Move to CMake 3.20.0.
At the Toolchain WG it was decided to move to CMake 3.20.0.
The main reason for increasing CMake version is better toolchain
support.
Better toolchain support is added in the following CMake versions:
- armclang, CMake 3.15
- Intel oneAPI, CMake 3.20
- IAR, CMake 3.15 and 3.20
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
The CMSIS-DSP matrix binary_q15 tests should not be excluded on the
mps2_an521 board, which is one of the integration platforms.
This test was likely excluded because of the computation precision-
related failures; but, this failure is no longer observed, so there is
no reason to exclude it anymore.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit adds new testcases for the CMSIS-DSP transform test that
enable testing with hardware FPU.
Note that the common properties are relocated to each testcase because
the twister does not support specifying extra filter and tags.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit adds a new testcase for the CMSIS-DSP svm test that
enables testing with hardware FPU.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit adds a new testcase for the CMSIS-DSP support test that
enables testing with hardware FPU.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit adds a new testcase for the CMSIS-DSP statistics test that
enables testing with hardware FPU.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit adds new testcases for the CMSIS-DSP matrix test that
enable testing with hardware FPU.
Note that the common properties are relocated to each testcase because
the twister does not support specifying extra filters and tags.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit adds new testcases for the CMSIS-DSP filtering test that
enable testing with hardware FPU.
Note that the common properties are relocated to each testcase because
the twister does not support specifying extra filters and tags.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit adds a new testcase for the CMSIS-DSP fastmath test that
enables testing with hardware FPU.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit adds a new testcase for the CMSIS-DSP distance test that
enables testing with hardware FPU.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit adds a new testcase for the CMSIS-DSP complexmath test that
enables testing with hardware FPU.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit adds a new testcase for the CMSIS-DSP bayes test that
enables testing with hardware FPU.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit adds a new testcase for the CMSIS-DSP basicmath test that
enables testing with hardware FPU.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
Added test cases that validates if ring buffer handles
correctly case with single producer single consumer from
different contexts.
Signed-off-by: Krzysztof Chruscinski <krzysztof.chruscinski@nordicsemi.no>
DT_PROP for a phandle property should return the node that phandle
points to (similar for DT_PROP_BY_IDX for phandles) and this wasn't
working as the define generator didn't create the proper defines for
phandle(s).
Fix the generator and add some tests to make sure this continues to
work correctly.
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
This commit adds the tests for the newlib retargetable locking
interface, as well as the tests for the internal lock functions that
are supposed to internally invoke the retargetable locking interface.
All of these tests must pass when the toolchain newlib is compiled with
the `retargetable-locking` and `multithread` options, which are
required to ensure that the newlib is thread-safe, enabled. If the
toolchain newlib is compiled with either of these options disabled,
this test will fail.
This commit also adds the userspace testcases to ensure that the newlib
is thread-safe in the user mode.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
This commit adds the `static` keyword to the test functions that are
not intended to be globally available.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
Add two new for-each macros:
- DT_FOREACH_STATUS_OKAY(compat, fn)
- DT_FOREACH_STATUS_OKAY_VARGS(compat, fn, ...)
These can be used to expand "fn" once for every status "okay" node in
the devicetree which has a given compatible. The intended use case is
to allow doing something in C once for each node of a compatible,
but outside of a device driver.
E.g. an application might want to collect an array of structures for a
compatible, where each structure is initialized from a node.
In such cases we don't want people to be forced into using
DT_DRV_COMPAT and instance numbers, because that's in general a hint
that you're doing something driver-like.
Signed-off-by: Martí Bolívar <marti.bolivar@nordicsemi.no>
Split the generic tests for DT_INST_FOREACH_STATUS_OKAY() out of the
function which tests it as applied to device instantiation.
This is just moving code around as prep work for expanding on the set
of macros which are in for-each-node style.
Signed-off-by: Martí Bolívar <marti.bolivar@nordicsemi.no>
The commit
6fe9e408ab (tests/cbprintf_fp: Filter on CONSOLE_HAS_DRIVER)
added filtering on Kconfig symbol incorrectly - it checked
CONSOLE_HAS_DRIVER instead CONFIG_CONSOLE_HAS_DRIVER so
this test was filtered out (and therefore skipped) on all
platforms.
Fix that.
Signed-off-by: Eugeniy Paltsev <Eugeniy.Paltsev@synopsys.com>
Signed-off-by: Evgeniy Paltsev <PaltsevEvgeniy@gmail.com>
Added tests for fully self-contained packages and use of
CBPRINTF_PACKAGE_ADD_STRING_IDXS flag.
Signed-off-by: Krzysztof Chruscinski <krzysztof.chruscinski@nordicsemi.no>
Added support for conversion from a standard package which contains
pointers to read only strings to fully self-contained (fsc) package.
Fsc package contains all strings associated with the package thus
access to read only strings is not needed to format a string.
In order to allow conversion to fsc package, standard package must
contain locations of all string pointers within the package. Appending
that information is optional and is controlled by flags parameter
which was added to packaging API. If option flag is set then
package contains header, arguments, locations of read only strings and
transient strings (each prefixed with string argument location).
Package header has been extended with field which contains number of
read only string locations.
A function for conversion to fsc package has been added
(cbprintf_fsc_package()).
Signed-off-by: Krzysztof Chruscinski <krzysztof.chruscinski@nordicsemi.no>
To be able to get a tokenize DT string without the quotes. Deprecate
also DT_ENUM_TOKEN and DT_ENUM_UPPER_TOKEN.
Signed-off-by: Carlo Caione <ccaione@baylibre.com>
`_FOREACH_` macros do not allow the caller to pass additional arguments
to the `fn`. A series of `_VARGS` variants have been added that allow
the caller to pass arbitrary number of arguments to the `fn`:
```
DT_FOREACH_CHILD_VARGS
DT_FOREACH_CHILD_STATUS_OKAY_VARGS
DT_FOREACH_PROP_ELEM_VARGS
DT_INST_FOREACH_CHILD_VARGS
DT_INST_FOREACH_STATUS_OKAY_VARGS
DT_INST_FOREACH_PROP_ELEM_VARGS
```
Signed-off-by: Arvin Farahmand <arvinf@ip-logix.com>
This commit increases the heap size for the newlib nano test to 2048,
which is the recommended minimum required to ensure proper operation
(see #35799).
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
As we don't use memory allocated in test_calloc, test_no_mem_malloc,
and test_no_mem_realloc. malloc call can be optimized away (that really
happens with ARC MWDT toolchain). That breaks the test. So disable
optimization for these functions.
Signed-off-by: Watson Zeng <zhiwei@synopsys.com>
The coverity report "Out-of-bounds access". The reason is
The room of src and dest buffer is less than the count wanted
to be copyied. So enlarged the src and dest buffer to solve this
issue.
Fixes:#35345
Fixes:#35346
Signed-off-by: Ying ming <mingx.ying@intel.com>
1. Refine some testcases for testing some character output function.
2. Delete some negative testcases which will not invoke fatal error.
Fixes#33822Fixes#33803Fixes#33801Fixes#33800Fixes#33790
Signed-off-by: Ying ming <mingx.ying@intel.com>
This commit adds a new test to verify the thread safety of the C
standard functions provided by newlib.
Only the memory management functions (malloc and free) are tested at
this time; this test will be further extended in the future, to verify
the thread safety and re-entrancy of all supported newlib functions.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
1. Refine some testcases for testing some libc functions, include
strlen(),strcmp(),strncmp(),memcmp(),memstr(),memcpy(),memmove()
strtol(),strtoul().
2. Change some code to make the code style consistent.
Fixes#33794Fixes#33810Fixes#33806
Signed-off-by: Ying ming <mingx.ying@intel.com>
Puts all filter in common section and fix filters. Platforms in
integration_platforms must also be allowed first, otherwise the tests
will never run.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Adding static to local functions. Using void * for %p argument
which often triggered compilation warning.
Signed-off-by: Krzysztof Chruscinski <krzysztof.chruscinski@nordicsemi.no>
The devicetree test will not link the hsdk platforms as
when CONFIG_GPIO=y this enable CONFIG_I2C as the board utilizes
a GPIO expander over I2C. So we exclude building this test on
the hsdk platforms.
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
These should run for minlibc configured with
CONFIG_MINIMAL_LIBC_MALLOC_ARENA_SIZE=0. In all other cases normal
testsuite should run.
How preprocessor #if's were put previously, some configurations,
e.g. prj_newlib.conf, had the testsuite completely skipped.
Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org>
