usr/src/uts/common/io/cmlb.c

/*
 * 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 2012 DEY Storage Systems, Inc.  All rights reserved.
 * Copyright 2010 Sun Microsystems, Inc.  All rights reserved.
 * Use is subject to license terms.
 * Copyright 2016 Toomas Soome <tsoome@me.com>
 */

/*
 * This module provides support for labeling operations for target
 * drivers.
 */

#include <sys/scsi/scsi.h>
#include <sys/sunddi.h>
#include <sys/dklabel.h>
#include <sys/dkio.h>
#include <sys/vtoc.h>
#include <sys/dktp/fdisk.h>
#include <sys/vtrace.h>
#include <sys/efi_partition.h>
#include <sys/cmlb.h>
#include <sys/cmlb_impl.h>
#if defined(__i386) || defined(__amd64)
#include <sys/fs/dv_node.h>
#endif
#include <sys/ddi_impldefs.h>

/*
 * Driver minor node structure and data table
 */
struct driver_minor_data {
	char	*name;
	minor_t	minor;
	int	type;
};

static struct driver_minor_data dk_minor_data[] = {
	{"a", 0, S_IFBLK},
	{"b", 1, S_IFBLK},
	{"c", 2, S_IFBLK},
	{"d", 3, S_IFBLK},
	{"e", 4, S_IFBLK},
	{"f", 5, S_IFBLK},
	{"g", 6, S_IFBLK},
	{"h", 7, S_IFBLK},
#if defined(_SUNOS_VTOC_16)
	{"i", 8, S_IFBLK},
	{"j", 9, S_IFBLK},
	{"k", 10, S_IFBLK},
	{"l", 11, S_IFBLK},
	{"m", 12, S_IFBLK},
	{"n", 13, S_IFBLK},
	{"o", 14, S_IFBLK},
	{"p", 15, S_IFBLK},
#endif			/* defined(_SUNOS_VTOC_16) */
#if defined(_FIRMWARE_NEEDS_FDISK)
	{"q", 16, S_IFBLK},
	{"r", 17, S_IFBLK},
	{"s", 18, S_IFBLK},
	{"t", 19, S_IFBLK},
	{"u", 20, S_IFBLK},
#endif			/* defined(_FIRMWARE_NEEDS_FDISK) */
	{"a,raw", 0, S_IFCHR},
	{"b,raw", 1, S_IFCHR},
	{"c,raw", 2, S_IFCHR},
	{"d,raw", 3, S_IFCHR},
	{"e,raw", 4, S_IFCHR},
	{"f,raw", 5, S_IFCHR},
	{"g,raw", 6, S_IFCHR},
	{"h,raw", 7, S_IFCHR},
#if defined(_SUNOS_VTOC_16)
	{"i,raw", 8, S_IFCHR},
	{"j,raw", 9, S_IFCHR},
	{"k,raw", 10, S_IFCHR},
	{"l,raw", 11, S_IFCHR},
	{"m,raw", 12, S_IFCHR},
	{"n,raw", 13, S_IFCHR},
	{"o,raw", 14, S_IFCHR},
	{"p,raw", 15, S_IFCHR},
#endif			/* defined(_SUNOS_VTOC_16) */
#if defined(_FIRMWARE_NEEDS_FDISK)
	{"q,raw", 16, S_IFCHR},
	{"r,raw", 17, S_IFCHR},
	{"s,raw", 18, S_IFCHR},
	{"t,raw", 19, S_IFCHR},
	{"u,raw", 20, S_IFCHR},
#endif			/* defined(_FIRMWARE_NEEDS_FDISK) */
	{0}
};

#if defined(__i386) || defined(__amd64)
#if defined(_FIRMWARE_NEEDS_FDISK)
static struct driver_minor_data dk_ext_minor_data[] = {
	{"p5", 21, S_IFBLK},
	{"p6", 22, S_IFBLK},
	{"p7", 23, S_IFBLK},
	{"p8", 24, S_IFBLK},
	{"p9", 25, S_IFBLK},
	{"p10", 26, S_IFBLK},
	{"p11", 27, S_IFBLK},
	{"p12", 28, S_IFBLK},
	{"p13", 29, S_IFBLK},
	{"p14", 30, S_IFBLK},
	{"p15", 31, S_IFBLK},
	{"p16", 32, S_IFBLK},
	{"p17", 33, S_IFBLK},
	{"p18", 34, S_IFBLK},
	{"p19", 35, S_IFBLK},
	{"p20", 36, S_IFBLK},
	{"p21", 37, S_IFBLK},
	{"p22", 38, S_IFBLK},
	{"p23", 39, S_IFBLK},
	{"p24", 40, S_IFBLK},
	{"p25", 41, S_IFBLK},
	{"p26", 42, S_IFBLK},
	{"p27", 43, S_IFBLK},
	{"p28", 44, S_IFBLK},
	{"p29", 45, S_IFBLK},
	{"p30", 46, S_IFBLK},
	{"p31", 47, S_IFBLK},
	{"p32", 48, S_IFBLK},
	{"p33", 49, S_IFBLK},
	{"p34", 50, S_IFBLK},
	{"p35", 51, S_IFBLK},
	{"p36", 52, S_IFBLK},
	{"p5,raw", 21, S_IFCHR},
	{"p6,raw", 22, S_IFCHR},
	{"p7,raw", 23, S_IFCHR},
	{"p8,raw", 24, S_IFCHR},
	{"p9,raw", 25, S_IFCHR},
	{"p10,raw", 26, S_IFCHR},
	{"p11,raw", 27, S_IFCHR},
	{"p12,raw", 28, S_IFCHR},
	{"p13,raw", 29, S_IFCHR},
	{"p14,raw", 30, S_IFCHR},
	{"p15,raw", 31, S_IFCHR},
	{"p16,raw", 32, S_IFCHR},
	{"p17,raw", 33, S_IFCHR},
	{"p18,raw", 34, S_IFCHR},
	{"p19,raw", 35, S_IFCHR},
	{"p20,raw", 36, S_IFCHR},
	{"p21,raw", 37, S_IFCHR},
	{"p22,raw", 38, S_IFCHR},
	{"p23,raw", 39, S_IFCHR},
	{"p24,raw", 40, S_IFCHR},
	{"p25,raw", 41, S_IFCHR},
	{"p26,raw", 42, S_IFCHR},
	{"p27,raw", 43, S_IFCHR},
	{"p28,raw", 44, S_IFCHR},
	{"p29,raw", 45, S_IFCHR},
	{"p30,raw", 46, S_IFCHR},
	{"p31,raw", 47, S_IFCHR},
	{"p32,raw", 48, S_IFCHR},
	{"p33,raw", 49, S_IFCHR},
	{"p34,raw", 50, S_IFCHR},
	{"p35,raw", 51, S_IFCHR},
	{"p36,raw", 52, S_IFCHR},
	{0}
};
#endif			/* defined(_FIRMWARE_NEEDS_FDISK) */
#endif			/* if defined(__i386) || defined(__amd64) */

static struct driver_minor_data dk_minor_data_efi[] = {
	{"a", 0, S_IFBLK},
	{"b", 1, S_IFBLK},
	{"c", 2, S_IFBLK},
	{"d", 3, S_IFBLK},
	{"e", 4, S_IFBLK},
	{"f", 5, S_IFBLK},
	{"g", 6, S_IFBLK},
	{"wd", 7, S_IFBLK},
#if defined(_SUNOS_VTOC_16)
	{"i", 8, S_IFBLK},
	{"j", 9, S_IFBLK},
	{"k", 10, S_IFBLK},
	{"l", 11, S_IFBLK},
	{"m", 12, S_IFBLK},
	{"n", 13, S_IFBLK},
	{"o", 14, S_IFBLK},
	{"p", 15, S_IFBLK},
#endif			/* defined(_SUNOS_VTOC_16) */
#if defined(_FIRMWARE_NEEDS_FDISK)
	{"q", 16, S_IFBLK},
	{"r", 17, S_IFBLK},
	{"s", 18, S_IFBLK},
	{"t", 19, S_IFBLK},
	{"u", 20, S_IFBLK},
#endif			/* defined(_FIRMWARE_NEEDS_FDISK) */
	{"a,raw", 0, S_IFCHR},
	{"b,raw", 1, S_IFCHR},
	{"c,raw", 2, S_IFCHR},
	{"d,raw", 3, S_IFCHR},
	{"e,raw", 4, S_IFCHR},
	{"f,raw", 5, S_IFCHR},
	{"g,raw", 6, S_IFCHR},
	{"wd,raw", 7, S_IFCHR},
#if defined(_SUNOS_VTOC_16)
	{"i,raw", 8, S_IFCHR},
	{"j,raw", 9, S_IFCHR},
	{"k,raw", 10, S_IFCHR},
	{"l,raw", 11, S_IFCHR},
	{"m,raw", 12, S_IFCHR},
	{"n,raw", 13, S_IFCHR},
	{"o,raw", 14, S_IFCHR},
	{"p,raw", 15, S_IFCHR},
#endif			/* defined(_SUNOS_VTOC_16) */
#if defined(_FIRMWARE_NEEDS_FDISK)
	{"q,raw", 16, S_IFCHR},
	{"r,raw", 17, S_IFCHR},
	{"s,raw", 18, S_IFCHR},
	{"t,raw", 19, S_IFCHR},
	{"u,raw", 20, S_IFCHR},
#endif			/* defined(_FIRMWARE_NEEDS_FDISK) */
	{0}
};

/*
 * Declare the dynamic properties implemented in prop_op(9E) implementation
 * that we want to have show up in a di_init(3DEVINFO) device tree snapshot
 * of drivers that call cmlb_attach().
 */
static i_ddi_prop_dyn_t cmlb_prop_dyn[] = {
	{"Nblocks",		DDI_PROP_TYPE_INT64,	S_IFBLK},
	{"Size",		DDI_PROP_TYPE_INT64,	S_IFCHR},
	{"device-nblocks",	DDI_PROP_TYPE_INT64},
	{"device-blksize",	DDI_PROP_TYPE_INT},
	{"device-solid-state",	DDI_PROP_TYPE_INT},
	{"device-rotational",	DDI_PROP_TYPE_INT},
	{NULL}
};

/*
 * This implies an upper limit of 8192 GPT partitions
 * in one transfer for GUID Partition Entry Array.
 */
len_t cmlb_tg_max_efi_xfer = 1024 * 1024;

/*
 * External kernel interfaces
 */
extern struct mod_ops mod_miscops;

extern int ddi_create_internal_pathname(dev_info_t *dip, char *name,
    int spec_type, minor_t minor_num);

/*
 * Global buffer and mutex for debug logging
 */
static char	cmlb_log_buffer[1024];
static kmutex_t	cmlb_log_mutex;


struct cmlb_lun *cmlb_debug_cl = NULL;
uint_t cmlb_level_mask = 0x0;

int cmlb_rot_delay = 4;	/* default rotational delay */

static struct modlmisc modlmisc = {
	&mod_miscops,   /* Type of module */
	"Common Labeling module"
};

static struct modlinkage modlinkage = {
	MODREV_1, (void *)&modlmisc, NULL
};

/* Local function prototypes */
static dev_t cmlb_make_device(struct cmlb_lun *cl);
static int cmlb_validate_geometry(struct cmlb_lun *cl, boolean_t forcerevalid,
    int flags, void *tg_cookie);
static void cmlb_resync_geom_caches(struct cmlb_lun *cl, diskaddr_t capacity,
    void *tg_cookie);
static int cmlb_read_fdisk(struct cmlb_lun *cl, diskaddr_t capacity,
    void *tg_cookie);
static void cmlb_swap_efi_gpt(efi_gpt_t *e);
static void cmlb_swap_efi_gpe(int nparts, efi_gpe_t *p);
static int cmlb_validate_efi(efi_gpt_t *labp);
static int cmlb_use_efi(struct cmlb_lun *cl, diskaddr_t capacity, int flags,
    void *tg_cookie);
static void cmlb_build_default_label(struct cmlb_lun *cl, void *tg_cookie);
static int  cmlb_uselabel(struct cmlb_lun *cl,  struct dk_label *l, int flags);
#if defined(_SUNOS_VTOC_8)
static void cmlb_build_user_vtoc(struct cmlb_lun *cl, struct vtoc *user_vtoc);
#endif
static int cmlb_build_label_vtoc(struct cmlb_lun *cl, struct vtoc *user_vtoc);
static int cmlb_write_label(struct cmlb_lun *cl, void *tg_cookie);
static int cmlb_set_vtoc(struct cmlb_lun *cl, struct dk_label *dkl,
    void *tg_cookie);
static void cmlb_clear_efi(struct cmlb_lun *cl, void *tg_cookie);
static void cmlb_clear_vtoc(struct cmlb_lun *cl, void *tg_cookie);
static void cmlb_setup_default_geometry(struct cmlb_lun *cl, void *tg_cookie);
static int cmlb_create_minor_nodes(struct cmlb_lun *cl);
static int cmlb_check_update_blockcount(struct cmlb_lun *cl, void *tg_cookie);
static boolean_t cmlb_check_efi_mbr(uchar_t *buf, boolean_t *is_mbr);

#if defined(__i386) || defined(__amd64)
static int cmlb_update_fdisk_and_vtoc(struct cmlb_lun *cl, void *tg_cookie);
#endif

#if defined(_FIRMWARE_NEEDS_FDISK)
static boolean_t  cmlb_has_max_chs_vals(struct ipart *fdp);
#endif

#if defined(_SUNOS_VTOC_16)
static void cmlb_convert_geometry(struct cmlb_lun *cl, diskaddr_t capacity,
    struct dk_geom *cl_g, void *tg_cookie);
#endif

static int cmlb_dkio_get_geometry(struct cmlb_lun *cl, caddr_t arg, int flag,
    void *tg_cookie);
static int cmlb_dkio_set_geometry(struct cmlb_lun *cl, caddr_t arg, int flag);
static int cmlb_dkio_get_partition(struct cmlb_lun *cl, caddr_t arg, int flag,
    void *tg_cookie);
static int cmlb_dkio_set_partition(struct cmlb_lun *cl, caddr_t arg, int flag);
static int cmlb_dkio_get_efi(struct cmlb_lun *cl, caddr_t arg, int flag,
    void *tg_cookie);
static int cmlb_dkio_set_efi(struct cmlb_lun *cl, dev_t dev, caddr_t arg,
    int flag, void *tg_cookie);
static int cmlb_dkio_get_vtoc(struct cmlb_lun *cl, caddr_t arg, int flag,
    void *tg_cookie);
static int cmlb_dkio_get_extvtoc(struct cmlb_lun *cl, caddr_t arg, int flag,
    void *tg_cookie);
static int cmlb_dkio_set_vtoc(struct cmlb_lun *cl, dev_t dev, caddr_t arg,
    int flag, void *tg_cookie);
static int cmlb_dkio_set_extvtoc(struct cmlb_lun *cl, dev_t dev, caddr_t arg,
    int flag, void *tg_cookie);
static int cmlb_dkio_get_mboot(struct cmlb_lun *cl, caddr_t arg, int flag,
    void *tg_cookie);
static int cmlb_dkio_set_mboot(struct cmlb_lun *cl, caddr_t arg, int flag,
    void *tg_cookie);
static int cmlb_dkio_partition(struct cmlb_lun *cl, caddr_t arg, int flag,
    void *tg_cookie);

#if defined(__i386) || defined(__amd64)
static int cmlb_dkio_set_ext_part(struct cmlb_lun *cl, caddr_t arg, int flag,
    void *tg_cookie);
static int cmlb_validate_ext_part(struct cmlb_lun *cl, int part, int epart,
    uint32_t start, uint32_t size);
static int cmlb_is_linux_swap(struct cmlb_lun *cl, uint32_t part_start,
    void *tg_cookie);
static int cmlb_dkio_get_virtgeom(struct cmlb_lun *cl, caddr_t arg, int flag);
static int cmlb_dkio_get_phygeom(struct cmlb_lun *cl, caddr_t  arg, int flag,
    void *tg_cookie);
static int cmlb_dkio_partinfo(struct cmlb_lun *cl, dev_t dev, caddr_t arg,
    int flag);
static int cmlb_dkio_extpartinfo(struct cmlb_lun *cl, dev_t dev, caddr_t arg,
    int flag);
#endif

static void cmlb_dbg(uint_t comp, struct cmlb_lun *cl, const char *fmt, ...);
static void cmlb_v_log(dev_info_t *dev, const char *label, uint_t level,
    const char *fmt, va_list ap);
static void cmlb_log(dev_info_t *dev, const char *label, uint_t level,
    const char *fmt, ...);

int
_init(void)
{
	mutex_init(&cmlb_log_mutex, NULL, MUTEX_DRIVER, NULL);
	return (mod_install(&modlinkage));
}

int
_info(struct modinfo *modinfop)
{
	return (mod_info(&modlinkage, modinfop));
}

int
_fini(void)
{
	int err;

	if ((err = mod_remove(&modlinkage)) != 0) {
		return (err);
	}

	mutex_destroy(&cmlb_log_mutex);
	return (err);
}

/*
 * cmlb_dbg is used for debugging to log additional info
 * Level of output is controlled via cmlb_level_mask setting.
 */
static void
cmlb_dbg(uint_t comp, struct cmlb_lun *cl, const char *fmt, ...)
{
	va_list		ap;
	dev_info_t	*dev;
	uint_t		level_mask = 0;

	ASSERT(cl != NULL);
	dev = CMLB_DEVINFO(cl);
	ASSERT(dev != NULL);
	/*
	 * Filter messages based on the global component and level masks,
	 * also print if cl matches the value of cmlb_debug_cl, or if
	 * cmlb_debug_cl is set to NULL.
	 */
	if (comp & CMLB_TRACE)
		level_mask |= CMLB_LOGMASK_TRACE;

	if (comp & CMLB_INFO)
		level_mask |= CMLB_LOGMASK_INFO;

	if (comp & CMLB_ERROR)
		level_mask |= CMLB_LOGMASK_ERROR;

	if ((cmlb_level_mask & level_mask) &&
	    ((cmlb_debug_cl == NULL) || (cmlb_debug_cl == cl))) {
		va_start(ap, fmt);
		cmlb_v_log(dev, CMLB_LABEL(cl), CE_CONT, fmt, ap);
		va_end(ap);
	}
}

/*
 * cmlb_log is basically a duplicate of scsi_log. It is redefined here
 * so that this module does not depend on scsi module.
 */
static void
cmlb_log(dev_info_t *dev, const char *label, uint_t level, const char *fmt, ...)
{
	va_list		ap;

	va_start(ap, fmt);
	cmlb_v_log(dev, label, level, fmt, ap);
	va_end(ap);
}

static void
cmlb_v_log(dev_info_t *dev, const char *label, uint_t level, const char *fmt,
    va_list ap)
{
	static char 	name[256];
	int 		log_only = 0;
	int 		boot_only = 0;
	int 		console_only = 0;

	mutex_enter(&cmlb_log_mutex);

	if (dev) {
		if (level == CE_PANIC || level == CE_WARN ||
		    level == CE_NOTE) {
			(void) sprintf(name, "%s (%s%d):\n",
			    ddi_pathname(dev, cmlb_log_buffer),
			    label, ddi_get_instance(dev));
		} else {
			name[0] = '\0';
		}
	} else {
		(void) sprintf(name, "%s:", label);
	}

	(void) vsprintf(cmlb_log_buffer, fmt, ap);

	switch (cmlb_log_buffer[0]) {
	case '!':
		log_only = 1;
		break;
	case '?':
		boot_only = 1;
		break;
	case '^':
		console_only = 1;
		break;
	}

	switch (level) {
	case CE_NOTE:
		level = CE_CONT;
		/* FALLTHROUGH */
	case CE_CONT:
	case CE_WARN:
	case CE_PANIC:
		if (boot_only) {
			cmn_err(level, "?%s\t%s", name, &cmlb_log_buffer[1]);
		} else if (console_only) {
			cmn_err(level, "^%s\t%s", name, &cmlb_log_buffer[1]);
		} else if (log_only) {
			cmn_err(level, "!%s\t%s", name, &cmlb_log_buffer[1]);
		} else {
			cmn_err(level, "%s\t%s", name, cmlb_log_buffer);
		}
		break;
	case CE_IGNORE:
		break;
	default:
		cmn_err(CE_CONT, "^DEBUG: %s\t%s", name, cmlb_log_buffer);
		break;
	}
	mutex_exit(&cmlb_log_mutex);
}


/*
 * cmlb_alloc_handle:
 *
 *	Allocates a handle.
 *
 * Arguments:
 *	cmlbhandlep	pointer to handle
 *
 * Notes:
 *	Allocates a handle and stores the allocated handle in the area
 *	pointed to by cmlbhandlep
 *
 * Context:
 *	Kernel thread only (can sleep).
 */
void
cmlb_alloc_handle(cmlb_handle_t *cmlbhandlep)
{
	struct cmlb_lun 	*cl;

	cl = kmem_zalloc(sizeof (struct cmlb_lun), KM_SLEEP);
	ASSERT(cmlbhandlep != NULL);

	cl->cl_state = CMLB_INITED;
	cl->cl_def_labeltype = CMLB_LABEL_UNDEF;
	mutex_init(CMLB_MUTEX(cl), NULL, MUTEX_DRIVER, NULL);

	*cmlbhandlep = (cmlb_handle_t)(cl);
}

/*
 * cmlb_free_handle
 *
 *	Frees handle.
 *
 * Arguments:
 *	cmlbhandlep	pointer to handle
 */
void
cmlb_free_handle(cmlb_handle_t *cmlbhandlep)
{
	struct cmlb_lun 	*cl;

	cl = (struct cmlb_lun *)*cmlbhandlep;
	if (cl != NULL) {
		mutex_destroy(CMLB_MUTEX(cl));
		kmem_free(cl, sizeof (struct cmlb_lun));
	}

}

/*
 * cmlb_attach:
 *
 *	Attach handle to device, create minor nodes for device.
 *
 * Arguments:
 * 	devi		pointer to device's dev_info structure.
 * 	tgopsp		pointer to array of functions cmlb can use to callback
 *			to target driver.
 *
 *	device_type	Peripheral device type as defined in
 *			scsi/generic/inquiry.h
 *
 *	is_removable	whether or not device is removable.
 *
 *	is_hotpluggable	whether or not device is hotpluggable.
 *
 *	node_type	minor node type (as used by ddi_create_minor_node)
 *
 *	alter_behavior
 *			bit flags:
 *
 *			CMLB_CREATE_ALTSLICE_VTOC_16_DTYPE_DIRECT: create
 *			an alternate slice for the default label, if
 *			device type is DTYPE_DIRECT an architectures default
 *			label type is VTOC16.
 *			Otherwise alternate slice will no be created.
 *
 *
 *			CMLB_FAKE_GEOM_LABEL_IOCTLS_VTOC8: report a default
 *			geometry and label for DKIOCGGEOM and DKIOCGVTOC
 *			on architecture with VTOC8 label types.
 *
 * 			CMLB_OFF_BY_ONE: do the workaround for legacy off-by-
 *                      one bug in obtaining capacity (in sd):
 *			SCSI READ_CAPACITY command returns the LBA number of the
 *			last logical block, but sd once treated this number as
 *			disks' capacity on x86 platform. And LBAs are addressed
 *			based 0. So the last block was lost on x86 platform.
 *
 *			Now, we remove this workaround. In order for present sd
 *			driver to work with disks which are labeled/partitioned
 *			via previous sd, we add workaround as follows:
 *
 *			1) Locate backup EFI label: cmlb searches the next to
 *			   last
 *			   block for backup EFI label. If fails, it will
 *			   turn to the last block for backup EFI label;
 *
 *			2) Clear backup EFI label: cmlb first search the last
 *			   block for backup EFI label, and will search the
 *			   next to last block only if failed for the last
 *			   block.
 *
 *			3) Calculate geometry:refer to cmlb_convert_geometry()
 *			   If capacity increasing by 1 causes disks' capacity
 *			   to cross over the limits in geometry calculation,
 *			   geometry info will change. This will raise an issue:
 *			   In case that primary VTOC label is destroyed, format
 *			   commandline can restore it via backup VTOC labels.
 *			   And format locates backup VTOC labels by use of
 *			   geometry. So changing geometry will
 *			   prevent format from finding backup VTOC labels. To
 *			   eliminate this side effect for compatibility,
 *			   sd uses (capacity -1) to calculate geometry;
 *
 *			4) 1TB disks: some important data structures use
 *			   32-bit signed long/int (for example, daddr_t),
 *			   so that sd doesn't support a disk with capacity
 *			   larger than 1TB on 32-bit platform. However,
 *			   for exactly 1TB disk, it was treated as (1T - 512)B
 *			   in the past, and could have valid Solaris
 *			   partitions. To workaround this, if an exactly 1TB
 *			   disk has Solaris fdisk partition, it will be allowed
 *			   to work with sd.
 *
 *
 *
 *			CMLB_FAKE_LABEL_ONE_PARTITION: create s0 and s2 covering
 *			the entire disk, if there is no valid partition info.
 *			If there is a valid Solaris partition, s0 and s2 will
 *			only cover the entire Solaris partition.
 *
 *			CMLB_CREATE_P0_MINOR_NODE: create p0 node covering
 *			the entire disk. Used by lofi to ensure presence of
 *			whole disk device node in case of LOFI_MAP_FILE ioctl.
 *
 *	cmlbhandle	cmlb handle associated with device
 *
 *	tg_cookie	cookie from target driver to be passed back to target
 *			driver when we call back to it through tg_ops.
 *
 * Notes:
 *	Assumes a default label based on capacity for non-removable devices.
 *	If capacity > 1TB, EFI is assumed otherwise VTOC (default VTOC
 *	for the architecture).
 *
 *	For removable devices, default label type is assumed to be VTOC
 *	type. Create minor nodes based on a default label type.
 *	Label on the media is not validated.
 *	minor number consists of:
 *		if _SUNOS_VTOC_8 is defined
 *			lowest 3 bits is taken as partition number
 *			the rest is instance number
 *		if _SUNOS_VTOC_16 is defined
 *			lowest 6 bits is taken as partition number
 *			the rest is instance number
 *
 *
 * Return values:
 *	0 	Success
 * 	ENXIO 	creating minor nodes failed.
 *	EINVAL  invalid arg, unsupported tg_ops version
 */
int
cmlb_attach(dev_info_t *devi, cmlb_tg_ops_t *tgopsp, int device_type,
    boolean_t is_removable, boolean_t is_hotpluggable, char *node_type,
    int alter_behavior, cmlb_handle_t cmlbhandle, void *tg_cookie)
{

	struct cmlb_lun	*cl = (struct cmlb_lun *)cmlbhandle;
	diskaddr_t	cap;
	int		status;

	ASSERT(VALID_BOOLEAN(is_removable));
	ASSERT(VALID_BOOLEAN(is_hotpluggable));

	if (tgopsp->tg_version < TG_DK_OPS_VERSION_1)
		return (EINVAL);

	mutex_enter(CMLB_MUTEX(cl));

	CMLB_DEVINFO(cl) = devi;
	cl->cmlb_tg_ops = tgopsp;
	cl->cl_device_type = device_type;
	cl->cl_is_removable = is_removable;
	cl->cl_is_hotpluggable = is_hotpluggable;
	cl->cl_node_type = node_type;
	cl->cl_sys_blocksize = DEV_BSIZE;
	cl->cl_f_geometry_is_valid = B_FALSE;
	cl->cl_def_labeltype = CMLB_LABEL_VTOC;
	cl->cl_alter_behavior = alter_behavior;
	cl->cl_reserved = -1;
	cl->cl_msglog_flag |= CMLB_ALLOW_2TB_WARN;
#if defined(__i386) || defined(__amd64)
	cl->cl_logical_drive_count = 0;
#endif

	if (!is_removable) {
		mutex_exit(CMLB_MUTEX(cl));
		status = DK_TG_GETCAP(cl, &cap, tg_cookie);
		mutex_enter(CMLB_MUTEX(cl));
		if (status == 0 && cap > CMLB_EXTVTOC_LIMIT) {
			/* set default EFI if > 2TB */
			cl->cl_def_labeltype = CMLB_LABEL_EFI;
		}
	}

	/* create minor nodes based on default label type */
	cl->cl_last_labeltype = CMLB_LABEL_UNDEF;
	cl->cl_cur_labeltype = CMLB_LABEL_UNDEF;

	if (cmlb_create_minor_nodes(cl) != 0) {
		mutex_exit(CMLB_MUTEX(cl));
		return (ENXIO);
	}

	/* Define the dynamic properties for devinfo spapshots. */
	i_ddi_prop_dyn_driver_set(CMLB_DEVINFO(cl), cmlb_prop_dyn);

	cl->cl_state = CMLB_ATTACHED;

	mutex_exit(CMLB_MUTEX(cl));
	return (0);
}

/*
 * cmlb_detach:
 *
 * Invalidate in-core labeling data and remove all minor nodes for
 * the device associate with handle.
 *
 * Arguments:
 *	cmlbhandle	cmlb handle associated with device.
 *
 *	tg_cookie	cookie from target driver to be passed back to target
 *			driver when we call back to it through tg_ops.
 *
 */
/*ARGSUSED1*/
void
cmlb_detach(cmlb_handle_t cmlbhandle, void *tg_cookie)
{
	struct cmlb_lun *cl = (struct cmlb_lun *)cmlbhandle;

	mutex_enter(CMLB_MUTEX(cl));
	cl->cl_def_labeltype = CMLB_LABEL_UNDEF;
	cl->cl_f_geometry_is_valid = B_FALSE;
	ddi_remove_minor_node(CMLB_DEVINFO(cl), NULL);
	i_ddi_prop_dyn_driver_set(CMLB_DEVINFO(cl), NULL);
	cl->cl_state = CMLB_INITED;
	mutex_exit(CMLB_MUTEX(cl));
}

/*
 * cmlb_validate:
 *
 *	Validates label.
 *
 * Arguments
 *	cmlbhandle	cmlb handle associated with device.
 *
 *	flags		operation flags. used for verbosity control
 *
 *	tg_cookie	cookie from target driver to be passed back to target
 *			driver when we call back to it through tg_ops.
 *
 *
 * Notes:
 *	If new label type is different from the current, adjust minor nodes
 *	accordingly.
 *
 * Return values:
 *	0		success
 *			Note: having fdisk but no solaris partition is assumed
 *			success.
 *
 *	ENOMEM		memory allocation failed
 *	EIO		i/o errors during read or get capacity
 * 	EACCESS		reservation conflicts
 * 	EINVAL		label was corrupt, or no default label was assumed
 *	ENXIO		invalid handle
 */
int
cmlb_validate(cmlb_handle_t cmlbhandle, int flags, void *tg_cookie)
{
	struct cmlb_lun *cl = (struct cmlb_lun *)cmlbhandle;
	int 		rval;
	int  		ret = 0;

	/*
	 * Temp work-around checking cl for NULL since there is a bug
	 * in sd_detach calling this routine from taskq_dispatch
	 * inited function.
	 */
	if (cl == NULL)
		return (ENXIO);

	mutex_enter(CMLB_MUTEX(cl));
	if (cl->cl_state < CMLB_ATTACHED) {
		mutex_exit(CMLB_MUTEX(cl));
		return (ENXIO);
	}

	rval = cmlb_validate_geometry((struct cmlb_lun *)cmlbhandle, B_TRUE,
	    flags, tg_cookie);

	if (rval == ENOTSUP) {
		if (cl->cl_f_geometry_is_valid) {
			cl->cl_cur_labeltype = CMLB_LABEL_EFI;
			ret = 0;
		} else {
			ret = EINVAL;
		}
	} else {
		ret = rval;
		if (ret == 0)
			cl->cl_cur_labeltype = CMLB_LABEL_VTOC;
	}

	if (ret == 0)
		(void) cmlb_create_minor_nodes(cl);

	mutex_exit(CMLB_MUTEX(cl));
	return (ret);
}

/*
 * cmlb_invalidate:
 *	Invalidate in core label data
 *
 * Arguments:
 *	cmlbhandle	cmlb handle associated with device.
 *	tg_cookie	cookie from target driver to be passed back to target
 *			driver when we call back to it through tg_ops.
 */
/*ARGSUSED1*/
void
cmlb_invalidate(cmlb_handle_t cmlbhandle, void *tg_cookie)
{
	struct cmlb_lun *cl = (struct cmlb_lun *)cmlbhandle;

	if (cl == NULL)
		return;

	mutex_enter(CMLB_MUTEX(cl));
	cl->cl_f_geometry_is_valid = B_FALSE;
	mutex_exit(CMLB_MUTEX(cl));
}

/*
 * cmlb_is_valid
 * 	Get status on whether the incore label/geom data is valid
 *
 * Arguments:
 *	cmlbhandle      cmlb handle associated with device.
 *
 * Return values:
 *	B_TRUE if incore label/geom data is valid.
 *	B_FALSE otherwise.
 *
 */


boolean_t
cmlb_is_valid(cmlb_handle_t cmlbhandle)
{
	struct cmlb_lun *cl = (struct cmlb_lun *)cmlbhandle;

	if (cmlbhandle == NULL)
		return (B_FALSE);

	return (cl->cl_f_geometry_is_valid);

}



/*
 * cmlb_close:
 *
 * Close the device, revert to a default label minor node for the device,
 * if it is removable.
 *
 * Arguments:
 *	cmlbhandle	cmlb handle associated with device.
 *
 *	tg_cookie	cookie from target driver to be passed back to target
 *			driver when we call back to it through tg_ops.
 * Return values:
 *	0	Success
 * 	ENXIO	Re-creating minor node failed.
 */
/*ARGSUSED1*/
int
cmlb_close(cmlb_handle_t cmlbhandle, void *tg_cookie)
{
	struct cmlb_lun *cl = (struct cmlb_lun *)cmlbhandle;

	mutex_enter(CMLB_MUTEX(cl));
	cl->cl_f_geometry_is_valid = B_FALSE;

	/* revert to default minor node for this device */
	if (ISREMOVABLE(cl)) {
		cl->cl_cur_labeltype = CMLB_LABEL_UNDEF;
		(void) cmlb_create_minor_nodes(cl);
	}

	mutex_exit(CMLB_MUTEX(cl));
	return (0);
}

/*
 * cmlb_get_devid_block:
 *	 get the block number where device id is stored.
 *
 * Arguments:
 *	cmlbhandle	cmlb handle associated with device.
 *	devidblockp	pointer to block number.
 *	tg_cookie	cookie from target driver to be passed back to target
 *			driver when we call back to it through tg_ops.
 *
 * Notes:
 *	It stores the block number of device id in the area pointed to
 *	by devidblockp.
 * 	with the block number of device id.
 *
 * Return values:
 *	0	success
 *	EINVAL 	device id does not apply to current label type.
 */
/*ARGSUSED2*/
int
cmlb_get_devid_block(cmlb_handle_t cmlbhandle, diskaddr_t *devidblockp,
    void *tg_cookie)
{
	daddr_t			spc, blk, head, cyl;
	struct cmlb_lun *cl = (struct cmlb_lun *)cmlbhandle;

	mutex_enter(CMLB_MUTEX(cl));
	if (cl->cl_state < CMLB_ATTACHED) {
		mutex_exit(CMLB_MUTEX(cl));
		return (EINVAL);
	}

	if ((!cl->cl_f_geometry_is_valid) ||
	    (cl->cl_solaris_size < DK_LABEL_LOC)) {
		mutex_exit(CMLB_MUTEX(cl));
		return (EINVAL);
	}

	if (cl->cl_cur_labeltype == CMLB_LABEL_EFI) {
		if (cl->cl_reserved != -1) {
			blk = cl->cl_map[cl->cl_reserved].dkl_cylno;
		} else {
			mutex_exit(CMLB_MUTEX(cl));
			return (EINVAL);
		}
	} else {
		/* if the disk is unlabeled, don't write a devid to it */
		if (cl->cl_label_from_media != CMLB_LABEL_VTOC) {
			mutex_exit(CMLB_MUTEX(cl));
			return (EINVAL);
		}

		/* this geometry doesn't allow us to write a devid */
		if (cl->cl_g.dkg_acyl < 2) {
			mutex_exit(CMLB_MUTEX(cl));
			return (EINVAL);
		}

		/*
		 * Subtract 2 guarantees that the next to last cylinder
		 * is used
		 */
		cyl  = cl->cl_g.dkg_ncyl  + cl->cl_g.dkg_acyl - 2;
		spc  = cl->cl_g.dkg_nhead * cl->cl_g.dkg_nsect;
		head = cl->cl_g.dkg_nhead - 1;
		blk  = cl->cl_solaris_offset +
		    (cyl * (spc - cl->cl_g.dkg_apc)) +
		    (head * cl->cl_g.dkg_nsect) + 1;
	}

	*devidblockp = blk;
	mutex_exit(CMLB_MUTEX(cl));
	return (0);
}

/*
 * cmlb_partinfo:
 *	Get partition info for specified partition number.
 *
 * Arguments:
 *	cmlbhandle	cmlb handle associated with device.
 *	part		partition number
 *	nblocksp	pointer to number of blocks
 *	startblockp	pointer to starting block
 *	partnamep	pointer to name of partition
 *	tagp		pointer to tag info
 *	tg_cookie	cookie from target driver to be passed back to target
 *			driver when we call back to it through tg_ops.
 *
 *
 * Notes:
 *	If in-core label is not valid, this functions tries to revalidate
 *	the label. If label is valid, it stores the total number of blocks
 *	in this partition in the area pointed to by nblocksp, starting
 *	block number in area pointed to by startblockp,  pointer to partition
 *	name in area pointed to by partnamep, and tag value in area
 *	pointed by tagp.
 *	For EFI labels, tag value will be set to 0.
 *
 *	For all nblocksp, startblockp and partnamep, tagp, a value of NULL
 *	indicates the corresponding info is not requested.
 *
 *
 * Return values:
 *	0	success
 *	EINVAL  no valid label or requested partition number is invalid.
 *
 */
int
cmlb_partinfo(cmlb_handle_t cmlbhandle, int part, diskaddr_t *nblocksp,
    diskaddr_t *startblockp, char **partnamep, uint16_t *tagp, void *tg_cookie)
{

	struct cmlb_lun *cl = (struct cmlb_lun *)cmlbhandle;
	int rval;
#if defined(__i386) || defined(__amd64)
	int ext_part;
#endif

	ASSERT(cl != NULL);
	mutex_enter(CMLB_MUTEX(cl));
	if (cl->cl_state < CMLB_ATTACHED) {
		mutex_exit(CMLB_MUTEX(cl));
		return (EINVAL);
	}

	if (part  < 0 || part >= MAXPART) {
		rval = EINVAL;
	} else {
		if (!cl->cl_f_geometry_is_valid)
			(void) cmlb_validate_geometry((struct cmlb_lun *)cl,
			    B_FALSE, 0, tg_cookie);

		if (((!cl->cl_f_geometry_is_valid) ||
		    (part < NDKMAP && cl->cl_solaris_size == 0)) &&
		    (part != P0_RAW_DISK)) {
			rval = EINVAL;
		} else {
			if (startblockp != NULL)
				*startblockp = (diskaddr_t)cl->cl_offset[part];

			if (nblocksp != NULL)
				*nblocksp = (diskaddr_t)
				    cl->cl_map[part].dkl_nblk;

			if (tagp != NULL)
				*tagp =
				    ((cl->cl_cur_labeltype == CMLB_LABEL_EFI) ||
				    (part >= NDKMAP)) ? V_UNASSIGNED :
				    cl->cl_vtoc.v_part[part].p_tag;
			rval = 0;
		}

		/* consistent with behavior of sd for getting minor name */
		if (partnamep != NULL) {
#if defined(__i386) || defined(__amd64)
#if defined(_FIRMWARE_NEEDS_FDISK)
		if (part > FDISK_P4) {
			ext_part = part-FDISK_P4-1;
			*partnamep = dk_ext_minor_data[ext_part].name;
		} else
#endif
#endif
			*partnamep = dk_minor_data[part].name;
		}

	}

	mutex_exit(CMLB_MUTEX(cl));
	return (rval);
}

/*
 * cmlb_efi_label_capacity:
 *	Get capacity stored in EFI disk label.
 *
 * Arguments:
 *	cmlbhandle	cmlb handle associated with device.
 *	capacity	pointer to capacity stored in EFI disk label.
 *	tg_cookie	cookie from target driver to be passed back to target
 *			driver when we call back to it through tg_ops.
 *
 *
 * Notes:
 *	If in-core label is not valid, this functions tries to revalidate
 *	the label. If label is valid and is an EFI label, it stores the capacity
 *      in disk label in the area pointed to by capacity.
 *
 *
 * Return values:
 *	0	success
 *	EINVAL  no valid EFI label or capacity is NULL.
 *
 */
int
cmlb_efi_label_capacity(cmlb_handle_t cmlbhandle, diskaddr_t *capacity,
    void *tg_cookie)
{
	struct cmlb_lun *cl = (struct cmlb_lun *)cmlbhandle;
	int rval;

	ASSERT(cl != NULL);
	mutex_enter(CMLB_MUTEX(cl));
	if (cl->cl_state < CMLB_ATTACHED) {
		mutex_exit(CMLB_MUTEX(cl));
		return (EINVAL);
	}

	if (!cl->cl_f_geometry_is_valid)
		(void) cmlb_validate_geometry((struct cmlb_lun *)cl, B_FALSE,
		    0, tg_cookie);

	if ((!cl->cl_f_geometry_is_valid) || (capacity == NULL) ||
	    (cl->cl_cur_labeltype != CMLB_LABEL_EFI)) {
		rval = EINVAL;
	} else {
		*capacity = (diskaddr_t)cl->cl_map[WD_NODE].dkl_nblk;
		rval = 0;
	}

	mutex_exit(CMLB_MUTEX(cl));
	return (rval);
}

/* Caller should make sure Test Unit Ready succeeds before calling this. */
/*ARGSUSED*/
int
cmlb_ioctl(cmlb_handle_t cmlbhandle, dev_t dev, int cmd, intptr_t arg,
    int flag, cred_t *cred_p, int *rval_p, void *tg_cookie)
{

	int err;
	struct cmlb_lun *cl;

	cl = (struct cmlb_lun *)cmlbhandle;

	ASSERT(cl != NULL);

	mutex_enter(CMLB_MUTEX(cl));
	if (cl->cl_state < CMLB_ATTACHED) {
		mutex_exit(CMLB_MUTEX(cl));
		return (EIO);
	}

	switch (cmd) {
		case DKIOCSEXTVTOC:
		case DKIOCSGEOM:
		case DKIOCSETEFI:
		case DKIOCSMBOOT:
#if defined(__i386) || defined(__amd64)
		case DKIOCSETEXTPART:
#endif
			break;
		case DKIOCSVTOC:
#if defined(__i386) || defined(__amd64)
		case DKIOCPARTINFO:
#endif
			if (cl->cl_blockcount > CMLB_OLDVTOC_LIMIT) {
				mutex_exit(CMLB_MUTEX(cl));
				return (EOVERFLOW);
			}
			break;
		default:
			(void) cmlb_validate_geometry(cl, 1, CMLB_SILENT,
			    tg_cookie);

			switch (cmd) {
			case DKIOCGVTOC:
			case DKIOCGAPART:
			case DKIOCSAPART:

				if (cl->cl_label_from_media == CMLB_LABEL_EFI) {
					/* GPT label on disk */
					mutex_exit(CMLB_MUTEX(cl));
					return (ENOTSUP);
				} else if
				    (cl->cl_blockcount > CMLB_OLDVTOC_LIMIT) {
					mutex_exit(CMLB_MUTEX(cl));
					return (EOVERFLOW);
				}
				break;

			case DKIOCGGEOM:
				if (cl->cl_label_from_media == CMLB_LABEL_EFI) {
					/* GPT label on disk */
					mutex_exit(CMLB_MUTEX(cl));
					return (ENOTSUP);
				}
				break;
			default:
				break;
			}
	}

	mutex_exit(CMLB_MUTEX(cl));

	switch (cmd) {
	case DKIOCGGEOM:
		cmlb_dbg(CMLB_TRACE, cl, "DKIOCGGEOM\n");
		err = cmlb_dkio_get_geometry(cl, (caddr_t)arg, flag, tg_cookie);
		break;

	case DKIOCSGEOM:
		cmlb_dbg(CMLB_TRACE, cl, "DKIOCSGEOM\n");
		err = cmlb_dkio_set_geometry(cl, (caddr_t)arg, flag);
		break;

	case DKIOCGAPART:
		cmlb_dbg(CMLB_TRACE, cl, "DKIOCGAPART\n");
		err = cmlb_dkio_get_partition(cl, (caddr_t)arg,
		    flag, tg_cookie);
		break;

	case DKIOCSAPART:
		cmlb_dbg(CMLB_TRACE, cl, "DKIOCSAPART\n");
		err = cmlb_dkio_set_partition(cl, (caddr_t)arg, flag);
		break;

	case DKIOCGVTOC:
		cmlb_dbg(CMLB_TRACE, cl, "DKIOCGVTOC\n");
		err = cmlb_dkio_get_vtoc(cl, (caddr_t)arg, flag, tg_cookie);
		break;

	case DKIOCGEXTVTOC:
		cmlb_dbg(CMLB_TRACE, cl, "DKIOCGVTOC\n");
		err = cmlb_dkio_get_extvtoc(cl, (caddr_t)arg, flag, tg_cookie);
		break;

	case DKIOCGETEFI:
		cmlb_dbg(CMLB_TRACE, cl, "DKIOCGETEFI\n");
		err = cmlb_dkio_get_efi(cl, (caddr_t)arg, flag, tg_cookie);
		break;

	case DKIOCPARTITION:
		cmlb_dbg(CMLB_TRACE, cl, "DKIOCPARTITION\n");
		err = cmlb_dkio_partition(cl, (caddr_t)arg, flag, tg_cookie);
		break;

	case DKIOCSVTOC:
		cmlb_dbg(CMLB_TRACE, cl, "DKIOCSVTOC\n");
		err = cmlb_dkio_set_vtoc(cl, dev, (caddr_t)arg, flag,
		    tg_cookie);
		break;

	case DKIOCSEXTVTOC:
		cmlb_dbg(CMLB_TRACE, cl, "DKIOCSVTOC\n");
		err = cmlb_dkio_set_extvtoc(cl, dev, (caddr_t)arg, flag,
		    tg_cookie);
		break;

	case DKIOCSETEFI:
		cmlb_dbg(CMLB_TRACE, cl, "DKIOCSETEFI\n");
		err = cmlb_dkio_set_efi(cl, dev, (caddr_t)arg, flag, tg_cookie);
		break;

	case DKIOCGMBOOT:
		cmlb_dbg(CMLB_TRACE, cl, "DKIOCGMBOOT\n");
		err = cmlb_dkio_get_mboot(cl, (caddr_t)arg, flag, tg_cookie);
		break;

	case DKIOCSMBOOT:
		cmlb_dbg(CMLB_TRACE, cl, "DKIOCSMBOOT\n");
		err = cmlb_dkio_set_mboot(cl, (caddr_t)arg, flag, tg_cookie);
		break;
	case DKIOCG_PHYGEOM:
		cmlb_dbg(CMLB_TRACE, cl, "DKIOCG_PHYGEOM\n");
#if defined(__i386) || defined(__amd64)
		err = cmlb_dkio_get_phygeom(cl, (caddr_t)arg, flag, tg_cookie);
#else
		err = ENOTTY;
#endif
		break;
	case DKIOCG_VIRTGEOM:
		cmlb_dbg(CMLB_TRACE, cl, "DKIOCG_VIRTGEOM\n");
#if defined(__i386) || defined(__amd64)
		err = cmlb_dkio_get_virtgeom(cl, (caddr_t)arg, flag);
#else
		err = ENOTTY;
#endif
		break;
	case DKIOCPARTINFO:
		cmlb_dbg(CMLB_TRACE, cl, "DKIOCPARTINFO");
#if defined(__i386) || defined(__amd64)
		err = cmlb_dkio_partinfo(cl, dev, (caddr_t)arg, flag);
#else
		err = ENOTTY;
#endif
		break;
	case DKIOCEXTPARTINFO:
		cmlb_dbg(CMLB_TRACE, cl, "DKIOCPARTINFO");
#if defined(__i386) || defined(__amd64)
		err = cmlb_dkio_extpartinfo(cl, dev, (caddr_t)arg, flag);
#else
		err = ENOTTY;
#endif
		break;
#if defined(__i386) || defined(__amd64)
	case DKIOCSETEXTPART:
		cmlb_dbg(CMLB_TRACE, cl, "DKIOCSETEXTPART");
		err = cmlb_dkio_set_ext_part(cl, (caddr_t)arg, flag, tg_cookie);
		break;
#endif
	default:
		err = ENOTTY;

	}

	/*
	 * An ioctl that succeeds and changed ('set') size(9P) information
	 * needs to invalidate the cached devinfo snapshot to avoid having
	 * old information being returned in a snapshots.
	 *
	 * NB: When available, call ddi_change_minor_node() to clear
	 * SSIZEVALID in specfs vnodes via spec_size_invalidate().
	 */
	if (err == 0) {
		switch (cmd) {
		case DKIOCSGEOM:
		case DKIOCSAPART:
		case DKIOCSVTOC:
		case DKIOCSEXTVTOC:
		case DKIOCSETEFI:
			i_ddi_prop_dyn_cache_invalidate(CMLB_DEVINFO(cl),
			    i_ddi_prop_dyn_driver_get(CMLB_DEVINFO(cl)));
		}
	}
	return (err);
}

dev_t
cmlb_make_device(struct cmlb_lun *cl)
{
	if (cl->cl_alter_behavior & CMLB_CREATE_P0_MINOR_NODE) {
		return (makedevice(ddi_driver_major(CMLB_DEVINFO(cl)),
		    ddi_get_instance(
		    CMLB_DEVINFO(cl)) << CMLBUNIT_FORCE_P0_SHIFT));
	} else {
		return (makedevice(ddi_driver_major(CMLB_DEVINFO(cl)),
		    ddi_get_instance(CMLB_DEVINFO(cl)) << CMLBUNIT_SHIFT));
	}
}

/*
 * Function: cmlb_check_update_blockcount
 *
 * Description: If current capacity value is invalid, obtains the
 *		current capacity from target driver.
 *
 * Return Code: 0	success
 *		EIO	failure
 */
static int
cmlb_check_update_blockcount(struct cmlb_lun *cl, void *tg_cookie)
{
	int status;
	diskaddr_t capacity;
	uint32_t lbasize;

	ASSERT(mutex_owned(CMLB_MUTEX(cl)));

	if (cl->cl_f_geometry_is_valid)
		return (0);

	mutex_exit(CMLB_MUTEX(cl));
	status = DK_TG_GETCAP(cl, &capacity, tg_cookie);
	if (status != 0) {
		mutex_enter(CMLB_MUTEX(cl));
		return (EIO);
	}

	status = DK_TG_GETBLOCKSIZE(cl, &lbasize, tg_cookie);
	mutex_enter(CMLB_MUTEX(cl));
	if (status != 0)
		return (EIO);

	if ((capacity != 0) && (lbasize != 0)) {
		cl->cl_blockcount = capacity;
		cl->cl_tgt_blocksize = lbasize;
		if (!cl->cl_is_removable) {
			cl->cl_sys_blocksize = lbasize;
		}
		return (0);
	} else {
		return (EIO);
	}
}

static int
cmlb_create_minor(dev_info_t *dip, char *name, int spec_type,
    minor_t minor_num, char *node_type, int flag, boolean_t internal)
{
	ASSERT(VALID_BOOLEAN(internal));

	if (internal)
		return (ddi_create_internal_pathname(dip,
		    name, spec_type, minor_num));
	else
		return (ddi_create_minor_node(dip,
		    name, spec_type, minor_num, node_type, flag));
}

/*
 *    Function: cmlb_create_minor_nodes
 *
 * Description: Create or adjust the minor device nodes for the instance.
 * 		Minor nodes are created based on default label type,
 *		current label type and last label type we created
 *		minor nodes based on.
 *
 *
 *   Arguments: cl - driver soft state (unit) structure
 *
 * Return Code: 0 success
 *		ENXIO	failure.
 *
 *     Context: Kernel thread context
 */
static int
cmlb_create_minor_nodes(struct cmlb_lun *cl)
{
	struct driver_minor_data	*dmdp;
	int				instance, shift;
	char				name[48];
	cmlb_label_t			newlabeltype;
	boolean_t			internal;

	ASSERT(cl != NULL);
	ASSERT(mutex_owned(CMLB_MUTEX(cl)));

	internal = VOID2BOOLEAN(
	    (cl->cl_alter_behavior & (CMLB_INTERNAL_MINOR_NODES)) != 0);

	if (cl->cl_alter_behavior & CMLB_CREATE_P0_MINOR_NODE)
		shift = CMLBUNIT_FORCE_P0_SHIFT;
	else
		shift = CMLBUNIT_SHIFT;

	/* check the most common case */
	if (cl->cl_cur_labeltype != CMLB_LABEL_UNDEF &&
	    cl->cl_last_labeltype == cl->cl_cur_labeltype) {
		/* do nothing */
		return (0);
	}

	if (cl->cl_def_labeltype == CMLB_LABEL_UNDEF) {
		/* we should never get here */
		return (ENXIO);
	}

	if (cl->cl_last_labeltype == CMLB_LABEL_UNDEF) {
		/* first time during attach */
		newlabeltype = cl->cl_def_labeltype;

		instance = ddi_get_instance(CMLB_DEVINFO(cl));

		/* Create all the minor nodes for this target. */
		dmdp = (newlabeltype == CMLB_LABEL_EFI) ? dk_minor_data_efi :
		    dk_minor_data;
		while (dmdp->name != NULL) {

			(void) sprintf(name, "%s", dmdp->name);

			if (cmlb_create_minor(CMLB_DEVINFO(cl), name,
			    dmdp->type,
			    (instance << shift) | dmdp->minor,
			    cl->cl_node_type, NULL, internal) == DDI_FAILURE) {
				/*
				 * Clean up any nodes that may have been
				 * created, in case this fails in the middle
				 * of the loop.
				 */
				ddi_remove_minor_node(CMLB_DEVINFO(cl), NULL);
				return (ENXIO);
			}
			dmdp++;
		}
		cl->cl_last_labeltype = newlabeltype;
#if defined(_SUNOS_VTOC_8)
		/*
		 * "emulate" p0 device for sparc, used by lofi
		 */
		if (cl->cl_alter_behavior & CMLB_CREATE_P0_MINOR_NODE) {
			if (cmlb_create_minor(CMLB_DEVINFO(cl), "q", S_IFBLK,
			    (instance << CMLBUNIT_FORCE_P0_SHIFT) | P0_RAW_DISK,
			    cl->cl_node_type, NULL, internal) == DDI_FAILURE) {
				ddi_remove_minor_node(CMLB_DEVINFO(cl), NULL);
				return (ENXIO);
			}

			if (cmlb_create_minor(CMLB_DEVINFO(cl), "q,raw",
			    S_IFCHR,
			    (instance << CMLBUNIT_FORCE_P0_SHIFT) | P0_RAW_DISK,
			    cl->cl_node_type, NULL, internal) == DDI_FAILURE) {
				ddi_remove_minor_node(CMLB_DEVINFO(cl), NULL);
				return (ENXIO);
			}
		}
#endif	/* defined(_SUNOS_VTOC_8) */
		return (0);
	}

	/* Not first time  */
	if (cl->cl_cur_labeltype == CMLB_LABEL_UNDEF) {
		if (cl->cl_last_labeltype != cl->cl_def_labeltype) {
			/* close time, revert to default. */
			newlabeltype = cl->cl_def_labeltype;
		} else {
			/*
			 * do nothing since the type for which we last created
			 * nodes matches the default
			 */
			return (0);
		}
	} else {
		if (cl->cl_cur_labeltype != cl->cl_last_labeltype) {
			/* We are not closing, use current label type */
			newlabeltype = cl->cl_cur_labeltype;
		} else {
			/*
			 * do nothing since the type for which we last created
			 * nodes matches the current label type
			 */
			return (0);
		}
	}

	instance = ddi_get_instance(CMLB_DEVINFO(cl));

	/*
	 * Currently we only fix up the s7 node when we are switching
	 * label types from or to EFI. This is consistent with
	 * current behavior of sd.
	 */
	if (newlabeltype == CMLB_LABEL_EFI &&
	    cl->cl_last_labeltype != CMLB_LABEL_EFI) {
		/* from vtoc to EFI */
		ddi_remove_minor_node(CMLB_DEVINFO(cl), "h");
		ddi_remove_minor_node(CMLB_DEVINFO(cl), "h,raw");
		(void) cmlb_create_minor(CMLB_DEVINFO(cl), "wd",
		    S_IFBLK, (instance << shift) | WD_NODE,
		    cl->cl_node_type, NULL, internal);
		(void) cmlb_create_minor(CMLB_DEVINFO(cl), "wd,raw",
		    S_IFCHR, (instance << shift) | WD_NODE,
		    cl->cl_node_type, NULL, internal);
	} else {
		/* from efi to vtoc */
		ddi_remove_minor_node(CMLB_DEVINFO(cl), "wd");
		ddi_remove_minor_node(CMLB_DEVINFO(cl), "wd,raw");
		(void) cmlb_create_minor(CMLB_DEVINFO(cl), "h",
		    S_IFBLK, (instance << shift) | WD_NODE,
		    cl->cl_node_type, NULL, internal);
		(void) cmlb_create_minor(CMLB_DEVINFO(cl), "h,raw",
		    S_IFCHR, (instance << shift) | WD_NODE,
		    cl->cl_node_type, NULL, internal);
	}

	cl->cl_last_labeltype = newlabeltype;
	return (0);
}

/*
 *    Function: cmlb_validate_geometry
 *
 * Description: Read the label from the disk (if present). Update the unit's
 *		geometry and vtoc information from the data in the label.
 *		Verify that the label is valid.
 *
 *   Arguments:
 *	cl		driver soft state (unit) structure
 *
 *	forcerevalid	force revalidation even if we are already valid.
 *	flags		operation flags from target driver. Used for verbosity
 *			control	at this time.
 *	tg_cookie	cookie from target driver to be passed back to target
 *			driver when we call back to it through tg_ops.
 *
 * Return Code: 0 - Successful completion
 *		EINVAL  - Invalid value in cl->cl_tgt_blocksize or
 *			  cl->cl_blockcount; or label on disk is corrupted
 *			  or unreadable.
 *		EACCES  - Reservation conflict at the device.
 *		ENOMEM  - Resource allocation error
 *		ENOTSUP - geometry not applicable
 *
 *     Context: Kernel thread only (can sleep).
 */
static int
cmlb_validate_geometry(struct cmlb_lun *cl, boolean_t forcerevalid, int flags,
    void *tg_cookie)
{
	int		label_error = 0;
	diskaddr_t	capacity;
	int		count;

	ASSERT(mutex_owned(CMLB_MUTEX(cl)));
	ASSERT(VALID_BOOLEAN(forcerevalid));

	if ((cl->cl_f_geometry_is_valid) && (!forcerevalid)) {
		if (cl->cl_cur_labeltype == CMLB_LABEL_EFI)
			return (ENOTSUP);
		return (0);
	}

	if (cmlb_check_update_blockcount(cl, tg_cookie) != 0)
		return (EIO);

	capacity = cl->cl_blockcount;

	/*
	 * Set up the "whole disk" fdisk partition; this should always
	 * exist, regardless of whether the disk contains an fdisk table
	 * or vtoc.
	 */
	cl->cl_map[P0_RAW_DISK].dkl_cylno = 0;
	cl->cl_offset[P0_RAW_DISK] = 0;
	/*
	 * note if capacity > int32_max(1TB) we are in 64bit environment
	 * so no truncation happens
	 */
	cl->cl_map[P0_RAW_DISK].dkl_nblk  = capacity;

	/*
	 * Refresh the logical and physical geometry caches.
	 * (data from MODE SENSE format/rigid disk geometry pages,
	 * and scsi_ifgetcap("geometry").
	 */
	cmlb_resync_geom_caches(cl, capacity, tg_cookie);

	cl->cl_label_from_media = CMLB_LABEL_UNDEF;
	label_error = cmlb_use_efi(cl, capacity, flags, tg_cookie);
	if (label_error == 0) {

		/* found a valid EFI label */
		cmlb_dbg(CMLB_TRACE, cl,
		    "cmlb_validate_geometry: found EFI label\n");
		/*
		 * solaris_size and geometry_is_valid are set in
		 * cmlb_use_efi
		 */
		return (ENOTSUP);
	}

	/* NO EFI label found */

	if (capacity > CMLB_EXTVTOC_LIMIT) {
		if (label_error == ESRCH) {
			/*
			 * they've configured a LUN over 2TB, but used
			 * format.dat to restrict format's view of the
			 * capacity to be under 2TB in some earlier Solaris
			 * release.
			 */
			/* i.e > 2TB with a VTOC < 2TB */
			if (!(flags & CMLB_SILENT) &&
			    (cl->cl_msglog_flag & CMLB_ALLOW_2TB_WARN)) {

				cmlb_log(CMLB_DEVINFO(cl), CMLB_LABEL(cl),
				    CE_NOTE, "!Disk (%s%d) is limited to 2 TB "
				    "due to VTOC label. To use the full "
				    "capacity of the disk, use format(1M) to "
				    "relabel the disk with EFI/GPT label.\n",
				    CMLB_LABEL(cl),
				    ddi_get_instance(CMLB_DEVINFO(cl)));

				cl->cl_msglog_flag &= ~CMLB_ALLOW_2TB_WARN;
			}
		} else {
				return (ENOTSUP);
		}
	}

	label_error = 0;

	/*
	 * at this point it is either labeled with a VTOC or it is
	 * under 1TB (<= 1TB actually for off-by-1)
	 */

	/*
	 * Only DIRECT ACCESS devices will have Scl labels.
	 * CD's supposedly have a Scl label, too
	 */
	if (cl->cl_device_type == DTYPE_DIRECT || ISREMOVABLE(cl)) {
		struct	dk_label *dkl;
		offset_t label_addr;
		int	rval;
		size_t	buffer_size;

		/*
		 * Note: This will set up cl->cl_solaris_size and
		 * cl->cl_solaris_offset.
		 */
		rval = cmlb_read_fdisk(cl, capacity, tg_cookie);
		if ((rval != 0) && !ISCD(cl)) {
			ASSERT(mutex_owned(CMLB_MUTEX(cl)));
			return (rval);
		}

		if (cl->cl_solaris_size <= DK_LABEL_LOC) {
			/*
			 * Found fdisk table but no Solaris partition entry,
			 * so don't call cmlb_uselabel() and don't create
			 * a default label.
			 */
			label_error = 0;
			cl->cl_f_geometry_is_valid = B_TRUE;
			goto no_solaris_partition;
		}

		label_addr = (daddr_t)(cl->cl_solaris_offset + DK_LABEL_LOC);

		buffer_size = cl->cl_sys_blocksize;

		cmlb_dbg(CMLB_TRACE, cl, "cmlb_validate_geometry: "
		    "label_addr: 0x%x allocation size: 0x%x\n",
		    label_addr, buffer_size);

		if ((dkl = kmem_zalloc(buffer_size, KM_NOSLEEP)) == NULL)
			return (ENOMEM);

		mutex_exit(CMLB_MUTEX(cl));
		rval = DK_TG_READ(cl, dkl, label_addr, buffer_size, tg_cookie);
		mutex_enter(CMLB_MUTEX(cl));

		switch (rval) {
		case 0:
			/*
			 * cmlb_uselabel will establish that the geometry
			 * is valid.
			 */
			if (cmlb_uselabel(cl,
			    (struct dk_label *)(uintptr_t)dkl, flags) !=
			    CMLB_LABEL_IS_VALID) {
				label_error = EINVAL;
			} else
				cl->cl_label_from_media = CMLB_LABEL_VTOC;
			break;
		case EACCES:
			label_error = EACCES;
			break;
		default:
			label_error = EINVAL;
			break;
		}

		kmem_free(dkl, buffer_size);
	}

	/*
	 * If a valid label was not found, AND if no reservation conflict
	 * was detected, then go ahead and create a default label (4069506).
	 *
	 * Note: currently, for VTOC_8 devices, the default label is created
	 * for removables and hotpluggables only.  For VTOC_16 devices, the
	 * default label will be created for all devices.
	 * (see cmlb_build_default_label)
	 */
#if defined(_SUNOS_VTOC_8)
	if ((ISREMOVABLE(cl) || ISHOTPLUGGABLE(cl)) &&
	    (label_error != EACCES)) {
#elif defined(_SUNOS_VTOC_16)
	if (label_error != EACCES) {
#endif
		if (!cl->cl_f_geometry_is_valid) {
			cmlb_build_default_label(cl, tg_cookie);
		}
		label_error = 0;
	}

no_solaris_partition:

#if defined(_SUNOS_VTOC_16)
	/*
	 * If we have valid geometry, set up the remaining fdisk partitions.
	 * Note that dkl_cylno is not used for the fdisk map entries, so
	 * we set it to an entirely bogus value.
	 */
	for (count = 0; count < FDISK_PARTS; count++) {
		cl->cl_map[FDISK_P1 + count].dkl_cylno = UINT16_MAX;
		cl->cl_map[FDISK_P1 + count].dkl_nblk =
		    cl->cl_fmap[count].fmap_nblk;

		cl->cl_offset[FDISK_P1 + count] =
		    cl->cl_fmap[count].fmap_start;
	}
#endif

	for (count = 0; count < NDKMAP; count++) {
#if defined(_SUNOS_VTOC_8)
		struct dk_map *lp  = &cl->cl_map[count];
		cl->cl_offset[count] =
		    cl->cl_g.dkg_nhead * cl->cl_g.dkg_nsect * lp->dkl_cylno;
#elif defined(_SUNOS_VTOC_16)
		struct dkl_partition *vp = &cl->cl_vtoc.v_part[count];

		cl->cl_offset[count] = vp->p_start + cl->cl_solaris_offset;
#else
#error "No VTOC format defined."
#endif
	}

	return (label_error);
}

#if defined(_SUNOS_VTOC_16)
/*
 *    Function: cmlb_convert_geometry
 *
 * Description: Convert physical geometry into a dk_geom structure. In
 *		other words, make sure we don't wrap 16-bit values.
 *		e.g. converting from geom_cache to dk_geom
 *
 *     Context: Kernel thread only
 */
static void
cmlb_convert_geometry(struct cmlb_lun *cl, diskaddr_t capacity,
    struct dk_geom *cl_g, void *tg_cookie)
{

	ASSERT(cl != NULL);
	ASSERT(mutex_owned(CMLB_MUTEX(cl)));

	/* Unlabeled SCSI floppy device */
	if (capacity < 160) {
		/* Less than 80K */
		cl_g->dkg_nhead = 1;
		cl_g->dkg_ncyl = capacity;
		cl_g->dkg_nsect = 1;
		return;
	} else if (capacity <= 0x1000) {
		cl_g->dkg_nhead = 2;
		cl_g->dkg_ncyl = 80;
		cl_g->dkg_nsect = capacity / (cl_g->dkg_nhead * cl_g->dkg_ncyl);
		return;
	}

	/*
	 * For all devices we calculate cylinders using the heads and sectors
	 * we assign based on capacity of the device.  The algorithm is
	 * designed to be compatible with the way other operating systems
	 * lay out fdisk tables for X86 and to insure that the cylinders never
	 * exceed 65535 to prevent problems with X86 ioctls that report
	 * geometry.
	 * For some smaller disk sizes we report geometry that matches those
	 * used by X86 BIOS usage. For larger disks, we use SPT that are
	 * multiples of 63, since other OSes that are not limited to 16-bits
	 * for cylinders stop at 63 SPT we make do by using multiples of 63 SPT.
	 *
	 * The following table (in order) illustrates some end result
	 * calculations:
	 *
	 * Maximum number of blocks 		nhead	nsect
	 *
	 * 2097152 (1GB)			64	32
	 * 16777216 (8GB)			128	32
	 * 1052819775 (502.02GB)		255  	63
	 * 2105639550 (0.98TB)			255	126
	 * 3158459325 (1.47TB)			255  	189
	 * 4211279100 (1.96TB)			255  	252
	 * 5264098875 (2.45TB)			255  	315
	 * ...
	 *
	 * For Solid State Drive(SSD), it uses 4K page size inside and may be
	 * double with every new generation. If the I/O is not aligned with
	 * page size on SSDs, SSDs perform a lot slower.
	 * By default, Solaris partition starts from cylinder 1. It will be
	 * misaligned even with 4K if using heads(255) and SPT(63). To
	 * workaround the problem, if the device is SSD, we use heads(224) and
	 * SPT multiple of 56. Thus the default Solaris partition starts from
	 * a position that aligns with 128K on a 512 bytes sector size SSD.
	 */

	if (capacity <= 0x200000) {
		cl_g->dkg_nhead = 64;
		cl_g->dkg_nsect = 32;
	} else if (capacity <= 0x01000000) {
		cl_g->dkg_nhead = 128;
		cl_g->dkg_nsect = 32;
	} else {
		tg_attribute_t tgattribute;
		int is_solid_state;
		unsigned short nhead;
		unsigned short nsect;

		bzero(&tgattribute, sizeof (tg_attribute_t));

		mutex_exit(CMLB_MUTEX(cl));
		is_solid_state =
		    (DK_TG_GETATTRIBUTE(cl, &tgattribute, tg_cookie) == 0) ?
		    tgattribute.media_is_solid_state : FALSE;
		mutex_enter(CMLB_MUTEX(cl));

		if (is_solid_state) {
			nhead = 224;
			nsect = 56;
		} else {
			nhead = 255;
			nsect = 63;
		}

		cl_g->dkg_nhead = nhead;

		/* make dkg_nsect be smallest multiple of nsect */
		cl_g->dkg_nsect = ((capacity +
		    (UINT16_MAX * nhead * nsect) - 1) /
		    (UINT16_MAX * nhead * nsect)) * nsect;

		if (cl_g->dkg_nsect == 0)
			cl_g->dkg_nsect = (UINT16_MAX / nsect) * nsect;
	}

}
#endif

/*
 *    Function: cmlb_resync_geom_caches
 *
 * Description: (Re)initialize both geometry caches: the virtual geometry
 *            information is extracted from the HBA (the "geometry"
 *            capability), and the physical geometry cache data is
 *            generated by issuing MODE SENSE commands.
 *
 *   Arguments:
 *	cl 		driver soft state (unit) structure
 *	capacity	disk capacity in #blocks
 *	tg_cookie	cookie from target driver to be passed back to target
 *			driver when we call back to it through tg_ops.
 *
 *     Context: Kernel thread only (can sleep).
 */
static void
cmlb_resync_geom_caches(struct cmlb_lun *cl, diskaddr_t capacity,
    void *tg_cookie)
{
	struct	cmlb_geom 	pgeom;
	struct	cmlb_geom	lgeom;
	struct 	cmlb_geom	*pgeomp = &pgeom;
	unsigned short 		nhead;
	unsigned short 		nsect;
	int 			spc;
	int			ret;

	ASSERT(cl != NULL);
	ASSERT(mutex_owned(CMLB_MUTEX(cl)));

	/*
	 * Ask the controller for its logical geometry.
	 * Note: if the HBA does not support scsi_ifgetcap("geometry"),
	 * then the lgeom cache will be invalid.
	 */
	mutex_exit(CMLB_MUTEX(cl));
	bzero(&lgeom, sizeof (struct cmlb_geom));
	ret = DK_TG_GETVIRTGEOM(cl, &lgeom, tg_cookie);
	mutex_enter(CMLB_MUTEX(cl));

	bcopy(&lgeom, &cl->cl_lgeom, sizeof (cl->cl_lgeom));

	/*
	 * Initialize the pgeom cache from lgeom, so that if MODE SENSE
	 * doesn't work, DKIOCG_PHYSGEOM can return reasonable values.
	 */
	if (ret != 0 || cl->cl_lgeom.g_nsect == 0 ||
	    cl->cl_lgeom.g_nhead == 0) {
		/*
		 * Note: Perhaps this needs to be more adaptive? The rationale
		 * is that, if there's no HBA geometry from the HBA driver, any
		 * guess is good, since this is the physical geometry. If MODE
		 * SENSE fails this gives a max cylinder size for non-LBA access
		 */
		nhead = 255;
		nsect = 63;
	} else {
		nhead = cl->cl_lgeom.g_nhead;
		nsect = cl->cl_lgeom.g_nsect;
	}

	if (ISCD(cl)) {
		pgeomp->g_nhead = 1;
		pgeomp->g_nsect = nsect * nhead;
	} else {
		pgeomp->g_nhead = nhead;
		pgeomp->g_nsect = nsect;
	}

	spc = pgeomp->g_nhead * pgeomp->g_nsect;
	pgeomp->g_capacity = capacity;
	if (spc == 0)
		pgeomp->g_ncyl = 0;
	else
		pgeomp->g_ncyl = pgeomp->g_capacity / spc;
	pgeomp->g_acyl = 0;

	/*
	 * Retrieve fresh geometry data from the hardware, stash it
	 * here temporarily before we rebuild the incore label.
	 *
	 * We want to use the MODE SENSE commands to derive the
	 * physical geometry of the device, but if either command
	 * fails, the logical geometry is used as the fallback for
	 * disk label geometry.
	 */

	mutex_exit(CMLB_MUTEX(cl));
	(void) DK_TG_GETPHYGEOM(cl, pgeomp, tg_cookie);
	mutex_enter(CMLB_MUTEX(cl));

	/*
	 * Now update the real copy while holding the mutex. This
	 * way the global copy is never in an inconsistent state.
	 */
	bcopy(pgeomp, &cl->cl_pgeom,  sizeof (cl->cl_pgeom));

	cmlb_dbg(CMLB_INFO, cl, "cmlb_resync_geom_caches: "
	    "(cached from lgeom)\n");
	cmlb_dbg(CMLB_INFO,  cl,
	    "   ncyl: %ld; acyl: %d; nhead: %d; nsect: %d\n",
	    cl->cl_pgeom.g_ncyl, cl->cl_pgeom.g_acyl,
	    cl->cl_pgeom.g_nhead, cl->cl_pgeom.g_nsect);
	cmlb_dbg(CMLB_INFO,  cl, "   lbasize: %d; capacity: %ld; "
	    "intrlv: %d; rpm: %d\n", cl->cl_pgeom.g_secsize,
	    cl->cl_pgeom.g_capacity, cl->cl_pgeom.g_intrlv,
	    cl->cl_pgeom.g_rpm);
}


#if defined(__i386) || defined(__amd64)
/*
 *    Function: cmlb_update_ext_minor_nodes
 *
 * Description: Routine to add/remove extended partition device nodes
 *
 *   Arguments:
 *	cl		driver soft state (unit) structure
 *	num_parts	Number of logical drives found on the LUN
 *
 * Should be called with the mutex held
 *
 * Return Code: 0 for success
 *
 *     Context: User and Kernel thread
 *
 */
static int
cmlb_update_ext_minor_nodes(struct cmlb_lun *cl, int num_parts)
{
	int				i, count, shift;
	char				name[48];
	int				instance;
	struct driver_minor_data	*demdp, *demdpr;
	char				*devnm;
	dev_info_t			*pdip;
	boolean_t 			internal;

	ASSERT(mutex_owned(CMLB_MUTEX(cl)));
	ASSERT(cl->cl_update_ext_minor_nodes == 1);

	internal = VOID2BOOLEAN(
	    (cl->cl_alter_behavior & (CMLB_INTERNAL_MINOR_NODES)) != 0);
	instance = ddi_get_instance(CMLB_DEVINFO(cl));
	demdp = dk_ext_minor_data;
	demdpr = &dk_ext_minor_data[MAX_EXT_PARTS];

	if (cl->cl_alter_behavior & CMLB_CREATE_P0_MINOR_NODE)
		shift = CMLBUNIT_FORCE_P0_SHIFT;
	else
		shift = CMLBUNIT_SHIFT;

	if (cl->cl_logical_drive_count) {
		for (i = 0; i < cl->cl_logical_drive_count; i++) {
			(void) sprintf(name, "%s", demdp->name);
			ddi_remove_minor_node(CMLB_DEVINFO(cl), name);
			(void) sprintf(name, "%s", demdpr->name);
			ddi_remove_minor_node(CMLB_DEVINFO(cl), name);
			demdp++;
			demdpr++;
		}
		/* There are existing device nodes. Remove them */
		devnm = kmem_alloc(MAXNAMELEN + 1, KM_SLEEP);
		(void) ddi_deviname(cl->cl_devi, devnm);
		pdip = ddi_get_parent(cl->cl_devi);
		(void) devfs_clean(pdip, devnm + 1, DV_CLEAN_FORCE);
		kmem_free(devnm, MAXNAMELEN + 1);
	}

	demdp = dk_ext_minor_data;
	demdpr = &dk_ext_minor_data[MAX_EXT_PARTS];

	for (i = 0; i < num_parts; i++) {
		(void) sprintf(name, "%s", demdp->name);
		if (cmlb_create_minor(CMLB_DEVINFO(cl), name,
		    demdp->type,
		    (instance << shift) | demdp->minor,
		    cl->cl_node_type, NULL, internal) == DDI_FAILURE) {
			/*
			 * Clean up any nodes that may have been
			 * created, in case this fails in the middle
			 * of the loop.
			 */
			ddi_remove_minor_node(CMLB_DEVINFO(cl), NULL);
			cl->cl_logical_drive_count = 0;
			return (ENXIO);
		}
		(void) sprintf(name, "%s", demdpr->name);
		if (ddi_create_minor_node(CMLB_DEVINFO(cl), name,
		    demdpr->type,
		    (instance << shift) | demdpr->minor,
		    cl->cl_node_type, NULL) == DDI_FAILURE) {
			/*
			 * Clean up any nodes that may have been
			 * created, in case this fails in the middle
			 * of the loop.
			 */
			ddi_remove_minor_node(CMLB_DEVINFO(cl), NULL);
			cl->cl_logical_drive_count = 0;
			return (ENXIO);
		}
		demdp++;
		demdpr++;
	}

	/* Update the cl_map array for logical drives */
	for (count = 0; count < MAX_EXT_PARTS; count++) {
		cl->cl_map[FDISK_P4 + 1 + count].dkl_cylno = UINT32_MAX;
		cl->cl_map[FDISK_P4 + 1 + count].dkl_nblk =
		    cl->cl_fmap[FD_NUMPART + count].fmap_nblk;
		cl->cl_offset[FDISK_P4 + 1 + count] =
		    cl->cl_fmap[FD_NUMPART + count].fmap_start;
	}

	cl->cl_logical_drive_count = i;
	cl->cl_update_ext_minor_nodes = 0;
	return (0);
}
/*
 *    Function: cmlb_validate_ext_part
 *
 * Description: utility routine to validate an extended partition's
 *		metadata as found on disk
 *
 *   Arguments:
 *	cl		driver soft state (unit) structure
 *	part		partition number of the extended partition
 *	epart		partition number of the logical drive
 *	start		absolute sector number of the start of the logical
 *			drive being validated
 *	size		size of logical drive being validated
 *
 * Return Code: 0 for success
 *
 *     Context: User and Kernel thread
 *
 * Algorithm :
 * Error cases are :
 *	1. If start block is lesser than or equal to the end block
 *	2. If either start block or end block is beyond the bounadry
 *	   of the extended partition.
 *	3. start or end block overlap with existing partitions.
 *		To check this, first make sure that the start block doesnt
 *		overlap with existing partitions. Then, calculate the
 *		possible end block for the given start block that doesnt
 *		overlap with existing partitions. This can be calculated by
 *		first setting the possible end block to the end of the
 *		extended partition (optimistic) and then, checking if there
 *		is any other partition that lies after the start of the
 *		partition being validated. If so, set the possible end to
 *		one block less than the beginning of the next nearest partition
 *		If the actual end block is greater than the calculated end
 *		block, we have an overlap.
 *
 */
static int
cmlb_validate_ext_part(struct cmlb_lun *cl, int part, int epart, uint32_t start,
    uint32_t size)
{
	int i;
	uint32_t end = start + size - 1;
	uint32_t ext_start = cl->cl_fmap[part].fmap_start;
	uint32_t ext_end = ext_start + cl->cl_fmap[part].fmap_nblk - 1;
	uint32_t ts, te;
	uint32_t poss_end = ext_end;

	if (end <= start) {
		return (1);
	}

	/*
	 * Check if the logical drive boundaries are within that of the
	 * extended partition.
	 */
	if (start <= ext_start || start > ext_end || end <= ext_start ||
	    end > ext_end) {
		return (1);
	}

	/*
	 * epart will be equal to FD_NUMPART if it is the first logical drive.
	 * There is no need to check for overlaps with other logical drives,
	 * since it is the only logical drive that we have come across so far.
	 */
	if (epart == FD_NUMPART) {
		return (0);
	}

	/* Check for overlaps with existing logical drives */
	i = FD_NUMPART;
	ts = cl->cl_fmap[FD_NUMPART].fmap_start;
	te = ts + cl->cl_fmap[FD_NUMPART].fmap_nblk - 1;

	while ((i < epart) && ts && te) {
		if (start >= ts && start <= te) {
			return (1);
		}

		if ((ts < poss_end) && (ts > start)) {
			poss_end = ts - 1;
		}

		i++;
		ts = cl->cl_fmap[i].fmap_start;
		te = ts + cl->cl_fmap[i].fmap_nblk - 1;
	}

	if (end > poss_end) {
		return (1);
	}

	return (0);
}


/*
 *    Function: cmlb_is_linux_swap
 *
 * Description: utility routine to verify if a partition is a linux swap
 *		partition or not.
 *
 *   Arguments:
 *	cl		driver soft state (unit) structure
 *	part_start	absolute sector number of the start of the partition
 *			being verified
 *	tg_cookie	cookie from target driver to be passed back to target
 *			driver when we call back to it through tg_ops.
 *
 * Return Code: 0 for success
 *
 *     Context: User and Kernel thread
 *
 * Notes:
 *	The linux swap magic "SWAP-SPACE" or "SWAPSPACE2" is found as the
 *	last 10 bytes of a disk block whose size is that of the linux page
 *	size. This disk block is found at the beginning of the swap partition.
 */
static int
cmlb_is_linux_swap(struct cmlb_lun *cl, uint32_t part_start, void *tg_cookie)
{
	int		i;
	int		rval = -1;
	uint32_t	seek_offset;
	uint32_t	linux_pg_size;
	char 		*buf, *linux_swap_magic;
	int		sec_sz = cl->cl_sys_blocksize;
	/* Known linux kernel page sizes */
	uint32_t	linux_pg_size_arr[] = {4096, };

	ASSERT(cl != NULL);
	ASSERT(mutex_owned(CMLB_MUTEX(cl)));

	if ((buf = kmem_zalloc(sec_sz, KM_NOSLEEP)) == NULL) {
		return (ENOMEM);
	}

	/*
	 * Check if there is a sane Solaris VTOC
	 * If there is a valid vtoc, no need to lookup
	 * for the linux swap signature.
	 */
	mutex_exit(CMLB_MUTEX(cl));
	rval = DK_TG_READ(cl, buf, part_start + DK_LABEL_LOC,
	    sec_sz, tg_cookie);
	mutex_enter(CMLB_MUTEX(cl));
	if (rval != 0) {
		cmlb_dbg(CMLB_ERROR,  cl,
		    "cmlb_is_linux_swap: disk vtoc read err\n");
		rval = EIO;
		goto done;
	}

	if ((((struct dk_label *)buf)->dkl_magic == DKL_MAGIC) &&
	    (((struct dk_label *)buf)->dkl_vtoc.v_sanity == VTOC_SANE)) {
		rval = -1;
		goto done;
	}


	/* No valid vtoc, so check for linux swap signature */
	linux_swap_magic = buf + sec_sz - 10;

	for (i = 0; i < sizeof (linux_pg_size_arr)/sizeof (uint32_t); i++) {
		linux_pg_size = linux_pg_size_arr[i];
		seek_offset = linux_pg_size/sec_sz - 1;
		seek_offset += part_start;

		mutex_exit(CMLB_MUTEX(cl));
		rval = DK_TG_READ(cl, buf, seek_offset, sec_sz, tg_cookie);
		mutex_enter(CMLB_MUTEX(cl));

		if (rval != 0) {
			cmlb_dbg(CMLB_ERROR,  cl,
			    "cmlb_is_linux_swap: disk read err\n");
			rval = EIO;
			break;
		}

		rval = -1;

		if ((strncmp(linux_swap_magic, "SWAP-SPACE", 10) == 0) ||
		    (strncmp(linux_swap_magic, "SWAPSPACE2", 10) == 0)) {
			/* Found a linux swap */
			rval = 0;
			break;
		}
	}

done:
	kmem_free(buf, sec_sz);
	return (rval);
}
#endif

/*
 *    Function: cmlb_read_fdisk
 *
 * Description: utility routine to read the fdisk table.
 *
 *   Arguments:
 *	cl		driver soft state (unit) structure
 *	capacity	disk capacity in #blocks
 *	tg_cookie	cookie from target driver to be passed back to target
 *			driver when we call back to it through tg_ops.
 *
 * Return Code: 0 for success (includes not reading for no_fdisk_present case
 *		errnos from tg_rw if failed to read the first block.
 *
 *     Context: Kernel thread only (can sleep).
 */
/*ARGSUSED*/
static int
cmlb_read_fdisk(struct cmlb_lun *cl, diskaddr_t capacity, void *tg_cookie)
{
#if defined(_NO_FDISK_PRESENT)

	cl->cl_solaris_offset = 0;
	cl->cl_solaris_size = capacity;
	bzero(cl->cl_fmap, sizeof (struct fmap) * FD_NUMPART);
	return (0);

#elif defined(_FIRMWARE_NEEDS_FDISK)

	struct ipart	*fdp;
	struct mboot	*mbp;
	struct ipart	fdisk[FD_NUMPART];
	int		i, k;
	char		sigbuf[2];
	caddr_t		bufp;
	int		uidx;
	int 		rval;
	int		lba = 0;
	uint_t		solaris_offset;	/* offset to solaris part. */
	daddr_t		solaris_size;	/* size of solaris partition */
	uint32_t	blocksize;
#if defined(__i386) || defined(__amd64)
	struct ipart	eparts[2];
	struct ipart	*efdp1 = &eparts[0];
	struct ipart	*efdp2 = &eparts[1];
	int		ext_part_exists = 0;
	int		ld_count = 0;
#endif

	ASSERT(cl != NULL);
	ASSERT(mutex_owned(CMLB_MUTEX(cl)));

	/*
	 * Start off assuming no fdisk table
	 */
	solaris_offset = 0;
	solaris_size   = capacity;

	blocksize = cl->cl_tgt_blocksize;

	bufp = kmem_zalloc(blocksize, KM_SLEEP);

	mutex_exit(CMLB_MUTEX(cl));
	rval = DK_TG_READ(cl, bufp, 0, blocksize, tg_cookie);
	mutex_enter(CMLB_MUTEX(cl));

	if (rval != 0) {
		cmlb_dbg(CMLB_ERROR,  cl,
		    "cmlb_read_fdisk: fdisk read err\n");
		bzero(cl->cl_fmap, sizeof (struct fmap) * FD_NUMPART);
		goto done;
	}

	mbp = (struct mboot *)bufp;

	/*
	 * The fdisk table does not begin on a 4-byte boundary within the
	 * master boot record, so we copy it to an aligned structure to avoid
	 * alignment exceptions on some processors.
	 */
	bcopy(&mbp->parts[0], fdisk, sizeof (fdisk));

	/*
	 * Check for lba support before verifying sig; sig might not be
	 * there, say on a blank disk, but the max_chs mark may still
	 * be present.
	 *
	 * Note: LBA support and BEFs are an x86-only concept but this
	 * code should work OK on SPARC as well.
	 */

	/*
	 * First, check for lba-access-ok on root node (or prom root node)
	 * if present there, don't need to search fdisk table.
	 */
	if (ddi_getprop(DDI_DEV_T_ANY, ddi_root_node(), 0,
	    "lba-access-ok", 0) != 0) {
		/* All drives do LBA; don't search fdisk table */
		lba = 1;
	} else {
		/* Okay, look for mark in fdisk table */
		for (fdp = fdisk, i = 0; i < FD_NUMPART; i++, fdp++) {
			/* accumulate "lba" value from all partitions */
			lba = (lba || cmlb_has_max_chs_vals(fdp));
		}
	}

	if (lba != 0) {
		dev_t dev = cmlb_make_device(cl);

		if (ddi_getprop(dev, CMLB_DEVINFO(cl), DDI_PROP_DONTPASS,
		    "lba-access-ok", 0) == 0) {
			/* not found; create it */
			if (ddi_prop_create(dev, CMLB_DEVINFO(cl), 0,
			    "lba-access-ok", (caddr_t)NULL, 0) !=
			    DDI_PROP_SUCCESS) {
				cmlb_dbg(CMLB_ERROR,  cl,
				    "cmlb_read_fdisk: Can't create lba "
				    "property for instance %d\n",
				    ddi_get_instance(CMLB_DEVINFO(cl)));
			}
		}
	}

	bcopy(&mbp->signature, sigbuf, sizeof (sigbuf));

	/*
	 * Endian-independent signature check
	 */
	if (((sigbuf[1] & 0xFF) != ((MBB_MAGIC >> 8) & 0xFF)) ||
	    (sigbuf[0] != (MBB_MAGIC & 0xFF))) {
		cmlb_dbg(CMLB_ERROR,  cl,
		    "cmlb_read_fdisk: no fdisk\n");
		bzero(cl->cl_fmap, sizeof (struct fmap) * FD_NUMPART);
		goto done;
	}

#ifdef CMLBDEBUG
	if (cmlb_level_mask & CMLB_LOGMASK_INFO) {
		fdp = fdisk;
		cmlb_dbg(CMLB_INFO,  cl, "cmlb_read_fdisk:\n");
		cmlb_dbg(CMLB_INFO,  cl, "         relsect    "
		    "numsect         sysid       bootid\n");
		for (i = 0; i < FD_NUMPART; i++, fdp++) {
			cmlb_dbg(CMLB_INFO,  cl,
			    "    %d:  %8d   %8d     0x%08x     0x%08x\n",
			    i, fdp->relsect, fdp->numsect,
			    fdp->systid, fdp->bootid);
		}
	}
#endif

	/*
	 * Try to find the unix partition
	 */
	uidx = -1;
	solaris_offset = 0;
	solaris_size   = 0;

	for (fdp = fdisk, i = 0; i < FD_NUMPART; i++, fdp++) {
		uint32_t relsect;
		uint32_t numsect;
		uchar_t systid;
#if defined(__i386) || defined(__amd64)
		/*
		 * Stores relative block offset from the beginning of the
		 * Extended Partition.
		 */
		int	ext_relsect = 0;
#endif

		if (fdp->numsect == 0) {
			cl->cl_fmap[i].fmap_start = 0;
			cl->cl_fmap[i].fmap_nblk  = 0;
			continue;
		}

		/*
		 * Data in the fdisk table is little-endian.
		 */
		relsect = LE_32(fdp->relsect);
		numsect = LE_32(fdp->numsect);

		cl->cl_fmap[i].fmap_start = relsect;
		cl->cl_fmap[i].fmap_nblk  = numsect;
		cl->cl_fmap[i].fmap_systid = LE_8(fdp->systid);

#if defined(__i386) || defined(__amd64)
		/* Support only one extended partition per LUN */
		if ((fdp->systid == EXTDOS || fdp->systid == FDISK_EXTLBA) &&
		    (ext_part_exists == 0)) {
			int j;
			uint32_t logdrive_offset;
			uint32_t ext_numsect;
			uint32_t abs_secnum;

			ext_part_exists = 1;

			for (j = FD_NUMPART; j < FDISK_PARTS; j++) {
				mutex_exit(CMLB_MUTEX(cl));
				rval = DK_TG_READ(cl, bufp,
				    (relsect + ext_relsect), blocksize,
				    tg_cookie);
				mutex_enter(CMLB_MUTEX(cl));

				if (rval != 0) {
					cmlb_dbg(CMLB_ERROR,  cl,
					    "cmlb_read_fdisk: Extended "
					    "partition read err\n");
					goto done;
				}
				/*
				 * The first ipart entry provides the offset
				 * at which the logical drive starts off from
				 * the beginning of the container partition
				 * and the size of the logical drive.
				 * The second ipart entry provides the offset
				 * of the next container partition from the
				 * beginning of the extended partition.
				 */
				bcopy(&bufp[FDISK_PART_TABLE_START], eparts,
				    sizeof (eparts));
				logdrive_offset = LE_32(efdp1->relsect);
				ext_numsect = LE_32(efdp1->numsect);
				systid = LE_8(efdp1->systid);
				if (logdrive_offset <= 0 || ext_numsect <= 0)
					break;
				abs_secnum = relsect + ext_relsect +
				    logdrive_offset;

				/* Boundary condition and overlap checking */
				if (cmlb_validate_ext_part(cl, i, j, abs_secnum,
				    ext_numsect)) {
					break;
				}

				if ((cl->cl_fmap[j].fmap_start != abs_secnum) ||
				    (cl->cl_fmap[j].fmap_nblk != ext_numsect) ||
				    (cl->cl_fmap[j].fmap_systid != systid)) {
					/*
					 * Indicates change from previous
					 * partinfo. Need to recreate
					 * logical device nodes.
					 */
					cl->cl_update_ext_minor_nodes = 1;
				}
				cl->cl_fmap[j].fmap_start = abs_secnum;
				cl->cl_fmap[j].fmap_nblk  = ext_numsect;
				cl->cl_fmap[j].fmap_systid = systid;
				ld_count++;

				if ((efdp1->systid == SUNIXOS &&
				    (cmlb_is_linux_swap(cl, abs_secnum,
				    tg_cookie) != 0)) ||
				    efdp1->systid == SUNIXOS2) {
					if (uidx == -1) {
						uidx = 0;
						solaris_offset = abs_secnum;
						solaris_size = ext_numsect;
					}
				}

				if ((ext_relsect = LE_32(efdp2->relsect)) == 0)
					break;
			}
		}

#endif

		if (fdp->systid != SUNIXOS &&
		    fdp->systid != SUNIXOS2 &&
		    fdp->systid != EFI_PMBR) {
			continue;
		}

		/*
		 * use the last active solaris partition id found
		 * (there should only be 1 active partition id)
		 *
		 * if there are no active solaris partition id
		 * then use the first inactive solaris partition id
		 */
		if ((uidx == -1) || (fdp->bootid == ACTIVE)) {
#if defined(__i386) || defined(__amd64)
			if (fdp->systid != SUNIXOS ||
			    (fdp->systid == SUNIXOS &&
			    (cmlb_is_linux_swap(cl, relsect,
			    tg_cookie) != 0))) {
#endif
				uidx = i;
				solaris_offset = relsect;
				solaris_size   = numsect;
#if defined(__i386) || defined(__amd64)
			}
#endif
		}
	}
#if defined(__i386) || defined(__amd64)
	if (ld_count < cl->cl_logical_drive_count) {
		/*
		 * Some/all logical drives were deleted. Clear out
		 * the fmap entries correspoding to those deleted drives.
		 */
		for (k = ld_count + FD_NUMPART;
		    k < cl->cl_logical_drive_count + FD_NUMPART; k++) {
			cl->cl_fmap[k].fmap_start = 0;
			cl->cl_fmap[k].fmap_nblk  = 0;
			cl->cl_fmap[k].fmap_systid = 0;
		}
		cl->cl_update_ext_minor_nodes = 1;
	}
	if (cl->cl_update_ext_minor_nodes) {
		rval = cmlb_update_ext_minor_nodes(cl, ld_count);
		if (rval != 0) {
			goto done;
		}
	}
#endif
	cmlb_dbg(CMLB_INFO,  cl, "fdisk 0x%x 0x%lx",
	    cl->cl_solaris_offset, cl->cl_solaris_size);
done:

	/*
	 * Clear the VTOC info, only if the Solaris partition entry
	 * has moved, changed size, been deleted, or if the size of
	 * the partition is too small to even fit the label sector.
	 */
	if ((cl->cl_solaris_offset != solaris_offset) ||
	    (cl->cl_solaris_size != solaris_size) ||
	    solaris_size <= DK_LABEL_LOC) {
		cmlb_dbg(CMLB_INFO,  cl, "fdisk moved 0x%x 0x%lx",
		    solaris_offset, solaris_size);
		bzero(&cl->cl_g, sizeof (struct dk_geom));
		bzero(&cl->cl_vtoc, sizeof (struct dk_vtoc));
		bzero(&cl->cl_map, NDKMAP * (sizeof (struct dk_map)));
		cl->cl_f_geometry_is_valid = B_FALSE;
	}
	cl->cl_solaris_offset = solaris_offset;
	cl->cl_solaris_size = solaris_size;
	kmem_free(bufp, blocksize);
	return (rval);

#else	/* #elif defined(_FIRMWARE_NEEDS_FDISK) */
#error "fdisk table presence undetermined for this platform."
#endif	/* #if defined(_NO_FDISK_PRESENT) */
}

static void
cmlb_swap_efi_gpt(efi_gpt_t *e)
{
	_NOTE(ASSUMING_PROTECTED(*e))
	e->efi_gpt_Signature = LE_64(e->efi_gpt_Signature);
	e->efi_gpt_Revision = LE_32(e->efi_gpt_Revision);
	e->efi_gpt_HeaderSize = LE_32(e->efi_gpt_HeaderSize);
	e->efi_gpt_HeaderCRC32 = LE_32(e->efi_gpt_HeaderCRC32);
	e->efi_gpt_MyLBA = LE_64(e->efi_gpt_MyLBA);
	e->efi_gpt_AlternateLBA = LE_64(e->efi_gpt_AlternateLBA);
	e->efi_gpt_FirstUsableLBA = LE_64(e->efi_gpt_FirstUsableLBA);
	e->efi_gpt_LastUsableLBA = LE_64(e->efi_gpt_LastUsableLBA);
	UUID_LE_CONVERT(e->efi_gpt_DiskGUID, e->efi_gpt_DiskGUID);
	e->efi_gpt_PartitionEntryLBA = LE_64(e->efi_gpt_PartitionEntryLBA);
	e->efi_gpt_NumberOfPartitionEntries =
	    LE_32(e->efi_gpt_NumberOfPartitionEntries);
	e->efi_gpt_SizeOfPartitionEntry =
	    LE_32(e->efi_gpt_SizeOfPartitionEntry);
	e->efi_gpt_PartitionEntryArrayCRC32 =
	    LE_32(e->efi_gpt_PartitionEntryArrayCRC32);
}

static void
cmlb_swap_efi_gpe(int nparts, efi_gpe_t *p)
{
	int i;

	_NOTE(ASSUMING_PROTECTED(*p))
	for (i = 0; i < nparts; i++) {
		UUID_LE_CONVERT(p[i].efi_gpe_PartitionTypeGUID,
		    p[i].efi_gpe_PartitionTypeGUID);
		p[i].efi_gpe_StartingLBA = LE_64(p[i].efi_gpe_StartingLBA);
		p[i].efi_gpe_EndingLBA = LE_64(p[i].efi_gpe_EndingLBA);
		/* PartitionAttrs */
	}
}

static int
cmlb_validate_efi(efi_gpt_t *labp)
{
	if (labp->efi_gpt_Signature != EFI_SIGNATURE)
		return (EINVAL);
	/* at least 96 bytes in this version of the spec. */
	if (sizeof (efi_gpt_t) - sizeof (labp->efi_gpt_Reserved2) >
	    labp->efi_gpt_HeaderSize)
		return (EINVAL);
	/* this should be 128 bytes */
	if (labp->efi_gpt_SizeOfPartitionEntry != sizeof (efi_gpe_t))
		return (EINVAL);
	return (0);
}

/*
 * This function returns B_FALSE if there is a valid MBR signature and no
 * partition table entries of type EFI_PMBR (0xEE). Otherwise it returns B_TRUE.
 *
 * The EFI spec (1.10 and later) requires having a Protective MBR (PMBR) to
 * recognize the disk as GPT partitioned. However, some other OS creates an MBR
 * where a PMBR entry is not the only one. Also, if the first block has been
 * corrupted, currently best attempt to allow data access would be to try to
 * check for GPT headers. Hence in case of more than one partition entry, but
 * at least one EFI_PMBR partition type or no valid magic number, the function
 * returns B_TRUE to continue with looking for GPT header.
 */

static boolean_t
cmlb_check_efi_mbr(uchar_t *buf, boolean_t *is_mbr)
{
	struct ipart	*fdp;
	struct mboot	*mbp = (struct mboot *)buf;
	struct ipart	fdisk[FD_NUMPART];
	int		i;

	if (is_mbr != NULL)
		*is_mbr = B_TRUE;

	if (LE_16(mbp->signature) != MBB_MAGIC) {
		if (is_mbr != NULL)
			*is_mbr = B_FALSE;
		return (B_TRUE);
	}

	bcopy(&mbp->parts[0], fdisk, sizeof (fdisk));

	for (fdp = fdisk, i = 0; i < FD_NUMPART; i++, fdp++) {
		if (fdp->systid == EFI_PMBR)
			return (B_TRUE);
	}

	return (B_FALSE);
}

static int
cmlb_use_efi(struct cmlb_lun *cl, diskaddr_t capacity, int flags,
    void *tg_cookie)
{
	int		i;
	int		rval = 0;
	efi_gpe_t	*partitions;
	uchar_t		*buf;
	uint_t		lbasize;	/* is really how much to read */
	diskaddr_t	cap = 0;
	uint_t		nparts;
	diskaddr_t	gpe_lba;
	diskaddr_t	alternate_lba;
	int		iofailed = 0;
	struct uuid	uuid_type_reserved = EFI_RESERVED;
#if defined(_FIRMWARE_NEEDS_FDISK)
	boolean_t 	is_mbr;
#endif

	ASSERT(mutex_owned(CMLB_MUTEX(cl)));

	lbasize = cl->cl_sys_blocksize;

	cl->cl_reserved = -1;
	mutex_exit(CMLB_MUTEX(cl));

	buf = kmem_zalloc(EFI_MIN_ARRAY_SIZE, KM_SLEEP);

	rval = DK_TG_READ(cl, buf,  0, lbasize, tg_cookie);
	if (rval) {
		iofailed = 1;