[thirdparty/xfsprogs-dev.git] / 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 "libxfs_priv.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_bmap.h"
#include "xfs_attr_leaf.h"
#include "xfs_attr_remote.h"
#include "xfs_trans_space.h"
#include "xfs_trace.h"
#include "xfs_cksum.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
9d6fabc0 DC	1	/*
9d6fabc0 DC	2	* Copyright (c) 2000-2005 Silicon Graphics, Inc.
59915981	3	* Copyright (c) 2013 Red Hat, Inc.
9d6fabc0 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	*/
9c799827	19	#include "libxfs_priv.h"
b626fb59 DC	20	#include "xfs_fs.h"
	21	#include "xfs_shared.h"
	22	#include "xfs_format.h"
	23	#include "xfs_log_format.h"
	24	#include "xfs_trans_resv.h"
	25	#include "xfs_bit.h"
	26	#include "xfs_mount.h"
	27	#include "xfs_da_format.h"
	28	#include "xfs_da_btree.h"
	29	#include "xfs_inode.h"
	30	#include "xfs_alloc.h"
	31	#include "xfs_trans.h"
	32	#include "xfs_bmap.h"
	33	#include "xfs_attr_leaf.h"
	34	#include "xfs_attr_remote.h"
	35	#include "xfs_trans_space.h"
	36	#include "xfs_trace.h"
	37	#include "xfs_cksum.h"
9d6fabc0 DC	38
	39	#define ATTR_RMTVALUE_MAPSIZE 1 /* # of map entries at once */
	40
59915981 DC	41	/*
	42	* Each contiguous block has a header, so it is not just a simple attribute
	43	* length to FSB conversion.
	44	*/
f08bc2a9	45	int
59915981 DC	46	xfs_attr3_rmt_blocks(
	47	struct xfs_mount *mp,
	48	int attrlen)
	49	{
72e78b51 DC	50	if (xfs_sb_version_hascrc(&mp->m_sb)) {
	51	int buflen = XFS_ATTR3_RMT_BUF_SPACE(mp, mp->m_sb.sb_blocksize);
	52	return (attrlen + buflen - 1) / buflen;
	53	}
	54	return XFS_B_TO_FSB(mp, attrlen);
59915981 DC	55	}
59915981 DC	56
f08bc2a9 DC	57	/*
	58	* Checking of the remote attribute header is split into two parts. The verifier
	59	* does CRC, location and bounds checking, the unpacking function checks the
	60	* attribute parameters and owner.
	61	*/
	62	static bool
	63	xfs_attr3_rmt_hdr_ok(
f08bc2a9 DC	64	void *ptr,
	65	xfs_ino_t ino,
	66	uint32_t offset,
	67	uint32_t size,
	68	xfs_daddr_t bno)
	69	{
	70	struct xfs_attr3_rmt_hdr *rmt = ptr;
	71
	72	if (bno != be64_to_cpu(rmt->rm_blkno))
	73	return false;
	74	if (offset != be32_to_cpu(rmt->rm_offset))
	75	return false;
	76	if (size != be32_to_cpu(rmt->rm_bytes))
	77	return false;
	78	if (ino != be64_to_cpu(rmt->rm_owner))
	79	return false;
	80
	81	/* ok */
	82	return true;
	83	}
	84
59915981 DC	85	static bool
59915981 DC	86	xfs_attr3_rmt_verify(
f08bc2a9 DC	87	struct xfs_mount *mp,
	88	void *ptr,
	89	int fsbsize,
	90	xfs_daddr_t bno)
59915981	91	{
f08bc2a9	92	struct xfs_attr3_rmt_hdr *rmt = ptr;
59915981 DC	93
	94	if (!xfs_sb_version_hascrc(&mp->m_sb))
	95	return false;
	96	if (rmt->rm_magic != cpu_to_be32(XFS_ATTR3_RMT_MAGIC))
	97	return false;
	98	if (!uuid_equal(&rmt->rm_uuid, &mp->m_sb.sb_uuid))
	99	return false;
f08bc2a9 DC	100	if (be64_to_cpu(rmt->rm_blkno) != bno)
	101	return false;
	102	if (be32_to_cpu(rmt->rm_bytes) > fsbsize - sizeof(*rmt))
59915981 DC	103	return false;
59915981 DC	104	if (be32_to_cpu(rmt->rm_offset) +
b5f5c2d1	105	be32_to_cpu(rmt->rm_bytes) > XATTR_SIZE_MAX)
59915981 DC	106	return false;
	107	if (rmt->rm_owner == 0)
	108	return false;
	109
	110	return true;
	111	}
	112
	113	static void
	114	xfs_attr3_rmt_read_verify(
	115	struct xfs_buf *bp)
	116	{
	117	struct xfs_mount *mp = bp->b_target->bt_mount;
f08bc2a9 DC	118	char *ptr;
f08bc2a9 DC	119	int len;
f08bc2a9	120	xfs_daddr_t bno;
ff105f75	121	int blksize = mp->m_attr_geo->blksize;
59915981 DC	122
	123	/* no verification of non-crc buffers */
	124	if (!xfs_sb_version_hascrc(&mp->m_sb))
	125	return;
	126
f08bc2a9 DC	127	ptr = bp->b_addr;
	128	bno = bp->b_bn;
	129	len = BBTOB(bp->b_length);
ff105f75	130	ASSERT(len >= blksize);
f08bc2a9 DC	131
f08bc2a9 DC	132	while (len > 0) {
ff105f75	133	if (!xfs_verify_cksum(ptr, blksize, XFS_ATTR3_RMT_CRC_OFF)) {
12b53197	134	xfs_buf_ioerror(bp, -EFSBADCRC);
f08bc2a9 DC	135	break;
f08bc2a9 DC	136	}
ff105f75	137	if (!xfs_attr3_rmt_verify(mp, ptr, blksize, bno)) {
12b53197	138	xfs_buf_ioerror(bp, -EFSCORRUPTED);
f08bc2a9 DC	139	break;
f08bc2a9 DC	140	}
ff105f75 DC	141	len -= blksize;
	142	ptr += blksize;
	143	bno += BTOBB(blksize);
f08bc2a9 DC	144	}
f08bc2a9 DC	145
45922933 DC	146	if (bp->b_error)
	147	xfs_verifier_error(bp);
	148	else
f08bc2a9	149	ASSERT(len == 0);
59915981 DC	150	}
	151
	152	static void
	153	xfs_attr3_rmt_write_verify(
	154	struct xfs_buf *bp)
	155	{
	156	struct xfs_mount *mp = bp->b_target->bt_mount;
	157	struct xfs_buf_log_item *bip = bp->b_fspriv;
f08bc2a9 DC	158	char *ptr;
	159	int len;
	160	xfs_daddr_t bno;
ff105f75	161	int blksize = mp->m_attr_geo->blksize;
59915981 DC	162
	163	/* no verification of non-crc buffers */
	164	if (!xfs_sb_version_hascrc(&mp->m_sb))
	165	return;
	166
f08bc2a9 DC	167	ptr = bp->b_addr;
	168	bno = bp->b_bn;
	169	len = BBTOB(bp->b_length);
ff105f75	170	ASSERT(len >= blksize);
f08bc2a9 DC	171
f08bc2a9 DC	172	while (len > 0) {
ff105f75	173	if (!xfs_attr3_rmt_verify(mp, ptr, blksize, bno)) {
12b53197	174	xfs_buf_ioerror(bp, -EFSCORRUPTED);
45922933	175	xfs_verifier_error(bp);
f08bc2a9 DC	176	return;
	177	}
	178	if (bip) {
	179	struct xfs_attr3_rmt_hdr *rmt;
59915981	180
f08bc2a9 DC	181	rmt = (struct xfs_attr3_rmt_hdr *)ptr;
	182	rmt->rm_lsn = cpu_to_be64(bip->bli_item.li_lsn);
	183	}
ff105f75	184	xfs_update_cksum(ptr, blksize, XFS_ATTR3_RMT_CRC_OFF);
f08bc2a9	185
ff105f75 DC	186	len -= blksize;
	187	ptr += blksize;
	188	bno += BTOBB(blksize);
59915981	189	}
f08bc2a9	190	ASSERT(len == 0);
59915981 DC	191	}
	192
	193	const struct xfs_buf_ops xfs_attr3_rmt_buf_ops = {
	194	.verify_read = xfs_attr3_rmt_read_verify,
	195	.verify_write = xfs_attr3_rmt_write_verify,
	196	};
	197
f08bc2a9	198	STATIC int
59915981 DC	199	xfs_attr3_rmt_hdr_set(
59915981 DC	200	struct xfs_mount *mp,
f08bc2a9	201	void *ptr,
59915981 DC	202	xfs_ino_t ino,
	203	uint32_t offset,
	204	uint32_t size,
f08bc2a9	205	xfs_daddr_t bno)
59915981	206	{
f08bc2a9	207	struct xfs_attr3_rmt_hdr *rmt = ptr;
59915981 DC	208
	209	if (!xfs_sb_version_hascrc(&mp->m_sb))
	210	return 0;
	211
	212	rmt->rm_magic = cpu_to_be32(XFS_ATTR3_RMT_MAGIC);
	213	rmt->rm_offset = cpu_to_be32(offset);
	214	rmt->rm_bytes = cpu_to_be32(size);
	215	uuid_copy(&rmt->rm_uuid, &mp->m_sb.sb_uuid);
	216	rmt->rm_owner = cpu_to_be64(ino);
f08bc2a9	217	rmt->rm_blkno = cpu_to_be64(bno);
59915981 DC	218
	219	return sizeof(struct xfs_attr3_rmt_hdr);
	220	}
	221
	222	/*
f08bc2a9	223	* Helper functions to copy attribute data in and out of the one disk extents
59915981	224	*/
f08bc2a9 DC	225	STATIC int
	226	xfs_attr_rmtval_copyout(
	227	struct xfs_mount *mp,
	228	struct xfs_buf *bp,
	229	xfs_ino_t ino,
	230	int *offset,
	231	int *valuelen,
6bddecbc	232	__uint8_t **dst)
59915981	233	{
f08bc2a9 DC	234	char *src = bp->b_addr;
	235	xfs_daddr_t bno = bp->b_bn;
	236	int len = BBTOB(bp->b_length);
ff105f75	237	int blksize = mp->m_attr_geo->blksize;
59915981	238
ff105f75	239	ASSERT(len >= blksize);
59915981	240
f08bc2a9 DC	241	while (len > 0 && *valuelen > 0) {
f08bc2a9 DC	242	int hdr_size = 0;
ff105f75	243	int byte_cnt = XFS_ATTR3_RMT_BUF_SPACE(mp, blksize);
f08bc2a9 DC	244
	245	byte_cnt = min(*valuelen, byte_cnt);
	246
	247	if (xfs_sb_version_hascrc(&mp->m_sb)) {
ff105f75	248	if (!xfs_attr3_rmt_hdr_ok(src, ino, *offset,
f08bc2a9 DC	249	byte_cnt, bno)) {
	250	xfs_alert(mp,
	251	"remote attribute header mismatch bno/off/len/owner (0x%llx/0x%x/Ox%x/0x%llx)",
	252	bno, *offset, byte_cnt, ino);
12b53197	253	return -EFSCORRUPTED;
f08bc2a9 DC	254	}
	255	hdr_size = sizeof(struct xfs_attr3_rmt_hdr);
	256	}
	257
	258	memcpy(*dst, src + hdr_size, byte_cnt);
	259
	260	/* roll buffer forwards */
ff105f75 DC	261	len -= blksize;
	262	src += blksize;
	263	bno += BTOBB(blksize);
f08bc2a9 DC	264
	265	/* roll attribute data forwards */
	266	*valuelen -= byte_cnt;
	267	*dst += byte_cnt;
	268	*offset += byte_cnt;
	269	}
	270	return 0;
	271	}
	272
	273	STATIC void
	274	xfs_attr_rmtval_copyin(
	275	struct xfs_mount *mp,
	276	struct xfs_buf *bp,
	277	xfs_ino_t ino,
	278	int *offset,
	279	int *valuelen,
6bddecbc	280	__uint8_t **src)
f08bc2a9 DC	281	{
	282	char *dst = bp->b_addr;
	283	xfs_daddr_t bno = bp->b_bn;
	284	int len = BBTOB(bp->b_length);
ff105f75	285	int blksize = mp->m_attr_geo->blksize;
f08bc2a9	286
ff105f75	287	ASSERT(len >= blksize);
f08bc2a9 DC	288
	289	while (len > 0 && *valuelen > 0) {
	290	int hdr_size;
ff105f75	291	int byte_cnt = XFS_ATTR3_RMT_BUF_SPACE(mp, blksize);
f08bc2a9 DC	292
	293	byte_cnt = min(*valuelen, byte_cnt);
	294	hdr_size = xfs_attr3_rmt_hdr_set(mp, dst, ino, *offset,
	295	byte_cnt, bno);
	296
	297	memcpy(dst + hdr_size, *src, byte_cnt);
	298
	299	/*
	300	* If this is the last block, zero the remainder of it.
	301	* Check that we are actually the last block, too.
	302	*/
ff105f75	303	if (byte_cnt + hdr_size < blksize) {
f08bc2a9	304	ASSERT(*valuelen - byte_cnt == 0);
ff105f75	305	ASSERT(len == blksize);
f08bc2a9	306	memset(dst + hdr_size + byte_cnt, 0,
ff105f75	307	blksize - hdr_size - byte_cnt);
f08bc2a9 DC	308	}
	309
	310	/* roll buffer forwards */
ff105f75 DC	311	len -= blksize;
	312	dst += blksize;
	313	bno += BTOBB(blksize);
f08bc2a9 DC	314
	315	/* roll attribute data forwards */
	316	*valuelen -= byte_cnt;
	317	*src += byte_cnt;
	318	*offset += byte_cnt;
	319	}
59915981 DC	320	}
59915981 DC	321
9d6fabc0 DC	322	/*
	323	* Read the value associated with an attribute from the out-of-line buffer
	324	* that we stored it in.
	325	*/
	326	int
59915981 DC	327	xfs_attr_rmtval_get(
59915981 DC	328	struct xfs_da_args *args)
9d6fabc0	329	{
59915981 DC	330	struct xfs_bmbt_irec map[ATTR_RMTVALUE_MAPSIZE];
	331	struct xfs_mount *mp = args->dp->i_mount;
	332	struct xfs_buf *bp;
59915981	333	xfs_dablk_t lblkno = args->rmtblkno;
6bddecbc	334	__uint8_t *dst = args->value;
ff105f75	335	int valuelen;
59915981 DC	336	int nmap;
59915981 DC	337	int error;
f08bc2a9	338	int blkcnt = args->rmtblkcnt;
59915981 DC	339	int i;
59915981 DC	340	int offset = 0;
9d6fabc0 DC	341
	342	trace_xfs_attr_rmtval_get(args);
	343
	344	ASSERT(!(args->flags & ATTR_KERNOVAL));
ff105f75	345	ASSERT(args->rmtvaluelen == args->valuelen);
9d6fabc0	346
ff105f75	347	valuelen = args->rmtvaluelen;
9d6fabc0 DC	348	while (valuelen > 0) {
	349	nmap = ATTR_RMTVALUE_MAPSIZE;
	350	error = xfs_bmapi_read(args->dp, (xfs_fileoff_t)lblkno,
72e78b51	351	blkcnt, map, &nmap,
9d6fabc0 DC	352	XFS_BMAPI_ATTRFORK);
9d6fabc0 DC	353	if (error)
59915981	354	return error;
9d6fabc0 DC	355	ASSERT(nmap >= 1);
	356
	357	for (i = 0; (i < nmap) && (valuelen > 0); i++) {
f08bc2a9 DC	358	xfs_daddr_t dblkno;
f08bc2a9 DC	359	int dblkcnt;
59915981	360
9d6fabc0 DC	361	ASSERT((map[i].br_startblock != DELAYSTARTBLOCK) &&
	362	(map[i].br_startblock != HOLESTARTBLOCK));
	363	dblkno = XFS_FSB_TO_DADDR(mp, map[i].br_startblock);
f08bc2a9	364	dblkcnt = XFS_FSB_TO_BB(mp, map[i].br_blockcount);
9d6fabc0	365	error = xfs_trans_read_buf(mp, NULL, mp->m_ddev_targp,
f08bc2a9	366	dblkno, dblkcnt, 0, &bp,
59915981	367	&xfs_attr3_rmt_buf_ops);
9d6fabc0	368	if (error)
59915981 DC	369	return error;
59915981 DC	370
f08bc2a9 DC	371	error = xfs_attr_rmtval_copyout(mp, bp, args->dp->i_ino,
	372	&offset, &valuelen,
	373	&dst);
9d6fabc0	374	xfs_buf_relse(bp);
f08bc2a9 DC	375	if (error)
f08bc2a9 DC	376	return error;
59915981	377
f08bc2a9	378	/* roll attribute extent map forwards */
9d6fabc0	379	lblkno += map[i].br_blockcount;
f08bc2a9	380	blkcnt -= map[i].br_blockcount;
9d6fabc0 DC	381	}
	382	}
	383	ASSERT(valuelen == 0);
59915981	384	return 0;
9d6fabc0 DC	385	}
	386
	387	/*
	388	* Write the value associated with an attribute into the out-of-line buffer
	389	* that we have defined for it.
	390	*/
	391	int
59915981 DC	392	xfs_attr_rmtval_set(
59915981 DC	393	struct xfs_da_args *args)
9d6fabc0	394	{
59915981 DC	395	struct xfs_inode *dp = args->dp;
	396	struct xfs_mount *mp = dp->i_mount;
	397	struct xfs_bmbt_irec map;
59915981 DC	398	xfs_dablk_t lblkno;
59915981 DC	399	xfs_fileoff_t lfileoff = 0;
6bddecbc	400	__uint8_t *src = args->value;
59915981 DC	401	int blkcnt;
	402	int valuelen;
	403	int nmap;
	404	int error;
59915981	405	int offset = 0;
9d6fabc0 DC	406
	407	trace_xfs_attr_rmtval_set(args);
	408
9d6fabc0 DC	409	/*
9d6fabc0 DC	410	* Find a "hole" in the attribute address space large enough for
59915981 DC	411	* us to drop the new attribute's value into. Because CRC enable
59915981 DC	412	* attributes have headers, we can't just do a straight byte to FSB
f08bc2a9	413	* conversion and have to take the header space into account.
9d6fabc0	414	*/
ff105f75	415	blkcnt = xfs_attr3_rmt_blocks(mp, args->rmtvaluelen);
9d6fabc0 DC	416	error = xfs_bmap_first_unused(args->trans, args->dp, blkcnt, &lfileoff,
9d6fabc0 DC	417	XFS_ATTR_FORK);
59915981 DC	418	if (error)
	419	return error;
	420
9d6fabc0 DC	421	args->rmtblkno = lblkno = (xfs_dablk_t)lfileoff;
	422	args->rmtblkcnt = blkcnt;
	423
	424	/*
	425	* Roll through the "value", allocating blocks on disk as required.
	426	*/
	427	while (blkcnt > 0) {
59915981 DC	428	int committed;
59915981 DC	429
9d6fabc0 DC	430	/*
	431	* Allocate a single extent, up to the size of the value.
	432	*/
	433	xfs_bmap_init(args->flist, args->firstblock);
	434	nmap = 1;
	435	error = xfs_bmapi_write(args->trans, dp, (xfs_fileoff_t)lblkno,
	436	blkcnt,
	437	XFS_BMAPI_ATTRFORK \| XFS_BMAPI_METADATA,
	438	args->firstblock, args->total, &map, &nmap,
	439	args->flist);
	440	if (!error) {
	441	error = xfs_bmap_finish(&args->trans, args->flist,
	442	&committed);
	443	}
	444	if (error) {
	445	ASSERT(committed);
	446	args->trans = NULL;
	447	xfs_bmap_cancel(args->flist);
af43ca9f	448	return error;
9d6fabc0 DC	449	}
	450
	451	/*
	452	* bmap_finish() may have committed the last trans and started
	453	* a new one. We need the inode to be in all transactions.
	454	*/
	455	if (committed)
	456	xfs_trans_ijoin(args->trans, dp, 0);
	457
	458	ASSERT(nmap == 1);
	459	ASSERT((map.br_startblock != DELAYSTARTBLOCK) &&
	460	(map.br_startblock != HOLESTARTBLOCK));
	461	lblkno += map.br_blockcount;
	462	blkcnt -= map.br_blockcount;
	463
	464	/*
	465	* Start the next trans in the chain.
	466	*/
	467	error = xfs_trans_roll(&args->trans, dp);
	468	if (error)
af43ca9f	469	return error;
9d6fabc0 DC	470	}
	471
	472	/*
	473	* Roll through the "value", copying the attribute value to the
	474	* already-allocated blocks. Blocks are written synchronously
	475	* so that we can know they are all on disk before we turn off
	476	* the INCOMPLETE flag.
	477	*/
	478	lblkno = args->rmtblkno;
4da23ca6	479	blkcnt = args->rmtblkcnt;
ff105f75	480	valuelen = args->rmtvaluelen;
9d6fabc0	481	while (valuelen > 0) {
f08bc2a9 DC	482	struct xfs_buf *bp;
	483	xfs_daddr_t dblkno;
	484	int dblkcnt;
	485
	486	ASSERT(blkcnt > 0);
9d6fabc0	487
9d6fabc0 DC	488	xfs_bmap_init(args->flist, args->firstblock);
	489	nmap = 1;
	490	error = xfs_bmapi_read(dp, (xfs_fileoff_t)lblkno,
4da23ca6	491	blkcnt, &map, &nmap,
9d6fabc0 DC	492	XFS_BMAPI_ATTRFORK);
9d6fabc0 DC	493	if (error)
af43ca9f	494	return error;
9d6fabc0 DC	495	ASSERT(nmap == 1);
	496	ASSERT((map.br_startblock != DELAYSTARTBLOCK) &&
	497	(map.br_startblock != HOLESTARTBLOCK));
	498
	499	dblkno = XFS_FSB_TO_DADDR(mp, map.br_startblock),
4da23ca6	500	dblkcnt = XFS_FSB_TO_BB(mp, map.br_blockcount);
9d6fabc0	501
4da23ca6	502	bp = xfs_buf_get(mp->m_ddev_targp, dblkno, dblkcnt, 0);
9d6fabc0	503	if (!bp)
12b53197	504	return -ENOMEM;
59915981 DC	505	bp->b_ops = &xfs_attr3_rmt_buf_ops;
59915981 DC	506
f08bc2a9 DC	507	xfs_attr_rmtval_copyin(mp, bp, args->dp->i_ino, &offset,
f08bc2a9 DC	508	&valuelen, &src);
9d6fabc0 DC	509
	510	error = xfs_bwrite(bp); /* GROT: NOTE: synchronous write */
	511	xfs_buf_relse(bp);
	512	if (error)
	513	return error;
59915981	514
9d6fabc0	515
f08bc2a9	516	/* roll attribute extent map forwards */
9d6fabc0	517	lblkno += map.br_blockcount;
4da23ca6	518	blkcnt -= map.br_blockcount;
9d6fabc0 DC	519	}
9d6fabc0 DC	520	ASSERT(valuelen == 0);
59915981	521	return 0;
9d6fabc0 DC	522	}
	523
	524	/*
	525	* Remove the value associated with an attribute by deleting the
	526	* out-of-line buffer that it is stored on.
	527	*/
	528	int
f08bc2a9 DC	529	xfs_attr_rmtval_remove(
f08bc2a9 DC	530	struct xfs_da_args *args)
9d6fabc0	531	{
f08bc2a9 DC	532	struct xfs_mount *mp = args->dp->i_mount;
	533	xfs_dablk_t lblkno;
	534	int blkcnt;
	535	int error;
	536	int done;
9d6fabc0 DC	537
	538	trace_xfs_attr_rmtval_remove(args);
	539
9d6fabc0	540	/*
526f9be5	541	* Roll through the "value", invalidating the attribute value's blocks.
9d6fabc0 DC	542	*/
9d6fabc0 DC	543	lblkno = args->rmtblkno;
f08bc2a9 DC	544	blkcnt = args->rmtblkcnt;
	545	while (blkcnt > 0) {
	546	struct xfs_bmbt_irec map;
	547	struct xfs_buf *bp;
	548	xfs_daddr_t dblkno;
	549	int dblkcnt;
	550	int nmap;
526f9be5	551
9d6fabc0 DC	552	/*
	553	* Try to remember where we decided to put the value.
	554	*/
	555	nmap = 1;
	556	error = xfs_bmapi_read(args->dp, (xfs_fileoff_t)lblkno,
526f9be5	557	blkcnt, &map, &nmap, XFS_BMAPI_ATTRFORK);
9d6fabc0	558	if (error)
af43ca9f	559	return error;
9d6fabc0 DC	560	ASSERT(nmap == 1);
	561	ASSERT((map.br_startblock != DELAYSTARTBLOCK) &&
	562	(map.br_startblock != HOLESTARTBLOCK));
	563
	564	dblkno = XFS_FSB_TO_DADDR(mp, map.br_startblock),
526f9be5	565	dblkcnt = XFS_FSB_TO_BB(mp, map.br_blockcount);
9d6fabc0 DC	566
	567	/*
	568	* If the "remote" value is in the cache, remove it.
	569	*/
526f9be5	570	bp = xfs_incore(mp->m_ddev_targp, dblkno, dblkcnt, XBF_TRYLOCK);
9d6fabc0 DC	571	if (bp) {
	572	xfs_buf_stale(bp);
	573	xfs_buf_relse(bp);
	574	bp = NULL;
	575	}
	576
9d6fabc0	577	lblkno += map.br_blockcount;
526f9be5	578	blkcnt -= map.br_blockcount;
9d6fabc0 DC	579	}
	580
	581	/*
	582	* Keep de-allocating extents until the remote-value region is gone.
	583	*/
	584	lblkno = args->rmtblkno;
f08bc2a9	585	blkcnt = args->rmtblkcnt;
9d6fabc0 DC	586	done = 0;
9d6fabc0 DC	587	while (!done) {
f08bc2a9 DC	588	int committed;
f08bc2a9 DC	589
9d6fabc0 DC	590	xfs_bmap_init(args->flist, args->firstblock);
	591	error = xfs_bunmapi(args->trans, args->dp, lblkno, blkcnt,
	592	XFS_BMAPI_ATTRFORK \| XFS_BMAPI_METADATA,
	593	1, args->firstblock, args->flist,
	594	&done);
	595	if (!error) {
	596	error = xfs_bmap_finish(&args->trans, args->flist,
	597	&committed);
	598	}
	599	if (error) {
	600	ASSERT(committed);
	601	args->trans = NULL;
	602	xfs_bmap_cancel(args->flist);
59915981	603	return error;
9d6fabc0 DC	604	}
	605
	606	/*
	607	* bmap_finish() may have committed the last trans and started
	608	* a new one. We need the inode to be in all transactions.
	609	*/
	610	if (committed)
	611	xfs_trans_ijoin(args->trans, args->dp, 0);
	612
	613	/*
	614	* Close out trans and start the next one in the chain.
	615	*/
	616	error = xfs_trans_roll(&args->trans, args->dp);
	617	if (error)
af43ca9f	618	return error;
9d6fabc0	619	}
af43ca9f	620	return 0;
9d6fabc0	621	}