**** BEGIN LOGGING AT Thu Mar 24 02:59:58 2011 Mar 24 03:23:36 build #75 of octeon is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/octeon/builds/75 Mar 24 04:30:30 build #76 of mpc52xx is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/mpc52xx/builds/76 Mar 24 05:49:44 build #105 of ubicom32 is complete: Failure [failed compile_3] Build details are at http://tksite.gotdns.org:8010/builders/ubicom32/builds/105 Mar 24 06:11:17 build #110 of at91 is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/at91/builds/110 Mar 24 07:20:13 i'd really like to fix the iftop problem, but ... like ... i am out of ideas? Mar 24 08:51:47 gmorning Mar 24 10:59:41 Hi. Which config options should I use to build backfire 10.03.1-rc4 with two SSID enabled? I have DLink Dir-320 router. Mar 24 11:25:55 acoul * r26282 /packages/net/iputils/Makefile: net/iputils: beautify brcm47xx fix (make owrt devs happy) Mar 24 11:30:37 acoul * r26283 /trunk/tools/cmake/Makefile: tools/cmake: update to version 2.8.4 Mar 24 13:41:56 nbd * r26284 /trunk/package/mac80211/patches/570-mac80211_fix_ibss_probereq_handling.patch: mac80211: fix dropping multicast proberequest frames for ad-hoc Mar 24 13:42:41 nbd * r26285 /branches/backfire/package/mac80211/patches/582-mac80211_fix_ibss_probereq_handling.patch: mac80211: fix dropping multicast proberequest frames for ad-hoc (backport of r26284) Mar 24 14:46:31 stintel: ping Mar 24 14:51:31 KanjiMonster: pang :) Mar 24 14:54:10 stintel: have you seen this ticket? -> https://dev.openwrt.org/ticket/9107 Mar 24 14:54:22 perhaps the workaround works for your tp-link, too Mar 24 14:55:02 KanjiMonster: interesting, I'll have to check it when I'm at home Mar 24 14:55:42 thanks, btw Mar 24 14:56:10 yw Mar 24 15:17:23 'lo KanjiMonster, FinboySlick Mar 24 15:19:45 EqUaTe: Yo dude :) Mar 24 15:20:11 EqUaTe: Had I told you about my little vrrp/conntrackd/lacp router project? Mar 24 15:25:23 don't remember.. possibly? Mar 24 15:29:45 I'm still in the planning phase but my only disappointment so far is that keepalived only takes second timing. No sub-second failover for me :P Mar 24 15:37:35 lol Mar 24 15:50:24 EqUaTe: It shouldn't be all that tough to hack in though. Using utime instead of time and you're pretty much set. Mar 24 16:30:47 jow_laptop: latest include/image.mk find expression w/build output here: http://paste.openwrt.org/m49f36246 Mar 24 16:31:19 markmentovai: ^^ better? Mar 24 16:31:23 ;) Mar 24 16:32:59 i don't like the duplication in the find condition, i wonder if you could pull that out into some variable? Mar 24 16:33:52 also the indent on the ( and ) still has me a little confused, i think it might help if you always put your -a and -o at the end of the line rather than at the beginning. (you've got one at a line end, the rest are at the beginning.) Mar 24 16:37:09 markmentovai: as in a MATCH variable which can be preceded by a '!'? Mar 24 16:37:17 exactly Mar 24 16:37:28 I'll have a look Mar 24 17:07:29 build #94 of ps3 is complete: Failure [failed compile_3] Build details are at http://tksite.gotdns.org:8010/builders/ps3/builds/94 Mar 24 17:38:43 build #95 of pxcab is complete: Failure [failed compile_3] Build details are at http://tksite.gotdns.org:8010/builders/pxcab/builds/95 Mar 24 18:09:44 is there a whois query tool for OpenWRT? Mar 24 18:14:15 build #109 of atheros is complete: Failure [failed compile_6] Build details are at http://tksite.gotdns.org:8010/builders/atheros/builds/109 Mar 24 19:05:50 nbd: while using xz instead of lzma for squashfs in 2.6.38 is nice, having a way of creating xz compressed squashfs images would be even nicer ;p (I'm currently looking into updating squashfs to 4.2, which includes xz support) Mar 24 19:16:52 KanjiMonster: wouldn't xz have a detrimental effect on boot time though? Mar 24 19:17:40 Still a worthy compromise as boot time isn't quite as important as space... But that's still something to consider in my opinion. Mar 24 19:17:42 FinboySlick: I don't know. Wasn't my idea of switching to xz for .38, and my experience with it is exactly zero Mar 24 19:17:58 FinboySlick: we already have lzma, xz shouldn't be visibly slower than that Mar 24 19:18:08 good point Mar 24 19:18:13 nbd: I guess so. Mar 24 19:18:23 I never played with xz myself. Mar 24 19:19:31 You guys ever run into a situation where the modules.d/* things just plain ignored params? If I load the module by hand it accepts the param, if I just add the param=val to the appropriate line, it's ignored. Mar 24 19:19:50 I don't quite get how, it's just a simple pipe of the contents to insmod. Mar 24 19:20:22 Oh now I remember why I didn't want to put stuff into online auctions anymore: http://pastebin.com/pzamUL7r Mar 24 19:20:43 heh Mar 24 19:20:47 (99% of the words are written incorrectly) Mar 24 19:26:56 KanjiMonster: I suspect that mkfs.jffs strips SUID/SGID bits btw. Mar 24 19:27:05 mksquashfs4 doesn't Mar 24 19:48:01 ermo: somehow I'm missing the context ... Mar 24 20:09:51 nbd: might this be relevant to the iftop problem? https://forum.openwrt.org/viewtopic.php?id=13848 Mar 24 20:10:16 interesting Mar 24 20:10:52 yes, it might be relevant Mar 24 20:10:54 please test it Mar 24 20:11:07 if you get it working with linking order changes, i'll dig into uclibc to try to find the source of this bug Mar 24 20:17:41 compile.txt shows "-lpcap -lm -lcurses -lpthread" Mar 24 20:19:59 hm Mar 24 20:20:13 try linking -lpthread first Mar 24 20:20:19 okay Mar 24 20:36:03 nbd: what's the best way to modify that order? Mar 24 20:39:13 no idea, check the makefile Mar 24 20:53:35 thepeople1: which reminds me, what's the state of my account request? ;) Mar 24 21:06:40 nbd, using CONFIGURE_VARS += LIBS="-lm -lpthread" i got the link order of curses and pthread reversed, no difference. Mar 24 21:30:05 build #103 of cobalt is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/cobalt/builds/103 Mar 24 21:46:30 moved -lpthread in front of -lpcap and still the same Mar 24 22:08:30 build #108 of s3c24xx is complete: Failure [failed shell compile_6] Build details are at http://tksite.gotdns.org:8010/builders/s3c24xx/builds/108 Mar 24 22:41:35 EqUaTe: Still awake? Mar 24 22:41:43 just about Mar 24 22:42:26 You aware of something that might give me better than 2-3 ms failover for a pair of redundant openwrt boxes? Mar 24 22:42:35 I mean 2-3 seconds. Mar 24 22:42:48 I'd like to be in the ms range, but it can be a bit higher than 3 ;) Mar 24 22:43:11 see if you can find a daemon that implements bfd Mar 24 22:43:15 and use ospf? Mar 24 22:43:27 bfd + ospf == very fast Mar 24 22:43:36 ospf w/o bfd can be a bit slow Mar 24 22:44:00 to be fair, i'm not actually sure what you're doing at the moment.. Mar 24 22:45:27 EqUaTe: More or less a standard NAT box (though it NATs and routes between several subnets), if it dies for some reason, I have to fail over to a hot standby twin. That implies conntrackd to preserve existing connections. Mar 24 22:46:00 ah ok. that's what the vrrp thing you mentioned earlier was in connection with then Mar 24 22:46:22 Yes. Honestly, I'd be fine with keepalived if it let me set values less than a second. Mar 24 22:46:58 i don't know if you can use bfd with vrrp or not.. might be worth looking into, but kinda dependant on if there's anything on openwrt that implements it at all Mar 24 22:51:01 EqUaTe: I'll certainly look into it. Thanks and try to get some good rest. Mar 24 22:51:12 yeah, hitting the sack now. Mar 24 22:51:15 gl Mar 24 22:51:22 Yup. Mar 24 22:55:07 KanjiMonster: you said "(I'm currently looking into updating squashfs to 4.2, which includes xz support)" -- I merely remarked that mksquashfs (which would presumably be updated if you updated squashfs to v4.2) preservers SGID/SUID bits while it appears that e.g. mkfs.jffs doesn't Mar 24 22:55:47 implying that if you knew how to make jffs / mkfs.jffs not strip suid/sgid bits it would be cool :) Mar 24 23:08:13 xMff: http://paste.openwrt.org/m57573ff <- how about that? Mar 24 23:08:51 jow_laptop: (you probably highlight on both) Mar 24 23:39:04 ermo: you were asking it's clear and easy to follow and I'd say it definitely is Mar 24 23:40:47 I also like the printing of what's change and that we ignored for debugging Mar 24 23:43:03 It might become unwieldy is my only concern (the size of that output of changed and ignored I mean). Mar 24 23:46:27 ermo: also, would it be possible to use some kind of $(foreach protected,$(CONFIG_IMAGE_PRESERVE_PERM),-path '$(protected)' -o \) type thing? Mar 24 23:47:36 ermo: in which the user could specify a space-separate list of paths to keep the permissions on? Mar 24 23:48:29 oh, skip the \ though Mar 25 00:58:58 nbd * r26286 /trunk/package/redboot-ar231x/Makefile: redboot-ar231x: mark as broken, the ecos host tool crap needs some rework for tcl on some systems Mar 25 01:27:57 thepeople1: ping Mar 25 01:28:02 OutBackDingo: pong **** ENDING LOGGING AT Fri Mar 25 02:59:57 2011