[linux.git] / fs / xfs / libxfs / xfs_attr_remote.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_shared.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_alloc.h"
#include "xfs_trans.h"
#include "xfs_inode_item.h"
#include "xfs_bmap.h"
#include "xfs_bmap_util.h"
#include "xfs_attr.h"
#include "xfs_attr_leaf.h"
#include "xfs_attr_remote.h"
#include "xfs_trans_space.h"
#include "xfs_trace.h"
#include "xfs_cksum.h"
#include "xfs_buf_item.h"
#include "xfs_error.h"

#define ATTR_RMTVALUE_MAPSIZE	1	/* # of map entries at once */

/*
 * Each contiguous block has a header, so it is not just a simple attribute
 * length to FSB conversion.
 */
int
xfs_attr3_rmt_blocks(
	struct xfs_mount *mp,
	int		attrlen)
{
	if (xfs_sb_version_hascrc(&mp->m_sb)) {
		int buflen = XFS_ATTR3_RMT_BUF_SPACE(mp, mp->m_sb.sb_blocksize);
		return (attrlen + buflen - 1) / buflen;
	}
	return XFS_B_TO_FSB(mp, attrlen);
}

/*
 * Checking of the remote attribute header is split into two parts. The verifier
 * does CRC, location and bounds checking, the unpacking function checks the
 * attribute parameters and owner.
 */
static bool
xfs_attr3_rmt_hdr_ok(
	void			*ptr,
	xfs_ino_t		ino,
	uint32_t		offset,
	uint32_t		size,
	xfs_daddr_t		bno)
{
	struct xfs_attr3_rmt_hdr *rmt = ptr;

	if (bno != be64_to_cpu(rmt->rm_blkno))
		return false;
	if (offset != be32_to_cpu(rmt->rm_offset))
		return false;
	if (size != be32_to_cpu(rmt->rm_bytes))
		return false;
	if (ino != be64_to_cpu(rmt->rm_owner))
		return false;

	/* ok */
	return true;
}

static bool
xfs_attr3_rmt_verify(
	struct xfs_mount	*mp,
	void			*ptr,
	int			fsbsize,
	xfs_daddr_t		bno)
{
	struct xfs_attr3_rmt_hdr *rmt = ptr;

	if (!xfs_sb_version_hascrc(&mp->m_sb))
		return false;
	if (rmt->rm_magic != cpu_to_be32(XFS_ATTR3_RMT_MAGIC))
		return false;
	if (!uuid_equal(&rmt->rm_uuid, &mp->m_sb.sb_uuid))
		return false;
	if (be64_to_cpu(rmt->rm_blkno) != bno)
		return false;
	if (be32_to_cpu(rmt->rm_bytes) > fsbsize - sizeof(*rmt))
		return false;
	if (be32_to_cpu(rmt->rm_offset) +
				be32_to_cpu(rmt->rm_bytes) > XATTR_SIZE_MAX)
		return false;
	if (rmt->rm_owner == 0)
		return false;

	return true;
}

static void
xfs_attr3_rmt_read_verify(
	struct xfs_buf	*bp)
{
	struct xfs_mount *mp = bp->b_target->bt_mount;
	char		*ptr;
	int		len;
	xfs_daddr_t	bno;
	int		blksize = mp->m_attr_geo->blksize;

	/* no verification of non-crc buffers */
	if (!xfs_sb_version_hascrc(&mp->m_sb))
		return;

	ptr = bp->b_addr;
	bno = bp->b_bn;
	len = BBTOB(bp->b_length);
	ASSERT(len >= blksize);

	while (len > 0) {
		if (!xfs_verify_cksum(ptr, blksize, XFS_ATTR3_RMT_CRC_OFF)) {
			xfs_buf_ioerror(bp, -EFSBADCRC);
			break;
		}
		if (!xfs_attr3_rmt_verify(mp, ptr, blksize, bno)) {
			xfs_buf_ioerror(bp, -EFSCORRUPTED);
			break;
		}
		len -= blksize;
		ptr += blksize;
		bno += BTOBB(blksize);
	}

	if (bp->b_error)
		xfs_verifier_error(bp);
	else
		ASSERT(len == 0);
}

static void
xfs_attr3_rmt_write_verify(
	struct xfs_buf	*bp)
{
	struct xfs_mount *mp = bp->b_target->bt_mount;
	struct xfs_buf_log_item	*bip = bp->b_fspriv;
	char		*ptr;
	int		len;
	xfs_daddr_t	bno;
	int		blksize = mp->m_attr_geo->blksize;

	/* no verification of non-crc buffers */
	if (!xfs_sb_version_hascrc(&mp->m_sb))
		return;

	ptr = bp->b_addr;
	bno = bp->b_bn;
	len = BBTOB(bp->b_length);
	ASSERT(len >= blksize);

	while (len > 0) {
		if (!xfs_attr3_rmt_verify(mp, ptr, blksize, bno)) {
			xfs_buf_ioerror(bp, -EFSCORRUPTED);
			xfs_verifier_error(bp);
			return;
		}
		if (bip) {
			struct xfs_attr3_rmt_hdr *rmt;

			rmt = (struct xfs_attr3_rmt_hdr *)ptr;
			rmt->rm_lsn = cpu_to_be64(bip->bli_item.li_lsn);
		}
		xfs_update_cksum(ptr, blksize, XFS_ATTR3_RMT_CRC_OFF);

		len -= blksize;
		ptr += blksize;
		bno += BTOBB(blksize);
	}
	ASSERT(len == 0);
}

const struct xfs_buf_ops xfs_attr3_rmt_buf_ops = {
	.verify_read = xfs_attr3_rmt_read_verify,
	.verify_write = xfs_attr3_rmt_write_verify,
};

STATIC int
xfs_attr3_rmt_hdr_set(
	struct xfs_mount	*mp,
	void			*ptr,
	xfs_ino_t		ino,
	uint32_t		offset,
	uint32_t		size,
	xfs_daddr_t		bno)
{
	struct xfs_attr3_rmt_hdr *rmt = ptr;

	if (!xfs_sb_version_hascrc(&mp->m_sb))
		return 0;

	rmt->rm_magic = cpu_to_be32(XFS_ATTR3_RMT_MAGIC);
	rmt->rm_offset = cpu_to_be32(offset);
	rmt->rm_bytes = cpu_to_be32(size);
	uuid_copy(&rmt->rm_uuid, &mp->m_sb.sb_uuid);
	rmt->rm_owner = cpu_to_be64(ino);
	rmt->rm_blkno = cpu_to_be64(bno);

	return sizeof(struct xfs_attr3_rmt_hdr);
}

/*
 * Helper functions to copy attribute data in and out of the one disk extents
 */
STATIC int
xfs_attr_rmtval_copyout(
	struct xfs_mount *mp,
	struct xfs_buf	*bp,
	xfs_ino_t	ino,
	int		*offset,
	int		*valuelen,
	__uint8_t	**dst)
{
	char		*src = bp->b_addr;
	xfs_daddr_t	bno = bp->b_bn;
	int		len = BBTOB(bp->b_length);
	int		blksize = mp->m_attr_geo->blksize;

	ASSERT(len >= blksize);

	while (len > 0 && *valuelen > 0) {
		int hdr_size = 0;
		int byte_cnt = XFS_ATTR3_RMT_BUF_SPACE(mp, blksize);

		byte_cnt = min(*valuelen, byte_cnt);

		if (xfs_sb_version_hascrc(&mp->m_sb)) {
			if (!xfs_attr3_rmt_hdr_ok(src, ino, *offset,
						  byte_cnt, bno)) {
				xfs_alert(mp,
"remote attribute header mismatch bno/off/len/owner (0x%llx/0x%x/Ox%x/0x%llx)",
					bno, *offset, byte_cnt, ino);
				return -EFSCORRUPTED;
			}
			hdr_size = sizeof(struct xfs_attr3_rmt_hdr);
		}

		memcpy(*dst, src + hdr_size, byte_cnt);

		/* roll buffer forwards */
		len -= blksize;
		src += blksize;
		bno += BTOBB(blksize);

		/* roll attribute data forwards */
		*valuelen -= byte_cnt;
		*dst += byte_cnt;
		*offset += byte_cnt;
	}
	return 0;
}

STATIC void
xfs_attr_rmtval_copyin(
	struct xfs_mount *mp,
	struct xfs_buf	*bp,
	xfs_ino_t	ino,
	int		*offset,
	int		*valuelen,
	__uint8_t	**src)
{
	char		*dst = bp->b_addr;
	xfs_daddr_t	bno = bp->b_bn;
	int		len = BBTOB(bp->b_length);
	int		blksize = mp->m_attr_geo->blksize;

	ASSERT(len >= blksize);

	while (len > 0 && *valuelen > 0) {
		int hdr_size;
		int byte_cnt = XFS_ATTR3_RMT_BUF_SPACE(mp, blksize);

		byte_cnt = min(*valuelen, byte_cnt);
		hdr_size = xfs_attr3_rmt_hdr_set(mp, dst, ino, *offset,
						 byte_cnt, bno);

		memcpy(dst + hdr_size, *src, byte_cnt);

		/*
		 * If this is the last block, zero the remainder of it.
		 * Check that we are actually the last block, too.
		 */
		if (byte_cnt + hdr_size < blksize) {
			ASSERT(*valuelen - byte_cnt == 0);
			ASSERT(len == blksize);
			memset(dst + hdr_size + byte_cnt, 0,
					blksize - hdr_size - byte_cnt);
		}

		/* roll buffer forwards */
		len -= blksize;
		dst += blksize;
		bno += BTOBB(blksize);

		/* roll attribute data forwards */
		*valuelen -= byte_cnt;
		*src += byte_cnt;
		*offset += byte_cnt;
	}
}

/*
 * Read the value associated with an attribute from the out-of-line buffer
 * that we stored it in.
 */
int
xfs_attr_rmtval_get(
	struct xfs_da_args	*args)
{
	struct xfs_bmbt_irec	map[ATTR_RMTVALUE_MAPSIZE];
	struct xfs_mount	*mp = args->dp->i_mount;
	struct xfs_buf		*bp;
	xfs_dablk_t		lblkno = args->rmtblkno;
	__uint8_t		*dst = args->value;
	int			valuelen;
	int			nmap;
	int			error;
	int			blkcnt = args->rmtblkcnt;
	int			i;
	int			offset = 0;

	trace_xfs_attr_rmtval_get(args);

	ASSERT(!(args->flags & ATTR_KERNOVAL));
	ASSERT(args->rmtvaluelen == args->valuelen);

	valuelen = args->rmtvaluelen;
	while (valuelen > 0) {
		nmap = ATTR_RMTVALUE_MAPSIZE;
		error = xfs_bmapi_read(args->dp, (xfs_fileoff_t)lblkno,
				       blkcnt, map, &nmap,
				       XFS_BMAPI_ATTRFORK);
		if (error)
			return error;
		ASSERT(nmap >= 1);

		for (i = 0; (i < nmap) && (valuelen > 0); i++) {
			xfs_daddr_t	dblkno;
			int		dblkcnt;

			ASSERT((map[i].br_startblock != DELAYSTARTBLOCK) &&
			       (map[i].br_startblock != HOLESTARTBLOCK));
			dblkno = XFS_FSB_TO_DADDR(mp, map[i].br_startblock);
			dblkcnt = XFS_FSB_TO_BB(mp, map[i].br_blockcount);
			error = xfs_trans_read_buf(mp, NULL, mp->m_ddev_targp,
						   dblkno, dblkcnt, 0, &bp,
						   &xfs_attr3_rmt_buf_ops);
			if (error)
				return error;

			error = xfs_attr_rmtval_copyout(mp, bp, args->dp->i_ino,
							&offset, &valuelen,
							&dst);
			xfs_buf_relse(bp);
			if (error)
				return error;

			/* roll attribute extent map forwards */
			lblkno += map[i].br_blockcount;
			blkcnt -= map[i].br_blockcount;
		}
	}
	ASSERT(valuelen == 0);
	return 0;
}

/*
 * Write the value associated with an attribute into the out-of-line buffer
 * that we have defined for it.
 */
int
xfs_attr_rmtval_set(
	struct xfs_da_args	*args)
{
	struct xfs_inode	*dp = args->dp;
	struct xfs_mount	*mp = dp->i_mount;
	struct xfs_bmbt_irec	map;
	xfs_dablk_t		lblkno;
	xfs_fileoff_t		lfileoff = 0;
	__uint8_t		*src = args->value;
	int			blkcnt;
	int			valuelen;
	int			nmap;
	int			error;
	int			offset = 0;

	trace_xfs_attr_rmtval_set(args);

	/*
	 * Find a "hole" in the attribute address space large enough for
	 * us to drop the new attribute's value into. Because CRC enable
	 * attributes have headers, we can't just do a straight byte to FSB
	 * conversion and have to take the header space into account.
	 */
	blkcnt = xfs_attr3_rmt_blocks(mp, args->rmtvaluelen);
	error = xfs_bmap_first_unused(args->trans, args->dp, blkcnt, &lfileoff,
						   XFS_ATTR_FORK);
	if (error)
		return error;

	args->rmtblkno = lblkno = (xfs_dablk_t)lfileoff;
	args->rmtblkcnt = blkcnt;

	/*
	 * Roll through the "value", allocating blocks on disk as required.
	 */
	while (blkcnt > 0) {
		int	committed;

		/*
		 * Allocate a single extent, up to the size of the value.
		 */
		xfs_bmap_init(args->flist, args->firstblock);
		nmap = 1;
		error = xfs_bmapi_write(args->trans, dp, (xfs_fileoff_t)lblkno,
				  blkcnt,
				  XFS_BMAPI_ATTRFORK | XFS_BMAPI_METADATA,
				  args->firstblock, args->total, &map, &nmap,
				  args->flist);
		if (!error) {
			error = xfs_bmap_finish(&args->trans, args->flist,
						&committed);
		}
		if (error) {
			ASSERT(committed);
			args->trans = NULL;
			xfs_bmap_cancel(args->flist);
			return error;
		}

		/*
		 * bmap_finish() may have committed the last trans and started
		 * a new one.  We need the inode to be in all transactions.
		 */
		if (committed)
			xfs_trans_ijoin(args->trans, dp, 0);

		ASSERT(nmap == 1);
		ASSERT((map.br_startblock != DELAYSTARTBLOCK) &&
		       (map.br_startblock != HOLESTARTBLOCK));
		lblkno += map.br_blockcount;
		blkcnt -= map.br_blockcount;

		/*
		 * Start the next trans in the chain.
		 */
		error = xfs_trans_roll(&args->trans, dp);
		if (error)
			return error;
	}

	/*
	 * Roll through the "value", copying the attribute value to the
	 * already-allocated blocks.  Blocks are written synchronously
	 * so that we can know they are all on disk before we turn off
	 * the INCOMPLETE flag.
	 */
	lblkno = args->rmtblkno;
	blkcnt = args->rmtblkcnt;
	valuelen = args->rmtvaluelen;
	while (valuelen > 0) {
		struct xfs_buf	*bp;
		xfs_daddr_t	dblkno;
		int		dblkcnt;

		ASSERT(blkcnt > 0);

		xfs_bmap_init(args->flist, args->firstblock);
		nmap = 1;
		error = xfs_bmapi_read(dp, (xfs_fileoff_t)lblkno,
				       blkcnt, &map, &nmap,
				       XFS_BMAPI_ATTRFORK);
		if (error)
			return error;
		ASSERT(nmap == 1);
		ASSERT((map.br_startblock != DELAYSTARTBLOCK) &&
		       (map.br_startblock != HOLESTARTBLOCK));

		dblkno = XFS_FSB_TO_DADDR(mp, map.br_startblock),
		dblkcnt = XFS_FSB_TO_BB(mp, map.br_blockcount);

		bp = xfs_buf_get(mp->m_ddev_targp, dblkno, dblkcnt, 0);
		if (!bp)
			return -ENOMEM;
		bp->b_ops = &xfs_attr3_rmt_buf_ops;

		xfs_attr_rmtval_copyin(mp, bp, args->dp->i_ino, &offset,
				       &valuelen, &src);

		error = xfs_bwrite(bp);	/* GROT: NOTE: synchronous write */
		xfs_buf_relse(bp);
		if (error)
			return error;


		/* roll attribute extent map forwards */
		lblkno += map.br_blockcount;
		blkcnt -= map.br_blockcount;
	}
	ASSERT(valuelen == 0);
	return 0;
}

/*
 * Remove the value associated with an attribute by deleting the
 * out-of-line buffer that it is stored on.
 */
int
xfs_attr_rmtval_remove(
	struct xfs_da_args	*args)
{
	struct xfs_mount	*mp = args->dp->i_mount;
	xfs_dablk_t		lblkno;
	int			blkcnt;
	int			error;
	int			done;

	trace_xfs_attr_rmtval_remove(args);

	/*
	 * Roll through the "value", invalidating the attribute value's blocks.
	 */
	lblkno = args->rmtblkno;
	blkcnt = args->rmtblkcnt;
	while (blkcnt > 0) {
		struct xfs_bmbt_irec	map;
		struct xfs_buf		*bp;
		xfs_daddr_t		dblkno;
		int			dblkcnt;
		int			nmap;

		/*
		 * Try to remember where we decided to put the value.
		 */
		nmap = 1;
		error = xfs_bmapi_read(args->dp, (xfs_fileoff_t)lblkno,
				       blkcnt, &map, &nmap, XFS_BMAPI_ATTRFORK);
		if (error)
			return error;
		ASSERT(nmap == 1);
		ASSERT((map.br_startblock != DELAYSTARTBLOCK) &&
		       (map.br_startblock != HOLESTARTBLOCK));

		dblkno = XFS_FSB_TO_DADDR(mp, map.br_startblock),
		dblkcnt = XFS_FSB_TO_BB(mp, map.br_blockcount);

		/*
		 * If the "remote" value is in the cache, remove it.
		 */
		bp = xfs_incore(mp->m_ddev_targp, dblkno, dblkcnt, XBF_TRYLOCK);
		if (bp) {
			xfs_buf_stale(bp);
			xfs_buf_relse(bp);
			bp = NULL;
		}

		lblkno += map.br_blockcount;
		blkcnt -= map.br_blockcount;
	}

	/*
	 * Keep de-allocating extents until the remote-value region is gone.
	 */
	lblkno = args->rmtblkno;
	blkcnt = args->rmtblkcnt;
	done = 0;
	while (!done) {
		int committed;

		xfs_bmap_init(args->flist, args->firstblock);
		error = xfs_bunmapi(args->trans, args->dp, lblkno, blkcnt,
				    XFS_BMAPI_ATTRFORK | XFS_BMAPI_METADATA,
				    1, args->firstblock, args->flist,
				    &done);
		if (!error) {
			error = xfs_bmap_finish(&args->trans, args->flist,
						&committed);
		}
		if (error) {
			ASSERT(committed);
			args->trans = NULL;
			xfs_bmap_cancel(args->flist);
			return error;
		}

		/*
		 * bmap_finish() may have committed the last trans and started
		 * a new one.  We need the inode to be in all transactions.
		 */
		if (committed)
			xfs_trans_ijoin(args->trans, args->dp, 0);

		/*
		 * Close out trans and start the next one in the chain.
		 */
		error = xfs_trans_roll(&args->trans, args->dp);
		if (error)
			return error;
	}
	return 0;
}
Commit	Line	Data
95920cd6 DC	1	/*
95920cd6 DC	2	* Copyright (c) 2000-2005 Silicon Graphics, Inc.
d2e448d5	3	* Copyright (c) 2013 Red Hat, Inc.
95920cd6 DC	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"
632b89e8	21	#include "xfs_shared.h"
a4fbe6ab	22	#include "xfs_format.h"
239880ef DC	23	#include "xfs_log_format.h"
239880ef DC	24	#include "xfs_trans_resv.h"
95920cd6	25	#include "xfs_bit.h"
95920cd6	26	#include "xfs_mount.h"
57062787	27	#include "xfs_da_format.h"
95920cd6	28	#include "xfs_da_btree.h"
95920cd6 DC	29	#include "xfs_inode.h"
95920cd6 DC	30	#include "xfs_alloc.h"
239880ef	31	#include "xfs_trans.h"
95920cd6 DC	32	#include "xfs_inode_item.h"
95920cd6 DC	33	#include "xfs_bmap.h"
68988114	34	#include "xfs_bmap_util.h"
95920cd6 DC	35	#include "xfs_attr.h"
	36	#include "xfs_attr_leaf.h"
	37	#include "xfs_attr_remote.h"
	38	#include "xfs_trans_space.h"
	39	#include "xfs_trace.h"
d2e448d5 DC	40	#include "xfs_cksum.h"
d2e448d5 DC	41	#include "xfs_buf_item.h"
a4fbe6ab	42	#include "xfs_error.h"
95920cd6 DC	43
	44	#define ATTR_RMTVALUE_MAPSIZE 1 /* # of map entries at once */
	45
d2e448d5 DC	46	/*
	47	* Each contiguous block has a header, so it is not just a simple attribute
	48	* length to FSB conversion.
	49	*/
7bc0dc27	50	int
d2e448d5 DC	51	xfs_attr3_rmt_blocks(
	52	struct xfs_mount *mp,
	53	int attrlen)
	54	{
551b382f DC	55	if (xfs_sb_version_hascrc(&mp->m_sb)) {
	56	int buflen = XFS_ATTR3_RMT_BUF_SPACE(mp, mp->m_sb.sb_blocksize);
	57	return (attrlen + buflen - 1) / buflen;
	58	}
	59	return XFS_B_TO_FSB(mp, attrlen);
d2e448d5 DC	60	}
d2e448d5 DC	61
7bc0dc27 DC	62	/*
	63	* Checking of the remote attribute header is split into two parts. The verifier
	64	* does CRC, location and bounds checking, the unpacking function checks the
	65	* attribute parameters and owner.
	66	*/
	67	static bool
	68	xfs_attr3_rmt_hdr_ok(
7bc0dc27 DC	69	void *ptr,
	70	xfs_ino_t ino,
	71	uint32_t offset,
	72	uint32_t size,
	73	xfs_daddr_t bno)
	74	{
	75	struct xfs_attr3_rmt_hdr *rmt = ptr;
	76
	77	if (bno != be64_to_cpu(rmt->rm_blkno))
	78	return false;
	79	if (offset != be32_to_cpu(rmt->rm_offset))
	80	return false;
	81	if (size != be32_to_cpu(rmt->rm_bytes))
	82	return false;
	83	if (ino != be64_to_cpu(rmt->rm_owner))
	84	return false;
	85
	86	/* ok */
	87	return true;
	88	}
	89
d2e448d5 DC	90	static bool
d2e448d5 DC	91	xfs_attr3_rmt_verify(
7bc0dc27 DC	92	struct xfs_mount *mp,
	93	void *ptr,
	94	int fsbsize,
	95	xfs_daddr_t bno)
d2e448d5	96	{
7bc0dc27	97	struct xfs_attr3_rmt_hdr *rmt = ptr;
d2e448d5 DC	98
	99	if (!xfs_sb_version_hascrc(&mp->m_sb))
	100	return false;
	101	if (rmt->rm_magic != cpu_to_be32(XFS_ATTR3_RMT_MAGIC))
	102	return false;
	103	if (!uuid_equal(&rmt->rm_uuid, &mp->m_sb.sb_uuid))
	104	return false;
7bc0dc27 DC	105	if (be64_to_cpu(rmt->rm_blkno) != bno)
	106	return false;
	107	if (be32_to_cpu(rmt->rm_bytes) > fsbsize - sizeof(*rmt))
d2e448d5 DC	108	return false;
d2e448d5 DC	109	if (be32_to_cpu(rmt->rm_offset) +
bba719b5	110	be32_to_cpu(rmt->rm_bytes) > XATTR_SIZE_MAX)
d2e448d5 DC	111	return false;
	112	if (rmt->rm_owner == 0)
	113	return false;
	114
	115	return true;
	116	}
	117
	118	static void
	119	xfs_attr3_rmt_read_verify(
	120	struct xfs_buf *bp)
	121	{
	122	struct xfs_mount *mp = bp->b_target->bt_mount;
7bc0dc27 DC	123	char *ptr;
7bc0dc27 DC	124	int len;
7bc0dc27	125	xfs_daddr_t bno;
c2c4c477	126	int blksize = mp->m_attr_geo->blksize;
d2e448d5 DC	127
	128	/* no verification of non-crc buffers */
	129	if (!xfs_sb_version_hascrc(&mp->m_sb))
	130	return;
	131
7bc0dc27 DC	132	ptr = bp->b_addr;
	133	bno = bp->b_bn;
	134	len = BBTOB(bp->b_length);
c2c4c477	135	ASSERT(len >= blksize);
7bc0dc27 DC	136
7bc0dc27 DC	137	while (len > 0) {
c2c4c477	138	if (!xfs_verify_cksum(ptr, blksize, XFS_ATTR3_RMT_CRC_OFF)) {
2451337d	139	xfs_buf_ioerror(bp, -EFSBADCRC);
7bc0dc27 DC	140	break;
7bc0dc27 DC	141	}
c2c4c477	142	if (!xfs_attr3_rmt_verify(mp, ptr, blksize, bno)) {
2451337d	143	xfs_buf_ioerror(bp, -EFSCORRUPTED);
7bc0dc27 DC	144	break;
7bc0dc27 DC	145	}
c2c4c477 DC	146	len -= blksize;
	147	ptr += blksize;
	148	bno += BTOBB(blksize);
7bc0dc27 DC	149	}
7bc0dc27 DC	150
ce5028cf ES	151	if (bp->b_error)
	152	xfs_verifier_error(bp);
	153	else
7bc0dc27	154	ASSERT(len == 0);
d2e448d5 DC	155	}
	156
	157	static void
	158	xfs_attr3_rmt_write_verify(
	159	struct xfs_buf *bp)
	160	{
	161	struct xfs_mount *mp = bp->b_target->bt_mount;
	162	struct xfs_buf_log_item *bip = bp->b_fspriv;
7bc0dc27 DC	163	char *ptr;
	164	int len;
	165	xfs_daddr_t bno;
c2c4c477	166	int blksize = mp->m_attr_geo->blksize;
d2e448d5 DC	167
	168	/* no verification of non-crc buffers */
	169	if (!xfs_sb_version_hascrc(&mp->m_sb))
	170	return;
	171
7bc0dc27 DC	172	ptr = bp->b_addr;
	173	bno = bp->b_bn;
	174	len = BBTOB(bp->b_length);
c2c4c477	175	ASSERT(len >= blksize);
7bc0dc27 DC	176
7bc0dc27 DC	177	while (len > 0) {
c2c4c477	178	if (!xfs_attr3_rmt_verify(mp, ptr, blksize, bno)) {
2451337d	179	xfs_buf_ioerror(bp, -EFSCORRUPTED);
ce5028cf	180	xfs_verifier_error(bp);
7bc0dc27 DC	181	return;
	182	}
	183	if (bip) {
	184	struct xfs_attr3_rmt_hdr *rmt;
d2e448d5	185
7bc0dc27 DC	186	rmt = (struct xfs_attr3_rmt_hdr *)ptr;
	187	rmt->rm_lsn = cpu_to_be64(bip->bli_item.li_lsn);
	188	}
c2c4c477	189	xfs_update_cksum(ptr, blksize, XFS_ATTR3_RMT_CRC_OFF);
7bc0dc27	190
c2c4c477 DC	191	len -= blksize;
	192	ptr += blksize;
	193	bno += BTOBB(blksize);
d2e448d5	194	}
7bc0dc27	195	ASSERT(len == 0);
d2e448d5 DC	196	}
	197
	198	const struct xfs_buf_ops xfs_attr3_rmt_buf_ops = {
	199	.verify_read = xfs_attr3_rmt_read_verify,
	200	.verify_write = xfs_attr3_rmt_write_verify,
	201	};
	202
7bc0dc27	203	STATIC int
d2e448d5 DC	204	xfs_attr3_rmt_hdr_set(
d2e448d5 DC	205	struct xfs_mount *mp,
7bc0dc27	206	void *ptr,
d2e448d5 DC	207	xfs_ino_t ino,
	208	uint32_t offset,
	209	uint32_t size,
7bc0dc27	210	xfs_daddr_t bno)
d2e448d5	211	{
7bc0dc27	212	struct xfs_attr3_rmt_hdr *rmt = ptr;
d2e448d5 DC	213
	214	if (!xfs_sb_version_hascrc(&mp->m_sb))
	215	return 0;
	216
	217	rmt->rm_magic = cpu_to_be32(XFS_ATTR3_RMT_MAGIC);
	218	rmt->rm_offset = cpu_to_be32(offset);
	219	rmt->rm_bytes = cpu_to_be32(size);
	220	uuid_copy(&rmt->rm_uuid, &mp->m_sb.sb_uuid);
	221	rmt->rm_owner = cpu_to_be64(ino);
7bc0dc27	222	rmt->rm_blkno = cpu_to_be64(bno);
d2e448d5 DC	223
	224	return sizeof(struct xfs_attr3_rmt_hdr);
	225	}
	226
	227	/*
7bc0dc27	228	* Helper functions to copy attribute data in and out of the one disk extents
d2e448d5	229	*/
7bc0dc27 DC	230	STATIC int
	231	xfs_attr_rmtval_copyout(
	232	struct xfs_mount *mp,
	233	struct xfs_buf *bp,
	234	xfs_ino_t ino,
	235	int *offset,
	236	int *valuelen,
836a94ad	237	__uint8_t **dst)
d2e448d5	238	{
7bc0dc27 DC	239	char *src = bp->b_addr;
	240	xfs_daddr_t bno = bp->b_bn;
	241	int len = BBTOB(bp->b_length);
c2c4c477	242	int blksize = mp->m_attr_geo->blksize;
d2e448d5	243
c2c4c477	244	ASSERT(len >= blksize);
d2e448d5	245
7bc0dc27 DC	246	while (len > 0 && *valuelen > 0) {
7bc0dc27 DC	247	int hdr_size = 0;
c2c4c477	248	int byte_cnt = XFS_ATTR3_RMT_BUF_SPACE(mp, blksize);
7bc0dc27	249
c5c249b4	250	byte_cnt = min(*valuelen, byte_cnt);
7bc0dc27 DC	251
7bc0dc27 DC	252	if (xfs_sb_version_hascrc(&mp->m_sb)) {
6d0081a3	253	if (!xfs_attr3_rmt_hdr_ok(src, ino, *offset,
7bc0dc27 DC	254	byte_cnt, bno)) {
	255	xfs_alert(mp,
	256	"remote attribute header mismatch bno/off/len/owner (0x%llx/0x%x/Ox%x/0x%llx)",
	257	bno, *offset, byte_cnt, ino);
2451337d	258	return -EFSCORRUPTED;
7bc0dc27 DC	259	}
	260	hdr_size = sizeof(struct xfs_attr3_rmt_hdr);
	261	}
	262
	263	memcpy(*dst, src + hdr_size, byte_cnt);
	264
	265	/* roll buffer forwards */
c2c4c477 DC	266	len -= blksize;
	267	src += blksize;
	268	bno += BTOBB(blksize);
7bc0dc27 DC	269
	270	/* roll attribute data forwards */
	271	*valuelen -= byte_cnt;
	272	*dst += byte_cnt;
	273	*offset += byte_cnt;
	274	}
	275	return 0;
	276	}
	277
	278	STATIC void
	279	xfs_attr_rmtval_copyin(
	280	struct xfs_mount *mp,
	281	struct xfs_buf *bp,
	282	xfs_ino_t ino,
	283	int *offset,
	284	int *valuelen,
836a94ad	285	__uint8_t **src)
7bc0dc27 DC	286	{
	287	char *dst = bp->b_addr;
	288	xfs_daddr_t bno = bp->b_bn;
	289	int len = BBTOB(bp->b_length);
c2c4c477	290	int blksize = mp->m_attr_geo->blksize;
7bc0dc27	291
c2c4c477	292	ASSERT(len >= blksize);
7bc0dc27 DC	293
	294	while (len > 0 && *valuelen > 0) {
	295	int hdr_size;
c2c4c477	296	int byte_cnt = XFS_ATTR3_RMT_BUF_SPACE(mp, blksize);
7bc0dc27 DC	297
	298	byte_cnt = min(*valuelen, byte_cnt);
	299	hdr_size = xfs_attr3_rmt_hdr_set(mp, dst, ino, *offset,
	300	byte_cnt, bno);
	301
	302	memcpy(dst + hdr_size, *src, byte_cnt);
	303
	304	/*
	305	* If this is the last block, zero the remainder of it.
	306	* Check that we are actually the last block, too.
	307	*/
c2c4c477	308	if (byte_cnt + hdr_size < blksize) {
7bc0dc27	309	ASSERT(*valuelen - byte_cnt == 0);
c2c4c477	310	ASSERT(len == blksize);
7bc0dc27	311	memset(dst + hdr_size + byte_cnt, 0,
c2c4c477	312	blksize - hdr_size - byte_cnt);
7bc0dc27 DC	313	}
	314
	315	/* roll buffer forwards */
c2c4c477 DC	316	len -= blksize;
	317	dst += blksize;
	318	bno += BTOBB(blksize);
7bc0dc27 DC	319
	320	/* roll attribute data forwards */
	321	*valuelen -= byte_cnt;
	322	*src += byte_cnt;
	323	*offset += byte_cnt;
	324	}
d2e448d5 DC	325	}
d2e448d5 DC	326
95920cd6 DC	327	/*
	328	* Read the value associated with an attribute from the out-of-line buffer
	329	* that we stored it in.
	330	*/
	331	int
d2e448d5 DC	332	xfs_attr_rmtval_get(
d2e448d5 DC	333	struct xfs_da_args *args)
95920cd6	334	{
d2e448d5 DC	335	struct xfs_bmbt_irec map[ATTR_RMTVALUE_MAPSIZE];
	336	struct xfs_mount *mp = args->dp->i_mount;
	337	struct xfs_buf *bp;
d2e448d5	338	xfs_dablk_t lblkno = args->rmtblkno;
836a94ad	339	__uint8_t *dst = args->value;
8275cdd0	340	int valuelen;
d2e448d5 DC	341	int nmap;
d2e448d5 DC	342	int error;
7bc0dc27	343	int blkcnt = args->rmtblkcnt;
d2e448d5 DC	344	int i;
d2e448d5 DC	345	int offset = 0;
95920cd6 DC	346
	347	trace_xfs_attr_rmtval_get(args);
	348
	349	ASSERT(!(args->flags & ATTR_KERNOVAL));
8275cdd0	350	ASSERT(args->rmtvaluelen == args->valuelen);
95920cd6	351
8275cdd0	352	valuelen = args->rmtvaluelen;
95920cd6 DC	353	while (valuelen > 0) {
	354	nmap = ATTR_RMTVALUE_MAPSIZE;
	355	error = xfs_bmapi_read(args->dp, (xfs_fileoff_t)lblkno,
551b382f	356	blkcnt, map, &nmap,
95920cd6 DC	357	XFS_BMAPI_ATTRFORK);
95920cd6 DC	358	if (error)
d2e448d5	359	return error;
95920cd6 DC	360	ASSERT(nmap >= 1);
	361
	362	for (i = 0; (i < nmap) && (valuelen > 0); i++) {
7bc0dc27 DC	363	xfs_daddr_t dblkno;
7bc0dc27 DC	364	int dblkcnt;
d2e448d5	365
95920cd6 DC	366	ASSERT((map[i].br_startblock != DELAYSTARTBLOCK) &&
	367	(map[i].br_startblock != HOLESTARTBLOCK));
	368	dblkno = XFS_FSB_TO_DADDR(mp, map[i].br_startblock);
7bc0dc27	369	dblkcnt = XFS_FSB_TO_BB(mp, map[i].br_blockcount);
95920cd6	370	error = xfs_trans_read_buf(mp, NULL, mp->m_ddev_targp,
7bc0dc27	371	dblkno, dblkcnt, 0, &bp,
d2e448d5	372	&xfs_attr3_rmt_buf_ops);
95920cd6	373	if (error)
d2e448d5 DC	374	return error;
d2e448d5 DC	375
7bc0dc27 DC	376	error = xfs_attr_rmtval_copyout(mp, bp, args->dp->i_ino,
	377	&offset, &valuelen,
	378	&dst);
95920cd6	379	xfs_buf_relse(bp);
7bc0dc27 DC	380	if (error)
7bc0dc27 DC	381	return error;
d2e448d5	382
7bc0dc27	383	/* roll attribute extent map forwards */
95920cd6	384	lblkno += map[i].br_blockcount;
7bc0dc27	385	blkcnt -= map[i].br_blockcount;
95920cd6 DC	386	}
	387	}
	388	ASSERT(valuelen == 0);
d2e448d5	389	return 0;
95920cd6 DC	390	}
	391
	392	/*
	393	* Write the value associated with an attribute into the out-of-line buffer
	394	* that we have defined for it.
	395	*/
	396	int
d2e448d5 DC	397	xfs_attr_rmtval_set(
d2e448d5 DC	398	struct xfs_da_args *args)
95920cd6	399	{
d2e448d5 DC	400	struct xfs_inode *dp = args->dp;
	401	struct xfs_mount *mp = dp->i_mount;
	402	struct xfs_bmbt_irec map;
d2e448d5 DC	403	xfs_dablk_t lblkno;
d2e448d5 DC	404	xfs_fileoff_t lfileoff = 0;
836a94ad	405	__uint8_t *src = args->value;
d2e448d5 DC	406	int blkcnt;
	407	int valuelen;
	408	int nmap;
	409	int error;
d2e448d5	410	int offset = 0;
95920cd6 DC	411
	412	trace_xfs_attr_rmtval_set(args);
	413
95920cd6 DC	414	/*
95920cd6 DC	415	* Find a "hole" in the attribute address space large enough for
d2e448d5 DC	416	* us to drop the new attribute's value into. Because CRC enable
d2e448d5 DC	417	* attributes have headers, we can't just do a straight byte to FSB
7bc0dc27	418	* conversion and have to take the header space into account.
95920cd6	419	*/
8275cdd0	420	blkcnt = xfs_attr3_rmt_blocks(mp, args->rmtvaluelen);
95920cd6 DC	421	error = xfs_bmap_first_unused(args->trans, args->dp, blkcnt, &lfileoff,
95920cd6 DC	422	XFS_ATTR_FORK);
d2e448d5 DC	423	if (error)
	424	return error;
	425
95920cd6 DC	426	args->rmtblkno = lblkno = (xfs_dablk_t)lfileoff;
	427	args->rmtblkcnt = blkcnt;
	428
	429	/*
	430	* Roll through the "value", allocating blocks on disk as required.
	431	*/
	432	while (blkcnt > 0) {
d2e448d5 DC	433	int committed;
d2e448d5 DC	434
95920cd6 DC	435	/*
	436	* Allocate a single extent, up to the size of the value.
	437	*/
	438	xfs_bmap_init(args->flist, args->firstblock);
	439	nmap = 1;
	440	error = xfs_bmapi_write(args->trans, dp, (xfs_fileoff_t)lblkno,
	441	blkcnt,
	442	XFS_BMAPI_ATTRFORK \| XFS_BMAPI_METADATA,
	443	args->firstblock, args->total, &map, &nmap,
	444	args->flist);
	445	if (!error) {
	446	error = xfs_bmap_finish(&args->trans, args->flist,
	447	&committed);
	448	}
	449	if (error) {
	450	ASSERT(committed);
	451	args->trans = NULL;
	452	xfs_bmap_cancel(args->flist);
d99831ff	453	return error;
95920cd6 DC	454	}
	455
	456	/*
	457	* bmap_finish() may have committed the last trans and started
	458	* a new one. We need the inode to be in all transactions.
	459	*/
	460	if (committed)
	461	xfs_trans_ijoin(args->trans, dp, 0);
	462
	463	ASSERT(nmap == 1);
	464	ASSERT((map.br_startblock != DELAYSTARTBLOCK) &&
	465	(map.br_startblock != HOLESTARTBLOCK));
	466	lblkno += map.br_blockcount;
	467	blkcnt -= map.br_blockcount;
	468
	469	/*
	470	* Start the next trans in the chain.
	471	*/
	472	error = xfs_trans_roll(&args->trans, dp);
	473	if (error)
d99831ff	474	return error;
95920cd6 DC	475	}
	476
	477	/*
	478	* Roll through the "value", copying the attribute value to the
	479	* already-allocated blocks. Blocks are written synchronously
	480	* so that we can know they are all on disk before we turn off
	481	* the INCOMPLETE flag.
	482	*/
	483	lblkno = args->rmtblkno;
26f71445	484	blkcnt = args->rmtblkcnt;
8275cdd0	485	valuelen = args->rmtvaluelen;
95920cd6	486	while (valuelen > 0) {
7bc0dc27 DC	487	struct xfs_buf *bp;
	488	xfs_daddr_t dblkno;
	489	int dblkcnt;
	490
	491	ASSERT(blkcnt > 0);
95920cd6	492
95920cd6 DC	493	xfs_bmap_init(args->flist, args->firstblock);
	494	nmap = 1;
	495	error = xfs_bmapi_read(dp, (xfs_fileoff_t)lblkno,
26f71445	496	blkcnt, &map, &nmap,
95920cd6 DC	497	XFS_BMAPI_ATTRFORK);
95920cd6 DC	498	if (error)
d99831ff	499	return error;
95920cd6 DC	500	ASSERT(nmap == 1);
	501	ASSERT((map.br_startblock != DELAYSTARTBLOCK) &&
	502	(map.br_startblock != HOLESTARTBLOCK));
	503
	504	dblkno = XFS_FSB_TO_DADDR(mp, map.br_startblock),
26f71445	505	dblkcnt = XFS_FSB_TO_BB(mp, map.br_blockcount);
95920cd6	506
26f71445	507	bp = xfs_buf_get(mp->m_ddev_targp, dblkno, dblkcnt, 0);
95920cd6	508	if (!bp)
2451337d	509	return -ENOMEM;
d2e448d5	510	bp->b_ops = &xfs_attr3_rmt_buf_ops;
26f71445	511
7bc0dc27 DC	512	xfs_attr_rmtval_copyin(mp, bp, args->dp->i_ino, &offset,
7bc0dc27 DC	513	&valuelen, &src);
95920cd6 DC	514
	515	error = xfs_bwrite(bp); /* GROT: NOTE: synchronous write */
	516	xfs_buf_relse(bp);
	517	if (error)
	518	return error;
d2e448d5	519
95920cd6	520
7bc0dc27	521	/* roll attribute extent map forwards */
95920cd6	522	lblkno += map.br_blockcount;
26f71445	523	blkcnt -= map.br_blockcount;
95920cd6 DC	524	}
95920cd6 DC	525	ASSERT(valuelen == 0);
d2e448d5	526	return 0;
95920cd6 DC	527	}
	528
	529	/*
	530	* Remove the value associated with an attribute by deleting the
	531	* out-of-line buffer that it is stored on.
	532	*/
	533	int
7bc0dc27 DC	534	xfs_attr_rmtval_remove(
7bc0dc27 DC	535	struct xfs_da_args *args)
95920cd6	536	{
7bc0dc27 DC	537	struct xfs_mount *mp = args->dp->i_mount;
	538	xfs_dablk_t lblkno;
	539	int blkcnt;
	540	int error;
	541	int done;
95920cd6 DC	542
	543	trace_xfs_attr_rmtval_remove(args);
	544
95920cd6	545	/*
58a72281	546	* Roll through the "value", invalidating the attribute value's blocks.
95920cd6 DC	547	*/
95920cd6 DC	548	lblkno = args->rmtblkno;
7bc0dc27 DC	549	blkcnt = args->rmtblkcnt;
	550	while (blkcnt > 0) {
	551	struct xfs_bmbt_irec map;
	552	struct xfs_buf *bp;
	553	xfs_daddr_t dblkno;
	554	int dblkcnt;
	555	int nmap;
58a72281	556
95920cd6 DC	557	/*
	558	* Try to remember where we decided to put the value.
	559	*/
	560	nmap = 1;
	561	error = xfs_bmapi_read(args->dp, (xfs_fileoff_t)lblkno,
58a72281	562	blkcnt, &map, &nmap, XFS_BMAPI_ATTRFORK);
95920cd6	563	if (error)
d99831ff	564	return error;
95920cd6 DC	565	ASSERT(nmap == 1);
	566	ASSERT((map.br_startblock != DELAYSTARTBLOCK) &&
	567	(map.br_startblock != HOLESTARTBLOCK));
	568
	569	dblkno = XFS_FSB_TO_DADDR(mp, map.br_startblock),
58a72281	570	dblkcnt = XFS_FSB_TO_BB(mp, map.br_blockcount);
95920cd6 DC	571
	572	/*
	573	* If the "remote" value is in the cache, remove it.
	574	*/
58a72281	575	bp = xfs_incore(mp->m_ddev_targp, dblkno, dblkcnt, XBF_TRYLOCK);
95920cd6 DC	576	if (bp) {
	577	xfs_buf_stale(bp);
	578	xfs_buf_relse(bp);
	579	bp = NULL;
	580	}
	581
95920cd6	582	lblkno += map.br_blockcount;
58a72281	583	blkcnt -= map.br_blockcount;
95920cd6 DC	584	}
	585
	586	/*
	587	* Keep de-allocating extents until the remote-value region is gone.
	588	*/
	589	lblkno = args->rmtblkno;
7bc0dc27	590	blkcnt = args->rmtblkcnt;
95920cd6 DC	591	done = 0;
95920cd6 DC	592	while (!done) {
7bc0dc27 DC	593	int committed;
7bc0dc27 DC	594
95920cd6 DC	595	xfs_bmap_init(args->flist, args->firstblock);
	596	error = xfs_bunmapi(args->trans, args->dp, lblkno, blkcnt,
	597	XFS_BMAPI_ATTRFORK \| XFS_BMAPI_METADATA,
	598	1, args->firstblock, args->flist,
	599	&done);
	600	if (!error) {
	601	error = xfs_bmap_finish(&args->trans, args->flist,
	602	&committed);
	603	}
	604	if (error) {
	605	ASSERT(committed);
	606	args->trans = NULL;
	607	xfs_bmap_cancel(args->flist);
d2e448d5	608	return error;
95920cd6 DC	609	}
	610
	611	/*
	612	* bmap_finish() may have committed the last trans and started
	613	* a new one. We need the inode to be in all transactions.
	614	*/
	615	if (committed)
	616	xfs_trans_ijoin(args->trans, args->dp, 0);
	617
	618	/*
	619	* Close out trans and start the next one in the chain.
	620	*/
	621	error = xfs_trans_roll(&args->trans, args->dp);
	622	if (error)
d99831ff	623	return error;
95920cd6	624	}
d99831ff	625	return 0;
95920cd6	626	}