clang doesn't understand fortify at all, provide no op macro,
in order to handle the request to fortify in a generic way.
Signed-off-by: Oleg Zhurakivskyy <oleg.zhurakivskyy@intel.com>
In case TOOLCHAIN_HOME isn't explicitly reuqested,
(or indirectly forced with CLANG_ROOT_DIR), detect
any host installed clang in the path.
Signed-off-by: Oleg Zhurakivskyy <oleg.zhurakivskyy@intel.com>
toolchain_ld_base() represents flags that are fundamental to linking or
otherwise does not belong in any further specified linker flag category.
No functional change expected.
This is motivated by the wish to abstract Zephyr's usage of toolchains,
permitting non-intrusive porting to other (commercial) toolchains.
Signed-off-by: Mark Ruvald Pedersen <mped@oticon.com>
Update the files which contain no license information with the
'Apache-2.0' SPDX license identifier. Many source files in the tree are
missing licensing information, which makes it harder for compliance
tools to determine the correct license.
By default all files without license information are under the default
license of Zephyr, which is Apache version 2.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Prepend -m32 to existing flags as opposed to silently override and
discard them:
- set(CMAKE_<tool>_FLAGS -m32 )
+ string(PREPEND CMAKE_<tool>_FLAGS " -m32")
This stops discarding additional flags passed with either cmake
-DCMAKE_<tool>_FLAGS="-fu -bar" or cmake -C params.cmake.
Note this bug was affecting only the combination of x86 and
CMAKE_<tool>_FLAGS. x86_64 wasn't affected and other, non-empty
CMAKE_<tool>_FLAGS_[DEBUG/RELEASE/etc.] weren't affected either.
Also convert flags string to list when invoking --print-libgcc-file-name
to support more than one CMAKE_C_FLAGS.
Signed-off-by: Marc Herbert <marc.herbert@intel.com>
When we build with newlib we don't set -nostdinc. In that case make
sure that we leave it to the toolchain to set the system include paths.
The one exception to leaving to the toolchain to set the system include
paths is the path to the newlib headers. Since we build
with -ffreestanding we need to make sure the newlib header path is the
before the toolchain headers. Otherwise the toolchain's 'freestanding'
headers get picked up and that causes issues (for example getting PRI*64
defined properly from inttypes.h due to __STDC_HOSTED__ being '0').
For newlib we accomplish this by having the only system header specified
by zephyr_system_include_directories() being just the newlib headers.
Note: for minlibc we leave things alone as things just happen to work as
the -I include of the libc headers takes precedence over -isystem so we
get the libc headers over the toolchain ones. For the newlib case it
appears that setting both -I and -isystem for the same dir causes the
-I to be ignored.
Fixes#14310
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
- newlib needs c standard includes (so no -nostdinc)
- xcc needs toolchain headers (so no -nostdinc)
- with host gcc:
- x86_64 should not build with standard includes (-nostdinc needed)
- native_posix should build with standard include (no -nostdinc)
..
..
To simplify, abstract this and move it to compilers/toolchains and still
depend on what the application wants.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
XCC is based on xcc, but is nothing like gcc and his many differences.
Instead of ifdeffing the gcc code with Xcc specifics, maintain it
standalone.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Introduce toolchain_cc_asm macro to capture toolchain specific flags
related to assembly.
-D_ASMLANGUAGE is kept common for all, assuming -D as define flag is
supported by all compilers (which is almost the case).
No functional change expected.
Clang's flags are compatible with gcc, and are thus inherited.
This is motivated by the wish to abstract Zephyr's usage of toolchains,
permitting easier porting to other (commercial) toolchains.
Signed-off-by: Mark Ruvald Pedersen <mped@oticon.com>
We where defining the variable NOSTDINC_F after we tried to use it.
Move the definition before the first reference fixes things. When
-nostdinc now enabled we need to explicitly add the compiler include
path for x86_64 based builds (x86_64 and ARCH_POSIX).
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
Introduce toolchain_cc_cpp_*-family of macros.
Move the following into the toolchain_cc_cpp_* macros:
* Common base set of flags
* C++ standard version
* RTTI
* Exceptions
These macros must be implemented by every compiler port.
These macros set the respective flags, but leaves logic and control to
the root CMakeLists.txt file.
No functional change expected.
Clang's C++ flags are compatible with gcc, and are thus
inherited.
This is motivated by the wish to abstract Zephyr's usage of toolchains,
permitting easier porting to other (commercial) toolchains.
Signed-off-by: Mark Ruvald Pedersen <mped@oticon.com>
Introduce toolchain_cc_optimize_for_* family of macros.
Each macro represents a general optimization class.
Each macro is then responsible for setting an output variable to that
class-of-optimization's flag.
The names of these output variables are decided from the root
CMakeLists.txt.
No functional change expected.
Clang's optimization flags are compatible with gcc, and are thus
inherited.
This is motivated by the wish to abstract Zephyr's usage of toolchains,
permitting easier porting to other (commercial) toolchains.
Signed-off-by: Mark Ruvald Pedersen <mped@oticon.com>
We were select host compiler from the generic gcc compiler section which
is used for cross toolchains. This arch should use the host-gcc compiler
definition instead.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
PR #9522 series ending with commit c2c9265b7d ("tests: cmsis: Disable
two cmsis portability tests on x86_64") added -mx32 support for the
x86_64 ARCH and qemu_x86_64. While this was implemented in
"compiler/gcc/target.cmake" as fall back from cross-compilation to the
host compiler, it worked with a direct ZEPHYR_TOOLCHAIN_VARIANT=host
too.
Later, -lgcc was added to "compiler/host-gcc/target.cmake" by PR #12674
to fix the -m32 x86 build. This broke the x86_64 build when using
ZEPHYR_TOOLCHAIN_VARIANT=host because even "multilib" packages usually
don't feature the -mx32 version of libgcc.
Fix this by excluding -lgcc in compiler/host-gcc/target.cmake just like
compiler/gcc/target.cmake always did for x86_64.
Signed-off-by: Marc Herbert <marc.herbert@intel.com>
Introduce the first of the toolchain_cc-family of macros:
toolchain_cc_security_fortify and toolchain_cc_security_canaries.
No functional change expected.
Fortify source is not supported by Clang, but this commit retains
current behavior.
This is motivated by the wish to abstract Zephyr's usage of toolchains,
permitting easier porting to other (commercial) toolchains.
Signed-off-by: Mark Ruvald Pedersen <mped@oticon.com>
Add missing -lgcc when compiling with ZEPHYR_TOOLCHAIN_VARIANT=host
merely copying some existing code from
'compiler/{clang,gcc}/target.cmake'.
This fixes compilation for the following boards with an x86
microprocessor:
galileo, minnowboard, qemu_x86, qemu_x86_nommu, up_squared,
up_squared_sbl
Compilation of the following boards with an X86_IAMCU microcontroller
still fail with a "cannot find -lgcc" error:
arduino_101, qemu_x86_iamcu, quark_d2000_crb, quark_se_c1000_devboard,
tinytile
This is _not_ a regression because these boards _already_ failed with
"undefined reference to __udivdi3" and other libgcc symbols.
Signed-off-by: Marc Herbert <marc.herbert@intel.com>
This fixes the following error on all x86 BOARDs (qemu_x86,
galileo,... see "make usage" for the complete list) when compiling
assembly files with ZEPHYR_TOOLCHAIN_VARIANT=host:
cc1: error: CPU you selected does not support x86-64 instruction set
This also fixes the following error when compiling minnowboard,
up_squared and up_squared_sbl boards with ZEPHYR_TOOLCHAIN_VARIANT=host:
cc1: error: -mpreferred-stack-boundary=2 is not between 3 and 12
This fix alone is not enough to compile any of these boards; however it
moves compilation much further to the next, unrelated failure(s),
namely: 'undefined __udivdi3' and other libgcc symbols for all x86
boards; + some other, additional issues for some boards. See next
commit.
Signed-off-by: Marc Herbert <marc.herbert@intel.com>
A GCC-based toolchain may require additional, toolchain-specific
values in CMAKE_REQUIRED_FLAGS to perform compiler checks properly,
but gcc.cmake clobbers any values the user provides.
Preserve them instead, allowing users to give their own compiler
checking flags at generation time.
The details for the particular issue that inspired this are described
in https://github.com/pulp-platform/pulpino/issues/240.
Signed-off-by: Marti Bolivar <marti@foundries.io>
Host toolchains don't tend to provide an x32 libgcc. But we don't
actually need one for existing code. This is fragile, but better to
work for all but obscure cases than break outright.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Cosmetics, no functional change expected.
Fixed leading space alignment.
Replaced tabs with spaces.
Emulation error message output is now aligned.
To locate tabs in cmake, the following bash is useful:
grep -PRil "\t" * | grep -i cmake | grep -v ^sanity
Signed-off-by: Mark Ruvald Pedersen <mped@oticon.com>
Split up the toolchain configuration into two phases, generic and
target. The 'generic' phase configures the toolchain just enough to be
able to preprocess DT files. The 'target' phase completes the
configuration with target-specific configuration.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
When building with clang symbols from the runtime library, like
__divdi3 were not being resolved. To fix this we add the "library path
for the currently used compiler runtime library" to the link path.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
Align 'clang' with 'gcc' by having it also use 'find_program' instead
of 'set' to assign toolchain paths.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
Align 'clang' with gcc by having it also set CMAKE_C_COMPILER in the
'compiler' build script instead of the 'toolchain' build script.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
The 'llvm' and 'clang' build scripts have been named strangely. It is
more natural for 'clang' to be the compiler, and 'llvm' to be the
toolchain.
This commit rectifies this by renaming the files.
This also fixes#11187
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
Fixed whitespace such that 'XTOOLS_TOOLCHAIN_PATH' is vertically
aligned.
Also, combined two cmake invocations of 'list' into one.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
Some toolchains are built with multilib enabled in order to provide
multiple versions of the same library, optimized for different ABI
or architecture. They require the the ABI/architecture/CPU selection
options to be passed at linked time. This is important for example
when linking with newlib.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
MSYS support was added as a stop-gap while native windows support was
unsupported. Now that Native windows support is stable we can drop
support for MSYS.
Dropping support for MSYS fixes#11260 and allows us to spend more
resources on native windows support.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
Not all compiles/linkers support the GCC flags to not include standard
defines, include files and libraries. So make these parameters such
that the toolchain can define them when supported.
Also, according to documentation, -nostdlib does the same for both
-nostartfiles and -nodefaultlibs. So remove the redundant ones.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
This re-applies a patch that was merged in pr #6989, but then
accidentally reverted in a bad merge conflict resolution in pr #6842.
The actual behaviour of CONIG_FP_SOFTABI has not been matching the
documenation. The Kconfig documentation states that floating point
instructions should be generated, but this Kconfig option has been
turning off floating point instructions instead.
This commit causes floating point instructions to be generated when
CONFIG_FP_SOFTABI is enabled, as was originally intended and
documented.
This commit can cause regressions if users have been relying on the
actual behaviour instead of the documented behaviour.
Kconfig documentation:
config FP_SOFTABI
help
This option selects the Floating point ABI in which hardware
floating point instructions are generated but soft-float calling
conventions.
GCC documentation:
Specifying ‘soft’ causes GCC to generate output containing library
calls for floating-point operations. ‘softfp’ allows the generation of
code using hardware floating-point instructions, but still uses the
soft-float calling conventions. ‘hard’ allows generation of
floating-point instructions and uses FPU-specific calling conventions.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
'find_program' should be used instead of 'set' for determining the
paths of host programs. It resolves#9582 and various other issues.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
When the SDK/toolchain has been incorrectly installed such that the
user does not have permissions to execute it an incomprehensible
errror message occurs.
This patch tests that executing 'gcc --version' does not give an error
code and errors out with an appropriate error message if it does.
This fixes#8977
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
Add an LLVM backend and a clang toolchain variant to support building
with llvm coming with popular Linux distributions.
This has been tested with X86 boards:
- quark_d2000_crb
- quark_se_c1000_devboard/Arduino 101
Use:
export ZEPHYR_TOOLCHAIN_VARIANT=clang
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
This patchset provides Xtensa's xcc compiler support for Xtensa
projects in Cmake. This requires the below environment variables
to be defined aptly. The appropriate xcc license information also
need to be supplied.
ZEPHYR_GCC_VARIANT=xcc
TOOLCHAIN_VER=RF-2015.3-linux
XTENSA_CORE=cavs21_LX6HiFi3_RF3_WB16
XTENSA_SYSTEM=/opt/xtensa/XtDevTools/install/tools/
RF-2015.3-linux/XtensaTools/config/
XTENSA_BUILD_PATHS=/opt/xtensa/XtDevTools/install/builds/
Change-Id: Ib3c10e8095439b0e32276ff37c00eca8420773ec
Signed-off-by: Rajavardhan Gundi <rajavardhan.gundi@intel.com>
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Make posix align with non-posix platforms by having CMAKE_C_COMPILER
and friends use absolute paths.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
Improve the error feedback when the toolchain is not found.
Currently, for some setups, presumably primarily SDK-less setups,
users can get an error message like this for certain types of
misconfigurations:
CMake Error at zephyr/cmake/extensions.cmake:984 (message):
No such file or directory: CMAKE_READELF: 'CMAKE_READELF-NOTFOUND'
Call Stack (most recent call first):
zephyr/cmake/compiler/gcc.cmake:14 (assert_exists)
zephyr/cmake/toolchain.cmake:38 (include)
zephyr/cmake/app/boilerplate.cmake:243 (include)
CMakeLists.txt:1 (include)
This commit will improve the error feedback to give a explanation for
what kind of failure has occured and will also dump relevant variables
to ease debugging.
This fixes#7075
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
The actual behaviour of CONIG_FP_SOFTABI has not been matching the
documenation. The Kconfig documentation states that floating point
instructions should be generated, but this Kconfig option has been
turning off floating point instructions instead.
This commit causes floating point instructions to be generated when
CONFIG_FP_SOFTABI is enabled, as was originally intended and
documented.
This commit can cause regressions if users have been relying on the
actual behaviour instead of the documented behaviour.
Kconfig documentation:
config FP_SOFTABI
help
This option selects the Floating point ABI in which hardware
floating point instructions are generated but soft-float calling
conventions.
GCC documentation:
Specifying ‘soft’ causes GCC to generate output containing library
calls for floating-point operations. ‘softfp’ allows the generation of
code using hardware floating-point instructions, but still uses the
soft-float calling conventions. ‘hard’ allows generation of
floating-point instructions and uses FPU-specific calling conventions.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
Move toolchain definitions into toolchain/ and move compilers to
compilers/
Usage of toolchain- for everything was confusing, there are the actual
compiler related definitions and there is the toolchain/SDK related
configs, so keeping them separate helps understand the different and
prevents confusion.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
