Implement SEEK_DATA and SEEK_HOLE extensions to lseek(2) as found in

OpenSolaris. For more information please refer to:

	http://blogs.sun.com/bonwick/entry/seek_hole_and_seek_data

include/stdio.h

@@ -200,6 +200,12 @@ __END_DECLS
 #ifndef SEEK_END
 #define	SEEK_END	2	/* set file offset to EOF plus offset */
 #endif
+#ifndef SEEK_DATA
+#define	SEEK_DATA	3	/* set file offset to next data past offset */
+#endif
+#ifndef SEEK_HOLE
+#define	SEEK_HOLE	4	/* set file offset to next hole past offset */
+#endif
 
 #define	stdin	__stdinp
 #define	stdout	__stdoutp

lib/libc/sys/lseek.2

@@ -87,6 +87,23 @@ the offset is set to the size of the
 file plus
 .Fa offset
 bytes.
+.It
+If
+.Fa whence
+is
+.Dv SEEK_HOLE ,
+the offset of the start of the next hole greater than or equal to the supplied
+.Fa offset
+is returned.
+The definition of a hole is provided below.
+.It
+If
+.Fa whence
+is
+.Dv SEEK_DATA ,
+the offset is set to the start of the next non-hole file region greater
+than or equal to the supplied
+.Fa offset .
 .El
 .Pp
 The
@@ -100,6 +117,33 @@ bytes of zeros (until data is actually written into the gap).
 Some devices are incapable of seeking.
 The value of the pointer
 associated with such a device is undefined.
+.Pp
+A
+.Qq hole
+is defined as a contiguous range of bytes in a file, all having the value of
+zero, but not all zeros in a file are guaranteed to be represented as holes
+returned with
+.Dv SEEK_HOLE .
+File systems are allowed to expose ranges of zeros with
+.Dv SEEK_HOLE ,
+but not required to.
+Applications can use
+.Dv SEEK_HOLE
+to optimise their behavior for ranges of zeros, but must not depend on it to
+find all such ranges in a file.
+The existence of a hole at the end of every data region allows for easy
+programming and implies that a virtual hole exists at the end of the file.
+Applications should use
+.Fn fpathconf _PC_MIN_HOLE_SIZE
+or
+.Fn pathconf _PC_MIN_HOLE_SIZE
+to determine if a file system supports
+.Dv SEEK_HOLE .
+See
+.Xr pathconf 2 .
+.Pp
+For file systems that do not supply information about holes, the file will be
+represented as one entire data region.
 .Sh RETURN VALUES
 Upon successful completion,
 .Fn lseek
@@ -128,6 +172,13 @@ argument
 is not a proper value
 or the resulting file offset would
 be negative for a non-character special file.
+.It Bq Er ENXIO
+For
+.Dv SEEK_DATA ,
+there are no more data regions past the supplied offset.
+For
+.Dv SEEK_HOLE ,
+there are no more holes past the supplied offset.
 .It Bq Er EOVERFLOW
 The resulting file offset would be a value which cannot be represented
 correctly in an object of type
@@ -140,7 +191,8 @@ is associated with a pipe, socket, or FIFO.
 .El
 .Sh SEE ALSO
 .Xr dup 2 ,
-.Xr open 2
+.Xr open 2 ,
+.Xr pathconf 2
 .Sh STANDARDS
 The
 .Fn lseek

lib/libc/sys/pathconf.2

@@ -144,6 +144,17 @@ otherwise 0.
 .It Li _PC_MAC_PRESENT
 Returns 1 if a Mandatory Access Control (MAC) label can be set on the
 specified file, otherwise 0.
+.It Li _PC_MIN_HOLE_SIZE
+If a file system supports the reporting of holes (see
+.Xr lseek 2 ,
+.Fn pathconf
+and
+.Fn fpathconf
+return a positive number that represents the minimum hole size returned in
+bytes.
+The offsets of holes returned will be aligned to this same value.
+A special value of 1 is returned if the file system does not specify the minimum
+hole size but still reports holes.  
 .El
 .Sh RETURN VALUES
 If the call to
@@ -216,6 +227,7 @@ is not a valid open file descriptor.
 An I/O error occurred while reading from or writing to the file system.
 .El
 .Sh SEE ALSO
+.Xr lseek 2 ,
 .Xr sysctl 3
 .Sh HISTORY
 The

sys/kern/vfs_syscalls.c

@@ -54,6 +54,7 @@ __FBSDID("$FreeBSD$");
 #include <sys/kernel.h>
 #include <sys/fcntl.h>
 #include <sys/file.h>
+#include <sys/filio.h>
 #include <sys/limits.h>
 #include <sys/linker.h>
 #include <sys/stat.h>
@@ -1740,6 +1741,12 @@ lseek(td, uap)
 		break;
 	case L_SET:
 		break;
+	case SEEK_DATA:
+		error = fo_ioctl(fp, FIOSEEKDATA, &offset, cred, td);
+		break;
+	case SEEK_HOLE:
+		error = fo_ioctl(fp, FIOSEEKHOLE, &offset, cred, td);
+		break;
 	default:
 		error = EINVAL;
 	}

sys/sys/filio.h

@@ -55,5 +55,8 @@ struct fiodgname_arg {
 	void	*buf;
 };
 #define	FIODGNAME	_IOW('f', 120, struct fiodgname_arg) /* get dev. name */
+/* Handle lseek SEEK_DATA and SEEK_HOLE for holey file knowledge. */
+#define	FIOSEEKDATA	_IOWR('f', 97, off_t)	/* SEEK_DATA */
+#define	FIOSEEKHOLE	_IOWR('f', 98, off_t)	/* SEEK_HOLE */
 
 #endif /* !_SYS_FILIO_H_ */

sys/sys/unistd.h

@@ -107,6 +107,8 @@
 #define	SEEK_SET	0	/* set file offset to offset */
 #define	SEEK_CUR	1	/* set file offset to current plus offset */
 #define	SEEK_END	2	/* set file offset to EOF plus offset */
+#define	SEEK_DATA	3	/* set file offset to next data past offset */
+#define	SEEK_HOLE	4	/* set file offset to next hole past offset */
 #endif
 
 #ifndef _POSIX_SOURCE
@@ -151,6 +153,9 @@
 #define	_PC_MAC_PRESENT		63
 #endif
 
+/* From OpenSolaris, used by SEEK_DATA/SEEK_HOLE. */
+#define	_PC_MIN_HOLE_SIZE	21
+
 #if __BSD_VISIBLE
 /*
  * rfork() options.