grep: fix CVE-2015-1345

Backport patch to fix CVE-2015-1345. The issue was introduced with
v2.18-90-g73893ff, and version 2.5.1a is not affected.

Replace tab with spaces in SRC_URI as well.

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

gpgme: fix CVE-2014-3564

Backport patch to fix CVE-2014-3564.

http://git.gnupg.org/cgi-bin/gitweb.cgi?p=gpgme.git;a=commit;h=2cbd76f

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

security_flags: Add comment about what it does and who uses it

It was pointed out that people couldn't easily see who used this or
why so add some comments about that.

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

lttng-ust: 2.6.0 -> 2.6.1

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>

nasm: 2.11.06 -> 2.11.08

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>

gnupg: 2.1.2 -> 2.1.4

Updated dirmngr-uses-libgpg-error.patch.

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>

man-pages: 3.83 -> 4.00

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>

cracklib: 2.9.3 -> 2.9.4

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>

python-git: 1.0.0 -> 1.0.1

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>

git: 2.3.7 -> 2.4.2

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>

debianutils: 4.4 -> 4.5

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>

harfbuzz: upgrade to 0.9.40

- Optimizations and refactoring;
- Bug fixes.

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

wpa-supplicant: upgrade to 2.4

1. upgrade to 2.4
2. update the checksum, and license checksum since date in it is changed
3. Backport a patch to fix CVE-2015-1863
4. remove two deprecated patches

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

ppp: Security Advisory - CVE-2015-3310

http://web.nvd.nist.gov/view/vuln/detail?vulnId=CVE-2015-3310

Buffer overflow in the rc_mksid function in plugins/radius/util.c in
Paul's PPP Package (ppp) 2.4.6 and earlier, when the PID for pppd is
greater than 65535, allows remote attackers to cause a denial of
service (crash) via a start accounting message to the RADIUS server.

oe-core is using ppp 2.4.7, and this CVE say ppp 2.4.7 was not
effected, but I found this buggy codes are same between 2.4.6 and
2.4.7, and 2.4.7 should have this issue.

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

boost: extend libraries with log and random

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

xkeyboard-config: bring back Philippines keyboard symbols

Now that upstream has relicensed the Philippines symbols to remove the GPLv3
license, backport the patch instead of removing the file.

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

gcc-source.inc: set PATH for gnu-configize, not for cd

Setting OE's PATH for the 'cd' command has no real effect.

In the normal case it has no effect for the gnu-configize command
either (since OE's PATH is already set in the context which runs
do_preconfigure) but it may be useful when manually re-running a
failed gnu-configize commandline copied from an error log, etc.

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

qemu: fix CVE-2015-3456

Backport patch to fix qemuc CVE issue CVE-2015-3456.

Refs:
https://web.nvd.nist.gov/view/vuln/detail?vulnId=CVE-2015-3456
http://git.qemu.org/?p=qemu.git;a=commit;h=e907746266721f305d67bc0718795fedee2e824c

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

apt: upgrade to 1.0.9.9

1. Upgrade to fix the several CVEs: CVE-2014-0488, CVE-2014-0490
2. Remove apt-0.9.9.4-CVE-2014-0478.patch, which was backport.
3. Romve no-ko-translation.patch, apt-1.0.9.9 has ko translation
4. Update use-host.patch no-curl.patch db_linking_hack.patch and
noconfigure.patch
5. Not build the test cases since it requires gtest
6. install libapt-private.so.* to libdir, otherwise this file is
not installed into sysroot for native, and apt-get will use host's,
and lead to fail
7. Revert apt commit[a2a75ff45]"always run 'dpkg --configure -a'
at the end of our dpkg callings" for native package, otherwise
the postscript for these installed packages will be run, and fail
since the rootfs dir is not considered
8. Add lzma dependency by PACKAGECONFIG for target, and add xz
dependency for native
9. Support to compile apt-native on centos6

Signed-off-by: Roy Li <rongqing.li@windriver.com>
Acked-by: Aníbal Limón <anibal.limon@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

waffle: Upgrade 1.3.0 -> 1.5.1

* Update SRC_URI to waffle-gl.org
* Add new cmake files to -dev package
* Inherit lib_package so the new wflinfo binary is packaged properly
* Fix PACKAGECONFIG line for gbm
* Add the not-found configuration for all PACKAGECONFIG options to make
  builds reproducible (the default builds if dependencies are there)

Signed-off-by: Jussi Kukkonen <jussi.kukkonen@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

wic: refactored processing of wic exceptions

All wic exceptions are now inherited from new base exception
class WicError. It makes them easy to maintain and catch.

Processing of exceptions is done this way:
    Known wic exceptions cause wic to print error message to stdout.
    Unknown exceptions are not catched anymore and produce standard
    python traceback.

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

wic: do not strip traceback

Printing only first 5 levels of wic traceback makes it almost
useless as the most valuable part of it is stripped.

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

update-rc.d: Improve RRECOMMENDS handling

Unfortunately the combination of:

RRECOMMENDS_${PN} = "X"
UPDATERCPN = "${PN}"
RRECOMMENDS_${UPDATERCPN}_append = "Y"

is tricky for bitbake to order correctly since RRECOMMENDS_${UPDATERCPN} can
become "Y" which can then completely overwrite RRECOMMENDS_${PN}.

Avoid these issues and improve handling in general by explictly setting
the RRECOMMENDS on the list of packages modified in the general
code.

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

bitbake.conf: Use immediate expansion for os.uname()

Use immediate expansion for BUILD_ARCH and BUILD_OS since there is no
point in repeatedly calling os.uname() throughout parsing. This is
worth around 2% of parsing time, small but measurable.

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

avahi: Fix key expansion variable masking

bitbake -b avahi_ -e | grep FILES_avahi=

shows this code from http://git.yoctoproject.org/cgit.cgi/poky/commit/?h=master&id=093149d22461a3a76980635bc46cdba1c7c0b181
doesn't do what is expected. This is due to key expansion. Change to use ${PN}
to avoid warnings with new versions of bitbake.

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

kernel: Build uImage only when really needed

Build the uImage file using the kernel build system only when
it is really required, which is only in case KEEPUIMAGE == yes.
Otherwise, just build zImage, since the Yocto build system will
handle the uImage generation for us.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Richard Purdie <richard.purdie@linuxfoundation.org>
Cc: Koen Kooi <koen@dominion.thruhere.net>
Cc: Paul Eggleton <paul.eggleton@linux.intel.com>
Cc: Ross Burton <ross.burton@intel.com>
Cc: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

kernel: Add basic fitImage support

This patch adds support for generating a kernel fitImage, which is
a a successor to the uImage format. Unlike uImage, which could only
contain the kernel image itself, the fitImage can contain all kinds
of artifacts, like the kernel image, device tree blobs, initramfs
images, binary firmwares etc. Furthermore, the fitImage supports
different kinds of checksums, not only CRC32 like the uImage did.
Last, but not least, fitImage supports signatures such that either
the whole image or it's parts can be signed and then in turn can
be verified by the bootloader.

So far we only add support for wrapping the kernel image and DTB
into the fitImage. The fitImage uses the sha1 checksum, which is
the default.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Richard Purdie <richard.purdie@linuxfoundation.org>
Cc: Koen Kooi <koen@dominion.thruhere.net>
Cc: Paul Eggleton <paul.eggleton@linux.intel.com>
Cc: Ross Burton <ross.burton@intel.com>
Cc: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

kernel: Build DTBs early

Pull out the compilation of the DTB blobs right after the kernel's
own do_compile function finishes. This makes them available just in
time for the kernel image construction functions.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Richard Purdie <richard.purdie@linuxfoundation.org>
Cc: Koen Kooi <koen@dominion.thruhere.net>
Cc: Paul Eggleton <paul.eggleton@linux.intel.com>
Cc: Ross Burton <ross.burton@intel.com>
Cc: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

kernel: Separate out uboot_prep_kimage

Separate the function which prepares the kernel for packing into
uImage into separate class, so this function can be reused by the
fitImage class.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Richard Purdie <richard.purdie@linuxfoundation.org>
Cc: Koen Kooi <koen@dominion.thruhere.net>
Cc: Paul Eggleton <paul.eggleton@linux.intel.com>
Cc: Ross Burton <ross.burton@intel.com>
Cc: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

kernel: Pull uImage generation into separate class

Pull the uImage image format generation from kernel.bbclass into
a separate kernel-uimage.bbclass. Introduce new KERNEL_CLASSES
variable, which allows registration of additional classes which
implement new kernel image types. The default value of is to
register kernel-uimage to preserve the original behavior.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Richard Purdie <richard.purdie@linuxfoundation.org>
Cc: Koen Kooi <koen@dominion.thruhere.net>
Cc: Paul Eggleton <paul.eggleton@linux.intel.com>
Cc: Ross Burton <ross.burton@intel.com>
Cc: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

kernel: Pull out the linux.bin generation

Pull the generation of linux.bin image, which is then packed into uImage,
into a separate function. No functional change.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Richard Purdie <richard.purdie@linuxfoundation.org>
Cc: Koen Kooi <koen@dominion.thruhere.net>
Cc: Paul Eggleton <paul.eggleton@linux.intel.com>
Cc: Ross Burton <ross.burton@intel.com>
Cc: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

kernel: Rework do_uboot_mkimage

Rework the function so part it's internals can be re-used by fitImage
image type. The name of the temporary file , linux.bin , is recycled
a little more as it's now used for both the case where it is gzip
compressed and where it is not. This should be fine, since the file
is temporary and removed after the uImage was created anyway.

There is no functional change here.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Richard Purdie <richard.purdie@linuxfoundation.org>
Cc: Koen Kooi <koen@dominion.thruhere.net>
Cc: Paul Eggleton <paul.eggleton@linux.intel.com>
Cc: Ross Burton <ross.burton@intel.com>
Cc: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

kernel: Clean up KERNEL_IMAGETYPE_FOR_MAKE

Remove the lambda function setting KERNEL_IMAGETYPE_FOR_MAKE and instead
set it in the anonymous python function. This also allows us to handle
image types which are not supported directly by kernel, but require some
other kernel target to be built. This is the case for example with the
fitImage, which is the uImage successor.

There is no functional change.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Richard Purdie <richard.purdie@linuxfoundation.org>
Cc: Koen Kooi <koen@dominion.thruhere.net>
Cc: Paul Eggleton <paul.eggleton@linux.intel.com>
Cc: Ross Burton <ross.burton@intel.com>
Cc: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

grep: Fix LSB NG Cases for gplv2 version.

-/tset/LI18NUX2K.L1/utils/egrep-tp/T.egrep-tp 5
-/tset/LI18NUX2K.L1/utils/fgrep/T.fgrep 5
-/tset/LI18NUX2K.L1/utils/grep-tp/T.grep-tp 5

The LSB core test requires grep egrep and fgrep can
perform pattern matching in searches without regard
to case if -i option is specified.

Signed-off-by: Li Xin <lixin.fnst@cn.fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

native: Improve PROVIDES handling

Since this class works on pre finalised data, the logic hasn't needed
to be exact. If we change the way the finalised data works, we find
that certain dependencies can be dropped (e.g. pn isn't in the name).

To fix this, restructure the function to alter each entry in turn
and not drop any entries.

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

autotools: Use space with prepend to clean up variable whitespace

Correct the whitespace in this variable to make this more readable when
debugging, no functional change.

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

packagegroup-core-x11-sato: add PulseAudio

This enables PulseAudio in the Sato images.

[YOCTO #7517]

Signed-off-by: Tanu Kaskinen <tanu.kaskinen@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

pulseaudio-client-conf-sato: initial recipe

This recipe is related to enabling PulseAudio in Sato images. We will
rely on PulseAudio's autospawn feature to automatically start the
daemon. The graphical session in Sato runs under root, however, and
PulseAudio disables autospawning for root by default. We provide a
client.conf fragment in Sato to change that default setting.

Signed-off-by: Tanu Kaskinen <tanu.kaskinen@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

pulseaudio: conf-parser: add support for .d directories

Sato images should set allow-autospawn-for-root=true in
/etc/pulse/client.conf, but in non-Sato images that option should be
disabled by default. I first tried to have two packages that ship
different versions of the client.conf file, but it turned out to be
tricky to ensure that the package manager always chooses the package
that provides the default version when the Sato package is not
explicitly requested. This patch allows the Sato specific
configuration to be installed in /etc/pulse/client.conf.d without
replacing the default version of client.conf, which makes packaging
much simpler.

Signed-off-by: Tanu Kaskinen <tanu.kaskinen@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

pulseaudio: client-conf: Add allow-autospawn-for-root

This is related to enabling PulseAudio in Sato images. Sato doesn't
have regular users; the graphical session is run as root. PulseAudio
disables autospawning for root, but in Sato that's not the desired
behaviour. This patch allows autospawning to be enabled for root in
systems where that makes sense.

Signed-off-by: Tanu Kaskinen <tanu.kaskinen@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

alsa-plugins: initial recipe

When PulseAudio is installed, alsa-plugins is needed to provide
compatibility for ALSA applications.

The dependencies are set up so that pulseaudio-server depends on
alsa-plugins-pulseaudio-conf, and alsa-plugins-pulseaudio-conf depends
on the PulseAudio plugins. This should make the ALSA->PulseAudio
compatibility configuration work out of the box, while leaving the
PulseAudio specific configuration out on systems that don't have
pulseaudio-server installed.

The alsa-plugins recipes in meta-guacamayo and meta-webos were used as
references, but this recipe is not a straight copy of either.

Signed-off-by: Tanu Kaskinen <tanu.kaskinen@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

eglibc-use-option-groups: Conditionally exclude c++ tests

Some test programs written in c++ are still included in spite of
"libc-cxx-tests" being omitted from DISTRO_FEATURES_LIBC.
All .cc programs are compiled with g++.
g++ automatically specifies linking against the C++ library.
This patch conditionally excludes the following tests as well:

  bug-atexit3-lib.cc
  tst-cancel24.cc
  tst-cancel24-static.cc
  tst-unique3lib.cc
  tst-unique3lib2.cc
  tst-unique4lib.cc
  tst-unique3.cc
  tst-unique4.cc

Tested with DISTRO_FEATURES_LIBC_remove = " libc-cxx-tests"

[YOCTO #7003]

Signed-off-by: Juro Bystricky <juro.bystricky@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

image.bbclass: Add a method for creating a companion debug filesystem

The companion debug filesystem contains only the package database and the
complementary *-dbg packages for the main filesystem component.  This is
useful in a production environment to produce a companion filesystem capable
of remote system debugging, without requiring corresponding debug symbols or
source code on the device.

Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
If dbg pkgs have already been installed to the rootfs image,
the installation to companion debug filesystem will fail,
because both of image creation make use of the same pm
database.

In this situation, try to copy installed dbg files from rootfs
image to companion debug filesystem.

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Acked-by: Christopher Larson <chris_larson@mentor.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

openssl: Fix build with gcc5 on mips64

Patch is submitted upstream as well

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

npth: Upgrade to 1.2

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

libassuan: Upgrade to 2.2.1

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

boost/bjam-native: Upgrade to 1.58.0

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

valgrind: remove arm tests that don't compile

[Yocto #7453]

Corrects the original commit for the patch that removed ARM ptest CFLAGS
settings. Since the flags could be set by a user, the flags should
be kept in place during compilation.  By keeping the original up-stream
CFLAGS for the tests, then additional tests successfully compile
for all tested ARM tunings.

However, there were still two tests listed below that did not compile
for any beaglebone tuning that is valid for valgrind. With the updated
patch, the set of excluded ARM ptests and their respective build
failures are:
  intdiv - fails for all beaglebone tunings with 2 errors:
  {standard input}:(40 or 41): Error: selected processor does not
       support Thumb mode `udiv r3,r9,r10'
  {standard input}:(72 or 73): Error: selected processor does not
       support Thumb mode `sdiv r3,r9,r10'

  vcvt_fixed_float_VFP - fails for all beaglebone tunings in one of
  two ways:
    with neon tuning (-mfpu=neon) fails with Internal Compiler Error
    without neon tuning fails with 3 errors:
    {standard input}:33: Error: selected FPU does not support
      instruction -- `vcvt.f32.s32 s15,s15,#1'
    {standard input}:58: Error: selected FPU does not support
      instruction -- `vcvt.f32.s32 s15,s15,#32'
    {standard input}:136: Error: selected FPU does not support
      instruction -- `vcvt.f32.u32 s15,s15,#1'

After applying this commit, the valgrind ARM ptests compile without
errors for tunings:
  armv7[t][hf][b][-neon] cortexa8[t][hf][-neon]
where the tuning [option] was successfully compiled, both with
and without the 'option', and in combination with all other options.

Signed-off-by: Dave Lerner <dave.lerner@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

package.bbclass: Include missing variables on PACKAGEVARS

PACKAGEVARS appended on this commit:

    LICENSE SECTION
    pkg_preinst
    pkg_prerm
    RREPLACES
    GROUPMEMS_PARAM
    SYSTEMD_AUTO_ENABLE

[Yocto #7754]

Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

libpng: Upgrade 1.6.16 -> 1.6.17

License files changes are not actual license changes.

Signed-off-by: Jussi Kukkonen <jussi.kukkonen@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

libpng12: Upgrade 1.2.52 -> 1.2.53

License file changes are not actual license changes.

Signed-off-by: Jussi Kukkonen <jussi.kukkonen@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

oe-selftest: Build wic runtime requirements and images before testing

Some native tools (syslinux, parted, mtools, etc) are required
by wic to produce images. Unit tests fail if the tools are
not available.

Baked tools and image-core-minimal used by wic before running tests.

[YOCTO #7730]

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

gettext-minimal-native: Disable the unnecessary check in iconv.m4

Disable the test "Test against HP-UX 11.11 bug: No converter from EUC-JP
to UTF-8 is provided" since we don't support HP-UX and if the euc-jp is
not installed on the host, the dependence will be built without iconv
support and will cause guile-native building fail.

The patch is similar as 0470bd7a9658d3[libunistring: remove the test to
convert euc-jp in configure]

Signed-off-by: Roy Li <rongqing.li@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

bitbake.conf: programatically generate COMBINED_FEATURES

Instead of hard-coding the set of features that can be considered in
COMBINED_FEATURES, simply generate the intersection of DISTRO_FEATURES and
MACHINE_FEATURES.

Signed-off-by: Ross Burton <ross.burton@intel.com>

utils: add helper to perform the intersection of two string lists

Useful for e.g. generating a COMBINED_FEATURES list from DISTRO_FEATURES and
MACHINE_FEATURES.

Signed-off-by: Ross Burton <ross.burton@intel.com>

openssl: drop the padlock_conf.patch

padlock_conf.patch will enable the padlock engine by default,
but this engine does not work on some 32bit machine, and lead
to openssl unable to work

Signed-off-by: Roy Li <rongqing.li@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

packagegroup-core-qt4e: make phonon-gstreamer plugin optional

Having disabled phonon by default in Qt4 we need to ensure this
packagegroup is still buildable when it's not available.

Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

gnutls: remove --with-libz-prefix configure option

The default search path in sysroot is sufficient to find zlib, so the
--with-libz-prefix configure option seems to be unnecessary.

For target builds, relying on sysroot also prevents an absolute path
from being hardcoded in the gnutls.pc pkg-config file.

Signed-off-by: Andre McCurdy <armccurdy@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

toolchain-scripts.bbclass: fix quoting for CCACHE_PATH

Single quotes prevent expansion of $sdkpathnative$bindir.

Signed-off-by: Andre McCurdy <armccurdy@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

squashfs-tools: build and install unsquashfs as well

* it's useful for debugging corrupt squashfs images from mksquashfs

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

wic: Make sure file exists before removing it

Bunch of os.remove calls were added to the partition.py lately.
They're causing wic to fail with OSError: [Errno 2] No such file or directory
if file doesn't exist.

Added check for file existence to all recently added calls of
os.remove. That should fix this regression.

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

classes/buildhistory: ensure that git user email and name are set

The git user.email and user.name settings must be set or the commit to
the buildhistory repo (when BUILDHISTORY_COMMIT is set to "1") will
fail. If the user hasn't set these, set some dummy values since the
actual value is unlikely to be particularly useful and it's far worse
for the commit not to be done leaving the repository in a partially set
up state.

Prompted by [YOCTO #6973].

Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

initrdscripts: make boot drive detection more generic

The init script that invokes install and install-efi scripts
passes the first parameter that identifies the boot drive but
in cases when this disk is labeled and kernel configurations
allow disk labeling under /run/media/ this would pass the disk
label.
The earlier implementation considered that the drive name will
be passed and in case the label is passed it fails and provides
the boot drive as an option for installation driver.
We now use a more generic approach to identify the boot drive
which can handle both drive name as well as label if passed.

Signed-off-by: Awais Belal <awais_belal@mentor.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

buildstats-summary: add eventmask

Add eventmask 'bb.event.BuildCompleted' for buildstats_summary handler to avoid
running the codes in the handler everytime there's an event.

The codes in buildstats_summary only need to run for bb.event.BuildCompleted.
So add this event mask.

Also, we remove the redundant isinstance() statement in the handler.

Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

rootfs.py: Make set -x not break rootfs creation

The rootfs logfile check errors out if it finds an "exit 1" in
a log. But consider the shell idiom:
if ! $command; then exit 1; fi
and consider a postinstall script with a "set -x" for debugging
purposes (to get log output in /var/log/postinstall.log.)

Solution: Ignore lines prefixed with a +, because those show shell
code even if a specific fragment won't be executed.

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

libepoxy: make buildable for x11-less environments / update to latest version

* Make glx part configurable by PACKAGECONFIG enabled by 'x11' distro feature
* Major upstream enhancement is using pkgconfig for egl. This should make
  those egl implementations happy that require compiler extra options for their
  egl headers.
  Example: meta-fsl-arm/imx-gpu-viv sets -DLINUX (and others) in egl.pc. These
  are mandatory for proper builds.

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

linux-yocto/3.19: NFC config, Braswell fixes and axxia support

Updating the 3.19 kernel SRCREVs for the following commits:

NFC:

   727cfce45077 meta: intel-common: enable generic NFC support
   f3e890065903 meta: features/nfc: add generic configuration for NFC support

Braswell:

   d3c0b958e688 drm/i915: New offset for reading frequencies on CHV.
   dac6babbba62 drm/i915/chv: Populate total EU count on Cherryview

Axxia:

   c706cb356dea arch/arm/axxia: Remove the axxia zImage.fm build
   45a1aaacf6cb drivers/hwmon: Remove adt75 redundant driver
   7203ee166c01 arch/arm/mach-axxia: Fixed L2 power up failure
   a1541e9f0a4c kernel/irq/manage.c: Fix irq_set_affinity to allow use with buslocks
   d42c0bd1c8be arch/arm/mach-axxia: Reverse checkpatch compatibility
   5dbd07431a7a arch/arm/mach-axxia: changed affinity parameter to cpu
   23e4ebc66acb char: hwrng: AXXIA TRNG driver added for AXM55xx
   61eb3c8ee1d7 arch/arm/mach-axxia: Updated Axxia 55xx PCIe driver to use correct Doorbell interrupt IRQs
   b03e0655094a AXM55xx RapidIO : Added support for Device revision and link down Monitor.
   7577ad26cf73 rionet: Fix the corruption of tx_cnt during Transmit.
   07fd2163ae56 arch/arm/mach-axxia: fixed NO SMP
   c9371e98cc80 arch/arm/mach-axxia: fixed compiler warning
   0a814fe7e747 arch/arm/mach-axxia: Updated PCIe driver doorbell support
   cd3d9b46bea4 misc: lsi-ncr: Only use AMP lock on PPC platforms.
   826c600c6ad6 fs: vmfs: Use generic mmap function
   e2d0047cb106 arch/arm/mach-axxia: Removed axxia_circular_queue
   f1ff06eaabed ARM: axxia: Add cluster and L2 power off in hotplug
   b98711344263 ARM: irq: Return error when set_affinity is called from hotplug
   9fd9306a30ae arch/arm/mach-axxia: Enable L3
   68989d446582 Revert "arch/arm/mach-axxia: Fixed affinity error on IRQ migration"
   0be0ef90a887 Revert "arch/arm/mach-axxia: fixed failure to power up cluster"
   1c6522a03e81 Revert "arch/arm/mach-axxia: fixed hang in preempt kernel"
   ef06a68393a3 fs/vmfs: Changes to add VMFS support for axxia.
   3025e9deede6 drivers/usb/host: Changes to support the axxia BSP
   b128b4f0d4e8 drivers/tty: Changes to support the axxia BSP
   d1b60a7e62ef drivers/hwmon: Changes made to support the axxia BSP
   2899ccbf1438 drivers/spi: Changes to support the axxia BSP
   8e9cff566254 drivers/rapidio/devices: Changes to support axxia BSP
   3424d2ca2653 drivers/net/ethernet: Changes to support the axxia BSP
   2353dc816d6b drivers/mtd: Changes to support the axxia BSP
   4b5fa2bf86cf drivers/misc: Changes made to support axxia BSP
   8a5bd53ec241 drivers/i2c: Changes to support axxia BSP
   02bfe1548d1a arch/arm/drivers/edac: Changes to support axxia BSP
   2431db7c0c90 arch/arm/drivers/dma: Changes to support the axxia BSP
   61fa3f268d5c arch/arm: arm changes to support the axxia BSP
   22e0fb7be665 arch/arm/mach-axxia: add power management support
   5cab63c5a1f2 arch/arm/mach-axxia: kernel files to support the mach-axxia
   1336aba51616 arch/arm/boot: Changes to support the axxia BSP
   97324fa920cb arch/arm/boot/dts: Files added to support axxia 5500 board

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>

linux-yocto/3.19: CONFIG_ATA_BMDMA and gcc5 ARM64

Integrating the following commit to remove a ATA configuration warning:

    common-pc-drivers: Enable CONFIG_ATA_BMDMA

    Enabling the ATA_BMDMA re-enables ATA_PIIX which was getting
    disabled

Signed-off-by: Saul Wold <sgw@linux.intel.com>
And the following to fix the ARM64 gcc5 boot:

    arm64: psci: move psci firmware calls out of line

    An arm64 allmodconfig fails to build with GCC 5 due to __asmeq
    assertions in the PSCI firmware calling code firing due to mcount
    preambles breaking our assumptions about register allocation of
    function
    arguments:

      /tmp/ccDqJsJ6.s: Assembler messages:
      /tmp/ccDqJsJ6.s:60: Error: .err encountered
      /tmp/ccDqJsJ6.s:61: Error: .err encountered
      /tmp/ccDqJsJ6.s:62: Error: .err encountered
      /tmp/ccDqJsJ6.s:99: Error: .err encountered
      /tmp/ccDqJsJ6.s:100: Error: .err encountered
      /tmp/ccDqJsJ6.s:101: Error: .err encountered

    This patch fixes the issue by moving the PSCI calls out-of-line into
    their own assembly files, which are safe from the compiler's
    meddling
    fingers.

Reported-by: Andy Whitcroft <apw@canonical.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Armin Kuter <akuster@mvista.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>

linux-yocto/3.19: Braswell DRM fixes

Updating the 3.19 kernel SRCREVs to integrate the following Braswell
changes:

  374b5d0e09ea drm/i915: Only wait for required lanes in vlv_wait_port_ready()
  fca99e8ee111 Revert "drm/i915: Hack to tie both common lanes together on chv"
  00682f31b612 drm/i915: Work around DISPLAY_PHY_CONTROL register corruption on CHV
  654b1a4497c5 drm/i915: Implement chv display PHY lane stagger setup

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>

linux-yocto: drop suggestion of devshell for patch failures

When a patch fails to apply, the kernel-yocto bbclass attempted to
be helpful and suggest that devshell be used to fix the issue.

The only problem is that you can't get to devshell if a patch is
failing.

We drop this bad advise and instead point to the linux source directory.

[YOCTO: #6202]

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>

linux-yocto-custom: clarify bbappend versus copy in comments

The kernel development manual recommends making a copy of
linux-yocto-custom, while the comments in the recipe talk about
bbappends.

To synchronize the two, we remove the bbappend references from the
recipe and instead indicate that a copy should be updated.

[YOCTO: #6925]

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>

linux-yocto/3.19: braswell DRM/i915 Graphics fixes

Updating the SRCREVs to integrate the following fixes:

   0befa35f4099 drm/i915/chv: Remove DPIO force latency causing interpair skew issue
   184e0374e4eb drm/i915: Fix chv cdclk support
   e2a99b9cd086 drm/i915: Increase the range of sideband address.
   9d5d55ede53b drm/i915: Disable DDR DVFS on CHV
   96cce945ac97 drm/i915: Enable the maxfifo PM5 mode when appropriate on CHV
   b5005319da56 drm/i915: Program PFI credits for VLV
   c7aa33eb3697 drm/i915: Rewrite VLV/CHV watermark code
   a421d8bcaa6d drm/i915: Make sure PND deadline mode is enabled on VLV/CHV
   631afc98c4b5 drm/i915: Read out display FIFO size on VLV/CHV
   e0dcdc019b8a drm/i915: Pass plane to vlv_compute_drain_latency()
   a6a5562b7754 drm/i915: Reorganize VLV DDL setup
   bb662a47ec0c drm/i915: Hide VLV DDL precision handling
   3d2d93239cd7 drm/i915: Simplify VLV drain latency computation
   f68614743b92 drm/i915: Kill DRAIN_LATENCY_PRECISION_* defines
   86c658c06ede drm/i915: Reduce CHV DDL multiplier to 16/8
   8c4cdd96a3f2 drm/i915: Allow pixel clock up to 95% of cdclk on CHV
   d9d4fb889c2a drm/i915: Reduce CHV DPLL min vco frequency to 4.8 GHz

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>

linux-yocto/3.14: cavium and configuration changes

Updating the 3.14 SRCREVs to integrate the latest round of cavium
support patches, as well as configuration backports from 3.19.

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>

xvinfo: upgrade to 1.1.3

Signed-off-by: Ross Burton <ross.burton@intel.com>

xkeyboard-config: upgrade to 2.14

Signed-off-by: Ross Burton <ross.burton@intel.com>

xdpyinfo: upgrade to 1.3.2

Signed-off-by: Ross Burton <ross.burton@intel.com>

puzzles: upgrade to latest srcrev

Signed-off-by: Ross Burton <ross.burton@intel.com>

sanity.bbclass: import subprocess in correct function

check_sanity() no longer needs the subprocess module but
sanity_handle_abichanges() does use subprocess.call().

Signed-off-by: Jussi Kukkonen <jussi.kukkonen@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

license_class: Add explicity avoid of CLOSED license in validations

The CLOSED license isn't a generic license it is a set and can
be any closed source license.

[YOCTO #7752]

Signed-off-by: Aníbal Limón <anibal.limon@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

gtk-engines: Make gtk-engines-dev installable by dropping dependency to not generated gtk-engines package.

Signed-off-by: Ross Burton <ross.burton@intel.com>

rootfs.py: Improve rpm log_check_regex

Current regex can cause false negatives if paths in bitbake log files
contain "ERR" or "Fail". do_rootfs fails with return code 1 in
this case.

Improved regexp is based on error messages produced by rpm.
Those are found by analyzing rpm source code.

[YOCTO #7789]

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

zlib: clean up base_libdir symlink

libz.so symbolic link created in ${libdir} is
../../${base_libdir}/libz.so.1.2.8.  This doesn't work if base_libdir or libdir
is changed, so use oe.path.relative to construct the correct path at build time.

Signed-off-by: Ross Burton <ross.burton@intel.com>

combo-layer: handle unset dest_dir in sanity_check()

The previous "clean up dest_dir checking" patch (f8cdbe7497) improved
handling of empty dest_dir but made handling of unset dest_dir worse:
instead showing the "Option dest_dir is not defined for component ..."
error, it fails with a Python exception.

Avoid that by providing a sane fallback for the unset case. With that
change, dest_dir is no longer strictly required, but the check for it
is kept to ensure that a combo-layer.conf also works with older
combo-layer versions.

[Yocto #7773]

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

libxdmcp: make libbsd usage deterministic

libxdmcp can optionally use arc4random_buf() from libbsd instead of it's own
random key generation code. As this is an optional dependency add a
PACKAGECONFIG for it and default to not using libbsd.

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

sstate: Add handling of do_shared_workdir task

Changing TMPDIR and rebuilding an image was resulting in rebuilds of
kernels due to dependencies on the shared_workdir task. If installed
from sstate, nothing needs this task so add it to the whitelisted
task patterns.

After this change, the kernel does not rebuild when a new TMPDIR and
hot sstate cache is used.

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

sqlite3: upgrade to 3.8.10

upgrade to include CVE fixes:
    CVE-2015-3414
    CVE-2015-3415
    CVE-2015-3416

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

midori: fix build with GLib 2.44

GLib 2.44 introduced a GListStore type which is ambiguous with GtkListStore when
using wildcard using statements.  Use the full type name to fix builds.

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

testapps: remove old gst-meta-audio/video dependencies

Unnecessary because gst-player already pulls in correct gstreamer packages anyway.

Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

sanity.bbclass: improve the sanity check for PATH and BBPATH

Build fails if "./" is in the PATH

To reproduce the issue:

  $ export PATH="./:$PATH"
  $ make -C build-tools elfutils

  Error: ./ld: unrecognized option '--sysroot=/'

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

pinentry: package update to 9.2

added a few more PACKAGECONIG options.

Signed-off-by: Armin Kuster <akuster808@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

alsa-utils: upgrade to 1.0.29

Upgrade alsa-utils from 1.0.28 to 1.0.29.

* remove PR
* rename package alsa-utils-alsaconf to alsa-utils-scripts, and add
  script alsa-info.sh

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

alsa-tools: upgrade to version 1.0.29

Upgrade alsa-tools from verion 1.0.28 to 1.0.29.

* update autotools.patch
* add file gitcompile which exists in git repo but missing in release
  tar ball for sub-directory hdajacksensetest

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

vala: upgrade to version 0.28.0

Upgrade vala from version 0.26.1 to 0.28.0.

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

gstreamer: remove 0.10 recipes

gstreamer 0.10 is still available from meta-openembedded/meta-multimedia

See also
https://bugzilla.yoctoproject.org/show_bug.cgi?id=6294
http://lists.openembedded.org/pipermail/openembedded-core/2015-April/104276.html

Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

webkitgtk: disable audio and video support

This allows removal of gstreamer 0.10 dependency from webkit recipe

https://bugzilla.yoctoproject.org/show_bug.cgi?id=6294

Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

qt4: disable Phonon framework

This allows removal of gstreamer 0.10 dependency from Qt4 recipe

https://bugzilla.yoctoproject.org/show_bug.cgi?id=6294
Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

qt4: remove Qt Mobility

Qt Mobility is an obsolete and unmaintained framework that is also
dependent on gstreamer 0.10

https://bugzilla.yoctoproject.org/show_bug.cgi?id=6294

Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

clutter-gst: update to latest upstream version 3.0.4

This also makes it GStreamer 1.0 compatible

https://bugzilla.yoctoproject.org/show_bug.cgi?id=6294

Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

python-gst: remove obsolete Python gstreamer 0.10 bindings

GStreamer 1.0 python bindings are gobject-introspection based,
so until that is fixed, they will not be available.

https://bugzilla.yoctoproject.org/show_bug.cgi?id=6294

Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

icu-native: fix hard paths in native icu

The icu recipe installs icu-native twice, once in the usual location,
and once for cross builds into the path given by --with-cross-build.

This latter path is not included in the list of paths recognised by
chrpath.bbclass, hence the binaries in there retain the rpath as used
during compilation. This causes the package to not be relocateable
from sstate
[YOCTO #6851]

We use the infrastructure that is in place already, and simply set
PREPROCESS_RELOCATE_DIRS as necessary, and things start to work.

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