**** BEGIN LOGGING AT Tue Nov 12 02:59:58 2013
Nov 12 08:05:56 <russell--> has anyone tested on wgt634u recently?
Nov 12 08:06:52 * russell-- just built an image using my customary kit of options and files tree, and it hangs at "Starting program at 0x80001000"
Nov 12 08:07:07 <russell--> trying a vanilla config now
Nov 12 08:24:50 <russell--> vanilla config hands in the same spot
Nov 12 08:25:06 * russell-- wonders if he needs to clean his tree somehow
Nov 12 08:25:18 <russell--> has been a few months since the last brcm47xx build
Nov 12 08:26:32 <builder> build #363 of kirkwood is complete: Failure [failed compile_8]  Build details are at http://buildbot.openwrt.org:8010/builders/kirkwood/builds/363
Nov 12 09:23:01 <eigma> JyZyXEL: fascinating. I've added some comments to the ticket.
Nov 12 09:28:24 * russell-- didn't see trouble until trunk bumped from 3.6 to 3.8 iirc
Nov 12 09:32:03 <eigma> can't offer any specific advice
Nov 12 09:32:12 <eigma> try clean build, or perhaps bisect
Nov 12 09:36:52 <eigma> I've had problems with upstream changes not properly causing rebuilds, so I tend to rely more on ccache rather than incremental builds
Nov 12 09:40:01 <Guest19371> wigyori r38775 trunk/target/linux/mxs/files/arch/arm/boot/dts/imx28-duckbill.dts * mxs: duckbill: adjust leds to final hardware
Nov 12 10:08:49 <Guest19371> wigyori r38776 trunk/package/boot/uboot-envtools/ Makefile files/mxs * uboot-envtools: add support for mxs target
Nov 12 10:11:12 <russell--> atheros, airrouter, buffalo wzr600dhp, tplink wdr3600, all "working" so far, just wgt634u screwed (again), doing an alix2 build now
Nov 12 10:11:49 <russell--> just got a 7-port usb hub, have them all wired up to serial consoles in the basement, debugging from couch
Nov 12 10:12:59 <wigyori> sounds like a low-power computing cluster
Nov 12 10:21:12 <Guest19371> wigyori r38777 trunk/target/linux/ (7 files in 5 dirs) * mxs: add basic board runtime detection support
Nov 12 10:31:18 <russell--> alix2 working fine too
Nov 12 10:31:39 * russell-- is still perplexed by weird /proc/meminfo math
Nov 12 10:34:56 <russell--> on some of the devices (wzr600dhp, wnd3600, airrouter ... all the ar71xx's) MemTotal minus MemFree,Buffers,Cached,AnonPages,Slab,Mapped,Shmem,KernelStack,PageTables gives a positive number (~1MB-~6MB), but on atheros and alix2, i get a negative number, wtf?
Nov 12 11:02:35 <Guest19371> jow r38778 trunk/package/network/ utils/ipset/Makefile utils/ipset/patches * Update ipset to version 6.20.1
Nov 12 11:02:57 <Guest19371> jow r38779 trunk/package/network/ utils/ipset/patches/001-timeout-size.patch utils/ipset/patches * Add patch to fix ipset timeout setting problem
Nov 12 12:49:56 <Guest19371> wigyori r38780 trunk/target/linux/ (5 files in 2 dirs)
Nov 12 12:49:56 <Guest19371> mxs: bump kernel to 3.12
Nov 12 12:49:56 <Guest19371>  - remove DCP patch as it's in mainline now (dts patch will stay)
Nov 12 12:49:56 <Guest19371>  - refresh Duckbill patch
Nov 12 12:49:56 <Guest19371>  - update kmod-usb-chipidea
Nov 12 12:49:58 <Guest19371> Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
Nov 12 14:41:19 <Guest19371> wigyori r38781 trunk/ package/boot/uboot-mxs/Makefile target/linux/mxs/profiles/01-olinuxino.mk
Nov 12 14:41:19 <Guest19371> mxs: - update uboot-mxs to 3.10
Nov 12 14:41:19 <Guest19371>  - enable building uboot for imx23 along sbImage by default
Nov 12 14:41:19 <Guest19371>  - (doc updated)
Nov 12 14:41:20 <Guest19371> Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
Nov 12 19:13:26 <russell--> built a clean checkout for wgt634u, it's busted too.  r38779.  I guess I start bisecting.
Nov 12 19:21:30 <builder> build #417 of uml is complete: Failure [failed compile_5]  Build details are at http://buildbot.openwrt.org:8010/builders/uml/builds/417
Nov 12 19:30:06 <russell--> Hauke: have you tested a wgt634u on trunk recently?  /me built HEAD last night, the console output stops at "Starting program at 0x80001000"
Nov 12 19:48:18 <Hauke> russell--: I do not own a wgt634u, I would assume this change is causing the problems: https://dev.openwrt.org/changeset/38309
Nov 12 19:49:02 <Hauke> russell--: could you please replace the SSB_CHIPCO_UART0_DATA with SSB_CHIPCO_UART1_DATA in arch/mips/bcm47xx/prom.c
Nov 12 21:06:03 <obinou> Hello, is someone working on the "Lite" versions of the microtik's RB433 (and alike) ?
Nov 12 21:06:15 <obinou> the ones without a functionning serial port
Nov 12 21:43:37 <Pteridium>  
Nov 12 22:13:13 <Guest19371> nbd r38782 trunk/package/network/ (5 files)
Nov 12 22:13:13 <Guest19371> hostapd: remove #ifdef MULTICALL around ap+sta support code (to support separately installed supplicant+hostapd)
Nov 12 22:13:20 <Guest19371> nbd r38783 trunk/package/kernel/ (56 files in 2 dirs) * mac80211: update to wireless-testing 2013-11-05
Nov 12 22:32:11 <builder> build #440 of orion is complete: Failure [failed compile_8]  Build details are at http://buildbot.openwrt.org:8010/builders/orion/builds/440
Nov 12 22:50:23 <falstaff_> Hm, lists.openwrt.org down?
Nov 12 22:50:56 <tripolar> hi
Nov 12 22:52:03 <tripolar> i have a bug with the AP Pro when using sysupgrade - it doesnt save the settings - the same image (compiled from the same sources) works as expected for the nanostation m2
Nov 12 22:52:11 <tripolar> here are the two sysupgrade ouputs
Nov 12 22:52:16 <tripolar> http://paste.debian.net/65528/
Nov 12 22:53:06 <tripolar> it seems the AP Pro writes too early it seems it writes the jffs2 data - and goes on with "Writing from <stdin> to firmware"
Nov 12 22:53:54 <tripolar> im now reading thorugh mtd.c
Nov 12 22:54:11 <tripolar> nbd: blogic any idea how to fix this?
Nov 12 23:01:50 <nbd> tripolar: ah, i see why
Nov 12 23:01:57 <tripolar> nbd why?
Nov 12 23:02:01 <nbd> because the kernel partition contains a jffs2 too
Nov 12 23:02:20 <nbd> mtd writes the config data to the first end-of-filesystem marker it can find
Nov 12 23:02:56 <nbd> so a way to fix this would be to not append the marker to the kernel jffs2
Nov 12 23:03:04 <nbd> but pad the image with 0xff up to the partition size
Nov 12 23:03:16 <nbd> hm, wait
Nov 12 23:03:19 <nbd> that still won't work
Nov 12 23:03:54 <nbd> i guess we need to add something to mtd to make it skip the first n number of bytes for jffs2 detection
Nov 12 23:05:03 <tripolar> nbd: where from do you know that the kernel partition contains a jffs2 too?
Nov 12 23:05:13 <tripolar> and why is it working on the nanostation?
Nov 12 23:05:26 <nbd> i looked at the image makefile
Nov 12 23:05:39 <nbd> only the ap pro has that quirk
Nov 12 23:05:45 <nbd> on the nano there is no kernel partition jffs2
Nov 12 23:05:52 <tripolar> okay thanks
Nov 12 23:06:16 <tripolar> so what would be a way to fix it?
Nov 12 23:06:34 <nbd> what i just said
Nov 12 23:06:35 <nbd> 00:03 < nbd> i guess we need to add something to mtd to make it skip the first n number of bytes for jffs2 detection
Nov 12 23:06:50 <nbd> then change sysupgrade to use that to make it ignore the first jffs2 data
Nov 12 23:08:02 <tripolar> but when i skip the first n number of bytes for jffs2 detection - why would it work then?
Nov 12 23:09:10 <nbd> because then it will not try to append the config data to the kernel partition
Nov 12 23:09:20 <nbd> and it will find the rootfs instead
Nov 12 23:10:28 <tripolar> so "n" must be as big as the kernel partition?
Nov 12 23:10:52 <nbd> yes
Nov 12 23:18:40 <tripolar> nbd: i will try to fix it
Nov 12 23:18:48 <tripolar> thanks for your help
Nov 12 23:19:49 <tripolar> nbd: one last question how do i find out how big the kernel image is? from /proc/mtd
Nov 12 23:19:52 <tripolar> mtd2: 00180000 00010000 "kernel"
Nov 12 23:20:05 <nbd> you can also look at the image makefile
Nov 12 23:20:37 <nbd> but yes, there you have the size in hex
Nov 12 23:21:50 <tripolar> nbd: is there somewhere a good documentation about flash handling and mtd devices in linux - programming wise? or just source code?
Nov 12 23:22:55 <nbd> what part of it?
Nov 12 23:23:03 <nbd> either way, i don't know any good documentation on it
Nov 12 23:24:11 <tripolar> okay thanks
Nov 12 23:25:07 <tripolar> as it seems i'm not the only one who avoids writing domcumentation but who likes to read it ;)
Nov 13 00:21:13 <russell--> weird, on one rev of my bisection, i got a no route to host message on another host on the LAN, but can get routed to a distant host, copying image off remote host.
Nov 13 00:21:37 <russell--> r38114
Nov 13 00:26:37 <russell--> r38217 also
Nov 13 00:27:25 <russell--> can ping distant hosts, not local ones
Nov 13 00:30:16 <russell--> tried turning off the firewall, no diff
Nov 13 00:45:05 <russell--> hmm, here's a kernel panic on atheros (accton mr3201a): http://pastebin.ca/2475713 (r38774)
Nov 13 00:54:53 <DonkeyHotei> russell--: would that be one of the fon devices?
Nov 13 00:56:45 <russell--> not specifically, but i think it's equivalent
Nov 13 00:56:57 <russell--> the fon devices had less ram, iirc
Nov 13 00:57:39 <DonkeyHotei> equivalent to which one?
Nov 13 00:58:50 * russell-- doesn't know
Nov 13 00:59:06 * russell-- only ever saw one
Nov 13 00:59:13 <russell--> of the fon's
Nov 13 01:00:09 <DonkeyHotei> i'm curious at to whether this would be correct for it: http://patchwork.openwrt.org/patch/4351/
Nov 13 01:00:16 <DonkeyHotei> *as to
Nov 13 01:02:37 * russell-- is busy bisecting my wgt634u not-booting problem, i just happened to notice the kernel panic, thought i'd report it
Nov 13 01:04:32 <russell--> each iteration takes about 24 minutes to build, 5-10 minutes to test, could be done tonight!
Nov 13 01:04:55 <Devastator> jesus..
Nov 13 01:15:54 <russell--> this not ping local hosts is pretty weird, not sure what would account for that
Nov 13 01:17:16 <Devastator> are you trying to ping ips, right?
Nov 13 01:17:30 <russell--> yes
Nov 13 01:17:33 <Devastator> I mean, of course, but I thought I should ask
Nov 13 01:17:35 <russell--> not a dns problem
Nov 13 01:17:58 <Devastator> routing? firewall on the other end?
Nov 13 01:18:29 <russell--> it's another host on the same /24
Nov 13 01:18:55 <Devastator> linux?
Nov 13 01:18:57 <russell--> yes
Nov 13 01:19:12 <Devastator> maybe it's dropping INPUT
Nov 13 01:19:16 <russell--> not just that host, but all hosts i've tried
Nov 13 01:19:35 <russell--> it's nothing on the destination, it sees the traffic
Nov 13 01:19:50 <russell--> and i can forward through the gateway router
Nov 13 01:20:08 <russell--> i've turned off firewall
Nov 13 01:20:11 <russell--> on the router
Nov 13 01:20:20 <russell--> the one i'm testing (not the gateway)
Nov 13 01:20:27 <Devastator> can you ping from a host to another host?
Nov 13 01:21:23 <Devastator> hum, then your gateway is causing something, to eliminate this hypotesis I would try to ping between hosts
Nov 13 01:21:39 <DonkeyHotei> is loopback down perchance?
Nov 13 01:21:46 <russell--> yes, that works fine, it's just this router i'm bisecting one and it just started
Nov 13 01:22:21 <russell--> lo is up 127.0.0.1
Nov 13 01:22:32 <Devastator> ah, have you tried adding iface parameter to ping?
Nov 13 01:22:35 <russell--> and i can ping to distant routers
Nov 13 01:22:53 <Devastator> I think it's ping -I
Nov 13 01:22:54 <russell--> like 4.2.2.2 etc
Nov 13 01:23:16 <russell--> the default gateway is on the network i can't reach
Nov 13 01:23:44 <russell--> i can ping it
Nov 13 01:23:53 <russell--> 192.168.x.1 is reachable
Nov 13 01:24:03 <russell--> 192.168.x.y is not (where y is not 1)
Nov 13 01:25:05 <Devastator> well, my bet is either the input of this router is dropping the packets or something on the gateway
Nov 13 01:25:54 <Devastator> if the hosts can see the traffic, then it's accepting it, but it gets dropped somewhere returning
Nov 13 01:26:15 <russell--> iptables (-t filter, -t nat and -t mangle) are empty
Nov 13 01:27:08 <Devastator> what about default policy?
Nov 13 01:28:40 <russell--> ACCEPT
Nov 13 01:29:28 <builder> build #319 of adm5120 is complete: Failure [failed svn compile_10]  Build details are at http://buildbot.openwrt.org:8010/builders/adm5120/builds/319
Nov 13 01:29:49 <Devastator> then it is something in the middle
Nov 13 01:30:19 <russell--> no, nothing in the middle has changed
Nov 13 01:30:33 <russell--> it's some kind of bug in the network stack, is my guess
Nov 13 01:30:48 <Devastator> tcpdump time :)
Nov 13 01:31:07 <russell--> when i'm done bisecting to booting problem i'll go back and find this one
Nov 13 01:31:25 <russell--> besides, there is nothing in the middle except switches
Nov 13 01:31:27 <russell--> dumb ones
Nov 13 01:32:08 <builder> build #301 of mpc52xx is complete: Failure [failed svn compile_10]  Build details are at http://buildbot.openwrt.org:8010/builders/mpc52xx/builds/301
Nov 13 01:34:56 <russell--> there was a switch problem earlier
Nov 13 01:35:17 <russell--> maybe that's it, but why distant hosts work?
Nov 13 01:36:07 <Devastator> maybe it uses another route?
Nov 13 01:36:20 <Devastator> wire route, that is
Nov 13 01:39:50 <russell--> nope
Nov 13 01:42:08 <russell--> i can't ping between devices that are on the same switch from the test router, but not from the test router
Nov 13 01:43:20 <russell--> sorry, i *can* ping between devices on the same switch as the test router, but not from the test router
Nov 13 01:43:40 <russell--> fingers--
Nov 13 01:44:13 <Devastator> hum, have you checked if there is no mac cloned or dumb mac?
Nov 13 01:45:01 <russell--> it's got the same mac address as it had on an earlier image, before it stopped working
Nov 13 01:45:41 <Devastator> hum..
Nov 13 01:47:15 <tripolar> ping nbd
**** ENDING LOGGING AT Wed Nov 13 02:59:58 2013