Maybe, I found a small bug in the spi_lld.c

[Alex]

In my opinion, procedure "spi_lld_start" must contain calling rccResetSPIx, to clear SPI FIFO of "garbage". At least in my case this action helped me. CPU - STM32F051R6C8
P.S. I use NIL, not ChibiOs, but I think they are similar

[Alex]

Additionaly information: SPI driver was used in SLAVE mode = "selfmade"

[Giovanni]

Alex, I think this problem is related to the slave mode, in master mode there is no way that the FIFO still contains data (it would be a bug should this happen).

Giovanni

[samtatasurya]

Additionaly information: SPI driver was used in SLAVE mode = "selfmade"

Hi Alex,

I'm currently working on SPI which requires me to set up one STM32F0 as SPI Slave mode. Do you mind sharing your working code for the Slave?

Your help is greatly appreciated. Thanks!

- Sam