Mixed type operations with arrays

Important facts when using arrays

Important facts:

An address is used to identify a memory location

An address is always 32 bits

A SSN for humans is always 9 digits:

123-45-7890 090-00-0012 <---- Leading 0's are significant !!! 000-00-0001 <---- Leading 0's are significant !!!

Therfore:

You must use:

move.l #Label, .... * Use .l for 32 bits address !!!

Assignment with arrays of mixed data type

Large = small

int A[10]; short B[10]; A[2] = B[3];
In assembler code: movea.l #B, a0 An address is always 32 bits, so use movea.l move.w 6(a0), d0 d0 = B[3] (16 bits) You are transfering B[3] which is a short, so move.w ext.l d0 convert B[3] into 32 bits movea.l #A, a0 An address is always 32 bits, so use movea.l move.l d0, 8(a0) You are transfering 32 bits so move.l

Small = large (dangerous conversion !)

int A[10]; short B[10]; B[3] = A[2];
In assembler code: movea.l #A, a0 An address is always 32 bits, so use movea.l move.l 8(a0), d0 d0 = A[2] (32 bits) movea.l #B, a0 An address is always 32 bits, so use movea.l move.2 d0, 6(a0) Transfer only the lower 16 bits of d0 into B[3]

Calculations with arrays of mixed type: convert smaller to larger representation first

Example 1: int = int + short

In Java: int A[10]; short B[10]; A[4] = A[7] + B[5];
In assembler: movea.l #A, a0 * Get the base address of array A in A0 move.l 28(a0), d0 * Get the int value A[7] in D0 movea.l #B, a1 * Get the base address of array B in A1 move.w 10(a1), d1 * Get the short value B[5] in D1 ext.l d1 * Convert 16 bit repr to 32 bit repr add.l d1, d0 * Add two 32 bit numbers together move.l d0, 28(a0) * Store the 32 bits result in A[4]

Example 2: short = int + short

In Java: int A[10]; short B[10]; B[4] = (short) (A[7] + B[5]);
In assembler: movea.l #A, a0 * Get the base address of array A in A0 move.l 28(a0), d0 * Get the int value A[7] in D0 movea.l #B, a1 * Get the base address of array B in A1 move.w 10(a1), d1 * Get the short value B[5] in D1 ext.l d1 * Convert 16 bit repr to 32 bit repr add.l d1, d0 * Add two 32 bit numbers together * The result is 32 bits !!! move.w d0, 8(a0) * Store only 16 bits from D0 in B[4]