llvm-project/llvm/lib/CodeGen/MachineDebugify.cpp

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

173 lines
6.5 KiB
C++
Raw Normal View History

Add MIR-level debugify with only locations support for now Summary: Re-used the IR-level debugify for the most part. The MIR-level code then adds locations to the MachineInstrs afterwards based on the LLVM-IR debug info. It's worth mentioning that the resulting locations make little sense as the range of line numbers used in a Function at the MIR level exceeds that of the equivelent IR level function. As such, MachineInstrs can appear to originate from outside the subprogram scope (and from other subprogram scopes). However, it doesn't seem worth worrying about as the source is imaginary anyway. There's a few high level goals this pass works towards: * We should be able to debugify our .ll/.mir in the lit tests without changing the checks and still pass them. I.e. Debug info should not change codegen. Combining this with a strip-debug pass should enable this. The main issue I ran into without the strip-debug pass was instructions with MMO's and checks on both the instruction and the MMO as the debug-location is between them. I currently have a simple hack in the MIRPrinter to resolve that but the more general solution is a proper strip-debug pass. * We should be able to test that GlobalISel does not lose debug info. I recently found that the legalizer can be unexpectedly lossy in seemingly simple cases (e.g. expanding one instr into many). I have a verifier (will be posted separately) that can be integrated with passes that use the observer interface and will catch location loss (it does not verify correctness, just that there's zero lossage). It is a little conservative as the line-0 locations that arise from conflicts do not track the conflicting locations but it can still catch a fair bit. Depends on D77439, D77438 Reviewers: aprantl, bogner, vsk Subscribers: mgorny, hiraditya, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D77446
2020-04-04 07:18:45 +08:00
//===- MachineDebugify.cpp - Attach synthetic debug info to everything ----===//
//
// 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
//
//===----------------------------------------------------------------------===//
///
/// \file This pass attaches synthetic debug info to everything. It can be used
/// to create targeted tests for debug info preservation, or test for CodeGen
/// differences with vs. without debug info.
Add MIR-level debugify with only locations support for now Summary: Re-used the IR-level debugify for the most part. The MIR-level code then adds locations to the MachineInstrs afterwards based on the LLVM-IR debug info. It's worth mentioning that the resulting locations make little sense as the range of line numbers used in a Function at the MIR level exceeds that of the equivelent IR level function. As such, MachineInstrs can appear to originate from outside the subprogram scope (and from other subprogram scopes). However, it doesn't seem worth worrying about as the source is imaginary anyway. There's a few high level goals this pass works towards: * We should be able to debugify our .ll/.mir in the lit tests without changing the checks and still pass them. I.e. Debug info should not change codegen. Combining this with a strip-debug pass should enable this. The main issue I ran into without the strip-debug pass was instructions with MMO's and checks on both the instruction and the MMO as the debug-location is between them. I currently have a simple hack in the MIRPrinter to resolve that but the more general solution is a proper strip-debug pass. * We should be able to test that GlobalISel does not lose debug info. I recently found that the legalizer can be unexpectedly lossy in seemingly simple cases (e.g. expanding one instr into many). I have a verifier (will be posted separately) that can be integrated with passes that use the observer interface and will catch location loss (it does not verify correctness, just that there's zero lossage). It is a little conservative as the line-0 locations that arise from conflicts do not track the conflicting locations but it can still catch a fair bit. Depends on D77439, D77438 Reviewers: aprantl, bogner, vsk Subscribers: mgorny, hiraditya, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D77446
2020-04-04 07:18:45 +08:00
///
/// This isn't intended to have feature parity with Debugify.
//===----------------------------------------------------------------------===//
#include "llvm/ADT/DenseMap.h"
#include "llvm/ADT/SmallVector.h"
Add MIR-level debugify with only locations support for now Summary: Re-used the IR-level debugify for the most part. The MIR-level code then adds locations to the MachineInstrs afterwards based on the LLVM-IR debug info. It's worth mentioning that the resulting locations make little sense as the range of line numbers used in a Function at the MIR level exceeds that of the equivelent IR level function. As such, MachineInstrs can appear to originate from outside the subprogram scope (and from other subprogram scopes). However, it doesn't seem worth worrying about as the source is imaginary anyway. There's a few high level goals this pass works towards: * We should be able to debugify our .ll/.mir in the lit tests without changing the checks and still pass them. I.e. Debug info should not change codegen. Combining this with a strip-debug pass should enable this. The main issue I ran into without the strip-debug pass was instructions with MMO's and checks on both the instruction and the MMO as the debug-location is between them. I currently have a simple hack in the MIRPrinter to resolve that but the more general solution is a proper strip-debug pass. * We should be able to test that GlobalISel does not lose debug info. I recently found that the legalizer can be unexpectedly lossy in seemingly simple cases (e.g. expanding one instr into many). I have a verifier (will be posted separately) that can be integrated with passes that use the observer interface and will catch location loss (it does not verify correctness, just that there's zero lossage). It is a little conservative as the line-0 locations that arise from conflicts do not track the conflicting locations but it can still catch a fair bit. Depends on D77439, D77438 Reviewers: aprantl, bogner, vsk Subscribers: mgorny, hiraditya, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D77446
2020-04-04 07:18:45 +08:00
#include "llvm/CodeGen/MachineFunctionPass.h"
#include "llvm/CodeGen/MachineInstrBuilder.h"
Add MIR-level debugify with only locations support for now Summary: Re-used the IR-level debugify for the most part. The MIR-level code then adds locations to the MachineInstrs afterwards based on the LLVM-IR debug info. It's worth mentioning that the resulting locations make little sense as the range of line numbers used in a Function at the MIR level exceeds that of the equivelent IR level function. As such, MachineInstrs can appear to originate from outside the subprogram scope (and from other subprogram scopes). However, it doesn't seem worth worrying about as the source is imaginary anyway. There's a few high level goals this pass works towards: * We should be able to debugify our .ll/.mir in the lit tests without changing the checks and still pass them. I.e. Debug info should not change codegen. Combining this with a strip-debug pass should enable this. The main issue I ran into without the strip-debug pass was instructions with MMO's and checks on both the instruction and the MMO as the debug-location is between them. I currently have a simple hack in the MIRPrinter to resolve that but the more general solution is a proper strip-debug pass. * We should be able to test that GlobalISel does not lose debug info. I recently found that the legalizer can be unexpectedly lossy in seemingly simple cases (e.g. expanding one instr into many). I have a verifier (will be posted separately) that can be integrated with passes that use the observer interface and will catch location loss (it does not verify correctness, just that there's zero lossage). It is a little conservative as the line-0 locations that arise from conflicts do not track the conflicting locations but it can still catch a fair bit. Depends on D77439, D77438 Reviewers: aprantl, bogner, vsk Subscribers: mgorny, hiraditya, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D77446
2020-04-04 07:18:45 +08:00
#include "llvm/CodeGen/MachineModuleInfo.h"
#include "llvm/CodeGen/Passes.h"
#include "llvm/CodeGen/TargetInstrInfo.h"
#include "llvm/CodeGen/TargetSubtargetInfo.h"
Add MIR-level debugify with only locations support for now Summary: Re-used the IR-level debugify for the most part. The MIR-level code then adds locations to the MachineInstrs afterwards based on the LLVM-IR debug info. It's worth mentioning that the resulting locations make little sense as the range of line numbers used in a Function at the MIR level exceeds that of the equivelent IR level function. As such, MachineInstrs can appear to originate from outside the subprogram scope (and from other subprogram scopes). However, it doesn't seem worth worrying about as the source is imaginary anyway. There's a few high level goals this pass works towards: * We should be able to debugify our .ll/.mir in the lit tests without changing the checks and still pass them. I.e. Debug info should not change codegen. Combining this with a strip-debug pass should enable this. The main issue I ran into without the strip-debug pass was instructions with MMO's and checks on both the instruction and the MMO as the debug-location is between them. I currently have a simple hack in the MIRPrinter to resolve that but the more general solution is a proper strip-debug pass. * We should be able to test that GlobalISel does not lose debug info. I recently found that the legalizer can be unexpectedly lossy in seemingly simple cases (e.g. expanding one instr into many). I have a verifier (will be posted separately) that can be integrated with passes that use the observer interface and will catch location loss (it does not verify correctness, just that there's zero lossage). It is a little conservative as the line-0 locations that arise from conflicts do not track the conflicting locations but it can still catch a fair bit. Depends on D77439, D77438 Reviewers: aprantl, bogner, vsk Subscribers: mgorny, hiraditya, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D77446
2020-04-04 07:18:45 +08:00
#include "llvm/IR/DIBuilder.h"
#include "llvm/IR/DebugInfo.h"
#include "llvm/IR/IntrinsicInst.h"
Add MIR-level debugify with only locations support for now Summary: Re-used the IR-level debugify for the most part. The MIR-level code then adds locations to the MachineInstrs afterwards based on the LLVM-IR debug info. It's worth mentioning that the resulting locations make little sense as the range of line numbers used in a Function at the MIR level exceeds that of the equivelent IR level function. As such, MachineInstrs can appear to originate from outside the subprogram scope (and from other subprogram scopes). However, it doesn't seem worth worrying about as the source is imaginary anyway. There's a few high level goals this pass works towards: * We should be able to debugify our .ll/.mir in the lit tests without changing the checks and still pass them. I.e. Debug info should not change codegen. Combining this with a strip-debug pass should enable this. The main issue I ran into without the strip-debug pass was instructions with MMO's and checks on both the instruction and the MMO as the debug-location is between them. I currently have a simple hack in the MIRPrinter to resolve that but the more general solution is a proper strip-debug pass. * We should be able to test that GlobalISel does not lose debug info. I recently found that the legalizer can be unexpectedly lossy in seemingly simple cases (e.g. expanding one instr into many). I have a verifier (will be posted separately) that can be integrated with passes that use the observer interface and will catch location loss (it does not verify correctness, just that there's zero lossage). It is a little conservative as the line-0 locations that arise from conflicts do not track the conflicting locations but it can still catch a fair bit. Depends on D77439, D77438 Reviewers: aprantl, bogner, vsk Subscribers: mgorny, hiraditya, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D77446
2020-04-04 07:18:45 +08:00
#include "llvm/InitializePasses.h"
#include "llvm/Transforms/Utils/Debugify.h"
#define DEBUG_TYPE "mir-debugify"
using namespace llvm;
namespace {
bool applyDebugifyMetadataToMachineFunction(MachineModuleInfo &MMI,
DIBuilder &DIB, Function &F) {
MachineFunction *MaybeMF = MMI.getMachineFunction(F);
if (!MaybeMF)
return false;
MachineFunction &MF = *MaybeMF;
const TargetInstrInfo &TII = *MF.getSubtarget().getInstrInfo();
Add MIR-level debugify with only locations support for now Summary: Re-used the IR-level debugify for the most part. The MIR-level code then adds locations to the MachineInstrs afterwards based on the LLVM-IR debug info. It's worth mentioning that the resulting locations make little sense as the range of line numbers used in a Function at the MIR level exceeds that of the equivelent IR level function. As such, MachineInstrs can appear to originate from outside the subprogram scope (and from other subprogram scopes). However, it doesn't seem worth worrying about as the source is imaginary anyway. There's a few high level goals this pass works towards: * We should be able to debugify our .ll/.mir in the lit tests without changing the checks and still pass them. I.e. Debug info should not change codegen. Combining this with a strip-debug pass should enable this. The main issue I ran into without the strip-debug pass was instructions with MMO's and checks on both the instruction and the MMO as the debug-location is between them. I currently have a simple hack in the MIRPrinter to resolve that but the more general solution is a proper strip-debug pass. * We should be able to test that GlobalISel does not lose debug info. I recently found that the legalizer can be unexpectedly lossy in seemingly simple cases (e.g. expanding one instr into many). I have a verifier (will be posted separately) that can be integrated with passes that use the observer interface and will catch location loss (it does not verify correctness, just that there's zero lossage). It is a little conservative as the line-0 locations that arise from conflicts do not track the conflicting locations but it can still catch a fair bit. Depends on D77439, D77438 Reviewers: aprantl, bogner, vsk Subscribers: mgorny, hiraditya, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D77446
2020-04-04 07:18:45 +08:00
DISubprogram *SP = F.getSubprogram();
assert(SP && "IR Debugify just created it?");
Module &M = *F.getParent();
LLVMContext &Ctx = M.getContext();
Add MIR-level debugify with only locations support for now Summary: Re-used the IR-level debugify for the most part. The MIR-level code then adds locations to the MachineInstrs afterwards based on the LLVM-IR debug info. It's worth mentioning that the resulting locations make little sense as the range of line numbers used in a Function at the MIR level exceeds that of the equivelent IR level function. As such, MachineInstrs can appear to originate from outside the subprogram scope (and from other subprogram scopes). However, it doesn't seem worth worrying about as the source is imaginary anyway. There's a few high level goals this pass works towards: * We should be able to debugify our .ll/.mir in the lit tests without changing the checks and still pass them. I.e. Debug info should not change codegen. Combining this with a strip-debug pass should enable this. The main issue I ran into without the strip-debug pass was instructions with MMO's and checks on both the instruction and the MMO as the debug-location is between them. I currently have a simple hack in the MIRPrinter to resolve that but the more general solution is a proper strip-debug pass. * We should be able to test that GlobalISel does not lose debug info. I recently found that the legalizer can be unexpectedly lossy in seemingly simple cases (e.g. expanding one instr into many). I have a verifier (will be posted separately) that can be integrated with passes that use the observer interface and will catch location loss (it does not verify correctness, just that there's zero lossage). It is a little conservative as the line-0 locations that arise from conflicts do not track the conflicting locations but it can still catch a fair bit. Depends on D77439, D77438 Reviewers: aprantl, bogner, vsk Subscribers: mgorny, hiraditya, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D77446
2020-04-04 07:18:45 +08:00
unsigned NextLine = SP->getLine();
Add MIR-level debugify with only locations support for now Summary: Re-used the IR-level debugify for the most part. The MIR-level code then adds locations to the MachineInstrs afterwards based on the LLVM-IR debug info. It's worth mentioning that the resulting locations make little sense as the range of line numbers used in a Function at the MIR level exceeds that of the equivelent IR level function. As such, MachineInstrs can appear to originate from outside the subprogram scope (and from other subprogram scopes). However, it doesn't seem worth worrying about as the source is imaginary anyway. There's a few high level goals this pass works towards: * We should be able to debugify our .ll/.mir in the lit tests without changing the checks and still pass them. I.e. Debug info should not change codegen. Combining this with a strip-debug pass should enable this. The main issue I ran into without the strip-debug pass was instructions with MMO's and checks on both the instruction and the MMO as the debug-location is between them. I currently have a simple hack in the MIRPrinter to resolve that but the more general solution is a proper strip-debug pass. * We should be able to test that GlobalISel does not lose debug info. I recently found that the legalizer can be unexpectedly lossy in seemingly simple cases (e.g. expanding one instr into many). I have a verifier (will be posted separately) that can be integrated with passes that use the observer interface and will catch location loss (it does not verify correctness, just that there's zero lossage). It is a little conservative as the line-0 locations that arise from conflicts do not track the conflicting locations but it can still catch a fair bit. Depends on D77439, D77438 Reviewers: aprantl, bogner, vsk Subscribers: mgorny, hiraditya, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D77446
2020-04-04 07:18:45 +08:00
for (MachineBasicBlock &MBB : MF) {
for (MachineInstr &MI : MBB) {
// This will likely emit line numbers beyond the end of the imagined
// source function and into subsequent ones. We don't do anything about
// that as it doesn't really matter to the compiler where the line is in
// the imaginary source code.
MI.setDebugLoc(DILocation::get(Ctx, NextLine++, 1, SP));
}
}
// Find local variables defined by debugify. No attempt is made to match up
// MIR-level regs to the 'correct' IR-level variables: there isn't a simple
// way to do that, and it isn't necessary to find interesting CodeGen bugs.
// Instead, simply keep track of one variable per line. Later, we can insert
// DBG_VALUE insts that point to these local variables. Emitting DBG_VALUEs
// which cover a wide range of lines can help stress the debug info passes:
// if we can't do that, fall back to using the local variable which precedes
// all the others.
Function *DbgValF = M.getFunction("llvm.dbg.value");
DbgValueInst *EarliestDVI = nullptr;
DenseMap<unsigned, DILocalVariable *> Line2Var;
DIExpression *Expr = nullptr;
if (DbgValF) {
for (const Use &U : DbgValF->uses()) {
auto *DVI = dyn_cast<DbgValueInst>(U.getUser());
if (!DVI || DVI->getFunction() != &F)
continue;
unsigned Line = DVI->getDebugLoc().getLine();
assert(Line != 0 && "debugify should not insert line 0 locations");
Line2Var[Line] = DVI->getVariable();
if (!EarliestDVI || Line < EarliestDVI->getDebugLoc().getLine())
EarliestDVI = DVI;
Expr = DVI->getExpression();
}
}
if (Line2Var.empty())
return true;
// Now, try to insert a DBG_VALUE instruction after each real instruction.
// Do this by introducing debug uses of each register definition. If that is
// not possible (e.g. we have a phi or a meta instruction), emit a constant.
uint64_t NextImm = 0;
const MCInstrDesc &DbgValDesc = TII.get(TargetOpcode::DBG_VALUE);
for (MachineBasicBlock &MBB : MF) {
MachineBasicBlock::iterator FirstNonPHIIt = MBB.getFirstNonPHI();
for (auto I = MBB.begin(), E = MBB.end(); I != E; ) {
MachineInstr &MI = *I;
++I;
// `I` may point to a DBG_VALUE created in the previous loop iteration.
if (MI.isDebugInstr())
continue;
// It's not allowed to insert DBG_VALUEs after a terminator.
if (MI.isTerminator())
continue;
// Find a suitable insertion point for the DBG_VALUE.
auto InsertBeforeIt = MI.isPHI() ? FirstNonPHIIt : I;
// Find a suitable local variable for the DBG_VALUE.
unsigned Line = MI.getDebugLoc().getLine();
if (!Line2Var.count(Line))
Line = EarliestDVI->getDebugLoc().getLine();
DILocalVariable *LocalVar = Line2Var[Line];
assert(LocalVar && "No variable for current line?");
// Emit DBG_VALUEs for register definitions.
SmallVector<MachineOperand *, 4> RegDefs;
for (MachineOperand &MO : MI.operands())
if (MO.isReg() && MO.isDef() && MO.getReg())
RegDefs.push_back(&MO);
for (MachineOperand *MO : RegDefs)
BuildMI(MBB, InsertBeforeIt, MI.getDebugLoc(), DbgValDesc,
/*IsIndirect=*/false, *MO, LocalVar, Expr);
// OK, failing that, emit a constant DBG_VALUE.
if (RegDefs.empty()) {
auto ImmOp = MachineOperand::CreateImm(NextImm++);
BuildMI(MBB, InsertBeforeIt, MI.getDebugLoc(), DbgValDesc,
/*IsIndirect=*/false, ImmOp, LocalVar, Expr);
}
}
}
Add MIR-level debugify with only locations support for now Summary: Re-used the IR-level debugify for the most part. The MIR-level code then adds locations to the MachineInstrs afterwards based on the LLVM-IR debug info. It's worth mentioning that the resulting locations make little sense as the range of line numbers used in a Function at the MIR level exceeds that of the equivelent IR level function. As such, MachineInstrs can appear to originate from outside the subprogram scope (and from other subprogram scopes). However, it doesn't seem worth worrying about as the source is imaginary anyway. There's a few high level goals this pass works towards: * We should be able to debugify our .ll/.mir in the lit tests without changing the checks and still pass them. I.e. Debug info should not change codegen. Combining this with a strip-debug pass should enable this. The main issue I ran into without the strip-debug pass was instructions with MMO's and checks on both the instruction and the MMO as the debug-location is between them. I currently have a simple hack in the MIRPrinter to resolve that but the more general solution is a proper strip-debug pass. * We should be able to test that GlobalISel does not lose debug info. I recently found that the legalizer can be unexpectedly lossy in seemingly simple cases (e.g. expanding one instr into many). I have a verifier (will be posted separately) that can be integrated with passes that use the observer interface and will catch location loss (it does not verify correctness, just that there's zero lossage). It is a little conservative as the line-0 locations that arise from conflicts do not track the conflicting locations but it can still catch a fair bit. Depends on D77439, D77438 Reviewers: aprantl, bogner, vsk Subscribers: mgorny, hiraditya, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D77446
2020-04-04 07:18:45 +08:00
return true;
}
/// ModulePass for attaching synthetic debug info to everything, used with the
/// legacy module pass manager.
struct DebugifyMachineModule : public ModulePass {
bool runOnModule(Module &M) override {
MachineModuleInfo &MMI =
getAnalysis<MachineModuleInfoWrapperPass>().getMMI();
return applyDebugifyMetadata(
M, M.functions(),
"ModuleDebugify: ", [&](DIBuilder &DIB, Function &F) -> bool {
return applyDebugifyMetadataToMachineFunction(MMI, DIB, F);
});
}
DebugifyMachineModule() : ModulePass(ID) {}
void getAnalysisUsage(AnalysisUsage &AU) const override {
AU.addRequired<MachineModuleInfoWrapperPass>();
AU.addPreserved<MachineModuleInfoWrapperPass>();
Add -debugify-and-strip-all to add debug info before a pass and remove it after Summary: This allows us to test each backend pass under the presence of debug info using pre-existing tests. The tests should not fail as a result of this so long as it's true that debug info does not affect CodeGen. In practice, a few tests are sensitive to this: * Tests that check the pass structure (e.g. O0-pipeline.ll) * Tests that check --debug output. Specifically instruction dumps containing MMO's (e.g. prelegalizercombiner-extends.ll) * Tests that contain debugify metadata as mir-strip-debug will remove it (e.g. fastisel-debugvalue-undef.ll) * Tests with partial debug info (e.g. patchable-function-entry-empty.mir had debug info but no !llvm.dbg.cu) * Tests that check optimization remarks overly strictly (e.g. prologue-epilogue-remarks.mir) * Tests that would inject the pass in an unsafe region (e.g. seqpairspill.mir would inject between register alloc and virt reg rewriter) In all cases, the checks can either be updated or --debugify-and-strip-all-safe=0 can be used to avoid being affected by something like llvm-lit -Dllc='llc --debugify-and-strip-all-safe' I tested this without the lost debug locations verifier to confirm that AArch64 behaviour is unaffected (with the fixes in this patch) and with it to confirm it finds the problems without the additional RUN lines we had before. Depends on D77886, D77887, D77747 Reviewers: aprantl, vsk, bogner Subscribers: qcolombet, kristof.beyls, hiraditya, danielkiss, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D77888
2020-04-09 04:48:40 +08:00
AU.setPreservesCFG();
Add MIR-level debugify with only locations support for now Summary: Re-used the IR-level debugify for the most part. The MIR-level code then adds locations to the MachineInstrs afterwards based on the LLVM-IR debug info. It's worth mentioning that the resulting locations make little sense as the range of line numbers used in a Function at the MIR level exceeds that of the equivelent IR level function. As such, MachineInstrs can appear to originate from outside the subprogram scope (and from other subprogram scopes). However, it doesn't seem worth worrying about as the source is imaginary anyway. There's a few high level goals this pass works towards: * We should be able to debugify our .ll/.mir in the lit tests without changing the checks and still pass them. I.e. Debug info should not change codegen. Combining this with a strip-debug pass should enable this. The main issue I ran into without the strip-debug pass was instructions with MMO's and checks on both the instruction and the MMO as the debug-location is between them. I currently have a simple hack in the MIRPrinter to resolve that but the more general solution is a proper strip-debug pass. * We should be able to test that GlobalISel does not lose debug info. I recently found that the legalizer can be unexpectedly lossy in seemingly simple cases (e.g. expanding one instr into many). I have a verifier (will be posted separately) that can be integrated with passes that use the observer interface and will catch location loss (it does not verify correctness, just that there's zero lossage). It is a little conservative as the line-0 locations that arise from conflicts do not track the conflicting locations but it can still catch a fair bit. Depends on D77439, D77438 Reviewers: aprantl, bogner, vsk Subscribers: mgorny, hiraditya, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D77446
2020-04-04 07:18:45 +08:00
}
static char ID; // Pass identification.
};
char DebugifyMachineModule::ID = 0;
} // end anonymous namespace
INITIALIZE_PASS_BEGIN(DebugifyMachineModule, DEBUG_TYPE,
"Machine Debugify Module", false, false)
INITIALIZE_PASS_END(DebugifyMachineModule, DEBUG_TYPE,
"Machine Debugify Module", false, false)
Add -debugify-and-strip-all to add debug info before a pass and remove it after Summary: This allows us to test each backend pass under the presence of debug info using pre-existing tests. The tests should not fail as a result of this so long as it's true that debug info does not affect CodeGen. In practice, a few tests are sensitive to this: * Tests that check the pass structure (e.g. O0-pipeline.ll) * Tests that check --debug output. Specifically instruction dumps containing MMO's (e.g. prelegalizercombiner-extends.ll) * Tests that contain debugify metadata as mir-strip-debug will remove it (e.g. fastisel-debugvalue-undef.ll) * Tests with partial debug info (e.g. patchable-function-entry-empty.mir had debug info but no !llvm.dbg.cu) * Tests that check optimization remarks overly strictly (e.g. prologue-epilogue-remarks.mir) * Tests that would inject the pass in an unsafe region (e.g. seqpairspill.mir would inject between register alloc and virt reg rewriter) In all cases, the checks can either be updated or --debugify-and-strip-all-safe=0 can be used to avoid being affected by something like llvm-lit -Dllc='llc --debugify-and-strip-all-safe' I tested this without the lost debug locations verifier to confirm that AArch64 behaviour is unaffected (with the fixes in this patch) and with it to confirm it finds the problems without the additional RUN lines we had before. Depends on D77886, D77887, D77747 Reviewers: aprantl, vsk, bogner Subscribers: qcolombet, kristof.beyls, hiraditya, danielkiss, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D77888
2020-04-09 04:48:40 +08:00
ModulePass *llvm::createDebugifyMachineModulePass() {
Add MIR-level debugify with only locations support for now Summary: Re-used the IR-level debugify for the most part. The MIR-level code then adds locations to the MachineInstrs afterwards based on the LLVM-IR debug info. It's worth mentioning that the resulting locations make little sense as the range of line numbers used in a Function at the MIR level exceeds that of the equivelent IR level function. As such, MachineInstrs can appear to originate from outside the subprogram scope (and from other subprogram scopes). However, it doesn't seem worth worrying about as the source is imaginary anyway. There's a few high level goals this pass works towards: * We should be able to debugify our .ll/.mir in the lit tests without changing the checks and still pass them. I.e. Debug info should not change codegen. Combining this with a strip-debug pass should enable this. The main issue I ran into without the strip-debug pass was instructions with MMO's and checks on both the instruction and the MMO as the debug-location is between them. I currently have a simple hack in the MIRPrinter to resolve that but the more general solution is a proper strip-debug pass. * We should be able to test that GlobalISel does not lose debug info. I recently found that the legalizer can be unexpectedly lossy in seemingly simple cases (e.g. expanding one instr into many). I have a verifier (will be posted separately) that can be integrated with passes that use the observer interface and will catch location loss (it does not verify correctness, just that there's zero lossage). It is a little conservative as the line-0 locations that arise from conflicts do not track the conflicting locations but it can still catch a fair bit. Depends on D77439, D77438 Reviewers: aprantl, bogner, vsk Subscribers: mgorny, hiraditya, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D77446
2020-04-04 07:18:45 +08:00
return new DebugifyMachineModule();
}