**** BEGIN LOGGING AT Wed Nov 13 02:59:58 2013
Nov 13 03:02:05 <builder> build #9 of mvebu is complete: Failure [failed compile_4]  Build details are at http://buildbot.openwrt.org:8010/builders/mvebu/builds/9
Nov 13 03:06:34 <russell--> r38309 didn't build
Nov 13 04:19:45 <russell--> r38308 boots, r38309 doesn't build, r38310 builds but doesn't boot on wgt634u
Nov 13 04:20:06 <russell--> where is Hauke?
Nov 13 04:21:40 <DonkeyHotei> germany, i think
Nov 13 04:22:03 <DonkeyHotei> it's early morning there
Nov 13 04:25:15 <russell--> yeah, i know
Nov 13 04:30:26 <MartijnVdS> very early morning
Nov 13 04:30:30 <MartijnVdS> (5:30)
Nov 13 04:47:40 <russell--> when i had jet lag last month in germany, i was waking up that early!
Nov 13 05:26:22 <MartijnVdS> Hmm.. ar8327_init_globals initializes the max MTU to 1518+8+2 (so, basically, 1500 for users)
Nov 13 05:26:39 <MartijnVdS> All other switch chips in that file get 9018+8+2 (MTU 9000)
Nov 13 05:27:46 <MartijnVdS> What would break if I increased it for ar8327? (I hope "probably nothing, as long as you don't set the MTU > 1500, and even MTU >1500 might just work")
Nov 13 06:16:16 <russell--> hmm.  looks like my weird local-destination problem is related to switch configuration
Nov 13 06:16:48 <russell--> if i remove the switch configuration and use the raw eth0 as a wan interface, everything works
Nov 13 08:28:01 <russell--> KanjiMonster: ping
Nov 13 09:54:57 <nbd> tripolar: pong
Nov 13 10:30:58 <Guest19371> nbd r38784 trunk/target/linux/ramips/Makefile * ramips: select kmod-rt2800-soc by default (if available)
Nov 13 10:51:09 <Guest19371> wigyori r38785 trunk/package/ boot/uboot-sunxi boot/uboot-sunxi/Makefile * [package]: uboot-sunxi: add U-boot builds for a couple sunxi boards
Nov 13 10:51:11 <Guest19371> blogic r38786 trunk/package/ system/procd/files/procd.sh system/procd/Makefile * procd: update to latest git head
Nov 13 10:51:28 <Guest19371> blogic r38787 trunk/package/ network/services/dropbear/files/dropbear.init utils/busybox/files/sysntpd * procd: convert services to the new validation api
Nov 13 10:59:03 <Guest19371> blogic r38788 trunk/package/system/procd/files/procd.sh
Nov 13 10:59:03 <Guest19371> procd: options with default values are not parsed properly in the ubus register code
Nov 13 11:09:11 <Guest19371> blogic r38789 trunk/package/ network/services/dropbear/files/dropbear.init utils/busybox/files/sysntpd * procd: fixup 2 wrong option types
Nov 13 11:51:36 <russell--> Hauke: i've been bisecting a problem on netgear wgt634u: r38308 boots, r38309 gave me a build error, r38310 builds but doesn't boot
Nov 13 11:54:58 <Hauke> russell--: thanks, have you tried to change SSB_CHIPCO_UART0_DATA to SSB_CHIPCO_UART1_DATA in arch/mips/bcm47xx/prom.c ?
Nov 13 11:56:18 <russell--> no
Nov 13 11:56:30 <russell--> but, i will
Nov 13 11:56:50 <Hauke> thanks
Nov 13 11:57:46 <russell--> there is also some kind of switch problem in r38308 and earlier ... but i'll see if i can fix the booting problem first
Nov 13 11:59:09 <Hauke> yes lets have a look at the boot problem first
Nov 13 12:00:45 <Devastator> damn, I really need to build a image for my netgear wndr4300
Nov 13 12:00:56 <Devastator> being lazy for too long
Nov 13 12:01:31 <russell--> Hauke: so, i'm just modifying the patch target/linux/brcm47xx/patches-3.10/082-MIPS-BCM47XX-add-EARLY_PRINTK_8250-support.patch
Nov 13 12:01:36 <russell--> by hand
Nov 13 12:01:49 <Hauke> russell--: yes that's ok
Nov 13 12:06:54 <builder> build #431 of lantiq is complete: Failure [failed compile_5]  Build details are at http://buildbot.openwrt.org:8010/builders/lantiq/builds/431
Nov 13 12:09:02 <builder> build #405 of rb532 is complete: Failure [failed shell_12]  Build details are at http://buildbot.openwrt.org:8010/builders/rb532/builds/405
Nov 13 12:11:05 <builder> build #405 of ppc44x is complete: Failure [failed shell_12]  Build details are at http://buildbot.openwrt.org:8010/builders/ppc44x/builds/405
Nov 13 12:14:18 <russell--> Hauke: that fixed the booting
Nov 13 12:14:54 <Hauke> nice
Nov 13 12:15:30 <russell--> yes, nice job, nailed that first try
Nov 13 12:16:06 <KanjiMonster> russell--: pong
Nov 13 12:16:48 <russell--> KanjiMonster: i was about to report the switch problem on wgt634u, which luckily i can now see again ;-)
Nov 13 12:16:53 <Guest19371> blogic r38790 trunk/package/system/procd/Makefile
Nov 13 12:16:53 <Guest19371> procd: remove obselete script install reference
Nov 13 12:16:53 <Guest19371> Signed-off-by: John Crispin <blogic@openwrt.org>
Nov 13 12:17:40 <russell--> i weirdly can only ping the default gateway on the local network, but all remote (off-local-network) hosts work fine
Nov 13 12:18:06 <Hauke> russell--: could you pastebin the output of dmesg
Nov 13 12:18:17 <russell--> until i blow away the switch configuration in /etc/config/network, get a plain eth0 and then it works again
Nov 13 12:18:33 <russell--> Hauke: yes, just a moment
Nov 13 12:21:27 <russell--> http://sprunge.us/aSHW
Nov 13 12:23:10 <builder> build #430 of ramips is complete: Failure [failed compile_5]  Build details are at http://buildbot.openwrt.org:8010/builders/ramips/builds/430
Nov 13 12:23:27 <Hauke> russell--: what is the output of: "swconfig dev switch0 show"
Nov 13 12:25:09 <russell--> http://sprunge.us/gMYJ
Nov 13 12:27:27 <Hauke> russell--: to get networking to work you have to change the network interface from eth0.1 to eth0 ?
Nov 13 12:28:26 <russell--> i just nuke all the config switch* stuff and fix up the interface names and then it works again
Nov 13 12:28:47 <KanjiMonster> with or without a reboot in between?
Nov 13 12:28:53 <russell--> with a reboot
Nov 13 12:29:39 <russell--> i'm editting the /etc/config/network, and that's the way i know how to reload the new config ;-)
Nov 13 12:35:11 <russell--> the problem seemed to be on the receive end.  looking at traffic on the other host, i was seeing traffic coming from the wgt, but it evidently wasn't making it back to the cpu.  tried with/without the firewall active
Nov 13 12:35:49 <russell--> weirdly only the hosts on the local network were affected
Nov 13 12:36:06 <russell--> the default gateway and everything beyond worked normally
Nov 13 12:36:39 <russell--> (i was getting new images on to it by uploading them to a remote server and then scp'ing them back down)
Nov 13 12:40:40 <russell--> not seeing the arp reply, maybe?
Nov 13 12:42:28 <russell--> hmm. i see the arp requests, and the reply going back and more arp requests, but never any ping packets.
Nov 13 12:44:40 <Hauke> russell--: what was your switch config before and after your change?
Nov 13 12:47:33 <russell--> Hauke: this is the default i get after a new flash: http://pastebin.ca/2475820
Nov 13 12:49:19 <russell--> and this is my fix: http://pastebin.ca/2475821
Nov 13 12:50:11 <builder> build #427 of ppc40x is complete: Failure [failed compile_5]  Build details are at http://buildbot.openwrt.org:8010/builders/ppc40x/builds/427
Nov 13 12:50:43 <russell--> and then i can talk to local hosts again
Nov 13 12:52:10 <Hauke> russell--: talking to wan also works with the automatically generated config?
Nov 13 12:53:12 <russell--> all the devices i'm talking to are on the wan interface
Nov 13 12:53:47 <russell--> there isn't anything connected to the lan interface
Nov 13 12:54:07 <russell--> the wifi isn't up and nothing is plugged into ethernet ports
Nov 13 12:54:23 <russell--> except the "wan" port
Nov 13 12:55:27 <russell--> dhcp works in both configurations
Nov 13 12:56:10 <russell--> the wan interface gets an ipaddr on the right network
Nov 13 13:03:00 <Hauke> russell--: could you please try this patch: http://hauke-m.de/files/openwrt/bcm5365-switch.patch
Nov 13 13:03:13 <russell--> yes, trying ...
Nov 13 13:04:23 <Hauke> KanjiMonster: for BCM5365 VA_UNTAG_S is 7 and not 6 in switch-robo.c is that correct?
Nov 13 13:05:27 <KanjiMonster> according to the datasheet it indeed is
Nov 13 13:05:50 <KanjiMonster> also there seems to be no "untag" bit for the cpu port
Nov 13 13:10:51 <Hauke> KanjiMonster: thanks
Nov 13 13:11:38 <Hauke> KanjiMonster: so a mask of 1F for the untagged ports is correct?
Nov 13 13:11:47 <KanjiMonster> yes
Nov 13 13:12:55 <KanjiMonster> but the reserved bit that would be the cpu untag bit is RW, so I wonder what happens if you use it anyway :p
Nov 13 13:14:43 <russell--> yay, i can ping my local network hosts now
Nov 13 13:14:46 <builder> build #418 of uml is complete: Failure [failed compile_5]  Build details are at http://buildbot.openwrt.org:8010/builders/uml/builds/418
Nov 13 13:17:03 <builder> build #376 of sibyte is complete: Failure [failed shell_12]  Build details are at http://buildbot.openwrt.org:8010/builders/sibyte/builds/376
Nov 13 13:17:22 <Hauke> russell--: I will commit the patches in the evening
Nov 13 13:17:38 <Hauke> so the regressions on your device are fixed with these patches?
Nov 13 13:19:42 <russell--> i think so
Nov 13 13:19:52 * russell-- will let you know if i encounter more
Nov 13 13:19:59 <russell--> thanks!
Nov 13 13:20:14 <KanjiMonster> Hauke: jow_laptop: I wonder how we want to/can handle switches properly that don't support cpu port untagged when vlan support is enabled
Nov 13 13:21:55 <Hauke> add some port status which contains the ability to work in tagged/untagged mode?
Nov 13 13:22:58 <Hauke> for the adm6996l you can only specify port to be generally untagged or generally tagged, this can not be done on a per vlan base
Nov 13 13:26:17 <KanjiMonster> Hauke: I meant the cpu port is always tagged, even when using only one vlan; only by setting the switch into the "dumb" switch mode you can use it untagged
Nov 13 13:27:43 <Hauke> I do not see a big problem there, just the user is able to configure the network in an other way that it does not work any more
Nov 13 13:28:39 <KanjiMonster> right, except it isn't obvious why it doesn't work
Nov 13 13:41:40 * russell-- sleeps
Nov 13 14:43:06 <tripolar> ping nbd
Nov 13 14:48:08 <Guest19371> kaloz r38791 trunk/package/kernel/mac80211/Makefile * [mac80211]: add another symbol for TI chipsets
Nov 13 14:49:18 <Guest19371> kaloz r38792 trunk/package/kernel/mac80211/Makefile * [mac80211]: upgrade linux-firmware git revision, bump package version
Nov 13 15:32:15 <tripolar> blogic: i have found a bug on the ap pro - its not possible to use the "sysupgrade" feature to append the saved config during a flash of the firmware - and with nbd's help i could locate the cause, it's because it has 2 jffs2 partitions (one for the kernel the other one for the overlayfs and mtd writes to the first jffs2 partition it finds (so it appends the data to the kernel partiton), i could create a fix in mtd.c - with an option  "-jp <Number
Nov 13 15:39:04 <nbd> tripolar: pong
Nov 13 15:47:45 <tripolar> nbd i created a patch but i don't know where to set the option in sysupgrade - i posted a long text a few lines above your "pong"
Nov 13 15:48:43 <nbd> i think your line was cut off
Nov 13 15:48:52 <nbd> it ends with: with an option  "-jp <Number
Nov 13 15:49:02 <tripolar> with an option  "-jp <Number>" added  which allows me to select the jffs2 parition   to use during the flash. i'm thinking about where i set this option in sysupgrade - as the firmware image is piped to mtd my idea to parse the file and find out how many
Nov 13 15:49:14 <tripolar>                      jffs2 partitons the file doesent worked as i can't lseek back in the stream. any ideas?
Nov 13 15:50:17 <nbd> the place to put it is target/linux/ar71xx/base-files/lib/upgrade/platform.sh
Nov 13 15:50:32 <nbd> you could make an environment variable for extra mtd command options
Nov 13 15:50:36 <nbd> and only set it for uap-pro
Nov 13 15:51:14 <tripolar> okay thanks
Nov 13 15:51:48 <nbd> btw. you should pick a one-letter option
Nov 13 15:51:52 <nbd> to make it work with getopt
Nov 13 15:52:02 <tripolar> okay
Nov 13 15:55:28 <tripolar> nbd: is "$ARGV" only holding on value when calling "default_do_upgrade" in platform.sh
Nov 13 15:55:43 <tripolar> as default_do_upgrade seems to use on $1
Nov 13 15:55:47 <tripolar> on/only
Nov 13 15:56:20 <tripolar> ahh my fault didn't think about the ""
Nov 13 16:15:35 <builder> build #386 of avr32 is complete: Failure [failed compile_5]  Build details are at http://buildbot.openwrt.org:8010/builders/avr32/builds/386
Nov 13 18:06:25 <Guest19371> hauke r38793 trunk/target/linux/ generic/files/drivers/net/phy/b53/b53_regs.h generic/files/drivers/net/phy/b53/b53_common.c * kernel: b53: fix untagged shift for BCM5365
Nov 13 18:36:17 <builder> build #325 of iop32x is complete: Failure [failed shell_12]  Build details are at http://buildbot.openwrt.org:8010/builders/iop32x/builds/325
Nov 13 18:40:05 <Guest19371> hauke r38794 trunk/target/linux/ brcm47xx/patches-3.10/830-huawei_e970_support.patch brcm47xx/patches-3.10/082-MIPS-BCM47XX-add-EARLY_PRINTK_8250-support.patch * brcm47xx: swap also early boot console if necessary.
Nov 13 18:44:32 <Guest19371> hauke r38795 trunk/package/system/zram-swap/Makefile * zram-swap: fix dependencies
Nov 13 18:47:28 <mazilo> When I tried to compile yate for AR71XX platform, the uClibc++ failed to compile as shown here (http://pastebin.com/neLqiuCq). Has anyone some ideas how to fix this?
Nov 13 19:28:17 <winni> hi
Nov 13 19:28:19 <blogic> winni: huhu
Nov 13 19:28:30 <winni> huhu :)
Nov 13 20:04:38 <apocn> Hello all, what's the best way to start the DHCP client (on the WAN interface) only when wpa_supplicant is done running and in Authorized state?
Nov 13 20:05:28 <apocn> Right now I added an infinite loop in the netifd/proto/dhcp.sh script, to start wpa_supplicant (if it's not already up) and to exit the loop only when "Authorized".
Nov 13 20:05:35 <jow_laptop> a custom program observing wpa_supp control socket events
Nov 13 20:08:17 <apocn> I'm using wpa_cli to get "suppPortStatus".
Nov 13 20:08:56 <apocn> but, is there a better solution? I don't feel good running this inside the dhcp.sh script…, it actually runs before proto_run_command "udhcpc".
Nov 13 20:09:23 <jow_laptop> not at the moment
Nov 13 20:10:38 <apocn> jow_laptop: ok. then I will create a crontab that will restart the dhcp if suppPortStatus is no longer authorized. Thanks!
Nov 13 21:14:09 <russell--> Hauke: just tested the serial fix for wgt634u, didn't work
Nov 13 21:14:20 <Hauke> :-(
Nov 13 21:15:26 <russell--> at least, i get a hung console in the same spot for the image i just flashed
Nov 13 21:16:01 <russell--> base on r38795
Nov 13 21:16:14 <russell--> based*
Nov 13 21:18:31 <russell--> did a git pull ; make target/linux/clean world
Nov 13 21:56:56 <builder> build #9 of mxs is complete: Failure [failed compile_4]  Build details are at http://buildbot.openwrt.org:8010/builders/mxs/builds/9
Nov 13 22:27:55 <Guest19371> hauke r38796 trunk/ include/target.mk target/linux/bcm53xx/Makefile * bcm53xx: this ARM SoC does not have a FPU, use soft float.
Nov 13 22:37:59 <Guest19371> hauke r38797 trunk/package/kernel/mac80211/files/regdb.txt * mac80211: update regdb.txt
Nov 13 22:57:15 <tripolar> nbd: i have this patch by now and i'm trying sinze hours to find out why mtd troughs a useage error when i use sysupgrade
Nov 13 22:57:18 <tripolar> http://paste.debian.net/65719/
Nov 13 22:57:25 <tripolar> any idea why this happens?
Nov 13 23:06:12 <nbd> tripolar: are you sure you have the patched version of mtd on the system?
Nov 13 23:06:35 <tripolar> yes i think so i will try to reflash it
Nov 13 23:06:46 <nbd> also, does it just show the usage, or that other error string you added as well?
Nov 13 23:07:10 <tripolar> some other error strings i included some removed them before posting
Nov 13 23:07:22 <tripolar> i "real" debugger would be nice :)
Nov 13 23:08:08 <nbd> oh, wait
Nov 13 23:08:25 <nbd> are you sure that busybox ash actually supports ${@:2}?
Nov 13 23:08:40 <tripolar> i tested it in a shell and there it worked
Nov 13 23:08:43 <nbd> ok
Nov 13 23:08:48 <tripolar> i also tried $2 $3
Nov 13 23:12:03 <nbd> in my tests, ${@:2} does not work properly
Nov 13 23:13:46 <tripolar> okay
Nov 13 23:14:02 <nbd> tripolar: http://pastebin.com/giiTpMbU
Nov 13 23:14:29 <nbd> no idea where that bogus 'o' argument came from
Nov 13 23:14:39 <nbd> oh, wait
Nov 13 23:14:50 <nbd> ${@:2} cuts off the first two characters
Nov 13 23:14:55 <nbd> not the first two arguments ;)
Nov 13 23:15:11 <tripolar> okay i will try it with $2 $3
Nov 13 23:19:12 <tripolar> nbd: the simple version works as expected
Nov 13 23:19:13 <tripolar> http://pastebin.com/SYcXsiad
Nov 13 23:19:41 <tripolar> no my fault "as" is a bad example :D
Nov 13 23:19:44 <nbd> ;)
Nov 13 23:20:09 <tripolar> i think that was my problem i tested it only with 1 or two chars for $1 :(
Nov 13 23:21:17 <tripolar> nbd: did you try gentoo on the mac by now?
Nov 13 23:25:01 <nbd> no
Nov 13 23:51:44 <tripolar> nbd: it works now, i sent a patch to the ml, could you apply it?
Nov 13 23:51:58 <tripolar> thanks for your help
Nov 14 00:05:51 <nbd> tripolar: btw. now that i think about it, it would probably be better to pass the extra arguments via a variable
Nov 14 00:06:02 <nbd> instead of passing them as arguments and hardcoding the number of parameters to be processed
Nov 14 00:06:57 <jow_laptop> whats wrong with  img="$1"; shift + get_image "$img" | mtd "$@" -j ... ?
Nov 14 00:07:11 <jow_laptop> that will also properly work for zero or 4 arguments
Nov 14 00:07:22 <tripolar> jow_laptop: didn't try it good idea
Nov 14 00:08:29 <tripolar> what do you mean with "pass the extra arguments via a variable"?
Nov 14 00:09:54 <tripolar> nbd you mean i set a variable in the shell script and read it within mtd?
Nov 14 00:10:24 <nbd> MTD_CONFIG_ARGS="-j 2" just before the default_do_upgrade
Nov 14 00:10:36 <nbd> and then $MTD_CONFIG_ARGS inside the function on the mtd command line
Nov 14 00:10:40 <Guest19371> luka r38798 packages/utils/lxc/ patches/015-getline.patch patches/020-lxc-checkconfig.patch Makefile patches/010-compile.patch * [packages] lxc: update to latest git version
Nov 14 00:11:02 <nbd> i think that's better than using $@ because the arguments are only applied to one of the mtd calls
Nov 14 00:11:22 <tripolar> okay
Nov 14 00:11:26 <tripolar> i will change it
Nov 14 00:30:43 <tripolar> nbd: sent
Nov 14 00:52:21 <Devastator> Guten Abend :)
Nov 14 01:00:29 <tripolar> nbd is the patch now okay?
Nov 14 01:05:36 <Devastator> tripolar Wo wohnen Sie? if you don't mind :)
Nov 14 01:37:09 <Guest19371> luka r38799 trunk/ include/image.mk Config.in * image: don't force ubifs journal size
Nov 14 01:37:10 <Guest19371> luka r38800 trunk/Config.in * Config.in: drop TARGET_UBIFS_SQUASH_UIDS, it's used by default now
**** ENDING LOGGING AT Thu Nov 14 02:59:59 2013