xfs_ialloc.c

linux 内核源代码

			"xfs_difree: agno >= mp->m_sb.sb_agcount (%d >= %d) on %s.  Returning EINVAL.",
			agno, mp->m_sb.sb_agcount, mp->m_fsname);
		ASSERT(0);
		return XFS_ERROR(EINVAL);
	}
	agino = XFS_INO_TO_AGINO(mp, inode);
	if (inode != XFS_AGINO_TO_INO(mp, agno, agino))  {
		cmn_err(CE_WARN,
			"xfs_difree: inode != XFS_AGINO_TO_INO() "
			"(%llu != %llu) on %s.  Returning EINVAL.",
			(unsigned long long)inode,
			(unsigned long long)XFS_AGINO_TO_INO(mp, agno, agino),
			mp->m_fsname);
		ASSERT(0);
		return XFS_ERROR(EINVAL);
	}
	agbno = XFS_AGINO_TO_AGBNO(mp, agino);
	if (agbno >= mp->m_sb.sb_agblocks)  {
		cmn_err(CE_WARN,
			"xfs_difree: agbno >= mp->m_sb.sb_agblocks (%d >= %d) on %s.  Returning EINVAL.",
			agbno, mp->m_sb.sb_agblocks, mp->m_fsname);
		ASSERT(0);
		return XFS_ERROR(EINVAL);
	}
	/*
	 * Get the allocation group header.
	 */
	down_read(&mp->m_peraglock);
	error = xfs_ialloc_read_agi(mp, tp, agno, &agbp);
	up_read(&mp->m_peraglock);
	if (error) {
		cmn_err(CE_WARN,
			"xfs_difree: xfs_ialloc_read_agi() returned an error %d on %s.  Returning error.",
			error, mp->m_fsname);
		return error;
	}
	agi = XFS_BUF_TO_AGI(agbp);
	ASSERT(be32_to_cpu(agi->agi_magicnum) == XFS_AGI_MAGIC);
	ASSERT(agbno < be32_to_cpu(agi->agi_length));
	/*
	 * Initialize the cursor.
	 */
	cur = xfs_btree_init_cursor(mp, tp, agbp, agno, XFS_BTNUM_INO,
		(xfs_inode_t *)0, 0);
#ifdef DEBUG
	if (cur->bc_nlevels == 1) {
		int freecount = 0;

		if ((error = xfs_inobt_lookup_ge(cur, 0, 0, 0, &i)))
			goto error0;
		do {
			if ((error = xfs_inobt_get_rec(cur, &rec.ir_startino,
					&rec.ir_freecount, &rec.ir_free, &i)))
				goto error0;
			if (i) {
				freecount += rec.ir_freecount;
				if ((error = xfs_inobt_increment(cur, 0, &i)))
					goto error0;
			}
		} while (i == 1);
		ASSERT(freecount == be32_to_cpu(agi->agi_freecount) ||
		       XFS_FORCED_SHUTDOWN(mp));
	}
#endif
	/*
	 * Look for the entry describing this inode.
	 */
	if ((error = xfs_inobt_lookup_le(cur, agino, 0, 0, &i))) {
		cmn_err(CE_WARN,
			"xfs_difree: xfs_inobt_lookup_le returned()  an error %d on %s.  Returning error.",
			error, mp->m_fsname);
		goto error0;
	}
	XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
	if ((error = xfs_inobt_get_rec(cur, &rec.ir_startino, &rec.ir_freecount,
			&rec.ir_free, &i))) {
		cmn_err(CE_WARN,
			"xfs_difree: xfs_inobt_get_rec()  returned an error %d on %s.  Returning error.",
			error, mp->m_fsname);
		goto error0;
	}
	XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
	/*
	 * Get the offset in the inode chunk.
	 */
	off = agino - rec.ir_startino;
	ASSERT(off >= 0 && off < XFS_INODES_PER_CHUNK);
	ASSERT(!XFS_INOBT_IS_FREE(&rec, off));
	/*
	 * Mark the inode free & increment the count.
	 */
	XFS_INOBT_SET_FREE(&rec, off);
	rec.ir_freecount++;

	/*
	 * When an inode cluster is free, it becomes eligible for removal
	 */
	if ((mp->m_flags & XFS_MOUNT_IDELETE) &&
	    (rec.ir_freecount == XFS_IALLOC_INODES(mp))) {

		*delete = 1;
		*first_ino = XFS_AGINO_TO_INO(mp, agno, rec.ir_startino);

		/*
		 * Remove the inode cluster from the AGI B+Tree, adjust the
		 * AGI and Superblock inode counts, and mark the disk space
		 * to be freed when the transaction is committed.
		 */
		ilen = XFS_IALLOC_INODES(mp);
		be32_add(&agi->agi_count, -ilen);
		be32_add(&agi->agi_freecount, -(ilen - 1));
		xfs_ialloc_log_agi(tp, agbp, XFS_AGI_COUNT | XFS_AGI_FREECOUNT);
		down_read(&mp->m_peraglock);
		mp->m_perag[agno].pagi_freecount -= ilen - 1;
		up_read(&mp->m_peraglock);
		xfs_trans_mod_sb(tp, XFS_TRANS_SB_ICOUNT, -ilen);
		xfs_trans_mod_sb(tp, XFS_TRANS_SB_IFREE, -(ilen - 1));

		if ((error = xfs_inobt_delete(cur, &i))) {
			cmn_err(CE_WARN, "xfs_difree: xfs_inobt_delete returned an error %d on %s.\n",
				error, mp->m_fsname);
			goto error0;
		}

		xfs_bmap_add_free(XFS_AGB_TO_FSB(mp,
				agno, XFS_INO_TO_AGBNO(mp,rec.ir_startino)),
				XFS_IALLOC_BLOCKS(mp), flist, mp);
	} else {
		*delete = 0;

		if ((error = xfs_inobt_update(cur, rec.ir_startino, rec.ir_freecount, rec.ir_free))) {
			cmn_err(CE_WARN,
				"xfs_difree: xfs_inobt_update()  returned an error %d on %s.  Returning error.",
				error, mp->m_fsname);
			goto error0;
		}
		/* 
		 * Change the inode free counts and log the ag/sb changes.
		 */
		be32_add(&agi->agi_freecount, 1);
		xfs_ialloc_log_agi(tp, agbp, XFS_AGI_FREECOUNT);
		down_read(&mp->m_peraglock);
		mp->m_perag[agno].pagi_freecount++;
		up_read(&mp->m_peraglock);
		xfs_trans_mod_sb(tp, XFS_TRANS_SB_IFREE, 1);
	}

#ifdef DEBUG
	if (cur->bc_nlevels == 1) {
		int freecount = 0;

		if ((error = xfs_inobt_lookup_ge(cur, 0, 0, 0, &i)))
			goto error0;
		do {
			if ((error = xfs_inobt_get_rec(cur,
					&rec.ir_startino,
					&rec.ir_freecount,
					&rec.ir_free, &i)))
				goto error0;
			if (i) {
				freecount += rec.ir_freecount;
				if ((error = xfs_inobt_increment(cur, 0, &i)))
					goto error0;
			}
		} while (i == 1);
		ASSERT(freecount == be32_to_cpu(agi->agi_freecount) ||
		       XFS_FORCED_SHUTDOWN(mp));
	}
#endif
	xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
	return 0;

error0:
	xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
	return error;
}

/*
 * Return the location of the inode in bno/off, for mapping it into a buffer.
 */
/*ARGSUSED*/
int
xfs_dilocate(
	xfs_mount_t	*mp,	/* file system mount structure */
	xfs_trans_t	*tp,	/* transaction pointer */
	xfs_ino_t	ino,	/* inode to locate */
	xfs_fsblock_t	*bno,	/* output: block containing inode */
	int		*len,	/* output: num blocks in inode cluster */
	int		*off,	/* output: index in block of inode */
	uint		flags)	/* flags concerning inode lookup */
{
	xfs_agblock_t	agbno;	/* block number of inode in the alloc group */
	xfs_buf_t	*agbp;	/* agi buffer */
	xfs_agino_t	agino;	/* inode number within alloc group */
	xfs_agnumber_t	agno;	/* allocation group number */
	int		blks_per_cluster; /* num blocks per inode cluster */
	xfs_agblock_t	chunk_agbno;	/* first block in inode chunk */
	xfs_agino_t	chunk_agino;	/* first agino in inode chunk */
	__int32_t	chunk_cnt;	/* count of free inodes in chunk */
	xfs_inofree_t	chunk_free;	/* mask of free inodes in chunk */
	xfs_agblock_t	cluster_agbno;	/* first block in inode cluster */
	xfs_btree_cur_t	*cur;	/* inode btree cursor */
	int		error;	/* error code */
	int		i;	/* temp state */
	int		offset;	/* index of inode in its buffer */
	int		offset_agbno;	/* blks from chunk start to inode */

	ASSERT(ino != NULLFSINO);
	/*
	 * Split up the inode number into its parts.
	 */
	agno = XFS_INO_TO_AGNO(mp, ino);
	agino = XFS_INO_TO_AGINO(mp, ino);
	agbno = XFS_AGINO_TO_AGBNO(mp, agino);
	if (agno >= mp->m_sb.sb_agcount || agbno >= mp->m_sb.sb_agblocks ||
	    ino != XFS_AGINO_TO_INO(mp, agno, agino)) {
#ifdef DEBUG
		/* no diagnostics for bulkstat, ino comes from userspace */
		if (flags & XFS_IMAP_BULKSTAT)
			return XFS_ERROR(EINVAL);
		if (agno >= mp->m_sb.sb_agcount) {
			xfs_fs_cmn_err(CE_ALERT, mp,
					"xfs_dilocate: agno (%d) >= "
					"mp->m_sb.sb_agcount (%d)",
					agno,  mp->m_sb.sb_agcount);
		}
		if (agbno >= mp->m_sb.sb_agblocks) {
			xfs_fs_cmn_err(CE_ALERT, mp,
					"xfs_dilocate: agbno (0x%llx) >= "
					"mp->m_sb.sb_agblocks (0x%lx)",
					(unsigned long long) agbno,
					(unsigned long) mp->m_sb.sb_agblocks);
		}
		if (ino != XFS_AGINO_TO_INO(mp, agno, agino)) {
			xfs_fs_cmn_err(CE_ALERT, mp,
					"xfs_dilocate: ino (0x%llx) != "
					"XFS_AGINO_TO_INO(mp, agno, agino) "
					"(0x%llx)",
					ino, XFS_AGINO_TO_INO(mp, agno, agino));
		}
		xfs_stack_trace();
#endif /* DEBUG */
		return XFS_ERROR(EINVAL);
	}
	if ((mp->m_sb.sb_blocksize >= XFS_INODE_CLUSTER_SIZE(mp)) ||
	    !(flags & XFS_IMAP_LOOKUP)) {
		offset = XFS_INO_TO_OFFSET(mp, ino);
		ASSERT(offset < mp->m_sb.sb_inopblock);
		*bno = XFS_AGB_TO_FSB(mp, agno, agbno);
		*off = offset;
		*len = 1;
		return 0;
	}
	blks_per_cluster = XFS_INODE_CLUSTER_SIZE(mp) >> mp->m_sb.sb_blocklog;
	if (*bno != NULLFSBLOCK) {
		offset = XFS_INO_TO_OFFSET(mp, ino);
		ASSERT(offset < mp->m_sb.sb_inopblock);
		cluster_agbno = XFS_FSB_TO_AGBNO(mp, *bno);
		*off = ((agbno - cluster_agbno) * mp->m_sb.sb_inopblock) +
			offset;
		*len = blks_per_cluster;
		return 0;
	}
	if (mp->m_inoalign_mask) {
		offset_agbno = agbno & mp->m_inoalign_mask;
		chunk_agbno = agbno - offset_agbno;
	} else {
		down_read(&mp->m_peraglock);
		error = xfs_ialloc_read_agi(mp, tp, agno, &agbp);
		up_read(&mp->m_peraglock);
		if (error) {
#ifdef DEBUG
			xfs_fs_cmn_err(CE_ALERT, mp, "xfs_dilocate: "
					"xfs_ialloc_read_agi() returned "
					"error %d, agno %d",
					error, agno);
#endif /* DEBUG */
			return error;
		}
		cur = xfs_btree_init_cursor(mp, tp, agbp, agno, XFS_BTNUM_INO,
			(xfs_inode_t *)0, 0);
		if ((error = xfs_inobt_lookup_le(cur, agino, 0, 0, &i))) {
#ifdef DEBUG
			xfs_fs_cmn_err(CE_ALERT, mp, "xfs_dilocate: "
					"xfs_inobt_lookup_le() failed");
#endif /* DEBUG */
			goto error0;
		}
		if ((error = xfs_inobt_get_rec(cur, &chunk_agino, &chunk_cnt,
				&chunk_free, &i))) {
#ifdef DEBUG
			xfs_fs_cmn_err(CE_ALERT, mp, "xfs_dilocate: "
					"xfs_inobt_get_rec() failed");
#endif /* DEBUG */
			goto error0;
		}
		if (i == 0) {
#ifdef DEBUG
			xfs_fs_cmn_err(CE_ALERT, mp, "xfs_dilocate: "
					"xfs_inobt_get_rec() failed");
#endif /* DEBUG */
			error = XFS_ERROR(EINVAL);
		}
		xfs_trans_brelse(tp, agbp);
		xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
		if (error)
			return error;
		chunk_agbno = XFS_AGINO_TO_AGBNO(mp, chunk_agino);
		offset_agbno = agbno - chunk_agbno;
	}
	ASSERT(agbno >= chunk_agbno);
	cluster_agbno = chunk_agbno +
		((offset_agbno / blks_per_cluster) * blks_per_cluster);
	offset = ((agbno - cluster_agbno) * mp->m_sb.sb_inopblock) +
		XFS_INO_TO_OFFSET(mp, ino);
	*bno = XFS_AGB_TO_FSB(mp, agno, cluster_agbno);
	*off = offset;
	*len = blks_per_cluster;
	return 0;
error0:
	xfs_trans_brelse(tp, agbp);
	xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
	return error;
}

/*
 * Compute and fill in value of m_in_maxlevels.
 */
void
xfs_ialloc_compute_maxlevels(
	xfs_mount_t	*mp)		/* file system mount structure */
{
	int		level;
	uint		maxblocks;
	uint		maxleafents;
	int		minleafrecs;
	int		minnoderecs;

	maxleafents = (1LL << XFS_INO_AGINO_BITS(mp)) >>
		XFS_INODES_PER_CHUNK_LOG;
	minleafrecs = mp->m_alloc_mnr[0];
	minnoderecs = mp->m_alloc_mnr[1];
	maxblocks = (maxleafents + minleafrecs - 1) / minleafrecs;
	for (level = 1; maxblocks > 1; level++)
		maxblocks = (maxblocks + minnoderecs - 1) / minnoderecs;
	mp->m_in_maxlevels = level;
}

/*
 * Log specified fields for the ag hdr (inode section)
 */
void
xfs_ialloc_log_agi(
	xfs_trans_t	*tp,		/* transaction pointer */
	xfs_buf_t	*bp,		/* allocation group header buffer */
	int		fields)		/* bitmask of fields to log */
{
	int			first;		/* first byte number */
	int			last;		/* last byte number */
	static const short	offsets[] = {	/* field starting offsets */
					/* keep in sync with bit definitions */
		offsetof(xfs_agi_t, agi_magicnum),
		offsetof(xfs_agi_t, agi_versionnum),
		offsetof(xfs_agi_t, agi_seqno),
		offsetof(xfs_agi_t, agi_length),
		offsetof(xfs_agi_t, agi_count),
		offsetof(xfs_agi_t, agi_root),
		offsetof(xfs_agi_t, agi_level),
		offsetof(xfs_agi_t, agi_freecount),
		offsetof(xfs_agi_t, agi_newino),
		offsetof(xfs_agi_t, agi_dirino),
		offsetof(xfs_agi_t, agi_unlinked),
		sizeof(xfs_agi_t)
	};
#ifdef DEBUG
	xfs_agi_t		*agi;	/* allocation group header */

	agi = XFS_BUF_TO_AGI(bp);
	ASSERT(be32_to_cpu(agi->agi_magicnum) == XFS_AGI_MAGIC);
#endif
	/*
	 * Compute byte offsets for the first and last fields.
	 */
	xfs_btree_offsets(fields, offsets, XFS_AGI_NUM_BITS, &first, &last);
	/*
	 * Log the allocation group inode header buffer.
	 */
	xfs_trans_log_buf(tp, bp, first, last);
}

/*
 * Read in the allocation group header (inode allocation section)
 */
int
xfs_ialloc_read_agi(
	xfs_mount_t	*mp,		/* file system mount structure */
	xfs_trans_t	*tp,		/* transaction pointer */
	xfs_agnumber_t	agno,		/* allocation group number */
	xfs_buf_t	**bpp)		/* allocation group hdr buf */
{
	xfs_agi_t	*agi;		/* allocation group header */
	int		agi_ok;		/* agi is consistent */
	xfs_buf_t	*bp;		/* allocation group hdr buf */
	xfs_perag_t	*pag;		/* per allocation group data */
	int		error;

	ASSERT(agno != NULLAGNUMBER);
	error = xfs_trans_read_buf(
			mp, tp, mp->m_ddev_targp,
			XFS_AG_DADDR(mp, agno, XFS_AGI_DADDR(mp)),
			XFS_FSS_TO_BB(mp, 1), 0, &bp);
	if (error)
		return error;
	ASSERT(bp && !XFS_BUF_GETERROR(bp));

	/*
	 * Validate the magic number of the agi block.
	 */
	agi = XFS_BUF_TO_AGI(bp);
	agi_ok =
		be32_to_cpu(agi->agi_magicnum) == XFS_AGI_MAGIC &&
		XFS_AGI_GOOD_VERSION(be32_to_cpu(agi->agi_versionnum));
	if (unlikely(XFS_TEST_ERROR(!agi_ok, mp, XFS_ERRTAG_IALLOC_READ_AGI,
			XFS_RANDOM_IALLOC_READ_AGI))) {
		XFS_CORRUPTION_ERROR("xfs_ialloc_read_agi", XFS_ERRLEVEL_LOW,
				     mp, agi);
		xfs_trans_brelse(tp, bp);
		return XFS_ERROR(EFSCORRUPTED);
	}
	pag = &mp->m_perag[agno];
	if (!pag->pagi_init) {
		pag->pagi_freecount = be32_to_cpu(agi->agi_freecount);
		pag->pagi_count = be32_to_cpu(agi->agi_count);
		pag->pagi_init = 1;
	} else {
		/*
		 * It's possible for these to be out of sync if
		 * we are in the middle of a forced shutdown.
		 */
		ASSERT(pag->pagi_freecount == be32_to_cpu(agi->agi_freecount) ||
			XFS_FORCED_SHUTDOWN(mp));
	}

#ifdef DEBUG
	{
		int	i;

		for (i = 0; i < XFS_AGI_UNLINKED_BUCKETS; i++)
			ASSERT(agi->agi_unlinked[i]);
	}
#endif

	XFS_BUF_SET_VTYPE_REF(bp, B_FS_AGI, XFS_AGI_REF);
	*bpp = bp;
	return 0;
}

/*
 * Read in the agi to initialise the per-ag data in the mount structure
 */
int
xfs_ialloc_pagi_init(
	xfs_mount_t	*mp,		/* file system mount structure */
	xfs_trans_t	*tp,		/* transaction pointer */
	xfs_agnumber_t	agno)		/* allocation group number */
{
	xfs_buf_t	*bp = NULL;
	int		error;

	error = xfs_ialloc_read_agi(mp, tp, agno, &bp);
	if (error)
		return error;
	if (bp)
		xfs_trans_brelse(tp, bp);
	return 0;
}