From: Chris Lattner Date: Sat, 19 Jan 2013 18:19:39 +0000 (+0000) Subject: BitstreamReader hasn't aged well. It's been hacked on by various people and X-Git-Url: http://demsky.eecs.uci.edu/git/?a=commitdiff_plain;h=312c7d9dc126e90ca1f40d30ba9a3ba22ade35c4;p=oota-llvm.git BitstreamReader hasn't aged well. It's been hacked on by various people and has past the point of making sense. Lets tidy things up: first step, moving a ton of big functions out of line. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@172904 91177308-0d34-0410-b5e6-96231b3b80d8 --- diff --git a/include/llvm/Bitcode/BitstreamReader.h b/include/llvm/Bitcode/BitstreamReader.h index 3a81de81a4f..bdc71ae406c 100644 --- a/include/llvm/Bitcode/BitstreamReader.h +++ b/include/llvm/Bitcode/BitstreamReader.h @@ -179,50 +179,10 @@ public: freeState(); } - void operator=(const BitstreamCursor &RHS) { - freeState(); - - BitStream = RHS.BitStream; - NextChar = RHS.NextChar; - CurWord = RHS.CurWord; - BitsInCurWord = RHS.BitsInCurWord; - CurCodeSize = RHS.CurCodeSize; - - // Copy abbreviations, and bump ref counts. - CurAbbrevs = RHS.CurAbbrevs; - for (unsigned i = 0, e = static_cast(CurAbbrevs.size()); - i != e; ++i) - CurAbbrevs[i]->addRef(); - - // Copy block scope and bump ref counts. - BlockScope = RHS.BlockScope; - for (unsigned S = 0, e = static_cast(BlockScope.size()); - S != e; ++S) { - std::vector &Abbrevs = BlockScope[S].PrevAbbrevs; - for (unsigned i = 0, e = static_cast(Abbrevs.size()); - i != e; ++i) - Abbrevs[i]->addRef(); - } - } - - void freeState() { - // Free all the Abbrevs. - for (unsigned i = 0, e = static_cast(CurAbbrevs.size()); - i != e; ++i) - CurAbbrevs[i]->dropRef(); - CurAbbrevs.clear(); - - // Free all the Abbrevs in the block scope. - for (unsigned S = 0, e = static_cast(BlockScope.size()); - S != e; ++S) { - std::vector &Abbrevs = BlockScope[S].PrevAbbrevs; - for (unsigned i = 0, e = static_cast(Abbrevs.size()); - i != e; ++i) - Abbrevs[i]->dropRef(); - } - BlockScope.clear(); - } + void operator=(const BitstreamCursor &RHS); + void freeState(); + /// GetAbbrevIDWidth - Return the number of bits used to encode an abbrev #. unsigned GetAbbrevIDWidth() const { return CurCodeSize; } @@ -243,12 +203,8 @@ public: } uint32_t getWord(size_t pos) { - uint8_t buf[sizeof(uint32_t)]; - memset(buf, 0xFF, sizeof(buf)); - BitStream->getBitcodeBytes().readBytes(pos, - sizeof(buf), - buf, - NULL); + uint8_t buf[4] = { 0xFF, 0xFF, 0xFF, 0xFF }; + BitStream->getBitcodeBytes().readBytes(pos, sizeof(buf), buf, NULL); return *reinterpret_cast(buf); } @@ -410,34 +366,8 @@ public: /// EnterSubBlock - Having read the ENTER_SUBBLOCK abbrevid, enter /// the block, and return true if the block has an error. - bool EnterSubBlock(unsigned BlockID, unsigned *NumWordsP = 0) { - // Save the current block's state on BlockScope. - BlockScope.push_back(Block(CurCodeSize)); - BlockScope.back().PrevAbbrevs.swap(CurAbbrevs); - - // Add the abbrevs specific to this block to the CurAbbrevs list. - if (const BitstreamReader::BlockInfo *Info = - BitStream->getBlockInfo(BlockID)) { - for (unsigned i = 0, e = static_cast(Info->Abbrevs.size()); - i != e; ++i) { - CurAbbrevs.push_back(Info->Abbrevs[i]); - CurAbbrevs.back()->addRef(); - } - } - - // Get the codesize of this block. - CurCodeSize = ReadVBR(bitc::CodeLenWidth); - SkipToWord(); - unsigned NumWords = Read(bitc::BlockSizeWidth); - if (NumWordsP) *NumWordsP = NumWords; - - // Validate that this block is sane. - if (CurCodeSize == 0 || AtEndOfStream()) - return true; - - return false; - } - + bool EnterSubBlock(unsigned BlockID, unsigned *NumWordsP = 0); + bool ReadBlockEnd() { if (BlockScope.empty()) return true; @@ -502,70 +432,8 @@ public: } unsigned ReadRecord(unsigned AbbrevID, SmallVectorImpl &Vals, - const char **BlobStart = 0, unsigned *BlobLen = 0) { - if (AbbrevID == bitc::UNABBREV_RECORD) { - unsigned Code = ReadVBR(6); - unsigned NumElts = ReadVBR(6); - for (unsigned i = 0; i != NumElts; ++i) - Vals.push_back(ReadVBR64(6)); - return Code; - } - - const BitCodeAbbrev *Abbv = getAbbrev(AbbrevID); - - for (unsigned i = 0, e = Abbv->getNumOperandInfos(); i != e; ++i) { - const BitCodeAbbrevOp &Op = Abbv->getOperandInfo(i); - if (Op.isLiteral()) { - ReadAbbreviatedLiteral(Op, Vals); - } else if (Op.getEncoding() == BitCodeAbbrevOp::Array) { - // Array case. Read the number of elements as a vbr6. - unsigned NumElts = ReadVBR(6); - - // Get the element encoding. - assert(i+2 == e && "array op not second to last?"); - const BitCodeAbbrevOp &EltEnc = Abbv->getOperandInfo(++i); - - // Read all the elements. - for (; NumElts; --NumElts) - ReadAbbreviatedField(EltEnc, Vals); - } else if (Op.getEncoding() == BitCodeAbbrevOp::Blob) { - // Blob case. Read the number of bytes as a vbr6. - unsigned NumElts = ReadVBR(6); - SkipToWord(); // 32-bit alignment - - // Figure out where the end of this blob will be including tail padding. - size_t NewEnd = NextChar+((NumElts+3)&~3); - - // If this would read off the end of the bitcode file, just set the - // record to empty and return. - if (!canSkipToPos(NewEnd)) { - Vals.append(NumElts, 0); - NextChar = BitStream->getBitcodeBytes().getExtent(); - break; - } - - // Otherwise, read the number of bytes. If we can return a reference to - // the data, do so to avoid copying it. - if (BlobStart) { - *BlobStart = (const char*)BitStream->getBitcodeBytes().getPointer( - NextChar, NumElts); - *BlobLen = NumElts; - } else { - for (; NumElts; ++NextChar, --NumElts) - Vals.push_back(getByte(NextChar)); - } - // Skip over tail padding. - NextChar = NewEnd; - } else { - ReadAbbreviatedField(Op, Vals); - } - } - - unsigned Code = (unsigned)Vals[0]; - Vals.erase(Vals.begin()); - return Code; - } - + const char **BlobStart = 0, unsigned *BlobLen = 0); + unsigned ReadRecord(unsigned AbbrevID, SmallVectorImpl &Vals, const char *&BlobStart, unsigned &BlobLen) { return ReadRecord(AbbrevID, Vals, &BlobStart, &BlobLen); @@ -576,91 +444,9 @@ public: // Abbrev Processing //===--------------------------------------------------------------------===// - void ReadAbbrevRecord() { - BitCodeAbbrev *Abbv = new BitCodeAbbrev(); - unsigned NumOpInfo = ReadVBR(5); - for (unsigned i = 0; i != NumOpInfo; ++i) { - bool IsLiteral = Read(1) ? true : false; - if (IsLiteral) { - Abbv->Add(BitCodeAbbrevOp(ReadVBR64(8))); - continue; - } - - BitCodeAbbrevOp::Encoding E = (BitCodeAbbrevOp::Encoding)Read(3); - if (BitCodeAbbrevOp::hasEncodingData(E)) - Abbv->Add(BitCodeAbbrevOp(E, ReadVBR64(5))); - else - Abbv->Add(BitCodeAbbrevOp(E)); - } - CurAbbrevs.push_back(Abbv); - } - -public: - - bool ReadBlockInfoBlock() { - // If this is the second stream to get to the block info block, skip it. - if (BitStream->hasBlockInfoRecords()) - return SkipBlock(); - - if (EnterSubBlock(bitc::BLOCKINFO_BLOCK_ID)) return true; - - SmallVector Record; - BitstreamReader::BlockInfo *CurBlockInfo = 0; - - // Read all the records for this module. - while (1) { - unsigned Code = ReadCode(); - if (Code == bitc::END_BLOCK) - return ReadBlockEnd(); - if (Code == bitc::ENTER_SUBBLOCK) { - ReadSubBlockID(); - if (SkipBlock()) return true; - continue; - } - - // Read abbrev records, associate them with CurBID. - if (Code == bitc::DEFINE_ABBREV) { - if (!CurBlockInfo) return true; - ReadAbbrevRecord(); - - // ReadAbbrevRecord installs the abbrev in CurAbbrevs. Move it to the - // appropriate BlockInfo. - BitCodeAbbrev *Abbv = CurAbbrevs.back(); - CurAbbrevs.pop_back(); - CurBlockInfo->Abbrevs.push_back(Abbv); - continue; - } - - // Read a record. - Record.clear(); - switch (ReadRecord(Code, Record)) { - default: break; // Default behavior, ignore unknown content. - case bitc::BLOCKINFO_CODE_SETBID: - if (Record.size() < 1) return true; - CurBlockInfo = &BitStream->getOrCreateBlockInfo((unsigned)Record[0]); - break; - case bitc::BLOCKINFO_CODE_BLOCKNAME: { - if (!CurBlockInfo) return true; - if (BitStream->isIgnoringBlockInfoNames()) break; // Ignore name. - std::string Name; - for (unsigned i = 0, e = Record.size(); i != e; ++i) - Name += (char)Record[i]; - CurBlockInfo->Name = Name; - break; - } - case bitc::BLOCKINFO_CODE_SETRECORDNAME: { - if (!CurBlockInfo) return true; - if (BitStream->isIgnoringBlockInfoNames()) break; // Ignore name. - std::string Name; - for (unsigned i = 1, e = Record.size(); i != e; ++i) - Name += (char)Record[i]; - CurBlockInfo->RecordNames.push_back(std::make_pair((unsigned)Record[0], - Name)); - break; - } - } - } - } + void ReadAbbrevRecord(); + + bool ReadBlockInfoBlock(); }; } // End llvm namespace diff --git a/lib/Bitcode/Reader/BitcodeReader.cpp b/lib/Bitcode/Reader/BitcodeReader.cpp index 3bd64a92d36..fba876936e3 100644 --- a/lib/Bitcode/Reader/BitcodeReader.cpp +++ b/lib/Bitcode/Reader/BitcodeReader.cpp @@ -6,10 +6,6 @@ // License. See LICENSE.TXT for details. // //===----------------------------------------------------------------------===// -// -// This header defines the BitcodeReader class. -// -//===----------------------------------------------------------------------===// #include "llvm/Bitcode/ReaderWriter.h" #include "BitcodeReader.h" diff --git a/lib/Bitcode/Reader/BitstreamReader.cpp b/lib/Bitcode/Reader/BitstreamReader.cpp new file mode 100644 index 00000000000..abc78acce72 --- /dev/null +++ b/lib/Bitcode/Reader/BitstreamReader.cpp @@ -0,0 +1,253 @@ +//===- BitstreamReader.cpp - BitstreamReader implementation ---------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// + +#include "llvm/Bitcode/BitstreamReader.h" + +using namespace llvm; + +//===----------------------------------------------------------------------===// +// BitstreamCursor implementation +//===----------------------------------------------------------------------===// + +void BitstreamCursor::operator=(const BitstreamCursor &RHS) { + freeState(); + + BitStream = RHS.BitStream; + NextChar = RHS.NextChar; + CurWord = RHS.CurWord; + BitsInCurWord = RHS.BitsInCurWord; + CurCodeSize = RHS.CurCodeSize; + + // Copy abbreviations, and bump ref counts. + CurAbbrevs = RHS.CurAbbrevs; + for (unsigned i = 0, e = static_cast(CurAbbrevs.size()); + i != e; ++i) + CurAbbrevs[i]->addRef(); + + // Copy block scope and bump ref counts. + BlockScope = RHS.BlockScope; + for (unsigned S = 0, e = static_cast(BlockScope.size()); + S != e; ++S) { + std::vector &Abbrevs = BlockScope[S].PrevAbbrevs; + for (unsigned i = 0, e = static_cast(Abbrevs.size()); + i != e; ++i) + Abbrevs[i]->addRef(); + } +} + +void BitstreamCursor::freeState() { + // Free all the Abbrevs. + for (unsigned i = 0, e = static_cast(CurAbbrevs.size()); + i != e; ++i) + CurAbbrevs[i]->dropRef(); + CurAbbrevs.clear(); + + // Free all the Abbrevs in the block scope. + for (unsigned S = 0, e = static_cast(BlockScope.size()); + S != e; ++S) { + std::vector &Abbrevs = BlockScope[S].PrevAbbrevs; + for (unsigned i = 0, e = static_cast(Abbrevs.size()); + i != e; ++i) + Abbrevs[i]->dropRef(); + } + BlockScope.clear(); +} + +/// EnterSubBlock - Having read the ENTER_SUBBLOCK abbrevid, enter +/// the block, and return true if the block has an error. +bool BitstreamCursor::EnterSubBlock(unsigned BlockID, unsigned *NumWordsP) { + // Save the current block's state on BlockScope. + BlockScope.push_back(Block(CurCodeSize)); + BlockScope.back().PrevAbbrevs.swap(CurAbbrevs); + + // Add the abbrevs specific to this block to the CurAbbrevs list. + if (const BitstreamReader::BlockInfo *Info = + BitStream->getBlockInfo(BlockID)) { + for (unsigned i = 0, e = static_cast(Info->Abbrevs.size()); + i != e; ++i) { + CurAbbrevs.push_back(Info->Abbrevs[i]); + CurAbbrevs.back()->addRef(); + } + } + + // Get the codesize of this block. + CurCodeSize = ReadVBR(bitc::CodeLenWidth); + SkipToWord(); + unsigned NumWords = Read(bitc::BlockSizeWidth); + if (NumWordsP) *NumWordsP = NumWords; + + // Validate that this block is sane. + if (CurCodeSize == 0 || AtEndOfStream()) + return true; + + return false; +} + + +unsigned BitstreamCursor::ReadRecord(unsigned AbbrevID, + SmallVectorImpl &Vals, + const char **BlobStart, unsigned *BlobLen){ + if (AbbrevID == bitc::UNABBREV_RECORD) { + unsigned Code = ReadVBR(6); + unsigned NumElts = ReadVBR(6); + for (unsigned i = 0; i != NumElts; ++i) + Vals.push_back(ReadVBR64(6)); + return Code; + } + + const BitCodeAbbrev *Abbv = getAbbrev(AbbrevID); + + for (unsigned i = 0, e = Abbv->getNumOperandInfos(); i != e; ++i) { + const BitCodeAbbrevOp &Op = Abbv->getOperandInfo(i); + if (Op.isLiteral()) { + ReadAbbreviatedLiteral(Op, Vals); + continue; + } + + if (Op.getEncoding() != BitCodeAbbrevOp::Array && + Op.getEncoding() != BitCodeAbbrevOp::Blob) { + ReadAbbreviatedField(Op, Vals); + continue; + } + + if (Op.getEncoding() == BitCodeAbbrevOp::Array) { + // Array case. Read the number of elements as a vbr6. + unsigned NumElts = ReadVBR(6); + + // Get the element encoding. + assert(i+2 == e && "array op not second to last?"); + const BitCodeAbbrevOp &EltEnc = Abbv->getOperandInfo(++i); + + // Read all the elements. + for (; NumElts; --NumElts) + ReadAbbreviatedField(EltEnc, Vals); + continue; + } + + assert(Op.getEncoding() == BitCodeAbbrevOp::Blob); + // Blob case. Read the number of bytes as a vbr6. + unsigned NumElts = ReadVBR(6); + SkipToWord(); // 32-bit alignment + + // Figure out where the end of this blob will be including tail padding. + size_t NewEnd = NextChar+((NumElts+3)&~3); + + // If this would read off the end of the bitcode file, just set the + // record to empty and return. + if (!canSkipToPos(NewEnd)) { + Vals.append(NumElts, 0); + NextChar = BitStream->getBitcodeBytes().getExtent(); + break; + } + + // Otherwise, read the number of bytes. If we can return a reference to + // the data, do so to avoid copying it. + if (BlobStart) { + *BlobStart = (const char*)BitStream->getBitcodeBytes().getPointer( + NextChar, NumElts); + *BlobLen = NumElts; + } else { + for (; NumElts; ++NextChar, --NumElts) + Vals.push_back(getByte(NextChar)); + } + // Skip over tail padding. + NextChar = NewEnd; + } + + unsigned Code = (unsigned)Vals[0]; + Vals.erase(Vals.begin()); + return Code; +} + + +void BitstreamCursor::ReadAbbrevRecord() { + BitCodeAbbrev *Abbv = new BitCodeAbbrev(); + unsigned NumOpInfo = ReadVBR(5); + for (unsigned i = 0; i != NumOpInfo; ++i) { + bool IsLiteral = Read(1) ? true : false; + if (IsLiteral) { + Abbv->Add(BitCodeAbbrevOp(ReadVBR64(8))); + continue; + } + + BitCodeAbbrevOp::Encoding E = (BitCodeAbbrevOp::Encoding)Read(3); + if (BitCodeAbbrevOp::hasEncodingData(E)) + Abbv->Add(BitCodeAbbrevOp(E, ReadVBR64(5))); + else + Abbv->Add(BitCodeAbbrevOp(E)); + } + CurAbbrevs.push_back(Abbv); +} + +bool BitstreamCursor::ReadBlockInfoBlock() { + // If this is the second stream to get to the block info block, skip it. + if (BitStream->hasBlockInfoRecords()) + return SkipBlock(); + + if (EnterSubBlock(bitc::BLOCKINFO_BLOCK_ID)) return true; + + SmallVector Record; + BitstreamReader::BlockInfo *CurBlockInfo = 0; + + // Read all the records for this module. + while (1) { + unsigned Code = ReadCode(); + if (Code == bitc::END_BLOCK) + return ReadBlockEnd(); + if (Code == bitc::ENTER_SUBBLOCK) { + ReadSubBlockID(); + if (SkipBlock()) return true; + continue; + } + + // Read abbrev records, associate them with CurBID. + if (Code == bitc::DEFINE_ABBREV) { + if (!CurBlockInfo) return true; + ReadAbbrevRecord(); + + // ReadAbbrevRecord installs the abbrev in CurAbbrevs. Move it to the + // appropriate BlockInfo. + BitCodeAbbrev *Abbv = CurAbbrevs.back(); + CurAbbrevs.pop_back(); + CurBlockInfo->Abbrevs.push_back(Abbv); + continue; + } + + // Read a record. + Record.clear(); + switch (ReadRecord(Code, Record)) { + default: break; // Default behavior, ignore unknown content. + case bitc::BLOCKINFO_CODE_SETBID: + if (Record.size() < 1) return true; + CurBlockInfo = &BitStream->getOrCreateBlockInfo((unsigned)Record[0]); + break; + case bitc::BLOCKINFO_CODE_BLOCKNAME: { + if (!CurBlockInfo) return true; + if (BitStream->isIgnoringBlockInfoNames()) break; // Ignore name. + std::string Name; + for (unsigned i = 0, e = Record.size(); i != e; ++i) + Name += (char)Record[i]; + CurBlockInfo->Name = Name; + break; + } + case bitc::BLOCKINFO_CODE_SETRECORDNAME: { + if (!CurBlockInfo) return true; + if (BitStream->isIgnoringBlockInfoNames()) break; // Ignore name. + std::string Name; + for (unsigned i = 1, e = Record.size(); i != e; ++i) + Name += (char)Record[i]; + CurBlockInfo->RecordNames.push_back(std::make_pair((unsigned)Record[0], + Name)); + break; + } + } + } +} + + diff --git a/lib/Bitcode/Reader/CMakeLists.txt b/lib/Bitcode/Reader/CMakeLists.txt index dfe7e1065c7..f614c9fd4a0 100644 --- a/lib/Bitcode/Reader/CMakeLists.txt +++ b/lib/Bitcode/Reader/CMakeLists.txt @@ -1,6 +1,7 @@ add_llvm_library(LLVMBitReader BitReader.cpp BitcodeReader.cpp + BitstreamReader.cpp ) add_dependencies(LLVMBitReader intrinsics_gen)