package.bbclass: Improve package_fixsymlinks

Improve package_fixsymlinks so we don't handle RDEPENDS for every single package
in PACKAGES.

(From OE-Core rev: 20ff8feb95d54e4db646f8c0cb006ce187e288af)

(From OE-Core rev: ca2ee871f82dd0ba4122a8373e4efd21cec5722b)

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

package.bbclass: Tweak PACKAGEFUNCS

Move package_rename_hook call into PACKAGEFUNCS and also move package_get_auto_pr
to a more appropriate execution point, grouping package metadata handling
functions together.

(From OE-Core rev: a3f41cfbc8923e54198d10db292a11ef2edda4d7)

(From OE-Core rev: 96d268166ab0d4640767bf3a0886358dcb663d0f)

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

package.bbclass: Split out package_fixsymlinks function

The symlink handling code doesn't need to being part of populate_packages
and is logically separate so split it out into a separate function,
package_fixsymlinks.

(From OE-Core rev: fc964ed2b558b08327e2bc8c93ada80ca9d97337)

(From OE-Core rev: 08a194c13857b02a2febf9b1484b3cdd67a18160)

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

package.bbclass: Move PKG handling code to emit_pkgdata

We only use the PKG variable in emit_pkgdata so we might as well move the
fallback code there, allowing restructuring of other parts of the metadata.

(From OE-Core rev: 0136ae9a5f719f0e7ba6e00bfd366c0e61b3b3f2)

(From OE-Core rev: df38695f0fbea10289e85fdcb7c2bdf566990577)

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

package.bbclass: Avoid copying the datastore for FILES handling

There is no real point in adjusting overrides and creating a copy of the datastore,
just to access a single variable. We can do this just as easily with a slightly
more complicated getVar call. This improves performance.

(From OE-Core rev: 69f4351809359a0c7c38e8f233f3e3f7680ed2e2)

(From OE-Core rev: b5d65f5d5e5b26d3a2c673e899114c90bcaf6bc8)

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

package.bbclass: Check FILES once, not once per loop iteration

There is no need to check FILES in each loop iteration, we can just check it once
at the start when we read the variable.

(From OE-Core rev: 1bcc5cba12cbb1b846b433719d1b0820f5a97105)

(From OE-Core rev: 40542f7a46bfb8431e31eca9af06adb4b343d810)

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

package.bbclass: Add a shortcut exit from the mkdir function

The mkdir function iterates over strings with many different operations,
even if ultimately the target already exists. This adds a check to the start
of the function so we don't waste time when the target already exists.

(From OE-Core rev: 06e188e673313f1eb9ae7d85ae64467c8d2a94b3)

(From OE-Core rev: 9509627e06ab4cf18c5bbf3f405ecf19a9f40287)

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

package.bbclass: Improve kernel module handling

Currently the kernel module handling consists of several special cases
and has its own path walking. This refactors the code to handle them in
a more standardised way which is also a bit more efficient.

(From OE-Core rev: ad51b54f0afe8c56033137b7cf9ba398877b2651)

(From OE-Core rev: cb24a72e97afb43de5e1e79ff807fd9e184df6a2)

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

package: Rename splitfile functions to something more descriptive

The splitfile and splitfile2 function names are confusing and the comments
are also misleading, hard to understand or plain incorrect. This tries to
improve things.

(From OE-Core rev: 46f3050a1f46f814e2d031e7e71600b2932d631c)

(From OE-Core rev: bc6f7085b42d8137957795e800e8b46d2f1eddac)

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

multilib.conf: Use BASELIB to set baselib, not hardcode the value

(From OE-Core rev: 1a97cab9d4856e8948025ce6f406c76a2732ec36)

(From OE-Core rev: b157ac5590a06405fd5622c7cf7c51c0a2d3cc5e)

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

sqlite3: Ensure MLPREFIX is applied to AUTO_LIBNAME_PKGS

(From OE-Core rev: c09777a99281e22f006b3e8bdc90b469ba12d9fe)

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

QT4: use '-no-neon' if cpu does not support neon

Some armv7a cpu does not support neon, like SPEAr1310,
so we should check against "neon" in TUNE_FEATURES.

(From OE-Core rev: 173d430ab2a06fb86f1c5bb0ca60376dd2fdbdfb)

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

dpkg-native:fix dpkg-scanpackages error at deb image creation time on Fedora 17

Invoking dpkg-scanpackages on Fedora 17 to build deb image meets such an error:

Can't locate Dpkg.pm in @INC  (@INC contains: /usr/local/lib64/perl5 /usr/loc
al/share/perl5 /usr/lib64/perl5/vendor_perl /usr/share/perl5/vendor_perl /usr
/lib64/perl5 /usr/share/perl5.) at tmp/sysroots/x86_64-linux/usr/bin/dpkg-sca
npackages line 27.

1. The dpkg-scanpackages is invoked at deb image creation time and it is built
by dpkg-native. The dpkg-scanpackages is a perl script and import module `Dpkg'.

2. The Dpkg.pm has been installed in perl's staging dir at dpkg-native compiling
time. Such as `tmp/sysroots/x86_64-linux/usr/lib/perl-native/perl/Dpkg.pm'

3. The dpkg-native uses create_wrapper to relocate to perl-native if perl-native
has been built. If perl-native isn't built, it will use the host Dpkg as default.

4. If dpkg-native is built on a new build envionment which means perl-native doesn't
exist before, and the build system doesn't install dpkg (such as Fedora 17), the
dpkg-scanpackages could not work correctly.

5. Modify dpkg-native's depends, change `perl-native-runtime' to `perl-native', Let
perl-native automatically be built before building dpkg-native.

[YOCTO #3817]

(From OE-Core rev: 7d194e72ee12cb6690bf07a8423cb461dd210719)

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

mini-x-session: Fix runtime when using without session contents

The script where behaving badly when the session directory does not
exist and also lacking a dependency to 'sudo'. Instead of adding a
dependency on 'sudo' we use 'su' as done in xserver-common.

(From OE-Core rev: 84b58ed421fd14796ddcbc4cc7fc13a5a098d6cc)

Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

connman-gnome: Add DEPENDS on intltool-native

When 'bitbake -c configure connman-gnome' is executed from scratch, it fails
with syntax error parsing `IT_PROG_INTLTOOL(0.35.0)' in configure.in. This
is caused by its definition is not included into aclocal.m4

Adding intltool-native to DEPENDS is gonna fix it.

[YOCTO #3815]
[ CQID: WIND00401118 ]

(From OE-Core rev: 283dd547b5ed657f871b9330ff4d8d5b3afd1e32)

Signed-off-by: Ming Liu <ming.liu@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libaio: Fix MIPS system call interface

The io_syscallX wrappers in syscall-mips.h discard error return status
by overwriting the value returned in v0 from the system call with -1.

Modify this behavior by returning the negative of the return value on
error (as identified by a3 != 0). This convention is consistent with
the behavior observed in syscall-ppc.h.

For a description of the MIPS system call interface, see:

http://www.linux-mips.org/wiki/Syscall

(From OE-Core rev: 0d8f50c33e5ff71dc17121b5c293718a8ad602bd)

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

multilib: skip packages that provide virtual/kernel

Rather than keying on recipes that inherit kernel.bbclass, we should
be checking for providers of virtual/kernel when skipping kernel
recipes in multlib builds.

Not all providers of virtual/kernel inherit kernel.bbclass (notably
linux-dummy), so checking on the provider is a more complete check.

We need to be sure to check for inheritance of module-base as well, this
allows for packages that provides modules to avoid the multilib renaming.

(From OE-Core rev: dc7d181ab03ceab87a24d932130109003334dbf8)

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

kernel: avoid copying unnecessary files during do_install

kernel_do_install() populates $kerneldir with files needed to build
external modules. To accomplish this there are several copy commands
to get source from the kernel source tree and build trees after which
a 'clean' is performed. Since we are copying from the build tree we
get about 1G of .o and .cmd files copied over only to have them
removed when we clean. This adds additional IO overhead as well as
overhead caused by pseudo. By avoiding copying these files in the
first place we get multiple gains:
  * avoid initial copy
  * avoid file deletes during clean
  * reduce pseudo overhead

Additionally we are making use of cpio vs cp which tends to be
significantly faster at performing copies.

With these changes I observe a 15-30% decrease in the time to complete
the do_install() operation on the kernel.

[YOCTO #3517]

(From OE-Core rev: c753f9d59f4d0a5af4ea5deb6e2b9609e05314e2)

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

mesa-dri: add extra dri drivers

 * By default enable only swrast.  This needs to
   be here or for some reason qemuarm tries to
   detect the intel dri libraries and fails.
 * For x86 and x86-64 explicitly set all of
   the supported drm drivers.

 builds properly with qemux86 qemuarm and qemux86-64

(From OE-Core rev: 3efcfa7e75dc0fe98b75ae1b7be8db5549182ff9)

Signed-off-by: Ian Reinhart Geiser <igeiser@devonit.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

insane.bbclass: removing outdated comment

This commit removes an outdated comment that recorded an issue that has been
resolved in the commits 1c75a5df6845090480deff20f469f177501d2182 and
d2d5456cd3b3bd3e52a5dedccca4d46e3a7986d1. Probably, the comment was missed
when changing the behavior in those commits.

(From OE-Core rev: fb1b51e902186b6a7d05e511fe3c3f03c10e3444)

Signed-off-by: Lukas Bulwahn <lukas.bulwahn@oss.bmw-carit.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

oprofile: avoid processing files under .pc

Fix the below issue:
| DEBUG: Executing shell function do_configure
| sed: can't read ./.pc/opstart.patch/doc/opstop.1.in: Permission denied
| sed: can't read ./.pc/opstart.patch/doc/opstart.1.in: Permission
denied
| sed: can't read ./.pc/opstart.patch/utils/opstart.c: Permission denied
| ERROR: Function failed: do_configure

(From OE-Core rev: d1003ea86543ce5d09c163db95e9e723aa1adc10)

Signed-off-by: Ting Liu <b28495@freescale.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: bitbake-layers: print the recipe's depends that crosses a layer boundary

Figure out the dependency between recipes that crosses a layer boundary

* Introduction:
  - For the .bb file, we need check the following depends:
    > Check the DEPENDS, RDEPENDS and inherits according to the
      cooker_data.
    > The cooker_data doesn't have the info about "require/include xxx",
      so we need check them manually.

  - For the .bbclass, .inc and .conf file, we can't get their file
    depends from the cooker_data, we need check them manually.

* Usage:
  $ bitbake-layer show-cross-depends

  I don't like the name "show-cross-depends", figure-out-cross-depends
  might be better, but most of the commands are "show-xxx" for bitbake-layers.

* Output: (snipped)
/work/poky/meta/recipes-support/libusb/libusb-compat_0.1.4.bb inherits /work/poky/meta-yocto/classes/poky-sanity.bbclass
/work/poky/meta-intel/meta-cedartrail/recipes-cdv-media/images/core-image-cdv-media.bb requires /work/poky/meta/recipes-sato/images/core-image-sato.bb
/work/poky/meta-yocto/recipes-core/tiny-init/tiny-init.bb RDEPENDS /work/poky/meta/recipes-core/busybox/busybox_1.20.2.bb
/work/poky/meta-intel/common/recipes-bsp/amt/lms_7.1.20.bb DEPENDS /work/poky/meta/recipes-devtools/autoconf/autoconf_2.69.bb
/work/poky/meta/classes/distrodata.bbclass includes /work/poky/meta-yocto/conf/distro/include/package_regex.inc
/work/poky/meta-intel/meta-sys940x/conf/machine/sys940x.conf requires /work/poky/meta/conf/machine/include/tune-atom.inc

[YOCTO #3387]

(Bitbake rev: 100502ef722bf19b6c125a5e017f31686ad18421)

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

bitbake: ssh.py: add example SRC_URI

(Bitbake rev: f838af8e7afebf279ffb00a3afa6592f061b703f)

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: ssh.py: throw ParameterError when someone tries ssh://foo; protocol=git

* taken from SFTP fetcher:
  http://patchwork.openembedded.org/patch/43027/

(Bitbake rev: 88e565855b52e905156d85c3f45b341cddfe2f55)

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: ssh: fix fetcher

* set localpath in urldata_init
  otherwise localpath and basename were None, when fetcher was trying to
  define .lock and .done paths
  basepath = d.expand("${DL_DIR}/%s" % os.path.basename(self.localpath or self.basename))
* remove "host" from localpath
  .done and .lock files are always using just basename, so if someone
  has 2 recipes with:
  SRC_URI = "ssh://foo/file.txt"
  SRC_URI = "ssh://bar/file.txt"
  then there will be only one file.txt.done in downloads anyway (and
  only first file.txt from first server will be returned on do_fetch

(Bitbake rev: 41208760d70a657297f9ecfb48b74e2c3b594e70)

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: hob: eliminate the "by recipe" grouping in the packages table

We have agreed to change how we show the packages. Now they are not grouped by
recipe. Until now, it was implemented using  TreeStore and I've changed it to
extend a ListStore. I have modified all the function in according to this.

[YOCTO #2180]
(Bitbake rev: 311e7ba8b2c88fbf6f16b6ffb1400226b155ddd4)

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

bitbake: hob: Use a GtkTreeView to present and configure the sstate mirrors

After some discussions about the design of the Shared State Settings tab in Hob,
we have decided that using a GtkTreeView to present and configure the sstate
mirrors is probably the best solution.

[YOCTO #3569]
(Bitbake rev: 2ff636910b9cf3099e8d961f4bbe71512c015ecc)

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

bitbake: tinfoil: support other fds, enable color support

Rather than only handling sys.stdout, also support any arbitrary file object,
and enable color for the formatter if that file is a tty.

(Bitbake rev: c46db4be4cc4dc53376ed3f574b2f1c868730f2a)

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

bitbake: hob: small fixes to parsing warnings dialog

Set the labels selectable; reset the warnings list when another machine is selected;
changed a label name

[YOCTO #3215]
(Bitbake rev: a4463d7b51828c32e55dea3c0dd51966d387abac)

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

bitbake: bitbake & hob: implement functions to assure consistency for configuration files

Added a new command in bitbake to save a variable in a file; added a function
in cooker which is called by this command.

Added new command in bitbake to enable/disable data tracking.

The function saveConfigurationVar from cooker.py saves a variable in the file that
is received by argument. It checks all the operations made on that variable, using the history.
If it's the first time when it does some changes on a variable,it comments the lines where
an operation is made on it, and it sets it in a line to the end of file. If it's not
the first time(it has a comment before), it replaces the line.

Made some changes in hob to save the variables from bblayers.conf and local.conf
using the bitbake command.

[YOCTO #2934]
(Bitbake rev: 55b814ccfa413d461d12956896364ab63eed70a8)

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

harfbuzz: fix install of version.h

version.h is listed both in pkginclude_HEADERS (via $HBHEADERS) and
nodist_pkginclude_HEADERS. This double listing is likely cause of
the make install error:
/usr/bin/install: cannot create regular file `.../harfbuzz/0.9.10-r0/image/usr/include/harfbuzz/hb-version.h': File exists
Just remove the nodist_pkginclude_HEADERS entry. We're not creating
the tarball, and listing version.h here wouldn't prevent the other
listing from including it to tarball anyway.

(From OE-Core rev: 1489a69cecb1dcc3502a4c24beaea81e6ca6dacc)

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>

rpm: Ensure native binaries are correctly wrapped

(From OE-Core rev: 1cdeff2c50a13c7238543ee1e4e1eb60753120e2)

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

package_ipk/deb/rpm: Ensure deploy staging directory is empty before rerunning task

If we don't do this, stale files can build up, particularly with the PR
server.

(From OE-Core rev: c30ae39229b35d72e2205040d76754b5120fa878)

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

directfb: Explictly disable mesa

Without this, directfb might build with mesa enabled if present.

(From OE-Core rev: 1531adb8d2230646e03f31ea30faa916d846ccfc)

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

insane.bbclass: Fix incorrect getVar call

On a trace I was a bit puzzled why getVar was making 180 calls to len(d).
This is an expensive operation that should be very rarely called and
certainly not by getVar. In perl's do_package it was resulting in
~1.5 million function calls from those 180 cases.

Ultimately this typo was why. Lets fix it and save the CPU cyles.

(From OE-Core rev: a8ba821d0002e4395fc5c80649fe14f93a7971fe)

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

sstate: Move debug comment to more logical place

The same log message gets output multiple times in the log which look
confusing and is rather pointless. Move the log message to the correct
level.

(From OE-Core rev: 3917409004a830e7ad0646f05ad7421385cbd1de)

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

bitbake: cache.py: Drop support for BROKEN variable

All it now does is function in a similar way to EXCLUDE_FROM_WORLD and
since we have a better named variable for this, lets just drop the
usage of BROKEN at the bitbake level.

(Bitbake rev: 8f4dcf794545273417d78ba18f51aa2b81606ae4)

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

alsa-tools: upgrade to 1.0.26.1

Removed the following tools:
- all related to hdsp (required gtk+ and fltk-config)
- ld10k1, qlo10k1 (required QT)
- hdajackretask

Fixed the automake issue for cross-compilation

(From OE-Core rev: 9a148e6b100fe8bc3e162d79630552df5eb78fc0)

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>

module.bbclass: Don't add pkg_postinst/pkg_prerm to all packages in recipe

The code in module.bbclass was appending the pkg_postinst and
pkg_prerm to all packages that are part of a given recipe, meaning
that the -lic, -dev, -doc, ... packages all got the scriptlet
This change uses  only which macthes with the RDEPENDS and FILES
already used in module.bbclass.

The failure was that rootfs creation would fail due to the -lic package
being installed before the kernel and the script would fail.

[YOCTO #3803]

(From OE-Core rev: cf05c4578c99c0cb885cf2706f7f2b39b100aeb8)

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

rootfs_rpm: Escape the backtick to ensure the ls runs on the target

This change ensures that the ls /etc/rpm-postinsts runs in the target
at first boot time, rather than at the creation time of the script on
the host.

This was causing the following error in the rootfs log:

+ install -d /srv/ssd/sgw/machines/fri2/tmp/work/fri2-poky-linux/core-image-minimal/1.0-r0/rootfs//etc/rcS.d
+ i=
+ ls /etc/rpm-postinsts/
ls: cannot access /etc/rpm-postinsts/: No such file or directory

(From OE-Core rev: e893cf0b3843701e80f5c9f47be04d1a88e5ed68)

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

bitbake: build.py: Dump out performance data of individual tasks

(Bitbake rev: 32aa49519e4f015e3c21466a7e5dc939f6369851)

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

bitbake: utils.py: Add function for processing profile output

(Bitbake rev: 0df64810e8d40e7761cfd5059c0617dda31a6641)

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

mtools : upgrade to 4.0.18

(From OE-Core rev: e11fdc145255c8b97c6c7ac6c5f7ce8aa50bdf0d)

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

openssh : upgrade to 6.1p1

(From OE-Core rev: 5534deb69b0a2835fbbf149a00d1f6ba61cc8160)

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

migrate_localcount.bbclass: use PRAUTOINX instead of PN & PV

- we need to use PRAUTOINX (it covers all cases)
- it addresses bumped PE issues

[YOCTO #3071]

(From OE-Core rev: ec81b985bb29a8562366f5343171e68a1f9cc4e9)

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

libcgroup: Update to 0.38

(From OE-Core rev: daf1298f6c34ce7efe4084ebbb1015b55c248383)

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

resolvconf: Update to 1.69

(From OE-Core rev: 990ac9be82b1657293a6bef89fd55bfdce3ad802)

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

util-linux: Update to 2.22.2

(From OE-Core rev: b4ec72a5efe7a22a00aa3ced7eeb2213233b848c)

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

dhcp: Update to 4.2.5

(From OE-Core rev: fb1a7cf8ac1b5ab8047f47ace28775685ba28804)

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

file: Update to 5.12

Remove the backported automake patch, now fixed upstream

(From OE-Core rev: a7d06efe4856ea76b1ffa15ed398099cc963d14a)

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

cracklib: Update to 2.8.22

(From OE-Core rev: e21c57c204024be06ac5ac0c56a065b271b9ba5c)

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

libidn: Update to 1.26

(From OE-Core rev: ba9bf2b01f8166860e724b416f413153de14a750)

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

man-pages: Update to 3.45

(From OE-Core rev: 61efbc44d5d49a56323cf2c018713b0cd7204eae)

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

mc: Update to 4.8.7

(From OE-Core rev: 5a3b6f5213e6fced8c6162d4fae40ed0bc8f3f65)

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

sysstat: Update to 10.1.3

(From OE-Core rev: a9a39d944046d29e4870b4449921095c797b55fa)

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

rgb: Update to 1.0.5

Remove the backported automake patch, now fixed upstream

(From OE-Core rev: 2356ce3b58b44551d51b3de0eddb81e32e907c45)

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

libnl: Update to 3.2.19

(From OE-Core rev: 7ccb2e6463a745e90aa40c7e4163f9ef3495f866)

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

systemd: Upgrade to 197

LGPL-2.1 licensce is updated to reflect the new FSF address
Drop patches to disable argparse and dbus-python

analyser has been implemented in C lets use that patch
and get rid of the python version

disable hostname detection for uclibc bases systems
since uclibc lacks NSS

(From OE-Core rev: af17e816dd4ee0c2a7401f26a148129ad8f8e1fa)

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Conflicts:
meta/recipes-core/systemd/systemd_197.bb
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

icu: Fix build on uclibc

_timezone is not implemented in uclibc therefore
bypass it.

(From OE-Core rev: 8fb1e3b74b01095cbc626497f1a7c71145343af3)

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>

quota: inherit gettext

Othereise it does not honor USE_NLS variable
and does not have proper dependencies. Which leads
to compile errors like missing liintl.h on uclibc

(From OE-Core rev: 606d1245cfa23bd51240af5112dc2e5257efbed1)

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>

eglibc: Backport upstream 2.17 patches

These two patches has been applied to eglibc 2.17 branch
since we created our snapshot.

(From OE-Core rev: 4fdec58c979bb9715a8ab92344ad05110aa57cc6)

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>

eglibc_2.17: Fix Tibetian locale generation error

This is fallout of BZ #13950 fix to glibc

(From OE-Core rev: 4b39d90c84b1c1577b6744e7117502069c5bf719)

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>

eglibc: Upgrade recipes 2.16 -> 2.17

Drop patches that are applied upstream

Fix the license checksums for changes in LICENSES file
the new changes add more copyright notices that were missing earlier

Moving ports is no longer needed since ports is now part of libc proper

Refresh tzselect-sh.patch to accomodate upstream changes

C++ headers discovery relative to target sysroot is fixed differently
upstream hence we drop use-sysroot-cxx-headers.patch

aarch64 support is already available in 2.17 hence drop the local
patches

(From OE-Core rev: 83b6fe6d91b924be5a7676e6ee973ce26b5eefc5)

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>

package.bbclass: remove RRECOMMENDS on base packages for locale packages

As long as we support only one locale package per recipe this leads to unwanted
packages in images. This patch was motivated by samba (meta-oe): only wanting
libsmbclient one gets full samba server into the image. For my standard image
the follwing package were removed by this patch:

-acl_2.2.51-r3_armv7a-vfp-neon.ipk
-alsa-states_0.2.0-r3_armv7a-vfp-neon.ipk
-alsa-utils_1.0.25-r3_armv7a-vfp-neon.ipk
-alsa-utils-aconnect_1.0.25-r3_armv7a-vfp-neon.ipk
-alsa-utils-alsaconf_1.0.25-r3_armv7a-vfp-neon.ipk
-alsa-utils-alsactl_1.0.25-r3_armv7a-vfp-neon.ipk
-alsa-utils-alsaloop_1.0.25-r3_armv7a-vfp-neon.ipk
 alsa-utils-alsamixer_1.0.25-r3_armv7a-vfp-neon.ipk
-alsa-utils-alsaucm_1.0.25-r3_armv7a-vfp-neon.ipk
-alsa-utils-amixer_1.0.25-r3_armv7a-vfp-neon.ipk
-alsa-utils-aplay_1.0.25-r3_armv7a-vfp-neon.ipk
-alsa-utils-aseqdump_1.0.25-r3_armv7a-vfp-neon.ipk
-alsa-utils-aseqnet_1.0.25-r3_armv7a-vfp-neon.ipk
-alsa-utils-iecset_1.0.25-r3_armv7a-vfp-neon.ipk
-alsa-utils-midi_1.0.25-r3_armv7a-vfp-neon.ipk
-alsa-utils-speakertest_1.0.25-r3_armv7a-vfp-neon.ipk
-attr_2.4.46-r4_armv7a-vfp-neon.ipk
-avahi_0.6.31-r7.1_armv7a-vfp-neon.ipk
-gtk-engines_2.20.2-r3_armv7a-vfp-neon.ipk
-pulseaudio_2.1-r15_armv7a-vfp-neon.ipk
-samba_3.6.8-r6_armv7a-vfp-neon.ipk
-vte_0.28.2-r5_armv7a-vfp-neon.ipk
-xz_5.1.2alpha-r0_armv7a-vfp-neon.ipk

Strange effect: For samba buildhistory's depends.dot says:
-samba -> libc6;
-samba -> libgcc1;
-samba -> libreadline6;
-samba -> libtalloc2;
-samba -> libtdb1;
-samba -> libwbclient0;
-samba -> libz1;
-samba -> update_rc_d;
-samba_locale_de -> samba [style=dotted];
 samba_locale_de -> virtual_locale_de [style=dotted];
-samba_locale_fr -> samba [style=dotted];
-samba_locale_fr -> samba [style=dotted];

This is conflicts to buildhistory's result on samba package not mentioning
any changes for samba package.

(From OE-Core rev: 50354ebece43a6c13f9ca592e7d230765a26fdf8)

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

buildhistory.bbclass: track also complementary package information

(From OE-Core rev: ffdb003590d5c1d49d6c6a6710adf71397e4fb9b)

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

libcgroup: fix the QA issue for pam_cgroup.so*

PAM modules in ${base_libdir}/security/ should be binary .so files,
not symlinks, so fix this. Since pam_cgroup.so is installed into
${base_libdir}/security, move libcgroup.so.* to ${base_libdir} to
avoid "unsafe-references-in-binaries" QA issue.

(From OE-Core rev: 35567eed06ebd12f7c8ee0a04b6cb28530cf85d7)

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

sysvinit & initscripts: use update-rc.d

Update-rc.d provides a cleaner interface for creating links for init scripts.
So we use update-rc.d to replace the redundant 'ln -sf xxx' statements.

[YOCTO #3708]

(From OE-Core rev: 66f3e3c15faedc9ee78532b4c59fa5d7148ddace)

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

valgrind: update to 3.8.1

Changed license copyright notice 2011->2012
remove unnecessary patches
built for x86, 86-64, ppc

(From OE-Core rev: df0d23c2bedafd534c5909a65afaf3373d7bc33e)

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

libaio: add aarch64 support

Picking up a patch from gentoo and adding aarch64 defines
is enough to fix libaio and pass the harness testsuite

(From OE-Core rev: 7255c43b6e545a4c15c3cd57f6c240668a77786a)

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

mtd-utils: split into multiple packages

This allows one not to install rarely used or fs specific files
As discussed in http://thread.gmane.org/gmane.comp.handhelds.openembedded.core/31744/focus=31804

(From OE-Core rev: ca9d10543f22b60ba32fd78130970591782b880c)

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

alsa-utils: upgrade to 1.0.26

ncursesfix.patch is no longer necessary,
build system fixed upstream.
ncursesfix.patch removed.

(From OE-Core rev: 1dccd16e29186e7baad6ebcf189b7ea2463c2275)

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>

xorg-minimal-fonts: depend on font-alias instead of shipping fonts.alias

| Collected errors:
|  * check_data_file_clashes: Package font-alias wants to install file /home/hrw/HDD/devel/canonical/aarch64/openembedded/build/tmp-eglibc/work/genericarmv8-oe-linux/core-image-lsb-dev/1.0-r0/rootfs/usr/share/fonts/X11/misc/fonts.alias
|       But that file is already provided by package  * xorg-minimal-fonts
|  * opkg_install_cmd: Cannot install package packagegroup-core-lsb-dev.
| ERROR: Function failed: do_rootfs (see /home/hrw/HDD/devel/canonical/aarch64/openembedded/build/tmp-eglibc/work/genericarmv8-oe-linux/core-image-lsb-dev/1.0-r0/temp/log.do_rootfs.7564 for further information)

(From OE-Core rev: e3d5cb2033245ecb6e25f90d9d71296a018c4342)

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

gtk-immodules-cache.bbclass:fix support postrm at image creation time

Let postrm use the same logic as the postinst to run on both build machine
and target

[YOCTO #3633]

(From OE-Core rev: dc1bb7a5532411bc9cb7e8678bc540c44fd2cd63)

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>

gtk-icon-cache.bbclass:fix support postrm at image creation time

When use postrm on the build machine, it installs the hook in intercept-scripts
directory and exit 0, the hook will be later invoked and it will properly call
gtk-update-icon-cache.

[YOCTO #3633]

(From OE-Core rev: 6ae0b8339134300d0c179bd47dc3062e0e1f2f0b)

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>

license.bbclass: package license texts to PN-lic when LICENSE_CREATE_PACKAGE is enabled

[YOCTO #3743]

(From OE-Core rev: dc78ef91a2bf01efb8028c9afbe69e506e016265)

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

license.bbclass: extract functionality to find license files to separate function

* move it from do_populate_lic to find_license_files so we can reuse it
  to populate license in package itself

[YOCTO #3743]

(From OE-Core rev: 833f8c239aa475b3e0cacbd448a90079ac287468)

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

gettext: Fix build with uclibc remove upstreamed patches

uclibc specific patches are not needed now as they are
available in 0.18.2 gettext now

(From OE-Core rev: ab692ed618f83df2ce452769165314e70b867345)

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

eglibc: fix evacuate_scripts for external toolchains

Not every external toolchain has mtrace/sotruss/xtrace scripts so check
their existance first.

(From OE-Core rev: 06ea50973a4446053d1ecac5b8e4cf425c8df5b1)

Signed-off-by: Marcin Juszkiewicz <marcin.juszkiewicz@linaro.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

eglibc-scripts: don't package mtrace, RDEPEND on eglibc-mtrace instead, fix RDEPENDS

* eglibc-mtrace and eglibc-scripts provide the same mtrace script
* remove mtrace from eglibc-scripts and replace it with RDEPENDS
* mtrace also needs perl (eglibc-mtrace has correct RDEPENDS), bash is
  for sotruss and xtrace
* remove invalid RDEPENDS_ldd, ldd is now using /bin/sh and is not
  packaged by eglibc-scripts
* fixes error in world-image build:
  * check_data_file_clashes: Package eglibc-mtrace wants to install file
    rootfs/usr/bin/mtrace
    But that file is already provided by package  * eglibc-scripts

(From OE-Core rev: a3599db48576043d7c335aff677471fc1afed8ab)

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

eglibc-package: remove infodir/dir if it exists

* fixes error in world-image build:
 * check_data_file_clashes: Package eglibc-doc wants to install file
   rootfs/usr/share/info/dir
   But that file is already provided by package  * gcc-doc

(From OE-Core rev: f6e0875e35b9149f38f3bbfab6bd7a26048dd689)

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gcc-configure-runtime: always remove info/dir

* move libgomp.info removal also outside that for cycle, can be removed outside if exists
* also use infodir variable instead of ${datadir}/info

(From OE-Core rev: 0b8fe539928c7a61ca01a4a0a0f54df47164ac6c)

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

automake: use autotools_do_install

* autotools_do_install removes /info/dir when it exists, overwritting whole
  do_install prevented that, use do_install_append instead.

(From OE-Core rev: 5c2efbaad15a8430b4a02f3fcffd92578e1bb1a3)

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

valgrind: explicitly disable MPI2 support

We don't have support for this in OE-Core, so ensure we don't pick this
up from the build host if e.g. openmpi development files happen to be
installed there.

Fixes [YOCTO #3726].

(From OE-Core rev: cc490d76aba0a778409ca1a3d0e1f2c308684c9b)

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

gawk: Add missing dependency on readline

building from shared state often ends up with

|
/b/kraj/angstrom/build/tmp-angstrom_next-uclibc/work/i586-angstrom-linux-uclibc/gawk/4.0.1-r1/gawk-4
.0.1/debug.c:4135: undefined reference to `rl_reset_terminal'
|
/b/kraj/angstrom/build/tmp-angstrom_next-uclibc/work/i586-angstrom-linux-uclibc/gawk/4.0.1-r1/gawk-4
.0.1/debug.c:4138: undefined reference to `rl_get_screen_size'
| debug.o: In function `unserialize':

Adding the missing readline dependency to fix such cases
where readline has not been populated in sysroot via some
other dependency and gawk would then use it or bail out.

(From OE-Core rev: 9a43a67993c21cedf6cf71138e3da9c691ebf4a7)

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

libpcap: Turn canusb support into PACKAGECONFIG

libpcap notices libusb in its configure if its sysrooted
and ignores it if its not. But this causes problems since
there is no explicit dependency on libusb. Rebuild from shared
state sometimes fails when libusb has not been yet populated
in sysroot and libpcap has. And do_configure of consumers of libpcap e.g. tcpdump
are probing for features on libpcap. All the tests fail since
none of link steps succeed and it sees as if libpcap does
not have required features and configure of tcpdump bails out.

turn canusb into PACKAGECONFIG and let it be disabled
by default.

Backport additional patches needed for disable-canusb to work

(From OE-Core rev: a7cfc0ce9885c51ccefcaba001e6672d7b3e36e5)

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

libtirpc: Disable DES functions on uclibc

uclibc does not provide these functions so dont expect them

/usr/lib/libtirpc.so: undefined reference to `key_encryptsession_pk'
/usr/lib/libtirpc.so: undefined reference to `getnetname'
/usr/lib/libtirpc.so: undefined reference to `cbc_crypt'
/usr/lib/libtirpc.so: undefined reference to `ecb_crypt'
/usr/lib/libtirpc.so: undefined reference to `getpublickey'
/usr/lib/libtirpc.so: undefined reference to `key_gendes'

(From OE-Core rev: c66a269f6f68d670ca3bf9e7b22a5dce4bb95cd0)

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

tcmode-default.inc: Switch to git version of uclibc by default

systemd related functionality is tested in latest git of uclibc
therefore lets use it as default provider for uclibc as its the
most tested version on master

(From OE-Core rev: db93f49c676f84d6d5ad54a9f1ed9be7ba6d5364)

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

uclibc-git/uClibc.distro: Enable UCLIBC_HAS_REENTRANT_RPC

The renterant functions are needed by quota

(From OE-Core rev: 406aaccfd115ef189e4372f581bf19e5b0906a15)

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

wpa-supplicant: upgrade to 2.0

- as of February 11, 2012, the project has chosen to use only
the BSD license option for future distribution
- wpa-supplicant-1.0 dir was not version specific,
as such it is generic now

(From OE-Core rev: 0f83cb8ca629fbf1af448781f28912bbc75e0d80)

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

systemd: remove usbutils dependancy

V2:

Dependancy was removed in version 196 according to following systemd commit.

http://cgit.freedesktop.org/systemd/systemd/commit/configure.ac?id=796b06c21b62d13c9021e2fbd9c58a5c6edb2764

V1:

No where in the systemd readme does it specify that usbutils
is a dependancy of systemd. I can only guess that it was added
due to an issue elsewhere in the chain or it was brought in
when udev was merged and the dependancies never checked.

I have build tested this with no issues.

(From OE-Core rev: 251627cdfc5a9aa69e469f3c9072264fd0bda73b)

Signed-off-by: Jack Mitchell <jack.mitchell@dbbroadcast.co.uk>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

yocto-bsp: qualify user files with machine name

The bblayer abstraction makes it where multiple layers can be
configured and used at the same time.  Some layers make changes to
support a specific machine, and should not have any affect when other
machines are in use.

For linux-yocto, all bsps are created with a user-config.cfg and
user-config.cfg and user-patches.scc.  This means that those files
will be pulled from the first location found, which might correspond
to files customized for a different machine.

Instead of using the names user-config.cfg and user-patches.scc, I
propose a machine specific name be used such as
{{=machine}}user-patches.scc and {{=machine}}user-config.cfg.  This
would necessitate that all references changed to these new names,
which would affect the yocto-bsp and yocto-kernel scripts.

With this change, it would be possible to have multiple machine BSPs
searched at the same time and to select which to build against by
using a command like MACHINE=qmeux86 bitbake core-image-sato to
override the default.

Note many of the standard BSPs do not seem to suffer this problem as
they do not use the common files user-config.cfg and user-patches.scc
that the yocto-* scripts depend upon.

Additions by Tom Zanussi:
 - renamed user-config.cfg to {{=machine}}-user-config.cfg everywhere
 - renamed user-patches.scc to {{=machine}}-user-patches.scc everywhere
 - added the user-config/patches SRC_URI items to the qemu -rt kernel recipes
 - fixed conflicts due to the new open_user_file() helper function
 - updated user filename conflicts caused by directory renaming
 - updated custom kernel files to match

Fixes [YOCTO #3731]

(From meta-yocto rev: c20bef60aa8d52971fb061d4b8d473ad19c03180)

Signed-off-by: Brian A. Lloyd <brian.lloyd@familyhonor.net>
Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

prservice.py: fix NameError: global name 'host' is not defined

* http://git.openembedded.org/openembedded-core/commit/?id=e00f49de8b1f79c3e07b887d257bd75a46052fa0
  removed host and port variables

(From OE-Core rev: 494439c95d33b2ad457cc205413680b0bf176560)

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: bitbake-diffigs: Don't pull in tinfoil unless we really need/plan to use it

This means we can run in parallel with other bitbake processed we're not in
task mode.

(Bitbake rev: 20222f2f16130e91eed5e68286188ee0e8f8f3bf)

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

staging: Drop obsolete comments

(From OE-Core rev: 44a2374dcfd68c8e158e0c901d75dbd7327652a5)

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

git: remove unpackaged perl-native files

${libdir} is not applicable for the install path of perl-native files,
files are always installed to /usr/lib no matter the target is 32/64
bits. After installing, remove unpackaged and unneeded perl-native
files to prevent warnings.

Fix warning:
    WARNING: For recipe git, the following files/directories were \
        installed but not shipped in any package:
    ...
    WARNING:   /usr/lib/perl-native/perl/5.14.2/Git.pm
    WARNING:   /usr/lib/perl-native/perl/5.14.2/perllocal.pod
    WARNING:   /usr/lib/perl-native/perl/5.14.2/Error.pm
    WARNING:   /usr/lib/perl-native/perl/5.14.2/auto

[YOCTO#3780]

(From OE-Core rev: cc6b8261fa47a049e501882e9bfc40f61e603b6f)

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

puzzles: upgrade to r9751

(From OE-Core rev: cf97241a18284180ff3b8dd880f219d09155bf09)

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

cairo: upgrade to 1.12.10

(From OE-Core rev: 5910cf254635545040df62d9db898abcf17e00d4)

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

perl: update to 5.14.3

There is a securty issue:
http://web.nvd.nist.gov/view/vuln/detail?vulnId=CVE-2012-5195
Update perl to 5.14.3 to resolve this problem.

Patches hurd-ccflags.diff, h2ph-multiarch.diff, index-tainting.diff and
hurd-hints.diff have been merged, so remove them from SRC_URI.
Update patches config.sh and Makefile.SH.patch with new PV.

[Yocto 3701]

(From OE-Core rev: b1fd25e05308cabb56afe1d4276470bf7380ea59)

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

perl-native: update to 5.14.3:

There is a securty issue:
http://web.nvd.nist.gov/view/vuln/detail?vulnId=CVE-2012-5195
Update perl to 5.14.3 could resolve this problem. And update perl-native
first.

(From OE-Core rev: 360401af6e7729a373d0a6d13995714aff121064)

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

meta: remove all mention of PCMCIA_MANAGER

Nothing appears to use this anymore, and it's been a very long time since there
was anyone expressing an interest in the alternatives.

(From OE-Core rev: f6f289c13b9da9c2793d1fd30456216db8afad64)

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