/* inffast.c -- fast decoding
- * Copyright (C) 1995-2006 Mark Adler
+ * Copyright (C) 1995-2008, 2010 Mark Adler
* For conditions of distribution and use, see copyright notice in zlib.h
*/
requires strm->avail_out >= 258 for each loop to avoid checking for
output space.
*/
-void inflate_fast(strm, start)
+void ZLIB_INTERNAL inflate_fast(strm, start)
z_streamp strm;
unsigned start; /* inflate()'s starting value for strm->avail_out */
{
#endif
unsigned wsize; /* window size or zero if not using window */
unsigned whave; /* valid bytes in the window */
- unsigned write; /* window write index */
+ unsigned wnext; /* window write index */
unsigned char FAR *window; /* allocated sliding window, if wsize != 0 */
unsigned long hold; /* local strm->hold */
unsigned bits; /* local strm->bits */
#endif
wsize = state->wsize;
whave = state->whave;
- write = state->write;
+ wnext = state->wnext;
window = state->window;
hold = state->hold;
bits = state->bits;
op = dist - op; /* distance back in window */
if (op > whave) {
if (state->sane) {
- strm->msg = (char *)"invalid distance too far back";
+ strm->msg =
+ (char *)"invalid distance too far back";
state->mode = BAD;
break;
}
#endif
}
from = window - OFF;
- if (write == 0) { /* very common case */
+ if (wnext == 0) { /* very common case */
from += wsize - op;
if (op < len) { /* some from window */
len -= op;
from = out - dist; /* rest from output */
}
}
- else if (write < op) { /* wrap around window */
- from += wsize + write - op;
- op -= write;
+ else if (wnext < op) { /* wrap around window */
+ from += wsize + wnext - op;
+ op -= wnext;
if (op < len) { /* some from end of window */
len -= op;
do {
PUP(out) = PUP(from);
} while (--op);
from = window - OFF;
- if (write < len) { /* some from start of window */
- op = write;
+ if (wnext < len) { /* some from start of window */
+ op = wnext;
len -= op;
do {
PUP(out) = PUP(from);
}
}
else { /* contiguous in window */
- from += write - op;
+ from += wnext - op;
if (op < len) { /* some from window */
len -= op;
do {
inflate_fast() speedups that turned out slower (on a PowerPC G3 750CXe):
- Using bit fields for code structure
- Different op definition to avoid & for extra bits (do & for table bits)
- - Three separate decoding do-loops for direct, window, and write == 0
+ - Three separate decoding do-loops for direct, window, and wnext == 0
- Special case for distance > 1 copies to do overlapped load and store copy
- Explicit branch predictions (based on measured branch probabilities)
- Deferring match copy and interspersed it with decoding subsequent codes