llvm/lib/CodeGen/AsmPrinter/DbgValueHistoryCalculator.cpp

256 lines
10 KiB
C++
Raw Normal View History

//===-- llvm/CodeGen/AsmPrinter/DbgValueHistoryCalculator.cpp -------------===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
#include "DbgValueHistoryCalculator.h"
#include "llvm/ADT/BitVector.h"
#include "llvm/ADT/SmallVector.h"
#include "llvm/CodeGen/MachineBasicBlock.h"
#include "llvm/CodeGen/MachineFunction.h"
#include "llvm/IR/DebugInfo.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/raw_ostream.h"
#include "llvm/Target/TargetLowering.h"
#include "llvm/Target/TargetRegisterInfo.h"
#include "llvm/Target/TargetSubtargetInfo.h"
#include <algorithm>
#include <map>
using namespace llvm;
#define DEBUG_TYPE "dwarfdebug"
// \brief If @MI is a DBG_VALUE with debug value described by a
// defined register, returns the number of this register.
// In the other case, returns 0.
static unsigned isDescribedByReg(const MachineInstr &MI) {
assert(MI.isDebugValue());
Move the complex address expression out of DIVariable and into an extra argument of the llvm.dbg.declare/llvm.dbg.value intrinsics. Previously, DIVariable was a variable-length field that has an optional reference to a Metadata array consisting of a variable number of complex address expressions. In the case of OpPiece expressions this is wasting a lot of storage in IR, because when an aggregate type is, e.g., SROA'd into all of its n individual members, the IR will contain n copies of the DIVariable, all alike, only differing in the complex address reference at the end. By making the complex address into an extra argument of the dbg.value/dbg.declare intrinsics, all of the pieces can reference the same variable and the complex address expressions can be uniqued across the CU, too. Down the road, this will allow us to move other flags, such as "indirection" out of the DIVariable, too. The new intrinsics look like this: declare void @llvm.dbg.declare(metadata %storage, metadata %var, metadata %expr) declare void @llvm.dbg.value(metadata %storage, i64 %offset, metadata %var, metadata %expr) This patch adds a new LLVM-local tag to DIExpressions, so we can detect and pretty-print DIExpression metadata nodes. What this patch doesn't do: This patch does not touch the "Indirect" field in DIVariable; but moving that into the expression would be a natural next step. http://reviews.llvm.org/D4919 rdar://problem/17994491 Thanks to dblaikie and dexonsmith for reviewing this patch! Note: I accidentally committed a bogus older version of this patch previously. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@218787 91177308-0d34-0410-b5e6-96231b3b80d8
2014-10-01 18:55:02 +00:00
assert(MI.getNumOperands() == 4);
// If location of variable is described using a register (directly or
// indirectly), this register is always a first operand.
return MI.getOperand(0).isReg() ? MI.getOperand(0).getReg() : 0;
}
void DbgValueHistoryMap::startInstrRange(InlinedVariable Var,
const MachineInstr &MI) {
// Instruction range should start with a DBG_VALUE instruction for the
// variable.
Move the complex address expression out of DIVariable and into an extra argument of the llvm.dbg.declare/llvm.dbg.value intrinsics. Previously, DIVariable was a variable-length field that has an optional reference to a Metadata array consisting of a variable number of complex address expressions. In the case of OpPiece expressions this is wasting a lot of storage in IR, because when an aggregate type is, e.g., SROA'd into all of its n individual members, the IR will contain n copies of the DIVariable, all alike, only differing in the complex address reference at the end. By making the complex address into an extra argument of the dbg.value/dbg.declare intrinsics, all of the pieces can reference the same variable and the complex address expressions can be uniqued across the CU, too. Down the road, this will allow us to move other flags, such as "indirection" out of the DIVariable, too. The new intrinsics look like this: declare void @llvm.dbg.declare(metadata %storage, metadata %var, metadata %expr) declare void @llvm.dbg.value(metadata %storage, i64 %offset, metadata %var, metadata %expr) This patch adds a new LLVM-local tag to DIExpressions, so we can detect and pretty-print DIExpression metadata nodes. What this patch doesn't do: This patch does not touch the "Indirect" field in DIVariable; but moving that into the expression would be a natural next step. http://reviews.llvm.org/D4919 rdar://problem/17994491 Thanks to dblaikie and dexonsmith for reviewing this patch! Note: I accidentally committed a bogus older version of this patch previously. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@218787 91177308-0d34-0410-b5e6-96231b3b80d8
2014-10-01 18:55:02 +00:00
assert(MI.isDebugValue() && "not a DBG_VALUE");
auto &Ranges = VarInstrRanges[Var];
if (!Ranges.empty() && Ranges.back().second == nullptr &&
Ranges.back().first->isIdenticalTo(MI)) {
DEBUG(dbgs() << "Coalescing identical DBG_VALUE entries:\n"
<< "\t" << Ranges.back().first << "\t" << MI << "\n");
return;
}
Ranges.push_back(std::make_pair(&MI, nullptr));
}
void DbgValueHistoryMap::endInstrRange(InlinedVariable Var,
const MachineInstr &MI) {
auto &Ranges = VarInstrRanges[Var];
// Verify that the current instruction range is not yet closed.
assert(!Ranges.empty() && Ranges.back().second == nullptr);
// For now, instruction ranges are not allowed to cross basic block
// boundaries.
assert(Ranges.back().first->getParent() == MI.getParent());
Ranges.back().second = &MI;
}
unsigned DbgValueHistoryMap::getRegisterForVar(InlinedVariable Var) const {
const auto &I = VarInstrRanges.find(Var);
if (I == VarInstrRanges.end())
return 0;
const auto &Ranges = I->second;
if (Ranges.empty() || Ranges.back().second != nullptr)
return 0;
return isDescribedByReg(*Ranges.back().first);
}
namespace {
// Maps physreg numbers to the variables they describe.
typedef DbgValueHistoryMap::InlinedVariable InlinedVariable;
typedef std::map<unsigned, SmallVector<InlinedVariable, 1>> RegDescribedVarsMap;
}
// \brief Claim that @Var is not described by @RegNo anymore.
static void dropRegDescribedVar(RegDescribedVarsMap &RegVars, unsigned RegNo,
InlinedVariable Var) {
const auto &I = RegVars.find(RegNo);
assert(RegNo != 0U && I != RegVars.end());
auto &VarSet = I->second;
const auto &VarPos = find(VarSet, Var);
assert(VarPos != VarSet.end());
VarSet.erase(VarPos);
// Don't keep empty sets in a map to keep it as small as possible.
if (VarSet.empty())
RegVars.erase(I);
}
// \brief Claim that @Var is now described by @RegNo.
static void addRegDescribedVar(RegDescribedVarsMap &RegVars, unsigned RegNo,
InlinedVariable Var) {
assert(RegNo != 0U);
auto &VarSet = RegVars[RegNo];
assert(!is_contained(VarSet, Var));
VarSet.push_back(Var);
}
// \brief Terminate the location range for variables described by register at
// @I by inserting @ClobberingInstr to their history.
static void clobberRegisterUses(RegDescribedVarsMap &RegVars,
RegDescribedVarsMap::iterator I,
DbgValueHistoryMap &HistMap,
const MachineInstr &ClobberingInstr) {
// Iterate over all variables described by this register and add this
// instruction to their history, clobbering it.
for (const auto &Var : I->second)
HistMap.endInstrRange(Var, ClobberingInstr);
RegVars.erase(I);
}
// \brief Terminate the location range for variables described by register
// @RegNo by inserting @ClobberingInstr to their history.
static void clobberRegisterUses(RegDescribedVarsMap &RegVars, unsigned RegNo,
DbgValueHistoryMap &HistMap,
const MachineInstr &ClobberingInstr) {
const auto &I = RegVars.find(RegNo);
if (I == RegVars.end())
return;
clobberRegisterUses(RegVars, I, HistMap, ClobberingInstr);
}
Generate better location ranges for some register-described variables. Don't terminate location ranges for register-described variables at the end of machine basic block if this register is never modified in the function body, except for the prologue and epilogue. Prologue location is guessed by FrameSetup flags on MachineInstructions, while epilogue location is deduced from debug locations of instructions in the basic blocks ending with return instructions. This patch is mostly targeted to fix non-trivial debug locations for variables addressed via stack and frame pointers. It is not really a generic fix. We can still produce poor debug info for register-described variables if this register *is* modified somewhere in the function, but in unrelated places. This might be the case for the debug info in optimized binaries (e.g. for local variables in inlined functions). LiveDebugVariables pass in CodeGen attempts to fix this problem by adjusting DBG_VALUE instructions, but this pass is tied to greedy register allocator, which is used in optimized builds only. Proper fix would likely involve generalizing LiveDebugVariables to all register allocators. See more discussion in http://reviews.llvm.org/D3933 review thread. I'm proceeding with this patch to fix immediate severe problems and important cases, e.g. fix completely broken debug info with AddressSanitizer and fix PR19307 (missing debug info for by-value std::string arguments). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@210492 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-09 21:53:47 +00:00
// \brief Returns the first instruction in @MBB which corresponds to
// the function epilogue, or nullptr if @MBB doesn't contain an epilogue.
static const MachineInstr *getFirstEpilogueInst(const MachineBasicBlock &MBB) {
auto LastMI = MBB.getLastNonDebugInstr();
if (LastMI == MBB.end() || !LastMI->isReturn())
return nullptr;
// Assume that epilogue starts with instruction having the same debug location
// as the return instruction.
DebugLoc LastLoc = LastMI->getDebugLoc();
auto Res = LastMI;
CodeGen: Give MachineBasicBlock::reverse_iterator a handle to the current MI Now that MachineBasicBlock::reverse_instr_iterator knows when it's at the end (since r281168 and r281170), implement MachineBasicBlock::reverse_iterator directly on top of an ilist::reverse_iterator by adding an IsReverse template parameter to MachineInstrBundleIterator. This replaces another hard-to-reason-about use of std::reverse_iterator on list iterators, matching the changes for ilist::reverse_iterator from r280032 (see the "out of scope" section at the end of that commit message). MachineBasicBlock::reverse_iterator now has a handle to the current node and has obvious invalidation semantics. r280032 has a more detailed explanation of how list-style reverse iterators (invalidated when the pointed-at node is deleted) are different from vector-style reverse iterators like std::reverse_iterator (invalidated on every operation). A great motivating example is this commit's changes to lib/CodeGen/DeadMachineInstructionElim.cpp. Note: If your out-of-tree backend deletes instructions while iterating on a MachineBasicBlock::reverse_iterator or converts between MachineBasicBlock::iterator and MachineBasicBlock::reverse_iterator, you'll need to update your code in similar ways to r280032. The following table might help: [Old] ==> [New] delete &*RI, RE = end() delete &*RI++ RI->erase(), RE = end() RI++->erase() reverse_iterator(I) std::prev(I).getReverse() reverse_iterator(I) ++I.getReverse() --reverse_iterator(I) I.getReverse() reverse_iterator(std::next(I)) I.getReverse() RI.base() std::prev(RI).getReverse() RI.base() ++RI.getReverse() --RI.base() RI.getReverse() std::next(RI).base() RI.getReverse() (For more details, have a look at r280032.) git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@281172 91177308-0d34-0410-b5e6-96231b3b80d8
2016-09-11 18:51:28 +00:00
for (MachineBasicBlock::const_reverse_iterator I = LastMI.getReverse(),
E = MBB.rend();
I != E; ++I) {
Generate better location ranges for some register-described variables. Don't terminate location ranges for register-described variables at the end of machine basic block if this register is never modified in the function body, except for the prologue and epilogue. Prologue location is guessed by FrameSetup flags on MachineInstructions, while epilogue location is deduced from debug locations of instructions in the basic blocks ending with return instructions. This patch is mostly targeted to fix non-trivial debug locations for variables addressed via stack and frame pointers. It is not really a generic fix. We can still produce poor debug info for register-described variables if this register *is* modified somewhere in the function, but in unrelated places. This might be the case for the debug info in optimized binaries (e.g. for local variables in inlined functions). LiveDebugVariables pass in CodeGen attempts to fix this problem by adjusting DBG_VALUE instructions, but this pass is tied to greedy register allocator, which is used in optimized builds only. Proper fix would likely involve generalizing LiveDebugVariables to all register allocators. See more discussion in http://reviews.llvm.org/D3933 review thread. I'm proceeding with this patch to fix immediate severe problems and important cases, e.g. fix completely broken debug info with AddressSanitizer and fix PR19307 (missing debug info for by-value std::string arguments). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@210492 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-09 21:53:47 +00:00
if (I->getDebugLoc() != LastLoc)
return &*Res;
Res = &*I;
Generate better location ranges for some register-described variables. Don't terminate location ranges for register-described variables at the end of machine basic block if this register is never modified in the function body, except for the prologue and epilogue. Prologue location is guessed by FrameSetup flags on MachineInstructions, while epilogue location is deduced from debug locations of instructions in the basic blocks ending with return instructions. This patch is mostly targeted to fix non-trivial debug locations for variables addressed via stack and frame pointers. It is not really a generic fix. We can still produce poor debug info for register-described variables if this register *is* modified somewhere in the function, but in unrelated places. This might be the case for the debug info in optimized binaries (e.g. for local variables in inlined functions). LiveDebugVariables pass in CodeGen attempts to fix this problem by adjusting DBG_VALUE instructions, but this pass is tied to greedy register allocator, which is used in optimized builds only. Proper fix would likely involve generalizing LiveDebugVariables to all register allocators. See more discussion in http://reviews.llvm.org/D3933 review thread. I'm proceeding with this patch to fix immediate severe problems and important cases, e.g. fix completely broken debug info with AddressSanitizer and fix PR19307 (missing debug info for by-value std::string arguments). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@210492 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-09 21:53:47 +00:00
}
// If all instructions have the same debug location, assume whole MBB is
// an epilogue.
return &*MBB.begin();
Generate better location ranges for some register-described variables. Don't terminate location ranges for register-described variables at the end of machine basic block if this register is never modified in the function body, except for the prologue and epilogue. Prologue location is guessed by FrameSetup flags on MachineInstructions, while epilogue location is deduced from debug locations of instructions in the basic blocks ending with return instructions. This patch is mostly targeted to fix non-trivial debug locations for variables addressed via stack and frame pointers. It is not really a generic fix. We can still produce poor debug info for register-described variables if this register *is* modified somewhere in the function, but in unrelated places. This might be the case for the debug info in optimized binaries (e.g. for local variables in inlined functions). LiveDebugVariables pass in CodeGen attempts to fix this problem by adjusting DBG_VALUE instructions, but this pass is tied to greedy register allocator, which is used in optimized builds only. Proper fix would likely involve generalizing LiveDebugVariables to all register allocators. See more discussion in http://reviews.llvm.org/D3933 review thread. I'm proceeding with this patch to fix immediate severe problems and important cases, e.g. fix completely broken debug info with AddressSanitizer and fix PR19307 (missing debug info for by-value std::string arguments). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@210492 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-09 21:53:47 +00:00
}
// \brief Collect registers that are modified in the function body (their
// contents is changed outside of the prologue and epilogue).
Generate better location ranges for some register-described variables. Don't terminate location ranges for register-described variables at the end of machine basic block if this register is never modified in the function body, except for the prologue and epilogue. Prologue location is guessed by FrameSetup flags on MachineInstructions, while epilogue location is deduced from debug locations of instructions in the basic blocks ending with return instructions. This patch is mostly targeted to fix non-trivial debug locations for variables addressed via stack and frame pointers. It is not really a generic fix. We can still produce poor debug info for register-described variables if this register *is* modified somewhere in the function, but in unrelated places. This might be the case for the debug info in optimized binaries (e.g. for local variables in inlined functions). LiveDebugVariables pass in CodeGen attempts to fix this problem by adjusting DBG_VALUE instructions, but this pass is tied to greedy register allocator, which is used in optimized builds only. Proper fix would likely involve generalizing LiveDebugVariables to all register allocators. See more discussion in http://reviews.llvm.org/D3933 review thread. I'm proceeding with this patch to fix immediate severe problems and important cases, e.g. fix completely broken debug info with AddressSanitizer and fix PR19307 (missing debug info for by-value std::string arguments). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@210492 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-09 21:53:47 +00:00
static void collectChangingRegs(const MachineFunction *MF,
const TargetRegisterInfo *TRI,
BitVector &Regs) {
Generate better location ranges for some register-described variables. Don't terminate location ranges for register-described variables at the end of machine basic block if this register is never modified in the function body, except for the prologue and epilogue. Prologue location is guessed by FrameSetup flags on MachineInstructions, while epilogue location is deduced from debug locations of instructions in the basic blocks ending with return instructions. This patch is mostly targeted to fix non-trivial debug locations for variables addressed via stack and frame pointers. It is not really a generic fix. We can still produce poor debug info for register-described variables if this register *is* modified somewhere in the function, but in unrelated places. This might be the case for the debug info in optimized binaries (e.g. for local variables in inlined functions). LiveDebugVariables pass in CodeGen attempts to fix this problem by adjusting DBG_VALUE instructions, but this pass is tied to greedy register allocator, which is used in optimized builds only. Proper fix would likely involve generalizing LiveDebugVariables to all register allocators. See more discussion in http://reviews.llvm.org/D3933 review thread. I'm proceeding with this patch to fix immediate severe problems and important cases, e.g. fix completely broken debug info with AddressSanitizer and fix PR19307 (missing debug info for by-value std::string arguments). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@210492 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-09 21:53:47 +00:00
for (const auto &MBB : *MF) {
auto FirstEpilogueInst = getFirstEpilogueInst(MBB);
Generate better location ranges for some register-described variables. Don't terminate location ranges for register-described variables at the end of machine basic block if this register is never modified in the function body, except for the prologue and epilogue. Prologue location is guessed by FrameSetup flags on MachineInstructions, while epilogue location is deduced from debug locations of instructions in the basic blocks ending with return instructions. This patch is mostly targeted to fix non-trivial debug locations for variables addressed via stack and frame pointers. It is not really a generic fix. We can still produce poor debug info for register-described variables if this register *is* modified somewhere in the function, but in unrelated places. This might be the case for the debug info in optimized binaries (e.g. for local variables in inlined functions). LiveDebugVariables pass in CodeGen attempts to fix this problem by adjusting DBG_VALUE instructions, but this pass is tied to greedy register allocator, which is used in optimized builds only. Proper fix would likely involve generalizing LiveDebugVariables to all register allocators. See more discussion in http://reviews.llvm.org/D3933 review thread. I'm proceeding with this patch to fix immediate severe problems and important cases, e.g. fix completely broken debug info with AddressSanitizer and fix PR19307 (missing debug info for by-value std::string arguments). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@210492 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-09 21:53:47 +00:00
for (const auto &MI : MBB) {
// Avoid looking at prologue or epilogue instructions.
if (&MI == FirstEpilogueInst)
break;
if (MI.getFlag(MachineInstr::FrameSetup))
continue;
// Look for register defs and register masks. Register masks are
// typically on calls and they clobber everything not in the mask.
for (const MachineOperand &MO : MI.operands()) {
// Skip virtual registers since they are handled by the parent.
if (MO.isReg() && MO.isDef() && MO.getReg() &&
!TRI->isVirtualRegister(MO.getReg())) {
for (MCRegAliasIterator AI(MO.getReg(), TRI, true); AI.isValid();
++AI)
Regs.set(*AI);
} else if (MO.isRegMask()) {
Regs.setBitsNotInMask(MO.getRegMask());
}
}
Generate better location ranges for some register-described variables. Don't terminate location ranges for register-described variables at the end of machine basic block if this register is never modified in the function body, except for the prologue and epilogue. Prologue location is guessed by FrameSetup flags on MachineInstructions, while epilogue location is deduced from debug locations of instructions in the basic blocks ending with return instructions. This patch is mostly targeted to fix non-trivial debug locations for variables addressed via stack and frame pointers. It is not really a generic fix. We can still produce poor debug info for register-described variables if this register *is* modified somewhere in the function, but in unrelated places. This might be the case for the debug info in optimized binaries (e.g. for local variables in inlined functions). LiveDebugVariables pass in CodeGen attempts to fix this problem by adjusting DBG_VALUE instructions, but this pass is tied to greedy register allocator, which is used in optimized builds only. Proper fix would likely involve generalizing LiveDebugVariables to all register allocators. See more discussion in http://reviews.llvm.org/D3933 review thread. I'm proceeding with this patch to fix immediate severe problems and important cases, e.g. fix completely broken debug info with AddressSanitizer and fix PR19307 (missing debug info for by-value std::string arguments). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@210492 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-09 21:53:47 +00:00
}
}
}
void llvm::calculateDbgValueHistory(const MachineFunction *MF,
const TargetRegisterInfo *TRI,
DbgValueHistoryMap &Result) {
BitVector ChangingRegs(TRI->getNumRegs());
Generate better location ranges for some register-described variables. Don't terminate location ranges for register-described variables at the end of machine basic block if this register is never modified in the function body, except for the prologue and epilogue. Prologue location is guessed by FrameSetup flags on MachineInstructions, while epilogue location is deduced from debug locations of instructions in the basic blocks ending with return instructions. This patch is mostly targeted to fix non-trivial debug locations for variables addressed via stack and frame pointers. It is not really a generic fix. We can still produce poor debug info for register-described variables if this register *is* modified somewhere in the function, but in unrelated places. This might be the case for the debug info in optimized binaries (e.g. for local variables in inlined functions). LiveDebugVariables pass in CodeGen attempts to fix this problem by adjusting DBG_VALUE instructions, but this pass is tied to greedy register allocator, which is used in optimized builds only. Proper fix would likely involve generalizing LiveDebugVariables to all register allocators. See more discussion in http://reviews.llvm.org/D3933 review thread. I'm proceeding with this patch to fix immediate severe problems and important cases, e.g. fix completely broken debug info with AddressSanitizer and fix PR19307 (missing debug info for by-value std::string arguments). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@210492 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-09 21:53:47 +00:00
collectChangingRegs(MF, TRI, ChangingRegs);
const TargetLowering *TLI = MF->getSubtarget().getTargetLowering();
unsigned SP = TLI->getStackPointerRegisterToSaveRestore();
Generate better location ranges for some register-described variables. Don't terminate location ranges for register-described variables at the end of machine basic block if this register is never modified in the function body, except for the prologue and epilogue. Prologue location is guessed by FrameSetup flags on MachineInstructions, while epilogue location is deduced from debug locations of instructions in the basic blocks ending with return instructions. This patch is mostly targeted to fix non-trivial debug locations for variables addressed via stack and frame pointers. It is not really a generic fix. We can still produce poor debug info for register-described variables if this register *is* modified somewhere in the function, but in unrelated places. This might be the case for the debug info in optimized binaries (e.g. for local variables in inlined functions). LiveDebugVariables pass in CodeGen attempts to fix this problem by adjusting DBG_VALUE instructions, but this pass is tied to greedy register allocator, which is used in optimized builds only. Proper fix would likely involve generalizing LiveDebugVariables to all register allocators. See more discussion in http://reviews.llvm.org/D3933 review thread. I'm proceeding with this patch to fix immediate severe problems and important cases, e.g. fix completely broken debug info with AddressSanitizer and fix PR19307 (missing debug info for by-value std::string arguments). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@210492 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-09 21:53:47 +00:00
RegDescribedVarsMap RegVars;
for (const auto &MBB : *MF) {
for (const auto &MI : MBB) {
if (!MI.isDebugValue()) {
// Not a DBG_VALUE instruction. It may clobber registers which describe
// some variables.
for (const MachineOperand &MO : MI.operands()) {
if (MO.isReg() && MO.isDef() && MO.getReg()) {
// If this is a virtual register, only clobber it since it doesn't
// have aliases.
if (TRI->isVirtualRegister(MO.getReg()))
clobberRegisterUses(RegVars, MO.getReg(), Result, MI);
// If this is a register def operand, it may end a debug value
// range.
else {
for (MCRegAliasIterator AI(MO.getReg(), TRI, true); AI.isValid();
++AI)
if (ChangingRegs.test(*AI))
clobberRegisterUses(RegVars, *AI, Result, MI);
}
} else if (MO.isRegMask()) {
// If this is a register mask operand, clobber all debug values in
// non-CSRs.
for (int I = ChangingRegs.find_first(); I != -1;
I = ChangingRegs.find_next(I)) {
// Don't consider SP to be clobbered by register masks.
if (unsigned(I) != SP && TRI->isPhysicalRegister(I) &&
MO.clobbersPhysReg(I)) {
clobberRegisterUses(RegVars, I, Result, MI);
}
}
}
}
continue;
}
assert(MI.getNumOperands() > 1 && "Invalid DBG_VALUE instruction!");
// Use the base variable (without any DW_OP_piece expressions)
// as index into History. The full variables including the
// piece expressions are attached to the MI.
const DILocalVariable *RawVar = MI.getDebugVariable();
assert(RawVar->isValidLocationForIntrinsic(MI.getDebugLoc()) &&
"Expected inlined-at fields to agree");
InlinedVariable Var(RawVar, MI.getDebugLoc()->getInlinedAt());
if (unsigned PrevReg = Result.getRegisterForVar(Var))
dropRegDescribedVar(RegVars, PrevReg, Var);
Result.startInstrRange(Var, MI);
if (unsigned NewReg = isDescribedByReg(MI))
addRegDescribedVar(RegVars, NewReg, Var);
}
// Make sure locations for register-described variables are valid only
// until the end of the basic block (unless it's the last basic block, in
// which case let their liveness run off to the end of the function).
if (!MBB.empty() && &MBB != &MF->back()) {
for (auto I = RegVars.begin(), E = RegVars.end(); I != E;) {
auto CurElem = I++; // CurElem can be erased below.
if (TRI->isVirtualRegister(CurElem->first) ||
ChangingRegs.test(CurElem->first))
clobberRegisterUses(RegVars, CurElem, Result, MBB.back());
}
Generate better location ranges for some register-described variables. Don't terminate location ranges for register-described variables at the end of machine basic block if this register is never modified in the function body, except for the prologue and epilogue. Prologue location is guessed by FrameSetup flags on MachineInstructions, while epilogue location is deduced from debug locations of instructions in the basic blocks ending with return instructions. This patch is mostly targeted to fix non-trivial debug locations for variables addressed via stack and frame pointers. It is not really a generic fix. We can still produce poor debug info for register-described variables if this register *is* modified somewhere in the function, but in unrelated places. This might be the case for the debug info in optimized binaries (e.g. for local variables in inlined functions). LiveDebugVariables pass in CodeGen attempts to fix this problem by adjusting DBG_VALUE instructions, but this pass is tied to greedy register allocator, which is used in optimized builds only. Proper fix would likely involve generalizing LiveDebugVariables to all register allocators. See more discussion in http://reviews.llvm.org/D3933 review thread. I'm proceeding with this patch to fix immediate severe problems and important cases, e.g. fix completely broken debug info with AddressSanitizer and fix PR19307 (missing debug info for by-value std::string arguments). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@210492 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-09 21:53:47 +00:00
}
}
}