SIFixSGPRCopies.cpp File Reference

Copies from VGPR to SGPR registers are illegal and the register coalescer will sometimes generate these illegal copies in situations like this: More...

#include "SIFixSGPRCopies.h"
#include "AMDGPU.h"
#include "GCNSubtarget.h"
#include "MCTargetDesc/AMDGPUMCTargetDesc.h"
#include "llvm/CodeGen/MachineDominators.h"
#include "llvm/InitializePasses.h"
#include "llvm/Target/TargetMachine.h"

Go to the source code of this file.

Macros
#define	DEBUG_TYPE   "si-fix-sgpr-copies"

Functions
	INITIALIZE_PASS_BEGIN (SIFixSGPRCopiesLegacy, DEBUG_TYPE, "SI Fix SGPR copies", false, false) INITIALIZE_PASS_END(SIFixSGPRCopiesLegacy
static std::pair< const TargetRegisterClass *, const TargetRegisterClass * >	getCopyRegClasses (const MachineInstr &Copy, const SIRegisterInfo &TRI, const MachineRegisterInfo &MRI)
static bool	isVGPRToSGPRCopy (const TargetRegisterClass *SrcRC, const TargetRegisterClass *DstRC, const SIRegisterInfo &TRI)
static bool	isSGPRToVGPRCopy (const TargetRegisterClass *SrcRC, const TargetRegisterClass *DstRC, const SIRegisterInfo &TRI)
static bool	tryChangeVGPRtoSGPRinCopy (MachineInstr &MI, const SIRegisterInfo *TRI, const SIInstrInfo *TII)
static bool	foldVGPRCopyIntoRegSequence (MachineInstr &MI, const SIRegisterInfo *TRI, const SIInstrInfo *TII, MachineRegisterInfo &MRI)
static bool	isSafeToFoldImmIntoCopy (const MachineInstr *Copy, const MachineInstr *MoveImm, const SIInstrInfo *TII, unsigned &SMovOp, int64_t &Imm)
template<class UnaryPredicate >
bool	searchPredecessors (const MachineBasicBlock *MBB, const MachineBasicBlock *CutOff, UnaryPredicate Predicate)
static bool	isReachable (const MachineInstr *From, const MachineInstr *To, const MachineBasicBlock *CutOff, MachineDominatorTree &MDT)
static MachineBasicBlock::iterator	getFirstNonPrologue (MachineBasicBlock *MBB, const TargetInstrInfo *TII)
static bool	hoistAndMergeSGPRInits (unsigned Reg, const MachineRegisterInfo &MRI, const TargetRegisterInfo *TRI, MachineDominatorTree &MDT, const TargetInstrInfo *TII)

Variables
static cl::opt< bool >	EnableM0Merge ("amdgpu-enable-merge-m0", cl::desc("Merge and hoist M0 initializations"), cl::init(true))
	DEBUG_TYPE
SI Fix SGPR	copies
SI Fix SGPR	false

Detailed Description

Copies from VGPR to SGPR registers are illegal and the register coalescer will sometimes generate these illegal copies in situations like this:

Register Class <vsrc> is the union of <vgpr> and <sgpr>

BB0: %0 <sgpr> = SCALAR_INST %1 <vsrc> = COPY %0 <sgpr> ... BRANCH cond BB1, BB2 BB1: %2 <vgpr> = VECTOR_INST %3 <vsrc> = COPY %2 <vgpr> BB2: %4 <vsrc> = PHI %1 <vsrc>, <bb.0>, %3 <vrsc>, <bb.1> %5 <vgpr> = VECTOR_INST %4 <vsrc>

The coalescer will begin at BB0 and eliminate its copy, then the resulting code will look like this:

BB0: %0 <sgpr> = SCALAR_INST ... BRANCH cond BB1, BB2 BB1: %2 <vgpr> = VECTOR_INST %3 <vsrc> = COPY %2 <vgpr> BB2: %4 <sgpr> = PHI %0 <sgpr>, <bb.0>, %3 <vsrc>, <bb.1> %5 <vgpr> = VECTOR_INST %4 <sgpr>

Now that the result of the PHI instruction is an SGPR, the register allocator is now forced to constrain the register class of %3 to <sgpr> so we end up with final code like this:

BB0: %0 <sgpr> = SCALAR_INST ... BRANCH cond BB1, BB2 BB1: %2 <vgpr> = VECTOR_INST %3 <sgpr> = COPY %2 <vgpr> BB2: %4 <sgpr> = PHI %0 <sgpr>, <bb.0>, %3 <sgpr>, <bb.1> %5 <vgpr> = VECTOR_INST %4 <sgpr>

Now this code contains an illegal copy from a VGPR to an SGPR.

In order to avoid this problem, this pass searches for PHI instructions which define a <vsrc> register and constrains its definition class to <vgpr> if the user of the PHI's definition register is a vector instruction. If the PHI's definition class is constrained to <vgpr> then the coalescer will be unable to perform the COPY removal from the above example which ultimately led to the creation of an illegal COPY.

Definition in file SIFixSGPRCopies.cpp.

Macro Definition Documentation

DEBUG_TYPE

#define DEBUG_TYPE   "si-fix-sgpr-copies"

Definition at line 77 of file SIFixSGPRCopies.cpp.

Function Documentation

foldVGPRCopyIntoRegSequence()

static bool foldVGPRCopyIntoRegSequence ( MachineInstr &  MI, const SIRegisterInfo *  TRI, const SIInstrInfo *  TII, MachineRegisterInfo &  MRI  )

static

Definition at line 276 of file SIFixSGPRCopies.cpp.

References llvm::MachineInstrBuilder::add(), llvm::MachineInstrBuilder::addReg(), assert(), llvm::BuildMI(), llvm::MachineInstr::eraseFromParent(), getCopyRegClasses(), llvm::MachineInstr::getOperand(), llvm::MachineOperand::getReg(), llvm::MachineOperand::getSubReg(), I, llvm::MachineInstr::isCopy(), llvm::Register::isPhysical(), isSGPRToVGPRCopy(), llvm::RegState::Kill, MI, MRI, N, SubReg, TII, TRI, and tryChangeVGPRtoSGPRinCopy().

getCopyRegClasses()

static std::pair< const TargetRegisterClass *, const TargetRegisterClass * > getCopyRegClasses ( const MachineInstr &  Copy, const SIRegisterInfo &  TRI, const MachineRegisterInfo &  MRI  )

static

Definition at line 201 of file SIFixSGPRCopies.cpp.

References llvm::Register::isVirtual(), MRI, and TRI.

Referenced by foldVGPRCopyIntoRegSequence().

getFirstNonPrologue()

static MachineBasicBlock::iterator getFirstNonPrologue ( MachineBasicBlock *  MBB, const TargetInstrInfo *  TII  )

static

Definition at line 437 of file SIFixSGPRCopies.cpp.

References llvm::MachineBasicBlock::end(), llvm::MachineBasicBlock::getFirstNonPHI(), I, MBB, and TII.

Referenced by hoistAndMergeSGPRInits().

hoistAndMergeSGPRInits()

static bool hoistAndMergeSGPRInits ( unsigned  Reg, const MachineRegisterInfo &  MRI, const TargetRegisterInfo *  TRI, MachineDominatorTree &  MDT, const TargetInstrInfo *  TII  )

static

Definition at line 449 of file SIFixSGPRCopies.cpp.

References llvm::any_of(), assert(), B, llvm::CallingConv::C, llvm::SmallSet< T, N, C >::count(), llvm::dbgs(), llvm::MachineDominatorTree::dominates(), llvm::DominatorTreeBase< NodeT, IsPostDom >::findNearestCommonDominator(), From, getFirstNonPrologue(), llvm::MachineBasicBlock::getParent(), llvm::MachineInstr::getParent(), I, llvm::SmallSet< T, N, C >::insert(), isReachable(), llvm::HexagonInstrInfo::isSchedulingBoundary(), LLVM_DEBUG, MBB, MI, MRI, llvm::printMBBReference(), llvm::DominatorTreeBase< NodeT, IsPostDom >::properlyDominates(), llvm::SmallVectorTemplateBase< T, bool >::push_back(), llvm::MachineBasicBlock::splice(), TII, and TRI.

INITIALIZE_PASS_BEGIN()

INITIALIZE_PASS_BEGIN	(	SIFixSGPRCopiesLegacy	,
		DEBUG_TYPE	,
		"SI Fix SGPR copies"	,
		false	,
		false
	)

isReachable()

static bool isReachable ( const MachineInstr *  From, const MachineInstr *  To, const MachineBasicBlock *  CutOff, MachineDominatorTree &  MDT  )

static

Definition at line 418 of file SIFixSGPRCopies.cpp.

References llvm::MachineDominatorTree::dominates(), From, llvm::MachineInstr::getParent(), MBB, and searchPredecessors().

Referenced by hoistAndMergeSGPRInits().

isSafeToFoldImmIntoCopy()

static bool isSafeToFoldImmIntoCopy ( const MachineInstr *  Copy, const MachineInstr *  MoveImm, const SIInstrInfo *  TII, unsigned &  SMovOp, int64_t &  Imm  )

static

Definition at line 355 of file SIFixSGPRCopies.cpp.

References llvm::MachineOperand::getImm(), llvm::MachineOperand::isImm(), and TII.

isSGPRToVGPRCopy()

static bool isSGPRToVGPRCopy ( const TargetRegisterClass *  SrcRC, const TargetRegisterClass *  DstRC, const SIRegisterInfo &  TRI  )

static

Definition at line 228 of file SIFixSGPRCopies.cpp.

References TRI.

Referenced by foldVGPRCopyIntoRegSequence().

isVGPRToSGPRCopy()

static bool isVGPRToSGPRCopy ( const TargetRegisterClass *  SrcRC, const TargetRegisterClass *  DstRC, const SIRegisterInfo &  TRI  )

static

Definition at line 221 of file SIFixSGPRCopies.cpp.

References TRI.

searchPredecessors()

template<class UnaryPredicate >

bool searchPredecessors	(	const MachineBasicBlock *	MBB,
		const MachineBasicBlock *	CutOff,
		UnaryPredicate	Predicate
	)

Definition at line 390 of file SIFixSGPRCopies.cpp.

References llvm::SmallVectorImpl< T >::append(), llvm::SmallVectorBase< Size_T >::empty(), llvm::detail::DenseSetImpl< ValueT, MapTy, ValueInfoT >::insert(), MBB, llvm::SmallVectorImpl< T >::pop_back_val(), llvm::MachineBasicBlock::pred_begin(), llvm::MachineBasicBlock::pred_end(), and llvm::MachineBasicBlock::predecessors().

Referenced by isReachable().

tryChangeVGPRtoSGPRinCopy()

static bool tryChangeVGPRtoSGPRinCopy ( MachineInstr &  MI, const SIRegisterInfo *  TRI, const SIInstrInfo *  TII  )

static

Definition at line 235 of file SIFixSGPRCopies.cpp.

References llvm::MachineInstr::getDesc(), llvm::MCInstrDesc::getNumOperands(), llvm::MachineInstr::getOpcode(), llvm::MachineInstr::getParent(), llvm::Register::isVirtual(), MI, MRI, TII, TRI, and UseMI.

Referenced by foldVGPRCopyIntoRegSequence().

Variable Documentation

copies

SI Fix SGPR copies

Definition at line 189 of file SIFixSGPRCopies.cpp.

DEBUG_TYPE

DEBUG_TYPE

Definition at line 189 of file SIFixSGPRCopies.cpp.

EnableM0Merge

cl::opt< bool > EnableM0Merge("amdgpu-enable-merge-m0", cl::desc("Merge and hoist M0 initializations"), cl::init(true)) ( "amdgpu-enable-merge-m0"  , cl::desc("Merge and hoist M0 initializations")  , cl::init(true)    )

static

false

SI Fix SGPR false

Definition at line 190 of file SIFixSGPRCopies.cpp.