sec Aug 13 16:36:30 Throughput: 4466040.53 bytes/sec Aug 13 16:36:40 to Daka. Aug 13 16:37:03 <[g2]> yeah... He's running APEX and trying 128 Aug 13 16:37:05 modprobe ext3 ? Aug 13 16:37:35 landslide already has a 128 bloatSlug Aug 13 16:37:54 just showing off IMO Aug 13 16:38:08 <[g2]> scalabilty is nice though Aug 13 16:38:18 mithro: you forget the -t ext3 options to turnup disk Aug 13 16:38:24 <[g2]> I think 64MB may be the sweet spot for a while Aug 13 16:38:36 do i actually need an external disk? Aug 13 16:38:49 mithro: basically yes, unless you run UcSLugC Aug 13 16:38:59 gtg Aug 13 16:39:07 <[g2]> blaster8, cheers Aug 13 16:39:28 hopefully this build will finally complete overnight Aug 13 16:39:30 cya Aug 13 16:42:08 how big is the flash in the slug? Aug 13 16:43:33 mithro: wiki/Info Aug 13 16:44:19 it's just my external USB HD is rather noise Aug 13 16:44:32 so i would love my slug to automatically disconnected if after a while Aug 13 16:45:39 mithro: if you can solve that problem (without using an enclosure which does spin-down on it's own, like the Maxtor OneTouch), then you will be worshipped by many nslu2 users. Aug 13 16:46:22 it would require unmount the drive and using usb to disconnect it :P Aug 13 16:47:00 btw i'm assuming you have tried using the standard scsi spindown commands? Aug 13 16:47:48 mithro: Yes. Aug 13 16:51:07 <[g2]> rwhitby, btw the Maxtor OneTouch has a usb chip in it and I think it just uses some extra commands Aug 13 16:51:37 you can disconnect power from the drive in software Aug 13 16:51:40 g2: Yes. It's nonstandard Aug 13 16:52:19 as I said, if anyone solves it they will be praised .... Aug 13 16:55:41 rwhitby: well it can be kind of solved, but it depends by what you mean solved Aug 13 16:57:00 all you need is a few small changes to an automounter Aug 13 16:57:44 mithro: are you talking about spinning disks down when they are idle for some time? Aug 13 16:58:21 yes Aug 13 16:58:29 i can think of one issue... Aug 13 16:58:33 it would be a bit more violent then just spinning them down Aug 13 16:58:41 really, more violent? Aug 13 16:58:49 you would umount them and then kill the power to the device Aug 13 16:58:52 ah Aug 13 16:59:13 but what if a cron job or something runs, and needs to write to /var/log/messages Aug 13 16:59:13 so that means no open files etc Aug 13 16:59:20 then you'd have to crank the hd up again Aug 13 16:59:26 all for 1 measly write Aug 13 17:00:21 The delay would be considerable Aug 13 17:00:36 delay? Aug 13 17:00:40 Yes Aug 13 17:00:49 it can't be good for the disk either Aug 13 17:00:56 turning on and off every hour for instance Aug 13 17:00:58 The average HD takes a few seconds from power up before you can start talking to it Aug 13 17:01:20 the best situation in many ways would be root fs on flash, and external hard disk acting purely as storage Aug 13 17:01:52 best solution would be for spindown on idle to work :) Aug 13 17:02:07 yeah i agree Aug 13 17:02:19 what is the general problem behind that? Aug 13 17:02:33 The interfaces don't support it Aug 13 17:02:47 usb interface? Aug 13 17:02:56 It's either the fact that USB drives appear as SCSI devicess in Linux or that the bridge chips don't allow you do pass that command through Aug 13 17:03:02 ah Aug 13 17:03:07 Never looked too closely but it's one of the two Aug 13 17:03:16 "you to pass" Aug 13 17:03:21 i reckon it may be the bridge thing Aug 13 17:03:26 That's my guess Aug 13 17:03:32 scsi drives can be spun down on a normal scsi bus Aug 13 17:03:36 Airw0lf: yes it would only work if root was on flash and disk purally for storage Aug 13 17:03:56 Airw0lf: I was more thinking that the USB MSC driver doesn't support those ioctls Aug 13 17:04:00 Not that the SCSI layer can't do it Aug 13 17:04:15 Tiersten: i think it does Aug 13 17:04:30 It does support the SCSI spindown commands? Aug 13 17:04:34 yes Aug 13 17:04:44 i think it's the cheap USB converters Aug 13 17:04:45 If so then we've so far not come across any bridge chip that accepts it Aug 13 17:04:46 sucky Aug 13 17:05:16 anyway i think i'm going to bed Aug 13 17:05:17 i'm trying to think if you can spin down a usb hd in windows Aug 13 17:05:18 I've got the docs for the Cypress ones which are reasonably popular. I should dig around in it to see what they say Aug 13 17:05:23 mithro: Night Aug 13 17:05:27 night Aug 13 17:05:35 tommorrow i'll get up and overclock this slug :P Aug 13 17:05:38 Airw0lf: You can spindown the Maxtor OneTouch ones but that's because you install a special driver Aug 13 17:05:57 then i need to get my wrt54g playing nice with none of the crappy linksys software Aug 13 17:05:57 The driver handles the extra commands that Maxtor added Aug 13 17:06:00 ah right Aug 13 17:06:34 i got some spam for the maxtor drives with my nslu2 docs Aug 13 17:06:36 Don't know what happens if you don't have the driver at all. I believe some people have experienced some success in enabling the APM features in the drives so it spins down by itself Aug 13 17:07:04 Yeah. The NSLU2 has some Maxtor patches in the kernel to understand what to do about the buttons on the OneTouch Aug 13 17:07:20 bbl, breakfast Aug 13 17:07:25 Enjoy Aug 13 17:34:38 damn wikiwords Aug 13 17:42:01 that's better... Aug 13 18:14:42 DaKa: ping? Aug 13 18:15:29 DaKa: thttpd doesn't use a config file by default. It serves out /srv/www by default. I haven't looked much into creating that config-file yet. Aug 13 18:15:41 * NAiL just came back from town, so .. Aug 13 23:35:01 NAiL: ah, I figured that Aug 13 23:45:45 thttpd actually works great Aug 13 23:46:24 I didn't think it would be that fast Aug 13 23:51:01 on a 1K file: Transaction rate: 566.46 trans/sec Aug 14 00:02:05 with some tweaking I can get apache to do 320.56 trans/sec Aug 14 00:02:44 but, as soon as the files get larger apache can do about the same as thttpd Aug 14 01:13:44 ~weather kcvo Aug 14 01:13:47 Corvallis, Corvallis Municipal Airport, OR, United States; (KCVO) 44-30N 123-17W; last updated: 2005.08.14 0655 UTC; Dew Point: 48 F (9 C); Pressure (altimeter): 29.87 in. Hg (1011 hPa); Relative Humidity: 40%; Sky conditions: clear; Temperature: 73 F (23 C); Visibility: 10 mile(s); Wind: from the W (280 degrees) at 6 MPH (5 KT) Aug 14 01:16:59 ~weather ksba Aug 14 01:17:01 Santa Barbara, Santa Barbara Municipal Airport, CA, United States; (KSBA) 34-25-34N 119-50-37W 3M; last updated: 2005.08.14 0653 UTC; Dew Point: 54.0 F (12.2 C); Pressure (altimeter): 29.87 in. Hg (1011 hPa); Relative Humidity: 89%; Sky conditions: clear; Temperature: 57.0 F (13.9 C); Visibility: 10 mile(s); Wind: Calm Aug 14 01:17:21 brrr Aug 14 01:18:14 ~weather phnl Aug 14 01:18:17 Honolulu, Honolulu International Airport, HI, United States; (PHNL) 21-19-39N 157-56-35W 4M; last updated: 2005.08.14 0653 UTC; Dew Point: 68.0 F (20.0 C); Pressure (altimeter): 30.01 in. Hg (1016 hPa); Relative Humidity: 66%; Sky conditions: clear; Temperature: 80.1 F (26.7 C); Visibility: 10 mile(s); Wind: from the E (080 degrees) at 13 MPH (11 KT) Aug 14 02:17:00 ~weather ypad Aug 14 02:17:03 Adelaide Airport, Australia; (YPAD) 34-56S 138-31E 4M; last updated: 2005.08.14 0700 UTC; Dew Point: 35 F (2 C); Pressure (altimeter): 29.68 in. Hg (1005 hPa); Relative Humidity: 41%; Sky conditions: overcast; Temperature: 59 F (15 C); Visibility: greater than 7 mile(s); Weather: Light rain; Wind: from the N (010 degrees) at 15 MPH (13 KT) Aug 14 03:00:18 CIA-8: wb Aug 14 03:35:05 anyone have experiences with openslug and a maxtor onetouch drive? I'm trying to figure out why it doesn't spin down even though I'm quite sure there isn't any disk activity Aug 14 03:57:29 hmm, might have been the --MARK-- messages syslogd enters ever now and then... Aug 14 04:11:27 billm: Max size for Jumbo frames is 9000 IIRC Aug 14 04:15:52 sorry, it isn't actually that predictable - different interfaces support different MTUs Aug 14 04:17:41 the em driver (with Intel chips) supports 16114, whilst the Realtek driver only supports 7422 Aug 14 04:18:09 updating Ticket 346 with this info, btw Aug 14 04:21:58 argh, wrong room Aug 14 04:22:11 need to get used to this new IRC client Aug 14 04:25:59 ~weather enva Aug 14 04:26:01 Trondheim / Vaernes, Norway; (ENVA) 63-28N 010-56E 17M; last updated: 2005.08.14 0820 UTC; Dew Point: 48 F (9 C); Pressure (altimeter): 29.94 in. Hg (1014 hPa); Relative Humidity: 81%; Sky conditions: partly cloudy; Temperature: 53 F (12 C); Visibility: greater than 7 mile(s); Wind: from the WSW (250 degrees) at 6 MPH (5 KT) (direction variable) Aug 14 04:34:36 neat Aug 14 04:34:48 someone has run linux and windows on the psp Aug 14 04:43:23 emulation, surely Aug 14 04:43:52 native linux would be interesting Aug 14 04:44:38 bochs Aug 14 04:44:50 but it's a step forward :) Aug 14 04:45:00 true, but a painfully slow one Aug 14 04:55:18 after the openslug build finishes, where is the actual package Aug 14 04:55:26 sorry, image Aug 14 04:56:41 tmp/deploy/images Aug 14 04:56:59 openslug--flashdisk.img Aug 14 04:58:01 ucslug -u openslug--flashdisk.img is correct ? Aug 14 04:58:12 sorry, upslug -u openslug--flashdisk.img Aug 14 05:35:11 blaster8: yup Aug 14 06:20:46 Hey all! Aug 14 06:21:06 Anyone here alive / awake? Aug 14 06:21:28 yes Aug 14 06:22:34 Dou you happen to know anything about kernel modules? Aug 14 06:23:01 I have foudn a rtl8150 USB-2-Ethernet thingie Aug 14 06:23:07 a bit Aug 14 06:23:18 I can get it to work with insmod, but not on boot-up Aug 14 06:23:29 Is this unslung or openslug Aug 14 06:23:30 ? Aug 14 06:23:50 oh, uhm, unslung Aug 14 06:24:30 (I only got the thing yesterday, at least I got Apache / SSH / de-underclock going) Aug 14 06:25:49 well, that's not bad ;) Aug 14 06:26:01 I don't know anything about how the unslung boot process is however Aug 14 06:26:11 nope, but I would like to use it as a ip_masq box Aug 14 06:26:27 the odd thing is, that modprobe throwws an error Aug 14 06:26:43 root@NSLUg:/kit# modprobe rtl8150 Aug 14 06:26:43 modprobe: could not parse modules.dep Aug 14 06:26:54 (yes, that's a bash# prompt ;-) ) Aug 14 06:27:06 hmm Aug 14 06:27:07 depmod -a Aug 14 06:27:18 If that's even installed ;) Aug 14 06:27:41 then try again Aug 14 06:27:51 root@NSLUg:/kit# depmod -a Aug 14 06:27:51 root@NSLUg:/kit# Aug 14 06:28:02 still same error Aug 14 06:28:04 hmm Aug 14 06:28:24 the module is installed and I can insert it with insmod, but it's hardly the official way Aug 14 06:28:26 could you cat modules.dep, and put it on http://pastebin.ca/ ? Aug 14 06:29:05 Strike_2: /unslung/rc.modules Aug 14 06:29:23 insmod it there, after reading up on Diversion Scripts in the wiki. Aug 14 06:29:24 modules.dep == empty Aug 14 06:29:33 aha Aug 14 06:29:36 listen to rwhitby ;) Aug 14 06:30:10 root@NSLUg:/kit# cat /unslung/rc.modules Aug 14 06:30:10 cat: /unslung/rc.modules: No such file or directory Aug 14 06:30:12 ile Aug 14 06:30:32 Strike_2: read up on Diversion Scripts first Aug 14 06:31:19 03nail 07org.openembedded.nslu2-linux * rd4efe752... 10/packages/meta/openslug-packages.bb: Add usbutils Aug 14 06:31:21 03nail 07org.openembedded.nslu2-linux * rf059cbad... 10/packages/zd1211/zd1211-20050315/Makefile.patch: Fixup patch, minor typo Aug 14 06:31:24 03nail 07org.openembedded.nslu2-linux * r878c33f6... 10/conf/distro/openslug-packages.conf: Add usbutils and zd1211 Aug 14 06:31:50 reading... Aug 14 06:33:49 OK, so should I just insmod with a diversion script? Aug 14 06:34:46 yes, in the file that i mentioned Aug 14 06:35:05 (which you will need to create, as is the case for all diversion scripts) Aug 14 06:35:44 ah, so the rc.modules is a script tha twill manually load a mocule Aug 14 06:36:50 It's odd that the /etc/modules trick in EthernetAdapter didn't work, or is that for OpenSlug? Aug 14 06:39:05 did you try /etc/modules? Aug 14 06:39:11 Yup Aug 14 06:39:23 root@NSLUg:/kit# cat /etc/modules Aug 14 06:39:23 rtl8150 alias eth0 Aug 14 06:39:43 maybe alias eth0 rtl8150 (RedHat style)? Aug 14 06:42:05 (Only now did I notice a NewUsersGuide, checking...) Aug 14 06:47:17 03rwhitby 07org.openembedded.nslu2-linux * r90995c6a... 10/files/device_table-unslung.txt: Expanded Unslung device table to match Linksys firmware 2.3R63 Aug 14 06:47:26 03rwhitby 07org.openembedded.nslu2-linux * r8dcbe859... 10/conf/machine/nslu2.conf: Removed the unused variables from nslu2.conf Aug 14 06:47:28 03rwhitby 07org.openembedded.nslu2-linux * r0bfafb15... 10/ (6 files in 2 dirs): Split nslu2_flashimg.bbclass into nslu2-jffs2-image.bbclass and nslu2-ramdisk-image.bbclass Aug 14 06:47:33 03rwhitby 07org.openembedded.nslu2-linux * rf27802cb... 10/ (16 files in 5 dirs): Added unslung-binary-kernel.conf distro for NSLU2 Linksys V2.3R63 firmware with NTFS read/write support. Aug 14 06:47:47 nice commit rwhitby Aug 14 06:48:14 btw, R63 removes the handy behavior where holding the reset button starts Upgrade mode Aug 14 06:49:22 *sigh* Aug 14 06:49:23 so rwhitby, why arn't we all using the uclibc stuff? Aug 14 06:49:35 mithro: ucslugc? Aug 14 06:49:42 mithro: 'we' are - look at UcSlugC Aug 14 06:49:52 mithro: when you've got a 200Gb disk attached, why bother with uclibc ? Aug 14 06:50:07 rwhitby: because you want to be able to disconnect the 200gb disk? :P Aug 14 06:50:12 and for those who want a diskless installation, there is UcSlugC. Aug 14 06:50:29 Same as OpenSlug, but with uClibc instead. Aug 14 06:50:30 rwhitby: Eventually, I want to have my entire root on the flash disk, with data on my HD Aug 14 06:50:40 uclugc has it's limitations... bind is one thing that won't compile Aug 14 06:50:46 neater Aug 14 06:50:47 (And probably never will) Aug 14 06:50:48 blaster8: thats what i want too Aug 14 06:51:26 mithro: waiting till UcSlugC is a little more stable, mind Aug 14 06:51:33 Basically, we couldn't be bothered working out all the uClibc incompatibilities when building stuff, so OpenSlug was done with glibc. UcSlugC is now on-line so that interested parties can sort out the problems. Aug 14 06:51:39 so now i have a nslu2 and a wrt54g, i wonder if i could have a mosix cluster! :P Aug 14 06:51:48 NAiL: No-one likes BIND (not secure) Aug 14 06:51:49 hehe Aug 14 06:52:00 blaster8: the alternatives aren't much better Aug 14 06:52:08 there is one good one Aug 14 06:52:23 the slug doesn't have any spair GPIO pins does it? Aug 14 06:53:02 blaster8: what? Aug 14 06:53:13 djb's piece of crap? Aug 14 06:53:36 what is the problem with it? Aug 14 06:53:51 One, it doesn't adhere with the FHS. And there's no way to fix it. Aug 14 06:54:02 Two, it requires lots of other stuff to be installed Aug 14 06:54:26 Three, it doesn't do proper zone-transfers Aug 14 06:54:42 Four, last time I checked, you could neither interface with mysql, nor use ssl Aug 14 06:55:19 but, feel free to get it to compile and work ;) Aug 14 06:55:31 * NAiL gotta go Aug 14 06:56:35 hmm, according to gentoo djbdns has 2 dependencies Aug 14 06:57:05 as opposed to BIND's 2 dependencies Aug 14 06:58:40 well i'm off see ya Aug 14 07:04:04 I'm not exactly an expert at DNS server configuration but these two pages seem fair: http://homepages.tesco.net./~J.deBoynePollard/FGA/djbdns-myths-dispelled.html http://homepages.tesco.net./~J.deBoynePollard/FGA/djbdns-problems.html Aug 14 07:21:18 03rwhitby 07org.openembedded.nslu2-linux * r52d3a2c8... 10/conf/distro/unslung-binary-kernel-packages.conf: Added unslung-binary-kernel-packages.conf Aug 14 07:21:18 blaster8: I don't care if others use djb's stuff. I personally won't though Aug 14 07:46:13 03rwhitby 07org.openembedded.nslu2-linux * r28750942... 10/ (7 files in 4 dirs): Added the unslung OVERRIDE for unslung-binary-kernel distro Aug 14 07:51:14 03nail 07org.openembedded.nslu2-linux * r407eb029... 10/packages/openslug-init/openslug-init-0.10/initscripts/zleds: Update so that the disk leds aren't switched on/off Aug 14 08:06:59 03rwhitby 07org.openembedded.nslu2-linux * r3b73d893... 10/packages/meta/unslung-binary-kernel-packages.bb: Added unslung-binary-kernel-packages.bb Aug 14 08:23:44 for trying to audio recording, libao was suggested by dyoung yesterday. it isn't in the ipkg list because it has "runtime problems". i have unslung, and did 'ipkg install python' as a test. how do i proceed with work on libao? Aug 14 08:26:20 03nail 07org.openembedded.nslu2-linux * r703248fa... 10/packages/openntpd/openntpd_3.7p1.bb: Use sysconfdir/localstatedir instead of hardcoded paths. Create homedir in postinst Aug 14 08:31:10 03nail 07org.openembedded.nslu2-linux * rdda7bbb0... 10/packages/openntpd/openntpd_3.7p1.bb: Check if the dir exists first ;-) Aug 14 08:41:07 03nail 07org.openembedded.nslu2-linux * rc344fe82... 10/packages/openntpd/openntpd_3.7p1.bb: Fixup typo Aug 14 09:52:51 Resolving www.you-need-to-get-this-one-yourself.com... failed: Name or service not known. Aug 14 09:52:54 hmm Aug 14 09:53:11 what should I do? Aug 14 09:53:34 anything in /etc/resolv.conf? Aug 14 09:53:53 ;) Try a make update ; make anything and you will see what I mean Aug 14 09:56:48 hahah Aug 14 09:59:31 jbowler: :) Aug 14 10:02:03 hmm, the only issue is that I did get it myself and put it in downloads Aug 14 10:02:12 but it still tries to grab it over the internet Aug 14 10:02:16 did you md5sum it? Aug 14 10:02:21 of course Aug 14 10:02:34 what's the command (still a n00b at heart) Aug 14 10:03:00 md5sum|cut -d " " -f 1 > filename.md5 Aug 14 10:03:56 hmm... VoodooZ.. you good with redboot? :) Aug 14 10:04:30 * NAiL needs to test the ram on his slug somehow Aug 14 10:05:49 thanks, NAiL - I'll add this stuff to the wiki Aug 14 10:06:08 np Aug 14 10:11:38 nope. I only use it to flash new loads. that's it. Aug 14 10:11:53 I would swap it for Apex anytime. Aug 14 10:13:08 well, yeah Aug 14 10:13:30 I'm not sure if I'm ready to flash with apex quite yet Aug 14 10:14:10 It'd be nice though Aug 14 10:14:29 but someone would have to test it before I'd even consider flashing Aug 14 10:17:23 03repvik * r176 10/releases/OpenSlug-2.5-beta/openembedded/packages/openntpd/openntpd_3.7p1.bb: Update openntpd to not use fixed paths, depend on ntp-tickadj Aug 14 10:22:33 anyone know how to make bash the default shell for new users? Aug 14 10:50:39 NAiL, Same here. I'm a bit chicken. This whole monotone merging issue slowed my robot's development enough as it is without bricking my slug! Aug 14 10:52:10 VoodooZ: does --lca help? Aug 14 10:59:37 * NAiL hugs ingeba Aug 14 10:59:47 not that he's here, but ccxstream works like a charm ;) Aug 14 11:00:10 * NAiL is currently streaming video from the slug to his XBox Aug 14 11:04:25 jbowler-away, you mean for merging? Aug 14 11:05:07 My monotone repo is damaged beyond repair so I'll have to start from scratch AGAIN. Aug 14 11:05:14 Hopefully it won't happen again. Aug 14 11:05:18 how do you do it? Aug 14 11:05:24 are you running 0.22? Aug 14 11:08:21 VoodooZ: yes, for merging Aug 14 11:08:39 Given that the changes aren't being propagated back it should help a lot... Aug 14 11:13:22 NAiL DaKa: I'd like to be able to use openslug-native with ucslugc, but at present it RDEPENDS glibc and has no DEPENDS, so I have a version which does the uclibc/glibc switch and correct (I think) DEPENDS - how do you feel about me checking this in? Aug 14 11:18:17 Nail: what are you using to stream th evideo? Aug 14 11:18:29 ccxstream Aug 14 11:18:35 * blaster8 thinks he is NAiL Aug 14 11:19:54 * copperbeech grins Aug 14 11:20:10 ccxstream is specifically for xbox? Aug 14 11:20:35 so how stable is the uclibc image? Aug 14 11:22:06 well, if you mean crashing stable then yes Aug 14 11:22:12 but if you mean feature stable, then no Aug 14 11:22:40 http://www.nslu2-linux.org/wiki/UcSlugC/Status Aug 14 11:25:19 jbowler-away, how would I call it though? I usually encounter the 3way merges while doing make update. Would I need to mod Makefile? Aug 14 11:27:08 jbowler-away: Not sure about the ramifications Aug 14 11:27:55 ccxstream is specifically for some XBox media centre software only, I think Aug 14 11:28:47 hi Aug 14 11:28:52 hey there Aug 14 11:29:30 has anybody hee tried to compile nslu2_optware on an FC4 machine? Aug 14 11:30:38 also is there something wrong with the monotone server? the makefile freezes .... Aug 14 11:31:15 or rather is reaaaallly slow Aug 14 11:34:15 probably something wrong with FC4 ;) Aug 14 11:35:44 Bijan: what monotone version are you using? Aug 14 11:36:01 VoodooZ: do it by hand, run monotone in the openembedded directory, that way you know it's safe to kill it if weirdness starts happening. Aug 14 11:36:40 NAiL: the main implication is that I added 'DEPENDS' to ensure that all the required packages are there. I don't know if there is a reason for not having it. Aug 14 11:37:14 The minor implication is that I had to make some of the dependencies TARGET_OS specific, so there isn't a single list any longer. Aug 14 11:38:11 03matt_mcneill * 10unslung/Makefile: correcting an appweb make file in the WGTESTING zone Aug 14 11:45:50 blaster8 : I got monotone 0.22 Aug 14 11:46:16 we don't really support Fedora Core - the devs use Debian and Gentoo mostly Aug 14 11:46:57 oh ok, am I correct assuming that those are not using gcc 4.x ? Aug 14 11:46:59 if i have the current OE monotone repository how do i pull nslu2? Aug 14 11:48:13 mithro: just use the makefile and make update Aug 14 11:48:20 Bijan: no Aug 14 11:48:30 hmmm Aug 14 11:49:00 ok, I guess no FC4 specific build questions anymore then ;-) Aug 14 11:49:07 GCC 4 really isn't ready for production use on x86 Aug 14 11:49:20 makefile? Aug 14 11:49:38 the master makefile - http://www.nslu2-linux.org/wiki/HowTo/UseTheMasterMakefile Aug 14 11:50:16 Describe UseTheMasterMakefile? here. Aug 14 11:50:31 different question : Openslug doesn't seem to come with an /etc/shadow file. Is that intended? Aug 14 11:51:32 YAY! I got my USB-Ethernet thingie working! Aug 14 11:51:47 sorry it moved: http://www.nslu2-linux.org/wiki/Development/MasterMakefile Aug 14 11:51:52 only one thing: it won't parse my ifcfg-eth0 Aug 14 11:52:01 Strike_2: Well done Aug 14 11:52:26 Strike_2: Most appreciated if you could write up how you managed it on the wiki Aug 14 11:52:29 I can happily report unslug 5.5 compatibility with Sitecom LN-013 (RTL8150) Aug 14 11:52:32 Will do Aug 14 11:52:42 thanks Aug 14 11:52:50 blaster8: so where do i get the MasterMakefile then :P Aug 14 11:53:03 follow the instructions: http://www.nslu2-linux.org/wiki/Development/MasterMakefile Aug 14 11:53:07 it's all there Aug 14 11:53:11 http://www.nslu2-linux.org/wiki/Development/MasterMakefile Aug 14 11:53:12 does anyone know which daemon / script parses ifcfg scripts? Aug 14 11:53:25 blaster8: sorry missed your line Aug 14 11:53:33 np :) Aug 14 11:54:45 is there any way to get sudo su - working? Aug 14 11:54:58 sudo works, but su doesn't 'do' anything Aug 14 11:55:13 use login Aug 14 11:55:17 ugly but true Aug 14 11:55:19 jbowler-away: sorry, had to go. Go ahead and commit Aug 14 11:55:29 and if you want a proper linux, then use Openslug Aug 14 11:56:02 I've only had the thing powered up for 24 hrs Aug 14 11:56:13 for the moment, I am using 'sudo bash' Aug 14 11:56:37 hrm that page doesn't describe how to get the uslugc one Aug 14 11:56:48 actually it does Aug 14 11:57:03 but anyway Aug 14 11:57:07 then we should all hail mithro the blind one :P Aug 14 11:57:08 http://www.nslu2-linux.org/Makefile Aug 14 11:57:20 search for ucslugc Aug 14 11:57:47 if you want a target, for now you should just search through the Makefile itself Aug 14 11:57:58 so load the link above in your browser and get looking Aug 14 11:58:09 actually the target you want is 'make ucslugc' Aug 14 11:58:11 but anyway Aug 14 11:58:14 yeah found it Aug 14 11:58:23 it is on the MasterMakefile page Aug 14 11:58:28 have you got the right software installed Aug 14 11:58:28 just a bit hidden Aug 14 11:58:34 i should Aug 14 11:58:37 don't use the setup target Aug 14 11:58:45 just make ucslugc Aug 14 11:58:51 unless you require something OE doesn't Aug 14 11:59:34 monotone 0.22/psyco/ccache are the main ones, but they should be OE standard Aug 14 11:59:49 hrm, i havn't install ccache yet Aug 14 11:59:53 get it, really Aug 14 12:00:13 when you need to rebuild it will be 5x faster, literally Aug 14 12:00:22 yeah i have it on my normal build machine Aug 14 12:00:31 but i'm now building on my laptop Aug 14 12:03:12 how do you enable ccache? Aug 14 12:07:20 you don't need to Aug 14 12:07:28 the scripts enable it automagically Aug 14 12:07:39 oh, neat Aug 14 12:07:44 for regular compiling and for portage you do need to muck around, however Aug 14 12:07:50 but it's not that hard Aug 14 12:08:08 Is this alright: http://www.nslu2-linux.org/wiki/Peripherals/EthernetAdapter Aug 14 12:08:10 iptables 1.3.3 builds (my first bitbake hack) Aug 14 12:08:17 ( my first wiki edit) Aug 14 12:08:40 great Aug 14 12:09:11 until USB Ethernet is a bit more mainstream, ppl will have to use anecdotal hints to get it right Aug 14 12:09:27 ;-) Aug 14 12:09:51 and remember, the wiki is as good as you make it :) Aug 14 12:10:08 I tried a lot of ways to auto-load the module on boot, but was stuck with a rc-script Aug 14 12:10:20 so anyone use one of those cheap skype usb phone things with a slug? Aug 14 12:11:21 I have one, but no use for skype on my Slug Aug 14 12:12:01 mithro: would love to Aug 14 12:12:14 but I don't have the first clue Aug 14 12:12:30 jbowler-away, ok. but wouldn't it be something that the Makefile and thus everybody could benefit from? Aug 14 12:13:02 why are you using a yahoo group!? Aug 14 12:13:09 historical reasons Aug 14 12:13:36 hysterical raisins Aug 14 12:13:40 and that Aug 14 12:13:49 how do you get ipkg to install a local ipkg Aug 14 12:14:01 ipkg install ./something.ipkg Aug 14 12:14:33 I think iptables will work, btw Aug 14 12:15:08 some of these captchas are getting a bit insane Aug 14 12:15:11 i can't even read them Aug 14 12:16:18 03nail 07org.openembedded.nslu2-linux * ra609ff1e... 10/ (2 files in 2 dirs): Added wireless-tools to feed Aug 14 12:16:34 NAiL: How do I get SFTP working on a default openslug install Aug 14 12:16:43 (still on dropbear) Aug 14 12:17:16 uh Aug 14 12:17:20 don't think that'll work Aug 14 12:17:30 hmm Aug 14 12:18:00 why not install openssh? Aug 14 12:18:08 will have to Aug 14 12:20:49 blaster8: http://www.nslu2-linux.org/wiki/HowTo/UseOpenSSHForRemoteAccess Aug 14 12:21:27 mithro: having problems readin the yahoo group did you say? Aug 14 12:21:47 http://www.nslu2-linux.org/wiki/HowTo/SwapFromDropbearToOpenSSH Aug 14 12:21:59 working on it Aug 14 12:22:01 copperbeech: no i just don't like yahoo groups :P Aug 14 12:22:47 me neither - that's why I use the gmane nntp interface. It's fantastic with thunderbird (esp with the view all threads with unread) Aug 14 12:24:21 yeah gmane is great Aug 14 12:25:05 wtf? Aug 14 12:25:17 I have ipx0 on 10.2.1.11 and eth0 on 10.2.1.12 Aug 14 12:25:33 when I disconnect the eth0 cable, neither IP os reachable Aug 14 12:25:35 ?? Aug 14 12:33:21 what's the latest info on the commercial software incorporated in the R63 linksys firmware? Aug 14 12:33:32 anyone know? Aug 14 12:34:08 where do i get this? http://www.you-need-to-get-this-one-yourself.com/NSLU2_V23R63.zip Aug 14 12:34:18 the linksys site only has V23R29 Aug 14 12:34:47 grr Aug 14 12:34:58 read the guide - I updated it 60 mins ago Aug 14 12:35:02 with a direct link Aug 14 12:35:31 sorry, just getting grouchy - need to concentrate on iptables Aug 14 12:35:48 http://www.nslu2-linux.org/wiki/Development/MasterMakefile Aug 14 12:35:50 NB 2 Aug 14 12:36:11 03nail 07org.openembedded.nslu2-linux * r933c360c... 10/packages/meta/openslug-packages.bb: Added zd1211 driver to feed Aug 14 12:37:24 blaster8: well i will soon be adding a light weight bittorrent client which is about 8 times faster then the real one and only uses about 6MB of memory for 400Kbytes/sec and 60 clients at once Aug 14 12:37:25 :P Aug 14 12:37:26 blaster8: thanks - I had already cast my eye over the recent changes, but hadn;t picked up that the info would be in the page for the MasterMakefile :-) Aug 14 12:38:16 mithro: what client? Aug 14 12:38:22 I'm using ctorrent atm Aug 14 12:38:27 Can anyone here guess what's going on? Aug 14 12:38:29 http://pastebin.com/336861 Aug 14 12:38:46 rtorrent Aug 14 12:39:06 http://libtorrent.rakshasa.no/ Aug 14 12:39:14 mithro: how's your progress thus far? Aug 14 12:39:23 <[cc]smart> NAiL: i have taken bug #256. i would suggest amending bootmisc.sh accordingly. do you agree with that ? Aug 14 12:40:28 [cc]smart: ah, yes Aug 14 12:40:29 NAiL: well i'm currently compiling ucslugc :P Aug 14 12:40:40 mithro: should work just nicely with openslug then Aug 14 12:41:21 i would like it to work off the flash downloading to the HD's Aug 14 12:41:33 [cc]smart: I'm aware that it's not gentoo specific, but I wasn't aware how it was actually done. It's just a neat feature :) Aug 14 12:41:37 <[cc]smart> Strike_2: output of arp is from second machine, aka windows, correct ? Aug 14 12:41:46 copperbeech: there was a commit by jbowler about 6 hours ago - he has added the paragon ntfs kernel module to the optware feed (I think) Aug 14 12:42:13 <[cc]smart> NAiL: but the way i suggest to do it is accepted ? Aug 14 12:42:25 [cc]smart: let me check, just to be sure Aug 14 12:42:48 [cc]smart: It's just one command, right? Aug 14 12:42:54 ok, how do ppl get files onto their slugs? Aug 14 12:43:00 blaster8: I usually use scp Aug 14 12:43:03 or samba :D Aug 14 12:43:07 has anyone actually collated the ins and outs of R63 - is paragon the only difference? Aug 14 12:43:09 <[cc]smart> more or less. it will verify existing/executable sysctl software and existing file sysctl.conf Aug 14 12:43:10 <[cc]smart> then do it Aug 14 12:43:20 [cc]smart: bootmisc.sh looks like a nice place to me Aug 14 12:43:22 copperbeech: It kills lots of useful functionality too Aug 14 12:43:28 <[cc]smart> ok, will do Aug 14 12:43:56 blaster8: it does? Not one to install then ? Aug 14 12:44:06 copperbeech: and until they release the GPL code there is only so much we can do Aug 14 12:44:29 copperbeech: removes the handy reset button trick for entering upgrade mode for as start Aug 14 12:44:35 which suggests a new bootloader Aug 14 12:45:02 blaster8: yep - we'll have to wait... new bootloader... pretty serious, anyone tried a binary decomposition Aug 14 12:45:07 i thought bootloader wasn't overwritten on update Aug 14 12:45:30 it can be - we don't know how the linksys cgi files work Aug 14 12:45:37 <[cc]smart> Strike_2: The output you get is fully correct and expected. Aug 14 12:45:46 jacques: upslug -U will overwrite the entire flash (!) Aug 14 12:46:10 hmm, ok I have never used upslug Aug 14 12:46:39 <[cc]smart> Strike_2: Or better, the behaiour is correct, the output is windowzish silly. Aug 14 12:47:31 <[cc]smart> Strike_2: my guess owuld be though, that for this case, linux prolly will have the same windowzish silly output :) Aug 14 12:56:07 blaster8: that was rwhitby's commit. Aug 14 12:56:14 sorry, couldn't remember Aug 14 12:56:36 The bootloader is not installed by either upslug or reflash - to overwrite it you have to hack upslug. Aug 14 12:57:04 I'm going to change ucslugc to avoid the need to download the linksys firmware, since it isn't really necessary. Aug 14 12:57:23 jbowler: good idea, for most ppl it is just a nuisance Aug 14 12:59:39 oops, sorry Aug 14 12:59:44 was playing with network config Aug 14 13:00:17 I was trying to set up the internal interface for local and USB interface for internet, but that doesn't really work Aug 14 13:00:44 somehow, the Linksys software overwrites the IP info in ifcfg-ixp0 before/after reboot Aug 14 13:00:57 blaster8: it does create a bootstrap problem, it's necessary to get a system with a working reflash, and that means upsluging something, but the binary openslug is fine... Aug 14 13:01:46 the good news: it will now (automagically) load teh correct kernel module (syntax in /etc/modules was probably incorrect) Aug 14 13:02:10 looks like more fixes for Linksys brokenness necessary in Unslung Aug 14 13:03:02 [cc]smart OK, arp-duplicate is now fixed (arp -d * and ipconfig /flushdns on Windoze side) Aug 14 13:03:37 <[cc]smart> Strike_2: Won't change anything :) Aug 14 13:03:42 ? Aug 14 13:03:46 <[cc]smart> Strike_2: Cause there is no fix to be applied. Aug 14 13:03:53 <[cc]smart> This is correct behaviour Aug 14 13:04:03 ? Aug 14 13:04:13 why? Aug 14 13:04:24 <[cc]smart> go to the unslung box and do netstat -r Aug 14 13:04:36 <[cc]smart> for as long as you have no special routing, that's it Aug 14 13:04:45 <[cc]smart> see ? Aug 14 13:05:21 so it does routing to the same subnet over either IP Aug 14 13:05:31 <[cc]smart> sure, it is the same subnet Aug 14 13:05:59 uhm, why does it override my settings? Aug 14 13:06:09 what pastebin do you guys here use? Aug 14 13:06:13 <[cc]smart> why do ou think it's overriding ? Aug 14 13:06:25 lemme show you some files Aug 14 13:06:42 <[cc]smart> i see this: http://pastebin.com/336861 Aug 14 13:06:51 <[cc]smart> now add netstat -r output Aug 14 13:06:52 oh, waitasec... Aug 14 13:07:03 I still have a script in /unslung Aug 14 13:07:09 I thought I had gotten rid of that... Aug 14 13:08:06 Why does Windows report the same MAC for both IP's, though? Aug 14 13:10:14 <[cc]smart> look at netstat -r once again Aug 14 13:10:23 jacques, long time no see. Aug 14 13:10:37 <[cc]smart> on the slug Aug 14 13:10:39 hi VoodooZ Aug 14 13:10:42 I saw the NSLU2 kiosk you guys put together. Cool stuff. Aug 14 13:11:02 the LWE booth? Aug 14 13:11:55 yeah. Aug 14 13:12:08 good exposure. Aug 14 13:12:57 yep I agree Aug 14 13:16:22 [cc]smart http://pastebin.com/336884 Aug 14 13:16:32 03jbowler 07org.openembedded.nslu2-linux * r2e318668... 10/ (14 files in 10 dirs): Aug 14 13:16:32 propagate from branch 'org.openembedded.dev' (head 90c52179101c43123ae384910c0aa1ee36375c11) Aug 14 13:16:32 to branch 'org.openembedded.nslu2-linux' (head 933c360c8ecf4315f45d81c4049112cb00b7f479) Aug 14 13:18:02 even weirder: If I disconnect the .11 line, only it is unreachable Aug 14 13:18:13 if I disconnect the .12 line, _both_ are unreachable Aug 14 13:19:02 jbowler: which depconfig would you add modules to to compile them into openslug? Aug 14 13:19:25 and how would you get the modules to be rebuilt? Aug 14 13:21:26 The defconfig is packages/linux/openslug-2.6.12.2/defconfig (at present). I just re-run temp/do_compile*. Aug 14 13:21:45 ok, I changed the correct defconfig Aug 14 13:21:46 [cc]smart : I think I got it: it receives on either, but it replies on one Aug 14 13:21:54 what does the last bit mean, jbowler? Aug 14 13:22:24 the fact that it receives both from one car confuses Windows (ping replies from both IP's from teh same MAC} Aug 14 13:22:48 s/car/card Aug 14 13:22:52 cd tmp/work/openslug-kernel*; $(tail $(ls -rt temp/run.do_compile*)) Aug 14 13:23:00 cd tmp/work/openslug-kernel*; $(tail -n 1 $(ls -rt temp/run.do_compile*)) Aug 14 13:23:04 (forgot the -n 1) Aug 14 13:23:52 That just builds them, and you probably have to edit the .config in the ${S} directory first. Aug 14 13:24:09 (I run menuconfig by hacking temp/run.do_configure, that guarantees I get the edit correct). Aug 14 13:24:31 clever Aug 14 13:24:31 Alternatively just bb -b packages/linux/openslug-kernel_2.6.12.bb -c clean Aug 14 13:24:38 Alternatively just bb -b packages/linux/openslug-kernel_2.6.12.2.bb Aug 14 13:24:56 (i.e. clean, rebuild) - that gets you the ipks too. Aug 14 13:25:16 off it goes Aug 14 13:26:08 btw, the bridge tools are present in the Openslug 2.6 beta feed without the kernel-module-bridge actually existing Aug 14 13:26:28 I'll add it once I've checked it works Aug 14 13:33:33 koen: did you write an init script for iptables Aug 14 13:37:06 bb in a bit Aug 14 13:37:06 Strike_2: Such behaviour is not uncommon with 2 adapters on the same subnet. When looking for a return path IP doesn't look at the packet that came in (after all the IP layer doesn't even kow if there was one) it just picks the lowest cost interfadce which has a route to the required host Aug 14 13:40:13 DaKa: How'd you set up the toolchain? Aug 14 13:41:24 for? Aug 14 13:41:43 NAiL: apex? Aug 14 13:42:39 yes Aug 14 13:42:51 well... install crosstool and build Aug 14 13:43:34 ok Aug 14 13:43:50 anyone set up samba to use ldap, btw? Aug 14 13:43:55 somehow the OE didn't work.. Aug 14 13:44:05 DaKa: Yes, I found that out... Aug 14 13:44:08 :) Aug 14 13:44:57 I'm not even using samba with ldap for my 1200+ users network.. Aug 14 13:45:00 probably should.. Aug 14 13:45:38 yes ;) Aug 14 13:45:57 Well... ldap can apparently be used for lots of useful stuff Aug 14 13:46:14 Damn, I need to get a slug connected permanently to teh intarweb Aug 14 13:46:50 DaKa: Is it possible to test if apex works with redboot? Aug 14 13:46:59 ie, load it to ram and run it? Aug 14 13:47:07 yes Aug 14 13:47:18 read on the apex wiki Aug 14 13:47:37 You wouldn't happen to have an image for a "normal" slug hanging around, would you? ;) Aug 14 13:47:55 hm.. no.. could build one if you like.. Aug 14 13:48:00 you have jtag? Aug 14 13:48:20 nope :P Aug 14 13:48:26 hm.. Aug 14 13:48:46 "Friends don't let friends flash custom bootloaders without confirmed JTAG access." Aug 14 13:48:55 :) Aug 14 13:48:58 You're not my friend, are you? Aug 14 13:49:08 hehe, ok then Aug 14 13:50:44 NAiL: http://david.thg.se/saker/apex.bin Aug 14 13:50:58 dunno if it works, but loading into ram shouldnt mess anything up. Aug 14 13:51:29 but.. I don't want to hear you flashing that without jtag :) Aug 14 13:51:42 dammit Aug 14 13:51:55 why the fsck does openslug try to download R63? Aug 14 13:53:11 DaKa: Ok, I won't tell you :P Aug 14 13:53:31 Although, I'm fully aware that if I flash and it fails, I have a brick. Aug 14 13:53:42 DaKa: Mind md5'ing? Aug 14 13:53:50 hm.. could do.. Aug 14 13:53:51 df05ab0e24655b93487c94a1d0e7f624 Aug 14 13:54:07 just have to find the file first... Aug 14 13:54:34 df05ab0e24655b93487c94a1d0e7f624 Aug 14 13:54:51 looks good :) Aug 14 13:55:00 (not that I have jtag anyway..) Aug 14 13:59:33 ugh Aug 14 13:59:37 My kernel oopses Aug 14 14:01:08 jbowler: still not building the modules Aug 14 14:02:08 can you tell me how you hack run.do_configure to use menuconfig? Aug 14 14:02:18 something must be wrong with my defconfig Aug 14 14:03:04 the changes I make don't get into log.do_configure Aug 14 14:04:44 well Aug 14 14:04:50 apex loads the kernel atleast :D Aug 14 14:05:30 jacques: Rod put the NTFS support which is a binary module in the R63 file into unslung - presumably he has or is about to with openslug. Aug 14 14:05:55 AdamBake1: The kernel module may not be 2.6 compatible Aug 14 14:05:59 personally I don't want it in openslug Aug 14 14:06:07 jacques: agreesd Aug 14 14:06:10 AFAIK that hasn't been discussed at all Aug 14 14:06:25 mke2fs -j -O dir_index /dev/sda1 is good enough for me Aug 14 14:06:33 blaster8, good point - it's a 2.4 module Aug 14 14:06:34 DaKa: works. Thanks a bunch ;) Aug 14 14:06:40 NAiL: :) Aug 14 14:06:55 I stand by my "there is no good reason for openslug to use R63" statement Aug 14 14:07:12 DaKa: This speeds up my dev work a bit too Aug 14 14:07:20 I don't have to use reflash to switch kernels all the time Aug 14 14:07:34 yep, apex is really really nice Aug 14 14:07:51 I'll switch when it is actually released Aug 14 14:08:28 it'll be great to have an alternative bootloader Aug 14 14:09:29 only problem is flashing the bootloader can turn your *slug into a brickslug Aug 14 14:09:51 well since you have jtag you can fix it Aug 14 14:10:07 (since that's one of the rules) Aug 14 14:10:07 exactly Aug 14 14:10:21 * NAiL coughs Aug 14 14:10:23 :) Aug 14 14:10:33 now, how to test the mem with apex ? ;) Aug 14 14:11:02 well, really no good way.. but using copy and compare works kind of good Aug 14 14:11:36 There's no easy way to test the whole of memory? Aug 14 14:11:43 I don't have my notes here, and no apex slug here either Aug 14 14:12:05 well, a small number of copy/compare pasted in should do the trick Aug 14 14:12:19 but no, no single command that does it all Aug 14 14:12:59 best way is copying the kernel to other places and comparing that Aug 14 14:17:19 DaKa: btw, 1.3.1 came out three days ago. "relevant" fixes include: Aug 14 14:17:19 # Improved error handling with copy command. Aug 14 14:17:19 # Fixed startup command parser. There was an undetected NULL pointer dereference bug. Aug 14 14:17:23 # Added ^C handling. It is especially useful with ethernet commands. Aug 14 14:18:05 hm.. Aug 14 14:18:30 should probably get that.. tomorrow Aug 14 14:18:53 hmm... apex doesn't do GPIO stuff on the slug, does it? Aug 14 14:18:59 right now Im doing what I hope will be my last apache test compile Aug 14 14:19:00 * NAiL just noticed the led is off Aug 14 14:19:07 great :D Aug 14 14:19:07 it does Aug 14 14:19:15 it flashes and stuff Aug 14 14:19:21 when doing some things Aug 14 14:19:23 iirc Aug 14 14:19:26 hmm... ok Aug 14 14:19:31 Haven't seen that yet ;) Aug 14 14:20:06 when doing xrecive i think.. Aug 14 14:20:45 neat Aug 14 14:21:01 apex loads the kernel directly off flash now, or? Aug 14 14:21:22 uh... think so... Aug 14 14:21:24 maybe.. Aug 14 14:21:40 maybe... :) Aug 14 14:21:50 that's why I'm not upgrading yet Aug 14 14:22:11 hehe Aug 14 14:22:37 I'll bug beewoolie or g2 when they're on Aug 14 14:23:14 heh, I was wondering why my apache compile died.. guess what, I disabled swap for some tests earlier :) Aug 14 14:23:43 heh Aug 14 14:24:19 there is a reason why step one for a native build machine is 'add swap' Aug 14 14:24:31 yep Aug 14 14:30:06 fixed the defconfig problem (damn you incompatible 2.4.x and 2.6.x kernel config files) Aug 14 14:43:41 03repvik * r177 10/releases/OpenSlug-2.5-beta/openembedded/packages/sudo/ (sudo.inc sudo_1.6.8p8.bb): Fix sudoers permission Aug 14 14:46:18 03nail 07org.openembedded.nslu2-linux * r8cf91caa... 10/packages/sudo/ (sudo.inc sudo_1.6.8p8.bb): Fix sudoers permission Aug 14 14:51:38 blaster8: cd tmp/work/openslug-kernel*; sed -i 's/yes | oldconfig/menuconfig' temp/run.do_conf* - then just run it, it will build menuconfig and bring it up, cp */.config to defconfig. Aug 14 14:51:52 thanks Aug 14 14:52:30 My problem was that I assumed 2.4 & 2.6 kernel configs were equivalent (just didn't realise) Aug 14 14:52:33 fixed now Aug 14 14:55:34 Someone who has a recent openslug build could try putting this in openslug.conf: Aug 14 14:56:01 PREFERRED_VERSION_nslu2-linksys-firmware = "2.3r29" Aug 14 14:56:47 jbowler-away: testing Aug 14 15:21:10 03nail 07org.openembedded.nslu2-linux * r97bdaac8... 10/packages/vsftpd/vsftpd_2.0.3.bb: Fix hardcoded paths mess. Create localstatedir/shared/empty if not exists Aug 14 15:26:06 03nail 07org.openembedded.nslu2-linux * r2cd93746... 10/ (2 files in 2 dirs): Add vsftpd to feed Aug 14 15:26:51 03repvik * r178 10/releases/OpenSlug-2.5-beta/openembedded/conf/distro/openslug-packages.conf: Add joe Aug 14 15:46:39 03rwhitby 07org.openembedded.nslu2-linux * r53c5b2e8... 10/conf/distro/ (openslug.conf ucslugc.conf): Fixed openslug and ucslugc to use R29 firmware, not the new R63 Aug 14 15:46:42 03nail 07org.openembedded.nslu2-linux * re6e39921... 10/ (9 files in 8 dirs): Aug 14 15:46:42 propagate from branch 'org.openembedded.dev' (head f54beea777293a8da6140fa15a4a7a54d04be965) Aug 14 15:46:42 to branch 'org.openembedded.nslu2-linux' (head 2cd93746e9de314a31afe6613e1982251236b43a) Aug 14 15:49:16 blaster8: the www.you-need-to-get-this-one-yourself.com... has been fixed - make update will get the fix for you. Aug 14 15:49:46 Do you still get the message for unslung? **** ENDING LOGGING AT Sun Aug 14 15:50:06 2005 **** BEGIN LOGGING AT Sun Aug 14 15:50:30 2005 Aug 14 15:51:51 or, getting 3-way merge issues Aug 14 15:51:56 what should I do? Aug 14 15:52:41 blaster8: on what files? Aug 14 15:52:46 (discarding local changes) Aug 14 15:52:49 doesn't say Aug 14 15:53:36 there are multiple heads on nslu2-linux Aug 14 15:53:54 http://pastebin.ca/20028 Aug 14 15:54:10 it asks me for a passphrase? Aug 14 15:54:37 blaster8: you can't fix this easily Aug 14 15:55:04 what's the protocol for fixing it 'at all' Aug 14 15:55:19 it will get fixed Aug 14 15:55:30 2sec, pushing Aug 14 15:55:36 I see Aug 14 15:55:57 part of the problem is that my DSL line is down, so I'm using a modem and this delays my syncs somewhat. Aug 14 15:56:21 I think I understand the problem Aug 14 15:56:25 (Particular because my pesky ISP drops the modem after 12hrs 35s...) Aug 14 15:56:36 blaster8: make update again, and see if you still get problmens Aug 14 15:56:39 problems Aug 14 15:58:19 looks good Aug 14 15:58:40 [cc]smart: ping Aug 14 15:58:46 it's worked, thanks Aug 14 15:58:51 np Aug 14 16:10:05 ~weather kcvo Aug 14 16:10:08 Corvallis, Corvallis Municipal Airport, OR, United States; (KCVO) 44-30N 123-17W; last updated: 2005.08.14 1855 UTC; Dew Point: 46 F (8 C); Pressure (altimeter): 29.88 in. Hg (1011 hPa); Relative Humidity: 22%; Sky conditions: clear; Temperature: 89 F (32 C); Visibility: 10 mile(s); Wind: from the ENE (060 degrees) at 12 MPH (10 KT) (direction variable) Aug 14 16:10:16 that is so wrong Aug 14 16:10:24 Currently: 97 F Aug 14 16:10:33 WTF Aug 14 16:11:30 "last updated: 2005.08.14 1855 UTC" ah that explains that Aug 14 16:12:39 59 here Aug 14 16:13:06 ~weather enva Aug 14 16:13:08 Trondheim / Vaernes, Norway; (ENVA) 63-28N 010-56E 17M; last updated: 2005.08.14 1850 UTC; Dew Point: 46 F (8 C); Pressure (altimeter): 29.94 in. Hg (1014 hPa); Relative Humidity: 67%; Sky conditions: partly cloudy; Temperature: 57 F (14 C); Visibility: greater than 7 mile(s); Wind: from the E (080 degrees) at 7 MPH (6 KT) Aug 14 16:13:17 ~weather oxford Aug 14 16:13:19 I can't find station code "OXFORD" (see http://www.nws.noaa.gov/oso/site.shtml or http://www.nws.noaa.gov/tg/siteloc.shtml for ICAO locations codes). Aug 14 16:14:10 ~weather benson Aug 14 16:14:12 I can't find station code "BENSON" (see http://www.nws.noaa.gov/oso/site.shtml or http://www.nws.noaa.gov/tg/siteloc.shtml for ICAO locations codes). Aug 14 16:14:16 ~weather eham Aug 14 16:14:19 Amsterdam Airport Schiphol, Netherlands; (EHAM) 52-18N 004-46E -2M; last updated: 2005.08.14 1902 UTC; Dew Point: 57 F (14 C); Pressure (altimeter): 29.97 in. Hg (1015 hPa); Relative Humidity: 87%; Sky conditions: mostly cloudy; Visibility: greater than 7 mile(s); Weather: 60 F (16 C); Wind: from the N (350 degrees) at 13 MPH (11 KT) (direction variable) Aug 14 16:14:23 ~weather EGUB Aug 14 16:14:27 Benson, United Kingdom; (EGUB) 51-37N 001-05W 0M; last updated: 2005.08.14 1550 UTC; Dew Point: 53 F (12 C); Pressure (altimeter): 30.15 in. Hg (1021 hPa); Relative Humidity: 63%; Sky conditions: mostly cloudy; Temperature: 66 F (19 C); Visibility: greater than 7 mile(s); Wind: from the N (350 degrees) at 13 MPH (11 KT) Aug 14 16:14:32 got it Aug 14 16:14:36 :-D Aug 14 16:14:51 that's about 4 miles from where I live Aug 14 16:15:20 ~weather ehrd Aug 14 16:15:23 Rotterdam Airport Zestienhoven, Netherlands; (EHRD) 51-57N 004-27E -4M; last updated: 2005.08.14 1857 UTC; Dew Point: 55 F (13 C); Pressure (altimeter): 29.97 in. Hg (1015 hPa); Relative Humidity: 87%; Sky conditions: mostly cloudy; Visibility: 3 mile(s); Weather: 59 F (15 C); Wind: from the N (350 degrees) at 14 MPH (12 KT) (direction variable) Aug 14 16:15:30 some 8 miles from where I am Aug 14 16:15:59 15C, just cold enough to keep my computers cool Aug 14 16:17:25 ~weather essb Aug 14 16:17:28 Stockholm / Bromma, Sweden; (ESSB) 59-21N 017-57E 14M; last updated: 2005.08.14 1850 UTC; Dew Point: 55 F (13 C); Pressure (altimeter): 29.88 in. Hg (1012 hPa); Relative Humidity: 77%; Sky conditions: mostly cloudy; Temperature: 62 F (17 C); Visibility: greater than 7 mile(s); Wind: from the S (170 degrees) at 5 MPH (4 KT) Aug 14 16:26:37 openslug feeds are updated .... should be back to the 15-30 minute commit->feeds schedule now Aug 14 16:39:49 ~weather eddi Aug 14 16:39:51 Berlin-Tempelhof, Germany; (EDDI) 52-28N 013-24E 49M; last updated: 2005.08.14 1920 UTC; Dew Point: 59 F (15 C); Pressure (altimeter): 29.88 in. Hg (1012 hPa); Relative Humidity: 93%; Sky conditions: mostly cloudy; Temperature: 60 F (16 C); Visibility: greater than 7 mile(s); Wind: from the E (100 degrees) at 6 MPH (5 KT) Aug 14 16:51:31 ~weather ksba Aug 14 16:51:34 Santa Barbara, Santa Barbara Municipal Airport, CA, United States; (KSBA) 34-25-34N 119-50-37W 3M; last updated: 2005.08.14 1853 UTC; Dew Point: 57.9 F (14.4 C); Pressure (altimeter): 29.91 in. Hg (1012 hPa); Relative Humidity: 77%; Sky conditions: overcast; Temperature: 64.9 F (18.3 C); Visibility: 5 mile(s); Weather: Mist; Wind: Calm Aug 14 16:52:13 Currently: 99 F Aug 14 16:59:04 ~weather phnl Aug 14 16:59:07 Honolulu, Honolulu International Airport, HI, United States; (PHNL) 21-19-39N 157-56-35W 4M; last updated: 2005.08.14 1853 UTC; Dew Point: 63.0 F (17.2 C); Pressure (altimeter): 30.04 in. Hg (1017 hPa); Relative Humidity: 49%; Sky conditions: mostly clear; Temperature: 84.0 F (28.9 C); Visibility: 10 mile(s); Wind: from the NE (050 degrees) at 15 MPH (13 KT) Aug 14 16:59:39 A wonderful day, I should be out. Aug 14 17:04:01 here I live 500m from the beach and I didn't even go this weekend. Aug 14 17:15:47 Is it possible to do ipmasq with unslung ipfilter? Aug 14 17:19:43 root@NSLUg:~$ iptables -t nat -A POSTROUTING -o ixp0 -j MASQUERADE Aug 14 17:19:43 iptables v1.2.11: can't initialize iptables table `nat': Table does not exist (do you need to insmod?) Aug 14 17:51:10 03nail 07org.openembedded.nslu2-linux * r65876316... 10/packages/ (6 files in 6 dirs): Update MAINTAINER and LICENCE Aug 14 19:19:17 How fast does the RAM on the slug work at? 66, 100 or 133Mhz? Aug 14 19:19:28 133MHz Aug 14 19:19:45 really? good. Aug 14 19:20:00 I'm trying to optimize some image copying function Aug 14 19:20:21 and I was trying to minimize ram accesses thinking it was much slower than the rest. Aug 14 19:20:50 do you know how many cycles it takes to read/write a 8/32 bits? Aug 14 19:22:37 jbowler-away is probably your best bet to ask that Aug 14 19:22:56 cool. Aug 14 19:23:13 I guess it would do me good to read some ARM manual or something too. Aug 14 19:24:34 That won't help - the XScale developer manual should have the timing. Aug 14 19:25:13 That's what I though, it will take me longer to find the right resource than asking around! :) Aug 14 19:25:18 But it is abysmal - Lennert did some timing tests (IIRC, may have been someone else) but I remember figures of 30-100 cycles/access. Aug 14 19:25:28 ouch! Aug 14 19:25:33 Exactly. Aug 14 19:25:45 Being negative, I assumed the worst. Aug 14 19:26:06 even with xscale optimized code? Aug 14 19:26:08 So I'm avoid byte accesses and grouping them in block of 32 bits to write once. Aug 14 19:26:11 Anything you do you want to fit in the cache - really, it must fit in the cache. Aug 14 19:26:39 I'm not too sure how the darn cache works actually. Aug 14 19:26:44 jacques: that's cycles to actually get data out of the DRAM, one word, or maybe one line - it's determined by the hardware. Aug 14 19:26:53 I was under the impression of of them was already claimed by linux Aug 14 19:26:57 The test was done by bypassing the cache. Aug 14 19:27:00 I'm not sure about the other 2. Aug 14 19:27:07 I see. Aug 14 19:27:31 so it should be better with cache on right Aug 14 19:27:31 The cache is just a set associative random replacement cache of lines read from the DRAM, linux shouldn't be messing with it. Aug 14 19:27:51 I guess that was the mini-cache I was refering to. Aug 14 19:28:10 so It should be transparent to my C programs right? Aug 14 19:28:40 as for Xscale optimization: shouldn't the armeb-linux-gcc cross-compiler generated by OE do that by default? Aug 14 19:30:52 The mini data cache is irrelevant, if you are reading/writing from the standard cache in the standard way you get 266M memory accesses (32 bits) per second. Aug 14 19:31:13 That's one instruction or one read or one write. Aug 14 19:31:38 assuming the cache covers that memory location right? Aug 14 19:31:46 (For thumb code you still only get one instruction even though they are only 16 bits each, likewise for byte or halfword access). Aug 14 19:32:16 ... assuming the data/instruction is in the cache... Aug 14 19:32:36 ok, but I'm not currently accessing the cache or doing anything special but using pointers in C. Aug 14 19:32:53 The only time it isn't likely to be is outside a loop or if you do something silly - like tranversing a block of data bigger than 1/2 the dcache in size. Aug 14 19:33:22 So for optimization purposes what size data chunks are optimal? 32bits right? Aug 14 19:33:28 I think it's 32k, (D cache), so if your data fits in 16K no problem. Aug 14 19:33:51 32 bits is the widest single cycle access - if you access by byte it takes 4x as long. Aug 14 19:34:00 well, part of it should fit as it's a 320x240x3 image so.. Aug 14 19:34:25 yeah, that's what I assumed. So I'm buffering to write 32bits. Aug 14 19:34:39 (buffering to a register buffer) Aug 14 19:34:47 So access no more than 16 lines of the image at once... Aug 14 19:35:00 shouldn't be a problem. Aug 14 19:35:21 So long as you avoid things which fsck gcc - like the keyword 'register' Aug 14 19:35:38 Actually, I think they might have fixed it to ignore register. Aug 14 19:35:43 The problem is that particular routine is a YUV422 Raw to Planar conversion so I a Y U Y V .... Aug 14 19:35:58 really? Aug 14 19:35:58 That's only 2 lines Aug 14 19:36:30 I saw a good boost in speed by using the register command on certain vars. Aug 14 19:36:47 That would be a bug in gcc then. Aug 14 19:37:05 It should know far better than you what to put into a register and what to spill. Aug 14 19:37:19 After all... how many registers are there? Aug 14 19:37:27 2 lines? YUYV is only for 2 pixels so.. And I need to copy this to 3 seperate buffers (Y, U & V) Aug 14 19:37:43 not sure. I assumed 32 Aug 14 19:38:29 422 depends on the layout - I can't remember but I thought the U/V contributed to adjacent lines. Aug 14 19:39:05 YUYV = 422 and I there's one U/V for each 2 pixels. Aug 14 19:39:21 There are 15 general purpose registers, of which in practice maybe 10 are useable in a leaf procedure. Aug 14 19:39:45 ummm. I don't know but it seems to help speed things up. Aug 14 19:40:15 Yeah, well, it isn't necessarily the best ARM compiler in the world... Aug 14 19:40:28 but it's free! Aug 14 19:41:09 Yes, and the code is publically available so it can be fixed.... maybe.... Aug 14 19:41:11 bbiab - support Aug 14 19:41:16 s/support/supper/ Aug 14 19:41:18 later Aug 14 19:41:24 thanks btw Aug 14 21:10:25 is it just me or does ucslugs have a lot more packages than openslug? Aug 14 21:11:27 uclugc that is Aug 14 21:11:32 ucslugc Aug 14 21:12:01 no maybe not Aug 14 21:13:28 openslug has about 217 and ucslugc is only up to 184 so far Aug 14 21:15:55 ucslugc build completed Aug 14 21:21:58 does unslung or openslug support this: http://www.nongnu.org/gmediaserver/ ? Aug 14 21:23:10 oh it's only for music - never mind Aug 14 21:45:53 anyone care to chat about ARM/Thumb ? Aug 15 03:16:12 Hey all! Aug 15 03:16:37 Does anyone know where I can get a iptables-conntrack module for unslug 5.5? Aug 15 03:17:59 compile it yourself Aug 15 03:18:12 I'm doing that for openslug now Aug 15 03:23:52 uhm, can I compile it without compiling the entire kernel~? Aug 15 03:24:14 somehow, the thought of compiling an entire kernel ona na arm9 doesn't really appeal to me Aug 15 03:24:41 I don't do native compiles - I run a cross compiler on a gentoo PC Aug 15 03:24:59 you probably can natively compile kernel modules on unslung Aug 15 03:25:22 Checking MAsterMake right now... Aug 15 03:34:21 Your primary build machine is a ARM9 ?! Aug 15 03:35:51 no, the NSLU2 is Aug 15 03:36:03 I don't have a build machine right now Aug 15 03:36:39 Oh. Aug 15 03:36:48 I have a FreeBSD, one RHEL, one FC2 and a mandrake machine, but none of them have any cross-compilers isntalled Aug 15 03:36:51 NSLU2 is XScale, armv5b Aug 15 03:36:56 I was hoping to Aug 15 03:37:00 oh, oops Aug 15 03:37:16 Your Linux machine when using the master makefile will make the cross compiler. Aug 15 03:37:26 for both unslung and optware packages. Aug 15 03:37:34 unslung kernel that is Aug 15 03:38:04 odoc, hows your Slug doing? Aug 15 03:40:05 It's rather annoying, that I managed to install SSHd, bash, Apache, dnsmasq, iptables, rtl8150 (USB Ethernet) (all working) from modules, but stumble over a basic iptables modules Aug 15 03:40:30 well, that's because people painstakingly built and tested all the othre stuff Aug 15 03:41:11 You'd expect that a lot more people see the advantage of a noiseless, low-power-usage, highly modular NAT gateway Aug 15 03:41:50 what is the specific module you need? Aug 15 03:41:50 (Energy prices here have gone up aome 30% Aug 15 03:41:58 iptables-conntrack Aug 15 03:42:06 strike-2: yes, but not everyone needs connection tracking Aug 15 03:42:07 it's required for NAT Aug 15 03:42:40 interesting. It used to be there. Aug 15 03:42:47 ? Aug 15 03:42:53 (I'm currently building all that for Openslug, anyway, it's a unique facet of the NSLU2-Linux that everything needs to be done twice :)) Aug 15 03:43:10 ghehe Aug 15 03:43:23 that's what you get for maintaining 2 distro's ;-) Aug 15 03:43:32 Hmm Aug 15 03:43:40 # CONFIG_IP_NF_CONNTRACK is not set Aug 15 03:44:15 I don't mind building the thing, but it takes a few days' of re-learning kernel-compiling, cross-compiling and other stuff, unless MasterMakeFile is fully-automated Aug 15 03:44:28 Makefile is fully automated Aug 15 03:44:45 but then it takes about 2 days of sitting on your behind while it compiles everything automatically Aug 15 03:44:52 ouch Aug 15 03:44:59 well, 1 day Aug 15 03:45:04 uhm, need I remind you my basic playbox is a PII-300? Aug 15 03:45:25 What about nf_queue ? Aug 15 03:45:30 I _do_ have a P4 XEON 2.8 w/ 1 GB, but that machine's running production Aug 15 03:45:41 nf_queue? Aug 15 03:47:31 doesn't netfilter bite with iptables (and I don't see a netfilter package in ipkg) Aug 15 03:48:33 netfilter is the name for all of the packet filtering options in the kernel Aug 15 03:49:06 of which kernel-module-ip-tables and your connection tracking (and NAT) modules are just one 'option' Aug 15 03:50:49 03dyoung 07org.openembedded.nslu2-linux * r8aca482c... 10/packages/linux/unslung-kernel/defconfig: add conntrack and queue modules for iptables Aug 15 03:50:52 03dyoung 07org.openembedded.nslu2-linux * r05bbc684... 10/packages/linux/unslung-kernel_2.4.22.l2.3r29.bb: Oops forgot to bump the version Aug 15 03:52:44 it'll be in the feed shortly. Aug 15 04:03:34 w00t w00t Aug 15 04:03:51 is that allowed in here? Aug 15 04:03:53 * strike-2 hands dyoung a big bowl of smarties Aug 15 04:03:54 ;) Aug 15 04:04:10 strike-2: you may want to update this http://www.nslu2-linux.org/wiki/HowTo/EnableFirewall Aug 15 04:04:21 noting the paragraph: Currently, the connection tracking kernel module is not in the feed, so a stateful firewall is not possible. If/when that changes I will update this page. Aug 15 04:04:56 and if you're happy about how quickly the build machine has built your module for you, then remember that it's been paid for by community donations .... Aug 15 04:05:01 or maybe create a new page with that one as a template, as, to be fair, what you want to do (NAT) is quite different from just a packet filter for to secure local services Aug 15 04:08:25 rwhitby-away PayPal's on its way Aug 15 04:09:13 thanks - it's all useful to keep this rather complicated machine ticking along Aug 15 04:10:24 urgh, no xs to internet from this site (I took it to the office from home Aug 15 04:10:45 is there any way to DHCP/pump an interface Aug 15 04:10:52 it's usually hard-set Aug 15 04:14:10 n/m, I have a 2nd interface, I can set the ixp0 to DHCP using the Linksys pages Aug 15 04:14:27 ...and still shell in on the second card (eth0) Aug 15 04:14:43 strike-2: Bear in mind that, for the complexity of what you are doing, I would normally say to use Openslug Aug 15 04:14:54 strike-2: http://www.nslu2-linux.org/wiki/Info/DonationList updated Aug 15 04:15:02 yay! Aug 15 04:15:10 complex? Aug 15 04:15:14 but Openslug iptables isn't ready yet - so you will have to make do Aug 15 04:15:22 hehe Aug 15 04:15:27 strike-2: Its going into the feeds right now. Aug 15 04:15:44 dyoung held it hostage until the donation was confirmed :-) Aug 15 04:15:46 strike-2: comlpex in the the sense that Unslung has to co-operate with the Linksys software - and if you want to do something completely different, it may argue Aug 15 04:15:58 oh, dear Aug 15 04:16:01 As I've said before, <30min from package push to feed! Aug 15 04:16:04 Hey Aug 15 04:16:07 I get double-beeps every 5 seconds... Aug 15 04:16:08 ~lart rwhitby Aug 15 04:16:08 * jbot DoSes rwhitby Aug 15 04:16:18 oh, there's no cable in there... Aug 15 04:16:37 crud, there went my shell (on other card) Aug 15 04:16:41 But We do thank you very much for your kind donation. :-) Aug 15 04:17:11 You're very welcome Aug 15 04:17:26 lemme know if it works the way its supposed to. Aug 15 04:17:43 kernel-module-ip-conntrack_2.4.22.l2.3r29-r3_nslu2.ipk Aug 15 04:18:20 blaster8 that is true, there is always an uncertainty when adding modules to an existing kernel Aug 15 04:18:51 strike-2: not thinking of the kernel, thinking of the assumptions built into the Linksys software about what the slug will be doing and how it should be configured Aug 15 04:19:28 hehe Aug 15 04:19:33 the old Win9x issue Aug 15 04:19:33 strike-2: we have the Linksys kernel source, which the modules are built directly from Aug 15 04:19:48 exactly - Me, on a network? Hardly.. Aug 15 04:20:14 the only computer I ever had which couldn't pick up a DHCP broadcast reliably from a FreeBSD server Aug 15 04:20:15 vQuake would literally re-program my Verite video card while Windows wasn't looking Aug 15 04:20:45 running the ISC DHCP server, no less Aug 15 04:20:59 ? Aug 15 04:21:19 I usually get that when I hard-assign an IP to a Windoze b0x Aug 15 04:21:37 I had to manually assign it, in the end Aug 15 04:21:52 back to breaking my Openslug installation, anyway Aug 15 04:21:54 ISC doesn't check if the address is already alive (wrong MAC?), but Windows detects a MAC conflict and requests another IP Aug 15 04:23:05 So, is there anyone here who is up to the task of working on Unslung 6.0 firmware ? Aug 15 04:23:23 so did the conntrack module load/work ? Aug 15 04:25:30 just a sec, I need to get the thing on-line first Aug 15 04:25:43 need 2 get a bigger hub, brb Aug 15 04:26:29 rwhitby: What are you looking for in Unslung 6? Aug 15 04:26:51 apart from as much R63 compatibility as seems sensible Aug 15 05:01:16 03nail 07org.openembedded.nslu2-linux * rb00bca80... 10/packages/linux/ (nslu2-kernel_2.6.12.2.bb openslug-kernel-2.6.12.2/defconfig): Add IPTABLES config and CONFIG_BRIDGE Aug 15 05:01:26 03nail 07org.openembedded.nslu2-linux * r72ced9eb... 10/ (2 files in 2 dirs): Add iptables to openslug feed Aug 15 05:02:07 yay, back Aug 15 05:02:36 Unslung 6 will basically be the features on Unslung 5.5, but based on R63 firmware. There will be a 6.x based on the R63 binary kernel, and then later a 6.x based on the R63 source code release. Aug 15 05:04:14 The 6.x based on the binary kernel will necessarily be missing any Unslung-specific kernel features (/proc/bus/usb is one example), and so some people will need to stick with 5.5 until Linksys releases the R63 kernel source code. Aug 15 05:08:20 YAY! internet access fixed Aug 15 05:09:38 kernel-module-ip-conntrack - 2.4.22.l2.3r29-r3 - ip-conntrack kernel module Aug 15 05:14:43 hey [g2] Aug 15 05:15:12 <[g2]> morning sunshine! Aug 15 05:15:12 morning sunshine. Aug 15 05:15:16 echo Aug 15 05:15:37 * [g2] hugs dyoung rwhitby Aug 15 05:15:51 * rwhitby looks sideways at [g2] :-) Aug 15 05:16:14 <[g2]> T1 guy is supposed to be here in the next 4-8 hours Aug 15 05:16:34 heh - a slug webserver on a T1 connection - why does that sound just not right ... Aug 15 05:16:57 <[g2]> yeah... it should be DS3 but I can't afford it :) Aug 15 05:17:08 <[g2]> or even OC3 Aug 15 05:17:12 [g2]: which web server will the loftweb slug be running? Aug 15 05:17:15 <[g2]> then I could use 2 slugs Aug 15 05:17:48 <[g2]> right now I'm leaning toward thttpd Aug 15 05:17:57 <[g2]> I'll probably run apache2 also Aug 15 05:18:17 [g2]: why not go for OC-12 .. Aug 15 05:18:39 [g2]: do you have apache2 building for openslug ? Aug 15 05:19:26 <[g2]> DaKa was testing yesterday with siege and a turbo slug was serving up nearly line rate Aug 15 05:20:02 <[g2]> about a month ago I had apache2/openssl/php5 all running on OpenSlug (built natively) Aug 15 05:20:25 <[g2]> it's not an ipkg :( Aug 15 05:20:45 <[g2]> but it's pretty straight forward to build Aug 15 05:21:52 <[g2]> I'll have multiple ip addrs, so I'll probably play with different servers as things get going and run multiple servers Aug 15 05:22:22 So if you've got a native build of apache2/openssl/php5, then how come no-one (including our new OpenSlug native feed manager) has put them in the native feed? Aug 15 05:23:00 [g2]: I apex'd a slug yesterday ;) Aug 15 05:23:25 <[g2]> NAiL, congrats! Aug 15 05:23:35 <[g2]> rwhitby, it's a straight native compile Aug 15 05:23:45 <[g2]> not a native-build Aug 15 05:23:47 Dunno if it boots from flash or jffs and stuff yet though ;) Aug 15 05:24:15 <[g2]> NAiL, by default it boots from the kernel in flash Aug 15 05:24:23 [g2]: yeah, but that should be really easy to turn into a .bb ... Aug 15 05:24:50 DaKa2 has that more or less worked out, doesn't he? Aug 15 05:25:07 <[g2]> he was working on apache2 last I chatted Aug 15 05:25:40 [g2] I think the time is mostly taken getting apache integrated into the OE system rather than just getting to to build Aug 15 05:27:08 <[g2]> blaster8, nod. I've been building/testing native compiles first then I'd know "what" to try and get in OE Aug 15 05:29:01 [g2]: so at what point do you think that apex will become mainsteam enough that it is discussed here instead of #openjtag ? Aug 15 05:29:14 [g2] exactly - so far I have been lucky that (at least on old version of) everything I need is already a .bb in OE Aug 15 05:31:10 03rwhitby 07org.openembedded.nslu2-linux * rc8d07f5c... 10/packages/nslu2-binary-only/nslu2-linksys-ramdisk_2.3r63.bb: Updated the location of the R63 rootfs which has the ufsd.o module removed (it's a commercially licensed module). Aug 15 05:31:36 <[g2]> rwhitby, I thing APEX should go mainstream when one of the following occurs Aug 15 05:31:49 <[g2]> 1) usb-serial is added to it Aug 15 05:32:01 <[g2]> 2) usb-networking is added to it Aug 15 05:32:25 <[g2]> beewoolie has been doing networking support for the sharp line Aug 15 05:32:55 <[g2]> We'd need to add pci/usb code to support the networking Aug 15 05:33:12 <[g2]> I think going the usb-serial console support is probably far easier Aug 15 05:33:35 <[g2]> 3) Require the serial console Aug 15 05:34:28 <[g2]> Option 3) may have some legs but it's secondary to me launching the "loft" Aug 15 05:35:19 * dyoung off to lofty dreams Aug 15 05:40:04 <[g2]> sweet dreams Aug 15 05:41:58 [g2] Good list - I look forward to a stable APEX Aug 15 05:42:26 <[g2]> blaster8, APEX is stable Aug 15 05:42:52 but there is not release packaged for the NSLU2 Aug 15 05:42:57 not=no Aug 15 05:43:56 Uhm, guyz.... Aug 15 05:43:58 [g2]: there is an issue that an apex built inside OE doesn't work, but one built from crosstool does work. Bewoolie is looking into it ... Aug 15 05:44:54 Has anyonehere played with ip_masq on unslung? Aug 15 05:45:28 not that I've heard of ... Aug 15 05:45:41 <[g2]> rwhitby, AH... I wasn't aware of that Aug 15 05:46:15 [g2]: familiar with ARM/Thumb at all ? Aug 15 05:46:17 <[g2]> rwhitby, does include just running the OE built compiler/tools Aug 15 05:46:22 iptables complains it doesn;t know table nat (understandable), but I need fo figure out what to hook the nat chain into Aug 15 05:46:28 <[g2]> hey vapier ! Aug 15 05:47:24 <[g2]> vapier, you've got a ARM/Thumb question ? Aug 15 05:47:26 strike-2: do you have the NAT kernel modules? Aug 15 05:47:45 <[g2]> the guys in #oe could probably answer it Aug 15 05:47:49 strike-2: btw I have just tested the full iptables implementation on Openslug Aug 15 05:48:03 all 47 kernel modules Aug 15 05:48:17 Module Size Used by Tainted: P Aug 15 05:48:17 ip_conntrack 18408 0 Aug 15 05:48:17 iptable_filter 1516 1 Aug 15 05:48:17 ip_tables 12912 1 [iptable_filter] Aug 15 05:48:31 I believe conntrack is the only module required for nat? Aug 15 05:48:37 [g2]: know which arm cpu first supported thumb ? Aug 15 05:48:57 <[g2]> blaster8, Great. Aug 15 05:48:59 * NAiL wonders what Thumb is Aug 15 05:49:04 root@NSLUg:~$ iptables -t nat -A POSTROUTING -o ixp0 -j MASQUERADE Aug 15 05:49:04 iptables v1.2.11: can't initialize iptables table `nat': Table does not exist (do you need to insmod?) Aug 15 05:49:04 Perhaps iptables or your kernel needs to be upgraded. Aug 15 05:49:19 <[g2]> Thumb is a 16-bit instruction set Aug 15 05:49:33 blaster8, you got it to work right on OpenSlug? Aug 15 05:49:58 NAiL: http://www.embedded.com/shared/printableArticle.jhtml?articleID=15200241 Aug 15 05:52:17 <[g2]> vapier, jbowler is the only guys that I think would know the original processor supporting Thumb of the top of his head Aug 15 05:52:32 strike-2: I can't directly test NAT but I tested stateful firewall rules and modprobed all of the NAT modules successfully - the patches are making their way into Openslug HEAD Aug 15 05:52:54 [g2]: yeah, i was hoping he'd be around, but he hasnt popped in yet :) Aug 15 05:53:55 <[g2]> I can ask in #oe Aug 15 05:54:03 <[g2]> or #elinux Aug 15 05:54:36 nah i'll just wait Aug 15 05:54:46 he's the reason i'm askin in the first place ;p Aug 15 05:54:56 <[g2]> I just asked in #oe Aug 15 05:55:03 aha Aug 15 05:55:04 <[g2]> I couldn't contain myslef Aug 15 05:55:10 so *that's Aug 15 05:55:18 <[g2]> vapier, piqued my interest Aug 15 05:55:25 what the CPU in one of my routers does... I noticed there was a T :P Aug 15 05:56:19 03rwhitby 07org.openembedded.nslu2-linux * r3e3bbfe0... 10/packages/nslu2-binary-only/ (4 files): Updated unslung-binary-kernel distro to use the expurgated version of the Linksys R63 firmware. Aug 15 05:57:54 Thtats a big word Aug 15 05:59:21 blaster8 did you simply build / load all modules? Aug 15 06:00:51 strike-2: built the modules on my cross-compile machine - sftped the ipks over to the NSLU2 and installed them manually (ipkg install ./kernelmodule.ipk) then installed iptables user-space tools (also cross-compiled) and tested a few simple rules Aug 15 06:01:16 will that also work on unslug? Aug 15 06:01:38 strike-2: no Aug 15 06:01:39 I have a nice, beefy remote FC2 machine, but it's running production Aug 15 06:01:44 d'oh! Aug 15 06:01:51 strike-2: totally different kernel - different config, the lot Aug 15 06:01:57 crud Aug 15 06:02:01 got caught out by that Aug 15 06:02:13 If i use MakeMaster, will I be able to build it? Aug 15 06:02:22 unslung-binary-kernel distro will now build from OE - ready for some early alpha testers. Aug 15 06:02:26 yes Aug 15 06:02:42 strike-2: cat /proc/sys/net/ipv4/ip_forward Aug 15 06:02:54 should be 1 Aug 15 06:03:02 uhm, 0 Aug 15 06:03:08 aah Aug 15 06:03:11 cat 1 > ip_forward? Aug 15 06:03:36 echo 1 >> /proc/sys/net/ipv4/ip_forward Aug 15 06:03:38 uhm, s/cat/echo Aug 15 06:03:43 I think Aug 15 06:03:47 is 1 now Aug 15 06:04:12 I think it wants a ip_nat module Aug 15 06:04:20 iptables v1.2.11: can't initialize iptables table `nat': Table does not exist (do you need to insmod?) Aug 15 06:04:21 Perhaps iptables or your kernel needs to be upgraded. Aug 15 06:04:46 cd /proc Aug 15 06:05:01 yes? Aug 15 06:05:09 wrong window Aug 15 06:05:13 ah Aug 15 06:05:37 iptables v1.2.11: can't initialize iptables table `nat': Table does not exist (do you need to insmod?) Aug 15 06:05:38 Perhaps iptables or your kernel needs to be upgraded. Aug 15 06:05:41 oops Aug 15 06:05:45 alias erp say 'erp' means: 'I just typed that in the wrong window (again). Regardless, I still maintain that my ircii/bitchx irc client is *far* superior to any gooey, mircy windows client. The only thing I can't figure out is how come the only people I ever see 'erp'ing are ircii/bitchx users. Aug 15 06:05:46 dyoung: can you add some more modules to the unslung feed? Aug 15 06:06:41 If iptables needs more modules, I'll be broke before the firewall's up Aug 15 06:06:42 ask someone else, I'm headed for sleep now Aug 15 06:06:56 have fun, dyoung Aug 15 06:07:03 but if it can wait until tommrow, I can do it. Aug 15 06:07:11 I got some time to spare Aug 15 06:07:23 <[g2]> ~rules Aug 15 06:07:23 rules are a requirement for having exceptions Aug 15 06:07:33 <[g2]> ~nslu2_rules Aug 15 06:07:59 ~nslu2-rules Aug 15 06:07:59 i heard nslu2-rules is 1) You must search the wiki before asking questions on IRC or mailing list. 2) You must follow the README steps precisely when flashing firmware. 3) Those who ask the questions, update the wiki with the answers. 4) Those who complain about doco, update the wiki to improve it. 5) Friends don't let friends flash custom firmware without ... Aug 15 06:09:00 <[g2]> we need to setup an iptables/ebtables howto for openslug Aug 15 06:09:24 [g2] someone who actually knows iptables should do it Aug 15 06:09:44 suffice to say that everything that needs to be there for iptables to work, is now in place in HEAD Aug 15 06:09:49 (ebtables pending) Aug 15 06:10:23 http://www.gentoo.org/doc/en/security/security-handbook.xml?part=1&chap=12 Aug 15 06:10:30 is a good start for advanced firewalls Aug 15 06:13:10 <[g2]> thx Aug 15 06:13:25 <[g2]> I'll be publishing my set up after I get it going Aug 15 06:14:37 <[g2]> I'll have a transparent logging bridge-fw between my web server and internet Aug 15 06:15:23 <[g2]> the real beauty of the setup will be JTAG access to verify/reprogram everything from the ground up Aug 15 06:15:33 Lat time I've set up a Linux firewall, RedHat 6 was new. 'Nuff said Aug 15 06:15:50 s/lat/last Aug 15 06:16:42 the good news: ipchains was also written by Rusty Russell Aug 15 06:16:51 the bad new: that was all the good news Aug 15 06:19:02 iek, nen Belg Aug 15 06:20:31 blaster8 Hmm, my FC2 box doesn't even have gcc installed... Aug 15 06:21:02 hehe :) Aug 15 06:21:08 CPU0: Intel(R) Xeon(TM) CPU 2.80GHz stepping 09 Aug 15 06:21:10 everyone loves Red Hat Aug 15 06:21:13 shouldn't take too long Aug 15 06:21:23 if it works at all Aug 15 06:21:26 CPU1: Intel P4/Xeon Extended MCE MSRs (12) available Aug 15 06:21:39 had someone, unable to do the make on FC4 yesterday Aug 15 06:21:40 j/l, it's a P4 HT Aug 15 06:21:45 wtf? Aug 15 06:21:51 well, it's FC2... Aug 15 06:21:59 hopefully, you will be fine Aug 15 06:22:10 I needed a decent Linux and didn't want to splash on RHEL Aug 15 06:22:14 Red Hat: The company that brought the world GCC 2.96 Aug 15 06:22:19 # CONFIG_IP_NF_NAT is not set Aug 15 06:22:19 lol Aug 15 06:22:33 if noone else fixes it in the next 16 hours I'll get to it then. Aug 15 06:22:35 otherwise known as the don't use "2.96 or your build will fail" 2.96 Aug 15 06:22:36 night Aug 15 06:22:44 g'night dyoung Aug 15 06:22:47 night Aug 15 06:23:03 you can get RHEL free, anyway Aug 15 06:23:16 http://www.centos.org/ as long as you don't need support Aug 15 06:23:18 yes, but not the updates Aug 15 06:23:27 the community are really quite to update it Aug 15 06:23:35 and FC has no upgrade path anyway Aug 15 06:23:41 (as I'm sure you know) Aug 15 06:23:45 Maybe if I use my _legal_ RHEL to download them and save them, then let the 'other' RHEL load those Aug 15 06:23:52 no Aug 15 06:24:07 they're machine-specific? Aug 15 06:24:13 you don't need to Aug 15 06:24:17 ? Aug 15 06:24:22 centos release all the updates you need themselves Aug 15 06:24:39 is CentOS production-quality alreday? Aug 15 06:25:00 well, it is just RHEL with the words Red Hat removed to avoid copyright restrictions Aug 15 06:25:10 the FC2 machine is running an imap server with mailboxen that makes a BSD buildworld look like peanuts Aug 15 06:25:11 "CentOS is an Enterprise-class Linux Distribution derived from sources freely provided to the public by a prominent North American Enterprise Linux vendor." Aug 15 06:25:35 I'll look into it with the next server-upgrade Aug 15 06:25:48 but when FC2 is not maintained any more and security hoes are found, what will you do? Aug 15 06:25:54 hehe hoes = holes Aug 15 06:25:59 ah,ok Aug 15 06:26:20 well, only 2 ports (https, imaps) end up on the box Aug 15 06:26:26 If you want to run only OS software, I'm in the 'just run Debian' camp Aug 15 06:26:43 nope, it needs to run Bynari InSight Server Aug 15 06:26:56 http://www.bynari.net/ Aug 15 06:27:01 fair enough Aug 15 06:27:24 it's a combination of Cyrus, Postfix, slapd, apache, all strung together with a decent management UI Aug 15 06:27:42 ...and an Outlook Connector with (almost) full Exchange functionality Aug 15 06:27:56 neat Aug 15 06:27:59 aight Aug 15 06:28:06 hope they donate back to those projects Aug 15 06:28:57 I think they do, but I never really checked Aug 15 06:29:14 they even have a sexy WebClient (http://www.bynari.net/index.php?id=3545) Aug 15 06:29:50 neat - there aren't enough good IMAP projects Aug 15 06:30:25 A friend of mine is working on a Dutch government-sponsored Exchange clone Aug 15 06:30:51 they base their work on MAPI, they even wrote a Linux MAPI stack Aug 15 06:53:27 strike-2: if you like I will try to find an exact list of the 2.4 modules you want enabled Aug 15 07:16:26 anybody has ideas: I'm getting this: monotone: trying 3-way merge Aug 15 07:16:26 monotone: misuse: unable to resolve file conflict 'packages/linux/linux-openzaurus_2.6.13-rc3-mm1.bb' -> 'packages/linux/linux-openzaurus_2.6.13-rc5-mm1.bb' vs. 'packages/linux/linux-openzaurus_2.6.13-rc3-mm3.bb' Aug 15 07:18:01 are you using the master makefile? Aug 15 07:19:20 VoodooZ_Work: are you using 0.22 and the merge --lca switch? Aug 15 07:19:33 tried it but it didn't seem to like it. **** ENDING LOGGING AT Mon Aug 15 07:19:38 2005