DWARF: Port support for parsing .debug_aranges section from LLDB and wire it up to llvm-dwarfdump.

This is only one half of it, the part that caches address ranges from the DIEs when .debug_aranges is
not available will be ported soon.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@139680 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
Benjamin Kramer 2011-09-14 01:09:52 +00:00
parent 836623420d
commit 358f4fd9ee
8 changed files with 556 additions and 3 deletions

View File

@ -4,6 +4,8 @@ add_llvm_library(LLVMDebugInfo
DWARFCompileUnit.cpp DWARFCompileUnit.cpp
DWARFContext.cpp DWARFContext.cpp
DWARFDebugAbbrev.cpp DWARFDebugAbbrev.cpp
DWARFDebugArangeSet.cpp
DWARFDebugAranges.cpp
DWARFDebugInfoEntry.cpp DWARFDebugInfoEntry.cpp
DWARFFormValue.cpp DWARFFormValue.cpp
) )

View File

@ -8,12 +8,23 @@
//===----------------------------------------------------------------------===// //===----------------------------------------------------------------------===//
#include "DWARFContext.h" #include "DWARFContext.h"
#include "llvm/Support/raw_ostream.h"
using namespace llvm; using namespace llvm;
void DWARFContext::dump(raw_ostream &OS) { void DWARFContext::dump(raw_ostream &OS) {
OS << ".debug_abbrev contents:\n";
getDebugAbbrev()->dump(OS); getDebugAbbrev()->dump(OS);
OS << "\n.debug_info contents:\n";
for (unsigned i = 0, e = getNumCompileUnits(); i != e; ++i) for (unsigned i = 0, e = getNumCompileUnits(); i != e; ++i)
getCompileUnitAtIndex(i)->dump(OS); getCompileUnitAtIndex(i)->dump(OS);
OS << "\n.debug_aranges contents:\n";
DataExtractor arangesData(getARangeSection(), isLittleEndian(), 0);
uint32_t offset = 0;
DWARFDebugArangeSet set;
while (set.extract(arangesData, &offset))
set.dump(OS);
} }
const DWARFDebugAbbrev *DWARFContext::getDebugAbbrev() { const DWARFDebugAbbrev *DWARFContext::getDebugAbbrev() {
@ -27,6 +38,17 @@ const DWARFDebugAbbrev *DWARFContext::getDebugAbbrev() {
return Abbrev.get(); return Abbrev.get();
} }
const DWARFDebugAranges *DWARFContext::getDebugAranges() {
if (Aranges)
return Aranges.get();
DataExtractor arangesData(getARangeSection(), isLittleEndian(), 0);
Aranges.reset(new DWARFDebugAranges());
Aranges->extract(arangesData);
return Aranges.get();
}
void DWARFContext::parseCompileUnits() { void DWARFContext::parseCompileUnits() {
uint32_t offset = 0; uint32_t offset = 0;
const DataExtractor &debug_info_data = DataExtractor(getInfoSection(), const DataExtractor &debug_info_data = DataExtractor(getInfoSection(),

View File

@ -11,14 +11,13 @@
#define LLVM_DEBUGINFO_DWARFCONTEXT_H #define LLVM_DEBUGINFO_DWARFCONTEXT_H
#include "DWARFCompileUnit.h" #include "DWARFCompileUnit.h"
#include "DWARFDebugAranges.h"
#include "llvm/DebugInfo/DIContext.h" #include "llvm/DebugInfo/DIContext.h"
#include "llvm/ADT/OwningPtr.h" #include "llvm/ADT/OwningPtr.h"
#include "llvm/ADT/SmallVector.h" #include "llvm/ADT/SmallVector.h"
namespace llvm { namespace llvm {
class DWARFDebugAbbrev;
/// DWARFContext /// DWARFContext
/// This data structure is the top level entity that deals with dwarf debug /// This data structure is the top level entity that deals with dwarf debug
/// information parsing. The actual data is supplied through pure virtual /// information parsing. The actual data is supplied through pure virtual
@ -28,6 +27,7 @@ class DWARFContext : public DIContext {
SmallVector<DWARFCompileUnit, 1> CUs; SmallVector<DWARFCompileUnit, 1> CUs;
OwningPtr<DWARFDebugAbbrev> Abbrev; OwningPtr<DWARFDebugAbbrev> Abbrev;
OwningPtr<DWARFDebugAranges> Aranges;
DWARFContext(DWARFContext &); // = delete DWARFContext(DWARFContext &); // = delete
DWARFContext &operator=(DWARFContext &); // = delete DWARFContext &operator=(DWARFContext &); // = delete
@ -54,6 +54,9 @@ public:
/// Get a pointer to the parsed DebugAbbrev object. /// Get a pointer to the parsed DebugAbbrev object.
const DWARFDebugAbbrev *getDebugAbbrev(); const DWARFDebugAbbrev *getDebugAbbrev();
/// Get a pointer to the parsed DebugAranges object.
const DWARFDebugAranges *getDebugAranges();
bool isLittleEndian() const { return IsLittleEndian; } bool isLittleEndian() const { return IsLittleEndian; }
virtual StringRef getInfoSection() = 0; virtual StringRef getInfoSection() = 0;

View File

@ -0,0 +1,140 @@
//===-- DWARFDebugArangeSet.cpp -------------------------------------------===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
#include "DWARFDebugArangeSet.h"
#include "llvm/Support/Format.h"
#include "llvm/Support/raw_ostream.h"
#include <cassert>
using namespace llvm;
void DWARFDebugArangeSet::clear() {
Offset = -1U;
std::memset(&Header, 0, sizeof(Header));
ArangeDescriptors.clear();
}
void DWARFDebugArangeSet::compact() {
if (ArangeDescriptors.empty())
return;
// Iterate through all arange descriptors and combine any ranges that
// overlap or have matching boundaries. The ArangeDescriptors are assumed
// to be in ascending order.
uint32_t i = 0;
while (i + 1 < ArangeDescriptors.size()) {
if (ArangeDescriptors[i].getEndAddress() >= ArangeDescriptors[i+1].Address){
// The current range ends at or exceeds the start of the next address
// range. Compute the max end address between the two and use that to
// make the new length.
const uint64_t max_end_addr =
std::max(ArangeDescriptors[i].getEndAddress(),
ArangeDescriptors[i+1].getEndAddress());
ArangeDescriptors[i].Length = max_end_addr - ArangeDescriptors[i].Address;
// Now remove the next entry as it was just combined with the previous one
ArangeDescriptors.erase(ArangeDescriptors.begin()+i+1);
} else {
// Discontiguous address range, just proceed to the next one.
++i;
}
}
}
bool
DWARFDebugArangeSet::extract(DataExtractor data, uint32_t *offset_ptr) {
if (data.isValidOffset(*offset_ptr)) {
ArangeDescriptors.clear();
Offset = *offset_ptr;
// 7.20 Address Range Table
//
// Each set of entries in the table of address ranges contained in
// the .debug_aranges section begins with a header consisting of: a
// 4-byte length containing the length of the set of entries for this
// compilation unit, not including the length field itself; a 2-byte
// version identifier containing the value 2 for DWARF Version 2; a
// 4-byte offset into the.debug_infosection; a 1-byte unsigned integer
// containing the size in bytes of an address (or the offset portion of
// an address for segmented addressing) on the target system; and a
// 1-byte unsigned integer containing the size in bytes of a segment
// descriptor on the target system. This header is followed by a series
// of tuples. Each tuple consists of an address and a length, each in
// the size appropriate for an address on the target architecture.
Header.Length = data.getU32(offset_ptr);
Header.Version = data.getU16(offset_ptr);
Header.CuOffset = data.getU32(offset_ptr);
Header.AddrSize = data.getU8(offset_ptr);
Header.SegSize = data.getU8(offset_ptr);
// The first tuple following the header in each set begins at an offset
// that is a multiple of the size of a single tuple (that is, twice the
// size of an address). The header is padded, if necessary, to the
// appropriate boundary.
const uint32_t header_size = *offset_ptr - Offset;
const uint32_t tuple_size = Header.AddrSize * 2;
uint32_t first_tuple_offset = 0;
while (first_tuple_offset < header_size)
first_tuple_offset += tuple_size;
*offset_ptr = Offset + first_tuple_offset;
Descriptor arangeDescriptor;
assert(sizeof(arangeDescriptor.Address) == sizeof(arangeDescriptor.Length));
assert(sizeof(arangeDescriptor.Address) >= Header.AddrSize);
while (data.isValidOffset(*offset_ptr)) {
arangeDescriptor.Address = data.getUnsigned(offset_ptr, Header.AddrSize);
arangeDescriptor.Length = data.getUnsigned(offset_ptr, Header.AddrSize);
// Each set of tuples is terminated by a 0 for the address and 0
// for the length.
if (arangeDescriptor.Address || arangeDescriptor.Length)
ArangeDescriptors.push_back(arangeDescriptor);
else
break; // We are done if we get a zero address and length
}
return !ArangeDescriptors.empty();
}
return false;
}
void DWARFDebugArangeSet::dump(raw_ostream &OS) const {
OS << format("Address Range Header: length = 0x%8.8x, version = 0x%4.4x, ",
Header.Length, Header.Version)
<< format("cu_offset = 0x%8.8x, addr_size = 0x%2.2x, seg_size = 0x%2.2x\n",
Header.CuOffset, Header.AddrSize, Header.SegSize);
const uint32_t hex_width = Header.AddrSize * 2;
for (DescriptorConstIter pos = ArangeDescriptors.begin(),
end = ArangeDescriptors.end(); pos != end; ++pos)
OS << format("[0x%*.*llx -", hex_width, hex_width, pos->Address)
<< format(" 0x%*.*llx)\n", hex_width, hex_width, pos->getEndAddress());
}
class DescriptorContainsAddress {
const uint64_t Address;
public:
DescriptorContainsAddress(uint64_t address) : Address(address) {}
bool operator()(const DWARFDebugArangeSet::Descriptor &desc) const {
return Address >= desc.Address && Address < (desc.Address + desc.Length);
}
};
uint32_t DWARFDebugArangeSet::findAddress(uint64_t address) const {
DescriptorConstIter end = ArangeDescriptors.end();
DescriptorConstIter pos =
std::find_if(ArangeDescriptors.begin(), end, // Range
DescriptorContainsAddress(address)); // Predicate
if (pos != end)
return Header.CuOffset;
return -1U;
}

View File

@ -0,0 +1,75 @@
//===-- DWARFDebugArangeSet.h -----------------------------------*- C++ -*-===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
#ifndef LLVM_DEBUGINFO_DWARFDEBUGARANGESET_H
#define LLVM_DEBUGINFO_DWARFDEBUGARANGESET_H
#include "llvm/Support/DataExtractor.h"
#include <vector>
namespace llvm {
class raw_ostream;
class DWARFDebugArangeSet {
public:
struct Header {
// The total length of the entries for that set, not including the length
// field itself.
uint32_t Length;
// The DWARF version number.
uint16_t Version;
// The offset from the beginning of the .debug_info section of the
// compilation unit entry referenced by the table.
uint32_t CuOffset;
// The size in bytes of an address on the target architecture. For segmented
// addressing, this is the size of the offset portion of the address.
uint8_t AddrSize;
// The size in bytes of a segment descriptor on the target architecture.
// If the target system uses a flat address space, this value is 0.
uint8_t SegSize;
};
struct Descriptor {
uint64_t Address;
uint64_t Length;
uint64_t getEndAddress() const { return Address + Length; }
};
private:
typedef std::vector<Descriptor> DescriptorColl;
typedef DescriptorColl::iterator DescriptorIter;
typedef DescriptorColl::const_iterator DescriptorConstIter;
uint32_t Offset;
Header Header;
DescriptorColl ArangeDescriptors;
public:
DWARFDebugArangeSet() { clear(); }
void clear();
void compact();
bool extract(DataExtractor data, uint32_t *offset_ptr);
void dump(raw_ostream &OS) const;
uint32_t getCompileUnitDIEOffset() const { return Header.CuOffset; }
uint32_t getOffsetOfNextEntry() const { return Offset + Header.Length + 4; }
uint32_t findAddress(uint64_t address) const;
uint32_t getNumDescriptors() const { return ArangeDescriptors.size(); }
const struct Header &getHeader() const { return Header; }
const Descriptor *getDescriptor(uint32_t i) const {
if (i < ArangeDescriptors.size())
return &ArangeDescriptors[i];
return NULL;
}
};
}
#endif

View File

@ -0,0 +1,210 @@
//===-- DWARFDebugAranges.cpp -----------------------------------*- C++ -*-===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
#include "DWARFDebugAranges.h"
#include "DWARFCompileUnit.h"
#include "DWARFContext.h"
#include "llvm/Support/Format.h"
#include "llvm/Support/raw_ostream.h"
#include <algorithm>
#include <cassert>
using namespace llvm;
// Compare function DWARFDebugAranges::Range structures
static bool RangeLessThan(const DWARFDebugAranges::Range &range1,
const DWARFDebugAranges::Range &range2) {
return range1.LoPC < range2.LoPC;
}
namespace {
class CountArangeDescriptors {
public:
CountArangeDescriptors(uint32_t &count_ref) : Count(count_ref) {}
void operator()(const DWARFDebugArangeSet &set) {
Count += set.getNumDescriptors();
}
uint32_t &Count;
};
class AddArangeDescriptors {
public:
AddArangeDescriptors(DWARFDebugAranges::RangeColl &ranges)
: RangeCollection(ranges) {}
void operator()(const DWARFDebugArangeSet& set) {
const DWARFDebugArangeSet::Descriptor* arange_desc_ptr;
DWARFDebugAranges::Range range;
range.Offset = set.getCompileUnitDIEOffset();
for (uint32_t i=0; (arange_desc_ptr = set.getDescriptor(i)) != NULL; ++i){
range.LoPC = arange_desc_ptr->Address;
range.Length = arange_desc_ptr->Length;
// Insert each item in increasing address order so binary searching
// can later be done!
DWARFDebugAranges::RangeColl::iterator insert_pos =
std::lower_bound(RangeCollection.begin(), RangeCollection.end(),
range, RangeLessThan);
RangeCollection.insert(insert_pos, range);
}
}
DWARFDebugAranges::RangeColl& RangeCollection;
};
}
bool DWARFDebugAranges::extract(DataExtractor debug_aranges_data) {
if (debug_aranges_data.isValidOffset(0)) {
uint32_t offset = 0;
typedef std::vector<DWARFDebugArangeSet> SetCollection;
typedef SetCollection::const_iterator SetCollectionIter;
SetCollection sets;
DWARFDebugArangeSet set;
Range range;
while (set.extract(debug_aranges_data, &offset))
sets.push_back(set);
uint32_t count = 0;
std::for_each(sets.begin(), sets.end(), CountArangeDescriptors(count));
if (count > 0) {
Aranges.reserve(count);
AddArangeDescriptors range_adder(Aranges);
std::for_each(sets.begin(), sets.end(), range_adder);
}
}
return false;
}
void DWARFDebugAranges::dump(raw_ostream &OS) const {
const uint32_t num_ranges = getNumRanges();
for (uint32_t i = 0; i < num_ranges; ++i) {
const Range &range = Aranges[i];
OS << format("0x%8.8x: [0x%8.8llx - 0x%8.8llx)", range.Offset,
(uint64_t)range.LoPC, (uint64_t)range.HiPC());
}
}
void DWARFDebugAranges::Range::dump(raw_ostream &OS) const {
OS << format("{0x%8.8x}: [0x%8.8llx - 0x%8.8llx)\n", Offset, LoPC, HiPC());
}
void DWARFDebugAranges::appendRange(uint32_t offset, uint64_t low_pc,
uint64_t high_pc) {
if (!Aranges.empty()) {
if (Aranges.back().Offset == offset && Aranges.back().HiPC() == low_pc) {
Aranges.back().setHiPC(high_pc);
return;
}
}
Aranges.push_back(Range(low_pc, high_pc, offset));
}
void DWARFDebugAranges::sort(bool minimize, uint32_t n) {
const size_t orig_arange_size = Aranges.size();
// Size of one? If so, no sorting is needed
if (orig_arange_size <= 1)
return;
// Sort our address range entries
std::stable_sort(Aranges.begin(), Aranges.end(), RangeLessThan);
if (!minimize)
return;
// Most address ranges are contiguous from function to function
// so our new ranges will likely be smaller. We calculate the size
// of the new ranges since although std::vector objects can be resized,
// the will never reduce their allocated block size and free any excesss
// memory, so we might as well start a brand new collection so it is as
// small as possible.
// First calculate the size of the new minimal arange vector
// so we don't have to do a bunch of re-allocations as we
// copy the new minimal stuff over to the new collection.
size_t minimal_size = 1;
for (size_t i = 1; i < orig_arange_size; ++i) {
if (!Range::SortedOverlapCheck(Aranges[i-1], Aranges[i], n))
++minimal_size;
}
// If the sizes are the same, then no consecutive aranges can be
// combined, we are done.
if (minimal_size == orig_arange_size)
return;
// Else, make a new RangeColl that _only_ contains what we need.
RangeColl minimal_aranges;
minimal_aranges.resize(minimal_size);
uint32_t j = 0;
minimal_aranges[j] = Aranges[0];
for (size_t i = 1; i < orig_arange_size; ++i) {
if(Range::SortedOverlapCheck (minimal_aranges[j], Aranges[i], n)) {
minimal_aranges[j].setHiPC (Aranges[i].HiPC());
} else {
// Only increment j if we aren't merging
minimal_aranges[++j] = Aranges[i];
}
}
assert (j+1 == minimal_size);
// Now swap our new minimal aranges into place. The local
// minimal_aranges will then contian the old big collection
// which will get freed.
minimal_aranges.swap(Aranges);
}
uint32_t DWARFDebugAranges::findAddress(uint64_t address) const {
if (!Aranges.empty()) {
Range range(address);
RangeCollIterator begin = Aranges.begin();
RangeCollIterator end = Aranges.end();
RangeCollIterator pos = lower_bound(begin, end, range, RangeLessThan);
if (pos != end && pos->LoPC <= address && address < pos->HiPC()) {
return pos->Offset;
} else if (pos != begin) {
--pos;
if (pos->LoPC <= address && address < pos->HiPC())
return (*pos).Offset;
}
}
return -1U;
}
bool
DWARFDebugAranges::allRangesAreContiguous(uint64_t &LoPC, uint64_t &HiPC) const{
if (Aranges.empty())
return false;
uint64_t next_addr = 0;
RangeCollIterator begin = Aranges.begin();
for (RangeCollIterator pos = begin, end = Aranges.end(); pos != end;
++pos) {
if (pos != begin && pos->LoPC != next_addr)
return false;
next_addr = pos->HiPC();
}
// We checked for empty at the start of function so front() will be valid.
LoPC = Aranges.front().LoPC;
// We checked for empty at the start of function so back() will be valid.
HiPC = Aranges.back().HiPC();
return true;
}
bool DWARFDebugAranges::getMaxRange(uint64_t &LoPC, uint64_t &HiPC) const {
if (Aranges.empty())
return false;
// We checked for empty at the start of function so front() will be valid.
LoPC = Aranges.front().LoPC;
// We checked for empty at the start of function so back() will be valid.
HiPC = Aranges.back().HiPC();
return true;
}

View File

@ -0,0 +1,97 @@
//===-- DWARFDebugAranges.h -------------------------------------*- C++ -*-===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
#ifndef LLVM_DEBUGINFO_DWARFDEBUGARANGES_H
#define LLVM_DEBUGINFO_DWARFDEBUGARANGES_H
#include "DWARFDebugArangeSet.h"
#include <list>
namespace llvm {
class DWARFContext;
class DWARFDebugAranges {
public:
struct Range {
explicit Range(uint64_t lo = -1ULL, uint64_t hi = -1ULL,
uint32_t off = -1U)
: LoPC(lo), Length(hi-lo), Offset(off) {}
void clear() {
LoPC = -1ULL;
Length = 0;
Offset = -1U;
}
void setHiPC(uint64_t HiPC) {
if (HiPC == -1ULL || HiPC <= LoPC)
Length = 0;
else
Length = HiPC - LoPC;
}
uint64_t HiPC() const {
if (Length)
return LoPC + Length;
return -1ULL;
}
bool isValidRange() const { return Length > 0; }
static bool SortedOverlapCheck(const Range &curr_range,
const Range &next_range, uint32_t n) {
if (curr_range.Offset != next_range.Offset)
return false;
return curr_range.HiPC() + n >= next_range.LoPC;
}
bool contains(const Range &range) const {
return LoPC <= range.LoPC && range.HiPC() <= HiPC();
}
void dump(raw_ostream &OS) const;
uint64_t LoPC; // Start of address range
uint32_t Length; // End of address range (not including this address)
uint32_t Offset; // Offset of the compile unit or die
};
void clear() { Aranges.clear(); }
bool allRangesAreContiguous(uint64_t& LoPC, uint64_t& HiPC) const;
bool getMaxRange(uint64_t& LoPC, uint64_t& HiPC) const;
bool extract(DataExtractor debug_aranges_data);
// Use append range multiple times and then call sort
void appendRange(uint32_t cu_offset, uint64_t low_pc, uint64_t high_pc);
void sort(bool minimize, uint32_t n);
const Range *rangeAtIndex(uint32_t idx) const {
if (idx < Aranges.size())
return &Aranges[idx];
return NULL;
}
void dump(raw_ostream &OS) const;
uint32_t findAddress(uint64_t address) const;
bool isEmpty() const { return Aranges.empty(); }
uint32_t getNumRanges() const { return Aranges.size(); }
uint32_t offsetAtIndex(uint32_t idx) const {
if (idx < Aranges.size())
return Aranges[idx].Offset;
return -1U;
}
typedef std::vector<Range> RangeColl;
typedef RangeColl::const_iterator RangeCollIterator;
private:
RangeColl Aranges;
};
}
#endif

View File

@ -52,6 +52,7 @@ static void DumpInput(const StringRef &Filename) {
StringRef DebugInfoSection; StringRef DebugInfoSection;
StringRef DebugAbbrevSection; StringRef DebugAbbrevSection;
StringRef DebugLineSection; StringRef DebugLineSection;
StringRef DebugArangesSection;
error_code ec; error_code ec;
for (ObjectFile::section_iterator i = Obj->begin_sections(), for (ObjectFile::section_iterator i = Obj->begin_sections(),
@ -67,11 +68,14 @@ static void DumpInput(const StringRef &Filename) {
DebugAbbrevSection = data; DebugAbbrevSection = data;
else if (name.endswith("debug_line")) else if (name.endswith("debug_line"))
DebugLineSection = data; DebugLineSection = data;
else if (name.endswith("debug_aranges"))
DebugArangesSection = data;
} }
OwningPtr<DIContext> dictx(DIContext::getDWARFContext(/*FIXME*/true, OwningPtr<DIContext> dictx(DIContext::getDWARFContext(/*FIXME*/true,
DebugInfoSection, DebugInfoSection,
DebugAbbrevSection)); DebugAbbrevSection,
DebugArangesSection));
dictx->dump(outs()); dictx->dump(outs());
} }