**** BEGIN LOGGING AT Wed Nov 07 03:00:02 2012
Nov 07 07:50:54 <russell--> that should be grub_dprintf's
Nov 07 07:51:47 <OpenWrt-Commits> kaloz r34104 trunk/target/linux/ixp4xx/patches-3.3/120-compex_support.patch * [ixp4xx]: fixup typos in the compex patch
Nov 07 09:01:25 <OpenWrt-Commits> jogo r34105 trunk/target/linux/ brcm63xx/patches-3.6/444-BCM63XX-add-endian-check-for-ath9k.patch brcm63xx/patches-3.3/444-BCM63XX-add-endian-check-for-ath9k.patch * bcm63xx: Enable endian check support for ath9k pci fixup.
Nov 07 09:41:35 <written_tabletpc> re :-)
Nov 07 16:15:44 <OpenWrt-Commits> jow r34106 trunk/ include/package.mk include/package-ipkg.mk package/Makefile * buildroot: isolate the .install stamp files for build variants (#12279)
Nov 07 19:37:18 <OpenWrt-Commits> blogic r34107 trunk/package/kernel/modules/netsupport.mk * [kernel] Load schedule modules
Nov 07 20:25:56 <OpenWrt-Commits> blogic r34108 trunk/target/linux/ (14 files in 2 dirs) * ramips: port patches to linux-3.6
Nov 07 20:26:17 <OpenWrt-Commits> blogic r34109 trunk/target/linux/ ramips/rt3883/config-3.6 ramips/rt288x/config-3.6 ramips/rt305x/config-3.6 * ramips: add linux-3.6 configuration for ramips socs
Nov 07 20:26:26 <OpenWrt-Commits> blogic r34110 trunk/target/linux/ramips/files/arch/mips/ralink/Kconfig * ramips: add HAVE_CLK to Kconfig on all ramips socs
Nov 07 20:26:44 <OpenWrt-Commits> blogic r34111 trunk/target/linux/ramips/Makefile * ramips: use kernel 3.6.6
Nov 07 20:26:53 <OpenWrt-Commits> blogic r34112 trunk/target/linux/ramips/rt305x/profiles/fon.mk * [ramips] adds fon20n profile
Nov 07 20:29:07 <blogic> dangole: ^
Nov 07 20:30:32 <dape> cool ! any news on dlink dir615 h1 wireless problem..?
Nov 07 20:30:47 <dape> the issue being that it isnt working.
Nov 07 20:53:22 <DonkeyHotei> blogic: openwrt now works on the fon20n? i have one! i should try it!
Nov 07 21:22:11 <russell--> weird, i've been having trouble with a mikrotik RB493G, where one of the ethernet interfaces just goes away.  It's bridged with a wireless interface, eth0 and wlan0, and wlan0 will still work, but eth0 (which has 4 external ports on ths switch) just becomes unreachable.
Nov 07 21:22:44 <russell--> it shows "UP" but no packets go out/in
Nov 07 21:23:18 <russell--> i tried ifconfig eth0 down ; ifconfig eth0 up and that didn't help.  only solution seems to be a reboot.
Nov 07 21:28:28 <russell--> the only clue i have is these that show up in the logs:
Nov 07 21:28:32 <russell--> Nov  6 20:29:11 192.168.80.1 kernel: [348988.000000] eth0: link up (10Mbps/Half duplex)
Nov 07 21:28:35 <russell--> Nov  6 20:29:13 192.168.80.1 kernel: [348990.010000] eth0: link up (1000Mbps/Full duplex)
Nov 07 21:31:39 <russell--> seems to happen after a few days of uptime
Nov 07 21:35:30 <russell--> last interval from reboot to failure was 4 days +/- an hour
Nov 07 21:42:47 <KanjiMonster> russell--: strange. I don't have that problem on mine.
Nov 07 21:45:10 <KanjiMonster> russell--: or maybe I do; I only use the switch on eht1; let me check
Nov 07 21:49:27 <KanjiMonster> russell--: hm nope. uptime 15 days, and it's working
Nov 07 22:08:31 <russell--> time to failure before that was 6.5 days approximately
Nov 07 22:09:08 <russell--> but there was a good stretch in october (4th->17th) that didn't see it)
Nov 07 22:09:38 <russell--> i have 100Mbps device plugged into the external ports
Nov 07 22:09:44 <russell--> i wonder if that's related?
Nov 07 22:10:24 <russell--> why would the port speed/duplex flap like that?
Nov 07 22:10:37 <russell--> KanjiMonster: ^^
Nov 07 22:12:00 <KanjiMonster> russell--: I have no idea. ideally it should not do that at all, especially since the link to the switches is fixed
Nov 07 22:16:37 <russell--> yeah
Nov 07 22:16:40 <russell--> Nov  2 10:03:14 192.168.80.1 kernel: [568886.860000] ar71xx: pll_reg 0xb8050014: 0x991099
Nov 07 22:16:43 <russell--> Nov  2 10:03:14 192.168.80.1 kernel: [568886.860000] eth0: link up (10Mbps/Half duplex)
Nov 07 22:16:46 <russell--> Nov  2 10:03:16 192.168.80.1 kernel: [568888.870000] ar71xx: pll_reg 0xb8050014: 0x110000
Nov 07 22:16:50 <russell--> Nov  2 10:03:16 192.168.80.1 kernel: [568888.870000] eth0: link up (1000Mbps/Full duplex)
Nov 07 22:17:01 <russell--> that happened in the neighborhood of the previous failure too
Nov 07 22:17:38 <KanjiMonster> I see that in my log, too; but it didn't break it for me
Nov 07 22:17:48 <russell--> yeah, not always fatal
Nov 07 22:18:03 <KanjiMonster> well, not quite. for me it was 1000 Mbps => 1000 Mbps
Nov 07 22:18:23 <KanjiMonster> well, 1000/half => 1000/full
Nov 07 22:18:41 * russell-- scanning my logs
Nov 07 22:18:46 <KanjiMonster> (no idea what happens there - 1000/half isn't even a supported mode)
Nov 07 22:22:43 <KanjiMonster> russell--: does eth1 ever do it?
Nov 07 22:23:05 <russell--> no
Nov 07 22:23:36 <russell--> always eth0
Nov 07 22:24:31 <russell--> my neighbors are getting annoyed, since i'm on the local wireless (wlan0) part of the bridge and that keeps working, i don't notice it until someone texts me or calls. ;-)
Nov 07 22:30:19 <KanjiMonster> strange
Nov 07 22:30:26 <russell--> yes
Nov 07 22:30:52 <KanjiMonster> but I can confirm that it goes down
Nov 07 22:30:57 <KanjiMonster> or rather up again
Nov 07 22:31:15 <russell--> do you have any advice what to look at to try to recover?
Nov 07 22:31:33 <russell--> so far, i've stared at it, can't figure out whats wrong, and reboot
Nov 07 22:31:37 <russell--> which fixes it
Nov 07 22:32:57 <russell--> "ip a" shows eth0 is UP, but i can't ping across it
Nov 07 22:33:33 <russell--> and as i said before ifconfig eth0 down ; ifconfig eth0 up didn't help
Nov 07 22:34:08 <russell--> although log messages show the bridge noticing it had gone down and came back up
Nov 07 22:34:41 <KanjiMonster> if the down/up doesn't fix it, I'm out of ideas for now
Nov 07 22:35:07 <russell--> is there a way to nail the port to 1000Mbps/Full duplex?
Nov 07 22:35:29 <KanjiMonster> the port is supposed to be nailed to 1000/Full
Nov 07 22:35:33 <russell--> heh
Nov 07 22:36:38 <russell--> how about a way to reset the switch chip, sort of rebooting the machine?
Nov 07 22:36:47 <russell--> short* of
Nov 07 22:37:49 <russell--> maybe i can test whether other devices on the same switch can still see each other, next time
Nov 07 22:38:15 <russell--> all i know now is that the CPU can't see across the switch
Nov 07 22:38:36 <russell--> but maybe devices on the other ports of the switch can still see each other
Nov 07 22:39:43 <russell--> also, i have a spare here, maybe i should try dropping that in and see if i still see outages
Nov 07 22:39:49 <KanjiMonster> I would guess devices still see each other
Nov 07 22:39:55 <russell--> okay
Nov 07 22:40:21 <KanjiMonster> what's probably failing is the rgmii connection
Nov 07 22:59:14 <dangole> @dape: you can try playing with vasilugin recently posted patch, place it in package/mac80211/patches/625-rt3352-ext-alc.patch
Nov 07 22:59:46 <dangole> dape: i'll have a closer look to it tomorrow and i would be glad to have someone with the actual hardware around for testing
Nov 07 23:10:41 <dape> dangole oh my god, im at the office tomorrow with the dir 615 h1 on my desk with serial on
Nov 07 23:10:51 <dape> i suppose you are gmt+1 +2 ?
Nov 07 23:10:55 <dangole> +2
Nov 07 23:11:10 <dape> awesome, so i'm from 9 to 6pm
Nov 07 23:12:28 <OpenWrt-Commits> luka r34113 trunk/target/linux/cns3xxx/base-files/lib/cns3xxx.sh * cns3xxx: fix coding style in board script
Nov 08 00:20:13 <dangole> KanjiMonster: 1000/Half has never been specified. gigabit afaik always do full duplex
Nov 08 00:20:54 <KanjiMonster> dangole: tell that my device: [1254156.880000] eth0: link up (1000Mbps/Half duplex)
Nov 08 00:21:33 <KanjiMonster> there's definitely something fishy going on
Nov 08 00:21:38 <dangole> wtf
Nov 08 00:23:58 <KanjiMonster> dangole: two seconds later it switched back to 1000/full, and it's obviously working
**** ENDING LOGGING AT Thu Nov 08 03:00:01 2012