**** BEGIN LOGGING AT Thu Jan 26 02:59:57 2012 **** ENDING LOGGING AT Thu Jan 26 03:37:48 2012 **** BEGIN LOGGING AT Thu Jan 26 03:55:12 2012 Jan 26 05:46:24 hi anyone alive Jan 26 05:46:57 when making a package can i do PKG_SOURCE=./ or something Jan 26 05:47:06 so i dont have to keep downloading Jan 26 09:35:31 build #114 of ppc44x is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/ppc44x/builds/114 Jan 26 09:36:38 so whats the story with building your own apps.. Jan 26 09:37:01 i.e if you need to patch apps to make them work.. the easiest way is to make a package Jan 26 09:37:12 install, test, change code build again Jan 26 09:44:37 vot: as part of the package source stuff, you include a checksum.. if the source matches the checksum it won't redownload it - it keeps the tarball from previous downloads unless you do distclean iirc Jan 26 09:44:39 vot: post your makefile that you have made Jan 26 09:55:30 luka12345 im looking at someone elses port of nut Jan 26 09:55:36 looking at updating it Jan 26 09:57:21 any recommendations on visual diff tools Jan 26 09:57:22 http://amjith.blogspot.com/2007/07/visual-diff-tools-in-linux.html Jan 26 09:57:27 those are some Jan 26 09:59:06 i think meld was ok Jan 26 10:10:19 depends if you want a gui tool or not.. if you do, xxdiff is good. Jan 26 10:26:44 whats -lm do in makefile.am used like this Jan 26 10:26:45 LDADD_COMMON = ../common/libcommon.a -lm Jan 26 10:26:57 some linking option i guess Jan 26 10:27:23 probably links libm or something along those lines Jan 26 10:37:54 EqUaTe link lib math Jan 26 10:44:25 juhosg * r29910 /trunk/target/linux/ramips/ (11 files in 9 dirs): (log message trimmed) Jan 26 10:44:25 ramips: add support for Sitecom WL-341v3 and other Sercomm IP1006RRv2 boards Jan 26 10:44:25 This patch adds support for the Sitecom WL-341 v3 and other Sercomm Jan 26 10:44:25 IP1006RRv2 based boards for sysupgrade support and for the initial Jan 26 10:44:25 flash through pushbutton initiated recovery mode with the special Jan 26 10:44:25 partition table and fixes for the quirks and things required by the Jan 26 10:44:26 modified bootloader. Jan 26 11:47:26 hmm Jan 26 11:47:34 isnt libusb a standard package Jan 26 11:48:40 wow something really old in https://dev.openwrt.org/browser/packages/libs/libusb?rev=9173 Jan 26 11:49:15 we have two libusb packages Jan 26 11:49:18 that's the old one Jan 26 11:49:23 there's also libusb-1.0 Jan 26 12:10:47 arent they built as a part of the standard builds ? Jan 26 12:12:03 yes Jan 26 12:13:13 oh my Jan 26 12:13:15 http://downloads.openwrt.org/backfire/10.03.1/ar71xx/packages/ Jan 26 12:13:21 my local mirror is totally busted Jan 26 12:13:24 their missing Jan 26 14:38:30 hi Jan 26 14:38:39 o/ Jan 26 14:38:58 I just packaged ndppd for Openwrt (http://www.priv.nu/projects/ndppd/) Jan 26 14:39:09 oh, yay Jan 26 14:39:11 how to submit it? Jan 26 14:39:36 i already maintain babel, polipo and ahcp Jan 26 14:39:40 send it to openwrt-devel, I'll look into it Jan 26 14:39:51 ok Jan 26 14:40:14 wasn't aware that there already is such a software Jan 26 14:40:25 I recall debating about the need to add ndp proxying to radvd a while ago Jan 26 14:43:21 jow_laptop: there is also another one, but much simpler, and probably a bit buggy Jan 26 14:43:42 can’t find it right now Jan 26 14:44:20 I did not test ndppd yet, I’ll send the patch once I’m sure it runs ok Jan 26 14:44:27 allright Jan 26 14:46:49 but upstream is responsive so I’m confident Jan 26 14:47:03 (fixed 2 issues already) Jan 26 14:48:52 jow_laptop: http://code.google.com/p/npd6/ Jan 26 14:49:42 but it replies blindly for all IP, not checking they are really available Jan 26 14:49:54 so a bit rude Jan 26 14:50:07 hm yes Jan 26 14:51:47 oh, and btw: any reason enable_vlan4k is disabled by default (at least on netgear wndr3700)? Jan 26 14:52:08 i lost an afternoon figuring this one out Jan 26 14:52:08 probably not Jan 26 14:52:25 I just unconditionally enable it from within the gui nowadays Jan 26 14:52:32 if it is available Jan 26 14:52:33 ok Jan 26 14:52:48 KanjiMonster might know if there is. Jan 26 14:52:51 I should have used the gui then Jan 26 14:52:55 ;) Jan 26 14:53:05 well I asked the same a while ago actually Jan 26 14:53:13 and it probably boils down to the fact that nobody cares Jan 26 14:55:24 I don't think there is Jan 26 14:55:56 because nobody uses vlan > 15, except when your ISP decided that IPTV should be on vlan 100 :-$ Jan 26 14:56:47 heh Jan 26 14:56:53 i prefer vlan numbers like 42 :P Jan 26 14:56:56 and higher. Jan 26 14:57:04 i very rarely use anything less than 10 Jan 26 16:15:46 jow_laptop: why would i get this the linker Jan 26 16:16:02 bcmxcp.c:(.text+0x278): undefined reference to `ldexp' Jan 26 16:16:08 thats in libc isnt it Jan 26 16:19:40 kids Jan 26 16:20:03 yes, it's an e/glibc function Jan 26 16:20:15 you'll have to change the code, or put in a function for uclibc Jan 26 16:20:59 or you're not linking -lm Jan 26 16:21:04 which seems more likely Jan 26 16:23:52 compiles file under linux Jan 26 16:24:07 just not sure how to modify to link it under the package make file Jan 26 16:24:23 openwrt *is* linux, just a different one Jan 26 16:24:37 i mean on ubuntu etc Jan 26 16:24:58 under glibc etc, libm is normally implicated in builds Jan 26 16:25:24 so how would you go about making it link -lm Jan 26 16:25:41 you want LIBS="-lm" Jan 26 16:25:54 for the package configure Jan 26 16:26:14 although it depends on the precise build system Jan 26 16:27:50 hey Jan 26 16:28:46 I noticed, that my mmc card is identified with a different dev (not /dev/sdc but /dev/sdd), when I start with usb-stick at the same time Jan 26 16:29:00 this fs up the fstab and so the mmc doesn't get mounted Jan 26 16:29:38 yes, typical. unless you have hotplug mappings or something, they'll be in the order they were detected/inserted Jan 26 16:29:41 so I want to ask you, why doesn't openwrt use unique ids? in mandriva they are called uuid and they help identifying devices without relying on the name in /dev Jan 26 16:29:41 Chocks looks like it has a LIBS= Jan 26 16:30:15 uuids are unwieldy IMO, but I presume they can be used in openwrt Jan 26 16:31:18 Chocks: why unwieldy, cause of their looks?? Jan 26 16:31:23 yes(??) Jan 26 16:31:27 stupid long string Jan 26 16:31:30 :) :) Jan 26 16:31:41 I agree there's an issue, but I don't agree that a long string is the answer Jan 26 16:31:52 just like IPv6 addresses Jan 26 16:31:55 for flash partitions, there's a simple name. that's an easy solution. Jan 26 16:32:16 ker: yes, although at least those have shortcuts. not that I know them Jan 26 16:32:21 kind of true, but they prevent mischief, but I will be happy if it's true that you can already use uuid Jan 26 16:32:36 * Chocks suggests you investigate and report back Jan 26 16:32:42 you won't be the first person to ask Jan 26 16:35:22 mounting by uuid or partition label should work Jan 26 16:37:54 jow_laptop: nice :-) Jan 26 17:01:43 juhosg * r29911 /trunk/target/linux/generic/files/include/linux/myloader.h: Jan 26 17:01:44 generic: add WPE72 ID to myloader.h Jan 26 17:01:44 Signed-off-by: Johnathan Boyce Jan 26 17:01:45 juhosg * r29912 /trunk/target/linux/ar71xx/ (11 files in 6 dirs): Jan 26 17:01:45 ar71xx: add support for Compex WPE72/WPE72NX Jan 26 17:01:45 This patch adds support for Compex WPE72 bare board and Compex WPE72NX Jan 26 17:01:45 Indoor Access Point. Jan 26 17:01:45 [juhosg: add support to 3.2 as well] Jan 26 17:01:46 Signed-off-by: Johnathan Boyce Jan 26 17:01:46 juhosg * r29913 /trunk/tools/firmware-utils/src/ (mkmylofw.c myloader.h): Jan 26 17:01:47 firmware-utils/mkmylofw: add WPE72 support Jan 26 17:01:47 Signed-off-by: Johnathan Boyce Jan 26 17:02:42 juhosg * r29914 /trunk/target/linux/ar71xx/ (generic/profiles/compex.mk image/Makefile): Jan 26 17:02:43 ar71xx: create a profile and build images for Compex WPE72 Jan 26 17:02:43 [juhosg: remove ath9k from the the profile package list, build WPE72 Jan 26 17:02:43 image with the Default profile as well] Jan 26 17:02:43 Signed-off-by: Johnathan Boyce **** ENDING LOGGING AT Thu Jan 26 17:46:57 2012 **** BEGIN LOGGING AT Thu Jan 26 17:48:20 2012 Jan 26 20:26:30 hey people Jan 26 20:26:47 just tried to use ruby in trunk, installed ruby, ruby-gems, ruby-ncurses Jan 26 20:27:18 tried to run an example.rb which depends on ruby-ncurses Jan 26 20:27:47 require fails, so it seems ruby is setup poorly and misses some module directories Jan 26 20:27:53 or something else is wrong Jan 26 20:51:30 ruby is a mess, in my limited experience Jan 26 20:51:44 unless you use the ruby tools, which are contrary to normal Linux packaging Jan 26 20:53:23 Chocks: the ruby gems tool is available in openwrt Jan 26 20:53:38 but it doesn't really work... it seems :-) Jan 26 20:53:49 I'm sure it is. doesn't mean it's not a mess. Ubuntu it's not much better Jan 26 20:54:04 it's like the PHP PEAR stuff, but worse Jan 26 20:54:33 I'd still like to find a solution Jan 26 20:54:50 you have to package everything, really Jan 26 20:55:05 the other modules may or may not be in other packages. Jan 26 20:55:08 nbd checked ruby into svn, maybe contact him Jan 26 20:55:20 Chocks: all the modules I need are available via opkg Jan 26 20:55:21 * Chocks slaps nbd around Jan 26 20:55:24 and installed Jan 26 20:55:37 are you saying some configure options need adding? it's possible Jan 26 20:55:51 PHP is like that. there's some additional options I had to add for some of the extensions to work Jan 26 20:55:59 but the require statement which is like include fails and I tried to supply the path, but there's really no reason not to find the module Jan 26 20:57:01 the ruby include rules are bizarre in the extreme Jan 26 20:57:25 don't know, didn't use ruby much, just wanted to play with it and learn Jan 26 20:57:26 * Chocks suggests use of strace Jan 26 20:57:33 but from the tutorials I really like its elegance Jan 26 20:57:44 good idea Jan 26 20:57:56 the language might be fine, but the install is not well thought out Jan 26 20:58:02 :-) Jan 26 20:58:49 package management is a mess in every OS AFAIK Jan 26 20:59:28 not that I tried many... Jan 26 20:59:43 yes, I like opkg, but in a perfect world Jan 26 20:59:43 it's well done in Debian Jan 26 20:59:51 opkg doesn't pretend to catch every case Jan 26 21:00:17 you can type install xyz, and it gets the SOURCES via GOOGLE, compiles it on some server farm with perfect options, downloads it in a neat package, and voila Jan 26 21:00:29 ^^ Jan 26 21:01:06 the "problem" with openwrt is that in an embedded system, packages tend to be more tightly integrated with options, etc. so mixing and matching packages doesn't always work Jan 26 21:01:16 opkg is good enough for installing new software, and that's usually all that';s required Jan 26 21:02:04 respect to opkg devs, it's really good for what it does Jan 26 21:02:21 it does it better than rpm, that's for sure Jan 26 21:02:53 rpm does a lot more though, but it has ugly baggage, urpmi gets around alot of the issues Jan 26 21:03:00 a lot Jan 26 21:03:13 rpm doesn't really do very much more than opkg Jan 26 21:03:38 I don't know, the things you can put in a .spec are pretty complex, you can sign packages and so on Jan 26 21:03:56 .spec files are a mess. anyway, you handle that all in Makefiles in openwrt Jan 26 21:04:18 seen it, looks simple, which is good Jan 26 21:04:28 the spec file syntax is a disaster Jan 26 21:04:37 blogic: you about? got another NAND question :) Jan 26 21:04:50 hi Jan 26 21:05:41 got most ok, but 'Attempt to write not page aligned data' on mts wrtie and DD. Have you had time to look at the driver; got some code for me to test? Jan 26 21:05:47 I agree, it is a disaster :-) do a rpm --help and see how many "features" there are, that's not to say they are good, but it's complex, maybe that's the problem Jan 26 21:06:01 anybody tried creating a real desktop distro based on opkg? Jan 26 21:06:16 I'm not sure it's worthwhile. just use Debian Jan 26 21:06:18 btsimonh: not yet Jan 26 21:06:22 busy with other stuff Jan 26 21:06:24 sorry, Ubuntu Jan 26 21:06:51 I don't know that everything you'd want in a Desktop system is in OpenWrt Jan 26 21:07:00 perhaps it is. a simple one, anyway Jan 26 21:07:33 ok, will try to find where it breaks. it's probably a 32/8 access I missed. will keep you informed if I get finalised with flash all good. Currently I'm getting read only mount of overlay. Jan 26 21:08:07 certainly such a system would be a lot more lightweight than ubuntu + gnome + moose Jan 26 21:08:27 well Jan 26 21:08:35 you need to make sure to use 8bit read Jan 26 21:08:44 the ebu is not sane for 32 bit reads Jan 26 21:08:56 I bed nbd has opinions on the matter Jan 26 21:09:00 bet Jan 26 21:09:18 turns out ruby tries to load some thread.so which is nowhere Jan 26 21:11:59 build #107 of avr32 is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/avr32/builds/107 Jan 26 21:12:06 ok, turns out after installing ruby-core via opkg the error has shifted to something else Jan 26 21:13:29 does something in nand 'read then write' to write short blocks? seems 'nand_write' just bombs if size is not page aligned? Jan 26 21:19:03 great, now in strace it claims not to find a file that is there Jan 26 21:19:27 either that's again an overlayfs problem, or there's a broad definition of file not found Jan 26 21:21:28 I know that from my exeperience with strace + ruby, it tries all kinds of paths, and you have to sift to find what was really missing Jan 26 21:21:46 true Jan 26 21:25:14 the point where I want to give up is approaching very fast now heh Jan 26 21:25:23 I can imagine Jan 26 21:26:44 would my chances be better with python? Jan 26 21:27:17 probably Jan 26 21:27:20 or lua Jan 26 21:27:32 given that luci is lua Jan 26 21:28:37 juhosg * r29915 /trunk/target/linux/ar71xx/image/Makefile: Jan 26 21:28:37 ar71xx: create the temporary directory before building initramfs images Jan 26 21:28:37 Thanks to Florian for reporting. Jan 26 21:38:50 boot log is here http://pastebin.com/Ds3pRfpu, read only overlay; pls take a lok to see if there are any ideas :) Jan 26 21:40:02 well, surely the warnings from mtd at boot are the problem? Jan 26 21:40:30 fa954 is definitely not rounded to a block Jan 26 21:44:39 should the build process have done this rounding of kernel to rootfs boundary?. I'd ignored that because rootfs_data is aligned :) Jan 26 21:44:58 Should I try a jffs2 based image, rather than squashfs, maybe? Jan 26 21:47:06 Chocks: lua is very interesting, ruby, python and the like have an extensive library though, don't believe it has that Jan 26 21:47:48 could be. it's the usual language debate. All I can say is that lua is well supported in openwrt, since luci uses it Jan 26 21:47:59 And PHP is well supported too, although that's a mess Jan 26 21:48:19 yeah, PHP is not bad, but no thanks Jan 26 21:48:42 I've not seen python used much on OpenWrt, but it might be packaged properly Jan 26 21:48:54 we use PHP extensively here on OpenWrt Jan 26 21:49:07 it's not bad, it's not just not great Jan 26 21:49:14 -one not Jan 26 21:50:19 I had to use it a lot for a project once, then I tried it for another project too, but with the switch to python it became all so much smoother Jan 26 21:50:47 let's stop the lang debate right here ;-) Jan 26 21:53:35 I never started ;-) I'm just making observations on what's in openwrt, if you're determined to code there. Jan 26 21:53:54 but I can't fault your choice of python from what I know Jan 26 22:00:07 you might find that the image build needs to do padding Jan 26 22:00:12 ^btsimonh: Jan 26 22:02:00 just tried a jffs2 based image. got lots of jffs2 errors with 64k version. now going to try 128k version. using u-boot to program to avoid linux programming issues for the moment. Jan 26 22:02:31 I suggesting staying with squashfs and fixing that alignment first Jan 26 22:02:43 if that's wrong, all bets are off Jan 26 22:03:07 btsimonh: using nand flash? Jan 26 22:03:18 mb_! nbd! Jan 26 22:03:29 where do I find image padding? nbd: yes, nand Jan 26 22:03:38 Chocks: lua has a few package websites, but I don't see API documentation anywhere, that's a mayor issue for me, even PHP is great for it's API docs alone Jan 26 22:03:39 should be in arch/xxx/image/Makefile Jan 26 22:03:51 well, nand is somewhat different from nor flash, so you need to generate the jffs2 images with nand headers Jan 26 22:04:04 major ;-) Jan 26 22:04:07 depending on the flash type Jan 26 22:04:14 nbd: OpenWrt desktop system. discuss. Jan 26 22:05:10 discuss what? Jan 26 22:05:36 how you would go about making an OpenWrt-based Desktop system Jan 26 22:06:00 what might be missing Jan 26 22:06:20 step 1: figure out what's missing. step 2: port what's missing. step 3: use it ;) Jan 26 22:06:25 true, true Jan 26 22:06:37 what about middleware stuff? Jan 26 22:06:51 *cough*dbus*cough* Jan 26 22:07:08 what about it? Jan 26 22:07:15 dunno. any issues you forsee? Jan 26 22:07:25 we have a package for dbus already Jan 26 22:07:28 including a python binding Jan 26 22:07:30 yeah, I know. Jan 26 22:07:32 just wondering. Jan 26 22:07:46 anyway, 3 = ????? 4 = Profit! Jan 26 22:07:49 i don't usually 'forsee' issues in random packages Jan 26 22:08:01 just try it and see how much of it works or is broken Jan 26 22:08:08 especially not about the future? Jan 26 22:08:36 don't think a root-only sys is good idea for desktop... but opkg might be able to replace rpm urpmi in mageia or something like that Jan 26 22:08:50 who says it has to be root-only? Jan 26 22:08:51 and then use ROM for standard tools and /overlay for updates and the like Jan 26 22:08:51 adduser Jan 26 22:09:01 no problem Jan 26 22:09:03 sudo Jan 26 22:09:04 etc Jan 26 22:09:13 that would be cool for desktop, because you can always fallback and remove /overlay Jan 26 22:09:15 Chocks: when i look into the future, i only see stuff that works, not stuff that breaks ;) Jan 26 22:09:32 or you can copy /overlay which makes for easy backups Jan 26 22:09:35 nbd: can I borrow your rose-tinted spectacles some time? Jan 26 22:09:49 dadada: you can just run ext3, no problem Jan 26 22:09:49 just make your own, they're free ;) Jan 26 22:10:18 Chocks: I run ext4 on top of overlay Jan 26 22:11:02 you'd have to go e/glibc if you wanted to use non-free stuff though, I think Jan 26 22:11:08 html5 videos/mp3 with google-chrome are running with 2x speed, somethings wrong with my system Jan 26 22:11:36 which license is used now? Jan 26 22:11:56 I'm talking about flash, nvidia stuff, etc Jan 26 22:12:05 those will have glibc bindings Jan 26 22:12:09 maybe we should run non-free stuff in a vm :-) Jan 26 22:13:00 I see, so in theory they can run with our libc, but not in practice Jan 26 22:13:17 couldn't we have two sets of libc libraries? Jan 26 22:13:36 yes, that's possible Jan 26 22:13:42 if you set it up right. Jan 26 22:13:46 or would that mess with symbols? Jan 26 22:14:23 you'd have to run e.g, Firefox against glibc too, since the plugin will be glibc-bindings. Jan 26 22:14:34 I don't think it's worthwhile Jan 26 22:14:51 the point of uclibc is to save space on very constrained systems. you don't have that issue on x86 with phat HDDs Jan 26 22:15:22 so we make openwrt desktop edition glibc only? Jan 26 22:15:35 it's "just" a build option Jan 26 22:15:39 as it is now Jan 26 22:17:32 I can imagine such a project to be successfull... it's just a lot of work Jan 26 22:17:50 it's not really any more working than porting new packages to openwrt Jan 26 22:17:58 90% of it's there already Jan 26 22:18:16 depends, ubuntu and the like have 10000s of packages Jan 26 22:18:23 how many do you use? Jan 26 22:18:26 you don't need it for a basic desktop Jan 26 22:18:41 but having everything available via packagecommand is huge sellingpoint Jan 26 22:18:47 there's only 10 major apps I use most of the time. Jan 26 22:18:49 I' Jan 26 22:18:54 it's an option, is all I'm saying Jan 26 22:18:59 and the 10 major apps I use may not be the same Jan 26 22:19:03 has uses for kiosks and such Jan 26 22:19:04 likewise for the next user Jan 26 22:19:15 oh, yeah, there I agree Jan 26 22:19:22 I don't doubt it'd be very personalized at a first cut Jan 26 22:19:48 why couldn't we take packages from another distro (say arch) and convert them to ours? Jan 26 22:20:14 or maybe something like gentoo or arch or whatever has package "makefiles", that could be converted to our's on the fly? Jan 26 22:20:28 then we can copy their repositories heh Jan 26 22:20:34 if you don't need to recompile, that is assuming x86/glibc, then you could convert packages, yes Jan 26 22:20:54 but the point of openwrt is to remove all the crap from packages, so really you want to make them into native ones Jan 26 22:21:13 the alien tool may already do some of this Jan 26 22:21:19 the point makes sense on a router, not sure on the desktop Jan 26 22:21:41 depends on the goal of your desktop. if you want a desktop with everything, stick with ubuntu Jan 26 22:22:13 so is it a microdesktop? Jan 26 22:22:22 whatever you want to call it Jan 26 22:22:53 opendesktop Jan 26 22:22:58 sounds good IMO Jan 26 22:24:32 doh.... found it - look at line 78 of previous pastebin http://pastebin.com/Ds3pRfpu and just call me stupid :) i left a rootfs_data in my nor defn i did as a test Jan 26 22:24:40 silly Jan 26 22:45:47 anyone use u-boot fw_setenv here? does it not come with uboot-envtools? Jan 26 22:50:45 btsimonh: yes Jan 26 22:51:03 but you need to enable it in 'make menuconfig' Jan 26 22:51:21 it's not in the current lantiq download package? Jan 26 22:51:27 but if i recall corectly it's only a symlink Jan 26 22:51:39 i wouldn't know Jan 26 22:51:45 i build my own images Jan 26 22:52:14 :) understood. i'll take a look & make it part of the image. Jan 26 22:52:33 ps. so far, jffs2 hold up. Jan 26 22:52:50 good to hear Jan 26 22:52:55 it should be there, what other use is envtools with set Jan 26 22:53:01 without Jan 26 22:53:17 you can read the values Jan 26 22:53:21 ;) that's what i thought. but it's not ... Jan 26 22:53:51 did not wantto reboot again into u-boot to make it so it would boot MY firmware from now on.... Jan 26 22:56:01 yes - fw_setenv is a symlink to fw_printenv. i guess opkg can't do this easily Jan 26 22:56:19 "<*> uboot-envtools................. read/modify U-Boot bootloader environment" should include modifying -> set Jan 27 01:13:58 someone i'm working with posted a kernel panic he's getting on a wgt634u during bittorrent Jan 27 01:21:03 involves linux-3.0.17/mm/swap.c:static void __put_compound_page(struct page *page) Jan 27 01:21:22 which makes it sound like a memory exhaustion problem Jan 27 01:21:40 but why is it panic'ing? there is no swap. Jan 27 01:22:21 also, we continue to see "kernel taint" messages in oops traces, even though lsmod says kernel isn't tainted Jan 27 01:22:49 seems like maybe a holdover from the madwifi days? all ath5k now. Jan 27 01:23:11 so why the message: "Disabling lock debugging due to kernel taint" Jan 27 01:36:57 http://pastebin.com/Fr3tTjqq Jan 27 01:43:30 looks like that comes from add_taint Jan 27 01:43:32 if (flag != TAINT_CRAP && flag != TAINT_WARN && __debug_locks_off()) Jan 27 01:43:36 printk(KERN_WARNING "Disabling lock debugging due to kernel taint\n"); Jan 27 01:45:53 good question Jan 27 01:48:54 root@k:/proc# cat /proc/sys/kernel/tainted Jan 27 01:48:55 0 Jan 27 01:49:10 i guess it's something that's happening as it's panic'ing maybe Jan 27 01:49:27 __debug_locks_off() Jan 27 01:49:34 sounds soething that's rather normal Jan 27 01:49:40 *like something Jan 27 01:49:55 so the msg would always occur Jan 27 01:50:20 yeah, but message gives a scary impression, i guess i'll learn to ignore Jan 27 01:50:43 any thoughts on the panic itself? Jan 27 01:50:47 hm Jan 27 01:51:11 philipp64|laptop reported unaligned access errors with ath5k as well Jan 27 01:51:27 apparently caused by some led support code or corresponding mac80211 update Jan 27 01:51:36 yup. Jan 27 01:52:42 this particular panic was observed at r29903 Jan 27 01:58:46 what are philipp64|laptop's details, i don't see them in my irc logs Jan 27 02:56:52 build #109 of ar7 is complete: Exception [exception interrupted] Build details are at http://buildbot.openwrt.org:8010/builders/ar7/builds/109 Jan 27 02:56:53 build #108 of sibyte is complete: Exception [exception interrupted] Build details are at http://buildbot.openwrt.org:8010/builders/sibyte/builds/108 Jan 27 02:56:55 build #107 of xburst is complete: Exception [exception interrupted] Build details are at http://buildbot.openwrt.org:8010/builders/xburst/builds/107 Jan 27 02:56:57 build #107 of ixp4xx is complete: Exception [exception interrupted] Build details are at http://buildbot.openwrt.org:8010/builders/ixp4xx/builds/107 **** ENDING LOGGING AT Fri Jan 27 02:59:57 2012