filename | src/sh4/sh4core.in |
changeset | 758:99ae000d4e09 |
prev | 740:dd11269ee48b |
next | 927:17b6b9e245d8 |
author | nkeynes |
date | Wed Oct 29 23:51:58 2008 +0000 (15 years ago) |
permissions | -rw-r--r-- |
last change | Use regparam calling conventions for all functions called from translated code, along with a few other high-use functions. Can probably extend this to all functions, but as it is this is a nice performance boost |
view | annotate | diff | log | raw |
1 /**
2 * $Id$
3 *
4 * SH4 emulation core, and parent module for all the SH4 peripheral
5 * modules.
6 *
7 * Copyright (c) 2005 Nathan Keynes.
8 *
9 * This program is free software; you can redistribute it and/or modify
10 * it under the terms of the GNU General Public License as published by
11 * the Free Software Foundation; either version 2 of the License, or
12 * (at your option) any later version.
13 *
14 * This program is distributed in the hope that it will be useful,
15 * but WITHOUT ANY WARRANTY; without even the implied warranty of
16 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
17 * GNU General Public License for more details.
18 */
20 #define MODULE sh4_module
21 #include <assert.h>
22 #include <math.h>
23 #include "dream.h"
24 #include "dreamcast.h"
25 #include "eventq.h"
26 #include "mem.h"
27 #include "clock.h"
28 #include "syscall.h"
29 #include "sh4/sh4core.h"
30 #include "sh4/sh4mmio.h"
31 #include "sh4/sh4stat.h"
32 #include "sh4/intc.h"
34 #define SH4_CALLTRACE 1
36 #define MAX_INT 0x7FFFFFFF
37 #define MIN_INT 0x80000000
38 #define MAX_INTF 2147483647.0
39 #define MIN_INTF -2147483648.0
41 /********************** SH4 Module Definition ****************************/
43 uint32_t sh4_emulate_run_slice( uint32_t nanosecs )
44 {
45 int i;
47 if( sh4_breakpoint_count == 0 ) {
48 for( ; sh4r.slice_cycle < nanosecs; sh4r.slice_cycle += sh4_cpu_period ) {
49 if( SH4_EVENT_PENDING() ) {
50 if( sh4r.event_types & PENDING_EVENT ) {
51 event_execute();
52 }
53 /* Eventq execute may (quite likely) deliver an immediate IRQ */
54 if( sh4r.event_types & PENDING_IRQ ) {
55 sh4_accept_interrupt();
56 }
57 }
58 if( !sh4_execute_instruction() ) {
59 break;
60 }
61 }
62 } else {
63 for( ;sh4r.slice_cycle < nanosecs; sh4r.slice_cycle += sh4_cpu_period ) {
64 if( SH4_EVENT_PENDING() ) {
65 if( sh4r.event_types & PENDING_EVENT ) {
66 event_execute();
67 }
68 /* Eventq execute may (quite likely) deliver an immediate IRQ */
69 if( sh4r.event_types & PENDING_IRQ ) {
70 sh4_accept_interrupt();
71 }
72 }
74 if( !sh4_execute_instruction() )
75 break;
76 #ifdef ENABLE_DEBUG_MODE
77 for( i=0; i<sh4_breakpoint_count; i++ ) {
78 if( sh4_breakpoints[i].address == sh4r.pc ) {
79 break;
80 }
81 }
82 if( i != sh4_breakpoint_count ) {
83 sh4_core_exit( CORE_EXIT_BREAKPOINT );
84 }
85 #endif
86 }
87 }
89 /* If we aborted early, but the cpu is still technically running,
90 * we're doing a hard abort - cut the timeslice back to what we
91 * actually executed
92 */
93 if( sh4r.slice_cycle != nanosecs && sh4r.sh4_state == SH4_STATE_RUNNING ) {
94 nanosecs = sh4r.slice_cycle;
95 }
96 if( sh4r.sh4_state != SH4_STATE_STANDBY ) {
97 TMU_run_slice( nanosecs );
98 SCIF_run_slice( nanosecs );
99 }
100 return nanosecs;
101 }
103 /********************** SH4 emulation core ****************************/
105 #define UNDEF(ir) return sh4_raise_slot_exception(EXC_ILLEGAL, EXC_SLOT_ILLEGAL)
106 #define UNIMP(ir) do{ ERROR( "Halted on unimplemented instruction at %08x, opcode = %04x", sh4r.pc, ir ); sh4_core_exit(CORE_EXIT_HALT); return FALSE; }while(0)
108 #if(SH4_CALLTRACE == 1)
109 #define MAX_CALLSTACK 32
110 static struct call_stack {
111 sh4addr_t call_addr;
112 sh4addr_t target_addr;
113 sh4addr_t stack_pointer;
114 } call_stack[MAX_CALLSTACK];
116 static int call_stack_depth = 0;
117 int sh4_call_trace_on = 0;
119 static inline void trace_call( sh4addr_t source, sh4addr_t dest )
120 {
121 if( call_stack_depth < MAX_CALLSTACK ) {
122 call_stack[call_stack_depth].call_addr = source;
123 call_stack[call_stack_depth].target_addr = dest;
124 call_stack[call_stack_depth].stack_pointer = sh4r.r[15];
125 }
126 call_stack_depth++;
127 }
129 static inline void trace_return( sh4addr_t source, sh4addr_t dest )
130 {
131 if( call_stack_depth > 0 ) {
132 call_stack_depth--;
133 }
134 }
136 void fprint_stack_trace( FILE *f )
137 {
138 int i = call_stack_depth -1;
139 if( i >= MAX_CALLSTACK )
140 i = MAX_CALLSTACK - 1;
141 for( ; i >= 0; i-- ) {
142 fprintf( f, "%d. Call from %08X => %08X, SP=%08X\n",
143 (call_stack_depth - i), call_stack[i].call_addr,
144 call_stack[i].target_addr, call_stack[i].stack_pointer );
145 }
146 }
148 #define TRACE_CALL( source, dest ) trace_call(source, dest)
149 #define TRACE_RETURN( source, dest ) trace_return(source, dest)
150 #else
151 #define TRACE_CALL( dest, rts )
152 #define TRACE_RETURN( source, dest )
153 #endif
155 #define CHECKPRIV() if( !IS_SH4_PRIVMODE() ) return sh4_raise_slot_exception( EXC_ILLEGAL, EXC_SLOT_ILLEGAL )
156 #define CHECKRALIGN16(addr) if( (addr)&0x01 ) return sh4_raise_exception( EXC_DATA_ADDR_READ )
157 #define CHECKRALIGN32(addr) if( (addr)&0x03 ) return sh4_raise_exception( EXC_DATA_ADDR_READ )
158 #define CHECKRALIGN64(addr) if( (addr)&0x07 ) return sh4_raise_exception( EXC_DATA_ADDR_READ )
159 #define CHECKWALIGN16(addr) if( (addr)&0x01 ) return sh4_raise_exception( EXC_DATA_ADDR_WRITE )
160 #define CHECKWALIGN32(addr) if( (addr)&0x03 ) return sh4_raise_exception( EXC_DATA_ADDR_WRITE )
161 #define CHECKWALIGN64(addr) if( (addr)&0x07 ) return sh4_raise_exception( EXC_DATA_ADDR_WRITE )
163 #define CHECKFPUEN() if( !IS_FPU_ENABLED() ) { if( ir == 0xFFFD ) { UNDEF(ir); } else { return sh4_raise_slot_exception( EXC_FPU_DISABLED, EXC_SLOT_FPU_DISABLED ); } }
164 #define CHECKDEST(p) if( (p) == 0 ) { ERROR( "%08X: Branch/jump to NULL, CPU halted", sh4r.pc ); sh4_core_exit(CORE_EXIT_HALT); return FALSE; }
165 #define CHECKSLOTILLEGAL() if(sh4r.in_delay_slot) return sh4_raise_exception(EXC_SLOT_ILLEGAL)
167 #define MEM_READ_BYTE( addr, val ) memtmp = mmu_vma_to_phys_read(addr); if( memtmp == MMU_VMA_ERROR ) { return TRUE; } else { val = sh4_read_byte(memtmp); }
168 #define MEM_READ_WORD( addr, val ) memtmp = mmu_vma_to_phys_read(addr); if( memtmp == MMU_VMA_ERROR ) { return TRUE; } else { val = sh4_read_word(memtmp); }
169 #define MEM_READ_LONG( addr, val ) memtmp = mmu_vma_to_phys_read(addr); if( memtmp == MMU_VMA_ERROR ) { return TRUE; } else { val = sh4_read_long(memtmp); }
170 #define MEM_WRITE_BYTE( addr, val ) memtmp = mmu_vma_to_phys_write(addr); if( memtmp == MMU_VMA_ERROR ) { return TRUE; } else { sh4_write_byte(memtmp, val); }
171 #define MEM_WRITE_WORD( addr, val ) memtmp = mmu_vma_to_phys_write(addr); if( memtmp == MMU_VMA_ERROR ) { return TRUE; } else { sh4_write_word(memtmp, val); }
172 #define MEM_WRITE_LONG( addr, val ) memtmp = mmu_vma_to_phys_write(addr); if( memtmp == MMU_VMA_ERROR ) { return TRUE; } else { sh4_write_long(memtmp, val); }
174 #define FP_WIDTH (IS_FPU_DOUBLESIZE() ? 8 : 4)
176 #define MEM_FP_READ( addr, reg ) \
177 if( IS_FPU_DOUBLESIZE() ) { \
178 CHECKRALIGN64(addr); \
179 memtmp = mmu_vma_to_phys_read(addr); \
180 if( memtmp == MMU_VMA_ERROR ) { \
181 return TRUE; \
182 } else { \
183 if( reg & 1 ) { \
184 *((uint32_t *)&XF((reg) & 0x0E)) = sh4_read_long(memtmp); \
185 *((uint32_t *)&XF(reg)) = sh4_read_long(memtmp+4); \
186 } else { \
187 *((uint32_t *)&FR(reg)) = sh4_read_long(memtmp); \
188 *((uint32_t *)&FR((reg) | 0x01)) = sh4_read_long(memtmp+4); \
189 } \
190 } \
191 } else { \
192 CHECKRALIGN32(addr); \
193 memtmp = mmu_vma_to_phys_read(addr); \
194 if( memtmp == MMU_VMA_ERROR ) { \
195 return TRUE; \
196 } else { \
197 *((uint32_t *)&FR(reg)) = sh4_read_long(memtmp); \
198 } \
199 }
200 #define MEM_FP_WRITE( addr, reg ) \
201 if( IS_FPU_DOUBLESIZE() ) { \
202 CHECKWALIGN64(addr); \
203 memtmp = mmu_vma_to_phys_write(addr); \
204 if( memtmp == MMU_VMA_ERROR ) { \
205 return TRUE; \
206 } else { \
207 if( reg & 1 ) { \
208 sh4_write_long( memtmp, *((uint32_t *)&XF((reg)&0x0E)) ); \
209 sh4_write_long( memtmp+4, *((uint32_t *)&XF(reg)) ); \
210 } else { \
211 sh4_write_long( memtmp, *((uint32_t *)&FR(reg)) ); \
212 sh4_write_long( memtmp+4, *((uint32_t *)&FR((reg)|0x01)) ); \
213 } \
214 } \
215 } else { \
216 CHECKWALIGN32(addr); \
217 memtmp = mmu_vma_to_phys_write(addr); \
218 if( memtmp == MMU_VMA_ERROR ) { \
219 return TRUE; \
220 } else { \
221 sh4_write_long( memtmp, *((uint32_t *)&FR((reg))) ); \
222 } \
223 }
225 gboolean sh4_execute_instruction( void )
226 {
227 uint32_t pc;
228 unsigned short ir;
229 uint32_t tmp;
230 float ftmp;
231 double dtmp;
232 int64_t memtmp; // temporary holder for memory reads
234 #define R0 sh4r.r[0]
235 pc = sh4r.pc;
236 if( pc > 0xFFFFFF00 ) {
237 /* SYSCALL Magic */
238 syscall_invoke( pc );
239 sh4r.in_delay_slot = 0;
240 pc = sh4r.pc = sh4r.pr;
241 sh4r.new_pc = sh4r.pc + 2;
242 return TRUE;
243 }
244 CHECKRALIGN16(pc);
246 #ifdef ENABLE_SH4STATS
247 sh4_stats_add_by_pc(sh4r.pc);
248 #endif
250 /* Read instruction */
251 if( !IS_IN_ICACHE(pc) ) {
252 if( !mmu_update_icache(pc) ) {
253 // Fault - look for the fault handler
254 if( !mmu_update_icache(sh4r.pc) ) {
255 // double fault - halt
256 ERROR( "Double fault - halting" );
257 sh4_core_exit(CORE_EXIT_HALT);
258 return FALSE;
259 }
260 }
261 pc = sh4r.pc;
262 }
263 assert( IS_IN_ICACHE(pc) );
264 ir = *(uint16_t *)GET_ICACHE_PTR(sh4r.pc);
265 %%
266 AND Rm, Rn {: sh4r.r[Rn] &= sh4r.r[Rm]; :}
267 AND #imm, R0 {: R0 &= imm; :}
268 AND.B #imm, @(R0, GBR) {: MEM_READ_BYTE(R0+sh4r.gbr, tmp); MEM_WRITE_BYTE( R0 + sh4r.gbr, imm & tmp ); :}
269 NOT Rm, Rn {: sh4r.r[Rn] = ~sh4r.r[Rm]; :}
270 OR Rm, Rn {: sh4r.r[Rn] |= sh4r.r[Rm]; :}
271 OR #imm, R0 {: R0 |= imm; :}
272 OR.B #imm, @(R0, GBR) {: MEM_READ_BYTE(R0+sh4r.gbr, tmp); MEM_WRITE_BYTE( R0 + sh4r.gbr, imm | tmp ); :}
273 TAS.B @Rn {:
274 MEM_READ_BYTE( sh4r.r[Rn], tmp );
275 sh4r.t = ( tmp == 0 ? 1 : 0 );
276 MEM_WRITE_BYTE( sh4r.r[Rn], tmp | 0x80 );
277 :}
278 TST Rm, Rn {: sh4r.t = (sh4r.r[Rn]&sh4r.r[Rm] ? 0 : 1); :}
279 TST #imm, R0 {: sh4r.t = (R0 & imm ? 0 : 1); :}
280 TST.B #imm, @(R0, GBR) {: MEM_READ_BYTE(R0+sh4r.gbr, tmp); sh4r.t = ( tmp & imm ? 0 : 1 ); :}
281 XOR Rm, Rn {: sh4r.r[Rn] ^= sh4r.r[Rm]; :}
282 XOR #imm, R0 {: R0 ^= imm; :}
283 XOR.B #imm, @(R0, GBR) {: MEM_READ_BYTE(R0+sh4r.gbr, tmp); MEM_WRITE_BYTE( R0 + sh4r.gbr, imm ^ tmp ); :}
284 XTRCT Rm, Rn {: sh4r.r[Rn] = (sh4r.r[Rn]>>16) | (sh4r.r[Rm]<<16); :}
286 ROTL Rn {:
287 sh4r.t = sh4r.r[Rn] >> 31;
288 sh4r.r[Rn] <<= 1;
289 sh4r.r[Rn] |= sh4r.t;
290 :}
291 ROTR Rn {:
292 sh4r.t = sh4r.r[Rn] & 0x00000001;
293 sh4r.r[Rn] >>= 1;
294 sh4r.r[Rn] |= (sh4r.t << 31);
295 :}
296 ROTCL Rn {:
297 tmp = sh4r.r[Rn] >> 31;
298 sh4r.r[Rn] <<= 1;
299 sh4r.r[Rn] |= sh4r.t;
300 sh4r.t = tmp;
301 :}
302 ROTCR Rn {:
303 tmp = sh4r.r[Rn] & 0x00000001;
304 sh4r.r[Rn] >>= 1;
305 sh4r.r[Rn] |= (sh4r.t << 31 );
306 sh4r.t = tmp;
307 :}
308 SHAD Rm, Rn {:
309 tmp = sh4r.r[Rm];
310 if( (tmp & 0x80000000) == 0 ) sh4r.r[Rn] <<= (tmp&0x1f);
311 else if( (tmp & 0x1F) == 0 )
312 sh4r.r[Rn] = ((int32_t)sh4r.r[Rn]) >> 31;
313 else
314 sh4r.r[Rn] = ((int32_t)sh4r.r[Rn]) >> (((~sh4r.r[Rm]) & 0x1F)+1);
315 :}
316 SHLD Rm, Rn {:
317 tmp = sh4r.r[Rm];
318 if( (tmp & 0x80000000) == 0 ) sh4r.r[Rn] <<= (tmp&0x1f);
319 else if( (tmp & 0x1F) == 0 ) sh4r.r[Rn] = 0;
320 else sh4r.r[Rn] >>= (((~tmp) & 0x1F)+1);
321 :}
322 SHAL Rn {:
323 sh4r.t = sh4r.r[Rn] >> 31;
324 sh4r.r[Rn] <<= 1;
325 :}
326 SHAR Rn {:
327 sh4r.t = sh4r.r[Rn] & 0x00000001;
328 sh4r.r[Rn] = ((int32_t)sh4r.r[Rn]) >> 1;
329 :}
330 SHLL Rn {: sh4r.t = sh4r.r[Rn] >> 31; sh4r.r[Rn] <<= 1; :}
331 SHLR Rn {: sh4r.t = sh4r.r[Rn] & 0x00000001; sh4r.r[Rn] >>= 1; :}
332 SHLL2 Rn {: sh4r.r[Rn] <<= 2; :}
333 SHLR2 Rn {: sh4r.r[Rn] >>= 2; :}
334 SHLL8 Rn {: sh4r.r[Rn] <<= 8; :}
335 SHLR8 Rn {: sh4r.r[Rn] >>= 8; :}
336 SHLL16 Rn {: sh4r.r[Rn] <<= 16; :}
337 SHLR16 Rn {: sh4r.r[Rn] >>= 16; :}
339 EXTU.B Rm, Rn {: sh4r.r[Rn] = sh4r.r[Rm]&0x000000FF; :}
340 EXTU.W Rm, Rn {: sh4r.r[Rn] = sh4r.r[Rm]&0x0000FFFF; :}
341 EXTS.B Rm, Rn {: sh4r.r[Rn] = SIGNEXT8( sh4r.r[Rm]&0x000000FF ); :}
342 EXTS.W Rm, Rn {: sh4r.r[Rn] = SIGNEXT16( sh4r.r[Rm]&0x0000FFFF ); :}
343 SWAP.B Rm, Rn {: sh4r.r[Rn] = (sh4r.r[Rm]&0xFFFF0000) | ((sh4r.r[Rm]&0x0000FF00)>>8) | ((sh4r.r[Rm]&0x000000FF)<<8); :}
344 SWAP.W Rm, Rn {: sh4r.r[Rn] = (sh4r.r[Rm]>>16) | (sh4r.r[Rm]<<16); :}
346 CLRT {: sh4r.t = 0; :}
347 SETT {: sh4r.t = 1; :}
348 CLRMAC {: sh4r.mac = 0; :}
349 LDTLB {: MMU_ldtlb(); :}
350 CLRS {: sh4r.s = 0; :}
351 SETS {: sh4r.s = 1; :}
352 MOVT Rn {: sh4r.r[Rn] = sh4r.t; :}
353 NOP {: /* NOP */ :}
355 PREF @Rn {:
356 tmp = sh4r.r[Rn];
357 if( (tmp & 0xFC000000) == 0xE0000000 ) {
358 sh4_flush_store_queue(tmp);
359 }
360 :}
361 OCBI @Rn {: :}
362 OCBP @Rn {: :}
363 OCBWB @Rn {: :}
364 MOVCA.L R0, @Rn {:
365 tmp = sh4r.r[Rn];
366 CHECKWALIGN32(tmp);
367 MEM_WRITE_LONG( tmp, R0 );
368 :}
369 MOV.B Rm, @(R0, Rn) {: MEM_WRITE_BYTE( R0 + sh4r.r[Rn], sh4r.r[Rm] ); :}
370 MOV.W Rm, @(R0, Rn) {:
371 CHECKWALIGN16( R0 + sh4r.r[Rn] );
372 MEM_WRITE_WORD( R0 + sh4r.r[Rn], sh4r.r[Rm] );
373 :}
374 MOV.L Rm, @(R0, Rn) {:
375 CHECKWALIGN32( R0 + sh4r.r[Rn] );
376 MEM_WRITE_LONG( R0 + sh4r.r[Rn], sh4r.r[Rm] );
377 :}
378 MOV.B @(R0, Rm), Rn {: MEM_READ_BYTE( R0 + sh4r.r[Rm], sh4r.r[Rn] ); :}
379 MOV.W @(R0, Rm), Rn {: CHECKRALIGN16( R0 + sh4r.r[Rm] );
380 MEM_READ_WORD( R0 + sh4r.r[Rm], sh4r.r[Rn] );
381 :}
382 MOV.L @(R0, Rm), Rn {: CHECKRALIGN32( R0 + sh4r.r[Rm] );
383 MEM_READ_LONG( R0 + sh4r.r[Rm], sh4r.r[Rn] );
384 :}
385 MOV.L Rm, @(disp, Rn) {:
386 tmp = sh4r.r[Rn] + disp;
387 CHECKWALIGN32( tmp );
388 MEM_WRITE_LONG( tmp, sh4r.r[Rm] );
389 :}
390 MOV.B Rm, @Rn {: MEM_WRITE_BYTE( sh4r.r[Rn], sh4r.r[Rm] ); :}
391 MOV.W Rm, @Rn {: CHECKWALIGN16( sh4r.r[Rn] ); MEM_WRITE_WORD( sh4r.r[Rn], sh4r.r[Rm] ); :}
392 MOV.L Rm, @Rn {: CHECKWALIGN32( sh4r.r[Rn] ); MEM_WRITE_LONG( sh4r.r[Rn], sh4r.r[Rm] ); :}
393 MOV.B Rm, @-Rn {: MEM_WRITE_BYTE( sh4r.r[Rn]-1, sh4r.r[Rm] ); sh4r.r[Rn]--; :}
394 MOV.W Rm, @-Rn {: CHECKWALIGN16( sh4r.r[Rn] ); MEM_WRITE_WORD( sh4r.r[Rn]-2, sh4r.r[Rm] ); sh4r.r[Rn] -= 2; :}
395 MOV.L Rm, @-Rn {: CHECKWALIGN32( sh4r.r[Rn] ); MEM_WRITE_LONG( sh4r.r[Rn]-4, sh4r.r[Rm] ); sh4r.r[Rn] -= 4; :}
396 MOV.L @(disp, Rm), Rn {:
397 tmp = sh4r.r[Rm] + disp;
398 CHECKRALIGN32( tmp );
399 MEM_READ_LONG( tmp, sh4r.r[Rn] );
400 :}
401 MOV.B @Rm, Rn {: MEM_READ_BYTE( sh4r.r[Rm], sh4r.r[Rn] ); :}
402 MOV.W @Rm, Rn {: CHECKRALIGN16( sh4r.r[Rm] ); MEM_READ_WORD( sh4r.r[Rm], sh4r.r[Rn] ); :}
403 MOV.L @Rm, Rn {: CHECKRALIGN32( sh4r.r[Rm] ); MEM_READ_LONG( sh4r.r[Rm], sh4r.r[Rn] ); :}
404 MOV Rm, Rn {: sh4r.r[Rn] = sh4r.r[Rm]; :}
405 MOV.B @Rm+, Rn {: MEM_READ_BYTE( sh4r.r[Rm], sh4r.r[Rn] ); sh4r.r[Rm] ++; :}
406 MOV.W @Rm+, Rn {: CHECKRALIGN16( sh4r.r[Rm] ); MEM_READ_WORD( sh4r.r[Rm], sh4r.r[Rn] ); sh4r.r[Rm] += 2; :}
407 MOV.L @Rm+, Rn {: CHECKRALIGN32( sh4r.r[Rm] ); MEM_READ_LONG( sh4r.r[Rm], sh4r.r[Rn] ); sh4r.r[Rm] += 4; :}
408 MOV.L @(disp, PC), Rn {:
409 CHECKSLOTILLEGAL();
410 tmp = (pc&0xFFFFFFFC) + disp + 4;
411 MEM_READ_LONG( tmp, sh4r.r[Rn] );
412 :}
413 MOV.B R0, @(disp, GBR) {: MEM_WRITE_BYTE( sh4r.gbr + disp, R0 ); :}
414 MOV.W R0, @(disp, GBR) {:
415 tmp = sh4r.gbr + disp;
416 CHECKWALIGN16( tmp );
417 MEM_WRITE_WORD( tmp, R0 );
418 :}
419 MOV.L R0, @(disp, GBR) {:
420 tmp = sh4r.gbr + disp;
421 CHECKWALIGN32( tmp );
422 MEM_WRITE_LONG( tmp, R0 );
423 :}
424 MOV.B @(disp, GBR), R0 {: MEM_READ_BYTE( sh4r.gbr + disp, R0 ); :}
425 MOV.W @(disp, GBR), R0 {:
426 tmp = sh4r.gbr + disp;
427 CHECKRALIGN16( tmp );
428 MEM_READ_WORD( tmp, R0 );
429 :}
430 MOV.L @(disp, GBR), R0 {:
431 tmp = sh4r.gbr + disp;
432 CHECKRALIGN32( tmp );
433 MEM_READ_LONG( tmp, R0 );
434 :}
435 MOV.B R0, @(disp, Rn) {: MEM_WRITE_BYTE( sh4r.r[Rn] + disp, R0 ); :}
436 MOV.W R0, @(disp, Rn) {:
437 tmp = sh4r.r[Rn] + disp;
438 CHECKWALIGN16( tmp );
439 MEM_WRITE_WORD( tmp, R0 );
440 :}
441 MOV.B @(disp, Rm), R0 {: MEM_READ_BYTE( sh4r.r[Rm] + disp, R0 ); :}
442 MOV.W @(disp, Rm), R0 {:
443 tmp = sh4r.r[Rm] + disp;
444 CHECKRALIGN16( tmp );
445 MEM_READ_WORD( tmp, R0 );
446 :}
447 MOV.W @(disp, PC), Rn {:
448 CHECKSLOTILLEGAL();
449 tmp = pc + 4 + disp;
450 MEM_READ_WORD( tmp, sh4r.r[Rn] );
451 :}
452 MOVA @(disp, PC), R0 {:
453 CHECKSLOTILLEGAL();
454 R0 = (pc&0xFFFFFFFC) + disp + 4;
455 :}
456 MOV #imm, Rn {: sh4r.r[Rn] = imm; :}
458 FMOV @(R0, Rm), FRn {: MEM_FP_READ( sh4r.r[Rm] + R0, FRn ); :}
459 FMOV FRm, @(R0, Rn) {: MEM_FP_WRITE( sh4r.r[Rn] + R0, FRm ); :}
460 FMOV @Rm, FRn {: MEM_FP_READ( sh4r.r[Rm], FRn ); :}
461 FMOV @Rm+, FRn {: MEM_FP_READ( sh4r.r[Rm], FRn ); sh4r.r[Rm] += FP_WIDTH; :}
462 FMOV FRm, @Rn {: MEM_FP_WRITE( sh4r.r[Rn], FRm ); :}
463 FMOV FRm, @-Rn {: MEM_FP_WRITE( sh4r.r[Rn] - FP_WIDTH, FRm ); sh4r.r[Rn] -= FP_WIDTH; :}
464 FMOV FRm, FRn {:
465 if( IS_FPU_DOUBLESIZE() )
466 DR(FRn) = DR(FRm);
467 else
468 FR(FRn) = FR(FRm);
469 :}
471 CMP/EQ #imm, R0 {: sh4r.t = ( R0 == imm ? 1 : 0 ); :}
472 CMP/EQ Rm, Rn {: sh4r.t = ( sh4r.r[Rm] == sh4r.r[Rn] ? 1 : 0 ); :}
473 CMP/GE Rm, Rn {: sh4r.t = ( ((int32_t)sh4r.r[Rn]) >= ((int32_t)sh4r.r[Rm]) ? 1 : 0 ); :}
474 CMP/GT Rm, Rn {: sh4r.t = ( ((int32_t)sh4r.r[Rn]) > ((int32_t)sh4r.r[Rm]) ? 1 : 0 ); :}
475 CMP/HI Rm, Rn {: sh4r.t = ( sh4r.r[Rn] > sh4r.r[Rm] ? 1 : 0 ); :}
476 CMP/HS Rm, Rn {: sh4r.t = ( sh4r.r[Rn] >= sh4r.r[Rm] ? 1 : 0 ); :}
477 CMP/PL Rn {: sh4r.t = ( ((int32_t)sh4r.r[Rn]) > 0 ? 1 : 0 ); :}
478 CMP/PZ Rn {: sh4r.t = ( ((int32_t)sh4r.r[Rn]) >= 0 ? 1 : 0 ); :}
479 CMP/STR Rm, Rn {:
480 /* set T = 1 if any byte in RM & RN is the same */
481 tmp = sh4r.r[Rm] ^ sh4r.r[Rn];
482 sh4r.t = ((tmp&0x000000FF)==0 || (tmp&0x0000FF00)==0 ||
483 (tmp&0x00FF0000)==0 || (tmp&0xFF000000)==0)?1:0;
484 :}
486 ADD Rm, Rn {: sh4r.r[Rn] += sh4r.r[Rm]; :}
487 ADD #imm, Rn {: sh4r.r[Rn] += imm; :}
488 ADDC Rm, Rn {:
489 tmp = sh4r.r[Rn];
490 sh4r.r[Rn] += sh4r.r[Rm] + sh4r.t;
491 sh4r.t = ( sh4r.r[Rn] < tmp || (sh4r.r[Rn] == tmp && sh4r.t != 0) ? 1 : 0 );
492 :}
493 ADDV Rm, Rn {:
494 tmp = sh4r.r[Rn] + sh4r.r[Rm];
495 sh4r.t = ( (sh4r.r[Rn]>>31) == (sh4r.r[Rm]>>31) && ((sh4r.r[Rn]>>31) != (tmp>>31)) );
496 sh4r.r[Rn] = tmp;
497 :}
498 DIV0U {: sh4r.m = sh4r.q = sh4r.t = 0; :}
499 DIV0S Rm, Rn {:
500 sh4r.q = sh4r.r[Rn]>>31;
501 sh4r.m = sh4r.r[Rm]>>31;
502 sh4r.t = sh4r.q ^ sh4r.m;
503 :}
504 DIV1 Rm, Rn {:
505 /* This is derived from the sh4 manual with some simplifications */
506 uint32_t tmp0, tmp1, tmp2, dir;
508 dir = sh4r.q ^ sh4r.m;
509 sh4r.q = (sh4r.r[Rn] >> 31);
510 tmp2 = sh4r.r[Rm];
511 sh4r.r[Rn] = (sh4r.r[Rn] << 1) | sh4r.t;
512 tmp0 = sh4r.r[Rn];
513 if( dir ) {
514 sh4r.r[Rn] += tmp2;
515 tmp1 = (sh4r.r[Rn]<tmp0 ? 1 : 0 );
516 } else {
517 sh4r.r[Rn] -= tmp2;
518 tmp1 = (sh4r.r[Rn]>tmp0 ? 1 : 0 );
519 }
520 sh4r.q ^= sh4r.m ^ tmp1;
521 sh4r.t = ( sh4r.q == sh4r.m ? 1 : 0 );
522 :}
523 DMULS.L Rm, Rn {: sh4r.mac = SIGNEXT32(sh4r.r[Rm]) * SIGNEXT32(sh4r.r[Rn]); :}
524 DMULU.L Rm, Rn {: sh4r.mac = ((uint64_t)sh4r.r[Rm]) * ((uint64_t)sh4r.r[Rn]); :}
525 DT Rn {:
526 sh4r.r[Rn] --;
527 sh4r.t = ( sh4r.r[Rn] == 0 ? 1 : 0 );
528 :}
529 MAC.W @Rm+, @Rn+ {:
530 int32_t stmp;
531 if( Rm == Rn ) {
532 CHECKRALIGN16(sh4r.r[Rn]);
533 MEM_READ_WORD( sh4r.r[Rn], tmp );
534 stmp = SIGNEXT16(tmp);
535 MEM_READ_WORD( sh4r.r[Rn]+2, tmp );
536 stmp *= SIGNEXT16(tmp);
537 sh4r.r[Rn] += 4;
538 } else {
539 CHECKRALIGN16( sh4r.r[Rn] );
540 CHECKRALIGN16( sh4r.r[Rm] );
541 MEM_READ_WORD(sh4r.r[Rn], tmp);
542 stmp = SIGNEXT16(tmp);
543 MEM_READ_WORD(sh4r.r[Rm], tmp);
544 stmp = stmp * SIGNEXT16(tmp);
545 sh4r.r[Rn] += 2;
546 sh4r.r[Rm] += 2;
547 }
548 if( sh4r.s ) {
549 int64_t tmpl = (int64_t)((int32_t)sh4r.mac) + (int64_t)stmp;
550 if( tmpl > (int64_t)0x000000007FFFFFFFLL ) {
551 sh4r.mac = 0x000000017FFFFFFFLL;
552 } else if( tmpl < (int64_t)0xFFFFFFFF80000000LL ) {
553 sh4r.mac = 0x0000000180000000LL;
554 } else {
555 sh4r.mac = (sh4r.mac & 0xFFFFFFFF00000000LL) |
556 ((uint32_t)(sh4r.mac + stmp));
557 }
558 } else {
559 sh4r.mac += SIGNEXT32(stmp);
560 }
561 :}
562 MAC.L @Rm+, @Rn+ {:
563 int64_t tmpl;
564 if( Rm == Rn ) {
565 CHECKRALIGN32( sh4r.r[Rn] );
566 MEM_READ_LONG(sh4r.r[Rn], tmp);
567 tmpl = SIGNEXT32(tmp);
568 MEM_READ_LONG(sh4r.r[Rn]+4, tmp);
569 tmpl = tmpl * SIGNEXT32(tmp) + sh4r.mac;
570 sh4r.r[Rn] += 8;
571 } else {
572 CHECKRALIGN32( sh4r.r[Rm] );
573 CHECKRALIGN32( sh4r.r[Rn] );
574 MEM_READ_LONG(sh4r.r[Rn], tmp);
575 tmpl = SIGNEXT32(tmp);
576 MEM_READ_LONG(sh4r.r[Rm], tmp);
577 tmpl = tmpl * SIGNEXT32(tmp) + sh4r.mac;
578 sh4r.r[Rn] += 4;
579 sh4r.r[Rm] += 4;
580 }
581 if( sh4r.s ) {
582 /* 48-bit Saturation. Yuch */
583 if( tmpl < (int64_t)0xFFFF800000000000LL )
584 tmpl = 0xFFFF800000000000LL;
585 else if( tmpl > (int64_t)0x00007FFFFFFFFFFFLL )
586 tmpl = 0x00007FFFFFFFFFFFLL;
587 }
588 sh4r.mac = tmpl;
589 :}
590 MUL.L Rm, Rn {: sh4r.mac = (sh4r.mac&0xFFFFFFFF00000000LL) |
591 (sh4r.r[Rm] * sh4r.r[Rn]); :}
592 MULU.W Rm, Rn {:
593 sh4r.mac = (sh4r.mac&0xFFFFFFFF00000000LL) |
594 (uint32_t)((sh4r.r[Rm]&0xFFFF) * (sh4r.r[Rn]&0xFFFF));
595 :}
596 MULS.W Rm, Rn {:
597 sh4r.mac = (sh4r.mac&0xFFFFFFFF00000000LL) |
598 (uint32_t)(SIGNEXT32(sh4r.r[Rm]&0xFFFF) * SIGNEXT32(sh4r.r[Rn]&0xFFFF));
599 :}
600 NEGC Rm, Rn {:
601 tmp = 0 - sh4r.r[Rm];
602 sh4r.r[Rn] = tmp - sh4r.t;
603 sh4r.t = ( 0<tmp || tmp<sh4r.r[Rn] ? 1 : 0 );
604 :}
605 NEG Rm, Rn {: sh4r.r[Rn] = 0 - sh4r.r[Rm]; :}
606 SUB Rm, Rn {: sh4r.r[Rn] -= sh4r.r[Rm]; :}
607 SUBC Rm, Rn {:
608 tmp = sh4r.r[Rn];
609 sh4r.r[Rn] = sh4r.r[Rn] - sh4r.r[Rm] - sh4r.t;
610 sh4r.t = (sh4r.r[Rn] > tmp || (sh4r.r[Rn] == tmp && sh4r.t == 1));
611 :}
613 BRAF Rn {:
614 CHECKSLOTILLEGAL();
615 CHECKDEST( pc + 4 + sh4r.r[Rn] );
616 sh4r.in_delay_slot = 1;
617 sh4r.pc = sh4r.new_pc;
618 sh4r.new_pc = pc + 4 + sh4r.r[Rn];
619 return TRUE;
620 :}
621 BSRF Rn {:
622 CHECKSLOTILLEGAL();
623 CHECKDEST( pc + 4 + sh4r.r[Rn] );
624 sh4r.in_delay_slot = 1;
625 sh4r.pr = sh4r.pc + 4;
626 sh4r.pc = sh4r.new_pc;
627 sh4r.new_pc = pc + 4 + sh4r.r[Rn];
628 TRACE_CALL( pc, sh4r.new_pc );
629 return TRUE;
630 :}
631 BT disp {:
632 CHECKSLOTILLEGAL();
633 if( sh4r.t ) {
634 CHECKDEST( sh4r.pc + disp + 4 )
635 sh4r.pc += disp + 4;
636 sh4r.new_pc = sh4r.pc + 2;
637 return TRUE;
638 }
639 :}
640 BF disp {:
641 CHECKSLOTILLEGAL();
642 if( !sh4r.t ) {
643 CHECKDEST( sh4r.pc + disp + 4 )
644 sh4r.pc += disp + 4;
645 sh4r.new_pc = sh4r.pc + 2;
646 return TRUE;
647 }
648 :}
649 BT/S disp {:
650 CHECKSLOTILLEGAL();
651 if( sh4r.t ) {
652 CHECKDEST( sh4r.pc + disp + 4 )
653 sh4r.in_delay_slot = 1;
654 sh4r.pc = sh4r.new_pc;
655 sh4r.new_pc = pc + disp + 4;
656 sh4r.in_delay_slot = 1;
657 return TRUE;
658 }
659 :}
660 BF/S disp {:
661 CHECKSLOTILLEGAL();
662 if( !sh4r.t ) {
663 CHECKDEST( sh4r.pc + disp + 4 )
664 sh4r.in_delay_slot = 1;
665 sh4r.pc = sh4r.new_pc;
666 sh4r.new_pc = pc + disp + 4;
667 return TRUE;
668 }
669 :}
670 BRA disp {:
671 CHECKSLOTILLEGAL();
672 CHECKDEST( sh4r.pc + disp + 4 );
673 sh4r.in_delay_slot = 1;
674 sh4r.pc = sh4r.new_pc;
675 sh4r.new_pc = pc + 4 + disp;
676 return TRUE;
677 :}
678 BSR disp {:
679 CHECKDEST( sh4r.pc + disp + 4 );
680 CHECKSLOTILLEGAL();
681 sh4r.in_delay_slot = 1;
682 sh4r.pr = pc + 4;
683 sh4r.pc = sh4r.new_pc;
684 sh4r.new_pc = pc + 4 + disp;
685 TRACE_CALL( pc, sh4r.new_pc );
686 return TRUE;
687 :}
688 TRAPA #imm {:
689 CHECKSLOTILLEGAL();
690 sh4r.pc += 2;
691 sh4_raise_trap( imm );
692 return TRUE;
693 :}
694 RTS {:
695 CHECKSLOTILLEGAL();
696 CHECKDEST( sh4r.pr );
697 sh4r.in_delay_slot = 1;
698 sh4r.pc = sh4r.new_pc;
699 sh4r.new_pc = sh4r.pr;
700 TRACE_RETURN( pc, sh4r.new_pc );
701 return TRUE;
702 :}
703 SLEEP {:
704 if( MMIO_READ( CPG, STBCR ) & 0x80 ) {
705 sh4r.sh4_state = SH4_STATE_STANDBY;
706 } else {
707 sh4r.sh4_state = SH4_STATE_SLEEP;
708 }
709 return FALSE; /* Halt CPU */
710 :}
711 RTE {:
712 CHECKPRIV();
713 CHECKDEST( sh4r.spc );
714 CHECKSLOTILLEGAL();
715 sh4r.in_delay_slot = 1;
716 sh4r.pc = sh4r.new_pc;
717 sh4r.new_pc = sh4r.spc;
718 sh4_write_sr( sh4r.ssr );
719 return TRUE;
720 :}
721 JMP @Rn {:
722 CHECKDEST( sh4r.r[Rn] );
723 CHECKSLOTILLEGAL();
724 sh4r.in_delay_slot = 1;
725 sh4r.pc = sh4r.new_pc;
726 sh4r.new_pc = sh4r.r[Rn];
727 return TRUE;
728 :}
729 JSR @Rn {:
730 CHECKDEST( sh4r.r[Rn] );
731 CHECKSLOTILLEGAL();
732 sh4r.in_delay_slot = 1;
733 sh4r.pc = sh4r.new_pc;
734 sh4r.new_pc = sh4r.r[Rn];
735 sh4r.pr = pc + 4;
736 TRACE_CALL( pc, sh4r.new_pc );
737 return TRUE;
738 :}
739 STS MACH, Rn {: sh4r.r[Rn] = (sh4r.mac>>32); :}
740 STS.L MACH, @-Rn {:
741 CHECKWALIGN32( sh4r.r[Rn] );
742 MEM_WRITE_LONG( sh4r.r[Rn]-4, (sh4r.mac>>32) );
743 sh4r.r[Rn] -= 4;
744 :}
745 STC.L SR, @-Rn {:
746 CHECKPRIV();
747 CHECKWALIGN32( sh4r.r[Rn] );
748 MEM_WRITE_LONG( sh4r.r[Rn]-4, sh4_read_sr() );
749 sh4r.r[Rn] -= 4;
750 :}
751 LDS.L @Rm+, MACH {:
752 CHECKRALIGN32( sh4r.r[Rm] );
753 MEM_READ_LONG(sh4r.r[Rm], tmp);
754 sh4r.mac = (sh4r.mac & 0x00000000FFFFFFFF) |
755 (((uint64_t)tmp)<<32);
756 sh4r.r[Rm] += 4;
757 :}
758 LDC.L @Rm+, SR {:
759 CHECKSLOTILLEGAL();
760 CHECKPRIV();
761 CHECKWALIGN32( sh4r.r[Rm] );
762 MEM_READ_LONG(sh4r.r[Rm], tmp);
763 sh4_write_sr( tmp );
764 sh4r.r[Rm] +=4;
765 :}
766 LDS Rm, MACH {:
767 sh4r.mac = (sh4r.mac & 0x00000000FFFFFFFF) |
768 (((uint64_t)sh4r.r[Rm])<<32);
769 :}
770 LDC Rm, SR {:
771 CHECKSLOTILLEGAL();
772 CHECKPRIV();
773 sh4_write_sr( sh4r.r[Rm] );
774 :}
775 LDC Rm, SGR {:
776 CHECKPRIV();
777 sh4r.sgr = sh4r.r[Rm];
778 :}
779 LDC.L @Rm+, SGR {:
780 CHECKPRIV();
781 CHECKRALIGN32( sh4r.r[Rm] );
782 MEM_READ_LONG(sh4r.r[Rm], sh4r.sgr);
783 sh4r.r[Rm] +=4;
784 :}
785 STS MACL, Rn {: sh4r.r[Rn] = (uint32_t)sh4r.mac; :}
786 STS.L MACL, @-Rn {:
787 CHECKWALIGN32( sh4r.r[Rn] );
788 MEM_WRITE_LONG( sh4r.r[Rn]-4, (uint32_t)sh4r.mac );
789 sh4r.r[Rn] -= 4;
790 :}
791 STC.L GBR, @-Rn {:
792 CHECKWALIGN32( sh4r.r[Rn] );
793 MEM_WRITE_LONG( sh4r.r[Rn]-4, sh4r.gbr );
794 sh4r.r[Rn] -= 4;
795 :}
796 LDS.L @Rm+, MACL {:
797 CHECKRALIGN32( sh4r.r[Rm] );
798 MEM_READ_LONG(sh4r.r[Rm], tmp);
799 sh4r.mac = (sh4r.mac & 0xFFFFFFFF00000000LL) |
800 (uint64_t)((uint32_t)tmp);
801 sh4r.r[Rm] += 4;
802 :}
803 LDC.L @Rm+, GBR {:
804 CHECKRALIGN32( sh4r.r[Rm] );
805 MEM_READ_LONG(sh4r.r[Rm], sh4r.gbr);
806 sh4r.r[Rm] +=4;
807 :}
808 LDS Rm, MACL {:
809 sh4r.mac = (sh4r.mac & 0xFFFFFFFF00000000LL) |
810 (uint64_t)((uint32_t)(sh4r.r[Rm]));
811 :}
812 LDC Rm, GBR {: sh4r.gbr = sh4r.r[Rm]; :}
813 STS PR, Rn {: sh4r.r[Rn] = sh4r.pr; :}
814 STS.L PR, @-Rn {:
815 CHECKWALIGN32( sh4r.r[Rn] );
816 MEM_WRITE_LONG( sh4r.r[Rn]-4, sh4r.pr );
817 sh4r.r[Rn] -= 4;
818 :}
819 STC.L VBR, @-Rn {:
820 CHECKPRIV();
821 CHECKWALIGN32( sh4r.r[Rn] );
822 MEM_WRITE_LONG( sh4r.r[Rn]-4, sh4r.vbr );
823 sh4r.r[Rn] -= 4;
824 :}
825 LDS.L @Rm+, PR {:
826 CHECKRALIGN32( sh4r.r[Rm] );
827 MEM_READ_LONG( sh4r.r[Rm], sh4r.pr );
828 sh4r.r[Rm] += 4;
829 :}
830 LDC.L @Rm+, VBR {:
831 CHECKPRIV();
832 CHECKRALIGN32( sh4r.r[Rm] );
833 MEM_READ_LONG(sh4r.r[Rm], sh4r.vbr);
834 sh4r.r[Rm] +=4;
835 :}
836 LDS Rm, PR {: sh4r.pr = sh4r.r[Rm]; :}
837 LDC Rm, VBR {:
838 CHECKPRIV();
839 sh4r.vbr = sh4r.r[Rm];
840 :}
841 STC SGR, Rn {:
842 CHECKPRIV();
843 sh4r.r[Rn] = sh4r.sgr;
844 :}
845 STC.L SGR, @-Rn {:
846 CHECKPRIV();
847 CHECKWALIGN32( sh4r.r[Rn] );
848 MEM_WRITE_LONG( sh4r.r[Rn]-4, sh4r.sgr );
849 sh4r.r[Rn] -= 4;
850 :}
851 STC.L SSR, @-Rn {:
852 CHECKPRIV();
853 CHECKWALIGN32( sh4r.r[Rn] );
854 MEM_WRITE_LONG( sh4r.r[Rn]-4, sh4r.ssr );
855 sh4r.r[Rn] -= 4;
856 :}
857 LDC.L @Rm+, SSR {:
858 CHECKPRIV();
859 CHECKRALIGN32( sh4r.r[Rm] );
860 MEM_READ_LONG(sh4r.r[Rm], sh4r.ssr);
861 sh4r.r[Rm] +=4;
862 :}
863 LDC Rm, SSR {:
864 CHECKPRIV();
865 sh4r.ssr = sh4r.r[Rm];
866 :}
867 STC.L SPC, @-Rn {:
868 CHECKPRIV();
869 CHECKWALIGN32( sh4r.r[Rn] );
870 MEM_WRITE_LONG( sh4r.r[Rn]-4, sh4r.spc );
871 sh4r.r[Rn] -= 4;
872 :}
873 LDC.L @Rm+, SPC {:
874 CHECKPRIV();
875 CHECKRALIGN32( sh4r.r[Rm] );
876 MEM_READ_LONG(sh4r.r[Rm], sh4r.spc);
877 sh4r.r[Rm] +=4;
878 :}
879 LDC Rm, SPC {:
880 CHECKPRIV();
881 sh4r.spc = sh4r.r[Rm];
882 :}
883 STS FPUL, Rn {:
884 CHECKFPUEN();
885 sh4r.r[Rn] = FPULi;
886 :}
887 STS.L FPUL, @-Rn {:
888 CHECKFPUEN();
889 CHECKWALIGN32( sh4r.r[Rn] );
890 MEM_WRITE_LONG( sh4r.r[Rn]-4, FPULi );
891 sh4r.r[Rn] -= 4;
892 :}
893 LDS.L @Rm+, FPUL {:
894 CHECKFPUEN();
895 CHECKRALIGN32( sh4r.r[Rm] );
896 MEM_READ_LONG(sh4r.r[Rm], FPULi);
897 sh4r.r[Rm] +=4;
898 :}
899 LDS Rm, FPUL {:
900 CHECKFPUEN();
901 FPULi = sh4r.r[Rm];
902 :}
903 STS FPSCR, Rn {:
904 CHECKFPUEN();
905 sh4r.r[Rn] = sh4r.fpscr;
906 :}
907 STS.L FPSCR, @-Rn {:
908 CHECKFPUEN();
909 CHECKWALIGN32( sh4r.r[Rn] );
910 MEM_WRITE_LONG( sh4r.r[Rn]-4, sh4r.fpscr );
911 sh4r.r[Rn] -= 4;
912 :}
913 LDS.L @Rm+, FPSCR {:
914 CHECKFPUEN();
915 CHECKRALIGN32( sh4r.r[Rm] );
916 MEM_READ_LONG(sh4r.r[Rm], tmp);
917 sh4r.r[Rm] +=4;
918 sh4_write_fpscr( tmp );
919 :}
920 LDS Rm, FPSCR {:
921 CHECKFPUEN();
922 sh4_write_fpscr( sh4r.r[Rm] );
923 :}
924 STC DBR, Rn {: CHECKPRIV(); sh4r.r[Rn] = sh4r.dbr; :}
925 STC.L DBR, @-Rn {:
926 CHECKPRIV();
927 CHECKWALIGN32( sh4r.r[Rn] );
928 MEM_WRITE_LONG( sh4r.r[Rn]-4, sh4r.dbr );
929 sh4r.r[Rn] -= 4;
930 :}
931 LDC.L @Rm+, DBR {:
932 CHECKPRIV();
933 CHECKRALIGN32( sh4r.r[Rm] );
934 MEM_READ_LONG(sh4r.r[Rm], sh4r.dbr);
935 sh4r.r[Rm] +=4;
936 :}
937 LDC Rm, DBR {:
938 CHECKPRIV();
939 sh4r.dbr = sh4r.r[Rm];
940 :}
941 STC.L Rm_BANK, @-Rn {:
942 CHECKPRIV();
943 CHECKWALIGN32( sh4r.r[Rn] );
944 MEM_WRITE_LONG( sh4r.r[Rn]-4, sh4r.r_bank[Rm_BANK] );
945 sh4r.r[Rn] -= 4;
946 :}
947 LDC.L @Rm+, Rn_BANK {:
948 CHECKPRIV();
949 CHECKRALIGN32( sh4r.r[Rm] );
950 MEM_READ_LONG( sh4r.r[Rm], sh4r.r_bank[Rn_BANK] );
951 sh4r.r[Rm] += 4;
952 :}
953 LDC Rm, Rn_BANK {:
954 CHECKPRIV();
955 sh4r.r_bank[Rn_BANK] = sh4r.r[Rm];
956 :}
957 STC SR, Rn {:
958 CHECKPRIV();
959 sh4r.r[Rn] = sh4_read_sr();
960 :}
961 STC GBR, Rn {:
962 sh4r.r[Rn] = sh4r.gbr;
963 :}
964 STC VBR, Rn {:
965 CHECKPRIV();
966 sh4r.r[Rn] = sh4r.vbr;
967 :}
968 STC SSR, Rn {:
969 CHECKPRIV();
970 sh4r.r[Rn] = sh4r.ssr;
971 :}
972 STC SPC, Rn {:
973 CHECKPRIV();
974 sh4r.r[Rn] = sh4r.spc;
975 :}
976 STC Rm_BANK, Rn {:
977 CHECKPRIV();
978 sh4r.r[Rn] = sh4r.r_bank[Rm_BANK];
979 :}
981 FADD FRm, FRn {:
982 CHECKFPUEN();
983 if( IS_FPU_DOUBLEPREC() ) {
984 DR(FRn) += DR(FRm);
985 } else {
986 FR(FRn) += FR(FRm);
987 }
988 :}
989 FSUB FRm, FRn {:
990 CHECKFPUEN();
991 if( IS_FPU_DOUBLEPREC() ) {
992 DR(FRn) -= DR(FRm);
993 } else {
994 FR(FRn) -= FR(FRm);
995 }
996 :}
998 FMUL FRm, FRn {:
999 CHECKFPUEN();
1000 if( IS_FPU_DOUBLEPREC() ) {
1001 DR(FRn) *= DR(FRm);
1002 } else {
1003 FR(FRn) *= FR(FRm);
1004 }
1005 :}
1007 FDIV FRm, FRn {:
1008 CHECKFPUEN();
1009 if( IS_FPU_DOUBLEPREC() ) {
1010 DR(FRn) /= DR(FRm);
1011 } else {
1012 FR(FRn) /= FR(FRm);
1013 }
1014 :}
1016 FCMP/EQ FRm, FRn {:
1017 CHECKFPUEN();
1018 if( IS_FPU_DOUBLEPREC() ) {
1019 sh4r.t = ( DR(FRn) == DR(FRm) ? 1 : 0 );
1020 } else {
1021 sh4r.t = ( FR(FRn) == FR(FRm) ? 1 : 0 );
1022 }
1023 :}
1025 FCMP/GT FRm, FRn {:
1026 CHECKFPUEN();
1027 if( IS_FPU_DOUBLEPREC() ) {
1028 sh4r.t = ( DR(FRn) > DR(FRm) ? 1 : 0 );
1029 } else {
1030 sh4r.t = ( FR(FRn) > FR(FRm) ? 1 : 0 );
1031 }
1032 :}
1034 FSTS FPUL, FRn {: CHECKFPUEN(); FR(FRn) = FPULf; :}
1035 FLDS FRm, FPUL {: CHECKFPUEN(); FPULf = FR(FRm); :}
1036 FLOAT FPUL, FRn {:
1037 CHECKFPUEN();
1038 if( IS_FPU_DOUBLEPREC() ) {
1039 if( FRn&1 ) { // No, really...
1040 dtmp = (double)FPULi;
1041 FR(FRn) = *(((float *)&dtmp)+1);
1042 } else {
1043 DRF(FRn>>1) = (double)FPULi;
1044 }
1045 } else {
1046 FR(FRn) = (float)FPULi;
1047 }
1048 :}
1049 FTRC FRm, FPUL {:
1050 CHECKFPUEN();
1051 if( IS_FPU_DOUBLEPREC() ) {
1052 if( FRm&1 ) {
1053 dtmp = 0;
1054 *(((float *)&dtmp)+1) = FR(FRm);
1055 } else {
1056 dtmp = DRF(FRm>>1);
1057 }
1058 if( dtmp >= MAX_INTF )
1059 FPULi = MAX_INT;
1060 else if( dtmp <= MIN_INTF )
1061 FPULi = MIN_INT;
1062 else
1063 FPULi = (int32_t)dtmp;
1064 } else {
1065 ftmp = FR(FRm);
1066 if( ftmp >= MAX_INTF )
1067 FPULi = MAX_INT;
1068 else if( ftmp <= MIN_INTF )
1069 FPULi = MIN_INT;
1070 else
1071 FPULi = (int32_t)ftmp;
1072 }
1073 :}
1074 FNEG FRn {:
1075 CHECKFPUEN();
1076 if( IS_FPU_DOUBLEPREC() ) {
1077 DR(FRn) = -DR(FRn);
1078 } else {
1079 FR(FRn) = -FR(FRn);
1080 }
1081 :}
1082 FABS FRn {:
1083 CHECKFPUEN();
1084 if( IS_FPU_DOUBLEPREC() ) {
1085 DR(FRn) = fabs(DR(FRn));
1086 } else {
1087 FR(FRn) = fabsf(FR(FRn));
1088 }
1089 :}
1090 FSQRT FRn {:
1091 CHECKFPUEN();
1092 if( IS_FPU_DOUBLEPREC() ) {
1093 DR(FRn) = sqrt(DR(FRn));
1094 } else {
1095 FR(FRn) = sqrtf(FR(FRn));
1096 }
1097 :}
1098 FLDI0 FRn {:
1099 CHECKFPUEN();
1100 if( IS_FPU_DOUBLEPREC() ) {
1101 DR(FRn) = 0.0;
1102 } else {
1103 FR(FRn) = 0.0;
1104 }
1105 :}
1106 FLDI1 FRn {:
1107 CHECKFPUEN();
1108 if( IS_FPU_DOUBLEPREC() ) {
1109 DR(FRn) = 1.0;
1110 } else {
1111 FR(FRn) = 1.0;
1112 }
1113 :}
1114 FMAC FR0, FRm, FRn {:
1115 CHECKFPUEN();
1116 if( IS_FPU_DOUBLEPREC() ) {
1117 DR(FRn) += DR(FRm)*DR(0);
1118 } else {
1119 FR(FRn) += FR(FRm)*FR(0);
1120 }
1121 :}
1122 FRCHG {:
1123 CHECKFPUEN();
1124 sh4r.fpscr ^= FPSCR_FR;
1125 sh4_switch_fr_banks();
1126 :}
1127 FSCHG {: CHECKFPUEN(); sh4r.fpscr ^= FPSCR_SZ; :}
1128 FCNVSD FPUL, FRn {:
1129 CHECKFPUEN();
1130 if( IS_FPU_DOUBLEPREC() && !IS_FPU_DOUBLESIZE() ) {
1131 DR(FRn) = (double)FPULf;
1132 }
1133 :}
1134 FCNVDS FRm, FPUL {:
1135 CHECKFPUEN();
1136 if( IS_FPU_DOUBLEPREC() && !IS_FPU_DOUBLESIZE() ) {
1137 FPULf = (float)DR(FRm);
1138 }
1139 :}
1141 FSRRA FRn {:
1142 CHECKFPUEN();
1143 if( !IS_FPU_DOUBLEPREC() ) {
1144 FR(FRn) = 1.0/sqrtf(FR(FRn));
1145 }
1146 :}
1147 FIPR FVm, FVn {:
1148 CHECKFPUEN();
1149 if( !IS_FPU_DOUBLEPREC() ) {
1150 int tmp2 = FVn<<2;
1151 tmp = FVm<<2;
1152 FR(tmp2+3) = FR(tmp)*FR(tmp2) +
1153 FR(tmp+1)*FR(tmp2+1) +
1154 FR(tmp+2)*FR(tmp2+2) +
1155 FR(tmp+3)*FR(tmp2+3);
1156 }
1157 :}
1158 FSCA FPUL, FRn {:
1159 CHECKFPUEN();
1160 if( !IS_FPU_DOUBLEPREC() ) {
1161 sh4_fsca( FPULi, (float *)&(DRF(FRn>>1)) );
1162 /*
1163 float angle = (((float)(FPULi&0xFFFF))/65536.0) * 2 * M_PI;
1164 FR(FRn) = sinf(angle);
1165 FR((FRn)+1) = cosf(angle);
1166 */
1167 }
1168 :}
1169 FTRV XMTRX, FVn {:
1170 CHECKFPUEN();
1171 if( !IS_FPU_DOUBLEPREC() ) {
1172 sh4_ftrv((float *)&(DRF(FVn<<1)) );
1173 }
1174 :}
1175 UNDEF {:
1176 UNDEF(ir);
1177 :}
1178 %%
1179 sh4r.pc = sh4r.new_pc;
1180 sh4r.new_pc += 2;
1181 sh4r.in_delay_slot = 0;
1182 return TRUE;
1183 }
.