systemd: Fix misc journald memory bugs

These set of patches fix journald exhibiting some issues
under load.

One of the prevelant issues is that when appending to journal
it is not able to allocate memory and starts taking 100% cpu
spewing errors like

systemd-journald[2934]: Failed to write entry (19 items, 452 bytes), ignoring: Cannot allocate memory

Other memory issues crept up with time e.g.vacuuming

(From OE-Core rev: b1bdc1c6fb6914d85f888acde9d806d5560c84d8)

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

bitbake: runqueue: Fix setscene hard dependency problems

Commit c54e738e2b5dc0d8e6fd8e93b284ed96e7a83051 added in the idea of hard dependencies
such as the case a setscene has a hard dependency on pseudo-native and that
dependency wasn't available from sstate for some reason.

Unfortunately the implementation was a bit too enthusiastic, causing rebuilds
of things when it wasn't necessary. A test case was:

bitbake quilt-native
bitbake quilt-native -c clean
bitbake <some-image>

and then you'd watch quilt-native get rebuilt for no good reason.

The clue to the problem is in the for loop where it never depends on
the item being iterated over.

The fix is to include the exact list of hard dependencies rather than
guessing. With these changes, the use case above works, the one in
the original commit also works.

This patch also adds in or cleans up various pieces of logging to
allow issues like this to be more easily debugged in future.

(Bitbake rev: 81bd475585ff1b44b390036b1eca0feae7c149eb)

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

image.py, package_manager.py, rootfs.py: dump command output on error

Print the entire command output in case of errors.

(From OE-Core rev: 2253c9ac2caa61dee0bd4fea04d4d77b79be7b36)

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

lib/oe/package_manager.py: RpmPM: fix issue with multilib builds

Use python sets instead of lists, to avoid duplicates. When doing a
multilib build, "smart channel --add" fails because it tries to add
'all' channel twice.

(From OE-Core rev: 730d675090eec5c03e444f0448d96a52035d0bef)

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

lib/oe/package_manager.py: check that package list is not empty

If we're just attempting to install packages and the package list is
empty, just return.

(From OE-Core rev: 8bf5d80e3502a71fdda688aace0799b47ae19b3b)

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

rootfs.py, package_manager.py, sdk.py: Fix building from feeds feature for opkg

When using opkg as the PM backend, one has the option to provide custom
feeds to create the rootfs from.

This commit:
 * fixes this in the refactored code;
 * moves the custom config creation code to python;
 * clean up the package-ipk.bbclass;

(From OE-Core rev: 19c538f57c8fa7c566e88a6dbe13ea4826d4f26c)

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

lib/oe/package_manager.py: OpkgPM, bad recommendation handling change

The following commit:

rootfs_ipk: Ensure that BAD_RECOMMENDATIONS are honoured for all
architectures

changed the way BAD_RECOMMENDATIONS are handled. Make the change in the
new code too.

(From OE-Core rev: 7e518e399da51de3b159bd6804735b2f14c39357)

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

lib/oe/package_manager.py: do not use stamps

Using stamps shortens rootfs/image creation because the package indexes
are not created again if no new package was installed in the deploy dir.
Unfortunately, there are some syncronization issues with
do_package_write_*() and the indexes are not properly created.

Will have to revisit this issue in the near future. For now, lose it.

(From OE-Core rev: 276f2a8a072ce243daa74210934ee63046e9c480)

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

lib/oe/rootfs.py: OpkgRootfs, fix issue in _get_delayed_postinsts()

Status file path was not created correctly.

(From OE-Core rev: 628c4bce36ca0bc2b316ec78cf58b02a42e7f35f)

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

lib/oe/package_manager.py: OpkgPM/DpkgPM fixes

This commit:
 * adds missing mark_packages() function for OpkgPM. This is needed to
 * manually change the packages installation status in the status file;
 * fix OpkgPM remove_packaging_data() issue;
 * fix OpkgPM handle_bad_recommendations() issue;
 * improve OpkgPM/DpkgPM mark_packages() regex pattern;
 * fix DpkgPM list_installed() problem;
 * fix DpkgPM _create_configs() problem: status file should not be
   truncated if it already exists/

(From OE-Core rev: d50a40100763dcbfabec2c147fdfcb46aa909329)

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

lib/oe/image.py: fix image size calculation routine

The IMAGE_ROOTFS_EXTRA_SPACE value was not correctly added to the base
size.

(From OE-Core rev: f3ac2a68ae0ae9bdf2a03c1b6d0f83c9918d58bd)

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

lib/oe/package_manager.py: DpkgPM, use lock on deploy dir

Lock deploy directory against concurrent index creation.

(From OE-Core rev: 127d19b84e78b5e650d0e6d7eaf29a8b7d799e5e)

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

scripts/postinst-intercepts: do not redirect errors to /dev/null

The redirection was necessary with the old bash code because the log
checking routine was searching for error strings in the log and abort
the build in case failures occured. With the new python code,
redirecting the intercept stderr is no longer necessary. This also makes
the intercept hooks easier to debug.

(From OE-Core rev: e2c3c59b26d84cd4052a953adafcd4d456264c26)

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

lib/oe/rootfs.py: warn the user if intercept hooks fail

The idea of postinstall intercept hooks is to run CPU intensive
postinstalls (like the ones generating font/pixbuf/icon caches) on host,
at rootfs generation time. So, ideally, the user would like to know if
the intercepts fail on host.

With this patch, the user will see warnings on console if intercept
hooks fail to execute.

(From OE-Core rev: 5666ce9720132a76b5ff6e99712e240b8dc8e47a)

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

lib/oe/package_manager.py: RpmPM fixes

This commit:
 * fixes a crash when handling interecept hook failures which happened when
   MULTILIB_GLOBAL_VARIANTS was not set;
 * convert dashes to underscores and use sets (so that we make sure the
   items are unique) when creating RPM repos;
 * uses a regex pattern to search for packages in the feeds list. The
   old method could match also strings in the middle. For example: 'rpm'
   matched 'kernel-module-lttng-probe-rpm" in qemux86_64 feeds;
 * issue a bb.fatal if smart returns error while installing packages.
   Otherwise we might end up with an incomplete image...
 * fixes the /etc/rpm/platform file creation;

(From OE-Core rev: b98c7e4945f1c36a6e4f98144a3af4f3049450ae)

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

lib/oe/package_manager.py: moving global SCRIPTLET_FORMAT and DB_CONFIG_CONTENT to RpmPM

The tweak made the code looks better.

(From OE-Core rev: 8e06be633222635c549d7f067218e8b833e76bda)

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

package_manger.py/rootfs.py: rename save_rpmpostinist with save_rpmpostinst

The tweak made the code looks better.

(From OE-Core rev: 61981755b4738ccc49be0e790de828f4166ef0e8)

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

populate_sdk_rpm.bbclass: remove old shell code

The old code no more needed since the job is done in python now.

(From OE-Core rev: 96bf2a93bb75fea23732014106039e14d72cd1c7)

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lib/oe/rootfs.py: fix invoking _run_intercepts failed

- The variable D is needed by intercepts scripts;

(From OE-Core rev: 3d37dcdb836b59d8a7f880ce22d06760eb8af232)

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lib/oe/package_manager.py: fix invoking rpm save_rpmpostinist failed

- Use self.target_rootfs instead of self.image_rootfs as saved dir;
- Remove redundant comment line;

(From OE-Core rev: 41d6edab676277afbcaa6cf41cd64a5663f09dc7)

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lib/oe/rootfs.py: fix invoking rpm _handle_intercept_failure failed

(From OE-Core rev: 7858c8ac141a7bf66c65ddc7e4b39ad937a7e084)

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

Cleanup rootfs_rpm,package_rpm bbclass files

This commit cleans up the functions that were ported to python.

(From OE-Core rev: a4fe5263a683b73d0318fe3c0c55382c084f25b5)

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

image.bbclass/populate_sdk_base.bbclass: shift position of sdk complementary install definition

The sdk complementary install operation was defined in image.bbclass,
but the sdk recipe (such as meta-toolchain.bb) didn't inherit this
bbclass but populate_sdk, and both of image and populate_sdk bbclass
inherited populate_sdk_base bbclass, so move the sdk complementary
install definition to populate_sdk_base bbclass fixed this issue.

(From OE-Core rev: 907458935bab391f5bfba8f581ea9835078548d9)

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lib/oe/sdk.py: support RpmRootfs

- Implementation RpmSdk class

(From OE-Core rev: 417b27ce5c1ea2fe2057caa898291cce0c043359)

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

license.bbclass: fix license_create_manifest failed

While rpm incremental image generation enabled, invoked
license_create_manifest failed.
...
| ln: failed to create symbolic link 'tmp/work/qemux86_
64-poky-linux/core-image-base/1.0-r0/rootfs/usr/share/
common-licenses/alsa-conf-base/generic_GPLv2':File exists
| WARNING:tmp/work/qemux86_64-poky-linux/core-image-base/
1.0-r0/temp/run.license_create_manifest.26327:1 exit 1 from
|   ln -s ../${lic}/home/jiahongxu/yocto/build-20140120-
yocto-qemux86-64/tmp/work/qemux86_64-poky-linux/core-image
-base/1.0-r0/rootfs/usr/share/common-licenses/${pkg}/${lic}
...

(From OE-Core rev: 39f8658033c886fedefc9038a673f9a5edad7d26)

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

image.bbclass/license.bbclass: ajustment list_installed_packages invoking

Since the list_installed_packages() function has refactored in
python, do the necessary adjustments to license_create_manifes-
t() and write_image_manifest() in license.bbclass and image.bb-
class respectively;

(From OE-Core rev: 4477991633d2971c17afbd0874b1ab2efb7d6eef)

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lib/oe/image.py: fix get rootfs_extra_space failed

The value of IMAGE_ROOTFS_EXTRA_SPACE is "0 + 51200",
we should use eval rather than int in python.

(From OE-Core rev: dd3418e2db732ca1bc78fd93efa3f08da88a1183)

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rootfs.py: strip kernel-abiversion useless readline

It failed to read Kernel version from kernel-abiversion
file, the reason was it didn't strip the readline.
...
 Error: Kernel version 3.10.25-yocto-standard
  does not match kernel-abiversion (3.10.25-yocto-standard)
...

(From OE-Core rev: 17a0a8f44fe344ab1ae49730d9f17705b43987e5)

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rootfs.py: fix uninstall uneeded pkgs failed

The refactor of shell function rootfs_uninstall_unneeded is incorrect,
it should check and update the installed_pkgs.txt file for the existance
of the packages that were removed.
...
rootfs_uninstall_unneeded () {
    if ${@base_contains("IMAGE_FEATURES", "package-management", "false", "true", d)}; then
        if [ -z "$(delayed_postinsts)" ]; then
            # All packages were successfully configured.
            # update-rc.d, base-passwd, run-postinsts are no further
            # use, remove them now
            remove_run_postinsts=false
            if [ -e ${IMAGE_ROOTFS}${sysconfdir}/init.d/run-postinsts ]; then
                remove_run_postinsts=true
            fi

            # Remove package only if it's installed
            pkgs_to_remove="update-rc.d base-passwd ${ROOTFS_BOOTSTRAP_INSTALL}"
            for pkg in $pkgs_to_remove; do
                # regexp for pkg, to be used in grep and sed
                pkg_regexp="^`echo $pkg | sed 's/\./\\\./'` "
                if grep -q "$pkg_regexp" ${WORKDIR}/installed_pkgs.txt; then
                rootfs_uninstall_packages $pkg
                sed -i "/$pkg_regexp/d" ${WORKDIR}/installed_pkgs.txt
                fi
            done
...

(From OE-Core rev: 9cdecb3935962653733705ad6313558bfd4fda29)

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lib/oe/rootfs.py: Tweak the rootfs removing operation

Move the bb.utils.remove(self.image_rootfs, True) from the base class
constructor, to Opkg/Ipkg constructors after super's constructor is
called.

(From OE-Core rev: 66d86c4bbb9423c9fe0fc89005530be54068d401)

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lib/oe/rootfs.py: support RpmRootfs

- Implementation RpmRootfs class
- Refactor list_installed_packages in python

(From OE-Core rev: b3eeb0edf3f0296d30d7e62628bde722402bdb9d)

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lib/oe/package_manager.py: support RpmPM

- Implementation RpmPM class
- Support rpm incremental image generation

(From OE-Core rev: ca5203d6b1bb0cbf7830ea4f46109c6867138efc)

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lib/oe/manifest.py: add rpm image manifest creation

Implementation RpmManifest class.

(From OE-Core rev: a0ef59ef9263653877db4853633883f2684d7a30)

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

populate_sdk_(deb|ipk).bbclass: remove old shell code

The old code no more needed since the job is done in python now.

(From OE-Core rev: 58f06621aa1a5618dc969e8d1b537e32763a9461)

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

image.bbclass: change POPULATE_SDK_POST_TARGET_COMMAND

Since rootfs_install_complementary() is now implemented in python for
each backend, remove it from POPULATE_SDK_POST_TARGET_COMMAND. Call it
directly in python.

(From OE-Core rev: c941fce37eb1ca7b77e5885c3e3e776f9538754b)

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

populate_sdk_base.bbclass: activate the new python routines

Lose the old shell populate_sdk_image() function and use the new python
implementation.

(From OE-Core rev: 5a81b8a0820ceac972a68af2caebdc8d2f3945a1)

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

lib/oe/sdk.py: add SDK class

This new file contains the python 'populate sdk' implementation of the
old bash populate_sdk_image() function for Opkg and Dpkg.

(From OE-Core rev: 6247efaba592db924e6466c39aef441f0e07c62a)

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

lib/oe/package_manager.py: fixes for dpkg backend

This commit contains the following fixes:
 * pass the apt config directory to the DpkgPM constructor, so one can
   instantiate this class multiple times and give it different config
   files (like for creating SDK);
 * change constructor argument name from 'dpkg_archs' to 'base_archs';
 * export APT_CONFIG environment variable before calling apt-get, not in
   constructor. If done in constructor, the last class instantiation,
   sets the environment, which is note desireable;

(From OE-Core rev: dc626cbcfd37c940bb8739b14d3ab8097e1760ea)

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

lib/oe/rootfs.py: use Manifest.INSTALL_ORDER

Since the Manifest class has this property, use it. This contains the
default package installation order.

(From OE-Core rev: bb1fb4ceb544c161ed30cd102155657e3771859b)

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

lib/oe/manifest.py: create manifests for SDK too

This commit contains several changes:
 * it is possible to create manifests for following types of images:
   regular image, target SDK and host SDK. To distinguish between these
   types of manifests, one has to pass the manifest_type argument to the
   contructor or create_manifest() wrapper. The manifest type can have
   the following values: image, sdk_host, sdk_target;
 * move image_rootfs variable to _create_dummy_initial() since it's used
   only here. This function will probably be removed in the future;
 * fix a bug in the Dpkg class;
 * add INSTALL_ORDER property to Manifest class which contains the
   default install order for the packages and will be used Rootfs/Sdk
   classes;

(From OE-Core rev: a8c1b7504bf9cd5625fdecfdc3c3adce53aa164c)

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

image.bbclass, lib/oe/rootfs.py: remove intercept_scripts directory

Remove the directory, manually, in the Rootfs.create() function.

(From OE-Core rev: 37a9dc05272e7e49c8a475ccb3d10880dcee4763)

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

Cleanup image,rootfs_ipk,package_ipk bbclass files

This commit cleans up the functions that were ported to python.

(From OE-Core rev: d950ef40a3eae4b54cc62828e66d84a62d78c447)

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

lib/oe/rootfs.py: add support for opkg backend

(From OE-Core rev: 0650e1f2354cdb6f925040d9d07e2c785274a592)

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

lib/oe/package_manager.py: add support for opkg backend

Additionaly, the commit contains a couple of minor changes
(comments, error printing, etc).

(From OE-Core rev: ef3faaef6b1a25c943a8d5594ce55b0e558b62f3)

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

lib/oe/package_manager.py: use the bitbake APT_ARGS variable

This commit will revert on using the bitbake APT_ARGS variable, so users
can alter the way apt is called without needing to change it in code.

(From OE-Core rev: ae337edc8ac441c947d1543c2800f37104a0be70)

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

lib/oe/package_manager.py: add deploy dir locking mechanism

This is needed in order to serialize the index file creation when
multiple do_rootfs tasks are running in the same time.

(From OE-Core rev: cb03d15482569c2e56232c921526938dcecfdb68)

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

lib/oe/manifest.py: create global variables for package types

Manifest class clients don't really need to know how package types are
encoded.

(From OE-Core rev: bac2e279005b601daff4d53549612ceb76a6a857)

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

package_(deb|ipk).bbclass: remove the stamp when creating package from cache

If the packages are created from cache, we need to remove the stamp so
that we re-generate the index files at do_rootfs time.

(From OE-Core rev: dc06a91144b79a152eb481f6d36f6c328321b7c4)

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

oe/manifest.py: Add manifest creation for opkg

In this commit:
 * add ability to create initial manifest for opkg;
 * make var_map available to all backends;

(From OE-Core rev: 8d14fbedcee02b723288004ae29a5c29524eec5a)

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

Activate the new python rootfs/image creation routines

This commit will:
 * remove old bash code common to all backends;
 * create a new do_rootfs() python function that will use the new
   rootfs/image creation routines;
 * allow creation of dpkg based images;
 * fail for rpm/opkg (not implemented yet);

(From OE-Core rev: a83144bac8d67704ff66f5dc0fc56f5b63979694)

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

lib/oe/rootfs.py: add new library for rootfs creation

This library will be used to generate the rootfs.

Recommended usage:
 create_rootfs(d, manifest_file)

this will determine the PM backend used, save the shell environment and
will call the appropriate backend implementation (DpkgRootfs(d,
manifest_file).create()).

NOTE: this commit adds Dpkg support.

(From OE-Core rev: 5ad8c21bc7f2213d1287cecf07d00a61d1456ff7)

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

lib/oe/package_manager.py: add new package management library

This new library is intended to be used by the new python rootfs
creation code.

It implements the rpm/dpkg/opkg package management backends: RpmPM,
DpkgPM and OpkgPM.

The base API is this:
  update()
  install()
  install_complementary()
  remove()
  write_index()
  remove_packaging_data()
  list_installed()

All implementations have to provide these functions. Some backends may
need to implement additional functions though.

(From OE-Core rev: 224e5053044b4c7966fea3b8d199a3dfd4fe438e)

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

lib/oe/manifest.py: add library for image manifest creation

This new library allows for the creation of 2 types of manifests:
 * initial manifest - used by the new rootfs creation routines to
   generate the rootfs;
 * final_manifest - this will contain all the packages in the image,
   after all installations finished;

Usage:
  Manifest(d, manifest_dir).create_initial()
  Manifest(d, manifest_dir).create_final()
or using the provided wrapper function:
  create_manifest(d, False, manifest_dir) -> creates initial manifest
  create_manifest(d, True, manifest_dir) -> creates final manifest

If manifest_dir argument is ommited, it defaults to ${WORKDIR}.

NOTE: this commit creates fixed manifests for minimal/sato/sato-sdk
images, for Rpm & Opkg backends, in order to help speed up
development of rootfs refactoring. Dpkg initial manifest creation is
implemented.

(From OE-Core rev: a9d8e5e5878d14b4804317a7f7ea6394fca5e010)

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

lib/oe/image.py: add new image creation library

This will replace the old bash image creation code. This needs the
rootfs to be already generated in order to work.

Usage:
  Image(d).create()
or using the provided wrapper function:
  create_image(d)

(From OE-Core rev: b75b78ce534fbf0d4de2f7f66af5b721d68b7471)

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

lib/oe/utils.py: add pre/post process helper function

This helper function will be used to execute pre/post process commands.

(From OE-Core rev: 23d409558cb12cbf0300156e67f768a13442910a)

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

bitbake: bitbake: build.py: create separate function for shell trap creation code

Currently, the shell trap code was created in exec_func_shell(). Split
the function so that we can create the code separately.

Also, some whitespaces were automatically deleted by my editor. Since
this is not necessarily a bad thing, leave these changes too.

(Bitbake rev: c712e622d20c61a07c9c172b60e9dc6beae14197)

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

python-pycurl: upgrade to 7.19.3

Changes:
 * Refreshed no-static-link.patch;
 * LGPL&MIT license files changed names;
 * MIT license file added a contributor to the list;
 * README.rst license snippet changed to reflect the filename changes
   and the new contributor;

(From OE-Core rev: b5b4898cd409036161c62891e9618d9ab3f891f9)

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

xserver-xorg: upgrade to 1.15.0

License file changed slightly: a new person has been added to the list
of authors.

Add dependency of presentproto.

(From OE-Core rev: b18fb872baa0deb89f1b8250883122fb5f6328c5)

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

xf86-video-intel: add compat patch for xorg >= 1.14.99.2

(From OE-Core rev: 0d80bec421b45dc5b997eecdce3fe21f00784cd7)

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

presentproto_git.bb: add recipe

This is needed by xserver-xorg 1.15. Since presentproto has just been
created, use a git recipe because the initial version is missing some
fixes, license file, etc.

(From OE-Core rev: 05026e3a3dbe1c8a652e5af03a13967c1e7a9817)

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

xf86-input-synaptics: upgrade to 1.7.3

(From OE-Core rev: 0551c0f90247ff14bbdd7405c999711241fa1fa6)

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

libmpc: upgrade to 1.0.2

(From OE-Core rev: 11e2c03411365fe1cc16c79e5902719dd25411bd)

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

xkeyboard-config: upgrade to 2.11

(From OE-Core rev: 633a6b1e6cc7b0d29b418267521ea2c560b210f5)

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

libdrm: upgrade to 2.4.52

(From OE-Core rev: b4a6fec9be9581a99768a26d55a106cb444ff598)

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

xtrans: upgrade to 1.3.3

(From OE-Core rev: 218beaf1072a29f7de60018c417cdbf4c1d0d26d)

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

xcb-util-wm: upgrade to 0.4.0

(From OE-Core rev: cc5de7844640f63bbf8aec6c0963d8d9a24af405)

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

gdb: upgrade to 7.6.2

(From OE-Core rev: bfeb7ce8b525890c150f0b1cc2c5df089aa4b052)

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

xextproto: upgrade to 7.3.0

(From OE-Core rev: 5bcf6a03c83cb91793d891f1e860c200eaa68be3)

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

nasm: upgrade to 2.11

(From OE-Core rev: 5847c2eb3fb03ce6de0842f69ae0c1d44ef3c8a0)

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

binconfig: mangle ${base_libdir}

Some recipes are installing libraries into ${base_libdir} (typically /lib) and
also use a foo-config binary to identify compile paths, for example
libusb-compat.  Without mangling ${base_libdir} the ${base_libdir} path is
passed to the compiler, where it looks like a host path and results in
compile-host-path QA errors.

(From OE-Core rev: ccd9abdccb84d713427541b6ee29a0e217360e74)

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

useradd.bbclass: Add ability to select a static uid/gid automatically

[YOCTO #5436]

Automatic selection of static uid/gid is needed for a dynamically generated
passwd and group file to have a deterministic outcome.

When a package is installed and instructs the system to add a new user or
group, unless it selects a static uid/gid value, the next available uid/gid
will be used.  The order in which packages are installed is dynamically
computed, and may change from one installation to the next.  This results
in a non-deterministic set of uid/gid values.

Enabling this code by adding USERADDEXTENSION = "useradd-staticids", and
adding a preconfigured passwd/group file will allow the continued dynamic
generation of the rootfs passwd/group files, but will ensure a deterministic
outcome.  (Dynamic generation is desired so that users and groups that have
no corresponding functionality are not present within the final system image.)

The rewrite params function will override each of the fields in the
useradd and groupadd calls with the values specified.  Note, the password
field is ignored as is the member groups field in the group file.  If the
field is empty, the value will not be overridden.  (Note, there is no way
to 'blank' a field, as this would only generally affect the 'comment' field
and there really is no reason to blank it.)

Enabling USERADD_ERROR_DYNAMIC will cause packages without static uid/gid
to generate an error and be skipped for the purpose of building.  This is
used to prevent non-deterministic behavior.

USERADD_UID_TABLES and USERADD_GID_TABLES may be used to specify the name
of the passwd and group files.  By default they are assumed to be
'files/passwd' and 'files/group'.  Layers are searched in BBPATH order.

(From OE-Core rev: 18c99dac52b746b88cd084eb4c2a2ef0329a6ff3)

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

useradd.bbclass: Fix build time install issues

When the system attempts to populate the sysroot's passwd/group files, it
does so in a single block.  However, with the way it was previously
implemented, the system would always run through the code necessary to
populate the sysroot, even in the case of target packages.  This had
the side effect that a cross-installed filesystem may not match a
target installed filesystem.

The code was slightly reorganized to ensure that the cross/target installed
pre-install script behavior is the same.  It also moves the block that
configures the sysroot parameters to the sysroot specific section of
the code.

Also some minor validation was occuring even on nativesdk packages.
Nativesdk packages should be skipped when processing useradd ops.

(From OE-Core rev: 99fd0f14fd774c9194f62795e6023880e3aa5612)

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

scripts/oe-selftest: return based on the test results

Regardless if the tests passed or not the script returned 0,
which isn't what one would expect.

(From OE-Core rev: c38f943c7fbb1fc077c875099dce8f73f41043b9)

Signed-off-by: Stefan Stanacar <stefanx.stanacar@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

minicom: updated download link

minicom tarball is now downloaded from
the recommended web location.

Final fix for [YOCTO #5781].

(From OE-Core rev: b01e4438a08a0b9c6950af666fa13eaf71b45fc9)

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

gdk-pixbuf: enable ptest

(From OE-Core rev: b78d9b2f059f3f87c474bba8eb6258bb7299cabc)

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

gnome-desktop-testing: upgrade to 2014.1

no-introspection.patch has been merged upstream.

Update license checksum as copyright dates have been updated.

(From OE-Core rev: c8272e900ed87761d95ba0065a954dc2bb9f760c)

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

wic: Honor --size for --source partititions

Instead of simply creating partitions large enough to contain the
contents of a --source partition (and adding a pre-specified amount of
padding), use the --size used in the partition .wks statement.

If --size isn't used, or is smaller than the actual --source size,
retain the current behavior.

(From OE-Core rev: 23b6c5ea4d48cdf731e5202991961a0e4b10ff29)

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

jpeg and directfb: cosmetic, fix perms for patches

(From OE-Core rev: c1375208b363d0ac281189889efd450685bc46bc)

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

rootfs_deb: Fix whitespace issue

Fixing the following warning:
WARNING: Variable rootfs_deb_bad_recommendations contains tabs, please remove these (/srv/ssd/sgw/poky/meta/recipes-extended/images/core-image-lsb-dev.bb)

(From OE-Core rev: aec771c113ae4347af7ef98bef8c289e3a5069e2)

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

systemtap: add aarch64 support

bump SRCREV to include the relevant commits for aarch64 support
update COMPATIBLE_HOST to include aarch64
cleanup md5sum/sha256sum since we use git

(From OE-Core rev: 1c2cbd3b54a9bf589d41c8320734a95af52c36d6)

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

oeqa/utils: targetbuild: don't use bb.fetch anymore

When running tests outside of the build system we can't use
bb.fetch anymore. It was nice but tests and their modules
need to rely on the data storage only as that gets exported.
This module is used by the oeqa/runtime/build* tests.

(From OE-Core rev: 3caf8e244ea94f62a93f3b40e73e15ea78fc2880)

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

lib/oeqa: add module for running tests outside of the build system

This script will run the exported tests outside of the build system.

Simplest way to test this is with a qemu image that you manually start.
For an already build image use this in local.conf:
TEST_EXPORT_ONLY = "1"
TEST_TARGET = "simpleremote"
TEST_TARGET_IP = "192.168.7.2"
TEST_SERVER_IP = "192.168.7.1"
Export the tests: bitbake core-image-sato -c testimage
Then: runqemu core-image-sato
And: cd build/tmp/testimage/core-image-sato
     ./runexported.py testdata.json

The contents of build/tmp/testimage/core-image-sato can be moved on another machine
as long as some paths are updated in the json.
The exported data contains paths to the build dir. We only care about DEPLOY_DIR/rpm (
if the rpm and smart tests are enabled), so running the tests on other machine
means that the user has to move the contents and call runexported with --deploy-dir PATH:
 ./runexported.py --deploy-dir /path/on/another/machine testdata.json
runexported.py accepts other arguments as well, see --help.

[YOCTO #5613]

(From OE-Core rev: 087ee840ad642bada6fe0b02311f05a595ea2e65)

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

testimage: add ability to export tests

Add the ability to export the tests so that they can run independently of
the build system, as is required if you want to be able to hand the test
execution off to a scheduler.
Booting/deployment of the target is still handled by the build system,
as before, only the execution of the tests happens outside of the build system.
Tests exported are the ones defined in TEST_SUITES.

No tests have been changed as interesting parts of the data store have been
exported and tests can continue to query them as before. Small adjustments were made
for a couple of oeqa modules though.

[YOCTO #5613]

(From OE-Core rev: 155dd52e0f707e06f50756584a50f744ba6b7844)

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

classes/populate_sdk_base: remove nostamp from do_populate_sdk

We've removed nostamp from do_rootfs in image.bbclass in OE-Core commit
8505008c115efb54d18e5f25441c7a938a32ffaf, and we should do the same for
do_populate_sdk here for the same reason - we can now rely on task
signatures so if nothing has changed, we don't need to re-run it.

(From OE-Core rev: de9b693f4ff311f1310a1c6005e0d5c225aabef6)

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>

bluez5: upgrade to 5.14

(From OE-Core rev: 97df23a15f875d4f98c579d707cb788d577de539)

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>

sbc: upgrade to 1.2

- allows high precision encoding;
- setup enc/dec A2DP streams;
- switched to ${BP};
- license was wrong, corrected;
- added Intel Corporation as a copyright owner,
triggering a change in copyright checksum.

(From OE-Core rev: 55644525da23b6cbb6ffc46400358a2bef9beec8)

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>

ossp-uuid: Use alternative source mirror

Original ftp location is inaccessible

Fixes

WARNING: Failed to fetch URL
ftp://ftp.ossp.org/pkg/lib/uuid/uuid-1.6.2.tar.gz, attempting MIRRORS if
available

(From OE-Core rev: 3109be6cca065c42ec0fddace23531219820205b)

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

openssh: drag in required PAM modules.

/etc/pam.d/opensshd lists keyinit and uid as required, so add them to
RDEPENDS when PAM is enabled.

(From OE-Core rev: b2f49f9d2d8cd033611108c2bfe4871d02df0887)

Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

openssh: package sshd PAM config inside openssh-sshd package

Without this PAM integration is broken after installing openssh-sshd

(From OE-Core rev: 2ba31c0f0fad6a1917e282dc381e8632db08200c)

Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

binutils: Add backports to resolve systemd (from git) breakage

Backport the patches for PR2404 and PR16476 from binutils to
resolve failures when compiling systemd from git.

(From OE-Core rev: 25e376f6c8c66cca95ae7fa7f512b7dadfdd6140)

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

systemd-compat-units: Handle nfs scripts

The nfs init scripts have both systemd and sysvinit versions with
non-matching names so we need to mask out the duplicate scripts
else we see NFS errors on bootup.

(From OE-Core rev: 6fadb3f27c48bb92f9f9de5977707a6b244aac54)

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

build-appliance-image: Update to poky commit b37dd451a52622d5b570183a81583cc34c2ff555

This update will add improvements to HOB,
fix issues encountered during build.

Fixes [YOCTO #5759].

(From OE-Core rev: cda502815c6acf789e1a0db7a7a1a7015b4ef71d)

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

pcmanfm: apply a patch to the source, not a generated file

no-desktop.patch was applying to a generated file, not the true source, which
meant it may not appear to be applied in the resulting package.

(From OE-Core rev: c6dee0c0388572aa3f28d363f94c0749e66f1289)

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

distcc: rename systemd service to distcc

The sysv init script is called distcc, so rename the service to match so that in
hybrid systemd/sysvinit images the service correctly masks the init script.

This prevents such images from failing to start the distccd unit with error code
102, as the ports it wants to bind to are already taken.

(From OE-Core rev: ee58b618ec68c02b3e8759086e1dcc45c1fe3970)

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

glib-2.0: fix race in parallel builds

There is a race condition in the makefile that can result in build failures like
this in parallel builds:

| ./gdbus-test-codegen-generated.h:7:0: error: unterminated #ifndef
|  #ifndef __GDBUS_TEST_CODEGEN_GENERATED_H__

Fix the rules to stop this happening.

(From OE-Core rev: 97ccd2b841c9dc598dbe39162f335bcde48a7c26)

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

systemd: add a hard dependency on udev=$EXTENDPKGV

If the systemd/sysvinit DISTRO_FEATURES are changed without wiping the package
feeds it's possible to build an image that pulls in mismatching versions of
systemd and udev.  This leads to images that are broken and don't boot
correctly.

Prevent this by adding a version-locked dependency on udev in systemd so that
images that attempt to install mis-matching versions don't build.

(From OE-Core rev: cb64f979ac4d792027a4a85fe086d0854e7bc9bc)

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

iproute2: split out package for tc

tc (the IP traffic control utility) isn't often used and makes up a
reasonably large part of the iproute2 package as well as having a
runtime dependency on iptables, so split it out into its own package.

(From OE-Core rev: 1d353cb30b93cd08d7a0f743534c1cd712bbe018)

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

systemd-compat-units: don't mask dbus, the recipe does this itself

(From OE-Core rev: a5436fa039d1e3153fb724a3fa59c5955995aa2d)

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

dbus: mask the dbus-1 init script if using systemd

If the image is built with both systemd and sysvinit there'll be a dbus-1 init
script and a dbus service. This means systemd can try and launch both, which
won't work.

There's a systemctl mask to stop this in systemd-compat-units, but the logical
place for it to be is in the dbus recipe so it can't be left out of an image.

(From OE-Core rev: c910aa17689077362a25938aeebee7fb24057e30)

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

classes/buildhistory: write out files in UTF-8 format

If the package information contained unicode characters (for example, in
the package file listing) then writing them out as ASCII would fail, so
write them out using codecs.open() instead of open() using UTF-8. This
fixes ca-certificates failing in do_packagedata when buildhistory is
enabled.

(From OE-Core rev: dcf228fe69bfee4e22baad477ad407248c0f9cdb)

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