X-Git-Url: http://demsky.eecs.uci.edu/git/?a=blobdiff_plain;f=include%2Fllvm%2FBitcode%2FBitstreamReader.h;h=3b7a9b61e6aaab5d8a14ae85cb9e20e08b2f83ad;hb=5eca075b74d62c621b160aa216b4cd50829a2cc7;hp=0ac71cd9da78a57a5511aa205ed482775048a50d;hpb=36d5e7d31be61f631ace0488f0d6cd71b8f31a16;p=oota-llvm.git diff --git a/include/llvm/Bitcode/BitstreamReader.h b/include/llvm/Bitcode/BitstreamReader.h index 0ac71cd9da7..3b7a9b61e6a 100644 --- a/include/llvm/Bitcode/BitstreamReader.h +++ b/include/llvm/Bitcode/BitstreamReader.h @@ -2,8 +2,8 @@ // // The LLVM Compiler Infrastructure // -// This file was developed by Chris Lattner and is distributed under -// the University of Illinois Open Source License. See LICENSE.TXT for details. +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. // //===----------------------------------------------------------------------===// // @@ -16,12 +16,16 @@ #define BITSTREAM_READER_H #include "llvm/Bitcode/BitCodes.h" +#include namespace llvm { + class Deserializer; + class BitstreamReader { const unsigned char *NextChar; const unsigned char *LastChar; + friend class Deserializer; /// CurWord - This is the current data we have pulled from the stream but have /// not returned to the client. @@ -34,20 +38,105 @@ class BitstreamReader { // CurCodeSize - This is the declared size of code values used for the current // block, in bits. unsigned CurCodeSize; + + /// CurAbbrevs - Abbrevs installed at in this block. + std::vector CurAbbrevs; + + struct Block { + unsigned PrevCodeSize; + std::vector PrevAbbrevs; + explicit Block(unsigned PCS) : PrevCodeSize(PCS) {} + }; /// BlockScope - This tracks the codesize of parent blocks. - SmallVector BlockScope; + SmallVector BlockScope; + + /// BlockInfo - This contains information emitted to BLOCKINFO_BLOCK blocks. + /// These describe abbreviations that all blocks of the specified ID inherit. + struct BlockInfo { + unsigned BlockID; + std::vector Abbrevs; + }; + std::vector BlockInfoRecords; + /// FirstChar - This remembers the first byte of the stream. + const unsigned char *FirstChar; public: - BitstreamReader(const unsigned char *Start, const unsigned char *End) - : NextChar(Start), LastChar(End) { + BitstreamReader() { + NextChar = FirstChar = LastChar = 0; + CurWord = 0; + BitsInCurWord = 0; + CurCodeSize = 0; + } + + BitstreamReader(const unsigned char *Start, const unsigned char *End) { + init(Start, End); + } + + void init(const unsigned char *Start, const unsigned char *End) { + NextChar = FirstChar = Start; + LastChar = End; assert(((End-Start) & 3) == 0 &&"Bitcode stream not a multiple of 4 bytes"); CurWord = 0; BitsInCurWord = 0; CurCodeSize = 2; } - bool AtEndOfStream() const { return NextChar == LastChar; } + ~BitstreamReader() { + // Abbrevs could still exist if the stream was broken. If so, don't leak + // them. + for (unsigned i = 0, e = static_cast(CurAbbrevs.size()); + i != e; ++i) + CurAbbrevs[i]->dropRef(); + + 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(); + } + + // Free the BlockInfoRecords. + while (!BlockInfoRecords.empty()) { + BlockInfo &Info = BlockInfoRecords.back(); + // Free blockinfo abbrev info. + for (unsigned i = 0, e = static_cast(Info.Abbrevs.size()); + i != e; ++i) + Info.Abbrevs[i]->dropRef(); + BlockInfoRecords.pop_back(); + } + } + + bool AtEndOfStream() const { + return NextChar == LastChar && BitsInCurWord == 0; + } + + /// GetCurrentBitNo - Return the bit # of the bit we are reading. + uint64_t GetCurrentBitNo() const { + return (NextChar-FirstChar)*8 + ((32-BitsInCurWord) & 31); + } + + /// JumpToBit - Reset the stream to the specified bit number. + void JumpToBit(uint64_t BitNo) { + uintptr_t ByteNo = uintptr_t(BitNo/8) & ~3; + uintptr_t WordBitNo = uintptr_t(BitNo) & 31; + assert(ByteNo < (uintptr_t)(LastChar-FirstChar) && "Invalid location"); + + // Move the cursor to the right word. + NextChar = FirstChar+ByteNo; + BitsInCurWord = 0; + CurWord = 0; + + // Skip over any bits that are already consumed. + if (WordBitNo) { + NextChar -= 4; + Read(static_cast(WordBitNo)); + } + } + + /// GetAbbrevIDWidth - Return the number of bits used to encode an abbrev #. + unsigned GetAbbrevIDWidth() const { return CurCodeSize; } uint32_t Read(unsigned NumBits) { // If the field is fully contained by CurWord, return it quickly. @@ -87,6 +176,13 @@ public: return R; } + uint64_t Read64(unsigned NumBits) { + if (NumBits <= 32) return Read(NumBits); + + uint64_t V = Read(32); + return V | (uint64_t)Read(NumBits-32) << 32; + } + uint32_t ReadVBR(unsigned NumBits) { uint32_t Piece = Read(NumBits); if ((Piece & (1U << (NumBits-1))) == 0) @@ -137,6 +233,23 @@ public: // Block Manipulation //===--------------------------------------------------------------------===// +private: + /// getBlockInfo - If there is block info for the specified ID, return it, + /// otherwise return null. + BlockInfo *getBlockInfo(unsigned BlockID) { + // Common case, the most recent entry matches BlockID. + if (!BlockInfoRecords.empty() && BlockInfoRecords.back().BlockID == BlockID) + return &BlockInfoRecords.back(); + + for (unsigned i = 0, e = static_cast(BlockInfoRecords.size()); + i != e; ++i) + if (BlockInfoRecords[i].BlockID == BlockID) + return &BlockInfoRecords[i]; + return 0; + } +public: + + // Block header: // [ENTER_SUBBLOCK, blockid, newcodelen, , blocklen] @@ -165,15 +278,27 @@ public: return false; } - /// EnterSubBlock - Having read the ENTER_SUBBLOCK abbrevid, read and enter - /// the block, returning the BlockID of the block we just entered. - bool EnterSubBlock() { - BlockScope.push_back(CurCodeSize); + /// EnterSubBlock - Having read the ENTER_SUBBLOCK abbrevid, enter + /// the block, and return true if the block is valid. + 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 (BlockInfo *Info = 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() || NextChar+NumWords*4 > LastChar) @@ -188,15 +313,51 @@ public: // Block tail: // [END_BLOCK, ] SkipToWord(); - CurCodeSize = BlockScope.back(); - BlockScope.pop_back(); + + PopBlockScope(); return false; } +private: + void PopBlockScope() { + CurCodeSize = BlockScope.back().PrevCodeSize; + + // Delete abbrevs from popped scope. + for (unsigned i = 0, e = static_cast(CurAbbrevs.size()); + i != e; ++i) + CurAbbrevs[i]->dropRef(); + + BlockScope.back().PrevAbbrevs.swap(CurAbbrevs); + BlockScope.pop_back(); + } + //===--------------------------------------------------------------------===// // Record Processing //===--------------------------------------------------------------------===// +private: + void ReadAbbreviatedField(const BitCodeAbbrevOp &Op, + SmallVectorImpl &Vals) { + if (Op.isLiteral()) { + // If the abbrev specifies the literal value to use, use it. + Vals.push_back(Op.getLiteralValue()); + } else { + // Decode the value as we are commanded. + switch (Op.getEncoding()) { + default: assert(0 && "Unknown encoding!"); + case BitCodeAbbrevOp::Fixed: + Vals.push_back(Read((unsigned)Op.getEncodingData())); + break; + case BitCodeAbbrevOp::VBR: + Vals.push_back(ReadVBR64((unsigned)Op.getEncodingData())); + break; + case BitCodeAbbrevOp::Char6: + Vals.push_back(BitCodeAbbrevOp::DecodeChar6(Read(6))); + break; + } + } + } +public: unsigned ReadRecord(unsigned AbbrevID, SmallVectorImpl &Vals) { if (AbbrevID == bitc::UNABBREV_RECORD) { unsigned Code = ReadVBR(6); @@ -206,10 +367,114 @@ public: return Code; } - assert(0 && "Reading with abbrevs not implemented!"); - return 0; + unsigned AbbrevNo = AbbrevID-bitc::FIRST_APPLICATION_ABBREV; + assert(AbbrevNo < CurAbbrevs.size() && "Invalid abbrev #!"); + BitCodeAbbrev *Abbv = CurAbbrevs[AbbrevNo]; + + for (unsigned i = 0, e = Abbv->getNumOperandInfos(); i != e; ++i) { + const BitCodeAbbrevOp &Op = Abbv->getOperandInfo(i); + if (Op.isLiteral() || Op.getEncoding() != BitCodeAbbrevOp::Array) { + ReadAbbreviatedField(Op, Vals); + } else { + // 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); + } + } + + unsigned Code = (unsigned)Vals[0]; + Vals.erase(Vals.begin()); + return Code; } + //===--------------------------------------------------------------------===// + // 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); + } + + //===--------------------------------------------------------------------===// + // BlockInfo Block Reading + //===--------------------------------------------------------------------===// + +private: + BlockInfo &getOrCreateBlockInfo(unsigned BlockID) { + if (BlockInfo *BI = getBlockInfo(BlockID)) + return *BI; + + // Otherwise, add a new record. + BlockInfoRecords.push_back(BlockInfo()); + BlockInfoRecords.back().BlockID = BlockID; + return BlockInfoRecords.back(); + } + +public: + + bool ReadBlockInfoBlock() { + if (EnterSubBlock(bitc::BLOCKINFO_BLOCK_ID)) return true; + + SmallVector Record; + 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 = &getOrCreateBlockInfo((unsigned)Record[0]); + break; + } + } + } }; } // End llvm namespace