mirror of
https://github.com/jart/cosmopolitan.git
synced 2025-07-05 02:38:31 +00:00
Make AARCH64 harder, better, faster, stronger
- Perform some housekeeping on scalar math function code - Import ARM's Optimized Routines for SIMD string processing - Upgrade to latest Chromium zlib and enable more SIMD optimizations
This commit is contained in:
Justine Tunney
parent
550b52abf6
commit
cc1732bc42
143 changed files with 15661 additions and 1329 deletions
233
third_party/zlib/deflate.c
vendored
233
third_party/zlib/deflate.c
vendored
|
|
@ -5,13 +5,15 @@
|
|||
* Copyright (C) 1995-2022 Jean-loup Gailly and Mark Adler
|
||||
* For conditions of distribution and use, see copyright notice in zlib.h
|
||||
*/
|
||||
|
||||
#include "libc/assert.h"
|
||||
#include "third_party/zlib/cpu_features.internal.h"
|
||||
#include "third_party/zlib/deflate.internal.h"
|
||||
#include "third_party/zlib/insert_string.internal.h"
|
||||
#include "third_party/zlib/internal.h"
|
||||
#include "third_party/zlib/zutil.internal.h"
|
||||
|
||||
asm(".ident\t\"\\n\\n\
|
||||
deflate 1.2.12.1 (zlib License)\\n\
|
||||
zlib 1.2.13 (zlib License)\\n\
|
||||
Copyright 1995-2022 Jean-loup Gailly and Mark Adler\\n\
|
||||
Invented 1990 Phillip Walter Katz\"");
|
||||
// clang-format off
|
||||
|
|
@ -62,6 +64,12 @@ Invented 1990 Phillip Walter Katz\"");
|
|||
|
||||
/* @(#) $Id$ */
|
||||
|
||||
#if defined(DEFLATE_SLIDE_HASH_SSE2) || defined(DEFLATE_SLIDE_HASH_NEON)
|
||||
#include "third_party/zlib/slide_hash_simd.inc"
|
||||
#endif
|
||||
|
||||
#include "third_party/zlib/insert_string.inc"
|
||||
|
||||
#ifdef FASTEST
|
||||
/* See http://crbug.com/1113596 */
|
||||
#error "FASTEST is not supported in Chromium's zlib."
|
||||
|
|
@ -101,13 +109,7 @@ local void lm_init OF((deflate_state *s));
|
|||
local void putShortMSB OF((deflate_state *s, uInt b));
|
||||
local void flush_pending OF((z_streamp strm));
|
||||
local unsigned read_buf OF((z_streamp strm, Bytef *buf, unsigned size));
|
||||
#ifdef ASMV
|
||||
# pragma message("Assembler code may have bugs -- use at your own risk")
|
||||
void match_init OF((void)); /* asm code initialization */
|
||||
uInt longest_match OF((deflate_state *s, IPos cur_match));
|
||||
#else
|
||||
local uInt longest_match OF((deflate_state *s, IPos cur_match));
|
||||
#endif
|
||||
|
||||
#ifdef ZLIB_DEBUG
|
||||
local void check_match OF((deflate_state *s, IPos start, IPos match,
|
||||
|
|
@ -181,9 +183,9 @@ local const config configuration_table[10] = {
|
|||
*/
|
||||
#define CLEAR_HASH(s) \
|
||||
do { \
|
||||
s->head[s->hash_size-1] = NIL; \
|
||||
s->head[s->hash_size - 1] = NIL; \
|
||||
zmemzero((Bytef *)s->head, \
|
||||
(unsigned)(s->hash_size-1)*sizeof(*s->head)); \
|
||||
(unsigned)(s->hash_size - 1)*sizeof(*s->head)); \
|
||||
} while (0)
|
||||
|
||||
/* ===========================================================================
|
||||
|
|
@ -245,6 +247,14 @@ int ZEXPORT deflateInit2(strm, level, method, windowBits, memLevel, strategy)
|
|||
deflate_state *s;
|
||||
int wrap = 1;
|
||||
|
||||
// Needed to activate optimized insert_string() that helps compression
|
||||
// for all wrapper formats (e.g. RAW, ZLIB, GZIP).
|
||||
// Feature detection is not triggered while using RAW mode (i.e. we never
|
||||
// call crc32() with a NULL buffer).
|
||||
#if defined(CRC32_ARMV8_CRC32) || defined(CRC32_SIMD_SSE42_PCLMUL)
|
||||
cpu_check_features();
|
||||
#endif
|
||||
|
||||
if (strm == Z_NULL) return Z_STREAM_ERROR;
|
||||
|
||||
strm->msg = Z_NULL;
|
||||
|
|
@ -271,6 +281,8 @@ int ZEXPORT deflateInit2(strm, level, method, windowBits, memLevel, strategy)
|
|||
|
||||
if (windowBits < 0) { /* suppress zlib wrapper */
|
||||
wrap = 0;
|
||||
if (windowBits < -15)
|
||||
return Z_STREAM_ERROR;
|
||||
windowBits = -windowBits;
|
||||
}
|
||||
#ifdef GZIP
|
||||
|
|
@ -300,7 +312,7 @@ int ZEXPORT deflateInit2(strm, level, method, windowBits, memLevel, strategy)
|
|||
s->chromium_zlib_hash = 0;
|
||||
#if !defined(USE_ZLIB_RABIN_KARP_ROLLING_HASH)
|
||||
#if defined(TARGET_CPU_WITH_CRC) && defined(CRC32_SIMD_SSE42_PCLMUL)
|
||||
if (X86_HAVE(SSE4_2))
|
||||
if (x86_cpu_enable_simd)
|
||||
s->chromium_zlib_hash = 1;
|
||||
#elif defined(TARGET_CPU_WITH_CRC) && defined(CRC32_ARMV8_CRC32)
|
||||
if (arm_cpu_enable_crc32)
|
||||
|
|
@ -315,16 +327,15 @@ int ZEXPORT deflateInit2(strm, level, method, windowBits, memLevel, strategy)
|
|||
|
||||
s->hash_size = 1 << s->hash_bits;
|
||||
s->hash_mask = s->hash_size - 1;
|
||||
s->hash_shift = ((s->hash_bits+MIN_MATCH-1)/MIN_MATCH);
|
||||
s->hash_shift = ((s->hash_bits + MIN_MATCH-1) / MIN_MATCH);
|
||||
|
||||
s->window = (Bytef *) ZALLOC(strm,
|
||||
s->w_size + window_padding,
|
||||
2*sizeof(Byte));
|
||||
|
||||
/* Avoid use of unitialized values in the window, see crbug.com/1137613 and
|
||||
* crbug.com/1144420 */
|
||||
if (s->window) { /* [jart] fix regression in malloc failure checking */
|
||||
zmemzero(s->window, (s->w_size + window_padding) * (2 * sizeof(Byte)));
|
||||
zmemzero(s->window, (s->w_size + window_padding) * (2 * sizeof(Byte)));
|
||||
}
|
||||
s->prev = (Posf *) ZALLOC(strm, s->w_size, sizeof(Pos));
|
||||
/* Avoid use of uninitialized value, see:
|
||||
|
|
@ -355,11 +366,11 @@ int ZEXPORT deflateInit2(strm, level, method, windowBits, memLevel, strategy)
|
|||
* sym_buf value to read moves forward three bytes. From that symbol, up to
|
||||
* 31 bits are written to pending_buf. The closest the written pending_buf
|
||||
* bits gets to the next sym_buf symbol to read is just before the last
|
||||
* code is written. At that time, 31*(n-2) bits have been written, just
|
||||
* after 24*(n-2) bits have been consumed from sym_buf. sym_buf starts at
|
||||
* 8*n bits into pending_buf. (Note that the symbol buffer fills when n-1
|
||||
* code is written. At that time, 31*(n - 2) bits have been written, just
|
||||
* after 24*(n - 2) bits have been consumed from sym_buf. sym_buf starts at
|
||||
* 8*n bits into pending_buf. (Note that the symbol buffer fills when n - 1
|
||||
* symbols are written.) The closest the writing gets to what is unread is
|
||||
* then n+14 bits. Here n is lit_bufsize, which is 16384 by default, and
|
||||
* then n + 14 bits. Here n is lit_bufsize, which is 16384 by default, and
|
||||
* can range from 128 to 32768.
|
||||
*
|
||||
* Therefore, at a minimum, there are 142 bits of space between what is
|
||||
|
|
@ -404,7 +415,7 @@ int ZEXPORT deflateInit2(strm, level, method, windowBits, memLevel, strategy)
|
|||
/* =========================================================================
|
||||
* Check for a valid deflate stream state. Return 0 if ok, 1 if not.
|
||||
*/
|
||||
local int deflateStateCheck (strm)
|
||||
local int deflateStateCheck(strm)
|
||||
z_streamp strm;
|
||||
{
|
||||
deflate_state *s;
|
||||
|
|
@ -427,7 +438,7 @@ local int deflateStateCheck (strm)
|
|||
}
|
||||
|
||||
/* ========================================================================= */
|
||||
int ZEXPORT deflateSetDictionary (strm, dictionary, dictLength)
|
||||
int ZEXPORT deflateSetDictionary(strm, dictionary, dictLength)
|
||||
z_streamp strm;
|
||||
const Bytef *dictionary;
|
||||
uInt dictLength;
|
||||
|
|
@ -492,7 +503,7 @@ int ZEXPORT deflateSetDictionary (strm, dictionary, dictLength)
|
|||
}
|
||||
|
||||
/* ========================================================================= */
|
||||
int ZEXPORT deflateGetDictionary (strm, dictionary, dictLength)
|
||||
int ZEXPORT deflateGetDictionary(strm, dictionary, dictLength)
|
||||
z_streamp strm;
|
||||
Bytef *dictionary;
|
||||
uInt *dictLength;
|
||||
|
|
@ -514,7 +525,7 @@ int ZEXPORT deflateGetDictionary (strm, dictionary, dictLength)
|
|||
}
|
||||
|
||||
/* ========================================================================= */
|
||||
int ZEXPORT deflateResetKeep (strm)
|
||||
int ZEXPORT deflateResetKeep(strm)
|
||||
z_streamp strm;
|
||||
{
|
||||
deflate_state *s;
|
||||
|
|
@ -552,7 +563,7 @@ int ZEXPORT deflateResetKeep (strm)
|
|||
}
|
||||
|
||||
/* ========================================================================= */
|
||||
int ZEXPORT deflateReset (strm)
|
||||
int ZEXPORT deflateReset(strm)
|
||||
z_streamp strm;
|
||||
{
|
||||
int ret;
|
||||
|
|
@ -564,7 +575,7 @@ int ZEXPORT deflateReset (strm)
|
|||
}
|
||||
|
||||
/* ========================================================================= */
|
||||
int ZEXPORT deflateSetHeader (strm, head)
|
||||
int ZEXPORT deflateSetHeader(strm, head)
|
||||
z_streamp strm;
|
||||
gz_headerp head;
|
||||
{
|
||||
|
|
@ -575,7 +586,7 @@ int ZEXPORT deflateSetHeader (strm, head)
|
|||
}
|
||||
|
||||
/* ========================================================================= */
|
||||
int ZEXPORT deflatePending (strm, pending, bits)
|
||||
int ZEXPORT deflatePending(strm, pending, bits)
|
||||
unsigned *pending;
|
||||
int *bits;
|
||||
z_streamp strm;
|
||||
|
|
@ -589,7 +600,7 @@ int ZEXPORT deflatePending (strm, pending, bits)
|
|||
}
|
||||
|
||||
/* ========================================================================= */
|
||||
int ZEXPORT deflatePrime (strm, bits, value)
|
||||
int ZEXPORT deflatePrime(strm, bits, value)
|
||||
z_streamp strm;
|
||||
int bits;
|
||||
int value;
|
||||
|
|
@ -684,36 +695,50 @@ int ZEXPORT deflateTune(strm, good_length, max_lazy, nice_length, max_chain)
|
|||
}
|
||||
|
||||
/* =========================================================================
|
||||
* For the default windowBits of 15 and memLevel of 8, this function returns
|
||||
* a close to exact, as well as small, upper bound on the compressed size.
|
||||
* They are coded as constants here for a reason--if the #define's are
|
||||
* changed, then this function needs to be changed as well. The return
|
||||
* value for 15 and 8 only works for those exact settings.
|
||||
* For the default windowBits of 15 and memLevel of 8, this function returns a
|
||||
* close to exact, as well as small, upper bound on the compressed size. This
|
||||
* is an expansion of ~0.03%, plus a small constant.
|
||||
*
|
||||
* For any setting other than those defaults for windowBits and memLevel,
|
||||
* the value returned is a conservative worst case for the maximum expansion
|
||||
* resulting from using fixed blocks instead of stored blocks, which deflate
|
||||
* can emit on compressed data for some combinations of the parameters.
|
||||
* For any setting other than those defaults for windowBits and memLevel, one
|
||||
* of two worst case bounds is returned. This is at most an expansion of ~4% or
|
||||
* ~13%, plus a small constant.
|
||||
*
|
||||
* This function could be more sophisticated to provide closer upper bounds for
|
||||
* every combination of windowBits and memLevel. But even the conservative
|
||||
* upper bound of about 14% expansion does not seem onerous for output buffer
|
||||
* allocation.
|
||||
* Both the 0.03% and 4% derive from the overhead of stored blocks. The first
|
||||
* one is for stored blocks of 16383 bytes (memLevel == 8), whereas the second
|
||||
* is for stored blocks of 127 bytes (the worst case memLevel == 1). The
|
||||
* expansion results from five bytes of header for each stored block.
|
||||
*
|
||||
* The larger expansion of 13% results from a window size less than or equal to
|
||||
* the symbols buffer size (windowBits <= memLevel + 7). In that case some of
|
||||
* the data being compressed may have slid out of the sliding window, impeding
|
||||
* a stored block from being emitted. Then the only choice is a fixed or
|
||||
* dynamic block, where a fixed block limits the maximum expansion to 9 bits
|
||||
* per 8-bit byte, plus 10 bits for every block. The smallest block size for
|
||||
* which this can occur is 255 (memLevel == 2).
|
||||
*
|
||||
* Shifts are used to approximate divisions, for speed.
|
||||
*/
|
||||
uLong ZEXPORT deflateBound(strm, sourceLen)
|
||||
z_streamp strm;
|
||||
uLong sourceLen;
|
||||
{
|
||||
deflate_state *s;
|
||||
uLong complen, wraplen;
|
||||
uLong fixedlen, storelen, wraplen;
|
||||
|
||||
/* conservative upper bound for compressed data */
|
||||
complen = sourceLen +
|
||||
((sourceLen + 7) >> 3) + ((sourceLen + 63) >> 6) + 5;
|
||||
/* upper bound for fixed blocks with 9-bit literals and length 255
|
||||
(memLevel == 2, which is the lowest that may not use stored blocks) --
|
||||
~13% overhead plus a small constant */
|
||||
fixedlen = sourceLen + (sourceLen >> 3) + (sourceLen >> 8) +
|
||||
(sourceLen >> 9) + 4;
|
||||
|
||||
/* if can't get parameters, return conservative bound plus zlib wrapper */
|
||||
/* upper bound for stored blocks with length 127 (memLevel == 1) --
|
||||
~4% overhead plus a small constant */
|
||||
storelen = sourceLen + (sourceLen >> 5) + (sourceLen >> 7) +
|
||||
(sourceLen >> 11) + 7;
|
||||
|
||||
/* if can't get parameters, return larger bound plus a zlib wrapper */
|
||||
if (deflateStateCheck(strm))
|
||||
return complen + 6;
|
||||
return (fixedlen > storelen ? fixedlen : storelen) + 6;
|
||||
|
||||
/* compute wrapper length */
|
||||
s = strm->state;
|
||||
|
|
@ -750,11 +775,12 @@ uLong ZEXPORT deflateBound(strm, sourceLen)
|
|||
wraplen = 6;
|
||||
}
|
||||
|
||||
/* if not default parameters, return conservative bound */
|
||||
/* if not default parameters, return one of the conservative bounds */
|
||||
if (s->w_bits != 15 || s->hash_bits != 8 + 7)
|
||||
return complen + wraplen;
|
||||
return (s->w_bits <= s->hash_bits ? fixedlen : storelen) + wraplen;
|
||||
|
||||
/* default settings: return tight bound for that case */
|
||||
/* default settings: return tight bound for that case -- ~0.03% overhead
|
||||
plus a small constant */
|
||||
return sourceLen + (sourceLen >> 12) + (sourceLen >> 14) +
|
||||
(sourceLen >> 25) + 13 - 6 + wraplen;
|
||||
}
|
||||
|
|
@ -764,7 +790,7 @@ uLong ZEXPORT deflateBound(strm, sourceLen)
|
|||
* IN assertion: the stream state is correct and there is enough room in
|
||||
* pending_buf.
|
||||
*/
|
||||
local void putShortMSB (s, b)
|
||||
local void putShortMSB(s, b)
|
||||
deflate_state *s;
|
||||
uInt b;
|
||||
{
|
||||
|
|
@ -811,7 +837,7 @@ local void flush_pending(strm)
|
|||
} while (0)
|
||||
|
||||
/* ========================================================================= */
|
||||
int ZEXPORT deflate (strm, flush)
|
||||
int ZEXPORT deflate(strm, flush)
|
||||
z_streamp strm;
|
||||
int flush;
|
||||
{
|
||||
|
|
@ -866,7 +892,7 @@ int ZEXPORT deflate (strm, flush)
|
|||
s->status = BUSY_STATE;
|
||||
if (s->status == INIT_STATE) {
|
||||
/* zlib header */
|
||||
uInt header = (Z_DEFLATED + ((s->w_bits-8)<<4)) << 8;
|
||||
uInt header = (Z_DEFLATED + ((s->w_bits - 8) << 4)) << 8;
|
||||
uInt level_flags;
|
||||
|
||||
if (s->strategy >= Z_HUFFMAN_ONLY || s->level < 2)
|
||||
|
|
@ -1127,7 +1153,7 @@ int ZEXPORT deflate (strm, flush)
|
|||
}
|
||||
|
||||
/* ========================================================================= */
|
||||
int ZEXPORT deflateEnd (strm)
|
||||
int ZEXPORT deflateEnd(strm)
|
||||
z_streamp strm;
|
||||
{
|
||||
int status;
|
||||
|
|
@ -1153,7 +1179,7 @@ int ZEXPORT deflateEnd (strm)
|
|||
* To simplify the source, this is not supported for 16-bit MSDOS (which
|
||||
* doesn't have enough memory anyway to duplicate compression states).
|
||||
*/
|
||||
int ZEXPORT deflateCopy (dest, source)
|
||||
int ZEXPORT deflateCopy(dest, source)
|
||||
z_streamp dest;
|
||||
z_streamp source;
|
||||
{
|
||||
|
|
@ -1243,7 +1269,7 @@ local unsigned read_buf(strm, buf, size)
|
|||
/* ===========================================================================
|
||||
* Initialize the "longest match" routines for a new zlib stream
|
||||
*/
|
||||
local void lm_init (s)
|
||||
local void lm_init(s)
|
||||
deflate_state *s;
|
||||
{
|
||||
s->window_size = (ulg)2L*s->w_size;
|
||||
|
|
@ -1264,11 +1290,6 @@ local void lm_init (s)
|
|||
s->match_length = s->prev_length = MIN_MATCH-1;
|
||||
s->match_available = 0;
|
||||
s->ins_h = 0;
|
||||
#ifndef FASTEST
|
||||
#ifdef ASMV
|
||||
match_init(); /* initialize the asm code */
|
||||
#endif
|
||||
#endif
|
||||
}
|
||||
|
||||
#ifndef FASTEST
|
||||
|
|
@ -1281,10 +1302,6 @@ local void lm_init (s)
|
|||
* string (strstart) and its distance is <= MAX_DIST, and prev_length >= 1
|
||||
* OUT assertion: the match length is not greater than s->lookahead.
|
||||
*/
|
||||
#ifndef ASMV
|
||||
/* For 80x86 and 680x0, an optimized version will be provided in match.asm or
|
||||
* match.S. The code will be functionally equivalent.
|
||||
*/
|
||||
local uInt longest_match(s, cur_match)
|
||||
deflate_state *s;
|
||||
IPos cur_match; /* current match */
|
||||
|
|
@ -1309,10 +1326,10 @@ local uInt longest_match(s, cur_match)
|
|||
*/
|
||||
register Bytef *strend = s->window + s->strstart + MAX_MATCH - 1;
|
||||
register ush scan_start = *(ushf*)scan;
|
||||
register ush scan_end = *(ushf*)(scan+best_len-1);
|
||||
register ush scan_end = *(ushf*)(scan + best_len - 1);
|
||||
#else
|
||||
register Bytef *strend = s->window + s->strstart + MAX_MATCH;
|
||||
register Byte scan_end1 = scan[best_len-1];
|
||||
register Byte scan_end1 = scan[best_len - 1];
|
||||
register Byte scan_end = scan[best_len];
|
||||
#endif
|
||||
|
||||
|
|
@ -1330,7 +1347,8 @@ local uInt longest_match(s, cur_match)
|
|||
*/
|
||||
if ((uInt)nice_match > s->lookahead) nice_match = (int)s->lookahead;
|
||||
|
||||
Assert((ulg)s->strstart <= s->window_size-MIN_LOOKAHEAD, "need lookahead");
|
||||
Assert((ulg)s->strstart <= s->window_size - MIN_LOOKAHEAD,
|
||||
"need lookahead");
|
||||
|
||||
do {
|
||||
Assert(cur_match < s->strstart, "no future");
|
||||
|
|
@ -1348,15 +1366,15 @@ local uInt longest_match(s, cur_match)
|
|||
/* This code assumes sizeof(unsigned short) == 2. Do not use
|
||||
* UNALIGNED_OK if your compiler uses a different size.
|
||||
*/
|
||||
if (*(ushf*)(match+best_len-1) != scan_end ||
|
||||
if (*(ushf*)(match + best_len - 1) != scan_end ||
|
||||
*(ushf*)match != scan_start) continue;
|
||||
|
||||
/* 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. Compare 2 bytes at a time at
|
||||
* strstart+3, +5, ... up to strstart+257. We check for insufficient
|
||||
* strstart + 3, + 5, up to strstart + 257. We check for insufficient
|
||||
* lookahead only every 4th comparison; the 128th check will be made
|
||||
* at strstart+257. If MAX_MATCH-2 is not a multiple of 8, it is
|
||||
* at strstart + 257. If MAX_MATCH-2 is not a multiple of 8, it is
|
||||
* necessary to put more guard bytes at the end of the window, or
|
||||
* to check more often for insufficient lookahead.
|
||||
*/
|
||||
|
|
@ -1372,28 +1390,29 @@ local uInt longest_match(s, cur_match)
|
|||
}
|
||||
scan++, match++;
|
||||
do {
|
||||
} while (*(ushf*)(scan+=2) == *(ushf*)(match+=2) &&
|
||||
*(ushf*)(scan+=2) == *(ushf*)(match+=2) &&
|
||||
*(ushf*)(scan+=2) == *(ushf*)(match+=2) &&
|
||||
*(ushf*)(scan+=2) == *(ushf*)(match+=2) &&
|
||||
} while (*(ushf*)(scan += 2) == *(ushf*)(match += 2) &&
|
||||
*(ushf*)(scan += 2) == *(ushf*)(match += 2) &&
|
||||
*(ushf*)(scan += 2) == *(ushf*)(match += 2) &&
|
||||
*(ushf*)(scan += 2) == *(ushf*)(match += 2) &&
|
||||
scan < strend);
|
||||
/* The funny "do {}" generates better code on most compilers */
|
||||
|
||||
/* Here, scan <= window+strstart+257 */
|
||||
Assert(scan <= s->window+(unsigned)(s->window_size-1), "wild scan");
|
||||
/* Here, scan <= window + strstart + 257 */
|
||||
Assert(scan <= s->window+(unsigned)(s->window_size - 1),
|
||||
"wild scan");
|
||||
if (*scan == *match) scan++;
|
||||
|
||||
len = (MAX_MATCH - 1) - (int)(strend-scan);
|
||||
len = (MAX_MATCH - 1) - (int)(strend - scan);
|
||||
scan = strend - (MAX_MATCH-1);
|
||||
|
||||
#else /* UNALIGNED_OK */
|
||||
|
||||
if (match[best_len] != scan_end ||
|
||||
match[best_len-1] != scan_end1 ||
|
||||
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
|
||||
/* 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
|
||||
|
|
@ -1412,7 +1431,7 @@ local uInt longest_match(s, cur_match)
|
|||
}
|
||||
|
||||
/* We check for insufficient lookahead only every 8th comparison;
|
||||
* the 256th check will be made at strstart+258.
|
||||
* the 256th check will be made at strstart + 258.
|
||||
*/
|
||||
do {
|
||||
} while (*++scan == *++match && *++scan == *++match &&
|
||||
|
|
@ -1421,7 +1440,8 @@ local uInt longest_match(s, cur_match)
|
|||
*++scan == *++match && *++scan == *++match &&
|
||||
scan < strend);
|
||||
|
||||
Assert(scan <= s->window+(unsigned)(s->window_size-1), "wild scan");
|
||||
Assert(scan <= s->window + (unsigned)(s->window_size - 1),
|
||||
"wild scan");
|
||||
|
||||
len = MAX_MATCH - (int)(strend - scan);
|
||||
scan = strend - MAX_MATCH;
|
||||
|
|
@ -1433,9 +1453,9 @@ local uInt longest_match(s, cur_match)
|
|||
best_len = len;
|
||||
if (len >= nice_match) break;
|
||||
#ifdef UNALIGNED_OK
|
||||
scan_end = *(ushf*)(scan+best_len-1);
|
||||
scan_end = *(ushf*)(scan + best_len - 1);
|
||||
#else
|
||||
scan_end1 = scan[best_len-1];
|
||||
scan_end1 = scan[best_len - 1];
|
||||
scan_end = scan[best_len];
|
||||
#endif
|
||||
}
|
||||
|
|
@ -1445,7 +1465,6 @@ local uInt longest_match(s, cur_match)
|
|||
if ((uInt)best_len <= s->lookahead) return (uInt)best_len;
|
||||
return s->lookahead;
|
||||
}
|
||||
#endif /* ASMV */
|
||||
|
||||
#else /* FASTEST */
|
||||
|
||||
|
|
@ -1466,7 +1485,8 @@ local uInt longest_match(s, cur_match)
|
|||
*/
|
||||
Assert(s->hash_bits >= 8 && MAX_MATCH == 258, "Code too clever");
|
||||
|
||||
Assert((ulg)s->strstart <= s->window_size-MIN_LOOKAHEAD, "need lookahead");
|
||||
Assert((ulg)s->strstart <= s->window_size - MIN_LOOKAHEAD,
|
||||
"need lookahead");
|
||||
|
||||
Assert(cur_match < s->strstart, "no future");
|
||||
|
||||
|
|
@ -1476,7 +1496,7 @@ local uInt longest_match(s, cur_match)
|
|||
*/
|
||||
if (match[0] != scan[0] || match[1] != scan[1]) return MIN_MATCH-1;
|
||||
|
||||
/* The check at best_len-1 can be removed because it will be made
|
||||
/* 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
|
||||
|
|
@ -1486,7 +1506,7 @@ local uInt longest_match(s, cur_match)
|
|||
Assert(*scan == *match, "match[2]?");
|
||||
|
||||
/* We check for insufficient lookahead only every 8th comparison;
|
||||
* the 256th check will be made at strstart+258.
|
||||
* the 256th check will be made at strstart + 258.
|
||||
*/
|
||||
do {
|
||||
} while (*++scan == *++match && *++scan == *++match &&
|
||||
|
|
@ -1495,7 +1515,7 @@ local uInt longest_match(s, cur_match)
|
|||
*++scan == *++match && *++scan == *++match &&
|
||||
scan < strend);
|
||||
|
||||
Assert(scan <= s->window+(unsigned)(s->window_size-1), "wild scan");
|
||||
Assert(scan <= s->window + (unsigned)(s->window_size - 1), "wild scan");
|
||||
|
||||
len = MAX_MATCH - (int)(strend - scan);
|
||||
|
||||
|
|
@ -1531,7 +1551,7 @@ local void check_match(s, start, match, length)
|
|||
z_error(__FILE__, __LINE__, "invalid match");
|
||||
}
|
||||
if (z_verbose > 1) {
|
||||
kprintf("\\[%d,%d]", start-match, length);
|
||||
kprintf("\\[%d,%d]", start - match, length);
|
||||
do { kprintf("%c", s->window[start++]); } while (--length != 0);
|
||||
}
|
||||
}
|
||||
|
|
@ -1577,9 +1597,9 @@ local void fill_window(s)
|
|||
/* 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 (s->strstart >= wsize+MAX_DIST(s)) {
|
||||
if (s->strstart >= wsize + MAX_DIST(s)) {
|
||||
|
||||
zmemcpy(s->window, s->window+wsize, (unsigned)wsize - more);
|
||||
zmemcpy(s->window, s->window + wsize, (unsigned)wsize - more);
|
||||
s->match_start -= wsize;
|
||||
s->strstart -= wsize; /* we now have strstart >= MAX_DIST */
|
||||
s->block_start -= (long) wsize;
|
||||
|
|
@ -1934,7 +1954,7 @@ local block_state deflate_fast(s, flush)
|
|||
if (s->lookahead == 0) break; /* flush the current block */
|
||||
}
|
||||
|
||||
/* Insert the string window[strstart .. strstart+2] in the
|
||||
/* Insert the string window[strstart .. strstart + 2] in the
|
||||
* dictionary, and set hash_head to the head of the hash chain:
|
||||
*/
|
||||
hash_head = NIL;
|
||||
|
|
@ -1984,7 +2004,7 @@ local block_state deflate_fast(s, flush)
|
|||
|
||||
if (!s->chromium_zlib_hash) {
|
||||
s->ins_h = s->window[s->strstart];
|
||||
UPDATE_HASH(s, s->ins_h, s->window[s->strstart+1]);
|
||||
UPDATE_HASH(s, s->ins_h, s->window[s->strstart + 1]);
|
||||
#if MIN_MATCH != 3
|
||||
Call UPDATE_HASH() MIN_MATCH-3 more times
|
||||
#endif
|
||||
|
|
@ -1996,7 +2016,7 @@ local block_state deflate_fast(s, flush)
|
|||
} else {
|
||||
/* No match, output a literal byte */
|
||||
Tracevv(("%c", s->window[s->strstart]));
|
||||
_tr_tally_lit (s, s->window[s->strstart], bflush);
|
||||
_tr_tally_lit(s, s->window[s->strstart], bflush);
|
||||
s->lookahead--;
|
||||
s->strstart++;
|
||||
}
|
||||
|
|
@ -2040,7 +2060,7 @@ local block_state deflate_slow(s, flush)
|
|||
if (s->lookahead == 0) break; /* flush the current block */
|
||||
}
|
||||
|
||||
/* Insert the string window[strstart .. strstart+2] in the
|
||||
/* Insert the string window[strstart .. strstart + 2] in the
|
||||
* dictionary, and set hash_head to the head of the hash chain:
|
||||
*/
|
||||
hash_head = NIL;
|
||||
|
|
@ -2085,20 +2105,20 @@ local block_state deflate_slow(s, flush)
|
|||
if (s->prev_match == -1) {
|
||||
/* The window has slid one byte past the previous match,
|
||||
* so the first byte cannot be compared. */
|
||||
check_match(s, s->strstart, s->prev_match+1, s->prev_length-1);
|
||||
check_match(s, s->strstart, s->prev_match + 1, s->prev_length - 1);
|
||||
} else {
|
||||
check_match(s, s->strstart-1, s->prev_match, s->prev_length);
|
||||
check_match(s, s->strstart - 1, s->prev_match, s->prev_length);
|
||||
}
|
||||
|
||||
_tr_tally_dist(s, s->strstart -1 - s->prev_match,
|
||||
_tr_tally_dist(s, s->strstart - 1 - s->prev_match,
|
||||
s->prev_length - MIN_MATCH, bflush);
|
||||
|
||||
/* Insert in hash table all strings up to the end of the match.
|
||||
* strstart-1 and strstart are already inserted. If there is not
|
||||
* strstart - 1 and strstart are already inserted. If there is not
|
||||
* enough lookahead, the last two strings are not inserted in
|
||||
* the hash table.
|
||||
*/
|
||||
s->lookahead -= s->prev_length-1;
|
||||
s->lookahead -= s->prev_length - 1;
|
||||
s->prev_length -= 2;
|
||||
do {
|
||||
if (++s->strstart <= max_insert) {
|
||||
|
|
@ -2116,8 +2136,8 @@ local block_state deflate_slow(s, flush)
|
|||
* single literal. If there was a match but the current match
|
||||
* is longer, truncate the previous match to a single literal.
|
||||
*/
|
||||
Tracevv(("%c", s->window[s->strstart-1]));
|
||||
_tr_tally_lit(s, s->window[s->strstart-1], bflush);
|
||||
Tracevv(("%c", s->window[s->strstart - 1]));
|
||||
_tr_tally_lit(s, s->window[s->strstart - 1], bflush);
|
||||
if (bflush) {
|
||||
FLUSH_BLOCK_ONLY(s, 0);
|
||||
}
|
||||
|
|
@ -2135,8 +2155,8 @@ local block_state deflate_slow(s, flush)
|
|||
}
|
||||
Assert (flush != Z_NO_FLUSH, "no flush?");
|
||||
if (s->match_available) {
|
||||
Tracevv(("%c", s->window[s->strstart-1]));
|
||||
_tr_tally_lit(s, s->window[s->strstart-1], bflush);
|
||||
Tracevv(("%c", s->window[s->strstart - 1]));
|
||||
_tr_tally_lit(s, s->window[s->strstart - 1], bflush);
|
||||
s->match_available = 0;
|
||||
}
|
||||
s->insert = s->strstart < MIN_MATCH-1 ? s->strstart : MIN_MATCH-1;
|
||||
|
|
@ -2193,7 +2213,8 @@ local block_state deflate_rle(s, flush)
|
|||
if (s->match_length > s->lookahead)
|
||||
s->match_length = s->lookahead;
|
||||
}
|
||||
Assert(scan <= s->window+(uInt)(s->window_size-1), "wild scan");
|
||||
Assert(scan <= s->window + (uInt)(s->window_size - 1),
|
||||
"wild scan");
|
||||
}
|
||||
|
||||
/* Emit match if have run of MIN_MATCH or longer, else emit literal */
|
||||
|
|
@ -2208,7 +2229,7 @@ local block_state deflate_rle(s, flush)
|
|||
} else {
|
||||
/* No match, output a literal byte */
|
||||
Tracevv(("%c", s->window[s->strstart]));
|
||||
_tr_tally_lit (s, s->window[s->strstart], bflush);
|
||||
_tr_tally_lit(s, s->window[s->strstart], bflush);
|
||||
s->lookahead--;
|
||||
s->strstart++;
|
||||
}
|
||||
|
|
@ -2248,7 +2269,7 @@ local block_state deflate_huff(s, flush)
|
|||
/* Output a literal byte */
|
||||
s->match_length = 0;
|
||||
Tracevv(("%c", s->window[s->strstart]));
|
||||
_tr_tally_lit (s, s->window[s->strstart], bflush);
|
||||
_tr_tally_lit(s, s->window[s->strstart], bflush);
|
||||
s->lookahead--;
|
||||
s->strstart++;
|
||||
if (bflush) FLUSH_BLOCK(s, 0);
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue