/*

 * 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 (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.

 */

/* Portions Copyright 2010 Robert Milkowski */

/*

 * 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 /dev filesystem (sdev_zvolops.c).

 * 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 <sys/zil_impl.h>

#include "zfs_namecheck.h"

void *zfsdev_state;

static char *zvol_tag = "zvol_tag";

#define	ZVOL_DUMPSIZE		"dumpsize"

/*

 * This lock protects the zfsdev_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.

 */

kmutex_t zfsdev_state_lock;

static uint32_t zvol_minors;

typedef struct zvol_extent {

	list_node_t	ze_node;

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

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

} zvol_extent_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, etc. */

	objset_t	*zv_objset;	/* objset handle */

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

	uint32_t	zv_total_opens;	/* total open count */

	zilog_t		*zv_zilog;	/* ZIL handle */

	list_t		zv_extents;	/* List of extents for dump */

	znode_t		zv_znode;	/* for range locking */

	dmu_buf_t	*zv_dbuf;	/* bonus handle */

} zvol_state_t;

/*

 * zvol specific flags

 */

#define	ZVOL_RDONLY	0x1

#define	ZVOL_DUMPIFIED	0x2

#define	ZVOL_EXCL	0x4

#define	ZVOL_WCE	0x8

/*

 * zvol maximum transfer in one DMU tx.

 */

int zvol_maxphys = DMU_MAX_ACCESS/2;

extern int zfs_set_prop_nvlist(const char *, zprop_source_t,

    nvlist_t *, nvlist_t **);

static int zvol_remove_zv(zvol_state_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(uint64_t volsize, major_t maj, minor_t min)

{

	dev_t dev = makedevice(maj, min);

	VERIFY(ddi_prop_update_int64(dev, zfs_dip,

	    "Size", volsize) == DDI_SUCCESS);

	VERIFY(ddi_prop_update_int64(dev, zfs_dip,

	    "Nblocks", lbtodb(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);

}

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);

}

static zvol_state_t *

zvol_minor_lookup(const char *name)

{

	minor_t minor;

	zvol_state_t *zv;

	ASSERT(MUTEX_HELD(&zfsdev_state_lock));

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

		zv = zfsdev_get_soft_state(minor, ZSST_ZVOL);

		if (zv == NULL)

			continue;

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

			return (zv);

	}

	return (NULL);

}

/* extent mapping arg */

struct maparg {

	zvol_state_t	*ma_zv;

	uint64_t	ma_blks;

};

/*ARGSUSED*/

static int

zvol_map_block(spa_t *spa, zilog_t *zilog, const blkptr_t *bp, arc_buf_t *pbuf,

    const zbookmark_t *zb, const dnode_phys_t *dnp, void *arg)

{

	struct maparg *ma = arg;

	zvol_extent_t *ze;

	int bs = ma->ma_zv->zv_volblocksize;

	if (bp == NULL || zb->zb_object != ZVOL_OBJ || zb->zb_level != 0)

		return (0);

	VERIFY3U(ma->ma_blks, ==, zb->zb_blkid);

	ma->ma_blks++;

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

	if (BP_IS_GANG(bp))

		return (EFRAGS);

	/*

	 * See if the block is at the end of the previous extent.

	 */

	ze = list_tail(&ma->ma_zv->zv_extents);

	if (ze &&

	    DVA_GET_VDEV(BP_IDENTITY(bp)) == DVA_GET_VDEV(&ze->ze_dva) &&

	    DVA_GET_OFFSET(BP_IDENTITY(bp)) ==

	    DVA_GET_OFFSET(&ze->ze_dva) + ze->ze_nblks * bs) {

		ze->ze_nblks++;

		return (0);

	}

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

	/* start a new extent */

	ze = kmem_zalloc(sizeof (zvol_extent_t), KM_SLEEP);

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

	ze->ze_nblks = 1;

	list_insert_tail(&ma->ma_zv->zv_extents, ze);

	return (0);

}

static void

zvol_free_extents(zvol_state_t *zv)

{

	zvol_extent_t *ze;

	while (ze = list_head(&zv->zv_extents)) {

		list_remove(&zv->zv_extents, ze);

		kmem_free(ze, sizeof (zvol_extent_t));

	}

}

static int

zvol_get_lbas(zvol_state_t *zv)

{

	objset_t *os = zv->zv_objset;

	struct maparg	ma;

	int		err;

	ma.ma_zv = zv;

	ma.ma_blks = 0;

	zvol_free_extents(zv);

	/* commit any in-flight changes before traversing the dataset */

	txg_wait_synced(dmu_objset_pool(os), 0);

	err = traverse_dataset(dmu_objset_ds(os), 0,

	    TRAVERSE_PRE | TRAVERSE_PREFETCH_METADATA, zvol_map_block, &ma);

	if (err || ma.ma_blks != (zv->zv_volsize / zv->zv_volblocksize)) {

		zvol_free_extents(zv);

		return (err ? err : EIO);

	}

	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 offset, length;

	dmu_tx_t *tx;

	int error;

	if (byteswap)

		byteswap_uint64_array(lr, sizeof (*lr));

	offset = lr->lr_offset;

	length = lr->lr_length;

	/* If it's a dmu_sync() block, write the whole block */

	if (lr->lr_common.lrc_reclen == sizeof (lr_write_t)) {

		uint64_t blocksize = BP_GET_LSIZE(&lr->lr_blkptr);

		if (length < blocksize) {

			offset -= offset % blocksize;

			length = blocksize;

		}

	}

	tx = dmu_tx_create(os);

	dmu_tx_hold_write(tx, ZVOL_OBJ, offset, length);

	error = dmu_tx_assign(tx, TXG_WAIT);

	if (error) {

		dmu_tx_abort(tx);

	} else {

		dmu_write(os, ZVOL_OBJ, offset, length, 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 */

	zvol_replay_err,	/* TX_CREATE_ACL */

	zvol_replay_err,	/* TX_CREATE_ATTR */

	zvol_replay_err,	/* TX_CREATE_ACL_ATTR */

	zvol_replay_err,	/* TX_MKDIR_ACL */

	zvol_replay_err,	/* TX_MKDIR_ATTR */

	zvol_replay_err,	/* TX_MKDIR_ACL_ATTR */

	zvol_replay_err,	/* TX_WRITE2 */

};

int

zvol_name2minor(const char *name, minor_t *minor)

{

	zvol_state_t *zv;

	mutex_enter(&zfsdev_state_lock);

	zv = zvol_minor_lookup(name);

	if (minor && zv)

		*minor = zv->zv_minor;

	mutex_exit(&zfsdev_state_lock);

	return (zv ? 0 : -1);

}

/*

 * Create a minor node (plus a whole lot more) for the specified volume.

 */

int

zvol_create_minor(const char *name)

{

	zfs_soft_state_t *zs;

	zvol_state_t *zv;

	objset_t *os;

	dmu_object_info_t doi;

	minor_t minor = 0;

	char chrbuf[30], blkbuf[30];

	int error;

	mutex_enter(&zfsdev_state_lock);

	if (zvol_minor_lookup(name) != NULL) {

		mutex_exit(&zfsdev_state_lock);

		return (EEXIST);

	}

	/* lie and say we're read-only */

	error = dmu_objset_own(name, DMU_OST_ZVOL, B_TRUE, zvol_tag, &os);

	if (error) {

		mutex_exit(&zfsdev_state_lock);

		return (error);

	}

	if ((minor = zfsdev_minor_alloc()) == 0) {

		dmu_objset_disown(os, zvol_tag);

		mutex_exit(&zfsdev_state_lock);

		return (ENXIO);

	}

	if (ddi_soft_state_zalloc(zfsdev_state, minor) != DDI_SUCCESS) {

		dmu_objset_disown(os, zvol_tag);

		mutex_exit(&zfsdev_state_lock);

		return (EAGAIN);

	}

	(void) ddi_prop_update_string(minor, zfs_dip, ZVOL_PROP_NAME,

	    (char *)name);

	(void) snprintf(chrbuf, sizeof (chrbuf), "%u,raw", minor);

	if (ddi_create_minor_node(zfs_dip, chrbuf, S_IFCHR,

	    minor, DDI_PSEUDO, 0) == DDI_FAILURE) {

		ddi_soft_state_free(zfsdev_state, minor);

		dmu_objset_disown(os, zvol_tag);

		mutex_exit(&zfsdev_state_lock);

		return (EAGAIN);

	}

	(void) snprintf(blkbuf, sizeof (blkbuf), "%u", minor);

	if (ddi_create_minor_node(zfs_dip, blkbuf, S_IFBLK,

	    minor, DDI_PSEUDO, 0) == DDI_FAILURE) {

		ddi_remove_minor_node(zfs_dip, chrbuf);

		ddi_soft_state_free(zfsdev_state, minor);

		dmu_objset_disown(os, zvol_tag);

		mutex_exit(&zfsdev_state_lock);

		return (EAGAIN);

	}