Update lzf to 3.2 fixes texturecache on 64 bit.

git-svn-id: https://svn.eduke32.com/eduke32@742 1a8010ca-5511-0410-912e-c29ae57300e0
This commit is contained in:
qbix79 2008-05-25 08:23:53 +00:00
parent 67e777a899
commit 43aa2bbc23
4 changed files with 337 additions and 258 deletions

View file

@ -1,5 +1,5 @@
/*
* Copyright (c) 2000-2005 Marc Alexander Lehmann <schmorp@schmorp.de>
* Copyright (c) 2000-2007 Marc Alexander Lehmann <schmorp@schmorp.de>
*
* Redistribution and use in source and binary forms, with or without modifica-
* tion, are permitted provided that the following conditions are met:
@ -11,9 +11,6 @@
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* 3. The name of the author may not be used to endorse or promote products
* derived from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR IMPLIED
* WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MER-
* CHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO
@ -26,14 +23,15 @@
* OF THE POSSIBILITY OF SUCH DAMAGE.
*
* Alternatively, the contents of this file may be used under the terms of
* the GNU General Public License version 2 (the "GPL"), in which case the
* provisions of the GPL are applicable instead of the above. If you wish to
* allow the use of your version of this file only under the terms of the
* GPL and not to allow others to use your version of this file under the
* BSD license, indicate your decision by deleting the provisions above and
* replace them with the notice and other provisions required by the GPL. If
* you do not delete the provisions above, a recipient may use your version
* of this file under either the BSD or the GPL.
* the GNU General Public License ("GPL") version 2 or any later version,
* in which case the provisions of the GPL are applicable instead of
* the above. If you wish to allow the use of your version of this file
* only under the terms of the GPL and not to allow others to use your
* version of this file under the BSD license, indicate your decision
* by deleting the provisions above and replace them with the notice
* and other provisions required by the GPL. If you do not delete the
* provisions above, a recipient may use your version of this file under
* either the BSD or the GPL.
*/
#ifndef LZF_H
@ -62,7 +60,7 @@
* the data uncompressed otherwise.
*
* lzf_compress might use different algorithms on different systems and
* even diferent runs, thus might result in different compressed strings
* even different runs, thus might result in different compressed strings
* depending on the phase of the moon or similar factors. However, all
* these strings are architecture-independent and will result in the
* original data when decompressed using lzf_decompress.

View file

@ -1,5 +1,5 @@
/*
* Copyright (c) 2000-2005 Marc Alexander Lehmann <schmorp@schmorp.de>
* Copyright (c) 2000-2007 Marc Alexander Lehmann <schmorp@schmorp.de>
*
* Redistribution and use in source and binary forms, with or without modifica-
* tion, are permitted provided that the following conditions are met:
@ -11,9 +11,6 @@
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* 3. The name of the author may not be used to endorse or promote products
* derived from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR IMPLIED
* WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MER-
* CHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO
@ -26,14 +23,15 @@
* OF THE POSSIBILITY OF SUCH DAMAGE.
*
* Alternatively, the contents of this file may be used under the terms of
* the GNU General Public License version 2 (the "GPL"), in which case the
* provisions of the GPL are applicable instead of the above. If you wish to
* allow the use of your version of this file only under the terms of the
* GPL and not to allow others to use your version of this file under the
* BSD license, indicate your decision by deleting the provisions above and
* replace them with the notice and other provisions required by the GPL. If
* you do not delete the provisions above, a recipient may use your version
* of this file under either the BSD or the GPL.
* the GNU General Public License ("GPL") version 2 or any later version,
* in which case the provisions of the GPL are applicable instead of
* the above. If you wish to allow the use of your version of this file
* only under the terms of the GPL and not to allow others to use your
* version of this file under the BSD license, indicate your decision
* by deleting the provisions above and replace them with the notice
* and other provisions required by the GPL. If you do not delete the
* provisions above, a recipient may use your version of this file under
* either the BSD or the GPL.
*/
#ifndef LZFP_h
@ -46,29 +44,28 @@
#endif
/*
* size of hashtable is (1 << HLOG) * sizeof (char *)
* Size of hashtable is (1 << HLOG) * sizeof (char *)
* decompression is independent of the hash table size
* the difference between 15 and 14 is very small
* for small blocks (and 14 is usually a but faster).
* for small blocks (and 14 is usually a bit faster).
* For a low-memory/faster configuration, use HLOG == 13;
* For best compression, use 15 or 16 (or more).
* For best compression, use 15 or 16 (or more, up to 23).
*/
#ifndef HLOG
# define HLOG 14
# define HLOG 16
#endif
/*
* sacrifice very little compression quality in favour of compression speed.
* Sacrifice very little compression quality in favour of compression speed.
* This gives almost the same compression as the default code, and is
* (very roughly) 15% faster. This is the preferable mode of operation.
* (very roughly) 15% faster. This is the preferred mode of operation.
*/
#ifndef VERY_FAST
# define VERY_FAST 1
#endif
/*
* sacrifice some more compression quality in favour of compression speed.
* Sacrifice some more compression quality in favour of compression speed.
* (roughly 1-2% worse compression for large blocks and
* 9-10% for small, redundant, blocks and >>20% better speed in both cases)
* In short: when in need for speed, enable this for binary data,
@ -79,30 +76,23 @@
#endif
/*
* unconditionally aligning does not cost very much, so do it if unsure
* Unconditionally aligning does not cost very much, so do it if unsure
*/
#ifndef STRICT_ALIGN
# define STRICT_ALIGN !(defined(__i386) || defined (__amd64))
#endif
/*
* use string functions to copy memory.
* this is usually a loss, even with glibc's optimized memcpy
*/
#ifndef USE_MEMCPY
# define USE_MEMCPY 0
#endif
/*
* you may choose to pre-set the hash table (might be faster on some
* modern cpus and large (>>64k) blocks)
* You may choose to pre-set the hash table (might be faster on some
* modern cpus and large (>>64k) blocks, and also makes compression
* deterministic/repeatable when the configuration otherwise is the same).
*/
#ifndef INIT_HTAB
# define INIT_HTAB 0
#endif
/*
* avoid assigning values to errno variable? for some embedding purposes
* Avoid assigning values to errno variable? for some embedding purposes
* (linux kernel for example), this is neccessary. NOTE: this breaks
* the documentation in lzf.h.
*/
@ -119,6 +109,18 @@
# define LZF_STATE_ARG 0
#endif
/*
* Wether to add extra checks for input validity in lzf_decompress
* and return EINVAL if the input stream has been corrupted. This
* only shields against overflowing the input buffer and will not
* detect most corrupted streams.
* This check is not normally noticable on modern hardware
* (<1% slowdown), but might slow down older cpus considerably.
*/
#ifndef CHECK_INPUT
# define CHECK_INPUT 1
#endif
/*****************************************************************************/
/* nothing should be changed below */
@ -145,7 +147,7 @@ typedef const u8 *LZF_STATE[1 << (HLOG)];
# endif
#endif
#if USE_MEMCPY || INIT_HTAB
#if INIT_HTAB
# ifdef __cplusplus
# include <cstring>
# else

View file

@ -1,5 +1,5 @@
/*
* Copyright (c) 2000-2005 Marc Alexander Lehmann <schmorp@schmorp.de>
* Copyright (c) 2000-2007 Marc Alexander Lehmann <schmorp@schmorp.de>
*
* Redistribution and use in source and binary forms, with or without modifica-
* tion, are permitted provided that the following conditions are met:
@ -11,9 +11,6 @@
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* 3. The name of the author may not be used to endorse or promote products
* derived from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR IMPLIED
* WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MER-
* CHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO
@ -26,14 +23,15 @@
* OF THE POSSIBILITY OF SUCH DAMAGE.
*
* Alternatively, the contents of this file may be used under the terms of
* the GNU General Public License version 2 (the "GPL"), in which case the
* provisions of the GPL are applicable instead of the above. If you wish to
* allow the use of your version of this file only under the terms of the
* GPL and not to allow others to use your version of this file under the
* BSD license, indicate your decision by deleting the provisions above and
* replace them with the notice and other provisions required by the GPL. If
* you do not delete the provisions above, a recipient may use your version
* of this file under either the BSD or the GPL.
* the GNU General Public License ("GPL") version 2 or any later version,
* in which case the provisions of the GPL are applicable instead of
* the above. If you wish to allow the use of your version of this file
* only under the terms of the GPL and not to allow others to use your
* version of this file under the BSD license, indicate your decision
* by deleting the provisions above and replace them with the notice
* and other provisions required by the GPL. If you do not delete the
* provisions above, a recipient may use your version of this file under
* either the BSD or the GPL.
*/
#include "lzfP.h"
@ -49,12 +47,18 @@
#ifndef FRST
# define FRST(p) (((p[0]) << 8) | p[1])
# define NEXT(v,p) (((v) << 8) | p[2])
# define IDX(h) ((((h ^ (h << 5)) >> (3*8 - HLOG)) - h*5) & (HSIZE - 1))
# if ULTRA_FAST
# define IDX(h) ((( h >> (3*8 - HLOG)) - h ) & (HSIZE - 1))
# elif VERY_FAST
# define IDX(h) ((( h >> (3*8 - HLOG)) - h*5) & (HSIZE - 1))
# else
# define IDX(h) ((((h ^ (h << 5)) >> (3*8 - HLOG)) - h*5) & (HSIZE - 1))
# endif
#endif
/*
* IDX works because it is very similar to a multiplicative hash, e.g.
* ((h * 57321 >> (3*8 - HLOG)) & (HSIZE - 1))
* the latter is also quite fast on newer CPUs, and sligthly better
* the latter is also quite fast on newer CPUs, and compresses similarly.
*
* the next one is also quite good, albeit slow ;)
* (int)(cos(h & 0xffffff) * 1e6)
@ -69,7 +73,18 @@
#define MAX_LIT (1 << 5)
#define MAX_OFF (1 << 13)
#define MAX_REF ((1 << 8) + (1 << 3))
#define MAX_REF ((1 << 8) + (1 << 3))
#if __GNUC__ >= 3
# define expect(expr,value) __builtin_expect ((expr),(value))
# define inline inline
#else
# define expect(expr,value) (expr)
# define inline static
#endif
#define expect_false(expr) expect ((expr) != 0, 0)
#define expect_true(expr) expect ((expr) != 0, 1)
/*
* compressed format
@ -81,162 +96,194 @@
*/
unsigned int
lzf_compress(const void *const in_data, unsigned int in_len,
void *out_data, unsigned int out_len
lzf_compress (const void *const in_data, unsigned int in_len,
void *out_data, unsigned int out_len
#if LZF_STATE_ARG
, LZF_STATE *htab
, LZF_STATE htab
#endif
)
)
{
#if !LZF_STATE_ARG
LZF_STATE htab;
LZF_STATE htab;
#endif
const u8 **hslot;
const u8 *ip = (const u8 *)in_data;
u8 *op = (u8 *)out_data;
const u8 *in_end = ip + in_len;
u8 *out_end = op + out_len;
const u8 *ref;
const u8 **hslot;
const u8 *ip = (const u8 *)in_data;
u8 *op = (u8 *)out_data;
const u8 *in_end = ip + in_len;
u8 *out_end = op + out_len;
const u8 *ref;
unsigned int hval = FRST(ip);
unsigned int off;
int lit = 0;
/* off requires a type wide enough to hold a general pointer difference.
* ISO C doesn't have that (size_t might not be enough and ptrdiff_t only
* works for differences within a single object). We also assume that no
* no bit pattern traps. Since the only platform that is both non-POSIX
* and fails to support both assumptions is windows 64 bit, we make a
* special workaround for it.
*/
#if defined (WIN32) && defined (_M_X64)
unsigned _int64 off; /* workaround for missing POSIX compliance */
#else
unsigned long off;
#endif
unsigned int hval;
int lit;
if (!in_len || !out_len)
return 0;
#if INIT_HTAB
# if USE_MEMCPY
memset(htab, 0, sizeof(htab));
# else
for (hslot = htab; hslot < htab + HSIZE; hslot++)
*hslot++ = ip;
memset (htab, 0, sizeof (htab));
# if 0
for (hslot = htab; hslot < htab + HSIZE; hslot++)
*hslot++ = ip;
# endif
#endif
for (;;)
{
if (ip < in_end - 2)
{
hval = NEXT(hval, ip);
hslot = htab + IDX(hval);
ref = *hslot; *hslot = ip;
lit = 0; op++; /* start run */
if (1
#if INIT_HTAB && !USE_MEMCPY
&& ref < ip /* the next test will actually take care of this, but this is faster */
hval = FRST (ip);
while (ip < in_end - 2)
{
hval = NEXT (hval, ip);
hslot = htab + IDX (hval);
ref = *hslot; *hslot = ip;
if (1
#if INIT_HTAB
&& ref < ip /* the next test will actually take care of this, but this is faster */
#endif
&& (off = ip - ref - 1) < MAX_OFF
&& ip + 4 < in_end
&& ref > (u8 *)in_data
&& (off = ip - ref - 1) < MAX_OFF
&& ip + 4 < in_end
&& ref > (u8 *)in_data
#if STRICT_ALIGN
&& ref[0] == ip[0]
&& ref[1] == ip[1]
&& ref[2] == ip[2]
&& ref[0] == ip[0]
&& ref[1] == ip[1]
&& ref[2] == ip[2]
#else
&& *(u16 *)ref == *(u16 *)ip
&& ref[2] == ip[2]
&& *(u16 *)ref == *(u16 *)ip
&& ref[2] == ip[2]
#endif
)
{
/* match found at *ref++ */
unsigned int len = 2;
unsigned int maxlen = in_end - ip - len;
maxlen = maxlen > MAX_REF ? MAX_REF : maxlen;
if (op + lit + 1 + 3 >= out_end)
return 0;
do
len++;
while (len < maxlen && ref[len] == ip[len]);
if (lit)
{
*op++ = lit - 1;
lit = -lit;
do
*op++ = ip[lit];
while (++lit);
}
len -= 2;
ip++;
if (len < 7)
{
*op++ = (off >> 8) + (len << 5);
}
else
{
*op++ = (off >> 8) + (7 << 5);
*op++ = len - 7;
}
*op++ = off;
#if ULTRA_FAST || VERY_FAST
ip += len;
#if VERY_FAST && !ULTRA_FAST
--ip;
#endif
hval = FRST(ip);
hval = NEXT(hval, ip);
htab[IDX(hval)] = ip;
ip++;
#if VERY_FAST && !ULTRA_FAST
hval = NEXT(hval, ip);
htab[IDX(hval)] = ip;
ip++;
#endif
#else
do
{
hval = NEXT(hval, ip);
htab[IDX(hval)] = ip;
ip++;
}
while (len--);
#endif
continue;
}
}
else if (ip == in_end)
break;
/* one more literal byte we must copy */
lit++;
ip++;
if (lit == MAX_LIT)
)
{
if (op + 1 + MAX_LIT >= out_end)
return 0;
/* match found at *ref++ */
unsigned int len = 2;
unsigned int maxlen = in_end - ip - len;
maxlen = maxlen > MAX_REF ? MAX_REF : maxlen;
*op++ = MAX_LIT - 1;
#if USE_MEMCPY
memcpy(op, ip - MAX_LIT, MAX_LIT);
op += MAX_LIT;
lit = 0;
#else
lit = -lit;
do
*op++ = ip[lit];
while (++lit);
#endif
}
}
op [- lit - 1] = lit - 1; /* stop run */
op -= !lit; /* undo run if length is zero */
if (lit)
{
if (op + lit + 1 >= out_end)
if (expect_false (op + 3 + 1 >= out_end))
return 0;
*op++ = lit - 1;
lit = -lit;
do
*op++ = ip[lit];
while (++lit);
for (;;)
{
if (expect_true (maxlen > 16))
{
len++; if (ref [len] != ip [len]) break;
len++; if (ref [len] != ip [len]) break;
len++; if (ref [len] != ip [len]) break;
len++; if (ref [len] != ip [len]) break;
len++; if (ref [len] != ip [len]) break;
len++; if (ref [len] != ip [len]) break;
len++; if (ref [len] != ip [len]) break;
len++; if (ref [len] != ip [len]) break;
len++; if (ref [len] != ip [len]) break;
len++; if (ref [len] != ip [len]) break;
len++; if (ref [len] != ip [len]) break;
len++; if (ref [len] != ip [len]) break;
len++; if (ref [len] != ip [len]) break;
len++; if (ref [len] != ip [len]) break;
len++; if (ref [len] != ip [len]) break;
len++; if (ref [len] != ip [len]) break;
}
do
len++;
while (len < maxlen && ref[len] == ip[len]);
break;
}
len -= 2;
ip++;
if (len < 7)
{
*op++ = (off >> 8) + (len << 5);
}
else
{
*op++ = (off >> 8) + ( 7 << 5);
*op++ = len - 7;
}
*op++ = off;
#if ULTRA_FAST || VERY_FAST
ip += len;
#if VERY_FAST && !ULTRA_FAST
--ip;
#endif
hval = FRST (ip);
hval = NEXT (hval, ip);
htab[IDX (hval)] = ip;
ip++;
#if VERY_FAST && !ULTRA_FAST
hval = NEXT (hval, ip);
htab[IDX (hval)] = ip;
ip++;
#endif
#else
do
{
hval = NEXT (hval, ip);
htab[IDX (hval)] = ip;
ip++;
}
while (len--);
#endif
lit = 0; op++; /* start run */
}
else
{
/* one more literal byte we must copy */
if (expect_false (op >= out_end))
return 0;
lit++; *op++ = *ip++;
if (expect_false (lit == MAX_LIT))
{
op [- lit - 1] = lit - 1; /* stop run */
lit = 0; op++; /* start run */
}
}
}
return op - (u8 *) out_data;
if (op + 3 > out_end) /* at most 3 bytes can be missing here */
return 0;
while (ip < in_end)
{
lit++; *op++ = *ip++;
if (expect_false (lit == MAX_LIT))
{
op [- lit - 1] = lit - 1; /* stop run */
lit = 0; op++; /* start run */
}
}
op [- lit - 1] = lit - 1; /* end run */
op -= !lit; /* undo run if length is zero */
return op - (u8 *)out_data;
}

View file

@ -1,5 +1,5 @@
/*
* Copyright (c) 2000-2005 Marc Alexander Lehmann <schmorp@schmorp.de>
* Copyright (c) 2000-2007 Marc Alexander Lehmann <schmorp@schmorp.de>
*
* Redistribution and use in source and binary forms, with or without modifica-
* tion, are permitted provided that the following conditions are met:
@ -11,9 +11,6 @@
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* 3. The name of the author may not be used to endorse or promote products
* derived from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR IMPLIED
* WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MER-
* CHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO
@ -26,14 +23,15 @@
* OF THE POSSIBILITY OF SUCH DAMAGE.
*
* Alternatively, the contents of this file may be used under the terms of
* the GNU General Public License version 2 (the "GPL"), in which case the
* provisions of the GPL are applicable instead of the above. If you wish to
* allow the use of your version of this file only under the terms of the
* GPL and not to allow others to use your version of this file under the
* BSD license, indicate your decision by deleting the provisions above and
* replace them with the notice and other provisions required by the GPL. If
* you do not delete the provisions above, a recipient may use your version
* of this file under either the BSD or the GPL.
* the GNU General Public License ("GPL") version 2 or any later version,
* in which case the provisions of the GPL are applicable instead of
* the above. If you wish to allow the use of your version of this file
* only under the terms of the GPL and not to allow others to use your
* version of this file under the BSD license, indicate your decision
* by deleting the provisions above and replace them with the notice
* and other provisions required by the GPL. If you do not delete the
* provisions above, a recipient may use your version of this file under
* either the BSD or the GPL.
*/
#include "lzfP.h"
@ -45,72 +43,106 @@
# define SET_ERRNO(n) errno = (n)
#endif
#if (__i386 || __amd64) && __GNUC__ >= 3
# define lzf_movsb(dst, src, len) \
asm ("rep movsb" \
: "=D" (dst), "=S" (src), "=c" (len) \
: "0" (dst), "1" (src), "2" (len));
#endif
unsigned int
lzf_decompress(const void *const in_data, unsigned int in_len,
void *out_data, unsigned int out_len)
lzf_decompress (const void *const in_data, unsigned int in_len,
void *out_data, unsigned int out_len)
{
u8 const *ip = (const u8 *)in_data;
u8 *op = (u8 *)out_data;
u8 const *const in_end = ip + in_len;
u8 *const out_end = op + out_len;
u8 const *ip = (const u8 *)in_data;
u8 *op = (u8 *)out_data;
u8 const *const in_end = ip + in_len;
u8 *const out_end = op + out_len;
do
do
{
unsigned int ctrl = *ip++;
unsigned int ctrl = *ip++;
if (ctrl < (1 << 5)) /* literal run */
if (ctrl < (1 << 5)) /* literal run */
{
ctrl++;
ctrl++;
if (op + ctrl > out_end)
if (op + ctrl > out_end)
{
SET_ERRNO(E2BIG);
return 0;
SET_ERRNO (E2BIG);
return 0;
}
#if USE_MEMCPY
memcpy(op, ip, ctrl);
op += ctrl;
ip += ctrl;
#if CHECK_INPUT
if (ip + ctrl > in_end)
{
SET_ERRNO (EINVAL);
return 0;
}
#endif
#ifdef lzf_movsb
lzf_movsb (op, ip, ctrl);
#else
do
*op++ = *ip++;
while (--ctrl);
do
*op++ = *ip++;
while (--ctrl);
#endif
}
else /* back reference */
else /* back reference */
{
unsigned int len = ctrl >> 5;
unsigned int len = ctrl >> 5;
u8 *ref = op - ((ctrl & 0x1f) << 8) - 1;
u8 *ref = op - ((ctrl & 0x1f) << 8) - 1;
if (len == 7)
len += *ip++;
ref -= *ip++;
if (op + len + 2 > out_end)
#if CHECK_INPUT
if (ip >= in_end)
{
SET_ERRNO(E2BIG);
return 0;
SET_ERRNO (EINVAL);
return 0;
}
#endif
if (len == 7)
{
len += *ip++;
#if CHECK_INPUT
if (ip >= in_end)
{
SET_ERRNO (EINVAL);
return 0;
}
#endif
}
if (ref < (u8 *)out_data)
ref -= *ip++;
if (op + len + 2 > out_end)
{
SET_ERRNO(EINVAL);
return 0;
SET_ERRNO (E2BIG);
return 0;
}
*op++ = *ref++;
*op++ = *ref++;
if (ref < (u8 *)out_data)
{
SET_ERRNO (EINVAL);
return 0;
}
do
*op++ = *ref++;
while (--len);
#ifdef lzf_movsb
len += 2;
lzf_movsb (op, ref, len);
#else
*op++ = *ref++;
*op++ = *ref++;
do
*op++ = *ref++;
while (--len);
#endif
}
}
while (op < out_end && ip < in_end);
while (ip < in_end);
return op - (u8 *)out_data;
return op - (u8 *)out_data;
}