mesa: remove unrecognized configure options

--disable-glut and --disable-glw are not recognized since Mesa 8.0.
--disable-glu is not recognized since Mesa 9.0.

(From OE-Core rev: dac1f95e03f897237e6551240c200370d5102807)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

glib-2.0: drop Makefile.in hunks from patch

Patching a generated file is prone to conflicts, and we'll be replacing this
file anyway when configuring, so drop the Makefile.in hunks.

(From OE-Core rev: a3fa323c4d24375be9b783281c3cf31c0707ecb7)

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

handle two-word commands with tmux

Trying to make a devshell using tmux can fail because "tmux new"
expects a single command, not a series of arguments. It does, however,
split strings in a suitable way. So you can quote the command.

The failure mode is particularly arcane, in that you end up
with a message like:

ERROR: Unable to spawn terminal auto: \
Execution of 'pseudo /bin/bash' failed with exit code 1:
usage: new-session [-d] [-n window-name] [-s session-name] \
[-t target-session] [command]

which is confusing because there's no "new-session" anywhere in
sight (that's actually "tmux new"), and because what failed to execute
wasn't either pseudo or bash.

(From OE-Core rev: f8ed7446755eeb88191e16749350efa1e7e6197c)

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

librsvg: configure needs a native gdk-pixbuf binary

Build failure was seen on the autobuilder, so add gdk-pixbuf-native to depends list

(From OE-Core rev: 6ea293a8696843779cb23d6176b54efcde0462ec)

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

rpm: replace rpm-postinsts dependency with run-postinsts

run-postinsts is generic and will handle rpm delayed postinsts just
fine. No need to use another recipe.

(From OE-Core rev: 1195a188f366fbbb04b9379d7ea3c7c02130cc98)

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

package_rpm.bbclass: handle pre/post remove scriptlets

Currently the scriptlet_wrapper is designed to handle only pre/post
install scriptlets. This patch will slightly change the wrapper script
to handle also pre/post remove scriptlets when we want to remove
packages from the final image.

(From OE-Core rev: aea47c77d69407b2e62f151cabba35293d179f0c)

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

rootfs_rpm.bbclass: switch to using run-postinsts

Since run-postinsts was made generic, use it for RPM too.

(From OE-Core rev: 1ba44302f522a42535c9ad7d2fcccc06ba4c20b3)

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

run-postinsts: make it generic

This recipe is used only when the package manager is not deployed on
target and there still are delayed postinstalls that need to be run on
target. At the moment, run-postinsts is targeted to dpkg/opkg. RPM uses another
recipe, called rpm-postinsts which, in turn, is used even when the rpm
package is deployed.

This patch intends to make run-postinsts generic and deprecate
rpm-postinsts. Here's why:
 * when opkg/dpkg are not deplyed, the meta-data files are useless.
   Hence, the awk script is not needed. The postinstall files are already
   saved in /etc/(deb|ipk)-postinsts directory and we can use the same
   procedure as in rpm's case;

Also, the recipe is moved out from meta/recipes-devel/dpkg to
meta/recipes-devel/run-postinsts.

(From OE-Core rev: 0222e6740ee4670bbf42d246ab3ecad415ffcdee)

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

core-image-minimal: do not remove packaging files in the rootfs postprocess

Since the removal of PM's meta-data was moved in image.bbclass, after
rootfs_*_do_rootfs() runs, there's no need to remove the files in rootfs
postprocess stage. Also, if the meta-data is removed then there's no way
we can save the postinstalls to /etc/(deb|ipk)-postinsts later.

For RPM, it doesn't really matter since the postinstalls are saved to
/etc/rpm-postinsts as soon as the postinstall fails.

[YOCTO #4484]

(From OE-Core rev: aac8a59216819759d6b97bc30a7c717dbee7060a)

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

image.bbclass: remove unneeded files from the image

This patch creates a new function, rootfs_remove_unneeded(), that will
be called after rootfs_*_do_rootfs is called and which will handle
delayed postinstalls in a generic fashion, for all backends in the same
way.

(From OE-Core rev: d3d810faf8e93ce2d3faaa70da6deeb25b1cec26)

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

rootfs_*.bbclass: add some helper functions

This patch adds the following helper functions:
 * delayed_postinsts - outputs the list of delayed postinstalls;
 * save_postinsts - this will save the delayed postinstalls for ipk/deb
   in /etc/(ipk|deb)_postinsts;
 * rootfs_remove_packages - removes packages from an image;

Additionaly, this patch will remove a piece of code in
rootfs_ipk_do_rootfs which will be moved to image.bbclass and used for
all backends;

[YOCTO #4484]

(From OE-Core rev: 928df79cd3964f775c4c6e4283ef84b8882f9328)

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

qt4: fix QMAKE_QT_CONFIG being overwritten with empty value

QMAKE_QT_CONFIG is set by the linux-oe-g++ mkspec but it is overwritten
with an empty value read from .qmake.cache. Avoid this by first checking
if the value from .qmake.cache is not empty before assigning it to
QMAKE_QT_CONFIG.

This allows variables from qconfig.pri such as QT_ARCH, QT_CONFIG,
QT_VERSION, etc. to be queried by qmake projects.

(From OE-Core rev: 5330905749828ea8befdee559c180754ac506b75)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

qt4-native: update configure paths to match target

qmake2 was being built with paths different to the target Qt.

The incorrect paths can be seen by running qmake2 -query:
QT_INSTALL_DOCS:${STAGING_DIR_NATIVE}/usr/doc
  should be /usr/share/doc/qt4
QT_INSTALL_PLUGINS:${STAGING_DIR_NATIVE}/usr/plugins
  should be /usr/lib/qt4/plugins
QT_INSTALL_IMPORTS:${STAGING_DIR_NATIVE}/usr/imports
  should be /usr/lib/qt4/imports
QT_INSTALL_TRANSLATIONS:${STAGING_DIR_NATIVE}/usr/translations
  should be /usr/share/qt4/translations
QT_INSTALL_EXAMPLES:${STAGING_DIR_NATIVE}/usr/examples
  should be /usr/bin/qt4/examples
QT_INSTALL_DEMOS:${STAGING_DIR_NATIVE}/usr/demos
  should be /usr/bin/qt4/demos
QMAKE_MKSPECS:${STAGING_DIR_NATIVE}/usr/mkspecs
  should be /usr/share/qt4/mkspecs

Use the same paths as target Qt for consistency. As the mkspecs are
installed to /usr/share/qt4/mkspecs not /usr/mkspecs, qmake2 was
unable to detect and load the webkit module (it is detected by loading
modules/qt_*.pri inside the mkspecs directory) so webkit was missing
from QT_CONFIG if a qmake project reads the QT_CONFIG variable.

(From OE-Core rev: 37b57b44ff83c718021da4f717fc66a28204c96c)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

classes/qmake_base: update .pro hack

This adds support for lconvert and xmlpatterns tools and tweaks the
sed expressions to handle more complex assignments in .pro files like:
LRELEASE = $$targetPath($$[QT_INSTALL_BINS]/lrelease)

(From OE-Core rev: 542fefa61f0ac1dcdacac2630fd44950ac11cdd9)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

classes/qmake2: export OE_QMAKE_LCONVERT and OE_QMAKE_XMLPATTERNS

These will be used by qmake_base.bbclass.

(From OE-Core rev: 405d1325ebbd31676d88a3e157f55d40d04cc3f4)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

qt4-native: build lconvert and xmlpatterns

These tools are needed by some Qt projects (e.g. Qt Creator).

(From OE-Core rev: e2f3232fdc2836685549e59dc5c71363d3bbab5f)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libxml2: added python dependency to nativesdk

This is necessary since libxml2 has python dependency.

This patch will fix this error:
...
/path/to/build/system/4.7.2/ld: cannot find -lpython2.7
...
ERROR: Task 4152 (virtual:nativesdk:meta/recipes-core/libxml/libxml2_2.9.0.bb, do_compile) failed with exit code '1'

(From OE-Core rev: 4f2b3e3831bdc5707eacdab571ab207d8b09953e)

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

classes/buildhistory: split SDK data by IMAGE_NAME

Data from SDKs produced from different images or different
meta-toolchain style recipes should go into separate directories,
since SDK_NAME doesn't include anything to differentiate these by
default (only pieces for the distro name and architectures).

Note you should delete or move existing data under the buildhistory/sdk/
directory manually after this change.

Fixes [YOCTO #4622].

(From OE-Core rev: 74658b66530184b0287152397341fbe411e90c71)

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

sudo : upgrade to 1.8.6p8

upgrade from 1.8.6p7 -> 1.8.6p8

 - removed crypt.patch because it was contained upstream

(From OE-Core rev: 198e0db0e840dd3ac719d0c2ea980e08bf1f3442)

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

openssh : upgrade to 6.2p2

upgrade from 6.2p1 -> 6.2p2

(From OE-Core rev: 752ad432a248befe7d7b1dd799209f7bcbe9231e)

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

cracklib: fix QA warning about unsafe references in binaries

If enable test 'unsafe-references-in-binaries', build libpam will get
following warnings:

WARNING: QA Issue: libpam: /lib/security/pam_cracklib.so, installed in \
the base_prefix, requires a shared library under exec_prefix (/usr): \
libcrack.so.2 => /usr/lib/libcrack.so.2

(From OE-Core rev: 7ffe63c4a4f6b3adfea97eed4c0452c287784663)

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

qt4-x11-free: update HOMEPAGE

(From OE-Core rev: cf5c9ee6128cbec3bf7d74210bd968fba9ac87da)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

qt4-native: update HOMEPAGE

(From OE-Core rev: b56ee1a32f047da288bc7b2435470609b70319fe)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

qt4-embedded: update HOMEPAGE

(From OE-Core rev: 246d16f453466ab609861e170d0b51d7dd23e91a)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

qt-mobility: update HOMEPAGE

(From OE-Core rev: 74fc4c0d17f72f5b59b8da6e03d3a50e8c66d76d)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

nativesdk-qt4-tools.inc: update HOMEPAGE

(From OE-Core rev: dbd5c1fda3b477e861a979e98a4f879071bebc37)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

trace-cmd: backport three patches

Backport three patches to fix running issues.

(From OE-Core rev: c00f67b5cd84adda7566505ec9372e11ad812ca1)

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

PPP: add default configuration file and fix hard-coded CFLAGS

1. Add three ppp configuration files, make "/etc/init.d/ppp start"
be able to work with little change. provider and pap are adopted from
Ubuntu. ppp_on_boot is adopted from http://www.wlug.org.nz.
2. Use build system CFLAGS when compiling

(From OE-Core rev: 6ab8adf980f90040d3a24dd9c0f12dc04d7aadeb)

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

chkconfig: fix QA warning about unsafe references in binaries

This test is disabled in oe-core by default, but if we enable it, we'll
get following warnings:

  + WARN_QA = "unsafe-references-in-binaries unsafe-references-in-scripts"

WARNING: QA Issue: chkconfig: /sbin/chkconfig, installed in the \
base_prefix, requires a shared library under exec_prefix (/usr): \
libpopt.so.0 => /usr/lib64/libpopt.so.0

(From OE-Core rev: 18f5e6b80e1465e00242a2513a4bd90e0a2f3ff7)

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

socat: enable termios

fix two cross-compile issue to enable termios:
1. Define ac_cv_ispeed_offset in EXTRA_OECONF to avoid to call AC_TRY_RUN()
on cross-compile environment.
2. fix autoheader to generate *_SHIFT macros

(From OE-Core rev: fa4d1526f639d50b578ff82b4e65ac45c612bcca)

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

ddimage: fix incompatibility with dash

On systems with dash as /bin/sh there were failures while invoking ddimage.
Fix to let it work with both bash and dash shells.

[YOCTO #4617]

(From OE-Core rev: 4c6f7a5d8bd6ada434b91037ecd5db06f3eac814)

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

init-live.sh: fix media automount failed after booting from usb-drive

1, This issue happens to BSP only.
After a BSP board is booted with Yocto linux from USB drive, "cat /proc/mounts" shows:
...
/dev/sda3 /media/sda3 ext3 rw,relatime,errors=continue,user_xattr,acl,barrier=1,data=ordered 0 0
/dev/sda1 /media/sda1 ext3 rw,relatime,errors=continue,barrier=1,data=ordered 0 0
/dev/sda2 /media/sda2 ext3 rw,relatime,errors=continue,user_xattr,acl,barrier=1,data=ordered 0 0
...

but actually the directory /media/sda1 doesn't exist at all, "df" shows:
...
df: /media/sda3: No such file or directory
df: /media/sda1: No such file or directory
df: /media/sda2: No such file or directory
...

2, This is because the mount data comes from proc setup during early boot
before the change root, which then uses a different root filesystem, the
media is not in the new root filesystem.

3, During early boot before switch_root, use `mount --move' to move all
medias to the new root filesystem could also fix this issue.

[YOCTO #2064]
[YOCTO #3705]

(From OE-Core rev: 79bd773cc5e8b8e873cabeb2b9a91f460501dad7)

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

gnomebase.bbclass: add support for archive name different from BPN

Use new variable GNOMEBN, defaulting to value of BPN, instead of
BPN directly when determining URL for source archive.

(From OE-Core rev: fe297a70c7d6b36a6595534589c71d32209d6a53)

Signed-off-by: Marko Lindqvist <cazfi74@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gnomebase.bbclass: implement gnome_verdir() using trim_version()

(From OE-Core rev: 2341c0b801b62d0249567806a70d9a38f4dfe17f)

Signed-off-by: Marko Lindqvist <cazfi74@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gstreamer: added GStreamer 1.0 recipes

Originally from the https://github.com/dv1/meta-gstreamer1.0 repository,
the patches were copied with only one change:
gstreamer1.0-plugins-package.inc was removed, since gst-plugins-package.inc
is identical except for the LIBV variable.

(From OE-Core rev: e200e61529fa5a78354cde67442e90ace4b3857c)

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

tar: don't mv tar for nativesdk

We move tar into /bin for target but it can stay in /usr/bin for nativesdk
this will also mean the paths are correct for the buildtools-tarball environment

(From OE-Core rev: b4f208f418d18f2a4e78a56bebacef481061d917)

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

buildtools-tarball: Add json, unittest and compression python module packages

We need unittest so bitbake-selftest can work, tarfile from the compression module
for opkg-utils and json for webkit builds.

(From OE-Core rev: bbcb7329ca4340dbd3d529567dc6b605347e3368)

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

bitbake: bitbake-layers: fix for move of data object to cooker

(Bitbake rev: 063ee01935767ac4c9518d7477c8c85da1c9969a)

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

bitbake: tinfoil: fix for move of data attribute to cooker

(Bitbake rev: c400fe36f7609d53fb413484dc03bbce307f31f9)

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

base/useradd/sstate: Ensure do_package setscene has correct fakeroot dependencies

The do_package_setscene task needs to depend on fakeroot in order to correctly
install its files.

We can whitelist the dependency in the sstate handling code for some
performance improvements since we only need this if we're installing the
package from sstate.

Also use an append operator in base.bbclass for clarity.

(From OE-Core rev: 0810ea2a72bdea67a3d8002c4e12fb20f45cf1d5)

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

bitbake: lib/bb/cache.py: invalidate cache when file checksum entry no longer exists

Go through the cached list of file checksums and check if any of the
files no longer exist; if any are missing then invalidate the cached
recipe, which will force it to be reparsed and thus force the list of
files to be collected again. This prevents a warning when moving a file
to a different location that is still picked up by the recipe, e.g.
moving a file from a "files" subdirectory to one named with the recipe
name (${BPN}).

Fixes [YOCTO #4474].

(Bitbake rev: e1aa103593c46612c4d9000524174983b5994dc9)

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

bitbake: cooker: Split data from configuration

The reasons for combining these objects is ancient history, it makes
sense to clean things up and separate them out now. This follows on
logically from the configuration cleansups and leads well into the
bitbake-worker changes.

(Bitbake rev: 89ffd62661ebcf2a97ce0c8dfd5e4d5bfbe27de7)

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

python-external-tarball: Improve DESCRIPTION/SUMMARY

Improve the DESCRIPTION for the recipe and add a SUMMMARY.

(From OE-Core rev: ecdccd0c19c1d27aee43a3cca8e4467f270d7f9c)

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

package_*.bbclass: Drop fakeroot from setscene calls

Back in 2010, I added these in commit: http://git.yoctoproject.org/cgit.cgi/poky/commit/?id=36f1ae42fe13dae174b7fb5eb85dc49d7d7b516b
however the package_write tasks sstate only consists of package files and no
fakeroot privileges are needed to write these out, only originally create them.

We can therefore drop these for some small performance gains and a less
convoluted depenency chain.

(From OE-Core rev: 3c760ce4dc15d85be07aafbfea896e7276e0c2c4)

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

python-external-tarball: Modernise and rename

The external-python-tarball was intended to provide python for those users who
didn't have python 2.6. It has been unloved for a while so this refreshes
it a bit. Specifically:

* Include git and tar since these can have version issues on some systems
* Rename to buildtools-tarball to better reflect its purpose
* Remove the site-config file from the tarball, its pointless
* Prune down the environment file to just cover PATH and OECORE_NATIVE
  (which is needed to correctly install the tarball)
* Add missing grp module (used by do_package) from the unixadmin package

(From OE-Core rev: c8de009e96930709efe8f6aa4e65dd04a48081f2)

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

git/tar/findutils: Add nativesdk versions

We need to be able to generate a standalone tarball containing tar/git so
add nativesdk versions of the appropriate recipes to allow this to be possible.

Tweak the git perl paths to avoid warnings when building the nativesdk version,
ensure the binaries are wrapped correctly and avoid update-alternatives in
nativesdk-tar.

(From OE-Core rev: c91bb8c76e3bd45690e66f3de79cd3adfe45f600)

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

libxi: Add libxfixes as dependency

A clean build found libxi missing this dependency, possible build order
issue.

(From OE-Core rev: 7f5a0f98b3449d0a95fd6c12f1d6fa61b835fc2b)

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

self-hosted: Add libssp libraries to packagegroup/image

These libraries are needed to ensure when the -fstack-protector flags
are used the compile will complete and not fail due to missing libraries

[YOCTO #4586]

(From OE-Core rev: 543ede707a268ca17f972be54860e455b4de9e0b)

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

zlib: update to upstream version 1.2.8

Only version number and release year changed in license text.

(From OE-Core rev: 776fa3e49db312bcfd71c6b0637c989ad36fc84b)

Signed-off-by: Marko Lindqvist <cazfi74@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

tune-cortexa5.inc: Add tune file for cortex-a5

(From OE-Core rev: 50bc63c5c377d9fbb87b3efefc8c6f5473ba642a)

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

tune-cortexa.inc: create a common include for cortex-a armv7a tuning

The tuning files for the cortex-a* processors are mostly identical for
the A7,A8,A9,A15 processors. Rework these files to use a CORTEX_ID
variable to setup the tuning for each specific processor.

(From OE-Core rev: 3e4f4a1cf07ff7cf4c71566492385f8fbf581789)

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

libpcap: upgrade to 1.4.0

0001-Fix-disable-canusb.patch,
0001-The-leading-comma-looked-weird-remove-it.patch,
0001-canusb-needs-lpthread.patch patches included in
upstream, no longer needed.

(From OE-Core rev: 329e73e56b227c5c9fb00e2201fdf1099607b473)

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>

iptables: upgrade to 1.4.19.1

fix-iptables-extensions-build-error.patch no longer needed.

(From OE-Core rev: 02971543527e993b60132ddb101a9093efa3f324)

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>

kernel.bbclass: Fix link for modules-${MACHINE}.tgz

The modules-${MACHINE}.tgz should link to ${MODULE_TARBALL_BASE_NAME}.
But now it links to an invalid file ${MODULE_TARBALL_BASE_NAME}.bin.
Remove the extra .bin suffix.

[YOCTO #4595]

(From OE-Core rev: ef9fad8fff1c413bc0eab15625f051d4e2d0a201)

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

image_types.bbclass: add ubifs to IMAGE_TYPES

* the IMAGE_TYPES variable is used exclusively by hob
* IMAGE_CMD_ubifs is already present in the class
* ubifs images can be flashed on existing ubi volumes with 'ubiupdatevol'

(From OE-Core rev: 0dc5a90b2d95ad21025769b0300390a4ff0052ac)

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

python: enable IPv6 based on DISTRO_FEATURES

Base this on DISTRO_FEATURES, because we don't want it
unconditionally enabled.

(From OE-Core rev: 6956fb01e0c7332e4d8b59cbcb1cd997c54191fc)

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

multilib.bbclass: fix the PACKAGEFUNCS_append

The PACKAGEFUNCS_append = "do_package_qa_multilib" lacks a "space",
which would cause unexpected errors.

[YOCTO #3190]
[YOCTO #4396]

(From OE-Core rev: acd5fc716bc3095d568bd1474b79f3a0fd616eea)

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

insane/package: refactor packaging sanity tests

Refactor packaging sanity tests from package.bbclass to insane.bbclass
so that the message can respect WARN_QA (print the warning message and
go on the task) and ERROR_QA (print the error message and fail the
task).

- For the bb.warn(), give it a message name and add it to WARN_QA, then
  use package_qa_handle_error() to handle it.

- For the bb.error(), give it a message name and add it to ERROR_QA,
  then use package_qa_handle_error() to handle it.

- All the bb.warn() and bb.error() have been replaced in
  package.bbclass.

- A few bb.warn() and bb.error() in insane.bbclass have been kept since
  they can not be replaced or doesn't have to, for example the
  bb.error() in package_qa_check_license(), it will print the error
  message and then invoke bb.fatal() to fail the task, I think that we
  don't have to replace it with package_qa_handle_error().

- Put all the WARN_QA and ERROR_QA in one line, so that they can be
  redefined by the user easily.

[YOCTO #3190]
[YOCTO #4396]

(From OE-Core rev: 2f117ee615b703db07d3274ac592e2bd653743dd)

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

defaultsetup.conf: remove INHERIT_INSANE

The insane has been inherited by package.bbclass and becomes a
requirement, so we can remove it from defaultsetup.conf.

Note:
You can decide whether to take this patch or not.

[YOCTO #3190]
[YOCTO #4396]

(From OE-Core rev: 875f31facd02b47afb867aed76fef6b89a7b17cf)

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

insane/package: let package.bbclass inherit insane.bbclass

RP's comment:
"What we're trying to do is move everything to use a standard mechanism
for reporting issues of this type (do_package). With insane.bbclass, you
can elect whether a given type of error is a warning or error and fails
the task."

* The package.bbclass had used package_qa_handle_error() which is from
  insane.bbclass, and we will use it for handling other warnings and
  errors, so let package.bbclass inherit insane.bbclass, this change will
  make the insane as a requirement (always included).

* Change the "PACKAGEFUNCS ?=" to "+=", otherwise there would be an
  error like:
  Exception: variable SUMMARY references itself!

  This is because we let package.bbclass inherit insane.bbclass, and
  PACKAGEFUNCS has been set in insane.bbclass, so the "PACKAGEFUNCS ?="
  will set nothing, then the "emit_pkgdata" doesn't run which will
  cause this error.

* Add a QA_SANE variable in insane.bbclass, once the error type
  is ERROR_QA, it will fail the task and stop the build.

[YOCTO #3190]
[YOCTO #4396]

(From OE-Core rev: 852dead71387c66ec0cba7c71e3814a74e409560)

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

connman: update to 1.15

(From OE-Core rev: b1cd1c692d7b13e6865353c9bb2f7cf331c9f774)

Signed-off-by: Jack Mitchell <jack.mitchell@dbbroadcast.co.uk>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

weston: update to 1.1.0

Upgrade to 1.1.0, and disable the RaspberryPi and RDP compositors.  Hopefully
the RPi community can send a patch to add a PACKAGECONFIG for them.

(From OE-Core rev: 2e8e31a6812368688138b53c0aa92982354f832f)

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

wayland: update to 1.1.0

(From OE-Core rev: e7856f5527368c09c78f8e093c20fefd0e9fc83e)

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

mesa: update git recipe to 9.1.3 release

It's not bleeding edge but it's better than the previous pre-9.1 commit.

(From OE-Core rev: a8b8e21afce050b677e9168e08b3f318aee72e1f)

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

mesa: upgrade to 9.1.3

Merge mesa-PV.bb and mesa-PV.inc as the is no point to the split now we're only
building one variant of Mesa.

Drop 0002-cross-compile.patch isn't needed as libtool is used instead of mklib.
However, as libtool is used instead of mklib add fix-glsl-cross.patch to work
around build failures caused by the cross libtool being used to build host
binaries.

Add EGL-Mutate-NativeDisplayType-depending-on-config.patch to build correctly in
a non-X11 environment.

Drop dont-fail-if-libX11-isnt-installed.patch and
fix-egl-compilation-without-x11-headers.patch as they were backports.

license.html had some formatting and typo fixes, update the checksum.

(From OE-Core rev: d7033f4441183e53bf7fc8aa2293a9ec156a28dd)

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

pixman: update to upstream version 0.30.0

(From OE-Core rev: 25a77c7c5793ac0c1b1c2c686bfe9d966fb49023)

Signed-off-by: Marko Lindqvist <cazfi74@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

icu: update to upstream version 51.2

Only thing changed in license.html is addition of 2013 to
(c) years.

(From OE-Core rev: f994c58e21bc400a98d532095633fba810251300)

Signed-off-by: Marko Lindqvist <cazfi74@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

test_utils: import functions directly for conciseness

(From OE-Core rev: 2a147008cf81838cfc569640a30df0c1bfd74e08)

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

test-utils: handle import bb failing and skip the test

Instead of reporting an error when bb cannot be imported, skip the test
instead. This makes it a lot easier to iterate a test suite when we don't care
about this particular test.

(From OE-Core rev: c4a5bd810ca92d57c334113c528bd1d233b3eac4)

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

matchbox-terminal: Update to git HEAD 452bca253492a97a587f440289b9ab27d217353e

(From OE-Core rev: 6ff15b794d3027d315712e6a5abae1bed0159d1a)

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

grub-0.97: fix the objcopy error in configure

Get patch from ubuntu to fix this:
error: GRUB requires a working absolute objcopy; upgrade your binutils

(From OE-Core rev: 18bb555268ecc11d0fb7d4f404a38f1b453f2928)

Signed-off-by: Jackie Huang <jackie.huang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bdwgc: enable full version name in recipe name

bdwgc has version names that may or may not contain a revision
(for example 7.2d or simply 7.2). The current recipe specifies
the revision as a separate variable, which confuses the automatic
version checker. This patch enables using the full version name
as part of the recipe name.

(From OE-Core rev: 662b485e155257695ac7cc47f4cd26ddf94720ba)

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

subversion: upgraded to 1.7.10

(From OE-Core rev: 6866fd80ec59ef1e2d24263827237be8ff21584f)

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

libpng : upgrade to 1.6.2

upgrade from 1.6.1 -> 1.6.2

- removed patch contained upstream
- minor changes in png.h and LICENSE file ( changes
regarding the date )

(From OE-Core rev: c71a175add6493a2a419e46ab0ef96cbd3fcb70a)

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

gdk-pixbuf: upgrade to 2.28.1

Something along the lines of pngversion.patch was merged upstream, so drop that
patch.

Also use the new trim_version() function for a future-proof SRC_URI.

(From OE-Core rev: 30b32addd189e7acce73965c9176227d5f392ab9)

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

weston: clean up and add comments

(From OE-Core rev: 46159f632513de1f14d03e425799a0c193532298)

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

weston: add a weston-launch group

For weston-launch to be used by a non-root user, the user either needs to be a
member of the weston-launch group or own an active systemd user session.  Create
this group so users can be members of it.

(From OE-Core rev: 02e486ec9321fd3d7769c70979207335847a8f92)

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

weston: install the examples into weston-examples

(From OE-Core rev: 04234dfc2bc4c1bc4f3840fa9a162b25ba6ba049)

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

core-image-weston: basic image that boots directly to Weston

This adds an image that boots directly to Weston on KMS/DRM, and includes the
Weston terminal and gtk+3-demo for incredibly basic testing.

(From OE-Core rev: 18d88072f1392bbd98036f877c46c213af5b9722)

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

weston-init: basic init script to start Weston on KMS/DRM

weston-init is a very basic init script to start Weston as root on KMS/DRM.

To re-iterate, this runs Weston as root.  This will be fixed to use
weston-launch shortly.

(From OE-Core rev: eba825e4698f6923c32c347eb306abe9d7f3519d)

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

gstreamer: moved LIBV out of gst-plugins-package.inc file

Since gst-plugins-package.inc is going to be used by the following
GStreamer 1.0 recipes as well, LIBV has to be set outside of the .inc file.

(From OE-Core rev: 3cda3f1e3081ce51d8e964feff29e44558076522)

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

gstreamer: fixed -meta package rdepends

The -meta packages were depending on non-existing packages,
like -glib and -apps. The fix checks if packages in the PACKAGES list
are empty. If so, they are omitted from the rdepends list.

(From OE-Core rev: 7c556a1f437c0a5b472727f89ff6c3d5f835b63d)

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

libpcre: upgraded to 8.33

The license didn't modify, just the license years (2012->2013).

(From OE-Core rev: 8085ccf585b7d7f70f47362308ddacf9aa577649)

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

diffstat: upgraded to 1.57

(From OE-Core rev: 36f14ee692a632234a4670f4b4ec2db8f4e32241)

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

linuxdoc-tools-native : upgrade to 0.9.69

upgrade from 0.9.66 -> 0.9.69

- COPYING file contains now more detailed information.
- reworked a patch because it didn't apply.

(From OE-Core rev: 376612cbf7fa5c68ae3adb6f4c1b4427e69a1871)

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

zlib: update to upstream version 1.2.8

Only version number and release year changed in license text.

(From OE-Core rev: 4688c905776b9d995b2510224da269ac85bc8253)

Signed-off-by: Marko Lindqvist <cazfi74@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: bitbake: Improve remote server handling

Two fixes in bitbake related to running remote servers -
* can now specify correctly the bind port
* the information in print conforms to common server infrastructure

(Bitbake rev: b657208ee15ae065e5fcc2dd6e0051e03d246727)

Signed-off-by: Alexandru DAMIAN <alexandru.damian@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: xmlrpc: Allow server to remain memory resident

This patch enables the XMLRPC server to remain
resident in memory after a task is run, and to accept
a new controlling client. To check the server after
task completion, do

lsof bitbake.lock

in the build directory. Kill the server with kill.

(Bitbake rev: e823e1f0675ff3794eb39ef0b4df2d7a220f4013)

Signed-off-by: Alexandru DAMIAN <alexandru.damian@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: bitbake server: create common server infrastructure

In an attempt to minimize code duplication, create
clear interfaces, and maximize code reuse through OOP,
bb.server adds base classes for the BitBakeServer,
BitBakeServerConnection and actual server implementations
instructed in particular server types.

These classes document the minimum interfaces that the
derived classes must implement, and provide boilerplate code.

Changes to None, Process and XMLRPC servers as to use
the common server infrastructure.

(Bitbake rev: 6db4a64cef20f8d0aba804db4c4e1eec7b112b46)

Signed-off-by: Alexandru DAMIAN <alexandru.damian@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lrzsz: check locale.h in configure

fix the build failure:
lrz.c:284:13: error: 'LC_ALL' undeclared (first use in this function)

(From OE-Core rev: 8a2ac668d99f7d64c2acffc3a39cedb2d152be6e)

Signed-off-by: Jackie Huang <jackie.huang@windriver.com>
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gcc: add build directory to include directories

Add build directory to include directories by -I${B}/include which will
be searched before standard system include directories.

Both libunwind and gcc supply unwind.h, once gcc use libunwind's unwind.h,
the compiling will fail.

This patch is generally not applicable to the upstream as they do not
use libunwind.

(From OE-Core rev: 2b47bce78536a800205b2385bba69038351545e5)

Signed-off-by: Roy.Li <rongqing.li@windriver.com>
Signed-off-by: Jeff Polk <jeff.polk@windriver.com>
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gcc: Workaround for compiler flag mismatch

When cross compiling the target compiler, both the cross-compiler
and the host compiler are used.  However, the -W options used were
discovered from the cross-compiler and may be incompatible with the
host compiler.

(From OE-Core rev: 1f40202ce3c1282674b6cea39ef709972275f201)

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

util-linux: Add support for older hosts

Older hosts don't support some of the features required by
the latet util-linux.  Add workarounds or revert changes to older
versions to make it work.

(From OE-Core rev: da5b23e45c7e4dea2f3802ff5af5c81b08aba201)

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

udev: do not recommend pciutils/usbutils ids

udev does not require those files to be operational and they add ~350KB
to rootfs

(From OE-Core rev: af85ad00d31db0cfe499af815357f7f118e7e546)

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

bitbake: Drop bitbake-runtask (removed upstream a while ago)

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

alsa-tools: Fix configure race

aclocal is being called here directly,
not called by autotools.bbclass wrapper.
aclocal files are installed in sysroot,
and are removed while build is still running.
This translates to a possible race condition during the build.

Fixes [YOCTO #4358].

(From OE-Core rev: dea66ade1184cef6aeb242d87867759ca44a8895)

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

autotools.bbclass: Factor out aclocal copying function

Some recipes may need to manually call the aclocal copying
functionality so factor this out into a function.

(From OE-Core rev: a5a08543c8cec43d993b2bba0ad6a9357c0a5e04)

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

yocto_kernel: modify the msg when adding duplicate items

Privious check-in "yocto_kernel: check current items before add a
new one" had been merged before I apply the feedback from Zanussi, Tom.
Now fix it as a new patch.

This fix modify the output message when customer adding duplicate
items.

[YOCTO #4558]

(From meta-yocto rev: 530c6efa85b1798d30db4c6c83a748b100b8c1c3)

Signed-off-by: Ning Zhang <ning.zhang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

xf86-video-modesetting: backport fix for graphics corruption

This fixes graphics corruption that can occur when using 32 bpp pixmaps
with 24 bpp framebuffer.

(From OE-Core rev: dc8ad21502549b33b4c59c31df66d15d2f656df7)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

makedevs: Make count actually behave as a count for device blocks

Previously count actually behaved as end, and did not take start into
account.

(From OE-Core rev: 4fe2b01bfd2831b002e7138dadbc0437df6e9ed6)

Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>