.. and switch makerfabs tft7 over to it as a test.
We have our existing way of "reserving" PSRAM for esp-idf (we actually
control it all but add back the "reserved" part). However, this does
not work with off the shelf esp_lcd, which only will allocate a
framebuffer in PSRAM if CONFIG_SPIRAM_USE_CAPS_ALLOC (or CONFIG_SPIRAM_USE_ALLOC)
is defined, not if CONFIG_SPIRAM_USE_MEMMAP is.
This new way is possibly compatible with more esp-idf code, but it complicates
CircuitPython's initial startup since nothing until port_heap_init is
permitted to use the CP heap or supervisor allocator. In practice this
seems to be OK today.
Right now this doesn't change the setting across all boards with PSRAM and so
it does not revert esp-idf to its prior state. Instead, what I'm thinking is
that we can do it during or just after the IDF5 update when sdkconfig files
will be getting an overhaul anyway.
This allows the web workflow send code to yield briefly when
waiting for more room to send in a socket. Waiting for an "interrupt"
could wait forever because the select task only waits for read and
error. Adding wait on write is tricky because much of the time we
don't care if the sockets are ready to write. Using yield avoids
this trickiness.
This uses the esp32-camera code instead of our own homebrewed camera code.
In theory it supports esp32, esp32-s2 and esp32-s3, as long as they have
PSRAM.
This is very basic and doesn't support changing any camera parameters,
including switching resolution or pixelformat.
This is tested on the Kaluga (ESP32-S2) and ESP32-S3-Eye boards.
First, reserve some PSRAM by putting this line in `CIRCUITPY/_env`:
```
CIRCUITPY_RESERVED_PSRAM=524288
```
and hard-reset the board for it to take effect.
Now, the following script will take a very low-resolution jpeg file and print
it in the REPL in escape coded form:
```python
import board
import esp32_camera
c = esp32_camera.Camera(
data_pins=board.CAMERA_DATA,
external_clock_pin=board.CAMERA_XCLK,
pixel_clock_pin=board.CAMERA_PCLK,
vsync_pin=board.CAMERA_VSYNC,
href_pin=board.CAMERA_HREF,
pixel_format=esp32_camera.PixelFormat.JPEG,
i2c=board.I2C(),
external_clock_frequency=20_000_000)
m = c.take()
if m is not None:
print(bytes(m))
```
Then on desktop open a python repl and run something like
```python
>>> with open("my.jpg", "wb") as f: f.write(<BIG PASTE FROM REPL>)
```
and open my.jpg in a viewer.
.. the primary user of which will be the camera, since the framebuffers
must be allocated via esp-idf allocation function and never from the
gc heap.
A board can have a default value, and the value can also be set in the
/.env file using the key CIRCUITPY_RESERVED_PSRAM with the value being
the reserved size in bytes.
Co-authored-by: Dan Halbert <halbert@adafruit.com>
The ringbuf used to fill up and the recv interrupt would block CP.
Now it'll disable the interrupt until we have enough room in the
ringbuf.
Fixes#6678
* Fixes#6221 - C3 hang on `import wifi`. Enabling the WiFi PHY was
disabling USB. Now boards that use it set CONFIG_ESP_PHY_ENABLE_USB
explicitly.
* Fixes#6655 - Allows pasting into the web serial page. Fixes reading
more than 0xf bytes at a time.
* Fixes#6653 - Fixes web socket encoding of payloads >125 bytes. Can
happen when printing a long string.
* Fixes C3 responsiveness when waiting for key to enter REPL. (It
now correctly stops sleeping.)
* Disables title bar updates when in raw REPL. Related to #6548.
* Adds version to title bar.
It had a tight 5ms timeout before that caused some characters to
drop. Now the wait is longer and reset after a successful transmit.
This follows what MicroPython does.
Fixes#6220
rather than setting the heap size statically, micropython allocates
the biggest contiguous chunk possible, but in no event more than half the
total internal sram. On esp32 this gives 123728 bytes of `gc.mem_free`
in the repl.
esp32 places the psram start at SOC_EXTRAM_DATA_LOW and it can extend
up to SOC_EXTRAM_DATA_SIZE. This is different than esp32-s2 and later,
which place the end at EXTRAM_DATA_HIGH and the limitation of
SOC_EXTRAM_DATA_SIZE was not previously identified as important.
Additionally, the esp32 has a reserved area within himem which was
not being accounted for.
With this change, the Feather ESP32 V2 feather can be used via thonny,
and the other "quick memory corruption tests" I was performing
also all succeed instead of failing.
Before this change, the incorrect address being used for spiram was
0x3fa00000..0x3fc00000 (2MiB). Now, it's 0x3f800000..0x3f9c0000 (1.75MiB)
due to the reserved area and the changed start address.
This is intended to be a no-effect change for other espressif chips besides
original esp32.
