Richard Purdie [Mon, 5 Dec 2011 23:00:30 +0000 (23:00 +0000)]
libsdl: Disable pulseaudio explicitly
Its not listed in DEPENDS so should never have been built. We could
configure this as a configuration option and I'll take a patch for
that but I like deterministic builds so force it off for now.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Nitin A Kamble [Fri, 2 Dec 2011 20:20:09 +0000 (12:20 -0800)]
libatomics-ops: patch source code for x32
This Fixes bug: [YOCTO #1418]
Remove the `q' suffix on x86-64 atomic instructions.
We don't need the `q' suffix on x86_64 atomic instructions for AO_t,
which is defined as "unsigned long". "unsigned long" is 32bit for x32
and 64bit for x86-64. The register operand in x86-64 atomic instructions
is sufficient to properly determine the register size.
Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com> Signed-off-by: H.J. Lu <hjl.tools@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Nitin A Kamble [Fri, 2 Dec 2011 20:20:08 +0000 (12:20 -0800)]
libaio: patch source code for x32
This Fixes bug: [YOCTO #1417]
Properly load arguments 5 an 6 for x86-64 syscall
Use asm ("r10") and asm ("r8") to load arguments 5 an 6 for x86-64
syscall so that it works with both x32 and x86-64.
Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com> Signed-Off-By: H.J. Lu <hjl.tools@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Nitin A Kamble [Fri, 2 Dec 2011 20:20:07 +0000 (12:20 -0800)]
xproto: fix compilation with x32 toolchain
Don't always define LONG64 for AMD64
X32 defines __amd64__/amd64 with 32bit long. We should simply check
__LP64__ before defining LONG64 without checking __amd64__/amd64.
This fixes compilation with x32 toolchain.
Signed-Off-By: H.J. Lu <hjl.tools@gmail.com> Signed-Off-By: Nitin A Kamble <nitin.a.kamble@intel.com> 2011/12/1 Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Nitin A Kamble [Fri, 2 Dec 2011 20:20:05 +0000 (12:20 -0800)]
libxt: fix compilatoin with x32 toolchain
This fixes compilation with x32 toolchain.
Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com> Signed-off-by: H.J. Lu <hjl.tools@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Nitin A Kamble [Fri, 2 Dec 2011 20:20:04 +0000 (12:20 -0800)]
glib-2.0: fix compilatoin with x32 toolchain
Pass along CC & CFLAGS vars so that the tune parameters set get used.
This fixes compilation with x32 toolchain.
Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com> Signed-off-by: H.J. Lu <hjl.tools@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Nitin A Kamble [Fri, 2 Dec 2011 20:20:03 +0000 (12:20 -0800)]
mesa-dri, mesa-xlib: fix compilation with x32 toolchain
Add support for building with x32 toolchain.
Simplified the use of SRC_URI & S vars across multiple files.
Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com> Signed-off-by: H.J. Lu <hjl.tools@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Nitin A Kamble [Fri, 2 Dec 2011 20:20:02 +0000 (12:20 -0800)]
gmp: fix the recipe for x32 target
Add support for building with x32 toolchain.
Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com> Signed-off-by: H.J. Lu <hjl.tools@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Nitin A Kamble [Fri, 2 Dec 2011 20:20:01 +0000 (12:20 -0800)]
openssl-1.0.0e: fix to wotk with x32 toolchain
Add BN_ADDR for address type instead of using BN_ULONG or unsigned long:
1. For W64, address type is unsigned long long, not unsigned long.
2. For x32, address type is unsigned long , not BN_ULONG.
Added a new targetlinux-x32 in the config file
The do_install() code to move lib/* to lib64 is not needed now with the
enhanced multilib support.
Make the x86-64 assembly syntax compatible with x32 compiler.
Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com> Signed-off-by: H.J. Lu <hjl.tools@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Paul Eggleton [Mon, 5 Dec 2011 18:28:12 +0000 (18:28 +0000)]
classes/buildhistory: do not save old packagehistory files by default
Disable storing package history as version named files unless
BUILDHISTORY_KEEP_VERSIONS is set to 1; otherwise the adds of these
files that duplicate what is already in git anyway is just noise in the
git log.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
With the recent set of version increments, the option being passed to
configure no longer matched the patch set. This corrects libx11-trim
so everything is using a consistent option.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Paul Eggleton [Thu, 1 Dec 2011 17:26:44 +0000 (17:26 +0000)]
classes/buildhistory: merge in package history functionality
Include package history collection from packagehistory.bbclass (thus
superseding it). The only change is to store package history under
BUILDHISTORY_DIR/packages and rename one of the functions.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Paul Eggleton [Wed, 30 Nov 2011 16:48:47 +0000 (16:48 +0000)]
classes/buildhistory: add new output history collection class
Create a new build output history reporting class, using testlab.bbclass
from meta-oe as a base. This records information from images produced by
the build process in text files structured suitably for tracking within
a git repository, thus enabling monitoring of changes over time.
Build history collection can be enabled simply by adding the following
to your local.conf:
INHERIT += "buildhistory"
The output after a build can then be found in BUILDHISTORY_DIR (defaults to
TMPDIR/buildhistory). If you set up this directory as a git repository and
set BUILDHISTORY_COMMIT to "1" in local.conf, the build history data will
be committed on every build.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Paul Eggleton [Mon, 21 Nov 2011 16:49:25 +0000 (16:49 +0000)]
classes/packagehistory: fix and extend
* Replace use of BASEPKG_TARGET_SYS which is no longer available
* Replace use of bb.data.getVar(...,d) with d.getVar(...)
* Change the file structure - use single files within PN/package subdirs
rather than having a subdir level for each part of the version. There
is a set of files for each recipe and for each package in directories
underneath.
* Record more information - PACKAGES, DEPENDS, RDEPENDS, RRECOMMENDS,
FILES, and the total size and a complete list of the packaged files.
* Record the values in simple text format. The "latest" file, rather
than a symlink has been changed to a copy of the latest file so that
if it is tracked in a VCS repository (e.g. git) you can compare it
easily to the previous version.
Implements [YOCTO #1565].
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Darren Hart [Fri, 2 Dec 2011 03:20:15 +0000 (19:20 -0800)]
bootimage: Use ${S} explicitly for generated config files
The syslinux and grub-efi classes were generating config files in the current
working directory. This caused a failure due to a race in the creation of the
directories leading to cwd changing and the build failing to find the config
files. While this has been addressed in bitbake, it is better to use an
explicit path.
While ${WORKDIR} may seem a more appropriate place, the recipe
already uses ${S} for the "hdd" and "cd" construction, so we use ${S}
here to keep things consolidated and consistent and address the issue
with minimal change.
Signed-off-by: Darren Hart <dvhart@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Bruce Ashfield [Tue, 22 Nov 2011 22:35:07 +0000 (17:35 -0500)]
kern-tools: use Makefile provided install rules
Previously the install of the kern-tools was manual and fully controlled by
whatever build system was integrating/using the utilities. To make this more
generic a Makefile is now provided to take care of installing and removing to
DESTDIR.
Updating the kern-tools recipe to take advantage of this new facility.
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Bruce Ashfield [Tue, 15 Nov 2011 21:05:39 +0000 (16:05 -0500)]
kern-tools: add pre_config and merge_config.sh to the list of tools
As a transition step to moving kern_tools into the kernel tree
itself this change adopts merge_config.sh as a common base for
merging configuration fragments. So we add merge_config.sh and
pre_config to the list of kern_tools.
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Bruce Ashfield [Sun, 20 Nov 2011 05:30:02 +0000 (00:30 -0500)]
linux-yocto: remove 2.6.34 recipes
To limit the number of active recipes in master to 3 (v2.6.37, v3.0 and
-dev), the 2.6.34 recipe can now be removed. It is available in the
previous release branches and will continue to work, but won't be
updated as part of ongoing dev efforts.
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Bruce Ashfield [Tue, 15 Nov 2011 21:10:06 +0000 (16:10 -0500)]
kernel-yocto: support allnoconfig base configuration
When creating a minimal config or creating a hard baseline for
incremental changes, it is often desired to disable all kernel
options and then begin building and enabling only what is required.
To support this workflow, a new variable KCONFIG_MODE is introduced
to contain a hint to the kernel configuration about how the kernel
config should be produced. This variable is passed directly to lkc
when it is invoked during configuration, so the contents of the
variable must be a valid option for the kernel config build.
Additionally, when a defconfig is detected, allnoconfig is enabled
as the default operation, unless otherwise specified by KCONFIG_MODE.
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Richard Purdie [Thu, 1 Dec 2011 23:00:52 +0000 (23:00 +0000)]
Revert "classes/buildhistory: add new output history collection class"
This reverts commit 508ff624fea705eb93cf2cc1e0c9c42cb817acf8. RP
accidentally pulled the wrong commit in, it was supposed to be a
*package*history bbclass change.
Nitin A Kamble [Tue, 29 Nov 2011 19:30:39 +0000 (11:30 -0800)]
binutils: fix for .debug files QA warnings
THis commit fixes these QA warnings for binutils recipe
WARNING: For recipe binutils, the following files/directories were installed but not shipped in any package:
WARNING: /usr/arm-oe-linux-gnueabi/bin/.debug
WARNING: /usr/arm-oe-linux-gnueabi/bin/.debug/strip
WARNING: /usr/arm-oe-linux-gnueabi/bin/.debug/objcopy
WARNING: /usr/arm-oe-linux-gnueabi/bin/.debug/objdump
WARNING: /usr/arm-oe-linux-gnueabi/bin/.debug/ld
WARNING: /usr/arm-oe-linux-gnueabi/bin/.debug/nm
WARNING: /usr/arm-oe-linux-gnueabi/bin/.debug/as
WARNING: /usr/arm-oe-linux-gnueabi/bin/.debug/ranlib
WARNING: /usr/arm-oe-linux-gnueabi/bin/.debug/ld.bfd
WARNING: /usr/arm-oe-linux-gnueabi/bin/.debug/ar
Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Nitin A Kamble [Tue, 29 Nov 2011 19:30:38 +0000 (11:30 -0800)]
gmp: also generate the libgmpcxx library & package it properly
configure runs few checks to make sure c++ compiler and runtime are working
as expected with the --enable-cxx=detect option. And it enables building
of libgmpxx library.
Same as earlier the libgmp.so.10.x file is packaged in the libgmp10 package,
and a new package named libgmpxx4 is added for libgmpxx.so.4.x file.
Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Koen Kooi [Thu, 1 Dec 2011 16:04:49 +0000 (17:04 +0100)]
zlib 1.2.5: install pkgconfig file, needed for e.g. modern webkit
Upstream has grown cmake support which would allow us to dump the OE autotools hack, but the cmakefile doesn't install the .pc file either and breaks with zlib-native
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Paul Eggleton [Wed, 30 Nov 2011 16:48:47 +0000 (16:48 +0000)]
classes/buildhistory: add new output history collection class
Create a new build output history reporting class, using testlab.bbclass
from meta-oe and packagehistory.bbclass as a base. This records information
from packages and images output from the build process in text files
structured suitably for tracking within a git repository, thus enabling
monitoring of changes over time.
Build history collection can be enabled simply by adding the following
to your local.conf:
INHERIT += "buildhistory"
The output after a build can then be found in BUILDHISTORY_DIR (defaults to
TMPDIR/buildhistory). If you set up this directory as a git repository and
set BUILDHISTORY_COMMIT to "1" in local.conf, the build history data will
be committed on every build.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Paul Eggleton [Wed, 30 Nov 2011 16:39:15 +0000 (16:39 +0000)]
classes/base: split out metadata revision code to separate function
Split out a new function get_layers_branch_rev() which returns a list
of the metadata layers and their revisions. This enables reuse in other
places, such as buildhistory.bbclass.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Darren Hart [Thu, 1 Dec 2011 16:07:53 +0000 (16:07 +0000)]
trace-cmd: Add blktrace_api compatibility for TC_BARRIER
Newer kernels replace TC_BARRIER with TC_FLUSH. Ensure trace-cmd
can build regardless of the linux-kernel-headers version.
This is intended as a stop-gap to get the builds working again. A
proper fix will need to be discussed with the trace-cmd community.
RP: Tweaked the SRC_URI to remove unnecessary path and added PR bump Signed-off-by: Darren Hart <dvhart@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Thu, 1 Dec 2011 14:26:44 +0000 (14:26 +0000)]
puzzles: Ensure to link against libm for math functions
Without this we can see failures like:
| make: *** [samegame] Error 1
| /media/build1/poky/build/tmp/sysroots/x86_64-linux/usr/libexec/i586-poky-linux/gcc/i586-poky-linux/4.6.3/ld: inertia.o: undefined reference to symbol 'sqrt@@GLIBC_2.0'
| /media/build1/poky/build/tmp/sysroots/x86_64-linux/usr/libexec/i586-poky-linux/gcc/i586-poky-linux/4.6.3/ld: note: 'sqrt@@GLIBC_2.0' is defined in DSO /media/build1/poky/build/tmp/sysroots/qemux86/lib/libm.so.6 so try adding it to the linker command line
| /media/build1/poky/build/tmp/sysroots/qemux86/lib/libm.so.6: could not read symbols: Invalid operation
| collect2: ld returned 1 exit status
| make: *** [inertia] Error 1
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Khem Raj [Thu, 1 Dec 2011 03:07:07 +0000 (19:07 -0800)]
classes, conf: Exclude native and nativesdk recipes from libc specific override
Based on suggestions from RP
libc-uclibc and libc-glibc overrides are for denoting
system C library in use on the target and not for the
host therefore we make sure that the override only takes
effect for target recipes only.
Signed-off-by: Khem Raj <raj.khem@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Thu, 1 Dec 2011 12:42:27 +0000 (12:42 +0000)]
gnu-config: Only apply path transformations in the non-native/non-nativesdk case
The BUILD_ARCH != TARGET_ARCH check isn't a safe one to detect native builds
and doesn't cover the nativesdk case. This converts the recipe to use PN
instead which is more accurate and ensures the correct entries making it
into the correct packages.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Martin Jansa [Thu, 1 Dec 2011 12:02:09 +0000 (13:02 +0100)]
coreutils: fix u-a for base64, mktemp and df
* busybox installs mktemp and df to base_bindir not bindir
SHR root@gjama ~ $ ll /bin/df /bin/mktemp /bin/base64
lrwxrwxrwx 1 root root 7 Nov 10 15:44 /bin/df -> busybox
lrwxrwxrwx 1 root root 7 Nov 10 15:44 /bin/mktemp -> busybox
lrwxrwxrwx 1 root root 7 Nov 28 15:48 /bin/base64 -> busybox
* so u-a for coreutils fails:
update-alternatives: Error: cannot register alternative base64 to /usr/bin/base64 since it is already registered to /bin/base64
update-alternatives: Error: cannot register alternative mktemp to /usr/bin/mktemp since it is already registered to /bin/mktemp
update-alternatives: Error: cannot register alternative df to /usr/bin/df since it is already registered to /bin/df
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Martin Jansa [Thu, 1 Dec 2011 12:02:08 +0000 (13:02 +0100)]
shadow: use u-a for /usr/bin/groups
* otherwise coreutils upgrade fails with
update-alternatives: Error: not linking //usr/bin/groups to groups.coreutils since //usr/bin/groups exists and is not a link
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Darren Hart [Thu, 1 Dec 2011 01:58:52 +0000 (17:58 -0800)]
trace-cmd: Update to 1.2 (includes kernelshark)
Update the trace-cmd and kernelshark recipes to version 1.2. Rename the recipes
to the more descriptive trace-cmd_1.2.bb and kernelshark_1.2.bb respectively.
Update addldflags.patch for 1.2 and current Upstream-Status (Accepted).
Add make-docs-optional.patch to avoid building the new doc target which
requires asciidocs. We should add asciidocs and properly package the
docs, but for now it's more important to get the sources current.
Signed-off-by: Darren Hart <dvhart@linux.intel.com>
Dexuan Cui [Thu, 1 Dec 2011 09:21:25 +0000 (17:21 +0800)]
make: expand MAKEFLAGS before we re-exec after rebuilding makefiles.
This patch was got from the upstream cvs repo of make to fix the bug of
make-3.82: http://savannah.gnu.org/bugs/?30723
RP: Tweaked patch status to Backport Signed-off-by: Dexuan Cui <dexuan.cui@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Dexuan Cui [Thu, 1 Dec 2011 09:21:26 +0000 (17:21 +0800)]
task-self-hosted: install sudo, tun.ko, iptables, libgl and libgl-dev into the target
sudo, kernel-module-tun and iptables are needed by runqemu.
strace has appeared in RDEPENDS_task-self-hosted-debug, so let's remove it
from RDEPENDS_task-self-hosted-extended.
install libglu and libgl-dev rather than mesa-dri and mesa-dri-dev due to the
recent commit "mesa: package gl/egl/osmesa to separate packages".
Signed-off-by: Dexuan Cui <dexuan.cui@intel.com> Signed-off-by: Zhai Edwin <edwin.zhai@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Julian Pidancet [Thu, 1 Dec 2011 00:01:56 +0000 (00:01 +0000)]
Fix multiarch DISTRO_FEATURE
Make get_gcc_multiarch_setting more elegant. Use a dictionnary
to store the config options and replace bb.data.getVar with d.getVar.
Remove i686 from the architecture list because it doesn't seem
to be a valid TARGET_ARCH any more in OE.
Configure gdb (gdb and gdb-cross) with --enable-64-bit-bfd if
multiarch DISTRO_FEATURE is present
Signed-off-by: Julian Pidancet <julian.pidancet@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Wed, 30 Nov 2011 22:57:47 +0000 (22:57 +0000)]
consolekit: Fix ${localstatedir} race
The change adding ${localstatedir}/log to the package caused failures
on meta-toolchain-gmae builds since ${localstatedir} might be a
symlink and it can be installation order dependent.
By adding in the RDEPENDS, we defer to base-files to ensure
the layout is correct and that packages install correctly.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Darren Hart [Wed, 23 Nov 2011 23:17:16 +0000 (23:17 +0000)]
grub: Use COMPATIBLE_HOST
Drop the anonymous python block in favor of the much simpler
COMPATIBLE_HOST mechanism.
Signed-off-by: Darren Hart <dvhart@linux.intel.com> CC: Tom Zanussi <tom.zanussi@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Darren Hart [Wed, 30 Nov 2011 21:49:46 +0000 (13:49 -0800)]
grub: Drop "apply=yes" from patch
The "apply=yes" doesn't appear to be necessary, drop it in favor of
a simpler SRC_URI specification.
Signed-off-by: Darren Hart <dvhart@linux.intel.com> CC: Tom Zanussi <tom.zanussi@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Darren Hart [Wed, 30 Nov 2011 21:49:44 +0000 (13:49 -0800)]
autotools.bbclass: Report the missing configure path
If the configure script isn't found, report the explicit path tried.
This can help debug subtle errors where the ${S} sourcedir may not
be exactly what is expected.
Signed-off-by: Darren Hart <dvhart@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
