nandland / spi-master Goto Github PK

The MAX_BYTES_PER_CS parameter setups the length of the r_TX_Count variable which determines the max amount of bytes per active level of chip select signal. As result the maximum number of bytes per CS becomes a 2^(len(r_TX_Count)). So the settings become like this:

Hi, I just ran a syntax check on your VHDL source code and it seems there is an error with the signal o_tx_ready. Essentially, you are reading an output signal, which should not be allowed. I haven't run the simulation though, so perhaps I might mistake. In that case, I do apologise.

Command used:

ghdl -a VHDL/source/*

Output:

VHDL/source/SPI_Master.vhd:161:10:error: port "o_tx_ready" cannot be read
VHDL/source/SPI_Master.vhd:187:10:error: port "o_tx_ready" cannot be read

  parameter SPI_MODE = 3;           // CPOL = 1, CPHA = 1
  parameter CLKS_PER_HALF_BIT = 4;  // 6.25 MHz
  parameter MAIN_CLK_DELAY = 2;     // 25 MHz
  parameter MAX_BYTES_PER_CS = 2;   // 2 bytes per chip select
  parameter CS_INACTIVE_CLKS = 10;  // Adds delay between bytes

Here the actual clockspeed to me seems 250 MHz instead of 25 MHz. Also with a CLKS_PER_HALF_BIT of 4 this turns out to be 250 MHz/8 = 31.25 MHz maybe a zero was deleted from the MAIN_CLK_DELAY? So you meat to set it to 20 ns? if you want on the other hands the values in the comment it would require

  parameter SPI_MODE = 3;           // CPOL = 1, CPHA = 1
  parameter CLKS_PER_HALF_BIT = 2;  // 6.25 MHz
  parameter MAIN_CLK_DELAY = 20;     // 25 MHz
  parameter MAX_BYTES_PER_CS = 2;   // 2 bytes per chip select
  parameter CS_INACTIVE_CLKS = 10;  // Adds delay between bytes

I guess this line should be using the std_logic_1164 function "to_bitvector", shouldn't it?
Vivado claims an invalid type-conversion here otherwise