maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/llvm/lib/Target/AArch64/AArch64FastISel.cpp

5085 lines
167 KiB
C++
Raw Normal View History

AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
//===-- AArch6464FastISel.cpp - AArch64 FastISel implementation -----------===//
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
// This file defines the AArch64-specific support for the FastISel class. Some
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// of the target-specific code is generated by tablegen in the file
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
// AArch64GenFastISel.inc, which is #included here.
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
//
//===----------------------------------------------------------------------===//
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
#include "AArch64.h"
AArch64: treat [N x Ty] as a block during procedure calls. The AAPCS treats small structs and homogeneous floating (or vector) aggregates specially, and guarantees they either get passed as a contiguous block of registers, or prevent any future use of those registers and get passed on the stack. This concept can fit quite neatly into LLVM's own type system, mapping an HFA to [N x float] and so on, and small structs to [N x i64]. Doing so allows front-ends to emit AAPCS compliant code without having to duplicate the register counting logic. llvm-svn: 222903
2014-11-28 05:02:42 +08:00
#include "AArch64CallingConvention.h"
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
#include "AArch64Subtarget.h"
Run sort_includes.py on the AArch64 backend. No functionality change. llvm-svn: 213938
2014-07-25 19:42:14 +08:00
#include "AArch64TargetMachine.h"
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
#include "MCTargetDesc/AArch64AddressingModes.h"
[FastISel][AArch64] Add branch weights. Add branch weights to branch instructions, so that the following passes can optimize based on it (i.e. basic block ordering). Fixes <rdar://problem/17887137>. llvm-svn: 214537
2014-08-02 02:39:24 +08:00
#include "llvm/Analysis/BranchProbabilityInfo.h"
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
#include "llvm/CodeGen/CallingConvLower.h"
#include "llvm/CodeGen/FastISel.h"
#include "llvm/CodeGen/FunctionLoweringInfo.h"
#include "llvm/CodeGen/MachineConstantPool.h"
#include "llvm/CodeGen/MachineFrameInfo.h"
#include "llvm/CodeGen/MachineInstrBuilder.h"
#include "llvm/CodeGen/MachineRegisterInfo.h"
#include "llvm/IR/CallingConv.h"
#include "llvm/IR/DataLayout.h"
#include "llvm/IR/DerivedTypes.h"
#include "llvm/IR/Function.h"
#include "llvm/IR/GetElementPtrTypeIterator.h"
#include "llvm/IR/GlobalAlias.h"
#include "llvm/IR/GlobalVariable.h"
#include "llvm/IR/Instructions.h"
#include "llvm/IR/IntrinsicInst.h"
#include "llvm/IR/Operator.h"
Use MCSymbols for FastISel. The summary is that it moves the mangling earlier and replaces a few calls to .addExternalSymbol with addSym. I originally wanted to replace all the uses of addExternalSymbol with addSym, but noticed it was a lot of work and doesn't need to be done all at once. llvm-svn: 240395
2015-06-23 20:21:54 +08:00
#include "llvm/MC/MCSymbol.h"
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
using namespace llvm;
namespace {
[FastISel][AArch64] Make AArch64FastISel class final. NFC. llvm-svn: 217840
2014-09-16 06:33:11 +08:00
class AArch64FastISel final : public FastISel {
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
class Address {
public:
typedef enum {
RegBase,
FrameIndexBase
} BaseKind;
private:
BaseKind Kind;
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
AArch64_AM::ShiftExtendType ExtType;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
union {
unsigned Reg;
int FI;
} Base;
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
unsigned OffsetReg;
unsigned Shift;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
int64_t Offset;
[FastISel][AArch64] Add MachO large code model support for function calls. Currently the large code model for MachO uses the GOT to make function calls. Emit the required adrp and ldr instructions to load the address from the GOT. Related to <rdar://problem/17733076>. llvm-svn: 214381
2014-07-31 12:10:40 +08:00
const GlobalValue *GV;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
public:
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
Address() : Kind(RegBase), ExtType(AArch64_AM::InvalidShiftExtend),
OffsetReg(0), Shift(0), Offset(0), GV(nullptr) { Base.Reg = 0; }
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
void setKind(BaseKind K) { Kind = K; }
BaseKind getKind() const { return Kind; }
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
void setExtendType(AArch64_AM::ShiftExtendType E) { ExtType = E; }
AArch64_AM::ShiftExtendType getExtendType() const { return ExtType; }
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
bool isRegBase() const { return Kind == RegBase; }
bool isFIBase() const { return Kind == FrameIndexBase; }
void setReg(unsigned Reg) {
assert(isRegBase() && "Invalid base register access!");
Base.Reg = Reg;
}
unsigned getReg() const {
assert(isRegBase() && "Invalid base register access!");
return Base.Reg;
}
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
void setOffsetReg(unsigned Reg) {
OffsetReg = Reg;
}
unsigned getOffsetReg() const {
return OffsetReg;
}
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
void setFI(unsigned FI) {
assert(isFIBase() && "Invalid base frame index access!");
Base.FI = FI;
}
unsigned getFI() const {
assert(isFIBase() && "Invalid base frame index access!");
return Base.FI;
}
void setOffset(int64_t O) { Offset = O; }
int64_t getOffset() { return Offset; }
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
void setShift(unsigned S) { Shift = S; }
unsigned getShift() { return Shift; }
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
[FastISel][AArch64] Add MachO large code model support for function calls. Currently the large code model for MachO uses the GOT to make function calls. Emit the required adrp and ldr instructions to load the address from the GOT. Related to <rdar://problem/17733076>. llvm-svn: 214381
2014-07-31 12:10:40 +08:00
void setGlobalValue(const GlobalValue *G) { GV = G; }
const GlobalValue *getGlobalValue() { return GV; }
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
};
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
/// Subtarget - Keep a pointer to the AArch64Subtarget around so that we can
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
/// make the right decision when generating code for different targets.
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
const AArch64Subtarget *Subtarget;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
LLVMContext *Context;
[FastISel] Rename public visible FastISel functions. NFC. This commit renames the following public FastISel functions: LowerArguments -> lowerArguments SelectInstruction -> selectInstruction TargetSelectInstruction -> fastSelectInstruction FastLowerArguments -> fastLowerArguments FastLowerCall -> fastLowerCall FastLowerIntrinsicCall -> fastLowerIntrinsicCall FastEmitZExtFromI1 -> fastEmitZExtFromI1 FastEmitBranch -> fastEmitBranch UpdateValueMap -> updateValueMap TargetMaterializeConstant -> fastMaterializeConstant TargetMaterializeAlloca -> fastMaterializeAlloca TargetMaterializeFloatZero -> fastMaterializeFloatZero LowerCallTo -> lowerCallTo Reviewed by Eric llvm-svn: 217074
2014-09-04 04:56:52 +08:00
bool fastLowerArguments() override;
bool fastLowerCall(CallLoweringInfo &CLI) override;
bool fastLowerIntrinsicCall(const IntrinsicInst *II) override;
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
private:
// Selection routines.
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
bool selectAddSub(const Instruction *I);
[FastISel][AArch64] Refactor selectAddSub, selectLogicalOp, and SelectShift. NFC. Small refactor to tidy up the code a little. llvm-svn: 217827
2014-09-16 05:27:56 +08:00
bool selectLogicalOp(const Instruction *I);
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool selectLoad(const Instruction *I);
bool selectStore(const Instruction *I);
bool selectBranch(const Instruction *I);
bool selectIndirectBr(const Instruction *I);
bool selectCmp(const Instruction *I);
bool selectSelect(const Instruction *I);
bool selectFPExt(const Instruction *I);
bool selectFPTrunc(const Instruction *I);
bool selectFPToInt(const Instruction *I, bool Signed);
bool selectIntToFP(const Instruction *I, bool Signed);
bool selectRem(const Instruction *I, unsigned ISDOpcode);
bool selectRet(const Instruction *I);
bool selectTrunc(const Instruction *I);
bool selectIntExt(const Instruction *I);
bool selectMul(const Instruction *I);
bool selectShift(const Instruction *I);
bool selectBitCast(const Instruction *I);
[FastISel][AArch64] Add lowering support for frem. This lowers frem to a runtime libcall inside fast-isel. The test case also checks the CallLoweringInfo bug that was exposed by this change. This fixes rdar://problem/18342783. llvm-svn: 217833
2014-09-16 06:07:49 +08:00
bool selectFRem(const Instruction *I);
[FastISel][AArch64] Custom lower sdiv by power-of-2. Emit an optimized instruction sequence for sdiv by power-of-2 depending on the exact flag. This fixes rdar://problem/18224511. llvm-svn: 217986
2014-09-18 05:55:55 +08:00
bool selectSDiv(const Instruction *I);
Reapply "[FastISel][AArch64] Add custom lowering for GEPs." This is mostly a copy of the existing FastISel GEP code, but we have to duplicate it for AArch64, because otherwise we would bail out even for simple cases. This is because the standard fastEmit functions don't cover MUL at all and ADD is lowered very inefficientily. The original commit had a bug in the add emit logic, which has been fixed. llvm-svn: 219831
2014-10-16 02:58:07 +08:00
bool selectGetElementPtr(const Instruction *I);
[AArch64][FastISel] Select -O0 legal cmpxchg. At -O0, cmpxchg survives AtomicExpand: it's mostly straightforward to select it in fast-isel, and let the pseudo be expanded later. extractvalues on the result are the tricky part: the generic logic only works for legal types (and it would be painful to make it support illegal types), so we can only support i32/i64 cmpxchg. llvm-svn: 276183
2016-07-21 05:12:32 +08:00
bool selectAtomicCmpXchg(const AtomicCmpXchgInst *I);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Utility helper routines.
bool isTypeLegal(Type *Ty, MVT &VT);
[FastISel][AArch64] Refactor code to use isTypeSupported. NFC. Gets rid of isLoadStoreTypeLegal and replace it with isTypeSupported. llvm-svn: 217826
2014-09-16 05:27:54 +08:00
bool isTypeSupported(Type *Ty, MVT &VT, bool IsVectorAllowed = false);
[FastISel][AArch64] Don't fold instructions that are not in the same basic block. This fix checks first if the instruction to be folded (e.g. sign-/zero-extend, or shift) is in the same machine basic block as the instruction we are folding into. Not doing so can result in incorrect code, because the value might not be live-out of the basic block, where the value is defined. This fixes rdar://problem/18169495. llvm-svn: 216700
2014-08-29 08:19:21 +08:00
bool isValueAvailable(const Value *V) const;
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool computeAddress(const Value *Obj, Address &Addr, Type *Ty = nullptr);
bool computeCallAddress(const Value *V, Address &Addr);
bool simplifyAddress(Address &Addr, MVT VT);
void addLoadStoreOperands(Address &Addr, const MachineInstrBuilder &MIB,
[CodeGen] Take a MachineMemOperand::Flags in MachineFunction::getMachineMemOperand. Summary: Previously we took an unsigned. Hooray for type-safety. Reviewers: chandlerc Subscribers: dsanders, llvm-commits Differential Revision: http://reviews.llvm.org/D22282 llvm-svn: 275591
2016-07-16 02:26:59 +08:00
MachineMemOperand::Flags Flags,
unsigned ScaleFactor, MachineMemOperand *MMO);
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool isMemCpySmall(uint64_t Len, unsigned Alignment);
bool tryEmitSmallMemCpy(Address Dest, Address Src, uint64_t Len,
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
unsigned Alignment);
[FastISel][AArch64] Add branch folding support for the XALU intrinsics. This improves the code generation for the XALU intrinsics when the condition is feeding a branch instruction. This is related to <rdar://problem/17831117>. llvm-svn: 214349
2014-07-31 06:04:34 +08:00
bool foldXALUIntrinsic(AArch64CC::CondCode &CC, const Instruction *I,
const Value *Cond);
[FastISel][AArch64] Fix sign-/zero-extend folding when SelectionDAG is involved. Sign-/zero-extend folding depended on the load and the integer extend to be both selected by FastISel. This cannot always be garantueed and SelectionDAG might interfer. This commit adds additonal checks to load and integer extend lowering to catch this. Related to rdar://problem/18495928. llvm-svn: 219716
2014-10-15 04:36:02 +08:00
bool optimizeIntExtLoad(const Instruction *I, MVT RetVT, MVT SrcVT);
[FastISel][AArch64] Optimize select when one of the operands is a 'true' or 'false' value. Optimize selects of i1 in the presence of 'true' and 'false' operands to simple logic operations. This fixes rdar://problem/18960150. llvm-svn: 221848
2014-11-13 08:36:46 +08:00
bool optimizeSelect(const SelectInst *SI);
[FastISel][AArch64] Don't bail during simple GEP instruction selection. The generic FastISel code would bail, because it can't emit a sign-extend for AArch64. This copies the code over and uses AArch64 specific emit functions. This is not ideal and 'computeAddress' should handles this, so it can fold the address computation into the memory operation. I plan to clean up 'computeAddress' anyways, so I will add that in a future commit. Related to rdar://problem/18962471. llvm-svn: 221923
2014-11-14 04:50:44 +08:00
std::pair<unsigned, bool> getRegForGEPIndex(const Value *Idx);
[FastISel][AArch64] Add branch folding support for the XALU intrinsics. This improves the code generation for the XALU intrinsics when the condition is feeding a branch instruction. This is related to <rdar://problem/17831117>. llvm-svn: 214349
2014-07-31 06:04:34 +08:00
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
// Emit helper routines.
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
unsigned emitAddSub(bool UseAdd, MVT RetVT, const Value *LHS,
const Value *RHS, bool SetFlags = false,
bool WantResult = true, bool IsZExt = false);
unsigned emitAddSub_rr(bool UseAdd, MVT RetVT, unsigned LHSReg,
bool LHSIsKill, unsigned RHSReg, bool RHSIsKill,
bool SetFlags = false, bool WantResult = true);
unsigned emitAddSub_ri(bool UseAdd, MVT RetVT, unsigned LHSReg,
bool LHSIsKill, uint64_t Imm, bool SetFlags = false,
bool WantResult = true);
[FastISel][AArch64] Fix address simplification. When a shift with extension or an add with shift and extension cannot be folded into the memory operation, then the address calculation has to be materialized separately. While doing so the code forgot to consider a possible sign-/zero- extension. This fix folds now also the sign-/zero-extension into the add or shift instruction which is used to materialize the address. This fixes rdar://problem/18141718. llvm-svn: 216511
2014-08-27 08:58:30 +08:00
unsigned emitAddSub_rs(bool UseAdd, MVT RetVT, unsigned LHSReg,
bool LHSIsKill, unsigned RHSReg, bool RHSIsKill,
AArch64_AM::ShiftExtendType ShiftType,
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
uint64_t ShiftImm, bool SetFlags = false,
bool WantResult = true);
[FastISel][AArch64] Fix address simplification. When a shift with extension or an add with shift and extension cannot be folded into the memory operation, then the address calculation has to be materialized separately. While doing so the code forgot to consider a possible sign-/zero- extension. This fix folds now also the sign-/zero-extension into the add or shift instruction which is used to materialize the address. This fixes rdar://problem/18141718. llvm-svn: 216511
2014-08-27 08:58:30 +08:00
unsigned emitAddSub_rx(bool UseAdd, MVT RetVT, unsigned LHSReg,
bool LHSIsKill, unsigned RHSReg, bool RHSIsKill,
AArch64_AM::ShiftExtendType ExtType,
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
uint64_t ShiftImm, bool SetFlags = false,
bool WantResult = true);
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Emit functions.
Recommit r218010 [FastISel][AArch64] Fold bit test and branch into TBZ and TBNZ. Note: This version fixed an issue with the TBZ/TBNZ instructions that were generated in FastISel. The issue was that the 64bit version of TBZ (TBZX) automagically sets the upper bit of the immediate field that is used to specify the bit we want to test. To test for any of the lower 32bits we have to first extract the subregister and use the 32bit version of the TBZ instruction (TBZW). Original commit message: Teach selectBranch to fold bit test and branch into a single instruction (TBZ or TBNZ). llvm-svn: 218693
2014-10-01 03:59:35 +08:00
bool emitCompareAndBranch(const BranchInst *BI);
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
bool emitCmp(const Value *LHS, const Value *RHS, bool IsZExt);
bool emitICmp(MVT RetVT, const Value *LHS, const Value *RHS, bool IsZExt);
bool emitICmp_ri(MVT RetVT, unsigned LHSReg, bool LHSIsKill, uint64_t Imm);
bool emitFCmp(MVT RetVT, const Value *LHS, const Value *RHS);
[FastISel][AArch64] Fix sign-/zero-extend folding when SelectionDAG is involved. Sign-/zero-extend folding depended on the load and the integer extend to be both selected by FastISel. This cannot always be garantueed and SelectionDAG might interfer. This commit adds additonal checks to load and integer extend lowering to catch this. Related to rdar://problem/18495928. llvm-svn: 219716
2014-10-15 04:36:02 +08:00
unsigned emitLoad(MVT VT, MVT ResultVT, Address Addr, bool WantZExt = true,
MachineMemOperand *MMO = nullptr);
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool emitStore(MVT VT, unsigned SrcReg, Address Addr,
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
MachineMemOperand *MMO = nullptr);
[AArch64][FastISel] Select atomic stores into STLR. llvm-svn: 276182
2016-07-21 05:12:27 +08:00
bool emitStoreRelease(MVT VT, unsigned SrcReg, unsigned AddrReg,
MachineMemOperand *MMO = nullptr);
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
unsigned emitIntExt(MVT SrcVT, unsigned SrcReg, MVT DestVT, bool isZExt);
unsigned emiti1Ext(unsigned SrcReg, MVT DestVT, bool isZExt);
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
unsigned emitAdd(MVT RetVT, const Value *LHS, const Value *RHS,
bool SetFlags = false, bool WantResult = true,
bool IsZExt = false);
[FastISel][AArch64] Factor out add with immediate emission into a helper function. NFC. Simplify add with immediate emission by factoring it out into a helper function. llvm-svn: 219830
2014-10-16 02:58:02 +08:00
unsigned emitAdd_ri_(MVT VT, unsigned Op0, bool Op0IsKill, int64_t Imm);
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
unsigned emitSub(MVT RetVT, const Value *LHS, const Value *RHS,
bool SetFlags = false, bool WantResult = true,
bool IsZExt = false);
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
unsigned emitSubs_rr(MVT RetVT, unsigned LHSReg, bool LHSIsKill,
unsigned RHSReg, bool RHSIsKill, bool WantResult = true);
unsigned emitSubs_rs(MVT RetVT, unsigned LHSReg, bool LHSIsKill,
unsigned RHSReg, bool RHSIsKill,
AArch64_AM::ShiftExtendType ShiftType, uint64_t ShiftImm,
bool WantResult = true);
[FastISel][AArch64] Add target-specific lowering for logical operations. This change adds support for immediate and shift-left folding into logical operations. This fixes rdar://problem/18223183. llvm-svn: 217118
2014-09-04 09:29:18 +08:00
unsigned emitLogicalOp(unsigned ISDOpc, MVT RetVT, const Value *LHS,
const Value *RHS);
unsigned emitLogicalOp_ri(unsigned ISDOpc, MVT RetVT, unsigned LHSReg,
bool LHSIsKill, uint64_t Imm);
unsigned emitLogicalOp_rs(unsigned ISDOpc, MVT RetVT, unsigned LHSReg,
bool LHSIsKill, unsigned RHSReg, bool RHSIsKill,
uint64_t ShiftImm);
unsigned emitAnd_ri(MVT RetVT, unsigned LHSReg, bool LHSIsKill, uint64_t Imm);
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
unsigned emitMul_rr(MVT RetVT, unsigned Op0, bool Op0IsKill,
unsigned Op1, bool Op1IsKill);
unsigned emitSMULL_rr(MVT RetVT, unsigned Op0, bool Op0IsKill,
unsigned Op1, bool Op1IsKill);
unsigned emitUMULL_rr(MVT RetVT, unsigned Op0, bool Op0IsKill,
unsigned Op1, bool Op1IsKill);
[FastISel][AArch64] Add support for variable shift. This adds the missing variable shift support for value type i8, i16, and i32. This fixes <rdar://problem/18095685>. llvm-svn: 216242
2014-08-22 07:06:07 +08:00
unsigned emitLSL_rr(MVT RetVT, unsigned Op0Reg, bool Op0IsKill,
unsigned Op1Reg, bool Op1IsKill);
[FastISel][AArch64] Fold Sign-/Zero-Extend into the shift immediate instruction. llvm-svn: 216510
2014-08-27 08:58:26 +08:00
unsigned emitLSL_ri(MVT RetVT, MVT SrcVT, unsigned Op0Reg, bool Op0IsKill,
uint64_t Imm, bool IsZExt = true);
[FastISel][AArch64] Add support for variable shift. This adds the missing variable shift support for value type i8, i16, and i32. This fixes <rdar://problem/18095685>. llvm-svn: 216242
2014-08-22 07:06:07 +08:00
unsigned emitLSR_rr(MVT RetVT, unsigned Op0Reg, bool Op0IsKill,
unsigned Op1Reg, bool Op1IsKill);
[FastISel][AArch64] Fold Sign-/Zero-Extend into the shift immediate instruction. llvm-svn: 216510
2014-08-27 08:58:26 +08:00
unsigned emitLSR_ri(MVT RetVT, MVT SrcVT, unsigned Op0Reg, bool Op0IsKill,
uint64_t Imm, bool IsZExt = true);
[FastISel][AArch64] Add support for variable shift. This adds the missing variable shift support for value type i8, i16, and i32. This fixes <rdar://problem/18095685>. llvm-svn: 216242
2014-08-22 07:06:07 +08:00
unsigned emitASR_rr(MVT RetVT, unsigned Op0Reg, bool Op0IsKill,
unsigned Op1Reg, bool Op1IsKill);
[FastISel][AArch64] Fold Sign-/Zero-Extend into the shift immediate instruction. llvm-svn: 216510
2014-08-27 08:58:26 +08:00
unsigned emitASR_ri(MVT RetVT, MVT SrcVT, unsigned Op0Reg, bool Op0IsKill,
uint64_t Imm, bool IsZExt = false);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
unsigned materializeInt(const ConstantInt *CI, MVT VT);
unsigned materializeFP(const ConstantFP *CFP, MVT VT);
unsigned materializeGV(const GlobalValue *GV);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Call handling routines.
private:
CCAssignFn *CCAssignFnForCall(CallingConv::ID CC) const;
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool processCallArgs(CallLoweringInfo &CLI, SmallVectorImpl<MVT> &ArgVTs,
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
unsigned &NumBytes);
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool finishCall(CallLoweringInfo &CLI, MVT RetVT, unsigned NumBytes);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
public:
// Backend specific FastISel code.
[FastISel] Rename public visible FastISel functions. NFC. This commit renames the following public FastISel functions: LowerArguments -> lowerArguments SelectInstruction -> selectInstruction TargetSelectInstruction -> fastSelectInstruction FastLowerArguments -> fastLowerArguments FastLowerCall -> fastLowerCall FastLowerIntrinsicCall -> fastLowerIntrinsicCall FastEmitZExtFromI1 -> fastEmitZExtFromI1 FastEmitBranch -> fastEmitBranch UpdateValueMap -> updateValueMap TargetMaterializeConstant -> fastMaterializeConstant TargetMaterializeAlloca -> fastMaterializeAlloca TargetMaterializeFloatZero -> fastMaterializeFloatZero LowerCallTo -> lowerCallTo Reviewed by Eric llvm-svn: 217074
2014-09-04 04:56:52 +08:00
unsigned fastMaterializeAlloca(const AllocaInst *AI) override;
unsigned fastMaterializeConstant(const Constant *C) override;
unsigned fastMaterializeFloatZero(const ConstantFP* CF) override;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
[FastISel][AArch64] Move over to target-dependent instruction selection only. This change moves FastISel for AArch64 to target-dependent instruction selection only. This change replicates the existing target-independent behavior, therefore there are no changes to the unit tests or new tests. Future changes will take advantage of this change and update functionality and unit tests. llvm-svn: 216955
2014-09-03 05:32:54 +08:00
explicit AArch64FastISel(FunctionLoweringInfo &FuncInfo,
Clean up some uses of getSubtarget in AArch64. llvm-svn: 227530
2015-01-30 09:10:24 +08:00
const TargetLibraryInfo *LibInfo)
[FastISel][AArch64] Move over to target-dependent instruction selection only. This change moves FastISel for AArch64 to target-dependent instruction selection only. This change replicates the existing target-independent behavior, therefore there are no changes to the unit tests or new tests. Future changes will take advantage of this change and update functionality and unit tests. llvm-svn: 216955
2014-09-03 05:32:54 +08:00
: FastISel(FuncInfo, LibInfo, /*SkipTargetIndependentISel=*/true) {
Clean up some uses of getSubtarget in AArch64. llvm-svn: 227530
2015-01-30 09:10:24 +08:00
Subtarget =
&static_cast<const AArch64Subtarget &>(FuncInfo.MF->getSubtarget());
[FastISel][AArch64] Move over to target-dependent instruction selection only. This change moves FastISel for AArch64 to target-dependent instruction selection only. This change replicates the existing target-independent behavior, therefore there are no changes to the unit tests or new tests. Future changes will take advantage of this change and update functionality and unit tests. llvm-svn: 216955
2014-09-03 05:32:54 +08:00
Context = &FuncInfo.Fn->getContext();
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
}
[FastISel] Rename public visible FastISel functions. NFC. This commit renames the following public FastISel functions: LowerArguments -> lowerArguments SelectInstruction -> selectInstruction TargetSelectInstruction -> fastSelectInstruction FastLowerArguments -> fastLowerArguments FastLowerCall -> fastLowerCall FastLowerIntrinsicCall -> fastLowerIntrinsicCall FastEmitZExtFromI1 -> fastEmitZExtFromI1 FastEmitBranch -> fastEmitBranch UpdateValueMap -> updateValueMap TargetMaterializeConstant -> fastMaterializeConstant TargetMaterializeAlloca -> fastMaterializeAlloca TargetMaterializeFloatZero -> fastMaterializeFloatZero LowerCallTo -> lowerCallTo Reviewed by Eric llvm-svn: 217074
2014-09-04 04:56:52 +08:00
bool fastSelectInstruction(const Instruction *I) override;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
#include "AArch64GenFastISel.inc"
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
};
} // end anonymous namespace
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
#include "AArch64GenCallingConv.inc"
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
/// \brief Check if the sign-/zero-extend will be a noop.
static bool isIntExtFree(const Instruction *I) {
assert((isa<ZExtInst>(I) || isa<SExtInst>(I)) &&
"Unexpected integer extend instruction.");
[FastISel][AArch64] Fix "Fold sign-/zero-extends into the load instruction." This commit fixes an issue with sign-/zero-extending loads that was discovered by Richard Barton. We use now the correct load instructions for sign-extending loads to 64bit. Also updated and added more unit tests. llvm-svn: 219185
2014-10-07 11:39:59 +08:00
assert(!I->getType()->isVectorTy() && I->getType()->isIntegerTy() &&
"Unexpected value type.");
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
bool IsZExt = isa<ZExtInst>(I);
if (const auto *LI = dyn_cast<LoadInst>(I->getOperand(0)))
if (LI->hasOneUse())
return true;
if (const auto *Arg = dyn_cast<Argument>(I->getOperand(0)))
if ((IsZExt && Arg->hasZExtAttr()) || (!IsZExt && Arg->hasSExtAttr()))
return true;
return false;
}
[FastISel][AArch64] Factor out scale factor calculation. NFC. Factor out the code that determines the implicit scale factor of memory operations for a given value type. llvm-svn: 218652
2014-09-30 08:49:54 +08:00
/// \brief Determine the implicit scale factor that is applied by a memory
/// operation for a given value type.
static unsigned getImplicitScaleFactor(MVT VT) {
switch (VT.SimpleTy) {
default:
return 0; // invalid
case MVT::i1: // fall-through
case MVT::i8:
return 1;
case MVT::i16:
return 2;
case MVT::i32: // fall-through
case MVT::f32:
return 4;
case MVT::i64: // fall-through
case MVT::f64:
return 8;
}
}
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
CCAssignFn *AArch64FastISel::CCAssignFnForCall(CallingConv::ID CC) const {
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
if (CC == CallingConv::WebKit_JS)
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
return CC_AArch64_WebKit_JS;
[AArch64] Implement GHC calling convention Original patch by Luke Iannini. Minor improvements and test added by Erik de Castro Lopo. Differential Revision: http://reviews.llvm.org/D6877 From: Erik de Castro Lopo <erikd@mega-nerd.com> llvm-svn: 226473
2015-01-20 01:40:05 +08:00
if (CC == CallingConv::GHC)
return CC_AArch64_GHC;
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
return Subtarget->isTargetDarwin() ? CC_AArch64_DarwinPCS : CC_AArch64_AAPCS;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
}
[FastISel] Rename public visible FastISel functions. NFC. This commit renames the following public FastISel functions: LowerArguments -> lowerArguments SelectInstruction -> selectInstruction TargetSelectInstruction -> fastSelectInstruction FastLowerArguments -> fastLowerArguments FastLowerCall -> fastLowerCall FastLowerIntrinsicCall -> fastLowerIntrinsicCall FastEmitZExtFromI1 -> fastEmitZExtFromI1 FastEmitBranch -> fastEmitBranch UpdateValueMap -> updateValueMap TargetMaterializeConstant -> fastMaterializeConstant TargetMaterializeAlloca -> fastMaterializeAlloca TargetMaterializeFloatZero -> fastMaterializeFloatZero LowerCallTo -> lowerCallTo Reviewed by Eric llvm-svn: 217074
2014-09-04 04:56:52 +08:00
unsigned AArch64FastISel::fastMaterializeAlloca(const AllocaInst *AI) {
Make TargetLowering::getPointerTy() taking DataLayout as an argument Summary: This change is part of a series of commits dedicated to have a single DataLayout during compilation by using always the one owned by the module. Reviewers: echristo Subscribers: jholewinski, ted, yaron.keren, rafael, llvm-commits Differential Revision: http://reviews.llvm.org/D11028 From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 241775
2015-07-09 10:09:04 +08:00
assert(TLI.getValueType(DL, AI->getType(), true) == MVT::i64 &&
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
"Alloca should always return a pointer.");
// Don't handle dynamic allocas.
if (!FuncInfo.StaticAllocaMap.count(AI))
return 0;
DenseMap<const AllocaInst *, int>::iterator SI =
FuncInfo.StaticAllocaMap.find(AI);
if (SI != FuncInfo.StaticAllocaMap.end()) {
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
unsigned ResultReg = createResultReg(&AArch64::GPR64spRegClass);
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(AArch64::ADDXri),
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
ResultReg)
.addFrameIndex(SI->second)
.addImm(0)
.addImm(0);
return ResultReg;
}
return 0;
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
unsigned AArch64FastISel::materializeInt(const ConstantInt *CI, MVT VT) {
Reapplying [FastISel][AArch64] Cleanup constant materialization code. NFCI. Note: This reapplies r215582 without any modifications. The refactoring wasn't responsible for the buildbot failures. Original commit message: Cleanup and prepare constant materialization code for future commits. llvm-svn: 215752
2014-08-16 02:55:52 +08:00
if (VT > MVT::i64)
return 0;
Reapply [FastISel][AArch64] Make use of the zero register when possible (r215591). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: This change materializes now the value "0" from the zero register. The zero register can be folded by several instruction, so no materialization is need at all. Fixes <rdar://problem/17924413>. llvm-svn: 216009
2014-08-20 03:44:02 +08:00
if (!CI->isZero())
[FastISel][tblgen] Rename tblgen generated FastISel functions. NFC. This is the final round of renaming. This changes tblgen to emit lower-case function names for FastEmitInst_* and FastEmit_*, and updates all its uses in the source code. Reviewed by Eric llvm-svn: 217075
2014-09-04 04:56:59 +08:00
return fastEmit_i(VT, VT, ISD::Constant, CI->getZExtValue());
Reapply [FastISel][AArch64] Make use of the zero register when possible (r215591). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: This change materializes now the value "0" from the zero register. The zero register can be folded by several instruction, so no materialization is need at all. Fixes <rdar://problem/17924413>. llvm-svn: 216009
2014-08-20 03:44:02 +08:00
// Create a copy from the zero register to materialize a "0" value.
const TargetRegisterClass *RC = (VT == MVT::i64) ? &AArch64::GPR64RegClass
: &AArch64::GPR32RegClass;
unsigned ZeroReg = (VT == MVT::i64) ? AArch64::XZR : AArch64::WZR;
unsigned ResultReg = createResultReg(RC);
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(TargetOpcode::COPY),
ResultReg).addReg(ZeroReg, getKillRegState(true));
Reapply [FastISel][AArch64] Make use of the zero register when possible (r215591). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: This change materializes now the value "0" from the zero register. The zero register can be folded by several instruction, so no materialization is need at all. Fixes <rdar://problem/17924413>. llvm-svn: 216009
2014-08-20 03:44:02 +08:00
return ResultReg;
Reapplying [FastISel][AArch64] Cleanup constant materialization code. NFCI. Note: This reapplies r215582 without any modifications. The refactoring wasn't responsible for the buildbot failures. Original commit message: Cleanup and prepare constant materialization code for future commits. llvm-svn: 215752
2014-08-16 02:55:52 +08:00
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
unsigned AArch64FastISel::materializeFP(const ConstantFP *CFP, MVT VT) {
[FastISel][AArch64] Refactor float zero materialization. NFCI. llvm-svn: 216403
2014-08-26 03:58:05 +08:00
// Positive zero (+0.0) has to be materialized with a fmov from the zero
// register, because the immediate version of fmov cannot encode zero.
if (CFP->isNullValue())
[FastISel] Rename public visible FastISel functions. NFC. This commit renames the following public FastISel functions: LowerArguments -> lowerArguments SelectInstruction -> selectInstruction TargetSelectInstruction -> fastSelectInstruction FastLowerArguments -> fastLowerArguments FastLowerCall -> fastLowerCall FastLowerIntrinsicCall -> fastLowerIntrinsicCall FastEmitZExtFromI1 -> fastEmitZExtFromI1 FastEmitBranch -> fastEmitBranch UpdateValueMap -> updateValueMap TargetMaterializeConstant -> fastMaterializeConstant TargetMaterializeAlloca -> fastMaterializeAlloca TargetMaterializeFloatZero -> fastMaterializeFloatZero LowerCallTo -> lowerCallTo Reviewed by Eric llvm-svn: 217074
2014-09-04 04:56:52 +08:00
return fastMaterializeFloatZero(CFP);
[FastISel][AArch64] Refactor float zero materialization. NFCI. llvm-svn: 216403
2014-08-26 03:58:05 +08:00
[ARM64][fast-isel] Fast-isel doesn't know how to handle f128. llvm-svn: 207659
2014-04-30 23:29:57 +08:00
if (VT != MVT::f32 && VT != MVT::f64)
return 0;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
const APFloat Val = CFP->getValueAPF();
Reapplying [FastISel][AArch64] Cleanup constant materialization code. NFCI. Note: This reapplies r215582 without any modifications. The refactoring wasn't responsible for the buildbot failures. Original commit message: Cleanup and prepare constant materialization code for future commits. llvm-svn: 215752
2014-08-16 02:55:52 +08:00
bool Is64Bit = (VT == MVT::f64);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// This checks to see if we can use FMOV instructions to materialize
// a constant, otherwise we have to materialize via the constant pool.
if (TLI.isFPImmLegal(Val, VT)) {
[FastISel][AArch64] Refactor float zero materialization. NFCI. llvm-svn: 216403
2014-08-26 03:58:05 +08:00
int Imm =
Is64Bit ? AArch64_AM::getFP64Imm(Val) : AArch64_AM::getFP32Imm(Val);
assert((Imm != -1) && "Cannot encode floating-point constant.");
Reapplying [FastISel][AArch64] Cleanup constant materialization code. NFCI. Note: This reapplies r215582 without any modifications. The refactoring wasn't responsible for the buildbot failures. Original commit message: Cleanup and prepare constant materialization code for future commits. llvm-svn: 215752
2014-08-16 02:55:52 +08:00
unsigned Opc = Is64Bit ? AArch64::FMOVDi : AArch64::FMOVSi;
[FastISel][tblgen] Rename tblgen generated FastISel functions. NFC. This is the final round of renaming. This changes tblgen to emit lower-case function names for FastEmitInst_* and FastEmit_*, and updates all its uses in the source code. Reviewed by Eric llvm-svn: 217075
2014-09-04 04:56:59 +08:00
return fastEmitInst_i(Opc, TLI.getRegClassFor(VT), Imm);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
}
[AArch64] MachO large code-model: Materialize FP constants in code. In the large code model we have to first get the address of the GOT entry, load the address of the constant, and then load the constant itself. To avoid these loads and the GOT entry alltogether this commit changes the way how FP constants are materialized in the large code model. The constats are now materialized in a GPR and then bitconverted/moved into the FPR. Reviewed by Tim Northover Fixes rdar://problem/16572564. llvm-svn: 223941
2014-12-11 03:43:32 +08:00
// For the MachO large code model materialize the FP constant in code.
if (Subtarget->isTargetMachO() && TM.getCodeModel() == CodeModel::Large) {
unsigned Opc1 = Is64Bit ? AArch64::MOVi64imm : AArch64::MOVi32imm;
const TargetRegisterClass *RC = Is64Bit ?
&AArch64::GPR64RegClass : &AArch64::GPR32RegClass;
unsigned TmpReg = createResultReg(RC);
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(Opc1), TmpReg)
.addImm(CFP->getValueAPF().bitcastToAPInt().getZExtValue());
unsigned ResultReg = createResultReg(TLI.getRegClassFor(VT));
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc,
TII.get(TargetOpcode::COPY), ResultReg)
.addReg(TmpReg, getKillRegState(true));
return ResultReg;
}
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Materialize via constant pool. MachineConstantPool wants an explicit
// alignment.
unsigned Align = DL.getPrefTypeAlignment(CFP->getType());
if (Align == 0)
Align = DL.getTypeAllocSize(CFP->getType());
Reapplying [FastISel][AArch64] Cleanup constant materialization code. NFCI. Note: This reapplies r215582 without any modifications. The refactoring wasn't responsible for the buildbot failures. Original commit message: Cleanup and prepare constant materialization code for future commits. llvm-svn: 215752
2014-08-16 02:55:52 +08:00
unsigned CPI = MCP.getConstantPoolIndex(cast<Constant>(CFP), Align);
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
unsigned ADRPReg = createResultReg(&AArch64::GPR64commonRegClass);
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(AArch64::ADRP),
[FastISel][AArch64] Refactor float zero materialization. NFCI. llvm-svn: 216403
2014-08-26 03:58:05 +08:00
ADRPReg).addConstantPoolIndex(CPI, 0, AArch64II::MO_PAGE);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
Reapplying [FastISel][AArch64] Cleanup constant materialization code. NFCI. Note: This reapplies r215582 without any modifications. The refactoring wasn't responsible for the buildbot failures. Original commit message: Cleanup and prepare constant materialization code for future commits. llvm-svn: 215752
2014-08-16 02:55:52 +08:00
unsigned Opc = Is64Bit ? AArch64::LDRDui : AArch64::LDRSui;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
unsigned ResultReg = createResultReg(TLI.getRegClassFor(VT));
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(Opc), ResultReg)
[FastISel][AArch64] Refactor float zero materialization. NFCI. llvm-svn: 216403
2014-08-26 03:58:05 +08:00
.addReg(ADRPReg)
.addConstantPoolIndex(CPI, 0, AArch64II::MO_PAGEOFF | AArch64II::MO_NC);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return ResultReg;
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
unsigned AArch64FastISel::materializeGV(const GlobalValue *GV) {
[pr19844] Add thread local mode to aliases. This matches gcc's behavior. It also seems natural given that aliases contain other properties that govern how it is accessed (linkage, visibility, dll storage). Clang still has to be updated to expose this feature to C. llvm-svn: 209759
2014-05-29 02:15:43 +08:00
// We can't handle thread-local variables quickly yet.
if (GV->isThreadLocal())
return 0;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
AArch64: disable FastISel for large code model. The code emitted is what would be expected for the small model, so it shouldn't be used when objects can be the full 64-bits away. This fixes MCJIT tests on Linux. llvm-svn: 209585
2014-05-25 03:45:41 +08:00
// MachO still uses GOT for large code-model accesses, but ELF requires
// movz/movk sequences, which FastISel doesn't handle yet.
if (TM.getCodeModel() != CodeModel::Small && !Subtarget->isTargetMachO())
return 0;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
unsigned char OpFlags = Subtarget->ClassifyGlobalReference(GV, TM);
Make TargetLowering::getPointerTy() taking DataLayout as an argument Summary: This change is part of a series of commits dedicated to have a single DataLayout during compilation by using always the one owned by the module. Reviewers: echristo Subscribers: jholewinski, ted, yaron.keren, rafael, llvm-commits Differential Revision: http://reviews.llvm.org/D11028 From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 241775
2015-07-09 10:09:04 +08:00
EVT DestEVT = TLI.getValueType(DL, GV->getType(), true);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
if (!DestEVT.isSimple())
return 0;
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
unsigned ADRPReg = createResultReg(&AArch64::GPR64commonRegClass);
ARM64: add constraints to various FastISel operations llvm-svn: 206284
2014-04-15 21:59:53 +08:00
unsigned ResultReg;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
if (OpFlags & AArch64II::MO_GOT) {
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// ADRP + LDRX
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(AArch64::ADRP),
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
ADRPReg)
Reapplying [FastISel][AArch64] Cleanup constant materialization code. NFCI. Note: This reapplies r215582 without any modifications. The refactoring wasn't responsible for the buildbot failures. Original commit message: Cleanup and prepare constant materialization code for future commits. llvm-svn: 215752
2014-08-16 02:55:52 +08:00
.addGlobalAddress(GV, 0, AArch64II::MO_GOT | AArch64II::MO_PAGE);
ARM64: add constraints to various FastISel operations llvm-svn: 206284
2014-04-15 21:59:53 +08:00
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
ResultReg = createResultReg(&AArch64::GPR64RegClass);
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(AArch64::LDRXui),
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
ResultReg)
Reapplying [FastISel][AArch64] Cleanup constant materialization code. NFCI. Note: This reapplies r215582 without any modifications. The refactoring wasn't responsible for the buildbot failures. Original commit message: Cleanup and prepare constant materialization code for future commits. llvm-svn: 215752
2014-08-16 02:55:52 +08:00
.addReg(ADRPReg)
.addGlobalAddress(GV, 0, AArch64II::MO_GOT | AArch64II::MO_PAGEOFF |
AArch64II::MO_NC);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
} else {
// ADRP + ADDX
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(AArch64::ADRP),
Reapplying [FastISel][AArch64] Cleanup constant materialization code. NFCI. Note: This reapplies r215582 without any modifications. The refactoring wasn't responsible for the buildbot failures. Original commit message: Cleanup and prepare constant materialization code for future commits. llvm-svn: 215752
2014-08-16 02:55:52 +08:00
ADRPReg)
.addGlobalAddress(GV, 0, AArch64II::MO_PAGE);
ARM64: add constraints to various FastISel operations llvm-svn: 206284
2014-04-15 21:59:53 +08:00
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
ResultReg = createResultReg(&AArch64::GPR64spRegClass);
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(AArch64::ADDXri),
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
ResultReg)
Reapplying [FastISel][AArch64] Cleanup constant materialization code. NFCI. Note: This reapplies r215582 without any modifications. The refactoring wasn't responsible for the buildbot failures. Original commit message: Cleanup and prepare constant materialization code for future commits. llvm-svn: 215752
2014-08-16 02:55:52 +08:00
.addReg(ADRPReg)
.addGlobalAddress(GV, 0, AArch64II::MO_PAGEOFF | AArch64II::MO_NC)
.addImm(0);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
}
return ResultReg;
}
[FastISel] Rename public visible FastISel functions. NFC. This commit renames the following public FastISel functions: LowerArguments -> lowerArguments SelectInstruction -> selectInstruction TargetSelectInstruction -> fastSelectInstruction FastLowerArguments -> fastLowerArguments FastLowerCall -> fastLowerCall FastLowerIntrinsicCall -> fastLowerIntrinsicCall FastEmitZExtFromI1 -> fastEmitZExtFromI1 FastEmitBranch -> fastEmitBranch UpdateValueMap -> updateValueMap TargetMaterializeConstant -> fastMaterializeConstant TargetMaterializeAlloca -> fastMaterializeAlloca TargetMaterializeFloatZero -> fastMaterializeFloatZero LowerCallTo -> lowerCallTo Reviewed by Eric llvm-svn: 217074
2014-09-04 04:56:52 +08:00
unsigned AArch64FastISel::fastMaterializeConstant(const Constant *C) {
Make TargetLowering::getPointerTy() taking DataLayout as an argument Summary: This change is part of a series of commits dedicated to have a single DataLayout during compilation by using always the one owned by the module. Reviewers: echristo Subscribers: jholewinski, ted, yaron.keren, rafael, llvm-commits Differential Revision: http://reviews.llvm.org/D11028 From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 241775
2015-07-09 10:09:04 +08:00
EVT CEVT = TLI.getValueType(DL, C->getType(), true);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Only handle simple types.
if (!CEVT.isSimple())
return 0;
MVT VT = CEVT.getSimpleVT();
Reapplying [FastISel][AArch64] Cleanup constant materialization code. NFCI. Note: This reapplies r215582 without any modifications. The refactoring wasn't responsible for the buildbot failures. Original commit message: Cleanup and prepare constant materialization code for future commits. llvm-svn: 215752
2014-08-16 02:55:52 +08:00
if (const auto *CI = dyn_cast<ConstantInt>(C))
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return materializeInt(CI, VT);
Reapplying [FastISel][AArch64] Cleanup constant materialization code. NFCI. Note: This reapplies r215582 without any modifications. The refactoring wasn't responsible for the buildbot failures. Original commit message: Cleanup and prepare constant materialization code for future commits. llvm-svn: 215752
2014-08-16 02:55:52 +08:00
else if (const ConstantFP *CFP = dyn_cast<ConstantFP>(C))
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return materializeFP(CFP, VT);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
else if (const GlobalValue *GV = dyn_cast<GlobalValue>(C))
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return materializeGV(GV);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return 0;
}
[FastISel] Rename public visible FastISel functions. NFC. This commit renames the following public FastISel functions: LowerArguments -> lowerArguments SelectInstruction -> selectInstruction TargetSelectInstruction -> fastSelectInstruction FastLowerArguments -> fastLowerArguments FastLowerCall -> fastLowerCall FastLowerIntrinsicCall -> fastLowerIntrinsicCall FastEmitZExtFromI1 -> fastEmitZExtFromI1 FastEmitBranch -> fastEmitBranch UpdateValueMap -> updateValueMap TargetMaterializeConstant -> fastMaterializeConstant TargetMaterializeAlloca -> fastMaterializeAlloca TargetMaterializeFloatZero -> fastMaterializeFloatZero LowerCallTo -> lowerCallTo Reviewed by Eric llvm-svn: 217074
2014-09-04 04:56:52 +08:00
unsigned AArch64FastISel::fastMaterializeFloatZero(const ConstantFP* CFP) {
[FastISel][AArch64] Refactor float zero materialization. NFCI. llvm-svn: 216403
2014-08-26 03:58:05 +08:00
assert(CFP->isNullValue() &&
"Floating-point constant is not a positive zero.");
MVT VT;
if (!isTypeLegal(CFP->getType(), VT))
return 0;
if (VT != MVT::f32 && VT != MVT::f64)
return 0;
bool Is64Bit = (VT == MVT::f64);
unsigned ZReg = Is64Bit ? AArch64::XZR : AArch64::WZR;
unsigned Opc = Is64Bit ? AArch64::FMOVXDr : AArch64::FMOVWSr;
[FastISel][tblgen] Rename tblgen generated FastISel functions. NFC. This is the final round of renaming. This changes tblgen to emit lower-case function names for FastEmitInst_* and FastEmit_*, and updates all its uses in the source code. Reviewed by Eric llvm-svn: 217075
2014-09-04 04:56:59 +08:00
return fastEmitInst_r(Opc, TLI.getRegClassFor(VT), ZReg, /*IsKill=*/true);
[FastISel][AArch64] Refactor float zero materialization. NFCI. llvm-svn: 216403
2014-08-26 03:58:05 +08:00
}
[FastISel][AArch64] Fold mul into the address computation of memory operations. Teach 'computeAddress' to also fold multiplies into the address computation (when possible). This fixes rdar://problem/18369443. llvm-svn: 217977
2014-09-18 03:19:31 +08:00
/// \brief Check if the multiply is by a power-of-2 constant.
static bool isMulPowOf2(const Value *I) {
if (const auto *MI = dyn_cast<MulOperator>(I)) {
if (const auto *C = dyn_cast<ConstantInt>(MI->getOperand(0)))
if (C->getValue().isPowerOf2())
return true;
if (const auto *C = dyn_cast<ConstantInt>(MI->getOperand(1)))
if (C->getValue().isPowerOf2())
return true;
}
return false;
}
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Computes the address to get to an object.
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool AArch64FastISel::computeAddress(const Value *Obj, Address &Addr, Type *Ty)
Revert "[FastISel][AArch64] Don't fold instructions too aggressively into the memory operation." Quentin pointed out that this is not the correct approach and there is a better and easier solution. llvm-svn: 216632
2014-08-28 07:09:40 +08:00
{
[C++] Use 'nullptr'. Target edition. llvm-svn: 207197
2014-04-25 13:30:21 +08:00
const User *U = nullptr;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
unsigned Opcode = Instruction::UserOp1;
if (const Instruction *I = dyn_cast<Instruction>(Obj)) {
// Don't walk into other basic blocks unless the object is an alloca from
// another block, otherwise it may not have a virtual register assigned.
if (FuncInfo.StaticAllocaMap.count(static_cast<const AllocaInst *>(Obj)) ||
FuncInfo.MBBMap[I->getParent()] == FuncInfo.MBB) {
Opcode = I->getOpcode();
U = I;
}
} else if (const ConstantExpr *C = dyn_cast<ConstantExpr>(Obj)) {
Opcode = C->getOpcode();
U = C;
}
De-constify pointers to Type since they can't be modified. NFC This was already done in most places a while ago. This just fixes the ones that crept in over time. llvm-svn: 243842
2015-08-02 06:20:21 +08:00
if (auto *Ty = dyn_cast<PointerType>(Obj->getType()))
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
if (Ty->getAddressSpace() > 255)
// Fast instruction selection doesn't support the special
// address spaces.
return false;
switch (Opcode) {
default:
break;
case Instruction::BitCast: {
// Look through bitcasts.
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return computeAddress(U->getOperand(0), Addr, Ty);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
}
case Instruction::IntToPtr: {
// Look past no-op inttoptrs.
Make TargetLowering::getPointerTy() taking DataLayout as an argument Summary: This change is part of a series of commits dedicated to have a single DataLayout during compilation by using always the one owned by the module. Reviewers: echristo Subscribers: jholewinski, ted, yaron.keren, rafael, llvm-commits Differential Revision: http://reviews.llvm.org/D11028 From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 241775
2015-07-09 10:09:04 +08:00
if (TLI.getValueType(DL, U->getOperand(0)->getType()) ==
TLI.getPointerTy(DL))
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return computeAddress(U->getOperand(0), Addr, Ty);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
break;
}
case Instruction::PtrToInt: {
Revert "[FastISel][AArch64] Don't fold instructions too aggressively into the memory operation." Quentin pointed out that this is not the correct approach and there is a better and easier solution. llvm-svn: 216632
2014-08-28 07:09:40 +08:00
// Look past no-op ptrtoints.
Make TargetLowering::getPointerTy() taking DataLayout as an argument Summary: This change is part of a series of commits dedicated to have a single DataLayout during compilation by using always the one owned by the module. Reviewers: echristo Subscribers: jholewinski, ted, yaron.keren, rafael, llvm-commits Differential Revision: http://reviews.llvm.org/D11028 From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 241775
2015-07-09 10:09:04 +08:00
if (TLI.getValueType(DL, U->getType()) == TLI.getPointerTy(DL))
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return computeAddress(U->getOperand(0), Addr, Ty);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
break;
}
case Instruction::GetElementPtr: {
Address SavedAddr = Addr;
uint64_t TmpOffset = Addr.getOffset();
// Iterate through the GEP folding the constants into offsets where
// we can.
[NFC] Replace several manual GEP loops with gep_type_iterator. Reviewers: dblaikie Subscribers: llvm-commits Differential Revision: http://reviews.llvm.org/D16335 llvm-svn: 258262
2016-01-20 08:26:52 +08:00
for (gep_type_iterator GTI = gep_type_begin(U), E = gep_type_end(U);
GTI != E; ++GTI) {
const Value *Op = GTI.getOperand();
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
if (StructType *STy = dyn_cast<StructType>(*GTI)) {
const StructLayout *SL = DL.getStructLayout(STy);
unsigned Idx = cast<ConstantInt>(Op)->getZExtValue();
TmpOffset += SL->getElementOffset(Idx);
} else {
uint64_t S = DL.getTypeAllocSize(GTI.getIndexedType());
for (;;) {
if (const ConstantInt *CI = dyn_cast<ConstantInt>(Op)) {
// Constant-offset addressing.
TmpOffset += CI->getSExtValue() * S;
break;
}
if (canFoldAddIntoGEP(U, Op)) {
// A compatible add with a constant operand. Fold the constant.
ConstantInt *CI =
cast<ConstantInt>(cast<AddOperator>(Op)->getOperand(1));
TmpOffset += CI->getSExtValue() * S;
// Iterate on the other operand.
Op = cast<AddOperator>(Op)->getOperand(0);
continue;
}
// Unsupported
goto unsupported_gep;
}
}
}
// Try to grab the base operand now.
Addr.setOffset(TmpOffset);
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
if (computeAddress(U->getOperand(0), Addr, Ty))
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return true;
// We failed, restore everything and try the other options.
Addr = SavedAddr;
unsupported_gep:
break;
}
case Instruction::Alloca: {
const AllocaInst *AI = cast<AllocaInst>(Obj);
DenseMap<const AllocaInst *, int>::iterator SI =
FuncInfo.StaticAllocaMap.find(AI);
if (SI != FuncInfo.StaticAllocaMap.end()) {
Addr.setKind(Address::FrameIndexBase);
Addr.setFI(SI->second);
return true;
}
break;
}
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
case Instruction::Add: {
[FastISel][AArch64] Fold offset into the memory operation. Fold simple offsets into the memory operation: add x0, x0, #8 ldr x0, [x0] --> ldr x0, [x0, #8] Fixes <rdar://problem/17887945>. llvm-svn: 214545
2014-08-02 03:40:16 +08:00
// Adds of constants are common and easy enough.
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
const Value *LHS = U->getOperand(0);
const Value *RHS = U->getOperand(1);
if (isa<ConstantInt>(LHS))
std::swap(LHS, RHS);
if (const ConstantInt *CI = dyn_cast<ConstantInt>(RHS)) {
[FastISel][AArch64] Teach the address computation to also fold sub instructions. Tiny enhancement to the address computation code to also fold sub instructions if the rhs is constant and can be folded into the offset. llvm-svn: 219186
2014-10-07 11:40:03 +08:00
Addr.setOffset(Addr.getOffset() + CI->getSExtValue());
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return computeAddress(LHS, Addr, Ty);
[FastISel][AArch64] Fold offset into the memory operation. Fold simple offsets into the memory operation: add x0, x0, #8 ldr x0, [x0] --> ldr x0, [x0, #8] Fixes <rdar://problem/17887945>. llvm-svn: 214545
2014-08-02 03:40:16 +08:00
}
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
Address Backup = Addr;
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
if (computeAddress(LHS, Addr, Ty) && computeAddress(RHS, Addr, Ty))
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
return true;
Addr = Backup;
[FastISel][AArch64] Fold offset into the memory operation. Fold simple offsets into the memory operation: add x0, x0, #8 ldr x0, [x0] --> ldr x0, [x0, #8] Fixes <rdar://problem/17887945>. llvm-svn: 214545
2014-08-02 03:40:16 +08:00
break;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
}
[FastISel][AArch64] Teach the address computation to also fold sub instructions. Tiny enhancement to the address computation code to also fold sub instructions if the rhs is constant and can be folded into the offset. llvm-svn: 219186
2014-10-07 11:40:03 +08:00
case Instruction::Sub: {
// Subs of constants are common and easy enough.
const Value *LHS = U->getOperand(0);
const Value *RHS = U->getOperand(1);
if (const ConstantInt *CI = dyn_cast<ConstantInt>(RHS)) {
Addr.setOffset(Addr.getOffset() - CI->getSExtValue());
return computeAddress(LHS, Addr, Ty);
}
break;
}
[FastIsel][AArch64] Fix a think-o in address computation. When looking through sign/zero-extensions the code would always assume there is such an extension instruction and use the wrong operand for the address. There was also a minor issue in the handling of 'AND' instructions. I accidentially used a 'cast' instead of a 'dyn_cast'. llvm-svn: 218161
2014-09-20 06:23:46 +08:00
case Instruction::Shl: {
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
if (Addr.getOffsetReg())
break;
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
const auto *CI = dyn_cast<ConstantInt>(U->getOperand(1));
if (!CI)
break;
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
unsigned Val = CI->getZExtValue();
if (Val < 1 || Val > 3)
break;
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
uint64_t NumBytes = 0;
if (Ty && Ty->isSized()) {
uint64_t NumBits = DL.getTypeSizeInBits(Ty);
NumBytes = NumBits / 8;
if (!isPowerOf2_64(NumBits))
NumBytes = 0;
}
if (NumBytes != (1ULL << Val))
break;
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
Addr.setShift(Val);
Addr.setExtendType(AArch64_AM::LSL);
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
const Value *Src = U->getOperand(0);
[AArch64] Fix sext/zext folding in address arithmetic. We were accidentally folding a sign/zero extend in to address arithmetic in a different BB when the extend wasn't available there. Cross BB fast-isel isn't safe, so restrict this to only when the extend is in the same BB as the use. llvm-svn: 236764
2015-05-08 03:21:36 +08:00
if (const auto *I = dyn_cast<Instruction>(Src)) {
if (FuncInfo.MBBMap[I->getParent()] == FuncInfo.MBB) {
// Fold the zext or sext when it won't become a noop.
if (const auto *ZE = dyn_cast<ZExtInst>(I)) {
if (!isIntExtFree(ZE) &&
ZE->getOperand(0)->getType()->isIntegerTy(32)) {
Addr.setExtendType(AArch64_AM::UXTW);
Src = ZE->getOperand(0);
}
} else if (const auto *SE = dyn_cast<SExtInst>(I)) {
if (!isIntExtFree(SE) &&
SE->getOperand(0)->getType()->isIntegerTy(32)) {
Addr.setExtendType(AArch64_AM::SXTW);
Src = SE->getOperand(0);
}
}
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
}
}
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
if (const auto *AI = dyn_cast<BinaryOperator>(Src))
if (AI->getOpcode() == Instruction::And) {
const Value *LHS = AI->getOperand(0);
const Value *RHS = AI->getOperand(1);
[FastISel][AArch64] Fold 'AND' instruction during the address computation. The 'AND' instruction could be used to mask out the lower 32 bits of a register. If this is done inside an address computation we might be able to fold the instruction into the memory instruction itself. and x1, x1, #0xffffffff ---> ldrb x0, [x0, w1, uxtw] ldrb x0, [x0, x1] llvm-svn: 218030
2014-09-18 13:40:41 +08:00
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
if (const auto *C = dyn_cast<ConstantInt>(LHS))
if (C->getValue() == 0xffffffff)
std::swap(LHS, RHS);
if (const auto *C = dyn_cast<ConstantInt>(RHS))
if (C->getValue() == 0xffffffff) {
Addr.setExtendType(AArch64_AM::UXTW);
unsigned Reg = getRegForValue(LHS);
if (!Reg)
return false;
bool RegIsKill = hasTrivialKill(LHS);
Reg = fastEmitInst_extractsubreg(MVT::i32, Reg, RegIsKill,
AArch64::sub_32);
Addr.setOffsetReg(Reg);
return true;
}
}
unsigned Reg = getRegForValue(Src);
if (!Reg)
return false;
Addr.setOffsetReg(Reg);
return true;
[FastIsel][AArch64] Fix a think-o in address computation. When looking through sign/zero-extensions the code would always assume there is such an extension instruction and use the wrong operand for the address. There was also a minor issue in the handling of 'AND' instructions. I accidentially used a 'cast' instead of a 'dyn_cast'. llvm-svn: 218161
2014-09-20 06:23:46 +08:00
}
[FastISel][AArch64] Fold mul into the address computation of memory operations. Teach 'computeAddress' to also fold multiplies into the address computation (when possible). This fixes rdar://problem/18369443. llvm-svn: 217977
2014-09-18 03:19:31 +08:00
case Instruction::Mul: {
if (Addr.getOffsetReg())
break;
if (!isMulPowOf2(U))
break;
const Value *LHS = U->getOperand(0);
const Value *RHS = U->getOperand(1);
// Canonicalize power-of-2 value to the RHS.
if (const auto *C = dyn_cast<ConstantInt>(LHS))
if (C->getValue().isPowerOf2())
std::swap(LHS, RHS);
assert(isa<ConstantInt>(RHS) && "Expected an ConstantInt.");
const auto *C = cast<ConstantInt>(RHS);
unsigned Val = C->getValue().logBase2();
if (Val < 1 || Val > 3)
break;
uint64_t NumBytes = 0;
if (Ty && Ty->isSized()) {
uint64_t NumBits = DL.getTypeSizeInBits(Ty);
NumBytes = NumBits / 8;
if (!isPowerOf2_64(NumBits))
NumBytes = 0;
}
if (NumBytes != (1ULL << Val))
break;
Addr.setShift(Val);
Addr.setExtendType(AArch64_AM::LSL);
[FastIsel][AArch64] Fix a think-o in address computation. When looking through sign/zero-extensions the code would always assume there is such an extension instruction and use the wrong operand for the address. There was also a minor issue in the handling of 'AND' instructions. I accidentially used a 'cast' instead of a 'dyn_cast'. llvm-svn: 218161
2014-09-20 06:23:46 +08:00
const Value *Src = LHS;
[AArch64] Fix sext/zext folding in address arithmetic. We were accidentally folding a sign/zero extend in to address arithmetic in a different BB when the extend wasn't available there. Cross BB fast-isel isn't safe, so restrict this to only when the extend is in the same BB as the use. llvm-svn: 236764
2015-05-08 03:21:36 +08:00
if (const auto *I = dyn_cast<Instruction>(Src)) {
if (FuncInfo.MBBMap[I->getParent()] == FuncInfo.MBB) {
// Fold the zext or sext when it won't become a noop.
if (const auto *ZE = dyn_cast<ZExtInst>(I)) {
if (!isIntExtFree(ZE) &&
ZE->getOperand(0)->getType()->isIntegerTy(32)) {
Addr.setExtendType(AArch64_AM::UXTW);
Src = ZE->getOperand(0);
}
} else if (const auto *SE = dyn_cast<SExtInst>(I)) {
if (!isIntExtFree(SE) &&
SE->getOperand(0)->getType()->isIntegerTy(32)) {
Addr.setExtendType(AArch64_AM::SXTW);
Src = SE->getOperand(0);
}
}
[FastISel][AArch64] Fold mul into the address computation of memory operations. Teach 'computeAddress' to also fold multiplies into the address computation (when possible). This fixes rdar://problem/18369443. llvm-svn: 217977
2014-09-18 03:19:31 +08:00
}
[FastIsel][AArch64] Fix a think-o in address computation. When looking through sign/zero-extensions the code would always assume there is such an extension instruction and use the wrong operand for the address. There was also a minor issue in the handling of 'AND' instructions. I accidentially used a 'cast' instead of a 'dyn_cast'. llvm-svn: 218161
2014-09-20 06:23:46 +08:00
}
[FastISel][AArch64] Fold mul into the address computation of memory operations. Teach 'computeAddress' to also fold multiplies into the address computation (when possible). This fixes rdar://problem/18369443. llvm-svn: 217977
2014-09-18 03:19:31 +08:00
[FastIsel][AArch64] Fix a think-o in address computation. When looking through sign/zero-extensions the code would always assume there is such an extension instruction and use the wrong operand for the address. There was also a minor issue in the handling of 'AND' instructions. I accidentially used a 'cast' instead of a 'dyn_cast'. llvm-svn: 218161
2014-09-20 06:23:46 +08:00
unsigned Reg = getRegForValue(Src);
[FastISel][AArch64] Fold mul into the address computation of memory operations. Teach 'computeAddress' to also fold multiplies into the address computation (when possible). This fixes rdar://problem/18369443. llvm-svn: 217977
2014-09-18 03:19:31 +08:00
if (!Reg)
return false;
Addr.setOffsetReg(Reg);
return true;
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
}
[FastISel][AArch64] Fold 'AND' instruction during the address computation. The 'AND' instruction could be used to mask out the lower 32 bits of a register. If this is done inside an address computation we might be able to fold the instruction into the memory instruction itself. and x1, x1, #0xffffffff ---> ldrb x0, [x0, w1, uxtw] ldrb x0, [x0, x1] llvm-svn: 218030
2014-09-18 13:40:41 +08:00
case Instruction::And: {
if (Addr.getOffsetReg())
break;
[FastISel][AArch64] Fix a missing nullptr check in 'computeAddress'. The load/store value type is currently not available when lowering the memcpy intrinsic. Add the missing nullptr check to support this in 'computeAddress'. Fixes rdar://problem/19178947. llvm-svn: 223818
2014-12-10 03:44:38 +08:00
if (!Ty || DL.getTypeSizeInBits(Ty) != 8)
[FastISel][AArch64] Fold 'AND' instruction during the address computation. The 'AND' instruction could be used to mask out the lower 32 bits of a register. If this is done inside an address computation we might be able to fold the instruction into the memory instruction itself. and x1, x1, #0xffffffff ---> ldrb x0, [x0, w1, uxtw] ldrb x0, [x0, x1] llvm-svn: 218030
2014-09-18 13:40:41 +08:00
break;
const Value *LHS = U->getOperand(0);
const Value *RHS = U->getOperand(1);
if (const auto *C = dyn_cast<ConstantInt>(LHS))
if (C->getValue() == 0xffffffff)
std::swap(LHS, RHS);
[FastIsel][AArch64] Fix a think-o in address computation. When looking through sign/zero-extensions the code would always assume there is such an extension instruction and use the wrong operand for the address. There was also a minor issue in the handling of 'AND' instructions. I accidentially used a 'cast' instead of a 'dyn_cast'. llvm-svn: 218161
2014-09-20 06:23:46 +08:00
if (const auto *C = dyn_cast<ConstantInt>(RHS))
[FastISel][AArch64] Fold 'AND' instruction during the address computation. The 'AND' instruction could be used to mask out the lower 32 bits of a register. If this is done inside an address computation we might be able to fold the instruction into the memory instruction itself. and x1, x1, #0xffffffff ---> ldrb x0, [x0, w1, uxtw] ldrb x0, [x0, x1] llvm-svn: 218030
2014-09-18 13:40:41 +08:00
if (C->getValue() == 0xffffffff) {
Addr.setShift(0);
Addr.setExtendType(AArch64_AM::LSL);
Addr.setExtendType(AArch64_AM::UXTW);
unsigned Reg = getRegForValue(LHS);
if (!Reg)
return false;
bool RegIsKill = hasTrivialKill(LHS);
Reg = fastEmitInst_extractsubreg(MVT::i32, Reg, RegIsKill,
AArch64::sub_32);
Addr.setOffsetReg(Reg);
return true;
}
break;
}
[FastISel][AArch64] Teach the address computation code to also fold sign-/zero-extends. The code already folds sign-/zero-extends, but only if they are arguments to mul and shift instructions. This extends the code to also fold them when they are direct inputs. llvm-svn: 219187
2014-10-07 11:40:06 +08:00
case Instruction::SExt:
case Instruction::ZExt: {
if (!Addr.getReg() || Addr.getOffsetReg())
break;
const Value *Src = nullptr;
// Fold the zext or sext when it won't become a noop.
if (const auto *ZE = dyn_cast<ZExtInst>(U)) {
if (!isIntExtFree(ZE) && ZE->getOperand(0)->getType()->isIntegerTy(32)) {
Addr.setExtendType(AArch64_AM::UXTW);
Src = ZE->getOperand(0);
}
} else if (const auto *SE = dyn_cast<SExtInst>(U)) {
if (!isIntExtFree(SE) && SE->getOperand(0)->getType()->isIntegerTy(32)) {
Addr.setExtendType(AArch64_AM::SXTW);
Src = SE->getOperand(0);
}
}
if (!Src)
break;
Addr.setShift(0);
unsigned Reg = getRegForValue(Src);
if (!Reg)
return false;
Addr.setOffsetReg(Reg);
return true;
}
[FastISel][AArch64] Fold mul into the address computation of memory operations. Teach 'computeAddress' to also fold multiplies into the address computation (when possible). This fixes rdar://problem/18369443. llvm-svn: 217977
2014-09-18 03:19:31 +08:00
} // end switch
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
[FastISel][AArch64] Fix load/store with frame indices. At higher optimization levels the LLVM IR may contain more complex patterns for loads/stores from/to frame indices. The 'computeAddress' function wasn't able to handle this and triggered an assertion. This fix extends the possible addressing modes for frame indices. This fixes rdar://problem/18783298. llvm-svn: 220700
2014-10-28 02:21:58 +08:00
if (Addr.isRegBase() && !Addr.getReg()) {
unsigned Reg = getRegForValue(Obj);
if (!Reg)
return false;
Addr.setReg(Reg);
return true;
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
}
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
[FastISel][AArch64] Fix load/store with frame indices. At higher optimization levels the LLVM IR may contain more complex patterns for loads/stores from/to frame indices. The 'computeAddress' function wasn't able to handle this and triggered an assertion. This fix extends the possible addressing modes for frame indices. This fixes rdar://problem/18783298. llvm-svn: 220700
2014-10-28 02:21:58 +08:00
if (!Addr.getOffsetReg()) {
unsigned Reg = getRegForValue(Obj);
if (!Reg)
return false;
Addr.setOffsetReg(Reg);
return true;
}
return false;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool AArch64FastISel::computeCallAddress(const Value *V, Address &Addr) {
[FastISel][AArch64] Add MachO large code model support for function calls. Currently the large code model for MachO uses the GOT to make function calls. Emit the required adrp and ldr instructions to load the address from the GOT. Related to <rdar://problem/17733076>. llvm-svn: 214381
2014-07-31 12:10:40 +08:00
const User *U = nullptr;
unsigned Opcode = Instruction::UserOp1;
bool InMBB = true;
if (const auto *I = dyn_cast<Instruction>(V)) {
Opcode = I->getOpcode();
U = I;
InMBB = I->getParent() == FuncInfo.MBB->getBasicBlock();
} else if (const auto *C = dyn_cast<ConstantExpr>(V)) {
Opcode = C->getOpcode();
U = C;
}
switch (Opcode) {
default: break;
case Instruction::BitCast:
// Look past bitcasts if its operand is in the same BB.
if (InMBB)
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return computeCallAddress(U->getOperand(0), Addr);
[FastISel][AArch64] Add MachO large code model support for function calls. Currently the large code model for MachO uses the GOT to make function calls. Emit the required adrp and ldr instructions to load the address from the GOT. Related to <rdar://problem/17733076>. llvm-svn: 214381
2014-07-31 12:10:40 +08:00
break;
case Instruction::IntToPtr:
// Look past no-op inttoptrs if its operand is in the same BB.
if (InMBB &&
Make TargetLowering::getPointerTy() taking DataLayout as an argument Summary: This change is part of a series of commits dedicated to have a single DataLayout during compilation by using always the one owned by the module. Reviewers: echristo Subscribers: jholewinski, ted, yaron.keren, rafael, llvm-commits Differential Revision: http://reviews.llvm.org/D11028 From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 241775
2015-07-09 10:09:04 +08:00
TLI.getValueType(DL, U->getOperand(0)->getType()) ==
TLI.getPointerTy(DL))
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return computeCallAddress(U->getOperand(0), Addr);
[FastISel][AArch64] Add MachO large code model support for function calls. Currently the large code model for MachO uses the GOT to make function calls. Emit the required adrp and ldr instructions to load the address from the GOT. Related to <rdar://problem/17733076>. llvm-svn: 214381
2014-07-31 12:10:40 +08:00
break;
case Instruction::PtrToInt:
// Look past no-op ptrtoints if its operand is in the same BB.
Make TargetLowering::getPointerTy() taking DataLayout as an argument Summary: This change is part of a series of commits dedicated to have a single DataLayout during compilation by using always the one owned by the module. Reviewers: echristo Subscribers: jholewinski, ted, yaron.keren, rafael, llvm-commits Differential Revision: http://reviews.llvm.org/D11028 From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 241775
2015-07-09 10:09:04 +08:00
if (InMBB && TLI.getValueType(DL, U->getType()) == TLI.getPointerTy(DL))
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return computeCallAddress(U->getOperand(0), Addr);
[FastISel][AArch64] Add MachO large code model support for function calls. Currently the large code model for MachO uses the GOT to make function calls. Emit the required adrp and ldr instructions to load the address from the GOT. Related to <rdar://problem/17733076>. llvm-svn: 214381
2014-07-31 12:10:40 +08:00
break;
}
if (const GlobalValue *GV = dyn_cast<GlobalValue>(V)) {
Addr.setGlobalValue(GV);
return true;
}
// If all else fails, try to materialize the value in a register.
if (!Addr.getGlobalValue()) {
Addr.setReg(getRegForValue(V));
return Addr.getReg() != 0;
}
return false;
}
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
bool AArch64FastISel::isTypeLegal(Type *Ty, MVT &VT) {
Make TargetLowering::getPointerTy() taking DataLayout as an argument Summary: This change is part of a series of commits dedicated to have a single DataLayout during compilation by using always the one owned by the module. Reviewers: echristo Subscribers: jholewinski, ted, yaron.keren, rafael, llvm-commits Differential Revision: http://reviews.llvm.org/D11028 From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 241775
2015-07-09 10:09:04 +08:00
EVT evt = TLI.getValueType(DL, Ty, true);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Only handle simple types.
if (evt == MVT::Other || !evt.isSimple())
return false;
VT = evt.getSimpleVT();
[ARM64][fast-isel] Fast-isel doesn't know how to handle f128. llvm-svn: 207659
2014-04-30 23:29:57 +08:00
// This is a legal type, but it's not something we handle in fast-isel.
if (VT == MVT::f128)
return false;
// Handle all other legal types, i.e. a register that will directly hold this
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// value.
return TLI.isTypeLegal(VT);
}
[FastISel][AArch64] Use a new helper function to determine if a value type is supported. NFCI. FastISel for AArch64 supports more value types than are actually legal. Use a dedicated helper function to reflect this. It is very similar to the isLoadStoreTypeLegal function, with the exception that vector types are not supported yet. llvm-svn: 216984
2014-09-03 06:33:53 +08:00
/// \brief Determine if the value type is supported by FastISel.
///
/// FastISel for AArch64 can handle more value types than are legal. This adds
/// simple value type such as i1, i8, and i16.
[FastISel][AArch64] Refactor code to use isTypeSupported. NFC. Gets rid of isLoadStoreTypeLegal and replace it with isTypeSupported. llvm-svn: 217826
2014-09-16 05:27:54 +08:00
bool AArch64FastISel::isTypeSupported(Type *Ty, MVT &VT, bool IsVectorAllowed) {
if (Ty->isVectorTy() && !IsVectorAllowed)
[FastISel][AArch64] Use a new helper function to determine if a value type is supported. NFCI. FastISel for AArch64 supports more value types than are actually legal. Use a dedicated helper function to reflect this. It is very similar to the isLoadStoreTypeLegal function, with the exception that vector types are not supported yet. llvm-svn: 216984
2014-09-03 06:33:53 +08:00
return false;
if (isTypeLegal(Ty, VT))
return true;
// If this is a type than can be sign or zero-extended to a basic operation
// go ahead and accept it now.
if (VT == MVT::i1 || VT == MVT::i8 || VT == MVT::i16)
return true;
return false;
}
[FastISel][AArch64] Don't fold instructions that are not in the same basic block. This fix checks first if the instruction to be folded (e.g. sign-/zero-extend, or shift) is in the same machine basic block as the instruction we are folding into. Not doing so can result in incorrect code, because the value might not be live-out of the basic block, where the value is defined. This fixes rdar://problem/18169495. llvm-svn: 216700
2014-08-29 08:19:21 +08:00
bool AArch64FastISel::isValueAvailable(const Value *V) const {
if (!isa<Instruction>(V))
return true;
const auto *I = cast<Instruction>(V);
Simplify some boolean conditional return statements in AArch64. http://reviews.llvm.org/D9979 Patch by Richard Thomson (and some conflict resolution by me). llvm-svn: 262266
2016-03-01 06:50:49 +08:00
return FuncInfo.MBBMap[I->getParent()] == FuncInfo.MBB;
[FastISel][AArch64] Don't fold instructions that are not in the same basic block. This fix checks first if the instruction to be folded (e.g. sign-/zero-extend, or shift) is in the same machine basic block as the instruction we are folding into. Not doing so can result in incorrect code, because the value might not be live-out of the basic block, where the value is defined. This fixes rdar://problem/18169495. llvm-svn: 216700
2014-08-29 08:19:21 +08:00
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool AArch64FastISel::simplifyAddress(Address &Addr, MVT VT) {
[FastISel][AArch64] Factor out scale factor calculation. NFC. Factor out the code that determines the implicit scale factor of memory operations for a given value type. llvm-svn: 218652
2014-09-30 08:49:54 +08:00
unsigned ScaleFactor = getImplicitScaleFactor(VT);
if (!ScaleFactor)
return false;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
bool ImmediateOffsetNeedsLowering = false;
bool RegisterOffsetNeedsLowering = false;
int64_t Offset = Addr.getOffset();
if (((Offset < 0) || (Offset & (ScaleFactor - 1))) && !isInt<9>(Offset))
ImmediateOffsetNeedsLowering = true;
else if (Offset > 0 && !(Offset & (ScaleFactor - 1)) &&
!isUInt<12>(Offset / ScaleFactor))
ImmediateOffsetNeedsLowering = true;
// Cannot encode an offset register and an immediate offset in the same
// instruction. Fold the immediate offset into the load/store instruction and
Fix some comment typos. llvm-svn: 244402
2015-08-09 02:27:36 +08:00
// emit an additional add to take care of the offset register.
[FastISel][AArch64] Fix load/store with frame indices. At higher optimization levels the LLVM IR may contain more complex patterns for loads/stores from/to frame indices. The 'computeAddress' function wasn't able to handle this and triggered an assertion. This fix extends the possible addressing modes for frame indices. This fixes rdar://problem/18783298. llvm-svn: 220700
2014-10-28 02:21:58 +08:00
if (!ImmediateOffsetNeedsLowering && Addr.getOffset() && Addr.getOffsetReg())
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
RegisterOffsetNeedsLowering = true;
[FastISel][AArch64] Fix simplify address when the address comes from a shift. When the address comes directly from a shift instruction then the address computation cannot be folded into the memory instruction, because the zero register is not available as a base register. Simplify addess needs to emit the shift instruction and use the result as base register. llvm-svn: 216621
2014-08-28 05:38:33 +08:00
// Cannot encode zero register as base.
if (Addr.isRegBase() && Addr.getOffsetReg() && !Addr.getReg())
RegisterOffsetNeedsLowering = true;
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
// If this is a stack pointer and the offset needs to be simplified then put
AArch64: teach FastISel how to handle offset FrameIndices Previously we were abandonning the attempt, leading to some combination of extra work (when selection of a load/store fails completely) and inferior code (when this leads to a real memcpy call instead of inlining). rdar://problem/17187463 llvm-svn: 210520
2014-06-10 17:52:44 +08:00
// the alloca address into a register, set the base type back to register and
// continue. This should almost never happen.
[FastISel][AArch64] Fix load/store with frame indices. At higher optimization levels the LLVM IR may contain more complex patterns for loads/stores from/to frame indices. The 'computeAddress' function wasn't able to handle this and triggered an assertion. This fix extends the possible addressing modes for frame indices. This fixes rdar://problem/18783298. llvm-svn: 220700
2014-10-28 02:21:58 +08:00
if ((ImmediateOffsetNeedsLowering || Addr.getOffsetReg()) && Addr.isFIBase())
{
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
unsigned ResultReg = createResultReg(&AArch64::GPR64spRegClass);
AArch64: teach FastISel how to handle offset FrameIndices Previously we were abandonning the attempt, leading to some combination of extra work (when selection of a load/store fails completely) and inferior code (when this leads to a real memcpy call instead of inlining). rdar://problem/17187463 llvm-svn: 210520
2014-06-10 17:52:44 +08:00
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(AArch64::ADDXri),
ResultReg)
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
.addFrameIndex(Addr.getFI())
.addImm(0)
.addImm(0);
AArch64: teach FastISel how to handle offset FrameIndices Previously we were abandonning the attempt, leading to some combination of extra work (when selection of a load/store fails completely) and inferior code (when this leads to a real memcpy call instead of inlining). rdar://problem/17187463 llvm-svn: 210520
2014-06-10 17:52:44 +08:00
Addr.setKind(Address::RegBase);
Addr.setReg(ResultReg);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
}
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
if (RegisterOffsetNeedsLowering) {
unsigned ResultReg = 0;
[FastISel][AArch64] Fix address simplification. When a shift with extension or an add with shift and extension cannot be folded into the memory operation, then the address calculation has to be materialized separately. While doing so the code forgot to consider a possible sign-/zero- extension. This fix folds now also the sign-/zero-extension into the add or shift instruction which is used to materialize the address. This fixes rdar://problem/18141718. llvm-svn: 216511
2014-08-27 08:58:30 +08:00
if (Addr.getReg()) {
if (Addr.getExtendType() == AArch64_AM::SXTW ||
Addr.getExtendType() == AArch64_AM::UXTW )
ResultReg = emitAddSub_rx(/*UseAdd=*/true, MVT::i64, Addr.getReg(),
/*TODO:IsKill=*/false, Addr.getOffsetReg(),
/*TODO:IsKill=*/false, Addr.getExtendType(),
Addr.getShift());
else
ResultReg = emitAddSub_rs(/*UseAdd=*/true, MVT::i64, Addr.getReg(),
/*TODO:IsKill=*/false, Addr.getOffsetReg(),
/*TODO:IsKill=*/false, AArch64_AM::LSL,
Addr.getShift());
} else {
if (Addr.getExtendType() == AArch64_AM::UXTW)
ResultReg = emitLSL_ri(MVT::i64, MVT::i32, Addr.getOffsetReg(),
/*Op0IsKill=*/false, Addr.getShift(),
/*IsZExt=*/true);
else if (Addr.getExtendType() == AArch64_AM::SXTW)
ResultReg = emitLSL_ri(MVT::i64, MVT::i32, Addr.getOffsetReg(),
/*Op0IsKill=*/false, Addr.getShift(),
/*IsZExt=*/false);
else
ResultReg = emitLSL_ri(MVT::i64, MVT::i64, Addr.getOffsetReg(),
/*Op0IsKill=*/false, Addr.getShift());
}
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
if (!ResultReg)
return false;
Addr.setReg(ResultReg);
Addr.setOffsetReg(0);
Addr.setShift(0);
[FastISel][AArch64] Fix address simplification. When a shift with extension or an add with shift and extension cannot be folded into the memory operation, then the address calculation has to be materialized separately. While doing so the code forgot to consider a possible sign-/zero- extension. This fix folds now also the sign-/zero-extension into the add or shift instruction which is used to materialize the address. This fixes rdar://problem/18141718. llvm-svn: 216511
2014-08-27 08:58:30 +08:00
Addr.setExtendType(AArch64_AM::InvalidShiftExtend);
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
}
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Since the offset is too large for the load/store instruction get the
// reg+offset into a register.
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
if (ImmediateOffsetNeedsLowering) {
[FastISel][AArch64] Followup commit for 218031 to handle negative offsets too. llvm-svn: 218032
2014-09-18 15:04:49 +08:00
unsigned ResultReg;
[FastISel][AArch64] Factor out add with immediate emission into a helper function. NFC. Simplify add with immediate emission by factoring it out into a helper function. llvm-svn: 219830
2014-10-16 02:58:02 +08:00
if (Addr.getReg())
[FastISel][AArch64] Try to fold the offset into the add instruction when simplifying a memory address. Small optimization in 'simplifyAddress'. When the offset cannot be encoded in the load/store instruction, then we need to materialize the address manually. The add instruction can encode a wider range of immediates than the load/store instructions. This change tries to fold the offset into the add instruction first before materializing the offset in a register. llvm-svn: 218031
2014-09-18 13:40:47 +08:00
// Try to fold the immediate into the add instruction.
[FastISel][AArch64] Factor out add with immediate emission into a helper function. NFC. Simplify add with immediate emission by factoring it out into a helper function. llvm-svn: 219830
2014-10-16 02:58:02 +08:00
ResultReg = emitAdd_ri_(MVT::i64, Addr.getReg(), /*IsKill=*/false, Offset);
else
[FastISel][tblgen] Rename tblgen generated FastISel functions. NFC. This is the final round of renaming. This changes tblgen to emit lower-case function names for FastEmitInst_* and FastEmit_*, and updates all its uses in the source code. Reviewed by Eric llvm-svn: 217075
2014-09-04 04:56:59 +08:00
ResultReg = fastEmit_i(MVT::i64, MVT::i64, ISD::Constant, Offset);
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
if (!ResultReg)
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
Addr.setReg(ResultReg);
Addr.setOffset(0);
}
return true;
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
void AArch64FastISel::addLoadStoreOperands(Address &Addr,
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
const MachineInstrBuilder &MIB,
[CodeGen] Take a MachineMemOperand::Flags in MachineFunction::getMachineMemOperand. Summary: Previously we took an unsigned. Hooray for type-safety. Reviewers: chandlerc Subscribers: dsanders, llvm-commits Differential Revision: http://reviews.llvm.org/D22282 llvm-svn: 275591
2016-07-16 02:26:59 +08:00
MachineMemOperand::Flags Flags,
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
unsigned ScaleFactor,
MachineMemOperand *MMO) {
int64_t Offset = Addr.getOffset() / ScaleFactor;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Frame base works a bit differently. Handle it separately.
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
if (Addr.isFIBase()) {
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
int FI = Addr.getFI();
// FIXME: We shouldn't be using getObjectSize/getObjectAlignment. The size
// and alignment should be based on the VT.
[FastISel][AArch64] Attach MachineMemOperands to load and store instructions. llvm-svn: 215231
2014-08-09 01:24:10 +08:00
MMO = FuncInfo.MF->getMachineMemOperand(
PseudoSourceValue: Replace global manager with a manager in a machine function. This commit removes the global manager variable which is responsible for storing and allocating pseudo source values and instead it introduces a new manager class named 'PseudoSourceValueManager'. Machine functions now own an instance of the pseudo source value manager class. This commit also modifies the 'get...' methods in the 'MachinePointerInfo' class to construct pseudo source values using the instance of the pseudo source value manager object from the machine function. This commit updates calls to the 'get...' methods from the 'MachinePointerInfo' class in a lot of different files because those calls now need to pass in a reference to a machine function to those methods. This change will make it easier to serialize pseudo source values as it will enable me to transform the mips specific MipsCallEntry PseudoSourceValue subclass into two target independent subclasses. Reviewers: Akira Hatanaka llvm-svn: 244693
2015-08-12 07:09:45 +08:00
MachinePointerInfo::getFixedStack(*FuncInfo.MF, FI, Offset), Flags,
MFI.getObjectSize(FI), MFI.getObjectAlignment(FI));
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Now add the rest of the operands.
[FastISel][AArch64] Attach MachineMemOperands to load and store instructions. llvm-svn: 215231
2014-08-09 01:24:10 +08:00
MIB.addFrameIndex(FI).addImm(Offset);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
} else {
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
assert(Addr.isRegBase() && "Unexpected address kind.");
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
const MCInstrDesc &II = MIB->getDesc();
unsigned Idx = (Flags & MachineMemOperand::MOStore) ? 1 : 0;
Addr.setReg(
constrainOperandRegClass(II, Addr.getReg(), II.getNumDefs()+Idx));
Addr.setOffsetReg(
constrainOperandRegClass(II, Addr.getOffsetReg(), II.getNumDefs()+Idx+1));
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
if (Addr.getOffsetReg()) {
assert(Addr.getOffset() == 0 && "Unexpected offset");
bool IsSigned = Addr.getExtendType() == AArch64_AM::SXTW ||
Addr.getExtendType() == AArch64_AM::SXTX;
MIB.addReg(Addr.getReg());
MIB.addReg(Addr.getOffsetReg());
MIB.addImm(IsSigned);
MIB.addImm(Addr.getShift() != 0);
[FastISel][AArch64] Fix load/store with frame indices. At higher optimization levels the LLVM IR may contain more complex patterns for loads/stores from/to frame indices. The 'computeAddress' function wasn't able to handle this and triggered an assertion. This fix extends the possible addressing modes for frame indices. This fixes rdar://problem/18783298. llvm-svn: 220700
2014-10-28 02:21:58 +08:00
} else
MIB.addReg(Addr.getReg()).addImm(Offset);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
}
[FastISel][AArch64] Attach MachineMemOperands to load and store instructions. llvm-svn: 215231
2014-08-09 01:24:10 +08:00
if (MMO)
MIB.addMemOperand(MMO);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
}
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
unsigned AArch64FastISel::emitAddSub(bool UseAdd, MVT RetVT, const Value *LHS,
const Value *RHS, bool SetFlags,
bool WantResult, bool IsZExt) {
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
AArch64_AM::ShiftExtendType ExtendType = AArch64_AM::InvalidShiftExtend;
[FastISel][AArch64] Don't fold the sign-/zero-extend from i1 into the compare. This fixes a bug I introduced in a previous commit (r216033). Sign-/Zero- extension from i1 cannot be folded into the ADDS/SUBS instructions. Instead both operands have to be sign-/zero-extended with separate instructions. Related to <rdar://problem/17913111>. llvm-svn: 216073
2014-08-21 00:34:15 +08:00
bool NeedExtend = false;
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
switch (RetVT.SimpleTy) {
[FastISel][AArch64] Don't fold the sign-/zero-extend from i1 into the compare. This fixes a bug I introduced in a previous commit (r216033). Sign-/Zero- extension from i1 cannot be folded into the ADDS/SUBS instructions. Instead both operands have to be sign-/zero-extended with separate instructions. Related to <rdar://problem/17913111>. llvm-svn: 216073
2014-08-21 00:34:15 +08:00
default:
return 0;
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
case MVT::i1:
[FastISel][AArch64] Don't fold the sign-/zero-extend from i1 into the compare. This fixes a bug I introduced in a previous commit (r216033). Sign-/Zero- extension from i1 cannot be folded into the ADDS/SUBS instructions. Instead both operands have to be sign-/zero-extended with separate instructions. Related to <rdar://problem/17913111>. llvm-svn: 216073
2014-08-21 00:34:15 +08:00
NeedExtend = true;
break;
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
case MVT::i8:
[FastISel][AArch64] Don't fold the sign-/zero-extend from i1 into the compare. This fixes a bug I introduced in a previous commit (r216033). Sign-/Zero- extension from i1 cannot be folded into the ADDS/SUBS instructions. Instead both operands have to be sign-/zero-extended with separate instructions. Related to <rdar://problem/17913111>. llvm-svn: 216073
2014-08-21 00:34:15 +08:00
NeedExtend = true;
ExtendType = IsZExt ? AArch64_AM::UXTB : AArch64_AM::SXTB;
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
break;
case MVT::i16:
[FastISel][AArch64] Don't fold the sign-/zero-extend from i1 into the compare. This fixes a bug I introduced in a previous commit (r216033). Sign-/Zero- extension from i1 cannot be folded into the ADDS/SUBS instructions. Instead both operands have to be sign-/zero-extended with separate instructions. Related to <rdar://problem/17913111>. llvm-svn: 216073
2014-08-21 00:34:15 +08:00
NeedExtend = true;
ExtendType = IsZExt ? AArch64_AM::UXTH : AArch64_AM::SXTH;
break;
case MVT::i32: // fall-through
case MVT::i64:
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
break;
}
[FastISel][AArch64] Don't fold the sign-/zero-extend from i1 into the compare. This fixes a bug I introduced in a previous commit (r216033). Sign-/Zero- extension from i1 cannot be folded into the ADDS/SUBS instructions. Instead both operands have to be sign-/zero-extended with separate instructions. Related to <rdar://problem/17913111>. llvm-svn: 216073
2014-08-21 00:34:15 +08:00
MVT SrcVT = RetVT;
RetVT.SimpleTy = std::max(RetVT.SimpleTy, MVT::i32);
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
// Canonicalize immediates to the RHS first.
[FastISel][AArch64] Emit immediate version of icmp (subs) for null pointer check. This is a minor change to use the immediate version when the operand is a null value. This should get rid of an unnecessary 'mov' instruction in debug builds and align the code more with the one generated by SelectionDAG. This fixes rdar://problem/18785125. llvm-svn: 220713
2014-10-28 03:58:36 +08:00
if (UseAdd && isa<Constant>(LHS) && !isa<Constant>(RHS))
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
std::swap(LHS, RHS);
[FastISel][AArch64] Fold mul into add/sub and logical operations. Try to fold the multiply into the add/sub or logical operations (when possible). This is related to rdar://problem/18369687. llvm-svn: 217978
2014-09-18 03:51:38 +08:00
// Canonicalize mul by power of 2 to the RHS.
if (UseAdd && LHS->hasOneUse() && isValueAvailable(LHS))
if (isMulPowOf2(LHS))
std::swap(LHS, RHS);
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
// Canonicalize shift immediate to the RHS.
[FastISel][AArch64] Fold mul into add/sub and logical operations. Try to fold the multiply into the add/sub or logical operations (when possible). This is related to rdar://problem/18369687. llvm-svn: 217978
2014-09-18 03:51:38 +08:00
if (UseAdd && LHS->hasOneUse() && isValueAvailable(LHS))
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
if (const auto *SI = dyn_cast<BinaryOperator>(LHS))
if (isa<ConstantInt>(SI->getOperand(1)))
if (SI->getOpcode() == Instruction::Shl ||
SI->getOpcode() == Instruction::LShr ||
SI->getOpcode() == Instruction::AShr )
std::swap(LHS, RHS);
unsigned LHSReg = getRegForValue(LHS);
if (!LHSReg)
return 0;
bool LHSIsKill = hasTrivialKill(LHS);
[FastISel][AArch64] Don't fold the sign-/zero-extend from i1 into the compare. This fixes a bug I introduced in a previous commit (r216033). Sign-/Zero- extension from i1 cannot be folded into the ADDS/SUBS instructions. Instead both operands have to be sign-/zero-extended with separate instructions. Related to <rdar://problem/17913111>. llvm-svn: 216073
2014-08-21 00:34:15 +08:00
if (NeedExtend)
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
LHSReg = emitIntExt(SrcVT, LHSReg, RetVT, IsZExt);
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
unsigned ResultReg = 0;
if (const auto *C = dyn_cast<ConstantInt>(RHS)) {
uint64_t Imm = IsZExt ? C->getZExtValue() : C->getSExtValue();
if (C->isNegative())
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
ResultReg = emitAddSub_ri(!UseAdd, RetVT, LHSReg, LHSIsKill, -Imm,
SetFlags, WantResult);
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
else
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
ResultReg = emitAddSub_ri(UseAdd, RetVT, LHSReg, LHSIsKill, Imm, SetFlags,
WantResult);
[FastISel][AArch64] Emit immediate version of icmp (subs) for null pointer check. This is a minor change to use the immediate version when the operand is a null value. This should get rid of an unnecessary 'mov' instruction in debug builds and align the code more with the one generated by SelectionDAG. This fixes rdar://problem/18785125. llvm-svn: 220713
2014-10-28 03:58:36 +08:00
} else if (const auto *C = dyn_cast<Constant>(RHS))
if (C->isNullValue())
ResultReg = emitAddSub_ri(UseAdd, RetVT, LHSReg, LHSIsKill, 0, SetFlags,
WantResult);
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
if (ResultReg)
return ResultReg;
[FastISel][AArch64] Don't fold the sign-/zero-extend from i1 into the compare. This fixes a bug I introduced in a previous commit (r216033). Sign-/Zero- extension from i1 cannot be folded into the ADDS/SUBS instructions. Instead both operands have to be sign-/zero-extended with separate instructions. Related to <rdar://problem/17913111>. llvm-svn: 216073
2014-08-21 00:34:15 +08:00
// Only extend the RHS within the instruction if there is a valid extend type.
[FastISel][AArch64] Fold mul into add/sub and logical operations. Try to fold the multiply into the add/sub or logical operations (when possible). This is related to rdar://problem/18369687. llvm-svn: 217978
2014-09-18 03:51:38 +08:00
if (ExtendType != AArch64_AM::InvalidShiftExtend && RHS->hasOneUse() &&
isValueAvailable(RHS)) {
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
if (const auto *SI = dyn_cast<BinaryOperator>(RHS))
if (const auto *C = dyn_cast<ConstantInt>(SI->getOperand(1)))
if ((SI->getOpcode() == Instruction::Shl) && (C->getZExtValue() < 4)) {
unsigned RHSReg = getRegForValue(SI->getOperand(0));
if (!RHSReg)
return 0;
bool RHSIsKill = hasTrivialKill(SI->getOperand(0));
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
return emitAddSub_rx(UseAdd, RetVT, LHSReg, LHSIsKill, RHSReg,
RHSIsKill, ExtendType, C->getZExtValue(),
SetFlags, WantResult);
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
}
unsigned RHSReg = getRegForValue(RHS);
if (!RHSReg)
return 0;
bool RHSIsKill = hasTrivialKill(RHS);
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
return emitAddSub_rx(UseAdd, RetVT, LHSReg, LHSIsKill, RHSReg, RHSIsKill,
ExtendType, 0, SetFlags, WantResult);
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
}
[FastISel][AArch64] Fold mul into add/sub and logical operations. Try to fold the multiply into the add/sub or logical operations (when possible). This is related to rdar://problem/18369687. llvm-svn: 217978
2014-09-18 03:51:38 +08:00
// Check if the mul can be folded into the instruction.
[AArch64][FastISel] Don't fold shifts with UB. We are already falling back to SelectionDAG when encountering an shift with UB. This adds the same checks for shifts with UB that get folded into arithmetic or logical operations. This fixes rdar://problem/22345295. llvm-svn: 245499
2015-08-20 04:52:55 +08:00
if (RHS->hasOneUse() && isValueAvailable(RHS)) {
[FastISel][AArch64] Fold mul into add/sub and logical operations. Try to fold the multiply into the add/sub or logical operations (when possible). This is related to rdar://problem/18369687. llvm-svn: 217978
2014-09-18 03:51:38 +08:00
if (isMulPowOf2(RHS)) {
const Value *MulLHS = cast<MulOperator>(RHS)->getOperand(0);
const Value *MulRHS = cast<MulOperator>(RHS)->getOperand(1);
if (const auto *C = dyn_cast<ConstantInt>(MulLHS))
if (C->getValue().isPowerOf2())
std::swap(MulLHS, MulRHS);
assert(isa<ConstantInt>(MulRHS) && "Expected a ConstantInt.");
uint64_t ShiftVal = cast<ConstantInt>(MulRHS)->getValue().logBase2();
unsigned RHSReg = getRegForValue(MulLHS);
if (!RHSReg)
return 0;
bool RHSIsKill = hasTrivialKill(MulLHS);
[AArch64][FastISel] Don't fold shifts with UB. We are already falling back to SelectionDAG when encountering an shift with UB. This adds the same checks for shifts with UB that get folded into arithmetic or logical operations. This fixes rdar://problem/22345295. llvm-svn: 245499
2015-08-20 04:52:55 +08:00
ResultReg = emitAddSub_rs(UseAdd, RetVT, LHSReg, LHSIsKill, RHSReg,
RHSIsKill, AArch64_AM::LSL, ShiftVal, SetFlags,
WantResult);
if (ResultReg)
return ResultReg;
[FastISel][AArch64] Fold mul into add/sub and logical operations. Try to fold the multiply into the add/sub or logical operations (when possible). This is related to rdar://problem/18369687. llvm-svn: 217978
2014-09-18 03:51:38 +08:00
}
[AArch64][FastISel] Don't fold shifts with UB. We are already falling back to SelectionDAG when encountering an shift with UB. This adds the same checks for shifts with UB that get folded into arithmetic or logical operations. This fixes rdar://problem/22345295. llvm-svn: 245499
2015-08-20 04:52:55 +08:00
}
[FastISel][AArch64] Fold mul into add/sub and logical operations. Try to fold the multiply into the add/sub or logical operations (when possible). This is related to rdar://problem/18369687. llvm-svn: 217978
2014-09-18 03:51:38 +08:00
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
// Check if the shift can be folded into the instruction.
[AArch64][FastISel] Don't fold shifts with UB. We are already falling back to SelectionDAG when encountering an shift with UB. This adds the same checks for shifts with UB that get folded into arithmetic or logical operations. This fixes rdar://problem/22345295. llvm-svn: 245499
2015-08-20 04:52:55 +08:00
if (RHS->hasOneUse() && isValueAvailable(RHS)) {
[FastISel][AArch64] Don't fold instructions that are not in the same basic block. This fix checks first if the instruction to be folded (e.g. sign-/zero-extend, or shift) is in the same machine basic block as the instruction we are folding into. Not doing so can result in incorrect code, because the value might not be live-out of the basic block, where the value is defined. This fixes rdar://problem/18169495. llvm-svn: 216700
2014-08-29 08:19:21 +08:00
if (const auto *SI = dyn_cast<BinaryOperator>(RHS)) {
if (const auto *C = dyn_cast<ConstantInt>(SI->getOperand(1))) {
AArch64_AM::ShiftExtendType ShiftType = AArch64_AM::InvalidShiftExtend;
switch (SI->getOpcode()) {
default: break;
case Instruction::Shl: ShiftType = AArch64_AM::LSL; break;
case Instruction::LShr: ShiftType = AArch64_AM::LSR; break;
case Instruction::AShr: ShiftType = AArch64_AM::ASR; break;
}
uint64_t ShiftVal = C->getZExtValue();
if (ShiftType != AArch64_AM::InvalidShiftExtend) {
unsigned RHSReg = getRegForValue(SI->getOperand(0));
if (!RHSReg)
return 0;
bool RHSIsKill = hasTrivialKill(SI->getOperand(0));
[AArch64][FastISel] Don't fold shifts with UB. We are already falling back to SelectionDAG when encountering an shift with UB. This adds the same checks for shifts with UB that get folded into arithmetic or logical operations. This fixes rdar://problem/22345295. llvm-svn: 245499
2015-08-20 04:52:55 +08:00
ResultReg = emitAddSub_rs(UseAdd, RetVT, LHSReg, LHSIsKill, RHSReg,
RHSIsKill, ShiftType, ShiftVal, SetFlags,
WantResult);
if (ResultReg)
return ResultReg;
[FastISel][AArch64] Don't fold instructions that are not in the same basic block. This fix checks first if the instruction to be folded (e.g. sign-/zero-extend, or shift) is in the same machine basic block as the instruction we are folding into. Not doing so can result in incorrect code, because the value might not be live-out of the basic block, where the value is defined. This fixes rdar://problem/18169495. llvm-svn: 216700
2014-08-29 08:19:21 +08:00
}
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
}
}
[AArch64][FastISel] Don't fold shifts with UB. We are already falling back to SelectionDAG when encountering an shift with UB. This adds the same checks for shifts with UB that get folded into arithmetic or logical operations. This fixes rdar://problem/22345295. llvm-svn: 245499
2015-08-20 04:52:55 +08:00
}
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
unsigned RHSReg = getRegForValue(RHS);
if (!RHSReg)
return 0;
bool RHSIsKill = hasTrivialKill(RHS);
[FastISel][AArch64] Don't fold the sign-/zero-extend from i1 into the compare. This fixes a bug I introduced in a previous commit (r216033). Sign-/Zero- extension from i1 cannot be folded into the ADDS/SUBS instructions. Instead both operands have to be sign-/zero-extended with separate instructions. Related to <rdar://problem/17913111>. llvm-svn: 216073
2014-08-21 00:34:15 +08:00
if (NeedExtend)
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
RHSReg = emitIntExt(SrcVT, RHSReg, RetVT, IsZExt);
[FastISel][AArch64] Don't fold the sign-/zero-extend from i1 into the compare. This fixes a bug I introduced in a previous commit (r216033). Sign-/Zero- extension from i1 cannot be folded into the ADDS/SUBS instructions. Instead both operands have to be sign-/zero-extended with separate instructions. Related to <rdar://problem/17913111>. llvm-svn: 216073
2014-08-21 00:34:15 +08:00
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
return emitAddSub_rr(UseAdd, RetVT, LHSReg, LHSIsKill, RHSReg, RHSIsKill,
SetFlags, WantResult);
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
}
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
unsigned AArch64FastISel::emitAddSub_rr(bool UseAdd, MVT RetVT, unsigned LHSReg,
bool LHSIsKill, unsigned RHSReg,
bool RHSIsKill, bool SetFlags,
bool WantResult) {
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
assert(LHSReg && RHSReg && "Invalid register number.");
if (RetVT != MVT::i32 && RetVT != MVT::i64)
return 0;
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
static const unsigned OpcTable[2][2][2] = {
{ { AArch64::SUBWrr, AArch64::SUBXrr },
{ AArch64::ADDWrr, AArch64::ADDXrr } },
{ { AArch64::SUBSWrr, AArch64::SUBSXrr },
{ AArch64::ADDSWrr, AArch64::ADDSXrr } }
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
};
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
bool Is64Bit = RetVT == MVT::i64;
unsigned Opc = OpcTable[SetFlags][UseAdd][Is64Bit];
const TargetRegisterClass *RC =
Is64Bit ? &AArch64::GPR64RegClass : &AArch64::GPR32RegClass;
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
unsigned ResultReg;
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
if (WantResult)
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
ResultReg = createResultReg(RC);
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
else
ResultReg = Is64Bit ? AArch64::XZR : AArch64::WZR;
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
const MCInstrDesc &II = TII.get(Opc);
LHSReg = constrainOperandRegClass(II, LHSReg, II.getNumDefs());
RHSReg = constrainOperandRegClass(II, RHSReg, II.getNumDefs() + 1);
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, II, ResultReg)
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
.addReg(LHSReg, getKillRegState(LHSIsKill))
.addReg(RHSReg, getKillRegState(RHSIsKill));
return ResultReg;
}
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
unsigned AArch64FastISel::emitAddSub_ri(bool UseAdd, MVT RetVT, unsigned LHSReg,
bool LHSIsKill, uint64_t Imm,
bool SetFlags, bool WantResult) {
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
assert(LHSReg && "Invalid register number.");
if (RetVT != MVT::i32 && RetVT != MVT::i64)
return 0;
unsigned ShiftImm;
if (isUInt<12>(Imm))
ShiftImm = 0;
else if ((Imm & 0xfff000) == Imm) {
ShiftImm = 12;
Imm >>= 12;
} else
return 0;
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
static const unsigned OpcTable[2][2][2] = {
{ { AArch64::SUBWri, AArch64::SUBXri },
{ AArch64::ADDWri, AArch64::ADDXri } },
{ { AArch64::SUBSWri, AArch64::SUBSXri },
{ AArch64::ADDSWri, AArch64::ADDSXri } }
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
};
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
bool Is64Bit = RetVT == MVT::i64;
unsigned Opc = OpcTable[SetFlags][UseAdd][Is64Bit];
const TargetRegisterClass *RC;
if (SetFlags)
RC = Is64Bit ? &AArch64::GPR64RegClass : &AArch64::GPR32RegClass;
else
RC = Is64Bit ? &AArch64::GPR64spRegClass : &AArch64::GPR32spRegClass;
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
unsigned ResultReg;
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
if (WantResult)
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
ResultReg = createResultReg(RC);
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
else
ResultReg = Is64Bit ? AArch64::XZR : AArch64::WZR;
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
const MCInstrDesc &II = TII.get(Opc);
LHSReg = constrainOperandRegClass(II, LHSReg, II.getNumDefs());
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, II, ResultReg)
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
.addReg(LHSReg, getKillRegState(LHSIsKill))
.addImm(Imm)
.addImm(getShifterImm(AArch64_AM::LSL, ShiftImm));
return ResultReg;
}
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
unsigned AArch64FastISel::emitAddSub_rs(bool UseAdd, MVT RetVT, unsigned LHSReg,
bool LHSIsKill, unsigned RHSReg,
bool RHSIsKill,
[FastISel][AArch64] Fix address simplification. When a shift with extension or an add with shift and extension cannot be folded into the memory operation, then the address calculation has to be materialized separately. While doing so the code forgot to consider a possible sign-/zero- extension. This fix folds now also the sign-/zero-extension into the add or shift instruction which is used to materialize the address. This fixes rdar://problem/18141718. llvm-svn: 216511
2014-08-27 08:58:30 +08:00
AArch64_AM::ShiftExtendType ShiftType,
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
uint64_t ShiftImm, bool SetFlags,
bool WantResult) {
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
assert(LHSReg && RHSReg && "Invalid register number.");
if (RetVT != MVT::i32 && RetVT != MVT::i64)
return 0;
[AArch64][FastISel] Don't fold shifts with UB. We are already falling back to SelectionDAG when encountering an shift with UB. This adds the same checks for shifts with UB that get folded into arithmetic or logical operations. This fixes rdar://problem/22345295. llvm-svn: 245499
2015-08-20 04:52:55 +08:00
// Don't deal with undefined shifts.
if (ShiftImm >= RetVT.getSizeInBits())
return 0;
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
static const unsigned OpcTable[2][2][2] = {
{ { AArch64::SUBWrs, AArch64::SUBXrs },
{ AArch64::ADDWrs, AArch64::ADDXrs } },
{ { AArch64::SUBSWrs, AArch64::SUBSXrs },
{ AArch64::ADDSWrs, AArch64::ADDSXrs } }
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
};
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
bool Is64Bit = RetVT == MVT::i64;
unsigned Opc = OpcTable[SetFlags][UseAdd][Is64Bit];
const TargetRegisterClass *RC =
Is64Bit ? &AArch64::GPR64RegClass : &AArch64::GPR32RegClass;
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
unsigned ResultReg;
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
if (WantResult)
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
ResultReg = createResultReg(RC);
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
else
ResultReg = Is64Bit ? AArch64::XZR : AArch64::WZR;
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
const MCInstrDesc &II = TII.get(Opc);
LHSReg = constrainOperandRegClass(II, LHSReg, II.getNumDefs());
RHSReg = constrainOperandRegClass(II, RHSReg, II.getNumDefs() + 1);
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, II, ResultReg)
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
.addReg(LHSReg, getKillRegState(LHSIsKill))
.addReg(RHSReg, getKillRegState(RHSIsKill))
.addImm(getShifterImm(ShiftType, ShiftImm));
return ResultReg;
}
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
unsigned AArch64FastISel::emitAddSub_rx(bool UseAdd, MVT RetVT, unsigned LHSReg,
bool LHSIsKill, unsigned RHSReg,
bool RHSIsKill,
[FastISel][AArch64] Fix address simplification. When a shift with extension or an add with shift and extension cannot be folded into the memory operation, then the address calculation has to be materialized separately. While doing so the code forgot to consider a possible sign-/zero- extension. This fix folds now also the sign-/zero-extension into the add or shift instruction which is used to materialize the address. This fixes rdar://problem/18141718. llvm-svn: 216511
2014-08-27 08:58:30 +08:00
AArch64_AM::ShiftExtendType ExtType,
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
uint64_t ShiftImm, bool SetFlags,
bool WantResult) {
[FastISel][AArch64] Fix address simplification. When a shift with extension or an add with shift and extension cannot be folded into the memory operation, then the address calculation has to be materialized separately. While doing so the code forgot to consider a possible sign-/zero- extension. This fix folds now also the sign-/zero-extension into the add or shift instruction which is used to materialize the address. This fixes rdar://problem/18141718. llvm-svn: 216511
2014-08-27 08:58:30 +08:00
assert(LHSReg && RHSReg && "Invalid register number.");
if (RetVT != MVT::i32 && RetVT != MVT::i64)
return 0;
[AArch64][FastISel] Don't fold shifts with UB. We are already falling back to SelectionDAG when encountering an shift with UB. This adds the same checks for shifts with UB that get folded into arithmetic or logical operations. This fixes rdar://problem/22345295. llvm-svn: 245499
2015-08-20 04:52:55 +08:00
if (ShiftImm >= 4)
return 0;
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
static const unsigned OpcTable[2][2][2] = {
{ { AArch64::SUBWrx, AArch64::SUBXrx },
{ AArch64::ADDWrx, AArch64::ADDXrx } },
{ { AArch64::SUBSWrx, AArch64::SUBSXrx },
{ AArch64::ADDSWrx, AArch64::ADDSXrx } }
[FastISel][AArch64] Fix address simplification. When a shift with extension or an add with shift and extension cannot be folded into the memory operation, then the address calculation has to be materialized separately. While doing so the code forgot to consider a possible sign-/zero- extension. This fix folds now also the sign-/zero-extension into the add or shift instruction which is used to materialize the address. This fixes rdar://problem/18141718. llvm-svn: 216511
2014-08-27 08:58:30 +08:00
};
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
bool Is64Bit = RetVT == MVT::i64;
unsigned Opc = OpcTable[SetFlags][UseAdd][Is64Bit];
const TargetRegisterClass *RC = nullptr;
if (SetFlags)
RC = Is64Bit ? &AArch64::GPR64RegClass : &AArch64::GPR32RegClass;
else
RC = Is64Bit ? &AArch64::GPR64spRegClass : &AArch64::GPR32spRegClass;
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
unsigned ResultReg;
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
if (WantResult)
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
ResultReg = createResultReg(RC);
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
else
ResultReg = Is64Bit ? AArch64::XZR : AArch64::WZR;
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
const MCInstrDesc &II = TII.get(Opc);
LHSReg = constrainOperandRegClass(II, LHSReg, II.getNumDefs());
RHSReg = constrainOperandRegClass(II, RHSReg, II.getNumDefs() + 1);
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, II, ResultReg)
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
.addReg(LHSReg, getKillRegState(LHSIsKill))
.addReg(RHSReg, getKillRegState(RHSIsKill))
.addImm(getArithExtendImm(ExtType, ShiftImm));
return ResultReg;
}
bool AArch64FastISel::emitCmp(const Value *LHS, const Value *RHS, bool IsZExt) {
Type *Ty = LHS->getType();
Make TargetLowering::getPointerTy() taking DataLayout as an argument Summary: This change is part of a series of commits dedicated to have a single DataLayout during compilation by using always the one owned by the module. Reviewers: echristo Subscribers: jholewinski, ted, yaron.keren, rafael, llvm-commits Differential Revision: http://reviews.llvm.org/D11028 From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 241775
2015-07-09 10:09:04 +08:00
EVT EVT = TLI.getValueType(DL, Ty, true);
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
if (!EVT.isSimple())
return false;
MVT VT = EVT.getSimpleVT();
switch (VT.SimpleTy) {
default:
return false;
case MVT::i1:
case MVT::i8:
case MVT::i16:
case MVT::i32:
case MVT::i64:
return emitICmp(VT, LHS, RHS, IsZExt);
case MVT::f32:
case MVT::f64:
return emitFCmp(VT, LHS, RHS);
}
}
bool AArch64FastISel::emitICmp(MVT RetVT, const Value *LHS, const Value *RHS,
bool IsZExt) {
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
return emitSub(RetVT, LHS, RHS, /*SetFlags=*/true, /*WantResult=*/false,
IsZExt) != 0;
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
}
bool AArch64FastISel::emitICmp_ri(MVT RetVT, unsigned LHSReg, bool LHSIsKill,
uint64_t Imm) {
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
return emitAddSub_ri(/*UseAdd=*/false, RetVT, LHSReg, LHSIsKill, Imm,
/*SetFlags=*/true, /*WantResult=*/false) != 0;
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
}
bool AArch64FastISel::emitFCmp(MVT RetVT, const Value *LHS, const Value *RHS) {
if (RetVT != MVT::f32 && RetVT != MVT::f64)
return false;
// Check to see if the 2nd operand is a constant that we can encode directly
// in the compare.
bool UseImm = false;
if (const auto *CFP = dyn_cast<ConstantFP>(RHS))
if (CFP->isZero() && !CFP->isNegative())
UseImm = true;
unsigned LHSReg = getRegForValue(LHS);
if (!LHSReg)
return false;
bool LHSIsKill = hasTrivialKill(LHS);
if (UseImm) {
unsigned Opc = (RetVT == MVT::f64) ? AArch64::FCMPDri : AArch64::FCMPSri;
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(Opc))
.addReg(LHSReg, getKillRegState(LHSIsKill));
return true;
}
unsigned RHSReg = getRegForValue(RHS);
if (!RHSReg)
return false;
bool RHSIsKill = hasTrivialKill(RHS);
unsigned Opc = (RetVT == MVT::f64) ? AArch64::FCMPDrr : AArch64::FCMPSrr;
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(Opc))
.addReg(LHSReg, getKillRegState(LHSIsKill))
.addReg(RHSReg, getKillRegState(RHSIsKill));
return true;
}
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
unsigned AArch64FastISel::emitAdd(MVT RetVT, const Value *LHS, const Value *RHS,
bool SetFlags, bool WantResult, bool IsZExt) {
return emitAddSub(/*UseAdd=*/true, RetVT, LHS, RHS, SetFlags, WantResult,
IsZExt);
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
}
[FastISel][AArch64] Factor out add with immediate emission into a helper function. NFC. Simplify add with immediate emission by factoring it out into a helper function. llvm-svn: 219830
2014-10-16 02:58:02 +08:00
/// \brief This method is a wrapper to simplify add emission.
///
/// First try to emit an add with an immediate operand using emitAddSub_ri. If
/// that fails, then try to materialize the immediate into a register and use
/// emitAddSub_rr instead.
unsigned AArch64FastISel::emitAdd_ri_(MVT VT, unsigned Op0, bool Op0IsKill,
int64_t Imm) {
unsigned ResultReg;
if (Imm < 0)
ResultReg = emitAddSub_ri(false, VT, Op0, Op0IsKill, -Imm);
else
ResultReg = emitAddSub_ri(true, VT, Op0, Op0IsKill, Imm);
if (ResultReg)
return ResultReg;
unsigned CReg = fastEmit_i(VT, VT, ISD::Constant, Imm);
if (!CReg)
return 0;
ResultReg = emitAddSub_rr(true, VT, Op0, Op0IsKill, CReg, true);
return ResultReg;
}
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
unsigned AArch64FastISel::emitSub(MVT RetVT, const Value *LHS, const Value *RHS,
bool SetFlags, bool WantResult, bool IsZExt) {
return emitAddSub(/*UseAdd=*/false, RetVT, LHS, RHS, SetFlags, WantResult,
IsZExt);
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
}
unsigned AArch64FastISel::emitSubs_rr(MVT RetVT, unsigned LHSReg,
bool LHSIsKill, unsigned RHSReg,
bool RHSIsKill, bool WantResult) {
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
return emitAddSub_rr(/*UseAdd=*/false, RetVT, LHSReg, LHSIsKill, RHSReg,
RHSIsKill, /*SetFlags=*/true, WantResult);
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
}
unsigned AArch64FastISel::emitSubs_rs(MVT RetVT, unsigned LHSReg,
bool LHSIsKill, unsigned RHSReg,
bool RHSIsKill,
AArch64_AM::ShiftExtendType ShiftType,
uint64_t ShiftImm, bool WantResult) {
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
return emitAddSub_rs(/*UseAdd=*/false, RetVT, LHSReg, LHSIsKill, RHSReg,
RHSIsKill, ShiftType, ShiftImm, /*SetFlags=*/true,
WantResult);
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
}
[FastISel][AArch64] Add target-specific lowering for logical operations. This change adds support for immediate and shift-left folding into logical operations. This fixes rdar://problem/18223183. llvm-svn: 217118
2014-09-04 09:29:18 +08:00
unsigned AArch64FastISel::emitLogicalOp(unsigned ISDOpc, MVT RetVT,
const Value *LHS, const Value *RHS) {
// Canonicalize immediates to the RHS first.
if (isa<ConstantInt>(LHS) && !isa<ConstantInt>(RHS))
std::swap(LHS, RHS);
[FastISel][AArch64] Fold mul into add/sub and logical operations. Try to fold the multiply into the add/sub or logical operations (when possible). This is related to rdar://problem/18369687. llvm-svn: 217978
2014-09-18 03:51:38 +08:00
// Canonicalize mul by power-of-2 to the RHS.
if (LHS->hasOneUse() && isValueAvailable(LHS))
if (isMulPowOf2(LHS))
std::swap(LHS, RHS);
[FastISel][AArch64] Add target-specific lowering for logical operations. This change adds support for immediate and shift-left folding into logical operations. This fixes rdar://problem/18223183. llvm-svn: 217118
2014-09-04 09:29:18 +08:00
// Canonicalize shift immediate to the RHS.
[FastISel][AArch64] Fold mul into add/sub and logical operations. Try to fold the multiply into the add/sub or logical operations (when possible). This is related to rdar://problem/18369687. llvm-svn: 217978
2014-09-18 03:51:38 +08:00
if (LHS->hasOneUse() && isValueAvailable(LHS))
if (const auto *SI = dyn_cast<ShlOperator>(LHS))
[FastISel][AArch64] Add target-specific lowering for logical operations. This change adds support for immediate and shift-left folding into logical operations. This fixes rdar://problem/18223183. llvm-svn: 217118
2014-09-04 09:29:18 +08:00
if (isa<ConstantInt>(SI->getOperand(1)))
[FastISel][AArch64] Fold mul into add/sub and logical operations. Try to fold the multiply into the add/sub or logical operations (when possible). This is related to rdar://problem/18369687. llvm-svn: 217978
2014-09-18 03:51:38 +08:00
std::swap(LHS, RHS);
[FastISel][AArch64] Add target-specific lowering for logical operations. This change adds support for immediate and shift-left folding into logical operations. This fixes rdar://problem/18223183. llvm-svn: 217118
2014-09-04 09:29:18 +08:00
unsigned LHSReg = getRegForValue(LHS);
if (!LHSReg)
return 0;
bool LHSIsKill = hasTrivialKill(LHS);
unsigned ResultReg = 0;
if (const auto *C = dyn_cast<ConstantInt>(RHS)) {
uint64_t Imm = C->getZExtValue();
ResultReg = emitLogicalOp_ri(ISDOpc, RetVT, LHSReg, LHSIsKill, Imm);
}
if (ResultReg)
return ResultReg;
[FastISel][AArch64] Fold mul into add/sub and logical operations. Try to fold the multiply into the add/sub or logical operations (when possible). This is related to rdar://problem/18369687. llvm-svn: 217978
2014-09-18 03:51:38 +08:00
// Check if the mul can be folded into the instruction.
[AArch64][FastISel] Don't fold shifts with UB. We are already falling back to SelectionDAG when encountering an shift with UB. This adds the same checks for shifts with UB that get folded into arithmetic or logical operations. This fixes rdar://problem/22345295. llvm-svn: 245499
2015-08-20 04:52:55 +08:00
if (RHS->hasOneUse() && isValueAvailable(RHS)) {
[FastISel][AArch64] Fold mul into add/sub and logical operations. Try to fold the multiply into the add/sub or logical operations (when possible). This is related to rdar://problem/18369687. llvm-svn: 217978
2014-09-18 03:51:38 +08:00
if (isMulPowOf2(RHS)) {
const Value *MulLHS = cast<MulOperator>(RHS)->getOperand(0);
const Value *MulRHS = cast<MulOperator>(RHS)->getOperand(1);
if (const auto *C = dyn_cast<ConstantInt>(MulLHS))
if (C->getValue().isPowerOf2())
std::swap(MulLHS, MulRHS);
assert(isa<ConstantInt>(MulRHS) && "Expected a ConstantInt.");
uint64_t ShiftVal = cast<ConstantInt>(MulRHS)->getValue().logBase2();
unsigned RHSReg = getRegForValue(MulLHS);
if (!RHSReg)
return 0;
bool RHSIsKill = hasTrivialKill(MulLHS);
[AArch64][FastISel] Don't fold shifts with UB. We are already falling back to SelectionDAG when encountering an shift with UB. This adds the same checks for shifts with UB that get folded into arithmetic or logical operations. This fixes rdar://problem/22345295. llvm-svn: 245499
2015-08-20 04:52:55 +08:00
ResultReg = emitLogicalOp_rs(ISDOpc, RetVT, LHSReg, LHSIsKill, RHSReg,
RHSIsKill, ShiftVal);
if (ResultReg)
return ResultReg;
[FastISel][AArch64] Fold mul into add/sub and logical operations. Try to fold the multiply into the add/sub or logical operations (when possible). This is related to rdar://problem/18369687. llvm-svn: 217978
2014-09-18 03:51:38 +08:00
}
[AArch64][FastISel] Don't fold shifts with UB. We are already falling back to SelectionDAG when encountering an shift with UB. This adds the same checks for shifts with UB that get folded into arithmetic or logical operations. This fixes rdar://problem/22345295. llvm-svn: 245499
2015-08-20 04:52:55 +08:00
}
[FastISel][AArch64] Fold mul into add/sub and logical operations. Try to fold the multiply into the add/sub or logical operations (when possible). This is related to rdar://problem/18369687. llvm-svn: 217978
2014-09-18 03:51:38 +08:00
[FastISel][AArch64] Add target-specific lowering for logical operations. This change adds support for immediate and shift-left folding into logical operations. This fixes rdar://problem/18223183. llvm-svn: 217118
2014-09-04 09:29:18 +08:00
// Check if the shift can be folded into the instruction.
[AArch64][FastISel] Don't fold shifts with UB. We are already falling back to SelectionDAG when encountering an shift with UB. This adds the same checks for shifts with UB that get folded into arithmetic or logical operations. This fixes rdar://problem/22345295. llvm-svn: 245499
2015-08-20 04:52:55 +08:00
if (RHS->hasOneUse() && isValueAvailable(RHS)) {
[FastISel][AArch64] Fold mul into add/sub and logical operations. Try to fold the multiply into the add/sub or logical operations (when possible). This is related to rdar://problem/18369687. llvm-svn: 217978
2014-09-18 03:51:38 +08:00
if (const auto *SI = dyn_cast<ShlOperator>(RHS))
if (const auto *C = dyn_cast<ConstantInt>(SI->getOperand(1))) {
uint64_t ShiftVal = C->getZExtValue();
unsigned RHSReg = getRegForValue(SI->getOperand(0));
if (!RHSReg)
return 0;
bool RHSIsKill = hasTrivialKill(SI->getOperand(0));
[AArch64][FastISel] Don't fold shifts with UB. We are already falling back to SelectionDAG when encountering an shift with UB. This adds the same checks for shifts with UB that get folded into arithmetic or logical operations. This fixes rdar://problem/22345295. llvm-svn: 245499
2015-08-20 04:52:55 +08:00
ResultReg = emitLogicalOp_rs(ISDOpc, RetVT, LHSReg, LHSIsKill, RHSReg,
RHSIsKill, ShiftVal);
if (ResultReg)
return ResultReg;
[FastISel][AArch64] Fold mul into add/sub and logical operations. Try to fold the multiply into the add/sub or logical operations (when possible). This is related to rdar://problem/18369687. llvm-svn: 217978
2014-09-18 03:51:38 +08:00
}
[AArch64][FastISel] Don't fold shifts with UB. We are already falling back to SelectionDAG when encountering an shift with UB. This adds the same checks for shifts with UB that get folded into arithmetic or logical operations. This fixes rdar://problem/22345295. llvm-svn: 245499
2015-08-20 04:52:55 +08:00
}
[FastISel][AArch64] Add target-specific lowering for logical operations. This change adds support for immediate and shift-left folding into logical operations. This fixes rdar://problem/18223183. llvm-svn: 217118
2014-09-04 09:29:18 +08:00
unsigned RHSReg = getRegForValue(RHS);
if (!RHSReg)
return 0;
bool RHSIsKill = hasTrivialKill(RHS);
[FastISel][AArch64] Add support for non-native types for logical ops. Extend the logical ops selection to also support non-native types such as i1, i8, and i16. Fixes rdar://problem/18330589. llvm-svn: 217732
2014-09-14 07:46:28 +08:00
MVT VT = std::max(MVT::i32, RetVT.SimpleTy);
ResultReg = fastEmit_rr(VT, VT, ISDOpc, LHSReg, LHSIsKill, RHSReg, RHSIsKill);
if (RetVT >= MVT::i8 && RetVT <= MVT::i16) {
uint64_t Mask = (RetVT == MVT::i8) ? 0xff : 0xffff;
ResultReg = emitAnd_ri(MVT::i32, ResultReg, /*IsKill=*/true, Mask);
}
return ResultReg;
[FastISel][AArch64] Add target-specific lowering for logical operations. This change adds support for immediate and shift-left folding into logical operations. This fixes rdar://problem/18223183. llvm-svn: 217118
2014-09-04 09:29:18 +08:00
}
unsigned AArch64FastISel::emitLogicalOp_ri(unsigned ISDOpc, MVT RetVT,
unsigned LHSReg, bool LHSIsKill,
uint64_t Imm) {
Apply clang-tidy's misc-static-assert where it makes sense. Also fold conditions into assert(0) where it makes sense. No functional change intended. llvm-svn: 270982
2016-05-27 19:36:04 +08:00
static_assert((ISD::AND + 1 == ISD::OR) && (ISD::AND + 2 == ISD::XOR),
"ISD nodes are not consecutive!");
[FastISel][AArch64] Add target-specific lowering for logical operations. This change adds support for immediate and shift-left folding into logical operations. This fixes rdar://problem/18223183. llvm-svn: 217118
2014-09-04 09:29:18 +08:00
static const unsigned OpcTable[3][2] = {
{ AArch64::ANDWri, AArch64::ANDXri },
{ AArch64::ORRWri, AArch64::ORRXri },
{ AArch64::EORWri, AArch64::EORXri }
};
const TargetRegisterClass *RC;
unsigned Opc;
unsigned RegSize;
[FastISel][AArch64] Factor out ANDWri instruction generation into a helper function. NFCI. llvm-svn: 216199
2014-08-22 02:02:25 +08:00
switch (RetVT.SimpleTy) {
default:
return 0;
[FastISel][AArch64] Add support for non-native types for logical ops. Extend the logical ops selection to also support non-native types such as i1, i8, and i16. Fixes rdar://problem/18330589. llvm-svn: 217732
2014-09-14 07:46:28 +08:00
case MVT::i1:
case MVT::i8:
case MVT::i16:
[FastISel][AArch64] Add target-specific lowering for logical operations. This change adds support for immediate and shift-left folding into logical operations. This fixes rdar://problem/18223183. llvm-svn: 217118
2014-09-04 09:29:18 +08:00
case MVT::i32: {
unsigned Idx = ISDOpc - ISD::AND;
Opc = OpcTable[Idx][0];
[FastISel][AArch64] Factor out ANDWri instruction generation into a helper function. NFCI. llvm-svn: 216199
2014-08-22 02:02:25 +08:00
RC = &AArch64::GPR32spRegClass;
RegSize = 32;
break;
[FastISel][AArch64] Add target-specific lowering for logical operations. This change adds support for immediate and shift-left folding into logical operations. This fixes rdar://problem/18223183. llvm-svn: 217118
2014-09-04 09:29:18 +08:00
}
[FastISel][AArch64] Factor out ANDWri instruction generation into a helper function. NFCI. llvm-svn: 216199
2014-08-22 02:02:25 +08:00
case MVT::i64:
[FastISel][AArch64] Add target-specific lowering for logical operations. This change adds support for immediate and shift-left folding into logical operations. This fixes rdar://problem/18223183. llvm-svn: 217118
2014-09-04 09:29:18 +08:00
Opc = OpcTable[ISDOpc - ISD::AND][1];
[FastISel][AArch64] Factor out ANDWri instruction generation into a helper function. NFCI. llvm-svn: 216199
2014-08-22 02:02:25 +08:00
RC = &AArch64::GPR64spRegClass;
RegSize = 64;
break;
}
if (!AArch64_AM::isLogicalImmediate(Imm, RegSize))
return 0;
[FastISel][AArch64] Add support for non-native types for logical ops. Extend the logical ops selection to also support non-native types such as i1, i8, and i16. Fixes rdar://problem/18330589. llvm-svn: 217732
2014-09-14 07:46:28 +08:00
unsigned ResultReg =
fastEmitInst_ri(Opc, RC, LHSReg, LHSIsKill,
AArch64_AM::encodeLogicalImmediate(Imm, RegSize));
if (RetVT >= MVT::i8 && RetVT <= MVT::i16 && ISDOpc != ISD::AND) {
uint64_t Mask = (RetVT == MVT::i8) ? 0xff : 0xffff;
ResultReg = emitAnd_ri(MVT::i32, ResultReg, /*IsKill=*/true, Mask);
}
return ResultReg;
[FastISel][AArch64] Factor out ANDWri instruction generation into a helper function. NFCI. llvm-svn: 216199
2014-08-22 02:02:25 +08:00
}
[FastISel][AArch64] Add target-specific lowering for logical operations. This change adds support for immediate and shift-left folding into logical operations. This fixes rdar://problem/18223183. llvm-svn: 217118
2014-09-04 09:29:18 +08:00
unsigned AArch64FastISel::emitLogicalOp_rs(unsigned ISDOpc, MVT RetVT,
unsigned LHSReg, bool LHSIsKill,
unsigned RHSReg, bool RHSIsKill,
uint64_t ShiftImm) {
Apply clang-tidy's misc-static-assert where it makes sense. Also fold conditions into assert(0) where it makes sense. No functional change intended. llvm-svn: 270982
2016-05-27 19:36:04 +08:00
static_assert((ISD::AND + 1 == ISD::OR) && (ISD::AND + 2 == ISD::XOR),
"ISD nodes are not consecutive!");
[FastISel][AArch64] Add target-specific lowering for logical operations. This change adds support for immediate and shift-left folding into logical operations. This fixes rdar://problem/18223183. llvm-svn: 217118
2014-09-04 09:29:18 +08:00
static const unsigned OpcTable[3][2] = {
{ AArch64::ANDWrs, AArch64::ANDXrs },
{ AArch64::ORRWrs, AArch64::ORRXrs },
{ AArch64::EORWrs, AArch64::EORXrs }
};
[AArch64][FastISel] Don't fold shifts with UB. We are already falling back to SelectionDAG when encountering an shift with UB. This adds the same checks for shifts with UB that get folded into arithmetic or logical operations. This fixes rdar://problem/22345295. llvm-svn: 245499
2015-08-20 04:52:55 +08:00
// Don't deal with undefined shifts.
if (ShiftImm >= RetVT.getSizeInBits())
return 0;
[FastISel][AArch64] Add target-specific lowering for logical operations. This change adds support for immediate and shift-left folding into logical operations. This fixes rdar://problem/18223183. llvm-svn: 217118
2014-09-04 09:29:18 +08:00
const TargetRegisterClass *RC;
unsigned Opc;
switch (RetVT.SimpleTy) {
[FastISel][AArch64] Add support for non-native types for logical ops. Extend the logical ops selection to also support non-native types such as i1, i8, and i16. Fixes rdar://problem/18330589. llvm-svn: 217732
2014-09-14 07:46:28 +08:00
default:
return 0;
case MVT::i1:
case MVT::i8:
case MVT::i16:
case MVT::i32:
Opc = OpcTable[ISDOpc - ISD::AND][0];
RC = &AArch64::GPR32RegClass;
break;
case MVT::i64:
Opc = OpcTable[ISDOpc - ISD::AND][1];
RC = &AArch64::GPR64RegClass;
break;
[FastISel][AArch64] Add target-specific lowering for logical operations. This change adds support for immediate and shift-left folding into logical operations. This fixes rdar://problem/18223183. llvm-svn: 217118
2014-09-04 09:29:18 +08:00
}
[FastISel][AArch64] Add support for non-native types for logical ops. Extend the logical ops selection to also support non-native types such as i1, i8, and i16. Fixes rdar://problem/18330589. llvm-svn: 217732
2014-09-14 07:46:28 +08:00
unsigned ResultReg =
fastEmitInst_rri(Opc, RC, LHSReg, LHSIsKill, RHSReg, RHSIsKill,
AArch64_AM::getShifterImm(AArch64_AM::LSL, ShiftImm));
if (RetVT >= MVT::i8 && RetVT <= MVT::i16) {
uint64_t Mask = (RetVT == MVT::i8) ? 0xff : 0xffff;
ResultReg = emitAnd_ri(MVT::i32, ResultReg, /*IsKill=*/true, Mask);
}
return ResultReg;
[FastISel][AArch64] Add target-specific lowering for logical operations. This change adds support for immediate and shift-left folding into logical operations. This fixes rdar://problem/18223183. llvm-svn: 217118
2014-09-04 09:29:18 +08:00
}
unsigned AArch64FastISel::emitAnd_ri(MVT RetVT, unsigned LHSReg, bool LHSIsKill,
uint64_t Imm) {
return emitLogicalOp_ri(ISD::AND, RetVT, LHSReg, LHSIsKill, Imm);
}
[FastISel][AArch64] Fix sign-/zero-extend folding when SelectionDAG is involved. Sign-/zero-extend folding depended on the load and the integer extend to be both selected by FastISel. This cannot always be garantueed and SelectionDAG might interfer. This commit adds additonal checks to load and integer extend lowering to catch this. Related to rdar://problem/18495928. llvm-svn: 219716
2014-10-15 04:36:02 +08:00
unsigned AArch64FastISel::emitLoad(MVT VT, MVT RetVT, Address Addr,
bool WantZExt, MachineMemOperand *MMO) {
fix formatting; NFC llvm-svn: 241175
2015-07-02 01:58:53 +08:00
if (!TLI.allowsMisalignedMemoryAccesses(VT))
On behalf of Alexandros Lamprineas: LLVM targeting aarch64 doesn't correctly produce aligned accesses for non-aligned data at -O0/fast-isel (-mno-unaligned-access). The root cause seems to be in fast-isel not producing unaligned access correctly for -mno-unaligned-access. The patch just aborts fast-isel for loads and stores when -mno-unaligned-access is present. The regression test is updated to check this new test case (-mno-unaligned-access together with fast-isel). Differential Revision: http://reviews.llvm.org/D10360 llvm-svn: 239732
2015-06-15 23:48:44 +08:00
return 0;
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
// Simplify this down to something we can handle.
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
if (!simplifyAddress(Addr, VT))
[FastISel][AArch64] Fix sign-/zero-extend folding when SelectionDAG is involved. Sign-/zero-extend folding depended on the load and the integer extend to be both selected by FastISel. This cannot always be garantueed and SelectionDAG might interfer. This commit adds additonal checks to load and integer extend lowering to catch this. Related to rdar://problem/18495928. llvm-svn: 219716
2014-10-15 04:36:02 +08:00
return 0;
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
[FastISel][AArch64] Factor out scale factor calculation. NFC. Factor out the code that determines the implicit scale factor of memory operations for a given value type. llvm-svn: 218652
2014-09-30 08:49:54 +08:00
unsigned ScaleFactor = getImplicitScaleFactor(VT);
if (!ScaleFactor)
llvm_unreachable("Unexpected value type.");
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Negative offsets require unscaled, 9-bit, signed immediate offsets.
// Otherwise, we try using scaled, 12-bit, unsigned immediate offsets.
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
bool UseScaled = true;
if ((Addr.getOffset() < 0) || (Addr.getOffset() & (ScaleFactor - 1))) {
UseScaled = false;
ScaleFactor = 1;
}
[FastISel][AArch64] Fix "Fold sign-/zero-extends into the load instruction." This commit fixes an issue with sign-/zero-extending loads that was discovered by Richard Barton. We use now the correct load instructions for sign-extending loads to 64bit. Also updated and added more unit tests. llvm-svn: 219185
2014-10-07 11:39:59 +08:00
static const unsigned GPOpcTable[2][8][4] = {
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
// Sign-extend.
[FastISel][AArch64] Fix "Fold sign-/zero-extends into the load instruction." This commit fixes an issue with sign-/zero-extending loads that was discovered by Richard Barton. We use now the correct load instructions for sign-extending loads to 64bit. Also updated and added more unit tests. llvm-svn: 219185
2014-10-07 11:39:59 +08:00
{ { AArch64::LDURSBWi, AArch64::LDURSHWi, AArch64::LDURWi,
AArch64::LDURXi },
{ AArch64::LDURSBXi, AArch64::LDURSHXi, AArch64::LDURSWi,
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
AArch64::LDURXi },
[FastISel][AArch64] Fix "Fold sign-/zero-extends into the load instruction." This commit fixes an issue with sign-/zero-extending loads that was discovered by Richard Barton. We use now the correct load instructions for sign-extending loads to 64bit. Also updated and added more unit tests. llvm-svn: 219185
2014-10-07 11:39:59 +08:00
{ AArch64::LDRSBWui, AArch64::LDRSHWui, AArch64::LDRWui,
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
AArch64::LDRXui },
[FastISel][AArch64] Fix "Fold sign-/zero-extends into the load instruction." This commit fixes an issue with sign-/zero-extending loads that was discovered by Richard Barton. We use now the correct load instructions for sign-extending loads to 64bit. Also updated and added more unit tests. llvm-svn: 219185
2014-10-07 11:39:59 +08:00
{ AArch64::LDRSBXui, AArch64::LDRSHXui, AArch64::LDRSWui,
AArch64::LDRXui },
{ AArch64::LDRSBWroX, AArch64::LDRSHWroX, AArch64::LDRWroX,
AArch64::LDRXroX },
{ AArch64::LDRSBXroX, AArch64::LDRSHXroX, AArch64::LDRSWroX,
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
AArch64::LDRXroX },
[FastISel][AArch64] Fix "Fold sign-/zero-extends into the load instruction." This commit fixes an issue with sign-/zero-extending loads that was discovered by Richard Barton. We use now the correct load instructions for sign-extending loads to 64bit. Also updated and added more unit tests. llvm-svn: 219185
2014-10-07 11:39:59 +08:00
{ AArch64::LDRSBWroW, AArch64::LDRSHWroW, AArch64::LDRWroW,
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
AArch64::LDRXroW },
[FastISel][AArch64] Fix "Fold sign-/zero-extends into the load instruction." This commit fixes an issue with sign-/zero-extending loads that was discovered by Richard Barton. We use now the correct load instructions for sign-extending loads to 64bit. Also updated and added more unit tests. llvm-svn: 219185
2014-10-07 11:39:59 +08:00
{ AArch64::LDRSBXroW, AArch64::LDRSHXroW, AArch64::LDRSWroW,
AArch64::LDRXroW }
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
},
// Zero-extend.
{ { AArch64::LDURBBi, AArch64::LDURHHi, AArch64::LDURWi,
AArch64::LDURXi },
[FastISel][AArch64] Fix "Fold sign-/zero-extends into the load instruction." This commit fixes an issue with sign-/zero-extending loads that was discovered by Richard Barton. We use now the correct load instructions for sign-extending loads to 64bit. Also updated and added more unit tests. llvm-svn: 219185
2014-10-07 11:39:59 +08:00
{ AArch64::LDURBBi, AArch64::LDURHHi, AArch64::LDURWi,
AArch64::LDURXi },
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
{ AArch64::LDRBBui, AArch64::LDRHHui, AArch64::LDRWui,
AArch64::LDRXui },
[FastISel][AArch64] Fix "Fold sign-/zero-extends into the load instruction." This commit fixes an issue with sign-/zero-extending loads that was discovered by Richard Barton. We use now the correct load instructions for sign-extending loads to 64bit. Also updated and added more unit tests. llvm-svn: 219185
2014-10-07 11:39:59 +08:00
{ AArch64::LDRBBui, AArch64::LDRHHui, AArch64::LDRWui,
AArch64::LDRXui },
{ AArch64::LDRBBroX, AArch64::LDRHHroX, AArch64::LDRWroX,
AArch64::LDRXroX },
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
{ AArch64::LDRBBroX, AArch64::LDRHHroX, AArch64::LDRWroX,
AArch64::LDRXroX },
[FastISel][AArch64] Fix "Fold sign-/zero-extends into the load instruction." This commit fixes an issue with sign-/zero-extending loads that was discovered by Richard Barton. We use now the correct load instructions for sign-extending loads to 64bit. Also updated and added more unit tests. llvm-svn: 219185
2014-10-07 11:39:59 +08:00
{ AArch64::LDRBBroW, AArch64::LDRHHroW, AArch64::LDRWroW,
AArch64::LDRXroW },
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
{ AArch64::LDRBBroW, AArch64::LDRHHroW, AArch64::LDRWroW,
AArch64::LDRXroW }
}
};
static const unsigned FPOpcTable[4][2] = {
{ AArch64::LDURSi, AArch64::LDURDi },
{ AArch64::LDRSui, AArch64::LDRDui },
{ AArch64::LDRSroX, AArch64::LDRDroX },
{ AArch64::LDRSroW, AArch64::LDRDroW }
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
};
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
unsigned Opc;
const TargetRegisterClass *RC;
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
bool UseRegOffset = Addr.isRegBase() && !Addr.getOffset() && Addr.getReg() &&
Addr.getOffsetReg();
unsigned Idx = UseRegOffset ? 2 : UseScaled ? 1 : 0;
if (Addr.getExtendType() == AArch64_AM::UXTW ||
Addr.getExtendType() == AArch64_AM::SXTW)
Idx++;
[FastISel][AArch64] Fix "Fold sign-/zero-extends into the load instruction." This commit fixes an issue with sign-/zero-extending loads that was discovered by Richard Barton. We use now the correct load instructions for sign-extending loads to 64bit. Also updated and added more unit tests. llvm-svn: 219185
2014-10-07 11:39:59 +08:00
bool IsRet64Bit = RetVT == MVT::i64;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
switch (VT.SimpleTy) {
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
default:
llvm_unreachable("Unexpected value type.");
case MVT::i1: // Intentional fall-through.
case MVT::i8:
[FastISel][AArch64] Fix "Fold sign-/zero-extends into the load instruction." This commit fixes an issue with sign-/zero-extending loads that was discovered by Richard Barton. We use now the correct load instructions for sign-extending loads to 64bit. Also updated and added more unit tests. llvm-svn: 219185
2014-10-07 11:39:59 +08:00
Opc = GPOpcTable[WantZExt][2 * Idx + IsRet64Bit][0];
RC = (IsRet64Bit && !WantZExt) ?
&AArch64::GPR64RegClass: &AArch64::GPR32RegClass;
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
break;
case MVT::i16:
[FastISel][AArch64] Fix "Fold sign-/zero-extends into the load instruction." This commit fixes an issue with sign-/zero-extending loads that was discovered by Richard Barton. We use now the correct load instructions for sign-extending loads to 64bit. Also updated and added more unit tests. llvm-svn: 219185
2014-10-07 11:39:59 +08:00
Opc = GPOpcTable[WantZExt][2 * Idx + IsRet64Bit][1];
RC = (IsRet64Bit && !WantZExt) ?
&AArch64::GPR64RegClass: &AArch64::GPR32RegClass;
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
break;
case MVT::i32:
[FastISel][AArch64] Fix "Fold sign-/zero-extends into the load instruction." This commit fixes an issue with sign-/zero-extending loads that was discovered by Richard Barton. We use now the correct load instructions for sign-extending loads to 64bit. Also updated and added more unit tests. llvm-svn: 219185
2014-10-07 11:39:59 +08:00
Opc = GPOpcTable[WantZExt][2 * Idx + IsRet64Bit][2];
RC = (IsRet64Bit && !WantZExt) ?
&AArch64::GPR64RegClass: &AArch64::GPR32RegClass;
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
break;
case MVT::i64:
[FastISel][AArch64] Fix "Fold sign-/zero-extends into the load instruction." This commit fixes an issue with sign-/zero-extending loads that was discovered by Richard Barton. We use now the correct load instructions for sign-extending loads to 64bit. Also updated and added more unit tests. llvm-svn: 219185
2014-10-07 11:39:59 +08:00
Opc = GPOpcTable[WantZExt][2 * Idx + IsRet64Bit][3];
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
RC = &AArch64::GPR64RegClass;
break;
case MVT::f32:
Opc = FPOpcTable[Idx][0];
RC = &AArch64::FPR32RegClass;
break;
case MVT::f64:
Opc = FPOpcTable[Idx][1];
RC = &AArch64::FPR64RegClass;
break;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
}
// Create the base instruction, then add the operands.
[FastISel][AArch64] Fix sign-/zero-extend folding when SelectionDAG is involved. Sign-/zero-extend folding depended on the load and the integer extend to be both selected by FastISel. This cannot always be garantueed and SelectionDAG might interfer. This commit adds additonal checks to load and integer extend lowering to catch this. Related to rdar://problem/18495928. llvm-svn: 219716
2014-10-15 04:36:02 +08:00
unsigned ResultReg = createResultReg(RC);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
MachineInstrBuilder MIB = BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc,
TII.get(Opc), ResultReg);
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
addLoadStoreOperands(Addr, MIB, MachineMemOperand::MOLoad, ScaleFactor, MMO);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
[FastISel][AArch64] Fix sign-/zero-extend folding when SelectionDAG is involved. Sign-/zero-extend folding depended on the load and the integer extend to be both selected by FastISel. This cannot always be garantueed and SelectionDAG might interfer. This commit adds additonal checks to load and integer extend lowering to catch this. Related to rdar://problem/18495928. llvm-svn: 219716
2014-10-15 04:36:02 +08:00
// Loading an i1 requires special handling.
if (VT == MVT::i1) {
unsigned ANDReg = emitAnd_ri(MVT::i32, ResultReg, /*IsKill=*/true, 1);
assert(ANDReg && "Unexpected AND instruction emission failure.");
ResultReg = ANDReg;
}
[FastISel][AArch64] Fix "Fold sign-/zero-extends into the load instruction." This commit fixes an issue with sign-/zero-extending loads that was discovered by Richard Barton. We use now the correct load instructions for sign-extending loads to 64bit. Also updated and added more unit tests. llvm-svn: 219185
2014-10-07 11:39:59 +08:00
// For zero-extending loads to 64bit we emit a 32bit load and then convert
[FastISel][AArch64] Fix sign-/zero-extend folding when SelectionDAG is involved. Sign-/zero-extend folding depended on the load and the integer extend to be both selected by FastISel. This cannot always be garantueed and SelectionDAG might interfer. This commit adds additonal checks to load and integer extend lowering to catch this. Related to rdar://problem/18495928. llvm-svn: 219716
2014-10-15 04:36:02 +08:00
// the 32bit reg to a 64bit reg.
[FastISel][AArch64] Fix "Fold sign-/zero-extends into the load instruction." This commit fixes an issue with sign-/zero-extending loads that was discovered by Richard Barton. We use now the correct load instructions for sign-extending loads to 64bit. Also updated and added more unit tests. llvm-svn: 219185
2014-10-07 11:39:59 +08:00
if (WantZExt && RetVT == MVT::i64 && VT <= MVT::i32) {
unsigned Reg64 = createResultReg(&AArch64::GPR64RegClass);
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc,
TII.get(AArch64::SUBREG_TO_REG), Reg64)
.addImm(0)
.addReg(ResultReg, getKillRegState(true))
.addImm(AArch64::sub_32);
ResultReg = Reg64;
}
[FastISel][AArch64] Fix sign-/zero-extend folding when SelectionDAG is involved. Sign-/zero-extend folding depended on the load and the integer extend to be both selected by FastISel. This cannot always be garantueed and SelectionDAG might interfer. This commit adds additonal checks to load and integer extend lowering to catch this. Related to rdar://problem/18495928. llvm-svn: 219716
2014-10-15 04:36:02 +08:00
return ResultReg;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
}
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
bool AArch64FastISel::selectAddSub(const Instruction *I) {
MVT VT;
[FastISel][AArch64] Refactor selectAddSub, selectLogicalOp, and SelectShift. NFC. Small refactor to tidy up the code a little. llvm-svn: 217827
2014-09-16 05:27:56 +08:00
if (!isTypeSupported(I->getType(), VT, /*IsVectorAllowed=*/true))
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
return false;
[FastISel][AArch64] Refactor selectAddSub, selectLogicalOp, and SelectShift. NFC. Small refactor to tidy up the code a little. llvm-svn: 217827
2014-09-16 05:27:56 +08:00
if (VT.isVector())
return selectOperator(I, I->getOpcode());
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
unsigned ResultReg;
[FastISel][AArch64] Refactor selectAddSub, selectLogicalOp, and SelectShift. NFC. Small refactor to tidy up the code a little. llvm-svn: 217827
2014-09-16 05:27:56 +08:00
switch (I->getOpcode()) {
default:
llvm_unreachable("Unexpected instruction.");
case Instruction::Add:
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
ResultReg = emitAdd(VT, I->getOperand(0), I->getOperand(1));
[FastISel][AArch64] Refactor selectAddSub, selectLogicalOp, and SelectShift. NFC. Small refactor to tidy up the code a little. llvm-svn: 217827
2014-09-16 05:27:56 +08:00
break;
case Instruction::Sub:
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
ResultReg = emitSub(VT, I->getOperand(0), I->getOperand(1));
[FastISel][AArch64] Refactor selectAddSub, selectLogicalOp, and SelectShift. NFC. Small refactor to tidy up the code a little. llvm-svn: 217827
2014-09-16 05:27:56 +08:00
break;
}
if (!ResultReg)
return false;
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
[FastISel] Rename public visible FastISel functions. NFC. This commit renames the following public FastISel functions: LowerArguments -> lowerArguments SelectInstruction -> selectInstruction TargetSelectInstruction -> fastSelectInstruction FastLowerArguments -> fastLowerArguments FastLowerCall -> fastLowerCall FastLowerIntrinsicCall -> fastLowerIntrinsicCall FastEmitZExtFromI1 -> fastEmitZExtFromI1 FastEmitBranch -> fastEmitBranch UpdateValueMap -> updateValueMap TargetMaterializeConstant -> fastMaterializeConstant TargetMaterializeAlloca -> fastMaterializeAlloca TargetMaterializeFloatZero -> fastMaterializeFloatZero LowerCallTo -> lowerCallTo Reviewed by Eric llvm-svn: 217074
2014-09-04 04:56:52 +08:00
updateValueMap(I, ResultReg);
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
return true;
}
[FastISel][AArch64] Refactor selectAddSub, selectLogicalOp, and SelectShift. NFC. Small refactor to tidy up the code a little. llvm-svn: 217827
2014-09-16 05:27:56 +08:00
bool AArch64FastISel::selectLogicalOp(const Instruction *I) {
[FastISel][AArch64] Add target-specific lowering for logical operations. This change adds support for immediate and shift-left folding into logical operations. This fixes rdar://problem/18223183. llvm-svn: 217118
2014-09-04 09:29:18 +08:00
MVT VT;
[FastISel][AArch64] Refactor selectAddSub, selectLogicalOp, and SelectShift. NFC. Small refactor to tidy up the code a little. llvm-svn: 217827
2014-09-16 05:27:56 +08:00
if (!isTypeSupported(I->getType(), VT, /*IsVectorAllowed=*/true))
[FastISel][AArch64] Add target-specific lowering for logical operations. This change adds support for immediate and shift-left folding into logical operations. This fixes rdar://problem/18223183. llvm-svn: 217118
2014-09-04 09:29:18 +08:00
return false;
[FastISel][AArch64] Refactor selectAddSub, selectLogicalOp, and SelectShift. NFC. Small refactor to tidy up the code a little. llvm-svn: 217827
2014-09-16 05:27:56 +08:00
if (VT.isVector())
return selectOperator(I, I->getOpcode());
unsigned ResultReg;
switch (I->getOpcode()) {
default:
llvm_unreachable("Unexpected instruction.");
case Instruction::And:
ResultReg = emitLogicalOp(ISD::AND, VT, I->getOperand(0), I->getOperand(1));
break;
case Instruction::Or:
ResultReg = emitLogicalOp(ISD::OR, VT, I->getOperand(0), I->getOperand(1));
break;
case Instruction::Xor:
ResultReg = emitLogicalOp(ISD::XOR, VT, I->getOperand(0), I->getOperand(1));
break;
}
[FastISel][AArch64] Add target-specific lowering for logical operations. This change adds support for immediate and shift-left folding into logical operations. This fixes rdar://problem/18223183. llvm-svn: 217118
2014-09-04 09:29:18 +08:00
if (!ResultReg)
return false;
updateValueMap(I, ResultReg);
return true;
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool AArch64FastISel::selectLoad(const Instruction *I) {
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
MVT VT;
// Verify we have a legal type before going any further. Currently, we handle
// simple types that will directly fit in a register (i32/f32/i64/f64) or
// those that can be sign or zero-extended to a basic operation (i1/i8/i16).
[FastISel][AArch64] Refactor code to use isTypeSupported. NFC. Gets rid of isLoadStoreTypeLegal and replace it with isTypeSupported. llvm-svn: 217826
2014-09-16 05:27:54 +08:00
if (!isTypeSupported(I->getType(), VT, /*IsVectorAllowed=*/true) ||
cast<LoadInst>(I)->isAtomic())
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
Swift Calling Convention: swifterror target support. Differential Revision: http://reviews.llvm.org/D18716 llvm-svn: 265997
2016-04-12 05:08:06 +08:00
const Value *SV = I->getOperand(0);
if (TLI.supportSwiftError()) {
// Swifterror values can come from either a function parameter with
// swifterror attribute or an alloca with swifterror attribute.
if (const Argument *Arg = dyn_cast<Argument>(SV)) {
if (Arg->hasSwiftErrorAttr())
return false;
}
if (const AllocaInst *Alloca = dyn_cast<AllocaInst>(SV)) {
if (Alloca->isSwiftError())
return false;
}
}
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// See if we can handle this address.
Address Addr;
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
if (!computeAddress(I->getOperand(0), Addr, I->getType()))
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
[FastISel][AArch64] Fix sign-/zero-extend folding when SelectionDAG is involved. Sign-/zero-extend folding depended on the load and the integer extend to be both selected by FastISel. This cannot always be garantueed and SelectionDAG might interfer. This commit adds additonal checks to load and integer extend lowering to catch this. Related to rdar://problem/18495928. llvm-svn: 219716
2014-10-15 04:36:02 +08:00
// Fold the following sign-/zero-extend into the load instruction.
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
bool WantZExt = true;
[FastISel][AArch64] Fix "Fold sign-/zero-extends into the load instruction." This commit fixes an issue with sign-/zero-extending loads that was discovered by Richard Barton. We use now the correct load instructions for sign-extending loads to 64bit. Also updated and added more unit tests. llvm-svn: 219185
2014-10-07 11:39:59 +08:00
MVT RetVT = VT;
[FastISel][AArch64] Fix sign-/zero-extend folding when SelectionDAG is involved. Sign-/zero-extend folding depended on the load and the integer extend to be both selected by FastISel. This cannot always be garantueed and SelectionDAG might interfer. This commit adds additonal checks to load and integer extend lowering to catch this. Related to rdar://problem/18495928. llvm-svn: 219716
2014-10-15 04:36:02 +08:00
const Value *IntExtVal = nullptr;
[FastISel][AArch64] Fix "Fold sign-/zero-extends into the load instruction." This commit fixes an issue with sign-/zero-extending loads that was discovered by Richard Barton. We use now the correct load instructions for sign-extending loads to 64bit. Also updated and added more unit tests. llvm-svn: 219185
2014-10-07 11:39:59 +08:00
if (I->hasOneUse()) {
if (const auto *ZE = dyn_cast<ZExtInst>(I->use_begin()->getUser())) {
[FastISel][AArch64] Fix sign-/zero-extend folding when SelectionDAG is involved. Sign-/zero-extend folding depended on the load and the integer extend to be both selected by FastISel. This cannot always be garantueed and SelectionDAG might interfer. This commit adds additonal checks to load and integer extend lowering to catch this. Related to rdar://problem/18495928. llvm-svn: 219716
2014-10-15 04:36:02 +08:00
if (isTypeSupported(ZE->getType(), RetVT))
IntExtVal = ZE;
else
[FastISel][AArch64] Fix "Fold sign-/zero-extends into the load instruction." This commit fixes an issue with sign-/zero-extending loads that was discovered by Richard Barton. We use now the correct load instructions for sign-extending loads to 64bit. Also updated and added more unit tests. llvm-svn: 219185
2014-10-07 11:39:59 +08:00
RetVT = VT;
} else if (const auto *SE = dyn_cast<SExtInst>(I->use_begin()->getUser())) {
[FastISel][AArch64] Fix sign-/zero-extend folding when SelectionDAG is involved. Sign-/zero-extend folding depended on the load and the integer extend to be both selected by FastISel. This cannot always be garantueed and SelectionDAG might interfer. This commit adds additonal checks to load and integer extend lowering to catch this. Related to rdar://problem/18495928. llvm-svn: 219716
2014-10-15 04:36:02 +08:00
if (isTypeSupported(SE->getType(), RetVT))
IntExtVal = SE;
else
[FastISel][AArch64] Fix "Fold sign-/zero-extends into the load instruction." This commit fixes an issue with sign-/zero-extending loads that was discovered by Richard Barton. We use now the correct load instructions for sign-extending loads to 64bit. Also updated and added more unit tests. llvm-svn: 219185
2014-10-07 11:39:59 +08:00
RetVT = VT;
WantZExt = false;
}
}
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
[FastISel][AArch64] Fix sign-/zero-extend folding when SelectionDAG is involved. Sign-/zero-extend folding depended on the load and the integer extend to be both selected by FastISel. This cannot always be garantueed and SelectionDAG might interfer. This commit adds additonal checks to load and integer extend lowering to catch this. Related to rdar://problem/18495928. llvm-svn: 219716
2014-10-15 04:36:02 +08:00
unsigned ResultReg =
emitLoad(VT, RetVT, Addr, WantZExt, createMachineMemOperandFor(I));
if (!ResultReg)
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
[FastISel][AArch64] Fix sign-/zero-extend folding when SelectionDAG is involved. Sign-/zero-extend folding depended on the load and the integer extend to be both selected by FastISel. This cannot always be garantueed and SelectionDAG might interfer. This commit adds additonal checks to load and integer extend lowering to catch this. Related to rdar://problem/18495928. llvm-svn: 219716
2014-10-15 04:36:02 +08:00
// There are a few different cases we have to handle, because the load or the
// sign-/zero-extend might not be selected by FastISel if we fall-back to
// SelectionDAG. There is also an ordering issue when both instructions are in
// different basic blocks.
// 1.) The load instruction is selected by FastISel, but the integer extend
// not. This usually happens when the integer extend is in a different
// basic block and SelectionDAG took over for that basic block.
// 2.) The load instruction is selected before the integer extend. This only
// happens when the integer extend is in a different basic block.
// 3.) The load instruction is selected by SelectionDAG and the integer extend
// by FastISel. This happens if there are instructions between the load
// and the integer extend that couldn't be selected by FastISel.
if (IntExtVal) {
// The integer extend hasn't been emitted yet. FastISel or SelectionDAG
// could select it. Emit a copy to subreg if necessary. FastISel will remove
// it when it selects the integer extend.
unsigned Reg = lookUpRegForValue(IntExtVal);
[AArch64][FastISel] Fix integer extend optimization. The integer extend optimization tries to fold the extend into the load instruction. This requires us to identify if the extend has already been emitted or not and act accordingly on it. The check that was originally performed for this was not sufficient. Besides checking the ValueMap for a mapped register we also need to check if the virtual register has already an associated machine instruction that defines it. This fixes rdar://problem/20470788. llvm-svn: 234529
2015-04-10 04:00:46 +08:00
auto *MI = MRI.getUniqueVRegDef(Reg);
if (!MI) {
[FastISel][AArch64] Fix sign-/zero-extend folding when SelectionDAG is involved. Sign-/zero-extend folding depended on the load and the integer extend to be both selected by FastISel. This cannot always be garantueed and SelectionDAG might interfer. This commit adds additonal checks to load and integer extend lowering to catch this. Related to rdar://problem/18495928. llvm-svn: 219716
2014-10-15 04:36:02 +08:00
if (RetVT == MVT::i64 && VT <= MVT::i32) {
if (WantZExt) {
// Delete the last emitted instruction from emitLoad (SUBREG_TO_REG).
std::prev(FuncInfo.InsertPt)->eraseFromParent();
ResultReg = std::prev(FuncInfo.InsertPt)->getOperand(0).getReg();
} else
ResultReg = fastEmitInst_extractsubreg(MVT::i32, ResultReg,
/*IsKill=*/true,
AArch64::sub_32);
}
updateValueMap(I, ResultReg);
return true;
}
// The integer extend has already been emitted - delete all the instructions
// that have been emitted by the integer extend lowering code and use the
// result from the load instruction directly.
[AArch64][FastISel] Fix integer extend optimization. The integer extend optimization tries to fold the extend into the load instruction. This requires us to identify if the extend has already been emitted or not and act accordingly on it. The check that was originally performed for this was not sufficient. Besides checking the ValueMap for a mapped register we also need to check if the virtual register has already an associated machine instruction that defines it. This fixes rdar://problem/20470788. llvm-svn: 234529
2015-04-10 04:00:46 +08:00
while (MI) {
[FastISel][AArch64] Fix sign-/zero-extend folding when SelectionDAG is involved. Sign-/zero-extend folding depended on the load and the integer extend to be both selected by FastISel. This cannot always be garantueed and SelectionDAG might interfer. This commit adds additonal checks to load and integer extend lowering to catch this. Related to rdar://problem/18495928. llvm-svn: 219716
2014-10-15 04:36:02 +08:00
Reg = 0;
for (auto &Opnd : MI->uses()) {
if (Opnd.isReg()) {
Reg = Opnd.getReg();
break;
}
}
MI->eraseFromParent();
[AArch64][FastISel] Fix integer extend optimization. The integer extend optimization tries to fold the extend into the load instruction. This requires us to identify if the extend has already been emitted or not and act accordingly on it. The check that was originally performed for this was not sufficient. Besides checking the ValueMap for a mapped register we also need to check if the virtual register has already an associated machine instruction that defines it. This fixes rdar://problem/20470788. llvm-svn: 234529
2015-04-10 04:00:46 +08:00
MI = nullptr;
if (Reg)
MI = MRI.getUniqueVRegDef(Reg);
[FastISel][AArch64] Fix sign-/zero-extend folding when SelectionDAG is involved. Sign-/zero-extend folding depended on the load and the integer extend to be both selected by FastISel. This cannot always be garantueed and SelectionDAG might interfer. This commit adds additonal checks to load and integer extend lowering to catch this. Related to rdar://problem/18495928. llvm-svn: 219716
2014-10-15 04:36:02 +08:00
}
updateValueMap(IntExtVal, ResultReg);
return true;
}
[FastISel] Rename public visible FastISel functions. NFC. This commit renames the following public FastISel functions: LowerArguments -> lowerArguments SelectInstruction -> selectInstruction TargetSelectInstruction -> fastSelectInstruction FastLowerArguments -> fastLowerArguments FastLowerCall -> fastLowerCall FastLowerIntrinsicCall -> fastLowerIntrinsicCall FastEmitZExtFromI1 -> fastEmitZExtFromI1 FastEmitBranch -> fastEmitBranch UpdateValueMap -> updateValueMap TargetMaterializeConstant -> fastMaterializeConstant TargetMaterializeAlloca -> fastMaterializeAlloca TargetMaterializeFloatZero -> fastMaterializeFloatZero LowerCallTo -> lowerCallTo Reviewed by Eric llvm-svn: 217074
2014-09-04 04:56:52 +08:00
updateValueMap(I, ResultReg);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return true;
}
[AArch64][FastISel] Select atomic stores into STLR. llvm-svn: 276182
2016-07-21 05:12:27 +08:00
bool AArch64FastISel::emitStoreRelease(MVT VT, unsigned SrcReg,
unsigned AddrReg,
MachineMemOperand *MMO) {
unsigned Opc;
switch (VT.SimpleTy) {
default: return false;
case MVT::i8: Opc = AArch64::STLRB; break;
case MVT::i16: Opc = AArch64::STLRH; break;
case MVT::i32: Opc = AArch64::STLRW; break;
case MVT::i64: Opc = AArch64::STLRX; break;
}
const MCInstrDesc &II = TII.get(Opc);
SrcReg = constrainOperandRegClass(II, SrcReg, 0);
AddrReg = constrainOperandRegClass(II, AddrReg, 1);
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, II)
.addReg(SrcReg)
.addReg(AddrReg)
.addMemOperand(MMO);
return true;
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool AArch64FastISel::emitStore(MVT VT, unsigned SrcReg, Address Addr,
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
MachineMemOperand *MMO) {
fix formatting; NFC llvm-svn: 241175
2015-07-02 01:58:53 +08:00
if (!TLI.allowsMisalignedMemoryAccesses(VT))
On behalf of Alexandros Lamprineas: LLVM targeting aarch64 doesn't correctly produce aligned accesses for non-aligned data at -O0/fast-isel (-mno-unaligned-access). The root cause seems to be in fast-isel not producing unaligned access correctly for -mno-unaligned-access. The patch just aborts fast-isel for loads and stores when -mno-unaligned-access is present. The regression test is updated to check this new test case (-mno-unaligned-access together with fast-isel). Differential Revision: http://reviews.llvm.org/D10360 llvm-svn: 239732
2015-06-15 23:48:44 +08:00
return false;
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
// Simplify this down to something we can handle.
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
if (!simplifyAddress(Addr, VT))
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
return false;
[FastISel][AArch64] Factor out scale factor calculation. NFC. Factor out the code that determines the implicit scale factor of memory operations for a given value type. llvm-svn: 218652
2014-09-30 08:49:54 +08:00
unsigned ScaleFactor = getImplicitScaleFactor(VT);
if (!ScaleFactor)
llvm_unreachable("Unexpected value type.");
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Negative offsets require unscaled, 9-bit, signed immediate offsets.
// Otherwise, we try using scaled, 12-bit, unsigned immediate offsets.
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
bool UseScaled = true;
if ((Addr.getOffset() < 0) || (Addr.getOffset() & (ScaleFactor - 1))) {
UseScaled = false;
ScaleFactor = 1;
}
static const unsigned OpcTable[4][6] = {
{ AArch64::STURBBi, AArch64::STURHHi, AArch64::STURWi, AArch64::STURXi,
AArch64::STURSi, AArch64::STURDi },
{ AArch64::STRBBui, AArch64::STRHHui, AArch64::STRWui, AArch64::STRXui,
AArch64::STRSui, AArch64::STRDui },
{ AArch64::STRBBroX, AArch64::STRHHroX, AArch64::STRWroX, AArch64::STRXroX,
AArch64::STRSroX, AArch64::STRDroX },
{ AArch64::STRBBroW, AArch64::STRHHroW, AArch64::STRWroW, AArch64::STRXroW,
AArch64::STRSroW, AArch64::STRDroW }
};
unsigned Opc;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
bool VTIsi1 = false;
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
bool UseRegOffset = Addr.isRegBase() && !Addr.getOffset() && Addr.getReg() &&
Addr.getOffsetReg();
unsigned Idx = UseRegOffset ? 2 : UseScaled ? 1 : 0;
if (Addr.getExtendType() == AArch64_AM::UXTW ||
Addr.getExtendType() == AArch64_AM::SXTW)
Idx++;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
switch (VT.SimpleTy) {
Reapply [FastISel][AArch64] Add support for more addressing modes (r215597). Note: This was originally reverted to track down a buildbot error. Reapply without any modifications. Original commit message: FastISel didn't take much advantage of the different addressing modes available to it on AArch64. This commit allows the ComputeAddress method to recognize more addressing modes that allows shifts and sign-/zero-extensions to be folded into the memory operation itself. For Example: lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3] ldr x0, [x0, x1] sxtw x1, w1 lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3] ldr x0, [x0, x1] llvm-svn: 216013
2014-08-20 03:44:17 +08:00
default: llvm_unreachable("Unexpected value type.");
case MVT::i1: VTIsi1 = true;
case MVT::i8: Opc = OpcTable[Idx][0]; break;
case MVT::i16: Opc = OpcTable[Idx][1]; break;
case MVT::i32: Opc = OpcTable[Idx][2]; break;
case MVT::i64: Opc = OpcTable[Idx][3]; break;
case MVT::f32: Opc = OpcTable[Idx][4]; break;
case MVT::f64: Opc = OpcTable[Idx][5]; break;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
}
// Storing an i1 requires special handling.
[FastISel][AArch64] Use the zero register for stores. Use the zero register directly when possible to avoid an unnecessary register copy and a wasted register at -O0. This also uses integer stores to store a positive floating-point zero. This saves us from materializing the positive zero in a register and then storing it. llvm-svn: 216617
2014-08-28 05:04:52 +08:00
if (VTIsi1 && SrcReg != AArch64::WZR) {
[FastISel][AArch64] Add target-specific lowering for logical operations. This change adds support for immediate and shift-left folding into logical operations. This fixes rdar://problem/18223183. llvm-svn: 217118
2014-09-04 09:29:18 +08:00
unsigned ANDReg = emitAnd_ri(MVT::i32, SrcReg, /*TODO:IsKill=*/false, 1);
[FastISel][AArch64] Factor out ANDWri instruction generation into a helper function. NFCI. llvm-svn: 216199
2014-08-22 02:02:25 +08:00
assert(ANDReg && "Unexpected AND instruction emission failure.");
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
SrcReg = ANDReg;
}
// Create the base instruction, then add the operands.
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
const MCInstrDesc &II = TII.get(Opc);
SrcReg = constrainOperandRegClass(II, SrcReg, II.getNumDefs());
MachineInstrBuilder MIB =
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, II).addReg(SrcReg);
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
addLoadStoreOperands(Addr, MIB, MachineMemOperand::MOStore, ScaleFactor, MMO);
[FastISel][AArch64] Attach MachineMemOperands to load and store instructions. llvm-svn: 215231
2014-08-09 01:24:10 +08:00
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return true;
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool AArch64FastISel::selectStore(const Instruction *I) {
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
MVT VT;
[FastISel][AArch64] Use the zero register for stores. Use the zero register directly when possible to avoid an unnecessary register copy and a wasted register at -O0. This also uses integer stores to store a positive floating-point zero. This saves us from materializing the positive zero in a register and then storing it. llvm-svn: 216617
2014-08-28 05:04:52 +08:00
const Value *Op0 = I->getOperand(0);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Verify we have a legal type before going any further. Currently, we handle
// simple types that will directly fit in a register (i32/f32/i64/f64) or
// those that can be sign or zero-extended to a basic operation (i1/i8/i16).
[AArch64][FastISel] Select atomic stores into STLR. llvm-svn: 276182
2016-07-21 05:12:27 +08:00
if (!isTypeSupported(Op0->getType(), VT, /*IsVectorAllowed=*/true))
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
Swift Calling Convention: swifterror target support. Differential Revision: http://reviews.llvm.org/D18716 llvm-svn: 265997
2016-04-12 05:08:06 +08:00
const Value *PtrV = I->getOperand(1);
if (TLI.supportSwiftError()) {
// Swifterror values can come from either a function parameter with
// swifterror attribute or an alloca with swifterror attribute.
if (const Argument *Arg = dyn_cast<Argument>(PtrV)) {
if (Arg->hasSwiftErrorAttr())
return false;
}
if (const AllocaInst *Alloca = dyn_cast<AllocaInst>(PtrV)) {
if (Alloca->isSwiftError())
return false;
}
}
[FastISel][AArch64] Use the zero register for stores. Use the zero register directly when possible to avoid an unnecessary register copy and a wasted register at -O0. This also uses integer stores to store a positive floating-point zero. This saves us from materializing the positive zero in a register and then storing it. llvm-svn: 216617
2014-08-28 05:04:52 +08:00
// Get the value to be stored into a register. Use the zero register directly
[FastISel][AArch64] Fix a comment in my previous commit (r216617). llvm-svn: 216622
2014-08-28 05:40:50 +08:00
// when possible to avoid an unnecessary copy and a wasted register.
[FastISel][AArch64] Use the zero register for stores. Use the zero register directly when possible to avoid an unnecessary register copy and a wasted register at -O0. This also uses integer stores to store a positive floating-point zero. This saves us from materializing the positive zero in a register and then storing it. llvm-svn: 216617
2014-08-28 05:04:52 +08:00
unsigned SrcReg = 0;
if (const auto *CI = dyn_cast<ConstantInt>(Op0)) {
if (CI->isZero())
SrcReg = (VT == MVT::i64) ? AArch64::XZR : AArch64::WZR;
} else if (const auto *CF = dyn_cast<ConstantFP>(Op0)) {
if (CF->isZero() && !CF->isNegative()) {
VT = MVT::getIntegerVT(VT.getSizeInBits());
SrcReg = (VT == MVT::i64) ? AArch64::XZR : AArch64::WZR;
}
}
if (!SrcReg)
SrcReg = getRegForValue(Op0);
if (!SrcReg)
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
[AArch64][FastISel] Select atomic stores into STLR. llvm-svn: 276182
2016-07-21 05:12:27 +08:00
auto *SI = cast<StoreInst>(I);
// Try to emit a STLR for seq_cst/release.
if (SI->isAtomic()) {
AtomicOrdering Ord = SI->getOrdering();
// The non-atomic instructions are sufficient for relaxed stores.
if (isReleaseOrStronger(Ord)) {
// The STLR addressing mode only supports a base reg; pass that directly.
unsigned AddrReg = getRegForValue(PtrV);
return emitStoreRelease(VT, SrcReg, AddrReg,
createMachineMemOperandFor(I));
}
}
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// See if we can handle this address.
Address Addr;
[AArch64][FastISel] Select atomic stores into STLR. llvm-svn: 276182
2016-07-21 05:12:27 +08:00
if (!computeAddress(PtrV, Addr, Op0->getType()))
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
if (!emitStore(VT, SrcReg, Addr, createMachineMemOperandFor(I)))
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
return true;
}
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
static AArch64CC::CondCode getCompareCC(CmpInst::Predicate Pred) {
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
switch (Pred) {
case CmpInst::FCMP_ONE:
case CmpInst::FCMP_UEQ:
default:
// AL is our "false" for now. The other two need more compares.
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
return AArch64CC::AL;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
case CmpInst::ICMP_EQ:
case CmpInst::FCMP_OEQ:
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
return AArch64CC::EQ;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
case CmpInst::ICMP_SGT:
case CmpInst::FCMP_OGT:
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
return AArch64CC::GT;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
case CmpInst::ICMP_SGE:
case CmpInst::FCMP_OGE:
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
return AArch64CC::GE;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
case CmpInst::ICMP_UGT:
case CmpInst::FCMP_UGT:
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
return AArch64CC::HI;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
case CmpInst::FCMP_OLT:
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
return AArch64CC::MI;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
case CmpInst::ICMP_ULE:
case CmpInst::FCMP_OLE:
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
return AArch64CC::LS;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
case CmpInst::FCMP_ORD:
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
return AArch64CC::VC;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
case CmpInst::FCMP_UNO:
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
return AArch64CC::VS;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
case CmpInst::FCMP_UGE:
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
return AArch64CC::PL;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
case CmpInst::ICMP_SLT:
case CmpInst::FCMP_ULT:
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
return AArch64CC::LT;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
case CmpInst::ICMP_SLE:
case CmpInst::FCMP_ULE:
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
return AArch64CC::LE;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
case CmpInst::FCMP_UNE:
case CmpInst::ICMP_NE:
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
return AArch64CC::NE;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
case CmpInst::ICMP_UGE:
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
return AArch64CC::HS;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
case CmpInst::ICMP_ULT:
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
return AArch64CC::LO;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
}
}
Recommit r218010 [FastISel][AArch64] Fold bit test and branch into TBZ and TBNZ. Note: This version fixed an issue with the TBZ/TBNZ instructions that were generated in FastISel. The issue was that the 64bit version of TBZ (TBZX) automagically sets the upper bit of the immediate field that is used to specify the bit we want to test. To test for any of the lower 32bits we have to first extract the subregister and use the 32bit version of the TBZ instruction (TBZW). Original commit message: Teach selectBranch to fold bit test and branch into a single instruction (TBZ or TBNZ). llvm-svn: 218693
2014-10-01 03:59:35 +08:00
/// \brief Try to emit a combined compare-and-branch instruction.
bool AArch64FastISel::emitCompareAndBranch(const BranchInst *BI) {
assert(isa<CmpInst>(BI->getCondition()) && "Expected cmp instruction");
const CmpInst *CI = cast<CmpInst>(BI->getCondition());
CmpInst::Predicate Predicate = optimizeCmpPredicate(CI);
[FastISel][AArch64] Fold compare with zero and branch into CBZ and CBNZ. This takes advanatage of the CBZ and CBNZ instruction to further optimize the common null check pattern into a single instruction. This is related to rdar://problem/18358882. llvm-svn: 217972
2014-09-18 02:05:34 +08:00
Recommit r218010 [FastISel][AArch64] Fold bit test and branch into TBZ and TBNZ. Note: This version fixed an issue with the TBZ/TBNZ instructions that were generated in FastISel. The issue was that the 64bit version of TBZ (TBZX) automagically sets the upper bit of the immediate field that is used to specify the bit we want to test. To test for any of the lower 32bits we have to first extract the subregister and use the 32bit version of the TBZ instruction (TBZW). Original commit message: Teach selectBranch to fold bit test and branch into a single instruction (TBZ or TBNZ). llvm-svn: 218693
2014-10-01 03:59:35 +08:00
const Value *LHS = CI->getOperand(0);
const Value *RHS = CI->getOperand(1);
[FastISel][AArch64] Use 'cbz' also for null values (pointers). The pattern matching for a 'ConstantInt' value was too restrictive. Checking for a 'Constant' with a bull value is sufficient for using an 'cbz/cbnz' instruction. This fixes rdar://problem/18784732. llvm-svn: 220709
2014-10-28 03:38:05 +08:00
MVT VT;
if (!isTypeSupported(LHS->getType(), VT))
return false;
[FastISel][AArch64] Fold compare with zero and branch into CBZ and CBNZ. This takes advanatage of the CBZ and CBNZ instruction to further optimize the common null check pattern into a single instruction. This is related to rdar://problem/18358882. llvm-svn: 217972
2014-09-18 02:05:34 +08:00
[FastISel][AArch64] Use 'cbz' also for null values (pointers). The pattern matching for a 'ConstantInt' value was too restrictive. Checking for a 'Constant' with a bull value is sufficient for using an 'cbz/cbnz' instruction. This fixes rdar://problem/18784732. llvm-svn: 220709
2014-10-28 03:38:05 +08:00
unsigned BW = VT.getSizeInBits();
if (BW > 64)
[FastISel][AArch64] Fold compare with zero and branch into CBZ and CBNZ. This takes advanatage of the CBZ and CBNZ instruction to further optimize the common null check pattern into a single instruction. This is related to rdar://problem/18358882. llvm-svn: 217972
2014-09-18 02:05:34 +08:00
return false;
Recommit r218010 [FastISel][AArch64] Fold bit test and branch into TBZ and TBNZ. Note: This version fixed an issue with the TBZ/TBNZ instructions that were generated in FastISel. The issue was that the 64bit version of TBZ (TBZX) automagically sets the upper bit of the immediate field that is used to specify the bit we want to test. To test for any of the lower 32bits we have to first extract the subregister and use the 32bit version of the TBZ instruction (TBZW). Original commit message: Teach selectBranch to fold bit test and branch into a single instruction (TBZ or TBNZ). llvm-svn: 218693
2014-10-01 03:59:35 +08:00
MachineBasicBlock *TBB = FuncInfo.MBBMap[BI->getSuccessor(0)];
MachineBasicBlock *FBB = FuncInfo.MBBMap[BI->getSuccessor(1)];
[FastISel][AArch64] Fold compare with zero and branch into CBZ and CBNZ. This takes advanatage of the CBZ and CBNZ instruction to further optimize the common null check pattern into a single instruction. This is related to rdar://problem/18358882. llvm-svn: 217972
2014-09-18 02:05:34 +08:00
Recommit r218010 [FastISel][AArch64] Fold bit test and branch into TBZ and TBNZ. Note: This version fixed an issue with the TBZ/TBNZ instructions that were generated in FastISel. The issue was that the 64bit version of TBZ (TBZX) automagically sets the upper bit of the immediate field that is used to specify the bit we want to test. To test for any of the lower 32bits we have to first extract the subregister and use the 32bit version of the TBZ instruction (TBZW). Original commit message: Teach selectBranch to fold bit test and branch into a single instruction (TBZ or TBNZ). llvm-svn: 218693
2014-10-01 03:59:35 +08:00
// Try to take advantage of fallthrough opportunities.
if (FuncInfo.MBB->isLayoutSuccessor(TBB)) {
std::swap(TBB, FBB);
Predicate = CmpInst::getInversePredicate(Predicate);
}
[FastISel][AArch64] Fold compare with zero and branch into CBZ and CBNZ. This takes advanatage of the CBZ and CBNZ instruction to further optimize the common null check pattern into a single instruction. This is related to rdar://problem/18358882. llvm-svn: 217972
2014-09-18 02:05:34 +08:00
Recommit r218010 [FastISel][AArch64] Fold bit test and branch into TBZ and TBNZ. Note: This version fixed an issue with the TBZ/TBNZ instructions that were generated in FastISel. The issue was that the 64bit version of TBZ (TBZX) automagically sets the upper bit of the immediate field that is used to specify the bit we want to test. To test for any of the lower 32bits we have to first extract the subregister and use the 32bit version of the TBZ instruction (TBZW). Original commit message: Teach selectBranch to fold bit test and branch into a single instruction (TBZ or TBNZ). llvm-svn: 218693
2014-10-01 03:59:35 +08:00
int TestBit = -1;
bool IsCmpNE;
[FastISel][AArch64] Fix and extend the tbz/tbnz pattern matching. The pattern matching failed to recognize all instances of "-1", because when comparing against "-1" we didn't use an APInt of the same bitwidth. This commit fixes this and also adds inverse versions of the conditon to catch more cases. llvm-svn: 222722
2014-11-25 12:16:15 +08:00
switch (Predicate) {
default:
return false;
case CmpInst::ICMP_EQ:
case CmpInst::ICMP_NE:
if (isa<Constant>(LHS) && cast<Constant>(LHS)->isNullValue())
std::swap(LHS, RHS);
Recommit r218010 [FastISel][AArch64] Fold bit test and branch into TBZ and TBNZ. Note: This version fixed an issue with the TBZ/TBNZ instructions that were generated in FastISel. The issue was that the 64bit version of TBZ (TBZX) automagically sets the upper bit of the immediate field that is used to specify the bit we want to test. To test for any of the lower 32bits we have to first extract the subregister and use the 32bit version of the TBZ instruction (TBZW). Original commit message: Teach selectBranch to fold bit test and branch into a single instruction (TBZ or TBNZ). llvm-svn: 218693
2014-10-01 03:59:35 +08:00
[FastISel][AArch64] Fix and extend the tbz/tbnz pattern matching. The pattern matching failed to recognize all instances of "-1", because when comparing against "-1" we didn't use an APInt of the same bitwidth. This commit fixes this and also adds inverse versions of the conditon to catch more cases. llvm-svn: 222722
2014-11-25 12:16:15 +08:00
if (!isa<Constant>(RHS) || !cast<Constant>(RHS)->isNullValue())
Recommit r218010 [FastISel][AArch64] Fold bit test and branch into TBZ and TBNZ. Note: This version fixed an issue with the TBZ/TBNZ instructions that were generated in FastISel. The issue was that the 64bit version of TBZ (TBZX) automagically sets the upper bit of the immediate field that is used to specify the bit we want to test. To test for any of the lower 32bits we have to first extract the subregister and use the 32bit version of the TBZ instruction (TBZW). Original commit message: Teach selectBranch to fold bit test and branch into a single instruction (TBZ or TBNZ). llvm-svn: 218693
2014-10-01 03:59:35 +08:00
return false;
if (const auto *AI = dyn_cast<BinaryOperator>(LHS))
[FastISel][AArch64] Don't fold the 'and' instruction into the 'tbz/tbnz' instruction if it is in a different basic block. This fixes a bug where the input register was not defined for the 'tbz/tbnz' instruction. This happened, because we folded the 'and' instruction from a different basic block. This fixes rdar://problem/18784013. llvm-svn: 220704
2014-10-28 03:16:48 +08:00
if (AI->getOpcode() == Instruction::And && isValueAvailable(AI)) {
Recommit r218010 [FastISel][AArch64] Fold bit test and branch into TBZ and TBNZ. Note: This version fixed an issue with the TBZ/TBNZ instructions that were generated in FastISel. The issue was that the 64bit version of TBZ (TBZX) automagically sets the upper bit of the immediate field that is used to specify the bit we want to test. To test for any of the lower 32bits we have to first extract the subregister and use the 32bit version of the TBZ instruction (TBZW). Original commit message: Teach selectBranch to fold bit test and branch into a single instruction (TBZ or TBNZ). llvm-svn: 218693
2014-10-01 03:59:35 +08:00
const Value *AndLHS = AI->getOperand(0);
const Value *AndRHS = AI->getOperand(1);
if (const auto *C = dyn_cast<ConstantInt>(AndLHS))
if (C->getValue().isPowerOf2())
std::swap(AndLHS, AndRHS);
if (const auto *C = dyn_cast<ConstantInt>(AndRHS))
if (C->getValue().isPowerOf2()) {
TestBit = C->getValue().logBase2();
LHS = AndLHS;
}
}
[FastISel][AArch64] Optimize compare-and-branch for i1 to use 'tbz'. Minor enhancement to use 'tbz' for i1 compare-and-branch to get rid of an 'and' instruction. This fixes rdar://problem/18784953. llvm-svn: 220712
2014-10-28 03:46:23 +08:00
if (VT == MVT::i1)
TestBit = 0;
Recommit r218010 [FastISel][AArch64] Fold bit test and branch into TBZ and TBNZ. Note: This version fixed an issue with the TBZ/TBNZ instructions that were generated in FastISel. The issue was that the 64bit version of TBZ (TBZX) automagically sets the upper bit of the immediate field that is used to specify the bit we want to test. To test for any of the lower 32bits we have to first extract the subregister and use the 32bit version of the TBZ instruction (TBZW). Original commit message: Teach selectBranch to fold bit test and branch into a single instruction (TBZ or TBNZ). llvm-svn: 218693
2014-10-01 03:59:35 +08:00
IsCmpNE = Predicate == CmpInst::ICMP_NE;
[FastISel][AArch64] Fix and extend the tbz/tbnz pattern matching. The pattern matching failed to recognize all instances of "-1", because when comparing against "-1" we didn't use an APInt of the same bitwidth. This commit fixes this and also adds inverse versions of the conditon to catch more cases. llvm-svn: 222722
2014-11-25 12:16:15 +08:00
break;
case CmpInst::ICMP_SLT:
case CmpInst::ICMP_SGE:
if (!isa<Constant>(RHS) || !cast<Constant>(RHS)->isNullValue())
Recommit r218010 [FastISel][AArch64] Fold bit test and branch into TBZ and TBNZ. Note: This version fixed an issue with the TBZ/TBNZ instructions that were generated in FastISel. The issue was that the 64bit version of TBZ (TBZX) automagically sets the upper bit of the immediate field that is used to specify the bit we want to test. To test for any of the lower 32bits we have to first extract the subregister and use the 32bit version of the TBZ instruction (TBZW). Original commit message: Teach selectBranch to fold bit test and branch into a single instruction (TBZ or TBNZ). llvm-svn: 218693
2014-10-01 03:59:35 +08:00
return false;
TestBit = BW - 1;
[FastISel][AArch64] Fix and extend the tbz/tbnz pattern matching. The pattern matching failed to recognize all instances of "-1", because when comparing against "-1" we didn't use an APInt of the same bitwidth. This commit fixes this and also adds inverse versions of the conditon to catch more cases. llvm-svn: 222722
2014-11-25 12:16:15 +08:00
IsCmpNE = Predicate == CmpInst::ICMP_SLT;
break;
case CmpInst::ICMP_SGT:
case CmpInst::ICMP_SLE:
Recommit r218010 [FastISel][AArch64] Fold bit test and branch into TBZ and TBNZ. Note: This version fixed an issue with the TBZ/TBNZ instructions that were generated in FastISel. The issue was that the 64bit version of TBZ (TBZX) automagically sets the upper bit of the immediate field that is used to specify the bit we want to test. To test for any of the lower 32bits we have to first extract the subregister and use the 32bit version of the TBZ instruction (TBZW). Original commit message: Teach selectBranch to fold bit test and branch into a single instruction (TBZ or TBNZ). llvm-svn: 218693
2014-10-01 03:59:35 +08:00
if (!isa<ConstantInt>(RHS))
return false;
[FastISel][AArch64] Fix and extend the tbz/tbnz pattern matching. The pattern matching failed to recognize all instances of "-1", because when comparing against "-1" we didn't use an APInt of the same bitwidth. This commit fixes this and also adds inverse versions of the conditon to catch more cases. llvm-svn: 222722
2014-11-25 12:16:15 +08:00
if (cast<ConstantInt>(RHS)->getValue() != APInt(BW, -1, true))
Recommit r218010 [FastISel][AArch64] Fold bit test and branch into TBZ and TBNZ. Note: This version fixed an issue with the TBZ/TBNZ instructions that were generated in FastISel. The issue was that the 64bit version of TBZ (TBZX) automagically sets the upper bit of the immediate field that is used to specify the bit we want to test. To test for any of the lower 32bits we have to first extract the subregister and use the 32bit version of the TBZ instruction (TBZW). Original commit message: Teach selectBranch to fold bit test and branch into a single instruction (TBZ or TBNZ). llvm-svn: 218693
2014-10-01 03:59:35 +08:00
return false;
TestBit = BW - 1;
[FastISel][AArch64] Fix and extend the tbz/tbnz pattern matching. The pattern matching failed to recognize all instances of "-1", because when comparing against "-1" we didn't use an APInt of the same bitwidth. This commit fixes this and also adds inverse versions of the conditon to catch more cases. llvm-svn: 222722
2014-11-25 12:16:15 +08:00
IsCmpNE = Predicate == CmpInst::ICMP_SLE;
break;
} // end switch
Recommit r218010 [FastISel][AArch64] Fold bit test and branch into TBZ and TBNZ. Note: This version fixed an issue with the TBZ/TBNZ instructions that were generated in FastISel. The issue was that the 64bit version of TBZ (TBZX) automagically sets the upper bit of the immediate field that is used to specify the bit we want to test. To test for any of the lower 32bits we have to first extract the subregister and use the 32bit version of the TBZ instruction (TBZW). Original commit message: Teach selectBranch to fold bit test and branch into a single instruction (TBZ or TBNZ). llvm-svn: 218693
2014-10-01 03:59:35 +08:00
static const unsigned OpcTable[2][2][2] = {
{ {AArch64::CBZW, AArch64::CBZX },
{AArch64::CBNZW, AArch64::CBNZX} },
{ {AArch64::TBZW, AArch64::TBZX },
{AArch64::TBNZW, AArch64::TBNZX} }
};
bool IsBitTest = TestBit != -1;
bool Is64Bit = BW == 64;
if (TestBit < 32 && TestBit >= 0)
Is64Bit = false;
[FastISel][AArch64] Don't fold the 'and' instruction into the 'tbz/tbnz' instruction if it is in a different basic block. This fixes a bug where the input register was not defined for the 'tbz/tbnz' instruction. This happened, because we folded the 'and' instruction from a different basic block. This fixes rdar://problem/18784013. llvm-svn: 220704
2014-10-28 03:16:48 +08:00
Recommit r218010 [FastISel][AArch64] Fold bit test and branch into TBZ and TBNZ. Note: This version fixed an issue with the TBZ/TBNZ instructions that were generated in FastISel. The issue was that the 64bit version of TBZ (TBZX) automagically sets the upper bit of the immediate field that is used to specify the bit we want to test. To test for any of the lower 32bits we have to first extract the subregister and use the 32bit version of the TBZ instruction (TBZW). Original commit message: Teach selectBranch to fold bit test and branch into a single instruction (TBZ or TBNZ). llvm-svn: 218693
2014-10-01 03:59:35 +08:00
unsigned Opc = OpcTable[IsBitTest][IsCmpNE][Is64Bit];
const MCInstrDesc &II = TII.get(Opc);
unsigned SrcReg = getRegForValue(LHS);
if (!SrcReg)
return false;
bool SrcIsKill = hasTrivialKill(LHS);
[FastISel][AArch64] Fix sign-/zero-extend folding when SelectionDAG is involved. Sign-/zero-extend folding depended on the load and the integer extend to be both selected by FastISel. This cannot always be garantueed and SelectionDAG might interfer. This commit adds additonal checks to load and integer extend lowering to catch this. Related to rdar://problem/18495928. llvm-svn: 219716
2014-10-15 04:36:02 +08:00
if (BW == 64 && !Is64Bit)
Recommit r218010 [FastISel][AArch64] Fold bit test and branch into TBZ and TBNZ. Note: This version fixed an issue with the TBZ/TBNZ instructions that were generated in FastISel. The issue was that the 64bit version of TBZ (TBZX) automagically sets the upper bit of the immediate field that is used to specify the bit we want to test. To test for any of the lower 32bits we have to first extract the subregister and use the 32bit version of the TBZ instruction (TBZW). Original commit message: Teach selectBranch to fold bit test and branch into a single instruction (TBZ or TBNZ). llvm-svn: 218693
2014-10-01 03:59:35 +08:00
SrcReg = fastEmitInst_extractsubreg(MVT::i32, SrcReg, SrcIsKill,
AArch64::sub_32);
[FastISel][AArch64] Use 'cbz' also for null values (pointers). The pattern matching for a 'ConstantInt' value was too restrictive. Checking for a 'Constant' with a bull value is sufficient for using an 'cbz/cbnz' instruction. This fixes rdar://problem/18784732. llvm-svn: 220709
2014-10-28 03:38:05 +08:00
if ((BW < 32) && !IsBitTest)
SrcReg = emitIntExt(VT, SrcReg, MVT::i32, /*IsZExt=*/true);
[AArch64] Fix fast-isel of cbz of i1, i8, i16 This fixes a miscompilation in the AArch64 fast-isel which was triggered when a branch is based on an icmp with condition eq or ne, and type i1, i8 or i16. The cbz instruction compares the whole 32-bit register, so values with the bottom 1, 8 or 16 bits clear would cause the wrong branch to be taken. llvm-svn: 220553
2014-10-24 17:54:41 +08:00
Recommit r218010 [FastISel][AArch64] Fold bit test and branch into TBZ and TBNZ. Note: This version fixed an issue with the TBZ/TBNZ instructions that were generated in FastISel. The issue was that the 64bit version of TBZ (TBZX) automagically sets the upper bit of the immediate field that is used to specify the bit we want to test. To test for any of the lower 32bits we have to first extract the subregister and use the 32bit version of the TBZ instruction (TBZW). Original commit message: Teach selectBranch to fold bit test and branch into a single instruction (TBZ or TBNZ). llvm-svn: 218693
2014-10-01 03:59:35 +08:00
// Emit the combined compare and branch instruction.
[FastISel][AArch64] Fix sign-/zero-extend folding when SelectionDAG is involved. Sign-/zero-extend folding depended on the load and the integer extend to be both selected by FastISel. This cannot always be garantueed and SelectionDAG might interfer. This commit adds additonal checks to load and integer extend lowering to catch this. Related to rdar://problem/18495928. llvm-svn: 219716
2014-10-15 04:36:02 +08:00
SrcReg = constrainOperandRegClass(II, SrcReg, II.getNumDefs());
Recommit r218010 [FastISel][AArch64] Fold bit test and branch into TBZ and TBNZ. Note: This version fixed an issue with the TBZ/TBNZ instructions that were generated in FastISel. The issue was that the 64bit version of TBZ (TBZX) automagically sets the upper bit of the immediate field that is used to specify the bit we want to test. To test for any of the lower 32bits we have to first extract the subregister and use the 32bit version of the TBZ instruction (TBZW). Original commit message: Teach selectBranch to fold bit test and branch into a single instruction (TBZ or TBNZ). llvm-svn: 218693
2014-10-01 03:59:35 +08:00
MachineInstrBuilder MIB =
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(Opc))
.addReg(SrcReg, getKillRegState(SrcIsKill));
if (IsBitTest)
MIB.addImm(TestBit);
MIB.addMBB(TBB);
FastISel: Factor out common code; NFC intended This should be no functional change but for the record: For three cases in X86FastISel this will change the order in which the FalseMBB and TrueMBB of a conditional branch is addedd to the successor/predecessor lists. llvm-svn: 245997
2015-08-26 09:38:00 +08:00
finishCondBranch(BI->getParent(), TBB, FBB);
Recommit r218010 [FastISel][AArch64] Fold bit test and branch into TBZ and TBNZ. Note: This version fixed an issue with the TBZ/TBNZ instructions that were generated in FastISel. The issue was that the 64bit version of TBZ (TBZX) automagically sets the upper bit of the immediate field that is used to specify the bit we want to test. To test for any of the lower 32bits we have to first extract the subregister and use the 32bit version of the TBZ instruction (TBZW). Original commit message: Teach selectBranch to fold bit test and branch into a single instruction (TBZ or TBNZ). llvm-svn: 218693
2014-10-01 03:59:35 +08:00
return true;
[FastISel][AArch64] Fold compare with zero and branch into CBZ and CBNZ. This takes advanatage of the CBZ and CBNZ instruction to further optimize the common null check pattern into a single instruction. This is related to rdar://problem/18358882. llvm-svn: 217972
2014-09-18 02:05:34 +08:00
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool AArch64FastISel::selectBranch(const Instruction *I) {
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
const BranchInst *BI = cast<BranchInst>(I);
[FastISel][AArch64] Move unconditional branch handling into 'SelectBranch'. NFC. llvm-svn: 217054
2014-09-04 01:58:10 +08:00
if (BI->isUnconditional()) {
MachineBasicBlock *MSucc = FuncInfo.MBBMap[BI->getSuccessor(0)];
[FastISel] Rename public visible FastISel functions. NFC. This commit renames the following public FastISel functions: LowerArguments -> lowerArguments SelectInstruction -> selectInstruction TargetSelectInstruction -> fastSelectInstruction FastLowerArguments -> fastLowerArguments FastLowerCall -> fastLowerCall FastLowerIntrinsicCall -> fastLowerIntrinsicCall FastEmitZExtFromI1 -> fastEmitZExtFromI1 FastEmitBranch -> fastEmitBranch UpdateValueMap -> updateValueMap TargetMaterializeConstant -> fastMaterializeConstant TargetMaterializeAlloca -> fastMaterializeAlloca TargetMaterializeFloatZero -> fastMaterializeFloatZero LowerCallTo -> lowerCallTo Reviewed by Eric llvm-svn: 217074
2014-09-04 04:56:52 +08:00
fastEmitBranch(MSucc, BI->getDebugLoc());
[FastISel][AArch64] Move unconditional branch handling into 'SelectBranch'. NFC. llvm-svn: 217054
2014-09-04 01:58:10 +08:00
return true;
}
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
MachineBasicBlock *TBB = FuncInfo.MBBMap[BI->getSuccessor(0)];
MachineBasicBlock *FBB = FuncInfo.MBBMap[BI->getSuccessor(1)];
if (const CmpInst *CI = dyn_cast<CmpInst>(BI->getCondition())) {
[FastISel][AArch64] Improve branch selection to support all FP conditions. This adds the last two missing floating-point condition codes (FCMP_UEQ and FCMP_ONE) also to the branch selection. In these two cases an additonal branch instruction is required. This also adds unit tests to checks all the different condition codes. This is related o rdar://problem/18358882. llvm-svn: 217966
2014-09-18 01:46:47 +08:00
if (CI->hasOneUse() && isValueAvailable(CI)) {
// Try to optimize or fold the cmp.
CmpInst::Predicate Predicate = optimizeCmpPredicate(CI);
switch (Predicate) {
default:
break;
case CmpInst::FCMP_FALSE:
fastEmitBranch(FBB, DbgLoc);
return true;
case CmpInst::FCMP_TRUE:
fastEmitBranch(TBB, DbgLoc);
return true;
}
Recommit r218010 [FastISel][AArch64] Fold bit test and branch into TBZ and TBNZ. Note: This version fixed an issue with the TBZ/TBNZ instructions that were generated in FastISel. The issue was that the 64bit version of TBZ (TBZX) automagically sets the upper bit of the immediate field that is used to specify the bit we want to test. To test for any of the lower 32bits we have to first extract the subregister and use the 32bit version of the TBZ instruction (TBZW). Original commit message: Teach selectBranch to fold bit test and branch into a single instruction (TBZ or TBNZ). llvm-svn: 218693
2014-10-01 03:59:35 +08:00
// Try to emit a combined compare-and-branch first.
if (emitCompareAndBranch(BI))
return true;
[FastISel][AArch64] Improve branch selection to support all FP conditions. This adds the last two missing floating-point condition codes (FCMP_UEQ and FCMP_ONE) also to the branch selection. In these two cases an additonal branch instruction is required. This also adds unit tests to checks all the different condition codes. This is related o rdar://problem/18358882. llvm-svn: 217966
2014-09-18 01:46:47 +08:00
// Try to take advantage of fallthrough opportunities.
if (FuncInfo.MBB->isLayoutSuccessor(TBB)) {
std::swap(TBB, FBB);
Predicate = CmpInst::getInversePredicate(Predicate);
}
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Emit the cmp.
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
if (!emitCmp(CI->getOperand(0), CI->getOperand(1), CI->isUnsigned()))
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
[FastISel][AArch64] Improve branch selection to support all FP conditions. This adds the last two missing floating-point condition codes (FCMP_UEQ and FCMP_ONE) also to the branch selection. In these two cases an additonal branch instruction is required. This also adds unit tests to checks all the different condition codes. This is related o rdar://problem/18358882. llvm-svn: 217966
2014-09-18 01:46:47 +08:00
// FCMP_UEQ and FCMP_ONE cannot be checked with a single branch
// instruction.
AArch64FastISel: Use cbz/cbnz to branch on i1 In the case of a conditional branch without a preceding cmp we used to emit a "and; cmp; b.eq/b.ne" sequence, use tbz/tbnz instead. Differential Revision: http://reviews.llvm.org/D15122 llvm-svn: 254621
2015-12-04 01:19:58 +08:00
AArch64CC::CondCode CC = getCompareCC(Predicate);
[FastISel][AArch64] Improve branch selection to support all FP conditions. This adds the last two missing floating-point condition codes (FCMP_UEQ and FCMP_ONE) also to the branch selection. In these two cases an additonal branch instruction is required. This also adds unit tests to checks all the different condition codes. This is related o rdar://problem/18358882. llvm-svn: 217966
2014-09-18 01:46:47 +08:00
AArch64CC::CondCode ExtraCC = AArch64CC::AL;
switch (Predicate) {
default:
break;
case CmpInst::FCMP_UEQ:
ExtraCC = AArch64CC::EQ;
CC = AArch64CC::VS;
break;
case CmpInst::FCMP_ONE:
ExtraCC = AArch64CC::MI;
CC = AArch64CC::GT;
break;
}
assert((CC != AArch64CC::AL) && "Unexpected condition code.");
// Emit the extra branch for FCMP_UEQ and FCMP_ONE.
if (ExtraCC != AArch64CC::AL) {
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(AArch64::Bcc))
.addImm(ExtraCC)
.addMBB(TBB);
}
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Emit the branch.
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(AArch64::Bcc))
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
.addImm(CC)
.addMBB(TBB);
[FastISel][AArch64] Add branch weights. Add branch weights to branch instructions, so that the following passes can optimize based on it (i.e. basic block ordering). Fixes <rdar://problem/17887137>. llvm-svn: 214537
2014-08-02 02:39:24 +08:00
FastISel: Factor out common code; NFC intended This should be no functional change but for the record: For three cases in X86FastISel this will change the order in which the FalseMBB and TrueMBB of a conditional branch is addedd to the successor/predecessor lists. llvm-svn: 245997
2015-08-26 09:38:00 +08:00
finishCondBranch(BI->getParent(), TBB, FBB);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return true;
}
[FastISel][AArch64] Improve branch selection to support all FP conditions. This adds the last two missing floating-point condition codes (FCMP_UEQ and FCMP_ONE) also to the branch selection. In these two cases an additonal branch instruction is required. This also adds unit tests to checks all the different condition codes. This is related o rdar://problem/18358882. llvm-svn: 217966
2014-09-18 01:46:47 +08:00
} else if (const auto *CI = dyn_cast<ConstantInt>(BI->getCondition())) {
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
uint64_t Imm = CI->getZExtValue();
MachineBasicBlock *Target = (Imm == 0) ? FBB : TBB;
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(AArch64::B))
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
.addMBB(Target);
[FastISel][AArch64] Add branch weights. Add branch weights to branch instructions, so that the following passes can optimize based on it (i.e. basic block ordering). Fixes <rdar://problem/17887137>. llvm-svn: 214537
2014-08-02 02:39:24 +08:00
Let SelectionDAG start to use probability-based interface to add successors. The patch in http://reviews.llvm.org/D13745 is broken into four parts: 1. New interfaces without functional changes. 2. Use new interfaces in SelectionDAG, while in other passes treat probabilities as weights. 3. Use new interfaces in all other passes. 4. Remove old interfaces. This the second patch above. In this patch SelectionDAG starts to use probability-based interfaces in MBB to add successors but other MC passes are still using weight-based interfaces. Therefore, we need to maintain correct weight list in MBB even when probability-based interfaces are used. This is done by updating weight list in probability-based interfaces by treating the numerator of probabilities as weights. This change affects many test cases that check successor weight values. I will update those test cases once this patch looks good to you. Differential revision: http://reviews.llvm.org/D14361 llvm-svn: 253965
2015-11-24 16:51:23 +08:00
// Obtain the branch probability and add the target to the successor list.
Create a new interface addSuccessorWithoutWeight(MBB*) in MBB to add successors when optimization is disabled. When optimization is disabled, edge weights that are stored in MBB won't be used so that we don't have to store them. Currently, this is done by adding successors with default weight 0, and if all successors have default weights, the weight list will be empty. But that the weight list is empty doesn't mean disabled optimization (as is stated several times in MachineBasicBlock.cpp): it may also mean all successors just have default weights. We should discourage using default weights when adding successors, because it is very easy for users to forget update the correct edge weights instead of using default ones (one exception is that the MBB only has one successor). In order to detect such usages, it is better to differentiate using default weights from the case when optimizations is disabled. In this patch, a new interface addSuccessorWithoutWeight(MBB*) is created for when optimization is disabled. In this case, MBB will try to maintain an empty weight list, but it cannot guarantee this as for many uses of addSuccessor() whether optimization is disabled or not is not checked. But it can guarantee that if optimization is enabled, then the weight list always has the same size of the successor list. Differential revision: http://reviews.llvm.org/D13963 llvm-svn: 251429
2015-10-28 01:59:36 +08:00
if (FuncInfo.BPI) {
Let SelectionDAG start to use probability-based interface to add successors. The patch in http://reviews.llvm.org/D13745 is broken into four parts: 1. New interfaces without functional changes. 2. Use new interfaces in SelectionDAG, while in other passes treat probabilities as weights. 3. Use new interfaces in all other passes. 4. Remove old interfaces. This the second patch above. In this patch SelectionDAG starts to use probability-based interfaces in MBB to add successors but other MC passes are still using weight-based interfaces. Therefore, we need to maintain correct weight list in MBB even when probability-based interfaces are used. This is done by updating weight list in probability-based interfaces by treating the numerator of probabilities as weights. This change affects many test cases that check successor weight values. I will update those test cases once this patch looks good to you. Differential revision: http://reviews.llvm.org/D14361 llvm-svn: 253965
2015-11-24 16:51:23 +08:00
auto BranchProbability = FuncInfo.BPI->getEdgeProbability(
BI->getParent(), Target->getBasicBlock());
FuncInfo.MBB->addSuccessor(Target, BranchProbability);
Create a new interface addSuccessorWithoutWeight(MBB*) in MBB to add successors when optimization is disabled. When optimization is disabled, edge weights that are stored in MBB won't be used so that we don't have to store them. Currently, this is done by adding successors with default weight 0, and if all successors have default weights, the weight list will be empty. But that the weight list is empty doesn't mean disabled optimization (as is stated several times in MachineBasicBlock.cpp): it may also mean all successors just have default weights. We should discourage using default weights when adding successors, because it is very easy for users to forget update the correct edge weights instead of using default ones (one exception is that the MBB only has one successor). In order to detect such usages, it is better to differentiate using default weights from the case when optimizations is disabled. In this patch, a new interface addSuccessorWithoutWeight(MBB*) is created for when optimization is disabled. In this case, MBB will try to maintain an empty weight list, but it cannot guarantee this as for many uses of addSuccessor() whether optimization is disabled or not is not checked. But it can guarantee that if optimization is enabled, then the weight list always has the same size of the successor list. Differential revision: http://reviews.llvm.org/D13963 llvm-svn: 251429
2015-10-28 01:59:36 +08:00
} else
Let SelectionDAG start to use probability-based interface to add successors. The patch in http://reviews.llvm.org/D13745 is broken into four parts: 1. New interfaces without functional changes. 2. Use new interfaces in SelectionDAG, while in other passes treat probabilities as weights. 3. Use new interfaces in all other passes. 4. Remove old interfaces. This the second patch above. In this patch SelectionDAG starts to use probability-based interfaces in MBB to add successors but other MC passes are still using weight-based interfaces. Therefore, we need to maintain correct weight list in MBB even when probability-based interfaces are used. This is done by updating weight list in probability-based interfaces by treating the numerator of probabilities as weights. This change affects many test cases that check successor weight values. I will update those test cases once this patch looks good to you. Differential revision: http://reviews.llvm.org/D14361 llvm-svn: 253965
2015-11-24 16:51:23 +08:00
FuncInfo.MBB->addSuccessorWithoutProb(Target);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return true;
AArch64FastISel: Use cbz/cbnz to branch on i1 In the case of a conditional branch without a preceding cmp we used to emit a "and; cmp; b.eq/b.ne" sequence, use tbz/tbnz instead. Differential Revision: http://reviews.llvm.org/D15122 llvm-svn: 254621
2015-12-04 01:19:58 +08:00
} else {
AArch64CC::CondCode CC = AArch64CC::NE;
if (foldXALUIntrinsic(CC, I, BI->getCondition())) {
// Fake request the condition, otherwise the intrinsic might be completely
// optimized away.
unsigned CondReg = getRegForValue(BI->getCondition());
if (!CondReg)
return false;
[FastISel][AArch64] Add branch folding support for the XALU intrinsics. This improves the code generation for the XALU intrinsics when the condition is feeding a branch instruction. This is related to <rdar://problem/17831117>. llvm-svn: 214349
2014-07-31 06:04:34 +08:00
AArch64FastISel: Use cbz/cbnz to branch on i1 In the case of a conditional branch without a preceding cmp we used to emit a "and; cmp; b.eq/b.ne" sequence, use tbz/tbnz instead. Differential Revision: http://reviews.llvm.org/D15122 llvm-svn: 254621
2015-12-04 01:19:58 +08:00
// Emit the branch.
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(AArch64::Bcc))
.addImm(CC)
.addMBB(TBB);
[FastISel][AArch64] Add branch weights. Add branch weights to branch instructions, so that the following passes can optimize based on it (i.e. basic block ordering). Fixes <rdar://problem/17887137>. llvm-svn: 214537
2014-08-02 02:39:24 +08:00
AArch64FastISel: Use cbz/cbnz to branch on i1 In the case of a conditional branch without a preceding cmp we used to emit a "and; cmp; b.eq/b.ne" sequence, use tbz/tbnz instead. Differential Revision: http://reviews.llvm.org/D15122 llvm-svn: 254621
2015-12-04 01:19:58 +08:00
finishCondBranch(BI->getParent(), TBB, FBB);
return true;
}
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
}
unsigned CondReg = getRegForValue(BI->getCondition());
if (CondReg == 0)
return false;
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
bool CondRegIsKill = hasTrivialKill(BI->getCondition());
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
AArch64FastISel: Use cbz/cbnz to branch on i1 In the case of a conditional branch without a preceding cmp we used to emit a "and; cmp; b.eq/b.ne" sequence, use tbz/tbnz instead. Differential Revision: http://reviews.llvm.org/D15122 llvm-svn: 254621
2015-12-04 01:19:58 +08:00
// i1 conditions come as i32 values, test the lowest bit with tb(n)z.
unsigned Opcode = AArch64::TBNZW;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
if (FuncInfo.MBB->isLayoutSuccessor(TBB)) {
std::swap(TBB, FBB);
AArch64FastISel: Use cbz/cbnz to branch on i1 In the case of a conditional branch without a preceding cmp we used to emit a "and; cmp; b.eq/b.ne" sequence, use tbz/tbnz instead. Differential Revision: http://reviews.llvm.org/D15122 llvm-svn: 254621
2015-12-04 01:19:58 +08:00
Opcode = AArch64::TBZW;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
}
AArch64FastISel: Use cbz/cbnz to branch on i1 In the case of a conditional branch without a preceding cmp we used to emit a "and; cmp; b.eq/b.ne" sequence, use tbz/tbnz instead. Differential Revision: http://reviews.llvm.org/D15122 llvm-svn: 254621
2015-12-04 01:19:58 +08:00
const MCInstrDesc &II = TII.get(Opcode);
unsigned ConstrainedCondReg
= constrainOperandRegClass(II, CondReg, II.getNumDefs());
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, II)
.addReg(ConstrainedCondReg, getKillRegState(CondRegIsKill))
.addImm(0)
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
.addMBB(TBB);
[FastISel][AArch64] Add branch weights. Add branch weights to branch instructions, so that the following passes can optimize based on it (i.e. basic block ordering). Fixes <rdar://problem/17887137>. llvm-svn: 214537
2014-08-02 02:39:24 +08:00
FastISel: Factor out common code; NFC intended This should be no functional change but for the record: For three cases in X86FastISel this will change the order in which the FalseMBB and TrueMBB of a conditional branch is addedd to the successor/predecessor lists. llvm-svn: 245997
2015-08-26 09:38:00 +08:00
finishCondBranch(BI->getParent(), TBB, FBB);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return true;
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool AArch64FastISel::selectIndirectBr(const Instruction *I) {
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
const IndirectBrInst *BI = cast<IndirectBrInst>(I);
unsigned AddrReg = getRegForValue(BI->getOperand(0));
if (AddrReg == 0)
return false;
// Emit the indirect branch.
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
const MCInstrDesc &II = TII.get(AArch64::BR);
AddrReg = constrainOperandRegClass(II, AddrReg, II.getNumDefs());
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, II).addReg(AddrReg);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Make sure the CFG is up-to-date.
Move BB succ_iterator to be inside TerminatorInst. NFC. To get the successors of a BB we currently do successors(BB) which ultimately walks the successors of the BB's terminator. This moves the iterator to TerminatorInst as thats what we're actually using to do the iteration, and adds a member function to TerminatorInst to allow us to iterate directly over successors given an instruction. For example, we can now do for (auto *Succ : BI->successors()) instead of for (unsigned i = 0, e = BI->getNumSuccessors(); i != e; ++i) Reviewed by Tobias Grosser. llvm-svn: 244074
2015-08-06 01:43:01 +08:00
for (auto *Succ : BI->successors())
FuncInfo.MBB->addSuccessor(FuncInfo.MBBMap[Succ]);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return true;
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool AArch64FastISel::selectCmp(const Instruction *I) {
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
const CmpInst *CI = cast<CmpInst>(I);
[AArch64][FastISel] Don't even try to select vector icmps. We used to try to constant-fold them to i32 immediates. Given that fast-isel doesn't otherwise support vNi1, when selecting the result users, we'd fallback to SDAG anyway. However, if the users were in another block, we'd insert broken cross-class copies (GPR32 to FPR64). Give up, let SDAG agree with itself on a vNi1 legalization strategy. llvm-svn: 252364
2015-11-07 07:16:53 +08:00
// Vectors of i1 are weird: bail out.
if (CI->getType()->isVectorTy())
return false;
[FastISel][AArch64] Improve floating-point compare support. Add support for the last two missing fcmp condition codes: UEQ and ONE. This fixes rdar://problem/18341575. llvm-svn: 217823
2014-09-16 04:47:16 +08:00
// Try to optimize or fold the cmp.
CmpInst::Predicate Predicate = optimizeCmpPredicate(CI);
unsigned ResultReg = 0;
switch (Predicate) {
default:
break;
case CmpInst::FCMP_FALSE:
ResultReg = createResultReg(&AArch64::GPR32RegClass);
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc,
TII.get(TargetOpcode::COPY), ResultReg)
.addReg(AArch64::WZR, getKillRegState(true));
break;
case CmpInst::FCMP_TRUE:
ResultReg = fastEmit_i(MVT::i32, MVT::i32, ISD::Constant, 1);
break;
}
if (ResultReg) {
updateValueMap(I, ResultReg);
return true;
}
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Emit the cmp.
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
if (!emitCmp(CI->getOperand(0), CI->getOperand(1), CI->isUnsigned()))
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
[FastISel][AArch64] Improve floating-point compare support. Add support for the last two missing fcmp condition codes: UEQ and ONE. This fixes rdar://problem/18341575. llvm-svn: 217823
2014-09-16 04:47:16 +08:00
ResultReg = createResultReg(&AArch64::GPR32RegClass);
// FCMP_UEQ and FCMP_ONE cannot be checked with a single instruction. These
// condition codes are inverted, because they are used by CSINC.
static unsigned CondCodeTable[2][2] = {
{ AArch64CC::NE, AArch64CC::VC },
{ AArch64CC::PL, AArch64CC::LE }
};
unsigned *CondCodes = nullptr;
switch (Predicate) {
default:
break;
case CmpInst::FCMP_UEQ:
CondCodes = &CondCodeTable[0][0];
break;
case CmpInst::FCMP_ONE:
CondCodes = &CondCodeTable[1][0];
break;
}
if (CondCodes) {
unsigned TmpReg1 = createResultReg(&AArch64::GPR32RegClass);
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(AArch64::CSINCWr),
TmpReg1)
.addReg(AArch64::WZR, getKillRegState(true))
.addReg(AArch64::WZR, getKillRegState(true))
.addImm(CondCodes[0]);
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(AArch64::CSINCWr),
ResultReg)
.addReg(TmpReg1, getKillRegState(true))
.addReg(AArch64::WZR, getKillRegState(true))
.addImm(CondCodes[1]);
updateValueMap(I, ResultReg);
return true;
}
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Now set a register based on the comparison.
[FastISel][AArch64] Improve floating-point compare support. Add support for the last two missing fcmp condition codes: UEQ and ONE. This fixes rdar://problem/18341575. llvm-svn: 217823
2014-09-16 04:47:16 +08:00
AArch64CC::CondCode CC = getCompareCC(Predicate);
assert((CC != AArch64CC::AL) && "Unexpected condition code.");
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
AArch64CC::CondCode invertedCC = getInvertedCondCode(CC);
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(AArch64::CSINCWr),
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
ResultReg)
[FastISel][AArch64] Improve floating-point compare support. Add support for the last two missing fcmp condition codes: UEQ and ONE. This fixes rdar://problem/18341575. llvm-svn: 217823
2014-09-16 04:47:16 +08:00
.addReg(AArch64::WZR, getKillRegState(true))
.addReg(AArch64::WZR, getKillRegState(true))
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
.addImm(invertedCC);
[FastISel] Rename public visible FastISel functions. NFC. This commit renames the following public FastISel functions: LowerArguments -> lowerArguments SelectInstruction -> selectInstruction TargetSelectInstruction -> fastSelectInstruction FastLowerArguments -> fastLowerArguments FastLowerCall -> fastLowerCall FastLowerIntrinsicCall -> fastLowerIntrinsicCall FastEmitZExtFromI1 -> fastEmitZExtFromI1 FastEmitBranch -> fastEmitBranch UpdateValueMap -> updateValueMap TargetMaterializeConstant -> fastMaterializeConstant TargetMaterializeAlloca -> fastMaterializeAlloca TargetMaterializeFloatZero -> fastMaterializeFloatZero LowerCallTo -> lowerCallTo Reviewed by Eric llvm-svn: 217074
2014-09-04 04:56:52 +08:00
updateValueMap(I, ResultReg);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return true;
}
[FastISel][AArch64] Optimize select when one of the operands is a 'true' or 'false' value. Optimize selects of i1 in the presence of 'true' and 'false' operands to simple logic operations. This fixes rdar://problem/18960150. llvm-svn: 221848
2014-11-13 08:36:46 +08:00
/// \brief Optimize selects of i1 if one of the operands has a 'true' or 'false'
/// value.
bool AArch64FastISel::optimizeSelect(const SelectInst *SI) {
if (!SI->getType()->isIntegerTy(1))
return false;
const Value *Src1Val, *Src2Val;
unsigned Opc = 0;
bool NeedExtraOp = false;
if (auto *CI = dyn_cast<ConstantInt>(SI->getTrueValue())) {
if (CI->isOne()) {
Src1Val = SI->getCondition();
Src2Val = SI->getFalseValue();
Opc = AArch64::ORRWrr;
} else {
assert(CI->isZero());
Src1Val = SI->getFalseValue();
Src2Val = SI->getCondition();
Opc = AArch64::BICWrr;
}
} else if (auto *CI = dyn_cast<ConstantInt>(SI->getFalseValue())) {
if (CI->isOne()) {
Src1Val = SI->getCondition();
Src2Val = SI->getTrueValue();
Opc = AArch64::ORRWrr;
NeedExtraOp = true;
} else {
assert(CI->isZero());
Src1Val = SI->getCondition();
Src2Val = SI->getTrueValue();
Opc = AArch64::ANDWrr;
}
}
if (!Opc)
return false;
unsigned Src1Reg = getRegForValue(Src1Val);
if (!Src1Reg)
return false;
bool Src1IsKill = hasTrivialKill(Src1Val);
unsigned Src2Reg = getRegForValue(Src2Val);
if (!Src2Reg)
return false;
bool Src2IsKill = hasTrivialKill(Src2Val);
if (NeedExtraOp) {
Src1Reg = emitLogicalOp_ri(ISD::XOR, MVT::i32, Src1Reg, Src1IsKill, 1);
Src1IsKill = true;
}
[AArch64][FastISel] Variant of the logical instructions that use two input registers cannot write on SP. rdar://problem/20748715 llvm-svn: 236352
2015-05-02 05:34:57 +08:00
unsigned ResultReg = fastEmitInst_rr(Opc, &AArch64::GPR32RegClass, Src1Reg,
[FastISel][AArch64] Optimize select when one of the operands is a 'true' or 'false' value. Optimize selects of i1 in the presence of 'true' and 'false' operands to simple logic operations. This fixes rdar://problem/18960150. llvm-svn: 221848
2014-11-13 08:36:46 +08:00
Src1IsKill, Src2Reg, Src2IsKill);
updateValueMap(SI, ResultReg);
return true;
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool AArch64FastISel::selectSelect(const Instruction *I) {
[FastISel][AArch64] Extend 'select' lowering to support also i1 to i16. Related to rdar://problem/18960150. llvm-svn: 221846
2014-11-13 08:36:38 +08:00
assert(isa<SelectInst>(I) && "Expected a select instruction.");
MVT VT;
if (!isTypeSupported(I->getType(), VT))
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
[FastISel][AArch64] Extend 'select' lowering to support also i1 to i16. Related to rdar://problem/18960150. llvm-svn: 221846
2014-11-13 08:36:38 +08:00
unsigned Opc;
const TargetRegisterClass *RC;
switch (VT.SimpleTy) {
default:
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
[FastISel][AArch64] Extend 'select' lowering to support also i1 to i16. Related to rdar://problem/18960150. llvm-svn: 221846
2014-11-13 08:36:38 +08:00
case MVT::i1:
case MVT::i8:
case MVT::i16:
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
case MVT::i32:
[FastISel][AArch64] Extend 'select' lowering to support also i1 to i16. Related to rdar://problem/18960150. llvm-svn: 221846
2014-11-13 08:36:38 +08:00
Opc = AArch64::CSELWr;
RC = &AArch64::GPR32RegClass;
break;
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
case MVT::i64:
[FastISel][AArch64] Extend 'select' lowering to support also i1 to i16. Related to rdar://problem/18960150. llvm-svn: 221846
2014-11-13 08:36:38 +08:00
Opc = AArch64::CSELXr;
RC = &AArch64::GPR64RegClass;
break;
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
case MVT::f32:
[FastISel][AArch64] Extend 'select' lowering to support also i1 to i16. Related to rdar://problem/18960150. llvm-svn: 221846
2014-11-13 08:36:38 +08:00
Opc = AArch64::FCSELSrrr;
RC = &AArch64::FPR32RegClass;
break;
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
case MVT::f64:
[FastISel][AArch64] Extend 'select' lowering to support also i1 to i16. Related to rdar://problem/18960150. llvm-svn: 221846
2014-11-13 08:36:38 +08:00
Opc = AArch64::FCSELDrrr;
RC = &AArch64::FPR64RegClass;
break;
[FastISel][AArch64] Add select folding support for the XALU intrinsics. This improves the code generation for the XALU intrinsics when the condition is feeding a select instruction. This also updates and enables the XALU unit tests for FastISel. This fixes <rdar://problem/17831117>. llvm-svn: 214350
2014-07-31 06:04:37 +08:00
}
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
[FastISel][AArch64] Extend 'select' lowering to support also i1 to i16. Related to rdar://problem/18960150. llvm-svn: 221846
2014-11-13 08:36:38 +08:00
const SelectInst *SI = cast<SelectInst>(I);
[FastISel][AArch64] Add select folding support for the XALU intrinsics. This improves the code generation for the XALU intrinsics when the condition is feeding a select instruction. This also updates and enables the XALU unit tests for FastISel. This fixes <rdar://problem/17831117>. llvm-svn: 214350
2014-07-31 06:04:37 +08:00
const Value *Cond = SI->getCondition();
AArch64CC::CondCode CC = AArch64CC::NE;
[FastISel][AArch64] Fold the cmp into the select when possible. This folds the compare emission into the select emission when possible, so we can directly use the flags and don't have to emit a separate compare. Related to rdar://problem/18960150. llvm-svn: 221847
2014-11-13 08:36:43 +08:00
AArch64CC::CondCode ExtraCC = AArch64CC::AL;
[ARM64] Correctly select ANDWri in FastISel. http://reviews.llvm.org/D3598 llvm-svn: 207917
2014-05-04 01:27:06 +08:00
[FastISel][AArch64] Optimize select when one of the operands is a 'true' or 'false' value. Optimize selects of i1 in the presence of 'true' and 'false' operands to simple logic operations. This fixes rdar://problem/18960150. llvm-svn: 221848
2014-11-13 08:36:46 +08:00
if (optimizeSelect(SI))
return true;
[FastISel][AArch64] Extend 'select' lowering to support also i1 to i16. Related to rdar://problem/18960150. llvm-svn: 221846
2014-11-13 08:36:38 +08:00
// Try to pickup the flags, so we don't have to emit another compare.
if (foldXALUIntrinsic(CC, I, Cond)) {
// Fake request the condition to force emission of the XALU intrinsic.
unsigned CondReg = getRegForValue(Cond);
if (!CondReg)
return false;
[FastISel][AArch64] Fold the cmp into the select when possible. This folds the compare emission into the select emission when possible, so we can directly use the flags and don't have to emit a separate compare. Related to rdar://problem/18960150. llvm-svn: 221847
2014-11-13 08:36:43 +08:00
} else if (isa<CmpInst>(Cond) && cast<CmpInst>(Cond)->hasOneUse() &&
isValueAvailable(Cond)) {
const auto *Cmp = cast<CmpInst>(Cond);
// Try to optimize or fold the cmp.
CmpInst::Predicate Predicate = optimizeCmpPredicate(Cmp);
const Value *FoldSelect = nullptr;
switch (Predicate) {
default:
break;
case CmpInst::FCMP_FALSE:
FoldSelect = SI->getFalseValue();
break;
case CmpInst::FCMP_TRUE:
FoldSelect = SI->getTrueValue();
break;
}
if (FoldSelect) {
unsigned SrcReg = getRegForValue(FoldSelect);
if (!SrcReg)
return false;
unsigned UseReg = lookUpRegForValue(SI);
if (UseReg)
MRI.clearKillFlags(UseReg);
updateValueMap(I, SrcReg);
return true;
}
// Emit the cmp.
if (!emitCmp(Cmp->getOperand(0), Cmp->getOperand(1), Cmp->isUnsigned()))
return false;
// FCMP_UEQ and FCMP_ONE cannot be checked with a single select instruction.
CC = getCompareCC(Predicate);
switch (Predicate) {
default:
break;
case CmpInst::FCMP_UEQ:
ExtraCC = AArch64CC::EQ;
CC = AArch64CC::VS;
break;
case CmpInst::FCMP_ONE:
ExtraCC = AArch64CC::MI;
CC = AArch64CC::GT;
break;
}
assert((CC != AArch64CC::AL) && "Unexpected condition code.");
[FastISel][AArch64] Extend 'select' lowering to support also i1 to i16. Related to rdar://problem/18960150. llvm-svn: 221846
2014-11-13 08:36:38 +08:00
} else {
unsigned CondReg = getRegForValue(Cond);
if (!CondReg)
return false;
bool CondIsKill = hasTrivialKill(Cond);
[FastISel][AArch64] Add select folding support for the XALU intrinsics. This improves the code generation for the XALU intrinsics when the condition is feeding a select instruction. This also updates and enables the XALU unit tests for FastISel. This fixes <rdar://problem/17831117>. llvm-svn: 214350
2014-07-31 06:04:37 +08:00
[AArch64] Fix bad register class constraint in fast-isel for TST instruction. rdar://problem/20748715 llvm-svn: 236273
2015-05-01 06:27:20 +08:00
const MCInstrDesc &II = TII.get(AArch64::ANDSWri);
CondReg = constrainOperandRegClass(II, CondReg, 1);
[FastISel][AArch64] Extend 'select' lowering to support also i1 to i16. Related to rdar://problem/18960150. llvm-svn: 221846
2014-11-13 08:36:38 +08:00
// Emit a TST instruction (ANDS wzr, reg, #imm).
[AArch64] Fix bad register class constraint in fast-isel for TST instruction. rdar://problem/20748715 llvm-svn: 236273
2015-05-01 06:27:20 +08:00
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, II,
[FastISel][AArch64] Extend 'select' lowering to support also i1 to i16. Related to rdar://problem/18960150. llvm-svn: 221846
2014-11-13 08:36:38 +08:00
AArch64::WZR)
.addReg(CondReg, getKillRegState(CondIsKill))
.addImm(AArch64_AM::encodeLogicalImmediate(1, 32));
[FastISel][AArch64] Add select folding support for the XALU intrinsics. This improves the code generation for the XALU intrinsics when the condition is feeding a select instruction. This also updates and enables the XALU unit tests for FastISel. This fixes <rdar://problem/17831117>. llvm-svn: 214350
2014-07-31 06:04:37 +08:00
}
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
[FastISel][AArch64] Extend 'select' lowering to support also i1 to i16. Related to rdar://problem/18960150. llvm-svn: 221846
2014-11-13 08:36:38 +08:00
unsigned Src1Reg = getRegForValue(SI->getTrueValue());
bool Src1IsKill = hasTrivialKill(SI->getTrueValue());
[FastISel][AArch64] Add select folding support for the XALU intrinsics. This improves the code generation for the XALU intrinsics when the condition is feeding a select instruction. This also updates and enables the XALU unit tests for FastISel. This fixes <rdar://problem/17831117>. llvm-svn: 214350
2014-07-31 06:04:37 +08:00
[FastISel][AArch64] Extend 'select' lowering to support also i1 to i16. Related to rdar://problem/18960150. llvm-svn: 221846
2014-11-13 08:36:38 +08:00
unsigned Src2Reg = getRegForValue(SI->getFalseValue());
bool Src2IsKill = hasTrivialKill(SI->getFalseValue());
[FastISel][AArch64] Add select folding support for the XALU intrinsics. This improves the code generation for the XALU intrinsics when the condition is feeding a select instruction. This also updates and enables the XALU unit tests for FastISel. This fixes <rdar://problem/17831117>. llvm-svn: 214350
2014-07-31 06:04:37 +08:00
[FastISel][AArch64] Extend 'select' lowering to support also i1 to i16. Related to rdar://problem/18960150. llvm-svn: 221846
2014-11-13 08:36:38 +08:00
if (!Src1Reg || !Src2Reg)
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
[FastISel][AArch64] Fold the cmp into the select when possible. This folds the compare emission into the select emission when possible, so we can directly use the flags and don't have to emit a separate compare. Related to rdar://problem/18960150. llvm-svn: 221847
2014-11-13 08:36:43 +08:00
if (ExtraCC != AArch64CC::AL) {
Src2Reg = fastEmitInst_rri(Opc, RC, Src1Reg, Src1IsKill, Src2Reg,
Src2IsKill, ExtraCC);
Src2IsKill = true;
}
[FastISel][AArch64] Extend 'select' lowering to support also i1 to i16. Related to rdar://problem/18960150. llvm-svn: 221846
2014-11-13 08:36:38 +08:00
unsigned ResultReg = fastEmitInst_rri(Opc, RC, Src1Reg, Src1IsKill, Src2Reg,
Src2IsKill, CC);
[FastISel] Rename public visible FastISel functions. NFC. This commit renames the following public FastISel functions: LowerArguments -> lowerArguments SelectInstruction -> selectInstruction TargetSelectInstruction -> fastSelectInstruction FastLowerArguments -> fastLowerArguments FastLowerCall -> fastLowerCall FastLowerIntrinsicCall -> fastLowerIntrinsicCall FastEmitZExtFromI1 -> fastEmitZExtFromI1 FastEmitBranch -> fastEmitBranch UpdateValueMap -> updateValueMap TargetMaterializeConstant -> fastMaterializeConstant TargetMaterializeAlloca -> fastMaterializeAlloca TargetMaterializeFloatZero -> fastMaterializeFloatZero LowerCallTo -> lowerCallTo Reviewed by Eric llvm-svn: 217074
2014-09-04 04:56:52 +08:00
updateValueMap(I, ResultReg);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return true;
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool AArch64FastISel::selectFPExt(const Instruction *I) {
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
Value *V = I->getOperand(0);
if (!I->getType()->isDoubleTy() || !V->getType()->isFloatTy())
return false;
unsigned Op = getRegForValue(V);
if (Op == 0)
return false;
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
unsigned ResultReg = createResultReg(&AArch64::FPR64RegClass);
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(AArch64::FCVTDSr),
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
ResultReg).addReg(Op);
[FastISel] Rename public visible FastISel functions. NFC. This commit renames the following public FastISel functions: LowerArguments -> lowerArguments SelectInstruction -> selectInstruction TargetSelectInstruction -> fastSelectInstruction FastLowerArguments -> fastLowerArguments FastLowerCall -> fastLowerCall FastLowerIntrinsicCall -> fastLowerIntrinsicCall FastEmitZExtFromI1 -> fastEmitZExtFromI1 FastEmitBranch -> fastEmitBranch UpdateValueMap -> updateValueMap TargetMaterializeConstant -> fastMaterializeConstant TargetMaterializeAlloca -> fastMaterializeAlloca TargetMaterializeFloatZero -> fastMaterializeFloatZero LowerCallTo -> lowerCallTo Reviewed by Eric llvm-svn: 217074
2014-09-04 04:56:52 +08:00
updateValueMap(I, ResultReg);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return true;
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool AArch64FastISel::selectFPTrunc(const Instruction *I) {
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
Value *V = I->getOperand(0);
if (!I->getType()->isFloatTy() || !V->getType()->isDoubleTy())
return false;
unsigned Op = getRegForValue(V);
if (Op == 0)
return false;
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
unsigned ResultReg = createResultReg(&AArch64::FPR32RegClass);
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(AArch64::FCVTSDr),
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
ResultReg).addReg(Op);
[FastISel] Rename public visible FastISel functions. NFC. This commit renames the following public FastISel functions: LowerArguments -> lowerArguments SelectInstruction -> selectInstruction TargetSelectInstruction -> fastSelectInstruction FastLowerArguments -> fastLowerArguments FastLowerCall -> fastLowerCall FastLowerIntrinsicCall -> fastLowerIntrinsicCall FastEmitZExtFromI1 -> fastEmitZExtFromI1 FastEmitBranch -> fastEmitBranch UpdateValueMap -> updateValueMap TargetMaterializeConstant -> fastMaterializeConstant TargetMaterializeAlloca -> fastMaterializeAlloca TargetMaterializeFloatZero -> fastMaterializeFloatZero LowerCallTo -> lowerCallTo Reviewed by Eric llvm-svn: 217074
2014-09-04 04:56:52 +08:00
updateValueMap(I, ResultReg);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return true;
}
// FPToUI and FPToSI
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool AArch64FastISel::selectFPToInt(const Instruction *I, bool Signed) {
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
MVT DestVT;
if (!isTypeLegal(I->getType(), DestVT) || DestVT.isVector())
return false;
unsigned SrcReg = getRegForValue(I->getOperand(0));
if (SrcReg == 0)
return false;
Make TargetLowering::getPointerTy() taking DataLayout as an argument Summary: This change is part of a series of commits dedicated to have a single DataLayout during compilation by using always the one owned by the module. Reviewers: echristo Subscribers: jholewinski, ted, yaron.keren, rafael, llvm-commits Differential Revision: http://reviews.llvm.org/D11028 From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 241775
2015-07-09 10:09:04 +08:00
EVT SrcVT = TLI.getValueType(DL, I->getOperand(0)->getType(), true);
[ARM64][fast-isel] Fast-isel doesn't know how to handle f128. llvm-svn: 207659
2014-04-30 23:29:57 +08:00
if (SrcVT == MVT::f128)
return false;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
unsigned Opc;
if (SrcVT == MVT::f64) {
if (Signed)
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
Opc = (DestVT == MVT::i32) ? AArch64::FCVTZSUWDr : AArch64::FCVTZSUXDr;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
else
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
Opc = (DestVT == MVT::i32) ? AArch64::FCVTZUUWDr : AArch64::FCVTZUUXDr;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
} else {
if (Signed)
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
Opc = (DestVT == MVT::i32) ? AArch64::FCVTZSUWSr : AArch64::FCVTZSUXSr;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
else
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
Opc = (DestVT == MVT::i32) ? AArch64::FCVTZUUWSr : AArch64::FCVTZUUXSr;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
}
ARM64: add constraints to various FastISel operations llvm-svn: 206284
2014-04-15 21:59:53 +08:00
unsigned ResultReg = createResultReg(
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
DestVT == MVT::i32 ? &AArch64::GPR32RegClass : &AArch64::GPR64RegClass);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(Opc), ResultReg)
.addReg(SrcReg);
[FastISel] Rename public visible FastISel functions. NFC. This commit renames the following public FastISel functions: LowerArguments -> lowerArguments SelectInstruction -> selectInstruction TargetSelectInstruction -> fastSelectInstruction FastLowerArguments -> fastLowerArguments FastLowerCall -> fastLowerCall FastLowerIntrinsicCall -> fastLowerIntrinsicCall FastEmitZExtFromI1 -> fastEmitZExtFromI1 FastEmitBranch -> fastEmitBranch UpdateValueMap -> updateValueMap TargetMaterializeConstant -> fastMaterializeConstant TargetMaterializeAlloca -> fastMaterializeAlloca TargetMaterializeFloatZero -> fastMaterializeFloatZero LowerCallTo -> lowerCallTo Reviewed by Eric llvm-svn: 217074
2014-09-04 04:56:52 +08:00
updateValueMap(I, ResultReg);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return true;
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool AArch64FastISel::selectIntToFP(const Instruction *I, bool Signed) {
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
MVT DestVT;
if (!isTypeLegal(I->getType(), DestVT) || DestVT.isVector())
return false;
[ARM64][fast-isel] Fast-isel doesn't know how to handle f128. llvm-svn: 207659
2014-04-30 23:29:57 +08:00
assert ((DestVT == MVT::f32 || DestVT == MVT::f64) &&
"Unexpected value type.");
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
unsigned SrcReg = getRegForValue(I->getOperand(0));
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
if (!SrcReg)
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
bool SrcIsKill = hasTrivialKill(I->getOperand(0));
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
Make TargetLowering::getPointerTy() taking DataLayout as an argument Summary: This change is part of a series of commits dedicated to have a single DataLayout during compilation by using always the one owned by the module. Reviewers: echristo Subscribers: jholewinski, ted, yaron.keren, rafael, llvm-commits Differential Revision: http://reviews.llvm.org/D11028 From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 241775
2015-07-09 10:09:04 +08:00
EVT SrcVT = TLI.getValueType(DL, I->getOperand(0)->getType(), true);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Handle sign-extension.
if (SrcVT == MVT::i16 || SrcVT == MVT::i8 || SrcVT == MVT::i1) {
SrcReg =
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
emitIntExt(SrcVT.getSimpleVT(), SrcReg, MVT::i32, /*isZExt*/ !Signed);
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
if (!SrcReg)
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
SrcIsKill = true;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
}
unsigned Opc;
if (SrcVT == MVT::i64) {
if (Signed)
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
Opc = (DestVT == MVT::f32) ? AArch64::SCVTFUXSri : AArch64::SCVTFUXDri;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
else
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
Opc = (DestVT == MVT::f32) ? AArch64::UCVTFUXSri : AArch64::UCVTFUXDri;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
} else {
if (Signed)
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
Opc = (DestVT == MVT::f32) ? AArch64::SCVTFUWSri : AArch64::SCVTFUWDri;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
else
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
Opc = (DestVT == MVT::f32) ? AArch64::UCVTFUWSri : AArch64::UCVTFUWDri;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
}
[FastISel][tblgen] Rename tblgen generated FastISel functions. NFC. This is the final round of renaming. This changes tblgen to emit lower-case function names for FastEmitInst_* and FastEmit_*, and updates all its uses in the source code. Reviewed by Eric llvm-svn: 217075
2014-09-04 04:56:59 +08:00
unsigned ResultReg = fastEmitInst_r(Opc, TLI.getRegClassFor(DestVT), SrcReg,
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
SrcIsKill);
[FastISel] Rename public visible FastISel functions. NFC. This commit renames the following public FastISel functions: LowerArguments -> lowerArguments SelectInstruction -> selectInstruction TargetSelectInstruction -> fastSelectInstruction FastLowerArguments -> fastLowerArguments FastLowerCall -> fastLowerCall FastLowerIntrinsicCall -> fastLowerIntrinsicCall FastEmitZExtFromI1 -> fastEmitZExtFromI1 FastEmitBranch -> fastEmitBranch UpdateValueMap -> updateValueMap TargetMaterializeConstant -> fastMaterializeConstant TargetMaterializeAlloca -> fastMaterializeAlloca TargetMaterializeFloatZero -> fastMaterializeFloatZero LowerCallTo -> lowerCallTo Reviewed by Eric llvm-svn: 217074
2014-09-04 04:56:52 +08:00
updateValueMap(I, ResultReg);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return true;
}
[FastISel] Rename public visible FastISel functions. NFC. This commit renames the following public FastISel functions: LowerArguments -> lowerArguments SelectInstruction -> selectInstruction TargetSelectInstruction -> fastSelectInstruction FastLowerArguments -> fastLowerArguments FastLowerCall -> fastLowerCall FastLowerIntrinsicCall -> fastLowerIntrinsicCall FastEmitZExtFromI1 -> fastEmitZExtFromI1 FastEmitBranch -> fastEmitBranch UpdateValueMap -> updateValueMap TargetMaterializeConstant -> fastMaterializeConstant TargetMaterializeAlloca -> fastMaterializeAlloca TargetMaterializeFloatZero -> fastMaterializeFloatZero LowerCallTo -> lowerCallTo Reviewed by Eric llvm-svn: 217074
2014-09-04 04:56:52 +08:00
bool AArch64FastISel::fastLowerArguments() {
[FastISel][AArch64] Implement the FastLowerArguments hook. This implements basic argument lowering for AArch64 in FastISel. It only handles a small subset of the C calling convention. It supports simple arguments that can be passed in GPR and FPR registers. This should cover most of the trivial cases without falling back to SelectionDAG. This fixes <rdar://problem/17890986>. llvm-svn: 214846
2014-08-05 13:43:48 +08:00
if (!FuncInfo.CanLowerReturn)
return false;
const Function *F = FuncInfo.Fn;
if (F->isVarArg())
return false;
CallingConv::ID CC = F->getCallingConv();
if (CC != CallingConv::C)
return false;
[FastISel][AArch64] Add vector support to argument lowering. Lower the first 8 vector arguments too. llvm-svn: 217850
2014-09-16 08:25:30 +08:00
// Only handle simple cases of up to 8 GPR and FPR each.
[FastISel][AArch64] Implement the FastLowerArguments hook. This implements basic argument lowering for AArch64 in FastISel. It only handles a small subset of the C calling convention. It supports simple arguments that can be passed in GPR and FPR registers. This should cover most of the trivial cases without falling back to SelectionDAG. This fixes <rdar://problem/17890986>. llvm-svn: 214846
2014-08-05 13:43:48 +08:00
unsigned GPRCnt = 0;
unsigned FPRCnt = 0;
unsigned Idx = 0;
for (auto const &Arg : F->args()) {
// The first argument is at index 1.
++Idx;
if (F->getAttributes().hasAttribute(Idx, Attribute::ByVal) ||
F->getAttributes().hasAttribute(Idx, Attribute::InReg) ||
F->getAttributes().hasAttribute(Idx, Attribute::StructRet) ||
Swift Calling Convention: add swiftself attribute. Differential Revision: http://reviews.llvm.org/D17866 llvm-svn: 264754
2016-03-30 01:37:21 +08:00
F->getAttributes().hasAttribute(Idx, Attribute::SwiftSelf) ||
Swift Calling Convention: swifterror target support. Differential Revision: http://reviews.llvm.org/D18716 llvm-svn: 265997
2016-04-12 05:08:06 +08:00
F->getAttributes().hasAttribute(Idx, Attribute::SwiftError) ||
[FastISel][AArch64] Implement the FastLowerArguments hook. This implements basic argument lowering for AArch64 in FastISel. It only handles a small subset of the C calling convention. It supports simple arguments that can be passed in GPR and FPR registers. This should cover most of the trivial cases without falling back to SelectionDAG. This fixes <rdar://problem/17890986>. llvm-svn: 214846
2014-08-05 13:43:48 +08:00
F->getAttributes().hasAttribute(Idx, Attribute::Nest))
return false;
Type *ArgTy = Arg.getType();
[FastISel][AArch64] Add vector support to argument lowering. Lower the first 8 vector arguments too. llvm-svn: 217850
2014-09-16 08:25:30 +08:00
if (ArgTy->isStructTy() || ArgTy->isArrayTy())
[FastISel][AArch64] Implement the FastLowerArguments hook. This implements basic argument lowering for AArch64 in FastISel. It only handles a small subset of the C calling convention. It supports simple arguments that can be passed in GPR and FPR registers. This should cover most of the trivial cases without falling back to SelectionDAG. This fixes <rdar://problem/17890986>. llvm-svn: 214846
2014-08-05 13:43:48 +08:00
return false;
Make TargetLowering::getPointerTy() taking DataLayout as an argument Summary: This change is part of a series of commits dedicated to have a single DataLayout during compilation by using always the one owned by the module. Reviewers: echristo Subscribers: jholewinski, ted, yaron.keren, rafael, llvm-commits Differential Revision: http://reviews.llvm.org/D11028 From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 241775
2015-07-09 10:09:04 +08:00
EVT ArgVT = TLI.getValueType(DL, ArgTy);
[FastISel][AArch64] Add vector support to argument lowering. Lower the first 8 vector arguments too. llvm-svn: 217850
2014-09-16 08:25:30 +08:00
if (!ArgVT.isSimple())
return false;
MVT VT = ArgVT.getSimpleVT().SimpleTy;
if (VT.isFloatingPoint() && !Subtarget->hasFPARMv8())
return false;
if (VT.isVector() &&
(!Subtarget->hasNEON() || !Subtarget->isLittleEndian()))
return false;
if (VT >= MVT::i1 && VT <= MVT::i64)
[FastISel][AArch64] Implement the FastLowerArguments hook. This implements basic argument lowering for AArch64 in FastISel. It only handles a small subset of the C calling convention. It supports simple arguments that can be passed in GPR and FPR registers. This should cover most of the trivial cases without falling back to SelectionDAG. This fixes <rdar://problem/17890986>. llvm-svn: 214846
2014-08-05 13:43:48 +08:00
++GPRCnt;
[FastISel][AArch64] Add vector support to argument lowering. Lower the first 8 vector arguments too. llvm-svn: 217850
2014-09-16 08:25:30 +08:00
else if ((VT >= MVT::f16 && VT <= MVT::f64) || VT.is64BitVector() ||
VT.is128BitVector())
[FastISel][AArch64] Implement the FastLowerArguments hook. This implements basic argument lowering for AArch64 in FastISel. It only handles a small subset of the C calling convention. It supports simple arguments that can be passed in GPR and FPR registers. This should cover most of the trivial cases without falling back to SelectionDAG. This fixes <rdar://problem/17890986>. llvm-svn: 214846
2014-08-05 13:43:48 +08:00
++FPRCnt;
[FastISel][AArch64] Add vector support to argument lowering. Lower the first 8 vector arguments too. llvm-svn: 217850
2014-09-16 08:25:30 +08:00
else
return false;
[FastISel][AArch64] Implement the FastLowerArguments hook. This implements basic argument lowering for AArch64 in FastISel. It only handles a small subset of the C calling convention. It supports simple arguments that can be passed in GPR and FPR registers. This should cover most of the trivial cases without falling back to SelectionDAG. This fixes <rdar://problem/17890986>. llvm-svn: 214846
2014-08-05 13:43:48 +08:00
if (GPRCnt > 8 || FPRCnt > 8)
return false;
}
[FastISel][AArch64] Add vector support to argument lowering. Lower the first 8 vector arguments too. llvm-svn: 217850
2014-09-16 08:25:30 +08:00
static const MCPhysReg Registers[6][8] = {
[FastISel][AArch64] Implement the FastLowerArguments hook. This implements basic argument lowering for AArch64 in FastISel. It only handles a small subset of the C calling convention. It supports simple arguments that can be passed in GPR and FPR registers. This should cover most of the trivial cases without falling back to SelectionDAG. This fixes <rdar://problem/17890986>. llvm-svn: 214846
2014-08-05 13:43:48 +08:00
{ AArch64::W0, AArch64::W1, AArch64::W2, AArch64::W3, AArch64::W4,
AArch64::W5, AArch64::W6, AArch64::W7 },
{ AArch64::X0, AArch64::X1, AArch64::X2, AArch64::X3, AArch64::X4,
AArch64::X5, AArch64::X6, AArch64::X7 },
{ AArch64::H0, AArch64::H1, AArch64::H2, AArch64::H3, AArch64::H4,
AArch64::H5, AArch64::H6, AArch64::H7 },
{ AArch64::S0, AArch64::S1, AArch64::S2, AArch64::S3, AArch64::S4,
AArch64::S5, AArch64::S6, AArch64::S7 },
{ AArch64::D0, AArch64::D1, AArch64::D2, AArch64::D3, AArch64::D4,
[FastISel][AArch64] Add vector support to argument lowering. Lower the first 8 vector arguments too. llvm-svn: 217850
2014-09-16 08:25:30 +08:00
AArch64::D5, AArch64::D6, AArch64::D7 },
{ AArch64::Q0, AArch64::Q1, AArch64::Q2, AArch64::Q3, AArch64::Q4,
AArch64::Q5, AArch64::Q6, AArch64::Q7 }
[FastISel][AArch64] Implement the FastLowerArguments hook. This implements basic argument lowering for AArch64 in FastISel. It only handles a small subset of the C calling convention. It supports simple arguments that can be passed in GPR and FPR registers. This should cover most of the trivial cases without falling back to SelectionDAG. This fixes <rdar://problem/17890986>. llvm-svn: 214846
2014-08-05 13:43:48 +08:00
};
unsigned GPRIdx = 0;
unsigned FPRIdx = 0;
for (auto const &Arg : F->args()) {
Make TargetLowering::getPointerTy() taking DataLayout as an argument Summary: This change is part of a series of commits dedicated to have a single DataLayout during compilation by using always the one owned by the module. Reviewers: echristo Subscribers: jholewinski, ted, yaron.keren, rafael, llvm-commits Differential Revision: http://reviews.llvm.org/D11028 From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 241775
2015-07-09 10:09:04 +08:00
MVT VT = TLI.getSimpleValueType(DL, Arg.getType());
[FastISel][AArch64] Implement the FastLowerArguments hook. This implements basic argument lowering for AArch64 in FastISel. It only handles a small subset of the C calling convention. It supports simple arguments that can be passed in GPR and FPR registers. This should cover most of the trivial cases without falling back to SelectionDAG. This fixes <rdar://problem/17890986>. llvm-svn: 214846
2014-08-05 13:43:48 +08:00
unsigned SrcReg;
[FastISel][AArch64] Add vector support to argument lowering. Lower the first 8 vector arguments too. llvm-svn: 217850
2014-09-16 08:25:30 +08:00
const TargetRegisterClass *RC;
if (VT >= MVT::i1 && VT <= MVT::i32) {
SrcReg = Registers[0][GPRIdx++];
RC = &AArch64::GPR32RegClass;
VT = MVT::i32;
} else if (VT == MVT::i64) {
SrcReg = Registers[1][GPRIdx++];
RC = &AArch64::GPR64RegClass;
} else if (VT == MVT::f16) {
SrcReg = Registers[2][FPRIdx++];
RC = &AArch64::FPR16RegClass;
} else if (VT == MVT::f32) {
SrcReg = Registers[3][FPRIdx++];
RC = &AArch64::FPR32RegClass;
} else if ((VT == MVT::f64) || VT.is64BitVector()) {
SrcReg = Registers[4][FPRIdx++];
RC = &AArch64::FPR64RegClass;
} else if (VT.is128BitVector()) {
SrcReg = Registers[5][FPRIdx++];
RC = &AArch64::FPR128RegClass;
} else
llvm_unreachable("Unexpected value type.");
[FastISel][AArch64] Implement the FastLowerArguments hook. This implements basic argument lowering for AArch64 in FastISel. It only handles a small subset of the C calling convention. It supports simple arguments that can be passed in GPR and FPR registers. This should cover most of the trivial cases without falling back to SelectionDAG. This fixes <rdar://problem/17890986>. llvm-svn: 214846
2014-08-05 13:43:48 +08:00
unsigned DstReg = FuncInfo.MF->addLiveIn(SrcReg, RC);
// FIXME: Unfortunately it's necessary to emit a copy from the livein copy.
// Without this, EmitLiveInCopies may eliminate the livein if its only
// use is a bitcast (which isn't turned into an instruction).
unsigned ResultReg = createResultReg(RC);
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc,
TII.get(TargetOpcode::COPY), ResultReg)
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
.addReg(DstReg, getKillRegState(true));
[FastISel] Rename public visible FastISel functions. NFC. This commit renames the following public FastISel functions: LowerArguments -> lowerArguments SelectInstruction -> selectInstruction TargetSelectInstruction -> fastSelectInstruction FastLowerArguments -> fastLowerArguments FastLowerCall -> fastLowerCall FastLowerIntrinsicCall -> fastLowerIntrinsicCall FastEmitZExtFromI1 -> fastEmitZExtFromI1 FastEmitBranch -> fastEmitBranch UpdateValueMap -> updateValueMap TargetMaterializeConstant -> fastMaterializeConstant TargetMaterializeAlloca -> fastMaterializeAlloca TargetMaterializeFloatZero -> fastMaterializeFloatZero LowerCallTo -> lowerCallTo Reviewed by Eric llvm-svn: 217074
2014-09-04 04:56:52 +08:00
updateValueMap(&Arg, ResultReg);
[FastISel][AArch64] Implement the FastLowerArguments hook. This implements basic argument lowering for AArch64 in FastISel. It only handles a small subset of the C calling convention. It supports simple arguments that can be passed in GPR and FPR registers. This should cover most of the trivial cases without falling back to SelectionDAG. This fixes <rdar://problem/17890986>. llvm-svn: 214846
2014-08-05 13:43:48 +08:00
}
return true;
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool AArch64FastISel::processCallArgs(CallLoweringInfo &CLI,
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
SmallVectorImpl<MVT> &OutVTs,
unsigned &NumBytes) {
CallingConv::ID CC = CLI.CallConv;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
SmallVector<CCValAssign, 16> ArgLocs;
Remove the target machine from CCState. Previously it was only used to get the subtarget and that's accessible from the MachineFunction now. This helps clear the way for smaller changes where we getting a subtarget will require passing in a MachineFunction/Function as well. llvm-svn: 214988
2014-08-07 02:45:26 +08:00
CCState CCInfo(CC, false, *FuncInfo.MF, ArgLocs, *Context);
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
CCInfo.AnalyzeCallOperands(OutVTs, CLI.OutFlags, CCAssignFnForCall(CC));
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Get a count of how many bytes are to be pushed on the stack.
NumBytes = CCInfo.getNextStackOffset();
// Issue CALLSEQ_START
unsigned AdjStackDown = TII.getCallFrameSetupOpcode();
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(AdjStackDown))
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
.addImm(NumBytes);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Process the args.
Convert some AArch64 code to foreach loops. NFC. Also converted a cast<> to dyn_cast while i was working on the same line of code. llvm-svn: 243894
2015-08-04 03:04:32 +08:00
for (CCValAssign &VA : ArgLocs) {
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
const Value *ArgVal = CLI.OutVals[VA.getValNo()];
MVT ArgVT = OutVTs[VA.getValNo()];
unsigned ArgReg = getRegForValue(ArgVal);
if (!ArgReg)
return false;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Handle arg promotion: SExt, ZExt, AExt.
switch (VA.getLocInfo()) {
case CCValAssign::Full:
break;
case CCValAssign::SExt: {
MVT DestVT = VA.getLocVT();
MVT SrcVT = ArgVT;
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
ArgReg = emitIntExt(SrcVT, ArgReg, DestVT, /*isZExt=*/false);
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
if (!ArgReg)
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
break;
}
case CCValAssign::AExt:
// Intentional fall-through.
case CCValAssign::ZExt: {
MVT DestVT = VA.getLocVT();
MVT SrcVT = ArgVT;
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
ArgReg = emitIntExt(SrcVT, ArgReg, DestVT, /*isZExt=*/true);
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
if (!ArgReg)
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
break;
}
default:
llvm_unreachable("Unknown arg promotion!");
}
// Now copy/store arg to correct locations.
if (VA.isRegLoc() && !VA.needsCustom()) {
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc,
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
TII.get(TargetOpcode::COPY), VA.getLocReg()).addReg(ArgReg);
CLI.OutRegs.push_back(VA.getLocReg());
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
} else if (VA.needsCustom()) {
// FIXME: Handle custom args.
return false;
} else {
assert(VA.isMemLoc() && "Assuming store on stack.");
[FastISel][AArch64 and X86] Don't emit stores for UNDEF arguments during function call lowering. UNDEF arguments are not ment to be touched - especially for the webkit_js calling convention. This fix reproduces the already existing behavior of SelectionDAG in FastISel. llvm-svn: 214366
2014-07-31 08:11:11 +08:00
// Don't emit stores for undef values.
if (isa<UndefValue>(ArgVal))
continue;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Need to store on the stack.
AArch64: mark small types (i1, i8, i16) as promoted This means the output of LowerFormalArguments returns a lowered SDValue with the correct type (expected in SelectionDAGBuilder). Without this, an assertion under a DEBUG macro triggers when those types are passed on the stack. llvm-svn: 210102
2014-06-03 21:54:53 +08:00
unsigned ArgSize = (ArgVT.getSizeInBits() + 7) / 8;
[ARM64-BE] Teach fast-isel about how to set up sub-word stack arguments for big endian calls. SelectionDAG already knows about this, but fast-isel was ignorant. llvm-svn: 208307
2014-05-08 20:53:50 +08:00
unsigned BEAlign = 0;
if (ArgSize < 8 && !Subtarget->isLittleEndian())
BEAlign = 8 - ArgSize;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
Address Addr;
Addr.setKind(Address::RegBase);
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
Addr.setReg(AArch64::SP);
[ARM64-BE] Teach fast-isel about how to set up sub-word stack arguments for big endian calls. SelectionDAG already knows about this, but fast-isel was ignorant. llvm-svn: 208307
2014-05-08 20:53:50 +08:00
Addr.setOffset(VA.getLocMemOffset() + BEAlign);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
[FastISel][AArch64] Attach MachineMemOperands to load and store instructions. llvm-svn: 215231
2014-08-09 01:24:10 +08:00
unsigned Alignment = DL.getABITypeAlignment(ArgVal->getType());
MachineMemOperand *MMO = FuncInfo.MF->getMachineMemOperand(
PseudoSourceValue: Replace global manager with a manager in a machine function. This commit removes the global manager variable which is responsible for storing and allocating pseudo source values and instead it introduces a new manager class named 'PseudoSourceValueManager'. Machine functions now own an instance of the pseudo source value manager class. This commit also modifies the 'get...' methods in the 'MachinePointerInfo' class to construct pseudo source values using the instance of the pseudo source value manager object from the machine function. This commit updates calls to the 'get...' methods from the 'MachinePointerInfo' class in a lot of different files because those calls now need to pass in a reference to a machine function to those methods. This change will make it easier to serialize pseudo source values as it will enable me to transform the mips specific MipsCallEntry PseudoSourceValue subclass into two target independent subclasses. Reviewers: Akira Hatanaka llvm-svn: 244693
2015-08-12 07:09:45 +08:00
MachinePointerInfo::getStack(*FuncInfo.MF, Addr.getOffset()),
MachineMemOperand::MOStore, ArgVT.getStoreSize(), Alignment);
[FastISel][AArch64] Attach MachineMemOperands to load and store instructions. llvm-svn: 215231
2014-08-09 01:24:10 +08:00
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
if (!emitStore(ArgVT, ArgReg, Addr, MMO))
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
}
}
return true;
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool AArch64FastISel::finishCall(CallLoweringInfo &CLI, MVT RetVT,
[FastISel][AArch64] Fix return type in FastLowerCall. I used the wrong method to obtain the return type inside FinishCall. This fix simply uses the return type from FastLowerCall, which we already determined to be a valid type. Reduced test case from Chad. Thanks. llvm-svn: 213788
2014-07-24 04:03:13 +08:00
unsigned NumBytes) {
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
CallingConv::ID CC = CLI.CallConv;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Issue CALLSEQ_END
unsigned AdjStackUp = TII.getCallFrameDestroyOpcode();
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(AdjStackUp))
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
.addImm(NumBytes).addImm(0);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Now the return value.
if (RetVT != MVT::isVoid) {
SmallVector<CCValAssign, 16> RVLocs;
Remove the target machine from CCState. Previously it was only used to get the subtarget and that's accessible from the MachineFunction now. This helps clear the way for smaller changes where we getting a subtarget will require passing in a MachineFunction/Function as well. llvm-svn: 214988
2014-08-07 02:45:26 +08:00
CCState CCInfo(CC, false, *FuncInfo.MF, RVLocs, *Context);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
CCInfo.AnalyzeCallResult(RetVT, CCAssignFnForCall(CC));
// Only handle a single return value.
if (RVLocs.size() != 1)
return false;
// Copy all of the result registers out of their specified physreg.
MVT CopyVT = RVLocs[0].getValVT();
Disable AArch64 fast-isel on big-endian call vector returns. A big-endian vector return needs a byte-swap which we aren't doing right now. For now just bail on these cases to get correctness back. llvm-svn: 235133
2015-04-17 05:19:36 +08:00
// TODO: Handle big-endian results
if (CopyVT.isVector() && !Subtarget->isLittleEndian())
return false;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
unsigned ResultReg = createResultReg(TLI.getRegClassFor(CopyVT));
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc,
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
TII.get(TargetOpcode::COPY), ResultReg)
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
.addReg(RVLocs[0].getLocReg());
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
CLI.InRegs.push_back(RVLocs[0].getLocReg());
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
CLI.ResultReg = ResultReg;
CLI.NumResultRegs = 1;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
}
return true;
}
[FastISel] Rename public visible FastISel functions. NFC. This commit renames the following public FastISel functions: LowerArguments -> lowerArguments SelectInstruction -> selectInstruction TargetSelectInstruction -> fastSelectInstruction FastLowerArguments -> fastLowerArguments FastLowerCall -> fastLowerCall FastLowerIntrinsicCall -> fastLowerIntrinsicCall FastEmitZExtFromI1 -> fastEmitZExtFromI1 FastEmitBranch -> fastEmitBranch UpdateValueMap -> updateValueMap TargetMaterializeConstant -> fastMaterializeConstant TargetMaterializeAlloca -> fastMaterializeAlloca TargetMaterializeFloatZero -> fastMaterializeFloatZero LowerCallTo -> lowerCallTo Reviewed by Eric llvm-svn: 217074
2014-09-04 04:56:52 +08:00
bool AArch64FastISel::fastLowerCall(CallLoweringInfo &CLI) {
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
CallingConv::ID CC = CLI.CallConv;
[AArch64, fast-isel] Fall back to SelectionDAG to select tail calls. Certain functions such as objc_autoreleaseReturnValue have to be called as tail-calls even at -O0. Since normal fast-isel doesn't emit calls as tail calls, we have to fall back to SelectionDAG to select calls that are marked as tail. <rdar://problem/17991614> llvm-svn: 215600
2014-08-14 07:23:58 +08:00
bool IsTailCall = CLI.IsTailCall;
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
bool IsVarArg = CLI.IsVarArg;
const Value *Callee = CLI.Callee;
Use MCSymbols for FastISel. The summary is that it moves the mangling earlier and replaces a few calls to .addExternalSymbol with addSym. I originally wanted to replace all the uses of addExternalSymbol with addSym, but noticed it was a lot of work and doesn't need to be done all at once. llvm-svn: 240395
2015-06-23 20:21:54 +08:00
MCSymbol *Symbol = CLI.Symbol;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
Use MCSymbols for FastISel. The summary is that it moves the mangling earlier and replaces a few calls to .addExternalSymbol with addSym. I originally wanted to replace all the uses of addExternalSymbol with addSym, but noticed it was a lot of work and doesn't need to be done all at once. llvm-svn: 240395
2015-06-23 20:21:54 +08:00
if (!Callee && !Symbol)
[FastISel][AArch64] Add lowering support for frem. This lowers frem to a runtime libcall inside fast-isel. The test case also checks the CallLoweringInfo bug that was exposed by this change. This fixes rdar://problem/18342783. llvm-svn: 217833
2014-09-16 06:07:49 +08:00
return false;
[AArch64, fast-isel] Fall back to SelectionDAG to select tail calls. Certain functions such as objc_autoreleaseReturnValue have to be called as tail-calls even at -O0. Since normal fast-isel doesn't emit calls as tail calls, we have to fall back to SelectionDAG to select calls that are marked as tail. <rdar://problem/17991614> llvm-svn: 215600
2014-08-14 07:23:58 +08:00
// Allow SelectionDAG isel to handle tail calls.
if (IsTailCall)
return false;
[FastISel][AArch64] Add MachO large code model support for function calls. Currently the large code model for MachO uses the GOT to make function calls. Emit the required adrp and ldr instructions to load the address from the GOT. Related to <rdar://problem/17733076>. llvm-svn: 214381
2014-07-31 12:10:40 +08:00
CodeModel::Model CM = TM.getCodeModel();
// Only support the small and large code model.
if (CM != CodeModel::Small && CM != CodeModel::Large)
return false;
// FIXME: Add large code model support for ELF.
if (CM == CodeModel::Large && !Subtarget->isTargetMachO())
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
// Let SDISel handle vararg functions.
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
if (IsVarArg)
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
// FIXME: Only handle *simple* calls for now.
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
MVT RetVT;
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
if (CLI.RetTy->isVoidTy())
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
RetVT = MVT::isVoid;
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
else if (!isTypeLegal(CLI.RetTy, RetVT))
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
for (auto Flag : CLI.OutFlags)
Swift Calling Convention: add swiftself attribute. Differential Revision: http://reviews.llvm.org/D17866 llvm-svn: 264754
2016-03-30 01:37:21 +08:00
if (Flag.isInReg() || Flag.isSRet() || Flag.isNest() || Flag.isByVal() ||
Swift Calling Convention: swifterror target support. Differential Revision: http://reviews.llvm.org/D18716 llvm-svn: 265997
2016-04-12 05:08:06 +08:00
Flag.isSwiftSelf() || Flag.isSwiftError())
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
// Set up the argument vectors.
SmallVector<MVT, 16> OutVTs;
OutVTs.reserve(CLI.OutVals.size());
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
for (auto *Val : CLI.OutVals) {
MVT VT;
if (!isTypeLegal(Val->getType(), VT) &&
!(VT == MVT::i1 || VT == MVT::i8 || VT == MVT::i16))
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
// We don't handle vector parameters yet.
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
if (VT.isVector() || VT.getSizeInBits() > 64)
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
OutVTs.push_back(VT);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
}
[FastISel][AArch64] Add MachO large code model support for function calls. Currently the large code model for MachO uses the GOT to make function calls. Emit the required adrp and ldr instructions to load the address from the GOT. Related to <rdar://problem/17733076>. llvm-svn: 214381
2014-07-31 12:10:40 +08:00
Address Addr;
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
if (Callee && !computeCallAddress(Callee, Addr))
[FastISel][AArch64] Add MachO large code model support for function calls. Currently the large code model for MachO uses the GOT to make function calls. Emit the required adrp and ldr instructions to load the address from the GOT. Related to <rdar://problem/17733076>. llvm-svn: 214381
2014-07-31 12:10:40 +08:00
return false;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Handle the arguments now that we've gotten them.
unsigned NumBytes;
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
if (!processCallArgs(CLI, OutVTs, NumBytes))
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
// Issue the call.
MachineInstrBuilder MIB;
[FastISel][AArch64] Add MachO large code model support for function calls. Currently the large code model for MachO uses the GOT to make function calls. Emit the required adrp and ldr instructions to load the address from the GOT. Related to <rdar://problem/17733076>. llvm-svn: 214381
2014-07-31 12:10:40 +08:00
if (CM == CodeModel::Small) {
[FastISel][AArch64] Use the correct register class for branches. Also constrain the register class for branches. This fixes rdar://problem/18181496. llvm-svn: 216804
2014-08-30 07:48:06 +08:00
const MCInstrDesc &II = TII.get(Addr.getReg() ? AArch64::BLR : AArch64::BL);
MIB = BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, II);
Use MCSymbols for FastISel. The summary is that it moves the mangling earlier and replaces a few calls to .addExternalSymbol with addSym. I originally wanted to replace all the uses of addExternalSymbol with addSym, but noticed it was a lot of work and doesn't need to be done all at once. llvm-svn: 240395
2015-06-23 20:21:54 +08:00
if (Symbol)
MIB.addSym(Symbol, 0);
[FastISel][AArch64] Add MachO large code model support for function calls. Currently the large code model for MachO uses the GOT to make function calls. Emit the required adrp and ldr instructions to load the address from the GOT. Related to <rdar://problem/17733076>. llvm-svn: 214381
2014-07-31 12:10:40 +08:00
else if (Addr.getGlobalValue())
MIB.addGlobalAddress(Addr.getGlobalValue(), 0, 0);
[FastISel][AArch64] Use the correct register class for branches. Also constrain the register class for branches. This fixes rdar://problem/18181496. llvm-svn: 216804
2014-08-30 07:48:06 +08:00
else if (Addr.getReg()) {
unsigned Reg = constrainOperandRegClass(II, Addr.getReg(), 0);
MIB.addReg(Reg);
} else
[FastISel][AArch64] Add MachO large code model support for function calls. Currently the large code model for MachO uses the GOT to make function calls. Emit the required adrp and ldr instructions to load the address from the GOT. Related to <rdar://problem/17733076>. llvm-svn: 214381
2014-07-31 12:10:40 +08:00
return false;
} else {
unsigned CallReg = 0;
Use MCSymbols for FastISel. The summary is that it moves the mangling earlier and replaces a few calls to .addExternalSymbol with addSym. I originally wanted to replace all the uses of addExternalSymbol with addSym, but noticed it was a lot of work and doesn't need to be done all at once. llvm-svn: 240395
2015-06-23 20:21:54 +08:00
if (Symbol) {
[FastISel][AArch64] Add MachO large code model support for function calls. Currently the large code model for MachO uses the GOT to make function calls. Emit the required adrp and ldr instructions to load the address from the GOT. Related to <rdar://problem/17733076>. llvm-svn: 214381
2014-07-31 12:10:40 +08:00
unsigned ADRPReg = createResultReg(&AArch64::GPR64commonRegClass);
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(AArch64::ADRP),
ADRPReg)
Use MCSymbols for FastISel. The summary is that it moves the mangling earlier and replaces a few calls to .addExternalSymbol with addSym. I originally wanted to replace all the uses of addExternalSymbol with addSym, but noticed it was a lot of work and doesn't need to be done all at once. llvm-svn: 240395
2015-06-23 20:21:54 +08:00
.addSym(Symbol, AArch64II::MO_GOT | AArch64II::MO_PAGE);
[FastISel][AArch64] Add MachO large code model support for function calls. Currently the large code model for MachO uses the GOT to make function calls. Emit the required adrp and ldr instructions to load the address from the GOT. Related to <rdar://problem/17733076>. llvm-svn: 214381
2014-07-31 12:10:40 +08:00
CallReg = createResultReg(&AArch64::GPR64RegClass);
Use MCSymbols for FastISel. The summary is that it moves the mangling earlier and replaces a few calls to .addExternalSymbol with addSym. I originally wanted to replace all the uses of addExternalSymbol with addSym, but noticed it was a lot of work and doesn't need to be done all at once. llvm-svn: 240395
2015-06-23 20:21:54 +08:00
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc,
TII.get(AArch64::LDRXui), CallReg)
.addReg(ADRPReg)
.addSym(Symbol,
AArch64II::MO_GOT | AArch64II::MO_PAGEOFF | AArch64II::MO_NC);
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
} else if (Addr.getGlobalValue())
CallReg = materializeGV(Addr.getGlobalValue());
else if (Addr.getReg())
[FastISel][AArch64] Add MachO large code model support for function calls. Currently the large code model for MachO uses the GOT to make function calls. Emit the required adrp and ldr instructions to load the address from the GOT. Related to <rdar://problem/17733076>. llvm-svn: 214381
2014-07-31 12:10:40 +08:00
CallReg = Addr.getReg();
if (!CallReg)
return false;
[FastISel][AArch64] Use the correct register class for branches. Also constrain the register class for branches. This fixes rdar://problem/18181496. llvm-svn: 216804
2014-08-30 07:48:06 +08:00
const MCInstrDesc &II = TII.get(AArch64::BLR);
CallReg = constrainOperandRegClass(II, CallReg, 0);
MIB = BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, II).addReg(CallReg);
[FastISel][AArch64] Add MachO large code model support for function calls. Currently the large code model for MachO uses the GOT to make function calls. Emit the required adrp and ldr instructions to load the address from the GOT. Related to <rdar://problem/17733076>. llvm-svn: 214381
2014-07-31 12:10:40 +08:00
}
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Add implicit physical register uses to the call.
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
for (auto Reg : CLI.OutRegs)
MIB.addReg(Reg, RegState::Implicit);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Add a register mask with the call-preserved registers.
// Proper defs for return values will be added by setPhysRegsDeadExcept().
Have getCallPreservedMask and getThisCallPreservedMask take a MachineFunction argument so that we can grab subtarget specific features off of it. llvm-svn: 231979
2015-03-12 06:42:13 +08:00
MIB.addRegMask(TRI.getCallPreservedMask(*FuncInfo.MF, CC));
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
[FastISel][AArch64] Add MachO large code model support for function calls. Currently the large code model for MachO uses the GOT to make function calls. Emit the required adrp and ldr instructions to load the address from the GOT. Related to <rdar://problem/17733076>. llvm-svn: 214381
2014-07-31 12:10:40 +08:00
CLI.Call = MIB;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Finish off the call including any return values.
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return finishCall(CLI, RetVT, NumBytes);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool AArch64FastISel::isMemCpySmall(uint64_t Len, unsigned Alignment) {
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
if (Alignment)
return Len / Alignment <= 4;
else
return Len < 32;
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool AArch64FastISel::tryEmitSmallMemCpy(Address Dest, Address Src,
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
uint64_t Len, unsigned Alignment) {
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Make sure we don't bloat code by inlining very large memcpy's.
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
if (!isMemCpySmall(Len, Alignment))
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
int64_t UnscaledOffset = 0;
Address OrigDest = Dest;
Address OrigSrc = Src;
while (Len) {
MVT VT;
if (!Alignment || Alignment >= 8) {
if (Len >= 8)
VT = MVT::i64;
else if (Len >= 4)
VT = MVT::i32;
else if (Len >= 2)
VT = MVT::i16;
else {
VT = MVT::i8;
}
} else {
// Bound based on alignment.
if (Len >= 4 && Alignment == 4)
VT = MVT::i32;
else if (Len >= 2 && Alignment == 2)
VT = MVT::i16;
else {
VT = MVT::i8;
}
}
[FastISel][AArch64] Fix sign-/zero-extend folding when SelectionDAG is involved. Sign-/zero-extend folding depended on the load and the integer extend to be both selected by FastISel. This cannot always be garantueed and SelectionDAG might interfer. This commit adds additonal checks to load and integer extend lowering to catch this. Related to rdar://problem/18495928. llvm-svn: 219716
2014-10-15 04:36:02 +08:00
unsigned ResultReg = emitLoad(VT, VT, Src);
if (!ResultReg)
AArch64: make FastISel memcpy emission more robust. We were hitting an assert if FastISel couldn't create the load or store we requested. Currently this happens for large frame-local addresses, though CodeGen could be improved there. rdar://problem/17187463 llvm-svn: 210519
2014-06-10 17:52:40 +08:00
return false;
[FastISel][AArch64] Fix sign-/zero-extend folding when SelectionDAG is involved. Sign-/zero-extend folding depended on the load and the integer extend to be both selected by FastISel. This cannot always be garantueed and SelectionDAG might interfer. This commit adds additonal checks to load and integer extend lowering to catch this. Related to rdar://problem/18495928. llvm-svn: 219716
2014-10-15 04:36:02 +08:00
if (!emitStore(VT, ResultReg, Dest))
AArch64: make FastISel memcpy emission more robust. We were hitting an assert if FastISel couldn't create the load or store we requested. Currently this happens for large frame-local addresses, though CodeGen could be improved there. rdar://problem/17187463 llvm-svn: 210519
2014-06-10 17:52:40 +08:00
return false;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
int64_t Size = VT.getSizeInBits() / 8;
Len -= Size;
UnscaledOffset += Size;
// We need to recompute the unscaled offset for each iteration.
Dest.setOffset(OrigDest.getOffset() + UnscaledOffset);
Src.setOffset(OrigSrc.getOffset() + UnscaledOffset);
}
return true;
}
[FastISel][AArch64] Add branch folding support for the XALU intrinsics. This improves the code generation for the XALU intrinsics when the condition is feeding a branch instruction. This is related to <rdar://problem/17831117>. llvm-svn: 214349
2014-07-31 06:04:34 +08:00
/// \brief Check if it is possible to fold the condition from the XALU intrinsic
/// into the user. The condition code will only be updated on success.
bool AArch64FastISel::foldXALUIntrinsic(AArch64CC::CondCode &CC,
const Instruction *I,
const Value *Cond) {
if (!isa<ExtractValueInst>(Cond))
return false;
const auto *EV = cast<ExtractValueInst>(Cond);
if (!isa<IntrinsicInst>(EV->getAggregateOperand()))
return false;
const auto *II = cast<IntrinsicInst>(EV->getAggregateOperand());
MVT RetVT;
const Function *Callee = II->getCalledFunction();
Type *RetTy =
cast<StructType>(Callee->getReturnType())->getTypeAtIndex(0U);
if (!isTypeLegal(RetTy, RetVT))
return false;
if (RetVT != MVT::i32 && RetVT != MVT::i64)
return false;
Fix previous commit: [FastISel][AArch64] Simplify XALU multiplies. When folding the intrinsic flag into the branch or select we also have to consider the fact if the intrinsic got simplified, because it changes the flag we have to check for. llvm-svn: 218034
2014-09-18 15:26:26 +08:00
const Value *LHS = II->getArgOperand(0);
const Value *RHS = II->getArgOperand(1);
// Canonicalize immediate to the RHS.
if (isa<ConstantInt>(LHS) && !isa<ConstantInt>(RHS) &&
isCommutativeIntrinsic(II))
std::swap(LHS, RHS);
// Simplify multiplies.
Change Function::getIntrinsicID() to return an Intrinsic::ID. NFC. Now that Intrinsic::ID is a typed enum, we can forward declare it and so return it from this method. This updates all users which were either using an unsigned to store it, or had a now unnecessary cast. llvm-svn: 237810
2015-05-21 01:16:39 +08:00
Intrinsic::ID IID = II->getIntrinsicID();
Fix previous commit: [FastISel][AArch64] Simplify XALU multiplies. When folding the intrinsic flag into the branch or select we also have to consider the fact if the intrinsic got simplified, because it changes the flag we have to check for. llvm-svn: 218034
2014-09-18 15:26:26 +08:00
switch (IID) {
default:
break;
case Intrinsic::smul_with_overflow:
if (const auto *C = dyn_cast<ConstantInt>(RHS))
if (C->getValue() == 2)
IID = Intrinsic::sadd_with_overflow;
break;
case Intrinsic::umul_with_overflow:
if (const auto *C = dyn_cast<ConstantInt>(RHS))
if (C->getValue() == 2)
IID = Intrinsic::uadd_with_overflow;
break;
}
[FastISel][AArch64] Add branch folding support for the XALU intrinsics. This improves the code generation for the XALU intrinsics when the condition is feeding a branch instruction. This is related to <rdar://problem/17831117>. llvm-svn: 214349
2014-07-31 06:04:34 +08:00
AArch64CC::CondCode TmpCC;
Fix previous commit: [FastISel][AArch64] Simplify XALU multiplies. When folding the intrinsic flag into the branch or select we also have to consider the fact if the intrinsic got simplified, because it changes the flag we have to check for. llvm-svn: 218034
2014-09-18 15:26:26 +08:00
switch (IID) {
default:
return false;
case Intrinsic::sadd_with_overflow:
case Intrinsic::ssub_with_overflow:
TmpCC = AArch64CC::VS;
break;
case Intrinsic::uadd_with_overflow:
TmpCC = AArch64CC::HS;
break;
case Intrinsic::usub_with_overflow:
TmpCC = AArch64CC::LO;
break;
case Intrinsic::smul_with_overflow:
case Intrinsic::umul_with_overflow:
TmpCC = AArch64CC::NE;
break;
[FastISel][AArch64] Add branch folding support for the XALU intrinsics. This improves the code generation for the XALU intrinsics when the condition is feeding a branch instruction. This is related to <rdar://problem/17831117>. llvm-svn: 214349
2014-07-31 06:04:34 +08:00
}
// Check if both instructions are in the same basic block.
[FastISel][AArch64] Improve branch selection to support all FP conditions. This adds the last two missing floating-point condition codes (FCMP_UEQ and FCMP_ONE) also to the branch selection. In these two cases an additonal branch instruction is required. This also adds unit tests to checks all the different condition codes. This is related o rdar://problem/18358882. llvm-svn: 217966
2014-09-18 01:46:47 +08:00
if (!isValueAvailable(II))
[FastISel][AArch64] Add branch folding support for the XALU intrinsics. This improves the code generation for the XALU intrinsics when the condition is feeding a branch instruction. This is related to <rdar://problem/17831117>. llvm-svn: 214349
2014-07-31 06:04:34 +08:00
return false;
// Make sure nothing is in the way
AArch64: Remove implicit ilist iterator conversions, NFC llvm-svn: 250216
2015-10-14 04:02:15 +08:00
BasicBlock::const_iterator Start(I);
BasicBlock::const_iterator End(II);
[FastISel][AArch64] Add branch folding support for the XALU intrinsics. This improves the code generation for the XALU intrinsics when the condition is feeding a branch instruction. This is related to <rdar://problem/17831117>. llvm-svn: 214349
2014-07-31 06:04:34 +08:00
for (auto Itr = std::prev(Start); Itr != End; --Itr) {
// We only expect extractvalue instructions between the intrinsic and the
// instruction to be selected.
if (!isa<ExtractValueInst>(Itr))
return false;
// Check that the extractvalue operand comes from the intrinsic.
const auto *EVI = cast<ExtractValueInst>(Itr);
if (EVI->getAggregateOperand() != II)
return false;
}
CC = TmpCC;
return true;
}
[FastISel] Rename public visible FastISel functions. NFC. This commit renames the following public FastISel functions: LowerArguments -> lowerArguments SelectInstruction -> selectInstruction TargetSelectInstruction -> fastSelectInstruction FastLowerArguments -> fastLowerArguments FastLowerCall -> fastLowerCall FastLowerIntrinsicCall -> fastLowerIntrinsicCall FastEmitZExtFromI1 -> fastEmitZExtFromI1 FastEmitBranch -> fastEmitBranch UpdateValueMap -> updateValueMap TargetMaterializeConstant -> fastMaterializeConstant TargetMaterializeAlloca -> fastMaterializeAlloca TargetMaterializeFloatZero -> fastMaterializeFloatZero LowerCallTo -> lowerCallTo Reviewed by Eric llvm-svn: 217074
2014-09-04 04:56:52 +08:00
bool AArch64FastISel::fastLowerIntrinsicCall(const IntrinsicInst *II) {
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// FIXME: Handle more intrinsics.
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
switch (II->getIntrinsicID()) {
[FastISel][AArch64] Add support for frameaddress intrinsic. This commit implements the frameaddress intrinsic for the AArch64 architecture in FastISel. There were two test cases that pretty much tested the same, so I combined them to a single test case. Fixes <rdar://problem/17811834> llvm-svn: 213959
2014-07-26 01:47:14 +08:00
default: return false;
case Intrinsic::frameaddress: {
MachineFunction: Return reference for getFrameInfo(); NFC getFrameInfo() never returns nullptr so we should use a reference instead of a pointer. llvm-svn: 277017
2016-07-29 02:40:00 +08:00
MachineFrameInfo &MFI = FuncInfo.MF->getFrameInfo();
MFI.setFrameAddressIsTaken(true);
[FastISel][AArch64] Add support for frameaddress intrinsic. This commit implements the frameaddress intrinsic for the AArch64 architecture in FastISel. There were two test cases that pretty much tested the same, so I combined them to a single test case. Fixes <rdar://problem/17811834> llvm-svn: 213959
2014-07-26 01:47:14 +08:00
const AArch64RegisterInfo *RegInfo =
Clean up some uses of getSubtarget in AArch64. llvm-svn: 227530
2015-01-30 09:10:24 +08:00
static_cast<const AArch64RegisterInfo *>(Subtarget->getRegisterInfo());
[FastISel][AArch64] Add support for frameaddress intrinsic. This commit implements the frameaddress intrinsic for the AArch64 architecture in FastISel. There were two test cases that pretty much tested the same, so I combined them to a single test case. Fixes <rdar://problem/17811834> llvm-svn: 213959
2014-07-26 01:47:14 +08:00
unsigned FramePtr = RegInfo->getFrameRegister(*(FuncInfo.MF));
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
unsigned SrcReg = MRI.createVirtualRegister(&AArch64::GPR64RegClass);
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc,
TII.get(TargetOpcode::COPY), SrcReg).addReg(FramePtr);
[FastISel][AArch64] Add support for frameaddress intrinsic. This commit implements the frameaddress intrinsic for the AArch64 architecture in FastISel. There were two test cases that pretty much tested the same, so I combined them to a single test case. Fixes <rdar://problem/17811834> llvm-svn: 213959
2014-07-26 01:47:14 +08:00
// Recursively load frame address
// ldr x0, [fp]
// ldr x0, [x0]
// ldr x0, [x0]
// ...
unsigned DestReg;
unsigned Depth = cast<ConstantInt>(II->getOperand(0))->getZExtValue();
while (Depth--) {
[FastISel][tblgen] Rename tblgen generated FastISel functions. NFC. This is the final round of renaming. This changes tblgen to emit lower-case function names for FastEmitInst_* and FastEmit_*, and updates all its uses in the source code. Reviewed by Eric llvm-svn: 217075
2014-09-04 04:56:59 +08:00
DestReg = fastEmitInst_ri(AArch64::LDRXui, &AArch64::GPR64RegClass,
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
SrcReg, /*IsKill=*/true, 0);
assert(DestReg && "Unexpected LDR instruction emission failure.");
[FastISel][AArch64] Add support for frameaddress intrinsic. This commit implements the frameaddress intrinsic for the AArch64 architecture in FastISel. There were two test cases that pretty much tested the same, so I combined them to a single test case. Fixes <rdar://problem/17811834> llvm-svn: 213959
2014-07-26 01:47:14 +08:00
SrcReg = DestReg;
}
[FastISel] Rename public visible FastISel functions. NFC. This commit renames the following public FastISel functions: LowerArguments -> lowerArguments SelectInstruction -> selectInstruction TargetSelectInstruction -> fastSelectInstruction FastLowerArguments -> fastLowerArguments FastLowerCall -> fastLowerCall FastLowerIntrinsicCall -> fastLowerIntrinsicCall FastEmitZExtFromI1 -> fastEmitZExtFromI1 FastEmitBranch -> fastEmitBranch UpdateValueMap -> updateValueMap TargetMaterializeConstant -> fastMaterializeConstant TargetMaterializeAlloca -> fastMaterializeAlloca TargetMaterializeFloatZero -> fastMaterializeFloatZero LowerCallTo -> lowerCallTo Reviewed by Eric llvm-svn: 217074
2014-09-04 04:56:52 +08:00
updateValueMap(II, SrcReg);
[FastISel][AArch64] Add support for frameaddress intrinsic. This commit implements the frameaddress intrinsic for the AArch64 architecture in FastISel. There were two test cases that pretty much tested the same, so I combined them to a single test case. Fixes <rdar://problem/17811834> llvm-svn: 213959
2014-07-26 01:47:14 +08:00
return true;
}
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
case Intrinsic::memcpy:
case Intrinsic::memmove: {
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
const auto *MTI = cast<MemTransferInst>(II);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Don't handle volatile.
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
if (MTI->isVolatile())
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
Revert "[FastISel][AArch64] Don't fold instructions too aggressively into the memory operation." Quentin pointed out that this is not the correct approach and there is a better and easier solution. llvm-svn: 216632
2014-08-28 07:09:40 +08:00
// Disable inlining for memmove before calls to ComputeAddress. Otherwise,
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// we would emit dead code because we don't currently handle memmoves.
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
bool IsMemCpy = (II->getIntrinsicID() == Intrinsic::memcpy);
if (isa<ConstantInt>(MTI->getLength()) && IsMemCpy) {
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Small memcpy's are common enough that we want to do them without a call
// if possible.
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
uint64_t Len = cast<ConstantInt>(MTI->getLength())->getZExtValue();
Revert "Change memcpy/memset/memmove to have dest and source alignments." This reverts commit r253511. This likely broke the bots in http://lab.llvm.org:8011/builders/clang-ppc64-elf-linux2/builds/20202 http://bb.pgr.jp/builders/clang-3stage-i686-linux/builds/3787 llvm-svn: 253543
2015-11-19 13:56:52 +08:00
unsigned Alignment = MTI->getAlignment();
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
if (isMemCpySmall(Len, Alignment)) {
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
Address Dest, Src;
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
if (!computeAddress(MTI->getRawDest(), Dest) ||
!computeAddress(MTI->getRawSource(), Src))
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
if (tryEmitSmallMemCpy(Dest, Src, Len, Alignment))
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return true;
}
}
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
if (!MTI->getLength()->getType()->isIntegerTy(64))
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
if (MTI->getSourceAddressSpace() > 255 || MTI->getDestAddressSpace() > 255)
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Fast instruction selection doesn't support the special
// address spaces.
return false;
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
const char *IntrMemName = isa<MemCpyInst>(II) ? "memcpy" : "memmove";
Revert "Change memcpy/memset/memmove to have dest and source alignments." This reverts commit r253511. This likely broke the bots in http://lab.llvm.org:8011/builders/clang-ppc64-elf-linux2/builds/20202 http://bb.pgr.jp/builders/clang-3stage-i686-linux/builds/3787 llvm-svn: 253543
2015-11-19 13:56:52 +08:00
return lowerCallTo(II, IntrMemName, II->getNumArgOperands() - 2);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
}
case Intrinsic::memset: {
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
const MemSetInst *MSI = cast<MemSetInst>(II);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Don't handle volatile.
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
if (MSI->isVolatile())
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
if (!MSI->getLength()->getType()->isIntegerTy(64))
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
[FastIsel][AArch64] Add support for the FastLowerCall and FastLowerIntrinsicCall target-hooks. This commit modifies the existing call lowering functions to be used as the FastLowerCall and FastLowerIntrinsicCall target-hooks instead. This enables patchpoint intrinsic lowering for AArch64. This fixes <rdar://problem/17733076> llvm-svn: 213704
2014-07-23 07:14:58 +08:00
if (MSI->getDestAddressSpace() > 255)
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Fast instruction selection doesn't support the special
// address spaces.
return false;
Revert "Change memcpy/memset/memmove to have dest and source alignments." This reverts commit r253511. This likely broke the bots in http://lab.llvm.org:8011/builders/clang-ppc64-elf-linux2/builds/20202 http://bb.pgr.jp/builders/clang-3stage-i686-linux/builds/3787 llvm-svn: 253543
2015-11-19 13:56:52 +08:00
return lowerCallTo(II, "memset", II->getNumArgOperands() - 2);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
}
[FastISel][AArch64] Lower sin/cos/pow to runtime lib calls. Also lower sin/cos/pow to runtime lib calls. This fixes rdar://problem/18343468. llvm-svn: 217839
2014-09-16 06:33:06 +08:00
case Intrinsic::sin:
case Intrinsic::cos:
case Intrinsic::pow: {
MVT RetVT;
if (!isTypeLegal(II->getType(), RetVT))
return false;
if (RetVT != MVT::f32 && RetVT != MVT::f64)
return false;
static const RTLIB::Libcall LibCallTable[3][2] = {
{ RTLIB::SIN_F32, RTLIB::SIN_F64 },
{ RTLIB::COS_F32, RTLIB::COS_F64 },
{ RTLIB::POW_F32, RTLIB::POW_F64 }
};
RTLIB::Libcall LC;
bool Is64Bit = RetVT == MVT::f64;
switch (II->getIntrinsicID()) {
default:
llvm_unreachable("Unexpected intrinsic.");
case Intrinsic::sin:
LC = LibCallTable[0][Is64Bit];
break;
case Intrinsic::cos:
LC = LibCallTable[1][Is64Bit];
break;
case Intrinsic::pow:
LC = LibCallTable[2][Is64Bit];
break;
}
ArgListTy Args;
Args.reserve(II->getNumArgOperands());
// Populate the argument list.
for (auto &Arg : II->arg_operands()) {
ArgListEntry Entry;
Entry.Val = Arg;
Entry.Ty = Arg->getType();
Args.push_back(Entry);
}
CallLoweringInfo CLI;
Use MCSymbols for FastISel. The summary is that it moves the mangling earlier and replaces a few calls to .addExternalSymbol with addSym. I originally wanted to replace all the uses of addExternalSymbol with addSym, but noticed it was a lot of work and doesn't need to be done all at once. llvm-svn: 240395
2015-06-23 20:21:54 +08:00
MCContext &Ctx = MF->getContext();
CLI.setCallee(DL, Ctx, TLI.getLibcallCallingConv(LC), II->getType(),
[FastISel][AArch64] Lower sin/cos/pow to runtime lib calls. Also lower sin/cos/pow to runtime lib calls. This fixes rdar://problem/18343468. llvm-svn: 217839
2014-09-16 06:33:06 +08:00
TLI.getLibcallName(LC), std::move(Args));
if (!lowerCallTo(CLI))
return false;
updateValueMap(II, CLI.ResultReg);
return true;
}
[FastISel][AArch64] Add support for fabs intrinsic. Lower the llvm.fabs intrinsic to the 'fabs' MI instruction. This fixes rdar://problem/18946552. llvm-svn: 221729
2014-11-12 07:10:44 +08:00
case Intrinsic::fabs: {
MVT VT;
if (!isTypeLegal(II->getType(), VT))
return false;
unsigned Opc;
switch (VT.SimpleTy) {
default:
return false;
case MVT::f32:
Opc = AArch64::FABSSr;
break;
case MVT::f64:
Opc = AArch64::FABSDr;
break;
}
unsigned SrcReg = getRegForValue(II->getOperand(0));
if (!SrcReg)
return false;
bool SrcRegIsKill = hasTrivialKill(II->getOperand(0));
unsigned ResultReg = createResultReg(TLI.getRegClassFor(VT));
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(Opc), ResultReg)
.addReg(SrcReg, getKillRegState(SrcRegIsKill));
updateValueMap(II, ResultReg);
return true;
}
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
case Intrinsic::trap: {
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(AArch64::BRK))
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
.addImm(1);
return true;
}
[FastISel][AArch64] Add sqrt intrinsic support. Fixes <rdar://problem/17867067>. llvm-svn: 214388
2014-07-31 14:25:33 +08:00
case Intrinsic::sqrt: {
Type *RetTy = II->getCalledFunction()->getReturnType();
MVT VT;
if (!isTypeLegal(RetTy, VT))
return false;
unsigned Op0Reg = getRegForValue(II->getOperand(0));
if (!Op0Reg)
return false;
bool Op0IsKill = hasTrivialKill(II->getOperand(0));
[FastISel][tblgen] Rename tblgen generated FastISel functions. NFC. This is the final round of renaming. This changes tblgen to emit lower-case function names for FastEmitInst_* and FastEmit_*, and updates all its uses in the source code. Reviewed by Eric llvm-svn: 217075
2014-09-04 04:56:59 +08:00
unsigned ResultReg = fastEmit_r(VT, VT, ISD::FSQRT, Op0Reg, Op0IsKill);
[FastISel][AArch64] Add sqrt intrinsic support. Fixes <rdar://problem/17867067>. llvm-svn: 214388
2014-07-31 14:25:33 +08:00
if (!ResultReg)
return false;
[FastISel] Rename public visible FastISel functions. NFC. This commit renames the following public FastISel functions: LowerArguments -> lowerArguments SelectInstruction -> selectInstruction TargetSelectInstruction -> fastSelectInstruction FastLowerArguments -> fastLowerArguments FastLowerCall -> fastLowerCall FastLowerIntrinsicCall -> fastLowerIntrinsicCall FastEmitZExtFromI1 -> fastEmitZExtFromI1 FastEmitBranch -> fastEmitBranch UpdateValueMap -> updateValueMap TargetMaterializeConstant -> fastMaterializeConstant TargetMaterializeAlloca -> fastMaterializeAlloca TargetMaterializeFloatZero -> fastMaterializeFloatZero LowerCallTo -> lowerCallTo Reviewed by Eric llvm-svn: 217074
2014-09-04 04:56:52 +08:00
updateValueMap(II, ResultReg);
[FastISel][AArch64] Add sqrt intrinsic support. Fixes <rdar://problem/17867067>. llvm-svn: 214388
2014-07-31 14:25:33 +08:00
return true;
}
[FastISel][AArch64] Add {s|u}{add|sub|mul}.with.overflow intrinsic support. This commit adds support for the {s|u}{add|sub|mul}.with.overflow intrinsics. The unit tests for FastISel will be enabled in a later commit, once there is also branch and select folding support. This is related to <rdar://problem/17831117>. llvm-svn: 214348
2014-07-31 06:04:31 +08:00
case Intrinsic::sadd_with_overflow:
case Intrinsic::uadd_with_overflow:
case Intrinsic::ssub_with_overflow:
case Intrinsic::usub_with_overflow:
case Intrinsic::smul_with_overflow:
case Intrinsic::umul_with_overflow: {
// This implements the basic lowering of the xalu with overflow intrinsics.
const Function *Callee = II->getCalledFunction();
auto *Ty = cast<StructType>(Callee->getReturnType());
Type *RetTy = Ty->getTypeAtIndex(0U);
MVT VT;
if (!isTypeLegal(RetTy, VT))
return false;
if (VT != MVT::i32 && VT != MVT::i64)
return false;
const Value *LHS = II->getArgOperand(0);
const Value *RHS = II->getArgOperand(1);
// Canonicalize immediate to the RHS.
if (isa<ConstantInt>(LHS) && !isa<ConstantInt>(RHS) &&
isCommutativeIntrinsic(II))
std::swap(LHS, RHS);
[FastISel][AArch64] Simplify XALU multiplies. Simplify {s|u}mul.with.overflow to {s|u}add.with.overflow when possible. llvm-svn: 218033
2014-09-18 15:04:54 +08:00
// Simplify multiplies.
Change Function::getIntrinsicID() to return an Intrinsic::ID. NFC. Now that Intrinsic::ID is a typed enum, we can forward declare it and so return it from this method. This updates all users which were either using an unsigned to store it, or had a now unnecessary cast. llvm-svn: 237810
2015-05-21 01:16:39 +08:00
Intrinsic::ID IID = II->getIntrinsicID();
[FastISel][AArch64] Simplify XALU multiplies. Simplify {s|u}mul.with.overflow to {s|u}add.with.overflow when possible. llvm-svn: 218033
2014-09-18 15:04:54 +08:00
switch (IID) {
default:
break;
case Intrinsic::smul_with_overflow:
if (const auto *C = dyn_cast<ConstantInt>(RHS))
if (C->getValue() == 2) {
IID = Intrinsic::sadd_with_overflow;
RHS = LHS;
}
break;
case Intrinsic::umul_with_overflow:
if (const auto *C = dyn_cast<ConstantInt>(RHS))
if (C->getValue() == 2) {
IID = Intrinsic::uadd_with_overflow;
RHS = LHS;
}
break;
}
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
unsigned ResultReg1 = 0, ResultReg2 = 0, MulReg = 0;
[FastISel][AArch64] Add {s|u}{add|sub|mul}.with.overflow intrinsic support. This commit adds support for the {s|u}{add|sub|mul}.with.overflow intrinsics. The unit tests for FastISel will be enabled in a later commit, once there is also branch and select folding support. This is related to <rdar://problem/17831117>. llvm-svn: 214348
2014-07-31 06:04:31 +08:00
AArch64CC::CondCode CC = AArch64CC::Invalid;
[FastISel][AArch64] Simplify XALU multiplies. Simplify {s|u}mul.with.overflow to {s|u}add.with.overflow when possible. llvm-svn: 218033
2014-09-18 15:04:54 +08:00
switch (IID) {
[FastISel][AArch64] Add {s|u}{add|sub|mul}.with.overflow intrinsic support. This commit adds support for the {s|u}{add|sub|mul}.with.overflow intrinsics. The unit tests for FastISel will be enabled in a later commit, once there is also branch and select folding support. This is related to <rdar://problem/17831117>. llvm-svn: 214348
2014-07-31 06:04:31 +08:00
default: llvm_unreachable("Unexpected intrinsic!");
case Intrinsic::sadd_with_overflow:
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
ResultReg1 = emitAdd(VT, LHS, RHS, /*SetFlags=*/true);
CC = AArch64CC::VS;
break;
[FastISel][AArch64] Add {s|u}{add|sub|mul}.with.overflow intrinsic support. This commit adds support for the {s|u}{add|sub|mul}.with.overflow intrinsics. The unit tests for FastISel will be enabled in a later commit, once there is also branch and select folding support. This is related to <rdar://problem/17831117>. llvm-svn: 214348
2014-07-31 06:04:31 +08:00
case Intrinsic::uadd_with_overflow:
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
ResultReg1 = emitAdd(VT, LHS, RHS, /*SetFlags=*/true);
CC = AArch64CC::HS;
break;
[FastISel][AArch64] Add {s|u}{add|sub|mul}.with.overflow intrinsic support. This commit adds support for the {s|u}{add|sub|mul}.with.overflow intrinsics. The unit tests for FastISel will be enabled in a later commit, once there is also branch and select folding support. This is related to <rdar://problem/17831117>. llvm-svn: 214348
2014-07-31 06:04:31 +08:00
case Intrinsic::ssub_with_overflow:
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
ResultReg1 = emitSub(VT, LHS, RHS, /*SetFlags=*/true);
CC = AArch64CC::VS;
break;
[FastISel][AArch64] Add {s|u}{add|sub|mul}.with.overflow intrinsic support. This commit adds support for the {s|u}{add|sub|mul}.with.overflow intrinsics. The unit tests for FastISel will be enabled in a later commit, once there is also branch and select folding support. This is related to <rdar://problem/17831117>. llvm-svn: 214348
2014-07-31 06:04:31 +08:00
case Intrinsic::usub_with_overflow:
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
ResultReg1 = emitSub(VT, LHS, RHS, /*SetFlags=*/true);
CC = AArch64CC::LO;
break;
[FastISel][AArch64] Add {s|u}{add|sub|mul}.with.overflow intrinsic support. This commit adds support for the {s|u}{add|sub|mul}.with.overflow intrinsics. The unit tests for FastISel will be enabled in a later commit, once there is also branch and select folding support. This is related to <rdar://problem/17831117>. llvm-svn: 214348
2014-07-31 06:04:31 +08:00
case Intrinsic::smul_with_overflow: {
CC = AArch64CC::NE;
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
unsigned LHSReg = getRegForValue(LHS);
if (!LHSReg)
return false;
bool LHSIsKill = hasTrivialKill(LHS);
unsigned RHSReg = getRegForValue(RHS);
[FastISel][AArch64] Fix the immediate versions of the {s|u}{add|sub}.with.overflow intrinsics. ADDS and SUBS cannot encode negative immediates or immediates larger than 12bit. This fix checks if the immediate version can be used under this constraints and if we can convert ADDS to SUBS or vice versa to support negative immediates. Also update the test cases to test the immediate versions. llvm-svn: 214470
2014-08-01 09:25:55 +08:00
if (!RHSReg)
return false;
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
bool RHSIsKill = hasTrivialKill(RHS);
[FastISel][AArch64] Fix the immediate versions of the {s|u}{add|sub}.with.overflow intrinsics. ADDS and SUBS cannot encode negative immediates or immediates larger than 12bit. This fix checks if the immediate version can be used under this constraints and if we can convert ADDS to SUBS or vice versa to support negative immediates. Also update the test cases to test the immediate versions. llvm-svn: 214470
2014-08-01 09:25:55 +08:00
[FastISel][AArch64] Add {s|u}{add|sub|mul}.with.overflow intrinsic support. This commit adds support for the {s|u}{add|sub|mul}.with.overflow intrinsics. The unit tests for FastISel will be enabled in a later commit, once there is also branch and select folding support. This is related to <rdar://problem/17831117>. llvm-svn: 214348
2014-07-31 06:04:31 +08:00
if (VT == MVT::i32) {
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
MulReg = emitSMULL_rr(MVT::i64, LHSReg, LHSIsKill, RHSReg, RHSIsKill);
[FastISel][AArch64] Fold Sign-/Zero-Extend into the shift immediate instruction. llvm-svn: 216510
2014-08-27 08:58:26 +08:00
unsigned ShiftReg = emitLSR_ri(MVT::i64, MVT::i64, MulReg,
/*IsKill=*/false, 32);
[FastISel][tblgen] Rename tblgen generated FastISel functions. NFC. This is the final round of renaming. This changes tblgen to emit lower-case function names for FastEmitInst_* and FastEmit_*, and updates all its uses in the source code. Reviewed by Eric llvm-svn: 217075
2014-09-04 04:56:59 +08:00
MulReg = fastEmitInst_extractsubreg(VT, MulReg, /*IsKill=*/true,
[FastISel][AArch64] Add {s|u}{add|sub|mul}.with.overflow intrinsic support. This commit adds support for the {s|u}{add|sub|mul}.with.overflow intrinsics. The unit tests for FastISel will be enabled in a later commit, once there is also branch and select folding support. This is related to <rdar://problem/17831117>. llvm-svn: 214348
2014-07-31 06:04:31 +08:00
AArch64::sub_32);
[FastISel][tblgen] Rename tblgen generated FastISel functions. NFC. This is the final round of renaming. This changes tblgen to emit lower-case function names for FastEmitInst_* and FastEmit_*, and updates all its uses in the source code. Reviewed by Eric llvm-svn: 217075
2014-09-04 04:56:59 +08:00
ShiftReg = fastEmitInst_extractsubreg(VT, ShiftReg, /*IsKill=*/true,
[FastISel][AArch64] Add {s|u}{add|sub|mul}.with.overflow intrinsic support. This commit adds support for the {s|u}{add|sub|mul}.with.overflow intrinsics. The unit tests for FastISel will be enabled in a later commit, once there is also branch and select folding support. This is related to <rdar://problem/17831117>. llvm-svn: 214348
2014-07-31 06:04:31 +08:00
AArch64::sub_32);
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
emitSubs_rs(VT, ShiftReg, /*IsKill=*/true, MulReg, /*IsKill=*/false,
AArch64_AM::ASR, 31, /*WantResult=*/false);
[FastISel][AArch64] Add {s|u}{add|sub|mul}.with.overflow intrinsic support. This commit adds support for the {s|u}{add|sub|mul}.with.overflow intrinsics. The unit tests for FastISel will be enabled in a later commit, once there is also branch and select folding support. This is related to <rdar://problem/17831117>. llvm-svn: 214348
2014-07-31 06:04:31 +08:00
} else {
assert(VT == MVT::i64 && "Unexpected value type.");
[AArch64][FastISel] Fix the setting of kill flags for MUL -> UMULH sequences. rdar://problem/20748715 llvm-svn: 236346
2015-05-02 04:57:11 +08:00
// LHSReg and RHSReg cannot be killed by this Mul, since they are
// reused in the next instruction.
MulReg = emitMul_rr(VT, LHSReg, /*IsKill=*/false, RHSReg,
/*IsKill=*/false);
[FastISel][tblgen] Rename tblgen generated FastISel functions. NFC. This is the final round of renaming. This changes tblgen to emit lower-case function names for FastEmitInst_* and FastEmit_*, and updates all its uses in the source code. Reviewed by Eric llvm-svn: 217075
2014-09-04 04:56:59 +08:00
unsigned SMULHReg = fastEmit_rr(VT, VT, ISD::MULHS, LHSReg, LHSIsKill,
[FastISel][AArch64] Add {s|u}{add|sub|mul}.with.overflow intrinsic support. This commit adds support for the {s|u}{add|sub|mul}.with.overflow intrinsics. The unit tests for FastISel will be enabled in a later commit, once there is also branch and select folding support. This is related to <rdar://problem/17831117>. llvm-svn: 214348
2014-07-31 06:04:31 +08:00
RHSReg, RHSIsKill);
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
emitSubs_rs(VT, SMULHReg, /*IsKill=*/true, MulReg, /*IsKill=*/false,
AArch64_AM::ASR, 63, /*WantResult=*/false);
[FastISel][AArch64] Add {s|u}{add|sub|mul}.with.overflow intrinsic support. This commit adds support for the {s|u}{add|sub|mul}.with.overflow intrinsics. The unit tests for FastISel will be enabled in a later commit, once there is also branch and select folding support. This is related to <rdar://problem/17831117>. llvm-svn: 214348
2014-07-31 06:04:31 +08:00
}
break;
}
case Intrinsic::umul_with_overflow: {
CC = AArch64CC::NE;
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
unsigned LHSReg = getRegForValue(LHS);
if (!LHSReg)
return false;
bool LHSIsKill = hasTrivialKill(LHS);
unsigned RHSReg = getRegForValue(RHS);
[FastISel][AArch64] Fix the immediate versions of the {s|u}{add|sub}.with.overflow intrinsics. ADDS and SUBS cannot encode negative immediates or immediates larger than 12bit. This fix checks if the immediate version can be used under this constraints and if we can convert ADDS to SUBS or vice versa to support negative immediates. Also update the test cases to test the immediate versions. llvm-svn: 214470
2014-08-01 09:25:55 +08:00
if (!RHSReg)
return false;
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
bool RHSIsKill = hasTrivialKill(RHS);
[FastISel][AArch64] Fix the immediate versions of the {s|u}{add|sub}.with.overflow intrinsics. ADDS and SUBS cannot encode negative immediates or immediates larger than 12bit. This fix checks if the immediate version can be used under this constraints and if we can convert ADDS to SUBS or vice versa to support negative immediates. Also update the test cases to test the immediate versions. llvm-svn: 214470
2014-08-01 09:25:55 +08:00
[FastISel][AArch64] Add {s|u}{add|sub|mul}.with.overflow intrinsic support. This commit adds support for the {s|u}{add|sub|mul}.with.overflow intrinsics. The unit tests for FastISel will be enabled in a later commit, once there is also branch and select folding support. This is related to <rdar://problem/17831117>. llvm-svn: 214348
2014-07-31 06:04:31 +08:00
if (VT == MVT::i32) {
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
MulReg = emitUMULL_rr(MVT::i64, LHSReg, LHSIsKill, RHSReg, RHSIsKill);
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
emitSubs_rs(MVT::i64, AArch64::XZR, /*IsKill=*/true, MulReg,
/*IsKill=*/false, AArch64_AM::LSR, 32,
/*WantResult=*/false);
[FastISel][tblgen] Rename tblgen generated FastISel functions. NFC. This is the final round of renaming. This changes tblgen to emit lower-case function names for FastEmitInst_* and FastEmit_*, and updates all its uses in the source code. Reviewed by Eric llvm-svn: 217075
2014-09-04 04:56:59 +08:00
MulReg = fastEmitInst_extractsubreg(VT, MulReg, /*IsKill=*/true,
[FastISel][AArch64] Add {s|u}{add|sub|mul}.with.overflow intrinsic support. This commit adds support for the {s|u}{add|sub|mul}.with.overflow intrinsics. The unit tests for FastISel will be enabled in a later commit, once there is also branch and select folding support. This is related to <rdar://problem/17831117>. llvm-svn: 214348
2014-07-31 06:04:31 +08:00
AArch64::sub_32);
} else {
assert(VT == MVT::i64 && "Unexpected value type.");
[AArch64][FastISel] Fix the setting of kill flags for MUL -> UMULH sequences. rdar://problem/20748715 llvm-svn: 236346
2015-05-02 04:57:11 +08:00
// LHSReg and RHSReg cannot be killed by this Mul, since they are
// reused in the next instruction.
MulReg = emitMul_rr(VT, LHSReg, /*IsKill=*/false, RHSReg,
/*IsKill=*/false);
[FastISel][tblgen] Rename tblgen generated FastISel functions. NFC. This is the final round of renaming. This changes tblgen to emit lower-case function names for FastEmitInst_* and FastEmit_*, and updates all its uses in the source code. Reviewed by Eric llvm-svn: 217075
2014-09-04 04:56:59 +08:00
unsigned UMULHReg = fastEmit_rr(VT, VT, ISD::MULHU, LHSReg, LHSIsKill,
[FastISel][AArch64] Add {s|u}{add|sub|mul}.with.overflow intrinsic support. This commit adds support for the {s|u}{add|sub|mul}.with.overflow intrinsics. The unit tests for FastISel will be enabled in a later commit, once there is also branch and select folding support. This is related to <rdar://problem/17831117>. llvm-svn: 214348
2014-07-31 06:04:31 +08:00
RHSReg, RHSIsKill);
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
emitSubs_rr(VT, AArch64::XZR, /*IsKill=*/true, UMULHReg,
/*IsKill=*/false, /*WantResult=*/false);
[FastISel][AArch64] Add {s|u}{add|sub|mul}.with.overflow intrinsic support. This commit adds support for the {s|u}{add|sub|mul}.with.overflow intrinsics. The unit tests for FastISel will be enabled in a later commit, once there is also branch and select folding support. This is related to <rdar://problem/17831117>. llvm-svn: 214348
2014-07-31 06:04:31 +08:00
}
break;
}
}
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
if (MulReg) {
ResultReg1 = createResultReg(TLI.getRegClassFor(VT));
[FastISel][AArch64] Add {s|u}{add|sub|mul}.with.overflow intrinsic support. This commit adds support for the {s|u}{add|sub|mul}.with.overflow intrinsics. The unit tests for FastISel will be enabled in a later commit, once there is also branch and select folding support. This is related to <rdar://problem/17831117>. llvm-svn: 214348
2014-07-31 06:04:31 +08:00
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc,
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
TII.get(TargetOpcode::COPY), ResultReg1).addReg(MulReg);
}
[FastISel][AArch64] Add {s|u}{add|sub|mul}.with.overflow intrinsic support. This commit adds support for the {s|u}{add|sub|mul}.with.overflow intrinsics. The unit tests for FastISel will be enabled in a later commit, once there is also branch and select folding support. This is related to <rdar://problem/17831117>. llvm-svn: 214348
2014-07-31 06:04:31 +08:00
[FastISel][tblgen] Rename tblgen generated FastISel functions. NFC. This is the final round of renaming. This changes tblgen to emit lower-case function names for FastEmitInst_* and FastEmit_*, and updates all its uses in the source code. Reviewed by Eric llvm-svn: 217075
2014-09-04 04:56:59 +08:00
ResultReg2 = fastEmitInst_rri(AArch64::CSINCWr, &AArch64::GPR32RegClass,
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
AArch64::WZR, /*IsKill=*/true, AArch64::WZR,
/*IsKill=*/true, getInvertedCondCode(CC));
Add fake use to suppress defined-but-unused warnings llvm-svn: 219045
2014-10-04 11:50:10 +08:00
(void)ResultReg2;
[FastISel][AArch64] Factor out ADDS/SUBS instruction emission and add support for extensions and shift folding. Factor out the ADDS/SUBS instruction emission code into helper functions and make the helper functions more clever to support most of the different ADDS/SUBS instructions the architecture support. This includes better immedediate support, shift folding, and sign-/zero-extend folding. This fixes <rdar://problem/17913111>. llvm-svn: 216033
2014-08-20 06:29:55 +08:00
assert((ResultReg1 + 1) == ResultReg2 &&
"Nonconsecutive result registers.");
[FastISel] Rename public visible FastISel functions. NFC. This commit renames the following public FastISel functions: LowerArguments -> lowerArguments SelectInstruction -> selectInstruction TargetSelectInstruction -> fastSelectInstruction FastLowerArguments -> fastLowerArguments FastLowerCall -> fastLowerCall FastLowerIntrinsicCall -> fastLowerIntrinsicCall FastEmitZExtFromI1 -> fastEmitZExtFromI1 FastEmitBranch -> fastEmitBranch UpdateValueMap -> updateValueMap TargetMaterializeConstant -> fastMaterializeConstant TargetMaterializeAlloca -> fastMaterializeAlloca TargetMaterializeFloatZero -> fastMaterializeFloatZero LowerCallTo -> lowerCallTo Reviewed by Eric llvm-svn: 217074
2014-09-04 04:56:52 +08:00
updateValueMap(II, ResultReg1, 2);
[FastISel][AArch64] Add {s|u}{add|sub|mul}.with.overflow intrinsic support. This commit adds support for the {s|u}{add|sub|mul}.with.overflow intrinsics. The unit tests for FastISel will be enabled in a later commit, once there is also branch and select folding support. This is related to <rdar://problem/17831117>. llvm-svn: 214348
2014-07-31 06:04:31 +08:00
return true;
}
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
}
return false;
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool AArch64FastISel::selectRet(const Instruction *I) {
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
const ReturnInst *Ret = cast<ReturnInst>(I);
const Function &F = *I->getParent()->getParent();
if (!FuncInfo.CanLowerReturn)
return false;
if (F.isVarArg())
return false;
Swift Calling Convention: swifterror target support. Differential Revision: http://reviews.llvm.org/D18716 llvm-svn: 265997
2016-04-12 05:08:06 +08:00
if (TLI.supportSwiftError() &&
F.getAttributes().hasAttrSomewhere(Attribute::SwiftError))
return false;
CXX_FAST_TLS calling convention: performance improvement for AArch64. The access function has a short entry and a short exit, the initialization block is only run the first time. To improve the performance, we want to have a short frame at the entry and exit. We explicitly handle most of the CSRs via copies. Only the CSRs that are not handled via copies will be in CSR_SaveList. Frame lowering and prologue/epilogue insertion will generate a short frame in the entry and exit according to CSR_SaveList. The majority of the CSRs will be handled by register allcoator. Register allocator will try to spill and reload them in the initialization block. We add CSRsViaCopy, it will be explicitly handled during lowering. 1> we first set FunctionLoweringInfo->SplitCSR if conditions are met (the target supports it for the given machine function and the function has only return exits). We also call TLI->initializeSplitCSR to perform initialization. 2> we call TLI->insertCopiesSplitCSR to insert copies from CSRsViaCopy to virtual registers at beginning of the entry block and copies from virtual registers to CSRsViaCopy at beginning of the exit blocks. 3> we also need to make sure the explicit copies will not be eliminated. The target independent portion was committed as r255353. rdar://problem/23557469 Differential Revision: http://reviews.llvm.org/D15341 llvm-svn: 255821
2015-12-17 05:04:19 +08:00
if (TLI.supportSplitCSR(FuncInfo.MF))
return false;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Build a list of return value registers.
SmallVector<unsigned, 4> RetRegs;
if (Ret->getNumOperands() > 0) {
CallingConv::ID CC = F.getCallingConv();
SmallVector<ISD::OutputArg, 4> Outs;
Redirect DataLayout from TargetMachine to Module in ComputeValueVTs() Summary: Avoid using the TargetMachine owned DataLayout and use the Module owned one instead. This requires passing the DataLayout up the stack to ComputeValueVTs(). This change is part of a series of commits dedicated to have a single DataLayout during compilation by using always the one owned by the module. Reviewers: echristo Subscribers: jholewinski, yaron.keren, rafael, llvm-commits Differential Revision: http://reviews.llvm.org/D11019 From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 241773
2015-07-09 09:57:34 +08:00
GetReturnInfo(F.getReturnType(), F.getAttributes(), Outs, TLI, DL);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Analyze operands of the call, assigning locations to each operand.
SmallVector<CCValAssign, 16> ValLocs;
Remove the target machine from CCState. Previously it was only used to get the subtarget and that's accessible from the MachineFunction now. This helps clear the way for smaller changes where we getting a subtarget will require passing in a MachineFunction/Function as well. llvm-svn: 214988
2014-08-07 02:45:26 +08:00
CCState CCInfo(CC, F.isVarArg(), *FuncInfo.MF, ValLocs, I->getContext());
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
CCAssignFn *RetCC = CC == CallingConv::WebKit_JS ? RetCC_AArch64_WebKit_JS
: RetCC_AArch64_AAPCS;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
CCInfo.AnalyzeReturn(Outs, RetCC);
// Only handle a single return value for now.
if (ValLocs.size() != 1)
return false;
CCValAssign &VA = ValLocs[0];
const Value *RV = Ret->getOperand(0);
// Don't bother handling odd stuff for now.
[FastISel][AArch64] Allow handling of vectors during return lowering for little endian machines. Allow handling of vectors during return lowering at least for little endian machines. This was restricted in r208200 to fix it for big endian machines (according to the comment), but it also disabled it for little endian too. llvm-svn: 217846
2014-09-16 07:40:10 +08:00
if ((VA.getLocInfo() != CCValAssign::Full) &&
(VA.getLocInfo() != CCValAssign::BCvt))
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
[FastISel][AArch64] Allow handling of vectors during return lowering for little endian machines. Allow handling of vectors during return lowering at least for little endian machines. This was restricted in r208200 to fix it for big endian machines (according to the comment), but it also disabled it for little endian too. llvm-svn: 217846
2014-09-16 07:40:10 +08:00
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Only handle register returns for now.
if (!VA.isRegLoc())
return false;
[FastISel][AArch64] Allow handling of vectors during return lowering for little endian machines. Allow handling of vectors during return lowering at least for little endian machines. This was restricted in r208200 to fix it for big endian machines (according to the comment), but it also disabled it for little endian too. llvm-svn: 217846
2014-09-16 07:40:10 +08:00
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
unsigned Reg = getRegForValue(RV);
if (Reg == 0)
return false;
unsigned SrcReg = Reg + VA.getValNo();
unsigned DestReg = VA.getLocReg();
// Avoid a cross-class copy. This is very unlikely.
if (!MRI.getRegClass(SrcReg)->contains(DestReg))
return false;
Make TargetLowering::getPointerTy() taking DataLayout as an argument Summary: This change is part of a series of commits dedicated to have a single DataLayout during compilation by using always the one owned by the module. Reviewers: echristo Subscribers: jholewinski, ted, yaron.keren, rafael, llvm-commits Differential Revision: http://reviews.llvm.org/D11028 From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 241775
2015-07-09 10:09:04 +08:00
EVT RVEVT = TLI.getValueType(DL, RV->getType());
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
if (!RVEVT.isSimple())
return false;
[ARM64-BE] Fix fast-isel, and add appropriate RUN lines to appropriate tests. llvm-svn: 208200
2014-05-07 20:33:55 +08:00
// Vectors (of > 1 lane) in big endian need tricky handling.
[FastISel][AArch64] Allow handling of vectors during return lowering for little endian machines. Allow handling of vectors during return lowering at least for little endian machines. This was restricted in r208200 to fix it for big endian machines (according to the comment), but it also disabled it for little endian too. llvm-svn: 217846
2014-09-16 07:40:10 +08:00
if (RVEVT.isVector() && RVEVT.getVectorNumElements() > 1 &&
!Subtarget->isLittleEndian())
[ARM64-BE] Fix fast-isel, and add appropriate RUN lines to appropriate tests. llvm-svn: 208200
2014-05-07 20:33:55 +08:00
return false;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
MVT RVVT = RVEVT.getSimpleVT();
[ARM64][fast-isel] Fast-isel doesn't know how to handle f128. llvm-svn: 207659
2014-04-30 23:29:57 +08:00
if (RVVT == MVT::f128)
return false;
[FastISel][AArch64] Allow handling of vectors during return lowering for little endian machines. Allow handling of vectors during return lowering at least for little endian machines. This was restricted in r208200 to fix it for big endian machines (according to the comment), but it also disabled it for little endian too. llvm-svn: 217846
2014-09-16 07:40:10 +08:00
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
MVT DestVT = VA.getValVT();
// Special handling for extended integers.
if (RVVT != DestVT) {
if (RVVT != MVT::i1 && RVVT != MVT::i8 && RVVT != MVT::i16)
return false;
if (!Outs[0].Flags.isZExt() && !Outs[0].Flags.isSExt())
return false;
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool IsZExt = Outs[0].Flags.isZExt();
SrcReg = emitIntExt(RVVT, SrcReg, DestVT, IsZExt);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
if (SrcReg == 0)
return false;
}
// Make the copy.
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc,
TII.get(TargetOpcode::COPY), DestReg).addReg(SrcReg);
// Add register to return instruction.
RetRegs.push_back(VA.getLocReg());
}
MachineInstrBuilder MIB = BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc,
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
TII.get(AArch64::RET_ReallyLR));
Convert some AArch64 code to foreach loops. NFC. Also converted a cast<> to dyn_cast while i was working on the same line of code. llvm-svn: 243894
2015-08-04 03:04:32 +08:00
for (unsigned RetReg : RetRegs)
MIB.addReg(RetReg, RegState::Implicit);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return true;
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool AArch64FastISel::selectTrunc(const Instruction *I) {
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
Type *DestTy = I->getType();
Value *Op = I->getOperand(0);
Type *SrcTy = Op->getType();
Make TargetLowering::getPointerTy() taking DataLayout as an argument Summary: This change is part of a series of commits dedicated to have a single DataLayout during compilation by using always the one owned by the module. Reviewers: echristo Subscribers: jholewinski, ted, yaron.keren, rafael, llvm-commits Differential Revision: http://reviews.llvm.org/D11028 From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 241775
2015-07-09 10:09:04 +08:00
EVT SrcEVT = TLI.getValueType(DL, SrcTy, true);
EVT DestEVT = TLI.getValueType(DL, DestTy, true);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
if (!SrcEVT.isSimple())
return false;
if (!DestEVT.isSimple())
return false;
MVT SrcVT = SrcEVT.getSimpleVT();
MVT DestVT = DestEVT.getSimpleVT();
if (SrcVT != MVT::i64 && SrcVT != MVT::i32 && SrcVT != MVT::i16 &&
SrcVT != MVT::i8)
return false;
if (DestVT != MVT::i32 && DestVT != MVT::i16 && DestVT != MVT::i8 &&
DestVT != MVT::i1)
return false;
unsigned SrcReg = getRegForValue(Op);
if (!SrcReg)
return false;
[FastISel][AArch64] Factor out ANDWri instruction generation into a helper function. NFCI. llvm-svn: 216199
2014-08-22 02:02:25 +08:00
bool SrcIsKill = hasTrivialKill(Op);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
Revert "[AArch64][FastISel] Add more truncation tests." and "[AArch64][FastISel] Always use an AND instruction when truncating to non-legal types." This reverts commit r243198 and 243304. Turns out this wasn't the correct fix for this problem. It works only within FastISel, but fails when the truncate is selected by SDAG. llvm-svn: 244287
2015-08-07 06:13:48 +08:00
// If we're truncating from i64 to a smaller non-legal type then generate an
// AND. Otherwise, we know the high bits are undefined and a truncate only
// generate a COPY. We cannot mark the source register also as result
// register, because this can incorrectly transfer the kill flag onto the
// source register.
unsigned ResultReg;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
if (SrcVT == MVT::i64) {
Revert "[AArch64][FastISel] Add more truncation tests." and "[AArch64][FastISel] Always use an AND instruction when truncating to non-legal types." This reverts commit r243198 and 243304. Turns out this wasn't the correct fix for this problem. It works only within FastISel, but fails when the truncate is selected by SDAG. llvm-svn: 244287
2015-08-07 06:13:48 +08:00
uint64_t Mask = 0;
switch (DestVT.SimpleTy) {
default:
// Trunc i64 to i32 is handled by the target-independent fast-isel.
return false;
case MVT::i1:
Mask = 0x1;
break;
case MVT::i8:
Mask = 0xff;
break;
case MVT::i16:
Mask = 0xffff;
break;
}
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Issue an extract_subreg to get the lower 32-bits.
Revert "[AArch64][FastISel] Add more truncation tests." and "[AArch64][FastISel] Always use an AND instruction when truncating to non-legal types." This reverts commit r243198 and 243304. Turns out this wasn't the correct fix for this problem. It works only within FastISel, but fails when the truncate is selected by SDAG. llvm-svn: 244287
2015-08-07 06:13:48 +08:00
unsigned Reg32 = fastEmitInst_extractsubreg(MVT::i32, SrcReg, SrcIsKill,
AArch64::sub_32);
// Create the AND instruction which performs the actual truncation.
ResultReg = emitAnd_ri(MVT::i32, Reg32, /*IsKill=*/true, Mask);
assert(ResultReg && "Unexpected AND instruction emission failure.");
} else {
ResultReg = createResultReg(&AArch64::GPR32RegClass);
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc,
TII.get(TargetOpcode::COPY), ResultReg)
.addReg(SrcReg, getKillRegState(SrcIsKill));
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
}
[FastISel] Rename public visible FastISel functions. NFC. This commit renames the following public FastISel functions: LowerArguments -> lowerArguments SelectInstruction -> selectInstruction TargetSelectInstruction -> fastSelectInstruction FastLowerArguments -> fastLowerArguments FastLowerCall -> fastLowerCall FastLowerIntrinsicCall -> fastLowerIntrinsicCall FastEmitZExtFromI1 -> fastEmitZExtFromI1 FastEmitBranch -> fastEmitBranch UpdateValueMap -> updateValueMap TargetMaterializeConstant -> fastMaterializeConstant TargetMaterializeAlloca -> fastMaterializeAlloca TargetMaterializeFloatZero -> fastMaterializeFloatZero LowerCallTo -> lowerCallTo Reviewed by Eric llvm-svn: 217074
2014-09-04 04:56:52 +08:00
updateValueMap(I, ResultReg);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return true;
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
unsigned AArch64FastISel::emiti1Ext(unsigned SrcReg, MVT DestVT, bool IsZExt) {
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
assert((DestVT == MVT::i8 || DestVT == MVT::i16 || DestVT == MVT::i32 ||
DestVT == MVT::i64) &&
"Unexpected value type.");
// Handle i8 and i16 as i32.
if (DestVT == MVT::i8 || DestVT == MVT::i16)
DestVT = MVT::i32;
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
if (IsZExt) {
[FastISel][AArch64] Add target-specific lowering for logical operations. This change adds support for immediate and shift-left folding into logical operations. This fixes rdar://problem/18223183. llvm-svn: 217118
2014-09-04 09:29:18 +08:00
unsigned ResultReg = emitAnd_ri(MVT::i32, SrcReg, /*TODO:IsKill=*/false, 1);
[FastISel][AArch64] Factor out ANDWri instruction generation into a helper function. NFCI. llvm-svn: 216199
2014-08-22 02:02:25 +08:00
assert(ResultReg && "Unexpected AND instruction emission failure.");
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
if (DestVT == MVT::i64) {
// We're ZExt i1 to i64. The ANDWri Wd, Ws, #1 implicitly clears the
// upper 32 bits. Emit a SUBREG_TO_REG to extend from Wd to Xd.
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
unsigned Reg64 = MRI.createVirtualRegister(&AArch64::GPR64RegClass);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc,
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
TII.get(AArch64::SUBREG_TO_REG), Reg64)
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
.addImm(0)
.addReg(ResultReg)
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
.addImm(AArch64::sub_32);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
ResultReg = Reg64;
}
return ResultReg;
} else {
if (DestVT == MVT::i64) {
// FIXME: We're SExt i1 to i64.
return 0;
}
[FastISel][tblgen] Rename tblgen generated FastISel functions. NFC. This is the final round of renaming. This changes tblgen to emit lower-case function names for FastEmitInst_* and FastEmit_*, and updates all its uses in the source code. Reviewed by Eric llvm-svn: 217075
2014-09-04 04:56:59 +08:00
return fastEmitInst_rii(AArch64::SBFMWri, &AArch64::GPR32RegClass, SrcReg,
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
/*TODO:IsKill=*/false, 0, 0);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
}
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
unsigned AArch64FastISel::emitMul_rr(MVT RetVT, unsigned Op0, bool Op0IsKill,
[FastISel][AArch64] Create helper functions to create the various multiplies on AArch64. llvm-svn: 214346
2014-07-31 06:04:25 +08:00
unsigned Op1, bool Op1IsKill) {
unsigned Opc, ZReg;
switch (RetVT.SimpleTy) {
default: return 0;
case MVT::i8:
case MVT::i16:
case MVT::i32:
RetVT = MVT::i32;
Opc = AArch64::MADDWrrr; ZReg = AArch64::WZR; break;
case MVT::i64:
Opc = AArch64::MADDXrrr; ZReg = AArch64::XZR; break;
}
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
const TargetRegisterClass *RC =
(RetVT == MVT::i64) ? &AArch64::GPR64RegClass : &AArch64::GPR32RegClass;
[FastISel][tblgen] Rename tblgen generated FastISel functions. NFC. This is the final round of renaming. This changes tblgen to emit lower-case function names for FastEmitInst_* and FastEmit_*, and updates all its uses in the source code. Reviewed by Eric llvm-svn: 217075
2014-09-04 04:56:59 +08:00
return fastEmitInst_rrr(Opc, RC, Op0, Op0IsKill, Op1, Op1IsKill,
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
/*IsKill=*/ZReg, true);
[FastISel][AArch64] Create helper functions to create the various multiplies on AArch64. llvm-svn: 214346
2014-07-31 06:04:25 +08:00
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
unsigned AArch64FastISel::emitSMULL_rr(MVT RetVT, unsigned Op0, bool Op0IsKill,
[FastISel][AArch64] Create helper functions to create the various multiplies on AArch64. llvm-svn: 214346
2014-07-31 06:04:25 +08:00
unsigned Op1, bool Op1IsKill) {
if (RetVT != MVT::i64)
return 0;
[FastISel][tblgen] Rename tblgen generated FastISel functions. NFC. This is the final round of renaming. This changes tblgen to emit lower-case function names for FastEmitInst_* and FastEmit_*, and updates all its uses in the source code. Reviewed by Eric llvm-svn: 217075
2014-09-04 04:56:59 +08:00
return fastEmitInst_rrr(AArch64::SMADDLrrr, &AArch64::GPR64RegClass,
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
Op0, Op0IsKill, Op1, Op1IsKill,
AArch64::XZR, /*IsKill=*/true);
[FastISel][AArch64] Create helper functions to create the various multiplies on AArch64. llvm-svn: 214346
2014-07-31 06:04:25 +08:00
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
unsigned AArch64FastISel::emitUMULL_rr(MVT RetVT, unsigned Op0, bool Op0IsKill,
[FastISel][AArch64] Create helper functions to create the various multiplies on AArch64. llvm-svn: 214346
2014-07-31 06:04:25 +08:00
unsigned Op1, bool Op1IsKill) {
if (RetVT != MVT::i64)
return 0;
[FastISel][tblgen] Rename tblgen generated FastISel functions. NFC. This is the final round of renaming. This changes tblgen to emit lower-case function names for FastEmitInst_* and FastEmit_*, and updates all its uses in the source code. Reviewed by Eric llvm-svn: 217075
2014-09-04 04:56:59 +08:00
return fastEmitInst_rrr(AArch64::UMADDLrrr, &AArch64::GPR64RegClass,
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
Op0, Op0IsKill, Op1, Op1IsKill,
AArch64::XZR, /*IsKill=*/true);
[FastISel][AArch64] Create helper functions to create the various multiplies on AArch64. llvm-svn: 214346
2014-07-31 06:04:25 +08:00
}
[FastISel][AArch64] Add support for variable shift. This adds the missing variable shift support for value type i8, i16, and i32. This fixes <rdar://problem/18095685>. llvm-svn: 216242
2014-08-22 07:06:07 +08:00
unsigned AArch64FastISel::emitLSL_rr(MVT RetVT, unsigned Op0Reg, bool Op0IsKill,
unsigned Op1Reg, bool Op1IsKill) {
unsigned Opc = 0;
bool NeedTrunc = false;
uint64_t Mask = 0;
switch (RetVT.SimpleTy) {
default: return 0;
case MVT::i8: Opc = AArch64::LSLVWr; NeedTrunc = true; Mask = 0xff; break;
case MVT::i16: Opc = AArch64::LSLVWr; NeedTrunc = true; Mask = 0xffff; break;
case MVT::i32: Opc = AArch64::LSLVWr; break;
case MVT::i64: Opc = AArch64::LSLVXr; break;
}
const TargetRegisterClass *RC =
(RetVT == MVT::i64) ? &AArch64::GPR64RegClass : &AArch64::GPR32RegClass;
if (NeedTrunc) {
[FastISel][AArch64] Add target-specific lowering for logical operations. This change adds support for immediate and shift-left folding into logical operations. This fixes rdar://problem/18223183. llvm-svn: 217118
2014-09-04 09:29:18 +08:00
Op1Reg = emitAnd_ri(MVT::i32, Op1Reg, Op1IsKill, Mask);
[FastISel][AArch64] Add support for variable shift. This adds the missing variable shift support for value type i8, i16, and i32. This fixes <rdar://problem/18095685>. llvm-svn: 216242
2014-08-22 07:06:07 +08:00
Op1IsKill = true;
}
[FastISel][tblgen] Rename tblgen generated FastISel functions. NFC. This is the final round of renaming. This changes tblgen to emit lower-case function names for FastEmitInst_* and FastEmit_*, and updates all its uses in the source code. Reviewed by Eric llvm-svn: 217075
2014-09-04 04:56:59 +08:00
unsigned ResultReg = fastEmitInst_rr(Opc, RC, Op0Reg, Op0IsKill, Op1Reg,
[FastISel][AArch64] Add support for variable shift. This adds the missing variable shift support for value type i8, i16, and i32. This fixes <rdar://problem/18095685>. llvm-svn: 216242
2014-08-22 07:06:07 +08:00
Op1IsKill);
if (NeedTrunc)
[FastISel][AArch64] Add target-specific lowering for logical operations. This change adds support for immediate and shift-left folding into logical operations. This fixes rdar://problem/18223183. llvm-svn: 217118
2014-09-04 09:29:18 +08:00
ResultReg = emitAnd_ri(MVT::i32, ResultReg, /*IsKill=*/true, Mask);
[FastISel][AArch64] Add support for variable shift. This adds the missing variable shift support for value type i8, i16, and i32. This fixes <rdar://problem/18095685>. llvm-svn: 216242
2014-08-22 07:06:07 +08:00
return ResultReg;
}
[FastISel][AArch64] Fold Sign-/Zero-Extend into the shift immediate instruction. llvm-svn: 216510
2014-08-27 08:58:26 +08:00
unsigned AArch64FastISel::emitLSL_ri(MVT RetVT, MVT SrcVT, unsigned Op0,
bool Op0IsKill, uint64_t Shift,
[FastISel][AArch64] Follow-up fix for "Fix shift-immediate emission for "zero" shifts." Shifts also perform sign-/zero-extends to larger types, which requires us to emit an integer extend instead of a simple COPY. Related to PR21594. llvm-svn: 222257
2014-11-19 05:20:17 +08:00
bool IsZExt) {
[FastISel][AArch64] Fold Sign-/Zero-Extend into the shift immediate instruction. llvm-svn: 216510
2014-08-27 08:58:26 +08:00
assert(RetVT.SimpleTy >= SrcVT.SimpleTy &&
"Unexpected source/return type pair.");
[FastISel][AArch64] Also allow folding of sign-/zero-extend and shift-left for booleans (i1). Shift-left immediate with sign-/zero-extensions also works for boolean values. Update the assert and the test cases to reflect that fact. This should fix a bug found by Chad. llvm-svn: 218275
2014-09-23 05:08:53 +08:00
assert((SrcVT == MVT::i1 || SrcVT == MVT::i8 || SrcVT == MVT::i16 ||
SrcVT == MVT::i32 || SrcVT == MVT::i64) &&
"Unexpected source value type.");
[FastISel][AArch64] Fold Sign-/Zero-Extend into the shift immediate instruction. llvm-svn: 216510
2014-08-27 08:58:26 +08:00
assert((RetVT == MVT::i8 || RetVT == MVT::i16 || RetVT == MVT::i32 ||
RetVT == MVT::i64) && "Unexpected return value type.");
bool Is64Bit = (RetVT == MVT::i64);
unsigned RegSize = Is64Bit ? 64 : 32;
unsigned DstBits = RetVT.getSizeInBits();
unsigned SrcBits = SrcVT.getSizeInBits();
[FastISel][AArch64] Fix shift-immediate emission for "zero" shifts. This change emits a COPY for a shift-immediate with a "zero" shift value. This fixes PR21594 where we emitted a shift instruction with an incorrect immediate operand. llvm-svn: 222247
2014-11-19 03:58:59 +08:00
const TargetRegisterClass *RC =
Is64Bit ? &AArch64::GPR64RegClass : &AArch64::GPR32RegClass;
// Just emit a copy for "zero" shifts.
if (Shift == 0) {
[FastISel][AArch64] Follow-up fix for "Fix shift-immediate emission for "zero" shifts." Shifts also perform sign-/zero-extends to larger types, which requires us to emit an integer extend instead of a simple COPY. Related to PR21594. llvm-svn: 222257
2014-11-19 05:20:17 +08:00
if (RetVT == SrcVT) {
unsigned ResultReg = createResultReg(RC);
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc,
TII.get(TargetOpcode::COPY), ResultReg)
.addReg(Op0, getKillRegState(Op0IsKill));
return ResultReg;
} else
return emitIntExt(SrcVT, Op0, RetVT, IsZExt);
[FastISel][AArch64] Fix shift-immediate emission for "zero" shifts. This change emits a COPY for a shift-immediate with a "zero" shift value. This fixes PR21594 where we emitted a shift instruction with an incorrect immediate operand. llvm-svn: 222247
2014-11-19 03:58:59 +08:00
}
[FastISel][AArch64] Add support for shift-immediate. Currently the shift-immediate versions are not supported by tblgen and hopefully this can be later removed, once the required support has been added to tblgen. llvm-svn: 214345
2014-07-31 06:04:22 +08:00
[FastISel][AArch64] Fold Sign-/Zero-Extend into the shift immediate instruction. llvm-svn: 216510
2014-08-27 08:58:26 +08:00
// Don't deal with undefined shifts.
if (Shift >= DstBits)
return 0;
// For immediate shifts we can fold the zero-/sign-extension into the shift.
// {S|U}BFM Wd, Wn, #r, #s
// Wd<32+s-r,32-r> = Wn<s:0> when r > s
// %1 = {s|z}ext i8 {0b1010_1010|0b0101_0101} to i16
// %2 = shl i16 %1, 4
// Wd<32+7-28,32-28> = Wn<7:0> <- clamp s to 7
// 0b1111_1111_1111_1111__1111_1010_1010_0000 sext
// 0b0000_0000_0000_0000__0000_0101_0101_0000 sext | zext
// 0b0000_0000_0000_0000__0000_1010_1010_0000 zext
// %1 = {s|z}ext i8 {0b1010_1010|0b0101_0101} to i16
// %2 = shl i16 %1, 8
// Wd<32+7-24,32-24> = Wn<7:0>
// 0b1111_1111_1111_1111__1010_1010_0000_0000 sext
// 0b0000_0000_0000_0000__0101_0101_0000_0000 sext | zext
// 0b0000_0000_0000_0000__1010_1010_0000_0000 zext
// %1 = {s|z}ext i8 {0b1010_1010|0b0101_0101} to i16
// %2 = shl i16 %1, 12
// Wd<32+3-20,32-20> = Wn<3:0>
// 0b1111_1111_1111_1111__1010_0000_0000_0000 sext
// 0b0000_0000_0000_0000__0101_0000_0000_0000 sext | zext
// 0b0000_0000_0000_0000__1010_0000_0000_0000 zext
unsigned ImmR = RegSize - Shift;
// Limit the width to the length of the source type.
unsigned ImmS = std::min<unsigned>(SrcBits - 1, DstBits - 1 - Shift);
static const unsigned OpcTable[2][2] = {
{AArch64::SBFMWri, AArch64::SBFMXri},
{AArch64::UBFMWri, AArch64::UBFMXri}
};
[FastISel][AArch64] Follow-up fix for "Fix shift-immediate emission for "zero" shifts." Shifts also perform sign-/zero-extends to larger types, which requires us to emit an integer extend instead of a simple COPY. Related to PR21594. llvm-svn: 222257
2014-11-19 05:20:17 +08:00
unsigned Opc = OpcTable[IsZExt][Is64Bit];
[FastISel][AArch64] Fold Sign-/Zero-Extend into the shift immediate instruction. llvm-svn: 216510
2014-08-27 08:58:26 +08:00
if (SrcVT.SimpleTy <= MVT::i32 && RetVT == MVT::i64) {
unsigned TmpReg = MRI.createVirtualRegister(RC);
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc,
TII.get(AArch64::SUBREG_TO_REG), TmpReg)
.addImm(0)
.addReg(Op0, getKillRegState(Op0IsKill))
.addImm(AArch64::sub_32);
Op0 = TmpReg;
Op0IsKill = true;
}
[FastISel][tblgen] Rename tblgen generated FastISel functions. NFC. This is the final round of renaming. This changes tblgen to emit lower-case function names for FastEmitInst_* and FastEmit_*, and updates all its uses in the source code. Reviewed by Eric llvm-svn: 217075
2014-09-04 04:56:59 +08:00
return fastEmitInst_rii(Opc, RC, Op0, Op0IsKill, ImmR, ImmS);
[FastISel][AArch64] Add support for shift-immediate. Currently the shift-immediate versions are not supported by tblgen and hopefully this can be later removed, once the required support has been added to tblgen. llvm-svn: 214345
2014-07-31 06:04:22 +08:00
}
[FastISel][AArch64] Add support for variable shift. This adds the missing variable shift support for value type i8, i16, and i32. This fixes <rdar://problem/18095685>. llvm-svn: 216242
2014-08-22 07:06:07 +08:00
unsigned AArch64FastISel::emitLSR_rr(MVT RetVT, unsigned Op0Reg, bool Op0IsKill,
unsigned Op1Reg, bool Op1IsKill) {
unsigned Opc = 0;
bool NeedTrunc = false;
uint64_t Mask = 0;
switch (RetVT.SimpleTy) {
default: return 0;
case MVT::i8: Opc = AArch64::LSRVWr; NeedTrunc = true; Mask = 0xff; break;
case MVT::i16: Opc = AArch64::LSRVWr; NeedTrunc = true; Mask = 0xffff; break;
case MVT::i32: Opc = AArch64::LSRVWr; break;
case MVT::i64: Opc = AArch64::LSRVXr; break;
}
const TargetRegisterClass *RC =
(RetVT == MVT::i64) ? &AArch64::GPR64RegClass : &AArch64::GPR32RegClass;
if (NeedTrunc) {
[FastISel][AArch64] Add target-specific lowering for logical operations. This change adds support for immediate and shift-left folding into logical operations. This fixes rdar://problem/18223183. llvm-svn: 217118
2014-09-04 09:29:18 +08:00
Op0Reg = emitAnd_ri(MVT::i32, Op0Reg, Op0IsKill, Mask);
Op1Reg = emitAnd_ri(MVT::i32, Op1Reg, Op1IsKill, Mask);
[FastISel][AArch64] Add support for variable shift. This adds the missing variable shift support for value type i8, i16, and i32. This fixes <rdar://problem/18095685>. llvm-svn: 216242
2014-08-22 07:06:07 +08:00
Op0IsKill = Op1IsKill = true;
}
[FastISel][tblgen] Rename tblgen generated FastISel functions. NFC. This is the final round of renaming. This changes tblgen to emit lower-case function names for FastEmitInst_* and FastEmit_*, and updates all its uses in the source code. Reviewed by Eric llvm-svn: 217075
2014-09-04 04:56:59 +08:00
unsigned ResultReg = fastEmitInst_rr(Opc, RC, Op0Reg, Op0IsKill, Op1Reg,
[FastISel][AArch64] Add support for variable shift. This adds the missing variable shift support for value type i8, i16, and i32. This fixes <rdar://problem/18095685>. llvm-svn: 216242
2014-08-22 07:06:07 +08:00
Op1IsKill);
if (NeedTrunc)
[FastISel][AArch64] Add target-specific lowering for logical operations. This change adds support for immediate and shift-left folding into logical operations. This fixes rdar://problem/18223183. llvm-svn: 217118
2014-09-04 09:29:18 +08:00
ResultReg = emitAnd_ri(MVT::i32, ResultReg, /*IsKill=*/true, Mask);
[FastISel][AArch64] Add support for variable shift. This adds the missing variable shift support for value type i8, i16, and i32. This fixes <rdar://problem/18095685>. llvm-svn: 216242
2014-08-22 07:06:07 +08:00
return ResultReg;
}
[FastISel][AArch64] Fold Sign-/Zero-Extend into the shift immediate instruction. llvm-svn: 216510
2014-08-27 08:58:26 +08:00
unsigned AArch64FastISel::emitLSR_ri(MVT RetVT, MVT SrcVT, unsigned Op0,
bool Op0IsKill, uint64_t Shift,
bool IsZExt) {
assert(RetVT.SimpleTy >= SrcVT.SimpleTy &&
"Unexpected source/return type pair.");
[FastISel][AArch64] Also allow folding of sign-/zero-extend and logical shift-right for booleans (i1). Logical shift-right immediate with sign-/zero-extensions also works for boolean values. Update the assert and the test cases to reflect that fact. llvm-svn: 222270
2014-11-19 06:38:42 +08:00
assert((SrcVT == MVT::i1 || SrcVT == MVT::i8 || SrcVT == MVT::i16 ||
SrcVT == MVT::i32 || SrcVT == MVT::i64) &&
"Unexpected source value type.");
[FastISel][AArch64] Fold Sign-/Zero-Extend into the shift immediate instruction. llvm-svn: 216510
2014-08-27 08:58:26 +08:00
assert((RetVT == MVT::i8 || RetVT == MVT::i16 || RetVT == MVT::i32 ||
RetVT == MVT::i64) && "Unexpected return value type.");
bool Is64Bit = (RetVT == MVT::i64);
unsigned RegSize = Is64Bit ? 64 : 32;
unsigned DstBits = RetVT.getSizeInBits();
unsigned SrcBits = SrcVT.getSizeInBits();
[FastISel][AArch64] Fix shift-immediate emission for "zero" shifts. This change emits a COPY for a shift-immediate with a "zero" shift value. This fixes PR21594 where we emitted a shift instruction with an incorrect immediate operand. llvm-svn: 222247
2014-11-19 03:58:59 +08:00
const TargetRegisterClass *RC =
Is64Bit ? &AArch64::GPR64RegClass : &AArch64::GPR32RegClass;
// Just emit a copy for "zero" shifts.
if (Shift == 0) {
[FastISel][AArch64] Follow-up fix for "Fix shift-immediate emission for "zero" shifts." Shifts also perform sign-/zero-extends to larger types, which requires us to emit an integer extend instead of a simple COPY. Related to PR21594. llvm-svn: 222257
2014-11-19 05:20:17 +08:00
if (RetVT == SrcVT) {
unsigned ResultReg = createResultReg(RC);
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc,
TII.get(TargetOpcode::COPY), ResultReg)
.addReg(Op0, getKillRegState(Op0IsKill));
return ResultReg;
} else
return emitIntExt(SrcVT, Op0, RetVT, IsZExt);
[FastISel][AArch64] Fix shift-immediate emission for "zero" shifts. This change emits a COPY for a shift-immediate with a "zero" shift value. This fixes PR21594 where we emitted a shift instruction with an incorrect immediate operand. llvm-svn: 222247
2014-11-19 03:58:59 +08:00
}
[FastISel][AArch64] Fold Sign-/Zero-Extend into the shift immediate instruction. llvm-svn: 216510
2014-08-27 08:58:26 +08:00
// Don't deal with undefined shifts.
if (Shift >= DstBits)
return 0;
// For immediate shifts we can fold the zero-/sign-extension into the shift.
// {S|U}BFM Wd, Wn, #r, #s
// Wd<s-r:0> = Wn<s:r> when r <= s
// %1 = {s|z}ext i8 {0b1010_1010|0b0101_0101} to i16
// %2 = lshr i16 %1, 4
// Wd<7-4:0> = Wn<7:4>
// 0b0000_0000_0000_0000__0000_1111_1111_1010 sext
// 0b0000_0000_0000_0000__0000_0000_0000_0101 sext | zext
// 0b0000_0000_0000_0000__0000_0000_0000_1010 zext
// %1 = {s|z}ext i8 {0b1010_1010|0b0101_0101} to i16
// %2 = lshr i16 %1, 8
// Wd<7-7,0> = Wn<7:7>
// 0b0000_0000_0000_0000__0000_0000_1111_1111 sext
// 0b0000_0000_0000_0000__0000_0000_0000_0000 sext
// 0b0000_0000_0000_0000__0000_0000_0000_0000 zext
// %1 = {s|z}ext i8 {0b1010_1010|0b0101_0101} to i16
// %2 = lshr i16 %1, 12
// Wd<7-7,0> = Wn<7:7> <- clamp r to 7
// 0b0000_0000_0000_0000__0000_0000_0000_1111 sext
// 0b0000_0000_0000_0000__0000_0000_0000_0000 sext
// 0b0000_0000_0000_0000__0000_0000_0000_0000 zext
if (Shift >= SrcBits && IsZExt)
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return materializeInt(ConstantInt::get(*Context, APInt(RegSize, 0)), RetVT);
[FastISel][AArch64] Fold Sign-/Zero-Extend into the shift immediate instruction. llvm-svn: 216510
2014-08-27 08:58:26 +08:00
// It is not possible to fold a sign-extend into the LShr instruction. In this
// case emit a sign-extend.
if (!IsZExt) {
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
Op0 = emitIntExt(SrcVT, Op0, RetVT, IsZExt);
[FastISel][AArch64] Fold Sign-/Zero-Extend into the shift immediate instruction. llvm-svn: 216510
2014-08-27 08:58:26 +08:00
if (!Op0)
return 0;
Op0IsKill = true;
SrcVT = RetVT;
SrcBits = SrcVT.getSizeInBits();
IsZExt = true;
[FastISel][AArch64] Add support for shift-immediate. Currently the shift-immediate versions are not supported by tblgen and hopefully this can be later removed, once the required support has been added to tblgen. llvm-svn: 214345
2014-07-31 06:04:22 +08:00
}
[FastISel][AArch64] Fold Sign-/Zero-Extend into the shift immediate instruction. llvm-svn: 216510
2014-08-27 08:58:26 +08:00
unsigned ImmR = std::min<unsigned>(SrcBits - 1, Shift);
unsigned ImmS = SrcBits - 1;
static const unsigned OpcTable[2][2] = {
{AArch64::SBFMWri, AArch64::SBFMXri},
{AArch64::UBFMWri, AArch64::UBFMXri}
};
unsigned Opc = OpcTable[IsZExt][Is64Bit];
if (SrcVT.SimpleTy <= MVT::i32 && RetVT == MVT::i64) {
unsigned TmpReg = MRI.createVirtualRegister(RC);
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc,
TII.get(AArch64::SUBREG_TO_REG), TmpReg)
.addImm(0)
.addReg(Op0, getKillRegState(Op0IsKill))
.addImm(AArch64::sub_32);
Op0 = TmpReg;
Op0IsKill = true;
}
[FastISel][tblgen] Rename tblgen generated FastISel functions. NFC. This is the final round of renaming. This changes tblgen to emit lower-case function names for FastEmitInst_* and FastEmit_*, and updates all its uses in the source code. Reviewed by Eric llvm-svn: 217075
2014-09-04 04:56:59 +08:00
return fastEmitInst_rii(Opc, RC, Op0, Op0IsKill, ImmR, ImmS);
[FastISel][AArch64] Add support for shift-immediate. Currently the shift-immediate versions are not supported by tblgen and hopefully this can be later removed, once the required support has been added to tblgen. llvm-svn: 214345
2014-07-31 06:04:22 +08:00
}
[FastISel][AArch64] Add support for variable shift. This adds the missing variable shift support for value type i8, i16, and i32. This fixes <rdar://problem/18095685>. llvm-svn: 216242
2014-08-22 07:06:07 +08:00
unsigned AArch64FastISel::emitASR_rr(MVT RetVT, unsigned Op0Reg, bool Op0IsKill,
unsigned Op1Reg, bool Op1IsKill) {
unsigned Opc = 0;
bool NeedTrunc = false;
uint64_t Mask = 0;
switch (RetVT.SimpleTy) {
default: return 0;
case MVT::i8: Opc = AArch64::ASRVWr; NeedTrunc = true; Mask = 0xff; break;
case MVT::i16: Opc = AArch64::ASRVWr; NeedTrunc = true; Mask = 0xffff; break;
case MVT::i32: Opc = AArch64::ASRVWr; break;
case MVT::i64: Opc = AArch64::ASRVXr; break;
}
const TargetRegisterClass *RC =
(RetVT == MVT::i64) ? &AArch64::GPR64RegClass : &AArch64::GPR32RegClass;
if (NeedTrunc) {
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
Op0Reg = emitIntExt(RetVT, Op0Reg, MVT::i32, /*IsZExt=*/false);
[FastISel][AArch64] Add target-specific lowering for logical operations. This change adds support for immediate and shift-left folding into logical operations. This fixes rdar://problem/18223183. llvm-svn: 217118
2014-09-04 09:29:18 +08:00
Op1Reg = emitAnd_ri(MVT::i32, Op1Reg, Op1IsKill, Mask);
[FastISel][AArch64] Add support for variable shift. This adds the missing variable shift support for value type i8, i16, and i32. This fixes <rdar://problem/18095685>. llvm-svn: 216242
2014-08-22 07:06:07 +08:00
Op0IsKill = Op1IsKill = true;
}
[FastISel][tblgen] Rename tblgen generated FastISel functions. NFC. This is the final round of renaming. This changes tblgen to emit lower-case function names for FastEmitInst_* and FastEmit_*, and updates all its uses in the source code. Reviewed by Eric llvm-svn: 217075
2014-09-04 04:56:59 +08:00
unsigned ResultReg = fastEmitInst_rr(Opc, RC, Op0Reg, Op0IsKill, Op1Reg,
[FastISel][AArch64] Add support for variable shift. This adds the missing variable shift support for value type i8, i16, and i32. This fixes <rdar://problem/18095685>. llvm-svn: 216242
2014-08-22 07:06:07 +08:00
Op1IsKill);
if (NeedTrunc)
[FastISel][AArch64] Add target-specific lowering for logical operations. This change adds support for immediate and shift-left folding into logical operations. This fixes rdar://problem/18223183. llvm-svn: 217118
2014-09-04 09:29:18 +08:00
ResultReg = emitAnd_ri(MVT::i32, ResultReg, /*IsKill=*/true, Mask);
[FastISel][AArch64] Add support for variable shift. This adds the missing variable shift support for value type i8, i16, and i32. This fixes <rdar://problem/18095685>. llvm-svn: 216242
2014-08-22 07:06:07 +08:00
return ResultReg;
}
[FastISel][AArch64] Fold Sign-/Zero-Extend into the shift immediate instruction. llvm-svn: 216510
2014-08-27 08:58:26 +08:00
unsigned AArch64FastISel::emitASR_ri(MVT RetVT, MVT SrcVT, unsigned Op0,
bool Op0IsKill, uint64_t Shift,
bool IsZExt) {
assert(RetVT.SimpleTy >= SrcVT.SimpleTy &&
"Unexpected source/return type pair.");
[FastISel][AArch64] Also allow folding of sign-/zero-extend and arithmetic shift-right for booleans (i1). Arithmetic shift-right immediate with sign-/zero-extensions also works for boolean values. Update the assert and the test cases to reflect that fact. llvm-svn: 222272
2014-11-19 06:41:49 +08:00
assert((SrcVT == MVT::i1 || SrcVT == MVT::i8 || SrcVT == MVT::i16 ||
SrcVT == MVT::i32 || SrcVT == MVT::i64) &&
"Unexpected source value type.");
[FastISel][AArch64] Fold Sign-/Zero-Extend into the shift immediate instruction. llvm-svn: 216510
2014-08-27 08:58:26 +08:00
assert((RetVT == MVT::i8 || RetVT == MVT::i16 || RetVT == MVT::i32 ||
RetVT == MVT::i64) && "Unexpected return value type.");
bool Is64Bit = (RetVT == MVT::i64);
unsigned RegSize = Is64Bit ? 64 : 32;
unsigned DstBits = RetVT.getSizeInBits();
unsigned SrcBits = SrcVT.getSizeInBits();
[FastISel][AArch64] Fix shift-immediate emission for "zero" shifts. This change emits a COPY for a shift-immediate with a "zero" shift value. This fixes PR21594 where we emitted a shift instruction with an incorrect immediate operand. llvm-svn: 222247
2014-11-19 03:58:59 +08:00
const TargetRegisterClass *RC =
Is64Bit ? &AArch64::GPR64RegClass : &AArch64::GPR32RegClass;
// Just emit a copy for "zero" shifts.
if (Shift == 0) {
[FastISel][AArch64] Follow-up fix for "Fix shift-immediate emission for "zero" shifts." Shifts also perform sign-/zero-extends to larger types, which requires us to emit an integer extend instead of a simple COPY. Related to PR21594. llvm-svn: 222257
2014-11-19 05:20:17 +08:00
if (RetVT == SrcVT) {
unsigned ResultReg = createResultReg(RC);
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc,
TII.get(TargetOpcode::COPY), ResultReg)
.addReg(Op0, getKillRegState(Op0IsKill));
return ResultReg;
} else
return emitIntExt(SrcVT, Op0, RetVT, IsZExt);
[FastISel][AArch64] Fix shift-immediate emission for "zero" shifts. This change emits a COPY for a shift-immediate with a "zero" shift value. This fixes PR21594 where we emitted a shift instruction with an incorrect immediate operand. llvm-svn: 222247
2014-11-19 03:58:59 +08:00
}
[FastISel][AArch64] Add support for shift-immediate. Currently the shift-immediate versions are not supported by tblgen and hopefully this can be later removed, once the required support has been added to tblgen. llvm-svn: 214345
2014-07-31 06:04:22 +08:00
[FastISel][AArch64] Fold Sign-/Zero-Extend into the shift immediate instruction. llvm-svn: 216510
2014-08-27 08:58:26 +08:00
// Don't deal with undefined shifts.
if (Shift >= DstBits)
return 0;
// For immediate shifts we can fold the zero-/sign-extension into the shift.
// {S|U}BFM Wd, Wn, #r, #s
// Wd<s-r:0> = Wn<s:r> when r <= s
// %1 = {s|z}ext i8 {0b1010_1010|0b0101_0101} to i16
// %2 = ashr i16 %1, 4
// Wd<7-4:0> = Wn<7:4>
// 0b1111_1111_1111_1111__1111_1111_1111_1010 sext
// 0b0000_0000_0000_0000__0000_0000_0000_0101 sext | zext
// 0b0000_0000_0000_0000__0000_0000_0000_1010 zext
// %1 = {s|z}ext i8 {0b1010_1010|0b0101_0101} to i16
// %2 = ashr i16 %1, 8
// Wd<7-7,0> = Wn<7:7>
// 0b1111_1111_1111_1111__1111_1111_1111_1111 sext
// 0b0000_0000_0000_0000__0000_0000_0000_0000 sext
// 0b0000_0000_0000_0000__0000_0000_0000_0000 zext
// %1 = {s|z}ext i8 {0b1010_1010|0b0101_0101} to i16
// %2 = ashr i16 %1, 12
// Wd<7-7,0> = Wn<7:7> <- clamp r to 7
// 0b1111_1111_1111_1111__1111_1111_1111_1111 sext
// 0b0000_0000_0000_0000__0000_0000_0000_0000 sext
// 0b0000_0000_0000_0000__0000_0000_0000_0000 zext
if (Shift >= SrcBits && IsZExt)
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return materializeInt(ConstantInt::get(*Context, APInt(RegSize, 0)), RetVT);
[FastISel][AArch64] Fold Sign-/Zero-Extend into the shift immediate instruction. llvm-svn: 216510
2014-08-27 08:58:26 +08:00
unsigned ImmR = std::min<unsigned>(SrcBits - 1, Shift);
unsigned ImmS = SrcBits - 1;
static const unsigned OpcTable[2][2] = {
{AArch64::SBFMWri, AArch64::SBFMXri},
{AArch64::UBFMWri, AArch64::UBFMXri}
};
unsigned Opc = OpcTable[IsZExt][Is64Bit];
if (SrcVT.SimpleTy <= MVT::i32 && RetVT == MVT::i64) {
unsigned TmpReg = MRI.createVirtualRegister(RC);
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc,
TII.get(AArch64::SUBREG_TO_REG), TmpReg)
.addImm(0)
.addReg(Op0, getKillRegState(Op0IsKill))
.addImm(AArch64::sub_32);
Op0 = TmpReg;
Op0IsKill = true;
}
[FastISel][tblgen] Rename tblgen generated FastISel functions. NFC. This is the final round of renaming. This changes tblgen to emit lower-case function names for FastEmitInst_* and FastEmit_*, and updates all its uses in the source code. Reviewed by Eric llvm-svn: 217075
2014-09-04 04:56:59 +08:00
return fastEmitInst_rii(Opc, RC, Op0, Op0IsKill, ImmR, ImmS);
[FastISel][AArch64] Add support for shift-immediate. Currently the shift-immediate versions are not supported by tblgen and hopefully this can be later removed, once the required support has been added to tblgen. llvm-svn: 214345
2014-07-31 06:04:22 +08:00
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
unsigned AArch64FastISel::emitIntExt(MVT SrcVT, unsigned SrcReg, MVT DestVT,
bool IsZExt) {
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
assert(DestVT != MVT::i1 && "ZeroExt/SignExt an i1?");
Allow AArch64FastISel to degrade graceully in the presence of an MVT::i128 Currently AArch64FastISel crashes if it tries to extend an integer into an MVT::i128. This can happen by creating 128 bit integers like so: typedef unsigned int uint128_t __attribute__((mode(TI))); typedef int sint128_t __attribute__((mode(TI))); This patch makes EmitIntExt check for their presence and then falls back to SelectionDAG. Tests included. rdar://17516686 llvm-svn: 212492
2014-07-08 05:37:51 +08:00
Make AArch64FastISel::EmitIntExt explicitly check its source and destination types This is a follow up to r212492. There should be no functional difference, but this patch makes it clear that SrcVT must be an i1/i8/16/i32 and DestVT must be an i8/i16/i32/i64. rdar://17516686 llvm-svn: 212633
2014-07-10 01:54:32 +08:00
// FastISel does not have plumbing to deal with extensions where the SrcVT or
// DestVT are odd things, so test to make sure that they are both types we can
// handle (i1/i8/i16/i32 for SrcVT and i8/i16/i32/i64 for DestVT), otherwise
// bail out to SelectionDAG.
if (((DestVT != MVT::i8) && (DestVT != MVT::i16) &&
(DestVT != MVT::i32) && (DestVT != MVT::i64)) ||
((SrcVT != MVT::i1) && (SrcVT != MVT::i8) &&
(SrcVT != MVT::i16) && (SrcVT != MVT::i32)))
Allow AArch64FastISel to degrade graceully in the presence of an MVT::i128 Currently AArch64FastISel crashes if it tries to extend an integer into an MVT::i128. This can happen by creating 128 bit integers like so: typedef unsigned int uint128_t __attribute__((mode(TI))); typedef int sint128_t __attribute__((mode(TI))); This patch makes EmitIntExt check for their presence and then falls back to SelectionDAG. Tests included. rdar://17516686 llvm-svn: 212492
2014-07-08 05:37:51 +08:00
return 0;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
unsigned Opc;
unsigned Imm = 0;
switch (SrcVT.SimpleTy) {
default:
return 0;
case MVT::i1:
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return emiti1Ext(SrcReg, DestVT, IsZExt);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
case MVT::i8:
if (DestVT == MVT::i64)
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
Opc = IsZExt ? AArch64::UBFMXri : AArch64::SBFMXri;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
else
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
Opc = IsZExt ? AArch64::UBFMWri : AArch64::SBFMWri;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
Imm = 7;
break;
case MVT::i16:
if (DestVT == MVT::i64)
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
Opc = IsZExt ? AArch64::UBFMXri : AArch64::SBFMXri;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
else
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
Opc = IsZExt ? AArch64::UBFMWri : AArch64::SBFMWri;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
Imm = 15;
break;
case MVT::i32:
assert(DestVT == MVT::i64 && "IntExt i32 to i32?!?");
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
Opc = IsZExt ? AArch64::UBFMXri : AArch64::SBFMXri;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
Imm = 31;
break;
}
// Handle i8 and i16 as i32.
if (DestVT == MVT::i8 || DestVT == MVT::i16)
DestVT = MVT::i32;
ARM64: make sure FastISel uses a GPR64 source in 64-bit extensions. llvm-svn: 207620
2014-04-30 17:32:01 +08:00
else if (DestVT == MVT::i64) {
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
unsigned Src64 = MRI.createVirtualRegister(&AArch64::GPR64RegClass);
ARM64: make sure FastISel uses a GPR64 source in 64-bit extensions. llvm-svn: 207620
2014-04-30 17:32:01 +08:00
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc,
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
TII.get(AArch64::SUBREG_TO_REG), Src64)
ARM64: make sure FastISel uses a GPR64 source in 64-bit extensions. llvm-svn: 207620
2014-04-30 17:32:01 +08:00
.addImm(0)
.addReg(SrcReg)
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
.addImm(AArch64::sub_32);
ARM64: make sure FastISel uses a GPR64 source in 64-bit extensions. llvm-svn: 207620
2014-04-30 17:32:01 +08:00
SrcReg = Src64;
}
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
const TargetRegisterClass *RC =
(DestVT == MVT::i64) ? &AArch64::GPR64RegClass : &AArch64::GPR32RegClass;
[FastISel][tblgen] Rename tblgen generated FastISel functions. NFC. This is the final round of renaming. This changes tblgen to emit lower-case function names for FastEmitInst_* and FastEmit_*, and updates all its uses in the source code. Reviewed by Eric llvm-svn: 217075
2014-09-04 04:56:59 +08:00
return fastEmitInst_rii(Opc, RC, SrcReg, /*TODO:IsKill=*/false, 0, Imm);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
}
[FastISel][AArch64] Fix sign-/zero-extend folding when SelectionDAG is involved. Sign-/zero-extend folding depended on the load and the integer extend to be both selected by FastISel. This cannot always be garantueed and SelectionDAG might interfer. This commit adds additonal checks to load and integer extend lowering to catch this. Related to rdar://problem/18495928. llvm-svn: 219716
2014-10-15 04:36:02 +08:00
static bool isZExtLoad(const MachineInstr *LI) {
switch (LI->getOpcode()) {
default:
return false;
case AArch64::LDURBBi:
case AArch64::LDURHHi:
case AArch64::LDURWi:
case AArch64::LDRBBui:
case AArch64::LDRHHui:
case AArch64::LDRWui:
case AArch64::LDRBBroX:
case AArch64::LDRHHroX:
case AArch64::LDRWroX:
case AArch64::LDRBBroW:
case AArch64::LDRHHroW:
case AArch64::LDRWroW:
return true;
}
}
static bool isSExtLoad(const MachineInstr *LI) {
switch (LI->getOpcode()) {
default:
return false;
case AArch64::LDURSBWi:
case AArch64::LDURSHWi:
case AArch64::LDURSBXi:
case AArch64::LDURSHXi:
case AArch64::LDURSWi:
case AArch64::LDRSBWui:
case AArch64::LDRSHWui:
case AArch64::LDRSBXui:
case AArch64::LDRSHXui:
case AArch64::LDRSWui:
case AArch64::LDRSBWroX:
case AArch64::LDRSHWroX:
case AArch64::LDRSBXroX:
case AArch64::LDRSHXroX:
case AArch64::LDRSWroX:
case AArch64::LDRSBWroW:
case AArch64::LDRSHWroW:
case AArch64::LDRSBXroW:
case AArch64::LDRSHXroW:
case AArch64::LDRSWroW:
return true;
}
}
bool AArch64FastISel::optimizeIntExtLoad(const Instruction *I, MVT RetVT,
MVT SrcVT) {
const auto *LI = dyn_cast<LoadInst>(I->getOperand(0));
if (!LI || !LI->hasOneUse())
return false;
// Check if the load instruction has already been selected.
unsigned Reg = lookUpRegForValue(LI);
if (!Reg)
return false;
MachineInstr *MI = MRI.getUniqueVRegDef(Reg);
if (!MI)
return false;
// Check if the correct load instruction has been emitted - SelectionDAG might
// have emitted a zero-extending load, but we need a sign-extending load.
bool IsZExt = isa<ZExtInst>(I);
const auto *LoadMI = MI;
if (LoadMI->getOpcode() == TargetOpcode::COPY &&
LoadMI->getOperand(1).getSubReg() == AArch64::sub_32) {
unsigned LoadReg = MI->getOperand(1).getReg();
LoadMI = MRI.getUniqueVRegDef(LoadReg);
assert(LoadMI && "Expected valid instruction");
}
if (!(IsZExt && isZExtLoad(LoadMI)) && !(!IsZExt && isSExtLoad(LoadMI)))
return false;
// Nothing to be done.
if (RetVT != MVT::i64 || SrcVT > MVT::i32) {
updateValueMap(I, Reg);
return true;
}
if (IsZExt) {
unsigned Reg64 = createResultReg(&AArch64::GPR64RegClass);
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc,
TII.get(AArch64::SUBREG_TO_REG), Reg64)
.addImm(0)
.addReg(Reg, getKillRegState(true))
.addImm(AArch64::sub_32);
Reg = Reg64;
} else {
assert((MI->getOpcode() == TargetOpcode::COPY &&
MI->getOperand(1).getSubReg() == AArch64::sub_32) &&
"Expected copy instruction");
Reg = MI->getOperand(1).getReg();
MI->eraseFromParent();
}
updateValueMap(I, Reg);
return true;
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool AArch64FastISel::selectIntExt(const Instruction *I) {
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
assert((isa<ZExtInst>(I) || isa<SExtInst>(I)) &&
"Unexpected integer extend instruction.");
MVT RetVT;
MVT SrcVT;
if (!isTypeSupported(I->getType(), RetVT))
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
if (!isTypeSupported(I->getOperand(0)->getType(), SrcVT))
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
[FastISel][AArch64] Fix sign-/zero-extend folding when SelectionDAG is involved. Sign-/zero-extend folding depended on the load and the integer extend to be both selected by FastISel. This cannot always be garantueed and SelectionDAG might interfer. This commit adds additonal checks to load and integer extend lowering to catch this. Related to rdar://problem/18495928. llvm-svn: 219716
2014-10-15 04:36:02 +08:00
// Try to optimize already sign-/zero-extended values from load instructions.
if (optimizeIntExtLoad(I, RetVT, SrcVT))
return true;
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
unsigned SrcReg = getRegForValue(I->getOperand(0));
if (!SrcReg)
return false;
[FastISel][AArch64] Fix "Fold sign-/zero-extends into the load instruction." This commit fixes an issue with sign-/zero-extending loads that was discovered by Richard Barton. We use now the correct load instructions for sign-extending loads to 64bit. Also updated and added more unit tests. llvm-svn: 219185
2014-10-07 11:39:59 +08:00
bool SrcIsKill = hasTrivialKill(I->getOperand(0));
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
[FastISel][AArch64] Fix sign-/zero-extend folding when SelectionDAG is involved. Sign-/zero-extend folding depended on the load and the integer extend to be both selected by FastISel. This cannot always be garantueed and SelectionDAG might interfer. This commit adds additonal checks to load and integer extend lowering to catch this. Related to rdar://problem/18495928. llvm-svn: 219716
2014-10-15 04:36:02 +08:00
// Try to optimize already sign-/zero-extended values from function arguments.
[FastISel][AArch64] Fix "Fold sign-/zero-extends into the load instruction." This commit fixes an issue with sign-/zero-extending loads that was discovered by Richard Barton. We use now the correct load instructions for sign-extending loads to 64bit. Also updated and added more unit tests. llvm-svn: 219185
2014-10-07 11:39:59 +08:00
bool IsZExt = isa<ZExtInst>(I);
if (const auto *Arg = dyn_cast<Argument>(I->getOperand(0))) {
if ((IsZExt && Arg->hasZExtAttr()) || (!IsZExt && Arg->hasSExtAttr())) {
if (RetVT == MVT::i64 && SrcVT != MVT::i64) {
unsigned ResultReg = createResultReg(&AArch64::GPR64RegClass);
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc,
TII.get(AArch64::SUBREG_TO_REG), ResultReg)
.addImm(0)
.addReg(SrcReg, getKillRegState(SrcIsKill))
.addImm(AArch64::sub_32);
SrcReg = ResultReg;
}
[AArch64][FastISel] Fix kill flags for integer extends. In the case we optimize an integer extend away and replace it directly with the source register, we also have to clear all kill flags at all its uses. This is necessary, because the orignal IR instruction might be trivially dead, but we replaced it with a nop at MI level. llvm-svn: 221628
2014-11-11 05:05:31 +08:00
// Conservatively clear all kill flags from all uses, because we are
// replacing a sign-/zero-extend instruction at IR level with a nop at MI
// level. The result of the instruction at IR level might have been
// trivially dead, which is now not longer true.
unsigned UseReg = lookUpRegForValue(I);
if (UseReg)
MRI.clearKillFlags(UseReg);
[FastISel][AArch64] Fix "Fold sign-/zero-extends into the load instruction." This commit fixes an issue with sign-/zero-extending loads that was discovered by Richard Barton. We use now the correct load instructions for sign-extending loads to 64bit. Also updated and added more unit tests. llvm-svn: 219185
2014-10-07 11:39:59 +08:00
updateValueMap(I, SrcReg);
return true;
}
}
[FastISel][AArch64] Don't perform sign-/zero-extension for function arguments that have already been sign-/zero-extended. llvm-svn: 214844
2014-08-05 13:43:44 +08:00
[FastISel][AArch64] Fix "Fold sign-/zero-extends into the load instruction." This commit fixes an issue with sign-/zero-extending loads that was discovered by Richard Barton. We use now the correct load instructions for sign-extending loads to 64bit. Also updated and added more unit tests. llvm-svn: 219185
2014-10-07 11:39:59 +08:00
unsigned ResultReg = emitIntExt(SrcVT, SrcReg, RetVT, IsZExt);
[FastISel][AArch64] Don't perform sign-/zero-extension for function arguments that have already been sign-/zero-extended. llvm-svn: 214844
2014-08-05 13:43:44 +08:00
if (!ResultReg)
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
[FastISel][AArch64] Don't perform sign-/zero-extension for function arguments that have already been sign-/zero-extended. llvm-svn: 214844
2014-08-05 13:43:44 +08:00
[FastISel] Rename public visible FastISel functions. NFC. This commit renames the following public FastISel functions: LowerArguments -> lowerArguments SelectInstruction -> selectInstruction TargetSelectInstruction -> fastSelectInstruction FastLowerArguments -> fastLowerArguments FastLowerCall -> fastLowerCall FastLowerIntrinsicCall -> fastLowerIntrinsicCall FastEmitZExtFromI1 -> fastEmitZExtFromI1 FastEmitBranch -> fastEmitBranch UpdateValueMap -> updateValueMap TargetMaterializeConstant -> fastMaterializeConstant TargetMaterializeAlloca -> fastMaterializeAlloca TargetMaterializeFloatZero -> fastMaterializeFloatZero LowerCallTo -> lowerCallTo Reviewed by Eric llvm-svn: 217074
2014-09-04 04:56:52 +08:00
updateValueMap(I, ResultReg);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return true;
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool AArch64FastISel::selectRem(const Instruction *I, unsigned ISDOpcode) {
Make TargetLowering::getPointerTy() taking DataLayout as an argument Summary: This change is part of a series of commits dedicated to have a single DataLayout during compilation by using always the one owned by the module. Reviewers: echristo Subscribers: jholewinski, ted, yaron.keren, rafael, llvm-commits Differential Revision: http://reviews.llvm.org/D11028 From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 241775
2015-07-09 10:09:04 +08:00
EVT DestEVT = TLI.getValueType(DL, I->getType(), true);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
if (!DestEVT.isSimple())
return false;
MVT DestVT = DestEVT.getSimpleVT();
if (DestVT != MVT::i64 && DestVT != MVT::i32)
return false;
unsigned DivOpc;
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool Is64bit = (DestVT == MVT::i64);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
switch (ISDOpcode) {
default:
return false;
case ISD::SREM:
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
DivOpc = Is64bit ? AArch64::SDIVXr : AArch64::SDIVWr;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
break;
case ISD::UREM:
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
DivOpc = Is64bit ? AArch64::UDIVXr : AArch64::UDIVWr;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
break;
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
unsigned MSubOpc = Is64bit ? AArch64::MSUBXrrr : AArch64::MSUBWrrr;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
unsigned Src0Reg = getRegForValue(I->getOperand(0));
if (!Src0Reg)
return false;
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
bool Src0IsKill = hasTrivialKill(I->getOperand(0));
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
unsigned Src1Reg = getRegForValue(I->getOperand(1));
if (!Src1Reg)
return false;
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
bool Src1IsKill = hasTrivialKill(I->getOperand(1));
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
const TargetRegisterClass *RC =
(DestVT == MVT::i64) ? &AArch64::GPR64RegClass : &AArch64::GPR32RegClass;
[FastISel][tblgen] Rename tblgen generated FastISel functions. NFC. This is the final round of renaming. This changes tblgen to emit lower-case function names for FastEmitInst_* and FastEmit_*, and updates all its uses in the source code. Reviewed by Eric llvm-svn: 217075
2014-09-04 04:56:59 +08:00
unsigned QuotReg = fastEmitInst_rr(DivOpc, RC, Src0Reg, /*IsKill=*/false,
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
Src1Reg, /*IsKill=*/false);
assert(QuotReg && "Unexpected DIV instruction emission failure.");
Replacing a non-ASCII character in a comment with an ASCII character. Fixes a C4819 warning in MSVC. llvm-svn: 206403
2014-04-17 01:09:20 +08:00
// The remainder is computed as numerator - (quotient * denominator) using the
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// MSUB instruction.
[FastISel][tblgen] Rename tblgen generated FastISel functions. NFC. This is the final round of renaming. This changes tblgen to emit lower-case function names for FastEmitInst_* and FastEmit_*, and updates all its uses in the source code. Reviewed by Eric llvm-svn: 217075
2014-09-04 04:56:59 +08:00
unsigned ResultReg = fastEmitInst_rrr(MSubOpc, RC, QuotReg, /*IsKill=*/true,
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
Src1Reg, Src1IsKill, Src0Reg,
Src0IsKill);
[FastISel] Rename public visible FastISel functions. NFC. This commit renames the following public FastISel functions: LowerArguments -> lowerArguments SelectInstruction -> selectInstruction TargetSelectInstruction -> fastSelectInstruction FastLowerArguments -> fastLowerArguments FastLowerCall -> fastLowerCall FastLowerIntrinsicCall -> fastLowerIntrinsicCall FastEmitZExtFromI1 -> fastEmitZExtFromI1 FastEmitBranch -> fastEmitBranch UpdateValueMap -> updateValueMap TargetMaterializeConstant -> fastMaterializeConstant TargetMaterializeAlloca -> fastMaterializeAlloca TargetMaterializeFloatZero -> fastMaterializeFloatZero LowerCallTo -> lowerCallTo Reviewed by Eric llvm-svn: 217074
2014-09-04 04:56:52 +08:00
updateValueMap(I, ResultReg);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return true;
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool AArch64FastISel::selectMul(const Instruction *I) {
[FastISel][AArch64] Simplify mul to shift when possible. This is related to rdar://problem/18369687. llvm-svn: 217980
2014-09-18 04:35:41 +08:00
MVT VT;
if (!isTypeSupported(I->getType(), VT, /*IsVectorAllowed=*/true))
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return false;
[FastISel][AArch64] Simplify mul to shift when possible. This is related to rdar://problem/18369687. llvm-svn: 217980
2014-09-18 04:35:41 +08:00
if (VT.isVector())
return selectBinaryOp(I, ISD::MUL);
const Value *Src0 = I->getOperand(0);
const Value *Src1 = I->getOperand(1);
if (const auto *C = dyn_cast<ConstantInt>(Src0))
if (C->getValue().isPowerOf2())
std::swap(Src0, Src1);
// Try to simplify to a shift instruction.
if (const auto *C = dyn_cast<ConstantInt>(Src1))
if (C->getValue().isPowerOf2()) {
uint64_t ShiftVal = C->getValue().logBase2();
MVT SrcVT = VT;
bool IsZExt = true;
if (const auto *ZExt = dyn_cast<ZExtInst>(Src0)) {
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
if (!isIntExtFree(ZExt)) {
MVT VT;
if (isValueAvailable(ZExt) && isTypeSupported(ZExt->getSrcTy(), VT)) {
SrcVT = VT;
IsZExt = true;
Src0 = ZExt->getOperand(0);
}
[FastISel][AArch64] Simplify mul to shift when possible. This is related to rdar://problem/18369687. llvm-svn: 217980
2014-09-18 04:35:41 +08:00
}
} else if (const auto *SExt = dyn_cast<SExtInst>(Src0)) {
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
if (!isIntExtFree(SExt)) {
MVT VT;
if (isValueAvailable(SExt) && isTypeSupported(SExt->getSrcTy(), VT)) {
SrcVT = VT;
IsZExt = false;
Src0 = SExt->getOperand(0);
}
[FastISel][AArch64] Simplify mul to shift when possible. This is related to rdar://problem/18369687. llvm-svn: 217980
2014-09-18 04:35:41 +08:00
}
}
unsigned Src0Reg = getRegForValue(Src0);
if (!Src0Reg)
return false;
bool Src0IsKill = hasTrivialKill(Src0);
unsigned ResultReg =
emitLSL_ri(VT, SrcVT, Src0Reg, Src0IsKill, ShiftVal, IsZExt);
if (ResultReg) {
updateValueMap(I, ResultReg);
return true;
}
}
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
unsigned Src0Reg = getRegForValue(I->getOperand(0));
if (!Src0Reg)
return false;
[FastISel][AArch64] Create helper functions to create the various multiplies on AArch64. llvm-svn: 214346
2014-07-31 06:04:25 +08:00
bool Src0IsKill = hasTrivialKill(I->getOperand(0));
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
unsigned Src1Reg = getRegForValue(I->getOperand(1));
if (!Src1Reg)
return false;
[FastISel][AArch64] Create helper functions to create the various multiplies on AArch64. llvm-svn: 214346
2014-07-31 06:04:25 +08:00
bool Src1IsKill = hasTrivialKill(I->getOperand(1));
[FastISel][AArch64] Simplify mul to shift when possible. This is related to rdar://problem/18369687. llvm-svn: 217980
2014-09-18 04:35:41 +08:00
unsigned ResultReg = emitMul_rr(VT, Src0Reg, Src0IsKill, Src1Reg, Src1IsKill);
[FastISel][AArch64] Create helper functions to create the various multiplies on AArch64. llvm-svn: 214346
2014-07-31 06:04:25 +08:00
if (!ResultReg)
return false;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
[FastISel] Rename public visible FastISel functions. NFC. This commit renames the following public FastISel functions: LowerArguments -> lowerArguments SelectInstruction -> selectInstruction TargetSelectInstruction -> fastSelectInstruction FastLowerArguments -> fastLowerArguments FastLowerCall -> fastLowerCall FastLowerIntrinsicCall -> fastLowerIntrinsicCall FastEmitZExtFromI1 -> fastEmitZExtFromI1 FastEmitBranch -> fastEmitBranch UpdateValueMap -> updateValueMap TargetMaterializeConstant -> fastMaterializeConstant TargetMaterializeAlloca -> fastMaterializeAlloca TargetMaterializeFloatZero -> fastMaterializeFloatZero LowerCallTo -> lowerCallTo Reviewed by Eric llvm-svn: 217074
2014-09-04 04:56:52 +08:00
updateValueMap(I, ResultReg);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
return true;
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool AArch64FastISel::selectShift(const Instruction *I) {
[FastISel][AArch64] Fold Sign-/Zero-Extend into the shift immediate instruction. llvm-svn: 216510
2014-08-27 08:58:26 +08:00
MVT RetVT;
[FastISel][AArch64] Refactor selectAddSub, selectLogicalOp, and SelectShift. NFC. Small refactor to tidy up the code a little. llvm-svn: 217827
2014-09-16 05:27:56 +08:00
if (!isTypeSupported(I->getType(), RetVT, /*IsVectorAllowed=*/true))
[FastISel][AArch64] Add support for shift-immediate. Currently the shift-immediate versions are not supported by tblgen and hopefully this can be later removed, once the required support has been added to tblgen. llvm-svn: 214345
2014-07-31 06:04:22 +08:00
return false;
[FastISel][AArch64] Refactor selectAddSub, selectLogicalOp, and SelectShift. NFC. Small refactor to tidy up the code a little. llvm-svn: 217827
2014-09-16 05:27:56 +08:00
if (RetVT.isVector())
return selectOperator(I, I->getOpcode());
[FastISel][AArch64] Add support for variable shift. This adds the missing variable shift support for value type i8, i16, and i32. This fixes <rdar://problem/18095685>. llvm-svn: 216242
2014-08-22 07:06:07 +08:00
if (const auto *C = dyn_cast<ConstantInt>(I->getOperand(1))) {
unsigned ResultReg = 0;
uint64_t ShiftVal = C->getZExtValue();
[FastISel][AArch64] Fold Sign-/Zero-Extend into the shift immediate instruction. llvm-svn: 216510
2014-08-27 08:58:26 +08:00
MVT SrcVT = RetVT;
Refactor: Simplify boolean expressions in AArch64 target Simplify boolean expressions using `true` and `false` with `clang-tidy` Patch by Richard Thomson. Reviewed By: rengolin Differential Revision: http://reviews.llvm.org/D8525 llvm-svn: 233089
2015-03-25 00:24:01 +08:00
bool IsZExt = I->getOpcode() != Instruction::AShr;
[FastISel][AArch64] Don't fold instructions that are not in the same basic block. This fix checks first if the instruction to be folded (e.g. sign-/zero-extend, or shift) is in the same machine basic block as the instruction we are folding into. Not doing so can result in incorrect code, because the value might not be live-out of the basic block, where the value is defined. This fixes rdar://problem/18169495. llvm-svn: 216700
2014-08-29 08:19:21 +08:00
const Value *Op0 = I->getOperand(0);
[FastISel][AArch64] Fold Sign-/Zero-Extend into the shift immediate instruction. llvm-svn: 216510
2014-08-27 08:58:26 +08:00
if (const auto *ZExt = dyn_cast<ZExtInst>(Op0)) {
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
if (!isIntExtFree(ZExt)) {
MVT TmpVT;
if (isValueAvailable(ZExt) && isTypeSupported(ZExt->getSrcTy(), TmpVT)) {
SrcVT = TmpVT;
IsZExt = true;
Op0 = ZExt->getOperand(0);
}
[FastISel][AArch64] Fold Sign-/Zero-Extend into the shift immediate instruction. llvm-svn: 216510
2014-08-27 08:58:26 +08:00
}
} else if (const auto *SExt = dyn_cast<SExtInst>(Op0)) {
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
if (!isIntExtFree(SExt)) {
MVT TmpVT;
if (isValueAvailable(SExt) && isTypeSupported(SExt->getSrcTy(), TmpVT)) {
SrcVT = TmpVT;
IsZExt = false;
Op0 = SExt->getOperand(0);
}
[FastISel][AArch64] Fold Sign-/Zero-Extend into the shift immediate instruction. llvm-svn: 216510
2014-08-27 08:58:26 +08:00
}
}
unsigned Op0Reg = getRegForValue(Op0);
if (!Op0Reg)
return false;
bool Op0IsKill = hasTrivialKill(Op0);
[FastISel][AArch64] Add support for variable shift. This adds the missing variable shift support for value type i8, i16, and i32. This fixes <rdar://problem/18095685>. llvm-svn: 216242
2014-08-22 07:06:07 +08:00
switch (I->getOpcode()) {
default: llvm_unreachable("Unexpected instruction.");
case Instruction::Shl:
[FastISel][AArch64] Fold Sign-/Zero-Extend into the shift immediate instruction. llvm-svn: 216510
2014-08-27 08:58:26 +08:00
ResultReg = emitLSL_ri(RetVT, SrcVT, Op0Reg, Op0IsKill, ShiftVal, IsZExt);
[FastISel][AArch64] Add support for variable shift. This adds the missing variable shift support for value type i8, i16, and i32. This fixes <rdar://problem/18095685>. llvm-svn: 216242
2014-08-22 07:06:07 +08:00
break;
case Instruction::AShr:
[FastISel][AArch64] Fold Sign-/Zero-Extend into the shift immediate instruction. llvm-svn: 216510
2014-08-27 08:58:26 +08:00
ResultReg = emitASR_ri(RetVT, SrcVT, Op0Reg, Op0IsKill, ShiftVal, IsZExt);
[FastISel][AArch64] Add support for variable shift. This adds the missing variable shift support for value type i8, i16, and i32. This fixes <rdar://problem/18095685>. llvm-svn: 216242
2014-08-22 07:06:07 +08:00
break;
case Instruction::LShr:
[FastISel][AArch64] Fold Sign-/Zero-Extend into the shift immediate instruction. llvm-svn: 216510
2014-08-27 08:58:26 +08:00
ResultReg = emitLSR_ri(RetVT, SrcVT, Op0Reg, Op0IsKill, ShiftVal, IsZExt);
[FastISel][AArch64] Add support for variable shift. This adds the missing variable shift support for value type i8, i16, and i32. This fixes <rdar://problem/18095685>. llvm-svn: 216242
2014-08-22 07:06:07 +08:00
break;
}
if (!ResultReg)
return false;
[FastISel][AArch64] Add support for shift-immediate. Currently the shift-immediate versions are not supported by tblgen and hopefully this can be later removed, once the required support has been added to tblgen. llvm-svn: 214345
2014-07-31 06:04:22 +08:00
[FastISel] Rename public visible FastISel functions. NFC. This commit renames the following public FastISel functions: LowerArguments -> lowerArguments SelectInstruction -> selectInstruction TargetSelectInstruction -> fastSelectInstruction FastLowerArguments -> fastLowerArguments FastLowerCall -> fastLowerCall FastLowerIntrinsicCall -> fastLowerIntrinsicCall FastEmitZExtFromI1 -> fastEmitZExtFromI1 FastEmitBranch -> fastEmitBranch UpdateValueMap -> updateValueMap TargetMaterializeConstant -> fastMaterializeConstant TargetMaterializeAlloca -> fastMaterializeAlloca TargetMaterializeFloatZero -> fastMaterializeFloatZero LowerCallTo -> lowerCallTo Reviewed by Eric llvm-svn: 217074
2014-09-04 04:56:52 +08:00
updateValueMap(I, ResultReg);
[FastISel][AArch64] Add support for variable shift. This adds the missing variable shift support for value type i8, i16, and i32. This fixes <rdar://problem/18095685>. llvm-svn: 216242
2014-08-22 07:06:07 +08:00
return true;
}
[FastISel][AArch64] Fold Sign-/Zero-Extend into the shift immediate instruction. llvm-svn: 216510
2014-08-27 08:58:26 +08:00
unsigned Op0Reg = getRegForValue(I->getOperand(0));
if (!Op0Reg)
return false;
bool Op0IsKill = hasTrivialKill(I->getOperand(0));
[FastISel][AArch64] Add support for variable shift. This adds the missing variable shift support for value type i8, i16, and i32. This fixes <rdar://problem/18095685>. llvm-svn: 216242
2014-08-22 07:06:07 +08:00
unsigned Op1Reg = getRegForValue(I->getOperand(1));
if (!Op1Reg)
return false;
bool Op1IsKill = hasTrivialKill(I->getOperand(1));
unsigned ResultReg = 0;
switch (I->getOpcode()) {
default: llvm_unreachable("Unexpected instruction.");
case Instruction::Shl:
ResultReg = emitLSL_rr(RetVT, Op0Reg, Op0IsKill, Op1Reg, Op1IsKill);
break;
case Instruction::AShr:
ResultReg = emitASR_rr(RetVT, Op0Reg, Op0IsKill, Op1Reg, Op1IsKill);
break;
case Instruction::LShr:
ResultReg = emitLSR_rr(RetVT, Op0Reg, Op0IsKill, Op1Reg, Op1IsKill);
break;
[FastISel][AArch64] Add support for shift-immediate. Currently the shift-immediate versions are not supported by tblgen and hopefully this can be later removed, once the required support has been added to tblgen. llvm-svn: 214345
2014-07-31 06:04:22 +08:00
}
if (!ResultReg)
return false;
[FastISel] Rename public visible FastISel functions. NFC. This commit renames the following public FastISel functions: LowerArguments -> lowerArguments SelectInstruction -> selectInstruction TargetSelectInstruction -> fastSelectInstruction FastLowerArguments -> fastLowerArguments FastLowerCall -> fastLowerCall FastLowerIntrinsicCall -> fastLowerIntrinsicCall FastEmitZExtFromI1 -> fastEmitZExtFromI1 FastEmitBranch -> fastEmitBranch UpdateValueMap -> updateValueMap TargetMaterializeConstant -> fastMaterializeConstant TargetMaterializeAlloca -> fastMaterializeAlloca TargetMaterializeFloatZero -> fastMaterializeFloatZero LowerCallTo -> lowerCallTo Reviewed by Eric llvm-svn: 217074
2014-09-04 04:56:52 +08:00
updateValueMap(I, ResultReg);
[FastISel][AArch64] Add support for shift-immediate. Currently the shift-immediate versions are not supported by tblgen and hopefully this can be later removed, once the required support has been added to tblgen. llvm-svn: 214345
2014-07-31 06:04:22 +08:00
return true;
}
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
bool AArch64FastISel::selectBitCast(const Instruction *I) {
[FastISel][AArch64] Add basic bitcast support for conversion between float and int. Fixes <rdar://problem/17867078>. llvm-svn: 214389
2014-07-31 14:25:37 +08:00
MVT RetVT, SrcVT;
if (!isTypeLegal(I->getOperand(0)->getType(), SrcVT))
return false;
if (!isTypeLegal(I->getType(), RetVT))
return false;
unsigned Opc;
if (RetVT == MVT::f32 && SrcVT == MVT::i32)
Opc = AArch64::FMOVWSr;
else if (RetVT == MVT::f64 && SrcVT == MVT::i64)
Opc = AArch64::FMOVXDr;
else if (RetVT == MVT::i32 && SrcVT == MVT::f32)
Opc = AArch64::FMOVSWr;
else if (RetVT == MVT::i64 && SrcVT == MVT::f64)
Opc = AArch64::FMOVDXr;
else
return false;
[FastISel][AArch64] Use the correct register class to make the MI verifier happy. This is mostly achieved by providing the correct register class manually, because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and MVT::i64. Also cleanup the code to use the FastEmitInst_* method whenever possible. This makes sure that the operands' register class is properly constrained. For all the remaining cases this adds the missing constrainOperandRegClass calls for each operand. llvm-svn: 216225
2014-08-22 04:57:57 +08:00
const TargetRegisterClass *RC = nullptr;
switch (RetVT.SimpleTy) {
default: llvm_unreachable("Unexpected value type.");
case MVT::i32: RC = &AArch64::GPR32RegClass; break;
case MVT::i64: RC = &AArch64::GPR64RegClass; break;
case MVT::f32: RC = &AArch64::FPR32RegClass; break;
case MVT::f64: RC = &AArch64::FPR64RegClass; break;
}
[FastISel][AArch64] Add basic bitcast support for conversion between float and int. Fixes <rdar://problem/17867078>. llvm-svn: 214389
2014-07-31 14:25:37 +08:00
unsigned Op0Reg = getRegForValue(I->getOperand(0));
if (!Op0Reg)
return false;
bool Op0IsKill = hasTrivialKill(I->getOperand(0));
[FastISel][tblgen] Rename tblgen generated FastISel functions. NFC. This is the final round of renaming. This changes tblgen to emit lower-case function names for FastEmitInst_* and FastEmit_*, and updates all its uses in the source code. Reviewed by Eric llvm-svn: 217075
2014-09-04 04:56:59 +08:00
unsigned ResultReg = fastEmitInst_r(Opc, RC, Op0Reg, Op0IsKill);
[FastISel][AArch64] Add basic bitcast support for conversion between float and int. Fixes <rdar://problem/17867078>. llvm-svn: 214389
2014-07-31 14:25:37 +08:00
if (!ResultReg)
return false;
[FastISel] Rename public visible FastISel functions. NFC. This commit renames the following public FastISel functions: LowerArguments -> lowerArguments SelectInstruction -> selectInstruction TargetSelectInstruction -> fastSelectInstruction FastLowerArguments -> fastLowerArguments FastLowerCall -> fastLowerCall FastLowerIntrinsicCall -> fastLowerIntrinsicCall FastEmitZExtFromI1 -> fastEmitZExtFromI1 FastEmitBranch -> fastEmitBranch UpdateValueMap -> updateValueMap TargetMaterializeConstant -> fastMaterializeConstant TargetMaterializeAlloca -> fastMaterializeAlloca TargetMaterializeFloatZero -> fastMaterializeFloatZero LowerCallTo -> lowerCallTo Reviewed by Eric llvm-svn: 217074
2014-09-04 04:56:52 +08:00
updateValueMap(I, ResultReg);
[FastISel][AArch64] Add basic bitcast support for conversion between float and int. Fixes <rdar://problem/17867078>. llvm-svn: 214389
2014-07-31 14:25:37 +08:00
return true;
}
[FastISel][AArch64] Add lowering support for frem. This lowers frem to a runtime libcall inside fast-isel. The test case also checks the CallLoweringInfo bug that was exposed by this change. This fixes rdar://problem/18342783. llvm-svn: 217833
2014-09-16 06:07:49 +08:00
bool AArch64FastISel::selectFRem(const Instruction *I) {
MVT RetVT;
if (!isTypeLegal(I->getType(), RetVT))
return false;
RTLIB::Libcall LC;
switch (RetVT.SimpleTy) {
default:
return false;
case MVT::f32:
LC = RTLIB::REM_F32;
break;
case MVT::f64:
LC = RTLIB::REM_F64;
break;
}
ArgListTy Args;
Args.reserve(I->getNumOperands());
// Populate the argument list.
for (auto &Arg : I->operands()) {
ArgListEntry Entry;
Entry.Val = Arg;
Entry.Ty = Arg->getType();
Args.push_back(Entry);
}
CallLoweringInfo CLI;
Use MCSymbols for FastISel. The summary is that it moves the mangling earlier and replaces a few calls to .addExternalSymbol with addSym. I originally wanted to replace all the uses of addExternalSymbol with addSym, but noticed it was a lot of work and doesn't need to be done all at once. llvm-svn: 240395
2015-06-23 20:21:54 +08:00
MCContext &Ctx = MF->getContext();
CLI.setCallee(DL, Ctx, TLI.getLibcallCallingConv(LC), I->getType(),
[FastISel][AArch64] Add lowering support for frem. This lowers frem to a runtime libcall inside fast-isel. The test case also checks the CallLoweringInfo bug that was exposed by this change. This fixes rdar://problem/18342783. llvm-svn: 217833
2014-09-16 06:07:49 +08:00
TLI.getLibcallName(LC), std::move(Args));
if (!lowerCallTo(CLI))
return false;
updateValueMap(I, CLI.ResultReg);
return true;
}
[FastISel][AArch64] Custom lower sdiv by power-of-2. Emit an optimized instruction sequence for sdiv by power-of-2 depending on the exact flag. This fixes rdar://problem/18224511. llvm-svn: 217986
2014-09-18 05:55:55 +08:00
bool AArch64FastISel::selectSDiv(const Instruction *I) {
MVT VT;
if (!isTypeLegal(I->getType(), VT))
return false;
if (!isa<ConstantInt>(I->getOperand(1)))
return selectBinaryOp(I, ISD::SDIV);
const APInt &C = cast<ConstantInt>(I->getOperand(1))->getValue();
if ((VT != MVT::i32 && VT != MVT::i64) || !C ||
!(C.isPowerOf2() || (-C).isPowerOf2()))
return selectBinaryOp(I, ISD::SDIV);
unsigned Lg2 = C.countTrailingZeros();
unsigned Src0Reg = getRegForValue(I->getOperand(0));
if (!Src0Reg)
return false;
bool Src0IsKill = hasTrivialKill(I->getOperand(0));
if (cast<BinaryOperator>(I)->isExact()) {
unsigned ResultReg = emitASR_ri(VT, VT, Src0Reg, Src0IsKill, Lg2);
if (!ResultReg)
return false;
updateValueMap(I, ResultReg);
return true;
}
[AArch64] Fix miscompile of sdiv-by-power-of-2. When the constant divisor was larger than 32bits, then the optimized code generated for the AArch64 backend would emit the wrong code, because the shift was defined as a shift of a 32bit constant '(1<<Lg2(divisor))' and we would loose the upper 32bits. This fixes rdar://problem/18678801. llvm-svn: 219934
2014-10-17 00:41:15 +08:00
int64_t Pow2MinusOne = (1ULL << Lg2) - 1;
unsigned AddReg = emitAdd_ri_(VT, Src0Reg, /*IsKill=*/false, Pow2MinusOne);
[FastISel][AArch64] Custom lower sdiv by power-of-2. Emit an optimized instruction sequence for sdiv by power-of-2 depending on the exact flag. This fixes rdar://problem/18224511. llvm-svn: 217986
2014-09-18 05:55:55 +08:00
if (!AddReg)
return false;
// (Src0 < 0) ? Pow2 - 1 : 0;
if (!emitICmp_ri(VT, Src0Reg, /*IsKill=*/false, 0))
return false;
unsigned SelectOpc;
const TargetRegisterClass *RC;
if (VT == MVT::i64) {
SelectOpc = AArch64::CSELXr;
RC = &AArch64::GPR64RegClass;
} else {
SelectOpc = AArch64::CSELWr;
RC = &AArch64::GPR32RegClass;
}
unsigned SelectReg =
fastEmitInst_rri(SelectOpc, RC, AddReg, /*IsKill=*/true, Src0Reg,
Src0IsKill, AArch64CC::LT);
if (!SelectReg)
return false;
// Divide by Pow2 --> ashr. If we're dividing by a negative value we must also
// negate the result.
unsigned ZeroReg = (VT == MVT::i64) ? AArch64::XZR : AArch64::WZR;
unsigned ResultReg;
if (C.isNegative())
ResultReg = emitAddSub_rs(/*UseAdd=*/false, VT, ZeroReg, /*IsKill=*/true,
SelectReg, /*IsKill=*/true, AArch64_AM::ASR, Lg2);
else
ResultReg = emitASR_ri(VT, VT, SelectReg, /*IsKill=*/true, Lg2);
if (!ResultReg)
return false;
updateValueMap(I, ResultReg);
return true;
}
[FastISel][AArch64] Don't bail during simple GEP instruction selection. The generic FastISel code would bail, because it can't emit a sign-extend for AArch64. This copies the code over and uses AArch64 specific emit functions. This is not ideal and 'computeAddress' should handles this, so it can fold the address computation into the memory operation. I plan to clean up 'computeAddress' anyways, so I will add that in a future commit. Related to rdar://problem/18962471. llvm-svn: 221923
2014-11-14 04:50:44 +08:00
/// This is mostly a copy of the existing FastISel getRegForGEPIndex code. We
/// have to duplicate it for AArch64, because otherwise we would fail during the
/// sign-extend emission.
std::pair<unsigned, bool> AArch64FastISel::getRegForGEPIndex(const Value *Idx) {
unsigned IdxN = getRegForValue(Idx);
if (IdxN == 0)
// Unhandled operand. Halt "fast" selection and bail.
return std::pair<unsigned, bool>(0, false);
bool IdxNIsKill = hasTrivialKill(Idx);
// If the index is smaller or larger than intptr_t, truncate or extend it.
Make TargetLowering::getPointerTy() taking DataLayout as an argument Summary: This change is part of a series of commits dedicated to have a single DataLayout during compilation by using always the one owned by the module. Reviewers: echristo Subscribers: jholewinski, ted, yaron.keren, rafael, llvm-commits Differential Revision: http://reviews.llvm.org/D11028 From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 241775
2015-07-09 10:09:04 +08:00
MVT PtrVT = TLI.getPointerTy(DL);
[FastISel][AArch64] Don't bail during simple GEP instruction selection. The generic FastISel code would bail, because it can't emit a sign-extend for AArch64. This copies the code over and uses AArch64 specific emit functions. This is not ideal and 'computeAddress' should handles this, so it can fold the address computation into the memory operation. I plan to clean up 'computeAddress' anyways, so I will add that in a future commit. Related to rdar://problem/18962471. llvm-svn: 221923
2014-11-14 04:50:44 +08:00
EVT IdxVT = EVT::getEVT(Idx->getType(), /*HandleUnknown=*/false);
if (IdxVT.bitsLT(PtrVT)) {
IdxN = emitIntExt(IdxVT.getSimpleVT(), IdxN, PtrVT, /*IsZExt=*/false);
IdxNIsKill = true;
} else if (IdxVT.bitsGT(PtrVT))
llvm_unreachable("AArch64 FastISel doesn't support types larger than i64");
return std::pair<unsigned, bool>(IdxN, IdxNIsKill);
}
Reapply "[FastISel][AArch64] Add custom lowering for GEPs." This is mostly a copy of the existing FastISel GEP code, but we have to duplicate it for AArch64, because otherwise we would bail out even for simple cases. This is because the standard fastEmit functions don't cover MUL at all and ADD is lowered very inefficientily. The original commit had a bug in the add emit logic, which has been fixed. llvm-svn: 219831
2014-10-16 02:58:07 +08:00
/// This is mostly a copy of the existing FastISel GEP code, but we have to
/// duplicate it for AArch64, because otherwise we would bail out even for
/// simple cases. This is because the standard fastEmit functions don't cover
/// MUL at all and ADD is lowered very inefficientily.
bool AArch64FastISel::selectGetElementPtr(const Instruction *I) {
unsigned N = getRegForValue(I->getOperand(0));
if (!N)
return false;
bool NIsKill = hasTrivialKill(I->getOperand(0));
// Keep a running tab of the total offset to coalesce multiple N = N + Offset
// into a single N = N + TotalOffset.
uint64_t TotalOffs = 0;
Make TargetLowering::getPointerTy() taking DataLayout as an argument Summary: This change is part of a series of commits dedicated to have a single DataLayout during compilation by using always the one owned by the module. Reviewers: echristo Subscribers: jholewinski, ted, yaron.keren, rafael, llvm-commits Differential Revision: http://reviews.llvm.org/D11028 From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 241775
2015-07-09 10:09:04 +08:00
MVT VT = TLI.getPointerTy(DL);
[NFC] Replace several manual GEP loops with gep_type_iterator. Reviewers: dblaikie Subscribers: llvm-commits Differential Revision: http://reviews.llvm.org/D16335 llvm-svn: 258262
2016-01-20 08:26:52 +08:00
for (gep_type_iterator GTI = gep_type_begin(I), E = gep_type_end(I);
GTI != E; ++GTI) {
const Value *Idx = GTI.getOperand();
if (auto *StTy = dyn_cast<StructType>(*GTI)) {
Reapply "[FastISel][AArch64] Add custom lowering for GEPs." This is mostly a copy of the existing FastISel GEP code, but we have to duplicate it for AArch64, because otherwise we would bail out even for simple cases. This is because the standard fastEmit functions don't cover MUL at all and ADD is lowered very inefficientily. The original commit had a bug in the add emit logic, which has been fixed. llvm-svn: 219831
2014-10-16 02:58:07 +08:00
unsigned Field = cast<ConstantInt>(Idx)->getZExtValue();
// N = N + Offset
if (Field)
TotalOffs += DL.getStructLayout(StTy)->getElementOffset(Field);
} else {
[NFC] Replace several manual GEP loops with gep_type_iterator. Reviewers: dblaikie Subscribers: llvm-commits Differential Revision: http://reviews.llvm.org/D16335 llvm-svn: 258262
2016-01-20 08:26:52 +08:00
Type *Ty = GTI.getIndexedType();
Reapply "[FastISel][AArch64] Add custom lowering for GEPs." This is mostly a copy of the existing FastISel GEP code, but we have to duplicate it for AArch64, because otherwise we would bail out even for simple cases. This is because the standard fastEmit functions don't cover MUL at all and ADD is lowered very inefficientily. The original commit had a bug in the add emit logic, which has been fixed. llvm-svn: 219831
2014-10-16 02:58:07 +08:00
// If this is a constant subscript, handle it quickly.
if (const auto *CI = dyn_cast<ConstantInt>(Idx)) {
if (CI->isZero())
continue;
// N = N + Offset
TotalOffs +=
DL.getTypeAllocSize(Ty) * cast<ConstantInt>(CI)->getSExtValue();
continue;
}
if (TotalOffs) {
N = emitAdd_ri_(VT, N, NIsKill, TotalOffs);
if (!N)
return false;
NIsKill = true;
TotalOffs = 0;
}
// N = N + Idx * ElementSize;
uint64_t ElementSize = DL.getTypeAllocSize(Ty);
std::pair<unsigned, bool> Pair = getRegForGEPIndex(Idx);
unsigned IdxN = Pair.first;
bool IdxNIsKill = Pair.second;
if (!IdxN)
return false;
if (ElementSize != 1) {
unsigned C = fastEmit_i(VT, VT, ISD::Constant, ElementSize);
if (!C)
return false;
IdxN = emitMul_rr(VT, IdxN, IdxNIsKill, C, true);
if (!IdxN)
return false;
IdxNIsKill = true;
}
N = fastEmit_rr(VT, VT, ISD::ADD, N, NIsKill, IdxN, IdxNIsKill);
if (!N)
return false;
}
}
if (TotalOffs) {
N = emitAdd_ri_(VT, N, NIsKill, TotalOffs);
if (!N)
return false;
}
updateValueMap(I, N);
return true;
}
[AArch64][FastISel] Select -O0 legal cmpxchg. At -O0, cmpxchg survives AtomicExpand: it's mostly straightforward to select it in fast-isel, and let the pseudo be expanded later. extractvalues on the result are the tricky part: the generic logic only works for legal types (and it would be painful to make it support illegal types), so we can only support i32/i64 cmpxchg. llvm-svn: 276183
2016-07-21 05:12:32 +08:00
bool AArch64FastISel::selectAtomicCmpXchg(const AtomicCmpXchgInst *I) {
assert(TM.getOptLevel() == CodeGenOpt::None &&
"cmpxchg survived AtomicExpand at optlevel > -O0");
auto *RetPairTy = cast<StructType>(I->getType());
Type *RetTy = RetPairTy->getTypeAtIndex(0U);
assert(RetPairTy->getTypeAtIndex(1U)->isIntegerTy(1) &&
"cmpxchg has a non-i1 status result");
MVT VT;
if (!isTypeLegal(RetTy, VT))
return false;
const TargetRegisterClass *ResRC;
unsigned Opc;
// This only supports i32/i64, because i8/i16 aren't legal, and the generic
// extractvalue selection doesn't support that.
if (VT == MVT::i32) {
Opc = AArch64::CMP_SWAP_32;
ResRC = &AArch64::GPR32RegClass;
} else if (VT == MVT::i64) {
Opc = AArch64::CMP_SWAP_64;
ResRC = &AArch64::GPR64RegClass;
} else {
return false;
}
const MCInstrDesc &II = TII.get(Opc);
const unsigned AddrReg = constrainOperandRegClass(
II, getRegForValue(I->getPointerOperand()), II.getNumDefs());
const unsigned DesiredReg = constrainOperandRegClass(
II, getRegForValue(I->getCompareOperand()), II.getNumDefs() + 1);
const unsigned NewReg = constrainOperandRegClass(
II, getRegForValue(I->getNewValOperand()), II.getNumDefs() + 2);
const unsigned ResultReg1 = createResultReg(ResRC);
const unsigned ResultReg2 = createResultReg(&AArch64::GPR32RegClass);
// FIXME: MachineMemOperand doesn't support cmpxchg yet.
BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, II)
.addReg(ResultReg1, RegState::Define)
.addReg(ResultReg2, RegState::Define)
.addReg(AddrReg)
.addReg(DesiredReg)
.addReg(NewReg);
assert((ResultReg1 + 1) == ResultReg2 && "Nonconsecutive result registers.");
updateValueMap(I, ResultReg1, 2);
return true;
}
[FastISel] Rename public visible FastISel functions. NFC. This commit renames the following public FastISel functions: LowerArguments -> lowerArguments SelectInstruction -> selectInstruction TargetSelectInstruction -> fastSelectInstruction FastLowerArguments -> fastLowerArguments FastLowerCall -> fastLowerCall FastLowerIntrinsicCall -> fastLowerIntrinsicCall FastEmitZExtFromI1 -> fastEmitZExtFromI1 FastEmitBranch -> fastEmitBranch UpdateValueMap -> updateValueMap TargetMaterializeConstant -> fastMaterializeConstant TargetMaterializeAlloca -> fastMaterializeAlloca TargetMaterializeFloatZero -> fastMaterializeFloatZero LowerCallTo -> lowerCallTo Reviewed by Eric llvm-svn: 217074
2014-09-04 04:56:52 +08:00
bool AArch64FastISel::fastSelectInstruction(const Instruction *I) {
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
switch (I->getOpcode()) {
default:
[FastISel][AArch64] Cleanup and simplify 'fastSelectInstruction'. NFC. llvm-svn: 217119
2014-09-04 09:29:21 +08:00
break;
[FastISel][AArch64] Move over to target-dependent instruction selection only. This change moves FastISel for AArch64 to target-dependent instruction selection only. This change replicates the existing target-independent behavior, therefore there are no changes to the unit tests or new tests. Future changes will take advantage of this change and update functionality and unit tests. llvm-svn: 216955
2014-09-03 05:32:54 +08:00
case Instruction::Add:
[FastISel][AArch64] Add target-dependent instruction selection for Add/Sub. There is already target-dependent instruction selection support for Adds/Subs to support compares and the intrinsics with overflow check. This takes advantage of the existing infrastructure to also support Add/Sub, which allows the folding of immediates, sign-/zero-extends, and shifts. This fixes rdar://problem/18207316. llvm-svn: 217007
2014-09-03 09:38:36 +08:00
case Instruction::Sub:
[FastISel][AArch64] Refactor selectAddSub, selectLogicalOp, and SelectShift. NFC. Small refactor to tidy up the code a little. llvm-svn: 217827
2014-09-16 05:27:56 +08:00
return selectAddSub(I);
[FastISel][AArch64] Move over to target-dependent instruction selection only. This change moves FastISel for AArch64 to target-dependent instruction selection only. This change replicates the existing target-independent behavior, therefore there are no changes to the unit tests or new tests. Future changes will take advantage of this change and update functionality and unit tests. llvm-svn: 216955
2014-09-03 05:32:54 +08:00
case Instruction::Mul:
[FastISel][AArch64] Simplify mul to shift when possible. This is related to rdar://problem/18369687. llvm-svn: 217980
2014-09-18 04:35:41 +08:00
return selectMul(I);
[FastISel][AArch64] Custom lower sdiv by power-of-2. Emit an optimized instruction sequence for sdiv by power-of-2 depending on the exact flag. This fixes rdar://problem/18224511. llvm-svn: 217986
2014-09-18 05:55:55 +08:00
case Instruction::SDiv:
return selectSDiv(I);
[FastISel][AArch64] Move over to target-dependent instruction selection only. This change moves FastISel for AArch64 to target-dependent instruction selection only. This change replicates the existing target-independent behavior, therefore there are no changes to the unit tests or new tests. Future changes will take advantage of this change and update functionality and unit tests. llvm-svn: 216955
2014-09-03 05:32:54 +08:00
case Instruction::SRem:
[FastISel] Some long overdue spring cleaning of FastISel. Things got a little bit messy over the years and it is time for a little bit spring cleaning. This first commit is focused on the FastISel base class itself. It doxyfies all comments, C++11fies the code where it makes sense, renames internal methods to adhere to the coding standard, and clang-formats the files. Reviewed by Eric llvm-svn: 217060
2014-09-04 02:46:45 +08:00
if (!selectBinaryOp(I, ISD::SREM))
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return selectRem(I, ISD::SREM);
[FastISel][AArch64] Move over to target-dependent instruction selection only. This change moves FastISel for AArch64 to target-dependent instruction selection only. This change replicates the existing target-independent behavior, therefore there are no changes to the unit tests or new tests. Future changes will take advantage of this change and update functionality and unit tests. llvm-svn: 216955
2014-09-03 05:32:54 +08:00
return true;
case Instruction::URem:
[FastISel] Some long overdue spring cleaning of FastISel. Things got a little bit messy over the years and it is time for a little bit spring cleaning. This first commit is focused on the FastISel base class itself. It doxyfies all comments, C++11fies the code where it makes sense, renames internal methods to adhere to the coding standard, and clang-formats the files. Reviewed by Eric llvm-svn: 217060
2014-09-04 02:46:45 +08:00
if (!selectBinaryOp(I, ISD::UREM))
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return selectRem(I, ISD::UREM);
[FastISel][AArch64] Move over to target-dependent instruction selection only. This change moves FastISel for AArch64 to target-dependent instruction selection only. This change replicates the existing target-independent behavior, therefore there are no changes to the unit tests or new tests. Future changes will take advantage of this change and update functionality and unit tests. llvm-svn: 216955
2014-09-03 05:32:54 +08:00
return true;
case Instruction::Shl:
case Instruction::LShr:
case Instruction::AShr:
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return selectShift(I);
[FastISel][AArch64] Move over to target-dependent instruction selection only. This change moves FastISel for AArch64 to target-dependent instruction selection only. This change replicates the existing target-independent behavior, therefore there are no changes to the unit tests or new tests. Future changes will take advantage of this change and update functionality and unit tests. llvm-svn: 216955
2014-09-03 05:32:54 +08:00
case Instruction::And:
case Instruction::Or:
case Instruction::Xor:
[FastISel][AArch64] Refactor selectAddSub, selectLogicalOp, and SelectShift. NFC. Small refactor to tidy up the code a little. llvm-svn: 217827
2014-09-16 05:27:56 +08:00
return selectLogicalOp(I);
[FastISel][AArch64] Move unconditional branch handling into 'SelectBranch'. NFC. llvm-svn: 217054
2014-09-04 01:58:10 +08:00
case Instruction::Br:
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return selectBranch(I);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
case Instruction::IndirectBr:
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return selectIndirectBr(I);
[FastISel][AArch64] Move over to target-dependent instruction selection only. This change moves FastISel for AArch64 to target-dependent instruction selection only. This change replicates the existing target-independent behavior, therefore there are no changes to the unit tests or new tests. Future changes will take advantage of this change and update functionality and unit tests. llvm-svn: 216955
2014-09-03 05:32:54 +08:00
case Instruction::BitCast:
[FastISel] Some long overdue spring cleaning of FastISel. Things got a little bit messy over the years and it is time for a little bit spring cleaning. This first commit is focused on the FastISel base class itself. It doxyfies all comments, C++11fies the code where it makes sense, renames internal methods to adhere to the coding standard, and clang-formats the files. Reviewed by Eric llvm-svn: 217060
2014-09-04 02:46:45 +08:00
if (!FastISel::selectBitCast(I))
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return selectBitCast(I);
[FastISel][AArch64] Move over to target-dependent instruction selection only. This change moves FastISel for AArch64 to target-dependent instruction selection only. This change replicates the existing target-independent behavior, therefore there are no changes to the unit tests or new tests. Future changes will take advantage of this change and update functionality and unit tests. llvm-svn: 216955
2014-09-03 05:32:54 +08:00
return true;
case Instruction::FPToSI:
[FastISel] Some long overdue spring cleaning of FastISel. Things got a little bit messy over the years and it is time for a little bit spring cleaning. This first commit is focused on the FastISel base class itself. It doxyfies all comments, C++11fies the code where it makes sense, renames internal methods to adhere to the coding standard, and clang-formats the files. Reviewed by Eric llvm-svn: 217060
2014-09-04 02:46:45 +08:00
if (!selectCast(I, ISD::FP_TO_SINT))
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return selectFPToInt(I, /*Signed=*/true);
[FastISel][AArch64] Move over to target-dependent instruction selection only. This change moves FastISel for AArch64 to target-dependent instruction selection only. This change replicates the existing target-independent behavior, therefore there are no changes to the unit tests or new tests. Future changes will take advantage of this change and update functionality and unit tests. llvm-svn: 216955
2014-09-03 05:32:54 +08:00
return true;
case Instruction::FPToUI:
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return selectFPToInt(I, /*Signed=*/false);
[FastISel][AArch64] Move over to target-dependent instruction selection only. This change moves FastISel for AArch64 to target-dependent instruction selection only. This change replicates the existing target-independent behavior, therefore there are no changes to the unit tests or new tests. Future changes will take advantage of this change and update functionality and unit tests. llvm-svn: 216955
2014-09-03 05:32:54 +08:00
case Instruction::ZExt:
case Instruction::SExt:
[FastISel][AArch64] Fold sign-/zero-extends into the load instruction. The sign-/zero-extension of the loaded value can be performed by the memory instruction for free. If the result of the load has only one use and the use is a sign-/zero-extend, then we emit the proper load instruction. The extend is only a register copy and will be optimized away later on. Other instructions that consume the sign-/zero-extended value are also made aware of this fact, so they don't fold the extend too. This fixes rdar://problem/18495928. llvm-svn: 218653
2014-09-30 08:49:58 +08:00
return selectIntExt(I);
[FastISel][AArch64] Move over to target-dependent instruction selection only. This change moves FastISel for AArch64 to target-dependent instruction selection only. This change replicates the existing target-independent behavior, therefore there are no changes to the unit tests or new tests. Future changes will take advantage of this change and update functionality and unit tests. llvm-svn: 216955
2014-09-03 05:32:54 +08:00
case Instruction::Trunc:
[FastISel] Some long overdue spring cleaning of FastISel. Things got a little bit messy over the years and it is time for a little bit spring cleaning. This first commit is focused on the FastISel base class itself. It doxyfies all comments, C++11fies the code where it makes sense, renames internal methods to adhere to the coding standard, and clang-formats the files. Reviewed by Eric llvm-svn: 217060
2014-09-04 02:46:45 +08:00
if (!selectCast(I, ISD::TRUNCATE))
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return selectTrunc(I);
[FastISel][AArch64] Move over to target-dependent instruction selection only. This change moves FastISel for AArch64 to target-dependent instruction selection only. This change replicates the existing target-independent behavior, therefore there are no changes to the unit tests or new tests. Future changes will take advantage of this change and update functionality and unit tests. llvm-svn: 216955
2014-09-03 05:32:54 +08:00
return true;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
case Instruction::FPExt:
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return selectFPExt(I);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
case Instruction::FPTrunc:
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return selectFPTrunc(I);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
case Instruction::SIToFP:
[FastISel] Some long overdue spring cleaning of FastISel. Things got a little bit messy over the years and it is time for a little bit spring cleaning. This first commit is focused on the FastISel base class itself. It doxyfies all comments, C++11fies the code where it makes sense, renames internal methods to adhere to the coding standard, and clang-formats the files. Reviewed by Eric llvm-svn: 217060
2014-09-04 02:46:45 +08:00
if (!selectCast(I, ISD::SINT_TO_FP))
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return selectIntToFP(I, /*Signed=*/true);
[FastISel][AArch64] Move over to target-dependent instruction selection only. This change moves FastISel for AArch64 to target-dependent instruction selection only. This change replicates the existing target-independent behavior, therefore there are no changes to the unit tests or new tests. Future changes will take advantage of this change and update functionality and unit tests. llvm-svn: 216955
2014-09-03 05:32:54 +08:00
return true;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
case Instruction::UIToFP:
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return selectIntToFP(I, /*Signed=*/false);
[FastISel][AArch64] Move over to target-dependent instruction selection only. This change moves FastISel for AArch64 to target-dependent instruction selection only. This change replicates the existing target-independent behavior, therefore there are no changes to the unit tests or new tests. Future changes will take advantage of this change and update functionality and unit tests. llvm-svn: 216955
2014-09-03 05:32:54 +08:00
case Instruction::Load:
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return selectLoad(I);
[FastISel][AArch64] Move over to target-dependent instruction selection only. This change moves FastISel for AArch64 to target-dependent instruction selection only. This change replicates the existing target-independent behavior, therefore there are no changes to the unit tests or new tests. Future changes will take advantage of this change and update functionality and unit tests. llvm-svn: 216955
2014-09-03 05:32:54 +08:00
case Instruction::Store:
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return selectStore(I);
[FastISel][AArch64] Move over to target-dependent instruction selection only. This change moves FastISel for AArch64 to target-dependent instruction selection only. This change replicates the existing target-independent behavior, therefore there are no changes to the unit tests or new tests. Future changes will take advantage of this change and update functionality and unit tests. llvm-svn: 216955
2014-09-03 05:32:54 +08:00
case Instruction::FCmp:
case Instruction::ICmp:
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return selectCmp(I);
[FastISel][AArch64] Move over to target-dependent instruction selection only. This change moves FastISel for AArch64 to target-dependent instruction selection only. This change replicates the existing target-independent behavior, therefore there are no changes to the unit tests or new tests. Future changes will take advantage of this change and update functionality and unit tests. llvm-svn: 216955
2014-09-03 05:32:54 +08:00
case Instruction::Select:
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return selectSelect(I);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
case Instruction::Ret:
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
return selectRet(I);
[FastISel][AArch64] Add lowering support for frem. This lowers frem to a runtime libcall inside fast-isel. The test case also checks the CallLoweringInfo bug that was exposed by this change. This fixes rdar://problem/18342783. llvm-svn: 217833
2014-09-16 06:07:49 +08:00
case Instruction::FRem:
return selectFRem(I);
Reapply "[FastISel][AArch64] Add custom lowering for GEPs." This is mostly a copy of the existing FastISel GEP code, but we have to duplicate it for AArch64, because otherwise we would bail out even for simple cases. This is because the standard fastEmit functions don't cover MUL at all and ADD is lowered very inefficientily. The original commit had a bug in the add emit logic, which has been fixed. llvm-svn: 219831
2014-10-16 02:58:07 +08:00
case Instruction::GetElementPtr:
return selectGetElementPtr(I);
[AArch64][FastISel] Select -O0 legal cmpxchg. At -O0, cmpxchg survives AtomicExpand: it's mostly straightforward to select it in fast-isel, and let the pseudo be expanded later. extractvalues on the result are the tricky part: the generic logic only works for legal types (and it would be painful to make it support illegal types), so we can only support i32/i64 cmpxchg. llvm-svn: 276183
2016-07-21 05:12:32 +08:00
case Instruction::AtomicCmpXchg:
return selectAtomicCmpXchg(cast<AtomicCmpXchgInst>(I));
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
}
[FastISel][AArch64] Move over to target-dependent instruction selection only. This change moves FastISel for AArch64 to target-dependent instruction selection only. This change replicates the existing target-independent behavior, therefore there are no changes to the unit tests or new tests. Future changes will take advantage of this change and update functionality and unit tests. llvm-svn: 216955
2014-09-03 05:32:54 +08:00
[FastISel][AArch64] Cleanup and simplify 'fastSelectInstruction'. NFC. llvm-svn: 217119
2014-09-04 09:29:21 +08:00
// fall-back to target-independent instruction selection.
return selectOperator(I, I->getOpcode());
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
// Silence warnings.
AArch64/ARM64: move ARM64 into AArch64's place This commit starts with a "git mv ARM64 AArch64" and continues out from there, renaming the C++ classes, intrinsics, and other target-local objects for consistency. "ARM64" test directories are also moved, and tests that began their life in ARM64 use an arm64 triple, those from AArch64 use an aarch64 triple. Both should be equivalent though. This finishes the AArch64 merge, and everyone should feel free to continue committing as normal now. llvm-svn: 209577
2014-05-24 20:50:23 +08:00
(void)&CC_AArch64_DarwinPCS_VarArg;
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
}
namespace llvm {
[FastISel][AArch64] Update function and variable names to follow the coding standard. NFC. llvm-svn: 217845
2014-09-16 07:20:17 +08:00
llvm::FastISel *AArch64::createFastISel(FunctionLoweringInfo &FuncInfo,
const TargetLibraryInfo *LibInfo) {
return new AArch64FastISel(FuncInfo, LibInfo);
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
}
}