summaryrefslogtreecommitdiff
path: root/src/arch/arm/insts/vfp.hh
blob: 964b62673ae2574127ee19706112637b4e8e308f (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
/*
 * Copyright (c) 2010 ARM Limited
 * All rights reserved
 *
 * The license below extends only to copyright in the software and shall
 * not be construed as granting a license to any other intellectual
 * property including but not limited to intellectual property relating
 * to a hardware implementation of the functionality of the software
 * licensed hereunder.  You may use the software subject to the license
 * terms below provided that you ensure that this notice is replicated
 * unmodified and in its entirety in all distributions of the software,
 * modified or unmodified, in source code or in binary form.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions are
 * met: redistributions of source code must retain the above copyright
 * notice, this list of conditions and the following disclaimer;
 * redistributions in binary form must reproduce the above copyright
 * notice, this list of conditions and the following disclaimer in the
 * documentation and/or other materials provided with the distribution;
 * neither the name of the copyright holders nor the names of its
 * contributors may be used to endorse or promote products derived from
 * this software without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
 * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
 * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 *
 * Authors: Gabe Black
 */

#ifndef __ARCH_ARM_INSTS_VFP_HH__
#define __ARCH_ARM_INSTS_VFP_HH__

#include "arch/arm/insts/misc.hh"
#include "arch/arm/miscregs.hh"
#include <fenv.h>
#include <cmath>

namespace ArmISA
{

enum VfpMicroMode {
    VfpNotAMicroop,
    VfpMicroop,
    VfpFirstMicroop,
    VfpLastMicroop
};

template<class T>
static inline void
setVfpMicroFlags(VfpMicroMode mode, T &flags)
{
    switch (mode) {
      case VfpMicroop:
        flags[StaticInst::IsMicroop] = true;
        break;
      case VfpFirstMicroop:
        flags[StaticInst::IsMicroop] =
            flags[StaticInst::IsFirstMicroop] = true;
        break;
      case VfpLastMicroop:
        flags[StaticInst::IsMicroop] =
            flags[StaticInst::IsLastMicroop] = true;
        break;
      case VfpNotAMicroop:
        break;
    }
    if (mode == VfpMicroop || mode == VfpFirstMicroop) {
        flags[StaticInst::IsDelayedCommit] = true;
    }
}

enum FeExceptionBit
{
    FeDivByZero = FE_DIVBYZERO,
    FeInexact = FE_INEXACT,
    FeInvalid = FE_INVALID,
    FeOverflow = FE_OVERFLOW,
    FeUnderflow = FE_UNDERFLOW,
    FeAllExceptions = FE_ALL_EXCEPT
};

enum FeRoundingMode
{
    FeRoundDown = FE_DOWNWARD,
    FeRoundNearest = FE_TONEAREST,
    FeRoundZero = FE_TOWARDZERO,
    FeRoundUpward = FE_UPWARD
};

enum VfpRoundingMode
{
    VfpRoundNearest = 0,
    VfpRoundUpward = 1,
    VfpRoundDown = 2,
    VfpRoundZero = 3
};

template <class fpType>
static inline bool
flushToZero(fpType &op)
{
    fpType junk = 0.0;
    if (std::fpclassify(op) == FP_SUBNORMAL) {
        uint64_t bitMask = ULL(0x1) << (sizeof(fpType) * 8 - 1);
        op = bitsToFp(fpToBits(op) & bitMask, junk);
        return true;
    }
    return false;
}

template <class fpType>
static inline bool
flushToZero(fpType &op1, fpType &op2)
{
    bool flush1 = flushToZero(op1);
    bool flush2 = flushToZero(op2);
    return flush1 || flush2;
}

template <class fpType>
static inline void
vfpFlushToZero(FPSCR &fpscr, fpType &op)
{
    if (fpscr.fz == 1 && flushToZero(op)) {
        fpscr.idc = 1;
    }
}

template <class fpType>
static inline void
vfpFlushToZero(FPSCR &fpscr, fpType &op1, fpType &op2)
{
    vfpFlushToZero(fpscr, op1);
    vfpFlushToZero(fpscr, op2);
}

static inline uint32_t
fpToBits(float fp)
{
    union
    {
        float fp;
        uint32_t bits;
    } val;
    val.fp = fp;
    return val.bits;
}

static inline uint64_t
fpToBits(double fp)
{
    union
    {
        double fp;
        uint64_t bits;
    } val;
    val.fp = fp;
    return val.bits;
}

static inline float
bitsToFp(uint64_t bits, float junk)
{
    union
    {
        float fp;
        uint32_t bits;
    } val;
    val.bits = bits;
    return val.fp;
}

static inline double
bitsToFp(uint64_t bits, double junk)
{
    union
    {
        double fp;
        uint64_t bits;
    } val;
    val.bits = bits;
    return val.fp;
}

template <class fpType>
static bool
isSnan(fpType val)
{
    const bool single = (sizeof(fpType) == sizeof(float));
    const uint64_t qnan =
        single ? 0x7fc00000 : ULL(0x7ff8000000000000);
    return std::isnan(val) && ((fpToBits(val) & qnan) != qnan);
}

typedef int VfpSavedState;

VfpSavedState prepFpState(uint32_t rMode);
void finishVfp(FPSCR &fpscr, VfpSavedState state, bool flush);

template <class fpType>
fpType fixDest(FPSCR fpscr, fpType val, fpType op1);

template <class fpType>
fpType fixDest(FPSCR fpscr, fpType val, fpType op1, fpType op2);

template <class fpType>
fpType fixDivDest(FPSCR fpscr, fpType val, fpType op1, fpType op2);

float fixFpDFpSDest(FPSCR fpscr, double val);
double fixFpSFpDDest(FPSCR fpscr, float val);

uint16_t vcvtFpSFpH(FPSCR &fpscr, bool flush, bool defaultNan,
                    uint32_t rMode, bool ahp, float op);
float vcvtFpHFpS(FPSCR &fpscr, bool defaultNan, bool ahp, uint16_t op);

static inline double
makeDouble(uint32_t low, uint32_t high)
{
    double junk = 0.0;
    return bitsToFp((uint64_t)low | ((uint64_t)high << 32), junk);
}

static inline uint32_t
lowFromDouble(double val)
{
    return fpToBits(val);
}

static inline uint32_t
highFromDouble(double val)
{
    return fpToBits(val) >> 32;
}

uint64_t vfpFpSToFixed(float val, bool isSigned, bool half,
                       uint8_t imm, bool rzero = true);
float vfpUFixedToFpS(bool flush, bool defaultNan,
        uint32_t val, bool half, uint8_t imm);
float vfpSFixedToFpS(bool flush, bool defaultNan,
        int32_t val, bool half, uint8_t imm);

uint64_t vfpFpDToFixed(double val, bool isSigned, bool half,
                       uint8_t imm, bool rzero = true);
double vfpUFixedToFpD(bool flush, bool defaultNan,
        uint32_t val, bool half, uint8_t imm);
double vfpSFixedToFpD(bool flush, bool defaultNan,
        int32_t val, bool half, uint8_t imm);

float fprSqrtEstimate(FPSCR &fpscr, float op);
uint32_t unsignedRSqrtEstimate(uint32_t op);

float fpRecipEstimate(FPSCR &fpscr, float op);
uint32_t unsignedRecipEstimate(uint32_t op);

class VfpMacroOp : public PredMacroOp
{
  public:
    static bool
    inScalarBank(IntRegIndex idx)
    {
        return (idx % 32) < 8;
    }

  protected:
    bool wide;

    VfpMacroOp(const char *mnem, ExtMachInst _machInst,
            OpClass __opClass, bool _wide) :
        PredMacroOp(mnem, _machInst, __opClass), wide(_wide)
    {}

    IntRegIndex addStride(IntRegIndex idx, unsigned stride);
    void nextIdxs(IntRegIndex &dest, IntRegIndex &op1, IntRegIndex &op2);
    void nextIdxs(IntRegIndex &dest, IntRegIndex &op1);
    void nextIdxs(IntRegIndex &dest);
};

static inline float
fpAddS(float a, float b)
{
    return a + b;
}

static inline double
fpAddD(double a, double b)
{
    return a + b;
}

static inline float
fpSubS(float a, float b)
{
    return a - b;
}

static inline double
fpSubD(double a, double b)
{
    return a - b;
}

static inline float
fpDivS(float a, float b)
{
    return a / b;
}

static inline double
fpDivD(double a, double b)
{
    return a / b;
}

static inline float
fpMulS(float a, float b)
{
    return a * b;
}

static inline double
fpMulD(double a, double b)
{
    return a * b;
}

static inline float
fpMaxS(float a, float b)
{
    // Handle comparisons of +0 and -0.
    if (!std::signbit(a) && std::signbit(b))
        return a;
    return fmaxf(a, b);
}

static inline float
fpMinS(float a, float b)
{
    // Handle comparisons of +0 and -0.
    if (std::signbit(a) && !std::signbit(b))
        return a;
    return fminf(a, b);
}

static inline float
fpRSqrtsS(float a, float b)
{
    int fpClassA = std::fpclassify(a);
    int fpClassB = std::fpclassify(b);
    float aXb;
    int fpClassAxB;

    if ((fpClassA == FP_ZERO && fpClassB == FP_INFINITE) ||
        (fpClassA == FP_INFINITE && fpClassB == FP_ZERO)) {
        return 1.5;
    }
    aXb = a*b;
    fpClassAxB = std::fpclassify(aXb);
    if(fpClassAxB == FP_SUBNORMAL) {
       feraiseexcept(FeUnderflow);
       return 1.5;
    }
    return (3.0 - (a * b)) / 2.0;
}

static inline float
fpRecpsS(float a, float b)
{
    int fpClassA = std::fpclassify(a);
    int fpClassB = std::fpclassify(b);
    float aXb;
    int fpClassAxB;

    if ((fpClassA == FP_ZERO && fpClassB == FP_INFINITE) ||
        (fpClassA == FP_INFINITE && fpClassB == FP_ZERO)) {
        return 2.0;
    }
    aXb = a*b;
    fpClassAxB = std::fpclassify(aXb);
    if(fpClassAxB == FP_SUBNORMAL) {
       feraiseexcept(FeUnderflow);
       return 2.0;
    }
    return 2.0 - (a * b);
}

class FpOp : public PredOp
{
  protected:
    FpOp(const char *mnem, ExtMachInst _machInst, OpClass __opClass) :
        PredOp(mnem, _machInst, __opClass)
    {}

    virtual float
    doOp(float op1, float op2) const
    {
        panic("Unimplemented version of doOp called.\n");
    }

    virtual float
    doOp(float op1) const
    {
        panic("Unimplemented version of doOp called.\n");
    }

    virtual double
    doOp(double op1, double op2) const
    {
        panic("Unimplemented version of doOp called.\n");
    }

    virtual double
    doOp(double op1) const
    {
        panic("Unimplemented version of doOp called.\n");
    }

    double
    dbl(uint32_t low, uint32_t high) const
    {
        double junk = 0.0;
        return bitsToFp((uint64_t)low | ((uint64_t)high << 32), junk);
    }

    uint32_t
    dblLow(double val) const
    {
        return fpToBits(val);
    }

    uint32_t
    dblHi(double val) const
    {
        return fpToBits(val) >> 32;
    }

    template <class fpType>
    fpType
    processNans(FPSCR &fpscr, bool &done, bool defaultNan,
                fpType op1, fpType op2) const;

    template <class fpType>
    fpType
    binaryOp(FPSCR &fpscr, fpType op1, fpType op2,
            fpType (*func)(fpType, fpType),
            bool flush, bool defaultNan, uint32_t rMode) const;

    template <class fpType>
    fpType
    unaryOp(FPSCR &fpscr, fpType op1,
            fpType (*func)(fpType),
            bool flush, uint32_t rMode) const;
};

class FpRegRegOp : public FpOp
{
  protected:
    IntRegIndex dest;
    IntRegIndex op1;

    FpRegRegOp(const char *mnem, ExtMachInst _machInst, OpClass __opClass,
               IntRegIndex _dest, IntRegIndex _op1,
               VfpMicroMode mode = VfpNotAMicroop) :
        FpOp(mnem, _machInst, __opClass), dest(_dest), op1(_op1)
    {
        setVfpMicroFlags(mode, flags);
    }

    std::string generateDisassembly(Addr pc, const SymbolTable *symtab) const;
};

class FpRegImmOp : public FpOp
{
  protected:
    IntRegIndex dest;
    uint64_t imm;

    FpRegImmOp(const char *mnem, ExtMachInst _machInst, OpClass __opClass,
               IntRegIndex _dest, uint64_t _imm,
               VfpMicroMode mode = VfpNotAMicroop) :
        FpOp(mnem, _machInst, __opClass), dest(_dest), imm(_imm)
    {
        setVfpMicroFlags(mode, flags);
    }

    std::string generateDisassembly(Addr pc, const SymbolTable *symtab) const;
};

class FpRegRegImmOp : public FpOp
{
  protected:
    IntRegIndex dest;
    IntRegIndex op1;
    uint64_t imm;

    FpRegRegImmOp(const char *mnem, ExtMachInst _machInst, OpClass __opClass,
                  IntRegIndex _dest, IntRegIndex _op1,
                  uint64_t _imm, VfpMicroMode mode = VfpNotAMicroop) :
        FpOp(mnem, _machInst, __opClass), dest(_dest), op1(_op1), imm(_imm)
    {
        setVfpMicroFlags(mode, flags);
    }

    std::string generateDisassembly(Addr pc, const SymbolTable *symtab) const;
};

class FpRegRegRegOp : public FpOp
{
  protected:
    IntRegIndex dest;
    IntRegIndex op1;
    IntRegIndex op2;

    FpRegRegRegOp(const char *mnem, ExtMachInst _machInst, OpClass __opClass,
                  IntRegIndex _dest, IntRegIndex _op1, IntRegIndex _op2,
                  VfpMicroMode mode = VfpNotAMicroop) :
        FpOp(mnem, _machInst, __opClass), dest(_dest), op1(_op1), op2(_op2)
    {
        setVfpMicroFlags(mode, flags);
    }

    std::string generateDisassembly(Addr pc, const SymbolTable *symtab) const;
};

class FpRegRegRegImmOp : public FpOp
{
  protected:
    IntRegIndex dest;
    IntRegIndex op1;
    IntRegIndex op2;
    uint64_t imm;

    FpRegRegRegImmOp(const char *mnem, ExtMachInst _machInst,
                     OpClass __opClass, IntRegIndex _dest,
                     IntRegIndex _op1, IntRegIndex _op2,
                     uint64_t _imm, VfpMicroMode mode = VfpNotAMicroop) :
        FpOp(mnem, _machInst, __opClass),
        dest(_dest), op1(_op1), op2(_op2), imm(_imm)
    {
        setVfpMicroFlags(mode, flags);
    }

    std::string generateDisassembly(Addr pc, const SymbolTable *symtab) const;
};

}

#endif //__ARCH_ARM_INSTS_VFP_HH__