llvm-project/llvm/tools/dsymutil/DwarfLinker.cpp

3151 lines
121 KiB
C++
Raw Normal View History

//===- tools/dsymutil/DwarfLinker.cpp - Dwarf debug info linker -----------===//
//
// The LLVM Linker
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
#include "DebugMap.h"
#include "BinaryHolder.h"
#include "DebugMap.h"
#include "dsymutil.h"
#include "llvm/ADT/IntervalMap.h"
#include "llvm/ADT/StringMap.h"
#include "llvm/ADT/STLExtras.h"
#include "llvm/CodeGen/AsmPrinter.h"
#include "llvm/CodeGen/DIE.h"
#include "llvm/Config/config.h"
#include "llvm/DebugInfo/DWARF/DWARFContext.h"
#include "llvm/DebugInfo/DWARF/DWARFDebugInfoEntry.h"
#include "llvm/DebugInfo/DWARF/DWARFFormValue.h"
#include "llvm/MC/MCAsmBackend.h"
#include "llvm/MC/MCAsmInfo.h"
#include "llvm/MC/MCContext.h"
#include "llvm/MC/MCCodeEmitter.h"
#include "llvm/MC/MCDwarf.h"
#include "llvm/MC/MCInstrInfo.h"
#include "llvm/MC/MCObjectFileInfo.h"
#include "llvm/MC/MCRegisterInfo.h"
#include "llvm/MC/MCStreamer.h"
#include "llvm/MC/MCSubtargetInfo.h"
#include "llvm/Object/MachO.h"
#include "llvm/Support/Dwarf.h"
#include "llvm/Support/LEB128.h"
#include "llvm/Support/TargetRegistry.h"
#include "llvm/Target/TargetMachine.h"
#include "llvm/Target/TargetOptions.h"
#include <string>
#include <tuple>
namespace llvm {
namespace dsymutil {
namespace {
void warn(const Twine &Warning, const Twine &Context) {
errs() << Twine("while processing ") + Context + ":\n";
errs() << Twine("warning: ") + Warning + "\n";
}
bool error(const Twine &Error, const Twine &Context) {
errs() << Twine("while processing ") + Context + ":\n";
errs() << Twine("error: ") + Error + "\n";
return false;
}
template <typename KeyT, typename ValT>
using HalfOpenIntervalMap =
IntervalMap<KeyT, ValT, IntervalMapImpl::NodeSizer<KeyT, ValT>::LeafSize,
IntervalMapHalfOpenInfo<KeyT>>;
typedef HalfOpenIntervalMap<uint64_t, int64_t> FunctionIntervals;
// FIXME: Delete this structure.
Reapply "AsmPrinter: Change DIEValue to be stored by value" This reverts commit r238350, effectively reapplying r238349 after fixing (all?) the problems, all somehow related to how I was using `AlignedArrayCharUnion<>` inside `DIEValue`: - MSVC can only handle `sizeof()` on types, not values. Change the assert. - GCC doesn't know the `is_trivially_copyable` type trait. Instead of asserting it, add destructors. - Call placement new even when constructing POD (i.e., the pointers). - Instead of copying the char buffer, copy the casted classes. I've left in a couple of `static_assert`s that I think both MSVC and GCC know how to handle. If the bots disagree with me, I'll remove them. - Check that the constructed type is either standard layout or a pointer. This protects against a programming error: we really want the "small" `DIEValue`s to be small and simple, so don't accidentally change them not to be. - Similarly, check that the size of the buffer is no bigger than a `uint64_t` or a pointer. (I thought checking against `sizeof(uint64_t)` would be good enough, but Chandler suggested that pointers might sometimes be bigger than that in the context of sanitizers.) I've also committed r238359 in the meantime, which introduces a DIEValue.def to simplify dispatching between the various types (thanks to a review comment by David Blaikie). Without that, this commit would be almost unintelligible. Here's the original commit message: -- Change `DIEValue` to be stored/passed/etc. by value, instead of reference. It's now a discriminated union, with a `Val` field storing the actual type. The classes that used to inherit from `DIEValue` no longer do. There are two categories of these: - Small values fit in a single pointer and are stored by value. - Large values require auxiliary storage, and are stored by reference. The only non-mechanical change is to tools/dsymutil/DwarfLinker.cpp. It was relying on `DIEInteger`s being passed around by reference, so I replaced that assumption with a `PatchLocation` type that stores a safe reference to where the `DIEInteger` lives instead. This commit causes a temporary regression in memory usage, since I've left merging `DIEAbbrevData` into `DIEValue` for a follow-up commit. I measured an increase from 845 MB to 879 MB, around 3.9%. The follow-up drops it lower than the starting point, and I've only recently brought the memory this low anyway, so I'm committing these changes separately to keep them incremental. (I also considered swapping the commits, but the other one first would cause a lot more code churn.) (I'm looking at `llc` memory usage on `verify-uselistorder.lto.opt.bc`; see r236629 for details.) -- llvm-svn: 238362
2015-05-28 06:14:58 +08:00
struct PatchLocation {
DIE::value_iterator I;
Reapply "AsmPrinter: Change DIEValue to be stored by value" This reverts commit r238350, effectively reapplying r238349 after fixing (all?) the problems, all somehow related to how I was using `AlignedArrayCharUnion<>` inside `DIEValue`: - MSVC can only handle `sizeof()` on types, not values. Change the assert. - GCC doesn't know the `is_trivially_copyable` type trait. Instead of asserting it, add destructors. - Call placement new even when constructing POD (i.e., the pointers). - Instead of copying the char buffer, copy the casted classes. I've left in a couple of `static_assert`s that I think both MSVC and GCC know how to handle. If the bots disagree with me, I'll remove them. - Check that the constructed type is either standard layout or a pointer. This protects against a programming error: we really want the "small" `DIEValue`s to be small and simple, so don't accidentally change them not to be. - Similarly, check that the size of the buffer is no bigger than a `uint64_t` or a pointer. (I thought checking against `sizeof(uint64_t)` would be good enough, but Chandler suggested that pointers might sometimes be bigger than that in the context of sanitizers.) I've also committed r238359 in the meantime, which introduces a DIEValue.def to simplify dispatching between the various types (thanks to a review comment by David Blaikie). Without that, this commit would be almost unintelligible. Here's the original commit message: -- Change `DIEValue` to be stored/passed/etc. by value, instead of reference. It's now a discriminated union, with a `Val` field storing the actual type. The classes that used to inherit from `DIEValue` no longer do. There are two categories of these: - Small values fit in a single pointer and are stored by value. - Large values require auxiliary storage, and are stored by reference. The only non-mechanical change is to tools/dsymutil/DwarfLinker.cpp. It was relying on `DIEInteger`s being passed around by reference, so I replaced that assumption with a `PatchLocation` type that stores a safe reference to where the `DIEInteger` lives instead. This commit causes a temporary regression in memory usage, since I've left merging `DIEAbbrevData` into `DIEValue` for a follow-up commit. I measured an increase from 845 MB to 879 MB, around 3.9%. The follow-up drops it lower than the starting point, and I've only recently brought the memory this low anyway, so I'm committing these changes separately to keep them incremental. (I also considered swapping the commits, but the other one first would cause a lot more code churn.) (I'm looking at `llc` memory usage on `verify-uselistorder.lto.opt.bc`; see r236629 for details.) -- llvm-svn: 238362
2015-05-28 06:14:58 +08:00
PatchLocation() = default;
PatchLocation(DIE::value_iterator I) : I(I) {}
Reapply "AsmPrinter: Change DIEValue to be stored by value" This reverts commit r238350, effectively reapplying r238349 after fixing (all?) the problems, all somehow related to how I was using `AlignedArrayCharUnion<>` inside `DIEValue`: - MSVC can only handle `sizeof()` on types, not values. Change the assert. - GCC doesn't know the `is_trivially_copyable` type trait. Instead of asserting it, add destructors. - Call placement new even when constructing POD (i.e., the pointers). - Instead of copying the char buffer, copy the casted classes. I've left in a couple of `static_assert`s that I think both MSVC and GCC know how to handle. If the bots disagree with me, I'll remove them. - Check that the constructed type is either standard layout or a pointer. This protects against a programming error: we really want the "small" `DIEValue`s to be small and simple, so don't accidentally change them not to be. - Similarly, check that the size of the buffer is no bigger than a `uint64_t` or a pointer. (I thought checking against `sizeof(uint64_t)` would be good enough, but Chandler suggested that pointers might sometimes be bigger than that in the context of sanitizers.) I've also committed r238359 in the meantime, which introduces a DIEValue.def to simplify dispatching between the various types (thanks to a review comment by David Blaikie). Without that, this commit would be almost unintelligible. Here's the original commit message: -- Change `DIEValue` to be stored/passed/etc. by value, instead of reference. It's now a discriminated union, with a `Val` field storing the actual type. The classes that used to inherit from `DIEValue` no longer do. There are two categories of these: - Small values fit in a single pointer and are stored by value. - Large values require auxiliary storage, and are stored by reference. The only non-mechanical change is to tools/dsymutil/DwarfLinker.cpp. It was relying on `DIEInteger`s being passed around by reference, so I replaced that assumption with a `PatchLocation` type that stores a safe reference to where the `DIEInteger` lives instead. This commit causes a temporary regression in memory usage, since I've left merging `DIEAbbrevData` into `DIEValue` for a follow-up commit. I measured an increase from 845 MB to 879 MB, around 3.9%. The follow-up drops it lower than the starting point, and I've only recently brought the memory this low anyway, so I'm committing these changes separately to keep them incremental. (I also considered swapping the commits, but the other one first would cause a lot more code churn.) (I'm looking at `llc` memory usage on `verify-uselistorder.lto.opt.bc`; see r236629 for details.) -- llvm-svn: 238362
2015-05-28 06:14:58 +08:00
void set(uint64_t New) const {
assert(I);
const auto &Old = *I;
assert(Old.getType() == DIEValue::isInteger);
*I = DIEValue(Old.getAttribute(), Old.getForm(), DIEInteger(New));
Reapply "AsmPrinter: Change DIEValue to be stored by value" This reverts commit r238350, effectively reapplying r238349 after fixing (all?) the problems, all somehow related to how I was using `AlignedArrayCharUnion<>` inside `DIEValue`: - MSVC can only handle `sizeof()` on types, not values. Change the assert. - GCC doesn't know the `is_trivially_copyable` type trait. Instead of asserting it, add destructors. - Call placement new even when constructing POD (i.e., the pointers). - Instead of copying the char buffer, copy the casted classes. I've left in a couple of `static_assert`s that I think both MSVC and GCC know how to handle. If the bots disagree with me, I'll remove them. - Check that the constructed type is either standard layout or a pointer. This protects against a programming error: we really want the "small" `DIEValue`s to be small and simple, so don't accidentally change them not to be. - Similarly, check that the size of the buffer is no bigger than a `uint64_t` or a pointer. (I thought checking against `sizeof(uint64_t)` would be good enough, but Chandler suggested that pointers might sometimes be bigger than that in the context of sanitizers.) I've also committed r238359 in the meantime, which introduces a DIEValue.def to simplify dispatching between the various types (thanks to a review comment by David Blaikie). Without that, this commit would be almost unintelligible. Here's the original commit message: -- Change `DIEValue` to be stored/passed/etc. by value, instead of reference. It's now a discriminated union, with a `Val` field storing the actual type. The classes that used to inherit from `DIEValue` no longer do. There are two categories of these: - Small values fit in a single pointer and are stored by value. - Large values require auxiliary storage, and are stored by reference. The only non-mechanical change is to tools/dsymutil/DwarfLinker.cpp. It was relying on `DIEInteger`s being passed around by reference, so I replaced that assumption with a `PatchLocation` type that stores a safe reference to where the `DIEInteger` lives instead. This commit causes a temporary regression in memory usage, since I've left merging `DIEAbbrevData` into `DIEValue` for a follow-up commit. I measured an increase from 845 MB to 879 MB, around 3.9%. The follow-up drops it lower than the starting point, and I've only recently brought the memory this low anyway, so I'm committing these changes separately to keep them incremental. (I also considered swapping the commits, but the other one first would cause a lot more code churn.) (I'm looking at `llc` memory usage on `verify-uselistorder.lto.opt.bc`; see r236629 for details.) -- llvm-svn: 238362
2015-05-28 06:14:58 +08:00
}
uint64_t get() const {
assert(I);
return I->getDIEInteger().getValue();
Reapply "AsmPrinter: Change DIEValue to be stored by value" This reverts commit r238350, effectively reapplying r238349 after fixing (all?) the problems, all somehow related to how I was using `AlignedArrayCharUnion<>` inside `DIEValue`: - MSVC can only handle `sizeof()` on types, not values. Change the assert. - GCC doesn't know the `is_trivially_copyable` type trait. Instead of asserting it, add destructors. - Call placement new even when constructing POD (i.e., the pointers). - Instead of copying the char buffer, copy the casted classes. I've left in a couple of `static_assert`s that I think both MSVC and GCC know how to handle. If the bots disagree with me, I'll remove them. - Check that the constructed type is either standard layout or a pointer. This protects against a programming error: we really want the "small" `DIEValue`s to be small and simple, so don't accidentally change them not to be. - Similarly, check that the size of the buffer is no bigger than a `uint64_t` or a pointer. (I thought checking against `sizeof(uint64_t)` would be good enough, but Chandler suggested that pointers might sometimes be bigger than that in the context of sanitizers.) I've also committed r238359 in the meantime, which introduces a DIEValue.def to simplify dispatching between the various types (thanks to a review comment by David Blaikie). Without that, this commit would be almost unintelligible. Here's the original commit message: -- Change `DIEValue` to be stored/passed/etc. by value, instead of reference. It's now a discriminated union, with a `Val` field storing the actual type. The classes that used to inherit from `DIEValue` no longer do. There are two categories of these: - Small values fit in a single pointer and are stored by value. - Large values require auxiliary storage, and are stored by reference. The only non-mechanical change is to tools/dsymutil/DwarfLinker.cpp. It was relying on `DIEInteger`s being passed around by reference, so I replaced that assumption with a `PatchLocation` type that stores a safe reference to where the `DIEInteger` lives instead. This commit causes a temporary regression in memory usage, since I've left merging `DIEAbbrevData` into `DIEValue` for a follow-up commit. I measured an increase from 845 MB to 879 MB, around 3.9%. The follow-up drops it lower than the starting point, and I've only recently brought the memory this low anyway, so I'm committing these changes separately to keep them incremental. (I also considered swapping the commits, but the other one first would cause a lot more code churn.) (I'm looking at `llc` memory usage on `verify-uselistorder.lto.opt.bc`; see r236629 for details.) -- llvm-svn: 238362
2015-05-28 06:14:58 +08:00
}
};
class CompileUnit;
struct DeclMapInfo;
class NonRelocatableStringpool;
/// A DeclContext is a named program scope that is used for ODR
/// uniquing of types.
/// The set of DeclContext for the ODR-subject parts of a Dwarf link
/// is expanded (and uniqued) with each new object file processed. We
/// need to determine the context of each DIE in an linked object file
/// to see if the corresponding type has already been emitted.
///
/// The contexts are conceptually organised as a tree (eg. a function
/// scope is contained in a namespace scope that contains other
/// scopes), but storing/accessing them in an actual tree is too
/// inefficient: we need to be able to very quickly query a context
/// for a given child context by name. Storing a StringMap in each
/// DeclContext would be too space inefficient.
/// The solution here is to give each DeclContext a link to its parent
/// (this allows to walk up the tree), but to query the existance of a
/// specific DeclContext using a separate DenseMap keyed on the hash
/// of the fully qualified name of the context.
class DeclContext {
unsigned QualifiedNameHash;
uint32_t Line;
uint32_t ByteSize;
uint16_t Tag;
StringRef Name;
StringRef File;
const DeclContext &Parent;
const DWARFDebugInfoEntryMinimal *LastSeenDIE;
uint32_t LastSeenCompileUnitID;
uint32_t CanonicalDIEOffset;
friend DeclMapInfo;
public:
typedef DenseSet<DeclContext *, DeclMapInfo> Map;
DeclContext()
: QualifiedNameHash(0), Line(0), ByteSize(0),
Tag(dwarf::DW_TAG_compile_unit), Name(), File(), Parent(*this),
LastSeenDIE(nullptr), LastSeenCompileUnitID(0), CanonicalDIEOffset(0) {}
DeclContext(unsigned Hash, uint32_t Line, uint32_t ByteSize, uint16_t Tag,
StringRef Name, StringRef File, const DeclContext &Parent,
const DWARFDebugInfoEntryMinimal *LastSeenDIE = nullptr,
unsigned CUId = 0)
: QualifiedNameHash(Hash), Line(Line), ByteSize(ByteSize), Tag(Tag),
Name(Name), File(File), Parent(Parent), LastSeenDIE(LastSeenDIE),
LastSeenCompileUnitID(CUId), CanonicalDIEOffset(0) {}
uint32_t getQualifiedNameHash() const { return QualifiedNameHash; }
bool setLastSeenDIE(CompileUnit &U, const DWARFDebugInfoEntryMinimal *Die);
uint32_t getCanonicalDIEOffset() const { return CanonicalDIEOffset; }
void setCanonicalDIEOffset(uint32_t Offset) { CanonicalDIEOffset = Offset; }
uint16_t getTag() const { return Tag; }
StringRef getName() const { return Name; }
};
/// Info type for the DenseMap storing the DeclContext pointers.
struct DeclMapInfo : private DenseMapInfo<DeclContext *> {
using DenseMapInfo<DeclContext *>::getEmptyKey;
using DenseMapInfo<DeclContext *>::getTombstoneKey;
static unsigned getHashValue(const DeclContext *Ctxt) {
return Ctxt->QualifiedNameHash;
}
static bool isEqual(const DeclContext *LHS, const DeclContext *RHS) {
if (RHS == getEmptyKey() || RHS == getTombstoneKey())
return RHS == LHS;
return LHS->QualifiedNameHash == RHS->QualifiedNameHash &&
LHS->Line == RHS->Line && LHS->ByteSize == RHS->ByteSize &&
LHS->Name.data() == RHS->Name.data() &&
LHS->File.data() == RHS->File.data() &&
LHS->Parent.QualifiedNameHash == RHS->Parent.QualifiedNameHash;
}
};
/// This class gives a tree-like API to the DenseMap that stores the
/// DeclContext objects. It also holds the BumpPtrAllocator where
/// these objects will be allocated.
class DeclContextTree {
BumpPtrAllocator Allocator;
DeclContext Root;
DeclContext::Map Contexts;
public:
/// Get the child of \a Context described by \a DIE in \a Unit. The
/// required strings will be interned in \a StringPool.
/// \returns The child DeclContext along with one bit that is set if
/// this context is invalid.
/// FIXME: the invalid bit along the return value is to emulate some
/// dsymutil-classic functionality. See the fucntion definition for
/// a more thorough discussion of its use.
PointerIntPair<DeclContext *, 1>
getChildDeclContext(DeclContext &Context,
const DWARFDebugInfoEntryMinimal *DIE, CompileUnit &Unit,
NonRelocatableStringpool &StringPool);
DeclContext &getRoot() { return Root; }
};
/// \brief Stores all information relating to a compile unit, be it in
/// its original instance in the object file to its brand new cloned
/// and linked DIE tree.
class CompileUnit {
public:
/// \brief Information gathered about a DIE in the object file.
struct DIEInfo {
int64_t AddrAdjust; ///< Address offset to apply to the described entity.
DeclContext *Ctxt; ///< ODR Declaration context.
DIE *Clone; ///< Cloned version of that DIE.
uint32_t ParentIdx; ///< The index of this DIE's parent.
bool Keep; ///< Is the DIE part of the linked output?
bool InDebugMap; ///< Was this DIE's entity found in the map?
};
CompileUnit(DWARFUnit &OrigUnit, unsigned ID, bool CanUseODR)
: OrigUnit(OrigUnit), ID(ID), LowPc(UINT64_MAX), HighPc(0), RangeAlloc(),
Reapply "AsmPrinter: Change DIEValue to be stored by value" This reverts commit r238350, effectively reapplying r238349 after fixing (all?) the problems, all somehow related to how I was using `AlignedArrayCharUnion<>` inside `DIEValue`: - MSVC can only handle `sizeof()` on types, not values. Change the assert. - GCC doesn't know the `is_trivially_copyable` type trait. Instead of asserting it, add destructors. - Call placement new even when constructing POD (i.e., the pointers). - Instead of copying the char buffer, copy the casted classes. I've left in a couple of `static_assert`s that I think both MSVC and GCC know how to handle. If the bots disagree with me, I'll remove them. - Check that the constructed type is either standard layout or a pointer. This protects against a programming error: we really want the "small" `DIEValue`s to be small and simple, so don't accidentally change them not to be. - Similarly, check that the size of the buffer is no bigger than a `uint64_t` or a pointer. (I thought checking against `sizeof(uint64_t)` would be good enough, but Chandler suggested that pointers might sometimes be bigger than that in the context of sanitizers.) I've also committed r238359 in the meantime, which introduces a DIEValue.def to simplify dispatching between the various types (thanks to a review comment by David Blaikie). Without that, this commit would be almost unintelligible. Here's the original commit message: -- Change `DIEValue` to be stored/passed/etc. by value, instead of reference. It's now a discriminated union, with a `Val` field storing the actual type. The classes that used to inherit from `DIEValue` no longer do. There are two categories of these: - Small values fit in a single pointer and are stored by value. - Large values require auxiliary storage, and are stored by reference. The only non-mechanical change is to tools/dsymutil/DwarfLinker.cpp. It was relying on `DIEInteger`s being passed around by reference, so I replaced that assumption with a `PatchLocation` type that stores a safe reference to where the `DIEInteger` lives instead. This commit causes a temporary regression in memory usage, since I've left merging `DIEAbbrevData` into `DIEValue` for a follow-up commit. I measured an increase from 845 MB to 879 MB, around 3.9%. The follow-up drops it lower than the starting point, and I've only recently brought the memory this low anyway, so I'm committing these changes separately to keep them incremental. (I also considered swapping the commits, but the other one first would cause a lot more code churn.) (I'm looking at `llc` memory usage on `verify-uselistorder.lto.opt.bc`; see r236629 for details.) -- llvm-svn: 238362
2015-05-28 06:14:58 +08:00
Ranges(RangeAlloc) {
Info.resize(OrigUnit.getNumDIEs());
const auto *CUDie = OrigUnit.getUnitDIE(false);
unsigned Lang = CUDie->getAttributeValueAsUnsignedConstant(
&OrigUnit, dwarf::DW_AT_language, 0);
HasODR = CanUseODR && (Lang == dwarf::DW_LANG_C_plus_plus ||
Lang == dwarf::DW_LANG_C_plus_plus_03 ||
Lang == dwarf::DW_LANG_C_plus_plus_11 ||
Lang == dwarf::DW_LANG_C_plus_plus_14 ||
Lang == dwarf::DW_LANG_ObjC_plus_plus);
}
CompileUnit(CompileUnit &&RHS)
: OrigUnit(RHS.OrigUnit), Info(std::move(RHS.Info)),
CUDie(std::move(RHS.CUDie)), StartOffset(RHS.StartOffset),
NextUnitOffset(RHS.NextUnitOffset), RangeAlloc(), Ranges(RangeAlloc) {
// The CompileUnit container has been 'reserve()'d with the right
// size. We cannot move the IntervalMap anyway.
llvm_unreachable("CompileUnits should not be moved.");
}
DWARFUnit &getOrigUnit() const { return OrigUnit; }
unsigned getUniqueID() const { return ID; }
DIE *getOutputUnitDIE() const { return CUDie; }
void setOutputUnitDIE(DIE *Die) { CUDie = Die; }
bool hasODR() const { return HasODR; }
DIEInfo &getInfo(unsigned Idx) { return Info[Idx]; }
const DIEInfo &getInfo(unsigned Idx) const { return Info[Idx]; }
uint64_t getStartOffset() const { return StartOffset; }
uint64_t getNextUnitOffset() const { return NextUnitOffset; }
void setStartOffset(uint64_t DebugInfoSize) { StartOffset = DebugInfoSize; }
uint64_t getLowPc() const { return LowPc; }
uint64_t getHighPc() const { return HighPc; }
Reapply "AsmPrinter: Change DIEValue to be stored by value" This reverts commit r238350, effectively reapplying r238349 after fixing (all?) the problems, all somehow related to how I was using `AlignedArrayCharUnion<>` inside `DIEValue`: - MSVC can only handle `sizeof()` on types, not values. Change the assert. - GCC doesn't know the `is_trivially_copyable` type trait. Instead of asserting it, add destructors. - Call placement new even when constructing POD (i.e., the pointers). - Instead of copying the char buffer, copy the casted classes. I've left in a couple of `static_assert`s that I think both MSVC and GCC know how to handle. If the bots disagree with me, I'll remove them. - Check that the constructed type is either standard layout or a pointer. This protects against a programming error: we really want the "small" `DIEValue`s to be small and simple, so don't accidentally change them not to be. - Similarly, check that the size of the buffer is no bigger than a `uint64_t` or a pointer. (I thought checking against `sizeof(uint64_t)` would be good enough, but Chandler suggested that pointers might sometimes be bigger than that in the context of sanitizers.) I've also committed r238359 in the meantime, which introduces a DIEValue.def to simplify dispatching between the various types (thanks to a review comment by David Blaikie). Without that, this commit would be almost unintelligible. Here's the original commit message: -- Change `DIEValue` to be stored/passed/etc. by value, instead of reference. It's now a discriminated union, with a `Val` field storing the actual type. The classes that used to inherit from `DIEValue` no longer do. There are two categories of these: - Small values fit in a single pointer and are stored by value. - Large values require auxiliary storage, and are stored by reference. The only non-mechanical change is to tools/dsymutil/DwarfLinker.cpp. It was relying on `DIEInteger`s being passed around by reference, so I replaced that assumption with a `PatchLocation` type that stores a safe reference to where the `DIEInteger` lives instead. This commit causes a temporary regression in memory usage, since I've left merging `DIEAbbrevData` into `DIEValue` for a follow-up commit. I measured an increase from 845 MB to 879 MB, around 3.9%. The follow-up drops it lower than the starting point, and I've only recently brought the memory this low anyway, so I'm committing these changes separately to keep them incremental. (I also considered swapping the commits, but the other one first would cause a lot more code churn.) (I'm looking at `llc` memory usage on `verify-uselistorder.lto.opt.bc`; see r236629 for details.) -- llvm-svn: 238362
2015-05-28 06:14:58 +08:00
Optional<PatchLocation> getUnitRangesAttribute() const {
return UnitRangeAttribute;
}
const FunctionIntervals &getFunctionRanges() const { return Ranges; }
Reapply "AsmPrinter: Change DIEValue to be stored by value" This reverts commit r238350, effectively reapplying r238349 after fixing (all?) the problems, all somehow related to how I was using `AlignedArrayCharUnion<>` inside `DIEValue`: - MSVC can only handle `sizeof()` on types, not values. Change the assert. - GCC doesn't know the `is_trivially_copyable` type trait. Instead of asserting it, add destructors. - Call placement new even when constructing POD (i.e., the pointers). - Instead of copying the char buffer, copy the casted classes. I've left in a couple of `static_assert`s that I think both MSVC and GCC know how to handle. If the bots disagree with me, I'll remove them. - Check that the constructed type is either standard layout or a pointer. This protects against a programming error: we really want the "small" `DIEValue`s to be small and simple, so don't accidentally change them not to be. - Similarly, check that the size of the buffer is no bigger than a `uint64_t` or a pointer. (I thought checking against `sizeof(uint64_t)` would be good enough, but Chandler suggested that pointers might sometimes be bigger than that in the context of sanitizers.) I've also committed r238359 in the meantime, which introduces a DIEValue.def to simplify dispatching between the various types (thanks to a review comment by David Blaikie). Without that, this commit would be almost unintelligible. Here's the original commit message: -- Change `DIEValue` to be stored/passed/etc. by value, instead of reference. It's now a discriminated union, with a `Val` field storing the actual type. The classes that used to inherit from `DIEValue` no longer do. There are two categories of these: - Small values fit in a single pointer and are stored by value. - Large values require auxiliary storage, and are stored by reference. The only non-mechanical change is to tools/dsymutil/DwarfLinker.cpp. It was relying on `DIEInteger`s being passed around by reference, so I replaced that assumption with a `PatchLocation` type that stores a safe reference to where the `DIEInteger` lives instead. This commit causes a temporary regression in memory usage, since I've left merging `DIEAbbrevData` into `DIEValue` for a follow-up commit. I measured an increase from 845 MB to 879 MB, around 3.9%. The follow-up drops it lower than the starting point, and I've only recently brought the memory this low anyway, so I'm committing these changes separately to keep them incremental. (I also considered swapping the commits, but the other one first would cause a lot more code churn.) (I'm looking at `llc` memory usage on `verify-uselistorder.lto.opt.bc`; see r236629 for details.) -- llvm-svn: 238362
2015-05-28 06:14:58 +08:00
const std::vector<PatchLocation> &getRangesAttributes() const {
return RangeAttributes;
}
Reapply "AsmPrinter: Change DIEValue to be stored by value" This reverts commit r238350, effectively reapplying r238349 after fixing (all?) the problems, all somehow related to how I was using `AlignedArrayCharUnion<>` inside `DIEValue`: - MSVC can only handle `sizeof()` on types, not values. Change the assert. - GCC doesn't know the `is_trivially_copyable` type trait. Instead of asserting it, add destructors. - Call placement new even when constructing POD (i.e., the pointers). - Instead of copying the char buffer, copy the casted classes. I've left in a couple of `static_assert`s that I think both MSVC and GCC know how to handle. If the bots disagree with me, I'll remove them. - Check that the constructed type is either standard layout or a pointer. This protects against a programming error: we really want the "small" `DIEValue`s to be small and simple, so don't accidentally change them not to be. - Similarly, check that the size of the buffer is no bigger than a `uint64_t` or a pointer. (I thought checking against `sizeof(uint64_t)` would be good enough, but Chandler suggested that pointers might sometimes be bigger than that in the context of sanitizers.) I've also committed r238359 in the meantime, which introduces a DIEValue.def to simplify dispatching between the various types (thanks to a review comment by David Blaikie). Without that, this commit would be almost unintelligible. Here's the original commit message: -- Change `DIEValue` to be stored/passed/etc. by value, instead of reference. It's now a discriminated union, with a `Val` field storing the actual type. The classes that used to inherit from `DIEValue` no longer do. There are two categories of these: - Small values fit in a single pointer and are stored by value. - Large values require auxiliary storage, and are stored by reference. The only non-mechanical change is to tools/dsymutil/DwarfLinker.cpp. It was relying on `DIEInteger`s being passed around by reference, so I replaced that assumption with a `PatchLocation` type that stores a safe reference to where the `DIEInteger` lives instead. This commit causes a temporary regression in memory usage, since I've left merging `DIEAbbrevData` into `DIEValue` for a follow-up commit. I measured an increase from 845 MB to 879 MB, around 3.9%. The follow-up drops it lower than the starting point, and I've only recently brought the memory this low anyway, so I'm committing these changes separately to keep them incremental. (I also considered swapping the commits, but the other one first would cause a lot more code churn.) (I'm looking at `llc` memory usage on `verify-uselistorder.lto.opt.bc`; see r236629 for details.) -- llvm-svn: 238362
2015-05-28 06:14:58 +08:00
const std::vector<std::pair<PatchLocation, int64_t>> &
getLocationAttributes() const {
return LocationAttributes;
}
/// \brief Compute the end offset for this unit. Must be
/// called after the CU's DIEs have been cloned.
/// \returns the next unit offset (which is also the current
/// debug_info section size).
uint64_t computeNextUnitOffset();
/// \brief Keep track of a forward reference to DIE \p Die in \p
/// RefUnit by \p Attr. The attribute should be fixed up later to
/// point to the absolute offset of \p Die in the debug_info section
/// or to the canonical offset of \p Ctxt if it is non-null.
void noteForwardReference(DIE *Die, const CompileUnit *RefUnit,
DeclContext *Ctxt, PatchLocation Attr);
/// \brief Apply all fixups recored by noteForwardReference().
void fixupForwardReferences();
/// \brief Add a function range [\p LowPC, \p HighPC) that is
/// relocatad by applying offset \p PCOffset.
void addFunctionRange(uint64_t LowPC, uint64_t HighPC, int64_t PCOffset);
/// \brief Keep track of a DW_AT_range attribute that we will need to
/// patch up later.
Reapply "AsmPrinter: Change DIEValue to be stored by value" This reverts commit r238350, effectively reapplying r238349 after fixing (all?) the problems, all somehow related to how I was using `AlignedArrayCharUnion<>` inside `DIEValue`: - MSVC can only handle `sizeof()` on types, not values. Change the assert. - GCC doesn't know the `is_trivially_copyable` type trait. Instead of asserting it, add destructors. - Call placement new even when constructing POD (i.e., the pointers). - Instead of copying the char buffer, copy the casted classes. I've left in a couple of `static_assert`s that I think both MSVC and GCC know how to handle. If the bots disagree with me, I'll remove them. - Check that the constructed type is either standard layout or a pointer. This protects against a programming error: we really want the "small" `DIEValue`s to be small and simple, so don't accidentally change them not to be. - Similarly, check that the size of the buffer is no bigger than a `uint64_t` or a pointer. (I thought checking against `sizeof(uint64_t)` would be good enough, but Chandler suggested that pointers might sometimes be bigger than that in the context of sanitizers.) I've also committed r238359 in the meantime, which introduces a DIEValue.def to simplify dispatching between the various types (thanks to a review comment by David Blaikie). Without that, this commit would be almost unintelligible. Here's the original commit message: -- Change `DIEValue` to be stored/passed/etc. by value, instead of reference. It's now a discriminated union, with a `Val` field storing the actual type. The classes that used to inherit from `DIEValue` no longer do. There are two categories of these: - Small values fit in a single pointer and are stored by value. - Large values require auxiliary storage, and are stored by reference. The only non-mechanical change is to tools/dsymutil/DwarfLinker.cpp. It was relying on `DIEInteger`s being passed around by reference, so I replaced that assumption with a `PatchLocation` type that stores a safe reference to where the `DIEInteger` lives instead. This commit causes a temporary regression in memory usage, since I've left merging `DIEAbbrevData` into `DIEValue` for a follow-up commit. I measured an increase from 845 MB to 879 MB, around 3.9%. The follow-up drops it lower than the starting point, and I've only recently brought the memory this low anyway, so I'm committing these changes separately to keep them incremental. (I also considered swapping the commits, but the other one first would cause a lot more code churn.) (I'm looking at `llc` memory usage on `verify-uselistorder.lto.opt.bc`; see r236629 for details.) -- llvm-svn: 238362
2015-05-28 06:14:58 +08:00
void noteRangeAttribute(const DIE &Die, PatchLocation Attr);
/// \brief Keep track of a location attribute pointing to a location
/// list in the debug_loc section.
Reapply "AsmPrinter: Change DIEValue to be stored by value" This reverts commit r238350, effectively reapplying r238349 after fixing (all?) the problems, all somehow related to how I was using `AlignedArrayCharUnion<>` inside `DIEValue`: - MSVC can only handle `sizeof()` on types, not values. Change the assert. - GCC doesn't know the `is_trivially_copyable` type trait. Instead of asserting it, add destructors. - Call placement new even when constructing POD (i.e., the pointers). - Instead of copying the char buffer, copy the casted classes. I've left in a couple of `static_assert`s that I think both MSVC and GCC know how to handle. If the bots disagree with me, I'll remove them. - Check that the constructed type is either standard layout or a pointer. This protects against a programming error: we really want the "small" `DIEValue`s to be small and simple, so don't accidentally change them not to be. - Similarly, check that the size of the buffer is no bigger than a `uint64_t` or a pointer. (I thought checking against `sizeof(uint64_t)` would be good enough, but Chandler suggested that pointers might sometimes be bigger than that in the context of sanitizers.) I've also committed r238359 in the meantime, which introduces a DIEValue.def to simplify dispatching between the various types (thanks to a review comment by David Blaikie). Without that, this commit would be almost unintelligible. Here's the original commit message: -- Change `DIEValue` to be stored/passed/etc. by value, instead of reference. It's now a discriminated union, with a `Val` field storing the actual type. The classes that used to inherit from `DIEValue` no longer do. There are two categories of these: - Small values fit in a single pointer and are stored by value. - Large values require auxiliary storage, and are stored by reference. The only non-mechanical change is to tools/dsymutil/DwarfLinker.cpp. It was relying on `DIEInteger`s being passed around by reference, so I replaced that assumption with a `PatchLocation` type that stores a safe reference to where the `DIEInteger` lives instead. This commit causes a temporary regression in memory usage, since I've left merging `DIEAbbrevData` into `DIEValue` for a follow-up commit. I measured an increase from 845 MB to 879 MB, around 3.9%. The follow-up drops it lower than the starting point, and I've only recently brought the memory this low anyway, so I'm committing these changes separately to keep them incremental. (I also considered swapping the commits, but the other one first would cause a lot more code churn.) (I'm looking at `llc` memory usage on `verify-uselistorder.lto.opt.bc`; see r236629 for details.) -- llvm-svn: 238362
2015-05-28 06:14:58 +08:00
void noteLocationAttribute(PatchLocation Attr, int64_t PcOffset);
/// \brief Add a name accelerator entry for \p Die with \p Name
/// which is stored in the string table at \p Offset.
void addNameAccelerator(const DIE *Die, const char *Name, uint32_t Offset,
bool SkipPubnamesSection = false);
/// \brief Add a type accelerator entry for \p Die with \p Name
/// which is stored in the string table at \p Offset.
void addTypeAccelerator(const DIE *Die, const char *Name, uint32_t Offset);
struct AccelInfo {
StringRef Name; ///< Name of the entry.
const DIE *Die; ///< DIE this entry describes.
uint32_t NameOffset; ///< Offset of Name in the string pool.
bool SkipPubSection; ///< Emit this entry only in the apple_* sections.
AccelInfo(StringRef Name, const DIE *Die, uint32_t NameOffset,
bool SkipPubSection = false)
: Name(Name), Die(Die), NameOffset(NameOffset),
SkipPubSection(SkipPubSection) {}
};
const std::vector<AccelInfo> &getPubnames() const { return Pubnames; }
const std::vector<AccelInfo> &getPubtypes() const { return Pubtypes; }
/// Get the full path for file \a FileNum in the line table
const char *getResolvedPath(unsigned FileNum) {
if (FileNum >= ResolvedPaths.size())
return nullptr;
return ResolvedPaths[FileNum].size() ? ResolvedPaths[FileNum].c_str()
: nullptr;
}
/// Set the fully resolved path for the line-table's file \a FileNum
/// to \a Path.
void setResolvedPath(unsigned FileNum, const std::string &Path) {
if (ResolvedPaths.size() <= FileNum)
ResolvedPaths.resize(FileNum + 1);
ResolvedPaths[FileNum] = Path;
}
private:
DWARFUnit &OrigUnit;
unsigned ID;
std::vector<DIEInfo> Info; ///< DIE info indexed by DIE index.
DIE *CUDie; ///< Root of the linked DIE tree.
uint64_t StartOffset;
uint64_t NextUnitOffset;
uint64_t LowPc;
uint64_t HighPc;
/// \brief A list of attributes to fixup with the absolute offset of
/// a DIE in the debug_info section.
///
/// The offsets for the attributes in this array couldn't be set while
/// cloning because for cross-cu forward refences the target DIE's
/// offset isn't known you emit the reference attribute.
std::vector<std::tuple<DIE *, const CompileUnit *, DeclContext *,
PatchLocation>> ForwardDIEReferences;
FunctionIntervals::Allocator RangeAlloc;
/// \brief The ranges in that interval map are the PC ranges for
/// functions in this unit, associated with the PC offset to apply
/// to the addresses to get the linked address.
FunctionIntervals Ranges;
/// \brief DW_AT_ranges attributes to patch after we have gathered
/// all the unit's function addresses.
/// @{
Reapply "AsmPrinter: Change DIEValue to be stored by value" This reverts commit r238350, effectively reapplying r238349 after fixing (all?) the problems, all somehow related to how I was using `AlignedArrayCharUnion<>` inside `DIEValue`: - MSVC can only handle `sizeof()` on types, not values. Change the assert. - GCC doesn't know the `is_trivially_copyable` type trait. Instead of asserting it, add destructors. - Call placement new even when constructing POD (i.e., the pointers). - Instead of copying the char buffer, copy the casted classes. I've left in a couple of `static_assert`s that I think both MSVC and GCC know how to handle. If the bots disagree with me, I'll remove them. - Check that the constructed type is either standard layout or a pointer. This protects against a programming error: we really want the "small" `DIEValue`s to be small and simple, so don't accidentally change them not to be. - Similarly, check that the size of the buffer is no bigger than a `uint64_t` or a pointer. (I thought checking against `sizeof(uint64_t)` would be good enough, but Chandler suggested that pointers might sometimes be bigger than that in the context of sanitizers.) I've also committed r238359 in the meantime, which introduces a DIEValue.def to simplify dispatching between the various types (thanks to a review comment by David Blaikie). Without that, this commit would be almost unintelligible. Here's the original commit message: -- Change `DIEValue` to be stored/passed/etc. by value, instead of reference. It's now a discriminated union, with a `Val` field storing the actual type. The classes that used to inherit from `DIEValue` no longer do. There are two categories of these: - Small values fit in a single pointer and are stored by value. - Large values require auxiliary storage, and are stored by reference. The only non-mechanical change is to tools/dsymutil/DwarfLinker.cpp. It was relying on `DIEInteger`s being passed around by reference, so I replaced that assumption with a `PatchLocation` type that stores a safe reference to where the `DIEInteger` lives instead. This commit causes a temporary regression in memory usage, since I've left merging `DIEAbbrevData` into `DIEValue` for a follow-up commit. I measured an increase from 845 MB to 879 MB, around 3.9%. The follow-up drops it lower than the starting point, and I've only recently brought the memory this low anyway, so I'm committing these changes separately to keep them incremental. (I also considered swapping the commits, but the other one first would cause a lot more code churn.) (I'm looking at `llc` memory usage on `verify-uselistorder.lto.opt.bc`; see r236629 for details.) -- llvm-svn: 238362
2015-05-28 06:14:58 +08:00
std::vector<PatchLocation> RangeAttributes;
Optional<PatchLocation> UnitRangeAttribute;
/// @}
/// \brief Location attributes that need to be transfered from th
/// original debug_loc section to the liked one. They are stored
/// along with the PC offset that is to be applied to their
/// function's address.
Reapply "AsmPrinter: Change DIEValue to be stored by value" This reverts commit r238350, effectively reapplying r238349 after fixing (all?) the problems, all somehow related to how I was using `AlignedArrayCharUnion<>` inside `DIEValue`: - MSVC can only handle `sizeof()` on types, not values. Change the assert. - GCC doesn't know the `is_trivially_copyable` type trait. Instead of asserting it, add destructors. - Call placement new even when constructing POD (i.e., the pointers). - Instead of copying the char buffer, copy the casted classes. I've left in a couple of `static_assert`s that I think both MSVC and GCC know how to handle. If the bots disagree with me, I'll remove them. - Check that the constructed type is either standard layout or a pointer. This protects against a programming error: we really want the "small" `DIEValue`s to be small and simple, so don't accidentally change them not to be. - Similarly, check that the size of the buffer is no bigger than a `uint64_t` or a pointer. (I thought checking against `sizeof(uint64_t)` would be good enough, but Chandler suggested that pointers might sometimes be bigger than that in the context of sanitizers.) I've also committed r238359 in the meantime, which introduces a DIEValue.def to simplify dispatching between the various types (thanks to a review comment by David Blaikie). Without that, this commit would be almost unintelligible. Here's the original commit message: -- Change `DIEValue` to be stored/passed/etc. by value, instead of reference. It's now a discriminated union, with a `Val` field storing the actual type. The classes that used to inherit from `DIEValue` no longer do. There are two categories of these: - Small values fit in a single pointer and are stored by value. - Large values require auxiliary storage, and are stored by reference. The only non-mechanical change is to tools/dsymutil/DwarfLinker.cpp. It was relying on `DIEInteger`s being passed around by reference, so I replaced that assumption with a `PatchLocation` type that stores a safe reference to where the `DIEInteger` lives instead. This commit causes a temporary regression in memory usage, since I've left merging `DIEAbbrevData` into `DIEValue` for a follow-up commit. I measured an increase from 845 MB to 879 MB, around 3.9%. The follow-up drops it lower than the starting point, and I've only recently brought the memory this low anyway, so I'm committing these changes separately to keep them incremental. (I also considered swapping the commits, but the other one first would cause a lot more code churn.) (I'm looking at `llc` memory usage on `verify-uselistorder.lto.opt.bc`; see r236629 for details.) -- llvm-svn: 238362
2015-05-28 06:14:58 +08:00
std::vector<std::pair<PatchLocation, int64_t>> LocationAttributes;
/// \brief Accelerator entries for the unit, both for the pub*
/// sections and the apple* ones.
/// @{
std::vector<AccelInfo> Pubnames;
std::vector<AccelInfo> Pubtypes;
/// @}
/// Cached resolved paths from the line table.
std::vector<std::string> ResolvedPaths;
/// Is this unit subject to the ODR rule?
bool HasODR;
};
uint64_t CompileUnit::computeNextUnitOffset() {
NextUnitOffset = StartOffset + 11 /* Header size */;
// The root DIE might be null, meaning that the Unit had nothing to
// contribute to the linked output. In that case, we will emit the
// unit header without any actual DIE.
if (CUDie)
NextUnitOffset += CUDie->getSize();
return NextUnitOffset;
}
/// \brief Keep track of a forward cross-cu reference from this unit
/// to \p Die that lives in \p RefUnit.
void CompileUnit::noteForwardReference(DIE *Die, const CompileUnit *RefUnit,
DeclContext *Ctxt, PatchLocation Attr) {
ForwardDIEReferences.emplace_back(Die, RefUnit, Ctxt, Attr);
}
/// \brief Apply all fixups recorded by noteForwardReference().
void CompileUnit::fixupForwardReferences() {
for (const auto &Ref : ForwardDIEReferences) {
DIE *RefDie;
const CompileUnit *RefUnit;
Reapply "AsmPrinter: Change DIEValue to be stored by value" This reverts commit r238350, effectively reapplying r238349 after fixing (all?) the problems, all somehow related to how I was using `AlignedArrayCharUnion<>` inside `DIEValue`: - MSVC can only handle `sizeof()` on types, not values. Change the assert. - GCC doesn't know the `is_trivially_copyable` type trait. Instead of asserting it, add destructors. - Call placement new even when constructing POD (i.e., the pointers). - Instead of copying the char buffer, copy the casted classes. I've left in a couple of `static_assert`s that I think both MSVC and GCC know how to handle. If the bots disagree with me, I'll remove them. - Check that the constructed type is either standard layout or a pointer. This protects against a programming error: we really want the "small" `DIEValue`s to be small and simple, so don't accidentally change them not to be. - Similarly, check that the size of the buffer is no bigger than a `uint64_t` or a pointer. (I thought checking against `sizeof(uint64_t)` would be good enough, but Chandler suggested that pointers might sometimes be bigger than that in the context of sanitizers.) I've also committed r238359 in the meantime, which introduces a DIEValue.def to simplify dispatching between the various types (thanks to a review comment by David Blaikie). Without that, this commit would be almost unintelligible. Here's the original commit message: -- Change `DIEValue` to be stored/passed/etc. by value, instead of reference. It's now a discriminated union, with a `Val` field storing the actual type. The classes that used to inherit from `DIEValue` no longer do. There are two categories of these: - Small values fit in a single pointer and are stored by value. - Large values require auxiliary storage, and are stored by reference. The only non-mechanical change is to tools/dsymutil/DwarfLinker.cpp. It was relying on `DIEInteger`s being passed around by reference, so I replaced that assumption with a `PatchLocation` type that stores a safe reference to where the `DIEInteger` lives instead. This commit causes a temporary regression in memory usage, since I've left merging `DIEAbbrevData` into `DIEValue` for a follow-up commit. I measured an increase from 845 MB to 879 MB, around 3.9%. The follow-up drops it lower than the starting point, and I've only recently brought the memory this low anyway, so I'm committing these changes separately to keep them incremental. (I also considered swapping the commits, but the other one first would cause a lot more code churn.) (I'm looking at `llc` memory usage on `verify-uselistorder.lto.opt.bc`; see r236629 for details.) -- llvm-svn: 238362
2015-05-28 06:14:58 +08:00
PatchLocation Attr;
DeclContext *Ctxt;
std::tie(RefDie, RefUnit, Ctxt, Attr) = Ref;
if (Ctxt && Ctxt->getCanonicalDIEOffset())
Attr.set(Ctxt->getCanonicalDIEOffset());
else
Attr.set(RefDie->getOffset() + RefUnit->getStartOffset());
}
}
void CompileUnit::addFunctionRange(uint64_t FuncLowPc, uint64_t FuncHighPc,
int64_t PcOffset) {
Ranges.insert(FuncLowPc, FuncHighPc, PcOffset);
this->LowPc = std::min(LowPc, FuncLowPc + PcOffset);
this->HighPc = std::max(HighPc, FuncHighPc + PcOffset);
}
Reapply "AsmPrinter: Change DIEValue to be stored by value" This reverts commit r238350, effectively reapplying r238349 after fixing (all?) the problems, all somehow related to how I was using `AlignedArrayCharUnion<>` inside `DIEValue`: - MSVC can only handle `sizeof()` on types, not values. Change the assert. - GCC doesn't know the `is_trivially_copyable` type trait. Instead of asserting it, add destructors. - Call placement new even when constructing POD (i.e., the pointers). - Instead of copying the char buffer, copy the casted classes. I've left in a couple of `static_assert`s that I think both MSVC and GCC know how to handle. If the bots disagree with me, I'll remove them. - Check that the constructed type is either standard layout or a pointer. This protects against a programming error: we really want the "small" `DIEValue`s to be small and simple, so don't accidentally change them not to be. - Similarly, check that the size of the buffer is no bigger than a `uint64_t` or a pointer. (I thought checking against `sizeof(uint64_t)` would be good enough, but Chandler suggested that pointers might sometimes be bigger than that in the context of sanitizers.) I've also committed r238359 in the meantime, which introduces a DIEValue.def to simplify dispatching between the various types (thanks to a review comment by David Blaikie). Without that, this commit would be almost unintelligible. Here's the original commit message: -- Change `DIEValue` to be stored/passed/etc. by value, instead of reference. It's now a discriminated union, with a `Val` field storing the actual type. The classes that used to inherit from `DIEValue` no longer do. There are two categories of these: - Small values fit in a single pointer and are stored by value. - Large values require auxiliary storage, and are stored by reference. The only non-mechanical change is to tools/dsymutil/DwarfLinker.cpp. It was relying on `DIEInteger`s being passed around by reference, so I replaced that assumption with a `PatchLocation` type that stores a safe reference to where the `DIEInteger` lives instead. This commit causes a temporary regression in memory usage, since I've left merging `DIEAbbrevData` into `DIEValue` for a follow-up commit. I measured an increase from 845 MB to 879 MB, around 3.9%. The follow-up drops it lower than the starting point, and I've only recently brought the memory this low anyway, so I'm committing these changes separately to keep them incremental. (I also considered swapping the commits, but the other one first would cause a lot more code churn.) (I'm looking at `llc` memory usage on `verify-uselistorder.lto.opt.bc`; see r236629 for details.) -- llvm-svn: 238362
2015-05-28 06:14:58 +08:00
void CompileUnit::noteRangeAttribute(const DIE &Die, PatchLocation Attr) {
if (Die.getTag() != dwarf::DW_TAG_compile_unit)
RangeAttributes.push_back(Attr);
else
UnitRangeAttribute = Attr;
}
Reapply "AsmPrinter: Change DIEValue to be stored by value" This reverts commit r238350, effectively reapplying r238349 after fixing (all?) the problems, all somehow related to how I was using `AlignedArrayCharUnion<>` inside `DIEValue`: - MSVC can only handle `sizeof()` on types, not values. Change the assert. - GCC doesn't know the `is_trivially_copyable` type trait. Instead of asserting it, add destructors. - Call placement new even when constructing POD (i.e., the pointers). - Instead of copying the char buffer, copy the casted classes. I've left in a couple of `static_assert`s that I think both MSVC and GCC know how to handle. If the bots disagree with me, I'll remove them. - Check that the constructed type is either standard layout or a pointer. This protects against a programming error: we really want the "small" `DIEValue`s to be small and simple, so don't accidentally change them not to be. - Similarly, check that the size of the buffer is no bigger than a `uint64_t` or a pointer. (I thought checking against `sizeof(uint64_t)` would be good enough, but Chandler suggested that pointers might sometimes be bigger than that in the context of sanitizers.) I've also committed r238359 in the meantime, which introduces a DIEValue.def to simplify dispatching between the various types (thanks to a review comment by David Blaikie). Without that, this commit would be almost unintelligible. Here's the original commit message: -- Change `DIEValue` to be stored/passed/etc. by value, instead of reference. It's now a discriminated union, with a `Val` field storing the actual type. The classes that used to inherit from `DIEValue` no longer do. There are two categories of these: - Small values fit in a single pointer and are stored by value. - Large values require auxiliary storage, and are stored by reference. The only non-mechanical change is to tools/dsymutil/DwarfLinker.cpp. It was relying on `DIEInteger`s being passed around by reference, so I replaced that assumption with a `PatchLocation` type that stores a safe reference to where the `DIEInteger` lives instead. This commit causes a temporary regression in memory usage, since I've left merging `DIEAbbrevData` into `DIEValue` for a follow-up commit. I measured an increase from 845 MB to 879 MB, around 3.9%. The follow-up drops it lower than the starting point, and I've only recently brought the memory this low anyway, so I'm committing these changes separately to keep them incremental. (I also considered swapping the commits, but the other one first would cause a lot more code churn.) (I'm looking at `llc` memory usage on `verify-uselistorder.lto.opt.bc`; see r236629 for details.) -- llvm-svn: 238362
2015-05-28 06:14:58 +08:00
void CompileUnit::noteLocationAttribute(PatchLocation Attr, int64_t PcOffset) {
LocationAttributes.emplace_back(Attr, PcOffset);
}
/// \brief Add a name accelerator entry for \p Die with \p Name
/// which is stored in the string table at \p Offset.
void CompileUnit::addNameAccelerator(const DIE *Die, const char *Name,
uint32_t Offset, bool SkipPubSection) {
Pubnames.emplace_back(Name, Die, Offset, SkipPubSection);
}
/// \brief Add a type accelerator entry for \p Die with \p Name
/// which is stored in the string table at \p Offset.
void CompileUnit::addTypeAccelerator(const DIE *Die, const char *Name,
uint32_t Offset) {
Pubtypes.emplace_back(Name, Die, Offset, false);
}
/// \brief A string table that doesn't need relocations.
///
/// We are doing a final link, no need for a string table that
/// has relocation entries for every reference to it. This class
/// provides this ablitity by just associating offsets with
/// strings.
class NonRelocatableStringpool {
public:
/// \brief Entries are stored into the StringMap and simply linked
/// together through the second element of this pair in order to
/// keep track of insertion order.
typedef StringMap<std::pair<uint32_t, StringMapEntryBase *>, BumpPtrAllocator>
MapTy;
NonRelocatableStringpool()
: CurrentEndOffset(0), Sentinel(0), Last(&Sentinel) {
// Legacy dsymutil puts an empty string at the start of the line
// table.
getStringOffset("");
}
/// \brief Get the offset of string \p S in the string table. This
/// can insert a new element or return the offset of a preexisitng
/// one.
uint32_t getStringOffset(StringRef S);
/// \brief Get permanent storage for \p S (but do not necessarily
/// emit \p S in the output section).
/// \returns The StringRef that points to permanent storage to use
/// in place of \p S.
StringRef internString(StringRef S);
// \brief Return the first entry of the string table.
const MapTy::MapEntryTy *getFirstEntry() const {
return getNextEntry(&Sentinel);
}
// \brief Get the entry following \p E in the string table or null
// if \p E was the last entry.
const MapTy::MapEntryTy *getNextEntry(const MapTy::MapEntryTy *E) const {
return static_cast<const MapTy::MapEntryTy *>(E->getValue().second);
}
uint64_t getSize() { return CurrentEndOffset; }
private:
MapTy Strings;
uint32_t CurrentEndOffset;
MapTy::MapEntryTy Sentinel, *Last;
};
/// \brief Get the offset of string \p S in the string table. This
/// can insert a new element or return the offset of a preexisitng
/// one.
uint32_t NonRelocatableStringpool::getStringOffset(StringRef S) {
if (S.empty() && !Strings.empty())
return 0;
std::pair<uint32_t, StringMapEntryBase *> Entry(0, nullptr);
MapTy::iterator It;
bool Inserted;
// A non-empty string can't be at offset 0, so if we have an entry
// with a 0 offset, it must be a previously interned string.
std::tie(It, Inserted) = Strings.insert(std::make_pair(S, Entry));
if (Inserted || It->getValue().first == 0) {
// Set offset and chain at the end of the entries list.
It->getValue().first = CurrentEndOffset;
CurrentEndOffset += S.size() + 1; // +1 for the '\0'.
Last->getValue().second = &*It;
Last = &*It;
}
return It->getValue().first;
}
/// \brief Put \p S into the StringMap so that it gets permanent
/// storage, but do not actually link it in the chain of elements
/// that go into the output section. A latter call to
/// getStringOffset() with the same string will chain it though.
StringRef NonRelocatableStringpool::internString(StringRef S) {
std::pair<uint32_t, StringMapEntryBase *> Entry(0, nullptr);
auto InsertResult = Strings.insert(std::make_pair(S, Entry));
return InsertResult.first->getKey();
}
/// \brief The Dwarf streaming logic
///
/// All interactions with the MC layer that is used to build the debug
/// information binary representation are handled in this class.
class DwarfStreamer {
/// \defgroup MCObjects MC layer objects constructed by the streamer
/// @{
std::unique_ptr<MCRegisterInfo> MRI;
std::unique_ptr<MCAsmInfo> MAI;
std::unique_ptr<MCObjectFileInfo> MOFI;
std::unique_ptr<MCContext> MC;
MCAsmBackend *MAB; // Owned by MCStreamer
std::unique_ptr<MCInstrInfo> MII;
std::unique_ptr<MCSubtargetInfo> MSTI;
MCCodeEmitter *MCE; // Owned by MCStreamer
MCStreamer *MS; // Owned by AsmPrinter
std::unique_ptr<TargetMachine> TM;
std::unique_ptr<AsmPrinter> Asm;
/// @}
/// \brief the file we stream the linked Dwarf to.
std::unique_ptr<raw_fd_ostream> OutFile;
uint32_t RangesSectionSize;
uint32_t LocSectionSize;
uint32_t LineSectionSize;
uint32_t FrameSectionSize;
/// \brief Emit the pubnames or pubtypes section contribution for \p
/// Unit into \p Sec. The data is provided in \p Names.
void emitPubSectionForUnit(MCSection *Sec, StringRef Name,
const CompileUnit &Unit,
const std::vector<CompileUnit::AccelInfo> &Names);
public:
/// \brief Actually create the streamer and the ouptut file.
///
/// This could be done directly in the constructor, but it feels
/// more natural to handle errors through return value.
bool init(Triple TheTriple, StringRef OutputFilename);
/// \brief Dump the file to the disk.
bool finish();
AsmPrinter &getAsmPrinter() const { return *Asm; }
/// \brief Set the current output section to debug_info and change
/// the MC Dwarf version to \p DwarfVersion.
void switchToDebugInfoSection(unsigned DwarfVersion);
/// \brief Emit the compilation unit header for \p Unit in the
/// debug_info section.
///
/// As a side effect, this also switches the current Dwarf version
/// of the MC layer to the one of U.getOrigUnit().
void emitCompileUnitHeader(CompileUnit &Unit);
/// \brief Recursively emit the DIE tree rooted at \p Die.
void emitDIE(DIE &Die);
/// \brief Emit the abbreviation table \p Abbrevs to the
/// debug_abbrev section.
void emitAbbrevs(const std::vector<DIEAbbrev *> &Abbrevs);
/// \brief Emit the string table described by \p Pool.
void emitStrings(const NonRelocatableStringpool &Pool);
/// \brief Emit debug_ranges for \p FuncRange by translating the
/// original \p Entries.
void emitRangesEntries(
int64_t UnitPcOffset, uint64_t OrigLowPc,
FunctionIntervals::const_iterator FuncRange,
const std::vector<DWARFDebugRangeList::RangeListEntry> &Entries,
unsigned AddressSize);
/// \brief Emit debug_aranges entries for \p Unit and if \p
/// DoRangesSection is true, also emit the debug_ranges entries for
/// the DW_TAG_compile_unit's DW_AT_ranges attribute.
void emitUnitRangesEntries(CompileUnit &Unit, bool DoRangesSection);
uint32_t getRangesSectionSize() const { return RangesSectionSize; }
/// \brief Emit the debug_loc contribution for \p Unit by copying
/// the entries from \p Dwarf and offseting them. Update the
/// location attributes to point to the new entries.
void emitLocationsForUnit(const CompileUnit &Unit, DWARFContext &Dwarf);
/// \brief Emit the line table described in \p Rows into the
/// debug_line section.
void emitLineTableForUnit(StringRef PrologueBytes, unsigned MinInstLength,
std::vector<DWARFDebugLine::Row> &Rows,
unsigned AdddressSize);
uint32_t getLineSectionSize() const { return LineSectionSize; }
/// \brief Emit the .debug_pubnames contribution for \p Unit.
void emitPubNamesForUnit(const CompileUnit &Unit);
/// \brief Emit the .debug_pubtypes contribution for \p Unit.
void emitPubTypesForUnit(const CompileUnit &Unit);
/// \brief Emit a CIE.
void emitCIE(StringRef CIEBytes);
/// \brief Emit an FDE with data \p Bytes.
void emitFDE(uint32_t CIEOffset, uint32_t AddreSize, uint32_t Address,
StringRef Bytes);
uint32_t getFrameSectionSize() const { return FrameSectionSize; }
};
bool DwarfStreamer::init(Triple TheTriple, StringRef OutputFilename) {
std::string ErrorStr;
std::string TripleName;
StringRef Context = "dwarf streamer init";
// Get the target.
const Target *TheTarget =
TargetRegistry::lookupTarget(TripleName, TheTriple, ErrorStr);
if (!TheTarget)
return error(ErrorStr, Context);
TripleName = TheTriple.getTriple();
// Create all the MC Objects.
MRI.reset(TheTarget->createMCRegInfo(TripleName));
if (!MRI)
return error(Twine("no register info for target ") + TripleName, Context);
MAI.reset(TheTarget->createMCAsmInfo(*MRI, TripleName));
if (!MAI)
return error("no asm info for target " + TripleName, Context);
MOFI.reset(new MCObjectFileInfo);
MC.reset(new MCContext(MAI.get(), MRI.get(), MOFI.get()));
MOFI->InitMCObjectFileInfo(TheTriple, Reloc::Default, CodeModel::Default,
*MC);
MAB = TheTarget->createMCAsmBackend(*MRI, TripleName, "");
if (!MAB)
return error("no asm backend for target " + TripleName, Context);
MII.reset(TheTarget->createMCInstrInfo());
if (!MII)
return error("no instr info info for target " + TripleName, Context);
MSTI.reset(TheTarget->createMCSubtargetInfo(TripleName, "", ""));
if (!MSTI)
return error("no subtarget info for target " + TripleName, Context);
MCE = TheTarget->createMCCodeEmitter(*MII, *MRI, *MC);
if (!MCE)
return error("no code emitter for target " + TripleName, Context);
// Create the output file.
std::error_code EC;
OutFile =
llvm::make_unique<raw_fd_ostream>(OutputFilename, EC, sys::fs::F_None);
if (EC)
return error(Twine(OutputFilename) + ": " + EC.message(), Context);
MS = TheTarget->createMCObjectStreamer(TheTriple, *MC, *MAB, *OutFile, MCE,
*MSTI, false,
/*DWARFMustBeAtTheEnd*/ false);
if (!MS)
return error("no object streamer for target " + TripleName, Context);
// Finally create the AsmPrinter we'll use to emit the DIEs.
TM.reset(TheTarget->createTargetMachine(TripleName, "", "", TargetOptions()));
if (!TM)
return error("no target machine for target " + TripleName, Context);
Asm.reset(TheTarget->createAsmPrinter(*TM, std::unique_ptr<MCStreamer>(MS)));
if (!Asm)
return error("no asm printer for target " + TripleName, Context);
RangesSectionSize = 0;
LocSectionSize = 0;
LineSectionSize = 0;
FrameSectionSize = 0;
return true;
}
bool DwarfStreamer::finish() {
MS->Finish();
return true;
}
/// \brief Set the current output section to debug_info and change
/// the MC Dwarf version to \p DwarfVersion.
void DwarfStreamer::switchToDebugInfoSection(unsigned DwarfVersion) {
MS->SwitchSection(MOFI->getDwarfInfoSection());
MC->setDwarfVersion(DwarfVersion);
}
/// \brief Emit the compilation unit header for \p Unit in the
/// debug_info section.
///
/// A Dwarf scetion header is encoded as:
/// uint32_t Unit length (omiting this field)
/// uint16_t Version
/// uint32_t Abbreviation table offset
/// uint8_t Address size
///
/// Leading to a total of 11 bytes.
void DwarfStreamer::emitCompileUnitHeader(CompileUnit &Unit) {
unsigned Version = Unit.getOrigUnit().getVersion();
switchToDebugInfoSection(Version);
// Emit size of content not including length itself. The size has
// already been computed in CompileUnit::computeOffsets(). Substract
// 4 to that size to account for the length field.
Asm->EmitInt32(Unit.getNextUnitOffset() - Unit.getStartOffset() - 4);
Asm->EmitInt16(Version);
// We share one abbreviations table across all units so it's always at the
// start of the section.
Asm->EmitInt32(0);
Asm->EmitInt8(Unit.getOrigUnit().getAddressByteSize());
}
/// \brief Emit the \p Abbrevs array as the shared abbreviation table
/// for the linked Dwarf file.
void DwarfStreamer::emitAbbrevs(const std::vector<DIEAbbrev *> &Abbrevs) {
MS->SwitchSection(MOFI->getDwarfAbbrevSection());
Asm->emitDwarfAbbrevs(Abbrevs);
}
/// \brief Recursively emit the DIE tree rooted at \p Die.
void DwarfStreamer::emitDIE(DIE &Die) {
MS->SwitchSection(MOFI->getDwarfInfoSection());
Asm->emitDwarfDIE(Die);
}
/// \brief Emit the debug_str section stored in \p Pool.
void DwarfStreamer::emitStrings(const NonRelocatableStringpool &Pool) {
Asm->OutStreamer->SwitchSection(MOFI->getDwarfStrSection());
for (auto *Entry = Pool.getFirstEntry(); Entry;
Entry = Pool.getNextEntry(Entry))
Asm->OutStreamer->EmitBytes(
StringRef(Entry->getKey().data(), Entry->getKey().size() + 1));
}
/// \brief Emit the debug_range section contents for \p FuncRange by
/// translating the original \p Entries. The debug_range section
/// format is totally trivial, consisting just of pairs of address
/// sized addresses describing the ranges.
void DwarfStreamer::emitRangesEntries(
int64_t UnitPcOffset, uint64_t OrigLowPc,
FunctionIntervals::const_iterator FuncRange,
const std::vector<DWARFDebugRangeList::RangeListEntry> &Entries,
unsigned AddressSize) {
MS->SwitchSection(MC->getObjectFileInfo()->getDwarfRangesSection());
// Offset each range by the right amount.
int64_t PcOffset = FuncRange.value() + UnitPcOffset;
for (const auto &Range : Entries) {
if (Range.isBaseAddressSelectionEntry(AddressSize)) {
warn("unsupported base address selection operation",
"emitting debug_ranges");
break;
}
// Do not emit empty ranges.
if (Range.StartAddress == Range.EndAddress)
continue;
// All range entries should lie in the function range.
if (!(Range.StartAddress + OrigLowPc >= FuncRange.start() &&
Range.EndAddress + OrigLowPc <= FuncRange.stop()))
warn("inconsistent range data.", "emitting debug_ranges");
MS->EmitIntValue(Range.StartAddress + PcOffset, AddressSize);
MS->EmitIntValue(Range.EndAddress + PcOffset, AddressSize);
RangesSectionSize += 2 * AddressSize;
}
// Add the terminator entry.
MS->EmitIntValue(0, AddressSize);
MS->EmitIntValue(0, AddressSize);
RangesSectionSize += 2 * AddressSize;
}
/// \brief Emit the debug_aranges contribution of a unit and
/// if \p DoDebugRanges is true the debug_range contents for a
/// compile_unit level DW_AT_ranges attribute (Which are basically the
/// same thing with a different base address).
/// Just aggregate all the ranges gathered inside that unit.
void DwarfStreamer::emitUnitRangesEntries(CompileUnit &Unit,
bool DoDebugRanges) {
unsigned AddressSize = Unit.getOrigUnit().getAddressByteSize();
// Gather the ranges in a vector, so that we can simplify them. The
// IntervalMap will have coalesced the non-linked ranges, but here
// we want to coalesce the linked addresses.
std::vector<std::pair<uint64_t, uint64_t>> Ranges;
const auto &FunctionRanges = Unit.getFunctionRanges();
for (auto Range = FunctionRanges.begin(), End = FunctionRanges.end();
Range != End; ++Range)
Ranges.push_back(std::make_pair(Range.start() + Range.value(),
Range.stop() + Range.value()));
// The object addresses where sorted, but again, the linked
// addresses might end up in a different order.
std::sort(Ranges.begin(), Ranges.end());
if (!Ranges.empty()) {
MS->SwitchSection(MC->getObjectFileInfo()->getDwarfARangesSection());
MCSymbol *BeginLabel = Asm->createTempSymbol("Barange");
MCSymbol *EndLabel = Asm->createTempSymbol("Earange");
unsigned HeaderSize =
sizeof(int32_t) + // Size of contents (w/o this field
sizeof(int16_t) + // DWARF ARange version number
sizeof(int32_t) + // Offset of CU in the .debug_info section
sizeof(int8_t) + // Pointer Size (in bytes)
sizeof(int8_t); // Segment Size (in bytes)
unsigned TupleSize = AddressSize * 2;
unsigned Padding = OffsetToAlignment(HeaderSize, TupleSize);
Asm->EmitLabelDifference(EndLabel, BeginLabel, 4); // Arange length
Asm->OutStreamer->EmitLabel(BeginLabel);
Asm->EmitInt16(dwarf::DW_ARANGES_VERSION); // Version number
Asm->EmitInt32(Unit.getStartOffset()); // Corresponding unit's offset
Asm->EmitInt8(AddressSize); // Address size
Asm->EmitInt8(0); // Segment size
Asm->OutStreamer->EmitFill(Padding, 0x0);
for (auto Range = Ranges.begin(), End = Ranges.end(); Range != End;
++Range) {
uint64_t RangeStart = Range->first;
MS->EmitIntValue(RangeStart, AddressSize);
while ((Range + 1) != End && Range->second == (Range + 1)->first)
++Range;
MS->EmitIntValue(Range->second - RangeStart, AddressSize);
}
// Emit terminator
Asm->OutStreamer->EmitIntValue(0, AddressSize);
Asm->OutStreamer->EmitIntValue(0, AddressSize);
Asm->OutStreamer->EmitLabel(EndLabel);
}
if (!DoDebugRanges)
return;
MS->SwitchSection(MC->getObjectFileInfo()->getDwarfRangesSection());
// Offset each range by the right amount.
int64_t PcOffset = -Unit.getLowPc();
// Emit coalesced ranges.
for (auto Range = Ranges.begin(), End = Ranges.end(); Range != End; ++Range) {
MS->EmitIntValue(Range->first + PcOffset, AddressSize);
while (Range + 1 != End && Range->second == (Range + 1)->first)
++Range;
MS->EmitIntValue(Range->second + PcOffset, AddressSize);
RangesSectionSize += 2 * AddressSize;
}
// Add the terminator entry.
MS->EmitIntValue(0, AddressSize);
MS->EmitIntValue(0, AddressSize);
RangesSectionSize += 2 * AddressSize;
}
/// \brief Emit location lists for \p Unit and update attribtues to
/// point to the new entries.
void DwarfStreamer::emitLocationsForUnit(const CompileUnit &Unit,
DWARFContext &Dwarf) {
Reapply "AsmPrinter: Change DIEValue to be stored by value" This reverts commit r238350, effectively reapplying r238349 after fixing (all?) the problems, all somehow related to how I was using `AlignedArrayCharUnion<>` inside `DIEValue`: - MSVC can only handle `sizeof()` on types, not values. Change the assert. - GCC doesn't know the `is_trivially_copyable` type trait. Instead of asserting it, add destructors. - Call placement new even when constructing POD (i.e., the pointers). - Instead of copying the char buffer, copy the casted classes. I've left in a couple of `static_assert`s that I think both MSVC and GCC know how to handle. If the bots disagree with me, I'll remove them. - Check that the constructed type is either standard layout or a pointer. This protects against a programming error: we really want the "small" `DIEValue`s to be small and simple, so don't accidentally change them not to be. - Similarly, check that the size of the buffer is no bigger than a `uint64_t` or a pointer. (I thought checking against `sizeof(uint64_t)` would be good enough, but Chandler suggested that pointers might sometimes be bigger than that in the context of sanitizers.) I've also committed r238359 in the meantime, which introduces a DIEValue.def to simplify dispatching between the various types (thanks to a review comment by David Blaikie). Without that, this commit would be almost unintelligible. Here's the original commit message: -- Change `DIEValue` to be stored/passed/etc. by value, instead of reference. It's now a discriminated union, with a `Val` field storing the actual type. The classes that used to inherit from `DIEValue` no longer do. There are two categories of these: - Small values fit in a single pointer and are stored by value. - Large values require auxiliary storage, and are stored by reference. The only non-mechanical change is to tools/dsymutil/DwarfLinker.cpp. It was relying on `DIEInteger`s being passed around by reference, so I replaced that assumption with a `PatchLocation` type that stores a safe reference to where the `DIEInteger` lives instead. This commit causes a temporary regression in memory usage, since I've left merging `DIEAbbrevData` into `DIEValue` for a follow-up commit. I measured an increase from 845 MB to 879 MB, around 3.9%. The follow-up drops it lower than the starting point, and I've only recently brought the memory this low anyway, so I'm committing these changes separately to keep them incremental. (I also considered swapping the commits, but the other one first would cause a lot more code churn.) (I'm looking at `llc` memory usage on `verify-uselistorder.lto.opt.bc`; see r236629 for details.) -- llvm-svn: 238362
2015-05-28 06:14:58 +08:00
const auto &Attributes = Unit.getLocationAttributes();
if (Attributes.empty())
return;
MS->SwitchSection(MC->getObjectFileInfo()->getDwarfLocSection());
unsigned AddressSize = Unit.getOrigUnit().getAddressByteSize();
const DWARFSection &InputSec = Dwarf.getLocSection();
DataExtractor Data(InputSec.Data, Dwarf.isLittleEndian(), AddressSize);
DWARFUnit &OrigUnit = Unit.getOrigUnit();
const auto *OrigUnitDie = OrigUnit.getUnitDIE(false);
int64_t UnitPcOffset = 0;
uint64_t OrigLowPc = OrigUnitDie->getAttributeValueAsAddress(
&OrigUnit, dwarf::DW_AT_low_pc, -1ULL);
if (OrigLowPc != -1ULL)
UnitPcOffset = int64_t(OrigLowPc) - Unit.getLowPc();
for (const auto &Attr : Attributes) {
Reapply "AsmPrinter: Change DIEValue to be stored by value" This reverts commit r238350, effectively reapplying r238349 after fixing (all?) the problems, all somehow related to how I was using `AlignedArrayCharUnion<>` inside `DIEValue`: - MSVC can only handle `sizeof()` on types, not values. Change the assert. - GCC doesn't know the `is_trivially_copyable` type trait. Instead of asserting it, add destructors. - Call placement new even when constructing POD (i.e., the pointers). - Instead of copying the char buffer, copy the casted classes. I've left in a couple of `static_assert`s that I think both MSVC and GCC know how to handle. If the bots disagree with me, I'll remove them. - Check that the constructed type is either standard layout or a pointer. This protects against a programming error: we really want the "small" `DIEValue`s to be small and simple, so don't accidentally change them not to be. - Similarly, check that the size of the buffer is no bigger than a `uint64_t` or a pointer. (I thought checking against `sizeof(uint64_t)` would be good enough, but Chandler suggested that pointers might sometimes be bigger than that in the context of sanitizers.) I've also committed r238359 in the meantime, which introduces a DIEValue.def to simplify dispatching between the various types (thanks to a review comment by David Blaikie). Without that, this commit would be almost unintelligible. Here's the original commit message: -- Change `DIEValue` to be stored/passed/etc. by value, instead of reference. It's now a discriminated union, with a `Val` field storing the actual type. The classes that used to inherit from `DIEValue` no longer do. There are two categories of these: - Small values fit in a single pointer and are stored by value. - Large values require auxiliary storage, and are stored by reference. The only non-mechanical change is to tools/dsymutil/DwarfLinker.cpp. It was relying on `DIEInteger`s being passed around by reference, so I replaced that assumption with a `PatchLocation` type that stores a safe reference to where the `DIEInteger` lives instead. This commit causes a temporary regression in memory usage, since I've left merging `DIEAbbrevData` into `DIEValue` for a follow-up commit. I measured an increase from 845 MB to 879 MB, around 3.9%. The follow-up drops it lower than the starting point, and I've only recently brought the memory this low anyway, so I'm committing these changes separately to keep them incremental. (I also considered swapping the commits, but the other one first would cause a lot more code churn.) (I'm looking at `llc` memory usage on `verify-uselistorder.lto.opt.bc`; see r236629 for details.) -- llvm-svn: 238362
2015-05-28 06:14:58 +08:00
uint32_t Offset = Attr.first.get();
Attr.first.set(LocSectionSize);
// This is the quantity to add to the old location address to get
// the correct address for the new one.
int64_t LocPcOffset = Attr.second + UnitPcOffset;
while (Data.isValidOffset(Offset)) {
uint64_t Low = Data.getUnsigned(&Offset, AddressSize);
uint64_t High = Data.getUnsigned(&Offset, AddressSize);
LocSectionSize += 2 * AddressSize;
if (Low == 0 && High == 0) {
Asm->OutStreamer->EmitIntValue(0, AddressSize);
Asm->OutStreamer->EmitIntValue(0, AddressSize);
break;
}
Asm->OutStreamer->EmitIntValue(Low + LocPcOffset, AddressSize);
Asm->OutStreamer->EmitIntValue(High + LocPcOffset, AddressSize);
uint64_t Length = Data.getU16(&Offset);
Asm->OutStreamer->EmitIntValue(Length, 2);
// Just copy the bytes over.
Asm->OutStreamer->EmitBytes(
StringRef(InputSec.Data.substr(Offset, Length)));
Offset += Length;
LocSectionSize += Length + 2;
}
}
}
void DwarfStreamer::emitLineTableForUnit(StringRef PrologueBytes,
unsigned MinInstLength,
std::vector<DWARFDebugLine::Row> &Rows,
unsigned PointerSize) {
// Switch to the section where the table will be emitted into.
MS->SwitchSection(MC->getObjectFileInfo()->getDwarfLineSection());
MCSymbol *LineStartSym = MC->createTempSymbol();
MCSymbol *LineEndSym = MC->createTempSymbol();
// The first 4 bytes is the total length of the information for this
// compilation unit (not including these 4 bytes for the length).
Asm->EmitLabelDifference(LineEndSym, LineStartSym, 4);
Asm->OutStreamer->EmitLabel(LineStartSym);
// Copy Prologue.
MS->EmitBytes(PrologueBytes);
LineSectionSize += PrologueBytes.size() + 4;
SmallString<128> EncodingBuffer;
raw_svector_ostream EncodingOS(EncodingBuffer);
if (Rows.empty()) {
// We only have the dummy entry, dsymutil emits an entry with a 0
// address in that case.
MCDwarfLineAddr::Encode(*MC, INT64_MAX, 0, EncodingOS);
MS->EmitBytes(EncodingOS.str());
LineSectionSize += EncodingBuffer.size();
MS->EmitLabel(LineEndSym);
return;
}
// Line table state machine fields
unsigned FileNum = 1;
unsigned LastLine = 1;
unsigned Column = 0;
unsigned IsStatement = 1;
unsigned Isa = 0;
uint64_t Address = -1ULL;
unsigned RowsSinceLastSequence = 0;
for (unsigned Idx = 0; Idx < Rows.size(); ++Idx) {
auto &Row = Rows[Idx];
int64_t AddressDelta;
if (Address == -1ULL) {
MS->EmitIntValue(dwarf::DW_LNS_extended_op, 1);
MS->EmitULEB128IntValue(PointerSize + 1);
MS->EmitIntValue(dwarf::DW_LNE_set_address, 1);
MS->EmitIntValue(Row.Address, PointerSize);
LineSectionSize += 2 + PointerSize + getULEB128Size(PointerSize + 1);
AddressDelta = 0;
} else {
AddressDelta = (Row.Address - Address) / MinInstLength;
}
// FIXME: code copied and transfromed from
// MCDwarf.cpp::EmitDwarfLineTable. We should find a way to share
// this code, but the current compatibility requirement with
// classic dsymutil makes it hard. Revisit that once this
// requirement is dropped.
if (FileNum != Row.File) {
FileNum = Row.File;
MS->EmitIntValue(dwarf::DW_LNS_set_file, 1);
MS->EmitULEB128IntValue(FileNum);
LineSectionSize += 1 + getULEB128Size(FileNum);
}
if (Column != Row.Column) {
Column = Row.Column;
MS->EmitIntValue(dwarf::DW_LNS_set_column, 1);
MS->EmitULEB128IntValue(Column);
LineSectionSize += 1 + getULEB128Size(Column);
}
// FIXME: We should handle the discriminator here, but dsymutil
// doesn' consider it, thus ignore it for now.
if (Isa != Row.Isa) {
Isa = Row.Isa;
MS->EmitIntValue(dwarf::DW_LNS_set_isa, 1);
MS->EmitULEB128IntValue(Isa);
LineSectionSize += 1 + getULEB128Size(Isa);
}
if (IsStatement != Row.IsStmt) {
IsStatement = Row.IsStmt;
MS->EmitIntValue(dwarf::DW_LNS_negate_stmt, 1);
LineSectionSize += 1;
}
if (Row.BasicBlock) {
MS->EmitIntValue(dwarf::DW_LNS_set_basic_block, 1);
LineSectionSize += 1;
}
if (Row.PrologueEnd) {
MS->EmitIntValue(dwarf::DW_LNS_set_prologue_end, 1);
LineSectionSize += 1;
}
if (Row.EpilogueBegin) {
MS->EmitIntValue(dwarf::DW_LNS_set_epilogue_begin, 1);
LineSectionSize += 1;
}
int64_t LineDelta = int64_t(Row.Line) - LastLine;
if (!Row.EndSequence) {
MCDwarfLineAddr::Encode(*MC, LineDelta, AddressDelta, EncodingOS);
MS->EmitBytes(EncodingOS.str());
LineSectionSize += EncodingBuffer.size();
EncodingBuffer.resize(0);
EncodingOS.resync();
Address = Row.Address;
LastLine = Row.Line;
RowsSinceLastSequence++;
} else {
if (LineDelta) {
MS->EmitIntValue(dwarf::DW_LNS_advance_line, 1);
MS->EmitSLEB128IntValue(LineDelta);
LineSectionSize += 1 + getSLEB128Size(LineDelta);
}
if (AddressDelta) {
MS->EmitIntValue(dwarf::DW_LNS_advance_pc, 1);
MS->EmitULEB128IntValue(AddressDelta);
LineSectionSize += 1 + getULEB128Size(AddressDelta);
}
MCDwarfLineAddr::Encode(*MC, INT64_MAX, 0, EncodingOS);
MS->EmitBytes(EncodingOS.str());
LineSectionSize += EncodingBuffer.size();
EncodingBuffer.resize(0);
EncodingOS.resync();
Address = -1ULL;
LastLine = FileNum = IsStatement = 1;
RowsSinceLastSequence = Column = Isa = 0;
}
}
if (RowsSinceLastSequence) {
MCDwarfLineAddr::Encode(*MC, INT64_MAX, 0, EncodingOS);
MS->EmitBytes(EncodingOS.str());
LineSectionSize += EncodingBuffer.size();
EncodingBuffer.resize(0);
EncodingOS.resync();
}
MS->EmitLabel(LineEndSym);
}
/// \brief Emit the pubnames or pubtypes section contribution for \p
/// Unit into \p Sec. The data is provided in \p Names.
void DwarfStreamer::emitPubSectionForUnit(
MCSection *Sec, StringRef SecName, const CompileUnit &Unit,
const std::vector<CompileUnit::AccelInfo> &Names) {
if (Names.empty())
return;
// Start the dwarf pubnames section.
Asm->OutStreamer->SwitchSection(Sec);
MCSymbol *BeginLabel = Asm->createTempSymbol("pub" + SecName + "_begin");
MCSymbol *EndLabel = Asm->createTempSymbol("pub" + SecName + "_end");
bool HeaderEmitted = false;
// Emit the pubnames for this compilation unit.
for (const auto &Name : Names) {
if (Name.SkipPubSection)
continue;
if (!HeaderEmitted) {
// Emit the header.
Asm->EmitLabelDifference(EndLabel, BeginLabel, 4); // Length
Asm->OutStreamer->EmitLabel(BeginLabel);
Asm->EmitInt16(dwarf::DW_PUBNAMES_VERSION); // Version
Asm->EmitInt32(Unit.getStartOffset()); // Unit offset
Asm->EmitInt32(Unit.getNextUnitOffset() - Unit.getStartOffset()); // Size
HeaderEmitted = true;
}
Asm->EmitInt32(Name.Die->getOffset());
Asm->OutStreamer->EmitBytes(
StringRef(Name.Name.data(), Name.Name.size() + 1));
}
if (!HeaderEmitted)
return;
Asm->EmitInt32(0); // End marker.
Asm->OutStreamer->EmitLabel(EndLabel);
}
/// \brief Emit .debug_pubnames for \p Unit.
void DwarfStreamer::emitPubNamesForUnit(const CompileUnit &Unit) {
emitPubSectionForUnit(MC->getObjectFileInfo()->getDwarfPubNamesSection(),
"names", Unit, Unit.getPubnames());
}
/// \brief Emit .debug_pubtypes for \p Unit.
void DwarfStreamer::emitPubTypesForUnit(const CompileUnit &Unit) {
emitPubSectionForUnit(MC->getObjectFileInfo()->getDwarfPubTypesSection(),
"types", Unit, Unit.getPubtypes());
}
/// \brief Emit a CIE into the debug_frame section.
void DwarfStreamer::emitCIE(StringRef CIEBytes) {
MS->SwitchSection(MC->getObjectFileInfo()->getDwarfFrameSection());
MS->EmitBytes(CIEBytes);
FrameSectionSize += CIEBytes.size();
}
/// \brief Emit a FDE into the debug_frame section. \p FDEBytes
/// contains the FDE data without the length, CIE offset and address
/// which will be replaced with the paramter values.
void DwarfStreamer::emitFDE(uint32_t CIEOffset, uint32_t AddrSize,
uint32_t Address, StringRef FDEBytes) {
MS->SwitchSection(MC->getObjectFileInfo()->getDwarfFrameSection());
MS->EmitIntValue(FDEBytes.size() + 4 + AddrSize, 4);
MS->EmitIntValue(CIEOffset, 4);
MS->EmitIntValue(Address, AddrSize);
MS->EmitBytes(FDEBytes);
FrameSectionSize += FDEBytes.size() + 8 + AddrSize;
}
/// \brief The core of the Dwarf linking logic.
///
/// The link of the dwarf information from the object files will be
/// driven by the selection of 'root DIEs', which are DIEs that
/// describe variables or functions that are present in the linked
/// binary (and thus have entries in the debug map). All the debug
/// information that will be linked (the DIEs, but also the line
/// tables, ranges, ...) is derived from that set of root DIEs.
///
/// The root DIEs are identified because they contain relocations that
/// correspond to a debug map entry at specific places (the low_pc for
/// a function, the location for a variable). These relocations are
/// called ValidRelocs in the DwarfLinker and are gathered as a very
/// first step when we start processing a DebugMapObject.
class DwarfLinker {
public:
DwarfLinker(StringRef OutputFilename, const LinkOptions &Options)
: OutputFilename(OutputFilename), Options(Options),
BinHolder(Options.Verbose), LastCIEOffset(0) {}
~DwarfLinker() {
for (auto *Abbrev : Abbreviations)
delete Abbrev;
}
/// \brief Link the contents of the DebugMap.
bool link(const DebugMap &);
private:
/// \brief Called at the start of a debug object link.
void startDebugObject(DWARFContext &, DebugMapObject &);
/// \brief Called at the end of a debug object link.
void endDebugObject();
/// \defgroup FindValidRelocations Translate debug map into a list
/// of relevant relocations
///
/// @{
struct ValidReloc {
uint32_t Offset;
uint32_t Size;
uint64_t Addend;
const DebugMapObject::DebugMapEntry *Mapping;
ValidReloc(uint32_t Offset, uint32_t Size, uint64_t Addend,
const DebugMapObject::DebugMapEntry *Mapping)
: Offset(Offset), Size(Size), Addend(Addend), Mapping(Mapping) {}
bool operator<(const ValidReloc &RHS) const { return Offset < RHS.Offset; }
};
/// \brief The valid relocations for the current DebugMapObject.
/// This vector is sorted by relocation offset.
std::vector<ValidReloc> ValidRelocs;
/// \brief Index into ValidRelocs of the next relocation to
/// consider. As we walk the DIEs in acsending file offset and as
/// ValidRelocs is sorted by file offset, keeping this index
/// uptodate is all we have to do to have a cheap lookup during the
/// root DIE selection and during DIE cloning.
unsigned NextValidReloc;
bool findValidRelocsInDebugInfo(const object::ObjectFile &Obj,
const DebugMapObject &DMO);
bool findValidRelocs(const object::SectionRef &Section,
const object::ObjectFile &Obj,
const DebugMapObject &DMO);
void findValidRelocsMachO(const object::SectionRef &Section,
const object::MachOObjectFile &Obj,
const DebugMapObject &DMO);
/// @}
/// \defgroup FindRootDIEs Find DIEs corresponding to debug map entries.
///
/// @{
/// \brief Recursively walk the \p DIE tree and look for DIEs to
/// keep. Store that information in \p CU's DIEInfo.
void lookForDIEsToKeep(const DWARFDebugInfoEntryMinimal &DIE,
const DebugMapObject &DMO, CompileUnit &CU,
unsigned Flags);
/// \brief Flags passed to DwarfLinker::lookForDIEsToKeep
enum TravesalFlags {
TF_Keep = 1 << 0, ///< Mark the traversed DIEs as kept.
TF_InFunctionScope = 1 << 1, ///< Current scope is a fucntion scope.
TF_DependencyWalk = 1 << 2, ///< Walking the dependencies of a kept DIE.
TF_ParentWalk = 1 << 3, ///< Walking up the parents of a kept DIE.
TF_ODR = 1 << 4, ///< Use the ODR whhile keeping dependants.
};
/// \brief Mark the passed DIE as well as all the ones it depends on
/// as kept.
void keepDIEAndDenpendencies(const DWARFDebugInfoEntryMinimal &DIE,
CompileUnit::DIEInfo &MyInfo,
const DebugMapObject &DMO, CompileUnit &CU,
bool UseODR);
unsigned shouldKeepDIE(const DWARFDebugInfoEntryMinimal &DIE,
CompileUnit &Unit, CompileUnit::DIEInfo &MyInfo,
unsigned Flags);
unsigned shouldKeepVariableDIE(const DWARFDebugInfoEntryMinimal &DIE,
CompileUnit &Unit,
CompileUnit::DIEInfo &MyInfo, unsigned Flags);
unsigned shouldKeepSubprogramDIE(const DWARFDebugInfoEntryMinimal &DIE,
CompileUnit &Unit,
CompileUnit::DIEInfo &MyInfo,
unsigned Flags);
bool hasValidRelocation(uint32_t StartOffset, uint32_t EndOffset,
CompileUnit::DIEInfo &Info);
/// @}
/// \defgroup Linking Methods used to link the debug information
///
/// @{
/// \brief Recursively clone \p InputDIE into an tree of DIE objects
/// where useless (as decided by lookForDIEsToKeep()) bits have been
/// stripped out and addresses have been rewritten according to the
/// debug map.
///
/// \param OutOffset is the offset the cloned DIE in the output
/// compile unit.
/// \param PCOffset (while cloning a function scope) is the offset
/// applied to the entry point of the function to get the linked address.
///
/// \returns the root of the cloned tree.
DIE *cloneDIE(const DWARFDebugInfoEntryMinimal &InputDIE, CompileUnit &U,
int64_t PCOffset, uint32_t OutOffset);
typedef DWARFAbbreviationDeclaration::AttributeSpec AttributeSpec;
/// \brief Information gathered and exchanged between the various
/// clone*Attributes helpers about the attributes of a particular DIE.
struct AttributesInfo {
const char *Name, *MangledName; ///< Names.
uint32_t NameOffset, MangledNameOffset; ///< Offsets in the string pool.
uint64_t OrigHighPc; ///< Value of AT_high_pc in the input DIE
int64_t PCOffset; ///< Offset to apply to PC addresses inside a function.
bool HasLowPc; ///< Does the DIE have a low_pc attribute?
bool IsDeclaration; ///< Is this DIE only a declaration?
AttributesInfo()
: Name(nullptr), MangledName(nullptr), NameOffset(0),
MangledNameOffset(0), OrigHighPc(0), PCOffset(0), HasLowPc(false),
IsDeclaration(false) {}
};
/// \brief Helper for cloneDIE.
unsigned cloneAttribute(DIE &Die, const DWARFDebugInfoEntryMinimal &InputDIE,
CompileUnit &U, const DWARFFormValue &Val,
const AttributeSpec AttrSpec, unsigned AttrSize,
AttributesInfo &AttrInfo);
/// \brief Helper for cloneDIE.
unsigned cloneStringAttribute(DIE &Die, AttributeSpec AttrSpec,
const DWARFFormValue &Val, const DWARFUnit &U);
/// \brief Helper for cloneDIE.
unsigned
cloneDieReferenceAttribute(DIE &Die,
const DWARFDebugInfoEntryMinimal &InputDIE,
AttributeSpec AttrSpec, unsigned AttrSize,
const DWARFFormValue &Val, CompileUnit &Unit);
/// \brief Helper for cloneDIE.
unsigned cloneBlockAttribute(DIE &Die, AttributeSpec AttrSpec,
const DWARFFormValue &Val, unsigned AttrSize);
/// \brief Helper for cloneDIE.
unsigned cloneAddressAttribute(DIE &Die, AttributeSpec AttrSpec,
const DWARFFormValue &Val,
const CompileUnit &Unit, AttributesInfo &Info);
/// \brief Helper for cloneDIE.
unsigned cloneScalarAttribute(DIE &Die,
const DWARFDebugInfoEntryMinimal &InputDIE,
CompileUnit &U, AttributeSpec AttrSpec,
const DWARFFormValue &Val, unsigned AttrSize,
AttributesInfo &Info);
/// \brief Helper for cloneDIE.
bool applyValidRelocs(MutableArrayRef<char> Data, uint32_t BaseOffset,
bool isLittleEndian);
/// \brief Assign an abbreviation number to \p Abbrev
void AssignAbbrev(DIEAbbrev &Abbrev);
/// \brief FoldingSet that uniques the abbreviations.
FoldingSet<DIEAbbrev> AbbreviationsSet;
/// \brief Storage for the unique Abbreviations.
/// This is passed to AsmPrinter::emitDwarfAbbrevs(), thus it cannot
/// be changed to a vecot of unique_ptrs.
std::vector<DIEAbbrev *> Abbreviations;
/// \brief Compute and emit debug_ranges section for \p Unit, and
/// patch the attributes referencing it.
void patchRangesForUnit(const CompileUnit &Unit, DWARFContext &Dwarf) const;
/// \brief Generate and emit the DW_AT_ranges attribute for a
/// compile_unit if it had one.
void generateUnitRanges(CompileUnit &Unit) const;
/// \brief Extract the line tables fromt he original dwarf, extract
/// the relevant parts according to the linked function ranges and
/// emit the result in the debug_line section.
void patchLineTableForUnit(CompileUnit &Unit, DWARFContext &OrigDwarf);
/// \brief Emit the accelerator entries for \p Unit.
void emitAcceleratorEntriesForUnit(CompileUnit &Unit);
/// \brief Patch the frame info for an object file and emit it.
void patchFrameInfoForObject(const DebugMapObject &, DWARFContext &,
unsigned AddressSize);
/// \brief DIELoc objects that need to be destructed (but not freed!).
std::vector<DIELoc *> DIELocs;
/// \brief DIEBlock objects that need to be destructed (but not freed!).
std::vector<DIEBlock *> DIEBlocks;
/// \brief Allocator used for all the DIEValue objects.
BumpPtrAllocator DIEAlloc;
/// @}
/// ODR Contexts for that link.
DeclContextTree ODRContexts;
/// \defgroup Helpers Various helper methods.
///
/// @{
const DWARFDebugInfoEntryMinimal *
resolveDIEReference(const DWARFFormValue &RefValue, const DWARFUnit &Unit,
const DWARFDebugInfoEntryMinimal &DIE,
CompileUnit *&ReferencedCU);
CompileUnit *getUnitForOffset(unsigned Offset);
bool getDIENames(const DWARFDebugInfoEntryMinimal &Die, DWARFUnit &U,
AttributesInfo &Info);
void reportWarning(const Twine &Warning, const DWARFUnit *Unit = nullptr,
const DWARFDebugInfoEntryMinimal *DIE = nullptr) const;
bool createStreamer(Triple TheTriple, StringRef OutputFilename);
/// \brief Attempt to load a debug object from disk.
ErrorOr<const object::ObjectFile &> loadObject(BinaryHolder &BinaryHolder,
DebugMapObject &Obj,
const DebugMap &Map);
/// @}
private:
std::string OutputFilename;
LinkOptions Options;
BinaryHolder BinHolder;
std::unique_ptr<DwarfStreamer> Streamer;
/// The units of the current debug map object.
std::vector<CompileUnit> Units;
/// The debug map object curently under consideration.
DebugMapObject *CurrentDebugObject;
/// \brief The Dwarf string pool
NonRelocatableStringpool StringPool;
/// \brief This map is keyed by the entry PC of functions in that
/// debug object and the associated value is a pair storing the
/// corresponding end PC and the offset to apply to get the linked
/// address.
///
/// See startDebugObject() for a more complete description of its use.
std::map<uint64_t, std::pair<uint64_t, int64_t>> Ranges;
/// \brief The CIEs that have been emitted in the output
/// section. The actual CIE data serves a the key to this StringMap,
/// this takes care of comparing the semantics of CIEs defined in
/// different object files.
StringMap<uint32_t> EmittedCIEs;
/// Offset of the last CIE that has been emitted in the output
/// debug_frame section.
uint32_t LastCIEOffset;
};
/// \brief Similar to DWARFUnitSection::getUnitForOffset(), but
/// returning our CompileUnit object instead.
CompileUnit *DwarfLinker::getUnitForOffset(unsigned Offset) {
auto CU =
std::upper_bound(Units.begin(), Units.end(), Offset,
[](uint32_t LHS, const CompileUnit &RHS) {
return LHS < RHS.getOrigUnit().getNextUnitOffset();
});
return CU != Units.end() ? &*CU : nullptr;
}
/// \brief Resolve the DIE attribute reference that has been
/// extracted in \p RefValue. The resulting DIE migh be in another
/// CompileUnit which is stored into \p ReferencedCU.
/// \returns null if resolving fails for any reason.
const DWARFDebugInfoEntryMinimal *DwarfLinker::resolveDIEReference(
const DWARFFormValue &RefValue, const DWARFUnit &Unit,
const DWARFDebugInfoEntryMinimal &DIE, CompileUnit *&RefCU) {
assert(RefValue.isFormClass(DWARFFormValue::FC_Reference));
uint64_t RefOffset = *RefValue.getAsReference(&Unit);
if ((RefCU = getUnitForOffset(RefOffset)))
if (const auto *RefDie = RefCU->getOrigUnit().getDIEForOffset(RefOffset))
return RefDie;
reportWarning("could not find referenced DIE", &Unit, &DIE);
return nullptr;
}
/// \returns whether the passed \a Attr type might contain a DIE
/// reference suitable for ODR uniquing.
static bool isODRAttribute(uint16_t Attr) {
switch (Attr) {
default:
return false;
case dwarf::DW_AT_type:
case dwarf::DW_AT_containing_type:
case dwarf::DW_AT_specification:
case dwarf::DW_AT_abstract_origin:
case dwarf::DW_AT_import:
return true;
}
llvm_unreachable("Improper attribute.");
}
/// Set the last DIE/CU a context was seen in and, possibly invalidate
/// the context if it is ambiguous.
///
/// In the current implementation, we don't handle overloaded
/// functions well, because the argument types are not taken into
/// account when computing the DeclContext tree.
///
/// Some of this is mitigated byt using mangled names that do contain
/// the arguments types, but sometimes (eg. with function templates)
/// we don't have that. In that case, just do not unique anything that
/// refers to the contexts we are not able to distinguish.
///
/// If a context that is not a namespace appears twice in the same CU,
/// we know it is ambiguous. Make it invalid.
bool DeclContext::setLastSeenDIE(CompileUnit &U,
const DWARFDebugInfoEntryMinimal *Die) {
if (LastSeenCompileUnitID == U.getUniqueID()) {
DWARFUnit &OrigUnit = U.getOrigUnit();
uint32_t FirstIdx = OrigUnit.getDIEIndex(LastSeenDIE);
U.getInfo(FirstIdx).Ctxt = nullptr;
return false;
}
LastSeenCompileUnitID = U.getUniqueID();
LastSeenDIE = Die;
return true;
}
/// Get the child context of \a Context corresponding to \a DIE.
///
/// \returns the child context or null if we shouldn't track children
/// contexts. It also returns an additional bit meaning 'invalid'. An
/// invalid context means it shouldn't be considered for uniquing, but
/// its not returning null, because some children of that context
/// might be uniquing candidates.
/// FIXME: this is for dsymutil-classic compatibility, I don't think
/// it buys us much.
PointerIntPair<DeclContext *, 1> DeclContextTree::getChildDeclContext(
DeclContext &Context, const DWARFDebugInfoEntryMinimal *DIE, CompileUnit &U,
NonRelocatableStringpool &StringPool) {
unsigned Tag = DIE->getTag();
// FIXME: dsymutil-classic compat: We should bail out here if we
// have a specification or an abstract_origin. We will get the
// parent context wrong here.
switch (Tag) {
default:
// By default stop gathering child contexts.
return PointerIntPair<DeclContext *, 1>(nullptr);
case dwarf::DW_TAG_compile_unit:
// FIXME: Add support for DW_TAG_module.
return PointerIntPair<DeclContext *, 1>(&Context);
case dwarf::DW_TAG_subprogram:
// Do not unique anything inside CU local functions.
if ((Context.getTag() == dwarf::DW_TAG_namespace ||
Context.getTag() == dwarf::DW_TAG_compile_unit) &&
!DIE->getAttributeValueAsUnsignedConstant(&U.getOrigUnit(),
dwarf::DW_AT_external, 0))
return PointerIntPair<DeclContext *, 1>(nullptr);
// Fallthrough
case dwarf::DW_TAG_member:
case dwarf::DW_TAG_namespace:
case dwarf::DW_TAG_structure_type:
case dwarf::DW_TAG_class_type:
case dwarf::DW_TAG_union_type:
case dwarf::DW_TAG_enumeration_type:
case dwarf::DW_TAG_typedef:
// Artificial things might be ambiguous, because they might be
// created on demand. For example implicitely defined constructors
// are ambiguous because of the way we identify contexts, and they
// won't be generated everytime everywhere.
if (DIE->getAttributeValueAsUnsignedConstant(&U.getOrigUnit(),
dwarf::DW_AT_artificial, 0))
return PointerIntPair<DeclContext *, 1>(nullptr);
break;
}
const char *Name = DIE->getName(&U.getOrigUnit(), DINameKind::LinkageName);
const char *ShortName = DIE->getName(&U.getOrigUnit(), DINameKind::ShortName);
StringRef NameRef;
StringRef ShortNameRef;
StringRef FileRef;
if (Name)
NameRef = StringPool.internString(Name);
else if (Tag == dwarf::DW_TAG_namespace)
// FIXME: For dsymutil-classic compatibility. I think uniquing
// within anonymous namespaces is wrong. There is no ODR guarantee
// there.
NameRef = StringPool.internString("(anonymous namespace)");
if (ShortName && ShortName != Name)
ShortNameRef = StringPool.internString(ShortName);
else
ShortNameRef = NameRef;
if (Tag != dwarf::DW_TAG_class_type && Tag != dwarf::DW_TAG_structure_type &&
Tag != dwarf::DW_TAG_union_type &&
Tag != dwarf::DW_TAG_enumeration_type && NameRef.empty())
return PointerIntPair<DeclContext *, 1>(nullptr);
std::string File;
unsigned Line = 0;
unsigned ByteSize = 0;
// Gather some discriminating data about the DeclContext we will be
// creating: File, line number and byte size. This shouldn't be
// necessary, because the ODR is just about names, but given that we
// do some approximations with overloaded functions and anonymous
// namespaces, use these additional data points to make the process safer.
ByteSize = DIE->getAttributeValueAsUnsignedConstant(
&U.getOrigUnit(), dwarf::DW_AT_byte_size, UINT64_MAX);
if (Tag != dwarf::DW_TAG_namespace || !Name) {
if (unsigned FileNum = DIE->getAttributeValueAsUnsignedConstant(
&U.getOrigUnit(), dwarf::DW_AT_decl_file, 0)) {
if (const auto *LT = U.getOrigUnit().getContext().getLineTableForUnit(
&U.getOrigUnit())) {
// FIXME: dsymutil-classic compatibility. I'd rather not
// unique anything in anonymous namespaces, but if we do, then
// verify that the file and line correspond.
if (!Name && Tag == dwarf::DW_TAG_namespace)
FileNum = 1;
// FIXME: Passing U.getOrigUnit().getCompilationDir()
// instead of "" would allow more uniquing, but for now, do
// it this way to match dsymutil-classic.
if (LT->getFileNameByIndex(
FileNum, "",
DILineInfoSpecifier::FileLineInfoKind::AbsoluteFilePath,
File)) {
Line = DIE->getAttributeValueAsUnsignedConstant(
&U.getOrigUnit(), dwarf::DW_AT_decl_line, 0);
#ifdef HAVE_REALPATH
// Cache the resolved paths, because calling realpath is expansive.
if (const char *ResolvedPath = U.getResolvedPath(FileNum)) {
File = ResolvedPath;
} else {
char RealPath[PATH_MAX + 1];
RealPath[PATH_MAX] = 0;
if (::realpath(File.c_str(), RealPath))
File = RealPath;
U.setResolvedPath(FileNum, File);
}
#endif
FileRef = StringPool.internString(File);
}
}
}
}
if (!Line && NameRef.empty())
return PointerIntPair<DeclContext *, 1>(nullptr);
// FIXME: dsymutil-classic compat won't unique the same type
// presented once as a struct and once as a class. Use the Tag in
// the fully qualified name hash to get the same effect.
// We hash NameRef, which is the mangled name, in order to get most
// overloaded functions resolvec correctly.
unsigned Hash = hash_combine(Context.getQualifiedNameHash(), Tag, NameRef);
// FIXME: dsymutil-classic compatibility: when we don't have a name,
// use the filename.
if (Tag == dwarf::DW_TAG_namespace && NameRef == "(anonymous namespace)")
Hash = hash_combine(Hash, FileRef);
// Now look if this context already exists.
DeclContext Key(Hash, Line, ByteSize, Tag, NameRef, FileRef, Context);
auto ContextIter = Contexts.find(&Key);
if (ContextIter == Contexts.end()) {
// The context wasn't found.
bool Inserted;
DeclContext *NewContext =
new (Allocator) DeclContext(Hash, Line, ByteSize, Tag, NameRef, FileRef,
Context, DIE, U.getUniqueID());
std::tie(ContextIter, Inserted) = Contexts.insert(NewContext);
assert(Inserted && "Failed to insert DeclContext");
(void)Inserted;
} else if (Tag != dwarf::DW_TAG_namespace &&
!(*ContextIter)->setLastSeenDIE(U, DIE)) {
// The context was found, but it is ambiguous with another context
// in the same file. Mark it invalid.
return PointerIntPair<DeclContext *, 1>(*ContextIter, /* Invalid= */ 1);
}
assert(ContextIter != Contexts.end());
// FIXME: dsymutil-classic compatibility. Union types aren't
// uniques, but their children might be.
if ((Tag == dwarf::DW_TAG_subprogram &&
Context.getTag() != dwarf::DW_TAG_structure_type &&
Context.getTag() != dwarf::DW_TAG_class_type) ||
(Tag == dwarf::DW_TAG_union_type))
return PointerIntPair<DeclContext *, 1>(*ContextIter, /* Invalid= */ 1);
return PointerIntPair<DeclContext *, 1>(*ContextIter);
}
/// \brief Get the potential name and mangled name for the entity
/// described by \p Die and store them in \Info if they are not
/// already there.
/// \returns is a name was found.
bool DwarfLinker::getDIENames(const DWARFDebugInfoEntryMinimal &Die,
DWARFUnit &U, AttributesInfo &Info) {
// FIXME: a bit wastefull as the first getName might return the
// short name.
if (!Info.MangledName &&
(Info.MangledName = Die.getName(&U, DINameKind::LinkageName)))
Info.MangledNameOffset = StringPool.getStringOffset(Info.MangledName);
if (!Info.Name && (Info.Name = Die.getName(&U, DINameKind::ShortName)))
Info.NameOffset = StringPool.getStringOffset(Info.Name);
return Info.Name || Info.MangledName;
}
/// \brief Report a warning to the user, optionaly including
/// information about a specific \p DIE related to the warning.
void DwarfLinker::reportWarning(const Twine &Warning, const DWARFUnit *Unit,
const DWARFDebugInfoEntryMinimal *DIE) const {
StringRef Context = "<debug map>";
if (CurrentDebugObject)
Context = CurrentDebugObject->getObjectFilename();
warn(Warning, Context);
if (!Options.Verbose || !DIE)
return;
errs() << " in DIE:\n";
DIE->dump(errs(), const_cast<DWARFUnit *>(Unit), 0 /* RecurseDepth */,
6 /* Indent */);
}
bool DwarfLinker::createStreamer(Triple TheTriple, StringRef OutputFilename) {
if (Options.NoOutput)
return true;
Streamer = llvm::make_unique<DwarfStreamer>();
return Streamer->init(TheTriple, OutputFilename);
}
/// \brief Recursive helper to gather the child->parent relationships in the
/// original compile unit.
static void gatherDIEParents(const DWARFDebugInfoEntryMinimal *DIE,
unsigned ParentIdx, CompileUnit &CU,
DeclContext *CurrentDeclContext,
NonRelocatableStringpool &StringPool,
DeclContextTree &Contexts) {
unsigned MyIdx = CU.getOrigUnit().getDIEIndex(DIE);
CompileUnit::DIEInfo &Info = CU.getInfo(MyIdx);
Info.ParentIdx = ParentIdx;
if (CU.hasODR()) {
if (CurrentDeclContext) {
auto PtrInvalidPair = Contexts.getChildDeclContext(*CurrentDeclContext,
DIE, CU, StringPool);
CurrentDeclContext = PtrInvalidPair.getPointer();
Info.Ctxt =
PtrInvalidPair.getInt() ? nullptr : PtrInvalidPair.getPointer();
} else
Info.Ctxt = CurrentDeclContext = nullptr;
}
if (DIE->hasChildren())
for (auto *Child = DIE->getFirstChild(); Child && !Child->isNULL();
Child = Child->getSibling())
gatherDIEParents(Child, MyIdx, CU, CurrentDeclContext, StringPool,
Contexts);
}
static bool dieNeedsChildrenToBeMeaningful(uint32_t Tag) {
switch (Tag) {
default:
return false;
case dwarf::DW_TAG_subprogram:
case dwarf::DW_TAG_lexical_block:
case dwarf::DW_TAG_subroutine_type:
case dwarf::DW_TAG_structure_type:
case dwarf::DW_TAG_class_type:
case dwarf::DW_TAG_union_type:
return true;
}
llvm_unreachable("Invalid Tag");
}
static unsigned getRefAddrSize(const DWARFUnit &U) {
if (U.getVersion() == 2)
return U.getAddressByteSize();
return 4;
}
void DwarfLinker::startDebugObject(DWARFContext &Dwarf, DebugMapObject &Obj) {
Units.reserve(Dwarf.getNumCompileUnits());
NextValidReloc = 0;
// Iterate over the debug map entries and put all the ones that are
// functions (because they have a size) into the Ranges map. This
// map is very similar to the FunctionRanges that are stored in each
// unit, with 2 notable differences:
// - obviously this one is global, while the other ones are per-unit.
// - this one contains not only the functions described in the DIE
// tree, but also the ones that are only in the debug map.
// The latter information is required to reproduce dsymutil's logic
// while linking line tables. The cases where this information
// matters look like bugs that need to be investigated, but for now
// we need to reproduce dsymutil's behavior.
// FIXME: Once we understood exactly if that information is needed,
// maybe totally remove this (or try to use it to do a real
// -gline-tables-only on Darwin.
for (const auto &Entry : Obj.symbols()) {
const auto &Mapping = Entry.getValue();
if (Mapping.Size)
Ranges[Mapping.ObjectAddress] = std::make_pair(
Mapping.ObjectAddress + Mapping.Size,
int64_t(Mapping.BinaryAddress) - Mapping.ObjectAddress);
}
}
void DwarfLinker::endDebugObject() {
Units.clear();
ValidRelocs.clear();
Ranges.clear();
for (auto I = DIEBlocks.begin(), E = DIEBlocks.end(); I != E; ++I)
(*I)->~DIEBlock();
for (auto I = DIELocs.begin(), E = DIELocs.end(); I != E; ++I)
(*I)->~DIELoc();
DIEBlocks.clear();
DIELocs.clear();
DIEAlloc.Reset();
}
/// \brief Iterate over the relocations of the given \p Section and
/// store the ones that correspond to debug map entries into the
/// ValidRelocs array.
void DwarfLinker::findValidRelocsMachO(const object::SectionRef &Section,
const object::MachOObjectFile &Obj,
const DebugMapObject &DMO) {
StringRef Contents;
Section.getContents(Contents);
DataExtractor Data(Contents, Obj.isLittleEndian(), 0);
for (const object::RelocationRef &Reloc : Section.relocations()) {
object::DataRefImpl RelocDataRef = Reloc.getRawDataRefImpl();
MachO::any_relocation_info MachOReloc = Obj.getRelocation(RelocDataRef);
unsigned RelocSize = 1 << Obj.getAnyRelocationLength(MachOReloc);
uint64_t Offset64 = Reloc.getOffset();
if ((RelocSize != 4 && RelocSize != 8)) {
reportWarning(" unsupported relocation in debug_info section.");
continue;
}
uint32_t Offset = Offset64;
// Mach-o uses REL relocations, the addend is at the relocation offset.
uint64_t Addend = Data.getUnsigned(&Offset, RelocSize);
auto Sym = Reloc.getSymbol();
if (Sym != Obj.symbol_end()) {
ErrorOr<StringRef> SymbolName = Sym->getName();
if (!SymbolName) {
reportWarning("error getting relocation symbol name.");
continue;
}
if (const auto *Mapping = DMO.lookupSymbol(*SymbolName))
ValidRelocs.emplace_back(Offset64, RelocSize, Addend, Mapping);
} else if (const auto *Mapping = DMO.lookupObjectAddress(Addend)) {
// Do not store the addend. The addend was the address of the
// symbol in the object file, the address in the binary that is
// stored in the debug map doesn't need to be offseted.
ValidRelocs.emplace_back(Offset64, RelocSize, 0, Mapping);
}
}
}
/// \brief Dispatch the valid relocation finding logic to the
/// appropriate handler depending on the object file format.
bool DwarfLinker::findValidRelocs(const object::SectionRef &Section,
const object::ObjectFile &Obj,
const DebugMapObject &DMO) {
// Dispatch to the right handler depending on the file type.
if (auto *MachOObj = dyn_cast<object::MachOObjectFile>(&Obj))
findValidRelocsMachO(Section, *MachOObj, DMO);
else
reportWarning(Twine("unsupported object file type: ") + Obj.getFileName());
if (ValidRelocs.empty())
return false;
// Sort the relocations by offset. We will walk the DIEs linearly in
// the file, this allows us to just keep an index in the relocation
// array that we advance during our walk, rather than resorting to
// some associative container. See DwarfLinker::NextValidReloc.
std::sort(ValidRelocs.begin(), ValidRelocs.end());
return true;
}
/// \brief Look for relocations in the debug_info section that match
/// entries in the debug map. These relocations will drive the Dwarf
/// link by indicating which DIEs refer to symbols present in the
/// linked binary.
/// \returns wether there are any valid relocations in the debug info.
bool DwarfLinker::findValidRelocsInDebugInfo(const object::ObjectFile &Obj,
const DebugMapObject &DMO) {
// Find the debug_info section.
for (const object::SectionRef &Section : Obj.sections()) {
StringRef SectionName;
Section.getName(SectionName);
SectionName = SectionName.substr(SectionName.find_first_not_of("._"));
if (SectionName != "debug_info")
continue;
return findValidRelocs(Section, Obj, DMO);
}
return false;
}
/// \brief Checks that there is a relocation against an actual debug
/// map entry between \p StartOffset and \p NextOffset.
///
/// This function must be called with offsets in strictly ascending
/// order because it never looks back at relocations it already 'went past'.
/// \returns true and sets Info.InDebugMap if it is the case.
bool DwarfLinker::hasValidRelocation(uint32_t StartOffset, uint32_t EndOffset,
CompileUnit::DIEInfo &Info) {
assert(NextValidReloc == 0 ||
StartOffset > ValidRelocs[NextValidReloc - 1].Offset);
if (NextValidReloc >= ValidRelocs.size())
return false;
uint64_t RelocOffset = ValidRelocs[NextValidReloc].Offset;
// We might need to skip some relocs that we didn't consider. For
// example the high_pc of a discarded DIE might contain a reloc that
// is in the list because it actually corresponds to the start of a
// function that is in the debug map.
while (RelocOffset < StartOffset && NextValidReloc < ValidRelocs.size() - 1)
RelocOffset = ValidRelocs[++NextValidReloc].Offset;
if (RelocOffset < StartOffset || RelocOffset >= EndOffset)
return false;
const auto &ValidReloc = ValidRelocs[NextValidReloc++];
const auto &Mapping = ValidReloc.Mapping->getValue();
if (Options.Verbose)
outs() << "Found valid debug map entry: " << ValidReloc.Mapping->getKey()
<< " " << format("\t%016" PRIx64 " => %016" PRIx64,
uint64_t(Mapping.ObjectAddress),
uint64_t(Mapping.BinaryAddress));
Info.AddrAdjust = int64_t(Mapping.BinaryAddress) + ValidReloc.Addend -
Mapping.ObjectAddress;
Info.InDebugMap = true;
return true;
}
/// \brief Get the starting and ending (exclusive) offset for the
/// attribute with index \p Idx descibed by \p Abbrev. \p Offset is
/// supposed to point to the position of the first attribute described
/// by \p Abbrev.
/// \return [StartOffset, EndOffset) as a pair.
static std::pair<uint32_t, uint32_t>
getAttributeOffsets(const DWARFAbbreviationDeclaration *Abbrev, unsigned Idx,
unsigned Offset, const DWARFUnit &Unit) {
DataExtractor Data = Unit.getDebugInfoExtractor();
for (unsigned i = 0; i < Idx; ++i)
DWARFFormValue::skipValue(Abbrev->getFormByIndex(i), Data, &Offset, &Unit);
uint32_t End = Offset;
DWARFFormValue::skipValue(Abbrev->getFormByIndex(Idx), Data, &End, &Unit);
return std::make_pair(Offset, End);
}
/// \brief Check if a variable describing DIE should be kept.
/// \returns updated TraversalFlags.
unsigned DwarfLinker::shouldKeepVariableDIE(
const DWARFDebugInfoEntryMinimal &DIE, CompileUnit &Unit,
CompileUnit::DIEInfo &MyInfo, unsigned Flags) {
const auto *Abbrev = DIE.getAbbreviationDeclarationPtr();
// Global variables with constant value can always be kept.
if (!(Flags & TF_InFunctionScope) &&
Abbrev->findAttributeIndex(dwarf::DW_AT_const_value) != -1U) {
MyInfo.InDebugMap = true;
return Flags | TF_Keep;
}
uint32_t LocationIdx = Abbrev->findAttributeIndex(dwarf::DW_AT_location);
if (LocationIdx == -1U)
return Flags;
uint32_t Offset = DIE.getOffset() + getULEB128Size(Abbrev->getCode());
const DWARFUnit &OrigUnit = Unit.getOrigUnit();
uint32_t LocationOffset, LocationEndOffset;
std::tie(LocationOffset, LocationEndOffset) =
getAttributeOffsets(Abbrev, LocationIdx, Offset, OrigUnit);
// See if there is a relocation to a valid debug map entry inside
// this variable's location. The order is important here. We want to
// always check in the variable has a valid relocation, so that the
// DIEInfo is filled. However, we don't want a static variable in a
// function to force us to keep the enclosing function.
if (!hasValidRelocation(LocationOffset, LocationEndOffset, MyInfo) ||
(Flags & TF_InFunctionScope))
return Flags;
if (Options.Verbose)
DIE.dump(outs(), const_cast<DWARFUnit *>(&OrigUnit), 0, 8 /* Indent */);
return Flags | TF_Keep;
}
/// \brief Check if a function describing DIE should be kept.
/// \returns updated TraversalFlags.
unsigned DwarfLinker::shouldKeepSubprogramDIE(
const DWARFDebugInfoEntryMinimal &DIE, CompileUnit &Unit,
CompileUnit::DIEInfo &MyInfo, unsigned Flags) {
const auto *Abbrev = DIE.getAbbreviationDeclarationPtr();
Flags |= TF_InFunctionScope;
uint32_t LowPcIdx = Abbrev->findAttributeIndex(dwarf::DW_AT_low_pc);
if (LowPcIdx == -1U)
return Flags;
uint32_t Offset = DIE.getOffset() + getULEB128Size(Abbrev->getCode());
const DWARFUnit &OrigUnit = Unit.getOrigUnit();
uint32_t LowPcOffset, LowPcEndOffset;
std::tie(LowPcOffset, LowPcEndOffset) =
getAttributeOffsets(Abbrev, LowPcIdx, Offset, OrigUnit);
uint64_t LowPc =
DIE.getAttributeValueAsAddress(&OrigUnit, dwarf::DW_AT_low_pc, -1ULL);
assert(LowPc != -1ULL && "low_pc attribute is not an address.");
if (LowPc == -1ULL ||
!hasValidRelocation(LowPcOffset, LowPcEndOffset, MyInfo))
return Flags;
if (Options.Verbose)
DIE.dump(outs(), const_cast<DWARFUnit *>(&OrigUnit), 0, 8 /* Indent */);
Flags |= TF_Keep;
DWARFFormValue HighPcValue;
if (!DIE.getAttributeValue(&OrigUnit, dwarf::DW_AT_high_pc, HighPcValue)) {
reportWarning("Function without high_pc. Range will be discarded.\n",
&OrigUnit, &DIE);
return Flags;
}
uint64_t HighPc;
if (HighPcValue.isFormClass(DWARFFormValue::FC_Address)) {
HighPc = *HighPcValue.getAsAddress(&OrigUnit);
} else {
assert(HighPcValue.isFormClass(DWARFFormValue::FC_Constant));
HighPc = LowPc + *HighPcValue.getAsUnsignedConstant();
}
// Replace the debug map range with a more accurate one.
Ranges[LowPc] = std::make_pair(HighPc, MyInfo.AddrAdjust);
Unit.addFunctionRange(LowPc, HighPc, MyInfo.AddrAdjust);
return Flags;
}
/// \brief Check if a DIE should be kept.
/// \returns updated TraversalFlags.
unsigned DwarfLinker::shouldKeepDIE(const DWARFDebugInfoEntryMinimal &DIE,
CompileUnit &Unit,
CompileUnit::DIEInfo &MyInfo,
unsigned Flags) {
switch (DIE.getTag()) {
case dwarf::DW_TAG_constant:
case dwarf::DW_TAG_variable:
return shouldKeepVariableDIE(DIE, Unit, MyInfo, Flags);
case dwarf::DW_TAG_subprogram:
return shouldKeepSubprogramDIE(DIE, Unit, MyInfo, Flags);
case dwarf::DW_TAG_module:
case dwarf::DW_TAG_imported_module:
case dwarf::DW_TAG_imported_declaration:
case dwarf::DW_TAG_imported_unit:
// We always want to keep these.
return Flags | TF_Keep;
}
return Flags;
}
/// \brief Mark the passed DIE as well as all the ones it depends on
/// as kept.
///
/// This function is called by lookForDIEsToKeep on DIEs that are
/// newly discovered to be needed in the link. It recursively calls
/// back to lookForDIEsToKeep while adding TF_DependencyWalk to the
/// TraversalFlags to inform it that it's not doing the primary DIE
/// tree walk.
void DwarfLinker::keepDIEAndDenpendencies(const DWARFDebugInfoEntryMinimal &DIE,
CompileUnit::DIEInfo &MyInfo,
const DebugMapObject &DMO,
CompileUnit &CU, bool UseODR) {
const DWARFUnit &Unit = CU.getOrigUnit();
MyInfo.Keep = true;
// First mark all the parent chain as kept.
unsigned AncestorIdx = MyInfo.ParentIdx;
while (!CU.getInfo(AncestorIdx).Keep) {
unsigned ODRFlag = UseODR ? TF_ODR : 0;
lookForDIEsToKeep(*Unit.getDIEAtIndex(AncestorIdx), DMO, CU,
TF_ParentWalk | TF_Keep | TF_DependencyWalk | ODRFlag);
AncestorIdx = CU.getInfo(AncestorIdx).ParentIdx;
}
// Then we need to mark all the DIEs referenced by this DIE's
// attributes as kept.
DataExtractor Data = Unit.getDebugInfoExtractor();
const auto *Abbrev = DIE.getAbbreviationDeclarationPtr();
uint32_t Offset = DIE.getOffset() + getULEB128Size(Abbrev->getCode());
// Mark all DIEs referenced through atttributes as kept.
for (const auto &AttrSpec : Abbrev->attributes()) {
DWARFFormValue Val(AttrSpec.Form);
if (!Val.isFormClass(DWARFFormValue::FC_Reference)) {
DWARFFormValue::skipValue(AttrSpec.Form, Data, &Offset, &Unit);
continue;
}
Val.extractValue(Data, &Offset, &Unit);
CompileUnit *ReferencedCU;
if (const auto *RefDIE =
resolveDIEReference(Val, Unit, DIE, ReferencedCU)) {
uint32_t RefIdx = ReferencedCU->getOrigUnit().getDIEIndex(RefDIE);
CompileUnit::DIEInfo &Info = ReferencedCU->getInfo(RefIdx);
// If the referenced DIE has a DeclContext that has already been
// emitted, then do not keep the one in this CU. We'll link to
// the canonical DIE in cloneDieReferenceAttribute.
// FIXME: compatibility with dsymutil-classic. UseODR shouldn't
// be necessary and could be advantageously replaced by
// ReferencedCU->hasODR() && CU.hasODR().
// FIXME: compatibility with dsymutil-classic. There is no
// reason not to unique ref_addr references.
if (AttrSpec.Form != dwarf::DW_FORM_ref_addr && UseODR && Info.Ctxt &&
Info.Ctxt != ReferencedCU->getInfo(Info.ParentIdx).Ctxt &&
Info.Ctxt->getCanonicalDIEOffset() && isODRAttribute(AttrSpec.Attr))
continue;
unsigned ODRFlag = UseODR ? TF_ODR : 0;
lookForDIEsToKeep(*RefDIE, DMO, *ReferencedCU,
TF_Keep | TF_DependencyWalk | ODRFlag);
}
}
}
/// \brief Recursively walk the \p DIE tree and look for DIEs to
/// keep. Store that information in \p CU's DIEInfo.
///
/// This function is the entry point of the DIE selection
/// algorithm. It is expected to walk the DIE tree in file order and
/// (though the mediation of its helper) call hasValidRelocation() on
/// each DIE that might be a 'root DIE' (See DwarfLinker class
/// comment).
/// While walking the dependencies of root DIEs, this function is
/// also called, but during these dependency walks the file order is
/// not respected. The TF_DependencyWalk flag tells us which kind of
/// traversal we are currently doing.
void DwarfLinker::lookForDIEsToKeep(const DWARFDebugInfoEntryMinimal &DIE,
const DebugMapObject &DMO, CompileUnit &CU,
unsigned Flags) {
unsigned Idx = CU.getOrigUnit().getDIEIndex(&DIE);
CompileUnit::DIEInfo &MyInfo = CU.getInfo(Idx);
bool AlreadyKept = MyInfo.Keep;
// If the Keep flag is set, we are marking a required DIE's
// dependencies. If our target is already marked as kept, we're all
// set.
if ((Flags & TF_DependencyWalk) && AlreadyKept)
return;
// We must not call shouldKeepDIE while called from keepDIEAndDenpendencies,
// because it would screw up the relocation finding logic.
if (!(Flags & TF_DependencyWalk))
Flags = shouldKeepDIE(DIE, CU, MyInfo, Flags);
// If it is a newly kept DIE mark it as well as all its dependencies as kept.
if (!AlreadyKept && (Flags & TF_Keep)) {
bool UseOdr = (Flags & TF_DependencyWalk) ? (Flags & TF_ODR) : CU.hasODR();
keepDIEAndDenpendencies(DIE, MyInfo, DMO, CU, UseOdr);
}
// The TF_ParentWalk flag tells us that we are currently walking up
// the parent chain of a required DIE, and we don't want to mark all
// the children of the parents as kept (consider for example a
// DW_TAG_namespace node in the parent chain). There are however a
// set of DIE types for which we want to ignore that directive and still
// walk their children.
if (dieNeedsChildrenToBeMeaningful(DIE.getTag()))
Flags &= ~TF_ParentWalk;
if (!DIE.hasChildren() || (Flags & TF_ParentWalk))
return;
for (auto *Child = DIE.getFirstChild(); Child && !Child->isNULL();
Child = Child->getSibling())
lookForDIEsToKeep(*Child, DMO, CU, Flags);
}
/// \brief Assign an abbreviation numer to \p Abbrev.
///
/// Our DIEs get freed after every DebugMapObject has been processed,
/// thus the FoldingSet we use to unique DIEAbbrevs cannot refer to
/// the instances hold by the DIEs. When we encounter an abbreviation
/// that we don't know, we create a permanent copy of it.
void DwarfLinker::AssignAbbrev(DIEAbbrev &Abbrev) {
// Check the set for priors.
FoldingSetNodeID ID;
Abbrev.Profile(ID);
void *InsertToken;
DIEAbbrev *InSet = AbbreviationsSet.FindNodeOrInsertPos(ID, InsertToken);
// If it's newly added.
if (InSet) {
// Assign existing abbreviation number.
Abbrev.setNumber(InSet->getNumber());
} else {
// Add to abbreviation list.
Abbreviations.push_back(
new DIEAbbrev(Abbrev.getTag(), Abbrev.hasChildren()));
for (const auto &Attr : Abbrev.getData())
Abbreviations.back()->AddAttribute(Attr.getAttribute(), Attr.getForm());
AbbreviationsSet.InsertNode(Abbreviations.back(), InsertToken);
// Assign the unique abbreviation number.
Abbrev.setNumber(Abbreviations.size());
Abbreviations.back()->setNumber(Abbreviations.size());
}
}
/// \brief Clone a string attribute described by \p AttrSpec and add
/// it to \p Die.
/// \returns the size of the new attribute.
unsigned DwarfLinker::cloneStringAttribute(DIE &Die, AttributeSpec AttrSpec,
const DWARFFormValue &Val,
const DWARFUnit &U) {
// Switch everything to out of line strings.
const char *String = *Val.getAsCString(&U);
unsigned Offset = StringPool.getStringOffset(String);
Die.addValue(DIEAlloc, dwarf::Attribute(AttrSpec.Attr), dwarf::DW_FORM_strp,
Reapply "AsmPrinter: Change DIEValue to be stored by value" This reverts commit r238350, effectively reapplying r238349 after fixing (all?) the problems, all somehow related to how I was using `AlignedArrayCharUnion<>` inside `DIEValue`: - MSVC can only handle `sizeof()` on types, not values. Change the assert. - GCC doesn't know the `is_trivially_copyable` type trait. Instead of asserting it, add destructors. - Call placement new even when constructing POD (i.e., the pointers). - Instead of copying the char buffer, copy the casted classes. I've left in a couple of `static_assert`s that I think both MSVC and GCC know how to handle. If the bots disagree with me, I'll remove them. - Check that the constructed type is either standard layout or a pointer. This protects against a programming error: we really want the "small" `DIEValue`s to be small and simple, so don't accidentally change them not to be. - Similarly, check that the size of the buffer is no bigger than a `uint64_t` or a pointer. (I thought checking against `sizeof(uint64_t)` would be good enough, but Chandler suggested that pointers might sometimes be bigger than that in the context of sanitizers.) I've also committed r238359 in the meantime, which introduces a DIEValue.def to simplify dispatching between the various types (thanks to a review comment by David Blaikie). Without that, this commit would be almost unintelligible. Here's the original commit message: -- Change `DIEValue` to be stored/passed/etc. by value, instead of reference. It's now a discriminated union, with a `Val` field storing the actual type. The classes that used to inherit from `DIEValue` no longer do. There are two categories of these: - Small values fit in a single pointer and are stored by value. - Large values require auxiliary storage, and are stored by reference. The only non-mechanical change is to tools/dsymutil/DwarfLinker.cpp. It was relying on `DIEInteger`s being passed around by reference, so I replaced that assumption with a `PatchLocation` type that stores a safe reference to where the `DIEInteger` lives instead. This commit causes a temporary regression in memory usage, since I've left merging `DIEAbbrevData` into `DIEValue` for a follow-up commit. I measured an increase from 845 MB to 879 MB, around 3.9%. The follow-up drops it lower than the starting point, and I've only recently brought the memory this low anyway, so I'm committing these changes separately to keep them incremental. (I also considered swapping the commits, but the other one first would cause a lot more code churn.) (I'm looking at `llc` memory usage on `verify-uselistorder.lto.opt.bc`; see r236629 for details.) -- llvm-svn: 238362
2015-05-28 06:14:58 +08:00
DIEInteger(Offset));
return 4;
}
/// \brief Clone an attribute referencing another DIE and add
/// it to \p Die.
/// \returns the size of the new attribute.
unsigned DwarfLinker::cloneDieReferenceAttribute(
DIE &Die, const DWARFDebugInfoEntryMinimal &InputDIE,
AttributeSpec AttrSpec, unsigned AttrSize, const DWARFFormValue &Val,
CompileUnit &Unit) {
const DWARFUnit &U = Unit.getOrigUnit();
uint32_t Ref = *Val.getAsReference(&U);
DIE *NewRefDie = nullptr;
CompileUnit *RefUnit = nullptr;
DeclContext *Ctxt = nullptr;
const DWARFDebugInfoEntryMinimal *RefDie =
resolveDIEReference(Val, U, InputDIE, RefUnit);
// If the referenced DIE is not found, drop the attribute.
if (!RefDie)
return 0;
unsigned Idx = RefUnit->getOrigUnit().getDIEIndex(RefDie);
CompileUnit::DIEInfo &RefInfo = RefUnit->getInfo(Idx);
// If we already have emitted an equivalent DeclContext, just point
// at it.
if (isODRAttribute(AttrSpec.Attr)) {
Ctxt = RefInfo.Ctxt;
if (Ctxt && Ctxt->getCanonicalDIEOffset()) {
DIEInteger Attr(Ctxt->getCanonicalDIEOffset());
Die.addValue(DIEAlloc, dwarf::Attribute(AttrSpec.Attr),
dwarf::DW_FORM_ref_addr, Attr);
return getRefAddrSize(U);
}
}
if (!RefInfo.Clone) {
assert(Ref > InputDIE.getOffset());
// We haven't cloned this DIE yet. Just create an empty one and
// store it. It'll get really cloned when we process it.
RefInfo.Clone = DIE::get(DIEAlloc, dwarf::Tag(RefDie->getTag()));
}
NewRefDie = RefInfo.Clone;
if (AttrSpec.Form == dwarf::DW_FORM_ref_addr ||
(Unit.hasODR() && isODRAttribute(AttrSpec.Attr))) {
// We cannot currently rely on a DIEEntry to emit ref_addr
// references, because the implementation calls back to DwarfDebug
// to find the unit offset. (We don't have a DwarfDebug)
// FIXME: we should be able to design DIEEntry reliance on
// DwarfDebug away.
Reapply "AsmPrinter: Change DIEValue to be stored by value" This reverts commit r238350, effectively reapplying r238349 after fixing (all?) the problems, all somehow related to how I was using `AlignedArrayCharUnion<>` inside `DIEValue`: - MSVC can only handle `sizeof()` on types, not values. Change the assert. - GCC doesn't know the `is_trivially_copyable` type trait. Instead of asserting it, add destructors. - Call placement new even when constructing POD (i.e., the pointers). - Instead of copying the char buffer, copy the casted classes. I've left in a couple of `static_assert`s that I think both MSVC and GCC know how to handle. If the bots disagree with me, I'll remove them. - Check that the constructed type is either standard layout or a pointer. This protects against a programming error: we really want the "small" `DIEValue`s to be small and simple, so don't accidentally change them not to be. - Similarly, check that the size of the buffer is no bigger than a `uint64_t` or a pointer. (I thought checking against `sizeof(uint64_t)` would be good enough, but Chandler suggested that pointers might sometimes be bigger than that in the context of sanitizers.) I've also committed r238359 in the meantime, which introduces a DIEValue.def to simplify dispatching between the various types (thanks to a review comment by David Blaikie). Without that, this commit would be almost unintelligible. Here's the original commit message: -- Change `DIEValue` to be stored/passed/etc. by value, instead of reference. It's now a discriminated union, with a `Val` field storing the actual type. The classes that used to inherit from `DIEValue` no longer do. There are two categories of these: - Small values fit in a single pointer and are stored by value. - Large values require auxiliary storage, and are stored by reference. The only non-mechanical change is to tools/dsymutil/DwarfLinker.cpp. It was relying on `DIEInteger`s being passed around by reference, so I replaced that assumption with a `PatchLocation` type that stores a safe reference to where the `DIEInteger` lives instead. This commit causes a temporary regression in memory usage, since I've left merging `DIEAbbrevData` into `DIEValue` for a follow-up commit. I measured an increase from 845 MB to 879 MB, around 3.9%. The follow-up drops it lower than the starting point, and I've only recently brought the memory this low anyway, so I'm committing these changes separately to keep them incremental. (I also considered swapping the commits, but the other one first would cause a lot more code churn.) (I'm looking at `llc` memory usage on `verify-uselistorder.lto.opt.bc`; see r236629 for details.) -- llvm-svn: 238362
2015-05-28 06:14:58 +08:00
uint64_t Attr;
if (Ref < InputDIE.getOffset()) {
// We must have already cloned that DIE.
uint32_t NewRefOffset =
RefUnit->getStartOffset() + NewRefDie->getOffset();
Reapply "AsmPrinter: Change DIEValue to be stored by value" This reverts commit r238350, effectively reapplying r238349 after fixing (all?) the problems, all somehow related to how I was using `AlignedArrayCharUnion<>` inside `DIEValue`: - MSVC can only handle `sizeof()` on types, not values. Change the assert. - GCC doesn't know the `is_trivially_copyable` type trait. Instead of asserting it, add destructors. - Call placement new even when constructing POD (i.e., the pointers). - Instead of copying the char buffer, copy the casted classes. I've left in a couple of `static_assert`s that I think both MSVC and GCC know how to handle. If the bots disagree with me, I'll remove them. - Check that the constructed type is either standard layout or a pointer. This protects against a programming error: we really want the "small" `DIEValue`s to be small and simple, so don't accidentally change them not to be. - Similarly, check that the size of the buffer is no bigger than a `uint64_t` or a pointer. (I thought checking against `sizeof(uint64_t)` would be good enough, but Chandler suggested that pointers might sometimes be bigger than that in the context of sanitizers.) I've also committed r238359 in the meantime, which introduces a DIEValue.def to simplify dispatching between the various types (thanks to a review comment by David Blaikie). Without that, this commit would be almost unintelligible. Here's the original commit message: -- Change `DIEValue` to be stored/passed/etc. by value, instead of reference. It's now a discriminated union, with a `Val` field storing the actual type. The classes that used to inherit from `DIEValue` no longer do. There are two categories of these: - Small values fit in a single pointer and are stored by value. - Large values require auxiliary storage, and are stored by reference. The only non-mechanical change is to tools/dsymutil/DwarfLinker.cpp. It was relying on `DIEInteger`s being passed around by reference, so I replaced that assumption with a `PatchLocation` type that stores a safe reference to where the `DIEInteger` lives instead. This commit causes a temporary regression in memory usage, since I've left merging `DIEAbbrevData` into `DIEValue` for a follow-up commit. I measured an increase from 845 MB to 879 MB, around 3.9%. The follow-up drops it lower than the starting point, and I've only recently brought the memory this low anyway, so I'm committing these changes separately to keep them incremental. (I also considered swapping the commits, but the other one first would cause a lot more code churn.) (I'm looking at `llc` memory usage on `verify-uselistorder.lto.opt.bc`; see r236629 for details.) -- llvm-svn: 238362
2015-05-28 06:14:58 +08:00
Attr = NewRefOffset;
Die.addValue(DIEAlloc, dwarf::Attribute(AttrSpec.Attr),
dwarf::DW_FORM_ref_addr, DIEInteger(Attr));
} else {
// A forward reference. Note and fixup later.
Reapply "AsmPrinter: Change DIEValue to be stored by value" This reverts commit r238350, effectively reapplying r238349 after fixing (all?) the problems, all somehow related to how I was using `AlignedArrayCharUnion<>` inside `DIEValue`: - MSVC can only handle `sizeof()` on types, not values. Change the assert. - GCC doesn't know the `is_trivially_copyable` type trait. Instead of asserting it, add destructors. - Call placement new even when constructing POD (i.e., the pointers). - Instead of copying the char buffer, copy the casted classes. I've left in a couple of `static_assert`s that I think both MSVC and GCC know how to handle. If the bots disagree with me, I'll remove them. - Check that the constructed type is either standard layout or a pointer. This protects against a programming error: we really want the "small" `DIEValue`s to be small and simple, so don't accidentally change them not to be. - Similarly, check that the size of the buffer is no bigger than a `uint64_t` or a pointer. (I thought checking against `sizeof(uint64_t)` would be good enough, but Chandler suggested that pointers might sometimes be bigger than that in the context of sanitizers.) I've also committed r238359 in the meantime, which introduces a DIEValue.def to simplify dispatching between the various types (thanks to a review comment by David Blaikie). Without that, this commit would be almost unintelligible. Here's the original commit message: -- Change `DIEValue` to be stored/passed/etc. by value, instead of reference. It's now a discriminated union, with a `Val` field storing the actual type. The classes that used to inherit from `DIEValue` no longer do. There are two categories of these: - Small values fit in a single pointer and are stored by value. - Large values require auxiliary storage, and are stored by reference. The only non-mechanical change is to tools/dsymutil/DwarfLinker.cpp. It was relying on `DIEInteger`s being passed around by reference, so I replaced that assumption with a `PatchLocation` type that stores a safe reference to where the `DIEInteger` lives instead. This commit causes a temporary regression in memory usage, since I've left merging `DIEAbbrevData` into `DIEValue` for a follow-up commit. I measured an increase from 845 MB to 879 MB, around 3.9%. The follow-up drops it lower than the starting point, and I've only recently brought the memory this low anyway, so I'm committing these changes separately to keep them incremental. (I also considered swapping the commits, but the other one first would cause a lot more code churn.) (I'm looking at `llc` memory usage on `verify-uselistorder.lto.opt.bc`; see r236629 for details.) -- llvm-svn: 238362
2015-05-28 06:14:58 +08:00
Attr = 0xBADDEF;
Unit.noteForwardReference(
NewRefDie, RefUnit, Ctxt,
Die.addValue(DIEAlloc, dwarf::Attribute(AttrSpec.Attr),
dwarf::DW_FORM_ref_addr, DIEInteger(Attr)));
}
return getRefAddrSize(U);
}
Die.addValue(DIEAlloc, dwarf::Attribute(AttrSpec.Attr),
dwarf::Form(AttrSpec.Form), DIEEntry(*NewRefDie));
return AttrSize;
}
/// \brief Clone an attribute of block form (locations, constants) and add
/// it to \p Die.
/// \returns the size of the new attribute.
unsigned DwarfLinker::cloneBlockAttribute(DIE &Die, AttributeSpec AttrSpec,
const DWARFFormValue &Val,
unsigned AttrSize) {
DIEValueList *Attr;
Reapply "AsmPrinter: Change DIEValue to be stored by value" This reverts commit r238350, effectively reapplying r238349 after fixing (all?) the problems, all somehow related to how I was using `AlignedArrayCharUnion<>` inside `DIEValue`: - MSVC can only handle `sizeof()` on types, not values. Change the assert. - GCC doesn't know the `is_trivially_copyable` type trait. Instead of asserting it, add destructors. - Call placement new even when constructing POD (i.e., the pointers). - Instead of copying the char buffer, copy the casted classes. I've left in a couple of `static_assert`s that I think both MSVC and GCC know how to handle. If the bots disagree with me, I'll remove them. - Check that the constructed type is either standard layout or a pointer. This protects against a programming error: we really want the "small" `DIEValue`s to be small and simple, so don't accidentally change them not to be. - Similarly, check that the size of the buffer is no bigger than a `uint64_t` or a pointer. (I thought checking against `sizeof(uint64_t)` would be good enough, but Chandler suggested that pointers might sometimes be bigger than that in the context of sanitizers.) I've also committed r238359 in the meantime, which introduces a DIEValue.def to simplify dispatching between the various types (thanks to a review comment by David Blaikie). Without that, this commit would be almost unintelligible. Here's the original commit message: -- Change `DIEValue` to be stored/passed/etc. by value, instead of reference. It's now a discriminated union, with a `Val` field storing the actual type. The classes that used to inherit from `DIEValue` no longer do. There are two categories of these: - Small values fit in a single pointer and are stored by value. - Large values require auxiliary storage, and are stored by reference. The only non-mechanical change is to tools/dsymutil/DwarfLinker.cpp. It was relying on `DIEInteger`s being passed around by reference, so I replaced that assumption with a `PatchLocation` type that stores a safe reference to where the `DIEInteger` lives instead. This commit causes a temporary regression in memory usage, since I've left merging `DIEAbbrevData` into `DIEValue` for a follow-up commit. I measured an increase from 845 MB to 879 MB, around 3.9%. The follow-up drops it lower than the starting point, and I've only recently brought the memory this low anyway, so I'm committing these changes separately to keep them incremental. (I also considered swapping the commits, but the other one first would cause a lot more code churn.) (I'm looking at `llc` memory usage on `verify-uselistorder.lto.opt.bc`; see r236629 for details.) -- llvm-svn: 238362
2015-05-28 06:14:58 +08:00
DIEValue Value;
DIELoc *Loc = nullptr;
DIEBlock *Block = nullptr;
// Just copy the block data over.
if (AttrSpec.Form == dwarf::DW_FORM_exprloc) {
Reapply "AsmPrinter: Change DIEValue to be stored by value" This reverts commit r238350, effectively reapplying r238349 after fixing (all?) the problems, all somehow related to how I was using `AlignedArrayCharUnion<>` inside `DIEValue`: - MSVC can only handle `sizeof()` on types, not values. Change the assert. - GCC doesn't know the `is_trivially_copyable` type trait. Instead of asserting it, add destructors. - Call placement new even when constructing POD (i.e., the pointers). - Instead of copying the char buffer, copy the casted classes. I've left in a couple of `static_assert`s that I think both MSVC and GCC know how to handle. If the bots disagree with me, I'll remove them. - Check that the constructed type is either standard layout or a pointer. This protects against a programming error: we really want the "small" `DIEValue`s to be small and simple, so don't accidentally change them not to be. - Similarly, check that the size of the buffer is no bigger than a `uint64_t` or a pointer. (I thought checking against `sizeof(uint64_t)` would be good enough, but Chandler suggested that pointers might sometimes be bigger than that in the context of sanitizers.) I've also committed r238359 in the meantime, which introduces a DIEValue.def to simplify dispatching between the various types (thanks to a review comment by David Blaikie). Without that, this commit would be almost unintelligible. Here's the original commit message: -- Change `DIEValue` to be stored/passed/etc. by value, instead of reference. It's now a discriminated union, with a `Val` field storing the actual type. The classes that used to inherit from `DIEValue` no longer do. There are two categories of these: - Small values fit in a single pointer and are stored by value. - Large values require auxiliary storage, and are stored by reference. The only non-mechanical change is to tools/dsymutil/DwarfLinker.cpp. It was relying on `DIEInteger`s being passed around by reference, so I replaced that assumption with a `PatchLocation` type that stores a safe reference to where the `DIEInteger` lives instead. This commit causes a temporary regression in memory usage, since I've left merging `DIEAbbrevData` into `DIEValue` for a follow-up commit. I measured an increase from 845 MB to 879 MB, around 3.9%. The follow-up drops it lower than the starting point, and I've only recently brought the memory this low anyway, so I'm committing these changes separately to keep them incremental. (I also considered swapping the commits, but the other one first would cause a lot more code churn.) (I'm looking at `llc` memory usage on `verify-uselistorder.lto.opt.bc`; see r236629 for details.) -- llvm-svn: 238362
2015-05-28 06:14:58 +08:00
Loc = new (DIEAlloc) DIELoc;
DIELocs.push_back(Loc);
} else {
Reapply "AsmPrinter: Change DIEValue to be stored by value" This reverts commit r238350, effectively reapplying r238349 after fixing (all?) the problems, all somehow related to how I was using `AlignedArrayCharUnion<>` inside `DIEValue`: - MSVC can only handle `sizeof()` on types, not values. Change the assert. - GCC doesn't know the `is_trivially_copyable` type trait. Instead of asserting it, add destructors. - Call placement new even when constructing POD (i.e., the pointers). - Instead of copying the char buffer, copy the casted classes. I've left in a couple of `static_assert`s that I think both MSVC and GCC know how to handle. If the bots disagree with me, I'll remove them. - Check that the constructed type is either standard layout or a pointer. This protects against a programming error: we really want the "small" `DIEValue`s to be small and simple, so don't accidentally change them not to be. - Similarly, check that the size of the buffer is no bigger than a `uint64_t` or a pointer. (I thought checking against `sizeof(uint64_t)` would be good enough, but Chandler suggested that pointers might sometimes be bigger than that in the context of sanitizers.) I've also committed r238359 in the meantime, which introduces a DIEValue.def to simplify dispatching between the various types (thanks to a review comment by David Blaikie). Without that, this commit would be almost unintelligible. Here's the original commit message: -- Change `DIEValue` to be stored/passed/etc. by value, instead of reference. It's now a discriminated union, with a `Val` field storing the actual type. The classes that used to inherit from `DIEValue` no longer do. There are two categories of these: - Small values fit in a single pointer and are stored by value. - Large values require auxiliary storage, and are stored by reference. The only non-mechanical change is to tools/dsymutil/DwarfLinker.cpp. It was relying on `DIEInteger`s being passed around by reference, so I replaced that assumption with a `PatchLocation` type that stores a safe reference to where the `DIEInteger` lives instead. This commit causes a temporary regression in memory usage, since I've left merging `DIEAbbrevData` into `DIEValue` for a follow-up commit. I measured an increase from 845 MB to 879 MB, around 3.9%. The follow-up drops it lower than the starting point, and I've only recently brought the memory this low anyway, so I'm committing these changes separately to keep them incremental. (I also considered swapping the commits, but the other one first would cause a lot more code churn.) (I'm looking at `llc` memory usage on `verify-uselistorder.lto.opt.bc`; see r236629 for details.) -- llvm-svn: 238362
2015-05-28 06:14:58 +08:00
Block = new (DIEAlloc) DIEBlock;
DIEBlocks.push_back(Block);
}
Attr = Loc ? static_cast<DIEValueList *>(Loc)
: static_cast<DIEValueList *>(Block);
if (Loc)
Value = DIEValue(dwarf::Attribute(AttrSpec.Attr),
dwarf::Form(AttrSpec.Form), Loc);
else
Value = DIEValue(dwarf::Attribute(AttrSpec.Attr),
dwarf::Form(AttrSpec.Form), Block);
ArrayRef<uint8_t> Bytes = *Val.getAsBlock();
for (auto Byte : Bytes)
Attr->addValue(DIEAlloc, static_cast<dwarf::Attribute>(0),
dwarf::DW_FORM_data1, DIEInteger(Byte));
// FIXME: If DIEBlock and DIELoc just reuses the Size field of
// the DIE class, this if could be replaced by
// Attr->setSize(Bytes.size()).
if (Streamer) {
if (Loc)
Loc->ComputeSize(&Streamer->getAsmPrinter());
else
Block->ComputeSize(&Streamer->getAsmPrinter());
}
Die.addValue(DIEAlloc, Value);
return AttrSize;
}
/// \brief Clone an address attribute and add it to \p Die.
/// \returns the size of the new attribute.
unsigned DwarfLinker::cloneAddressAttribute(DIE &Die, AttributeSpec AttrSpec,
const DWARFFormValue &Val,
const CompileUnit &Unit,
AttributesInfo &Info) {
uint64_t Addr = *Val.getAsAddress(&Unit.getOrigUnit());
if (AttrSpec.Attr == dwarf::DW_AT_low_pc) {
if (Die.getTag() == dwarf::DW_TAG_inlined_subroutine ||
Die.getTag() == dwarf::DW_TAG_lexical_block)
Addr += Info.PCOffset;
else if (Die.getTag() == dwarf::DW_TAG_compile_unit) {
Addr = Unit.getLowPc();
if (Addr == UINT64_MAX)
return 0;
}
Info.HasLowPc = true;
} else if (AttrSpec.Attr == dwarf::DW_AT_high_pc) {
if (Die.getTag() == dwarf::DW_TAG_compile_unit) {
if (uint64_t HighPc = Unit.getHighPc())
Addr = HighPc;
else
return 0;
} else
// If we have a high_pc recorded for the input DIE, use
// it. Otherwise (when no relocations where applied) just use the
// one we just decoded.
Addr = (Info.OrigHighPc ? Info.OrigHighPc : Addr) + Info.PCOffset;
}
Die.addValue(DIEAlloc, static_cast<dwarf::Attribute>(AttrSpec.Attr),
Reapply "AsmPrinter: Change DIEValue to be stored by value" This reverts commit r238350, effectively reapplying r238349 after fixing (all?) the problems, all somehow related to how I was using `AlignedArrayCharUnion<>` inside `DIEValue`: - MSVC can only handle `sizeof()` on types, not values. Change the assert. - GCC doesn't know the `is_trivially_copyable` type trait. Instead of asserting it, add destructors. - Call placement new even when constructing POD (i.e., the pointers). - Instead of copying the char buffer, copy the casted classes. I've left in a couple of `static_assert`s that I think both MSVC and GCC know how to handle. If the bots disagree with me, I'll remove them. - Check that the constructed type is either standard layout or a pointer. This protects against a programming error: we really want the "small" `DIEValue`s to be small and simple, so don't accidentally change them not to be. - Similarly, check that the size of the buffer is no bigger than a `uint64_t` or a pointer. (I thought checking against `sizeof(uint64_t)` would be good enough, but Chandler suggested that pointers might sometimes be bigger than that in the context of sanitizers.) I've also committed r238359 in the meantime, which introduces a DIEValue.def to simplify dispatching between the various types (thanks to a review comment by David Blaikie). Without that, this commit would be almost unintelligible. Here's the original commit message: -- Change `DIEValue` to be stored/passed/etc. by value, instead of reference. It's now a discriminated union, with a `Val` field storing the actual type. The classes that used to inherit from `DIEValue` no longer do. There are two categories of these: - Small values fit in a single pointer and are stored by value. - Large values require auxiliary storage, and are stored by reference. The only non-mechanical change is to tools/dsymutil/DwarfLinker.cpp. It was relying on `DIEInteger`s being passed around by reference, so I replaced that assumption with a `PatchLocation` type that stores a safe reference to where the `DIEInteger` lives instead. This commit causes a temporary regression in memory usage, since I've left merging `DIEAbbrevData` into `DIEValue` for a follow-up commit. I measured an increase from 845 MB to 879 MB, around 3.9%. The follow-up drops it lower than the starting point, and I've only recently brought the memory this low anyway, so I'm committing these changes separately to keep them incremental. (I also considered swapping the commits, but the other one first would cause a lot more code churn.) (I'm looking at `llc` memory usage on `verify-uselistorder.lto.opt.bc`; see r236629 for details.) -- llvm-svn: 238362
2015-05-28 06:14:58 +08:00
static_cast<dwarf::Form>(AttrSpec.Form), DIEInteger(Addr));
return Unit.getOrigUnit().getAddressByteSize();
}
/// \brief Clone a scalar attribute and add it to \p Die.
/// \returns the size of the new attribute.
unsigned DwarfLinker::cloneScalarAttribute(
DIE &Die, const DWARFDebugInfoEntryMinimal &InputDIE, CompileUnit &Unit,
AttributeSpec AttrSpec, const DWARFFormValue &Val, unsigned AttrSize,
AttributesInfo &Info) {
uint64_t Value;
if (AttrSpec.Attr == dwarf::DW_AT_high_pc &&
Die.getTag() == dwarf::DW_TAG_compile_unit) {
if (Unit.getLowPc() == -1ULL)
return 0;
// Dwarf >= 4 high_pc is an size, not an address.
Value = Unit.getHighPc() - Unit.getLowPc();
} else if (AttrSpec.Form == dwarf::DW_FORM_sec_offset)
Value = *Val.getAsSectionOffset();
else if (AttrSpec.Form == dwarf::DW_FORM_sdata)
Value = *Val.getAsSignedConstant();
else if (auto OptionalValue = Val.getAsUnsignedConstant())
Value = *OptionalValue;
else {
reportWarning("Unsupported scalar attribute form. Dropping attribute.",
&Unit.getOrigUnit(), &InputDIE);
return 0;
}
PatchLocation Patch =
Die.addValue(DIEAlloc, dwarf::Attribute(AttrSpec.Attr),
dwarf::Form(AttrSpec.Form), DIEInteger(Value));
if (AttrSpec.Attr == dwarf::DW_AT_ranges)
Unit.noteRangeAttribute(Die, Patch);
// A more generic way to check for location attributes would be
// nice, but it's very unlikely that any other attribute needs a
// location list.
else if (AttrSpec.Attr == dwarf::DW_AT_location ||
AttrSpec.Attr == dwarf::DW_AT_frame_base)
Unit.noteLocationAttribute(Patch, Info.PCOffset);
else if (AttrSpec.Attr == dwarf::DW_AT_declaration && Value)
Info.IsDeclaration = true;
return AttrSize;
}
/// \brief Clone \p InputDIE's attribute described by \p AttrSpec with
/// value \p Val, and add it to \p Die.
/// \returns the size of the cloned attribute.
unsigned DwarfLinker::cloneAttribute(DIE &Die,
const DWARFDebugInfoEntryMinimal &InputDIE,
CompileUnit &Unit,
const DWARFFormValue &Val,
const AttributeSpec AttrSpec,
unsigned AttrSize, AttributesInfo &Info) {
const DWARFUnit &U = Unit.getOrigUnit();
switch (AttrSpec.Form) {
case dwarf::DW_FORM_strp:
case dwarf::DW_FORM_string:
return cloneStringAttribute(Die, AttrSpec, Val, U);
case dwarf::DW_FORM_ref_addr:
case dwarf::DW_FORM_ref1:
case dwarf::DW_FORM_ref2:
case dwarf::DW_FORM_ref4:
case dwarf::DW_FORM_ref8:
return cloneDieReferenceAttribute(Die, InputDIE, AttrSpec, AttrSize, Val,
Unit);
case dwarf::DW_FORM_block:
case dwarf::DW_FORM_block1:
case dwarf::DW_FORM_block2:
case dwarf::DW_FORM_block4:
case dwarf::DW_FORM_exprloc:
return cloneBlockAttribute(Die, AttrSpec, Val, AttrSize);
case dwarf::DW_FORM_addr:
return cloneAddressAttribute(Die, AttrSpec, Val, Unit, Info);
case dwarf::DW_FORM_data1:
case dwarf::DW_FORM_data2:
case dwarf::DW_FORM_data4:
case dwarf::DW_FORM_data8:
case dwarf::DW_FORM_udata:
case dwarf::DW_FORM_sdata:
case dwarf::DW_FORM_sec_offset:
case dwarf::DW_FORM_flag:
case dwarf::DW_FORM_flag_present:
return cloneScalarAttribute(Die, InputDIE, Unit, AttrSpec, Val, AttrSize,
Info);
default:
reportWarning("Unsupported attribute form in cloneAttribute. Dropping.", &U,
&InputDIE);
}
return 0;
}
/// \brief Apply the valid relocations found by findValidRelocs() to
/// the buffer \p Data, taking into account that Data is at \p BaseOffset
/// in the debug_info section.
///
/// Like for findValidRelocs(), this function must be called with
/// monotonic \p BaseOffset values.
///
/// \returns wether any reloc has been applied.
bool DwarfLinker::applyValidRelocs(MutableArrayRef<char> Data,
uint32_t BaseOffset, bool isLittleEndian) {
assert((NextValidReloc == 0 ||
BaseOffset > ValidRelocs[NextValidReloc - 1].Offset) &&
"BaseOffset should only be increasing.");
if (NextValidReloc >= ValidRelocs.size())
return false;
// Skip relocs that haven't been applied.
while (NextValidReloc < ValidRelocs.size() &&
ValidRelocs[NextValidReloc].Offset < BaseOffset)
++NextValidReloc;
bool Applied = false;
uint64_t EndOffset = BaseOffset + Data.size();
while (NextValidReloc < ValidRelocs.size() &&
ValidRelocs[NextValidReloc].Offset >= BaseOffset &&
ValidRelocs[NextValidReloc].Offset < EndOffset) {
const auto &ValidReloc = ValidRelocs[NextValidReloc++];
assert(ValidReloc.Offset - BaseOffset < Data.size());
assert(ValidReloc.Offset - BaseOffset + ValidReloc.Size <= Data.size());
char Buf[8];
uint64_t Value = ValidReloc.Mapping->getValue().BinaryAddress;
Value += ValidReloc.Addend;
for (unsigned i = 0; i != ValidReloc.Size; ++i) {
unsigned Index = isLittleEndian ? i : (ValidReloc.Size - i - 1);
Buf[i] = uint8_t(Value >> (Index * 8));
}
assert(ValidReloc.Size <= sizeof(Buf));
memcpy(&Data[ValidReloc.Offset - BaseOffset], Buf, ValidReloc.Size);
Applied = true;
}
return Applied;
}
static bool isTypeTag(uint16_t Tag) {
switch (Tag) {
case dwarf::DW_TAG_array_type:
case dwarf::DW_TAG_class_type:
case dwarf::DW_TAG_enumeration_type:
case dwarf::DW_TAG_pointer_type:
case dwarf::DW_TAG_reference_type:
case dwarf::DW_TAG_string_type:
case dwarf::DW_TAG_structure_type:
case dwarf::DW_TAG_subroutine_type:
case dwarf::DW_TAG_typedef:
case dwarf::DW_TAG_union_type:
case dwarf::DW_TAG_ptr_to_member_type:
case dwarf::DW_TAG_set_type:
case dwarf::DW_TAG_subrange_type:
case dwarf::DW_TAG_base_type:
case dwarf::DW_TAG_const_type:
case dwarf::DW_TAG_constant:
case dwarf::DW_TAG_file_type:
case dwarf::DW_TAG_namelist:
case dwarf::DW_TAG_packed_type:
case dwarf::DW_TAG_volatile_type:
case dwarf::DW_TAG_restrict_type:
case dwarf::DW_TAG_interface_type:
case dwarf::DW_TAG_unspecified_type:
case dwarf::DW_TAG_shared_type:
return true;
default:
break;
}
return false;
}
/// \brief Recursively clone \p InputDIE's subtrees that have been
/// selected to appear in the linked output.
///
/// \param OutOffset is the Offset where the newly created DIE will
/// lie in the linked compile unit.
///
/// \returns the cloned DIE object or null if nothing was selected.
DIE *DwarfLinker::cloneDIE(const DWARFDebugInfoEntryMinimal &InputDIE,
CompileUnit &Unit, int64_t PCOffset,
uint32_t OutOffset) {
DWARFUnit &U = Unit.getOrigUnit();
unsigned Idx = U.getDIEIndex(&InputDIE);
CompileUnit::DIEInfo &Info = Unit.getInfo(Idx);
// Should the DIE appear in the output?
if (!Unit.getInfo(Idx).Keep)
return nullptr;
uint32_t Offset = InputDIE.getOffset();
// The DIE might have been already created by a forward reference
// (see cloneDieReferenceAttribute()).
DIE *Die = Info.Clone;
if (!Die)
Die = Info.Clone = DIE::get(DIEAlloc, dwarf::Tag(InputDIE.getTag()));
assert(Die->getTag() == InputDIE.getTag());
Die->setOffset(OutOffset);
if (Unit.hasODR() && Die->getTag() != dwarf::DW_TAG_namespace && Info.Ctxt &&
Info.Ctxt != Unit.getInfo(Info.ParentIdx).Ctxt &&
!Info.Ctxt->getCanonicalDIEOffset()) {
// We are about to emit a DIE that is the root of its own valid
// DeclContext tree. Make the current offset the canonical offset
// for this context.
Info.Ctxt->setCanonicalDIEOffset(OutOffset + Unit.getStartOffset());
}
// Extract and clone every attribute.
DataExtractor Data = U.getDebugInfoExtractor();
uint32_t NextOffset = U.getDIEAtIndex(Idx + 1)->getOffset();
AttributesInfo AttrInfo;
// We could copy the data only if we need to aply a relocation to
// it. After testing, it seems there is no performance downside to
// doing the copy unconditionally, and it makes the code simpler.
SmallString<40> DIECopy(Data.getData().substr(Offset, NextOffset - Offset));
Data = DataExtractor(DIECopy, Data.isLittleEndian(), Data.getAddressSize());
// Modify the copy with relocated addresses.
if (applyValidRelocs(DIECopy, Offset, Data.isLittleEndian())) {
// If we applied relocations, we store the value of high_pc that was
// potentially stored in the input DIE. If high_pc is an address
// (Dwarf version == 2), then it might have been relocated to a
// totally unrelated value (because the end address in the object
// file might be start address of another function which got moved
// independantly by the linker). The computation of the actual
// high_pc value is done in cloneAddressAttribute().
AttrInfo.OrigHighPc =
InputDIE.getAttributeValueAsAddress(&U, dwarf::DW_AT_high_pc, 0);
}
// Reset the Offset to 0 as we will be working on the local copy of
// the data.
Offset = 0;
const auto *Abbrev = InputDIE.getAbbreviationDeclarationPtr();
Offset += getULEB128Size(Abbrev->getCode());
// We are entering a subprogram. Get and propagate the PCOffset.
if (Die->getTag() == dwarf::DW_TAG_subprogram)
PCOffset = Info.AddrAdjust;
AttrInfo.PCOffset = PCOffset;
for (const auto &AttrSpec : Abbrev->attributes()) {
DWARFFormValue Val(AttrSpec.Form);
uint32_t AttrSize = Offset;
Val.extractValue(Data, &Offset, &U);
AttrSize = Offset - AttrSize;
OutOffset +=
cloneAttribute(*Die, InputDIE, Unit, Val, AttrSpec, AttrSize, AttrInfo);
}
// Look for accelerator entries.
uint16_t Tag = InputDIE.getTag();
// FIXME: This is slightly wrong. An inline_subroutine without a
// low_pc, but with AT_ranges might be interesting to get into the
// accelerator tables too. For now stick with dsymutil's behavior.
if ((Info.InDebugMap || AttrInfo.HasLowPc) &&
Tag != dwarf::DW_TAG_compile_unit &&
getDIENames(InputDIE, Unit.getOrigUnit(), AttrInfo)) {
if (AttrInfo.MangledName && AttrInfo.MangledName != AttrInfo.Name)
Unit.addNameAccelerator(Die, AttrInfo.MangledName,
AttrInfo.MangledNameOffset,
Tag == dwarf::DW_TAG_inlined_subroutine);
if (AttrInfo.Name)
Unit.addNameAccelerator(Die, AttrInfo.Name, AttrInfo.NameOffset,
Tag == dwarf::DW_TAG_inlined_subroutine);
} else if (isTypeTag(Tag) && !AttrInfo.IsDeclaration &&
getDIENames(InputDIE, Unit.getOrigUnit(), AttrInfo)) {
Unit.addTypeAccelerator(Die, AttrInfo.Name, AttrInfo.NameOffset);
}
DIEAbbrev NewAbbrev = Die->generateAbbrev();
// If a scope DIE is kept, we must have kept at least one child. If
// it's not the case, we'll just be emitting one wasteful end of
// children marker, but things won't break.
if (InputDIE.hasChildren())
NewAbbrev.setChildrenFlag(dwarf::DW_CHILDREN_yes);
// Assign a permanent abbrev number
AssignAbbrev(NewAbbrev);
Die->setAbbrevNumber(NewAbbrev.getNumber());
// Add the size of the abbreviation number to the output offset.
OutOffset += getULEB128Size(Die->getAbbrevNumber());
if (!Abbrev->hasChildren()) {
// Update our size.
Die->setSize(OutOffset - Die->getOffset());
return Die;
}
// Recursively clone children.
for (auto *Child = InputDIE.getFirstChild(); Child && !Child->isNULL();
Child = Child->getSibling()) {
if (DIE *Clone = cloneDIE(*Child, Unit, PCOffset, OutOffset)) {
Die->addChild(Clone);
OutOffset = Clone->getOffset() + Clone->getSize();
}
}
// Account for the end of children marker.
OutOffset += sizeof(int8_t);
// Update our size.
Die->setSize(OutOffset - Die->getOffset());
return Die;
}
/// \brief Patch the input object file relevant debug_ranges entries
/// and emit them in the output file. Update the relevant attributes
/// to point at the new entries.
void DwarfLinker::patchRangesForUnit(const CompileUnit &Unit,
DWARFContext &OrigDwarf) const {
DWARFDebugRangeList RangeList;
const auto &FunctionRanges = Unit.getFunctionRanges();
unsigned AddressSize = Unit.getOrigUnit().getAddressByteSize();
DataExtractor RangeExtractor(OrigDwarf.getRangeSection(),
OrigDwarf.isLittleEndian(), AddressSize);
auto InvalidRange = FunctionRanges.end(), CurrRange = InvalidRange;
DWARFUnit &OrigUnit = Unit.getOrigUnit();
const auto *OrigUnitDie = OrigUnit.getUnitDIE(false);
uint64_t OrigLowPc = OrigUnitDie->getAttributeValueAsAddress(
&OrigUnit, dwarf::DW_AT_low_pc, -1ULL);
// Ranges addresses are based on the unit's low_pc. Compute the
// offset we need to apply to adapt to the the new unit's low_pc.
int64_t UnitPcOffset = 0;
if (OrigLowPc != -1ULL)
UnitPcOffset = int64_t(OrigLowPc) - Unit.getLowPc();
for (const auto &RangeAttribute : Unit.getRangesAttributes()) {
Reapply "AsmPrinter: Change DIEValue to be stored by value" This reverts commit r238350, effectively reapplying r238349 after fixing (all?) the problems, all somehow related to how I was using `AlignedArrayCharUnion<>` inside `DIEValue`: - MSVC can only handle `sizeof()` on types, not values. Change the assert. - GCC doesn't know the `is_trivially_copyable` type trait. Instead of asserting it, add destructors. - Call placement new even when constructing POD (i.e., the pointers). - Instead of copying the char buffer, copy the casted classes. I've left in a couple of `static_assert`s that I think both MSVC and GCC know how to handle. If the bots disagree with me, I'll remove them. - Check that the constructed type is either standard layout or a pointer. This protects against a programming error: we really want the "small" `DIEValue`s to be small and simple, so don't accidentally change them not to be. - Similarly, check that the size of the buffer is no bigger than a `uint64_t` or a pointer. (I thought checking against `sizeof(uint64_t)` would be good enough, but Chandler suggested that pointers might sometimes be bigger than that in the context of sanitizers.) I've also committed r238359 in the meantime, which introduces a DIEValue.def to simplify dispatching between the various types (thanks to a review comment by David Blaikie). Without that, this commit would be almost unintelligible. Here's the original commit message: -- Change `DIEValue` to be stored/passed/etc. by value, instead of reference. It's now a discriminated union, with a `Val` field storing the actual type. The classes that used to inherit from `DIEValue` no longer do. There are two categories of these: - Small values fit in a single pointer and are stored by value. - Large values require auxiliary storage, and are stored by reference. The only non-mechanical change is to tools/dsymutil/DwarfLinker.cpp. It was relying on `DIEInteger`s being passed around by reference, so I replaced that assumption with a `PatchLocation` type that stores a safe reference to where the `DIEInteger` lives instead. This commit causes a temporary regression in memory usage, since I've left merging `DIEAbbrevData` into `DIEValue` for a follow-up commit. I measured an increase from 845 MB to 879 MB, around 3.9%. The follow-up drops it lower than the starting point, and I've only recently brought the memory this low anyway, so I'm committing these changes separately to keep them incremental. (I also considered swapping the commits, but the other one first would cause a lot more code churn.) (I'm looking at `llc` memory usage on `verify-uselistorder.lto.opt.bc`; see r236629 for details.) -- llvm-svn: 238362
2015-05-28 06:14:58 +08:00
uint32_t Offset = RangeAttribute.get();
RangeAttribute.set(Streamer->getRangesSectionSize());
RangeList.extract(RangeExtractor, &Offset);
const auto &Entries = RangeList.getEntries();
const DWARFDebugRangeList::RangeListEntry &First = Entries.front();
if (CurrRange == InvalidRange || First.StartAddress < CurrRange.start() ||
First.StartAddress >= CurrRange.stop()) {
CurrRange = FunctionRanges.find(First.StartAddress + OrigLowPc);
if (CurrRange == InvalidRange ||
CurrRange.start() > First.StartAddress + OrigLowPc) {
reportWarning("no mapping for range.");
continue;
}
}
Streamer->emitRangesEntries(UnitPcOffset, OrigLowPc, CurrRange, Entries,
AddressSize);
}
}
/// \brief Generate the debug_aranges entries for \p Unit and if the
/// unit has a DW_AT_ranges attribute, also emit the debug_ranges
/// contribution for this attribute.
/// FIXME: this could actually be done right in patchRangesForUnit,
/// but for the sake of initial bit-for-bit compatibility with legacy
/// dsymutil, we have to do it in a delayed pass.
void DwarfLinker::generateUnitRanges(CompileUnit &Unit) const {
Reapply "AsmPrinter: Change DIEValue to be stored by value" This reverts commit r238350, effectively reapplying r238349 after fixing (all?) the problems, all somehow related to how I was using `AlignedArrayCharUnion<>` inside `DIEValue`: - MSVC can only handle `sizeof()` on types, not values. Change the assert. - GCC doesn't know the `is_trivially_copyable` type trait. Instead of asserting it, add destructors. - Call placement new even when constructing POD (i.e., the pointers). - Instead of copying the char buffer, copy the casted classes. I've left in a couple of `static_assert`s that I think both MSVC and GCC know how to handle. If the bots disagree with me, I'll remove them. - Check that the constructed type is either standard layout or a pointer. This protects against a programming error: we really want the "small" `DIEValue`s to be small and simple, so don't accidentally change them not to be. - Similarly, check that the size of the buffer is no bigger than a `uint64_t` or a pointer. (I thought checking against `sizeof(uint64_t)` would be good enough, but Chandler suggested that pointers might sometimes be bigger than that in the context of sanitizers.) I've also committed r238359 in the meantime, which introduces a DIEValue.def to simplify dispatching between the various types (thanks to a review comment by David Blaikie). Without that, this commit would be almost unintelligible. Here's the original commit message: -- Change `DIEValue` to be stored/passed/etc. by value, instead of reference. It's now a discriminated union, with a `Val` field storing the actual type. The classes that used to inherit from `DIEValue` no longer do. There are two categories of these: - Small values fit in a single pointer and are stored by value. - Large values require auxiliary storage, and are stored by reference. The only non-mechanical change is to tools/dsymutil/DwarfLinker.cpp. It was relying on `DIEInteger`s being passed around by reference, so I replaced that assumption with a `PatchLocation` type that stores a safe reference to where the `DIEInteger` lives instead. This commit causes a temporary regression in memory usage, since I've left merging `DIEAbbrevData` into `DIEValue` for a follow-up commit. I measured an increase from 845 MB to 879 MB, around 3.9%. The follow-up drops it lower than the starting point, and I've only recently brought the memory this low anyway, so I'm committing these changes separately to keep them incremental. (I also considered swapping the commits, but the other one first would cause a lot more code churn.) (I'm looking at `llc` memory usage on `verify-uselistorder.lto.opt.bc`; see r236629 for details.) -- llvm-svn: 238362
2015-05-28 06:14:58 +08:00
auto Attr = Unit.getUnitRangesAttribute();
if (Attr)
Reapply "AsmPrinter: Change DIEValue to be stored by value" This reverts commit r238350, effectively reapplying r238349 after fixing (all?) the problems, all somehow related to how I was using `AlignedArrayCharUnion<>` inside `DIEValue`: - MSVC can only handle `sizeof()` on types, not values. Change the assert. - GCC doesn't know the `is_trivially_copyable` type trait. Instead of asserting it, add destructors. - Call placement new even when constructing POD (i.e., the pointers). - Instead of copying the char buffer, copy the casted classes. I've left in a couple of `static_assert`s that I think both MSVC and GCC know how to handle. If the bots disagree with me, I'll remove them. - Check that the constructed type is either standard layout or a pointer. This protects against a programming error: we really want the "small" `DIEValue`s to be small and simple, so don't accidentally change them not to be. - Similarly, check that the size of the buffer is no bigger than a `uint64_t` or a pointer. (I thought checking against `sizeof(uint64_t)` would be good enough, but Chandler suggested that pointers might sometimes be bigger than that in the context of sanitizers.) I've also committed r238359 in the meantime, which introduces a DIEValue.def to simplify dispatching between the various types (thanks to a review comment by David Blaikie). Without that, this commit would be almost unintelligible. Here's the original commit message: -- Change `DIEValue` to be stored/passed/etc. by value, instead of reference. It's now a discriminated union, with a `Val` field storing the actual type. The classes that used to inherit from `DIEValue` no longer do. There are two categories of these: - Small values fit in a single pointer and are stored by value. - Large values require auxiliary storage, and are stored by reference. The only non-mechanical change is to tools/dsymutil/DwarfLinker.cpp. It was relying on `DIEInteger`s being passed around by reference, so I replaced that assumption with a `PatchLocation` type that stores a safe reference to where the `DIEInteger` lives instead. This commit causes a temporary regression in memory usage, since I've left merging `DIEAbbrevData` into `DIEValue` for a follow-up commit. I measured an increase from 845 MB to 879 MB, around 3.9%. The follow-up drops it lower than the starting point, and I've only recently brought the memory this low anyway, so I'm committing these changes separately to keep them incremental. (I also considered swapping the commits, but the other one first would cause a lot more code churn.) (I'm looking at `llc` memory usage on `verify-uselistorder.lto.opt.bc`; see r236629 for details.) -- llvm-svn: 238362
2015-05-28 06:14:58 +08:00
Attr->set(Streamer->getRangesSectionSize());
Streamer->emitUnitRangesEntries(Unit, static_cast<bool>(Attr));
}
/// \brief Insert the new line info sequence \p Seq into the current
/// set of already linked line info \p Rows.
static void insertLineSequence(std::vector<DWARFDebugLine::Row> &Seq,
std::vector<DWARFDebugLine::Row> &Rows) {
if (Seq.empty())
return;
if (!Rows.empty() && Rows.back().Address < Seq.front().Address) {
Rows.insert(Rows.end(), Seq.begin(), Seq.end());
Seq.clear();
return;
}
auto InsertPoint = std::lower_bound(
Rows.begin(), Rows.end(), Seq.front(),
[](const DWARFDebugLine::Row &LHS, const DWARFDebugLine::Row &RHS) {
return LHS.Address < RHS.Address;
});
// FIXME: this only removes the unneeded end_sequence if the
// sequences have been inserted in order. using a global sort like
// described in patchLineTableForUnit() and delaying the end_sequene
// elimination to emitLineTableForUnit() we can get rid of all of them.
if (InsertPoint != Rows.end() &&
InsertPoint->Address == Seq.front().Address && InsertPoint->EndSequence) {
*InsertPoint = Seq.front();
Rows.insert(InsertPoint + 1, Seq.begin() + 1, Seq.end());
} else {
Rows.insert(InsertPoint, Seq.begin(), Seq.end());
}
Seq.clear();
}
static void patchStmtList(DIE &Die, DIEInteger Offset) {
for (auto &V : Die.values())
if (V.getAttribute() == dwarf::DW_AT_stmt_list) {
V = DIEValue(V.getAttribute(), V.getForm(), Offset);
return;
}
llvm_unreachable("Didn't find DW_AT_stmt_list in cloned DIE!");
}
/// \brief Extract the line table for \p Unit from \p OrigDwarf, and
/// recreate a relocated version of these for the address ranges that
/// are present in the binary.
void DwarfLinker::patchLineTableForUnit(CompileUnit &Unit,
DWARFContext &OrigDwarf) {
const DWARFDebugInfoEntryMinimal *CUDie = Unit.getOrigUnit().getUnitDIE();
uint64_t StmtList = CUDie->getAttributeValueAsSectionOffset(
&Unit.getOrigUnit(), dwarf::DW_AT_stmt_list, -1ULL);
if (StmtList == -1ULL)
return;
// Update the cloned DW_AT_stmt_list with the correct debug_line offset.
if (auto *OutputDIE = Unit.getOutputUnitDIE())
patchStmtList(*OutputDIE, DIEInteger(Streamer->getLineSectionSize()));
// Parse the original line info for the unit.
DWARFDebugLine::LineTable LineTable;
uint32_t StmtOffset = StmtList;
StringRef LineData = OrigDwarf.getLineSection().Data;
DataExtractor LineExtractor(LineData, OrigDwarf.isLittleEndian(),
Unit.getOrigUnit().getAddressByteSize());
LineTable.parse(LineExtractor, &OrigDwarf.getLineSection().Relocs,
&StmtOffset);
// This vector is the output line table.
std::vector<DWARFDebugLine::Row> NewRows;
NewRows.reserve(LineTable.Rows.size());
// Current sequence of rows being extracted, before being inserted
// in NewRows.
std::vector<DWARFDebugLine::Row> Seq;
const auto &FunctionRanges = Unit.getFunctionRanges();
auto InvalidRange = FunctionRanges.end(), CurrRange = InvalidRange;
// FIXME: This logic is meant to generate exactly the same output as
// Darwin's classic dsynutil. There is a nicer way to implement this
// by simply putting all the relocated line info in NewRows and simply
// sorting NewRows before passing it to emitLineTableForUnit. This
// should be correct as sequences for a function should stay
// together in the sorted output. There are a few corner cases that
// look suspicious though, and that required to implement the logic
// this way. Revisit that once initial validation is finished.
// Iterate over the object file line info and extract the sequences
// that correspond to linked functions.
for (auto &Row : LineTable.Rows) {
// Check wether we stepped out of the range. The range is
// half-open, but consider accept the end address of the range if
// it is marked as end_sequence in the input (because in that
// case, the relocation offset is accurate and that entry won't
// serve as the start of another function).
if (CurrRange == InvalidRange || Row.Address < CurrRange.start() ||
Row.Address > CurrRange.stop() ||
(Row.Address == CurrRange.stop() && !Row.EndSequence)) {
// We just stepped out of a known range. Insert a end_sequence
// corresponding to the end of the range.
uint64_t StopAddress = CurrRange != InvalidRange
? CurrRange.stop() + CurrRange.value()
: -1ULL;
CurrRange = FunctionRanges.find(Row.Address);
bool CurrRangeValid =
CurrRange != InvalidRange && CurrRange.start() <= Row.Address;
if (!CurrRangeValid) {
CurrRange = InvalidRange;
if (StopAddress != -1ULL) {
// Try harder by looking in the DebugMapObject function
// ranges map. There are corner cases where this finds a
// valid entry. It's unclear if this is right or wrong, but
// for now do as dsymutil.
// FIXME: Understand exactly what cases this addresses and
// potentially remove it along with the Ranges map.
auto Range = Ranges.lower_bound(Row.Address);
if (Range != Ranges.begin() && Range != Ranges.end())
--Range;
if (Range != Ranges.end() && Range->first <= Row.Address &&
Range->second.first >= Row.Address) {
StopAddress = Row.Address + Range->second.second;
}
}
}
if (StopAddress != -1ULL && !Seq.empty()) {
// Insert end sequence row with the computed end address, but
// the same line as the previous one.
Seq.emplace_back(Seq.back());
Seq.back().Address = StopAddress;
Seq.back().EndSequence = 1;
Seq.back().PrologueEnd = 0;
Seq.back().BasicBlock = 0;
Seq.back().EpilogueBegin = 0;
insertLineSequence(Seq, NewRows);
}
if (!CurrRangeValid)
continue;
}
// Ignore empty sequences.
if (Row.EndSequence && Seq.empty())
continue;
// Relocate row address and add it to the current sequence.
Row.Address += CurrRange.value();
Seq.emplace_back(Row);
if (Row.EndSequence)
insertLineSequence(Seq, NewRows);
}
// Finished extracting, now emit the line tables.
uint32_t PrologueEnd = StmtList + 10 + LineTable.Prologue.PrologueLength;
// FIXME: LLVM hardcodes it's prologue values. We just copy the
// prologue over and that works because we act as both producer and
// consumer. It would be nicer to have a real configurable line
// table emitter.
if (LineTable.Prologue.Version != 2 ||
LineTable.Prologue.DefaultIsStmt != DWARF2_LINE_DEFAULT_IS_STMT ||
LineTable.Prologue.LineBase != -5 || LineTable.Prologue.LineRange != 14 ||
LineTable.Prologue.OpcodeBase != 13)
reportWarning("line table paramters mismatch. Cannot emit.");
else
Streamer->emitLineTableForUnit(LineData.slice(StmtList + 4, PrologueEnd),
LineTable.Prologue.MinInstLength, NewRows,
Unit.getOrigUnit().getAddressByteSize());
}
void DwarfLinker::emitAcceleratorEntriesForUnit(CompileUnit &Unit) {
Streamer->emitPubNamesForUnit(Unit);
Streamer->emitPubTypesForUnit(Unit);
}
/// \brief Read the frame info stored in the object, and emit the
/// patched frame descriptions for the linked binary.
///
/// This is actually pretty easy as the data of the CIEs and FDEs can
/// be considered as black boxes and moved as is. The only thing to do
/// is to patch the addresses in the headers.
void DwarfLinker::patchFrameInfoForObject(const DebugMapObject &DMO,
DWARFContext &OrigDwarf,
unsigned AddrSize) {
StringRef FrameData = OrigDwarf.getDebugFrameSection();
if (FrameData.empty())
return;
DataExtractor Data(FrameData, OrigDwarf.isLittleEndian(), 0);
uint32_t InputOffset = 0;
// Store the data of the CIEs defined in this object, keyed by their
// offsets.
DenseMap<uint32_t, StringRef> LocalCIES;
while (Data.isValidOffset(InputOffset)) {
uint32_t EntryOffset = InputOffset;
uint32_t InitialLength = Data.getU32(&InputOffset);
if (InitialLength == 0xFFFFFFFF)
return reportWarning("Dwarf64 bits no supported");
uint32_t CIEId = Data.getU32(&InputOffset);
if (CIEId == 0xFFFFFFFF) {
// This is a CIE, store it.
StringRef CIEData = FrameData.substr(EntryOffset, InitialLength + 4);
LocalCIES[EntryOffset] = CIEData;
// The -4 is to account for the CIEId we just read.
InputOffset += InitialLength - 4;
continue;
}
uint32_t Loc = Data.getUnsigned(&InputOffset, AddrSize);
// Some compilers seem to emit frame info that doesn't start at
// the function entry point, thus we can't just lookup the address
// in the debug map. Use the linker's range map to see if the FDE
// describes something that we can relocate.
auto Range = Ranges.upper_bound(Loc);
if (Range != Ranges.begin())
--Range;
if (Range == Ranges.end() || Range->first > Loc ||
Range->second.first <= Loc) {
// The +4 is to account for the size of the InitialLength field itself.
InputOffset = EntryOffset + InitialLength + 4;
continue;
}
// This is an FDE, and we have a mapping.
// Have we already emitted a corresponding CIE?
StringRef CIEData = LocalCIES[CIEId];
if (CIEData.empty())
return reportWarning("Inconsistent debug_frame content. Dropping.");
// Look if we already emitted a CIE that corresponds to the
// referenced one (the CIE data is the key of that lookup).
auto IteratorInserted = EmittedCIEs.insert(
std::make_pair(CIEData, Streamer->getFrameSectionSize()));
// If there is no CIE yet for this ID, emit it.
if (IteratorInserted.second ||
// FIXME: dsymutil-classic only caches the last used CIE for
// reuse. Mimic that behavior for now. Just removing that
// second half of the condition and the LastCIEOffset variable
// makes the code DTRT.
LastCIEOffset != IteratorInserted.first->getValue()) {
LastCIEOffset = Streamer->getFrameSectionSize();
IteratorInserted.first->getValue() = LastCIEOffset;
Streamer->emitCIE(CIEData);
}
// Emit the FDE with updated address and CIE pointer.
// (4 + AddrSize) is the size of the CIEId + initial_location
// fields that will get reconstructed by emitFDE().
unsigned FDERemainingBytes = InitialLength - (4 + AddrSize);
Streamer->emitFDE(IteratorInserted.first->getValue(), AddrSize,
Loc + Range->second.second,
FrameData.substr(InputOffset, FDERemainingBytes));
InputOffset += FDERemainingBytes;
}
}
ErrorOr<const object::ObjectFile &>
DwarfLinker::loadObject(BinaryHolder &BinaryHolder, DebugMapObject &Obj,
const DebugMap &Map) {
auto ErrOrObjs =
BinaryHolder.GetObjectFiles(Obj.getObjectFilename(), Obj.getTimestamp());
if (std::error_code EC = ErrOrObjs.getError())
reportWarning(Twine(Obj.getObjectFilename()) + ": " + EC.message());
auto ErrOrObj = BinaryHolder.Get(Map.getTriple());
if (std::error_code EC = ErrOrObj.getError())
reportWarning(Twine(Obj.getObjectFilename()) + ": " + EC.message());
return ErrOrObj;
}
bool DwarfLinker::link(const DebugMap &Map) {
if (Map.begin() == Map.end()) {
errs() << "Empty debug map.\n";
return false;
}
if (!createStreamer(Map.getTriple(), OutputFilename))
return false;
// Size of the DIEs (and headers) generated for the linked output.
uint64_t OutputDebugInfoSize = 0;
// A unique ID that identifies each compile unit.
unsigned UnitID = 0;
for (const auto &Obj : Map.objects()) {
CurrentDebugObject = Obj.get();
if (Options.Verbose)
outs() << "DEBUG MAP OBJECT: " << Obj->getObjectFilename() << "\n";
auto ErrOrObj = loadObject(BinHolder, *Obj, Map);
if (!ErrOrObj)
continue;
// Look for relocations that correspond to debug map entries.
if (!findValidRelocsInDebugInfo(*ErrOrObj, *Obj)) {
if (Options.Verbose)
outs() << "No valid relocations found. Skipping.\n";
continue;
}
// Setup access to the debug info.
DWARFContextInMemory DwarfContext(*ErrOrObj);
startDebugObject(DwarfContext, *Obj);
// In a first phase, just read in the debug info and store the DIE
// parent links that we will use during the next phase.
for (const auto &CU : DwarfContext.compile_units()) {
auto *CUDie = CU->getUnitDIE(false);
if (Options.Verbose) {
outs() << "Input compilation unit:";
CUDie->dump(outs(), CU.get(), 0);
}
Units.emplace_back(*CU, UnitID++, !Options.NoODR);
gatherDIEParents(CUDie, 0, Units.back(), &ODRContexts.getRoot(),
StringPool, ODRContexts);
}
// Then mark all the DIEs that need to be present in the linked
// output and collect some information about them. Note that this
// loop can not be merged with the previous one becaue cross-cu
// references require the ParentIdx to be setup for every CU in
// the object file before calling this.
for (auto &CurrentUnit : Units)
lookForDIEsToKeep(*CurrentUnit.getOrigUnit().getUnitDIE(), *Obj,
CurrentUnit, 0);
// The calls to applyValidRelocs inside cloneDIE will walk the
// reloc array again (in the same way findValidRelocsInDebugInfo()
// did). We need to reset the NextValidReloc index to the beginning.
NextValidReloc = 0;
// Construct the output DIE tree by cloning the DIEs we chose to
// keep above. If there are no valid relocs, then there's nothing
// to clone/emit.
if (!ValidRelocs.empty())
for (auto &CurrentUnit : Units) {
const auto *InputDIE = CurrentUnit.getOrigUnit().getUnitDIE();
CurrentUnit.setStartOffset(OutputDebugInfoSize);
DIE *OutputDIE = cloneDIE(*InputDIE, CurrentUnit, 0 /* PCOffset */,
11 /* Unit Header size */);
CurrentUnit.setOutputUnitDIE(OutputDIE);
OutputDebugInfoSize = CurrentUnit.computeNextUnitOffset();
if (Options.NoOutput)
continue;
// FIXME: for compatibility with the classic dsymutil, we emit
// an empty line table for the unit, even if the unit doesn't
// actually exist in the DIE tree.
patchLineTableForUnit(CurrentUnit, DwarfContext);
if (!OutputDIE)
continue;
patchRangesForUnit(CurrentUnit, DwarfContext);
Streamer->emitLocationsForUnit(CurrentUnit, DwarfContext);
emitAcceleratorEntriesForUnit(CurrentUnit);
}
// Emit all the compile unit's debug information.
if (!ValidRelocs.empty() && !Options.NoOutput)
for (auto &CurrentUnit : Units) {
generateUnitRanges(CurrentUnit);
CurrentUnit.fixupForwardReferences();
Streamer->emitCompileUnitHeader(CurrentUnit);
if (!CurrentUnit.getOutputUnitDIE())
continue;
Streamer->emitDIE(*CurrentUnit.getOutputUnitDIE());
}
if (!ValidRelocs.empty() && !Options.NoOutput && !Units.empty())
patchFrameInfoForObject(*Obj, DwarfContext,
Units[0].getOrigUnit().getAddressByteSize());
// Clean-up before starting working on the next object.
endDebugObject();
}
// Emit everything that's global.
if (!Options.NoOutput) {
Streamer->emitAbbrevs(Abbreviations);
Streamer->emitStrings(StringPool);
}
return Options.NoOutput ? true : Streamer->finish();
}
}
bool linkDwarf(StringRef OutputFilename, const DebugMap &DM,
const LinkOptions &Options) {
DwarfLinker Linker(OutputFilename, Options);
return Linker.link(DM);
}
}
}