diff options
Diffstat (limited to 'sux.h')
-rw-r--r-- | sux.h | 338 |
1 files changed, 138 insertions, 200 deletions
@@ -25,9 +25,37 @@ extern uint8_t subdbg; extern uint8_t step; extern uint8_t esc; +static inline uint8_t get_opsize(uint8_t prefix, uint8_t addrmode); + +enum sw_type {RS, AM, BYTE}; + #define setflag(flag, bit) ((flag)) ? (cpu->ps.u8[thread] |= bit) : (cpu->ps.u8[thread] &= ~bit) #define getflag(bit) (cpu->ps.u8[thread] & bit) +/* reg_expr, and val_expr are the arithmetic expressions + * that will be used for either the value, or register, such as '+', or '-'. + */ +#define setreg(reg, reg_expr, reg_idx, val, val_expr, val_idx, size) {\ + switch (size) {\ + case 7: reg[(reg_idx) reg_expr 7] = val[(val_idx) val_expr 7];\ + case 6: reg[(reg_idx) reg_expr 6] = val[(val_idx) val_expr 6];\ + case 5: reg[(reg_idx) reg_expr 5] = val[(val_idx) val_expr 5];\ + case 4: reg[(reg_idx) reg_expr 4] = val[(val_idx) val_expr 4];\ + case 3: reg[(reg_idx) reg_expr 3] = val[(val_idx) val_expr 3];\ + case 2: reg[(reg_idx) reg_expr 2] = val[(val_idx) val_expr 2];\ + case 1: reg[(reg_idx) reg_expr 1] = val[(val_idx) val_expr 1];\ + case 0: reg[(reg_idx) reg_expr 0] = val[(val_idx) val_expr 0];\ + }\ +} + +#define setreg_sw(reg, reg_idx, val, val_idx, prefix, addrmode, type) {\ + switch (type) {\ + case RS: setreg(reg, +, reg_idx, val, +, val_idx, (1 << (prefix >> 4))-1); break;\ + case AM: setreg(reg, +, reg_idx, val, +, val_idx, get_opsize(prefix, addrmode)); break;\ + }\ +} + + extern pthread_mutex_t mutex; extern pthread_mutex_t main_mutex; extern pthread_cond_t cond; @@ -39,6 +67,29 @@ extern void disasm(struct sux *cpu, uint64_t *operands, uint8_t lines, uint8_t o extern void io(uint64_t address, uint8_t rw); +static inline uint8_t get_opsize(uint8_t prefix, uint8_t addrmode) { + uint8_t id = (prefix & 0x0C) >> 2; + switch (addrmode) { + case ZM: + switch (id) { + case 2: return 5; + case 3: return 3; + case 1: return 2; + case 0: return 0; + } + break; + case ABS: + switch (id) { + case 3: return 7; + case 2: return 6; + case 1: return 4; + case 0: return 1; + } + break; + } + return 0; +} + static inline uint64_t get_addr(struct sux *cpu, uint64_t *tmpaddr, uint8_t opcode, uint8_t prefix, uint8_t thread) { union reg address; union reg value; @@ -49,7 +100,7 @@ static inline uint64_t get_addr(struct sux *cpu, uint64_t *tmpaddr, uint8_t opco case IMPL: break; case IMM: - address.u64 = cpu->pc[thread]; + address.u64 = cpu->pc; switch (opcode) { case PHB: case PHP: @@ -67,8 +118,8 @@ static inline uint64_t get_addr(struct sux *cpu, uint64_t *tmpaddr, uint8_t opco case ROL: case ROR: case ASR: - case ENT: ++cpu->pc[thread]; break; - default : cpu->pc[thread]+=(1 << (prefix >> 4)); + case ENT: ++cpu->pc; break; + default : cpu->pc+=(1 << (prefix >> 4)); /* Falls Through. */ case TXS: break; } break; @@ -79,21 +130,11 @@ static inline uint64_t get_addr(struct sux *cpu, uint64_t *tmpaddr, uint8_t opco case INDX: case INDY: tmp = 0; - address.u8[0] = addr[cpu->pc[thread]]; + address.u8[0] = addr[cpu->pc]; /* Unroll Loop by implementing Duff's Device. */ - switch ((prefix & 0x0C) >> 2) { - case 2: - address.u8[5] = addr[cpu->pc[thread]+5];++tmp; - address.u8[4] = addr[cpu->pc[thread]+4];++tmp; - case 3: - address.u8[3] = addr[cpu->pc[thread]+3];++tmp; - case 1: - address.u8[2] = addr[cpu->pc[thread]+2];++tmp; - address.u8[1] = addr[cpu->pc[thread]+1];++tmp; - case 0: - ++tmp; - } - cpu->pc[thread]+=tmp; + tmp = get_opsize(prefix, ZM)+1; + setreg_sw(address.u8, 0, addr, cpu->pc, prefix, ZM, AM); + cpu->pc+=tmp; #if debug && !bench *tmpaddr = address.u64; #endif @@ -103,43 +144,32 @@ static inline uint64_t get_addr(struct sux *cpu, uint64_t *tmpaddr, uint8_t opco uint64_t reg = 0; switch (optype[opcode]) { case ZMX: - address.u64 += cpu->x[thread]; + address.u64 += cpu->x; #if getclk iclk++; #endif break; case ZMY: - address.u64 += cpu->y[thread]; + address.u64 += cpu->y; #if getclk iclk++; #endif break; case INDX: - address.u64 += cpu->x[thread]; - #if getclk - iclk++; - #endif - /* Falls Through. */ case INDY: - /* Did we fall through? */ if (optype[opcode] == INDX) { - reg = 0; /* Yes, so set reg back to zero. */ + address.u64 += cpu->x; + #if getclk + iclk++; + #endif } else { - reg = cpu->y[thread]; /* No, so set reg to Y. */ + reg = cpu->y; #if getclk iclk++; #endif } - /* Falls Through. */ case IND: - value.u8[0] = addr[address.u64 ]; - value.u8[1] = addr[address.u64+1]; - value.u8[2] = addr[address.u64+2]; - value.u8[3] = addr[address.u64+3]; - value.u8[4] = addr[address.u64+4]; - value.u8[5] = addr[address.u64+5]; - value.u8[6] = addr[address.u64+6]; - value.u8[7] = addr[address.u64+7]; + setreg(value.u8, +, 0, addr, +, address.u64, 7); #if getclk iclk++; #endif @@ -149,26 +179,9 @@ static inline uint64_t get_addr(struct sux *cpu, uint64_t *tmpaddr, uint8_t opco } break; case ABS: - tmp = 0; - address.u8[0] = addr[cpu->pc[thread]];++tmp; - /* Unroll Loop by implementing Duff's Device. */ - switch ((prefix & 0x0C) >> 2) { - case 3: - address.u8[7] = addr[cpu->pc[thread]+7];++tmp; - case 2: - address.u8[6] = addr[cpu->pc[thread]+6];++tmp; - address.u8[5] = addr[cpu->pc[thread]+5];++tmp; - #if getclk - iclk++; - #endif - case 1: - address.u8[4] = addr[cpu->pc[thread]+4];++tmp; - address.u8[3] = addr[cpu->pc[thread]+3];++tmp; - address.u8[2] = addr[cpu->pc[thread]+2];++tmp; - case 0: - address.u8[1] = addr[cpu->pc[thread]+1];++tmp; - } - cpu->pc[thread]+=tmp; + tmp = get_opsize(prefix, ABS)+1; + setreg_sw(address.u8, 0, addr, cpu->pc, prefix, ABS, AM); + cpu->pc+=tmp; #if getclk iclk++; #endif @@ -179,39 +192,39 @@ static inline uint64_t get_addr(struct sux *cpu, uint64_t *tmpaddr, uint8_t opco } static inline void adc(struct sux *cpu, uint64_t value, uint8_t thread) { - uint64_t sum = cpu->a[thread]+value+getflag(C); + uint64_t sum = cpu->a+value+getflag(C); setflag(sum == 0, Z); setflag((sum >> 63), N); - setflag(((cpu->a[thread]^value) >> 63) && ((cpu->a[thread]^sum) >> 63), V); + setflag(((cpu->a^value) >> 63) && ((cpu->a^sum) >> 63), V); setflag((sum < value), C); - cpu->a[thread] = sum; + cpu->a = sum; } static inline void sbc(struct sux *cpu, uint64_t value, uint8_t thread) { - uint64_t sum = cpu->a[thread]-value-!getflag(C); + uint64_t sum = cpu->a-value-!getflag(C); setflag(sum == 0, Z); setflag(sum >> 63, N); - setflag(((cpu->a[thread]^value) >> 63) && ((cpu->a[thread]^sum) >> 63), V); + setflag(((cpu->a^value) >> 63) && ((cpu->a^sum) >> 63), V); setflag((sum > value), C); - cpu->a[thread] = sum; + cpu->a = sum; } static inline void transfer(struct sux *cpu, uint64_t value, uint8_t opcode, uint8_t prefix, uint8_t thread) { uint64_t reg; switch (opcode) { - case TBA: cpu->a[thread] = cpu->b[thread]; reg = cpu->a[thread]; break; - case TXA: cpu->a[thread] = cpu->x[thread]; reg = cpu->a[thread]; break; - case TYA: cpu->a[thread] = cpu->y[thread]; reg = cpu->a[thread]; break; - case TAB: cpu->b[thread] = cpu->a[thread]; reg = cpu->b[thread]; break; - case TAY: cpu->y[thread] = cpu->a[thread]; reg = cpu->y[thread]; break; - case TXY: cpu->y[thread] = cpu->x[thread]; reg = cpu->y[thread]; break; - case TAX: cpu->x[thread] = cpu->a[thread]; reg = cpu->x[thread]; break; - case TYX: cpu->x[thread] = cpu->y[thread]; reg = cpu->x[thread]; break; - case TSX: cpu->x[thread] = cpu->sp[thread] & 0xFFFF; cpu->x[thread] = cpu->stk_st[thread] << 16; break; - case TXS: cpu->sp[thread] = cpu->x[thread]; + case TBA: cpu->a = cpu->b; reg = cpu->a; break; + case TXA: cpu->a = cpu->x; reg = cpu->a; break; + case TYA: cpu->a = cpu->y; reg = cpu->a; break; + case TAB: cpu->b = cpu->a; reg = cpu->b; break; + case TAY: cpu->y = cpu->a; reg = cpu->y; break; + case TXY: cpu->y = cpu->x; reg = cpu->y; break; + case TAX: cpu->x = cpu->a; reg = cpu->x; break; + case TYX: cpu->x = cpu->y; reg = cpu->x; break; + case TSX: cpu->x = cpu->sp & 0xFFFF; cpu->x = cpu->stk_st << 16; break; + case TXS: cpu->sp = cpu->x; if (prefix == 0x13 && (value == thread+1 || value > 8)) { - cpu->stk_st[thread] = value & 0xFF; - cpu->stk_st[thread] += value << 16; - cpu->pc[thread]+=2; + cpu->stk_st = value & 0xFF; + cpu->stk_st += value << 16; + cpu->pc+=2; } break; } @@ -219,125 +232,100 @@ static inline void transfer(struct sux *cpu, uint64_t value, uint8_t opcode, uin setflag(reg >> 63, N); } -static inline void push(struct sux *cpu, uint64_t value, uint64_t reg, uint8_t thread) { - union { - uint64_t reg; - uint8_t byte[8]; - } r; - r.reg = reg; - uint8_t size = (value > 0) ? value-1 : 0; +static inline void push(struct sux *cpu, uint64_t size, uint64_t value, uint8_t thread) { + union reg reg; + reg.u64 = value; + size = (size > 0) ? size-1 : 0; uint8_t tmp = (size <= 7) ? size : 7; - /* Unroll Loop by implementing Duff's Device. */ - switch (tmp) { - case 7: addr[(cpu->stk_st[thread] << 16)+cpu->sp[thread]] = r.byte[7];cpu->sp[thread]--; - case 6: addr[(cpu->stk_st[thread] << 16)+cpu->sp[thread]] = r.byte[6];cpu->sp[thread]--; - case 5: addr[(cpu->stk_st[thread] << 16)+cpu->sp[thread]] = r.byte[5];cpu->sp[thread]--; - case 4: addr[(cpu->stk_st[thread] << 16)+cpu->sp[thread]] = r.byte[4];cpu->sp[thread]--; - case 3: addr[(cpu->stk_st[thread] << 16)+cpu->sp[thread]] = r.byte[3];cpu->sp[thread]--; - case 2: addr[(cpu->stk_st[thread] << 16)+cpu->sp[thread]] = r.byte[2];cpu->sp[thread]--; - case 1: addr[(cpu->stk_st[thread] << 16)+cpu->sp[thread]] = r.byte[1];cpu->sp[thread]--; - case 0: addr[(cpu->stk_st[thread] << 16)+cpu->sp[thread]] = r.byte[0];cpu->sp[thread]--; - } + setreg(addr, -, (cpu->stk_st << 16)+cpu->sp, reg.u8, +, 0, tmp); + cpu->sp -= (tmp+1); } -static inline uint64_t pull(struct sux *cpu, uint64_t value, uint8_t thread) { - union { - uint64_t reg; - uint8_t byte[8]; - } r; - r.reg = 0; - uint8_t size = (value > 0) ? value-1 : 0; +static inline uint64_t pull(struct sux *cpu, uint64_t size, uint8_t thread) { + union reg reg; + reg.u64 = 0; + size = (size > 0) ? size-1 : 0; uint8_t tmp = (size <= 7) ? size : 7; - uint8_t tmp2 = 0; - /* Unroll Loop by implementing Duff's Device. */ - cpu->sp[thread]++;r.byte[tmp2] = addr[(cpu->stk_st[thread] << 16)+cpu->sp[thread]]; - switch (tmp) { - case 7: cpu->sp[thread]++;tmp2++;r.byte[tmp2] |= addr[(cpu->stk_st[thread] << 16)+cpu->sp[thread]]; - case 6: cpu->sp[thread]++;tmp2++;r.byte[tmp2] |= addr[(cpu->stk_st[thread] << 16)+cpu->sp[thread]]; - case 5: cpu->sp[thread]++;tmp2++;r.byte[tmp2] |= addr[(cpu->stk_st[thread] << 16)+cpu->sp[thread]]; - case 4: cpu->sp[thread]++;tmp2++;r.byte[tmp2] |= addr[(cpu->stk_st[thread] << 16)+cpu->sp[thread]]; - case 3: cpu->sp[thread]++;tmp2++;r.byte[tmp2] |= addr[(cpu->stk_st[thread] << 16)+cpu->sp[thread]]; - case 2: cpu->sp[thread]++;tmp2++;r.byte[tmp2] |= addr[(cpu->stk_st[thread] << 16)+cpu->sp[thread]]; - case 1: cpu->sp[thread]++;tmp2++;r.byte[tmp2] |= addr[(cpu->stk_st[thread] << 16)+cpu->sp[thread]]; - } - return r.reg; + cpu->sp += (tmp+1); + setreg(reg.u8, +, 0, addr, -, (cpu->stk_st << 16)+cpu->sp, tmp); + return reg.u64; } static inline void and(struct sux *cpu, uint64_t value, uint8_t thread) { - cpu->a[thread] &= value; - setflag(cpu->a[thread] == 0, Z); - setflag(cpu->a[thread] >> 63, N); + cpu->a &= value; + setflag(cpu->a == 0, Z); + setflag(cpu->a >> 63, N); } static inline void or(struct sux *cpu, uint64_t value, uint8_t thread) { - cpu->a[thread] |= value; - setflag(cpu->a[thread] == 0, Z); - setflag(cpu->a[thread] >> 63, N); + cpu->a |= value; + setflag(cpu->a == 0, Z); + setflag(cpu->a >> 63, N); } static inline void xor(struct sux *cpu, uint64_t value, uint8_t thread) { - cpu->a[thread] ^= value; - setflag(cpu->a[thread] == 0, Z); - setflag(cpu->a[thread] >> 63, N); + cpu->a ^= value; + setflag(cpu->a == 0, Z); + setflag(cpu->a >> 63, N); } static inline void lsl(struct sux *cpu, uint64_t value, uint8_t thread) { - uint64_t sum = (value < 64) ? cpu->a[thread] << value : 0; + uint64_t sum = (value < 64) ? cpu->a << value : 0; setflag(sum == 0, Z); setflag(sum >> 63, N); - setflag(cpu->a[thread] >> (64-value), C); - cpu->a[thread] = sum; + setflag(cpu->a >> (64-value), C); + cpu->a = sum; } static inline void lsr(struct sux *cpu, uint64_t value, uint8_t thread) { - uint64_t sum = (value < 64) ? cpu->a[thread] >> value : 0; + uint64_t sum = (value < 64) ? cpu->a >> value : 0; setflag(sum == 0, Z); setflag(sum >> 63, N); - setflag(cpu->a[thread] & 1, C); - cpu->a[thread] = sum; + setflag(cpu->a & 1, C); + cpu->a = sum; } static inline void asr(struct sux *cpu, uint64_t value, uint8_t thread) { - uint8_t sign = cpu->a[thread] >> 63; - uint64_t sum = (value < 64) ? (cpu->a[thread] >> value) | ((uint64_t)sign << 63) : 0; + uint8_t sign = cpu->a >> 63; + uint64_t sum = (value < 64) ? (cpu->a >> value) | ((uint64_t)sign << 63) : 0; setflag(sum == 0, Z); setflag(sum >> 63, N); - setflag(cpu->a[thread] & 1, C); - cpu->a[thread] = sum; + setflag(cpu->a & 1, C); + cpu->a = sum; } static inline void rol(struct sux *cpu, uint64_t value, uint8_t thread) { - uint64_t sum = cpu->a[thread] << value; + uint64_t sum = cpu->a << value; sum |= getflag(C); setflag(sum == 0, Z); setflag(sum >> 63, N); - setflag(cpu->a[thread] >> (uint64_t)(64-value), C); - cpu->a[thread] = sum; + setflag(cpu->a >> (uint64_t)(64-value), C); + cpu->a = sum; } static inline void ror(struct sux *cpu, uint64_t value, uint8_t thread) { - uint64_t sum = cpu->a[thread] >> value; + uint64_t sum = cpu->a >> value; sum |= (uint64_t)getflag(C) << (uint64_t)(64-value); setflag(sum == 0, Z); setflag(sum >> 63, N); - setflag(cpu->a[thread] & 1, C); - cpu->a[thread] = sum; + setflag(cpu->a & 1, C); + cpu->a = sum; } static inline void mul(struct sux *cpu, uint64_t value, uint8_t thread) { - uint64_t sum = cpu->a[thread]*value; - cpu->a[thread] = sum; + uint64_t sum = cpu->a*value; + cpu->a = sum; setflag(sum == 0, Z); setflag(sum >> 63, N); - setflag(!((cpu->a[thread]^value) >> 63) && ((cpu->a[thread]^sum) >> 63), V); + setflag(!((cpu->a^value) >> 63) && ((cpu->a^sum) >> 63), V); } static inline void divd(struct sux *cpu, uint64_t value, uint8_t opcode, uint8_t thread) { - uint64_t sum = cpu->a[thread]/value; + uint64_t sum = cpu->a/value; if (opcode != DAB) { - cpu->b[thread] = cpu->a[thread] % value; + cpu->b = cpu->a % value; } else { - value = cpu->b[thread]; - cpu->x[thread] = cpu->a[thread] % value; + value = cpu->b; + cpu->x = cpu->a % value; } - cpu->a[thread] = sum; + cpu->a = sum; setflag(sum == 0, Z); setflag((sum >> 63), N); } @@ -366,19 +354,7 @@ static inline void idm(struct sux *cpu, uint64_t address, uint8_t prefix, uint8_ union reg value; value.u64 = 0; /* Unroll Loop by implementing Duff's Device. */ - value.u8[0] = addr[address]; - switch (1 << (prefix >> 4)) { - case 8: - value.u8[7] = addr[address+7]; - value.u8[6] = addr[address+6]; - value.u8[5] = addr[address+5]; - value.u8[4] = addr[address+4]; - case 4: - value.u8[3] = addr[address+3]; - value.u8[2] = addr[address+2]; - case 2: - value.u8[1] = addr[address+1]; - } + setreg_sw(value.u8, 0, addr, address, prefix, 0, RS); if (inc) { value.u64++; } else { @@ -386,40 +362,15 @@ static inline void idm(struct sux *cpu, uint64_t address, uint8_t prefix, uint8_ } setflag(value.u64 == 0, Z); setflag(value.u64 >> 7, N); - addr[address] = value.u8[0]; + setreg_sw(addr, address, value.u8, 0, prefix, 0, RS); io(address, 0); - switch (1 << (prefix >> 4)) { - case 8: - addr[address+7] = value.u8[7]; - addr[address+6] = value.u8[6]; - addr[address+5] = value.u8[5]; - addr[address+4] = value.u8[4]; - case 4: - addr[address+3] = value.u8[3]; - addr[address+2] = value.u8[2]; - case 2: - addr[address+1] = value.u8[1]; - } } static inline uint64_t load(struct sux *cpu, uint64_t address, uint64_t reg, uint8_t prefix, uint8_t thread) { io(address, 1); union reg value; value.u64 = reg; - /* Unroll Loop by implementing Duff's Device. */ - value.u8[0] = addr[address]; - switch (1 << (prefix >> 4)) { - case 8: - value.u8[7] = addr[address+7]; - value.u8[6] = addr[address+6]; - value.u8[5] = addr[address+5]; - value.u8[4] = addr[address+4]; - case 4: - value.u8[3] = addr[address+3]; - value.u8[2] = addr[address+2]; - case 2: - value.u8[1] = addr[address+1]; - } + setreg_sw(value.u8, 0, addr, address, prefix, 0, RS); setflag(value.u64 == 0, Z); setflag(value.u64 >> 63, N); return value.u64; @@ -428,7 +379,7 @@ static inline uint64_t load(struct sux *cpu, uint64_t address, uint64_t reg, uin static inline void store(struct sux *cpu, uint64_t address, uint64_t reg, uint8_t prefix, uint8_t thread) { union reg value; value.u64 = reg; - addr[address] = value.u8[0]; + setreg_sw(addr, address, value.u8, 0, prefix, 0, RS); #if (IO || debug) && !branch #if keypoll pthread_mutex_lock(&mutex); @@ -438,18 +389,5 @@ static inline void store(struct sux *cpu, uint64_t address, uint64_t reg, uint8_ pthread_mutex_unlock(&mutex); #endif #endif - /* Unroll Loop by implementing Duff's Device. */ - switch (1 << (prefix >> 4)) { - case 8: - addr[address+7] = value.u8[7]; - addr[address+6] = value.u8[6]; - addr[address+5] = value.u8[5]; - addr[address+4] = value.u8[4]; - case 4: - addr[address+3] = value.u8[3]; - addr[address+2] = value.u8[2]; - case 2: - addr[address+1] = value.u8[1]; - } } |