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.
Note: Added a small fix to exclude esp_clk.c from bootloader builds in the
legacy build system (this is already the case in CMake).
define: lock_bits = (lock->status & LOCK_MASK) >> LOCK_SHIFT; This `lock_bits` is the Bit 29-20 of the lock->status
1. spi_hdl_1:
acquire_end_core():
uint32_t status = lock_status_clear(lock, dev_handle->mask & LOCK_MASK);
Becuase this is the first `spi_hdl_1`, so after this , lock_bits == 0`b0. status == 0
2. spi_hdl_2:
acquire_core:
uint32_t status = lock_status_fetch_set(lock, dev_handle->mask & LOCK_MASK);
Then here status is 0`b0, but lock_bits == 0`b10. Because this is the `spi_hdl_2`
3. spi_hdl_2:
`acquire_core` return true, because status == 0. `spi_bus_lock_acquire_start(spi_hdl_2)` then won't block.
4. spi_hdl_2:
spi_device_polling_end(spi_hdl_2).
5. spi_hdl_1:
acquire_end_core:
status is 0, so it cleas the lock->acquiring_dev
6. spi_hdl_2:
spi_device_polling_end:
assert(handle == get_acquiring_dev(host)); Fail
Closes https://github.com/espressif/esp-idf/issues/8179
update gpio_sig at `spics_out` array in each spi_periph.c of chips later than s2
then `spi_bus_add_device` can correctly distribute gpio_signals for cs_signal
Closes https://github.com/espressif/esp-idf/issues/8876
Due to unit test, we can't specify runners, for adc related tests.
Whereas adc tests are easily influenced by board hw condition, e.g.
pullup/down value.
This commit increases the thresh to 200mV.
Fixed calling esp_ble_get_cur_sendable_packets_num() sometimes crashes when bluetooth is disconnecting (backport v4.4)
See merge request espressif/esp-idf!20042
- Fixed the performance issue of BLE sync scan in coexistence scenario when PLL track is enabledFixed the usage of PLL track related baseband IRQs
- Fixed connect failed due to error sync found signal
- Fixed BLE disconnect failed due to terminate_ind have sent but no ACK is received
- Added ADV mode error check in ADV rx isr handler function
- Fixed BLE RAL_UNDERRUN and TX_CRYPT_ERR assert
- Disable scan continue
- Fixed scan forever if scan continue is disable
bt:Fixed deadlock due to wrong parameter when calling btc_transfer_context() after calling esp_bluedroid_disable()(v4.4)
See merge request espressif/esp-idf!19951
If a windowoverflow8 happened after changing the SP, the exception handler would look for
the extra save area by looking at the previous frame's SP. This SP would be a garbage value
and could cause the windowoverflow exception to write to invalid memory ares.
