**** BEGIN LOGGING AT Thu Mar 13 02:59:58 2014
Mar 13 12:53:07 <rogue2> I need genius
Mar 13 12:53:25 <pompomJuice> for some reason
Mar 13 12:53:34 <pompomJuice> when I build my kernel
Mar 13 12:53:41 <pompomJuice> and it gest to the install stage
Mar 13 12:53:50 <pompomJuice> I get make errors that looks like this
Mar 13 12:54:02 <pompomJuice> make[3]: *** /mnt/w/danny/setup-scripts/build/tmp-angstrom_v2013_06-eglibc/work/impedo_duo-angstrom-linux-gnueabi/linux-android-omap3/2.6.37-psp04.04.00.01-r46/image/usr/src/kernel/drivers/usb/atm/Makefile: Is a directory.  Stop.
Mar 13 12:54:19 <pompomJuice> after investigations
Mar 13 12:55:25 <pompomJuice> it seems that kernel.bbclass does a copy of the kernel src into ${D}
Mar 13 12:55:33 <pompomJuice> it does so with this command:
Mar 13 12:55:44 <pompomJuice> find . -depth -not -name "*.cmd" -not -name "*.o" -not -path "./.*" -print0 | cpio --null -pdu $kerneldir
Mar 13 12:55:51 <pompomJuice> shortly after this command
Mar 13 12:55:55 <pompomJuice> it does this command
Mar 13 12:56:05 <pompomJuice> oe_runmake -C $kerneldir CC="${KERNEL_CC}" LD="${KERNEL_LD}" clean
Mar 13 12:56:10 <pompomJuice> therefor
Mar 13 12:56:13 <pompomJuice> it copies the source
Mar 13 12:56:25 <pompomJuice> then runs a clean on it (hell knows why not the other way around)
Mar 13 12:56:30 <pompomJuice> the problem is
Mar 13 12:56:39 <pompomJuice> always makefiles end up as directories
Mar 13 12:57:01 <pompomJuice> somehow the cpio creates directories instead of files and it happens randomly
Mar 13 12:57:05 <pompomJuice> not always the same
Mar 13 12:57:19 <pompomJuice> most of the times it never happens
Mar 13 12:57:37 <pompomJuice> if I do a bitbake -c cleansstate virtual/kernel
Mar 13 12:57:46 <pompomJuice> then bitbake virtual/kernel
Mar 13 12:58:06 <pompomJuice> then makefile directory error followed by another cleasstate and nother build that would succeed
Mar 13 12:58:10 <pompomJuice> so its pretty random
Mar 13 12:58:21 <pompomJuice> what the hell
Mar 13 13:06:45 <XorA> pompomJuice: reiserfs?
Mar 13 13:08:27 <pompomJuice> uuh
Mar 13 13:08:34 <pompomJuice> on my build system
Mar 13 13:08:35 <pompomJuice> let me check
Mar 13 13:08:45 <XorA> very unlikely if you dont know
Mar 13 13:09:02 <pompomJuice> ext4
Mar 13 13:09:21 <pompomJuice> I jsut had a brainwave
Mar 13 13:09:34 <XorA> that is new and interesting behaviour, there are build machines crunching those recipes
Mar 13 13:10:03 <pompomJuice> the other day llvm failed misrably because I have my PARALLEL_MAKE     = "-j8"
Mar 13 13:10:35 <pompomJuice> not sure if that setting gets suck into oe_runmake?
Mar 13 13:10:38 <pompomJuice> most likely
Mar 13 13:10:45 <pompomJuice> let me override it
Mar 13 13:10:47 <pompomJuice> see what happens
Mar 13 13:11:00 <pompomJuice> but I only want to verride for the install stage
Mar 13 13:11:01 <pompomJuice> not sure how
Mar 13 13:11:03 <XorA> the equates to make -j8
Mar 13 13:11:07 <pompomJuice> yea
Mar 13 13:11:17 <pompomJuice> make clean -j8?
Mar 13 13:11:24 <XorA> kernel Makefile is tested thousands of times a day for parrellel make
Mar 13 13:11:26 <pompomJuice> that could cause issues on a kernel folder?
Mar 13 13:11:32 <pompomJuice> this kernel is old
Mar 13 13:11:51 <pompomJuice> do they run clean with -j8 though?
Mar 13 13:11:55 <XorA> ah I see its a TI kernel, they do love to break things
Mar 13 13:11:57 <pompomJuice> since that does not make any sense
Mar 13 13:12:02 <pompomJuice> yea
Mar 13 13:12:06 <pompomJuice> unfortunately
Mar 13 13:12:09 <XorA> I run clean with -j4 all the time
Mar 13 13:12:20 <XorA> cant be bothered with different macro
Mar 13 13:13:05 <pompomJuice> damn
Mar 13 13:14:36 <XorA> very very odd, bring this up on #oe
Mar 13 13:25:03 <pompomJuice> will do
Mar 13 13:27:44 <pompomJuice> see 
Mar 13 13:27:51 <pompomJuice> there the build now succeeded
Mar 13 13:27:58 <pompomJuice> must be a race condition of some sort
**** ENDING LOGGING AT Fri Mar 14 02:59:59 2014