MAX() and MIN() were evaluating arguments twice. If arguments are
functions they were called twice which resulted in bigger code
and potential misbehavior.
Added alternative macros (Z_MAX, Z_MIN) which can be used instead.
Macros have usage limitations thus they are not replacements. They
are also relying on GCC extension thus placed in gcc.h
Signed-off-by: Krzysztof Chruscinski <krzysztof.chruscinski@nordicsemi.no>
__BYTE_ORDER__ preprocessor definition is not defined by older versions
of GCC. The definitions for __ORDER_BIG_ENDIAN__,
__ORDER_LITTLE_ENDIAN__ and __BYTE_ORDER__ by automatic detection using
arch-specific endianness definitions have been added.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
With the upcoming riscv64 support, it is best to use "riscv" as the
subdirectory name and common symbols as riscv32 and riscv64 support
code is almost identical. Then later decide whether 32-bit or 64-bit
compilation is wanted.
Redirects for the web documentation are also included.
Then zephyrbot complained about this:
"
New files added that are not covered in CODEOWNERS:
dts/riscv/microsemi-miv.dtsi
dts/riscv/riscv32-fe310.dtsi
Please add one or more entries in the CODEOWNERS file to cover
those files
"
So I assigned them to those who created them. Feel free to readjust
as necessary.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
Ensure that xcc is at parity with gcc and clang by inferring missing
definitions based on those that it already provides.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
Compilers (at least gcc and clang) already provide definitions to
create standard types and their range. For example, __INT16_TYPE__ is
normally defined as a short to be used with the int16_t typedef, and
__INT16_MAX__ is defined as 32767. So it makes sense to rely on them
rather than hardcoding our own, especially for the fast types where
the compiler itself knows what basic type is best.
Using compiler provided definitions makes even more sense when dealing
with 64-bit targets where some types such as intptr_t and size_t must
have a different size and range. Those definitions are then adjusted
by the compiler directly.
However there are two cases for which we should override those
definitions:
* The __INT32_TYPE__ definition on 32-bit targets vary between an int
and a long int depending on the architecture and configuration.
Notably, all compilers shipped with the Zephyr SDK, except for the
i586-zephyr-elfiamcu variant, define __INT32_TYPE__ to a long int.
Whereas, all Linux configurations for gcc, both 32-bit and 64-bit,
always define __INT32_TYPE__ as an int. Having variability here is
not welcome as pointers to a long int and to an int are not deemed
compatible by the compiler, and printing an int32_t defined with a
long using %d makes the compiler to complain, even if they're the
same size on 32-bit targets. Given that an int is always 32 bits
on all targets we might care about, and given that Zephyr hardcoded
int32_t to an int before, then we just redefine __INT32_TYPE__ and
derrivatives to an int to keep the peace in the code.
* The confusion also exists with __INTPTR_TYPE__. Looking again at the
Zephyr SDK, it is defined as an int, even even when __INT32_TYPE__ is
initially a long int. One notable exception is i586-zephyr-elf where
__INTPTR_TYPE__ is a long int even when using -m32. On 64-bit targets
this is always a long int. So let's redefine __INTPTR_TYPE__ to always
be a long int on Zephyr which simplifies the code, works for both
32-bit and 64-bit targets, and mimics what the Linux kernel does.
Only a few print format strings needed adjustment.
In those two cases, there is a safeguard to ensure the type we're
enforcing has the right size and fail the build otherwise.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
Found a few annoying typos and figured I better run script and
fix anything it can find, here are the results...
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Given that the section name and boundary simbols can be inferred from
the struct object name, it makes sense to create an iterator that
abstracts away the access details and reduce the possibility for
mistakes.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
This convenience macro wraps Z_DECL_ALIGN() and __in_section() to
simplify static definitions of structure instances gathered in dedicated
sections. Most of the time those go together, and the section name is
already closely related to the struct type, so abstracting things behind
a simpler interface reduces probability of mistakes and makes the code
clearer. A few input section names have been adjusted accordingly.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
The alignment fix on struct device definitions should be done to all
such linker list tricks. Let's abstract the declaration plus alignment
with a macro and apply it to all concerned cases.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
Use the new math_extras functions instead of calling builtins directly.
Change a few local variables to size_t after checking that all uses of
the variable actually expects a size_t.
Signed-off-by: Jakob Olesen <jolesen@fb.com>
Compilers based on Clang provide a __has_builtin(x) macro which can be
used to detect in the preprocessor if a given builtin function is
supported by the compiler.
For other compilers (notably GCC), we provide an alternative definition
of HAS_BUILTIN(x) that depends on the toolchain-specific header file to
declare which builtin functions are supported based on the current
compiler version.
Signed-off-by: Jakob Olesen <jolesen@fb.com>
Previous commit c31e659165 changed
the ALWAYS_INLINE macros to avoid functions being inlined for
the purpose of code coverage. This has a side effect of causing
the text sections of kobject_hash.c and priv_stacks_hash.c
to ballon more than 10 times in size. This is caused by
attaching the unused attribute, which results in all those
functions being in the text sections though they are never
used. So just keep the "inline" there.
This also removes -fno-inline from NO_COVERAGE_FLAGS so these
two files are not compiled with flags related to code coverage.
Fixes#15009
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
This adds a compiler option -fno-inline for code coverage on
architectures which supports doing code coverage. This also
modifies the ALWAYS_INLINE macro to not do any inlining. This
needs to be done so code coverage can count the number of
executions to the correct lines.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Update reserved function names starting with one underscore, replacing
them as follows:
'_k_' with 'z_'
'_K_' with 'Z_'
'_handler_' with 'z_handl_'
'_Cstart' with 'z_cstart'
'_Swap' with 'z_swap'
This renaming is done on both global and those static function names
in kernel/include and include/. Other static function names in kernel/
are renamed by removing the leading underscore. Other function names
not starting with any prefix listed above are renamed starting with
a 'z_' or 'Z_' prefix.
Function names starting with two or three leading underscores are not
automatcally renamed since these names will collide with the variants
with two or three leading underscores.
Various generator scripts have also been updated as well as perf,
linker and usb files. These are
drivers/serial/uart_handlers.c
include/linker/kobject-text.ld
kernel/include/syscall_handler.h
scripts/gen_kobject_list.py
scripts/gen_syscall_header.py
Signed-off-by: Patrik Flykt <patrik.flykt@intel.com>
Instead of having to enable ramfunc support manually, just make it
transparently available to users, keeping the MPU region disabled if not
used to not waste a MPU region. This however wastes 24 bytes of code
area when the MPU is disabled and 48 bytes when it is enabled, and
probably a dozen of CPU cycles during boot. I believe it is something
acceptable.
Note that when XIP is used, code is already in RAM, so the __ramfunc
keyword does nothing, but does not generate an error.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Using __ramfunc to places a function in RAM instead of Flash.
Code that for example reprograms flash at runtime can't execute
from flash, in that case must placing code into RAM.
This commit create a new section named '.ramfunc' in link scripts,
all functions has __ramfunc keyword saved in thats sections and
will load from flash to sram after the system booted.
Fixes: #10253
Signed-off-by: qianfan Zhao <qianfanguijin@163.com>
This was never a long-term solution, more of a gross hack
to get test cases working until we could figure out a good
end-to-end solution for memory domains that generated
appropriate linker sections. Now that we have this with
the app shared memory feature, and have converted all tests
to remove it, delete this feature.
To date all userspace APIs have been tagged as 'experimental'
which sidesteps deprecation policies.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
This patch adds a x86_64 architecture and qemu_x86_64 board to Zephyr.
Only the basic architecture support needed to run 64 bit code is
added; no drivers are added, though a low-level console exists and is
wired to printk().
The support is built on top of a "X86 underkernel" layer, which can be
built in isolation as a unit test on a Linux host.
Limitations:
+ Right now the SDK lacks an x86_64 toolchain. The build will fall
back to a host toolchain if it finds no cross compiler defined,
which is tested to work on gcc 8.2.1 right now.
+ No x87/SSE/AVX usage is allowed. This is a stronger limitation than
other architectures where the instructions work from one thread even
if the context switch code doesn't support it. We are passing
-no-sse to prevent gcc from automatically generating SSE
instructions for non-floating-point purposes, which has the side
effect of changing the ABI. Future work to handle the FPU registers
will need to be combined with an "application" ABI distinct from the
kernel one (or just to require USERSPACE).
+ Paging is enabled (it has to be in long mode), but is a 1:1 mapping
of all memory. No MMU/USERSPACE support yet.
+ We are building with -mno-red-zone for stack size reasons, but this
is a valuable optimization. Enabling it requires automatic stack
switching, which requires a TSS, which means it has to happen after
MMU support.
+ The OS runs in 64 bit mode, but for compatibility reasons is
compiled to the 32 bit "X32" ABI. So while the full 64 bit
registers and instruction set are available, C pointers are 32 bits
long and Zephyr is constrained to run in the bottom 4G of memory.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
The builtin_*_overflow functions when available return a
boolean. Change internal definitions to return same type.
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
Functions are declared as noreturn but they do in fact return (when
control reaches the end of the body, since it doesn't enter an infinite
loop, it doesn't call other "noreturn" functions, etc.)
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
XCC doesn't provide these builtins so we have to define them
with minimal functionality for testing.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
The macros likely() and unlikely() used by the compiler for
optimization are always used inside an if condition.
According to MISRA we need to have bool type and not long.
Signed-off-by: Adithya Baglody <adithya.nagaraj.baglody@intel.com>
Any word started with underscore followed by and uppercase letter or a
second underscore is a reserved word according with C99.
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
compiler_barrier() is itself defined down in this file. Without
adding it, newer versions of GCC (7+) for ARM Cortex-M may mistakenly
coalesce multiple strb/strh/str (store byte/half-word/word)
instructions, which support unaligned access on some
sub-architectures (Cortex-M3 and higher, but not on Cortex-M0),
into strd (store double), which doesn't support unaligned access.
Fixes: #6307
Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org>
The logic for using _Static_assert() was a little broken. We were
using it when on GCC 4.6+ AND when __STDC_VERSION__ said we were on
C99 or better. But it's not a C99 feature, it's a C11 feature. And
if GCC provides it as an extension, that's unrelated to a particular
language version. This should have been "GCC 4.6+ OR C11+".
This actually broke on the ESP-32 IDF toolchain, where (when using
-std=c99) the compiler was actually defining a C99 macro instead of
the C11 one, and choosing to use the wrong (and independently broken)
handling incorrectly. Fixes#8093.
Signed-off-by: Andy Ross <andrew.j.ross@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>
BUILD_ASSERT() was always defining a type with the name
__build_assert_failure, causing issues if more than one assertion were
used in the same scope.
Also, use an enum instead of a typedef to avoid (possibly spurious)
warnings such as these:
variably modified ‘__build_assert_failure1’ at file scope [-Werror]
Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
This macro has been deprecated in favor of K_DECLARE_STACK; should have
been removed by 1.11.
Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
This patch lets a C++ application use more of Zephyr by adding guards
and changeing some constructs to the C++11 equivalent.
Changes include:
- Adding guards
- Switching to static_assert
- Switching to a template for ARRAY_SIZE as g++ doesn't have the
builtin.
- Re-ordering designated initialisers to match the struct field order
as G++ only supports simple designated initialisers.
Signed-off-by: Michael Hope <mlhx@google.com>
Added guards to the toolchain/gcc.h and toolchain/common.h
Those files should never be included directly, but a guard is useful
regardless.
Fixes#5130
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
A new arch (posix) which relies on pthreads to emulate the context
switching
A new soc for it (inf_clock) which emulates a CPU running at an
infinely high clock (so when the CPU is awaken it runs till completion
in 0 time)
A new board, which provides a trivial system tick timer and
irq generation.
Origin: Original
Fixes#1891
Signed-off-by: Alberto Escolar Piedras <alpi@oticon.com>
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
This is a temporary hack until #5006 is resolved (possibly using
https://github.com/zephyrproject-rtos/zephyr/issues/5006
Unit testing (BOARD == unit_testing) doesn't need the system call
definitions. Because we foward declare with __syscall them as "static
inline" (from common.h), the compilers will complain that the
definition is missing.
Change to only define __syscall as "static inline" if we are not
builing a unit test to avoid said warnings.
Signed-off-by: Inaky Perez-Gonzalez <inaky.perez-gonzalez@intel.com>
This used to exist because in earlier versions of the system call
interfaces, an "extern" declaration of the system call implementation
function would precede the real inline version of the implementation.
The compiler would not like this and would throw "static declaration
of ‘foo’ follows non-static declaration". So alternate macros were
needed which declare the implementation function as 'static inline'
instead of extern.
However, currently the inline version of these system call
implementations appear first, the K_SYSCALL_DECLARE() macros appear in
the header generated by gen_syscalls.py, which is always included at the
end of the header file. The compiler does not complain if a
static inline function is succeeded by an extern prototype of the
same function. This lets us simplify the generated system call
macros and just use __syscall everywhere.
The disassembly of this was checked on x86 to ensure that for
kernel-only or CONFIG_USERSPACE=n scenarios, everything is still being
inlined as expected.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
To define a system call, it's now sufficient to simply tag the inline
prototype with "__syscall" or "__syscall_inline" and include a special
generated header at the end of the header file.
The system call dispatch table and enumeration of system call IDs is now
automatically generated.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
XCC assembler freaks out if a section name has __FILE__ in it,
forward slashes and quotation marks confuse it and result in
build errors.
This is not a perfect fix, its possible for two sections to collide,
but at worst this will result is some unnecessary space in noinit,
fooling gc-sections.
XCC also doesn't support __COUNTER__, use __LINE__ as a substitute.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
_FILE_PATH_HASH appears to be a legacy Diab-ism that doesn't
expand to anything in GCC.
As a result, when linking the combined binary, it's quite
possible that objects in separate C files would be merged
instead of truly being in their own section. This can confound
--gc-sections and result in unused objects still being in
the final binary if one of the other objects with the same
generated section name was actually used.
We instead just use __FILE__. This results in sometimes absurdly-
long section names in the intermediate .o files, but there is no
actual limit to how long section names in ELF binaries can be;
they are not stored directly in headers but instead referenced
as an offset in the .shstrtab section, which has all the section
names stored in it.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Add a macro which signals to the compiler that use of the macro is
deprecated.
Example:
#define FOO __DEPRECATED_MACRO bar
Defines FOO to 'bar' but emits a warning if used in code.
Cannot filter out with -Wno-deprecated, so be careful with -Werror.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
These are macros that are expected to be defined at all times by
the compiler. We need them at the very beginning of kernel.h for
the k_thread definition, before it's possible to include arch.h.
Make a special toolchain header for XCC compiler and place these
defines in there. Otherwise inherit all the other GCC defines.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
__stack is defined as a C language attribute for on-stack arrays. Don't
define it outside C source code.
This definition conflicts with __stack symbol defined in xtensa linker.ld
files.
Change-Id: I59fe34603bc2bb5732ed45c7974de5f8b25d77ed
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Provide a BUILD_ASSERT and BUILD_ASSERT_MSG that are based on
_Static_assert when that's available, as its output is easier to read.
Change-Id: Ifa96d5073b1341cab2a90e4dcd04752ee80c69bb
Signed-off-by: Marti Bolivar <marti.bolivar@linaro.org>
Like BUILD_ASSERT(), but with a message to emit on failure.
The base implementation swallows the message; compiler headers can
override it when they can do better.
Change-Id: Ib724e48554da77a51afa01468b1d5b7806f9de6b
Signed-off-by: Marti Bolivar <marti.bolivar@linaro.org>
Krzysztof Chruscinski