**** BEGIN LOGGING AT Fri Mar 03 10:59:57 2006 Mar 03 12:43:06 * philpem is trying to install his new hotplug package with ipkg, but ipkg has other plans Mar 03 12:43:36 "An error ocurred, return value: 4" / "Cannot find package hotplug." Mar 03 12:44:15 i gave it a URL, it grabbed the file, then it refused to install it... Mar 03 13:19:06 dwery: It's just assumptions based on the (n)ftl references in the kernel Mar 03 13:20:00 yay, sparkfun has sent my package Mar 03 13:22:42 philpem: How far did you get on that dsl-thingy you worked on? Mar 03 14:10:02 NAiL: eciadsl? Mar 03 14:10:06 it works :) Mar 03 14:10:38 i had to compile the n_hdlc kernel module (CONFIG_N_HDLC=m and CONFIG_SERIAL_NONSTANDARD=y), but that's it. Mar 03 14:11:27 what is a little more troublesome is that my hotplug scripts won't work with hotplug-ng. Mar 03 14:12:28 but anyway.. i've just done the turboslug hack (remove a resistor to gain a 2x speedup - nice!) and my version of the autopower hack Mar 03 14:13:23 oh, how is your version of the hack? Mar 03 14:13:39 and could you send me a diff of the changes necessary to get eciadsl to work? Mar 03 14:16:05 NAiL: in the kernel config file, search for "CONFIG_SERIAL_NONSTANDARD". it'll be one of those 'x is turned off' comments. add "CONFIG_SERIAL_NONSTANDARD=y" and "CONFIG_N_HDLC=m" below the comment. delete the comment Mar 03 14:16:26 save, then compile the kernel and n_hdlc will build as a package (kernel-module-n-hdlc) Mar 03 14:18:28 ok, I'll see if I can get this into 3.6-beta. Not sure if it'll be possible to add to 2.7-beta Mar 03 14:19:09 fair enough Mar 03 14:19:24 i've got a 2.7beta build here, with the relevant packages. Mar 03 14:20:32 The config_serial_nonstandard... does that change the kernel? Mar 03 14:21:04 because there can't be any changes to the kernel, since that requires reflashing to use certain modules, which would be a headache Mar 03 14:21:23 NAiL: not sure. i think it just enables support for a few additional modules. Mar 03 14:21:47 n_hdlc does work fine with the stock kernel flashimage Mar 03 14:21:56 cool Mar 03 14:22:22 i know because i'm using a stock 2.7 kernel+initrd with n_hdlc at the moment :) Mar 03 14:22:37 i've got the diff here if you want it... Mar 03 14:22:56 That's be nice. Send it either to the mailinglist or nail @ nslu2-linux.org Mar 03 14:23:31 which mailinglist - nslu2-linux@yahoogroups/ Mar 03 14:23:31 ? Mar 03 14:25:42 yeah Mar 03 14:26:06 preferrably the ml, so others can take up the ball if I'm not able to somehow ;) Mar 03 14:31:52 sent. Mar 03 14:32:12 thanks Mar 03 14:32:27 wow, an amazing *3 lines* of patch code. my eciadsl-0.11.bb is bigger than that, even though it just pulls in autoconf :) Mar 03 14:33:13 hehe Mar 03 14:33:40 who's responsible for http://ipkgfind.nslu2-linux.org/ anyway? the free-text search is broken. Mar 03 14:34:04 I don't think ipkgfind has been updated for a while Mar 03 14:34:15 ah. that explains the 'Warning: mysql_num_rows(): supplied argument is not a valid MySQL result resource in /home/nslu/ipkgfind/htdocs/result.php on line 127' :) Mar 03 14:37:15 NAiL: I have a really horrible ugly hack of a patch and a change to the .bb file for zd1211 to submit (allows zd1211 to build for Unslung) - you're listed as the package maintainer. Do you want me to sent you the diffs, or post them to the linux or developers lists? Mar 03 14:37:22 03repvik * r278 10/releases/slugos-3.6-beta/openembedded/ (19 files in 5 dirs): Upstream upgrades Mar 03 14:37:37 mwester: send to dev list, cc me Mar 03 14:37:46 Makes it easier for me to track ;) Mar 03 14:37:46 Ok :) Mar 03 14:48:33 NAiL, we're keeping you busy! :) Mar 03 14:49:52 definitely Mar 03 14:53:20 philpem: Ok, I see your mail. I'll look into it in a little while and see if I can get it to compile on all releases. Mar 03 14:53:40 Got a headache for some reason Mar 03 14:58:38 dwery: Heh, reading stuff on the other freecom products. Their lan/usb2 enclosure has this text: "No IP address so the drive is invisible to anyone outside of your network: Your data is 100% secure!" Mar 03 15:00:28 NAiL: ROFL! Mar 03 15:01:01 they'll be using IPX then :) Mar 03 15:01:46 or worst NetBEUI! Mar 03 15:01:56 oh yes, IPX makes it secure! :) Mar 03 15:02:37 I dunno what they're using, but claiming 100% security is just sad Mar 03 15:03:03 A blatant invitation for hackers Mar 03 15:03:36 Hmm. That kind of marketing statement is almost a challenge for the hackers! Mar 03 15:03:54 yeah Mar 03 15:04:12 Kindof tempted to sue them for false advertising .-P Mar 03 15:07:59 "100% secure? I can see its network traffic on Ethereal!" Mar 03 15:11:03 I kinda like their FSG-3 product though Mar 03 15:11:46 hmm, it's based on ximeta ndas. Mar 03 15:12:05 http://www.ximeta.com/support/faq/index.php -> "NetDisk does not use IP Addressing and does not run TCP/IP. We have developed a very unique technology called NDAS- Network Direct Attached Storage that is designed to run with lower overhead and eliminates the need for a processor. In turn, this makes speeds much faster and your data more secure." Mar 03 15:12:30 so in other words, it is to NAS what winmodems are to modems. Mar 03 15:13:13 bleh Mar 03 15:13:24 I bet it's just some iSCSI-spinoff Mar 03 15:13:31 or ATAoE Mar 03 15:13:43 just not using TCP/IP Mar 03 15:18:03 "...elminates the need for a processor..." ??? say what? Mar 03 15:18:38 c/elminates/eliminates/ Mar 03 15:18:49 so basically it's an ethernet frame decoder with some form of recognition circuitry slapped onto the front end. ATA over ethernet in other words. Mar 03 15:19:20 i'd love to get my mitts on one of those things for a few hours. ethereal + homebrew packet decoder :) Mar 03 15:19:29 So it doesn't work on anything other than stock standard ethernet... Mar 03 15:20:00 And certainly not over wifi, surely... Mar 03 15:20:42 That's exactly what I'm thinking. Which would knock out every member of my family, since my servers and my tower are the only wired systems. Mar 03 15:21:07 Doesn't sound like a good direction to me; everything is moving to wireless now... Mar 03 15:21:40 NAiL: Patch sent, I hope I did it correctly! Mar 03 15:21:52 mwester: at the moment, most of my lan is 100baseT, with a netgear wlan bridge hooked up. Mar 03 15:22:20 mwester: The only necessary change is the "if kernel-version = the unslung-kernel", right? Mar 03 15:22:56 And the change to the .bb file -- the extra set of quotes broke the build Mar 03 15:23:10 odd Mar 03 15:24:22 Yeah - really odd. I would have thought they would have been stripped out, but the compile log made it pretty clear that Make was tryign to execute a command named "ccache gcc -option -option..." instead of just "ccache" :( Mar 03 15:24:57 It builds on slugos, you see Mar 03 15:26:23 NAiL: :-D Mar 03 15:29:12 I'm open for other ideas! But other than a different set of paths being passed in, I dont' think there's any other place unique to Unslung. Unless Unslung already quotes the args while slugos doesn't? Mar 03 15:29:33 dwery: yes? :)' Mar 03 15:30:05 NAiL: the no TCP/IP thingy :) Mar 03 15:30:14 i've read it only now Mar 03 15:30:28 hehe Mar 03 15:35:50 AARG! the drive in my usb caddy just died! Mar 03 15:36:05 that sucks Mar 03 15:36:18 It does seem to be an issue though Mar 03 15:36:26 now i know why the server i pulled it from kept freezing - the drive randomly drops off the ATA bus. Mar 03 15:36:40 I wonder what causes drives in USB-enclosures to die faster than other drives.... Mar 03 15:36:48 shitty cooling. Mar 03 15:37:12 shove a couple of 40mm fans in the drive caddy and see if that helps. Mar 03 15:37:12 maybe Mar 03 15:37:25 I think my enclosure is rather well-ventilated though Mar 03 15:37:57 ah, there we go. nothing a quick power-off/power-on couldn't cure :) Mar 03 15:38:06 heh Mar 03 15:39:37 Still, regardless of enclosure issues, there's much that better USB->IDE chips could do. Just monitoring the SMART status could alert to overtemp problems (and I bet it wouldn't be too hard for a USB chipset to powercycle philpem's HD as a 'last resort' error recovery means...) Mar 03 15:40:14 In that case I'd rather prefer a "error" led Mar 03 15:40:43 Even if it's "General error", it's a nice way to let the user know that the drive is not-ok. Mar 03 15:40:44 and a beeper. not one as annoying as the one on my asus mobo though - that thing plays what sounds like beethoven when the cpu overheats :) Mar 03 15:40:53 haha Mar 03 15:40:59 hmm Mar 03 15:41:10 Maybe I should build one myself Mar 03 15:41:14 That'd be a nice project ;) Mar 03 15:41:32 i found that out when i knocked the fan power cable off while i was working on it. fired up UT and got beethoven coming from the motherboard beeper :) Mar 03 15:41:51 Mine won't boot without the fan... Mar 03 15:42:04 * NAiL doesn't have a fan ;) Mar 03 15:42:25 I use watercooling :-P Mar 03 15:42:27 a few seconds later Probe woke up and told me the cpufan was dead and that the cpu was up to 60 C... Mar 03 15:42:43 at which point i noticed the power cable dangling in free space :P Mar 03 15:42:50 Sheeesh! Watercooling - I HATE PLUMBING! :D Mar 03 15:43:55 hehe Mar 03 15:44:13 mwester: same here... the idea of large quantities of water being one split seal away from drenching my machine does not appeal to me. Mar 03 15:44:57 hehe Mar 03 15:45:03 hence why cheetah (ath64-3200+/512m/100g/gf6600gt) is cooled the traditional way.. with lots of big fans :P Mar 03 15:46:42 Hmmmm. That's odd - I think I do not understand something important. "build optware" builds stuff for various systems/kernels (to generalize), "build unslung" builds the core Unslung kernel --- what do I build and where do I specify if I need to build Unlsung-kernel-specific packages (or do they just get added to the "make unslung" build)? Mar 03 15:49:03 * mwester wanders off to see if he can find the link to the oil-cooled PC experiment... Mar 03 15:49:32 hehe Mar 03 15:49:37 Mineral oil Mar 03 15:49:42 :D http://www.tomshardware.com/2006/01/09/strip_out_the_fans/ Mar 03 15:50:18 My Ath64-3700+ idles at 26C and maxes out at 32C with watercooling :) Mar 03 15:52:44 29 idle, 40 max with cool-n-quiet enabled. Mar 03 15:52:56 37 idle, 42 max with cnq disabled Mar 03 16:08:22 woo, getting ever closer to the bit where i have to strip down the old server and pilfer its hard drive :) Mar 03 16:22:47 then i get the lovely job of moving the root partition from the 2gig drive onto the 20gig. Mar 03 16:23:53 what i'll probably do is put it in the server, then move all the stuff on the server's 20gig data drive onto the 4gig boot drive. then partition, dd, resize2fs and done. Mar 03 16:29:06 03repvik * r279 10/releases/slugos-3.6-beta/openembedded/packages/icecast/icecast_2.2.0.bb: Fix DEPENDS Mar 03 16:45:50 hi can anyone help me here with syslog ?? Mar 03 16:47:58 kami22: whats the problem? Mar 03 16:51:34 i am using Openslug and cron.info is always posting messages in my /var/log/messages how can i stop this i thought about a solution to transfer all cron.info to /dev/nul but how? Mar 03 18:08:32 dwery: btw, do you know how hard it would be to add led triggers to the scsi subsystem? Mar 03 18:09:50 or to USB Mar 03 18:10:00 to make the disk-1 and disk-2 leds make more sense Mar 03 18:48:51 anyone experience with the kernel image provided by Debian from the linux-image-2.6.15-1-nslu2 package in testing? Mar 03 18:49:05 is it possible to actually is this one, and if so, how? Mar 03 18:49:15 i was thinking that the following should work: Mar 03 18:49:30 afaik, that kernel contains a subset of the patches in the current OE tree Mar 03 18:49:37 boot to RAM (touch /.recovery, reboot, turnup ram, reboot) Mar 03 18:49:46 reflash -k /mnt/boot/vmlinuz... Mar 03 18:50:02 cp modules from the included initrd to the flash Mar 03 18:50:36 advantage is that on the disk i then have a complete set of modules available Mar 03 18:51:23 tbm: Do you know how well that kernel works on the slug? Mar 03 18:52:45 NAiL: not too hard. you just have to identify the right place Mar 03 18:53:04 hmm Mar 03 18:53:05 i'd add them in the scsi disk subsystem Mar 03 18:53:11 yeah Mar 03 18:53:28 well... people are connecting cd/dvd-rw to their slugs ;) Mar 03 18:55:23 NAiL: the kernel lacks RTC and LEDS support but works fine apart from that Mar 03 18:55:43 ah Mar 03 18:55:53 I don't know exactly what you have to do to use it on an non-debian system tho Mar 03 18:56:15 flash it, copy /lib/modules, and then get the IXP400 stuff from somewhere Mar 03 18:56:32 we should probably put them (copiled against the debian kernel) as a .tar on slug-firmware Mar 03 19:00:40 but someone should write a short howto about this... rwhitby maybe Mar 03 19:24:18 are the IXP400 modules versioned? Mar 03 19:24:32 can modules build against 2.6.16 be used on 2.6.15? Mar 03 19:24:52 no Mar 03 19:24:55 Atleast, not yet Mar 03 19:25:28 k, so if i really want to use the Debian 2.6.15 kernel i first have to get them properly compiled against that kernel Mar 03 19:25:38 CONFIG_MODVERSIONS is enabled on the kernel in cvs, which will probably make it into slugos-3.6-beta Mar 03 22:12:35 hello, I have a problem with the OpenSlug build system. Is there anything known to be broken in the current version? Mar 03 22:12:45 libtool-native-1.5.10-r3 fails to build Mar 03 22:13:01 configure: error: source directory already configured; run "make distclean" there first Mar 03 22:13:50 03repvik * r280 10/releases/OpenSlug-2.7-beta/openembedded/packages/linux/ (2 files in 2 dirs): Add HDLC as module Mar 03 22:20:51 03repvik * r281 10/releases/OpenSlug-2.7-beta/ (7 files in 5 dirs): Add eciadsl package Mar 03 22:28:21 03repvik * r282 10/releases/slugos-3.6-beta/openembedded/ (8 files in 6 dirs): Add eciadsl and HDLC support Mar 03 22:29:59 03repvik * r283 10/releases/slugos-3.6-beta/openembedded/packages/eciadsl/eciadsl_0.11.bb: Update RDEPENDS Mar 03 22:30:15 03repvik * r284 10/releases/OpenSlug-2.7-beta/openembedded/packages/eciadsl/eciadsl_0.11.bb: Update RDEPENDS Mar 03 22:34:52 can nobody help? Mar 03 22:41:26 03orepvik * 10kernel/2.6.16/defconfig: defconfig: Add PPP_SYNC_TTY, MPPE, and HDLC Mar 03 22:43:54 03orepvik * 10kernel/2.6.16/KERNEL: Update to rc5-git7 Mar 03 23:01:53 mwester: ping? Mar 03 23:03:58 NAiL: pong (sorta, new irc client which only kinda works... :( Mar 03 23:04:26 heh, Mar 03 23:04:51 The KERN26=y hack should be fixed though... It should just properly detect 2.4.x vs 2.6.x Mar 03 23:05:14 I agree, but don't know the "correct" way to do so. Mar 03 23:05:28 try asking in OE Mar 03 23:06:37 Of course, the original KERN26=y is the original Makefile; we can't change that (or is that what you wish to have changed with a more comprehensive patch?) Mar 03 23:07:13 what we can do is to change the Makefile to read KERN26 ?= y, and set the variable in the .bb instead Mar 03 23:07:20 ie, like it is done with CC and LD Mar 03 23:08:26 Well, that we can do - but it just moves the ugliness up to the .bb file where we have to set that var based on the distro being built, right? Mar 03 23:08:34 yes Mar 03 23:08:45 but that makes it possible to actually upstream the Makefile patch Mar 03 23:08:54 aaah! Mar 03 23:08:57 I see. Mar 03 23:09:29 The Makefile will simply default to "sane" values, unless otherwise specified in the environment Mar 03 23:11:38 Then I'd need to figure out how to set KERN26 in the .bb file based on distro -- any suggestions? Or perhaps I'll check other .bb files for hints, this cannot be the only place where such a decision is made based on the distro or machine... Mar 03 23:12:12 no, hence asking in #oe isn't a bad idea. Someone once gave me a .bb as an example of how to do it. Mar 03 23:12:22 Unfortunately, I never looked at it, nor do I remember which one. Mar 03 23:12:52 Ok - I'll mess with that. How about the quotes - can that change in the .bb file go in as is? Mar 03 23:13:04 I've committed that change already Mar 03 23:13:12 So it's only the KERN26 change left Mar 03 23:13:29 * NAiL has been on a wild committing spree this evening Mar 03 23:13:34 Excellent. I'll post another patch to the list when I find a somewhat more elegant way to resolve that. Thanks! Mar 03 23:14:29 np Mar 03 23:14:37 * mwester has noticed the steady stream of commits... Mar 03 23:14:55 yeah, there's a few different repos for me to update Mar 03 23:15:09 CVS, OE, 2.7-beta and 3.6-beta ;) Mar 03 23:17:12 Hehehe - and the only thing I've been working on seems to be the "nearly-obsolete" stuff like the 2.4 kernel! Mar 03 23:17:56 The 2.4 kernel (RedHat) and Windows NT4 at work during the day... I seem to specialize in "antiquities" :D Mar 03 23:18:03 ouch Mar 03 23:18:17 And I'm trying out 2.6.16-rc5-git7 ;) Mar 03 23:19:11 I bet you see kernel panics more often than most that way! Mar 03 23:19:37 well, I see a few, yes Mar 03 23:19:50 mostly because of the DiskOnChip that I've tried to get working Mar 03 23:20:46 Is it working yet? Last I saw there seemed to much work involving byte-swapping... Mar 03 23:21:33 well, it kindof worked Mar 03 23:21:42 the problem now is that I overwrote redboot Mar 03 23:22:13 I solved the problem: the image was built inside a symlinke'd directory tree Mar 03 23:22:51 Break out the jtag, I guess :( Mar 03 23:23:22 yeah. Unfortunately, there's no support for the DoC Mar 03 23:23:38 so it appears I'll have to work around that somehow Mar 03 23:24:45 The redboot was in the DoC on that machine? Mar 03 23:25:18 yeah Mar 03 23:26:08 Well, I guess they got the redboot in there somehow, so there has to be a way to put it back! Mar 03 23:40:07 bulk programmer? ;) Mar 03 23:40:46 * NAiL just has 140 commits left before he crosses the 1k border ;) Mar 04 00:03:27 ~weather enva Mar 04 00:04:49 Apparently there is no weather in enva? Mar 04 00:08:20 heh Mar 04 00:08:24 -11C Mar 04 00:08:50 ~weather enva Mar 04 00:09:03 Trondheim / Vaernes, Norway; (ENVA) 63-28N 010-56E 17M; last updated: 2006.03.03 2250 UTC; Dew Point: 5 F (-15 C); Pressure (altimeter): 29.50 in. Hg (0999 hPa); Relative Humidity: 72%; Temperature: 12 F (-11 C); Visibility: greater than 7 mile(s); Wind: from the SE (130 degrees) at 6 MPH (5 KT); Windchill: 3 F (-16 C) Mar 04 00:09:03 hmm Mar 04 00:09:11 there we are Mar 04 00:28:03 hello everyone Mar 04 00:28:29 i have installed cyrus-imapd but have a slight problem....anyone else missing cyradm too? Mar 04 05:21:06 03bzhou * 10unslung/make/samba.mk: upgrade to 3.0.21c Mar 04 06:25:04 03bzhou * 10unslung/make/py-curl.mk: upgrade to 7.15.2 Mar 04 06:28:00 03bzhou * 10unslung/make/py-mssql.mk: upgrade to 0.7.4 Mar 04 06:32:48 03bzhou * 10unslung/make/py-nose.mk: upgrade to 0.8.7.1 Mar 04 06:35:20 03bzhou * 10unslung/make/py-roundup.mk: upgrade to 1.1.1 Mar 04 07:59:04 hello. Mar 04 07:59:10 or good morning. Mar 04 08:00:15 I got some problems with my nslu2, it's not so much capable of being a.. storage device, it more like a temporary storage device. Mar 04 08:00:31 Since while disk is attached to it, unslung can't fsck it. Mar 04 08:01:16 and this is already a third time that I loose files while doing a fsck on another machine for disk since unslung wasn't able to boot anymore. Mar 04 08:01:34 Ok, no, actually I don't loose them. Files are in /lost+found Mar 04 08:02:16 but some are corrupted, they lack their locating and their filename.. and did I forgot to mention.. There's nearly a million of those files.. :) Mar 04 08:02:26 So yeah, I kinda lost them.. Mar 04 08:02:37 So, I was here wondering something.. Mar 04 08:02:53 When you nóormally boot, let's say, slackware linux.. Mar 04 08:03:33 on bootup it checks that, oh, disk has been mounted already N times without being checked.. or there's already N days since last fsck.. Mar 04 08:03:52 unslung 6.x could be having a similar setup too. Mar 04 08:04:08 while disk hasn Mar 04 08:04:26 't yet been mounted, unslung would be able to do a fsck for it. Mar 04 08:05:09 as you all propably know, clicking "scandisk" (or what ever it was) in web interface does NOT work. Mar 04 08:17:58 hopefully this was stored to someone's log as I think it ain't too hard for developers to implement, but I decided not to write it on the wiki's wishlist, as page says that it's more like a developers' meeting place instead of end-user wishlist page and as I am too busy with my other projects and ainä't a unslung developer, I wasn't eligible to type it to there. Mar 04 08:22:27 and I know, fsck'ing big disks is quite slow.. but would we rather wait some times during bootup instead of losing files? **** ENDING LOGGING AT Sat Mar 04 10:59:57 2006