diff options
author | Kate Stone <katherine.stone@apple.com> | 2016-09-06 20:57:50 +0000 |
---|---|---|
committer | Kate Stone <katherine.stone@apple.com> | 2016-09-06 20:57:50 +0000 |
commit | b9c1b51e45b845debb76d8658edabca70ca56079 (patch) | |
tree | dfcb5a13ef2b014202340f47036da383eaee74aa /lldb/source/Plugins/SymbolFile/DWARF/HashedNameToDIE.cpp | |
parent | d5aa73376966339caad04013510626ec2e42c760 (diff) | |
download | llvm-b9c1b51e45b845debb76d8658edabca70ca56079.zip llvm-b9c1b51e45b845debb76d8658edabca70ca56079.tar.gz llvm-b9c1b51e45b845debb76d8658edabca70ca56079.tar.bz2 |
*** This commit represents a complete reformatting of the LLDB source code
*** to conform to clang-format’s LLVM style. This kind of mass change has
*** two obvious implications:
Firstly, merging this particular commit into a downstream fork may be a huge
effort. Alternatively, it may be worth merging all changes up to this commit,
performing the same reformatting operation locally, and then discarding the
merge for this particular commit. The commands used to accomplish this
reformatting were as follows (with current working directory as the root of
the repository):
find . \( -iname "*.c" -or -iname "*.cpp" -or -iname "*.h" -or -iname "*.mm" \) -exec clang-format -i {} +
find . -iname "*.py" -exec autopep8 --in-place --aggressive --aggressive {} + ;
The version of clang-format used was 3.9.0, and autopep8 was 1.2.4.
Secondly, “blame” style tools will generally point to this commit instead of
a meaningful prior commit. There are alternatives available that will attempt
to look through this change and find the appropriate prior commit. YMMV.
llvm-svn: 280751
Diffstat (limited to 'lldb/source/Plugins/SymbolFile/DWARF/HashedNameToDIE.cpp')
-rw-r--r-- | lldb/source/Plugins/SymbolFile/DWARF/HashedNameToDIE.cpp | 1249 |
1 files changed, 569 insertions, 680 deletions
diff --git a/lldb/source/Plugins/SymbolFile/DWARF/HashedNameToDIE.cpp b/lldb/source/Plugins/SymbolFile/DWARF/HashedNameToDIE.cpp index 12e1e89..16c0b90 100644 --- a/lldb/source/Plugins/SymbolFile/DWARF/HashedNameToDIE.cpp +++ b/lldb/source/Plugins/SymbolFile/DWARF/HashedNameToDIE.cpp @@ -9,746 +9,635 @@ #include "HashedNameToDIE.h" -void -DWARFMappedHash::ExtractDIEArray (const DIEInfoArray &die_info_array, DIEArray &die_offsets) -{ +void DWARFMappedHash::ExtractDIEArray(const DIEInfoArray &die_info_array, + DIEArray &die_offsets) { + const size_t count = die_info_array.size(); + for (size_t i = 0; i < count; ++i) + die_offsets.emplace_back(die_info_array[i].cu_offset, + die_info_array[i].offset); +} + +void DWARFMappedHash::ExtractDIEArray(const DIEInfoArray &die_info_array, + const dw_tag_t tag, + DIEArray &die_offsets) { + if (tag == 0) { + ExtractDIEArray(die_info_array, die_offsets); + } else { const size_t count = die_info_array.size(); - for (size_t i=0; i<count; ++i) - die_offsets.emplace_back(die_info_array[i].cu_offset, die_info_array[i].offset); -} - -void -DWARFMappedHash::ExtractDIEArray (const DIEInfoArray &die_info_array, - const dw_tag_t tag, - DIEArray &die_offsets) -{ - if (tag == 0) - { - ExtractDIEArray (die_info_array, die_offsets); - } - else - { - const size_t count = die_info_array.size(); - for (size_t i=0; i<count; ++i) - { - const dw_tag_t die_tag = die_info_array[i].tag; - bool tag_matches = die_tag == 0 || tag == die_tag; - if (!tag_matches) - { - if (die_tag == DW_TAG_class_type || die_tag == DW_TAG_structure_type) - tag_matches = tag == DW_TAG_structure_type || tag == DW_TAG_class_type; - } - if (tag_matches) - die_offsets.emplace_back(die_info_array[i].cu_offset, die_info_array[i].offset); - } + for (size_t i = 0; i < count; ++i) { + const dw_tag_t die_tag = die_info_array[i].tag; + bool tag_matches = die_tag == 0 || tag == die_tag; + if (!tag_matches) { + if (die_tag == DW_TAG_class_type || die_tag == DW_TAG_structure_type) + tag_matches = + tag == DW_TAG_structure_type || tag == DW_TAG_class_type; + } + if (tag_matches) + die_offsets.emplace_back(die_info_array[i].cu_offset, + die_info_array[i].offset); } + } } -void -DWARFMappedHash::ExtractDIEArray (const DIEInfoArray &die_info_array, - const dw_tag_t tag, - const uint32_t qualified_name_hash, - DIEArray &die_offsets) -{ - if (tag == 0) - { - ExtractDIEArray (die_info_array, die_offsets); - } - else - { - const size_t count = die_info_array.size(); - for (size_t i=0; i<count; ++i) - { - if (qualified_name_hash != die_info_array[i].qualified_name_hash) - continue; - const dw_tag_t die_tag = die_info_array[i].tag; - bool tag_matches = die_tag == 0 || tag == die_tag; - if (!tag_matches) - { - if (die_tag == DW_TAG_class_type || die_tag == DW_TAG_structure_type) - tag_matches = tag == DW_TAG_structure_type || tag == DW_TAG_class_type; - } - if (tag_matches) - die_offsets.emplace_back(die_info_array[i].cu_offset, die_info_array[i].offset); - } - } -} - -void -DWARFMappedHash::ExtractClassOrStructDIEArray (const DIEInfoArray &die_info_array, - bool return_implementation_only_if_available, - DIEArray &die_offsets) -{ +void DWARFMappedHash::ExtractDIEArray(const DIEInfoArray &die_info_array, + const dw_tag_t tag, + const uint32_t qualified_name_hash, + DIEArray &die_offsets) { + if (tag == 0) { + ExtractDIEArray(die_info_array, die_offsets); + } else { const size_t count = die_info_array.size(); - for (size_t i=0; i<count; ++i) - { - const dw_tag_t die_tag = die_info_array[i].tag; - if (die_tag == 0 || die_tag == DW_TAG_class_type || die_tag == DW_TAG_structure_type) - { - if (die_info_array[i].type_flags & eTypeFlagClassIsImplementation) - { - if (return_implementation_only_if_available) - { - // We found the one true definition for this class, so - // only return that - die_offsets.clear(); - die_offsets.emplace_back(die_info_array[i].cu_offset, die_info_array[i].offset); - return; - } - else - { - // Put the one true definition as the first entry so it - // matches first - die_offsets.emplace(die_offsets.begin(), die_info_array[i].cu_offset, die_info_array[i].offset); - } - } - else - { - die_offsets.emplace_back(die_info_array[i].cu_offset, die_info_array[i].offset); - } - } + for (size_t i = 0; i < count; ++i) { + if (qualified_name_hash != die_info_array[i].qualified_name_hash) + continue; + const dw_tag_t die_tag = die_info_array[i].tag; + bool tag_matches = die_tag == 0 || tag == die_tag; + if (!tag_matches) { + if (die_tag == DW_TAG_class_type || die_tag == DW_TAG_structure_type) + tag_matches = + tag == DW_TAG_structure_type || tag == DW_TAG_class_type; + } + if (tag_matches) + die_offsets.emplace_back(die_info_array[i].cu_offset, + die_info_array[i].offset); } -} - -void -DWARFMappedHash::ExtractTypesFromDIEArray (const DIEInfoArray &die_info_array, - uint32_t type_flag_mask, - uint32_t type_flag_value, - DIEArray &die_offsets) -{ - const size_t count = die_info_array.size(); - for (size_t i=0; i<count; ++i) - { - if ((die_info_array[i].type_flags & type_flag_mask) == type_flag_value) - die_offsets.emplace_back(die_info_array[i].cu_offset, die_info_array[i].offset); + } +} + +void DWARFMappedHash::ExtractClassOrStructDIEArray( + const DIEInfoArray &die_info_array, + bool return_implementation_only_if_available, DIEArray &die_offsets) { + const size_t count = die_info_array.size(); + for (size_t i = 0; i < count; ++i) { + const dw_tag_t die_tag = die_info_array[i].tag; + if (die_tag == 0 || die_tag == DW_TAG_class_type || + die_tag == DW_TAG_structure_type) { + if (die_info_array[i].type_flags & eTypeFlagClassIsImplementation) { + if (return_implementation_only_if_available) { + // We found the one true definition for this class, so + // only return that + die_offsets.clear(); + die_offsets.emplace_back(die_info_array[i].cu_offset, + die_info_array[i].offset); + return; + } else { + // Put the one true definition as the first entry so it + // matches first + die_offsets.emplace(die_offsets.begin(), die_info_array[i].cu_offset, + die_info_array[i].offset); + } + } else { + die_offsets.emplace_back(die_info_array[i].cu_offset, + die_info_array[i].offset); + } } + } +} + +void DWARFMappedHash::ExtractTypesFromDIEArray( + const DIEInfoArray &die_info_array, uint32_t type_flag_mask, + uint32_t type_flag_value, DIEArray &die_offsets) { + const size_t count = die_info_array.size(); + for (size_t i = 0; i < count; ++i) { + if ((die_info_array[i].type_flags & type_flag_mask) == type_flag_value) + die_offsets.emplace_back(die_info_array[i].cu_offset, + die_info_array[i].offset); + } +} + +const char *DWARFMappedHash::GetAtomTypeName(uint16_t atom) { + switch (atom) { + case eAtomTypeNULL: + return "NULL"; + case eAtomTypeDIEOffset: + return "die-offset"; + case eAtomTypeCUOffset: + return "cu-offset"; + case eAtomTypeTag: + return "die-tag"; + case eAtomTypeNameFlags: + return "name-flags"; + case eAtomTypeTypeFlags: + return "type-flags"; + case eAtomTypeQualNameHash: + return "qualified-name-hash"; + } + return "<invalid>"; +} + +DWARFMappedHash::DIEInfo::DIEInfo() + : cu_offset(DW_INVALID_OFFSET), offset(DW_INVALID_OFFSET), tag(0), + type_flags(0), qualified_name_hash(0) {} + +DWARFMappedHash::DIEInfo::DIEInfo(dw_offset_t c, dw_offset_t o, dw_tag_t t, + uint32_t f, uint32_t h) + : cu_offset(c), offset(o), tag(t), type_flags(f), qualified_name_hash(h) {} + +DWARFMappedHash::Prologue::Prologue(dw_offset_t _die_base_offset) + : die_base_offset(_die_base_offset), atoms(), atom_mask(0), + min_hash_data_byte_size(0), hash_data_has_fixed_byte_size(true) { + // Define an array of DIE offsets by first defining an array, + // and then define the atom type for the array, in this case + // we have an array of DIE offsets + AppendAtom(eAtomTypeDIEOffset, DW_FORM_data4); +} + +void DWARFMappedHash::Prologue::ClearAtoms() { + hash_data_has_fixed_byte_size = true; + min_hash_data_byte_size = 0; + atom_mask = 0; + atoms.clear(); +} + +bool DWARFMappedHash::Prologue::ContainsAtom(AtomType atom_type) const { + return (atom_mask & (1u << atom_type)) != 0; +} + +void DWARFMappedHash::Prologue::Clear() { + die_base_offset = 0; + ClearAtoms(); +} + +void DWARFMappedHash::Prologue::AppendAtom(AtomType type, dw_form_t form) { + atoms.push_back({type, form}); + atom_mask |= 1u << type; + switch (form) { + case DW_FORM_indirect: + case DW_FORM_exprloc: + case DW_FORM_flag_present: + case DW_FORM_ref_sig8: + assert(!"Unhandled atom form"); + break; + + case DW_FORM_string: + case DW_FORM_block: + case DW_FORM_block1: + case DW_FORM_sdata: + case DW_FORM_udata: + case DW_FORM_ref_udata: + case DW_FORM_GNU_addr_index: + case DW_FORM_GNU_str_index: + hash_data_has_fixed_byte_size = false; + LLVM_FALLTHROUGH; + case DW_FORM_flag: + case DW_FORM_data1: + case DW_FORM_ref1: + case DW_FORM_sec_offset: + min_hash_data_byte_size += 1; + break; + + case DW_FORM_block2: + hash_data_has_fixed_byte_size = false; + LLVM_FALLTHROUGH; + case DW_FORM_data2: + case DW_FORM_ref2: + min_hash_data_byte_size += 2; + break; + + case DW_FORM_block4: + hash_data_has_fixed_byte_size = false; + LLVM_FALLTHROUGH; + case DW_FORM_data4: + case DW_FORM_ref4: + case DW_FORM_addr: + case DW_FORM_ref_addr: + case DW_FORM_strp: + min_hash_data_byte_size += 4; + break; + + case DW_FORM_data8: + case DW_FORM_ref8: + min_hash_data_byte_size += 8; + break; + } } -const char * -DWARFMappedHash::GetAtomTypeName (uint16_t atom) -{ - switch (atom) - { - case eAtomTypeNULL: return "NULL"; - case eAtomTypeDIEOffset: return "die-offset"; - case eAtomTypeCUOffset: return "cu-offset"; - case eAtomTypeTag: return "die-tag"; - case eAtomTypeNameFlags: return "name-flags"; - case eAtomTypeTypeFlags: return "type-flags"; - case eAtomTypeQualNameHash: return "qualified-name-hash"; +lldb::offset_t +DWARFMappedHash::Prologue::Read(const lldb_private::DataExtractor &data, + lldb::offset_t offset) { + ClearAtoms(); + + die_base_offset = data.GetU32(&offset); + + const uint32_t atom_count = data.GetU32(&offset); + if (atom_count == 0x00060003u) { + // Old format, deal with contents of old pre-release format + while (data.GetU32(&offset)) + /* do nothing */; + + // Hardcode to the only known value for now. + AppendAtom(eAtomTypeDIEOffset, DW_FORM_data4); + } else { + for (uint32_t i = 0; i < atom_count; ++i) { + AtomType type = (AtomType)data.GetU16(&offset); + dw_form_t form = (dw_form_t)data.GetU16(&offset); + AppendAtom(type, form); } - return "<invalid>"; + } + return offset; } -DWARFMappedHash::DIEInfo::DIEInfo () : - cu_offset (DW_INVALID_OFFSET), - offset (DW_INVALID_OFFSET), - tag (0), - type_flags (0), - qualified_name_hash (0) -{ +size_t DWARFMappedHash::Prologue::GetByteSize() const { + // Add an extra count to the atoms size for the zero termination Atom that + // gets + // written to disk + return sizeof(die_base_offset) + sizeof(uint32_t) + + atoms.size() * sizeof(Atom); } -DWARFMappedHash::DIEInfo::DIEInfo (dw_offset_t c, - dw_offset_t o, - dw_tag_t t, - uint32_t f, - uint32_t h) : - cu_offset (c), - offset (o), - tag (t), - type_flags (f), - qualified_name_hash (h) -{ +size_t DWARFMappedHash::Prologue::GetMinimumHashDataByteSize() const { + return min_hash_data_byte_size; } -DWARFMappedHash::Prologue::Prologue (dw_offset_t _die_base_offset) : - die_base_offset (_die_base_offset), - atoms(), - atom_mask (0), - min_hash_data_byte_size(0), - hash_data_has_fixed_byte_size(true) -{ - // Define an array of DIE offsets by first defining an array, - // and then define the atom type for the array, in this case - // we have an array of DIE offsets - AppendAtom (eAtomTypeDIEOffset, DW_FORM_data4); +bool DWARFMappedHash::Prologue::HashDataHasFixedByteSize() const { + return hash_data_has_fixed_byte_size; } -void -DWARFMappedHash::Prologue::ClearAtoms () -{ - hash_data_has_fixed_byte_size = true; - min_hash_data_byte_size = 0; - atom_mask = 0; - atoms.clear(); +size_t DWARFMappedHash::Header::GetByteSize(const HeaderData &header_data) { + return header_data.GetByteSize(); } -bool -DWARFMappedHash::Prologue::ContainsAtom (AtomType atom_type) const -{ - return (atom_mask & (1u << atom_type)) != 0; +lldb::offset_t DWARFMappedHash::Header::Read(lldb_private::DataExtractor &data, + lldb::offset_t offset) { + offset = MappedHash::Header<Prologue>::Read(data, offset); + if (offset != UINT32_MAX) { + offset = header_data.Read(data, offset); + } + return offset; } -void -DWARFMappedHash::Prologue::Clear () -{ - die_base_offset = 0; - ClearAtoms (); -} +bool DWARFMappedHash::Header::Read(const lldb_private::DWARFDataExtractor &data, + lldb::offset_t *offset_ptr, + DIEInfo &hash_data) const { + const size_t num_atoms = header_data.atoms.size(); + if (num_atoms == 0) + return false; -void -DWARFMappedHash::Prologue::AppendAtom (AtomType type, dw_form_t form) -{ - atoms.push_back ({type, form}); - atom_mask |= 1u << type; - switch (form) - { - case DW_FORM_indirect: - case DW_FORM_exprloc: - case DW_FORM_flag_present: - case DW_FORM_ref_sig8: - assert (!"Unhandled atom form"); - break; - - case DW_FORM_string: - case DW_FORM_block: - case DW_FORM_block1: - case DW_FORM_sdata: - case DW_FORM_udata: - case DW_FORM_ref_udata: - case DW_FORM_GNU_addr_index: - case DW_FORM_GNU_str_index: - hash_data_has_fixed_byte_size = false; - LLVM_FALLTHROUGH; - case DW_FORM_flag: - case DW_FORM_data1: - case DW_FORM_ref1: - case DW_FORM_sec_offset: - min_hash_data_byte_size += 1; - break; - - case DW_FORM_block2: - hash_data_has_fixed_byte_size = false; - LLVM_FALLTHROUGH; - case DW_FORM_data2: - case DW_FORM_ref2: - min_hash_data_byte_size += 2; - break; - - case DW_FORM_block4: - hash_data_has_fixed_byte_size = false; - LLVM_FALLTHROUGH; - case DW_FORM_data4: - case DW_FORM_ref4: - case DW_FORM_addr: - case DW_FORM_ref_addr: - case DW_FORM_strp: - min_hash_data_byte_size += 4; - break; - - case DW_FORM_data8: - case DW_FORM_ref8: - min_hash_data_byte_size += 8; - break; - - } -} + for (size_t i = 0; i < num_atoms; ++i) { + DWARFFormValue form_value(NULL, header_data.atoms[i].form); -lldb::offset_t -DWARFMappedHash::Prologue::Read (const lldb_private::DataExtractor &data, - lldb::offset_t offset) -{ - ClearAtoms (); - - die_base_offset = data.GetU32 (&offset); - - const uint32_t atom_count = data.GetU32 (&offset); - if (atom_count == 0x00060003u) - { - // Old format, deal with contents of old pre-release format - while (data.GetU32(&offset)) - /* do nothing */; - - // Hardcode to the only known value for now. - AppendAtom (eAtomTypeDIEOffset, DW_FORM_data4); - } - else - { - for (uint32_t i=0; i<atom_count; ++i) - { - AtomType type = (AtomType)data.GetU16 (&offset); - dw_form_t form = (dw_form_t)data.GetU16 (&offset); - AppendAtom (type, form); - } - } - return offset; -} + if (!form_value.ExtractValue(data, offset_ptr)) + return false; -size_t -DWARFMappedHash::Prologue::GetByteSize () const -{ - // Add an extra count to the atoms size for the zero termination Atom that gets - // written to disk - return sizeof(die_base_offset) + sizeof(uint32_t) + atoms.size() * sizeof(Atom); -} + switch (header_data.atoms[i].type) { + case eAtomTypeDIEOffset: // DIE offset, check form for encoding + hash_data.offset = + (dw_offset_t)form_value.Reference(header_data.die_base_offset); + break; -size_t -DWARFMappedHash::Prologue::GetMinimumHashDataByteSize () const -{ - return min_hash_data_byte_size; -} + case eAtomTypeTag: // DW_TAG value for the DIE + hash_data.tag = (dw_tag_t)form_value.Unsigned(); + break; -bool -DWARFMappedHash::Prologue::HashDataHasFixedByteSize() const -{ - return hash_data_has_fixed_byte_size; -} + case eAtomTypeTypeFlags: // Flags from enum TypeFlags + hash_data.type_flags = (uint32_t)form_value.Unsigned(); + break; -size_t -DWARFMappedHash::Header::GetByteSize (const HeaderData &header_data) -{ - return header_data.GetByteSize(); -} + case eAtomTypeQualNameHash: // Flags from enum TypeFlags + hash_data.qualified_name_hash = form_value.Unsigned(); + break; -lldb::offset_t -DWARFMappedHash::Header::Read (lldb_private::DataExtractor &data, lldb::offset_t offset) -{ - offset = MappedHash::Header<Prologue>::Read (data, offset); - if (offset != UINT32_MAX) - { - offset = header_data.Read (data, offset); + default: + // We can always skip atoms we don't know about + break; } - return offset; + } + return true; } -bool -DWARFMappedHash::Header::Read (const lldb_private::DWARFDataExtractor &data, - lldb::offset_t *offset_ptr, - DIEInfo &hash_data) const -{ - const size_t num_atoms = header_data.atoms.size(); - if (num_atoms == 0) - return false; - - for (size_t i=0; i<num_atoms; ++i) - { - DWARFFormValue form_value (NULL, header_data.atoms[i].form); - - if (!form_value.ExtractValue(data, offset_ptr)) - return false; - - switch (header_data.atoms[i].type) - { - case eAtomTypeDIEOffset: // DIE offset, check form for encoding - hash_data.offset = (dw_offset_t)form_value.Reference (header_data.die_base_offset); - break; - - case eAtomTypeTag: // DW_TAG value for the DIE - hash_data.tag = (dw_tag_t)form_value.Unsigned (); - break; - - case eAtomTypeTypeFlags: // Flags from enum TypeFlags - hash_data.type_flags = (uint32_t)form_value.Unsigned (); - break; - - case eAtomTypeQualNameHash: // Flags from enum TypeFlags - hash_data.qualified_name_hash = form_value.Unsigned (); - break; - - default: - // We can always skip atoms we don't know about - break; - } - } - return true; -} +void DWARFMappedHash::Header::Dump(lldb_private::Stream &strm, + const DIEInfo &hash_data) const { + const size_t num_atoms = header_data.atoms.size(); + for (size_t i = 0; i < num_atoms; ++i) { + if (i > 0) + strm.PutCString(", "); + + DWARFFormValue form_value(NULL, header_data.atoms[i].form); + switch (header_data.atoms[i].type) { + case eAtomTypeDIEOffset: // DIE offset, check form for encoding + strm.Printf("{0x%8.8x}", hash_data.offset); + break; -void -DWARFMappedHash::Header::Dump (lldb_private::Stream& strm, const DIEInfo &hash_data) const -{ - const size_t num_atoms = header_data.atoms.size(); - for (size_t i=0; i<num_atoms; ++i) + case eAtomTypeTag: // DW_TAG value for the DIE { - if (i > 0) - strm.PutCString (", "); - - DWARFFormValue form_value (NULL, header_data.atoms[i].form); - switch (header_data.atoms[i].type) - { - case eAtomTypeDIEOffset: // DIE offset, check form for encoding - strm.Printf ("{0x%8.8x}", hash_data.offset); - break; - - case eAtomTypeTag: // DW_TAG value for the DIE - { - const char *tag_cstr = lldb_private::DW_TAG_value_to_name (hash_data.tag); - if (tag_cstr) - strm.PutCString (tag_cstr); - else - strm.Printf ("DW_TAG_(0x%4.4x)", hash_data.tag); - } - break; - - case eAtomTypeTypeFlags: // Flags from enum TypeFlags - strm.Printf ("0x%2.2x", hash_data.type_flags); - if (hash_data.type_flags) - { - strm.PutCString (" ("); - if (hash_data.type_flags & eTypeFlagClassIsImplementation) - strm.PutCString (" implementation"); - strm.PutCString (" )"); - } - break; - - case eAtomTypeQualNameHash: // Flags from enum TypeFlags - strm.Printf ("0x%8.8x", hash_data.qualified_name_hash); - break; - - default: - strm.Printf ("AtomType(0x%x)", header_data.atoms[i].type); - break; - } + const char *tag_cstr = lldb_private::DW_TAG_value_to_name(hash_data.tag); + if (tag_cstr) + strm.PutCString(tag_cstr); + else + strm.Printf("DW_TAG_(0x%4.4x)", hash_data.tag); + } break; + + case eAtomTypeTypeFlags: // Flags from enum TypeFlags + strm.Printf("0x%2.2x", hash_data.type_flags); + if (hash_data.type_flags) { + strm.PutCString(" ("); + if (hash_data.type_flags & eTypeFlagClassIsImplementation) + strm.PutCString(" implementation"); + strm.PutCString(" )"); + } + break; + + case eAtomTypeQualNameHash: // Flags from enum TypeFlags + strm.Printf("0x%8.8x", hash_data.qualified_name_hash); + break; + + default: + strm.Printf("AtomType(0x%x)", header_data.atoms[i].type); + break; } + } } -DWARFMappedHash::MemoryTable::MemoryTable (lldb_private::DWARFDataExtractor &table_data, - const lldb_private::DWARFDataExtractor &string_table, - const char *name) : - MappedHash::MemoryTable<uint32_t, Header, DIEInfoArray> (table_data), - m_data (table_data), - m_string_table (string_table), - m_name (name) -{ -} +DWARFMappedHash::MemoryTable::MemoryTable( + lldb_private::DWARFDataExtractor &table_data, + const lldb_private::DWARFDataExtractor &string_table, const char *name) + : MappedHash::MemoryTable<uint32_t, Header, DIEInfoArray>(table_data), + m_data(table_data), m_string_table(string_table), m_name(name) {} const char * -DWARFMappedHash::MemoryTable::GetStringForKeyType (KeyType key) const -{ - // The key in the DWARF table is the .debug_str offset for the string - return m_string_table.PeekCStr (key); -} - -bool -DWARFMappedHash::MemoryTable::ReadHashData (uint32_t hash_data_offset, HashData &hash_data) const -{ - lldb::offset_t offset = hash_data_offset; - offset += 4; // Skip string table offset that contains offset of hash name in .debug_str - const uint32_t count = m_data.GetU32 (&offset); - if (count > 0) - { - hash_data.resize(count); - for (uint32_t i=0; i<count; ++i) - { - if (!m_header.Read(m_data, &offset, hash_data[i])) - return false; - } +DWARFMappedHash::MemoryTable::GetStringForKeyType(KeyType key) const { + // The key in the DWARF table is the .debug_str offset for the string + return m_string_table.PeekCStr(key); +} + +bool DWARFMappedHash::MemoryTable::ReadHashData(uint32_t hash_data_offset, + HashData &hash_data) const { + lldb::offset_t offset = hash_data_offset; + offset += 4; // Skip string table offset that contains offset of hash name in + // .debug_str + const uint32_t count = m_data.GetU32(&offset); + if (count > 0) { + hash_data.resize(count); + for (uint32_t i = 0; i < count; ++i) { + if (!m_header.Read(m_data, &offset, hash_data[i])) + return false; } - else - hash_data.clear(); - return true; + } else + hash_data.clear(); + return true; } DWARFMappedHash::MemoryTable::Result -DWARFMappedHash::MemoryTable::GetHashDataForName (const char *name, - lldb::offset_t* hash_data_offset_ptr, - Pair &pair) const -{ - pair.key = m_data.GetU32 (hash_data_offset_ptr); - pair.value.clear(); - - // If the key is zero, this terminates our chain of HashData objects - // for this hash value. - if (pair.key == 0) - return eResultEndOfHashData; - - // There definitely should be a string for this string offset, if - // there isn't, there is something wrong, return and error - const char *strp_cstr = m_string_table.PeekCStr (pair.key); - if (strp_cstr == NULL) - { - *hash_data_offset_ptr = UINT32_MAX; - return eResultError; - } - - const uint32_t count = m_data.GetU32 (hash_data_offset_ptr); - const size_t min_total_hash_data_size = count * m_header.header_data.GetMinimumHashDataByteSize(); - if (count > 0 && m_data.ValidOffsetForDataOfSize (*hash_data_offset_ptr, min_total_hash_data_size)) - { - // We have at least one HashData entry, and we have enough - // data to parse at least "count" HashData entries. - - // First make sure the entire C string matches... - const bool match = strcmp (name, strp_cstr) == 0; - - if (!match && m_header.header_data.HashDataHasFixedByteSize()) - { - // If the string doesn't match and we have fixed size data, - // we can just add the total byte size of all HashData objects - // to the hash data offset and be done... - *hash_data_offset_ptr += min_total_hash_data_size; +DWARFMappedHash::MemoryTable::GetHashDataForName( + const char *name, lldb::offset_t *hash_data_offset_ptr, Pair &pair) const { + pair.key = m_data.GetU32(hash_data_offset_ptr); + pair.value.clear(); + + // If the key is zero, this terminates our chain of HashData objects + // for this hash value. + if (pair.key == 0) + return eResultEndOfHashData; + + // There definitely should be a string for this string offset, if + // there isn't, there is something wrong, return and error + const char *strp_cstr = m_string_table.PeekCStr(pair.key); + if (strp_cstr == NULL) { + *hash_data_offset_ptr = UINT32_MAX; + return eResultError; + } + + const uint32_t count = m_data.GetU32(hash_data_offset_ptr); + const size_t min_total_hash_data_size = + count * m_header.header_data.GetMinimumHashDataByteSize(); + if (count > 0 && + m_data.ValidOffsetForDataOfSize(*hash_data_offset_ptr, + min_total_hash_data_size)) { + // We have at least one HashData entry, and we have enough + // data to parse at least "count" HashData entries. + + // First make sure the entire C string matches... + const bool match = strcmp(name, strp_cstr) == 0; + + if (!match && m_header.header_data.HashDataHasFixedByteSize()) { + // If the string doesn't match and we have fixed size data, + // we can just add the total byte size of all HashData objects + // to the hash data offset and be done... + *hash_data_offset_ptr += min_total_hash_data_size; + } else { + // If the string does match, or we don't have fixed size data + // then we need to read the hash data as a stream. If the + // string matches we also append all HashData objects to the + // value array. + for (uint32_t i = 0; i < count; ++i) { + DIEInfo die_info; + if (m_header.Read(m_data, hash_data_offset_ptr, die_info)) { + // Only happened if the HashData of the string matched... + if (match) + pair.value.push_back(die_info); + } else { + // Something went wrong while reading the data + *hash_data_offset_ptr = UINT32_MAX; + return eResultError; } - else - { - // If the string does match, or we don't have fixed size data - // then we need to read the hash data as a stream. If the - // string matches we also append all HashData objects to the - // value array. - for (uint32_t i=0; i<count; ++i) - { - DIEInfo die_info; - if (m_header.Read(m_data, hash_data_offset_ptr, die_info)) - { - // Only happened if the HashData of the string matched... - if (match) - pair.value.push_back (die_info); - } - else - { - // Something went wrong while reading the data - *hash_data_offset_ptr = UINT32_MAX; - return eResultError; - } - } - } - // Return the correct response depending on if the string matched - // or not... - if (match) - return eResultKeyMatch; // The key (cstring) matches and we have lookup results! - else - return eResultKeyMismatch; // The key doesn't match, this function will get called - // again for the next key/value or the key terminator - // which in our case is a zero .debug_str offset. + } } + // Return the correct response depending on if the string matched + // or not... + if (match) + return eResultKeyMatch; // The key (cstring) matches and we have lookup + // results! else - { - *hash_data_offset_ptr = UINT32_MAX; - return eResultError; - } + return eResultKeyMismatch; // The key doesn't match, this function will + // get called + // again for the next key/value or the key terminator + // which in our case is a zero .debug_str offset. + } else { + *hash_data_offset_ptr = UINT32_MAX; + return eResultError; + } } DWARFMappedHash::MemoryTable::Result -DWARFMappedHash::MemoryTable::AppendHashDataForRegularExpression ( - const lldb_private::RegularExpression& regex, - lldb::offset_t* hash_data_offset_ptr, - Pair &pair) const -{ - pair.key = m_data.GetU32 (hash_data_offset_ptr); - // If the key is zero, this terminates our chain of HashData objects - // for this hash value. - if (pair.key == 0) - return eResultEndOfHashData; - - // There definitely should be a string for this string offset, if - // there isn't, there is something wrong, return and error - const char *strp_cstr = m_string_table.PeekCStr (pair.key); - if (strp_cstr == NULL) - return eResultError; - - const uint32_t count = m_data.GetU32 (hash_data_offset_ptr); - const size_t min_total_hash_data_size = count * m_header.header_data.GetMinimumHashDataByteSize(); - if (count > 0 && m_data.ValidOffsetForDataOfSize (*hash_data_offset_ptr, min_total_hash_data_size)) - { - const bool match = regex.Execute(strp_cstr); - - if (!match && m_header.header_data.HashDataHasFixedByteSize()) - { - // If the regex doesn't match and we have fixed size data, - // we can just add the total byte size of all HashData objects - // to the hash data offset and be done... - *hash_data_offset_ptr += min_total_hash_data_size; - } - else - { - // If the string does match, or we don't have fixed size data - // then we need to read the hash data as a stream. If the - // string matches we also append all HashData objects to the - // value array. - for (uint32_t i=0; i<count; ++i) - { - DIEInfo die_info; - if (m_header.Read(m_data, hash_data_offset_ptr, die_info)) - { - // Only happened if the HashData of the string matched... - if (match) - pair.value.push_back (die_info); - } - else - { - // Something went wrong while reading the data - *hash_data_offset_ptr = UINT32_MAX; - return eResultError; - } - } +DWARFMappedHash::MemoryTable::AppendHashDataForRegularExpression( + const lldb_private::RegularExpression ®ex, + lldb::offset_t *hash_data_offset_ptr, Pair &pair) const { + pair.key = m_data.GetU32(hash_data_offset_ptr); + // If the key is zero, this terminates our chain of HashData objects + // for this hash value. + if (pair.key == 0) + return eResultEndOfHashData; + + // There definitely should be a string for this string offset, if + // there isn't, there is something wrong, return and error + const char *strp_cstr = m_string_table.PeekCStr(pair.key); + if (strp_cstr == NULL) + return eResultError; + + const uint32_t count = m_data.GetU32(hash_data_offset_ptr); + const size_t min_total_hash_data_size = + count * m_header.header_data.GetMinimumHashDataByteSize(); + if (count > 0 && + m_data.ValidOffsetForDataOfSize(*hash_data_offset_ptr, + min_total_hash_data_size)) { + const bool match = regex.Execute(strp_cstr); + + if (!match && m_header.header_data.HashDataHasFixedByteSize()) { + // If the regex doesn't match and we have fixed size data, + // we can just add the total byte size of all HashData objects + // to the hash data offset and be done... + *hash_data_offset_ptr += min_total_hash_data_size; + } else { + // If the string does match, or we don't have fixed size data + // then we need to read the hash data as a stream. If the + // string matches we also append all HashData objects to the + // value array. + for (uint32_t i = 0; i < count; ++i) { + DIEInfo die_info; + if (m_header.Read(m_data, hash_data_offset_ptr, die_info)) { + // Only happened if the HashData of the string matched... + if (match) + pair.value.push_back(die_info); + } else { + // Something went wrong while reading the data + *hash_data_offset_ptr = UINT32_MAX; + return eResultError; } - // Return the correct response depending on if the string matched - // or not... - if (match) - return eResultKeyMatch; // The key (cstring) matches and we have lookup results! - else - return eResultKeyMismatch; // The key doesn't match, this function will get called - // again for the next key/value or the key terminator - // which in our case is a zero .debug_str offset. + } } + // Return the correct response depending on if the string matched + // or not... + if (match) + return eResultKeyMatch; // The key (cstring) matches and we have lookup + // results! else - { - *hash_data_offset_ptr = UINT32_MAX; - return eResultError; + return eResultKeyMismatch; // The key doesn't match, this function will + // get called + // again for the next key/value or the key terminator + // which in our case is a zero .debug_str offset. + } else { + *hash_data_offset_ptr = UINT32_MAX; + return eResultError; + } +} + +size_t DWARFMappedHash::MemoryTable::AppendAllDIEsThatMatchingRegex( + const lldb_private::RegularExpression ®ex, + DIEInfoArray &die_info_array) const { + const uint32_t hash_count = m_header.hashes_count; + Pair pair; + for (uint32_t offset_idx = 0; offset_idx < hash_count; ++offset_idx) { + lldb::offset_t hash_data_offset = GetHashDataOffset(offset_idx); + while (hash_data_offset != UINT32_MAX) { + const lldb::offset_t prev_hash_data_offset = hash_data_offset; + Result hash_result = + AppendHashDataForRegularExpression(regex, &hash_data_offset, pair); + if (prev_hash_data_offset == hash_data_offset) + break; + + // Check the result of getting our hash data + switch (hash_result) { + case eResultKeyMatch: + case eResultKeyMismatch: + // Whether we matches or not, it doesn't matter, we + // keep looking. + break; + + case eResultEndOfHashData: + case eResultError: + hash_data_offset = UINT32_MAX; + break; + } } -} - -size_t -DWARFMappedHash::MemoryTable::AppendAllDIEsThatMatchingRegex ( - const lldb_private::RegularExpression& regex, - DIEInfoArray &die_info_array) const -{ - const uint32_t hash_count = m_header.hashes_count; - Pair pair; - for (uint32_t offset_idx=0; offset_idx<hash_count; ++offset_idx) - { - lldb::offset_t hash_data_offset = GetHashDataOffset (offset_idx); - while (hash_data_offset != UINT32_MAX) - { - const lldb::offset_t prev_hash_data_offset = hash_data_offset; - Result hash_result = AppendHashDataForRegularExpression (regex, &hash_data_offset, pair); - if (prev_hash_data_offset == hash_data_offset) - break; - - // Check the result of getting our hash data - switch (hash_result) - { - case eResultKeyMatch: - case eResultKeyMismatch: - // Whether we matches or not, it doesn't matter, we - // keep looking. - break; - - case eResultEndOfHashData: - case eResultError: - hash_data_offset = UINT32_MAX; - break; - } + } + die_info_array.swap(pair.value); + return die_info_array.size(); +} + +size_t DWARFMappedHash::MemoryTable::AppendAllDIEsInRange( + const uint32_t die_offset_start, const uint32_t die_offset_end, + DIEInfoArray &die_info_array) const { + const uint32_t hash_count = m_header.hashes_count; + for (uint32_t offset_idx = 0; offset_idx < hash_count; ++offset_idx) { + bool done = false; + lldb::offset_t hash_data_offset = GetHashDataOffset(offset_idx); + while (!done && hash_data_offset != UINT32_MAX) { + KeyType key = m_data.GetU32(&hash_data_offset); + // If the key is zero, this terminates our chain of HashData objects + // for this hash value. + if (key == 0) + break; + + const uint32_t count = m_data.GetU32(&hash_data_offset); + for (uint32_t i = 0; i < count; ++i) { + DIEInfo die_info; + if (m_header.Read(m_data, &hash_data_offset, die_info)) { + if (die_info.offset == 0) + done = true; + if (die_offset_start <= die_info.offset && + die_info.offset < die_offset_end) + die_info_array.push_back(die_info); } + } } - die_info_array.swap (pair.value); - return die_info_array.size(); -} - -size_t -DWARFMappedHash::MemoryTable::AppendAllDIEsInRange (const uint32_t die_offset_start, - const uint32_t die_offset_end, - DIEInfoArray &die_info_array) const -{ - const uint32_t hash_count = m_header.hashes_count; - for (uint32_t offset_idx=0; offset_idx<hash_count; ++offset_idx) - { - bool done = false; - lldb::offset_t hash_data_offset = GetHashDataOffset (offset_idx); - while (!done && hash_data_offset != UINT32_MAX) - { - KeyType key = m_data.GetU32 (&hash_data_offset); - // If the key is zero, this terminates our chain of HashData objects - // for this hash value. - if (key == 0) - break; - - const uint32_t count = m_data.GetU32 (&hash_data_offset); - for (uint32_t i=0; i<count; ++i) - { - DIEInfo die_info; - if (m_header.Read(m_data, &hash_data_offset, die_info)) - { - if (die_info.offset == 0) - done = true; - if (die_offset_start <= die_info.offset && die_info.offset < die_offset_end) - die_info_array.push_back(die_info); - } - } - } - } - return die_info_array.size(); -} - -size_t -DWARFMappedHash::MemoryTable::FindByName (const char *name, DIEArray &die_offsets) -{ - if (!name || !name[0]) - return 0; - - DIEInfoArray die_info_array; - if (FindByName(name, die_info_array)) - DWARFMappedHash::ExtractDIEArray (die_info_array, die_offsets); - return die_info_array.size(); + } + return die_info_array.size(); } -size_t -DWARFMappedHash::MemoryTable::FindByNameAndTag (const char *name, - const dw_tag_t tag, - DIEArray &die_offsets) -{ - DIEInfoArray die_info_array; - if (FindByName(name, die_info_array)) - DWARFMappedHash::ExtractDIEArray (die_info_array, tag, die_offsets); - return die_info_array.size(); -} - -size_t -DWARFMappedHash::MemoryTable::FindByNameAndTagAndQualifiedNameHash (const char *name, - const dw_tag_t tag, - const uint32_t qualified_name_hash, - DIEArray &die_offsets) -{ - DIEInfoArray die_info_array; - if (FindByName(name, die_info_array)) - DWARFMappedHash::ExtractDIEArray (die_info_array, tag, qualified_name_hash, die_offsets); - return die_info_array.size(); -} +size_t DWARFMappedHash::MemoryTable::FindByName(const char *name, + DIEArray &die_offsets) { + if (!name || !name[0]) + return 0; -size_t -DWARFMappedHash::MemoryTable::FindCompleteObjCClassByName (const char *name, - DIEArray &die_offsets, - bool must_be_implementation) -{ - DIEInfoArray die_info_array; - if (FindByName(name, die_info_array)) - { - if (must_be_implementation && GetHeader().header_data.ContainsAtom (eAtomTypeTypeFlags)) - { - // If we have two atoms, then we have the DIE offset and - // the type flags so we can find the objective C class - // efficiently. - DWARFMappedHash::ExtractTypesFromDIEArray (die_info_array, - UINT32_MAX, - eTypeFlagClassIsImplementation, - die_offsets); - } - else - { - // We don't only want the one true definition, so try and see - // what we can find, and only return class or struct DIEs. - // If we do have the full implementation, then return it alone, - // else return all possible matches. - const bool return_implementation_only_if_available = true; - DWARFMappedHash::ExtractClassOrStructDIEArray (die_info_array, - return_implementation_only_if_available, - die_offsets); - } + DIEInfoArray die_info_array; + if (FindByName(name, die_info_array)) + DWARFMappedHash::ExtractDIEArray(die_info_array, die_offsets); + return die_info_array.size(); +} + +size_t DWARFMappedHash::MemoryTable::FindByNameAndTag(const char *name, + const dw_tag_t tag, + DIEArray &die_offsets) { + DIEInfoArray die_info_array; + if (FindByName(name, die_info_array)) + DWARFMappedHash::ExtractDIEArray(die_info_array, tag, die_offsets); + return die_info_array.size(); +} + +size_t DWARFMappedHash::MemoryTable::FindByNameAndTagAndQualifiedNameHash( + const char *name, const dw_tag_t tag, const uint32_t qualified_name_hash, + DIEArray &die_offsets) { + DIEInfoArray die_info_array; + if (FindByName(name, die_info_array)) + DWARFMappedHash::ExtractDIEArray(die_info_array, tag, qualified_name_hash, + die_offsets); + return die_info_array.size(); +} + +size_t DWARFMappedHash::MemoryTable::FindCompleteObjCClassByName( + const char *name, DIEArray &die_offsets, bool must_be_implementation) { + DIEInfoArray die_info_array; + if (FindByName(name, die_info_array)) { + if (must_be_implementation && + GetHeader().header_data.ContainsAtom(eAtomTypeTypeFlags)) { + // If we have two atoms, then we have the DIE offset and + // the type flags so we can find the objective C class + // efficiently. + DWARFMappedHash::ExtractTypesFromDIEArray(die_info_array, UINT32_MAX, + eTypeFlagClassIsImplementation, + die_offsets); + } else { + // We don't only want the one true definition, so try and see + // what we can find, and only return class or struct DIEs. + // If we do have the full implementation, then return it alone, + // else return all possible matches. + const bool return_implementation_only_if_available = true; + DWARFMappedHash::ExtractClassOrStructDIEArray( + die_info_array, return_implementation_only_if_available, die_offsets); } - return die_offsets.size(); + } + return die_offsets.size(); } -size_t -DWARFMappedHash::MemoryTable::FindByName (const char *name, DIEInfoArray &die_info_array) -{ - if (!name || !name[0]) - return 0; - - Pair kv_pair; - size_t old_size = die_info_array.size(); - if (Find (name, kv_pair)) - { - die_info_array.swap(kv_pair.value); - return die_info_array.size() - old_size; - } +size_t DWARFMappedHash::MemoryTable::FindByName(const char *name, + DIEInfoArray &die_info_array) { + if (!name || !name[0]) return 0; + + Pair kv_pair; + size_t old_size = die_info_array.size(); + if (Find(name, kv_pair)) { + die_info_array.swap(kv_pair.value); + return die_info_array.size() - old_size; + } + return 0; } |