* Add flash for vtable destination, make it default
Add an option for placing vtables in flash to complement the existing
iram and heap options. "make flash"
Now that there is a way to change it, move to vtables in flash as default
as only users with interrupts which use vtables require the vtable to
be in RAM. For those users, if the tables are small enough they can put
them in IRAM and save heap space for their app. If not, then the vtables
can be placed in HEAP which supports much larger tables.
* Add VTable menu, FLASH as default, remove Makefile
Convert from manual "make" operated app.ld creation to runtime creation
whose options are selected from the build menu.
Use a prelink recipe to create the output app.ld file each run, without
need for any special tools.
Update the boards.txt.py script to generate this new config.
Enables I2S stereo input via DMA using new API calls:
. i2s_rxtx_begin(bool rx, rool tx);
. i2s_read_sample(uint32_t *l, uint32_t *r);
Original API calls will only enable TX, so this is backwards compatible.
Add simple I2S input example code using Arduino serial plotter.
Add UDP transmit of I2S microphone data to a PC (remote microphone).
Clean up and reorganize code to share RX and TX logic as much as
possible. Fix a potential WDT error while in blocking sample read
and write.
* Fix for #4565 (rx fifo length), protect access to rx_buffer
* Fix typo
* reworked to separate safe from unsafe functions, factorized some code, constness
* additional rework for uart_rx_fifo_available()
* swapped unsafe function definition order
* Remove static for overrun string
* Some shorthand for perf and readability
* I2S driver fixes for IRQs, protocol, factoring
All redundant ICACHE_FLASH_ATTR decorators were removed, we already do
this by default for all routines, anyway,
The actual ISR and its called function moved to to IRAM. Used to be in flash
due to the decorator, which could lead to crashes. Use ets_memset to mute
buffers in ISR.
Fix the I2S on-the-wire protocol by enabling the transmit delay I2STMS because
I2S is supposed to send the MSB one clock after LRCLK toggles. This was
causing I2S to be twice as loud as intended in the best of cases, and causing
garbage/noise output when the MSB was set since data was effectively shifted.
Refactor the clock divider setting to be done in one function only, as there
is no reason to do the same complicated bit setting in two spots.
* Comment some add'l registers, use optimstic_yield
Comment the known and unknown I2S register settings for posterity, using
the ESP32 guide as a basis.
Use optimistic_yield() instead of esp_wdt_disable/enable when busy
waiting in blocking writes to ensure we don't hog the CPU completely.
Move the constant IO pins to #defines for easier understanding.
Generate and preserve a linker .MAP file for the project in the usual
build directory. This map file is useful for finding out where FLASH,
IROM, and IRAM are being used as it shows gcc internally generated
things such as jump tables and constant function-local variables.
With this patch the set up clock rate survives a re-init that is done by many libraries several times.
This makes e.g. an accelerated OLED display possible with the adafruit libraries
* device test for #4516 (WiFiClient leaking)
* simply count the number of connections that do not change heap
* remove unused heapLost
* + python server (to not depend on gateway's opened tcp port)
* remove old&bad define
* fix another old address
* Fix WebServerSecure streamFile()
ESP8266WebServerSecure's streamFile was using the base class' method
which did not use SSL encrypt before transmitting, leading to failure.
Add a new template method and required support for
WiFiClientSecure::write(Stream&) (using a local temp buffer since the
SSL libs do not grok Arduino Streams at all).
Fixes#4544
* Match ClientContext buffer and yield() behavior
ClientContext sends out 256 bytes at a time and gives a yield after
each chunk to ensure the WDT doesn't fire. Mimic that behavior in
WiFiClientSecure::write(Stream&).
* inefficient Print::write(data,len) shows message if used (only in debug mode)
* make HardwareSerial's write(data,len) efficient
* HardwareSerial: remove duplicate tests, move trivial code from .cpp to .h
* tests/device/Makefile: use bourne shell compatible "." instead ot "source" (ubuntu's dash does not understand it)
* tests/device: make BEGINTIMEOUT a new error return value instead of a test-process-breaking exception
* Update readme.rst
fix broken link to file system flash layout
* linkage repair
my second shot at fixing the external page link for "flash layout". It appears that the on-page link to "Update Process - Memory View" was broken (not sure if I got this one right or not).
* Change argument to Esp.deepSleep from uint32 to uint64 to match SDK, add deepSleepMax based on the cali_proc function per SDK
* SPISlave.end() added
* Nameoftherose patch for Issue #2435 (#4256)
* WiFiTelnetToSerial Example - Minor Issues #2435
* WiFiTelnetToSerial Example - Minor Issues #2435
Patch to rectify issue #2435
* Check that pins needed by Wire are defined at compile-time (#4261)
Move all exception strings to IRAM and out of both PMEM (illegal) and add
output of any assert() failinf conditions.
The exception handler may be called while the SPI interface is in a bad
state. This means no PROGMEM reads are allowed, and all data and functions
used must be in system RAM or IRAM.
Add a new helper macro, ets_printf_P(), which places a constant string in
IRAM and copies it to the stack before calling the real ets_printf().
This makes the code simpler to read as no unwieldy combinations of
ets_putc/ets_printf/... are required to output anything.
The old handler also mistakenly used PSTR() strings in some places, so
fix those with this patch as well.
Gives back ~180 bytes of heap to every sketch built as the exception handler
is always included an application.
At runtime, whenever a realloc, malloc, calloc, or new call fails to
find enough memory, record the calling function and size requested.
Does not print anything or call any heavyweight functions on this, as it
is legal to return NULL to an alloc-type call.
If the main application handles this NULL properly, there should be no
panic and nothing different will happen.
If the main application panics() at any later point in time, this record
will be printed out as part of the crash log for processing with an
updated EspExceptionDecoder.java.
This adds 2-3 instructions overhead in the normal case, and around 10-12
instructions in the failing case, and requires an additional 8 bytes of
.BSS to function.
Only a single address is kept, the final failing malloc-type function call
before a panic, but it is trivial to extend to a fifo with little overhead
in the common, non-failing case.
The default Print::write(byte, count) method was continuing to send
bytes one-by-one even when a prior write returned 0. Because the buffer
pointer was incremented no matter success or fail, this leads to data
corruption as you'll not send some bytes in the middle and will then
send extra bytes past the end of the passed in buffer.
Because there's no concept of timeout, just stop on the first time
write(byte) fails and return the total bytes successfully written
and let the user worry about retrying or handling an error.
Found by @d-a-v and discussed on gitter.
Remove stray, older version of sigma_delta.h
An older version of sigma_delta.h was committed to the root directory. Remove it, as we only want the later, new one already present in cores/esp8266. Thanks, @reaper7, for noticing it.
Provides a simple interface to attach/detach pins to the sigma-delta generator, and get/set the 2 parameters prescaler & target. Working example that fades the onboard LED is provided. Code and sample by @stefaandesmet2003.
