Conceptual questions regarding SPI master for nRF51822

Question

I have successfully tested the SPI master loopback example on pca10001 eval board. I have a few questions regarding a real-world scenario when there is a slave SPI present.
In the loop-back test the RXD byte is immediately available when TXD is written. In a non-loopback scenario, SPI master needs to wait for a valid RX byte. I don't know how to implement this wait. Should I keep reading RXD and clear EVENTS_READY till I get a valid RX byte or is "spi_master_tx_rx" smart enough to throw away blank RX bytes until it receives a real RX byte and put it in the RXD? 
What triggers receiving a new RX byte? Clearing EVENTS_READY or reading the RXD? or both?
Could you please point me to a simple example of how this is done?

Chris Lovett · Accepted Answer

spi_master_tx_rx does both, so for example, to read a register from my sensor over SPI I have the following code that works: 
 uint8_t readSpiByte(uint32_t *spi_base_address, uint32_t slaveSelect, uint8_t reg)
{
// write a byte from tx_data, then read a byte into rx_data.
// since these are done sequentially we need a 2 byte buffer to
// capture the result.
uint8_t tx_data[2];
tx_data[0] = READ_BIT | (reg << 3);
tx_data[1] = tx_data[0];
uint8_t rx_data[2]; 
 // Transmit TX_RX_MSG_LENGTH bytes from tx_data and receive same 
// number of bytes and data into rx_data
if(!spi_master_tx_rx(spi_base_address, slaveSelect, 2, tx_data, rx_data) )
{
 return 0;
}	
return rx_data[1];
 
 }