[apple/xnu.git] / bsd / ufs / ufs / ufs_bmap.c

/*
 * Copyright (c) 2000 Apple Computer, Inc. All rights reserved.
 *
 * @APPLE_OSREFERENCE_LICENSE_HEADER_START@
 * 
 * This file contains Original Code and/or Modifications of Original Code
 * as defined in and that are subject to the Apple Public Source License
 * Version 2.0 (the 'License'). You may not use this file except in
 * compliance with the License. The rights granted to you under the License
 * may not be used to create, or enable the creation or redistribution of,
 * unlawful or unlicensed copies of an Apple operating system, or to
 * circumvent, violate, or enable the circumvention or violation of, any
 * terms of an Apple operating system software license agreement.
 * 
 * Please obtain a copy of the License at
 * http://www.opensource.apple.com/apsl/ and read it before using this file.
 * 
 * The Original Code and all software distributed under the License are
 * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
 * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
 * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
 * Please see the License for the specific language governing rights and
 * limitations under the License.
 * 
 * @APPLE_OSREFERENCE_LICENSE_HEADER_END@
 */
/* Copyright (c) 1995 NeXT Computer, Inc. All Rights Reserved */
/*
 * Copyright (c) 1989, 1991, 1993
 *	The Regents of the University of California.  All rights reserved.
 * (c) UNIX System Laboratories, Inc.
 * All or some portions of this file are derived from material licensed
 * to the University of California by American Telephone and Telegraph
 * Co. or Unix System Laboratories, Inc. and are reproduced herein with
 * the permission of UNIX System Laboratories, Inc.
 *
 * 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.
 *
 *	@(#)ufs_bmap.c	8.7 (Berkeley) 3/21/95
 */
/*
 * HISTORY
 * 11-July-97  Umesh Vaishampayan (umeshv@apple.com)
 *	Cleanup. Fixed compilation error when tracing is turned on.
 */
#include <rev_endian_fs.h>
#include <sys/param.h>
#include <sys/buf.h>
#include <sys/proc_internal.h>	/* for p_stats */
#include <sys/vnode_internal.h>
#include <sys/mount_internal.h>
#include <sys/resourcevar.h>
#include <sys/trace.h>
#include <sys/quota.h>

#include <miscfs/specfs/specdev.h>

#include <ufs/ufs/quota.h>
#include <ufs/ufs/inode.h>
#include <ufs/ufs/ufsmount.h>
#include <ufs/ufs/ufs_extern.h>
#if REV_ENDIAN_FS
#include <ufs/ufs/ufs_byte_order.h>
#include <libkern/OSByteOrder.h>
#endif /* REV_ENDIAN_FS */
#include <libkern/OSAtomic.h>


/*
 * Indirect blocks are now on the vnode for the file.  They are given negative
 * logical block numbers.  Indirect blocks are addressed by the negative
 * address of the first data block to which they point.  Double indirect blocks
 * are addressed by one less than the address of the first indirect block to
 * which they point.  Triple indirect blocks are addressed by one less than
 * the address of the first double indirect block to which they point.
 *
 * ufs_bmaparray does the bmap conversion, and if requested returns the
 * array of logical blocks which must be traversed to get to a block.
 * Each entry contains the offset into that block that gets you to the
 * next block and the disk address of the block (if it is assigned).
 */

int
ufs_bmaparray(vp, bn, bnp, ap, nump, runp)
	vnode_t	    vp;
	ufs_daddr_t bn;
	ufs_daddr_t *bnp;
	struct indir *ap;
	int *nump;
	int *runp;
{
	register struct inode *ip;
	struct buf *bp;
	struct ufsmount *ump;
	struct mount *mp;
	struct vnode *devvp;
	struct indir a[NIADDR], *xap;
	ufs_daddr_t daddr;
	long metalbn;
	int error, maxrun, num;
#if REV_ENDIAN_FS
	int rev_endian=0;
#endif /* REV_ENDIAN_FS */

	ip = VTOI(vp);
	mp = vp->v_mount;
	ump = VFSTOUFS(mp);

#if REV_ENDIAN_FS
	rev_endian=(mp->mnt_flag & MNT_REVEND);
#endif /* REV_ENDIAN_FS */

#if DIAGNOSTIC
	if (ap != NULL && nump == NULL || ap == NULL && nump != NULL)
		panic("ufs_bmaparray: invalid arguments");
#endif

	if (runp) {
		/*
		 * XXX
		 * If MAXPHYSIO is the largest transfer the disks can handle,
		 * we probably want maxrun to be 1 block less so that we
		 * don't create a block larger than the device can handle.
		 */
		*runp = 0;
		maxrun = MAXPHYSIO / mp->mnt_vfsstat.f_iosize - 1;
	}

	xap = ap == NULL ? a : ap;
	if (!nump)
		nump = &num;
	if (error = ufs_getlbns(vp, bn, xap, nump))
		return (error);

	num = *nump;
	if (num == 0) {
		*bnp = blkptrtodb(ump, ip->i_db[bn]);
		if (*bnp == 0)
			*bnp = -1;
		else if (runp)
			for (++bn; bn < NDADDR && *runp < maxrun &&
			    is_sequential(ump, ip->i_db[bn - 1], ip->i_db[bn]);
			    ++bn, ++*runp);
		return (0);
	}


	/* Get disk address out of indirect block array */
	daddr = ip->i_ib[xap->in_off];

	devvp = VFSTOUFS(vp->v_mount)->um_devvp;
	for (bp = NULL, ++xap; --num; ++xap) {
	        ufs_daddr_t *dataptr;
		int bop;

	        if ((metalbn = xap->in_lbn) == bn)
		       /*	
			* found the indirect block we were
			* looking for... exit the loop
			*/
		       break;
		
		if (daddr == 0)
		        bop = BLK_ONLYVALID | BLK_META;
		else
		        bop = BLK_META;

		if (bp)
			buf_brelse(bp);
		bp = buf_getblk(vp, (daddr64_t)((unsigned)metalbn), mp->mnt_vfsstat.f_iosize, 0, 0, bop);

		if (bp == 0) {
		        /* 
			 * Exit the loop if there is no disk address assigned yet and
			 * the indirect block isn't in the cache
			 */
			break;
		}
		/*
		 * If we get here, we've either got the block in the cache
		 * or we have a disk address for it, go fetch it.
		 */
		xap->in_exists = 1;

		if (buf_valid(bp)) {
			trace(TR_BREADHIT, pack(vp, mp->mnt_vfsstat.f_iosize), metalbn);
		}
		else {
			trace(TR_BREADMISS, pack(vp, mp->mnt_vfsstat.f_iosize), metalbn);
			buf_setblkno(bp, blkptrtodb(ump, (daddr64_t)((unsigned)daddr)));
			buf_setflags(bp, B_READ);
			VNOP_STRATEGY(bp);
			OSIncrementAtomic(&current_proc()->p_stats->p_ru.ru_inblock);
			if (error = (int)buf_biowait(bp)) {
				buf_brelse(bp);
				return (error);
			}
		}
		dataptr = (ufs_daddr_t *)buf_dataptr(bp);
		daddr = dataptr[xap->in_off];
#if REV_ENDIAN_FS
		if (rev_endian)
			daddr = OSSwapInt32(daddr);
#endif /* REV_ENDIAN_FS */
		if (num == 1 && daddr && runp) {
#if REV_ENDIAN_FS
		if (rev_endian) {
			for (bn = xap->in_off + 1;
			    bn < MNINDIR(ump) && *runp < maxrun &&
			    is_sequential(ump,
			    OSSwapInt32(dataptr[bn - 1]),
			    OSSwapInt32(dataptr[bn]));
			    ++bn, ++*runp);
		 } else {
#endif /* REV_ENDIAN_FS */
			for (bn = xap->in_off + 1;
			    bn < MNINDIR(ump) && *runp < maxrun &&
			    is_sequential(ump,
			    dataptr[bn - 1],
			    dataptr[bn]);
			    ++bn, ++*runp);
#if REV_ENDIAN_FS
		}
#endif /* REV_ENDIAN_FS */
		}
	}
	if (bp)
		buf_brelse(bp);

	daddr = blkptrtodb(ump, daddr);
	*bnp = daddr == 0 ? -1 : daddr;
	return (0);
}

/*
 * Create an array of logical block number/offset pairs which represent the
 * path of indirect blocks required to access a data block.  The first "pair"
 * contains the logical block number of the appropriate single, double or
 * triple indirect block and the offset into the inode indirect block array.
 * Note, the logical block number of the inode single/double/triple indirect
 * block appears twice in the array, once with the offset into the i_ib and
 * once with the offset into the page itself.
 */
int
ufs_getlbns(vp, bn, ap, nump)
	struct vnode *vp;
	ufs_daddr_t bn;
	struct indir *ap;
	int *nump;
{
	long metalbn, realbn;
	struct ufsmount *ump;
	int blockcnt, i, numlevels, off;

	ump = VFSTOUFS(vp->v_mount);
	if (nump)
		*nump = 0;
	numlevels = 0;
	realbn = bn;
	if ((long)bn < 0)
		bn = -(long)bn;

	/* The first NDADDR blocks are direct blocks. */
	if (bn < NDADDR)
		return (0);

	/* 
	 * Determine the number of levels of indirection.  After this loop
	 * is done, blockcnt indicates the number of data blocks possible
	 * at the given level of indirection, and NIADDR - i is the number
	 * of levels of indirection needed to locate the requested block.
	 */
	for (blockcnt = 1, i = NIADDR, bn -= NDADDR;; i--, bn -= blockcnt) {
		if (i == 0)
			return (EFBIG);
		blockcnt *= MNINDIR(ump);
		if (bn < blockcnt)
			break;
	}

	/* Calculate the address of the first meta-block. */
	if (realbn >= 0)
		metalbn = -(realbn - bn + NIADDR - i);
	else
		metalbn = -(-realbn - bn + NIADDR - i);

	/* 
	 * At each iteration, off is the offset into the bap array which is
	 * an array of disk addresses at the current level of indirection.
	 * The logical block number and the offset in that block are stored
	 * into the argument array.
	 */
	ap->in_lbn = metalbn;
	ap->in_off = off = NIADDR - i;
	ap->in_exists = 0;
	ap++;
	for (++numlevels; i <= NIADDR; i++) {
		/* If searching for a meta-data block, quit when found. */
		if (metalbn == realbn)
			break;

		blockcnt /= MNINDIR(ump);
		off = (bn / blockcnt) % MNINDIR(ump);

		++numlevels;
		ap->in_lbn = metalbn;
		ap->in_off = off;
		ap->in_exists = 0;
		++ap;

		metalbn -= -1 + off * blockcnt;
	}
	if (nump)
		*nump = numlevels;
	return (0);
}
/*
 * blockmap converts a file offsetto its physical block
 * number on the disk... it optionally returns the physically
 * contiguous size.
 */
int
ufs_blockmap(ap)
	struct vnop_blockmap_args /* {
		struct vnode *a_vp;
		off_t a_foffset;    
		size_t a_size;
		daddr64_t *a_bpn;
		size_t *a_run;
		void *a_poff;
		int a_flags;
	} */ *ap;
{
	vnode_t	    vp   = ap->a_vp;
	daddr64_t * bnp  = ap->a_bpn;
	size_t    * runp = ap->a_run;
	int	    size = ap->a_size;
	struct fs * fs;
	struct inode *ip;
	ufs_daddr_t lbn;
	ufs_daddr_t daddr = 0;
	int	devBlockSize = 0;
	int	retsize = 0;
	int	error = 0;
	int	nblks;
	int	lblk_offset;

	ip = VTOI(vp);
	fs = ip->i_fs;
	
	devBlockSize = vfs_devblocksize(vnode_mount(vp));

	if (ap->a_foffset % devBlockSize)
		panic("ufs_blockmap; allocation requested inside a device block");

	if (size % devBlockSize)
		panic("ufs_blockmap: size is not multiple of device block size\n");

	/*
	 * round down to the beginning of a filesystem block
	 */
	lbn = (ufs_daddr_t)lblkno(fs, ap->a_foffset);

	lblk_offset = (int)(ap->a_foffset - lblktosize(fs, lbn));

	if ((error = ufs_bmaparray(vp, lbn, &daddr, NULL, NULL, &nblks)))
	        return (error);

	if (ap->a_poff) 
		*(int *)ap->a_poff = 0;

        if (lbn < 0) {
	        /*
		 * we're dealing with the indirect blocks
		 * which are always fs_bsize in size
		 */
	        retsize = (nblks + 1) * fs->fs_bsize;
	} else if (daddr == -1 || nblks == 0) {
	        /*
		 * we're dealing with a 'hole'... UFS doesn't
		 * have a clean way to determine it's size
		 * or
		 * there's are no physically contiguous blocks
		 * so
		 * just return the size of the lbn we started with
		 */
	        retsize = blksize(fs, ip, lbn);
	} else {
	        /*
		 * we have 1 or more blocks that are physically contiguous
		 * to our starting block number... the orignal block + (nblks - 1)
		 * blocks must be full sized since only the last block can be 
		 * composed of fragments...
		 */
		 retsize = nblks * fs->fs_bsize;

		 /*
		  * now compute the size of the last block and add it in
		  */
		 retsize += blksize(fs, ip, (lbn + nblks));
	}
	if (lblk_offset) {
	        if (daddr != -1)
	                daddr += (lblk_offset / devBlockSize);

		if (retsize > lblk_offset)
		        retsize -= lblk_offset;
		else {
		        retsize = 0;
			daddr = -1;
		}
	}
	if (runp) {
		if (retsize < size)
		        *runp = retsize;
		else
		        *runp = size;
	}
	if (bnp)
		*bnp = (daddr64_t)daddr;

	return (0);
}
Commit	Line	Data
1c79356b	1	/*
5d5c5d0d A	2	* Copyright (c) 2000 Apple Computer, Inc. All rights reserved.
5d5c5d0d A	3	*
2d21ac55	4	* @APPLE_OSREFERENCE_LICENSE_HEADER_START@
1c79356b	5	*
2d21ac55 A	6	* This file contains Original Code and/or Modifications of Original Code
	7	* as defined in and that are subject to the Apple Public Source License
	8	* Version 2.0 (the 'License'). You may not use this file except in
	9	* compliance with the License. The rights granted to you under the License
	10	* may not be used to create, or enable the creation or redistribution of,
	11	* unlawful or unlicensed copies of an Apple operating system, or to
	12	* circumvent, violate, or enable the circumvention or violation of, any
	13	* terms of an Apple operating system software license agreement.
8f6c56a5	14	*
2d21ac55 A	15	* Please obtain a copy of the License at
	16	* http://www.opensource.apple.com/apsl/ and read it before using this file.
	17	*
	18	* The Original Code and all software distributed under the License are
	19	* distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
8f6c56a5 A	20	* EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
8f6c56a5 A	21	* INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
2d21ac55 A	22	* FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
	23	* Please see the License for the specific language governing rights and
	24	* limitations under the License.
8f6c56a5	25	*
2d21ac55	26	* @APPLE_OSREFERENCE_LICENSE_HEADER_END@
1c79356b A	27	*/
	28	/* Copyright (c) 1995 NeXT Computer, Inc. All Rights Reserved */
	29	/*
	30	* Copyright (c) 1989, 1991, 1993
	31	* The Regents of the University of California. All rights reserved.
	32	* (c) UNIX System Laboratories, Inc.
	33	* All or some portions of this file are derived from material licensed
	34	* to the University of California by American Telephone and Telegraph
	35	* Co. or Unix System Laboratories, Inc. and are reproduced herein with
	36	* the permission of UNIX System Laboratories, Inc.
	37	*
	38	* Redistribution and use in source and binary forms, with or without
	39	* modification, are permitted provided that the following conditions
	40	* are met:
	41	* 1. Redistributions of source code must retain the above copyright
	42	* notice, this list of conditions and the following disclaimer.
	43	* 2. Redistributions in binary form must reproduce the above copyright
	44	* notice, this list of conditions and the following disclaimer in the
	45	* documentation and/or other materials provided with the distribution.
	46	* 3. All advertising materials mentioning features or use of this software
	47	* must display the following acknowledgement:
	48	* This product includes software developed by the University of
	49	* California, Berkeley and its contributors.
	50	* 4. Neither the name of the University nor the names of its contributors
	51	* may be used to endorse or promote products derived from this software
	52	* without specific prior written permission.
	53	*
	54	* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
	55	* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
	56	* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
	57	* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
	58	* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
	59	* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
	60	* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
	61	* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
	62	* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
	63	* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
	64	* SUCH DAMAGE.
	65	*
	66	* @(#)ufs_bmap.c 8.7 (Berkeley) 3/21/95
	67	*/
	68	/*
	69	* HISTORY
	70	* 11-July-97 Umesh Vaishampayan (umeshv@apple.com)
	71	* Cleanup. Fixed compilation error when tracing is turned on.
	72	*/
	73	#include <rev_endian_fs.h>
	74	#include <sys/param.h>
	75	#include <sys/buf.h>
91447636 A	76	#include <sys/proc_internal.h> /* for p_stats */
	77	#include <sys/vnode_internal.h>
	78	#include <sys/mount_internal.h>
1c79356b A	79	#include <sys/resourcevar.h>
1c79356b A	80	#include <sys/trace.h>
9bccf70c	81	#include <sys/quota.h>
1c79356b A	82
	83	#include <miscfs/specfs/specdev.h>
	84
	85	#include <ufs/ufs/quota.h>
	86	#include <ufs/ufs/inode.h>
	87	#include <ufs/ufs/ufsmount.h>
	88	#include <ufs/ufs/ufs_extern.h>
	89	#if REV_ENDIAN_FS
	90	#include <ufs/ufs/ufs_byte_order.h>
0c530ab8	91	#include <libkern/OSByteOrder.h>
1c79356b	92	#endif /* REV_ENDIAN_FS */
2d21ac55	93	#include <libkern/OSAtomic.h>
1c79356b	94
1c79356b A	95
	96	/*
	97	* Indirect blocks are now on the vnode for the file. They are given negative
	98	* logical block numbers. Indirect blocks are addressed by the negative
	99	* address of the first data block to which they point. Double indirect blocks
	100	* are addressed by one less than the address of the first indirect block to
	101	* which they point. Triple indirect blocks are addressed by one less than
	102	* the address of the first double indirect block to which they point.
	103	*
	104	* ufs_bmaparray does the bmap conversion, and if requested returns the
	105	* array of logical blocks which must be traversed to get to a block.
	106	* Each entry contains the offset into that block that gets you to the
	107	* next block and the disk address of the block (if it is assigned).
	108	*/
	109
	110	int
	111	ufs_bmaparray(vp, bn, bnp, ap, nump, runp)
91447636	112	vnode_t vp;
1c79356b A	113	ufs_daddr_t bn;
	114	ufs_daddr_t *bnp;
	115	struct indir *ap;
	116	int *nump;
	117	int *runp;
	118	{
	119	register struct inode *ip;
	120	struct buf *bp;
	121	struct ufsmount *ump;
	122	struct mount *mp;
	123	struct vnode *devvp;
	124	struct indir a[NIADDR], *xap;
	125	ufs_daddr_t daddr;
	126	long metalbn;
	127	int error, maxrun, num;
	128	#if REV_ENDIAN_FS
	129	int rev_endian=0;
	130	#endif /* REV_ENDIAN_FS */
	131
	132	ip = VTOI(vp);
	133	mp = vp->v_mount;
	134	ump = VFSTOUFS(mp);
	135
	136	#if REV_ENDIAN_FS
	137	rev_endian=(mp->mnt_flag & MNT_REVEND);
	138	#endif /* REV_ENDIAN_FS */
	139
	140	#if DIAGNOSTIC
	141	if (ap != NULL && nump == NULL \|\| ap == NULL && nump != NULL)
	142	panic("ufs_bmaparray: invalid arguments");
	143	#endif
	144
	145	if (runp) {
	146	/*
	147	* XXX
	148	* If MAXPHYSIO is the largest transfer the disks can handle,
	149	* we probably want maxrun to be 1 block less so that we
	150	* don't create a block larger than the device can handle.
	151	*/
	152	*runp = 0;
91447636	153	maxrun = MAXPHYSIO / mp->mnt_vfsstat.f_iosize - 1;
1c79356b A	154	}
	155
	156	xap = ap == NULL ? a : ap;
	157	if (!nump)
	158	nump = &num;
	159	if (error = ufs_getlbns(vp, bn, xap, nump))
	160	return (error);
	161
	162	num = *nump;
	163	if (num == 0) {
	164	*bnp = blkptrtodb(ump, ip->i_db[bn]);
	165	if (*bnp == 0)
	166	*bnp = -1;
	167	else if (runp)
	168	for (++bn; bn < NDADDR && *runp < maxrun &&
	169	is_sequential(ump, ip->i_db[bn - 1], ip->i_db[bn]);
	170	++bn, ++*runp);
	171	return (0);
	172	}
	173
	174
	175	/* Get disk address out of indirect block array */
	176	daddr = ip->i_ib[xap->in_off];
	177
	178	devvp = VFSTOUFS(vp->v_mount)->um_devvp;
	179	for (bp = NULL, ++xap; --num; ++xap) {
91447636 A	180	ufs_daddr_t *dataptr;
	181	int bop;
	182
	183	if ((metalbn = xap->in_lbn) == bn)
	184	/*
	185	* found the indirect block we were
	186	* looking for... exit the loop
	187	*/
	188	break;
	189
	190	if (daddr == 0)
	191	bop = BLK_ONLYVALID \| BLK_META;
	192	else
	193	bop = BLK_META;
1c79356b	194
91447636 A	195	if (bp)
	196	buf_brelse(bp);
	197	bp = buf_getblk(vp, (daddr64_t)((unsigned)metalbn), mp->mnt_vfsstat.f_iosize, 0, 0, bop);
	198
	199	if (bp == 0) {
	200	/*
	201	* Exit the loop if there is no disk address assigned yet and
	202	* the indirect block isn't in the cache
	203	*/
1c79356b	204	break;
91447636	205	}
1c79356b A	206	/*
	207	* If we get here, we've either got the block in the cache
	208	* or we have a disk address for it, go fetch it.
	209	*/
1c79356b	210	xap->in_exists = 1;
91447636 A	211
	212	if (buf_valid(bp)) {
	213	trace(TR_BREADHIT, pack(vp, mp->mnt_vfsstat.f_iosize), metalbn);
1c79356b	214	}
1c79356b	215	else {
91447636 A	216	trace(TR_BREADMISS, pack(vp, mp->mnt_vfsstat.f_iosize), metalbn);
	217	buf_setblkno(bp, blkptrtodb(ump, (daddr64_t)((unsigned)daddr)));
	218	buf_setflags(bp, B_READ);
	219	VNOP_STRATEGY(bp);
2d21ac55	220	OSIncrementAtomic(&current_proc()->p_stats->p_ru.ru_inblock);
91447636 A	221	if (error = (int)buf_biowait(bp)) {
91447636 A	222	buf_brelse(bp);
1c79356b A	223	return (error);
	224	}
	225	}
91447636 A	226	dataptr = (ufs_daddr_t *)buf_dataptr(bp);
91447636 A	227	daddr = dataptr[xap->in_off];
1c79356b A	228	#if REV_ENDIAN_FS
1c79356b A	229	if (rev_endian)
0c530ab8	230	daddr = OSSwapInt32(daddr);
1c79356b A	231	#endif /* REV_ENDIAN_FS */
	232	if (num == 1 && daddr && runp) {
	233	#if REV_ENDIAN_FS
	234	if (rev_endian) {
	235	for (bn = xap->in_off + 1;
	236	bn < MNINDIR(ump) && *runp < maxrun &&
	237	is_sequential(ump,
0c530ab8 A	238	OSSwapInt32(dataptr[bn - 1]),
0c530ab8 A	239	OSSwapInt32(dataptr[bn]));
1c79356b A	240	++bn, ++*runp);
	241	} else {
	242	#endif /* REV_ENDIAN_FS */
	243	for (bn = xap->in_off + 1;
	244	bn < MNINDIR(ump) && *runp < maxrun &&
	245	is_sequential(ump,
91447636 A	246	dataptr[bn - 1],
91447636 A	247	dataptr[bn]);
1c79356b A	248	++bn, ++*runp);
	249	#if REV_ENDIAN_FS
	250	}
	251	#endif /* REV_ENDIAN_FS */
	252	}
	253	}
	254	if (bp)
91447636	255	buf_brelse(bp);
1c79356b A	256
	257	daddr = blkptrtodb(ump, daddr);
	258	*bnp = daddr == 0 ? -1 : daddr;
	259	return (0);
	260	}
	261
	262	/*
	263	* Create an array of logical block number/offset pairs which represent the
	264	* path of indirect blocks required to access a data block. The first "pair"
	265	* contains the logical block number of the appropriate single, double or
	266	* triple indirect block and the offset into the inode indirect block array.
	267	* Note, the logical block number of the inode single/double/triple indirect
	268	* block appears twice in the array, once with the offset into the i_ib and
	269	* once with the offset into the page itself.
	270	*/
	271	int
	272	ufs_getlbns(vp, bn, ap, nump)
	273	struct vnode *vp;
	274	ufs_daddr_t bn;
	275	struct indir *ap;
	276	int *nump;
	277	{
	278	long metalbn, realbn;
	279	struct ufsmount *ump;
	280	int blockcnt, i, numlevels, off;
	281
	282	ump = VFSTOUFS(vp->v_mount);
	283	if (nump)
	284	*nump = 0;
	285	numlevels = 0;
	286	realbn = bn;
	287	if ((long)bn < 0)
	288	bn = -(long)bn;
	289
	290	/* The first NDADDR blocks are direct blocks. */
	291	if (bn < NDADDR)
	292	return (0);
	293
	294	/*
	295	* Determine the number of levels of indirection. After this loop
	296	* is done, blockcnt indicates the number of data blocks possible
	297	* at the given level of indirection, and NIADDR - i is the number
	298	* of levels of indirection needed to locate the requested block.
	299	*/
	300	for (blockcnt = 1, i = NIADDR, bn -= NDADDR;; i--, bn -= blockcnt) {
	301	if (i == 0)
	302	return (EFBIG);
	303	blockcnt *= MNINDIR(ump);
	304	if (bn < blockcnt)
	305	break;
	306	}
	307
	308	/* Calculate the address of the first meta-block. */
	309	if (realbn >= 0)
	310	metalbn = -(realbn - bn + NIADDR - i);
	311	else
	312	metalbn = -(-realbn - bn + NIADDR - i);
	313
	314	/*
	315	* At each iteration, off is the offset into the bap array which is
	316	* an array of disk addresses at the current level of indirection.
	317	* The logical block number and the offset in that block are stored
	318	* into the argument array.
	319	*/
320	ap->in_lbn = metalbn;
321	ap->in_off = off = NIADDR - i;
322	ap->in_exists = 0;
323	ap++;
324	for (++numlevels; i <= NIADDR; i++) {
325	/* If searching for a meta-data block, quit when found. */
326	if (metalbn == realbn)
327	break;
328
329	blockcnt /= MNINDIR(ump);
330	off = (bn / blockcnt) % MNINDIR(ump);
331
332	++numlevels;
333	ap->in_lbn = metalbn;
334	ap->in_off = off;
335	ap->in_exists = 0;
336	++ap;
337
338	metalbn -= -1 + off * blockcnt;
339	}
340	if (nump)
341	*nump = numlevels;
342	return (0);
343	}
344	/*
91447636 A	345	* blockmap converts a file offsetto its physical block
	346	* number on the disk... it optionally returns the physically
	347	* contiguous size.
1c79356b A	348	*/
1c79356b A	349	int
91447636 A	350	ufs_blockmap(ap)
91447636 A	351	struct vnop_blockmap_args /* {
1c79356b A	352	struct vnode *a_vp;
	353	off_t a_foffset;
	354	size_t a_size;
91447636	355	daddr64_t *a_bpn;
1c79356b A	356	size_t *a_run;
1c79356b A	357	void *a_poff;
91447636	358	int a_flags;
1c79356b A	359	} / ap;
1c79356b A	360	{
91447636 A	361	vnode_t vp = ap->a_vp;
	362	daddr64_t * bnp = ap->a_bpn;
	363	size_t * runp = ap->a_run;
	364	int size = ap->a_size;
	365	struct fs * fs;
	366	struct inode *ip;
	367	ufs_daddr_t lbn;
1c79356b	368	ufs_daddr_t daddr = 0;
91447636 A	369	int devBlockSize = 0;
	370	int retsize = 0;
	371	int error = 0;
	372	int nblks;
2d21ac55	373	int lblk_offset;
1c79356b A	374
	375	ip = VTOI(vp);
	376	fs = ip->i_fs;
	377
91447636	378	devBlockSize = vfs_devblocksize(vnode_mount(vp));
1c79356b	379
2d21ac55 A	380	if (ap->a_foffset % devBlockSize)
2d21ac55 A	381	panic("ufs_blockmap; allocation requested inside a device block");
1c79356b	382
91447636 A	383	if (size % devBlockSize)
91447636 A	384	panic("ufs_blockmap: size is not multiple of device block size\n");
1c79356b	385
2d21ac55 A	386	/*
	387	* round down to the beginning of a filesystem block
	388	*/
	389	lbn = (ufs_daddr_t)lblkno(fs, ap->a_foffset);
	390
	391	lblk_offset = (int)(ap->a_foffset - lblktosize(fs, lbn));
	392
91447636 A	393	if ((error = ufs_bmaparray(vp, lbn, &daddr, NULL, NULL, &nblks)))
91447636 A	394	return (error);
1c79356b	395
1c79356b A	396	if (ap->a_poff)
	397	(int )ap->a_poff = 0;
	398
2d21ac55 A	399	if (lbn < 0) {
	400	/*
	401	* we're dealing with the indirect blocks
	402	* which are always fs_bsize in size
	403	*/
	404	retsize = (nblks + 1) * fs->fs_bsize;
	405	} else if (daddr == -1 \|\| nblks == 0) {
	406	/*
	407	* we're dealing with a 'hole'... UFS doesn't
	408	* have a clean way to determine it's size
	409	* or
	410	* there's are no physically contiguous blocks
	411	* so
	412	* just return the size of the lbn we started with
	413	*/
	414	retsize = blksize(fs, ip, lbn);
	415	} else {
	416	/*
	417	* we have 1 or more blocks that are physically contiguous
	418	* to our starting block number... the orignal block + (nblks - 1)
	419	* blocks must be full sized since only the last block can be
	420	* composed of fragments...
	421	*/
	422	retsize = nblks * fs->fs_bsize;
	423
	424	/*
	425	* now compute the size of the last block and add it in
	426	*/
	427	retsize += blksize(fs, ip, (lbn + nblks));
	428	}
	429	if (lblk_offset) {
	430	if (daddr != -1)
	431	daddr += (lblk_offset / devBlockSize);
91447636	432
2d21ac55 A	433	if (retsize > lblk_offset)
	434	retsize -= lblk_offset;
	435	else {
	436	retsize = 0;
	437	daddr = -1;
1c79356b	438	}
2d21ac55 A	439	}
2d21ac55 A	440	if (runp) {
91447636 A	441	if (retsize < size)
	442	*runp = retsize;
	443	else
	444	*runp = size;
1c79356b	445	}
2d21ac55 A	446	if (bnp)
	447	*bnp = (daddr64_t)daddr;
	448
1c79356b A	449	return (0);
1c79356b A	450	}