SPI library not read output after transfer() call

[acrist]

Boards

AMB23/SPI

External Hardware

28J60 ethernet shield or MCP1251 CAN shield

Hardware Configuration

Shield connected to SPI pins 12-15 due to official pinout

Version

v4.0.6

IDE Name

Arduino IDE 2

Operating System

macOS

Auto Flash Mode

Disable

Erase All Flash Memory (16MB)

Disable

Standard Lib

Arduino_STD_PRINTF

Upload Speed

2000000

Description

I'm trying to connect ethernet shield and modify library to support AMB82 but never had luck. After digging inside library has found that SPI function transfer() always return 0, but with description should return response from slave: Code from library:

byte SPIClass::transfer(uint8_t _data, SPITransferMode _mode){ // transfer 1 byte data without SS
    (void)_mode;
     spi_master_write((spi_t *)pSpiMaster, _data);
    //printf("Master write: %02X\n\r", _data);
    return 0;
}

Same time, _mode has no effect in this point as well.

Description official API says: Returns This function either returns NA or data of 8-bits that transferred through SPI master to the slave. but from code, it seams that 0 should always returned.

Function spi_master_write returning data, has read after data transfer so far I found inside SDK: (file: component/mbed/targets/hal/rtl8735b/spi_api.c)

static inline void ssi_write(spi_t *obj, int value)
{
    phal_ssi_adaptor_t phal_ssi_adaptor = &(obj->hal_ssi_adaptor);
        while (!hal_ssi_writable(phal_ssi_adaptor));
    hal_ssi_write(phal_ssi_adaptor, value);
}

static inline int ssi_read(spi_t *obj)
{
    phal_ssi_adaptor_t phal_ssi_adaptor = &(obj->hal_ssi_adaptor);

    while (!hal_ssi_readable(phal_ssi_adaptor));
    return (int)hal_ssi_read(phal_ssi_adaptor);
}

int spi_master_write(spi_t *obj, int value)
{
    ssi_write(obj, value);
    return ssi_read(obj);
}

Which means that data has been received after transfer should be received.

Meanwhile, I've tried to modify SPI.cpp file with some like this:

byte SPIClass::transfer(uint8_t data, SPITransferMode mode) { // transfer 1 byte data without SS
    (void)mode;
    spi_master_write(pSpiMaster, data);
    u_int8_t _rcvData = spi_master_write(pSpiMaster, data);
    printf("\r\n[INFO] Master write: %02X receive %02X\n", data, _rcvData);

    return _rcvData;
}

and now basic loopback test (jumped SI/SO pins) works, but anyway in general drivers for SPI devices that works not only in "TRANSFER_ONLY" mode, it's not enough because driver stacks at something like:

17:45:28.529 -> 
[INFO] Master write: 00 receive FF
17:45:28.529 -> 

17:45:28.529 -> 
[INFO] Master write: 00 receive FF
17:45:28.529 -> 

17:45:28.529 -> 
[INFO] Master write: 0A receive FF
17:45:28.529 -> 

17:45:28.529 -> 
[INFO] Master write: 00 receive FF
17:45:28.529 -> 

17:45:28.529 -> 
[INFO] Master write: 00 receive FF
17:45:28.529 -> 

17:45:28.529 -> 
[INFO] Master write: 0A receive FF
17:45:28.529 -> 

17:45:28.529 -> 
[INFO] Master write: 00 receive FF

Which means that data response not transferred back or it works in some mode, that a'm not familiar yet.

Same time, function transfer16 will never works well, such as inside method data respect to receiver response from transfer function which is always 0:

uint16_t SPIClass::transfer16(byte pin, uint16_t data, SPITransferMode mode) {
    union {
        uint16_t val;
        struct {
            uint8_t lsb;
            uint8_t msb;
        };
    } t;

    t.val = data;

    if (_bitOrder == LSBFIRST) {
        t.lsb = transfer(pin, t.lsb, SPI_CONTINUE);// <<<< ----- HERE
        t.msb = transfer(pin, t.msb, mode); // <<<< ----- HERE
    } else {
        t.msb = transfer(pin, t.msb, SPI_CONTINUE); // <<<< ----- HERE
        t.lsb = transfer(pin, t.lsb, mode); // <<<< ----- HERE
    }
    //printf("\r\n[INFO] Master write: %04X\n", t.val);

    data = t.val;
    return data;
}

Sketch

/*
 * A sketch to perform a LOOPBACK
 */ 
#include <SPI.h>

uint16_t caunter;
uint16_t inn;

void setup() {

    Serial.begin(115200);
    Serial.println("SPI Loopback Test!"); 

    SPI.begin();
    delay(10);  
}

void loop(){

  for (caunter = 0; caunter <= 1000; caunter++) {

    inn = SPI.transfer(caunter,SPI_CONTINUE);

     Serial.print (caunter, DEC);
     Serial.print(" -- ");
     Serial.println(inn, HEX);

delay(500);

   }
}

Error/Debug Message

18:27:59.405 -> store fcs data for application 
18:27:59.405 -> RAM TM_STATUS 0x00bf1208 err 0x00001208
18:27:59.405 ->  read fcs_status 0x000000bf
18:27:59.438 ->  read fcs_status 0x000000bf
18:27:59.438 -> === Process VOE IMG ===
18:27:59.438 -> [Image Start Table @ 0x20106200]
18:27:59.438 -> RAM Load @ 0x80fd100->0x20106200, 0x5c48
18:27:59.438 -> DDR Load @ 0x8103080->0x70100000, 0x180e4
18:27:59.438 -> === FW Load Done ===
18:27:59.438 -> 
18:27:59.438 -> Boot Loader <==
18:27:59.438 -> 
18:27:59.438 -> == RAM Start ==
18:27:59.470 -> Build @ 17:05:27, Dec 20 2023
18:27:59.470 -> 
18:27:59.470 -> $8735b>SPI Loopback Test!
18:27:59.470 -> 0 -- 0
18:27:59.963 -> 1 -- 0
18:28:00.457 -> 2 -- 0
18:28:00.986 -> 3 -- 0
18:28:01.477 -> 4 -- 0
18:28:01.966 -> 5 -- 0
18:28:02.456 -> 6 -- 0
18:28:02.983 -> 7 -- 0
18:28:03.474 -> 8 -- 0
18:28:03.967 -> 9 -- 0
18:28:04.458 -> 10 -- 0
18:28:04.984 -> 11 -- 0
18:28:05.481 -> 12 -- 0
18:28:05.972 -> 13 -- 0

Reproduce remarks

No response

I have checked online documentation, FAQ, GitHub Wiki and existing/closed issues.

• [X] I confirm I have checked online documentation, FAQ, GitHub Wiki and existing/closed issues.

[github-actions[bot]]

Hello, hope this message finds you well. Congrats to your first Issue! We will review it as soon as possiable. Feel free to have a look at https://www.amebaiot.com/en/ameba-arduino-summary/ for more information

[github-actions[bot]]

This issue is stale because it has been open for 14 days with no activity.

[M-ichae-l]

the board hardware is

Boards AMB23/SPI

Could I double check the hardware?　

[acrist]

Board is AMB82-MINI