recipes.txt: Add entries for "recipes-rt" and "recipes-support"

(From OE-Core rev: 9705ccc10892b15fa3269ed071bb41d45987694c)

Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

pixman: upgrade to 0.26.2

Also, disable Loongson MMI for now because Pixman's build system will
autodetect a Loongson2f architecture revision and the build will fail.

(From OE-Core rev: 7dded4da52bc73677583da740620b0c81e594d62)

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

xinput: upgrade to 1.6.0

Licence file changed, an entire part has been removed.

See diff below:

<<
 TORTIOUS  ACTION, ARISING    OUT OF OR   IN  CONNECTION  WITH THE USE OR
 PERFORMANCE OF THIS SOFTWARE.

-
-Copyright 2007 Peter Hutterer
-
-Permission to use, copy, modify, distribute, and sell this software and its
-documentation for any purpose is hereby granted without fee, provided that
-the above copyright notice appear in all copies and that both that
-copyright notice and this permission notice appear in supporting
-documentation.
-
-The above copyright notice and this permission notice shall be included
-in all copies or substantial portions of the Software.
-
-THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
-OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
-MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
-IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY CLAIM, DAMAGES OR
-OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
-ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
-OTHER DEALINGS IN THE SOFTWARE.
-
-Except as contained in this notice, the name of the author shall
-not be used in advertising or otherwise to promote the sale, use or
-other dealings in this Software without prior written authorization
-from the author.
-
-
+Copyright © 2007 Peter Hutterer
Copyright © 2009 Red Hat, Inc.

Permission is hereby granted, free of charge, to any person obtaining a
>>

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

libxi: upgrade to 1.6.1

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

libx11-diet: upgrade to 1.5.0

The X18NCMSstubs.diff was modified to please the -Werror=return-type
compiler option. Otherwise, it will throw an error.

(From OE-Core rev: 708785266d1d671f5d2a8e8af81a77e34c70662f)

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

libx11-trim: upgrade to 1.5.0

(From OE-Core rev: 817742fa1a5261b3b2c697de3c14e602d55713e2)

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

libx11: upgrade to 1.5.0

Removed one backported patch.

(From OE-Core rev: 52d4ff7b35602faa52360a3e7a800891ef60c739)

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

directfb-examples: upgrade to 1.6.0

(From OE-Core rev: 37a1e3c0ed5347c77ace9f3ce262f34d9ac53528)

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

directfb: upgrade to 1.6.1

Removed two backported patches that already exist in the current
version and fixed a compilation issue when zlib is used.

(From OE-Core rev: 47800f9bff7a799489851156c341151ee9cf364f)

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

multilib: Enable multilib remapping for SDK generation

Enable the remapping for SDK generation, this is required to be able to
create an SDK that targets an alternative multilib.  Note, this work does
not finish SDK/multilib support, but it is one more step toward making
it work properly.

(From OE-Core rev: d66d2bdf6214939e8b131f47db2f35d3c64e7dd2)

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

package_rpm: Avoid duplicate package generation failures

When constructing a multilib based image, it's possible to end up with
the same package listed in multiple steps of the install.  During resolution
we use the --replacepkgs option to avoid errors, as the resulting install
solution will be as expected.  (We do not enable the replacepkgs option
for the final install step, only the generation of the install solution.)

(From OE-Core rev: 68ea7d827710ae7c5cf402887d4c12319038e064)

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

populate_sdk_rpm: Sync multilib configuration with rootfs_rpm

An anonymous python chunk configures some multilib settings for both
populate_sdk_rpm and rootfs_rpm.  The two classes should contain identical
versions to ensure that the generated multilib configuration is the same
for the SDK and the rootfs.

(From OE-Core rev: 0f60722ef460fa826e4f02e7457004e64ca41d76)

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

rpm: Fix PACKAGECONFIG dependencies

neon requires expat to be built first

lua similarly requires expat if enabled

selinux package name should be "libselinux" (from meta-selinux).

(From OE-Core rev: 6b3c9802b659ca1d3b7638ff80e6492cc0e16d6f)

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

gnutls: make sure native is patched for gettext version

(From OE-Core rev: 92ede8a994b9a24387d3d7dc79e80705f6be4524)

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

lsbinitscripts: increase ALTERNATIVE_PRIORITY

[YOCTO #2133]

Increase ALTERNATIVE_PRIORITY of file functions to be higher than
the value in package initscript. Then when lsbinitscripts installed,
file functions provided by lsbinitscripts will be used.

(From OE-Core rev: 33f9abed7658bd5eae39e987b210651d202e1f4e)

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

initscripts: use update-alternative to handle file functions

lsb need a more abundant /etc/init.d/functions file to handle lsb test.
Use update-alternative to install file funtions in package initscripts,
when package lsbinitscripts installed the file functions of
lsbinitscripts will be used.

(From OE-Core rev: 478f821462665a9dc6f1755059cdbf642aebbf4d)

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

matchbox-panel-2: bump PR because libxcb-util soname was changed

(From OE-Core rev: 31b8fe8d1a9921e4efc64b6ab00c26e4bd09ed2c)

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

startup-notification: bump PR because libxcb-util soname was changed

(From OE-Core rev: 1dc4100dd35c050b38e4733cb98752c1323ef7b1)

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

bitbake: usermanual: Fix missing markup

(Bitbake rev: 68af2b09315ba35eae24933a599014a662789c2c)

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

bootimg: Use STAGING_KERNEL_DIR

bootimg.bbclass using STAGING_DIR_HOST/kernel instead of
STAGING_KERNEL_DIR, resulting in build failure of live images.

| install: cannot stat `/usr/local/dev/yocto/fishriver-test/build/tmp/sysroots/fishriver/kernel/bzImage': No such file or directory

Replace it with STAGING_KERNEL_DIR.

(From OE-Core rev: 8f16811a8d51982a8b3d70e6087aef4a41926840)

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

perf: remove unconditional python-ext install

This shouldn't be unconditional - a later patch made it so, but that's
not yet pulled in.  In the meantime, to fix build failures remove the
unconditional install.

(From OE-Core rev: dcc43c34bfa9304233ecfd3f518a84ddfdc2fd90)

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

dpkg 1.15.8.7: bad interpreter when long tmpdir

When use PACKAGE_CLASSES = "package_deb", and the length of tmpdir is
longer than 177, the error occurs during the image generation:

/very/long/path/totmp/sysroots/x86_64-linux/usr/bin/dpkg-scanpackages:
bad interpreter: No such file or directory

The interpreter is perl, and it does exist, this is because the first
line:

  #!/very/long/path/to/perl

in the script can't be very long.

Create a wrapper for it would fix the problem.

[YOCTO #2640]

(From OE-Core rev: 7d3281be18567207de8ce2cabce03957bf2cbca8)

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

netbase: update to ver. 5.0

(From OE-Core rev: 5beb7176d361d4bfa279bbfab5a17b8fe4fa3b73)

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

bluez4: update to ver. 4.101

(From OE-Core rev: 1b2f6a21dde6045c4604ca5463e6c4db3499ee06)

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

ofono: update to ver. 1.8

(From OE-Core rev: 80f05ff2afed7ef24d17209d39a17cf8139855bb)

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

gst-plugin-bluetooth: update to ver. 4.101

(From OE-Core rev: 5d0e81b370a4c6b0473efd7989b6881a16e02f29)

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

telepathy-glib: update to ver. 0.19.2

(From OE-Core rev: 2cc263bf5402e21429a5057c183a10918491da48)

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

u-boot: make FILESDIR a shared setting via FILESPATH

The setting is the same in all recipes, so move it to
the shared settings in u-boot.inc

Since FILESDIR is also being phased out, use the FILESPATH
setting as suggested by Richard Purdie.

Cc: Richard Purdie <richard.purdie@linuxfoundation.org>
(From OE-Core rev: 1b78904b115b50172cf82948d8a68dc3a98005ce)

Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

linux-firware: bump PE because git SRC_URI was changed resulting in LOCALCOUNT reset

(From OE-Core rev: 088b59e7888ae670710be19c868d0fe90ca44b78)

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

bitbake: data_smart: Fix multiple override interaction with append and prepend operators

Variables which used multiple overrides and the append/prepend operators were
not functioning correctly. This change fixes that.

This fixes the testcase:

OVERRIDES = "linux:x86"
TESTVAR = "original"
TESTVAR_append_x86 = " x86"
TESTVAR_append_x86_linux = " x86+linux"
TESTVAR_append_linux_x86 = " linux+x86"

[YOCTO #2672]

(Bitbake rev: dc35a2e506e15fb7ddbf74c3b3280e9e83ab33bb)

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

lib/oe/sstatesig.py: Exclude subversion-native from sstate checksums

If we don't do this, target and cross recipes end up with different sstate
checksums for shared work directory tasks which is bad in the case of gcc.
It leads to multiple fetch/unpack tasks against the shared directory
which ends up with build failures/races.

(From OE-Core rev: 793ce6cd9aa632e0f13789c8293770a86085d28d)

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

gcc: fix collect2 host contamination problem properly

We added the autoconf cache line a while back to ensure that configure doesn't
poke into some hardcoded host paths looking for things it shouldn't. Applying
it as part of do_configure wasn't getting it to the do_compile tasks
where much of the configure scripts are run by gcc. This changes it to a simple
export to ensure it reaches the places it needs to and truly gets rid of
the cross compile badness messages from the logs.

(From OE-Core rev: af58d5c3a5c5e632aa20b1d007d47dbbfb4ed6e6)

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

bitbake: runqueue.py: Fix recursive task pruning to only prune self referencing tasks

(Bitbake rev: 4962a59793504b26b06cf058dda600a07fbbd951)

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

prexport.bbclass: Fix typo in warning: "AUROPR" -> "AUTOPR".

(From OE-Core rev: 0b79c730f3e5c3c5329e439dffd137f2605b6cba)

Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lsbinitiscripts: Add glib-2.0 to DEPENDS list

ppp-watch requires it.

x86_64-poky-linux-gcc    -m64 --sysroot=/intel/poky2/builds/fetch/tmp/sysroots/qemux86-64 -O2 -pipe -g -feliminate-unused-debug-types  -Wall -D_GNU_SOURCE `pkg-config glib-2.0 --cflags` -c ppp-watch.c -o ppp-watch.o
Package glib-2.0 was not found in the pkg-config search path.
Perhaps you should add the directory containing `glib-2.0.pc'
to the PKG_CONFIG_PATH environment variable
No package 'glib-2.0' found
ppp-watch.c:78:18: fatal error: glib.h: No such file or directory
compilation terminated.
make[1]: *** [ppp-watch.o] Error 1
make[1]: Leaving directory `/intel/poky2/builds/fetch/tmp/work/x86_64-poky-linux/lsbinitscripts-9.03-r0/initscripts-9.03/src'

(From OE-Core rev: 488610f4a5b7591a989ff1a6137f276a85fd7a85)

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

qemumachines: make MACHINE_FEATURES append follow qemu.inc include

qemu.inc does a straight assign to MACHINE_FEATURES so overwriting the
preceding append to MACHINE_FEATURES, so the MACHINE_FEATURES append
needs to be moved after the include.

This situation came about as a result of commit 71a4bf386:

    qemumachines: Enable xserver-xorg as default xserver

    For qemux86 and qemux86-64 include qemu.inc after defining XSERVER

which missed this side-effect (and maybe others).

(From OE-Core rev: 4f336e5f416df382fdd2b405314741164d537b22)

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

perf: add libexec/perf-core and contents

libexec/perf-core contains all the pre-canned scripts and modules
needed by both the Perl and Python bindigs.  Add libexec/perf-core
along with all the pre-defined perf scripts underneath it.

(From OE-Core rev: 00f9e8aed9f111d351c2dc71cd8605ccbd11f68f)

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

perf: enable Perl binding

Add support to enable the perf Perl binding.

The build depends on perl-native to retrieve the configuration
settings needed for the binding.  cpan-base adds some useful functions
like is_target() and get_perl_version() that we need for
PERLCONFIGTARGET and related settings, which allow us to use the
target's Config_heavy.pl settings for ExtUtils:Embed when building for
the target.

Also adds the perl-modules dependency to give the target the perl
modules that scripts using the binding need.

(From OE-Core rev: 73a8eed010a821fc4f5c0e71ef749938ed780b44)

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

perf: enable Python bindings

Add support to enable the perf python bindings.  The combination of
these changes and the changes in the python-config sections in the
kernel Makefile enable all the python bindings currently available in
perf.

(From OE-Core rev: 487c3a72f0ac7121b79d2f78b81f1d4732bc258a)

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

perl: use @STAGINGDIR@ in config.sh

archlibexp in the snapshotted config.sh needs to be relative to
staging dir - it gets generated into Config_heavy.pl as one of the
settings and is used by ExtUtils::Embed during the target build.

The substituted setting doesn't make it into the target package
because it's removed by perl_package_preprocess() before packaging.

(From OE-Core rev: b647d2ab84ff0eb94f4b6a1b731d74aaf8c60b9d)

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

perl: add @STAGINGDIR@ for config.sh substitions

For some config.sh substitions, we'd like to be able to use
STAGING_DIR_HOST; add @STAGINGDIR@ to allow for that.

(From OE-Core rev: cc594be0877b2a6616453c402db0a58ef265bc8d)

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

perl: keep original libperl location

Things like ExtTools expect to find libperl at the original
(${libdir}/perl/${PV}/CORE/libperl.so) location, so keep it there but
symlink to it for the original (?) reason.

(From OE-Core rev: deb5d5ededfc404f4c54b18f3dcf8192e1c835e0)

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

mkelfimage: Add stable git build (initial recipe)

(From OE-Core rev: c04136393325967cbd491636b582d3e0f394de16)

Signed-off-by: Raymond Danks <ray.danks@se-eng.com>
Acked-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

image_types: Add elf image type

On x86, an ELF image file may be stored as a coreboot payload.
The image file is constructed, using the mkelfimage utility,
from a kernel and an initrd.

(From OE-Core rev: 93e9de4e27919f59b6783d53f314df6e49ccb436)

Signed-off-by: Raymond Danks <ray.danks@se-eng.com>
Acked-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

package.bbclass: enable the use of package_qa_handle_error

This will allow the reporting of these errors as either WARNINGs (default)
or ERRORs if installed_vs_shipped is added to the ERROR_QA of the policy
file (such as a <distro_name>.conf file.

V2: found the code I had intended to send instead of that other junk,
    was just not watching what I pushed on that one, sorry. (this is edit in
    no in the actual commit message)

(From OE-Core rev: 20d4205a4e408b6a99db392d4df458156113106a)

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

oprofile: Change /root to /home/root

OE uses /home/root instead of /root for the root user's directory.  Update
oprofile to match.

Upstream-status: Inappropriate [OE Specific]

(From OE-Core rev: d50c44585e72057532b8e0d9221ebf058e514672)

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

speex: use lib_package bbclass

This is a clean-up following the recent addition of PN-bin
to bitbake.conf and lib_package

(From OE-Core rev: 4c41954227144606d2d22fbdf07231f4c7207ac5)

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

gnutls: use lib_package bbclass

This is a clean-up following the recent addition of PN-bin
to bitbake.conf and lib_package

(From OE-Core rev: b9553b5fe8e06032770fd34e26ab97e6834292ab)

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

db: use lib_package bbclass

This is a clean-up following the recent addition of PN-bin
to bitbake.conf and lib_package

(From OE-Core rev: b3bfeec4db7d8fa6b8cc4bb2153c7f94cbe1dcc3)

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

pulseaudio: use lib_package bbclass

This is a clean-up following the recent addition of PN-bin
to bitbake.conf and lib_package

(From OE-Core rev: 42dfedc01beb90adfa222a0c671ca8876849f938)

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

libacpi: use lib_package bbclass

This is a clean-up following the recent addition of PN-bin
to bitbake.conf and lib_package

(From OE-Core rev: 6597f7c77e1ccb21d73f8e3cec25bb27a901b500)

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

lib_package: use PACKAGE_BEFORE_PN

Since this check effective is not a change becase the package
order that was in this file has not really changed no PR Bumps
are needed.

(From OE-Core rev: d807d5dd5bd0e161057ee115256dc050dbe8609b)

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

bitbake.conf: Add PACKAGE_BEFORE_PN to PACKAGES

By doing this we can easily add addtional packages that can be
greedy before PN, but after some of the other standard packages.

This will also allow us to simplify the lib_package class to set
this variable.

(From OE-Core rev: b41b2bc584756956a48c9c5f91e7337e821105b2)

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

kernel: Add kernel headers to kernel-dev package

[YOCTO #1614]

Add the kernel headers to the kernel-dev package. This packages what was
already built and kept in sysroots for building modules with bitbake.
Making this available on the target requires removing some additional
host binaries.

Move the location to /usr/src/kernel

Before use on the target, the user will need to:

    # cd /usr/src/kernel
    # make scripts

This renders the kernel-misc recipe empty, so remove it.

As we use /usr/src/kernel in several places (and I missed one in the
previous version), add a KERNEL_SRC_DIR variable and use that throughout
the class to avoid update errors in the future.

Now that we package the kernel headers, drop the
kernel_package_preprocess function which removed them from PKGD.

All *-sdk image recipes include dev-pkgs, so the kernel-dev package will
be installed by default on all such images.

(From OE-Core rev: 6125ea40d4483965f793bd847b3ce14b668a5b1e)

Signed-off-by: Darren Hart <dvhart@linux.intel.com>
CC: Bruce Ashfield <bruce.ashfield@windriver.com>
CC: Tom Zanussi <tom.zanussi@intel.com>
CC: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

linux-yocto-tiny: add KMETA specification

With recent kern tools changes, having a meta branch is opt-in. The
other linux-yocto recipes were updated, but linux-yocto-tiny was
missed.

Without specifying a meta branch, the kernel configuration audit
won't return any useful results, and throws a warning message.

 WARNING: Can't find any BSP hardware or required configuration fragments.
 WARNING: Looked at //cfg///hdw_frags.txt and //cfg///required_frags.txt in directory: //cfg//
 NOTE: package linux-yocto-tiny-3.2.18+git1+ee78519365bdb25287703bbc31c06b193263c654_1+27b68a93eb791e830da8d3a2c0fc99780897ad89-r3.0: task do_kernel

(From OE-Core rev: 29dbe7d7ba66c66e9e5c8dc438f89d9b46c509a0)

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

linux-yocto/3.4: update preempt-rt to 3.4.4-rt13

Updating preempt-rt in the 3.4 kernel to the latest upstream version
3.4.4-rt13.

(From OE-Core rev: f5ed8bc876fd24e4f08d29da06454037eec818e6)

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

linux-yocto/3.4: v3.4.4 + configuration updates

Bumping the kernel version to v3.4.4 and importing the following
meta branch commits:

 9b7c74b meta: bump kver to v3.4.4
 8231dec ck: relocate config post patches
 26b965e net_sched: remove dupicate configuration option
 df8bf19 meta: remove non applying routerstationpro patch
 292d8ea arm: import upstream versatile io.h fix
 f4824b1 net_sched: select ACT as a dependency of INGRESS
 30ae722 cgroups: delete obsolete namespace option
 51f94f0 meta: add qemumipsel mapping

Built and boot tested on qemu*

(From OE-Core rev: 3d33b8c8b3be5dcdac30c115b9fec427269db2b4)

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

linux-yocto/3.4: add qemumipsel machine mapping

Updating the meta branch with a board description and configuration
that can be found from the linux-yocto recipes without doing a
MACHINE -> KMACHINE mapping.

From the meta branch commit:

    The mti malta board description for a little endian config can't
    be found by default unless it has a matching KMACHINE, or the
    KMACHINE is set in a kernel recipe. In this case, it is easier to
    just add qemumipsel in the KMACHINE list, so the recipe doesn't
    need to do the mapping.

The little endian mips support reuses everything but the endianess
settings from the big endian mti malta board emulation.

This change adds the required mappings and SRCREVs, but does not
update the machine compatibilty, since official support is still
pending. The compatibility can be set in addon layers until that
time.

(From OE-Core rev: de2535b2bb47b80bdaeb3053b0ac0d1bbb9c5d53)

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

telepathy-glib: update to ver. 0.19.1

(From OE-Core rev: 5109954afefc48d8f44776ec1f2d8a94a21c04f6)

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

pcmcia-cs: Remove references to long-deprecated pcmcia-cs.

pcmciautils is the appropriate PCMCIA package these days.

(From OE-Core rev: 5066f7e9750253bc4678f2884e15f0333e60932b)

Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lighttpd: upgrade to 1.4.31

(From OE-Core rev: 913f4831aeb83e2c7f08daa47a31089dbf06f2be)

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

xcb-util-keysyms: upgrade to 0.3.9

(From OE-Core rev: f89834e7277a77d46b21a363a82371085f028969)

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

xcb-util-image: upgrade to 0.3.9

(From OE-Core rev: 3a88f34e9aa04f5aa988f9f77cf68189da169084)

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

xcb-util-wm: upgrade to 0.3.9

(From OE-Core rev: fba0788bb5ef1743aeceb9d874eb099f7240d118)

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

libxft: upgrade to 2.3.1

(From OE-Core rev: d1e7823c33eafbdd8ca086b8ba48f61faf391f20)

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

ethtool: upgrade to 3.4.1

(From OE-Core rev: 5550c437b50ea9124ce06fa7d17166f2be6d4ee5)

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

libdrm: upgrade to 2.4.35

(From OE-Core rev: 4b640c4737806cbc035fa49bb22b3ee39e372e6a)

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

xcb-util: upgrade to 0.3.9

(From OE-Core rev: 8c6acf7edabb95051384025d3793f6346e234b71)

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

xf86-input-synaptics: upgrade to 1.6.2

(From OE-Core rev: 1e42302426780d8603517418163f46a7eef9931b)

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

xf86-input-vmmouse: upgrade to 12.9.0

(From OE-Core rev: 886c923c1f0259204399582eeb8523fd146c2c62)

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

bitbake: fetch2/svn: Enhance to cope with subversion 1.7 upgrade

svn changed working checkout formats between 1.6 and 1.7. Its convoluted to
detect what format a given working copy is in so the simplest solution is simply
to run "svn upgrade" within the working copy.

The base svn command variable is relocated slightly to enable this new code to
work effectively.

(Bitbake rev: ebd3ecdb5f3c6d3fe1fad27cbed4d80f4277e92e)

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

bitbake: usermanual: Improve task dependency documentation to match reality

(Bitbake rev: f67c0606fc681359fb0c68be55cfc9f11d410f17)

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

bitbake: runqueue.py: Allow recrdeptasks that have self references

In some cases we want to pull in DEPENDS and RDEPENDS of recrdeptask
dependencies but we need a way to trigger or avoid this behaviour
depending on context. The logical syntax to trigger such behaviour
would be a self referencing recrdeptask:

do_a[recrdeptask] = "do_a do_b"

The dependency chains already recurse this kind of expression correctly, the
missing piece is to avoid any circular reference errors. Since the dependencies
have already been recursively resolved, simply removing any recrdeptask
references is enough to break the circular references.

This patch therefore removes any circular references using the set
difference_update() operator. There will be metadata tweaks required to
add any references needed to the extra taskname.

(Bitbake rev: a5324da9b8a0c9307a6c511ea9009f34be70c92b)

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

bitbake: Hob: fixed some variables not clean bug in detail page

[YOCTO #2679]

(Bitbake rev: 73801f571e040dcdfeb15a15b9a484cbefaae70c)

Signed-off-by: Liming An <limingx.l.an@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: git: do not execute 'git remote prune'

'git remote prune' at this location does not make much sense because
the following 'git remote rm' will prune stale and non-stale branches.

The 'prune' can cause trouble because it will access the network
bypassing the no-network code in bitbake. When this operation fails and
throws an exception, the next command (--> 'git remote rm') will be
skipped.  This in turn, will make all the following operations fail,
because they assume that the remote does not exist yet.

(Bitbake rev: 2ba23df5fad4b94d38a6aed97f7822226d72eb89)

Signed-off-by: Enrico Scholz <enrico.scholz@sigma-chemnitz.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

meta/classes: Update recrdeptask fields for recursive dependency handling changes in bitbake

This also deletes the buildall task since I seen usecases for it.

(From OE-Core rev: 8229fb5d7205f5e5b198ab2860fbcc02054476eb)

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

subversion-native: Enable subversion-native 1.7 and remove it from ASSUME_PROVIDED

This enables a switch to subversion 1.7 now bitbake is able to cope with
upgrading existing working copies. The impact of this change should be
minimal since we don't have many subversion recipes now.

(From OE-Core rev: ac0aa35ba6d7a21636bdd23d45ae0bf8112bdaa8)

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

task-core-tools-debug: Added openssh-sftp-server.

[YOCTO #2605].
Added openssh-sftp-server to the tools-debug IMAGE_FEATURE.

(From OE-Core rev: a6b6df1350149c116050cb93c3c7b4802c709d31)

Signed-off-by: Lianhao Lu <lianhao.lu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

image/core-image: Handle conflicting IMAGE_FEATURES.

IMAGE_FEATURES such as 'ssh-server-dropbear' and 'ssh-server-openssh'
can't be both enabled. User can use the following variables to define
the relationship of image features:

IMAGE_FEATURES_REPLACES_foo = "bar" means including image feature "foo"
would replace the image feature "bar".

IMAGE_FEATURES_CONFLICTS_foo = "bar" means including both image features
"foo" and "bar" would cause an parsing error.

(From OE-Core rev: e36d12a9c1cf69540079e48a1dfadbc343758e48)

Signed-off-by: Lianhao Lu <lianhao.lu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

openjade-native: fix undefined Getopts error, use std namespace

Using Gentoo Linux as the build host, it fails without this patch
Use Getopt::Std in place of getopts.pl.

https://bugs.gentoo.org/show_bug.cgi?id=420083

which following error:
/usr/bin/perl -w ./../msggen.pl -l jstyleModule InterpreterMessages.msg
/usr/bin/perl -w ./../msggen.pl -l jstyleModule DssslAppMessages.msg
Undefined subroutine &main::Getopts called at ./../msggen.pl line 22.
make[2]: *** [InterpreterMessages.h] Error 2
make[2]: *** Waiting for unfinished jobs....
Undefined subroutine &main::Getopts called at ./../msggen.pl line 22.
make[2]: *** [DssslAppMessages.h] Error 2

(From OE-Core rev: 169a89b10817b742c063fcd76721e4dbbcca6199)

Signed-off-by: Dennis Lan <dennis.yxun@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

base.bbclass: Add cross-canadian-${TRANSLATED_TARGET_ARCH} to the license exclusion list

This appears to be an oversight in the original implementation.  All of the
host package types were being ignored except for the SDK cross-canadian type.

(From OE-Core rev: 750f99d4e2c805985cd87a2358b0625a808ecf4d)

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

image.bbclass: Add support to build the SDK in parallel with the image

When building an image recipe, you can now build a companion SDK by
calling the populate_sdk task:
  bitbake -c populate_sdk core-image-minimal

Note: there are still issues w/ the SDK not working completely with
multilibs.

A lock is required between rootfs and populate_sdk activities to prevent
configuration file clashes and similar package management problems in ipk
and deb based systems.  (RPM already had a lock for a different reason.)

(From OE-Core rev: a0de2a56f19ae4d8cd88e46e96917a7a019fe1ab)

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

package.bbclass: Allow a package to skip per file dependency generation

Most of the time skipping per file dependency generation is a bad idea, but
when building a nativesdk or similar you may be required to pickup host
dependencies.  These host dependencies can not always be reconciled within
the scope of other nativesdk components, so if we skip them we can facilitate
this unique situation.

(From OE-Core rev: a1ca125de55a8e7cab402e2ba5737a56d4e890c2)

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

libc-common: Allow -dbg package to append, and not just set package variables

The libc-common attempts to rewrite the package information in a way similar
to debian.bbclass.  When it does this, it should be appending to the
dependency variables (RPROVIDES, RREPLACES, and RCONFLICTS), instead
of simply setting a hard coded value.  Otherwise the lib package can not
tailor the dependency variables to suite it's needs.

(From OE-Core rev: fa1c7b797593cbd5e82dc264bde2667620eb0515)

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

autotools.bbclass: When configure fails, dump the config.log files the logs

When configure fails, it usually says "see config.log" yet nobody ever shares
the config.log file meaning the person trying to help invariably has to ask
for more information.

This patch dumps all the config.log files into the main bitbake log files when
configure fails, meaning all the information is present to help someone debug
such failures. It does make the log rather larger but this is preferable to
not having enough information in most cases.

[YOCTO #2463]

(From OE-Core rev: 7ad08f5b73aa949a877adc5641b4bb1d007de750)

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

u-boot: do not clobber PARALLEL_MAKE setting

The u-boot tree is fully capable of parallel builds, so this
setting should not exist as a blanket setting for all of the
recipes.  Going forward, if there is a parallelism issue
in u-boot, it needs to be reported and fixed there, and not
with the "make -j1" band-aid approach.

(From OE-Core rev: 7fdd1f82803df9752e908f01f7643d66b82a690e)

Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

u-boot: Don't make the -Os removal part of global settings.

We don't want to force everyone to be stripping the -Os
flags from their u-boot builds.  Remove it, since it pertains
to an old toolchain issue that is no longer relevant, and it
breaks the powerpc mpc8315.

(From OE-Core rev: 5e3f777f9fcc986ae66203651830b5765bca49fd)

Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

populate_sdk: enable basic multilib support

In order to enable basic multilib support for the export of an SDK
image, a number of minor changes had to be made:

The value of MULTIMACH_TARGET_SYS needs to be variable.  This way we
can define the value to the appropriate multilib.  (Also in some cases
the default PACKAGE_ARCH is set to MACHINE_ARCH which is incorrect for
the SDK.)  Add a companion REAL_MULTIMACH_TARGET_SYS, based on code
from meta-environment, to allow for this.

We have to convert the do_populate_sdk into a python call, and then
break up the previous items into three parts.
  * Image construction
  * Setup of environment files
  * Generation of the tarball

Then we can iterate over the multilibs to populate the environment files.

Finally, matching changes were needed in the toolchain-scripts file.  And
what I presume is an optimization of immediate evaluation for
TOOLCHAIN_CONFIGSITE_NOCACHE and ..._SYSROOTCACHE needed to be done at
runtime, otherwise the wrong values may be used.

(From OE-Core rev: 42545ffbb37f2646a2a8c20999c21d3300e24f59)

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

populate_sdk_base.bbclass: Change to using task specific depends

If we combine the do_populate_sdk with the image generation, we want
to avoid the dependency processing unless do_populate_sdk is run.

This requires the bitbake change to implement task based rdepends.

(From OE-Core rev: 5fa9f980248e7813ce74f48a29c4b7d94e308cf9)

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

populate_sdk: Allow for attempt only packages in the SDK

We want to be able to supply attempt only packages in the SDK in order to
support more advanced SDK images that more closely match specific image
recipes.

(From OE-Core rev: 9b9efa96537f4977b158c29151e53d02600d2294)

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

populate_sdk.bbclass: Split into two parts

populate_sdk was renamed to populate_sdk_base in order to allow for changes
that may break existing SDK recipes.  Any such changes need an analog
in populate_sdk (new version) to restore previous desired behavior.

In addition to the rename, one minor change was made.  The _base version
only had the do_populate_sdk as an added task, but no before or after defined.

For compatibility, populate_sdk has do_populate_sdk defined as "after"
do_install and before do_build, this is identical to the original behavior.

(From OE-Core rev: bde224ba44c16edc1892cea1b33ab973ae971115)

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

Fix manual log file paths

When a recent change, the path to log files may be contained within an
arbitrary directory.  To generate the manual log files in the correct path
we should be using the ${BB_LOGFILE}'s path instead of always assuming the
logs go into ${WORKDIR}/temp.

(From OE-Core rev: 779db325d407f0bade84572ef99fdad4d0c88011)

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

core-image*sdk: Add kernel-dev to sdk images

While sdk images already have dev-pkgs included, the kernel-dev package
is special and does not get installed. Add kernel-dev explicitly to the
*sdk image recipes.

Note: We have to be careful with "IMAGE_INSTALL +=" as it must appear
after "inherit core-image" which makes an "IMAGE_INSTALL ?=" assignment.

(From OE-Core rev: 337428316de3612a8e1efa70b3a4ba7cb6e058d2)

Signed-off-by: Darren Hart <dvhart@linux.intel.com>
CC: Jessica Zhang <jessica.zhang@intel.com>
CC: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

linux-yocto-custom: Remove linux-tools

Remove the linux-tools include as perf is now packaged independently.
Without this, including meta-skeleton in bblayers.conf results in a
build failure as linux-tools.inc cannot be found.

(From OE-Core rev: 9d27daeafcc6ea892dee0ee5d0816069b04f6771)

Signed-off-by: Darren Hart <dvhart@linux.intel.com>
Acked-by: Bruce Ashfield <bruce.ashfield@windriver.com>
CC: Tom Zanussi <tom.zanussi@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

icu 3.6: do_install failed: Segmentation fault

There is a "Segmentation fault" error when the tmpdir is longer than 470
(or less), this is because it used "char cmd[1024]" which is not enough
for the command line. Allocate a larger memory size should fix this problem.

[YOCTO #2664]

(From OE-Core rev: 3fd5d0334439b6231a6afd2a544a16b87ae81c99)

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

kconfig-frontends: New recipe

kconfig-frontends provides the kconfig infrastructure
thats used for linux kernel as a standalone package. Its
not intended to be used for kernel itself but can be
used by packages who are based on kconfig e.g. busybox
uclibc or new packages who wants to use kconfig infra
for the package build

kconfig-frontends names it's shared library as -3.4.0.so instead of .so.3.4.0,
this causes the -dev package to grab it since it matches on *.so. This error
would also have shown up with the old PACKAGE order.

(From OE-Core rev: 705baf06d151eb4f9dcae7b23c7f3e57c0c09a74)

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: Add missing PKGR, PKGV, PKGE and ALTERNATIVE variables from PACKAGEVARS

These variables all take package name extensions and should be considered in sstate
dependency processing.

This may also fix some multilib alternatives issues (unconfirmed).

[YOCTO #2056]

(From OE-Core rev: 64422f7c5da160050a5454817c8fa9d070104b34)

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

package_deb: When searching for E:, anchor expression

If we don't do this it will match on expressions like "NOTE:" which
are not fatal errors.

(From OE-Core rev: 8f4e160ce0b0c755545d62b06db14decd45b9f1e)

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

EXTRA_IMAGEDEPENDS is used to ensure things like qemu-native are built and these are not strictly dependencies of the do_rootfs task. This patch moves them to the do_build task allowing bitbake a little more flexibility about when things need to get built by.

I noticed this when qemu-native failed to build and a rootfs was not generated
which is not the intended behaviour.

Also update the syntax to use appendVarFlag instead of get/setVarFlag

(From OE-Core rev: fa13e83ec3f91dce866ac212e91b62db24b6486d)

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

kernel-arch: Remove test for long-dead arm26 architecture.

(From OE-Core rev: ff3f912eed0270afa14d706fc1e57f8a1de2614b)

Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>