…library sources + (electrical) testing)

brought to you by Martin Turski, company owner of EirDev (turningtides@outlook.de)
this adds support for the Makerbase MKS TS35-R V2.0 screen in combination with the MKS Robin E3D V1.1 board. proper testing has shown that this hardware combination works very well with the Marlin firmware. previously the bus negotiations of various Marlin components was conflicting so much that this use-case was a head-ache!

Adds proper baudrate determination between hardware specifications and board clock frequencies.
- new configuration option: TOUCH_BAUDRATE
- new configuration option: TFT_BAUDRATE_READ
- new configuration option: TFT_BAUDRATE_WRITE

Buy MKS Makerbase TS35-R V2.0 screen: https://de.aliexpress.com/item/1005003634231986.html
Buy MKS Makerbase Robin E3D V1.1 board: https://de.aliexpress.com/item/4000781744682.html

https://green-candy.osdn.jp/external/promo/mks_robin_e3d_v1_1_with_ts35r_v2_0.png

Simply connect the screen across the EXP1/EXP2 connectors, flash the properly configured firmware and give it a go!

Interupt-driven SPI DMA has not been tested. Also I have not tested every configuration combination. I would appreciate some verification by the Marlin FW community.

…MC drivers on the same SPI bus as TFT, Touch, SD, etc

brought to you by Martin Turski, company owner of EirDev (turningtides@outlook.de). In this commit the SPI support has been improved in all of Marlin - across ALL the boards! - so that the big problem of connecting peripherals on the same SPI bus is tackled. I have tested the commit on my MKS Makerbase Robin E3D V1.1 board and all SPI components have successfully initialized, are successfully communicating and it works quite nicely. I invite you to test your own board using this commit.

I tested the TMC2130 drivers in SPI mode and there is no issue when using them. Also connected as a MKS Makerbase TS35-R V2.0 screen.

Have fun!

…ompilation for other people aswell

…l lacks them)

… support chaining)

… into bugfix-2.1.x

…ons)

- added more reliability ensurances to the SD card stuff

… (oops)

…ompilation fail

…rd SPI pins (for now?)

…e (the standard SD card read block retry count)

…ique (reference) implementations

This is the big HAL SPI update you have been waiting for! It fixes the SPI bus sharing problem for good, laying a strong foundation for the Marlin FW SPI into the future. All brought to you by Martin Turski, company owner of EirDev.

CHANGELOG:
* added ESP32 unique HAL SPI
* added LPC1768 unique HAL SPI (needs testing because I got a broken board)
* tft_spi.cpp is now a generic class (SPI ccreen support for virtually every SPI capable controller board)
* xpt2046.cpp is not a generic class (generic touch screen support comes closer!)
* further global SPI refactoring, removing use of the generic Arduino SPIClass in favor of the new & better HAL SPI interface
* removed obsolete code (MarlinSPI)
* added HAL NVIC (ARM) for proofed interrupt management (shared between STM32 and LPC1768)
* removed some arbitrary Marlin FW limits related to shared SPI bus (may not have caught all thus I need the support for the Marlin FW community for testing)
* made drawing to the screen asynchronous for color UI (if it is supported by enabling USE_SPI_DMA_TC, only on supported boards!)
* merged ARM exception handling code with the new HAL NVIC code
* added some ARM exception management helpers (BEEP_ON_SYSTEM_EXCEPTION)
* added HALSPI_DO_ERRORBEEPS, HALSPI_DO_LOOPBEEPS, HALSPI_ESP32_STATIC_DMADESCS, HALSPI_DMA_ALWAYS, HALSPI_DEBUG, HALSPI_DISABLE_DMA, HALSPI_LOOPBEEP_TIMEOUT, HALSPI_DMA_THRESHOLD configuration preprocessor definitions for HAL SPI
* added HAL_SPI_SUPPORTS_ASYNC option: if defined then the SPI async interface (spiWriteAsync, spiWriteAsync16, spiAsyncAbort, spiAsyncJoin, spiAsyncIsRunning) is available
* improved SD_WRITE_TIMEOUT, SD_ERASE_TIMEOUT, SD_INIT_TIMEOUT, SD_READ_TIMEOUT SD-card timeout variables: they can now be disabled to allow for no-timeout at SD card operations (use with caution!)
* added SD_DISABLE_TIMEOUTS preprocessor option: if defined in your configuration then the default SD card timeout variables are not applied, if previous config of them is missing

HAL SPI has been thoroughly tested on the STM32 and ESP32 implementations! The generic SPI has been working fine on the LPC1768 (needs further testing, I will get a fully working board soon).

Have fun!

….1.x

… into bugfix-2.1.x

…mpiles)

- removed some compilation warnings
- disabled async SPI support it HALSPI_HW_GENERIC is enabled (the generic implementations are supposed to be very easy)
- SPI fixes for AVR

… into bugfix-2.1.x

…561)

The _ATmega_savePinAlternate function saves the device state which affects the provided GPIO pin and disables it. This way the provided pin can be safely used for GPIO purposes until the _ATmega_restorePinAlternate function is called. Then the device state is reset prior to the call of _ATmega_savePinAlternate, assuming that no other device state change related to the disabled peripherals was undertaken inbetween the function calls.

- added REPRAP_FULL_GRAPHIC_SMART_CONTROLLER to MKS TinyBee V1.0

…lity update

…ibility list (that should cover all of the AVR devices already supported by Marlin)

…ctions mapped to the given digital pin, enabled as soon as the powering MCU component is enabled

…PI implementation into it's own file

- improved the maintainability of dozen AVR board pin header files (actual pin names instead of internal numbers)

Please help me convert the rest of them so that we can have a clean Marlin experience!

- added spiEstablish function which forces kick-up of SPI signals + peripherals that would otherwise be implicitly done by transmission functions (if required)
- added more debug logic to ESP32 (further work pending to get DMA working)

…ssage

…presence of no status updates was unchecked, causing a jump to corrupt memory locations

- various adjustments

- added generic HW implementation for 8bit AVR MCUs based on the ArduinoCore

new preprocessor macro AVR_CHIPOSCILLATOR_FREQ: please look into your board manufacturer schematics to find the XTAL1/XTAL2 pins, then create a pin header board preprocessor definition with the Hz number (typically in MHz) to enable the optimized AVR HAL SPI implementation.

- removed an unnecessary comment that I added in Sd2Card

….1.x

… into bugfix-2.1.x

…ternal numbers

- refactored the sanguino 8bit AVR pin headers to use pin names instead of internal numbers

- refactored the teensy2 8bit AVR board pin header files to use pin names instead of internal numbers

…for code-graph construction

- improved the pin device state saving, adding support for powered-down peripherals, adding the _ATmega_savePinAlternates function to save the combined device state of multiple pins

…sage of a pin as SPI-protocol chip-select (and no other purpose!)

- searched for SPI CS initializations and converted them to use the new spiSetupChipSelect function

…llator frequency definitions

- fixed pin definitions in AVR fastio which were used by external libraries, such as LiquidCrystal
- added chip oscillator frequencies for RAMPS and MKS GEN L 1.0

…rnal numbering

… into bugfix-2.1.x

….1.x

…ino and teensy2 pin header files

- added missing AVR_CHIPOSCILLATOR_FREQ and LPC_MAINOSCILLATOR_FREQ board oscillator definitions where they were missing, according to the board schematics & made a guess whenever the schematics are missing (please help me get the remaining schematics for good future proofed support of Marlin FW!)
- various adjustments

… into bugfix-2.1.x

….1.x

…d during code formatting on Dec 12, 18:12

- added missing board oscillator freq (forgot to commit it?)

….1.x

…or my hardware partner

- grammar fix for SOFTWARE_SPI on LPC1768
- various stability adjustments for the heck of it (it is better to be safe than sorry!)
- bugfix for DYNAMIC_VECTORTABLE exception handling on LPC1768 (vector table entry 7 does not contain default handler)

…e not on the hardware SPI bus then we force their TMC SPI mode to software mode

…uild process

- fixed some issues regarding volatile memory access semantics across AVR, LPC1768, ESP32
- upgraded MKS TinyBee compilation to espressif 6.0.0 + GCC 11.2.0 (very recently released)
 -> code changes to make Marlin FW compatible with the latest toolchain
- overhaul of ESP32 HW HAL SPI layer, making use of Eir SDK BitManager code-graphs to speed up 64 byte / 512 bit SPI buffer fetch & fill (further optimizations to the software model are outstanding to a point past pull request approval)
- tiny adjustments regarding SPI FLASH & MAX31865 SPI temperature controller (next project)

- upgraded the atmelavr PlatformIO package to 4.0.1
- changed the build configuration to compile Marlin FW code with the GNU++1z standard (constexpr-if is useful in the embedded field)
- further "fixes" to the volatile problem

…mmon instead of common_avr8

- set GNU++20 standard for LPC176x

…st transfer through DMA but the MCU peripheral supports MSBFIRST transfer only

 -> solved by falling back to async-SPI-through-interrupts or generic SPI (special case is repeat-by-DMA which still works using DMAC!)
- optimized the _flip_bits function for LPC176x by use of RBIT assembly instruction (ARM)
- reduced the Wno-volatile warning spam

… connected again

…n terms of TFT_IO async operation availability (write_sequence_async method)

…sing the latest toolchain

- improved generic_create_variant.py to avoid wrecking official/generic variant build configurations

…stem

 -> experimentally used for selection of font files for building
- added TFT_FONT_* Marlin feature entries that map to the appropriate source files
 -> also applied TFT_FONT checks inside the font-data source files for robustness
- fixed compilation for the maple STM32F1 environments
- updated the maple STM32F1 configurations to use the latest platform definitions (15.4.1)
- fixed some bugs due to newer GCC compiler (duplicate inline)
- linked the STM32 HAL SPI with the STM32F1 maple build configurations since these are compatible (do not expect me to update deprecated code!)

…inal source file convergence phase on the filter list instead of by each feature, fixing bugs that broke the software model

… variable visible to the acquisition logic

- cleaned up platformio.ini source filters by removing deprecated ones, generalizing according to source layout (it could still be improved but I did my best! Scott should decide further steps)

…STM32F4xx)

- implemented HAL_NVIC for STM32H7xx
- improved building for STM32, removing TFT source filesfrom building stage if they are not needed
- bugfix for 1024x600 resolution (TCT)

Merge branch 'bugfix-2.1.x' of https://github.com/MarlinFirmware/Marlin into bugfix-2.1.x

…ssume NOTOSANS as default, which it is in the sources anyway)

- removed game menus from feature HAS_MARLINUI_MENU (they should be optionally enabled)

…OTOR_CURRENT_* where the gcode feature source files were missing

…future

- improved generic_create_variant.py to not mess with official board variant configurations (only touch if board_build.variant value is prefixed with MARLIN_)

…ard)

…e STM32 framework (it is excluded by default) [chitu_f103 fix]

…32 to fix compilation issues on the latest Espressif toolchains

When the PR luc-github/ESP3DLib#51 has been merged then please change it back!

- (temporarily) updated the "ESP Async WebServer" dependency for MKS TinyBee / ESP32 to fix compilation issues on the latest Espressif toolchains

Please watch PR me-no-dev/ESPAsyncWebServer#1142 and change back once merged.

- fixed another ESP32 bug due to toolchain update

…uilding again (shell passthrough special characters)

…spressif32 (please fix dependency once 2.0.7 of that package is out on PlatformIO registry!)

…tions

Please check PR rhapsodyv/Arduino_Core_STM32#1 and if @rhapsodyv has merged then adjust the dependency to an official one!

…moving Marlin internals out of variants-headers and making them optional (new platdefs STM32 HAL folder for Marlin specific board specializations + _STM32_PLATDEFS define)

@ellensp said that PINS_DEBUGGING support should be rewritten in another PR; I agree but this is a necessity to unbreak compilation + PINS_DEBUGGING is not a release feature so don't obstruct progress, please!

…t is a bad idea to rely on such internal thus adjusted some things.

…x that was previously allowed due to compiler-forgiveness

…nto bugfix-2.1.x

- added YouTube videos about the Arduino pin layout and the AVR alternate pin functions to the comments

- fixed a bug in ESP32 HAL SPI where interrupt code cannot be assumed to be entirely in IRAM due to access to peripheral memory, possibility of having a compiler-internal function inside the generated code-graph