This commit replaces the use of portNUM_PROCESSORS and configNUM_CORES
macros in all of ESP-IDF. These macros are needed to realize an SMP
scenario by fetching the number of active cores FreeRTOS is running on.
Instead, a new Kconfig option, CONFIG_FREERTOS_NUMBER_OF_CORES, has been
added as a proxy for the FreeRTOS config option, configNUMBER_OF_CORES.
This new commit is now used to realize an SMP scenario in various places
in ESP-IDF.
[Sudeep Mohanty: Added new Kconfig option CONFIG_FREERTOS_NUMBER_OF_CORES]
Signed-off-by: Sudeep Mohanty <sudeep.mohanty@espressif.com>
* All components which won't build (yet) on Linux are excluded.
This enables switching to Linux in an application without
explicitly setting COMPONENTS to main in the main
CMakeLists.txt.
* ESP Timer provides headers for Linux now
* automatically disabling LWIP in Kconfig if it is not available
doc(linux): brought section
"Component Linux/Mock Support Overview" up to date
Fix false positive in coverity where it thinks memory buffers are freed for
statically created ring buffers, leading to a "USE_AFTER_FREE" warning in
vRingbufferDeleteWithCaps().
This commit adds APIS to create a ring buffer with specific memory caps. The
following APIs have been added:
- xRingbufferGetStaticBuffer()
- xRingbufferCreateWithCaps()
- vRingbufferDeleteWithCaps()
When building on linux/host compilers (e.g., GCC), the compiler may add padding
depending on the size and order of the member types.
This commit fixes the ordering or the StaticRingbuffer_t such that it matches
the internal Ringbuffer_t. The "_Static_assert" is always enabled for all
compilers.
Closes https://github.com/espressif/esp-idf/issues/11726
Ring buffers permit sending items of 0 size where:
- No-split/allow-split buffers will send an item (i.e., header) but the item
will have a 0 sized data section
- Byte buffers will simply return pdTRUE
This commit fixes an incorrect assert w.r.t 0 item size in
xRingbufferSendAcquire() and also adds a note to the API descriptions regarding
0 item sizes. A test case has also been added.
This commit replaces the existing TX/RX semaphores with FreeRTOS task event
lists (similar to how FreeRTOS queues implement task blocking). Using task
event lists allows the ring buffer object to be smaller and various ring buffer
functions to be faster. The ring buffer queue set impelementation was also
updated as a result.
xSemaphoreGive won't be executed in configASSERT and semaphore will stay locked if NDEBUG (idf v5) or CONFIG_FREERTOS_ASSERT_DISABLE (idf v3, v4) are defined.
esp-ringbuf funtion placement is now controlled by its own configs:
CONFIG_RINGBUF_PLACE_FUNCTIONS_INTO_FLASH and CONFIG_RINGBUF_PLACE_ISR_FUNCTIONS_INTO_FLASH
Closes https://github.com/espressif/esp-idf/issues/9198
This commit removes the usage of all legacy FreeRTOS data types that
are exposed via configENABLE_BACKWARD_COMPATIBILITY. Legacy types can
still be used by enabling CONFIG_FREERTOS_ENABLE_BACKWARD_COMPATIBILITY.
This commit updates the chip independent system chapters of the
programming guide for esp32s3.
Signed-off-by: Sudeep Mohanty <sudeep.mohanty@espressif.com>
Add TRY_ENTRY_CRITICAL() API to all for timeouts when entering critical sections.
The following port API were added:
- portTRY_ENTER_CRITICAL()
- portTRY_ENTER_CRITICAL_ISR()
- portTRY_ENTER_CRITICAL_SAFE()
Deprecated legacy spinlock API in favor of spinlock.h. The following API were deprecated:
- vPortCPUInitializeMutex()
- vPortCPUAcquireMutex()
- vPortCPUAcquireMutexTimeout()
- vPortCPUReleaseMutex()
Other Changes:
- Added portMUX_INITIALIZE() to replace vPortCPUInitializeMutex()
- The assembly of the critical section functions ends up being about 50 instructions longer,
thus the spinlock test pass threshold had to be increased to account for the extra runtime.
Closes https://github.com/espressif/esp-idf/issues/5301
This commit fixes a bug in no-split and allow-split ring buffers free buffer size calculation.
When the free size available in the buffers less than the size of one item header,
the function prvGetCurMaxSizeNoSplit/AllowSplit() incorrectly returned the maxItemSize instead of 0.
This is due to the comparision between a negative and a positive value
where both operands are treated as unsigned during the comparision operation,
thereby treating the negative operand as a large integer.
Also added new unit tests to test buffer-full and almost-full conditions
where this scenario is likely to be hit.
Closes https://github.com/espressif/esp-idf/issues/7344
Closes https://github.com/espressif/esp-idf/pull/7371
The release of the semaphore indicating the item was successfully sent must be the last semaphore released. The receiver may be in another task and may delete the Ringbuffer (such as with a return code across tasks design pattern) if they are through with the Ringbuffer.
The function xRingbufferSendAcquire followed by xRingbufferSendComplete had the semaphores released in the proper order and that same pattern should have been used in xRingbufferSend and xRingbufferSendFromISR. This commit fixes this order.
Issue (IDFGH-6030) #7716 describes the problem in more detail.
Closes IDFGH-6030, https://github.com/espressif/esp-idf/issues/7716
Closes IDFGH-6036, https://github.com/espressif/esp-idf/pull/7721
* Target components pull in xtensa component directly
* Use CPU HAL where applicable
* Remove unnecessary xtensa headers
* Compilation changes necessary to support non-xtensa gcc types (ie int32_t/uint32_t is no
longer signed/unsigned int).
Changes come from internal branch commit a6723fc
