changing spi master mode gives wrong value about caravel_openframe_project HOT 2 OPEN

I'm not sure what I'm looking at in the above graphs. . . SCK and SDO both from the SPI master? In principle, it looks right; the SDO is aligned with the rising edge of the clock in one mode and the falling edge of the clock in the other mode. But do you have an SPI slave in the testbench that can also be put in mode 0 or mode 1? The modes of master and slave have to match, or else yes, there will definitely be race conditions and misinterpretation of data.

from caravel_openframe_project.

yes, sck and sdo are from the SPI master.
yes the slave works with different modes.

The mode that align with rising edge is functionally right only the last pose edge reads 1 at this example. However the mode that aligns with negative edge reads 1 at edges number 6 and 7 (counting from 0)

Here are some other trials sending byte in this mode to provide more details about the shift between clock driving and data sending.

try sending 0x00 but it sends 0x1

try sending 0x7F but it sends 0xFF

try sending 0x02 but it sends 0x05

try sending 0x03 but it sends 0x07

from caravel_openframe_project.