Keystone II DDR3Register Calc spreadsheet usage

Topi,

You misunderstand the intent of the read-modify-write mask.  It is a negative-logic mask.  For instance the PGCR1 mask is 0x0180_0184.  It would be used to CLEAR all of the bits set.  Therefore, during the init sequence, you would invert it (ones complement) and then AND it with the value read from the register.  Then once the bits of interest are cleared, you want to OR the desired values into these bits.  Here is an example that explicitly masks each field:

DDR3A_PGCR1 &= ~( 0x01800184);
DDR3A_PGCR1 |= ((1 << 2) & 0x00000004); //WLSTEP = [1]
DDR3A_PGCR1 |= ((1 << 7) & 0x00000180);  //IODDRM = [01]
DDR3A_PGCR1 |= ((1 << 23) & 0x01800000);  //ZCKSEL= [01]

The Init App Note breaks this into pieces and simplifies but does the same thing:

// Program WLSTEP=1, IODDRM=1, and ZCKSEL in the PHY General Configuration Register 1
// (address offset 0x00C). All other fields must be left at their default values by using a
// read-modify-write sequence to preserve the other bits. ZCKSEL is chosen from the register
// calculation spreadsheet based on the main PLL rate.
DDR3A_PGCR1 |= (1 << 2); //WLSTEP = 1
DDR3A_PGCR1 &= ~( 0x00000180);
DDR3A_PGCR1 |= (( 1 << 7) & 0x00000180);
DDR3A_PGCR1 &= ~( 0x01800000);
DDR3A_PGCR1 |= (( 1 << 23) & 0x01800000);

Tom