/*

 * CDDL HEADER START

 *

 * The contents of this file are subject to the terms of the

 * Common Development and Distribution License (the "License").

 * You may not use this file except in compliance with the License.

 *

 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE

 * or http://www.opensolaris.org/os/licensing.

 * See the License for the specific language governing permissions

 * and limitations under the License.

 *

 * When distributing Covered Code, include this CDDL HEADER in each

 * file and include the License file at usr/src/OPENSOLARIS.LICENSE.

 * If applicable, add the following below this CDDL HEADER, with the

 * fields enclosed by brackets "[]" replaced with your own identifying

 * information: Portions Copyright [yyyy] [name of copyright owner]

 *

 * CDDL HEADER END

 */

/*

 * Copyright 2008 Sun Microsystems, Inc.  All rights reserved.

 * Use is subject to license terms.

 */

#pragma ident	"%Z%%M%	%I%	%E% SMI"

/*

 * ZFS volume emulation driver.

 *

 * Makes a DMU object look like a volume of arbitrary size, up to 2^64 bytes.

 * Volumes are accessed through the symbolic links named:

 *

 * /dev/zvol/dsk/<pool_name>/<dataset_name>

 * /dev/zvol/rdsk/<pool_name>/<dataset_name>

 *

 * These links are created by the ZFS-specific devfsadm link generator.

 * Volumes are persistent through reboot.  No user command needs to be

 * run before opening and using a device.

 */

#include <sys/types.h>

#include <sys/param.h>

#include <sys/errno.h>

#include <sys/uio.h>

#include <sys/buf.h>

#include <sys/modctl.h>

#include <sys/open.h>

#include <sys/kmem.h>

#include <sys/conf.h>

#include <sys/cmn_err.h>

#include <sys/stat.h>

#include <sys/zap.h>

#include <sys/spa.h>

#include <sys/zio.h>

#include <sys/dmu_traverse.h>

#include <sys/dnode.h>

#include <sys/dsl_dataset.h>

#include <sys/dsl_prop.h>

#include <sys/dkio.h>

#include <sys/efi_partition.h>

#include <sys/byteorder.h>

#include <sys/pathname.h>

#include <sys/ddi.h>

#include <sys/sunddi.h>

#include <sys/crc32.h>

#include <sys/dirent.h>

#include <sys/policy.h>

#include <sys/fs/zfs.h>

#include <sys/zfs_ioctl.h>

#include <sys/mkdev.h>

#include <sys/zil.h>

#include <sys/refcount.h>

#include <sys/zfs_znode.h>

#include <sys/zfs_rlock.h>

#include <sys/vdev_disk.h>

#include <sys/vdev_impl.h>

#include <sys/zvol.h>

#include <sys/dumphdr.h>

#include "zfs_namecheck.h"

static void *zvol_state;

#define	ZVOL_DUMPSIZE		"dumpsize"

/*

 * This lock protects the zvol_state structure from being modified

 * while it's being used, e.g. an open that comes in before a create

 * finishes.  It also protects temporary opens of the dataset so that,

 * e.g., an open doesn't get a spurious EBUSY.

 */

static kmutex_t zvol_state_lock;

static uint32_t zvol_minors;

#define	NUM_EXTENTS	((SPA_MAXBLOCKSIZE) / sizeof (zvol_extent_t))

typedef struct zvol_extent {

	dva_t		ze_dva;		/* dva associated with this extent */

	uint64_t	ze_stride;	/* extent stride */

	uint64_t	ze_size;	/* number of blocks in extent */

} zvol_extent_t;

/*

 * The list of extents associated with the dump device

 */

typedef struct zvol_ext_list {

	zvol_extent_t		zl_extents[NUM_EXTENTS];

	struct zvol_ext_list	*zl_next;

} zvol_ext_list_t;

/*

 * The in-core state of each volume.

 */

typedef struct zvol_state {

	char		zv_name[MAXPATHLEN]; /* pool/dd name */

	uint64_t	zv_volsize;	/* amount of space we advertise */

	uint64_t	zv_volblocksize; /* volume block size */

	minor_t		zv_minor;	/* minor number */

	uint8_t		zv_min_bs;	/* minimum addressable block shift */

	uint8_t		zv_flags;	/* readonly; dumpified */

	objset_t	*zv_objset;	/* objset handle */

	uint32_t	zv_mode;	/* DS_MODE_* flags at open time */

	uint32_t	zv_open_count[OTYPCNT];	/* open counts */

	uint32_t	zv_total_opens;	/* total open count */

	zilog_t		*zv_zilog;	/* ZIL handle */

	zvol_ext_list_t	*zv_list;	/* List of extents for dump */

	uint64_t	zv_txg_assign;	/* txg to assign during ZIL replay */

	znode_t		zv_znode;	/* for range locking */

} zvol_state_t;

/*

 * zvol specific flags

 */

#define	ZVOL_RDONLY	0x1

#define	ZVOL_DUMPIFIED	0x2

/*

 * zvol maximum transfer in one DMU tx.

 */

int zvol_maxphys = DMU_MAX_ACCESS/2;

extern int zfs_set_prop_nvlist(const char *, nvlist_t *);

static int zvol_get_data(void *arg, lr_write_t *lr, char *buf, zio_t *zio);

static int zvol_dumpify(zvol_state_t *zv);

static int zvol_dump_fini(zvol_state_t *zv);

static int zvol_dump_init(zvol_state_t *zv, boolean_t resize);

static void

zvol_size_changed(zvol_state_t *zv, major_t maj)

{

	dev_t dev = makedevice(maj, zv->zv_minor);

	VERIFY(ddi_prop_update_int64(dev, zfs_dip,

	    "Size", zv->zv_volsize) == DDI_SUCCESS);

	VERIFY(ddi_prop_update_int64(dev, zfs_dip,

	    "Nblocks", lbtodb(zv->zv_volsize)) == DDI_SUCCESS);

	/* Notify specfs to invalidate the cached size */

	spec_size_invalidate(dev, VBLK);

	spec_size_invalidate(dev, VCHR);

}

int

zvol_check_volsize(uint64_t volsize, uint64_t blocksize)

{

	if (volsize == 0)

		return (EINVAL);

	if (volsize % blocksize != 0)

		return (EINVAL);

#ifdef _ILP32

	if (volsize - 1 > SPEC_MAXOFFSET_T)

		return (EOVERFLOW);

#endif

	return (0);

}

int

zvol_check_volblocksize(uint64_t volblocksize)

{

	if (volblocksize < SPA_MINBLOCKSIZE ||

	    volblocksize > SPA_MAXBLOCKSIZE ||

	    !ISP2(volblocksize))

		return (EDOM);

	return (0);

}

static void

zvol_readonly_changed_cb(void *arg, uint64_t newval)

{

	zvol_state_t *zv = arg;

	if (newval)

		zv->zv_flags |= ZVOL_RDONLY;

	else

		zv->zv_flags &= ~ZVOL_RDONLY;

}

int

zvol_get_stats(objset_t *os, nvlist_t *nv)

{

	int error;

	dmu_object_info_t doi;

	uint64_t val;

	error = zap_lookup(os, ZVOL_ZAP_OBJ, "size", 8, 1, &val);

	if (error)

		return (error);

	dsl_prop_nvlist_add_uint64(nv, ZFS_PROP_VOLSIZE, val);

	error = dmu_object_info(os, ZVOL_OBJ, &doi);

	if (error == 0) {

		dsl_prop_nvlist_add_uint64(nv, ZFS_PROP_VOLBLOCKSIZE,

		    doi.doi_data_block_size);

	}

	return (error);

}

/*

 * Find a free minor number.

 */

static minor_t

zvol_minor_alloc(void)

{

	minor_t minor;

	ASSERT(MUTEX_HELD(&zvol_state_lock));

	for (minor = 1; minor <= ZVOL_MAX_MINOR; minor++)

		if (ddi_get_soft_state(zvol_state, minor) == NULL)

			return (minor);

	return (0);

}

static zvol_state_t *

zvol_minor_lookup(const char *name)

{

	minor_t minor;

	zvol_state_t *zv;

	ASSERT(MUTEX_HELD(&zvol_state_lock));

	for (minor = 1; minor <= ZVOL_MAX_MINOR; minor++) {

		zv = ddi_get_soft_state(zvol_state, minor);

		if (zv == NULL)

			continue;

		if (strcmp(zv->zv_name, name) == 0)

			break;

	}

	return (zv);

}

void

zvol_init_extent(zvol_extent_t *ze, blkptr_t *bp)

{

	ze->ze_dva = bp->blk_dva[0];	/* structure assignment */

	ze->ze_stride = 0;

	ze->ze_size = 1;

}

/* extent mapping arg */

struct maparg {

	zvol_ext_list_t	*ma_list;

	zvol_extent_t	*ma_extent;

	int		ma_gang;

};

/*ARGSUSED*/

static int

zvol_map_block(traverse_blk_cache_t *bc, spa_t *spa, void *arg)

{

	zbookmark_t *zb = &bc->bc_bookmark;

	blkptr_t *bp = &bc->bc_blkptr;

	void *data = bc->bc_data;

	dnode_phys_t *dnp = bc->bc_dnode;

	struct maparg *ma = (struct maparg *)arg;

	uint64_t stride;

	/* If there is an error, then keep trying to make progress */

	if (bc->bc_errno)

		return (ERESTART);

#ifdef ZFS_DEBUG

	if (zb->zb_level == -1) {

		ASSERT3U(BP_GET_TYPE(bp), ==, DMU_OT_OBJSET);

		ASSERT3U(BP_GET_LEVEL(bp), ==, 0);

	} else {

		ASSERT3U(BP_GET_TYPE(bp), ==, dnp->dn_type);

		ASSERT3U(BP_GET_LEVEL(bp), ==, zb->zb_level);

	}

	if (zb->zb_level > 0) {

		uint64_t fill = 0;

		blkptr_t *bpx, *bpend;

		for (bpx = data, bpend = bpx + BP_GET_LSIZE(bp) / sizeof (*bpx);

		    bpx < bpend; bpx++) {

			if (bpx->blk_birth != 0) {

				fill += bpx->blk_fill;

			} else {

				ASSERT(bpx->blk_fill == 0);

			}

		}

		ASSERT3U(fill, ==, bp->blk_fill);

	}

	if (zb->zb_level == 0 && dnp->dn_type == DMU_OT_DNODE) {

		uint64_t fill = 0;

		dnode_phys_t *dnx, *dnend;

		for (dnx = data, dnend = dnx + (BP_GET_LSIZE(bp)>>DNODE_SHIFT);

		    dnx < dnend; dnx++) {

			if (dnx->dn_type != DMU_OT_NONE)

				fill++;

		}

		ASSERT3U(fill, ==, bp->blk_fill);

	}

#endif

	if (zb->zb_level || dnp->dn_type == DMU_OT_DNODE)

		return (0);

	/* Abort immediately if we have encountered gang blocks */

	if (BP_IS_GANG(bp)) {

		ma->ma_gang++;

		return (EINTR);

	}

	/* first time? */

	if (ma->ma_extent->ze_size == 0) {

		zvol_init_extent(ma->ma_extent, bp);

		return (0);

	}

	stride = (DVA_GET_OFFSET(&bp->blk_dva[0])) -

	    ((DVA_GET_OFFSET(&ma->ma_extent->ze_dva)) +

	    (ma->ma_extent->ze_size - 1) * (ma->ma_extent->ze_stride));

	if (DVA_GET_VDEV(BP_IDENTITY(bp)) ==

	    DVA_GET_VDEV(&ma->ma_extent->ze_dva)) {

		if (ma->ma_extent->ze_stride == 0) {

			/* second block in this extent */

			ma->ma_extent->ze_stride = stride;

			ma->ma_extent->ze_size++;

			return (0);

		} else if (ma->ma_extent->ze_stride == stride) {

			/*

			 * the block we allocated has the same

			 * stride

			 */

			ma->ma_extent->ze_size++;

			return (0);

		}

	}

	/*

	 * dtrace -n 'zfs-dprintf

	 * /stringof(arg0) == "zvol.c"/

	 * {

	 *	printf("%s: %s", stringof(arg1), stringof(arg3))

	 * } '

	 */

	dprintf("ma_extent 0x%lx mrstride 0x%lx stride %lx\n",

	    ma->ma_extent->ze_size, ma->ma_extent->ze_stride, stride);

	dprintf_bp(bp, "%s", "next blkptr:");

	/* start a new extent */

	if (ma->ma_extent == &ma->ma_list->zl_extents[NUM_EXTENTS - 1]) {

		ma->ma_list->zl_next = kmem_zalloc(sizeof (zvol_ext_list_t),

		    KM_SLEEP);

		ma->ma_list = ma->ma_list->zl_next;

		ma->ma_extent = &ma->ma_list->zl_extents[0];

	} else {

		ma->ma_extent++;

	}

	zvol_init_extent(ma->ma_extent, bp);

	return (0);

}

/* ARGSUSED */

void

zvol_create_cb(objset_t *os, void *arg, cred_t *cr, dmu_tx_t *tx)

{

	zfs_creat_t *zct = arg;

	nvlist_t *nvprops = zct->zct_props;

	int error;

	uint64_t volblocksize, volsize;

	VERIFY(nvlist_lookup_uint64(nvprops,

	    zfs_prop_to_name(ZFS_PROP_VOLSIZE), &volsize) == 0);

	if (nvlist_lookup_uint64(nvprops,

	    zfs_prop_to_name(ZFS_PROP_VOLBLOCKSIZE), &volblocksize) != 0)

		volblocksize = zfs_prop_default_numeric(ZFS_PROP_VOLBLOCKSIZE);

	/*

	 * These properties must be removed from the list so the generic

	 * property setting step won't apply to them.

	 */

	VERIFY(nvlist_remove_all(nvprops,

	    zfs_prop_to_name(ZFS_PROP_VOLSIZE)) == 0);

	(void) nvlist_remove_all(nvprops,

	    zfs_prop_to_name(ZFS_PROP_VOLBLOCKSIZE));

	error = dmu_object_claim(os, ZVOL_OBJ, DMU_OT_ZVOL, volblocksize,

	    DMU_OT_NONE, 0, tx);

	ASSERT(error == 0);

	error = zap_create_claim(os, ZVOL_ZAP_OBJ, DMU_OT_ZVOL_PROP,

	    DMU_OT_NONE, 0, tx);

	ASSERT(error == 0);

	error = zap_update(os, ZVOL_ZAP_OBJ, "size", 8, 1, &volsize, tx);

	ASSERT(error == 0);

}

/*

 * Replay a TX_WRITE ZIL transaction that didn't get committed

 * after a system failure

 */

static int

zvol_replay_write(zvol_state_t *zv, lr_write_t *lr, boolean_t byteswap)

{

	objset_t *os = zv->zv_objset;

	char *data = (char *)(lr + 1);	/* data follows lr_write_t */

	uint64_t off = lr->lr_offset;

	uint64_t len = lr->lr_length;

	dmu_tx_t *tx;

	int error;

	if (byteswap)

		byteswap_uint64_array(lr, sizeof (*lr));

	tx = dmu_tx_create(os);

	dmu_tx_hold_write(tx, ZVOL_OBJ, off, len);

	error = dmu_tx_assign(tx, zv->zv_txg_assign);

	if (error) {

		dmu_tx_abort(tx);

	} else {

		dmu_write(os, ZVOL_OBJ, off, len, data, tx);

		dmu_tx_commit(tx);

	}

	return (error);

}

/* ARGSUSED */

static int

zvol_replay_err(zvol_state_t *zv, lr_t *lr, boolean_t byteswap)

{

	return (ENOTSUP);

}

/*

 * Callback vectors for replaying records.

 * Only TX_WRITE is needed for zvol.

 */

zil_replay_func_t *zvol_replay_vector[TX_MAX_TYPE] = {

	zvol_replay_err,	/* 0 no such transaction type */

	zvol_replay_err,	/* TX_CREATE */

	zvol_replay_err,	/* TX_MKDIR */

	zvol_replay_err,	/* TX_MKXATTR */

	zvol_replay_err,	/* TX_SYMLINK */

	zvol_replay_err,	/* TX_REMOVE */

	zvol_replay_err,	/* TX_RMDIR */

	zvol_replay_err,	/* TX_LINK */

	zvol_replay_err,	/* TX_RENAME */

	zvol_replay_write,	/* TX_WRITE */

	zvol_replay_err,	/* TX_TRUNCATE */

	zvol_replay_err,	/* TX_SETATTR */

	zvol_replay_err,	/* TX_ACL */

};

/*

 * reconstruct dva that gets us to the desired offset (offset

 * is in bytes)

 */

int

zvol_get_dva(zvol_state_t *zv, uint64_t offset, dva_t *dva)

{

	zvol_ext_list_t	*zl;

	zvol_extent_t	*ze;

	int		idx;

	uint64_t	tmp;

	if ((zl = zv->zv_list) == NULL)

		return (EIO);

	idx = 0;

	ze =  &zl->zl_extents[0];

	while (offset >= ze->ze_size * zv->zv_volblocksize) {

		offset -= ze->ze_size * zv->zv_volblocksize;

		if (idx == NUM_EXTENTS - 1) {

			/* we've reached the end of this array */

			ASSERT(zl->zl_next != NULL);

			if (zl->zl_next == NULL)