ARMmbed
/
mbed-os
mirror of https://github.com/ARMmbed/mbed-os.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

Simplify spi Asynch transfer implementation 

Following discussion on:

it seems now clear that the transfer API is meant to be used either
with only Rx, or only Tx or Rx and Tx with the same lenth.

Therefore we're removing support of transfers of Rx and Tx with different
lenghts - this makes porting to HAL more direct and simpler.
pull/2852/head
Laurent MEUNIER 2016-09-13 17:01:23 +02:00 committed by Anna Bridge
parent bcf7ee2e51
commit 9323ad204a
1 changed files with 14 additions and 45 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

59
hal/targets/hal/TARGET_STM/TARGET_STM32F4/spi_api.c
View File

@ -433,7 +433,7 @@ static int spi_master_start_asynch_transfer(spi_t *obj, transfer_type_t transfer
NVIC_EnableIRQ(irq_n);
// enable the right hal transfer
static uint16_t sink;
//static uint16_t sink;
int rc = 0;
switch(transfer_type) {
case SPI_TRANSFER_TYPE_TXRX:
@ -442,9 +442,9 @@ static int spi_master_start_asynch_transfer(spi_t *obj, transfer_type_t transfer
case SPI_TRANSFER_TYPE_TX:
// TODO: we do not use `HAL_SPI_Transmit_IT`, since it has some unknown bug
// and makes the HAL keep some state and then that fails successive transfers
// rc = HAL_SPI_Transmit_IT(handle, (uint8_t*)tx, words);
rc = HAL_SPI_TransmitReceive_IT(handle, (uint8_t*)tx, (uint8_t*)&sink, 1);
length = is16bit ? 2 : 1;
rc = HAL_SPI_Transmit_IT(handle, (uint8_t*)tx, words);
//rc = HAL_SPI_TransmitReceive_IT(handle, (uint8_t*)tx, (uint8_t*)&sink, 1);
//length = is16bit ? 2 : 1;
break;
case SPI_TRANSFER_TYPE_RX:
// the receive function also "transmits" the receive buffer so in order
@ -503,8 +503,13 @@ void spi_master_transfer(spi_t *obj, const void *tx, size_t tx_length, void *rx,
// enable the right hal transfer
if (use_tx && use_rx) {
// transfer with the min(tx, rx), then later either transmit _or_ receive the remainder
// we cannot manage different rx / tx sizes, let's use smaller one
size_t size = (tx_length < rx_length)? tx_length : rx_length;
if(tx_length != rx_length) {
DEBUG_PRINTF("SPI: Full duplex transfer only 1 size: %d\n", size);
obj->tx_buff.length = size;
obj->rx_buff.length = size;
}
spi_master_start_asynch_transfer(obj, SPI_TRANSFER_TYPE_TXRX, tx, rx, size);
} else if (use_tx) {
spi_master_start_asynch_transfer(obj, SPI_TRANSFER_TYPE_TX, tx, NULL, tx_length);
@ -524,25 +529,7 @@ uint32_t spi_irq_handler_asynch(spi_t *obj)
HAL_SPI_IRQHandler(handle);
if (HAL_SPI_GetState(handle) == HAL_SPI_STATE_READY) {
// adjust buffer positions
size_t tx_size = (handle->TxXferSize - handle->TxXferCount);
size_t rx_size = (handle->RxXferSize - handle->RxXferCount);
// 16 bit transfers need to be doubled to get bytes
if (handle->Init.DataSize == SPI_DATASIZE_16BIT) {
tx_size *= 2;
rx_size *= 2;
}
// adjust buffer positions
if (obj->spi.transfer_type != SPI_TRANSFER_TYPE_RX) {
obj->tx_buff.pos += tx_size;
}
if (obj->spi.transfer_type != SPI_TRANSFER_TYPE_TX) {
obj->rx_buff.pos += rx_size;
}
if (handle->TxXferCount > 0) {DEBUG_PRINTF("SPI: TxXferCount: %u\n", handle->TxXferCount);}
if (handle->RxXferCount > 0) {DEBUG_PRINTF("SPI: RxXferCount: %u\n", handle->RxXferCount);}
// When HAL SPI is back to READY state, check if there was an error
int error = HAL_SPI_GetError(handle);
if(error != HAL_SPI_ERROR_NONE) {
// something went wrong and the transfer has definitely completed
@ -553,27 +540,9 @@ uint32_t spi_irq_handler_asynch(spi_t *obj)
event |= SPI_EVENT_RX_OVERFLOW;
}
} else {
// figure out if we need to transfer more data:
if (obj->tx_buff.pos < obj->tx_buff.length) {
//DEBUG_PRINTF("t%u ", obj->tx_buff.pos);
// we need to transfer more data
spi_master_start_asynch_transfer(obj, SPI_TRANSFER_TYPE_TX,
obj->tx_buff.buffer + obj->tx_buff.pos, // offset the initial buffer by the position
NULL, // there is no receive buffer
obj->tx_buff.length - obj->tx_buff.pos); // transfer the remaining bytes only
} else if (obj->rx_buff.pos < obj->rx_buff.length) {
//DEBUG_PRINTF("r%u ", obj->rx_buff.pos);
// we need to receive more data
spi_master_start_asynch_transfer(obj, SPI_TRANSFER_TYPE_RX,
NULL, // there is no transmit buffer
obj->rx_buff.buffer + obj->rx_buff.pos, // offset the initial buffer by the position
obj->rx_buff.length - obj->rx_buff.pos); // transfer one byte at a time, until we received everything
} else {
// everything is ok, nothing else needs to be transferred
event = SPI_EVENT_COMPLETE | SPI_EVENT_INTERNAL_TRANSFER_COMPLETE;
DEBUG_PRINTF("SPI: Done: %u, %u\n", obj->tx_buff.pos, obj->rx_buff.pos);
}
}
// else we're done
event = SPI_EVENT_COMPLETE | SPI_EVENT_INTERNAL_TRANSFER_COMPLETE;
}
}
if (event) DEBUG_PRINTF("SPI: Event: 0x%x\n", event);