void gencfc1(void)
{
#ifdef INTERPRET_CFC1
gencallinterp((native_type)cached_interpreter_table.CFC1, 0);
#else
gencheck_cop1_unusable();
#ifdef __x86_64__
if(dst->f.r.nrd == 31)
mov_xreg32_m32rel(EAX, (unsigned int*)&FCR31);
else
mov_xreg32_m32rel(EAX, (unsigned int*)&FCR0);
mov_m32rel_xreg32((unsigned int*)dst->f.r.rt, EAX);
sar_reg32_imm8(EAX, 31);
mov_m32rel_xreg32(((unsigned int*)dst->f.r.rt)+1, EAX);
#else
if(dst->f.r.nrd == 31)
mov_eax_memoffs32((unsigned int*)&FCR31);
else
mov_eax_memoffs32((unsigned int*)&FCR0);
mov_memoffs32_eax((unsigned int*)dst->f.r.rt);
sar_reg32_imm8(EAX, 31);
mov_memoffs32_eax(((unsigned int*)dst->f.r.rt)+1);
#endif
#endif
}
void gencfc1(void)
{
#ifdef INTERPRET_CFC1
gencallinterp((unsigned int)CFC1, 0);
#else
gencheck_cop1_unusable();
if(dst->f.r.nrd == 31) mov_eax_memoffs32((unsigned int*)&FCR31);
else mov_eax_memoffs32((unsigned int*)&FCR0);
mov_memoffs32_eax((unsigned int*)dst->f.r.rt);
sar_reg32_imm8(EAX, 31);
mov_memoffs32_eax(((unsigned int*)dst->f.r.rt)+1);
#endif
}
void genctc1(void)
{
#ifdef INTERPRET_CTC1
gencallinterp((unsigned int)CTC1, 0);
#else
gencheck_cop1_unusable();
if (dst->f.r.nrd != 31) return;
mov_eax_memoffs32((unsigned int*)dst->f.r.rt);
mov_memoffs32_eax((unsigned int*)&FCR31);
and_eax_imm32(3);
cmp_eax_imm32(0);
jne_rj(12);
mov_m32_imm32((unsigned int*)&rounding_mode, 0x33F); // 10
jmp_imm_short(48); // 2
cmp_eax_imm32(1); // 5
jne_rj(12); // 2
mov_m32_imm32((unsigned int*)&rounding_mode, 0xF3F); // 10
jmp_imm_short(29); // 2
cmp_eax_imm32(2); // 5
jne_rj(12); // 2
mov_m32_imm32((unsigned int*)&rounding_mode, 0xB3F); // 10
jmp_imm_short(10); // 2
mov_m32_imm32((unsigned int*)&rounding_mode, 0x73F); // 10
fldcw_m16((unsigned short*)&rounding_mode);
#endif
}
void genctc1(void)
{
#ifdef INTERPRET_CTC1
gencallinterp((native_type)cached_interpreter_table.CTC1, 0);
#else
gencheck_cop1_unusable();
if (dst->f.r.nrd != 31)
return;
#ifdef __x86_64__
mov_xreg32_m32rel(EAX, (unsigned int*)dst->f.r.rt);
mov_m32rel_xreg32((unsigned int*)&FCR31, EAX);
and_eax_imm32(3);
cmp_eax_imm32(0);
jne_rj(13);
mov_m32rel_imm32((unsigned int*)&rounding_mode, 0x33F); // 11
jmp_imm_short(51); // 2
cmp_eax_imm32(1); // 5
jne_rj(13); // 2
mov_m32rel_imm32((unsigned int*)&rounding_mode, 0xF3F); // 11
jmp_imm_short(31); // 2
cmp_eax_imm32(2); // 5
jne_rj(13); // 2
mov_m32rel_imm32((unsigned int*)&rounding_mode, 0xB3F); // 11
jmp_imm_short(11); // 2
mov_m32rel_imm32((unsigned int*)&rounding_mode, 0x73F); // 11
fldcw_m16rel((unsigned short*)&rounding_mode);
#else
mov_eax_memoffs32((unsigned int*)dst->f.r.rt);
mov_memoffs32_eax((unsigned int*)&FCR31);
and_eax_imm32(3);
cmp_eax_imm32(0);
jne_rj(12);
mov_m32_imm32((unsigned int*)&rounding_mode, 0x33F); // 10
jmp_imm_short(48); // 2
cmp_eax_imm32(1); // 5
jne_rj(12); // 2
mov_m32_imm32((unsigned int*)&rounding_mode, 0xF3F); // 10
jmp_imm_short(29); // 2
cmp_eax_imm32(2); // 5
jne_rj(12); // 2
mov_m32_imm32((unsigned int*)&rounding_mode, 0xB3F); // 10
jmp_imm_short(10); // 2
mov_m32_imm32((unsigned int*)&rounding_mode, 0x73F); // 10
fldcw_m16((unsigned short*)&rounding_mode);
#endif
#endif
}
void gendmultu()
{
#ifdef INTERPRET_DMULTU
gencallinterp((unsigned long)DMULTU, 0);
#else
free_all_registers();
simplify_access();
mov_eax_memoffs32((unsigned long *)dst->f.r.rs);
mul_m32((unsigned long *)dst->f.r.rt); // EDX:EAX = temp1
mov_memoffs32_eax((unsigned long *)(&lo));
mov_reg32_reg32(EBX, EDX); // EBX = temp1>>32
mov_eax_memoffs32((unsigned long *)dst->f.r.rs);
mul_m32((unsigned long *)(dst->f.r.rt)+1);
add_reg32_reg32(EBX, EAX);
adc_reg32_imm32(EDX, 0);
mov_reg32_reg32(ECX, EDX); // ECX:EBX = temp2
mov_eax_memoffs32((unsigned long *)(dst->f.r.rs)+1);
mul_m32((unsigned long *)dst->f.r.rt); // EDX:EAX = temp3
add_reg32_reg32(EBX, EAX);
adc_reg32_imm32(ECX, 0); // ECX:EBX = result2
mov_m32_reg32((unsigned long*)(&lo)+1, EBX);
mov_reg32_reg32(ESI, EDX); // ESI = temp3>>32
mov_eax_memoffs32((unsigned long *)(dst->f.r.rs)+1);
mul_m32((unsigned long *)(dst->f.r.rt)+1);
add_reg32_reg32(EAX, ESI);
adc_reg32_imm32(EDX, 0); // EDX:EAX = temp4
add_reg32_reg32(EAX, ECX);
adc_reg32_imm32(EDX, 0); // EDX:EAX = result3
mov_memoffs32_eax((unsigned long *)(&hi));
mov_m32_reg32((unsigned long *)(&hi)+1, EDX);
#endif
}
void gendmultu(usf_state_t * state)
{
#ifdef INTERPRET_DMULTU
gencallinterp(state, (unsigned int)state->current_instruction_table.DMULTU, 0);
#else
free_all_registers(state);
simplify_access(state);
mov_eax_memoffs32(state, (unsigned int *)state->dst->f.r.rs);
mul_m32(state, (unsigned int *)state->dst->f.r.rt); // EDX:EAX = temp1
mov_memoffs32_eax(state, (unsigned int *)(&state->lo));
mov_reg32_reg32(state, EBX, EDX); // EBX = temp1>>32
mov_eax_memoffs32(state, (unsigned int *)state->dst->f.r.rs);
mul_m32(state, (unsigned int *)(state->dst->f.r.rt)+1);
add_reg32_reg32(state, EBX, EAX);
adc_reg32_imm32(state, EDX, 0);
mov_reg32_reg32(state, ECX, EDX); // ECX:EBX = temp2
mov_eax_memoffs32(state, (unsigned int *)(state->dst->f.r.rs)+1);
mul_m32(state, (unsigned int *)state->dst->f.r.rt); // EDX:EAX = temp3
add_reg32_reg32(state, EBX, EAX);
adc_reg32_imm32(state, ECX, 0); // ECX:EBX = result2
mov_m32_reg32(state, (unsigned int*)(&state->lo)+1, EBX);
mov_reg32_reg32(state, EDI, EDX); // EDI = temp3>>32
mov_eax_memoffs32(state, (unsigned int *)(state->dst->f.r.rs)+1);
mul_m32(state, (unsigned int *)(state->dst->f.r.rt)+1);
add_reg32_reg32(state, EAX, EDI);
adc_reg32_imm32(state, EDX, 0); // EDX:EAX = temp4
add_reg32_reg32(state, EAX, ECX);
adc_reg32_imm32(state, EDX, 0); // EDX:EAX = result3
mov_memoffs32_eax(state,(unsigned int *)(&state->hi));
mov_m32_reg32(state, (unsigned int *)(&state->hi)+1, EDX);
#endif
}
void gendmultu(void)
{
#ifdef INTERPRET_DMULTU
gencallinterp((unsigned int)cached_interpreter_table.DMULTU, 0);
#else
free_all_registers();
simplify_access();
mov_eax_memoffs32((unsigned int *)g_dev.r4300.recomp.dst->f.r.rs);
mul_m32((unsigned int *)g_dev.r4300.recomp.dst->f.r.rt); // EDX:EAX = temp1
mov_memoffs32_eax((unsigned int *)(r4300_mult_lo()));
mov_reg32_reg32(EBX, EDX); // EBX = temp1>>32
mov_eax_memoffs32((unsigned int *)g_dev.r4300.recomp.dst->f.r.rs);
mul_m32((unsigned int *)(g_dev.r4300.recomp.dst->f.r.rt)+1);
add_reg32_reg32(EBX, EAX);
adc_reg32_imm32(EDX, 0);
mov_reg32_reg32(ECX, EDX); // ECX:EBX = temp2
mov_eax_memoffs32((unsigned int *)(g_dev.r4300.recomp.dst->f.r.rs)+1);
mul_m32((unsigned int *)g_dev.r4300.recomp.dst->f.r.rt); // EDX:EAX = temp3
add_reg32_reg32(EBX, EAX);
adc_reg32_imm32(ECX, 0); // ECX:EBX = result2
mov_m32_reg32((unsigned int*)(r4300_mult_lo())+1, EBX);
mov_reg32_reg32(ESI, EDX); // ESI = temp3>>32
mov_eax_memoffs32((unsigned int *)(g_dev.r4300.recomp.dst->f.r.rs)+1);
mul_m32((unsigned int *)(g_dev.r4300.recomp.dst->f.r.rt)+1);
add_reg32_reg32(EAX, ESI);
adc_reg32_imm32(EDX, 0); // EDX:EAX = temp4
add_reg32_reg32(EAX, ECX);
adc_reg32_imm32(EDX, 0); // EDX:EAX = result3
mov_memoffs32_eax((unsigned int *)(r4300_mult_hi()));
mov_m32_reg32((unsigned int *)(r4300_mult_hi())+1, EDX);
#endif
}
void genjalr()
{
#ifdef INTERPRET_JALR
gencallinterp((unsigned long)JALR, 0);
#else
static unsigned long precomp_instr_size = sizeof(precomp_instr);
unsigned long diff =
(unsigned long)(&dst->local_addr) - (unsigned long)(dst);
unsigned long diff_need =
(unsigned long)(&dst->reg_cache_infos.need_map) - (unsigned long)(dst);
unsigned long diff_wrap =
(unsigned long)(&dst->reg_cache_infos.jump_wrapper) - (unsigned long)(dst);
unsigned long temp, temp2;
if (((dst->addr & 0xFFF) == 0xFFC &&
(dst->addr < 0x80000000 || dst->addr >= 0xC0000000))||no_compiled_jump)
{
gencallinterp((unsigned long)JALR, 1);
return;
}
free_all_registers();
simplify_access();
mov_eax_memoffs32((unsigned long *)dst->f.r.rs);
mov_memoffs32_eax((unsigned long *)&local_rs);
gendelayslot();
mov_m32_imm32((unsigned long *)(dst-1)->f.r.rd, dst->addr+4);
if ((dst->addr+4) & 0x80000000)
mov_m32_imm32(((unsigned long *)(dst-1)->f.r.rd)+1, 0xFFFFFFFF);
else
mov_m32_imm32(((unsigned long *)(dst-1)->f.r.rd)+1, 0);
mov_eax_memoffs32((unsigned long *)&local_rs);
mov_memoffs32_eax((unsigned long *)&last_addr);
gencheck_interupt_reg();
mov_eax_memoffs32((unsigned long *)&local_rs);
mov_reg32_reg32(EBX, EAX);
and_eax_imm32(0xFFFFF000);
cmp_eax_imm32(dst_block->start & 0xFFFFF000);
je_near_rj(0);
temp = code_length;
mov_m32_reg32(&jump_to_address, EBX);
mov_m32_imm32((unsigned long*)(&PC), (unsigned long)(dst+1));
mov_reg32_imm32(EAX, (unsigned long)jump_to_func);
call_reg32(EAX);
temp2 = code_length;
code_length = temp-4;
put32(temp2 - temp);
code_length = temp2;
mov_reg32_reg32(EAX, EBX);
sub_eax_imm32(dst_block->start);
shr_reg32_imm8(EAX, 2);
mul_m32((unsigned long *)(&precomp_instr_size));
mov_reg32_preg32pimm32(EBX, EAX, (unsigned long)(dst_block->block)+diff_need);
cmp_reg32_imm32(EBX, 1);
jne_rj(7);
add_eax_imm32((unsigned long)(dst_block->block)+diff_wrap); // 5
jmp_reg32(EAX); // 2
mov_reg32_preg32pimm32(EAX, EAX, (unsigned long)(dst_block->block)+diff);
add_reg32_m32(EAX, (unsigned long *)(&dst_block->code));
jmp_reg32(EAX);
#endif
}
void genjr(void)
{
#ifdef INTERPRET_JR
gencallinterp((unsigned int)cached_interpreter_table.JR, 1);
#else
static unsigned int precomp_instr_size = sizeof(precomp_instr);
unsigned int diff =
(unsigned int)(&dst->local_addr) - (unsigned int)(dst);
unsigned int diff_need =
(unsigned int)(&dst->reg_cache_infos.need_map) - (unsigned int)(dst);
unsigned int diff_wrap =
(unsigned int)(&dst->reg_cache_infos.jump_wrapper) - (unsigned int)(dst);
if (((dst->addr & 0xFFF) == 0xFFC &&
(dst->addr < 0x80000000 || dst->addr >= 0xC0000000))||no_compiled_jump)
{
gencallinterp((unsigned int)cached_interpreter_table.JR, 1);
return;
}
free_all_registers();
simplify_access();
mov_eax_memoffs32((unsigned int *)dst->f.i.rs);
mov_memoffs32_eax((unsigned int *)&local_rs);
gendelayslot();
mov_eax_memoffs32((unsigned int *)&local_rs);
mov_memoffs32_eax((unsigned int *)&last_addr);
gencheck_interupt_reg();
mov_eax_memoffs32((unsigned int *)&local_rs);
mov_reg32_reg32(EBX, EAX);
and_eax_imm32(0xFFFFF000);
cmp_eax_imm32(dst_block->start & 0xFFFFF000);
je_near_rj(0);
jump_start_rel32();
mov_m32_reg32(&jump_to_address, EBX);
mov_m32_imm32((unsigned int*)(&PC), (unsigned int)(dst+1));
mov_reg32_imm32(EAX, (unsigned int)jump_to_func);
call_reg32(EAX);
jump_end_rel32();
mov_reg32_reg32(EAX, EBX);
sub_eax_imm32(dst_block->start);
shr_reg32_imm8(EAX, 2);
mul_m32((unsigned int *)(&precomp_instr_size));
mov_reg32_preg32pimm32(EBX, EAX, (unsigned int)(dst_block->block)+diff_need);
cmp_reg32_imm32(EBX, 1);
jne_rj(7);
add_eax_imm32((unsigned int)(dst_block->block)+diff_wrap); // 5
jmp_reg32(EAX); // 2
mov_reg32_preg32pimm32(EAX, EAX, (unsigned int)(dst_block->block)+diff);
add_reg32_m32(EAX, (unsigned int *)(&dst_block->code));
jmp_reg32(EAX);
#endif
}
void genjalr(usf_state_t * state)
{
#ifdef INTERPRET_JALR
gencallinterp(state, (unsigned int)state->current_instruction_table.JALR, 0);
#else
unsigned int diff =
(unsigned int)(&state->dst->local_addr) - (unsigned int)(state->dst);
unsigned int diff_need =
(unsigned int)(&state->dst->reg_cache_infos.need_map) - (unsigned int)(state->dst);
unsigned int diff_wrap =
(unsigned int)(&state->dst->reg_cache_infos.jump_wrapper) - (unsigned int)(state->dst);
if (((state->dst->addr & 0xFFF) == 0xFFC &&
(state->dst->addr < 0x80000000 || state->dst->addr >= 0xC0000000))||state->no_compiled_jump)
{
gencallinterp(state, (unsigned int)state->current_instruction_table.JALR, 1);
return;
}
free_all_registers(state);
simplify_access(state);
mov_eax_memoffs32(state, (unsigned int *)state->dst->f.r.rs);
mov_memoffs32_eax(state, (unsigned int *)&state->local_rs);
gendelayslot(state);
mov_m32_imm32(state, (unsigned int *)(state->dst-1)->f.r.rd, state->dst->addr+4);
if ((state->dst->addr+4) & 0x80000000)
mov_m32_imm32(state, ((unsigned int *)(state->dst-1)->f.r.rd)+1, 0xFFFFFFFF);
else
mov_m32_imm32(state, ((unsigned int *)(state->dst-1)->f.r.rd)+1, 0);
mov_eax_memoffs32(state, (unsigned int *)&state->local_rs);
mov_memoffs32_eax(state, (unsigned int *)&state->last_addr);
gencheck_interupt_reg(state);
mov_eax_memoffs32(state, (unsigned int *)&state->local_rs);
mov_reg32_reg32(state, EBX, EAX);
and_eax_imm32(state, 0xFFFFF000);
cmp_eax_imm32(state, state->dst_block->start & 0xFFFFF000);
je_near_rj(state, 0);
jump_start_rel32(state);
mov_m32_reg32(state, &state->jump_to_address, EBX);
mov_m32_imm32(state, (unsigned int*)(&state->PC), (unsigned int)(state->dst+1));
mov_reg32_imm32(state, EBX, (unsigned int)jump_to_func);
mov_reg32_reg32(state, RP0, ESI);
call_reg32(state, EBX);
jump_end_rel32(state);
mov_reg32_reg32(state, EAX, EBX);
sub_eax_imm32(state, state->dst_block->start);
shr_reg32_imm8(state, EAX, 2);
mul_m32(state, (unsigned int *)(&state->precomp_instr_size));
mov_reg32_preg32pimm32(state, EBX, EAX, (unsigned int)(state->dst_block->block)+diff_need);
cmp_reg32_imm32(state, EBX, 1);
jne_rj(state, 7);
add_eax_imm32(state, (unsigned int)(state->dst_block->block)+diff_wrap); // 5
jmp_reg32(state, EAX); // 2
mov_reg32_preg32pimm32(state, EAX, EAX, (unsigned int)(state->dst_block->block)+diff);
add_reg32_m32(state, EAX, (unsigned int *)(&state->dst_block->code));
jmp_reg32(state, EAX);
#endif
}
void genjalr(void)
{
#ifdef INTERPRET_JALR
gencallinterp((unsigned int)cached_interpreter_table.JALR, 0);
#else
unsigned int diff =
(unsigned int)(&g_dev.r4300.recomp.dst->local_addr) - (unsigned int)(g_dev.r4300.recomp.dst);
unsigned int diff_need =
(unsigned int)(&g_dev.r4300.recomp.dst->reg_cache_infos.need_map) - (unsigned int)(g_dev.r4300.recomp.dst);
unsigned int diff_wrap =
(unsigned int)(&g_dev.r4300.recomp.dst->reg_cache_infos.jump_wrapper) - (unsigned int)(g_dev.r4300.recomp.dst);
if (((g_dev.r4300.recomp.dst->addr & 0xFFF) == 0xFFC &&
(g_dev.r4300.recomp.dst->addr < 0x80000000 || g_dev.r4300.recomp.dst->addr >= 0xC0000000))||g_dev.r4300.recomp.no_compiled_jump)
{
gencallinterp((unsigned int)cached_interpreter_table.JALR, 1);
return;
}
free_all_registers();
simplify_access();
mov_eax_memoffs32((unsigned int *)g_dev.r4300.recomp.dst->f.r.rs);
mov_memoffs32_eax((unsigned int *)&g_dev.r4300.local_rs);
gendelayslot();
mov_m32_imm32((unsigned int *)(g_dev.r4300.recomp.dst-1)->f.r.rd, g_dev.r4300.recomp.dst->addr+4);
if ((g_dev.r4300.recomp.dst->addr+4) & 0x80000000)
mov_m32_imm32(((unsigned int *)(g_dev.r4300.recomp.dst-1)->f.r.rd)+1, 0xFFFFFFFF);
else
mov_m32_imm32(((unsigned int *)(g_dev.r4300.recomp.dst-1)->f.r.rd)+1, 0);
mov_eax_memoffs32((unsigned int *)&g_dev.r4300.local_rs);
mov_memoffs32_eax((unsigned int *)&g_dev.r4300.cp0.last_addr);
gencheck_interupt_reg();
mov_eax_memoffs32((unsigned int *)&g_dev.r4300.local_rs);
mov_reg32_reg32(EBX, EAX);
and_eax_imm32(0xFFFFF000);
cmp_eax_imm32(g_dev.r4300.recomp.dst_block->start & 0xFFFFF000);
je_near_rj(0);
jump_start_rel32();
mov_m32_reg32(&g_dev.r4300.cached_interp.jump_to_address, EBX);
mov_m32_imm32((unsigned int*)(&(*r4300_pc_struct())), (unsigned int)(g_dev.r4300.recomp.dst+1));
mov_reg32_imm32(EAX, (unsigned int)jump_to_func);
call_reg32(EAX);
jump_end_rel32();
mov_reg32_reg32(EAX, EBX);
sub_eax_imm32(g_dev.r4300.recomp.dst_block->start);
shr_reg32_imm8(EAX, 2);
mul_m32((unsigned int *)(&precomp_instr_size));
mov_reg32_preg32pimm32(EBX, EAX, (unsigned int)(g_dev.r4300.recomp.dst_block->block)+diff_need);
cmp_reg32_imm32(EBX, 1);
jne_rj(7);
add_eax_imm32((unsigned int)(g_dev.r4300.recomp.dst_block->block)+diff_wrap); // 5
jmp_reg32(EAX); // 2
mov_reg32_preg32pimm32(EAX, EAX, (unsigned int)(g_dev.r4300.recomp.dst_block->block)+diff);
add_reg32_m32(EAX, (unsigned int *)(&g_dev.r4300.recomp.dst_block->code));
jmp_reg32(EAX);
#endif
}
