Previously esp_restart would stall the other CPU before enabling RTC_WDT.
If the other CPU was executing an s32c1i instruction, the lock signal
from CPU to the arbiter would still be held after CPU was stalled. If
the CPU running esp_restart would then try to access the same locked
memory pool, it would be stuck, because lock signal would never be
released.
With this change, esp_restart resets the other CPU before stalling it.
Ideally, we would want to reset the CPU and keep it in reset, but the
hardware doesn't have such feature for PRO_CPU (it is possible to hold
APP_CPU in reset using DPORT register). Given that ROM code will not use
s32c1i in the first few hundred cycles, doing reset and then stall seems
to be safe.
In addition to than, RTC_WDT initialization is moved to the beginning of
the function, to prevent possible lock-up if CPU stalling still has any
issue.
1. Make sure that 8MD256 clock used to estimate XTAL frequency is enabled
before trying to use rtc_clk_cal_ratio.
This fixes "Bogus XTAL frequency: 0 MHz" warnings after software reset.
2. Don't call rtc_clk_xtal_freq_estimate if XTAL frequency is already
known. This reduces startup time after deep sleep or software reset.
3. Compare known XTAL frequency and estimated one before printing a
warning. This fixes "Possibly invalid CONFIG_ESP32_XTAL_FREQ setting
(40MHz). Detected 40 MHz." warnings.
Previous implementation waited for 20us after setting
RTC_CNTL_SOC_CLK_SEL_XTL register, using ets_delay_us, assuming that
the CPU was running at XTAL frequency. In reality, clock switch happened
on the next RTC_SLOW_CLK cycle, and CPU could be running at the previous
frequency (for example, 240 MHz) until then.
ets_delay_us would wait for 20 us * 40 cycles per us = 800 CPU cycles
(assuming 40 MHz XTAL; even less with a 26 MHz XTAL).
But if CPU was running at 240 MHz, 800 cycles would pass in just 3.3us,
while SLOW_CLK cycle could happen as much as 1/150kHz = 6.7us after
RTC_CNTL_SOC_CLK_SEL_XTL was set. So the software would not actually wait
long enough for the clock switch to happen, and would disable the PLL
while CPU was still clocked from PLL, leading to a halt.
This implementation uses rtc_clk_wait_for_slow_cycle() function to wait
until the clock switch, removing the need to wait for a fixed number of
CPU cycles.
Some RTC features are synchronized to RTC_SLOW_CLK, so sometimes
software needs to wait for the next slow clock cycle.
This function implements waiting using Timer Group clock calibration
feature.
append adc support and api
- esp_err_t adc2_config_width(adc_bits_width_t width_bit);
- esp_err_t adc2_config_channel_atten(adc2_channel_t channel, adc_atten_t atten);
- int adc2_get_voltage(adc2_channel_t channel);
The mutex is common across all the threads. It needn't be held across
the init_routine() call as long as the 'once' behaviour is guaranteed
Saw a deadlock case, where init_routine of one thread was waiting for
the completion of init_routine in another thread.
t2: wait for command
t1: pthread_once:
lock once_mux
init_routine:
inform thread t2
wait for signal from t2
t2: received command
pthread_once
lock once_mux (already held by t1)
---- Deadlock ----
Reported from:
https://github.com/espressif/esp-idf/issues/703https://github.com/espressif/esp-idf/issues/917
In uart driver we didn't change the default value of tx idle num, so there would be a delay after tx FIFO is empty.
1. Add API to set tx idle interval before next data transmission. (The UART hardware can add an interval after tx FIFO is empty).
2. Set default tx idle interval to zero.
3. Add hardware disable in uart driver delete function.
1. add sw interrupt cause osi to controller.
2. modify the kconfig to improve the option view.
3. add option of the cpu core which bluedroid run.
4. add option of the cpu core which bluetooth controller run.
Previously ioctl was declared as a macro in lwip/sockets.h.
Disabling LWIP_POSIX_SOCKETS_IO_NAMES removed that declaration.
This adds sys/ioctl.h file and the missing declaration.
Also adds missing includes in vfs.c.
Reported from different sources from github or bbs:
https://github.com/espressif/esp-idf/issues/680https://github.com/espressif/esp-idf/issues/922
We tested reading several sensor or other I2C slave devices, if the power and SDA/SCL wires are in proper condition, everything works find with reading the slave.
If we remove the power supply for the slave during I2C is reading, or directly connect SDA or SCL to ground, this would cause the I2C FSM get stuck in wrong state, all we can do is the reset the I2C hardware in this case.
After this commit, no matter whether the power supply of I2C slave is removed or SDA / SCL are shorted to ground, the driver can recover from wrong state.
We are not sure whether this the save issue with the reported one yet, but to make the driver more robust.
Further information:
1. For I2C master mode, we have tested different situations, e.g., to short the SDA/SCL directly to GND/VCC, to short the SDA to SCL, to un-plug the slave device, to power off the slave device. Under all of those situations, this version of driver can recover and keep working.
2. Some slave device will die by accident and keep the SDA in low level, in this case, master should send several clock to make the slave release the bus.
3. Slave mode of ESP32 might also get in wrong state that held the SDA low, in this case, master device could send a stop signal to make esp32 slave release the bus.
Modifications:
1. Disable I2C_MASTER_TRAN_COMP interrupt to void extra interrupt.
2. Disable un-used timeout interrupt for slave.
3. Add bus reset if error detected for master mode.
4. Add bus clear if SDA level is low when error detected.
5. Modify the argument type of i2c_set_pin.
6. add API to set timeout value
7. add parameter check for timing APIs
This commit makes configUSE_TRACE_FACILITY and
configUSE_STATS_FORMATTING_FUNCTIONS configurable in kconfig. Test cases fro the
functions enabled by the two configurations above have also been added.
Test cases for the following functions have been added...
- uxTaskGetSystemState()
- uxTaskGetTaskNumber()
- vTaskSetTaskNumber()
- xEventGroupClearBitsFromISR()
- xEventGroupSetBitsFromISR()
- uxEventGroupGetNumber()
- uxQueueGetQueueNumber()
- vQueueSetQueueNumber()
- ucQueueGetQueueType()
Test cases for the following functions were not required...
- prvListTaskWithinSingleList()
- prvWriteNameToBuffer()
- vTaskList()
Internal byte accessible memory starts with Internal ROM 1 at 0x3FF90000.
Region of RTC fast memory starting at 0x3FF80000 is not used in IDF as
it is mapped to PRO CPU only.
During a call to multi_heap_malloc(), if both these conditions were true:
- That heap only has one block large enough for the allocation
(this is always the case if the heap is unfragmented).
- Another allocation is simultaneously occurring in the same heap.
... multi_heap_malloc() could incorrectly return NULL.
This caused IDF heap_caps_malloc() and malloc() to also fail, particularly
often if only one or two heaps had space for the allocation (otherwise
heap_caps_malloc() fails over to the next heap).
1. Fix WiFi timer enable twice issue
2. Fix a softAP compatibility issue
3. Fix ebuf alloc counter issue
4. remove nvs_flash_init() in esp_wifi_init()
5. Fix scan only find few APs after set bssid
6. Fix can not connect to hidden SoftAP
7. Remove group key entry before connecting ot AP
- _FS_TINY: disables per-file caches
- _FS_TIMEOUT: locking timeout for concurrent access
- _FS_LOCK: prevents operations which are not allowed on open files
Also sets _MAX_SS based on sector size configured for wear_levelling.
This reduces memory usage of FATFS if wear_levelling is using 512 byte
sectors.
Moves the ets_timer_arm() / ets_timer_disarm() code paths to RAM
Overhead is 740 bytes of IRAM, 0 bytes DRAM
(For comparison: If all of esp_timer.c is moved to RAM, overhead is 1068 bytes IRAM and 480 bytes DRAM.)
Allows fcntl() implementation in vfs to be used.
Closes https://github.com/espressif/esp-idf/issues/1070
Note this is different to the other libc file-related syscalls, as there is no
reent structure involved.
Some frameworks based on ESP-IDF need to be able to decide whether to
initialize SPI RAM after the application has started. This change splits
out part of esp_spiram_init which manipulate cache MMU into a separate
function. Applications can disable cache, call esp_spiram_init_cache,
re-enable cache, and then call esp_spiram_init.
Disabling and re-enabling the cache can be achieved using functions
provided in esp_spi_flash.h.
This commit fixes bugs with the dual core changes to freeRTOS tick and idle hooks.
Interrupt watchdog now registers tick hooks to both cores.
API for cross core tick and idle hook registration were also added
Do some refactor to socket ping:
1. Add ping_deinit to kill ping thread
2. Apply configuration set by esp_ping_set_target to socket ping
3. Limit only one ping thread
Previously the timeout was set to the same value (1000ms) for all kinds
of commands. In some cases, such as with slow cards, write commands
failed to complete in time.
This change makes command timeouts configurable via sdmmc_host_t
structure, and also makes default timeouts different for ordinary
commands and write commands.
Closes https://github.com/espressif/esp-idf/issues/1093
Ref TW15774.
Previously if multiple tasks had been added to xPendingReadyList for the CPU, only the first one was resumed.
Includes a test case for resuming multiple (pending) tasks on xTaskResumeAll().
Document the limitation that while scheduler is suspended on one CPU, it can't wake tasks on either CPU.
The way these hooks are implemented at the moment, once you register a tick or idle task hook, it is run by both cores, leading to intermittent crashes that are difficult to explain (and debug). One solution is to register the hook with the core that is currently running, which is what this patch attempts to do.
Merges https://github.com/espressif/esp-idf/pull/781
using apll_param to setup APLL
new apll calculation method, much faster
validate freq calculation
Ensure that the i2s frequency is greater than the hardware limit
Add description of how to calculate apll clock, support apll for other 16-bits audio, check rev0 chip
correct space
* Only direct invocation of cypto_hash_sha256 & crypto_hash_sha512 produced correct output.
* Some remaining header file mixups, so changed approach to wrapping the state structures.
* Fixes https://github.com/espressif/esp-idf/issues/1044 (crypto_sign problems)
* Add option to disable mbedTLS implementation for SHA256 & SHA512 in libsodium
All options that were enabled via CONFIG_MDNS are now in menuconfig, with
the default values set the same as with CONFIG_MDNS enabled (meaning existing
projects that were using CONFIG_MDNS do not need to change).
* setsockopt(s, IPV6_ONLY, &one, sizeof(int)) will disable IPV6-only
mode. Incoming/outgoing IPV4 packets are dropped.
* Otherwise, sockets bound to IPV6_ANY_ADDR can receive unicast packets
for IPV4 or IPV6.
* sendto() a IPV6-mapped-IPV4 address on a UDP socket works correctly
(not supported for RAW or TCP sockets.)
* getaddrinfo() option AI_V4MAPPED is implemented.
As well as extending support to TCP & RAW, there is some potential improvement
to dropping incoming packets - the drop happens a bit late in the process and
there is no "ICMP port unreachable" response sent.
Legacy API of task watchdog used the same function esp_task_wdt_feed() to add
and feed a task. This caused issues of implicitly adding a task to the wdt list
if the function was used in shared code.
The new API introduces init, adding, feeding, deleting, deinit functions. Tasks
must now be explicitly added to the task watchdog using their handles. Deletion
must also be explicit using task handles. This resolves the issue of implicit
task additions to the task watchdog due to shared code calling
esp_task_wdt_feed().
Task watchdog is now fully configurable at runtime by calling the init and
deinit functions.
Also added functions to get the handles of idle tasks of the other core. This
helps when adding idle tasks to the watchdog at run time.
Configuring the task watchdog using menu config is still available, however
menu config will only result in calling the init and add functions for idle
tasks shortly after the scheduler starts.
Menu config also allows for using legacy behavior, however the legacy behavior
willcall the new API functions but with slight variations to make them legacy
compatible.
Documentation and example have also been updated
gcov_rtio.c headers updated to prevent error of freertos header files being
included in the wrong order.
Resolves issue TW#13265
The two task watchdog timer bugs are as follows...
1) If only a single task existed on the wdt task list, and esp_task_wdt_feed()
was only called once, the watchdog triggers but fails to print task name
2) If a single task already exists on the task wdt list, and another task calls
esp_task_wdt_feed() once, the watchdog fails to trigger
Problem stemmed from the loop responsible for resetting the watchdog timer
having incorrect loop parameters. The loop failed to traverse the full length
of the task wdt list
In release mode, switches in rtc_clk_bbpll_set would be converted to
jump tables. These tables would be placed into .rodata, which caused
rtc_clk_cpu_freq_set to be unusable while cache is disabled. This
manifested itself in crashes when exiting from light sleep:
https://esp32.com/posting.php?mode=reply&f=13&t=3089#pr14590
get IP address after connecting to the softap.
Now the default netif is station in station+softap mode. The UDP pcb of
DHCP server is binded to IP 0.0.0.0. When other device connects to softap, the
DHCP offer will be sent by softap.But because the destination IP of DHCP offer
is broadcast IP and the source IP of it is 0.0.0.0, it will be sent from default
netif, that is station interface. Other device can not receive the DHCP offer
sent from station interface. As a result, other device can not get IP address.
The resolution is that bind UDP pcb to the IP address of softap. So, the source
IP of DHCP offer is that of softap. DHCP offer can be sent from softap interface
by source IP route method.
Added documentation about the ESP-IDF changes to FreeRTOS.
The documentation covers changes to the following FreeRTOS aspects.
- Task Creation
- Affects on scheduling (Task skipping, scheduler suspension, tick synchronicity)
- Critical sections and disabling interrupts
- Thread Local Storage Pointers and deletion callbacks
- Configuring ESP-IDF FreeRTOS
Previously, nvs_flash_init_custom would not be called if Storage for a
particular partition was already created. This caused issues if the
first call to nvs_flash_init failed due to Storage init error. This issue
exhibited itself as random failures of NVS CI test.
With this change, storage object is deleted (and not added to storage
list) if initialization fails.
squash the commit.
component/bt:Added the update_white_list & set_prefer_conn_params API to the bt project.
component/bt: Added the get white list size API & delete the read adv tx power API.
component/bt: Change the esp_gap_ble_api.h esp_ble_gap_set_prefer_conn_params API docs.
changed btm_find_dev to btm_find_or_alloc_dev
component/bt: Change the rssi API deep copy method.
component/bt: Change the code as comment in the gitlab.
squash the commit.
component/bt:Added the update_white_list & set_prefer_conn_params API to the bt project.
component/bt: Added the get white list size API & delete the read adv tx power API.
component/bt: Change the esp_gap_ble_api.h esp_ble_gap_set_prefer_conn_params API docs.
changed btm_find_dev to btm_find_or_alloc_dev
component/bt: Change the rssi API deep copy method.
component/bt: Change the code as comment in the gitlab.
component/bt: change the ESP_BLE_CONN_PARAM_UNDEF redefinition error.
component/bt: Change the ESP_BLE_IS_VALID_PARAM function with the compile error.
New multi_heap code has proven effective at aborting when buffer overruns occur,
but it's currently hard to debug the stack traces from these failures.
