This commit reverts the previous "taskENTER_CRTIICAL();" so that the argument
is now provided for better code readability. The names of the spinlocks have
also been updated.
The execution order of the "task_event_source" and "application_task" tasks
on startup can vary. This will result a different order of logs, thus leading
to the example test failing.
This commit synchronizes both tasks on startup.
This commit updates various FreeRTOS functions to call the newly added
prvCheckForYield() and prvCheckForYieldUsingPriority() when checking for
yielding. This allows the source code to match upstream more closely.
When a FreeRTOS function unblocks a task, that function will check whether
the unblocked task requires a yield to be called. This is currently done by
having each function individually check if the unblocked task has a higher
priority than the both cores, and yielding the appropriate core.
This commit adds the macros list below to abstract away the yielding checking
procedure. This will allow the code to match upstream more closely.
- prvCheckForYield()
- prvCheckForYieldUsingPriority()
esp_light_sleep_start() will stall the other CPU via esp_ipc_isr_stall_other_cpu(). After stalling the other CPU,
will call esp_clk_... API which themselves take locks. If the other stalled CPU is holding those locks, this will
result in a deadlock.
This commit adds a workaround calling esp_clk_private_lock() to take the lock before stalling the other CPU.
esp_clk used to be protected using _lock_t (i.e., a FreeRTOS Mutex). However, esp_clk API is current called from
from critical sections, thus mutex should not be used (as they can be blocking).
This commit updates esp_clk to use spinlocks for critical sections instead.
The event group unit tests would previously use a single call bit (the
"BIT_CALL" macro) to unblock all the test tasks. However, if one or more tasks
were delayed in calling xEventGroupWaitBits(), the other tasks would clear the
test bits, leading to the test deadlocking.
This commit updates the bits used so that each task gets their own CALL and
RESPONSE bits.
ESP32-S3 stacks have currently have less useable space compared to other targets (due to
AI coprocessor context registers). This has caused a stack overflow in the IPC tasks under
some configurations. This commit increases the default IPC stack size for the ESP32-S3.
Spinlocks themselves do not constitute critical sections as after a spinlock is acquired, interrupts
can remain enabled. However, there are some places where spinlocks are used direclty instead of using
the portMUX_TYPE and portENTER_CRITICAL_...() APIs. This commit fixes those calls.
This function removes the following legacy atomic CAS functions:
From compare_set.h (file removed):
- compare_and_set_native()
- compare_and_set_extram()
From portmacro.h
- uxPortCompareSet()
- uxPortCompareSetExtram()
Users should call esp_cpu_compare_and_set() instead as this function hides the details
of atomic CAS on internal and external RAM addresses.
Due to the removal of compare_set.h, some missing header includes are also fixed in this commit.
esp_cpu_compare_and_set() abstracts the atomic compare-and-set instruction by
hiding the details of whether the target variable is in internal or external
RAM. This commit updates "spinlocks.h" as follows:
- esp_cpu_compare_and_set() is now called instead of "compare_set.h"
- Refactored spinlock logic to be more optimized and have more stringent sanity checks
This commit fixes esp_cpu_compare_and_set() in the following ways
- Removed call to esp_ptr_external_ram() as it incurred > 80 CPU cycles (due to multiple nested
function calls, and those functions not being in IRAM). We now check manually if the pointer
is in external RAM for increased speed.
- Fixed infinite wait when attempting to get the external_ram_cas_lock. The function should
return immediatley if any part of the compare and set call fails.
- The preprocessor conditions of esp_cpu_compare_and_set() to depend on CONFIG_SPIRAM instead
of SOC_SPIRAM_SUPPORTED. Even if the target supports SPIRAM, we only need the external RAM
compare and set feature if SPIRAM is enabled.
Also fixed incorrect inclusion of esp_intr_alloc.h in esp_cpu.h
This commit fixes vTaskSuspendAll() and xTaskResumeAll() in the following ways.
- For vTaskSuspendAll()
- Sync function source code with upstream single core version
- Clearly mark IDF additions
- For xTaskResumeAll()
- Sync function source code with upstream single core version
- Clearly mark IDF additions
- Fix bug where cores other than core 0 were allowed to unwind pended ticks
This commit fixes the following issues:
- Clearly mark IDF additions in vTaskStepTick() and xTaskCatchUpTicks()
- Add back vTaskSuspendAll()/xTaskResumeAll() to xTaskCatchUpTicks(). Scheduler
suspension and resumption is required to unwind the pended ticks.
Previously, xTaskIncrementTick() would always trigger a yield (i.e., return pdTRUE)
when called on core 1. This would mean core 1 would call vTaskSwitchContext() on
every tick, leading to uneccesary CPU overhead.
This commit fixes xTaskIncrementTick() in the following ways:
- Clearly mark IDF additions in xTaskIncrementTick()
- Moved esp_vApplicationTickHook() to be called direclty in xPortSysTickHandler() in order to
simplify xTaskIncrementTick().
- Only core 0 calls xTaskIncrementTick() in order to simplify the function's logic. Core 0 is
solely responsible for managing xTickCount and xPendedTicks
- All other cores call xTaskIncrementTickOtherCores() instead which is a simplified version of
xTaskIncrementTick() and handles the following:
- Check if core 0 has unblocked a higher priority task to run
- Check if the current core requires time slicing
- Call vApplicationTickHook()
FreeRTOS uses a single "xSchedulerRunning" variable to tack whether the
scheduler has started, and this variable is set to "pdTRUE" by core 0
via calling vTaskStartScheduler().
However, with SMP FreeRTOS, there is a race condition where core 0 has
already started the scheduler and another core has not called xPortStartScheduler()
yet and calls some FreeRTOS API. Thus the resultant FreeRTOS API can
cause errors as it thinks the scheduler has started.
This commit adds a temporary workaround (by having each core maintain their
own "xSchedulerRunning" variable.
The "FreeRTOS Event Groups" main task will only wait a single tick for the created
tasks to set their response bits. This short delay may not be sufficent if the tick
frequency is high.
This commit updates the test so that
- the main task waits indefinitely for all the response bits to be set.
- created tasks are cleand up by the main task
FreeRTOS synchronization primitives (e.g., queues, eventgroups) use various event lists (i.e., task lists) to track what
tasks are blocked on a current primitive. Usually these event lists are accessed via one of the event lists functions
(such as vTask[PlaceOn|RemoveFrom]UnorderedEventList()), which in turn ensure that the global task list spinlock
(xTaskQueueMutex) is taken when accessing these lists.
However, some functions in event_groups.c manually traverse their event lists. Thus if a tick interrupt occurs on
another core during traversal and that tick interrupt unblocks a task on the event list being traversed, the event list
will be corrupted.
This commit modifies the following event_groups.c functions so that they take the global task list lock before
traversing their event list.
- xEventGroupSetBits()
- vEventGroupDelete()
The CONFIG_FREERTOS_PLACE_FUNCTIONS_INTO_FLASH allows some FreeRTOS functions to be placed
into flash, thus saving some IRAM. However, this feature was previously not supported for
SMP FreeRTOS.
This commit adds CONFIG_FREERTOS_PLACE_FUNCTIONS_INTO_FLASH support for SMP FreeRTOS. There
are now separate linker fragment files for IDF FreeRTOS and SMP FreeRTOS, named "linker.lf"
and "linker_smp.lf" respectively.
This commit fixes typos that are breaking changes (i.e., macros, types, and functions)
in the bluedroid API. Migration notes for v5.0 have also been added.
Closes https://github.com/espressif/esp-idf/issues/6887
The following functions are only defined when CONFIG_PM_ENABLE is defined:
- wifi_apb80m_request()
- wifi_apb80m_release()
This commit makes their linker script mapping statements also dependent on CONFIG_PM_ENABLE
thus fixing some linker errors when CONFIG_PM_ENABLE is disabled.
The TLSP deletion callback feature is not compatible with the CONFIG_FREERTOS_ENABLE_STATIC_TASK_CLEAN_UP
option. However, the "freertos_options" unit test configuration will enable that option.
This commit disables all CONFIG_FREERTOS_ENABLE_STATIC_TASK_CLEAN_UP related features when compiling with
SMP FreeRTOS.
SMP FreeRTOS uses a single kernel lock for all critical sections. There is a known
issue with esp_ipc_isr_stall_other_cpu() that can cause dead if the other CPU is
already in a critical section.
This commit adds a temporary workaround to reduce the chance of deadlock by taking
the SMP FreeRTOS kernel lock first before stalling the other CPU.
See IDF-5257 for more details.
The Xtensa FreeRTOS port does not save the threadptr register when
doing a voluntary yield. This can result in a crash when multiple
tasks used the threadptr register and call "taskYIELD()".
This commit adds the threadptr register to the solicited stack frame.
This commit makes changes to cpu_ll.h, cpu_hal.h, and interrupt_controller_hal.h:
- Moved to esp_hw_support in order to be deprecated in the future
- HAL/LL API now route their calls to esp_cpu.h functions instead
Also updated soc_hal.h as follows:
- Removed __SOC_HAL_..._OTHER_CORES() macros as they dependend on cpu_hal.h
- Made soc_hal.h and soc_ll.h interfaces always inline, and removed soc_hal.c.
This commit also updates the XCHAL_ERRATUM_572 workaround by
- Removing it's HAL function and invoking the workaround it directly the bootloader
- Added missing workaround for the ESP32-S3
This commit moves the dedicated GPIO LL and HAL functions from
cpu_ll.h to dedic_gpio_cpu_ll.h.
- cpu_ll_enable_cycle_count() has also been removed due to lack of feasible usage scenarios
This commit updates the esp_cpu.h API. The new API presents a new
abstraction of the CPU where CPU presents the following interfaces:
- CPU Control (to stall/unstall/reset the CPU)
- CPU Registers (to read registers commonly used in SW such as SP, PC)
- CPU Interrupts (to inquire/allocate/control the CPUs 32 interrupts)
- Memory Port (to configure the CPU's memory bus for memory protection)
- Debugging (to configure/control the CPU's debugging port)
Note: Also added FORCE_INLINE_ATTR to the DoxyFile in order to pass doc
builds for esp_cpu.h
Previously the RV port was routing IDF style critical section API to call FreeRTOS style critical section API.
For example, a call to "portENTER_CRITICAL(mux)" would eventually call `vTaskEnterCritical()" via the following call flow:
- portENTER_CRITICAL(mux)
- vPortEnterCritical()
- portSET_INTERRUPT_MASK_FROM_ISR()
- vTaskEnterCritical()
This commit fixes the IDF style critical section by making sure that they are completely orthogonal to FreeRTOS critical sections
This commit fixes the following build errors in SMP FreeRTOS when building with GCC 11:
- "-Wattributes" warning due to conflicting attributes on xPortInterruptedFromISRContext()
- "-Wtype-limits" warning due to taskVALID_CORE_ID() not casting xCoreID
The behavior of vTaskSuspendAll() has changed in SMP FreeRTOS. This commit
disables scheduler suspension related unit tests for SMP FreeRTOS.
The tests will be updated in IDF-5201
The behavior of portSET_INTERRUPT_MASK_FROM_ISR() has changed in SMP FreeRTOS. It's
previous behavior is now implemented in portDISABLE_INTERRUPTS() and portRESTORE_INTERRUPTS().
This commit replaces all portSET_INTERRUPT_MASK_FROM_ISR() and portCLEAR_INTERRUPT_MASK_FROM_ISR()
calls with portDISABLE_INTERRUPTS() and portRESTORE_INTERRUPTS() respectively
This commit fixes the systimer_counter_value_t by adding a resreved field so that the
type fills 64-bits.
Without the reserved field, when compiling with -O0 optimization, the unoccupied high
bits would not be initalized by the compiler, leading to systimer_hal_get_counter_value()
returning a garbage value.
Build system tests check that xtensa_vectors.S is rebuilt in some of the CMake build system
tests. However, the path of this file may change when FreeRTOS SMP is enabled.
This commit fixes the CMake build system test by selecting another ".S" file who's path
does not change based on configuration.
When configNUM_CORES = 1, vTaskCoreAffinityGet() is not defined. This
commit fixes the TWDT to omit calls to vTaskCoreAffinityGet() when building
for unicore.
The "Test Queue sets thread safety" tests relied on IDF FreeRTOS specific
scheduling behavior, thus resulting in the test getting stuck when running
with SMP FreeRTOS.
This commit updates and refactors the tests to work on both IDF and SMP FreeRTOS.
This commit fixes how SMP FreeRTOS feeds the INT WDT. Previously, the
port used vApplicationTickHook(). Now, all cores will feed the INT WDT
in xPortSysTickHandler().
This commit updates the watchdog documentation as follows:
- Removes IWDT API reference as it is now considered private API
- Rewrite IWDT section
- Fix headers, types, and formatting
This commit does the following:
- Update existing unit tests that use the TWDT to call the new
esp_task_wdt_init() API
- Add a set of dedicate TWDT unit tests
- Updates the TWDT example
This commit moidifies the TWDT as follows:
- Adds a feature to allows subscribing arbitrary users to the TWDT
- Changes esp_task_wdt_init() API to accept configuration structure
- Changes esp_task_wdt_init() and esp_task_wdt_deinit() to subscribe/unsubscribe
idle tasks of various cores.
- Adds support for SMP FreeRTOS idle tasks
- Updates startup code TWDT initialization
- Updates API documentation
This commit refactors the task watchdog as follows:
- Renamed variables, types, and functions
- Replaced manual linked list implementation with SLIST()
- Moved calloc()/free() calls out of critical sections
- Shortened ISR critical sections
- Updated API description
- Updated code formatting
This commit refactors FreeRTOS Kconfig options as follows:
- Grouped them into Kernel and Port submenus
- Renamed Kernel option promppts to match upstream names
- Simplified some option descriptions
- Added "Todo" markers related to SMP support
This commit updates and refactors the AFL host test mock as follows:
- Isolates the mock so that it only relies on lwip component and linux stdlib headers
- Grouped all mock functions in 'esp32_mock.h/c'
- Updates the lwip esp32 port files to include 'esp32_mock.h' when building for linux
- Use host stdlib headers instead of xtensa stdlib headers
This commit moves the IDF API additions from task.h/task.c to seperate header/source files.
- Declarations moved to "idf_additions.h"
- Definitions moved to "freertos_task_c_additions.h"
The API descriptions have also been updated.
This commit disables portUSING_MPU_WRAPPERS for the FreeRTOS SMP xtensa port.
This was previously enabled due to the need to keep a CPSA (coprocessor save
area pointer) in the TCB. The CPSA pointer is now calculated at run time.
Darian