RE: SPI receive

jhellen — Thu, 09 Jan 2025 15:45:11 GMT

The case can be closed. I found the problem. the SPI uses one of the same pins as the near field communication peripheral I was assuming that a disabled NFC peripheral within the device tree also released the pins. This is not the case I has to be done explicitly:

# NO NFC, docs.zephyrproject.org/.../CONFIG_NFCT_PINS_AS_GPIOS.html

CONFIG_NFCT_PINS_AS_GPIOS=y

This took me a long time to figure out. It would be really nice if there was a database addressing these kind of issues.

RE: SPI receive

jhellen — Wed, 08 Jan 2025 12:57:27 GMT

Sorry don't know what you mean,

I've tried all kind of setups. I understand that the both Tx and Rx is active at the same time. Keeping the issue

https://github.com/zephyrproject-rtos/zephyr/issues/65512 in mind I first tried the following as bases, This does not work either.

____________________________________________________________________________________________

static int spi_transfer(const struct device *spi_dev, struct spi_config *spi_cfg, uint8_t *tx_buf, size_t tx_len, uint8_t *rx_buf, size_t rx_len)

{

/// READ github.com/.../65512

struct spi_buf tx_buffer[1] = {

{

.buf = tx_buf,

.len = tx_len,

}

};

struct spi_buf rx_buffer[1] = {

{

.buf = rx_buf,

.len = rx_len,

};

struct spi_buf_set tx_set = {

.buffers = tx_buffer,

.count = 1,

};

struct spi_buf_set rx_set = {

.buffers = rx_buffer,

.count = 1,

};

#if defined(CONFIG_SPI_ASYNC)

return spi_transceive_cb(spi_dev, spi_cfg, &tx_set, &rx_set, spi_rdy, NULL);

#else

/* Synchronous SPI transfer */

return spi_transceive(spi_dev, spi_cfg, &tx_set, &rx_set);

#endif

}

static int read_id(const struct device *spi_dev, struct spi_config *spi_cfg)

{

uint8_t id[5] = { 0xFF };

uint8_t cmd[1] = { MB85RS4MT_RDID };

if (0 != spi_transfer(spi_dev, spi_cfg, cmd, sizeof(cmd), id, sizeof(id)))

{

return -EIO;

}

.....

}

___________________________________________________________________

the tx_buf is 1 byte containing the command. The receive buffer rx_buf contains the ID memory space of 4 byte + 1 byte for reception during transmission of the RDID command. Same result applies. The complete ID[5] content is zero. I should at least expect some data which is not zero in for index 1..4

see logic analyzer for the configuration above. Maybe my interpretation of the complete Zephyr SPI concept is wrong but it is very difficult to find proper background information. Especially the tx/rx config part.

RE: SPI receive

Turbo J — Wed, 08 Jan 2025 10:59:18 GMT

Have you even read the code that you posted? The zero overwrite is expected when looking how the rx_buffers[2] are set up. It completely overwrites the tx_buf memory.

Rember that SPI will always both transmit and receive at the same time.