/*

 * 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 2007 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/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 "zfs_namecheck.h"

#define	ZVOL_OBJ		1ULL

#define	ZVOL_ZAP_OBJ		2ULL

static void *zvol_state;

/*

 * 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;

/*

 * 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_readonly;	/* hard readonly; like write-protect */

	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 */

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

	znode_t		zv_znode;	/* for range locking */

} zvol_state_t;

/*

 * zvol maximum transfer in one DMU tx.

 */

int zvol_maxphys = DMU_MAX_ACCESS/2;

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

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);

}

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;

	zv->zv_readonly = (uint8_t)newval;

}

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);

}

/* ARGSUSED */

void

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

{

	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 properites 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 */

};

/*

 * Create a minor node for the specified volume.

 */

int

zvol_create_minor(const char *name, major_t maj)

{

	zvol_state_t *zv;

	objset_t *os;

	dmu_object_info_t doi;

	uint64_t volsize;

	minor_t minor = 0;

	struct pathname linkpath;

	int ds_mode = DS_MODE_PRIMARY;

	vnode_t *vp = NULL;

	char *devpath;

	size_t devpathlen = strlen(ZVOL_FULL_DEV_DIR) + 1 + strlen(name) + 1;

	char chrbuf[30], blkbuf[30];

	int error;

	mutex_enter(&zvol_state_lock);

	if ((zv = zvol_minor_lookup(name)) != NULL) {

		mutex_exit(&zvol_state_lock);

		return (EEXIST);

	}

	if (strchr(name, '@') != 0)

		ds_mode |= DS_MODE_READONLY;

	error = dmu_objset_open(name, DMU_OST_ZVOL, ds_mode, &os);

	if (error) {

		mutex_exit(&zvol_state_lock);

		return (error);

	}

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

	if (error) {

		dmu_objset_close(os);

		mutex_exit(&zvol_state_lock);

		return (error);

	}

	/*

	 * If there's an existing /dev/zvol symlink, try to use the

	 * same minor number we used last time.

	 */

	devpath = kmem_alloc(devpathlen, KM_SLEEP);

	(void) sprintf(devpath, "%s/%s", ZVOL_FULL_DEV_DIR, name);

	error = lookupname(devpath, UIO_SYSSPACE, NO_FOLLOW, NULL, &vp);

	kmem_free(devpath, devpathlen);

	if (error == 0 && vp->v_type != VLNK)

		error = EINVAL;

	if (error == 0) {

		pn_alloc(&linkpath);

		error = pn_getsymlink(vp, &linkpath, kcred);

		if (error == 0) {

			char *ms = strstr(linkpath.pn_path, ZVOL_PSEUDO_DEV);

			if (ms != NULL) {

				ms += strlen(ZVOL_PSEUDO_DEV);

				minor = stoi(&ms);

			}

		}

		pn_free(&linkpath);

	}

	if (vp != NULL)

		VN_RELE(vp);

	/*

	 * If we found a minor but it's already in use, we must pick a new one.

	 */

	if (minor != 0 && ddi_get_soft_state(zvol_state, minor) != NULL)

		minor = 0;

	if (minor == 0)

		minor = zvol_minor_alloc();

	if (minor == 0) {

		dmu_objset_close(os);

		mutex_exit(&zvol_state_lock);

		return (ENXIO);

	}

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

		dmu_objset_close(os);

		mutex_exit(&zvol_state_lock);

		return (EAGAIN);

	}

	(void) ddi_prop_update_string(minor, zfs_dip, ZVOL_PROP_NAME,

	    (char *)name);

	(void) sprintf(chrbuf, "%uc,raw", minor);

	if (ddi_create_minor_node(zfs_dip, chrbuf, S_IFCHR,

	    minor, DDI_PSEUDO, 0) == DDI_FAILURE) {

		ddi_soft_state_free(zvol_state, minor);

		dmu_objset_close(os);

		mutex_exit(&zvol_state_lock);

		return (EAGAIN);

	}

	(void) sprintf(blkbuf, "%uc", 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(zvol_state, minor);

		dmu_objset_close(os);

		mutex_exit(&zvol_state_lock);

		return (EAGAIN);

	}

	zv = ddi_get_soft_state(zvol_state, minor);

	(void) strcpy(zv->zv_name, name);

	zv->zv_min_bs = DEV_BSHIFT;

	zv->zv_minor = minor;

	zv->zv_volsize = volsize;

	zv->zv_objset = os;

	zv->zv_mode = ds_mode;

	zv->zv_zilog = zil_open(os, zvol_get_data);

	mutex_init(&zv->zv_znode.z_range_lock, NULL, MUTEX_DEFAULT, NULL);

	avl_create(&zv->zv_znode.z_range_avl, zfs_range_compare,

	    sizeof (rl_t), offsetof(rl_t, r_node));

	/* get and cache the blocksize */

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

	ASSERT(error == 0);

	zv->zv_volblocksize = doi.doi_data_block_size;

	zil_replay(os, zv, &zv->zv_txg_assign, zvol_replay_vector);

	zvol_size_changed(zv, maj);

	/* XXX this should handle the possible i/o error */

	VERIFY(dsl_prop_register(dmu_objset_ds(zv->zv_objset),

	    "readonly", zvol_readonly_changed_cb, zv) == 0);

	zvol_minors++;

	mutex_exit(&zvol_state_lock);

	return (0);

}

/*

 * Remove minor node for the specified volume.

 */

int

zvol_remove_minor(const char *name)

{

	zvol_state_t *zv;

	char namebuf[30];

	mutex_enter(&zvol_state_lock);

	if ((zv = zvol_minor_lookup(name)) == NULL) {

		mutex_exit(&zvol_state_lock);

		return (ENXIO);

	}

	if (zv->zv_total_opens != 0) {

		mutex_exit(&zvol_state_lock);

		return (EBUSY);

	}

	(void) sprintf(namebuf, "%uc,raw", zv->zv_minor);

	ddi_remove_minor_node(zfs_dip, namebuf);

	(void) sprintf(namebuf, "%uc", zv->zv_minor);

	ddi_remove_minor_node(zfs_dip, namebuf);

	VERIFY(dsl_prop_unregister(dmu_objset_ds(zv->zv_objset),

	    "readonly", zvol_readonly_changed_cb, zv) == 0);

	zil_close(zv->zv_zilog);

	zv->zv_zilog = NULL;

	dmu_objset_close(zv->zv_objset);

	zv->zv_objset = NULL;

	avl_destroy(&zv->zv_znode.z_range_avl);

	mutex_destroy(&zv->zv_znode.z_range_lock);

	ddi_soft_state_free(zvol_state, zv->zv_minor);

	zvol_minors--;

	mutex_exit(&zvol_state_lock);

	return (0);

}