**** BEGIN LOGGING AT Wed Oct 02 02:59:58 2013
Oct 02 03:25:03 <eigma> are there emulators for ar71xx? (qemu?)
Oct 02 05:13:14 <builder> build #338 of kirkwood is complete: Success [build successful]  Build details are at http://buildbot.openwrt.org:8010/builders/kirkwood/builds/338
Oct 02 05:29:53 <eigma> sigh, ag71xx kernel panics when loaded as a module
Oct 02 06:15:45 <OpenWrt-Commits> cyrus r38283 trunk/package/network/ ipv6/6relayd/files/6relayd.config ipv6/6relayd/Makefile * 6relayd: remove compat ULA setting by default (too many bad sideeffects)
Oct 02 07:12:02 <eigma> I think 305-mips_module_reloc.patch breaks CONFIG_KALLSYMS=y
Oct 02 07:12:41 <eigma> somehow because of the way the module is allocated and laid out, the loop in add_kallsyms clobbers the struct module
Oct 02 09:00:39 <soma> hi, if i want to get the default gateway from specific routing tables in a c program, what would be the best way? should work on openwrt and normal linux. netlink socket?
Oct 02 09:02:03 <cyrusff> soma: yes
Oct 02 09:03:50 <builder> build #301 of ep93xx is complete: Success [build successful]  Build details are at http://buildbot.openwrt.org:8010/builders/ep93xx/builds/301
Oct 02 09:43:44 <builder> build #388 of lantiq is complete: Failure [failed compile_5]  Build details are at http://buildbot.openwrt.org:8010/builders/lantiq/builds/388
Oct 02 12:00:05 <mwalle> luka: i saw your latest patches for kirkwood 3.10 support. nice :)
Oct 02 12:00:41 <mwalle> luka: who is supposed to deliver a device tree binary? will it build by openwrt?
Oct 02 12:00:51 <mwalle> +be
Oct 02 12:02:57 <mwalle> the reason i'm asking, is because i would like to support the Buffalo Linkstation LS-XHL
Oct 02 12:03:17 <mwalle> and the support is already in the vanilla kernel
Oct 02 12:06:38 <KanjiMonster> mwalle: it's partially by me
Oct 02 12:07:49 <KanjiMonster> ah, there isn't everything yet
Oct 02 12:13:17 <OpenWrt-Commits> cyrus r38284 trunk/ (5 files in 5 dirs) * Add package signing infrastructure
Oct 02 12:18:35 <mwalle> KanjiMonster: will there be seperate dtbs or will they be appended to the kernel? or will this depend on the target (because the bootloader can't handle device trees, for example)
Oct 02 12:21:32 <KanjiMonster> mwalle: most devices will need appended dtbs
Oct 02 12:22:26 <luka> KanjiMonster: ah... i'm waiting till new uboot is out
Oct 02 12:22:57 <KanjiMonster> luka: but new uboot isn't an option for most targets
Oct 02 12:23:37 <luka> i know
Oct 02 12:23:51 <luka> maybe we should chainload new uboot
Oct 02 12:24:12 <KanjiMonster> what's wrong with appending dtb to the kernel?
Oct 02 12:24:50 <KanjiMonster> even the device tree maintainers think that that should have been the default
Oct 02 12:25:52 <KanjiMonster> and allowing bootloader passed dtbs was a mistake before the bindings are complete (or at least stable enough)
Oct 02 12:26:43 <luka> default uboots dont support ubi
Oct 02 12:27:22 <luka> i havent slept much and i'm talking nonsense
Oct 02 12:28:13 <KanjiMonster> either you have to read a second stage uboot raw from flash or the kernel; I don't see much of a difference ;p
Oct 02 12:28:15 <luka> KanjiMonster: did you meant to upstream the patch which is the last in the series
Oct 02 12:28:40 <mwalle> btw the hardware detection in target/linux/kirkwood/base-files/lib/kirkwood.sh doesn't work with device trees anymore
Oct 02 12:28:42 <luka> i was thinking making uimage out of uboot
Oct 02 12:28:52 <luka> mwalle: we know
Oct 02 12:28:58 <mwalle> ok :)
Oct 02 12:29:24 <mwalle> and chainloading uboot isnt recommended on kirkwood, iirc
Oct 02 12:29:44 <luka> KanjiMonster has added a script that manages both DT and non-DT but i would prefer if we had only DT variant so i didnt push it
Oct 02 12:30:09 * luka does not have device with original uboot
Oct 02 12:31:10 <mwalle> btw where do i define the buttons? eg the one which is used for failsafe mode?
Oct 02 12:32:48 <KanjiMonster> luka: the non-DT is for bootloader passed DTBs (else the embedded DT would override the bootlader passed one I think)
Oct 02 12:33:08 <KanjiMonster> mwalle: all of them are used for failsafe ;)
Oct 02 12:33:46 <mwalle> KanjiMonster: ah so any hotplug button event will trigger failsafe mode?
Oct 02 12:34:21 <luka> KanjiMonster: i was refering to the /proc/devicetree stuff
Oct 02 12:34:23 <KanjiMonster> right
Oct 02 12:34:55 <mwalle> ah, cool. so where do i define the leds? i guess that go into the uci defaults?
Oct 02 12:35:05 <KanjiMonster> luka: hm? oh, that one was for backward compatibility to 3.3 (so you don't have to change that part when switching from 3.3 to 3.10)
Oct 02 12:35:13 <KanjiMonster> also to keep the changes at a minimum
Oct 02 12:36:44 <luka> but once we drop 3.3 it wont be needed
Oct 02 12:36:52 <luka> that is why i didn't push that
Oct 02 12:37:12 <luka> i'll add it once 3.10 is default
Oct 02 12:37:27 <KanjiMonster> then you have a revising where kirkwood is broken
Oct 02 12:38:08 <luka> i think we could live with that
Oct 02 12:40:11 <KanjiMonster> also I vote for leaving kirkwood_hardware_name in place instead of editing all callers
Oct 02 12:40:17 <KanjiMonster> (even if it is just a one liner)
Oct 02 12:41:37 <luka> i was thinking to move that function to base-files or somewhere
Oct 02 12:41:46 <luka> so it can be reused by all DT targets
Oct 02 12:41:56 <luka> and change the name of course
Oct 02 12:42:11 <KanjiMonster> yes
Oct 02 12:42:45 <luka> and i wonder why there is no /proc/devicetree on mips
Oct 02 12:42:53 <KanjiMonster> there is
Oct 02 12:43:00 <luka> hmm, are you sure
Oct 02 12:43:10 <luka> last time i checked on lantiq it was not there
Oct 02 12:43:12 <KanjiMonster> well it's a generic kernel config options
Oct 02 12:43:17 <KanjiMonster> -s
Oct 02 12:43:27 <luka> yes, but on the device itself
Oct 02 12:43:35 <KanjiMonster> I already used and accessed it in bcm63xx
Oct 02 12:43:42 <luka> i will double check on lantiq then
Oct 02 12:43:53 <KanjiMonster> sounds like it isn't turned on then
Oct 02 12:44:12 <luka> i thought i did that... anyway i'll check
Oct 02 12:44:41 <KanjiMonster> at least there's nothing preventing proc devicetree on mips
Oct 02 12:44:57 <luka> kk
Oct 02 13:18:43 <mwalle> i'm currently looking at the boot concept of openwrt. my board has only the bootloader in an 512k SPI flash. The actual linux system is loaded from the hard disk. does that fit the openwrt boot concept? i saw, that some targets already have a harddisk flavor.
Oct 02 13:19:38 <mwalle> there seems to be some magic in mount_root/extroot which searches for mtd partitions named rootfs and rootfs_data
Oct 02 13:55:11 <KanjiMonster> mwalle: but this requires the rootfs already being mounted successfully
Oct 02 13:55:49 <KanjiMonster> mwalle: you probably need to provide your own kernel command line to point the rootfs to the harddisk (or wherever it is supposed to be)
Oct 02 14:03:34 <KanjiMonster> you can define one in your dts file (the chosen node allows setting the command line)
Oct 02 14:05:09 <Devastator> device tree is so practical
Oct 02 14:05:35 <DonkeyHotei> indeed
Oct 02 14:09:20 <karlp> it's almost like people took years of experience with things that weren't good enough, and then came up with a solutoin based on that collected wisdom :)
Oct 02 14:17:12 <OpenWrt-Commits> cyrus r38285 trunk/package/network/ipv6/odhcp6c/Makefile * odhcp6c: validate hop-limit on RAs
Oct 02 14:18:07 <OpenWrt-Commits> cyrus r38286 branches/attitude_adjustment/package/odhcp6c/Makefile * AA: odhcp6c: validate hop-limit on RAs
Oct 02 14:22:24 <KanjiMonster> device tree has its downsides
Oct 02 14:22:44 <KanjiMonster> especially since it's basically an ABI, so you have all the legacy crap issues
Oct 02 14:23:01 <KanjiMonster> and there was a lot of crap in the early days
Oct 02 14:40:40 <Devastator> KanjiMonster ABI stands for..?
Oct 02 14:42:59 <Hauke> application binary interface
Oct 02 14:43:00 <KanjiMonster> application binary interface
Oct 02 14:44:21 <KanjiMonster> Devastator: basically every last binding ever supported once by the kernel must be supported now and forever regardless how "broken" or "wrong" it is, because there might be a bootloader out there passing a dtb using that binding
Oct 02 14:47:30 <Devastator> will look into it
Oct 02 14:48:06 <Devastator> btw, just finished the buy of wndr3400, I think it will be a good unit for testings :)
Oct 02 14:55:39 <Devastator> not that great, no gigabit heheh
Oct 02 14:56:26 <Devastator> but it has usb and well, I will find some place to use it
Oct 02 16:39:36 <OpenWrt-Commits> cyrus r38287 trunk/package/Makefile * package/index: fix index creating when building without signing
Oct 02 16:58:12 <mwalle> KanjiMonster: "this" ?
Oct 02 17:00:40 <mwalle> i already have root=/dev/sda1 set. so i don't need any additional settings? and sth like 'reset to factory defaults' works too?
Oct 02 17:06:09 <mwalle> luka: do you have a public repository where i can find the rest of the kirkwood 3.10 patches?
Oct 02 17:06:46 <mwalle> i guess there are still some patches left, eg. the kernel version is still 3.3.8 in the makefile
Oct 02 17:32:14 <builder> build #419 of brcm63xx is complete: Success [build successful]  Build details are at http://buildbot.openwrt.org:8010/builders/brcm63xx/builds/419
Oct 02 17:49:12 <builder> build #367 of rb532 is complete: Failure [failed shell_12]  Build details are at http://buildbot.openwrt.org:8010/builders/rb532/builds/367
Oct 02 17:51:43 <builder> build #367 of ppc44x is complete: Failure [failed shell_12]  Build details are at http://buildbot.openwrt.org:8010/builders/ppc44x/builds/367
Oct 02 18:12:00 <KanjiMonster> mwalle: http://88.73.137.4/~jonas/kw-3.10/ is my wip state (which luka uses as a base)
Oct 02 18:15:38 <mwalle> KanjiMonster: thx, if i add an subtarget "harddisk" to kirkwood, i'll need to add also a subtarget "generic", right?
Oct 02 18:15:54 <KanjiMonster> yes
Oct 02 18:15:59 <builder> build #415 of at91 is complete: Success [build successful]  Build details are at http://buildbot.openwrt.org:8010/builders/at91/builds/415
Oct 02 18:16:19 <mwalle> i guess the generic subtarget can be empty, except for the target.mk descrption?
Oct 02 18:46:57 <luka> mwalle: dont add subtargets to kirkwood
Oct 02 19:45:59 <builder> build #346 of xburst is complete: Failure [failed compile_9]  Build details are at http://buildbot.openwrt.org:8010/builders/xburst/builds/346
Oct 02 19:50:45 <builder> build #347 of sibyte is complete: Failure [failed shell_12]  Build details are at http://buildbot.openwrt.org:8010/builders/sibyte/builds/347
Oct 02 20:05:24 <builder> build #389 of lantiq is complete: Failure [failed compile_5]  Build details are at http://buildbot.openwrt.org:8010/builders/lantiq/builds/389
Oct 02 20:42:38 <mwalle> luka: so what should i do instead?
Oct 02 20:42:54 <mwalle> add config_ata and the like to the config-3.10?
Oct 02 22:25:09 <OpenWrt-Commits> hauke r38288 trunk/target/linux/ (13 files in 3 dirs) * kernel: bgmac: move bgmac patches already applied in mainline kernel to generic
Oct 02 22:25:38 <builder> build #388 of ramips is complete: Success [build successful]  Build details are at http://buildbot.openwrt.org:8010/builders/ramips/builds/388
Oct 02 22:26:33 <OpenWrt-Commits> hauke r38289 trunk/target/linux/ (8 files in 3 dirs) * kernel: bgmac: move other bgmac patches to generic
Oct 02 22:27:22 <OpenWrt-Commits> hauke r38290 trunk/target/linux/ (5 files in 5 dirs) * kernel: bcma: update to wireless-testing master-2013-10-01
Oct 02 22:27:50 <OpenWrt-Commits> hauke r38291 trunk/target/linux/ (11 files in 4 dirs) * kernel: refresh patches
Oct 02 23:13:53 <OpenWrt-Commits> hauke r38292 trunk/target/linux/generic/patches-3.10/772-bgmac-add-supprot-for-BCM4707.patch * kernel: bgmac: fix BCM4707 patch
Oct 02 23:15:15 <OpenWrt-Commits> hauke r38293 trunk/package/kernel/broadcom-diag/Makefile * broadcom-diag: fix 'autoload' number of diag.ko
Oct 02 23:23:27 <OpenWrt-Commits> hauke r38294 trunk/target/linux/brcm47xx/base-files/etc/init.d/netconfig * brcm47xx: add support for ipv6 in default network config
Oct 02 23:39:04 <builder> build #380 of uml is complete: Failure [failed compile_5]  Build details are at http://buildbot.openwrt.org:8010/builders/uml/builds/380
Oct 03 02:37:23 <builder> build #357 of avr32 is complete: Failure [failed compile_5]  Build details are at http://buildbot.openwrt.org:8010/builders/avr32/builds/357
Oct 03 02:51:08 <OpenWrt-Commits> luka r38295 trunk/target/linux/ (43 files in 2 dirs) * kirkwood: update patches and configuration (3.10)
Oct 03 02:51:10 <OpenWrt-Commits> luka r38296 trunk/target/linux/kirkwood/image/Makefile * kirkwood: fix initramfs images
Oct 03 02:51:11 <OpenWrt-Commits> luka r38297 trunk/target/linux/kirkwood/image/Makefile * kirkwood: enable dts compilation
**** ENDING LOGGING AT Thu Oct 03 02:59:58 2013