Programming the 65816 Including the 6502, 65C02, and 65802 (1986)

Part V. Reference

18. The Instruction Sets

This chapter devotes a page to each of the 94 different 65816 operations. Each operation may have more than one addressing mode available to it; these are detailed for each instruction. The symbols in Table 18.1 are used to express the different kinds of values that instruction operands may have. The effect of each operation on the status flags varies. The symbols in Table 18.2 are used to indicate the flags that are affected by a given operation.

Table 18.1. Operand Symbols.

Table 18.2. 65x Flags.

Add the data located at the effective address specified by the operand to the contents of the accumulator; add one to the result if the carry flag is set, and store the final result in the accumulator.

The 65x processors have no add instruction which does not involve the carry. To avoid adding the carry flag to the result, you must either be sure that it is already clear, or you must explicitly clear it (using CLC) prior to executing the ADC instruction.

In a multi-precision (multi-word) addition, the carry should be cleared before the low-order words are added; the addition of the low word will generate a new carry flag value based on that addition. This new value in the carry flag is added into the next (middle-order or high-order) addition; each intermediate result will correctly reflect the carry from the previous addition.

d flag clear; Binary addition is performed.

d flag set; Binary coded decimal (BCD) addition is performed.

8-bit accumulator (all processors): Data added from memory is eight-bit.

16-bit accumulator (65802/65816 only, m = 0): Data added from memory is sixteen-bit: the low-order eight bits are located at the effective address; the high-order eight bits are located at the effective address plus one.

Codes:

^{+ +} ADC, a Primary Group Instruction, has available all of the Primary Group addressing modes and bit patterns

* Add 1 byte if m = 0 (16-bit memory/accumulator)

1 Add 1 cycle if m = 0 (16-bit memory/accumulator)

2 Add 1 cycle if low byte of Direct Page register is other than zero (DL< > 0)

3 Add 1 cycle if adding index crosses a page boundary

4 Add 1 cycle if 65C02 and d = 1 (decimal mode, 65C02)

Bitwise logical AND the data located at the effective address specified by the operand with the contents of the accumulator. Each bit in the accumulator is ANDed with the corresponding bit in memory, with the result being stored in the respective accumulator bit.

The truth table for the logical AND operation is:

Figure 18.1. AND Truth Table.

That is, a 1 or logical true results in a given bit being true only if both elements of the respective bits being ANDed are Is, or logically true.

8-bit accumulator (all processors): Data ANDed from memory is eight-bit.

16-bit accumulator (65802/65816 only, m = 0): Data ANDed from memory is sixteen-bit: the low-order byte is located at the effective address; the high-order byte is located at the effective address plus one.

Codes:

^{+ +} AND, a Primary Group Instruction, has available all of the Primary Group addressing modes and bit patterns

* Add 1 byte if m = 0 (16-bit memory/accumulator)

1 Add 1 cycle if m = 0 (16-bit memory/accumulator)

2 Add 1 cycle if low byte of Direct Page register is other than zero (DL< > 0)

3 Add 1 cycle if adding index crosses a page boundary

Shift the contents of the location specified by the operand left one bit. That is, bit one takes on the value originally found in bit zero, bit two takes the value originally in bit one, and so on; the leftmost bit (bit 7 on the 6502 and 65C02 or if m = 1 on the 65802/65816, or bit 15 ifm = 0) is transferred into the carry flag; the rightmost bit, bit zero, is cleared. The arithmetic result of the operation is an unsigned multiplication by two.

Figure 18.2. ASL.

8-bit accumulator/memory (all processors): Data shifted is eight bits.

16-bit accumulator/memory (65802/65816 only, m = 0): Data shifted is sixteen bits: if in memory, the low-order eight bits are located at the effective address; the high-order eight bits are located at the effective address plus one.

Codes:

1 Add 2 cycles if m = 0 (16-bit memory/accumulator)

2 Add 1 cycle if low byte of Direct Page register is other than zero (DL< >0)

3 Subtract 1 cycle if 65C02 and no page boundary crossed

The carry flag in the P status register is tested. If it is clear, a branch is taken; if it is set, the instruction immediately following the two-byte BCC instruction is executed.

If the branch is taken, a one-byte signed displacement, fetched from the second byte of the instruction, is sign-extended to sixteen bits and added to the program counter. Once the branch address has been calculated, the result is loaded into the program counter, transferring control to that location.

The allowable range of the displacement is — 128 to +127 (from the instruction immediately following the branch).

BCC may be used in several ways: to test the result of a shift into the carry; to determine if the result of a comparison is either less than (in which case a branch will be taken), or greater than or equal (which causes control to fall through the branch instruction); or to determine if further operations are needed in multi-precision arithmetic.

Because the BCC instruction causes a branch to be taken after a comparison or subtraction if the accumulator is less than the memory operand (since the carry flag will always be cleared as a result), many assemblers allow an alternate mnemonic for the BCC instruction: BLT, or Branch if Less Than.

Flags Affected: -------------------

Codes:

1 Add 1 cycle if branch is taken

2 Add 1 more cycle if branch taken crosses page boundary on 6502, 65C02, or 65816/65802's 6502 emulation mode (e = 1)

The carry flag in the P status register is tested. If it is set, a branch is taken; if it is clear, the instruction immediately following the two-byte BCS instruction is executed.

If the branch is taken, a one-byte signed displacement, fetched from the second byte of the instruction, is sign-extended to sixteen bits and added to the program counter. Once the branch address has been calculated, the result is loaded into the program counter, transferring control to that location.

The allowable range of the displacement is — 128 to + 127 (from the instruction immediately following the branch).

BCS is used in several ways: to test the result of a shift into the carry; to determine if the result of a comparison is either greater than or equal (which causes the branch to be taken) or less than; or to determine if further operations are needed in multi-precision arithmetic operations.

Because the BCS instruction causes a branch to be taken after a comparison or subtraction if the accumulator is greater than or equal to the memory operand (since the carry flag will always be set as a result), many assemblers allow an alternate mnemonic for the BCSinstruction: BGE or Branch if Greater or Equal.

Flags Affected: -------------------

Codes:

1 Add 1 cycle if branch is taken

2 Add 1 more cycle if branch taken crosses page boundary on 6502, 65C02, or 65816/65802's 6502 emulation mode (e = 1)

The zero flag in the P status register is tested. If it is set, meaning that the last value tested (which affected the zero flag) was zero, a branch is taken; if it is clear, meaning the value tested was non-zero, the instruction immediately following the two-byte BEQ instruction is executed.

If the branch is taken, a one-byte signed displacement, fetched from the second byte of the instruction, is sign-extended to sixteen bits and added to the program counter. Once the branch address has been calculated, the result is loaded into the program counter, transferring control to that location.

The allowable range of the displacement is — 128 to +127 (from the instruction immediately following the branch).

BEQ may be used in several ways: to determine if the result of a comparison is zero (the two values compared are equal), for example, or if a value just loaded, pulled, shifted, incremented or decremented is zero; or to determine if further operations are needed in multi-precision arithmetic operations. Because testing for equality to zero does not require a previous comparison with zero, it is generally most efficient for loop counters to count downwards, exiting when zero is reached.

Flags Affected: -------------------

Codes:

1 Add 1 cycle if branch is taken

2 Add 1 more cycle if branch taken crosses page boundary on 6502, 65C02, or 65816/65802's 6502 emulation mode (e = 1)

BIT sets the P status register flags based on the result of two different operations, making it a dual-purpose instruction:

First, it sets or clears the n flag to reflect the value of the high bit of the data located at the effective address specified by the operand, and sets or clears the v flag to reflect the contents of the next-to-highest bit of the data addressed.

Second, it logically ANDs the data located at the effective address with the contents of the accumulator; it changes neither value, but sets the z flag if the result is zero, or clears it if the result is non-zero.

BIT is usually used immediately preceding a conditional branch instruction: to test a memory value's highest or next-to-highest bits; with a mask in the accumulator, to test any bits of the memory operand; or with a constant as the mask (using immediate addressing) or a mask in memory, to test any bits in the accumulator. All of these tests are non-destructive of the data in the accumulator or in memory. When the BIT instruction is used with the immediate addressing mode, the n and v flags are unaffected.

8-bit accumulator/memory (all processors): Data in memory is eight-bit; bit 7 is moved into the n flag; bit 6 is moved into the v flag.

16-bit accumulator/memory (65802/65816 only, m = 0): Data in memory is sixteen-bit: the low-order eight bits are located at the effective address; the high-order eight bits are located at the effective address plus one. Bit 15 is moved into the n flag; bit 14 is moved into the v flag.

Codes:

* Add 1 byte if m = 0 (16-bit memory/accumulator)

1 Add 1 cycle if m = 0 (16-bit memory/accumulator)

2 Add 1 cycle if low byte of Direct Page register is other than zero (DL< >0)

3 Add 1 cycle if adding index crosses a page boundary

The negative flag in the P status register is tested. If it is set, the high bit of the value which most recently affected the n flag was set, and a branch is taken. A number with its high bit set may be interpreted as a negative two's-complement number, so this instruction tests, among other things, for the sign of two's-complement numbers. If the negative flag is clear, the high bit of the value which most recently affected the flag was clear, or, in the two's-complement system, was a positive number, and the instruction immediately following the two-byte BMI instruction is executed.

If the branch is taken, a one-byte signed displacement, fetched from the second byte of the instruction, is sign-extended to sixteen bits and added to the program counter. Once the branch address has been calculated, the result is loaded into the program counter, transferring control to that location.

The allowable range of the displacement is - 128 to + 127 (from the instruction immediately following the branch).

BMI is primarily used to either determine, in two's-complement arithmetic, if a value is negative or, in logic situations, if the high bit of the value is set. It can also be used when looping down through zero (the loop counter must have a positive initial value) to determine if zero has been passed and to effect an exit from the loop.

Flags Affected: -------------------

Codes:

1 Add 1 cycle if branch is taken

2 Add 1 more cycle if branch taken crosses page boundary on 6502, 65C02, or 65816/65802's 6502 emulation mode (e = 1)

The zero flag in the P status register is tested. If it is clear (meaning the value just tested is non-zero), a branch is taken; if it is set (meaning the value tested is zero), the instruction immediately following the two-byte BNE instruction is executed.

If the branch is taken, a one-byte signed displacement, fetched from the second byte of the instruction, is sign-extended to sixteen bits and added to the program counter. Once the branch address has been calculated, the result is loaded into the program counter, transferring control to that location.

The allowable range of the displacement is —128 to +1 2 7 (from the instruction immediately following the branch).

BNE may be used in several ways: to determine if the result of a comparison is non-zero (the two values compared are not equal), for example, or if the value just loaded or pulled from the stack is non-zero, or to determine if further operations are needed in multi-precision arithmetic operations.

Flags Affected: -------------------

Codes:

1 Add 1 cycle if branch is taken

2 Add 1 more cycle if branch taken crosses page boundary on 6502, 65C02, or 65816/65802's 6502 emulation mode (e = 1)

The negative flag in the P status register is tested. If it is clear—meaning that the last value which affected the zero flag had its high bit clear—a branch is taken. In the two's-complement system, values with their high bit clear are interpreted as positive numbers. If the flag is set, meaning the high bit of the last value was set, the branch is not taken; it is a two's-complement negative number, and the instruction immediately following the two-byte BPL instruction is executed.

If the branch is taken, a one-byte signed displacement, fetched from the second byte of the instruction, is sign-extended to sixteen bits and added to the program counter. Once the branch address has been calculated, the result is loaded into the program counter, transferring control to that location.

The allowable range of the displacement is —128 to -I-127 (from the instruction immediately following the branch).

BPL is used primarily to determine, in two's-complement arithmetic, if a value is positive or not or, in logic situations, if the high bit of the value is clear.

Flags Affected: -------------------

Codes:

1 Add 1 cycle if branch is taken

2 Add 1 more cycle if branch taken crosses page boundary on 6502, 65C02, or 65816/65802's 6502 emulation mode (e = 1)

A branch is always taken, and no testing is done: in effect, an unconditional JMP is executed, but since signed displacements are used, the instruction is only two bytes, rather than the three bytes of a JMP. Additionally, using displacements from the program counter makes theBRA instruction relocatable. Unlike a JMP instruction, the BRA is limited to targets that lie within the range of the one-byte signed displacement of the conditional branches: — 128 to + 127 bytes from the first byte following the BRA instruction.

To branch, a one-byte signed displacement, fetched from the second byte of the instruction, is sign-extended to sixteen bits and added to the program counter. Once the branch address has been calculated, the result is loaded into the program counter, transferring control to that location.

Flags Affected: -------------------

Codes:

1 Add 1 cycle if branch crosses page boundary on 65C02 or in 65816/65802's 6502 emulation mode (e = 1)

Force a software interrupt. BRK is unaffected by the i interrupt disable flag.

Although BRK is a one-byte instruction, the program counter (which is pushed onto the stack by the instruction) is incremented by two; this lets you follow the break instruction with a one-byte signature byte indicating which break caused the interrupt. Even if a signature byte is not needed, either the byte following the BRK instruction must be padded with some value or the break-handling routine must decrement the return address on the stack to let an RTI (return from interrupt) instruction execute correctly.

6502, 65C02, and Emulation Mode (e = 1): The program counter is incremented by two, then pushed onto the stack; the status register, with the b break flag set, is pushed onto the stack; the interrupt disable flag is set; and the program counter is loaded from the interrupt vector at $FFFE-FFFF. It is up to the interrupt handling routine at this address to check the b flag in the stacked status register to determine if the interrupt was caused by a software interrupt (BRK) or by a hardware IRQ, which shares the BRK vector but pushes the status register onto the stack with the b break flag clear. For example,

Fragment 18.1.

65802/65816 Native Mode (e = 0): The program counter bank register is pushed onto the stack; the program counter is incremented by two and pushed onto the stack; the status register is pushed onto the stack; the interrupt disable flag is set; the program bank register is cleared to zero; and the program counter is loaded from the break vector at $OOFFE6-OOFFE7.

6502: The d decimal flag is not modified after a break is executed.

65C02 and 65816/65802: The d decimal flag is reset to 0 after a break is executed.

Figure 18.3. 65802/65816 Stack After BRK.

Codes:

* BRK is 1 byte, but program counter value pushed onto stack is incremented by 2 allowing for optional signature byte

1 Add 1 cycle for 65802/65816 native mode (e = 0)

A branch is always taken, similar to the BRA instruction. However, BRL is a three-byte instruction; the two bytes immediately following the opcode form a sixteen-bit signed displacement from the program counter. Once the branch address has been calculated, the result is loaded into the program counter, transferring control to that location.

The allowable range of the displacement is anywhere within the current 64K program bank.

The long branch provides an unconditional transfer of control similar to the JMP instruction, with one major advantage: the branch instruction is relocatable while jump instructions are not. However, the (non-relocatable) jump absolute instruction executes one cycle faster.

Flags Affected: -------------------

Codes:

The overflow flag in the P status register is tested. If it is clear, a branch is taken; if it is set, the instruction immediately following the two-byte BVC instruction is executed.

If the branch is taken, a one-byte signed displacement, fetched from the second byte of the instruction, is sign-extended to sixteen bits and added to the program counter. Once the branch address has been calculated, the result is loaded into the program counter, transferring control to that location.

The allowable range of the displacement is —128 to -1-127 (from the instruction immediately following the branch).

The overflow flag is altered by only four instructions on the 6502 and 65C02—addition, subtraction, the CLV clear-the-flag instruction, and the BIT bit-testing instruction. In addition, all the flags are restored from the stack by the PLP and RTI instructions. On the 65802/65816, however, the SEP and REP instructions can also modify the v flag.

BVC is used almost exclusively to check that a two's-complement arithmetic calculation has not overflowed, much as the carry is used to determine if an unsigned arithmetic calculation has overflowed. (Note, however, that the compare instructions do not affect the overflow flag.) You can also use BVC to test the second—highest bit in a value by using it after the BIT instruction, which moves the second-highest bit of the tested value into the v flag.

The overflow flag can also be set by the Set Overflow hardware signal on the 6502, 65C02, and 65802; on many systems, however, there is no connection to this pin.

Flags Affected: -------------------

Codes:

1 Add 1 cycle if branch is taken

2 Add 1 more cycle if branch taken crosses page boundary on 6502, 65C02, or 65816/65802's 6502 emulation mode (e = 1)

The overflow flag in the P status register is tested. If it is set, a branch is taken; if it is clear, the instruction immediately following the two-byte BVS instruction is executed.

If the branch is taken, a one-byte signed displacement, fetched from the second byte of the instruction, is sign-extended to sixteen bits and added to the program counter. Once the branch address has been calculated, the result is loaded into the program counter, transferring control to that location.

The allowable range of the displacement is —128 to + 127 (from the instruction immediately following the branch).

The overflow flag is altered by only four instructions on the 6502 and 65C02—addition, subtraction, the CLV clear-the-flag instruction, and the BIT bit-testing instruction. In addition, all the flags are restored from the stack by the PLP and RTI instructions. On the 65802/65816, the SEP and REP instructions can also modify the v flag.

BVS is used almost exclusively to determine if a two's-complement arithmetic calculation has overflowed, much as the carry is used to determine if an unsigned arithmetic calculation has overflowed. (Note, however, that the compare instructions do not affect the overflow flag.) You can also use BVS to test the second-highest bit in a value by using it after the BIT instruction, which moves the second-highest bit of the tested value into the v flag.

The overflow flag can also be set by the Set Overflow hardware signal on the 6502, 65C02, and 65802; on many systems, however, there is no hardware connection to this signal.

Flags Affected: -------------------

Codes:

1 Add 1 cycle if branch is taken

2 Add 1 more cycle if branch taken crosses page boundary on 6502, 65C02, or 65816/65802's 6502 emulation mode (e = 1)

Clear the carry flag in the status register.

CLC is used prior to addition (using the 65x's ADC instruction) to keep the carry flag from affecting the result; prior to a BCC (branch on carry clear) instruction on the 6502 to force a branch-always; and prior to an XCE (exchange carry flag with emulation bit) instruction to put the 65802 or 65816 into native mode.

Codes:

Clear the decimal mode flag in the status register.

CLD is used to shift 65x processors back into binary mode from decimal mode, so that the ADC and SBC instructions will correctly operate on binary rather than BCD data.

Codes:

Clear the interrupt disable flag in the status register.

CLI is used to re-enable hardware interrupt (IRQ) processing. (When the i bit is set, hardware interrupts are ignored.) The processor itself sets the i flag when it begins servicing an interrupt, so interrupt handling routines must re-enable interrupts with CLI if the interrupt-service routine is designed to service interrupts that occur while a previous interrupt is still being handled; otherwise, the RTI instruction will restore a clear i flag from the stack, and CLI is not necessary. CLI is also used to re-enable interrupts if they have been disabled during the execution of time-critical or other code which cannot be interrupted.

Codes:

Clear the overflow flag in the status register.

CLV is sometimes used prior to a BVC (branch on overflow clear) to force a branch-always on the 6502. Unlike the other clear flag instructions, there is no complementary "set flag" instruction to set the overflow flag, although the overflow flag can be set by hardware via the Set Overflow input pin on the processor. This signal, however, is often unconnected. The 65802/65816 REP instruction can, of course, clear the overflow flag; on the 6502 and 65C02, a BIT instruction with a mask in memory that has bit 6 set can be used to set the overflow flag.

Codes:

Subtract the data located at the effective address specified by the operand from the contents of the accumulator, setting the carry, zero, and negative flags based on the result, but without altering the contents of either the memory location or the accumulator. That is, the result is not saved. The comparison is of unsigned binary values only.

The CMP instruction differs from the SBC instruction in several ways. First, the result is not saved. Second, the value in the carry prior to the operation is irrelevant to the operation; that is, the carry does not have to be set prior to a compare as it is with 65x subtractions. Third, the compare instruction does not set the overflow flag, so it cannot be used for signed comparisons. Although decimal mode does not affect the CMP instruction, decimal comparisons are effective, since the equivalent binary values maintain the same magnitude relationships as the decimal values have, for example, $99 > $04 just as 99 > 4.

The primary use for the compare instruction is to set the flags so that a conditional branch can then be executed.

8-bit accumulator (all processors); Data compared is eight-bit.

16-bit accumulator (65802/65816 only, m = 0): Data compared is sixteen-bit: the low-order eight bits of the data in memory are located at the effective address; the high-order eight bits are located at the effective address plus one.

Codes:

+ + CMP, a Primary Group Instruction, has available all of the Primary Group addressing modes and bit patterns

* Add 1 byte if m = 0 (16-bit memory/accumulator)

1 Add 1 cycle if m = 0 (16-bit memory/accumulator)

2 Add 1 cycle if low byte of Direct Page register is other than zero (DL< > 0)

3 Add 1 cycle if adding index crosses a page boundary

Execution of COP causes a software interrupt, similarly to BRK, but through the separate COP vector. Alternatively, COP may be trapped by a co-processor, such as a floating point or graphics processor, to call a co-processor function. COP is unaffected by the i interrupt disable flag.

COP is much like BRK, with the program counter value pushed on the stack being incremented by two; this lets you follow the co-processor instruction with a signature byte to indicate to the co-processor or co-processor handling routine which operation to execute. Unlike theBRK instruction, 65816 assemblers require you to follow the COP instruction with such a signature byte. Signature bytes in the range $80 - $FF are reserved by the Western Design Center for implementation of co-processor control; signatures in the range $00 - $7F are available for use with software-implemented COP handlers.

6502 Emulation Mode (65802/65816, e= 1): The program counter is incremented by two and pushed onto the stack; the status register is pushed onto the stack; the interrupt disable flag is set; and the program counter is loaded from the emulation mode co-processor vector at $FFF4-FFF5. The d decimal flag is cleared after a COP is executed.

65802/65816 Native Mode (e = 0): The program counter bank register is pushed onto the stack; the program counter is incremented by two and pushed onto the stack; the status register is pushed onto the stack; the interrupt disable flag is set; the program bank register is cleared to zero; and the program counter is loaded from the native mode co-processor vector at $00FFE4-00FFE5. The d decimal flag is reset to 0 after a COP is executed.

Figure 18.4. Stack after COP.

Codes:

* COP is 1 byte, but program counter value pushed onto stack is incremented by 2 allowing for optional code byte

1 Add 1 cycle for 65816/65802 native mode (e = 0)

Subtract the data located at the effective address specified by the operand from the contents of the X register, setting the carry, zero, and negative flags based on the result, but without altering the contents of either the memory location or the register. The result is not saved. The comparison is of unsigned values only (except for signed comparison for equality).

The primary use for the CPX instruction is to test the value of the X index register against loop boundaries, setting the flags so that a conditional branch can be executed.

8-bit index registers (all processors): Data compared is eight-bit.

16-bit index registers (65802/65816 only, x = 0): Data compared is sixteen-bit: the low-order eight bits of the data in memory are located at the effective address; the high-order eight bits are located at the effective address plus one.

Codes:

* Add 1 byte if x = 0 (16-bit index registers)

1 Add 1 cycle if x = 0 (16-bit index registers)

2 Add 1 cycle if low byte of Direct Page register is other than zero (DL< >0)

Subtract the data located at the effective address specified by the operand from the contents of the Y register, setting the carry, zero, and negative flags based on the result, but without altering the contents of either the memory location or the register. The comparison is of unsigned values only (except for signed comparison for equality).

The primary use for the CPY instruction is to test the value of the Y index register against loop boundaries, setting the flags so that a conditional branch can be executed.

8-bit index registers (all processors): Data compared is eight-bit.

16-bit index registers (65802/65816 only, x = 0): Data compared is sixteen-bit: the low-order eight bits of the data in memory is located at the effective address; the high-order eight bits are located at the effective address plus one.

Codes:

* Add 1 byte if x = 0 (16-bit index registers)

1 Add 1 cycle if x = 0 (16-bit index registers)

2 Add 1 cycle if low byte of Direct Page register is other than zero (DL< > 0)

Decrement by one the contents of the location specified by the operand (subtract one from the value).

Unlike subtracting a one using the SBC instruction, the decrement instruction is neither affected by nor affects the carry flag. You can test for wraparound only by testing after every decrement to see if the value is zero or negative. On the other hand, you don't need to set the carry before decrementing.

DEC is unaffected by the setting of the d (decimal) flag.

8-bit accumulator/memory (all processors): Data decremented is eight-bit.

16-bit accumulator/memory (65802/65816 only, m = 0): Data decremented is sixteen-bit: if in memory, the low-order eight bits are located at the effective address; the high-order eight bits are located at the effective address plus one.

Codes:

1 Add 2 cycles if m = 0 (16-bit memory/accumulator)

2 Add 1 cycle if low byte of Direct Page register is other than zero (DL< > 0)

3 Subtract 1 cycle if 65C02 and no page boundary crossed

Decrement by one the contents of index register X (subtract one from the value). This is a special purpose, implied addressing form of the DEC instruction.

Unlike using SBC to subtract a one from the value, the DEX instruction does not affect the carry flag; you can test for wraparound only by testing after every decrement to see if the value is zero or negative. On the other hand, you don't need to set the carry before decrementing.

DEX is unaffected by the setting of the d (decimal) flag.

8-bit index registers (all processors): Data decremented is eight-bit.

16-bit index registers (65802/65816 only, x = 0): Data decremented is sixteen-bit.

Codes:

Decrement by one the contents of index register Y (subtract one from the value). This is a special purpose, implied addressing form of the DEC instruction.

Unlike using SBC to subtract a one from the value, the DEY instruction does not affect the carry flag; you can test for wraparound only by testing after every decrement to see if the value is zero or negative. On the other hand, you don't need to set the carry before decrementing.

DEY is unaffected by the setting of the d (decimal) flag.

8-bit index registers (all processors): Data decremented is eight-bit.

16-bit index registers (65802/65816 only, x = 0): Data decremented is sixteen-bit.

Codes:

Bitwise logical Exclusive-OR the data located at the effective address specified by the operand with the contents of the accumulator. Each bit in the accumulator is exclusive-ORed with the corresponding bit in memory, and the result is stored into the same accumulator bit.

The truth table for the logical exclusive-OR operation is:

Figure 18.5. Exclusive OR Truth Table.

A 1 or logical true results only if the two elements of the Exclusive-OR operation are different.

8-bit accumulator (all processors): Data exclusive-ORed from memory is eight-bit.

16-bit accumulator (65802/65816 only, m = 0): Data exclusive-ORed from memory is sixteen-bit: the low-order eight bits are located at the effective address; the high-order eight bits are located at the effective address plus one.

Codes:

^{+ +} EOR, a Primary Group Instruction, has available all of the Primary Group addressing modes and bit patterns

* Add 1 byte if m = 0 (16-bit memory/accumulator)

1 Add 1 cycle if m = 0 (16-bit memory/accumulator)

2 Add 1 cycle if low byte of Direct Page register is other than zero (DL< > 0)

3 Add 1 cycle if adding index crosses a page boundary

Increment by one the contents of the location specified by the operand (add one to the value).

Unlike adding a one with the ADC instruction, however, the increment instruction is neither affected by nor affects the carry flag. You can test for wraparound only by testing after every increment to see if the result is zero or positive. On the other hand, you don't have to clear the carry before incrementing.

The INC instruction is unaffected by the d (decimal) flag.

8-bit accumulator/memory (all processors): Data incremented is eight-bit.

16-bit accumulator/memory (65802/65816 only, m = 0): Data incremented is sixteen-bit: if in memory, the low-order eight bits are located at the effective address; the high-order eight-bits are located at the effective address plus one.

Codes:

1 Add 2 cycles if m = 0 (16-bit memory/accumulator)

2 Add 1 cycle if low byte of Direct Page register is other than zero (DL< >0)

3 Subtract 1 cycle if 65C02 and no page boundary crossed

Increment by one the contents of index register X (add one to the value). This is a special purpose, implied addressing form of the INC instruction.

Unlike using ADC to add a one to the value, the INX instruction does not affect the carry flag. You can execute it without first clearing the carry. But you can test for wraparound only by testing after every increment to see if the result is zero or positive. The INX instruction is unaffected by the d (decimal) flag.

8-bit index registers (all processors): Data incremented is eight-bit.

16-bit index registers (65802/65816 only, x = 0): Data incremented is sixteen-bit.

Codes:

Increment by one the contents of index register Y (add one to the value). This is a special purpose, implied addressing form of the INC instruction.

Unlike using ADC to add one to the value, the INY instruction does not affect the carry flag. You can execute it without first clearing the carry. But you can test for wraparound only by testing after every increment to see if the value is zero or positive. The INY instruction is unaffected by the d (decimal) flag.

8-bit index registers (all processors): Data incremented is eight-bit.

16-bit index registers (65802/65816 only, x = 0): Data incremented is sixteen-bit.

Codes:

Transfer control to the address specified by the operand field.

The program counter is loaded with the target address. If a long JMP is executed, the program counter bank is loaded from the third byte of the target address specified by the operand.

Flags Affected: -------------------

Codes:

1 Add 1 cycle if 65C02

2 6502: If low byte of addr is $FF (i.e., addr is SxxFF): yields incorrect result

Jump-to-subroutine with long (24-bit) addressing: transfer control to the subroutine at the 24-bit address which is the operand, after first pushing a 24-bit (long) return address onto the stack. This return address is the address of the last instruction byte (the fourth instruction byte, or the third operand byte), not the address of the next instruction; it is the return address minus one.

The current program counter bank is pushed onto the stack first, then the high-order byte of the return address and then the low-order byte of the address are pushed on the stack in standard 65x order (low byte in the lowest address, bank byte in the highest address). The stack pointer is adjusted after each byte is pushed to point to the next lower byte (the next available stack location). The program counter bank register and program counter are then loaded with the operand values, and control is transferred to the specified location.

Flags Affected: -------------------

Codes:

Transfer control to the subroutine at the location specified by the operand, after first pushing onto the stack, as a return address, the current program counter value, that is, the address of the last instruction byte (the third byte of a three-byte instruction, the fourth byte of a four-byte instruction), not the address of the next instruction.

If an absolute operand is coded and is less than or equal to $FFFF, absolute addressing is assumed by the assembler; if the value is greater than $FFFF, absolute long addressing is used.

If long addressing is used, the current program counter bank is pushed onto the stack first. Next—or first in the more normal case of intra-bank addressing—the high order byte of the return address is pushed, followed by the low order byte. This leaves it on the stack in standard 65x order (lowest byte at the lowest address, highest byte at the highest address). After the return address is pushed, the stack pointer points to the next available location (next lower byte) on the stack. Finally, the program counter (and, in the case of long addressing, the program counter bank register) is loaded with the values specified by the operand, and control is transferred to the target location.

Flags Affected: -------------------

Codes:

Load the accumulator with the data located at the effective address specified by the operand.

8-bit accumulator (all processors): Data is eight-bit.

16-bit accumulator (65802/65816 only, m = 0): Data is sixteen-bit; the low-order eight bits are located at the effective address; the high-order eight bits are located at the effective address plus one.

Codes:

^{+ +} LDA, a Primary Group Instruction, has available all of the Primary Group addressing modes and bit patterns

* Add 1 byte if m = 0 (16-bit memory/accumulator)

1 Add 1 cycle if m = 0 (16-bit memory/accumulator)

2 Add 1 cycle if low byte of Direct Page register is other than zero (DL< >0)

3 Add 1 cycle if adding index crosses a page boundary

Load index register X with the data located at the effective address specified by the operand.

8-bit index registers (all processors): Data is eight-bit.

16-bit index registers (65802/65816 only, x = 0): Data is sixteen-bit: the low-order eight bits are located at the effective address; the high-order eight bits are located at the effective address plus one.

Codes:

* Add 1 byte if x = 0 (16-bit index registers)

1 Add 1 cycle if x = 0 (16-bit index registers)

2 Add 1 cycle if low byte of Direct Page register is other than zero (DL < >0)

3 Add 1 cycle if adding index crosses a page boundary

Load index register Y with the data located at the effective address specified by the operand.

8-bit index registers (all processors): Data is eight-bit.

16-bit index registers (65802/65816 only, x = 0): Data is sixteen-bit: the low-order eight bits are located at the effective address; the high-order eight bits are located at the effective address plus one.

Codes:

* Add 1 byte if x = 0 (16-bit index registers)

1 Add 1 cycle if x = 0 (16-bit index registers)

2 Add 1 cycle if low byte of Direct Page register is other than zero (DL< >0)

3 Add 1 cycle if adding index crosses a page boundary

Logical shift the contents of the location specified by the operand right one bit. That is, bit zero takes on the value originally found in bit one, bit one takes the value originally found in bit two, and so on; the leftmost bit (bit 7 if the m memory select flag is one when the instruction is executed or bit 15 if it is zero) is cleared; the rightmost bit, bit zero, is transferred to the carry flag. This is the arithmetic equivalent of unsigned division by two.

Figure 18.6. LSR.

8-bit accumulator/memory (all processors): Data shifted is eight-bit.

16-bit accumulator/memory (65802/65816 only, m = 0): Data shifted is sixteen-bit: if in memory, the low-order eight bits are located at the effective address; the high-order eight bits are located at the effective address plus one.

Codes:

1 Add 2 cycles if m = 0 (16-bit memory/accumulator)

2 Add 1 cycle if low byte of Direct Page register is other than zero (DL< >0)

3 Subtract 1 cycle if 65C02 and no page boundary crossed

Moves (copies) a block of memory to a new location. The source, destination and length operands of this instruction are taken from the X, Y, and C (double accumulator) registers; these should be loaded with the correct values before executing the MVN instruction.

The source address for MVN, taken from the X register, should be the starting address (lowest in memory) of the block to be moved. The destination address, in the Y register, should be the new starting address for the moved block. The length, loaded into the double accumulator (the value in C is always used, regardless of the setting of the m flag) should be the length of the block to be moved minus one; if C contains $0005, six bytes will be moved. The two operand bytes of the MVN instruction specify the banks holding the two blocks of memory: the first operand byte (of object code) specifies the destination bank; the second operand byte specifies the source bank.

The execution sequence is: the first byte is moved from the address in X to the address in Y; then X and Y are incremented, C is decremented, and the next byte is moved; this process continues until the number of bytes specified by the value in C plus one is moved. In other words, until the value in C is $FFFF.

If the source and destination blocks do not overlap, then the source block remains intact after it has been copied to the destination.

If the source and destination blocks do overlap, then MVN should be used only if the destination is lower than the source to avoid overwriting source bytes before they've been copied to the destination. If the destination is higher, then the MVP instruction should be used instead.

When execution is complete, the value in C is $FFFF, registers X and Y each point one byte past the end of the blocks to which they were pointing, and the data bank register holds the destination bank value (the first operand byte).

Assembler syntax for the block move instruction calls for the operand field to be coded as two addresses, source first, then destination—the more intuitive ordering, but the opposite of the actual operand order in the object code. The assembler strips the bank bytes from the addresses (ignoring the rest) and reverses them to object code order. If a block move instruction is interrupted, it may be resumed automatically via execution of an RTI if all of the registers are restored or intact. The value pushed onto the stack when a block move is interrupted is the address of the block move instruction. The current byte-move is completed before the interrupt is serviced.

If the index registers are in eight-bit mode (x = 1), or the processor is in 6502 emulation mode (e = 1), then the blocks being specified must necessarily be in page zero since the high bytes of the index registers will contain zeroes.

Flags Affected:

Codes:

* 7 cycles per byte moved

Moves (copies) a block of memory to a new location. The source, destination and length operands of this instruction are taken from the X, Y, and C (double accumulator) registers; these should be loaded with the correct values before executing the MVP instruction.

The source address for MVP, taken from the X register, should be the ending address (highest in memory) of the block to be moved. The destination address, in the Y register, should be the new ending address for the moved block. The length, loaded into the double accumulator (the value in C is always used, regardless of the setting of the m flag) should be the length of the block to be moved minus one; if C contains $0005, six bytes will be moved. The two operand bytes of the MVP instruction specify the banks holding the two blocks of memory: the first operand byte (of object code) specifies the destination bank; the second operand byte specifies the source bank.

The execution sequence is: the first byte is moved from the address in X to the address in Y; then X and Y are decremented, C is decremented, and the previous byte is moved; this process continues until the number of bytes specified by the value in C plus one is moved. In other words, until the value in C is $FFFF.

If the source and destination blocks do not overlap, then the source block remains intact after it has been copied to the destination.

If the source and destination blocks do overlap, then MVP should be used only if the destination is higher than the source to avoid overwriting source bytes before they've been copied to the destination. If the destination is lower, then the MVN instruction should be used instead.

When execution is complete, the value in C is SFFFF, registers X and Y each point one byte past the beginning of the blocks to which they were pointing, and the data bank register holds the destination bank value (the first operand byte).

Assembler syntax for the block move instruction calls for the operand field to be coded as two addresses, source first, then destination—the more intuitive ordering, but the opposite of the actual operand order in the object code. The assembler strips the bank bytes from the addresses (ignoring the rest) and reverses them to object code order. If a block move instruction is interrupted, it may be resumed automatically via execution of an RTI if all of the registers are restored or intact. The value pushed onto the stack when a block move is interrupted is the address of the block move instruction. The current byte-move is completed before the interrupt is serviced.

If the index registers are in eight-bit mode (x = 1), or the processor is in 6502 emulation mode (e = 1), then the blocks being specified must necessarily be in page zero since the high bytes of the index registers will contain zeroes.

Flags Affected:

Codes:

* 7 cycles per byte moved

Executing a NOP takes no action; it has no effect on any 65x registers or memory, except the program counter, which is incremented once to point to the next instruction.

Its primary uses are during debugging, where it is used to "patch out" unwanted code, or as a place-holder, included in the assembler source, where you anticipate you may have to "patch in" instructions, and want to leave a "hole" for the patch.

NOP may also be used to expand timing loops—each NOP instruction takes two cycles to execute, so adding one or more may help fine tune a timing loop.

Flags Affected: -------------------

Codes:

Bitwise logical OR the data located at the effective address specified by the operand with the contents of the accumulator. Each bit in the accumulator is ORed with the corresponding bit in memory. The result is stored into the same accumulator bit.

The truth table for the logical OR operation is:

Figure 18.7. Logical OR Truth Table.

A 1 or logical true results if either of the two operands of the OR operation is true.

8-bit accumulator (all processors): Data ORed from memory is eight-bit.

16-bit accumulator (65802/65816 only, m = 0): Data ORed from memory is sixteen-bit: the low-order eight bits are located at the effective address; the high-order eight bits are located at the effective address plus one.

Codes:

^{+ +} ORA, a Primary Group Instruction, has available all of the Primary Group addressing modes and bit patterns

* Add 1 byte if m = 0 (16-bit memory/accumulator)

1 Add 1 cycle if m = 0 (16-bit memory/accumulator)

2 Add 1 cycle if low byte of Direct Page register is other than zero (DL< >0)

3 Add 1 cycle if adding index crosses a page boundary

Push the sixteen-bit operand (typically an absolute address) onto the stack. The stack pointer is decremented twice. This operation always pushes sixteen bits of data, irrespective of the settings of the m and x mode select flags.

Although the mnemonic suggests that the sixteen-bit value pushed on the stack be considered an address, the instruction may also be considered a "push sixteen-bit immediate data" instruction, although the syntax of immediate addressing is not used. The assembler syntax is that of the absolute addressing mode, that is, a label or sixteen-bit value in the operand field. Unlike all other instructions that use this assembler syntax, the effective address itself, rather than the data stored at the effective address, is what is accessed (and in this case, pushed onto the stack).

Flags Affected: -------------------

Codes:

Push the sixteen-bit value located at the address formed by adding the direct page offset specified by the operand to the direct page register. The mnemonic implies that the sixteen-bit data pushed is considered an address, although it can be any sixteen-bit data. This operation always pushes sixteen bits of data, irrespective of the settings of the m and x mode select flags.

The first byte pushed is the byte at the direct page offset plus one (the high byte of the double byte stored at the direct page offset). The byte at the direct page offset itself (the low byte) is pushed next. The stack pointer now points to the next available stack location, directly below the last byte pushed.

The assembler syntax is that of direct page indirect; however, unlike other instructions which use this assembler syntax, the effective indirect address, rather than the data stored at that address, is what is accessed and pushed onto the stack.

Flags Affected: -------------------

Codes:

1 Add 1 cycle if low byte of Direct Page register is other than zero (DL< >0)

Add the current value of the program counter to the sixteen-bit signed displacement in the operand, and push the result on the stack. This operation always pushes sixteen bits of data, irrespective of the settings of the m and x mode select flags.

The high byte of the sum is pushed first, then the low byte is pushed. After the instruction is completed, the stack pointer points to the next available stack location, immediately below the last byte pushed.

Because PER's operand is a displacement relative to the current value of the program counter (as with the branch instructions), this instruction is helpful in writing self-relocatable code in which an address within the program (typically of a data area) must be accessed. The address pushed onto the stack will be the run-time address of the data area, regardless of where the program was loaded in memory; it may be pulled into a register, stored in an indirect pointer, or used on the stack with the stack relative indirect indexed addressing mode to access the data at that location.

As is the case with the branch instructions, the syntax used is to specify as the operand the label of the data area you want to reference. This location must be in the program bank, since the displacement is relative to the program counter. The assembler converts the assembly-time label into a displacement from the assembly-time address of the next instruction.

The value of the program counter used in the addition is the address of the next instruction, that is, the instruction following the PER instruction.

PER may also be used to push return addresses on the stack, either as part of a simulated branch-to-subroutine or to place the return address beneath the stacked parameters to a subroutine call; always remember that a pushed return address should be the desired return addressminus one.

Flags Affected: -------------------

Codes:

Push the accumulator onto the stack. The accumulator itself is unchanged.

8-bit accumulator (all processors): The single byte contents of the accumulator are pushed—they are stored to the location pointed to by the stack pointer and the stack pointer is decremented.

16-bit accumulator (65802/65816 only, m = 0): Both accumulator bytes are pushed. The high byte is pushed first, then the low byte. The stack pointer now points to the next available stack location, directly below the last byte pushed.

Flags Affected: -------------------

Codes:

1 Add 1 cycle if m = 0 (16-bit memory/accumulator)

Push the contents of the data bank register onto the stack.

The single-byte contents of the data bank register are pushed onto the stack; the stack pointer now points to the next available stack location, directly below the byte pushed. The data bank register itself is unchanged. Since the data bank register is an eight-bit register, only one byte is pushed onto the stack, regardless of the settings of the m and x mode select flags.

While the 65816 always generates 24-bit addresses, most memory references are specified by a sixteen-bit address. These addresses are concatenated with the contents of the data bank register to form a full 24-bit address. This instruction lets the current value of the data bank register be saved prior to loading a new value.

Flags Affected: -------------------

Codes:

Push the contents of the direct page register D onto the stack.

Since the direct page register is always a sixteen-bit register, this is always a sixteen-bit operation, regardless of the settings of the m and x mode select flags. The high byte of the direct page register is pushed first, then the low byte. The direct page register itself is unchanged. The stack pointer now points to the next available stack location, directly below the last byte pushed.

By pushing the D register onto the stack, the local environment of a calling subroutine may easily be saved by a called subroutine before modifying the D register to provide itself with its own direct page memory.

Flags Affected: -------------------

Codes:

Push the program bank register onto the stack.

The single-byte contents of the program bank register are pushed. The program bank register itself is unchanged. The stack pointer now points to the next available stack location, directly below the byte pushed. Since the program bank register is an eight-bit register, only one byte is pushed onto the stack, regardless of the settings of the m and x mode select flags.

While the 65816 always generates 24-bit addresses, most jumps and branches specify only a sixteen-bit address. These addresses are concatenated with the contents of the program bank register to form a full 24-bit address. This instruction lets you determine the current value of the program bank register—for example, if you want the data bank to be set to the same value as the program bank.

Flags Affected: -------------------

Codes:

Push the contents of the processor status register P onto the stack.

Since the status register is always an eight-bit register, this is always an eight-bit operation, regardless of the settings of the m and x mode select flags on the 65802/65816. The status register contents are not changed by the operation. The stack pointer now points to the next available stack location, directly below the byte pushed.

This provides the means for saving either the current mode settings or a particular set of status flags so they may be restored or in some other way used later.

Note, however, that the e bit (the 6502 emulation mode flag on the 65802/65816) is not pushed onto the stack or otherwise accessed or saved. The only access to the e flag is via the XCE instruction.

Flags Affected: -------------------

Codes:

Push the contents of the X index register onto the stack. The register itself is unchanged.

8-bit index registers (all processors): The eight-bit contents of the index register are pushed onto the stack. The stack pointer now points to the next available stack location, directly below the byte pushed.

16-bit index registers (65802/65816 only, x = 0): The sixteen-bit contents of the index register are pushed. The high byte is pushed first, then the low byte. The stack pointer now points to the next available stack location, directly below the last byte pushed.

Flags Affected: -------------------

Codes:

1 Add 1 cycle if x = 0 (16-bit index registers)

Push the contents of the Y index register onto the stack. The register itself is unchanged.

8-bit index registers (all processors): The eight-bit contents of the index register are pushed onto the stack. The stack pointer now points to the next available stack location, directly below the byte pushed.

16-bit index registers (65802/65816 only, x = 0): The sixteen-bit contents of the index register are pushed. The high byte is pushed first, then the low byte. The stack pointer now points to the next available stack location, directly below the last byte pushed.

Flags Affected: -------------------

Codes:

1 Add 1 cycle if x = 0 (16-bit index registers)

Pull the value on the top of the stack into the accumulator. The previous contents of the accumulator are destroyed.

8-bit accumulator (all processors): The stack pointer is first incremented. Then the byte pointed to by the stack pointer is loaded into the accumulator.

16-bit accumulator (65802/65816 only, m = 0): Both accumulator bytes are pulled. The accumulator's low byte is pulled first, then the high byte is pulled.

Note that unlike some other microprocessors, the 65x pull instructions set the negative and zero flags.

Codes:

1 Add 1 cycle if m = 0 (16-bit memory/accumulator)

Pull the eight-bit value on top of the stack into the data bank register B, switching the data bank to that value. All instructions which reference data that specify only sixteen-bit addresses will get their bank address from the value pulled into the data bank register. This is the only instruction that can modify the data bank register.

Since the bank register is an eight-bit register, only one byte is pulled from the stack, regardless of the settings of the m and x mode select flags. The stack pointer is first incremented. Then the byte pointed to by the stack pointer is loaded into the register.

Codes:

Pull the sixteen-bit value on top of the stack into the direct page register D, switching the direct page to that value.

PLD is typically used to restore the direct page register to a previous value.

Since the direct page register is a sixteen-bit register, two bytes are pulled from the stack, regardless of the settings of the m and x mode select flags. The low byte of the direct page register is pulled first, then the high byte. The stack pointer now points to where the high byte just pulled was stored; this is now the next available stack location.

Codes:

Pull the eight-bit value on the top of the stack into the processor status register P, switching the status byte to that value.

Since the status register is an eight-bit register, only one byte is pulled from the stack, regardless of the settings of the m and x mode select flags on the 65802/65816. The stack pointer is first incremented. Then the byte pointed to by the stack pointer is loaded into the status register.

This provides the means for restoring either previous mode settings or a particular set of status flags that reflect the result of a previous operation.

Note, however, that the e flag—the 6502 emulation mode flag on the 65802/65816—is not on the stack so cannot be pulled from it. The only means of setting the e flag is the XCE instruction.

Codes:

Pull the value on the top of the stack into the X index register. The previous contents of the register are destroyed.

8-bit index registers (all processors): The stack pointer is first incremented. Then the byte pointed to by the stack pointer is loaded into the register.

16-bit index registers (65802/65816 only, x = 0): Both bytes of the index register are pulled. First the low-order byte of the index register is pulled, then the high-order byte of the index register is pulled.

Unlike some other microprocessors, the 65x instructions to pull an index register affect the negative and zero flags.

Codes:

1 Add 1 cycle if x = 0 (16-bit index registers)

Pull the value on the top of the stack into the Y index register. The previous contents of the register are destroyed.

8-bit index registers (all processors): The stack pointer is first incremented. Then the byte pointed to by the stack pointer is loaded into the register.

16-bit index registers (65802/65816 only, x = 0): Both bytes of the index register are pulled. First the low-order byte of the index register is pulled, then the high-order byte of the index register is pulled.

Unlike some other microprocessors, the 65x instructions to pull an index register affect the negative and zero flags.

Codes:

1 Add 1 cycle if x = 0 (16-bit index registers)

For each bit set to one in the operand byte, reset the corresponding bit in the status register to zero. For example, if bit three is set in the operand byte, bit three in the status register (the decimal flag) is reset to zero by this instruction. Zeroes in the operand byte cause no change to their corresponding status register bits.

This instruction lets you reset any flag or flags in the status register with a single two-byte instruction. Further, it is the only direct means of resetting several of the flags, including the m and x mode select flags (although instructions that pull the P status register affect the mand x mode select flags).

6502 emulation mode (65802/65816, e = 1): Neither the break flag nor bit five (the 6502's undefined flag bit) are affected by REP.

Codes:

Rotate the contents of the location specified by the operand left one bit. Bit one takes on the value originally found in bit zero, bit two takes the value originally in bit one, and so on; the rightmost bit, bit zero, takes the value in the carry flag; the leftmost bit (bit 7 on the 6502 and 65C02 or if m = 1 on the 65802/65816, or bit 15 if m = 0) is transferred into the carry flag.

Figure 18.8. ROL.

8-bit accumulator/memory (all processors): Data rotated is eight bits, plus carry.

16-bit accumulator/memory (65802/65816 only, m = 0): Data rotated is sixteen bits, plus carry: if in memory, the low-order eight bits are located at the effective address; the high eight bits are located at the effective address plus one.

Codes:

1 Add 2 cycles if m = 0 (16-bit memory/accumulator)

2 Add 1 cycle if low byte of Direct Page register is other than zero (DL< >0)

3 Subtract 1 cycle if 65C02 and no page boundary crossed

Rotate the contents of the location specified by the operand right one bit. Bit zero takes on the value originally found in bit one, bit one takes the value originally in bit two, and so on; the leftmost bit (bit 7 on the 6502 and 65C02 or if m = 1 on the 65802/65816, or bit 15 if m = 0) takes the value in the carry flag; the rightmost bit, bit zero, is transferred into the carry flag.

Figure 18.9. ROR.

8-bit accumulator/memory (all processors): Data rotated is eight bits, plus carry.

16-bit accumulator/memory (65802/65816 only, m = 0): Data rotated is sixteen bits, plus carry: if in memory, the low-order eight bits are located at the effective address; the high-order eight bits are located at the effective address plus one.

Codes:

1 Add 2 cycles if m = 0 (16-bit memory/accumulator)

2 Add 1 cycle if low byte of Direct Page register is other than zero (DL< >0)

3 Subtract 1 cycle if 65C02 and no page boundary crossed

Pull the status register and the program counter from the stack. If the 65802/65816 is set to native mode (e = 0), also pull the program bank register from the stack.

RTI pulls values off the stack in the reverse order they were pushed onto it by hardware or software interrupts. The RTI instruction, however, has no way of knowing whether the values pulled off the stack into the status register and the program counter are valid—or even, for that matter, that an interrupt has ever occurred. It blindly pulls the first three (or four) bytes off the top of the stack and stores them into the various registers.

Unlike the RTS instruction, the program counter address pulled off the stack is the exact address to return to; the value on the stack is the value loaded into the program counter. It does not need to be incremented as a subroutine's return address does.

Pulling the status register gives the status flags the values they had immediately prior to the start of interrupt-processing.

One extra byte is pulled in the 65802/65816 native mode than in emulation mode, the same extra byte that is pushed by interrupts in native mode, the program bank register. It is therefore essential that the return from interrupt be executed in the same mode (emulation or native) as the original interrupt.

6502, 65C02, and Emulation Mode (e = 1): The status register is pulled from the stack, then the program counter is pulled from the stack (three bytes are pulled).

65802/65816 Native Mode (e = 0): The status register is pulled from the stack, then the program counter is pulled from the stack, then the program bank register is pulled from the stack (four bytes are pulled).

Figure 18.10. Native Mode Stack before RTI.

Pull the program counter (incrementing the stacked, sixteen-bit value by one before loading the program counter with it), then the program bank register from the stack.

When a subroutine in another bank is called (via a jump to subroutine long instruction), the current bank address is pushed onto the stack along with the return address. To return to the calling bank, a long return instruction must be executed, which first pulls the return address from the stack, increments it, and loads the program counter with it, then pulls the calling bank from the stack and loads the program bank register. This transfers control to the instruction immediately following the original jump to subroutine long.

Figure 18.11. Stack before RTL.

Flags Affected:

Codes:

Pull the program counter, incrementing the stacked, sixteen-bit value by one before loading the program counter with it.

When a subroutine is called (via a jump to subroutine instruction), the current return address is pushed onto the stack. To return to the code following the subroutine call, a return instruction must be executed, which pulls the return address from the stack, increments it, and loads the program counter with it, transferring control to the instruction immediately following the jump to subroutine.

Figure 18.12. Stack before RTS.

Flags Affected: -------------------

Codes:

Subtract the data located at the effective address specified by the operand from the contents of the accumulator; subtract one more if the carry flag is clear, and store the result in the accumulator.

The 65x processors have no subtract instruction that does not involve the carry. To avoid subtracting the carry flag from the result, either you must be sure it is set or you must explicitly set it (using SEC) prior to executing the SBC instruction.

In a multi-precision (multi-word) subtract, you set the carry before the low words are subtracted. The low word subtraction generates a new carry flag value based on the subtraction. The carry is set if no borrow was required and cleared if borrow was required. The complement of the new carry flag (one if the carry is clear) is subtracted during the next subtraction, and so on. Each result thus correctly reflects the borrow from the previous subtraction.

Note that this use of the carry flag is the opposite of the way the borrow flag is used by some other processors, which clear (not set) the carry if no borrow was required.

d flag clear: Binary subtraction is performed.

d flag set: Binary coded decimal (BCD) subtraction is performed.

8-bit accumulator (all processors): Data subtracted from memory is eight-bit.

16-bit accumulator (65802/65816 only, m = 0): Data subtracted from memory is sixteen-bit: the low eight bits is located at the effective address; the high eight bits is located at the effective address plus one.

Codes:

+ + SBC, a Primary Group Instruction, has available all of the Primary Group addressing modes and bit patterns

* Add 1 byte if m = 0 (16-bit memory/accumulator)

1 Add 1 cycle if m = 0 (16-bit memory/accumulator)

2 Add 1 cycle if low byte of Direct Page register is other than zero (DL< >0)

3 Add 1 cycle if adding index crosses a page boundary

4 Add 1 cycle if 65C02 and d = 1 (decimal mode, 65C02)

Set the carry flag in the status register.

SEC is used prior to subtraction (using the 65x's SBC instruction) to keep the carry flag from affecting the result, and prior to an XCE (exchange carry flag with emulation bit) instruction to put the 65802 or 65816 into 6502 emulation mode.

Codes:

Set the decimal mode flag in the status register.

SED is used to shift 65x processors into decimal mode from binary mode, so that the ADC and SBC instructions will operate correctly on BCD data, performing automatic decimal adjustment.

Codes:

Set the interrupt disable flag in the status register.

SEI is used to disable hardware interrupt processing. When the i bit is set, mask-able hardware interrupts (IRQ') are ignored. The processor itself sets the i flag when it begins servicing an interrupt, so interrupt handling routines that are intended to be interruptable must reenable interrupts with CLI. If interrupts are to remain blocked during the interrupt service, exiting the routine via RTI will automatically restore the status register with the i flag clear, re-enabling interrupts.

Codes:

For each one-bit in the operand byte, set the corresponding bit in the status register to one. For example, if bit three is set in the operand byte, bit three in the status register (the decimal flag) is set to one by this instruction. Zeroes in the operand byte cause no change to their corresponding status register bits.

This instruction lets you set any flag or flags in the status register with a single two-byte instruction. Furthermore, it is the only direct means of setting the m and x mode select flags. (Instructions that pull the P status register indirectly affect the m and x mode select flags).

6502 emulation mode (65802/65816, e = 1): Neither the break flag nor bit five (the 6502's non-flag bit) is affected by SEP.

Codes:

Store the value in the accumulator to the effective address specified by the operand.

8-bit accumulator (all processors): Value is eight-bit.

16-bit accumulator (65802/65816 only, m = 0): Value is sixteen-bit: the low-order eight bits are stored to the effective address; the high-order eight bits are stored to the effective address plus one.

The 65x flags are unaffected by store instructions.

Flags Affected: -------------------

Codes:

+ +STA, a Primary Group Instruction, has available all of the Primary Group addressing modes and bit patterns

1 Add 1 cycle if m = 0 (16-bit memory/accumulator)

2 Add 1 cycle if low byte of Direct Page register is other than zero (DL < >0)

During the processor's next phase 2 clock cycle, stop the processor's oscillator input; the processor is effectively shut down until a reset occurs (until the RES' pin is pulled low).

STP is designed to put the processor to sleep while it's not (actively) in use in order to reduce power consumption. Since power consumption is a function of frequency with CMOS circuits, stopping the clock cuts power to almost nil.

Your reset handling routine (pointed to by the reset vector, $00:FFFC-FD) should be designed to either reinitialize the system or resume control through a previously-installed reset handler.

Remember that reset is an interrupt-like signal that causes the emulation bit to be set to one. It also causes the direct page register to be reset to zero; stack high to be set to one (forcing the stack pointer to page one); and the mode select flags to be set to one (eight-bit registers; a side effect is that the high bytes of the index registers are zeroed). STP is useful only in hardware systems (such as battery-powered systems) specifically designed to support a low-power mode.

Flags Affected; -------------------

Codes:

1 Uses 3 cycles to shut the processor down; additional cycles are required by reset to restart it

Store the value in index register X to the effective address specified by the operand.

8-bit index registers (all processors): Value is eight-bit.

16-bit index registers (65802/65816 only, x = 0): Value is sixteen-bit: the low-order eight bits are stored to the effective address; the high-order eight bits are stored to the effective address plus one.

The 65x flags are unaffected by store instructions.

Flags Affected: -------------------

Codes:

1 Add 1 cycle if x = 0 (16-bit index registers)

2 Add 1 cycle if low byte of Direct Page register is other than zero (DL< >0)

Store the value in index register Y to the effective address specified by the operand.

8-bit index registers (all processors): Value is eight-bit.

16-bit index registers (65802/65816 only, x = 0): Value is sixteen-bit: the low-order eight bits are stored to the effective address; the high-order eight bits are stored to the effective address plus one.

The 65x flags are unaffected by store instructions.

Flags Affected: -------------------

Codes:

1 Add 1 cycle if x = 0 (16-bit index registers)

2 Add 1 cycle if low byte of Direct Page register is other than zero (DL< >0)

Store zero to the effective address specified by the operand.

8-bit accumulator (all processors): Zero is stored at the effective address.

16-bit accumulator/memory (65802/65816 only, m = 0): Zero is stored at the effective address and at the effective address plus one.

The 65x store zero instruction does not affect the flags.

Flags Affected: -------------------

Codes:

1 Add 1 cycle if m = 0 (16-bit memory/accumulator)

2 Add 1 cycle if low byte of Direct Page register is other than zero (DL< >0)

Transfer the value in the accumulator to index register X. If the registers are different sizes, the nature of the transfer is determined by the destination register. The value in the accumulator is not changed by the operation.

8-bit accumulator, 8-bit index registers (all processors): Value transferred is eight-bit.

8-bit accumulator, 16-bit index registers (65802/65816 only, m = 1, x = 0): Value transferred is sixteen-bit; the eight-bit A accumulator becomes the low byte of the index register; the hidden eight-bit B accumulator becomes the high byte of the index register.

16-bit accumulator, 8-bit index registers (65802/65816 only, m = 0, x = 1): Value transferred to the eight-bit index register is eight-bit, the low byte of the accumulator.

16-bit accumulator, 16-bit index registers (65802/65816 only, m = 0, x = 0): Value transferred to the sixteen-bit index register is sixteen-bit, the full sixteen-bit accumulator.

Codes:

Transfer the value in the accumulator to index register Y. If the registers are different sizes, the nature of the transfer is determined by the destination register. The value in the accumulator is not changed by the operation.

8-bit accumulator, 8-bit index registers (all processors): Value transferred is eight-bit.

8-bit accumulator, 16-bit index registers (65802/65816 only, m = 1, x = 0): Value transferred is sixteen-bit; the eight-bit A accumulator becomes the low byte of the index register; the hidden eight-bit B accumulator becomes the high byte of the index register.

16-bit accumulator, 8-bit index registers (65802/65816 only, m = 0, x = 1): Value transferred to the eight-bit index register is eight-bit, the low byte of the accumulator.

16-bit accumulator, 16-bit index registers (65802/65816 only, m = 0, x = 0): Value transferred to the sixteen-bit index register is sixteen-bit, the full sixteen-bit accumulator.

Codes:

Transfer the value in the sixteen-bit accumulator C to the direct page register D, regardless of the setting of the accumulator/memory mode flag.

An alternate mnemonic is TAD, (transfer the value in the A accumulator to the direct page register).

In TCD, the "C" is used to indicate that sixteen bits are transferred regardless of the m flag. If the A accumulator is set to just eight bits (whether because the m flag is set, or because the processor is in 6502 emulation mode), then its value becomes the low byte of the direct page register and the value in the hidden B accumulator becomes the high byte of the direct page register.

The accumulator's sixteen-bit value is unchanged by the operation.

Codes:

Transfer the value in the accumulator to the stack pointer S. The accumulator's value is unchanged by the operation.

An alternate mnemonic is TAS (transfer the value in the A accumulator to the stack pointer).

In TCS, the "C" is used to indicate that, in native mode, sixteen bits are transferred regardless of the m flag. If the A accumulator is set to just eight bits (because the m flag is set), then its value is transferred to the low byte of the stack pointer and the value in the hidden Baccumulator is transferred to the high byte of the stack pointer. In emulation mode, only the eight-bit A accumulator is transferred, since the high stack pointer byte is forced to one (the stack is confined to page one).

TCS, along with TXS, are the only two instructions for changing the value in the stack pointer. The two are also the only two transfer instructions not to alter the flags.

Flags Affected: -------------------

Codes:

Transfer the value in the sixteen-bit direct page register D to the sixteen-bit accumulator C, regardless of the setting of the accumulator/memory mode flag.

An alternate mnemonic is TDA (transfer the value in the direct page register to the A accumulator).

In TDC, the "C" is used to indicate that sixteen bits are transferred regardless of the m flag. If the A accumulator is set to just eight bits (whether because the m flag is set, or because the processor is in 6502 emulation mode), then it takes the value of the low byte of the direct page register and the hidden B accumulator takes the value of the high byte of the direct page register.

The direct page register's sixteen-bit value is unchanged by the operation.

Codes:

Logically AND together the complement of the value in the accumulator with the data at the effective address specified by the operand. Store the result at the memory location.

This has the effect of clearing each memory bit for which the corresponding accumulator bit is set, while leaving unchanged all memory bits in which the corresponding accumulator bits are zeroes.

Unlike the BIT instruction, TRB is a read-modify-write instruction, not only calculating a result and modifying a flag, but also storing the result to memory as well.

The z zero flag is set based on a second and different operation, the ANDing of the accumulator value (not its complement) with the memory value (the same way the BIT instruction affects the zero flag). The result of this second operation is not saved; only the zero flag is affected by it.

8-bit accumulator/memory (65C02; 65802/65816, m = 1): Values in accumulator and memory are eight-bit.

16-bit accumulator/memory (65802/65816 only, m = 0): Values in accumulator and memory are sixteen-bit: the low-order eight bits are located at the effective address; the high-order eight bits are at the effective address plus one.

Codes:

1 Add 2 cycles if m = 0 (16-bit memory/accumulator)

2 Add 1 cycle if low byte of Direct Page register is other than zero (DL< >0)

Logically OR together the value in the accumulator with the data at the effective address specified by the operand. Store the result at the memory location.

This has the effect of setting each memory bit for which the corresponding accumulator bit is set, while leaving unchanged all memory bits in which the corresponding accumulator bits are zeroes.

Unlike the BIT instruction, TSB is a read-modify-write instruction, not only calculating a result and modifying a flag, but storing the result to memory as well.

The z zero flag is set based on a second different operation, the ANDing of the accumulator value with the memory value (the same way the BIT instruction affects the zero flag). The result of this second operation is not saved; only the zero flag is affected by it.

8-bit accumulator/memory (65C02; 65802/65816, m = 1): Values in accumulator and memory are eight-bit.

16-bit accumulator/memory (65802/65816 only, m = 0); Values in accumulator and memory are sixteen-bit: the low-order eight bits are located at the effective address; the high-order eight bits are at the effective address plus one.

Codes:

1 Add 2 cycles if m = 0 (16-bit memory/accumulator)

2 Add 1 cycle if low byte of Direct Page register is other than zero (DL< >0)

Transfer the value in the sixteen-bit stack pointer S to the sixteen-bit accumulator C, regardless of the setting of the accumulator/memory mode flag.

An alternate mnemonic is TSA (transfer the value in the stack pointer to the A accumulator).

In TSC, the "C" is used to indicate that sixteen bits are transferred regardless of the m flag. If the A accumulator is set to just eight bits (whether because the m flag is set, or because the processor is in 6502 emulation mode), then it takes the value of the low byte of the stack pointer and the hidden B accumulator takes the value of the high byte of the stack pointer. (In emulation mode, B will always take a value of one, since the stack is confined to page one.)

The stack pointer's value is unchanged by the operation.

Codes:

Transfer the value in the stack pointer S to index register X. The stack pointer's value is not changed by the operation.

8-bit index registers (all processors): Only the low byte of the value in the stack pointer is transferred to the X register. In the 6502, the 65C02, and the 6502 emulation mode, the stack pointer and the index registers are only a single byte each, so the byte in the stack pointer is transferred to the eight-bit X register. In 65802/65816 native mode, the stack pointer is sixteen bits, so its most significant byte is not transferred if the index registers are in eight-bit mode.

16-bit index registers (65802/65816 only, x = 0): The full sixteen-bit value in the stack pointer is transferred to the X register.

Codes:

Transfer the value in index register X to the accumulator. If the registers are different sizes, the nature of the transfer is determined by the destination (the accumulator). The value in the index register is not changed by the operation.

8-bit index registers, 8-bit accumulator (all processors): Value transferred is eight-bit.

16-bit index registers, 8-bit accumulator (65802/65816 only, x = 0, m = 1): Value transferred to the eight-bit accumulator is eight-bit, the low byte of the index register; the hidden eight-bit accumulator B is not affected by the transfer.

8-bit index registers, 16-bit accumulator (65802/65816 only, x = 1, m = 0): The eight-bit index register becomes the low byte of the accumulator; the high accumulator byte is zeroed.

16-bit index registers, 16-bit accumulator (65802/65816 only, x = 0, m = 0):Value transferred to the sixteen-bit accumulator is sixteen-bit, the full sixteen-bit index register.

Codes:

Transfer the value in index register X to the stack pointer, S. The index register's value is not changed by the operation.

TXS, along with TCS, are the only two instructions for changing the value in the stack pointer. The two are also the only two transfer instructions that do not alter the flags.

6502, 65C02, and 6502 emulation mode(65802/65816, e = 1): The stack pointer is only eight bits (it is concatenated to a high byte of one, confining the stack to page one), and the index registers are only eight bits. The byte in X is transferred to the eight-bit stack pointer.

8-bit index registers (65802/65816 native mode, x = 1): The stack pointer is sixteen bits but the index registers are only eight bits. A copy of the byte in X is transferred to the low stack pointer byte and the high stack pointer byte is zeroed.

16-bit index registers (65802/65816 native mode, x = 0): The full sixteen-bit value in X is transferred to the sixteen-bit stack pointer.

Flags Affected: -------------------

Codes:

Transfer the value in index register X to index register Y. The value in index register X is not changed by the operation. Note that the two registers are never different sizes.

8-bit index registers (x = 1): Value transferred is eight-bit.

16-bit index registers (x = 0): Value transferred is sixteen-bit.

Codes:

Transfer the value in index register Y to the accumulator. If the registers are different sizes, the nature of the transfer is determined by the destination (the accumulator). The value in the index register is not changed by the operation.

8-bit index registers, 8-bit accumulator (all processors): Value transferred is eight-bit.

16-bit index registers, 8-bit accumulator (65802/65816 only, x = 0, m = 1): Value transferred to the eight-bit accumulator is eight-bit, the low byte of the index register; the hidden eight-bit accumulator B is not affected by the transfer.

8-bit index registers, 16-bit accumulator (65802/65816 only, x = 1, m = 0): The eight-bit index register becomes the low byte of the accumulator; the high accumulator byte is zeroed.

16-bit index registers, 16-bit accumulator (65802/65816 only, x = 0, m = 0): Value transferred to the sixteen-bit accumulator is sixteen-bit, the full sixteen-bit index register.

Codes:

Transfer the value in index register Y to index register X. The value in index register Y is not changed by the operation. Note that the two registers are never different sizes.

8-bit index registers (x = 1): Value transferred is eight-bit.

16-bit index registers (x = 0): Value transferred is sixteen-bit.

Codes:

Pull the RDY pin low. Power consumption is reduced and RDY remains low until an external hardware interrupt (NMI, IRQ, ABORT, or RESET) is received.

WAI is designed to put the processor to sleep during an external event to reduce its power consumption, to allow it to be synchronized with an external event, and/or to reduce interrupt latency (an interrupt occurring during execution of an instruction is not acted upon until execution of the instruction is complete, perhaps many cycles later; WAI ensures that an interrupt is recognized immediately).

Once an interrupt is received, control is vectored through one of the hardware interrupt vectors; an RTI from the interrupt handling routine will return control to the instruction following the original WAI. However, if by setting the i flag, interrupts have been disabled prior to the execution of the WAI instruction, and IRQ' is asserted, the "wait” condition is terminated and control resumes with the next instruction, rather than through the interrupt vectors. This provides the quickest response to an interrupt, allowing synchronization with external events. WAI also frees up the bus; since RDY is pulled low in the third instruction cycle, the processor may be disconnected from the bus if BE is also pulled low.

Flags Affected: -------------------

Codes:

1 Uses 3 cycles to shut the processor down; additional cycles are required by interrupt to restart it

The 65802 and 65816 use 255 of the 256 possible eight-bit opcodes. One was reserved; it provides an "escape hatch" for future 65x processors to expand their instruction set to sixteen bit opcodes; this opcode would signal that the next byte is an opcode in the expanded instruction set. This reserved byte for future two-byte opcodes was given a temporary mnemonic, WDM, which happen to be the initials of the processors' designer—William D. Mensch, Jr.

WDM should never be used in a program, since it would render the object program incompatible with any future 65x processors.

If the 65802/65816 WDM instruction is accidentally executed, it will act like a two-byte NOP instruction.

Codes;

*Byte and cycle counts subject to change in future processors which expand WDM into 2-byte opcode portions of instructions of varying lengths

B represents the high-order byte of the sixteen-bit C accumulator, and A in this case represents the low-order byte. XBA swaps the contents of the low-order and high-order bytes of C.

An alternate mnemonic is SWA (swap the high and low bytes of the sixteen-bit A accumulator).

XBA can be used to invert the low-order, high-order arrangement of a sixteen-bit value, or to temporarily store an eight-bit value from the A accumulator into B. Since it is an exchange, the previous contents of both accumulators are changed, replaced by the previous contents of the other.

Neither the mode select flags nor the emulation mode flag affects this operation.

The flags are changed based on the new value of the low byte, the A accumulator (that is, on the former value of the high byte, the B accumulator), even in sixteen-bit accumulator mode.

Codes:

This instruction is the only means provided by the 65802 and 65816 to shift between 6502 emulation mode and the full, sixteen-bit native mode.

The emulation mode is used to provide hardware and software compatibility between the 6502 and 65802/65816.

If the processor is in emulation mode, then to switch to native mode, first clear the carry bit, then execute an XCE. Since it is an exchange operation, the carry flag will reflect the previous state of the emulation bit. Switching to native mode causes bit five to stop functioning as the break flag, and function instead as the x mode select flag. A second mode select flag, m, uses bit six, which was unused in emulation mode. Both mode select flags are initially set to one (eight-bit modes). There are also other differences described in the text.

If the processor is in native mode, then to switch to emulation mode, you first set the carry bit, then execute an XCE. Switching to emulation mode causes the mode select flags (m and x) to be lost from the status register, with x replaced by the b break flag. This forces the accumulator to eight bits, but the high accumulator byte is preserved in the hidden B accumulator. It also forces the index registers to eight bits, causing the loss of values in their high bytes, and the stack to page one, causing the loss of the high byte of the previous stack address. There are also other differences described in the text.

Codes: