| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
As of GCC 8, plugin support is available with mingw. If the plugins are
enabled in the regular binutils/gcc configurations, carry over that
configuration with the mingw versions as well.
When plugins are enabled, they produce a file called [plugin].dll.a. This
causes the debugsrc processor to find the file and attempt to pull out dwarf
information. However, we need to prevent this as the files are not ELF, but
in Microsoft PE format, and trigger an error if processed.
Signed-off-by: Mark Hatle <mark.hatle@kernel.crashing.org>
Signed-off-by: Joshua Watt <JPEWhacker@gmail.com>
|
|
|
|
|
|
| |
OE-Core dropped gcc-initial, we can do the same for mingw.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
|
|
|
|
|
|
|
| |
Use the winpthreads library available in the mingw64 and
change the thread model from win32 to posix to get access
to c++11 features.
Signed-off-by: Samuli Piippo <samuli.piippo@qt.io>
Signed-off-by: Joshua Watt <JPEWhacker@gmail.com>
|
|
|
|
| |
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
|
|
|
|
| |
oe_filter_out has been removed from oe-core so use the replacement function.
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
|
|
|
|
|
|
|
| |
Relocate and package the dll's provided by libgcc in bindir. It is
convention that on mingw/windows dll files are installed into the
bindir, this is done to avoid issues with search paths and allows for
execution of binaries without the need to point at a libdir.
Signed-off-by: Nathan Rossi <nathan@nathanrossi.com>
|
|
|
|
|
|
|
|
| |
Intel Memory Protection Extension library for x86 builds are now enabled
by default. However, the library does not build for mingw32, so remove it
from mingw builds.
Signed-off-by: Juro Bystricky <juro.bystricky@intel.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Default configuration for gcc-crosssddk-initial is now to enable initfini-array.
However, this works only for Linux so we disable it for mingw32.
Otherwise, (with SDKMACHINE=x86_64-mingw32) we will eventually encounter errors such as:
Assembler messages:
Error: invalid instruction suffix for `push'
Error: invalid instruction suffix for `pop'
Signed-off-by: Juro Bystricky <juro.bystricky@intel.com>
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
|
|
|
|
|
|
|
|
|
|
|
| |
Default configuration for gcc-crosssddk is now to enable initfini-array.
However, this works only for Linux so we disable it for mingw32.
Otherwise we will eventually encounter build error such as:
multiple definition of `__do_global_dtors'
Signed-off-by: Juro Bystricky <juro.bystricky@intel.com>
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Issue: LIN8-1615
Workaround for the failure:
e:/windri~1/lx8.sdk/qemux8~1.800/sysroots/i686-w~1/usr/bin/i586-w~3/../../libexec/i586-wrs-linux/gcc/i586-wrs-linux/5.2.0/real-ld.exe: unrecognized option '-plugin'
This disables the LTO support in the compiler, but does allow things to
compile and link properly.
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
|
|
|
|
|
|
|
| |
libitm mingw build presently not supported, so don't build it.
Signed-off-by: Juro Bystricky <juro.bystricky@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
|
|
|
|
|
| |
Also we skip the staticdev sanity check, as the '.a' file should be packaged
as listed.
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
|
|
|
|
|
|
|
|
|
| |
Rename the bbappends to use % so they can be used with both 4.x
and 5.x. None of these changes appear to be specific to a given
version of gcc.
Signed-off-by: Peter Seebach <peter.seebach@windriver.com>
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
|
|
|
|
|
|
|
| |
The mingw32 runtime has issues building libgomp, so drop it for now.
Signed-off-by: Peter Seebach <peter.seebach@windriver.com>
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
|
|
|
|
|
|
|
| |
Uprev gcc to match master, and resync with latest mingw runtime.
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
|
|
|
|
|
|
|
|
|
| |
The symlinking is really an artefact of the way the headers are
provided. Move all the code into one place.
Also have the initial compiler depend on the headers else the build
can fail. High parallelism prevented the issue from being seen
in most cases.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
|
|
| |
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
|