xfs_alloc.c

linux 内核源代码

	xfs_extlen_t	delta;	/* new blocks needed in freelist */
	int		error;	/* error result code */
	xfs_extlen_t	longest;/* longest extent in allocation group */
	xfs_mount_t	*mp;	/* file system mount point structure */
	xfs_extlen_t	need;	/* total blocks needed in freelist */
	xfs_perag_t	*pag;	/* per-ag information structure */
	xfs_alloc_arg_t	targs;	/* local allocation arguments */
	xfs_trans_t	*tp;	/* transaction pointer */

	mp = args->mp;

	pag = args->pag;
	tp = args->tp;
	if (!pag->pagf_init) {
		if ((error = xfs_alloc_read_agf(mp, tp, args->agno, flags,
				&agbp)))
			return error;
		if (!pag->pagf_init) {
			ASSERT(flags & XFS_ALLOC_FLAG_TRYLOCK);
			ASSERT(!(flags & XFS_ALLOC_FLAG_FREEING));
			args->agbp = NULL;
			return 0;
		}
	} else
		agbp = NULL;

	/*
	 * If this is a metadata preferred pag and we are user data
	 * then try somewhere else if we are not being asked to
	 * try harder at this point
	 */
	if (pag->pagf_metadata && args->userdata &&
	    (flags & XFS_ALLOC_FLAG_TRYLOCK)) {
		ASSERT(!(flags & XFS_ALLOC_FLAG_FREEING));
		args->agbp = NULL;
		return 0;
	}

	if (!(flags & XFS_ALLOC_FLAG_FREEING)) {
		need = XFS_MIN_FREELIST_PAG(pag, mp);
		delta = need > pag->pagf_flcount ? need - pag->pagf_flcount : 0;
		/*
		 * If it looks like there isn't a long enough extent, or enough
		 * total blocks, reject it.
		 */
		longest = (pag->pagf_longest > delta) ?
			(pag->pagf_longest - delta) :
			(pag->pagf_flcount > 0 || pag->pagf_longest > 0);
		if ((args->minlen + args->alignment + args->minalignslop - 1) >
				longest ||
		    ((int)(pag->pagf_freeblks + pag->pagf_flcount -
			   need - args->total) < (int)args->minleft)) {
			if (agbp)
				xfs_trans_brelse(tp, agbp);
			args->agbp = NULL;
			return 0;
		}
	}

	/*
	 * Get the a.g. freespace buffer.
	 * Can fail if we're not blocking on locks, and it's held.
	 */
	if (agbp == NULL) {
		if ((error = xfs_alloc_read_agf(mp, tp, args->agno, flags,
				&agbp)))
			return error;
		if (agbp == NULL) {
			ASSERT(flags & XFS_ALLOC_FLAG_TRYLOCK);
			ASSERT(!(flags & XFS_ALLOC_FLAG_FREEING));
			args->agbp = NULL;
			return 0;
		}
	}
	/*
	 * Figure out how many blocks we should have in the freelist.
	 */
	agf = XFS_BUF_TO_AGF(agbp);
	need = XFS_MIN_FREELIST(agf, mp);
	/*
	 * If there isn't enough total or single-extent, reject it.
	 */
	if (!(flags & XFS_ALLOC_FLAG_FREEING)) {
		delta = need > be32_to_cpu(agf->agf_flcount) ?
			(need - be32_to_cpu(agf->agf_flcount)) : 0;
		longest = be32_to_cpu(agf->agf_longest);
		longest = (longest > delta) ? (longest - delta) :
			(be32_to_cpu(agf->agf_flcount) > 0 || longest > 0);
		if ((args->minlen + args->alignment + args->minalignslop - 1) >
				longest ||
		    ((int)(be32_to_cpu(agf->agf_freeblks) +
		     be32_to_cpu(agf->agf_flcount) - need - args->total) <
				(int)args->minleft)) {
			xfs_trans_brelse(tp, agbp);
			args->agbp = NULL;
			return 0;
		}
	}
	/*
	 * Make the freelist shorter if it's too long.
	 */
	while (be32_to_cpu(agf->agf_flcount) > need) {
		xfs_buf_t	*bp;

		error = xfs_alloc_get_freelist(tp, agbp, &bno, 0);
		if (error)
			return error;
		if ((error = xfs_free_ag_extent(tp, agbp, args->agno, bno, 1, 1)))
			return error;
		bp = xfs_btree_get_bufs(mp, tp, args->agno, bno, 0);
		xfs_trans_binval(tp, bp);
	}
	/*
	 * Initialize the args structure.
	 */
	targs.tp = tp;
	targs.mp = mp;
	targs.agbp = agbp;
	targs.agno = args->agno;
	targs.mod = targs.minleft = targs.wasdel = targs.userdata =
		targs.minalignslop = 0;
	targs.alignment = targs.minlen = targs.prod = targs.isfl = 1;
	targs.type = XFS_ALLOCTYPE_THIS_AG;
	targs.pag = pag;
	if ((error = xfs_alloc_read_agfl(mp, tp, targs.agno, &agflbp)))
		return error;
	/*
	 * Make the freelist longer if it's too short.
	 */
	while (be32_to_cpu(agf->agf_flcount) < need) {
		targs.agbno = 0;
		targs.maxlen = need - be32_to_cpu(agf->agf_flcount);
		/*
		 * Allocate as many blocks as possible at once.
		 */
		if ((error = xfs_alloc_ag_vextent(&targs))) {
			xfs_trans_brelse(tp, agflbp);
			return error;
		}
		/*
		 * Stop if we run out.  Won't happen if callers are obeying
		 * the restrictions correctly.  Can happen for free calls
		 * on a completely full ag.
		 */
		if (targs.agbno == NULLAGBLOCK) {
			if (flags & XFS_ALLOC_FLAG_FREEING)
				break;
			xfs_trans_brelse(tp, agflbp);
			args->agbp = NULL;
			return 0;
		}
		/*
		 * Put each allocated block on the list.
		 */
		for (bno = targs.agbno; bno < targs.agbno + targs.len; bno++) {
			error = xfs_alloc_put_freelist(tp, agbp,
							agflbp, bno, 0);
			if (error)
				return error;
		}
	}
	xfs_trans_brelse(tp, agflbp);
	args->agbp = agbp;
	return 0;
}

/*
 * Get a block from the freelist.
 * Returns with the buffer for the block gotten.
 */
int				/* error */
xfs_alloc_get_freelist(
	xfs_trans_t	*tp,	/* transaction pointer */
	xfs_buf_t	*agbp,	/* buffer containing the agf structure */
	xfs_agblock_t	*bnop,	/* block address retrieved from freelist */
	int		btreeblk) /* destination is a AGF btree */
{
	xfs_agf_t	*agf;	/* a.g. freespace structure */
	xfs_agfl_t	*agfl;	/* a.g. freelist structure */
	xfs_buf_t	*agflbp;/* buffer for a.g. freelist structure */
	xfs_agblock_t	bno;	/* block number returned */
	int		error;
	int		logflags;
	xfs_mount_t	*mp;	/* mount structure */
	xfs_perag_t	*pag;	/* per allocation group data */

	agf = XFS_BUF_TO_AGF(agbp);
	/*
	 * Freelist is empty, give up.
	 */
	if (!agf->agf_flcount) {
		*bnop = NULLAGBLOCK;
		return 0;
	}
	/*
	 * Read the array of free blocks.
	 */
	mp = tp->t_mountp;
	if ((error = xfs_alloc_read_agfl(mp, tp,
			be32_to_cpu(agf->agf_seqno), &agflbp)))
		return error;
	agfl = XFS_BUF_TO_AGFL(agflbp);
	/*
	 * Get the block number and update the data structures.
	 */
	bno = be32_to_cpu(agfl->agfl_bno[be32_to_cpu(agf->agf_flfirst)]);
	be32_add(&agf->agf_flfirst, 1);
	xfs_trans_brelse(tp, agflbp);
	if (be32_to_cpu(agf->agf_flfirst) == XFS_AGFL_SIZE(mp))
		agf->agf_flfirst = 0;
	pag = &mp->m_perag[be32_to_cpu(agf->agf_seqno)];
	be32_add(&agf->agf_flcount, -1);
	xfs_trans_agflist_delta(tp, -1);
	pag->pagf_flcount--;

	logflags = XFS_AGF_FLFIRST | XFS_AGF_FLCOUNT;
	if (btreeblk) {
		be32_add(&agf->agf_btreeblks, 1);
		pag->pagf_btreeblks++;
		logflags |= XFS_AGF_BTREEBLKS;
	}

	TRACE_MODAGF(NULL, agf, logflags);
	xfs_alloc_log_agf(tp, agbp, logflags);
	*bnop = bno;

	/*
	 * As blocks are freed, they are added to the per-ag busy list
	 * and remain there until the freeing transaction is committed to
	 * disk.  Now that we have allocated blocks, this list must be
	 * searched to see if a block is being reused.  If one is, then
	 * the freeing transaction must be pushed to disk NOW by forcing
	 * to disk all iclogs up that transaction's LSN.
	 */
	xfs_alloc_search_busy(tp, be32_to_cpu(agf->agf_seqno), bno, 1);
	return 0;
}

/*
 * Log the given fields from the agf structure.
 */
void
xfs_alloc_log_agf(
	xfs_trans_t	*tp,	/* transaction pointer */
	xfs_buf_t	*bp,	/* buffer for a.g. freelist header */
	int		fields)	/* mask of fields to be logged (XFS_AGF_...) */
{
	int	first;		/* first byte offset */
	int	last;		/* last byte offset */
	static const short	offsets[] = {
		offsetof(xfs_agf_t, agf_magicnum),
		offsetof(xfs_agf_t, agf_versionnum),
		offsetof(xfs_agf_t, agf_seqno),
		offsetof(xfs_agf_t, agf_length),
		offsetof(xfs_agf_t, agf_roots[0]),
		offsetof(xfs_agf_t, agf_levels[0]),
		offsetof(xfs_agf_t, agf_flfirst),
		offsetof(xfs_agf_t, agf_fllast),
		offsetof(xfs_agf_t, agf_flcount),
		offsetof(xfs_agf_t, agf_freeblks),
		offsetof(xfs_agf_t, agf_longest),
		offsetof(xfs_agf_t, agf_btreeblks),
		sizeof(xfs_agf_t)
	};

	xfs_btree_offsets(fields, offsets, XFS_AGF_NUM_BITS, &first, &last);
	xfs_trans_log_buf(tp, bp, (uint)first, (uint)last);
}

/*
 * Interface for inode allocation to force the pag data to be initialized.
 */
int					/* error */
xfs_alloc_pagf_init(
	xfs_mount_t		*mp,	/* file system mount structure */
	xfs_trans_t		*tp,	/* transaction pointer */
	xfs_agnumber_t		agno,	/* allocation group number */
	int			flags)	/* XFS_ALLOC_FLAGS_... */
{
	xfs_buf_t		*bp;
	int			error;

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

/*
 * Put the block on the freelist for the allocation group.
 */
int					/* error */
xfs_alloc_put_freelist(
	xfs_trans_t		*tp,	/* transaction pointer */
	xfs_buf_t		*agbp,	/* buffer for a.g. freelist header */
	xfs_buf_t		*agflbp,/* buffer for a.g. free block array */
	xfs_agblock_t		bno,	/* block being freed */
	int			btreeblk) /* block came from a AGF btree */
{
	xfs_agf_t		*agf;	/* a.g. freespace structure */
	xfs_agfl_t		*agfl;	/* a.g. free block array */
	__be32			*blockp;/* pointer to array entry */
	int			error;
	int			logflags;
	xfs_mount_t		*mp;	/* mount structure */
	xfs_perag_t		*pag;	/* per allocation group data */

	agf = XFS_BUF_TO_AGF(agbp);
	mp = tp->t_mountp;

	if (!agflbp && (error = xfs_alloc_read_agfl(mp, tp,
			be32_to_cpu(agf->agf_seqno), &agflbp)))
		return error;
	agfl = XFS_BUF_TO_AGFL(agflbp);
	be32_add(&agf->agf_fllast, 1);
	if (be32_to_cpu(agf->agf_fllast) == XFS_AGFL_SIZE(mp))
		agf->agf_fllast = 0;
	pag = &mp->m_perag[be32_to_cpu(agf->agf_seqno)];
	be32_add(&agf->agf_flcount, 1);
	xfs_trans_agflist_delta(tp, 1);
	pag->pagf_flcount++;

	logflags = XFS_AGF_FLLAST | XFS_AGF_FLCOUNT;
	if (btreeblk) {
		be32_add(&agf->agf_btreeblks, -1);
		pag->pagf_btreeblks--;
		logflags |= XFS_AGF_BTREEBLKS;
	}

	TRACE_MODAGF(NULL, agf, logflags);
	xfs_alloc_log_agf(tp, agbp, logflags);

	ASSERT(be32_to_cpu(agf->agf_flcount) <= XFS_AGFL_SIZE(mp));
	blockp = &agfl->agfl_bno[be32_to_cpu(agf->agf_fllast)];
	*blockp = cpu_to_be32(bno);
	TRACE_MODAGF(NULL, agf, logflags);
	xfs_alloc_log_agf(tp, agbp, logflags);
	xfs_trans_log_buf(tp, agflbp,
		(int)((xfs_caddr_t)blockp - (xfs_caddr_t)agfl),
		(int)((xfs_caddr_t)blockp - (xfs_caddr_t)agfl +
			sizeof(xfs_agblock_t) - 1));
	return 0;
}

/*
 * Read in the allocation group header (free/alloc section).
 */
int					/* error */
xfs_alloc_read_agf(
	xfs_mount_t	*mp,		/* mount point structure */
	xfs_trans_t	*tp,		/* transaction pointer */
	xfs_agnumber_t	agno,		/* allocation group number */
	int		flags,		/* XFS_ALLOC_FLAG_... */
	xfs_buf_t	**bpp)		/* buffer for the ag freelist header */
{
	xfs_agf_t	*agf;		/* ag freelist header */
	int		agf_ok;		/* set if agf is consistent */
	xfs_buf_t	*bp;		/* return value */
	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_AGF_DADDR(mp)),
			XFS_FSS_TO_BB(mp, 1),
			(flags & XFS_ALLOC_FLAG_TRYLOCK) ? XFS_BUF_TRYLOCK : 0U,
			&bp);
	if (error)
		return error;
	ASSERT(!bp || !XFS_BUF_GETERROR(bp));
	if (!bp) {
		*bpp = NULL;
		return 0;
	}
	/*
	 * Validate the magic number of the agf block.
	 */
	agf = XFS_BUF_TO_AGF(bp);
	agf_ok =
		be32_to_cpu(agf->agf_magicnum) == XFS_AGF_MAGIC &&
		XFS_AGF_GOOD_VERSION(be32_to_cpu(agf->agf_versionnum)) &&
		be32_to_cpu(agf->agf_freeblks) <= be32_to_cpu(agf->agf_length) &&
		be32_to_cpu(agf->agf_flfirst) < XFS_AGFL_SIZE(mp) &&
		be32_to_cpu(agf->agf_fllast) < XFS_AGFL_SIZE(mp) &&
		be32_to_cpu(agf->agf_flcount) <= XFS_AGFL_SIZE(mp);
	if (unlikely(XFS_TEST_ERROR(!agf_ok, mp, XFS_ERRTAG_ALLOC_READ_AGF,
			XFS_RANDOM_ALLOC_READ_AGF))) {
		XFS_CORRUPTION_ERROR("xfs_alloc_read_agf",
				     XFS_ERRLEVEL_LOW, mp, agf);
		xfs_trans_brelse(tp, bp);
		return XFS_ERROR(EFSCORRUPTED);
	}
	pag = &mp->m_perag[agno];
	if (!pag->pagf_init) {
		pag->pagf_freeblks = be32_to_cpu(agf->agf_freeblks);
		pag->pagf_btreeblks = be32_to_cpu(agf->agf_btreeblks);
		pag->pagf_flcount = be32_to_cpu(agf->agf_flcount);
		pag->pagf_longest = be32_to_cpu(agf->agf_longest);
		pag->pagf_levels[XFS_BTNUM_BNOi] =
			be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNOi]);
		pag->pagf_levels[XFS_BTNUM_CNTi] =
			be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNTi]);
		spinlock_init(&pag->pagb_lock, "xfspagb");
		pag->pagb_list = kmem_zalloc(XFS_PAGB_NUM_SLOTS *
					sizeof(xfs_perag_busy_t), KM_SLEEP);
		pag->pagf_init = 1;
	}
#ifdef DEBUG
	else if (!XFS_FORCED_SHUTDOWN(mp)) {
		ASSERT(pag->pagf_freeblks == be32_to_cpu(agf->agf_freeblks));
		ASSERT(pag->pagf_flcount == be32_to_cpu(agf->agf_flcount));
		ASSERT(pag->pagf_longest == be32_to_cpu(agf->agf_longest));
		ASSERT(pag->pagf_levels[XFS_BTNUM_BNOi] ==
		       be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNOi]));
		ASSERT(pag->pagf_levels[XFS_BTNUM_CNTi] ==
		       be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNTi]));
	}
#endif
	XFS_BUF_SET_VTYPE_REF(bp, B_FS_AGF, XFS_AGF_REF);
	*bpp = bp;
	return 0;
}

/*
 * Allocate an extent (variable-size).
 * Depending on the allocation type, we either look in a single allocation
 * group or loop over the allocation groups to find the result.
 */
int				/* error */
xfs_alloc_vextent(
	xfs_alloc_arg_t	*args)	/* allocation argument structure */
{
	xfs_agblock_t	agsize;	/* allocation group size */
	int		error;
	int		flags;	/* XFS_ALLOC_FLAG_... locking flags */
	xfs_extlen_t	minleft;/* minimum left value, temp copy */
	xfs_mount_t	*mp;	/* mount structure pointer */
	xfs_agnumber_t	sagno;	/* starting allocation group number */
	xfs_alloctype_t	type;	/* input allocation type */
	int		bump_rotor = 0;
	int		no_min = 0;
	xfs_agnumber_t	rotorstep = xfs_rotorstep; /* inode32 agf stepper */

	mp = args->mp;
	type = args->otype = args->type;
	args-