aboutsummaryrefslogtreecommitdiffstats
path: root/src/gallium/drivers/radeon/R600InstrInfo.cpp
blob: 3c9e4eb1de61b93f5734b1412e12a5d694815138 (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
//===-- R600InstrInfo.cpp - R600 Instruction Information ------------------===//
//
//                     The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// R600 Implementation of TargetInstrInfo.
//
//===----------------------------------------------------------------------===//

#include "R600InstrInfo.h"
#include "AMDGPUTargetMachine.h"
#include "AMDGPUSubtarget.h"
#include "R600RegisterInfo.h"
#include "llvm/CodeGen/MachineInstrBuilder.h"

#define GET_INSTRINFO_CTOR
#include "AMDGPUGenDFAPacketizer.inc"

using namespace llvm;

R600InstrInfo::R600InstrInfo(AMDGPUTargetMachine &tm)
  : AMDGPUInstrInfo(tm),
    RI(tm, *this),
    TM(tm)
  { }

const R600RegisterInfo &R600InstrInfo::getRegisterInfo() const
{
  return RI;
}

bool R600InstrInfo::isTrig(const MachineInstr &MI) const
{
  return get(MI.getOpcode()).TSFlags & R600_InstFlag::TRIG;
}

bool R600InstrInfo::isVector(const MachineInstr &MI) const
{
  return get(MI.getOpcode()).TSFlags & R600_InstFlag::VECTOR;
}

void
R600InstrInfo::copyPhysReg(MachineBasicBlock &MBB,
                           MachineBasicBlock::iterator MI, DebugLoc DL,
                           unsigned DestReg, unsigned SrcReg,
                           bool KillSrc) const
{

  unsigned subRegMap[4] = {AMDGPU::sel_x, AMDGPU::sel_y,
                           AMDGPU::sel_z, AMDGPU::sel_w};

  if (AMDGPU::R600_Reg128RegClass.contains(DestReg)
      && AMDGPU::R600_Reg128RegClass.contains(SrcReg)) {
    for (unsigned i = 0; i < 4; i++) {
      BuildMI(MBB, MI, DL, get(AMDGPU::MOV))
              .addReg(RI.getSubReg(DestReg, subRegMap[i]), RegState::Define)
              .addReg(RI.getSubReg(SrcReg, subRegMap[i]))
              .addReg(DestReg, RegState::Define | RegState::Implicit);
    }
  } else {

    /* We can't copy vec4 registers */
    assert(!AMDGPU::R600_Reg128RegClass.contains(DestReg)
           && !AMDGPU::R600_Reg128RegClass.contains(SrcReg));

    BuildMI(MBB, MI, DL, get(AMDGPU::MOV), DestReg)
      .addReg(SrcReg, getKillRegState(KillSrc));
  }
}

MachineInstr * R600InstrInfo::getMovImmInstr(MachineFunction *MF,
                                             unsigned DstReg, int64_t Imm) const
{
  MachineInstr * MI = MF->CreateMachineInstr(get(AMDGPU::MOV), DebugLoc());
  MachineInstrBuilder(MI).addReg(DstReg, RegState::Define);
  MachineInstrBuilder(MI).addReg(AMDGPU::ALU_LITERAL_X);
  MachineInstrBuilder(MI).addImm(Imm);

  return MI;
}

unsigned R600InstrInfo::getIEQOpcode() const
{
  return AMDGPU::SETE_INT;
}

bool R600InstrInfo::isMov(unsigned Opcode) const
{
  switch(Opcode) {
  default: return false;
  case AMDGPU::MOV:
  case AMDGPU::MOV_IMM_F32:
  case AMDGPU::MOV_IMM_I32:
    return true;
  }
}

// Some instructions act as place holders to emulate operations that the GPU
// hardware does automatically. This function can be used to check if
// an opcode falls into this category.
bool R600InstrInfo::isPlaceHolderOpcode(unsigned opcode) const
{
  switch (opcode) {
  default: return false;
  case AMDGPU::RETURN:
  case AMDGPU::LAST:
  case AMDGPU::MASK_WRITE:
  case AMDGPU::RESERVE_REG:
    return true;
  }
}

bool R600InstrInfo::isTexOp(unsigned opcode) const
{
  switch(opcode) {
  default: return false;
  case AMDGPU::TEX_LD:
  case AMDGPU::TEX_GET_TEXTURE_RESINFO:
  case AMDGPU::TEX_SAMPLE:
  case AMDGPU::TEX_SAMPLE_C:
  case AMDGPU::TEX_SAMPLE_L:
  case AMDGPU::TEX_SAMPLE_C_L:
  case AMDGPU::TEX_SAMPLE_LB:
  case AMDGPU::TEX_SAMPLE_C_LB:
  case AMDGPU::TEX_SAMPLE_G:
  case AMDGPU::TEX_SAMPLE_C_G:
  case AMDGPU::TEX_GET_GRADIENTS_H:
  case AMDGPU::TEX_GET_GRADIENTS_V:
  case AMDGPU::TEX_SET_GRADIENTS_H:
  case AMDGPU::TEX_SET_GRADIENTS_V:
    return true;
  }
}

bool R600InstrInfo::isReductionOp(unsigned opcode) const
{
  switch(opcode) {
    default: return false;
    case AMDGPU::DOT4_r600:
    case AMDGPU::DOT4_eg:
      return true;
  }
}

bool R600InstrInfo::isCubeOp(unsigned opcode) const
{
  switch(opcode) {
    default: return false;
    case AMDGPU::CUBE_r600:
    case AMDGPU::CUBE_eg:
      return true;
  }
}


bool R600InstrInfo::isFCOp(unsigned opcode) const
{
  switch(opcode) {
  default: return false;
  case AMDGPU::BREAK_LOGICALZ_f32:
  case AMDGPU::BREAK_LOGICALNZ_i32:
  case AMDGPU::BREAK_LOGICALZ_i32:
  case AMDGPU::BREAK_LOGICALNZ_f32:
  case AMDGPU::CONTINUE_LOGICALNZ_f32:
  case AMDGPU::IF_LOGICALNZ_i32:
  case AMDGPU::IF_LOGICALZ_f32:
  case AMDGPU::ELSE:
  case AMDGPU::ENDIF:
  case AMDGPU::ENDLOOP:
  case AMDGPU::IF_LOGICALNZ_f32:
  case AMDGPU::WHILELOOP:
    return true;
  }
}

DFAPacketizer *R600InstrInfo::CreateTargetScheduleState(const TargetMachine *TM,
    const ScheduleDAG *DAG) const
{
  const InstrItineraryData *II = TM->getInstrItineraryData();
  return TM->getSubtarget<AMDGPUSubtarget>().createDFAPacketizer(II);
}