1 /*
2 * iwMMXt micro operations for XScale.
3 *
4 * Copyright (c) 2007 OpenedHand, Ltd.
5 * Written by Andrzej Zaborowski <andrew@openedhand.com>
6 * Copyright (c) 2008 CodeSourcery
7 *
8 * This library is free software; you can redistribute it and/or
9 * modify it under the terms of the GNU Lesser General Public
10 * License as published by the Free Software Foundation; either
11 * version 2 of the License, or (at your option) any later version.
12 *
13 * This library is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
16 * Lesser General Public License for more details.
17 *
18 * You should have received a copy of the GNU Lesser General Public
19 * License along with this library; if not, write to the Free Software
20 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston MA 02110-1301 USA
21 */
22
23 #include <stdlib.h>
24 #include <stdio.h>
25
26 #include "cpu.h"
27 #include "exec-all.h"
28 #include "helpers.h"
29
30 /* iwMMXt macros extracted from GNU gdb. */
31
32 /* Set the SIMD wCASF flags for 8, 16, 32 or 64-bit operations. */
33 #define SIMD8_SET( v, n, b) ((v != 0) << ((((b) + 1) * 4) + (n)))
34 #define SIMD16_SET(v, n, h) ((v != 0) << ((((h) + 1) * 8) + (n)))
35 #define SIMD32_SET(v, n, w) ((v != 0) << ((((w) + 1) * 16) + (n)))
36 #define SIMD64_SET(v, n) ((v != 0) << (32 + (n)))
37 /* Flags to pass as "n" above. */
38 #define SIMD_NBIT -1
39 #define SIMD_ZBIT -2
40 #define SIMD_CBIT -3
41 #define SIMD_VBIT -4
42 /* Various status bit macros. */
43 #define NBIT8(x) ((x) & 0x80)
44 #define NBIT16(x) ((x) & 0x8000)
45 #define NBIT32(x) ((x) & 0x80000000)
46 #define NBIT64(x) ((x) & 0x8000000000000000ULL)
47 #define ZBIT8(x) (((x) & 0xff) == 0)
48 #define ZBIT16(x) (((x) & 0xffff) == 0)
49 #define ZBIT32(x) (((x) & 0xffffffff) == 0)
50 #define ZBIT64(x) (x == 0)
51 /* Sign extension macros. */
52 #define EXTEND8H(a) ((uint16_t) (int8_t) (a))
53 #define EXTEND8(a) ((uint32_t) (int8_t) (a))
54 #define EXTEND16(a) ((uint32_t) (int16_t) (a))
55 #define EXTEND16S(a) ((int32_t) (int16_t) (a))
56 #define EXTEND32(a) ((uint64_t) (int32_t) (a))
57
HELPER(iwmmxt_maddsq)58 uint64_t HELPER(iwmmxt_maddsq)(uint64_t a, uint64_t b)
59 {
60 a = ((
61 EXTEND16S((a >> 0) & 0xffff) * EXTEND16S((b >> 0) & 0xffff) +
62 EXTEND16S((a >> 16) & 0xffff) * EXTEND16S((b >> 16) & 0xffff)
63 ) & 0xffffffff) | ((uint64_t) (
64 EXTEND16S((a >> 32) & 0xffff) * EXTEND16S((b >> 32) & 0xffff) +
65 EXTEND16S((a >> 48) & 0xffff) * EXTEND16S((b >> 48) & 0xffff)
66 ) << 32);
67 return a;
68 }
69
HELPER(iwmmxt_madduq)70 uint64_t HELPER(iwmmxt_madduq)(uint64_t a, uint64_t b)
71 {
72 a = ((
73 ((a >> 0) & 0xffff) * ((b >> 0) & 0xffff) +
74 ((a >> 16) & 0xffff) * ((b >> 16) & 0xffff)
75 ) & 0xffffffff) | ((
76 ((a >> 32) & 0xffff) * ((b >> 32) & 0xffff) +
77 ((a >> 48) & 0xffff) * ((b >> 48) & 0xffff)
78 ) << 32);
79 return a;
80 }
81
HELPER(iwmmxt_sadb)82 uint64_t HELPER(iwmmxt_sadb)(uint64_t a, uint64_t b)
83 {
84 #define abs(x) (((x) >= 0) ? x : -x)
85 #define SADB(SHR) abs((int) ((a >> SHR) & 0xff) - (int) ((b >> SHR) & 0xff))
86 return
87 SADB(0) + SADB(8) + SADB(16) + SADB(24) +
88 SADB(32) + SADB(40) + SADB(48) + SADB(56);
89 #undef SADB
90 }
91
HELPER(iwmmxt_sadw)92 uint64_t HELPER(iwmmxt_sadw)(uint64_t a, uint64_t b)
93 {
94 #define SADW(SHR) \
95 abs((int) ((a >> SHR) & 0xffff) - (int) ((b >> SHR) & 0xffff))
96 return SADW(0) + SADW(16) + SADW(32) + SADW(48);
97 #undef SADW
98 }
99
HELPER(iwmmxt_mulslw)100 uint64_t HELPER(iwmmxt_mulslw)(uint64_t a, uint64_t b)
101 {
102 #define MULS(SHR) ((uint64_t) ((( \
103 EXTEND16S((a >> SHR) & 0xffff) * EXTEND16S((b >> SHR) & 0xffff) \
104 ) >> 0) & 0xffff) << SHR)
105 return MULS(0) | MULS(16) | MULS(32) | MULS(48);
106 #undef MULS
107 }
108
HELPER(iwmmxt_mulshw)109 uint64_t HELPER(iwmmxt_mulshw)(uint64_t a, uint64_t b)
110 {
111 #define MULS(SHR) ((uint64_t) ((( \
112 EXTEND16S((a >> SHR) & 0xffff) * EXTEND16S((b >> SHR) & 0xffff) \
113 ) >> 16) & 0xffff) << SHR)
114 return MULS(0) | MULS(16) | MULS(32) | MULS(48);
115 #undef MULS
116 }
117
HELPER(iwmmxt_mululw)118 uint64_t HELPER(iwmmxt_mululw)(uint64_t a, uint64_t b)
119 {
120 #define MULU(SHR) ((uint64_t) ((( \
121 ((a >> SHR) & 0xffff) * ((b >> SHR) & 0xffff) \
122 ) >> 0) & 0xffff) << SHR)
123 return MULU(0) | MULU(16) | MULU(32) | MULU(48);
124 #undef MULU
125 }
126
HELPER(iwmmxt_muluhw)127 uint64_t HELPER(iwmmxt_muluhw)(uint64_t a, uint64_t b)
128 {
129 #define MULU(SHR) ((uint64_t) ((( \
130 ((a >> SHR) & 0xffff) * ((b >> SHR) & 0xffff) \
131 ) >> 16) & 0xffff) << SHR)
132 return MULU(0) | MULU(16) | MULU(32) | MULU(48);
133 #undef MULU
134 }
135
HELPER(iwmmxt_macsw)136 uint64_t HELPER(iwmmxt_macsw)(uint64_t a, uint64_t b)
137 {
138 #define MACS(SHR) ( \
139 EXTEND16((a >> SHR) & 0xffff) * EXTEND16S((b >> SHR) & 0xffff))
140 return (int64_t) (MACS(0) + MACS(16) + MACS(32) + MACS(48));
141 #undef MACS
142 }
143
HELPER(iwmmxt_macuw)144 uint64_t HELPER(iwmmxt_macuw)(uint64_t a, uint64_t b)
145 {
146 #define MACU(SHR) ( \
147 (uint32_t) ((a >> SHR) & 0xffff) * \
148 (uint32_t) ((b >> SHR) & 0xffff))
149 return MACU(0) + MACU(16) + MACU(32) + MACU(48);
150 #undef MACU
151 }
152
153 #define NZBIT8(x, i) \
154 SIMD8_SET(NBIT8((x) & 0xff), SIMD_NBIT, i) | \
155 SIMD8_SET(ZBIT8((x) & 0xff), SIMD_ZBIT, i)
156 #define NZBIT16(x, i) \
157 SIMD16_SET(NBIT16((x) & 0xffff), SIMD_NBIT, i) | \
158 SIMD16_SET(ZBIT16((x) & 0xffff), SIMD_ZBIT, i)
159 #define NZBIT32(x, i) \
160 SIMD32_SET(NBIT32((x) & 0xffffffff), SIMD_NBIT, i) | \
161 SIMD32_SET(ZBIT32((x) & 0xffffffff), SIMD_ZBIT, i)
162 #define NZBIT64(x) \
163 SIMD64_SET(NBIT64(x), SIMD_NBIT) | \
164 SIMD64_SET(ZBIT64(x), SIMD_ZBIT)
165 #define IWMMXT_OP_UNPACK(S, SH0, SH1, SH2, SH3) \
166 uint64_t HELPER(glue(iwmmxt_unpack, glue(S, b)))(CPUState *env, \
167 uint64_t a, uint64_t b) \
168 { \
169 a = \
170 (((a >> SH0) & 0xff) << 0) | (((b >> SH0) & 0xff) << 8) | \
171 (((a >> SH1) & 0xff) << 16) | (((b >> SH1) & 0xff) << 24) | \
172 (((a >> SH2) & 0xff) << 32) | (((b >> SH2) & 0xff) << 40) | \
173 (((a >> SH3) & 0xff) << 48) | (((b >> SH3) & 0xff) << 56); \
174 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \
175 NZBIT8(a >> 0, 0) | NZBIT8(a >> 8, 1) | \
176 NZBIT8(a >> 16, 2) | NZBIT8(a >> 24, 3) | \
177 NZBIT8(a >> 32, 4) | NZBIT8(a >> 40, 5) | \
178 NZBIT8(a >> 48, 6) | NZBIT8(a >> 56, 7); \
179 return a; \
180 } \
181 uint64_t HELPER(glue(iwmmxt_unpack, glue(S, w)))(CPUState *env, \
182 uint64_t a, uint64_t b) \
183 { \
184 a = \
185 (((a >> SH0) & 0xffff) << 0) | \
186 (((b >> SH0) & 0xffff) << 16) | \
187 (((a >> SH2) & 0xffff) << 32) | \
188 (((b >> SH2) & 0xffff) << 48); \
189 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \
190 NZBIT8(a >> 0, 0) | NZBIT8(a >> 16, 1) | \
191 NZBIT8(a >> 32, 2) | NZBIT8(a >> 48, 3); \
192 return a; \
193 } \
194 uint64_t HELPER(glue(iwmmxt_unpack, glue(S, l)))(CPUState *env, \
195 uint64_t a, uint64_t b) \
196 { \
197 a = \
198 (((a >> SH0) & 0xffffffff) << 0) | \
199 (((b >> SH0) & 0xffffffff) << 32); \
200 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \
201 NZBIT32(a >> 0, 0) | NZBIT32(a >> 32, 1); \
202 return a; \
203 } \
204 uint64_t HELPER(glue(iwmmxt_unpack, glue(S, ub)))(CPUState *env, \
205 uint64_t x) \
206 { \
207 x = \
208 (((x >> SH0) & 0xff) << 0) | \
209 (((x >> SH1) & 0xff) << 16) | \
210 (((x >> SH2) & 0xff) << 32) | \
211 (((x >> SH3) & 0xff) << 48); \
212 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \
213 NZBIT16(x >> 0, 0) | NZBIT16(x >> 16, 1) | \
214 NZBIT16(x >> 32, 2) | NZBIT16(x >> 48, 3); \
215 return x; \
216 } \
217 uint64_t HELPER(glue(iwmmxt_unpack, glue(S, uw)))(CPUState *env, \
218 uint64_t x) \
219 { \
220 x = \
221 (((x >> SH0) & 0xffff) << 0) | \
222 (((x >> SH2) & 0xffff) << 32); \
223 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \
224 NZBIT32(x >> 0, 0) | NZBIT32(x >> 32, 1); \
225 return x; \
226 } \
227 uint64_t HELPER(glue(iwmmxt_unpack, glue(S, ul)))(CPUState *env, \
228 uint64_t x) \
229 { \
230 x = (((x >> SH0) & 0xffffffff) << 0); \
231 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = NZBIT64(x >> 0); \
232 return x; \
233 } \
234 uint64_t HELPER(glue(iwmmxt_unpack, glue(S, sb)))(CPUState *env, \
235 uint64_t x) \
236 { \
237 x = \
238 ((uint64_t) EXTEND8H((x >> SH0) & 0xff) << 0) | \
239 ((uint64_t) EXTEND8H((x >> SH1) & 0xff) << 16) | \
240 ((uint64_t) EXTEND8H((x >> SH2) & 0xff) << 32) | \
241 ((uint64_t) EXTEND8H((x >> SH3) & 0xff) << 48); \
242 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \
243 NZBIT16(x >> 0, 0) | NZBIT16(x >> 16, 1) | \
244 NZBIT16(x >> 32, 2) | NZBIT16(x >> 48, 3); \
245 return x; \
246 } \
247 uint64_t HELPER(glue(iwmmxt_unpack, glue(S, sw)))(CPUState *env, \
248 uint64_t x) \
249 { \
250 x = \
251 ((uint64_t) EXTEND16((x >> SH0) & 0xffff) << 0) | \
252 ((uint64_t) EXTEND16((x >> SH2) & 0xffff) << 32); \
253 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \
254 NZBIT32(x >> 0, 0) | NZBIT32(x >> 32, 1); \
255 return x; \
256 } \
257 uint64_t HELPER(glue(iwmmxt_unpack, glue(S, sl)))(CPUState *env, \
258 uint64_t x) \
259 { \
260 x = EXTEND32((x >> SH0) & 0xffffffff); \
261 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = NZBIT64(x >> 0); \
262 return x; \
263 }
264 IWMMXT_OP_UNPACK(l, 0, 8, 16, 24)
265 IWMMXT_OP_UNPACK(h, 32, 40, 48, 56)
266
267 #define IWMMXT_OP_CMP(SUFF, Tb, Tw, Tl, O) \
268 uint64_t HELPER(glue(iwmmxt_, glue(SUFF, b)))(CPUState *env, \
269 uint64_t a, uint64_t b) \
270 { \
271 a = \
272 CMP(0, Tb, O, 0xff) | CMP(8, Tb, O, 0xff) | \
273 CMP(16, Tb, O, 0xff) | CMP(24, Tb, O, 0xff) | \
274 CMP(32, Tb, O, 0xff) | CMP(40, Tb, O, 0xff) | \
275 CMP(48, Tb, O, 0xff) | CMP(56, Tb, O, 0xff); \
276 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \
277 NZBIT8(a >> 0, 0) | NZBIT8(a >> 8, 1) | \
278 NZBIT8(a >> 16, 2) | NZBIT8(a >> 24, 3) | \
279 NZBIT8(a >> 32, 4) | NZBIT8(a >> 40, 5) | \
280 NZBIT8(a >> 48, 6) | NZBIT8(a >> 56, 7); \
281 return a; \
282 } \
283 uint64_t HELPER(glue(iwmmxt_, glue(SUFF, w)))(CPUState *env, \
284 uint64_t a, uint64_t b) \
285 { \
286 a = CMP(0, Tw, O, 0xffff) | CMP(16, Tw, O, 0xffff) | \
287 CMP(32, Tw, O, 0xffff) | CMP(48, Tw, O, 0xffff); \
288 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \
289 NZBIT16(a >> 0, 0) | NZBIT16(a >> 16, 1) | \
290 NZBIT16(a >> 32, 2) | NZBIT16(a >> 48, 3); \
291 return a; \
292 } \
293 uint64_t HELPER(glue(iwmmxt_, glue(SUFF, l)))(CPUState *env, \
294 uint64_t a, uint64_t b) \
295 { \
296 a = CMP(0, Tl, O, 0xffffffff) | \
297 CMP(32, Tl, O, 0xffffffff); \
298 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \
299 NZBIT32(a >> 0, 0) | NZBIT32(a >> 32, 1); \
300 return a; \
301 }
302 #define CMP(SHR, TYPE, OPER, MASK) ((((TYPE) ((a >> SHR) & MASK) OPER \
303 (TYPE) ((b >> SHR) & MASK)) ? (uint64_t) MASK : 0) << SHR)
304 IWMMXT_OP_CMP(cmpeq, uint8_t, uint16_t, uint32_t, ==)
305 IWMMXT_OP_CMP(cmpgts, int8_t, int16_t, int32_t, >)
306 IWMMXT_OP_CMP(cmpgtu, uint8_t, uint16_t, uint32_t, >)
307 #undef CMP
308 #define CMP(SHR, TYPE, OPER, MASK) ((((TYPE) ((a >> SHR) & MASK) OPER \
309 (TYPE) ((b >> SHR) & MASK)) ? a : b) & ((uint64_t) MASK << SHR))
310 IWMMXT_OP_CMP(mins, int8_t, int16_t, int32_t, <)
311 IWMMXT_OP_CMP(minu, uint8_t, uint16_t, uint32_t, <)
312 IWMMXT_OP_CMP(maxs, int8_t, int16_t, int32_t, >)
313 IWMMXT_OP_CMP(maxu, uint8_t, uint16_t, uint32_t, >)
314 #undef CMP
315 #define CMP(SHR, TYPE, OPER, MASK) ((uint64_t) (((TYPE) ((a >> SHR) & MASK) \
316 OPER (TYPE) ((b >> SHR) & MASK)) & MASK) << SHR)
317 IWMMXT_OP_CMP(subn, uint8_t, uint16_t, uint32_t, -)
318 IWMMXT_OP_CMP(addn, uint8_t, uint16_t, uint32_t, +)
319 #undef CMP
320 /* TODO Signed- and Unsigned-Saturation */
321 #define CMP(SHR, TYPE, OPER, MASK) ((uint64_t) (((TYPE) ((a >> SHR) & MASK) \
322 OPER (TYPE) ((b >> SHR) & MASK)) & MASK) << SHR)
323 IWMMXT_OP_CMP(subu, uint8_t, uint16_t, uint32_t, -)
324 IWMMXT_OP_CMP(addu, uint8_t, uint16_t, uint32_t, +)
325 IWMMXT_OP_CMP(subs, int8_t, int16_t, int32_t, -)
326 IWMMXT_OP_CMP(adds, int8_t, int16_t, int32_t, +)
327 #undef CMP
328 #undef IWMMXT_OP_CMP
329
330 #define AVGB(SHR) ((( \
331 ((a >> SHR) & 0xff) + ((b >> SHR) & 0xff) + round) >> 1) << SHR)
332 #define IWMMXT_OP_AVGB(r) \
333 uint64_t HELPER(iwmmxt_avgb##r)(CPUState *env, uint64_t a, uint64_t b) \
334 { \
335 const int round = r; \
336 a = AVGB(0) | AVGB(8) | AVGB(16) | AVGB(24) | \
337 AVGB(32) | AVGB(40) | AVGB(48) | AVGB(56); \
338 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \
339 SIMD8_SET(ZBIT8((a >> 0) & 0xff), SIMD_ZBIT, 0) | \
340 SIMD8_SET(ZBIT8((a >> 8) & 0xff), SIMD_ZBIT, 1) | \
341 SIMD8_SET(ZBIT8((a >> 16) & 0xff), SIMD_ZBIT, 2) | \
342 SIMD8_SET(ZBIT8((a >> 24) & 0xff), SIMD_ZBIT, 3) | \
343 SIMD8_SET(ZBIT8((a >> 32) & 0xff), SIMD_ZBIT, 4) | \
344 SIMD8_SET(ZBIT8((a >> 40) & 0xff), SIMD_ZBIT, 5) | \
345 SIMD8_SET(ZBIT8((a >> 48) & 0xff), SIMD_ZBIT, 6) | \
346 SIMD8_SET(ZBIT8((a >> 56) & 0xff), SIMD_ZBIT, 7); \
347 return a; \
348 }
349 IWMMXT_OP_AVGB(0)
350 IWMMXT_OP_AVGB(1)
351 #undef IWMMXT_OP_AVGB
352 #undef AVGB
353
354 #define AVGW(SHR) ((( \
355 ((a >> SHR) & 0xffff) + ((b >> SHR) & 0xffff) + round) >> 1) << SHR)
356 #define IWMMXT_OP_AVGW(r) \
357 uint64_t HELPER(iwmmxt_avgw##r)(CPUState *env, uint64_t a, uint64_t b) \
358 { \
359 const int round = r; \
360 a = AVGW(0) | AVGW(16) | AVGW(32) | AVGW(48); \
361 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \
362 SIMD16_SET(ZBIT16((a >> 0) & 0xffff), SIMD_ZBIT, 0) | \
363 SIMD16_SET(ZBIT16((a >> 16) & 0xffff), SIMD_ZBIT, 1) | \
364 SIMD16_SET(ZBIT16((a >> 32) & 0xffff), SIMD_ZBIT, 2) | \
365 SIMD16_SET(ZBIT16((a >> 48) & 0xffff), SIMD_ZBIT, 3); \
366 return a; \
367 }
368 IWMMXT_OP_AVGW(0)
369 IWMMXT_OP_AVGW(1)
370 #undef IWMMXT_OP_AVGW
371 #undef AVGW
372
HELPER(iwmmxt_msadb)373 uint64_t HELPER(iwmmxt_msadb)(uint64_t a, uint64_t b)
374 {
375 a = ((((a >> 0 ) & 0xffff) * ((b >> 0) & 0xffff) +
376 ((a >> 16) & 0xffff) * ((b >> 16) & 0xffff)) & 0xffffffff) |
377 ((((a >> 32) & 0xffff) * ((b >> 32) & 0xffff) +
378 ((a >> 48) & 0xffff) * ((b >> 48) & 0xffff)) << 32);
379 return a;
380 }
381
HELPER(iwmmxt_align)382 uint64_t HELPER(iwmmxt_align)(uint64_t a, uint64_t b, uint32_t n)
383 {
384 a >>= n << 3;
385 a |= b << (64 - (n << 3));
386 return a;
387 }
388
HELPER(iwmmxt_insr)389 uint64_t HELPER(iwmmxt_insr)(uint64_t x, uint32_t a, uint32_t b, uint32_t n)
390 {
391 x &= ~((uint64_t) b << n);
392 x |= (uint64_t) (a & b) << n;
393 return x;
394 }
395
HELPER(iwmmxt_setpsr_nz)396 uint32_t HELPER(iwmmxt_setpsr_nz)(uint64_t x)
397 {
398 return SIMD64_SET((x == 0), SIMD_ZBIT) |
399 SIMD64_SET((x & (1ULL << 63)), SIMD_NBIT);
400 }
401
HELPER(iwmmxt_bcstb)402 uint64_t HELPER(iwmmxt_bcstb)(uint32_t arg)
403 {
404 arg &= 0xff;
405 return
406 ((uint64_t) arg << 0 ) | ((uint64_t) arg << 8 ) |
407 ((uint64_t) arg << 16) | ((uint64_t) arg << 24) |
408 ((uint64_t) arg << 32) | ((uint64_t) arg << 40) |
409 ((uint64_t) arg << 48) | ((uint64_t) arg << 56);
410 }
411
HELPER(iwmmxt_bcstw)412 uint64_t HELPER(iwmmxt_bcstw)(uint32_t arg)
413 {
414 arg &= 0xffff;
415 return
416 ((uint64_t) arg << 0 ) | ((uint64_t) arg << 16) |
417 ((uint64_t) arg << 32) | ((uint64_t) arg << 48);
418 }
419
HELPER(iwmmxt_bcstl)420 uint64_t HELPER(iwmmxt_bcstl)(uint32_t arg)
421 {
422 return arg | ((uint64_t) arg << 32);
423 }
424
HELPER(iwmmxt_addcb)425 uint64_t HELPER(iwmmxt_addcb)(uint64_t x)
426 {
427 return
428 ((x >> 0) & 0xff) + ((x >> 8) & 0xff) +
429 ((x >> 16) & 0xff) + ((x >> 24) & 0xff) +
430 ((x >> 32) & 0xff) + ((x >> 40) & 0xff) +
431 ((x >> 48) & 0xff) + ((x >> 56) & 0xff);
432 }
433
HELPER(iwmmxt_addcw)434 uint64_t HELPER(iwmmxt_addcw)(uint64_t x)
435 {
436 return
437 ((x >> 0) & 0xffff) + ((x >> 16) & 0xffff) +
438 ((x >> 32) & 0xffff) + ((x >> 48) & 0xffff);
439 }
440
HELPER(iwmmxt_addcl)441 uint64_t HELPER(iwmmxt_addcl)(uint64_t x)
442 {
443 return (x & 0xffffffff) + (x >> 32);
444 }
445
HELPER(iwmmxt_msbb)446 uint32_t HELPER(iwmmxt_msbb)(uint64_t x)
447 {
448 return
449 ((x >> 7) & 0x01) | ((x >> 14) & 0x02) |
450 ((x >> 21) & 0x04) | ((x >> 28) & 0x08) |
451 ((x >> 35) & 0x10) | ((x >> 42) & 0x20) |
452 ((x >> 49) & 0x40) | ((x >> 56) & 0x80);
453 }
454
HELPER(iwmmxt_msbw)455 uint32_t HELPER(iwmmxt_msbw)(uint64_t x)
456 {
457 return
458 ((x >> 15) & 0x01) | ((x >> 30) & 0x02) |
459 ((x >> 45) & 0x04) | ((x >> 52) & 0x08);
460 }
461
HELPER(iwmmxt_msbl)462 uint32_t HELPER(iwmmxt_msbl)(uint64_t x)
463 {
464 return ((x >> 31) & 0x01) | ((x >> 62) & 0x02);
465 }
466
467 /* FIXME: Split wCASF setting into a separate op to avoid env use. */
HELPER(iwmmxt_srlw)468 uint64_t HELPER(iwmmxt_srlw)(CPUState *env, uint64_t x, uint32_t n)
469 {
470 x = (((x & (0xffffll << 0)) >> n) & (0xffffll << 0)) |
471 (((x & (0xffffll << 16)) >> n) & (0xffffll << 16)) |
472 (((x & (0xffffll << 32)) >> n) & (0xffffll << 32)) |
473 (((x & (0xffffll << 48)) >> n) & (0xffffll << 48));
474 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] =
475 NZBIT16(x >> 0, 0) | NZBIT16(x >> 16, 1) |
476 NZBIT16(x >> 32, 2) | NZBIT16(x >> 48, 3);
477 return x;
478 }
479
HELPER(iwmmxt_srll)480 uint64_t HELPER(iwmmxt_srll)(CPUState *env, uint64_t x, uint32_t n)
481 {
482 x = ((x & (0xffffffffll << 0)) >> n) |
483 ((x >> n) & (0xffffffffll << 32));
484 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] =
485 NZBIT32(x >> 0, 0) | NZBIT32(x >> 32, 1);
486 return x;
487 }
488
HELPER(iwmmxt_srlq)489 uint64_t HELPER(iwmmxt_srlq)(CPUState *env, uint64_t x, uint32_t n)
490 {
491 x >>= n;
492 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = NZBIT64(x);
493 return x;
494 }
495
HELPER(iwmmxt_sllw)496 uint64_t HELPER(iwmmxt_sllw)(CPUState *env, uint64_t x, uint32_t n)
497 {
498 x = (((x & (0xffffll << 0)) << n) & (0xffffll << 0)) |
499 (((x & (0xffffll << 16)) << n) & (0xffffll << 16)) |
500 (((x & (0xffffll << 32)) << n) & (0xffffll << 32)) |
501 (((x & (0xffffll << 48)) << n) & (0xffffll << 48));
502 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] =
503 NZBIT16(x >> 0, 0) | NZBIT16(x >> 16, 1) |
504 NZBIT16(x >> 32, 2) | NZBIT16(x >> 48, 3);
505 return x;
506 }
507
HELPER(iwmmxt_slll)508 uint64_t HELPER(iwmmxt_slll)(CPUState *env, uint64_t x, uint32_t n)
509 {
510 x = ((x << n) & (0xffffffffll << 0)) |
511 ((x & (0xffffffffll << 32)) << n);
512 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] =
513 NZBIT32(x >> 0, 0) | NZBIT32(x >> 32, 1);
514 return x;
515 }
516
HELPER(iwmmxt_sllq)517 uint64_t HELPER(iwmmxt_sllq)(CPUState *env, uint64_t x, uint32_t n)
518 {
519 x <<= n;
520 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = NZBIT64(x);
521 return x;
522 }
523
HELPER(iwmmxt_sraw)524 uint64_t HELPER(iwmmxt_sraw)(CPUState *env, uint64_t x, uint32_t n)
525 {
526 x = ((uint64_t) ((EXTEND16(x >> 0) >> n) & 0xffff) << 0) |
527 ((uint64_t) ((EXTEND16(x >> 16) >> n) & 0xffff) << 16) |
528 ((uint64_t) ((EXTEND16(x >> 32) >> n) & 0xffff) << 32) |
529 ((uint64_t) ((EXTEND16(x >> 48) >> n) & 0xffff) << 48);
530 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] =
531 NZBIT16(x >> 0, 0) | NZBIT16(x >> 16, 1) |
532 NZBIT16(x >> 32, 2) | NZBIT16(x >> 48, 3);
533 return x;
534 }
535
HELPER(iwmmxt_sral)536 uint64_t HELPER(iwmmxt_sral)(CPUState *env, uint64_t x, uint32_t n)
537 {
538 x = (((EXTEND32(x >> 0) >> n) & 0xffffffff) << 0) |
539 (((EXTEND32(x >> 32) >> n) & 0xffffffff) << 32);
540 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] =
541 NZBIT32(x >> 0, 0) | NZBIT32(x >> 32, 1);
542 return x;
543 }
544
HELPER(iwmmxt_sraq)545 uint64_t HELPER(iwmmxt_sraq)(CPUState *env, uint64_t x, uint32_t n)
546 {
547 x = (int64_t) x >> n;
548 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = NZBIT64(x);
549 return x;
550 }
551
HELPER(iwmmxt_rorw)552 uint64_t HELPER(iwmmxt_rorw)(CPUState *env, uint64_t x, uint32_t n)
553 {
554 x = ((((x & (0xffffll << 0)) >> n) |
555 ((x & (0xffffll << 0)) << (16 - n))) & (0xffffll << 0)) |
556 ((((x & (0xffffll << 16)) >> n) |
557 ((x & (0xffffll << 16)) << (16 - n))) & (0xffffll << 16)) |
558 ((((x & (0xffffll << 32)) >> n) |
559 ((x & (0xffffll << 32)) << (16 - n))) & (0xffffll << 32)) |
560 ((((x & (0xffffll << 48)) >> n) |
561 ((x & (0xffffll << 48)) << (16 - n))) & (0xffffll << 48));
562 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] =
563 NZBIT16(x >> 0, 0) | NZBIT16(x >> 16, 1) |
564 NZBIT16(x >> 32, 2) | NZBIT16(x >> 48, 3);
565 return x;
566 }
567
HELPER(iwmmxt_rorl)568 uint64_t HELPER(iwmmxt_rorl)(CPUState *env, uint64_t x, uint32_t n)
569 {
570 x = ((x & (0xffffffffll << 0)) >> n) |
571 ((x >> n) & (0xffffffffll << 32)) |
572 ((x << (32 - n)) & (0xffffffffll << 0)) |
573 ((x & (0xffffffffll << 32)) << (32 - n));
574 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] =
575 NZBIT32(x >> 0, 0) | NZBIT32(x >> 32, 1);
576 return x;
577 }
578
HELPER(iwmmxt_rorq)579 uint64_t HELPER(iwmmxt_rorq)(CPUState *env, uint64_t x, uint32_t n)
580 {
581 x = (x >> n) | (x << (64 - n));
582 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = NZBIT64(x);
583 return x;
584 }
585
HELPER(iwmmxt_shufh)586 uint64_t HELPER(iwmmxt_shufh)(CPUState *env, uint64_t x, uint32_t n)
587 {
588 x = (((x >> ((n << 4) & 0x30)) & 0xffff) << 0) |
589 (((x >> ((n << 2) & 0x30)) & 0xffff) << 16) |
590 (((x >> ((n << 0) & 0x30)) & 0xffff) << 32) |
591 (((x >> ((n >> 2) & 0x30)) & 0xffff) << 48);
592 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] =
593 NZBIT16(x >> 0, 0) | NZBIT16(x >> 16, 1) |
594 NZBIT16(x >> 32, 2) | NZBIT16(x >> 48, 3);
595 return x;
596 }
597
598 /* TODO: Unsigned-Saturation */
HELPER(iwmmxt_packuw)599 uint64_t HELPER(iwmmxt_packuw)(CPUState *env, uint64_t a, uint64_t b)
600 {
601 a = (((a >> 0) & 0xff) << 0) | (((a >> 16) & 0xff) << 8) |
602 (((a >> 32) & 0xff) << 16) | (((a >> 48) & 0xff) << 24) |
603 (((b >> 0) & 0xff) << 32) | (((b >> 16) & 0xff) << 40) |
604 (((b >> 32) & 0xff) << 48) | (((b >> 48) & 0xff) << 56);
605 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] =
606 NZBIT8(a >> 0, 0) | NZBIT8(a >> 8, 1) |
607 NZBIT8(a >> 16, 2) | NZBIT8(a >> 24, 3) |
608 NZBIT8(a >> 32, 4) | NZBIT8(a >> 40, 5) |
609 NZBIT8(a >> 48, 6) | NZBIT8(a >> 56, 7);
610 return a;
611 }
612
HELPER(iwmmxt_packul)613 uint64_t HELPER(iwmmxt_packul)(CPUState *env, uint64_t a, uint64_t b)
614 {
615 a = (((a >> 0) & 0xffff) << 0) | (((a >> 32) & 0xffff) << 16) |
616 (((b >> 0) & 0xffff) << 32) | (((b >> 32) & 0xffff) << 48);
617 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] =
618 NZBIT16(a >> 0, 0) | NZBIT16(a >> 16, 1) |
619 NZBIT16(a >> 32, 2) | NZBIT16(a >> 48, 3);
620 return a;
621 }
622
HELPER(iwmmxt_packuq)623 uint64_t HELPER(iwmmxt_packuq)(CPUState *env, uint64_t a, uint64_t b)
624 {
625 a = (a & 0xffffffff) | ((b & 0xffffffff) << 32);
626 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] =
627 NZBIT32(a >> 0, 0) | NZBIT32(a >> 32, 1);
628 return a;
629 }
630
631 /* TODO: Signed-Saturation */
HELPER(iwmmxt_packsw)632 uint64_t HELPER(iwmmxt_packsw)(CPUState *env, uint64_t a, uint64_t b)
633 {
634 a = (((a >> 0) & 0xff) << 0) | (((a >> 16) & 0xff) << 8) |
635 (((a >> 32) & 0xff) << 16) | (((a >> 48) & 0xff) << 24) |
636 (((b >> 0) & 0xff) << 32) | (((b >> 16) & 0xff) << 40) |
637 (((b >> 32) & 0xff) << 48) | (((b >> 48) & 0xff) << 56);
638 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] =
639 NZBIT8(a >> 0, 0) | NZBIT8(a >> 8, 1) |
640 NZBIT8(a >> 16, 2) | NZBIT8(a >> 24, 3) |
641 NZBIT8(a >> 32, 4) | NZBIT8(a >> 40, 5) |
642 NZBIT8(a >> 48, 6) | NZBIT8(a >> 56, 7);
643 return a;
644 }
645
HELPER(iwmmxt_packsl)646 uint64_t HELPER(iwmmxt_packsl)(CPUState *env, uint64_t a, uint64_t b)
647 {
648 a = (((a >> 0) & 0xffff) << 0) | (((a >> 32) & 0xffff) << 16) |
649 (((b >> 0) & 0xffff) << 32) | (((b >> 32) & 0xffff) << 48);
650 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] =
651 NZBIT16(a >> 0, 0) | NZBIT16(a >> 16, 1) |
652 NZBIT16(a >> 32, 2) | NZBIT16(a >> 48, 3);
653 return a;
654 }
655
HELPER(iwmmxt_packsq)656 uint64_t HELPER(iwmmxt_packsq)(CPUState *env, uint64_t a, uint64_t b)
657 {
658 a = (a & 0xffffffff) | ((b & 0xffffffff) << 32);
659 env->iwmmxt.cregs[ARM_IWMMXT_wCASF] =
660 NZBIT32(a >> 0, 0) | NZBIT32(a >> 32, 1);
661 return a;
662 }
663
HELPER(iwmmxt_muladdsl)664 uint64_t HELPER(iwmmxt_muladdsl)(uint64_t c, uint32_t a, uint32_t b)
665 {
666 return c + ((int32_t) EXTEND32(a) * (int32_t) EXTEND32(b));
667 }
668
HELPER(iwmmxt_muladdsw)669 uint64_t HELPER(iwmmxt_muladdsw)(uint64_t c, uint32_t a, uint32_t b)
670 {
671 c += EXTEND32(EXTEND16S((a >> 0) & 0xffff) *
672 EXTEND16S((b >> 0) & 0xffff));
673 c += EXTEND32(EXTEND16S((a >> 16) & 0xffff) *
674 EXTEND16S((b >> 16) & 0xffff));
675 return c;
676 }
677
HELPER(iwmmxt_muladdswl)678 uint64_t HELPER(iwmmxt_muladdswl)(uint64_t c, uint32_t a, uint32_t b)
679 {
680 return c + (EXTEND32(EXTEND16S(a & 0xffff) *
681 EXTEND16S(b & 0xffff)));
682 }
683