**** BEGIN LOGGING AT Wed Jan 20 02:59:59 2016
Jan 20 08:41:30 <OpenWrt-Commits> nbd r48389 trunk/target/linux/lantiq/config-4.4 * lantiq: Fix enabling CONFIG_MTD_UIMAGE_SPLIT
Jan 20 10:16:06 <OpenWrt-Commits> nbd r48390 trunk/package/libs/uclient/Makefile
Jan 20 10:16:06 <OpenWrt-Commits> uclient: update to the latest version, adds an auth reconnect handling fix and a small uclient-fetch command line fix
Jan 20 10:16:11 <OpenWrt-Commits> nbd r48391 trunk/package/network/ipv6/ 6in4/Makefile 6in4/files/6in4.sh * 6in4: use uclient-fetch instead of wget/curl
Jan 20 10:16:41 <jow_laptop> thanks nbd
Jan 20 10:32:59 <OpenWrt-Commits> nbd r48392 branches/chaos_calmer/ target/linux/generic/patches-3.18/904-debloat_dma_buf.patch package/kernel/linux/modules/video.mk package/kernel/linux/modules/other.mk * kernel: modularize dma shared buffer code to eliminate kernel bloat
Jan 20 10:33:31 <txomon> hey, I have updated https://wiki.openwrt.org/doc/techref/libubox with descriptions, so if someone could check and ACK I would be grateful!
Jan 20 13:50:09 <stintel> nbd: fyi, I have no objections against dropping oprofile, but I still couldn't compile perf on ar71xx yesterday, iirc
Jan 20 13:50:56 <stintel> actually I think it was one of the dependencies, I'll let you know when I run into it again
Jan 20 13:51:11 <stintel> or create a ticket, if I find my trac login :P
Jan 20 13:51:26 <nbd> ok
Jan 20 13:51:30 <nbd> show me the error and i'll fix it ;)
Jan 20 14:18:32 <tharvey> nbd: were there changes to my fbcon and imx-drm patch you wanted as well? patchwork says 'changes requested' but I see none other then a request to move backlight to video.mk
Jan 20 14:20:04 <nbd> i assumed you were going to resend the series
Jan 20 14:20:17 <tharvey> ok, I will do it now
Jan 20 14:21:31 <tharvey> nbd: thank you for giving cns3xxx some love
Jan 20 14:21:54 <nbd> no problem
Jan 20 14:22:10 <tharvey> we will run it through some testing
Jan 20 14:23:04 <nbd> most work went into tracking down nasty user space crashes
Jan 20 14:23:22 <nbd> which turned out to be an issue known since october last year that went unfixed
Jan 20 14:23:31 <tharvey> which were caused by something due to porting?
Jan 20 14:23:51 <nbd> there's a security feature that's broken on arm11 mpcore
Jan 20 14:24:00 <nbd> and no one (not even arm people) knows why
Jan 20 14:24:08 <tharvey> oh fun...
Jan 20 14:24:23 <nbd> http://thread.gmane.org/gmane.linux.ports.arm.kernel/470447
Jan 20 14:31:47 <tharvey> is there something special to be done in owrt to get a getty to run on a tty1?
Jan 20 14:32:35 <nbd> throw it in inittab
Jan 20 14:35:28 <tharvey> it gave me a login prompt on fb0 but after entering user just stalled
Jan 20 14:41:11 <tharvey> is procd handling the getty on console?
Jan 20 14:42:09 <tharvey> nbd: patches sent
Jan 20 15:02:58 <tharvey> nbd: something in the last 24 hours has broken glib building - bisecting
Jan 20 15:03:16 <nbd> i recently made some changes to it
Jan 20 15:03:27 <tharvey> ya, saw that
Jan 20 15:03:41 <tharvey> noticed while building gstreamer
Jan 20 15:05:01 <nbd> i'll try to build it on a different host
Jan 20 15:06:40 <nbd> did you do ./scripts/feeds update?
Jan 20 15:06:55 <tharvey> ya
Jan 20 15:07:09 <nbd> and also an update of your tree?
Jan 20 15:07:17 <tharvey> my tree?
Jan 20 15:07:22 <nbd> the openwrt tree
Jan 20 15:07:31 <nbd> both need to be in sync for my changes to work
Jan 20 15:07:58 <tharvey> I did a fresh git pull if thats what you mean
Jan 20 15:08:15 <nbd> ok
Jan 20 15:08:42 <tharvey> but you bring up a good point... when I bisect I shoud remove/update feeds again to keep in sync
Jan 20 15:09:04 <nbd> can you show me the error that you're getting?
Jan 20 15:09:38 <tharvey> ya, give me a few mins to reproduce - it had to do with inconv but I don't think it was that last patch about changing the include guard
Jan 20 15:10:38 <nbd> i see the error now
Jan 20 15:11:02 <tharvey> ok
Jan 20 15:11:11 <nbd> ah
Jan 20 15:11:15 <nbd> the issue is in the host build
Jan 20 15:11:51 <nbd> should be easy to fix
Jan 20 15:11:52 <nbd> testing now
Jan 20 15:15:27 <nbd> tharvey: should be fixed now
Jan 20 15:15:32 <stintel> nbd: saw this flooding on my screen during ar71xx build, should I worry? http://pastebin.com/QrMCrg5c
Jan 20 15:17:14 <tharvey> nbd: where do you push to originally? I seem to not be able to find the freshest git repo... found yesterday its not github/openwrt
Jan 20 15:17:31 <nbd> tharvey: i pushed to the package feed
Jan 20 15:17:36 <tharvey> ahhh
Jan 20 15:18:14 <nbd> and other changes that i make i push to svn
Jan 20 15:18:28 <nbd> which gets synced to github and git.openwrt.org every few minutes
Jan 20 15:22:35 <tharvey> hmm... ok yesterday seemed like github/openwrt was out of date, likely my mistake
Jan 20 15:23:02 <tharvey> currently git://git.openwrt.org/packages.git is way out of date
Jan 20 15:23:35 <mamarley> nbd: I just realized, one thing I forgot to do in qos-scripts was bump the PKG_VERSION.  Maybe it would be a better idea to wait though since I also submitted some other patches?
Jan 20 15:24:32 <tharvey> as github is now used in feeds.default.conf seems like https://dev.openwrt.org/wiki/GetSource should be updated to point to github
Jan 20 15:24:35 <tharvey> ?
Jan 20 15:25:39 <stintel> nbd: and here is the perf build failure on ar71xx: http://pastebin.com/GcunAH7a
Jan 20 15:32:27 <nbd> tharvey: git://git.openwrt.org/packages.git is the repo of old unmaintained packages
Jan 20 15:33:06 <nbd> tharvey: and packages.git isn't mentioned on GetSource
Jan 20 15:33:18 <nbd> because it's just one of multiple feeds
Jan 20 15:54:26 <tharvey> nbd: confirmed glib fine now thx
Jan 20 16:16:50 <KevinDB> Also seeing ar71xx build failure with perf enabled as per http://pastebin.com/GcunAH7a
Jan 20 19:12:23 <OpenWrt-Commits> nbd r48397 trunk/package/network/config/netifd/Makefile
Jan 20 19:12:24 <OpenWrt-Commits> netifd: update to the latest version, adds a cosmetic fix for a wpa related variable
Jan 20 19:12:35 <OpenWrt-Commits> nbd r48398 trunk/target/linux/ oxnas/base-files/etc/board.d/01_leds oxnas/base-files/lib/oxnas.sh oxnas/base-files/etc/diag.sh * oxnas: finally make board names consistent
Jan 20 19:12:43 <OpenWrt-Commits> nbd r48399 trunk/ toolchain/glibc/common.mk package/libs/toolchain/Makefile * toolchain: Reverse glibc/eglibc conditionals to check for eglibc
Jan 20 19:12:47 <OpenWrt-Commits> nbd r48400 trunk/ rules.mk include/subdir.mk * build: move host build directories of package/* to build_dir/target-*/host/
Jan 20 19:12:56 <OpenWrt-Commits> nbd r48401 trunk/ rules.mk include/host-build.mk
Jan 20 19:12:56 <OpenWrt-Commits> host-build.mk: change default installation prefix of package/* host builds to staging_dir/target-*/host
Jan 20 19:13:04 <OpenWrt-Commits> nbd r48402 trunk/package/libs/libiconv/Makefile * libiconv: install to the new prefix
Jan 20 19:13:18 <OpenWrt-Commits> nbd r48403 trunk/include/autotools.mk * include/autoconf.mk: decouple host fixups from target fixups
Jan 20 19:32:07 <tharvey> nbd: '8b152c5 kernel: modularize dma shared buffer code to eliminate kernel bloat' breaks my drm patch - I'm not clear why yet but drm requires that to be static and not a module, any ideas?
Jan 20 19:32:24 <nbd> any details on the breakage?
Jan 20 19:32:35 <nbd> maybe it could also just be the module load order?
Jan 20 19:32:45 <OpenWrt-Commits> rmilecki r48404 branches/chaos_calmer/package/kernel/mac80211/patches/863-brcmfmac-fix-setting-primary-channel-for-80-MHz-widt.patch * mac80211: fix brcmfmac picking wrong channel when using 80 MHz
Jan 20 19:32:55 <tharvey> drm-shared-buf fails loading due to missing symbols from DEBUG_FS
Jan 20 19:33:38 <nbd> what symbols are missing?
Jan 20 19:35:25 <tharvey> http://pastebin.com/TfanebNL
Jan 20 19:35:39 <nbd> ah
Jan 20 19:35:44 <tharvey> doesn't make sense to me... debugfs is static
Jan 20 19:35:49 <nbd> looks like it simply needs a MODULE_LICENSE("GPL")
Jan 20 19:36:01 <nbd> since it depends on EXPORT_SYMBOL_GPL symbols
Jan 20 19:36:03 <tharvey> yes, saw the taint as well
Jan 20 19:36:15 <tharvey> ah... if non gpl it won't link with gpl module debugfs?
Jan 20 19:36:40 <nbd> right
Jan 20 19:36:51 <OpenWrt-Commits> nbd r48405 trunk/package/utils/lua/Makefile * lua: use $(STAGING_DIR)/host instead of $(STAGING_DIR_HOST)
Jan 20 19:36:56 <OpenWrt-Commits> nbd r48406 trunk/package/libs/gettext/Makefile * gettext: use $(STAGING_DIR)/host instead of $(STAGING_DIR_HOST)
Jan 20 19:37:00 <OpenWrt-Commits> nbd r48407 trunk/package/libs/gettext-full/Makefile * gettext-full: use $(STAGING_DIR)/host instead of $(STAGING_DIR_HOST)
Jan 20 19:37:04 <OpenWrt-Commits> nbd r48408 trunk/package/utils/bzip2/Makefile * bzip2: use $(STAGING_DIR)/host instead of $(STAGING_DIR_HOST)
Jan 20 19:37:15 <tharvey> yuck
Jan 20 19:37:42 <tharvey> how come you didn't see that?
Jan 20 19:38:09 <nbd> probably was distracted and didn't actually try to load the module
Jan 20 19:38:15 <nbd> just let the kernel check the cross dependencies
Jan 20 19:38:18 <nbd> normally it complains if a symbol is missing
Jan 20 19:38:25 <nbd> but it doesn't validate the license stuff at compile time
Jan 20 19:38:30 <tharvey> heh... when did EXPORT_SYMBOL change to EXPORT_SYMOBOL_GPL? that's what caused this right?
Jan 20 19:38:40 <nbd> it was always EXPORT_SYMBOL_GPL
Jan 20 19:38:52 <nbd> in-kernel code doesn't need MODULE_LICENSE
Jan 20 19:38:58 <nbd> since it's always considered GPL
Jan 20 19:39:05 <nbd> only with modules there's runtime linking involved
Jan 20 19:39:08 <nbd> and that checks the license
Jan 20 19:39:14 <tharvey> true, the symbol checking must not be working properly
Jan 20 19:41:53 <tharvey> hmmm... why hasn't mainline seen this? Should see it with CONFIG_DRM=m
Jan 20 19:43:00 <nbd> because it only breaks with my patch
Jan 20 19:43:10 <nbd> mainline forces the dma-buf code to always be linked into the kernel
Jan 20 19:43:16 <nbd> my patch turns it into a module
Jan 20 19:44:46 <tharvey> right... drm/Kconfig has 'select DMA_SHARED_BUFFER' which would make it static regardless of DRM=m
Jan 20 19:45:16 <nbd> it's the fact that DMA_SHARED_BUFFER is a bool that accomplishes that
Jan 20 19:45:46 <nbd> i'll have a patch in a minute
Jan 20 19:45:49 <tharvey> makes sense... just how much bloat was this causing?
Jan 20 19:45:49 <nbd> can you test it for me?
Jan 20 19:45:52 <tharvey> sure
Jan 20 19:46:01 <nbd> it was causing about 5k after lzma
Jan 20 19:46:04 <tharvey> you going to add a MODULE_LICENSE("GPL"); to end of dma-buf.c?
Jan 20 19:46:06 <nbd> yes
Jan 20 19:46:58 <tharvey> ok afterward will you push my drm patchset before you break something else? ;P
Jan 20 19:46:59 <nbd> http://nbd.name/dma-buf-fix.patch
Jan 20 19:47:28 <tharvey> thought I was going crazy that my drm patch worked yesterday and not today!
Jan 20 19:47:31 <nbd> we'll see. there's probably more stuff that i need to break ;)
Jan 20 19:47:37 <tharvey> he
Jan 20 19:47:39 <tharvey> heh
Jan 20 19:48:02 <nbd> currently working on a pretty intrusive change
Jan 20 19:48:12 <nbd> to properly separate tools/ host-builds from package/ host-builds
Jan 20 19:49:10 <tharvey> is there a timeframe for next release? lots of big changes happening :)
Jan 20 19:49:36 <nbd> no real timeframe
Jan 20 19:49:44 <nbd> before the next release can happen, we need to get all targets on 4.4 anyway
Jan 20 20:08:18 <plntyk> yeah that different host build thing is slightly confusing - and i think package/host builds do not have build logging compared to tools building
Jan 20 20:09:39 <tharvey> nbd: patch is good
Jan 20 20:10:25 <OpenWrt-Commits> nbd r48409 trunk/target/linux/ generic/patches-4.3/904-debloat_dma_buf.patch generic/patches-3.18/904-debloat_dma_buf.patch generic/patches-4.1/904-debloat_dma_buf.patch generic/patches-4.4/904-debloat_dma_buf.patch * kernel: fix module load issue in the dma-buf debloat patch
Jan 20 20:14:22 <OpenWrt-Commits> nbd r48410 trunk/ rules.mk include/subdir.mk * build: use a separate variable checking the subdir path (for host packages)
Jan 20 20:17:40 <nbd> plntyk: i'll look into that
Jan 20 20:27:49 <OpenWrt-Commits> nbd r48411 branches/chaos_calmer/target/linux/generic/patches-3.18/904-debloat_dma_buf.patch * kernel: fix module load issue in the dma-buf debloat patch
Jan 20 20:47:01 <OpenWrt-Commits> nbd r48412 trunk/include/autotools.mk * Revert "include/autoconf.mk: decouple host fixups from target fixups"
Jan 20 21:15:14 <tharvey> nbd did you break mklibs too? ;)
Jan 20 21:20:08 <nbd> yeah, but i fixed it again
Jan 20 21:20:09 <nbd> ;)
Jan 20 21:20:35 <nbd> i'd recommend a make dirclean at this point
Jan 20 21:20:45 <tharvey> i've been doing that each repo update ;)
Jan 20 21:20:48 <nbd> ah
Jan 20 21:20:51 <tharvey> distclean
Jan 20 21:20:57 <nbd> distclean is unnecessary
Jan 20 21:21:04 <nbd> dirclean is enough
Jan 20 21:28:31 <KevinDB> Oooh! It compiles again...except for perf :)
Jan 20 21:29:57 <tharvey> still seeing breakages in host tools here... make-ext4fs refusing to build
Jan 20 21:32:15 <OpenWrt-Commits> nbd r48413 trunk/tools/Makefile * tools: compile flock before everything else
Jan 20 21:32:20 <OpenWrt-Commits> nbd r48414 trunk/rules.mk
Jan 20 21:32:20 <OpenWrt-Commits> rules.mk: make the locked template available even if flock has not been built yet (fall back to unlocked shell command)
Jan 20 21:32:25 <OpenWrt-Commits> nbd r48415 trunk/include/subdir.mk * build: split BUILD_LOG logs per variant
Jan 20 21:32:32 <OpenWrt-Commits> nbd r48416 trunk/include/download.mk * build: add locking for downloads (fixes race conditions with multiple variants)
Jan 20 21:32:38 <OpenWrt-Commits> nbd r48417 trunk/include/subdir.mk * build: add logging for package host builds
Jan 20 21:32:50 <nbd> tharvey: what rev did you build?
Jan 20 21:33:06 <tharvey> 48412
Jan 20 21:33:15 <nbd> hm
Jan 20 21:33:19 <nbd> please post the build error log
Jan 20 21:34:03 <nbd> i need to head out now, will look at it maybe in an hour or so
Jan 20 21:34:15 <tharvey> http://pastebin.com/T7RM8LDa
Jan 20 21:59:41 <tharvey> nbd: a make distclean seems to have resolved my build failure
Jan 20 22:00:39 <tharvey> hi luka! there's a new breakage with my drm patch due to a change yesterday to modularize dma-shared-buf - posting update now
Jan 20 22:07:18 <tharvey> luka nbd drm patches updated - perhaps you can look at them before something gets broken again :)
Jan 20 23:00:10 <OpenWrt-Commits> nbd r48418 trunk/package/utils/bzip2/Makefile * bzip2: install to STAGING_DIR/host instead of STAGING_DIR/host/usr
Jan 20 23:00:16 <OpenWrt-Commits> nbd r48419 trunk/package/devel/valgrind/patches/130-mips_fix_soft_float.patch * valgrind: fix build with mips soft-float
Jan 20 23:21:45 <OpenWrt-Commits> nbd r48420 trunk/package/libs/gettext-full/patches/000-relocatable.patch * gettext-full: fix relocatable patch
Jan 20 23:21:51 <OpenWrt-Commits> nbd r48421 trunk/include/autotools.mk * include/autotools.mk: fix gettext installation path
Jan 20 23:51:17 <OpenWrt-Commits> nbd r48422 trunk/package/devel/binutils/Makefile * binutils: use nls.mk to fix libbfd link errors in other packages
**** ENDING LOGGING AT Thu Jan 21 02:59:58 2016