winamp/Src/external_dependencies/openmpt-trunk/include/unrar/unpack30.cpp

766 lines
19 KiB
C++
Raw Normal View History

2024-09-24 08:54:57 -04:00
// We use it instead of direct PPM.DecodeChar call to be sure that
// we reset PPM structures in case of corrupt data. It is important,
// because these structures can be invalid after PPM.DecodeChar returned -1.
inline int Unpack::SafePPMDecodeChar()
{
int Ch=PPM.DecodeChar();
if (Ch==-1) // Corrupt PPM data found.
{
PPM.CleanUp(); // Reset possibly corrupt PPM data structures.
UnpBlockType=BLOCK_LZ; // Set faster and more fail proof LZ mode.
}
return(Ch);
}
void Unpack::Unpack29(bool Solid)
{
static unsigned char LDecode[]={0,1,2,3,4,5,6,7,8,10,12,14,16,20,24,28,32,40,48,56,64,80,96,112,128,160,192,224};
static unsigned char LBits[]= {0,0,0,0,0,0,0,0,1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 4, 4, 5, 5, 5, 5};
static int DDecode[DC];
static byte DBits[DC];
static int DBitLengthCounts[]= {4,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,14,0,12};
static unsigned char SDDecode[]={0,4,8,16,32,64,128,192};
static unsigned char SDBits[]= {2,2,3, 4, 5, 6, 6, 6};
unsigned int Bits;
if (DDecode[1]==0)
{
int Dist=0,BitLength=0,Slot=0;
for (int I=0;I<ASIZE(DBitLengthCounts);I++,BitLength++)
for (int J=0;J<DBitLengthCounts[I];J++,Slot++,Dist+=(1<<BitLength))
{
DDecode[Slot]=Dist;
DBits[Slot]=BitLength;
}
}
FileExtracted=true;
if (!Suspended)
{
UnpInitData(Solid);
if (!UnpReadBuf30())
return;
if ((!Solid || !TablesRead3) && !ReadTables30())
return;
}
while (true)
{
UnpPtr&=MaxWinMask;
if (Inp.InAddr>ReadBorder)
{
if (!UnpReadBuf30())
break;
}
if (((WrPtr-UnpPtr) & MaxWinMask)<260 && WrPtr!=UnpPtr)
{
UnpWriteBuf30();
if (WrittenFileSize>DestUnpSize)
return;
if (Suspended)
{
FileExtracted=false;
return;
}
}
if (UnpBlockType==BLOCK_PPM)
{
// Here speed is critical, so we do not use SafePPMDecodeChar,
// because sometimes even the inline function can introduce
// some additional penalty.
int Ch=PPM.DecodeChar();
if (Ch==-1) // Corrupt PPM data found.
{
PPM.CleanUp(); // Reset possibly corrupt PPM data structures.
UnpBlockType=BLOCK_LZ; // Set faster and more fail proof LZ mode.
break;
}
if (Ch==PPMEscChar)
{
int NextCh=SafePPMDecodeChar();
if (NextCh==0) // End of PPM encoding.
{
if (!ReadTables30())
break;
continue;
}
if (NextCh==-1) // Corrupt PPM data found.
break;
if (NextCh==2) // End of file in PPM mode.
break;
if (NextCh==3) // Read VM code.
{
if (!ReadVMCodePPM())
break;
continue;
}
if (NextCh==4) // LZ inside of PPM.
{
unsigned int Distance=0,Length;
bool Failed=false;
for (int I=0;I<4 && !Failed;I++)
{
int Ch=SafePPMDecodeChar();
if (Ch==-1)
Failed=true;
else
if (I==3)
Length=(byte)Ch;
else
Distance=(Distance<<8)+(byte)Ch;
}
if (Failed)
break;
CopyString(Length+32,Distance+2);
continue;
}
if (NextCh==5) // One byte distance match (RLE) inside of PPM.
{
int Length=SafePPMDecodeChar();
if (Length==-1)
break;
CopyString(Length+4,1);
continue;
}
// If we are here, NextCh must be 1, what means that current byte
// is equal to our 'escape' byte, so we just store it to Window.
}
Window[UnpPtr++]=Ch;
continue;
}
uint Number=DecodeNumber(Inp,&BlockTables.LD);
if (Number<256)
{
Window[UnpPtr++]=(byte)Number;
continue;
}
if (Number>=271)
{
uint Length=LDecode[Number-=271]+3;
if ((Bits=LBits[Number])>0)
{
Length+=Inp.getbits()>>(16-Bits);
Inp.addbits(Bits);
}
uint DistNumber=DecodeNumber(Inp,&BlockTables.DD);
uint Distance=DDecode[DistNumber]+1;
if ((Bits=DBits[DistNumber])>0)
{
if (DistNumber>9)
{
if (Bits>4)
{
Distance+=((Inp.getbits()>>(20-Bits))<<4);
Inp.addbits(Bits-4);
}
if (LowDistRepCount>0)
{
LowDistRepCount--;
Distance+=PrevLowDist;
}
else
{
uint LowDist=DecodeNumber(Inp,&BlockTables.LDD);
if (LowDist==16)
{
LowDistRepCount=LOW_DIST_REP_COUNT-1;
Distance+=PrevLowDist;
}
else
{
Distance+=LowDist;
PrevLowDist=LowDist;
}
}
}
else
{
Distance+=Inp.getbits()>>(16-Bits);
Inp.addbits(Bits);
}
}
if (Distance>=0x2000)
{
Length++;
if (Distance>=0x40000)
Length++;
}
InsertOldDist(Distance);
LastLength=Length;
CopyString(Length,Distance);
continue;
}
if (Number==256)
{
if (!ReadEndOfBlock())
break;
continue;
}
if (Number==257)
{
if (!ReadVMCode())
break;
continue;
}
if (Number==258)
{
if (LastLength!=0)
CopyString(LastLength,OldDist[0]);
continue;
}
if (Number<263)
{
uint DistNum=Number-259;
uint Distance=OldDist[DistNum];
for (uint I=DistNum;I>0;I--)
OldDist[I]=OldDist[I-1];
OldDist[0]=Distance;
uint LengthNumber=DecodeNumber(Inp,&BlockTables.RD);
int Length=LDecode[LengthNumber]+2;
if ((Bits=LBits[LengthNumber])>0)
{
Length+=Inp.getbits()>>(16-Bits);
Inp.addbits(Bits);
}
LastLength=Length;
CopyString(Length,Distance);
continue;
}
if (Number<272)
{
uint Distance=SDDecode[Number-=263]+1;
if ((Bits=SDBits[Number])>0)
{
Distance+=Inp.getbits()>>(16-Bits);
Inp.addbits(Bits);
}
InsertOldDist(Distance);
LastLength=2;
CopyString(2,Distance);
continue;
}
}
UnpWriteBuf30();
}
// Return 'false' to quit unpacking the current file or 'true' to continue.
bool Unpack::ReadEndOfBlock()
{
uint BitField=Inp.getbits();
bool NewTable,NewFile=false;
// "1" - no new file, new table just here.
// "00" - new file, no new table.
// "01" - new file, new table (in beginning of next file).
if ((BitField & 0x8000)!=0)
{
NewTable=true;
Inp.addbits(1);
}
else
{
NewFile=true;
NewTable=(BitField & 0x4000)!=0;
Inp.addbits(2);
}
TablesRead3=!NewTable;
// Quit immediately if "new file" flag is set. If "new table" flag
// is present, we'll read the table in beginning of next file
// based on 'TablesRead3' 'false' value.
if (NewFile)
return false;
return ReadTables30(); // Quit only if we failed to read tables.
}
bool Unpack::ReadVMCode()
{
// Entire VM code is guaranteed to fully present in block defined
// by current Huffman table. Compressor checks that VM code does not cross
// Huffman block boundaries.
uint FirstByte=Inp.getbits()>>8;
Inp.addbits(8);
uint Length=(FirstByte & 7)+1;
if (Length==7)
{
Length=(Inp.getbits()>>8)+7;
Inp.addbits(8);
}
else
if (Length==8)
{
Length=Inp.getbits();
Inp.addbits(16);
}
if (Length==0)
return false;
Array<byte> VMCode(Length);
for (uint I=0;I<Length;I++)
{
// Try to read the new buffer if only one byte is left.
// But if we read all bytes except the last, one byte is enough.
if (Inp.InAddr>=ReadTop-1 && !UnpReadBuf30() && I<Length-1)
return false;
VMCode[I]=Inp.getbits()>>8;
Inp.addbits(8);
}
return AddVMCode(FirstByte,&VMCode[0],Length);
}
bool Unpack::ReadVMCodePPM()
{
uint FirstByte=SafePPMDecodeChar();
if ((int)FirstByte==-1)
return false;
uint Length=(FirstByte & 7)+1;
if (Length==7)
{
int B1=SafePPMDecodeChar();
if (B1==-1)
return false;
Length=B1+7;
}
else
if (Length==8)
{
int B1=SafePPMDecodeChar();
if (B1==-1)
return false;
int B2=SafePPMDecodeChar();
if (B2==-1)
return false;
Length=B1*256+B2;
}
if (Length==0)
return false;
Array<byte> VMCode(Length);
for (uint I=0;I<Length;I++)
{
int Ch=SafePPMDecodeChar();
if (Ch==-1)
return false;
VMCode[I]=Ch;
}
return AddVMCode(FirstByte,&VMCode[0],Length);
}
bool Unpack::AddVMCode(uint FirstByte,byte *Code,uint CodeSize)
{
VMCodeInp.InitBitInput();
memcpy(VMCodeInp.InBuf,Code,Min(BitInput::MAX_SIZE,CodeSize));
VM.Init();
uint FiltPos;
if ((FirstByte & 0x80)!=0)
{
FiltPos=RarVM::ReadData(VMCodeInp);
if (FiltPos==0)
InitFilters30(false);
else
FiltPos--;
}
else
FiltPos=LastFilter; // Use the same filter as last time.
if (FiltPos>Filters30.Size() || FiltPos>OldFilterLengths.Size())
return false;
LastFilter=FiltPos;
bool NewFilter=(FiltPos==Filters30.Size());
UnpackFilter30 *StackFilter=new UnpackFilter30; // New filter for PrgStack.
UnpackFilter30 *Filter;
if (NewFilter) // New filter code, never used before since VM reset.
{
if (FiltPos>MAX3_UNPACK_FILTERS)
{
// Too many different filters, corrupt archive.
delete StackFilter;
return false;
}
Filters30.Add(1);
Filters30[Filters30.Size()-1]=Filter=new UnpackFilter30;
StackFilter->ParentFilter=(uint)(Filters30.Size()-1);
// Reserve one item to store the data block length of our new filter
// entry. We'll set it to real block length below, after reading it.
// But we need to initialize it now, because when processing corrupt
// data, we can access this item even before we set it to real value.
OldFilterLengths.Push(0);
}
else // Filter was used in the past.
{
Filter=Filters30[FiltPos];
StackFilter->ParentFilter=FiltPos;
}
uint EmptyCount=0;
for (uint I=0;I<PrgStack.Size();I++)
{
PrgStack[I-EmptyCount]=PrgStack[I];
if (PrgStack[I]==NULL)
EmptyCount++;
if (EmptyCount>0)
PrgStack[I]=NULL;
}
if (EmptyCount==0)
{
if (PrgStack.Size()>MAX3_UNPACK_FILTERS)
{
delete StackFilter;
return false;
}
PrgStack.Add(1);
EmptyCount=1;
}
size_t StackPos=PrgStack.Size()-EmptyCount;
PrgStack[StackPos]=StackFilter;
uint BlockStart=RarVM::ReadData(VMCodeInp);
if ((FirstByte & 0x40)!=0)
BlockStart+=258;
StackFilter->BlockStart=(uint)((BlockStart+UnpPtr)&MaxWinMask);
if ((FirstByte & 0x20)!=0)
{
StackFilter->BlockLength=RarVM::ReadData(VMCodeInp);
// Store the last data block length for current filter.
OldFilterLengths[FiltPos]=StackFilter->BlockLength;
}
else
{
// Set the data block size to same value as the previous block size
// for same filter. It is possible for corrupt data to access a new
// and not filled yet item of OldFilterLengths array here. This is why
// we set new OldFilterLengths items to zero above.
StackFilter->BlockLength=FiltPos<OldFilterLengths.Size() ? OldFilterLengths[FiltPos]:0;
}
StackFilter->NextWindow=WrPtr!=UnpPtr && ((WrPtr-UnpPtr)&MaxWinMask)<=BlockStart;
// DebugLog("\nNextWindow: UnpPtr=%08x WrPtr=%08x BlockStart=%08x",UnpPtr,WrPtr,BlockStart);
memset(StackFilter->Prg.InitR,0,sizeof(StackFilter->Prg.InitR));
StackFilter->Prg.InitR[4]=StackFilter->BlockLength;
if ((FirstByte & 0x10)!=0) // Set registers to optional parameters if any.
{
uint InitMask=VMCodeInp.fgetbits()>>9;
VMCodeInp.faddbits(7);
for (uint I=0;I<7;I++)
if (InitMask & (1<<I))
StackFilter->Prg.InitR[I]=RarVM::ReadData(VMCodeInp);
}
if (NewFilter)
{
uint VMCodeSize=RarVM::ReadData(VMCodeInp);
if (VMCodeSize>=0x10000 || VMCodeSize==0 || VMCodeInp.InAddr+VMCodeSize>CodeSize)
return false;
Array<byte> VMCode(VMCodeSize);
for (uint I=0;I<VMCodeSize;I++)
{
if (VMCodeInp.Overflow(3))
return false;
VMCode[I]=VMCodeInp.fgetbits()>>8;
VMCodeInp.faddbits(8);
}
VM.Prepare(&VMCode[0],VMCodeSize,&Filter->Prg);
}
StackFilter->Prg.Type=Filter->Prg.Type;
return true;
}
bool Unpack::UnpReadBuf30()
{
int DataSize=ReadTop-Inp.InAddr; // Data left to process.
if (DataSize<0)
return false;
if (Inp.InAddr>BitInput::MAX_SIZE/2)
{
// If we already processed more than half of buffer, let's move
// remaining data into beginning to free more space for new data
// and ensure that calling function does not cross the buffer border
// even if we did not read anything here. Also it ensures that read size
// is not less than CRYPT_BLOCK_SIZE, so we can align it without risk
// to make it zero.
if (DataSize>0)
memmove(Inp.InBuf,Inp.InBuf+Inp.InAddr,DataSize);
Inp.InAddr=0;
ReadTop=DataSize;
}
else
DataSize=ReadTop;
int ReadCode=UnpIO->UnpRead(Inp.InBuf+DataSize,BitInput::MAX_SIZE-DataSize);
if (ReadCode>0)
ReadTop+=ReadCode;
ReadBorder=ReadTop-30;
return ReadCode!=-1;
}
void Unpack::UnpWriteBuf30()
{
uint WrittenBorder=(uint)WrPtr;
uint WriteSize=(uint)((UnpPtr-WrittenBorder)&MaxWinMask);
for (size_t I=0;I<PrgStack.Size();I++)
{
// Here we apply filters to data which we need to write.
// We always copy data to virtual machine memory before processing.
// We cannot process them just in place in Window buffer, because
// these data can be used for future string matches, so we must
// preserve them in original form.
UnpackFilter30 *flt=PrgStack[I];
if (flt==NULL)
continue;
if (flt->NextWindow)
{
flt->NextWindow=false;
continue;
}
unsigned int BlockStart=flt->BlockStart;
unsigned int BlockLength=flt->BlockLength;
if (((BlockStart-WrittenBorder)&MaxWinMask)<WriteSize)
{
if (WrittenBorder!=BlockStart)
{
UnpWriteArea(WrittenBorder,BlockStart);
WrittenBorder=BlockStart;
WriteSize=(uint)((UnpPtr-WrittenBorder)&MaxWinMask);
}
if (BlockLength<=WriteSize)
{
uint BlockEnd=(BlockStart+BlockLength)&MaxWinMask;
if (BlockStart<BlockEnd || BlockEnd==0)
VM.SetMemory(0,Window+BlockStart,BlockLength);
else
{
uint FirstPartLength=uint(MaxWinSize-BlockStart);
VM.SetMemory(0,Window+BlockStart,FirstPartLength);
VM.SetMemory(FirstPartLength,Window,BlockEnd);
}
VM_PreparedProgram *ParentPrg=&Filters30[flt->ParentFilter]->Prg;
VM_PreparedProgram *Prg=&flt->Prg;
ExecuteCode(Prg);
byte *FilteredData=Prg->FilteredData;
unsigned int FilteredDataSize=Prg->FilteredDataSize;
delete PrgStack[I];
PrgStack[I]=NULL;
while (I+1<PrgStack.Size())
{
UnpackFilter30 *NextFilter=PrgStack[I+1];
// It is required to check NextWindow here.
if (NextFilter==NULL || NextFilter->BlockStart!=BlockStart ||
NextFilter->BlockLength!=FilteredDataSize || NextFilter->NextWindow)
break;
// Apply several filters to same data block.
VM.SetMemory(0,FilteredData,FilteredDataSize);
VM_PreparedProgram *ParentPrg=&Filters30[NextFilter->ParentFilter]->Prg;
VM_PreparedProgram *NextPrg=&NextFilter->Prg;
ExecuteCode(NextPrg);
FilteredData=NextPrg->FilteredData;
FilteredDataSize=NextPrg->FilteredDataSize;
I++;
delete PrgStack[I];
PrgStack[I]=NULL;
}
UnpIO->UnpWrite(FilteredData,FilteredDataSize);
UnpSomeRead=true;
WrittenFileSize+=FilteredDataSize;
WrittenBorder=BlockEnd;
WriteSize=uint((UnpPtr-WrittenBorder)&MaxWinMask);
}
else
{
// Current filter intersects the window write border, so we adjust
// the window border to process this filter next time, not now.
for (size_t J=I;J<PrgStack.Size();J++)
{
UnpackFilter30 *flt=PrgStack[J];
if (flt!=NULL && flt->NextWindow)
flt->NextWindow=false;
}
WrPtr=WrittenBorder;
return;
}
}
}
UnpWriteArea(WrittenBorder,UnpPtr);
WrPtr=UnpPtr;
}
void Unpack::ExecuteCode(VM_PreparedProgram *Prg)
{
Prg->InitR[6]=(uint)WrittenFileSize;
VM.Execute(Prg);
}
bool Unpack::ReadTables30()
{
byte BitLength[BC];
byte Table[HUFF_TABLE_SIZE30];
if (Inp.InAddr>ReadTop-25)
if (!UnpReadBuf30())
return(false);
Inp.faddbits((8-Inp.InBit)&7);
uint BitField=Inp.fgetbits();
if (BitField & 0x8000)
{
UnpBlockType=BLOCK_PPM;
return(PPM.DecodeInit(this,PPMEscChar));
}
UnpBlockType=BLOCK_LZ;
PrevLowDist=0;
LowDistRepCount=0;
if (!(BitField & 0x4000))
memset(UnpOldTable,0,sizeof(UnpOldTable));
Inp.faddbits(2);
for (uint I=0;I<BC;I++)
{
uint Length=(byte)(Inp.fgetbits() >> 12);
Inp.faddbits(4);
if (Length==15)
{
uint ZeroCount=(byte)(Inp.fgetbits() >> 12);
Inp.faddbits(4);
if (ZeroCount==0)
BitLength[I]=15;
else
{
ZeroCount+=2;
while (ZeroCount-- > 0 && I<ASIZE(BitLength))
BitLength[I++]=0;
I--;
}
}
else
BitLength[I]=Length;
}
MakeDecodeTables(BitLength,&BlockTables.BD,BC30);
const uint TableSize=HUFF_TABLE_SIZE30;
for (uint I=0;I<TableSize;)
{
if (Inp.InAddr>ReadTop-5)
if (!UnpReadBuf30())
return(false);
uint Number=DecodeNumber(Inp,&BlockTables.BD);
if (Number<16)
{
Table[I]=(Number+UnpOldTable[I]) & 0xf;
I++;
}
else
if (Number<18)
{
uint N;
if (Number==16)
{
N=(Inp.fgetbits() >> 13)+3;
Inp.faddbits(3);
}
else
{
N=(Inp.fgetbits() >> 9)+11;
Inp.faddbits(7);
}
if (I==0)
return false; // We cannot have "repeat previous" code at the first position.
else
while (N-- > 0 && I<TableSize)
{
Table[I]=Table[I-1];
I++;
}
}
else
{
uint N;
if (Number==18)
{
N=(Inp.fgetbits() >> 13)+3;
Inp.faddbits(3);
}
else
{
N=(Inp.fgetbits() >> 9)+11;
Inp.faddbits(7);
}
while (N-- > 0 && I<TableSize)
Table[I++]=0;
}
}
TablesRead3=true;
if (Inp.InAddr>ReadTop)
return false;
MakeDecodeTables(&Table[0],&BlockTables.LD,NC30);
MakeDecodeTables(&Table[NC30],&BlockTables.DD,DC30);
MakeDecodeTables(&Table[NC30+DC30],&BlockTables.LDD,LDC30);
MakeDecodeTables(&Table[NC30+DC30+LDC30],&BlockTables.RD,RC30);
memcpy(UnpOldTable,Table,sizeof(UnpOldTable));
return true;
}
void Unpack::UnpInitData30(bool Solid)
{
if (!Solid)
{
TablesRead3=false;
memset(UnpOldTable,0,sizeof(UnpOldTable));
PPMEscChar=2;
UnpBlockType=BLOCK_LZ;
}
InitFilters30(Solid);
}
void Unpack::InitFilters30(bool Solid)
{
if (!Solid)
{
OldFilterLengths.SoftReset();
LastFilter=0;
for (size_t I=0;I<Filters30.Size();I++)
delete Filters30[I];
Filters30.SoftReset();
}
for (size_t I=0;I<PrgStack.Size();I++)
delete PrgStack[I];
PrgStack.SoftReset();
}