1. Modify esp_bt_controller_mem_release() to release BTDM BSS and Data to heap if
ESP_BT_MODE_BTDM mode is passed to it
2. Add a new API esp_bt_mem_release() which internally calls
esp_bt_controller_mem_release() with the provided mode and then if mode
is ESP_BT_MODE_BTDM, releases BT BSS and Data to heap.
Background:
For Wi-Fi and BT/BLE applications, for e.g. the usecase is like when
Bluetooth is used for provisioning and once the device is connected to the Wi-Fi
AP, we can turn off Bluetooth completely. In such scenarios, it should be possible to
reclaim all the memory of Bluetooth. Although, currently this does not
happen.
Experiment:
Made the following modifications to examples/bluetooth/gatt_server :
1. Added support of simple_wifi to it
2. Moved all the bluetooth related code under CONFIG_BT_ENABLED config
option
3. Calculated the free heap in 2 similar scenarios:
i. Disabled BT (CONFIG_BT_ENABLED undefined) and checked the free
heap after STA connected
ii. Kept BT enabled and disabled it after STA connected and checked
the free heap
Ideally, the numbers for i., ii. above should have been similar. But
there was a delta of almost 30-31K. (i. > ii.)
4. Through make size-components checked the common BSS for libbta.a and libbtdm_app.a
and found it to be almost 30K. Data is around 1K
Solution:
1. Modified the linker script to mark the BSS and Data for these libraries and
free it when ESP_BT_MODE_BTDM mode is passed to mem release APIs.
2. Verified that the free heap is comparable for i. and ii. above.
Note: It is known that once this is done, Bluetooth can only be used
again post reboot.
Signed-off-by: Hrishikesh Dhayagude <hrishi@espressif.com>
Added .rtc_noinit and .noinit section definitions into linker file /esp32/ld/esp32.common.ld.
The macro __NOINIT_ATTR, RTC_NOINIT_ATTR declared in esp32/esp_attr.h file.
Added unit test file to test added behavior for noinit variables and its attributes.
Added documentation changes for new added attributes.
Make some corrections after code review:
The linker file has been corrected to place noinit section before bss_start to make it safer.
Documentation file has been modified to clarify reset behavior of allocated data .
Corrected typos in test_noinit.c and removed assertion of noinit variable to avoid possible issues with ROM boot loader memory allocation.
The linker file has been corrected to place noinit section before bss_start to make it safer.
Documentation file has been modified to clarify reset behavior of allocated data .
Corrected typos in test_noinit.c and removed assertion of noinit variable to avoid possible issues with ROM boot loader memory allocation.
Update test_noinit.c file to address RTCWDT_RTC_RESET reset reason instead of POWERON_RESET.
Test optimized to pass automated unit testing.
esp32: Add .noinit and .rtc_noinit sections to the linker script
Update of general-notes.rst documentation to fomat examples as code and attributes as identifiers.
Test test_noinit.c corrected to pass automated testing (support ofTEST_CASE_MULTIPLE_STAGES())
https://ezredmine.espressif.cn:8765/issues/15878
1. Put some lwip udp rx/tx relating functions to IRAM
2. Put some wifi rx/tx relating functions to IRAMa
3. Reduce wifi dynamic malloc from 4 to 1 for each ebuf
4. Update iperf example accordingly
5. Update libphy.a to v383
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
Specifying -fexceptions for the compiler is not enough.
- add necessary zero padding after .eh_frame section
- link .gcc_except_table_table in a way flash script does not complain
- call __registrer_frame_info before global constructors
Kudos jcmvbkbc for the necessary help.
Implements support for system level traces compatible with SEGGER
SystemView tool on top of ESP32 application tracing module.
That kind of traces can help to analyse program's behaviour.
SystemView can show timeline of tasks/ISRs execution, context switches,
statistics related to the CPUs' load distribution etc.
Also this commit adds useful feature to ESP32 application tracing module:
- Trace data buffering is implemented to handle temporary peaks of events load
There are some RODATAs of libphy.a that are called in ISR. So need
to put them into DRAM to avoid access them when R/W SPI flash. Due
to the RODATAs which are called in ISR haven't been picked out to
put into DRAM, put all of the RODATA of libphy.a into DRAM. This
will be optimized in the future.
- Implements application tracing module which allows to send arbitrary
data to host over JTAG. This feature is useful for analyzing
program modules behavior, dumping run-time application data etc.
- Implements printf-like logging functions on top of apptrace module.
This feature is a kind of semihosted printf functionality with lower
overhead and impact on system behaviour as compared to standard printf.
Introduce soc component, add source of rtc_clk and rtc_pm libraries
This MR adds parts of the RTC library source code (initialization, clock selection functions, sleep functions). WiFi-related power management functions are kept inside the precompiled library. Most of RTC library APIs have been renamed.
Default CPU frequency option in Kconfig is set to 160MHz, pending qualification of 240MHz mode at high temperatures.
Register header files are moved into the new soc component, which will contain chip-specific header files and low-level non-RTOS-aware APIs (such as rtc_ APIs). Some of the files from ESP32 component were also moved: cpu_util.c, brownout.c, and the corresponding header files. Further refactoring of ESP32 component into more meaningful layers (chip-specific low level functions; chip-specific RTOS aware functions; framework-specific RTOS-related functions) will be done in future MRs.
See merge request !633
1) fixed SPI_read_status: added check for flash busy flag in matrix mode
2) fixed SPI_page_program: enable write before writing data to SPI FIFO
3) SPI flash ROM funcs replacement is controlled via menuconfig option
Reset WIFI mac when wifi start and update wifi lib
1. reset wifi mac when wifi start
2. roll back rx hung workaround for beacon timeout
3. fix amsdu ap interface wrong issue
4. fix amsdu header parse error
5. fix amsdu flag wrong issue
6. PHY: V350, fix BT rssi bug
7. RTC: V225, fix bt will be not work when wifi is reset
See merge request !606
SPI flash operations lock for shorter periods
Based on bug report here:
https://esp32.com/viewtopic.php?f=13&t=1489&p=6816#p6773
Long SPI flash erase (and possibly write/read) cycles block all tasks on both CPUs for an extended period, and can lead to WiFi dissassociation (and general priority inversion.)
This MR inserts preemption breaks into large operations (all erases, writes every 8KB, reads every 16KB).
Overhead of a single spi_flash_guart_start()/spi_flash_guard_end() cycle measured at approx 67us (assuming no preemption.)
See merge request !600
1. reset wifi mac when wifi start
2. roll back rx hung workaround for beacon timeout
3. fix amsdu ap interface wrong issue
4. fix amsdu header parse error
5. fix amsdu flag wrong issue
6. PHY: V350, fix BT rssi bug
7. RTC: V225, fix bt will be not work when wifi is reset
This is to avoid confusion when idf_monitor prints the first symbol in each section, ie "WindowOverflow4" or similar,
when bootloader prints the section mapping address.
Closes#447https://github.com/espressif/esp-idf/issues/447
1. Add disable PHY and RF when WiFi and BT are both disabled(including call sniffer disable API).
2. Do not init PHY and RF when cpu start. Init PHY and RF when call Wifi or BT start APIs(including sniffer enable API).
3. Add a temporary lib: librtc_clk.a and will delete it when CPU frequency switching function is done.
4. Add an function to get OS tick rate.
5. Do not put the whole pp.a in iram0, only put lmac.o, ieee80211_misc.o, ets_time.o and wdev.o in iram0.
This change reduces chances that a large .bss segment will push .data all the way into
0x3ffe1320 — 0x3ffe5320 range where the bootloader stack is, creating a problem when
bootloader will be loading application into memory.
With this change, .data would need to be at least 200k big to cause problems.
Deep sleep: Any source named rtc_wake_stub* is linked as RTC wake stub code
Also move esp_deepsleep.h documentation out to docs/deep-sleep-stub.rst
See merge request !142
App can contain a stub program resident in RTC fast memory. Bootloader
will load the stub on initial boot. If the device wakes from deep sleep,
the stub is run immediately (before any other data is loaded, etc.)
To implement a custom wake stub, implement a function in your program:
```
void RTC_IRAM_ATTR esp_wake_deep_sleep(void)
{
esp_default_wake_deep_sleep();
// other wake logic
}
```
... and it will replace the default implementation.
