When compiling meta-toolchain-qt5 on cortexa8, the compiler throws an
internal compiler error:

...
qttools-opensource-src-5.3.2/src/linguist/shared/po.cpp:
In function 'bool loadPO(Translator&, QIODevice&, ConversionData&)':
qttools-opensource-src-5.3.2/src/linguist/shared/po.cpp:717:1:
internal compiler error: in add_stores, at var-tracking.c:6000
...

Tracking this down led to https://bugs.linaro.org/show_bug.cgi?id=534
It seems the bug is well know and fixed upstream. So backporting from
trunk seems to be the right solution. This fixes the compiler problem
on cortexa8 and does not seem to be very invasive. The original commit
can be found at:

git-svn-id: svn+ssh://gcc.gnu.org/svn/gcc/trunk@212178 138bc75d-0d04-0410-961f-82ee72b054a4

(From OE-Core master rev: 6751ef78694783fb86e55c77afefae750ab1b610)

Signed-off-by: Stefan Müller-Klieser <s.mueller-klieser@phytec.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Joshua Lock <joshua.lock@collabora.co.uk>

In a baremetal build, TARGET_ENDIAN_OPTION isn't set leading to build
failures. Add in ifdefs to avoid this.

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

This seems to have been disabled since the dawn of time for no good reason.
Enable the .debug stripping and packaging allowing for a smaller SDK.

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

gcc-source don't do do_package_write_rpm, so we should set
PACKAGES = "" to avoid the building error if we want generate srpm,
otherwise, we get the error as below.

ERROR: Task do_deploy_archives in meta/recipes-devtools/gcc/gcc-source_4.9.bb \
       depends upon non-existent task do_package_write_rpm in \
       meta/recipes-devtools/gcc/gcc-source_4.9.bb

Signed-off-by: Bian Naimeng <biannm@cn.fujitsu.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>

Delete backported patch which are present in 4.9.2

backport patched from upstream gcc trunk to fix

[YOCTO #6824]

Change-Id: Ia0067940471d4c5d9d62089bf6f18f3a9c2bfedd
Signed-off-by: Khem Raj <raj.khem@gmail.com>

The ${STAGING_INCDIR_NATIVE}/gcc-build-internal-$mtarget may not exist
when use the external sdk toolchain, we need check before link for it.

Fixed:
run.do_configure.12538: 149: cd: can't cd to sysroots/x86_64-linux/usr/include/gcc-build-internal-x86_64-wrs-linux

(LOCAL REV: NOT UPSTREAM) -- Sent to oe-core on 20150204

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

Installing /usr/lib/gcc/* means we'd have two copies, one from gcc-cross
and one from here. These can confuse gcc cross where includes use #include_next
and builds track file dependencies (e.g. perl and its makedepends code).
For determinism we don't install this to the sysroot, ever and rely on the
copy from gcc-cross.

[YOCTO #7287]

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

It is not needed any more since sed-native had been dropped.

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

[YOCTO #7230]

In certain system configurations TRANSLATED_TARGET_ARCH will not
expand in the right order for gcc-cross-candian-mips64n32 to be
generated properly.

This will cause SDKs to fail to generate properly.

Changing the global definition of TRANSLATED_TARGET_ARCH always
expands the ABIEXTENSION, which causes the OVERRIDES to pick it up
as well.  This effectively defines a new class of overrides for the 'n32'.

The side effect is that we need to duplicate some mips64 overrides, and
redefine others that were previously 'n32' or 'mips64' exclusive to have
the correct semantics.

Signed-off-by: Mark Hatle <mark.hatle@windriver.com>

The sanitizer runtime library is dual-licensed under the NCSA
and MIT licenses.

Also make nativesdk-gcc-sanitizers use SDKGCCVERSION by default
instead of GCCVERSION

Signed-off-by: Dan McGregor <dan.mcgregor@usask.ca>
Signed-off-by: Ross Burton <ross.burton@intel.com>