it seems the Chip Select is not taken into account when decoding SPI data, see this capture :
- Blue is MOSI
- Red is CS
- Green is SCK
- Gold is MISO
Taking a look at ~50us we can see CS is deasserted during a byte cycle (only 2 clock cycles have been done when CS is deasserted).
The decoding engine should abort when CS is released and start over when it is asserted again.
What we see here is the decoding engine takes the 6 first next clock cycles and merges them to the unfinnished previous word to build the next byte. Thus all the decoded data is shifted after this event.
Here are my settings :
Did I do something wrong ?