// // 68000Implementation.hpp // Clock Signal // // Created by Thomas Harte on 10/03/2019. // Copyright © 2019 Thomas Harte. All rights reserved. // #define ccr() \ ( \ (carry_flag_ ? 0x0001 : 0x0000) | \ (overflow_flag_ ? 0x0002 : 0x0000) | \ (zero_result_ ? 0x0000 : 0x0004) | \ (negative_flag_ ? 0x0008 : 0x0000) | \ (extend_flag_ ? 0x0010 : 0x0000) \ ) #define status() \ uint16_t( \ ccr() | \ (interrupt_level_ << 8) | \ (trace_flag_ ? 0x8000 : 0x0000) | \ (is_supervisor_ << 13) \ ) #define apply_ccr(x) \ carry_flag_ = (x) & 0x0001; \ overflow_flag_ = (x) & 0x0002; \ zero_result_ = ((x) & 0x0004) ^ 0x0004; \ negative_flag_ = (x) & 0x0008; \ extend_flag_ = (x) & 0x0010; #define apply_status(x) \ apply_ccr(x) \ interrupt_level_ = ((x) >> 8) & 7; \ trace_flag_ = (x) & 0x8000; \ set_is_supervisor(!!(((x) >> 13) & 1)); #define get_bus_code() \ uint16_t( \ ((active_step_->microcycle.operation & Microcycle::IsProgram) ? 0x02 : 0x01) | \ (is_supervisor_ << 2) | \ (active_program_ ? 0x08 : 0) | \ ((active_step_->microcycle.operation & Microcycle::Read) ? 0x10 : 0) | \ (decoded_instruction_.full & 0xffe0) \ ) #define u_extend16(x) uint32_t(int16_t(x)) #define u_extend8(x) uint32_t(int8_t(x)) #define s_extend16(x) int32_t(int16_t(x)) #define s_extend8(x) int32_t(int8_t(x)) #define convert_to_bit_count_16(x) \ x = ((x & 0xaaaa) >> 1) + (x & 0x5555); \ x = ((x & 0xcccc) >> 2) + (x & 0x3333); \ x = ((x & 0xf0f0) >> 4) + (x & 0x0f0f); \ x = ((x & 0xff00) >> 8) + (x & 0x00ff); // Sets the length of the next microcycle; if this is a debug build, also confirms // that the microcycle being adjusted is the one that it's permissible to adjust. #ifdef NDEBUG #define set_next_microcycle_length(x) \ bus_program->microcycle.length = x #else #define set_next_microcycle_length(x) \ assert(bus_program->microcycle.is_resizeable); \ bus_program->microcycle.length = x #endif template void Processor::run_for(HalfCycles duration) { const HalfCycles remaining_duration = duration + half_cycles_left_to_run_; // This loop counts upwards rather than downwards because it simplifies calculation of // E as and when required. HalfCycles cycles_run_for; while(cycles_run_for < remaining_duration) { /* PERFORM THE CURRENT BUS STEP'S MICROCYCLE. */ switch(execution_state_) { default: case ExecutionState::Executing: // Check for entry into the halted state. if(halt_ && active_step_[0].microcycle.operation & Microcycle::NewAddress) { execution_state_ = ExecutionState::Halted; continue; } if(active_step_->microcycle.data_select_active()) { // Check whether the processor needs to await DTack. if(!dtack_is_implicit && !dtack_ && !bus_error_) { execution_state_ = ExecutionState::WaitingForDTack; dtack_cycle_ = active_step_->microcycle; dtack_cycle_.length = HalfCycles(2); dtack_cycle_.operation &= ~(Microcycle::SelectByte | Microcycle::SelectWord); continue; } // Check for bus error. if(bus_error_ && !is_starting_interrupt_) { const auto offending_address = *active_step_->microcycle.address; active_program_ = nullptr; active_micro_op_ = long_exception_micro_ops_; populate_bus_error_steps(2, status(), get_bus_code(), offending_address); program_counter_.full -= 4; active_step_ = &all_bus_steps_[active_micro_op_->bus_program]; } } // Check for an address error. Which I have assumed happens before the microcycle that // would nominate the new address. if( (active_step_[0].microcycle.operation & Microcycle::NewAddress) && (active_step_[1].microcycle.operation & Microcycle::SelectWord) && *active_step_->microcycle.address & 1) { const auto offending_address = *active_step_->microcycle.address; active_program_ = nullptr; active_micro_op_ = long_exception_micro_ops_; populate_bus_error_steps(3, status(), get_bus_code(), offending_address); program_counter_.full -= 4; active_step_ = &all_bus_steps_[active_micro_op_->bus_program]; // TODO: the above is only correct prior to the final microcycle of an // instruction. If an exception occurs in the final microcycle then // the next instruction will already have moved into the current instruction // slot (decoded_instruction_ in my terms). // TODO: it's also not correct for a bus error that occurs during another exception. } // Perform the microcycle if it is of non-zero length. If this is an operation that // would normally strobe one of the data selects and VPA is active, it will also need // stretching. if(active_step_->microcycle.length != HalfCycles(0)) { if(is_peripheral_address_ && active_step_->microcycle.data_select_active()) { auto cycle_copy = active_step_->microcycle; cycle_copy.operation |= Microcycle::IsPeripheral; // Length will be: (i) distance to next E cycle, plus (ii) difference between // current length and a whole E cycle. const auto phase_now = (e_clock_phase_ + cycles_run_for) % 20; const auto time_to_boundary = (HalfCycles(20) - phase_now) % HalfCycles(20); cycle_copy.length = HalfCycles(20) + time_to_boundary; cycles_run_for += cycle_copy.length + bus_handler_.perform_bus_operation(cycle_copy, is_supervisor_); } else { cycles_run_for += active_step_->microcycle.length + bus_handler_.perform_bus_operation(active_step_->microcycle, is_supervisor_); } } #ifdef LOG_TRACE if(should_log && !(active_step_->microcycle.operation & Microcycle::IsProgram)) { switch(active_step_->microcycle.operation & (Microcycle::SelectWord | Microcycle::SelectByte | Microcycle::Read)) { default: break; case Microcycle::SelectWord | Microcycle::Read: printf("[%08x -> %04x] ", *active_step_->microcycle.address, active_step_->microcycle.value->full); break; case Microcycle::SelectByte | Microcycle::Read: printf("[%08x -> %02x] ", *active_step_->microcycle.address, active_step_->microcycle.value->halves.low); break; case Microcycle::SelectWord: printf("{%04x -> %08x} ", active_step_->microcycle.value->full, *active_step_->microcycle.address); break; case Microcycle::SelectByte: printf("{%02x -> %08x} ", active_step_->microcycle.value->halves.low, *active_step_->microcycle.address); break; } } #endif /* PERFORM THE BUS STEP'S ACTION. */ switch(active_step_->action) { default: std::cerr << "Unimplemented 68000 bus step action: " << int(active_step_->action) << std::endl; return; break; case BusStep::Action::None: break; case BusStep::Action::IncrementEffectiveAddress0: effective_address_[0].full += 2; break; case BusStep::Action::IncrementEffectiveAddress1: effective_address_[1].full += 2; break; case BusStep::Action::DecrementEffectiveAddress0: effective_address_[0].full -= 2; break; case BusStep::Action::DecrementEffectiveAddress1: effective_address_[1].full -= 2; break; case BusStep::Action::IncrementProgramCounter: program_counter_.full += 2; break; case BusStep::Action::IncrementEffectiveAddress0AlignStackPointer: effective_address_[0].full += 2; address_[7].full &= 0xffff'fffe; break; case BusStep::Action::AdvancePrefetch: prefetch_queue_.halves.high = prefetch_queue_.halves.low; // During prefetch advance seems to be the only time the interrupt inputs are sampled; // TODO: determine whether this really happens on *every* advance. if(bus_interrupt_level_ > interrupt_level_) { pending_interrupt_level_ = bus_interrupt_level_; } break; } // Move to the next bus step. ++ active_step_; break; case ExecutionState::Stopped: // If an interrupt (TODO: or reset) has finally arrived that will be serviced, // exit the STOP. if(bus_interrupt_level_ > interrupt_level_) { pending_interrupt_level_ = bus_interrupt_level_; program_counter_.full += 4; // Don't return to this stop. execution_state_ = ExecutionState::WillBeginInterrupt; continue; } // Otherwise continue being stopped. cycles_run_for += stop_cycle_.length + bus_handler_.perform_bus_operation(stop_cycle_, is_supervisor_); continue; case ExecutionState::WaitingForDTack: // If DTack or bus error has been signalled, stop waiting. if(dtack_ || bus_error_) { execution_state_ = ExecutionState::Executing; continue; } // Otherwise, signal another cycle of wait. cycles_run_for += dtack_cycle_.length + bus_handler_.perform_bus_operation(dtack_cycle_, is_supervisor_); continue; case ExecutionState::Halted: if(!halt_) { execution_state_ = ExecutionState::Executing; continue; } cycles_run_for += stop_cycle_.length + bus_handler_.perform_bus_operation(stop_cycle_, is_supervisor_); continue; case ExecutionState::WillBeginInterrupt: #ifdef LOG_TRACE // should_log = true; #endif active_program_ = nullptr; active_micro_op_ = interrupt_micro_ops_; execution_state_ = ExecutionState::Executing; active_step_ = &all_bus_steps_[active_micro_op_->bus_program]; is_starting_interrupt_ = true; break; } /* FIND THE NEXT MICRO-OP IF UNKNOWN. */ if(active_step_->is_terminal()) { while(true) { // If there are any more micro-operations available, just move onwards. if(active_micro_op_ && !active_micro_op_->is_terminal()) { ++active_micro_op_; } else { // Either the micro-operations for this instruction have been exhausted, or // no instruction was ongoing. Either way, do a standard instruction operation. if(pending_interrupt_level_) { execution_state_ = ExecutionState::WillBeginInterrupt; break; } if(trace_flag_ && last_trace_flag_) { // The user has set the trace bit in the status register. active_program_ = nullptr; active_micro_op_ = short_exception_micro_ops_; populate_trap_steps(9, status()); program_counter_.full -= 4; } else { #ifdef LOG_TRACE if(should_log) { std::cout << std::setfill('0'); std::cout << (extend_flag_ ? 'x' : '-') << (negative_flag_ ? 'n' : '-') << (zero_result_ ? '-' : 'z'); std::cout << (overflow_flag_ ? 'v' : '-') << (carry_flag_ ? 'c' : '-') << '\t'; std::cout << (is_supervisor_ ? 's' : 'u') << '\t'; for(int c = 0; c < 8; ++ c) std::cout << "d" << c << ":" << std::setw(8) << data_[c].full << " "; for(int c = 0; c < 8; ++ c) std::cout << "a" << c << ":" << std::setw(8) << address_[c].full << " "; if(is_supervisor_) { std::cout << "usp:" << std::setw(8) << std::setfill('0') << stack_pointers_[0].full << " "; } else { std::cout << "ssp:" << std::setw(8) << std::setfill('0') << stack_pointers_[1].full << " "; } std::cout << '\n'; } #endif decoded_instruction_.full = prefetch_queue_.halves.high.full; #ifndef NDEBUG /* Debugging feature: reset the effective addresses and data latches, so that it's more obvious if some of the instructions aren't properly feeding them. */ effective_address_[0].full = effective_address_[1].full = source_bus_data_.full = destination_bus_data_.full = 0x12344321; #endif #ifdef LOG_TRACE if(should_log) { std::cout << std::hex << (program_counter_.full - 4) << ": " << std::setw(4) << decoded_instruction_.full << '\t'; } #endif if constexpr (signal_will_perform) { bus_handler_.will_perform(program_counter_.full - 4, decoded_instruction_.full); } #ifdef LOG_TRACE // const uint32_t fetched_pc = (program_counter_.full - 4)&0xffffff; // should_log |= fetched_pc == 0x6d9c; // should_log = (fetched_pc >= 0x41806A && fetched_pc <= 0x418618); // should_log |= fetched_pc == 0x4012A2; // should_log &= fetched_pc != 0x4012AE; // should_log = (fetched_pc >= 0x408D66) && (fetched_pc <= 0x408D84); #endif if(instructions[decoded_instruction_.full].micro_operations != std::numeric_limits::max()) { if((instructions[decoded_instruction_.full].source_dest & 0x80) && !is_supervisor_) { // A privilege violation has been detected. active_program_ = nullptr; active_micro_op_ = short_exception_micro_ops_; populate_trap_steps(8, status()); // The location of the failed instruction is what should end up on the stack. program_counter_.full -= 4; } else { // Standard instruction dispatch. active_program_ = &instructions[decoded_instruction_.full]; active_micro_op_ = &all_micro_ops_[active_program_->micro_operations]; } } else { // The opcode fetched isn't valid. active_program_ = nullptr; active_micro_op_ = short_exception_micro_ops_; // The location of the failed instruction is what should end up on the stack. program_counter_.full -= 4; #ifdef LOG_TRACE // should_log = true; #endif // The vector used depends on whether this is a vanilla unrecognised instruction, // or one on the A or F lines. switch(decoded_instruction_.full >> 12) { default: populate_trap_steps(4, status()); break; case 0xa: populate_trap_steps(10, status()); break; case 0xf: populate_trap_steps(11, status()); break; } } } last_trace_flag_ = trace_flag_; } auto bus_program = &all_bus_steps_[active_micro_op_->bus_program]; using int_type = decltype(active_micro_op_->action); switch(active_micro_op_->action) { default: std::cerr << "Unhandled 68000 micro op action " << std::hex << active_micro_op_->action << " within instruction " << decoded_instruction_.full << std::endl; break; case int_type(MicroOp::Action::None): break; #define offset_pointer(x) reinterpret_cast(&reinterpret_cast(static_cast(this))[x]) #define source() offset_pointer(active_program_->source_offset) #define source_address() address_[(active_program_->source_dest >> 4) & 7] #define destination() offset_pointer(active_program_->destination_offset) #define destination_address() address_[active_program_->source_dest & 7] case int_type(MicroOp::Action::PerformOperation): #define sub_overflow() ((result ^ destination) & (destination ^ source)) #define add_overflow() ((result ^ destination) & ~(destination ^ source)) switch(active_program_->operation) { /* ABCD adds the lowest bytes form the source and destination using BCD arithmetic, obeying the extend flag. */ case Operation::ABCD: { // Pull out the two halves, for simplicity. const uint8_t source = source()->halves.low.halves.low; const uint8_t destination = destination()->halves.low.halves.low; // Perform the BCD add by evaluating the two nibbles separately. const int unadjusted_result = destination + source + (extend_flag_ ? 1 : 0); int result = (destination & 0xf) + (source & 0xf) + (extend_flag_ ? 1 : 0); if(result > 0x09) result += 0x06; result += (destination & 0xf0) + (source & 0xf0); if(result > 0x99) result += 0x60; // Set all flags essentially as if this were normal addition. zero_result_ |= result & 0xff; extend_flag_ = carry_flag_ = uint_fast32_t(result & ~0xff); negative_flag_ = result & 0x80; overflow_flag_ = ~unadjusted_result & result & 0x80; // Store the result. destination()->halves.low.halves.low = uint8_t(result); } break; // ADD and ADDA add two quantities, the latter sign extending and without setting any flags; // ADDQ and SUBQ act as ADD and SUB, but taking the second argument from the instruction code. #define addop(a, b, x) a + b + (x ? 1 : 0) #define subop(a, b, x) a - b - (x ? 1 : 0) #define z_set(a, b) a = b #define z_or(a, b) a |= b #define addsubb(a, b, dest, op, overflow, x, zero_op) \ const int source = a; \ const int destination = b; \ const auto result = op(destination, source, x); \ \ dest = uint8_t(result); \ zero_op(zero_result_, dest); \ extend_flag_ = carry_flag_ = uint_fast32_t(result & ~0xff); \ negative_flag_ = result & 0x80; \ overflow_flag_ = overflow() & 0x80; #define addsubw(a, b, dest, op, overflow, x, zero_op) \ const int source = a; \ const int destination = b; \ const auto result = op(destination, source, x); \ \ dest = uint16_t(result); \ zero_op(zero_result_, dest); \ extend_flag_ = carry_flag_ = uint_fast32_t(result & ~0xffff); \ negative_flag_ = result & 0x8000; \ overflow_flag_ = overflow() & 0x8000; #define addsubl(a, b, dest, op, overflow, x, zero_op) \ const uint64_t source = a; \ const uint64_t destination = b; \ const auto result = op(destination, source, x); \ \ dest = uint32_t(result); \ zero_op(zero_result_, dest); \ extend_flag_ = carry_flag_ = uint_fast32_t(result >> 32); \ negative_flag_ = result & 0x80000000; \ overflow_flag_ = overflow() & 0x80000000; #define addb(a, b, dest, x, z) addsubb(a, b, dest, addop, add_overflow, x, z) #define subb(a, b, dest, x, z) addsubb(a, b, dest, subop, sub_overflow, x, z) #define addw(a, b, dest, x, z) addsubw(a, b, dest, addop, add_overflow, x, z) #define subw(a, b, dest, x, z) addsubw(a, b, dest, subop, sub_overflow, x, z) #define addl(a, b, dest, x, z) addsubl(a, b, dest, addop, add_overflow, x, z) #define subl(a, b, dest, x, z) addsubl(a, b, dest, subop, sub_overflow, x, z) #define no_extend(op, a, b, c) op(a, b, c, 0, z_set) #define extend(op, a, b, c) op(a, b, c, extend_flag_, z_or) #define q() (((decoded_instruction_.full >> 9)&7) ? ((decoded_instruction_.full >> 9)&7) : 8) case Operation::ADDb: { no_extend( addb, source()->halves.low.halves.low, destination()->halves.low.halves.low, destination()->halves.low.halves.low); } break; case Operation::ADDXb: { extend( addb, source()->halves.low.halves.low, destination()->halves.low.halves.low, destination()->halves.low.halves.low); } break; case Operation::ADDQb: { no_extend( addb, q(), destination()->halves.low.halves.low, destination()->halves.low.halves.low); } break; case Operation::ADDw: { no_extend( addw, source()->halves.low.full, destination()->halves.low.full, destination()->halves.low.full); } break; case Operation::ADDXw: { extend( addw, source()->halves.low.full, destination()->halves.low.full, destination()->halves.low.full); } break; case Operation::ADDQw: { no_extend( addw, q(), destination()->halves.low.full, destination()->halves.low.full); } break; case Operation::ADDl: { no_extend( addl, source()->full, destination()->full, destination()->full); } break; case Operation::ADDXl: { extend( addl, source()->full, destination()->full, destination()->full); } break; case Operation::ADDQl: { no_extend( addl, q(), destination()->full, destination()->full); } break; case Operation::SUBb: { no_extend( subb, source()->halves.low.halves.low, destination()->halves.low.halves.low, destination()->halves.low.halves.low); } break; case Operation::SUBXb: { extend( subb, source()->halves.low.halves.low, destination()->halves.low.halves.low, destination()->halves.low.halves.low); } break; case Operation::SUBQb: { no_extend( subb, q(), destination()->halves.low.halves.low, destination()->halves.low.halves.low); } break; case Operation::SUBw: { no_extend( subw, source()->halves.low.full, destination()->halves.low.full, destination()->halves.low.full); } break; case Operation::SUBXw: { extend( subw, source()->halves.low.full, destination()->halves.low.full, destination()->halves.low.full); } break; case Operation::SUBQw: { no_extend( subw, q(), destination()->halves.low.full, destination()->halves.low.full); } break; case Operation::SUBl: { no_extend( subl, source()->full, destination()->full, destination()->full); } break; case Operation::SUBXl: { extend( subl, source()->full, destination()->full, destination()->full); } break; case Operation::SUBQl: { no_extend( subl, q(), destination()->full, destination()->full); } break; case Operation::ADDQAl: destination()->full += q(); break; case Operation::SUBQAl: destination()->full -= q(); break; #undef addl #undef addw #undef addb #undef subl #undef subw #undef subb #undef addsubl #undef addsubw #undef addsubb #undef q #undef z_set #undef z_or #undef no_extend #undef extend #undef addop #undef subop case Operation::ADDAw: destination()->full += u_extend16(source()->halves.low.full); break; case Operation::ADDAl: destination()->full += source()->full; break; case Operation::SUBAw: destination()->full -= u_extend16(source()->halves.low.full); break; case Operation::SUBAl: destination()->full -= source()->full; break; // BRA: alters the program counter, exclusively via the prefetch queue. case Operation::BRA: { const int8_t byte_offset = int8_t(prefetch_queue_.halves.high.halves.low); // A non-zero offset byte branches by just that amount; otherwise use the word // after as an offset. In both cases, treat as signed. if(byte_offset) { program_counter_.full += uint32_t(byte_offset); } else { program_counter_.full += u_extend16(prefetch_queue_.halves.low.full); } program_counter_.full -= 2; } break; // Two BTSTs: set the zero flag according to the value of the destination masked by // the bit named in the source modulo the operation size. case Operation::BTSTb: zero_result_ = destination()->full & (1 << (source()->full & 7)); break; case Operation::BTSTl: zero_result_ = destination()->full & (1 << (source()->full & 31)); break; case Operation::BCLRb: zero_result_ = destination()->full & (1 << (source()->full & 7)); destination()->full &= ~(1 << (source()->full & 7)); break; case Operation::BCLRl: zero_result_ = destination()->full & (1 << (source()->full & 31)); destination()->full &= ~(1 << (source()->full & 31)); // Clearing in the top word requires an extra four cycles. set_next_microcycle_length(HalfCycles(8 + ((source()->full & 31) / 16) * 4)); break; case Operation::BCHGl: zero_result_ = destination()->full & (1 << (source()->full & 31)); destination()->full ^= 1 << (source()->full & 31); set_next_microcycle_length(HalfCycles(4 + (((source()->full & 31) / 16) * 4))); break; case Operation::BCHGb: zero_result_ = destination()->halves.low.halves.low & (1 << (source()->full & 7)); destination()->halves.low.halves.low ^= 1 << (source()->full & 7); break; case Operation::BSETl: zero_result_ = destination()->full & (1 << (source()->full & 31)); destination()->full |= 1 << (source()->full & 31); set_next_microcycle_length(HalfCycles(4 + (((source()->full & 31) / 16) * 4))); break; case Operation::BSETb: zero_result_ = destination()->halves.low.halves.low & (1 << (source()->full & 7)); destination()->halves.low.halves.low |= 1 << (source()->full & 7); break; // Bcc: ordinarily evaluates the relevant condition and displacement size and then: // if condition is false, schedules bus operations to get past this instruction; // otherwise applies the offset and schedules bus operations to refill the prefetch queue. // // Special case: the condition code is 1, which is ordinarily false. In that case this // is the trailing step of a BSR. case Operation::Bcc: { // Grab the 8-bit offset. const int8_t byte_offset = int8_t(prefetch_queue_.halves.high.halves.low); // Check whether this is secretly BSR. const bool is_bsr = ((decoded_instruction_.full >> 8) & 0xf) == 1; // Test the conditional, treating 'false' as true. const bool should_branch = is_bsr || evaluate_condition(decoded_instruction_.full >> 8); // Schedule something appropriate, by rewriting the program for this instruction temporarily. if(should_branch) { if(byte_offset) { program_counter_.full += decltype(program_counter_.full)(byte_offset); } else { program_counter_.full += u_extend16(prefetch_queue_.halves.low.full); } program_counter_.full -= 2; bus_program = is_bsr ? bsr_bus_steps_ : branch_taken_bus_steps_; } else { if(byte_offset) { bus_program = branch_byte_not_taken_bus_steps_; } else { bus_program = branch_word_not_taken_bus_steps_; } } } break; case Operation::DBcc: { // Decide what sort of DBcc this is. if(!evaluate_condition(decoded_instruction_.full >> 8)) { -- source()->halves.low.full; const auto target_program_counter = program_counter_.full + u_extend16(prefetch_queue_.halves.low.full) - 2; if(source()->halves.low.full == 0xffff) { // This DBcc will be ignored as the counter has underflowed. // Schedule n np np np and continue. Assumed: the first np // is from where the branch would have been if taken? bus_program = dbcc_condition_false_no_branch_steps_; dbcc_false_address_ = target_program_counter; } else { // Take the branch. Change PC and schedule n np np. bus_program = dbcc_condition_false_branch_steps_; program_counter_.full = target_program_counter; } } else { // This DBcc will be ignored as the condition is true; // perform nn np np and continue. bus_program = dbcc_condition_true_steps_; } } break; case Operation::Scc: { destination()->halves.low.halves.low = evaluate_condition(decoded_instruction_.full >> 8) ? 0xff : 0x00; } break; /* CLRs: store 0 to the destination, set the zero flag, and clear negative, overflow and carry. */ case Operation::CLRb: destination()->halves.low.halves.low = 0; negative_flag_ = overflow_flag_ = carry_flag_ = zero_result_ = 0; break; case Operation::CLRw: destination()->halves.low.full = 0; negative_flag_ = overflow_flag_ = carry_flag_ = zero_result_ = 0; break; case Operation::CLRl: destination()->full = 0; negative_flag_ = overflow_flag_ = carry_flag_ = zero_result_ = 0; break; /* CMP.b, CMP.l and CMP.w: sets the condition flags (other than extend) based on a subtraction of the source from the destination; the result of the subtraction is not stored. */ case Operation::CMPb: { const uint8_t source = source()->halves.low.halves.low; const uint8_t destination = destination()->halves.low.halves.low; const int result = destination - source; zero_result_ = result & 0xff; carry_flag_ = decltype(carry_flag_)(result & ~0xff); negative_flag_ = result & 0x80; overflow_flag_ = sub_overflow() & 0x80; } break; case Operation::CMPw: { const uint16_t source = source()->halves.low.full; const uint16_t destination = destination()->halves.low.full; const int result = destination - source; zero_result_ = result & 0xffff; carry_flag_ = decltype(carry_flag_)(result & ~0xffff); negative_flag_ = result & 0x8000; overflow_flag_ = sub_overflow() & 0x8000; } break; case Operation::CMPAw: { const auto source = uint64_t(u_extend16(source()->halves.low.full)); const uint64_t destination = destination()->full; const auto result = destination - source; zero_result_ = uint32_t(result); carry_flag_ = result >> 32; negative_flag_ = result & 0x80000000; overflow_flag_ = sub_overflow() & 0x80000000; } break; case Operation::CMPl: { const auto source = uint64_t(source()->full); const auto destination = uint64_t(destination()->full); const auto result = destination - source; zero_result_ = uint32_t(result); carry_flag_ = result >> 32; negative_flag_ = result & 0x80000000; overflow_flag_ = sub_overflow() & 0x80000000; } break; // JMP: copies EA(0) to the program counter. case Operation::JMP: program_counter_ = effective_address_[0]; break; // JMP: copies the source bus data to the program counter. case Operation::RTS: program_counter_ = source_bus_data_; break; /* MOVE.b, MOVE.l and MOVE.w: move the least significant byte or word, or the entire long word, and set negative, zero, overflow and carry as appropriate. */ case Operation::MOVEb: zero_result_ = destination()->halves.low.halves.low = source()->halves.low.halves.low; negative_flag_ = zero_result_ & 0x80; overflow_flag_ = carry_flag_ = 0; break; case Operation::MOVEw: zero_result_ = destination()->halves.low.full = source()->halves.low.full; negative_flag_ = zero_result_ & 0x8000; overflow_flag_ = carry_flag_ = 0; break; case Operation::MOVEl: zero_result_ = destination()->full = source()->full; negative_flag_ = zero_result_ & 0x80000000; overflow_flag_ = carry_flag_ = 0; break; /* MOVE.q: a single byte is moved from the current instruction, and sign extended. */ case Operation::MOVEq: zero_result_ = destination()->full = prefetch_queue_.halves.high.halves.low; negative_flag_ = zero_result_ & 0x80; overflow_flag_ = carry_flag_ = 0; destination()->full |= negative_flag_ ? 0xffffff00 : 0; break; /* MOVEA.l: move the entire long word; MOVEA.w: move the least significant word and sign extend it. Neither sets any flags. */ case Operation::MOVEAw: destination()->halves.low.full = source()->halves.low.full; destination()->halves.high.full = (destination()->halves.low.full & 0x8000) ? 0xffff : 0; break; case Operation::MOVEAl: destination()->full = source()->full; break; case Operation::PEA: destination_bus_data_ = effective_address_[0]; break; /* Status word moves and manipulations. */ case Operation::MOVEtoSR: apply_status(source()->full); break; case Operation::MOVEfromSR: destination()->halves.low.full = status(); break; case Operation::MOVEtoCCR: apply_ccr(source()->full); break; case Operation::EXTbtow: destination()->halves.low.halves.high = (destination()->halves.low.halves.low & 0x80) ? 0xff : 0x00; overflow_flag_ = carry_flag_ = 0; zero_result_ = destination()->halves.low.full; negative_flag_ = zero_result_ & 0x8000; break; case Operation::EXTwtol: destination()->halves.high.full = (destination()->halves.low.full & 0x8000) ? 0xffff : 0x0000; overflow_flag_ = carry_flag_ = 0; zero_result_ = destination()->full; negative_flag_ = zero_result_ & 0x80000000; break; #define and_op(a, b) a &= b #define or_op(a, b) a |= b #define eor_op(a, b) a ^= b #define apply(op, func) {\ auto status = status(); \ op(status, prefetch_queue_.halves.high.full); \ func(status); \ program_counter_.full -= 2; \ } #define apply_op_sr(op) apply(op, apply_status) #define apply_op_ccr(op) apply(op, apply_ccr) case Operation::ANDItoSR: apply_op_sr(and_op); break; case Operation::EORItoSR: apply_op_sr(eor_op); break; case Operation::ORItoSR: apply_op_sr(or_op); break; case Operation::ANDItoCCR: apply_op_ccr(and_op); break; case Operation::EORItoCCR: apply_op_ccr(eor_op); break; case Operation::ORItoCCR: apply_op_ccr(or_op); break; #undef apply_op_ccr #undef apply_op_sr #undef apply #undef eor_op #undef or_op #undef and_op /* Multiplications. */ case Operation::MULU: { destination()->full = destination()->halves.low.full * source()->halves.low.full; carry_flag_ = overflow_flag_ = 0; zero_result_ = destination()->full; negative_flag_ = zero_result_ & 0x80000000; int number_of_ones = source()->halves.low.full; convert_to_bit_count_16(number_of_ones); // Time taken = 38 cycles + 2 cycles per 1 in the source. set_next_microcycle_length(HalfCycles(4 * number_of_ones + 34*2)); } break; case Operation::MULS: { destination()->full = u_extend16(destination()->halves.low.full) * u_extend16(source()->halves.low.full); carry_flag_ = overflow_flag_ = 0; zero_result_ = destination()->full; negative_flag_ = zero_result_ & 0x80000000; // Find the number of 01 or 10 pairs in the 17-bit number // formed by the source value with a 0 suffix. int number_of_pairs = source()->halves.low.full; number_of_pairs = (number_of_pairs ^ (number_of_pairs << 1)) & 0xffff; convert_to_bit_count_16(number_of_pairs); // Time taken = 38 cycles + 2 cycles per 1 in the source. set_next_microcycle_length(HalfCycles(4 * number_of_pairs + 34*2)); } break; /* Divisions. */ #define announce_divide_by_zero() \ negative_flag_ = overflow_flag_ = 0; \ zero_result_ = 1; \ active_program_ = nullptr; \ active_micro_op_ = short_exception_micro_ops_; \ bus_program = &all_bus_steps_[active_micro_op_->bus_program]; \ \ populate_trap_steps(5, status()); \ bus_program->microcycle.length = HalfCycles(20); \ \ program_counter_.full -= 6; case Operation::DIVU: { carry_flag_ = 0; // An attempt to divide by zero schedules an exception. if(!source()->halves.low.full) { // Schedule a divide-by-zero exception. announce_divide_by_zero(); break; } uint32_t dividend = destination()->full; uint32_t divisor = source()->halves.low.full; const auto quotient = dividend / divisor; // If overflow would occur, appropriate flags are set and the result is not written back. if(quotient > 65535) { overflow_flag_ = zero_result_ = negative_flag_ = 1; set_next_microcycle_length(HalfCycles(3*2*2)); break; } const uint16_t remainder = uint16_t(dividend % divisor); destination()->halves.high.full = remainder; destination()->halves.low.full = uint16_t(quotient); overflow_flag_ = 0; zero_result_ = quotient; negative_flag_ = zero_result_ & 0x8000; // Calculate cost; this is based on the flowchart in yacht.txt. // I could actually calculate the division result here, since this is // a classic divide algorithm, but would rather that errors produce // incorrect timing only, not incorrect timing plus incorrect results. int cycles_expended = 12; // Covers the nn n to get into the loop. divisor <<= 16; for(int c = 0; c < 15; ++c) { if(dividend & 0x80000000) { dividend = (dividend << 1) - divisor; cycles_expended += 4; // Easy; just the fixed nn iteration cost. } else { dividend <<= 1; // Yacht.txt, and indeed a real microprogram, would just subtract here // and test the sign of the result, but this is easier to follow: if (dividend >= divisor) { dividend -= divisor; cycles_expended += 6; // i.e. the original nn plus one further n before going down the MSB=0 route. } else { cycles_expended += 8; // The costliest path (since in real life it's a subtraction and then a step // back from there) — all costs accrue. So the fixed nn loop plus another n, // plus another one. } } } set_next_microcycle_length(HalfCycles(cycles_expended * 2)); } break; case Operation::DIVS: { carry_flag_ = 0; // An attempt to divide by zero schedules an exception. if(!source()->halves.low.full) { // Schedule a divide-by-zero exception. announce_divide_by_zero() break; } const int32_t signed_dividend = int32_t(destination()->full); const int32_t signed_divisor = s_extend16(source()->halves.low.full); const auto result_sign = ( (0 <= signed_dividend) - (signed_dividend < 0) ) * ( (0 <= signed_divisor) - (signed_divisor < 0) ); const uint32_t dividend = uint32_t(abs(signed_dividend)); const uint32_t divisor = uint32_t(abs(signed_divisor)); int cycles_expended = 12; // Covers the nn nnn n to get beyond the sign test. if(signed_dividend < 0) { cycles_expended += 2; // An additional microycle applies if the dividend is negative. } // Check for overflow. If it exists, work here is already done. const auto quotient = dividend / divisor; if(quotient > 32767) { overflow_flag_ = 1; set_next_microcycle_length(HalfCycles(6*2*2)); break; } const uint16_t remainder = uint16_t(signed_dividend % signed_divisor); const int signed_quotient = result_sign*int(quotient); destination()->halves.high.full = remainder; destination()->halves.low.full = uint16_t(signed_quotient); zero_result_ = decltype(zero_result_)(signed_quotient); negative_flag_ = zero_result_ & 0x8000; overflow_flag_ = 0; // Algorithm here: there is a fixed cost per unset bit // in the first 15 bits of the unsigned quotient. auto positive_quotient_bits = ~quotient & 0xfffe; convert_to_bit_count_16(positive_quotient_bits); cycles_expended += 2 * positive_quotient_bits; // There's then no way to terminate the loop that isn't at least ten cycles long; // there's also a fixed overhead per bit. The two together add up to the 104 below. cycles_expended += 104; // This picks up at 'No more bits' in yacht.txt's diagram. if(signed_divisor < 0) { cycles_expended += 2; } else if(signed_dividend < 0) { cycles_expended += 4; } set_next_microcycle_length(HalfCycles(cycles_expended * 2)); } break; #undef announce_divide_by_zero /* MOVEP: move words and long-words a byte at a time. */ case Operation::MOVEPtoMw: // Write pattern is nW+ nw, which should write the low word of the source in big-endian form. destination_bus_data_.halves.high.full = source()->halves.low.halves.high; destination_bus_data_.halves.low.full = source()->halves.low.halves.low; break; case Operation::MOVEPtoMl: // Write pattern is nW+ nWr+ nw+ nwr, which should write the source in big-endian form. destination_bus_data_.halves.high.full = source()->halves.high.halves.high; source_bus_data_.halves.high.full = source()->halves.high.halves.low; destination_bus_data_.halves.low.full = source()->halves.low.halves.high; source_bus_data_.halves.low.full = source()->halves.low.halves.low; break; case Operation::MOVEPtoRw: // Read pattern is nRd+ nrd. source()->halves.low.halves.high = destination_bus_data_.halves.high.halves.low; source()->halves.low.halves.low = destination_bus_data_.halves.low.halves.low; break; case Operation::MOVEPtoRl: // Read pattern is nRd+ nR+ nrd+ nr. source()->halves.high.halves.high = destination_bus_data_.halves.high.halves.low; source()->halves.high.halves.low = source_bus_data_.halves.high.halves.low; source()->halves.low.halves.high = destination_bus_data_.halves.low.halves.low; source()->halves.low.halves.low = source_bus_data_.halves.low.halves.low; break; /* MOVEM: multi-word moves. */ #define setup_movem(words_per_reg, base) \ /* Count the number of long words to move. */ \ size_t total_to_move = 0; \ auto mask = next_word_; \ while(mask) { \ total_to_move += mask&1; \ mask >>= 1; \ } \ \ /* Twice that many words plus one will need to be moved */ \ bus_program = base + (64 - total_to_move*words_per_reg)*2; \ \ /* Fill in the proper addresses and targets. */ \ const auto mode = (decoded_instruction_.full >> 3) & 7; \ uint32_t start_address; \ if(mode <= 4) { \ start_address = destination_address().full; \ } else { \ start_address = effective_address_[1].full; \ } \ \ auto step = bus_program; \ uint32_t *address_storage = precomputed_addresses_; \ mask = next_word_; \ int offset = 0; #define inc_action(x, v) x += v #define dec_action(x, v) x -= v #define write_address_sequence_long(action, l) \ while(mask) { \ if(mask&1) { \ address_storage[0] = start_address; \ action(start_address, 2); \ address_storage[1] = start_address; \ action(start_address, 2); \ \ step[0].microcycle.address = step[1].microcycle.address = address_storage; \ step[2].microcycle.address = step[3].microcycle.address = address_storage + 1; \ \ const auto target = (offset > 7) ? &address_[offset&7] : &data_[offset]; \ step[l].microcycle.value = step[l+1].microcycle.value = &target->halves.high; \ step[(l^2)].microcycle.value = step[(l^2)+1].microcycle.value = &target->halves.low; \ \ address_storage += 2; \ step += 4; \ } \ mask >>= 1; \ action(offset, 1); \ } #define write_address_sequence_word(action) \ while(mask) { \ if(mask&1) { \ address_storage[0] = start_address; \ action(start_address, 2); \ \ step[0].microcycle.address = step[1].microcycle.address = address_storage; \ \ const auto target = (offset > 7) ? &address_[offset&7] : &data_[offset]; \ step[0].microcycle.value = step[1].microcycle.value = &target->halves.low; \ \ ++ address_storage; \ step += 2; \ } \ mask >>= 1; \ action(offset, 1); \ } case Operation::MOVEMtoRl: { setup_movem(2, movem_read_steps_); // Everything for move to registers is based on an incrementing // address; per M68000PRM: // // "[If using the postincrement addressing mode then] the incremented address // register contains the address of the last operand loaded plus the operand length. // If the addressing register is also loaded from memory, the memory value is ignored // and the register is written with the postincremented effective address." // // The latter part is dealt with by MicroOp::Action::MOVEMtoRComplete, which also // does any necessary sign extension. write_address_sequence_long(inc_action, 0); // MOVEM to R always reads one word too many. address_storage[0] = start_address; step[0].microcycle.address = step[1].microcycle.address = address_storage; step[0].microcycle.value = step[1].microcycle.value = &throwaway_value_; movem_final_address_ = start_address; } break; case Operation::MOVEMtoRw: { setup_movem(1, movem_read_steps_); write_address_sequence_word(inc_action); // MOVEM to R always reads one word too many. address_storage[0] = start_address; step[0].microcycle.address = step[1].microcycle.address = address_storage; step[0].microcycle.value = step[1].microcycle.value = &throwaway_value_; movem_final_address_ = start_address; } break; case Operation::MOVEMtoMl: { setup_movem(2, movem_write_steps_); // MOVEM to M counts downwards and enumerates the registers in reverse order // if subject to the predecrementing mode; otherwise it counts upwards and // operates exactly as does MOVEM to R. // // Note also: "The MC68000 and MC68010 write the initial register value // (not decremented) [when writing a register that is providing // pre-decrementing addressing]." // // Hence the decrementing register (if any) is updated // by MicroOp::Action::MOVEMtoMComplete. if(mode == 4) { offset = 15; start_address -= 2; write_address_sequence_long(dec_action, 2); movem_final_address_ = start_address + 2; } else { write_address_sequence_long(inc_action, 0); } } break; case Operation::MOVEMtoMw: { setup_movem(1, movem_write_steps_); if(mode == 4) { offset = 15; start_address -= 2; write_address_sequence_word(dec_action); movem_final_address_ = start_address + 2; } else { write_address_sequence_word(inc_action); } } break; #undef setup_movem #undef write_address_sequence_long #undef write_address_sequence_word #undef inc_action #undef dec_action // TRAP, which is a nicer form of ILLEGAL. case Operation::TRAP: { // Select the trap steps as next; the initial microcycle should be 4 cycles long. bus_program = trap_steps_; populate_trap_steps((decoded_instruction_.full & 15) + 32, status()); set_next_microcycle_length(HalfCycles(12)); // The program counter to push is actually one slot ago. program_counter_.full -= 2; } break; case Operation::TRAPV: { if(overflow_flag_) { // Select the trap steps as next; the initial microcycle should be skipped. bus_program = trap_steps_; populate_trap_steps(7, status()); set_next_microcycle_length(HalfCycles(4)); // Push the address after the TRAPV. program_counter_.full -= 4; } } break; case Operation::CHK: { const bool is_under = s_extend16(destination()->halves.low.full) < 0; const bool is_over = s_extend16(destination()->halves.low.full) > s_extend16(source()->halves.low.full); overflow_flag_ = carry_flag_ = 0; zero_result_ = destination()->halves.low.full; // Test applied for N: // // if Dn < 0, set negative flag; // otherwise, if Dn > , reset negative flag. if(is_over) negative_flag_ = 0; if(is_under) negative_flag_ = 1; // No exception is the default course of action; deviate only if an // exception is necessary. if(is_under || is_over) { bus_program = trap_steps_; populate_trap_steps(6, status()); if(is_over) { set_next_microcycle_length(HalfCycles(20)); } else { set_next_microcycle_length(HalfCycles(24)); } // The program counter to push is two slots ago as whatever was the correct prefetch // to continue without an exception has already happened, just in case. program_counter_.full -= 4; } } break; /* NEGs: negatives the destination, setting the zero, negative, overflow and carry flags appropriate, and extend. NB: since the same logic as SUB is used to calculate overflow, and SUB calculates `destination - source`, the NEGs deliberately label 'source' and 'destination' differently from Motorola. */ case Operation::NEGb: { const int destination = 0; const int source = destination()->halves.low.halves.low; const auto result = destination - source; destination()->halves.low.halves.low = uint8_t(result); zero_result_ = result & 0xff; extend_flag_ = carry_flag_ = decltype(carry_flag_)(result & ~0xff); negative_flag_ = result & 0x80; overflow_flag_ = sub_overflow() & 0x80; } break; case Operation::NEGw: { const int destination = 0; const int source = destination()->halves.low.full; const auto result = destination - source; destination()->halves.low.full = uint16_t(result); zero_result_ = result & 0xffff; extend_flag_ = carry_flag_ = decltype(carry_flag_)(result & ~0xffff); negative_flag_ = result & 0x8000; overflow_flag_ = sub_overflow() & 0x8000; } break; case Operation::NEGl: { const uint64_t destination = 0; const uint64_t source = destination()->full; const auto result = destination - source; destination()->full = uint32_t(result); zero_result_ = uint_fast32_t(result); extend_flag_ = carry_flag_ = result >> 32; negative_flag_ = result & 0x80000000; overflow_flag_ = sub_overflow() & 0x80000000; } break; /* NEGXs: NEG, with extend. */ case Operation::NEGXb: { const int source = destination()->halves.low.halves.low; const int destination = 0; const auto result = destination - source - (extend_flag_ ? 1 : 0); destination()->halves.low.halves.low = uint8_t(result); zero_result_ |= result & 0xff; extend_flag_ = carry_flag_ = decltype(carry_flag_)(result & ~0xff); negative_flag_ = result & 0x80; overflow_flag_ = sub_overflow() & 0x80; } break; case Operation::NEGXw: { const int source = destination()->halves.low.full; const int destination = 0; const auto result = destination - source - (extend_flag_ ? 1 : 0); destination()->halves.low.full = uint16_t(result); zero_result_ |= result & 0xffff; extend_flag_ = carry_flag_ = decltype(carry_flag_)(result & ~0xffff); negative_flag_ = result & 0x8000; overflow_flag_ = sub_overflow() & 0x8000; } break; case Operation::NEGXl: { const uint64_t source = destination()->full; const uint64_t destination = 0; const auto result = destination - source - (extend_flag_ ? 1 : 0); destination()->full = uint32_t(result); zero_result_ |= uint_fast32_t(result); extend_flag_ = carry_flag_ = result >> 32; negative_flag_ = result & 0x80000000; overflow_flag_ = sub_overflow() & 0x80000000; } break; /* The no-op. */ case Operation::None: break; /* LINK and UNLINK help with stack frames, allowing a certain amount of stack space to be allocated or deallocated. */ case Operation::LINK: // Make space for the new long-word value, and set up // the proper target address for the stack operations to follow. address_[7].full -= 4; effective_address_[1].full = address_[7].full; // The current value of the address register will be pushed. destination_bus_data_.full = source()->full; // The address register will then contain the bottom of the stack, // and the stack pointer will be offset. source()->full = address_[7].full; address_[7].full += u_extend16(prefetch_queue_.halves.low.full); break; case Operation::UNLINK: address_[7].full = effective_address_[1].full + 2; destination()->full = destination_bus_data_.full; break; /* TAS: sets zero and negative depending on the current value of the destination, and sets the high bit. */ case Operation::TAS: overflow_flag_ = carry_flag_ = 0; zero_result_ = destination()->halves.low.halves.low; negative_flag_ = destination()->halves.low.halves.low & 0x80; destination()->halves.low.halves.low |= 0x80; break; /* Bitwise operators: AND, OR and EOR. All three clear the overflow and carry flags, and set zero and negative appropriately. */ #define op_and(x, y) x &= y #define op_or(x, y) x |= y #define op_eor(x, y) x ^= y #define bitwise(source, dest, sign_mask, operator) \ operator(dest, source); \ overflow_flag_ = carry_flag_ = 0; \ zero_result_ = dest; \ negative_flag_ = dest & sign_mask; #define andx(source, dest, sign_mask) bitwise(source, dest, sign_mask, op_and) #define eorx(source, dest, sign_mask) bitwise(source, dest, sign_mask, op_eor) #define orx(source, dest, sign_mask) bitwise(source, dest, sign_mask, op_or) #define op_bwl(name, op) \ case Operation::name##b: op(source()->halves.low.halves.low, destination()->halves.low.halves.low, 0x80); break; \ case Operation::name##w: op(source()->halves.low.full, destination()->halves.low.full, 0x8000); break; \ case Operation::name##l: op(source()->full, destination()->full, 0x80000000); break; op_bwl(AND, andx); op_bwl(EOR, eorx); op_bwl(OR, orx); #undef op_bwl #undef orx #undef eorx #undef andx #undef bitwise #undef op_eor #undef op_or #undef op_and // NOTs: take the logical inverse, affecting the negative and zero flags. case Operation::NOTb: destination()->halves.low.halves.low ^= 0xff; zero_result_ = destination()->halves.low.halves.low; negative_flag_ = zero_result_ & 0x80; overflow_flag_ = carry_flag_ = 0; break; case Operation::NOTw: destination()->halves.low.full ^= 0xffff; zero_result_ = destination()->halves.low.full; negative_flag_ = zero_result_ & 0x8000; overflow_flag_ = carry_flag_ = 0; break; case Operation::NOTl: destination()->full ^= 0xffffffff; zero_result_ = destination()->full; negative_flag_ = zero_result_ & 0x80000000; overflow_flag_ = carry_flag_ = 0; break; #define sbcd() \ /* Perform the BCD arithmetic by evaluating the two nibbles separately. */ \ const int unadjusted_result = destination - source - (extend_flag_ ? 1 : 0); \ int result = (destination & 0xf) - (source & 0xf) - (extend_flag_ ? 1 : 0); \ if((result & 0x1f) > 0x09) result -= 0x06; \ result += (destination & 0xf0) - (source & 0xf0); \ extend_flag_ = carry_flag_ = decltype(carry_flag_)((result & 0x1ff) > 0x99); \ if(carry_flag_) result -= 0x60; \ \ /* Set all flags essentially as if this were normal subtraction. */ \ zero_result_ |= result & 0xff; \ negative_flag_ = result & 0x80; \ overflow_flag_ = unadjusted_result & ~result & 0x80; \ \ /* Store the result. */ \ destination()->halves.low.halves.low = uint8_t(result); /* SBCD subtracts the lowest byte of the source from that of the destination using BCD arithmetic, obeying the extend flag. */ case Operation::SBCD: { const uint8_t source = source()->halves.low.halves.low; const uint8_t destination = destination()->halves.low.halves.low; sbcd(); } break; /* NBCD is like SBCD except that the result is 0 - destination rather than destination - source. */ case Operation::NBCD: { const uint8_t source = destination()->halves.low.halves.low; const uint8_t destination = 0; sbcd(); } break; // EXG and SWAP exchange/swap words or long words. case Operation::EXG: { const auto temporary = source()->full; source()->full = destination()->full; destination()->full = temporary; } break; case Operation::SWAP: { const auto temporary = destination()->halves.low.full; destination()->halves.low.full = destination()->halves.high.full; destination()->halves.high.full = temporary; zero_result_ = destination()->full; negative_flag_ = temporary & 0x8000; overflow_flag_ = carry_flag_ = 0; } break; /* Shifts and rotates. */ #define set_neg_zero(v, m) \ zero_result_ = decltype(zero_result_)(v); \ negative_flag_ = zero_result_ & decltype(negative_flag_)(m); #define set_neg_zero_overflow(v, m) \ set_neg_zero(v, m); \ overflow_flag_ = (decltype(zero_result_)(value) ^ zero_result_) & decltype(overflow_flag_)(m); #define decode_shift_count() \ int shift_count = (decoded_instruction_.full & 32) ? data_[(decoded_instruction_.full >> 9) & 7].full&63 : ( ((decoded_instruction_.full >> 9)&7) ? ((decoded_instruction_.full >> 9)&7) : 8) ; \ set_next_microcycle_length(HalfCycles(4 * shift_count)); #define set_flags_b(t) set_flags(destination()->halves.low.halves.low, 0x80, t) #define set_flags_w(t) set_flags(destination()->halves.low.full, 0x8000, t) #define set_flags_l(t) set_flags(destination()->full, 0x80000000, t) #define asl(destination, size) {\ decode_shift_count(); \ const auto value = destination; \ \ if(!shift_count) { \ carry_flag_ = overflow_flag_ = 0; \ } else { \ destination = (shift_count < size) ? decltype(destination)(value << shift_count) : 0; \ extend_flag_ = carry_flag_ = decltype(carry_flag_)(value) & decltype(carry_flag_)( (1u << (size - 1)) >> (shift_count - 1) ); \ \ if(shift_count >= size) overflow_flag_ = value && (value != decltype(value)(-1)); \ else { \ const auto mask = decltype(destination)(0xffffffff << (size - shift_count)); \ overflow_flag_ = mask & value && ((mask & value) != mask); \ } \ } \ \ set_neg_zero(destination, 1 << (size - 1)); \ } case Operation::ASLm: { const auto value = destination()->halves.low.full; destination()->halves.low.full = uint16_t(value << 1); extend_flag_ = carry_flag_ = value & 0x8000; set_neg_zero_overflow(destination()->halves.low.full, 0x8000); } break; case Operation::ASLb: asl(destination()->halves.low.halves.low, 8); break; case Operation::ASLw: asl(destination()->halves.low.full, 16); break; case Operation::ASLl: asl(destination()->full, 32); break; #define asr(destination, size) {\ decode_shift_count(); \ const auto value = destination; \ \ if(!shift_count) { \ carry_flag_ = 0; \ } else { \ destination = (shift_count < size) ? \ decltype(destination)(\ (value >> shift_count) | \ ((value & decltype(value)(1 << (size - 1)) ? 0xffffffff : 0x000000000) << (size - shift_count)) \ ) : \ decltype(destination)( \ (value & decltype(value)(1 << (size - 1))) ? 0xffffffff : 0x000000000 \ ); \ extend_flag_ = carry_flag_ = decltype(carry_flag_)(value) & decltype(carry_flag_)(1 << (shift_count - 1)); \ } \ \ set_neg_zero_overflow(destination, 1 << (size - 1)); \ } case Operation::ASRm: { const auto value = destination()->halves.low.full; destination()->halves.low.full = (value&0x8000) | (value >> 1); extend_flag_ = carry_flag_ = value & 1; set_neg_zero_overflow(destination()->halves.low.full, 0x8000); } break; case Operation::ASRb: asr(destination()->halves.low.halves.low, 8); break; case Operation::ASRw: asr(destination()->halves.low.full, 16); break; case Operation::ASRl: asr(destination()->full, 32); break; #undef set_neg_zero_overflow #define set_neg_zero_overflow(v, m) \ set_neg_zero(v, m); \ overflow_flag_ = 0; #undef set_flags #define set_flags(v, m, t) \ zero_result_ = v; \ negative_flag_ = zero_result_ & (m); \ overflow_flag_ = 0; \ carry_flag_ = value & (t); #define lsl(destination, size) {\ decode_shift_count(); \ const auto value = destination; \ \ if(!shift_count) { \ carry_flag_ = 0; \ } else { \ destination = (shift_count < size) ? decltype(destination)(value << shift_count) : 0; \ extend_flag_ = carry_flag_ = decltype(carry_flag_)(value) & decltype(carry_flag_)( (1u << (size - 1)) >> (shift_count - 1) ); \ } \ \ set_neg_zero_overflow(destination, 1 << (size - 1)); \ } case Operation::LSLm: { const auto value = destination()->halves.low.full; destination()->halves.low.full = uint16_t(value << 1); extend_flag_ = carry_flag_ = value & 0x8000; set_neg_zero_overflow(destination()->halves.low.full, 0x8000); } break; case Operation::LSLb: lsl(destination()->halves.low.halves.low, 8); break; case Operation::LSLw: lsl(destination()->halves.low.full, 16); break; case Operation::LSLl: lsl(destination()->full, 32); break; #define lsr(destination, size) {\ decode_shift_count(); \ const auto value = destination; \ \ if(!shift_count) { \ carry_flag_ = 0; \ } else { \ destination = (shift_count < size) ? (value >> shift_count) : 0; \ extend_flag_ = carry_flag_ = value & decltype(carry_flag_)(1 << (shift_count - 1)); \ } \ \ set_neg_zero_overflow(destination, 1 << (size - 1)); \ } case Operation::LSRm: { const auto value = destination()->halves.low.full; destination()->halves.low.full = value >> 1; extend_flag_ = carry_flag_ = value & 1; set_neg_zero_overflow(destination()->halves.low.full, 0x8000); } break; case Operation::LSRb: lsr(destination()->halves.low.halves.low, 8); break; case Operation::LSRw: lsr(destination()->halves.low.full, 16); break; case Operation::LSRl: lsr(destination()->full, 32); break; #define rol(destination, size) { \ decode_shift_count(); \ const auto value = destination; \ \ if(!shift_count) { \ carry_flag_ = 0; \ } else { \ shift_count &= (size - 1); \ destination = decltype(destination)( \ (value << shift_count) | \ (value >> (size - shift_count)) \ ); \ carry_flag_ = decltype(carry_flag_)(destination & 1); \ } \ \ set_neg_zero_overflow(destination, 1 << (size - 1)); \ } case Operation::ROLm: { const auto value = destination()->halves.low.full; destination()->halves.low.full = uint16_t((value << 1) | (value >> 15)); carry_flag_ = destination()->halves.low.full & 1; set_neg_zero_overflow(destination()->halves.low.full, 0x8000); } break; case Operation::ROLb: rol(destination()->halves.low.halves.low, 8); break; case Operation::ROLw: rol(destination()->halves.low.full, 16); break; case Operation::ROLl: rol(destination()->full, 32); break; #define ror(destination, size) { \ decode_shift_count(); \ const auto value = destination; \ \ if(!shift_count) { \ carry_flag_ = 0; \ } else { \ shift_count &= (size - 1); \ destination = decltype(destination)(\ (value >> shift_count) | \ (value << (size - shift_count)) \ );\ carry_flag_ = destination & decltype(carry_flag_)(1 << (size - 1)); \ } \ \ set_neg_zero_overflow(destination, 1 << (size - 1)); \ } case Operation::RORm: { const auto value = destination()->halves.low.full; destination()->halves.low.full = uint16_t((value >> 1) | (value << 15)); carry_flag_ = destination()->halves.low.full & 0x8000; set_neg_zero_overflow(destination()->halves.low.full, 0x8000); } break; case Operation::RORb: ror(destination()->halves.low.halves.low, 8); break; case Operation::RORw: ror(destination()->halves.low.full, 16); break; case Operation::RORl: ror(destination()->full, 32); break; #define roxl(destination, size) { \ decode_shift_count(); \ \ shift_count %= (size + 1); \ uint64_t compound = uint64_t(destination) | (extend_flag_ ? (1ull << size) : 0); \ compound = \ (compound << shift_count) | \ (compound >> (size + 1 - shift_count)); \ carry_flag_ = extend_flag_ = decltype(carry_flag_)((compound >> size) & 1); \ destination = decltype(destination)(compound); \ \ set_neg_zero_overflow(destination, 1 << (size - 1)); \ } case Operation::ROXLm: { const auto value = destination()->halves.low.full; destination()->halves.low.full = uint16_t((value << 1) | (extend_flag_ ? 0x0001 : 0x0000)); extend_flag_ = value & 0x8000; set_flags_w(0x8000); } break; case Operation::ROXLb: roxl(destination()->halves.low.halves.low, 8); break; case Operation::ROXLw: roxl(destination()->halves.low.full, 16); break; case Operation::ROXLl: roxl(destination()->full, 32); break; #define roxr(destination, size) { \ decode_shift_count(); \ \ shift_count %= (size + 1); \ uint64_t compound = uint64_t(destination) | (extend_flag_ ? (1ull << size) : 0); \ compound = \ (compound >> shift_count) | \ (compound << (size + 1 - shift_count)); \ carry_flag_ = extend_flag_ = decltype(carry_flag_)((compound >> size) & 1); \ destination = decltype(destination)(compound); \ \ set_neg_zero_overflow(destination, 1 << (size - 1)); \ } case Operation::ROXRm: { const auto value = destination()->halves.low.full; destination()->halves.low.full = (value >> 1) | (extend_flag_ ? 0x8000 : 0x0000); extend_flag_ = value & 0x0001; set_flags_w(0x0001); } break; case Operation::ROXRb: roxr(destination()->halves.low.halves.low, 8); break; case Operation::ROXRw: roxr(destination()->halves.low.full, 16); break; case Operation::ROXRl: roxr(destination()->full, 32); break; #undef roxr #undef roxl #undef ror #undef rol #undef asr #undef lsr #undef lsl #undef asl #undef set_flags #undef decode_shift_count #undef set_flags_b #undef set_flags_w #undef set_flags_l #undef set_neg_zero_overflow #undef set_net_zero /* RTE and RTR share an implementation. */ case Operation::RTE_RTR: // If this is RTR, patch out the supervisor half of the status register. if(decoded_instruction_.full == 0x4e77) { const auto current_status = status(); source_bus_data_.halves.low.halves.high = uint8_t(current_status >> 8); } apply_status(source_bus_data_.full); break; /* TSTs: compare to zero. */ case Operation::TSTb: carry_flag_ = overflow_flag_ = 0; zero_result_ = source()->halves.low.halves.low; negative_flag_ = zero_result_ & 0x80; break; case Operation::TSTw: carry_flag_ = overflow_flag_ = 0; zero_result_ = source()->halves.low.full; negative_flag_ = zero_result_ & 0x8000; break; case Operation::TSTl: carry_flag_ = overflow_flag_ = 0; zero_result_ = source()->full; negative_flag_ = zero_result_ & 0x80000000; break; case Operation::STOP: apply_status(prefetch_queue_.halves.low.full); execution_state_ = ExecutionState::Stopped; break; /* Development period debugging. */ default: std::cerr << "Should do something with program operation " << int(active_program_->operation) << std::endl; break; } #undef sub_overflow #undef add_overflow break; case int_type(MicroOp::Action::MOVEMtoRComplete): { // If this was a word-sized move, perform sign extension. if(active_program_->operation == Operation::MOVEMtoRw) { auto mask = next_word_; int offset = 0; while(mask) { if(mask&1) { const auto target = (offset > 7) ? &address_[offset&7] : &data_[offset]; target->halves.high.full = (target->halves.low.full & 0x8000) ? 0xffff : 0x0000; } mask >>= 1; ++offset; } } // If the post-increment mode was used, overwrite the source register. const auto mode = (decoded_instruction_.full >> 3) & 7; if(mode == 3) { const auto reg = decoded_instruction_.full & 7; address_[reg] = movem_final_address_; } } break; case int_type(MicroOp::Action::MOVEMtoMComplete): { const auto mode = (decoded_instruction_.full >> 3) & 7; if(mode == 4) { const auto reg = decoded_instruction_.full & 7; address_[reg] = movem_final_address_; } } break; case int_type(MicroOp::Action::PrepareJSR): { const auto mode = (decoded_instruction_.full >> 3) & 7; // Determine the proper resumption address. switch(mode) { case 2: destination_bus_data_.full = program_counter_.full - 2; break; /* (An) */ default: destination_bus_data_.full = program_counter_.full; /* Everything other than (An) */ break; } address_[7].full -= 4; effective_address_[1].full = address_[7].full; } break; case int_type(MicroOp::Action::PrepareBSR): destination_bus_data_.full = (decoded_instruction_.full & 0xff) ? program_counter_.full - 2 : program_counter_.full; address_[7].full -= 4; effective_address_[1].full = address_[7].full; break; case int_type(MicroOp::Action::PrepareRTS): effective_address_[0].full = address_[7].full; address_[7].full += 4; break; case int_type(MicroOp::Action::PrepareRTE_RTR): precomputed_addresses_[0] = address_[7].full + 2; precomputed_addresses_[1] = address_[7].full; precomputed_addresses_[2] = address_[7].full + 4; address_[7].full += 6; break; case int_type(MicroOp::Action::PrepareINT): // The INT sequence uses the same storage as the TRAP steps, so this'll get // the necessary stack work set up. populate_trap_steps(0, status()); // Mutate neessary internal state — effective_address_[0] is exposed // on the data bus as the accepted interrupt number during the interrupt // acknowledge cycle, with all other bits set, including the low bit as // a real 68000 uses the lower data strobe to collect the corresponding vector byte. // // Cf. M68000 8-/16-/32-BIT MICROPROCESSORS USER'S MANUAL 5.1.4. accepted_interrupt_level_ = interrupt_level_ = pending_interrupt_level_; pending_interrupt_level_ = 0; effective_address_[0].full = 0xfffffff1 | uint32_t(accepted_interrupt_level_ << 1); // Recede the program counter to where it would have been were there no // prefetch; that's where the reading stream should pick up upon RTE. program_counter_.full -= 4; break; case int_type(MicroOp::Action::PrepareINTVector): // Let bus error go back to causing exceptions. is_starting_interrupt_ = false; // Bus error => spurious interrupt. if(bus_error_) { effective_address_[0].full = 24 << 2; break; } // Valid peripheral address => autovectored interrupt. if(is_peripheral_address_) { effective_address_[0].full = uint32_t(24 + accepted_interrupt_level_) << 2; break; } // Otherwise, the vector is whatever we were just told it is. effective_address_[0].full = uint32_t(source_bus_data_.halves.low.halves.low << 2); // printf("Interrupt vector: %06x\n", effective_address_[0].full); break; case int_type(MicroOp::Action::CopyNextWord): next_word_ = prefetch_queue_.halves.low.full; break; // Increments and decrements. #define op_add(x, y) x += y #define op_sub(x, y) x -= y #define Adjust(op, quantity, effect) \ case int_type(op) | MicroOp::SourceMask: effect(source_address().full, quantity); break; \ case int_type(op) | MicroOp::DestinationMask: effect(destination_address().full, quantity); break; \ case int_type(op) | MicroOp::SourceMask | MicroOp::DestinationMask: \ effect(destination_address().full, quantity); \ effect(source_address().full, quantity); \ break; Adjust(MicroOp::Action::Decrement1, 1, op_sub); Adjust(MicroOp::Action::Decrement2, 2, op_sub); Adjust(MicroOp::Action::Decrement4, 4, op_sub); Adjust(MicroOp::Action::Increment1, 1, op_add); Adjust(MicroOp::Action::Increment2, 2, op_add); Adjust(MicroOp::Action::Increment4, 4, op_add); #undef Adjust #undef op_add #undef op_sub case int_type(MicroOp::Action::SignExtendWord): if(active_micro_op_->action & MicroOp::SourceMask) { source()->halves.high.full = (source()->halves.low.full & 0x8000) ? 0xffff : 0x0000; } if(active_micro_op_->action & MicroOp::DestinationMask) { destination()->halves.high.full = (destination()->halves.low.full & 0x8000) ? 0xffff : 0x0000; } break; case int_type(MicroOp::Action::SignExtendByte): if(active_micro_op_->action & MicroOp::SourceMask) { source()->full = (source()->full & 0xff) | (source()->full & 0x80) ? 0xffffff : 0x000000; } if(active_micro_op_->action & MicroOp::DestinationMask) { destination()->full = (destination()->full & 0xff) | (destination()->full & 0x80) ? 0xffffff : 0x000000; } break; // 16-bit offset addressing modes. case int_type(MicroOp::Action::CalcD16PC) | MicroOp::SourceMask: // The address the low part of the prefetch queue was read from was two bytes ago, hence // the subtraction of 2. effective_address_[0] = u_extend16(prefetch_queue_.halves.low.full) + program_counter_.full - 2; break; case int_type(MicroOp::Action::CalcD16PC) | MicroOp::DestinationMask: effective_address_[1] = u_extend16(prefetch_queue_.halves.low.full) + program_counter_.full - 2; break; case int_type(MicroOp::Action::CalcD16PC) | MicroOp::SourceMask | MicroOp::DestinationMask: // Similar logic applies here to above, but the high part of the prefetch queue was four bytes // ago rather than merely two. effective_address_[0] = u_extend16(prefetch_queue_.halves.high.full) + program_counter_.full - 4; effective_address_[1] = u_extend16(prefetch_queue_.halves.low.full) + program_counter_.full - 2; break; case int_type(MicroOp::Action::CalcD16An) | MicroOp::SourceMask: effective_address_[0] = u_extend16(prefetch_queue_.halves.low.full) + source_address().full; break; case int_type(MicroOp::Action::CalcD16An) | MicroOp::DestinationMask: effective_address_[1] = u_extend16(prefetch_queue_.halves.low.full) + destination_address().full; break; case int_type(MicroOp::Action::CalcD16An) | MicroOp::SourceMask | MicroOp::DestinationMask: effective_address_[0] = u_extend16(prefetch_queue_.halves.high.full) + source_address().full; effective_address_[1] = u_extend16(prefetch_queue_.halves.low.full) + destination_address().full; break; #define CalculateD8AnXn(data, source, target) {\ const auto register_index = (data.full >> 12) & 7; \ const RegisterPair32 &displacement = (data.full & 0x8000) ? address_[register_index] : data_[register_index]; \ target.full = u_extend8(data.halves.low) + source; \ \ if(data.full & 0x800) { \ target.full += displacement.full; \ } else { \ target.full += u_extend16(displacement.halves.low.full); \ } \ } case int_type(MicroOp::Action::CalcD8AnXn) | MicroOp::SourceMask: { CalculateD8AnXn(prefetch_queue_.halves.low, source_address().full, effective_address_[0]); } break; case int_type(MicroOp::Action::CalcD8AnXn) | MicroOp::DestinationMask: { CalculateD8AnXn(prefetch_queue_.halves.low, destination_address().full, effective_address_[1]); } break; case int_type(MicroOp::Action::CalcD8AnXn) | MicroOp::SourceMask | MicroOp::DestinationMask: { CalculateD8AnXn(prefetch_queue_.halves.high, source_address().full, effective_address_[0]); CalculateD8AnXn(prefetch_queue_.halves.low, destination_address().full, effective_address_[1]); } break; case int_type(MicroOp::Action::CalcD8PCXn) | MicroOp::SourceMask: { CalculateD8AnXn(prefetch_queue_.halves.low, program_counter_.full - 2, effective_address_[0]); } break; case int_type(MicroOp::Action::CalcD8PCXn) | MicroOp::DestinationMask: { CalculateD8AnXn(prefetch_queue_.halves.low, program_counter_.full - 2, effective_address_[1]); } break; case int_type(MicroOp::Action::CalcD8PCXn) | MicroOp::SourceMask | MicroOp::DestinationMask: { CalculateD8AnXn(prefetch_queue_.halves.high, program_counter_.full - 4, effective_address_[0]); CalculateD8AnXn(prefetch_queue_.halves.low, program_counter_.full - 2, effective_address_[1]); } break; #undef CalculateD8AnXn case int_type(MicroOp::Action::AssembleWordAddressFromPrefetch) | MicroOp::SourceMask: effective_address_[0] = u_extend16(prefetch_queue_.halves.low.full); break; case int_type(MicroOp::Action::AssembleWordAddressFromPrefetch) | MicroOp::DestinationMask: effective_address_[1] = u_extend16(prefetch_queue_.halves.low.full); break; case int_type(MicroOp::Action::AssembleLongWordAddressFromPrefetch) | MicroOp::SourceMask: effective_address_[0] = prefetch_queue_.full; break; case int_type(MicroOp::Action::AssembleLongWordAddressFromPrefetch) | MicroOp::DestinationMask: effective_address_[1] = prefetch_queue_.full; break; case int_type(MicroOp::Action::AssembleWordDataFromPrefetch) | MicroOp::SourceMask: source_bus_data_ = prefetch_queue_.halves.low.full; break; case int_type(MicroOp::Action::AssembleWordDataFromPrefetch) | MicroOp::DestinationMask: destination_bus_data_ = prefetch_queue_.halves.low.full; break; case int_type(MicroOp::Action::AssembleLongWordDataFromPrefetch) | MicroOp::SourceMask: source_bus_data_ = prefetch_queue_.full; break; case int_type(MicroOp::Action::AssembleLongWordDataFromPrefetch) | MicroOp::DestinationMask: destination_bus_data_ = prefetch_queue_.full; break; case int_type(MicroOp::Action::CopyToEffectiveAddress) | MicroOp::SourceMask: effective_address_[0] = source_address(); break; case int_type(MicroOp::Action::CopyToEffectiveAddress) | MicroOp::DestinationMask: effective_address_[1] = destination_address(); break; case int_type(MicroOp::Action::CopyToEffectiveAddress) | MicroOp::SourceMask | MicroOp::DestinationMask: effective_address_[0] = source_address(); effective_address_[1] = destination_address(); break; } // If we've got to a micro-op that includes bus steps, break out of this loop. if(!active_micro_op_->is_terminal()) { active_step_ = bus_program; if(!active_step_->is_terminal()) break; } } } } #undef source #undef source_address #undef destination #undef destination_address bus_handler_.flush(); e_clock_phase_ = (e_clock_phase_ + cycles_run_for) % 20; half_cycles_left_to_run_ = remaining_duration - cycles_run_for; } template ProcessorState Processor::get_state() { write_back_stack_pointer(); State state; memcpy(state.data, data_, sizeof(state.data)); memcpy(state.address, address_, sizeof(state.address)); state.user_stack_pointer = stack_pointers_[0].full; state.supervisor_stack_pointer = stack_pointers_[1].full; state.program_counter = program_counter_.full; state.status = status(); return state; } template void Processor::set_state(const ProcessorState &state) { memcpy(data_, state.data, sizeof(state.data)); memcpy(address_, state.address, sizeof(state.address)); apply_status(state.status); stack_pointers_[0].full = state.user_stack_pointer; stack_pointers_[1].full = state.supervisor_stack_pointer; address_[7] = stack_pointers_[is_supervisor_]; } uint16_t ProcessorStorage::get_status() const { return status(); } void ProcessorStorage::set_status(uint16_t status) { apply_status(status); } #undef status #undef apply_status #undef apply_ccr #undef ccr #undef u_extend16 #undef u_extend8 #undef s_extend16 #undef s_extend8 #undef set_next_microcycle_length #undef convert_to_bit_count_16