bugfix: esp32s3 DCache data memory is retention dma inaccessible
Closes IDFCI-1409, IDFCI-1410, IDFCI-1411, IDFCI-1412, and IDFCI-1413
See merge request espressif/esp-idf!19365
Since cpu retention dma use rc fast as clk source, so rc_fast_digi
will be enabled when we config to pd cpu. And cpu retention does not
need rc fast keep on during light sleep. So, if we use rc_fast_digi
to determine whether rc fast can be powered down, then cpu and and
rc fast cannot pd at the same time.
- Remove esp_cpu_in_ocd_mode() from esp_cpu.h. Users should call esp_cpu_dbgr_is_attached() instead.
- Remove esp_cpu_get_ccount() from esp_cpu.h. Users should call esp_cpu_get_cycle_count() instead.
- Remove esp_cpu_set_ccount() from esp_cpu.h. Users should call esp_cpu_set_cycle_count() instead.
- Other IDF components updated to call esp_cpu_dbgr_is_attached(), esp_cpu_get_cycle_count() and esp_cpu_set_cycle_count() as well.
This commit marks all functions in soc_hal.h and soc_ll.h as deprecated.
Users should use functions from esp_cpu.h instead.
Also added missing wrap funcions for esp_cpu_stall() in test_panic.c files.
This commit marks all functions in interrupt_controller_hal.h, cpu_ll.h and cpu_hal.h as deprecated.
Users should use functions from esp_cpu.h instead.
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
The following two functions in bootloader_support are private now:
* esp_secure_boot_verify_sbv2_signature_block()
* esp_secure_boot_verify_rsa_signature_block()
They have been moved into private header files
inside bootloader_private/
* Removed bootloader_reset_reason.h and
bootloader_common_get_reset_reason() completely.
Alternative in ROM component is available.
* made esp_efuse.h independent of target-specific rom header
Some esp32c2 boards will be produced with a 26 MHz XTAL. This commit
adds the basic Kconfig option for this type of hardware.
Support for CONFIG_ESP32C2_XTAL_FREQ_26 in other areas of IDF will be
implemented in subsequent commits.
Some of the Memory Protection (internal) API functions dealing with per-CPU operations were missing appropriate handling of the CPU count actually configured by CONFIG_FREERTOS_UNICORE. The flaw was fixed across all the places found in the code as the issue was of general type
When creating G0 layer, some regi2c_*.h headers were moved out from
esp_hw_support (G1) to soc (G0). In order to be consistent with that change,
move all the remaining regi2c_*.h headers to soc too.
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 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
ESP32C3/ESP32S3: Fix cpu crash bug when wakeup from lightsleep for memory data miss
Closes IDF-162 and IDF-4923
See merge request espressif/esp-idf!17823
When enable sleep reject before this fix, we have two limitations:
1. it must be light sleep
2. RTC GPIO wakeup source must be set
We require light sleep because `esp_deep_sleep_start` function has
been declared with "noreturn" attribute, So developers don't expect
that this function may return (due to an error or a sleep reject).
But the requirement for RTC GPIO wakeup source is not reasonable for
all chips. This requirement exists because ESP32 only supports RTC GPIO
and SDIO sleep reject sources. But later chips support all sleep reject
sources.
This fix brings the following changes:
for ESP32: RTC GPIO and SDIO sleep reject sources can be enabled
when corresponding wakeup source is set.
for later chips: all sleep reject sources can be enabled when
corresponding wakeup source is set.
Since ulp wakeup signal are connected to ulp int raw(except esp32), we
need to clear ulp int raw before sleep when ulp wakeup enabled. Otherwise,
if the ulp int raw is already set, chip will not sleep properly.
Closes https://github.com/espressif/esp-idf/issues/6229
introduced in e44ead5356
1. The int8M power domain config by default is PD. While LEDC is using
RTC8M as clock source, this power domain will be kept on.
But when 8MD256 is used as RTC clock source, the power domain should
also be kept on.
On ESP32, there was protection for it, but broken by commit
e44ead5356. Currently the power domain
will be forced on when LEDC is using RTC8M as clock source &&
!int8m_pd_en (user enable ESP_PDP_DOMAIN_RTC8M in lightsleep). Otherwise
the power domain will be powered off, regardless of RTC clock source.
In other words, int8M domain will be forced off (even when 8MD256
used as RTC clock source) if LEDC not using RTC8M as clock source, user
doesn't enable ESP_PDP_DOMAIN_RTC8M, or in deep sleep.
On later chips, there's no such protection, so 8MD256 could't be used as
RTC clock source in sleep modes.
This commit adds protection of 8MD256 clock to other chips. Fixes the
incorrect protection logic overriding on ESP32. Now the power domain
will be determiend by the logic below (order by priority):
1. When RTC clock source uses 8MD256, power up
2. When LEDC uses RTC8M clock source, power up
3. In deepsleep, power down
4. Otherwise determined by user config of ESP_PDP_DOMAIN_RTC8M,
power down by default. (This is preferred to have highest
priority, but it's kept as is because of current code structure.)
2. Before, after the macro `RTC_SLEEP_CONFIG_DEFAULT` decides dbias, the
protection above may force the int8m PU. This may cause the inconsistent
of dbias and the int8m PU status.
This commit lifts the logic of pd int8m/xtal fpu logic to upper layer
(sleep_modes.c).
Related: https://github.com/espressif/esp-idf/issues/8007, https://github.com/espressif/esp-idf/pull/8089
temp
check_kconfigs.py would only correctly handle "source", but not "rsource", "osource"
or "orsource".
Fixed any warnings detected after updating the test.
Some components were including esp_timer.h without declaring a
dependency on esp_timer component. This used to work due to a
transitive public dependency on esp_timer from freertos component.
Add explicit dependencies where needed.
Also some source files were using esp_timer functions without
including the header file. This used to work because esp_timer.h was
included from freertos port header file. This commit adds esp_timer.h
includes where needed.
