Per-CPU interrupt handlers and args
Up to now, the interrupt handlers and args were shared between CPUs, that is, if you set an interrupt handler on CPU0, CPU1 would invoke the same interrupt handler on that interrupt. This code gives every CPU its own space for interrupt handlers.
See merge request !192
nvs: fix Page::findItem and Storage::findItem regression
When read caching was added, Page::findItem started modifying itemIndex reference argument even if item wasn't found.
Incidentally, Storage::findItem reused itemIndex when starting search at next page.
So,
- if the first page had a cached index (findItem was called for that page), and it pointed to a non-zero index,
- first page has a few empty items at the end (but is marked full),
- next search looked up the item on the second page,
- index of the item on the second page was less than the cached index on the first page,
then the search would fail because cached starting index was reused.
This change fixes both sides of the problem:
- Page::findItem shouldn't modify itemIndex argument if item is not found
- Storage::findItem should not reuse itemIndex between pages
Two tests have been added.
ref. TW8406
See merge request !171
New makefile component_wrapper.mk allows some variables to be set
before component.mk is evaluated. This properly fixes problems with
sdkconfig being hard to access in all phases of the build.
Including component_common.mk is no longer necessary and will print a
deprecation warning for components which use it.
Reduces number of make invocations, allows variables exported in project
to be seen in all component make processes, not just the main ones.
Also makes a no-op build about 3x faster than it was.
Fixes for github issues
This MR contains a bunch of small fixes for issues raised on Github and esp32.com forum:
- vfs doesn't check error code returned by open syscall
- spi_flash doesn't work when built in release mode
- duplicate definition of O_NONBLOCK when combining LwIP socket.h with sys/fcntl.h
- wrong order of creation of standard streams
- `_times_r` returning incorrect values, causing `clock`to return double of the actual time
- driver/gpio.h: comment fix
- wifi event handlers: fix incorrect MAC address printed by logging statements
- move some functions out of IRAM when compiling for bootloader
Please check commit descriptions for links to issues/forum posts and more details.
See merge request !183
newlib: fix compilation error when no timers are enabled in menuconfig
I have checked that three options (RTC/RTC+FRC1/FRC1) work, but didn't check "None" option.
This adds some missing ifdef blocks to remove "unused function/variable" warnings and fixes the use of `__errno_r` in a non-reentrant function.
See merge request !174
LwIP will define O_NONBLOCK in sockets.h if it isn't defined yet.
If sys/fcntl.h is included after socket.h, there will be duplicate definition.
Work around by including sys/fcntl.h into lwipopts.h.
https://github.com/espressif/esp-idf/issues/75
esp32/make: add detailed return error code for wifi APIs
1. Add detailed return error code for wifi APIs
2. Add description about error code in esp_wifi.h
3. Modify esp_wifi_reg_rxcb to esp_wifi_internal_reg_rxcb
4. Modify esp_wifi_set_sta_ip to esp_wifi_internal_set_sta_ip
5. Mark system_init as deprecated API
See merge request !187
1. Add detailed return error code for wifi APIs
2. Add description about error code in esp_wifi.h
3. Modify esp_wifi_reg_rxcb to esp_wifi_internal_reg_rxcb
4. Modify esp_wifi_set_sta_ip to esp_wifi_internal_set_sta_ip
5. Mark system_init as deprecated API
wpa_supplicant: replace pre-built crypto library with source code
This change replaces libcrypto.a pre-built library with source code from wpa_supplicant.
For now i've decided to keep the modified layout of source and header files which is used in the esp32_rtos_sdk_core project. This will make it easier to move other parts of wpa_supplicant from esp32_rtos_sdk_core into esp-idf. Once we move everything, we can rearrange the files and upgrade wpa_supplicant to latest version.
See merge request !177
1. remove "\n" when calling ESP_LOGX APIs.
2. modify uart_event_t for uart rx data.
3. use MICRO for uart inverse value
4. add uart_tx_data_t for internal tx function.
1. Set XXX_TAG static, remove extern XXX_TAG in uart.h/ledc.h/gpio.h
2. I removed uart_set/get_print_port() functions, these functions are not well tested, I removed them for now.
3. Modify some function names for uart_read/write_bytes
4. Modify uart_write_bytes and uart_write_bytes_with_break.
1. fix bug in ringbuffer.c:
When return an dummy item, free_ptr might exceed rd_ptr, so the write_ptr would overwrite rd_ptr in this case.
2. Delete UART tx task in buffer mode. UART ISR will copy the data from tx buffer to fifo.
When read caching was added, Page::findItem started modifying itemIndex reference argument even if item wasn't found.
Incidentally, Storage::findItem reused itemIndex when starting search at next page.
So,
- if the first page had a cached index (findItem was called for that page), and it pointed to a non-zero index,
- first page has a few empty items at the end (but is marked full),
- next search looked up the item on the second page,
- index of the item on the second page was less than the cached index on the first page,
then the search would fail because cached starting index was reused.
This change fixes both sides of the problem:
- Page::findItem shouldn't modify itemIndex argument if item is not found
- Storage::findItem should not reuse itemIndex between pages
Two tests have been added.
implement time syscalls
This change adds optional support for libc time functions.
New menuconfig option allows selecting FRC1, RTC, both, or none to be used for timekeeping.
- If only FRC1 timer is used, gettimeofday will provide time at
microsecond resolution. Time will not be preserved when going
into deep sleep mode.
- If both FRC1 and RTC timers are used, timekeeping will
continue in deep sleep. Time will be reported at 1 microsecond
resolution.
- If only RTC timer is used, timekeeping will continue in
deep sleep, but time will be measured at 6.(6) microsecond
resolution. Also the gettimeofday function itself may take
longer to run.
- If no timers are used, gettimeofday function will return -1 and
set errno to ENOSYS.
`times` function returns time derived from FreeRTOS ticks. It reports all as 'system' time, 'user' time is reported as zero.
`settimeofday` function may be used to set current time.
LwIP SNTP module is hooked up into the system through `settimeofday`/`gettimeofday`.
Example demonstrating this functionality is also added.
ref. TW6415
See merge request !168
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
Fix spi_flash_write regression and nvs error recovery
In the previous set of changes related to spi_flash, new alignment checks were added. One of these checks, word alignment of `src` buffer in `nvs_flas_write`, was unnecessary. ROM `SPIWrite` function can handle unaligned source buffers.
This change caused an error to be returned to nvs for some legitimate write operations. Due to a bug in nvs, further write operations were possible, while the internal state of `Page` instance was broken. In WiFi functional tests, this inflicted havoc on the nvs partition, creating multiple duplicate items in the affected page. Because multiple duplicate items per page were never supposed to be handled during page load, duplicates were not removed. Thankfully this caused an assert at a later point in the loading process, otherwise this bug would be very difficult to detect.
This change set fixes the original spi_flash regression, handling of `INVALID` state of `nvs::Page`, and handling of duplicate items.
See merge request !161
Feature/tw8155 optimize tx flow control
1. Remove TX flow control in LWIP
2. Make the return value of esp_wifi_internal_tx consistent with LWIP error code
See merge request !164
146f5962 - Make the return value of esp_wifi_internal_tx consistent with LWIP error code so that
the up-layer can detect the out-of-memory error and take action accordingly, such do flow control.
1. Add a ring buffer for UART TX.
If the buffer size is set to zero, driver will not use a buffer. But we need a task to send data from buffer to fifo. I tried directly copy data in ISR, but the code looked too long for ISR.
2. Modify the format in uart.h
Due to previous flash write bug it was possible to create multiple duplicate entries in a single page.
Recovery logic detected that case and bailed out with an assert.
This change adds graceful recovery from this condition.
Tests included.
lwip: fix tcp rx abnormal issue(tw8242)
In tcp_alloc(), initialize per_soc_tcp_wnd before initializing recv_wnd because recv_wnd depends on per_soc_tcp_wnd.
See merge request !160
Currently a restart is required to recover a page from invalid state.
The long-term solution is to detect such a condition and recover automatically (without a restart). This will be implemented in a separate change set.
Previously the test bench would check failure recovery by introducing error after each write operation.
This makes checks a bit more extensive (and much longer) by failing after every word written. Surprisingly, this change didn't expose any bugs.
ROM SPIWrite routine can work with unaligned sources, so this check is unnecessary.
Furthermore, it breaks nvs_set_str and nvs_get_blob when data pointer is unaligned.
Also fix stray backslash in COUNTER_STOP macro
nvs: fix memory leaks in HashList and nvs_close
Fixes TW8162.
Associated test case is run under Instruments on macOS, until I set up valgrind to test this automatically on Linux.
See merge request !150
Add cross-core int to accelerate task being awoken from another CPU.
This adds a per-CPU interrupt that can be used to poke the CPU to go do something. In this case all that is implemented is a request to yield the current task, used in case a CPU unblocks a task that runs on another CPU. This gets rid of the limitation that inter-CPU communication using queues, muxes etc can take up to a FreeRTOS tick to happen.
Specs!
Sending an in in a queue of length 1 (essentially a semaphore) as quickly as possible (just a small delay in the sender, to make sure the receiver task gets swapped out) for 10 seconds. Number indicates the amount of ints transferred
Old code:
CPU0->CPU0: 42986
CPU0->CPU1,: 2999
New code:
CPU0->CPU0: 42868
CPU0->CPU1: 62073
See merge request !155
Fix the things that broke when adding the new WDTs
Seemingly, I broke a bunch of things when adding the interrupt WDTs and moved the panic handler to the esp32 directory. This fixes that, as well as the issue where flashing would trigger the int wdt. It also bodges in a fix for a merge artifact breaking the halt-on-first-thread-when-openocd-is-connected; that fix should be refined later.
See merge request !157
lwip: support max 16 sockets
Since the customers need more sockets in their application, support max 16 sockets,
in other words, the total socket number of UDP/TCP/RAW sockets should not exceed 16.
See merge request !156
1. change the description of Makefile.projbuild
2. remove the license header in the API document
3. add private inlcuding header code in the components file
Since the customers need more sockets in their application, support max 16 sockets,
in other words, the total socket number of UDP/TCP/RAW sockets should not exceed 16.
