1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-25 12:12:47 +01:00
llvm-mirror/lib/Target/X86/X86FixupSetCC.cpp
Harald van Dijk 2ccb3d8978 [X86] Do not allow FixupSetCC to relax constraints
The build bots caught two additional pre-existing problems exposed by the test change part of my change https://reviews.llvm.org/D91339, when expensive checks are enabled. https://reviews.llvm.org/D91924 fixes one of them, this fixes the other.

FixupSetCC will change code in the form of

  %setcc = SETCCr ...
  %ext1 = MOVZX32rr8 %setcc

to

  %zero = MOV32r0
  %setcc = SETCCr ...
  %ext2 = INSERT_SUBREG %zero, %setcc, %subreg.sub_8bit

and replace uses of %ext1 with %ext2.

The register class for %ext2 did not take into account any constraints on %ext1, which may have been required by its uses. This change ensures that the original constraints are honoured, by instead of creating a new %ext2 register, reusing %ext1 and further constraining it as needed. This requires a slight reorganisation to account for the fact that it is possible for the constraining to fail, in which case no changes should be made.

Reviewed By: RKSimon

Differential Revision: https://reviews.llvm.org/D91933
2020-11-28 17:46:56 +00:00

134 lines
4.6 KiB
C++

//===---- X86FixupSetCC.cpp - optimize usage of LEA instructions ----------===//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
//
// This file defines a pass that fixes zero-extension of setcc patterns.
// X86 setcc instructions are modeled to have no input arguments, and a single
// GR8 output argument. This is consistent with other similar instructions
// (e.g. movb), but means it is impossible to directly generate a setcc into
// the lower GR8 of a specified GR32.
// This means that ISel must select (zext (setcc)) into something like
// seta %al; movzbl %al, %eax.
// Unfortunately, this can cause a stall due to the partial register write
// performed by the setcc. Instead, we can use:
// xor %eax, %eax; seta %al
// This both avoids the stall, and encodes shorter.
//===----------------------------------------------------------------------===//
#include "X86.h"
#include "X86InstrInfo.h"
#include "X86Subtarget.h"
#include "llvm/ADT/Statistic.h"
#include "llvm/CodeGen/MachineFunctionPass.h"
#include "llvm/CodeGen/MachineInstrBuilder.h"
#include "llvm/CodeGen/MachineRegisterInfo.h"
using namespace llvm;
#define DEBUG_TYPE "x86-fixup-setcc"
STATISTIC(NumSubstZexts, "Number of setcc + zext pairs substituted");
namespace {
class X86FixupSetCCPass : public MachineFunctionPass {
public:
static char ID;
X86FixupSetCCPass() : MachineFunctionPass(ID) {}
StringRef getPassName() const override { return "X86 Fixup SetCC"; }
bool runOnMachineFunction(MachineFunction &MF) override;
private:
MachineRegisterInfo *MRI = nullptr;
const X86InstrInfo *TII = nullptr;
enum { SearchBound = 16 };
};
} // end anonymous namespace
char X86FixupSetCCPass::ID = 0;
INITIALIZE_PASS(X86FixupSetCCPass, DEBUG_TYPE, DEBUG_TYPE, false, false)
FunctionPass *llvm::createX86FixupSetCC() { return new X86FixupSetCCPass(); }
bool X86FixupSetCCPass::runOnMachineFunction(MachineFunction &MF) {
bool Changed = false;
MRI = &MF.getRegInfo();
TII = MF.getSubtarget<X86Subtarget>().getInstrInfo();
SmallVector<MachineInstr*, 4> ToErase;
for (auto &MBB : MF) {
MachineInstr *FlagsDefMI = nullptr;
for (auto &MI : MBB) {
// Remember the most recent preceding eflags defining instruction.
if (MI.definesRegister(X86::EFLAGS))
FlagsDefMI = &MI;
// Find a setcc that is used by a zext.
// This doesn't have to be the only use, the transformation is safe
// regardless.
if (MI.getOpcode() != X86::SETCCr)
continue;
MachineInstr *ZExt = nullptr;
for (auto &Use : MRI->use_instructions(MI.getOperand(0).getReg()))
if (Use.getOpcode() == X86::MOVZX32rr8)
ZExt = &Use;
if (!ZExt)
continue;
if (!FlagsDefMI)
continue;
// We'd like to put something that clobbers eflags directly before
// FlagsDefMI. This can't hurt anything after FlagsDefMI, because
// it, itself, by definition, clobbers eflags. But it may happen that
// FlagsDefMI also *uses* eflags, in which case the transformation is
// invalid.
if (FlagsDefMI->readsRegister(X86::EFLAGS))
continue;
// On 32-bit, we need to be careful to force an ABCD register.
const TargetRegisterClass *RC = MF.getSubtarget<X86Subtarget>().is64Bit()
? &X86::GR32RegClass
: &X86::GR32_ABCDRegClass;
if (!MRI->constrainRegClass(ZExt->getOperand(0).getReg(), RC)) {
// If we cannot constrain the register, we would need an additional copy
// and are better off keeping the MOVZX32rr8 we have now.
continue;
}
++NumSubstZexts;
Changed = true;
// Initialize a register with 0. This must go before the eflags def
Register ZeroReg = MRI->createVirtualRegister(RC);
BuildMI(MBB, FlagsDefMI, MI.getDebugLoc(), TII->get(X86::MOV32r0),
ZeroReg);
// X86 setcc only takes an output GR8, so fake a GR32 input by inserting
// the setcc result into the low byte of the zeroed register.
BuildMI(*ZExt->getParent(), ZExt, ZExt->getDebugLoc(),
TII->get(X86::INSERT_SUBREG), ZExt->getOperand(0).getReg())
.addReg(ZeroReg)
.addReg(MI.getOperand(0).getReg())
.addImm(X86::sub_8bit);
ToErase.push_back(ZExt);
}
}
for (auto &I : ToErase)
I->eraseFromParent();
return Changed;
}