xref: /src/sys/fs/efs/efs_subr.c

/*	$NetBSD: efs_subr.c,v 1.5.4.3 2007/10/27 11:35:01 yamt Exp $	*/

/*
 * Copyright (c) 2006 Stephen M. Rumble <rumble (at) ephemeral.org>
 *
 * Permission to use, copy, modify, and distribute this software for any
 * purpose with or without fee is hereby granted, provided that the above
 * copyright notice and this permission notice appear in all copies.
 *
 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
 * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
 * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
 */

#include <sys/cdefs.h>
__KERNEL_RCSID(0, "$NetBSD: efs_subr.c,v 1.5.4.3 2007/10/27 11:35:01 yamt Exp $");

#include <sys/param.h>
#include <sys/kauth.h>
#include <sys/lwp.h>
#include <sys/proc.h>
#include <sys/buf.h>
#include <sys/mount.h>
#include <sys/vnode.h>
#include <sys/namei.h>
#include <sys/stat.h>
#include <sys/malloc.h>

#include <miscfs/genfs/genfs_node.h>

#include <fs/efs/efs.h>
#include <fs/efs/efs_sb.h>
#include <fs/efs/efs_dir.h>
#include <fs/efs/efs_genfs.h>
#include <fs/efs/efs_mount.h>
#include <fs/efs/efs_extent.h>
#include <fs/efs/efs_dinode.h>
#include <fs/efs/efs_inode.h>
#include <fs/efs/efs_subr.h>

struct pool efs_inode_pool;

/*
 * Calculate a checksum for the provided superblock in __host byte order__.
 *
 * At some point SGI changed the checksum algorithm slightly, which can be
 * enabled with the 'new' flag.
 *
 * Presumably this change occured on or before 24 Oct 1988 (around IRIX 3.1),
 * so we're pretty unlikely to ever actually see an old checksum. Further, it
 * means that EFS_NEWMAGIC filesystems (IRIX >= 3.3) must match the new
 * checksum whereas EFS_MAGIC filesystems could potentially use either
 * algorithm.
 *
 * See comp.sys.sgi <1991Aug9.050838.16876 (at) odin.corp.sgi.com>
 */
int32_t
efs_sb_checksum(struct efs_sb *esb, int new)
{
	int i;
	int32_t cksum;
	uint16_t *sbarray = (uint16_t *)esb;

	KASSERT((EFS_SB_CHECKSUM_SIZE % 2) == 0);

	for (i = cksum = 0; i < (EFS_SB_CHECKSUM_SIZE / 2); i++) {
		cksum ^= be16toh(sbarray[i]);
		cksum  = (cksum << 1) | (new && cksum < 0);
	}

	return (cksum);
}

/*
 * Determine if the superblock is valid.
 *
 * Returns 0 if valid, else invalid. If invalid, 'why' is set to an
 * explanation.
 */
int
efs_sb_validate(struct efs_sb *esb, const char **why)
{
	uint32_t ocksum, ncksum;

	*why = NULL;

	if (be32toh(esb->sb_magic) != EFS_SB_MAGIC &&
	    be32toh(esb->sb_magic) != EFS_SB_NEWMAGIC) {
		*why = "sb_magic invalid";
		return (1);
	}

	ocksum = htobe32(efs_sb_checksum(esb, 0));
	ncksum = htobe32(efs_sb_checksum(esb, 1));
	if (esb->sb_checksum != ocksum && esb->sb_checksum != ncksum) {
		*why = "sb_checksum invalid";
		return (1);
	}

	if (be32toh(esb->sb_size) > EFS_SIZE_MAX) {
		*why = "sb_size > EFS_SIZE_MAX";
		return (1);
	}

	if (be32toh(esb->sb_firstcg) <= EFS_BB_BITMAP) {
		*why = "sb_firstcg <= EFS_BB_BITMAP";
		return (1);
	}

	/* XXX - add better sb consistency checks here */
	if (esb->sb_cgfsize == 0 ||
	    esb->sb_cgisize == 0 ||
	    esb->sb_ncg == 0 ||
	    esb->sb_bmsize == 0) {
		*why = "something bad happened";
		return (1);
	}

	return (0);
}

/*
 * Determine the basic block offset and inode index within that block, given
 * the inode 'ino' and filesystem parameters _in host byte order_. The inode
 * will live at byte address 'bboff' * EFS_BB_SIZE + 'index' * EFS_DINODE_SIZE.
 */
void
efs_locate_inode(ino_t ino, struct efs_sb *sbp, uint32_t *bboff, int *index)
{
	uint32_t cgfsize, firstcg;
	uint16_t cgisize;

	cgisize = be16toh(sbp->sb_cgisize);
	cgfsize = be32toh(sbp->sb_cgfsize);
	firstcg = be32toh(sbp->sb_firstcg),

	*bboff = firstcg + ((ino / (cgisize * EFS_DINODES_PER_BB)) * cgfsize) +
	    ((ino % (cgisize * EFS_DINODES_PER_BB)) / EFS_DINODES_PER_BB);
	*index = ino & (EFS_DINODES_PER_BB - 1);
}

/*
 * Read in an inode from disk.
 *
 * We actually take in four inodes at a time. Hopefully these will stick
 * around in the buffer cache and get used without going to disk.
 *
 * Returns 0 on success.
 */
int
efs_read_inode(struct efs_mount *emp, ino_t ino, struct lwp *l,
    struct efs_dinode *di)
{
	struct efs_sb *sbp;
	struct buf *bp;
	int index, err;
	uint32_t bboff;

	sbp = &emp->em_sb;
	efs_locate_inode(ino, sbp, &bboff, &index);

	err = efs_bread(emp, bboff, l, &bp);
	if (err) {
		brelse(bp, 0);
		return (err);
	}
	memcpy(di, ((struct efs_dinode *)bp->b_data) + index, sizeof(*di));
	brelse(bp, 0);

	return (0);
}

/*
 * Perform a read from our device handling the potential DEV_BSIZE
 * messiness (although as of 19.2.2006, all ports appear to use 512) as
 * we as EFS block sizing.
 *
 * bboff: basic block offset
 *
 * Returns 0 on success.
 */
int
efs_bread(struct efs_mount *emp, uint32_t bboff, struct lwp *l, struct buf **bp)
{
	KASSERT(bboff < EFS_SIZE_MAX);

	return (bread(emp->em_devvp, (daddr_t)bboff * (EFS_BB_SIZE / DEV_BSIZE),
	    EFS_BB_SIZE, (l == NULL) ? NOCRED : l->l_cred, bp));
}

/*
 * Synchronise the in-core, host ordered and typed inode fields with their
 * corresponding on-disk, EFS ordered and typed copies.
 *
 * This is the inverse of efs_dinode_sync_inode(), and should be called when
 * an inode is loaded from disk.
 */
void
efs_sync_dinode_to_inode(struct efs_inode *ei)
{

	ei->ei_mode		= be16toh(ei->ei_di.di_mode);	/*same as nbsd*/
	ei->ei_nlink		= be16toh(ei->ei_di.di_nlink);
	ei->ei_uid		= be16toh(ei->ei_di.di_uid);
	ei->ei_gid		= be16toh(ei->ei_di.di_gid);
	ei->ei_size		= be32toh(ei->ei_di.di_size);
	ei->ei_atime		= be32toh(ei->ei_di.di_atime);
	ei->ei_mtime		= be32toh(ei->ei_di.di_mtime);
	ei->ei_ctime		= be32toh(ei->ei_di.di_ctime);
	ei->ei_gen		= be32toh(ei->ei_di.di_gen);
	ei->ei_numextents 	= be16toh(ei->ei_di.di_numextents);
	ei->ei_version		= ei->ei_di.di_version;
}

/*
 * Synchronise the on-disk, EFS ordered and typed inode fields with their
 * corresponding in-core, host ordered and typed copies.
 *
 * This is the inverse of efs_inode_sync_dinode(), and should be called before
 * an inode is flushed to disk.
 */
void
efs_sync_inode_to_dinode(struct efs_inode *ei)
{

	panic("readonly -- no need to call me");
}

#ifdef DIAGNOSTIC
/*
 * Ensure that the in-core inode's host cached fields match its on-disk copy.
 *
 * Returns 0 if they match.
 */
static int
efs_is_inode_synced(struct efs_inode *ei)
{
	int s;

	s = 0;
	/* XXX -- see above remarks about assumption */
	s += (ei->ei_mode	!= be16toh(ei->ei_di.di_mode));
	s += (ei->ei_nlink	!= be16toh(ei->ei_di.di_nlink));
	s += (ei->ei_uid	!= be16toh(ei->ei_di.di_uid));
	s += (ei->ei_gid	!= be16toh(ei->ei_di.di_gid));
	s += (ei->ei_size	!= be32toh(ei->ei_di.di_size));
	s += (ei->ei_atime	!= be32toh(ei->ei_di.di_atime));
	s += (ei->ei_mtime	!= be32toh(ei->ei_di.di_mtime));
	s += (ei->ei_ctime	!= be32toh(ei->ei_di.di_ctime));
	s += (ei->ei_gen	!= be32toh(ei->ei_di.di_gen));
	s += (ei->ei_numextents	!= be16toh(ei->ei_di.di_numextents));
	s += (ei->ei_version	!= ei->ei_di.di_version);

	return (s);
}
#endif

/*
 * Given an efs_dirblk structure and a componentname to search for, return the
 * corresponding inode if it is found.
 *
 * Returns 0 on success.
 */
static int
efs_dirblk_lookup(struct efs_dirblk *dir, struct componentname *cn,
    ino_t *inode)
{
	struct efs_dirent *de;
	int i, slot, offset;

	KASSERT(cn->cn_namelen <= EFS_DIRENT_NAMELEN_MAX);

	slot = offset = 0;

	for (i = 0; i < dir->db_slots; i++) {
		offset = EFS_DIRENT_OFF_EXPND(dir->db_space[i]);

		if (offset == EFS_DIRBLK_SLOT_FREE)
			continue;

		de = (struct efs_dirent *)((char *)dir + offset);
		if (de->de_namelen == cn->cn_namelen &&
		   (strncmp(cn->cn_nameptr, de->de_name, cn->cn_namelen) == 0)){
			slot = i;
			break;
		}
	}
	if (i == dir->db_slots)
		return (ENOENT);

	KASSERT(slot < offset && offset < EFS_DIRBLK_SPACE_SIZE);
	de = (struct efs_dirent *)((char *)dir + offset);
	*inode = be32toh(de->de_inumber);

	return (0);
}

/*
 * Given an extent descriptor that represents a directory, look up
 * componentname within its efs_dirblk's. If it is found, return the
 * corresponding inode in 'ino'.
 *
 * Returns 0 on success.
 */
static int
efs_extent_lookup(struct efs_mount *emp, struct efs_extent *ex,
    struct componentname *cn, ino_t *ino)
{
	struct efs_dirblk *db;
	struct buf *bp;
	int i, err;

	/*
	 * Read in each of the dirblks until we find our entry.
	 * If we don't, return ENOENT.
	 */
	for (i = 0; i < ex->ex_length; i++) {
		err = efs_bread(emp, ex->ex_bn + i, NULL, &bp);
		if (err) {
			printf("efs: warning: invalid extent descriptor\n");
			brelse(bp, 0);
			return (err);
		}

		db = (struct efs_dirblk *)bp->b_data;
		if (efs_dirblk_lookup(db, cn, ino) == 0) {
			brelse(bp, 0);
			return (0);
		}
		brelse(bp, 0);
	}

	return (ENOENT);
}

/*
 * Given the provided in-core inode, look up the pathname requested. If
 * we find it, 'ino' reflects its corresponding on-disk inode number.
 *
 * Returns 0 on success.
 */
int
efs_inode_lookup(struct efs_mount *emp, struct efs_inode *ei,
    struct componentname *cn, ino_t *ino)
{
	struct efs_extent ex;
	struct efs_extent_iterator exi;
	int ret;

	KASSERT(VOP_ISLOCKED(ei->ei_vp));
	KASSERT(efs_is_inode_synced(ei) == 0);
	KASSERT((ei->ei_mode & S_IFMT) == S_IFDIR);

	efs_extent_iterator_init(&exi, ei, 0);
	while ((ret = efs_extent_iterator_next(&exi, &ex)) == 0) {
		if (efs_extent_lookup(emp, &ex, cn, ino) == 0) {
			return (0);
		}
	}

	return ((ret == -1) ? ENOENT : ret);
}

/*
 * Convert on-disk extent structure to in-core format.
 */
void
efs_dextent_to_extent(struct efs_dextent *dex, struct efs_extent *ex)
{

	KASSERT(dex != NULL && ex != NULL);

	ex->ex_magic	= dex->ex_bytes[0];
	ex->ex_bn	= be32toh(dex->ex_words[0]) & 0x00ffffff;
	ex->ex_length	= dex->ex_bytes[4];
	ex->ex_offset	= be32toh(dex->ex_words[1]) & 0x00ffffff;
}

/*
 * Convert in-core extent format to on-disk structure.
 */
void
efs_extent_to_dextent(struct efs_extent *ex, struct efs_dextent *dex)
{

	KASSERT(ex != NULL && dex != NULL);
	KASSERT(ex->ex_magic == EFS_EXTENT_MAGIC);
	KASSERT((ex->ex_bn & ~EFS_EXTENT_BN_MASK) == 0);
	KASSERT((ex->ex_offset & ~EFS_EXTENT_OFFSET_MASK) == 0);

	dex->ex_words[0] = htobe32(ex->ex_bn);
	dex->ex_bytes[0] = ex->ex_magic;
	dex->ex_words[1] = htobe32(ex->ex_offset);
	dex->ex_bytes[4] = ex->ex_length;
}

/*
 * Initialise an extent iterator.
 *
 * If start_hint is non-0, attempt to set up the iterator beginning with the
 * extent descriptor in which the start_hint'th byte exists. Callers must not
 * expect success (this is simply an optimisation), so we reserve the right
 * to start from the beginning.
 */
void
efs_extent_iterator_init(struct efs_extent_iterator *exi, struct efs_inode *eip,
    off_t start_hint)
{
	struct efs_extent ex, ex2;
	struct buf *bp;
	struct efs_mount *emp = VFSTOEFS(eip->ei_vp->v_mount);
	off_t offset, length, next;
	int i, err, numextents, numinextents;
	int hi, lo, mid;
	int indir;

	exi->exi_eip	= eip;
	exi->exi_next	= 0;
	exi->exi_dnext	= 0;
	exi->exi_innext	= 0;

	if (start_hint == 0)
		return;

	/* force iterator to end if hint is too big */
	if (start_hint >= eip->ei_size) {
		exi->exi_next = eip->ei_numextents;
		return;
	}

	/*
	 * Use start_hint to jump to the right extent descriptor. We'll
	 * iterate over the 12 indirect extents because it's cheap, then
	 * bring the appropriate vector into core and binary search it.
	 */

	/*
	 * Handle the small file case separately first...
	 */
	if (eip->ei_numextents <= EFS_DIRECTEXTENTS) {
		for (i = 0; i < eip->ei_numextents; i++) {
			efs_dextent_to_extent(&eip->ei_di.di_extents[i], &ex);

			offset = ex.ex_offset * EFS_BB_SIZE;
			length = ex.ex_length * EFS_BB_SIZE;

			if (start_hint >= offset &&
			    start_hint < (offset + length)) {
				exi->exi_next = exi->exi_dnext = i;
				return;
			}
		}

		/* shouldn't get here, no? */
		EFS_DPRINTF(("efs_extent_iterator_init: bad direct extents\n"));
		return;
	}

	/*
	 * Now do the large files with indirect extents...
	 *
	 * The first indirect extent's ex_offset field contains the
	 * number of indirect extents used.
	 */
	efs_dextent_to_extent(&eip->ei_di.di_extents[0], &ex);

	numinextents = ex.ex_offset;
	if (numinextents < 1 || numinextents >= EFS_DIRECTEXTENTS) {
		EFS_DPRINTF(("efs_extent_iterator_init: bad ex.ex_offset\n"));
		return;
	}

	next = 0;
	indir = -1;
	numextents = 0;
	for (i = 0; i < numinextents; i++) {
		efs_dextent_to_extent(&eip->ei_di.di_extents[i], &ex);

		err = efs_bread(emp, ex.ex_bn, NULL, &bp);
		if (err) {
			brelse(bp, 0);
			return;
		}

		efs_dextent_to_extent((struct efs_dextent *)bp->b_data, &ex2);
		brelse(bp, 0);

		offset = ex2.ex_offset * EFS_BB_SIZE;

		if (offset > start_hint) {
			indir = MAX(0, i - 1);
			break;
		}

		/* number of extents prior to this indirect vector of extents */
		next += numextents;

		/* number of extents within this indirect vector of extents */
		numextents = ex.ex_length * EFS_EXTENTS_PER_BB;
		numextents = MIN(numextents, eip->ei_numextents - next);
	}

	/*
	 * We hit the end, so assume it's in the last extent.
	 */
	if (indir == -1)
		indir = numinextents - 1;

	/*
	 * Binary search to find our desired direct extent.
	 */
	lo = 0;
	mid = 0;
	hi = numextents - 1;
	efs_dextent_to_extent(&eip->ei_di.di_extents[indir], &ex);
	while (lo <= hi) {
		int bboff;
		int index;

		mid = (lo + hi) / 2;

		bboff = mid / EFS_EXTENTS_PER_BB;
		index = mid % EFS_EXTENTS_PER_BB;

		err = efs_bread(emp, ex.ex_bn + bboff, NULL, &bp);
		if (err) {
			brelse(bp, 0);
			EFS_DPRINTF(("efs_extent_iterator_init: bsrch read\n"));
			return;
		}

		efs_dextent_to_extent((struct efs_dextent *)bp->b_data + index,
		    &ex2);
		brelse(bp, 0);

		offset = ex2.ex_offset * EFS_BB_SIZE;
		length = ex2.ex_length * EFS_BB_SIZE;

		if (start_hint >= offset && start_hint < (offset + length))
			break;

		if (start_hint < offset)
			hi = mid - 1;
		else
			lo = mid + 1;
	}

	/*
	 * This is bad. Either the hint is bogus (which shouldn't
	 * happen) or the extent list must be screwed up. We
	 * have to abort.
	 */
	if (lo > hi) {
		EFS_DPRINTF(("efs_extent_iterator_init: bsearch "
		    "failed to find extent\n"));
		return;
	}

	exi->exi_next	= next + mid;
	exi->exi_dnext	= indir;
	exi->exi_innext	= mid;
}

/*
 * Return the next EFS extent.
 *
 * Returns 0 if another extent was iterated, -1 if we've exhausted all
 * extents, or an error number. If 'exi' is non-NULL, the next extent is
 * written to it (should it exist).
 */
int
efs_extent_iterator_next(struct efs_extent_iterator *exi,
    struct efs_extent *exp)
{
	struct efs_extent ex;
	struct efs_dextent *dexp;
	struct efs_inode *eip = exi->exi_eip;
	struct buf *bp;
	int err, bboff, index;

	if (exi->exi_next++ >= eip->ei_numextents)
		return (-1);

	/* direct or indirect extents? */
	if (eip->ei_numextents <= EFS_DIRECTEXTENTS) {
		if (exp != NULL) {
			dexp = &eip->ei_di.di_extents[exi->exi_dnext++];
			efs_dextent_to_extent(dexp, exp);
		}
	} else {
		efs_dextent_to_extent(
		    &eip->ei_di.di_extents[exi->exi_dnext], &ex);

		bboff	= exi->exi_innext / EFS_EXTENTS_PER_BB;
		index	= exi->exi_innext % EFS_EXTENTS_PER_BB;

		err = efs_bread(VFSTOEFS(eip->ei_vp->v_mount),
		    ex.ex_bn + bboff, NULL, &bp);
		if (err) {
			EFS_DPRINTF(("efs_extent_iterator_next: "
			    "efs_bread failed: %d\n", err));
			brelse(bp, 0);
			return (err);
		}

		if (exp != NULL) {
			dexp = (struct efs_dextent *)bp->b_data + index;
			efs_dextent_to_extent(dexp, exp);
		}
		brelse(bp, 0);

		bboff = exi->exi_innext++ / EFS_EXTENTS_PER_BB;
		if (bboff >= ex.ex_length) {
			exi->exi_innext = 0;
			exi->exi_dnext++;
		}
	}

	return (0);
}