482 lines
13 KiB
C
482 lines
13 KiB
C
/* vim: tabstop=4 shiftwidth=4 noexpandtab
|
|
* This file is part of ToaruOS and is released under the terms
|
|
* of the NCSA / University of Illinois License - see LICENSE.md
|
|
* Copyright (C) 2020 K. Lange
|
|
*
|
|
* libtoaru_inflate: Methods for decompressing DEFLATE and gzip payloads.
|
|
*/
|
|
#include <stdint.h>
|
|
#include <stddef.h>
|
|
|
|
#ifndef _BOOT_LOADER
|
|
#include <toaru/inflate.h>
|
|
#endif
|
|
|
|
/**
|
|
* Decoded Huffman table
|
|
*/
|
|
struct huff {
|
|
uint16_t counts[16]; /* Number of symbols of each length */
|
|
uint16_t symbols[288]; /* Ordered symbols */
|
|
};
|
|
|
|
/**
|
|
* 32K ringbuffer for backwards lookup
|
|
*/
|
|
struct huff_ring {
|
|
size_t pointer;
|
|
uint8_t data[32768];
|
|
};
|
|
|
|
/**
|
|
* Fixed Huffman code tables, generated later.
|
|
*/
|
|
struct huff fixed_lengths;
|
|
struct huff fixed_dists;
|
|
|
|
/**
|
|
* Read a little-endian short from the input.
|
|
*/
|
|
static uint16_t read_16le(struct inflate_context * ctx) {
|
|
uint16_t a, b;
|
|
a = ctx->get_input(ctx);
|
|
b = ctx->get_input(ctx);
|
|
return (a << 0) | (b << 8);
|
|
}
|
|
|
|
/**
|
|
* Read a single bit from the source.
|
|
* Fills the byte buffer with one byte when it runs out.
|
|
*/
|
|
static uint8_t read_bit(struct inflate_context * ctx) {
|
|
|
|
/* When we run out of bits... */
|
|
if (ctx->buffer_size == 0) {
|
|
/* Refill from the next input byte */
|
|
ctx->bit_buffer = ctx->get_input(ctx);
|
|
/* And restore bit buffer size to 8 bits */
|
|
ctx->buffer_size = 8;
|
|
}
|
|
|
|
/* Get the next available bit */
|
|
int out = ctx->bit_buffer & 1;
|
|
|
|
/* Shift the bit buffer forward */
|
|
ctx->bit_buffer >>= 1;
|
|
|
|
/* There is now one less bit available */
|
|
ctx->buffer_size--;
|
|
|
|
return out;
|
|
}
|
|
|
|
/**
|
|
* Read multible bits, in bit order, from the source.
|
|
*/
|
|
static uint32_t read_bits(struct inflate_context * ctx, unsigned int count) {
|
|
uint32_t out = 0;
|
|
for (unsigned int bit = 0; bit < count; bit++) {
|
|
/* Read one bit at a time, from least to most significant */
|
|
out |= (read_bit(ctx) << bit);
|
|
}
|
|
return out;
|
|
}
|
|
|
|
/**
|
|
* Build a Huffman table from an array of lengths.
|
|
*/
|
|
static void build_huffman(uint8_t * lengths, size_t size, struct huff * out) {
|
|
|
|
uint16_t offsets[16];
|
|
unsigned int count = 0;
|
|
|
|
/* Zero symbol counts */
|
|
for (unsigned int i = 0; i < 16; ++i) out->counts[i] = 0;
|
|
|
|
/* Count symbols */
|
|
for (unsigned int i = 0; i < size; ++i) out->counts[lengths[i]]++;
|
|
|
|
/* Special case... */
|
|
out->counts[0] = 0;
|
|
|
|
/* Figure out offsets */
|
|
for (unsigned int i = 0; i < 16; ++i) {
|
|
offsets[i] = count;
|
|
count += out->counts[i];
|
|
}
|
|
|
|
/* Build symbol ordering */
|
|
for (unsigned int i = 0; i < size; ++i) {
|
|
if (lengths[i]) out->symbols[offsets[lengths[i]]++] = i;
|
|
}
|
|
}
|
|
|
|
/**
|
|
* Build the fixed Huffman tables
|
|
*/
|
|
static void build_fixed(void) {
|
|
/* From 3.2.6:
|
|
* Lit Value Bits Codes
|
|
* --------- ---- -----
|
|
* 0 - 143 8 00110000 through
|
|
* 10111111
|
|
* 144 - 255 9 110010000 through
|
|
* 111111111
|
|
* 256 - 279 7 0000000 through
|
|
* 0010111
|
|
* 280 - 287 8 11000000 through
|
|
* 11000111
|
|
*/
|
|
uint8_t lengths[288];
|
|
for (int i = 0; i < 144; ++i) lengths[i] = 8;
|
|
for (int i = 144; i < 256; ++i) lengths[i] = 9;
|
|
for (int i = 256; i < 280; ++i) lengths[i] = 7;
|
|
for (int i = 280; i < 288; ++i) lengths[i] = 8;
|
|
build_huffman(lengths, 288, &fixed_lengths);
|
|
|
|
/* Continued from 3.2.6:
|
|
* Distance codes 0-31 are represented by (fixed-length) 5-bit
|
|
* codes, with possible additional bits as shown in the table
|
|
* shown in Paragraph 3.2.5, above. Note that distance codes 30-
|
|
* 31 will never actually occur in the compressed data.
|
|
*/
|
|
for (int i = 0; i < 30; ++i) lengths[i] = 5;
|
|
build_huffman(lengths, 30, &fixed_dists);
|
|
}
|
|
|
|
|
|
/**
|
|
* Decode a symbol from the source using a Huffman table.
|
|
*/
|
|
static int decode(struct inflate_context * ctx, struct huff * huff) {
|
|
int count = 0, cur = 0;
|
|
for (int i = 1; cur >= 0; i++) {
|
|
cur = (cur << 1) | read_bit(ctx); /* Shift */
|
|
count += huff->counts[i];
|
|
cur -= huff->counts[i];
|
|
}
|
|
return huff->symbols[count + cur];
|
|
}
|
|
|
|
/**
|
|
* Emit one byte to the output, maintaining the ringbuffer.
|
|
* The ringbuffer ensures we can always look back 32K bytes
|
|
* while keeping output streaming.
|
|
*/
|
|
static void emit(struct inflate_context * ctx, unsigned char byte) {
|
|
if (ctx->ring->pointer == 32768) {
|
|
ctx->ring->pointer = 0;
|
|
}
|
|
|
|
ctx->ring->data[ctx->ring->pointer] = byte;
|
|
ctx->write_output(ctx, byte);
|
|
ctx->ring->pointer++;
|
|
}
|
|
|
|
/**
|
|
* Look backwards in the output ring buffer.
|
|
*/
|
|
static uint8_t peek(struct inflate_context * ctx, int offset) {
|
|
return ctx->ring->data[(ctx->ring->pointer - offset) % 32768];
|
|
}
|
|
|
|
/**
|
|
* Decompress a block of Huffman-encoded data.
|
|
*/
|
|
static int inflate(struct inflate_context * ctx, struct huff * huff_len, struct huff * huff_dist) {
|
|
|
|
/* These are the extra bits for lengths from the tables in section 3.2.5
|
|
* Extra Extra Extra
|
|
* Code Bits Length(s) Code Bits Lengths Code Bits Length(s)
|
|
* ---- ---- ------ ---- ---- ------- ---- ---- -------
|
|
* 257 0 3 267 1 15,16 277 4 67-82
|
|
* 258 0 4 268 1 17,18 278 4 83-98
|
|
* 259 0 5 269 2 19-22 279 4 99-114
|
|
* 260 0 6 270 2 23-26 280 4 115-130
|
|
* 261 0 7 271 2 27-30 281 5 131-162
|
|
* 262 0 8 272 2 31-34 282 5 163-194
|
|
* 263 0 9 273 3 35-42 283 5 195-226
|
|
* 264 0 10 274 3 43-50 284 5 227-257
|
|
* 265 1 11,12 275 3 51-58 285 0 258
|
|
* 266 1 13,14 276 3 59-66
|
|
*/
|
|
static const uint16_t lens[] = {
|
|
3, 4, 5, 6, 7, 8, 9, 10, 11, 13, 15, 17, 19, 23, 27, 31, 35, 43, 51,
|
|
59, 67, 83, 99, 115, 131, 163, 195, 227, 258
|
|
};
|
|
static const uint16_t lext[] = {
|
|
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
|
|
};
|
|
|
|
/* Extra bits for distances....
|
|
* Extra Extra Extra
|
|
* Code Bits Dist Code Bits Dist Code Bits Distance
|
|
* ---- ---- ---- ---- ---- ------ ---- ---- --------
|
|
* 0 0 1 10 4 33-48 20 9 1025-1536
|
|
* 1 0 2 11 4 49-64 21 9 1537-2048
|
|
* 2 0 3 12 5 65-96 22 10 2049-3072
|
|
* 3 0 4 13 5 97-128 23 10 3073-4096
|
|
* 4 1 5,6 14 6 129-192 24 11 4097-6144
|
|
* 5 1 7,8 15 6 193-256 25 11 6145-8192
|
|
* 6 2 9-12 16 7 257-384 26 12 8193-12288
|
|
* 7 2 13-16 17 7 385-512 27 12 12289-16384
|
|
* 8 3 17-24 18 8 513-768 28 13 16385-24576
|
|
* 9 3 25-32 19 8 769-1024 29 13 24577-32768
|
|
*/
|
|
static const uint16_t dists[] = {
|
|
1, 2, 3, 4, 5, 7, 9, 13, 17, 25, 33, 49, 65, 97, 129, 193, 257, 385,
|
|
513, 769, 1025, 1537, 2049, 3073, 4097, 6145, 8193, 12289, 16385, 24577
|
|
};
|
|
static const uint16_t dext[] = {
|
|
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
|
|
};
|
|
|
|
while (1) {
|
|
int symbol = decode(ctx, huff_len);
|
|
|
|
if (symbol == 256) {
|
|
break;
|
|
}
|
|
|
|
if (symbol < 256) {
|
|
emit(ctx, symbol);
|
|
} else if (symbol == 256) {
|
|
/* "The literal/length symbol 256 (end of data), ..." */
|
|
break;
|
|
} else {
|
|
int length, distance, offset;
|
|
|
|
symbol -= 257;
|
|
length = read_bits(ctx, lext[symbol]) + lens[symbol];
|
|
distance = decode(ctx, huff_dist);
|
|
offset = read_bits(ctx, dext[distance]) + dists[distance];
|
|
|
|
for (int i = 0; i < length; ++i) {
|
|
uint8_t b = peek(ctx, offset);
|
|
emit(ctx, b);
|
|
}
|
|
}
|
|
}
|
|
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* Decode a dynamic Huffman block.
|
|
*/
|
|
static void decode_huffman(struct inflate_context * ctx) {
|
|
|
|
/* Ordering of code length codes:
|
|
* (HCLEN + 4) x 3 bits: code lengths for the code length
|
|
* alphabet given just above, in the order: ...
|
|
*/
|
|
static const uint8_t clens[] = {
|
|
16, 17, 18, 0, 8, 7, 9, 6, 10, 5, 11, 4, 12, 3, 13, 2, 14, 1, 15
|
|
};
|
|
|
|
unsigned int literals, distances, clengths;
|
|
uint8_t lengths[320] = {0};
|
|
|
|
literals = 257 + read_bits(ctx, 5); /* 5 Bits: HLIT ... 257 */
|
|
distances = 1 + read_bits(ctx, 5); /* 5 Bits: HDIST ... 1 */
|
|
clengths = 4 + read_bits(ctx, 4); /* 4 Bits: HCLEN ... 4 */
|
|
|
|
/* (HCLEN + 4) x 3 bits... */
|
|
for (unsigned int i = 0; i < clengths; ++i) {
|
|
lengths[clens[i]] = read_bits(ctx, 3);
|
|
}
|
|
|
|
struct huff codes;
|
|
build_huffman(lengths, 19, &codes);
|
|
|
|
/* Decode symbols:
|
|
* HLIT + 257 code lengths for the literal/length alphabet...
|
|
* HDIST + 1 code lengths for the distance alphabet...
|
|
*/
|
|
unsigned int count = 0;
|
|
while (count < literals + distances) {
|
|
int symbol = decode(ctx, &codes);
|
|
|
|
if (symbol < 16) {
|
|
/* 0 - 15: Represent code lengths of 0-15 */
|
|
lengths[count++] = symbol;
|
|
} else if (symbol < 19) {
|
|
int rep = 0, length;
|
|
if (symbol == 16) {
|
|
/* 16: Copy the previous code length 3-6 times */
|
|
rep = lengths[count-1];
|
|
length = read_bits(ctx, 2) + 3; /* The next 2 bits indicate repeat length */
|
|
} else if (symbol == 17) {
|
|
/* Repeat a code length of 0 for 3 - 10 times */
|
|
length = read_bits(ctx, 3) + 3; /* 3 bits of length */
|
|
} else if (symbol == 18) {
|
|
/* Repeat a code length of 0 for 11 - 138 times */
|
|
length = read_bits(ctx, 7) + 11; /* 7 bits of length */
|
|
}
|
|
do {
|
|
lengths[count++] = rep;
|
|
length--;
|
|
} while (length);
|
|
} else {
|
|
break;
|
|
}
|
|
}
|
|
|
|
/* Build tables from lenghts decoded above */
|
|
struct huff huff_len;
|
|
build_huffman(lengths, literals, &huff_len);
|
|
struct huff huff_dist;
|
|
build_huffman(lengths + literals, distances, &huff_dist);
|
|
|
|
inflate(ctx, &huff_len, &huff_dist);
|
|
}
|
|
|
|
/**
|
|
* Decode an uncompressed block.
|
|
*/
|
|
static int uncompressed(struct inflate_context * ctx) {
|
|
/* Reset byte alignment */
|
|
ctx->bit_buffer = 0;
|
|
ctx->buffer_size = 0;
|
|
|
|
/* "The rest of the block consists of the following information:"
|
|
* 0 1 2 3 4...
|
|
* +---+---+---+---+================================+
|
|
* | LEN | NLEN |... LEN bytes of literal data...|
|
|
* +---+---+---+---+================================+
|
|
*/
|
|
uint16_t len = read_16le(ctx); /* "the number of data bytes in the block" */
|
|
uint16_t nlen = read_16le(ctx); /* "the one's complement of LEN */
|
|
|
|
/* Sanity check - does the ones-complement length actually match? */
|
|
if ((nlen & 0xFFFF) != (~len & 0xFFFF)) {
|
|
return 1;
|
|
}
|
|
|
|
/* Emit LEN bytes from the source to the output */
|
|
for (int i = 0; i < len; ++i) {
|
|
emit(ctx, ctx->get_input(ctx));
|
|
}
|
|
|
|
return 0;
|
|
}
|
|
|
|
static struct huff_ring data = {0, {0}};
|
|
|
|
/**
|
|
* Decompress DEFLATE-compressed data.
|
|
*/
|
|
int deflate_decompress(struct inflate_context * ctx) {
|
|
ctx->bit_buffer = 0;
|
|
ctx->buffer_size = 0;
|
|
|
|
build_fixed();
|
|
|
|
if (!ctx->ring) {
|
|
ctx->ring = &data;
|
|
}
|
|
|
|
/* read compressed data */
|
|
while (1) {
|
|
/* Read bit */
|
|
|
|
int is_final = read_bit(ctx);
|
|
int type = read_bits(ctx, 2);
|
|
|
|
switch (type) {
|
|
case 0x00: /* BTYPE=00 Non-compressed blocks */
|
|
uncompressed(ctx);
|
|
break;
|
|
case 0x01: /* BYTPE=01 Compressed with fixed Huffman codes */
|
|
inflate(ctx, &fixed_lengths, &fixed_dists);
|
|
break;
|
|
case 0x02: /* BTYPE=02 Compression with dynamic Huffman codes */
|
|
decode_huffman(ctx);
|
|
break;
|
|
case 0x03:
|
|
return 1;
|
|
}
|
|
|
|
if (is_final) {
|
|
break;
|
|
}
|
|
}
|
|
|
|
return 0;
|
|
}
|
|
|
|
#define GZIP_FLAG_TEXT (1 << 0)
|
|
#define GZIP_FLAG_HCRC (1 << 1)
|
|
#define GZIP_FLAG_EXTR (1 << 2)
|
|
#define GZIP_FLAG_NAME (1 << 3)
|
|
#define GZIP_FLAG_COMM (1 << 4)
|
|
|
|
static unsigned int read_32le(struct inflate_context * ctx) {
|
|
unsigned int a, b, c, d;
|
|
a = ctx->get_input(ctx);
|
|
b = ctx->get_input(ctx);
|
|
c = ctx->get_input(ctx);
|
|
d = ctx->get_input(ctx);
|
|
|
|
return (d << 24) | (c << 16) | (b << 8) | (a << 0);
|
|
}
|
|
|
|
int gzip_decompress(struct inflate_context * ctx) {
|
|
|
|
/* Read gzip headers */
|
|
if (ctx->get_input(ctx) != 0x1F) return 1;
|
|
if (ctx->get_input(ctx) != 0x8B) return 1;
|
|
|
|
unsigned int cm = ctx->get_input(ctx);
|
|
if (cm != 8) return 1;
|
|
|
|
unsigned int flags = ctx->get_input(ctx);
|
|
|
|
/* Read mtime */
|
|
unsigned int mtime = read_32le(ctx);
|
|
(void)mtime;
|
|
|
|
/* Read extra flags */
|
|
unsigned int xflags = ctx->get_input(ctx);
|
|
(void)xflags;
|
|
|
|
/* Read and discord OS flag */
|
|
unsigned int os = ctx->get_input(ctx);
|
|
(void)os;
|
|
|
|
/* Extra bytes */
|
|
if (flags & GZIP_FLAG_EXTR) {
|
|
unsigned short size = read_16le(ctx);
|
|
for (unsigned int i = 0; i < size; ++i) ctx->get_input(ctx);
|
|
}
|
|
|
|
if (flags & GZIP_FLAG_NAME) {
|
|
unsigned int c;
|
|
while ((c = ctx->get_input(ctx)) != 0);
|
|
}
|
|
|
|
if (flags & GZIP_FLAG_COMM) {
|
|
unsigned int c;
|
|
while ((c = ctx->get_input(ctx)) != 0);
|
|
}
|
|
|
|
unsigned int crc16 = 0;
|
|
if (flags & GZIP_FLAG_HCRC) {
|
|
crc16 = read_16le(ctx);
|
|
}
|
|
(void)crc16;
|
|
|
|
int status = deflate_decompress(ctx);
|
|
|
|
/* Read CRC and decompressed size from end of input */
|
|
unsigned int crc32 = read_32le(ctx);
|
|
unsigned int dsize = read_32le(ctx);
|
|
|
|
(void)crc32;
|
|
(void)dsize;
|
|
|
|
return status;
|
|
}
|