OHHH YEAH!!!! #368
Replies: 10 comments 27 replies
-
|
I can't wait to try it ;-) |
Beta Was this translation helpful? Give feedback.
-
|
I am going to copy the files over to my clone of the repo on a new branch and commit/push the changes. Give me like 20 minutes or so to do that.... |
Beta Was this translation helpful? Give feedback.
-
|
Take your Time, as you know i’ll need one hour to compile it |
Beta Was this translation helpful? Give feedback.
-
|
Well I pushed the commit. I more than likely broke the build tho because I changed a heap load of other things before I did the push and I didn't test the changes. |
Beta Was this translation helpful? Give feedback.
-
|
81 files changed/added/removed |
Beta Was this translation helpful? Give feedback.
-
|
uhhh thanks !! |
Beta Was this translation helpful? Give feedback.
-
|
Nothing, no error, no raise, but callbacks function are just not called (on my p4). this will cause a problem for the task_hander calls. |
Beta Was this translation helpful? Give feedback.
-
|
I'm pressed for time for a project, I found a temporary solution with asyncio, the result is pretty cool: |
Beta Was this translation helpful? Give feedback.
-
|
I am going to look into it now. I do not recommend using this binding for any kind of a project where it needs to run without any bugs. It is still very much an alpha test version and it's being changed a lot. |
Beta Was this translation helpful? Give feedback.
-
|
I just updated the code for the machine.Timer. I know it compiles bit I have not tested the code. Give it a try and see if it works... I did alter one thing in how the timer behaves. Previously MicroPython would scheduled the callback to get called from the main task. I thought this is a bad idea if a user is wanting a more precise timer because who known how long it can be before there is either a time.sleep that gets called or the code in the main task goes to a REPL prompt. There are other places in the code where the main task has it's scheduled tasks run but they are not all that often. Depending on what the user has for code that is running the timer callback would end up being called whenever their was time for MicroPython to do it. Now it is going to get called as soon as the timer ISR occurs. There is no task that gets scheduled. So be aware that in timer callbacks no memory allocation can occur and you also want to keep the amount of code that is running in the ISR context as small as possible. |
Beta Was this translation helpful? Give feedback.
-
Sure, this is proof of concept ;-) |
Beta Was this translation helpful? Give feedback.
-
|
This is nice ! |
Beta Was this translation helpful? Give feedback.
-
|
This is really good news. |
Beta Was this translation helpful? Give feedback.
-
|
I am going to wait a while before updating to the newest version of LVGL. Only because they have added a bunch of new things and when that gets done bugs get introduced. so I am expecting a patch version within a month or so to be released. when that happens I will update to the patch version. |
Beta Was this translation helpful? Give feedback.
-
|
How can I compile the firmware for the ESP32-P4? I have an official ESP32-P4 development board from Espressif, called the ESP32-P4-Function-EV-Board. This board is equipped with an ESP32-C6-MINI-1 module for Wi-Fi and Bluetooth communication, a 7-inch capacitive touchscreen with a resolution of 1024 x 600, a 2-megapixel MIPI CSI camera, 16 MB of onboard flash, 32 MB of PSRAM integrated within the ESP32-P4 chip, and 4 MB of flash built into the ESP32-C6-MINI-1. What commands should I use to compile it? |
Beta Was this translation helpful? Give feedback.
-
The current version of micropython is not officially compatible with P4. All this is very recent. We're experimenting with a partially functional version. I've been pushing patches on the micropython git to get things moving, but they haven't been validated yet. Normally you can compile micropython for P4 with my patch : https://github.com/favnec5/micropython/tree/favnec5-esp32p4-1 with LVGL, it is much more complicated, and what's more with the right driver for the screen (ILI9881c) |
Beta Was this translation helpful? Give feedback.
-
|
Thank you for your reply. I haven't learned how to configure the development board yet. I think I'll just wait for Kevin. |
Beta Was this translation helpful? Give feedback.
-
|
Hello Kevin, After I updated the repository using the command When I run the previous example, the development board restarts and displays the following error: Could you please help me figure out what the issue might be? Thank you. |
Beta Was this translation helpful? Give feedback.
-
|
git pull is not going to update the submodules. You need to use I also suggest you stop using thonny as it doesn't play nice with this binding. I have been trying to get it to work properly but with zero luck so far. |
Beta Was this translation helpful? Give feedback.
-
|
It sure would have been nice if there weren't so many dedicated pins. Things like the DSI and CSI. Those 2 things alone have 13 pins that can only be used to connect to a display and a camera. There are also another 2 pins dedicated to high speed USB. |
Beta Was this translation helpful? Give feedback.
-
|
Here are the pin definitions for that development board. I got that done at least. C6_EN = 54 # RMII_RXER
PA_CTRL = 53 # RMII_RXD1
ETH_MDIO = 52 # RMII_RXD0
ETH_RSTN = 51 # RMII_RXDV
ETH_CLK = 50 # RMII_CLK
ETH_TXEN = 49 # RMII_TXEN
CNN_GPIO48 = 48 # RMII_RXER
CNN_GPIO47 = 47 # RMII_RXD1
CNN_GPIO46 = 46 # RMII_RXD0
SD_PWRn = 45 # RMII_RXDV
SD_CMD = 44 # RMII_CLK
SD_CLK = 43 # RMII_TXER
SD_D3 = 42 # RMII_TXD1
SD_D2 = 41 # RMII_TXD0
SD_D1 = 40 # RMII_TXEN
SD_D0 = 39 # REF50M
# pins used to control boot mode
CNN_GPIO38 = 38 # U0RXD
CNN_GPIO37 = 37 # U0TXD
CNN_GPIO36 = 36 # RMII_TXER
ETH_TXD1 = 35 # RMII_TXD1
ETH_TXD0 = 34 # RMII_TXD0
CNN_GPIO33 = 33 # RMII_TXEN
CNN_GPIO32 = 32 # RMII_CLK
ETH_MDC = 31 # RMII_RXER
ETH_RXD1 = 30 # RMII_RXD1
ETH__RXD0 = 29 # RMII_RXD0
ETH_RXDV = 28 # RMII_RXDV
# Full Speed USB
USBLS_DATA_POS = 27 # USB1P1_P1
USBLS_DATA_NEG = 26 # USB1P1_N1
# JTAG
JTAG_DATA_POS = 25 # USB1P1_P0
JTAG_DATA_NEG = 24 # USB1P1_N0
# Addl GPIO's
CNN_GPIO23 = 23
CNN_GPIO22 = 22
CNN_GPIO21 = 21
CNN_GPIO20 = 20
# ESP32-C3
C3_CMD = 19
C3_CLK = 18
C3_D3 = 17
C3_D2 = 16
C3_D1 = 15
C3_D0 = 14
# I2S
I2S_MCLK = 13
I2S_SCLK = 12
I2S_ASDOUT = 11
I2S_LRCK = 10
I2S_DSDIN = 9
# I2C
ESP_I2C_SCL = 8
ESP_I2C_SDA = 7
C6_WAKEUP = 6
# ADDL GPIO's'
CNN_GPIO5 = 5
CNN_GPIO4 = 4
CNN_GPIO3 = 3
CNN_GPIO2 = 2
CNN_GPIO1 = 1
CNN_GPIO0 = 0
# GPIO Pin Header
HEADER3 = ESP_I2C_SDA
HEADER5 = ESP_I2C_SCL
HEADER7 = CNN_GPIO23
# HEADER9 = GND
HEADER11 = CNN_GPIO21
HAEDER13 = CNN_GPIO20
HEADER15 = C6_WAKEUP
# HEADER17 = +3v3
HEADER19 = CNN_GPIO3
HEADER21 = CNN_GPIO2
HEADER23 = CNN_GPIO0
# HEADER25 = - GND
# HEADER27 = N/C
HEADER29 = CNN_GPIO33
HEADER31 = USBLS_DATA_NEG
HEADER33 = CNN_GPIO48
HEADER35 = PA_CTRL
HEADER37 = CNN_GPIO47
# HEADER39 = GND
# HEADER2 = +5V
# HEADER4 = +5V
# HEADER6 = GND
HEADER8 = CNN_GPIO37
HEADER10 = CNN_GPIO38
HEADER12 = CNN_GPIO22
# HEADER14 = GND
HEADER16 = CNN_GPIO5
HEADER18 = CNN_GPIO4
# HEADER20 = GND
HEADER22 = CNN_GPIO1
HEADER24 = CNN_GPIO36
HEADER26 = CNN_GPIO32
# HEADER28 = N/C
# HEADER30 = GND
HEADER32 = C6_EN
# HEADER34 = GND
HEADER36 = CNN_GPIO46
HEADER38 = USBLS_DATA_POS
HEADER40 = SD_PWRn |
Beta Was this translation helpful? Give feedback.
-
|
Thank you for your reply. I'll try cloning and compiling again. |
Beta Was this translation helpful? Give feedback.
-
I have updated the repository and submodules, recompiled, and obtained the firmware for LVGL 9.2.2 + MicroPython 1.25. When I tested it with the previous example, the development board kept restarting, regardless of whether I used Thonny or not. Below is one of the examples: import lvgl as lv
import screen
scrn = lv.screen_active() # Get the currently active screen object
scrn.set_style_bg_color(lv.color_hex(0x000000), 0) # Set the screen background color to black
def lv_example_menu_3():
# Create a menu object
menu = lv.menu(scrn)
# Set the menu size to the screen resolution
menu.set_size(320, 240)
menu.center() # Center the display
# Modify the back button text (using the correct function name)
back_btn = menu.get_main_header_back_button() # Corrected function name
back_button_label = lv.label(back_btn)
back_button_label.set_text("Back") # Display "Back" on the back button
# Create three subpages
# First subpage
sub_1_page = lv.menu_page(menu, "")
cont = lv.menu_cont(sub_1_page)
label = lv.label(cont)
label.set_text("Hello, PY Learning Notes") # Content of subpage 1
# Second subpage
sub_2_page = lv.menu_page(menu, "")
cont = lv.menu_cont(sub_2_page)
label = lv.label(cont)
label.set_text("Hello, Little Car") # Content of subpage 2
# Third subpage
sub_3_page = lv.menu_page(menu, "")
cont = lv.menu_cont(sub_3_page)
label = lv.label(cont)
label.set_text("LVGL-MicroPython Menu Example") # Content of subpage 3
# Create the main page (without a title)
main_page = lv.menu_page(menu, None)
# Add three clickable menu items to the main page
# First menu item
cont = lv.menu_cont(main_page)
label = lv.label(cont)
label.set_text("PY Learning Notes") # Text of menu item 1
menu.set_load_page_event(cont, sub_1_page) # Click to jump to subpage 1
# Second menu item
cont = lv.menu_cont(main_page)
label = lv.label(cont)
label.set_text("LVGL Example") # Text of menu item 2
menu.set_load_page_event(cont, sub_2_page) # Click to jump to subpage 2
# Third menu item
cont = lv.menu_cont(main_page)
label = lv.label(cont)
label.set_text("About") # Text of menu item 3
menu.set_load_page_event(cont, sub_3_page) # Click to jump to subpage 3
# Set the main page as the default display
menu.set_page(main_page)
# Call the function to run the example
lv_example_menu_3() |
Beta Was this translation helpful? Give feedback.
-
|
A clean compile on OSX!! At runtime there looks to be a missing parameter in the data_bus.init(). Which can be fixed (I think) with: This allows the "import test" to run, but with no display occurring that's likely work in progress still. |
Beta Was this translation helpful? Give feedback.
-
|
and what are you compiling? are you compiling the update_rotation branch or are you using the main branch. The update_rotation branch has the ESP32-P4 board added to it and changes are made to that branch to allow the P4 board to work. The main branch doesn't have the P4 updates. The update_rotation branch also has a large portion of the display bus drivers that have been rewritten. I am thinking you are using the update_rotation branch and I forgot to update the SDL driver to pass |
Beta Was this translation helpful? Give feedback.
-
|
Apologies, yes, I am compiling the branch "update_rotation" for SDL on Mac/OSX. To me, it looks like the buffer size was missing, as the code seems to have False already. I also have an Elecrow 7" ESP32 panel I can test with when things are ready on the branch. |
Beta Was this translation helpful? Give feedback.
-
|
This is the call that is being made after I updated the code. and these are the parameters with the SDL driver the buffer size is always going to be That being said. There are some extra methods added to the bus driver to handle pumping for window events and this MUST be done using the main thread. so there is no way around those additional methods unfortunately. I have updated the code to fix that issue with SDL. I am sure there is going to be other issues with that branch I am sure. I don't have a DSI display to test on my P4 yet. I am going to put an order in for one now. This will allow me to hammer out the issues with the rotation and freertos portions of the new bus drivers. I am still working out in my head how to handle the single core ESP32's. I am hoping that everything will iron itself out because of how MicroPython has been written to use Core 1 and the ESP-IDF uses core 0 for all IO related tasks. The tast that is set up is set to use core 0 so when a single core ESP32 is used the task is not being created on a core that doesn't exist. I think the only thing I may have to add is to yield in the tx_color and tx_params bus methods and that should only need to be done if the ESP32 is a single core. So it should be pretty simple to handle. I will work on that once I get it running for the S3 and P4. |
Beta Was this translation helpful? Give feedback.
-
|
Thanks for update (and answer). Also thanks for the awesome work being done wrt micropython + lvgl. I (and we community) really appreciate it. |
Beta Was this translation helpful? Give feedback.
-
|
When this branch is ready, I'm willing to test SDL on Mac, SDL on Windows Linux Subsystem and on a Crowpanel 7" with an ESP32 |
Beta Was this translation helpful? Give feedback.
-
|
Thank you for all your time and hard work. I got little experience in compiling projects and I was never able to compile lvgl + mpy before. With your binding it finally worked. |
Beta Was this translation helpful? Give feedback.
-
|
glad to hear that it compiled for ya. I did my best to make it as painless as possible. The only thing the user needs to do is to collect the OS level requirements like Python3 and build essentials, the build script collects anything else that is needed. The script also allows a user to inject a custom board without needing to modify anything in MicroPython to do so. There are also some configuration parameters that can be passed in the build command to tweak the build for a specific board. Things like adjusting the flash size and type. |
Beta Was this translation helpful? Give feedback.
-
|
Thanks. I think you made the whole compilation progress as easy as possible and it will help many people. |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
So I have spent some time working on getting things ironed out with the new update to ESP-IDF 5.4 and MicroPython 1.25. One of the things I have been working on is completing the LED bus driver. It compiles now but I have to test it. I also added in support for the ESP32-P4. I want to thank @favnec5 for his work with getting things started with the P4. I did manage to get the PWM sorted out and also the ADC. There are only a few things that don't work yet like the touch sensor which honestly is not really al that important when there is a touch screen display attached. I already have a DSI bus driver written and it compiles. I do not have a DSI display that I am able to test with, I do have a P4 which is a start.
I have completely reworked all of the bus drivers. They all now run on the second core of an ESP32 if there is another core available. I am basically pushing all of the IO over to that second core to lighten the load on the first core so there is better performance for the GUI rendering. By doing this it also allowed be to provide software rotation to all but one bus driver. The LED bus has a fixed rotation of 0. I also added dithering for RGB565 and RGB888 as well. the RGB565 byte swap, dithering and rotation are handled in a single iteration over the buffer data.All or Any combination of them is handled that way... Dithering is able to be turned on and off while the UI is running.This is useful if you know you will be rendering a gradient or displaying an rgb565 image. When those kinds of things are being displayed you can turn it on and when they are no longer being displayed you can turn it off.
Because DSI display function much like an RGB display does in terms of frame buffer handling there is going to be a 4 buffer setup just like what is seen with the RGB bus. 2 full frame buffers behind the scenes and 2 partial buffers the user creates and passes off to LVGL.
I have done some work to tighten up the code to reduce both flash and ram consumption.. I have not looked to see what the results of that are...
I will be pushing the new code to a branch for testing and bug fixing. once it checks out I will merge it with the master branch. I have done pretty much everything I know how to do to squeeze every drop of performance out of the ESP32's when running MicroPython without writing all of the display drivers in C code. Doing that would make it a compile time choice for the display and not a runtime one. I like the current design of being able to change out displays simply by plugging in a new display and uploading a new python script and it's off a running, no compiling the firmware is needed.
I am super happy about adding support for the P4. This is going to be a game changing MCU for running displays. 400mhz dual core with a 2 lane DSI running at 1.5 gHz per lane. There is also a 2 lane CSI (Camera Serial Interface) and I believe there is a way to leverage that to get a 4 lane DSI. I was only able to get my hands on a preliminary release of the data sheet for the P4 so things may have changed.
Beta Was this translation helpful? Give feedback.
All reactions