freebsd-dev/usr.bin/more/ch.c
Tim Vanderhoek e965d6a895 Rather than use an LRU-ordered circular list to store buffered data,
simply keep an index into the last access on the circular list and begin
searches at that point.  An LRU list is inappropriate here since the
vast majority of accesses will occur in the same order that the list
is created in.  The only case where an LRU is remotely useful here is when
reading from a file and the user is jumping to randomish positions and
constantly returning to some central position.  Even for this case it is
such a small optimization as not to be noticed in an interactive program
such as more(1).

This change results in a _tremendously_ noticable speed-up when reading long
files through a pipe (where long = ~200k, machine = ~2.5h single-disk
worldstone).
1999-12-28 05:47:24 +00:00

466 lines
11 KiB
C

/*
* Copyright (c) 1988 Mark Nudleman
* Copyright (c) 1988, 1993
* The Regents of the University of California. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* This product includes software developed by the University of
* California, Berkeley and its contributors.
* 4. Neither the name of the University nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*/
#ifndef lint
static char sccsid[] = "@(#)ch.c 8.1 (Berkeley) 6/6/93";
#endif /* not lint */
#ifndef lint
static const char rcsid[] =
"$FreeBSD$";
#endif /* not lint */
/*
* Low level character input from the input file.
* We use these special purpose routines which optimize moving
* both forward and backward from the current read pointer.
*/
#include <sys/types.h>
#include <sys/file.h>
#include <unistd.h>
#include <stdio.h>
#include <less.h>
int file = -1; /* File descriptor of the input file */
/*
* Pool of buffers holding the most recently used blocks of the input file.
*/
struct buf {
struct buf *next, *prev;
long block;
int datasize;
char data[BUFSIZ];
};
int nbufs;
/*
* The buffer pool is kept as a doubly-linked circular list. For the ispipe
* case, this list will always be ordered from highest-numbered block downto
* lowest-numbered block, skipping no blocks. For the !ispipe case,
* it may become disordered. It is not clear that this is a feature.
*/
#define END_OF_CHAIN ((struct buf *)&buf_anchor)
#define buf_head buf_anchor.next
#define buf_tail buf_anchor.prev
static struct {
struct buf *next, *prev;
long block; /* this is never changed from -1 */
} buf_anchor = { END_OF_CHAIN, END_OF_CHAIN, (long)-1 };
/*
* The last buffer in the circular list that was accessed, and correspondingly
* the most likely to be accessed in the future.
*/
static struct buf *buf_lastacc = END_OF_CHAIN;
extern int ispipe, cbufs, sigs;
/*
* Current position in file.
* Stored as a block number and an offset into the block.
*/
static long ch_block;
static int ch_offset;
/* Length of file, needed if input is a pipe. */
static off_t ch_fsize;
/* Number of bytes read, if input is standard input (a pipe). */
static off_t last_piped_pos;
/*
* Get the character pointed to by the read pointer. ch_get() is a macro
* which is more efficient to call than fch_get (the function), in the usual
* case that the block desired is at the head of the chain.
*/
#define ch_get() \
((buf_lastacc->block == ch_block && \
ch_offset < buf_lastacc->datasize) ? \
(unsigned char)buf_lastacc->data[ch_offset] : fch_get())
static
fch_get()
{
register struct buf *bp;
register char *p, *t;
int n, gofor;
off_t pos, lseek();
/*
* look for a buffer holding the desired block.
*/
if (abs(buf_lastacc->next->block - ch_block) <
abs(buf_lastacc->prev->block - ch_block))
gofor = 1; /* Look forwards through the buffer queue */
else
gofor = 0; /* Look backwards through the buffer queue */
bp = buf_lastacc;
do {
if (bp->block == ch_block) {
buf_lastacc = bp;
if (ch_offset >= bp->datasize)
goto read_more;
return((unsigned char)bp->data[ch_offset]);
}
if (gofor)
bp = bp->next;
else
bp = bp->prev;
} while (bp != buf_lastacc);
/*
* Block is not in a buffer. Take the buffer from the tail and
* read the desired block into it. If the input is a pipe, we try
* to buffer as much input as possible since the input will be
* permanently lost if we throw it from the buffer queue.
*/
if (ispipe && buf_tail->block != (long)(-1))
(void)ch_addbuf(1);
bp = buf_tail;
bp->block = ch_block;
bp->datasize = 0;
read_more:
pos = (ch_block * BUFSIZ) + bp->datasize;
if (ispipe) {
/*
* The data requested should be immediately after
* the last data read from the pipe.
*/
if (pos != last_piped_pos) {
error("pipe error");
quit();
}
} else
(void)lseek(file, pos, L_SET);
/*
* Read the block.
*
* If we read less than a full block, we just return the
* partial block and pick up the rest next time.
*/
n = iread(file, &bp->data[bp->datasize], BUFSIZ - bp->datasize);
if (n == READ_INTR)
return (EOI);
if (n < 0) {
error("read error");
quit();
}
if (ispipe)
last_piped_pos += n;
bp->datasize += n;
if (n == 0) {
ch_fsize = pos;
bp->data[bp->datasize++] = EOI;
}
/*
* Turn other EOI (nul) chars into 0200 since EOI has special meaning.
*/
for (p = &bp->data[bp->datasize]; --n >= 0;) {
--p;
if (*p == EOI)
*p = 0200;
}
found:
if (buf_head != bp) {
/*
* Move the buffer to the head of the buffer chain. This
* ensures correct order for the ispipe case and prevents
* needless buffer thrashing for the !ispipe case. It's not
* clear that buffer thrashing isn't desirable in this latter
* case, since the VM should probably be handling the file
* buffer...
*/
bp->next->prev = bp->prev;
bp->prev->next = bp->next;
bp->next = buf_head;
bp->prev = END_OF_CHAIN;
buf_head->prev = bp;
buf_head = bp;
}
if (ch_offset >= bp->datasize)
/*
* After all that, we still don't have enough data.
* Go back and try again.
*/
goto read_more;
return((unsigned char)bp->data[ch_offset]);
}
/*
* Determine if a specific block is currently in one of the buffers.
*
* In general, this function is only called for the ispipe case. For the
* !ispipe case, ch.c generally assumes that any given block is accessible
* through ch_get(), even though ch_get() may not have it buffered.
*/
static
buffered(block)
long block;
{
register struct buf *bp;
/* For the ispipe case, we know that the buffer queue is sequentially
* ordered from tail to head. */
if (ispipe && (block <= buf_head->block && block >= buf_tail->block))
return(1);
/*
* XXX This is dead code.
*/
for (bp = buf_head; bp != END_OF_CHAIN; bp = bp->next)
if (bp->block == block)
return(1);
return(0);
}
/*
* Seek to a specified position in the file.
* Return 0 if successful, non-zero if can't seek there.
*/
ch_seek(pos)
register off_t pos;
{
long new_block;
new_block = pos / BUFSIZ;
if (!ispipe || pos == last_piped_pos || buffered(new_block)) {
/*
* Set read pointer.
*/
ch_block = new_block;
ch_offset = pos % BUFSIZ;
return(0);
}
return(1);
}
/*
* Seek to the end of the file.
*/
ch_end_seek()
{
off_t ch_length();
if (!ispipe)
return(ch_seek(ch_length()));
/*
* Do it the slow way: read till end of data.
*/
while (ch_forw_get() != EOI)
if (sigs)
return(1);
return(0);
}
/*
* Seek to the beginning of the file, or as close to it as we can get.
* We may not be able to seek there if input is a pipe and the
* beginning of the pipe is no longer buffered.
*/
ch_beg_seek()
{
register struct buf *bp, *firstbp;
/*
* Try a plain ch_seek first.
*/
if (ch_seek((off_t)0) == 0)
return(0);
/*
* Can't get to position 0.
* Look thru the buffers for the one closest to position 0.
*
* This should use the obvious optimization that applies for the
* ispipe case (which is also the only case under which this
* code will be executed, ie. the only case under which ch_seek()
* will fail).
*/
firstbp = bp = buf_head;
if (bp == END_OF_CHAIN)
return(1);
while ((bp = bp->next) != END_OF_CHAIN)
if (bp->block < firstbp->block)
firstbp = bp;
ch_block = firstbp->block;
ch_offset = 0;
return(0);
}
/*
* Return the length of the file, if known.
*/
off_t
ch_length()
{
off_t lseek();
if (ispipe)
return(ch_fsize);
return((off_t)(lseek(file, (off_t)0, L_XTND)));
}
/*
* Return the current position in the file.
*/
off_t
ch_tell()
{
return(ch_block * BUFSIZ + ch_offset);
}
/*
* Get the current char and post-increment the read pointer.
*/
ch_forw_get()
{
register int c;
c = ch_get();
if (c != EOI && ++ch_offset >= BUFSIZ) {
ch_offset = 0;
++ch_block;
}
return(c);
}
/*
* Pre-decrement the read pointer and get the new current char.
*/
ch_back_get()
{
if (--ch_offset < 0) {
if (ch_block <= 0 || (ispipe && !buffered(ch_block-1))) {
ch_offset = 0;
return(EOI);
}
ch_offset = BUFSIZ - 1;
ch_block--;
}
return(ch_get());
}
/*
* Allocate buffers.
* Caller wants us to have a total of at least want_nbufs buffers.
* keep==1 means keep the data in the current buffers;
* otherwise discard the old data.
*/
ch_init(want_nbufs, keep)
int want_nbufs;
int keep;
{
register struct buf *bp;
char message[80];
cbufs = nbufs;
if (nbufs < want_nbufs && ch_addbuf(want_nbufs - nbufs)) {
/*
* Cannot allocate enough buffers.
* If we don't have ANY, then quit.
* Otherwise, just report the error and return.
*/
(void)snprintf(message, sizeof(message),
"cannot allocate %d buffers", want_nbufs - nbufs);
error(message);
if (nbufs == 0)
quit();
return;
}
if (keep)
return;
/*
* We don't want to keep the old data,
* so initialize all the buffers now.
*/
for (bp = buf_head; bp != END_OF_CHAIN; bp = bp->next)
bp->block = (long)(-1);
last_piped_pos = (off_t)0;
ch_fsize = NULL_POSITION;
(void)ch_seek((off_t)0);
}
/*
* Allocate some new buffers.
* The buffers are added to the tail of the buffer chain.
*/
ch_addbuf(nnew)
int nnew;
{
register struct buf *bp;
register struct buf *newbufs;
char *calloc();
/*
* We don't have enough buffers.
* Allocate some new ones.
*/
newbufs = (struct buf *)calloc((u_int)nnew, sizeof(struct buf));
if (newbufs == NULL)
return(1);
/*
* Initialize the new buffers and link them together.
* Link them all onto the tail of the buffer list.
*/
nbufs += nnew;
cbufs = nbufs;
for (bp = &newbufs[0]; bp < &newbufs[nnew]; bp++) {
bp->next = bp + 1;
bp->prev = bp - 1;
bp->block = (long)(-1);
}
newbufs[nnew-1].next = END_OF_CHAIN;
newbufs[0].prev = buf_tail;
buf_tail->next = &newbufs[0];
buf_tail = &newbufs[nnew-1];
return(0);
}