[mirror_ubuntu-artful-kernel.git] / fs / xfs / xfs_dir2_readdir.c

/*
 * Copyright (c) 2000-2005 Silicon Graphics, Inc.
 * Copyright (c) 2013 Red Hat, Inc.
 * All Rights Reserved.
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public License as
 * published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it would be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write the Free Software Foundation,
 * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 */
#include "xfs.h"
#include "xfs_fs.h"
#include "xfs_format.h"
#include "xfs_log_format.h"
#include "xfs_trans_resv.h"
#include "xfs_bit.h"
#include "xfs_mount.h"
#include "xfs_da_format.h"
#include "xfs_da_btree.h"
#include "xfs_inode.h"
#include "xfs_dir2.h"
#include "xfs_dir2_priv.h"
#include "xfs_error.h"
#include "xfs_trace.h"
#include "xfs_bmap.h"
#include "xfs_trans.h"

/*
 * Directory file type support functions
 */
static unsigned char xfs_dir3_filetype_table[] = {
	DT_UNKNOWN, DT_REG, DT_DIR, DT_CHR, DT_BLK,
	DT_FIFO, DT_SOCK, DT_LNK, DT_WHT,
};

static unsigned char
xfs_dir3_get_dtype(
	struct xfs_mount	*mp,
	uint8_t			filetype)
{
	if (!xfs_sb_version_hasftype(&mp->m_sb))
		return DT_UNKNOWN;

	if (filetype >= XFS_DIR3_FT_MAX)
		return DT_UNKNOWN;

	return xfs_dir3_filetype_table[filetype];
}

STATIC int
xfs_dir2_sf_getdents(
	struct xfs_da_args	*args,
	struct dir_context	*ctx)
{
	int			i;		/* shortform entry number */
	struct xfs_inode	*dp = args->dp;	/* incore directory inode */
	xfs_dir2_dataptr_t	off;		/* current entry's offset */
	xfs_dir2_sf_entry_t	*sfep;		/* shortform directory entry */
	xfs_dir2_sf_hdr_t	*sfp;		/* shortform structure */
	xfs_dir2_dataptr_t	dot_offset;
	xfs_dir2_dataptr_t	dotdot_offset;
	xfs_ino_t		ino;
	struct xfs_da_geometry	*geo = args->geo;

	ASSERT(dp->i_df.if_flags & XFS_IFINLINE);
	ASSERT(dp->i_df.if_bytes == dp->i_d.di_size);
	ASSERT(dp->i_df.if_u1.if_data != NULL);

	sfp = (xfs_dir2_sf_hdr_t *)dp->i_df.if_u1.if_data;

	/*
	 * If the block number in the offset is out of range, we're done.
	 */
	if (xfs_dir2_dataptr_to_db(geo, ctx->pos) > geo->datablk)
		return 0;

	/*
	 * Precalculate offsets for . and .. as we will always need them.
	 *
	 * XXX(hch): the second argument is sometimes 0 and sometimes
	 * geo->datablk
	 */
	dot_offset = xfs_dir2_db_off_to_dataptr(geo, geo->datablk,
						dp->d_ops->data_dot_offset);
	dotdot_offset = xfs_dir2_db_off_to_dataptr(geo, geo->datablk,
						dp->d_ops->data_dotdot_offset);

	/*
	 * Put . entry unless we're starting past it.
	 */
	if (ctx->pos <= dot_offset) {
		ctx->pos = dot_offset & 0x7fffffff;
		if (!dir_emit(ctx, ".", 1, dp->i_ino, DT_DIR))
			return 0;
	}

	/*
	 * Put .. entry unless we're starting past it.
	 */
	if (ctx->pos <= dotdot_offset) {
		ino = dp->d_ops->sf_get_parent_ino(sfp);
		ctx->pos = dotdot_offset & 0x7fffffff;
		if (!dir_emit(ctx, "..", 2, ino, DT_DIR))
			return 0;
	}

	/*
	 * Loop while there are more entries and put'ing works.
	 */
	sfep = xfs_dir2_sf_firstentry(sfp);
	for (i = 0; i < sfp->count; i++) {
		uint8_t filetype;

		off = xfs_dir2_db_off_to_dataptr(geo, geo->datablk,
				xfs_dir2_sf_get_offset(sfep));

		if (ctx->pos > off) {
			sfep = dp->d_ops->sf_nextentry(sfp, sfep);
			continue;
		}

		ino = dp->d_ops->sf_get_ino(sfp, sfep);
		filetype = dp->d_ops->sf_get_ftype(sfep);
		ctx->pos = off & 0x7fffffff;
		if (!dir_emit(ctx, (char *)sfep->name, sfep->namelen, ino,
			    xfs_dir3_get_dtype(dp->i_mount, filetype)))
			return 0;
		sfep = dp->d_ops->sf_nextentry(sfp, sfep);
	}

	ctx->pos = xfs_dir2_db_off_to_dataptr(geo, geo->datablk + 1, 0) &
								0x7fffffff;
	return 0;
}

/*
 * Readdir for block directories.
 */
STATIC int
xfs_dir2_block_getdents(
	struct xfs_da_args	*args,
	struct dir_context	*ctx)
{
	struct xfs_inode	*dp = args->dp;	/* incore directory inode */
	xfs_dir2_data_hdr_t	*hdr;		/* block header */
	struct xfs_buf		*bp;		/* buffer for block */
	xfs_dir2_block_tail_t	*btp;		/* block tail */
	xfs_dir2_data_entry_t	*dep;		/* block data entry */
	xfs_dir2_data_unused_t	*dup;		/* block unused entry */
	char			*endptr;	/* end of the data entries */
	int			error;		/* error return value */
	char			*ptr;		/* current data entry */
	int			wantoff;	/* starting block offset */
	xfs_off_t		cook;
	struct xfs_da_geometry	*geo = args->geo;
	int			lock_mode;

	/*
	 * If the block number in the offset is out of range, we're done.
	 */
	if (xfs_dir2_dataptr_to_db(geo, ctx->pos) > geo->datablk)
		return 0;

	lock_mode = xfs_ilock_data_map_shared(dp);
	error = xfs_dir3_block_read(args->trans, dp, &bp);
	xfs_iunlock(dp, lock_mode);
	if (error)
		return error;

	/*
	 * Extract the byte offset we start at from the seek pointer.
	 * We'll skip entries before this.
	 */
	wantoff = xfs_dir2_dataptr_to_off(geo, ctx->pos);
	hdr = bp->b_addr;
	xfs_dir3_data_check(dp, bp);
	/*
	 * Set up values for the loop.
	 */
	btp = xfs_dir2_block_tail_p(geo, hdr);
	ptr = (char *)dp->d_ops->data_entry_p(hdr);
	endptr = (char *)xfs_dir2_block_leaf_p(btp);

	/*
	 * Loop over the data portion of the block.
	 * Each object is a real entry (dep) or an unused one (dup).
	 */
	while (ptr < endptr) {
		uint8_t filetype;

		dup = (xfs_dir2_data_unused_t *)ptr;
		/*
		 * Unused, skip it.
		 */
		if (be16_to_cpu(dup->freetag) == XFS_DIR2_DATA_FREE_TAG) {
			ptr += be16_to_cpu(dup->length);
			continue;
		}

		dep = (xfs_dir2_data_entry_t *)ptr;

		/*
		 * Bump pointer for the next iteration.
		 */
		ptr += dp->d_ops->data_entsize(dep->namelen);
		/*
		 * The entry is before the desired starting point, skip it.
		 */
		if ((char *)dep - (char *)hdr < wantoff)
			continue;

		cook = xfs_dir2_db_off_to_dataptr(geo, geo->datablk,
					    (char *)dep - (char *)hdr);

		ctx->pos = cook & 0x7fffffff;
		filetype = dp->d_ops->data_get_ftype(dep);
		/*
		 * If it didn't fit, set the final offset to here & return.
		 */
		if (!dir_emit(ctx, (char *)dep->name, dep->namelen,
			    be64_to_cpu(dep->inumber),
			    xfs_dir3_get_dtype(dp->i_mount, filetype))) {
			xfs_trans_brelse(args->trans, bp);
			return 0;
		}
	}

	/*
	 * Reached the end of the block.
	 * Set the offset to a non-existent block 1 and return.
	 */
	ctx->pos = xfs_dir2_db_off_to_dataptr(geo, geo->datablk + 1, 0) &
								0x7fffffff;
	xfs_trans_brelse(args->trans, bp);
	return 0;
}

/*
 * Read a directory block and initiate readahead for blocks beyond that.
 * We maintain a sliding readahead window of the remaining space in the
 * buffer rounded up to the nearest block.
 */
STATIC int
xfs_dir2_leaf_readbuf(
	struct xfs_da_args	*args,
	size_t			bufsize,
	xfs_dir2_off_t		*cur_off,
	xfs_dablk_t		*ra_blk,
	struct xfs_buf		**bpp)
{
	struct xfs_inode	*dp = args->dp;
	struct xfs_buf		*bp = NULL;
	struct xfs_da_geometry	*geo = args->geo;
	struct xfs_ifork	*ifp = XFS_IFORK_PTR(dp, XFS_DATA_FORK);
	struct xfs_bmbt_irec	map;
	struct blk_plug		plug;
	xfs_dir2_off_t		new_off;
	xfs_dablk_t		next_ra;
	xfs_dablk_t		map_off;
	xfs_dablk_t		last_da;
	xfs_extnum_t		idx;
	int			ra_want;
	int			error = 0;

	if (!(ifp->if_flags & XFS_IFEXTENTS)) {
		error = xfs_iread_extents(args->trans, dp, XFS_DATA_FORK);
		if (error)
			goto out;
	}

	/*
	 * Look for mapped directory blocks at or above the current offset.
	 * Truncate down to the nearest directory block to start the scanning
	 * operation.
	 */
	last_da = xfs_dir2_byte_to_da(geo, XFS_DIR2_LEAF_OFFSET);
	map_off = xfs_dir2_db_to_da(geo, xfs_dir2_byte_to_db(geo, *cur_off));
	if (!xfs_iext_lookup_extent(dp, ifp, map_off, &idx, &map))
		goto out;
	if (map.br_startoff >= last_da)
		goto out;
	xfs_trim_extent(&map, map_off, last_da - map_off);

	/* Read the directory block of that first mapping. */
	new_off = xfs_dir2_da_to_byte(geo, map.br_startoff);
	if (new_off > *cur_off)
		*cur_off = new_off;
	error = xfs_dir3_data_read(args->trans, dp, map.br_startoff, -1, &bp);
	if (error)
		goto out;

	/*
	 * Start readahead for the next bufsize's worth of dir data blocks.
	 * We may have already issued readahead for some of that range;
	 * ra_blk tracks the last block we tried to read(ahead).
	 */
	ra_want = howmany(bufsize + geo->blksize, (1 << geo->fsblog));
	if (*ra_blk >= last_da)
		goto out;
	else if (*ra_blk == 0)
		*ra_blk = map.br_startoff;
	next_ra = map.br_startoff + geo->fsbcount;
	if (next_ra >= last_da)
		goto out_no_ra;
	if (map.br_blockcount < geo->fsbcount &&
	    !xfs_iext_get_extent(ifp, ++idx, &map))
		goto out_no_ra;
	if (map.br_startoff >= last_da)
		goto out_no_ra;
	xfs_trim_extent(&map, next_ra, last_da - next_ra);

	/* Start ra for each dir (not fs) block that has a mapping. */
	blk_start_plug(&plug);
	while (ra_want > 0) {
		next_ra = roundup((xfs_dablk_t)map.br_startoff, geo->fsbcount);
		while (ra_want > 0 &&
		       next_ra < map.br_startoff + map.br_blockcount) {
			if (next_ra >= last_da) {
				*ra_blk = last_da;
				break;
			}
			if (next_ra > *ra_blk) {
				xfs_dir3_data_readahead(dp, next_ra, -2);
				*ra_blk = next_ra;
			}
			ra_want -= geo->fsbcount;
			next_ra += geo->fsbcount;
		}
		if (!xfs_iext_get_extent(ifp, ++idx, &map)) {
			*ra_blk = last_da;
			break;
		}
	}
	blk_finish_plug(&plug);

out:
	*bpp = bp;
	return error;
out_no_ra:
	*ra_blk = last_da;
	goto out;
}

/*
 * Getdents (readdir) for leaf and node directories.
 * This reads the data blocks only, so is the same for both forms.
 */
STATIC int
xfs_dir2_leaf_getdents(
	struct xfs_da_args	*args,
	struct dir_context	*ctx,
	size_t			bufsize)
{
	struct xfs_inode	*dp = args->dp;
	struct xfs_buf		*bp = NULL;	/* data block buffer */
	xfs_dir2_data_hdr_t	*hdr;		/* data block header */
	xfs_dir2_data_entry_t	*dep;		/* data entry */
	xfs_dir2_data_unused_t	*dup;		/* unused entry */
	char			*ptr = NULL;	/* pointer to current data */
	struct xfs_da_geometry	*geo = args->geo;
	xfs_dablk_t		rablk = 0;	/* current readahead block */
	xfs_dir2_off_t		curoff;		/* current overall offset */
	int			length;		/* temporary length value */
	int			byteoff;	/* offset in current block */
	int			lock_mode;
	int			error = 0;	/* error return value */

	/*
	 * If the offset is at or past the largest allowed value,
	 * give up right away.
	 */
	if (ctx->pos >= XFS_DIR2_MAX_DATAPTR)
		return 0;

	/*
	 * Inside the loop we keep the main offset value as a byte offset
	 * in the directory file.
	 */
	curoff = xfs_dir2_dataptr_to_byte(ctx->pos);

	/*
	 * Loop over directory entries until we reach the end offset.
	 * Get more blocks and readahead as necessary.
	 */
	while (curoff < XFS_DIR2_LEAF_OFFSET) {
		uint8_t filetype;

		/*
		 * If we have no buffer, or we're off the end of the
		 * current buffer, need to get another one.
		 */
		if (!bp || ptr >= (char *)bp->b_addr + geo->blksize) {
			if (bp) {
				xfs_trans_brelse(args->trans, bp);
				bp = NULL;
			}

			lock_mode = xfs_ilock_data_map_shared(dp);
			error = xfs_dir2_leaf_readbuf(args, bufsize, &curoff,
					&rablk, &bp);
			xfs_iunlock(dp, lock_mode);
			if (error || !bp)
				break;

			hdr = bp->b_addr;
			xfs_dir3_data_check(dp, bp);
			/*
			 * Find our position in the block.
			 */
			ptr = (char *)dp->d_ops->data_entry_p(hdr);
			byteoff = xfs_dir2_byte_to_off(geo, curoff);
			/*
			 * Skip past the header.
			 */
			if (byteoff == 0)
				curoff += dp->d_ops->data_entry_offset;
			/*
			 * Skip past entries until we reach our offset.
			 */
			else {
				while ((char *)ptr - (char *)hdr < byteoff) {
					dup = (xfs_dir2_data_unused_t *)ptr;

					if (be16_to_cpu(dup->freetag)
						  == XFS_DIR2_DATA_FREE_TAG) {

						length = be16_to_cpu(dup->length);
						ptr += length;
						continue;
					}
					dep = (xfs_dir2_data_entry_t *)ptr;
					length =
					   dp->d_ops->data_entsize(dep->namelen);
					ptr += length;
				}
				/*
				 * Now set our real offset.
				 */
				curoff =
					xfs_dir2_db_off_to_byte(geo,
					    xfs_dir2_byte_to_db(geo, curoff),
					    (char *)ptr - (char *)hdr);
				if (ptr >= (char *)hdr + geo->blksize) {
					continue;
				}
			}
		}
		/*
		 * We have a pointer to an entry.
		 * Is it a live one?
		 */
		dup = (xfs_dir2_data_unused_t *)ptr;
		/*
		 * No, it's unused, skip over it.
		 */
		if (be16_to_cpu(dup->freetag) == XFS_DIR2_DATA_FREE_TAG) {
			length = be16_to_cpu(dup->length);
			ptr += length;
			curoff += length;
			continue;
		}

		dep = (xfs_dir2_data_entry_t *)ptr;
		length = dp->d_ops->data_entsize(dep->namelen);
		filetype = dp->d_ops->data_get_ftype(dep);

		ctx->pos = xfs_dir2_byte_to_dataptr(curoff) & 0x7fffffff;
		if (!dir_emit(ctx, (char *)dep->name, dep->namelen,
			    be64_to_cpu(dep->inumber),
			    xfs_dir3_get_dtype(dp->i_mount, filetype)))
			break;

		/*
		 * Advance to next entry in the block.
		 */
		ptr += length;
		curoff += length;
		/* bufsize may have just been a guess; don't go negative */
		bufsize = bufsize > length ? bufsize - length : 0;
	}

	/*
	 * All done.  Set output offset value to current offset.
	 */
	if (curoff > xfs_dir2_dataptr_to_byte(XFS_DIR2_MAX_DATAPTR))
		ctx->pos = XFS_DIR2_MAX_DATAPTR & 0x7fffffff;
	else
		ctx->pos = xfs_dir2_byte_to_dataptr(curoff) & 0x7fffffff;
	if (bp)
		xfs_trans_brelse(args->trans, bp);
	return error;
}

/*
 * Read a directory.
 *
 * If supplied, the transaction collects locked dir buffers to avoid
 * nested buffer deadlocks.  This function does not dirty the
 * transaction.  The caller should ensure that the inode is locked
 * before calling this function.
 */
int
xfs_readdir(
	struct xfs_trans	*tp,
	struct xfs_inode	*dp,
	struct dir_context	*ctx,
	size_t			bufsize)
{
	struct xfs_da_args	args = { NULL };
	int			rval;
	int			v;

	trace_xfs_readdir(dp);

	if (XFS_FORCED_SHUTDOWN(dp->i_mount))
		return -EIO;

	ASSERT(S_ISDIR(VFS_I(dp)->i_mode));
	XFS_STATS_INC(dp->i_mount, xs_dir_getdents);

	args.dp = dp;
	args.geo = dp->i_mount->m_dir_geo;
	args.trans = tp;

	if (dp->i_d.di_format == XFS_DINODE_FMT_LOCAL)
		rval = xfs_dir2_sf_getdents(&args, ctx);
	else if ((rval = xfs_dir2_isblock(&args, &v)))
		;
	else if (v)
		rval = xfs_dir2_block_getdents(&args, ctx);
	else
		rval = xfs_dir2_leaf_getdents(&args, ctx, bufsize);

	return rval;
}
Commit	Line	Data
4a8af273 DC	1	/*
	2	* Copyright (c) 2000-2005 Silicon Graphics, Inc.
	3	* Copyright (c) 2013 Red Hat, Inc.
	4	* All Rights Reserved.
	5	*
	6	* This program is free software; you can redistribute it and/or
	7	* modify it under the terms of the GNU General Public License as
	8	* published by the Free Software Foundation.
	9	*
	10	* This program is distributed in the hope that it would be useful,
	11	* but WITHOUT ANY WARRANTY; without even the implied warranty of
	12	* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
	13	* GNU General Public License for more details.
	14	*
	15	* You should have received a copy of the GNU General Public License
	16	* along with this program; if not, write the Free Software Foundation,
	17	* Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
	18	*/
	19	#include "xfs.h"
	20	#include "xfs_fs.h"
a4fbe6ab	21	#include "xfs_format.h"
239880ef DC	22	#include "xfs_log_format.h"
239880ef DC	23	#include "xfs_trans_resv.h"
4a8af273	24	#include "xfs_bit.h"
4a8af273	25	#include "xfs_mount.h"
57062787	26	#include "xfs_da_format.h"
4a8af273	27	#include "xfs_da_btree.h"
4a8af273	28	#include "xfs_inode.h"
2b9ab5ab	29	#include "xfs_dir2.h"
4a8af273 DC	30	#include "xfs_dir2_priv.h"
	31	#include "xfs_error.h"
	32	#include "xfs_trace.h"
	33	#include "xfs_bmap.h"
239880ef	34	#include "xfs_trans.h"
4a8af273	35
0cb97766 DC	36	/*
	37	* Directory file type support functions
	38	*/
	39	static unsigned char xfs_dir3_filetype_table[] = {
	40	DT_UNKNOWN, DT_REG, DT_DIR, DT_CHR, DT_BLK,
	41	DT_FIFO, DT_SOCK, DT_LNK, DT_WHT,
	42	};
	43
1b767ee3	44	static unsigned char
0cb97766 DC	45	xfs_dir3_get_dtype(
0cb97766 DC	46	struct xfs_mount *mp,
c8ce540d	47	uint8_t filetype)
0cb97766 DC	48	{
	49	if (!xfs_sb_version_hasftype(&mp->m_sb))
	50	return DT_UNKNOWN;
	51
	52	if (filetype >= XFS_DIR3_FT_MAX)
	53	return DT_UNKNOWN;
	54
	55	return xfs_dir3_filetype_table[filetype];
	56	}
0cb97766	57
4a8af273 DC	58	STATIC int
4a8af273 DC	59	xfs_dir2_sf_getdents(
53f82db0	60	struct xfs_da_args *args,
4a8af273 DC	61	struct dir_context *ctx)
	62	{
	63	int i; /* shortform entry number */
53f82db0	64	struct xfs_inode dp = args->dp; / incore directory inode */
4a8af273 DC	65	xfs_dir2_dataptr_t off; /* current entry's offset */
	66	xfs_dir2_sf_entry_t sfep; / shortform directory entry */
	67	xfs_dir2_sf_hdr_t sfp; / shortform structure */
	68	xfs_dir2_dataptr_t dot_offset;
	69	xfs_dir2_dataptr_t dotdot_offset;
	70	xfs_ino_t ino;
53f82db0	71	struct xfs_da_geometry *geo = args->geo;
4a8af273 DC	72
4a8af273 DC	73	ASSERT(dp->i_df.if_flags & XFS_IFINLINE);
4a8af273 DC	74	ASSERT(dp->i_df.if_bytes == dp->i_d.di_size);
	75	ASSERT(dp->i_df.if_u1.if_data != NULL);
	76
	77	sfp = (xfs_dir2_sf_hdr_t *)dp->i_df.if_u1.if_data;
	78
4a8af273 DC	79	/*
	80	* If the block number in the offset is out of range, we're done.
	81	*/
7dda6e86	82	if (xfs_dir2_dataptr_to_db(geo, ctx->pos) > geo->datablk)
4a8af273 DC	83	return 0;
	84
	85	/*
	86	* Precalculate offsets for . and .. as we will always need them.
	87	*
	88	* XXX(hch): the second argument is sometimes 0 and sometimes
7dda6e86	89	* geo->datablk
4a8af273	90	*/
7dda6e86	91	dot_offset = xfs_dir2_db_off_to_dataptr(geo, geo->datablk,
1c9a5b2e	92	dp->d_ops->data_dot_offset);
7dda6e86	93	dotdot_offset = xfs_dir2_db_off_to_dataptr(geo, geo->datablk,
1c9a5b2e	94	dp->d_ops->data_dotdot_offset);
4a8af273 DC	95
	96	/*
	97	* Put . entry unless we're starting past it.
	98	*/
	99	if (ctx->pos <= dot_offset) {
	100	ctx->pos = dot_offset & 0x7fffffff;
	101	if (!dir_emit(ctx, ".", 1, dp->i_ino, DT_DIR))
	102	return 0;
	103	}
	104
	105	/*
	106	* Put .. entry unless we're starting past it.
	107	*/
	108	if (ctx->pos <= dotdot_offset) {
4740175e	109	ino = dp->d_ops->sf_get_parent_ino(sfp);
4a8af273 DC	110	ctx->pos = dotdot_offset & 0x7fffffff;
	111	if (!dir_emit(ctx, "..", 2, ino, DT_DIR))
	112	return 0;
	113	}
	114
	115	/*
	116	* Loop while there are more entries and put'ing works.
	117	*/
	118	sfep = xfs_dir2_sf_firstentry(sfp);
	119	for (i = 0; i < sfp->count; i++) {
c8ce540d	120	uint8_t filetype;
0cb97766	121
7dda6e86	122	off = xfs_dir2_db_off_to_dataptr(geo, geo->datablk,
4a8af273 DC	123	xfs_dir2_sf_get_offset(sfep));
	124
	125	if (ctx->pos > off) {
32c5483a	126	sfep = dp->d_ops->sf_nextentry(sfp, sfep);
4a8af273 DC	127	continue;
	128	}
	129
4740175e DC	130	ino = dp->d_ops->sf_get_ino(sfp, sfep);
4740175e DC	131	filetype = dp->d_ops->sf_get_ftype(sfep);
4a8af273	132	ctx->pos = off & 0x7fffffff;
0cb97766	133	if (!dir_emit(ctx, (char *)sfep->name, sfep->namelen, ino,
53f82db0	134	xfs_dir3_get_dtype(dp->i_mount, filetype)))
4a8af273	135	return 0;
32c5483a	136	sfep = dp->d_ops->sf_nextentry(sfp, sfep);
4a8af273 DC	137	}
4a8af273 DC	138
7dda6e86	139	ctx->pos = xfs_dir2_db_off_to_dataptr(geo, geo->datablk + 1, 0) &
53f82db0	140	0x7fffffff;
4a8af273 DC	141	return 0;
	142	}
	143
	144	/*
	145	* Readdir for block directories.
	146	*/
	147	STATIC int
	148	xfs_dir2_block_getdents(
53f82db0	149	struct xfs_da_args *args,
4a8af273 DC	150	struct dir_context *ctx)
4a8af273 DC	151	{
53f82db0	152	struct xfs_inode dp = args->dp; / incore directory inode */
4a8af273 DC	153	xfs_dir2_data_hdr_t hdr; / block header */
	154	struct xfs_buf bp; / buffer for block */
	155	xfs_dir2_block_tail_t btp; / block tail */
	156	xfs_dir2_data_entry_t dep; / block data entry */
	157	xfs_dir2_data_unused_t dup; / block unused entry */
	158	char endptr; / end of the data entries */
	159	int error; /* error return value */
4a8af273 DC	160	char ptr; / current data entry */
	161	int wantoff; /* starting block offset */
	162	xfs_off_t cook;
53f82db0	163	struct xfs_da_geometry *geo = args->geo;
dbad7c99	164	int lock_mode;
4a8af273	165
4a8af273 DC	166	/*
	167	* If the block number in the offset is out of range, we're done.
	168	*/
7dda6e86	169	if (xfs_dir2_dataptr_to_db(geo, ctx->pos) > geo->datablk)
4a8af273 DC	170	return 0;
4a8af273 DC	171
dbad7c99	172	lock_mode = xfs_ilock_data_map_shared(dp);
acb9553c	173	error = xfs_dir3_block_read(args->trans, dp, &bp);
dbad7c99	174	xfs_iunlock(dp, lock_mode);
4a8af273 DC	175	if (error)
	176	return error;
	177
	178	/*
	179	* Extract the byte offset we start at from the seek pointer.
	180	* We'll skip entries before this.
	181	*/
30028030	182	wantoff = xfs_dir2_dataptr_to_off(geo, ctx->pos);
4a8af273 DC	183	hdr = bp->b_addr;
	184	xfs_dir3_data_check(dp, bp);
	185	/*
	186	* Set up values for the loop.
	187	*/
8f66193c	188	btp = xfs_dir2_block_tail_p(geo, hdr);
2ca98774	189	ptr = (char *)dp->d_ops->data_entry_p(hdr);
4a8af273 DC	190	endptr = (char *)xfs_dir2_block_leaf_p(btp);
	191
	192	/*
	193	* Loop over the data portion of the block.
	194	* Each object is a real entry (dep) or an unused one (dup).
	195	*/
	196	while (ptr < endptr) {
c8ce540d	197	uint8_t filetype;
0cb97766	198
4a8af273 DC	199	dup = (xfs_dir2_data_unused_t *)ptr;
	200	/*
	201	* Unused, skip it.
	202	*/
	203	if (be16_to_cpu(dup->freetag) == XFS_DIR2_DATA_FREE_TAG) {
	204	ptr += be16_to_cpu(dup->length);
	205	continue;
	206	}
	207
	208	dep = (xfs_dir2_data_entry_t *)ptr;
	209
	210	/*
	211	* Bump pointer for the next iteration.
	212	*/
9d23fc85	213	ptr += dp->d_ops->data_entsize(dep->namelen);
4a8af273 DC	214	/*
	215	* The entry is before the desired starting point, skip it.
	216	*/
	217	if ((char )dep - (char )hdr < wantoff)
	218	continue;
	219
7dda6e86	220	cook = xfs_dir2_db_off_to_dataptr(geo, geo->datablk,
4a8af273 DC	221	(char )dep - (char )hdr);
	222
	223	ctx->pos = cook & 0x7fffffff;
9d23fc85	224	filetype = dp->d_ops->data_get_ftype(dep);
4a8af273 DC	225	/*
	226	* If it didn't fit, set the final offset to here & return.
	227	*/
	228	if (!dir_emit(ctx, (char *)dep->name, dep->namelen,
0cb97766	229	be64_to_cpu(dep->inumber),
53f82db0	230	xfs_dir3_get_dtype(dp->i_mount, filetype))) {
acb9553c	231	xfs_trans_brelse(args->trans, bp);
4a8af273 DC	232	return 0;
	233	}
	234	}
	235
	236	/*
	237	* Reached the end of the block.
	238	* Set the offset to a non-existent block 1 and return.
	239	*/
7dda6e86	240	ctx->pos = xfs_dir2_db_off_to_dataptr(geo, geo->datablk + 1, 0) &
53f82db0	241	0x7fffffff;
acb9553c	242	xfs_trans_brelse(args->trans, bp);
4a8af273 DC	243	return 0;
	244	}
	245
d205a7d0 DW	246	/*
	247	* Read a directory block and initiate readahead for blocks beyond that.
	248	* We maintain a sliding readahead window of the remaining space in the
	249	* buffer rounded up to the nearest block.
	250	*/
4a8af273 DC	251	STATIC int
4a8af273 DC	252	xfs_dir2_leaf_readbuf(
53f82db0	253	struct xfs_da_args *args,
4a8af273	254	size_t bufsize,
d205a7d0 DW	255	xfs_dir2_off_t *cur_off,
	256	xfs_dablk_t *ra_blk,
	257	struct xfs_buf **bpp)
4a8af273	258	{
53f82db0	259	struct xfs_inode *dp = args->dp;
9f541801	260	struct xfs_buf *bp = NULL;
d205a7d0 DW	261	struct xfs_da_geometry *geo = args->geo;
	262	struct xfs_ifork *ifp = XFS_IFORK_PTR(dp, XFS_DATA_FORK);
	263	struct xfs_bmbt_irec map;
4a8af273	264	struct blk_plug plug;
d205a7d0 DW	265	xfs_dir2_off_t new_off;
	266	xfs_dablk_t next_ra;
	267	xfs_dablk_t map_off;
	268	xfs_dablk_t last_da;
	269	xfs_extnum_t idx;
	270	int ra_want;
4a8af273	271	int error = 0;
4a8af273	272
d205a7d0 DW	273	if (!(ifp->if_flags & XFS_IFEXTENTS)) {
d205a7d0 DW	274	error = xfs_iread_extents(args->trans, dp, XFS_DATA_FORK);
4a8af273	275	if (error)
d205a7d0	276	goto out;
4a8af273 DC	277	}
	278
	279	/*
d205a7d0 DW	280	* Look for mapped directory blocks at or above the current offset.
	281	* Truncate down to the nearest directory block to start the scanning
	282	* operation.
4a8af273	283	*/
d205a7d0 DW	284	last_da = xfs_dir2_byte_to_da(geo, XFS_DIR2_LEAF_OFFSET);
	285	map_off = xfs_dir2_db_to_da(geo, xfs_dir2_byte_to_db(geo, *cur_off));
	286	if (!xfs_iext_lookup_extent(dp, ifp, map_off, &idx, &map))
4a8af273	287	goto out;
d205a7d0 DW	288	if (map.br_startoff >= last_da)
	289	goto out;
	290	xfs_trim_extent(&map, map_off, last_da - map_off);
4a8af273	291
d205a7d0 DW	292	/* Read the directory block of that first mapping. */
	293	new_off = xfs_dir2_da_to_byte(geo, map.br_startoff);
	294	if (new_off > *cur_off)
	295	*cur_off = new_off;
	296	error = xfs_dir3_data_read(args->trans, dp, map.br_startoff, -1, &bp);
4a8af273	297	if (error)
d205a7d0	298	goto out;
4a8af273 DC	299
4a8af273 DC	300	/*
d205a7d0 DW	301	* Start readahead for the next bufsize's worth of dir data blocks.
	302	* We may have already issued readahead for some of that range;
	303	* ra_blk tracks the last block we tried to read(ahead).
4a8af273	304	*/
d205a7d0 DW	305	ra_want = howmany(bufsize + geo->blksize, (1 << geo->fsblog));
	306	if (*ra_blk >= last_da)
	307	goto out;
	308	else if (*ra_blk == 0)
	309	*ra_blk = map.br_startoff;
	310	next_ra = map.br_startoff + geo->fsbcount;
	311	if (next_ra >= last_da)
	312	goto out_no_ra;
	313	if (map.br_blockcount < geo->fsbcount &&
	314	!xfs_iext_get_extent(ifp, ++idx, &map))
	315	goto out_no_ra;
	316	if (map.br_startoff >= last_da)
	317	goto out_no_ra;
	318	xfs_trim_extent(&map, next_ra, last_da - next_ra);
	319
	320	/* Start ra for each dir (not fs) block that has a mapping. */
4a8af273	321	blk_start_plug(&plug);
d205a7d0 DW	322	while (ra_want > 0) {
	323	next_ra = roundup((xfs_dablk_t)map.br_startoff, geo->fsbcount);
	324	while (ra_want > 0 &&
	325	next_ra < map.br_startoff + map.br_blockcount) {
	326	if (next_ra >= last_da) {
	327	*ra_blk = last_da;
	328	break;
	329	}
	330	if (next_ra > *ra_blk) {
	331	xfs_dir3_data_readahead(dp, next_ra, -2);
	332	*ra_blk = next_ra;
4a8af273	333	}
d205a7d0 DW	334	ra_want -= geo->fsbcount;
	335	next_ra += geo->fsbcount;
	336	}
	337	if (!xfs_iext_get_extent(ifp, ++idx, &map)) {
	338	*ra_blk = last_da;
	339	break;
4a8af273 DC	340	}
	341	}
	342	blk_finish_plug(&plug);
	343
	344	out:
	345	*bpp = bp;
	346	return error;
d205a7d0 DW	347	out_no_ra:
	348	*ra_blk = last_da;
	349	goto out;
4a8af273 DC	350	}
	351
	352	/*
	353	* Getdents (readdir) for leaf and node directories.
	354	* This reads the data blocks only, so is the same for both forms.
	355	*/
	356	STATIC int
	357	xfs_dir2_leaf_getdents(
53f82db0	358	struct xfs_da_args *args,
4a8af273 DC	359	struct dir_context *ctx,
	360	size_t bufsize)
	361	{
53f82db0	362	struct xfs_inode *dp = args->dp;
4a8af273 DC	363	struct xfs_buf bp = NULL; / data block buffer */
	364	xfs_dir2_data_hdr_t hdr; / data block header */
	365	xfs_dir2_data_entry_t dep; / data entry */
	366	xfs_dir2_data_unused_t dup; / unused entry */
4a8af273	367	char ptr = NULL; / pointer to current data */
53f82db0	368	struct xfs_da_geometry *geo = args->geo;
d205a7d0 DW	369	xfs_dablk_t rablk = 0; /* current readahead block */
	370	xfs_dir2_off_t curoff; /* current overall offset */
	371	int length; /* temporary length value */
	372	int byteoff; /* offset in current block */
	373	int lock_mode;
	374	int error = 0; /* error return value */
4a8af273 DC	375
	376	/*
	377	* If the offset is at or past the largest allowed value,
	378	* give up right away.
	379	*/
	380	if (ctx->pos >= XFS_DIR2_MAX_DATAPTR)
	381	return 0;
	382
4a8af273 DC	383	/*
	384	* Inside the loop we keep the main offset value as a byte offset
	385	* in the directory file.
	386	*/
25994053	387	curoff = xfs_dir2_dataptr_to_byte(ctx->pos);
4a8af273	388
4a8af273 DC	389	/*
	390	* Loop over directory entries until we reach the end offset.
	391	* Get more blocks and readahead as necessary.
	392	*/
	393	while (curoff < XFS_DIR2_LEAF_OFFSET) {
c8ce540d	394	uint8_t filetype;
0cb97766	395
4a8af273 DC	396	/*
	397	* If we have no buffer, or we're off the end of the
	398	* current buffer, need to get another one.
	399	*/
8f66193c	400	if (!bp \|\| ptr >= (char *)bp->b_addr + geo->blksize) {
9f541801	401	if (bp) {
d205a7d0	402	xfs_trans_brelse(args->trans, bp);
9f541801	403	bp = NULL;
9f541801	404	}
4a8af273	405
dbad7c99	406	lock_mode = xfs_ilock_data_map_shared(dp);
d205a7d0 DW	407	error = xfs_dir2_leaf_readbuf(args, bufsize, &curoff,
d205a7d0 DW	408	&rablk, &bp);
dbad7c99	409	xfs_iunlock(dp, lock_mode);
d205a7d0	410	if (error \|\| !bp)
4a8af273 DC	411	break;
4a8af273 DC	412
4a8af273 DC	413	hdr = bp->b_addr;
	414	xfs_dir3_data_check(dp, bp);
	415	/*
	416	* Find our position in the block.
	417	*/
2ca98774	418	ptr = (char *)dp->d_ops->data_entry_p(hdr);
53f82db0	419	byteoff = xfs_dir2_byte_to_off(geo, curoff);
4a8af273 DC	420	/*
	421	* Skip past the header.
	422	*/
	423	if (byteoff == 0)
1c9a5b2e	424	curoff += dp->d_ops->data_entry_offset;
4a8af273 DC	425	/*
	426	* Skip past entries until we reach our offset.
	427	*/
	428	else {
	429	while ((char )ptr - (char )hdr < byteoff) {
	430	dup = (xfs_dir2_data_unused_t *)ptr;
	431
	432	if (be16_to_cpu(dup->freetag)
	433	== XFS_DIR2_DATA_FREE_TAG) {
	434
	435	length = be16_to_cpu(dup->length);
	436	ptr += length;
	437	continue;
	438	}
	439	dep = (xfs_dir2_data_entry_t *)ptr;
	440	length =
9d23fc85	441	dp->d_ops->data_entsize(dep->namelen);
4a8af273 DC	442	ptr += length;
	443	}
	444	/*
	445	* Now set our real offset.
	446	*/
	447	curoff =
30028030 DC	448	xfs_dir2_db_off_to_byte(geo,
30028030 DC	449	xfs_dir2_byte_to_db(geo, curoff),
4a8af273	450	(char )ptr - (char )hdr);
8f66193c	451	if (ptr >= (char *)hdr + geo->blksize) {
4a8af273 DC	452	continue;
	453	}
	454	}
	455	}
	456	/*
	457	* We have a pointer to an entry.
	458	* Is it a live one?
	459	*/
	460	dup = (xfs_dir2_data_unused_t *)ptr;
	461	/*
	462	* No, it's unused, skip over it.
	463	*/
	464	if (be16_to_cpu(dup->freetag) == XFS_DIR2_DATA_FREE_TAG) {
	465	length = be16_to_cpu(dup->length);
	466	ptr += length;
	467	curoff += length;
	468	continue;
	469	}
	470
	471	dep = (xfs_dir2_data_entry_t *)ptr;
9d23fc85 DC	472	length = dp->d_ops->data_entsize(dep->namelen);
9d23fc85 DC	473	filetype = dp->d_ops->data_get_ftype(dep);
4a8af273	474
25994053	475	ctx->pos = xfs_dir2_byte_to_dataptr(curoff) & 0x7fffffff;
4a8af273	476	if (!dir_emit(ctx, (char *)dep->name, dep->namelen,
0cb97766	477	be64_to_cpu(dep->inumber),
53f82db0	478	xfs_dir3_get_dtype(dp->i_mount, filetype)))
4a8af273 DC	479	break;
	480
	481	/*
	482	* Advance to next entry in the block.
	483	*/
	484	ptr += length;
	485	curoff += length;
	486	/* bufsize may have just been a guess; don't go negative */
	487	bufsize = bufsize > length ? bufsize - length : 0;
	488	}
	489
	490	/*
	491	* All done. Set output offset value to current offset.
	492	*/
25994053	493	if (curoff > xfs_dir2_dataptr_to_byte(XFS_DIR2_MAX_DATAPTR))
4a8af273 DC	494	ctx->pos = XFS_DIR2_MAX_DATAPTR & 0x7fffffff;
4a8af273 DC	495	else
25994053	496	ctx->pos = xfs_dir2_byte_to_dataptr(curoff) & 0x7fffffff;
4a8af273	497	if (bp)
acb9553c	498	xfs_trans_brelse(args->trans, bp);
4a8af273 DC	499	return error;
	500	}
	501
	502	/*
	503	* Read a directory.
acb9553c DW	504	*
	505	* If supplied, the transaction collects locked dir buffers to avoid
	506	* nested buffer deadlocks. This function does not dirty the
	507	* transaction. The caller should ensure that the inode is locked
	508	* before calling this function.
4a8af273 DC	509	*/
	510	int
	511	xfs_readdir(
acb9553c	512	struct xfs_trans *tp,
53f82db0 DC	513	struct xfs_inode *dp,
	514	struct dir_context *ctx,
	515	size_t bufsize)
4a8af273	516	{
35f46c5f	517	struct xfs_da_args args = { NULL };
53f82db0 DC	518	int rval;
53f82db0 DC	519	int v;
4a8af273 DC	520
	521	trace_xfs_readdir(dp);
	522
	523	if (XFS_FORCED_SHUTDOWN(dp->i_mount))
2451337d	524	return -EIO;
4a8af273	525
c19b3b05	526	ASSERT(S_ISDIR(VFS_I(dp)->i_mode));
ff6d6af2	527	XFS_STATS_INC(dp->i_mount, xs_dir_getdents);
4a8af273	528
53f82db0 DC	529	args.dp = dp;
53f82db0 DC	530	args.geo = dp->i_mount->m_dir_geo;
acb9553c	531	args.trans = tp;
53f82db0	532
4a8af273	533	if (dp->i_d.di_format == XFS_DINODE_FMT_LOCAL)
53f82db0 DC	534	rval = xfs_dir2_sf_getdents(&args, ctx);
53f82db0 DC	535	else if ((rval = xfs_dir2_isblock(&args, &v)))
4a8af273 DC	536	;
4a8af273 DC	537	else if (v)
53f82db0	538	rval = xfs_dir2_block_getdents(&args, ctx);
4a8af273	539	else
53f82db0	540	rval = xfs_dir2_leaf_getdents(&args, ctx, bufsize);
40194ecc	541
4a8af273 DC	542	return rval;
4a8af273 DC	543	}