xfs_ialloc.c

linux 内核源代码

/*
 * Allocate an inode on disk.
 * Mode is used to tell whether the new inode will need space, and whether
 * it is a directory.
 *
 * The arguments IO_agbp and alloc_done are defined to work within
 * the constraint of one allocation per transaction.
 * xfs_dialloc() is designed to be called twice if it has to do an
 * allocation to make more free inodes.  On the first call,
 * IO_agbp should be set to NULL. If an inode is available,
 * i.e., xfs_dialloc() did not need to do an allocation, an inode
 * number is returned.  In this case, IO_agbp would be set to the
 * current ag_buf and alloc_done set to false.
 * If an allocation needed to be done, xfs_dialloc would return
 * the current ag_buf in IO_agbp and set alloc_done to true.
 * The caller should then commit the current transaction, allocate a new
 * transaction, and call xfs_dialloc() again, passing in the previous
 * value of IO_agbp.  IO_agbp should be held across the transactions.
 * Since the agbp is locked across the two calls, the second call is
 * guaranteed to have a free inode available.
 *
 * Once we successfully pick an inode its number is returned and the
 * on-disk data structures are updated.  The inode itself is not read
 * in, since doing so would break ordering constraints with xfs_reclaim.
 */
int
xfs_dialloc(
	xfs_trans_t	*tp,		/* transaction pointer */
	xfs_ino_t	parent,		/* parent inode (directory) */
	mode_t		mode,		/* mode bits for new inode */
	int		okalloc,	/* ok to allocate more space */
	xfs_buf_t	**IO_agbp,	/* in/out ag header's buffer */
	boolean_t	*alloc_done,	/* true if we needed to replenish
					   inode freelist */
	xfs_ino_t	*inop)		/* inode number allocated */
{
	xfs_agnumber_t	agcount;	/* number of allocation groups */
	xfs_buf_t	*agbp;		/* allocation group header's buffer */
	xfs_agnumber_t	agno;		/* allocation group number */
	xfs_agi_t	*agi;		/* allocation group header structure */
	xfs_btree_cur_t	*cur;		/* inode allocation btree cursor */
	int		error;		/* error return value */
	int		i;		/* result code */
	int		ialloced;	/* inode allocation status */
	int		noroom = 0;	/* no space for inode blk allocation */
	xfs_ino_t	ino;		/* fs-relative inode to be returned */
	/* REFERENCED */
	int		j;		/* result code */
	xfs_mount_t	*mp;		/* file system mount structure */
	int		offset;		/* index of inode in chunk */
	xfs_agino_t	pagino;		/* parent's a.g. relative inode # */
	xfs_agnumber_t	pagno;		/* parent's allocation group number */
	xfs_inobt_rec_incore_t rec;	/* inode allocation record */
	xfs_agnumber_t	tagno;		/* testing allocation group number */
	xfs_btree_cur_t	*tcur;		/* temp cursor */
	xfs_inobt_rec_incore_t trec;	/* temp inode allocation record */


	if (*IO_agbp == NULL) {
		/*
		 * We do not have an agbp, so select an initial allocation
		 * group for inode allocation.
		 */
		agbp = xfs_ialloc_ag_select(tp, parent, mode, okalloc);
		/*
		 * Couldn't find an allocation group satisfying the
		 * criteria, give up.
		 */
		if (!agbp) {
			*inop = NULLFSINO;
			return 0;
		}
		agi = XFS_BUF_TO_AGI(agbp);
		ASSERT(be32_to_cpu(agi->agi_magicnum) == XFS_AGI_MAGIC);
	} else {
		/*
		 * Continue where we left off before.  In this case, we
		 * know that the allocation group has free inodes.
		 */
		agbp = *IO_agbp;
		agi = XFS_BUF_TO_AGI(agbp);
		ASSERT(be32_to_cpu(agi->agi_magicnum) == XFS_AGI_MAGIC);
		ASSERT(be32_to_cpu(agi->agi_freecount) > 0);
	}
	mp = tp->t_mountp;
	agcount = mp->m_sb.sb_agcount;
	agno = be32_to_cpu(agi->agi_seqno);
	tagno = agno;
	pagno = XFS_INO_TO_AGNO(mp, parent);
	pagino = XFS_INO_TO_AGINO(mp, parent);

	/*
	 * If we have already hit the ceiling of inode blocks then clear
	 * okalloc so we scan all available agi structures for a free
	 * inode.
	 */

	if (mp->m_maxicount &&
	    mp->m_sb.sb_icount + XFS_IALLOC_INODES(mp) > mp->m_maxicount) {
		noroom = 1;
		okalloc = 0;
	}

	/*
	 * Loop until we find an allocation group that either has free inodes
	 * or in which we can allocate some inodes.  Iterate through the
	 * allocation groups upward, wrapping at the end.
	 */
	*alloc_done = B_FALSE;
	while (!agi->agi_freecount) {
		/*
		 * Don't do anything if we're not supposed to allocate
		 * any blocks, just go on to the next ag.
		 */
		if (okalloc) {
			/*
			 * Try to allocate some new inodes in the allocation
			 * group.
			 */
			if ((error = xfs_ialloc_ag_alloc(tp, agbp, &ialloced))) {
				xfs_trans_brelse(tp, agbp);
				if (error == ENOSPC) {
					*inop = NULLFSINO;
					return 0;
				} else
					return error;
			}
			if (ialloced) {
				/*
				 * We successfully allocated some inodes, return
				 * the current context to the caller so that it
				 * can commit the current transaction and call
				 * us again where we left off.
				 */
				ASSERT(be32_to_cpu(agi->agi_freecount) > 0);
				*alloc_done = B_TRUE;
				*IO_agbp = agbp;
				*inop = NULLFSINO;
				return 0;
			}
		}
		/*
		 * If it failed, give up on this ag.
		 */
		xfs_trans_brelse(tp, agbp);
		/*
		 * Go on to the next ag: get its ag header.
		 */
nextag:
		if (++tagno == agcount)
			tagno = 0;
		if (tagno == agno) {
			*inop = NULLFSINO;
			return noroom ? ENOSPC : 0;
		}
		down_read(&mp->m_peraglock);
		if (mp->m_perag[tagno].pagi_inodeok == 0) {
			up_read(&mp->m_peraglock);
			goto nextag;
		}
		error = xfs_ialloc_read_agi(mp, tp, tagno, &agbp);
		up_read(&mp->m_peraglock);
		if (error)
			goto nextag;
		agi = XFS_BUF_TO_AGI(agbp);
		ASSERT(be32_to_cpu(agi->agi_magicnum) == XFS_AGI_MAGIC);
	}
	/*
	 * Here with an allocation group that has a free inode.
	 * Reset agno since we may have chosen a new ag in the
	 * loop above.
	 */
	agno = tagno;
	*IO_agbp = NULL;
	cur = xfs_btree_init_cursor(mp, tp, agbp, be32_to_cpu(agi->agi_seqno),
				    XFS_BTNUM_INO, (xfs_inode_t *)0, 0);
	/*
	 * If pagino is 0 (this is the root inode allocation) use newino.
	 * This must work because we've just allocated some.
	 */
	if (!pagino)
		pagino = be32_to_cpu(agi->agi_newino);
#ifdef DEBUG
	if (cur->bc_nlevels == 1) {
		int	freecount = 0;

		if ((error = xfs_inobt_lookup_ge(cur, 0, 0, 0, &i)))
			goto error0;
		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
		do {
			if ((error = xfs_inobt_get_rec(cur, &rec.ir_startino,
					&rec.ir_freecount, &rec.ir_free, &i)))
				goto error0;
			XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
			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
	/*
	 * If in the same a.g. as the parent, try to get near the parent.
	 */
	if (pagno == agno) {
		if ((error = xfs_inobt_lookup_le(cur, pagino, 0, 0, &i)))
			goto error0;
		if (i != 0 &&
		    (error = xfs_inobt_get_rec(cur, &rec.ir_startino,
			    &rec.ir_freecount, &rec.ir_free, &j)) == 0 &&
		    j == 1 &&
		    rec.ir_freecount > 0) {
			/*
			 * Found a free inode in the same chunk
			 * as parent, done.
			 */
		}
		/*
		 * In the same a.g. as parent, but parent's chunk is full.
		 */
		else {
			int	doneleft;	/* done, to the left */
			int	doneright;	/* done, to the right */

			if (error)
				goto error0;
			ASSERT(i == 1);
			ASSERT(j == 1);
			/*
			 * Duplicate the cursor, search left & right
			 * simultaneously.
			 */
			if ((error = xfs_btree_dup_cursor(cur, &tcur)))
				goto error0;
			/*
			 * Search left with tcur, back up 1 record.
			 */
			if ((error = xfs_inobt_decrement(tcur, 0, &i)))
				goto error1;
			doneleft = !i;
			if (!doneleft) {
				if ((error = xfs_inobt_get_rec(tcur,
						&trec.ir_startino,
						&trec.ir_freecount,
						&trec.ir_free, &i)))
					goto error1;
				XFS_WANT_CORRUPTED_GOTO(i == 1, error1);
			}
			/*
			 * Search right with cur, go forward 1 record.
			 */
			if ((error = xfs_inobt_increment(cur, 0, &i)))
				goto error1;
			doneright = !i;
			if (!doneright) {
				if ((error = xfs_inobt_get_rec(cur,
						&rec.ir_startino,
						&rec.ir_freecount,
						&rec.ir_free, &i)))
					goto error1;
				XFS_WANT_CORRUPTED_GOTO(i == 1, error1);
			}
			/*
			 * Loop until we find the closest inode chunk
			 * with a free one.
			 */
			while (!doneleft || !doneright) {
				int	useleft;  /* using left inode
						     chunk this time */

				/*
				 * Figure out which block is closer,
				 * if both are valid.
				 */
				if (!doneleft && !doneright)
					useleft =
						pagino -
						(trec.ir_startino +
						 XFS_INODES_PER_CHUNK - 1) <
						 rec.ir_startino - pagino;
				else
					useleft = !doneleft;
				/*
				 * If checking the left, does it have
				 * free inodes?
				 */
				if (useleft && trec.ir_freecount) {
					/*
					 * Yes, set it up as the chunk to use.
					 */
					rec = trec;
					xfs_btree_del_cursor(cur,
						XFS_BTREE_NOERROR);
					cur = tcur;
					break;
				}
				/*
				 * If checking the right, does it have
				 * free inodes?
				 */
				if (!useleft && rec.ir_freecount) {
					/*
					 * Yes, it's already set up.
					 */
					xfs_btree_del_cursor(tcur,
						XFS_BTREE_NOERROR);
					break;
				}
				/*
				 * If used the left, get another one
				 * further left.
				 */
				if (useleft) {
					if ((error = xfs_inobt_decrement(tcur, 0,
							&i)))
						goto error1;
					doneleft = !i;
					if (!doneleft) {
						if ((error = xfs_inobt_get_rec(
							    tcur,
							    &trec.ir_startino,
							    &trec.ir_freecount,
							    &trec.ir_free, &i)))
							goto error1;
						XFS_WANT_CORRUPTED_GOTO(i == 1,
							error1);
					}
				}
				/*
				 * If used the right, get another one
				 * further right.
				 */
				else {
					if ((error = xfs_inobt_increment(cur, 0,
							&i)))
						goto error1;
					doneright = !i;
					if (!doneright) {
						if ((error = xfs_inobt_get_rec(
							    cur,
							    &rec.ir_startino,
							    &rec.ir_freecount,
							    &rec.ir_free, &i)))
							goto error1;
						XFS_WANT_CORRUPTED_GOTO(i == 1,
							error1);
					}
				}
			}
			ASSERT(!doneleft || !doneright);
		}
	}
	/*
	 * In a different a.g. from the parent.
	 * See if the most recently allocated block has any free.
	 */
	else if (be32_to_cpu(agi->agi_newino) != NULLAGINO) {
		if ((error = xfs_inobt_lookup_eq(cur,
				be32_to_cpu(agi->agi_newino), 0, 0, &i)))
			goto error0;
		if (i == 1 &&
		    (error = xfs_inobt_get_rec(cur, &rec.ir_startino,
			    &rec.ir_freecount, &rec.ir_free, &j)) == 0 &&
		    j == 1 &&
		    rec.ir_freecount > 0) {
			/*
			 * The last chunk allocated in the group still has
			 * a free inode.
			 */
		}
		/*
		 * None left in the last group, search the whole a.g.
		 */
		else {
			if (error)
				goto error0;
			if ((error = xfs_inobt_lookup_ge(cur, 0, 0, 0, &i)))
				goto error0;
			ASSERT(i == 1);
			for (;;) {
				if ((error = xfs_inobt_get_rec(cur,
						&rec.ir_startino,
						&rec.ir_freecount, &rec.ir_free,
						&i)))
					goto error0;
				XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
				if (rec.ir_freecount > 0)
					break;
				if ((error = xfs_inobt_increment(cur, 0, &i)))
					goto error0;
				XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
			}
		}
	}
	offset = XFS_IALLOC_FIND_FREE(&rec.ir_free);
	ASSERT(offset >= 0);
	ASSERT(offset < XFS_INODES_PER_CHUNK);
	ASSERT((XFS_AGINO_TO_OFFSET(mp, rec.ir_startino) %
				   XFS_INODES_PER_CHUNK) == 0);
	ino = XFS_AGINO_TO_INO(mp, agno, rec.ir_startino + offset);
	XFS_INOBT_CLR_FREE(&rec, offset);
	rec.ir_freecount--;
	if ((error = xfs_inobt_update(cur, rec.ir_startino, rec.ir_freecount,
			rec.ir_free)))
		goto error0;
	be32_add(&agi->agi_freecount, -1);
	xfs_ialloc_log_agi(tp, agbp, XFS_AGI_FREECOUNT);
	down_read(&mp->m_peraglock);
	mp->m_perag[tagno].pagi_freecount--;
	up_read(&mp->m_peraglock);
#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;
			XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
			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);
	xfs_trans_mod_sb(tp, XFS_TRANS_SB_IFREE, -1);
	*inop = ino;
	return 0;
error1:
	xfs_btree_del_cursor(tcur, XFS_BTREE_ERROR);
error0:
	xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
	return error;
}

/*
 * Free disk inode.  Carefully avoids touching the incore inode, all
 * manipulations incore are the caller's responsibility.
 * The on-disk inode is not changed by this operation, only the
 * btree (free inode mask) is changed.
 */
int
xfs_difree(
	xfs_trans_t	*tp,		/* transaction pointer */
	xfs_ino_t	inode,		/* inode to be freed */
	xfs_bmap_free_t	*flist,		/* extents to free */
	int		*delete,	/* set if inode cluster was deleted */
	xfs_ino_t	*first_ino)	/* first inode in deleted cluster */
{
	/* REFERENCED */
	xfs_agblock_t	agbno;	/* block number containing inode */
	xfs_buf_t	*agbp;	/* buffer containing allocation group header */
	xfs_agino_t	agino;	/* inode number relative to allocation group */
	xfs_agnumber_t	agno;	/* allocation group number */
	xfs_agi_t	*agi;	/* allocation group header */
	xfs_btree_cur_t	*cur;	/* inode btree cursor */
	int		error;	/* error return value */
	int		i;	/* result code */
	int		ilen;	/* inodes in an inode cluster */
	xfs_mount_t	*mp;	/* mount structure for filesystem */
	int		off;	/* offset of inode in inode chunk */
	xfs_inobt_rec_incore_t rec;	/* btree record */

	mp = tp->t_mountp;

	/*
	 * Break up inode number into its components.
	 */
	agno = XFS_INO_TO_AGNO(mp, inode);
	if (agno >= mp->m_sb.sb_agcount)  {
		cmn_err(CE_WARN,