mirror of
https://github.com/perfect-dark-pc-port/perfect_dark.git
synced 2026-10-24 00:02:43 +02:00
Introduce mkrom
This commit is contained in:
Ryan Dwyer committed
1 parent
39a92c5970
commit
91ae15a32d
22 files changed
+2836
-707
No files matched your search
@@ -0,0 +1,7 @@
|
||||
C_FILES = $(wildcard *.c)
|
||||
|
||||
%.o: %.c
|
||||
gcc -O3 %< -o $@
|
||||
|
||||
mkrom: $(C_FILES)
|
||||
gcc -O3 $(C_FILES) -o mkrom
|
||||
@@ -0,0 +1,152 @@
|
||||
#include <arpa/inet.h>
|
||||
#include <memory.h>
|
||||
#include <stdint.h>
|
||||
#include <stdio.h>
|
||||
#include <stdlib.h>
|
||||
#include "mkrom.h"
|
||||
|
||||
extern struct state state;
|
||||
|
||||
/**
|
||||
* This file handles creation of the gamezips segment.
|
||||
*
|
||||
* Before mkrom is called, the game segment (uncompressed) is placed past the
|
||||
* end of the ROM by ld. This segment must be split into 4KB chunks. Each chunk
|
||||
* is zipped and then placed in its final location within the ROM.
|
||||
*
|
||||
* The format of the gamezips segment is:
|
||||
* - Array of offsets to each chunk, where each offset is 4 bytes and relative
|
||||
* to the start of the gamezips segment.
|
||||
* - A final offset that points to the end of the final chunk.
|
||||
* - Chunk data, where each chunk consists of:
|
||||
* - A 2 byte checksum of the uncompressed chunk.
|
||||
* - Zip data (starting with 0x1173001000).
|
||||
* - Optional single byte to align it to the next 2 byte boundary.
|
||||
* The added byte is data from the previous chunk.
|
||||
*/
|
||||
|
||||
/**
|
||||
* Calculate the checksum of this chunk's raw data.
|
||||
*
|
||||
* The game never verifies this, but it exists in the ROM so it has to be
|
||||
* calculated by mkrom.
|
||||
*
|
||||
* It's a simple sum of each word, but then it gets stored as a short so the
|
||||
* upper half is lost.
|
||||
*/
|
||||
static uint32_t crc(uint8_t *buffer, size_t len)
|
||||
{
|
||||
uint32_t sum = 0;
|
||||
uint32_t offset;
|
||||
|
||||
for (offset = 0; offset < len; offset += 4) {
|
||||
sum += ntohl(*(uint32_t *) &buffer[offset]);
|
||||
}
|
||||
|
||||
return sum;
|
||||
}
|
||||
|
||||
/**
|
||||
* Create a game chunk. We just calculate and prepend the checksum, then call
|
||||
* rarezip() which does the zipping and adding of the 0x1173 header.
|
||||
*/
|
||||
static void create_chunk(uint8_t *outbuf, size_t *outlen, uint8_t *inbuf, size_t inlen)
|
||||
{
|
||||
uint32_t sum = crc(inbuf, inlen);
|
||||
|
||||
outbuf[0] = (sum >> 8) & 0xff;
|
||||
outbuf[1] = sum & 0xff;
|
||||
|
||||
rarezip(&outbuf[2], outlen, inbuf, inlen, state.zipmagic);
|
||||
|
||||
*outlen += 2;
|
||||
}
|
||||
|
||||
/**
|
||||
* Generate the gamezips segment.
|
||||
*
|
||||
* This segment starts with an offset table followed by the chunk data.
|
||||
* However, we need to keep the offset table zeroed here and build the table
|
||||
* in a different allocation because the ROM is packed with a duplicate of
|
||||
* this segment which has an empty offset table.
|
||||
*
|
||||
* So this function creates:
|
||||
* - state.gamezips, which is big enough to hold the offset table plus zips,
|
||||
* but has the offset table zeroed
|
||||
* - state.gametable, which is just the table and has the entries populated.
|
||||
*
|
||||
* Each chunk is aligned to an even byte. If alignment needs to occur, the extra
|
||||
* byte is taken from the same offset in the previous chunk. In other words, the
|
||||
* same output buffer is used for every chunk and it's not cleared between uses.
|
||||
*/
|
||||
void game_zip(void)
|
||||
{
|
||||
uint32_t end;
|
||||
size_t len;
|
||||
size_t num_chunks;
|
||||
uint32_t tableoffset;
|
||||
uint32_t dataoffset;
|
||||
size_t len_remaining;
|
||||
uint8_t outscratch[0x1100];
|
||||
uint32_t offset;
|
||||
uint32_t tablelen;
|
||||
|
||||
memset(outscratch, 0, 0x1000);
|
||||
|
||||
// Find the game's position in the ROM and calculate the number of chunks
|
||||
map_get_segment_rompos("game", &offset, &end);
|
||||
|
||||
len = end - offset;
|
||||
num_chunks = len / 0x1000;
|
||||
|
||||
if (len % 0x1000) {
|
||||
num_chunks++;
|
||||
}
|
||||
|
||||
// Allocate buffers
|
||||
tablelen = num_chunks * 4 + 4;
|
||||
|
||||
state.gamezips = malloc(tablelen + len);
|
||||
state.gametable = malloc(tablelen);
|
||||
state.gametablelen = tablelen;
|
||||
|
||||
memset(state.gamezips, 0, tablelen);
|
||||
|
||||
tableoffset = 0;
|
||||
dataoffset = tablelen;
|
||||
|
||||
len_remaining = len;
|
||||
|
||||
// Generate the chunks
|
||||
while (offset < end) {
|
||||
size_t chunkoriglen = len_remaining >= 0x1000 ? 0x1000 : len_remaining;
|
||||
size_t chunklen;
|
||||
|
||||
// Write the table entry
|
||||
state.gametable[tableoffset + 0] = (dataoffset >> 24) & 0xff;
|
||||
state.gametable[tableoffset + 1] = (dataoffset >> 16) & 0xff;
|
||||
state.gametable[tableoffset + 2] = (dataoffset >> 8) & 0xff;
|
||||
state.gametable[tableoffset + 3] = dataoffset & 0xff;
|
||||
tableoffset += 4;
|
||||
|
||||
// Write the data
|
||||
create_chunk(outscratch, &chunklen, &state.rom[offset], chunkoriglen);
|
||||
|
||||
chunklen += chunklen % 2;
|
||||
|
||||
memcpy(&state.gamezips[dataoffset], outscratch, chunklen);
|
||||
|
||||
dataoffset += chunklen;
|
||||
|
||||
len_remaining -= 0x1000;
|
||||
offset += 0x1000;
|
||||
}
|
||||
|
||||
// The table contains an additional pointer to the end of the final chunk
|
||||
state.gametable[tableoffset + 0] = (dataoffset >> 24) & 0xff;
|
||||
state.gametable[tableoffset + 1] = (dataoffset >> 16) & 0xff;
|
||||
state.gametable[tableoffset + 2] = (dataoffset >> 8) & 0xff;
|
||||
state.gametable[tableoffset + 3] = dataoffset & 0xff;
|
||||
|
||||
state.gamezipslen = dataoffset;
|
||||
}
|
||||
@@ -0,0 +1,178 @@
|
||||
/* gzip.h -- common declarations for all gzip modules
|
||||
* Copyright (C) 1992-1993 Jean-loup Gailly.
|
||||
* This is free software; you can redistribute it and/or modify it under the
|
||||
* terms of the GNU General Public License, see the file COPYING.
|
||||
*/
|
||||
|
||||
#if defined(__STDC__)
|
||||
# define OF(args) args
|
||||
#else
|
||||
# define OF(args) ()
|
||||
#endif
|
||||
|
||||
#ifdef __STDC__
|
||||
typedef void *voidp;
|
||||
#else
|
||||
typedef char *voidp;
|
||||
#endif
|
||||
|
||||
/* I don't like nested includes, but the string and io functions are used
|
||||
* too often
|
||||
*/
|
||||
#include <stdio.h>
|
||||
#if !defined(NO_STRING_H) || defined(STDC_HEADERS)
|
||||
# include <string.h>
|
||||
# if !defined(STDC_HEADERS) && !defined(NO_MEMORY_H) && !defined(__GNUC__)
|
||||
# include <memory.h>
|
||||
# endif
|
||||
# define memzero(s, n) memset ((voidp)(s), 0, (n))
|
||||
#else
|
||||
# include <strings.h>
|
||||
# define strchr index
|
||||
# define strrchr rindex
|
||||
# define memcpy(d, s, n) bcopy((s), (d), (n))
|
||||
# define memcmp(s1, s2, n) bcmp((s1), (s2), (n))
|
||||
# define memzero(s, n) bzero((s), (n))
|
||||
#endif
|
||||
|
||||
#ifndef RETSIGTYPE
|
||||
# define RETSIGTYPE void
|
||||
#endif
|
||||
|
||||
#define local static
|
||||
|
||||
typedef unsigned char uch;
|
||||
typedef unsigned short ush;
|
||||
typedef unsigned long ulg;
|
||||
|
||||
/* Compression methods (see algorithm.doc) */
|
||||
#define STORED 0
|
||||
#define COMPRESSED 1
|
||||
#define PACKED 2
|
||||
#define LZHED 3
|
||||
/* methods 4 to 7 reserved */
|
||||
#define DEFLATED 8
|
||||
#define MAX_METHODS 9
|
||||
|
||||
#ifndef INBUFSIZ
|
||||
# ifdef SMALL_MEM
|
||||
# define INBUFSIZ 0x2000 /* input buffer size */
|
||||
# else
|
||||
# define INBUFSIZ 0x8000 /* input buffer size */
|
||||
# endif
|
||||
#endif
|
||||
#define INBUF_EXTRA 64 /* required by unlzw() */
|
||||
|
||||
#ifndef OUTBUFSIZ
|
||||
# ifdef SMALL_MEM
|
||||
# define OUTBUFSIZ 8192 /* output buffer size */
|
||||
# else
|
||||
# define OUTBUFSIZ 16384 /* output buffer size */
|
||||
# endif
|
||||
#endif
|
||||
#define OUTBUF_EXTRA 2048 /* required by unlzw() */
|
||||
|
||||
#ifndef DIST_BUFSIZE
|
||||
# ifdef SMALL_MEM
|
||||
# define DIST_BUFSIZE 0x2000 /* buffer for distances, see trees.c */
|
||||
# else
|
||||
# define DIST_BUFSIZE 0x8000 /* buffer for distances, see trees.c */
|
||||
# endif
|
||||
#endif
|
||||
|
||||
#define EXTERN(type, array) extern type array[]
|
||||
#define DECLARE(type, array, size) type array[size]
|
||||
#define ALLOC(type, array, size)
|
||||
#define FREE(array)
|
||||
|
||||
EXTERN(uch, inbuf); /* input buffer */
|
||||
EXTERN(uch, outbuf); /* output buffer */
|
||||
EXTERN(ush, d_buf); /* buffer for distances, see trees.c */
|
||||
EXTERN(uch, window); /* Sliding window and suffix table (unlzw) */
|
||||
#define tab_suffix window
|
||||
#ifndef MAXSEG_64K
|
||||
# define tab_prefix prev /* hash link (see deflate.c) */
|
||||
# define head (prev+WSIZE) /* hash head (see deflate.c) */
|
||||
EXTERN(ush, tab_prefix); /* prefix code (see unlzw.c) */
|
||||
#else
|
||||
# define tab_prefix0 prev
|
||||
# define head tab_prefix1
|
||||
EXTERN(ush, tab_prefix0); /* prefix for even codes */
|
||||
EXTERN(ush, tab_prefix1); /* prefix for odd codes */
|
||||
#endif
|
||||
|
||||
extern unsigned outcnt; /* bytes in output buffer */
|
||||
|
||||
/* for compatibility with old zip sources (to be cleaned) */
|
||||
|
||||
typedef int file_t; /* Do not use stdio */
|
||||
#define NO_FILE (-1) /* in memory compression */
|
||||
|
||||
/* internal file attribute */
|
||||
#define UNKNOWN 0xffff
|
||||
#define BINARY 0
|
||||
#define ASCII 1
|
||||
|
||||
#ifndef WSIZE
|
||||
# define WSIZE 0x8000 /* window size--must be a power of two, and */
|
||||
#endif /* at least 32K for zip's deflate method */
|
||||
|
||||
#define MIN_MATCH 3
|
||||
#define MAX_MATCH 258
|
||||
/* The minimum and maximum match lengths */
|
||||
|
||||
#define MIN_LOOKAHEAD (MAX_MATCH+MIN_MATCH+1)
|
||||
/* Minimum amount of lookahead, except at the end of the input file.
|
||||
* See deflate.c for comments about the MIN_MATCH+1.
|
||||
*/
|
||||
|
||||
#define MAX_DIST (WSIZE-MIN_LOOKAHEAD)
|
||||
/* In order to simplify the code, particularly on 16 bit machines, match
|
||||
* distances are limited to MAX_DIST instead of WSIZE.
|
||||
*/
|
||||
|
||||
/* put_byte is used for the compressed output, put_ubyte for the
|
||||
* uncompressed output. However unlzw() uses window for its
|
||||
* suffix table instead of its output buffer, so it does not use put_ubyte
|
||||
* (to be cleaned up).
|
||||
*/
|
||||
#define put_byte(c) {outbuf[outcnt++]=(uch)(c); if (outcnt==OUTBUFSIZ)\
|
||||
flush_outbuf();}
|
||||
|
||||
/* Output a 16 bit value, lsb first */
|
||||
#define put_short(w) \
|
||||
{ if (outcnt < OUTBUFSIZ-2) { \
|
||||
outbuf[outcnt++] = (uch) ((w) & 0xff); \
|
||||
outbuf[outcnt++] = (uch) ((ush)(w) >> 8); \
|
||||
} else { \
|
||||
put_byte((uch)((w) & 0xff)); \
|
||||
put_byte((uch)((ush)(w) >> 8)); \
|
||||
} \
|
||||
}
|
||||
|
||||
#define seekable() 0 /* force sequential output */
|
||||
|
||||
/* in gzip.c */
|
||||
RETSIGTYPE abort_gzip OF((void));
|
||||
|
||||
/* in deflate.c */
|
||||
void lm_init OF((void));
|
||||
ulg deflate OF((void));
|
||||
|
||||
/* in trees.c */
|
||||
void ct_init OF((void));
|
||||
int ct_tally OF((int dist, int lc));
|
||||
ulg flush_block OF((char *buf, ulg stored_len, int eof));
|
||||
|
||||
/* in bits.c */
|
||||
void bi_init OF((void));
|
||||
void send_bits OF((int value, int length));
|
||||
unsigned bi_reverse OF((unsigned value, int length));
|
||||
void bi_windup OF((void));
|
||||
void copy_block OF((char *buf, unsigned len, int header));
|
||||
extern int read_buf OF((char *buf, unsigned size));
|
||||
|
||||
/* in util.c: */
|
||||
extern void flush_outbuf OF((void));
|
||||
extern void write_buf OF((voidp buf, unsigned cnt));
|
||||
extern void warn OF((char *a, char *b));
|
||||
@@ -0,0 +1,118 @@
|
||||
/* bits.c -- output variable-length bit strings
|
||||
* Copyright (C) 1992-1993 Jean-loup Gailly
|
||||
* This is free software; you can redistribute it and/or modify it under the
|
||||
* terms of the GNU General Public License, see the file COPYING.
|
||||
*/
|
||||
|
||||
#include "gzip.h"
|
||||
#include "crypt.h"
|
||||
|
||||
/* ===========================================================================
|
||||
* Local data used by the "bit string" routines.
|
||||
*/
|
||||
|
||||
local unsigned short bi_buf;
|
||||
/* Output buffer. bits are inserted starting at the bottom (least significant
|
||||
* bits).
|
||||
*/
|
||||
|
||||
#define Buf_size (8 * 2*sizeof(char))
|
||||
/* Number of bits used within bi_buf. (bi_buf might be implemented on
|
||||
* more than 16 bits on some systems.)
|
||||
*/
|
||||
|
||||
local int bi_valid;
|
||||
/* Number of valid bits in bi_buf. All bits above the last valid bit
|
||||
* are always zero.
|
||||
*/
|
||||
|
||||
/* ===========================================================================
|
||||
* Initialize the bit string routines.
|
||||
*/
|
||||
void bi_init(void)
|
||||
{
|
||||
bi_buf = 0;
|
||||
bi_valid = 0;
|
||||
}
|
||||
|
||||
/* ===========================================================================
|
||||
* Send a value on a given number of bits.
|
||||
* IN assertion: length <= 16 and value fits in length bits.
|
||||
*/
|
||||
void send_bits(value, length)
|
||||
int value; /* value to send */
|
||||
int length; /* number of bits */
|
||||
{
|
||||
/* If not enough room in bi_buf, use (valid) bits from bi_buf and
|
||||
* (16 - bi_valid) bits from value, leaving (width - (16-bi_valid))
|
||||
* unused bits in value.
|
||||
*/
|
||||
if (bi_valid > (int)Buf_size - length) {
|
||||
bi_buf |= (value << bi_valid);
|
||||
put_short(bi_buf);
|
||||
bi_buf = (ush)value >> (Buf_size - bi_valid);
|
||||
bi_valid += length - Buf_size;
|
||||
} else {
|
||||
bi_buf |= value << bi_valid;
|
||||
bi_valid += length;
|
||||
}
|
||||
}
|
||||
|
||||
/* ===========================================================================
|
||||
* Reverse the first len bits of a code, using straightforward code (a faster
|
||||
* method would use a table)
|
||||
* IN assertion: 1 <= len <= 15
|
||||
*/
|
||||
unsigned bi_reverse(code, len)
|
||||
unsigned code; /* the value to invert */
|
||||
int len; /* its bit length */
|
||||
{
|
||||
register unsigned res = 0;
|
||||
do {
|
||||
res |= code & 1;
|
||||
code >>= 1, res <<= 1;
|
||||
} while (--len > 0);
|
||||
return res >> 1;
|
||||
}
|
||||
|
||||
/* ===========================================================================
|
||||
* Write out any remaining bits in an incomplete byte.
|
||||
*/
|
||||
void bi_windup()
|
||||
{
|
||||
if (bi_valid > 8) {
|
||||
put_short(bi_buf);
|
||||
} else if (bi_valid > 0) {
|
||||
put_byte(bi_buf);
|
||||
}
|
||||
bi_buf = 0;
|
||||
bi_valid = 0;
|
||||
}
|
||||
|
||||
/* ===========================================================================
|
||||
* Copy a stored block to the zip file, storing first the length and its
|
||||
* one's complement if requested.
|
||||
*/
|
||||
void copy_block(buf, len, header)
|
||||
char *buf; /* the input data */
|
||||
unsigned len; /* its length */
|
||||
int header; /* true if block header must be written */
|
||||
{
|
||||
bi_windup(); /* align on byte boundary */
|
||||
|
||||
if (header) {
|
||||
put_short((ush)len);
|
||||
put_short((ush)~len);
|
||||
}
|
||||
while (len--) {
|
||||
put_byte(*buf++);
|
||||
}
|
||||
}
|
||||
|
||||
void flush_outbuf()
|
||||
{
|
||||
if (outcnt == 0) return;
|
||||
|
||||
write_buf((char *)outbuf, outcnt);
|
||||
outcnt = 0;
|
||||
}
|
||||
@@ -0,0 +1,481 @@
|
||||
#include <stdio.h>
|
||||
#include "gzip.h"
|
||||
|
||||
/**
|
||||
* This file is based off gzip 1.2.4, but with unused functionality removed.
|
||||
* No functionality has been added or changed.
|
||||
*/
|
||||
|
||||
/* Compile with MEDIUM_MEM to reduce the memory requirements or
|
||||
* with SMALL_MEM to use as little memory as possible. Use BIG_MEM if the
|
||||
* entire input file can be held in memory (not possible on 16 bit systems).
|
||||
* Warning: defining these symbols affects HASH_BITS (see below) and thus
|
||||
* affects the compression ratio. The compressed output
|
||||
* is still correct, and might even be smaller in some cases.
|
||||
*/
|
||||
|
||||
#ifndef HASH_BITS
|
||||
# define HASH_BITS 15
|
||||
/* For portability to 16 bit machines, do not use values above 15. */
|
||||
#endif
|
||||
|
||||
/* To save space (see unlzw.c), we overlay prev+head with tab_prefix and
|
||||
* window with tab_suffix. Check that we can do this:
|
||||
*/
|
||||
#define HASH_SIZE (unsigned)(1<<HASH_BITS)
|
||||
#define HASH_MASK (HASH_SIZE-1)
|
||||
#define WMASK (WSIZE-1)
|
||||
/* HASH_SIZE and WSIZE must be powers of two */
|
||||
|
||||
#define NIL 0
|
||||
/* Tail of hash chains */
|
||||
|
||||
#define FAST 4
|
||||
#define SLOW 2
|
||||
/* speed options for the general purpose bit flag */
|
||||
|
||||
#ifndef TOO_FAR
|
||||
# define TOO_FAR 4096
|
||||
#endif
|
||||
/* Matches of length 3 are discarded if their distance exceeds TOO_FAR */
|
||||
|
||||
/* ===========================================================================
|
||||
* Local data used by the "longest match" routines.
|
||||
*/
|
||||
|
||||
typedef ush Pos;
|
||||
typedef unsigned IPos;
|
||||
/* A Pos is an index in the character window. We use short instead of int to
|
||||
* save space in the various tables. IPos is used only for parameter passing.
|
||||
*/
|
||||
|
||||
/* DECLARE(uch, window, 2L*WSIZE); */
|
||||
/* Sliding window. Input bytes are read into the second half of the window,
|
||||
* and move to the first half later to keep a dictionary of at least WSIZE
|
||||
* bytes. With this organization, matches are limited to a distance of
|
||||
* WSIZE-MAX_MATCH bytes, but this ensures that IO is always
|
||||
* performed with a length multiple of the block size. Also, it limits
|
||||
* the window size to 64K, which is quite useful on MSDOS.
|
||||
* To do: limit the window size to WSIZE+BSZ if SMALL_MEM (the code would
|
||||
* be less efficient).
|
||||
*/
|
||||
|
||||
/* DECLARE(Pos, prev, WSIZE); */
|
||||
/* Link to older string with same hash index. To limit the size of this
|
||||
* array to 64K, this link is maintained only for the last 32K strings.
|
||||
* An index in this array is thus a window index modulo 32K.
|
||||
*/
|
||||
|
||||
/* DECLARE(Pos, head, 1<<HASH_BITS); */
|
||||
/* Heads of the hash chains or NIL. */
|
||||
|
||||
ulg window_size = (ulg)2*WSIZE;
|
||||
/* window size, 2*WSIZE except for MMAP or BIG_MEM, where it is the
|
||||
* input file length plus MIN_LOOKAHEAD.
|
||||
*/
|
||||
|
||||
long block_start;
|
||||
/* window position at the beginning of the current output block. Gets
|
||||
* negative when the window is moved backwards.
|
||||
*/
|
||||
|
||||
local unsigned ins_h; /* hash index of string to be inserted */
|
||||
|
||||
#define H_SHIFT ((HASH_BITS+MIN_MATCH-1)/MIN_MATCH)
|
||||
/* Number of bits by which ins_h and del_h must be shifted at each
|
||||
* input step. It must be such that after MIN_MATCH steps, the oldest
|
||||
* byte no longer takes part in the hash key, that is:
|
||||
* H_SHIFT * MIN_MATCH >= HASH_BITS
|
||||
*/
|
||||
|
||||
unsigned int prev_length;
|
||||
/* Length of the best match at previous step. Matches not greater than this
|
||||
* are discarded. This is used in the lazy match evaluation.
|
||||
*/
|
||||
|
||||
unsigned strstart; /* start of string to insert */
|
||||
unsigned match_start; /* start of matching string */
|
||||
local int eofile; /* flag set at end of input file */
|
||||
unsigned lookahead; /* number of valid bytes ahead in window */
|
||||
|
||||
unsigned max_chain_length;
|
||||
/* To speed up deflation, hash chains are never searched beyond this length.
|
||||
* A higher limit improves compression ratio but degrades the speed.
|
||||
*/
|
||||
|
||||
local unsigned int max_lazy_match;
|
||||
/* Attempt to find a better match only when the current match is strictly
|
||||
* smaller than this value. This mechanism is used only for compression
|
||||
* levels >= 4.
|
||||
*/
|
||||
#define max_insert_length max_lazy_match
|
||||
/* Insert new strings in the hash table only if the match length
|
||||
* is not greater than this length. This saves time but degrades compression.
|
||||
* max_insert_length is used only for compression levels <= 3.
|
||||
*/
|
||||
|
||||
unsigned good_match;
|
||||
/* Use a faster search when the previous match is longer than this */
|
||||
|
||||
|
||||
/* Values for max_lazy_match, good_match and max_chain_length, depending on
|
||||
* the desired pack level (0..9). The values given below have been tuned to
|
||||
* exclude worst case performance for pathological files. Better values may be
|
||||
* found for specific files.
|
||||
*/
|
||||
|
||||
typedef struct config {
|
||||
ush good_length; /* reduce lazy search above this match length */
|
||||
ush max_lazy; /* do not perform lazy search above this match length */
|
||||
ush nice_length; /* quit search above this match length */
|
||||
ush max_chain;
|
||||
} config;
|
||||
|
||||
#ifdef FULL_SEARCH
|
||||
# define nice_match MAX_MATCH
|
||||
#else
|
||||
int nice_match; /* Stop searching when current match exceeds this */
|
||||
#endif
|
||||
|
||||
local config configuration_table[10] = {
|
||||
/* good lazy nice chain */
|
||||
/* 0 */ {0, 0, 0, 0}, /* store only */
|
||||
/* 1 */ {4, 4, 8, 4}, /* maximum speed, no lazy matches */
|
||||
/* 2 */ {4, 5, 16, 8},
|
||||
/* 3 */ {4, 6, 32, 32},
|
||||
|
||||
/* 4 */ {4, 4, 16, 16}, /* lazy matches */
|
||||
/* 5 */ {8, 16, 32, 32},
|
||||
/* 6 */ {8, 16, 128, 128},
|
||||
/* 7 */ {8, 32, 128, 256},
|
||||
/* 8 */ {32, 128, 258, 1024},
|
||||
/* 9 */ {32, 258, 258, 4096}}; /* maximum compression */
|
||||
|
||||
/* Note: the deflate() code requires max_lazy >= MIN_MATCH and max_chain >= 4
|
||||
* For deflate_fast() (levels <= 3) good is ignored and lazy has a different
|
||||
* meaning.
|
||||
*/
|
||||
|
||||
#define EQUAL 0
|
||||
/* result of memcmp for equal strings */
|
||||
|
||||
/* ===========================================================================
|
||||
* Prototypes for local functions.
|
||||
*/
|
||||
local void fill_window OF((void));
|
||||
|
||||
int longest_match OF((IPos cur_match));
|
||||
|
||||
/* ===========================================================================
|
||||
* Update a hash value with the given input byte
|
||||
* IN assertion: all calls to to UPDATE_HASH are made with consecutive
|
||||
* input characters, so that a running hash key can be computed from the
|
||||
* previous key instead of complete recalculation each time.
|
||||
*/
|
||||
#define UPDATE_HASH(h,c) (h = (((h)<<H_SHIFT) ^ (c)) & HASH_MASK)
|
||||
|
||||
/* ===========================================================================
|
||||
* Insert string s in the dictionary and set match_head to the previous head
|
||||
* of the hash chain (the most recent string with same hash key). Return
|
||||
* the previous length of the hash chain.
|
||||
* IN assertion: all calls to to INSERT_STRING are made with consecutive
|
||||
* input characters and the first MIN_MATCH bytes of s are valid
|
||||
* (except for the last MIN_MATCH-1 bytes of the input file).
|
||||
*/
|
||||
#define INSERT_STRING(s, match_head) \
|
||||
(UPDATE_HASH(ins_h, window[(s) + MIN_MATCH-1]), \
|
||||
prev[(s) & WMASK] = match_head = head[ins_h], \
|
||||
head[ins_h] = (s))
|
||||
|
||||
/* ===========================================================================
|
||||
* Initialize the "longest match" routines for a new file
|
||||
*/
|
||||
void lm_init (void)
|
||||
{
|
||||
register unsigned j;
|
||||
|
||||
/* Initialize the hash table. */
|
||||
memzero((char*)head, HASH_SIZE*sizeof(*head));
|
||||
|
||||
/* prev will be initialized on the fly */
|
||||
|
||||
/* Set the default configuration parameters: */
|
||||
max_lazy_match = configuration_table[9].max_lazy;
|
||||
good_match = configuration_table[9].good_length;
|
||||
nice_match = configuration_table[9].nice_length;
|
||||
max_chain_length = configuration_table[9].max_chain;
|
||||
/* ??? reduce max_chain_length for binary files */
|
||||
|
||||
strstart = 0;
|
||||
block_start = 0L;
|
||||
|
||||
lookahead = read_buf((char*)window,
|
||||
sizeof(int) <= 2 ? (unsigned)WSIZE : 2*WSIZE);
|
||||
|
||||
if (lookahead == 0 || lookahead == (unsigned)EOF) {
|
||||
eofile = 1, lookahead = 0;
|
||||
return;
|
||||
}
|
||||
eofile = 0;
|
||||
/* Make sure that we always have enough lookahead. This is important
|
||||
* if input comes from a device such as a tty.
|
||||
*/
|
||||
while (lookahead < MIN_LOOKAHEAD && !eofile) fill_window();
|
||||
|
||||
ins_h = 0;
|
||||
for (j=0; j<MIN_MATCH-1; j++) UPDATE_HASH(ins_h, window[j]);
|
||||
/* If lookahead < MIN_MATCH, ins_h is garbage, but this is
|
||||
* not important since only literal bytes will be emitted.
|
||||
*/
|
||||
}
|
||||
|
||||
/* ===========================================================================
|
||||
* Set match_start to the longest match starting at the given string and
|
||||
* return its length. Matches shorter or equal to prev_length are discarded,
|
||||
* in which case the result is equal to prev_length and match_start is
|
||||
* garbage.
|
||||
* IN assertions: cur_match is the head of the hash chain for the current
|
||||
* string (strstart) and its distance is <= MAX_DIST, and prev_length >= 1
|
||||
*/
|
||||
|
||||
/* For MSDOS, OS/2 and 386 Unix, an optimized version is in match.asm or
|
||||
* match.s. The code is functionally equivalent, so you can use the C version
|
||||
* if desired.
|
||||
*/
|
||||
int longest_match(cur_match)
|
||||
IPos cur_match; /* current match */
|
||||
{
|
||||
unsigned chain_length = max_chain_length; /* max hash chain length */
|
||||
register uch *scan = window + strstart; /* current string */
|
||||
register uch *match; /* matched string */
|
||||
register int len; /* length of current match */
|
||||
int best_len = prev_length; /* best match length so far */
|
||||
IPos limit = strstart > (IPos)MAX_DIST ? strstart - (IPos)MAX_DIST : NIL;
|
||||
|
||||
/* Stop when cur_match becomes <= limit. To simplify the code,
|
||||
* we prevent matches with the string of window index 0.
|
||||
*/
|
||||
|
||||
/* The code is optimized for HASH_BITS >= 8 and MAX_MATCH-2 multiple of 16.
|
||||
* It is easy to get rid of this optimization if necessary.
|
||||
*/
|
||||
register uch *strend = window + strstart + MAX_MATCH;
|
||||
register uch scan_end1 = scan[best_len-1];
|
||||
register uch scan_end = scan[best_len];
|
||||
|
||||
/* Do not waste too much time if we already have a good match: */
|
||||
if (prev_length >= good_match) {
|
||||
chain_length >>= 2;
|
||||
}
|
||||
|
||||
// Iterate backwards through the buffer
|
||||
do {
|
||||
match = window + cur_match;
|
||||
|
||||
/* Skip to next match if the match length cannot increase
|
||||
* or if the match length is less than 2:
|
||||
*/
|
||||
if (match[best_len] != scan_end ||
|
||||
match[best_len-1] != scan_end1 ||
|
||||
*match != *scan ||
|
||||
*++match != scan[1]) {
|
||||
continue;
|
||||
}
|
||||
|
||||
/* The check at best_len-1 can be removed because it will be made
|
||||
* again later. (This heuristic is not always a win.)
|
||||
* It is not necessary to compare scan[2] and match[2] since they
|
||||
* are always equal when the other bytes match, given that
|
||||
* the hash keys are equal and that HASH_BITS >= 8.
|
||||
*/
|
||||
scan += 2, match++;
|
||||
|
||||
/* We check for insufficient lookahead only every 8th comparison;
|
||||
* the 256th check will be made at strstart+258.
|
||||
*/
|
||||
do {
|
||||
} while (*++scan == *++match && *++scan == *++match &&
|
||||
*++scan == *++match && *++scan == *++match &&
|
||||
*++scan == *++match && *++scan == *++match &&
|
||||
*++scan == *++match && *++scan == *++match &&
|
||||
scan < strend);
|
||||
|
||||
len = MAX_MATCH - (int)(strend - scan);
|
||||
|
||||
scan = strend - MAX_MATCH;
|
||||
|
||||
if (len > best_len) {
|
||||
match_start = cur_match;
|
||||
best_len = len;
|
||||
if (len >= nice_match) break;
|
||||
scan_end1 = scan[best_len-1];
|
||||
scan_end = scan[best_len];
|
||||
}
|
||||
} while ((cur_match = prev[cur_match & WMASK]) > limit
|
||||
&& --chain_length != 0);
|
||||
|
||||
return best_len;
|
||||
}
|
||||
|
||||
/* ===========================================================================
|
||||
* Fill the window when the lookahead becomes insufficient.
|
||||
* Updates strstart and lookahead, and sets eofile if end of input file.
|
||||
* IN assertion: lookahead < MIN_LOOKAHEAD && strstart + lookahead > 0
|
||||
* OUT assertions: at least one byte has been read, or eofile is set;
|
||||
* file reads are performed for at least two bytes (required for the
|
||||
* translate_eol option).
|
||||
*/
|
||||
local void fill_window()
|
||||
{
|
||||
register unsigned n, m;
|
||||
unsigned more = (unsigned)(window_size - (ulg)lookahead - (ulg)strstart);
|
||||
/* Amount of free space at the end of the window. */
|
||||
|
||||
/* If the window is almost full and there is insufficient lookahead,
|
||||
* move the upper half to the lower one to make room in the upper half.
|
||||
*/
|
||||
if (more == (unsigned)EOF) {
|
||||
/* Very unlikely, but possible on 16 bit machine if strstart == 0
|
||||
* and lookahead == 1 (input done one byte at time)
|
||||
*/
|
||||
more--;
|
||||
} else if (strstart >= WSIZE+MAX_DIST) {
|
||||
/* By the IN assertion, the window is not empty so we can't confuse
|
||||
* more == 0 with more == 64K on a 16 bit machine.
|
||||
*/
|
||||
memcpy((char*)window, (char*)window+WSIZE, (unsigned)WSIZE);
|
||||
match_start -= WSIZE;
|
||||
strstart -= WSIZE; /* we now have strstart >= MAX_DIST: */
|
||||
|
||||
block_start -= (long) WSIZE;
|
||||
|
||||
for (n = 0; n < HASH_SIZE; n++) {
|
||||
m = head[n];
|
||||
head[n] = (Pos)(m >= WSIZE ? m-WSIZE : NIL);
|
||||
}
|
||||
for (n = 0; n < WSIZE; n++) {
|
||||
m = prev[n];
|
||||
prev[n] = (Pos)(m >= WSIZE ? m-WSIZE : NIL);
|
||||
/* If n is not on any hash chain, prev[n] is garbage but
|
||||
* its value will never be used.
|
||||
*/
|
||||
}
|
||||
more += WSIZE;
|
||||
}
|
||||
/* At this point, more >= 2 */
|
||||
if (!eofile) {
|
||||
n = read_buf((char*)window+strstart+lookahead, more);
|
||||
if (n == 0 || n == (unsigned)EOF) {
|
||||
eofile = 1;
|
||||
} else {
|
||||
lookahead += n;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
/* ===========================================================================
|
||||
* Flush the current block, with given end-of-file flag.
|
||||
* IN assertion: strstart is set to the end of the current match.
|
||||
*/
|
||||
#define FLUSH_BLOCK(eof) \
|
||||
flush_block(block_start >= 0L ? (char*)&window[(unsigned)block_start] : \
|
||||
(char*)NULL, (long)strstart - block_start, (eof))
|
||||
|
||||
/* ===========================================================================
|
||||
* Processes a new input file and return its compressed length. We use a lazy
|
||||
* evaluation for matches: a match is finally adopted only if there is
|
||||
* no better match at the next window position.
|
||||
*/
|
||||
ulg deflate()
|
||||
{
|
||||
IPos hash_head; /* head of hash chain */
|
||||
IPos prev_match; /* previous match */
|
||||
int flush; /* set if current block must be flushed */
|
||||
int match_available = 0; /* set if previous match exists */
|
||||
register unsigned match_length = MIN_MATCH-1; /* length of best match */
|
||||
|
||||
/* Process the input block. */
|
||||
while (lookahead != 0) {
|
||||
/* Insert the string window[strstart .. strstart+2] in the
|
||||
* dictionary, and set hash_head to the head of the hash chain:
|
||||
*/
|
||||
INSERT_STRING(strstart, hash_head);
|
||||
|
||||
/* Find the longest match, discarding those <= prev_length. */
|
||||
prev_length = match_length, prev_match = match_start;
|
||||
match_length = MIN_MATCH-1;
|
||||
|
||||
if (hash_head != NIL && prev_length < max_lazy_match &&
|
||||
strstart - hash_head <= MAX_DIST) {
|
||||
/* To simplify the code, we prevent matches with the string
|
||||
* of window index 0 (in particular we have to avoid a match
|
||||
* of the string with itself at the start of the input file).
|
||||
*/
|
||||
match_length = longest_match (hash_head);
|
||||
/* longest_match() sets match_start */
|
||||
if (match_length > lookahead) match_length = lookahead;
|
||||
|
||||
/* Ignore a length 3 match if it is too distant: */
|
||||
if (match_length == MIN_MATCH && strstart-match_start > TOO_FAR){
|
||||
/* If prev_match is also MIN_MATCH, match_start is garbage
|
||||
* but we will ignore the current match anyway.
|
||||
*/
|
||||
match_length--;
|
||||
}
|
||||
}
|
||||
/* If there was a match at the previous step and the current
|
||||
* match is not better, output the previous match:
|
||||
*/
|
||||
if (prev_length >= MIN_MATCH && match_length <= prev_length) {
|
||||
|
||||
flush = ct_tally(strstart-1-prev_match, prev_length - MIN_MATCH);
|
||||
|
||||
/* Insert in hash table all strings up to the end of the match.
|
||||
* strstart-1 and strstart are already inserted.
|
||||
*/
|
||||
lookahead -= prev_length-1;
|
||||
prev_length -= 2;
|
||||
do {
|
||||
strstart++;
|
||||
INSERT_STRING(strstart, hash_head);
|
||||
/* strstart never exceeds WSIZE-MAX_MATCH, so there are
|
||||
* always MIN_MATCH bytes ahead. If lookahead < MIN_MATCH
|
||||
* these bytes are garbage, but it does not matter since the
|
||||
* next lookahead bytes will always be emitted as literals.
|
||||
*/
|
||||
} while (--prev_length != 0);
|
||||
match_available = 0;
|
||||
match_length = MIN_MATCH-1;
|
||||
strstart++;
|
||||
if (flush) FLUSH_BLOCK(0), block_start = strstart;
|
||||
|
||||
} else if (match_available) {
|
||||
/* If there was no match at the previous position, output a
|
||||
* single literal. If there was a match but the current match
|
||||
* is longer, truncate the previous match to a single literal.
|
||||
*/
|
||||
if (ct_tally (0, window[strstart-1])) {
|
||||
FLUSH_BLOCK(0), block_start = strstart;
|
||||
}
|
||||
strstart++;
|
||||
lookahead--;
|
||||
} else {
|
||||
/* There is no previous match to compare with, wait for
|
||||
* the next step to decide.
|
||||
*/
|
||||
match_available = 1;
|
||||
strstart++;
|
||||
lookahead--;
|
||||
}
|
||||
|
||||
/* Make sure that we always have enough lookahead, except
|
||||
* at the end of the input file. We need MAX_MATCH bytes
|
||||
* for the next match, plus MIN_MATCH bytes to insert the
|
||||
* string following the next match.
|
||||
*/
|
||||
while (lookahead < MIN_LOOKAHEAD && !eofile) fill_window();
|
||||
}
|
||||
if (match_available) ct_tally (0, window[strstart-1]);
|
||||
|
||||
return FLUSH_BLOCK(1); /* eof */
|
||||
}
|
||||
@@ -0,0 +1,911 @@
|
||||
/* trees.c -- output deflated data using Huffman coding
|
||||
* Copyright (C) 1992-1993 Jean-loup Gailly
|
||||
* This is free software; you can redistribute it and/or modify it under the
|
||||
* terms of the GNU General Public License, see the file COPYING.
|
||||
*/
|
||||
|
||||
#include <ctype.h>
|
||||
#include "gzip.h"
|
||||
|
||||
/* ===========================================================================
|
||||
* Constants
|
||||
*/
|
||||
|
||||
#define MAX_BITS 15
|
||||
/* All codes must not exceed MAX_BITS bits */
|
||||
|
||||
#define MAX_BL_BITS 7
|
||||
/* Bit length codes must not exceed MAX_BL_BITS bits */
|
||||
|
||||
#define LENGTH_CODES 29
|
||||
/* number of length codes, not counting the special END_BLOCK code */
|
||||
|
||||
#define LITERALS 256
|
||||
/* number of literal bytes 0..255 */
|
||||
|
||||
#define END_BLOCK 256
|
||||
/* end of block literal code */
|
||||
|
||||
#define L_CODES (LITERALS+1+LENGTH_CODES)
|
||||
/* number of Literal or Length codes, including the END_BLOCK code */
|
||||
|
||||
#define D_CODES 30
|
||||
/* number of distance codes */
|
||||
|
||||
#define BL_CODES 19
|
||||
/* number of codes used to transfer the bit lengths */
|
||||
|
||||
|
||||
local int extra_lbits[LENGTH_CODES] /* extra bits for each length code */
|
||||
= {0,0,0,0,0,0,0,0,1,1,1,1,2,2,2,2,3,3,3,3,4,4,4,4,5,5,5,5,0};
|
||||
|
||||
local int extra_dbits[D_CODES] /* extra bits for each distance code */
|
||||
= {0,0,0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,11,11,12,12,13,13};
|
||||
|
||||
local int extra_blbits[BL_CODES]/* extra bits for each bit length code */
|
||||
= {0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,2,3,7};
|
||||
|
||||
#define STORED_BLOCK 0
|
||||
#define STATIC_TREES 1
|
||||
#define DYN_TREES 2
|
||||
/* The three kinds of block type */
|
||||
|
||||
#ifndef LIT_BUFSIZE
|
||||
# ifdef SMALL_MEM
|
||||
# define LIT_BUFSIZE 0x2000
|
||||
# else
|
||||
# ifdef MEDIUM_MEM
|
||||
# define LIT_BUFSIZE 0x4000
|
||||
# else
|
||||
# define LIT_BUFSIZE 0x8000
|
||||
# endif
|
||||
# endif
|
||||
#endif
|
||||
#ifndef DIST_BUFSIZE
|
||||
# define DIST_BUFSIZE LIT_BUFSIZE
|
||||
#endif
|
||||
/* Sizes of match buffers for literals/lengths and distances. There are
|
||||
* 4 reasons for limiting LIT_BUFSIZE to 64K:
|
||||
* - frequencies can be kept in 16 bit counters
|
||||
* - if compression is not successful for the first block, all input data is
|
||||
* still in the window so we can still emit a stored block even when input
|
||||
* comes from standard input. (This can also be done for all blocks if
|
||||
* LIT_BUFSIZE is not greater than 32K.)
|
||||
* - if compression is not successful for a file smaller than 64K, we can
|
||||
* even emit a stored file instead of a stored block (saving 5 bytes).
|
||||
* - creating new Huffman trees less frequently may not provide fast
|
||||
* adaptation to changes in the input data statistics. (Take for
|
||||
* example a binary file with poorly compressible code followed by
|
||||
* a highly compressible string table.) Smaller buffer sizes give
|
||||
* fast adaptation but have of course the overhead of transmitting trees
|
||||
* more frequently.
|
||||
* - I can't count above 4
|
||||
* The current code is general and allows DIST_BUFSIZE < LIT_BUFSIZE (to save
|
||||
* memory at the expense of compression). Some optimizations would be possible
|
||||
* if we rely on DIST_BUFSIZE == LIT_BUFSIZE.
|
||||
*/
|
||||
#if LIT_BUFSIZE > INBUFSIZ
|
||||
error cannot overlay l_buf and inbuf
|
||||
#endif
|
||||
|
||||
#define REP_3_6 16
|
||||
/* repeat previous bit length 3-6 times (2 bits of repeat count) */
|
||||
|
||||
#define REPZ_3_10 17
|
||||
/* repeat a zero length 3-10 times (3 bits of repeat count) */
|
||||
|
||||
#define REPZ_11_138 18
|
||||
/* repeat a zero length 11-138 times (7 bits of repeat count) */
|
||||
|
||||
/* ===========================================================================
|
||||
* Local data
|
||||
*/
|
||||
|
||||
/* Data structure describing a single value and its code string. */
|
||||
typedef struct ct_data {
|
||||
union {
|
||||
ush freq; /* frequency count */
|
||||
ush code; /* bit string */
|
||||
} fc;
|
||||
union {
|
||||
ush dad; /* father node in Huffman tree */
|
||||
ush len; /* length of bit string */
|
||||
} dl;
|
||||
} ct_data;
|
||||
|
||||
#define Freq fc.freq
|
||||
#define Code fc.code
|
||||
#define Dad dl.dad
|
||||
#define Len dl.len
|
||||
|
||||
#define HEAP_SIZE (2*L_CODES+1)
|
||||
/* maximum heap size */
|
||||
|
||||
local ct_data dyn_ltree[HEAP_SIZE]; /* literal and length tree */
|
||||
local ct_data dyn_dtree[2*D_CODES+1]; /* distance tree */
|
||||
|
||||
local ct_data static_ltree[L_CODES+2];
|
||||
/* The static literal tree. Since the bit lengths are imposed, there is no
|
||||
* need for the L_CODES extra codes used during heap construction. However
|
||||
* The codes 286 and 287 are needed to build a canonical tree (see ct_init
|
||||
* below).
|
||||
*/
|
||||
|
||||
local ct_data static_dtree[D_CODES];
|
||||
/* The static distance tree. (Actually a trivial tree since all codes use
|
||||
* 5 bits.)
|
||||
*/
|
||||
|
||||
local ct_data bl_tree[2*BL_CODES+1];
|
||||
/* Huffman tree for the bit lengths */
|
||||
|
||||
typedef struct tree_desc {
|
||||
ct_data *dyn_tree; /* the dynamic tree */
|
||||
ct_data *static_tree; /* corresponding static tree or NULL */
|
||||
int *extra_bits; /* extra bits for each code or NULL */
|
||||
int extra_base; /* base index for extra_bits */
|
||||
int elems; /* max number of elements in the tree */
|
||||
int max_length; /* max bit length for the codes */
|
||||
int max_code; /* largest code with non zero frequency */
|
||||
} tree_desc;
|
||||
|
||||
local tree_desc l_desc =
|
||||
{dyn_ltree, static_ltree, extra_lbits, LITERALS+1, L_CODES, MAX_BITS, 0};
|
||||
|
||||
local tree_desc d_desc =
|
||||
{dyn_dtree, static_dtree, extra_dbits, 0, D_CODES, MAX_BITS, 0};
|
||||
|
||||
local tree_desc bl_desc =
|
||||
{bl_tree, (ct_data *)0, extra_blbits, 0, BL_CODES, MAX_BL_BITS, 0};
|
||||
|
||||
|
||||
local ush bl_count[MAX_BITS+1];
|
||||
/* number of codes at each bit length for an optimal tree */
|
||||
|
||||
local uch bl_order[BL_CODES]
|
||||
= {16,17,18,0,8,7,9,6,10,5,11,4,12,3,13,2,14,1,15};
|
||||
/* The lengths of the bit length codes are sent in order of decreasing
|
||||
* probability, to avoid transmitting the lengths for unused bit length codes.
|
||||
*/
|
||||
|
||||
local int heap[2*L_CODES+1]; /* heap used to build the Huffman trees */
|
||||
local int heap_len; /* number of elements in the heap */
|
||||
local int heap_max; /* element of largest frequency */
|
||||
/* The sons of heap[n] are heap[2*n] and heap[2*n+1]. heap[0] is not used.
|
||||
* The same heap array is used to build all trees.
|
||||
*/
|
||||
|
||||
local uch depth[2*L_CODES+1];
|
||||
/* Depth of each subtree used as tie breaker for trees of equal frequency */
|
||||
|
||||
local uch length_code[MAX_MATCH-MIN_MATCH+1];
|
||||
/* length code for each normalized match length (0 == MIN_MATCH) */
|
||||
|
||||
local uch dist_code[512];
|
||||
/* distance codes. The first 256 values correspond to the distances
|
||||
* 3 .. 258, the last 256 values correspond to the top 8 bits of
|
||||
* the 15 bit distances.
|
||||
*/
|
||||
|
||||
local int base_length[LENGTH_CODES];
|
||||
/* First normalized length for each code (0 = MIN_MATCH) */
|
||||
|
||||
local int base_dist[D_CODES];
|
||||
/* First normalized distance for each code (0 = distance of 1) */
|
||||
|
||||
#define l_buf inbuf
|
||||
/* DECLARE(uch, l_buf, LIT_BUFSIZE); buffer for literals or lengths */
|
||||
|
||||
/* DECLARE(ush, d_buf, DIST_BUFSIZE); buffer for distances */
|
||||
|
||||
local uch flag_buf[(LIT_BUFSIZE/8)];
|
||||
/* flag_buf is a bit array distinguishing literals from lengths in
|
||||
* l_buf, thus indicating the presence or absence of a distance.
|
||||
*/
|
||||
|
||||
local unsigned last_lit; /* running index in l_buf */
|
||||
local unsigned last_dist; /* running index in d_buf */
|
||||
local unsigned last_flags; /* running index in flag_buf */
|
||||
local uch flags; /* current flags not yet saved in flag_buf */
|
||||
local uch flag_bit; /* current bit used in flags */
|
||||
/* bits are filled in flags starting at bit 0 (least significant).
|
||||
* Note: these flags are overkill in the current code since we don't
|
||||
* take advantage of DIST_BUFSIZE == LIT_BUFSIZE.
|
||||
*/
|
||||
|
||||
local ulg opt_len; /* bit length of current block with optimal trees */
|
||||
local ulg static_len; /* bit length of current block with static trees */
|
||||
|
||||
local ulg compressed_len; /* total bit length of compressed file */
|
||||
|
||||
local ulg input_len; /* total byte length of input file */
|
||||
/* input_len is for debugging only since we can get it by other means. */
|
||||
|
||||
extern long block_start; /* window offset of current block */
|
||||
extern unsigned strstart; /* window offset of current string */
|
||||
|
||||
/* ===========================================================================
|
||||
* Local (static) routines in this file.
|
||||
*/
|
||||
|
||||
local void init_block OF((void));
|
||||
local void pqdownheap OF((ct_data *tree, int k));
|
||||
local void gen_bitlen OF((tree_desc *desc));
|
||||
local void gen_codes OF((ct_data *tree, int max_code));
|
||||
local void build_tree OF((tree_desc *desc));
|
||||
local void scan_tree OF((ct_data *tree, int max_code));
|
||||
local void send_tree OF((ct_data *tree, int max_code));
|
||||
local int build_bl_tree OF((void));
|
||||
local void send_all_trees OF((int lcodes, int dcodes, int blcodes));
|
||||
local void compress_block OF((ct_data *ltree, ct_data *dtree));
|
||||
|
||||
|
||||
#define send_code(c, tree) send_bits(tree[c].Code, tree[c].Len)
|
||||
/* Send a code of the given tree. c and tree must not have side effects */
|
||||
|
||||
#define d_code(dist) \
|
||||
((dist) < 256 ? dist_code[dist] : dist_code[256+((dist)>>7)])
|
||||
/* Mapping from a distance to a distance code. dist is the distance - 1 and
|
||||
* must not have side effects. dist_code[256] and dist_code[257] are never
|
||||
* used.
|
||||
*/
|
||||
|
||||
#define MAX(a,b) (a >= b ? a : b)
|
||||
/* the arguments must not have side effects */
|
||||
|
||||
/* ===========================================================================
|
||||
* Allocate the match buffer, initialize the various tables and save the
|
||||
* location of the internal file attribute (ascii/binary) and method
|
||||
* (DEFLATE/STORE).
|
||||
*/
|
||||
void ct_init(void)
|
||||
{
|
||||
int n; /* iterates over tree elements */
|
||||
int bits; /* bit counter */
|
||||
int length; /* length value */
|
||||
int code; /* code value */
|
||||
int dist; /* distance index */
|
||||
|
||||
compressed_len = input_len = 0L;
|
||||
|
||||
if (static_dtree[0].Len != 0) return; /* ct_init already called */
|
||||
|
||||
/* Initialize the mapping length (0..255) -> length code (0..28) */
|
||||
length = 0;
|
||||
for (code = 0; code < LENGTH_CODES-1; code++) {
|
||||
base_length[code] = length;
|
||||
for (n = 0; n < (1<<extra_lbits[code]); n++) {
|
||||
length_code[length++] = (uch)code;
|
||||
}
|
||||
}
|
||||
/* Note that the length 255 (match length 258) can be represented
|
||||
* in two different ways: code 284 + 5 bits or code 285, so we
|
||||
* overwrite length_code[255] to use the best encoding:
|
||||
*/
|
||||
length_code[length-1] = (uch)code;
|
||||
|
||||
/* Initialize the mapping dist (0..32K) -> dist code (0..29) */
|
||||
dist = 0;
|
||||
for (code = 0 ; code < 16; code++) {
|
||||
base_dist[code] = dist;
|
||||
for (n = 0; n < (1<<extra_dbits[code]); n++) {
|
||||
dist_code[dist++] = (uch)code;
|
||||
}
|
||||
}
|
||||
dist >>= 7; /* from now on, all distances are divided by 128 */
|
||||
for ( ; code < D_CODES; code++) {
|
||||
base_dist[code] = dist << 7;
|
||||
for (n = 0; n < (1<<(extra_dbits[code]-7)); n++) {
|
||||
dist_code[256 + dist++] = (uch)code;
|
||||
}
|
||||
}
|
||||
|
||||
/* Construct the codes of the static literal tree */
|
||||
for (bits = 0; bits <= MAX_BITS; bits++) bl_count[bits] = 0;
|
||||
n = 0;
|
||||
while (n <= 143) static_ltree[n++].Len = 8, bl_count[8]++;
|
||||
while (n <= 255) static_ltree[n++].Len = 9, bl_count[9]++;
|
||||
while (n <= 279) static_ltree[n++].Len = 7, bl_count[7]++;
|
||||
while (n <= 287) static_ltree[n++].Len = 8, bl_count[8]++;
|
||||
/* Codes 286 and 287 do not exist, but we must include them in the
|
||||
* tree construction to get a canonical Huffman tree (longest code
|
||||
* all ones)
|
||||
*/
|
||||
gen_codes((ct_data *)static_ltree, L_CODES+1);
|
||||
|
||||
/* The static distance tree is trivial: */
|
||||
for (n = 0; n < D_CODES; n++) {
|
||||
static_dtree[n].Len = 5;
|
||||
static_dtree[n].Code = bi_reverse(n, 5);
|
||||
}
|
||||
|
||||
/* Initialize the first block of the first file: */
|
||||
init_block();
|
||||
}
|
||||
|
||||
/* ===========================================================================
|
||||
* Initialize a new block.
|
||||
*/
|
||||
local void init_block()
|
||||
{
|
||||
int n; /* iterates over tree elements */
|
||||
|
||||
/* Initialize the trees. */
|
||||
for (n = 0; n < L_CODES; n++) dyn_ltree[n].Freq = 0;
|
||||
for (n = 0; n < D_CODES; n++) dyn_dtree[n].Freq = 0;
|
||||
for (n = 0; n < BL_CODES; n++) bl_tree[n].Freq = 0;
|
||||
|
||||
dyn_ltree[END_BLOCK].Freq = 1;
|
||||
opt_len = static_len = 0L;
|
||||
last_lit = last_dist = last_flags = 0;
|
||||
flags = 0; flag_bit = 1;
|
||||
}
|
||||
|
||||
#define SMALLEST 1
|
||||
/* Index within the heap array of least frequent node in the Huffman tree */
|
||||
|
||||
|
||||
/* ===========================================================================
|
||||
* Remove the smallest element from the heap and recreate the heap with
|
||||
* one less element. Updates heap and heap_len.
|
||||
*/
|
||||
#define pqremove(tree, top) \
|
||||
{\
|
||||
top = heap[SMALLEST]; \
|
||||
heap[SMALLEST] = heap[heap_len--]; \
|
||||
pqdownheap(tree, SMALLEST); \
|
||||
}
|
||||
|
||||
/* ===========================================================================
|
||||
* Compares to subtrees, using the tree depth as tie breaker when
|
||||
* the subtrees have equal frequency. This minimizes the worst case length.
|
||||
*/
|
||||
#define smaller(tree, n, m) \
|
||||
(tree[n].Freq < tree[m].Freq || \
|
||||
(tree[n].Freq == tree[m].Freq && depth[n] <= depth[m]))
|
||||
|
||||
/* ===========================================================================
|
||||
* Restore the heap property by moving down the tree starting at node k,
|
||||
* exchanging a node with the smallest of its two sons if necessary, stopping
|
||||
* when the heap property is re-established (each father smaller than its
|
||||
* two sons).
|
||||
*/
|
||||
local void pqdownheap(tree, k)
|
||||
ct_data *tree; /* the tree to restore */
|
||||
int k; /* node to move down */
|
||||
{
|
||||
int v = heap[k];
|
||||
int j = k << 1; /* left son of k */
|
||||
while (j <= heap_len) {
|
||||
/* Set j to the smallest of the two sons: */
|
||||
if (j < heap_len && smaller(tree, heap[j+1], heap[j])) j++;
|
||||
|
||||
/* Exit if v is smaller than both sons */
|
||||
if (smaller(tree, v, heap[j])) break;
|
||||
|
||||
/* Exchange v with the smallest son */
|
||||
heap[k] = heap[j]; k = j;
|
||||
|
||||
/* And continue down the tree, setting j to the left son of k */
|
||||
j <<= 1;
|
||||
}
|
||||
heap[k] = v;
|
||||
}
|
||||
|
||||
/* ===========================================================================
|
||||
* Compute the optimal bit lengths for a tree and update the total bit length
|
||||
* for the current block.
|
||||
* IN assertion: the fields freq and dad are set, heap[heap_max] and
|
||||
* above are the tree nodes sorted by increasing frequency.
|
||||
* OUT assertions: the field len is set to the optimal bit length, the
|
||||
* array bl_count contains the frequencies for each bit length.
|
||||
* The length opt_len is updated; static_len is also updated if stree is
|
||||
* not null.
|
||||
*/
|
||||
local void gen_bitlen(desc)
|
||||
tree_desc *desc; /* the tree descriptor */
|
||||
{
|
||||
ct_data *tree = desc->dyn_tree;
|
||||
int *extra = desc->extra_bits;
|
||||
int base = desc->extra_base;
|
||||
int max_code = desc->max_code;
|
||||
int max_length = desc->max_length;
|
||||
ct_data *stree = desc->static_tree;
|
||||
int h; /* heap index */
|
||||
int n, m; /* iterate over the tree elements */
|
||||
int bits; /* bit length */
|
||||
int xbits; /* extra bits */
|
||||
ush f; /* frequency */
|
||||
int overflow = 0; /* number of elements with bit length too large */
|
||||
|
||||
for (bits = 0; bits <= MAX_BITS; bits++) bl_count[bits] = 0;
|
||||
|
||||
/* In a first pass, compute the optimal bit lengths (which may
|
||||
* overflow in the case of the bit length tree).
|
||||
*/
|
||||
tree[heap[heap_max]].Len = 0; /* root of the heap */
|
||||
|
||||
for (h = heap_max+1; h < HEAP_SIZE; h++) {
|
||||
n = heap[h];
|
||||
bits = tree[tree[n].Dad].Len + 1;
|
||||
if (bits > max_length) bits = max_length, overflow++;
|
||||
tree[n].Len = (ush)bits;
|
||||
/* We overwrite tree[n].Dad which is no longer needed */
|
||||
|
||||
if (n > max_code) continue; /* not a leaf node */
|
||||
|
||||
bl_count[bits]++;
|
||||
xbits = 0;
|
||||
if (n >= base) xbits = extra[n-base];
|
||||
f = tree[n].Freq;
|
||||
opt_len += (ulg)f * (bits + xbits);
|
||||
if (stree) static_len += (ulg)f * (stree[n].Len + xbits);
|
||||
}
|
||||
if (overflow == 0) return;
|
||||
|
||||
/* Find the first bit length which could increase: */
|
||||
do {
|
||||
bits = max_length-1;
|
||||
while (bl_count[bits] == 0) bits--;
|
||||
bl_count[bits]--; /* move one leaf down the tree */
|
||||
bl_count[bits+1] += 2; /* move one overflow item as its brother */
|
||||
bl_count[max_length]--;
|
||||
/* The brother of the overflow item also moves one step up,
|
||||
* but this does not affect bl_count[max_length]
|
||||
*/
|
||||
overflow -= 2;
|
||||
} while (overflow > 0);
|
||||
|
||||
/* Now recompute all bit lengths, scanning in increasing frequency.
|
||||
* h is still equal to HEAP_SIZE. (It is simpler to reconstruct all
|
||||
* lengths instead of fixing only the wrong ones. This idea is taken
|
||||
* from 'ar' written by Haruhiko Okumura.)
|
||||
*/
|
||||
for (bits = max_length; bits != 0; bits--) {
|
||||
n = bl_count[bits];
|
||||
while (n != 0) {
|
||||
m = heap[--h];
|
||||
if (m > max_code) continue;
|
||||
if (tree[m].Len != (unsigned) bits) {
|
||||
opt_len += ((long)bits-(long)tree[m].Len)*(long)tree[m].Freq;
|
||||
tree[m].Len = (ush)bits;
|
||||
}
|
||||
n--;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
/* ===========================================================================
|
||||
* Generate the codes for a given tree and bit counts (which need not be
|
||||
* optimal).
|
||||
* IN assertion: the array bl_count contains the bit length statistics for
|
||||
* the given tree and the field len is set for all tree elements.
|
||||
* OUT assertion: the field code is set for all tree elements of non
|
||||
* zero code length.
|
||||
*/
|
||||
local void gen_codes (tree, max_code)
|
||||
ct_data *tree; /* the tree to decorate */
|
||||
int max_code; /* largest code with non zero frequency */
|
||||
{
|
||||
ush next_code[MAX_BITS+1]; /* next code value for each bit length */
|
||||
ush code = 0; /* running code value */
|
||||
int bits; /* bit index */
|
||||
int n; /* code index */
|
||||
|
||||
/* The distribution counts are first used to generate the code values
|
||||
* without bit reversal.
|
||||
*/
|
||||
for (bits = 1; bits <= MAX_BITS; bits++) {
|
||||
next_code[bits] = code = (code + bl_count[bits-1]) << 1;
|
||||
}
|
||||
/* Check that the bit counts in bl_count are consistent. The last code
|
||||
* must be all ones.
|
||||
*/
|
||||
|
||||
for (n = 0; n <= max_code; n++) {
|
||||
int len = tree[n].Len;
|
||||
if (len == 0) continue;
|
||||
/* Now reverse the bits */
|
||||
tree[n].Code = bi_reverse(next_code[len]++, len);
|
||||
}
|
||||
}
|
||||
|
||||
/* ===========================================================================
|
||||
* Construct one Huffman tree and assigns the code bit strings and lengths.
|
||||
* Update the total bit length for the current block.
|
||||
* IN assertion: the field freq is set for all tree elements.
|
||||
* OUT assertions: the fields len and code are set to the optimal bit length
|
||||
* and corresponding code. The length opt_len is updated; static_len is
|
||||
* also updated if stree is not null. The field max_code is set.
|
||||
*/
|
||||
local void build_tree(desc)
|
||||
tree_desc *desc; /* the tree descriptor */
|
||||
{
|
||||
ct_data *tree = desc->dyn_tree;
|
||||
ct_data *stree = desc->static_tree;
|
||||
int elems = desc->elems;
|
||||
int n, m; /* iterate over heap elements */
|
||||
int max_code = -1; /* largest code with non zero frequency */
|
||||
int node = elems; /* next internal node of the tree */
|
||||
|
||||
/* Construct the initial heap, with least frequent element in
|
||||
* heap[SMALLEST]. The sons of heap[n] are heap[2*n] and heap[2*n+1].
|
||||
* heap[0] is not used.
|
||||
*/
|
||||
heap_len = 0, heap_max = HEAP_SIZE;
|
||||
|
||||
for (n = 0; n < elems; n++) {
|
||||
if (tree[n].Freq != 0) {
|
||||
heap[++heap_len] = max_code = n;
|
||||
depth[n] = 0;
|
||||
} else {
|
||||
tree[n].Len = 0;
|
||||
}
|
||||
}
|
||||
|
||||
/* The pkzip format requires that at least one distance code exists,
|
||||
* and that at least one bit should be sent even if there is only one
|
||||
* possible code. So to avoid special checks later on we force at least
|
||||
* two codes of non zero frequency.
|
||||
*/
|
||||
while (heap_len < 2) {
|
||||
int new = heap[++heap_len] = (max_code < 2 ? ++max_code : 0);
|
||||
tree[new].Freq = 1;
|
||||
depth[new] = 0;
|
||||
opt_len--; if (stree) static_len -= stree[new].Len;
|
||||
/* new is 0 or 1 so it does not have extra bits */
|
||||
}
|
||||
desc->max_code = max_code;
|
||||
|
||||
/* The elements heap[heap_len/2+1 .. heap_len] are leaves of the tree,
|
||||
* establish sub-heaps of increasing lengths:
|
||||
*/
|
||||
for (n = heap_len/2; n >= 1; n--) pqdownheap(tree, n);
|
||||
|
||||
/* Construct the Huffman tree by repeatedly combining the least two
|
||||
* frequent nodes.
|
||||
*/
|
||||
do {
|
||||
pqremove(tree, n); /* n = node of least frequency */
|
||||
m = heap[SMALLEST]; /* m = node of next least frequency */
|
||||
|
||||
heap[--heap_max] = n; /* keep the nodes sorted by frequency */
|
||||
heap[--heap_max] = m;
|
||||
|
||||
/* Create a new node father of n and m */
|
||||
tree[node].Freq = tree[n].Freq + tree[m].Freq;
|
||||
depth[node] = (uch) (MAX(depth[n], depth[m]) + 1);
|
||||
tree[n].Dad = tree[m].Dad = (ush)node;
|
||||
/* and insert the new node in the heap */
|
||||
heap[SMALLEST] = node++;
|
||||
pqdownheap(tree, SMALLEST);
|
||||
|
||||
} while (heap_len >= 2);
|
||||
|
||||
heap[--heap_max] = heap[SMALLEST];
|
||||
|
||||
/* At this point, the fields freq and dad are set. We can now
|
||||
* generate the bit lengths.
|
||||
*/
|
||||
gen_bitlen((tree_desc *)desc);
|
||||
|
||||
/* The field len is now set, we can generate the bit codes */
|
||||
gen_codes ((ct_data *)tree, max_code);
|
||||
}
|
||||
|
||||
/* ===========================================================================
|
||||
* Scan a literal or distance tree to determine the frequencies of the codes
|
||||
* in the bit length tree. Updates opt_len to take into account the repeat
|
||||
* counts. (The contribution of the bit length codes will be added later
|
||||
* during the construction of bl_tree.)
|
||||
*/
|
||||
local void scan_tree (tree, max_code)
|
||||
ct_data *tree; /* the tree to be scanned */
|
||||
int max_code; /* and its largest code of non zero frequency */
|
||||
{
|
||||
int n; /* iterates over all tree elements */
|
||||
int prevlen = -1; /* last emitted length */
|
||||
int curlen; /* length of current code */
|
||||
int nextlen = tree[0].Len; /* length of next code */
|
||||
int count = 0; /* repeat count of the current code */
|
||||
int max_count = 7; /* max repeat count */
|
||||
int min_count = 4; /* min repeat count */
|
||||
|
||||
if (nextlen == 0) max_count = 138, min_count = 3;
|
||||
tree[max_code+1].Len = (ush)0xffff; /* guard */
|
||||
|
||||
for (n = 0; n <= max_code; n++) {
|
||||
curlen = nextlen; nextlen = tree[n+1].Len;
|
||||
if (++count < max_count && curlen == nextlen) {
|
||||
continue;
|
||||
} else if (count < min_count) {
|
||||
bl_tree[curlen].Freq += count;
|
||||
} else if (curlen != 0) {
|
||||
if (curlen != prevlen) bl_tree[curlen].Freq++;
|
||||
bl_tree[REP_3_6].Freq++;
|
||||
} else if (count <= 10) {
|
||||
bl_tree[REPZ_3_10].Freq++;
|
||||
} else {
|
||||
bl_tree[REPZ_11_138].Freq++;
|
||||
}
|
||||
count = 0; prevlen = curlen;
|
||||
if (nextlen == 0) {
|
||||
max_count = 138, min_count = 3;
|
||||
} else if (curlen == nextlen) {
|
||||
max_count = 6, min_count = 3;
|
||||
} else {
|
||||
max_count = 7, min_count = 4;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
/* ===========================================================================
|
||||
* Send a literal or distance tree in compressed form, using the codes in
|
||||
* bl_tree.
|
||||
*/
|
||||
local void send_tree (tree, max_code)
|
||||
ct_data *tree; /* the tree to be scanned */
|
||||
int max_code; /* and its largest code of non zero frequency */
|
||||
{
|
||||
int n; /* iterates over all tree elements */
|
||||
int prevlen = -1; /* last emitted length */
|
||||
int curlen; /* length of current code */
|
||||
int nextlen = tree[0].Len; /* length of next code */
|
||||
int count = 0; /* repeat count of the current code */
|
||||
int max_count = 7; /* max repeat count */
|
||||
int min_count = 4; /* min repeat count */
|
||||
|
||||
/* tree[max_code+1].Len = -1; */ /* guard already set */
|
||||
if (nextlen == 0) max_count = 138, min_count = 3;
|
||||
|
||||
for (n = 0; n <= max_code; n++) {
|
||||
curlen = nextlen; nextlen = tree[n+1].Len;
|
||||
if (++count < max_count && curlen == nextlen) {
|
||||
continue;
|
||||
} else if (count < min_count) {
|
||||
do { send_code(curlen, bl_tree); } while (--count != 0);
|
||||
|
||||
} else if (curlen != 0) {
|
||||
if (curlen != prevlen) {
|
||||
send_code(curlen, bl_tree); count--;
|
||||
}
|
||||
send_code(REP_3_6, bl_tree); send_bits(count-3, 2);
|
||||
|
||||
} else if (count <= 10) {
|
||||
send_code(REPZ_3_10, bl_tree); send_bits(count-3, 3);
|
||||
|
||||
} else {
|
||||
send_code(REPZ_11_138, bl_tree); send_bits(count-11, 7);
|
||||
}
|
||||
count = 0; prevlen = curlen;
|
||||
if (nextlen == 0) {
|
||||
max_count = 138, min_count = 3;
|
||||
} else if (curlen == nextlen) {
|
||||
max_count = 6, min_count = 3;
|
||||
} else {
|
||||
max_count = 7, min_count = 4;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
/* ===========================================================================
|
||||
* Construct the Huffman tree for the bit lengths and return the index in
|
||||
* bl_order of the last bit length code to send.
|
||||
*/
|
||||
local int build_bl_tree()
|
||||
{
|
||||
int max_blindex; /* index of last bit length code of non zero freq */
|
||||
|
||||
/* Determine the bit length frequencies for literal and distance trees */
|
||||
scan_tree((ct_data *)dyn_ltree, l_desc.max_code);
|
||||
scan_tree((ct_data *)dyn_dtree, d_desc.max_code);
|
||||
|
||||
/* Build the bit length tree: */
|
||||
build_tree((tree_desc *)(&bl_desc));
|
||||
/* opt_len now includes the length of the tree representations, except
|
||||
* the lengths of the bit lengths codes and the 5+5+4 bits for the counts.
|
||||
*/
|
||||
|
||||
/* Determine the number of bit length codes to send. The pkzip format
|
||||
* requires that at least 4 bit length codes be sent. (appnote.txt says
|
||||
* 3 but the actual value used is 4.)
|
||||
*/
|
||||
for (max_blindex = BL_CODES-1; max_blindex >= 3; max_blindex--) {
|
||||
if (bl_tree[bl_order[max_blindex]].Len != 0) break;
|
||||
}
|
||||
/* Update opt_len to include the bit length tree and counts */
|
||||
opt_len += 3*(max_blindex+1) + 5+5+4;
|
||||
|
||||
return max_blindex;
|
||||
}
|
||||
|
||||
/* ===========================================================================
|
||||
* Send the header for a block using dynamic Huffman trees: the counts, the
|
||||
* lengths of the bit length codes, the literal tree and the distance tree.
|
||||
* IN assertion: lcodes >= 257, dcodes >= 1, blcodes >= 4.
|
||||
*/
|
||||
local void send_all_trees(lcodes, dcodes, blcodes)
|
||||
int lcodes, dcodes, blcodes; /* number of codes for each tree */
|
||||
{
|
||||
int rank; /* index in bl_order */
|
||||
|
||||
send_bits(lcodes-257, 5); /* not +255 as stated in appnote.txt */
|
||||
send_bits(dcodes-1, 5);
|
||||
send_bits(blcodes-4, 4); /* not -3 as stated in appnote.txt */
|
||||
for (rank = 0; rank < blcodes; rank++) {
|
||||
send_bits(bl_tree[bl_order[rank]].Len, 3);
|
||||
}
|
||||
|
||||
send_tree((ct_data *)dyn_ltree, lcodes-1); /* send the literal tree */
|
||||
|
||||
send_tree((ct_data *)dyn_dtree, dcodes-1); /* send the distance tree */
|
||||
}
|
||||
|
||||
/* ===========================================================================
|
||||
* Determine the best encoding for the current block: dynamic trees, static
|
||||
* trees or store, and output the encoded block to the zip file. This function
|
||||
* returns the total compressed length for the file so far.
|
||||
*/
|
||||
ulg flush_block(buf, stored_len, eof)
|
||||
char *buf; /* input block, or NULL if too old */
|
||||
ulg stored_len; /* length of input block */
|
||||
int eof; /* true if this is the last block for a file */
|
||||
{
|
||||
ulg opt_lenb, static_lenb; /* opt_len and static_len in bytes */
|
||||
int max_blindex; /* index of last bit length code of non zero freq */
|
||||
|
||||
flag_buf[last_flags] = flags; /* Save the flags for the last 8 items */
|
||||
|
||||
/* Construct the literal and distance trees */
|
||||
build_tree((tree_desc *)(&l_desc));
|
||||
|
||||
build_tree((tree_desc *)(&d_desc));
|
||||
/* At this point, opt_len and static_len are the total bit lengths of
|
||||
* the compressed block data, excluding the tree representations.
|
||||
*/
|
||||
|
||||
/* Build the bit length tree for the above two trees, and get the index
|
||||
* in bl_order of the last bit length code to send.
|
||||
*/
|
||||
max_blindex = build_bl_tree();
|
||||
|
||||
/* Determine the best encoding. Compute first the block length in bytes */
|
||||
opt_lenb = (opt_len+3+7)>>3;
|
||||
static_lenb = (static_len+3+7)>>3;
|
||||
input_len += stored_len; /* for debugging only */
|
||||
|
||||
if (static_lenb <= opt_lenb) opt_lenb = static_lenb;
|
||||
|
||||
/* If compression failed and this is the first and last block,
|
||||
* and if the zip file can be seeked (to rewrite the local header),
|
||||
* the whole file is transformed into a stored file:
|
||||
*/
|
||||
if (stored_len <= opt_lenb && eof && compressed_len == 0L && seekable()) {
|
||||
copy_block(buf, (unsigned)stored_len, 0); /* without header */
|
||||
compressed_len = stored_len << 3;
|
||||
} else if (stored_len+4 <= opt_lenb && buf != (char*)0) {
|
||||
/* 4: two words for the lengths */
|
||||
/* The test buf != NULL is only necessary if LIT_BUFSIZE > WSIZE.
|
||||
* Otherwise we can't have processed more than WSIZE input bytes since
|
||||
* the last block flush, because compression would have been
|
||||
* successful. If LIT_BUFSIZE <= WSIZE, it is never too late to
|
||||
* transform a block into a stored block.
|
||||
*/
|
||||
send_bits((STORED_BLOCK<<1)+eof, 3); /* send block type */
|
||||
compressed_len = (compressed_len + 3 + 7) & ~7L;
|
||||
compressed_len += (stored_len + 4) << 3;
|
||||
|
||||
copy_block(buf, (unsigned)stored_len, 1); /* with header */
|
||||
} else if (static_lenb == opt_lenb) {
|
||||
send_bits((STATIC_TREES<<1)+eof, 3);
|
||||
compress_block((ct_data *)static_ltree, (ct_data *)static_dtree);
|
||||
compressed_len += 3 + static_len;
|
||||
} else {
|
||||
send_bits((DYN_TREES<<1)+eof, 3);
|
||||
send_all_trees(l_desc.max_code+1, d_desc.max_code+1, max_blindex+1);
|
||||
compress_block((ct_data *)dyn_ltree, (ct_data *)dyn_dtree);
|
||||
compressed_len += 3 + opt_len;
|
||||
}
|
||||
init_block();
|
||||
|
||||
if (eof) {
|
||||
bi_windup();
|
||||
compressed_len += 7; /* align on byte boundary */
|
||||
}
|
||||
|
||||
return compressed_len >> 3;
|
||||
}
|
||||
|
||||
/* ===========================================================================
|
||||
* Save the match info and tally the frequency counts. Return true if
|
||||
* the current block must be flushed.
|
||||
*/
|
||||
int ct_tally (dist, lc)
|
||||
int dist; /* distance of matched string */
|
||||
int lc; /* match length-MIN_MATCH or unmatched char (if dist==0) */
|
||||
{
|
||||
l_buf[last_lit++] = (uch)lc;
|
||||
if (dist == 0) {
|
||||
/* lc is the unmatched char */
|
||||
dyn_ltree[lc].Freq++;
|
||||
} else {
|
||||
/* Here, lc is the match length - MIN_MATCH */
|
||||
dist--; /* dist = match distance - 1 */
|
||||
dyn_ltree[length_code[lc]+LITERALS+1].Freq++;
|
||||
dyn_dtree[d_code(dist)].Freq++;
|
||||
|
||||
d_buf[last_dist++] = (ush)dist;
|
||||
flags |= flag_bit;
|
||||
}
|
||||
flag_bit <<= 1;
|
||||
|
||||
/* Output the flags if they fill a byte: */
|
||||
if ((last_lit & 7) == 0) {
|
||||
flag_buf[last_flags++] = flags;
|
||||
flags = 0, flag_bit = 1;
|
||||
}
|
||||
/* Try to guess if it is profitable to stop the current block here */
|
||||
if ((last_lit & 0xfff) == 0) {
|
||||
/* Compute an upper bound for the compressed length */
|
||||
ulg out_length = (ulg)last_lit*8L;
|
||||
ulg in_length = (ulg)strstart-block_start;
|
||||
int dcode;
|
||||
for (dcode = 0; dcode < D_CODES; dcode++) {
|
||||
out_length += (ulg)dyn_dtree[dcode].Freq*(5L+extra_dbits[dcode]);
|
||||
}
|
||||
out_length >>= 3;
|
||||
if (last_dist < last_lit/2 && out_length < in_length/2) return 1;
|
||||
}
|
||||
return (last_lit == LIT_BUFSIZE-1 || last_dist == DIST_BUFSIZE);
|
||||
/* We avoid equality with LIT_BUFSIZE because of wraparound at 64K
|
||||
* on 16 bit machines and because stored blocks are restricted to
|
||||
* 64K-1 bytes.
|
||||
*/
|
||||
}
|
||||
|
||||
/* ===========================================================================
|
||||
* Send the block data compressed using the given Huffman trees
|
||||
*/
|
||||
local void compress_block(ltree, dtree)
|
||||
ct_data *ltree; /* literal tree */
|
||||
ct_data *dtree; /* distance tree */
|
||||
{
|
||||
unsigned dist; /* distance of matched string */
|
||||
int lc; /* match length or unmatched char (if dist == 0) */
|
||||
unsigned lx = 0; /* running index in l_buf */
|
||||
unsigned dx = 0; /* running index in d_buf */
|
||||
unsigned fx = 0; /* running index in flag_buf */
|
||||
uch flag = 0; /* current flags */
|
||||
unsigned code; /* the code to send */
|
||||
int extra; /* number of extra bits to send */
|
||||
|
||||
if (last_lit != 0) do {
|
||||
if ((lx & 7) == 0) flag = flag_buf[fx++];
|
||||
lc = l_buf[lx++];
|
||||
if ((flag & 1) == 0) {
|
||||
send_code(lc, ltree); /* send a literal byte */
|
||||
} else {
|
||||
/* Here, lc is the match length - MIN_MATCH */
|
||||
code = length_code[lc];
|
||||
send_code(code+LITERALS+1, ltree); /* send the length code */
|
||||
extra = extra_lbits[code];
|
||||
if (extra != 0) {
|
||||
lc -= base_length[code];
|
||||
send_bits(lc, extra); /* send the extra length bits */
|
||||
}
|
||||
dist = d_buf[dx++];
|
||||
/* Here, dist is the match distance - 1 */
|
||||
code = d_code(dist);
|
||||
|
||||
send_code(code, dtree); /* send the distance code */
|
||||
extra = extra_dbits[code];
|
||||
if (extra != 0) {
|
||||
dist -= base_dist[code];
|
||||
send_bits(dist, extra); /* send the extra distance bits */
|
||||
}
|
||||
} /* literal or match pair ? */
|
||||
flag >>= 1;
|
||||
} while (lx < last_lit);
|
||||
|
||||
send_code(END_BLOCK, ltree);
|
||||
}
|
||||
@@ -0,0 +1,73 @@
|
||||
#include <stdio.h>
|
||||
#include <stdlib.h>
|
||||
#include "mkrom.h"
|
||||
|
||||
struct state state;
|
||||
|
||||
/**
|
||||
* mkrom - do ROM finalisation steps
|
||||
*
|
||||
* mkrom <romfile> <mapfile> <piracychecks> <zipmagic> <outfile>
|
||||
*
|
||||
* <romfile>
|
||||
* This is the path to the stage1 binary. This file is similar to the final ROM,
|
||||
* but the allocations for compressed segments are vacant, and the compressed
|
||||
* segments themselves are found uncompressed past the 32MB mark of the ROM.
|
||||
* This program will compress those segments, put them in their allocated spaces
|
||||
* and truncate the ROM to 32MB.
|
||||
*
|
||||
* <mapfile>
|
||||
* This is the path to the linker map, which is used to determine where the
|
||||
* uncompressed segments are and where they should be placed.
|
||||
*
|
||||
* <piracychecks>
|
||||
* This should be 0 or 1 to indicate whether this version of the ROM contains
|
||||
* piracy checks or not. Some piracy checks work by checksumming functions in
|
||||
* memory at runtime and comparing it with a known value. If set to 1, mkrom
|
||||
* will calculate the checksums for these functions and patch them into the
|
||||
* piracy checks.
|
||||
*
|
||||
* <zipmagic>
|
||||
* This is a two byte value which is used when zipping the game segments.
|
||||
* The original code was influenced by uninitialised data. These two bytes are
|
||||
* just setting that uninitialised data.
|
||||
*
|
||||
* <outfile>
|
||||
* The file to write the final ROM file to.
|
||||
*
|
||||
* eg. mkrom stage1.bin pd.map 1 0x1234 pd.z64
|
||||
*/
|
||||
int main(int argc, char **argv)
|
||||
{
|
||||
if (argc < 6) {
|
||||
fprintf(stderr, "Usage: mkrom <romfile> <mapfile> <piracychecks> <zipmagic> <outfile>\n");
|
||||
exit(1);
|
||||
}
|
||||
|
||||
rom_load(argv[1]);
|
||||
map_open(argv[2]);
|
||||
|
||||
state.piracychecks = atoi(argv[3]);
|
||||
state.zipmagic = strtol(argv[4], NULL, 16);
|
||||
|
||||
// Compute piracy checksums if requested
|
||||
if (state.piracychecks) {
|
||||
piracy_patch();
|
||||
}
|
||||
|
||||
// Slice the game segment into chunks and zip each of them to create the
|
||||
// gamezips segment
|
||||
game_zip();
|
||||
|
||||
// Pack each segment into their final locations
|
||||
pack_lib();
|
||||
pack_data();
|
||||
pack_game();
|
||||
pack_fill();
|
||||
|
||||
rom_update_crc();
|
||||
|
||||
rom_write(argv[5]);
|
||||
|
||||
return 0;
|
||||
}
|
||||
@@ -0,0 +1,151 @@
|
||||
#include <ctype.h>
|
||||
#include <stdint.h>
|
||||
#include <stdio.h>
|
||||
#include <stdlib.h>
|
||||
#include <string.h>
|
||||
#include "mkrom.h"
|
||||
|
||||
/**
|
||||
* This file handles reading and parsing the linker map.
|
||||
*/
|
||||
|
||||
extern struct state state;
|
||||
|
||||
void map_open(char *filename)
|
||||
{
|
||||
state.mapfd = fopen(filename, "r");
|
||||
|
||||
if (!state.mapfd) {
|
||||
fprintf(stderr, "Unable to open map file \"%s\" for reading\n", filename);
|
||||
exit(1);
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Find the start and end offsets of the given function in the ROM and write
|
||||
* their offsets to the start and end pointers.
|
||||
*
|
||||
* Return true if the function was found, false if not.
|
||||
*/
|
||||
bool map_get_function_rompos(char *funcname, uint32_t *start, uint32_t *end)
|
||||
{
|
||||
char line[1024];
|
||||
char *ptr;
|
||||
uint32_t segramaddr = 0;
|
||||
uint32_t segromoffset = 0;
|
||||
uint32_t ramaddr;
|
||||
char find[1024];
|
||||
bool lookingforend = false;
|
||||
|
||||
snprintf(find, sizeof(find), " %s\n", funcname);
|
||||
|
||||
fseek(state.mapfd, 0, SEEK_SET);
|
||||
|
||||
while (!feof(state.mapfd)) {
|
||||
fgets(line, 1024, state.mapfd);
|
||||
|
||||
if (lookingforend) {
|
||||
ptr = line;
|
||||
|
||||
while (isspace(*ptr)) {
|
||||
ptr++;
|
||||
}
|
||||
|
||||
if (ptr[0] == '0' && ptr[1] == 'x') {
|
||||
ramaddr = strtoul(ptr, NULL, 16);
|
||||
|
||||
*end = ramaddr - segramaddr + segromoffset;
|
||||
return true;
|
||||
}
|
||||
} else if (line[0] == '.') {
|
||||
// Start of a segment
|
||||
// ".game 0x000000007f000000 0x1b99e0 load address 0x00000000020ac170"
|
||||
|
||||
// Jump to RAM address
|
||||
ptr = strstr(line, "0x");
|
||||
segramaddr = strtoul(ptr, NULL, 16);
|
||||
|
||||
// Jump to length
|
||||
ptr++;
|
||||
ptr = strstr(ptr, "0x");
|
||||
|
||||
// Jump to ROM offset
|
||||
ptr++;
|
||||
ptr = strstr(ptr, "0x");
|
||||
segromoffset = strtoul(ptr, NULL, 16);
|
||||
} else if (strstr(line, find)) {
|
||||
// Found the function
|
||||
// "0x000000007f15d9a8 bgInflate"
|
||||
ptr = strstr(line, "0x");
|
||||
ramaddr = strtoul(ptr, NULL, 16);
|
||||
|
||||
*start = ramaddr - segramaddr + segromoffset;
|
||||
lookingforend = true;
|
||||
}
|
||||
}
|
||||
|
||||
fprintf(stderr, "Unable to find function \"%s\" in linker map\n", funcname);
|
||||
|
||||
return false;
|
||||
}
|
||||
|
||||
/**
|
||||
* Find the start and end offsets of the given segment in the ROM and write
|
||||
* their offsets to the start and end pointers.
|
||||
*
|
||||
* Either point may be NULL.
|
||||
*
|
||||
* Return true if the segment was found, false if not.
|
||||
*/
|
||||
bool map_get_segment_rompos(char *segname, uint32_t *start, uint32_t *end)
|
||||
{
|
||||
char startstring[64];
|
||||
char endstring[64];
|
||||
char line[1024];
|
||||
bool found_start = false;
|
||||
bool found_end = false;
|
||||
|
||||
snprintf(startstring, sizeof(startstring), "_%sSegmentRomStart = ", segname);
|
||||
snprintf(endstring, sizeof(endstring), "_%sSegmentRomEnd = ", segname);
|
||||
|
||||
fseek(state.mapfd, 0, SEEK_SET);
|
||||
|
||||
// Find lines like this:
|
||||
// " 0x0000000000001050 _libzipSegmentRomStart = __rompos"
|
||||
|
||||
while (!feof(state.mapfd)) {
|
||||
fgets(line, 1024, state.mapfd);
|
||||
|
||||
if (!found_start && strstr(line, startstring)) {
|
||||
char *ptr = strstr(line, "0x");
|
||||
|
||||
if (start != NULL) {
|
||||
*start = strtoul(ptr, NULL, 16);
|
||||
}
|
||||
|
||||
if (found_end) {
|
||||
return true;
|
||||
}
|
||||
|
||||
found_start = true;
|
||||
}
|
||||
|
||||
if (!found_end && strstr(line, endstring)) {
|
||||
char *ptr = strstr(line, "0x");
|
||||
|
||||
if (end != NULL) {
|
||||
*end = strtoul(ptr, NULL, 16);
|
||||
}
|
||||
|
||||
if (found_start) {
|
||||
return true;
|
||||
}
|
||||
|
||||
found_end = true;
|
||||
}
|
||||
}
|
||||
|
||||
fprintf(stderr, "Unable to find segment \"%s\" in linker map\n", segname);
|
||||
|
||||
return false;
|
||||
}
|
||||
@@ -0,0 +1,76 @@
|
||||
#include <stdbool.h>
|
||||
#include <stdio.h>
|
||||
#include <stdlib.h>
|
||||
#include <stdint.h>
|
||||
|
||||
struct state {
|
||||
/**
|
||||
* A pointer to the full working ROM area in memory.
|
||||
*/
|
||||
unsigned char *rom;
|
||||
|
||||
/**
|
||||
* The size of the above rom allocation in bytes.
|
||||
* The value is the same filesize as the stage1 binary,
|
||||
* which is 32MB plus some uncompressed segments on the end.
|
||||
*/
|
||||
size_t romlen;
|
||||
|
||||
/**
|
||||
* Whether piracy checks are enabled for this build or not.
|
||||
* If enabled, mkrom will recalculate piracy-related checksums.
|
||||
*/
|
||||
bool piracychecks;
|
||||
|
||||
/**
|
||||
* Two bytes that are used to seed some uninitialised data in the input
|
||||
* buffer when zipping game chunks.
|
||||
*/
|
||||
unsigned short zipmagic;
|
||||
|
||||
/**
|
||||
* File descriptor for the linker map.
|
||||
*/
|
||||
FILE *mapfd;
|
||||
|
||||
/**
|
||||
* A pointer to the gamezips segment, once created.
|
||||
* The gamezips segment is the full segment but with the offest table
|
||||
* zeroed.
|
||||
*/
|
||||
unsigned char *gamezips;
|
||||
|
||||
/**
|
||||
* Size of the above gamezips allocation in bytes.
|
||||
*/
|
||||
size_t gamezipslen;
|
||||
|
||||
/**
|
||||
* Pointer to a separate allocation for the gamezips offset table.
|
||||
*/
|
||||
unsigned char *gametable;
|
||||
|
||||
/**
|
||||
* Size of the above gametable allocation in bytes.
|
||||
*/
|
||||
size_t gametablelen;
|
||||
};
|
||||
|
||||
void game_zip(void);
|
||||
|
||||
void map_open(char *filename);
|
||||
bool map_get_function_rompos(char *funcname, uint32_t *start, uint32_t *end);
|
||||
bool map_get_segment_rompos(char *funcname, uint32_t *start, uint32_t *end);
|
||||
|
||||
void pack_lib(void);
|
||||
void pack_data(void);
|
||||
void pack_game(void);
|
||||
void pack_fill(void);
|
||||
|
||||
void piracy_patch(void);
|
||||
|
||||
void rarezip(uint8_t *outbuffer, size_t *outlen, uint8_t *inbuffer, size_t inlen, uint32_t magic);
|
||||
|
||||
void rom_load(char *filename);
|
||||
void rom_update_crc(void);
|
||||
void rom_write(char *filename);
|
||||
@@ -0,0 +1,153 @@
|
||||
#include <stdio.h>
|
||||
#include <stdlib.h>
|
||||
#include <string.h>
|
||||
#include "mkrom.h"
|
||||
|
||||
extern struct state state;
|
||||
|
||||
static void copy(char *segname, uint8_t *payload, size_t len, char *constname)
|
||||
{
|
||||
uint32_t start;
|
||||
uint32_t end;
|
||||
char zipsegname[32];
|
||||
uint32_t allocation;
|
||||
|
||||
snprintf(zipsegname, sizeof(zipsegname), "%szip", segname);
|
||||
|
||||
map_get_segment_rompos(zipsegname, &start, &end);
|
||||
|
||||
allocation = end - start;
|
||||
|
||||
if (len > allocation) {
|
||||
fprintf(stderr, "The %s segment is too big after compression to fit the allocation of 0x%x.\n", segname, allocation);
|
||||
fprintf(stderr, "In ld/pd.ld, increase the value of %s to 0x%x or higher.\n", constname, len);
|
||||
exit(1);
|
||||
}
|
||||
|
||||
memcpy(&state.rom[start], payload, len);
|
||||
}
|
||||
|
||||
/**
|
||||
* To pack the data segment, zip it in full and copy it to the datazip segment.
|
||||
*/
|
||||
void pack_data(void)
|
||||
{
|
||||
uint32_t start;
|
||||
uint32_t end;
|
||||
size_t ziplen;
|
||||
|
||||
map_get_segment_rompos("data", &start, &end);
|
||||
|
||||
uint8_t *buffer = malloc(end - start);
|
||||
|
||||
rarezip(buffer, &ziplen, &state.rom[start], end - start, 0);
|
||||
|
||||
copy("data", buffer, ziplen, "ROMALLOCATION_DATA");
|
||||
|
||||
free(buffer);
|
||||
}
|
||||
|
||||
/**
|
||||
* On the ROM, the gamezips segment exists with its offset table and zip data.
|
||||
* Then after that comes a copy of the gamezips segment but with the offset
|
||||
* table cleared. The second segment is garbage data.
|
||||
*
|
||||
* mkrom has already built the gamezips segment but with a zeroed table.
|
||||
* The real table is pointed to by state.gametable.
|
||||
*
|
||||
* So we have to copy the gamezips segment twice, then paste the gametable
|
||||
* segment over the start of the first one.
|
||||
*/
|
||||
void pack_game(void)
|
||||
{
|
||||
uint32_t gamezipstart;
|
||||
uint32_t gamezipend;
|
||||
size_t truncatedlen;
|
||||
|
||||
// Copy the gamezips segment
|
||||
copy("game", state.gamezips, state.gamezipslen, "ROMALLOCATION_GAME");
|
||||
|
||||
// Paste over the offset table
|
||||
map_get_segment_rompos("gamezip", &gamezipstart, &gamezipend);
|
||||
|
||||
memcpy(&state.rom[gamezipstart], state.gametable, state.gametablelen);
|
||||
|
||||
// Paste the second segment, truncating it to fit the allocation
|
||||
truncatedlen = gamezipend - gamezipstart - state.gamezipslen;
|
||||
|
||||
if (truncatedlen > state.gamezipslen) {
|
||||
truncatedlen = state.gamezipslen;
|
||||
}
|
||||
|
||||
memcpy(&state.rom[gamezipstart + state.gamezipslen], state.gamezips, truncatedlen);
|
||||
|
||||
// The final two bytes from the real segment are duplicated into
|
||||
// the first two bytes of the second segment's offset table
|
||||
state.rom[gamezipstart + state.gamezipslen + 0] = state.rom[gamezipstart + state.gamezipslen - 2];
|
||||
state.rom[gamezipstart + state.gamezipslen + 1] = state.rom[gamezipstart + state.gamezipslen - 1];
|
||||
}
|
||||
|
||||
/**
|
||||
* The lib segment is zipped from 0x2000 onwards.
|
||||
*
|
||||
* It's placed twice in a row in the ROM within its allocation, where the second
|
||||
* one is truncated and unused.
|
||||
*/
|
||||
void pack_lib(void)
|
||||
{
|
||||
uint32_t libzipstart;
|
||||
uint32_t libzipend;
|
||||
uint32_t libstart;
|
||||
uint32_t libend;
|
||||
size_t ziplen;
|
||||
size_t seglen;
|
||||
size_t truncatedlen;
|
||||
|
||||
map_get_segment_rompos("lib", &libstart, &libend);
|
||||
|
||||
uint8_t *buffer = malloc(libend - libstart);
|
||||
|
||||
// Read the first 0x2000 into a buffer
|
||||
memcpy(buffer, &state.rom[libstart], 0x2000);
|
||||
|
||||
// Compress the remainder from ROM, appending to the buffer
|
||||
rarezip(&buffer[0x2000], &ziplen, &state.rom[libstart + 0x2000], libend - libstart - 0x2000, 0);
|
||||
seglen = ziplen + 0x2000;
|
||||
|
||||
// Copy the buffer to its real spot in the ROM
|
||||
copy("lib", buffer, seglen, "ROMALLOCATION_LIB");
|
||||
|
||||
// Copy it truncated to its fake spot
|
||||
map_get_segment_rompos("libzip", &libzipstart, &libzipend);
|
||||
|
||||
truncatedlen = libzipend - libzipstart - seglen;
|
||||
|
||||
if (truncatedlen > seglen) {
|
||||
truncatedlen = seglen;
|
||||
}
|
||||
|
||||
memcpy(state.rom + libzipstart + seglen, buffer, truncatedlen);
|
||||
|
||||
free(buffer);
|
||||
}
|
||||
|
||||
/**
|
||||
* Fill from the end of the last segment to the end of the ROM with 0xff.
|
||||
*/
|
||||
void pack_fill(void)
|
||||
{
|
||||
uint32_t offset;
|
||||
|
||||
map_get_segment_rompos("accessingpak", NULL, &offset);
|
||||
|
||||
if (offset == 0) {
|
||||
// We're probably building ntsc-beta, which doesn't have the
|
||||
// accessingpak segment.
|
||||
map_get_segment_rompos("copyright", NULL, &offset);
|
||||
}
|
||||
|
||||
while (offset < 1024 * 1024 * 32) {
|
||||
state.rom[offset] = 0xff;
|
||||
offset++;
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,170 @@
|
||||
#include <arpa/inet.h>
|
||||
#include <stdio.h>
|
||||
#include <stdlib.h>
|
||||
#include <stdint.h>
|
||||
#include "mkrom.h"
|
||||
|
||||
#define CHECKSUM_PLACEHOLDER 0x99aabbcc
|
||||
|
||||
extern struct state state;
|
||||
|
||||
typedef uint32_t (*Algo)(uint32_t sum, uint32_t word);
|
||||
|
||||
static uint32_t algo01(uint32_t sum, uint32_t word) { return sum ^ word; }
|
||||
static uint32_t algo02(uint32_t sum, uint32_t word) { return sum ^ ~word; }
|
||||
static uint32_t algo03(uint32_t sum, uint32_t word) { return (sum + word) * 2; }
|
||||
static uint32_t algo04(uint32_t sum, uint32_t word) { return sum + ~word; }
|
||||
static uint32_t algo05(uint32_t sum, uint32_t word) { return sum * 2 + word; }
|
||||
static uint32_t algo06(uint32_t sum, uint32_t word) { return sum + word; }
|
||||
static uint32_t algo07(uint32_t sum, uint32_t word) { return (sum << 1) ^ word; }
|
||||
static uint32_t algo08(uint32_t sum, uint32_t word) { return (sum + word) + (word >> 1); }
|
||||
static uint32_t algo09(uint32_t sum, uint32_t word) { return sum - ~word; }
|
||||
static uint32_t algo10(uint32_t sum, uint32_t word) { return (sum ^ word) << 1; }
|
||||
static uint32_t algo11(uint32_t sum, uint32_t word) { return (sum ^ ~word) << 1; }
|
||||
|
||||
static uint32_t algo12(uint32_t sum, uint32_t word) {
|
||||
sum ^= ~word;
|
||||
sum ^= word << 5;
|
||||
sum ^= word >> 15;
|
||||
return sum;
|
||||
}
|
||||
|
||||
/**
|
||||
* Calculate the checksum of sumfunc.
|
||||
*
|
||||
* We just iterate each word in the function and run the algo function on each.
|
||||
*/
|
||||
static uint32_t calc_sum(char *sumfunc, Algo algo)
|
||||
{
|
||||
uint32_t start;
|
||||
uint32_t end;
|
||||
uint32_t sum = 0;
|
||||
uint32_t offset;
|
||||
|
||||
if (!map_get_function_rompos(sumfunc, &start, &end)) {
|
||||
fprintf(stderr, "Unable to find function \"%s\" in map file\n", sumfunc);
|
||||
exit(1);
|
||||
}
|
||||
|
||||
for (offset = start; offset < end; offset += 4) {
|
||||
sum = algo(sum, ntohl(*(uint32_t *) &state.rom[offset]));
|
||||
}
|
||||
|
||||
return sum;
|
||||
}
|
||||
|
||||
static bool is_branch_likely(uint32_t word)
|
||||
{
|
||||
uint32_t op = word & 0xfc000000;
|
||||
|
||||
if (op == 0x50000000) { // beql
|
||||
return true;
|
||||
}
|
||||
|
||||
if (op == 0x54000000) { // bnel
|
||||
return true;
|
||||
}
|
||||
|
||||
if (op == 0x58000000) { // blezl
|
||||
return true;
|
||||
}
|
||||
|
||||
if (op == 0x5c000000) { // bgtzl
|
||||
return true;
|
||||
}
|
||||
|
||||
if (op == 0x01000000 && (word & 0x001f0000) == 0x00020000) { // bltzl
|
||||
return true;
|
||||
}
|
||||
|
||||
if (op == 0x01000000 && (word & 0x001f0000) == 0x00030000) { // bgezl
|
||||
return true;
|
||||
}
|
||||
|
||||
return false;
|
||||
}
|
||||
|
||||
/**
|
||||
* Search the patchfunc for the placeholder checksum and replace it with the one
|
||||
* we calculated.
|
||||
*
|
||||
* Checksums are always written into $at with lui and ori instructions.
|
||||
*
|
||||
* 3c0199aa lui $at,0x99aa
|
||||
* 3421bbcc ori $at,$at,0xbbcc
|
||||
*/
|
||||
static void write_sum(char *patchfunc, uint32_t sum)
|
||||
{
|
||||
uint32_t start;
|
||||
uint32_t end;
|
||||
|
||||
if (!map_get_function_rompos(patchfunc, &start, &end)) {
|
||||
fprintf(stderr, "Unable to find function \"%s\" in map file\n", patchfunc);
|
||||
exit(1);
|
||||
}
|
||||
|
||||
bool in_branchlikely = false;
|
||||
|
||||
uint32_t upperpos = 0;
|
||||
uint32_t lowerpos = 0;
|
||||
uint32_t offset;
|
||||
|
||||
for (offset = start; offset < end && (!upperpos || !lowerpos); offset += 4) {
|
||||
uint32_t word = ntohl(*(uint32_t *) &state.rom[offset]);
|
||||
|
||||
if (in_branchlikely) {
|
||||
in_branchlikely = false;
|
||||
} else {
|
||||
if (is_branch_likely(word)) {
|
||||
in_branchlikely = true;
|
||||
} else if (word == (0x3c010000 | (CHECKSUM_PLACEHOLDER >> 16))) {
|
||||
upperpos = offset;
|
||||
} else if (upperpos && word == (0x34210000 | (CHECKSUM_PLACEHOLDER & 0xffff))) {
|
||||
lowerpos = offset;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
if (!upperpos || !lowerpos) {
|
||||
fprintf(stderr, "Unable to find placeholder checksum in %s.\n", patchfunc);
|
||||
fprintf(stderr, "This can happen if you've turned PIRACYCHECKS off, built the files, then turned it on without rebuilding.\n");
|
||||
fprintf(stderr, "To fix, try running the following:\n");
|
||||
fprintf(stderr, "\n");
|
||||
fprintf(stderr, " touch $(grep -lr PIRACYCHECKS src)\n");
|
||||
fprintf(stderr, " make\n");
|
||||
fprintf(stderr, "\n");
|
||||
exit(1);
|
||||
}
|
||||
|
||||
state.rom[upperpos + 2] = (sum >> 24) & 0xff;
|
||||
state.rom[upperpos + 3] = (sum >> 16) & 0xff;
|
||||
state.rom[lowerpos + 2] = (sum >> 8) & 0xff;
|
||||
state.rom[lowerpos + 3] = sum & 0xff;
|
||||
}
|
||||
|
||||
static void patch(Algo algo, char *patchfunc, char *sumfunc)
|
||||
{
|
||||
uint32_t sum = calc_sum(sumfunc, algo);
|
||||
|
||||
write_sum(patchfunc, sum);
|
||||
}
|
||||
|
||||
/**
|
||||
* Patch all the piracy functions in the game.
|
||||
*/
|
||||
void piracy_patch(void)
|
||||
{
|
||||
// algorithm, patch function, sum function
|
||||
patch(algo01, "__scHandleTasks", "bootPhase1");
|
||||
patch(algo02, "cheatMenuHandleDialog", "__scHandleTasks");
|
||||
patch(algo03, "propobjHandlePickupByAibot", "func0f08e2ac");
|
||||
patch(algo04, "chrUncloak", "propobjHandlePickupByAibot");
|
||||
patch(algo05, "chrsCheckForNoise", "__scHandleRetrace");
|
||||
patch(algo06, "lvInit", "lvGetSlowMotionType");
|
||||
patch(algo07, "propAllocateEyespy", "lvInit");
|
||||
patch(algo08, "chrConsiderGrenadeThrow", "bgInit");
|
||||
patch(algo09, "bgun0f09e144", "tagsAllocatePtrs");
|
||||
patch(algo10, "explosionAlertChrs", "glassDestroy");
|
||||
patch(algo11, "func0f0069dc", "mtxGetObfuscatedRomBase");
|
||||
patch(algo12, "func0f15c920", "func0f0069dc");
|
||||
}
|
||||
@@ -0,0 +1,107 @@
|
||||
#include <stdint.h>
|
||||
#include <stdio.h>
|
||||
#include <stdlib.h>
|
||||
#include <string.h>
|
||||
#include "gzip.h"
|
||||
|
||||
/**
|
||||
* This file is mkrom's interface to gzip.
|
||||
*
|
||||
* The key thing here is that we need to be able to set the uninitialised data
|
||||
* in the window buffer that gzip uses. This is required for a matching build
|
||||
* and is the reason why mkrom must be written in C.
|
||||
*
|
||||
* Much of the gzip code has been ripped out, including file functionality and
|
||||
* many, many global variables that it uses.
|
||||
*/
|
||||
|
||||
#define BITS 16
|
||||
|
||||
// This global variable is required by gzip
|
||||
unsigned outcnt; /* bytes in gzip's output buffer (not ours) */
|
||||
|
||||
DECLARE(uch, inbuf, INBUFSIZ +INBUF_EXTRA);
|
||||
DECLARE(uch, outbuf, OUTBUFSIZ+OUTBUF_EXTRA);
|
||||
DECLARE(ush, d_buf, DIST_BUFSIZE);
|
||||
DECLARE(uch, window, 2L*WSIZE);
|
||||
DECLARE(ush, tab_prefix, 1L<<BITS);
|
||||
|
||||
// These global variables are used to allow
|
||||
// us to work with gzip's input/output system
|
||||
unsigned len_remaining;
|
||||
uint8_t *inptr;
|
||||
uint8_t *outptr;
|
||||
|
||||
static void zip(uint8_t *outbuffer, size_t *outlen, uint8_t *inbuffer, size_t inlen, uint32_t magic)
|
||||
{
|
||||
// Set up pointers to the buffers.
|
||||
// We'll move the pointers forward as gzip requests data be read or written
|
||||
inptr = inbuffer;
|
||||
outptr = outbuffer;
|
||||
len_remaining = inlen;
|
||||
|
||||
// Here we're setting some "uninitialised" data in the window.
|
||||
// Older versions of gzip are influenced by whatever data happened
|
||||
// to be in the window immediately after the end of the input.
|
||||
// Each game zip is 0x1000 long uncompressed.
|
||||
// The magic values differ per ROM version and are specified as mkrom args.
|
||||
window[0x1000] = (magic >> 8) & 0xff;
|
||||
window[0x1001] = magic & 0xff;
|
||||
|
||||
outcnt = 0;
|
||||
|
||||
// Begin calling gzip routines
|
||||
bi_init();
|
||||
ct_init();
|
||||
lm_init();
|
||||
|
||||
deflate();
|
||||
flush_outbuf();
|
||||
|
||||
*outlen = outptr - outbuffer;
|
||||
}
|
||||
|
||||
/**
|
||||
* Write the RareZip header (0x1173 followed by original file size)
|
||||
* to the buffer followed by the compressed data.
|
||||
*
|
||||
* It's up to the caller to allocate an output buffer big enough.
|
||||
*/
|
||||
void rarezip(uint8_t *outbuffer, size_t *outlen, uint8_t *inbuffer, size_t inlen, uint32_t magic)
|
||||
{
|
||||
outbuffer[0] = 0x11;
|
||||
outbuffer[1] = 0x73;
|
||||
outbuffer[2] = (inlen >> 16) & 0xff;
|
||||
outbuffer[3] = (inlen >> 8) & 0xff;
|
||||
outbuffer[4] = inlen & 0xff;
|
||||
|
||||
zip(&outbuffer[5], outlen, inbuffer, inlen, magic);
|
||||
|
||||
*outlen += 5;
|
||||
}
|
||||
|
||||
/**
|
||||
* This function is called by gzip when it wants more data.
|
||||
*/
|
||||
int read_buf(char *buf, unsigned size)
|
||||
{
|
||||
if (size > len_remaining) {
|
||||
size = len_remaining;
|
||||
}
|
||||
|
||||
memcpy(buf, inptr, size);
|
||||
|
||||
len_remaining -= size;
|
||||
inptr += size;
|
||||
|
||||
return size;
|
||||
}
|
||||
|
||||
/**
|
||||
* This function is called by gzip when it wants to output data.
|
||||
*/
|
||||
void write_buf(voidp buf, unsigned size)
|
||||
{
|
||||
memcpy(outptr, buf, size);
|
||||
outptr += size;
|
||||
}
|
||||
@@ -0,0 +1,123 @@
|
||||
#include <stdio.h>
|
||||
#include <stdlib.h>
|
||||
#include "mkrom.h"
|
||||
|
||||
extern struct state state;
|
||||
|
||||
/**
|
||||
* Load the stage1 ROM into memory.
|
||||
*/
|
||||
void rom_load(char *filename)
|
||||
{
|
||||
FILE *fp = fopen(filename, "rb");
|
||||
|
||||
if (!fp) {
|
||||
fprintf(stderr, "Unable to open \"%s\" for reading\n", filename);
|
||||
exit(1);
|
||||
}
|
||||
|
||||
fseek(fp, 0, SEEK_END);
|
||||
state.romlen = ftell(fp);
|
||||
|
||||
state.rom = malloc(state.romlen);
|
||||
|
||||
if (!state.rom) {
|
||||
fprintf(stderr, "Unable to allocate memory for ROM\n");
|
||||
exit(1);
|
||||
}
|
||||
|
||||
fseek(fp, 0, SEEK_SET);
|
||||
fread(state.rom, state.romlen, 1, fp);
|
||||
fclose(fp);
|
||||
}
|
||||
|
||||
/**
|
||||
* Write the ROM to the given filename and truncate it to 32MB.
|
||||
*/
|
||||
void rom_write(char *filename)
|
||||
{
|
||||
FILE *fp = fopen(filename, "wb");
|
||||
|
||||
if (!fp) {
|
||||
fprintf(stderr, "Unable to open \"%s\" for writing\n", filename);
|
||||
exit(1);
|
||||
}
|
||||
|
||||
fwrite(state.rom, 1024 * 1024 * 32, 1, fp);
|
||||
fclose(fp);
|
||||
}
|
||||
|
||||
static uint32_t rol(uint32_t i, uint32_t b)
|
||||
{
|
||||
return (i << b) | (i >> (32 - b));
|
||||
}
|
||||
|
||||
static uint32_t r4(unsigned char *b)
|
||||
{
|
||||
return b[0] * 0x1000000 + b[1] * 0x10000 + b[2] * 0x100 + b[3];
|
||||
}
|
||||
|
||||
static void crc(unsigned char *rom, uint32_t *crc1, uint32_t *crc2)
|
||||
{
|
||||
uint32_t seed = 0xdf26f436;
|
||||
uint32_t t1 = seed;
|
||||
uint32_t t2 = seed;
|
||||
uint32_t t3 = seed;
|
||||
uint32_t t4 = seed;
|
||||
uint32_t t5 = seed;
|
||||
uint32_t t6 = seed;
|
||||
uint32_t offset;
|
||||
uint32_t d;
|
||||
uint32_t r;
|
||||
uint32_t temp;
|
||||
|
||||
unsigned char *lookup = &rom[0x40 + 0x0710];
|
||||
|
||||
for (offset = 0x1000; offset < 0x101000; offset += 4) {
|
||||
d = r4(&rom[offset]);
|
||||
|
||||
if ((t6 + d) < t6) {
|
||||
t4++;
|
||||
}
|
||||
|
||||
t6 += d;
|
||||
t3 ^= d;
|
||||
|
||||
r = rol(d, d & 0x1f);
|
||||
|
||||
t5 += r;
|
||||
|
||||
if (t2 > d) {
|
||||
t2 ^= r;
|
||||
} else {
|
||||
t2 ^= t6 ^ d;
|
||||
}
|
||||
|
||||
temp = r4(&lookup[offset & 0xff]);
|
||||
t1 += temp ^ d;
|
||||
}
|
||||
|
||||
*crc1 = t6 ^ t4 ^ t3;
|
||||
*crc2 = t5 ^ t2 ^ t1;
|
||||
}
|
||||
|
||||
/**
|
||||
* Calculate the checksum of the ROM and write it to the ROM header.
|
||||
*/
|
||||
void rom_update_crc(void)
|
||||
{
|
||||
uint32_t crc1;
|
||||
uint32_t crc2;
|
||||
|
||||
crc(state.rom, &crc1, &crc2);
|
||||
|
||||
state.rom[0x10] = (crc1 >> 24) & 0xff;
|
||||
state.rom[0x11] = (crc1 >> 16) & 0xff;
|
||||
state.rom[0x12] = (crc1 >> 8) & 0xff;
|
||||
state.rom[0x13] = crc1 & 0xff;
|
||||
|
||||
state.rom[0x14] = (crc2 >> 24) & 0xff;
|
||||
state.rom[0x15] = (crc2 >> 16) & 0xff;
|
||||
state.rom[0x16] = (crc2 >> 8) & 0xff;
|
||||
state.rom[0x17] = crc2 & 0xff;
|
||||
}
|
||||
Reference in new issue
Block a user