**** BEGIN LOGGING AT Wed Aug 08 02:59:58 2012 Aug 08 03:38:35 a vanilla image does the same thing Aug 08 03:39:43 (same thing == kernel panic) Aug 08 05:03:12 build #52 of pxcab is complete: Failure [failed shell_7 shell_11] Build details are at http://buildbot.openwrt.org:8010/builders/pxcab/builds/52 Aug 08 05:03:40 * russell-- trying again with kernel debugging options all on, hoping to see the source of the panic Aug 08 05:18:22 build #51 of rb532 is complete: Failure [failed shell_11] Build details are at http://buildbot.openwrt.org:8010/builders/rb532/builds/51 Aug 08 05:24:17 build #51 of ppc44x is complete: Failure [failed shell_11] Build details are at http://buildbot.openwrt.org:8010/builders/ppc44x/builds/51 Aug 08 06:41:29 hmm. no more verbosity was forthcoming. Aug 08 06:51:15 russell--: kallsyms log ? Aug 08 06:52:16 ah Aug 08 06:52:21 your rootfs is not detected Aug 08 07:04:10 blogic: which line is that? Aug 08 07:05:24 i was wrong Aug 08 07:05:28 no idea Aug 08 07:06:05 and here you got my hopes up ;-) Aug 08 07:06:22 * russell-- will try bisecting the problem tomorrow Aug 08 08:09:16 build #42 of sibyte is complete: Failure [failed shell_11] Build details are at http://buildbot.openwrt.org:8010/builders/sibyte/builds/42 Aug 08 08:31:52 build #40 of xburst is complete: Failure [failed shell_17] Build details are at http://buildbot.openwrt.org:8010/builders/xburst/builds/40 Aug 08 08:41:01 ping dangole Aug 08 08:52:09 pong jow_laptop Aug 08 08:54:01 jow_laptop: postpone, got to go for an hour or two, i'll get back to you later Aug 08 09:49:41 build #36 of au1000 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/au1000/builds/36 Aug 08 09:57:52 is there a way for a package using PKG_SOURCE_PROT:=git, to specify the branch that should be checked out? Aug 08 09:59:03 PKG_SOURCE_VERSION:=origin/branchname Aug 08 10:04:01 <[florian]> karlp: specifying the commit should be enough no? Aug 08 10:05:20 I guess? Aug 08 10:05:33 I don't always remember that detail of git Aug 08 10:06:02 jow_laptop: will PKG_SOURCE_VERSION with a branchname get HEAD of that branch then I take it? Aug 08 10:06:12 probably Aug 08 10:20:50 <[florian]> karlp: the commit id fully specifies the branch as well Aug 08 10:25:41 ping jow_laptop Aug 08 10:33:23 dangole: pong Aug 08 10:33:57 dangole: I've been looking through your rssi led code and wondered why you're not using quality_max to infer the range Aug 08 10:35:04 you mean as the upper limit? Aug 08 10:35:45 i see it can make sense to allow the configuration to contain percentage values and then scale them to be x% of quality_max Aug 08 10:37:31 jow * r33052 /packages/net/multiwan/ (Makefile files/usr/bin/multiwan): [packages] multiwan: switch to /lib/functions/network.sh to derive network state (#11780) Aug 08 10:40:28 yes, something like that Aug 08 10:40:59 however I am not sure whether the quality indicators are a good choice Aug 08 10:41:21 for wl you only have 0..5 and madwifi rarely reports something greater than 20/70 or so Aug 08 10:41:56 maybe you should just fetch noise floor and avg rssi and the apply the mac80211 algo on that Aug 08 10:43:30 I mena this: https://dev.openwrt.org/browser/trunk/package/iwinfo/src/iwinfo_nl80211.c#L1032 Aug 08 10:43:53 afaik e.g. rt2x00 doesn't report a meaningful noise floor, but does report a meaningful signal quality value Aug 08 10:44:00 I see Aug 08 10:44:13 via wext? Aug 08 10:44:52 nl80211 via libiwinfo Aug 08 10:44:57 and wext as well Aug 08 10:45:01 ok Aug 08 10:45:10 i can basically just use Aug 08 10:45:16 ( qual * 100 ) / qual_max Aug 08 10:45:20 yes Aug 08 10:45:23 instead of returning qual directly Aug 08 10:45:52 but according to what you state on madwifi and wl drivers, that wouldn't make things less confusing in practise Aug 08 10:46:13 but in theory it's nicer :) Aug 08 10:47:13 besides that, there are two more things i want to go over again Aug 08 10:47:51 - maybe use a timer call-back instead of usleep Aug 08 10:48:05 - add a debug parameter Aug 08 10:48:18 - make the thing fork() by itself instead of doing that in bash Aug 08 10:49:01 (i'm not sure if that's actually nicer, maybe fork'ing it and writing the pid file in shell is just as good) Aug 08 10:49:24 we have service wrappers for that Aug 08 10:49:33 so no need to reimplement it Aug 08 10:49:45 ah ok, i'll have a look at them then Aug 08 10:50:19 besides that, any hints? Aug 08 10:50:40 yeah, you can check if using "radio0" instead of a specific wifi iface gets you meaningful quality values Aug 08 10:50:53 in theory libiwinfo will try to find a vap on the phy to work with Aug 08 10:51:09 then you can maybe ditch the retry stuff Aug 08 10:52:59 ok, doesn't work Aug 08 10:54:09 yeah, and also there is no consistent scheme for accessing/naming the phy radio in different drivers Aug 08 10:54:21 using the vif is least-confusing imho Aug 08 10:55:09 the retry stuff works nicely btw, tested it a lot, making the interface go down, disappear entirely, ... it survives all of that and comes back on Aug 08 10:55:27 i guess quality_max is driver-specific and doesn't change Aug 08 10:55:28 good Aug 08 10:55:41 so i can cache it instead of getting it from iwinfo every time i'm polling the quality, right? Aug 08 10:57:21 yes, you can cache it Aug 08 10:57:36 and i'll implement signal handlers, so SIG_TERM would make it exit cleanly Aug 08 10:59:37 or maybe not, doesn't really matter in a way, killing it is good enough Aug 08 11:13:15 jow * r33053 /trunk/package/hostapd/ (Makefile files/wpa_supplicant.sh): [package] hostapd: fix ap/sta regression which got accidentally introduced by r32829 (#11914) Aug 08 11:14:27 dangole: yes, exit will clean everything up, no need to worry about teardown Aug 08 11:42:35 jow, I'm using XHR.poll(url, {data to post}, callback) but the data to post is always the same? Aug 08 11:42:39 http://pastebin.com/YXssednL Aug 08 11:42:54 is there anything extra to send different data on each loop? Aug 08 11:43:17 build #40 of ixp4xx is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/ixp4xx/builds/40 Aug 08 11:44:13 it looks like poll() does _q.push({data:data}) so it always uses the same data each time? Aug 08 12:14:45 blogic * r33054 /trunk/target/linux/lantiq/patches-3.3/0005-pci-support.patch: Aug 08 12:14:45 [lantiq] pci gpio mapping Aug 08 12:14:45 This patch fixes the GPIO ALTSEL settings for some of the GPIOs used by the PCI Aug 08 12:14:45 subsystem in Lantiq Danube. These changes are required for more than one PCI Aug 08 12:14:45 device to work. Tested with an ARV7510PW having two PCI-devices; a VIA USB Aug 08 12:14:46 controller and a Ralink WLAN mini-PCI card. Aug 08 12:14:47 Signed-off-by: Matti Laakso Aug 08 12:56:46 nbd * r33055 /trunk/package/mac80211/patches/573-ath9k_fix_reset_hang.patch: ath9k: fix hang issues on hw reset caused by interrupt storms Aug 08 14:12:11 hcg * r33056 /packages/net/wget/Makefile: [packages] wget: add required dependency (librt) Aug 08 14:38:25 build #35 of iop32x is complete: Failure [failed shell_11] Build details are at http://buildbot.openwrt.org:8010/builders/iop32x/builds/35 Aug 08 14:44:16 build #35 of etrax is complete: Failure [failed shell_11] Build details are at http://buildbot.openwrt.org:8010/builders/etrax/builds/35 Aug 08 15:05:45 build #34 of gemini is complete: Failure [failed shell_11] Build details are at http://buildbot.openwrt.org:8010/builders/gemini/builds/34 Aug 08 15:37:36 florian * r33057 /trunk/package/base-files/ (Makefile files/etc/profile): [package] base-files: check for ldd presence and executability (#11991) Aug 08 15:38:41 florian * r33058 /packages/mail/xmail/ (Makefile files/xmail.init): [package] xmail: cope with empty flags options (#9452) Aug 08 15:38:42 florian * r33059 /packages/multimedia/mjpg-streamer/ (Makefile files/mjpg-streamer.config files/mjpg-streamer.init): [package] mjpg-streamer: install web pages (#6499) Aug 08 15:38:47 florian * r33060 /packages/net/peervpn/ (. Makefile): (log message trimmed) Aug 08 15:38:47 [package] add peervpn package Makefile Aug 08 15:38:47 PeerVPN (http://www.peervpn.net/), by Tobias Volk, is software that Aug 08 15:38:47 builds virtual ethernet networks between multiple computers. A virtual Aug 08 15:38:48 network built by PeerVPN uses a full mesh topology. All nodes talk Aug 08 15:38:48 directly to each other, there is no need for a central server. If one Aug 08 15:38:48 node goes down, the rest of the network is unaffected. Aug 08 15:38:51 florian * r33061 /packages/net/kamailio3/ (. Makefile files/ files/kamailio.default files/kamailio.init): (log message trimmed) Aug 08 15:38:51 [package] add kamailio3 Aug 08 15:38:51 Official kamailio 3.0 was released on January 11, 2010. Aug 08 15:38:51 It includes an astonishing number of improvements and new features. Aug 08 15:38:51 Kamailio v3.3.0 was released on June 18, 2012. This patch contains Aug 08 15:38:52 last version of Kamailio. Compared to previous version of kamailio3 Aug 08 15:39:06 package (sent http://patchwork.openwrt.org/patch/2284/ ) there is no Aug 08 15:39:18 jow_laptop: I'm running luci locally with make runuhttpd, and it only _seems_ to be adding index() methods from a single controller. Aug 08 15:39:37 I have ./luasrc/controller/rme/blahx,y,z.lua, Aug 08 15:39:48 and pages from one of those files works, but not the others. Aug 08 15:39:53 never seen such a bheaviour so no idea, sorry Aug 08 15:40:01 they all work ok when actually bundled into an image and run on the router itself Aug 08 15:40:12 maybe another host system issue then Aug 08 15:40:13 was just trying to have a stdout/stderr to help debugging Aug 08 15:40:24 assert or whatever Aug 08 15:40:32 you need to investigate, cannot reproduce it here Aug 08 15:40:37 yeah, ok, fair enough Aug 08 15:41:37 any other tips for debugging controllers on the router? beyond asserting and io.writing to a log file or something? Aug 08 15:41:54 is there an easy way to get stdout/stderr going to /tmp or syslog? Aug 08 15:44:32 nope Aug 08 15:50:18 uhttpd -h /www -f -p 8080 seems to work somewhat ok... Aug 08 16:49:52 build #32 of octeon is complete: Failure [failed shell_11] Build details are at http://buildbot.openwrt.org:8010/builders/octeon/builds/32 Aug 08 18:23:02 nbd * r33062 /trunk/package/mac80211/patches/300-pending_work.patch: ath5k: revert a buggy tx power related patch Aug 08 18:23:05 nbd * r33063 /trunk/package/mac80211/patches/300-pending_work.patch: ath9k: fix handling of decryption errors, should improve link stability Aug 08 19:35:16 build #57 of s3c24xx is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/s3c24xx/builds/57 **** ENDING LOGGING AT Wed Aug 08 20:13:49 2012 **** BEGIN LOGGING AT Wed Aug 08 20:14:16 2012 Aug 08 20:40:20 marek * r33064 /packages/net/batman-adv/ (2 files in 2 dirs): Aug 08 20:40:20 batman-adv: bla2 & gateway mode interaction fix Aug 08 20:40:20 Signed-off-by: Marek Lindner **** ENDING LOGGING AT Wed Aug 08 20:41:31 2012 **** BEGIN LOGGING AT Wed Aug 08 20:42:22 2012 Aug 08 21:50:56 florian * r33065 /trunk/target/linux/mcs814x/image/Makefile: [mcs814x] cleanup image generation Aug 08 21:51:01 florian * r33066 /trunk/target/linux/mcs814x/image/Makefile: Aug 08 21:51:01 [mcs814x] build a Devolo dLAN USB Extender specific image Aug 08 21:51:01 The kernel has to be rounded up to the nearest 1MB, then we append Aug 08 21:51:01 the root filesystem at its end. Aug 08 21:51:06 florian * r33067 /trunk/target/linux/mcs814x/files-3.3/drivers/net/ethernet/mcs8140/nuport_mac.c: [mcs814x] nuport-mac: check dma_map_single address return values Aug 08 21:51:10 florian * r33068 /trunk/target/linux/mcs814x/files-3.3/drivers/net/ethernet/mcs8140/nuport_mac.c: [mcs814x] nuport-mac: unmap the DMA buffer when closing device Aug 08 21:51:15 florian * r33069 /trunk/target/linux/mcs814x/files-3.3/drivers/net/ethernet/mcs8140/nuport_mac.c: [mcs814x] nuport-mac: close possible race condition during start Aug 08 21:51:20 florian * r33070 /trunk/target/linux/mcs814x/files-3.3/drivers/net/ethernet/mcs8140/nuport_mac.c: [mcs814x] nuport-mac: rework PHY interrupt configuration to avoid races Aug 09 00:18:25 * danmackay wonders why my last build didnt have a working luci/openssh install and goes to make a new build Aug 09 00:55:22 http://pastebin.com/w0h2jdG3 well fuck, really **** ENDING LOGGING AT Thu Aug 09 02:59:59 2012