wic: Add mkgummidisk kickstart file

This is the same as mkefidisk but uses gummiboot instead of grub-efi.

(From OE-Core rev: 5979409ebfab0bb07b3c2b2fcf14a722c441f07b)

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

wic: Add sourceparams to mkefidisk.wks

The bootimg-efi plugin now requires a loader param, so supply it to
retain existing behavior.

(From OE-Core rev: ccef1385cb51ce2b9b75493b314e38599a8ae10e)

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

wic: Add gummiboot support to bootimg-efi

Add gummiboot support to bootimg-efi, which retains the existing
grub-efi support (though requires an update to the .wks file).

(From OE-Core rev: f1a81fcefa493540a9faac549fdd513b86f8f497)

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

wic: Parse and pass sourceparams to partition plugin methods

Add code to parse the sourceparams and pass them to the partition
plugin methods.

(From OE-Core rev: 769a6f920fb672c50d76bd4bdea4a65fdc18a7b3)

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

wic: Add sourceparam param to partition plugin methods

The sourceparam param allows source plugins to be parameterized
generically (via --sourceparams="key=val[,key=val], implemented
previously).

(From OE-Core rev: 2abbcc843ba888782f6d68748d930c81e6ae7040)

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

wic: Add utility function for parsing sourceparams

Parses strings of the form key1=val1[,key2=val2,...] and returns a
dict.  Also accepts valueless keys i.e. without =.

(From OE-Core rev: 36f258ee6e60c26fd44b9bc71c318363cec71f42)

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

wic: Add '--sourceparams' partition option

To go along with '--source' as a way to parameterize source plugins.

(From OE-Core rev: f5b9ef65453b3f66282c49f5e2584ad33ac6d2d7)

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

image.bbclass: image_types.bbclass is a must

The image_types.bbclass is a must since we use this in image.bbclass:

[snip]
python () {
    deps = " " + imagetypes_getdepends(d)
    d.appendVarFlag('do_rootfs', 'depends', deps)
[snip]

The imagetypes_getdepends() is defined in image_types.bbclass. Use "+="
to replace "?=" since it is a must, so that the user can use
"IMAGE_CLASSES = foo.bbclass" in local.conf to add their own image
class.

NOTE: the IMAGE_CLASSES_append = " foo" doesn't work since we use this
in image.bbclass:
 IMAGE_CLASSES += "image_types"
 inherit ${IMAGE_CLASSES}

I think that it is because inherit takes effect before append?

Another way to fix the problem is:
IMAGE_CLASSES ?= ""
inherit image_types ${IMAGE_CLASSES}

But it seems that we need another name for IMAGE_CLASSES, for example
IMAGE_CLASSES_EXTRA, and also need update the doc, which would make it
complicated.

(From OE-Core rev: c225613d997ef40dedf64b880b5ca4ae89adfded)

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

gmp: uprev it to 6.0.0

Uprev gmp from 5.1.1 to 6.0.0, and remove the 4.2.1 version which is GPLv2,
since gmp-6.0.0 is dual-licensing, LGPLv3 or GPLv2;

(From OE-Core rev: f181c6ce8b364fbf761a456d998ab78fbd751f35)

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

Support image type "squashfs-lzo"

Add "squashfs-lzo" to the image types. LZO compression support has been
in both kernel and squashfs tools for many years, but OE never enabled
it.

(From OE-Core rev: ab659b1c1e13fd6863893abff86990d88aa36790)

Signed-off-by: Mike Looijmans <mike.looijmans@topic.nl>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

squashfs-tools: Add support for LZO and LZ4 compression

For systems that want to optimize for speed rather than size, LZO is
usually a better choice than gzip or XZ. Kernel support for LZO has
been available since 2.6.29.

LZ4 support isn't in the mainline kernel yet, but we might as well add
it now for those who want to experiment with it.

(From OE-Core rev: 6c95440ed5c181754bb7c04da678081498840683)

Signed-off-by: Mike Looijmans <mike.looijmans@topic.nl>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

strace: apply ptrace.h conflict workaround

Apply patch from strace upstream to workaround ptrace.h
header file conflict. This patch is not available in
strace-4.8 tarball pulled in by recipe.

* patch from strace upstream
 - Work around conflict between <sys/ptrace.h> and
   <linux/ptrace.h>

(From OE-Core rev: 3c7301adce142bab64d49b5be7c39d8b223591f5)

Signed-off-by: Yasir-Khan <yasir_khan@mentor.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

curl: --with-random is only applicable with openssl

(From OE-Core rev: 482493b54d97c455bf4849efed3e543340412d7b)

Signed-off-by: Andre McCurdy <armccurdy@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

curl: let configure find gnutls via pkg-config

(From OE-Core rev: 3682d661f3b3a6fa7d9ef37968746cbaf1ede078)

Signed-off-by: Andre McCurdy <armccurdy@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

curl: add zlib PACKAGECONFIG and remove hardcoded DEPENDS

Add a zlib PACKAGECONFIG control and update PACKAGECONFIG[ssl] to
include the openssl dependency. Older hardcoded DEPENDS can then
be removed.

(From OE-Core rev: e668c79de927eff635f29fb5ff001f6b106ccc81)

Signed-off-by: Andre McCurdy <armccurdy@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

python-pygtk: fix native python path

Fixed:
1) Set IMAGE_INSTALL_append = " python-pygtk-demo" in local.conf
2) $ bitbake core-image-minimal
[snip]
Computing transaction...error: Can't install python-pygtk-demo-2.24.0-r1@core2_64: no package provides /usr/bin/python-native/python
[snip]

(From OE-Core rev: d70bc02a80f8508995cdf8ae536a0153cebee8c9)

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

consolekit: depend on glib-2.0-native

This is required for glib-gettextize.

(From OE-Core rev: ef5e0c2ab57cedee64c4a470de170dc1f098be60)

Signed-off-by: Christopher Larson <kergoth@gmail.com>
Signed-off-by: Yasir-Khan <yasir_khan@mentor.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

consolekit: allow explicit enable/disable of polkit

Adds the configure option to enable/disable policykit.

(From OE-Core rev: 66d53c40f1eb614803a34dfa3741b170785c637d)

Signed-off-by: Christopher Larson <kergoth@gmail.com>
Signed-off-by: Yasir-Khan <yasir_khan@mentor.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

udev-extraconf: Unmount SD card after ejection

Adds udev rule to unmount SD card partitions in case of improper
ejection from card reader.

When SD card is ejected from card reader without being unmounted
first, kernel does not generate a REMOVE event, instead it
generates a CHANGE event(only if polling is enabled
/sys/module/block/parameters/events_dfl_poll_msecs) and we don't
have any udev rule in automount.rules to handle this event,so
partitions never get unmounted. Unmounting of partitions can be
done if udev rules handle this CHANGE event.

(From OE-Core rev: 140d188b3665f914396834168feeff6ffcb64d76)

Signed-off-by: Abbas Raza <Abbas_Raza@mentor.com>
Signed-off-by: Yasir-Khan <yasir_khan@mentor.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

opkg: Protect add-exclude.patch from malloc failure

In the code added by add-exclude.patch, the return values of malloc and realloc
were not checked before being dereferenced.

In opkg we can use xmalloc and xrealloc instead of malloc and realloc. These
functions terminate the program instead of returning NULL if memory allocation
fails.

(From OE-Core rev: 64048d45d5c52d354396e357e765f0fd8b5e56fd)

Signed-off-by: Paul Barker <paul@paulbarker.me.uk>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

multilib_global.bbclass: PREFERRED_PROVIDERS for multilibs

The code in base.bbclass to spread PREFERRED_PROVIDERS values
to multilibs doesn't work for things which rely on TARGET_PREFIX,
such as virtual/${TARGET_PREFIX}gcc. This is because the expansion
of TARGET_PREFIX produces the wrong value if executed prior to
the assignment of TARGET_VENDOR_virtclass-multilib-libxx, which
will always happen since that assignment doesn't happen until recipe
parsing, but the PREFERRED_PROVIDERS expansion is happening
around ConfigParsed.

To solve this, we make a couple of changes. First, the creation
of the TARGET_VENDOR override values is moved into a new ConfigParsed
event handler in multilib_global. Second, the preferred_ml_updates()
function's code is moved into that function too. It seems safe to
assume that PREFERRED_PROVIDER values only need to be spread to
other multilibs when multilibs are in use.

I don't think this directly affects any use cases that don't involve
third-party or alternative toolchains.

(From OE-Core rev: 513f72274460e54fd35dda5ef70fa42ba2b284f8)

Signed-off-by: Peter Seebach <peter.seebach@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gcc: update compiler architecture to match gcc-runtime (armv6, armv7a)

The gcc-runtime recipe builds the gcc libraries including libstdc++ with
$TARGET_CC_ARCH flags, which include -march=FOO flags that affect
whether atomic instructions are available.  This causes an ABI
incompatibility when the compiler by default generates code for less
capable architectures.  For example, gcc-runtime libraries on a
Cortex-A8 are built with a different C++11/C++14 mutex implementation
than is used code compiled outside OE and without architecture-specific
flags.

This commit fixes the problem specifically for ABI issues related to
atomic instructions available in ARMV6 and subsequent architectures.
Other ABI incompatibilities may remain in other architectures.

See: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62100

(From OE-Core rev: 0ba6ab39f187ecd4261f08e768f365f461384a3a)

Signed-off-by: Peter A. Bigot <pab@pabigot.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gcc: backport patch affecting Linux kernel builds

A long-standing bug in gcc turns out to cause problems with unpatched
Linux versions due to improved optimization enabled by gcc 4.9.  The
upstream fix missed the gcc-4.9.1 cut-off.  It's also been applied
upstream to the 4.8 branch so is being added for OE's 4.8 as well.

(From OE-Core rev: 06f911894a367f395139c2b0d6c2ba6371398478)

Signed-off-by: Peter A. Bigot <pab@pabigot.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gcc: Abstract long double configuration into python function

musl does not support IBM 128 long double for ppc, instead of
doing complex overrides move it into a pythong snippet which
is easier to read and more compact.

(From OE-Core rev: e7011429e40ae96b9c9f1e7f3c6f4c1f1102607f)

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

sdk: change EXTRA_OECONF_FPU to EXTRA_OECONF_GCC_FLOAT

This variable is used to ensure the proper version of --with-float=FOO
is passed to gcc's configure script.  gcc also has a --with-fpu=FOO
option that means something different.  To avoid confusion, change the
names to be consistent.

(From OE-Core rev: c17d883fa99b6967d83c3796d22fc0c1dbe704e6)

Signed-off-by: Peter A. Bigot <pab@pabigot.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gcc-target: make --enable-clocale consistent with gcc-runtime

(From OE-Core rev: 9ec30be63ad6d991646a7ce0ee22acdad7a81184)

Signed-off-by: Peter A. Bigot <pab@pabigot.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gcc: remove outdated configuration option

--enable-libunwind-exceptions was removed from gcc at release 3.4.3
about ten years ago.

(From OE-Core rev: 285d3579727177e6962d7ad16677429e7dec65f4)

Signed-off-by: Peter A. Bigot <pab@pabigot.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gcc-4.9: Ensure c++ includes are in /usr/include/c++/${BINV}

Apply to gcc 4.9 the recent fix to the --with-gxx-include-dir override.

Original OE-Core rev: 5a2ff3e8f7cd7a47a5ab4e581847ecc4df87fca

(From OE-Core rev: 5fec278316fa9466241b9134c4553bad6db1c1a9)

Signed-off-by: Peter A. Bigot <pab@pabigot.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gcc: remove inappropriate patch

0037-gcc-4.8-PR56797.patch was originally added as an OE backport during
4.8.0.  Upstream merged it in 4.8.1, and it was present in 4.9.0.

The original patch still applies to 4.9.1 (and presumably 4.8.2), but
now is modifying store_multiple_sequence instead of
load_multiple_sequence (the two functions are nearly identical).  It may
or may not be necessary in store_multiple_sequence, but absent a bug
report upstream supporting its application in this case, or a least an
updated comment and upstream status in the patch, I think this patch
should be dropped.

(From OE-Core rev: c89443e0f98249b9f9ea33f686c27babe35fd024)

Signed-off-by: Peter A. Bigot <pab@pabigot.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gcc: recipe whitespace changes

Consistent use of whitespace in multi-line assignment, with special
focus on OECONF modifications.  Quotes on separate lines, four-space
indentation, one value per line.

(From OE-Core rev: d971db8b2259e4c35b871cccf130fba193849560)

Signed-off-by: Peter A. Bigot <pab@pabigot.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

tcmode-default.inc: Add eglibc-scripts and eglibc-mtrace

Add the specific scripts and mtrace versions to the preferred version list.

(From OE-Core rev: 6fa356c6a808a4876e23722ab86b80a4c6bd072d)

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

prelink_git: Update to current head of cross-prelink development

Update to the current version of cross-prelink development:

faa069deec99bf61418d0bab831c83d7c1b797ca

This adds a number of minor features, such as S390 support (irrelevant)
to ARM TLS DESC relocations.

(From OE-Core rev: e87a56bbdfbae56bc027ab66b350cc8fa7b65810)

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

autoconf: add rdep on perl-module-data-dumper

This patch adds rdepends on perl-module-data-dumper
for autoconf.

(From OE-Core rev: 8b3a911aac460a137099ca2b3a0a84fedea31309)

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

gcc-cross-initial: Use good old bfd linker by default

We already indicate our intentions to use ld.bfd by
specifying it in configure using --with-ld which works
ok unless here where we manually create symlinks to
binutils-cross components, when we use ld-is-gold feature
default ld points to gold and this symlinking has to be
aware of the fact that we configured binutils and gcc-cross to use
gold as default ld but gcc-cross-initial uses BFD ld

This would be visible when using gold and rebuilding
eglibc

(From OE-Core rev: 77cab553ee6caa940e21cca46ff134f84e65c171)

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

cmake: drop -fpermissive

* it was dropped from default CXXFLAGS in:
  commit 24dd8e129447013ee98609f3892ec414b1b21340
  Author: Richard Purdie <richard.purdie@linuxfoundation.org>
  Date:   Sun Mar 2 17:38:33 2014 +0000

    bitbake.conf: Drop -fpermissive

    Drop the -fpermissive C++ compiler flag. We've had this around for years, most
    code should have been fixed long ago. Its possible some recipes may fail
    however we can (and should) just use the flag where needed.

* I haven't build world with this yet, but maybe it's time to
  drop it here as well at least for consistency

(From OE-Core rev: 578124aedd3dbffd79ea01862d57223ffaa7216a)

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

gstreamer1.0-rtsp-server: Upgrade to 1.4.0

(From OE-Core rev: 466268fc86c3a08fab0ba45243c32bfaad613a96)

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

libgcrypt: Do not remove m4 files so aggressively

Removing all m4 files in the m4 directory led to a number of non-fatal
errors while running configure when the expected m4 macros could not
be found.

(From OE-Core rev: b43b6bae5b164f641233dcd2e373f3cae26273a6)

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

python-scons-native: Make it useable if old host install exists

This was recently fixed to work on Fedora 17 if no scons is installed
on the host by setting the PYTHONPATH to where BitBake has installed
scons. However, if an older version of scons than 2.3.0 is installed,
then it still breaks. This is due to how scons tries to determine its
installation by searching through standard paths. If it finds an old
installation it prepends that path to sys.path thereby causing it to
ignore the PYTHONPATH. The solution is to instead set SCONS_LIB_PATH
which works both if scons is not installed and if an older version is
installed.

(From OE-Core rev: e16c968ffb96fac3177bb885872c2b5cdde87239)

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

glib-networking: Inherit gettext

Build fails otherwise if USE_NLS is set to "no".

(From OE-Core rev: 4c1f678e4e5383baf40f367e6957dfd33fe52342)

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

libsoup-2.4: Inherit gettext

Build fails otherwise if USE_NLS is set to "no".

(From OE-Core rev: 07f9215d9615dfc26c7c1fccb98adbe593a0d967)

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

systemd: Inherit gettext

Build fails otherwise if USE_NLS is set to "no".

(From OE-Core rev: 7799774a76757be97e76b0568acdf5295ce73536)

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

qemu: add PACKAGECONF lzo

Add PACKAGECONF 'lzo' for qemu to fix QA warning:

WARNING: QA Issue: qemu rdepends on lzo, but it isn't a build dependency? [build-deps]

(From OE-Core rev: b05cf36a1fcb0fc2d3795ab43481f6de01de2af2)

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

qemu: enable aarch64 support

qemu 2.1.0 support aarch64 targets, so add aarch64 to QEMU_TARGETS to
enable qemu aarch64 support.

[YOCTO #6487]

(From OE-Core rev: 2ff11438b763ddebfd798ccfe93d0d981d596202)

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

python: Fix build on musl

musl has posix_close which conflicts in python
so lets rename it.

(From OE-Core rev: 9de4f6eb07696f618d4762c6eeb34dc9ea3080bd)

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

gstreamer1.0-plugins-good: Add a PACKAGECONFIG for udev

This solves these floating dependencies:

WARNING: QA Issue: gstreamer1.0-plugins-good-video4linux2 rdepends on
libudev, but it isn't a build dependency? [build-deps]
WARNING: QA Issue: gstreamer1.0-plugins-good-video4linux2 rdepends on
libgudev-1.0, but it isn't a build dependency? [build-deps]

(From OE-Core rev: b23a9c94e574fec0f0dcda498c1f5a0592114055)

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

gdb: Add a dependency on zlib

This solves the following warning:

WARNING: QA Issue: gdb rdepends on zlib, but it isn't a build
dependency? [build-deps]

(From OE-Core rev: 3dc99615b265999206264be33af1258a9c41c7b3)

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

cml1.bbclass: fix diffconfig diff command

Commit 6a7b98993350d0d24eae0058ae26ae19cfdf7c4c introduced a typo in
command, a whitespace was missing between --new-line-format and path to
original file. As a result, diffconfig task did not work, diff failed
with exit status 2, resulting in empty fragment.cfg.

(From OE-Core rev: 8c75aa7c30cc5a796962d099896ca2a9a16c3a06)

Signed-off-by: Maciej Borzecki <maciej.borzecki@open-rnd.pl>
Signed-off-by: Maciek Borzecki <maciek.borzecki@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libav: update 9.x version to 9.16

(From OE-Core rev: 32e95316db40d20bb2aa7dd48e35114f3ea7a0ee)

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

libav: update 0.8 version to 0.8.15

(From OE-Core rev: 1bfa027e4667248eeb74ee12fad84e82415efc4b)

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

qemu: Overwrite KVM explicitly for darwin/mingw

Force KVM to disabled on mingw/darwin systems since this makes no
sense there.

(From OE-Core rev: 0bbafc1e0250bc4df6c6fcc749ab9bbbd7de4a89)

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

qemu: Simplify DEPENDS

These DEPENDS are now identical and the native/nativesdk variants can
be dropped in favour of letting PACKAGECONFIG sort it out.

(From OE-Core rev: 67c8f70e84ac25adb5ea54dbde83d2377a994f97)

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

qemu: Add PACKAGECONFIG for glx

With the configuration for sdl and glx, the virtual/libx11 in DEPENDS
no longer makes any sense at all, the dependency comes through the
various graphics backends. Therefore drop it. The glx PACKAGECONFIG
is added for completeness. We avoid a floating dependency on mesa.

(From OE-Core rev: ddb88df0eca85b0b2afca68a23c018d39a4c3d78)

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

qemu: Add alsa PACKAGECONFIG and cleanup native/nativesdk DEPENDS

Converting this dependency to become a PACKAGECONFIG makes sense. Since
there is a "," in the configuration option, its not quite as straightforward
as normal.

Also clean up the native and nativesdk DEPENDS and set the PACKAGECONFIG
entries to match the old behaviour by default.

(From OE-Core rev: f724085b4947ec8437d48f77acd2e9e9d19e39ef)

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

qemu: Fixup target sdl configuration to use PACKAGECONFIG

SDL is now controlled by PACKAGECONFIG, remove the DEPENDS remnant
set PACKAGECONFIG accordingly.

(From OE-Core rev: 5bb25628e408f8738b38baba1f8609ab22cd5c51)

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

qemu: Move dtc(fdt) configuration to PACKAGECONFIG

Converting this dependency to become a PACKAGECONFIG makes sense.

(From OE-Core rev: 15d7dc8df4b508acfeac43a8d0f9e15889ec3e7d)

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

site: Add common-darwin

We need this to ensure there are some correct values when building with
meta-darwin. This matches the addition of a common-mingw site file in
the core.

(From OE-Core rev: 8c2ade6f48dcc522c68f798286d1c43bd2c1c8e7)

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

glib-2.0: Allow correct printf configuration on darwin/mingw

mingw and darwin builds need to change the value of printf from no
to yes. Whilst you can do this from a bbappend, its currently a bit ugly
and assumes knowledge of the other conf arguments. Fixing this in the glib
recipe directly is cleaner in this case.

(From OE-Core rev: ba37c147e82e6466030cf40a9e86442ac7592185)

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

glib-2.0: Make the gtester-report installation conditional

This isn't built/present under mingw builds so make installation
conditional upon its presence.

(From OE-Core rev: 53903afae4fe841c0d394172201660d5e3fad5d4)

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

populate_sdk_base: Move toolchain installation script to a separate file

It makes sense to move the script to a separate file, making the bbclass
clearer and allowing the end user to more easily customise or replace
it. There are no functionality changes, just the addition of some
substitution variables.

(From OE-Core rev: f99732a29689c65083ad09abb302f372042c8cd4)

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

rootfs_*: Refactor common depends/lockfile task flags

(From OE-Core rev: 96427656ef73bb50753a33ec6ad5e73de711ceb3)

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

classes: Combine rootfs_<pkgtype> and populate_sdk_<pkgtype>

There is no real reason to have these separate files any longer. It does
mean in meta-toolchain type recipes some extra variables are defined but
it also means the common code can be grouped and maintained together
which I believe is more beneficial. We therefore merge the classes.

(From OE-Core rev: 2cf42b49003494e1b10775523c9a2547eaf16ea4)

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

SIGGEN_EXCLUDERECIPES_ABISAFE: add initscripts

* update-rc.d now adds initscripts dependency even to allarch recipes
  making them effectively TUNE_PKGARCH, "fix" it by excluding
  initscripts from all signatures

(From OE-Core rev: 4321c553d5ae816e566234e981a0815bba046d39)

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

oe.package_manager: fix use of PACKAGE_EXCLUDE for dpkg

It was iterating over the variable character-by-character rather than
word-by-word.

(From OE-Core rev: 207f2cc0b0fc53b22e1dedfa26905ab143fb0de6)

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

python-dbus: use PACKAGECONFIG for doc generation

Adds 'docs' (for html doc generation) and 'api-docs' (for API doc generation)
configurations and leaves them both disabled by default. This avoids
autodetected dependency upon docutils.

Fixes [YOCTO #6530]

(From OE-Core rev: 7216ddad59f6a2315323befa69eebdbf07625e25)

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

oeqa/runtime: add new cpp test and file

This tests out the limits header which we've noticed does have problems in
some SDK builds.

(From OE-Core rev: 63cbed337241191f33fe951662a39ce59dce6774)

Signed-off-by: Corneliu Stoicescu <corneliux.stoicescu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lib/oeqa/selftest: Don't match log level in output

To facilitate changing the log level of the "Fetcher failure" message,
search only for the message without the "Error:" prefix.

(From OE-Core rev: 947e6f9005abc71f499f23a4dd3a5a9f8386a369)

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

sstate-sysroot-cruft.sh: Improve to use it from CI

* strip tmpdir prefix, so that we have shorter paths which aren't
  builder specific
* use '#' for regexp delimiter so that we don't need to prefix
  forward slashes in paths
* extend default whitelist to cover typical cases
* add parameter for external whitelist file
* use number of found paths as return code, so that CI can easily
  report error when new untracked files are found
* use .txt suffix for all output files, so that they can be easily
  viewed in browser
* add populate_sysroot task, because somewhere between dora and daisy
  the populate-sysroot files in sstate-control were renamed to have
  underscore instead of dash
* only few entries not covered by this default whitelist were found
  in world build (but I'll leave these for people to whitelist, because
  they are not generated in most builds)
  * [^/]*/home/builder
    home directory from meta/recipes-graphics/builder/builder_0.1.bb
  * [^/]*/usr/src/kernel/patches
  * [^/]*/usr/lib/gdk-pixbuf-2.0/.*/loaders.cache
    3 places are using this, not sure which one creates it
    meta/recipes-gnome/gdk-pixbuf/gdk-pixbuf_2.30.8.bb:
      GDK_PIXBUF_MODULE_FILE=${STAGING_LIBDIR_NATIVE}/gdk-pixbuf-2.0/${LIBV}/loaders.cache
    meta/recipes-gnome/gtk+/gtk-update-icon-cache-native_3.4.4.bb:
      GDK_PIXBUF_MODULE_FILE=${STAGING_LIBDIR_NATIVE}/gdk-pixbuf-2.0/2.10.0/loaders.cache
    scripts/postinst-intercepts/update_pixbuf_cache:
      >$GDK_PIXBUF_MODULEDIR/../loaders.cache && \
      sed -i -e "s:$D::g" $GDK_PIXBUF_MODULEDIR/../loaders.cache

(From OE-Core rev: b1bfec63949e16abe8c11c34530dfbfb176c04cd)

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

qt4: Avoid duplicate flags in the g++-unix.conf

The removed flags from this g++.conf file, which is installed to the
target as g++-unix.conf are added by gcc-base.conf . Do not add those
flags twice.

(From OE-Core rev: 3c49c575632cdac7bf950439b33773f09ae63354)

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Eric Bénard <eric@eukrea.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

tslib: Delete unnecessary "SRC_URI_OVERRIDES_PACKAGE_ARCH = 0"

Since the machine-specific files for tslib were removed quite some
time ago, there is no need for this directive anymore.

(From OE-Core rev: 8d7491d22a11632a6bfc2478cb4f040d6e53c16a)

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

trace-cmd: fix recompile error

Fixed:
make: *** No rule to make target `/path/to/sysroot/4.9.0/include/stddef.h', needed by `parse-events.o'.  Stop.
make: *** Waiting for unfinished jobs....
ERROR: oe_runmake failed

This happens when upgrade gcc from 4.9.0 to 4.9.1, and the
.parse-events.d isn't regenerated when recompile, the content of it are:

[snip]
parse-events.o: /path/to/sysroot/4.9.0/include/stddef.h
[snip]

And Makefile includes the .parse-events.d file if it exists, so there
would be errors when /path/to/sysroot/4.9.0/include/stddef.h doesn't
exist.

Remove .*.d (a few .d files, its Makefile uses this wildcard) will fix
the problem.

(From OE-Core rev: d56657cf5def88682954a97b4d94603ad81fd6e5)

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

openssl: upgrade to 1.0.1i

Removed one patch merged upstream.

(From OE-Core rev: fc1d2b4ec7e7f5c5e2b3434bc8208967ead6f336)

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

directfb: add PACKAGECONFIG for webp

* fixes floating dependency:
  directfb-1.7.4: directfb rdepends on libwebp, but it isn't a build dependency? [build-deps]

(From OE-Core rev: ee73ee5a37da284025364f322bbcf429b0896727)

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

kernelshark: add dependency on libxml2

* fixes floating dependency:
  kernelshark-1.2+gitAUTOINC+7055ffd37b: kernelshark rdepends on libxml2, but it isn't a build dependency? [build-deps]

(From OE-Core rev: 99300ef40c4b1f7eeaf5503fa293114d7b602d30)

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

gnupg: Added PACKAGECONFIG[curl]

gnupg can be built with or without curl. This should be explicitly defined
to avoid race between curl and gnupg.

(From OE-Core rev: 4e5f6f95e8e791c9cfc9e3bd91f1018d83ab252f)

Signed-off-by: Mikhail Durnev <mikhail_durnev@mentor.com>
Signed-off-by: Shrikant Bobade <Shrikant_Bobade@mentor.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

curl: add ssl to PACKAGECONFIG

This will allow curl run as nativesdk and fixes the following:

fatal: unable to access 'https://...': Protocol https not supported or disabled in libcurl

(From OE-Core rev: 76a702f4cde7ca8dd2946633f489386e43b6be26)

Signed-off-by: João Henrique Ferreira de Freitas <joaohf@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

base.bbclass: Remove old-style definition of do_build() task.

(From OE-Core rev: f31316b96a97e72d866f12a8e4bb33eaaa54b923)

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

guile: add dependency on ncurses and readline

* fixes floating dependency:
  guile/guile/latest lost dependency on  ncurses-libncurses readline

(From OE-Core rev: 39f98dd8280a17a400055ae20c48cf7b7f81e050)

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

midori: add dependency on libxscrnsaver

* fixes floating dependency:
  midori/midori/latest lost dependency on  libxscrnsaver

(From OE-Core rev: deec41394baf73bc7daf322a965bfb93ce87fdcf)

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

libarchive: add PACKAGECONFIG for nettle

* fixes following floating dependencies:
  libarchive/libarchive/latest lost dependency on  nettle
  libarchive/libarchive-bin/latest lost dependency on  libxml2 nettle

(From OE-Core rev: a4dd641f54f12d454ba9c6db624b94df63f7d220)

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

mpeg2dec: add PACKAGECONFIG for x11 and fix dependencies

* the configure script checks for Xext and Xv when X/libvo is enabled
* fixes following warnings:
  WARN: mpeg2dec: mpeg2dec rdepends on libxext but it isn't a build dependency?
  WARN: mpeg2dec: mpeg2dec rdepends on libxv but it isn't a build dependency?

(From OE-Core rev: 1876548c14945c317d4f0a3a2b8df44ea1ed019f)

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

gst-fluendo*: add --disable-debug only to gst-fluendo-mp3

* fixes following QA warnings:
  gst-fluendo-mpegdemux-0.10.72: gst-fluendo-mpegdemux: configure was
    passed unrecognised options: --disable-debug
    [unknown-configure-option]

(From OE-Core rev: e6e20fba657e37e1d2b7440a00cfb3c395bde19f)

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

test-dependencies, insane.bbclass: improve the message

(From OE-Core rev: b73e6159aed1a9768d02d12e7894ca0f1e72be69)

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

directfb: add PACKAGECONFIG for drmkms and tiff

* fixes following floating dependencies:
  directfb/directfb/latest lost dependency on  libdrm libdrm-kms liblzma tiff

(From OE-Core rev: c8268bbafc338cf73c5488df9839a241a3cb6ab4)

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

Remove remnants of deleted "do_package_write" task.

As the do_package_write() task is listed as deleted, remove the few
remaining references.

(From OE-Core rev: 201d572ab5c57cda1b332356a3b7711bc346696e)

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

wic: Bump the version to 2.0

This represents a complete rework, and if we assume the previous
version should have been 1.0 (instead of a silly .1x version), then
this should be 2.0.

(From OE-Core rev: d263cb055333e6432454a13403eda471c514a3e6)

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

wic: Rename /mic to /wic

As well as any other stray instances of mic in the codebase that can
be removed.

We don't really need to carry around legacy naming, and the history is
in git.

(From OE-Core rev: 598b120406dc1d2b7e377bd1ab6f0acbef034b22)

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

wic: Update Disk description

Remove obsolete comment.

(From OE-Core rev: d39c76f24a50c9359840c0959fb50fc6da0960cd)

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

wic: Rename MountError

wic doesn't mount anything, so can't have a mount error; rename it to
something more appropriate.

(From OE-Core rev: e1edee656fc9c0a791c0eb62796d1afa483be34e)

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

wic: Rename PartitionedMount

It's actually a container for disks and partitions, and wic doesn't
mount anything, so rename it to match what it really is.

(From OE-Core rev: 28c5249b4df46ba3ac227d0c4dc6e545b40dbaf3)

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

wic: Update/rename configure-related code

The wic code inherited a basic image-creation flow based on
image-configuration, but wic doesn't actually configure anything, so
rename parts of the code dealing with configuration to something more
appropriate.

(From OE-Core rev: 58dc8879bba31542e2386aaaa70034621b2b1e4e)

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

wic: Update/rename install-related code

The wic code inherited a basic image-creation flow based on installing
packages, but wic doesn't actually install anything, so rename parts
of the code dealing with installing to something more appropriate.

(From OE-Core rev: b4232041534a79236eb8d8ab5c0024a0ef4da649)

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

wic: Update cleanup/unmount-related code

The wic code inherited a basic image-creation flow based on mounting
loop devices, but wic doesn't actually mount or unmount anything, so
get rid of unmount() and consolidate whatever it did do with
cleanup().

(From OE-Core rev: 1b103bf4a7c7527d4fc57bad1bbd1d5a5bbddb4a)

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

wic: Update/rename/delete mount-related code

The wic code inherited a basic image-creation flow based on mounting
loop devices, but wic doesn't actually mount anything, so rename parts
of the code dealing with mounting to something more appropriate, and
remove related unused code.

(From OE-Core rev: 94e15c18c011b0d7d71276cd4566be2417c2c6be)

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

wic: Remove Mount object

It doesn't do anything we need, so remove it and fix up callers/base
classes.

(From OE-Core rev: fb2a162d8756ab69c9c29a0715b033f18620341d)

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

wic: Remove unused conf support

Also fix up users such as imager functions.

(From OE-Core rev: eb77b9c11bd9b8dc90aacfbd5b5bc5568a233525)

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

wic: Update 'Background and Motivation' help section

Remove help text regarding the source and future intentions of the wic
codebase, since the code prompting those comments has now been mostly
removed.

(From OE-Core rev: dfa8626700269141f8d2f5be12c8758db7ca6473)

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

wic: Remove unused command versioning support

The default is F16 and there's no reason to change that, so remove
everything else.

(From OE-Core rev: e0ec12d012e568c9943614fc1190c143912180b3)

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

wic: Clean up Creator

wic doesn't use a lot of Creator, so remove it.

(From OE-Core rev: ef6bc7a3b58ba8b9c94f4c1e41c0f6d7d51c0bf5)

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

wic: Clean up PartitionedMount

wic doesn't use a lot of ParitionedMount, so remove it.

(From OE-Core rev: 0f793ff0767b0cce032c20b677974806c65610d6)

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

wic: Clean up DirectImageCreator

Remove what wic doesn't use from DirectImageCreator.

(From OE-Core rev: dbbe6d04f6c606c94ff17151aee1214e4e24ce7d)

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