llvm-project/llvm/lib/Target/XCore/XCoreTargetMachine.cpp

116 lines
3.5 KiB
C++
Raw Normal View History

2008-11-07 18:59:00 +08:00
//===-- XCoreTargetMachine.cpp - Define TargetMachine for XCore -----------===//
//
// 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
2008-11-07 18:59:00 +08:00
//
//===----------------------------------------------------------------------===//
//
//
//===----------------------------------------------------------------------===//
#include "XCoreTargetMachine.h"
#include "MCTargetDesc/XCoreMCTargetDesc.h"
#include "TargetInfo/XCoreTargetInfo.h"
#include "XCore.h"
#include "XCoreTargetObjectFile.h"
#include "XCoreTargetTransformInfo.h"
#include "llvm/ADT/Optional.h"
#include "llvm/ADT/STLExtras.h"
#include "llvm/Analysis/TargetTransformInfo.h"
#include "llvm/CodeGen/Passes.h"
#include "llvm/CodeGen/TargetPassConfig.h"
#include "llvm/Support/CodeGen.h"
#include "llvm/Support/TargetRegistry.h"
2008-11-07 18:59:00 +08:00
using namespace llvm;
static Reloc::Model getEffectiveRelocModel(Optional<Reloc::Model> RM) {
if (!RM.hasValue())
return Reloc::Static;
return *RM;
}
static CodeModel::Model
getEffectiveXCoreCodeModel(Optional<CodeModel::Model> CM) {
if (CM) {
if (*CM != CodeModel::Small && *CM != CodeModel::Large)
report_fatal_error("Target only supports CodeModel Small or Large");
return *CM;
}
return CodeModel::Small;
}
/// Create an ILP32 architecture model
2008-11-07 18:59:00 +08:00
///
XCoreTargetMachine::XCoreTargetMachine(const Target &T, const Triple &TT,
StringRef CPU, StringRef FS,
const TargetOptions &Options,
Optional<Reloc::Model> RM,
Optional<CodeModel::Model> CM,
CodeGenOpt::Level OL, bool JIT)
: LLVMTargetMachine(
T, "e-m:e-p:32:32-i1:8:32-i8:8:32-i16:16:32-i64:32-f64:32-a:0:32-n32",
TT, CPU, FS, Options, getEffectiveRelocModel(RM),
getEffectiveXCoreCodeModel(CM), OL),
TLOF(std::make_unique<XCoreTargetObjectFile>()),
Subtarget(TT, CPU, FS, *this) {
initAsmInfo();
2008-11-07 18:59:00 +08:00
}
XCoreTargetMachine::~XCoreTargetMachine() = default;
namespace {
/// XCore Code Generator Pass Configuration Options.
class XCorePassConfig : public TargetPassConfig {
public:
XCorePassConfig(XCoreTargetMachine &TM, PassManagerBase &PM)
: TargetPassConfig(TM, PM) {}
XCoreTargetMachine &getXCoreTargetMachine() const {
return getTM<XCoreTargetMachine>();
}
Erase fence insertion from SelectionDAGBuilder.cpp (NFC) Summary: Backends can use setInsertFencesForAtomic to signal to the middle-end that montonic is the only memory ordering they can accept for stores/loads/rmws/cmpxchg. The code lowering those accesses with a stronger ordering to fences + monotonic accesses is currently living in SelectionDAGBuilder.cpp. In this patch I propose moving this logic out of it for several reasons: - There is lots of redundancy to avoid: extremely similar logic already exists in AtomicExpand. - The current code in SelectionDAGBuilder does not use any target-hooks, it does the same transformation for every backend that requires it - As a result it is plain *unsound*, as it was apparently designed for ARM. It happens to mostly work for the other targets because they are extremely conservative, but Power for example had to switch to AtomicExpand to be able to use lwsync safely (see r218331). - Because it produces IR-level fences, it cannot be made sound ! This is noted in the C++11 standard (section 29.3, page 1140): ``` Fences cannot, in general, be used to restore sequential consistency for atomic operations with weaker ordering semantics. ``` It can also be seen by the following example (called IRIW in the litterature): ``` atomic<int> x = y = 0; int r1, r2, r3, r4; Thread 0: x.store(1); Thread 1: y.store(1); Thread 2: r1 = x.load(); r2 = y.load(); Thread 3: r3 = y.load(); r4 = x.load(); ``` r1 = r3 = 1 and r2 = r4 = 0 is impossible as long as the accesses are all seq_cst. But if they are lowered to monotonic accesses, no amount of fences can prevent it.. This patch does three things (I could cut it into parts, but then some of them would not be tested/testable, please tell me if you would prefer that): - it provides a default implementation for emitLeadingFence/emitTrailingFence in terms of IR-level fences, that mimic the original logic of SelectionDAGBuilder. As we saw above, this is unsound, but the best that can be done without knowing the targets well (and there is a comment warning about this risk). - it then switches Mips/Sparc/XCore to use AtomicExpand, relying on this default implementation (that exactly replicates the logic of SelectionDAGBuilder, so no functional change) - it finally erase this logic from SelectionDAGBuilder as it is dead-code. Ideally, each target would define its own override for emitLeading/TrailingFence using target-specific fences, but I do not know the Sparc/Mips/XCore memory model well enough to do this, and they appear to be dealing fine with the ARM-inspired default expansion for now (probably because they are overly conservative, as Power was). If anyone wants to compile fences more agressively on these platforms, the long comment should make it clear why he should first override emitLeading/TrailingFence. Test Plan: make check-all, no functional change Reviewers: jfb, t.p.northover Subscribers: aemerson, llvm-commits Differential Revision: http://reviews.llvm.org/D5474 llvm-svn: 219957
2014-10-17 04:34:57 +08:00
void addIRPasses() override;
bool addPreISel() override;
bool addInstSelector() override;
void addPreEmitPass() override;
};
} // end anonymous namespace
TargetPassConfig *XCoreTargetMachine::createPassConfig(PassManagerBase &PM) {
return new XCorePassConfig(*this, PM);
}
Erase fence insertion from SelectionDAGBuilder.cpp (NFC) Summary: Backends can use setInsertFencesForAtomic to signal to the middle-end that montonic is the only memory ordering they can accept for stores/loads/rmws/cmpxchg. The code lowering those accesses with a stronger ordering to fences + monotonic accesses is currently living in SelectionDAGBuilder.cpp. In this patch I propose moving this logic out of it for several reasons: - There is lots of redundancy to avoid: extremely similar logic already exists in AtomicExpand. - The current code in SelectionDAGBuilder does not use any target-hooks, it does the same transformation for every backend that requires it - As a result it is plain *unsound*, as it was apparently designed for ARM. It happens to mostly work for the other targets because they are extremely conservative, but Power for example had to switch to AtomicExpand to be able to use lwsync safely (see r218331). - Because it produces IR-level fences, it cannot be made sound ! This is noted in the C++11 standard (section 29.3, page 1140): ``` Fences cannot, in general, be used to restore sequential consistency for atomic operations with weaker ordering semantics. ``` It can also be seen by the following example (called IRIW in the litterature): ``` atomic<int> x = y = 0; int r1, r2, r3, r4; Thread 0: x.store(1); Thread 1: y.store(1); Thread 2: r1 = x.load(); r2 = y.load(); Thread 3: r3 = y.load(); r4 = x.load(); ``` r1 = r3 = 1 and r2 = r4 = 0 is impossible as long as the accesses are all seq_cst. But if they are lowered to monotonic accesses, no amount of fences can prevent it.. This patch does three things (I could cut it into parts, but then some of them would not be tested/testable, please tell me if you would prefer that): - it provides a default implementation for emitLeadingFence/emitTrailingFence in terms of IR-level fences, that mimic the original logic of SelectionDAGBuilder. As we saw above, this is unsound, but the best that can be done without knowing the targets well (and there is a comment warning about this risk). - it then switches Mips/Sparc/XCore to use AtomicExpand, relying on this default implementation (that exactly replicates the logic of SelectionDAGBuilder, so no functional change) - it finally erase this logic from SelectionDAGBuilder as it is dead-code. Ideally, each target would define its own override for emitLeading/TrailingFence using target-specific fences, but I do not know the Sparc/Mips/XCore memory model well enough to do this, and they appear to be dealing fine with the ARM-inspired default expansion for now (probably because they are overly conservative, as Power was). If anyone wants to compile fences more agressively on these platforms, the long comment should make it clear why he should first override emitLeading/TrailingFence. Test Plan: make check-all, no functional change Reviewers: jfb, t.p.northover Subscribers: aemerson, llvm-commits Differential Revision: http://reviews.llvm.org/D5474 llvm-svn: 219957
2014-10-17 04:34:57 +08:00
void XCorePassConfig::addIRPasses() {
addPass(createAtomicExpandPass());
Erase fence insertion from SelectionDAGBuilder.cpp (NFC) Summary: Backends can use setInsertFencesForAtomic to signal to the middle-end that montonic is the only memory ordering they can accept for stores/loads/rmws/cmpxchg. The code lowering those accesses with a stronger ordering to fences + monotonic accesses is currently living in SelectionDAGBuilder.cpp. In this patch I propose moving this logic out of it for several reasons: - There is lots of redundancy to avoid: extremely similar logic already exists in AtomicExpand. - The current code in SelectionDAGBuilder does not use any target-hooks, it does the same transformation for every backend that requires it - As a result it is plain *unsound*, as it was apparently designed for ARM. It happens to mostly work for the other targets because they are extremely conservative, but Power for example had to switch to AtomicExpand to be able to use lwsync safely (see r218331). - Because it produces IR-level fences, it cannot be made sound ! This is noted in the C++11 standard (section 29.3, page 1140): ``` Fences cannot, in general, be used to restore sequential consistency for atomic operations with weaker ordering semantics. ``` It can also be seen by the following example (called IRIW in the litterature): ``` atomic<int> x = y = 0; int r1, r2, r3, r4; Thread 0: x.store(1); Thread 1: y.store(1); Thread 2: r1 = x.load(); r2 = y.load(); Thread 3: r3 = y.load(); r4 = x.load(); ``` r1 = r3 = 1 and r2 = r4 = 0 is impossible as long as the accesses are all seq_cst. But if they are lowered to monotonic accesses, no amount of fences can prevent it.. This patch does three things (I could cut it into parts, but then some of them would not be tested/testable, please tell me if you would prefer that): - it provides a default implementation for emitLeadingFence/emitTrailingFence in terms of IR-level fences, that mimic the original logic of SelectionDAGBuilder. As we saw above, this is unsound, but the best that can be done without knowing the targets well (and there is a comment warning about this risk). - it then switches Mips/Sparc/XCore to use AtomicExpand, relying on this default implementation (that exactly replicates the logic of SelectionDAGBuilder, so no functional change) - it finally erase this logic from SelectionDAGBuilder as it is dead-code. Ideally, each target would define its own override for emitLeading/TrailingFence using target-specific fences, but I do not know the Sparc/Mips/XCore memory model well enough to do this, and they appear to be dealing fine with the ARM-inspired default expansion for now (probably because they are overly conservative, as Power was). If anyone wants to compile fences more agressively on these platforms, the long comment should make it clear why he should first override emitLeading/TrailingFence. Test Plan: make check-all, no functional change Reviewers: jfb, t.p.northover Subscribers: aemerson, llvm-commits Differential Revision: http://reviews.llvm.org/D5474 llvm-svn: 219957
2014-10-17 04:34:57 +08:00
TargetPassConfig::addIRPasses();
}
bool XCorePassConfig::addPreISel() {
addPass(createXCoreLowerThreadLocalPass());
return false;
}
bool XCorePassConfig::addInstSelector() {
addPass(createXCoreISelDag(getXCoreTargetMachine(), getOptLevel()));
2008-11-07 18:59:00 +08:00
return false;
}
void XCorePassConfig::addPreEmitPass() {
addPass(createXCoreFrameToArgsOffsetEliminationPass(), false);
}
// Force static initialization.
extern "C" void LLVMInitializeXCoreTarget() {
RegisterTargetMachine<XCoreTargetMachine> X(getTheXCoreTarget());
}
TargetTransformInfo
XCoreTargetMachine::getTargetTransformInfo(const Function &F) {
return TargetTransformInfo(XCoreTTIImpl(this, F));
}