Fix tail to work on files bigger than 2GB.

PR:		14786
Reviewed by:	iedowse

4 files changed, 146 insertions, 65 deletions

diff --git a/usr.bin/tail/extern.h b/usr.bin/tail/extern.h
index 0ec272d..cdf6929 100644
--- a/usr.bin/tail/extern.h
+++ b/usr.bin/tail/extern.h
@@ -31,11 +31,24 @@
  * SUCH DAMAGE.
  *
  *	@(#)extern.h	8.1 (Berkeley) 6/6/93
+ *
+ * $FreeBSD$
  */
 
-#define	WR(p, size) \
+#define	WR(p, size) do { \
 	if (write(STDOUT_FILENO, p, size) != size) \
-		oerr();
+		oerr(); \
+	} while(0)
+
+#define TAILMAPLEN (4<<20)
+
+struct mapinfo {
+	off_t	mapoff;
+	off_t	maxoff;
+	size_t	maplen;
+	char	*start;
+	int	fd;
+};
 
 enum STYLE { NOTSET = 0, FBYTES, FLINES, RBYTES, RLINES, REVERSE };
 
@@ -47,6 +60,8 @@ int lines __P((FILE *, off_t));
 
 void ierr __P((void));
 void oerr __P((void));
+int mapprint __P((struct mapinfo *, off_t, off_t));
+int maparound __P((struct mapinfo *, off_t));
 
 extern int Fflag, fflag, rflag, rval;
 extern char *fname;
diff --git a/usr.bin/tail/forward.c b/usr.bin/tail/forward.c
index ef046cc..ffca06a 100644
--- a/usr.bin/tail/forward.c
+++ b/usr.bin/tail/forward.c
@@ -267,55 +267,46 @@ rlines(fp, off, sbp)
 	long off;
 	struct stat *sbp;
 {
-	off_t size;
-	char *p;
-	char *start;
+	struct mapinfo map;
+	off_t curoff, size;
+	int i;
 
 	if (!(size = sbp->st_size))
 		return;
+	map.start = NULL;
+	map.fd = fileno(fp);
+	map.mapoff = map.maxoff = size;
 
 	/*
-	 * size not passed directly to mmap() below because unclear error
-	 * diagnostic "Invalid argument".
+	 * Last char is special, ignore whether newline or not. Note that
+	 * size == 0 is dealt with above, and size == 1 sets curoff to -1.
 	 */
-	if (size > SIZE_T_MAX || size < 0) {
-		errno = EFBIG;
-		ierr();
-		exit(1);
-	}
-
-	/*
-	 * XXX: FIXME - mmap() not support files over 2GB
-	 * Large file processing require alternative implementation,
-	 * for now print nice error diagnostic at least.
-	 */
-	if (size > SSIZE_MAX) {
-		errno = EFBIG;
-		ierr();
-		exit(1);
+	curoff = size - 2;
+	while (curoff >= 0) {
+		if (curoff < map.mapoff && maparound(&map, curoff) != 0) {
+			ierr();
+			return;
+		}
+		for (i = curoff - map.mapoff; i >= 0; i--)
+			if (map.start[i] == '\n' && --off == 0)
+				break;
+		/* `i' is either the map offset of a '\n', or -1. */
+		curoff = map.mapoff + i;
+		if (i >= 0)
+			break;
 	}
-
-	if ((start = mmap(NULL, (size_t)size,
-	    PROT_READ, MAP_SHARED, fileno(fp), (off_t)0)) == MAP_FAILED) {
+	curoff++;
+	if (mapprint(&map, curoff, size - curoff) != 0) {
 		ierr();
 		exit(1);
 	}
 
-	/* Last char is special, ignore whether newline or not. */
-	for (p = start + size - 1; --size;)
-		if (*--p == '\n' && !--off) {
-			++p;
-			break;
-		}
-
 	/* Set the file pointer to reflect the length displayed. */
-	size = sbp->st_size - size;
-	WR(p, size);
-	if (fseek(fp, 0L, SEEK_END) == -1) {
+	if (fseeko(fp, sbp->st_size, SEEK_SET) == -1) {
 		ierr();
 		return;
 	}
-	if (munmap(start, (size_t)sbp->st_size)) {
+	if (map.start != NULL && munmap(map.start, map.maplen)) {
 		ierr();
 		return;
 	}
diff --git a/usr.bin/tail/misc.c b/usr.bin/tail/misc.c
index c4edfc3..275acb9 100644
--- a/usr.bin/tail/misc.c
+++ b/usr.bin/tail/misc.c
@@ -44,6 +44,7 @@ static const char rcsid[] =
 
 #include <sys/types.h>
 #include <sys/stat.h>
+#include <sys/mman.h>
 #include <errno.h>
 #include <unistd.h>
 #include <stdio.h>
@@ -64,3 +65,57 @@ oerr()
 {
 	err(1, "stdout");
 }
+
+/*
+ * Print `len' bytes from the file associated with `mip', starting at
+ * absolute file offset `startoff'. May move map window.
+ */
+int
+mapprint(mip, startoff, len)
+	struct mapinfo *mip;
+	off_t startoff, len;
+{
+	int n;
+
+	while (len > 0) {
+		if (startoff < mip->mapoff || startoff >= mip->mapoff +
+		    mip->maplen) {
+			if (maparound(mip, startoff) != 0)
+				return (1);
+		}
+		n = (mip->mapoff + mip->maplen) - startoff;
+		if (n > len)
+			n = len;
+		WR(mip->start + (startoff - mip->mapoff), n);
+		startoff += n;
+		len -= n;
+	}
+	return (0);
+}
+
+/*
+ * Move the map window so that it contains the byte at absolute file
+ * offset `offset'. The start of the map window will be TAILMAPLEN
+ * aligned.
+ */
+int
+maparound(mip, offset)
+	struct mapinfo *mip;
+	off_t offset;
+{
+
+	if (mip->start != NULL && munmap(mip->start, mip->maplen) != 0)
+		return (1);
+
+	mip->mapoff = offset & ~((off_t)TAILMAPLEN - 1);
+	mip->maplen = TAILMAPLEN;
+	if (mip->maplen > mip->maxoff - mip->mapoff)
+		mip->maplen = mip->maxoff - mip->mapoff;
+	if (mip->maplen <= 0)
+		abort();
+	if ((mip->start = mmap(NULL, mip->maplen, PROT_READ, MAP_SHARED,
+	     mip->fd, mip->mapoff)) == MAP_FAILED)
+		return (1);
+
+	return (0);
+}
diff --git a/usr.bin/tail/reverse.c b/usr.bin/tail/reverse.c
index bd2d3ee..24468dd 100644
--- a/usr.bin/tail/reverse.c
+++ b/usr.bin/tail/reverse.c
@@ -114,43 +114,63 @@ r_reg(fp, style, off, sbp)
 	long off;
 	struct stat *sbp;
 {
-	off_t size;
-	int llen;
-	char *p;
-	char *start;
+	struct mapinfo map;
+	off_t curoff, size, lineend;
+	int i;
 
 	if (!(size = sbp->st_size))
 		return;
 
-	if (size > SIZE_T_MAX) {
-		errno = EFBIG;
-		ierr();
-		return;
-	}
+	map.start = NULL;
+	map.mapoff = map.maxoff = size;
+	map.fd = fileno(fp);
 
-	if ((start = mmap(NULL, (size_t)size,
-	    PROT_READ, MAP_SHARED, fileno(fp), (off_t)0)) == MAP_FAILED) {
-		ierr();
-		return;
-	}
-	p = start + size - 1;
+	/*
+	 * Last char is special, ignore whether newline or not. Note that
+	 * size == 0 is dealt with above, and size == 1 sets curoff to -1.
+	 */
+	curoff = size - 2;
+	lineend = size;
+	while (curoff >= 0) {
+		if (curoff < map.mapoff || curoff >= map.mapoff + map.maplen) {
+			if (maparound(&map, curoff) != 0) {
+				ierr();
+				return;
+			}
+		}
+		for (i = curoff - map.mapoff; i >= 0; i--) {
+			if (style == RBYTES && --off == 0)
+				break;
+			if (map.start[i] == '\n')
+				break;
+		}
+		/* `i' is either the map offset of a '\n', or -1. */
+		curoff = map.mapoff + i;
+		if (i < 0)
+			continue;
 
-	if (style == RBYTES && off < size)
-		size = off;
+		/* Print the line and update offsets. */
+		if (mapprint(&map, curoff + 1, lineend - curoff - 1) != 0) {
+			ierr();
+			return;
+		}
+		lineend = curoff + 1;
+		curoff--;
 
-	/* Last char is special, ignore whether newline or not. */
-	for (llen = 1; --size; ++llen)
-		if (*--p == '\n') {
-			WR(p + 1, llen);
-			llen = 0;
-			if (style == RLINES && !--off) {
-				++p;
-				break;
-			}
+		if (style == RLINES)
+			off--;
+
+		if (off == 0 && style != REVERSE) {
+			/* Avoid printing anything below. */
+			curoff = 0;
+			break;
 		}
-	if (llen)
-		WR(p, llen);
-	if (munmap(start, (size_t)sbp->st_size))
+	}
+	if (curoff < 0 && mapprint(&map, 0, lineend) != 0) {
+		ierr();
+		return;
+	}
+	if (map.start != NULL && munmap(map.start, map.maplen))
 		ierr();
 }