Setting the send status flag `sndchange` after calling the can->ops->sendmsg function
could lead to a race condition if a transmission timeout occurs, resulting in incorrect state handling.
This patch moves the operation of setting the `sndchange` flag to before the call to can->ops->sendmsg.
This ensures that the mailbox's status is correctly marked as "sending" before the hardware begins transmission,
making the driver's state management more robust and reliable, especially in handling exceptions like timeouts.
Additionally, new macros for CAN filter modes have been added in dev_can.h.
Use "Test" instead of "Unit Testcases" to make string shorter.
Use uppercase to make it look more eye-catching.
Signed-off-by: Chen Wang <unicorn_wang@outlook.com>
Also:
- Add "RT_" prefix for utest config options.
- Rename the case names to following the naming rule.
Signed-off-by: Chen Wang <unicorn_wang@outlook.com>
Also,
- Rename UTEST_SMP_CALL_FUNC to RT_UTEST_SMP_CALL_FUNC.
- Add depends on RT_USING_SMP in Kconfig for RT_UTEST_SMP_CALL_FUNC
- Follow the naming rules of utest case to update the cases' name.
Signed-off-by: Chen Wang <unicorn_wang@outlook.com>
* doxygen: remove @return command if function return void
When doxygen is upgraded to v1.9.8 (on ubuntu 24.04), doxygen
reports: "doxygen error: found documented return type for xxx
that does not return anything" for those functions which return
void but declare "@return" in doxygen comment.
Solution: remove "@return" for those cases, and update
guide document for how to write doxgen comment for functions.
Signed-off-by: Chen Wang <unicorn_wang@outlook.com>
* doxygen: fixed a minor typo for uart doc
This error is found when building with doxygen 1.9.8, but not
detecetd on 1.9.1.
Signed-off-by: Chen Wang <unicorn_wang@outlook.com>
The names and number of macro parameters in doxygen comments
must match the actual macro definition; otherwise, a warning
will be generated.
Our goal is to eliminate all warnings.
Signed-off-by: Chen Wang <unicorn_wang@outlook.com>
- Added support for non-blocking mode CAN message sending, including software ring buffer management and dynamic memory allocation options.
- Improved related comments and error handling.
- Updated example code to demonstrate the usage of both blocking and non-blocking sending modes, and corrected some structure field naming and macro definitions.
src/utest/device_tc.c is testing API: rt_device_find(),
which is a function defined in components/drivers/core/device.c.
So it should be a testcase for drivers core, not for core.
Move it to under components/drivers/core.
Signed-off-by: Chen Wang <unicorn_wang@outlook.com>
Many modules' utests currently don't support enabling all
tests at once. Furthermore, some modules' tests are complex,
for example due to their numerous dependencies on other
modules. This makes it nearly impossible to enable all
tests with a single global switch. Consequently, the
previously defined `RT_UTEST_USING_ALL_CASES` has lost
its original meaning.
We recommend deprecating this configuration switch. If a
module needs to enable a group of functional tests through
its own configuration, this local enable all switch should
be implemented by the module itself, and a global RTT enable
switch will no longer be provided.
If such a requirement arises in the future, we recommend
careful design, especially considering how to ensure that
turning on a single switch enables all dependencies for
all involved modules for ease of use.
Signed-off-by: Chen Wang <unicorn_wang@outlook.com>
The names of the utest cases defined in their respective
modules have been updated.
Some utest case names are not yet in their respective modules
because their paths need to be finalized before their unique
names can be determined. Currently, these names do not
appear to conflict with the unified names. These include:
- utest cases still in examples
- bsp/qemu-virt64-riscv/applications/test/test_vector/test_vector.c.
The entire test case should probably be placed in libcpu/risc-v
rather than bsp.
Signed-off-by: Chen Wang <unicorn_wang@outlook.com>
Append a null statement for "Any statement may be preceded by a prefix that declares an identifier as a label name."
Signed-off-by: GuEe-GUI <2991707448@qq.com>
