libpcre: 8.36 -> 8.37

The LICENSE's md5sum has changed mainly because the new version added
this line:
The data in the testdata directory is not copyrighted and is in the
public domain.

The license is the same, so just update the md5sum.

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

coreutils_6.9.bb: Fix LSB NG Cases.

-/tset/LI18NUX2K.L1/utils/sort/T.sort #1/3/17/19/33/35
-/tset/LI18NUX2K.L1/utils/comm/T.comm #1/2

The LSB test require comm can handle multibyte characters and
sort can interpret blank and alphanumeric character according
to the current locale and so on.

Signed-off-by: Li Xin <lixin.fnst@cn.fujitsu.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

bind: separate B and S

Fix B != S and separate them.

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

xf86-video-cirrus: Upgrade 1.5.2 -> 1.5.3

Signed-off-by: Jussi Kukkonen <jussi.kukkonen@intel.com>

packagegroup-core-qt: add x11 to REQUIRED_DISTRO_FEATURES

It can't be built withou x11 in DISTRO_FEATURES.

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

at-spi2-atk: add x11 to REQUIRED_DISTRO_FEATURES

It can't be built withou x11 in DISTRO_FEATURES.

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

recipes-sato: add x11 to REQUIRED_DISTRO_FEATURES

They can't be built withou x11 in DISTRO_FEATURES.

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

gst-plugins-gl: add x11 to REQUIRED_DISTRO_FEATURES

It can't be built withou x11 in DISTRO_FEATURES.

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

recipes-graphics: add x11 to REQUIRED_DISTRO_FEATURES

They can't be built withou x11 in DISTRO_FEATURES.

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

recipes-extended: add x11 to REQUIRED_DISTRO_FEATURES

They can't be built withou x11 in DISTRO_FEATURES.

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

libowl/settings-daemon/libnotify: add x11 to REQUIRED_DISTRO_FEATURES

The libowl, settings-daemon and libnotify requires gdk/gdkx.h which is
provided by gtk when x11 in DISTRO_FEATURES, so add x11 to
REQUIRED_DISTRO_FEATURES.

The leafpad, settings-daemon and oh-puzzles requires
libowl or settings-daemon, so add xx to REQUIRED_DISTRO_FEATURES for
them too.

The leafpad can't be built without libowl, so depends in directly rather
than use DEPENDS_append_poky.

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

libsdl: depends on libglu when x11

Let it depend on libglu when x11 rather than opengl, since libglu can't
be built without x11.

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

libglu: add x11 to REQUIRED_DISTRO_FEATURES

It requires libGL.so which is provided by mesa when x11 in
DISTRO_FEATURES.

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

mesa-demos: only enable glu when x11 in DISTRO_FEATURES

Fixed when no x11 in DISTRO_FEATURES:
checking for GL/glu.h... no
configure: error: GLU not found

The GL/glu.h is provided by libglu, and libglu requires libGL.so which
is provided by mesa, but mesa doesn't build out libGL.so without x11 in
DISTRO_FEATURES, so only enable glu when x11 in DISTRO_FEATURES.

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

mesa: fix do_install_append

"ifdef MESA_EGL_NO_X11_HEADERS" -> "if defined(MESA_EGL_NO_X11_HEADERS)"

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

libxml2: fix python path and add libxml2-python

We have libxml2-python for native and nativesdk, but don't have it for
target, and can't find the reason from the git log, libxml2-python is
widely used, after looked into it's configure.in, we can add it
by PACKAGECONFIG.

The previous --with-python=${STAGING_BINDIR}/python is incorrect, it
acted as work becase it's conigure can check automatically, python is in
${STAGING_BINDIR_NATIVE}/${PYTHON_PN}-native/${PYTHON_PN}, as known as
${PYTHON}.

Add python to PACKAGECONFIG, since createrepo rdepends on
libxml2-python, otherwise the target createrepo can't work.

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

ncurses: inherit pkgconfig

Without pkgconfig being present at build time it's possible for the .pc files to
not be installed correctly.

Signed-off-by: Christophe Guillon <christophe.guillon@st.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

qt4-x11-free: make virtual/libgl dependency conditional

Make qt4-x11-free depend on virtual/libgl only if the DISTRO_FEATURES
contains opengl. Otherwise there will be no package to fullfil this
dependency.

Signed-off-by: Dmitry Eremin-Solenikov <dmitry_eremin@mentor.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

sanity.bbclass: Increased verbosity for connectivity check

The connectivity sanity error doesn't tell you which URL failed to fetch nor how
it failed.

This provides the URL that failed and why it failed using BBFetchException
messages during the connectivity check.

[YOCTO #7592]

Signed-off-by: Mariano Lopez <mariano.lopez@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

unzip: Security Advisory -CVE-2014-9636 and CVE-2015-1315

http://web.nvd.nist.gov/view/vuln/detail?vulnId=CVE-2014-9636

unzip 6.0 allows remote attackers to cause a denial of service
(out-of-bounds read or write and crash) via an extra field with
an uncompressed size smaller than the compressed field size in a
zip archive that advertises STORED method compression.

http://web.nvd.nist.gov/view/vuln/detail?vulnId=CVE-2015-1315

Buffer overflow in the charset_to_intern function in unix/unix.c in
Info-Zip UnZip 6.10b allows remote attackers to execute arbitrary code
via a crafted string, as demonstrated by converting a string from CP866
to UTF-8.

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

kernelshark: fix for rebuild

Fixed when rebuild:
| NOTE: make prefix=/usr bindir_relative=bin libdir=lib NO_PYTHON=1 gui clean
| make: *** No rule to make target `sysroots/x86_64-linux/usr/lib/i586-poky-linux/gcc/i586-poky-linux/4.9.2/include/stddef.h', needed by `event-parse.o'.  Stop.
| ERROR: oe_runmake failed

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

gcc-5: fix installed-vs-shipped

gcc-5.1.0: gcc: Files/directories were installed but not shipped in any package:
  /usr/bin/i586-poky-linux-gcov-tool

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

module/module-base/lttng-modules: Improve kernel module dependencies

If we have DEPENDS = "virtual/kernel" is means that the kernel module
depends on the kernel's do_populate_sysroot task. This is not entirely
desireable since that depends on do_install which depends on
do_compile_kernelmodules and so on. In a situation where rm_work in involved
this can cause some pretty length build cycles after the kernel workdir
has been cleaned up by rm_work.

As well as removing this, take the opportunity to clean up duplicated
dependency lines, tweak the dependency of make_scripts for the same
reason and generally try and make things more readable.

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

openssl/orc: virtclass-native -> class-native

The virtclass-native is out of date.

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

stat: fix SRC_URI

The old SRC_URI is redirected to the new one, fixed:
WARNING: Failed to fetch URL ftp://metalab.unc.edu/pub/Linux/utils/file/stat-3.3.tar.gz, attempting MIRRORS if available

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

elfutils: 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 guild-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: Richard Purdie <richard.purdie@linuxfoundation.org>

gcc5: Add back g++ sysroot patch

Without this, g++/c++ compilation doesn't work on target due to missing
header files. Automated sanity tests fail. Add back the gcc4 patch to
address this.

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

gcc5: Add PR65779 patch to fix powerpc compile issues

This fixes compile issues on powerpc with gcc 5 which show up with
errors like:

| make[2]: Entering directory '/media/build1/poky/build/tmp/work/ppc7400-poky-linux/xprop/1_1.2.2-r0/build'
| powerpc-poky-linux-gcc  -m32 -mhard-float -mcpu=7400 --sysroot=/media/build1/poky/build/tmp/sysroots/qemuppc -Wall -Wpointer-arith -Wmissing-declarations -Wformat=2 -Wstrict-prototypes -Wmissing-prototypes -Wnested-externs -Wbad-function-cast -Wold-style-definition -Wdeclaration-after-statement -Wunused -Wuninitialized -Wshadow -Wmissing-noreturn -Wmissing-format-attribute -Wredundant-decls -Wlogical-op -Werror=implicit -Werror=nonnull -Werror=init-self -Werror=main -Werror=missing-braces -Werror=sequence-point -Werror=return-type -Werror=trigraphs -Werror=array-bounds -Werror=write-strings -Werror=address -Werror=int-to-pointer-cast -Werror=pointer-to-int-cast -fno-strict-aliasing  -O2 -pipe -g -feliminate-unused-debug-types  -Wl,-O1 -Wl,--hash-style=gnu -Wl,--as-needed -o xprop dsimple.o clientwin.o xprop.o -lX11
| /media/build1/poky/build/tmp/sysroots/qemuppc/usr/lib/../lib/libX11.so: undefined reference to `.LCL2'
| collect2: error: ld returned 1 exit status

[YOCTO #7721]

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

insserv: Remove

Remove insserv from OE-Core. It did have uses but we've optimised the sysvinit
scripts as needed and there are other directions init systems are moving now.
It no longer belongs in the core.

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

dbus: Fix do_install files outside ${D}

do_install should not access the sysroot directly. Add missing ${D}
reference.

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

license_class: license_create_manifest improvment

Use image_list_installed_packges instead of open manually the package
manifest and iterate over it.

Signed-off-by: Aníbal Limón <anibal.limon@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

license: Add support for handle INCOMPATIBLE_LICENSE in manifest creation

When INCOMPATIBLE_LICENSE's is specified it need to be removed from
license.manifest and also avoid copy to target image.

Add ManifestVisitor that walk the license string searching for
INCOMPATIBLE_LICENSE's if found remove it.

[YOCTO #6765]

Signed-off-by: Aníbal Limón <anibal.limon@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

license: Split visit_string in LicenseVisitor

Create get_elements and visit_elements in LicenseVisitor based
on visit_string this allow to do modifications on elements before
parsing with AST.

Signed-off-by: Aníbal Limón <anibal.limon@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

license_class: Generalize license_ok function

Add dont_want_licenses as parameter to license_ok function and move it
to oe.license module in order to use in other modules.

Signed-off-by: Aníbal Limón <anibal.limon@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

license_class: Reimplemented manifest creation in python

Reimplemented license_manifest_create from shell to python for
INCOMPATIBLE_LICENSE handle using oe.license module.

Optimizations are made to avoid license copy now uses a hardlink
and symbolic link this helps to save space during build.

Signed-off-by: Aníbal Limón <anibal.limon@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

distro_features_check: add any of test

Add a test for distro features including one or more
items in a list.  This is useful when, for example, we
need either x11 or directfb as a feature.

Signed-off-by: Joe Slater <jslater@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

combo-layer: improve merge commit handling

When the head of a branch is a merge commit, combo-layer did not
record that commit as last_revision because it only considers applied
patches, and the merge commit never gets applied.

This causes problems when the merge commit leads to multiple patches
and the commit id that gets recorded only reaches some of these
patches. The next run then will try to re-apply the other patches.

This special case is now detected and dealt with by bumping
last_revision to the branch commit. The behavior where the head is a
normal commit is intentionally not changed, because some users might
prefer the traditional behavior.

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

weston-init: support system's configuration file

Look for OPTARGS variable in /etc/default/weston
and set it as weston's service default options.

This can be used to force system's supported backend.

Change-Id: I0562c9326df5b46226093199873ef58d77aeae75
Signed-off-by: Philippe Coval <philippe.coval@open.eurogiciel.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

dbus: upgrade to 1.8.16

Upgrade to fix CVE-2015-0245

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

gcc-4.9, gcc-5: Use variable SYSTEMLIBS_DIR instead of hardcoding it for aarch64

Change-Id: I54dc82a569f02d489137d88f16d6b768c4ab779b
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gcc: Add 5 recipes

Drop backports obviously
additionally drop local patches around
parallel compilation since it got reworked upstream

To select set

GCCVERSION = "5.1%"
SDKGCCVERSION = "5.1%"

in local.conf

Change-Id: Icdfa61017a617244128f361b022e7c8f1f9e0610
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

kernel-yocto: propagate in tree defconfigs to WORKDIR

As reported by Steffen Pankratz <Steffen.Pankratz@elektrobit.com>, the
previous logic of KBUILD_DEFCONFIG processing would not propagate an in
tree defcofig to WORKDIR if one was not already present.

We fix the propagation by copying the in tee config if a defconfig is
not already in WORKDIR.

Additionally we only warn (versus copying) if an in tree configuration
is specified, is different than the WORKDIR version and isn't copied.

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

linux-yocto/3.14: fix qemumips build error

Updating the SRCREVs to import the following fix:

    mips: define cpu_has_saa in common features include

    To avoid build failures such as the following on non-cavium
    platforms:

     | arch/mips/include/asm/atomic.h: In function 'atomic_add':
     | arch/mips/include/asm/atomic.h:52:6: error: 'cpu_has_saa'
     | undeclared (first use in this function)
     |   if (cpu_has_saa) {
     |       ^

    We define a disabled cpu_has_saa unless the machine specific feature
    overrides define a value.

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

linux-yocto: fix race between checkout and meta data generation

There are two tasks that must run before a linux-yocto kernel is built.

  - Kernel checkout and relocation to work-shared (kernel_checkout)
  - Meta data gathering and configuration prep (kernel_metadata)

The current task definitions for both are simply "before do_patch",
which is correct, but kernel_checkout must run before and not race with
kernel_metadata.

So we set the definition of kernel_checkout to be more specific and
enforce the proper ordering.

[YOCTO: #7731]

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

linux-yocto/3.19: Braswell support and bug fixes

Updating to include the following commits:

   a4d0c407cced dmaengine: dw: don't handle interrupt when dmaengine is not used
   e92b2ce791b2 dmaengine: dw: define DW_DMA_MAX_NR_MASTERS
   931304a6567e dmaengine: dw: provide DMA capabilities
   468bad4f7a6f dmaengine: dw: Split device_control
   b4afd7710db7 i2c: i801: Use managed pcim_* PCI device initialization and reservation
   d81a8a11ecba i2c: i801: Remove pci_enable_device() call from i801_resume()
   40e18604e70c i2c: i801: Use managed devm_* memory and irq allocation
   b54f65dbe57b i2c: i801: Remove i801_driver forward declaration
   e95740d4d079 i2c: i801: Don't break user-visible strings
   423e98721e04 ACPI: Introduce has_acpi_companion()
   291f620dc052 i2c: designware: Suppress error message if platform_get_irq() < 0
   cf5ff51a8e3f i2c: designware-pci: no need to provide clk_khz
   4f583ce420d3 i2c: designware-pci: remove Moorestown support
   e000c549c9d8 i2c: designware: Add Intel Baytrail PMIC I2C bus support
   7ffbd9ca19a7 i2c: designware: fixup return handling of wait_for_completion_timeout
   5758d5a1df32 i2c: designware: Do not calculate SCL timing parameters needlessly
   2f58fcae92db i2c: designware: Add i2c bus locking support
   84a73e51e900 i2c: designware: use {readl|writel}_relaxed instead of readl/writel
   f672bb8424e6 serial: 8250_dw: Fix get_mctrl behaviour
   91bd64585489 serial: 8250: add support for ACPI-probed serial port for X-Gene platform
   1190cba71f09 serial:8250:8250_pci: delete unneeded quirk entries
   6405a4b71451 serial:8250:8250_pci: fix redundant entry report for WCH_CH352_2S
   0a1a31bbbc19 serial: 8250_pci: remove one useless explicit type conversion
   4edc52a55f82 intel_idle: Add support for the Airmont Core in the Cherrytrail and Braswell SOCs
   934f85e8bfdb x86/irq, ACPI: Implement ACPI driver to support IOAPIC hotplug
   c6a3440252a8 ACPI / LPSS: check the result of ioremap()
   6aacc0c931b7 pinctrl: update direction_output function of cherryview driver
   cb4a43a2177d pinctrl: cherryview: Configure HiZ pins to be input when requested as GPIOs
   0df22c007ce1 pinctrl: intel: drop owner assignment from platform_drivers
   07b16f04700b pinctrl: cherryview: Save and restore pin configs over system sleep
   e8e5cfffa231 pinctrl: baytrail: Save pin context over system sleep
   04cb3cc0ff21 pinctrl: baytrail: Rework interrupt handling
   4cac25d2574d pinctrl: baytrail: Clear interrupt triggering from pins that are in GPIO mode
   eacab9ab234a pinctrl: baytrail: Relax GPIO request rules

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

xserver-xorg: Make xinerama selectable via PACKAGECONFIG

If xinerara can be selected via PACKAGECONFIG, xserver-xorg recipe can
be easier modified via bbappend.

xinerama is needed by the fglrx driver (from AMD).

Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gstreamer1.0: Make check selectable via PACKAGECONFIG

This way, this configuration can be easily changed via .bbappend file
without having to re-define the whole EXTRA_OECONF.

With --disable-check  libgstcheck is not build.

Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gstreamer1.0: convert GSTREAMER_1_DEBUG to PACKAGECONFIG

It is more elegant

Suggested-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bind: disable the HTTP statistics service by default

Rename the "libxml2" PACKAGECONFIG to "httpstats" so that is is meaningful, and
disable it by default as a web frontend to the server statistics shouldn't be
enabled by default.

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

bind: update libxml2 detection patch

Refresh the libxml2 detection patch to directly call pkg-config instead of
attempting to use xml2-config, which will always return an error in OE.

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

bind: upgrade to 9.10.2

1. Remove two unneeded CVE patches, 9.10.2 fixed the CVE-2015-1349 also
2. Remove bind-subdirs-run-serially.patch and cross-build-fix.patch,
similar fixes are merged into 9.10.2
3. update the dont-test-on-host.patch
4. update the Copyright file checksum, since the date in it has been changed.
5. replace the hardcode lib dir with $base_libdir in bind-add-crosscripts*.patch

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

oe-selftest: devtool: add a proper test to see if tap devices exist

Check up front in test_devtool_deploy_target whether the tap devices
exist and skip if not. If we don't do this we get a significantly less
comprehensible error via pexpect.

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

devtool: fix for rename of gcc-source

After OE-Core commit 67db7182faf6742b0d971d61d8c5ba34f69d2e12, PV is
appended to the end of the gcc-source PN, thus we need to handle that in
devtool and the corresponding test.

Part of the fix for [YOCTO #7729].

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

oe-selftest: devtool: fix broken URL in test_devtool_add_fetch

I already had the file fetched from some previous work and thus it
didn't attempt to download the invalid URL when I tested it earlier.

Part of the fix for [YOCTO #7729].

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

bluez: update to 5.30

Signed-off-by: Kevron Rees <kevron.m.rees@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

e2fsprogs: fix the CVE-2015-1572

Backport a patch to fix CVE-2015-1572
http://web.nvd.nist.gov/view/vuln/detail?vulnId=CVE-2015-1572

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

curl: upgrade to 7.42.1

Upgrade to include a security fixes for CVE-2015-3153

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

insane: Support aarch64 on musl

musl 1.1.8 supports aarch64

Change-Id: Ie8d39a29dcb7a41c8d59d2a00528b6add8a7655c
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libart-lgpl: Fix cross compiling

We use standard ISO C data types and get rid of configure guess work
which is iffy in cross compiled environment. Cleans up the re-build as well
since we were deleting art_config.h without considering if the file is
there in first place or not.

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

kernel.bbclass: Fix race condition

Race condition between do_compile_kernelmodules and do_shared_workdir
tasks occurs when do_compile_kernelmodules changes files in
include/generated/* while do_shared_workdir tries to copy them to
shared working directory.

Fixed race by moving do_shared_workdir after do_compile but before
do_compile_kernelmodules.

[YOCTO #7321]

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

classes/buildstats: misc cleanups

* Rename BNFILE and DEVFILE so that they are clearly specific to this
  class, since they end up in the global scope
* Use "with open" when opening files consistently
* Use getVar('PF', True) instead of expand('${PF}')
* Drop some unnecessary assignments

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

classes/buildstats: don't catch all errors from mkdirhier

mkdirhier() doesn't raise an exception if the directory exists, so if it
does raise one we should just be failing, otherwise we're just going
to hit errors later anyway.

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

classes/buildstats: whitespace cleanup

* Drop trailing whitespace
* Use spaces around equals in assignments
* Replace an errant tab with spaces

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

runqemu-internal: correctly set format for root FS as raw

qemu guesses via probing the format of root FS, but
gives a warning and restricts write operations on block 0.
So fix it by setting correctly the format as raw.

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

qemu: upgrade to 2.3.0

Qemu-Arm-versatilepb-Add-memory-size-checking.patch
brought up to date;
37ed3bf1ee07bb1a26adca0df8718f601f231c0b.patch removed,
integrated upstream;
glx enable config option changed to opengl enable,
update accordingly.

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

x11perf: Upgrade 1.5.4 -> 1.6.0

* remove PR

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

xhost: Upgrade 1.0.6 -> 1.0.7

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

xmodmap: Upgrade 1.0.8 -> 1.0.9

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

xev: Upgrade 1.2.1 -> 1.2.2

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

serf: update SRC_URI to use debian mirrors instead of googlecode

Update SRC_URI to use debian mirrors as googlecode will be closed
in near future. Also added a summary to the recipe.

Signed-off-by: Maxin B. John <maxin.john@enea.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libarchive: fix out of tree builds

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

kbproto: Upgrade 1.0.6 -> 1.0.7

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

randrproto: Upgrade 1.4.0 -> 1.4.1

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

xproto: Upgrade 7.0.26 -> 7.0.27

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

python3: Disable pip to fix issues on the Autobuilders

As Python-3.4 pip is required and installed by default, pips installation
is bootstrapped using python wheels, this makes it complicated to patch,
by default the installation uses an absolute path to /tmp, on most cases
this would not be a problem, but since on the Autobuilders several Python
installations (for different archs) may be happening at the same time,
this shared resource utilization results in errors at do_install(), this
patch disables pip installation by default for now, until we have a
python3-pip package which we can easily patch in these situations.

Signed-off-by: Alejandro Hernandez <alejandro.hernandez@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

python3-native: Disable pip on native builds to fix Autobuilder issues

As Python-3.4 pip is required and installed by default, pips installation
is bootstrapped using python wheels, this makes it complicated to patch
and causes errors on the Autobuilders, since we don't need pip for
python3-native, this disables it fixing the issues.

Signed-off-by: Alejandro Hernandez <alejandro.hernandez@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

python3-native: Fix pip install issue due to unclean build directory

When installing python3-native sometimes pips default build
directory (which is on the host and is user dependant) is left unclean,
due to this, when python3-core is being installed it tries to use
the same directory producing an error, this explicitly removes
what the previous installation might have left behind, fixing the issue.

Signed-off-by: Alejandro Hernandez <alejandro.hernandez@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

python3: Change python BASEVERSION to 3.4

Changes PYTHON_BASEVERSION = 3.4 to fix errors when installing
python3-distribute and probably other libraries

Signed-off-by: Alejandro Hernandez <alejandro.hernandez@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

generate-manifest-3.4.py: Add fractions to ${PN}-numbers

python-cryptography requires fractions module which is currently unpackaged.

Signed-off-by: Tim Orling <TicoTimo@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

python3: Upgrade from 3.3.3 to 3.4.2

Modifies:
python3-native_3.3.3.bb -> python3-native_3.4.2.bb: Updates checksums, version, patches
and SRC_URIs to use xz format, LICENSE did not change, dates were updated, adds fix for
ctypes/libffi needed for pip <- required since Python 3.4 (see PEP #453).
python3_3.3.3.bb -> python3_3.4.2.bb: Updates checksums, version, patches and SRC_URIs
to use xz format, adds fix for ctypes/libffi, fixes oe_multilib_header() call,
adds python3-core as RDEPENDS to python3-misc to avoid QA issue.
generate-manifest-3.4.py - > generate-manifest-3.4.py: fixes reprlib,
adds _sitebuiltins and _collections_abc <- formerly part of collections, separated since 3.4
python-3.3-manifest.inc -> python-3.4-manifest.inc

Rebases:
12-distutils-prefix-is-inside-staging-area.patch
python-3.3-multilib.patch
04-default-is-optimized.patch
avoid-ncursesw-include-path.patch

Deletes (fixed upstream):
fix-ast.h-dependency.patch
python3-fix-build-error-with-Readline-6.3.patch
06-ctypes-libffi-fix-configure.patch
remove_sqlite_rpath.patch

[YOCTO #6857]

Signed-off-by: Alejandro Hernandez <alejandro.hernandez@linux.intel.com>

connman: upgrade to 1.29

- Add support for Ethernet and VLAN usage;
- Fixes.

Added build-create-dirs-before-putting-files-in-them.patch,
already submitted upstream.

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

connman: code cleanup

- cleanup multiple inherit lines;
- RDEPENDS cleanup, use PACKAGECONFIG[] instead.

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

ghostscript: Add MicroBlaze support

* Add objarch.h for MicroBlaze big and little endian

Signed-off-by: Nathan Rossi <nathan.rossi@xilinx.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

siteinfo.bbclass: Add MicroBlaze big endian target info

* Add the target info for the 'microblazeeb' target which matches the
  default MicroBlaze target info

Signed-off-by: Nathan Rossi <nathan.rossi@xilinx.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

insane.bbclass: Add MicroBlaze big endian target info

* Add the target info for the 'microblazeeb' target which matches the
  default MicroBlaze target info

Signed-off-by: Nathan Rossi <nathan.rossi@xilinx.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

kernel-arch: Update mapping for MicroBlaze architectures

* Add pattern matching for the 'microblazeeb' architecture which maps to
  'microblaze' in the same way as 'microblazeel'

Signed-off-by: Nathan Rossi <nathan.rossi@xilinx.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

mirrors.bbclass: drop support for svk sources

The svk fetcher was removed from bitbake 1.22, so svk support should
logically have been removed from oe-core Daisy onwards.

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

shared-mime-info: Upgrade 1.3 -> 1.4

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

pseudo 1.6.5: less pointlessly chatty

There was a stupid logic error controlling the diagnostic for
a "possible" mismatch involving trailing slashes and whether or
not a node was believed to be a directory. Specifically, a diagnostic
got printed any time a lookup for a directory *didn't* have a
trailing slash, as well as in the (actually intended) case where
a non-directory lookup *did*.

No other changes, but that one is probably significant.

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

bitbake.conf/base.bbclass: Drop PRINC support

PRINC is no longer needed since we have the PR service instead
which people should be using. Drop the PRINC code.

This has been deprecated with a warning and then an error for quite some
time so we can remove it now.

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

directfb: Correct bashism

Fix builds of directfb on systems with dash as /bin/sh

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

directfb: Improve CFLAGS handling for x86

The problem with filter_out is that it expands the variable meaning overrides
to sub components can get lost. This has happening to the security flags
meaning directfb failed in lsb builds.

Use _remove instead of filter_out since it gives much more predictable results.

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

security_flags: Add python-numpy to pie incompatible list

With poky-lsb (security flags enabled), python-numpy doesn't build
with pie flags.

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

bitbake.conf: backfill bluez5 feature

If bluez4 is not explicitly passed as a
distro feature, bluez5 should be used as
the default, so backfill it.

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

bluez: remove bluez4 recipes

BlueZ 4.x and associated recipes
are now obsolete.
(bluez4,bluez-hcidump, gst-plugin-bluetooth).
Will be moved into recipes-connectivity
collection of meta-openembedded.
BlueZ 4.x is still usable in poky/YP.

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

bluez4: add proper systemd support

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

multilib: Tweak value of PN used for OVERRIDES

Currently, PN is used in overrides which is expanded to have a MLPREFIX. This
means and pn- overrides without the prefix would be ignored which is not what
is usually expected.

We noticed huge problems using poky-lsb with multilib since the per recipe
overrides were not applied. This adds in handling for PN with and without
the prefix. This should unbreak world-lsb builds on the autobuilder.

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

systemd: bring back the patch to customise root's $HOME

This patch appears to have been accidently dropped in the move to 219-stable,
probably because it didn't apply.  Update the patch and re-add it.

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

bind: remove 5 backport patches

They are backport patches, and verified that the patches are in the
source.

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

libart_lgpl: Allow art_config.h to be generated at compile time

* Remove patch that prevents the Makefile from generating the
  art_config.h file
* Remove pre-generated art_config.h for various architecture
* Add a do_compile prepend to remove the art_config.h that is provided
  in the source so that the compilation will regenerate the file

Signed-off-by: Nathan Rossi <nathan.rossi@xilinx.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>