Submitted By: Robert Connolly (ashes) Date: 2005-06-08 Initial Package Version: 2.4 Upstream Status: Rejected upstream Origin: http://www.geocities.com/mctiew/ffw/lzma-vmlinuz.patch.gz Description: This adds 7zip compression to the linux kernel, which is 20-30% better than bzip2 or gzip. You need the lzma program to compress the kernel. This patch is part of the 7zip.txt LFS hint. Also see: http://www.7-zip.org/sdk.html diff -Naur linux-old/arch/i386/boot/compressed/LzmaDecode.c linux-lzma/arch/i386/boot/compressed/LzmaDecode.c --- linux-old/arch/i386/boot/compressed/LzmaDecode.c 1969-12-31 17:00:00.000000000 -0700 +++ linux-lzma/arch/i386/boot/compressed/LzmaDecode.c 2005-03-19 00:22:55.000000000 -0700 @@ -0,0 +1,663 @@ +/* + LzmaDecode.c + LZMA Decoder + + LZMA SDK 4.05 Copyright (c) 1999-2004 Igor Pavlov (2004-08-25) + http://www.7-zip.org/ + + LZMA SDK is licensed under two licenses: + 1) GNU Lesser General Public License (GNU LGPL) + 2) Common Public License (CPL) + It means that you can select one of these two licenses and + follow rules of that license. + + SPECIAL EXCEPTION: + Igor Pavlov, as the author of this code, expressly permits you to + statically or dynamically link your code (or bind by name) to the + interfaces of this file without subjecting your linked code to the + terms of the CPL or GNU LGPL. Any modifications or additions + to this file, however, are subject to the LGPL or CPL terms. +*/ + +#include "LzmaDecode.h" + +#ifndef Byte +#define Byte unsigned char +#endif + +#define kNumTopBits 24 +#define kTopValue ((UInt32)1 << kNumTopBits) + +#define kNumBitModelTotalBits 11 +#define kBitModelTotal (1 << kNumBitModelTotalBits) +#define kNumMoveBits 5 + +typedef struct _CRangeDecoder +{ + Byte *Buffer; + Byte *BufferLim; + UInt32 Range; + UInt32 Code; + #ifdef _LZMA_IN_CB + ILzmaInCallback *InCallback; + int Result; + #endif + int ExtraBytes; +} CRangeDecoder; + +Byte RangeDecoderReadByte(CRangeDecoder *rd) +{ + if (rd->Buffer == rd->BufferLim) + { + #ifdef _LZMA_IN_CB + UInt32 size; + rd->Result = rd->InCallback->Read(rd->InCallback, &rd->Buffer, &size); + rd->BufferLim = rd->Buffer + size; + if (size == 0) + #endif + { + rd->ExtraBytes = 1; + return 0xFF; + } + } + return (*rd->Buffer++); +} + +/* #define ReadByte (*rd->Buffer++) */ +#define ReadByte (RangeDecoderReadByte(rd)) + +void RangeDecoderInit(CRangeDecoder *rd, + #ifdef _LZMA_IN_CB + ILzmaInCallback *inCallback + #else + Byte *stream, UInt32 bufferSize + #endif + ) +{ + int i; + #ifdef _LZMA_IN_CB + rd->InCallback = inCallback; + rd->Buffer = rd->BufferLim = 0; + #else + rd->Buffer = stream; + rd->BufferLim = stream + bufferSize; + #endif + rd->ExtraBytes = 0; + rd->Code = 0; + rd->Range = (0xFFFFFFFF); + for(i = 0; i < 5; i++) + rd->Code = (rd->Code << 8) | ReadByte; +} + +#define RC_INIT_VAR UInt32 range = rd->Range; UInt32 code = rd->Code; +#define RC_FLUSH_VAR rd->Range = range; rd->Code = code; +#define RC_NORMALIZE if (range < kTopValue) { range <<= 8; code = (code << 8) | ReadByte; } + +UInt32 RangeDecoderDecodeDirectBits(CRangeDecoder *rd, int numTotalBits) +{ + RC_INIT_VAR + UInt32 result = 0; + int i; + for (i = numTotalBits; i > 0; i--) + { + /* UInt32 t; */ + range >>= 1; + + result <<= 1; + if (code >= range) + { + code -= range; + result |= 1; + } + /* + t = (code - range) >> 31; + t &= 1; + code -= range & (t - 1); + result = (result + result) | (1 - t); + */ + RC_NORMALIZE + } + RC_FLUSH_VAR + return result; +} + +int RangeDecoderBitDecode(CProb *prob, CRangeDecoder *rd) +{ + UInt32 bound = (rd->Range >> kNumBitModelTotalBits) * *prob; + if (rd->Code < bound) + { + rd->Range = bound; + *prob += (kBitModelTotal - *prob) >> kNumMoveBits; + if (rd->Range < kTopValue) + { + rd->Code = (rd->Code << 8) | ReadByte; + rd->Range <<= 8; + } + return 0; + } + else + { + rd->Range -= bound; + rd->Code -= bound; + *prob -= (*prob) >> kNumMoveBits; + if (rd->Range < kTopValue) + { + rd->Code = (rd->Code << 8) | ReadByte; + rd->Range <<= 8; + } + return 1; + } +} + +#define RC_GET_BIT2(prob, mi, A0, A1) \ + UInt32 bound = (range >> kNumBitModelTotalBits) * *prob; \ + if (code < bound) \ + { A0; range = bound; *prob += (kBitModelTotal - *prob) >> kNumMoveBits; mi <<= 1; } \ + else \ + { A1; range -= bound; code -= bound; *prob -= (*prob) >> kNumMoveBits; mi = (mi + mi) + 1; } \ + RC_NORMALIZE + +#define RC_GET_BIT(prob, mi) RC_GET_BIT2(prob, mi, ; , ;) + +int RangeDecoderBitTreeDecode(CProb *probs, int numLevels, CRangeDecoder *rd) +{ + int mi = 1; + int i; + #ifdef _LZMA_LOC_OPT + RC_INIT_VAR + #endif + for(i = numLevels; i > 0; i--) + { + #ifdef _LZMA_LOC_OPT + CProb *prob = probs + mi; + RC_GET_BIT(prob, mi) + #else + mi = (mi + mi) + RangeDecoderBitDecode(probs + mi, rd); + #endif + } + #ifdef _LZMA_LOC_OPT + RC_FLUSH_VAR + #endif + return mi - (1 << numLevels); +} + +int RangeDecoderReverseBitTreeDecode(CProb *probs, int numLevels, CRangeDecoder *rd) +{ + int mi = 1; + int i; + int symbol = 0; + #ifdef _LZMA_LOC_OPT + RC_INIT_VAR + #endif + for(i = 0; i < numLevels; i++) + { + #ifdef _LZMA_LOC_OPT + CProb *prob = probs + mi; + RC_GET_BIT2(prob, mi, ; , symbol |= (1 << i)) + #else + int bit = RangeDecoderBitDecode(probs + mi, rd); + mi = mi + mi + bit; + symbol |= (bit << i); + #endif + } + #ifdef _LZMA_LOC_OPT + RC_FLUSH_VAR + #endif + return symbol; +} + +Byte LzmaLiteralDecode(CProb *probs, CRangeDecoder *rd) +{ + int symbol = 1; + #ifdef _LZMA_LOC_OPT + RC_INIT_VAR + #endif + do + { + #ifdef _LZMA_LOC_OPT + CProb *prob = probs + symbol; + RC_GET_BIT(prob, symbol) + #else + symbol = (symbol + symbol) | RangeDecoderBitDecode(probs + symbol, rd); + #endif + } + while (symbol < 0x100); + #ifdef _LZMA_LOC_OPT + RC_FLUSH_VAR + #endif + return symbol; +} + +Byte LzmaLiteralDecodeMatch(CProb *probs, CRangeDecoder *rd, Byte matchByte) +{ + int symbol = 1; + #ifdef _LZMA_LOC_OPT + RC_INIT_VAR + #endif + do + { + int bit; + int matchBit = (matchByte >> 7) & 1; + matchByte <<= 1; + #ifdef _LZMA_LOC_OPT + { + CProb *prob = probs + ((1 + matchBit) << 8) + symbol; + RC_GET_BIT2(prob, symbol, bit = 0, bit = 1) + } + #else + bit = RangeDecoderBitDecode(probs + ((1 + matchBit) << 8) + symbol, rd); + symbol = (symbol << 1) | bit; + #endif + if (matchBit != bit) + { + while (symbol < 0x100) + { + #ifdef _LZMA_LOC_OPT + CProb *prob = probs + symbol; + RC_GET_BIT(prob, symbol) + #else + symbol = (symbol + symbol) | RangeDecoderBitDecode(probs + symbol, rd); + #endif + } + break; + } + } + while (symbol < 0x100); + #ifdef _LZMA_LOC_OPT + RC_FLUSH_VAR + #endif + return symbol; +} + +#define kNumPosBitsMax 4 +#define kNumPosStatesMax (1 << kNumPosBitsMax) + +#define kLenNumLowBits 3 +#define kLenNumLowSymbols (1 << kLenNumLowBits) +#define kLenNumMidBits 3 +#define kLenNumMidSymbols (1 << kLenNumMidBits) +#define kLenNumHighBits 8 +#define kLenNumHighSymbols (1 << kLenNumHighBits) + +#define LenChoice 0 +#define LenChoice2 (LenChoice + 1) +#define LenLow (LenChoice2 + 1) +#define LenMid (LenLow + (kNumPosStatesMax << kLenNumLowBits)) +#define LenHigh (LenMid + (kNumPosStatesMax << kLenNumMidBits)) +#define kNumLenProbs (LenHigh + kLenNumHighSymbols) + +int LzmaLenDecode(CProb *p, CRangeDecoder *rd, int posState) +{ + if(RangeDecoderBitDecode(p + LenChoice, rd) == 0) + return RangeDecoderBitTreeDecode(p + LenLow + + (posState << kLenNumLowBits), kLenNumLowBits, rd); + if(RangeDecoderBitDecode(p + LenChoice2, rd) == 0) + return kLenNumLowSymbols + RangeDecoderBitTreeDecode(p + LenMid + + (posState << kLenNumMidBits), kLenNumMidBits, rd); + return kLenNumLowSymbols + kLenNumMidSymbols + + RangeDecoderBitTreeDecode(p + LenHigh, kLenNumHighBits, rd); +} + +#define kNumStates 12 + +#define kStartPosModelIndex 4 +#define kEndPosModelIndex 14 +#define kNumFullDistances (1 << (kEndPosModelIndex >> 1)) + +#define kNumPosSlotBits 6 +#define kNumLenToPosStates 4 + +#define kNumAlignBits 4 +#define kAlignTableSize (1 << kNumAlignBits) + +#define kMatchMinLen 2 + +#define IsMatch 0 +#define IsRep (IsMatch + (kNumStates << kNumPosBitsMax)) +#define IsRepG0 (IsRep + kNumStates) +#define IsRepG1 (IsRepG0 + kNumStates) +#define IsRepG2 (IsRepG1 + kNumStates) +#define IsRep0Long (IsRepG2 + kNumStates) +#define PosSlot (IsRep0Long + (kNumStates << kNumPosBitsMax)) +#define SpecPos (PosSlot + (kNumLenToPosStates << kNumPosSlotBits)) +#define Align (SpecPos + kNumFullDistances - kEndPosModelIndex) +#define LenCoder (Align + kAlignTableSize) +#define RepLenCoder (LenCoder + kNumLenProbs) +#define Literal (RepLenCoder + kNumLenProbs) + +#if Literal != LZMA_BASE_SIZE +StopCompilingDueBUG +#endif + +#ifdef _LZMA_OUT_READ + +typedef struct _LzmaVarState +{ + CRangeDecoder RangeDecoder; + Byte *Dictionary; + UInt32 DictionarySize; + UInt32 DictionaryPos; + UInt32 GlobalPos; + UInt32 Reps[4]; + int lc; + int lp; + int pb; + int State; + int PreviousIsMatch; + int RemainLen; +} LzmaVarState; + +int LzmaDecoderInit( + unsigned char *buffer, UInt32 bufferSize, + int lc, int lp, int pb, + unsigned char *dictionary, UInt32 dictionarySize, + #ifdef _LZMA_IN_CB + ILzmaInCallback *inCallback + #else + unsigned char *inStream, UInt32 inSize + #endif + ) +{ + LzmaVarState *vs = (LzmaVarState *)buffer; + CProb *p = (CProb *)(buffer + sizeof(LzmaVarState)); + UInt32 numProbs = Literal + ((UInt32)LZMA_LIT_SIZE << (lc + lp)); + UInt32 i; + if (bufferSize < numProbs * sizeof(CProb) + sizeof(LzmaVarState)) + return LZMA_RESULT_NOT_ENOUGH_MEM; + vs->Dictionary = dictionary; + vs->DictionarySize = dictionarySize; + vs->DictionaryPos = 0; + vs->GlobalPos = 0; + vs->Reps[0] = vs->Reps[1] = vs->Reps[2] = vs->Reps[3] = 1; + vs->lc = lc; + vs->lp = lp; + vs->pb = pb; + vs->State = 0; + vs->PreviousIsMatch = 0; + vs->RemainLen = 0; + dictionary[dictionarySize - 1] = 0; + for (i = 0; i < numProbs; i++) + p[i] = kBitModelTotal >> 1; + RangeDecoderInit(&vs->RangeDecoder, + #ifdef _LZMA_IN_CB + inCallback + #else + inStream, inSize + #endif + ); + return LZMA_RESULT_OK; +} + +int LzmaDecode(unsigned char *buffer, + unsigned char *outStream, UInt32 outSize, + UInt32 *outSizeProcessed) +{ + LzmaVarState *vs = (LzmaVarState *)buffer; + CProb *p = (CProb *)(buffer + sizeof(LzmaVarState)); + CRangeDecoder rd = vs->RangeDecoder; + int state = vs->State; + int previousIsMatch = vs->PreviousIsMatch; + Byte previousByte; + UInt32 rep0 = vs->Reps[0], rep1 = vs->Reps[1], rep2 = vs->Reps[2], rep3 = vs->Reps[3]; + UInt32 nowPos = 0; + UInt32 posStateMask = (1 << (vs->pb)) - 1; + UInt32 literalPosMask = (1 << (vs->lp)) - 1; + int lc = vs->lc; + int len = vs->RemainLen; + UInt32 globalPos = vs->GlobalPos; + + Byte *dictionary = vs->Dictionary; + UInt32 dictionarySize = vs->DictionarySize; + UInt32 dictionaryPos = vs->DictionaryPos; + + if (len == -1) + { + *outSizeProcessed = 0; + return LZMA_RESULT_OK; + } + + while(len > 0 && nowPos < outSize) + { + UInt32 pos = dictionaryPos - rep0; + if (pos >= dictionarySize) + pos += dictionarySize; + outStream[nowPos++] = dictionary[dictionaryPos] = dictionary[pos]; + if (++dictionaryPos == dictionarySize) + dictionaryPos = 0; + len--; + } + if (dictionaryPos == 0) + previousByte = dictionary[dictionarySize - 1]; + else + previousByte = dictionary[dictionaryPos - 1]; +#else + +int LzmaDecode( + Byte *buffer, UInt32 bufferSize, + int lc, int lp, int pb, + #ifdef _LZMA_IN_CB + ILzmaInCallback *inCallback, + #else + unsigned char *inStream, UInt32 inSize, + #endif + unsigned char *outStream, UInt32 outSize, + UInt32 *outSizeProcessed) +{ + UInt32 numProbs = Literal + ((UInt32)LZMA_LIT_SIZE << (lc + lp)); + CProb *p = (CProb *)buffer; + CRangeDecoder rd; + UInt32 i; + int state = 0; + int previousIsMatch = 0; + Byte previousByte = 0; + UInt32 rep0 = 1, rep1 = 1, rep2 = 1, rep3 = 1; + UInt32 nowPos = 0; + UInt32 posStateMask = (1 << pb) - 1; + UInt32 literalPosMask = (1 << lp) - 1; + int len = 0; + if (bufferSize < numProbs * sizeof(CProb)) + return LZMA_RESULT_NOT_ENOUGH_MEM; + for (i = 0; i < numProbs; i++) + p[i] = kBitModelTotal >> 1; + RangeDecoderInit(&rd, + #ifdef _LZMA_IN_CB + inCallback + #else + inStream, inSize + #endif + ); +#endif + + *outSizeProcessed = 0; + while(nowPos < outSize) + { + int posState = (int)( + (nowPos + #ifdef _LZMA_OUT_READ + + globalPos + #endif + ) + & posStateMask); + #ifdef _LZMA_IN_CB + if (rd.Result != LZMA_RESULT_OK) + return rd.Result; + #endif + if (rd.ExtraBytes != 0) + return LZMA_RESULT_DATA_ERROR; + if (RangeDecoderBitDecode(p + IsMatch + (state << kNumPosBitsMax) + posState, &rd) == 0) + { + CProb *probs = p + Literal + (LZMA_LIT_SIZE * + ((( + (nowPos + #ifdef _LZMA_OUT_READ + + globalPos + #endif + ) + & literalPosMask) << lc) + (previousByte >> (8 - lc)))); + + if (state < 4) state = 0; + else if (state < 10) state -= 3; + else state -= 6; + if (previousIsMatch) + { + Byte matchByte; + #ifdef _LZMA_OUT_READ + UInt32 pos = dictionaryPos - rep0; + if (pos >= dictionarySize) + pos += dictionarySize; + matchByte = dictionary[pos]; + #else + matchByte = outStream[nowPos - rep0]; + #endif + previousByte = LzmaLiteralDecodeMatch(probs, &rd, matchByte); + previousIsMatch = 0; + } + else + previousByte = LzmaLiteralDecode(probs, &rd); + outStream[nowPos++] = previousByte; + #ifdef _LZMA_OUT_READ + dictionary[dictionaryPos] = previousByte; + if (++dictionaryPos == dictionarySize) + dictionaryPos = 0; + #endif + } + else + { + previousIsMatch = 1; + if (RangeDecoderBitDecode(p + IsRep + state, &rd) == 1) + { + if (RangeDecoderBitDecode(p + IsRepG0 + state, &rd) == 0) + { + if (RangeDecoderBitDecode(p + IsRep0Long + (state << kNumPosBitsMax) + posState, &rd) == 0) + { + #ifdef _LZMA_OUT_READ + UInt32 pos; + #endif + if ( + (nowPos + #ifdef _LZMA_OUT_READ + + globalPos + #endif + ) + == 0) + return LZMA_RESULT_DATA_ERROR; + state = state < 7 ? 9 : 11; + #ifdef _LZMA_OUT_READ + pos = dictionaryPos - rep0; + if (pos >= dictionarySize) + pos += dictionarySize; + previousByte = dictionary[pos]; + dictionary[dictionaryPos] = previousByte; + if (++dictionaryPos == dictionarySize) + dictionaryPos = 0; + #else + previousByte = outStream[nowPos - rep0]; + #endif + outStream[nowPos++] = previousByte; + continue; + } + } + else + { + UInt32 distance; + if(RangeDecoderBitDecode(p + IsRepG1 + state, &rd) == 0) + distance = rep1; + else + { + if(RangeDecoderBitDecode(p + IsRepG2 + state, &rd) == 0) + distance = rep2; + else + { + distance = rep3; + rep3 = rep2; + } + rep2 = rep1; + } + rep1 = rep0; + rep0 = distance; + } + len = LzmaLenDecode(p + RepLenCoder, &rd, posState); + state = state < 7 ? 8 : 11; + } + else + { + int posSlot; + rep3 = rep2; + rep2 = rep1; + rep1 = rep0; + state = state < 7 ? 7 : 10; + len = LzmaLenDecode(p + LenCoder, &rd, posState); + posSlot = RangeDecoderBitTreeDecode(p + PosSlot + + ((len < kNumLenToPosStates ? len : kNumLenToPosStates - 1) << + kNumPosSlotBits), kNumPosSlotBits, &rd); + if (posSlot >= kStartPosModelIndex) + { + int numDirectBits = ((posSlot >> 1) - 1); + rep0 = ((2 | ((UInt32)posSlot & 1)) << numDirectBits); + if (posSlot < kEndPosModelIndex) + { + rep0 += RangeDecoderReverseBitTreeDecode( + p + SpecPos + rep0 - posSlot - 1, numDirectBits, &rd); + } + else + { + rep0 += RangeDecoderDecodeDirectBits(&rd, + numDirectBits - kNumAlignBits) << kNumAlignBits; + rep0 += RangeDecoderReverseBitTreeDecode(p + Align, kNumAlignBits, &rd); + } + } + else + rep0 = posSlot; + rep0++; + } + if (rep0 == (UInt32)(0)) + { + /* it's for stream version */ + len = -1; + break; + } + if (rep0 > nowPos + #ifdef _LZMA_OUT_READ + + globalPos + #endif + ) + { + return LZMA_RESULT_DATA_ERROR; + } + len += kMatchMinLen; + do + { + #ifdef _LZMA_OUT_READ + UInt32 pos = dictionaryPos - rep0; + if (pos >= dictionarySize) + pos += dictionarySize; + previousByte = dictionary[pos]; + dictionary[dictionaryPos] = previousByte; + if (++dictionaryPos == dictionarySize) + dictionaryPos = 0; + #else + previousByte = outStream[nowPos - rep0]; + #endif + outStream[nowPos++] = previousByte; + len--; + } + while(len > 0 && nowPos < outSize); + } + } + + #ifdef _LZMA_OUT_READ + vs->RangeDecoder = rd; + vs->DictionaryPos = dictionaryPos; + vs->GlobalPos = globalPos + nowPos; + vs->Reps[0] = rep0; + vs->Reps[1] = rep1; + vs->Reps[2] = rep2; + vs->Reps[3] = rep3; + vs->State = state; + vs->PreviousIsMatch = previousIsMatch; + vs->RemainLen = len; + #endif + + *outSizeProcessed = nowPos; + return LZMA_RESULT_OK; +} diff -Naur linux-old/arch/i386/boot/compressed/LzmaDecode.h linux-lzma/arch/i386/boot/compressed/LzmaDecode.h --- linux-old/arch/i386/boot/compressed/LzmaDecode.h 1969-12-31 17:00:00.000000000 -0700 +++ linux-lzma/arch/i386/boot/compressed/LzmaDecode.h 2005-03-19 00:22:58.000000000 -0700 @@ -0,0 +1,100 @@ +/* + LzmaDecode.h + LZMA Decoder interface + + LZMA SDK 4.05 Copyright (c) 1999-2004 Igor Pavlov (2004-08-25) + http://www.7-zip.org/ + + LZMA SDK is licensed under two licenses: + 1) GNU Lesser General Public License (GNU LGPL) + 2) Common Public License (CPL) + It means that you can select one of these two licenses and + follow rules of that license. + + SPECIAL EXCEPTION: + Igor Pavlov, as the author of this code, expressly permits you to + statically or dynamically link your code (or bind by name) to the + interfaces of this file without subjecting your linked code to the + terms of the CPL or GNU LGPL. Any modifications or additions + to this file, however, are subject to the LGPL or CPL terms. +*/ + +#ifndef __LZMADECODE_H +#define __LZMADECODE_H + +/* #define _LZMA_IN_CB */ +/* Use callback for input data */ + +/* #define _LZMA_OUT_READ */ +/* Use read function for output data */ + +/* #define _LZMA_PROB32 */ +/* It can increase speed on some 32-bit CPUs, + but memory usage will be doubled in that case */ + +/* #define _LZMA_LOC_OPT */ +/* Enable local speed optimizations inside code */ + +#ifndef UInt32 +#ifdef _LZMA_UINT32_IS_ULONG +#define UInt32 unsigned long +#else +#define UInt32 unsigned int +#endif +#endif + +#ifdef _LZMA_PROB32 +#define CProb UInt32 +#else +#define CProb unsigned short +#endif + +#define LZMA_RESULT_OK 0 +#define LZMA_RESULT_DATA_ERROR 1 +#define LZMA_RESULT_NOT_ENOUGH_MEM 2 + +#ifdef _LZMA_IN_CB +typedef struct _ILzmaInCallback +{ + int (*Read)(void *object, unsigned char **buffer, UInt32 *bufferSize); +} ILzmaInCallback; +#endif + +#define LZMA_BASE_SIZE 1846 +#define LZMA_LIT_SIZE 768 + +/* +bufferSize = (LZMA_BASE_SIZE + (LZMA_LIT_SIZE << (lc + lp)))* sizeof(CProb) +bufferSize += 100 in case of _LZMA_OUT_READ +by default CProb is unsigned short, +but if specify _LZMA_PROB_32, CProb will be UInt32(unsigned int) +*/ + +#ifdef _LZMA_OUT_READ +int LzmaDecoderInit( + unsigned char *buffer, UInt32 bufferSize, + int lc, int lp, int pb, + unsigned char *dictionary, UInt32 dictionarySize, + #ifdef _LZMA_IN_CB + ILzmaInCallback *inCallback + #else + unsigned char *inStream, UInt32 inSize + #endif +); +#endif + +int LzmaDecode( + unsigned char *buffer, + #ifndef _LZMA_OUT_READ + UInt32 bufferSize, + int lc, int lp, int pb, + #ifdef _LZMA_IN_CB + ILzmaInCallback *inCallback, + #else + unsigned char *inStream, UInt32 inSize, + #endif + #endif + unsigned char *outStream, UInt32 outSize, + UInt32 *outSizeProcessed); + +#endif diff -Naur linux-old/arch/i386/boot/compressed/Makefile linux-lzma/arch/i386/boot/compressed/Makefile --- linux-old/arch/i386/boot/compressed/Makefile 2005-03-21 05:43:02.000000000 -0700 +++ linux-lzma/arch/i386/boot/compressed/Makefile 2005-03-21 07:34:57.000000000 -0700 @@ -1,13 +1,16 @@ # # linux/arch/i386/boot/compressed/Makefile # -# create a compressed vmlinux image from the original vmlinux +# Create a compressed vmlinux image from the original vmlinux +# +# Patched by Ming-Ching Tiew +# # HEAD = head.o SYSTEM = $(TOPDIR)/vmlinux -OBJECTS = $(HEAD) misc.o +OBJECTS = $(HEAD) lzma_misc.o ZLDFLAGS = -e startup_32 @@ -37,11 +40,14 @@ misc.o: misc.c $(CC) $(CFLAGS) -DKBUILD_BASENAME=$(subst $(comma),_,$(subst -,_,$(*F))) -c misc.c +lzma_misc.o: lzma_misc.c + $(CC) $(CFLAGS) -DKBUILD_BASENAME=$(subst $(comma),_,$(subst -,_,$(*F))) -c lzma_misc.c + piggy.o: $(SYSTEM) tmppiggy=_tmp_$$$$piggy; \ rm -f $$tmppiggy $$tmppiggy.gz $$tmppiggy.lnk; \ $(OBJCOPY) $(SYSTEM) $$tmppiggy; \ - gzip -f -9 < $$tmppiggy > $$tmppiggy.gz; \ + lzma e $$tmppiggy $$tmppiggy.gz -lc7 -lp0 -pb0; \ echo "SECTIONS { .data : { input_len = .; LONG(input_data_end - input_data) input_data = .; *(.data) input_data_end = .; }}" > $$tmppiggy.lnk; \ $(LD) -r -o piggy.o -b binary $$tmppiggy.gz -b elf32-i386 -T $$tmppiggy.lnk; \ rm -f $$tmppiggy $$tmppiggy.gz $$tmppiggy.lnk diff -Naur linux-old/arch/i386/boot/compressed/lzma_misc.c linux-lzma/arch/i386/boot/compressed/lzma_misc.c --- linux-old/arch/i386/boot/compressed/lzma_misc.c 1969-12-31 17:00:00.000000000 -0700 +++ linux-lzma/arch/i386/boot/compressed/lzma_misc.c 2005-03-21 07:24:55.000000000 -0700 @@ -0,0 +1,412 @@ +/* + * lzma_misc.c + * + * Decompress LZMA compressed vmlinuz + * Version 0.8 Copyright (c) Ming-Ching Tiew mctiew@yahoo.com + * + * Program adapted from misc.c + * Date: 21 March 2005 + * Source released under GPL + */ + +#define STANDALONE + +#include +#include +#include +#include + +#define OF(args) args +#define STATIC static + +#undef memset +#undef memcpy + +/* + * Why do we do this? Don't ask me.. + * + * Incomprehensible are the ways of bootloaders. + */ +static void* memcpy(void *, __const void *, size_t); + +typedef unsigned char uch; +typedef unsigned short ush; +typedef unsigned long ulg; + +#define WSIZE 0x8000 /* Window size must be at least 32k, */ + /* and a power of two */ + +static uch *inbuf; /* input buffer */ + +static unsigned insize = 0; /* valid bytes in inbuf */ +static unsigned inptr = 0; /* index of next byte to be processed in inbuf */ + +#define get_byte() (inptr < insize ? inbuf[inptr++] : fill_inbuf()) + +/* Diagnostic functions */ +#ifdef DEBUG +# define Assert(cond,msg) {if(!(cond)) error(msg);} +# define Trace(x) fprintf x +# define Tracev(x) {if (verbose) fprintf x ;} +# define Tracevv(x) {if (verbose>1) fprintf x ;} +# define Tracec(c,x) {if (verbose && (c)) fprintf x ;} +# define Tracecv(c,x) {if (verbose>1 && (c)) fprintf x ;} +#else +# define Assert(cond,msg) +# define Trace(x) +# define Tracev(x) +# define Tracevv(x) +# define Tracec(c,x) +# define Tracecv(c,x) +#endif + +static int fill_inbuf(void); +static void error(char *m); + +/* + * This is set up by the setup-routine at boot-time + */ +static unsigned char *real_mode; /* Pointer to real-mode data */ + +#define EXT_MEM_K (*(unsigned short *)(real_mode + 0x2)) +#ifndef STANDARD_MEMORY_BIOS_CALL +#define ALT_MEM_K (*(unsigned long *)(real_mode + 0x1e0)) +#endif +#define SCREEN_INFO (*(struct screen_info *)(real_mode+0)) + +extern char input_data[]; +extern int input_len; + +static long bytes_out = 0; +static uch *output_data; + +static void puts(const char *); + +extern int end; +static long free_mem_ptr = (long)&end; +static long free_mem_end_ptr; + +#define INPLACE_MOVE_ROUTINE 0x1000 +#define LOW_BUFFER_START 0x2000 +#define LOW_BUFFER_MAX 0x90000 +#define HEAP_SIZE 0x3000 +static unsigned int low_buffer_end, low_buffer_size; +static int high_loaded =0; +static uch *high_buffer_start /* = (uch *)(((ulg)&end) + HEAP_SIZE)*/; + +static char *vidmem = (char *)0xb8000; +static int vidport; +static int lines, cols; + +static void scroll(void) +{ + int i; + + memcpy ( vidmem, vidmem + cols * 2, ( lines - 1 ) * cols * 2 ); + for ( i = ( lines - 1 ) * cols * 2; i < lines * cols * 2; i += 2 ) + vidmem[i] = ' '; +} + +static void puts(const char *s) +{ + int x,y,pos; + char c; + + x = SCREEN_INFO.orig_x; + y = SCREEN_INFO.orig_y; + + while ( ( c = *s++ ) != '\0' ) { + if ( c == '\n' ) { + x = 0; + if ( ++y >= lines ) { + scroll(); + y--; + } + } else { + vidmem [ ( x + cols * y ) * 2 ] = c; + if ( ++x >= cols ) { + x = 0; + if ( ++y >= lines ) { + scroll(); + y--; + } + } + } + } + + SCREEN_INFO.orig_x = x; + SCREEN_INFO.orig_y = y; + + pos = (x + cols * y) * 2; /* Update cursor position */ + outb_p(14, vidport); + outb_p(0xff & (pos >> 9), vidport+1); + outb_p(15, vidport); + outb_p(0xff & (pos >> 1), vidport+1); +} + +static void* memcpy(void* __dest, __const void* __src, + size_t __n) +{ + int i; + char *d = (char *)__dest, *s = (char *)__src; + + for (i=0;i<__n;i++) d[i] = s[i]; + return __dest; +} + +/* =========================================================================== + * Fill the input buffer. This is called only when the buffer is empty + * and at least one byte is really needed. + */ +static int fill_inbuf(void) +{ + if (insize != 0) { + error("ran out of input data\n"); + } + + inbuf = input_data; + insize = input_len; + inptr = 1; + return inbuf[0]; +} + +static void error(char *x) +{ + puts("\n\n"); + puts(x); + puts("\n\n -- System halted"); + + while(1); /* Halt */ +} + +#define STACK_SIZE (4096) + +long user_stack [STACK_SIZE]; + +struct { + long * a; + short b; + } stack_start = { & user_stack [STACK_SIZE] , __KERNEL_DS }; + +static void setup_normal_output_buffer(void) +{ +#ifdef STANDARD_MEMORY_BIOS_CALL + if (EXT_MEM_K < 1024) error("Less than 2MB of memory.\n"); +#else + if ((ALT_MEM_K > EXT_MEM_K ? ALT_MEM_K : EXT_MEM_K) < 1024) error("Less than 2MB of memory.\n"); +#endif + output_data = (char *)0x100000; /* Points to 1M */ + free_mem_end_ptr = (long)real_mode; +} + +struct moveparams { + uch *low_buffer_start; int lcount; + uch *high_buffer_start; int hcount; +}; + +static void setup_output_buffer_if_we_run_high(struct moveparams *mv) +{ + high_buffer_start = (uch *)(((ulg)&end) + HEAP_SIZE); +#ifdef STANDARD_MEMORY_BIOS_CALL + if (EXT_MEM_K < (3*1024)) error("Less than 4MB of memory.\n"); +#else + if ((ALT_MEM_K > EXT_MEM_K ? ALT_MEM_K : EXT_MEM_K) < (3*1024)) error("Less than 4MB of memory.\n"); +#endif + mv->low_buffer_start = output_data = (char *)LOW_BUFFER_START; + low_buffer_end = ((unsigned int)real_mode > LOW_BUFFER_MAX + ? LOW_BUFFER_MAX : (unsigned int)real_mode) & ~0xfff; + low_buffer_size = low_buffer_end - LOW_BUFFER_START; + high_loaded = 1; + free_mem_end_ptr = (long)high_buffer_start; + if ( (0x100000 + low_buffer_size) > ((ulg)high_buffer_start)) { + high_buffer_start = (uch *)(0x100000 + low_buffer_size); + mv->hcount = 0; /* say: we need not to move high_buffer */ + } + else mv->hcount = -1; + mv->high_buffer_start = high_buffer_start; +} + +static void close_output_buffer_if_we_run_high(struct moveparams *mv) +{ + if (bytes_out > low_buffer_size) { + mv->lcount = low_buffer_size; + if (mv->hcount) + mv->hcount = bytes_out - low_buffer_size; + } else { + mv->lcount = bytes_out; + mv->hcount = 0; + } +} + +// When using LZMA in callback, the compressed length is not needed. +// Otherwise you need a special version of lzma compression program +// which will pad the compressed length in the header. +#define _LZMA_IN_CB +#include "LzmaDecode.h" +#include "LzmaDecode.c" + +#ifdef _LZMA_IN_CB +static int read_byte(void *object, unsigned char **buffer, UInt32 *bufferSize); +#endif + + +/* + * Do the lzma decompression + * When using LZMA in callback, the end of input stream is automatically determined + */ +static int lzma_unzip(void) +{ + + unsigned int i; /* temp value */ + unsigned int lc; /* literal context bits */ + unsigned int lp; /* literal pos state bits */ + unsigned int pb; /* pos state bits */ + unsigned char* workspace; + unsigned int uncompressedSize = 0; + unsigned char* p; + +#ifdef _LZMA_IN_CB + ILzmaInCallback callback; + callback.Read = read_byte; +#else + unsigned char* inputbuf; + unsigned int lzma_workspace_size; + unsigned int compressedSize = 0; +#endif + + /* lzma args */ + i = get_byte(); + lc = i % 9, i = i / 9; + lp = i % 5, pb = i / 5; + + /* skip dictionary size */ + for (i = 0; i < 4; i++) + get_byte(); + // get uncompressedSize + p= (char*)&uncompressedSize; + for (i = 0; i < 4; i++) + *p++ = get_byte(); + + //get compressedSize +#ifdef _LZMA_IN_CB + for (i = 0; i < 4; i++) + get_byte(); +#else + p= (char*)&compressedSize; + for (i = 0; i < 4; i++) + *p++ = get_byte(); +#endif + +#if 0 + if ( (lc == 5 ) && (lp == 0 ) && ( pb == 0 )) + { + puts("got prop!\n"); + } + +#ifndef _LZMA_IN_CB + if( compressedSize == 496722 ) + { + puts( "got the right sizes\n"); + } + else if ( compressedSize > 496722 ) + { + puts( "greater !\n"); + } + else if ( compressedSize < 496722 ) + puts( "smaller!\n"); + +#endif + if ( uncompressedSize == 1187168 ) + { + puts( "got the right uncompressed size \n"); + }else if ( uncompressedSize > 1187168 ) + { + puts( "uncompressedSize greater!\n"); + }else + puts( "uncompressedSize smaller!|n"); +#endif + + // point it beyond uncompresedSize + workspace = high_buffer_start + uncompressedSize; + // +#ifndef _LZMA_IN_CB + lzma_workspace_size = (LZMA_BASE_SIZE + (LZMA_LIT_SIZE << (lc + lp))) * sizeof(CProb); + inputbuf = high_buffer_start + uncompressedSize + lzma_workspace_size; + // read the compressed data + for( i=0; i < compressedSize; i++) + { + if ( i % ( 1024 * 10 ) == 0 ) + puts("."); + *inputbuf++ = get_byte(); + } +#endif + + /* decompress kernel */ +#ifdef _LZMA_IN_CB + if (LzmaDecode(workspace, ~0, lc, lp, pb, + &callback, +#else + if (LzmaDecode(workspace, lzma_workspace_size, lc, lp, pb, + inputbuf - compressedSize, compressedSize, +#endif + (unsigned char*)high_buffer_start, uncompressedSize, &i) == LZMA_RESULT_OK) + { + if ( i != uncompressedSize ) + error( "kernel corrupted!\n"); + //copy it back to low_buffer + if( uncompressedSize > low_buffer_size ) + { + memcpy((char*)LOW_BUFFER_START, high_buffer_start, low_buffer_size); + memcpy(high_buffer_start, high_buffer_start+low_buffer_size, + uncompressedSize-low_buffer_size); + } + else + memcpy((char*)LOW_BUFFER_START, high_buffer_start, uncompressedSize ); + bytes_out = i; + return 0; + } + return 1; +} + + +#ifdef _LZMA_IN_CB +static int read_byte(void *object, unsigned char **buffer, UInt32 *bufferSize) +{ + static unsigned int i = 0; + static unsigned char val; + *bufferSize = 1; + val = get_byte(); + *buffer = &val; + if ( i++ % ( 1024 * 50 ) == 0 ) + puts("."); + return LZMA_RESULT_OK; +} +#endif + +asmlinkage int decompress_kernel(struct moveparams *mv, void *rmode) +{ + real_mode = rmode; + + if (SCREEN_INFO.orig_video_mode == 7) { + vidmem = (char *) 0xb0000; + vidport = 0x3b4; + } else { + vidmem = (char *) 0xb8000; + vidport = 0x3d4; + } + + lines = SCREEN_INFO.orig_video_lines; + cols = SCREEN_INFO.orig_video_cols; + + if (free_mem_ptr < 0x100000) setup_normal_output_buffer(); + else setup_output_buffer_if_we_run_high(mv); + + puts("LZMA vmlinuz: Ming-Ching Tiew ..."); + if( lzma_unzip() != 0 ) + { + error("inflate error\n"); + } + puts("Ok, booting the kernel.\n"); + if (high_loaded) close_output_buffer_if_we_run_high(mv); + return high_loaded; +}