Google Git
Sign in
cos / third_party / overlays / chromiumos-overlay / c8ee175d572d3111dd27c308d654fc67d9018248 / . / sys-boot / edk2 / files / 19_BaseTools-Lzma-Update-LZMA-SDK-version-to-18.05.patch
blob: 9bd0639f32409a5d5b7b5c0666b64727715653f0 [file] [log] [blame]
From 5ec5a236d1fe64b2c40075b6738b74bdf7e1acdc Mon Sep 17 00:00:00 2001
From: Liming Gao <liming.gao@intel.com>
Date: Wed, 29 Aug 2018 08:51:28 +0800
Subject: [PATCH] BaseTools Lzma: Update LZMA SDK version to 18.05
https://bugzilla.tianocore.org/show_bug.cgi?id=1006
New formal release in https://www.7-zip.org/sdk.html is 18.05.
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Liming Gao <liming.gao@intel.com>
Reviewed-by: Star Zeng <star.zeng@intel.com>
---
.../Source/C/LzmaCompress/LZMA-SDK-README.txt | 4 +-
.../Source/C/LzmaCompress/LzmaCompress.c | 8 +-
.../Source/C/LzmaCompress/Sdk/C/7zFile.c | 28 +-
.../Source/C/LzmaCompress/Sdk/C/7zFile.h | 8 +-
.../Source/C/LzmaCompress/Sdk/C/7zStream.c | 111 +-
.../Source/C/LzmaCompress/Sdk/C/7zTypes.h | 220 +-
.../Source/C/LzmaCompress/Sdk/C/7zVersion.h | 22 +-
BaseTools/Source/C/LzmaCompress/Sdk/C/Alloc.c | 399 ++-
BaseTools/Source/C/LzmaCompress/Sdk/C/Alloc.h | 20 +-
BaseTools/Source/C/LzmaCompress/Sdk/C/Bra86.c | 4 +-
.../Source/C/LzmaCompress/Sdk/C/Compiler.h | 3 +-
.../Source/C/LzmaCompress/Sdk/C/CpuArch.h | 166 +-
.../Source/C/LzmaCompress/Sdk/C/LzFind.c | 163 +-
.../Source/C/LzmaCompress/Sdk/C/LzFind.h | 12 +-
.../Source/C/LzmaCompress/Sdk/C/LzFindMt.c | 67 +-
.../Source/C/LzmaCompress/Sdk/C/LzFindMt.h | 6 +-
.../Source/C/LzmaCompress/Sdk/C/LzmaDec.c | 401 +--
.../Source/C/LzmaCompress/Sdk/C/LzmaDec.h | 47 +-
.../Source/C/LzmaCompress/Sdk/C/LzmaEnc.c | 2334 ++++++++++-------
.../Source/C/LzmaCompress/Sdk/C/LzmaEnc.h | 48 +-
.../Source/C/LzmaCompress/Sdk/C/Threads.c | 14 +-
.../Source/C/LzmaCompress/Sdk/C/Threads.h | 5 +-
.../C/LzmaCompress/Sdk/DOC/lzma-history.txt | 63 +-
.../C/LzmaCompress/Sdk/DOC/lzma-sdk.txt | 2 +-
24 files changed, 2684 insertions(+), 1471 deletions(-)
diff --git a/BaseTools/Source/C/LzmaCompress/LZMA-SDK-README.txt b/BaseTools/Source/C/LzmaCompress/LZMA-SDK-README.txt
index bc9e29fd9f..7cf967a774 100644
--- a/BaseTools/Source/C/LzmaCompress/LZMA-SDK-README.txt
+++ b/BaseTools/Source/C/LzmaCompress/LZMA-SDK-README.txt
@@ -1,3 +1,3 @@
-LzmaCompress is based on the LZMA SDK 16.04. LZMA SDK 16.04
-was placed in the public domain on 2016-10-04. It was
+LzmaCompress is based on the LZMA SDK 18.05. LZMA SDK 18.05
+was placed in the public domain on 2018-04-30. It was
released on the http://www.7-zip.org/sdk.html website.
diff --git a/BaseTools/Source/C/LzmaCompress/LzmaCompress.c b/BaseTools/Source/C/LzmaCompress/LzmaCompress.c
index ceb6a5bc77..da09b9f748 100644
--- a/BaseTools/Source/C/LzmaCompress/LzmaCompress.c
+++ b/BaseTools/Source/C/LzmaCompress/LzmaCompress.c
@@ -1,9 +1,9 @@
/** @file
LZMA Compress/Decompress tool (LzmaCompress)
- Based on LZMA SDK 16.04:
+ Based on LZMA SDK 18.05:
LzmaUtil.c -- Test application for LZMA compression
- 2016-10-04 : Igor Pavlov : Public domain
+ 2018-04-30 : Igor Pavlov : Public domain
Copyright (c) 2006 - 2018, Intel Corporation. All rights reserved.<BR>
This program and the accompanying materials
@@ -341,14 +341,14 @@ int main2(int numArgs, const char *args[], char *rs)
if (!mQuietMode) {
printf("Encoding\n");
}
- res = Encode(&outStream.s, &inStream.s, fileSize);
+ res = Encode(&outStream.vt, &inStream.vt, fileSize);
}
else
{
if (!mQuietMode) {
printf("Decoding\n");
}
- res = Decode(&outStream.s, &inStream.s, fileSize);
+ res = Decode(&outStream.vt, &inStream.vt, fileSize);
}
File_Close(&outStream.file);
diff --git a/BaseTools/Source/C/LzmaCompress/Sdk/C/7zFile.c b/BaseTools/Source/C/LzmaCompress/Sdk/C/7zFile.c
index b2fed368a4..e486901e30 100644
--- a/BaseTools/Source/C/LzmaCompress/Sdk/C/7zFile.c
+++ b/BaseTools/Source/C/LzmaCompress/Sdk/C/7zFile.c
@@ -1,5 +1,5 @@
/* 7zFile.c -- File IO
-2009-11-24 : Igor Pavlov : Public domain */
+2017-04-03 : Igor Pavlov : Public domain */
#include "Precomp.h"
@@ -213,7 +213,7 @@ WRes File_GetLength(CSzFile *p, UInt64 *length)
{
#ifdef USE_WINDOWS_FILE
- DWORD sizeHigh = 0;
+ DWORD sizeHigh;
DWORD sizeLow = GetFileSize(p->handle, &sizeHigh);
if (sizeLow == 0xFFFFFFFF)
{
@@ -238,49 +238,49 @@ WRes File_GetLength(CSzFile *p, UInt64 *length)
/* ---------- FileSeqInStream ---------- */
-static SRes FileSeqInStream_Read(void *pp, void *buf, size_t *size)
+static SRes FileSeqInStream_Read(const ISeqInStream *pp, void *buf, size_t *size)
{
- CFileSeqInStream *p = (CFileSeqInStream *)pp;
+ CFileSeqInStream *p = CONTAINER_FROM_VTBL(pp, CFileSeqInStream, vt);
return File_Read(&p->file, buf, size) == 0 ? SZ_OK : SZ_ERROR_READ;
}
void FileSeqInStream_CreateVTable(CFileSeqInStream *p)
{
- p->s.Read = FileSeqInStream_Read;
+ p->vt.Read = FileSeqInStream_Read;
}
/* ---------- FileInStream ---------- */
-static SRes FileInStream_Read(void *pp, void *buf, size_t *size)
+static SRes FileInStream_Read(const ISeekInStream *pp, void *buf, size_t *size)
{
- CFileInStream *p = (CFileInStream *)pp;
+ CFileInStream *p = CONTAINER_FROM_VTBL(pp, CFileInStream, vt);
return (File_Read(&p->file, buf, size) == 0) ? SZ_OK : SZ_ERROR_READ;
}
-static SRes FileInStream_Seek(void *pp, Int64 *pos, ESzSeek origin)
+static SRes FileInStream_Seek(const ISeekInStream *pp, Int64 *pos, ESzSeek origin)
{
- CFileInStream *p = (CFileInStream *)pp;
+ CFileInStream *p = CONTAINER_FROM_VTBL(pp, CFileInStream, vt);
return File_Seek(&p->file, pos, origin);
}
void FileInStream_CreateVTable(CFileInStream *p)
{
- p->s.Read = FileInStream_Read;
- p->s.Seek = FileInStream_Seek;
+ p->vt.Read = FileInStream_Read;
+ p->vt.Seek = FileInStream_Seek;
}
/* ---------- FileOutStream ---------- */
-static size_t FileOutStream_Write(void *pp, const void *data, size_t size)
+static size_t FileOutStream_Write(const ISeqOutStream *pp, const void *data, size_t size)
{
- CFileOutStream *p = (CFileOutStream *)pp;
+ CFileOutStream *p = CONTAINER_FROM_VTBL(pp, CFileOutStream, vt);
File_Write(&p->file, data, &size);
return size;
}
void FileOutStream_CreateVTable(CFileOutStream *p)
{
- p->s.Write = FileOutStream_Write;
+ p->vt.Write = FileOutStream_Write;
}
diff --git a/BaseTools/Source/C/LzmaCompress/Sdk/C/7zFile.h b/BaseTools/Source/C/LzmaCompress/Sdk/C/7zFile.h
index d62a192609..7e263bea1b 100644
--- a/BaseTools/Source/C/LzmaCompress/Sdk/C/7zFile.h
+++ b/BaseTools/Source/C/LzmaCompress/Sdk/C/7zFile.h
@@ -1,5 +1,5 @@
/* 7zFile.h -- File IO
-2013-01-18 : Igor Pavlov : Public domain */
+2017-04-03 : Igor Pavlov : Public domain */
#ifndef __7Z_FILE_H
#define __7Z_FILE_H
@@ -54,7 +54,7 @@ WRes File_GetLength(CSzFile *p, UInt64 *length);
typedef struct
{
- ISeqInStream s;
+ ISeqInStream vt;
CSzFile file;
} CFileSeqInStream;
@@ -63,7 +63,7 @@ void FileSeqInStream_CreateVTable(CFileSeqInStream *p);
typedef struct
{
- ISeekInStream s;
+ ISeekInStream vt;
CSzFile file;
} CFileInStream;
@@ -72,7 +72,7 @@ void FileInStream_CreateVTable(CFileInStream *p);
typedef struct
{
- ISeqOutStream s;
+ ISeqOutStream vt;
CSzFile file;
} CFileOutStream;
diff --git a/BaseTools/Source/C/LzmaCompress/Sdk/C/7zStream.c b/BaseTools/Source/C/LzmaCompress/Sdk/C/7zStream.c
index 5a92d532cc..579741fadc 100644
--- a/BaseTools/Source/C/LzmaCompress/Sdk/C/7zStream.c
+++ b/BaseTools/Source/C/LzmaCompress/Sdk/C/7zStream.c
@@ -1,5 +1,5 @@
/* 7zStream.c -- 7z Stream functions
-2013-11-12 : Igor Pavlov : Public domain */
+2017-04-03 : Igor Pavlov : Public domain */
#include "Precomp.h"
@@ -7,12 +7,12 @@
#include "7zTypes.h"
-SRes SeqInStream_Read2(ISeqInStream *stream, void *buf, size_t size, SRes errorType)
+SRes SeqInStream_Read2(const ISeqInStream *stream, void *buf, size_t size, SRes errorType)
{
while (size != 0)
{
size_t processed = size;
- RINOK(stream->Read(stream, buf, &processed));
+ RINOK(ISeqInStream_Read(stream, buf, &processed));
if (processed == 0)
return errorType;
buf = (void *)((Byte *)buf + processed);
@@ -21,40 +21,42 @@ SRes SeqInStream_Read2(ISeqInStream *stream, void *buf, size_t size, SRes errorT
return SZ_OK;
}
-SRes SeqInStream_Read(ISeqInStream *stream, void *buf, size_t size)
+SRes SeqInStream_Read(const ISeqInStream *stream, void *buf, size_t size)
{
return SeqInStream_Read2(stream, buf, size, SZ_ERROR_INPUT_EOF);
}
-SRes SeqInStream_ReadByte(ISeqInStream *stream, Byte *buf)
+SRes SeqInStream_ReadByte(const ISeqInStream *stream, Byte *buf)
{
size_t processed = 1;
- RINOK(stream->Read(stream, buf, &processed));
+ RINOK(ISeqInStream_Read(stream, buf, &processed));
return (processed == 1) ? SZ_OK : SZ_ERROR_INPUT_EOF;
}
-SRes LookInStream_SeekTo(ILookInStream *stream, UInt64 offset)
+
+
+SRes LookInStream_SeekTo(const ILookInStream *stream, UInt64 offset)
{
Int64 t = offset;
- return stream->Seek(stream, &t, SZ_SEEK_SET);
+ return ILookInStream_Seek(stream, &t, SZ_SEEK_SET);
}
-SRes LookInStream_LookRead(ILookInStream *stream, void *buf, size_t *size)
+SRes LookInStream_LookRead(const ILookInStream *stream, void *buf, size_t *size)
{
const void *lookBuf;
if (*size == 0)
return SZ_OK;
- RINOK(stream->Look(stream, &lookBuf, size));
+ RINOK(ILookInStream_Look(stream, &lookBuf, size));
memcpy(buf, lookBuf, *size);
- return stream->Skip(stream, *size);
+ return ILookInStream_Skip(stream, *size);
}
-SRes LookInStream_Read2(ILookInStream *stream, void *buf, size_t size, SRes errorType)
+SRes LookInStream_Read2(const ILookInStream *stream, void *buf, size_t size, SRes errorType)
{
while (size != 0)
{
size_t processed = size;
- RINOK(stream->Read(stream, buf, &processed));
+ RINOK(ILookInStream_Read(stream, buf, &processed));
if (processed == 0)
return errorType;
buf = (void *)((Byte *)buf + processed);
@@ -63,61 +65,67 @@ SRes LookInStream_Read2(ILookInStream *stream, void *buf, size_t size, SRes erro
return SZ_OK;
}
-SRes LookInStream_Read(ILookInStream *stream, void *buf, size_t size)
+SRes LookInStream_Read(const ILookInStream *stream, void *buf, size_t size)
{
return LookInStream_Read2(stream, buf, size, SZ_ERROR_INPUT_EOF);
}
-static SRes LookToRead_Look_Lookahead(void *pp, const void **buf, size_t *size)
+
+
+#define GET_LookToRead2 CLookToRead2 *p = CONTAINER_FROM_VTBL(pp, CLookToRead2, vt);
+
+static SRes LookToRead2_Look_Lookahead(const ILookInStream *pp, const void **buf, size_t *size)
{
SRes res = SZ_OK;
- CLookToRead *p = (CLookToRead *)pp;
+ GET_LookToRead2
size_t size2 = p->size - p->pos;
- if (size2 == 0 && *size > 0)
+ if (size2 == 0 && *size != 0)
{
p->pos = 0;
- size2 = LookToRead_BUF_SIZE;
- res = p->realStream->Read(p->realStream, p->buf, &size2);
+ p->size = 0;
+ size2 = p->bufSize;
+ res = ISeekInStream_Read(p->realStream, p->buf, &size2);
p->size = size2;
}
- if (size2 < *size)
+ if (*size > size2)
*size = size2;
*buf = p->buf + p->pos;
return res;
}
-static SRes LookToRead_Look_Exact(void *pp, const void **buf, size_t *size)
+static SRes LookToRead2_Look_Exact(const ILookInStream *pp, const void **buf, size_t *size)
{
SRes res = SZ_OK;
- CLookToRead *p = (CLookToRead *)pp;
+ GET_LookToRead2
size_t size2 = p->size - p->pos;
- if (size2 == 0 && *size > 0)
+ if (size2 == 0 && *size != 0)
{
p->pos = 0;
- if (*size > LookToRead_BUF_SIZE)
- *size = LookToRead_BUF_SIZE;
- res = p->realStream->Read(p->realStream, p->buf, size);
+ p->size = 0;
+ if (*size > p->bufSize)
+ *size = p->bufSize;
+ res = ISeekInStream_Read(p->realStream, p->buf, size);
size2 = p->size = *size;
}
- if (size2 < *size)
+ if (*size > size2)
*size = size2;
*buf = p->buf + p->pos;
return res;
}
-static SRes LookToRead_Skip(void *pp, size_t offset)
+static SRes LookToRead2_Skip(const ILookInStream *pp, size_t offset)
{
- CLookToRead *p = (CLookToRead *)pp;
+ GET_LookToRead2
p->pos += offset;
return SZ_OK;
}
-static SRes LookToRead_Read(void *pp, void *buf, size_t *size)
+static SRes LookToRead2_Read(const ILookInStream *pp, void *buf, size_t *size)
{
- CLookToRead *p = (CLookToRead *)pp;
+ GET_LookToRead2
size_t rem = p->size - p->pos;
if (rem == 0)
- return p->realStream->Read(p->realStream, buf, size);
+ return ISeekInStream_Read(p->realStream, buf, size);
if (rem > *size)
rem = *size;
memcpy(buf, p->buf + p->pos, rem);
@@ -126,46 +134,43 @@ static SRes LookToRead_Read(void *pp, void *buf, size_t *size)
return SZ_OK;
}
-static SRes LookToRead_Seek(void *pp, Int64 *pos, ESzSeek origin)
+static SRes LookToRead2_Seek(const ILookInStream *pp, Int64 *pos, ESzSeek origin)
{
- CLookToRead *p = (CLookToRead *)pp;
+ GET_LookToRead2
p->pos = p->size = 0;
- return p->realStream->Seek(p->realStream, pos, origin);
+ return ISeekInStream_Seek(p->realStream, pos, origin);
}
-void LookToRead_CreateVTable(CLookToRead *p, int lookahead)
+void LookToRead2_CreateVTable(CLookToRead2 *p, int lookahead)
{
- p->s.Look = lookahead ?
- LookToRead_Look_Lookahead :
- LookToRead_Look_Exact;
- p->s.Skip = LookToRead_Skip;
- p->s.Read = LookToRead_Read;
- p->s.Seek = LookToRead_Seek;
+ p->vt.Look = lookahead ?
+ LookToRead2_Look_Lookahead :
+ LookToRead2_Look_Exact;
+ p->vt.Skip = LookToRead2_Skip;
+ p->vt.Read = LookToRead2_Read;
+ p->vt.Seek = LookToRead2_Seek;
}
-void LookToRead_Init(CLookToRead *p)
-{
- p->pos = p->size = 0;
-}
-static SRes SecToLook_Read(void *pp, void *buf, size_t *size)
+
+static SRes SecToLook_Read(const ISeqInStream *pp, void *buf, size_t *size)
{
- CSecToLook *p = (CSecToLook *)pp;
+ CSecToLook *p = CONTAINER_FROM_VTBL(pp, CSecToLook, vt);
return LookInStream_LookRead(p->realStream, buf, size);
}
void SecToLook_CreateVTable(CSecToLook *p)
{
- p->s.Read = SecToLook_Read;
+ p->vt.Read = SecToLook_Read;
}
-static SRes SecToRead_Read(void *pp, void *buf, size_t *size)
+static SRes SecToRead_Read(const ISeqInStream *pp, void *buf, size_t *size)
{
- CSecToRead *p = (CSecToRead *)pp;
- return p->realStream->Read(p->realStream, buf, size);
+ CSecToRead *p = CONTAINER_FROM_VTBL(pp, CSecToRead, vt);
+ return ILookInStream_Read(p->realStream, buf, size);
}
void SecToRead_CreateVTable(CSecToRead *p)
{
- p->s.Read = SecToRead_Read;
+ p->vt.Read = SecToRead_Read;
}
diff --git a/BaseTools/Source/C/LzmaCompress/Sdk/C/7zTypes.h b/BaseTools/Source/C/LzmaCompress/Sdk/C/7zTypes.h
index 903047b10f..4977cdaa66 100644
--- a/BaseTools/Source/C/LzmaCompress/Sdk/C/7zTypes.h
+++ b/BaseTools/Source/C/LzmaCompress/Sdk/C/7zTypes.h
@@ -1,5 +1,5 @@
/* 7zTypes.h -- Basic types
-2013-11-12 : Igor Pavlov : Public domain */
+2017-07-17 : Igor Pavlov : Public domain */
#ifndef __7Z_TYPES_H
#define __7Z_TYPES_H
@@ -42,13 +42,23 @@ EXTERN_C_BEGIN
typedef int SRes;
+
#ifdef _WIN32
+
/* typedef DWORD WRes; */
typedef unsigned WRes;
+#define MY_SRes_HRESULT_FROM_WRes(x) HRESULT_FROM_WIN32(x)
+
#else
+
typedef int WRes;
+#define MY__FACILITY_WIN32 7
+#define MY__FACILITY__WRes MY__FACILITY_WIN32
+#define MY_SRes_HRESULT_FROM_WRes(x) ((HRESULT)(x) <= 0 ? ((HRESULT)(x)) : ((HRESULT) (((x) & 0x0000FFFF) | (MY__FACILITY__WRes << 16) | 0x80000000)))
+
#endif
+
#ifndef RINOK
#define RINOK(x) { int __result__ = (x); if (__result__ != 0) return __result__; }
#endif
@@ -112,48 +122,72 @@ typedef int Bool;
#define MY_NO_INLINE
#endif
+#define MY_FORCE_INLINE __forceinline
+
#define MY_CDECL __cdecl
#define MY_FAST_CALL __fastcall
#else
#define MY_NO_INLINE
+#define MY_FORCE_INLINE
#define MY_CDECL
#define MY_FAST_CALL
+/* inline keyword : for C++ / C99 */
+
+/* GCC, clang: */
+/*
+#if defined (__GNUC__) && (__GNUC__ >= 4)
+#define MY_FORCE_INLINE __attribute__((always_inline))
+#define MY_NO_INLINE __attribute__((noinline))
+#endif
+*/
+
#endif
/* The following interfaces use first parameter as pointer to structure */
-typedef struct
+typedef struct IByteIn IByteIn;
+struct IByteIn
{
- Byte (*Read)(void *p); /* reads one byte, returns 0 in case of EOF or error */
-} IByteIn;
+ Byte (*Read)(const IByteIn *p); /* reads one byte, returns 0 in case of EOF or error */
+};
+#define IByteIn_Read(p) (p)->Read(p)
-typedef struct
+
+typedef struct IByteOut IByteOut;
+struct IByteOut
{
- void (*Write)(void *p, Byte b);
-} IByteOut;
+ void (*Write)(const IByteOut *p, Byte b);
+};
+#define IByteOut_Write(p, b) (p)->Write(p, b)
-typedef struct
+
+typedef struct ISeqInStream ISeqInStream;
+struct ISeqInStream
{
- SRes (*Read)(void *p, void *buf, size_t *size);
+ SRes (*Read)(const ISeqInStream *p, void *buf, size_t *size);
/* if (input(*size) != 0 && output(*size) == 0) means end_of_stream.
(output(*size) < input(*size)) is allowed */
-} ISeqInStream;
+};
+#define ISeqInStream_Read(p, buf, size) (p)->Read(p, buf, size)
/* it can return SZ_ERROR_INPUT_EOF */
-SRes SeqInStream_Read(ISeqInStream *stream, void *buf, size_t size);
-SRes SeqInStream_Read2(ISeqInStream *stream, void *buf, size_t size, SRes errorType);
-SRes SeqInStream_ReadByte(ISeqInStream *stream, Byte *buf);
+SRes SeqInStream_Read(const ISeqInStream *stream, void *buf, size_t size);
+SRes SeqInStream_Read2(const ISeqInStream *stream, void *buf, size_t size, SRes errorType);
+SRes SeqInStream_ReadByte(const ISeqInStream *stream, Byte *buf);
-typedef struct
+
+typedef struct ISeqOutStream ISeqOutStream;
+struct ISeqOutStream
{
- size_t (*Write)(void *p, const void *buf, size_t size);
+ size_t (*Write)(const ISeqOutStream *p, const void *buf, size_t size);
/* Returns: result - the number of actually written bytes.
(result < size) means error */
-} ISeqOutStream;
+};
+#define ISeqOutStream_Write(p, buf, size) (p)->Write(p, buf, size)
typedef enum
{
@@ -162,78 +196,162 @@ typedef enum
SZ_SEEK_END = 2
} ESzSeek;
-typedef struct
+
+typedef struct ISeekInStream ISeekInStream;
+struct ISeekInStream
{
- SRes (*Read)(void *p, void *buf, size_t *size); /* same as ISeqInStream::Read */
- SRes (*Seek)(void *p, Int64 *pos, ESzSeek origin);
-} ISeekInStream;
+ SRes (*Read)(const ISeekInStream *p, void *buf, size_t *size); /* same as ISeqInStream::Read */
+ SRes (*Seek)(const ISeekInStream *p, Int64 *pos, ESzSeek origin);
+};
+#define ISeekInStream_Read(p, buf, size) (p)->Read(p, buf, size)
+#define ISeekInStream_Seek(p, pos, origin) (p)->Seek(p, pos, origin)
-typedef struct
+
+typedef struct ILookInStream ILookInStream;
+struct ILookInStream
{
- SRes (*Look)(void *p, const void **buf, size_t *size);
+ SRes (*Look)(const ILookInStream *p, const void **buf, size_t *size);
/* if (input(*size) != 0 && output(*size) == 0) means end_of_stream.
(output(*size) > input(*size)) is not allowed
(output(*size) < input(*size)) is allowed */
- SRes (*Skip)(void *p, size_t offset);
+ SRes (*Skip)(const ILookInStream *p, size_t offset);
/* offset must be <= output(*size) of Look */
- SRes (*Read)(void *p, void *buf, size_t *size);
+ SRes (*Read)(const ILookInStream *p, void *buf, size_t *size);
/* reads directly (without buffer). It's same as ISeqInStream::Read */
- SRes (*Seek)(void *p, Int64 *pos, ESzSeek origin);
-} ILookInStream;
+ SRes (*Seek)(const ILookInStream *p, Int64 *pos, ESzSeek origin);
+};
-SRes LookInStream_LookRead(ILookInStream *stream, void *buf, size_t *size);
-SRes LookInStream_SeekTo(ILookInStream *stream, UInt64 offset);
+#define ILookInStream_Look(p, buf, size) (p)->Look(p, buf, size)
+#define ILookInStream_Skip(p, offset) (p)->Skip(p, offset)
+#define ILookInStream_Read(p, buf, size) (p)->Read(p, buf, size)
+#define ILookInStream_Seek(p, pos, origin) (p)->Seek(p, pos, origin)
+
+
+SRes LookInStream_LookRead(const ILookInStream *stream, void *buf, size_t *size);
+SRes LookInStream_SeekTo(const ILookInStream *stream, UInt64 offset);
/* reads via ILookInStream::Read */
-SRes LookInStream_Read2(ILookInStream *stream, void *buf, size_t size, SRes errorType);
-SRes LookInStream_Read(ILookInStream *stream, void *buf, size_t size);
+SRes LookInStream_Read2(const ILookInStream *stream, void *buf, size_t size, SRes errorType);
+SRes LookInStream_Read(const ILookInStream *stream, void *buf, size_t size);
+
-#define LookToRead_BUF_SIZE (1 << 14)
typedef struct
{
- ILookInStream s;
- ISeekInStream *realStream;
+ ILookInStream vt;
+ const ISeekInStream *realStream;
+
size_t pos;
- size_t size;
- Byte buf[LookToRead_BUF_SIZE];
-} CLookToRead;
+ size_t size; /* it's data size */
+
+ /* the following variables must be set outside */
+ Byte *buf;
+ size_t bufSize;
+} CLookToRead2;
+
+void LookToRead2_CreateVTable(CLookToRead2 *p, int lookahead);
+
+#define LookToRead2_Init(p) { (p)->pos = (p)->size = 0; }
-void LookToRead_CreateVTable(CLookToRead *p, int lookahead);
-void LookToRead_Init(CLookToRead *p);
typedef struct
{
- ISeqInStream s;
- ILookInStream *realStream;
+ ISeqInStream vt;
+ const ILookInStream *realStream;
} CSecToLook;
void SecToLook_CreateVTable(CSecToLook *p);
+
+
typedef struct
{
- ISeqInStream s;
- ILookInStream *realStream;
+ ISeqInStream vt;
+ const ILookInStream *realStream;
} CSecToRead;
void SecToRead_CreateVTable(CSecToRead *p);
-typedef struct
+
+typedef struct ICompressProgress ICompressProgress;
+
+struct ICompressProgress
{
- SRes (*Progress)(void *p, UInt64 inSize, UInt64 outSize);
+ SRes (*Progress)(const ICompressProgress *p, UInt64 inSize, UInt64 outSize);
/* Returns: result. (result != SZ_OK) means break.
Value (UInt64)(Int64)-1 for size means unknown value. */
-} ICompressProgress;
+};
+#define ICompressProgress_Progress(p, inSize, outSize) (p)->Progress(p, inSize, outSize)
-typedef struct
+
+
+typedef struct ISzAlloc ISzAlloc;
+typedef const ISzAlloc * ISzAllocPtr;
+
+struct ISzAlloc
{
- void *(*Alloc)(void *p, size_t size);
- void (*Free)(void *p, void *address); /* address can be 0 */
-} ISzAlloc;
+ void *(*Alloc)(ISzAllocPtr p, size_t size);
+ void (*Free)(ISzAllocPtr p, void *address); /* address can be 0 */
+};
+
+#define ISzAlloc_Alloc(p, size) (p)->Alloc(p, size)
+#define ISzAlloc_Free(p, a) (p)->Free(p, a)
+
+/* deprecated */
+#define IAlloc_Alloc(p, size) ISzAlloc_Alloc(p, size)
+#define IAlloc_Free(p, a) ISzAlloc_Free(p, a)
+
+
+
+
+
+#ifndef MY_offsetof
+ #ifdef offsetof
+ #define MY_offsetof(type, m) offsetof(type, m)
+ /*
+ #define MY_offsetof(type, m) FIELD_OFFSET(type, m)
+ */
+ #else
+ #define MY_offsetof(type, m) ((size_t)&(((type *)0)->m))
+ #endif
+#endif
+
+
+
+#ifndef MY_container_of
+
+/*
+#define MY_container_of(ptr, type, m) container_of(ptr, type, m)
+#define MY_container_of(ptr, type, m) CONTAINING_RECORD(ptr, type, m)
+#define MY_container_of(ptr, type, m) ((type *)((char *)(ptr) - offsetof(type, m)))
+#define MY_container_of(ptr, type, m) (&((type *)0)->m == (ptr), ((type *)(((char *)(ptr)) - MY_offsetof(type, m))))
+*/
+
+/*
+ GCC shows warning: "perhaps the 'offsetof' macro was used incorrectly"
+ GCC 3.4.4 : classes with constructor
+ GCC 4.8.1 : classes with non-public variable members"
+*/
+
+#define MY_container_of(ptr, type, m) ((type *)((char *)(1 ? (ptr) : &((type *)0)->m) - MY_offsetof(type, m)))
+
+
+#endif
+
+#define CONTAINER_FROM_VTBL_SIMPLE(ptr, type, m) ((type *)(ptr))
+
+/*
+#define CONTAINER_FROM_VTBL(ptr, type, m) CONTAINER_FROM_VTBL_SIMPLE(ptr, type, m)
+*/
+#define CONTAINER_FROM_VTBL(ptr, type, m) MY_container_of(ptr, type, m)
+
+#define CONTAINER_FROM_VTBL_CLS(ptr, type, m) CONTAINER_FROM_VTBL_SIMPLE(ptr, type, m)
+/*
+#define CONTAINER_FROM_VTBL_CLS(ptr, type, m) CONTAINER_FROM_VTBL(ptr, type, m)
+*/
+
-#define IAlloc_Alloc(p, size) (p)->Alloc((p), size)
-#define IAlloc_Free(p, a) (p)->Free((p), a)
#ifdef _WIN32
diff --git a/BaseTools/Source/C/LzmaCompress/Sdk/C/7zVersion.h b/BaseTools/Source/C/LzmaCompress/Sdk/C/7zVersion.h
index acb67a94e5..ed3aa94270 100644
--- a/BaseTools/Source/C/LzmaCompress/Sdk/C/7zVersion.h
+++ b/BaseTools/Source/C/LzmaCompress/Sdk/C/7zVersion.h
@@ -1,14 +1,21 @@
-#define MY_VER_MAJOR 16
-#define MY_VER_MINOR 04
+#define MY_VER_MAJOR 18
+#define MY_VER_MINOR 05
#define MY_VER_BUILD 0
-#define MY_VERSION_NUMBERS "16.04"
-#define MY_VERSION "16.04"
-#define MY_DATE "2016-10-04"
+#define MY_VERSION_NUMBERS "18.05"
+#define MY_VERSION MY_VERSION_NUMBERS
+
+#ifdef MY_CPU_NAME
+ #define MY_VERSION_CPU MY_VERSION " (" MY_CPU_NAME ")"
+#else
+ #define MY_VERSION_CPU MY_VERSION
+#endif
+
+#define MY_DATE "2018-04-30"
#undef MY_COPYRIGHT
#undef MY_VERSION_COPYRIGHT_DATE
#define MY_AUTHOR_NAME "Igor Pavlov"
#define MY_COPYRIGHT_PD "Igor Pavlov : Public domain"
-#define MY_COPYRIGHT_CR "Copyright (c) 1999-2016 Igor Pavlov"
+#define MY_COPYRIGHT_CR "Copyright (c) 1999-2018 Igor Pavlov"
#ifdef USE_COPYRIGHT_CR
#define MY_COPYRIGHT MY_COPYRIGHT_CR
@@ -16,4 +23,5 @@
#define MY_COPYRIGHT MY_COPYRIGHT_PD
#endif
-#define MY_VERSION_COPYRIGHT_DATE MY_VERSION " : " MY_COPYRIGHT " : " MY_DATE
+#define MY_COPYRIGHT_DATE MY_COPYRIGHT " : " MY_DATE
+#define MY_VERSION_COPYRIGHT_DATE MY_VERSION_CPU " : " MY_COPYRIGHT " : " MY_DATE
diff --git a/BaseTools/Source/C/LzmaCompress/Sdk/C/Alloc.c b/BaseTools/Source/C/LzmaCompress/Sdk/C/Alloc.c
index 9f1d036afe..30b499e5ff 100644
--- a/BaseTools/Source/C/LzmaCompress/Sdk/C/Alloc.c
+++ b/BaseTools/Source/C/LzmaCompress/Sdk/C/Alloc.c
@@ -1,8 +1,10 @@
/* Alloc.c -- Memory allocation functions
-2015-02-21 : Igor Pavlov : Public domain */
+2018-04-27 : Igor Pavlov : Public domain */
#include "Precomp.h"
+#include <stdio.h>
+
#ifdef _WIN32
#include <windows.h>
#endif
@@ -14,20 +16,127 @@
/* use _SZ_ALLOC_DEBUG to debug alloc/free operations */
#ifdef _SZ_ALLOC_DEBUG
+
#include <stdio.h>
int g_allocCount = 0;
int g_allocCountMid = 0;
int g_allocCountBig = 0;
+
+
+#define CONVERT_INT_TO_STR(charType, tempSize) \
+ unsigned char temp[tempSize]; unsigned i = 0; \
+ while (val >= 10) { temp[i++] = (unsigned char)('0' + (unsigned)(val % 10)); val /= 10; } \
+ *s++ = (charType)('0' + (unsigned)val); \
+ while (i != 0) { i--; *s++ = temp[i]; } \
+ *s = 0;
+
+static void ConvertUInt64ToString(UInt64 val, char *s)
+{
+ CONVERT_INT_TO_STR(char, 24);
+}
+
+#define GET_HEX_CHAR(t) ((char)(((t < 10) ? ('0' + t) : ('A' + (t - 10)))))
+
+static void ConvertUInt64ToHex(UInt64 val, char *s)
+{
+ UInt64 v = val;
+ unsigned i;
+ for (i = 1;; i++)
+ {
+ v >>= 4;
+ if (v == 0)
+ break;
+ }
+ s[i] = 0;
+ do
+ {
+ unsigned t = (unsigned)(val & 0xF);
+ val >>= 4;
+ s[--i] = GET_HEX_CHAR(t);
+ }
+ while (i);
+}
+
+#define DEBUG_OUT_STREAM stderr
+
+static void Print(const char *s)
+{
+ fputs(s, DEBUG_OUT_STREAM);
+}
+
+static void PrintAligned(const char *s, size_t align)
+{
+ size_t len = strlen(s);
+ for(;;)
+ {
+ fputc(' ', DEBUG_OUT_STREAM);
+ if (len >= align)
+ break;
+ ++len;
+ }
+ Print(s);
+}
+
+static void PrintLn()
+{
+ Print("\n");
+}
+
+static void PrintHex(UInt64 v, size_t align)
+{
+ char s[32];
+ ConvertUInt64ToHex(v, s);
+ PrintAligned(s, align);
+}
+
+static void PrintDec(UInt64 v, size_t align)
+{
+ char s[32];
+ ConvertUInt64ToString(v, s);
+ PrintAligned(s, align);
+}
+
+static void PrintAddr(void *p)
+{
+ PrintHex((UInt64)(size_t)(ptrdiff_t)p, 12);
+}
+
+
+#define PRINT_ALLOC(name, cnt, size, ptr) \
+ Print(name " "); \
+ PrintDec(cnt++, 10); \
+ PrintHex(size, 10); \
+ PrintAddr(ptr); \
+ PrintLn();
+
+#define PRINT_FREE(name, cnt, ptr) if (ptr) { \
+ Print(name " "); \
+ PrintDec(--cnt, 10); \
+ PrintAddr(ptr); \
+ PrintLn(); }
+
+#else
+
+#define PRINT_ALLOC(name, cnt, size, ptr)
+#define PRINT_FREE(name, cnt, ptr)
+#define Print(s)
+#define PrintLn()
+#define PrintHex(v, align)
+#define PrintDec(v, align)
+#define PrintAddr(p)
+
#endif
+
+
void *MyAlloc(size_t size)
{
if (size == 0)
- return 0;
+ return NULL;
#ifdef _SZ_ALLOC_DEBUG
{
void *p = malloc(size);
- fprintf(stderr, "\nAlloc %10d bytes, count = %10d, addr = %8X", size, g_allocCount++, (unsigned)p);
+ PRINT_ALLOC("Alloc ", g_allocCount, size, p);
return p;
}
#else
@@ -37,10 +146,8 @@ void *MyAlloc(size_t size)
void MyFree(void *address)
{
- #ifdef _SZ_ALLOC_DEBUG
- if (address != 0)
- fprintf(stderr, "\nFree; count = %10d, addr = %8X", --g_allocCount, (unsigned)address);
- #endif
+ PRINT_FREE("Free ", g_allocCount, address);
+
free(address);
}
@@ -49,20 +156,18 @@ void MyFree(void *address)
void *MidAlloc(size_t size)
{
if (size == 0)
- return 0;
- #ifdef _SZ_ALLOC_DEBUG
- fprintf(stderr, "\nAlloc_Mid %10d bytes; count = %10d", size, g_allocCountMid++);
- #endif
- return VirtualAlloc(0, size, MEM_COMMIT, PAGE_READWRITE);
+ return NULL;
+
+ PRINT_ALLOC("Alloc-Mid", g_allocCountMid, size, NULL);
+
+ return VirtualAlloc(NULL, size, MEM_COMMIT, PAGE_READWRITE);
}
void MidFree(void *address)
{
- #ifdef _SZ_ALLOC_DEBUG
- if (address != 0)
- fprintf(stderr, "\nFree_Mid; count = %10d", --g_allocCountMid);
- #endif
- if (address == 0)
+ PRINT_FREE("Free-Mid", g_allocCountMid, address);
+
+ if (!address)
return;
VirtualFree(address, 0, MEM_RELEASE);
}
@@ -79,10 +184,10 @@ typedef SIZE_T (WINAPI *GetLargePageMinimumP)();
void SetLargePageSize()
{
#ifdef _7ZIP_LARGE_PAGES
- SIZE_T size = 0;
+ SIZE_T size;
GetLargePageMinimumP largePageMinimum = (GetLargePageMinimumP)
GetProcAddress(GetModuleHandle(TEXT("kernel32.dll")), "GetLargePageMinimum");
- if (largePageMinimum == 0)
+ if (!largePageMinimum)
return;
size = largePageMinimum();
if (size == 0 || (size & (size - 1)) != 0)
@@ -95,31 +200,36 @@ void SetLargePageSize()
void *BigAlloc(size_t size)
{
if (size == 0)
- return 0;
- #ifdef _SZ_ALLOC_DEBUG
- fprintf(stderr, "\nAlloc_Big %10d bytes; count = %10d", size, g_allocCountBig++);
- #endif
+ return NULL;
+
+ PRINT_ALLOC("Alloc-Big", g_allocCountBig, size, NULL);
#ifdef _7ZIP_LARGE_PAGES
- if (g_LargePageSize != 0 && g_LargePageSize <= (1 << 30) && size >= (1 << 18))
{
- void *res = VirtualAlloc(0, (size + g_LargePageSize - 1) & (~(g_LargePageSize - 1)),
- MEM_COMMIT | MEM_LARGE_PAGES, PAGE_READWRITE);
- if (res != 0)
- return res;
+ SIZE_T ps = g_LargePageSize;
+ if (ps != 0 && ps <= (1 << 30) && size > (ps / 2))
+ {
+ size_t size2;
+ ps--;
+ size2 = (size + ps) & ~ps;
+ if (size2 >= size)
+ {
+ void *res = VirtualAlloc(NULL, size2, MEM_COMMIT | MEM_LARGE_PAGES, PAGE_READWRITE);
+ if (res)
+ return res;
+ }
+ }
}
#endif
- return VirtualAlloc(0, size, MEM_COMMIT, PAGE_READWRITE);
+
+ return VirtualAlloc(NULL, size, MEM_COMMIT, PAGE_READWRITE);
}
void BigFree(void *address)
{
- #ifdef _SZ_ALLOC_DEBUG
- if (address != 0)
- fprintf(stderr, "\nFree_Big; count = %10d", --g_allocCountBig);
- #endif
+ PRINT_FREE("Free-Big", g_allocCountBig, address);
- if (address == 0)
+ if (!address)
return;
VirtualFree(address, 0, MEM_RELEASE);
}
@@ -127,10 +237,219 @@ void BigFree(void *address)
#endif
-static void *SzAlloc(void *p, size_t size) { UNUSED_VAR(p); return MyAlloc(size); }
-static void SzFree(void *p, void *address) { UNUSED_VAR(p); MyFree(address); }
-ISzAlloc g_Alloc = { SzAlloc, SzFree };
+static void *SzAlloc(ISzAllocPtr p, size_t size) { UNUSED_VAR(p); return MyAlloc(size); }
+static void SzFree(ISzAllocPtr p, void *address) { UNUSED_VAR(p); MyFree(address); }
+const ISzAlloc g_Alloc = { SzAlloc, SzFree };
+
+static void *SzMidAlloc(ISzAllocPtr p, size_t size) { UNUSED_VAR(p); return MidAlloc(size); }
+static void SzMidFree(ISzAllocPtr p, void *address) { UNUSED_VAR(p); MidFree(address); }
+const ISzAlloc g_MidAlloc = { SzMidAlloc, SzMidFree };
+
+static void *SzBigAlloc(ISzAllocPtr p, size_t size) { UNUSED_VAR(p); return BigAlloc(size); }
+static void SzBigFree(ISzAllocPtr p, void *address) { UNUSED_VAR(p); BigFree(address); }
+const ISzAlloc g_BigAlloc = { SzBigAlloc, SzBigFree };
+
+
+/*
+ uintptr_t : <stdint.h> C99 (optional)
+ : unsupported in VS6
+*/
+
+#ifdef _WIN32
+ typedef UINT_PTR UIntPtr;
+#else
+ /*
+ typedef uintptr_t UIntPtr;
+ */
+ typedef ptrdiff_t UIntPtr;
+#endif
+
+
+#define ADJUST_ALLOC_SIZE 0
+/*
+#define ADJUST_ALLOC_SIZE (sizeof(void *) - 1)
+*/
+/*
+ Use (ADJUST_ALLOC_SIZE = (sizeof(void *) - 1)), if
+ MyAlloc() can return address that is NOT multiple of sizeof(void *).
+*/
+
+
+/*
+#define MY_ALIGN_PTR_DOWN(p, align) ((void *)((char *)(p) - ((size_t)(UIntPtr)(p) & ((align) - 1))))
+*/
+#define MY_ALIGN_PTR_DOWN(p, align) ((void *)((((UIntPtr)(p)) & ~((UIntPtr)(align) - 1))))
+
+#define MY_ALIGN_PTR_UP_PLUS(p, align) MY_ALIGN_PTR_DOWN(((char *)(p) + (align) + ADJUST_ALLOC_SIZE), align)
+
+
+#if (_POSIX_C_SOURCE >= 200112L) && !defined(_WIN32)
+ #define USE_posix_memalign
+#endif
+
+/*
+ This posix_memalign() is for test purposes only.
+ We also need special Free() function instead of free(),
+ if this posix_memalign() is used.
+*/
+
+/*
+static int posix_memalign(void **ptr, size_t align, size_t size)
+{
+ size_t newSize = size + align;
+ void *p;
+ void *pAligned;
+ *ptr = NULL;
+ if (newSize < size)
+ return 12; // ENOMEM
+ p = MyAlloc(newSize);
+ if (!p)
+ return 12; // ENOMEM
+ pAligned = MY_ALIGN_PTR_UP_PLUS(p, align);
+ ((void **)pAligned)[-1] = p;
+ *ptr = pAligned;
+ return 0;
+}
+*/
+
+/*
+ ALLOC_ALIGN_SIZE >= sizeof(void *)
+ ALLOC_ALIGN_SIZE >= cache_line_size
+*/
+
+#define ALLOC_ALIGN_SIZE ((size_t)1 << 7)
+
+static void *SzAlignedAlloc(ISzAllocPtr pp, size_t size)
+{
+ #ifndef USE_posix_memalign
+
+ void *p;
+ void *pAligned;
+ size_t newSize;
+ UNUSED_VAR(pp);
+
+ /* also we can allocate additional dummy ALLOC_ALIGN_SIZE bytes after aligned
+ block to prevent cache line sharing with another allocated blocks */
+
+ newSize = size + ALLOC_ALIGN_SIZE * 1 + ADJUST_ALLOC_SIZE;
+ if (newSize < size)
+ return NULL;
+
+ p = MyAlloc(newSize);
+
+ if (!p)
+ return NULL;
+ pAligned = MY_ALIGN_PTR_UP_PLUS(p, ALLOC_ALIGN_SIZE);
+
+ Print(" size="); PrintHex(size, 8);
+ Print(" a_size="); PrintHex(newSize, 8);
+ Print(" ptr="); PrintAddr(p);
+ Print(" a_ptr="); PrintAddr(pAligned);
+ PrintLn();
+
+ ((void **)pAligned)[-1] = p;
+
+ return pAligned;
+
+ #else
+
+ void *p;
+ UNUSED_VAR(pp);
+ if (posix_memalign(&p, ALLOC_ALIGN_SIZE, size))
+ return NULL;
+
+ Print(" posix_memalign="); PrintAddr(p);
+ PrintLn();
+
+ return p;
+
+ #endif
+}
+
+
+static void SzAlignedFree(ISzAllocPtr pp, void *address)
+{
+ UNUSED_VAR(pp);
+ #ifndef USE_posix_memalign
+ if (address)
+ MyFree(((void **)address)[-1]);
+ #else
+ free(address);
+ #endif
+}
+
+
+const ISzAlloc g_AlignedAlloc = { SzAlignedAlloc, SzAlignedFree };
+
+
+
+#define MY_ALIGN_PTR_DOWN_1(p) MY_ALIGN_PTR_DOWN(p, sizeof(void *))
+
+/* we align ptr to support cases where CAlignOffsetAlloc::offset is not multiply of sizeof(void *) */
+#define REAL_BLOCK_PTR_VAR(p) ((void **)MY_ALIGN_PTR_DOWN_1(p))[-1]
+/*
+#define REAL_BLOCK_PTR_VAR(p) ((void **)(p))[-1]
+*/
+
+static void *AlignOffsetAlloc_Alloc(ISzAllocPtr pp, size_t size)
+{
+ CAlignOffsetAlloc *p = CONTAINER_FROM_VTBL(pp, CAlignOffsetAlloc, vt);
+ void *adr;
+ void *pAligned;
+ size_t newSize;
+ size_t extra;
+ size_t alignSize = (size_t)1 << p->numAlignBits;
+
+ if (alignSize < sizeof(void *))
+ alignSize = sizeof(void *);
+
+ if (p->offset >= alignSize)
+ return NULL;
+
+ /* also we can allocate additional dummy ALLOC_ALIGN_SIZE bytes after aligned
+ block to prevent cache line sharing with another allocated blocks */
+ extra = p->offset & (sizeof(void *) - 1);
+ newSize = size + alignSize + extra + ADJUST_ALLOC_SIZE;
+ if (newSize < size)
+ return NULL;
-static void *SzBigAlloc(void *p, size_t size) { UNUSED_VAR(p); return BigAlloc(size); }
-static void SzBigFree(void *p, void *address) { UNUSED_VAR(p); BigFree(address); }
-ISzAlloc g_BigAlloc = { SzBigAlloc, SzBigFree };
+ adr = ISzAlloc_Alloc(p->baseAlloc, newSize);
+
+ if (!adr)
+ return NULL;
+
+ pAligned = (char *)MY_ALIGN_PTR_DOWN((char *)adr +
+ alignSize - p->offset + extra + ADJUST_ALLOC_SIZE, alignSize) + p->offset;
+
+ PrintLn();
+ Print("- Aligned: ");
+ Print(" size="); PrintHex(size, 8);
+ Print(" a_size="); PrintHex(newSize, 8);
+ Print(" ptr="); PrintAddr(adr);
+ Print(" a_ptr="); PrintAddr(pAligned);
+ PrintLn();
+
+ REAL_BLOCK_PTR_VAR(pAligned) = adr;
+
+ return pAligned;
+}
+
+
+static void AlignOffsetAlloc_Free(ISzAllocPtr pp, void *address)
+{
+ if (address)
+ {
+ CAlignOffsetAlloc *p = CONTAINER_FROM_VTBL(pp, CAlignOffsetAlloc, vt);
+ PrintLn();
+ Print("- Aligned Free: ");
+ PrintLn();
+ ISzAlloc_Free(p->baseAlloc, REAL_BLOCK_PTR_VAR(address));
+ }
+}
+
+
+void AlignOffsetAlloc_CreateVTable(CAlignOffsetAlloc *p)
+{
+ p->vt.Alloc = AlignOffsetAlloc_Alloc;
+ p->vt.Free = AlignOffsetAlloc_Free;
+}
diff --git a/BaseTools/Source/C/LzmaCompress/Sdk/C/Alloc.h b/BaseTools/Source/C/LzmaCompress/Sdk/C/Alloc.h
index 73b282a071..3d796e5eee 100644
--- a/BaseTools/Source/C/LzmaCompress/Sdk/C/Alloc.h
+++ b/BaseTools/Source/C/LzmaCompress/Sdk/C/Alloc.h
@@ -1,5 +1,5 @@
/* Alloc.h -- Memory allocation functions
-2015-02-21 : Igor Pavlov : Public domain */
+2018-02-19 : Igor Pavlov : Public domain */
#ifndef __COMMON_ALLOC_H
#define __COMMON_ALLOC_H
@@ -29,8 +29,22 @@ void BigFree(void *address);
#endif
-extern ISzAlloc g_Alloc;
-extern ISzAlloc g_BigAlloc;
+extern const ISzAlloc g_Alloc;
+extern const ISzAlloc g_BigAlloc;
+extern const ISzAlloc g_MidAlloc;
+extern const ISzAlloc g_AlignedAlloc;
+
+
+typedef struct
+{
+ ISzAlloc vt;
+ ISzAllocPtr baseAlloc;
+ unsigned numAlignBits; /* ((1 << numAlignBits) >= sizeof(void *)) */
+ size_t offset; /* (offset == (k * sizeof(void *)) && offset < (1 << numAlignBits) */
+} CAlignOffsetAlloc;
+
+void AlignOffsetAlloc_CreateVTable(CAlignOffsetAlloc *p);
+
EXTERN_C_END
diff --git a/BaseTools/Source/C/LzmaCompress/Sdk/C/Bra86.c b/BaseTools/Source/C/LzmaCompress/Sdk/C/Bra86.c
index 8dd3ed48d9..a6463c63ba 100644
--- a/BaseTools/Source/C/LzmaCompress/Sdk/C/Bra86.c
+++ b/BaseTools/Source/C/LzmaCompress/Sdk/C/Bra86.c
@@ -1,5 +1,5 @@
/* Bra86.c -- Converter for x86 code (BCJ)
-2013-11-12 : Igor Pavlov : Public domain */
+2017-04-03 : Igor Pavlov : Public domain */
#include "Precomp.h"
@@ -37,7 +37,7 @@ SizeT x86_Convert(Byte *data, SizeT size, UInt32 ip, UInt32 *state, int encoding
else
{
mask >>= (unsigned)d;
- if (mask != 0 && (mask > 4 || mask == 3 || Test86MSByte(p[(mask >> 1) + 1])))
+ if (mask != 0 && (mask > 4 || mask == 3 || Test86MSByte(p[(size_t)(mask >> 1) + 1])))
{
mask = (mask >> 1) | 4;
pos++;
diff --git a/BaseTools/Source/C/LzmaCompress/Sdk/C/Compiler.h b/BaseTools/Source/C/LzmaCompress/Sdk/C/Compiler.h
index de8fab3749..c788648cd2 100644
--- a/BaseTools/Source/C/LzmaCompress/Sdk/C/Compiler.h
+++ b/BaseTools/Source/C/LzmaCompress/Sdk/C/Compiler.h
@@ -1,5 +1,5 @@
/* Compiler.h
-2015-08-02 : Igor Pavlov : Public domain */
+2017-04-03 : Igor Pavlov : Public domain */
#ifndef __7Z_COMPILER_H
#define __7Z_COMPILER_H
@@ -21,6 +21,7 @@
#pragma warning(disable : 4514) // unreferenced inline function has been removed
#pragma warning(disable : 4702) // unreachable code
#pragma warning(disable : 4710) // not inlined
+ #pragma warning(disable : 4714) // function marked as __forceinline not inlined
#pragma warning(disable : 4786) // identifier was truncated to '255' characters in the debug information
#endif
diff --git a/BaseTools/Source/C/LzmaCompress/Sdk/C/CpuArch.h b/BaseTools/Source/C/LzmaCompress/Sdk/C/CpuArch.h
index ef6083c3b8..7fb27282c7 100644
--- a/BaseTools/Source/C/LzmaCompress/Sdk/C/CpuArch.h
+++ b/BaseTools/Source/C/LzmaCompress/Sdk/C/CpuArch.h
@@ -1,5 +1,5 @@
/* CpuArch.h -- CPU specific code
-2016-06-09: Igor Pavlov : Public domain */
+2017-09-04 : Igor Pavlov : Public domain */
#ifndef __CPU_ARCH_H
#define __CPU_ARCH_H
@@ -16,48 +16,122 @@ If MY_CPU_LE and MY_CPU_BE are not defined, we don't know about ENDIANNESS of pl
MY_CPU_LE_UNALIGN means that CPU is LITTLE ENDIAN and CPU supports unaligned memory accesses.
*/
-#if defined(_M_X64) \
- || defined(_M_AMD64) \
- || defined(__x86_64__) \
- || defined(__AMD64__) \
- || defined(__amd64__)
+#if defined(_M_X64) \
+ || defined(_M_AMD64) \
+ || defined(__x86_64__) \
+ || defined(__AMD64__) \
+ || defined(__amd64__)
#define MY_CPU_AMD64
+ #ifdef __ILP32__
+ #define MY_CPU_NAME "x32"
+ #else
+ #define MY_CPU_NAME "x64"
+ #endif
+ #define MY_CPU_64BIT
#endif
-#if defined(MY_CPU_AMD64) \
- || defined(_M_IA64) \
- || defined(__AARCH64EL__) \
- || defined(__AARCH64EB__)
+
+#if defined(_M_IX86) \
+ || defined(__i386__)
+ #define MY_CPU_X86
+ #define MY_CPU_NAME "x86"
+ #define MY_CPU_32BIT
+#endif
+
+
+#if defined(_M_ARM64) \
+ || defined(__AARCH64EL__) \
+ || defined(__AARCH64EB__) \
+ || defined(__aarch64__)
+ #define MY_CPU_ARM64
+ #define MY_CPU_NAME "arm64"
#define MY_CPU_64BIT
#endif
-#if defined(_M_IX86) || defined(__i386__)
-#define MY_CPU_X86
+
+#if defined(_M_ARM) \
+ || defined(_M_ARM_NT) \
+ || defined(_M_ARMT) \
+ || defined(__arm__) \
+ || defined(__thumb__) \
+ || defined(__ARMEL__) \
+ || defined(__ARMEB__) \
+ || defined(__THUMBEL__) \
+ || defined(__THUMBEB__)
+ #define MY_CPU_ARM
+ #define MY_CPU_NAME "arm"
+ #define MY_CPU_32BIT
#endif
-#if defined(MY_CPU_X86) || defined(MY_CPU_AMD64)
-#define MY_CPU_X86_OR_AMD64
+
+#if defined(_M_IA64) \
+ || defined(__ia64__)
+ #define MY_CPU_IA64
+ #define MY_CPU_NAME "ia64"
+ #define MY_CPU_64BIT
#endif
-#if defined(MY_CPU_X86) \
- || defined(_M_ARM) \
- || defined(__ARMEL__) \
- || defined(__THUMBEL__) \
- || defined(__ARMEB__) \
- || defined(__THUMBEB__)
+
+#if defined(__mips64) \
+ || defined(__mips64__) \
+ || (defined(__mips) && (__mips == 64 || __mips == 4 || __mips == 3))
+ #define MY_CPU_NAME "mips64"
+ #define MY_CPU_64BIT
+#elif defined(__mips__)
+ #define MY_CPU_NAME "mips"
+ /* #define MY_CPU_32BIT */
+#endif
+
+
+#if defined(__ppc64__) \
+ || defined(__powerpc64__)
+ #ifdef __ILP32__
+ #define MY_CPU_NAME "ppc64-32"
+ #else
+ #define MY_CPU_NAME "ppc64"
+ #endif
+ #define MY_CPU_64BIT
+#elif defined(__ppc__) \
+ || defined(__powerpc__)
+ #define MY_CPU_NAME "ppc"
#define MY_CPU_32BIT
#endif
-#if defined(_WIN32) && defined(_M_ARM)
-#define MY_CPU_ARM_LE
+
+#if defined(__sparc64__)
+ #define MY_CPU_NAME "sparc64"
+ #define MY_CPU_64BIT
+#elif defined(__sparc__)
+ #define MY_CPU_NAME "sparc"
+ /* #define MY_CPU_32BIT */
#endif
-#if defined(_WIN32) && defined(_M_IA64)
-#define MY_CPU_IA64_LE
+
+#if defined(MY_CPU_X86) || defined(MY_CPU_AMD64)
+#define MY_CPU_X86_OR_AMD64
#endif
+
+#ifdef _WIN32
+
+ #ifdef MY_CPU_ARM
+ #define MY_CPU_ARM_LE
+ #endif
+
+ #ifdef MY_CPU_ARM64
+ #define MY_CPU_ARM64_LE
+ #endif
+
+ #ifdef _M_IA64
+ #define MY_CPU_IA64_LE
+ #endif
+
+#endif
+
+
#if defined(MY_CPU_X86_OR_AMD64) \
|| defined(MY_CPU_ARM_LE) \
+ || defined(MY_CPU_ARM64_LE) \
|| defined(MY_CPU_IA64_LE) \
|| defined(__LITTLE_ENDIAN__) \
|| defined(__ARMEL__) \
@@ -86,14 +160,37 @@ MY_CPU_LE_UNALIGN means that CPU is LITTLE ENDIAN and CPU supports unaligned mem
#define MY_CPU_BE
#endif
+
#if defined(MY_CPU_LE) && defined(MY_CPU_BE)
-Stop_Compiling_Bad_Endian
+ #error Stop_Compiling_Bad_Endian
+#endif
+
+
+#if defined(MY_CPU_32BIT) && defined(MY_CPU_64BIT)
+ #error Stop_Compiling_Bad_32_64_BIT
#endif
+#ifndef MY_CPU_NAME
+ #ifdef MY_CPU_LE
+ #define MY_CPU_NAME "LE"
+ #elif defined(MY_CPU_BE)
+ #define MY_CPU_NAME "BE"
+ #else
+ /*
+ #define MY_CPU_NAME ""
+ */
+ #endif
+#endif
+
+
+
+
+
#ifdef MY_CPU_LE
#if defined(MY_CPU_X86_OR_AMD64) \
- /* || defined(__AARCH64EL__) */
+ || defined(MY_CPU_ARM64) \
+ || defined(__ARM_FEATURE_UNALIGNED)
#define MY_CPU_LE_UNALIGN
#endif
#endif
@@ -139,6 +236,11 @@ Stop_Compiling_Bad_Endian
#endif
+#ifdef __has_builtin
+ #define MY__has_builtin(x) __has_builtin(x)
+#else
+ #define MY__has_builtin(x) 0
+#endif
#if defined(MY_CPU_LE_UNALIGN) && /* defined(_WIN64) && */ (_MSC_VER >= 1300)
@@ -146,15 +248,21 @@ Stop_Compiling_Bad_Endian
#include <stdlib.h>
+#pragma intrinsic(_byteswap_ushort)
#pragma intrinsic(_byteswap_ulong)
#pragma intrinsic(_byteswap_uint64)
+
+/* #define GetBe16(p) _byteswap_ushort(*(const UInt16 *)(const Byte *)(p)) */
#define GetBe32(p) _byteswap_ulong(*(const UInt32 *)(const Byte *)(p))
#define GetBe64(p) _byteswap_uint64(*(const UInt64 *)(const Byte *)(p))
#define SetBe32(p, v) (*(UInt32 *)(void *)(p)) = _byteswap_ulong(v)
-#elif defined(MY_CPU_LE_UNALIGN) && defined (__GNUC__) && (__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 3))
+#elif defined(MY_CPU_LE_UNALIGN) && ( \
+ (defined(__GNUC__) && (__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 3))) \
+ || (defined(__clang__) && MY__has_builtin(__builtin_bswap16)) )
+/* #define GetBe16(p) __builtin_bswap16(*(const UInt16 *)(const Byte *)(p)) */
#define GetBe32(p) __builtin_bswap32(*(const UInt32 *)(const Byte *)(p))
#define GetBe64(p) __builtin_bswap64(*(const UInt64 *)(const Byte *)(p))
@@ -179,10 +287,14 @@ Stop_Compiling_Bad_Endian
#endif
+#ifndef GetBe16
+
#define GetBe16(p) ( (UInt16) ( \
((UInt16)((const Byte *)(p))[0] << 8) | \
((const Byte *)(p))[1] ))
+#endif
+
#ifdef MY_CPU_X86_OR_AMD64
diff --git a/BaseTools/Source/C/LzmaCompress/Sdk/C/LzFind.c b/BaseTools/Source/C/LzmaCompress/Sdk/C/LzFind.c
index c335d363ce..6ea82a9b53 100644
--- a/BaseTools/Source/C/LzmaCompress/Sdk/C/LzFind.c
+++ b/BaseTools/Source/C/LzmaCompress/Sdk/C/LzFind.c
@@ -1,5 +1,5 @@
/* LzFind.c -- Match finder for LZ algorithms
-2015-10-15 : Igor Pavlov : Public domain */
+2017-06-10 : Igor Pavlov : Public domain */
#include "Precomp.h"
@@ -16,18 +16,18 @@
#define kStartMaxLen 3
-static void LzInWindow_Free(CMatchFinder *p, ISzAlloc *alloc)
+static void LzInWindow_Free(CMatchFinder *p, ISzAllocPtr alloc)
{
if (!p->directInput)
{
- alloc->Free(alloc, p->bufferBase);
+ ISzAlloc_Free(alloc, p->bufferBase);
p->bufferBase = NULL;
}
}
/* keepSizeBefore + keepSizeAfter + keepSizeReserv must be < 4G) */
-static int LzInWindow_Create(CMatchFinder *p, UInt32 keepSizeReserv, ISzAlloc *alloc)
+static int LzInWindow_Create(CMatchFinder *p, UInt32 keepSizeReserv, ISzAllocPtr alloc)
{
UInt32 blockSize = p->keepSizeBefore + p->keepSizeAfter + keepSizeReserv;
if (p->directInput)
@@ -39,7 +39,7 @@ static int LzInWindow_Create(CMatchFinder *p, UInt32 keepSizeReserv, ISzAlloc *a
{
LzInWindow_Free(p, alloc);
p->blockSize = blockSize;
- p->bufferBase = (Byte *)alloc->Alloc(alloc, (size_t)blockSize);
+ p->bufferBase = (Byte *)ISzAlloc_Alloc(alloc, (size_t)blockSize);
}
return (p->bufferBase != NULL);
}
@@ -81,7 +81,7 @@ static void MatchFinder_ReadBlock(CMatchFinder *p)
if (size == 0)
return;
- p->result = p->stream->Read(p->stream, dest, &size);
+ p->result = ISeqInStream_Read(p->stream, dest, &size);
if (p->result != SZ_OK)
return;
if (size == 0)
@@ -142,6 +142,7 @@ void MatchFinder_Construct(CMatchFinder *p)
p->bufferBase = NULL;
p->directInput = 0;
p->hash = NULL;
+ p->expectedDataSize = (UInt64)(Int64)-1;
MatchFinder_SetDefaultSettings(p);
for (i = 0; i < 256; i++)
@@ -149,34 +150,34 @@ void MatchFinder_Construct(CMatchFinder *p)
UInt32 r = i;
unsigned j;
for (j = 0; j < 8; j++)
- r = (r >> 1) ^ (kCrcPoly & ~((r & 1) - 1));
+ r = (r >> 1) ^ (kCrcPoly & ((UInt32)0 - (r & 1)));
p->crc[i] = r;
}
}
-static void MatchFinder_FreeThisClassMemory(CMatchFinder *p, ISzAlloc *alloc)
+static void MatchFinder_FreeThisClassMemory(CMatchFinder *p, ISzAllocPtr alloc)
{
- alloc->Free(alloc, p->hash);
+ ISzAlloc_Free(alloc, p->hash);
p->hash = NULL;
}
-void MatchFinder_Free(CMatchFinder *p, ISzAlloc *alloc)
+void MatchFinder_Free(CMatchFinder *p, ISzAllocPtr alloc)
{
MatchFinder_FreeThisClassMemory(p, alloc);
LzInWindow_Free(p, alloc);
}
-static CLzRef* AllocRefs(size_t num, ISzAlloc *alloc)
+static CLzRef* AllocRefs(size_t num, ISzAllocPtr alloc)
{
size_t sizeInBytes = (size_t)num * sizeof(CLzRef);
if (sizeInBytes / sizeof(CLzRef) != num)
return NULL;
- return (CLzRef *)alloc->Alloc(alloc, sizeInBytes);
+ return (CLzRef *)ISzAlloc_Alloc(alloc, sizeInBytes);
}
int MatchFinder_Create(CMatchFinder *p, UInt32 historySize,
UInt32 keepAddBufferBefore, UInt32 matchMaxLen, UInt32 keepAddBufferAfter,
- ISzAlloc *alloc)
+ ISzAllocPtr alloc)
{
UInt32 sizeReserv;
@@ -208,7 +209,11 @@ int MatchFinder_Create(CMatchFinder *p, UInt32 historySize,
hs = (1 << 16) - 1;
else
{
- hs = historySize - 1;
+ hs = historySize;
+ if (hs > p->expectedDataSize)
+ hs = (UInt32)p->expectedDataSize;
+ if (hs != 0)
+ hs--;
hs |= (hs >> 1);
hs |= (hs >> 2);
hs |= (hs >> 4);
@@ -292,17 +297,33 @@ static void MatchFinder_SetLimits(CMatchFinder *p)
p->posLimit = p->pos + limit;
}
-void MatchFinder_Init_2(CMatchFinder *p, int readData)
+
+void MatchFinder_Init_LowHash(CMatchFinder *p)
+{
+ size_t i;
+ CLzRef *items = p->hash;
+ size_t numItems = p->fixedHashSize;
+ for (i = 0; i < numItems; i++)
+ items[i] = kEmptyHashValue;
+}
+
+
+void MatchFinder_Init_HighHash(CMatchFinder *p)
+{
+ size_t i;
+ CLzRef *items = p->hash + p->fixedHashSize;
+ size_t numItems = (size_t)p->hashMask + 1;
+ for (i = 0; i < numItems; i++)
+ items[i] = kEmptyHashValue;
+}
+
+
+void MatchFinder_Init_3(CMatchFinder *p, int readData)
{
- UInt32 i;
- UInt32 *hash = p->hash;
- UInt32 num = p->hashSizeSum;
- for (i = 0; i < num; i++)
- hash[i] = kEmptyHashValue;
-
p->cyclicBufferPos = 0;
p->buffer = p->bufferBase;
- p->pos = p->streamPos = p->cyclicBufferSize;
+ p->pos =
+ p->streamPos = p->cyclicBufferSize;
p->result = SZ_OK;
p->streamEndWasReached = 0;
@@ -312,10 +333,14 @@ void MatchFinder_Init_2(CMatchFinder *p, int readData)
MatchFinder_SetLimits(p);
}
+
void MatchFinder_Init(CMatchFinder *p)
{
- MatchFinder_Init_2(p, True);
+ MatchFinder_Init_HighHash(p);
+ MatchFinder_Init_LowHash(p);
+ MatchFinder_Init_3(p, True);
}
+
static UInt32 MatchFinder_GetSubValue(CMatchFinder *p)
{
@@ -558,10 +583,10 @@ static UInt32 Bt3_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances)
d2 = pos - hash[h2];
- curMatch = hash[kFix3HashSize + hv];
+ curMatch = (hash + kFix3HashSize)[hv];
hash[h2] = pos;
- hash[kFix3HashSize + hv] = pos;
+ (hash + kFix3HashSize)[hv] = pos;
maxLen = 2;
offset = 0;
@@ -594,13 +619,13 @@ static UInt32 Bt4_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances)
pos = p->pos;
d2 = pos - hash[ h2];
- d3 = pos - hash[kFix3HashSize + h3];
+ d3 = pos - (hash + kFix3HashSize)[h3];
- curMatch = hash[kFix4HashSize + hv];
+ curMatch = (hash + kFix4HashSize)[hv];
hash[ h2] = pos;
- hash[kFix3HashSize + h3] = pos;
- hash[kFix4HashSize + hv] = pos;
+ (hash + kFix3HashSize)[h3] = pos;
+ (hash + kFix4HashSize)[hv] = pos;
maxLen = 0;
offset = 0;
@@ -615,7 +640,7 @@ static UInt32 Bt4_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances)
if (d2 != d3 && d3 < p->cyclicBufferSize && *(cur - d3) == *cur)
{
maxLen = 3;
- distances[offset + 1] = d3 - 1;
+ distances[(size_t)offset + 1] = d3 - 1;
offset += 2;
d2 = d3;
}
@@ -623,7 +648,7 @@ static UInt32 Bt4_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances)
if (offset != 0)
{
UPDATE_maxLen
- distances[offset - 2] = maxLen;
+ distances[(size_t)offset - 2] = maxLen;
if (maxLen == lenLimit)
{
SkipMatchesSpec(lenLimit, curMatch, MF_PARAMS(p));
@@ -650,15 +675,15 @@ static UInt32 Bt5_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances)
pos = p->pos;
d2 = pos - hash[ h2];
- d3 = pos - hash[kFix3HashSize + h3];
- d4 = pos - hash[kFix4HashSize + h4];
+ d3 = pos - (hash + kFix3HashSize)[h3];
+ d4 = pos - (hash + kFix4HashSize)[h4];
- curMatch = hash[kFix5HashSize + hv];
+ curMatch = (hash + kFix5HashSize)[hv];
hash[ h2] = pos;
- hash[kFix3HashSize + h3] = pos;
- hash[kFix4HashSize + h4] = pos;
- hash[kFix5HashSize + hv] = pos;
+ (hash + kFix3HashSize)[h3] = pos;
+ (hash + kFix4HashSize)[h4] = pos;
+ (hash + kFix5HashSize)[hv] = pos;
maxLen = 0;
offset = 0;
@@ -691,7 +716,7 @@ static UInt32 Bt5_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances)
&& *(cur - d4 + 3) == *(cur + 3))
{
maxLen = 4;
- distances[offset + 1] = d4 - 1;
+ distances[(size_t)offset + 1] = d4 - 1;
offset += 2;
d2 = d4;
}
@@ -699,7 +724,7 @@ static UInt32 Bt5_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances)
if (offset != 0)
{
UPDATE_maxLen
- distances[offset - 2] = maxLen;
+ distances[(size_t)offset - 2] = maxLen;
if (maxLen == lenLimit)
{
SkipMatchesSpec(lenLimit, curMatch, MF_PARAMS(p));
@@ -726,13 +751,13 @@ static UInt32 Hc4_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances)
pos = p->pos;
d2 = pos - hash[ h2];
- d3 = pos - hash[kFix3HashSize + h3];
+ d3 = pos - (hash + kFix3HashSize)[h3];
- curMatch = hash[kFix4HashSize + hv];
+ curMatch = (hash + kFix4HashSize)[hv];
hash[ h2] = pos;
- hash[kFix3HashSize + h3] = pos;
- hash[kFix4HashSize + hv] = pos;
+ (hash + kFix3HashSize)[h3] = pos;
+ (hash + kFix4HashSize)[hv] = pos;
maxLen = 0;
offset = 0;
@@ -747,7 +772,7 @@ static UInt32 Hc4_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances)
if (d2 != d3 && d3 < p->cyclicBufferSize && *(cur - d3) == *cur)
{
maxLen = 3;
- distances[offset + 1] = d3 - 1;
+ distances[(size_t)offset + 1] = d3 - 1;
offset += 2;
d2 = d3;
}
@@ -755,7 +780,7 @@ static UInt32 Hc4_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances)
if (offset != 0)
{
UPDATE_maxLen
- distances[offset - 2] = maxLen;
+ distances[(size_t)offset - 2] = maxLen;
if (maxLen == lenLimit)
{
p->son[p->cyclicBufferPos] = curMatch;
@@ -784,15 +809,15 @@ static UInt32 Hc5_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances)
pos = p->pos;
d2 = pos - hash[ h2];
- d3 = pos - hash[kFix3HashSize + h3];
- d4 = pos - hash[kFix4HashSize + h4];
+ d3 = pos - (hash + kFix3HashSize)[h3];
+ d4 = pos - (hash + kFix4HashSize)[h4];
- curMatch = hash[kFix5HashSize + hv];
+ curMatch = (hash + kFix5HashSize)[hv];
hash[ h2] = pos;
- hash[kFix3HashSize + h3] = pos;
- hash[kFix4HashSize + h4] = pos;
- hash[kFix5HashSize + hv] = pos;
+ (hash + kFix3HashSize)[h3] = pos;
+ (hash + kFix4HashSize)[h4] = pos;
+ (hash + kFix5HashSize)[hv] = pos;
maxLen = 0;
offset = 0;
@@ -825,7 +850,7 @@ static UInt32 Hc5_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances)
&& *(cur - d4 + 3) == *(cur + 3))
{
maxLen = 4;
- distances[offset + 1] = d4 - 1;
+ distances[(size_t)offset + 1] = d4 - 1;
offset += 2;
d2 = d4;
}
@@ -833,7 +858,7 @@ static UInt32 Hc5_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances)
if (offset != 0)
{
UPDATE_maxLen
- distances[offset - 2] = maxLen;
+ distances[(size_t)offset - 2] = maxLen;
if (maxLen == lenLimit)
{
p->son[p->cyclicBufferPos] = curMatch;
@@ -897,9 +922,9 @@ static void Bt3_MatchFinder_Skip(CMatchFinder *p, UInt32 num)
SKIP_HEADER(3)
HASH3_CALC;
hash = p->hash;
- curMatch = hash[kFix3HashSize + hv];
+ curMatch = (hash + kFix3HashSize)[hv];
hash[h2] =
- hash[kFix3HashSize + hv] = p->pos;
+ (hash + kFix3HashSize)[hv] = p->pos;
SKIP_FOOTER
}
while (--num != 0);
@@ -914,10 +939,10 @@ static void Bt4_MatchFinder_Skip(CMatchFinder *p, UInt32 num)
SKIP_HEADER(4)
HASH4_CALC;
hash = p->hash;
- curMatch = hash[kFix4HashSize + hv];
+ curMatch = (hash + kFix4HashSize)[hv];
hash[ h2] =
- hash[kFix3HashSize + h3] =
- hash[kFix4HashSize + hv] = p->pos;
+ (hash + kFix3HashSize)[h3] =
+ (hash + kFix4HashSize)[hv] = p->pos;
SKIP_FOOTER
}
while (--num != 0);
@@ -933,11 +958,11 @@ static void Bt5_MatchFinder_Skip(CMatchFinder *p, UInt32 num)
SKIP_HEADER(5)
HASH5_CALC;
hash = p->hash;
- curMatch = hash[kFix5HashSize + hv];
+ curMatch = (hash + kFix5HashSize)[hv];
hash[ h2] =
- hash[kFix3HashSize + h3] =
- hash[kFix4HashSize + h4] =
- hash[kFix5HashSize + hv] = p->pos;
+ (hash + kFix3HashSize)[h3] =
+ (hash + kFix4HashSize)[h4] =
+ (hash + kFix5HashSize)[hv] = p->pos;
SKIP_FOOTER
}
while (--num != 0);
@@ -953,10 +978,10 @@ static void Hc4_MatchFinder_Skip(CMatchFinder *p, UInt32 num)
SKIP_HEADER(4)
HASH4_CALC;
hash = p->hash;
- curMatch = hash[kFix4HashSize + hv];
+ curMatch = (hash + kFix4HashSize)[hv];
hash[ h2] =
- hash[kFix3HashSize + h3] =
- hash[kFix4HashSize + hv] = p->pos;
+ (hash + kFix3HashSize)[h3] =
+ (hash + kFix4HashSize)[hv] = p->pos;
p->son[p->cyclicBufferPos] = curMatch;
MOVE_POS
}
@@ -973,11 +998,11 @@ static void Hc5_MatchFinder_Skip(CMatchFinder *p, UInt32 num)
SKIP_HEADER(5)
HASH5_CALC;
hash = p->hash;
- curMatch = p->hash[kFix5HashSize + hv];
+ curMatch = hash + kFix5HashSize)[hv];
hash[ h2] =
- hash[kFix3HashSize + h3] =
- hash[kFix4HashSize + h4] =
- hash[kFix5HashSize + hv] = p->pos;
+ (hash + kFix3HashSize)[h3] =
+ (hash + kFix4HashSize)[h4] =
+ (hash + kFix5HashSize)[hv] = p->pos;
p->son[p->cyclicBufferPos] = curMatch;
MOVE_POS
}
diff --git a/BaseTools/Source/C/LzmaCompress/Sdk/C/LzFind.h b/BaseTools/Source/C/LzmaCompress/Sdk/C/LzFind.h
index 2ff6673771..c77added7b 100644
--- a/BaseTools/Source/C/LzmaCompress/Sdk/C/LzFind.h
+++ b/BaseTools/Source/C/LzmaCompress/Sdk/C/LzFind.h
@@ -1,5 +1,5 @@
/* LzFind.h -- Match finder for LZ algorithms
-2015-10-15 : Igor Pavlov : Public domain */
+2017-06-10 : Igor Pavlov : Public domain */
#ifndef __LZ_FIND_H
#define __LZ_FIND_H
@@ -47,6 +47,8 @@ typedef struct _CMatchFinder
SRes result;
UInt32 crc[256];
size_t numRefs;
+
+ UInt64 expectedDataSize;
} CMatchFinder;
#define Inline_MatchFinder_GetPointerToCurrentPos(p) ((p)->buffer)
@@ -71,8 +73,8 @@ void MatchFinder_Construct(CMatchFinder *p);
*/
int MatchFinder_Create(CMatchFinder *p, UInt32 historySize,
UInt32 keepAddBufferBefore, UInt32 matchMaxLen, UInt32 keepAddBufferAfter,
- ISzAlloc *alloc);
-void MatchFinder_Free(CMatchFinder *p, ISzAlloc *alloc);
+ ISzAllocPtr alloc);
+void MatchFinder_Free(CMatchFinder *p, ISzAllocPtr alloc);
void MatchFinder_Normalize3(UInt32 subValue, CLzRef *items, size_t numItems);
void MatchFinder_ReduceOffsets(CMatchFinder *p, UInt32 subValue);
@@ -103,7 +105,9 @@ typedef struct _IMatchFinder
void MatchFinder_CreateVTable(CMatchFinder *p, IMatchFinder *vTable);
-void MatchFinder_Init_2(CMatchFinder *p, int readData);
+void MatchFinder_Init_LowHash(CMatchFinder *p);
+void MatchFinder_Init_HighHash(CMatchFinder *p);
+void MatchFinder_Init_3(CMatchFinder *p, int readData);
void MatchFinder_Init(CMatchFinder *p);
UInt32 Bt3Zip_MatchFinder_GetMatches(CMatchFinder *p, UInt32 *distances);
diff --git a/BaseTools/Source/C/LzmaCompress/Sdk/C/LzFindMt.c b/BaseTools/Source/C/LzmaCompress/Sdk/C/LzFindMt.c
index cb61e0953a..2563824fcd 100644
--- a/BaseTools/Source/C/LzmaCompress/Sdk/C/LzFindMt.c
+++ b/BaseTools/Source/C/LzmaCompress/Sdk/C/LzFindMt.c
@@ -1,5 +1,5 @@
/* LzFindMt.c -- multithreaded Match finder for LZ algorithms
-2015-10-15 : Igor Pavlov : Public domain */
+2017-06-10 : Igor Pavlov : Public domain */
#include "Precomp.h"
@@ -33,6 +33,8 @@ static void MtSync_GetNextBlock(CMtSync *p)
Event_Set(&p->canStart);
Event_Wait(&p->wasStarted);
+
+ // if (mt) MatchFinder_Init_LowHash(mt->MatchFinder);
}
else
{
@@ -155,6 +157,9 @@ static void HashThreadFunc(CMatchFinderMt *mt)
UInt32 numProcessedBlocks = 0;
Event_Wait(&p->canStart);
Event_Set(&p->wasStarted);
+
+ MatchFinder_Init_HighHash(mt->MatchFinder);
+
for (;;)
{
if (p->exit)
@@ -205,7 +210,7 @@ static void HashThreadFunc(CMatchFinderMt *mt)
if (num > kMtHashBlockSize - 2)
num = kMtHashBlockSize - 2;
mt->GetHeadsFunc(mf->buffer, mf->pos, mf->hash + mf->fixedHashSize, mf->hashMask, heads + 2, num, mf->crc);
- heads[0] += num;
+ heads[0] = 2 + num;
}
mf->pos += num;
mf->buffer += num;
@@ -443,13 +448,13 @@ void MatchFinderMt_Construct(CMatchFinderMt *p)
MtSync_Construct(&p->btSync);
}
-static void MatchFinderMt_FreeMem(CMatchFinderMt *p, ISzAlloc *alloc)
+static void MatchFinderMt_FreeMem(CMatchFinderMt *p, ISzAllocPtr alloc)
{
- alloc->Free(alloc, p->hashBuf);
+ ISzAlloc_Free(alloc, p->hashBuf);
p->hashBuf = NULL;
}
-void MatchFinderMt_Destruct(CMatchFinderMt *p, ISzAlloc *alloc)
+void MatchFinderMt_Destruct(CMatchFinderMt *p, ISzAllocPtr alloc)
{
MtSync_Destruct(&p->hashSync);
MtSync_Destruct(&p->btSync);
@@ -472,7 +477,7 @@ static THREAD_FUNC_RET_TYPE THREAD_FUNC_CALL_TYPE BtThreadFunc2(void *p)
}
SRes MatchFinderMt_Create(CMatchFinderMt *p, UInt32 historySize, UInt32 keepAddBufferBefore,
- UInt32 matchMaxLen, UInt32 keepAddBufferAfter, ISzAlloc *alloc)
+ UInt32 matchMaxLen, UInt32 keepAddBufferAfter, ISzAllocPtr alloc)
{
CMatchFinder *mf = p->MatchFinder;
p->historySize = historySize;
@@ -480,7 +485,7 @@ SRes MatchFinderMt_Create(CMatchFinderMt *p, UInt32 historySize, UInt32 keepAddB
return SZ_ERROR_PARAM;
if (!p->hashBuf)
{
- p->hashBuf = (UInt32 *)alloc->Alloc(alloc, (kHashBufferSize + kBtBufferSize) * sizeof(UInt32));
+ p->hashBuf = (UInt32 *)ISzAlloc_Alloc(alloc, (kHashBufferSize + kBtBufferSize) * sizeof(UInt32));
if (!p->hashBuf)
return SZ_ERROR_MEM;
p->btBuf = p->hashBuf + kHashBufferSize;
@@ -496,14 +501,18 @@ SRes MatchFinderMt_Create(CMatchFinderMt *p, UInt32 historySize, UInt32 keepAddB
}
/* Call it after ReleaseStream / SetStream */
-void MatchFinderMt_Init(CMatchFinderMt *p)
+static void MatchFinderMt_Init(CMatchFinderMt *p)
{
CMatchFinder *mf = p->MatchFinder;
- p->btBufPos = p->btBufPosLimit = 0;
- p->hashBufPos = p->hashBufPosLimit = 0;
+
+ p->btBufPos =
+ p->btBufPosLimit = 0;
+ p->hashBufPos =
+ p->hashBufPosLimit = 0;
/* Init without data reading. We don't want to read data in this thread */
- MatchFinder_Init_2(mf, False);
+ MatchFinder_Init_3(mf, False);
+ MatchFinder_Init_LowHash(mf);
p->pointerToCurPos = Inline_MatchFinder_GetPointerToCurrentPos(mf);
p->btNumAvailBytes = 0;
@@ -591,10 +600,10 @@ static UInt32 * MixMatches3(CMatchFinderMt *p, UInt32 matchMinPos, UInt32 *dista
MT_HASH3_CALC
curMatch2 = hash[ h2];
- curMatch3 = hash[kFix3HashSize + h3];
+ curMatch3 = (hash + kFix3HashSize)[h3];
hash[ h2] = lzPos;
- hash[kFix3HashSize + h3] = lzPos;
+ (hash + kFix3HashSize)[h3] = lzPos;
if (curMatch2 >= matchMinPos && cur[(ptrdiff_t)curMatch2 - lzPos] == cur[0])
{
@@ -627,12 +636,12 @@ static UInt32 *MixMatches4(CMatchFinderMt *p, UInt32 matchMinPos, UInt32 *distan
MT_HASH4_CALC
curMatch2 = hash[ h2];
- curMatch3 = hash[kFix3HashSize + h3];
- curMatch4 = hash[kFix4HashSize + h4];
+ curMatch3 = (hash + kFix3HashSize)[h3];
+ curMatch4 = (hash + kFix4HashSize)[h4];
hash[ h2] = lzPos;
- hash[kFix3HashSize + h3] = lzPos;
- hash[kFix4HashSize + h4] = lzPos;
+ (hash + kFix3HashSize)[h3] = lzPos;
+ (hash + kFix4HashSize)[h4] = lzPos;
if (curMatch2 >= matchMinPos && cur[(ptrdiff_t)curMatch2 - lzPos] == cur[0])
{
@@ -684,8 +693,12 @@ static UInt32 MatchFinderMt2_GetMatches(CMatchFinderMt *p, UInt32 *distances)
UInt32 i;
for (i = 0; i < len; i += 2)
{
- *distances++ = *btBuf++;
- *distances++ = *btBuf++;
+ UInt32 v0 = btBuf[0];
+ UInt32 v1 = btBuf[1];
+ btBuf += 2;
+ distances[0] = v0;
+ distances[1] = v1;
+ distances += 2;
}
}
INCREASE_LZ_POS
@@ -712,8 +725,12 @@ static UInt32 MatchFinderMt_GetMatches(CMatchFinderMt *p, UInt32 *distances)
distances2 = p->MixMatchesFunc(p, p->lzPos - btBuf[1], distances);
do
{
- *distances2++ = *btBuf++;
- *distances2++ = *btBuf++;
+ UInt32 v0 = btBuf[0];
+ UInt32 v1 = btBuf[1];
+ btBuf += 2;
+ distances2[0] = v0;
+ distances2[1] = v1;
+ distances2 += 2;
}
while ((len -= 2) != 0);
len = (UInt32)(distances2 - (distances));
@@ -746,7 +763,7 @@ static void MatchFinderMt3_Skip(CMatchFinderMt *p, UInt32 num)
SKIP_HEADER_MT(3)
UInt32 h2, h3;
MT_HASH3_CALC
- hash[kFix3HashSize + h3] =
+ (hash + kFix3HashSize)[h3] =
hash[ h2] =
p->lzPos;
SKIP_FOOTER_MT
@@ -758,8 +775,8 @@ static void MatchFinderMt4_Skip(CMatchFinderMt *p, UInt32 num)
SKIP_HEADER_MT(4)
UInt32 h2, h3, h4;
MT_HASH4_CALC
- hash[kFix4HashSize + h4] =
- hash[kFix3HashSize + h3] =
+ (hash + kFix4HashSize)[h4] =
+ (hash + kFix3HashSize)[h3] =
hash[ h2] =
p->lzPos;
SKIP_FOOTER_MT
@@ -777,7 +794,7 @@ void MatchFinderMt_CreateVTable(CMatchFinderMt *p, IMatchFinder *vTable)
{
case 2:
p->GetHeadsFunc = GetHeads2;
- p->MixMatchesFunc = (Mf_Mix_Matches)0;
+ p->MixMatchesFunc = (Mf_Mix_Matches)NULL;
vTable->Skip = (Mf_Skip_Func)MatchFinderMt0_Skip;
vTable->GetMatches = (Mf_GetMatches_Func)MatchFinderMt2_GetMatches;
break;
diff --git a/BaseTools/Source/C/LzmaCompress/Sdk/C/LzFindMt.h b/BaseTools/Source/C/LzmaCompress/Sdk/C/LzFindMt.h
index 46b6924ad7..3d86c788f3 100644
--- a/BaseTools/Source/C/LzmaCompress/Sdk/C/LzFindMt.h
+++ b/BaseTools/Source/C/LzmaCompress/Sdk/C/LzFindMt.h
@@ -1,5 +1,5 @@
/* LzFindMt.h -- multithreaded Match finder for LZ algorithms
-2015-05-03 : Igor Pavlov : Public domain */
+2017-04-03 : Igor Pavlov : Public domain */
#ifndef __LZ_FIND_MT_H
#define __LZ_FIND_MT_H
@@ -90,9 +90,9 @@ typedef struct _CMatchFinderMt
} CMatchFinderMt;
void MatchFinderMt_Construct(CMatchFinderMt *p);
-void MatchFinderMt_Destruct(CMatchFinderMt *p, ISzAlloc *alloc);
+void MatchFinderMt_Destruct(CMatchFinderMt *p, ISzAllocPtr alloc);
SRes MatchFinderMt_Create(CMatchFinderMt *p, UInt32 historySize, UInt32 keepAddBufferBefore,
- UInt32 matchMaxLen, UInt32 keepAddBufferAfter, ISzAlloc *alloc);
+ UInt32 matchMaxLen, UInt32 keepAddBufferAfter, ISzAllocPtr alloc);
void MatchFinderMt_CreateVTable(CMatchFinderMt *p, IMatchFinder *vTable);
void MatchFinderMt_ReleaseStream(CMatchFinderMt *p);
diff --git a/BaseTools/Source/C/LzmaCompress/Sdk/C/LzmaDec.c b/BaseTools/Source/C/LzmaCompress/Sdk/C/LzmaDec.c
index 64f1164f3d..962b94bb63 100644
--- a/BaseTools/Source/C/LzmaCompress/Sdk/C/LzmaDec.c
+++ b/BaseTools/Source/C/LzmaCompress/Sdk/C/LzmaDec.c
@@ -1,8 +1,9 @@
/* LzmaDec.c -- LZMA Decoder
-2016-05-16 : Igor Pavlov : Public domain */
+2018-02-28 : Igor Pavlov : Public domain */
#include "Precomp.h"
+/* #include "CpuArch.h" */
#include "LzmaDec.h"
#include <string.h>
@@ -24,9 +25,16 @@
#define GET_BIT2(p, i, A0, A1) IF_BIT_0(p) \
{ UPDATE_0(p); i = (i + i); A0; } else \
{ UPDATE_1(p); i = (i + i) + 1; A1; }
-#define GET_BIT(p, i) GET_BIT2(p, i, ; , ;)
-#define TREE_GET_BIT(probs, i) { GET_BIT((probs + i), i); }
+#define TREE_GET_BIT(probs, i) { GET_BIT2(probs + i, i, ;, ;); }
+
+#define REV_BIT(p, i, A0, A1) IF_BIT_0(p + i) \
+ { UPDATE_0(p + i); A0; } else \
+ { UPDATE_1(p + i); A1; }
+#define REV_BIT_VAR( p, i, m) REV_BIT(p, i, i += m; m += m, m += m; i += m; )
+#define REV_BIT_CONST(p, i, m) REV_BIT(p, i, i += m; , i += m * 2; )
+#define REV_BIT_LAST( p, i, m) REV_BIT(p, i, i -= m , ; )
+
#define TREE_DECODE(probs, limit, i) \
{ i = 1; do { TREE_GET_BIT(probs, i); } while (i < limit); i -= limit; }
@@ -46,12 +54,15 @@
i -= 0x40; }
#endif
-#define NORMAL_LITER_DEC GET_BIT(prob + symbol, symbol)
+#define NORMAL_LITER_DEC TREE_GET_BIT(prob, symbol)
#define MATCHED_LITER_DEC \
- matchByte <<= 1; \
- bit = (matchByte & offs); \
- probLit = prob + offs + bit + symbol; \
- GET_BIT2(probLit, symbol, offs &= ~bit, offs &= bit)
+ matchByte += matchByte; \
+ bit = offs; \
+ offs &= matchByte; \
+ probLit = prob + (offs + bit + symbol); \
+ GET_BIT2(probLit, symbol, offs ^= bit; , ;)
+
+
#define NORMALIZE_CHECK if (range < kTopValue) { if (buf >= bufLimit) return DUMMY_ERROR; range <<= 8; code = (code << 8) | (*buf++); }
@@ -66,25 +77,28 @@
{ i = 1; do { GET_BIT_CHECK(probs + i, i) } while (i < limit); i -= limit; }
+#define REV_BIT_CHECK(p, i, m) IF_BIT_0_CHECK(p + i) \
+ { UPDATE_0_CHECK; i += m; m += m; } else \
+ { UPDATE_1_CHECK; m += m; i += m; }
+
+
#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 LenLow 0
+#define LenHigh (LenLow + 2 * (kNumPosStatesMax << kLenNumLowBits))
#define kNumLenProbs (LenHigh + kLenNumHighSymbols)
+#define LenChoice LenLow
+#define LenChoice2 (LenLow + (1 << kLenNumLowBits))
#define kNumStates 12
+#define kNumStates2 16
#define kNumLitStates 7
#define kStartPosModelIndex 4
@@ -98,54 +112,117 @@
#define kAlignTableSize (1 << kNumAlignBits)
#define kMatchMinLen 2
-#define kMatchSpecLenStart (kMatchMinLen + kLenNumLowSymbols + kLenNumMidSymbols + kLenNumHighSymbols)
+#define kMatchSpecLenStart (kMatchMinLen + kLenNumLowSymbols * 2 + kLenNumHighSymbols)
-#define IsMatch 0
-#define IsRep (IsMatch + (kNumStates << kNumPosBitsMax))
+/* External ASM code needs same CLzmaProb array layout. So don't change it. */
+
+/* (probs_1664) is faster and better for code size at some platforms */
+/*
+#ifdef MY_CPU_X86_OR_AMD64
+*/
+#define kStartOffset 1664
+#define GET_PROBS p->probs_1664
+/*
+#define GET_PROBS p->probs + kStartOffset
+#else
+#define kStartOffset 0
+#define GET_PROBS p->probs
+#endif
+*/
+
+#define SpecPos (-kStartOffset)
+#define IsRep0Long (SpecPos + kNumFullDistances)
+#define RepLenCoder (IsRep0Long + (kNumStates2 << kNumPosBitsMax))
+#define LenCoder (RepLenCoder + kNumLenProbs)
+#define IsMatch (LenCoder + kNumLenProbs)
+#define Align (IsMatch + (kNumStates2 << kNumPosBitsMax))
+#define IsRep (Align + kAlignTableSize)
#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)
-
-#define LZMA_BASE_SIZE 1846
-#define LZMA_LIT_SIZE 0x300
+#define PosSlot (IsRepG2 + kNumStates)
+#define Literal (PosSlot + (kNumLenToPosStates << kNumPosSlotBits))
+#define NUM_BASE_PROBS (Literal + kStartOffset)
-#if Literal != LZMA_BASE_SIZE
-StopCompilingDueBUG
+#if Align != 0 && kStartOffset != 0
+ #error Stop_Compiling_Bad_LZMA_kAlign
#endif
-#define LzmaProps_GetNumProbs(p) (Literal + ((UInt32)LZMA_LIT_SIZE << ((p)->lc + (p)->lp)))
+#if NUM_BASE_PROBS != 1984
+ #error Stop_Compiling_Bad_LZMA_PROBS
+#endif
+
+
+#define LZMA_LIT_SIZE 0x300
+
+#define LzmaProps_GetNumProbs(p) (NUM_BASE_PROBS + ((UInt32)LZMA_LIT_SIZE << ((p)->lc + (p)->lp)))
+
+
+#define CALC_POS_STATE(processedPos, pbMask) (((processedPos) & (pbMask)) << 4)
+#define COMBINED_PS_STATE (posState + state)
+#define GET_LEN_STATE (posState)
#define LZMA_DIC_MIN (1 << 12)
-/* First LZMA-symbol is always decoded.
-And it decodes new LZMA-symbols while (buf < bufLimit), but "buf" is without last normalization
+/*
+p->remainLen : shows status of LZMA decoder:
+ < kMatchSpecLenStart : normal remain
+ = kMatchSpecLenStart : finished
+ = kMatchSpecLenStart + 1 : need init range coder
+ = kMatchSpecLenStart + 2 : need init range coder and state
+*/
+
+/* ---------- LZMA_DECODE_REAL ---------- */
+/*
+LzmaDec_DecodeReal_3() can be implemented in external ASM file.
+3 - is the code compatibility version of that function for check at link time.
+*/
+
+#define LZMA_DECODE_REAL LzmaDec_DecodeReal_3
+
+/*
+LZMA_DECODE_REAL()
+In:
+ RangeCoder is normalized
+ if (p->dicPos == limit)
+ {
+ LzmaDec_TryDummy() was called before to exclude LITERAL and MATCH-REP cases.
+ So first symbol can be only MATCH-NON-REP. And if that MATCH-NON-REP symbol
+ is not END_OF_PAYALOAD_MARKER, then function returns error code.
+ }
+
+Processing:
+ first LZMA symbol will be decoded in any case
+ All checks for limits are at the end of main loop,
+ It will decode new LZMA-symbols while (p->buf < bufLimit && dicPos < limit),
+ RangeCoder is still without last normalization when (p->buf < bufLimit) is being checked.
+
Out:
+ RangeCoder is normalized
Result:
SZ_OK - OK
SZ_ERROR_DATA - Error
p->remainLen:
< kMatchSpecLenStart : normal remain
= kMatchSpecLenStart : finished
- = kMatchSpecLenStart + 1 : Flush marker (unused now)
- = kMatchSpecLenStart + 2 : State Init Marker (unused now)
*/
-static int MY_FAST_CALL LzmaDec_DecodeReal(CLzmaDec *p, SizeT limit, const Byte *bufLimit)
-{
- CLzmaProb *probs = p->probs;
- unsigned state = p->state;
+#ifdef _LZMA_DEC_OPT
+
+int MY_FAST_CALL LZMA_DECODE_REAL(CLzmaDec *p, SizeT limit, const Byte *bufLimit);
+
+#else
+
+static
+int MY_FAST_CALL LZMA_DECODE_REAL(CLzmaDec *p, SizeT limit, const Byte *bufLimit)
+{
+ CLzmaProb *probs = GET_PROBS;
+ unsigned state = (unsigned)p->state;
UInt32 rep0 = p->reps[0], rep1 = p->reps[1], rep2 = p->reps[2], rep3 = p->reps[3];
unsigned pbMask = ((unsigned)1 << (p->prop.pb)) - 1;
- unsigned lpMask = ((unsigned)1 << (p->prop.lp)) - 1;
unsigned lc = p->prop.lc;
+ unsigned lpMask = ((unsigned)0x100 << p->prop.lp) - ((unsigned)0x100 >> lc);
Byte *dic = p->dic;
SizeT dicBufSize = p->dicBufSize;
@@ -164,17 +241,16 @@ static int MY_FAST_CALL LzmaDec_DecodeReal(CLzmaDec *p, SizeT limit, const Byte
CLzmaProb *prob;
UInt32 bound;
unsigned ttt;
- unsigned posState = processedPos & pbMask;
+ unsigned posState = CALC_POS_STATE(processedPos, pbMask);
- prob = probs + IsMatch + (state << kNumPosBitsMax) + posState;
+ prob = probs + IsMatch + COMBINED_PS_STATE;
IF_BIT_0(prob)
{
unsigned symbol;
UPDATE_0(prob);
prob = probs + Literal;
if (processedPos != 0 || checkDicSize != 0)
- prob += ((UInt32)LZMA_LIT_SIZE * (((processedPos & lpMask) << lc) +
- (dic[(dicPos == 0 ? dicBufSize : dicPos) - 1] >> (8 - lc))));
+ prob += (UInt32)3 * ((((processedPos << 8) + dic[(dicPos == 0 ? dicBufSize : dicPos) - 1]) & lpMask) << lc);
processedPos++;
if (state < kNumLitStates)
@@ -240,13 +316,16 @@ static int MY_FAST_CALL LzmaDec_DecodeReal(CLzmaDec *p, SizeT limit, const Byte
else
{
UPDATE_1(prob);
+ /*
+ // that case was checked before with kBadRepCode
if (checkDicSize == 0 && processedPos == 0)
return SZ_ERROR_DATA;
+ */
prob = probs + IsRepG0 + state;
IF_BIT_0(prob)
{
UPDATE_0(prob);
- prob = probs + IsRep0Long + (state << kNumPosBitsMax) + posState;
+ prob = probs + IsRep0Long + COMBINED_PS_STATE;
IF_BIT_0(prob)
{
UPDATE_0(prob);
@@ -299,7 +378,7 @@ static int MY_FAST_CALL LzmaDec_DecodeReal(CLzmaDec *p, SizeT limit, const Byte
IF_BIT_0(probLen)
{
UPDATE_0(probLen);
- probLen = prob + LenLow + (posState << kLenNumLowBits);
+ probLen = prob + LenLow + GET_LEN_STATE;
offset = 0;
lim = (1 << kLenNumLowBits);
}
@@ -310,15 +389,15 @@ static int MY_FAST_CALL LzmaDec_DecodeReal(CLzmaDec *p, SizeT limit, const Byte
IF_BIT_0(probLen)
{
UPDATE_0(probLen);
- probLen = prob + LenMid + (posState << kLenNumMidBits);
+ probLen = prob + LenLow + GET_LEN_STATE + (1 << kLenNumLowBits);
offset = kLenNumLowSymbols;
- lim = (1 << kLenNumMidBits);
+ lim = (1 << kLenNumLowBits);
}
else
{
UPDATE_1(probLen);
probLen = prob + LenHigh;
- offset = kLenNumLowSymbols + kLenNumMidSymbols;
+ offset = kLenNumLowSymbols * 2;
lim = (1 << kLenNumHighBits);
}
}
@@ -331,7 +410,7 @@ static int MY_FAST_CALL LzmaDec_DecodeReal(CLzmaDec *p, SizeT limit, const Byte
IF_BIT_0(probLen)
{
UPDATE_0(probLen);
- probLen = prob + LenLow + (posState << kLenNumLowBits);
+ probLen = prob + LenLow + GET_LEN_STATE;
len = 1;
TREE_GET_BIT(probLen, len);
TREE_GET_BIT(probLen, len);
@@ -345,7 +424,7 @@ static int MY_FAST_CALL LzmaDec_DecodeReal(CLzmaDec *p, SizeT limit, const Byte
IF_BIT_0(probLen)
{
UPDATE_0(probLen);
- probLen = prob + LenMid + (posState << kLenNumMidBits);
+ probLen = prob + LenLow + GET_LEN_STATE + (1 << kLenNumLowBits);
len = 1;
TREE_GET_BIT(probLen, len);
TREE_GET_BIT(probLen, len);
@@ -356,7 +435,7 @@ static int MY_FAST_CALL LzmaDec_DecodeReal(CLzmaDec *p, SizeT limit, const Byte
UPDATE_1(probLen);
probLen = prob + LenHigh;
TREE_DECODE(probLen, (1 << kLenNumHighBits), len);
- len += kLenNumLowSymbols + kLenNumMidSymbols;
+ len += kLenNumLowSymbols * 2;
}
}
}
@@ -376,16 +455,16 @@ static int MY_FAST_CALL LzmaDec_DecodeReal(CLzmaDec *p, SizeT limit, const Byte
if (posSlot < kEndPosModelIndex)
{
distance <<= numDirectBits;
- prob = probs + SpecPos + distance - posSlot - 1;
+ prob = probs + SpecPos;
{
- UInt32 mask = 1;
- unsigned i = 1;
+ UInt32 m = 1;
+ distance++;
do
{
- GET_BIT2(prob + i, i, ; , distance |= mask);
- mask <<= 1;
+ REV_BIT_VAR(prob, distance, m);
}
- while (--numDirectBits != 0);
+ while (--numDirectBits);
+ distance -= m;
}
}
else
@@ -412,19 +491,20 @@ static int MY_FAST_CALL LzmaDec_DecodeReal(CLzmaDec *p, SizeT limit, const Byte
}
*/
}
- while (--numDirectBits != 0);
+ while (--numDirectBits);
prob = probs + Align;
distance <<= kNumAlignBits;
{
unsigned i = 1;
- GET_BIT2(prob + i, i, ; , distance |= 1);
- GET_BIT2(prob + i, i, ; , distance |= 2);
- GET_BIT2(prob + i, i, ; , distance |= 4);
- GET_BIT2(prob + i, i, ; , distance |= 8);
+ REV_BIT_CONST(prob, i, 1);
+ REV_BIT_CONST(prob, i, 2);
+ REV_BIT_CONST(prob, i, 4);
+ REV_BIT_LAST (prob, i, 8);
+ distance |= i;
}
if (distance == (UInt32)0xFFFFFFFF)
{
- len += kMatchSpecLenStart;
+ len = kMatchSpecLenStart;
state -= kNumStates;
break;
}
@@ -435,20 +515,12 @@ static int MY_FAST_CALL LzmaDec_DecodeReal(CLzmaDec *p, SizeT limit, const Byte
rep2 = rep1;
rep1 = rep0;
rep0 = distance + 1;
- if (checkDicSize == 0)
- {
- if (distance >= processedPos)
- {
- p->dicPos = dicPos;
- return SZ_ERROR_DATA;
- }
- }
- else if (distance >= checkDicSize)
+ state = (state < kNumStates + kNumLitStates) ? kNumLitStates : kNumLitStates + 3;
+ if (distance >= (checkDicSize == 0 ? processedPos: checkDicSize))
{
p->dicPos = dicPos;
return SZ_ERROR_DATA;
}
- state = (state < kNumStates + kNumLitStates) ? kNumLitStates : kNumLitStates + 3;
}
len += kMatchMinLen;
@@ -511,6 +583,7 @@ static int MY_FAST_CALL LzmaDec_DecodeReal(CLzmaDec *p, SizeT limit, const Byte
return SZ_OK;
}
+#endif
static void MY_FAST_CALL LzmaDec_WriteRem(CLzmaDec *p, SizeT limit)
{
@@ -519,7 +592,7 @@ static void MY_FAST_CALL LzmaDec_WriteRem(CLzmaDec *p, SizeT limit)
Byte *dic = p->dic;
SizeT dicPos = p->dicPos;
SizeT dicBufSize = p->dicBufSize;
- unsigned len = p->remainLen;
+ unsigned len = (unsigned)p->remainLen;
SizeT rep0 = p->reps[0]; /* we use SizeT to avoid the BUG of VC14 for AMD64 */
SizeT rem = limit - dicPos;
if (rem < len)
@@ -540,6 +613,14 @@ static void MY_FAST_CALL LzmaDec_WriteRem(CLzmaDec *p, SizeT limit)
}
}
+
+#define kRange0 0xFFFFFFFF
+#define kBound0 ((kRange0 >> kNumBitModelTotalBits) << (kNumBitModelTotalBits - 1))
+#define kBadRepCode (kBound0 + (((kRange0 - kBound0) >> kNumBitModelTotalBits) << (kNumBitModelTotalBits - 1)))
+#if kBadRepCode != (0xC0000000 - 0x400)
+ #error Stop_Compiling_Bad_LZMA_Check
+#endif
+
static int MY_FAST_CALL LzmaDec_DecodeReal2(CLzmaDec *p, SizeT limit, const Byte *bufLimit)
{
do
@@ -550,9 +631,13 @@ static int MY_FAST_CALL LzmaDec_DecodeReal2(CLzmaDec *p, SizeT limit, const Byte
UInt32 rem = p->prop.dicSize - p->processedPos;
if (limit - p->dicPos > rem)
limit2 = p->dicPos + rem;
+
+ if (p->processedPos == 0)
+ if (p->code >= kBadRepCode)
+ return SZ_ERROR_DATA;
}
-
- RINOK(LzmaDec_DecodeReal(p, limit2, bufLimit));
+
+ RINOK(LZMA_DECODE_REAL(p, limit2, bufLimit));
if (p->checkDicSize == 0 && p->processedPos >= p->prop.dicSize)
p->checkDicSize = p->prop.dicSize;
@@ -561,9 +646,6 @@ static int MY_FAST_CALL LzmaDec_DecodeReal2(CLzmaDec *p, SizeT limit, const Byte
}
while (p->dicPos < limit && p->buf < bufLimit && p->remainLen < kMatchSpecLenStart);
- if (p->remainLen > kMatchSpecLenStart)
- p->remainLen = kMatchSpecLenStart;
-
return 0;
}
@@ -580,17 +662,17 @@ static ELzmaDummy LzmaDec_TryDummy(const CLzmaDec *p, const Byte *buf, SizeT inS
UInt32 range = p->range;
UInt32 code = p->code;
const Byte *bufLimit = buf + inSize;
- const CLzmaProb *probs = p->probs;
- unsigned state = p->state;
+ const CLzmaProb *probs = GET_PROBS;
+ unsigned state = (unsigned)p->state;
ELzmaDummy res;
{
const CLzmaProb *prob;
UInt32 bound;
unsigned ttt;
- unsigned posState = (p->processedPos) & ((1 << p->prop.pb) - 1);
+ unsigned posState = CALC_POS_STATE(p->processedPos, (1 << p->prop.pb) - 1);
- prob = probs + IsMatch + (state << kNumPosBitsMax) + posState;
+ prob = probs + IsMatch + COMBINED_PS_STATE;
IF_BIT_0_CHECK(prob)
{
UPDATE_0_CHECK
@@ -618,10 +700,11 @@ static ELzmaDummy LzmaDec_TryDummy(const CLzmaDec *p, const Byte *buf, SizeT inS
{
unsigned bit;
const CLzmaProb *probLit;
- matchByte <<= 1;
- bit = (matchByte & offs);
- probLit = prob + offs + bit + symbol;
- GET_BIT2_CHECK(probLit, symbol, offs &= ~bit, offs &= bit)
+ matchByte += matchByte;
+ bit = offs;
+ offs &= matchByte;
+ probLit = prob + (offs + bit + symbol);
+ GET_BIT2_CHECK(probLit, symbol, offs ^= bit; , ; )
}
while (symbol < 0x100);
}
@@ -648,7 +731,7 @@ static ELzmaDummy LzmaDec_TryDummy(const CLzmaDec *p, const Byte *buf, SizeT inS
IF_BIT_0_CHECK(prob)
{
UPDATE_0_CHECK;
- prob = probs + IsRep0Long + (state << kNumPosBitsMax) + posState;
+ prob = probs + IsRep0Long + COMBINED_PS_STATE;
IF_BIT_0_CHECK(prob)
{
UPDATE_0_CHECK;
@@ -691,7 +774,7 @@ static ELzmaDummy LzmaDec_TryDummy(const CLzmaDec *p, const Byte *buf, SizeT inS
IF_BIT_0_CHECK(probLen)
{
UPDATE_0_CHECK;
- probLen = prob + LenLow + (posState << kLenNumLowBits);
+ probLen = prob + LenLow + GET_LEN_STATE;
offset = 0;
limit = 1 << kLenNumLowBits;
}
@@ -702,15 +785,15 @@ static ELzmaDummy LzmaDec_TryDummy(const CLzmaDec *p, const Byte *buf, SizeT inS
IF_BIT_0_CHECK(probLen)
{
UPDATE_0_CHECK;
- probLen = prob + LenMid + (posState << kLenNumMidBits);
+ probLen = prob + LenLow + GET_LEN_STATE + (1 << kLenNumLowBits);
offset = kLenNumLowSymbols;
- limit = 1 << kLenNumMidBits;
+ limit = 1 << kLenNumLowBits;
}
else
{
UPDATE_1_CHECK;
probLen = prob + LenHigh;
- offset = kLenNumLowSymbols + kLenNumMidSymbols;
+ offset = kLenNumLowSymbols * 2;
limit = 1 << kLenNumHighBits;
}
}
@@ -722,7 +805,7 @@ static ELzmaDummy LzmaDec_TryDummy(const CLzmaDec *p, const Byte *buf, SizeT inS
{
unsigned posSlot;
prob = probs + PosSlot +
- ((len < kNumLenToPosStates ? len : kNumLenToPosStates - 1) <<
+ ((len < kNumLenToPosStates - 1 ? len : kNumLenToPosStates - 1) <<
kNumPosSlotBits);
TREE_DECODE_CHECK(prob, 1 << kNumPosSlotBits, posSlot);
if (posSlot >= kStartPosModelIndex)
@@ -733,7 +816,7 @@ static ELzmaDummy LzmaDec_TryDummy(const CLzmaDec *p, const Byte *buf, SizeT inS
if (posSlot < kEndPosModelIndex)
{
- prob = probs + SpecPos + ((2 | (posSlot & 1)) << numDirectBits) - posSlot - 1;
+ prob = probs + SpecPos + ((2 | (posSlot & 1)) << numDirectBits);
}
else
{
@@ -745,17 +828,18 @@ static ELzmaDummy LzmaDec_TryDummy(const CLzmaDec *p, const Byte *buf, SizeT inS
code -= range & (((code - range) >> 31) - 1);
/* if (code >= range) code -= range; */
}
- while (--numDirectBits != 0);
+ while (--numDirectBits);
prob = probs + Align;
numDirectBits = kNumAlignBits;
}
{
unsigned i = 1;
+ unsigned m = 1;
do
{
- GET_BIT_CHECK(prob + i, i);
+ REV_BIT_CHECK(prob, i, m);
}
- while (--numDirectBits != 0);
+ while (--numDirectBits);
}
}
}
@@ -768,18 +852,17 @@ static ELzmaDummy LzmaDec_TryDummy(const CLzmaDec *p, const Byte *buf, SizeT inS
void LzmaDec_InitDicAndState(CLzmaDec *p, Bool initDic, Bool initState)
{
- p->needFlush = 1;
- p->remainLen = 0;
+ p->remainLen = kMatchSpecLenStart + 1;
p->tempBufSize = 0;
if (initDic)
{
p->processedPos = 0;
p->checkDicSize = 0;
- p->needInitState = 1;
+ p->remainLen = kMatchSpecLenStart + 2;
}
if (initState)
- p->needInitState = 1;
+ p->remainLen = kMatchSpecLenStart + 2;
}
void LzmaDec_Init(CLzmaDec *p)
@@ -788,53 +871,54 @@ void LzmaDec_Init(CLzmaDec *p)
LzmaDec_InitDicAndState(p, True, True);
}
-static void LzmaDec_InitStateReal(CLzmaDec *p)
-{
- SizeT numProbs = LzmaProps_GetNumProbs(&p->prop);
- SizeT i;
- CLzmaProb *probs = p->probs;
- for (i = 0; i < numProbs; i++)
- probs[i] = kBitModelTotal >> 1;
- p->reps[0] = p->reps[1] = p->reps[2] = p->reps[3] = 1;
- p->state = 0;
- p->needInitState = 0;
-}
SRes LzmaDec_DecodeToDic(CLzmaDec *p, SizeT dicLimit, const Byte *src, SizeT *srcLen,
ELzmaFinishMode finishMode, ELzmaStatus *status)
{
SizeT inSize = *srcLen;
(*srcLen) = 0;
- LzmaDec_WriteRem(p, dicLimit);
*status = LZMA_STATUS_NOT_SPECIFIED;
- while (p->remainLen != kMatchSpecLenStart)
+ if (p->remainLen > kMatchSpecLenStart)
{
- int checkEndMarkNow;
+ for (; inSize > 0 && p->tempBufSize < RC_INIT_SIZE; (*srcLen)++, inSize--)
+ p->tempBuf[p->tempBufSize++] = *src++;
+ if (p->tempBufSize != 0 && p->tempBuf[0] != 0)
+ return SZ_ERROR_DATA;
+ if (p->tempBufSize < RC_INIT_SIZE)
+ {
+ *status = LZMA_STATUS_NEEDS_MORE_INPUT;
+ return SZ_OK;
+ }
+ p->code =
+ ((UInt32)p->tempBuf[1] << 24)
+ | ((UInt32)p->tempBuf[2] << 16)
+ | ((UInt32)p->tempBuf[3] << 8)
+ | ((UInt32)p->tempBuf[4]);
+ p->range = 0xFFFFFFFF;
+ p->tempBufSize = 0;
+
+ if (p->remainLen > kMatchSpecLenStart + 1)
+ {
+ SizeT numProbs = LzmaProps_GetNumProbs(&p->prop);
+ SizeT i;
+ CLzmaProb *probs = p->probs;
+ for (i = 0; i < numProbs; i++)
+ probs[i] = kBitModelTotal >> 1;
+ p->reps[0] = p->reps[1] = p->reps[2] = p->reps[3] = 1;
+ p->state = 0;
+ }
- if (p->needFlush)
- {
- for (; inSize > 0 && p->tempBufSize < RC_INIT_SIZE; (*srcLen)++, inSize--)
- p->tempBuf[p->tempBufSize++] = *src++;
- if (p->tempBufSize < RC_INIT_SIZE)
- {
- *status = LZMA_STATUS_NEEDS_MORE_INPUT;
- return SZ_OK;
- }
- if (p->tempBuf[0] != 0)
- return SZ_ERROR_DATA;
- p->code =
- ((UInt32)p->tempBuf[1] << 24)
- | ((UInt32)p->tempBuf[2] << 16)
- | ((UInt32)p->tempBuf[3] << 8)
- | ((UInt32)p->tempBuf[4]);
- p->range = 0xFFFFFFFF;
- p->needFlush = 0;
- p->tempBufSize = 0;
- }
+ p->remainLen = 0;
+ }
+
+ LzmaDec_WriteRem(p, dicLimit);
+
+ while (p->remainLen != kMatchSpecLenStart)
+ {
+ int checkEndMarkNow = 0;
- checkEndMarkNow = 0;
if (p->dicPos >= dicLimit)
{
if (p->remainLen == 0 && p->code == 0)
@@ -855,9 +939,6 @@ SRes LzmaDec_DecodeToDic(CLzmaDec *p, SizeT dicLimit, const Byte *src, SizeT *sr
checkEndMarkNow = 1;
}
- if (p->needInitState)
- LzmaDec_InitStateReal(p);
-
if (p->tempBufSize == 0)
{
SizeT processed;
@@ -930,11 +1011,14 @@ SRes LzmaDec_DecodeToDic(CLzmaDec *p, SizeT dicLimit, const Byte *src, SizeT *sr
p->tempBufSize = 0;
}
}
- if (p->code == 0)
- *status = LZMA_STATUS_FINISHED_WITH_MARK;
- return (p->code == 0) ? SZ_OK : SZ_ERROR_DATA;
+
+ if (p->code != 0)
+ return SZ_ERROR_DATA;
+ *status = LZMA_STATUS_FINISHED_WITH_MARK;
+ return SZ_OK;
}
+
SRes LzmaDec_DecodeToBuf(CLzmaDec *p, Byte *dest, SizeT *destLen, const Byte *src, SizeT *srcLen, ELzmaFinishMode finishMode, ELzmaStatus *status)
{
SizeT outSize = *destLen;
@@ -975,19 +1059,19 @@ SRes LzmaDec_DecodeToBuf(CLzmaDec *p, Byte *dest, SizeT *destLen, const Byte *sr
}
}
-void LzmaDec_FreeProbs(CLzmaDec *p, ISzAlloc *alloc)
+void LzmaDec_FreeProbs(CLzmaDec *p, ISzAllocPtr alloc)
{
- alloc->Free(alloc, p->probs);
+ ISzAlloc_Free(alloc, p->probs);
p->probs = NULL;
}
-static void LzmaDec_FreeDict(CLzmaDec *p, ISzAlloc *alloc)
+static void LzmaDec_FreeDict(CLzmaDec *p, ISzAllocPtr alloc)
{
- alloc->Free(alloc, p->dic);
+ ISzAlloc_Free(alloc, p->dic);
p->dic = NULL;
}
-void LzmaDec_Free(CLzmaDec *p, ISzAlloc *alloc)
+void LzmaDec_Free(CLzmaDec *p, ISzAllocPtr alloc)
{
LzmaDec_FreeProbs(p, alloc);
LzmaDec_FreeDict(p, alloc);
@@ -1011,29 +1095,30 @@ SRes LzmaProps_Decode(CLzmaProps *p, const Byte *data, unsigned size)
if (d >= (9 * 5 * 5))
return SZ_ERROR_UNSUPPORTED;
- p->lc = d % 9;
+ p->lc = (Byte)(d % 9);
d /= 9;
- p->pb = d / 5;
- p->lp = d % 5;
+ p->pb = (Byte)(d / 5);
+ p->lp = (Byte)(d % 5);
return SZ_OK;
}
-static SRes LzmaDec_AllocateProbs2(CLzmaDec *p, const CLzmaProps *propNew, ISzAlloc *alloc)
+static SRes LzmaDec_AllocateProbs2(CLzmaDec *p, const CLzmaProps *propNew, ISzAllocPtr alloc)
{
UInt32 numProbs = LzmaProps_GetNumProbs(propNew);
if (!p->probs || numProbs != p->numProbs)
{
LzmaDec_FreeProbs(p, alloc);
- p->probs = (CLzmaProb *)alloc->Alloc(alloc, numProbs * sizeof(CLzmaProb));
- p->numProbs = numProbs;
+ p->probs = (CLzmaProb *)ISzAlloc_Alloc(alloc, numProbs * sizeof(CLzmaProb));
if (!p->probs)
return SZ_ERROR_MEM;
+ p->probs_1664 = p->probs + 1664;
+ p->numProbs = numProbs;
}
return SZ_OK;
}
-SRes LzmaDec_AllocateProbs(CLzmaDec *p, const Byte *props, unsigned propsSize, ISzAlloc *alloc)
+SRes LzmaDec_AllocateProbs(CLzmaDec *p, const Byte *props, unsigned propsSize, ISzAllocPtr alloc)
{
CLzmaProps propNew;
RINOK(LzmaProps_Decode(&propNew, props, propsSize));
@@ -1042,7 +1127,7 @@ SRes LzmaDec_AllocateProbs(CLzmaDec *p, const Byte *props, unsigned propsSize, I
return SZ_OK;
}
-SRes LzmaDec_Allocate(CLzmaDec *p, const Byte *props, unsigned propsSize, ISzAlloc *alloc)
+SRes LzmaDec_Allocate(CLzmaDec *p, const Byte *props, unsigned propsSize, ISzAllocPtr alloc)
{
CLzmaProps propNew;
SizeT dicBufSize;
@@ -1062,7 +1147,7 @@ SRes LzmaDec_Allocate(CLzmaDec *p, const Byte *props, unsigned propsSize, ISzAll
if (!p->dic || dicBufSize != p->dicBufSize)
{
LzmaDec_FreeDict(p, alloc);
- p->dic = (Byte *)alloc->Alloc(alloc, dicBufSize);
+ p->dic = (Byte *)ISzAlloc_Alloc(alloc, dicBufSize);
if (!p->dic)
{
LzmaDec_FreeProbs(p, alloc);
@@ -1076,7 +1161,7 @@ SRes LzmaDec_Allocate(CLzmaDec *p, const Byte *props, unsigned propsSize, ISzAll
SRes LzmaDecode(Byte *dest, SizeT *destLen, const Byte *src, SizeT *srcLen,
const Byte *propData, unsigned propSize, ELzmaFinishMode finishMode,
- ELzmaStatus *status, ISzAlloc *alloc)
+ ELzmaStatus *status, ISzAllocPtr alloc)
{
CLzmaDec p;
SRes res;
diff --git a/BaseTools/Source/C/LzmaCompress/Sdk/C/LzmaDec.h b/BaseTools/Source/C/LzmaCompress/Sdk/C/LzmaDec.h
index 2633abeac9..28ce60c3ea 100644
--- a/BaseTools/Source/C/LzmaCompress/Sdk/C/LzmaDec.h
+++ b/BaseTools/Source/C/LzmaCompress/Sdk/C/LzmaDec.h
@@ -1,5 +1,5 @@
/* LzmaDec.h -- LZMA Decoder
-2013-01-18 : Igor Pavlov : Public domain */
+2018-04-21 : Igor Pavlov : Public domain */
#ifndef __LZMA_DEC_H
#define __LZMA_DEC_H
@@ -12,11 +12,13 @@ EXTERN_C_BEGIN
/* _LZMA_PROB32 can increase the speed on some CPUs,
but memory usage for CLzmaDec::probs will be doubled in that case */
+typedef
#ifdef _LZMA_PROB32
-#define CLzmaProb UInt32
+ UInt32
#else
-#define CLzmaProb UInt16
+ UInt16
#endif
+ CLzmaProb;
/* ---------- LZMA Properties ---------- */
@@ -25,7 +27,10 @@ EXTERN_C_BEGIN
typedef struct _CLzmaProps
{
- unsigned lc, lp, pb;
+ Byte lc;
+ Byte lp;
+ Byte pb;
+ Byte _pad_;
UInt32 dicSize;
} CLzmaProps;
@@ -47,32 +52,34 @@ SRes LzmaProps_Decode(CLzmaProps *p, const Byte *data, unsigned size);
typedef struct
{
+ /* Don't change this structure. ASM code can use it. */
CLzmaProps prop;
CLzmaProb *probs;
+ CLzmaProb *probs_1664;
Byte *dic;
- const Byte *buf;
- UInt32 range, code;
- SizeT dicPos;
SizeT dicBufSize;
+ SizeT dicPos;
+ const Byte *buf;
+ UInt32 range;
+ UInt32 code;
UInt32 processedPos;
UInt32 checkDicSize;
- unsigned state;
UInt32 reps[4];
- unsigned remainLen;
- int needFlush;
- int needInitState;
+ UInt32 state;
+ UInt32 remainLen;
+
UInt32 numProbs;
unsigned tempBufSize;
Byte tempBuf[LZMA_REQUIRED_INPUT_MAX];
} CLzmaDec;
-#define LzmaDec_Construct(p) { (p)->dic = 0; (p)->probs = 0; }
+#define LzmaDec_Construct(p) { (p)->dic = NULL; (p)->probs = NULL; }
void LzmaDec_Init(CLzmaDec *p);
/* There are two types of LZMA streams:
- 0) Stream with end mark. That end mark adds about 6 bytes to compressed size.
- 1) Stream without end mark. You must know exact uncompressed size to decompress such stream. */
+ - Stream with end mark. That end mark adds about 6 bytes to compressed size.
+ - Stream without end mark. You must know exact uncompressed size to decompress such stream. */
typedef enum
{
@@ -129,11 +136,11 @@ LzmaDec_Allocate* can return:
SZ_ERROR_UNSUPPORTED - Unsupported properties
*/
-SRes LzmaDec_AllocateProbs(CLzmaDec *p, const Byte *props, unsigned propsSize, ISzAlloc *alloc);
-void LzmaDec_FreeProbs(CLzmaDec *p, ISzAlloc *alloc);
+SRes LzmaDec_AllocateProbs(CLzmaDec *p, const Byte *props, unsigned propsSize, ISzAllocPtr alloc);
+void LzmaDec_FreeProbs(CLzmaDec *p, ISzAllocPtr alloc);
-SRes LzmaDec_Allocate(CLzmaDec *state, const Byte *prop, unsigned propsSize, ISzAlloc *alloc);
-void LzmaDec_Free(CLzmaDec *state, ISzAlloc *alloc);
+SRes LzmaDec_Allocate(CLzmaDec *p, const Byte *props, unsigned propsSize, ISzAllocPtr alloc);
+void LzmaDec_Free(CLzmaDec *p, ISzAllocPtr alloc);
/* ---------- Dictionary Interface ---------- */
@@ -142,7 +149,7 @@ void LzmaDec_Free(CLzmaDec *state, ISzAlloc *alloc);
You must work with CLzmaDec variables directly in this interface.
STEPS:
- LzmaDec_Constr()
+ LzmaDec_Construct()
LzmaDec_Allocate()
for (each new stream)
{
@@ -220,7 +227,7 @@ Returns:
SRes LzmaDecode(Byte *dest, SizeT *destLen, const Byte *src, SizeT *srcLen,
const Byte *propData, unsigned propSize, ELzmaFinishMode finishMode,
- ELzmaStatus *status, ISzAlloc *alloc);
+ ELzmaStatus *status, ISzAllocPtr alloc);
EXTERN_C_END
diff --git a/BaseTools/Source/C/LzmaCompress/Sdk/C/LzmaEnc.c b/BaseTools/Source/C/LzmaCompress/Sdk/C/LzmaEnc.c
index 462ca67565..bebe664d3e 100644
--- a/BaseTools/Source/C/LzmaCompress/Sdk/C/LzmaEnc.c
+++ b/BaseTools/Source/C/LzmaCompress/Sdk/C/LzmaEnc.c
@@ -1,5 +1,5 @@
/* LzmaEnc.c -- LZMA Encoder
-2016-05-16 : Igor Pavlov : Public domain */
+2018-04-29 : Igor Pavlov : Public domain */
#include "Precomp.h"
@@ -23,17 +23,8 @@
static unsigned g_STAT_OFFSET = 0;
#endif
-#define kMaxHistorySize ((UInt32)3 << 29)
-/* #define kMaxHistorySize ((UInt32)7 << 29) */
-
-#define kBlockSizeMax ((1 << LZMA_NUM_BLOCK_SIZE_BITS) - 1)
-
-#define kBlockSize (9 << 10)
-#define kUnpackBlockSize (1 << 18)
-#define kMatchArraySize (1 << 21)
-#define kMatchRecordMaxSize ((LZMA_MATCH_LEN_MAX * 2 + 3) * LZMA_MATCH_LEN_MAX)
-
-#define kNumMaxDirectBits (31)
+#define kLzmaMaxHistorySize ((UInt32)3 << 29)
+/* #define kLzmaMaxHistorySize ((UInt32)7 << 29) */
#define kNumTopBits 24
#define kTopValue ((UInt32)1 << kNumTopBits)
@@ -62,14 +53,15 @@ void LzmaEncProps_Normalize(CLzmaEncProps *p)
if (level < 0) level = 5;
p->level = level;
- if (p->dictSize == 0) p->dictSize = (level <= 5 ? (1 << (level * 2 + 14)) : (level == 6 ? (1 << 25) : (1 << 26)));
+ if (p->dictSize == 0) p->dictSize = (level <= 5 ? (1 << (level * 2 + 14)) : (level <= 7 ? (1 << 25) : (1 << 26)));
if (p->dictSize > p->reduceSize)
{
unsigned i;
+ UInt32 reduceSize = (UInt32)p->reduceSize;
for (i = 11; i <= 30; i++)
{
- if ((UInt32)p->reduceSize <= ((UInt32)2 << i)) { p->dictSize = ((UInt32)2 << i); break; }
- if ((UInt32)p->reduceSize <= ((UInt32)3 << i)) { p->dictSize = ((UInt32)3 << i); break; }
+ if (reduceSize <= ((UInt32)2 << i)) { p->dictSize = ((UInt32)2 << i); break; }
+ if (reduceSize <= ((UInt32)3 << i)) { p->dictSize = ((UInt32)3 << i); break; }
}
}
@@ -110,9 +102,9 @@ UInt32 LzmaEncProps_GetDictSize(const CLzmaEncProps *props2)
#define BSR2_RET(pos, res) { unsigned long zz; _BitScanReverse(&zz, (pos)); res = (zz + zz) + ((pos >> (zz - 1)) & 1); }
-static UInt32 GetPosSlot1(UInt32 pos)
+static unsigned GetPosSlot1(UInt32 pos)
{
- UInt32 res;
+ unsigned res;
BSR2_RET(pos, res);
return res;
}
@@ -145,18 +137,18 @@ static void LzmaEnc_FastPosInit(Byte *g_FastPos)
/* we can use ((limit - pos) >> 31) only if (pos < ((UInt32)1 << 31)) */
/*
-#define BSR2_RET(pos, res) { UInt32 zz = 6 + ((kNumLogBits - 1) & \
+#define BSR2_RET(pos, res) { unsigned zz = 6 + ((kNumLogBits - 1) & \
(0 - (((((UInt32)1 << (kNumLogBits + 6)) - 1) - pos) >> 31))); \
res = p->g_FastPos[pos >> zz] + (zz * 2); }
*/
/*
-#define BSR2_RET(pos, res) { UInt32 zz = 6 + ((kNumLogBits - 1) & \
+#define BSR2_RET(pos, res) { unsigned zz = 6 + ((kNumLogBits - 1) & \
(0 - (((((UInt32)1 << (kNumLogBits)) - 1) - (pos >> 6)) >> 31))); \
res = p->g_FastPos[pos >> zz] + (zz * 2); }
*/
-#define BSR2_RET(pos, res) { UInt32 zz = (pos < (1 << (kNumLogBits + 6))) ? 6 : 6 + kNumLogBits - 1; \
+#define BSR2_RET(pos, res) { unsigned zz = (pos < (1 << (kNumLogBits + 6))) ? 6 : 6 + kNumLogBits - 1; \
res = p->g_FastPos[pos >> zz] + (zz * 2); }
/*
@@ -167,32 +159,32 @@ static void LzmaEnc_FastPosInit(Byte *g_FastPos)
#define GetPosSlot1(pos) p->g_FastPos[pos]
#define GetPosSlot2(pos, res) { BSR2_RET(pos, res); }
-#define GetPosSlot(pos, res) { if (pos < kNumFullDistances) res = p->g_FastPos[pos]; else BSR2_RET(pos, res); }
+#define GetPosSlot(pos, res) { if (pos < kNumFullDistances) res = p->g_FastPos[pos & (kNumFullDistances - 1)]; else BSR2_RET(pos, res); }
#endif
#define LZMA_NUM_REPS 4
-typedef unsigned CState;
+typedef UInt16 CState;
+typedef UInt16 CExtra;
typedef struct
{
UInt32 price;
-
CState state;
- int prev1IsChar;
- int prev2;
-
- UInt32 posPrev2;
- UInt32 backPrev2;
-
- UInt32 posPrev;
- UInt32 backPrev;
- UInt32 backs[LZMA_NUM_REPS];
+ CExtra extra;
+ // 0 : normal
+ // 1 : LIT : MATCH
+ // > 1 : MATCH (extra-1) : LIT : REP0 (len)
+ UInt32 len;
+ UInt32 dist;
+ UInt32 reps[LZMA_NUM_REPS];
} COptimal;
+
#define kNumOpts (1 << 12)
+#define kPackReserve (1 + kNumOpts * 2)
#define kNumLenToPosStates 4
#define kNumPosSlotBits 6
@@ -200,22 +192,21 @@ typedef struct
#define kDicLogSizeMax 32
#define kDistTableSizeMax (kDicLogSizeMax * 2)
-
#define kNumAlignBits 4
#define kAlignTableSize (1 << kNumAlignBits)
#define kAlignMask (kAlignTableSize - 1)
#define kStartPosModelIndex 4
#define kEndPosModelIndex 14
-#define kNumPosModels (kEndPosModelIndex - kStartPosModelIndex)
-
#define kNumFullDistances (1 << (kEndPosModelIndex >> 1))
+typedef
#ifdef _LZMA_PROB32
-#define CLzmaProb UInt32
+ UInt32
#else
-#define CLzmaProb UInt16
+ UInt16
#endif
+ CLzmaProb;
#define LZMA_PB_MAX 4
#define LZMA_LC_MAX 8
@@ -223,15 +214,11 @@ typedef struct
#define LZMA_NUM_PB_STATES_MAX (1 << LZMA_PB_MAX)
-
#define kLenNumLowBits 3
#define kLenNumLowSymbols (1 << kLenNumLowBits)
-#define kLenNumMidBits 3
-#define kLenNumMidSymbols (1 << kLenNumMidBits)
#define kLenNumHighBits 8
#define kLenNumHighSymbols (1 << kLenNumHighBits)
-
-#define kLenNumSymbolsTotal (kLenNumLowSymbols + kLenNumMidSymbols + kLenNumHighSymbols)
+#define kLenNumSymbolsTotal (kLenNumLowSymbols * 2 + kLenNumHighSymbols)
#define LZMA_MATCH_LEN_MIN 2
#define LZMA_MATCH_LEN_MAX (LZMA_MATCH_LEN_MIN + kLenNumSymbolsTotal - 1)
@@ -241,27 +228,23 @@ typedef struct
typedef struct
{
- CLzmaProb choice;
- CLzmaProb choice2;
- CLzmaProb low[LZMA_NUM_PB_STATES_MAX << kLenNumLowBits];
- CLzmaProb mid[LZMA_NUM_PB_STATES_MAX << kLenNumMidBits];
+ CLzmaProb low[LZMA_NUM_PB_STATES_MAX << (kLenNumLowBits + 1)];
CLzmaProb high[kLenNumHighSymbols];
} CLenEnc;
typedef struct
{
- CLenEnc p;
- UInt32 tableSize;
+ unsigned tableSize;
+ unsigned counters[LZMA_NUM_PB_STATES_MAX];
UInt32 prices[LZMA_NUM_PB_STATES_MAX][kLenNumSymbolsTotal];
- UInt32 counters[LZMA_NUM_PB_STATES_MAX];
} CLenPriceEnc;
typedef struct
{
UInt32 range;
- Byte cache;
+ unsigned cache;
UInt64 low;
UInt64 cacheSize;
Byte *buf;
@@ -277,48 +260,54 @@ typedef struct
{
CLzmaProb *litProbs;
- UInt32 state;
+ unsigned state;
UInt32 reps[LZMA_NUM_REPS];
- CLzmaProb isMatch[kNumStates][LZMA_NUM_PB_STATES_MAX];
+ CLzmaProb posAlignEncoder[1 << kNumAlignBits];
CLzmaProb isRep[kNumStates];
CLzmaProb isRepG0[kNumStates];
CLzmaProb isRepG1[kNumStates];
CLzmaProb isRepG2[kNumStates];
+ CLzmaProb isMatch[kNumStates][LZMA_NUM_PB_STATES_MAX];
CLzmaProb isRep0Long[kNumStates][LZMA_NUM_PB_STATES_MAX];
CLzmaProb posSlotEncoder[kNumLenToPosStates][1 << kNumPosSlotBits];
- CLzmaProb posEncoders[kNumFullDistances - kEndPosModelIndex];
- CLzmaProb posAlignEncoder[1 << kNumAlignBits];
+ CLzmaProb posEncoders[kNumFullDistances];
- CLenPriceEnc lenEnc;
- CLenPriceEnc repLenEnc;
+ CLenEnc lenProbs;
+ CLenEnc repLenProbs;
+
} CSaveState;
+typedef UInt32 CProbPrice;
+
+
typedef struct
{
void *matchFinderObj;
IMatchFinder matchFinder;
- UInt32 optimumEndIndex;
- UInt32 optimumCurrentIndex;
+ unsigned optCur;
+ unsigned optEnd;
- UInt32 longestMatchLength;
- UInt32 numPairs;
+ unsigned longestMatchLen;
+ unsigned numPairs;
UInt32 numAvail;
- UInt32 numFastBytes;
- UInt32 additionalOffset;
+ unsigned state;
+ unsigned numFastBytes;
+ unsigned additionalOffset;
UInt32 reps[LZMA_NUM_REPS];
- UInt32 state;
+ unsigned lpMask, pbMask;
+ CLzmaProb *litProbs;
+ CRangeEnc rc;
+
+ UInt32 backRes;
unsigned lc, lp, pb;
- unsigned lpMask, pbMask;
unsigned lclp;
- CLzmaProb *litProbs;
-
Bool fastMode;
Bool writeEndMark;
Bool finished;
@@ -327,19 +316,19 @@ typedef struct
UInt64 nowPos64;
- UInt32 matchPriceCount;
- UInt32 alignPriceCount;
+ unsigned matchPriceCount;
+ unsigned alignPriceCount;
- UInt32 distTableSize;
+ unsigned distTableSize;
UInt32 dictSize;
SRes result;
- CRangeEnc rc;
-
#ifndef _7ZIP_ST
Bool mtMode;
+ // begin of CMatchFinderMt is used in LZ thread
CMatchFinderMt matchFinderMt;
+ // end of CMatchFinderMt is used in BT and HASH threads
#endif
CMatchFinder matchFinderBase;
@@ -348,33 +337,37 @@ typedef struct
Byte pad[128];
#endif
- COptimal opt[kNumOpts];
-
- #ifndef LZMA_LOG_BSR
- Byte g_FastPos[1 << kNumLogBits];
- #endif
+ // LZ thread
+ CProbPrice ProbPrices[kBitModelTotal >> kNumMoveReducingBits];
- UInt32 ProbPrices[kBitModelTotal >> kNumMoveReducingBits];
UInt32 matches[LZMA_MATCH_LEN_MAX * 2 + 2 + 1];
+ UInt32 alignPrices[kAlignTableSize];
UInt32 posSlotPrices[kNumLenToPosStates][kDistTableSizeMax];
UInt32 distancesPrices[kNumLenToPosStates][kNumFullDistances];
- UInt32 alignPrices[kAlignTableSize];
- CLzmaProb isMatch[kNumStates][LZMA_NUM_PB_STATES_MAX];
+ CLzmaProb posAlignEncoder[1 << kNumAlignBits];
CLzmaProb isRep[kNumStates];
CLzmaProb isRepG0[kNumStates];
CLzmaProb isRepG1[kNumStates];
CLzmaProb isRepG2[kNumStates];
+ CLzmaProb isMatch[kNumStates][LZMA_NUM_PB_STATES_MAX];
CLzmaProb isRep0Long[kNumStates][LZMA_NUM_PB_STATES_MAX];
-
CLzmaProb posSlotEncoder[kNumLenToPosStates][1 << kNumPosSlotBits];
- CLzmaProb posEncoders[kNumFullDistances - kEndPosModelIndex];
- CLzmaProb posAlignEncoder[1 << kNumAlignBits];
+ CLzmaProb posEncoders[kNumFullDistances];
+ CLenEnc lenProbs;
+ CLenEnc repLenProbs;
+
+ #ifndef LZMA_LOG_BSR
+ Byte g_FastPos[1 << kNumLogBits];
+ #endif
+
CLenPriceEnc lenEnc;
CLenPriceEnc repLenEnc;
+ COptimal opt[kNumOpts];
+
CSaveState saveState;
#ifndef _7ZIP_ST
@@ -383,58 +376,62 @@ typedef struct
} CLzmaEnc;
+
+#define COPY_ARR(dest, src, arr) memcpy(dest->arr, src->arr, sizeof(src->arr));
+
void LzmaEnc_SaveState(CLzmaEncHandle pp)
{
CLzmaEnc *p = (CLzmaEnc *)pp;
CSaveState *dest = &p->saveState;
- int i;
- dest->lenEnc = p->lenEnc;
- dest->repLenEnc = p->repLenEnc;
+
dest->state = p->state;
+
+ dest->lenProbs = p->lenProbs;
+ dest->repLenProbs = p->repLenProbs;
+
+ COPY_ARR(dest, p, reps);
+
+ COPY_ARR(dest, p, posAlignEncoder);
+ COPY_ARR(dest, p, isRep);
+ COPY_ARR(dest, p, isRepG0);
+ COPY_ARR(dest, p, isRepG1);
+ COPY_ARR(dest, p, isRepG2);
+ COPY_ARR(dest, p, isMatch);
+ COPY_ARR(dest, p, isRep0Long);
+ COPY_ARR(dest, p, posSlotEncoder);
+ COPY_ARR(dest, p, posEncoders);
- for (i = 0; i < kNumStates; i++)
- {
- memcpy(dest->isMatch[i], p->isMatch[i], sizeof(p->isMatch[i]));
- memcpy(dest->isRep0Long[i], p->isRep0Long[i], sizeof(p->isRep0Long[i]));
- }
- for (i = 0; i < kNumLenToPosStates; i++)
- memcpy(dest->posSlotEncoder[i], p->posSlotEncoder[i], sizeof(p->posSlotEncoder[i]));
- memcpy(dest->isRep, p->isRep, sizeof(p->isRep));
- memcpy(dest->isRepG0, p->isRepG0, sizeof(p->isRepG0));
- memcpy(dest->isRepG1, p->isRepG1, sizeof(p->isRepG1));
- memcpy(dest->isRepG2, p->isRepG2, sizeof(p->isRepG2));
- memcpy(dest->posEncoders, p->posEncoders, sizeof(p->posEncoders));
- memcpy(dest->posAlignEncoder, p->posAlignEncoder, sizeof(p->posAlignEncoder));
- memcpy(dest->reps, p->reps, sizeof(p->reps));
memcpy(dest->litProbs, p->litProbs, ((UInt32)0x300 << p->lclp) * sizeof(CLzmaProb));
}
+
void LzmaEnc_RestoreState(CLzmaEncHandle pp)
{
CLzmaEnc *dest = (CLzmaEnc *)pp;
const CSaveState *p = &dest->saveState;
- int i;
- dest->lenEnc = p->lenEnc;
- dest->repLenEnc = p->repLenEnc;
+
dest->state = p->state;
- for (i = 0; i < kNumStates; i++)
- {
- memcpy(dest->isMatch[i], p->isMatch[i], sizeof(p->isMatch[i]));
- memcpy(dest->isRep0Long[i], p->isRep0Long[i], sizeof(p->isRep0Long[i]));
- }
- for (i = 0; i < kNumLenToPosStates; i++)
- memcpy(dest->posSlotEncoder[i], p->posSlotEncoder[i], sizeof(p->posSlotEncoder[i]));
- memcpy(dest->isRep, p->isRep, sizeof(p->isRep));
- memcpy(dest->isRepG0, p->isRepG0, sizeof(p->isRepG0));
- memcpy(dest->isRepG1, p->isRepG1, sizeof(p->isRepG1));
- memcpy(dest->isRepG2, p->isRepG2, sizeof(p->isRepG2));
- memcpy(dest->posEncoders, p->posEncoders, sizeof(p->posEncoders));
- memcpy(dest->posAlignEncoder, p->posAlignEncoder, sizeof(p->posAlignEncoder));
- memcpy(dest->reps, p->reps, sizeof(p->reps));
+ dest->lenProbs = p->lenProbs;
+ dest->repLenProbs = p->repLenProbs;
+
+ COPY_ARR(dest, p, reps);
+
+ COPY_ARR(dest, p, posAlignEncoder);
+ COPY_ARR(dest, p, isRep);
+ COPY_ARR(dest, p, isRepG0);
+ COPY_ARR(dest, p, isRepG1);
+ COPY_ARR(dest, p, isRepG2);
+ COPY_ARR(dest, p, isMatch);
+ COPY_ARR(dest, p, isRep0Long);
+ COPY_ARR(dest, p, posSlotEncoder);
+ COPY_ARR(dest, p, posEncoders);
+
memcpy(dest->litProbs, p->litProbs, ((UInt32)0x300 << dest->lclp) * sizeof(CLzmaProb));
}
+
+
SRes LzmaEnc_SetProps(CLzmaEncHandle pp, const CLzmaEncProps *props2)
{
CLzmaEnc *p = (CLzmaEnc *)pp;
@@ -445,7 +442,7 @@ SRes LzmaEnc_SetProps(CLzmaEncHandle pp, const CLzmaEncProps *props2)
|| props.lp > LZMA_LP_MAX
|| props.pb > LZMA_PB_MAX
|| props.dictSize > ((UInt64)1 << kDicLogSizeMaxCompress)
- || props.dictSize > kMaxHistorySize)
+ || props.dictSize > kLzmaMaxHistorySize)
return SZ_ERROR_PARAM;
p->dictSize = props.dictSize;
@@ -463,7 +460,7 @@ SRes LzmaEnc_SetProps(CLzmaEncHandle pp, const CLzmaEncProps *props2)
p->fastMode = (props.algo == 0);
p->matchFinderBase.btMode = (Byte)(props.btMode ? 1 : 0);
{
- UInt32 numHashBytes = 4;
+ unsigned numHashBytes = 4;
if (props.btMode)
{
if (props.numHashBytes < 2)
@@ -492,13 +489,27 @@ SRes LzmaEnc_SetProps(CLzmaEncHandle pp, const CLzmaEncProps *props2)
return SZ_OK;
}
-static const int kLiteralNextStates[kNumStates] = {0, 0, 0, 0, 1, 2, 3, 4, 5, 6, 4, 5};
-static const int kMatchNextStates[kNumStates] = {7, 7, 7, 7, 7, 7, 7, 10, 10, 10, 10, 10};
-static const int kRepNextStates[kNumStates] = {8, 8, 8, 8, 8, 8, 8, 11, 11, 11, 11, 11};
-static const int kShortRepNextStates[kNumStates]= {9, 9, 9, 9, 9, 9, 9, 11, 11, 11, 11, 11};
-#define IsCharState(s) ((s) < 7)
+void LzmaEnc_SetDataSize(CLzmaEncHandle pp, UInt64 expectedDataSiize)
+{
+ CLzmaEnc *p = (CLzmaEnc *)pp;
+ p->matchFinderBase.expectedDataSize = expectedDataSiize;
+}
+
+#define kState_Start 0
+#define kState_LitAfterMatch 4
+#define kState_LitAfterRep 5
+#define kState_MatchAfterLit 7
+#define kState_RepAfterLit 8
+
+static const Byte kLiteralNextStates[kNumStates] = {0, 0, 0, 0, 1, 2, 3, 4, 5, 6, 4, 5};
+static const Byte kMatchNextStates[kNumStates] = {7, 7, 7, 7, 7, 7, 7, 10, 10, 10, 10, 10};
+static const Byte kRepNextStates[kNumStates] = {8, 8, 8, 8, 8, 8, 8, 11, 11, 11, 11, 11};
+static const Byte kShortRepNextStates[kNumStates]= {9, 9, 9, 9, 9, 9, 9, 11, 11, 11, 11, 11};
+
+#define IsLitState(s) ((s) < 7)
+#define GetLenToPosState2(len) (((len) < kNumLenToPosStates - 1) ? (len) : kNumLenToPosStates - 1)
#define GetLenToPosState(len) (((len) < kNumLenToPosStates + 1) ? (len) - 2 : kNumLenToPosStates - 1)
#define kInfinityPrice (1 << 30)
@@ -509,14 +520,16 @@ static void RangeEnc_Construct(CRangeEnc *p)
p->bufBase = NULL;
}
-#define RangeEnc_GetProcessed(p) ((p)->processed + ((p)->buf - (p)->bufBase) + (p)->cacheSize)
+#define RangeEnc_GetProcessed(p) ((p)->processed + ((p)->buf - (p)->bufBase) + (p)->cacheSize)
+#define RangeEnc_GetProcessed_sizet(p) ((size_t)(p)->processed + ((p)->buf - (p)->bufBase) + (size_t)(p)->cacheSize)
#define RC_BUF_SIZE (1 << 16)
-static int RangeEnc_Alloc(CRangeEnc *p, ISzAlloc *alloc)
+
+static int RangeEnc_Alloc(CRangeEnc *p, ISzAllocPtr alloc)
{
if (!p->bufBase)
{
- p->bufBase = (Byte *)alloc->Alloc(alloc, RC_BUF_SIZE);
+ p->bufBase = (Byte *)ISzAlloc_Alloc(alloc, RC_BUF_SIZE);
if (!p->bufBase)
return 0;
p->bufLim = p->bufBase + RC_BUF_SIZE;
@@ -524,19 +537,19 @@ static int RangeEnc_Alloc(CRangeEnc *p, ISzAlloc *alloc)
return 1;
}
-static void RangeEnc_Free(CRangeEnc *p, ISzAlloc *alloc)
+static void RangeEnc_Free(CRangeEnc *p, ISzAllocPtr alloc)
{
- alloc->Free(alloc, p->bufBase);
+ ISzAlloc_Free(alloc, p->bufBase);
p->bufBase = 0;
}
static void RangeEnc_Init(CRangeEnc *p)
{
/* Stream.Init(); */
- p->low = 0;
p->range = 0xFFFFFFFF;
- p->cacheSize = 1;
p->cache = 0;
+ p->low = 0;
+ p->cacheSize = 0;
p->buf = p->bufBase;
@@ -544,37 +557,48 @@ static void RangeEnc_Init(CRangeEnc *p)
p->res = SZ_OK;
}
-static void RangeEnc_FlushStream(CRangeEnc *p)
+MY_NO_INLINE static void RangeEnc_FlushStream(CRangeEnc *p)
{
size_t num;
if (p->res != SZ_OK)
return;
num = p->buf - p->bufBase;
- if (num != p->outStream->Write(p->outStream, p->bufBase, num))
+ if (num != ISeqOutStream_Write(p->outStream, p->bufBase, num))
p->res = SZ_ERROR_WRITE;
p->processed += num;
p->buf = p->bufBase;
}
-static void MY_FAST_CALL RangeEnc_ShiftLow(CRangeEnc *p)
+MY_NO_INLINE static void MY_FAST_CALL RangeEnc_ShiftLow(CRangeEnc *p)
{
- if ((UInt32)p->low < (UInt32)0xFF000000 || (unsigned)(p->low >> 32) != 0)
+ UInt32 low = (UInt32)p->low;
+ unsigned high = (unsigned)(p->low >> 32);
+ p->low = (UInt32)(low << 8);
+ if (low < (UInt32)0xFF000000 || high != 0)
{
- Byte temp = p->cache;
- do
{
Byte *buf = p->buf;
- *buf++ = (Byte)(temp + (Byte)(p->low >> 32));
+ *buf++ = (Byte)(p->cache + high);
+ p->cache = (unsigned)(low >> 24);
p->buf = buf;
if (buf == p->bufLim)
RangeEnc_FlushStream(p);
- temp = 0xFF;
+ if (p->cacheSize == 0)
+ return;
+ }
+ high += 0xFF;
+ for (;;)
+ {
+ Byte *buf = p->buf;
+ *buf++ = (Byte)(high);
+ p->buf = buf;
+ if (buf == p->bufLim)
+ RangeEnc_FlushStream(p);
+ if (--p->cacheSize == 0)
+ return;
}
- while (--p->cacheSize != 0);
- p->cache = (Byte)((UInt32)p->low >> 24);
}
p->cacheSize++;
- p->low = (UInt32)p->low << 8;
}
static void RangeEnc_FlushData(CRangeEnc *p)
@@ -584,78 +608,121 @@ static void RangeEnc_FlushData(CRangeEnc *p)
RangeEnc_ShiftLow(p);
}
-static void RangeEnc_EncodeDirectBits(CRangeEnc *p, UInt32 value, unsigned numBits)
-{
- do
- {
- p->range >>= 1;
- p->low += p->range & (0 - ((value >> --numBits) & 1));
- if (p->range < kTopValue)
- {
- p->range <<= 8;
- RangeEnc_ShiftLow(p);
- }
- }
- while (numBits != 0);
-}
+#define RC_NORM(p) if (range < kTopValue) { range <<= 8; RangeEnc_ShiftLow(p); }
-static void RangeEnc_EncodeBit(CRangeEnc *p, CLzmaProb *prob, UInt32 symbol)
-{
- UInt32 ttt = *prob;
- UInt32 newBound = (p->range >> kNumBitModelTotalBits) * ttt;
- if (symbol == 0)
- {
- p->range = newBound;
- ttt += (kBitModelTotal - ttt) >> kNumMoveBits;
- }
- else
- {
- p->low += newBound;
- p->range -= newBound;
- ttt -= ttt >> kNumMoveBits;
+#define RC_BIT_PRE(p, prob) \
+ ttt = *(prob); \
+ newBound = (range >> kNumBitModelTotalBits) * ttt;
+
+// #define _LZMA_ENC_USE_BRANCH
+
+#ifdef _LZMA_ENC_USE_BRANCH
+
+#define RC_BIT(p, prob, symbol) { \
+ RC_BIT_PRE(p, prob) \
+ if (symbol == 0) { range = newBound; ttt += (kBitModelTotal - ttt) >> kNumMoveBits; } \
+ else { (p)->low += newBound; range -= newBound; ttt -= ttt >> kNumMoveBits; } \
+ *(prob) = (CLzmaProb)ttt; \
+ RC_NORM(p) \
}
- *prob = (CLzmaProb)ttt;
- if (p->range < kTopValue)
- {
- p->range <<= 8;
- RangeEnc_ShiftLow(p);
+
+#else
+
+#define RC_BIT(p, prob, symbol) { \
+ UInt32 mask; \
+ RC_BIT_PRE(p, prob) \
+ mask = 0 - (UInt32)symbol; \
+ range &= mask; \
+ mask &= newBound; \
+ range -= mask; \
+ (p)->low += mask; \
+ mask = (UInt32)symbol - 1; \
+ range += newBound & mask; \
+ mask &= (kBitModelTotal - ((1 << kNumMoveBits) - 1)); \
+ mask += ((1 << kNumMoveBits) - 1); \
+ ttt += (Int32)(mask - ttt) >> kNumMoveBits; \
+ *(prob) = (CLzmaProb)ttt; \
+ RC_NORM(p) \
}
+
+#endif
+
+
+
+
+#define RC_BIT_0_BASE(p, prob) \
+ range = newBound; *(prob) = (CLzmaProb)(ttt + ((kBitModelTotal - ttt) >> kNumMoveBits));
+
+#define RC_BIT_1_BASE(p, prob) \
+ range -= newBound; (p)->low += newBound; *(prob) = (CLzmaProb)(ttt - (ttt >> kNumMoveBits)); \
+
+#define RC_BIT_0(p, prob) \
+ RC_BIT_0_BASE(p, prob) \
+ RC_NORM(p)
+
+#define RC_BIT_1(p, prob) \
+ RC_BIT_1_BASE(p, prob) \
+ RC_NORM(p)
+
+static void RangeEnc_EncodeBit_0(CRangeEnc *p, CLzmaProb *prob)
+{
+ UInt32 range, ttt, newBound;
+ range = p->range;
+ RC_BIT_PRE(p, prob)
+ RC_BIT_0(p, prob)
+ p->range = range;
}
static void LitEnc_Encode(CRangeEnc *p, CLzmaProb *probs, UInt32 symbol)
{
+ UInt32 range = p->range;
symbol |= 0x100;
do
{
- RangeEnc_EncodeBit(p, probs + (symbol >> 8), (symbol >> 7) & 1);
+ UInt32 ttt, newBound;
+ // RangeEnc_EncodeBit(p, probs + (symbol >> 8), (symbol >> 7) & 1);
+ CLzmaProb *prob = probs + (symbol >> 8);
+ UInt32 bit = (symbol >> 7) & 1;
symbol <<= 1;
+ RC_BIT(p, prob, bit);
}
while (symbol < 0x10000);
+ p->range = range;
}
static void LitEnc_EncodeMatched(CRangeEnc *p, CLzmaProb *probs, UInt32 symbol, UInt32 matchByte)
{
+ UInt32 range = p->range;
UInt32 offs = 0x100;
symbol |= 0x100;
do
{
+ UInt32 ttt, newBound;
+ CLzmaProb *prob;
+ UInt32 bit;
matchByte <<= 1;
- RangeEnc_EncodeBit(p, probs + (offs + (matchByte & offs) + (symbol >> 8)), (symbol >> 7) & 1);
+ // RangeEnc_EncodeBit(p, probs + (offs + (matchByte & offs) + (symbol >> 8)), (symbol >> 7) & 1);
+ prob = probs + (offs + (matchByte & offs) + (symbol >> 8));
+ bit = (symbol >> 7) & 1;
symbol <<= 1;
offs &= ~(matchByte ^ symbol);
+ RC_BIT(p, prob, bit);
}
while (symbol < 0x10000);
+ p->range = range;
}
-static void LzmaEnc_InitPriceTables(UInt32 *ProbPrices)
+
+
+static void LzmaEnc_InitPriceTables(CProbPrice *ProbPrices)
{
UInt32 i;
- for (i = (1 << kNumMoveReducingBits) / 2; i < kBitModelTotal; i += (1 << kNumMoveReducingBits))
+ for (i = 0; i < (kBitModelTotal >> kNumMoveReducingBits); i++)
{
- const int kCyclesBits = kNumBitPriceShiftBits;
- UInt32 w = i;
- UInt32 bitCount = 0;
- int j;
+ const unsigned kCyclesBits = kNumBitPriceShiftBits;
+ UInt32 w = (i << kNumMoveReducingBits) + (1 << (kNumMoveReducingBits - 1));
+ unsigned bitCount = 0;
+ unsigned j;
for (j = 0; j < kCyclesBits; j++)
{
w = w * w;
@@ -666,37 +733,41 @@ static void LzmaEnc_InitPriceTables(UInt32 *ProbPrices)
bitCount++;
}
}
- ProbPrices[i >> kNumMoveReducingBits] = ((kNumBitModelTotalBits << kCyclesBits) - 15 - bitCount);
+ ProbPrices[i] = (CProbPrice)((kNumBitModelTotalBits << kCyclesBits) - 15 - bitCount);
+ // printf("\n%3d: %5d", i, ProbPrices[i]);
}
}
#define GET_PRICE(prob, symbol) \
- p->ProbPrices[((prob) ^ (((-(int)(symbol))) & (kBitModelTotal - 1))) >> kNumMoveReducingBits];
+ p->ProbPrices[((prob) ^ (unsigned)(((-(int)(symbol))) & (kBitModelTotal - 1))) >> kNumMoveReducingBits];
#define GET_PRICEa(prob, symbol) \
- ProbPrices[((prob) ^ ((-((int)(symbol))) & (kBitModelTotal - 1))) >> kNumMoveReducingBits];
+ ProbPrices[((prob) ^ (unsigned)((-((int)(symbol))) & (kBitModelTotal - 1))) >> kNumMoveReducingBits];
#define GET_PRICE_0(prob) p->ProbPrices[(prob) >> kNumMoveReducingBits]
#define GET_PRICE_1(prob) p->ProbPrices[((prob) ^ (kBitModelTotal - 1)) >> kNumMoveReducingBits]
-#define GET_PRICE_0a(prob) ProbPrices[(prob) >> kNumMoveReducingBits]
-#define GET_PRICE_1a(prob) ProbPrices[((prob) ^ (kBitModelTotal - 1)) >> kNumMoveReducingBits]
+#define GET_PRICEa_0(prob) ProbPrices[(prob) >> kNumMoveReducingBits]
+#define GET_PRICEa_1(prob) ProbPrices[((prob) ^ (kBitModelTotal - 1)) >> kNumMoveReducingBits]
+
-static UInt32 LitEnc_GetPrice(const CLzmaProb *probs, UInt32 symbol, const UInt32 *ProbPrices)
+static UInt32 LitEnc_GetPrice(const CLzmaProb *probs, UInt32 symbol, const CProbPrice *ProbPrices)
{
UInt32 price = 0;
symbol |= 0x100;
do
{
- price += GET_PRICEa(probs[symbol >> 8], (symbol >> 7) & 1);
- symbol <<= 1;
+ unsigned bit = symbol & 1;
+ symbol >>= 1;
+ price += GET_PRICEa(probs[symbol], bit);
}
- while (symbol < 0x10000);
+ while (symbol >= 2);
return price;
}
-static UInt32 LitEnc_GetPriceMatched(const CLzmaProb *probs, UInt32 symbol, UInt32 matchByte, const UInt32 *ProbPrices)
+
+static UInt32 LitEnc_Matched_GetPrice(const CLzmaProb *probs, UInt32 symbol, UInt32 matchByte, const CProbPrice *ProbPrices)
{
UInt32 price = 0;
UInt32 offs = 0x100;
@@ -713,520 +784,525 @@ static UInt32 LitEnc_GetPriceMatched(const CLzmaProb *probs, UInt32 symbol, UInt
}
-static void RcTree_Encode(CRangeEnc *rc, CLzmaProb *probs, int numBitLevels, UInt32 symbol)
+static void RcTree_ReverseEncode(CRangeEnc *rc, CLzmaProb *probs, unsigned numBits, UInt32 symbol)
{
- UInt32 m = 1;
- int i;
- for (i = numBitLevels; i != 0;)
- {
- UInt32 bit;
- i--;
- bit = (symbol >> i) & 1;
- RangeEnc_EncodeBit(rc, probs + m, bit);
- m = (m << 1) | bit;
- }
-}
-
-static void RcTree_ReverseEncode(CRangeEnc *rc, CLzmaProb *probs, int numBitLevels, UInt32 symbol)
-{
- UInt32 m = 1;
- int i;
- for (i = 0; i < numBitLevels; i++)
- {
- UInt32 bit = symbol & 1;
- RangeEnc_EncodeBit(rc, probs + m, bit);
- m = (m << 1) | bit;
- symbol >>= 1;
- }
-}
-
-static UInt32 RcTree_GetPrice(const CLzmaProb *probs, int numBitLevels, UInt32 symbol, const UInt32 *ProbPrices)
-{
- UInt32 price = 0;
- symbol |= (1 << numBitLevels);
- while (symbol != 1)
- {
- price += GET_PRICEa(probs[symbol >> 1], symbol & 1);
- symbol >>= 1;
- }
- return price;
-}
-
-static UInt32 RcTree_ReverseGetPrice(const CLzmaProb *probs, int numBitLevels, UInt32 symbol, const UInt32 *ProbPrices)
-{
- UInt32 price = 0;
- UInt32 m = 1;
- int i;
- for (i = numBitLevels; i != 0; i--)
+ UInt32 range = rc->range;
+ unsigned m = 1;
+ do
{
- UInt32 bit = symbol & 1;
+ UInt32 ttt, newBound;
+ unsigned bit = symbol & 1;
+ // RangeEnc_EncodeBit(rc, probs + m, bit);
symbol >>= 1;
- price += GET_PRICEa(probs[m], bit);
+ RC_BIT(rc, probs + m, bit);
m = (m << 1) | bit;
}
- return price;
+ while (--numBits);
+ rc->range = range;
}
+
static void LenEnc_Init(CLenEnc *p)
{
unsigned i;
- p->choice = p->choice2 = kProbInitValue;
- for (i = 0; i < (LZMA_NUM_PB_STATES_MAX << kLenNumLowBits); i++)
+ for (i = 0; i < (LZMA_NUM_PB_STATES_MAX << (kLenNumLowBits + 1)); i++)
p->low[i] = kProbInitValue;
- for (i = 0; i < (LZMA_NUM_PB_STATES_MAX << kLenNumMidBits); i++)
- p->mid[i] = kProbInitValue;
for (i = 0; i < kLenNumHighSymbols; i++)
p->high[i] = kProbInitValue;
}
-static void LenEnc_Encode(CLenEnc *p, CRangeEnc *rc, UInt32 symbol, UInt32 posState)
+static void LenEnc_Encode(CLenEnc *p, CRangeEnc *rc, unsigned symbol, unsigned posState)
{
- if (symbol < kLenNumLowSymbols)
+ UInt32 range, ttt, newBound;
+ CLzmaProb *probs = p->low;
+ range = rc->range;
+ RC_BIT_PRE(rc, probs);
+ if (symbol >= kLenNumLowSymbols)
{
- RangeEnc_EncodeBit(rc, &p->choice, 0);
- RcTree_Encode(rc, p->low + (posState << kLenNumLowBits), kLenNumLowBits, symbol);
- }
- else
- {
- RangeEnc_EncodeBit(rc, &p->choice, 1);
- if (symbol < kLenNumLowSymbols + kLenNumMidSymbols)
+ RC_BIT_1(rc, probs);
+ probs += kLenNumLowSymbols;
+ RC_BIT_PRE(rc, probs);
+ if (symbol >= kLenNumLowSymbols * 2)
{
- RangeEnc_EncodeBit(rc, &p->choice2, 0);
- RcTree_Encode(rc, p->mid + (posState << kLenNumMidBits), kLenNumMidBits, symbol - kLenNumLowSymbols);
- }
- else
- {
- RangeEnc_EncodeBit(rc, &p->choice2, 1);
- RcTree_Encode(rc, p->high, kLenNumHighBits, symbol - kLenNumLowSymbols - kLenNumMidSymbols);
+ RC_BIT_1(rc, probs);
+ rc->range = range;
+ // RcTree_Encode(rc, p->high, kLenNumHighBits, symbol - kLenNumLowSymbols * 2);
+ LitEnc_Encode(rc, p->high, symbol - kLenNumLowSymbols * 2);
+ return;
}
+ symbol -= kLenNumLowSymbols;
}
-}
-static void LenEnc_SetPrices(CLenEnc *p, UInt32 posState, UInt32 numSymbols, UInt32 *prices, const UInt32 *ProbPrices)
-{
- UInt32 a0 = GET_PRICE_0a(p->choice);
- UInt32 a1 = GET_PRICE_1a(p->choice);
- UInt32 b0 = a1 + GET_PRICE_0a(p->choice2);
- UInt32 b1 = a1 + GET_PRICE_1a(p->choice2);
- UInt32 i = 0;
- for (i = 0; i < kLenNumLowSymbols; i++)
+ // RcTree_Encode(rc, probs + (posState << kLenNumLowBits), kLenNumLowBits, symbol);
{
- if (i >= numSymbols)
- return;
- prices[i] = a0 + RcTree_GetPrice(p->low + (posState << kLenNumLowBits), kLenNumLowBits, i, ProbPrices);
+ unsigned m;
+ unsigned bit;
+ RC_BIT_0(rc, probs);
+ probs += (posState << (1 + kLenNumLowBits));
+ bit = (symbol >> 2) ; RC_BIT(rc, probs + 1, bit); m = (1 << 1) + bit;
+ bit = (symbol >> 1) & 1; RC_BIT(rc, probs + m, bit); m = (m << 1) + bit;
+ bit = symbol & 1; RC_BIT(rc, probs + m, bit);
+ rc->range = range;
}
- for (; i < kLenNumLowSymbols + kLenNumMidSymbols; i++)
- {
- if (i >= numSymbols)
- return;
- prices[i] = b0 + RcTree_GetPrice(p->mid + (posState << kLenNumMidBits), kLenNumMidBits, i - kLenNumLowSymbols, ProbPrices);
- }
- for (; i < numSymbols; i++)
- prices[i] = b1 + RcTree_GetPrice(p->high, kLenNumHighBits, i - kLenNumLowSymbols - kLenNumMidSymbols, ProbPrices);
}
-static void MY_FAST_CALL LenPriceEnc_UpdateTable(CLenPriceEnc *p, UInt32 posState, const UInt32 *ProbPrices)
+static void SetPrices_3(const CLzmaProb *probs, UInt32 startPrice, UInt32 *prices, const CProbPrice *ProbPrices)
{
- LenEnc_SetPrices(&p->p, posState, p->tableSize, p->prices[posState], ProbPrices);
- p->counters[posState] = p->tableSize;
+ unsigned i;
+ for (i = 0; i < 8; i += 2)
+ {
+ UInt32 price = startPrice;
+ UInt32 prob;
+ price += GET_PRICEa(probs[1 ], (i >> 2));
+ price += GET_PRICEa(probs[2 + (i >> 2)], (i >> 1) & 1);
+ prob = probs[4 + (i >> 1)];
+ prices[i ] = price + GET_PRICEa_0(prob);
+ prices[i + 1] = price + GET_PRICEa_1(prob);
+ }
}
-static void LenPriceEnc_UpdateTables(CLenPriceEnc *p, UInt32 numPosStates, const UInt32 *ProbPrices)
+
+MY_NO_INLINE static void MY_FAST_CALL LenPriceEnc_UpdateTable(
+ CLenPriceEnc *p, unsigned posState,
+ const CLenEnc *enc,
+ const CProbPrice *ProbPrices)
{
- UInt32 posState;
- for (posState = 0; posState < numPosStates; posState++)
- LenPriceEnc_UpdateTable(p, posState, ProbPrices);
+ // int y; for (y = 0; y < 100; y++) {
+ UInt32 a;
+ unsigned i, numSymbols;
+
+ UInt32 *prices = p->prices[posState];
+ {
+ const CLzmaProb *probs = enc->low + (posState << (1 + kLenNumLowBits));
+ SetPrices_3(probs, GET_PRICEa_0(enc->low[0]), prices, ProbPrices);
+ a = GET_PRICEa_1(enc->low[0]);
+ SetPrices_3(probs + kLenNumLowSymbols, a + GET_PRICEa_0(enc->low[kLenNumLowSymbols]), prices + kLenNumLowSymbols, ProbPrices);
+ a += GET_PRICEa_1(enc->low[kLenNumLowSymbols]);
+ }
+ numSymbols = p->tableSize;
+ p->counters[posState] = numSymbols;
+ for (i = kLenNumLowSymbols * 2; i < numSymbols; i += 1)
+ {
+ prices[i] = a +
+ // RcTree_GetPrice(enc->high, kLenNumHighBits, i - kLenNumLowSymbols * 2, ProbPrices);
+ LitEnc_GetPrice(enc->high, i - kLenNumLowSymbols * 2, ProbPrices);
+ /*
+ unsigned sym = (i - kLenNumLowSymbols * 2) >> 1;
+ UInt32 price = a + RcTree_GetPrice(enc->high, kLenNumHighBits - 1, sym, ProbPrices);
+ UInt32 prob = enc->high[(1 << 7) + sym];
+ prices[i ] = price + GET_PRICEa_0(prob);
+ prices[i + 1] = price + GET_PRICEa_1(prob);
+ */
+ }
+ // }
}
-static void LenEnc_Encode2(CLenPriceEnc *p, CRangeEnc *rc, UInt32 symbol, UInt32 posState, Bool updatePrice, const UInt32 *ProbPrices)
+static void LenPriceEnc_UpdateTables(CLenPriceEnc *p, unsigned numPosStates,
+ const CLenEnc *enc,
+ const CProbPrice *ProbPrices)
{
- LenEnc_Encode(&p->p, rc, symbol, posState);
- if (updatePrice)
- if (--p->counters[posState] == 0)
- LenPriceEnc_UpdateTable(p, posState, ProbPrices);
+ unsigned posState;
+ for (posState = 0; posState < numPosStates; posState++)
+ LenPriceEnc_UpdateTable(p, posState, enc, ProbPrices);
}
-
-
-static void MovePos(CLzmaEnc *p, UInt32 num)
-{
+/*
#ifdef SHOW_STAT
g_STAT_OFFSET += num;
printf("\n MovePos %u", num);
#endif
+*/
- if (num != 0)
- {
- p->additionalOffset += num;
- p->matchFinder.Skip(p->matchFinderObj, num);
- }
-}
+#define MOVE_POS(p, num) { \
+ p->additionalOffset += (num); \
+ p->matchFinder.Skip(p->matchFinderObj, (num)); }
+
-static UInt32 ReadMatchDistances(CLzmaEnc *p, UInt32 *numDistancePairsRes)
+static unsigned ReadMatchDistances(CLzmaEnc *p, unsigned *numPairsRes)
{
- UInt32 lenRes = 0, numPairs;
+ unsigned numPairs;
+
+ p->additionalOffset++;
p->numAvail = p->matchFinder.GetNumAvailableBytes(p->matchFinderObj);
numPairs = p->matchFinder.GetMatches(p->matchFinderObj, p->matches);
+ *numPairsRes = numPairs;
#ifdef SHOW_STAT
printf("\n i = %u numPairs = %u ", g_STAT_OFFSET, numPairs / 2);
g_STAT_OFFSET++;
{
- UInt32 i;
+ unsigned i;
for (i = 0; i < numPairs; i += 2)
printf("%2u %6u | ", p->matches[i], p->matches[i + 1]);
}
#endif
- if (numPairs > 0)
+ if (numPairs == 0)
+ return 0;
{
- lenRes = p->matches[numPairs - 2];
- if (lenRes == p->numFastBytes)
+ unsigned len = p->matches[(size_t)numPairs - 2];
+ if (len != p->numFastBytes)
+ return len;
{
UInt32 numAvail = p->numAvail;
if (numAvail > LZMA_MATCH_LEN_MAX)
numAvail = LZMA_MATCH_LEN_MAX;
{
- const Byte *pbyCur = p->matchFinder.GetPointerToCurrentPos(p->matchFinderObj) - 1;
- const Byte *pby = pbyCur + lenRes;
- ptrdiff_t dif = (ptrdiff_t)-1 - p->matches[numPairs - 1];
- const Byte *pbyLim = pbyCur + numAvail;
- for (; pby != pbyLim && *pby == pby[dif]; pby++);
- lenRes = (UInt32)(pby - pbyCur);
+ const Byte *p1 = p->matchFinder.GetPointerToCurrentPos(p->matchFinderObj) - 1;
+ const Byte *p2 = p1 + len;
+ ptrdiff_t dif = (ptrdiff_t)-1 - p->matches[(size_t)numPairs - 1];
+ const Byte *lim = p1 + numAvail;
+ for (; p2 != lim && *p2 == p2[dif]; p2++);
+ return (unsigned)(p2 - p1);
}
}
}
- p->additionalOffset++;
- *numDistancePairsRes = numPairs;
- return lenRes;
}
+#define MARK_LIT ((UInt32)(Int32)-1)
-#define MakeAsChar(p) (p)->backPrev = (UInt32)(-1); (p)->prev1IsChar = False;
-#define MakeAsShortRep(p) (p)->backPrev = 0; (p)->prev1IsChar = False;
-#define IsShortRep(p) ((p)->backPrev == 0)
+#define MakeAs_Lit(p) { (p)->dist = MARK_LIT; (p)->extra = 0; }
+#define MakeAs_ShortRep(p) { (p)->dist = 0; (p)->extra = 0; }
+#define IsShortRep(p) ((p)->dist == 0)
-static UInt32 GetRepLen1Price(CLzmaEnc *p, UInt32 state, UInt32 posState)
-{
- return
- GET_PRICE_0(p->isRepG0[state]) +
- GET_PRICE_0(p->isRep0Long[state][posState]);
-}
-static UInt32 GetPureRepPrice(CLzmaEnc *p, UInt32 repIndex, UInt32 state, UInt32 posState)
+#define GetPrice_ShortRep(p, state, posState) \
+ ( GET_PRICE_0(p->isRepG0[state]) + GET_PRICE_0(p->isRep0Long[state][posState]))
+
+#define GetPrice_Rep_0(p, state, posState) ( \
+ GET_PRICE_1(p->isMatch[state][posState]) \
+ + GET_PRICE_1(p->isRep0Long[state][posState])) \
+ + GET_PRICE_1(p->isRep[state]) \
+ + GET_PRICE_0(p->isRepG0[state])
+
+
+static UInt32 GetPrice_PureRep(const CLzmaEnc *p, unsigned repIndex, size_t state, size_t posState)
{
UInt32 price;
+ UInt32 prob = p->isRepG0[state];
if (repIndex == 0)
{
- price = GET_PRICE_0(p->isRepG0[state]);
+ price = GET_PRICE_0(prob);
price += GET_PRICE_1(p->isRep0Long[state][posState]);
}
else
{
- price = GET_PRICE_1(p->isRepG0[state]);
+ price = GET_PRICE_1(prob);
+ prob = p->isRepG1[state];
if (repIndex == 1)
- price += GET_PRICE_0(p->isRepG1[state]);
+ price += GET_PRICE_0(prob);
else
{
- price += GET_PRICE_1(p->isRepG1[state]);
+ price += GET_PRICE_1(prob);
price += GET_PRICE(p->isRepG2[state], repIndex - 2);
}
}
return price;
}
-static UInt32 GetRepPrice(CLzmaEnc *p, UInt32 repIndex, UInt32 len, UInt32 state, UInt32 posState)
-{
- return p->repLenEnc.prices[posState][len - LZMA_MATCH_LEN_MIN] +
- GetPureRepPrice(p, repIndex, state, posState);
-}
-static UInt32 Backward(CLzmaEnc *p, UInt32 *backRes, UInt32 cur)
+static unsigned Backward(CLzmaEnc *p, unsigned cur)
{
- UInt32 posMem = p->opt[cur].posPrev;
- UInt32 backMem = p->opt[cur].backPrev;
- p->optimumEndIndex = cur;
- do
+ unsigned wr = cur + 1;
+ p->optEnd = wr;
+
+ for (;;)
{
- if (p->opt[cur].prev1IsChar)
+ UInt32 dist = p->opt[cur].dist;
+ UInt32 len = p->opt[cur].len;
+ UInt32 extra = p->opt[cur].extra;
+ cur -= len;
+
+ if (extra)
{
- MakeAsChar(&p->opt[posMem])
- p->opt[posMem].posPrev = posMem - 1;
- if (p->opt[cur].prev2)
+ wr--;
+ p->opt[wr].len = len;
+ cur -= extra;
+ len = extra;
+ if (extra == 1)
+ {
+ p->opt[wr].dist = dist;
+ dist = MARK_LIT;
+ }
+ else
{
- p->opt[posMem - 1].prev1IsChar = False;
- p->opt[posMem - 1].posPrev = p->opt[cur].posPrev2;
- p->opt[posMem - 1].backPrev = p->opt[cur].backPrev2;
+ p->opt[wr].dist = 0;
+ len--;
+ wr--;
+ p->opt[wr].dist = MARK_LIT;
+ p->opt[wr].len = 1;
}
}
+
+ if (cur == 0)
{
- UInt32 posPrev = posMem;
- UInt32 backCur = backMem;
-
- backMem = p->opt[posPrev].backPrev;
- posMem = p->opt[posPrev].posPrev;
-
- p->opt[posPrev].backPrev = backCur;
- p->opt[posPrev].posPrev = cur;
- cur = posPrev;
+ p->backRes = dist;
+ p->optCur = wr;
+ return len;
}
+
+ wr--;
+ p->opt[wr].dist = dist;
+ p->opt[wr].len = len;
}
- while (cur != 0);
- *backRes = p->opt[0].backPrev;
- p->optimumCurrentIndex = p->opt[0].posPrev;
- return p->optimumCurrentIndex;
}
-#define LIT_PROBS(pos, prevByte) (p->litProbs + ((((pos) & p->lpMask) << p->lc) + ((prevByte) >> (8 - p->lc))) * (UInt32)0x300)
-static UInt32 GetOptimum(CLzmaEnc *p, UInt32 position, UInt32 *backRes)
-{
- UInt32 lenEnd, cur;
- UInt32 reps[LZMA_NUM_REPS], repLens[LZMA_NUM_REPS];
- UInt32 *matches;
- {
+#define LIT_PROBS(pos, prevByte) \
+ (p->litProbs + (UInt32)3 * (((((pos) << 8) + (prevByte)) & p->lpMask) << p->lc))
- UInt32 numAvail, mainLen, numPairs, repMaxIndex, i, posState, len;
- UInt32 matchPrice, repMatchPrice, normalMatchPrice;
- const Byte *data;
- Byte curByte, matchByte;
- if (p->optimumEndIndex != p->optimumCurrentIndex)
- {
- const COptimal *opt = &p->opt[p->optimumCurrentIndex];
- UInt32 lenRes = opt->posPrev - p->optimumCurrentIndex;
- *backRes = opt->backPrev;
- p->optimumCurrentIndex = opt->posPrev;
- return lenRes;
- }
- p->optimumCurrentIndex = p->optimumEndIndex = 0;
-
- if (p->additionalOffset == 0)
- mainLen = ReadMatchDistances(p, &numPairs);
- else
- {
- mainLen = p->longestMatchLength;
- numPairs = p->numPairs;
- }
-
- numAvail = p->numAvail;
- if (numAvail < 2)
- {
- *backRes = (UInt32)(-1);
- return 1;
- }
- if (numAvail > LZMA_MATCH_LEN_MAX)
- numAvail = LZMA_MATCH_LEN_MAX;
+static unsigned GetOptimum(CLzmaEnc *p, UInt32 position)
+{
+ unsigned last, cur;
+ UInt32 reps[LZMA_NUM_REPS];
+ unsigned repLens[LZMA_NUM_REPS];
+ UInt32 *matches;
- data = p->matchFinder.GetPointerToCurrentPos(p->matchFinderObj) - 1;
- repMaxIndex = 0;
- for (i = 0; i < LZMA_NUM_REPS; i++)
{
- UInt32 lenTest;
- const Byte *data2;
- reps[i] = p->reps[i];
- data2 = data - reps[i] - 1;
- if (data[0] != data2[0] || data[1] != data2[1])
+ UInt32 numAvail;
+ unsigned numPairs, mainLen, repMaxIndex, i, posState;
+ UInt32 matchPrice, repMatchPrice;
+ const Byte *data;
+ Byte curByte, matchByte;
+
+ p->optCur = p->optEnd = 0;
+
+ if (p->additionalOffset == 0)
+ mainLen = ReadMatchDistances(p, &numPairs);
+ else
{
- repLens[i] = 0;
- continue;
+ mainLen = p->longestMatchLen;
+ numPairs = p->numPairs;
}
- for (lenTest = 2; lenTest < numAvail && data[lenTest] == data2[lenTest]; lenTest++);
- repLens[i] = lenTest;
- if (lenTest > repLens[repMaxIndex])
- repMaxIndex = i;
- }
- if (repLens[repMaxIndex] >= p->numFastBytes)
- {
- UInt32 lenRes;
- *backRes = repMaxIndex;
- lenRes = repLens[repMaxIndex];
- MovePos(p, lenRes - 1);
- return lenRes;
- }
-
- matches = p->matches;
- if (mainLen >= p->numFastBytes)
- {
- *backRes = matches[numPairs - 1] + LZMA_NUM_REPS;
- MovePos(p, mainLen - 1);
- return mainLen;
- }
- curByte = *data;
- matchByte = *(data - (reps[0] + 1));
-
- if (mainLen < 2 && curByte != matchByte && repLens[repMaxIndex] < 2)
- {
- *backRes = (UInt32)-1;
- return 1;
- }
-
- p->opt[0].state = (CState)p->state;
-
- posState = (position & p->pbMask);
-
- {
- const CLzmaProb *probs = LIT_PROBS(position, *(data - 1));
- p->opt[1].price = GET_PRICE_0(p->isMatch[p->state][posState]) +
- (!IsCharState(p->state) ?
- LitEnc_GetPriceMatched(probs, curByte, matchByte, p->ProbPrices) :
- LitEnc_GetPrice(probs, curByte, p->ProbPrices));
- }
-
- MakeAsChar(&p->opt[1]);
-
- matchPrice = GET_PRICE_1(p->isMatch[p->state][posState]);
- repMatchPrice = matchPrice + GET_PRICE_1(p->isRep[p->state]);
-
- if (matchByte == curByte)
- {
- UInt32 shortRepPrice = repMatchPrice + GetRepLen1Price(p, p->state, posState);
- if (shortRepPrice < p->opt[1].price)
+
+ numAvail = p->numAvail;
+ if (numAvail < 2)
{
- p->opt[1].price = shortRepPrice;
- MakeAsShortRep(&p->opt[1]);
+ p->backRes = MARK_LIT;
+ return 1;
}
- }
- lenEnd = ((mainLen >= repLens[repMaxIndex]) ? mainLen : repLens[repMaxIndex]);
-
- if (lenEnd < 2)
- {
- *backRes = p->opt[1].backPrev;
- return 1;
- }
-
- p->opt[1].posPrev = 0;
- for (i = 0; i < LZMA_NUM_REPS; i++)
- p->opt[0].backs[i] = reps[i];
-
- len = lenEnd;
- do
- p->opt[len--].price = kInfinityPrice;
- while (len >= 2);
-
- for (i = 0; i < LZMA_NUM_REPS; i++)
- {
- UInt32 repLen = repLens[i];
- UInt32 price;
- if (repLen < 2)
- continue;
- price = repMatchPrice + GetPureRepPrice(p, i, p->state, posState);
- do
+ if (numAvail > LZMA_MATCH_LEN_MAX)
+ numAvail = LZMA_MATCH_LEN_MAX;
+
+ data = p->matchFinder.GetPointerToCurrentPos(p->matchFinderObj) - 1;
+ repMaxIndex = 0;
+
+ for (i = 0; i < LZMA_NUM_REPS; i++)
{
- UInt32 curAndLenPrice = price + p->repLenEnc.prices[posState][repLen - 2];
- COptimal *opt = &p->opt[repLen];
- if (curAndLenPrice < opt->price)
+ unsigned len;
+ const Byte *data2;
+ reps[i] = p->reps[i];
+ data2 = data - reps[i];
+ if (data[0] != data2[0] || data[1] != data2[1])
{
- opt->price = curAndLenPrice;
- opt->posPrev = 0;
- opt->backPrev = i;
- opt->prev1IsChar = False;
+ repLens[i] = 0;
+ continue;
}
+ for (len = 2; len < numAvail && data[len] == data2[len]; len++);
+ repLens[i] = len;
+ if (len > repLens[repMaxIndex])
+ repMaxIndex = i;
}
- while (--repLen >= 2);
- }
-
- normalMatchPrice = matchPrice + GET_PRICE_0(p->isRep[p->state]);
-
- len = ((repLens[0] >= 2) ? repLens[0] + 1 : 2);
- if (len <= mainLen)
- {
- UInt32 offs = 0;
- while (len > matches[offs])
- offs += 2;
- for (; ; len++)
+
+ if (repLens[repMaxIndex] >= p->numFastBytes)
{
- COptimal *opt;
- UInt32 distance = matches[offs + 1];
-
- UInt32 curAndLenPrice = normalMatchPrice + p->lenEnc.prices[posState][len - LZMA_MATCH_LEN_MIN];
- UInt32 lenToPosState = GetLenToPosState(len);
- if (distance < kNumFullDistances)
- curAndLenPrice += p->distancesPrices[lenToPosState][distance];
- else
+ unsigned len;
+ p->backRes = repMaxIndex;
+ len = repLens[repMaxIndex];
+ MOVE_POS(p, len - 1)
+ return len;
+ }
+
+ matches = p->matches;
+
+ if (mainLen >= p->numFastBytes)
+ {
+ p->backRes = matches[(size_t)numPairs - 1] + LZMA_NUM_REPS;
+ MOVE_POS(p, mainLen - 1)
+ return mainLen;
+ }
+
+ curByte = *data;
+ matchByte = *(data - reps[0]);
+
+ if (mainLen < 2 && curByte != matchByte && repLens[repMaxIndex] < 2)
+ {
+ p->backRes = MARK_LIT;
+ return 1;
+ }
+
+ p->opt[0].state = (CState)p->state;
+
+ posState = (position & p->pbMask);
+
+ {
+ const CLzmaProb *probs = LIT_PROBS(position, *(data - 1));
+ p->opt[1].price = GET_PRICE_0(p->isMatch[p->state][posState]) +
+ (!IsLitState(p->state) ?
+ LitEnc_Matched_GetPrice(probs, curByte, matchByte, p->ProbPrices) :
+ LitEnc_GetPrice(probs, curByte, p->ProbPrices));
+ }
+
+ MakeAs_Lit(&p->opt[1]);
+
+ matchPrice = GET_PRICE_1(p->isMatch[p->state][posState]);
+ repMatchPrice = matchPrice + GET_PRICE_1(p->isRep[p->state]);
+
+ if (matchByte == curByte)
+ {
+ UInt32 shortRepPrice = repMatchPrice + GetPrice_ShortRep(p, p->state, posState);
+ if (shortRepPrice < p->opt[1].price)
{
- UInt32 slot;
- GetPosSlot2(distance, slot);
- curAndLenPrice += p->alignPrices[distance & kAlignMask] + p->posSlotPrices[lenToPosState][slot];
+ p->opt[1].price = shortRepPrice;
+ MakeAs_ShortRep(&p->opt[1]);
}
- opt = &p->opt[len];
- if (curAndLenPrice < opt->price)
+ }
+
+ last = (mainLen >= repLens[repMaxIndex] ? mainLen : repLens[repMaxIndex]);
+
+ if (last < 2)
+ {
+ p->backRes = p->opt[1].dist;
+ return 1;
+ }
+
+ p->opt[1].len = 1;
+
+ p->opt[0].reps[0] = reps[0];
+ p->opt[0].reps[1] = reps[1];
+ p->opt[0].reps[2] = reps[2];
+ p->opt[0].reps[3] = reps[3];
+
+ {
+ unsigned len = last;
+ do
+ p->opt[len--].price = kInfinityPrice;
+ while (len >= 2);
+ }
+
+ // ---------- REP ----------
+
+ for (i = 0; i < LZMA_NUM_REPS; i++)
+ {
+ unsigned repLen = repLens[i];
+ UInt32 price;
+ if (repLen < 2)
+ continue;
+ price = repMatchPrice + GetPrice_PureRep(p, i, p->state, posState);
+ do
{
- opt->price = curAndLenPrice;
- opt->posPrev = 0;
- opt->backPrev = distance + LZMA_NUM_REPS;
- opt->prev1IsChar = False;
+ UInt32 price2 = price + p->repLenEnc.prices[posState][(size_t)repLen - 2];
+ COptimal *opt = &p->opt[repLen];
+ if (price2 < opt->price)
+ {
+ opt->price = price2;
+ opt->len = repLen;
+ opt->dist = i;
+ opt->extra = 0;
+ }
}
- if (len == matches[offs])
+ while (--repLen >= 2);
+ }
+
+
+ // ---------- MATCH ----------
+ {
+ unsigned len = ((repLens[0] >= 2) ? repLens[0] + 1 : 2);
+ if (len <= mainLen)
{
- offs += 2;
- if (offs == numPairs)
- break;
+ unsigned offs = 0;
+ UInt32 normalMatchPrice = matchPrice + GET_PRICE_0(p->isRep[p->state]);
+
+ while (len > matches[offs])
+ offs += 2;
+
+ for (; ; len++)
+ {
+ COptimal *opt;
+ UInt32 dist = matches[(size_t)offs + 1];
+ UInt32 price2 = normalMatchPrice + p->lenEnc.prices[posState][(size_t)len - LZMA_MATCH_LEN_MIN];
+ unsigned lenToPosState = GetLenToPosState(len);
+
+ if (dist < kNumFullDistances)
+ price2 += p->distancesPrices[lenToPosState][dist & (kNumFullDistances - 1)];
+ else
+ {
+ unsigned slot;
+ GetPosSlot2(dist, slot);
+ price2 += p->alignPrices[dist & kAlignMask];
+ price2 += p->posSlotPrices[lenToPosState][slot];
+ }
+
+ opt = &p->opt[len];
+
+ if (price2 < opt->price)
+ {
+ opt->price = price2;
+ opt->len = len;
+ opt->dist = dist + LZMA_NUM_REPS;
+ opt->extra = 0;
+ }
+
+ if (len == matches[offs])
+ {
+ offs += 2;
+ if (offs == numPairs)
+ break;
+ }
+ }
}
}
- }
+
- cur = 0;
+ cur = 0;
#ifdef SHOW_STAT2
/* if (position >= 0) */
{
unsigned i;
printf("\n pos = %4X", position);
- for (i = cur; i <= lenEnd; i++)
+ for (i = cur; i <= last; i++)
printf("\nprice[%4X] = %u", position - cur + i, p->opt[i].price);
}
#endif
+ }
+
- }
+
+ // ---------- Optimal Parsing ----------
for (;;)
{
- UInt32 numAvail;
- UInt32 numAvailFull, newLen, numPairs, posPrev, state, posState, startLen;
- UInt32 curPrice, curAnd1Price, matchPrice, repMatchPrice;
- Bool nextIsChar;
+ UInt32 numAvail, numAvailFull;
+ unsigned newLen, numPairs, prev, state, posState, startLen;
+ UInt32 curPrice, litPrice, matchPrice, repMatchPrice;
+ Bool nextIsLit;
Byte curByte, matchByte;
const Byte *data;
- COptimal *curOpt;
- COptimal *nextOpt;
+ COptimal *curOpt, *nextOpt;
- cur++;
- if (cur == lenEnd)
- return Backward(p, backRes, cur);
+ if (++cur == last)
+ return Backward(p, cur);
newLen = ReadMatchDistances(p, &numPairs);
+
if (newLen >= p->numFastBytes)
{
p->numPairs = numPairs;
- p->longestMatchLength = newLen;
- return Backward(p, backRes, cur);
+ p->longestMatchLen = newLen;
+ return Backward(p, cur);
}
- position++;
+
curOpt = &p->opt[cur];
- posPrev = curOpt->posPrev;
- if (curOpt->prev1IsChar)
- {
- posPrev--;
- if (curOpt->prev2)
- {
- state = p->opt[curOpt->posPrev2].state;
- if (curOpt->backPrev2 < LZMA_NUM_REPS)
- state = kRepNextStates[state];
- else
- state = kMatchNextStates[state];
- }
- else
- state = p->opt[posPrev].state;
- state = kLiteralNextStates[state];
- }
- else
- state = p->opt[posPrev].state;
- if (posPrev == cur - 1)
+ prev = cur - curOpt->len;
+
+ if (curOpt->len == 1)
{
+ state = p->opt[prev].state;
if (IsShortRep(curOpt))
state = kShortRepNextStates[state];
else
@@ -1234,92 +1310,136 @@ static UInt32 GetOptimum(CLzmaEnc *p, UInt32 position, UInt32 *backRes)
}
else
{
- UInt32 pos;
const COptimal *prevOpt;
- if (curOpt->prev1IsChar && curOpt->prev2)
+ UInt32 b0;
+ UInt32 dist = curOpt->dist;
+
+ if (curOpt->extra)
{
- posPrev = curOpt->posPrev2;
- pos = curOpt->backPrev2;
- state = kRepNextStates[state];
+ prev -= curOpt->extra;
+ state = kState_RepAfterLit;
+ if (curOpt->extra == 1)
+ state = (dist < LZMA_NUM_REPS) ? kState_RepAfterLit : kState_MatchAfterLit;
}
else
{
- pos = curOpt->backPrev;
- if (pos < LZMA_NUM_REPS)
+ state = p->opt[prev].state;
+ if (dist < LZMA_NUM_REPS)
state = kRepNextStates[state];
else
state = kMatchNextStates[state];
}
- prevOpt = &p->opt[posPrev];
- if (pos < LZMA_NUM_REPS)
+
+ prevOpt = &p->opt[prev];
+ b0 = prevOpt->reps[0];
+
+ if (dist < LZMA_NUM_REPS)
{
- UInt32 i;
- reps[0] = prevOpt->backs[pos];
- for (i = 1; i <= pos; i++)
- reps[i] = prevOpt->backs[i - 1];
- for (; i < LZMA_NUM_REPS; i++)
- reps[i] = prevOpt->backs[i];
+ if (dist == 0)
+ {
+ reps[0] = b0;
+ reps[1] = prevOpt->reps[1];
+ reps[2] = prevOpt->reps[2];
+ reps[3] = prevOpt->reps[3];
+ }
+ else
+ {
+ reps[1] = b0;
+ b0 = prevOpt->reps[1];
+ if (dist == 1)
+ {
+ reps[0] = b0;
+ reps[2] = prevOpt->reps[2];
+ reps[3] = prevOpt->reps[3];
+ }
+ else
+ {
+ reps[2] = b0;
+ reps[0] = prevOpt->reps[dist];
+ reps[3] = prevOpt->reps[dist ^ 1];
+ }
+ }
}
else
{
- UInt32 i;
- reps[0] = (pos - LZMA_NUM_REPS);
- for (i = 1; i < LZMA_NUM_REPS; i++)
- reps[i] = prevOpt->backs[i - 1];
+ reps[0] = (dist - LZMA_NUM_REPS + 1);
+ reps[1] = b0;
+ reps[2] = prevOpt->reps[1];
+ reps[3] = prevOpt->reps[2];
}
}
+
curOpt->state = (CState)state;
+ curOpt->reps[0] = reps[0];
+ curOpt->reps[1] = reps[1];
+ curOpt->reps[2] = reps[2];
+ curOpt->reps[3] = reps[3];
- curOpt->backs[0] = reps[0];
- curOpt->backs[1] = reps[1];
- curOpt->backs[2] = reps[2];
- curOpt->backs[3] = reps[3];
-
- curPrice = curOpt->price;
- nextIsChar = False;
data = p->matchFinder.GetPointerToCurrentPos(p->matchFinderObj) - 1;
curByte = *data;
- matchByte = *(data - (reps[0] + 1));
+ matchByte = *(data - reps[0]);
+ position++;
posState = (position & p->pbMask);
- curAnd1Price = curPrice + GET_PRICE_0(p->isMatch[state][posState]);
- {
- const CLzmaProb *probs = LIT_PROBS(position, *(data - 1));
- curAnd1Price +=
- (!IsCharState(state) ?
- LitEnc_GetPriceMatched(probs, curByte, matchByte, p->ProbPrices) :
- LitEnc_GetPrice(probs, curByte, p->ProbPrices));
- }
+ /*
+ The order of Price checks:
+ < LIT
+ <= SHORT_REP
+ < LIT : REP_0
+ < REP [ : LIT : REP_0 ]
+ < MATCH [ : LIT : REP_0 ]
+ */
+
+ curPrice = curOpt->price;
+ litPrice = curPrice + GET_PRICE_0(p->isMatch[state][posState]);
- nextOpt = &p->opt[cur + 1];
+ nextOpt = &p->opt[(size_t)cur + 1];
+ nextIsLit = False;
- if (curAnd1Price < nextOpt->price)
+ // if (litPrice >= nextOpt->price) litPrice = 0; else // 18.new
{
- nextOpt->price = curAnd1Price;
- nextOpt->posPrev = cur;
- MakeAsChar(nextOpt);
- nextIsChar = True;
+ const CLzmaProb *probs = LIT_PROBS(position, *(data - 1));
+ litPrice += (!IsLitState(state) ?
+ LitEnc_Matched_GetPrice(probs, curByte, matchByte, p->ProbPrices) :
+ LitEnc_GetPrice(probs, curByte, p->ProbPrices));
+
+ if (litPrice < nextOpt->price)
+ {
+ nextOpt->price = litPrice;
+ nextOpt->len = 1;
+ MakeAs_Lit(nextOpt);
+ nextIsLit = True;
+ }
}
matchPrice = curPrice + GET_PRICE_1(p->isMatch[state][posState]);
repMatchPrice = matchPrice + GET_PRICE_1(p->isRep[state]);
- if (matchByte == curByte && !(nextOpt->posPrev < cur && nextOpt->backPrev == 0))
+ // ---------- SHORT_REP ----------
+ // if (IsLitState(state)) // 18.new
+ if (matchByte == curByte)
+ // if (repMatchPrice < nextOpt->price) // 18.new
+ if (nextOpt->len < 2
+ || (nextOpt->dist != 0
+ && nextOpt->extra <= 1 // 17.old
+ ))
{
- UInt32 shortRepPrice = repMatchPrice + GetRepLen1Price(p, state, posState);
- if (shortRepPrice <= nextOpt->price)
+ UInt32 shortRepPrice = repMatchPrice + GetPrice_ShortRep(p, state, posState);
+ if (shortRepPrice <= nextOpt->price) // 17.old
+ // if (shortRepPrice < nextOpt->price) // 18.new
{
nextOpt->price = shortRepPrice;
- nextOpt->posPrev = cur;
- MakeAsShortRep(nextOpt);
- nextIsChar = True;
+ nextOpt->len = 1;
+ MakeAs_ShortRep(nextOpt);
+ nextIsLit = False;
}
}
+
numAvailFull = p->numAvail;
{
UInt32 temp = kNumOpts - 1 - cur;
- if (temp < numAvailFull)
+ if (numAvailFull > temp)
numAvailFull = temp;
}
@@ -1327,41 +1447,53 @@ static UInt32 GetOptimum(CLzmaEnc *p, UInt32 position, UInt32 *backRes)
continue;
numAvail = (numAvailFull <= p->numFastBytes ? numAvailFull : p->numFastBytes);
- if (!nextIsChar && matchByte != curByte) /* speed optimization */
+ // numAvail <= p->numFastBytes
+
+ // ---------- LIT : REP_0 ----------
+
+ if (
+ // litPrice != 0 && // 18.new
+ !nextIsLit
+ && matchByte != curByte
+ && numAvailFull > 2)
{
- /* try Literal + rep0 */
- UInt32 temp;
- UInt32 lenTest2;
- const Byte *data2 = data - reps[0] - 1;
- UInt32 limit = p->numFastBytes + 1;
- if (limit > numAvailFull)
- limit = numAvailFull;
-
- for (temp = 1; temp < limit && data[temp] == data2[temp]; temp++);
- lenTest2 = temp - 1;
- if (lenTest2 >= 2)
+ const Byte *data2 = data - reps[0];
+ if (data[1] == data2[1] && data[2] == data2[2])
{
- UInt32 state2 = kLiteralNextStates[state];
- UInt32 posStateNext = (position + 1) & p->pbMask;
- UInt32 nextRepMatchPrice = curAnd1Price +
- GET_PRICE_1(p->isMatch[state2][posStateNext]) +
- GET_PRICE_1(p->isRep[state2]);
- /* for (; lenTest2 >= 2; lenTest2--) */
+ unsigned len;
+ unsigned limit = p->numFastBytes + 1;
+ if (limit > numAvailFull)
+ limit = numAvailFull;
+ for (len = 3; len < limit && data[len] == data2[len]; len++);
+
{
- UInt32 curAndLenPrice;
- COptimal *opt;
- UInt32 offset = cur + 1 + lenTest2;
- while (lenEnd < offset)
- p->opt[++lenEnd].price = kInfinityPrice;
- curAndLenPrice = nextRepMatchPrice + GetRepPrice(p, 0, lenTest2, state2, posStateNext);
- opt = &p->opt[offset];
- if (curAndLenPrice < opt->price)
+ unsigned state2 = kLiteralNextStates[state];
+ unsigned posState2 = (position + 1) & p->pbMask;
+ UInt32 price = litPrice + GetPrice_Rep_0(p, state2, posState2);
{
- opt->price = curAndLenPrice;
- opt->posPrev = cur + 1;
- opt->backPrev = 0;
- opt->prev1IsChar = True;
- opt->prev2 = False;
+ unsigned offset = cur + len;
+ while (last < offset)
+ p->opt[++last].price = kInfinityPrice;
+
+ // do
+ {
+ UInt32 price2;
+ COptimal *opt;
+ len--;
+ // price2 = price + GetPrice_Len_Rep_0(p, len, state2, posState2);
+ price2 = price + p->repLenEnc.prices[posState2][len - LZMA_MATCH_LEN_MIN];
+
+ opt = &p->opt[offset];
+ // offset--;
+ if (price2 < opt->price)
+ {
+ opt->price = price2;
+ opt->len = len;
+ opt->dist = 0;
+ opt->extra = 1;
+ }
+ }
+ // while (len >= 3);
}
}
}
@@ -1369,87 +1501,105 @@ static UInt32 GetOptimum(CLzmaEnc *p, UInt32 position, UInt32 *backRes)
startLen = 2; /* speed optimization */
{
- UInt32 repIndex;
- for (repIndex = 0; repIndex < LZMA_NUM_REPS; repIndex++)
- {
- UInt32 lenTest;
- UInt32 lenTestTemp;
- UInt32 price;
- const Byte *data2 = data - reps[repIndex] - 1;
- if (data[0] != data2[0] || data[1] != data2[1])
- continue;
- for (lenTest = 2; lenTest < numAvail && data[lenTest] == data2[lenTest]; lenTest++);
- while (lenEnd < cur + lenTest)
- p->opt[++lenEnd].price = kInfinityPrice;
- lenTestTemp = lenTest;
- price = repMatchPrice + GetPureRepPrice(p, repIndex, state, posState);
- do
+ // ---------- REP ----------
+ unsigned repIndex = 0; // 17.old
+ // unsigned repIndex = IsLitState(state) ? 0 : 1; // 18.notused
+ for (; repIndex < LZMA_NUM_REPS; repIndex++)
{
- UInt32 curAndLenPrice = price + p->repLenEnc.prices[posState][lenTest - 2];
- COptimal *opt = &p->opt[cur + lenTest];
- if (curAndLenPrice < opt->price)
+ unsigned len;
+ UInt32 price;
+ const Byte *data2 = data - reps[repIndex];
+ if (data[0] != data2[0] || data[1] != data2[1])
+ continue;
+
+ for (len = 2; len < numAvail && data[len] == data2[len]; len++);
+
+ // if (len < startLen) continue; // 18.new: speed optimization
+
+ while (last < cur + len)
+ p->opt[++last].price = kInfinityPrice;
{
- opt->price = curAndLenPrice;
- opt->posPrev = cur;
- opt->backPrev = repIndex;
- opt->prev1IsChar = False;
+ unsigned len2 = len;
+ price = repMatchPrice + GetPrice_PureRep(p, repIndex, state, posState);
+ do
+ {
+ UInt32 price2 = price + p->repLenEnc.prices[posState][(size_t)len2 - 2];
+ COptimal *opt = &p->opt[cur + len2];
+ if (price2 < opt->price)
+ {
+ opt->price = price2;
+ opt->len = len2;
+ opt->dist = repIndex;
+ opt->extra = 0;
+ }
+ }
+ while (--len2 >= 2);
}
- }
- while (--lenTest >= 2);
- lenTest = lenTestTemp;
-
- if (repIndex == 0)
- startLen = lenTest + 1;
- /* if (_maxMode) */
+ if (repIndex == 0) startLen = len + 1; // 17.old
+ // startLen = len + 1; // 18.new
+
+ /* if (_maxMode) */
{
- UInt32 lenTest2 = lenTest + 1;
- UInt32 limit = lenTest2 + p->numFastBytes;
+ // ---------- REP : LIT : REP_0 ----------
+ // numFastBytes + 1 + numFastBytes
+
+ unsigned len2 = len + 1;
+ unsigned limit = len2 + p->numFastBytes;
if (limit > numAvailFull)
limit = numAvailFull;
- for (; lenTest2 < limit && data[lenTest2] == data2[lenTest2]; lenTest2++);
- lenTest2 -= lenTest + 1;
- if (lenTest2 >= 2)
+
+ for (; len2 < limit && data[len2] == data2[len2]; len2++);
+
+ len2 -= len;
+ if (len2 >= 3)
{
- UInt32 nextRepMatchPrice;
- UInt32 state2 = kRepNextStates[state];
- UInt32 posStateNext = (position + lenTest) & p->pbMask;
- UInt32 curAndLenCharPrice =
- price + p->repLenEnc.prices[posState][lenTest - 2] +
- GET_PRICE_0(p->isMatch[state2][posStateNext]) +
- LitEnc_GetPriceMatched(LIT_PROBS(position + lenTest, data[lenTest - 1]),
- data[lenTest], data2[lenTest], p->ProbPrices);
- state2 = kLiteralNextStates[state2];
- posStateNext = (position + lenTest + 1) & p->pbMask;
- nextRepMatchPrice = curAndLenCharPrice +
- GET_PRICE_1(p->isMatch[state2][posStateNext]) +
- GET_PRICE_1(p->isRep[state2]);
+ unsigned state2 = kRepNextStates[state];
+ unsigned posState2 = (position + len) & p->pbMask;
+ price +=
+ p->repLenEnc.prices[posState][(size_t)len - 2]
+ + GET_PRICE_0(p->isMatch[state2][posState2])
+ + LitEnc_Matched_GetPrice(LIT_PROBS(position + len, data[(size_t)len - 1]),
+ data[len], data2[len], p->ProbPrices);
- /* for (; lenTest2 >= 2; lenTest2--) */
+ // state2 = kLiteralNextStates[state2];
+ state2 = kState_LitAfterRep;
+ posState2 = (posState2 + 1) & p->pbMask;
+
+
+ price += GetPrice_Rep_0(p, state2, posState2);
{
- UInt32 curAndLenPrice;
- COptimal *opt;
- UInt32 offset = cur + lenTest + 1 + lenTest2;
- while (lenEnd < offset)
- p->opt[++lenEnd].price = kInfinityPrice;
- curAndLenPrice = nextRepMatchPrice + GetRepPrice(p, 0, lenTest2, state2, posStateNext);
- opt = &p->opt[offset];
- if (curAndLenPrice < opt->price)
+ unsigned offset = cur + len + len2;
+ while (last < offset)
+ p->opt[++last].price = kInfinityPrice;
+ // do
{
- opt->price = curAndLenPrice;
- opt->posPrev = cur + lenTest + 1;
- opt->backPrev = 0;
- opt->prev1IsChar = True;
- opt->prev2 = True;
- opt->posPrev2 = cur;
- opt->backPrev2 = repIndex;
+ unsigned price2;
+ COptimal *opt;
+ len2--;
+ // price2 = price + GetPrice_Len_Rep_0(p, len2, state2, posState2);
+ price2 = price + p->repLenEnc.prices[posState2][len2 - LZMA_MATCH_LEN_MIN];
+
+ opt = &p->opt[offset];
+ // offset--;
+ if (price2 < opt->price)
+ {
+ opt->price = price2;
+ opt->len = len2;
+ opt->extra = (CExtra)(len + 1);
+ opt->dist = repIndex;
+ }
}
+ // while (len2 >= 3);
}
}
}
+ }
}
- }
- /* for (UInt32 lenTest = 2; lenTest <= newLen; lenTest++) */
+
+
+ // ---------- MATCH ----------
+ /* for (unsigned len = 2; len <= newLen; len++) */
if (newLen > numAvail)
{
newLen = numAvail;
@@ -1457,134 +1607,148 @@ static UInt32 GetOptimum(CLzmaEnc *p, UInt32 position, UInt32 *backRes)
matches[numPairs] = newLen;
numPairs += 2;
}
+
if (newLen >= startLen)
{
UInt32 normalMatchPrice = matchPrice + GET_PRICE_0(p->isRep[state]);
- UInt32 offs, curBack, posSlot;
- UInt32 lenTest;
- while (lenEnd < cur + newLen)
- p->opt[++lenEnd].price = kInfinityPrice;
+ UInt32 dist;
+ unsigned offs, posSlot, len;
+ while (last < cur + newLen)
+ p->opt[++last].price = kInfinityPrice;
offs = 0;
while (startLen > matches[offs])
offs += 2;
- curBack = matches[offs + 1];
- GetPosSlot2(curBack, posSlot);
- for (lenTest = /*2*/ startLen; ; lenTest++)
+ dist = matches[(size_t)offs + 1];
+
+ // if (dist >= kNumFullDistances)
+ GetPosSlot2(dist, posSlot);
+
+ for (len = /*2*/ startLen; ; len++)
{
- UInt32 curAndLenPrice = normalMatchPrice + p->lenEnc.prices[posState][lenTest - LZMA_MATCH_LEN_MIN];
- {
- UInt32 lenToPosState = GetLenToPosState(lenTest);
- COptimal *opt;
- if (curBack < kNumFullDistances)
- curAndLenPrice += p->distancesPrices[lenToPosState][curBack];
- else
- curAndLenPrice += p->posSlotPrices[lenToPosState][posSlot] + p->alignPrices[curBack & kAlignMask];
-
- opt = &p->opt[cur + lenTest];
- if (curAndLenPrice < opt->price)
+ UInt32 price = normalMatchPrice + p->lenEnc.prices[posState][(size_t)len - LZMA_MATCH_LEN_MIN];
{
- opt->price = curAndLenPrice;
- opt->posPrev = cur;
- opt->backPrev = curBack + LZMA_NUM_REPS;
- opt->prev1IsChar = False;
- }
+ COptimal *opt;
+ unsigned lenToPosState = len - 2; lenToPosState = GetLenToPosState2(lenToPosState);
+ if (dist < kNumFullDistances)
+ price += p->distancesPrices[lenToPosState][dist & (kNumFullDistances - 1)];
+ else
+ price += p->posSlotPrices[lenToPosState][posSlot] + p->alignPrices[dist & kAlignMask];
+
+ opt = &p->opt[cur + len];
+ if (price < opt->price)
+ {
+ opt->price = price;
+ opt->len = len;
+ opt->dist = dist + LZMA_NUM_REPS;
+ opt->extra = 0;
+ }
}
- if (/*_maxMode && */lenTest == matches[offs])
+ if (/*_maxMode && */ len == matches[offs])
{
- /* Try Match + Literal + Rep0 */
- const Byte *data2 = data - curBack - 1;
- UInt32 lenTest2 = lenTest + 1;
- UInt32 limit = lenTest2 + p->numFastBytes;
+ // MATCH : LIT : REP_0
+
+ const Byte *data2 = data - dist - 1;
+ unsigned len2 = len + 1;
+ unsigned limit = len2 + p->numFastBytes;
if (limit > numAvailFull)
limit = numAvailFull;
- for (; lenTest2 < limit && data[lenTest2] == data2[lenTest2]; lenTest2++);
- lenTest2 -= lenTest + 1;
- if (lenTest2 >= 2)
+
+ for (; len2 < limit && data[len2] == data2[len2]; len2++);
+
+ len2 -= len;
+
+ if (len2 >= 3)
{
- UInt32 nextRepMatchPrice;
- UInt32 state2 = kMatchNextStates[state];
- UInt32 posStateNext = (position + lenTest) & p->pbMask;
- UInt32 curAndLenCharPrice = curAndLenPrice +
- GET_PRICE_0(p->isMatch[state2][posStateNext]) +
- LitEnc_GetPriceMatched(LIT_PROBS(position + lenTest, data[lenTest - 1]),
- data[lenTest], data2[lenTest], p->ProbPrices);
- state2 = kLiteralNextStates[state2];
- posStateNext = (posStateNext + 1) & p->pbMask;
- nextRepMatchPrice = curAndLenCharPrice +
- GET_PRICE_1(p->isMatch[state2][posStateNext]) +
- GET_PRICE_1(p->isRep[state2]);
-
- /* for (; lenTest2 >= 2; lenTest2--) */
+ unsigned state2 = kMatchNextStates[state];
+ unsigned posState2 = (position + len) & p->pbMask;
+ unsigned offset;
+ price += GET_PRICE_0(p->isMatch[state2][posState2]);
+ price += LitEnc_Matched_GetPrice(LIT_PROBS(position + len, data[(size_t)len - 1]),
+ data[len], data2[len], p->ProbPrices);
+
+ // state2 = kLiteralNextStates[state2];
+ state2 = kState_LitAfterMatch;
+
+ posState2 = (posState2 + 1) & p->pbMask;
+ price += GetPrice_Rep_0(p, state2, posState2);
+
+ offset = cur + len + len2;
+ while (last < offset)
+ p->opt[++last].price = kInfinityPrice;
+ // do
{
- UInt32 offset = cur + lenTest + 1 + lenTest2;
- UInt32 curAndLenPrice2;
+ UInt32 price2;
COptimal *opt;
- while (lenEnd < offset)
- p->opt[++lenEnd].price = kInfinityPrice;
- curAndLenPrice2 = nextRepMatchPrice + GetRepPrice(p, 0, lenTest2, state2, posStateNext);
+ len2--;
+ // price2 = price + GetPrice_Len_Rep_0(p, len2, state2, posState2);
+ price2 = price + p->repLenEnc.prices[posState2][len2 - LZMA_MATCH_LEN_MIN];
opt = &p->opt[offset];
- if (curAndLenPrice2 < opt->price)
+ // offset--;
+ if (price2 < opt->price)
{
- opt->price = curAndLenPrice2;
- opt->posPrev = cur + lenTest + 1;
- opt->backPrev = 0;
- opt->prev1IsChar = True;
- opt->prev2 = True;
- opt->posPrev2 = cur;
- opt->backPrev2 = curBack + LZMA_NUM_REPS;
+ opt->price = price2;
+ opt->len = len2;
+ opt->extra = (CExtra)(len + 1);
+ opt->dist = dist + LZMA_NUM_REPS;
}
}
+ // while (len2 >= 3);
}
+
offs += 2;
if (offs == numPairs)
break;
- curBack = matches[offs + 1];
- if (curBack >= kNumFullDistances)
- GetPosSlot2(curBack, posSlot);
+ dist = matches[(size_t)offs + 1];
+ // if (dist >= kNumFullDistances)
+ GetPosSlot2(dist, posSlot);
}
}
}
}
}
+
+
#define ChangePair(smallDist, bigDist) (((bigDist) >> 7) > (smallDist))
-static UInt32 GetOptimumFast(CLzmaEnc *p, UInt32 *backRes)
+
+
+static unsigned GetOptimumFast(CLzmaEnc *p)
{
- UInt32 numAvail, mainLen, mainDist, numPairs, repIndex, repLen, i;
+ UInt32 numAvail, mainDist;
+ unsigned mainLen, numPairs, repIndex, repLen, i;
const Byte *data;
- const UInt32 *matches;
if (p->additionalOffset == 0)
mainLen = ReadMatchDistances(p, &numPairs);
else
{
- mainLen = p->longestMatchLength;
+ mainLen = p->longestMatchLen;
numPairs = p->numPairs;
}
numAvail = p->numAvail;
- *backRes = (UInt32)-1;
+ p->backRes = MARK_LIT;
if (numAvail < 2)
return 1;
if (numAvail > LZMA_MATCH_LEN_MAX)
numAvail = LZMA_MATCH_LEN_MAX;
data = p->matchFinder.GetPointerToCurrentPos(p->matchFinderObj) - 1;
-
repLen = repIndex = 0;
+
for (i = 0; i < LZMA_NUM_REPS; i++)
{
- UInt32 len;
- const Byte *data2 = data - p->reps[i] - 1;
+ unsigned len;
+ const Byte *data2 = data - p->reps[i];
if (data[0] != data2[0] || data[1] != data2[1])
continue;
for (len = 2; len < numAvail && data[len] == data2[len]; len++);
if (len >= p->numFastBytes)
{
- *backRes = i;
- MovePos(p, len - 1);
+ p->backRes = i;
+ MOVE_POS(p, len - 1)
return len;
}
if (len > repLen)
@@ -1594,84 +1758,152 @@ static UInt32 GetOptimumFast(CLzmaEnc *p, UInt32 *backRes)
}
}
- matches = p->matches;
if (mainLen >= p->numFastBytes)
{
- *backRes = matches[numPairs - 1] + LZMA_NUM_REPS;
- MovePos(p, mainLen - 1);
+ p->backRes = p->matches[(size_t)numPairs - 1] + LZMA_NUM_REPS;
+ MOVE_POS(p, mainLen - 1)
return mainLen;
}
mainDist = 0; /* for GCC */
+
if (mainLen >= 2)
{
- mainDist = matches[numPairs - 1];
- while (numPairs > 2 && mainLen == matches[numPairs - 4] + 1)
+ mainDist = p->matches[(size_t)numPairs - 1];
+ while (numPairs > 2)
{
- if (!ChangePair(matches[numPairs - 3], mainDist))
+ UInt32 dist2;
+ if (mainLen != p->matches[(size_t)numPairs - 4] + 1)
+ break;
+ dist2 = p->matches[(size_t)numPairs - 3];
+ if (!ChangePair(dist2, mainDist))
break;
numPairs -= 2;
- mainLen = matches[numPairs - 2];
- mainDist = matches[numPairs - 1];
+ mainLen--;
+ mainDist = dist2;
}
if (mainLen == 2 && mainDist >= 0x80)
mainLen = 1;
}
- if (repLen >= 2 && (
- (repLen + 1 >= mainLen) ||
- (repLen + 2 >= mainLen && mainDist >= (1 << 9)) ||
- (repLen + 3 >= mainLen && mainDist >= (1 << 15))))
+ if (repLen >= 2)
+ if ( repLen + 1 >= mainLen
+ || (repLen + 2 >= mainLen && mainDist >= (1 << 9))
+ || (repLen + 3 >= mainLen && mainDist >= (1 << 15)))
{
- *backRes = repIndex;
- MovePos(p, repLen - 1);
+ p->backRes = repIndex;
+ MOVE_POS(p, repLen - 1)
return repLen;
}
if (mainLen < 2 || numAvail <= 2)
return 1;
- p->longestMatchLength = ReadMatchDistances(p, &p->numPairs);
- if (p->longestMatchLength >= 2)
{
- UInt32 newDistance = matches[p->numPairs - 1];
- if ((p->longestMatchLength >= mainLen && newDistance < mainDist) ||
- (p->longestMatchLength == mainLen + 1 && !ChangePair(mainDist, newDistance)) ||
- (p->longestMatchLength > mainLen + 1) ||
- (p->longestMatchLength + 1 >= mainLen && mainLen >= 3 && ChangePair(newDistance, mainDist)))
- return 1;
+ unsigned len1 = ReadMatchDistances(p, &p->numPairs);
+ p->longestMatchLen = len1;
+
+ if (len1 >= 2)
+ {
+ UInt32 newDist = p->matches[(size_t)p->numPairs - 1];
+ if ( (len1 >= mainLen && newDist < mainDist)
+ || (len1 == mainLen + 1 && !ChangePair(mainDist, newDist))
+ || (len1 > mainLen + 1)
+ || (len1 + 1 >= mainLen && mainLen >= 3 && ChangePair(newDist, mainDist)))
+ return 1;
+ }
}
data = p->matchFinder.GetPointerToCurrentPos(p->matchFinderObj) - 1;
+
for (i = 0; i < LZMA_NUM_REPS; i++)
{
- UInt32 len, limit;
- const Byte *data2 = data - p->reps[i] - 1;
+ unsigned len, limit;
+ const Byte *data2 = data - p->reps[i];
if (data[0] != data2[0] || data[1] != data2[1])
continue;
limit = mainLen - 1;
- for (len = 2; len < limit && data[len] == data2[len]; len++);
- if (len >= limit)
- return 1;
+ for (len = 2;; len++)
+ {
+ if (len >= limit)
+ return 1;
+ if (data[len] != data2[len])
+ break;
+ }
+ }
+
+ p->backRes = mainDist + LZMA_NUM_REPS;
+ if (mainLen != 2)
+ {
+ MOVE_POS(p, mainLen - 2)
}
- *backRes = mainDist + LZMA_NUM_REPS;
- MovePos(p, mainLen - 2);
return mainLen;
}
-static void WriteEndMarker(CLzmaEnc *p, UInt32 posState)
+
+
+
+static void WriteEndMarker(CLzmaEnc *p, unsigned posState)
{
- UInt32 len;
- RangeEnc_EncodeBit(&p->rc, &p->isMatch[p->state][posState], 1);
- RangeEnc_EncodeBit(&p->rc, &p->isRep[p->state], 0);
+ UInt32 range;
+ range = p->rc.range;
+ {
+ UInt32 ttt, newBound;
+ CLzmaProb *prob = &p->isMatch[p->state][posState];
+ RC_BIT_PRE(&p->rc, prob)
+ RC_BIT_1(&p->rc, prob)
+ prob = &p->isRep[p->state];
+ RC_BIT_PRE(&p->rc, prob)
+ RC_BIT_0(&p->rc, prob)
+ }
p->state = kMatchNextStates[p->state];
- len = LZMA_MATCH_LEN_MIN;
- LenEnc_Encode2(&p->lenEnc, &p->rc, len - LZMA_MATCH_LEN_MIN, posState, !p->fastMode, p->ProbPrices);
- RcTree_Encode(&p->rc, p->posSlotEncoder[GetLenToPosState(len)], kNumPosSlotBits, (1 << kNumPosSlotBits) - 1);
- RangeEnc_EncodeDirectBits(&p->rc, (((UInt32)1 << 30) - 1) >> kNumAlignBits, 30 - kNumAlignBits);
- RcTree_ReverseEncode(&p->rc, p->posAlignEncoder, kNumAlignBits, kAlignMask);
+
+ p->rc.range = range;
+ LenEnc_Encode(&p->lenProbs, &p->rc, 0, posState);
+ range = p->rc.range;
+
+ {
+ // RcTree_Encode_PosSlot(&p->rc, p->posSlotEncoder[0], (1 << kNumPosSlotBits) - 1);
+ CLzmaProb *probs = p->posSlotEncoder[0];
+ unsigned m = 1;
+ do
+ {
+ UInt32 ttt, newBound;
+ RC_BIT_PRE(p, probs + m)
+ RC_BIT_1(&p->rc, probs + m);
+ m = (m << 1) + 1;
+ }
+ while (m < (1 << kNumPosSlotBits));
+ }
+ {
+ // RangeEnc_EncodeDirectBits(&p->rc, ((UInt32)1 << (30 - kNumAlignBits)) - 1, 30 - kNumAlignBits); UInt32 range = p->range;
+ unsigned numBits = 30 - kNumAlignBits;
+ do
+ {
+ range >>= 1;
+ p->rc.low += range;
+ RC_NORM(&p->rc)
+ }
+ while (--numBits);
+ }
+
+ {
+ // RcTree_ReverseEncode(&p->rc, p->posAlignEncoder, kNumAlignBits, kAlignMask);
+ CLzmaProb *probs = p->posAlignEncoder;
+ unsigned m = 1;
+ do
+ {
+ UInt32 ttt, newBound;
+ RC_BIT_PRE(p, probs + m)
+ RC_BIT_1(&p->rc, probs + m);
+ m = (m << 1) + 1;
+ }
+ while (m < kAlignTableSize);
+ }
+ p->rc.range = range;
}
+
static SRes CheckErrors(CLzmaEnc *p)
{
if (p->result != SZ_OK)
@@ -1685,7 +1917,8 @@ static SRes CheckErrors(CLzmaEnc *p)
return p->result;
}
-static SRes Flush(CLzmaEnc *p, UInt32 nowPos)
+
+MY_NO_INLINE static SRes Flush(CLzmaEnc *p, UInt32 nowPos)
{
/* ReleaseMFStream(); */
p->finished = True;
@@ -1696,47 +1929,108 @@ static SRes Flush(CLzmaEnc *p, UInt32 nowPos)
return CheckErrors(p);
}
+
+
static void FillAlignPrices(CLzmaEnc *p)
{
- UInt32 i;
- for (i = 0; i < kAlignTableSize; i++)
- p->alignPrices[i] = RcTree_ReverseGetPrice(p->posAlignEncoder, kNumAlignBits, i, p->ProbPrices);
+ unsigned i;
+ const CProbPrice *ProbPrices = p->ProbPrices;
+ const CLzmaProb *probs = p->posAlignEncoder;
p->alignPriceCount = 0;
+ for (i = 0; i < kAlignTableSize / 2; i++)
+ {
+ UInt32 price = 0;
+ unsigned symbol = i;
+ unsigned m = 1;
+ unsigned bit;
+ UInt32 prob;
+ bit = symbol & 1; symbol >>= 1; price += GET_PRICEa(probs[m], bit); m = (m << 1) + bit;
+ bit = symbol & 1; symbol >>= 1; price += GET_PRICEa(probs[m], bit); m = (m << 1) + bit;
+ bit = symbol & 1; symbol >>= 1; price += GET_PRICEa(probs[m], bit); m = (m << 1) + bit;
+ prob = probs[m];
+ p->alignPrices[i ] = price + GET_PRICEa_0(prob);
+ p->alignPrices[i + 8] = price + GET_PRICEa_1(prob);
+ // p->alignPrices[i] = RcTree_ReverseGetPrice(p->posAlignEncoder, kNumAlignBits, i, p->ProbPrices);
+ }
}
+
static void FillDistancesPrices(CLzmaEnc *p)
{
UInt32 tempPrices[kNumFullDistances];
- UInt32 i, lenToPosState;
+ unsigned i, lenToPosState;
+
+ const CProbPrice *ProbPrices = p->ProbPrices;
+ p->matchPriceCount = 0;
+
for (i = kStartPosModelIndex; i < kNumFullDistances; i++)
{
- UInt32 posSlot = GetPosSlot1(i);
- UInt32 footerBits = ((posSlot >> 1) - 1);
- UInt32 base = ((2 | (posSlot & 1)) << footerBits);
- tempPrices[i] = RcTree_ReverseGetPrice(p->posEncoders + base - posSlot - 1, footerBits, i - base, p->ProbPrices);
+ unsigned posSlot = GetPosSlot1(i);
+ unsigned footerBits = ((posSlot >> 1) - 1);
+ unsigned base = ((2 | (posSlot & 1)) << footerBits);
+ // tempPrices[i] = RcTree_ReverseGetPrice(p->posEncoders + base, footerBits, i - base, p->ProbPrices);
+
+ const CLzmaProb *probs = p->posEncoders + base;
+ UInt32 price = 0;
+ unsigned m = 1;
+ unsigned symbol = i - base;
+ do
+ {
+ unsigned bit = symbol & 1;
+ symbol >>= 1;
+ price += GET_PRICEa(probs[m], bit);
+ m = (m << 1) + bit;
+ }
+ while (--footerBits);
+ tempPrices[i] = price;
}
for (lenToPosState = 0; lenToPosState < kNumLenToPosStates; lenToPosState++)
{
- UInt32 posSlot;
+ unsigned posSlot;
const CLzmaProb *encoder = p->posSlotEncoder[lenToPosState];
UInt32 *posSlotPrices = p->posSlotPrices[lenToPosState];
- for (posSlot = 0; posSlot < p->distTableSize; posSlot++)
- posSlotPrices[posSlot] = RcTree_GetPrice(encoder, kNumPosSlotBits, posSlot, p->ProbPrices);
- for (posSlot = kEndPosModelIndex; posSlot < p->distTableSize; posSlot++)
- posSlotPrices[posSlot] += ((((posSlot >> 1) - 1) - kNumAlignBits) << kNumBitPriceShiftBits);
+ unsigned distTableSize = p->distTableSize;
+ const CLzmaProb *probs = encoder;
+ for (posSlot = 0; posSlot < distTableSize; posSlot += 2)
+ {
+ // posSlotPrices[posSlot] = RcTree_GetPrice(encoder, kNumPosSlotBits, posSlot, p->ProbPrices);
+ UInt32 price = 0;
+ unsigned bit;
+ unsigned symbol = (posSlot >> 1) + (1 << (kNumPosSlotBits - 1));
+ UInt32 prob;
+ bit = symbol & 1; symbol >>= 1; price += GET_PRICEa(probs[symbol], bit);
+ bit = symbol & 1; symbol >>= 1; price += GET_PRICEa(probs[symbol], bit);
+ bit = symbol & 1; symbol >>= 1; price += GET_PRICEa(probs[symbol], bit);
+ bit = symbol & 1; symbol >>= 1; price += GET_PRICEa(probs[symbol], bit);
+ bit = symbol & 1; symbol >>= 1; price += GET_PRICEa(probs[symbol], bit);
+ prob = probs[(posSlot >> 1) + (1 << (kNumPosSlotBits - 1))];
+ posSlotPrices[posSlot ] = price + GET_PRICEa_0(prob);
+ posSlotPrices[posSlot + 1] = price + GET_PRICEa_1(prob);
+ }
+ for (posSlot = kEndPosModelIndex; posSlot < distTableSize; posSlot++)
+ posSlotPrices[posSlot] += ((UInt32)(((posSlot >> 1) - 1) - kNumAlignBits) << kNumBitPriceShiftBits);
{
UInt32 *distancesPrices = p->distancesPrices[lenToPosState];
- for (i = 0; i < kStartPosModelIndex; i++)
- distancesPrices[i] = posSlotPrices[i];
- for (; i < kNumFullDistances; i++)
- distancesPrices[i] = posSlotPrices[GetPosSlot1(i)] + tempPrices[i];
+ {
+ distancesPrices[0] = posSlotPrices[0];
+ distancesPrices[1] = posSlotPrices[1];
+ distancesPrices[2] = posSlotPrices[2];
+ distancesPrices[3] = posSlotPrices[3];
+ }
+ for (i = 4; i < kNumFullDistances; i += 2)
+ {
+ UInt32 slotPrice = posSlotPrices[GetPosSlot1(i)];
+ distancesPrices[i ] = slotPrice + tempPrices[i];
+ distancesPrices[i + 1] = slotPrice + tempPrices[i + 1];
+ }
}
}
- p->matchPriceCount = 0;
}
+
+
void LzmaEnc_Construct(CLzmaEnc *p)
{
RangeEnc_Construct(&p->rc);
@@ -1760,26 +2054,27 @@ void LzmaEnc_Construct(CLzmaEnc *p)
LzmaEnc_InitPriceTables(p->ProbPrices);
p->litProbs = NULL;
p->saveState.litProbs = NULL;
+
}
-CLzmaEncHandle LzmaEnc_Create(ISzAlloc *alloc)
+CLzmaEncHandle LzmaEnc_Create(ISzAllocPtr alloc)
{
void *p;
- p = alloc->Alloc(alloc, sizeof(CLzmaEnc));
+ p = ISzAlloc_Alloc(alloc, sizeof(CLzmaEnc));
if (p)
LzmaEnc_Construct((CLzmaEnc *)p);
return p;
}
-void LzmaEnc_FreeLits(CLzmaEnc *p, ISzAlloc *alloc)
+void LzmaEnc_FreeLits(CLzmaEnc *p, ISzAllocPtr alloc)
{
- alloc->Free(alloc, p->litProbs);
- alloc->Free(alloc, p->saveState.litProbs);
+ ISzAlloc_Free(alloc, p->litProbs);
+ ISzAlloc_Free(alloc, p->saveState.litProbs);
p->litProbs = NULL;
p->saveState.litProbs = NULL;
}
-void LzmaEnc_Destruct(CLzmaEnc *p, ISzAlloc *alloc, ISzAlloc *allocBig)
+void LzmaEnc_Destruct(CLzmaEnc *p, ISzAllocPtr alloc, ISzAllocPtr allocBig)
{
#ifndef _7ZIP_ST
MatchFinderMt_Destruct(&p->matchFinderMt, allocBig);
@@ -1790,13 +2085,14 @@ void LzmaEnc_Destruct(CLzmaEnc *p, ISzAlloc *alloc, ISzAlloc *allocBig)
RangeEnc_Free(&p->rc, alloc);
}
-void LzmaEnc_Destroy(CLzmaEncHandle p, ISzAlloc *alloc, ISzAlloc *allocBig)
+void LzmaEnc_Destroy(CLzmaEncHandle p, ISzAllocPtr alloc, ISzAllocPtr allocBig)
{
LzmaEnc_Destruct((CLzmaEnc *)p, alloc, allocBig);
- alloc->Free(alloc, p);
+ ISzAlloc_Free(alloc, p);
}
-static SRes LzmaEnc_CodeOneBlock(CLzmaEnc *p, Bool useLimits, UInt32 maxPackSize, UInt32 maxUnpackSize)
+
+static SRes LzmaEnc_CodeOneBlock(CLzmaEnc *p, UInt32 maxPackSize, UInt32 maxUnpackSize)
{
UInt32 nowPos32, startPos32;
if (p->needInit)
@@ -1814,13 +2110,13 @@ static SRes LzmaEnc_CodeOneBlock(CLzmaEnc *p, Bool useLimits, UInt32 maxPackSize
if (p->nowPos64 == 0)
{
- UInt32 numPairs;
+ unsigned numPairs;
Byte curByte;
if (p->matchFinder.GetNumAvailableBytes(p->matchFinderObj) == 0)
return Flush(p, nowPos32);
ReadMatchDistances(p, &numPairs);
- RangeEnc_EncodeBit(&p->rc, &p->isMatch[p->state][0], 0);
- p->state = kLiteralNextStates[p->state];
+ RangeEnc_EncodeBit_0(&p->rc, &p->isMatch[kState_Start][0]);
+ // p->state = kLiteralNextStates[p->state];
curByte = *(p->matchFinder.GetPointerToCurrentPos(p->matchFinderObj) - p->additionalOffset);
LitEnc_Encode(&p->rc, p->litProbs, curByte);
p->additionalOffset--;
@@ -1828,109 +2124,225 @@ static SRes LzmaEnc_CodeOneBlock(CLzmaEnc *p, Bool useLimits, UInt32 maxPackSize
}
if (p->matchFinder.GetNumAvailableBytes(p->matchFinderObj) != 0)
+
for (;;)
{
- UInt32 pos, len, posState;
-
+ UInt32 dist;
+ unsigned len, posState;
+ UInt32 range, ttt, newBound;
+ CLzmaProb *probs;
+
if (p->fastMode)
- len = GetOptimumFast(p, &pos);
+ len = GetOptimumFast(p);
else
- len = GetOptimum(p, nowPos32, &pos);
+ {
+ unsigned oci = p->optCur;
+ if (p->optEnd == oci)
+ len = GetOptimum(p, nowPos32);
+ else
+ {
+ const COptimal *opt = &p->opt[oci];
+ len = opt->len;
+ p->backRes = opt->dist;
+ p->optCur = oci + 1;
+ }
+ }
+
+ posState = (unsigned)nowPos32 & p->pbMask;
+ range = p->rc.range;
+ probs = &p->isMatch[p->state][posState];
+
+ RC_BIT_PRE(&p->rc, probs)
+
+ dist = p->backRes;
#ifdef SHOW_STAT2
- printf("\n pos = %4X, len = %u pos = %u", nowPos32, len, pos);
+ printf("\n pos = %6X, len = %3u pos = %6u", nowPos32, len, dist);
#endif
- posState = nowPos32 & p->pbMask;
- if (len == 1 && pos == (UInt32)-1)
+ if (dist == MARK_LIT)
{
Byte curByte;
- CLzmaProb *probs;
const Byte *data;
+ unsigned state;
- RangeEnc_EncodeBit(&p->rc, &p->isMatch[p->state][posState], 0);
+ RC_BIT_0(&p->rc, probs);
+ p->rc.range = range;
data = p->matchFinder.GetPointerToCurrentPos(p->matchFinderObj) - p->additionalOffset;
- curByte = *data;
probs = LIT_PROBS(nowPos32, *(data - 1));
- if (IsCharState(p->state))
+ curByte = *data;
+ state = p->state;
+ p->state = kLiteralNextStates[state];
+ if (IsLitState(state))
LitEnc_Encode(&p->rc, probs, curByte);
else
- LitEnc_EncodeMatched(&p->rc, probs, curByte, *(data - p->reps[0] - 1));
- p->state = kLiteralNextStates[p->state];
+ LitEnc_EncodeMatched(&p->rc, probs, curByte, *(data - p->reps[0]));
}
else
{
- RangeEnc_EncodeBit(&p->rc, &p->isMatch[p->state][posState], 1);
- if (pos < LZMA_NUM_REPS)
+ RC_BIT_1(&p->rc, probs);
+ probs = &p->isRep[p->state];
+ RC_BIT_PRE(&p->rc, probs)
+
+ if (dist < LZMA_NUM_REPS)
{
- RangeEnc_EncodeBit(&p->rc, &p->isRep[p->state], 1);
- if (pos == 0)
+ RC_BIT_1(&p->rc, probs);
+ probs = &p->isRepG0[p->state];
+ RC_BIT_PRE(&p->rc, probs)
+ if (dist == 0)
{
- RangeEnc_EncodeBit(&p->rc, &p->isRepG0[p->state], 0);
- RangeEnc_EncodeBit(&p->rc, &p->isRep0Long[p->state][posState], ((len == 1) ? 0 : 1));
+ RC_BIT_0(&p->rc, probs);
+ probs = &p->isRep0Long[p->state][posState];
+ RC_BIT_PRE(&p->rc, probs)
+ if (len != 1)
+ {
+ RC_BIT_1_BASE(&p->rc, probs);
+ }
+ else
+ {
+ RC_BIT_0_BASE(&p->rc, probs);
+ p->state = kShortRepNextStates[p->state];
+ }
}
else
{
- UInt32 distance = p->reps[pos];
- RangeEnc_EncodeBit(&p->rc, &p->isRepG0[p->state], 1);
- if (pos == 1)
- RangeEnc_EncodeBit(&p->rc, &p->isRepG1[p->state], 0);
+ RC_BIT_1(&p->rc, probs);
+ probs = &p->isRepG1[p->state];
+ RC_BIT_PRE(&p->rc, probs)
+ if (dist == 1)
+ {
+ RC_BIT_0_BASE(&p->rc, probs);
+ dist = p->reps[1];
+ }
else
{
- RangeEnc_EncodeBit(&p->rc, &p->isRepG1[p->state], 1);
- RangeEnc_EncodeBit(&p->rc, &p->isRepG2[p->state], pos - 2);
- if (pos == 3)
+ RC_BIT_1(&p->rc, probs);
+ probs = &p->isRepG2[p->state];
+ RC_BIT_PRE(&p->rc, probs)
+ if (dist == 2)
+ {
+ RC_BIT_0_BASE(&p->rc, probs);
+ dist = p->reps[2];
+ }
+ else
+ {
+ RC_BIT_1_BASE(&p->rc, probs);
+ dist = p->reps[3];
p->reps[3] = p->reps[2];
+ }
p->reps[2] = p->reps[1];
}
p->reps[1] = p->reps[0];
- p->reps[0] = distance;
+ p->reps[0] = dist;
}
- if (len == 1)
- p->state = kShortRepNextStates[p->state];
- else
+
+ RC_NORM(&p->rc)
+
+ p->rc.range = range;
+
+ if (len != 1)
{
- LenEnc_Encode2(&p->repLenEnc, &p->rc, len - LZMA_MATCH_LEN_MIN, posState, !p->fastMode, p->ProbPrices);
+ LenEnc_Encode(&p->repLenProbs, &p->rc, len - LZMA_MATCH_LEN_MIN, posState);
+ if (!p->fastMode)
+ if (--p->repLenEnc.counters[posState] == 0)
+ LenPriceEnc_UpdateTable(&p->repLenEnc, posState, &p->repLenProbs, p->ProbPrices);
+
p->state = kRepNextStates[p->state];
}
}
else
{
- UInt32 posSlot;
- RangeEnc_EncodeBit(&p->rc, &p->isRep[p->state], 0);
+ unsigned posSlot;
+ RC_BIT_0(&p->rc, probs);
+ p->rc.range = range;
p->state = kMatchNextStates[p->state];
- LenEnc_Encode2(&p->lenEnc, &p->rc, len - LZMA_MATCH_LEN_MIN, posState, !p->fastMode, p->ProbPrices);
- pos -= LZMA_NUM_REPS;
- GetPosSlot(pos, posSlot);
- RcTree_Encode(&p->rc, p->posSlotEncoder[GetLenToPosState(len)], kNumPosSlotBits, posSlot);
+
+ LenEnc_Encode(&p->lenProbs, &p->rc, len - LZMA_MATCH_LEN_MIN, posState);
+ if (!p->fastMode)
+ if (--p->lenEnc.counters[posState] == 0)
+ LenPriceEnc_UpdateTable(&p->lenEnc, posState, &p->lenProbs, p->ProbPrices);
+
+ dist -= LZMA_NUM_REPS;
+ p->reps[3] = p->reps[2];
+ p->reps[2] = p->reps[1];
+ p->reps[1] = p->reps[0];
+ p->reps[0] = dist + 1;
- if (posSlot >= kStartPosModelIndex)
+ p->matchPriceCount++;
+ GetPosSlot(dist, posSlot);
+ // RcTree_Encode_PosSlot(&p->rc, p->posSlotEncoder[GetLenToPosState(len)], posSlot);
+ {
+ UInt32 symbol = posSlot + (1 << kNumPosSlotBits);
+ range = p->rc.range;
+ probs = p->posSlotEncoder[GetLenToPosState(len)];
+ do
+ {
+ CLzmaProb *prob = probs + (symbol >> kNumPosSlotBits);
+ UInt32 bit = (symbol >> (kNumPosSlotBits - 1)) & 1;
+ symbol <<= 1;
+ RC_BIT(&p->rc, prob, bit);
+ }
+ while (symbol < (1 << kNumPosSlotBits * 2));
+ p->rc.range = range;
+ }
+
+ if (dist >= kStartPosModelIndex)
{
- UInt32 footerBits = ((posSlot >> 1) - 1);
- UInt32 base = ((2 | (posSlot & 1)) << footerBits);
- UInt32 posReduced = pos - base;
+ unsigned footerBits = ((posSlot >> 1) - 1);
- if (posSlot < kEndPosModelIndex)
- RcTree_ReverseEncode(&p->rc, p->posEncoders + base - posSlot - 1, footerBits, posReduced);
+ if (dist < kNumFullDistances)
+ {
+ unsigned base = ((2 | (posSlot & 1)) << footerBits);
+ RcTree_ReverseEncode(&p->rc, p->posEncoders + base, footerBits, dist - base);
+ }
else
{
- RangeEnc_EncodeDirectBits(&p->rc, posReduced >> kNumAlignBits, footerBits - kNumAlignBits);
- RcTree_ReverseEncode(&p->rc, p->posAlignEncoder, kNumAlignBits, posReduced & kAlignMask);
- p->alignPriceCount++;
+ UInt32 pos2 = (dist | 0xF) << (32 - footerBits);
+ range = p->rc.range;
+ // RangeEnc_EncodeDirectBits(&p->rc, posReduced >> kNumAlignBits, footerBits - kNumAlignBits);
+ /*
+ do
+ {
+ range >>= 1;
+ p->rc.low += range & (0 - ((dist >> --footerBits) & 1));
+ RC_NORM(&p->rc)
+ }
+ while (footerBits > kNumAlignBits);
+ */
+ do
+ {
+ range >>= 1;
+ p->rc.low += range & (0 - (pos2 >> 31));
+ pos2 += pos2;
+ RC_NORM(&p->rc)
+ }
+ while (pos2 != 0xF0000000);
+
+
+ // RcTree_ReverseEncode(&p->rc, p->posAlignEncoder, kNumAlignBits, posReduced & kAlignMask);
+
+ {
+ unsigned m = 1;
+ unsigned bit;
+ bit = dist & 1; dist >>= 1; RC_BIT(&p->rc, p->posAlignEncoder + m, bit); m = (m << 1) + bit;
+ bit = dist & 1; dist >>= 1; RC_BIT(&p->rc, p->posAlignEncoder + m, bit); m = (m << 1) + bit;
+ bit = dist & 1; dist >>= 1; RC_BIT(&p->rc, p->posAlignEncoder + m, bit); m = (m << 1) + bit;
+ bit = dist & 1; RC_BIT(&p->rc, p->posAlignEncoder + m, bit);
+ p->rc.range = range;
+ p->alignPriceCount++;
+ }
}
}
- p->reps[3] = p->reps[2];
- p->reps[2] = p->reps[1];
- p->reps[1] = p->reps[0];
- p->reps[0] = pos;
- p->matchPriceCount++;
}
}
- p->additionalOffset -= len;
+
nowPos32 += len;
+ p->additionalOffset -= len;
+
if (p->additionalOffset == 0)
{
UInt32 processed;
+
if (!p->fastMode)
{
if (p->matchPriceCount >= (1 << 7))
@@ -1938,13 +2350,15 @@ static SRes LzmaEnc_CodeOneBlock(CLzmaEnc *p, Bool useLimits, UInt32 maxPackSize
if (p->alignPriceCount >= kAlignTableSize)
FillAlignPrices(p);
}
+
if (p->matchFinder.GetNumAvailableBytes(p->matchFinderObj) == 0)
break;
processed = nowPos32 - startPos32;
- if (useLimits)
+
+ if (maxPackSize)
{
- if (processed + kNumOpts + 300 >= maxUnpackSize ||
- RangeEnc_GetProcessed(&p->rc) + kNumOpts * 2 >= maxPackSize)
+ if (processed + kNumOpts + 300 >= maxUnpackSize
+ || RangeEnc_GetProcessed_sizet(&p->rc) + kPackReserve >= maxPackSize)
break;
}
else if (processed >= (1 << 17))
@@ -1954,13 +2368,16 @@ static SRes LzmaEnc_CodeOneBlock(CLzmaEnc *p, Bool useLimits, UInt32 maxPackSize
}
}
}
+
p->nowPos64 += nowPos32 - startPos32;
return Flush(p, nowPos32);
}
+
+
#define kBigHashDicLimit ((UInt32)1 << 24)
-static SRes LzmaEnc_Alloc(CLzmaEnc *p, UInt32 keepWindowSize, ISzAlloc *alloc, ISzAlloc *allocBig)
+static SRes LzmaEnc_Alloc(CLzmaEnc *p, UInt32 keepWindowSize, ISzAllocPtr alloc, ISzAllocPtr allocBig)
{
UInt32 beforeSize = kNumOpts;
if (!RangeEnc_Alloc(&p->rc, alloc))
@@ -1975,8 +2392,8 @@ static SRes LzmaEnc_Alloc(CLzmaEnc *p, UInt32 keepWindowSize, ISzAlloc *alloc, I
if (!p->litProbs || !p->saveState.litProbs || p->lclp != lclp)
{
LzmaEnc_FreeLits(p, alloc);
- p->litProbs = (CLzmaProb *)alloc->Alloc(alloc, ((UInt32)0x300 << lclp) * sizeof(CLzmaProb));
- p->saveState.litProbs = (CLzmaProb *)alloc->Alloc(alloc, ((UInt32)0x300 << lclp) * sizeof(CLzmaProb));
+ p->litProbs = (CLzmaProb *)ISzAlloc_Alloc(alloc, ((UInt32)0x300 << lclp) * sizeof(CLzmaProb));
+ p->saveState.litProbs = (CLzmaProb *)ISzAlloc_Alloc(alloc, ((UInt32)0x300 << lclp) * sizeof(CLzmaProb));
if (!p->litProbs || !p->saveState.litProbs)
{
LzmaEnc_FreeLits(p, alloc);
@@ -1994,8 +2411,13 @@ static SRes LzmaEnc_Alloc(CLzmaEnc *p, UInt32 keepWindowSize, ISzAlloc *alloc, I
#ifndef _7ZIP_ST
if (p->mtMode)
{
- RINOK(MatchFinderMt_Create(&p->matchFinderMt, p->dictSize, beforeSize, p->numFastBytes, LZMA_MATCH_LEN_MAX, allocBig));
+ RINOK(MatchFinderMt_Create(&p->matchFinderMt, p->dictSize, beforeSize, p->numFastBytes,
+ LZMA_MATCH_LEN_MAX
+ + 1 /* 18.04 */
+ , allocBig));
p->matchFinderObj = &p->matchFinderMt;
+ p->matchFinderBase.bigHash = (Byte)(
+ (p->dictSize > kBigHashDicLimit && p->matchFinderBase.hashMask >= 0xFFFFFF) ? 1 : 0);
MatchFinderMt_CreateVTable(&p->matchFinderMt, &p->matchFinder);
}
else
@@ -2012,17 +2434,21 @@ static SRes LzmaEnc_Alloc(CLzmaEnc *p, UInt32 keepWindowSize, ISzAlloc *alloc, I
void LzmaEnc_Init(CLzmaEnc *p)
{
- UInt32 i;
+ unsigned i;
p->state = 0;
- for (i = 0 ; i < LZMA_NUM_REPS; i++)
- p->reps[i] = 0;
+ p->reps[0] =
+ p->reps[1] =
+ p->reps[2] =
+ p->reps[3] = 1;
RangeEnc_Init(&p->rc);
+ for (i = 0; i < (1 << kNumAlignBits); i++)
+ p->posAlignEncoder[i] = kProbInitValue;
for (i = 0; i < kNumStates; i++)
{
- UInt32 j;
+ unsigned j;
for (j = 0; j < LZMA_NUM_PB_STATES_MAX; j++)
{
p->isMatch[i][j] = kProbInitValue;
@@ -2034,39 +2460,38 @@ void LzmaEnc_Init(CLzmaEnc *p)
p->isRepG2[i] = kProbInitValue;
}
- {
- UInt32 num = (UInt32)0x300 << (p->lp + p->lc);
- CLzmaProb *probs = p->litProbs;
- for (i = 0; i < num; i++)
- probs[i] = kProbInitValue;
- }
-
{
for (i = 0; i < kNumLenToPosStates; i++)
{
CLzmaProb *probs = p->posSlotEncoder[i];
- UInt32 j;
+ unsigned j;
for (j = 0; j < (1 << kNumPosSlotBits); j++)
probs[j] = kProbInitValue;
}
}
{
- for (i = 0; i < kNumFullDistances - kEndPosModelIndex; i++)
+ for (i = 0; i < kNumFullDistances; i++)
p->posEncoders[i] = kProbInitValue;
}
- LenEnc_Init(&p->lenEnc.p);
- LenEnc_Init(&p->repLenEnc.p);
+ {
+ UInt32 num = (UInt32)0x300 << (p->lp + p->lc);
+ UInt32 k;
+ CLzmaProb *probs = p->litProbs;
+ for (k = 0; k < num; k++)
+ probs[k] = kProbInitValue;
+ }
- for (i = 0; i < (1 << kNumAlignBits); i++)
- p->posAlignEncoder[i] = kProbInitValue;
- p->optimumEndIndex = 0;
- p->optimumCurrentIndex = 0;
+ LenEnc_Init(&p->lenProbs);
+ LenEnc_Init(&p->repLenProbs);
+
+ p->optEnd = 0;
+ p->optCur = 0;
p->additionalOffset = 0;
p->pbMask = (1 << p->pb) - 1;
- p->lpMask = (1 << p->lp) - 1;
+ p->lpMask = ((UInt32)0x100 << p->lp) - ((unsigned)0x100 >> p->lc);
}
void LzmaEnc_InitPrices(CLzmaEnc *p)
@@ -2080,14 +2505,14 @@ void LzmaEnc_InitPrices(CLzmaEnc *p)
p->lenEnc.tableSize =
p->repLenEnc.tableSize =
p->numFastBytes + 1 - LZMA_MATCH_LEN_MIN;
- LenPriceEnc_UpdateTables(&p->lenEnc, 1 << p->pb, p->ProbPrices);
- LenPriceEnc_UpdateTables(&p->repLenEnc, 1 << p->pb, p->ProbPrices);
+ LenPriceEnc_UpdateTables(&p->lenEnc, 1 << p->pb, &p->lenProbs, p->ProbPrices);
+ LenPriceEnc_UpdateTables(&p->repLenEnc, 1 << p->pb, &p->repLenProbs, p->ProbPrices);
}
-static SRes LzmaEnc_AllocAndInit(CLzmaEnc *p, UInt32 keepWindowSize, ISzAlloc *alloc, ISzAlloc *allocBig)
+static SRes LzmaEnc_AllocAndInit(CLzmaEnc *p, UInt32 keepWindowSize, ISzAllocPtr alloc, ISzAllocPtr allocBig)
{
- UInt32 i;
- for (i = 0; i < (UInt32)kDicLogSizeMaxCompress; i++)
+ unsigned i;
+ for (i = kEndPosModelIndex / 2; i < kDicLogSizeMax; i++)
if (p->dictSize <= ((UInt32)1 << i))
break;
p->distTableSize = i * 2;
@@ -2102,7 +2527,7 @@ static SRes LzmaEnc_AllocAndInit(CLzmaEnc *p, UInt32 keepWindowSize, ISzAlloc *a
}
static SRes LzmaEnc_Prepare(CLzmaEncHandle pp, ISeqOutStream *outStream, ISeqInStream *inStream,
- ISzAlloc *alloc, ISzAlloc *allocBig)
+ ISzAllocPtr alloc, ISzAllocPtr allocBig)
{
CLzmaEnc *p = (CLzmaEnc *)pp;
p->matchFinderBase.stream = inStream;
@@ -2113,7 +2538,7 @@ static SRes LzmaEnc_Prepare(CLzmaEncHandle pp, ISeqOutStream *outStream, ISeqInS
SRes LzmaEnc_PrepareForLzma2(CLzmaEncHandle pp,
ISeqInStream *inStream, UInt32 keepWindowSize,
- ISzAlloc *alloc, ISzAlloc *allocBig)
+ ISzAllocPtr alloc, ISzAllocPtr allocBig)
{
CLzmaEnc *p = (CLzmaEnc *)pp;
p->matchFinderBase.stream = inStream;
@@ -2129,12 +2554,13 @@ static void LzmaEnc_SetInputBuf(CLzmaEnc *p, const Byte *src, SizeT srcLen)
}
SRes LzmaEnc_MemPrepare(CLzmaEncHandle pp, const Byte *src, SizeT srcLen,
- UInt32 keepWindowSize, ISzAlloc *alloc, ISzAlloc *allocBig)
+ UInt32 keepWindowSize, ISzAllocPtr alloc, ISzAllocPtr allocBig)
{
CLzmaEnc *p = (CLzmaEnc *)pp;
LzmaEnc_SetInputBuf(p, src, srcLen);
p->needInit = 1;
+ LzmaEnc_SetDataSize(pp, srcLen);
return LzmaEnc_AllocAndInit(p, keepWindowSize, alloc, allocBig);
}
@@ -2152,15 +2578,15 @@ void LzmaEnc_Finish(CLzmaEncHandle pp)
typedef struct
{
- ISeqOutStream funcTable;
+ ISeqOutStream vt;
Byte *data;
SizeT rem;
Bool overflow;
-} CSeqOutStreamBuf;
+} CLzmaEnc_SeqOutStreamBuf;
-static size_t MyWrite(void *pp, const void *data, size_t size)
+static size_t SeqOutStreamBuf_Write(const ISeqOutStream *pp, const void *data, size_t size)
{
- CSeqOutStreamBuf *p = (CSeqOutStreamBuf *)pp;
+ CLzmaEnc_SeqOutStreamBuf *p = CONTAINER_FROM_VTBL(pp, CLzmaEnc_SeqOutStreamBuf, vt);
if (p->rem < size)
{
size = p->rem;
@@ -2193,9 +2619,9 @@ SRes LzmaEnc_CodeOneMemBlock(CLzmaEncHandle pp, Bool reInit,
CLzmaEnc *p = (CLzmaEnc *)pp;
UInt64 nowPos64;
SRes res;
- CSeqOutStreamBuf outStream;
+ CLzmaEnc_SeqOutStreamBuf outStream;
- outStream.funcTable.Write = MyWrite;
+ outStream.vt.Write = SeqOutStreamBuf_Write;
outStream.data = dest;
outStream.rem = *destLen;
outStream.overflow = False;
@@ -2207,11 +2633,15 @@ SRes LzmaEnc_CodeOneMemBlock(CLzmaEncHandle pp, Bool reInit,
if (reInit)
LzmaEnc_Init(p);
LzmaEnc_InitPrices(p);
+
nowPos64 = p->nowPos64;
RangeEnc_Init(&p->rc);
- p->rc.outStream = &outStream.funcTable;
+ p->rc.outStream = &outStream.vt;
+
+ if (desiredPackSize == 0)
+ return SZ_ERROR_OUTPUT_EOF;
- res = LzmaEnc_CodeOneBlock(p, True, desiredPackSize, *unpackSize);
+ res = LzmaEnc_CodeOneBlock(p, desiredPackSize, *unpackSize);
*unpackSize = (UInt32)(p->nowPos64 - nowPos64);
*destLen -= outStream.rem;
@@ -2234,12 +2664,12 @@ static SRes LzmaEnc_Encode2(CLzmaEnc *p, ICompressProgress *progress)
for (;;)
{
- res = LzmaEnc_CodeOneBlock(p, False, 0, 0);
+ res = LzmaEnc_CodeOneBlock(p, 0, 0);
if (res != SZ_OK || p->finished)
break;
if (progress)
{
- res = progress->Progress(progress, p->nowPos64, RangeEnc_GetProcessed(&p->rc));
+ res = ICompressProgress_Progress(progress, p->nowPos64, RangeEnc_GetProcessed(&p->rc));
if (res != SZ_OK)
{
res = SZ_ERROR_PROGRESS;
@@ -2251,7 +2681,7 @@ static SRes LzmaEnc_Encode2(CLzmaEnc *p, ICompressProgress *progress)
LzmaEnc_Finish(p);
/*
- if (res == S_OK && !Inline_MatchFinder_IsFinishedOK(&p->matchFinderBase))
+ if (res == SZ_OK && !Inline_MatchFinder_IsFinishedOK(&p->matchFinderBase))
res = SZ_ERROR_FAIL;
}
*/
@@ -2261,7 +2691,7 @@ static SRes LzmaEnc_Encode2(CLzmaEnc *p, ICompressProgress *progress)
SRes LzmaEnc_Encode(CLzmaEncHandle pp, ISeqOutStream *outStream, ISeqInStream *inStream, ICompressProgress *progress,
- ISzAlloc *alloc, ISzAlloc *allocBig)
+ ISzAllocPtr alloc, ISzAllocPtr allocBig)
{
RINOK(LzmaEnc_Prepare(pp, outStream, inStream, alloc, allocBig));
return LzmaEnc_Encode2((CLzmaEnc *)pp, progress);
@@ -2296,21 +2726,27 @@ SRes LzmaEnc_WriteProperties(CLzmaEncHandle pp, Byte *props, SizeT *size)
}
+unsigned LzmaEnc_IsWriteEndMark(CLzmaEncHandle pp)
+{
+ return ((CLzmaEnc *)pp)->writeEndMark;
+}
+
+
SRes LzmaEnc_MemEncode(CLzmaEncHandle pp, Byte *dest, SizeT *destLen, const Byte *src, SizeT srcLen,
- int writeEndMark, ICompressProgress *progress, ISzAlloc *alloc, ISzAlloc *allocBig)
+ int writeEndMark, ICompressProgress *progress, ISzAllocPtr alloc, ISzAllocPtr allocBig)
{
SRes res;
CLzmaEnc *p = (CLzmaEnc *)pp;
- CSeqOutStreamBuf outStream;
+ CLzmaEnc_SeqOutStreamBuf outStream;
- outStream.funcTable.Write = MyWrite;
+ outStream.vt.Write = SeqOutStreamBuf_Write;
outStream.data = dest;
outStream.rem = *destLen;
outStream.overflow = False;
p->writeEndMark = writeEndMark;
- p->rc.outStream = &outStream.funcTable;
+ p->rc.outStream = &outStream.vt;
res = LzmaEnc_MemPrepare(pp, src, srcLen, 0, alloc, allocBig);
@@ -2330,7 +2766,7 @@ SRes LzmaEnc_MemEncode(CLzmaEncHandle pp, Byte *dest, SizeT *destLen, const Byte
SRes LzmaEncode(Byte *dest, SizeT *destLen, const Byte *src, SizeT srcLen,
const CLzmaEncProps *props, Byte *propsEncoded, SizeT *propsSize, int writeEndMark,
- ICompressProgress *progress, ISzAlloc *alloc, ISzAlloc *allocBig)
+ ICompressProgress *progress, ISzAllocPtr alloc, ISzAllocPtr allocBig)
{
CLzmaEnc *p = (CLzmaEnc *)LzmaEnc_Create(alloc);
SRes res;
diff --git a/BaseTools/Source/C/LzmaCompress/Sdk/C/LzmaEnc.h b/BaseTools/Source/C/LzmaCompress/Sdk/C/LzmaEnc.h
index c2806b45f4..c9938f04bc 100644
--- a/BaseTools/Source/C/LzmaCompress/Sdk/C/LzmaEnc.h
+++ b/BaseTools/Source/C/LzmaCompress/Sdk/C/LzmaEnc.h
@@ -1,5 +1,5 @@
/* LzmaEnc.h -- LZMA Encoder
-2013-01-18 : Igor Pavlov : Public domain */
+2017-07-27 : Igor Pavlov : Public domain */
#ifndef __LZMA_ENC_H
#define __LZMA_ENC_H
@@ -12,12 +12,10 @@ EXTERN_C_BEGIN
typedef struct _CLzmaEncProps
{
- int level; /* 0 <= level <= 9 */
+ int level; /* 0 <= level <= 9 */
UInt32 dictSize; /* (1 << 12) <= dictSize <= (1 << 27) for 32-bit version
- (1 << 12) <= dictSize <= (1 << 30) for 64-bit version
- default = (1 << 24) */
- UInt64 reduceSize; /* estimated size of data that will be compressed. default = 0xFFFFFFFF.
- Encoder uses this value to reduce dictionary size */
+ (1 << 12) <= dictSize <= (3 << 29) for 64-bit version
+ default = (1 << 24) */
int lc; /* 0 <= lc <= 8, default = 3 */
int lp; /* 0 <= lp <= 4, default = 0 */
int pb; /* 0 <= pb <= 4, default = 2 */
@@ -25,9 +23,12 @@ typedef struct _CLzmaEncProps
int fb; /* 5 <= fb <= 273, default = 32 */
int btMode; /* 0 - hashChain Mode, 1 - binTree mode - normal, default = 1 */
int numHashBytes; /* 2, 3 or 4, default = 4 */
- UInt32 mc; /* 1 <= mc <= (1 << 30), default = 32 */
+ UInt32 mc; /* 1 <= mc <= (1 << 30), default = 32 */
unsigned writeEndMark; /* 0 - do not write EOPM, 1 - write EOPM, default = 0 */
int numThreads; /* 1 or 2, default = 2 */
+
+ UInt64 reduceSize; /* estimated size of data that will be compressed. default = (UInt64)(Int64)-1.
+ Encoder uses this value to reduce dictionary size */
} CLzmaEncProps;
void LzmaEncProps_Init(CLzmaEncProps *p);
@@ -37,41 +38,38 @@ UInt32 LzmaEncProps_GetDictSize(const CLzmaEncProps *props2);
/* ---------- CLzmaEncHandle Interface ---------- */
-/* LzmaEnc_* functions can return the following exit codes:
-Returns:
+/* LzmaEnc* functions can return the following exit codes:
+SRes:
SZ_OK - OK
SZ_ERROR_MEM - Memory allocation error
SZ_ERROR_PARAM - Incorrect paramater in props
- SZ_ERROR_WRITE - Write callback error.
+ SZ_ERROR_WRITE - ISeqOutStream write callback error
+ SZ_ERROR_OUTPUT_EOF - output buffer overflow - version with (Byte *) output
SZ_ERROR_PROGRESS - some break from progress callback
- SZ_ERROR_THREAD - errors in multithreading functions (only for Mt version)
+ SZ_ERROR_THREAD - error in multithreading functions (only for Mt version)
*/
typedef void * CLzmaEncHandle;
-CLzmaEncHandle LzmaEnc_Create(ISzAlloc *alloc);
-void LzmaEnc_Destroy(CLzmaEncHandle p, ISzAlloc *alloc, ISzAlloc *allocBig);
+CLzmaEncHandle LzmaEnc_Create(ISzAllocPtr alloc);
+void LzmaEnc_Destroy(CLzmaEncHandle p, ISzAllocPtr alloc, ISzAllocPtr allocBig);
+
SRes LzmaEnc_SetProps(CLzmaEncHandle p, const CLzmaEncProps *props);
+void LzmaEnc_SetDataSize(CLzmaEncHandle p, UInt64 expectedDataSiize);
SRes LzmaEnc_WriteProperties(CLzmaEncHandle p, Byte *properties, SizeT *size);
+unsigned LzmaEnc_IsWriteEndMark(CLzmaEncHandle p);
+
SRes LzmaEnc_Encode(CLzmaEncHandle p, ISeqOutStream *outStream, ISeqInStream *inStream,
- ICompressProgress *progress, ISzAlloc *alloc, ISzAlloc *allocBig);
+ ICompressProgress *progress, ISzAllocPtr alloc, ISzAllocPtr allocBig);
SRes LzmaEnc_MemEncode(CLzmaEncHandle p, Byte *dest, SizeT *destLen, const Byte *src, SizeT srcLen,
- int writeEndMark, ICompressProgress *progress, ISzAlloc *alloc, ISzAlloc *allocBig);
+ int writeEndMark, ICompressProgress *progress, ISzAllocPtr alloc, ISzAllocPtr allocBig);
-/* ---------- One Call Interface ---------- */
-/* LzmaEncode
-Return code:
- SZ_OK - OK
- SZ_ERROR_MEM - Memory allocation error
- SZ_ERROR_PARAM - Incorrect paramater
- SZ_ERROR_OUTPUT_EOF - output buffer overflow
- SZ_ERROR_THREAD - errors in multithreading functions (only for Mt version)
-*/
+/* ---------- One Call Interface ---------- */
SRes LzmaEncode(Byte *dest, SizeT *destLen, const Byte *src, SizeT srcLen,
const CLzmaEncProps *props, Byte *propsEncoded, SizeT *propsSize, int writeEndMark,
- ICompressProgress *progress, ISzAlloc *alloc, ISzAlloc *allocBig);
+ ICompressProgress *progress, ISzAllocPtr alloc, ISzAllocPtr allocBig);
EXTERN_C_END
diff --git a/BaseTools/Source/C/LzmaCompress/Sdk/C/Threads.c b/BaseTools/Source/C/LzmaCompress/Sdk/C/Threads.c
index ece07e618a..8fd86f224b 100644
--- a/BaseTools/Source/C/LzmaCompress/Sdk/C/Threads.c
+++ b/BaseTools/Source/C/LzmaCompress/Sdk/C/Threads.c
@@ -1,5 +1,5 @@
/* Threads.c -- multithreading library
-2014-09-21 : Igor Pavlov : Public domain */
+2017-06-26 : Igor Pavlov : Public domain */
#include "Precomp.h"
@@ -12,18 +12,20 @@
static WRes GetError()
{
DWORD res = GetLastError();
- return (res) ? (WRes)(res) : 1;
+ return res ? (WRes)res : 1;
}
-WRes HandleToWRes(HANDLE h) { return (h != 0) ? 0 : GetError(); }
-WRes BOOLToWRes(BOOL v) { return v ? 0 : GetError(); }
+static WRes HandleToWRes(HANDLE h) { return (h != NULL) ? 0 : GetError(); }
+static WRes BOOLToWRes(BOOL v) { return v ? 0 : GetError(); }
WRes HandlePtr_Close(HANDLE *p)
{
if (*p != NULL)
+ {
if (!CloseHandle(*p))
return GetError();
- *p = NULL;
+ *p = NULL;
+ }
return 0;
}
@@ -49,7 +51,7 @@ WRes Thread_Create(CThread *p, THREAD_FUNC_TYPE func, LPVOID param)
return HandleToWRes(*p);
}
-WRes Event_Create(CEvent *p, BOOL manualReset, int signaled)
+static WRes Event_Create(CEvent *p, BOOL manualReset, int signaled)
{
*p = CreateEvent(NULL, manualReset, (signaled ? TRUE : FALSE), NULL);
return HandleToWRes(*p);
diff --git a/BaseTools/Source/C/LzmaCompress/Sdk/C/Threads.h b/BaseTools/Source/C/LzmaCompress/Sdk/C/Threads.h
index e927208d79..f913241aea 100644
--- a/BaseTools/Source/C/LzmaCompress/Sdk/C/Threads.h
+++ b/BaseTools/Source/C/LzmaCompress/Sdk/C/Threads.h
@@ -1,5 +1,5 @@
/* Threads.h -- multithreading library
-2013-11-12 : Igor Pavlov : Public domain */
+2017-06-18 : Igor Pavlov : Public domain */
#ifndef __7Z_THREADS_H
#define __7Z_THREADS_H
@@ -49,7 +49,8 @@ WRes AutoResetEvent_Create(CAutoResetEvent *p, int signaled);
WRes AutoResetEvent_CreateNotSignaled(CAutoResetEvent *p);
typedef HANDLE CSemaphore;
-#define Semaphore_Construct(p) (*p) = NULL
+#define Semaphore_Construct(p) *(p) = NULL
+#define Semaphore_IsCreated(p) (*(p) != NULL)
#define Semaphore_Close(p) HandlePtr_Close(p)
#define Semaphore_Wait(p) Handle_WaitObject(*(p))
WRes Semaphore_Create(CSemaphore *p, UInt32 initCount, UInt32 maxCount);
diff --git a/BaseTools/Source/C/LzmaCompress/Sdk/DOC/lzma-history.txt b/BaseTools/Source/C/LzmaCompress/Sdk/DOC/lzma-history.txt
index 7aaeb07ff4..f4793264a5 100644
--- a/BaseTools/Source/C/LzmaCompress/Sdk/DOC/lzma-history.txt
+++ b/BaseTools/Source/C/LzmaCompress/Sdk/DOC/lzma-history.txt
@@ -1,6 +1,67 @@
HISTORY of the LZMA SDK
-----------------------
+18.05 2018-04-30
+-------------------------
+- The speed for LZMA/LZMA2 compressing was increased
+ by 8% for fastest/fast compression levels and
+ by 3% for normal/maximum compression levels.
+- Previous versions of 7-Zip could work incorrectly in "Large memory pages" mode in
+ Windows 10 because of some BUG with "Large Pages" in Windows 10.
+ Now 7-Zip doesn't use "Large Pages" on Windows 10 up to revision 1709 (16299).
+- The BUG was fixed in Lzma2Enc.c
+ Lzma2Enc_Encode2() function worked incorretly,
+ if (inStream == NULL) and the number of block threads is more than 1.
+
+
+18.03 beta 2018-03-04
+-------------------------
+- Asm\x86\LzmaDecOpt.asm: new optimized LZMA decoder written in asm
+ for x64 with about 30% higher speed than main version of LZMA decoder written in C.
+- The speed for single-thread LZMA/LZMA2 decoder written in C was increased by 3%.
+- 7-Zip now can use multi-threading for 7z/LZMA2 decoding,
+ if there are multiple independent data chunks in LZMA2 stream.
+- 7-Zip now can use multi-threading for xz decoding,
+ if there are multiple blocks in xz stream.
+
+
+18.01 2019-01-28
+-------------------------
+- The BUG in 17.01 - 18.00 beta was fixed:
+ XzDec.c : random block unpacking and XzUnpacker_IsBlockFinished()
+ didn't work correctly for xz archives without checksum (CRC).
+
+
+18.00 beta 2019-01-10
+-------------------------
+- The BUG in xz encoder was fixed:
+ There was memory leak of 16 KB for each file compressed with
+ xz compression method, if additional filter was used.
+
+
+17.01 beta 2017-08-28
+-------------------------
+- Minor speed optimization for LZMA2 (xz and 7z) multi-threading compression.
+ 7-Zip now uses additional memory buffers for multi-block LZMA2 compression.
+ CPU utilization was slightly improved.
+- 7-zip now creates multi-block xz archives by default. Block size can be
+ specified with -ms[Size]{m|g} switch.
+- xz decoder now can unpack random block from multi-block xz archives.
+- 7-Zip command line: @listfile now doesn't work after -- switch.
+ Use -i@listfile before -- switch instead.
+- The BUGs were fixed:
+ 7-Zip 17.00 beta crashed for commands that write anti-item to 7z archive.
+
+
+17.00 beta 2017-04-29
+-------------------------
+- NewHandler.h / NewHandler.cpp:
+ now it redefines operator new() only for old MSVC compilers (_MSC_VER < 1900).
+- C/7zTypes.h : the names of variables in interface structures were changed (vt).
+- Some bugs were fixed. 7-Zip could crash in some cases.
+- Some internal changes in code.
+
+
16.04 2016-10-04
-------------------------
- The bug was fixed in DllSecur.c.
@@ -168,7 +229,7 @@ HISTORY of the LZMA SDK
4.57 2007-12-12
-------------------------
-- Speed optimizations in C++ LZMA Decoder.
+- Speed optimizations in ?++ LZMA Decoder.
- Small changes for more compatibility with some C/C++ compilers.
diff --git a/BaseTools/Source/C/LzmaCompress/Sdk/DOC/lzma-sdk.txt b/BaseTools/Source/C/LzmaCompress/Sdk/DOC/lzma-sdk.txt
index 86fef248f4..01521e9398 100644
--- a/BaseTools/Source/C/LzmaCompress/Sdk/DOC/lzma-sdk.txt
+++ b/BaseTools/Source/C/LzmaCompress/Sdk/DOC/lzma-sdk.txt
@@ -1,4 +1,4 @@
-LZMA SDK 16.04
+LZMA SDK 18.05
--------------
LZMA SDK provides the documentation, samples, header files,
--
2.34.1.307.g9b7440fafd-goog