core-image-gtk-directfb: rename to core-image-directfb

- DirectFB is not supported by gtk versions >= 2.18
- use packagegroup-core-directfb instead of packagegroup-core-gtk-directfb

[YOCTO #3642]

(From OE-Core rev: f4724b132de09d0c3db3755214ca83bb5acf5550)

Signed-off-by: Constantin Musca <constantinx.musca@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

packagegroup-core-gtk-directfb: rename to packagegroup-core-directfb

- DirectFB is not supported by gtk versions >= 2.18
- remove the backwards compatibility statements (packagegroup-core-directfb
is an independent packagegroup)
- remove all gtk dependencies

[YOCTO #3642]

(From OE-Core rev: 423a397b2bf1a32519761ae3f0b048e0595c59e8)

Signed-off-by: Constantin Musca <constantinx.musca@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

init-live.sh: move /media/xxx over to the real root filesystem

Previously, the /media/xxx which contains the root image was not
moved to be under the real root filesystem. Because of this, the
output of the 'mount' command is somewhat confusing, that is, it
has some mount point that is not even in the filesystem.

Besides, on some machine, it caused the recovery procedure when we
booted it next time.

This patches fixes this issue by moving the /media/xxx over to be
under the real root filesystem.

[YOCTO #2064]
[YOCTO #3705]

(From OE-Core rev: 261d21d36298bb7822ee7370c30441f42ef2b093)

Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

prserv: change PRSERV_HOST semantics

- remove PRSERV_PORT variable
- use 'hostname:port' as PRSERV_HOST format
- remove USE_PR_SERV variable
- one can activate PRS by setting PRSERV_HOST

[YOCTO #3744]

(From OE-Core rev: e00f49de8b1f79c3e07b887d257bd75a46052fa0)

Signed-off-by: Constantin Musca <constantinx.musca@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

base.bblass : Add support for cmake to PACKAGECONFIG.

The cmake class uses EXTRA_OECMAKE to adjust the cmake configuration. This
patch adds support for this, if the cmake class is used by the recipe.

(From OE-Core rev: d8c0ce5a4a27c8aa1d07fc15d6e000af725a51e6)

Signed-off-by: Philip Balister <philip@balister.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

nfs-utils: Upgrade 1.2.3 -> 1.2.8-rc3

Disable nfsv4l since it needs LVM2 which
is not available in OE-Core

Disable nfsdcltrack since its configure time check
for sqlite3 is not cross compiling safe

It can support ipv6 but thats disabled since
we disable libtirpc which is needed for ipv6
support in nfs-utils

Patches imported from fedora to take us to 1.2.8-rc3

(From OE-Core rev: 658bfe4690005d0a68dc4e3ca2bc741ff4c89f28)

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gcc: Enable OpenMP compiling and library

Multicore embedded systems are getting more and more common.

Remove "--disable-openmp" from the GCC configuration options and
always build libgomp. This only creates a "bigger" compiler but
has no effect on the compiled binaries that don't use openmp.

Tested a clean build on mips32el and arm7a, no problems encountered.

Autoconf will not detect OpenMP after this change, because it will
build and run a target binary on the build system. In order to use
OpenMP, the variable ac_cv_prog_c_openmp=-fopenmp must be set.

(From OE-Core rev: fdf37b792c13a1fc3f63faf471f0403396414a35)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

perl-rprovides: Fix perl-module-config anf perl-module-build providers

The rearrangement of various file contents left these empty modules which
don't get packaged and hence lead to image creation errors on anything using
them. This patch ensures the dependencies are attached to the packages
which contain the config files or the build module infrastructure as
appropriate. It also resolves some confusion over whether the name is
perl-module-build or perl-module-module-build, the latter being the
autogenerated dependency which previously couldn't be found.

(From OE-Core rev: 442f161ecd498513b72e7f1ca2668325c9389d24)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

perl: Add dyanloader build hack

Patch perl to allow builds to work where a native perl running against target modules
attempts to load a dynamic binary module. We assume that a native version of the
module exists as it would for the target and perform an appropriate path
substitution.

(From OE-Core rev: b5ea12fec329fe419bc3672ed1e1d5733ff2a9d3)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

perl-native: Add errno_ver.diff patch to native perl

Apply the errno_ver.diff patch to the native perl version. This ensures that when
we're building target modules, running the native perl binary but in a target
configuration we don't see error messages.

libencode-detect-perl is a module which tests this.

(From OE-Core rev: 52354c863a1a2e467ea1b55168bf4273f31bbba4)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

cpan_build: Fix cmaker Build.PL module builds

Certain cpan Build.PL based modules were failing to build due to
the host's config information being used instead of the targets.

This patch sets the appropriate variables to ensure the modules
get the correct configuration from the target config_heavy.pl script.

It also ensures the linker is CCLD, not LD since the LDFLAGS contain
-Wl,XXX options and build failures result if LD is used.

libencode-detect-perl is a module which tests this class extensively.

(From OE-Core rev: 7fac8bffcd250aec90c0b675aeb2bbe7896d0ae1)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: bitbake-layers: fix get_file_layer

The get_file_layer can't handle the nested layer correctly, e.g.:

meta-intel/conf/layer.conf
meta-intel/meta-cedartrail/conf/layer.conf

They are two layers, the file:
meta-intel/common/recipes-bsp/amt/lms_7.1.20.bb
belongs to meta-intel's layer, but the get_file_layer() may return
meta-cedartrail accordig to BBLAYERS' order, since it uses:

for layerdir in self.bblayers:
    if regex.match(os.path.join(layerdir, 'test')):
        return self.get_layer_name(layerdir)

which doesn't care the filename, re-match the filename would fix the
problem.

[YOCTO #3387]

(Bitbake rev: 7a31b9eac4d894c302f0e052a132c31b17d13d1f)

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: cache.py: remove the duplicated self.file_depends

There are two "self.file_depends =" lines in cache.py::CoreRecipeInfo:

class CoreRecipeInfo(RecipeInfoCommon):
    __slots__ = ()

    cachefile = "bb_cache.dat"

    def __init__(self, filename, metadata):
        self.file_depends = metadata.getVar('__depends', False)
    [snip]
        self.file_depends = metadata.getVar('__depends', False)

They are duplicated, remove the last one.

[YOCTO #3795]

(Bitbake rev: 409a4aa0e166c25a8cfcd91e54f1b61cced74e87)

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: hob: hob do not use .hob dir for variables from local.conf and bblayers.conf

If .hob exists, hob takes the variable from there. Change to not override anymore
the variables which comes from bblayers.conf and local.conf

[YOCTO #2934]
(Bitbake rev: 00533e83038e3d5e3cd5d1be73c45afece92ec0c)

Signed-off-by: Cristiana Voicu <cristiana.voicu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: prserv: use only PRSERV_HOST

- remove PRSERV_PORT variable
- use 'hostname:port' as PRSERV_HOST format
- use 'localhost:0' for enabling the local PRServer

[YOCTO #3744]

(Bitbake rev: ad62eed9e1f0867a406e9fbfa44916a0f1ad1282)

Signed-off-by: Constantin Musca <constantinx.musca@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rpm: properly disable perl file dependency checking for rpmdeps

We don't want the perl module dependencies being added to each package
by rpmdeps because it causes too many problems when the modules aren't
available at install time (since they get added in do_package which is
too late for the build system to ensure they are available), and it
appears that the perl dependency gathering isn't working reliably in any
case judging by how they are detected on some build hosts and not
others.

This was being disabled previously but it seems like the macros moved
to a different file at some point and nobody noticed; thus make sure we
disable the macros in all RPM macro files.

Should fix [YOCTO #3699].

(From OE-Core rev: e43a72af9055b76eb2ad7e2e42e6a5f3c8854efa)

Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

systemd: Fix build on ppc

ppc discovers an interesting issue in linking systemd where a library is
missing in link cmdline and linker barfs

./.libs/libsystemd-core.a(libsystemd_core_la-manager.o): In function
`manager_check_finished':
/builddir/build/BUILD/systemd-196/src/core/manager.c:2092: undefined
reference to `sd_notifyf'
collect2: error: ld returned 1 exit status

For more info see

https://bugzilla.redhat.com/show_bug.cgi?id=888255

(From OE-Core rev: 2e2a91025ca9c113365c810ce08b48201fe792b8)

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libcheck: fix HOMEPAGE

This was pointing to the zypper homepage, presumably a copy-paste error.

(From OE-Core rev: 5f363483253af712f8ccb173c4cc9d77b7c1ae8c)

Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

classes/image-swab: remove zypper-related references

Remove references to libzypp, zypper and sat-solver which have been
removed.

(From OE-Core rev: cc9144f99844a8485b2b96182bdf40429286ae66)

Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

sat-solver: remove

This was only needed by libzypp, which has itself been removed.

(From OE-Core rev: 905f0d407a259f89e420ccdbee5a471cb0fea8f5)

Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libzypp: remove

This was only needed by zypper, which has itself been removed.

(From OE-Core rev: 5c8169bd1c114aab76001a38ee4dd68d00d825e0)

Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

zypper: remove

Functionality provided by Zypper is now covered more effectively by
Smart (python-smartpm) which is now being used for package management
on the target when enabled and RPM packaging is selected.

Fixes [YOCTO #3349].

(From OE-Core rev: cc866433341532a0b5cf8cd5f67998960ea96499)

Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

packagedata: Add error message if multiple recipes try to write the same package

If multiple recipes try and write the same package it resulted in a
rather confusing traceback and unintuitive error. This patch
prints a human readable error instead.

[YOCTO #3645]

(From OE-Core rev: 6f817779af77fdb0b861297f0f43c4c6607ce6f0)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

sanity.bbclass: Ensure tmpdir exists when running the check

This avoids tracebacks from bitbake if the directory doesn't already
exist.

[YOCTO #3640]

(From OE-Core rev: 1a60e96b21d328fe936e594f3061459e8f2f1b5e)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

core-image: Add hwcodecs IMAGE_FEATURE

Currently hardware codecs are being injected in rather ugly ways and end up
either in no or all images. This adds a dedicated IMAGE_FEATURE for them based
on the MACHINE_HWCODECS variable. We may need to refine this in due course
but this patch at least illustrates the concept.

(From OE-Core rev: 1cb370587fd72e3f0f69678748108cc4116767fb)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

packagegroup-core-x11-xserver: Add XSERVERCODECS variable

At present its hard for the system to provide extra x11 hardware codecs.
We could lump them in with the XSERVER variable but this is suboptimal as
in some use cases the user might not want to include them in some images.

With this patch we keep our options open about separating usage of the two
in the future.

(From OE-Core rev: 285c7577813cf75a3b554f8a5d6d7ba002127ef4)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

base: make feature backfilling happen earlier

Backfilling DISTRO_FEATURES and MACHINE_FEATURES with _append statements happens
too late to use those variables with conditional inherits, like this:

inherit ${@base_contains('DISTRO_FEATURES','sysvinit','update-rc.d_real','',d)}

Instead, do the backfilling at ConfigParse time so that it happens earlier in
the parse, which results in that inherit behaving as expected when sysvinit was
backfilled.

(From OE-Core rev: 22429cdf79ed952072707a929643c7386fa7e056)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake.conf: unbreak all builds with custom DISTRO_FEATURES

Commit 9e7c64ca9afbf27edd0d35a1830ce55ee6d778ab broke all builds where
custom DISTRO_FEATURES were used.

Resulting images ended in non-bootable state due to lack of initscripts
(unless someone had sysvinit or systemd in D_E already).

https://bugs.launchpad.net/linaro-oe/+bug/1102910
https://bugs.launchpad.net/linaro-oe/+bug/1099405/comments/12

(From OE-Core rev: 211e473432230765a48d7af9c66c8737a08cdec7)

Signed-off-by: Marcin Juszkiewicz <marcin.juszkiewicz@linaro.org>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

core-image-lsb-sdk/dev: Inherit the base image instead of duplicating

(From OE-Core rev: de2ad361e7dfe553e2c49c73199da58ce44d89e4)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

prserv: add LOCALCOUNT to AUTOINCs migration feature

- use migrate_localcount.bbclass to generate AUTOINC entries
which are exported to LOCALCOUNT_DUMPFILE
- import the generated AUTOINC entries
- one can migrate LOCALCOUNT to AUTOINC by executing:
    bitbake-prserv-tool migrate_localcount

[YOCTO #3071]

(From OE-Core rev: ffab86f13cafb10d8d6273b6af8cd9a3c84eae20)

Signed-off-by: Constantin Musca <constantinx.musca@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

qemu.inc: Define preferred providers for libgl, libgles1 and libgles2

This is needed to to boot efl images on qemuarm,qemumips and qemuppc
these options were already defined for qemux86 and qemux86-64 and
therefore the images were booting fine for these two machines

(From OE-Core rev: 60e73068cf542c2134106fe6cfc5971874bbc766)

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

systemd: Fix build on uclibc

Systemd has some uclibc specific patches which
needed to be forwarded ported to 196

(From OE-Core rev: e39f01ba4f6d9d9f1d24fd01745530cde9e92131)

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

uclibc-git: Upgrades to latest git

Drop upstream accepted patches

(From OE-Core rev: b4ee580d37d15e699dd6efc6551f4c7f8e902e2a)

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rootfs_rpm.bbclass: fix the unexpected postinst error

If the /etc/rpm-postinsts/ directory was empty, the following error
would occur at system startup.

      ERROR: postinst /etc/rpm-postinsts/* failed

This patch fixes this issue.

[YOCTO #3767]

(From OE-Core rev: 202263aeca837dc7b6615a3dc34569c199163733)

Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

blktrace: add back do_install method

The recent change to blktrace removed the do_install method,
since this recipe does not use autotools it needs to specifiy
a do_install activity.

(From OE-Core rev: f8832d969746fb879e63d1b28e6beef5ab82a39d)

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

populate_sdk_base.bbclass: Allow installation of ix86 SDK on x86_64 host

Commit c04f5435 "populate_sdk_base.bbclass: use SDK_ARCH instead
of SDKMACHINE" prevents not only the installation of 64 bit SDK
configurations on 32 bit hosts (which indeed cannot work), but also
the legitimate installation of a 32 bit SDK on a 64 bit host.

Fix this.

While there, also make sure we use the same patterns ("i[3-6]86" resp.
"x86[-_]64" to get unified strings for both INST_ARCH and SDK_ARCH.

(From OE-Core rev: 7ddd97b9b09fe7a327916ea88908a63375556ae6)

Signed-off-by: Wolfgang Denk <wd@denx.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: crumbs: Add file lost from previous commit

(Bitbake rev: 90aabeb87d6c1f00d6333cc9930626d36b6d9709)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

createrepo-native: update dependency

"RDEPENDS_${PN}_class-native" doesn't work for native recipe because no
-native package will be created. Then the dependencies listed here are
not built when bitbake createrepo-native.
Update and use DEPENDS_class-native instead.

(From OE-Core rev: 0b8dc781e4df0edffa10ac989f8745ac3f684de0)

Signed-off-by: Kang Kai <kai.kang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bind: disable nslookup man page

The nslookup is not shipped for now, we should disable its man page if
the binary tool is not installed.

[YOCTO#3754]

(From OE-Core rev: 1a31918733a7f8479220c01fc99dd9099fdcaf9b)

Signed-off-by: Wenzong Fan <wenzong.fan@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: hob: combo box updated correctly

The 'select a base image' option is no longer displayed once
an image is selected

[YOCTO #3671]
(Bitbake rev: 27f2247b4c227c3b5adb1ca33d0cd7f7b492e170)

Signed-off-by: Cristiana Voicu <cristiana.voicu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: persist_data: add get_by_pattern method to API

- one can use get_by_pattern to get a list of values associated
with keys that match the specified pattern

(Bitbake rev: 6ee1f58698e2d782c54ce5aec271bcec26107eac)

Signed-off-by: Constantin Musca <constantinx.musca@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: parse_py: add ~ to config_regexp

- enable the '~' character in bitbake variables

(Bitbake rev: 7c15ff1d50d7b601414f1d55c90e3c59981a0876)

Signed-off-by: Constantin Musca <constantinx.musca@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: hob: Hob should display warnings generated during parsing

-now Hob catches the warnings generated during parsing, and after the parsing
is completed, if there are any warnings, it shows a bar that contains a message
with how many warnings has encountered, and a button "View warnings"
-when "View warnings" button is clicked, Hob shows a dialog with the warnings;
if there more than 1 warning, you can use "Previous" and "Next" button to see them

[YOCTO #3215]
(Bitbake rev: d7b5311d35b3974398fecabfb5ecf1effa85c27e)

Signed-off-by: Cristiana Voicu <cristiana.voicu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: hob: progress bar changed to busy cursor when you open log file

-the first implementation for this bug used a progress bar, that is shown
during open file process; it revelead that the progress bar stops earlier
-now I have implemented using gtk.show_uri() method, that shows itself a
busy cursor when it opens a file;
-deleted the code for the first implementation

[YOCTO #2997]
(Bitbake rev: 09d1c4c2db124104b9da460547b20a2c2ff07bb3)

Signed-off-by: Cristiana Voicu <cristiana.voicu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

systemd: remove unused SERIAL_CONSOLE variable

(From OE-Core rev: 751c11e998431613389f094264f096fac5a1c1ae)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

systemd-serialgetty: skip package when not using systemd

As this package RDEPENDS on systemd it wants to build systemd, but if systemd
isn't a DISTRO_FEATURE then that package is skipped so world builds fail.  Solve
this by skipping this package too.

(From OE-Core rev: 192efd033c38205e114f62e9f62696a9a49b0970)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

systemd-serialgetty: remove redundant statements

(From OE-Core rev: e5820f56becb3bcee44a461f7fc5640159679a15)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

perl: Fix errors if configure is reattempted

If configure of perl is reattempted it currently fails as it tries to edit
files outside ${S}. Chaging from ${WORKDIR} to ${S} avoids this issue
and allows rebuilds to work.

(From OE-Core rev: 1b66c30eb6085aedce118ce086efbd2e562d0d6c)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

systemd: update uclibc patch headers

Discussion has revealed that upstream has formally rejected patches to support
uclibc, so mark this in the headers.

(From OE-Core rev: caf6aa1c996a949ca85f4beaf40d2a19433d18c8)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

packagegroup-sdk-gmae: remove GUPnP stack as it's no longer in oe-core

GUPnP is now in meta-multimedia, so remove it from this SDK packagegroup.

(From OE-Core rev: 356a5ee5a74f9cee112eac2c595c13ef707eca82)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gupnp: remove, migrated to meta-multimedia

The GUPnP stack was stagnating in oe-core and untested.  It's now moved to
meta-multimedia where it's also been upgraded.

(From OE-Core rev: 51e548a4e61709d9d87a424586529638873835b3)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

site: add x86-32/64 alignment values for at-spi2-core

This package isn't yet in oe-core, but GTK+ 3.6 depends on it so will be here at
some point.

(From OE-Core rev: 8c46ec2edc0197b32d32e0f27d5b60271338b600)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

systemd: add missing patch headers

The reverts were by Koen and are literal revert commits, the uclibc patches were
mostly introduced by Khem.

(From OE-Core rev: 8925501890cbe630258d9c6fe348f1e3c821fc71)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

systemd: use new GTKDOC_DOCDIR and clean up

Now that gtk-doc.bbclass supports overriding the documentation directory,
systemd can just inherit gtk-doc and all of the manual operations can be
removed.

(From OE-Core rev: 8b7816804dcfdc380fbebe6dc8c81f431a7fc450)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gtk-doc: parameterise the documentation directory

Not every package puts the gtk-doc infrastructure into ${S}, so allow this to be
overridden.

(From OE-Core rev: 40353c9ef83f986a0d49cc6eb654b3ace6ef4dc4)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

systemd: fix typo in FILES

Accidently managed to corrupt FILES_udev.

(From OE-Core rev: b1d8187f66d9646a53928fe288085f4605bc43ef)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

udev: skip in systemd distros, to fix world builds

(From OE-Core rev: 1179e43d67c96367480e563a36684d550d83fcbe)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

systemd: skip in non-systemd distros to fix world builds

(From OE-Core rev: a2553b5f8b80ca0cd578afa2d2857ebb2b70ee3c)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: BBHandler/ConfHandler: Merge fix for multiline comments

This was meant to be squashed into the previous commit for multiline comment
handling. It fixes the case the commented multiline is followed by an empty
line which was resulting in a traceback instead of a sane error message.

(Bitbake rev: 7e7d692e244fe8dca533f842ca143b9c821e317c)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: bitbake: BBHandler/ConfHandler: Improve multiline comment handling

Faced with an expression like:

 # Some comment \
 FOO = "bar"

what should bitbake do? Technically, the \ character means its multiline and
currently the code treats this as a continuation of the comment. This can
surprise some people and is not intuitive.

This patch makes bitbake simply error and asks the user to be clearer
about what they mean.

(Bitbake rev: 589d31ce41e019ee6a7cb6527d67bc76c0b6382a)

(Bitbake rev: 79c00fabe08b4c210a3bd81cfaffbc47ffdc2e2b)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

dbus: upgrade to 1.6.8

dbus, dbus-ptest upgraded to v1.6.8

(From OE-Core rev: 272661a7fca7e0ee8f6b0d7a4ba97f09a82a540e)

Signed-off-by: Cristian Iorga <cristian.iorga@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

boost: Upgrade 1.51 -> 1.52

Drop hash_enums.patch which is already applied upstream

(From OE-Core rev: c4d3f5596662d6aced46f270f566b510a36e4952)

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gtk+3: refresh cross.patch

cross.patch was accepted upstream, so update this patch with the commit that was
merged.

(From OE-Core rev: 8e5ffa370c4259f925a0454fe7fd81c80ef2174b)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bb-matrix: Fix min and max calculations

The BB and PM ranges were originally intended to use leading 0s to
ensure all the values were the same string length, making for nice log
filenames and columnar dat files. However, not everyone will do this -
especially if it isn't documented.

Document the intent. Make the generation and parsing of dat files robust
to either method.

(From OE-Core rev: 90dc44c8246f2a580fe4a41ce28c201e52307500)

Signed-off-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

oprofile: backport patches to fix ppc build issues

Fixes:

| operf_utils.cpp: In function 'bool _op_get_event_codes(std::vector*)':
| operf_utils.cpp:151:21: error: 'pfm_initialize' was not declared in this scope
| operf_utils.cpp:151:26: error: 'PFM_SUCCESS' was not declared in this scope
| operf_utils.cpp:166:45: error: 'PFM_PLM3' was not declared in this scope
| operf_utils.cpp:166:55: error: 'PFM_OS_NONE' was not declared in this scope
| operf_utils.cpp:166:72: error: 'pfm_get_os_event_encoding' was not declared in this scope
| operf_utils.cpp:167:14: error: 'PFM_SUCCESS' was not declared in this scope

[YOCTO #3717]

(From OE-Core rev: 121cb96964fe2f374d814bf39036119bd63b9589)

Signed-off-by: Matthew McClintock <msm@freescale.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

package.bbclass: drop unnecessary and broken packages.remove()

(From OE-Core rev: 5480371d0473482437ce4350e6593f006101ef40)

Signed-off-by: Christopher Larson <chris_larson@mentor.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

oprofile: set correct kernel path

oprofile's "configure" was always looking at the configuration of the host
kernel, not the target kernel. This patch instructs oprofile to use the
target kernel to configure itself.

(From OE-Core rev: 5892db9a34c188b8a9bdaa86e19ba62a7f4793b5)

Signed-off-by: Bogdan Marinescu <bogdan.a.marinescu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rpm2cpio: Update to match rpm4 specs for decompression

[YOCTO #3753]

(From OE-Core rev: 1377278187d20853304fe2689105f2848d796d86)

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

update-alternatives: Add a build-time dependency

We need to add a build time dependency on virtual/update-alternatives,
however we can't just do DEPENDS +=, or we end up with various problems.  To
work around this, in the anonymous python space we ensure we only do the
addition when the package does not provide virtual/update-alternatives and
it is a target package.

Also the system wide PREFERRED_PROVIDER was incorrect.  It references a
runtime package, and not the recipe it should have.  This has been corrected.

[YOCTO #3691]

(From OE-Core rev: 56a59ef12936dcc6464cf1d43dda6957a5aa8c65)

Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Tested-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

populate_sdk_rpm: Re-add a few system provides to the SDK

When building the target SDK portion, we can safely ignore various
package rdepends.  These dependencies are not required on a build
only environment like an SDK.

[YOCTO #3691]

(From OE-Core rev: 1edd44bed9e9829f7c2466928bc3f45be102b84b)

Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Tested-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

scripts/lib/bsp/engine.py: add handling for JSON strings

Normally pre-canned properties are supplied as JSON from a file, which
the user can specify using e.g. the -i option.

We can reuse that basic functionality for dedicated command-line
parameters by sythesizing a JSON string containing those param values
on the fly and passing that in instead.

This adds the ability for the common creation code to accept JSON
strings as well as JSON files.

(From meta-yocto rev: 5a2e840b24822e018de94ec4f554363b59c4e8bd)

Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

yocto-layer: add optional layer priority param

If the user specifies a layer priority following the layer name, layer
creation will proceed without further queries using the specified
layer priority and the remaining values defaulted.

(From meta-yocto rev: 84a0bd8940f82fb938972d7b026367d40c9472e7)

Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

scripts/lib/bsp/engine.py: refactor bsp-creation code

This does a bit of refactoring of the bsp-generation code to make it
generically reusable for generating non-bsp layers.

The first user remains the existing yocto-bsp tool; these changes
allow a second user, the new yocto-layer tool, to use the same code.

(From meta-yocto rev: 1527a0ee7bce08a527c9d80516531b17816dff17)

Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

yocto-layer: add 'layer' template data

Add a 'layer' target containing all the data that will be used to
generate a generic yocto layer.

(From meta-yocto rev: 198a85f61ebd6435830285b2a9b1b925aea6779e)

Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

yocto-layer: add help/usage

This is essentially 'the documentation' for the yocto-layer tool.

(From meta-yocto rev: 34229b931bad8fc0e4d4431bb5cb46fccbea03bf)

Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

yocto-layer: new script

Implementation of the 'yocto-layer' command-line tool, for creating
generic layers and listing their input properties.

(From meta-yocto rev: 8170bea55379d1a25acc0fea108675526eeab6f6)

Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

scripts/lib/bsp/engine.py: add yocto_layer_create()

Add a new yocto_layer_create() function that will be used to generate
a generic yocto layer (for the new 'yocto-layer' command).

(From meta-yocto rev: 44acd01bf47c2e0a777e686c9339a6ff951fc972)

Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

attr: convert to generic syscall numbers

xattr system calls are old enough that we can switch
to just pulling the system call numbers from kernel
headers.

Upstream-Status: Submitted acl-devel@nongnu.org

(From OE-Core rev: 4dab91a4109a452f43483ef11ea9037ca8df1b33)

Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

nspr: add aarch64 support

Add needed defines for Aarch64 support.

Upstream-Status: Submitted https://bugzilla.mozilla.org/show_bug.cgi?id=827271

(From OE-Core rev: 53b4f10bf7e8e83275a64da6700a52c5a4ba533d)

Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

cpan_build: Fix bashism

Fix a build failure caused by a bashism and improve quoting whilst in
the area.

(From OE-Core rev: 9eb78340c61acddb0c2012e25c76c18735febb4d)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libtool-cross: Add missing libltdl components to install

Without this some software fails to build complaining it can't find these
utilities.

(From OE-Core rev: db84eaf851b22b262d9dc48eb55bd5224a00fdd2)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

liburi-perl: Add BBCLASSEXTEND for native version

(From OE-Core rev: cac29d930910c257293ed2165faa5397f0a5d07c)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

package.bbclass: Fix shlibs cross package-arch contamination

I found that if I had two packages architectures built (say core2 and i586),
that the shlibs data from i586 was being used in the core2 build. This
lead to odd dependency issues if the i586 build was out of date and core2
was being targetted.

When loading shlibs files, the last loaded wins so we need to iterate the
lists from least to strongest weighting, not the other way around. With this
patch applied, the contamination issue is resolved.

(From OE-Core rev: acd7b1c2ee9f3ff84c45a723dba5e0bc5f6a100c)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

dropbear: Deal with truncated host keys by removing them

Dropbear does not start when the host key is empty and it is possible
that a device is switched off before the host key is generated. This
is possible because the dropbearkey code doesn't create a temporary
file first. Detect truncated keys and then remove them which will lead
to the re-generation. This way the dropbear process will always start.

(From OE-Core rev: 16b57e352f5844f301cc6c7ea4f87bf750c11d67)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

populate_sdk_base.bbclass: Work around bugs with gnu tar < 1.24

The gnu tar versions less than 1.24 pose a problem with the SDK
extraction.  A typical SDK will have hard links.  The SDK is created
with a newer version of tar but extracted with the host version of tar
which may be much older.  The specific defect is that older version of
gnu tar do not properly handle the "--strip-components=" argument when
processing hard links and the paths do not get properly truncated and
hooked up with the corresponding real file payloads.  This leads to
errors like the following during the SDK install:

tar: sysroots/qemux86-yocto-linux/usr/bin/getconf: Cannot hard link to `./opt/yocto/SDK-qemux86/sysroots/qemux86-yocto-linux/usr/bin/POSIX_V6_ILP32_OFFBIG': No such file or directory

The simple way to work around the problem is to just not save the path
information that is going to get chopped off anyway.  This has the
added benefit that it saves a small amount of space in the tar archive
as well.

(From OE-Core rev: 0f4112016bac07f0294f2a4d0fde4659c0fed2d2)

Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

package.bbclass: replace AUTOINC directly in PKGV

b27b438221e16ac3df6ac66d761b77e3bd43db67 introduced a regression when
PKGV is modified by the recipe.  E.g. when having something like

| PV = "1.7.33+gitr${SRCPV}"
| PKGV = "1.7.33+gitr${GITPKGV}"
| RDEPENDS_${PN}-subpkg += "${PN} (= ${EXTENDPKGV})"
| inherit gitpkgv

The resulting core package has

| Version: 1.7.33+git234+cd01e8c-r11

and because the subpackage has

| Depends: vdr (= 1.7.33+git0+cd01e8c18d494150730214f033f84e5282ce7639-r11)

image build fails with

|  * satisfy_dependencies_for: Cannot satisfy the following dependencies ...
|  *    vdr (= 1.7.33+git0+cd01e8c18d494150730214f033f84e5282ce7639-r11)

Patch avoids the 'PKGV = "cleanup(${PV})"' operation and replaces
AUTOINC directly within PKGV.  This might introduce a regression too
when 'AUTOINC' is added to PKGV intentionally but impact is small and
can be fixed trivially by using another string.

(From OE-Core rev: 6e46664085fa49bc204fc944ca23ed6383e2ec51)

Signed-off-by: Enrico Scholz <enrico.scholz@informatik.tu-chemnitz.de>
Cc: Constantin Musca <constantinx.musca@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

linux-yocto/3.4: update intel graphics features/configs

Updating the meta branch to import the following commits that cleanup and
reorganize the features that make up the graphics features of the Intel
BSPs

  c0b3904 meta: add config fragment for gma600 graphics driver
  f41638f meta: add config fragment for gma500 graphics driver
  ff8ad3e meta: sys940x: drop vesa fragment from bsp config
  fcb82ac meta: sys940x: remove emgd config from bsp config
  c1677bb meta: fri2: drop vesa fragment from bsp config
  7324440 meta: fri2: remove emgd config from bsp config
  dd435c6 meta: emenlow: drop vesa fragment from bsp config
  243b515 meta: emenlow: remove emgd config from bsp config
  3e1be82 meta: crownbay: drop vesa fragment from bsp config
  fc5ed4e meta: crownbay: remove emgd config from bsp config
  d46236b meta: drm-emgd.cfg: add configs for dependent routines
  db4bc5e meta: relocate git-merge of emgd branch

(From OE-Core rev: 8d0074ef2b61bf022ed1bb3b5f1afe526d8c921b)

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

linux-yocto/3.4: update to 3.4.26 and 3.4.25-rt37

Updating the yocto 3.4 kernel tree to the 3.4.25 and 3.4.26 -stable updates
as well as refreshing the preempt-rt branches to 3.4.25-rt37.

(From OE-Core rev: da536703bdcf31c8c2ef17f5a49d06280fe77600)

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

consolekit_0.4.5.bb: Package unpackaged systemd files

(From OE-Core rev: 8ac35f3bba5d00549f5255acf17a8be914402d50)

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

kernel.bbclass, module-base.bbclass: Use CC to form KERNEL_CC

kernel compiler is not special and we currently have it so
we want to pass -march and -mtune options as CFLAGS to kernel
build so that compiler picks the right subarch flags when
compiling assembly files in particular. Otherwise defaults
are chosen which may not be right in many case e.g. when
compiling kernel for collie machine we should use arch=armv4
but it uses toolchain/as defaults which is armv5te

in some case e.g. thumb1 we know that kernel can not be compiled
in thumb1 mode so we can provide that information e.g. -marm
option through KERNEL_HOST_CC_ARCH variable as we do now

(From OE-Core rev: c49f967384ccbfe131bbb33ee518014f3fc4b38f)

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libpam: register PAM session with logind

This make screen/tmux/etc work as intended.

(From OE-Core rev: 58731bbdbd4ab4cfd560f14758a65efdfad2e28f)

Signed-off-by: Martin Donnelly <martin.donnelly@ge.com>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

packagegroup-core-boot: only install initscripts if we're using sysvinit

(From OE-Core rev: 9e7c64ca9afbf27edd0d35a1830ce55ee6d778ab)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

base-files: add fstab for systemd based systems

The default fstab has entries which are not necessary
in systemd based systems so add a simpler default in
this case.

(From OE-Core rev: 77bbb839ba25b974a538b90d346b454ccd5deefd)

Signed-off-by: Martin Donnelly <martin.donnelly@ge.com>
Signed-off-by: Radu Moisan <radu.moisan@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

update-rc.d: disable update-rc.d.bbclass when systemd enabled

update-rc.d is not necessary when systemd is enabled,
systemctl is the replacement

(From OE-Core rev: eaa82c842d3197e0251a887ba16028c22f9d8d94)

Signed-off-by: Radu Moisan <radu.moisan@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

packagegroup-core-boot: install systemd-compat-units on systemd images

systemd-compat-units is needed to run postinsts for example.

(From OE-Core rev: cf783377767c70db544da57593616cfc952dbfb9)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

default-providers: Add systemd option to PREFERRED_PROVIDER_udev

(From OE-Core rev: fb8e5285bbc6d5aac5d4103c0a983b7b6c54d73d)

Signed-off-by: Radu Moisan <radu.moisan@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

systemd: add systemd recipes

Add systemd recipes and associated support recipes.

Mostly based on meta-oe/meta-systemd, so almost all credit should go to:

Andreas Müller <schnitzeltony@googlemail.com>
Denis 'GNUtoo' Carikli <GNUtoo@no-log.org>
Holger Hans Peter Freyther <holger@moiji-mobile.com>
Khem Raj <raj.khem@gmail.com>
Koen Kooi <koen@dominion.thruhere.net>
Martin Jansa <Martin.Jansa@gmail.com>

(From OE-Core rev: 6a8a48b4d0d0b9b8d8af46cae11245bcb870bbc3)

Signed-off-by: Radu Moisan <radu.moisan@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

dbus: respect systemd distro feature

Only configure for systemd if we are using systemd, and remove the dbus-systemd
package by putting the service files into PN itself.

(From OE-Core rev: 06e262c9b4c406b886db6ca8eee55ab441599151)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

default-providers: Automatically set PREFERRED_PROVIDER_udev

This is a more generic way to set preferred provider for udev.
We expect to have multiple choices once we integrate other init
managers, and this way we can automatically set it considering
distro settings.

(From OE-Core rev: da13562af73d144b5782ee8d755e2249cd9d2d8c)

Signed-off-by: Radu Moisan <radu.moisan@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

default-distrovars: Add DISTRO_FEATURES_INITMAN to DISTRO_FEATURES

This is a more generic way to set the init manager since we
plan to support other init managers as well.
I will use this variable as a switch to turn on/off any
init scheme that we might support in the future.
By default we use sysvinit.

(From OE-Core rev: 87f06346728bda000c0c0f95312b6a0a1b149ab4)

Signed-off-by: Radu Moisan <radu.moisan@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>