[deliverable/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_sb.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_sb.h"
95920cd6	27	#include "xfs_mount.h"
57062787	28	#include "xfs_da_format.h"
95920cd6	29	#include "xfs_da_btree.h"
95920cd6 DC	30	#include "xfs_inode.h"
95920cd6 DC	31	#include "xfs_alloc.h"
239880ef	32	#include "xfs_trans.h"
95920cd6 DC	33	#include "xfs_inode_item.h"
95920cd6 DC	34	#include "xfs_bmap.h"
68988114	35	#include "xfs_bmap_util.h"
95920cd6 DC	36	#include "xfs_attr.h"
	37	#include "xfs_attr_leaf.h"
	38	#include "xfs_attr_remote.h"
	39	#include "xfs_trans_space.h"
	40	#include "xfs_trace.h"
d2e448d5 DC	41	#include "xfs_cksum.h"
d2e448d5 DC	42	#include "xfs_buf_item.h"
a4fbe6ab	43	#include "xfs_error.h"
95920cd6 DC	44
	45	#define ATTR_RMTVALUE_MAPSIZE 1 /* # of map entries at once */
	46
d2e448d5 DC	47	/*
	48	* Each contiguous block has a header, so it is not just a simple attribute
	49	* length to FSB conversion.
	50	*/
7bc0dc27	51	int
d2e448d5 DC	52	xfs_attr3_rmt_blocks(
	53	struct xfs_mount *mp,
	54	int attrlen)
	55	{
551b382f DC	56	if (xfs_sb_version_hascrc(&mp->m_sb)) {
	57	int buflen = XFS_ATTR3_RMT_BUF_SPACE(mp, mp->m_sb.sb_blocksize);
	58	return (attrlen + buflen - 1) / buflen;
	59	}
	60	return XFS_B_TO_FSB(mp, attrlen);
d2e448d5 DC	61	}
d2e448d5 DC	62
7bc0dc27 DC	63	/*
	64	* Checking of the remote attribute header is split into two parts. The verifier
	65	* does CRC, location and bounds checking, the unpacking function checks the
	66	* attribute parameters and owner.
	67	*/
	68	static bool
	69	xfs_attr3_rmt_hdr_ok(
7bc0dc27 DC	70	void *ptr,
	71	xfs_ino_t ino,
	72	uint32_t offset,
	73	uint32_t size,
	74	xfs_daddr_t bno)
	75	{
	76	struct xfs_attr3_rmt_hdr *rmt = ptr;
	77
	78	if (bno != be64_to_cpu(rmt->rm_blkno))
	79	return false;
	80	if (offset != be32_to_cpu(rmt->rm_offset))
	81	return false;
	82	if (size != be32_to_cpu(rmt->rm_bytes))
	83	return false;
	84	if (ino != be64_to_cpu(rmt->rm_owner))
	85	return false;
	86
	87	/* ok */
	88	return true;
	89	}
	90
d2e448d5 DC	91	static bool
d2e448d5 DC	92	xfs_attr3_rmt_verify(
7bc0dc27 DC	93	struct xfs_mount *mp,
	94	void *ptr,
	95	int fsbsize,
	96	xfs_daddr_t bno)
d2e448d5	97	{
7bc0dc27	98	struct xfs_attr3_rmt_hdr *rmt = ptr;
d2e448d5 DC	99
	100	if (!xfs_sb_version_hascrc(&mp->m_sb))
	101	return false;
	102	if (rmt->rm_magic != cpu_to_be32(XFS_ATTR3_RMT_MAGIC))
	103	return false;
	104	if (!uuid_equal(&rmt->rm_uuid, &mp->m_sb.sb_uuid))
	105	return false;
7bc0dc27 DC	106	if (be64_to_cpu(rmt->rm_blkno) != bno)
	107	return false;
	108	if (be32_to_cpu(rmt->rm_bytes) > fsbsize - sizeof(*rmt))
d2e448d5 DC	109	return false;
d2e448d5 DC	110	if (be32_to_cpu(rmt->rm_offset) +
bba719b5	111	be32_to_cpu(rmt->rm_bytes) > XATTR_SIZE_MAX)
d2e448d5 DC	112	return false;
	113	if (rmt->rm_owner == 0)
	114	return false;
	115
	116	return true;
	117	}
	118
	119	static void
	120	xfs_attr3_rmt_read_verify(
	121	struct xfs_buf *bp)
	122	{
	123	struct xfs_mount *mp = bp->b_target->bt_mount;
7bc0dc27 DC	124	char *ptr;
7bc0dc27 DC	125	int len;
7bc0dc27	126	xfs_daddr_t bno;
c2c4c477	127	int blksize = mp->m_attr_geo->blksize;
d2e448d5 DC	128
	129	/* no verification of non-crc buffers */
	130	if (!xfs_sb_version_hascrc(&mp->m_sb))
	131	return;
	132
7bc0dc27 DC	133	ptr = bp->b_addr;
	134	bno = bp->b_bn;
	135	len = BBTOB(bp->b_length);
c2c4c477	136	ASSERT(len >= blksize);
7bc0dc27 DC	137
7bc0dc27 DC	138	while (len > 0) {
c2c4c477	139	if (!xfs_verify_cksum(ptr, blksize, XFS_ATTR3_RMT_CRC_OFF)) {
2451337d	140	xfs_buf_ioerror(bp, -EFSBADCRC);
7bc0dc27 DC	141	break;
7bc0dc27 DC	142	}
c2c4c477	143	if (!xfs_attr3_rmt_verify(mp, ptr, blksize, bno)) {
2451337d	144	xfs_buf_ioerror(bp, -EFSCORRUPTED);
7bc0dc27 DC	145	break;
7bc0dc27 DC	146	}
c2c4c477 DC	147	len -= blksize;
	148	ptr += blksize;
	149	bno += BTOBB(blksize);
7bc0dc27 DC	150	}
7bc0dc27 DC	151
ce5028cf ES	152	if (bp->b_error)
	153	xfs_verifier_error(bp);
	154	else
7bc0dc27	155	ASSERT(len == 0);
d2e448d5 DC	156	}
	157
	158	static void
	159	xfs_attr3_rmt_write_verify(
	160	struct xfs_buf *bp)
	161	{
	162	struct xfs_mount *mp = bp->b_target->bt_mount;
	163	struct xfs_buf_log_item *bip = bp->b_fspriv;
7bc0dc27 DC	164	char *ptr;
	165	int len;
	166	xfs_daddr_t bno;
c2c4c477	167	int blksize = mp->m_attr_geo->blksize;
d2e448d5 DC	168
	169	/* no verification of non-crc buffers */
	170	if (!xfs_sb_version_hascrc(&mp->m_sb))
	171	return;
	172
7bc0dc27 DC	173	ptr = bp->b_addr;
	174	bno = bp->b_bn;
	175	len = BBTOB(bp->b_length);
c2c4c477	176	ASSERT(len >= blksize);
7bc0dc27 DC	177
7bc0dc27 DC	178	while (len > 0) {
c2c4c477	179	if (!xfs_attr3_rmt_verify(mp, ptr, blksize, bno)) {
2451337d	180	xfs_buf_ioerror(bp, -EFSCORRUPTED);
ce5028cf	181	xfs_verifier_error(bp);
7bc0dc27 DC	182	return;
	183	}
	184	if (bip) {
	185	struct xfs_attr3_rmt_hdr *rmt;
d2e448d5	186
7bc0dc27 DC	187	rmt = (struct xfs_attr3_rmt_hdr *)ptr;
	188	rmt->rm_lsn = cpu_to_be64(bip->bli_item.li_lsn);
	189	}
c2c4c477	190	xfs_update_cksum(ptr, blksize, XFS_ATTR3_RMT_CRC_OFF);
7bc0dc27	191
c2c4c477 DC	192	len -= blksize;
	193	ptr += blksize;
	194	bno += BTOBB(blksize);
d2e448d5	195	}
7bc0dc27	196	ASSERT(len == 0);
d2e448d5 DC	197	}
	198
	199	const struct xfs_buf_ops xfs_attr3_rmt_buf_ops = {
	200	.verify_read = xfs_attr3_rmt_read_verify,
	201	.verify_write = xfs_attr3_rmt_write_verify,
	202	};
	203
7bc0dc27	204	STATIC int
d2e448d5 DC	205	xfs_attr3_rmt_hdr_set(
d2e448d5 DC	206	struct xfs_mount *mp,
7bc0dc27	207	void *ptr,
d2e448d5 DC	208	xfs_ino_t ino,
	209	uint32_t offset,
	210	uint32_t size,
7bc0dc27	211	xfs_daddr_t bno)
d2e448d5	212	{
7bc0dc27	213	struct xfs_attr3_rmt_hdr *rmt = ptr;
d2e448d5 DC	214
	215	if (!xfs_sb_version_hascrc(&mp->m_sb))
	216	return 0;
	217
	218	rmt->rm_magic = cpu_to_be32(XFS_ATTR3_RMT_MAGIC);
	219	rmt->rm_offset = cpu_to_be32(offset);
	220	rmt->rm_bytes = cpu_to_be32(size);
	221	uuid_copy(&rmt->rm_uuid, &mp->m_sb.sb_uuid);
	222	rmt->rm_owner = cpu_to_be64(ino);
7bc0dc27	223	rmt->rm_blkno = cpu_to_be64(bno);
d2e448d5 DC	224
	225	return sizeof(struct xfs_attr3_rmt_hdr);
	226	}
	227
	228	/*
7bc0dc27	229	* Helper functions to copy attribute data in and out of the one disk extents
d2e448d5	230	*/
7bc0dc27 DC	231	STATIC int
	232	xfs_attr_rmtval_copyout(
	233	struct xfs_mount *mp,
	234	struct xfs_buf *bp,
	235	xfs_ino_t ino,
	236	int *offset,
	237	int *valuelen,
836a94ad	238	__uint8_t **dst)
d2e448d5	239	{
7bc0dc27 DC	240	char *src = bp->b_addr;
	241	xfs_daddr_t bno = bp->b_bn;
	242	int len = BBTOB(bp->b_length);
c2c4c477	243	int blksize = mp->m_attr_geo->blksize;
d2e448d5	244
c2c4c477	245	ASSERT(len >= blksize);
d2e448d5	246
7bc0dc27 DC	247	while (len > 0 && *valuelen > 0) {
7bc0dc27 DC	248	int hdr_size = 0;
c2c4c477	249	int byte_cnt = XFS_ATTR3_RMT_BUF_SPACE(mp, blksize);
7bc0dc27	250
c5c249b4	251	byte_cnt = min(*valuelen, byte_cnt);
7bc0dc27 DC	252
7bc0dc27 DC	253	if (xfs_sb_version_hascrc(&mp->m_sb)) {
6d0081a3	254	if (!xfs_attr3_rmt_hdr_ok(src, ino, *offset,
7bc0dc27 DC	255	byte_cnt, bno)) {
	256	xfs_alert(mp,
	257	"remote attribute header mismatch bno/off/len/owner (0x%llx/0x%x/Ox%x/0x%llx)",
	258	bno, *offset, byte_cnt, ino);
2451337d	259	return -EFSCORRUPTED;
7bc0dc27 DC	260	}
	261	hdr_size = sizeof(struct xfs_attr3_rmt_hdr);
	262	}
	263
	264	memcpy(*dst, src + hdr_size, byte_cnt);
	265
	266	/* roll buffer forwards */
c2c4c477 DC	267	len -= blksize;
	268	src += blksize;
	269	bno += BTOBB(blksize);
7bc0dc27 DC	270
	271	/* roll attribute data forwards */
	272	*valuelen -= byte_cnt;
	273	*dst += byte_cnt;
	274	*offset += byte_cnt;
	275	}
	276	return 0;
	277	}
	278
	279	STATIC void
	280	xfs_attr_rmtval_copyin(
	281	struct xfs_mount *mp,
	282	struct xfs_buf *bp,
	283	xfs_ino_t ino,
	284	int *offset,
	285	int *valuelen,
836a94ad	286	__uint8_t **src)
7bc0dc27 DC	287	{
	288	char *dst = bp->b_addr;
	289	xfs_daddr_t bno = bp->b_bn;
	290	int len = BBTOB(bp->b_length);
c2c4c477	291	int blksize = mp->m_attr_geo->blksize;
7bc0dc27	292
c2c4c477	293	ASSERT(len >= blksize);
7bc0dc27 DC	294
	295	while (len > 0 && *valuelen > 0) {
	296	int hdr_size;
c2c4c477	297	int byte_cnt = XFS_ATTR3_RMT_BUF_SPACE(mp, blksize);
7bc0dc27 DC	298
	299	byte_cnt = min(*valuelen, byte_cnt);
	300	hdr_size = xfs_attr3_rmt_hdr_set(mp, dst, ino, *offset,
	301	byte_cnt, bno);
	302
	303	memcpy(dst + hdr_size, *src, byte_cnt);
	304
	305	/*
	306	* If this is the last block, zero the remainder of it.
	307	* Check that we are actually the last block, too.
	308	*/
c2c4c477	309	if (byte_cnt + hdr_size < blksize) {
7bc0dc27	310	ASSERT(*valuelen - byte_cnt == 0);
c2c4c477	311	ASSERT(len == blksize);
7bc0dc27	312	memset(dst + hdr_size + byte_cnt, 0,
c2c4c477	313	blksize - hdr_size - byte_cnt);
7bc0dc27 DC	314	}
	315
	316	/* roll buffer forwards */
c2c4c477 DC	317	len -= blksize;
	318	dst += blksize;
	319	bno += BTOBB(blksize);
7bc0dc27 DC	320
	321	/* roll attribute data forwards */
	322	*valuelen -= byte_cnt;
	323	*src += byte_cnt;
	324	*offset += byte_cnt;
	325	}
d2e448d5 DC	326	}
d2e448d5 DC	327
95920cd6 DC	328	/*
	329	* Read the value associated with an attribute from the out-of-line buffer
	330	* that we stored it in.
	331	*/
	332	int
d2e448d5 DC	333	xfs_attr_rmtval_get(
d2e448d5 DC	334	struct xfs_da_args *args)
95920cd6	335	{
d2e448d5 DC	336	struct xfs_bmbt_irec map[ATTR_RMTVALUE_MAPSIZE];
	337	struct xfs_mount *mp = args->dp->i_mount;
	338	struct xfs_buf *bp;
d2e448d5	339	xfs_dablk_t lblkno = args->rmtblkno;
836a94ad	340	__uint8_t *dst = args->value;
8275cdd0	341	int valuelen;
d2e448d5 DC	342	int nmap;
d2e448d5 DC	343	int error;
7bc0dc27	344	int blkcnt = args->rmtblkcnt;
d2e448d5 DC	345	int i;
d2e448d5 DC	346	int offset = 0;
95920cd6 DC	347
	348	trace_xfs_attr_rmtval_get(args);
	349
	350	ASSERT(!(args->flags & ATTR_KERNOVAL));
8275cdd0	351	ASSERT(args->rmtvaluelen == args->valuelen);
95920cd6	352
8275cdd0	353	valuelen = args->rmtvaluelen;
95920cd6 DC	354	while (valuelen > 0) {
	355	nmap = ATTR_RMTVALUE_MAPSIZE;
	356	error = xfs_bmapi_read(args->dp, (xfs_fileoff_t)lblkno,
551b382f	357	blkcnt, map, &nmap,
95920cd6 DC	358	XFS_BMAPI_ATTRFORK);
95920cd6 DC	359	if (error)
d2e448d5	360	return error;
95920cd6 DC	361	ASSERT(nmap >= 1);
	362
	363	for (i = 0; (i < nmap) && (valuelen > 0); i++) {
7bc0dc27 DC	364	xfs_daddr_t dblkno;
7bc0dc27 DC	365	int dblkcnt;
d2e448d5	366
95920cd6 DC	367	ASSERT((map[i].br_startblock != DELAYSTARTBLOCK) &&
	368	(map[i].br_startblock != HOLESTARTBLOCK));
	369	dblkno = XFS_FSB_TO_DADDR(mp, map[i].br_startblock);
7bc0dc27	370	dblkcnt = XFS_FSB_TO_BB(mp, map[i].br_blockcount);
95920cd6	371	error = xfs_trans_read_buf(mp, NULL, mp->m_ddev_targp,
7bc0dc27	372	dblkno, dblkcnt, 0, &bp,
d2e448d5	373	&xfs_attr3_rmt_buf_ops);
95920cd6	374	if (error)
d2e448d5 DC	375	return error;
d2e448d5 DC	376
7bc0dc27 DC	377	error = xfs_attr_rmtval_copyout(mp, bp, args->dp->i_ino,
	378	&offset, &valuelen,
	379	&dst);
95920cd6	380	xfs_buf_relse(bp);
7bc0dc27 DC	381	if (error)
7bc0dc27 DC	382	return error;
d2e448d5	383
7bc0dc27	384	/* roll attribute extent map forwards */
95920cd6	385	lblkno += map[i].br_blockcount;
7bc0dc27	386	blkcnt -= map[i].br_blockcount;
95920cd6 DC	387	}
	388	}
	389	ASSERT(valuelen == 0);
d2e448d5	390	return 0;
95920cd6 DC	391	}
	392
	393	/*
	394	* Write the value associated with an attribute into the out-of-line buffer
	395	* that we have defined for it.
	396	*/
	397	int
d2e448d5 DC	398	xfs_attr_rmtval_set(
d2e448d5 DC	399	struct xfs_da_args *args)
95920cd6	400	{
d2e448d5 DC	401	struct xfs_inode *dp = args->dp;
	402	struct xfs_mount *mp = dp->i_mount;
	403	struct xfs_bmbt_irec map;
d2e448d5 DC	404	xfs_dablk_t lblkno;
d2e448d5 DC	405	xfs_fileoff_t lfileoff = 0;
836a94ad	406	__uint8_t *src = args->value;
d2e448d5 DC	407	int blkcnt;
	408	int valuelen;
	409	int nmap;
	410	int error;
d2e448d5	411	int offset = 0;
95920cd6 DC	412
	413	trace_xfs_attr_rmtval_set(args);
	414
95920cd6 DC	415	/*
95920cd6 DC	416	* Find a "hole" in the attribute address space large enough for
d2e448d5 DC	417	* us to drop the new attribute's value into. Because CRC enable
d2e448d5 DC	418	* attributes have headers, we can't just do a straight byte to FSB
7bc0dc27	419	* conversion and have to take the header space into account.
95920cd6	420	*/
8275cdd0	421	blkcnt = xfs_attr3_rmt_blocks(mp, args->rmtvaluelen);
95920cd6 DC	422	error = xfs_bmap_first_unused(args->trans, args->dp, blkcnt, &lfileoff,
95920cd6 DC	423	XFS_ATTR_FORK);
d2e448d5 DC	424	if (error)
	425	return error;
	426
95920cd6 DC	427	args->rmtblkno = lblkno = (xfs_dablk_t)lfileoff;
	428	args->rmtblkcnt = blkcnt;
	429
	430	/*
	431	* Roll through the "value", allocating blocks on disk as required.
	432	*/
	433	while (blkcnt > 0) {
d2e448d5 DC	434	int committed;
d2e448d5 DC	435
95920cd6 DC	436	/*
	437	* Allocate a single extent, up to the size of the value.
	438	*/
	439	xfs_bmap_init(args->flist, args->firstblock);
	440	nmap = 1;
	441	error = xfs_bmapi_write(args->trans, dp, (xfs_fileoff_t)lblkno,
	442	blkcnt,
	443	XFS_BMAPI_ATTRFORK \| XFS_BMAPI_METADATA,
	444	args->firstblock, args->total, &map, &nmap,
	445	args->flist);
	446	if (!error) {
	447	error = xfs_bmap_finish(&args->trans, args->flist,
	448	&committed);
	449	}
	450	if (error) {
	451	ASSERT(committed);
	452	args->trans = NULL;
	453	xfs_bmap_cancel(args->flist);
d99831ff	454	return error;
95920cd6 DC	455	}
	456
	457	/*
	458	* bmap_finish() may have committed the last trans and started
	459	* a new one. We need the inode to be in all transactions.
	460	*/
	461	if (committed)
	462	xfs_trans_ijoin(args->trans, dp, 0);
	463
	464	ASSERT(nmap == 1);
	465	ASSERT((map.br_startblock != DELAYSTARTBLOCK) &&
	466	(map.br_startblock != HOLESTARTBLOCK));
	467	lblkno += map.br_blockcount;
	468	blkcnt -= map.br_blockcount;
	469
	470	/*
	471	* Start the next trans in the chain.
	472	*/
	473	error = xfs_trans_roll(&args->trans, dp);
	474	if (error)
d99831ff	475	return error;
95920cd6 DC	476	}
	477
	478	/*
	479	* Roll through the "value", copying the attribute value to the
	480	* already-allocated blocks. Blocks are written synchronously
	481	* so that we can know they are all on disk before we turn off
	482	* the INCOMPLETE flag.
	483	*/
	484	lblkno = args->rmtblkno;
26f71445	485	blkcnt = args->rmtblkcnt;
8275cdd0	486	valuelen = args->rmtvaluelen;
95920cd6	487	while (valuelen > 0) {
7bc0dc27 DC	488	struct xfs_buf *bp;
	489	xfs_daddr_t dblkno;
	490	int dblkcnt;
	491
	492	ASSERT(blkcnt > 0);
95920cd6	493
95920cd6 DC	494	xfs_bmap_init(args->flist, args->firstblock);
	495	nmap = 1;
	496	error = xfs_bmapi_read(dp, (xfs_fileoff_t)lblkno,
26f71445	497	blkcnt, &map, &nmap,
95920cd6 DC	498	XFS_BMAPI_ATTRFORK);
95920cd6 DC	499	if (error)
d99831ff	500	return error;
95920cd6 DC	501	ASSERT(nmap == 1);
	502	ASSERT((map.br_startblock != DELAYSTARTBLOCK) &&
	503	(map.br_startblock != HOLESTARTBLOCK));
	504
	505	dblkno = XFS_FSB_TO_DADDR(mp, map.br_startblock),
26f71445	506	dblkcnt = XFS_FSB_TO_BB(mp, map.br_blockcount);
95920cd6	507
26f71445	508	bp = xfs_buf_get(mp->m_ddev_targp, dblkno, dblkcnt, 0);
95920cd6	509	if (!bp)
2451337d	510	return -ENOMEM;
d2e448d5	511	bp->b_ops = &xfs_attr3_rmt_buf_ops;
26f71445	512
7bc0dc27 DC	513	xfs_attr_rmtval_copyin(mp, bp, args->dp->i_ino, &offset,
7bc0dc27 DC	514	&valuelen, &src);
95920cd6 DC	515
	516	error = xfs_bwrite(bp); /* GROT: NOTE: synchronous write */
	517	xfs_buf_relse(bp);
	518	if (error)
	519	return error;
d2e448d5	520
95920cd6	521
7bc0dc27	522	/* roll attribute extent map forwards */
95920cd6	523	lblkno += map.br_blockcount;
26f71445	524	blkcnt -= map.br_blockcount;
95920cd6 DC	525	}
95920cd6 DC	526	ASSERT(valuelen == 0);
d2e448d5	527	return 0;
95920cd6 DC	528	}
	529
	530	/*
	531	* Remove the value associated with an attribute by deleting the
	532	* out-of-line buffer that it is stored on.
	533	*/
	534	int
7bc0dc27 DC	535	xfs_attr_rmtval_remove(
7bc0dc27 DC	536	struct xfs_da_args *args)
95920cd6	537	{
7bc0dc27 DC	538	struct xfs_mount *mp = args->dp->i_mount;
	539	xfs_dablk_t lblkno;
	540	int blkcnt;
	541	int error;
	542	int done;
95920cd6 DC	543
	544	trace_xfs_attr_rmtval_remove(args);
	545
95920cd6	546	/*
58a72281	547	* Roll through the "value", invalidating the attribute value's blocks.
95920cd6 DC	548	*/
95920cd6 DC	549	lblkno = args->rmtblkno;
7bc0dc27 DC	550	blkcnt = args->rmtblkcnt;
	551	while (blkcnt > 0) {
	552	struct xfs_bmbt_irec map;
	553	struct xfs_buf *bp;
	554	xfs_daddr_t dblkno;
	555	int dblkcnt;
	556	int nmap;
58a72281	557
95920cd6 DC	558	/*
	559	* Try to remember where we decided to put the value.
	560	*/
	561	nmap = 1;
	562	error = xfs_bmapi_read(args->dp, (xfs_fileoff_t)lblkno,
58a72281	563	blkcnt, &map, &nmap, XFS_BMAPI_ATTRFORK);
95920cd6	564	if (error)
d99831ff	565	return error;
95920cd6 DC	566	ASSERT(nmap == 1);
	567	ASSERT((map.br_startblock != DELAYSTARTBLOCK) &&
	568	(map.br_startblock != HOLESTARTBLOCK));
	569
	570	dblkno = XFS_FSB_TO_DADDR(mp, map.br_startblock),
58a72281	571	dblkcnt = XFS_FSB_TO_BB(mp, map.br_blockcount);
95920cd6 DC	572
	573	/*
	574	* If the "remote" value is in the cache, remove it.
	575	*/
58a72281	576	bp = xfs_incore(mp->m_ddev_targp, dblkno, dblkcnt, XBF_TRYLOCK);
95920cd6 DC	577	if (bp) {
	578	xfs_buf_stale(bp);
	579	xfs_buf_relse(bp);
	580	bp = NULL;
	581	}
	582
95920cd6	583	lblkno += map.br_blockcount;
58a72281	584	blkcnt -= map.br_blockcount;
95920cd6 DC	585	}
	586
	587	/*
	588	* Keep de-allocating extents until the remote-value region is gone.
	589	*/
	590	lblkno = args->rmtblkno;
7bc0dc27	591	blkcnt = args->rmtblkcnt;
95920cd6 DC	592	done = 0;
95920cd6 DC	593	while (!done) {
7bc0dc27 DC	594	int committed;
7bc0dc27 DC	595
95920cd6 DC	596	xfs_bmap_init(args->flist, args->firstblock);
	597	error = xfs_bunmapi(args->trans, args->dp, lblkno, blkcnt,
	598	XFS_BMAPI_ATTRFORK \| XFS_BMAPI_METADATA,
	599	1, args->firstblock, args->flist,
	600	&done);
	601	if (!error) {
	602	error = xfs_bmap_finish(&args->trans, args->flist,
	603	&committed);
	604	}
	605	if (error) {
	606	ASSERT(committed);
	607	args->trans = NULL;
	608	xfs_bmap_cancel(args->flist);
d2e448d5	609	return error;
95920cd6 DC	610	}
	611
	612	/*
	613	* bmap_finish() may have committed the last trans and started
	614	* a new one. We need the inode to be in all transactions.
	615	*/
	616	if (committed)
	617	xfs_trans_ijoin(args->trans, args->dp, 0);
	618
	619	/*
	620	* Close out trans and start the next one in the chain.
	621	*/
	622	error = xfs_trans_roll(&args->trans, args->dp);
	623	if (error)
d99831ff	624	return error;
95920cd6	625	}
d99831ff	626	return 0;
95920cd6	627	}