summaryrefslogtreecommitdiffstats
path: root/contrib/opensolaris/lib
diff options
context:
space:
mode:
authorpjd <pjd@FreeBSD.org>2007-04-06 01:09:06 +0000
committerpjd <pjd@FreeBSD.org>2007-04-06 01:09:06 +0000
commit3b005d330261f33318ca1ee3fef1940237fd788b (patch)
tree3061c8734d9ce560165e672836837a0f411a83c9 /contrib/opensolaris/lib
parent3be454b8211f48e634e6587f53807d3b5013e973 (diff)
downloadFreeBSD-src-3b005d330261f33318ca1ee3fef1940237fd788b.zip
FreeBSD-src-3b005d330261f33318ca1ee3fef1940237fd788b.tar.gz
Please welcome ZFS - The last word in file systems.
ZFS file system was ported from OpenSolaris operating system. The code in under CDDL license. I'd like to thank all SUN developers that created this great piece of software. Supported by: Wheel LTD (http://www.wheel.pl/) Supported by: The FreeBSD Foundation (http://www.freebsdfoundation.org/) Supported by: Sentex (http://www.sentex.net/)
Diffstat (limited to 'contrib/opensolaris/lib')
-rw-r--r--contrib/opensolaris/lib/libnvpair/libnvpair.c266
-rw-r--r--contrib/opensolaris/lib/libnvpair/libnvpair.h46
-rw-r--r--contrib/opensolaris/lib/libnvpair/nvpair_alloc_system.c59
-rw-r--r--contrib/opensolaris/lib/libuutil/common/libuutil.h384
-rw-r--r--contrib/opensolaris/lib/libuutil/common/libuutil_common.h46
-rw-r--r--contrib/opensolaris/lib/libuutil/common/libuutil_impl.h181
-rw-r--r--contrib/opensolaris/lib/libuutil/common/uu_alloc.c85
-rw-r--r--contrib/opensolaris/lib/libuutil/common/uu_avl.c567
-rw-r--r--contrib/opensolaris/lib/libuutil/common/uu_dprintf.c128
-rw-r--r--contrib/opensolaris/lib/libuutil/common/uu_ident.c122
-rw-r--r--contrib/opensolaris/lib/libuutil/common/uu_list.c711
-rw-r--r--contrib/opensolaris/lib/libuutil/common/uu_misc.c250
-rw-r--r--contrib/opensolaris/lib/libuutil/common/uu_open.c70
-rw-r--r--contrib/opensolaris/lib/libuutil/common/uu_pname.c205
-rw-r--r--contrib/opensolaris/lib/libuutil/common/uu_strtoint.c300
-rw-r--r--contrib/opensolaris/lib/libzfs/common/libzfs.h441
-rw-r--r--contrib/opensolaris/lib/libzfs/common/libzfs_changelist.c606
-rw-r--r--contrib/opensolaris/lib/libzfs/common/libzfs_config.c374
-rw-r--r--contrib/opensolaris/lib/libzfs/common/libzfs_dataset.c3753
-rw-r--r--contrib/opensolaris/lib/libzfs/common/libzfs_graph.c646
-rw-r--r--contrib/opensolaris/lib/libzfs/common/libzfs_impl.h171
-rw-r--r--contrib/opensolaris/lib/libzfs/common/libzfs_import.c994
-rw-r--r--contrib/opensolaris/lib/libzfs/common/libzfs_mount.c986
-rw-r--r--contrib/opensolaris/lib/libzfs/common/libzfs_pool.c2055
-rw-r--r--contrib/opensolaris/lib/libzfs/common/libzfs_status.c289
-rw-r--r--contrib/opensolaris/lib/libzfs/common/libzfs_util.c827
-rw-r--r--contrib/opensolaris/lib/libzpool/common/kernel.c830
-rw-r--r--contrib/opensolaris/lib/libzpool/common/sys/zfs_context.h497
-rw-r--r--contrib/opensolaris/lib/libzpool/common/taskq.c250
-rw-r--r--contrib/opensolaris/lib/libzpool/common/util.c144
30 files changed, 16283 insertions, 0 deletions
diff --git a/contrib/opensolaris/lib/libnvpair/libnvpair.c b/contrib/opensolaris/lib/libnvpair/libnvpair.c
new file mode 100644
index 0000000..58a47f1
--- /dev/null
+++ b/contrib/opensolaris/lib/libnvpair/libnvpair.c
@@ -0,0 +1,266 @@
+/*
+ * CDDL HEADER START
+ *
+ * The contents of this file are subject to the terms of the
+ * Common Development and Distribution License, Version 1.0 only
+ * (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 2004 Sun Microsystems, Inc. All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#pragma ident "%Z%%M% %I% %E% SMI"
+
+#include <unistd.h>
+#include <strings.h>
+#include "libnvpair.h"
+
+/*
+ * libnvpair - A tools library for manipulating <name, value> pairs.
+ *
+ * This library provides routines packing an unpacking nv pairs
+ * for transporting data across process boundaries, transporting
+ * between kernel and userland, and possibly saving onto disk files.
+ */
+
+static void
+indent(FILE *fp, int depth)
+{
+ while (depth-- > 0)
+ (void) fprintf(fp, "\t");
+}
+
+/*
+ * nvlist_print - Prints elements in an event buffer
+ */
+static
+void
+nvlist_print_with_indent(FILE *fp, nvlist_t *nvl, int depth)
+{
+ int i;
+ char *name;
+ uint_t nelem;
+ nvpair_t *nvp;
+
+ if (nvl == NULL)
+ return;
+
+ indent(fp, depth);
+ (void) fprintf(fp, "nvlist version: %d\n", NVL_VERSION(nvl));
+
+ nvp = nvlist_next_nvpair(nvl, NULL);
+
+ while (nvp) {
+ data_type_t type = nvpair_type(nvp);
+
+ indent(fp, depth);
+ name = nvpair_name(nvp);
+ (void) fprintf(fp, "\t%s =", name);
+ nelem = 0;
+ switch (type) {
+ case DATA_TYPE_BOOLEAN: {
+ (void) fprintf(fp, " 1");
+ break;
+ }
+ case DATA_TYPE_BOOLEAN_VALUE: {
+ boolean_t val;
+ (void) nvpair_value_boolean_value(nvp, &val);
+ (void) fprintf(fp, " %d", val);
+ break;
+ }
+ case DATA_TYPE_BYTE: {
+ uchar_t val;
+ (void) nvpair_value_byte(nvp, &val);
+ (void) fprintf(fp, " 0x%2.2x", val);
+ break;
+ }
+ case DATA_TYPE_INT8: {
+ int8_t val;
+ (void) nvpair_value_int8(nvp, &val);
+ (void) fprintf(fp, " %d", val);
+ break;
+ }
+ case DATA_TYPE_UINT8: {
+ uint8_t val;
+ (void) nvpair_value_uint8(nvp, &val);
+ (void) fprintf(fp, " 0x%x", val);
+ break;
+ }
+ case DATA_TYPE_INT16: {
+ int16_t val;
+ (void) nvpair_value_int16(nvp, &val);
+ (void) fprintf(fp, " %d", val);
+ break;
+ }
+ case DATA_TYPE_UINT16: {
+ uint16_t val;
+ (void) nvpair_value_uint16(nvp, &val);
+ (void) fprintf(fp, " 0x%x", val);
+ break;
+ }
+ case DATA_TYPE_INT32: {
+ int32_t val;
+ (void) nvpair_value_int32(nvp, &val);
+ (void) fprintf(fp, " %d", val);
+ break;
+ }
+ case DATA_TYPE_UINT32: {
+ uint32_t val;
+ (void) nvpair_value_uint32(nvp, &val);
+ (void) fprintf(fp, " 0x%x", val);
+ break;
+ }
+ case DATA_TYPE_INT64: {
+ int64_t val;
+ (void) nvpair_value_int64(nvp, &val);
+ (void) fprintf(fp, " %lld", (longlong_t)val);
+ break;
+ }
+ case DATA_TYPE_UINT64: {
+ uint64_t val;
+ (void) nvpair_value_uint64(nvp, &val);
+ (void) fprintf(fp, " 0x%llx", (u_longlong_t)val);
+ break;
+ }
+ case DATA_TYPE_STRING: {
+ char *val;
+ (void) nvpair_value_string(nvp, &val);
+ (void) fprintf(fp, " %s", val);
+ break;
+ }
+ case DATA_TYPE_BOOLEAN_ARRAY: {
+ boolean_t *val;
+ (void) nvpair_value_boolean_array(nvp, &val, &nelem);
+ for (i = 0; i < nelem; i++)
+ (void) fprintf(fp, " %d", val[i]);
+ break;
+ }
+ case DATA_TYPE_BYTE_ARRAY: {
+ uchar_t *val;
+ (void) nvpair_value_byte_array(nvp, &val, &nelem);
+ for (i = 0; i < nelem; i++)
+ (void) fprintf(fp, " 0x%2.2x", val[i]);
+ break;
+ }
+ case DATA_TYPE_INT8_ARRAY: {
+ int8_t *val;
+ (void) nvpair_value_int8_array(nvp, &val, &nelem);
+ for (i = 0; i < nelem; i++)
+ (void) fprintf(fp, " %d", val[i]);
+ break;
+ }
+ case DATA_TYPE_UINT8_ARRAY: {
+ uint8_t *val;
+ (void) nvpair_value_uint8_array(nvp, &val, &nelem);
+ for (i = 0; i < nelem; i++)
+ (void) fprintf(fp, " 0x%x", val[i]);
+ break;
+ }
+ case DATA_TYPE_INT16_ARRAY: {
+ int16_t *val;
+ (void) nvpair_value_int16_array(nvp, &val, &nelem);
+ for (i = 0; i < nelem; i++)
+ (void) fprintf(fp, " %d", val[i]);
+ break;
+ }
+ case DATA_TYPE_UINT16_ARRAY: {
+ uint16_t *val;
+ (void) nvpair_value_uint16_array(nvp, &val, &nelem);
+ for (i = 0; i < nelem; i++)
+ (void) fprintf(fp, " 0x%x", val[i]);
+ break;
+ }
+ case DATA_TYPE_INT32_ARRAY: {
+ int32_t *val;
+ (void) nvpair_value_int32_array(nvp, &val, &nelem);
+ for (i = 0; i < nelem; i++)
+ (void) fprintf(fp, " %d", val[i]);
+ break;
+ }
+ case DATA_TYPE_UINT32_ARRAY: {
+ uint32_t *val;
+ (void) nvpair_value_uint32_array(nvp, &val, &nelem);
+ for (i = 0; i < nelem; i++)
+ (void) fprintf(fp, " 0x%x", val[i]);
+ break;
+ }
+ case DATA_TYPE_INT64_ARRAY: {
+ int64_t *val;
+ (void) nvpair_value_int64_array(nvp, &val, &nelem);
+ for (i = 0; i < nelem; i++)
+ (void) fprintf(fp, " %lld", (longlong_t)val[i]);
+ break;
+ }
+ case DATA_TYPE_UINT64_ARRAY: {
+ uint64_t *val;
+ (void) nvpair_value_uint64_array(nvp, &val, &nelem);
+ for (i = 0; i < nelem; i++)
+ (void) fprintf(fp, " 0x%llx",
+ (u_longlong_t)val[i]);
+ break;
+ }
+ case DATA_TYPE_STRING_ARRAY: {
+ char **val;
+ (void) nvpair_value_string_array(nvp, &val, &nelem);
+ for (i = 0; i < nelem; i++)
+ (void) fprintf(fp, " %s", val[i]);
+ break;
+ }
+ case DATA_TYPE_HRTIME: {
+ hrtime_t val;
+ (void) nvpair_value_hrtime(nvp, &val);
+ (void) fprintf(fp, " 0x%llx", val);
+ break;
+ }
+ case DATA_TYPE_NVLIST: {
+ nvlist_t *val;
+ (void) nvpair_value_nvlist(nvp, &val);
+ (void) fprintf(fp, " (embedded nvlist)\n");
+ nvlist_print_with_indent(fp, val, depth + 1);
+ indent(fp, depth + 1);
+ (void) fprintf(fp, "(end %s)\n", name);
+ break;
+ }
+ case DATA_TYPE_NVLIST_ARRAY: {
+ nvlist_t **val;
+ (void) nvpair_value_nvlist_array(nvp, &val, &nelem);
+ (void) fprintf(fp, " (array of embedded nvlists)\n");
+ for (i = 0; i < nelem; i++) {
+ indent(fp, depth + 1);
+ (void) fprintf(fp,
+ "(start %s[%d])\n", name, i);
+ nvlist_print_with_indent(fp, val[i], depth + 1);
+ indent(fp, depth + 1);
+ (void) fprintf(fp, "(end %s[%d])\n", name, i);
+ }
+ break;
+ }
+ default:
+ (void) fprintf(fp, " unknown data type (%d)", type);
+ break;
+ }
+ (void) fprintf(fp, "\n");
+ nvp = nvlist_next_nvpair(nvl, nvp);
+ }
+}
+
+void
+nvlist_print(FILE *fp, nvlist_t *nvl)
+{
+ nvlist_print_with_indent(fp, nvl, 0);
+}
diff --git a/contrib/opensolaris/lib/libnvpair/libnvpair.h b/contrib/opensolaris/lib/libnvpair/libnvpair.h
new file mode 100644
index 0000000..d1d25ea
--- /dev/null
+++ b/contrib/opensolaris/lib/libnvpair/libnvpair.h
@@ -0,0 +1,46 @@
+/*
+ * CDDL HEADER START
+ *
+ * The contents of this file are subject to the terms of the
+ * Common Development and Distribution License, Version 1.0 only
+ * (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 2005 Sun Microsystems, Inc. All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#ifndef _LIBNVPAIR_H
+#define _LIBNVPAIR_H
+
+#pragma ident "%Z%%M% %I% %E% SMI"
+
+#include <sys/nvpair.h>
+#include <stdlib.h>
+#include <stdio.h>
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+void nvlist_print(FILE *, nvlist_t *);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* _LIBNVPAIR_H */
diff --git a/contrib/opensolaris/lib/libnvpair/nvpair_alloc_system.c b/contrib/opensolaris/lib/libnvpair/nvpair_alloc_system.c
new file mode 100644
index 0000000..1aefc10
--- /dev/null
+++ b/contrib/opensolaris/lib/libnvpair/nvpair_alloc_system.c
@@ -0,0 +1,59 @@
+/*
+ * CDDL HEADER START
+ *
+ * The contents of this file are subject to the terms of the
+ * Common Development and Distribution License, Version 1.0 only
+ * (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 2004 Sun Microsystems, Inc. All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#pragma ident "%Z%%M% %I% %E% SMI"
+
+#include <sys/nvpair.h>
+#include <stdlib.h>
+
+/*ARGSUSED*/
+static void *
+nv_alloc_sys(nv_alloc_t *nva, size_t size)
+{
+ return (malloc(size));
+}
+
+/*ARGSUSED*/
+static void
+nv_free_sys(nv_alloc_t *nva, void *buf, size_t size)
+{
+ free(buf);
+}
+
+const nv_alloc_ops_t system_ops_def = {
+ NULL, /* nv_ao_init() */
+ NULL, /* nv_ao_fini() */
+ nv_alloc_sys, /* nv_ao_alloc() */
+ nv_free_sys, /* nv_ao_free() */
+ NULL /* nv_ao_reset() */
+};
+
+nv_alloc_t nv_alloc_nosleep_def = {
+ &system_ops_def,
+ NULL
+};
+
+nv_alloc_t *nv_alloc_nosleep = &nv_alloc_nosleep_def;
diff --git a/contrib/opensolaris/lib/libuutil/common/libuutil.h b/contrib/opensolaris/lib/libuutil/common/libuutil.h
new file mode 100644
index 0000000..fcd5e79
--- /dev/null
+++ b/contrib/opensolaris/lib/libuutil/common/libuutil.h
@@ -0,0 +1,384 @@
+/*
+ * CDDL HEADER START
+ *
+ * The contents of this file are subject to the terms of the
+ * Common Development and Distribution License, Version 1.0 only
+ * (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 2005 Sun Microsystems, Inc. All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#ifndef _LIBUUTIL_H
+#define _LIBUUTIL_H
+
+#pragma ident "%Z%%M% %I% %E% SMI"
+
+#include <solaris.h>
+#include <sys/types.h>
+#include <stdarg.h>
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/*
+ * Standard flags codes.
+ */
+#define UU_DEFAULT 0
+
+/*
+ * Standard error codes.
+ */
+#define UU_ERROR_NONE 0 /* no error */
+#define UU_ERROR_INVALID_ARGUMENT 1 /* invalid argument */
+#define UU_ERROR_UNKNOWN_FLAG 2 /* passed flag invalid */
+#define UU_ERROR_NO_MEMORY 3 /* out of memory */
+#define UU_ERROR_CALLBACK_FAILED 4 /* callback-initiated error */
+#define UU_ERROR_NOT_SUPPORTED 5 /* operation not supported */
+#define UU_ERROR_EMPTY 6 /* no value provided */
+#define UU_ERROR_UNDERFLOW 7 /* value is too small */
+#define UU_ERROR_OVERFLOW 8 /* value is too value */
+#define UU_ERROR_INVALID_CHAR 9 /* value contains unexpected char */
+#define UU_ERROR_INVALID_DIGIT 10 /* value contains digit not in base */
+
+#define UU_ERROR_SYSTEM 99 /* underlying system error */
+#define UU_ERROR_UNKNOWN 100 /* error status not known */
+
+/*
+ * Standard program exit codes.
+ */
+#define UU_EXIT_OK (*(uu_exit_ok()))
+#define UU_EXIT_FATAL (*(uu_exit_fatal()))
+#define UU_EXIT_USAGE (*(uu_exit_usage()))
+
+/*
+ * Exit status profiles.
+ */
+#define UU_PROFILE_DEFAULT 0
+#define UU_PROFILE_LAUNCHER 1
+
+/*
+ * Error reporting functions.
+ */
+uint32_t uu_error(void);
+const char *uu_strerror(uint32_t);
+
+/*
+ * Program notification functions.
+ */
+extern void uu_alt_exit(int);
+extern const char *uu_setpname(char *);
+extern const char *uu_getpname(void);
+/*PRINTFLIKE1*/
+extern void uu_warn(const char *, ...);
+extern void uu_vwarn(const char *, va_list);
+/*PRINTFLIKE1*/
+extern void uu_die(const char *, ...) __NORETURN;
+extern void uu_vdie(const char *, va_list) __NORETURN;
+/*PRINTFLIKE2*/
+extern void uu_xdie(int, const char *, ...) __NORETURN;
+extern void uu_vxdie(int, const char *, va_list) __NORETURN;
+
+/*
+ * Exit status functions (not to be used directly)
+ */
+extern int *uu_exit_ok(void);
+extern int *uu_exit_fatal(void);
+extern int *uu_exit_usage(void);
+
+/*
+ * string->number conversions
+ */
+extern int uu_strtoint(const char *, void *, size_t, int, int64_t, int64_t);
+extern int uu_strtouint(const char *, void *, size_t, int, uint64_t, uint64_t);
+
+/*
+ * Debug print facility functions.
+ */
+typedef struct uu_dprintf uu_dprintf_t;
+
+typedef enum {
+ UU_DPRINTF_SILENT,
+ UU_DPRINTF_FATAL,
+ UU_DPRINTF_WARNING,
+ UU_DPRINTF_NOTICE,
+ UU_DPRINTF_INFO,
+ UU_DPRINTF_DEBUG
+} uu_dprintf_severity_t;
+
+extern uu_dprintf_t *uu_dprintf_create(const char *, uu_dprintf_severity_t,
+ uint_t);
+/*PRINTFLIKE3*/
+extern void uu_dprintf(uu_dprintf_t *, uu_dprintf_severity_t,
+ const char *, ...);
+extern void uu_dprintf_destroy(uu_dprintf_t *);
+extern const char *uu_dprintf_getname(uu_dprintf_t *);
+
+/*
+ * Identifier test flags and function.
+ */
+#define UU_NAME_DOMAIN 0x1 /* allow SUNW, or com.sun, prefix */
+#define UU_NAME_PATH 0x2 /* allow '/'-delimited paths */
+
+int uu_check_name(const char *, uint_t);
+
+/*
+ * File creation functions.
+ */
+extern int uu_open_tmp(const char *dir, uint_t uflags);
+
+/*
+ * Convenience functions.
+ */
+/*PRINTFLIKE1*/
+extern char *uu_msprintf(const char *format, ...);
+extern void *uu_zalloc(size_t);
+extern void uu_free(void *);
+
+/*
+ * Comparison function type definition.
+ * Developers should be careful in their use of the _private argument. If you
+ * break interface guarantees, you get undefined behavior.
+ */
+typedef int uu_compare_fn_t(const void *__left, const void *__right,
+ void *__private);
+
+/*
+ * Walk variant flags.
+ * A data structure need not provide support for all variants and
+ * combinations. Refer to the appropriate documentation.
+ */
+#define UU_WALK_ROBUST 0x00000001 /* walk can survive removes */
+#define UU_WALK_REVERSE 0x00000002 /* reverse walk order */
+
+#define UU_WALK_PREORDER 0x00000010 /* walk tree in pre-order */
+#define UU_WALK_POSTORDER 0x00000020 /* walk tree in post-order */
+
+/*
+ * Walk callback function return codes.
+ */
+#define UU_WALK_ERROR -1
+#define UU_WALK_NEXT 0
+#define UU_WALK_DONE 1
+
+/*
+ * Walk callback function type definition.
+ */
+typedef int uu_walk_fn_t(void *_elem, void *_private);
+
+/*
+ * lists: opaque structures
+ */
+typedef struct uu_list_pool uu_list_pool_t;
+typedef struct uu_list uu_list_t;
+
+typedef struct uu_list_node {
+ uintptr_t uln_opaque[2];
+} uu_list_node_t;
+
+typedef struct uu_list_walk uu_list_walk_t;
+
+typedef uintptr_t uu_list_index_t;
+
+/*
+ * lists: interface
+ *
+ * basic usage:
+ * typedef struct foo {
+ * ...
+ * uu_list_node_t foo_node;
+ * ...
+ * } foo_t;
+ *
+ * static int
+ * foo_compare(void *l_arg, void *r_arg, void *private)
+ * {
+ * foo_t *l = l_arg;
+ * foo_t *r = r_arg;
+ *
+ * if (... l greater than r ...)
+ * return (1);
+ * if (... l less than r ...)
+ * return (-1);
+ * return (0);
+ * }
+ *
+ * ...
+ * // at initialization time
+ * foo_pool = uu_list_pool_create("foo_pool",
+ * sizeof (foo_t), offsetof(foo_t, foo_node), foo_compare,
+ * debugging? 0 : UU_AVL_POOL_DEBUG);
+ * ...
+ */
+uu_list_pool_t *uu_list_pool_create(const char *, size_t, size_t,
+ uu_compare_fn_t *, uint32_t);
+#define UU_LIST_POOL_DEBUG 0x00000001
+
+void uu_list_pool_destroy(uu_list_pool_t *);
+
+/*
+ * usage:
+ *
+ * foo_t *a;
+ * a = malloc(sizeof(*a));
+ * uu_list_node_init(a, &a->foo_list, pool);
+ * ...
+ * uu_list_node_fini(a, &a->foo_list, pool);
+ * free(a);
+ */
+void uu_list_node_init(void *, uu_list_node_t *, uu_list_pool_t *);
+void uu_list_node_fini(void *, uu_list_node_t *, uu_list_pool_t *);
+
+uu_list_t *uu_list_create(uu_list_pool_t *, void *_parent, uint32_t);
+#define UU_LIST_DEBUG 0x00000001
+#define UU_LIST_SORTED 0x00000002 /* list is sorted */
+
+void uu_list_destroy(uu_list_t *); /* list must be empty */
+
+size_t uu_list_numnodes(uu_list_t *);
+
+void *uu_list_first(uu_list_t *);
+void *uu_list_last(uu_list_t *);
+
+void *uu_list_next(uu_list_t *, void *);
+void *uu_list_prev(uu_list_t *, void *);
+
+int uu_list_walk(uu_list_t *, uu_walk_fn_t *, void *, uint32_t);
+
+uu_list_walk_t *uu_list_walk_start(uu_list_t *, uint32_t);
+void *uu_list_walk_next(uu_list_walk_t *);
+void uu_list_walk_end(uu_list_walk_t *);
+
+void *uu_list_find(uu_list_t *, void *, void *, uu_list_index_t *);
+void uu_list_insert(uu_list_t *, void *, uu_list_index_t);
+
+void *uu_list_nearest_next(uu_list_t *, uu_list_index_t);
+void *uu_list_nearest_prev(uu_list_t *, uu_list_index_t);
+
+void *uu_list_teardown(uu_list_t *, void **);
+
+void uu_list_remove(uu_list_t *, void *);
+
+/*
+ * lists: interfaces for non-sorted lists only
+ */
+int uu_list_insert_before(uu_list_t *, void *_target, void *_elem);
+int uu_list_insert_after(uu_list_t *, void *_target, void *_elem);
+
+/*
+ * avl trees: opaque structures
+ */
+typedef struct uu_avl_pool uu_avl_pool_t;
+typedef struct uu_avl uu_avl_t;
+
+typedef struct uu_avl_node {
+#ifdef _LP64
+ uintptr_t uan_opaque[3];
+#else
+ uintptr_t uan_opaque[4];
+#endif
+} uu_avl_node_t;
+
+typedef struct uu_avl_walk uu_avl_walk_t;
+
+typedef uintptr_t uu_avl_index_t;
+
+/*
+ * avl trees: interface
+ *
+ * basic usage:
+ * typedef struct foo {
+ * ...
+ * uu_avl_node_t foo_node;
+ * ...
+ * } foo_t;
+ *
+ * static int
+ * foo_compare(void *l_arg, void *r_arg, void *private)
+ * {
+ * foo_t *l = l_arg;
+ * foo_t *r = r_arg;
+ *
+ * if (... l greater than r ...)
+ * return (1);
+ * if (... l less than r ...)
+ * return (-1);
+ * return (0);
+ * }
+ *
+ * ...
+ * // at initialization time
+ * foo_pool = uu_avl_pool_create("foo_pool",
+ * sizeof (foo_t), offsetof(foo_t, foo_node), foo_compare,
+ * debugging? 0 : UU_AVL_POOL_DEBUG);
+ * ...
+ */
+uu_avl_pool_t *uu_avl_pool_create(const char *, size_t, size_t,
+ uu_compare_fn_t *, uint32_t);
+#define UU_AVL_POOL_DEBUG 0x00000001
+
+void uu_avl_pool_destroy(uu_avl_pool_t *);
+
+/*
+ * usage:
+ *
+ * foo_t *a;
+ * a = malloc(sizeof(*a));
+ * uu_avl_node_init(a, &a->foo_avl, pool);
+ * ...
+ * uu_avl_node_fini(a, &a->foo_avl, pool);
+ * free(a);
+ */
+void uu_avl_node_init(void *, uu_avl_node_t *, uu_avl_pool_t *);
+void uu_avl_node_fini(void *, uu_avl_node_t *, uu_avl_pool_t *);
+
+uu_avl_t *uu_avl_create(uu_avl_pool_t *, void *_parent, uint32_t);
+#define UU_AVL_DEBUG 0x00000001
+
+void uu_avl_destroy(uu_avl_t *); /* list must be empty */
+
+size_t uu_avl_numnodes(uu_avl_t *);
+
+void *uu_avl_first(uu_avl_t *);
+void *uu_avl_last(uu_avl_t *);
+
+void *uu_avl_next(uu_avl_t *, void *);
+void *uu_avl_prev(uu_avl_t *, void *);
+
+int uu_avl_walk(uu_avl_t *, uu_walk_fn_t *, void *, uint32_t);
+
+uu_avl_walk_t *uu_avl_walk_start(uu_avl_t *, uint32_t);
+void *uu_avl_walk_next(uu_avl_walk_t *);
+void uu_avl_walk_end(uu_avl_walk_t *);
+
+void *uu_avl_find(uu_avl_t *, void *, void *, uu_avl_index_t *);
+void uu_avl_insert(uu_avl_t *, void *, uu_avl_index_t);
+
+void *uu_avl_nearest_next(uu_avl_t *, uu_avl_index_t);
+void *uu_avl_nearest_prev(uu_avl_t *, uu_avl_index_t);
+
+void *uu_avl_teardown(uu_avl_t *, void **);
+
+void uu_avl_remove(uu_avl_t *, void *);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* _LIBUUTIL_H */
diff --git a/contrib/opensolaris/lib/libuutil/common/libuutil_common.h b/contrib/opensolaris/lib/libuutil/common/libuutil_common.h
new file mode 100644
index 0000000..fc9cc7c
--- /dev/null
+++ b/contrib/opensolaris/lib/libuutil/common/libuutil_common.h
@@ -0,0 +1,46 @@
+/*
+ * CDDL HEADER START
+ *
+ * The contents of this file are subject to the terms of the
+ * Common Development and Distribution License, Version 1.0 only
+ * (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 2006 Sun Microsystems, Inc. All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#ifndef _LIBUUTIL_COMMON_H
+#define _LIBUUTIL_COMMON_H
+
+#pragma ident "%Z%%M% %I% %E% SMI"
+
+#include <solaris.h>
+
+/*
+ * We don't bind to the internal libc interfaces if this is a
+ * native build.
+ */
+#ifndef NATIVE_BUILD
+#include "c_synonyms.h"
+#endif
+
+#include <libuutil.h>
+#include <libuutil_impl.h>
+
+#endif /* _LIBUUTIL_COMMON_H */
diff --git a/contrib/opensolaris/lib/libuutil/common/libuutil_impl.h b/contrib/opensolaris/lib/libuutil/common/libuutil_impl.h
new file mode 100644
index 0000000..9466e59
--- /dev/null
+++ b/contrib/opensolaris/lib/libuutil/common/libuutil_impl.h
@@ -0,0 +1,181 @@
+/*
+ * CDDL HEADER START
+ *
+ * The contents of this file are subject to the terms of the
+ * Common Development and Distribution License, Version 1.0 only
+ * (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 2005 Sun Microsystems, Inc. All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#ifndef _LIBUUTIL_IMPL_H
+#define _LIBUUTIL_IMPL_H
+
+#pragma ident "%Z%%M% %I% %E% SMI"
+
+#include <libuutil.h>
+#include <pthread.h>
+
+#include <sys/avl_impl.h>
+#include <sys/byteorder.h>
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+void uu_set_error(uint_t);
+#pragma rarely_called(uu_set_error)
+
+/*PRINTFLIKE1*/
+void uu_panic(const char *format, ...);
+#pragma rarely_called(uu_panic)
+
+struct uu_dprintf {
+ char *uud_name;
+ uu_dprintf_severity_t uud_severity;
+ uint_t uud_flags;
+};
+
+/*
+ * For debugging purposes, libuutil keeps around linked lists of all uu_lists
+ * and uu_avls, along with pointers to their parents. These can cause false
+ * negatives when looking for memory leaks, so we encode the pointers by
+ * storing them with swapped endianness; this is not perfect, but it's about
+ * the best we can do without wasting a lot of space.
+ */
+#ifdef _LP64
+#define UU_PTR_ENCODE(ptr) BSWAP_64((uintptr_t)(void *)(ptr))
+#else
+#define UU_PTR_ENCODE(ptr) BSWAP_32((uintptr_t)(void *)(ptr))
+#endif
+
+#define UU_PTR_DECODE(ptr) ((void *)UU_PTR_ENCODE(ptr))
+
+/*
+ * uu_list structures
+ */
+typedef struct uu_list_node_impl {
+ struct uu_list_node_impl *uln_next;
+ struct uu_list_node_impl *uln_prev;
+} uu_list_node_impl_t;
+
+struct uu_list_walk {
+ uu_list_walk_t *ulw_next;
+ uu_list_walk_t *ulw_prev;
+
+ uu_list_t *ulw_list;
+ int8_t ulw_dir;
+ uint8_t ulw_robust;
+ uu_list_node_impl_t *ulw_next_result;
+};
+
+struct uu_list {
+ uintptr_t ul_next_enc;
+ uintptr_t ul_prev_enc;
+
+ uu_list_pool_t *ul_pool;
+ uintptr_t ul_parent_enc; /* encoded parent pointer */
+ size_t ul_offset;
+ size_t ul_numnodes;
+ uint8_t ul_debug;
+ uint8_t ul_sorted;
+ uint8_t ul_index; /* mark for uu_list_index_ts */
+
+ uu_list_node_impl_t ul_null_node;
+ uu_list_walk_t ul_null_walk; /* for robust walkers */
+};
+
+#define UU_LIST_PTR(ptr) ((uu_list_t *)UU_PTR_DECODE(ptr))
+
+#define UU_LIST_POOL_MAXNAME 64
+
+struct uu_list_pool {
+ uu_list_pool_t *ulp_next;
+ uu_list_pool_t *ulp_prev;
+
+ char ulp_name[UU_LIST_POOL_MAXNAME];
+ size_t ulp_nodeoffset;
+ size_t ulp_objsize;
+ uu_compare_fn_t *ulp_cmp;
+ uint8_t ulp_debug;
+ uint8_t ulp_last_index;
+ pthread_mutex_t ulp_lock; /* protects null_list */
+ uu_list_t ulp_null_list;
+};
+
+/*
+ * uu_avl structures
+ */
+typedef struct avl_node uu_avl_node_impl_t;
+
+struct uu_avl_walk {
+ uu_avl_walk_t *uaw_next;
+ uu_avl_walk_t *uaw_prev;
+
+ uu_avl_t *uaw_avl;
+ void *uaw_next_result;
+ int8_t uaw_dir;
+ uint8_t uaw_robust;
+};
+
+struct uu_avl {
+ uintptr_t ua_next_enc;
+ uintptr_t ua_prev_enc;
+
+ uu_avl_pool_t *ua_pool;
+ uintptr_t ua_parent_enc;
+ uint8_t ua_debug;
+ uint8_t ua_index; /* mark for uu_avl_index_ts */
+
+ struct avl_tree ua_tree;
+ uu_avl_walk_t ua_null_walk;
+};
+
+#define UU_AVL_PTR(x) ((uu_avl_t *)UU_PTR_DECODE(x))
+
+#define UU_AVL_POOL_MAXNAME 64
+
+struct uu_avl_pool {
+ uu_avl_pool_t *uap_next;
+ uu_avl_pool_t *uap_prev;
+
+ char uap_name[UU_AVL_POOL_MAXNAME];
+ size_t uap_nodeoffset;
+ size_t uap_objsize;
+ uu_compare_fn_t *uap_cmp;
+ uint8_t uap_debug;
+ uint8_t uap_last_index;
+ pthread_mutex_t uap_lock; /* protects null_avl */
+ uu_avl_t uap_null_avl;
+};
+
+/*
+ * atfork() handlers
+ */
+void uu_avl_lockup(void);
+void uu_avl_release(void);
+
+void uu_list_lockup(void);
+void uu_list_release(void);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* _LIBUUTIL_IMPL_H */
diff --git a/contrib/opensolaris/lib/libuutil/common/uu_alloc.c b/contrib/opensolaris/lib/libuutil/common/uu_alloc.c
new file mode 100644
index 0000000..7cdbf01
--- /dev/null
+++ b/contrib/opensolaris/lib/libuutil/common/uu_alloc.c
@@ -0,0 +1,85 @@
+/*
+ * CDDL HEADER START
+ *
+ * The contents of this file are subject to the terms of the
+ * Common Development and Distribution License, Version 1.0 only
+ * (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 2004 Sun Microsystems, Inc. All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#pragma ident "%Z%%M% %I% %E% SMI"
+
+#include "libuutil_common.h"
+
+#include <stdarg.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+void *
+uu_zalloc(size_t n)
+{
+ void *p = malloc(n);
+
+ if (p == NULL) {
+ uu_set_error(UU_ERROR_SYSTEM);
+ return (NULL);
+ }
+
+ (void) memset(p, 0, n);
+
+ return (p);
+}
+
+void
+uu_free(void *p)
+{
+ free(p);
+}
+
+char *
+uu_msprintf(const char *format, ...)
+{
+ va_list args;
+ char attic[1];
+ uint_t M, m;
+ char *b;
+
+ va_start(args, format);
+ M = vsnprintf(attic, 1, format, args);
+ va_end(args);
+
+ for (;;) {
+ m = M;
+ if ((b = uu_zalloc(m + 1)) == NULL)
+ return (NULL);
+
+ va_start(args, format);
+ M = vsnprintf(b, m + 1, format, args);
+ va_end(args);
+
+ if (M == m)
+ break; /* sizes match */
+
+ uu_free(b);
+ }
+
+ return (b);
+}
diff --git a/contrib/opensolaris/lib/libuutil/common/uu_avl.c b/contrib/opensolaris/lib/libuutil/common/uu_avl.c
new file mode 100644
index 0000000..93feea9
--- /dev/null
+++ b/contrib/opensolaris/lib/libuutil/common/uu_avl.c
@@ -0,0 +1,567 @@
+/*
+ * 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 2006 Sun Microsystems, Inc. All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#pragma ident "%Z%%M% %I% %E% SMI"
+
+#include "libuutil_common.h"
+
+#include <stdlib.h>
+#include <string.h>
+#include <unistd.h>
+#include <sys/avl.h>
+
+static uu_avl_pool_t uu_null_apool = { &uu_null_apool, &uu_null_apool };
+static pthread_mutex_t uu_apool_list_lock = PTHREAD_MUTEX_INITIALIZER;
+
+/*
+ * The index mark change on every insert and delete, to catch stale
+ * references.
+ *
+ * We leave the low bit alone, since the avl code uses it.
+ */
+#define INDEX_MAX (sizeof (uintptr_t) - 2)
+#define INDEX_NEXT(m) (((m) == INDEX_MAX)? 2 : ((m) + 2) & INDEX_MAX)
+
+#define INDEX_DECODE(i) ((i) & ~INDEX_MAX)
+#define INDEX_ENCODE(p, n) (((n) & ~INDEX_MAX) | (p)->ua_index)
+#define INDEX_VALID(p, i) (((i) & INDEX_MAX) == (p)->ua_index)
+#define INDEX_CHECK(i) (((i) & INDEX_MAX) != 0)
+
+/*
+ * When an element is inactive (not in a tree), we keep a marked pointer to
+ * its containing pool in its first word, and a NULL pointer in its second.
+ *
+ * On insert, we use these to verify that it comes from the correct pool.
+ */
+#define NODE_ARRAY(p, n) ((uintptr_t *)((uintptr_t)(n) + \
+ (pp)->uap_nodeoffset))
+
+#define POOL_TO_MARKER(pp) (((uintptr_t)(pp) | 1))
+
+#define DEAD_MARKER 0xc4
+
+uu_avl_pool_t *
+uu_avl_pool_create(const char *name, size_t objsize, size_t nodeoffset,
+ uu_compare_fn_t *compare_func, uint32_t flags)
+{
+ uu_avl_pool_t *pp, *next, *prev;
+
+ if (name == NULL ||
+ uu_check_name(name, UU_NAME_DOMAIN) == -1 ||
+ nodeoffset + sizeof (uu_avl_node_t) > objsize ||
+ compare_func == NULL) {
+ uu_set_error(UU_ERROR_INVALID_ARGUMENT);
+ return (NULL);
+ }
+
+ if (flags & ~UU_AVL_POOL_DEBUG) {
+ uu_set_error(UU_ERROR_UNKNOWN_FLAG);
+ return (NULL);
+ }
+
+ pp = uu_zalloc(sizeof (uu_avl_pool_t));
+ if (pp == NULL) {
+ uu_set_error(UU_ERROR_NO_MEMORY);
+ return (NULL);
+ }
+
+ (void) strlcpy(pp->uap_name, name, sizeof (pp->uap_name));
+ pp->uap_nodeoffset = nodeoffset;
+ pp->uap_objsize = objsize;
+ pp->uap_cmp = compare_func;
+ if (flags & UU_AVL_POOL_DEBUG)
+ pp->uap_debug = 1;
+ pp->uap_last_index = 0;
+
+ (void) pthread_mutex_init(&pp->uap_lock, NULL);
+
+ pp->uap_null_avl.ua_next_enc = UU_PTR_ENCODE(&pp->uap_null_avl);
+ pp->uap_null_avl.ua_prev_enc = UU_PTR_ENCODE(&pp->uap_null_avl);
+
+ (void) pthread_mutex_lock(&uu_apool_list_lock);
+ pp->uap_next = next = &uu_null_apool;
+ pp->uap_prev = prev = next->uap_prev;
+ next->uap_prev = pp;
+ prev->uap_next = pp;
+ (void) pthread_mutex_unlock(&uu_apool_list_lock);
+
+ return (pp);
+}
+
+void
+uu_avl_pool_destroy(uu_avl_pool_t *pp)
+{
+ if (pp->uap_debug) {
+ if (pp->uap_null_avl.ua_next_enc !=
+ UU_PTR_ENCODE(&pp->uap_null_avl) ||
+ pp->uap_null_avl.ua_prev_enc !=
+ UU_PTR_ENCODE(&pp->uap_null_avl)) {
+ uu_panic("uu_avl_pool_destroy: Pool \"%.*s\" (%p) has "
+ "outstanding avls, or is corrupt.\n",
+ sizeof (pp->uap_name), pp->uap_name, pp);
+ }
+ }
+ (void) pthread_mutex_lock(&uu_apool_list_lock);
+ pp->uap_next->uap_prev = pp->uap_prev;
+ pp->uap_prev->uap_next = pp->uap_next;
+ (void) pthread_mutex_unlock(&uu_apool_list_lock);
+ pp->uap_prev = NULL;
+ pp->uap_next = NULL;
+ uu_free(pp);
+}
+
+void
+uu_avl_node_init(void *base, uu_avl_node_t *np, uu_avl_pool_t *pp)
+{
+ uintptr_t *na = (uintptr_t *)np;
+
+ if (pp->uap_debug) {
+ uintptr_t offset = (uintptr_t)np - (uintptr_t)base;
+ if (offset + sizeof (*np) > pp->uap_objsize) {
+ uu_panic("uu_avl_node_init(%p, %p, %p (\"%s\")): "
+ "offset %ld doesn't fit in object (size %ld)\n",
+ base, np, pp, pp->uap_name, offset,
+ pp->uap_objsize);
+ }
+ if (offset != pp->uap_nodeoffset) {
+ uu_panic("uu_avl_node_init(%p, %p, %p (\"%s\")): "
+ "offset %ld doesn't match pool's offset (%ld)\n",
+ base, np, pp, pp->uap_name, offset,
+ pp->uap_objsize);
+ }
+ }
+
+ na[0] = POOL_TO_MARKER(pp);
+ na[1] = 0;
+}
+
+void
+uu_avl_node_fini(void *base, uu_avl_node_t *np, uu_avl_pool_t *pp)
+{
+ uintptr_t *na = (uintptr_t *)np;
+
+ if (pp->uap_debug) {
+ if (na[0] == DEAD_MARKER && na[1] == DEAD_MARKER) {
+ uu_panic("uu_avl_node_fini(%p, %p, %p (\"%s\")): "
+ "node already finied\n",
+ base, np, pp, pp->uap_name);
+ }
+ if (na[0] != POOL_TO_MARKER(pp) || na[1] != 0) {
+ uu_panic("uu_avl_node_fini(%p, %p, %p (\"%s\")): "
+ "node corrupt, in tree, or in different pool\n",
+ base, np, pp, pp->uap_name);
+ }
+ }
+
+ na[0] = DEAD_MARKER;
+ na[1] = DEAD_MARKER;
+ na[2] = DEAD_MARKER;
+}
+
+struct uu_avl_node_compare_info {
+ uu_compare_fn_t *ac_compare;
+ void *ac_private;
+ void *ac_right;
+ void *ac_found;
+};
+
+static int
+uu_avl_node_compare(const void *l, const void *r)
+{
+ struct uu_avl_node_compare_info *info =
+ (struct uu_avl_node_compare_info *)l;
+
+ int res = info->ac_compare(r, info->ac_right, info->ac_private);
+
+ if (res == 0) {
+ if (info->ac_found == NULL)
+ info->ac_found = (void *)r;
+ return (-1);
+ }
+ if (res < 0)
+ return (1);
+ return (-1);
+}
+
+uu_avl_t *
+uu_avl_create(uu_avl_pool_t *pp, void *parent, uint32_t flags)
+{
+ uu_avl_t *ap, *next, *prev;
+
+ if (flags & ~UU_AVL_DEBUG) {
+ uu_set_error(UU_ERROR_UNKNOWN_FLAG);
+ return (NULL);
+ }
+
+ ap = uu_zalloc(sizeof (*ap));
+ if (ap == NULL) {
+ uu_set_error(UU_ERROR_NO_MEMORY);
+ return (NULL);
+ }
+
+ ap->ua_pool = pp;
+ ap->ua_parent_enc = UU_PTR_ENCODE(parent);
+ ap->ua_debug = pp->uap_debug || (flags & UU_AVL_DEBUG);
+ ap->ua_index = (pp->uap_last_index = INDEX_NEXT(pp->uap_last_index));
+
+ avl_create(&ap->ua_tree, &uu_avl_node_compare, pp->uap_objsize,
+ pp->uap_nodeoffset);
+
+ ap->ua_null_walk.uaw_next = &ap->ua_null_walk;
+ ap->ua_null_walk.uaw_prev = &ap->ua_null_walk;
+
+ (void) pthread_mutex_lock(&pp->uap_lock);
+ next = &pp->uap_null_avl;
+ prev = UU_PTR_DECODE(next->ua_prev_enc);
+ ap->ua_next_enc = UU_PTR_ENCODE(next);
+ ap->ua_prev_enc = UU_PTR_ENCODE(prev);
+ next->ua_prev_enc = UU_PTR_ENCODE(ap);
+ prev->ua_next_enc = UU_PTR_ENCODE(ap);
+ (void) pthread_mutex_unlock(&pp->uap_lock);
+
+ return (ap);
+}
+
+void
+uu_avl_destroy(uu_avl_t *ap)
+{
+ uu_avl_pool_t *pp = ap->ua_pool;
+
+ if (ap->ua_debug) {
+ if (avl_numnodes(&ap->ua_tree) != 0) {
+ uu_panic("uu_avl_destroy(%p): tree not empty\n", ap);
+ }
+ if (ap->ua_null_walk.uaw_next != &ap->ua_null_walk ||
+ ap->ua_null_walk.uaw_prev != &ap->ua_null_walk) {
+ uu_panic("uu_avl_destroy(%p): outstanding walkers\n",
+ ap);
+ }
+ }
+ (void) pthread_mutex_lock(&pp->uap_lock);
+ UU_AVL_PTR(ap->ua_next_enc)->ua_prev_enc = ap->ua_prev_enc;
+ UU_AVL_PTR(ap->ua_prev_enc)->ua_next_enc = ap->ua_next_enc;
+ (void) pthread_mutex_unlock(&pp->uap_lock);
+ ap->ua_prev_enc = UU_PTR_ENCODE(NULL);
+ ap->ua_next_enc = UU_PTR_ENCODE(NULL);
+
+ ap->ua_pool = NULL;
+ avl_destroy(&ap->ua_tree);
+
+ uu_free(ap);
+}
+
+size_t
+uu_avl_numnodes(uu_avl_t *ap)
+{
+ return (avl_numnodes(&ap->ua_tree));
+}
+
+void *
+uu_avl_first(uu_avl_t *ap)
+{
+ return (avl_first(&ap->ua_tree));
+}
+
+void *
+uu_avl_last(uu_avl_t *ap)
+{
+ return (avl_last(&ap->ua_tree));
+}
+
+void *
+uu_avl_next(uu_avl_t *ap, void *node)
+{
+ return (AVL_NEXT(&ap->ua_tree, node));
+}
+
+void *
+uu_avl_prev(uu_avl_t *ap, void *node)
+{
+ return (AVL_PREV(&ap->ua_tree, node));
+}
+
+static void
+_avl_walk_init(uu_avl_walk_t *wp, uu_avl_t *ap, uint32_t flags)
+{
+ uu_avl_walk_t *next, *prev;
+
+ int robust = (flags & UU_WALK_ROBUST);
+ int direction = (flags & UU_WALK_REVERSE)? -1 : 1;
+
+ (void) memset(wp, 0, sizeof (*wp));
+ wp->uaw_avl = ap;
+ wp->uaw_robust = robust;
+ wp->uaw_dir = direction;
+
+ if (direction > 0)
+ wp->uaw_next_result = avl_first(&ap->ua_tree);
+ else
+ wp->uaw_next_result = avl_last(&ap->ua_tree);
+
+ if (ap->ua_debug || robust) {
+ wp->uaw_next = next = &ap->ua_null_walk;
+ wp->uaw_prev = prev = next->uaw_prev;
+ next->uaw_prev = wp;
+ prev->uaw_next = wp;
+ }
+}
+
+static void *
+_avl_walk_advance(uu_avl_walk_t *wp, uu_avl_t *ap)
+{
+ void *np = wp->uaw_next_result;
+
+ avl_tree_t *t = &ap->ua_tree;
+
+ if (np == NULL)
+ return (NULL);
+
+ wp->uaw_next_result = (wp->uaw_dir > 0)? AVL_NEXT(t, np) :
+ AVL_PREV(t, np);
+
+ return (np);
+}
+
+static void
+_avl_walk_fini(uu_avl_walk_t *wp)
+{
+ if (wp->uaw_next != NULL) {
+ wp->uaw_next->uaw_prev = wp->uaw_prev;
+ wp->uaw_prev->uaw_next = wp->uaw_next;
+ wp->uaw_next = NULL;
+ wp->uaw_prev = NULL;
+ }
+ wp->uaw_avl = NULL;
+ wp->uaw_next_result = NULL;
+}
+
+uu_avl_walk_t *
+uu_avl_walk_start(uu_avl_t *ap, uint32_t flags)
+{
+ uu_avl_walk_t *wp;
+
+ if (flags & ~(UU_WALK_ROBUST | UU_WALK_REVERSE)) {
+ uu_set_error(UU_ERROR_UNKNOWN_FLAG);
+ return (NULL);
+ }
+
+ wp = uu_zalloc(sizeof (*wp));
+ if (wp == NULL) {
+ uu_set_error(UU_ERROR_NO_MEMORY);
+ return (NULL);
+ }
+
+ _avl_walk_init(wp, ap, flags);
+ return (wp);
+}
+
+void *
+uu_avl_walk_next(uu_avl_walk_t *wp)
+{
+ return (_avl_walk_advance(wp, wp->uaw_avl));
+}
+
+void
+uu_avl_walk_end(uu_avl_walk_t *wp)
+{
+ _avl_walk_fini(wp);
+ uu_free(wp);
+}
+
+int
+uu_avl_walk(uu_avl_t *ap, uu_walk_fn_t *func, void *private, uint32_t flags)
+{
+ void *e;
+ uu_avl_walk_t my_walk;
+
+ int status = UU_WALK_NEXT;
+
+ if (flags & ~(UU_WALK_ROBUST | UU_WALK_REVERSE)) {
+ uu_set_error(UU_ERROR_UNKNOWN_FLAG);
+ return (-1);
+ }
+
+ _avl_walk_init(&my_walk, ap, flags);
+ while (status == UU_WALK_NEXT &&
+ (e = _avl_walk_advance(&my_walk, ap)) != NULL)
+ status = (*func)(e, private);
+ _avl_walk_fini(&my_walk);
+
+ if (status >= 0)
+ return (0);
+ uu_set_error(UU_ERROR_CALLBACK_FAILED);
+ return (-1);
+}
+
+void
+uu_avl_remove(uu_avl_t *ap, void *elem)
+{
+ uu_avl_walk_t *wp;
+ uu_avl_pool_t *pp = ap->ua_pool;
+ uintptr_t *na = NODE_ARRAY(pp, elem);
+
+ if (ap->ua_debug) {
+ /*
+ * invalidate outstanding uu_avl_index_ts.
+ */
+ ap->ua_index = INDEX_NEXT(ap->ua_index);
+ }
+
+ /*
+ * Robust walkers most be advanced, if we are removing the node
+ * they are currently using. In debug mode, non-robust walkers
+ * are also on the walker list.
+ */
+ for (wp = ap->ua_null_walk.uaw_next; wp != &ap->ua_null_walk;
+ wp = wp->uaw_next) {
+ if (wp->uaw_robust) {
+ if (elem == wp->uaw_next_result)
+ (void) _avl_walk_advance(wp, ap);
+ } else if (wp->uaw_next_result != NULL) {
+ uu_panic("uu_avl_remove(%p, %p): active non-robust "
+ "walker\n", ap, elem);
+ }
+ }
+
+ avl_remove(&ap->ua_tree, elem);
+
+ na[0] = POOL_TO_MARKER(pp);
+ na[1] = 0;
+}
+
+void *
+uu_avl_teardown(uu_avl_t *ap, void **cookie)
+{
+ void *elem = avl_destroy_nodes(&ap->ua_tree, cookie);
+
+ if (elem != NULL) {
+ uu_avl_pool_t *pp = ap->ua_pool;
+ uintptr_t *na = NODE_ARRAY(pp, elem);
+
+ na[0] = POOL_TO_MARKER(pp);
+ na[1] = 0;
+ }
+ return (elem);
+}
+
+void *
+uu_avl_find(uu_avl_t *ap, void *elem, void *private, uu_avl_index_t *out)
+{
+ struct uu_avl_node_compare_info info;
+ void *result;
+
+ info.ac_compare = ap->ua_pool->uap_cmp;
+ info.ac_private = private;
+ info.ac_right = elem;
+ info.ac_found = NULL;
+
+ result = avl_find(&ap->ua_tree, &info, out);
+ if (out != NULL)
+ *out = INDEX_ENCODE(ap, *out);
+
+ if (ap->ua_debug && result != NULL)
+ uu_panic("uu_avl_find: internal error: avl_find succeeded\n");
+
+ return (info.ac_found);
+}
+
+void
+uu_avl_insert(uu_avl_t *ap, void *elem, uu_avl_index_t idx)
+{
+ if (ap->ua_debug) {
+ uu_avl_pool_t *pp = ap->ua_pool;
+ uintptr_t *na = NODE_ARRAY(pp, elem);
+
+ if (na[1] != 0)
+ uu_panic("uu_avl_insert(%p, %p, %p): node already "
+ "in tree, or corrupt\n",
+ ap, elem, idx);
+ if (na[0] == 0)
+ uu_panic("uu_avl_insert(%p, %p, %p): node not "
+ "initialized\n",
+ ap, elem, idx);
+ if (na[0] != POOL_TO_MARKER(pp))
+ uu_panic("uu_avl_insert(%p, %p, %p): node from "
+ "other pool, or corrupt\n",
+ ap, elem, idx);
+
+ if (!INDEX_VALID(ap, idx))
+ uu_panic("uu_avl_insert(%p, %p, %p): %s\n",
+ ap, elem, idx,
+ INDEX_CHECK(idx)? "outdated index" :
+ "invalid index");
+
+ /*
+ * invalidate outstanding uu_avl_index_ts.
+ */
+ ap->ua_index = INDEX_NEXT(ap->ua_index);
+ }
+ avl_insert(&ap->ua_tree, elem, INDEX_DECODE(idx));
+}
+
+void *
+uu_avl_nearest_next(uu_avl_t *ap, uu_avl_index_t idx)
+{
+ if (ap->ua_debug && !INDEX_VALID(ap, idx))
+ uu_panic("uu_avl_nearest_next(%p, %p): %s\n",
+ ap, idx, INDEX_CHECK(idx)? "outdated index" :
+ "invalid index");
+ return (avl_nearest(&ap->ua_tree, INDEX_DECODE(idx), AVL_AFTER));
+}
+
+void *
+uu_avl_nearest_prev(uu_avl_t *ap, uu_avl_index_t idx)
+{
+ if (ap->ua_debug && !INDEX_VALID(ap, idx))
+ uu_panic("uu_avl_nearest_prev(%p, %p): %s\n",
+ ap, idx, INDEX_CHECK(idx)? "outdated index" :
+ "invalid index");
+ return (avl_nearest(&ap->ua_tree, INDEX_DECODE(idx), AVL_BEFORE));
+}
+
+/*
+ * called from uu_lockup() and uu_release(), as part of our fork1()-safety.
+ */
+void
+uu_avl_lockup(void)
+{
+ uu_avl_pool_t *pp;
+
+ (void) pthread_mutex_lock(&uu_apool_list_lock);
+ for (pp = uu_null_apool.uap_next; pp != &uu_null_apool;
+ pp = pp->uap_next)
+ (void) pthread_mutex_lock(&pp->uap_lock);
+}
+
+void
+uu_avl_release(void)
+{
+ uu_avl_pool_t *pp;
+
+ for (pp = uu_null_apool.uap_next; pp != &uu_null_apool;
+ pp = pp->uap_next)
+ (void) pthread_mutex_unlock(&pp->uap_lock);
+ (void) pthread_mutex_unlock(&uu_apool_list_lock);
+}
diff --git a/contrib/opensolaris/lib/libuutil/common/uu_dprintf.c b/contrib/opensolaris/lib/libuutil/common/uu_dprintf.c
new file mode 100644
index 0000000..5b990a5
--- /dev/null
+++ b/contrib/opensolaris/lib/libuutil/common/uu_dprintf.c
@@ -0,0 +1,128 @@
+/*
+ * CDDL HEADER START
+ *
+ * The contents of this file are subject to the terms of the
+ * Common Development and Distribution License, Version 1.0 only
+ * (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 2004 Sun Microsystems, Inc. All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#pragma ident "%Z%%M% %I% %E% SMI"
+
+#include "libuutil_common.h"
+
+#include <errno.h>
+#include <libintl.h>
+#include <stdarg.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <strings.h>
+
+#define FACILITY_FMT "%s (%s): "
+
+#if !defined(TEXT_DOMAIN)
+#define TEXT_DOMAIN "SYS_TEST"
+#endif
+
+static const char *
+strseverity(uu_dprintf_severity_t severity)
+{
+ switch (severity) {
+ case UU_DPRINTF_SILENT:
+ return (dgettext(TEXT_DOMAIN, "silent"));
+ case UU_DPRINTF_FATAL:
+ return (dgettext(TEXT_DOMAIN, "FATAL"));
+ case UU_DPRINTF_WARNING:
+ return (dgettext(TEXT_DOMAIN, "WARNING"));
+ case UU_DPRINTF_NOTICE:
+ return (dgettext(TEXT_DOMAIN, "note"));
+ case UU_DPRINTF_INFO:
+ return (dgettext(TEXT_DOMAIN, "info"));
+ case UU_DPRINTF_DEBUG:
+ return (dgettext(TEXT_DOMAIN, "debug"));
+ default:
+ return (dgettext(TEXT_DOMAIN, "unspecified"));
+ }
+}
+
+uu_dprintf_t *
+uu_dprintf_create(const char *name, uu_dprintf_severity_t severity,
+ uint_t flags)
+{
+ uu_dprintf_t *D;
+
+ if (uu_check_name(name, UU_NAME_DOMAIN) == -1) {
+ uu_set_error(UU_ERROR_INVALID_ARGUMENT);
+ return (NULL);
+ }
+
+ if ((D = uu_zalloc(sizeof (uu_dprintf_t))) == NULL)
+ return (NULL);
+
+ if (name != NULL) {
+ D->uud_name = strdup(name);
+ if (D->uud_name == NULL) {
+ uu_free(D);
+ return (NULL);
+ }
+ } else {
+ D->uud_name = NULL;
+ }
+
+ D->uud_severity = severity;
+ D->uud_flags = flags;
+
+ return (D);
+}
+
+/*PRINTFLIKE3*/
+void
+uu_dprintf(uu_dprintf_t *D, uu_dprintf_severity_t severity,
+ const char *format, ...)
+{
+ va_list alist;
+
+ /* XXX Assert that severity is not UU_DPRINTF_SILENT. */
+
+ if (severity > D->uud_severity)
+ return;
+
+ (void) fprintf(stderr, FACILITY_FMT, D->uud_name,
+ strseverity(severity));
+
+ va_start(alist, format);
+ (void) vfprintf(stderr, format, alist);
+ va_end(alist);
+}
+
+void
+uu_dprintf_destroy(uu_dprintf_t *D)
+{
+ if (D->uud_name)
+ free(D->uud_name);
+
+ uu_free(D);
+}
+
+const char *
+uu_dprintf_getname(uu_dprintf_t *D)
+{
+ return (D->uud_name);
+}
diff --git a/contrib/opensolaris/lib/libuutil/common/uu_ident.c b/contrib/opensolaris/lib/libuutil/common/uu_ident.c
new file mode 100644
index 0000000..9a64384
--- /dev/null
+++ b/contrib/opensolaris/lib/libuutil/common/uu_ident.c
@@ -0,0 +1,122 @@
+/*
+ * CDDL HEADER START
+ *
+ * The contents of this file are subject to the terms of the
+ * Common Development and Distribution License, Version 1.0 only
+ * (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 2004 Sun Microsystems, Inc. All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#pragma ident "%Z%%M% %I% %E% SMI"
+
+#include "libuutil_common.h"
+
+#include <string.h>
+
+/*
+ * We require names of the form:
+ * [provider,]identifier[/[provider,]identifier]...
+ *
+ * Where provider is either a stock symbol (SUNW) or a java-style reversed
+ * domain name (com.sun).
+ *
+ * Both providers and identifiers must start with a letter, and may
+ * only contain alphanumerics, dashes, and underlines. Providers
+ * may also contain periods.
+ *
+ * Note that we do _not_ use the macros in <ctype.h>, since they are affected
+ * by the current locale settings.
+ */
+
+#define IS_ALPHA(c) \
+ (((c) >= 'a' && (c) <= 'z') || ((c) >= 'A' && (c) <= 'Z'))
+
+#define IS_DIGIT(c) \
+ ((c) >= '0' && (c) <= '9')
+
+static int
+is_valid_ident(const char *s, const char *e, int allowdot)
+{
+ char c;
+
+ if (s >= e)
+ return (0); /* name is empty */
+
+ c = *s++;
+ if (!IS_ALPHA(c))
+ return (0); /* does not start with letter */
+
+ while (s < e && (c = *s++) != 0) {
+ if (IS_ALPHA(c) || IS_DIGIT(c) || c == '-' || c == '_' ||
+ (allowdot && c == '.'))
+ continue;
+ return (0); /* invalid character */
+ }
+ return (1);
+}
+
+static int
+is_valid_component(const char *b, const char *e, uint_t flags)
+{
+ char *sp;
+
+ if (flags & UU_NAME_DOMAIN) {
+ sp = strchr(b, ',');
+ if (sp != NULL && sp < e) {
+ if (!is_valid_ident(b, sp, 1))
+ return (0);
+ b = sp + 1;
+ }
+ }
+
+ return (is_valid_ident(b, e, 0));
+}
+
+int
+uu_check_name(const char *name, uint_t flags)
+{
+ const char *end = name + strlen(name);
+ const char *p;
+
+ if (flags & ~(UU_NAME_DOMAIN | UU_NAME_PATH)) {
+ uu_set_error(UU_ERROR_UNKNOWN_FLAG);
+ return (-1);
+ }
+
+ if (!(flags & UU_NAME_PATH)) {
+ if (!is_valid_component(name, end, flags))
+ goto bad;
+ return (0);
+ }
+
+ while ((p = strchr(name, '/')) != NULL) {
+ if (!is_valid_component(name, p - 1, flags))
+ goto bad;
+ name = p + 1;
+ }
+ if (!is_valid_component(name, end, flags))
+ goto bad;
+
+ return (0);
+
+bad:
+ uu_set_error(UU_ERROR_INVALID_ARGUMENT);
+ return (-1);
+}
diff --git a/contrib/opensolaris/lib/libuutil/common/uu_list.c b/contrib/opensolaris/lib/libuutil/common/uu_list.c
new file mode 100644
index 0000000..d9dc86f
--- /dev/null
+++ b/contrib/opensolaris/lib/libuutil/common/uu_list.c
@@ -0,0 +1,711 @@
+/*
+ * CDDL HEADER START
+ *
+ * The contents of this file are subject to the terms of the
+ * Common Development and Distribution License, Version 1.0 only
+ * (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 2005 Sun Microsystems, Inc. All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#pragma ident "%Z%%M% %I% %E% SMI"
+
+#include "libuutil_common.h"
+
+#include <stdlib.h>
+#include <string.h>
+#include <unistd.h>
+#include <sys/time.h>
+
+#define ELEM_TO_NODE(lp, e) \
+ ((uu_list_node_impl_t *)((uintptr_t)(e) + (lp)->ul_offset))
+
+#define NODE_TO_ELEM(lp, n) \
+ ((void *)((uintptr_t)(n) - (lp)->ul_offset))
+
+/*
+ * uu_list_index_ts define a location for insertion. They are simply a
+ * pointer to the object after the insertion point. We store a mark
+ * in the low-bits of the index, to help prevent mistakes.
+ *
+ * When debugging, the index mark changes on every insert and delete, to
+ * catch stale references.
+ */
+#define INDEX_MAX (sizeof (uintptr_t) - 1)
+#define INDEX_NEXT(m) (((m) == INDEX_MAX)? 1 : ((m) + 1) & INDEX_MAX)
+
+#define INDEX_TO_NODE(i) ((uu_list_node_impl_t *)((i) & ~INDEX_MAX))
+#define NODE_TO_INDEX(p, n) (((uintptr_t)(n) & ~INDEX_MAX) | (p)->ul_index)
+#define INDEX_VALID(p, i) (((i) & INDEX_MAX) == (p)->ul_index)
+#define INDEX_CHECK(i) (((i) & INDEX_MAX) != 0)
+
+#define POOL_TO_MARKER(pp) ((void *)((uintptr_t)(pp) | 1))
+
+static uu_list_pool_t uu_null_lpool = { &uu_null_lpool, &uu_null_lpool };
+static pthread_mutex_t uu_lpool_list_lock = PTHREAD_MUTEX_INITIALIZER;
+
+uu_list_pool_t *
+uu_list_pool_create(const char *name, size_t objsize,
+ size_t nodeoffset, uu_compare_fn_t *compare_func, uint32_t flags)
+{
+ uu_list_pool_t *pp, *next, *prev;
+
+ if (name == NULL ||
+ uu_check_name(name, UU_NAME_DOMAIN) == -1 ||
+ nodeoffset + sizeof (uu_list_node_t) > objsize) {
+ uu_set_error(UU_ERROR_INVALID_ARGUMENT);
+ return (NULL);
+ }
+
+ if (flags & ~UU_LIST_POOL_DEBUG) {
+ uu_set_error(UU_ERROR_UNKNOWN_FLAG);
+ return (NULL);
+ }
+
+ pp = uu_zalloc(sizeof (uu_list_pool_t));
+ if (pp == NULL) {
+ uu_set_error(UU_ERROR_NO_MEMORY);
+ return (NULL);
+ }
+
+ (void) strlcpy(pp->ulp_name, name, sizeof (pp->ulp_name));
+ pp->ulp_nodeoffset = nodeoffset;
+ pp->ulp_objsize = objsize;
+ pp->ulp_cmp = compare_func;
+ if (flags & UU_LIST_POOL_DEBUG)
+ pp->ulp_debug = 1;
+ pp->ulp_last_index = 0;
+
+ (void) pthread_mutex_init(&pp->ulp_lock, NULL);
+
+ pp->ulp_null_list.ul_next_enc = UU_PTR_ENCODE(&pp->ulp_null_list);
+ pp->ulp_null_list.ul_prev_enc = UU_PTR_ENCODE(&pp->ulp_null_list);
+
+ (void) pthread_mutex_lock(&uu_lpool_list_lock);
+ pp->ulp_next = next = &uu_null_lpool;
+ pp->ulp_prev = prev = next->ulp_prev;
+ next->ulp_prev = pp;
+ prev->ulp_next = pp;
+ (void) pthread_mutex_unlock(&uu_lpool_list_lock);
+
+ return (pp);
+}
+
+void
+uu_list_pool_destroy(uu_list_pool_t *pp)
+{
+ if (pp->ulp_debug) {
+ if (pp->ulp_null_list.ul_next_enc !=
+ UU_PTR_ENCODE(&pp->ulp_null_list) ||
+ pp->ulp_null_list.ul_prev_enc !=
+ UU_PTR_ENCODE(&pp->ulp_null_list)) {
+ uu_panic("uu_list_pool_destroy: Pool \"%.*s\" (%p) has "
+ "outstanding lists, or is corrupt.\n",
+ sizeof (pp->ulp_name), pp->ulp_name, pp);
+ }
+ }
+ (void) pthread_mutex_lock(&uu_lpool_list_lock);
+ pp->ulp_next->ulp_prev = pp->ulp_prev;
+ pp->ulp_prev->ulp_next = pp->ulp_next;
+ (void) pthread_mutex_unlock(&uu_lpool_list_lock);
+ pp->ulp_prev = NULL;
+ pp->ulp_next = NULL;
+ uu_free(pp);
+}
+
+void
+uu_list_node_init(void *base, uu_list_node_t *np_arg, uu_list_pool_t *pp)
+{
+ uu_list_node_impl_t *np = (uu_list_node_impl_t *)np_arg;
+
+ if (pp->ulp_debug) {
+ uintptr_t offset = (uintptr_t)np - (uintptr_t)base;
+ if (offset + sizeof (*np) > pp->ulp_objsize) {
+ uu_panic("uu_list_node_init(%p, %p, %p (\"%s\")): "
+ "offset %ld doesn't fit in object (size %ld)\n",
+ base, np, pp, pp->ulp_name, offset,
+ pp->ulp_objsize);
+ }
+ if (offset != pp->ulp_nodeoffset) {
+ uu_panic("uu_list_node_init(%p, %p, %p (\"%s\")): "
+ "offset %ld doesn't match pool's offset (%ld)\n",
+ base, np, pp, pp->ulp_name, offset,
+ pp->ulp_objsize);
+ }
+ }
+ np->uln_next = POOL_TO_MARKER(pp);
+ np->uln_prev = NULL;
+}
+
+void
+uu_list_node_fini(void *base, uu_list_node_t *np_arg, uu_list_pool_t *pp)
+{
+ uu_list_node_impl_t *np = (uu_list_node_impl_t *)np_arg;
+
+ if (pp->ulp_debug) {
+ if (np->uln_next == NULL &&
+ np->uln_prev == NULL) {
+ uu_panic("uu_list_node_fini(%p, %p, %p (\"%s\")): "
+ "node already finied\n",
+ base, np_arg, pp, pp->ulp_name);
+ }
+ if (np->uln_next != POOL_TO_MARKER(pp) ||
+ np->uln_prev != NULL) {
+ uu_panic("uu_list_node_fini(%p, %p, %p (\"%s\")): "
+ "node corrupt or on list\n",
+ base, np_arg, pp, pp->ulp_name);
+ }
+ }
+ np->uln_next = NULL;
+ np->uln_prev = NULL;
+}
+
+uu_list_t *
+uu_list_create(uu_list_pool_t *pp, void *parent, uint32_t flags)
+{
+ uu_list_t *lp, *next, *prev;
+
+ if (flags & ~(UU_LIST_DEBUG | UU_LIST_SORTED)) {
+ uu_set_error(UU_ERROR_UNKNOWN_FLAG);
+ return (NULL);
+ }
+
+ if ((flags & UU_LIST_SORTED) && pp->ulp_cmp == NULL) {
+ if (pp->ulp_debug)
+ uu_panic("uu_list_create(%p, ...): requested "
+ "UU_LIST_SORTED, but pool has no comparison func\n",
+ pp);
+ uu_set_error(UU_ERROR_NOT_SUPPORTED);
+ return (NULL);
+ }
+
+ lp = uu_zalloc(sizeof (*lp));
+ if (lp == NULL) {
+ uu_set_error(UU_ERROR_NO_MEMORY);
+ return (NULL);
+ }
+
+ lp->ul_pool = pp;
+ lp->ul_parent_enc = UU_PTR_ENCODE(parent);
+ lp->ul_offset = pp->ulp_nodeoffset;
+ lp->ul_debug = pp->ulp_debug || (flags & UU_LIST_DEBUG);
+ lp->ul_sorted = (flags & UU_LIST_SORTED);
+ lp->ul_numnodes = 0;
+ lp->ul_index = (pp->ulp_last_index = INDEX_NEXT(pp->ulp_last_index));
+
+ lp->ul_null_node.uln_next = &lp->ul_null_node;
+ lp->ul_null_node.uln_prev = &lp->ul_null_node;
+
+ lp->ul_null_walk.ulw_next = &lp->ul_null_walk;
+ lp->ul_null_walk.ulw_prev = &lp->ul_null_walk;
+
+ (void) pthread_mutex_lock(&pp->ulp_lock);
+ next = &pp->ulp_null_list;
+ prev = UU_PTR_DECODE(next->ul_prev_enc);
+ lp->ul_next_enc = UU_PTR_ENCODE(next);
+ lp->ul_prev_enc = UU_PTR_ENCODE(prev);
+ next->ul_prev_enc = UU_PTR_ENCODE(lp);
+ prev->ul_next_enc = UU_PTR_ENCODE(lp);
+ (void) pthread_mutex_unlock(&pp->ulp_lock);
+
+ return (lp);
+}
+
+void
+uu_list_destroy(uu_list_t *lp)
+{
+ uu_list_pool_t *pp = lp->ul_pool;
+
+ if (lp->ul_debug) {
+ if (lp->ul_null_node.uln_next != &lp->ul_null_node ||
+ lp->ul_null_node.uln_prev != &lp->ul_null_node) {
+ uu_panic("uu_list_destroy(%p): list not empty\n",
+ lp);
+ }
+ if (lp->ul_numnodes != 0) {
+ uu_panic("uu_list_destroy(%p): numnodes is nonzero, "
+ "but list is empty\n", lp);
+ }
+ if (lp->ul_null_walk.ulw_next != &lp->ul_null_walk ||
+ lp->ul_null_walk.ulw_prev != &lp->ul_null_walk) {
+ uu_panic("uu_list_destroy(%p): outstanding walkers\n",
+ lp);
+ }
+ }
+
+ (void) pthread_mutex_lock(&pp->ulp_lock);
+ UU_LIST_PTR(lp->ul_next_enc)->ul_prev_enc = lp->ul_prev_enc;
+ UU_LIST_PTR(lp->ul_prev_enc)->ul_next_enc = lp->ul_next_enc;
+ (void) pthread_mutex_unlock(&pp->ulp_lock);
+ lp->ul_prev_enc = UU_PTR_ENCODE(NULL);
+ lp->ul_next_enc = UU_PTR_ENCODE(NULL);
+ lp->ul_pool = NULL;
+ uu_free(lp);
+}
+
+static void
+list_insert(uu_list_t *lp, uu_list_node_impl_t *np, uu_list_node_impl_t *prev,
+ uu_list_node_impl_t *next)
+{
+ if (lp->ul_debug) {
+ if (next->uln_prev != prev || prev->uln_next != next)
+ uu_panic("insert(%p): internal error: %p and %p not "
+ "neighbors\n", lp, next, prev);
+
+ if (np->uln_next != POOL_TO_MARKER(lp->ul_pool) ||
+ np->uln_prev != NULL) {
+ uu_panic("insert(%p): elem %p node %p corrupt, "
+ "not initialized, or already in a list.\n",
+ lp, NODE_TO_ELEM(lp, np), np);
+ }
+ /*
+ * invalidate outstanding uu_list_index_ts.
+ */
+ lp->ul_index = INDEX_NEXT(lp->ul_index);
+ }
+ np->uln_next = next;
+ np->uln_prev = prev;
+ next->uln_prev = np;
+ prev->uln_next = np;
+
+ lp->ul_numnodes++;
+}
+
+void
+uu_list_insert(uu_list_t *lp, void *elem, uu_list_index_t idx)
+{
+ uu_list_node_impl_t *np;
+
+ np = INDEX_TO_NODE(idx);
+ if (np == NULL)
+ np = &lp->ul_null_node;
+
+ if (lp->ul_debug) {
+ if (!INDEX_VALID(lp, idx))
+ uu_panic("uu_list_insert(%p, %p, %p): %s\n",
+ lp, elem, idx,
+ INDEX_CHECK(idx)? "outdated index" :
+ "invalid index");
+ if (np->uln_prev == NULL)
+ uu_panic("uu_list_insert(%p, %p, %p): out-of-date "
+ "index\n", lp, elem, idx);
+ }
+
+ list_insert(lp, ELEM_TO_NODE(lp, elem), np->uln_prev, np);
+}
+
+void *
+uu_list_find(uu_list_t *lp, void *elem, void *private, uu_list_index_t *out)
+{
+ int sorted = lp->ul_sorted;
+ uu_compare_fn_t *func = lp->ul_pool->ulp_cmp;
+ uu_list_node_impl_t *np;
+
+ if (func == NULL) {
+ if (out != NULL)
+ *out = 0;
+ uu_set_error(UU_ERROR_NOT_SUPPORTED);
+ return (NULL);
+ }
+ for (np = lp->ul_null_node.uln_next; np != &lp->ul_null_node;
+ np = np->uln_next) {
+ void *ep = NODE_TO_ELEM(lp, np);
+ int cmp = func(ep, elem, private);
+ if (cmp == 0) {
+ if (out != NULL)
+ *out = NODE_TO_INDEX(lp, np);
+ return (ep);
+ }
+ if (sorted && cmp > 0) {
+ if (out != NULL)
+ *out = NODE_TO_INDEX(lp, np);
+ return (NULL);
+ }
+ }
+ if (out != NULL)
+ *out = NODE_TO_INDEX(lp, 0);
+ return (NULL);
+}
+
+void *
+uu_list_nearest_next(uu_list_t *lp, uu_list_index_t idx)
+{
+ uu_list_node_impl_t *np = INDEX_TO_NODE(idx);
+
+ if (np == NULL)
+ np = &lp->ul_null_node;
+
+ if (lp->ul_debug) {
+ if (!INDEX_VALID(lp, idx))
+ uu_panic("uu_list_nearest_next(%p, %p): %s\n",
+ lp, idx, INDEX_CHECK(idx)? "outdated index" :
+ "invalid index");
+ if (np->uln_prev == NULL)
+ uu_panic("uu_list_nearest_next(%p, %p): out-of-date "
+ "index\n", lp, idx);
+ }
+
+ if (np == &lp->ul_null_node)
+ return (NULL);
+ else
+ return (NODE_TO_ELEM(lp, np));
+}
+
+void *
+uu_list_nearest_prev(uu_list_t *lp, uu_list_index_t idx)
+{
+ uu_list_node_impl_t *np = INDEX_TO_NODE(idx);
+
+ if (np == NULL)
+ np = &lp->ul_null_node;
+
+ if (lp->ul_debug) {
+ if (!INDEX_VALID(lp, idx))
+ uu_panic("uu_list_nearest_prev(%p, %p): %s\n",
+ lp, idx, INDEX_CHECK(idx)? "outdated index" :
+ "invalid index");
+ if (np->uln_prev == NULL)
+ uu_panic("uu_list_nearest_prev(%p, %p): out-of-date "
+ "index\n", lp, idx);
+ }
+
+ if ((np = np->uln_prev) == &lp->ul_null_node)
+ return (NULL);
+ else
+ return (NODE_TO_ELEM(lp, np));
+}
+
+static void
+list_walk_init(uu_list_walk_t *wp, uu_list_t *lp, uint32_t flags)
+{
+ uu_list_walk_t *next, *prev;
+
+ int robust = (flags & UU_WALK_ROBUST);
+ int direction = (flags & UU_WALK_REVERSE)? -1 : 1;
+
+ (void) memset(wp, 0, sizeof (*wp));
+ wp->ulw_list = lp;
+ wp->ulw_robust = robust;
+ wp->ulw_dir = direction;
+ if (direction > 0)
+ wp->ulw_next_result = lp->ul_null_node.uln_next;
+ else
+ wp->ulw_next_result = lp->ul_null_node.uln_prev;
+
+ if (lp->ul_debug || robust) {
+ wp->ulw_next = next = &lp->ul_null_walk;
+ wp->ulw_prev = prev = next->ulw_prev;
+ next->ulw_prev = wp;
+ prev->ulw_next = wp;
+ }
+}
+
+static uu_list_node_impl_t *
+list_walk_advance(uu_list_walk_t *wp, uu_list_t *lp)
+{
+ uu_list_node_impl_t *np = wp->ulw_next_result;
+ uu_list_node_impl_t *next;
+
+ if (np == &lp->ul_null_node)
+ return (NULL);
+
+ next = (wp->ulw_dir > 0)? np->uln_next : np->uln_prev;
+
+ wp->ulw_next_result = next;
+ return (np);
+}
+
+static void
+list_walk_fini(uu_list_walk_t *wp)
+{
+ /* GLXXX debugging? */
+ if (wp->ulw_next != NULL) {
+ wp->ulw_next->ulw_prev = wp->ulw_prev;
+ wp->ulw_prev->ulw_next = wp->ulw_next;
+ wp->ulw_next = NULL;
+ wp->ulw_prev = NULL;
+ }
+ wp->ulw_list = NULL;
+ wp->ulw_next_result = NULL;
+}
+
+uu_list_walk_t *
+uu_list_walk_start(uu_list_t *lp, uint32_t flags)
+{
+ uu_list_walk_t *wp;
+
+ if (flags & ~(UU_WALK_ROBUST | UU_WALK_REVERSE)) {
+ uu_set_error(UU_ERROR_UNKNOWN_FLAG);
+ return (NULL);
+ }
+
+ wp = uu_zalloc(sizeof (*wp));
+ if (wp == NULL) {
+ uu_set_error(UU_ERROR_NO_MEMORY);
+ return (NULL);
+ }
+
+ list_walk_init(wp, lp, flags);
+ return (wp);
+}
+
+void *
+uu_list_walk_next(uu_list_walk_t *wp)
+{
+ uu_list_t *lp = wp->ulw_list;
+ uu_list_node_impl_t *np = list_walk_advance(wp, lp);
+
+ if (np == NULL)
+ return (NULL);
+
+ return (NODE_TO_ELEM(lp, np));
+}
+
+void
+uu_list_walk_end(uu_list_walk_t *wp)
+{
+ list_walk_fini(wp);
+ uu_free(wp);
+}
+
+int
+uu_list_walk(uu_list_t *lp, uu_walk_fn_t *func, void *private, uint32_t flags)
+{
+ uu_list_node_impl_t *np;
+
+ int status = UU_WALK_NEXT;
+
+ int robust = (flags & UU_WALK_ROBUST);
+ int reverse = (flags & UU_WALK_REVERSE);
+
+ if (flags & ~(UU_WALK_ROBUST | UU_WALK_REVERSE)) {
+ uu_set_error(UU_ERROR_UNKNOWN_FLAG);
+ return (-1);
+ }
+
+ if (lp->ul_debug || robust) {
+ uu_list_walk_t my_walk;
+ void *e;
+
+ list_walk_init(&my_walk, lp, flags);
+ while (status == UU_WALK_NEXT &&
+ (e = uu_list_walk_next(&my_walk)) != NULL)
+ status = (*func)(e, private);
+ list_walk_fini(&my_walk);
+ } else {
+ if (!reverse) {
+ for (np = lp->ul_null_node.uln_next;
+ status == UU_WALK_NEXT && np != &lp->ul_null_node;
+ np = np->uln_next) {
+ status = (*func)(NODE_TO_ELEM(lp, np), private);
+ }
+ } else {
+ for (np = lp->ul_null_node.uln_prev;
+ status == UU_WALK_NEXT && np != &lp->ul_null_node;
+ np = np->uln_prev) {
+ status = (*func)(NODE_TO_ELEM(lp, np), private);
+ }
+ }
+ }
+ if (status >= 0)
+ return (0);
+ uu_set_error(UU_ERROR_CALLBACK_FAILED);
+ return (-1);
+}
+
+void
+uu_list_remove(uu_list_t *lp, void *elem)
+{
+ uu_list_node_impl_t *np = ELEM_TO_NODE(lp, elem);
+ uu_list_walk_t *wp;
+
+ if (lp->ul_debug) {
+ if (np->uln_prev == NULL)
+ uu_panic("uu_list_remove(%p, %p): elem not on list\n",
+ lp, elem);
+ /*
+ * invalidate outstanding uu_list_index_ts.
+ */
+ lp->ul_index = INDEX_NEXT(lp->ul_index);
+ }
+
+ /*
+ * robust walkers must be advanced. In debug mode, non-robust
+ * walkers are also on the list. If there are any, it's an error.
+ */
+ for (wp = lp->ul_null_walk.ulw_next; wp != &lp->ul_null_walk;
+ wp = wp->ulw_next) {
+ if (wp->ulw_robust) {
+ if (np == wp->ulw_next_result)
+ (void) list_walk_advance(wp, lp);
+ } else if (wp->ulw_next_result != NULL) {
+ uu_panic("uu_list_remove(%p, %p): active non-robust "
+ "walker\n", lp, elem);
+ }
+ }
+
+ np->uln_next->uln_prev = np->uln_prev;
+ np->uln_prev->uln_next = np->uln_next;
+
+ lp->ul_numnodes--;
+
+ np->uln_next = POOL_TO_MARKER(lp->ul_pool);
+ np->uln_prev = NULL;
+}
+
+void *
+uu_list_teardown(uu_list_t *lp, void **cookie)
+{
+ void *ep;
+
+ /*
+ * XXX: disable list modification until list is empty
+ */
+ if (lp->ul_debug && *cookie != NULL)
+ uu_panic("uu_list_teardown(%p, %p): unexpected cookie\n", lp,
+ cookie);
+
+ ep = uu_list_first(lp);
+ if (ep)
+ uu_list_remove(lp, ep);
+ return (ep);
+}
+
+int
+uu_list_insert_before(uu_list_t *lp, void *target, void *elem)
+{
+ uu_list_node_impl_t *np = ELEM_TO_NODE(lp, target);
+
+ if (target == NULL)
+ np = &lp->ul_null_node;
+
+ if (lp->ul_debug) {
+ if (np->uln_prev == NULL)
+ uu_panic("uu_list_insert_before(%p, %p, %p): %p is "
+ "not currently on a list\n",
+ lp, target, elem, target);
+ }
+ if (lp->ul_sorted) {
+ if (lp->ul_debug)
+ uu_panic("uu_list_insert_before(%p, ...): list is "
+ "UU_LIST_SORTED\n", lp);
+ uu_set_error(UU_ERROR_NOT_SUPPORTED);
+ return (-1);
+ }
+
+ list_insert(lp, ELEM_TO_NODE(lp, elem), np->uln_prev, np);
+ return (0);
+}
+
+int
+uu_list_insert_after(uu_list_t *lp, void *target, void *elem)
+{
+ uu_list_node_impl_t *np = ELEM_TO_NODE(lp, target);
+
+ if (target == NULL)
+ np = &lp->ul_null_node;
+
+ if (lp->ul_debug) {
+ if (np->uln_prev == NULL)
+ uu_panic("uu_list_insert_after(%p, %p, %p): %p is "
+ "not currently on a list\n",
+ lp, target, elem, target);
+ }
+ if (lp->ul_sorted) {
+ if (lp->ul_debug)
+ uu_panic("uu_list_insert_after(%p, ...): list is "
+ "UU_LIST_SORTED\n", lp);
+ uu_set_error(UU_ERROR_NOT_SUPPORTED);
+ return (-1);
+ }
+
+ list_insert(lp, ELEM_TO_NODE(lp, elem), np, np->uln_next);
+ return (0);
+}
+
+size_t
+uu_list_numnodes(uu_list_t *lp)
+{
+ return (lp->ul_numnodes);
+}
+
+void *
+uu_list_first(uu_list_t *lp)
+{
+ uu_list_node_impl_t *n = lp->ul_null_node.uln_next;
+ if (n == &lp->ul_null_node)
+ return (NULL);
+ return (NODE_TO_ELEM(lp, n));
+}
+
+void *
+uu_list_last(uu_list_t *lp)
+{
+ uu_list_node_impl_t *n = lp->ul_null_node.uln_prev;
+ if (n == &lp->ul_null_node)
+ return (NULL);
+ return (NODE_TO_ELEM(lp, n));
+}
+
+void *
+uu_list_next(uu_list_t *lp, void *elem)
+{
+ uu_list_node_impl_t *n = ELEM_TO_NODE(lp, elem);
+
+ n = n->uln_next;
+ if (n == &lp->ul_null_node)
+ return (NULL);
+ return (NODE_TO_ELEM(lp, n));
+}
+
+void *
+uu_list_prev(uu_list_t *lp, void *elem)
+{
+ uu_list_node_impl_t *n = ELEM_TO_NODE(lp, elem);
+
+ n = n->uln_prev;
+ if (n == &lp->ul_null_node)
+ return (NULL);
+ return (NODE_TO_ELEM(lp, n));
+}
+
+/*
+ * called from uu_lockup() and uu_release(), as part of our fork1()-safety.
+ */
+void
+uu_list_lockup(void)
+{
+ uu_list_pool_t *pp;
+
+ (void) pthread_mutex_lock(&uu_lpool_list_lock);
+ for (pp = uu_null_lpool.ulp_next; pp != &uu_null_lpool;
+ pp = pp->ulp_next)
+ (void) pthread_mutex_lock(&pp->ulp_lock);
+}
+
+void
+uu_list_release(void)
+{
+ uu_list_pool_t *pp;
+
+ for (pp = uu_null_lpool.ulp_next; pp != &uu_null_lpool;
+ pp = pp->ulp_next)
+ (void) pthread_mutex_unlock(&pp->ulp_lock);
+ (void) pthread_mutex_unlock(&uu_lpool_list_lock);
+}
diff --git a/contrib/opensolaris/lib/libuutil/common/uu_misc.c b/contrib/opensolaris/lib/libuutil/common/uu_misc.c
new file mode 100644
index 0000000..fb0c32b
--- /dev/null
+++ b/contrib/opensolaris/lib/libuutil/common/uu_misc.c
@@ -0,0 +1,250 @@
+/*
+ * 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"
+
+#include "libuutil_common.h"
+
+#include <assert.h>
+#include <errno.h>
+#include <libintl.h>
+#include <pthread.h>
+#include <stdarg.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/debug.h>
+#include <thread.h>
+#include <unistd.h>
+
+#if !defined(TEXT_DOMAIN)
+#define TEXT_DOMAIN "SYS_TEST"
+#endif
+
+/*
+ * All of the old code under !defined(PTHREAD_ONCE_KEY_NP)
+ * is here to enable the building of a native version of
+ * libuutil.so when the build machine has not yet been upgraded
+ * to a version of libc that provides pthread_key_create_once_np().
+ * It should all be deleted when solaris_nevada ships.
+ * The code is not MT-safe in a relaxed memory model.
+ */
+
+#if defined(PTHREAD_ONCE_KEY_NP)
+static pthread_key_t uu_error_key = PTHREAD_ONCE_KEY_NP;
+#else /* PTHREAD_ONCE_KEY_NP */
+static pthread_key_t uu_error_key = 0;
+static pthread_mutex_t uu_key_lock = PTHREAD_MUTEX_INITIALIZER;
+#endif /* PTHREAD_ONCE_KEY_NP */
+
+static int uu_error_key_setup = 0;
+
+static pthread_mutex_t uu_panic_lock = PTHREAD_MUTEX_INITIALIZER;
+/* LINTED static unused */
+static const char *uu_panic_format;
+/* LINTED static unused */
+static va_list uu_panic_args;
+static pthread_t uu_panic_thread;
+
+static uint32_t _uu_main_error;
+
+void
+uu_set_error(uint_t code)
+{
+
+#if defined(PTHREAD_ONCE_KEY_NP)
+ if (pthread_key_create_once_np(&uu_error_key, NULL) != 0)
+ uu_error_key_setup = -1;
+ else
+ uu_error_key_setup = 1;
+#else /* PTHREAD_ONCE_KEY_NP */
+ if (uu_error_key_setup == 0) {
+ (void) pthread_mutex_lock(&uu_key_lock);
+ if (uu_error_key_setup == 0) {
+ if (pthread_key_create(&uu_error_key, NULL) != 0)
+ uu_error_key_setup = -1;
+ else
+ uu_error_key_setup = 1;
+ }
+ (void) pthread_mutex_unlock(&uu_key_lock);
+ }
+#endif /* PTHREAD_ONCE_KEY_NP */
+ if (uu_error_key_setup > 0)
+ (void) pthread_setspecific(uu_error_key,
+ (void *)(uintptr_t)code);
+}
+
+uint32_t
+uu_error(void)
+{
+
+ if (uu_error_key_setup < 0) /* can't happen? */
+ return (UU_ERROR_UNKNOWN);
+
+ /*
+ * Because UU_ERROR_NONE == 0, if uu_set_error() was
+ * never called, then this will return UU_ERROR_NONE:
+ */
+ return ((uint32_t)(uintptr_t)pthread_getspecific(uu_error_key));
+}
+
+const char *
+uu_strerror(uint32_t code)
+{
+ const char *str;
+
+ switch (code) {
+ case UU_ERROR_NONE:
+ str = dgettext(TEXT_DOMAIN, "No error");
+ break;
+
+ case UU_ERROR_INVALID_ARGUMENT:
+ str = dgettext(TEXT_DOMAIN, "Invalid argument");
+ break;
+
+ case UU_ERROR_UNKNOWN_FLAG:
+ str = dgettext(TEXT_DOMAIN, "Unknown flag passed");
+ break;
+
+ case UU_ERROR_NO_MEMORY:
+ str = dgettext(TEXT_DOMAIN, "Out of memory");
+ break;
+
+ case UU_ERROR_CALLBACK_FAILED:
+ str = dgettext(TEXT_DOMAIN, "Callback-initiated failure");
+ break;
+
+ case UU_ERROR_NOT_SUPPORTED:
+ str = dgettext(TEXT_DOMAIN, "Operation not supported");
+ break;
+
+ case UU_ERROR_EMPTY:
+ str = dgettext(TEXT_DOMAIN, "No value provided");
+ break;
+
+ case UU_ERROR_UNDERFLOW:
+ str = dgettext(TEXT_DOMAIN, "Value too small");
+ break;
+
+ case UU_ERROR_OVERFLOW:
+ str = dgettext(TEXT_DOMAIN, "Value too large");
+ break;
+
+ case UU_ERROR_INVALID_CHAR:
+ str = dgettext(TEXT_DOMAIN,
+ "Value contains unexpected character");
+ break;
+
+ case UU_ERROR_INVALID_DIGIT:
+ str = dgettext(TEXT_DOMAIN,
+ "Value contains digit not in base");
+ break;
+
+ case UU_ERROR_SYSTEM:
+ str = dgettext(TEXT_DOMAIN, "Underlying system error");
+ break;
+
+ case UU_ERROR_UNKNOWN:
+ str = dgettext(TEXT_DOMAIN, "Error status not known");
+ break;
+
+ default:
+ errno = ESRCH;
+ str = NULL;
+ break;
+ }
+ return (str);
+}
+
+void
+uu_panic(const char *format, ...)
+{
+ va_list args;
+
+ va_start(args, format);
+
+ (void) pthread_mutex_lock(&uu_panic_lock);
+ if (uu_panic_thread == 0) {
+ uu_panic_thread = pthread_self();
+ uu_panic_format = format;
+ va_copy(uu_panic_args, args);
+ }
+ (void) pthread_mutex_unlock(&uu_panic_lock);
+
+ (void) vfprintf(stderr, format, args);
+
+ if (uu_panic_thread == pthread_self())
+ abort();
+ else
+ for (;;)
+ (void) pause();
+}
+
+int
+assfail(const char *astring, const char *file, int line)
+{
+ __assert(astring, file, line);
+ /*NOTREACHED*/
+ return (0);
+}
+
+static void
+uu_lockup(void)
+{
+ (void) pthread_mutex_lock(&uu_panic_lock);
+#if !defined(PTHREAD_ONCE_KEY_NP)
+ (void) pthread_mutex_lock(&uu_key_lock);
+#endif
+ uu_avl_lockup();
+ uu_list_lockup();
+}
+
+static void
+uu_release(void)
+{
+ (void) pthread_mutex_unlock(&uu_panic_lock);
+#if !defined(PTHREAD_ONCE_KEY_NP)
+ (void) pthread_mutex_unlock(&uu_key_lock);
+#endif
+ uu_avl_release();
+ uu_list_release();
+}
+
+static void
+uu_release_child(void)
+{
+ uu_panic_format = NULL;
+ uu_panic_thread = 0;
+
+ uu_release();
+}
+
+#pragma init(uu_init)
+static void
+uu_init(void)
+{
+ (void) pthread_atfork(uu_lockup, uu_release, uu_release_child);
+}
diff --git a/contrib/opensolaris/lib/libuutil/common/uu_open.c b/contrib/opensolaris/lib/libuutil/common/uu_open.c
new file mode 100644
index 0000000..7256662
--- /dev/null
+++ b/contrib/opensolaris/lib/libuutil/common/uu_open.c
@@ -0,0 +1,70 @@
+/*
+ * CDDL HEADER START
+ *
+ * The contents of this file are subject to the terms of the
+ * Common Development and Distribution License, Version 1.0 only
+ * (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 2004 Sun Microsystems, Inc. All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#pragma ident "%Z%%M% %I% %E% SMI"
+
+#include "libuutil_common.h"
+
+#include <sys/time.h>
+
+#include <errno.h>
+#include <fcntl.h>
+#include <limits.h>
+#include <stdio.h>
+#include <unistd.h>
+
+#ifdef _LP64
+#define TMPPATHFMT "%s/uu%ld"
+#else /* _LP64 */
+#define TMPPATHFMT "%s/uu%lld"
+#endif /* _LP64 */
+
+/*ARGSUSED*/
+int
+uu_open_tmp(const char *dir, uint_t uflags)
+{
+ int f;
+ char *fname = uu_zalloc(PATH_MAX);
+
+ if (fname == NULL)
+ return (-1);
+
+ for (;;) {
+ (void) snprintf(fname, PATH_MAX, "%s/uu%lld", dir, gethrtime());
+
+ f = open(fname, O_CREAT | O_EXCL | O_RDWR, 0600);
+
+ if (f >= 0 || errno != EEXIST)
+ break;
+ }
+
+ if (f >= 0)
+ (void) unlink(fname);
+
+ uu_free(fname);
+
+ return (f);
+}
diff --git a/contrib/opensolaris/lib/libuutil/common/uu_pname.c b/contrib/opensolaris/lib/libuutil/common/uu_pname.c
new file mode 100644
index 0000000..20626ac
--- /dev/null
+++ b/contrib/opensolaris/lib/libuutil/common/uu_pname.c
@@ -0,0 +1,205 @@
+/*
+ * CDDL HEADER START
+ *
+ * The contents of this file are subject to the terms of the
+ * Common Development and Distribution License, Version 1.0 only
+ * (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 2005 Sun Microsystems, Inc. All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#pragma ident "%Z%%M% %I% %E% SMI"
+
+#include "libuutil_common.h"
+
+#include <libintl.h>
+#include <limits.h>
+#include <string.h>
+#include <stdlib.h>
+#include <stdarg.h>
+#include <stdio.h>
+#include <errno.h>
+#include <wchar.h>
+#include <unistd.h>
+
+static const char PNAME_FMT[] = "%s: ";
+static const char ERRNO_FMT[] = ": %s\n";
+
+static const char *pname;
+
+static void
+uu_die_internal(int status, const char *format, va_list alist) __NORETURN;
+
+int uu_exit_ok_value = EXIT_SUCCESS;
+int uu_exit_fatal_value = EXIT_FAILURE;
+int uu_exit_usage_value = 2;
+
+int *
+uu_exit_ok(void)
+{
+ return (&uu_exit_ok_value);
+}
+
+int *
+uu_exit_fatal(void)
+{
+ return (&uu_exit_fatal_value);
+}
+
+int *
+uu_exit_usage(void)
+{
+ return (&uu_exit_usage_value);
+}
+
+void
+uu_alt_exit(int profile)
+{
+ switch (profile) {
+ case UU_PROFILE_DEFAULT:
+ uu_exit_ok_value = EXIT_SUCCESS;
+ uu_exit_fatal_value = EXIT_FAILURE;
+ uu_exit_usage_value = 2;
+ break;
+ case UU_PROFILE_LAUNCHER:
+ uu_exit_ok_value = EXIT_SUCCESS;
+ uu_exit_fatal_value = 124;
+ uu_exit_usage_value = 125;
+ break;
+ }
+}
+
+static void
+uu_warn_internal(int err, const char *format, va_list alist)
+{
+ if (pname != NULL)
+ (void) fprintf(stderr, PNAME_FMT, pname);
+
+ (void) vfprintf(stderr, format, alist);
+
+ if (strrchr(format, '\n') == NULL)
+ (void) fprintf(stderr, ERRNO_FMT, strerror(err));
+}
+
+void
+uu_vwarn(const char *format, va_list alist)
+{
+ uu_warn_internal(errno, format, alist);
+}
+
+/*PRINTFLIKE1*/
+void
+uu_warn(const char *format, ...)
+{
+ va_list alist;
+ va_start(alist, format);
+ uu_warn_internal(errno, format, alist);
+ va_end(alist);
+}
+
+static void
+uu_die_internal(int status, const char *format, va_list alist)
+{
+ uu_warn_internal(errno, format, alist);
+#ifdef DEBUG
+ {
+ char *cp;
+
+ if (!issetugid()) {
+ cp = getenv("UU_DIE_ABORTS");
+ if (cp != NULL && *cp != '\0')
+ abort();
+ }
+ }
+#endif
+ exit(status);
+}
+
+void
+uu_vdie(const char *format, va_list alist)
+{
+ uu_die_internal(UU_EXIT_FATAL, format, alist);
+}
+
+/*PRINTFLIKE1*/
+void
+uu_die(const char *format, ...)
+{
+ va_list alist;
+ va_start(alist, format);
+ uu_die_internal(UU_EXIT_FATAL, format, alist);
+ va_end(alist);
+}
+
+void
+uu_vxdie(int status, const char *format, va_list alist)
+{
+ uu_die_internal(status, format, alist);
+}
+
+/*PRINTFLIKE2*/
+void
+uu_xdie(int status, const char *format, ...)
+{
+ va_list alist;
+ va_start(alist, format);
+ uu_die_internal(status, format, alist);
+ va_end(alist);
+}
+
+const char *
+uu_setpname(char *arg0)
+{
+ /*
+ * Having a NULL argv[0], while uncommon, is possible. It
+ * makes more sense to handle this event in uu_setpname rather
+ * than in each of its consumers.
+ */
+ if (arg0 == NULL) {
+ pname = "unknown_command";
+ return (pname);
+ }
+
+ /*
+ * Guard against '/' at end of command invocation.
+ */
+ for (;;) {
+ char *p = strrchr(arg0, '/');
+ if (p == NULL) {
+ pname = arg0;
+ break;
+ } else {
+ if (*(p + 1) == '\0') {
+ *p = '\0';
+ continue;
+ }
+
+ pname = p + 1;
+ break;
+ }
+ }
+
+ return (pname);
+}
+
+const char *
+uu_getpname(void)
+{
+ return (pname);
+}
diff --git a/contrib/opensolaris/lib/libuutil/common/uu_strtoint.c b/contrib/opensolaris/lib/libuutil/common/uu_strtoint.c
new file mode 100644
index 0000000..8fd1148
--- /dev/null
+++ b/contrib/opensolaris/lib/libuutil/common/uu_strtoint.c
@@ -0,0 +1,300 @@
+/*
+ * CDDL HEADER START
+ *
+ * The contents of this file are subject to the terms of the
+ * Common Development and Distribution License, Version 1.0 only
+ * (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 2004 Sun Microsystems, Inc. All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#pragma ident "%Z%%M% %I% %E% SMI"
+
+#include "libuutil_common.h"
+
+#include <limits.h>
+#include <ctype.h>
+
+#define MAX_BASE 36
+
+#define IS_DIGIT(x) ((x) >= '0' && (x) <= '9')
+
+#define CTOI(x) (((x) >= '0' && (x) <= '9') ? (x) - '0' : \
+ ((x) >= 'a' && (x) <= 'z') ? (x) + 10 - 'a' : (x) + 10 - 'A')
+
+static int
+strtoint(const char *s_arg, uint64_t *out, uint32_t base, int sign)
+{
+ const unsigned char *s = (const unsigned char *)s_arg;
+
+ uint64_t val = 0;
+ uint64_t multmax;
+
+ unsigned c, i;
+
+ int neg = 0;
+
+ int bad_digit = 0;
+ int bad_char = 0;
+ int overflow = 0;
+
+ if (s == NULL || base == 1 || base > MAX_BASE) {
+ uu_set_error(UU_ERROR_INVALID_ARGUMENT);
+ return (-1);
+ }
+
+ while ((c = *s) != 0 && isspace(c))
+ s++;
+
+ switch (c) {
+ case '-':
+ if (!sign)
+ overflow = 1; /* becomes underflow below */
+ neg = 1;
+ /*FALLTHRU*/
+ case '+':
+ c = *++s;
+ break;
+ default:
+ break;
+ }
+
+ if (c == '\0') {
+ uu_set_error(UU_ERROR_EMPTY);
+ return (-1);
+ }
+
+ if (base == 0) {
+ if (c != '0')
+ base = 10;
+ else if (s[1] == 'x' || s[1] == 'X')
+ base = 16;
+ else
+ base = 8;
+ }
+
+ if (base == 16 && c == '0' && (s[1] == 'x' || s[1] == 'X'))
+ c = *(s += 2);
+
+ if ((val = CTOI(c)) >= base) {
+ if (IS_DIGIT(c))
+ bad_digit = 1;
+ else
+ bad_char = 1;
+ val = 0;
+ }
+
+ multmax = (uint64_t)UINT64_MAX / (uint64_t)base;
+
+ for (c = *++s; c != '\0'; c = *++s) {
+ if ((i = CTOI(c)) >= base) {
+ if (isspace(c))
+ break;
+ if (IS_DIGIT(c))
+ bad_digit = 1;
+ else
+ bad_char = 1;
+ i = 0;
+ }
+
+ if (val > multmax)
+ overflow = 1;
+
+ val *= base;
+ if ((uint64_t)UINT64_MAX - val < (uint64_t)i)
+ overflow = 1;
+
+ val += i;
+ }
+
+ while ((c = *s) != 0) {
+ if (!isspace(c))
+ bad_char = 1;
+ s++;
+ }
+
+ if (sign) {
+ if (neg) {
+ if (val > -(uint64_t)INT64_MIN)
+ overflow = 1;
+ } else {
+ if (val > INT64_MAX)
+ overflow = 1;
+ }
+ }
+
+ if (neg)
+ val = -val;
+
+ if (bad_char | bad_digit | overflow) {
+ if (bad_char)
+ uu_set_error(UU_ERROR_INVALID_CHAR);
+ else if (bad_digit)
+ uu_set_error(UU_ERROR_INVALID_DIGIT);
+ else if (overflow) {
+ if (neg)
+ uu_set_error(UU_ERROR_UNDERFLOW);
+ else
+ uu_set_error(UU_ERROR_OVERFLOW);
+ }
+ return (-1);
+ }
+
+ *out = val;
+ return (0);
+}
+
+int
+uu_strtoint(const char *s, void *v, size_t sz, int base,
+ int64_t min, int64_t max)
+{
+ uint64_t val_u;
+ int64_t val;
+
+ if (min > max)
+ goto bad_argument;
+
+ switch (sz) {
+ case 1:
+ if (max > INT8_MAX || min < INT8_MIN)
+ goto bad_argument;
+ break;
+ case 2:
+ if (max > INT16_MAX || min < INT16_MIN)
+ goto bad_argument;
+ break;
+ case 4:
+ if (max > INT32_MAX || min < INT32_MIN)
+ goto bad_argument;
+ break;
+ case 8:
+ if (max > INT64_MAX || min < INT64_MIN)
+ goto bad_argument;
+ break;
+ default:
+ goto bad_argument;
+ }
+
+ if (min == 0 && max == 0) {
+ min = -(1ULL << (8 * sz - 1));
+ max = (1ULL << (8 * sz - 1)) - 1;
+ }
+
+ if (strtoint(s, &val_u, base, 1) == -1)
+ return (-1);
+
+ val = (int64_t)val_u;
+
+ if (val < min) {
+ uu_set_error(UU_ERROR_UNDERFLOW);
+ return (-1);
+ } else if (val > max) {
+ uu_set_error(UU_ERROR_OVERFLOW);
+ return (-1);
+ }
+
+ switch (sz) {
+ case 1:
+ *(int8_t *)v = val;
+ return (0);
+ case 2:
+ *(int16_t *)v = val;
+ return (0);
+ case 4:
+ *(int32_t *)v = val;
+ return (0);
+ case 8:
+ *(int64_t *)v = val;
+ return (0);
+ default:
+ break; /* fall through to bad_argument */
+ }
+
+bad_argument:
+ uu_set_error(UU_ERROR_INVALID_ARGUMENT);
+ return (-1);
+}
+
+int
+uu_strtouint(const char *s, void *v, size_t sz, int base,
+ uint64_t min, uint64_t max)
+{
+ uint64_t val;
+
+ if (min > max)
+ goto bad_argument;
+
+ switch (sz) {
+ case 1:
+ if (max > UINT8_MAX)
+ goto bad_argument;
+ break;
+ case 2:
+ if (max > UINT16_MAX)
+ goto bad_argument;
+ break;
+ case 4:
+ if (max > UINT32_MAX)
+ goto bad_argument;
+ break;
+ case 8:
+ if (max > UINT64_MAX)
+ goto bad_argument;
+ break;
+ default:
+ goto bad_argument;
+ }
+
+ if (min == 0 && max == 0) {
+ /* we have to be careful, since << can overflow */
+ max = (1ULL << (8 * sz - 1)) * 2 - 1;
+ }
+
+ if (strtoint(s, &val, base, 0) == -1)
+ return (-1);
+
+ if (val < min) {
+ uu_set_error(UU_ERROR_UNDERFLOW);
+ return (-1);
+ } else if (val > max) {
+ uu_set_error(UU_ERROR_OVERFLOW);
+ return (-1);
+ }
+
+ switch (sz) {
+ case 1:
+ *(uint8_t *)v = val;
+ return (0);
+ case 2:
+ *(uint16_t *)v = val;
+ return (0);
+ case 4:
+ *(uint32_t *)v = val;
+ return (0);
+ case 8:
+ *(uint64_t *)v = val;
+ return (0);
+ default:
+ break; /* shouldn't happen, fall through */
+ }
+
+bad_argument:
+ uu_set_error(UU_ERROR_INVALID_ARGUMENT);
+ return (-1);
+}
diff --git a/contrib/opensolaris/lib/libzfs/common/libzfs.h b/contrib/opensolaris/lib/libzfs/common/libzfs.h
new file mode 100644
index 0000000..8e9a848
--- /dev/null
+++ b/contrib/opensolaris/lib/libzfs/common/libzfs.h
@@ -0,0 +1,441 @@
+/*
+ * 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.
+ */
+
+#ifndef _LIBZFS_H
+#define _LIBZFS_H
+
+#pragma ident "%Z%%M% %I% %E% SMI"
+
+#include <assert.h>
+#include <libnvpair.h>
+#include <sys/param.h>
+#include <sys/types.h>
+#include <sys/varargs.h>
+#include <sys/fs/zfs.h>
+#include <sys/zfs_ioctl.h>
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/*
+ * Miscellaneous ZFS constants
+ */
+#define ZFS_MAXNAMELEN MAXNAMELEN
+#define ZPOOL_MAXNAMELEN MAXNAMELEN
+#define ZFS_MAXPROPLEN MAXPATHLEN
+
+/*
+ * libzfs errors
+ */
+enum {
+ EZFS_NOMEM = 2000, /* out of memory */
+ EZFS_BADPROP, /* invalid property value */
+ EZFS_PROPREADONLY, /* cannot set readonly property */
+ EZFS_PROPTYPE, /* property does not apply to dataset type */
+ EZFS_PROPNONINHERIT, /* property is not inheritable */
+ EZFS_PROPSPACE, /* bad quota or reservation */
+ EZFS_BADTYPE, /* dataset is not of appropriate type */
+ EZFS_BUSY, /* pool or dataset is busy */
+ EZFS_EXISTS, /* pool or dataset already exists */
+ EZFS_NOENT, /* no such pool or dataset */
+ EZFS_BADSTREAM, /* bad backup stream */
+ EZFS_DSREADONLY, /* dataset is readonly */
+ EZFS_VOLTOOBIG, /* volume is too large for 32-bit system */
+ EZFS_VOLHASDATA, /* volume already contains data */
+ EZFS_INVALIDNAME, /* invalid dataset name */
+ EZFS_BADRESTORE, /* unable to restore to destination */
+ EZFS_BADBACKUP, /* backup failed */
+ EZFS_BADTARGET, /* bad attach/detach/replace target */
+ EZFS_NODEVICE, /* no such device in pool */
+ EZFS_BADDEV, /* invalid device to add */
+ EZFS_NOREPLICAS, /* no valid replicas */
+ EZFS_RESILVERING, /* currently resilvering */
+ EZFS_BADVERSION, /* unsupported version */
+ EZFS_POOLUNAVAIL, /* pool is currently unavailable */
+ EZFS_DEVOVERFLOW, /* too many devices in one vdev */
+ EZFS_BADPATH, /* must be an absolute path */
+ EZFS_CROSSTARGET, /* rename or clone across pool or dataset */
+ EZFS_ZONED, /* used improperly in local zone */
+ EZFS_MOUNTFAILED, /* failed to mount dataset */
+ EZFS_UMOUNTFAILED, /* failed to unmount dataset */
+ EZFS_UNSHARENFSFAILED, /* unshare(1M) failed */
+ EZFS_SHARENFSFAILED, /* share(1M) failed */
+ EZFS_DEVLINKS, /* failed to create zvol links */
+ EZFS_PERM, /* permission denied */
+ EZFS_NOSPC, /* out of space */
+ EZFS_IO, /* I/O error */
+ EZFS_INTR, /* signal received */
+ EZFS_ISSPARE, /* device is a hot spare */
+ EZFS_INVALCONFIG, /* invalid vdev configuration */
+ EZFS_RECURSIVE, /* recursive dependency */
+ EZFS_NOHISTORY, /* no history object */
+ EZFS_UNSHAREISCSIFAILED, /* iscsitgtd failed request to unshare */
+ EZFS_SHAREISCSIFAILED, /* iscsitgtd failed request to share */
+ EZFS_POOLPROPS, /* couldn't retrieve pool props */
+ EZFS_POOL_NOTSUP, /* ops not supported for this type of pool */
+ EZFS_POOL_INVALARG, /* invalid argument for this pool operation */
+ EZFS_UNKNOWN
+};
+
+/*
+ * Basic handle types
+ */
+typedef struct zfs_handle zfs_handle_t;
+typedef struct zpool_handle zpool_handle_t;
+typedef struct libzfs_handle libzfs_handle_t;
+
+/*
+ * Library initialization
+ */
+extern libzfs_handle_t *libzfs_init(void);
+extern void libzfs_fini(libzfs_handle_t *);
+
+extern libzfs_handle_t *zpool_get_handle(zpool_handle_t *);
+extern libzfs_handle_t *zfs_get_handle(zfs_handle_t *);
+
+extern void libzfs_print_on_error(libzfs_handle_t *, boolean_t);
+
+extern int libzfs_errno(libzfs_handle_t *);
+extern const char *libzfs_error_action(libzfs_handle_t *);
+extern const char *libzfs_error_description(libzfs_handle_t *);
+
+/*
+ * Basic handle functions
+ */
+extern zpool_handle_t *zpool_open(libzfs_handle_t *, const char *);
+extern zpool_handle_t *zpool_open_canfail(libzfs_handle_t *, const char *);
+extern void zpool_close(zpool_handle_t *);
+extern const char *zpool_get_name(zpool_handle_t *);
+extern uint64_t zpool_get_guid(zpool_handle_t *);
+extern uint64_t zpool_get_space_used(zpool_handle_t *);
+extern uint64_t zpool_get_space_total(zpool_handle_t *);
+extern int zpool_get_root(zpool_handle_t *, char *, size_t);
+extern int zpool_get_state(zpool_handle_t *);
+extern uint64_t zpool_get_version(zpool_handle_t *);
+
+/*
+ * Iterate over all active pools in the system.
+ */
+typedef int (*zpool_iter_f)(zpool_handle_t *, void *);
+extern int zpool_iter(libzfs_handle_t *, zpool_iter_f, void *);
+
+/*
+ * Functions to create and destroy pools
+ */
+extern int zpool_create(libzfs_handle_t *, const char *, nvlist_t *,
+ const char *);
+extern int zpool_destroy(zpool_handle_t *);
+extern int zpool_add(zpool_handle_t *, nvlist_t *);
+
+/*
+ * Functions to manipulate pool and vdev state
+ */
+extern int zpool_scrub(zpool_handle_t *, pool_scrub_type_t);
+
+extern int zpool_vdev_online(zpool_handle_t *, const char *);
+extern int zpool_vdev_offline(zpool_handle_t *, const char *, int);
+extern int zpool_vdev_attach(zpool_handle_t *, const char *, const char *,
+ nvlist_t *, int);
+extern int zpool_vdev_detach(zpool_handle_t *, const char *);
+extern int zpool_vdev_remove(zpool_handle_t *, const char *);
+extern int zpool_clear(zpool_handle_t *, const char *);
+extern nvlist_t *zpool_find_vdev(zpool_handle_t *, const char *, boolean_t *);
+
+/*
+ * Functions to manage pool properties
+ */
+extern int zpool_set_prop(zpool_handle_t *, const char *, const char *);
+extern int zpool_get_prop(zpool_handle_t *, zfs_prop_t, char *,
+ size_t proplen, zfs_source_t *);
+extern const char *zpool_prop_to_name(zpool_prop_t);
+extern const char *zpool_prop_values(zpool_prop_t);
+
+/*
+ * Pool health statistics.
+ */
+typedef enum {
+ /*
+ * The following correspond to faults as defined in the (fault.fs.zfs.*)
+ * event namespace. Each is associated with a corresponding message ID.
+ */
+ ZPOOL_STATUS_CORRUPT_CACHE, /* corrupt /kernel/drv/zpool.cache */
+ ZPOOL_STATUS_MISSING_DEV_R, /* missing device with replicas */
+ ZPOOL_STATUS_MISSING_DEV_NR, /* missing device with no replicas */
+ ZPOOL_STATUS_CORRUPT_LABEL_R, /* bad device label with replicas */
+ ZPOOL_STATUS_CORRUPT_LABEL_NR, /* bad device label with no replicas */
+ ZPOOL_STATUS_BAD_GUID_SUM, /* sum of device guids didn't match */
+ ZPOOL_STATUS_CORRUPT_POOL, /* pool metadata is corrupted */
+ ZPOOL_STATUS_CORRUPT_DATA, /* data errors in user (meta)data */
+ ZPOOL_STATUS_FAILING_DEV, /* device experiencing errors */
+ ZPOOL_STATUS_VERSION_NEWER, /* newer on-disk version */
+
+ /*
+ * The following are not faults per se, but still an error possibly
+ * requiring administrative attention. There is no corresponding
+ * message ID.
+ */
+ ZPOOL_STATUS_VERSION_OLDER, /* older on-disk version */
+ ZPOOL_STATUS_RESILVERING, /* device being resilvered */
+ ZPOOL_STATUS_OFFLINE_DEV, /* device online */
+
+ /*
+ * Finally, the following indicates a healthy pool.
+ */
+ ZPOOL_STATUS_OK
+} zpool_status_t;
+
+extern zpool_status_t zpool_get_status(zpool_handle_t *, char **);
+extern zpool_status_t zpool_import_status(nvlist_t *, char **);
+
+/*
+ * Statistics and configuration functions.
+ */
+extern nvlist_t *zpool_get_config(zpool_handle_t *, nvlist_t **);
+extern int zpool_refresh_stats(zpool_handle_t *, boolean_t *);
+extern int zpool_get_errlog(zpool_handle_t *, nvlist_t **);
+
+/*
+ * Import and export functions
+ */
+extern int zpool_export(zpool_handle_t *);
+extern int zpool_import(libzfs_handle_t *, nvlist_t *, const char *,
+ const char *);
+
+/*
+ * Search for pools to import
+ */
+extern nvlist_t *zpool_find_import(libzfs_handle_t *, int, char **);
+
+/*
+ * Miscellaneous pool functions
+ */
+extern char *zpool_vdev_name(libzfs_handle_t *, zpool_handle_t *, nvlist_t *);
+extern int zpool_upgrade(zpool_handle_t *);
+extern int zpool_get_history(zpool_handle_t *, nvlist_t **);
+extern void zpool_log_history(libzfs_handle_t *, int, char **, const char *,
+ boolean_t, boolean_t);
+extern void zpool_obj_to_path(zpool_handle_t *, uint64_t, uint64_t, char *,
+ size_t len);
+
+/*
+ * Basic handle manipulations. These functions do not create or destroy the
+ * underlying datasets, only the references to them.
+ */
+extern zfs_handle_t *zfs_open(libzfs_handle_t *, const char *, int);
+extern void zfs_close(zfs_handle_t *);
+extern zfs_type_t zfs_get_type(const zfs_handle_t *);
+extern const char *zfs_get_name(const zfs_handle_t *);
+
+/*
+ * Property management functions. Some functions are shared with the kernel,
+ * and are found in sys/fs/zfs.h.
+ */
+extern const char *zfs_prop_to_name(zfs_prop_t);
+extern int zfs_prop_set(zfs_handle_t *, const char *, const char *);
+extern int zfs_prop_get(zfs_handle_t *, zfs_prop_t, char *, size_t,
+ zfs_source_t *, char *, size_t, boolean_t);
+extern int zfs_prop_get_numeric(zfs_handle_t *, zfs_prop_t, uint64_t *,
+ zfs_source_t *, char *, size_t);
+extern uint64_t zfs_prop_get_int(zfs_handle_t *, zfs_prop_t);
+extern const char *zfs_prop_get_string(zfs_handle_t *, zfs_prop_t);
+extern int zfs_prop_inherit(zfs_handle_t *, const char *);
+extern const char *zfs_prop_values(zfs_prop_t);
+extern int zfs_prop_valid_for_type(zfs_prop_t, int);
+extern const char *zfs_prop_default_string(zfs_prop_t prop);
+extern uint64_t zfs_prop_default_numeric(zfs_prop_t);
+extern int zfs_prop_is_string(zfs_prop_t prop);
+extern const char *zfs_prop_column_name(zfs_prop_t);
+extern boolean_t zfs_prop_align_right(zfs_prop_t);
+extern void nicebool(int value, char *buf, size_t buflen);
+
+typedef struct zfs_proplist {
+ zfs_prop_t pl_prop;
+ char *pl_user_prop;
+ struct zfs_proplist *pl_next;
+ boolean_t pl_all;
+ size_t pl_width;
+ boolean_t pl_fixed;
+} zfs_proplist_t;
+
+typedef zfs_proplist_t zpool_proplist_t;
+
+extern int zfs_get_proplist(libzfs_handle_t *, char *, zfs_proplist_t **);
+extern int zpool_get_proplist(libzfs_handle_t *, char *, zpool_proplist_t **);
+extern int zfs_expand_proplist(zfs_handle_t *, zfs_proplist_t **);
+extern int zpool_expand_proplist(zpool_handle_t *, zpool_proplist_t **);
+extern void zfs_free_proplist(zfs_proplist_t *);
+extern nvlist_t *zfs_get_user_props(zfs_handle_t *);
+
+#define ZFS_MOUNTPOINT_NONE "none"
+#define ZFS_MOUNTPOINT_LEGACY "legacy"
+
+/*
+ * Functions for printing properties from zfs/zpool
+ */
+typedef struct libzfs_get_cbdata {
+ int cb_sources;
+ int cb_columns[4];
+ int cb_colwidths[5];
+ boolean_t cb_scripted;
+ boolean_t cb_literal;
+ boolean_t cb_first;
+ zfs_proplist_t *cb_proplist;
+} libzfs_get_cbdata_t;
+
+void libzfs_print_one_property(const char *, libzfs_get_cbdata_t *,
+ const char *, const char *, zfs_source_t, const char *);
+
+#define GET_COL_NAME 1
+#define GET_COL_PROPERTY 2
+#define GET_COL_VALUE 3
+#define GET_COL_SOURCE 4
+
+/*
+ * Iterator functions.
+ */
+typedef int (*zfs_iter_f)(zfs_handle_t *, void *);
+extern int zfs_iter_root(libzfs_handle_t *, zfs_iter_f, void *);
+extern int zfs_iter_children(zfs_handle_t *, zfs_iter_f, void *);
+extern int zfs_iter_dependents(zfs_handle_t *, boolean_t, zfs_iter_f, void *);
+extern int zfs_iter_filesystems(zfs_handle_t *, zfs_iter_f, void *);
+extern int zfs_iter_snapshots(zfs_handle_t *, zfs_iter_f, void *);
+
+/*
+ * Functions to create and destroy datasets.
+ */
+extern int zfs_create(libzfs_handle_t *, const char *, zfs_type_t,
+ nvlist_t *);
+extern int zfs_destroy(zfs_handle_t *);
+extern int zfs_destroy_snaps(zfs_handle_t *, char *);
+extern int zfs_clone(zfs_handle_t *, const char *, nvlist_t *);
+extern int zfs_snapshot(libzfs_handle_t *, const char *, boolean_t);
+extern int zfs_rollback(zfs_handle_t *, zfs_handle_t *, int);
+extern int zfs_rename(zfs_handle_t *, const char *);
+extern int zfs_send(zfs_handle_t *, const char *, int);
+extern int zfs_receive(libzfs_handle_t *, const char *, int, int, int,
+ boolean_t, int);
+extern int zfs_promote(zfs_handle_t *);
+
+/*
+ * Miscellaneous functions.
+ */
+extern const char *zfs_type_to_name(zfs_type_t);
+extern void zfs_refresh_properties(zfs_handle_t *);
+extern int zfs_name_valid(const char *, zfs_type_t);
+extern int zfs_disable(zfs_handle_t *);
+extern int zfs_enable(zfs_handle_t *);
+extern zfs_handle_t *zfs_path_to_zhandle(libzfs_handle_t *, char *, zfs_type_t);
+
+/*
+ * Mount support functions.
+ */
+extern boolean_t is_mounted(libzfs_handle_t *, const char *special, char **);
+extern boolean_t zfs_is_mounted(zfs_handle_t *, char **);
+extern int zfs_mount(zfs_handle_t *, const char *, int);
+extern int zfs_unmount(zfs_handle_t *, const char *, int);
+extern int zfs_unmountall(zfs_handle_t *, int);
+
+/*
+ * Share support functions.
+ */
+extern boolean_t zfs_is_shared(zfs_handle_t *);
+extern int zfs_share(zfs_handle_t *);
+extern int zfs_unshare(zfs_handle_t *);
+
+/*
+ * Protocol-specifc share support functions.
+ */
+extern boolean_t zfs_is_shared_nfs(zfs_handle_t *, char **);
+extern int zfs_share_nfs(zfs_handle_t *);
+extern int zfs_unshare_nfs(zfs_handle_t *, const char *);
+extern int zfs_unshareall_nfs(zfs_handle_t *);
+extern boolean_t zfs_is_shared_iscsi(zfs_handle_t *);
+extern int zfs_share_iscsi(zfs_handle_t *);
+extern int zfs_unshare_iscsi(zfs_handle_t *);
+
+/*
+ * FreeBSD-specific jail support function.
+ */
+extern int zfs_jail(zfs_handle_t *, int, int);
+
+/*
+ * When dealing with nvlists, verify() is extremely useful
+ */
+#ifndef verify
+#ifdef NDEBUG
+#define verify(EX) ((void)(EX))
+#else
+#define verify(EX) assert(EX)
+#endif
+#endif
+
+/*
+ * Utility function to convert a number to a human-readable form.
+ */
+extern void zfs_nicenum(uint64_t, char *, size_t);
+extern int zfs_nicestrtonum(libzfs_handle_t *, const char *, uint64_t *);
+
+/*
+ * Pool destroy special. Remove the device information without destroying
+ * the underlying dataset.
+ */
+extern int zfs_remove_link(zfs_handle_t *);
+
+/*
+ * Given a device or file, determine if it is part of a pool.
+ */
+extern int zpool_in_use(libzfs_handle_t *, int, pool_state_t *, char **,
+ boolean_t *);
+
+/*
+ * ftyp special. Read the label from a given device.
+ */
+extern int zpool_read_label(int, nvlist_t **);
+
+/*
+ * Create and remove zvol /dev links.
+ */
+extern int zpool_create_zvol_links(zpool_handle_t *);
+extern int zpool_remove_zvol_links(zpool_handle_t *);
+
+/*
+ * Enable and disable datasets within a pool by mounting/unmounting and
+ * sharing/unsharing them.
+ */
+extern int zpool_enable_datasets(zpool_handle_t *, const char *, int);
+extern int zpool_disable_datasets(zpool_handle_t *, boolean_t);
+
+#ifdef __FreeBSD__
+extern int zmount(const char *, const char *, int, char *, char *, int, char *,
+ int);
+#endif
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* _LIBZFS_H */
diff --git a/contrib/opensolaris/lib/libzfs/common/libzfs_changelist.c b/contrib/opensolaris/lib/libzfs/common/libzfs_changelist.c
new file mode 100644
index 0000000..d68c420
--- /dev/null
+++ b/contrib/opensolaris/lib/libzfs/common/libzfs_changelist.c
@@ -0,0 +1,606 @@
+/*
+ * 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"
+
+#include <libintl.h>
+#include <libuutil.h>
+#include <stddef.h>
+#include <stdlib.h>
+#include <string.h>
+#include <unistd.h>
+#include <zone.h>
+
+#include <libzfs.h>
+
+#include "libzfs_impl.h"
+
+/*
+ * Structure to keep track of dataset state. Before changing the 'sharenfs' or
+ * 'mountpoint' property, we record whether the filesystem was previously
+ * mounted/shared. This prior state dictates whether we remount/reshare the
+ * dataset after the property has been changed.
+ *
+ * The interface consists of the following sequence of functions:
+ *
+ * changelist_gather()
+ * changelist_prefix()
+ * < change property >
+ * changelist_postfix()
+ * changelist_free()
+ *
+ * Other interfaces:
+ *
+ * changelist_remove() - remove a node from a gathered list
+ * changelist_rename() - renames all datasets appropriately when doing a rename
+ * changelist_unshare() - unshares all the nodes in a given changelist
+ * changelist_haszonedchild() - check if there is any child exported to
+ * a local zone
+ */
+typedef struct prop_changenode {
+ zfs_handle_t *cn_handle;
+ int cn_shared;
+ int cn_mounted;
+ int cn_zoned;
+ uu_list_node_t cn_listnode;
+} prop_changenode_t;
+
+struct prop_changelist {
+ zfs_prop_t cl_prop;
+ zfs_prop_t cl_realprop;
+ uu_list_pool_t *cl_pool;
+ uu_list_t *cl_list;
+ boolean_t cl_waslegacy;
+ boolean_t cl_allchildren;
+ boolean_t cl_alldependents;
+ int cl_flags;
+ boolean_t cl_haszonedchild;
+ boolean_t cl_sorted;
+};
+
+/*
+ * If the property is 'mountpoint', go through and unmount filesystems as
+ * necessary. We don't do the same for 'sharenfs', because we can just re-share
+ * with different options without interrupting service.
+ */
+int
+changelist_prefix(prop_changelist_t *clp)
+{
+ prop_changenode_t *cn;
+ int ret = 0;
+
+ if (clp->cl_prop != ZFS_PROP_MOUNTPOINT)
+ return (0);
+
+ for (cn = uu_list_first(clp->cl_list); cn != NULL;
+ cn = uu_list_next(clp->cl_list, cn)) {
+ /*
+ * If we are in the global zone, but this dataset is exported
+ * to a local zone, do nothing.
+ */
+ if (getzoneid() == GLOBAL_ZONEID && cn->cn_zoned)
+ continue;
+
+ if (ZFS_IS_VOLUME(cn->cn_handle)) {
+ switch (clp->cl_realprop) {
+ case ZFS_PROP_NAME:
+ /*
+ * If this was a rename, unshare the zvol, and
+ * remove the /dev/zvol links.
+ */
+ (void) zfs_unshare_iscsi(cn->cn_handle);
+
+ if (zvol_remove_link(cn->cn_handle->zfs_hdl,
+ cn->cn_handle->zfs_name) != 0)
+ ret = -1;
+ break;
+
+ case ZFS_PROP_VOLSIZE:
+ /*
+ * If this was a change to the volume size, we
+ * need to unshare and reshare the volume.
+ */
+ (void) zfs_unshare_iscsi(cn->cn_handle);
+ break;
+ }
+ } else if (zfs_unmount(cn->cn_handle, NULL, clp->cl_flags) != 0)
+ ret = -1;
+ }
+
+ return (ret);
+}
+
+/*
+ * If the property is 'mountpoint' or 'sharenfs', go through and remount and/or
+ * reshare the filesystems as necessary. In changelist_gather() we recorded
+ * whether the filesystem was previously shared or mounted. The action we take
+ * depends on the previous state, and whether the value was previously 'legacy'.
+ * For non-legacy properties, we only remount/reshare the filesystem if it was
+ * previously mounted/shared. Otherwise, we always remount/reshare the
+ * filesystem.
+ */
+int
+changelist_postfix(prop_changelist_t *clp)
+{
+ prop_changenode_t *cn;
+ char shareopts[ZFS_MAXPROPLEN];
+ int ret = 0;
+
+ /*
+ * If we're changing the mountpoint, attempt to destroy the underlying
+ * mountpoint. All other datasets will have inherited from this dataset
+ * (in which case their mountpoints exist in the filesystem in the new
+ * location), or have explicit mountpoints set (in which case they won't
+ * be in the changelist).
+ */
+ if ((cn = uu_list_last(clp->cl_list)) == NULL)
+ return (0);
+
+ if (clp->cl_prop == ZFS_PROP_MOUNTPOINT)
+ remove_mountpoint(cn->cn_handle);
+
+ /*
+ * We walk the datasets in reverse, because we want to mount any parent
+ * datasets before mounting the children.
+ */
+ for (cn = uu_list_last(clp->cl_list); cn != NULL;
+ cn = uu_list_prev(clp->cl_list, cn)) {
+ /*
+ * If we are in the global zone, but this dataset is exported
+ * to a local zone, do nothing.
+ */
+ if (getzoneid() == GLOBAL_ZONEID && cn->cn_zoned)
+ continue;
+
+ zfs_refresh_properties(cn->cn_handle);
+
+ if (ZFS_IS_VOLUME(cn->cn_handle)) {
+ /*
+ * If we're doing a rename, recreate the /dev/zvol
+ * links.
+ */
+ if (clp->cl_realprop == ZFS_PROP_NAME &&
+ zvol_create_link(cn->cn_handle->zfs_hdl,
+ cn->cn_handle->zfs_name) != 0) {
+ ret = -1;
+ } else if (cn->cn_shared ||
+ clp->cl_prop == ZFS_PROP_SHAREISCSI) {
+ if (zfs_prop_get(cn->cn_handle,
+ ZFS_PROP_SHAREISCSI, shareopts,
+ sizeof (shareopts), NULL, NULL, 0,
+ B_FALSE) == 0 &&
+ strcmp(shareopts, "off") == 0) {
+ ret = zfs_unshare_iscsi(cn->cn_handle);
+ } else {
+ ret = zfs_share_iscsi(cn->cn_handle);
+ }
+ }
+
+ continue;
+ }
+
+ if ((clp->cl_waslegacy || cn->cn_mounted) &&
+ !zfs_is_mounted(cn->cn_handle, NULL) &&
+ zfs_mount(cn->cn_handle, NULL, 0) != 0)
+ ret = -1;
+
+ /*
+ * We always re-share even if the filesystem is currently
+ * shared, so that we can adopt any new options.
+ */
+ if (cn->cn_shared ||
+ (clp->cl_prop == ZFS_PROP_SHARENFS && clp->cl_waslegacy)) {
+ if (zfs_prop_get(cn->cn_handle, ZFS_PROP_SHARENFS,
+ shareopts, sizeof (shareopts), NULL, NULL, 0,
+ B_FALSE) == 0 && strcmp(shareopts, "off") == 0) {
+ ret = zfs_unshare_nfs(cn->cn_handle, NULL);
+ } else {
+ ret = zfs_share_nfs(cn->cn_handle);
+ }
+ }
+ }
+
+ return (ret);
+}
+
+/*
+ * Is this "dataset" a child of "parent"?
+ */
+static boolean_t
+isa_child_of(const char *dataset, const char *parent)
+{
+ int len;
+
+ len = strlen(parent);
+
+ if (strncmp(dataset, parent, len) == 0 &&
+ (dataset[len] == '@' || dataset[len] == '/' ||
+ dataset[len] == '\0'))
+ return (B_TRUE);
+ else
+ return (B_FALSE);
+
+}
+
+/*
+ * If we rename a filesystem, child filesystem handles are no longer valid
+ * since we identify each dataset by its name in the ZFS namespace. As a
+ * result, we have to go through and fix up all the names appropriately. We
+ * could do this automatically if libzfs kept track of all open handles, but
+ * this is a lot less work.
+ */
+void
+changelist_rename(prop_changelist_t *clp, const char *src, const char *dst)
+{
+ prop_changenode_t *cn;
+ char newname[ZFS_MAXNAMELEN];
+
+ for (cn = uu_list_first(clp->cl_list); cn != NULL;
+ cn = uu_list_next(clp->cl_list, cn)) {
+ /*
+ * Do not rename a clone that's not in the source hierarchy.
+ */
+ if (!isa_child_of(cn->cn_handle->zfs_name, src))
+ continue;
+
+ /*
+ * Destroy the previous mountpoint if needed.
+ */
+ remove_mountpoint(cn->cn_handle);
+
+ (void) strlcpy(newname, dst, sizeof (newname));
+ (void) strcat(newname, cn->cn_handle->zfs_name + strlen(src));
+
+ (void) strlcpy(cn->cn_handle->zfs_name, newname,
+ sizeof (cn->cn_handle->zfs_name));
+ }
+}
+
+/*
+ * Given a gathered changelist for the 'sharenfs' property, unshare all the
+ * datasets in the list.
+ */
+int
+changelist_unshare(prop_changelist_t *clp)
+{
+ prop_changenode_t *cn;
+ int ret = 0;
+
+ if (clp->cl_prop != ZFS_PROP_SHARENFS)
+ return (0);
+
+ for (cn = uu_list_first(clp->cl_list); cn != NULL;
+ cn = uu_list_next(clp->cl_list, cn)) {
+ if (zfs_unshare_nfs(cn->cn_handle, NULL) != 0)
+ ret = -1;
+ }
+
+ return (ret);
+}
+
+/*
+ * Check if there is any child exported to a local zone in a given changelist.
+ * This information has already been recorded while gathering the changelist
+ * via changelist_gather().
+ */
+int
+changelist_haszonedchild(prop_changelist_t *clp)
+{
+ return (clp->cl_haszonedchild);
+}
+
+/*
+ * Remove a node from a gathered list.
+ */
+void
+changelist_remove(zfs_handle_t *zhp, prop_changelist_t *clp)
+{
+ prop_changenode_t *cn;
+
+ for (cn = uu_list_first(clp->cl_list); cn != NULL;
+ cn = uu_list_next(clp->cl_list, cn)) {
+
+ if (strcmp(cn->cn_handle->zfs_name, zhp->zfs_name) == 0) {
+ uu_list_remove(clp->cl_list, cn);
+ zfs_close(cn->cn_handle);
+ free(cn);
+ return;
+ }
+ }
+}
+
+/*
+ * Release any memory associated with a changelist.
+ */
+void
+changelist_free(prop_changelist_t *clp)
+{
+ prop_changenode_t *cn;
+ uu_list_walk_t *walk;
+
+ if (clp->cl_list) {
+ verify((walk = uu_list_walk_start(clp->cl_list,
+ UU_WALK_ROBUST)) != NULL);
+
+ while ((cn = uu_list_walk_next(walk)) != NULL) {
+
+ uu_list_remove(clp->cl_list, cn);
+
+ zfs_close(cn->cn_handle);
+ free(cn);
+ }
+
+ uu_list_walk_end(walk);
+
+ uu_list_destroy(clp->cl_list);
+ }
+ if (clp->cl_pool)
+ uu_list_pool_destroy(clp->cl_pool);
+
+ free(clp);
+}
+
+static int
+change_one(zfs_handle_t *zhp, void *data)
+{
+ prop_changelist_t *clp = data;
+ char property[ZFS_MAXPROPLEN];
+ char where[64];
+ prop_changenode_t *cn;
+ zfs_source_t sourcetype;
+
+ /*
+ * We only want to unmount/unshare those filesystems that may inherit
+ * from the target filesystem. If we find any filesystem with a
+ * locally set mountpoint, we ignore any children since changing the
+ * property will not affect them. If this is a rename, we iterate
+ * over all children regardless, since we need them unmounted in
+ * order to do the rename. Also, if this is a volume and we're doing
+ * a rename, then always add it to the changelist.
+ */
+
+ if (!(ZFS_IS_VOLUME(zhp) && clp->cl_realprop == ZFS_PROP_NAME) &&
+ zfs_prop_get(zhp, clp->cl_prop, property,
+ sizeof (property), &sourcetype, where, sizeof (where),
+ B_FALSE) != 0) {
+ zfs_close(zhp);
+ return (0);
+ }
+
+ if (clp->cl_alldependents || clp->cl_allchildren ||
+ sourcetype == ZFS_SRC_DEFAULT || sourcetype == ZFS_SRC_INHERITED) {
+ if ((cn = zfs_alloc(zfs_get_handle(zhp),
+ sizeof (prop_changenode_t))) == NULL) {
+ zfs_close(zhp);
+ return (-1);
+ }
+
+ cn->cn_handle = zhp;
+ cn->cn_mounted = zfs_is_mounted(zhp, NULL);
+ cn->cn_shared = zfs_is_shared(zhp);
+ cn->cn_zoned = zfs_prop_get_int(zhp, ZFS_PROP_ZONED);
+
+ /* Indicate if any child is exported to a local zone. */
+ if (getzoneid() == GLOBAL_ZONEID && cn->cn_zoned)
+ clp->cl_haszonedchild = B_TRUE;
+
+ uu_list_node_init(cn, &cn->cn_listnode, clp->cl_pool);
+
+ if (clp->cl_sorted) {
+ uu_list_index_t idx;
+
+ (void) uu_list_find(clp->cl_list, cn, NULL,
+ &idx);
+ uu_list_insert(clp->cl_list, cn, idx);
+ } else {
+ ASSERT(!clp->cl_alldependents);
+ verify(uu_list_insert_before(clp->cl_list,
+ uu_list_first(clp->cl_list), cn) == 0);
+ }
+
+ if (!clp->cl_alldependents)
+ return (zfs_iter_children(zhp, change_one, data));
+ } else {
+ zfs_close(zhp);
+ }
+
+ return (0);
+}
+
+/*ARGSUSED*/
+static int
+compare_mountpoints(const void *a, const void *b, void *unused)
+{
+ const prop_changenode_t *ca = a;
+ const prop_changenode_t *cb = b;
+
+ char mounta[MAXPATHLEN];
+ char mountb[MAXPATHLEN];
+
+ boolean_t hasmounta, hasmountb;
+
+ /*
+ * When unsharing or unmounting filesystems, we need to do it in
+ * mountpoint order. This allows the user to have a mountpoint
+ * hierarchy that is different from the dataset hierarchy, and still
+ * allow it to be changed. However, if either dataset doesn't have a
+ * mountpoint (because it is a volume or a snapshot), we place it at the
+ * end of the list, because it doesn't affect our change at all.
+ */
+ hasmounta = (zfs_prop_get(ca->cn_handle, ZFS_PROP_MOUNTPOINT, mounta,
+ sizeof (mounta), NULL, NULL, 0, B_FALSE) == 0);
+ hasmountb = (zfs_prop_get(cb->cn_handle, ZFS_PROP_MOUNTPOINT, mountb,
+ sizeof (mountb), NULL, NULL, 0, B_FALSE) == 0);
+
+ if (!hasmounta && hasmountb)
+ return (-1);
+ else if (hasmounta && !hasmountb)
+ return (1);
+ else if (!hasmounta && !hasmountb)
+ return (0);
+ else
+ return (strcmp(mountb, mounta));
+}
+
+/*
+ * Given a ZFS handle and a property, construct a complete list of datasets
+ * that need to be modified as part of this process. For anything but the
+ * 'mountpoint' and 'sharenfs' properties, this just returns an empty list.
+ * Otherwise, we iterate over all children and look for any datasets that
+ * inherit the property. For each such dataset, we add it to the list and
+ * mark whether it was shared beforehand.
+ */
+prop_changelist_t *
+changelist_gather(zfs_handle_t *zhp, zfs_prop_t prop, int flags)
+{
+ prop_changelist_t *clp;
+ prop_changenode_t *cn;
+ zfs_handle_t *temp;
+ char property[ZFS_MAXPROPLEN];
+ uu_compare_fn_t *compare = NULL;
+
+ if ((clp = zfs_alloc(zhp->zfs_hdl, sizeof (prop_changelist_t))) == NULL)
+ return (NULL);
+
+ /*
+ * For mountpoint-related tasks, we want to sort everything by
+ * mountpoint, so that we mount and unmount them in the appropriate
+ * order, regardless of their position in the hierarchy.
+ */
+ if (prop == ZFS_PROP_NAME || prop == ZFS_PROP_ZONED ||
+ prop == ZFS_PROP_MOUNTPOINT || prop == ZFS_PROP_SHARENFS) {
+ compare = compare_mountpoints;
+ clp->cl_sorted = B_TRUE;
+ }
+
+ clp->cl_pool = uu_list_pool_create("changelist_pool",
+ sizeof (prop_changenode_t),
+ offsetof(prop_changenode_t, cn_listnode),
+ compare, 0);
+ if (clp->cl_pool == NULL) {
+ assert(uu_error() == UU_ERROR_NO_MEMORY);
+ (void) zfs_error(zhp->zfs_hdl, EZFS_NOMEM, "internal error");
+ changelist_free(clp);
+ return (NULL);
+ }
+
+ clp->cl_list = uu_list_create(clp->cl_pool, NULL,
+ clp->cl_sorted ? UU_LIST_SORTED : 0);
+ clp->cl_flags = flags;
+
+ if (clp->cl_list == NULL) {
+ assert(uu_error() == UU_ERROR_NO_MEMORY);
+ (void) zfs_error(zhp->zfs_hdl, EZFS_NOMEM, "internal error");
+ changelist_free(clp);
+ return (NULL);
+ }
+
+ /*
+ * If this is a rename or the 'zoned' property, we pretend we're
+ * changing the mountpoint and flag it so we can catch all children in
+ * change_one().
+ *
+ * Flag cl_alldependents to catch all children plus the dependents
+ * (clones) that are not in the hierarchy.
+ */
+ if (prop == ZFS_PROP_NAME) {
+ clp->cl_prop = ZFS_PROP_MOUNTPOINT;
+ clp->cl_alldependents = B_TRUE;
+ } else if (prop == ZFS_PROP_ZONED) {
+ clp->cl_prop = ZFS_PROP_MOUNTPOINT;
+ clp->cl_allchildren = B_TRUE;
+ } else if (prop == ZFS_PROP_CANMOUNT) {
+ clp->cl_prop = ZFS_PROP_MOUNTPOINT;
+ } else if (prop == ZFS_PROP_VOLSIZE) {
+ clp->cl_prop = ZFS_PROP_MOUNTPOINT;
+ } else {
+ clp->cl_prop = prop;
+ }
+ clp->cl_realprop = prop;
+
+ if (clp->cl_prop != ZFS_PROP_MOUNTPOINT &&
+ clp->cl_prop != ZFS_PROP_SHARENFS &&
+ clp->cl_prop != ZFS_PROP_SHAREISCSI)
+ return (clp);
+
+ if (clp->cl_alldependents) {
+ if (zfs_iter_dependents(zhp, B_TRUE, change_one, clp) != 0) {
+ changelist_free(clp);
+ return (NULL);
+ }
+ } else if (zfs_iter_children(zhp, change_one, clp) != 0) {
+ changelist_free(clp);
+ return (NULL);
+ }
+
+ /*
+ * We have to re-open ourselves because we auto-close all the handles
+ * and can't tell the difference.
+ */
+ if ((temp = zfs_open(zhp->zfs_hdl, zfs_get_name(zhp),
+ ZFS_TYPE_ANY)) == NULL) {
+ changelist_free(clp);
+ return (NULL);
+ }
+
+ /*
+ * Always add ourself to the list. We add ourselves to the end so that
+ * we're the last to be unmounted.
+ */
+ if ((cn = zfs_alloc(zhp->zfs_hdl,
+ sizeof (prop_changenode_t))) == NULL) {
+ zfs_close(temp);
+ changelist_free(clp);
+ return (NULL);
+ }
+
+ cn->cn_handle = temp;
+ cn->cn_mounted = zfs_is_mounted(temp, NULL);
+ cn->cn_shared = zfs_is_shared(temp);
+ cn->cn_zoned = zfs_prop_get_int(zhp, ZFS_PROP_ZONED);
+
+ uu_list_node_init(cn, &cn->cn_listnode, clp->cl_pool);
+ if (clp->cl_sorted) {
+ uu_list_index_t idx;
+ (void) uu_list_find(clp->cl_list, cn, NULL, &idx);
+ uu_list_insert(clp->cl_list, cn, idx);
+ } else {
+ verify(uu_list_insert_after(clp->cl_list,
+ uu_list_last(clp->cl_list), cn) == 0);
+ }
+
+ /*
+ * If the property was previously 'legacy' or 'none', record this fact,
+ * as the behavior of changelist_postfix() will be different.
+ */
+ if (zfs_prop_get(zhp, prop, property, sizeof (property),
+ NULL, NULL, 0, B_FALSE) == 0 &&
+ (strcmp(property, "legacy") == 0 || strcmp(property, "none") == 0 ||
+ strcmp(property, "off") == 0))
+ clp->cl_waslegacy = B_TRUE;
+
+ return (clp);
+}
diff --git a/contrib/opensolaris/lib/libzfs/common/libzfs_config.c b/contrib/opensolaris/lib/libzfs/common/libzfs_config.c
new file mode 100644
index 0000000..45e2920
--- /dev/null
+++ b/contrib/opensolaris/lib/libzfs/common/libzfs_config.c
@@ -0,0 +1,374 @@
+/*
+ * 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"
+
+/*
+ * The pool configuration repository is stored in /etc/zfs/zpool.cache as a
+ * single packed nvlist. While it would be nice to just read in this
+ * file from userland, this wouldn't work from a local zone. So we have to have
+ * a zpool ioctl to return the complete configuration for all pools. In the
+ * global zone, this will be identical to reading the file and unpacking it in
+ * userland.
+ */
+
+#include <errno.h>
+#include <sys/stat.h>
+#include <fcntl.h>
+#include <stddef.h>
+#include <string.h>
+#include <unistd.h>
+#include <libintl.h>
+#include <libuutil.h>
+
+#include "libzfs_impl.h"
+
+typedef struct config_node {
+ char *cn_name;
+ nvlist_t *cn_config;
+ uu_avl_node_t cn_avl;
+} config_node_t;
+
+/* ARGSUSED */
+static int
+config_node_compare(const void *a, const void *b, void *unused)
+{
+ int ret;
+
+ const config_node_t *ca = (config_node_t *)a;
+ const config_node_t *cb = (config_node_t *)b;
+
+ ret = strcmp(ca->cn_name, cb->cn_name);
+
+ if (ret < 0)
+ return (-1);
+ else if (ret > 0)
+ return (1);
+ else
+ return (0);
+}
+
+void
+namespace_clear(libzfs_handle_t *hdl)
+{
+ if (hdl->libzfs_ns_avl) {
+ uu_avl_walk_t *walk;
+ config_node_t *cn;
+
+ if ((walk = uu_avl_walk_start(hdl->libzfs_ns_avl,
+ UU_WALK_ROBUST)) == NULL)
+ return;
+
+ while ((cn = uu_avl_walk_next(walk)) != NULL) {
+ uu_avl_remove(hdl->libzfs_ns_avl, cn);
+ nvlist_free(cn->cn_config);
+ free(cn->cn_name);
+ free(cn);
+ }
+
+ uu_avl_walk_end(walk);
+
+ uu_avl_destroy(hdl->libzfs_ns_avl);
+ hdl->libzfs_ns_avl = NULL;
+ }
+
+ if (hdl->libzfs_ns_avlpool) {
+ uu_avl_pool_destroy(hdl->libzfs_ns_avlpool);
+ hdl->libzfs_ns_avlpool = NULL;
+ }
+}
+
+/*
+ * Loads the pool namespace, or re-loads it if the cache has changed.
+ */
+static int
+namespace_reload(libzfs_handle_t *hdl)
+{
+ nvlist_t *config;
+ config_node_t *cn;
+ nvpair_t *elem;
+ zfs_cmd_t zc = { 0 };
+ uu_avl_walk_t *walk;
+
+ if (hdl->libzfs_ns_gen == 0) {
+ /*
+ * This is the first time we've accessed the configuration
+ * cache. Initialize the AVL tree and then fall through to the
+ * common code.
+ */
+ if ((hdl->libzfs_ns_avlpool = uu_avl_pool_create("config_pool",
+ sizeof (config_node_t),
+ offsetof(config_node_t, cn_avl),
+ config_node_compare, UU_DEFAULT)) == NULL)
+ return (no_memory(hdl));
+
+ if ((hdl->libzfs_ns_avl = uu_avl_create(hdl->libzfs_ns_avlpool,
+ NULL, UU_DEFAULT)) == NULL)
+ return (no_memory(hdl));
+ }
+
+ if (zcmd_alloc_dst_nvlist(hdl, &zc, 0) != 0)
+ return (-1);
+
+ for (;;) {
+ zc.zc_cookie = hdl->libzfs_ns_gen;
+ if (ioctl(hdl->libzfs_fd, ZFS_IOC_POOL_CONFIGS, &zc) != 0) {
+ switch (errno) {
+ case EEXIST:
+ /*
+ * The namespace hasn't changed.
+ */
+ zcmd_free_nvlists(&zc);
+ return (0);
+
+ case ENOMEM:
+ if (zcmd_expand_dst_nvlist(hdl, &zc) != 0) {
+ zcmd_free_nvlists(&zc);
+ return (-1);
+ }
+ break;
+
+ default:
+ zcmd_free_nvlists(&zc);
+ return (zfs_standard_error(hdl, errno,
+ dgettext(TEXT_DOMAIN, "failed to read "
+ "pool configuration")));
+ }
+ } else {
+ hdl->libzfs_ns_gen = zc.zc_cookie;
+ break;
+ }
+ }
+
+ if (zcmd_read_dst_nvlist(hdl, &zc, &config) != 0) {
+ zcmd_free_nvlists(&zc);
+ return (-1);
+ }
+
+ zcmd_free_nvlists(&zc);
+
+ /*
+ * Clear out any existing configuration information.
+ */
+ if ((walk = uu_avl_walk_start(hdl->libzfs_ns_avl,
+ UU_WALK_ROBUST)) == NULL) {
+ nvlist_free(config);
+ return (no_memory(hdl));
+ }
+
+ while ((cn = uu_avl_walk_next(walk)) != NULL) {
+ uu_avl_remove(hdl->libzfs_ns_avl, cn);
+ nvlist_free(cn->cn_config);
+ free(cn->cn_name);
+ free(cn);
+ }
+
+ uu_avl_walk_end(walk);
+
+ elem = NULL;
+ while ((elem = nvlist_next_nvpair(config, elem)) != NULL) {
+ nvlist_t *child;
+ uu_avl_index_t where;
+
+ if ((cn = zfs_alloc(hdl, sizeof (config_node_t))) == NULL) {
+ nvlist_free(config);
+ return (-1);
+ }
+
+ if ((cn->cn_name = zfs_strdup(hdl,
+ nvpair_name(elem))) == NULL) {
+ free(cn);
+ nvlist_free(config);
+ return (-1);
+ }
+
+ verify(nvpair_value_nvlist(elem, &child) == 0);
+ if (nvlist_dup(child, &cn->cn_config, 0) != 0) {
+ free(cn->cn_name);
+ free(cn);
+ nvlist_free(config);
+ return (no_memory(hdl));
+ }
+ verify(uu_avl_find(hdl->libzfs_ns_avl, cn, NULL, &where)
+ == NULL);
+
+ uu_avl_insert(hdl->libzfs_ns_avl, cn, where);
+ }
+
+ nvlist_free(config);
+ return (0);
+}
+
+/*
+ * Retrive the configuration for the given pool. The configuration is a nvlist
+ * describing the vdevs, as well as the statistics associated with each one.
+ */
+nvlist_t *
+zpool_get_config(zpool_handle_t *zhp, nvlist_t **oldconfig)
+{
+ if (oldconfig)
+ *oldconfig = zhp->zpool_old_config;
+ return (zhp->zpool_config);
+}
+
+/*
+ * Refresh the vdev statistics associated with the given pool. This is used in
+ * iostat to show configuration changes and determine the delta from the last
+ * time the function was called. This function can fail, in case the pool has
+ * been destroyed.
+ */
+int
+zpool_refresh_stats(zpool_handle_t *zhp, boolean_t *missing)
+{
+ zfs_cmd_t zc = { 0 };
+ int error;
+ nvlist_t *config;
+ libzfs_handle_t *hdl = zhp->zpool_hdl;
+
+ *missing = B_FALSE;
+ (void) strcpy(zc.zc_name, zhp->zpool_name);
+
+ if (zhp->zpool_config_size == 0)
+ zhp->zpool_config_size = 1 << 16;
+
+ if (zcmd_alloc_dst_nvlist(hdl, &zc, zhp->zpool_config_size) != 0)
+ return (-1);
+
+ for (;;) {
+ if (ioctl(zhp->zpool_hdl->libzfs_fd, ZFS_IOC_POOL_STATS,
+ &zc) == 0) {
+ /*
+ * The real error is returned in the zc_cookie field.
+ */
+ error = zc.zc_cookie;
+ break;
+ }
+
+ if (errno == ENOMEM) {
+ if (zcmd_expand_dst_nvlist(hdl, &zc) != 0) {
+ zcmd_free_nvlists(&zc);
+ return (-1);
+ }
+ } else {
+ zcmd_free_nvlists(&zc);
+ if (errno == ENOENT || errno == EINVAL)
+ *missing = B_TRUE;
+ zhp->zpool_state = POOL_STATE_UNAVAIL;
+ return (0);
+ }
+ }
+
+ if (zcmd_read_dst_nvlist(hdl, &zc, &config) != 0) {
+ zcmd_free_nvlists(&zc);
+ return (-1);
+ }
+
+ zcmd_free_nvlists(&zc);
+
+ zhp->zpool_config_size = zc.zc_nvlist_dst_size;
+
+ if (zhp->zpool_config != NULL) {
+ uint64_t oldtxg, newtxg;
+
+ verify(nvlist_lookup_uint64(zhp->zpool_config,
+ ZPOOL_CONFIG_POOL_TXG, &oldtxg) == 0);
+ verify(nvlist_lookup_uint64(config,
+ ZPOOL_CONFIG_POOL_TXG, &newtxg) == 0);
+
+ if (zhp->zpool_old_config != NULL)
+ nvlist_free(zhp->zpool_old_config);
+
+ if (oldtxg != newtxg) {
+ nvlist_free(zhp->zpool_config);
+ zhp->zpool_old_config = NULL;
+ } else {
+ zhp->zpool_old_config = zhp->zpool_config;
+ }
+ }
+
+ zhp->zpool_config = config;
+ if (error)
+ zhp->zpool_state = POOL_STATE_UNAVAIL;
+ else
+ zhp->zpool_state = POOL_STATE_ACTIVE;
+
+ return (0);
+}
+
+/*
+ * Iterate over all pools in the system.
+ */
+int
+zpool_iter(libzfs_handle_t *hdl, zpool_iter_f func, void *data)
+{
+ config_node_t *cn;
+ zpool_handle_t *zhp;
+ int ret;
+
+ if (namespace_reload(hdl) != 0)
+ return (-1);
+
+ for (cn = uu_avl_first(hdl->libzfs_ns_avl); cn != NULL;
+ cn = uu_avl_next(hdl->libzfs_ns_avl, cn)) {
+
+ if (zpool_open_silent(hdl, cn->cn_name, &zhp) != 0)
+ return (-1);
+
+ if (zhp == NULL)
+ continue;
+
+ if ((ret = func(zhp, data)) != 0)
+ return (ret);
+ }
+
+ return (0);
+}
+
+/*
+ * Iterate over root datasets, calling the given function for each. The zfs
+ * handle passed each time must be explicitly closed by the callback.
+ */
+int
+zfs_iter_root(libzfs_handle_t *hdl, zfs_iter_f func, void *data)
+{
+ config_node_t *cn;
+ zfs_handle_t *zhp;
+ int ret;
+
+ if (namespace_reload(hdl) != 0)
+ return (-1);
+
+ for (cn = uu_avl_first(hdl->libzfs_ns_avl); cn != NULL;
+ cn = uu_avl_next(hdl->libzfs_ns_avl, cn)) {
+
+ if ((zhp = make_dataset_handle(hdl, cn->cn_name)) == NULL)
+ continue;
+
+ if ((ret = func(zhp, data)) != 0)
+ return (ret);
+ }
+
+ return (0);
+}
diff --git a/contrib/opensolaris/lib/libzfs/common/libzfs_dataset.c b/contrib/opensolaris/lib/libzfs/common/libzfs_dataset.c
new file mode 100644
index 0000000..5bc6c7b
--- /dev/null
+++ b/contrib/opensolaris/lib/libzfs/common/libzfs_dataset.c
@@ -0,0 +1,3753 @@
+/*
+ * 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"
+
+#include <assert.h>
+#include <ctype.h>
+#include <errno.h>
+#include <libintl.h>
+#include <math.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <strings.h>
+#include <unistd.h>
+#include <zone.h>
+#include <fcntl.h>
+#include <sys/mntent.h>
+#include <sys/mnttab.h>
+#include <sys/mount.h>
+
+#include <sys/spa.h>
+#include <sys/zio.h>
+#include <sys/zap.h>
+#include <libzfs.h>
+
+#include "zfs_namecheck.h"
+#include "zfs_prop.h"
+#include "libzfs_impl.h"
+
+/*
+ * Given a single type (not a mask of types), return the type in a human
+ * readable form.
+ */
+const char *
+zfs_type_to_name(zfs_type_t type)
+{
+ switch (type) {
+ case ZFS_TYPE_FILESYSTEM:
+ return (dgettext(TEXT_DOMAIN, "filesystem"));
+ case ZFS_TYPE_SNAPSHOT:
+ return (dgettext(TEXT_DOMAIN, "snapshot"));
+ case ZFS_TYPE_VOLUME:
+ return (dgettext(TEXT_DOMAIN, "volume"));
+ }
+
+ return (NULL);
+}
+
+/*
+ * Given a path and mask of ZFS types, return a string describing this dataset.
+ * This is used when we fail to open a dataset and we cannot get an exact type.
+ * We guess what the type would have been based on the path and the mask of
+ * acceptable types.
+ */
+static const char *
+path_to_str(const char *path, int types)
+{
+ /*
+ * When given a single type, always report the exact type.
+ */
+ if (types == ZFS_TYPE_SNAPSHOT)
+ return (dgettext(TEXT_DOMAIN, "snapshot"));
+ if (types == ZFS_TYPE_FILESYSTEM)
+ return (dgettext(TEXT_DOMAIN, "filesystem"));
+ if (types == ZFS_TYPE_VOLUME)
+ return (dgettext(TEXT_DOMAIN, "volume"));
+
+ /*
+ * The user is requesting more than one type of dataset. If this is the
+ * case, consult the path itself. If we're looking for a snapshot, and
+ * a '@' is found, then report it as "snapshot". Otherwise, remove the
+ * snapshot attribute and try again.
+ */
+ if (types & ZFS_TYPE_SNAPSHOT) {
+ if (strchr(path, '@') != NULL)
+ return (dgettext(TEXT_DOMAIN, "snapshot"));
+ return (path_to_str(path, types & ~ZFS_TYPE_SNAPSHOT));
+ }
+
+
+ /*
+ * The user has requested either filesystems or volumes.
+ * We have no way of knowing a priori what type this would be, so always
+ * report it as "filesystem" or "volume", our two primitive types.
+ */
+ if (types & ZFS_TYPE_FILESYSTEM)
+ return (dgettext(TEXT_DOMAIN, "filesystem"));
+
+ assert(types & ZFS_TYPE_VOLUME);
+ return (dgettext(TEXT_DOMAIN, "volume"));
+}
+
+/*
+ * Validate a ZFS path. This is used even before trying to open the dataset, to
+ * provide a more meaningful error message. We place a more useful message in
+ * 'buf' detailing exactly why the name was not valid.
+ */
+static int
+zfs_validate_name(libzfs_handle_t *hdl, const char *path, int type)
+{
+ namecheck_err_t why;
+ char what;
+
+ if (dataset_namecheck(path, &why, &what) != 0) {
+ if (hdl != NULL) {
+ switch (why) {
+ case NAME_ERR_TOOLONG:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "name is too long"));
+ break;
+
+ case NAME_ERR_LEADING_SLASH:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "leading slash in name"));
+ break;
+
+ case NAME_ERR_EMPTY_COMPONENT:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "empty component in name"));
+ break;
+
+ case NAME_ERR_TRAILING_SLASH:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "trailing slash in name"));
+ break;
+
+ case NAME_ERR_INVALCHAR:
+ zfs_error_aux(hdl,
+ dgettext(TEXT_DOMAIN, "invalid character "
+ "'%c' in name"), what);
+ break;
+
+ case NAME_ERR_MULTIPLE_AT:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "multiple '@' delimiters in name"));
+ break;
+
+ case NAME_ERR_NOLETTER:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "pool doesn't begin with a letter"));
+ break;
+
+ case NAME_ERR_RESERVED:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "name is reserved"));
+ break;
+
+ case NAME_ERR_DISKLIKE:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "reserved disk name"));
+ break;
+ }
+ }
+
+ return (0);
+ }
+
+ if (!(type & ZFS_TYPE_SNAPSHOT) && strchr(path, '@') != NULL) {
+ if (hdl != NULL)
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "snapshot delimiter '@' in filesystem name"));
+ return (0);
+ }
+
+ if (type == ZFS_TYPE_SNAPSHOT && strchr(path, '@') == NULL) {
+ if (hdl != NULL)
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "missing '@' delimiter in snapshot name"));
+ return (0);
+ }
+
+ return (-1);
+}
+
+int
+zfs_name_valid(const char *name, zfs_type_t type)
+{
+ return (zfs_validate_name(NULL, name, type));
+}
+
+/*
+ * This function takes the raw DSL properties, and filters out the user-defined
+ * properties into a separate nvlist.
+ */
+static int
+process_user_props(zfs_handle_t *zhp)
+{
+ libzfs_handle_t *hdl = zhp->zfs_hdl;
+ nvpair_t *elem;
+ nvlist_t *propval;
+
+ nvlist_free(zhp->zfs_user_props);
+
+ if (nvlist_alloc(&zhp->zfs_user_props, NV_UNIQUE_NAME, 0) != 0)
+ return (no_memory(hdl));
+
+ elem = NULL;
+ while ((elem = nvlist_next_nvpair(zhp->zfs_props, elem)) != NULL) {
+ if (!zfs_prop_user(nvpair_name(elem)))
+ continue;
+
+ verify(nvpair_value_nvlist(elem, &propval) == 0);
+ if (nvlist_add_nvlist(zhp->zfs_user_props,
+ nvpair_name(elem), propval) != 0)
+ return (no_memory(hdl));
+ }
+
+ return (0);
+}
+
+/*
+ * Utility function to gather stats (objset and zpl) for the given object.
+ */
+static int
+get_stats(zfs_handle_t *zhp)
+{
+ zfs_cmd_t zc = { 0 };
+ libzfs_handle_t *hdl = zhp->zfs_hdl;
+
+ (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name));
+
+ if (zcmd_alloc_dst_nvlist(hdl, &zc, 0) != 0)
+ return (-1);
+
+ while (ioctl(zhp->zfs_hdl->libzfs_fd, ZFS_IOC_OBJSET_STATS, &zc) != 0) {
+ if (errno == ENOMEM) {
+ if (zcmd_expand_dst_nvlist(hdl, &zc) != 0) {
+ zcmd_free_nvlists(&zc);
+ return (-1);
+ }
+ } else {
+ zcmd_free_nvlists(&zc);
+ return (-1);
+ }
+ }
+
+ zhp->zfs_dmustats = zc.zc_objset_stats; /* structure assignment */
+
+ (void) strlcpy(zhp->zfs_root, zc.zc_value, sizeof (zhp->zfs_root));
+
+ if (zhp->zfs_props) {
+ nvlist_free(zhp->zfs_props);
+ zhp->zfs_props = NULL;
+ }
+
+ if (zcmd_read_dst_nvlist(hdl, &zc, &zhp->zfs_props) != 0) {
+ zcmd_free_nvlists(&zc);
+ return (-1);
+ }
+
+ zcmd_free_nvlists(&zc);
+
+ if (process_user_props(zhp) != 0)
+ return (-1);
+
+ return (0);
+}
+
+/*
+ * Refresh the properties currently stored in the handle.
+ */
+void
+zfs_refresh_properties(zfs_handle_t *zhp)
+{
+ (void) get_stats(zhp);
+}
+
+/*
+ * Makes a handle from the given dataset name. Used by zfs_open() and
+ * zfs_iter_* to create child handles on the fly.
+ */
+zfs_handle_t *
+make_dataset_handle(libzfs_handle_t *hdl, const char *path)
+{
+ zfs_handle_t *zhp = calloc(sizeof (zfs_handle_t), 1);
+
+ if (zhp == NULL)
+ return (NULL);
+
+ zhp->zfs_hdl = hdl;
+
+top:
+ (void) strlcpy(zhp->zfs_name, path, sizeof (zhp->zfs_name));
+
+ if (get_stats(zhp) != 0) {
+ free(zhp);
+ return (NULL);
+ }
+
+ if (zhp->zfs_dmustats.dds_inconsistent) {
+ zfs_cmd_t zc = { 0 };
+
+ /*
+ * If it is dds_inconsistent, then we've caught it in
+ * the middle of a 'zfs receive' or 'zfs destroy', and
+ * it is inconsistent from the ZPL's point of view, so
+ * can't be mounted. However, it could also be that we
+ * have crashed in the middle of one of those
+ * operations, in which case we need to get rid of the
+ * inconsistent state. We do that by either rolling
+ * back to the previous snapshot (which will fail if
+ * there is none), or destroying the filesystem. Note
+ * that if we are still in the middle of an active
+ * 'receive' or 'destroy', then the rollback and destroy
+ * will fail with EBUSY and we will drive on as usual.
+ */
+
+ (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name));
+
+ if (zhp->zfs_dmustats.dds_type == DMU_OST_ZVOL) {
+ (void) zvol_remove_link(hdl, zhp->zfs_name);
+ zc.zc_objset_type = DMU_OST_ZVOL;
+ } else {
+ zc.zc_objset_type = DMU_OST_ZFS;
+ }
+
+ /* If we can successfully roll it back, reget the stats */
+ if (ioctl(hdl->libzfs_fd, ZFS_IOC_ROLLBACK, &zc) == 0)
+ goto top;
+ /*
+ * If we can sucessfully destroy it, pretend that it
+ * never existed.
+ */
+ if (ioctl(hdl->libzfs_fd, ZFS_IOC_DESTROY, &zc) == 0) {
+ free(zhp);
+ errno = ENOENT;
+ return (NULL);
+ }
+ }
+
+ /*
+ * We've managed to open the dataset and gather statistics. Determine
+ * the high-level type.
+ */
+ if (zhp->zfs_dmustats.dds_type == DMU_OST_ZVOL)
+ zhp->zfs_head_type = ZFS_TYPE_VOLUME;
+ else if (zhp->zfs_dmustats.dds_type == DMU_OST_ZFS)
+ zhp->zfs_head_type = ZFS_TYPE_FILESYSTEM;
+ else
+ abort();
+
+ if (zhp->zfs_dmustats.dds_is_snapshot)
+ zhp->zfs_type = ZFS_TYPE_SNAPSHOT;
+ else if (zhp->zfs_dmustats.dds_type == DMU_OST_ZVOL)
+ zhp->zfs_type = ZFS_TYPE_VOLUME;
+ else if (zhp->zfs_dmustats.dds_type == DMU_OST_ZFS)
+ zhp->zfs_type = ZFS_TYPE_FILESYSTEM;
+ else
+ abort(); /* we should never see any other types */
+
+ return (zhp);
+}
+
+/*
+ * Opens the given snapshot, filesystem, or volume. The 'types'
+ * argument is a mask of acceptable types. The function will print an
+ * appropriate error message and return NULL if it can't be opened.
+ */
+zfs_handle_t *
+zfs_open(libzfs_handle_t *hdl, const char *path, int types)
+{
+ zfs_handle_t *zhp;
+ char errbuf[1024];
+
+ (void) snprintf(errbuf, sizeof (errbuf),
+ dgettext(TEXT_DOMAIN, "cannot open '%s'"), path);
+
+ /*
+ * Validate the name before we even try to open it.
+ */
+ if (!zfs_validate_name(hdl, path, ZFS_TYPE_ANY)) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "invalid dataset name"));
+ (void) zfs_error(hdl, EZFS_INVALIDNAME, errbuf);
+ return (NULL);
+ }
+
+ /*
+ * Try to get stats for the dataset, which will tell us if it exists.
+ */
+ errno = 0;
+ if ((zhp = make_dataset_handle(hdl, path)) == NULL) {
+ (void) zfs_standard_error(hdl, errno, errbuf);
+ return (NULL);
+ }
+
+ if (!(types & zhp->zfs_type)) {
+ (void) zfs_error(hdl, EZFS_BADTYPE, errbuf);
+ zfs_close(zhp);
+ return (NULL);
+ }
+
+ return (zhp);
+}
+
+/*
+ * Release a ZFS handle. Nothing to do but free the associated memory.
+ */
+void
+zfs_close(zfs_handle_t *zhp)
+{
+ if (zhp->zfs_mntopts)
+ free(zhp->zfs_mntopts);
+ nvlist_free(zhp->zfs_props);
+ nvlist_free(zhp->zfs_user_props);
+ free(zhp);
+}
+
+/*
+ * Given a numeric suffix, convert the value into a number of bits that the
+ * resulting value must be shifted.
+ */
+static int
+str2shift(libzfs_handle_t *hdl, const char *buf)
+{
+ const char *ends = "BKMGTPEZ";
+ int i;
+
+ if (buf[0] == '\0')
+ return (0);
+ for (i = 0; i < strlen(ends); i++) {
+ if (toupper(buf[0]) == ends[i])
+ break;
+ }
+ if (i == strlen(ends)) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "invalid numeric suffix '%s'"), buf);
+ return (-1);
+ }
+
+ /*
+ * We want to allow trailing 'b' characters for 'GB' or 'Mb'. But don't
+ * allow 'BB' - that's just weird.
+ */
+ if (buf[1] == '\0' || (toupper(buf[1]) == 'B' && buf[2] == '\0' &&
+ toupper(buf[0]) != 'B'))
+ return (10*i);
+
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "invalid numeric suffix '%s'"), buf);
+ return (-1);
+}
+
+/*
+ * Convert a string of the form '100G' into a real number. Used when setting
+ * properties or creating a volume. 'buf' is used to place an extended error
+ * message for the caller to use.
+ */
+static int
+nicestrtonum(libzfs_handle_t *hdl, const char *value, uint64_t *num)
+{
+ char *end;
+ int shift;
+
+ *num = 0;
+
+ /* Check to see if this looks like a number. */
+ if ((value[0] < '0' || value[0] > '9') && value[0] != '.') {
+ if (hdl)
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "bad numeric value '%s'"), value);
+ return (-1);
+ }
+
+ /* Rely on stroll() to process the numeric portion. */
+ errno = 0;
+ *num = strtoll(value, &end, 10);
+
+ /*
+ * Check for ERANGE, which indicates that the value is too large to fit
+ * in a 64-bit value.
+ */
+ if (errno == ERANGE) {
+ if (hdl)
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "numeric value is too large"));
+ return (-1);
+ }
+
+ /*
+ * If we have a decimal value, then do the computation with floating
+ * point arithmetic. Otherwise, use standard arithmetic.
+ */
+ if (*end == '.') {
+ double fval = strtod(value, &end);
+
+ if ((shift = str2shift(hdl, end)) == -1)
+ return (-1);
+
+ fval *= pow(2, shift);
+
+ if (fval > UINT64_MAX) {
+ if (hdl)
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "numeric value is too large"));
+ return (-1);
+ }
+
+ *num = (uint64_t)fval;
+ } else {
+ if ((shift = str2shift(hdl, end)) == -1)
+ return (-1);
+
+ /* Check for overflow */
+ if (shift >= 64 || (*num << shift) >> shift != *num) {
+ if (hdl)
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "numeric value is too large"));
+ return (-1);
+ }
+
+ *num <<= shift;
+ }
+
+ return (0);
+}
+
+int
+zfs_nicestrtonum(libzfs_handle_t *hdl, const char *str, uint64_t *val)
+{
+ return (nicestrtonum(hdl, str, val));
+}
+
+/*
+ * The prop_parse_*() functions are designed to allow flexibility in callers
+ * when setting properties. At the DSL layer, all properties are either 64-bit
+ * numbers or strings. We want the user to be able to ignore this fact and
+ * specify properties as native values (boolean, for example) or as strings (to
+ * simplify command line utilities). This also handles converting index types
+ * (compression, checksum, etc) from strings to their on-disk index.
+ */
+
+static int
+prop_parse_boolean(libzfs_handle_t *hdl, nvpair_t *elem, uint64_t *val)
+{
+ uint64_t ret;
+
+ switch (nvpair_type(elem)) {
+ case DATA_TYPE_STRING:
+ {
+ char *value;
+ verify(nvpair_value_string(elem, &value) == 0);
+
+ if (strcmp(value, "on") == 0) {
+ ret = 1;
+ } else if (strcmp(value, "off") == 0) {
+ ret = 0;
+ } else {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "property '%s' must be 'on' or 'off'"),
+ nvpair_name(elem));
+ return (-1);
+ }
+ break;
+ }
+
+ case DATA_TYPE_UINT64:
+ {
+ verify(nvpair_value_uint64(elem, &ret) == 0);
+ if (ret > 1) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "'%s' must be a boolean value"),
+ nvpair_name(elem));
+ return (-1);
+ }
+ break;
+ }
+
+ case DATA_TYPE_BOOLEAN_VALUE:
+ {
+ boolean_t value;
+ verify(nvpair_value_boolean_value(elem, &value) == 0);
+ ret = value;
+ break;
+ }
+
+ default:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "'%s' must be a boolean value"),
+ nvpair_name(elem));
+ return (-1);
+ }
+
+ *val = ret;
+ return (0);
+}
+
+static int
+prop_parse_number(libzfs_handle_t *hdl, nvpair_t *elem, zfs_prop_t prop,
+ uint64_t *val)
+{
+ uint64_t ret;
+ boolean_t isnone = B_FALSE;
+
+ switch (nvpair_type(elem)) {
+ case DATA_TYPE_STRING:
+ {
+ char *value;
+ (void) nvpair_value_string(elem, &value);
+ if (strcmp(value, "none") == 0) {
+ isnone = B_TRUE;
+ ret = 0;
+ } else if (nicestrtonum(hdl, value, &ret) != 0) {
+ return (-1);
+ }
+ break;
+ }
+
+ case DATA_TYPE_UINT64:
+ (void) nvpair_value_uint64(elem, &ret);
+ break;
+
+ default:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "'%s' must be a number"),
+ nvpair_name(elem));
+ return (-1);
+ }
+
+ /*
+ * Quota special: force 'none' and don't allow 0.
+ */
+ if (ret == 0 && !isnone && prop == ZFS_PROP_QUOTA) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "use 'none' to disable quota"));
+ return (-1);
+ }
+
+ *val = ret;
+ return (0);
+}
+
+static int
+prop_parse_index(libzfs_handle_t *hdl, nvpair_t *elem, zfs_prop_t prop,
+ uint64_t *val)
+{
+ char *propname = nvpair_name(elem);
+ char *value;
+
+ if (nvpair_type(elem) != DATA_TYPE_STRING) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "'%s' must be a string"), propname);
+ return (-1);
+ }
+
+ (void) nvpair_value_string(elem, &value);
+
+ if (zfs_prop_string_to_index(prop, value, val) != 0) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "'%s' must be one of '%s'"), propname,
+ zfs_prop_values(prop));
+ return (-1);
+ }
+
+ return (0);
+}
+
+/*
+ * Check if the bootfs name has the same pool name as it is set to.
+ * Assuming bootfs is a valid dataset name.
+ */
+static boolean_t
+bootfs_poolname_valid(char *pool, char *bootfs)
+{
+ char ch, *pname;
+
+ /* get the pool name from the bootfs name */
+ pname = bootfs;
+ while (*bootfs && !isspace(*bootfs) && *bootfs != '/')
+ bootfs++;
+
+ ch = *bootfs;
+ *bootfs = 0;
+
+ if (strcmp(pool, pname) == 0) {
+ *bootfs = ch;
+ return (B_TRUE);
+ }
+
+ *bootfs = ch;
+ return (B_FALSE);
+}
+
+/*
+ * Given an nvlist of properties to set, validates that they are correct, and
+ * parses any numeric properties (index, boolean, etc) if they are specified as
+ * strings.
+ */
+nvlist_t *
+zfs_validate_properties(libzfs_handle_t *hdl, zfs_type_t type, char *pool_name,
+ nvlist_t *nvl, uint64_t zoned, zfs_handle_t *zhp, const char *errbuf)
+{
+ nvpair_t *elem;
+ const char *propname;
+ zfs_prop_t prop;
+ uint64_t intval;
+ char *strval;
+ nvlist_t *ret;
+ int isuser;
+
+ if (nvlist_alloc(&ret, NV_UNIQUE_NAME, 0) != 0) {
+ (void) no_memory(hdl);
+ return (NULL);
+ }
+
+ if (type == ZFS_TYPE_SNAPSHOT) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "snapshot properties cannot be modified"));
+ (void) zfs_error(hdl, EZFS_PROPTYPE, errbuf);
+ goto error;
+ }
+
+ elem = NULL;
+ while ((elem = nvlist_next_nvpair(nvl, elem)) != NULL) {
+ propname = nvpair_name(elem);
+
+ /*
+ * Make sure this property is valid and applies to this type.
+ */
+ if ((prop = zfs_name_to_prop_common(propname, type))
+ == ZFS_PROP_INVAL) {
+ isuser = zfs_prop_user(propname);
+ if (!isuser || (isuser && (type & ZFS_TYPE_POOL))) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "invalid property '%s'"),
+ propname);
+ (void) zfs_error(hdl, EZFS_BADPROP, errbuf);
+ goto error;
+ } else {
+ /*
+ * If this is a user property, make sure it's a
+ * string, and that it's less than
+ * ZAP_MAXNAMELEN.
+ */
+ if (nvpair_type(elem) != DATA_TYPE_STRING) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "'%s' must be a string"),
+ propname);
+ (void) zfs_error(hdl, EZFS_BADPROP,
+ errbuf);
+ goto error;
+ }
+
+ if (strlen(nvpair_name(elem)) >=
+ ZAP_MAXNAMELEN) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "property name '%s' is too long"),
+ propname);
+ (void) zfs_error(hdl, EZFS_BADPROP,
+ errbuf);
+ goto error;
+ }
+ }
+
+ (void) nvpair_value_string(elem, &strval);
+ if (nvlist_add_string(ret, propname, strval) != 0) {
+ (void) no_memory(hdl);
+ goto error;
+ }
+ continue;
+ }
+
+ /*
+ * Normalize the name, to get rid of shorthand abbrevations.
+ */
+ propname = zfs_prop_to_name(prop);
+
+ if (!zfs_prop_valid_for_type(prop, type)) {
+ zfs_error_aux(hdl,
+ dgettext(TEXT_DOMAIN, "'%s' does not "
+ "apply to datasets of this type"), propname);
+ (void) zfs_error(hdl, EZFS_PROPTYPE, errbuf);
+ goto error;
+ }
+
+ if (zfs_prop_readonly(prop) &&
+ (prop != ZFS_PROP_VOLBLOCKSIZE || zhp != NULL)) {
+ zfs_error_aux(hdl,
+ dgettext(TEXT_DOMAIN, "'%s' is readonly"),
+ propname);
+ (void) zfs_error(hdl, EZFS_PROPREADONLY, errbuf);
+ goto error;
+ }
+
+ /*
+ * Convert any properties to the internal DSL value types.
+ */
+ strval = NULL;
+ switch (zfs_prop_get_type(prop)) {
+ case prop_type_boolean:
+ if (prop_parse_boolean(hdl, elem, &intval) != 0) {
+ (void) zfs_error(hdl, EZFS_BADPROP, errbuf);
+ goto error;
+ }
+ break;
+
+ case prop_type_string:
+ if (nvpair_type(elem) != DATA_TYPE_STRING) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "'%s' must be a string"),
+ propname);
+ (void) zfs_error(hdl, EZFS_BADPROP, errbuf);
+ goto error;
+ }
+ (void) nvpair_value_string(elem, &strval);
+ if (strlen(strval) >= ZFS_MAXPROPLEN) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "'%s' is too long"), propname);
+ (void) zfs_error(hdl, EZFS_BADPROP, errbuf);
+ goto error;
+ }
+ break;
+
+ case prop_type_number:
+ if (prop_parse_number(hdl, elem, prop, &intval) != 0) {
+ (void) zfs_error(hdl, EZFS_BADPROP, errbuf);
+ goto error;
+ }
+ break;
+
+ case prop_type_index:
+ if (prop_parse_index(hdl, elem, prop, &intval) != 0) {
+ (void) zfs_error(hdl, EZFS_BADPROP, errbuf);
+ goto error;
+ }
+ break;
+
+ default:
+ abort();
+ }
+
+ /*
+ * Add the result to our return set of properties.
+ */
+ if (strval) {
+ if (nvlist_add_string(ret, propname, strval) != 0) {
+ (void) no_memory(hdl);
+ goto error;
+ }
+ } else if (nvlist_add_uint64(ret, propname, intval) != 0) {
+ (void) no_memory(hdl);
+ goto error;
+ }
+
+ /*
+ * Perform some additional checks for specific properties.
+ */
+ switch (prop) {
+ case ZFS_PROP_RECORDSIZE:
+ case ZFS_PROP_VOLBLOCKSIZE:
+ /* must be power of two within SPA_{MIN,MAX}BLOCKSIZE */
+ if (intval < SPA_MINBLOCKSIZE ||
+ intval > SPA_MAXBLOCKSIZE || !ISP2(intval)) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "'%s' must be power of 2 from %u "
+ "to %uk"), propname,
+ (uint_t)SPA_MINBLOCKSIZE,
+ (uint_t)SPA_MAXBLOCKSIZE >> 10);
+ (void) zfs_error(hdl, EZFS_BADPROP, errbuf);
+ goto error;
+ }
+ break;
+
+ case ZFS_PROP_SHAREISCSI:
+ if (strcmp(strval, "off") != 0 &&
+ strcmp(strval, "on") != 0 &&
+ strcmp(strval, "type=disk") != 0) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "'%s' must be 'on', 'off', or 'type=disk'"),
+ propname);
+ (void) zfs_error(hdl, EZFS_BADPROP, errbuf);
+ goto error;
+ }
+
+ break;
+
+ case ZFS_PROP_MOUNTPOINT:
+ if (strcmp(strval, ZFS_MOUNTPOINT_NONE) == 0 ||
+ strcmp(strval, ZFS_MOUNTPOINT_LEGACY) == 0)
+ break;
+
+ if (strval[0] != '/') {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "'%s' must be an absolute path, "
+ "'none', or 'legacy'"), propname);
+ (void) zfs_error(hdl, EZFS_BADPROP, errbuf);
+ goto error;
+ }
+ /*FALLTHRU*/
+
+ case ZFS_PROP_SHARENFS:
+ /*
+ * For the mountpoint and sharenfs properties, check if
+ * it can be set in a global/non-global zone based on
+ * the zoned property value:
+ *
+ * global zone non-global zone
+ * --------------------------------------------------
+ * zoned=on mountpoint (no) mountpoint (yes)
+ * sharenfs (no) sharenfs (no)
+ *
+ * zoned=off mountpoint (yes) N/A
+ * sharenfs (yes)
+ */
+ if (zoned) {
+ if (getzoneid() == GLOBAL_ZONEID) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "'%s' cannot be set on "
+ "dataset in a non-global zone"),
+ propname);
+ (void) zfs_error(hdl, EZFS_ZONED,
+ errbuf);
+ goto error;
+ } else if (prop == ZFS_PROP_SHARENFS) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "'%s' cannot be set in "
+ "a non-global zone"), propname);
+ (void) zfs_error(hdl, EZFS_ZONED,
+ errbuf);
+ goto error;
+ }
+ } else if (getzoneid() != GLOBAL_ZONEID) {
+ /*
+ * If zoned property is 'off', this must be in
+ * a globle zone. If not, something is wrong.
+ */
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "'%s' cannot be set while dataset "
+ "'zoned' property is set"), propname);
+ (void) zfs_error(hdl, EZFS_ZONED, errbuf);
+ goto error;
+ }
+
+ break;
+
+ case ZFS_PROP_BOOTFS:
+ /*
+ * bootfs property value has to be a dataset name and
+ * the dataset has to be in the same pool as it sets to.
+ */
+ if (strval[0] != '\0' && (!zfs_name_valid(strval,
+ ZFS_TYPE_FILESYSTEM) || !bootfs_poolname_valid(
+ pool_name, strval))) {
+
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' "
+ "is an invalid name"), strval);
+ (void) zfs_error(hdl, EZFS_INVALIDNAME, errbuf);
+ goto error;
+ }
+ break;
+ }
+
+ /*
+ * For changes to existing volumes, we have some additional
+ * checks to enforce.
+ */
+ if (type == ZFS_TYPE_VOLUME && zhp != NULL) {
+ uint64_t volsize = zfs_prop_get_int(zhp,
+ ZFS_PROP_VOLSIZE);
+ uint64_t blocksize = zfs_prop_get_int(zhp,
+ ZFS_PROP_VOLBLOCKSIZE);
+ char buf[64];
+
+ switch (prop) {
+ case ZFS_PROP_RESERVATION:
+ if (intval > volsize) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "'%s' is greater than current "
+ "volume size"), propname);
+ (void) zfs_error(hdl, EZFS_BADPROP,
+ errbuf);
+ goto error;
+ }
+ break;
+
+ case ZFS_PROP_VOLSIZE:
+ if (intval % blocksize != 0) {
+ zfs_nicenum(blocksize, buf,
+ sizeof (buf));
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "'%s' must be a multiple of "
+ "volume block size (%s)"),
+ propname, buf);
+ (void) zfs_error(hdl, EZFS_BADPROP,
+ errbuf);
+ goto error;
+ }
+
+ if (intval == 0) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "'%s' cannot be zero"),
+ propname);
+ (void) zfs_error(hdl, EZFS_BADPROP,
+ errbuf);
+ goto error;
+ }
+ break;
+ }
+ }
+ }
+
+ /*
+ * If this is an existing volume, and someone is setting the volsize,
+ * make sure that it matches the reservation, or add it if necessary.
+ */
+ if (zhp != NULL && type == ZFS_TYPE_VOLUME &&
+ nvlist_lookup_uint64(ret, zfs_prop_to_name(ZFS_PROP_VOLSIZE),
+ &intval) == 0) {
+ uint64_t old_volsize = zfs_prop_get_int(zhp,
+ ZFS_PROP_VOLSIZE);
+ uint64_t old_reservation = zfs_prop_get_int(zhp,
+ ZFS_PROP_RESERVATION);
+ uint64_t new_reservation;
+
+ if (old_volsize == old_reservation &&
+ nvlist_lookup_uint64(ret,
+ zfs_prop_to_name(ZFS_PROP_RESERVATION),
+ &new_reservation) != 0) {
+ if (nvlist_add_uint64(ret,
+ zfs_prop_to_name(ZFS_PROP_RESERVATION),
+ intval) != 0) {
+ (void) no_memory(hdl);
+ goto error;
+ }
+ }
+ }
+
+ return (ret);
+
+error:
+ nvlist_free(ret);
+ return (NULL);
+}
+
+/*
+ * Given a property name and value, set the property for the given dataset.
+ */
+int
+zfs_prop_set(zfs_handle_t *zhp, const char *propname, const char *propval)
+{
+ zfs_cmd_t zc = { 0 };
+ int ret = -1;
+ prop_changelist_t *cl = NULL;
+ char errbuf[1024];
+ libzfs_handle_t *hdl = zhp->zfs_hdl;
+ nvlist_t *nvl = NULL, *realprops;
+ zfs_prop_t prop;
+
+ (void) snprintf(errbuf, sizeof (errbuf),
+ dgettext(TEXT_DOMAIN, "cannot set property for '%s'"),
+ zhp->zfs_name);
+
+ if (nvlist_alloc(&nvl, NV_UNIQUE_NAME, 0) != 0 ||
+ nvlist_add_string(nvl, propname, propval) != 0) {
+ (void) no_memory(hdl);
+ goto error;
+ }
+
+ if ((realprops = zfs_validate_properties(hdl, zhp->zfs_type, NULL, nvl,
+ zfs_prop_get_int(zhp, ZFS_PROP_ZONED), zhp, errbuf)) == NULL)
+ goto error;
+ nvlist_free(nvl);
+ nvl = realprops;
+
+ prop = zfs_name_to_prop(propname);
+
+ /* We don't support those properties on FreeBSD. */
+ switch (prop) {
+ case ZFS_PROP_SHAREISCSI:
+ case ZFS_PROP_DEVICES:
+ case ZFS_PROP_ACLMODE:
+ case ZFS_PROP_ACLINHERIT:
+ case ZFS_PROP_ISCSIOPTIONS:
+ (void) snprintf(errbuf, sizeof (errbuf),
+ "property '%s' not supported on FreeBSD", propname);
+ ret = zfs_error(hdl, EZFS_PERM, errbuf);
+ goto error;
+ }
+
+ if ((cl = changelist_gather(zhp, prop, 0)) == NULL)
+ goto error;
+
+ if (prop == ZFS_PROP_MOUNTPOINT && changelist_haszonedchild(cl)) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "child dataset with inherited mountpoint is used "
+ "in a non-global zone"));
+ ret = zfs_error(hdl, EZFS_ZONED, errbuf);
+ goto error;
+ }
+
+ if ((ret = changelist_prefix(cl)) != 0)
+ goto error;
+
+ /*
+ * Execute the corresponding ioctl() to set this property.
+ */
+ (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name));
+
+ if (zcmd_write_src_nvlist(hdl, &zc, nvl, NULL) != 0)
+ goto error;
+
+ ret = ioctl(hdl->libzfs_fd, ZFS_IOC_SET_PROP, &zc);
+
+ if (ret != 0) {
+ switch (errno) {
+
+ case ENOSPC:
+ /*
+ * For quotas and reservations, ENOSPC indicates
+ * something different; setting a quota or reservation
+ * doesn't use any disk space.
+ */
+ switch (prop) {
+ case ZFS_PROP_QUOTA:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "size is less than current used or "
+ "reserved space"));
+ (void) zfs_error(hdl, EZFS_PROPSPACE, errbuf);
+ break;
+
+ case ZFS_PROP_RESERVATION:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "size is greater than available space"));
+ (void) zfs_error(hdl, EZFS_PROPSPACE, errbuf);
+ break;
+
+ default:
+ (void) zfs_standard_error(hdl, errno, errbuf);
+ break;
+ }
+ break;
+
+ case EBUSY:
+ if (prop == ZFS_PROP_VOLBLOCKSIZE)
+ (void) zfs_error(hdl, EZFS_VOLHASDATA, errbuf);
+ else
+ (void) zfs_standard_error(hdl, EBUSY, errbuf);
+ break;
+
+ case EROFS:
+ (void) zfs_error(hdl, EZFS_DSREADONLY, errbuf);
+ break;
+
+ case ENOTSUP:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "pool must be upgraded to allow gzip compression"));
+ (void) zfs_error(hdl, EZFS_BADVERSION, errbuf);
+ break;
+
+ case EOVERFLOW:
+ /*
+ * This platform can't address a volume this big.
+ */
+#ifdef _ILP32
+ if (prop == ZFS_PROP_VOLSIZE) {
+ (void) zfs_error(hdl, EZFS_VOLTOOBIG, errbuf);
+ break;
+ }
+#endif
+ /* FALLTHROUGH */
+ default:
+ (void) zfs_standard_error(hdl, errno, errbuf);
+ }
+ } else {
+ /*
+ * Refresh the statistics so the new property value
+ * is reflected.
+ */
+ if ((ret = changelist_postfix(cl)) == 0)
+ (void) get_stats(zhp);
+ }
+
+error:
+ nvlist_free(nvl);
+ zcmd_free_nvlists(&zc);
+ if (cl)
+ changelist_free(cl);
+ return (ret);
+}
+
+/*
+ * Given a property, inherit the value from the parent dataset.
+ */
+int
+zfs_prop_inherit(zfs_handle_t *zhp, const char *propname)
+{
+ zfs_cmd_t zc = { 0 };
+ int ret;
+ prop_changelist_t *cl;
+ libzfs_handle_t *hdl = zhp->zfs_hdl;
+ char errbuf[1024];
+ zfs_prop_t prop;
+
+ (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
+ "cannot inherit %s for '%s'"), propname, zhp->zfs_name);
+
+ if ((prop = zfs_name_to_prop(propname)) == ZFS_PROP_INVAL) {
+ /*
+ * For user properties, the amount of work we have to do is very
+ * small, so just do it here.
+ */
+ if (!zfs_prop_user(propname)) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "invalid property"));
+ return (zfs_error(hdl, EZFS_BADPROP, errbuf));
+ }
+
+ (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name));
+ (void) strlcpy(zc.zc_value, propname, sizeof (zc.zc_value));
+
+ if (ioctl(zhp->zfs_hdl->libzfs_fd,
+ ZFS_IOC_SET_PROP, &zc) != 0)
+ return (zfs_standard_error(hdl, errno, errbuf));
+
+ return (0);
+ }
+
+ /*
+ * Verify that this property is inheritable.
+ */
+ if (zfs_prop_readonly(prop))
+ return (zfs_error(hdl, EZFS_PROPREADONLY, errbuf));
+
+ if (!zfs_prop_inheritable(prop))
+ return (zfs_error(hdl, EZFS_PROPNONINHERIT, errbuf));
+
+ /*
+ * Check to see if the value applies to this type
+ */
+ if (!zfs_prop_valid_for_type(prop, zhp->zfs_type))
+ return (zfs_error(hdl, EZFS_PROPTYPE, errbuf));
+
+ /*
+ * Normalize the name, to get rid of shorthand abbrevations.
+ */
+ propname = zfs_prop_to_name(prop);
+ (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name));
+ (void) strlcpy(zc.zc_value, propname, sizeof (zc.zc_value));
+
+ if (prop == ZFS_PROP_MOUNTPOINT && getzoneid() == GLOBAL_ZONEID &&
+ zfs_prop_get_int(zhp, ZFS_PROP_ZONED)) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "dataset is used in a non-global zone"));
+ return (zfs_error(hdl, EZFS_ZONED, errbuf));
+ }
+
+ /*
+ * Determine datasets which will be affected by this change, if any.
+ */
+ if ((cl = changelist_gather(zhp, prop, 0)) == NULL)
+ return (-1);
+
+ if (prop == ZFS_PROP_MOUNTPOINT && changelist_haszonedchild(cl)) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "child dataset with inherited mountpoint is used "
+ "in a non-global zone"));
+ ret = zfs_error(hdl, EZFS_ZONED, errbuf);
+ goto error;
+ }
+
+ if ((ret = changelist_prefix(cl)) != 0)
+ goto error;
+
+ if ((ret = ioctl(zhp->zfs_hdl->libzfs_fd,
+ ZFS_IOC_SET_PROP, &zc)) != 0) {
+ return (zfs_standard_error(hdl, errno, errbuf));
+ } else {
+
+ if ((ret = changelist_postfix(cl)) != 0)
+ goto error;
+
+ /*
+ * Refresh the statistics so the new property is reflected.
+ */
+ (void) get_stats(zhp);
+ }
+
+error:
+ changelist_free(cl);
+ return (ret);
+}
+
+void
+nicebool(int value, char *buf, size_t buflen)
+{
+ if (value)
+ (void) strlcpy(buf, "on", buflen);
+ else
+ (void) strlcpy(buf, "off", buflen);
+}
+
+/*
+ * True DSL properties are stored in an nvlist. The following two functions
+ * extract them appropriately.
+ */
+static uint64_t
+getprop_uint64(zfs_handle_t *zhp, zfs_prop_t prop, char **source)
+{
+ nvlist_t *nv;
+ uint64_t value;
+
+ *source = NULL;
+ if (nvlist_lookup_nvlist(zhp->zfs_props,
+ zfs_prop_to_name(prop), &nv) == 0) {
+ verify(nvlist_lookup_uint64(nv, ZFS_PROP_VALUE, &value) == 0);
+ (void) nvlist_lookup_string(nv, ZFS_PROP_SOURCE, source);
+ } else {
+ value = zfs_prop_default_numeric(prop);
+ *source = "";
+ }
+
+ return (value);
+}
+
+static char *
+getprop_string(zfs_handle_t *zhp, zfs_prop_t prop, char **source)
+{
+ nvlist_t *nv;
+ char *value;
+
+ *source = NULL;
+ if (nvlist_lookup_nvlist(zhp->zfs_props,
+ zfs_prop_to_name(prop), &nv) == 0) {
+ verify(nvlist_lookup_string(nv, ZFS_PROP_VALUE, &value) == 0);
+ (void) nvlist_lookup_string(nv, ZFS_PROP_SOURCE, source);
+ } else {
+ if ((value = (char *)zfs_prop_default_string(prop)) == NULL)
+ value = "";
+ *source = "";
+ }
+
+ return (value);
+}
+
+/*
+ * Internal function for getting a numeric property. Both zfs_prop_get() and
+ * zfs_prop_get_int() are built using this interface.
+ *
+ * Certain properties can be overridden using 'mount -o'. In this case, scan
+ * the contents of the /etc/mnttab entry, searching for the appropriate options.
+ * If they differ from the on-disk values, report the current values and mark
+ * the source "temporary".
+ */
+static int
+get_numeric_property(zfs_handle_t *zhp, zfs_prop_t prop, zfs_source_t *src,
+ char **source, uint64_t *val)
+{
+ struct mnttab mnt;
+ char *mntopt_on = NULL;
+ char *mntopt_off = NULL;
+
+ *source = NULL;
+
+ switch (prop) {
+ case ZFS_PROP_ATIME:
+ mntopt_on = MNTOPT_ATIME;
+ mntopt_off = MNTOPT_NOATIME;
+ break;
+
+ case ZFS_PROP_DEVICES:
+ mntopt_on = MNTOPT_DEVICES;
+ mntopt_off = MNTOPT_NODEVICES;
+ break;
+
+ case ZFS_PROP_EXEC:
+ mntopt_on = MNTOPT_EXEC;
+ mntopt_off = MNTOPT_NOEXEC;
+ break;
+
+ case ZFS_PROP_READONLY:
+ mntopt_on = MNTOPT_RO;
+ mntopt_off = MNTOPT_RW;
+ break;
+
+ case ZFS_PROP_SETUID:
+ mntopt_on = MNTOPT_SETUID;
+ mntopt_off = MNTOPT_NOSETUID;
+ break;
+
+ case ZFS_PROP_XATTR:
+ mntopt_on = MNTOPT_XATTR;
+ mntopt_off = MNTOPT_NOXATTR;
+ break;
+ }
+
+ /*
+ * Because looking up the mount options is potentially expensive
+ * (iterating over all of /etc/mnttab), we defer its calculation until
+ * we're looking up a property which requires its presence.
+ */
+ if (!zhp->zfs_mntcheck &&
+ (mntopt_on != NULL || prop == ZFS_PROP_MOUNTED)) {
+ struct mnttab entry, search = { 0 };
+ FILE *mnttab = zhp->zfs_hdl->libzfs_mnttab;
+
+ search.mnt_special = (char *)zhp->zfs_name;
+ search.mnt_fstype = MNTTYPE_ZFS;
+ rewind(mnttab);
+
+ if (getmntany(mnttab, &entry, &search) == 0) {
+ zhp->zfs_mntopts = zfs_strdup(zhp->zfs_hdl,
+ entry.mnt_mntopts);
+ if (zhp->zfs_mntopts == NULL)
+ return (-1);
+ }
+
+ zhp->zfs_mntcheck = B_TRUE;
+ }
+
+ if (zhp->zfs_mntopts == NULL)
+ mnt.mnt_mntopts = "";
+ else
+ mnt.mnt_mntopts = zhp->zfs_mntopts;
+
+ switch (prop) {
+ case ZFS_PROP_ATIME:
+ case ZFS_PROP_DEVICES:
+ case ZFS_PROP_EXEC:
+ case ZFS_PROP_READONLY:
+ case ZFS_PROP_SETUID:
+ case ZFS_PROP_XATTR:
+ *val = getprop_uint64(zhp, prop, source);
+
+ if (hasmntopt(&mnt, mntopt_on) && !*val) {
+ *val = B_TRUE;
+ if (src)
+ *src = ZFS_SRC_TEMPORARY;
+ } else if (hasmntopt(&mnt, mntopt_off) && *val) {
+ *val = B_FALSE;
+ if (src)
+ *src = ZFS_SRC_TEMPORARY;
+ }
+ break;
+
+ case ZFS_PROP_RECORDSIZE:
+ case ZFS_PROP_COMPRESSION:
+ case ZFS_PROP_ZONED:
+ case ZFS_PROP_CREATION:
+ case ZFS_PROP_COMPRESSRATIO:
+ case ZFS_PROP_REFERENCED:
+ case ZFS_PROP_USED:
+ case ZFS_PROP_CREATETXG:
+ case ZFS_PROP_AVAILABLE:
+ case ZFS_PROP_VOLSIZE:
+ case ZFS_PROP_VOLBLOCKSIZE:
+ *val = getprop_uint64(zhp, prop, source);
+ break;
+
+ case ZFS_PROP_CANMOUNT:
+ *val = getprop_uint64(zhp, prop, source);
+ if (*val == 0)
+ *source = zhp->zfs_name;
+ else
+ *source = ""; /* default */
+ break;
+
+ case ZFS_PROP_QUOTA:
+ case ZFS_PROP_RESERVATION:
+ *val = getprop_uint64(zhp, prop, source);
+ if (*val == 0)
+ *source = ""; /* default */
+ else
+ *source = zhp->zfs_name;
+ break;
+
+ case ZFS_PROP_MOUNTED:
+ *val = (zhp->zfs_mntopts != NULL);
+ break;
+
+ case ZFS_PROP_NUMCLONES:
+ *val = zhp->zfs_dmustats.dds_num_clones;
+ break;
+
+ default:
+ zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
+ "cannot get non-numeric property"));
+ return (zfs_error(zhp->zfs_hdl, EZFS_BADPROP,
+ dgettext(TEXT_DOMAIN, "internal error")));
+ }
+
+ return (0);
+}
+
+/*
+ * Calculate the source type, given the raw source string.
+ */
+static void
+get_source(zfs_handle_t *zhp, zfs_source_t *srctype, char *source,
+ char *statbuf, size_t statlen)
+{
+ if (statbuf == NULL || *srctype == ZFS_SRC_TEMPORARY)
+ return;
+
+ if (source == NULL) {
+ *srctype = ZFS_SRC_NONE;
+ } else if (source[0] == '\0') {
+ *srctype = ZFS_SRC_DEFAULT;
+ } else {
+ if (strcmp(source, zhp->zfs_name) == 0) {
+ *srctype = ZFS_SRC_LOCAL;
+ } else {
+ (void) strlcpy(statbuf, source, statlen);
+ *srctype = ZFS_SRC_INHERITED;
+ }
+ }
+
+}
+
+/*
+ * Retrieve a property from the given object. If 'literal' is specified, then
+ * numbers are left as exact values. Otherwise, numbers are converted to a
+ * human-readable form.
+ *
+ * Returns 0 on success, or -1 on error.
+ */
+int
+zfs_prop_get(zfs_handle_t *zhp, zfs_prop_t prop, char *propbuf, size_t proplen,
+ zfs_source_t *src, char *statbuf, size_t statlen, boolean_t literal)
+{
+ char *source = NULL;
+ uint64_t val;
+ char *str;
+ const char *root;
+ const char *strval;
+
+ /*
+ * Check to see if this property applies to our object
+ */
+ if (!zfs_prop_valid_for_type(prop, zhp->zfs_type))
+ return (-1);
+
+ if (src)
+ *src = ZFS_SRC_NONE;
+
+ switch (prop) {
+ case ZFS_PROP_ATIME:
+ case ZFS_PROP_READONLY:
+ case ZFS_PROP_SETUID:
+ case ZFS_PROP_ZONED:
+ case ZFS_PROP_DEVICES:
+ case ZFS_PROP_EXEC:
+ case ZFS_PROP_CANMOUNT:
+ case ZFS_PROP_XATTR:
+ /*
+ * Basic boolean values are built on top of
+ * get_numeric_property().
+ */
+ if (get_numeric_property(zhp, prop, src, &source, &val) != 0)
+ return (-1);
+ nicebool(val, propbuf, proplen);
+
+ break;
+
+ case ZFS_PROP_AVAILABLE:
+ case ZFS_PROP_RECORDSIZE:
+ case ZFS_PROP_CREATETXG:
+ case ZFS_PROP_REFERENCED:
+ case ZFS_PROP_USED:
+ case ZFS_PROP_VOLSIZE:
+ case ZFS_PROP_VOLBLOCKSIZE:
+ case ZFS_PROP_NUMCLONES:
+ /*
+ * Basic numeric values are built on top of
+ * get_numeric_property().
+ */
+ if (get_numeric_property(zhp, prop, src, &source, &val) != 0)
+ return (-1);
+ if (literal)
+ (void) snprintf(propbuf, proplen, "%llu",
+ (u_longlong_t)val);
+ else
+ zfs_nicenum(val, propbuf, proplen);
+ break;
+
+ case ZFS_PROP_COMPRESSION:
+ case ZFS_PROP_CHECKSUM:
+ case ZFS_PROP_SNAPDIR:
+#ifdef ZFS_NO_ACL
+ case ZFS_PROP_ACLMODE:
+ case ZFS_PROP_ACLINHERIT:
+ case ZFS_PROP_COPIES:
+ val = getprop_uint64(zhp, prop, &source);
+ verify(zfs_prop_index_to_string(prop, val, &strval) == 0);
+ (void) strlcpy(propbuf, strval, proplen);
+ break;
+#else /* ZFS_NO_ACL */
+ case ZFS_PROP_ACLMODE:
+ case ZFS_PROP_ACLINHERIT:
+ (void) strlcpy(propbuf, "<unsupported>", proplen);
+ break;
+#endif /* ZFS_NO_ACL */
+
+ case ZFS_PROP_CREATION:
+ /*
+ * 'creation' is a time_t stored in the statistics. We convert
+ * this into a string unless 'literal' is specified.
+ */
+ {
+ val = getprop_uint64(zhp, prop, &source);
+ time_t time = (time_t)val;
+ struct tm t;
+
+ if (literal ||
+ localtime_r(&time, &t) == NULL ||
+ strftime(propbuf, proplen, "%a %b %e %k:%M %Y",
+ &t) == 0)
+ (void) snprintf(propbuf, proplen, "%llu", val);
+ }
+ break;
+
+ case ZFS_PROP_MOUNTPOINT:
+ /*
+ * Getting the precise mountpoint can be tricky.
+ *
+ * - for 'none' or 'legacy', return those values.
+ * - for default mountpoints, construct it as /zfs/<dataset>
+ * - for inherited mountpoints, we want to take everything
+ * after our ancestor and append it to the inherited value.
+ *
+ * If the pool has an alternate root, we want to prepend that
+ * root to any values we return.
+ */
+ root = zhp->zfs_root;
+ str = getprop_string(zhp, prop, &source);
+
+ if (str[0] == '\0') {
+ (void) snprintf(propbuf, proplen, "%s/zfs/%s",
+ root, zhp->zfs_name);
+ } else if (str[0] == '/') {
+ const char *relpath = zhp->zfs_name + strlen(source);
+
+ if (relpath[0] == '/')
+ relpath++;
+ if (str[1] == '\0')
+ str++;
+
+ if (relpath[0] == '\0')
+ (void) snprintf(propbuf, proplen, "%s%s",
+ root, str);
+ else
+ (void) snprintf(propbuf, proplen, "%s%s%s%s",
+ root, str, relpath[0] == '@' ? "" : "/",
+ relpath);
+ } else {
+ /* 'legacy' or 'none' */
+ (void) strlcpy(propbuf, str, proplen);
+ }
+
+ break;
+
+ case ZFS_PROP_SHARENFS:
+ case ZFS_PROP_SHAREISCSI:
+ case ZFS_PROP_ISCSIOPTIONS:
+ (void) strlcpy(propbuf, getprop_string(zhp, prop, &source),
+ proplen);
+ break;
+
+ case ZFS_PROP_ORIGIN:
+ (void) strlcpy(propbuf, getprop_string(zhp, prop, &source),
+ proplen);
+ /*
+ * If there is no parent at all, return failure to indicate that
+ * it doesn't apply to this dataset.
+ */
+ if (propbuf[0] == '\0')
+ return (-1);
+ break;
+
+ case ZFS_PROP_QUOTA:
+ case ZFS_PROP_RESERVATION:
+ if (get_numeric_property(zhp, prop, src, &source, &val) != 0)
+ return (-1);
+
+ /*
+ * If quota or reservation is 0, we translate this into 'none'
+ * (unless literal is set), and indicate that it's the default
+ * value. Otherwise, we print the number nicely and indicate
+ * that its set locally.
+ */
+ if (val == 0) {
+ if (literal)
+ (void) strlcpy(propbuf, "0", proplen);
+ else
+ (void) strlcpy(propbuf, "none", proplen);
+ } else {
+ if (literal)
+ (void) snprintf(propbuf, proplen, "%llu",
+ (u_longlong_t)val);
+ else
+ zfs_nicenum(val, propbuf, proplen);
+ }
+ break;
+
+ case ZFS_PROP_COMPRESSRATIO:
+ if (get_numeric_property(zhp, prop, src, &source, &val) != 0)
+ return (-1);
+ (void) snprintf(propbuf, proplen, "%lld.%02lldx", (longlong_t)
+ val / 100, (longlong_t)val % 100);
+ break;
+
+ case ZFS_PROP_TYPE:
+ switch (zhp->zfs_type) {
+ case ZFS_TYPE_FILESYSTEM:
+ str = "filesystem";
+ break;
+ case ZFS_TYPE_VOLUME:
+ str = "volume";
+ break;
+ case ZFS_TYPE_SNAPSHOT:
+ str = "snapshot";
+ break;
+ default:
+ abort();
+ }
+ (void) snprintf(propbuf, proplen, "%s", str);
+ break;
+
+ case ZFS_PROP_MOUNTED:
+ /*
+ * The 'mounted' property is a pseudo-property that described
+ * whether the filesystem is currently mounted. Even though
+ * it's a boolean value, the typical values of "on" and "off"
+ * don't make sense, so we translate to "yes" and "no".
+ */
+ if (get_numeric_property(zhp, ZFS_PROP_MOUNTED,
+ src, &source, &val) != 0)
+ return (-1);
+ if (val)
+ (void) strlcpy(propbuf, "yes", proplen);
+ else
+ (void) strlcpy(propbuf, "no", proplen);
+ break;
+
+ case ZFS_PROP_NAME:
+ /*
+ * The 'name' property is a pseudo-property derived from the
+ * dataset name. It is presented as a real property to simplify
+ * consumers.
+ */
+ (void) strlcpy(propbuf, zhp->zfs_name, proplen);
+ break;
+
+ default:
+ abort();
+ }
+
+ get_source(zhp, src, source, statbuf, statlen);
+
+ return (0);
+}
+
+/*
+ * Utility function to get the given numeric property. Does no validation that
+ * the given property is the appropriate type; should only be used with
+ * hard-coded property types.
+ */
+uint64_t
+zfs_prop_get_int(zfs_handle_t *zhp, zfs_prop_t prop)
+{
+ char *source;
+ zfs_source_t sourcetype = ZFS_SRC_NONE;
+ uint64_t val;
+
+ (void) get_numeric_property(zhp, prop, &sourcetype, &source, &val);
+
+ return (val);
+}
+
+/*
+ * Similar to zfs_prop_get(), but returns the value as an integer.
+ */
+int
+zfs_prop_get_numeric(zfs_handle_t *zhp, zfs_prop_t prop, uint64_t *value,
+ zfs_source_t *src, char *statbuf, size_t statlen)
+{
+ char *source;
+
+ /*
+ * Check to see if this property applies to our object
+ */
+ if (!zfs_prop_valid_for_type(prop, zhp->zfs_type))
+ return (zfs_error_fmt(zhp->zfs_hdl, EZFS_PROPTYPE,
+ dgettext(TEXT_DOMAIN, "cannot get property '%s'"),
+ zfs_prop_to_name(prop)));
+
+ if (src)
+ *src = ZFS_SRC_NONE;
+
+ if (get_numeric_property(zhp, prop, src, &source, value) != 0)
+ return (-1);
+
+ get_source(zhp, src, source, statbuf, statlen);
+
+ return (0);
+}
+
+/*
+ * Returns the name of the given zfs handle.
+ */
+const char *
+zfs_get_name(const zfs_handle_t *zhp)
+{
+ return (zhp->zfs_name);
+}
+
+/*
+ * Returns the type of the given zfs handle.
+ */
+zfs_type_t
+zfs_get_type(const zfs_handle_t *zhp)
+{
+ return (zhp->zfs_type);
+}
+
+/*
+ * Iterate over all child filesystems
+ */
+int
+zfs_iter_filesystems(zfs_handle_t *zhp, zfs_iter_f func, void *data)
+{
+ zfs_cmd_t zc = { 0 };
+ zfs_handle_t *nzhp;
+ int ret;
+
+ for ((void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name));
+ ioctl(zhp->zfs_hdl->libzfs_fd, ZFS_IOC_DATASET_LIST_NEXT, &zc) == 0;
+ (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name))) {
+ /*
+ * Ignore private dataset names.
+ */
+ if (dataset_name_hidden(zc.zc_name))
+ continue;
+
+ /*
+ * Silently ignore errors, as the only plausible explanation is
+ * that the pool has since been removed.
+ */
+ if ((nzhp = make_dataset_handle(zhp->zfs_hdl,
+ zc.zc_name)) == NULL)
+ continue;
+
+ if ((ret = func(nzhp, data)) != 0)
+ return (ret);
+ }
+
+ /*
+ * An errno value of ESRCH indicates normal completion. If ENOENT is
+ * returned, then the underlying dataset has been removed since we
+ * obtained the handle.
+ */
+ if (errno != ESRCH && errno != ENOENT)
+ return (zfs_standard_error(zhp->zfs_hdl, errno,
+ dgettext(TEXT_DOMAIN, "cannot iterate filesystems")));
+
+ return (0);
+}
+
+/*
+ * Iterate over all snapshots
+ */
+int
+zfs_iter_snapshots(zfs_handle_t *zhp, zfs_iter_f func, void *data)
+{
+ zfs_cmd_t zc = { 0 };
+ zfs_handle_t *nzhp;
+ int ret;
+
+ for ((void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name));
+ ioctl(zhp->zfs_hdl->libzfs_fd, ZFS_IOC_SNAPSHOT_LIST_NEXT,
+ &zc) == 0;
+ (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name))) {
+
+ if ((nzhp = make_dataset_handle(zhp->zfs_hdl,
+ zc.zc_name)) == NULL)
+ continue;
+
+ if ((ret = func(nzhp, data)) != 0)
+ return (ret);
+ }
+
+ /*
+ * An errno value of ESRCH indicates normal completion. If ENOENT is
+ * returned, then the underlying dataset has been removed since we
+ * obtained the handle. Silently ignore this case, and return success.
+ */
+ if (errno != ESRCH && errno != ENOENT)
+ return (zfs_standard_error(zhp->zfs_hdl, errno,
+ dgettext(TEXT_DOMAIN, "cannot iterate filesystems")));
+
+ return (0);
+}
+
+/*
+ * Iterate over all children, snapshots and filesystems
+ */
+int
+zfs_iter_children(zfs_handle_t *zhp, zfs_iter_f func, void *data)
+{
+ int ret;
+
+ if ((ret = zfs_iter_filesystems(zhp, func, data)) != 0)
+ return (ret);
+
+ return (zfs_iter_snapshots(zhp, func, data));
+}
+
+/*
+ * Given a complete name, return just the portion that refers to the parent.
+ * Can return NULL if this is a pool.
+ */
+static int
+parent_name(const char *path, char *buf, size_t buflen)
+{
+ char *loc;
+
+ if ((loc = strrchr(path, '/')) == NULL)
+ return (-1);
+
+ (void) strncpy(buf, path, MIN(buflen, loc - path));
+ buf[loc - path] = '\0';
+
+ return (0);
+}
+
+/*
+ * Checks to make sure that the given path has a parent, and that it exists. We
+ * also fetch the 'zoned' property, which is used to validate property settings
+ * when creating new datasets.
+ */
+static int
+check_parents(libzfs_handle_t *hdl, const char *path, uint64_t *zoned)
+{
+ zfs_cmd_t zc = { 0 };
+ char parent[ZFS_MAXNAMELEN];
+ char *slash;
+ zfs_handle_t *zhp;
+ char errbuf[1024];
+
+ (void) snprintf(errbuf, sizeof (errbuf), "cannot create '%s'",
+ path);
+
+ /* get parent, and check to see if this is just a pool */
+ if (parent_name(path, parent, sizeof (parent)) != 0) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "missing dataset name"));
+ return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf));
+ }
+
+ /* check to see if the pool exists */
+ if ((slash = strchr(parent, '/')) == NULL)
+ slash = parent + strlen(parent);
+ (void) strncpy(zc.zc_name, parent, slash - parent);
+ zc.zc_name[slash - parent] = '\0';
+ if (ioctl(hdl->libzfs_fd, ZFS_IOC_OBJSET_STATS, &zc) != 0 &&
+ errno == ENOENT) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "no such pool '%s'"), zc.zc_name);
+ return (zfs_error(hdl, EZFS_NOENT, errbuf));
+ }
+
+ /* check to see if the parent dataset exists */
+ if ((zhp = make_dataset_handle(hdl, parent)) == NULL) {
+ switch (errno) {
+ case ENOENT:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "parent does not exist"));
+ return (zfs_error(hdl, EZFS_NOENT, errbuf));
+
+ default:
+ return (zfs_standard_error(hdl, errno, errbuf));
+ }
+ }
+
+ *zoned = zfs_prop_get_int(zhp, ZFS_PROP_ZONED);
+ /* we are in a non-global zone, but parent is in the global zone */
+ if (getzoneid() != GLOBAL_ZONEID && !(*zoned)) {
+ (void) zfs_standard_error(hdl, EPERM, errbuf);
+ zfs_close(zhp);
+ return (-1);
+ }
+
+ /* make sure parent is a filesystem */
+ if (zfs_get_type(zhp) != ZFS_TYPE_FILESYSTEM) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "parent is not a filesystem"));
+ (void) zfs_error(hdl, EZFS_BADTYPE, errbuf);
+ zfs_close(zhp);
+ return (-1);
+ }
+
+ zfs_close(zhp);
+ return (0);
+}
+
+/*
+ * Create a new filesystem or volume.
+ */
+int
+zfs_create(libzfs_handle_t *hdl, const char *path, zfs_type_t type,
+ nvlist_t *props)
+{
+ zfs_cmd_t zc = { 0 };
+ int ret;
+ uint64_t size = 0;
+ uint64_t blocksize = zfs_prop_default_numeric(ZFS_PROP_VOLBLOCKSIZE);
+ char errbuf[1024];
+ uint64_t zoned;
+
+ (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
+ "cannot create '%s'"), path);
+
+ /* validate the path, taking care to note the extended error message */
+ if (!zfs_validate_name(hdl, path, type))
+ return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf));
+
+ /* validate parents exist */
+ if (check_parents(hdl, path, &zoned) != 0)
+ return (-1);
+
+ /*
+ * The failure modes when creating a dataset of a different type over
+ * one that already exists is a little strange. In particular, if you
+ * try to create a dataset on top of an existing dataset, the ioctl()
+ * will return ENOENT, not EEXIST. To prevent this from happening, we
+ * first try to see if the dataset exists.
+ */
+ (void) strlcpy(zc.zc_name, path, sizeof (zc.zc_name));
+ if (ioctl(hdl->libzfs_fd, ZFS_IOC_OBJSET_STATS, &zc) == 0) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "dataset already exists"));
+ return (zfs_error(hdl, EZFS_EXISTS, errbuf));
+ }
+
+ if (type == ZFS_TYPE_VOLUME)
+ zc.zc_objset_type = DMU_OST_ZVOL;
+ else
+ zc.zc_objset_type = DMU_OST_ZFS;
+
+ if (props && (props = zfs_validate_properties(hdl, type, NULL, props,
+ zoned, NULL, errbuf)) == 0)
+ return (-1);
+
+ if (type == ZFS_TYPE_VOLUME) {
+ /*
+ * If we are creating a volume, the size and block size must
+ * satisfy a few restraints. First, the blocksize must be a
+ * valid block size between SPA_{MIN,MAX}BLOCKSIZE. Second, the
+ * volsize must be a multiple of the block size, and cannot be
+ * zero.
+ */
+ if (props == NULL || nvlist_lookup_uint64(props,
+ zfs_prop_to_name(ZFS_PROP_VOLSIZE), &size) != 0) {
+ nvlist_free(props);
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "missing volume size"));
+ return (zfs_error(hdl, EZFS_BADPROP, errbuf));
+ }
+
+ if ((ret = nvlist_lookup_uint64(props,
+ zfs_prop_to_name(ZFS_PROP_VOLBLOCKSIZE),
+ &blocksize)) != 0) {
+ if (ret == ENOENT) {
+ blocksize = zfs_prop_default_numeric(
+ ZFS_PROP_VOLBLOCKSIZE);
+ } else {
+ nvlist_free(props);
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "missing volume block size"));
+ return (zfs_error(hdl, EZFS_BADPROP, errbuf));
+ }
+ }
+
+ if (size == 0) {
+ nvlist_free(props);
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "volume size cannot be zero"));
+ return (zfs_error(hdl, EZFS_BADPROP, errbuf));
+ }
+
+ if (size % blocksize != 0) {
+ nvlist_free(props);
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "volume size must be a multiple of volume block "
+ "size"));
+ return (zfs_error(hdl, EZFS_BADPROP, errbuf));
+ }
+ }
+
+ if (props &&
+ zcmd_write_src_nvlist(hdl, &zc, props, NULL) != 0)
+ return (-1);
+ nvlist_free(props);
+
+ /* create the dataset */
+ ret = ioctl(hdl->libzfs_fd, ZFS_IOC_CREATE, &zc);
+
+ if (ret == 0 && type == ZFS_TYPE_VOLUME) {
+ ret = zvol_create_link(hdl, path);
+ if (ret) {
+ (void) zfs_standard_error(hdl, errno,
+ dgettext(TEXT_DOMAIN,
+ "Volume successfully created, but device links "
+ "were not created"));
+ zcmd_free_nvlists(&zc);
+ return (-1);
+ }
+ }
+
+ zcmd_free_nvlists(&zc);
+
+ /* check for failure */
+ if (ret != 0) {
+ char parent[ZFS_MAXNAMELEN];
+ (void) parent_name(path, parent, sizeof (parent));
+
+ switch (errno) {
+ case ENOENT:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "no such parent '%s'"), parent);
+ return (zfs_error(hdl, EZFS_NOENT, errbuf));
+
+ case EINVAL:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "parent '%s' is not a filesystem"), parent);
+ return (zfs_error(hdl, EZFS_BADTYPE, errbuf));
+
+ case EDOM:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "volume block size must be power of 2 from "
+ "%u to %uk"),
+ (uint_t)SPA_MINBLOCKSIZE,
+ (uint_t)SPA_MAXBLOCKSIZE >> 10);
+
+ return (zfs_error(hdl, EZFS_BADPROP, errbuf));
+
+#ifdef _ILP32
+ case EOVERFLOW:
+ /*
+ * This platform can't address a volume this big.
+ */
+ if (type == ZFS_TYPE_VOLUME)
+ return (zfs_error(hdl, EZFS_VOLTOOBIG,
+ errbuf));
+#endif
+ /* FALLTHROUGH */
+ default:
+ return (zfs_standard_error(hdl, errno, errbuf));
+ }
+ }
+
+ return (0);
+}
+
+/*
+ * Destroys the given dataset. The caller must make sure that the filesystem
+ * isn't mounted, and that there are no active dependents.
+ */
+int
+zfs_destroy(zfs_handle_t *zhp)
+{
+ zfs_cmd_t zc = { 0 };
+
+ (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name));
+
+ if (ZFS_IS_VOLUME(zhp)) {
+ /*
+ * Unconditionally unshare this zvol ignoring failure as it
+ * indicates only that the volume wasn't shared initially.
+ */
+ (void) zfs_unshare_iscsi(zhp);
+
+ if (zvol_remove_link(zhp->zfs_hdl, zhp->zfs_name) != 0)
+ return (-1);
+
+ zc.zc_objset_type = DMU_OST_ZVOL;
+ } else {
+ zc.zc_objset_type = DMU_OST_ZFS;
+ }
+
+ if (ioctl(zhp->zfs_hdl->libzfs_fd, ZFS_IOC_DESTROY, &zc) != 0) {
+ return (zfs_standard_error_fmt(zhp->zfs_hdl, errno,
+ dgettext(TEXT_DOMAIN, "cannot destroy '%s'"),
+ zhp->zfs_name));
+ }
+
+ remove_mountpoint(zhp);
+
+ return (0);
+}
+
+struct destroydata {
+ char *snapname;
+ boolean_t gotone;
+ boolean_t closezhp;
+};
+
+static int
+zfs_remove_link_cb(zfs_handle_t *zhp, void *arg)
+{
+ struct destroydata *dd = arg;
+ zfs_handle_t *szhp;
+ char name[ZFS_MAXNAMELEN];
+ boolean_t closezhp = dd->closezhp;
+ int rv;
+
+ (void) strlcpy(name, zhp->zfs_name, sizeof (name));
+ (void) strlcat(name, "@", sizeof (name));
+ (void) strlcat(name, dd->snapname, sizeof (name));
+
+ szhp = make_dataset_handle(zhp->zfs_hdl, name);
+ if (szhp) {
+ dd->gotone = B_TRUE;
+ zfs_close(szhp);
+ }
+
+ if (zhp->zfs_type == ZFS_TYPE_VOLUME) {
+ (void) zvol_remove_link(zhp->zfs_hdl, name);
+ /*
+ * NB: this is simply a best-effort. We don't want to
+ * return an error, because then we wouldn't visit all
+ * the volumes.
+ */
+ }
+
+ dd->closezhp = B_TRUE;
+ rv = zfs_iter_filesystems(zhp, zfs_remove_link_cb, arg);
+ if (closezhp)
+ zfs_close(zhp);
+ return (rv);
+}
+
+/*
+ * Destroys all snapshots with the given name in zhp & descendants.
+ */
+int
+zfs_destroy_snaps(zfs_handle_t *zhp, char *snapname)
+{
+ zfs_cmd_t zc = { 0 };
+ int ret;
+ struct destroydata dd = { 0 };
+
+ dd.snapname = snapname;
+ (void) zfs_remove_link_cb(zhp, &dd);
+
+ if (!dd.gotone) {
+ return (zfs_standard_error_fmt(zhp->zfs_hdl, ENOENT,
+ dgettext(TEXT_DOMAIN, "cannot destroy '%s@%s'"),
+ zhp->zfs_name, snapname));
+ }
+
+ (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name));
+ (void) strlcpy(zc.zc_value, snapname, sizeof (zc.zc_value));
+
+ ret = ioctl(zhp->zfs_hdl->libzfs_fd, ZFS_IOC_DESTROY_SNAPS, &zc);
+ if (ret != 0) {
+ char errbuf[1024];
+
+ (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
+ "cannot destroy '%s@%s'"), zc.zc_name, snapname);
+
+ switch (errno) {
+ case EEXIST:
+ zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
+ "snapshot is cloned"));
+ return (zfs_error(zhp->zfs_hdl, EZFS_EXISTS, errbuf));
+
+ default:
+ return (zfs_standard_error(zhp->zfs_hdl, errno,
+ errbuf));
+ }
+ }
+
+ return (0);
+}
+
+/*
+ * Clones the given dataset. The target must be of the same type as the source.
+ */
+int
+zfs_clone(zfs_handle_t *zhp, const char *target, nvlist_t *props)
+{
+ zfs_cmd_t zc = { 0 };
+ char parent[ZFS_MAXNAMELEN];
+ int ret;
+ char errbuf[1024];
+ libzfs_handle_t *hdl = zhp->zfs_hdl;
+ zfs_type_t type;
+ uint64_t zoned;
+
+ assert(zhp->zfs_type == ZFS_TYPE_SNAPSHOT);
+
+ (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
+ "cannot create '%s'"), target);
+
+ /* validate the target name */
+ if (!zfs_validate_name(hdl, target, ZFS_TYPE_FILESYSTEM))
+ return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf));
+
+ /* validate parents exist */
+ if (check_parents(hdl, target, &zoned) != 0)
+ return (-1);
+
+ (void) parent_name(target, parent, sizeof (parent));
+
+ /* do the clone */
+ if (ZFS_IS_VOLUME(zhp)) {
+ zc.zc_objset_type = DMU_OST_ZVOL;
+ type = ZFS_TYPE_VOLUME;
+ } else {
+ zc.zc_objset_type = DMU_OST_ZFS;
+ type = ZFS_TYPE_FILESYSTEM;
+ }
+
+ if (props) {
+ if ((props = zfs_validate_properties(hdl, type, NULL, props,
+ zoned, zhp, errbuf)) == NULL)
+ return (-1);
+
+ if (zcmd_write_src_nvlist(hdl, &zc, props, NULL) != 0) {
+ nvlist_free(props);
+ return (-1);
+ }
+
+ nvlist_free(props);
+ }
+
+ (void) strlcpy(zc.zc_name, target, sizeof (zc.zc_name));
+ (void) strlcpy(zc.zc_value, zhp->zfs_name, sizeof (zc.zc_value));
+ ret = ioctl(zhp->zfs_hdl->libzfs_fd, ZFS_IOC_CREATE, &zc);
+
+ zcmd_free_nvlists(&zc);
+
+ if (ret != 0) {
+ switch (errno) {
+
+ case ENOENT:
+ /*
+ * The parent doesn't exist. We should have caught this
+ * above, but there may a race condition that has since
+ * destroyed the parent.
+ *
+ * At this point, we don't know whether it's the source
+ * that doesn't exist anymore, or whether the target
+ * dataset doesn't exist.
+ */
+ zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
+ "no such parent '%s'"), parent);
+ return (zfs_error(zhp->zfs_hdl, EZFS_NOENT, errbuf));
+
+ case EXDEV:
+ zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
+ "source and target pools differ"));
+ return (zfs_error(zhp->zfs_hdl, EZFS_CROSSTARGET,
+ errbuf));
+
+ default:
+ return (zfs_standard_error(zhp->zfs_hdl, errno,
+ errbuf));
+ }
+ } else if (ZFS_IS_VOLUME(zhp)) {
+ ret = zvol_create_link(zhp->zfs_hdl, target);
+ }
+
+ return (ret);
+}
+
+typedef struct promote_data {
+ char cb_mountpoint[MAXPATHLEN];
+ const char *cb_target;
+ const char *cb_errbuf;
+ uint64_t cb_pivot_txg;
+} promote_data_t;
+
+static int
+promote_snap_cb(zfs_handle_t *zhp, void *data)
+{
+ promote_data_t *pd = data;
+ zfs_handle_t *szhp;
+ char snapname[MAXPATHLEN];
+ int rv = 0;
+
+ /* We don't care about snapshots after the pivot point */
+ if (zfs_prop_get_int(zhp, ZFS_PROP_CREATETXG) > pd->cb_pivot_txg) {
+ zfs_close(zhp);
+ return (0);
+ }
+
+ /* Remove the device link if it's a zvol. */
+ if (ZFS_IS_VOLUME(zhp))
+ (void) zvol_remove_link(zhp->zfs_hdl, zhp->zfs_name);
+
+ /* Check for conflicting names */
+ (void) strlcpy(snapname, pd->cb_target, sizeof (snapname));
+ (void) strlcat(snapname, strchr(zhp->zfs_name, '@'), sizeof (snapname));
+ szhp = make_dataset_handle(zhp->zfs_hdl, snapname);
+ if (szhp != NULL) {
+ zfs_close(szhp);
+ zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
+ "snapshot name '%s' from origin \n"
+ "conflicts with '%s' from target"),
+ zhp->zfs_name, snapname);
+ rv = zfs_error(zhp->zfs_hdl, EZFS_EXISTS, pd->cb_errbuf);
+ }
+ zfs_close(zhp);
+ return (rv);
+}
+
+static int
+promote_snap_done_cb(zfs_handle_t *zhp, void *data)
+{
+ promote_data_t *pd = data;
+
+ /* We don't care about snapshots after the pivot point */
+ if (zfs_prop_get_int(zhp, ZFS_PROP_CREATETXG) <= pd->cb_pivot_txg) {
+ /* Create the device link if it's a zvol. */
+ if (ZFS_IS_VOLUME(zhp))
+ (void) zvol_create_link(zhp->zfs_hdl, zhp->zfs_name);
+ }
+
+ zfs_close(zhp);
+ return (0);
+}
+
+/*
+ * Promotes the given clone fs to be the clone parent.
+ */
+int
+zfs_promote(zfs_handle_t *zhp)
+{
+ libzfs_handle_t *hdl = zhp->zfs_hdl;
+ zfs_cmd_t zc = { 0 };
+ char parent[MAXPATHLEN];
+ char *cp;
+ int ret;
+ zfs_handle_t *pzhp;
+ promote_data_t pd;
+ char errbuf[1024];
+
+ (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
+ "cannot promote '%s'"), zhp->zfs_name);
+
+ if (zhp->zfs_type == ZFS_TYPE_SNAPSHOT) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "snapshots can not be promoted"));
+ return (zfs_error(hdl, EZFS_BADTYPE, errbuf));
+ }
+
+ (void) strlcpy(parent, zhp->zfs_dmustats.dds_clone_of, sizeof (parent));
+ if (parent[0] == '\0') {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "not a cloned filesystem"));
+ return (zfs_error(hdl, EZFS_BADTYPE, errbuf));
+ }
+ cp = strchr(parent, '@');
+ *cp = '\0';
+
+ /* Walk the snapshots we will be moving */
+ pzhp = zfs_open(hdl, zhp->zfs_dmustats.dds_clone_of, ZFS_TYPE_SNAPSHOT);
+ if (pzhp == NULL)
+ return (-1);
+ pd.cb_pivot_txg = zfs_prop_get_int(pzhp, ZFS_PROP_CREATETXG);
+ zfs_close(pzhp);
+ pd.cb_target = zhp->zfs_name;
+ pd.cb_errbuf = errbuf;
+ pzhp = zfs_open(hdl, parent, ZFS_TYPE_ANY);
+ if (pzhp == NULL)
+ return (-1);
+ (void) zfs_prop_get(pzhp, ZFS_PROP_MOUNTPOINT, pd.cb_mountpoint,
+ sizeof (pd.cb_mountpoint), NULL, NULL, 0, FALSE);
+ ret = zfs_iter_snapshots(pzhp, promote_snap_cb, &pd);
+ if (ret != 0) {
+ zfs_close(pzhp);
+ return (-1);
+ }
+
+ /* issue the ioctl */
+ (void) strlcpy(zc.zc_value, zhp->zfs_dmustats.dds_clone_of,
+ sizeof (zc.zc_value));
+ (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name));
+ ret = ioctl(hdl->libzfs_fd, ZFS_IOC_PROMOTE, &zc);
+
+ if (ret != 0) {
+ int save_errno = errno;
+
+ (void) zfs_iter_snapshots(pzhp, promote_snap_done_cb, &pd);
+ zfs_close(pzhp);
+
+ switch (save_errno) {
+ case EEXIST:
+ /*
+ * There is a conflicting snapshot name. We
+ * should have caught this above, but they could
+ * have renamed something in the mean time.
+ */
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "conflicting snapshot name from parent '%s'"),
+ parent);
+ return (zfs_error(hdl, EZFS_EXISTS, errbuf));
+
+ default:
+ return (zfs_standard_error(hdl, save_errno, errbuf));
+ }
+ } else {
+ (void) zfs_iter_snapshots(zhp, promote_snap_done_cb, &pd);
+ }
+
+ zfs_close(pzhp);
+ return (ret);
+}
+
+static int
+zfs_create_link_cb(zfs_handle_t *zhp, void *arg)
+{
+ char *snapname = arg;
+ int ret;
+
+ if (zhp->zfs_type == ZFS_TYPE_VOLUME) {
+ char name[MAXPATHLEN];
+
+ (void) strlcpy(name, zhp->zfs_name, sizeof (name));
+ (void) strlcat(name, "@", sizeof (name));
+ (void) strlcat(name, snapname, sizeof (name));
+ (void) zvol_create_link(zhp->zfs_hdl, name);
+ /*
+ * NB: this is simply a best-effort. We don't want to
+ * return an error, because then we wouldn't visit all
+ * the volumes.
+ */
+ }
+
+ ret = zfs_iter_filesystems(zhp, zfs_create_link_cb, snapname);
+
+ zfs_close(zhp);
+
+ return (ret);
+}
+
+/*
+ * Takes a snapshot of the given dataset.
+ */
+int
+zfs_snapshot(libzfs_handle_t *hdl, const char *path, boolean_t recursive)
+{
+ const char *delim;
+ char *parent;
+ zfs_handle_t *zhp;
+ zfs_cmd_t zc = { 0 };
+ int ret;
+ char errbuf[1024];
+
+ (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
+ "cannot snapshot '%s'"), path);
+
+ /* validate the target name */
+ if (!zfs_validate_name(hdl, path, ZFS_TYPE_SNAPSHOT))
+ return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf));
+
+ /* make sure the parent exists and is of the appropriate type */
+ delim = strchr(path, '@');
+ if ((parent = zfs_alloc(hdl, delim - path + 1)) == NULL)
+ return (-1);
+ (void) strncpy(parent, path, delim - path);
+ parent[delim - path] = '\0';
+
+ if ((zhp = zfs_open(hdl, parent, ZFS_TYPE_FILESYSTEM |
+ ZFS_TYPE_VOLUME)) == NULL) {
+ free(parent);
+ return (-1);
+ }
+
+ (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name));
+ (void) strlcpy(zc.zc_value, delim+1, sizeof (zc.zc_value));
+ zc.zc_cookie = recursive;
+ ret = ioctl(zhp->zfs_hdl->libzfs_fd, ZFS_IOC_SNAPSHOT, &zc);
+
+ /*
+ * if it was recursive, the one that actually failed will be in
+ * zc.zc_name.
+ */
+ (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
+ "cannot create snapshot '%s@%s'"), zc.zc_name, zc.zc_value);
+ if (ret == 0 && recursive) {
+ (void) zfs_iter_filesystems(zhp,
+ zfs_create_link_cb, (char *)delim+1);
+ }
+ if (ret == 0 && zhp->zfs_type == ZFS_TYPE_VOLUME) {
+ ret = zvol_create_link(zhp->zfs_hdl, path);
+ if (ret != 0) {
+ (void) ioctl(zhp->zfs_hdl->libzfs_fd, ZFS_IOC_DESTROY,
+ &zc);
+ }
+ }
+
+ if (ret != 0)
+ (void) zfs_standard_error(hdl, errno, errbuf);
+
+ free(parent);
+ zfs_close(zhp);
+
+ return (ret);
+}
+
+/*
+ * Dumps a backup of the given snapshot (incremental from fromsnap if it's not
+ * NULL) to the file descriptor specified by outfd.
+ */
+int
+zfs_send(zfs_handle_t *zhp, const char *fromsnap, int outfd)
+{
+ zfs_cmd_t zc = { 0 };
+ char errbuf[1024];
+ libzfs_handle_t *hdl = zhp->zfs_hdl;
+
+ assert(zhp->zfs_type == ZFS_TYPE_SNAPSHOT);
+
+ (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name));
+ if (fromsnap)
+ (void) strlcpy(zc.zc_value, fromsnap, sizeof (zc.zc_name));
+ zc.zc_cookie = outfd;
+
+ if (ioctl(zhp->zfs_hdl->libzfs_fd, ZFS_IOC_SENDBACKUP, &zc) != 0) {
+ (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
+ "cannot send '%s'"), zhp->zfs_name);
+
+ switch (errno) {
+
+ case EXDEV:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "not an earlier snapshot from the same fs"));
+ return (zfs_error(hdl, EZFS_CROSSTARGET, errbuf));
+
+ case EDQUOT:
+ case EFBIG:
+ case EIO:
+ case ENOLINK:
+ case ENOSPC:
+ case ENXIO:
+ case EPIPE:
+ case ERANGE:
+ case EFAULT:
+ case EROFS:
+ zfs_error_aux(hdl, strerror(errno));
+ return (zfs_error(hdl, EZFS_BADBACKUP, errbuf));
+
+ default:
+ return (zfs_standard_error(hdl, errno, errbuf));
+ }
+ }
+
+ return (0);
+}
+
+/*
+ * Create ancestors of 'target', but not target itself, and not
+ * ancestors whose names are shorter than prefixlen. Die if
+ * prefixlen-ancestor does not exist.
+ */
+static int
+create_parents(libzfs_handle_t *hdl, char *target, int prefixlen)
+{
+ zfs_handle_t *h;
+ char *cp;
+
+ /* make sure prefix exists */
+ cp = strchr(target + prefixlen, '/');
+ *cp = '\0';
+ h = zfs_open(hdl, target, ZFS_TYPE_FILESYSTEM);
+ *cp = '/';
+ if (h == NULL)
+ return (-1);
+ zfs_close(h);
+
+ /*
+ * Attempt to create, mount, and share any ancestor filesystems,
+ * up to the prefixlen-long one.
+ */
+ for (cp = target + prefixlen + 1;
+ cp = strchr(cp, '/'); *cp = '/', cp++) {
+ const char *opname;
+
+ *cp = '\0';
+
+ h = make_dataset_handle(hdl, target);
+ if (h) {
+ /* it already exists, nothing to do here */
+ zfs_close(h);
+ continue;
+ }
+
+ opname = dgettext(TEXT_DOMAIN, "create");
+ if (zfs_create(hdl, target, ZFS_TYPE_FILESYSTEM,
+ NULL) != 0)
+ goto ancestorerr;
+
+ opname = dgettext(TEXT_DOMAIN, "open");
+ h = zfs_open(hdl, target, ZFS_TYPE_FILESYSTEM);
+ if (h == NULL)
+ goto ancestorerr;
+
+ opname = dgettext(TEXT_DOMAIN, "mount");
+ if (zfs_mount(h, NULL, 0) != 0)
+ goto ancestorerr;
+
+ opname = dgettext(TEXT_DOMAIN, "share");
+ if (zfs_share(h) != 0)
+ goto ancestorerr;
+
+ zfs_close(h);
+
+ continue;
+ancestorerr:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "failed to %s ancestor '%s'"), opname, target);
+ return (-1);
+ }
+
+ return (0);
+}
+
+/*
+ * Restores a backup of tosnap from the file descriptor specified by infd.
+ */
+int
+zfs_receive(libzfs_handle_t *hdl, const char *tosnap, int isprefix,
+ int verbose, int dryrun, boolean_t force, int infd)
+{
+ zfs_cmd_t zc = { 0 };
+ time_t begin_time;
+ int ioctl_err, err, bytes, size, choplen;
+ char *cp;
+ dmu_replay_record_t drr;
+ struct drr_begin *drrb = &zc.zc_begin_record;
+ char errbuf[1024];
+ prop_changelist_t *clp;
+ char chopprefix[ZFS_MAXNAMELEN];
+
+ begin_time = time(NULL);
+
+ (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
+ "cannot receive"));
+
+ /* read in the BEGIN record */
+ cp = (char *)&drr;
+ bytes = 0;
+ do {
+ size = read(infd, cp, sizeof (drr) - bytes);
+ cp += size;
+ bytes += size;
+ } while (size > 0);
+
+ if (size < 0 || bytes != sizeof (drr)) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid "
+ "stream (failed to read first record)"));
+ return (zfs_error(hdl, EZFS_BADSTREAM, errbuf));
+ }
+
+ zc.zc_begin_record = drr.drr_u.drr_begin;
+
+ if (drrb->drr_magic != DMU_BACKUP_MAGIC &&
+ drrb->drr_magic != BSWAP_64(DMU_BACKUP_MAGIC)) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid "
+ "stream (bad magic number)"));
+ return (zfs_error(hdl, EZFS_BADSTREAM, errbuf));
+ }
+
+ if (drrb->drr_version != DMU_BACKUP_VERSION &&
+ drrb->drr_version != BSWAP_64(DMU_BACKUP_VERSION)) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "only version "
+ "0x%llx is supported (stream is version 0x%llx)"),
+ DMU_BACKUP_VERSION, drrb->drr_version);
+ return (zfs_error(hdl, EZFS_BADSTREAM, errbuf));
+ }
+
+ if (strchr(drr.drr_u.drr_begin.drr_toname, '@') == NULL) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid "
+ "stream (bad snapshot name)"));
+ return (zfs_error(hdl, EZFS_BADSTREAM, errbuf));
+ }
+ /*
+ * Determine how much of the snapshot name stored in the stream
+ * we are going to tack on to the name they specified on the
+ * command line, and how much we are going to chop off.
+ *
+ * If they specified a snapshot, chop the entire name stored in
+ * the stream.
+ */
+ (void) strcpy(chopprefix, drr.drr_u.drr_begin.drr_toname);
+ if (isprefix) {
+ /*
+ * They specified a fs with -d, we want to tack on
+ * everything but the pool name stored in the stream
+ */
+ if (strchr(tosnap, '@')) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid "
+ "argument - snapshot not allowed with -d"));
+ return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf));
+ }
+ cp = strchr(chopprefix, '/');
+ if (cp == NULL)
+ cp = strchr(chopprefix, '@');
+ *cp = '\0';
+ } else if (strchr(tosnap, '@') == NULL) {
+ /*
+ * If they specified a filesystem without -d, we want to
+ * tack on everything after the fs specified in the
+ * first name from the stream.
+ */
+ cp = strchr(chopprefix, '@');
+ *cp = '\0';
+ }
+ choplen = strlen(chopprefix);
+
+ /*
+ * Determine name of destination snapshot, store in zc_value.
+ */
+ (void) strcpy(zc.zc_value, tosnap);
+ (void) strncat(zc.zc_value, drr.drr_u.drr_begin.drr_toname+choplen,
+ sizeof (zc.zc_value));
+ if (!zfs_validate_name(hdl, zc.zc_value, ZFS_TYPE_SNAPSHOT))
+ return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf));
+
+ (void) strcpy(zc.zc_name, zc.zc_value);
+ if (drrb->drr_fromguid) {
+ /* incremental backup stream */
+ zfs_handle_t *h;
+
+ /* do the recvbackup ioctl to the containing fs */
+ *strchr(zc.zc_name, '@') = '\0';
+
+ /* make sure destination fs exists */
+ h = zfs_open(hdl, zc.zc_name,
+ ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME);
+ if (h == NULL)
+ return (-1);
+ if (!dryrun) {
+ /*
+ * We need to unmount all the dependents of the dataset
+ * and the dataset itself. If it's a volume
+ * then remove device link.
+ */
+ if (h->zfs_type == ZFS_TYPE_FILESYSTEM) {
+ clp = changelist_gather(h, ZFS_PROP_NAME, 0);
+ if (clp == NULL)
+ return (-1);
+ if (changelist_prefix(clp) != 0) {
+ changelist_free(clp);
+ return (-1);
+ }
+ } else {
+ (void) zvol_remove_link(hdl, h->zfs_name);
+ }
+ }
+ zfs_close(h);
+ } else {
+ /* full backup stream */
+
+ /* Make sure destination fs does not exist */
+ *strchr(zc.zc_name, '@') = '\0';
+ if (ioctl(hdl->libzfs_fd, ZFS_IOC_OBJSET_STATS, &zc) == 0) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "destination '%s' exists"), zc.zc_name);
+ return (zfs_error(hdl, EZFS_EXISTS, errbuf));
+ }
+
+ if (strchr(zc.zc_name, '/') == NULL) {
+ /*
+ * they're trying to do a recv into a
+ * nonexistant topmost filesystem.
+ */
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "destination does not exist"), zc.zc_name);
+ return (zfs_error(hdl, EZFS_EXISTS, errbuf));
+ }
+
+ /* Do the recvbackup ioctl to the fs's parent. */
+ *strrchr(zc.zc_name, '/') = '\0';
+
+ if (isprefix && (err = create_parents(hdl,
+ zc.zc_value, strlen(tosnap))) != 0) {
+ return (zfs_error(hdl, EZFS_BADRESTORE, errbuf));
+ }
+
+ }
+
+ zc.zc_cookie = infd;
+ zc.zc_guid = force;
+ if (verbose) {
+ (void) printf("%s %s stream of %s into %s\n",
+ dryrun ? "would receive" : "receiving",
+ drrb->drr_fromguid ? "incremental" : "full",
+ drr.drr_u.drr_begin.drr_toname,
+ zc.zc_value);
+ (void) fflush(stdout);
+ }
+ if (dryrun)
+ return (0);
+ err = ioctl_err = ioctl(hdl->libzfs_fd, ZFS_IOC_RECVBACKUP, &zc);
+ if (ioctl_err != 0) {
+ switch (errno) {
+ case ENODEV:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "most recent snapshot does not match incremental "
+ "source"));
+ (void) zfs_error(hdl, EZFS_BADRESTORE, errbuf);
+ break;
+ case ETXTBSY:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "destination has been modified since most recent "
+ "snapshot"));
+ (void) zfs_error(hdl, EZFS_BADRESTORE, errbuf);
+ break;
+ case EEXIST:
+ if (drrb->drr_fromguid == 0) {
+ /* it's the containing fs that exists */
+ cp = strchr(zc.zc_value, '@');
+ *cp = '\0';
+ }
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "destination already exists"));
+ (void) zfs_error_fmt(hdl, EZFS_EXISTS,
+ dgettext(TEXT_DOMAIN, "cannot restore to %s"),
+ zc.zc_value);
+ break;
+ case EINVAL:
+ (void) zfs_error(hdl, EZFS_BADSTREAM, errbuf);
+ break;
+ case ECKSUM:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "invalid stream (checksum mismatch)"));
+ (void) zfs_error(hdl, EZFS_BADSTREAM, errbuf);
+ break;
+ default:
+ (void) zfs_standard_error(hdl, errno, errbuf);
+ }
+ }
+
+ /*
+ * Mount or recreate the /dev links for the target filesystem
+ * (if created, or if we tore them down to do an incremental
+ * restore), and the /dev links for the new snapshot (if
+ * created). Also mount any children of the target filesystem
+ * if we did an incremental receive.
+ */
+ cp = strchr(zc.zc_value, '@');
+ if (cp && (ioctl_err == 0 || drrb->drr_fromguid)) {
+ zfs_handle_t *h;
+
+ *cp = '\0';
+ h = zfs_open(hdl, zc.zc_value,
+ ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME);
+ *cp = '@';
+ if (h) {
+ if (h->zfs_type == ZFS_TYPE_VOLUME) {
+ err = zvol_create_link(hdl, h->zfs_name);
+ if (err == 0 && ioctl_err == 0)
+ err = zvol_create_link(hdl,
+ zc.zc_value);
+ } else {
+ if (drrb->drr_fromguid) {
+ err = changelist_postfix(clp);
+ changelist_free(clp);
+ } else {
+ err = zfs_mount(h, NULL, 0);
+ }
+ }
+ zfs_close(h);
+ }
+ }
+
+ if (err || ioctl_err)
+ return (-1);
+
+ if (verbose) {
+ char buf1[64];
+ char buf2[64];
+ uint64_t bytes = zc.zc_cookie;
+ time_t delta = time(NULL) - begin_time;
+ if (delta == 0)
+ delta = 1;
+ zfs_nicenum(bytes, buf1, sizeof (buf1));
+ zfs_nicenum(bytes/delta, buf2, sizeof (buf1));
+
+ (void) printf("received %sb stream in %lu seconds (%sb/sec)\n",
+ buf1, delta, buf2);
+ }
+
+ return (0);
+}
+
+/*
+ * Destroy any more recent snapshots. We invoke this callback on any dependents
+ * of the snapshot first. If the 'cb_dependent' member is non-zero, then this
+ * is a dependent and we should just destroy it without checking the transaction
+ * group.
+ */
+typedef struct rollback_data {
+ const char *cb_target; /* the snapshot */
+ uint64_t cb_create; /* creation time reference */
+ prop_changelist_t *cb_clp; /* changelist pointer */
+ int cb_error;
+ boolean_t cb_dependent;
+} rollback_data_t;
+
+static int
+rollback_destroy(zfs_handle_t *zhp, void *data)
+{
+ rollback_data_t *cbp = data;
+
+ if (!cbp->cb_dependent) {
+ if (strcmp(zhp->zfs_name, cbp->cb_target) != 0 &&
+ zfs_get_type(zhp) == ZFS_TYPE_SNAPSHOT &&
+ zfs_prop_get_int(zhp, ZFS_PROP_CREATETXG) >
+ cbp->cb_create) {
+
+ cbp->cb_dependent = B_TRUE;
+ if (zfs_iter_dependents(zhp, B_FALSE, rollback_destroy,
+ cbp) != 0)
+ cbp->cb_error = 1;
+ cbp->cb_dependent = B_FALSE;
+
+ if (zfs_destroy(zhp) != 0)
+ cbp->cb_error = 1;
+ else
+ changelist_remove(zhp, cbp->cb_clp);
+ }
+ } else {
+ if (zfs_destroy(zhp) != 0)
+ cbp->cb_error = 1;
+ else
+ changelist_remove(zhp, cbp->cb_clp);
+ }
+
+ zfs_close(zhp);
+ return (0);
+}
+
+/*
+ * Rollback the dataset to its latest snapshot.
+ */
+static int
+do_rollback(zfs_handle_t *zhp)
+{
+ int ret;
+ zfs_cmd_t zc = { 0 };
+
+ assert(zhp->zfs_type == ZFS_TYPE_FILESYSTEM ||
+ zhp->zfs_type == ZFS_TYPE_VOLUME);
+
+ if (zhp->zfs_type == ZFS_TYPE_VOLUME &&
+ zvol_remove_link(zhp->zfs_hdl, zhp->zfs_name) != 0)
+ return (-1);
+
+ (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name));
+
+ if (ZFS_IS_VOLUME(zhp))
+ zc.zc_objset_type = DMU_OST_ZVOL;
+ else
+ zc.zc_objset_type = DMU_OST_ZFS;
+
+ /*
+ * We rely on the consumer to verify that there are no newer snapshots
+ * for the given dataset. Given these constraints, we can simply pass
+ * the name on to the ioctl() call. There is still an unlikely race
+ * condition where the user has taken a snapshot since we verified that
+ * this was the most recent.
+ */
+ if ((ret = ioctl(zhp->zfs_hdl->libzfs_fd, ZFS_IOC_ROLLBACK,
+ &zc)) != 0) {
+ (void) zfs_standard_error_fmt(zhp->zfs_hdl, errno,
+ dgettext(TEXT_DOMAIN, "cannot rollback '%s'"),
+ zhp->zfs_name);
+ } else if (zhp->zfs_type == ZFS_TYPE_VOLUME) {
+ ret = zvol_create_link(zhp->zfs_hdl, zhp->zfs_name);
+ }
+
+ return (ret);
+}
+
+/*
+ * Given a dataset, rollback to a specific snapshot, discarding any
+ * data changes since then and making it the active dataset.
+ *
+ * Any snapshots more recent than the target are destroyed, along with
+ * their dependents.
+ */
+int
+zfs_rollback(zfs_handle_t *zhp, zfs_handle_t *snap, int flag)
+{
+ int ret;
+ rollback_data_t cb = { 0 };
+ prop_changelist_t *clp;
+
+ /*
+ * Unmount all dependendents of the dataset and the dataset itself.
+ * The list we need to gather is the same as for doing rename
+ */
+ clp = changelist_gather(zhp, ZFS_PROP_NAME, flag ? MS_FORCE: 0);
+ if (clp == NULL)
+ return (-1);
+
+ if ((ret = changelist_prefix(clp)) != 0)
+ goto out;
+
+ /*
+ * Destroy all recent snapshots and its dependends.
+ */
+ cb.cb_target = snap->zfs_name;
+ cb.cb_create = zfs_prop_get_int(snap, ZFS_PROP_CREATETXG);
+ cb.cb_clp = clp;
+ (void) zfs_iter_children(zhp, rollback_destroy, &cb);
+
+ if ((ret = cb.cb_error) != 0) {
+ (void) changelist_postfix(clp);
+ goto out;
+ }
+
+ /*
+ * Now that we have verified that the snapshot is the latest,
+ * rollback to the given snapshot.
+ */
+ ret = do_rollback(zhp);
+
+ if (ret != 0) {
+ (void) changelist_postfix(clp);
+ goto out;
+ }
+
+ /*
+ * We only want to re-mount the filesystem if it was mounted in the
+ * first place.
+ */
+ ret = changelist_postfix(clp);
+
+out:
+ changelist_free(clp);
+ return (ret);
+}
+
+/*
+ * Iterate over all dependents for a given dataset. This includes both
+ * hierarchical dependents (children) and data dependents (snapshots and
+ * clones). The bulk of the processing occurs in get_dependents() in
+ * libzfs_graph.c.
+ */
+int
+zfs_iter_dependents(zfs_handle_t *zhp, boolean_t allowrecursion,
+ zfs_iter_f func, void *data)
+{
+ char **dependents;
+ size_t count;
+ int i;
+ zfs_handle_t *child;
+ int ret = 0;
+
+ if (get_dependents(zhp->zfs_hdl, allowrecursion, zhp->zfs_name,
+ &dependents, &count) != 0)
+ return (-1);
+
+ for (i = 0; i < count; i++) {
+ if ((child = make_dataset_handle(zhp->zfs_hdl,
+ dependents[i])) == NULL)
+ continue;
+
+ if ((ret = func(child, data)) != 0)
+ break;
+ }
+
+ for (i = 0; i < count; i++)
+ free(dependents[i]);
+ free(dependents);
+
+ return (ret);
+}
+
+/*
+ * Renames the given dataset.
+ */
+int
+zfs_rename(zfs_handle_t *zhp, const char *target)
+{
+ int ret;
+ zfs_cmd_t zc = { 0 };
+ char *delim;
+ prop_changelist_t *cl;
+ char parent[ZFS_MAXNAMELEN];
+ libzfs_handle_t *hdl = zhp->zfs_hdl;
+ char errbuf[1024];
+
+ /* if we have the same exact name, just return success */
+ if (strcmp(zhp->zfs_name, target) == 0)
+ return (0);
+
+ (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
+ "cannot rename to '%s'"), target);
+
+ /*
+ * Make sure the target name is valid
+ */
+ if (zhp->zfs_type == ZFS_TYPE_SNAPSHOT) {
+ if ((strchr(target, '@') == NULL) ||
+ *target == '@') {
+ /*
+ * Snapshot target name is abbreviated,
+ * reconstruct full dataset name
+ */
+ (void) strlcpy(parent, zhp->zfs_name,
+ sizeof (parent));
+ delim = strchr(parent, '@');
+ if (strchr(target, '@') == NULL)
+ *(++delim) = '\0';
+ else
+ *delim = '\0';
+ (void) strlcat(parent, target, sizeof (parent));
+ target = parent;
+ } else {
+ /*
+ * Make sure we're renaming within the same dataset.
+ */
+ delim = strchr(target, '@');
+ if (strncmp(zhp->zfs_name, target, delim - target)
+ != 0 || zhp->zfs_name[delim - target] != '@') {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "snapshots must be part of same "
+ "dataset"));
+ return (zfs_error(hdl, EZFS_CROSSTARGET,
+ errbuf));
+ }
+ }
+ if (!zfs_validate_name(hdl, target, zhp->zfs_type))
+ return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf));
+ } else {
+ if (!zfs_validate_name(hdl, target, zhp->zfs_type))
+ return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf));
+ uint64_t unused;
+
+ /* validate parents */
+ if (check_parents(hdl, target, &unused) != 0)
+ return (-1);
+
+ (void) parent_name(target, parent, sizeof (parent));
+
+ /* make sure we're in the same pool */
+ verify((delim = strchr(target, '/')) != NULL);
+ if (strncmp(zhp->zfs_name, target, delim - target) != 0 ||
+ zhp->zfs_name[delim - target] != '/') {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "datasets must be within same pool"));
+ return (zfs_error(hdl, EZFS_CROSSTARGET, errbuf));
+ }
+
+ /* new name cannot be a child of the current dataset name */
+ if (strncmp(parent, zhp->zfs_name,
+ strlen(zhp->zfs_name)) == 0) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "New dataset name cannot be a descendent of "
+ "current dataset name"));
+ return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf));
+ }
+ }
+
+ (void) snprintf(errbuf, sizeof (errbuf),
+ dgettext(TEXT_DOMAIN, "cannot rename '%s'"), zhp->zfs_name);
+
+ if (getzoneid() == GLOBAL_ZONEID &&
+ zfs_prop_get_int(zhp, ZFS_PROP_ZONED)) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "dataset is used in a non-global zone"));
+ return (zfs_error(hdl, EZFS_ZONED, errbuf));
+ }
+
+ if ((cl = changelist_gather(zhp, ZFS_PROP_NAME, 0)) == NULL)
+ return (-1);
+
+ if (changelist_haszonedchild(cl)) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "child dataset with inherited mountpoint is used "
+ "in a non-global zone"));
+ (void) zfs_error(hdl, EZFS_ZONED, errbuf);
+ goto error;
+ }
+
+ if ((ret = changelist_prefix(cl)) != 0)
+ goto error;
+
+ if (ZFS_IS_VOLUME(zhp))
+ zc.zc_objset_type = DMU_OST_ZVOL;
+ else
+ zc.zc_objset_type = DMU_OST_ZFS;
+
+ (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name));
+ (void) strlcpy(zc.zc_value, target, sizeof (zc.zc_value));
+
+ if ((ret = ioctl(zhp->zfs_hdl->libzfs_fd, ZFS_IOC_RENAME, &zc)) != 0) {
+ (void) zfs_standard_error(zhp->zfs_hdl, errno, errbuf);
+
+ /*
+ * On failure, we still want to remount any filesystems that
+ * were previously mounted, so we don't alter the system state.
+ */
+ (void) changelist_postfix(cl);
+ } else {
+ changelist_rename(cl, zfs_get_name(zhp), target);
+
+ ret = changelist_postfix(cl);
+ }
+
+error:
+ changelist_free(cl);
+ return (ret);
+}
+
+/*
+ * Given a zvol dataset, issue the ioctl to create the appropriate minor node,
+ * poke devfsadm to create the /dev link, and then wait for the link to appear.
+ */
+int
+zvol_create_link(libzfs_handle_t *hdl, const char *dataset)
+{
+ zfs_cmd_t zc = { 0 };
+#if 0
+ di_devlink_handle_t dhdl;
+#endif
+
+ (void) strlcpy(zc.zc_name, dataset, sizeof (zc.zc_name));
+
+ /*
+ * Issue the appropriate ioctl.
+ */
+ if (ioctl(hdl->libzfs_fd, ZFS_IOC_CREATE_MINOR, &zc) != 0) {
+ switch (errno) {
+ case EEXIST:
+ /*
+ * Silently ignore the case where the link already
+ * exists. This allows 'zfs volinit' to be run multiple
+ * times without errors.
+ */
+ return (0);
+
+ default:
+ return (zfs_standard_error_fmt(hdl, errno,
+ dgettext(TEXT_DOMAIN, "cannot create device links "
+ "for '%s'"), dataset));
+ }
+ }
+
+#if 0
+ /*
+ * Call devfsadm and wait for the links to magically appear.
+ */
+ if ((dhdl = di_devlink_init(ZFS_DRIVER, DI_MAKE_LINK)) == NULL) {
+ zfs_error_aux(hdl, strerror(errno));
+ (void) zfs_error_fmt(hdl, EZFS_DEVLINKS,
+ dgettext(TEXT_DOMAIN, "cannot create device links "
+ "for '%s'"), dataset);
+ (void) ioctl(hdl->libzfs_fd, ZFS_IOC_REMOVE_MINOR, &zc);
+ return (-1);
+ } else {
+ (void) di_devlink_fini(&dhdl);
+ }
+#endif
+
+ return (0);
+}
+
+/*
+ * Remove a minor node for the given zvol and the associated /dev links.
+ */
+int
+zvol_remove_link(libzfs_handle_t *hdl, const char *dataset)
+{
+ zfs_cmd_t zc = { 0 };
+
+ (void) strlcpy(zc.zc_name, dataset, sizeof (zc.zc_name));
+
+ if (ioctl(hdl->libzfs_fd, ZFS_IOC_REMOVE_MINOR, &zc) != 0) {
+ switch (errno) {
+ case ENXIO:
+ /*
+ * Silently ignore the case where the link no longer
+ * exists, so that 'zfs volfini' can be run multiple
+ * times without errors.
+ */
+ return (0);
+
+ default:
+ return (zfs_standard_error_fmt(hdl, errno,
+ dgettext(TEXT_DOMAIN, "cannot remove device "
+ "links for '%s'"), dataset));
+ }
+ }
+
+ return (0);
+}
+
+nvlist_t *
+zfs_get_user_props(zfs_handle_t *zhp)
+{
+ return (zhp->zfs_user_props);
+}
+
+/*
+ * Given a comma-separated list of properties, contruct a property list
+ * containing both user-defined and native properties. This function will
+ * return a NULL list if 'all' is specified, which can later be expanded on a
+ * per-dataset basis by zfs_expand_proplist().
+ */
+int
+zfs_get_proplist_common(libzfs_handle_t *hdl, char *fields,
+ zfs_proplist_t **listp, zfs_type_t type)
+{
+ size_t len;
+ char *s, *p;
+ char c;
+ zfs_prop_t prop;
+ zfs_proplist_t *entry;
+ zfs_proplist_t **last;
+
+ *listp = NULL;
+ last = listp;
+
+ /*
+ * If 'all' is specified, return a NULL list.
+ */
+ if (strcmp(fields, "all") == 0)
+ return (0);
+
+ /*
+ * If no fields were specified, return an error.
+ */
+ if (fields[0] == '\0') {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "no properties specified"));
+ return (zfs_error(hdl, EZFS_BADPROP, dgettext(TEXT_DOMAIN,
+ "bad property list")));
+ }
+
+ /*
+ * It would be nice to use getsubopt() here, but the inclusion of column
+ * aliases makes this more effort than it's worth.
+ */
+ s = fields;
+ while (*s != '\0') {
+ if ((p = strchr(s, ',')) == NULL) {
+ len = strlen(s);
+ p = s + len;
+ } else {
+ len = p - s;
+ }
+
+ /*
+ * Check for empty options.
+ */
+ if (len == 0) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "empty property name"));
+ return (zfs_error(hdl, EZFS_BADPROP,
+ dgettext(TEXT_DOMAIN, "bad property list")));
+ }
+
+ /*
+ * Check all regular property names.
+ */
+ c = s[len];
+ s[len] = '\0';
+ prop = zfs_name_to_prop_common(s, type);
+
+ if (prop != ZFS_PROP_INVAL &&
+ !zfs_prop_valid_for_type(prop, type))
+ prop = ZFS_PROP_INVAL;
+
+ /*
+ * When no property table entry can be found, return failure if
+ * this is a pool property or if this isn't a user-defined
+ * dataset property,
+ */
+ if (prop == ZFS_PROP_INVAL &&
+ (type & ZFS_TYPE_POOL || !zfs_prop_user(s))) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "invalid property '%s'"), s);
+ return (zfs_error(hdl, EZFS_BADPROP,
+ dgettext(TEXT_DOMAIN, "bad property list")));
+ }
+
+ if ((entry = zfs_alloc(hdl, sizeof (zfs_proplist_t))) == NULL)
+ return (-1);
+
+ entry->pl_prop = prop;
+ if (prop == ZFS_PROP_INVAL) {
+ if ((entry->pl_user_prop =
+ zfs_strdup(hdl, s)) == NULL) {
+ free(entry);
+ return (-1);
+ }
+ entry->pl_width = strlen(s);
+ } else {
+ entry->pl_width = zfs_prop_width(prop,
+ &entry->pl_fixed);
+ }
+
+ *last = entry;
+ last = &entry->pl_next;
+
+ s = p;
+ if (c == ',')
+ s++;
+ }
+
+ return (0);
+}
+
+int
+zfs_get_proplist(libzfs_handle_t *hdl, char *fields, zfs_proplist_t **listp)
+{
+ return (zfs_get_proplist_common(hdl, fields, listp, ZFS_TYPE_ANY));
+}
+
+void
+zfs_free_proplist(zfs_proplist_t *pl)
+{
+ zfs_proplist_t *next;
+
+ while (pl != NULL) {
+ next = pl->pl_next;
+ free(pl->pl_user_prop);
+ free(pl);
+ pl = next;
+ }
+}
+
+typedef struct expand_data {
+ zfs_proplist_t **last;
+ libzfs_handle_t *hdl;
+} expand_data_t;
+
+static zfs_prop_t
+zfs_expand_proplist_cb(zfs_prop_t prop, void *cb)
+{
+ zfs_proplist_t *entry;
+ expand_data_t *edp = cb;
+
+ if ((entry = zfs_alloc(edp->hdl, sizeof (zfs_proplist_t))) == NULL)
+ return (ZFS_PROP_INVAL);
+
+ entry->pl_prop = prop;
+ entry->pl_width = zfs_prop_width(prop, &entry->pl_fixed);
+ entry->pl_all = B_TRUE;
+
+ *(edp->last) = entry;
+ edp->last = &entry->pl_next;
+
+ return (ZFS_PROP_CONT);
+}
+
+int
+zfs_expand_proplist_common(libzfs_handle_t *hdl, zfs_proplist_t **plp,
+ zfs_type_t type)
+{
+ zfs_proplist_t *entry;
+ zfs_proplist_t **last;
+ expand_data_t exp;
+
+ if (*plp == NULL) {
+ /*
+ * If this is the very first time we've been called for an 'all'
+ * specification, expand the list to include all native
+ * properties.
+ */
+ last = plp;
+
+ exp.last = last;
+ exp.hdl = hdl;
+
+ if (zfs_prop_iter_common(zfs_expand_proplist_cb, &exp, type,
+ B_FALSE) == ZFS_PROP_INVAL)
+ return (-1);
+
+ /*
+ * Add 'name' to the beginning of the list, which is handled
+ * specially.
+ */
+ if ((entry = zfs_alloc(hdl,
+ sizeof (zfs_proplist_t))) == NULL)
+ return (-1);
+
+ entry->pl_prop = ZFS_PROP_NAME;
+ entry->pl_width = zfs_prop_width(ZFS_PROP_NAME,
+ &entry->pl_fixed);
+ entry->pl_all = B_TRUE;
+ entry->pl_next = *plp;
+ *plp = entry;
+ }
+ return (0);
+}
+
+/*
+ * This function is used by 'zfs list' to determine the exact set of columns to
+ * display, and their maximum widths. This does two main things:
+ *
+ * - If this is a list of all properties, then expand the list to include
+ * all native properties, and set a flag so that for each dataset we look
+ * for new unique user properties and add them to the list.
+ *
+ * - For non fixed-width properties, keep track of the maximum width seen
+ * so that we can size the column appropriately.
+ */
+int
+zfs_expand_proplist(zfs_handle_t *zhp, zfs_proplist_t **plp)
+{
+ libzfs_handle_t *hdl = zhp->zfs_hdl;
+ zfs_proplist_t *entry;
+ zfs_proplist_t **last, **start;
+ nvlist_t *userprops, *propval;
+ nvpair_t *elem;
+ char *strval;
+ char buf[ZFS_MAXPROPLEN];
+
+ if (zfs_expand_proplist_common(hdl, plp, ZFS_TYPE_ANY) != 0)
+ return (-1);
+
+ userprops = zfs_get_user_props(zhp);
+
+ entry = *plp;
+ if (entry->pl_all && nvlist_next_nvpair(userprops, NULL) != NULL) {
+ /*
+ * Go through and add any user properties as necessary. We
+ * start by incrementing our list pointer to the first
+ * non-native property.
+ */
+ start = plp;
+ while (*start != NULL) {
+ if ((*start)->pl_prop == ZFS_PROP_INVAL)
+ break;
+ start = &(*start)->pl_next;
+ }
+
+ elem = NULL;
+ while ((elem = nvlist_next_nvpair(userprops, elem)) != NULL) {
+ /*
+ * See if we've already found this property in our list.
+ */
+ for (last = start; *last != NULL;
+ last = &(*last)->pl_next) {
+ if (strcmp((*last)->pl_user_prop,
+ nvpair_name(elem)) == 0)
+ break;
+ }
+
+ if (*last == NULL) {
+ if ((entry = zfs_alloc(hdl,
+ sizeof (zfs_proplist_t))) == NULL ||
+ ((entry->pl_user_prop = zfs_strdup(hdl,
+ nvpair_name(elem)))) == NULL) {
+ free(entry);
+ return (-1);
+ }
+
+ entry->pl_prop = ZFS_PROP_INVAL;
+ entry->pl_width = strlen(nvpair_name(elem));
+ entry->pl_all = B_TRUE;
+ *last = entry;
+ }
+ }
+ }
+
+ /*
+ * Now go through and check the width of any non-fixed columns
+ */
+ for (entry = *plp; entry != NULL; entry = entry->pl_next) {
+ if (entry->pl_fixed)
+ continue;
+
+ if (entry->pl_prop != ZFS_PROP_INVAL) {
+ if (zfs_prop_get(zhp, entry->pl_prop,
+ buf, sizeof (buf), NULL, NULL, 0, B_FALSE) == 0) {
+ if (strlen(buf) > entry->pl_width)
+ entry->pl_width = strlen(buf);
+ }
+ } else if (nvlist_lookup_nvlist(userprops,
+ entry->pl_user_prop, &propval) == 0) {
+ verify(nvlist_lookup_string(propval,
+ ZFS_PROP_VALUE, &strval) == 0);
+ if (strlen(strval) > entry->pl_width)
+ entry->pl_width = strlen(strval);
+ }
+ }
+
+ return (0);
+}
+
+/*
+ * Attach/detach the given filesystem to/from the given jail.
+ */
+int
+zfs_jail(zfs_handle_t *zhp, int jailid, int attach)
+{
+ libzfs_handle_t *hdl = zhp->zfs_hdl;
+ zfs_cmd_t zc = { 0 };
+ char errbuf[1024];
+ int cmd, ret;
+
+ if (attach) {
+ (void) snprintf(errbuf, sizeof (errbuf),
+ dgettext(TEXT_DOMAIN, "cannot jail '%s'"), zhp->zfs_name);
+ } else {
+ (void) snprintf(errbuf, sizeof (errbuf),
+ dgettext(TEXT_DOMAIN, "cannot jail '%s'"), zhp->zfs_name);
+ }
+
+ switch (zhp->zfs_type) {
+ case ZFS_TYPE_VOLUME:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "volumes can not be jailed"));
+ return (zfs_error(hdl, EZFS_BADTYPE, errbuf));
+ case ZFS_TYPE_SNAPSHOT:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "snapshots can not be jailed"));
+ return (zfs_error(hdl, EZFS_BADTYPE, errbuf));
+ }
+ assert(zhp->zfs_type == ZFS_TYPE_FILESYSTEM);
+
+ (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name));
+ zc.zc_objset_type = DMU_OST_ZFS;
+ zc.zc_jailid = jailid;
+
+ cmd = attach ? ZFS_IOC_JAIL : ZFS_IOC_UNJAIL;
+ if ((ret = ioctl(hdl->libzfs_fd, cmd, &zc)) != 0)
+ zfs_standard_error(hdl, errno, errbuf);
+
+ return (ret);
+}
diff --git a/contrib/opensolaris/lib/libzfs/common/libzfs_graph.c b/contrib/opensolaris/lib/libzfs/common/libzfs_graph.c
new file mode 100644
index 0000000..c283016
--- /dev/null
+++ b/contrib/opensolaris/lib/libzfs/common/libzfs_graph.c
@@ -0,0 +1,646 @@
+/*
+ * 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 2006 Sun Microsystems, Inc. All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#pragma ident "%Z%%M% %I% %E% SMI"
+
+/*
+ * Iterate over all children of the current object. This includes the normal
+ * dataset hierarchy, but also arbitrary hierarchies due to clones. We want to
+ * walk all datasets in the pool, and construct a directed graph of the form:
+ *
+ * home
+ * |
+ * +----+----+
+ * | |
+ * v v ws
+ * bar baz |
+ * | |
+ * v v
+ * @yesterday ----> foo
+ *
+ * In order to construct this graph, we have to walk every dataset in the pool,
+ * because the clone parent is stored as a property of the child, not the
+ * parent. The parent only keeps track of the number of clones.
+ *
+ * In the normal case (without clones) this would be rather expensive. To avoid
+ * unnecessary computation, we first try a walk of the subtree hierarchy
+ * starting from the initial node. At each dataset, we construct a node in the
+ * graph and an edge leading from its parent. If we don't see any snapshots
+ * with a non-zero clone count, then we are finished.
+ *
+ * If we do find a cloned snapshot, then we finish the walk of the current
+ * subtree, but indicate that we need to do a complete walk. We then perform a
+ * global walk of all datasets, avoiding the subtree we already processed.
+ *
+ * At the end of this, we'll end up with a directed graph of all relevant (and
+ * possible some irrelevant) datasets in the system. We need to both find our
+ * limiting subgraph and determine a safe ordering in which to destroy the
+ * datasets. We do a topological ordering of our graph starting at our target
+ * dataset, and then walk the results in reverse.
+ *
+ * It's possible for the graph to have cycles if, for example, the user renames
+ * a clone to be the parent of its origin snapshot. The user can request to
+ * generate an error in this case, or ignore the cycle and continue.
+ *
+ * When removing datasets, we want to destroy the snapshots in chronological
+ * order (because this is the most efficient method). In order to accomplish
+ * this, we store the creation transaction group with each vertex and keep each
+ * vertex's edges sorted according to this value. The topological sort will
+ * automatically walk the snapshots in the correct order.
+ */
+
+#include <assert.h>
+#include <libintl.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <strings.h>
+#include <unistd.h>
+
+#include <libzfs.h>
+
+#include "libzfs_impl.h"
+#include "zfs_namecheck.h"
+
+#define MIN_EDGECOUNT 4
+
+/*
+ * Vertex structure. Indexed by dataset name, this structure maintains a list
+ * of edges to other vertices.
+ */
+struct zfs_edge;
+typedef struct zfs_vertex {
+ char zv_dataset[ZFS_MAXNAMELEN];
+ struct zfs_vertex *zv_next;
+ int zv_visited;
+ uint64_t zv_txg;
+ struct zfs_edge **zv_edges;
+ int zv_edgecount;
+ int zv_edgealloc;
+} zfs_vertex_t;
+
+enum {
+ VISIT_SEEN = 1,
+ VISIT_SORT_PRE,
+ VISIT_SORT_POST
+};
+
+/*
+ * Edge structure. Simply maintains a pointer to the destination vertex. There
+ * is no need to store the source vertex, since we only use edges in the context
+ * of the source vertex.
+ */
+typedef struct zfs_edge {
+ zfs_vertex_t *ze_dest;
+ struct zfs_edge *ze_next;
+} zfs_edge_t;
+
+#define ZFS_GRAPH_SIZE 1027 /* this could be dynamic some day */
+
+/*
+ * Graph structure. Vertices are maintained in a hash indexed by dataset name.
+ */
+typedef struct zfs_graph {
+ zfs_vertex_t **zg_hash;
+ size_t zg_size;
+ size_t zg_nvertex;
+} zfs_graph_t;
+
+/*
+ * Allocate a new edge pointing to the target vertex.
+ */
+static zfs_edge_t *
+zfs_edge_create(libzfs_handle_t *hdl, zfs_vertex_t *dest)
+{
+ zfs_edge_t *zep = zfs_alloc(hdl, sizeof (zfs_edge_t));
+
+ if (zep == NULL)
+ return (NULL);
+
+ zep->ze_dest = dest;
+
+ return (zep);
+}
+
+/*
+ * Destroy an edge.
+ */
+static void
+zfs_edge_destroy(zfs_edge_t *zep)
+{
+ free(zep);
+}
+
+/*
+ * Allocate a new vertex with the given name.
+ */
+static zfs_vertex_t *
+zfs_vertex_create(libzfs_handle_t *hdl, const char *dataset)
+{
+ zfs_vertex_t *zvp = zfs_alloc(hdl, sizeof (zfs_vertex_t));
+
+ if (zvp == NULL)
+ return (NULL);
+
+ assert(strlen(dataset) < ZFS_MAXNAMELEN);
+
+ (void) strlcpy(zvp->zv_dataset, dataset, sizeof (zvp->zv_dataset));
+
+ if ((zvp->zv_edges = zfs_alloc(hdl,
+ MIN_EDGECOUNT * sizeof (void *))) == NULL) {
+ free(zvp);
+ return (NULL);
+ }
+
+ zvp->zv_edgealloc = MIN_EDGECOUNT;
+
+ return (zvp);
+}
+
+/*
+ * Destroy a vertex. Frees up any associated edges.
+ */
+static void
+zfs_vertex_destroy(zfs_vertex_t *zvp)
+{
+ int i;
+
+ for (i = 0; i < zvp->zv_edgecount; i++)
+ zfs_edge_destroy(zvp->zv_edges[i]);
+
+ free(zvp->zv_edges);
+ free(zvp);
+}
+
+/*
+ * Given a vertex, add an edge to the destination vertex.
+ */
+static int
+zfs_vertex_add_edge(libzfs_handle_t *hdl, zfs_vertex_t *zvp,
+ zfs_vertex_t *dest)
+{
+ zfs_edge_t *zep = zfs_edge_create(hdl, dest);
+
+ if (zep == NULL)
+ return (-1);
+
+ if (zvp->zv_edgecount == zvp->zv_edgealloc) {
+ void *ptr;
+
+ if ((ptr = zfs_realloc(hdl, zvp->zv_edges,
+ zvp->zv_edgealloc * sizeof (void *),
+ zvp->zv_edgealloc * 2 * sizeof (void *))) == NULL)
+ return (-1);
+
+ zvp->zv_edges = ptr;
+ zvp->zv_edgealloc *= 2;
+ }
+
+ zvp->zv_edges[zvp->zv_edgecount++] = zep;
+
+ return (0);
+}
+
+static int
+zfs_edge_compare(const void *a, const void *b)
+{
+ const zfs_edge_t *ea = *((zfs_edge_t **)a);
+ const zfs_edge_t *eb = *((zfs_edge_t **)b);
+
+ if (ea->ze_dest->zv_txg < eb->ze_dest->zv_txg)
+ return (-1);
+ if (ea->ze_dest->zv_txg > eb->ze_dest->zv_txg)
+ return (1);
+ return (0);
+}
+
+/*
+ * Sort the given vertex edges according to the creation txg of each vertex.
+ */
+static void
+zfs_vertex_sort_edges(zfs_vertex_t *zvp)
+{
+ if (zvp->zv_edgecount == 0)
+ return;
+
+ qsort(zvp->zv_edges, zvp->zv_edgecount, sizeof (void *),
+ zfs_edge_compare);
+}
+
+/*
+ * Construct a new graph object. We allow the size to be specified as a
+ * parameter so in the future we can size the hash according to the number of
+ * datasets in the pool.
+ */
+static zfs_graph_t *
+zfs_graph_create(libzfs_handle_t *hdl, size_t size)
+{
+ zfs_graph_t *zgp = zfs_alloc(hdl, sizeof (zfs_graph_t));
+
+ if (zgp == NULL)
+ return (NULL);
+
+ zgp->zg_size = size;
+ if ((zgp->zg_hash = zfs_alloc(hdl,
+ size * sizeof (zfs_vertex_t *))) == NULL) {
+ free(zgp);
+ return (NULL);
+ }
+
+ return (zgp);
+}
+
+/*
+ * Destroy a graph object. We have to iterate over all the hash chains,
+ * destroying each vertex in the process.
+ */
+static void
+zfs_graph_destroy(zfs_graph_t *zgp)
+{
+ int i;
+ zfs_vertex_t *current, *next;
+
+ for (i = 0; i < zgp->zg_size; i++) {
+ current = zgp->zg_hash[i];
+ while (current != NULL) {
+ next = current->zv_next;
+ zfs_vertex_destroy(current);
+ current = next;
+ }
+ }
+
+ free(zgp->zg_hash);
+ free(zgp);
+}
+
+/*
+ * Graph hash function. Classic bernstein k=33 hash function, taken from
+ * usr/src/cmd/sgs/tools/common/strhash.c
+ */
+static size_t
+zfs_graph_hash(zfs_graph_t *zgp, const char *str)
+{
+ size_t hash = 5381;
+ int c;
+
+ while ((c = *str++) != 0)
+ hash = ((hash << 5) + hash) + c; /* hash * 33 + c */
+
+ return (hash % zgp->zg_size);
+}
+
+/*
+ * Given a dataset name, finds the associated vertex, creating it if necessary.
+ */
+static zfs_vertex_t *
+zfs_graph_lookup(libzfs_handle_t *hdl, zfs_graph_t *zgp, const char *dataset,
+ uint64_t txg)
+{
+ size_t idx = zfs_graph_hash(zgp, dataset);
+ zfs_vertex_t *zvp;
+
+ for (zvp = zgp->zg_hash[idx]; zvp != NULL; zvp = zvp->zv_next) {
+ if (strcmp(zvp->zv_dataset, dataset) == 0) {
+ if (zvp->zv_txg == 0)
+ zvp->zv_txg = txg;
+ return (zvp);
+ }
+ }
+
+ if ((zvp = zfs_vertex_create(hdl, dataset)) == NULL)
+ return (NULL);
+
+ zvp->zv_next = zgp->zg_hash[idx];
+ zvp->zv_txg = txg;
+ zgp->zg_hash[idx] = zvp;
+ zgp->zg_nvertex++;
+
+ return (zvp);
+}
+
+/*
+ * Given two dataset names, create an edge between them. For the source vertex,
+ * mark 'zv_visited' to indicate that we have seen this vertex, and not simply
+ * created it as a destination of another edge. If 'dest' is NULL, then this
+ * is an individual vertex (i.e. the starting vertex), so don't add an edge.
+ */
+static int
+zfs_graph_add(libzfs_handle_t *hdl, zfs_graph_t *zgp, const char *source,
+ const char *dest, uint64_t txg)
+{
+ zfs_vertex_t *svp, *dvp;
+
+ if ((svp = zfs_graph_lookup(hdl, zgp, source, 0)) == NULL)
+ return (-1);
+ svp->zv_visited = VISIT_SEEN;
+ if (dest != NULL) {
+ dvp = zfs_graph_lookup(hdl, zgp, dest, txg);
+ if (dvp == NULL)
+ return (-1);
+ if (zfs_vertex_add_edge(hdl, svp, dvp) != 0)
+ return (-1);
+ }
+
+ return (0);
+}
+
+/*
+ * Iterate over all children of the given dataset, adding any vertices as
+ * necessary. Returns 0 if no cloned snapshots were seen, -1 if there was an
+ * error, or 1 otherwise. This is a simple recursive algorithm - the ZFS
+ * namespace typically is very flat. We manually invoke the necessary ioctl()
+ * calls to avoid the overhead and additional semantics of zfs_open().
+ */
+static int
+iterate_children(libzfs_handle_t *hdl, zfs_graph_t *zgp, const char *dataset)
+{
+ zfs_cmd_t zc = { 0 };
+ int ret = 0, err;
+ zfs_vertex_t *zvp;
+
+ /*
+ * Look up the source vertex, and avoid it if we've seen it before.
+ */
+ zvp = zfs_graph_lookup(hdl, zgp, dataset, 0);
+ if (zvp == NULL)
+ return (-1);
+ if (zvp->zv_visited == VISIT_SEEN)
+ return (0);
+
+ /*
+ * We check the clone parent here instead of within the loop, so that if
+ * the root dataset has been promoted from a clone, we find its parent
+ * appropriately.
+ */
+ (void) strlcpy(zc.zc_name, dataset, sizeof (zc.zc_name));
+ if (ioctl(hdl->libzfs_fd, ZFS_IOC_OBJSET_STATS, &zc) == 0 &&
+ zc.zc_objset_stats.dds_clone_of[0] != '\0') {
+ if (zfs_graph_add(hdl, zgp, zc.zc_objset_stats.dds_clone_of,
+ zc.zc_name, zc.zc_objset_stats.dds_creation_txg) != 0)
+ return (-1);
+ }
+
+ for ((void) strlcpy(zc.zc_name, dataset, sizeof (zc.zc_name));
+ ioctl(hdl->libzfs_fd, ZFS_IOC_DATASET_LIST_NEXT, &zc) == 0;
+ (void) strlcpy(zc.zc_name, dataset, sizeof (zc.zc_name))) {
+
+ /*
+ * Ignore private dataset names.
+ */
+ if (dataset_name_hidden(zc.zc_name))
+ continue;
+
+ /*
+ * Get statistics for this dataset, to determine the type of the
+ * dataset and clone statistics. If this fails, the dataset has
+ * since been removed, and we're pretty much screwed anyway.
+ */
+ if (ioctl(hdl->libzfs_fd, ZFS_IOC_OBJSET_STATS, &zc) != 0)
+ continue;
+
+ /*
+ * Add an edge between the parent and the child.
+ */
+ if (zfs_graph_add(hdl, zgp, dataset, zc.zc_name,
+ zc.zc_objset_stats.dds_creation_txg) != 0)
+ return (-1);
+
+ /*
+ * Iterate over all children
+ */
+ err = iterate_children(hdl, zgp, zc.zc_name);
+ if (err == -1)
+ return (-1);
+ else if (err == 1)
+ ret = 1;
+
+ /*
+ * Indicate if we found a dataset with a non-zero clone count.
+ */
+ if (zc.zc_objset_stats.dds_num_clones != 0)
+ ret = 1;
+ }
+
+ /*
+ * Now iterate over all snapshots.
+ */
+ bzero(&zc, sizeof (zc));
+
+ for ((void) strlcpy(zc.zc_name, dataset, sizeof (zc.zc_name));
+ ioctl(hdl->libzfs_fd, ZFS_IOC_SNAPSHOT_LIST_NEXT, &zc) == 0;
+ (void) strlcpy(zc.zc_name, dataset, sizeof (zc.zc_name))) {
+
+ /*
+ * Get statistics for this dataset, to determine the type of the
+ * dataset and clone statistics. If this fails, the dataset has
+ * since been removed, and we're pretty much screwed anyway.
+ */
+ if (ioctl(hdl->libzfs_fd, ZFS_IOC_OBJSET_STATS, &zc) != 0)
+ continue;
+
+ /*
+ * Add an edge between the parent and the child.
+ */
+ if (zfs_graph_add(hdl, zgp, dataset, zc.zc_name,
+ zc.zc_objset_stats.dds_creation_txg) != 0)
+ return (-1);
+
+ /*
+ * Indicate if we found a dataset with a non-zero clone count.
+ */
+ if (zc.zc_objset_stats.dds_num_clones != 0)
+ ret = 1;
+ }
+
+ zvp->zv_visited = VISIT_SEEN;
+
+ return (ret);
+}
+
+/*
+ * Construct a complete graph of all necessary vertices. First, we iterate over
+ * only our object's children. If we don't find any cloned snapshots, then we
+ * simple return that. Otherwise, we have to start at the pool root and iterate
+ * over all datasets.
+ */
+static zfs_graph_t *
+construct_graph(libzfs_handle_t *hdl, const char *dataset)
+{
+ zfs_graph_t *zgp = zfs_graph_create(hdl, ZFS_GRAPH_SIZE);
+ zfs_cmd_t zc = { 0 };
+ int ret = 0;
+
+ if (zgp == NULL)
+ return (zgp);
+
+ /*
+ * We need to explicitly check whether this dataset has clones or not,
+ * since iterate_children() only checks the children.
+ */
+ (void) strlcpy(zc.zc_name, dataset, sizeof (zc.zc_name));
+ (void) ioctl(hdl->libzfs_fd, ZFS_IOC_OBJSET_STATS, &zc);
+
+ if (zc.zc_objset_stats.dds_num_clones != 0 ||
+ (ret = iterate_children(hdl, zgp, dataset)) != 0) {
+ /*
+ * Determine pool name and try again.
+ */
+ char *pool, *slash;
+
+ if ((slash = strchr(dataset, '/')) != NULL ||
+ (slash = strchr(dataset, '@')) != NULL) {
+ pool = zfs_alloc(hdl, slash - dataset + 1);
+ if (pool == NULL) {
+ zfs_graph_destroy(zgp);
+ return (NULL);
+ }
+ (void) strncpy(pool, dataset, slash - dataset);
+ pool[slash - dataset] = '\0';
+
+ if (iterate_children(hdl, zgp, pool) == -1 ||
+ zfs_graph_add(hdl, zgp, pool, NULL, 0) != 0) {
+ free(pool);
+ zfs_graph_destroy(zgp);
+ return (NULL);
+ }
+
+ free(pool);
+ }
+ }
+
+ if (ret == -1 || zfs_graph_add(hdl, zgp, dataset, NULL, 0) != 0) {
+ zfs_graph_destroy(zgp);
+ return (NULL);
+ }
+
+ return (zgp);
+}
+
+/*
+ * Given a graph, do a recursive topological sort into the given array. This is
+ * really just a depth first search, so that the deepest nodes appear first.
+ * hijack the 'zv_visited' marker to avoid visiting the same vertex twice.
+ */
+static int
+topo_sort(libzfs_handle_t *hdl, boolean_t allowrecursion, char **result,
+ size_t *idx, zfs_vertex_t *zgv)
+{
+ int i;
+
+ if (zgv->zv_visited == VISIT_SORT_PRE && !allowrecursion) {
+ /*
+ * If we've already seen this vertex as part of our depth-first
+ * search, then we have a cyclic dependency, and we must return
+ * an error.
+ */
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "recursive dependency at '%s'"),
+ zgv->zv_dataset);
+ return (zfs_error(hdl, EZFS_RECURSIVE,
+ dgettext(TEXT_DOMAIN,
+ "cannot determine dependent datasets")));
+ } else if (zgv->zv_visited >= VISIT_SORT_PRE) {
+ /*
+ * If we've already processed this as part of the topological
+ * sort, then don't bother doing so again.
+ */
+ return (0);
+ }
+
+ zgv->zv_visited = VISIT_SORT_PRE;
+
+ /* avoid doing a search if we don't have to */
+ zfs_vertex_sort_edges(zgv);
+ for (i = 0; i < zgv->zv_edgecount; i++) {
+ if (topo_sort(hdl, allowrecursion, result, idx,
+ zgv->zv_edges[i]->ze_dest) != 0)
+ return (-1);
+ }
+
+ /* we may have visited this in the course of the above */
+ if (zgv->zv_visited == VISIT_SORT_POST)
+ return (0);
+
+ if ((result[*idx] = zfs_alloc(hdl,
+ strlen(zgv->zv_dataset) + 1)) == NULL)
+ return (-1);
+
+ (void) strcpy(result[*idx], zgv->zv_dataset);
+ *idx += 1;
+ zgv->zv_visited = VISIT_SORT_POST;
+ return (0);
+}
+
+/*
+ * The only public interface for this file. Do the dirty work of constructing a
+ * child list for the given object. Construct the graph, do the toplogical
+ * sort, and then return the array of strings to the caller.
+ *
+ * The 'allowrecursion' parameter controls behavior when cycles are found. If
+ * it is set, the the cycle is ignored and the results returned as if the cycle
+ * did not exist. If it is not set, then the routine will generate an error if
+ * a cycle is found.
+ */
+int
+get_dependents(libzfs_handle_t *hdl, boolean_t allowrecursion,
+ const char *dataset, char ***result, size_t *count)
+{
+ zfs_graph_t *zgp;
+ zfs_vertex_t *zvp;
+
+ if ((zgp = construct_graph(hdl, dataset)) == NULL)
+ return (-1);
+
+ if ((*result = zfs_alloc(hdl,
+ zgp->zg_nvertex * sizeof (char *))) == NULL) {
+ zfs_graph_destroy(zgp);
+ return (-1);
+ }
+
+ if ((zvp = zfs_graph_lookup(hdl, zgp, dataset, 0)) == NULL) {
+ free(*result);
+ zfs_graph_destroy(zgp);
+ return (-1);
+ }
+
+ *count = 0;
+ if (topo_sort(hdl, allowrecursion, *result, count, zvp) != 0) {
+ free(*result);
+ zfs_graph_destroy(zgp);
+ return (-1);
+ }
+
+ /*
+ * Get rid of the last entry, which is our starting vertex and not
+ * strictly a dependent.
+ */
+ assert(*count > 0);
+ free((*result)[*count - 1]);
+ (*count)--;
+
+ zfs_graph_destroy(zgp);
+
+ return (0);
+}
diff --git a/contrib/opensolaris/lib/libzfs/common/libzfs_impl.h b/contrib/opensolaris/lib/libzfs/common/libzfs_impl.h
new file mode 100644
index 0000000..9581331
--- /dev/null
+++ b/contrib/opensolaris/lib/libzfs/common/libzfs_impl.h
@@ -0,0 +1,171 @@
+/*
+ * 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.
+ */
+
+#ifndef _LIBFS_IMPL_H
+#define _LIBFS_IMPL_H
+
+#pragma ident "%Z%%M% %I% %E% SMI"
+
+#include <sys/dmu.h>
+#include <sys/fs/zfs.h>
+#include <sys/zfs_ioctl.h>
+#include <sys/zfs_acl.h>
+#include <sys/nvpair.h>
+
+#include <libuutil.h>
+#include <libzfs.h>
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+struct libzfs_handle {
+ int libzfs_error;
+ int libzfs_fd;
+ FILE *libzfs_mnttab;
+ FILE *libzfs_sharetab;
+ uu_avl_pool_t *libzfs_ns_avlpool;
+ uu_avl_t *libzfs_ns_avl;
+ uint64_t libzfs_ns_gen;
+ int libzfs_desc_active;
+ char libzfs_action[1024];
+ char libzfs_desc[1024];
+ int libzfs_printerr;
+};
+
+struct zfs_handle {
+ libzfs_handle_t *zfs_hdl;
+ char zfs_name[ZFS_MAXNAMELEN];
+ zfs_type_t zfs_type; /* type including snapshot */
+ zfs_type_t zfs_head_type; /* type excluding snapshot */
+ dmu_objset_stats_t zfs_dmustats;
+ nvlist_t *zfs_props;
+ nvlist_t *zfs_user_props;
+ boolean_t zfs_mntcheck;
+ char *zfs_mntopts;
+ char zfs_root[MAXPATHLEN];
+};
+
+/*
+ * This is different from checking zfs_type, because it will also catch
+ * snapshots of volumes.
+ */
+#define ZFS_IS_VOLUME(zhp) ((zhp)->zfs_head_type == ZFS_TYPE_VOLUME)
+
+struct zpool_handle {
+ libzfs_handle_t *zpool_hdl;
+ char zpool_name[ZPOOL_MAXNAMELEN];
+ int zpool_state;
+ size_t zpool_config_size;
+ nvlist_t *zpool_config;
+ nvlist_t *zpool_old_config;
+ nvlist_t *zpool_props;
+};
+
+int zfs_error(libzfs_handle_t *, int, const char *);
+int zfs_error_fmt(libzfs_handle_t *, int, const char *, ...);
+void zfs_error_aux(libzfs_handle_t *, const char *, ...);
+void *zfs_alloc(libzfs_handle_t *, size_t);
+void *zfs_realloc(libzfs_handle_t *, void *, size_t, size_t);
+char *zfs_strdup(libzfs_handle_t *, const char *);
+int no_memory(libzfs_handle_t *);
+
+int zfs_standard_error(libzfs_handle_t *, int, const char *);
+int zfs_standard_error_fmt(libzfs_handle_t *, int, const char *, ...);
+int zpool_standard_error(libzfs_handle_t *, int, const char *);
+int zpool_standard_error_fmt(libzfs_handle_t *, int, const char *, ...);
+
+int get_dependents(libzfs_handle_t *, boolean_t, const char *, char ***,
+ size_t *);
+
+int zfs_expand_proplist_common(libzfs_handle_t *, zfs_proplist_t **,
+ zfs_type_t);
+int zfs_get_proplist_common(libzfs_handle_t *, char *, zfs_proplist_t **,
+ zfs_type_t);
+zfs_prop_t zfs_prop_iter_common(zfs_prop_f, void *, zfs_type_t, boolean_t);
+zfs_prop_t zfs_name_to_prop_common(const char *, zfs_type_t);
+
+nvlist_t *zfs_validate_properties(libzfs_handle_t *, zfs_type_t, char *,
+ nvlist_t *, uint64_t, zfs_handle_t *zhp, const char *errbuf);
+
+typedef struct prop_changelist prop_changelist_t;
+
+int zcmd_alloc_dst_nvlist(libzfs_handle_t *, zfs_cmd_t *, size_t);
+int zcmd_write_src_nvlist(libzfs_handle_t *, zfs_cmd_t *, nvlist_t *, size_t *);
+int zcmd_expand_dst_nvlist(libzfs_handle_t *, zfs_cmd_t *);
+int zcmd_read_dst_nvlist(libzfs_handle_t *, zfs_cmd_t *, nvlist_t **);
+void zcmd_free_nvlists(zfs_cmd_t *);
+
+int changelist_prefix(prop_changelist_t *);
+int changelist_postfix(prop_changelist_t *);
+void changelist_rename(prop_changelist_t *, const char *, const char *);
+void changelist_remove(zfs_handle_t *, prop_changelist_t *);
+void changelist_free(prop_changelist_t *);
+prop_changelist_t *changelist_gather(zfs_handle_t *, zfs_prop_t, int);
+int changelist_unshare(prop_changelist_t *);
+int changelist_haszonedchild(prop_changelist_t *);
+
+void remove_mountpoint(zfs_handle_t *);
+
+zfs_handle_t *make_dataset_handle(libzfs_handle_t *, const char *);
+
+int zpool_open_silent(libzfs_handle_t *, const char *, zpool_handle_t **);
+
+int zvol_create_link(libzfs_handle_t *, const char *);
+int zvol_remove_link(libzfs_handle_t *, const char *);
+int zpool_iter_zvol(zpool_handle_t *, int (*)(const char *, void *), void *);
+
+void namespace_clear(libzfs_handle_t *);
+
+#ifdef __FreeBSD__
+/*
+ * This is FreeBSD version of ioctl, because Solaris' ioctl() updates
+ * zc_nvlist_dst_size even if an error is returned, on FreeBSD if an
+ * error is returned zc_nvlist_dst_size won't be updated.
+ */
+static __inline int
+zcmd_ioctl(int fd, unsigned long cmd, zfs_cmd_t *zc)
+{
+ size_t oldsize;
+ int ret;
+
+ oldsize = zc->zc_nvlist_dst_size;
+ ret = ioctl(fd, cmd, zc);
+ if (ret == 0 && oldsize < zc->zc_nvlist_dst_size) {
+ ret = -1;
+ errno = ENOMEM;
+ }
+
+ return (ret);
+}
+#define ioctl(fd, cmd, zc) zcmd_ioctl((fd), (cmd), (zc))
+#endif
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* _LIBFS_IMPL_H */
diff --git a/contrib/opensolaris/lib/libzfs/common/libzfs_import.c b/contrib/opensolaris/lib/libzfs/common/libzfs_import.c
new file mode 100644
index 0000000..cab969b
--- /dev/null
+++ b/contrib/opensolaris/lib/libzfs/common/libzfs_import.c
@@ -0,0 +1,994 @@
+/*
+ * 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"
+
+/*
+ * Pool import support functions.
+ *
+ * To import a pool, we rely on reading the configuration information from the
+ * ZFS label of each device. If we successfully read the label, then we
+ * organize the configuration information in the following hierarchy:
+ *
+ * pool guid -> toplevel vdev guid -> label txg
+ *
+ * Duplicate entries matching this same tuple will be discarded. Once we have
+ * examined every device, we pick the best label txg config for each toplevel
+ * vdev. We then arrange these toplevel vdevs into a complete pool config, and
+ * update any paths that have changed. Finally, we attempt to import the pool
+ * using our derived config, and record the results.
+ */
+
+#include <devid.h>
+#include <dirent.h>
+#include <errno.h>
+#include <libintl.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/stat.h>
+#include <unistd.h>
+#include <fcntl.h>
+#include <libgeom.h>
+
+#include <sys/vdev_impl.h>
+
+#include "libzfs.h"
+#include "libzfs_impl.h"
+
+/*
+ * Intermediate structures used to gather configuration information.
+ */
+typedef struct config_entry {
+ uint64_t ce_txg;
+ nvlist_t *ce_config;
+ struct config_entry *ce_next;
+} config_entry_t;
+
+typedef struct vdev_entry {
+ uint64_t ve_guid;
+ config_entry_t *ve_configs;
+ struct vdev_entry *ve_next;
+} vdev_entry_t;
+
+typedef struct pool_entry {
+ uint64_t pe_guid;
+ vdev_entry_t *pe_vdevs;
+ struct pool_entry *pe_next;
+} pool_entry_t;
+
+typedef struct name_entry {
+ char *ne_name;
+ uint64_t ne_guid;
+ struct name_entry *ne_next;
+} name_entry_t;
+
+typedef struct pool_list {
+ pool_entry_t *pools;
+ name_entry_t *names;
+} pool_list_t;
+
+static char *
+get_devid(const char *path)
+{
+ int fd;
+ ddi_devid_t devid;
+ char *minor, *ret;
+
+ if ((fd = open(path, O_RDONLY)) < 0)
+ return (NULL);
+
+ minor = NULL;
+ ret = NULL;
+ if (devid_get(fd, &devid) == 0) {
+ if (devid_get_minor_name(fd, &minor) == 0)
+ ret = devid_str_encode(devid, minor);
+ if (minor != NULL)
+ devid_str_free(minor);
+ devid_free(devid);
+ }
+ (void) close(fd);
+
+ return (ret);
+}
+
+/*
+ * Go through and fix up any path and/or devid information for the given vdev
+ * configuration.
+ */
+static int
+fix_paths(nvlist_t *nv, name_entry_t *names)
+{
+ nvlist_t **child;
+ uint_t c, children;
+ uint64_t guid;
+ name_entry_t *ne, *best;
+ char *path, *devid;
+ int matched;
+
+ if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
+ &child, &children) == 0) {
+ for (c = 0; c < children; c++)
+ if (fix_paths(child[c], names) != 0)
+ return (-1);
+ return (0);
+ }
+
+ /*
+ * This is a leaf (file or disk) vdev. In either case, go through
+ * the name list and see if we find a matching guid. If so, replace
+ * the path and see if we can calculate a new devid.
+ *
+ * There may be multiple names associated with a particular guid, in
+ * which case we have overlapping slices or multiple paths to the same
+ * disk. If this is the case, then we want to pick the path that is
+ * the most similar to the original, where "most similar" is the number
+ * of matching characters starting from the end of the path. This will
+ * preserve slice numbers even if the disks have been reorganized, and
+ * will also catch preferred disk names if multiple paths exist.
+ */
+ verify(nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) == 0);
+ if (nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &path) != 0)
+ path = NULL;
+
+ matched = 0;
+ best = NULL;
+ for (ne = names; ne != NULL; ne = ne->ne_next) {
+ if (ne->ne_guid == guid) {
+ const char *src, *dst;
+ int count;
+
+ if (path == NULL) {
+ best = ne;
+ break;
+ }
+
+ src = ne->ne_name + strlen(ne->ne_name) - 1;
+ dst = path + strlen(path) - 1;
+ for (count = 0; src >= ne->ne_name && dst >= path;
+ src--, dst--, count++)
+ if (*src != *dst)
+ break;
+
+ /*
+ * At this point, 'count' is the number of characters
+ * matched from the end.
+ */
+ if (count > matched || best == NULL) {
+ best = ne;
+ matched = count;
+ }
+ }
+ }
+
+ if (best == NULL)
+ return (0);
+
+ if (nvlist_add_string(nv, ZPOOL_CONFIG_PATH, best->ne_name) != 0)
+ return (-1);
+
+ if ((devid = get_devid(best->ne_name)) == NULL) {
+ (void) nvlist_remove_all(nv, ZPOOL_CONFIG_DEVID);
+ } else {
+ if (nvlist_add_string(nv, ZPOOL_CONFIG_DEVID, devid) != 0)
+ return (-1);
+ devid_str_free(devid);
+ }
+
+ return (0);
+}
+
+/*
+ * Add the given configuration to the list of known devices.
+ */
+static int
+add_config(libzfs_handle_t *hdl, pool_list_t *pl, const char *path,
+ nvlist_t *config)
+{
+ uint64_t pool_guid, vdev_guid, top_guid, txg, state;
+ pool_entry_t *pe;
+ vdev_entry_t *ve;
+ config_entry_t *ce;
+ name_entry_t *ne;
+
+ /*
+ * If this is a hot spare not currently in use, add it to the list of
+ * names to translate, but don't do anything else.
+ */
+ if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_STATE,
+ &state) == 0 && state == POOL_STATE_SPARE &&
+ nvlist_lookup_uint64(config, ZPOOL_CONFIG_GUID, &vdev_guid) == 0) {
+ if ((ne = zfs_alloc(hdl, sizeof (name_entry_t))) == NULL)
+ return (-1);
+
+ if ((ne->ne_name = zfs_strdup(hdl, path)) == NULL) {
+ free(ne);
+ return (-1);
+ }
+ ne->ne_guid = vdev_guid;
+ ne->ne_next = pl->names;
+ pl->names = ne;
+ return (0);
+ }
+
+ /*
+ * If we have a valid config but cannot read any of these fields, then
+ * it means we have a half-initialized label. In vdev_label_init()
+ * we write a label with txg == 0 so that we can identify the device
+ * in case the user refers to the same disk later on. If we fail to
+ * create the pool, we'll be left with a label in this state
+ * which should not be considered part of a valid pool.
+ */
+ if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID,
+ &pool_guid) != 0 ||
+ nvlist_lookup_uint64(config, ZPOOL_CONFIG_GUID,
+ &vdev_guid) != 0 ||
+ nvlist_lookup_uint64(config, ZPOOL_CONFIG_TOP_GUID,
+ &top_guid) != 0 ||
+ nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_TXG,
+ &txg) != 0 || txg == 0) {
+ nvlist_free(config);
+ return (0);
+ }
+
+ /*
+ * First, see if we know about this pool. If not, then add it to the
+ * list of known pools.
+ */
+ for (pe = pl->pools; pe != NULL; pe = pe->pe_next) {
+ if (pe->pe_guid == pool_guid)
+ break;
+ }
+
+ if (pe == NULL) {
+ if ((pe = zfs_alloc(hdl, sizeof (pool_entry_t))) == NULL) {
+ nvlist_free(config);
+ return (-1);
+ }
+ pe->pe_guid = pool_guid;
+ pe->pe_next = pl->pools;
+ pl->pools = pe;
+ }
+
+ /*
+ * Second, see if we know about this toplevel vdev. Add it if its
+ * missing.
+ */
+ for (ve = pe->pe_vdevs; ve != NULL; ve = ve->ve_next) {
+ if (ve->ve_guid == top_guid)
+ break;
+ }
+
+ if (ve == NULL) {
+ if ((ve = zfs_alloc(hdl, sizeof (vdev_entry_t))) == NULL) {
+ nvlist_free(config);
+ return (-1);
+ }
+ ve->ve_guid = top_guid;
+ ve->ve_next = pe->pe_vdevs;
+ pe->pe_vdevs = ve;
+ }
+
+ /*
+ * Third, see if we have a config with a matching transaction group. If
+ * so, then we do nothing. Otherwise, add it to the list of known
+ * configs.
+ */
+ for (ce = ve->ve_configs; ce != NULL; ce = ce->ce_next) {
+ if (ce->ce_txg == txg)
+ break;
+ }
+
+ if (ce == NULL) {
+ if ((ce = zfs_alloc(hdl, sizeof (config_entry_t))) == NULL) {
+ nvlist_free(config);
+ return (-1);
+ }
+ ce->ce_txg = txg;
+ ce->ce_config = config;
+ ce->ce_next = ve->ve_configs;
+ ve->ve_configs = ce;
+ } else {
+ nvlist_free(config);
+ }
+
+ /*
+ * At this point we've successfully added our config to the list of
+ * known configs. The last thing to do is add the vdev guid -> path
+ * mappings so that we can fix up the configuration as necessary before
+ * doing the import.
+ */
+ if ((ne = zfs_alloc(hdl, sizeof (name_entry_t))) == NULL)
+ return (-1);
+
+ if ((ne->ne_name = zfs_strdup(hdl, path)) == NULL) {
+ free(ne);
+ return (-1);
+ }
+
+ ne->ne_guid = vdev_guid;
+ ne->ne_next = pl->names;
+ pl->names = ne;
+
+ return (0);
+}
+
+/*
+ * Returns true if the named pool matches the given GUID.
+ */
+static int
+pool_active(libzfs_handle_t *hdl, const char *name, uint64_t guid,
+ boolean_t *isactive)
+{
+ zpool_handle_t *zhp;
+ uint64_t theguid;
+
+ if (zpool_open_silent(hdl, name, &zhp) != 0)
+ return (-1);
+
+ if (zhp == NULL) {
+ *isactive = B_FALSE;
+ return (0);
+ }
+
+ verify(nvlist_lookup_uint64(zhp->zpool_config, ZPOOL_CONFIG_POOL_GUID,
+ &theguid) == 0);
+
+ zpool_close(zhp);
+
+ *isactive = (theguid == guid);
+ return (0);
+}
+
+/*
+ * Convert our list of pools into the definitive set of configurations. We
+ * start by picking the best config for each toplevel vdev. Once that's done,
+ * we assemble the toplevel vdevs into a full config for the pool. We make a
+ * pass to fix up any incorrect paths, and then add it to the main list to
+ * return to the user.
+ */
+static nvlist_t *
+get_configs(libzfs_handle_t *hdl, pool_list_t *pl)
+{
+ pool_entry_t *pe;
+ vdev_entry_t *ve;
+ config_entry_t *ce;
+ nvlist_t *ret = NULL, *config = NULL, *tmp, *nvtop, *nvroot;
+ nvlist_t **spares;
+ uint_t i, nspares;
+ boolean_t config_seen;
+ uint64_t best_txg;
+ char *name;
+ zfs_cmd_t zc = { 0 };
+ uint64_t version, guid;
+ size_t len;
+ int err;
+ uint_t children = 0;
+ nvlist_t **child = NULL;
+ uint_t c;
+ boolean_t isactive;
+
+ if (nvlist_alloc(&ret, 0, 0) != 0)
+ goto nomem;
+
+ for (pe = pl->pools; pe != NULL; pe = pe->pe_next) {
+ uint64_t id;
+
+ if (nvlist_alloc(&config, NV_UNIQUE_NAME, 0) != 0)
+ goto nomem;
+ config_seen = B_FALSE;
+
+ /*
+ * Iterate over all toplevel vdevs. Grab the pool configuration
+ * from the first one we find, and then go through the rest and
+ * add them as necessary to the 'vdevs' member of the config.
+ */
+ for (ve = pe->pe_vdevs; ve != NULL; ve = ve->ve_next) {
+
+ /*
+ * Determine the best configuration for this vdev by
+ * selecting the config with the latest transaction
+ * group.
+ */
+ best_txg = 0;
+ for (ce = ve->ve_configs; ce != NULL;
+ ce = ce->ce_next) {
+
+ if (ce->ce_txg > best_txg) {
+ tmp = ce->ce_config;
+ best_txg = ce->ce_txg;
+ }
+ }
+
+ if (!config_seen) {
+ /*
+ * Copy the relevant pieces of data to the pool
+ * configuration:
+ *
+ * version
+ * pool guid
+ * name
+ * pool state
+ */
+ uint64_t state;
+
+ verify(nvlist_lookup_uint64(tmp,
+ ZPOOL_CONFIG_VERSION, &version) == 0);
+ if (nvlist_add_uint64(config,
+ ZPOOL_CONFIG_VERSION, version) != 0)
+ goto nomem;
+ verify(nvlist_lookup_uint64(tmp,
+ ZPOOL_CONFIG_POOL_GUID, &guid) == 0);
+ if (nvlist_add_uint64(config,
+ ZPOOL_CONFIG_POOL_GUID, guid) != 0)
+ goto nomem;
+ verify(nvlist_lookup_string(tmp,
+ ZPOOL_CONFIG_POOL_NAME, &name) == 0);
+ if (nvlist_add_string(config,
+ ZPOOL_CONFIG_POOL_NAME, name) != 0)
+ goto nomem;
+ verify(nvlist_lookup_uint64(tmp,
+ ZPOOL_CONFIG_POOL_STATE, &state) == 0);
+ if (nvlist_add_uint64(config,
+ ZPOOL_CONFIG_POOL_STATE, state) != 0)
+ goto nomem;
+
+ config_seen = B_TRUE;
+ }
+
+ /*
+ * Add this top-level vdev to the child array.
+ */
+ verify(nvlist_lookup_nvlist(tmp,
+ ZPOOL_CONFIG_VDEV_TREE, &nvtop) == 0);
+ verify(nvlist_lookup_uint64(nvtop, ZPOOL_CONFIG_ID,
+ &id) == 0);
+ if (id >= children) {
+ nvlist_t **newchild;
+
+ newchild = zfs_alloc(hdl, (id + 1) *
+ sizeof (nvlist_t *));
+ if (newchild == NULL)
+ goto nomem;
+
+ for (c = 0; c < children; c++)
+ newchild[c] = child[c];
+
+ free(child);
+ child = newchild;
+ children = id + 1;
+ }
+ if (nvlist_dup(nvtop, &child[id], 0) != 0)
+ goto nomem;
+
+ }
+
+ verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID,
+ &guid) == 0);
+
+ /*
+ * Look for any missing top-level vdevs. If this is the case,
+ * create a faked up 'missing' vdev as a placeholder. We cannot
+ * simply compress the child array, because the kernel performs
+ * certain checks to make sure the vdev IDs match their location
+ * in the configuration.
+ */
+ for (c = 0; c < children; c++)
+ if (child[c] == NULL) {
+ nvlist_t *missing;
+ if (nvlist_alloc(&missing, NV_UNIQUE_NAME,
+ 0) != 0)
+ goto nomem;
+ if (nvlist_add_string(missing,
+ ZPOOL_CONFIG_TYPE,
+ VDEV_TYPE_MISSING) != 0 ||
+ nvlist_add_uint64(missing,
+ ZPOOL_CONFIG_ID, c) != 0 ||
+ nvlist_add_uint64(missing,
+ ZPOOL_CONFIG_GUID, 0ULL) != 0) {
+ nvlist_free(missing);
+ goto nomem;
+ }
+ child[c] = missing;
+ }
+
+ /*
+ * Put all of this pool's top-level vdevs into a root vdev.
+ */
+ if (nvlist_alloc(&nvroot, NV_UNIQUE_NAME, 0) != 0)
+ goto nomem;
+ if (nvlist_add_string(nvroot, ZPOOL_CONFIG_TYPE,
+ VDEV_TYPE_ROOT) != 0 ||
+ nvlist_add_uint64(nvroot, ZPOOL_CONFIG_ID, 0ULL) != 0 ||
+ nvlist_add_uint64(nvroot, ZPOOL_CONFIG_GUID, guid) != 0 ||
+ nvlist_add_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN,
+ child, children) != 0) {
+ nvlist_free(nvroot);
+ goto nomem;
+ }
+
+ for (c = 0; c < children; c++)
+ nvlist_free(child[c]);
+ free(child);
+ children = 0;
+ child = NULL;
+
+ /*
+ * Go through and fix up any paths and/or devids based on our
+ * known list of vdev GUID -> path mappings.
+ */
+ if (fix_paths(nvroot, pl->names) != 0) {
+ nvlist_free(nvroot);
+ goto nomem;
+ }
+
+ /*
+ * Add the root vdev to this pool's configuration.
+ */
+ if (nvlist_add_nvlist(config, ZPOOL_CONFIG_VDEV_TREE,
+ nvroot) != 0) {
+ nvlist_free(nvroot);
+ goto nomem;
+ }
+ nvlist_free(nvroot);
+
+ /*
+ * Determine if this pool is currently active, in which case we
+ * can't actually import it.
+ */
+ verify(nvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME,
+ &name) == 0);
+ verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID,
+ &guid) == 0);
+
+ if (pool_active(hdl, name, guid, &isactive) != 0)
+ goto error;
+
+ if (isactive) {
+ nvlist_free(config);
+ config = NULL;
+ continue;
+ }
+
+ /*
+ * Try to do the import in order to get vdev state.
+ */
+ if (zcmd_write_src_nvlist(hdl, &zc, config, &len) != 0)
+ goto error;
+
+ nvlist_free(config);
+ config = NULL;
+
+ if (zcmd_alloc_dst_nvlist(hdl, &zc, len * 2) != 0) {
+ zcmd_free_nvlists(&zc);
+ goto error;
+ }
+
+ while ((err = ioctl(hdl->libzfs_fd, ZFS_IOC_POOL_TRYIMPORT,
+ &zc)) != 0 && errno == ENOMEM) {
+ if (zcmd_expand_dst_nvlist(hdl, &zc) != 0) {
+ zcmd_free_nvlists(&zc);
+ goto error;
+ }
+ }
+
+ if (err) {
+ (void) zpool_standard_error(hdl, errno,
+ dgettext(TEXT_DOMAIN, "cannot discover pools"));
+ zcmd_free_nvlists(&zc);
+ goto error;
+ }
+
+ if (zcmd_read_dst_nvlist(hdl, &zc, &config) != 0) {
+ zcmd_free_nvlists(&zc);
+ goto error;
+ }
+
+ zcmd_free_nvlists(&zc);
+
+ /*
+ * Go through and update the paths for spares, now that we have
+ * them.
+ */
+ verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE,
+ &nvroot) == 0);
+ if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES,
+ &spares, &nspares) == 0) {
+ for (i = 0; i < nspares; i++) {
+ if (fix_paths(spares[i], pl->names) != 0)
+ goto nomem;
+ }
+ }
+
+ /*
+ * Add this pool to the list of configs.
+ */
+ verify(nvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME,
+ &name) == 0);
+ if (nvlist_add_nvlist(ret, name, config) != 0)
+ goto nomem;
+
+ nvlist_free(config);
+ config = NULL;
+ }
+
+ return (ret);
+
+nomem:
+ (void) no_memory(hdl);
+error:
+ nvlist_free(config);
+ nvlist_free(ret);
+ for (c = 0; c < children; c++)
+ nvlist_free(child[c]);
+ free(child);
+
+ return (NULL);
+}
+
+/*
+ * Return the offset of the given label.
+ */
+static uint64_t
+label_offset(size_t size, int l)
+{
+ return (l * sizeof (vdev_label_t) + (l < VDEV_LABELS / 2 ?
+ 0 : size - VDEV_LABELS * sizeof (vdev_label_t)));
+}
+
+/*
+ * Given a file descriptor, read the label information and return an nvlist
+ * describing the configuration, if there is one.
+ */
+int
+zpool_read_label(int fd, nvlist_t **config)
+{
+ struct stat64 statbuf;
+ int l;
+ vdev_label_t *label;
+ uint64_t state, txg;
+
+ *config = NULL;
+
+ if (fstat64(fd, &statbuf) == -1)
+ return (0);
+
+ if ((label = malloc(sizeof (vdev_label_t))) == NULL)
+ return (-1);
+
+ for (l = 0; l < VDEV_LABELS; l++) {
+ if (pread(fd, label, sizeof (vdev_label_t),
+ label_offset(statbuf.st_size, l)) != sizeof (vdev_label_t))
+ continue;
+
+ if (nvlist_unpack(label->vl_vdev_phys.vp_nvlist,
+ sizeof (label->vl_vdev_phys.vp_nvlist), config, 0) != 0)
+ continue;
+
+ if (nvlist_lookup_uint64(*config, ZPOOL_CONFIG_POOL_STATE,
+ &state) != 0 || state > POOL_STATE_SPARE) {
+ nvlist_free(*config);
+ continue;
+ }
+
+ if (state != POOL_STATE_SPARE &&
+ (nvlist_lookup_uint64(*config, ZPOOL_CONFIG_POOL_TXG,
+ &txg) != 0 || txg == 0)) {
+ nvlist_free(*config);
+ continue;
+ }
+
+ free(label);
+ return (0);
+ }
+
+ free(label);
+ *config = NULL;
+ return (0);
+}
+
+/*
+ * Given a list of directories to search, find all pools stored on disk. This
+ * includes partial pools which are not available to import. If no args are
+ * given (argc is 0), then the default directory (/dev) is searched.
+ */
+nvlist_t *
+zpool_find_import(libzfs_handle_t *hdl, int argc, char **argv)
+{
+ int i;
+ DIR *dirp;
+ char path[MAXPATHLEN];
+ nvlist_t *ret = NULL, *config;
+ int fd;
+ pool_list_t pools = { 0 };
+ pool_entry_t *pe, *penext;
+ vdev_entry_t *ve, *venext;
+ config_entry_t *ce, *cenext;
+ name_entry_t *ne, *nenext;
+ struct gmesh mesh;
+ struct gclass *mp;
+ struct ggeom *gp;
+ struct gprovider *pp;
+
+ /*
+ * Go through and read the label configuration information from every
+ * possible device, organizing the information according to pool GUID
+ * and toplevel GUID.
+ */
+
+ fd = geom_gettree(&mesh);
+ assert(fd == 0);
+
+ LIST_FOREACH(mp, &mesh.lg_class, lg_class) {
+ LIST_FOREACH(gp, &mp->lg_geom, lg_geom) {
+ LIST_FOREACH(pp, &gp->lg_provider, lg_provider) {
+
+ (void) snprintf(path, sizeof (path), "%s%s",
+ _PATH_DEV, pp->lg_name);
+
+ if ((fd = open64(path, O_RDONLY)) < 0)
+ continue;
+
+ if ((zpool_read_label(fd, &config)) != 0) {
+ (void) no_memory(hdl);
+ goto error;
+ }
+
+ (void) close(fd);
+
+ if (config == NULL)
+ continue;
+
+ if (add_config(hdl, &pools, path, config) != 0)
+ goto error;
+ }
+ }
+ }
+
+ geom_deletetree(&mesh);
+
+ ret = get_configs(hdl, &pools);
+
+error:
+ for (pe = pools.pools; pe != NULL; pe = penext) {
+ penext = pe->pe_next;
+ for (ve = pe->pe_vdevs; ve != NULL; ve = venext) {
+ venext = ve->ve_next;
+ for (ce = ve->ve_configs; ce != NULL; ce = cenext) {
+ cenext = ce->ce_next;
+ if (ce->ce_config)
+ nvlist_free(ce->ce_config);
+ free(ce);
+ }
+ free(ve);
+ }
+ free(pe);
+ }
+
+ for (ne = pools.names; ne != NULL; ne = nenext) {
+ nenext = ne->ne_next;
+ if (ne->ne_name)
+ free(ne->ne_name);
+ free(ne);
+ }
+
+
+ return (ret);
+}
+
+boolean_t
+find_guid(nvlist_t *nv, uint64_t guid)
+{
+ uint64_t tmp;
+ nvlist_t **child;
+ uint_t c, children;
+
+ verify(nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &tmp) == 0);
+ if (tmp == guid)
+ return (B_TRUE);
+
+ if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
+ &child, &children) == 0) {
+ for (c = 0; c < children; c++)
+ if (find_guid(child[c], guid))
+ return (B_TRUE);
+ }
+
+ return (B_FALSE);
+}
+
+typedef struct spare_cbdata {
+ uint64_t cb_guid;
+ zpool_handle_t *cb_zhp;
+} spare_cbdata_t;
+
+static int
+find_spare(zpool_handle_t *zhp, void *data)
+{
+ spare_cbdata_t *cbp = data;
+ nvlist_t **spares;
+ uint_t i, nspares;
+ uint64_t guid;
+ nvlist_t *nvroot;
+
+ verify(nvlist_lookup_nvlist(zhp->zpool_config, ZPOOL_CONFIG_VDEV_TREE,
+ &nvroot) == 0);
+
+ if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES,
+ &spares, &nspares) == 0) {
+ for (i = 0; i < nspares; i++) {
+ verify(nvlist_lookup_uint64(spares[i],
+ ZPOOL_CONFIG_GUID, &guid) == 0);
+ if (guid == cbp->cb_guid) {
+ cbp->cb_zhp = zhp;
+ return (1);
+ }
+ }
+ }
+
+ zpool_close(zhp);
+ return (0);
+}
+
+/*
+ * Determines if the pool is in use. If so, it returns true and the state of
+ * the pool as well as the name of the pool. Both strings are allocated and
+ * must be freed by the caller.
+ */
+int
+zpool_in_use(libzfs_handle_t *hdl, int fd, pool_state_t *state, char **namestr,
+ boolean_t *inuse)
+{
+ nvlist_t *config;
+ char *name;
+ boolean_t ret;
+ uint64_t guid, vdev_guid;
+ zpool_handle_t *zhp;
+ nvlist_t *pool_config;
+ uint64_t stateval, isspare;
+ spare_cbdata_t cb = { 0 };
+ boolean_t isactive;
+
+ *inuse = B_FALSE;
+
+ if (zpool_read_label(fd, &config) != 0) {
+ (void) no_memory(hdl);
+ return (-1);
+ }
+
+ if (config == NULL)
+ return (0);
+
+ verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_STATE,
+ &stateval) == 0);
+ verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_GUID,
+ &vdev_guid) == 0);
+
+ if (stateval != POOL_STATE_SPARE) {
+ verify(nvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME,
+ &name) == 0);
+ verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID,
+ &guid) == 0);
+ }
+
+ switch (stateval) {
+ case POOL_STATE_EXPORTED:
+ ret = B_TRUE;
+ break;
+
+ case POOL_STATE_ACTIVE:
+ /*
+ * For an active pool, we have to determine if it's really part
+ * of a currently active pool (in which case the pool will exist
+ * and the guid will be the same), or whether it's part of an
+ * active pool that was disconnected without being explicitly
+ * exported.
+ */
+ if (pool_active(hdl, name, guid, &isactive) != 0) {
+ nvlist_free(config);
+ return (-1);
+ }
+
+ if (isactive) {
+ /*
+ * Because the device may have been removed while
+ * offlined, we only report it as active if the vdev is
+ * still present in the config. Otherwise, pretend like
+ * it's not in use.
+ */
+ if ((zhp = zpool_open_canfail(hdl, name)) != NULL &&
+ (pool_config = zpool_get_config(zhp, NULL))
+ != NULL) {
+ nvlist_t *nvroot;
+
+ verify(nvlist_lookup_nvlist(pool_config,
+ ZPOOL_CONFIG_VDEV_TREE, &nvroot) == 0);
+ ret = find_guid(nvroot, vdev_guid);
+ } else {
+ ret = B_FALSE;
+ }
+
+ /*
+ * If this is an active spare within another pool, we
+ * treat it like an unused hot spare. This allows the
+ * user to create a pool with a hot spare that currently
+ * in use within another pool. Since we return B_TRUE,
+ * libdiskmgt will continue to prevent generic consumers
+ * from using the device.
+ */
+ if (ret && nvlist_lookup_uint64(config,
+ ZPOOL_CONFIG_IS_SPARE, &isspare) == 0 && isspare)
+ stateval = POOL_STATE_SPARE;
+
+ if (zhp != NULL)
+ zpool_close(zhp);
+ } else {
+ stateval = POOL_STATE_POTENTIALLY_ACTIVE;
+ ret = B_TRUE;
+ }
+ break;
+
+ case POOL_STATE_SPARE:
+ /*
+ * For a hot spare, it can be either definitively in use, or
+ * potentially active. To determine if it's in use, we iterate
+ * over all pools in the system and search for one with a spare
+ * with a matching guid.
+ *
+ * Due to the shared nature of spares, we don't actually report
+ * the potentially active case as in use. This means the user
+ * can freely create pools on the hot spares of exported pools,
+ * but to do otherwise makes the resulting code complicated, and
+ * we end up having to deal with this case anyway.
+ */
+ cb.cb_zhp = NULL;
+ cb.cb_guid = vdev_guid;
+ if (zpool_iter(hdl, find_spare, &cb) == 1) {
+ name = (char *)zpool_get_name(cb.cb_zhp);
+ ret = TRUE;
+ } else {
+ ret = FALSE;
+ }
+ break;
+
+ default:
+ ret = B_FALSE;
+ }
+
+
+ if (ret) {
+ if ((*namestr = zfs_strdup(hdl, name)) == NULL) {
+ nvlist_free(config);
+ return (-1);
+ }
+ *state = (pool_state_t)stateval;
+ }
+
+ if (cb.cb_zhp)
+ zpool_close(cb.cb_zhp);
+
+ nvlist_free(config);
+ *inuse = ret;
+ return (0);
+}
diff --git a/contrib/opensolaris/lib/libzfs/common/libzfs_mount.c b/contrib/opensolaris/lib/libzfs/common/libzfs_mount.c
new file mode 100644
index 0000000..b4bc945
--- /dev/null
+++ b/contrib/opensolaris/lib/libzfs/common/libzfs_mount.c
@@ -0,0 +1,986 @@
+/*
+ * 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"
+
+/*
+ * Routines to manage ZFS mounts. We separate all the nasty routines that have
+ * to deal with the OS. The following functions are the main entry points --
+ * they are used by mount and unmount and when changing a filesystem's
+ * mountpoint.
+ *
+ * zfs_is_mounted()
+ * zfs_mount()
+ * zfs_unmount()
+ * zfs_unmountall()
+ *
+ * This file also contains the functions used to manage sharing filesystems via
+ * NFS and iSCSI:
+ *
+ * zfs_is_shared()
+ * zfs_share()
+ * zfs_unshare()
+ *
+ * zfs_is_shared_nfs()
+ * zfs_share_nfs()
+ * zfs_unshare_nfs()
+ * zfs_unshareall_nfs()
+ * zfs_is_shared_iscsi()
+ * zfs_share_iscsi()
+ * zfs_unshare_iscsi()
+ *
+ * The following functions are available for pool consumers, and will
+ * mount/unmount and share/unshare all datasets within pool:
+ *
+ * zpool_enable_datasets()
+ * zpool_disable_datasets()
+ */
+
+#include <dirent.h>
+#include <dlfcn.h>
+#include <errno.h>
+#include <libgen.h>
+#include <libintl.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <strings.h>
+#include <unistd.h>
+#include <zone.h>
+#include <sys/mntent.h>
+#include <sys/mnttab.h>
+#include <sys/mount.h>
+#include <sys/stat.h>
+
+#include <libzfs.h>
+
+#include "libzfs_impl.h"
+
+static int (*iscsitgt_zfs_share)(const char *);
+static int (*iscsitgt_zfs_unshare)(const char *);
+static int (*iscsitgt_zfs_is_shared)(const char *);
+
+#pragma init(zfs_iscsi_init)
+static void
+zfs_iscsi_init(void)
+{
+ void *libiscsitgt;
+
+ if ((libiscsitgt = dlopen("/lib/libiscsitgt.so.1",
+ RTLD_LAZY | RTLD_GLOBAL)) == NULL ||
+ (iscsitgt_zfs_share = (int (*)(const char *))dlsym(libiscsitgt,
+ "iscsitgt_zfs_share")) == NULL ||
+ (iscsitgt_zfs_unshare = (int (*)(const char *))dlsym(libiscsitgt,
+ "iscsitgt_zfs_unshare")) == NULL ||
+ (iscsitgt_zfs_is_shared = (int (*)(const char *))dlsym(libiscsitgt,
+ "iscsitgt_zfs_is_shared")) == NULL) {
+ iscsitgt_zfs_share = NULL;
+ iscsitgt_zfs_unshare = NULL;
+ iscsitgt_zfs_is_shared = NULL;
+ }
+}
+
+/*
+ * Search the sharetab for the given mountpoint, returning true if it is found.
+ */
+static boolean_t
+is_shared(libzfs_handle_t *hdl, const char *mountpoint)
+{
+ char buf[MAXPATHLEN], *tab;
+
+ if (hdl->libzfs_sharetab == NULL)
+ return (0);
+
+ (void) fseek(hdl->libzfs_sharetab, 0, SEEK_SET);
+
+ while (fgets(buf, sizeof (buf), hdl->libzfs_sharetab) != NULL) {
+
+ /* the mountpoint is the first entry on each line */
+ if ((tab = strchr(buf, '\t')) != NULL) {
+ *tab = '\0';
+ if (strcmp(buf, mountpoint) == 0)
+ return (B_TRUE);
+ }
+ }
+
+ return (B_FALSE);
+}
+
+#if 0
+/*
+ * Returns true if the specified directory is empty. If we can't open the
+ * directory at all, return true so that the mount can fail with a more
+ * informative error message.
+ */
+static boolean_t
+dir_is_empty(const char *dirname)
+{
+ DIR *dirp;
+ struct dirent64 *dp;
+
+ if ((dirp = opendir(dirname)) == NULL)
+ return (B_TRUE);
+
+ while ((dp = readdir64(dirp)) != NULL) {
+
+ if (strcmp(dp->d_name, ".") == 0 ||
+ strcmp(dp->d_name, "..") == 0)
+ continue;
+
+ (void) closedir(dirp);
+ return (B_FALSE);
+ }
+
+ (void) closedir(dirp);
+ return (B_TRUE);
+}
+#endif
+
+/*
+ * Checks to see if the mount is active. If the filesystem is mounted, we fill
+ * in 'where' with the current mountpoint, and return 1. Otherwise, we return
+ * 0.
+ */
+boolean_t
+is_mounted(libzfs_handle_t *zfs_hdl, const char *special, char **where)
+{
+ struct mnttab search = { 0 }, entry;
+
+ /*
+ * Search for the entry in /etc/mnttab. We don't bother getting the
+ * mountpoint, as we can just search for the special device. This will
+ * also let us find mounts when the mountpoint is 'legacy'.
+ */
+ search.mnt_special = (char *)special;
+ search.mnt_fstype = MNTTYPE_ZFS;
+
+ rewind(zfs_hdl->libzfs_mnttab);
+ if (getmntany(zfs_hdl->libzfs_mnttab, &entry, &search) != 0)
+ return (B_FALSE);
+
+ if (where != NULL)
+ *where = zfs_strdup(zfs_hdl, entry.mnt_mountp);
+
+ return (B_TRUE);
+}
+
+boolean_t
+zfs_is_mounted(zfs_handle_t *zhp, char **where)
+{
+ return (is_mounted(zhp->zfs_hdl, zfs_get_name(zhp), where));
+}
+
+/*
+ * Returns true if the given dataset is mountable, false otherwise. Returns the
+ * mountpoint in 'buf'.
+ */
+static boolean_t
+zfs_is_mountable(zfs_handle_t *zhp, char *buf, size_t buflen,
+ zfs_source_t *source)
+{
+ char sourceloc[ZFS_MAXNAMELEN];
+ zfs_source_t sourcetype;
+
+ if (!zfs_prop_valid_for_type(ZFS_PROP_MOUNTPOINT, zhp->zfs_type))
+ return (B_FALSE);
+
+ verify(zfs_prop_get(zhp, ZFS_PROP_MOUNTPOINT, buf, buflen,
+ &sourcetype, sourceloc, sizeof (sourceloc), B_FALSE) == 0);
+
+ if (strcmp(buf, ZFS_MOUNTPOINT_NONE) == 0 ||
+ strcmp(buf, ZFS_MOUNTPOINT_LEGACY) == 0)
+ return (B_FALSE);
+
+ if (!zfs_prop_get_int(zhp, ZFS_PROP_CANMOUNT))
+ return (B_FALSE);
+
+ if (zfs_prop_get_int(zhp, ZFS_PROP_ZONED) &&
+ getzoneid() == GLOBAL_ZONEID)
+ return (B_FALSE);
+
+ if (source)
+ *source = sourcetype;
+
+ return (B_TRUE);
+}
+
+/*
+ * Mount the given filesystem.
+ */
+int
+zfs_mount(zfs_handle_t *zhp, const char *options, int flags)
+{
+ struct stat buf;
+ char mountpoint[ZFS_MAXPROPLEN];
+ char mntopts[MNT_LINE_MAX];
+ libzfs_handle_t *hdl = zhp->zfs_hdl;
+
+ if (options == NULL)
+ mntopts[0] = '\0';
+ else
+ (void) strlcpy(mntopts, options, sizeof (mntopts));
+
+ if (!zfs_is_mountable(zhp, mountpoint, sizeof (mountpoint), NULL))
+ return (0);
+
+ /* Create the directory if it doesn't already exist */
+ if (lstat(mountpoint, &buf) != 0) {
+ if (mkdirp(mountpoint, 0755) != 0) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "failed to create mountpoint"));
+ return (zfs_error_fmt(hdl, EZFS_MOUNTFAILED,
+ dgettext(TEXT_DOMAIN, "cannot mount '%s'"),
+ mountpoint));
+ }
+ }
+
+#if 0 /* FreeBSD: overlay mounts are not checked. */
+ /*
+ * Determine if the mountpoint is empty. If so, refuse to perform the
+ * mount. We don't perform this check if MS_OVERLAY is specified, which
+ * would defeat the point. We also avoid this check if 'remount' is
+ * specified.
+ */
+ if ((flags & MS_OVERLAY) == 0 &&
+ strstr(mntopts, MNTOPT_REMOUNT) == NULL &&
+ !dir_is_empty(mountpoint)) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "directory is not empty"));
+ return (zfs_error_fmt(hdl, EZFS_MOUNTFAILED,
+ dgettext(TEXT_DOMAIN, "cannot mount '%s'"), mountpoint));
+ }
+#endif
+
+ /* perform the mount */
+ if (zmount(zfs_get_name(zhp), mountpoint, flags,
+ MNTTYPE_ZFS, NULL, 0, mntopts, sizeof (mntopts)) != 0) {
+ /*
+ * Generic errors are nasty, but there are just way too many
+ * from mount(), and they're well-understood. We pick a few
+ * common ones to improve upon.
+ */
+ if (errno == EBUSY)
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "mountpoint or dataset is busy"));
+ else
+ zfs_error_aux(hdl, strerror(errno));
+
+ return (zfs_error_fmt(hdl, EZFS_MOUNTFAILED,
+ dgettext(TEXT_DOMAIN, "cannot mount '%s'"),
+ zhp->zfs_name));
+ }
+
+ return (0);
+}
+
+/*
+ * Unmount a single filesystem.
+ */
+static int
+unmount_one(libzfs_handle_t *hdl, const char *mountpoint, int flags)
+{
+ if (unmount(mountpoint, flags) != 0) {
+ zfs_error_aux(hdl, strerror(errno));
+ return (zfs_error_fmt(hdl, EZFS_UMOUNTFAILED,
+ dgettext(TEXT_DOMAIN, "cannot unmount '%s'"),
+ mountpoint));
+ }
+
+ return (0);
+}
+
+/*
+ * Unmount the given filesystem.
+ */
+int
+zfs_unmount(zfs_handle_t *zhp, const char *mountpoint, int flags)
+{
+ struct mnttab search = { 0 }, entry;
+
+ /* check to see if need to unmount the filesystem */
+ search.mnt_special = zhp->zfs_name;
+ search.mnt_fstype = MNTTYPE_ZFS;
+ rewind(zhp->zfs_hdl->libzfs_mnttab);
+ if (mountpoint != NULL || ((zfs_get_type(zhp) == ZFS_TYPE_FILESYSTEM) &&
+ getmntany(zhp->zfs_hdl->libzfs_mnttab, &entry, &search) == 0)) {
+
+ if (mountpoint == NULL)
+ mountpoint = entry.mnt_mountp;
+
+ /*
+ * Unshare and unmount the filesystem
+ */
+ if (zfs_unshare_nfs(zhp, mountpoint) != 0 ||
+ unmount_one(zhp->zfs_hdl, mountpoint, flags) != 0)
+ return (-1);
+ }
+
+ return (0);
+}
+
+/*
+ * Unmount this filesystem and any children inheriting the mountpoint property.
+ * To do this, just act like we're changing the mountpoint property, but don't
+ * remount the filesystems afterwards.
+ */
+int
+zfs_unmountall(zfs_handle_t *zhp, int flags)
+{
+ prop_changelist_t *clp;
+ int ret;
+
+ clp = changelist_gather(zhp, ZFS_PROP_MOUNTPOINT, flags);
+ if (clp == NULL)
+ return (-1);
+
+ ret = changelist_prefix(clp);
+ changelist_free(clp);
+
+ return (ret);
+}
+
+boolean_t
+zfs_is_shared(zfs_handle_t *zhp)
+{
+ if (ZFS_IS_VOLUME(zhp))
+ return (zfs_is_shared_iscsi(zhp));
+
+ return (zfs_is_shared_nfs(zhp, NULL));
+}
+
+int
+zfs_share(zfs_handle_t *zhp)
+{
+ if (ZFS_IS_VOLUME(zhp))
+ return (zfs_share_iscsi(zhp));
+
+ return (zfs_share_nfs(zhp));
+}
+
+int
+zfs_unshare(zfs_handle_t *zhp)
+{
+ if (ZFS_IS_VOLUME(zhp))
+ return (zfs_unshare_iscsi(zhp));
+
+ return (zfs_unshare_nfs(zhp, NULL));
+}
+
+/*
+ * Check to see if the filesystem is currently shared.
+ */
+boolean_t
+zfs_is_shared_nfs(zfs_handle_t *zhp, char **where)
+{
+ char *mountpoint;
+
+ if (!zfs_is_mounted(zhp, &mountpoint))
+ return (B_FALSE);
+
+ if (is_shared(zhp->zfs_hdl, mountpoint)) {
+ if (where != NULL)
+ *where = mountpoint;
+ else
+ free(mountpoint);
+ return (B_TRUE);
+ } else {
+ free(mountpoint);
+ return (B_FALSE);
+ }
+}
+
+/*
+ * Share the given filesystem according to the options in 'sharenfs'. We rely
+ * on share(1M) to the dirty work for us.
+ */
+int
+zfs_share_nfs(zfs_handle_t *zhp)
+{
+ char mountpoint[ZFS_MAXPROPLEN];
+ char shareopts[ZFS_MAXPROPLEN];
+ char buf[MAXPATHLEN];
+ FILE *fp;
+ libzfs_handle_t *hdl = zhp->zfs_hdl;
+
+ if (!zfs_is_mountable(zhp, mountpoint, sizeof (mountpoint), NULL))
+ return (0);
+
+ /*
+ * Return success if there are no share options.
+ */
+ if (zfs_prop_get(zhp, ZFS_PROP_SHARENFS, shareopts, sizeof (shareopts),
+ NULL, NULL, 0, B_FALSE) != 0 ||
+ strcmp(shareopts, "off") == 0)
+ return (0);
+
+ /*
+ * If the 'zoned' property is set, then zfs_is_mountable() will have
+ * already bailed out if we are in the global zone. But local
+ * zones cannot be NFS servers, so we ignore it for local zones as well.
+ */
+ if (zfs_prop_get_int(zhp, ZFS_PROP_ZONED))
+ return (0);
+
+#ifdef __FreeBSD__
+ {
+ int error;
+
+ if (strcmp(shareopts, "on") == 0)
+ error = fsshare(ZFS_EXPORTS_PATH, mountpoint, "");
+ else
+ error = fsshare(ZFS_EXPORTS_PATH, mountpoint, shareopts);
+ if (error != 0) {
+ zfs_error_aux(hdl, "%s", strerror(error));
+ (void) zfs_error_fmt(hdl, EZFS_SHARENFSFAILED,
+ dgettext(TEXT_DOMAIN, "cannot share '%s'"),
+ zfs_get_name(zhp));
+ return (-1);
+ }
+ }
+#else
+ /*
+ * Invoke the share(1M) command. We always do this, even if it's
+ * currently shared, as the options may have changed.
+ */
+ if (strcmp(shareopts, "on") == 0)
+ (void) snprintf(buf, sizeof (buf), "/usr/sbin/share "
+ "-F nfs \"%s\" 2>&1", mountpoint);
+ else
+ (void) snprintf(buf, sizeof (buf), "/usr/sbin/share "
+ "-F nfs -o \"%s\" \"%s\" 2>&1", shareopts,
+ mountpoint);
+
+ if ((fp = popen(buf, "r")) == NULL)
+ return (zfs_error_fmt(hdl, EZFS_SHARENFSFAILED,
+ dgettext(TEXT_DOMAIN, "cannot share '%s'"),
+ zfs_get_name(zhp)));
+
+ /*
+ * share(1M) should only produce output if there is some kind
+ * of error. All output begins with "share_nfs: ", so we trim
+ * this off to get to the real error.
+ */
+ if (fgets(buf, sizeof (buf), fp) != NULL) {
+ char *colon = strchr(buf, ':');
+
+ while (buf[strlen(buf) - 1] == '\n')
+ buf[strlen(buf) - 1] = '\0';
+
+ if (colon != NULL)
+ zfs_error_aux(hdl, colon + 2);
+
+ (void) zfs_error_fmt(hdl, EZFS_SHARENFSFAILED,
+ dgettext(TEXT_DOMAIN, "cannot share '%s'"),
+ zfs_get_name(zhp));
+
+ verify(pclose(fp) != 0);
+ return (-1);
+ }
+
+ verify(pclose(fp) == 0);
+#endif
+
+ return (0);
+}
+
+/*
+ * Unshare a filesystem by mountpoint.
+ */
+static int
+unshare_one(libzfs_handle_t *hdl, const char *name, const char *mountpoint)
+{
+ char buf[MAXPATHLEN];
+ FILE *fp;
+
+#ifdef __FreeBSD__
+ {
+ int error;
+
+ error = fsunshare(ZFS_EXPORTS_PATH, mountpoint);
+ if (error != 0) {
+ zfs_error_aux(hdl, "%s", strerror(error));
+ return (zfs_error_fmt(hdl, EZFS_UNSHARENFSFAILED,
+ dgettext(TEXT_DOMAIN,
+ "cannot unshare '%s'"), name));
+ }
+ }
+#else
+ (void) snprintf(buf, sizeof (buf),
+ "/usr/sbin/unshare \"%s\" 2>&1",
+ mountpoint);
+
+ if ((fp = popen(buf, "r")) == NULL)
+ return (zfs_error_fmt(hdl, EZFS_UNSHARENFSFAILED,
+ dgettext(TEXT_DOMAIN,
+ "cannot unshare '%s'"), name));
+
+ /*
+ * unshare(1M) should only produce output if there is
+ * some kind of error. All output begins with "unshare
+ * nfs: ", so we trim this off to get to the real error.
+ */
+ if (fgets(buf, sizeof (buf), fp) != NULL) {
+ char *colon = strchr(buf, ':');
+
+ while (buf[strlen(buf) - 1] == '\n')
+ buf[strlen(buf) - 1] = '\0';
+
+ if (colon != NULL)
+ zfs_error_aux(hdl, colon + 2);
+
+ verify(pclose(fp) != 0);
+
+ return (zfs_error_fmt(hdl, EZFS_UNSHARENFSFAILED,
+ dgettext(TEXT_DOMAIN,
+ "cannot unshare '%s'"), name));
+ }
+
+ verify(pclose(fp) == 0);
+#endif
+
+ return (0);
+}
+
+/*
+ * Unshare the given filesystem.
+ */
+int
+zfs_unshare_nfs(zfs_handle_t *zhp, const char *mountpoint)
+{
+ struct mnttab search = { 0 }, entry;
+
+ /* check to see if need to unmount the filesystem */
+ search.mnt_special = (char *)zfs_get_name(zhp);
+ search.mnt_fstype = MNTTYPE_ZFS;
+ rewind(zhp->zfs_hdl->libzfs_mnttab);
+ if (mountpoint != NULL || ((zfs_get_type(zhp) == ZFS_TYPE_FILESYSTEM) &&
+ getmntany(zhp->zfs_hdl->libzfs_mnttab, &entry, &search) == 0)) {
+
+ if (mountpoint == NULL)
+ mountpoint = entry.mnt_mountp;
+
+ if (is_shared(zhp->zfs_hdl, mountpoint) &&
+ unshare_one(zhp->zfs_hdl, zhp->zfs_name, mountpoint) != 0)
+ return (-1);
+ }
+
+ return (0);
+}
+
+/*
+ * Same as zfs_unmountall(), but for NFS unshares.
+ */
+int
+zfs_unshareall_nfs(zfs_handle_t *zhp)
+{
+ prop_changelist_t *clp;
+ int ret;
+
+ clp = changelist_gather(zhp, ZFS_PROP_SHARENFS, 0);
+ if (clp == NULL)
+ return (-1);
+
+ ret = changelist_unshare(clp);
+ changelist_free(clp);
+
+ return (ret);
+}
+
+/*
+ * Remove the mountpoint associated with the current dataset, if necessary.
+ * We only remove the underlying directory if:
+ *
+ * - The mountpoint is not 'none' or 'legacy'
+ * - The mountpoint is non-empty
+ * - The mountpoint is the default or inherited
+ * - The 'zoned' property is set, or we're in a local zone
+ *
+ * Any other directories we leave alone.
+ */
+void
+remove_mountpoint(zfs_handle_t *zhp)
+{
+ char mountpoint[ZFS_MAXPROPLEN];
+ zfs_source_t source;
+
+ if (!zfs_is_mountable(zhp, mountpoint, sizeof (mountpoint),
+ &source))
+ return;
+
+ if (source == ZFS_SRC_DEFAULT ||
+ source == ZFS_SRC_INHERITED) {
+ /*
+ * Try to remove the directory, silently ignoring any errors.
+ * The filesystem may have since been removed or moved around,
+ * and this error isn't really useful to the administrator in
+ * any way.
+ */
+ (void) rmdir(mountpoint);
+ }
+}
+
+boolean_t
+zfs_is_shared_iscsi(zfs_handle_t *zhp)
+{
+ return (iscsitgt_zfs_is_shared != NULL &&
+ iscsitgt_zfs_is_shared(zhp->zfs_name) != 0);
+}
+
+int
+zfs_share_iscsi(zfs_handle_t *zhp)
+{
+ char shareopts[ZFS_MAXPROPLEN];
+ const char *dataset = zhp->zfs_name;
+ libzfs_handle_t *hdl = zhp->zfs_hdl;
+
+ /*
+ * Return success if there are no share options.
+ */
+ if (zfs_prop_get(zhp, ZFS_PROP_SHAREISCSI, shareopts,
+ sizeof (shareopts), NULL, NULL, 0, B_FALSE) != 0 ||
+ strcmp(shareopts, "off") == 0)
+ return (0);
+
+/* We don't support iSCSI on FreeBSD yet. */
+#ifdef TODO
+ if (iscsitgt_zfs_share == NULL || iscsitgt_zfs_share(dataset) != 0)
+ return (zfs_error_fmt(hdl, EZFS_SHAREISCSIFAILED,
+ dgettext(TEXT_DOMAIN, "cannot share '%s'"), dataset));
+#endif
+
+ return (0);
+}
+
+int
+zfs_unshare_iscsi(zfs_handle_t *zhp)
+{
+ const char *dataset = zfs_get_name(zhp);
+ libzfs_handle_t *hdl = zhp->zfs_hdl;
+
+/* We don't support iSCSI on FreeBSD yet. */
+#ifdef TODO
+ /*
+ * Return if the volume is not shared
+ */
+ if (!zfs_is_shared_iscsi(zhp))
+ return (0);
+
+ /*
+ * If this fails with ENODEV it indicates that zvol wasn't shared so
+ * we should return success in that case.
+ */
+ if (iscsitgt_zfs_unshare == NULL ||
+ (iscsitgt_zfs_unshare(dataset) != 0 && errno != ENODEV))
+ return (zfs_error_fmt(hdl, EZFS_UNSHAREISCSIFAILED,
+ dgettext(TEXT_DOMAIN, "cannot unshare '%s'"), dataset));
+#endif
+
+ return (0);
+}
+
+typedef struct mount_cbdata {
+ zfs_handle_t **cb_datasets;
+ int cb_used;
+ int cb_alloc;
+} mount_cbdata_t;
+
+static int
+mount_cb(zfs_handle_t *zhp, void *data)
+{
+ mount_cbdata_t *cbp = data;
+
+ if (!(zfs_get_type(zhp) & (ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME))) {
+ zfs_close(zhp);
+ return (0);
+ }
+
+ if (cbp->cb_alloc == cbp->cb_used) {
+ void *ptr;
+
+ if ((ptr = zfs_realloc(zhp->zfs_hdl,
+ cbp->cb_datasets, cbp->cb_alloc * sizeof (void *),
+ cbp->cb_alloc * 2 * sizeof (void *))) == NULL)
+ return (-1);
+ cbp->cb_datasets = ptr;
+
+ cbp->cb_alloc *= 2;
+ }
+
+ cbp->cb_datasets[cbp->cb_used++] = zhp;
+
+ return (zfs_iter_children(zhp, mount_cb, cbp));
+}
+
+static int
+dataset_cmp(const void *a, const void *b)
+{
+ zfs_handle_t **za = (zfs_handle_t **)a;
+ zfs_handle_t **zb = (zfs_handle_t **)b;
+ char mounta[MAXPATHLEN];
+ char mountb[MAXPATHLEN];
+ boolean_t gota, gotb;
+
+ if ((gota = (zfs_get_type(*za) == ZFS_TYPE_FILESYSTEM)) != 0)
+ verify(zfs_prop_get(*za, ZFS_PROP_MOUNTPOINT, mounta,
+ sizeof (mounta), NULL, NULL, 0, B_FALSE) == 0);
+ if ((gotb = (zfs_get_type(*zb) == ZFS_TYPE_FILESYSTEM)) != 0)
+ verify(zfs_prop_get(*zb, ZFS_PROP_MOUNTPOINT, mountb,
+ sizeof (mountb), NULL, NULL, 0, B_FALSE) == 0);
+
+ if (gota && gotb)
+ return (strcmp(mounta, mountb));
+
+ if (gota)
+ return (-1);
+ if (gotb)
+ return (1);
+
+ return (strcmp(zfs_get_name(a), zfs_get_name(b)));
+}
+
+/*
+ * Mount and share all datasets within the given pool. This assumes that no
+ * datasets within the pool are currently mounted. Because users can create
+ * complicated nested hierarchies of mountpoints, we first gather all the
+ * datasets and mountpoints within the pool, and sort them by mountpoint. Once
+ * we have the list of all filesystems, we iterate over them in order and mount
+ * and/or share each one.
+ */
+#pragma weak zpool_mount_datasets = zpool_enable_datasets
+int
+zpool_enable_datasets(zpool_handle_t *zhp, const char *mntopts, int flags)
+{
+ mount_cbdata_t cb = { 0 };
+ libzfs_handle_t *hdl = zhp->zpool_hdl;
+ zfs_handle_t *zfsp;
+ int i, ret = -1;
+
+ /*
+ * Gather all datasets within the pool.
+ */
+ if ((cb.cb_datasets = zfs_alloc(hdl, 4 * sizeof (void *))) == NULL)
+ return (-1);
+ cb.cb_alloc = 4;
+
+ if ((zfsp = zfs_open(hdl, zhp->zpool_name, ZFS_TYPE_ANY)) == NULL)
+ goto out;
+
+ cb.cb_datasets[0] = zfsp;
+ cb.cb_used = 1;
+
+ if (zfs_iter_children(zfsp, mount_cb, &cb) != 0)
+ goto out;
+
+ /*
+ * Sort the datasets by mountpoint.
+ */
+ qsort(cb.cb_datasets, cb.cb_used, sizeof (void *), dataset_cmp);
+
+ /*
+ * And mount all the datasets.
+ */
+ ret = 0;
+ for (i = 0; i < cb.cb_used; i++) {
+ if (zfs_mount(cb.cb_datasets[i], mntopts, flags) != 0 ||
+ zfs_share(cb.cb_datasets[i]) != 0)
+ ret = -1;
+ }
+
+out:
+ for (i = 0; i < cb.cb_used; i++)
+ zfs_close(cb.cb_datasets[i]);
+ free(cb.cb_datasets);
+
+ return (ret);
+}
+
+
+static int
+zvol_cb(const char *dataset, void *data)
+{
+ libzfs_handle_t *hdl = data;
+ zfs_handle_t *zhp;
+
+ /*
+ * Ignore snapshots and ignore failures from non-existant datasets.
+ */
+ if (strchr(dataset, '@') != NULL ||
+ (zhp = zfs_open(hdl, dataset, ZFS_TYPE_VOLUME)) == NULL)
+ return (0);
+
+ (void) zfs_unshare_iscsi(zhp);
+
+ zfs_close(zhp);
+
+ return (0);
+}
+
+static int
+mountpoint_compare(const void *a, const void *b)
+{
+ const char *mounta = *((char **)a);
+ const char *mountb = *((char **)b);
+
+ return (strcmp(mountb, mounta));
+}
+
+/*
+ * Unshare and unmount all datasets within the given pool. We don't want to
+ * rely on traversing the DSL to discover the filesystems within the pool,
+ * because this may be expensive (if not all of them are mounted), and can fail
+ * arbitrarily (on I/O error, for example). Instead, we walk /etc/mnttab and
+ * gather all the filesystems that are currently mounted.
+ */
+#pragma weak zpool_unmount_datasets = zpool_disable_datasets
+int
+zpool_disable_datasets(zpool_handle_t *zhp, boolean_t force)
+{
+ int used, alloc;
+ struct statfs *sfs;
+ size_t namelen;
+ char **mountpoints = NULL;
+ zfs_handle_t **datasets = NULL;
+ libzfs_handle_t *hdl = zhp->zpool_hdl;
+ int i, j, n;
+ int ret = -1;
+ int flags = (force ? MS_FORCE : 0);
+
+ /*
+ * First unshare all zvols.
+ */
+ if (zpool_iter_zvol(zhp, zvol_cb, hdl) != 0)
+ return (-1);
+
+ namelen = strlen(zhp->zpool_name);
+
+ used = alloc = 0;
+ if ((n = getmntinfo(&sfs, MNT_WAIT)) == 0) {
+ fprintf(stderr, "getmntinfo(): %s\n", strerror(errno));
+ return (-1);
+ }
+ for (j = 0; j < n; j++) {
+ /*
+ * Ignore non-ZFS entries.
+ */
+ if (strcmp(sfs[j].f_fstypename, MNTTYPE_ZFS) != 0)
+ continue;
+
+ /*
+ * Ignore filesystems not within this pool.
+ */
+ if (strncmp(sfs[j].f_mntfromname, zhp->zpool_name, namelen) != 0 ||
+ (sfs[j].f_mntfromname[namelen] != '/' &&
+ sfs[j].f_mntfromname[namelen] != '\0'))
+ continue;
+
+ /*
+ * At this point we've found a filesystem within our pool. Add
+ * it to our growing list.
+ */
+ if (used == alloc) {
+ if (alloc == 0) {
+ if ((mountpoints = zfs_alloc(hdl,
+ 8 * sizeof (void *))) == NULL)
+ goto out;
+
+ if ((datasets = zfs_alloc(hdl,
+ 8 * sizeof (void *))) == NULL)
+ goto out;
+
+ alloc = 8;
+ } else {
+ void *ptr;
+
+ if ((ptr = zfs_realloc(hdl, mountpoints,
+ alloc * sizeof (void *),
+ alloc * 2 * sizeof (void *))) == NULL)
+ goto out;
+ mountpoints = ptr;
+
+ if ((ptr = zfs_realloc(hdl, datasets,
+ alloc * sizeof (void *),
+ alloc * 2 * sizeof (void *))) == NULL)
+ goto out;
+ datasets = ptr;
+
+ alloc *= 2;
+ }
+ }
+
+ if ((mountpoints[used] = zfs_strdup(hdl,
+ sfs[j].f_mntonname)) == NULL)
+ goto out;
+
+ /*
+ * This is allowed to fail, in case there is some I/O error. It
+ * is only used to determine if we need to remove the underlying
+ * mountpoint, so failure is not fatal.
+ */
+ datasets[used] = make_dataset_handle(hdl, sfs[j].f_mntfromname);
+
+ used++;
+ }
+
+ /*
+ * At this point, we have the entire list of filesystems, so sort it by
+ * mountpoint.
+ */
+ qsort(mountpoints, used, sizeof (char *), mountpoint_compare);
+
+ /*
+ * Walk through and first unshare everything.
+ */
+ for (i = 0; i < used; i++) {
+ if (is_shared(hdl, mountpoints[i]) &&
+ unshare_one(hdl, mountpoints[i], mountpoints[i]) != 0)
+ goto out;
+ }
+
+ /*
+ * Now unmount everything, removing the underlying directories as
+ * appropriate.
+ */
+ for (i = 0; i < used; i++) {
+ if (unmount_one(hdl, mountpoints[i], flags) != 0)
+ goto out;
+ }
+
+ for (i = 0; i < used; i++) {
+ if (datasets[i])
+ remove_mountpoint(datasets[i]);
+ }
+
+ ret = 0;
+out:
+ for (i = 0; i < used; i++) {
+ if (datasets[i])
+ zfs_close(datasets[i]);
+ free(mountpoints[i]);
+ }
+ free(datasets);
+ free(mountpoints);
+
+ return (ret);
+}
diff --git a/contrib/opensolaris/lib/libzfs/common/libzfs_pool.c b/contrib/opensolaris/lib/libzfs/common/libzfs_pool.c
new file mode 100644
index 0000000..8580837
--- /dev/null
+++ b/contrib/opensolaris/lib/libzfs/common/libzfs_pool.c
@@ -0,0 +1,2055 @@
+/*
+ * 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"
+
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <assert.h>
+#include <ctype.h>
+#include <errno.h>
+#include <devid.h>
+#include <dirent.h>
+#include <fcntl.h>
+#include <libintl.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <strings.h>
+#include <unistd.h>
+#include <sys/zfs_ioctl.h>
+#include <sys/zio.h>
+#include <strings.h>
+#include <umem.h>
+
+#include "zfs_namecheck.h"
+#include "zfs_prop.h"
+#include "libzfs_impl.h"
+
+/*
+ * Validate the given pool name, optionally putting an extended error message in
+ * 'buf'.
+ */
+static boolean_t
+zpool_name_valid(libzfs_handle_t *hdl, boolean_t isopen, const char *pool)
+{
+ namecheck_err_t why;
+ char what;
+ int ret;
+
+ ret = pool_namecheck(pool, &why, &what);
+
+ /*
+ * The rules for reserved pool names were extended at a later point.
+ * But we need to support users with existing pools that may now be
+ * invalid. So we only check for this expanded set of names during a
+ * create (or import), and only in userland.
+ */
+ if (ret == 0 && !isopen &&
+ (strncmp(pool, "mirror", 6) == 0 ||
+ strncmp(pool, "raidz", 5) == 0 ||
+ strncmp(pool, "spare", 5) == 0)) {
+ zfs_error_aux(hdl,
+ dgettext(TEXT_DOMAIN, "name is reserved"));
+ return (B_FALSE);
+ }
+
+
+ if (ret != 0) {
+ if (hdl != NULL) {
+ switch (why) {
+ case NAME_ERR_TOOLONG:
+ zfs_error_aux(hdl,
+ dgettext(TEXT_DOMAIN, "name is too long"));
+ break;
+
+ case NAME_ERR_INVALCHAR:
+ zfs_error_aux(hdl,
+ dgettext(TEXT_DOMAIN, "invalid character "
+ "'%c' in pool name"), what);
+ break;
+
+ case NAME_ERR_NOLETTER:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "name must begin with a letter"));
+ break;
+
+ case NAME_ERR_RESERVED:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "name is reserved"));
+ break;
+
+ case NAME_ERR_DISKLIKE:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "pool name is reserved"));
+ break;
+
+ case NAME_ERR_LEADING_SLASH:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "leading slash in name"));
+ break;
+
+ case NAME_ERR_EMPTY_COMPONENT:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "empty component in name"));
+ break;
+
+ case NAME_ERR_TRAILING_SLASH:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "trailing slash in name"));
+ break;
+
+ case NAME_ERR_MULTIPLE_AT:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "multiple '@' delimiters in name"));
+ break;
+
+ }
+ }
+ return (B_FALSE);
+ }
+
+ return (B_TRUE);
+}
+
+static int
+zpool_get_all_props(zpool_handle_t *zhp)
+{
+ zfs_cmd_t zc = { 0 };
+ libzfs_handle_t *hdl = zhp->zpool_hdl;
+
+ (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
+
+ if (zcmd_alloc_dst_nvlist(hdl, &zc, 0) != 0)
+ return (-1);
+
+ while (ioctl(hdl->libzfs_fd, ZFS_IOC_POOL_GET_PROPS, &zc) != 0) {
+ if (errno == ENOMEM) {
+ if (zcmd_expand_dst_nvlist(hdl, &zc) != 0) {
+ zcmd_free_nvlists(&zc);
+ return (-1);
+ }
+ } else {
+ zcmd_free_nvlists(&zc);
+ return (-1);
+ }
+ }
+
+ if (zcmd_read_dst_nvlist(hdl, &zc, &zhp->zpool_props) != 0) {
+ zcmd_free_nvlists(&zc);
+ return (-1);
+ }
+
+ zcmd_free_nvlists(&zc);
+
+ return (0);
+}
+
+/*
+ * Open a handle to the given pool, even if the pool is currently in the FAULTED
+ * state.
+ */
+zpool_handle_t *
+zpool_open_canfail(libzfs_handle_t *hdl, const char *pool)
+{
+ zpool_handle_t *zhp;
+ boolean_t missing;
+
+ /*
+ * Make sure the pool name is valid.
+ */
+ if (!zpool_name_valid(hdl, B_TRUE, pool)) {
+ (void) zfs_error_fmt(hdl, EZFS_INVALIDNAME,
+ dgettext(TEXT_DOMAIN, "cannot open '%s'"),
+ pool);
+ return (NULL);
+ }
+
+ if ((zhp = zfs_alloc(hdl, sizeof (zpool_handle_t))) == NULL)
+ return (NULL);
+
+ zhp->zpool_hdl = hdl;
+ (void) strlcpy(zhp->zpool_name, pool, sizeof (zhp->zpool_name));
+
+ if (zpool_refresh_stats(zhp, &missing) != 0) {
+ zpool_close(zhp);
+ return (NULL);
+ }
+
+ if (missing) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "no such pool"));
+ (void) zfs_error_fmt(hdl, EZFS_NOENT,
+ dgettext(TEXT_DOMAIN, "cannot open '%s'"),
+ pool);
+ zpool_close(zhp);
+ return (NULL);
+ }
+
+ return (zhp);
+}
+
+/*
+ * Like the above, but silent on error. Used when iterating over pools (because
+ * the configuration cache may be out of date).
+ */
+int
+zpool_open_silent(libzfs_handle_t *hdl, const char *pool, zpool_handle_t **ret)
+{
+ zpool_handle_t *zhp;
+ boolean_t missing;
+
+ if ((zhp = zfs_alloc(hdl, sizeof (zpool_handle_t))) == NULL)
+ return (-1);
+
+ zhp->zpool_hdl = hdl;
+ (void) strlcpy(zhp->zpool_name, pool, sizeof (zhp->zpool_name));
+
+ if (zpool_refresh_stats(zhp, &missing) != 0) {
+ zpool_close(zhp);
+ return (-1);
+ }
+
+ if (missing) {
+ zpool_close(zhp);
+ *ret = NULL;
+ return (0);
+ }
+
+ *ret = zhp;
+ return (0);
+}
+
+/*
+ * Similar to zpool_open_canfail(), but refuses to open pools in the faulted
+ * state.
+ */
+zpool_handle_t *
+zpool_open(libzfs_handle_t *hdl, const char *pool)
+{
+ zpool_handle_t *zhp;
+
+ if ((zhp = zpool_open_canfail(hdl, pool)) == NULL)
+ return (NULL);
+
+ if (zhp->zpool_state == POOL_STATE_UNAVAIL) {
+ (void) zfs_error_fmt(hdl, EZFS_POOLUNAVAIL,
+ dgettext(TEXT_DOMAIN, "cannot open '%s'"), zhp->zpool_name);
+ zpool_close(zhp);
+ return (NULL);
+ }
+
+ return (zhp);
+}
+
+/*
+ * Close the handle. Simply frees the memory associated with the handle.
+ */
+void
+zpool_close(zpool_handle_t *zhp)
+{
+ if (zhp->zpool_config)
+ nvlist_free(zhp->zpool_config);
+ if (zhp->zpool_old_config)
+ nvlist_free(zhp->zpool_old_config);
+ if (zhp->zpool_props)
+ nvlist_free(zhp->zpool_props);
+ free(zhp);
+}
+
+/*
+ * Return the name of the pool.
+ */
+const char *
+zpool_get_name(zpool_handle_t *zhp)
+{
+ return (zhp->zpool_name);
+}
+
+/*
+ * Return the GUID of the pool.
+ */
+uint64_t
+zpool_get_guid(zpool_handle_t *zhp)
+{
+ uint64_t guid;
+
+ verify(nvlist_lookup_uint64(zhp->zpool_config, ZPOOL_CONFIG_POOL_GUID,
+ &guid) == 0);
+ return (guid);
+}
+
+/*
+ * Return the version of the pool.
+ */
+uint64_t
+zpool_get_version(zpool_handle_t *zhp)
+{
+ uint64_t version;
+
+ verify(nvlist_lookup_uint64(zhp->zpool_config, ZPOOL_CONFIG_VERSION,
+ &version) == 0);
+
+ return (version);
+}
+
+/*
+ * Return the amount of space currently consumed by the pool.
+ */
+uint64_t
+zpool_get_space_used(zpool_handle_t *zhp)
+{
+ nvlist_t *nvroot;
+ vdev_stat_t *vs;
+ uint_t vsc;
+
+ verify(nvlist_lookup_nvlist(zhp->zpool_config, ZPOOL_CONFIG_VDEV_TREE,
+ &nvroot) == 0);
+ verify(nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_STATS,
+ (uint64_t **)&vs, &vsc) == 0);
+
+ return (vs->vs_alloc);
+}
+
+/*
+ * Return the total space in the pool.
+ */
+uint64_t
+zpool_get_space_total(zpool_handle_t *zhp)
+{
+ nvlist_t *nvroot;
+ vdev_stat_t *vs;
+ uint_t vsc;
+
+ verify(nvlist_lookup_nvlist(zhp->zpool_config, ZPOOL_CONFIG_VDEV_TREE,
+ &nvroot) == 0);
+ verify(nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_STATS,
+ (uint64_t **)&vs, &vsc) == 0);
+
+ return (vs->vs_space);
+}
+
+/*
+ * Return the alternate root for this pool, if any.
+ */
+int
+zpool_get_root(zpool_handle_t *zhp, char *buf, size_t buflen)
+{
+ zfs_cmd_t zc = { 0 };
+
+ (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
+ if (ioctl(zhp->zpool_hdl->libzfs_fd, ZFS_IOC_OBJSET_STATS, &zc) != 0 ||
+ zc.zc_value[0] == '\0')
+ return (-1);
+
+ (void) strlcpy(buf, zc.zc_value, buflen);
+
+ return (0);
+}
+
+/*
+ * Return the state of the pool (ACTIVE or UNAVAILABLE)
+ */
+int
+zpool_get_state(zpool_handle_t *zhp)
+{
+ return (zhp->zpool_state);
+}
+
+/*
+ * Create the named pool, using the provided vdev list. It is assumed
+ * that the consumer has already validated the contents of the nvlist, so we
+ * don't have to worry about error semantics.
+ */
+int
+zpool_create(libzfs_handle_t *hdl, const char *pool, nvlist_t *nvroot,
+ const char *altroot)
+{
+ zfs_cmd_t zc = { 0 };
+ char msg[1024];
+
+ (void) snprintf(msg, sizeof (msg), dgettext(TEXT_DOMAIN,
+ "cannot create '%s'"), pool);
+
+ if (!zpool_name_valid(hdl, B_FALSE, pool))
+ return (zfs_error(hdl, EZFS_INVALIDNAME, msg));
+
+ if (altroot != NULL && altroot[0] != '/')
+ return (zfs_error_fmt(hdl, EZFS_BADPATH,
+ dgettext(TEXT_DOMAIN, "bad alternate root '%s'"), altroot));
+
+ if (zcmd_write_src_nvlist(hdl, &zc, nvroot, NULL) != 0)
+ return (-1);
+
+ (void) strlcpy(zc.zc_name, pool, sizeof (zc.zc_name));
+
+ if (altroot != NULL)
+ (void) strlcpy(zc.zc_value, altroot, sizeof (zc.zc_value));
+
+ if (ioctl(hdl->libzfs_fd, ZFS_IOC_POOL_CREATE, &zc) != 0) {
+ zcmd_free_nvlists(&zc);
+
+ switch (errno) {
+ case EBUSY:
+ /*
+ * This can happen if the user has specified the same
+ * device multiple times. We can't reliably detect this
+ * until we try to add it and see we already have a
+ * label.
+ */
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "one or more vdevs refer to the same device"));
+ return (zfs_error(hdl, EZFS_BADDEV, msg));
+
+ case EOVERFLOW:
+ /*
+ * This occurs when one of the devices is below
+ * SPA_MINDEVSIZE. Unfortunately, we can't detect which
+ * device was the problem device since there's no
+ * reliable way to determine device size from userland.
+ */
+ {
+ char buf[64];
+
+ zfs_nicenum(SPA_MINDEVSIZE, buf, sizeof (buf));
+
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "one or more devices is less than the "
+ "minimum size (%s)"), buf);
+ }
+ return (zfs_error(hdl, EZFS_BADDEV, msg));
+
+ case ENOSPC:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "one or more devices is out of space"));
+ return (zfs_error(hdl, EZFS_BADDEV, msg));
+
+ default:
+ return (zpool_standard_error(hdl, errno, msg));
+ }
+ }
+
+ zcmd_free_nvlists(&zc);
+
+ /*
+ * If this is an alternate root pool, then we automatically set the
+ * mountpoint of the root dataset to be '/'.
+ */
+ if (altroot != NULL) {
+ zfs_handle_t *zhp;
+
+ verify((zhp = zfs_open(hdl, pool, ZFS_TYPE_ANY)) != NULL);
+ verify(zfs_prop_set(zhp, zfs_prop_to_name(ZFS_PROP_MOUNTPOINT),
+ "/") == 0);
+
+ zfs_close(zhp);
+ }
+
+ return (0);
+}
+
+/*
+ * Destroy the given pool. It is up to the caller to ensure that there are no
+ * datasets left in the pool.
+ */
+int
+zpool_destroy(zpool_handle_t *zhp)
+{
+ zfs_cmd_t zc = { 0 };
+ zfs_handle_t *zfp = NULL;
+ libzfs_handle_t *hdl = zhp->zpool_hdl;
+ char msg[1024];
+
+ if (zhp->zpool_state == POOL_STATE_ACTIVE &&
+ (zfp = zfs_open(zhp->zpool_hdl, zhp->zpool_name,
+ ZFS_TYPE_FILESYSTEM)) == NULL)
+ return (-1);
+
+ if (zpool_remove_zvol_links(zhp) != 0)
+ return (-1);
+
+ (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
+
+ if (ioctl(zhp->zpool_hdl->libzfs_fd, ZFS_IOC_POOL_DESTROY, &zc) != 0) {
+ (void) snprintf(msg, sizeof (msg), dgettext(TEXT_DOMAIN,
+ "cannot destroy '%s'"), zhp->zpool_name);
+
+ if (errno == EROFS) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "one or more devices is read only"));
+ (void) zfs_error(hdl, EZFS_BADDEV, msg);
+ } else {
+ (void) zpool_standard_error(hdl, errno, msg);
+ }
+
+ if (zfp)
+ zfs_close(zfp);
+ return (-1);
+ }
+
+ if (zfp) {
+ remove_mountpoint(zfp);
+ zfs_close(zfp);
+ }
+
+ return (0);
+}
+
+/*
+ * Add the given vdevs to the pool. The caller must have already performed the
+ * necessary verification to ensure that the vdev specification is well-formed.
+ */
+int
+zpool_add(zpool_handle_t *zhp, nvlist_t *nvroot)
+{
+ zfs_cmd_t zc = { 0 };
+ int ret;
+ libzfs_handle_t *hdl = zhp->zpool_hdl;
+ char msg[1024];
+ nvlist_t **spares;
+ uint_t nspares;
+
+ (void) snprintf(msg, sizeof (msg), dgettext(TEXT_DOMAIN,
+ "cannot add to '%s'"), zhp->zpool_name);
+
+ if (zpool_get_version(zhp) < ZFS_VERSION_SPARES &&
+ nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES,
+ &spares, &nspares) == 0) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "pool must be "
+ "upgraded to add hot spares"));
+ return (zfs_error(hdl, EZFS_BADVERSION, msg));
+ }
+
+ if (zcmd_write_src_nvlist(hdl, &zc, nvroot, NULL) != 0)
+ return (-1);
+ (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
+
+ if (ioctl(zhp->zpool_hdl->libzfs_fd, ZFS_IOC_VDEV_ADD, &zc) != 0) {
+ switch (errno) {
+ case EBUSY:
+ /*
+ * This can happen if the user has specified the same
+ * device multiple times. We can't reliably detect this
+ * until we try to add it and see we already have a
+ * label.
+ */
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "one or more vdevs refer to the same device"));
+ (void) zfs_error(hdl, EZFS_BADDEV, msg);
+ break;
+
+ case EOVERFLOW:
+ /*
+ * This occurrs when one of the devices is below
+ * SPA_MINDEVSIZE. Unfortunately, we can't detect which
+ * device was the problem device since there's no
+ * reliable way to determine device size from userland.
+ */
+ {
+ char buf[64];
+
+ zfs_nicenum(SPA_MINDEVSIZE, buf, sizeof (buf));
+
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "device is less than the minimum "
+ "size (%s)"), buf);
+ }
+ (void) zfs_error(hdl, EZFS_BADDEV, msg);
+ break;
+
+ case ENOTSUP:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "pool must be upgraded to add raidz2 vdevs"));
+ (void) zfs_error(hdl, EZFS_BADVERSION, msg);
+ break;
+
+ case EDOM:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "root pool can not have concatenated devices"));
+ (void) zfs_error(hdl, EZFS_POOL_NOTSUP, msg);
+ break;
+
+ default:
+ (void) zpool_standard_error(hdl, errno, msg);
+ }
+
+ ret = -1;
+ } else {
+ ret = 0;
+ }
+
+ zcmd_free_nvlists(&zc);
+
+ return (ret);
+}
+
+/*
+ * Exports the pool from the system. The caller must ensure that there are no
+ * mounted datasets in the pool.
+ */
+int
+zpool_export(zpool_handle_t *zhp)
+{
+ zfs_cmd_t zc = { 0 };
+
+ if (zpool_remove_zvol_links(zhp) != 0)
+ return (-1);
+
+ (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
+
+ if (ioctl(zhp->zpool_hdl->libzfs_fd, ZFS_IOC_POOL_EXPORT, &zc) != 0)
+ return (zpool_standard_error_fmt(zhp->zpool_hdl, errno,
+ dgettext(TEXT_DOMAIN, "cannot export '%s'"),
+ zhp->zpool_name));
+ return (0);
+}
+
+/*
+ * Import the given pool using the known configuration. The configuration
+ * should have come from zpool_find_import(). The 'newname' and 'altroot'
+ * parameters control whether the pool is imported with a different name or with
+ * an alternate root, respectively.
+ */
+int
+zpool_import(libzfs_handle_t *hdl, nvlist_t *config, const char *newname,
+ const char *altroot)
+{
+ zfs_cmd_t zc = { 0 };
+ char *thename;
+ char *origname;
+ int ret;
+
+ verify(nvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME,
+ &origname) == 0);
+
+ if (newname != NULL) {
+ if (!zpool_name_valid(hdl, B_FALSE, newname))
+ return (zfs_error_fmt(hdl, EZFS_INVALIDNAME,
+ dgettext(TEXT_DOMAIN, "cannot import '%s'"),
+ newname));
+ thename = (char *)newname;
+ } else {
+ thename = origname;
+ }
+
+ if (altroot != NULL && altroot[0] != '/')
+ return (zfs_error_fmt(hdl, EZFS_BADPATH,
+ dgettext(TEXT_DOMAIN, "bad alternate root '%s'"),
+ altroot));
+
+ (void) strlcpy(zc.zc_name, thename, sizeof (zc.zc_name));
+
+ if (altroot != NULL)
+ (void) strlcpy(zc.zc_value, altroot, sizeof (zc.zc_value));
+ else
+ zc.zc_value[0] = '\0';
+
+ verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID,
+ &zc.zc_guid) == 0);
+
+ if (zcmd_write_src_nvlist(hdl, &zc, config, NULL) != 0)
+ return (-1);
+
+ ret = 0;
+ if (ioctl(hdl->libzfs_fd, ZFS_IOC_POOL_IMPORT, &zc) != 0) {
+ char desc[1024];
+ if (newname == NULL)
+ (void) snprintf(desc, sizeof (desc),
+ dgettext(TEXT_DOMAIN, "cannot import '%s'"),
+ thename);
+ else
+ (void) snprintf(desc, sizeof (desc),
+ dgettext(TEXT_DOMAIN, "cannot import '%s' as '%s'"),
+ origname, thename);
+
+ switch (errno) {
+ case ENOTSUP:
+ /*
+ * Unsupported version.
+ */
+ (void) zfs_error(hdl, EZFS_BADVERSION, desc);
+ break;
+
+ case EINVAL:
+ (void) zfs_error(hdl, EZFS_INVALCONFIG, desc);
+ break;
+
+ default:
+ (void) zpool_standard_error(hdl, errno, desc);
+ }
+
+ ret = -1;
+ } else {
+ zpool_handle_t *zhp;
+ /*
+ * This should never fail, but play it safe anyway.
+ */
+ if (zpool_open_silent(hdl, thename, &zhp) != 0) {
+ ret = -1;
+ } else if (zhp != NULL) {
+ ret = zpool_create_zvol_links(zhp);
+ zpool_close(zhp);
+ }
+ }
+
+ zcmd_free_nvlists(&zc);
+ return (ret);
+}
+
+/*
+ * Scrub the pool.
+ */
+int
+zpool_scrub(zpool_handle_t *zhp, pool_scrub_type_t type)
+{
+ zfs_cmd_t zc = { 0 };
+ char msg[1024];
+ libzfs_handle_t *hdl = zhp->zpool_hdl;
+
+ (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
+ zc.zc_cookie = type;
+
+ if (ioctl(zhp->zpool_hdl->libzfs_fd, ZFS_IOC_POOL_SCRUB, &zc) == 0)
+ return (0);
+
+ (void) snprintf(msg, sizeof (msg),
+ dgettext(TEXT_DOMAIN, "cannot scrub %s"), zc.zc_name);
+
+ if (errno == EBUSY)
+ return (zfs_error(hdl, EZFS_RESILVERING, msg));
+ else
+ return (zpool_standard_error(hdl, errno, msg));
+}
+
+/*
+ * 'avail_spare' is set to TRUE if the provided guid refers to an AVAIL
+ * spare; but FALSE if its an INUSE spare.
+ */
+static nvlist_t *
+vdev_to_nvlist_iter(nvlist_t *nv, const char *search, uint64_t guid,
+ boolean_t *avail_spare)
+{
+ uint_t c, children;
+ nvlist_t **child;
+ uint64_t theguid, present;
+ char *path;
+ uint64_t wholedisk = 0;
+ nvlist_t *ret;
+
+ verify(nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &theguid) == 0);
+
+ if (search == NULL &&
+ nvlist_lookup_uint64(nv, ZPOOL_CONFIG_NOT_PRESENT, &present) == 0) {
+ /*
+ * If the device has never been present since import, the only
+ * reliable way to match the vdev is by GUID.
+ */
+ if (theguid == guid)
+ return (nv);
+ } else if (search != NULL &&
+ nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &path) == 0) {
+ (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_WHOLE_DISK,
+ &wholedisk);
+ if (wholedisk) {
+ /*
+ * For whole disks, the internal path has 's0', but the
+ * path passed in by the user doesn't.
+ */
+ if (strlen(search) == strlen(path) - 2 &&
+ strncmp(search, path, strlen(search)) == 0)
+ return (nv);
+ } else if (strcmp(search, path) == 0) {
+ return (nv);
+ }
+ }
+
+ if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
+ &child, &children) != 0)
+ return (NULL);
+
+ for (c = 0; c < children; c++)
+ if ((ret = vdev_to_nvlist_iter(child[c], search, guid,
+ avail_spare)) != NULL)
+ return (ret);
+
+ if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_SPARES,
+ &child, &children) == 0) {
+ for (c = 0; c < children; c++) {
+ if ((ret = vdev_to_nvlist_iter(child[c], search, guid,
+ avail_spare)) != NULL) {
+ *avail_spare = B_TRUE;
+ return (ret);
+ }
+ }
+ }
+
+ return (NULL);
+}
+
+nvlist_t *
+zpool_find_vdev(zpool_handle_t *zhp, const char *path, boolean_t *avail_spare)
+{
+ char buf[MAXPATHLEN];
+ const char *search;
+ char *end;
+ nvlist_t *nvroot;
+ uint64_t guid;
+
+ guid = strtoull(path, &end, 10);
+ if (guid != 0 && *end == '\0') {
+ search = NULL;
+ } else if (path[0] != '/') {
+ (void) snprintf(buf, sizeof (buf), "%s%s", _PATH_DEV, path);
+ search = buf;
+ } else {
+ search = path;
+ }
+
+ verify(nvlist_lookup_nvlist(zhp->zpool_config, ZPOOL_CONFIG_VDEV_TREE,
+ &nvroot) == 0);
+
+ *avail_spare = B_FALSE;
+ return (vdev_to_nvlist_iter(nvroot, search, guid, avail_spare));
+}
+
+/*
+ * Returns TRUE if the given guid corresponds to a spare (INUSE or not).
+ */
+static boolean_t
+is_spare(zpool_handle_t *zhp, uint64_t guid)
+{
+ uint64_t spare_guid;
+ nvlist_t *nvroot;
+ nvlist_t **spares;
+ uint_t nspares;
+ int i;
+
+ verify(nvlist_lookup_nvlist(zhp->zpool_config, ZPOOL_CONFIG_VDEV_TREE,
+ &nvroot) == 0);
+ if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES,
+ &spares, &nspares) == 0) {
+ for (i = 0; i < nspares; i++) {
+ verify(nvlist_lookup_uint64(spares[i],
+ ZPOOL_CONFIG_GUID, &spare_guid) == 0);
+ if (guid == spare_guid)
+ return (B_TRUE);
+ }
+ }
+
+ return (B_FALSE);
+}
+
+/*
+ * Bring the specified vdev online
+ */
+int
+zpool_vdev_online(zpool_handle_t *zhp, const char *path)
+{
+ zfs_cmd_t zc = { 0 };
+ char msg[1024];
+ nvlist_t *tgt;
+ boolean_t avail_spare;
+ libzfs_handle_t *hdl = zhp->zpool_hdl;
+
+ (void) snprintf(msg, sizeof (msg),
+ dgettext(TEXT_DOMAIN, "cannot online %s"), path);
+
+ (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
+ if ((tgt = zpool_find_vdev(zhp, path, &avail_spare)) == NULL)
+ return (zfs_error(hdl, EZFS_NODEVICE, msg));
+
+ verify(nvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID, &zc.zc_guid) == 0);
+
+ if (avail_spare || is_spare(zhp, zc.zc_guid) == B_TRUE)
+ return (zfs_error(hdl, EZFS_ISSPARE, msg));
+
+ if (ioctl(zhp->zpool_hdl->libzfs_fd, ZFS_IOC_VDEV_ONLINE, &zc) == 0)
+ return (0);
+
+ return (zpool_standard_error(hdl, errno, msg));
+}
+
+/*
+ * Take the specified vdev offline
+ */
+int
+zpool_vdev_offline(zpool_handle_t *zhp, const char *path, int istmp)
+{
+ zfs_cmd_t zc = { 0 };
+ char msg[1024];
+ nvlist_t *tgt;
+ boolean_t avail_spare;
+ libzfs_handle_t *hdl = zhp->zpool_hdl;
+
+ (void) snprintf(msg, sizeof (msg),
+ dgettext(TEXT_DOMAIN, "cannot offline %s"), path);
+
+ (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
+ if ((tgt = zpool_find_vdev(zhp, path, &avail_spare)) == NULL)
+ return (zfs_error(hdl, EZFS_NODEVICE, msg));
+
+ verify(nvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID, &zc.zc_guid) == 0);
+
+ if (avail_spare || is_spare(zhp, zc.zc_guid) == B_TRUE)
+ return (zfs_error(hdl, EZFS_ISSPARE, msg));
+
+ zc.zc_cookie = istmp;
+
+ if (ioctl(zhp->zpool_hdl->libzfs_fd, ZFS_IOC_VDEV_OFFLINE, &zc) == 0)
+ return (0);
+
+ switch (errno) {
+ case EBUSY:
+
+ /*
+ * There are no other replicas of this device.
+ */
+ return (zfs_error(hdl, EZFS_NOREPLICAS, msg));
+
+ default:
+ return (zpool_standard_error(hdl, errno, msg));
+ }
+}
+
+/*
+ * Returns TRUE if the given nvlist is a vdev that was originally swapped in as
+ * a hot spare.
+ */
+static boolean_t
+is_replacing_spare(nvlist_t *search, nvlist_t *tgt, int which)
+{
+ nvlist_t **child;
+ uint_t c, children;
+ char *type;
+
+ if (nvlist_lookup_nvlist_array(search, ZPOOL_CONFIG_CHILDREN, &child,
+ &children) == 0) {
+ verify(nvlist_lookup_string(search, ZPOOL_CONFIG_TYPE,
+ &type) == 0);
+
+ if (strcmp(type, VDEV_TYPE_SPARE) == 0 &&
+ children == 2 && child[which] == tgt)
+ return (B_TRUE);
+
+ for (c = 0; c < children; c++)
+ if (is_replacing_spare(child[c], tgt, which))
+ return (B_TRUE);
+ }
+
+ return (B_FALSE);
+}
+
+/*
+ * Attach new_disk (fully described by nvroot) to old_disk.
+ * If 'replacing' is specified, tne new disk will replace the old one.
+ */
+int
+zpool_vdev_attach(zpool_handle_t *zhp,
+ const char *old_disk, const char *new_disk, nvlist_t *nvroot, int replacing)
+{
+ zfs_cmd_t zc = { 0 };
+ char msg[1024];
+ int ret;
+ nvlist_t *tgt;
+ boolean_t avail_spare;
+ uint64_t val;
+ char *path;
+ nvlist_t **child;
+ uint_t children;
+ nvlist_t *config_root;
+ libzfs_handle_t *hdl = zhp->zpool_hdl;
+
+ if (replacing)
+ (void) snprintf(msg, sizeof (msg), dgettext(TEXT_DOMAIN,
+ "cannot replace %s with %s"), old_disk, new_disk);
+ else
+ (void) snprintf(msg, sizeof (msg), dgettext(TEXT_DOMAIN,
+ "cannot attach %s to %s"), new_disk, old_disk);
+
+ (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
+ if ((tgt = zpool_find_vdev(zhp, old_disk, &avail_spare)) == 0)
+ return (zfs_error(hdl, EZFS_NODEVICE, msg));
+
+ if (avail_spare)
+ return (zfs_error(hdl, EZFS_ISSPARE, msg));
+
+ verify(nvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID, &zc.zc_guid) == 0);
+ zc.zc_cookie = replacing;
+
+ if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN,
+ &child, &children) != 0 || children != 1) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "new device must be a single disk"));
+ return (zfs_error(hdl, EZFS_INVALCONFIG, msg));
+ }
+
+ verify(nvlist_lookup_nvlist(zpool_get_config(zhp, NULL),
+ ZPOOL_CONFIG_VDEV_TREE, &config_root) == 0);
+
+ /*
+ * If the target is a hot spare that has been swapped in, we can only
+ * replace it with another hot spare.
+ */
+ if (replacing &&
+ nvlist_lookup_uint64(tgt, ZPOOL_CONFIG_IS_SPARE, &val) == 0 &&
+ nvlist_lookup_string(child[0], ZPOOL_CONFIG_PATH, &path) == 0 &&
+ (zpool_find_vdev(zhp, path, &avail_spare) == NULL ||
+ !avail_spare) && is_replacing_spare(config_root, tgt, 1)) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "can only be replaced by another hot spare"));
+ return (zfs_error(hdl, EZFS_BADTARGET, msg));
+ }
+
+ /*
+ * If we are attempting to replace a spare, it canot be applied to an
+ * already spared device.
+ */
+ if (replacing &&
+ nvlist_lookup_string(child[0], ZPOOL_CONFIG_PATH, &path) == 0 &&
+ zpool_find_vdev(zhp, path, &avail_spare) != NULL && avail_spare &&
+ is_replacing_spare(config_root, tgt, 0)) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "device has already been replaced with a spare"));
+ return (zfs_error(hdl, EZFS_BADTARGET, msg));
+ }
+
+ if (zcmd_write_src_nvlist(hdl, &zc, nvroot, NULL) != 0)
+ return (-1);
+
+ ret = ioctl(zhp->zpool_hdl->libzfs_fd, ZFS_IOC_VDEV_ATTACH, &zc);
+
+ zcmd_free_nvlists(&zc);
+
+ if (ret == 0)
+ return (0);
+
+ switch (errno) {
+ case ENOTSUP:
+ /*
+ * Can't attach to or replace this type of vdev.
+ */
+ if (replacing)
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "cannot replace a replacing device"));
+ else
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "can only attach to mirrors and top-level "
+ "disks"));
+ (void) zfs_error(hdl, EZFS_BADTARGET, msg);
+ break;
+
+ case EINVAL:
+ /*
+ * The new device must be a single disk.
+ */
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "new device must be a single disk"));
+ (void) zfs_error(hdl, EZFS_INVALCONFIG, msg);
+ break;
+
+ case EBUSY:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "%s is busy"),
+ new_disk);
+ (void) zfs_error(hdl, EZFS_BADDEV, msg);
+ break;
+
+ case EOVERFLOW:
+ /*
+ * The new device is too small.
+ */
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "device is too small"));
+ (void) zfs_error(hdl, EZFS_BADDEV, msg);
+ break;
+
+ case EDOM:
+ /*
+ * The new device has a different alignment requirement.
+ */
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "devices have different sector alignment"));
+ (void) zfs_error(hdl, EZFS_BADDEV, msg);
+ break;
+
+ case ENAMETOOLONG:
+ /*
+ * The resulting top-level vdev spec won't fit in the label.
+ */
+ (void) zfs_error(hdl, EZFS_DEVOVERFLOW, msg);
+ break;
+
+ default:
+ (void) zpool_standard_error(hdl, errno, msg);
+ }
+
+ return (-1);
+}
+
+/*
+ * Detach the specified device.
+ */
+int
+zpool_vdev_detach(zpool_handle_t *zhp, const char *path)
+{
+ zfs_cmd_t zc = { 0 };
+ char msg[1024];
+ nvlist_t *tgt;
+ boolean_t avail_spare;
+ libzfs_handle_t *hdl = zhp->zpool_hdl;
+
+ (void) snprintf(msg, sizeof (msg),
+ dgettext(TEXT_DOMAIN, "cannot detach %s"), path);
+
+ (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
+ if ((tgt = zpool_find_vdev(zhp, path, &avail_spare)) == 0)
+ return (zfs_error(hdl, EZFS_NODEVICE, msg));
+
+ if (avail_spare)
+ return (zfs_error(hdl, EZFS_ISSPARE, msg));
+
+ verify(nvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID, &zc.zc_guid) == 0);
+
+ if (ioctl(hdl->libzfs_fd, ZFS_IOC_VDEV_DETACH, &zc) == 0)
+ return (0);
+
+ switch (errno) {
+
+ case ENOTSUP:
+ /*
+ * Can't detach from this type of vdev.
+ */
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "only "
+ "applicable to mirror and replacing vdevs"));
+ (void) zfs_error(zhp->zpool_hdl, EZFS_BADTARGET, msg);
+ break;
+
+ case EBUSY:
+ /*
+ * There are no other replicas of this device.
+ */
+ (void) zfs_error(hdl, EZFS_NOREPLICAS, msg);
+ break;
+
+ default:
+ (void) zpool_standard_error(hdl, errno, msg);
+ }
+
+ return (-1);
+}
+
+/*
+ * Remove the given device. Currently, this is supported only for hot spares.
+ */
+int
+zpool_vdev_remove(zpool_handle_t *zhp, const char *path)
+{
+ zfs_cmd_t zc = { 0 };
+ char msg[1024];
+ nvlist_t *tgt;
+ boolean_t avail_spare;
+ libzfs_handle_t *hdl = zhp->zpool_hdl;
+
+ (void) snprintf(msg, sizeof (msg),
+ dgettext(TEXT_DOMAIN, "cannot remove %s"), path);
+
+ (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
+ if ((tgt = zpool_find_vdev(zhp, path, &avail_spare)) == 0)
+ return (zfs_error(hdl, EZFS_NODEVICE, msg));
+
+ if (!avail_spare) {
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "only inactive hot spares can be removed"));
+ return (zfs_error(hdl, EZFS_NODEVICE, msg));
+ }
+
+ verify(nvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID, &zc.zc_guid) == 0);
+
+ if (ioctl(hdl->libzfs_fd, ZFS_IOC_VDEV_REMOVE, &zc) == 0)
+ return (0);
+
+ return (zpool_standard_error(hdl, errno, msg));
+}
+
+/*
+ * Clear the errors for the pool, or the particular device if specified.
+ */
+int
+zpool_clear(zpool_handle_t *zhp, const char *path)
+{
+ zfs_cmd_t zc = { 0 };
+ char msg[1024];
+ nvlist_t *tgt;
+ boolean_t avail_spare;
+ libzfs_handle_t *hdl = zhp->zpool_hdl;
+
+ if (path)
+ (void) snprintf(msg, sizeof (msg),
+ dgettext(TEXT_DOMAIN, "cannot clear errors for %s"),
+ path);
+ else
+ (void) snprintf(msg, sizeof (msg),
+ dgettext(TEXT_DOMAIN, "cannot clear errors for %s"),
+ zhp->zpool_name);
+
+ (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
+ if (path) {
+ if ((tgt = zpool_find_vdev(zhp, path, &avail_spare)) == 0)
+ return (zfs_error(hdl, EZFS_NODEVICE, msg));
+
+ if (avail_spare)
+ return (zfs_error(hdl, EZFS_ISSPARE, msg));
+
+ verify(nvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID,
+ &zc.zc_guid) == 0);
+ }
+
+ if (ioctl(hdl->libzfs_fd, ZFS_IOC_CLEAR, &zc) == 0)
+ return (0);
+
+ return (zpool_standard_error(hdl, errno, msg));
+}
+
+/*
+ * Iterate over all zvols in a given pool by walking the /dev/zvol/dsk/<pool>
+ * hierarchy.
+ */
+int
+zpool_iter_zvol(zpool_handle_t *zhp, int (*cb)(const char *, void *),
+ void *data)
+{
+ libzfs_handle_t *hdl = zhp->zpool_hdl;
+ char (*paths)[MAXPATHLEN];
+ char path[MAXPATHLEN];
+ size_t size = 4;
+ int curr, fd, base, ret = 0;
+ DIR *dirp;
+ struct dirent *dp;
+ struct stat st;
+
+ if ((base = open(ZVOL_FULL_DEV_DIR, O_RDONLY)) < 0)
+ return (errno == ENOENT ? 0 : -1);
+
+ snprintf(path, sizeof(path), "%s/%s", ZVOL_FULL_DEV_DIR,
+ zhp->zpool_name);
+ if (stat(path, &st) != 0) {
+ int err = errno;
+ (void) close(base);
+ return (err == ENOENT ? 0 : -1);
+ }
+
+ /*
+ * Oddly this wasn't a directory -- ignore that failure since we
+ * know there are no links lower in the (non-existant) hierarchy.
+ */
+ if (!S_ISDIR(st.st_mode)) {
+ (void) close(base);
+ return (0);
+ }
+
+ if ((paths = zfs_alloc(hdl, size * sizeof (paths[0]))) == NULL) {
+ (void) close(base);
+ return (-1);
+ }
+
+ (void) strlcpy(paths[0], zhp->zpool_name, sizeof (paths[0]));
+ curr = 0;
+
+ while (curr >= 0) {
+ snprintf(path, sizeof(path), "%s/%s", ZVOL_FULL_DEV_DIR,
+ paths[curr]);
+ if (lstat(path, &st) != 0)
+ goto err;
+
+ if (S_ISDIR(st.st_mode)) {
+ if ((dirp = opendir(path)) == NULL) {
+ goto err;
+ }
+
+ while ((dp = readdir(dirp)) != NULL) {
+ if (dp->d_name[0] == '.')
+ continue;
+
+ if (curr + 1 == size) {
+ paths = zfs_realloc(hdl, paths,
+ size * sizeof (paths[0]),
+ size * 2 * sizeof (paths[0]));
+ if (paths == NULL) {
+ (void) closedir(dirp);
+ goto err;
+ }
+
+ size *= 2;
+ }
+
+ (void) strlcpy(paths[curr + 1], paths[curr],
+ sizeof (paths[curr + 1]));
+ (void) strlcat(paths[curr], "/",
+ sizeof (paths[curr]));
+ (void) strlcat(paths[curr], dp->d_name,
+ sizeof (paths[curr]));
+ curr++;
+ }
+
+ (void) closedir(dirp);
+
+ } else {
+ if ((ret = cb(paths[curr], data)) != 0)
+ break;
+ }
+
+ curr--;
+ }
+
+ free(paths);
+ (void) close(base);
+
+ return (ret);
+
+err:
+ free(paths);
+ (void) close(base);
+ return (-1);
+}
+
+typedef struct zvol_cb {
+ zpool_handle_t *zcb_pool;
+ boolean_t zcb_create;
+} zvol_cb_t;
+
+/*ARGSUSED*/
+static int
+do_zvol_create(zfs_handle_t *zhp, void *data)
+{
+ int ret;
+
+ if (ZFS_IS_VOLUME(zhp))
+ (void) zvol_create_link(zhp->zfs_hdl, zhp->zfs_name);
+
+ ret = zfs_iter_children(zhp, do_zvol_create, NULL);
+
+ zfs_close(zhp);
+
+ return (ret);
+}
+
+/*
+ * Iterate over all zvols in the pool and make any necessary minor nodes.
+ */
+int
+zpool_create_zvol_links(zpool_handle_t *zhp)
+{
+ zfs_handle_t *zfp;
+ int ret;
+
+ /*
+ * If the pool is unavailable, just return success.
+ */
+ if ((zfp = make_dataset_handle(zhp->zpool_hdl,
+ zhp->zpool_name)) == NULL)
+ return (0);
+
+ ret = zfs_iter_children(zfp, do_zvol_create, NULL);
+
+ zfs_close(zfp);
+ return (ret);
+}
+
+static int
+do_zvol_remove(const char *dataset, void *data)
+{
+ zpool_handle_t *zhp = data;
+
+ return (zvol_remove_link(zhp->zpool_hdl, dataset));
+}
+
+/*
+ * Iterate over all zvols in the pool and remove any minor nodes. We iterate
+ * by examining the /dev links so that a corrupted pool doesn't impede this
+ * operation.
+ */
+int
+zpool_remove_zvol_links(zpool_handle_t *zhp)
+{
+ return (zpool_iter_zvol(zhp, do_zvol_remove, zhp));
+}
+
+/*
+ * Convert from a devid string to a path.
+ */
+static char *
+devid_to_path(char *devid_str)
+{
+ ddi_devid_t devid;
+ char *minor;
+ char *path;
+ devid_nmlist_t *list = NULL;
+ int ret;
+
+ if (devid_str_decode(devid_str, &devid, &minor) != 0)
+ return (NULL);
+
+ ret = devid_deviceid_to_nmlist("/dev", devid, minor, &list);
+
+ devid_str_free(minor);
+ devid_free(devid);
+
+ if (ret != 0)
+ return (NULL);
+
+ if ((path = strdup(list[0].devname)) == NULL)
+ return (NULL);
+
+ devid_free_nmlist(list);
+
+ return (path);
+}
+
+/*
+ * Convert from a path to a devid string.
+ */
+static char *
+path_to_devid(const char *path)
+{
+ int fd;
+ ddi_devid_t devid;
+ char *minor, *ret;
+
+ if ((fd = open(path, O_RDONLY)) < 0)
+ return (NULL);
+
+ minor = NULL;
+ ret = NULL;
+ if (devid_get(fd, &devid) == 0) {
+ if (devid_get_minor_name(fd, &minor) == 0)
+ ret = devid_str_encode(devid, minor);
+ if (minor != NULL)
+ devid_str_free(minor);
+ devid_free(devid);
+ }
+ (void) close(fd);
+
+ return (ret);
+}
+
+/*
+ * Issue the necessary ioctl() to update the stored path value for the vdev. We
+ * ignore any failure here, since a common case is for an unprivileged user to
+ * type 'zpool status', and we'll display the correct information anyway.
+ */
+static void
+set_path(zpool_handle_t *zhp, nvlist_t *nv, const char *path)
+{
+ zfs_cmd_t zc = { 0 };
+
+ (void) strncpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
+ (void) strncpy(zc.zc_value, path, sizeof (zc.zc_value));
+ verify(nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID,
+ &zc.zc_guid) == 0);
+
+ (void) ioctl(zhp->zpool_hdl->libzfs_fd, ZFS_IOC_VDEV_SETPATH, &zc);
+}
+
+/*
+ * Given a vdev, return the name to display in iostat. If the vdev has a path,
+ * we use that, stripping off any leading "/dev/dsk/"; if not, we use the type.
+ * We also check if this is a whole disk, in which case we strip off the
+ * trailing 's0' slice name.
+ *
+ * This routine is also responsible for identifying when disks have been
+ * reconfigured in a new location. The kernel will have opened the device by
+ * devid, but the path will still refer to the old location. To catch this, we
+ * first do a path -> devid translation (which is fast for the common case). If
+ * the devid matches, we're done. If not, we do a reverse devid -> path
+ * translation and issue the appropriate ioctl() to update the path of the vdev.
+ * If 'zhp' is NULL, then this is an exported pool, and we don't need to do any
+ * of these checks.
+ */
+char *
+zpool_vdev_name(libzfs_handle_t *hdl, zpool_handle_t *zhp, nvlist_t *nv)
+{
+ char *path, *devid;
+ uint64_t value;
+ char buf[64];
+
+ if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_NOT_PRESENT,
+ &value) == 0) {
+ verify(nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID,
+ &value) == 0);
+ (void) snprintf(buf, sizeof (buf), "%llu",
+ (u_longlong_t)value);
+ path = buf;
+ } else if (nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &path) == 0) {
+
+ if (zhp != NULL &&
+ nvlist_lookup_string(nv, ZPOOL_CONFIG_DEVID, &devid) == 0) {
+ /*
+ * Determine if the current path is correct.
+ */
+ char *newdevid = path_to_devid(path);
+
+ if (newdevid == NULL ||
+ strcmp(devid, newdevid) != 0) {
+ char *newpath;
+
+ if ((newpath = devid_to_path(devid)) != NULL) {
+ /*
+ * Update the path appropriately.
+ */
+ set_path(zhp, nv, newpath);
+ if (nvlist_add_string(nv,
+ ZPOOL_CONFIG_PATH, newpath) == 0)
+ verify(nvlist_lookup_string(nv,
+ ZPOOL_CONFIG_PATH,
+ &path) == 0);
+ free(newpath);
+ }
+ }
+
+ if (newdevid)
+ devid_str_free(newdevid);
+ }
+
+ if (strncmp(path, _PATH_DEV, sizeof(_PATH_DEV) - 1) == 0)
+ path += sizeof(_PATH_DEV) - 1;
+
+ if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_WHOLE_DISK,
+ &value) == 0 && value) {
+ char *tmp = zfs_strdup(hdl, path);
+ if (tmp == NULL)
+ return (NULL);
+ tmp[strlen(path) - 2] = '\0';
+ return (tmp);
+ }
+ } else {
+ verify(nvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE, &path) == 0);
+
+ /*
+ * If it's a raidz device, we need to stick in the parity level.
+ */
+ if (strcmp(path, VDEV_TYPE_RAIDZ) == 0) {
+ verify(nvlist_lookup_uint64(nv, ZPOOL_CONFIG_NPARITY,
+ &value) == 0);
+ (void) snprintf(buf, sizeof (buf), "%s%llu", path,
+ (u_longlong_t)value);
+ path = buf;
+ }
+ }
+
+ return (zfs_strdup(hdl, path));
+}
+
+static int
+zbookmark_compare(const void *a, const void *b)
+{
+ return (memcmp(a, b, sizeof (zbookmark_t)));
+}
+
+/*
+ * Retrieve the persistent error log, uniquify the members, and return to the
+ * caller.
+ */
+int
+zpool_get_errlog(zpool_handle_t *zhp, nvlist_t **nverrlistp)
+{
+ zfs_cmd_t zc = { 0 };
+ uint64_t count;
+ zbookmark_t *zb = NULL;
+ int i;
+
+ /*
+ * Retrieve the raw error list from the kernel. If the number of errors
+ * has increased, allocate more space and continue until we get the
+ * entire list.
+ */
+ verify(nvlist_lookup_uint64(zhp->zpool_config, ZPOOL_CONFIG_ERRCOUNT,
+ &count) == 0);
+ if ((zc.zc_nvlist_dst = (uintptr_t)zfs_alloc(zhp->zpool_hdl,
+ count * sizeof (zbookmark_t))) == (uintptr_t)NULL)
+ return (-1);
+ zc.zc_nvlist_dst_size = count;
+ (void) strcpy(zc.zc_name, zhp->zpool_name);
+ for (;;) {
+ if (ioctl(zhp->zpool_hdl->libzfs_fd, ZFS_IOC_ERROR_LOG,
+ &zc) != 0) {
+ free((void *)(uintptr_t)zc.zc_nvlist_dst);
+ if (errno == ENOMEM) {
+ count = zc.zc_nvlist_dst_size;
+ if ((zc.zc_nvlist_dst = (uintptr_t)
+ zfs_alloc(zhp->zpool_hdl, count *
+ sizeof (zbookmark_t))) == (uintptr_t)NULL)
+ return (-1);
+ } else {
+ return (-1);
+ }
+ } else {
+ break;
+ }
+ }
+
+ /*
+ * Sort the resulting bookmarks. This is a little confusing due to the
+ * implementation of ZFS_IOC_ERROR_LOG. The bookmarks are copied last
+ * to first, and 'zc_nvlist_dst_size' indicates the number of boomarks
+ * _not_ copied as part of the process. So we point the start of our
+ * array appropriate and decrement the total number of elements.
+ */
+ zb = ((zbookmark_t *)(uintptr_t)zc.zc_nvlist_dst) +
+ zc.zc_nvlist_dst_size;
+ count -= zc.zc_nvlist_dst_size;
+
+ qsort(zb, count, sizeof (zbookmark_t), zbookmark_compare);
+
+ verify(nvlist_alloc(nverrlistp, 0, KM_SLEEP) == 0);
+
+ /*
+ * Fill in the nverrlistp with nvlist's of dataset and object numbers.
+ */
+ for (i = 0; i < count; i++) {
+ nvlist_t *nv;
+
+ /* ignoring zb_blkid and zb_level for now */
+ if (i > 0 && zb[i-1].zb_objset == zb[i].zb_objset &&
+ zb[i-1].zb_object == zb[i].zb_object)
+ continue;
+
+ if (nvlist_alloc(&nv, NV_UNIQUE_NAME, KM_SLEEP) != 0)
+ goto nomem;
+ if (nvlist_add_uint64(nv, ZPOOL_ERR_DATASET,
+ zb[i].zb_objset) != 0) {
+ nvlist_free(nv);
+ goto nomem;
+ }
+ if (nvlist_add_uint64(nv, ZPOOL_ERR_OBJECT,
+ zb[i].zb_object) != 0) {
+ nvlist_free(nv);
+ goto nomem;
+ }
+ if (nvlist_add_nvlist(*nverrlistp, "ejk", nv) != 0) {
+ nvlist_free(nv);
+ goto nomem;
+ }
+ nvlist_free(nv);
+ }
+
+ free((void *)(uintptr_t)zc.zc_nvlist_dst);
+ return (0);
+
+nomem:
+ free((void *)(uintptr_t)zc.zc_nvlist_dst);
+ return (no_memory(zhp->zpool_hdl));
+}
+
+/*
+ * Upgrade a ZFS pool to the latest on-disk version.
+ */
+int
+zpool_upgrade(zpool_handle_t *zhp)
+{
+ zfs_cmd_t zc = { 0 };
+ libzfs_handle_t *hdl = zhp->zpool_hdl;
+
+ (void) strcpy(zc.zc_name, zhp->zpool_name);
+ if (ioctl(hdl->libzfs_fd, ZFS_IOC_POOL_UPGRADE, &zc) != 0)
+ return (zpool_standard_error_fmt(hdl, errno,
+ dgettext(TEXT_DOMAIN, "cannot upgrade '%s'"),
+ zhp->zpool_name));
+
+ return (0);
+}
+
+/*
+ * Log command history.
+ *
+ * 'pool' is B_TRUE if we are logging a command for 'zpool'; B_FALSE
+ * otherwise ('zfs'). 'pool_create' is B_TRUE if we are logging the creation
+ * of the pool; B_FALSE otherwise. 'path' is the pathanme containing the
+ * poolname. 'argc' and 'argv' are used to construct the command string.
+ */
+void
+zpool_log_history(libzfs_handle_t *hdl, int argc, char **argv, const char *path,
+ boolean_t pool, boolean_t pool_create)
+{
+ char cmd_buf[HIS_MAX_RECORD_LEN];
+ char *dspath;
+ zfs_cmd_t zc = { 0 };
+ int i;
+
+ /* construct the command string */
+ (void) strcpy(cmd_buf, pool ? "zpool" : "zfs");
+ for (i = 0; i < argc; i++) {
+ if (strlen(cmd_buf) + 1 + strlen(argv[i]) > HIS_MAX_RECORD_LEN)
+ break;
+ (void) strcat(cmd_buf, " ");
+ (void) strcat(cmd_buf, argv[i]);
+ }
+
+ /* figure out the poolname */
+ dspath = strpbrk(path, "/@");
+ if (dspath == NULL) {
+ (void) strcpy(zc.zc_name, path);
+ } else {
+ (void) strncpy(zc.zc_name, path, dspath - path);
+ zc.zc_name[dspath-path] = '\0';
+ }
+
+ zc.zc_history = (uint64_t)(uintptr_t)cmd_buf;
+ zc.zc_history_len = strlen(cmd_buf);
+
+ /* overloading zc_history_offset */
+ zc.zc_history_offset = pool_create;
+
+ (void) ioctl(hdl->libzfs_fd, ZFS_IOC_POOL_LOG_HISTORY, &zc);
+}
+
+/*
+ * Perform ioctl to get some command history of a pool.
+ *
+ * 'buf' is the buffer to fill up to 'len' bytes. 'off' is the
+ * logical offset of the history buffer to start reading from.
+ *
+ * Upon return, 'off' is the next logical offset to read from and
+ * 'len' is the actual amount of bytes read into 'buf'.
+ */
+static int
+get_history(zpool_handle_t *zhp, char *buf, uint64_t *off, uint64_t *len)
+{
+ zfs_cmd_t zc = { 0 };
+ libzfs_handle_t *hdl = zhp->zpool_hdl;
+
+ (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
+
+ zc.zc_history = (uint64_t)(uintptr_t)buf;
+ zc.zc_history_len = *len;
+ zc.zc_history_offset = *off;
+
+ if (ioctl(hdl->libzfs_fd, ZFS_IOC_POOL_GET_HISTORY, &zc) != 0) {
+ switch (errno) {
+ case EPERM:
+ return (zfs_error_fmt(hdl, EZFS_PERM,
+ dgettext(TEXT_DOMAIN,
+ "cannot show history for pool '%s'"),
+ zhp->zpool_name));
+ case ENOENT:
+ return (zfs_error_fmt(hdl, EZFS_NOHISTORY,
+ dgettext(TEXT_DOMAIN, "cannot get history for pool "
+ "'%s'"), zhp->zpool_name));
+ case ENOTSUP:
+ return (zfs_error_fmt(hdl, EZFS_BADVERSION,
+ dgettext(TEXT_DOMAIN, "cannot get history for pool "
+ "'%s', pool must be upgraded"), zhp->zpool_name));
+ default:
+ return (zpool_standard_error_fmt(hdl, errno,
+ dgettext(TEXT_DOMAIN,
+ "cannot get history for '%s'"), zhp->zpool_name));
+ }
+ }
+
+ *len = zc.zc_history_len;
+ *off = zc.zc_history_offset;
+
+ return (0);
+}
+
+/*
+ * Process the buffer of nvlists, unpacking and storing each nvlist record
+ * into 'records'. 'leftover' is set to the number of bytes that weren't
+ * processed as there wasn't a complete record.
+ */
+static int
+zpool_history_unpack(char *buf, uint64_t bytes_read, uint64_t *leftover,
+ nvlist_t ***records, uint_t *numrecords)
+{
+ uint64_t reclen;
+ nvlist_t *nv;
+ int i;
+
+ while (bytes_read > sizeof (reclen)) {
+
+ /* get length of packed record (stored as little endian) */
+ for (i = 0, reclen = 0; i < sizeof (reclen); i++)
+ reclen += (uint64_t)(((uchar_t *)buf)[i]) << (8*i);
+
+ if (bytes_read < sizeof (reclen) + reclen)
+ break;
+
+ /* unpack record */
+ if (nvlist_unpack(buf + sizeof (reclen), reclen, &nv, 0) != 0)
+ return (ENOMEM);
+ bytes_read -= sizeof (reclen) + reclen;
+ buf += sizeof (reclen) + reclen;
+
+ /* add record to nvlist array */
+ (*numrecords)++;
+ if (ISP2(*numrecords + 1)) {
+ *records = realloc(*records,
+ *numrecords * 2 * sizeof (nvlist_t *));
+ }
+ (*records)[*numrecords - 1] = nv;
+ }
+
+ *leftover = bytes_read;
+ return (0);
+}
+
+#define HIS_BUF_LEN (128*1024)
+
+/*
+ * Retrieve the command history of a pool.
+ */
+int
+zpool_get_history(zpool_handle_t *zhp, nvlist_t **nvhisp)
+{
+ char buf[HIS_BUF_LEN];
+ uint64_t off = 0;
+ nvlist_t **records = NULL;
+ uint_t numrecords = 0;
+ int err, i;
+
+ do {
+ uint64_t bytes_read = sizeof (buf);
+ uint64_t leftover;
+
+ if ((err = get_history(zhp, buf, &off, &bytes_read)) != 0)
+ break;
+
+ /* if nothing else was read in, we're at EOF, just return */
+ if (!bytes_read)
+ break;
+
+ if ((err = zpool_history_unpack(buf, bytes_read,
+ &leftover, &records, &numrecords)) != 0)
+ break;
+ off -= leftover;
+
+ /* CONSTCOND */
+ } while (1);
+
+ if (!err) {
+ verify(nvlist_alloc(nvhisp, NV_UNIQUE_NAME, 0) == 0);
+ verify(nvlist_add_nvlist_array(*nvhisp, ZPOOL_HIST_RECORD,
+ records, numrecords) == 0);
+ }
+ for (i = 0; i < numrecords; i++)
+ nvlist_free(records[i]);
+ free(records);
+
+ return (err);
+}
+
+void
+zpool_obj_to_path(zpool_handle_t *zhp, uint64_t dsobj, uint64_t obj,
+ char *pathname, size_t len)
+{
+ zfs_cmd_t zc = { 0 };
+ boolean_t mounted = B_FALSE;
+ char *mntpnt = NULL;
+ char dsname[MAXNAMELEN];
+
+ if (dsobj == 0) {
+ /* special case for the MOS */
+ (void) snprintf(pathname, len, "<metadata>:<0x%llx>", obj);
+ return;
+ }
+
+ /* get the dataset's name */
+ (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
+ zc.zc_obj = dsobj;
+ if (ioctl(zhp->zpool_hdl->libzfs_fd,
+ ZFS_IOC_DSOBJ_TO_DSNAME, &zc) != 0) {
+ /* just write out a path of two object numbers */
+ (void) snprintf(pathname, len, "<0x%llx>:<0x%llx>",
+ dsobj, obj);
+ return;
+ }
+ (void) strlcpy(dsname, zc.zc_value, sizeof (dsname));
+
+ /* find out if the dataset is mounted */
+ mounted = is_mounted(zhp->zpool_hdl, dsname, &mntpnt);
+
+ /* get the corrupted object's path */
+ (void) strlcpy(zc.zc_name, dsname, sizeof (zc.zc_name));
+ zc.zc_obj = obj;
+ if (ioctl(zhp->zpool_hdl->libzfs_fd, ZFS_IOC_OBJ_TO_PATH,
+ &zc) == 0) {
+ if (mounted) {
+ (void) snprintf(pathname, len, "%s%s", mntpnt,
+ zc.zc_value);
+ } else {
+ (void) snprintf(pathname, len, "%s:%s",
+ dsname, zc.zc_value);
+ }
+ } else {
+ (void) snprintf(pathname, len, "%s:<0x%llx>", dsname, obj);
+ }
+ free(mntpnt);
+}
+
+int
+zpool_set_prop(zpool_handle_t *zhp, const char *propname, const char *propval)
+{
+ zfs_cmd_t zc = { 0 };
+ int ret = -1;
+ char errbuf[1024];
+ nvlist_t *nvl = NULL;
+ nvlist_t *realprops;
+
+ (void) snprintf(errbuf, sizeof (errbuf),
+ dgettext(TEXT_DOMAIN, "cannot set property for '%s'"),
+ zhp->zpool_name);
+
+ if (zpool_get_version(zhp) < ZFS_VERSION_BOOTFS) {
+ zfs_error_aux(zhp->zpool_hdl,
+ dgettext(TEXT_DOMAIN, "pool must be "
+ "upgraded to support pool properties"));
+ return (zfs_error(zhp->zpool_hdl, EZFS_BADVERSION, errbuf));
+ }
+
+ if (zhp->zpool_props == NULL && zpool_get_all_props(zhp))
+ return (zfs_error(zhp->zpool_hdl, EZFS_POOLPROPS, errbuf));
+
+ if (nvlist_alloc(&nvl, NV_UNIQUE_NAME, 0) != 0 ||
+ nvlist_add_string(nvl, propname, propval) != 0) {
+ return (no_memory(zhp->zpool_hdl));
+ }
+
+ if ((realprops = zfs_validate_properties(zhp->zpool_hdl, ZFS_TYPE_POOL,
+ zhp->zpool_name, nvl, 0, NULL, errbuf)) == NULL) {
+ nvlist_free(nvl);
+ return (-1);
+ }
+
+ nvlist_free(nvl);
+ nvl = realprops;
+
+ /*
+ * Execute the corresponding ioctl() to set this property.
+ */
+ (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
+
+ if (zcmd_write_src_nvlist(zhp->zpool_hdl, &zc, nvl, NULL) != 0)
+ return (-1);
+
+ ret = ioctl(zhp->zpool_hdl->libzfs_fd, ZFS_IOC_POOL_SET_PROPS, &zc);
+ zcmd_free_nvlists(&zc);
+
+ if (ret)
+ (void) zpool_standard_error(zhp->zpool_hdl, errno, errbuf);
+
+ return (ret);
+}
+
+int
+zpool_get_prop(zpool_handle_t *zhp, zpool_prop_t prop, char *propbuf,
+ size_t proplen, zfs_source_t *srctype)
+{
+ uint64_t value;
+ char msg[1024], *strvalue;
+ nvlist_t *nvp;
+ zfs_source_t src = ZFS_SRC_NONE;
+
+ (void) snprintf(msg, sizeof (msg), dgettext(TEXT_DOMAIN,
+ "cannot get property '%s'"), zpool_prop_to_name(prop));
+
+ if (zpool_get_version(zhp) < ZFS_VERSION_BOOTFS) {
+ zfs_error_aux(zhp->zpool_hdl,
+ dgettext(TEXT_DOMAIN, "pool must be "
+ "upgraded to support pool properties"));
+ return (zfs_error(zhp->zpool_hdl, EZFS_BADVERSION, msg));
+ }
+
+ if (zhp->zpool_props == NULL && zpool_get_all_props(zhp))
+ return (zfs_error(zhp->zpool_hdl, EZFS_POOLPROPS, msg));
+
+ /*
+ * the "name" property is special cased
+ */
+ if (!zfs_prop_valid_for_type(prop, ZFS_TYPE_POOL) &&
+ prop != ZFS_PROP_NAME)
+ return (-1);
+
+ switch (prop) {
+ case ZFS_PROP_NAME:
+ (void) strlcpy(propbuf, zhp->zpool_name, proplen);
+ break;
+
+ case ZFS_PROP_BOOTFS:
+ if (nvlist_lookup_nvlist(zhp->zpool_props,
+ zpool_prop_to_name(prop), &nvp) != 0) {
+ strvalue = (char *)zfs_prop_default_string(prop);
+ if (strvalue == NULL)
+ strvalue = "-";
+ src = ZFS_SRC_DEFAULT;
+ } else {
+ VERIFY(nvlist_lookup_uint64(nvp,
+ ZFS_PROP_SOURCE, &value) == 0);
+ src = value;
+ VERIFY(nvlist_lookup_string(nvp, ZFS_PROP_VALUE,
+ &strvalue) == 0);
+ if (strlen(strvalue) >= proplen)
+ return (-1);
+ }
+ (void) strcpy(propbuf, strvalue);
+ break;
+
+ default:
+ return (-1);
+ }
+ if (srctype)
+ *srctype = src;
+ return (0);
+}
+
+int
+zpool_get_proplist(libzfs_handle_t *hdl, char *fields, zpool_proplist_t **listp)
+{
+ return (zfs_get_proplist_common(hdl, fields, listp, ZFS_TYPE_POOL));
+}
+
+
+int
+zpool_expand_proplist(zpool_handle_t *zhp, zpool_proplist_t **plp)
+{
+ libzfs_handle_t *hdl = zhp->zpool_hdl;
+ zpool_proplist_t *entry;
+ char buf[ZFS_MAXPROPLEN];
+
+ if (zfs_expand_proplist_common(hdl, plp, ZFS_TYPE_POOL) != 0)
+ return (-1);
+
+ for (entry = *plp; entry != NULL; entry = entry->pl_next) {
+
+ if (entry->pl_fixed)
+ continue;
+
+ if (entry->pl_prop != ZFS_PROP_INVAL &&
+ zpool_get_prop(zhp, entry->pl_prop, buf, sizeof (buf),
+ NULL) == 0) {
+ if (strlen(buf) > entry->pl_width)
+ entry->pl_width = strlen(buf);
+ }
+ }
+
+ return (0);
+}
diff --git a/contrib/opensolaris/lib/libzfs/common/libzfs_status.c b/contrib/opensolaris/lib/libzfs/common/libzfs_status.c
new file mode 100644
index 0000000..2a41649
--- /dev/null
+++ b/contrib/opensolaris/lib/libzfs/common/libzfs_status.c
@@ -0,0 +1,289 @@
+/*
+ * 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 2006 Sun Microsystems, Inc. All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#pragma ident "%Z%%M% %I% %E% SMI"
+
+/*
+ * This file contains the functions which analyze the status of a pool. This
+ * include both the status of an active pool, as well as the status exported
+ * pools. Returns one of the ZPOOL_STATUS_* defines describing the status of
+ * the pool. This status is independent (to a certain degree) from the state of
+ * the pool. A pool's state descsribes only whether or not it is capable of
+ * providing the necessary fault tolerance for data. The status describes the
+ * overall status of devices. A pool that is online can still have a device
+ * that is experiencing errors.
+ *
+ * Only a subset of the possible faults can be detected using 'zpool status',
+ * and not all possible errors correspond to a FMA message ID. The explanation
+ * is left up to the caller, depending on whether it is a live pool or an
+ * import.
+ */
+
+#include <libzfs.h>
+#include <string.h>
+#include "libzfs_impl.h"
+
+/*
+ * Message ID table. This must be kep in sync with the ZPOOL_STATUS_* defines
+ * in libzfs.h. Note that there are some status results which go past the end
+ * of this table, and hence have no associated message ID.
+ */
+static char *msgid_table[] = {
+ "ZFS-8000-14",
+ "ZFS-8000-2Q",
+ "ZFS-8000-3C",
+ "ZFS-8000-4J",
+ "ZFS-8000-5E",
+ "ZFS-8000-6X",
+ "ZFS-8000-72",
+ "ZFS-8000-8A",
+ "ZFS-8000-9P",
+ "ZFS-8000-A5"
+};
+
+/*
+ * If the pool is active, a certain class of static errors is overridden by the
+ * faults as analayzed by FMA. These faults have separate knowledge articles,
+ * and the article referred to by 'zpool status' must match that indicated by
+ * the syslog error message. We override missing data as well as corrupt pool.
+ */
+static char *msgid_table_active[] = {
+ "ZFS-8000-14",
+ "ZFS-8000-D3", /* overridden */
+ "ZFS-8000-D3", /* overridden */
+ "ZFS-8000-4J",
+ "ZFS-8000-5E",
+ "ZFS-8000-6X",
+ "ZFS-8000-CS", /* overridden */
+ "ZFS-8000-8A",
+ "ZFS-8000-9P",
+ "ZFS-8000-CS", /* overridden */
+};
+
+#define NMSGID (sizeof (msgid_table) / sizeof (msgid_table[0]))
+
+/* ARGSUSED */
+static int
+vdev_missing(uint64_t state, uint64_t aux, uint64_t errs)
+{
+ return (state == VDEV_STATE_CANT_OPEN &&
+ aux == VDEV_AUX_OPEN_FAILED);
+}
+
+/* ARGSUSED */
+static int
+vdev_errors(uint64_t state, uint64_t aux, uint64_t errs)
+{
+ return (errs != 0);
+}
+
+/* ARGSUSED */
+static int
+vdev_broken(uint64_t state, uint64_t aux, uint64_t errs)
+{
+ return (state == VDEV_STATE_CANT_OPEN);
+}
+
+/* ARGSUSED */
+static int
+vdev_offlined(uint64_t state, uint64_t aux, uint64_t errs)
+{
+ return (state == VDEV_STATE_OFFLINE);
+}
+
+/*
+ * Detect if any leaf devices that have seen errors or could not be opened.
+ */
+static boolean_t
+find_vdev_problem(nvlist_t *vdev, int (*func)(uint64_t, uint64_t, uint64_t))
+{
+ nvlist_t **child;
+ vdev_stat_t *vs;
+ uint_t c, children;
+ char *type;
+
+ /*
+ * Ignore problems within a 'replacing' vdev, since we're presumably in
+ * the process of repairing any such errors, and don't want to call them
+ * out again. We'll pick up the fact that a resilver is happening
+ * later.
+ */
+ verify(nvlist_lookup_string(vdev, ZPOOL_CONFIG_TYPE, &type) == 0);
+ if (strcmp(type, VDEV_TYPE_REPLACING) == 0)
+ return (B_FALSE);
+
+ if (nvlist_lookup_nvlist_array(vdev, ZPOOL_CONFIG_CHILDREN, &child,
+ &children) == 0) {
+ for (c = 0; c < children; c++)
+ if (find_vdev_problem(child[c], func))
+ return (B_TRUE);
+ } else {
+ verify(nvlist_lookup_uint64_array(vdev, ZPOOL_CONFIG_STATS,
+ (uint64_t **)&vs, &c) == 0);
+
+ if (func(vs->vs_state, vs->vs_aux,
+ vs->vs_read_errors +
+ vs->vs_write_errors +
+ vs->vs_checksum_errors))
+ return (B_TRUE);
+ }
+
+ return (B_FALSE);
+}
+
+/*
+ * Active pool health status.
+ *
+ * To determine the status for a pool, we make several passes over the config,
+ * picking the most egregious error we find. In order of importance, we do the
+ * following:
+ *
+ * - Check for a complete and valid configuration
+ * - Look for any missing devices in a non-replicated config
+ * - Check for any data errors
+ * - Check for any missing devices in a replicated config
+ * - Look for any devices showing errors
+ * - Check for any resilvering devices
+ *
+ * There can obviously be multiple errors within a single pool, so this routine
+ * only picks the most damaging of all the current errors to report.
+ */
+static zpool_status_t
+check_status(nvlist_t *config, boolean_t isimport)
+{
+ nvlist_t *nvroot;
+ vdev_stat_t *vs;
+ uint_t vsc;
+ uint64_t nerr;
+ uint64_t version;
+
+ verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION,
+ &version) == 0);
+ verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE,
+ &nvroot) == 0);
+ verify(nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_STATS,
+ (uint64_t **)&vs, &vsc) == 0);
+
+ /*
+ * Newer on-disk version.
+ */
+ if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
+ vs->vs_aux == VDEV_AUX_VERSION_NEWER)
+ return (ZPOOL_STATUS_VERSION_NEWER);
+
+ /*
+ * Check that the config is complete.
+ */
+ if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
+ vs->vs_aux == VDEV_AUX_BAD_GUID_SUM)
+ return (ZPOOL_STATUS_BAD_GUID_SUM);
+
+ /*
+ * Missing devices in non-replicated config.
+ */
+ if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
+ find_vdev_problem(nvroot, vdev_missing))
+ return (ZPOOL_STATUS_MISSING_DEV_NR);
+
+ if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
+ find_vdev_problem(nvroot, vdev_broken))
+ return (ZPOOL_STATUS_CORRUPT_LABEL_NR);
+
+ /*
+ * Corrupted pool metadata
+ */
+ if (vs->vs_state == VDEV_STATE_CANT_OPEN &&
+ vs->vs_aux == VDEV_AUX_CORRUPT_DATA)
+ return (ZPOOL_STATUS_CORRUPT_POOL);
+
+ /*
+ * Persistent data errors.
+ */
+ if (!isimport) {
+ if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_ERRCOUNT,
+ &nerr) == 0 && nerr != 0)
+ return (ZPOOL_STATUS_CORRUPT_DATA);
+ }
+
+ /*
+ * Missing devices in a replicated config.
+ */
+ if (find_vdev_problem(nvroot, vdev_missing))
+ return (ZPOOL_STATUS_MISSING_DEV_R);
+ if (find_vdev_problem(nvroot, vdev_broken))
+ return (ZPOOL_STATUS_CORRUPT_LABEL_R);
+
+ /*
+ * Devices with errors
+ */
+ if (!isimport && find_vdev_problem(nvroot, vdev_errors))
+ return (ZPOOL_STATUS_FAILING_DEV);
+
+ /*
+ * Offlined devices
+ */
+ if (find_vdev_problem(nvroot, vdev_offlined))
+ return (ZPOOL_STATUS_OFFLINE_DEV);
+
+ /*
+ * Currently resilvering
+ */
+ if (!vs->vs_scrub_complete && vs->vs_scrub_type == POOL_SCRUB_RESILVER)
+ return (ZPOOL_STATUS_RESILVERING);
+
+ /*
+ * Outdated, but usable, version
+ */
+ if (version < ZFS_VERSION)
+ return (ZPOOL_STATUS_VERSION_OLDER);
+
+ return (ZPOOL_STATUS_OK);
+}
+
+zpool_status_t
+zpool_get_status(zpool_handle_t *zhp, char **msgid)
+{
+ zpool_status_t ret = check_status(zhp->zpool_config, B_FALSE);
+
+ if (ret >= NMSGID)
+ *msgid = NULL;
+ else
+ *msgid = msgid_table_active[ret];
+
+ return (ret);
+}
+
+zpool_status_t
+zpool_import_status(nvlist_t *config, char **msgid)
+{
+ zpool_status_t ret = check_status(config, B_TRUE);
+
+ if (ret >= NMSGID)
+ *msgid = NULL;
+ else
+ *msgid = msgid_table[ret];
+
+ return (ret);
+}
diff --git a/contrib/opensolaris/lib/libzfs/common/libzfs_util.c b/contrib/opensolaris/lib/libzfs/common/libzfs_util.c
new file mode 100644
index 0000000..279835b
--- /dev/null
+++ b/contrib/opensolaris/lib/libzfs/common/libzfs_util.c
@@ -0,0 +1,827 @@
+/*
+ * 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"
+
+/*
+ * Internal utility routines for the ZFS library.
+ */
+
+#include <errno.h>
+#include <fcntl.h>
+#include <libintl.h>
+#include <stdarg.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <strings.h>
+#include <unistd.h>
+#include <sys/mnttab.h>
+#include <sys/mntent.h>
+#include <sys/types.h>
+
+#include <libzfs.h>
+
+#include "libzfs_impl.h"
+
+int
+libzfs_errno(libzfs_handle_t *hdl)
+{
+ return (hdl->libzfs_error);
+}
+
+const char *
+libzfs_error_action(libzfs_handle_t *hdl)
+{
+ return (hdl->libzfs_action);
+}
+
+const char *
+libzfs_error_description(libzfs_handle_t *hdl)
+{
+ if (hdl->libzfs_desc[0] != '\0')
+ return (hdl->libzfs_desc);
+
+ switch (hdl->libzfs_error) {
+ case EZFS_NOMEM:
+ return (dgettext(TEXT_DOMAIN, "out of memory"));
+ case EZFS_BADPROP:
+ return (dgettext(TEXT_DOMAIN, "invalid property value"));
+ case EZFS_PROPREADONLY:
+ return (dgettext(TEXT_DOMAIN, "read only property"));
+ case EZFS_PROPTYPE:
+ return (dgettext(TEXT_DOMAIN, "property doesn't apply to "
+ "datasets of this type"));
+ case EZFS_PROPNONINHERIT:
+ return (dgettext(TEXT_DOMAIN, "property cannot be inherited"));
+ case EZFS_PROPSPACE:
+ return (dgettext(TEXT_DOMAIN, "invalid quota or reservation"));
+ case EZFS_BADTYPE:
+ return (dgettext(TEXT_DOMAIN, "operation not applicable to "
+ "datasets of this type"));
+ case EZFS_BUSY:
+ return (dgettext(TEXT_DOMAIN, "pool or dataset is busy"));
+ case EZFS_EXISTS:
+ return (dgettext(TEXT_DOMAIN, "pool or dataset exists"));
+ case EZFS_NOENT:
+ return (dgettext(TEXT_DOMAIN, "no such pool or dataset"));
+ case EZFS_BADSTREAM:
+ return (dgettext(TEXT_DOMAIN, "invalid backup stream"));
+ case EZFS_DSREADONLY:
+ return (dgettext(TEXT_DOMAIN, "dataset is read only"));
+ case EZFS_VOLTOOBIG:
+ return (dgettext(TEXT_DOMAIN, "volume size exceeds limit for "
+ "this system"));
+ case EZFS_VOLHASDATA:
+ return (dgettext(TEXT_DOMAIN, "volume has data"));
+ case EZFS_INVALIDNAME:
+ return (dgettext(TEXT_DOMAIN, "invalid name"));
+ case EZFS_BADRESTORE:
+ return (dgettext(TEXT_DOMAIN, "unable to restore to "
+ "destination"));
+ case EZFS_BADBACKUP:
+ return (dgettext(TEXT_DOMAIN, "backup failed"));
+ case EZFS_BADTARGET:
+ return (dgettext(TEXT_DOMAIN, "invalid target vdev"));
+ case EZFS_NODEVICE:
+ return (dgettext(TEXT_DOMAIN, "no such device in pool"));
+ case EZFS_BADDEV:
+ return (dgettext(TEXT_DOMAIN, "invalid device"));
+ case EZFS_NOREPLICAS:
+ return (dgettext(TEXT_DOMAIN, "no valid replicas"));
+ case EZFS_RESILVERING:
+ return (dgettext(TEXT_DOMAIN, "currently resilvering"));
+ case EZFS_BADVERSION:
+ return (dgettext(TEXT_DOMAIN, "unsupported version"));
+ case EZFS_POOLUNAVAIL:
+ return (dgettext(TEXT_DOMAIN, "pool is unavailable"));
+ case EZFS_DEVOVERFLOW:
+ return (dgettext(TEXT_DOMAIN, "too many devices in one vdev"));
+ case EZFS_BADPATH:
+ return (dgettext(TEXT_DOMAIN, "must be an absolute path"));
+ case EZFS_CROSSTARGET:
+ return (dgettext(TEXT_DOMAIN, "operation crosses datasets or "
+ "pools"));
+ case EZFS_ZONED:
+ return (dgettext(TEXT_DOMAIN, "dataset in use by local zone"));
+ case EZFS_MOUNTFAILED:
+ return (dgettext(TEXT_DOMAIN, "mount failed"));
+ case EZFS_UMOUNTFAILED:
+ return (dgettext(TEXT_DOMAIN, "umount failed"));
+ case EZFS_UNSHARENFSFAILED:
+ return (dgettext(TEXT_DOMAIN, "unshare(1M) failed"));
+ case EZFS_SHARENFSFAILED:
+ return (dgettext(TEXT_DOMAIN, "share(1M) failed"));
+ case EZFS_DEVLINKS:
+ return (dgettext(TEXT_DOMAIN, "failed to create /dev links"));
+ case EZFS_PERM:
+ return (dgettext(TEXT_DOMAIN, "permission denied"));
+ case EZFS_NOSPC:
+ return (dgettext(TEXT_DOMAIN, "out of space"));
+ case EZFS_IO:
+ return (dgettext(TEXT_DOMAIN, "I/O error"));
+ case EZFS_INTR:
+ return (dgettext(TEXT_DOMAIN, "signal received"));
+ case EZFS_ISSPARE:
+ return (dgettext(TEXT_DOMAIN, "device is reserved as a hot "
+ "spare"));
+ case EZFS_INVALCONFIG:
+ return (dgettext(TEXT_DOMAIN, "invalid vdev configuration"));
+ case EZFS_RECURSIVE:
+ return (dgettext(TEXT_DOMAIN, "recursive dataset dependency"));
+ case EZFS_NOHISTORY:
+ return (dgettext(TEXT_DOMAIN, "no history available"));
+ case EZFS_UNSHAREISCSIFAILED:
+ return (dgettext(TEXT_DOMAIN,
+ "iscsitgtd failed request to unshare"));
+ case EZFS_SHAREISCSIFAILED:
+ return (dgettext(TEXT_DOMAIN,
+ "iscsitgtd failed request to share"));
+ case EZFS_POOLPROPS:
+ return (dgettext(TEXT_DOMAIN, "failed to retrieve "
+ "pool properties"));
+ case EZFS_POOL_NOTSUP:
+ return (dgettext(TEXT_DOMAIN, "operation not supported "
+ "on this type of pool"));
+ case EZFS_POOL_INVALARG:
+ return (dgettext(TEXT_DOMAIN, "invalid argument for "
+ "this pool operation"));
+ case EZFS_UNKNOWN:
+ return (dgettext(TEXT_DOMAIN, "unknown error"));
+ default:
+ assert(hdl->libzfs_error == 0);
+ return (dgettext(TEXT_DOMAIN, "no error"));
+ }
+}
+
+/*PRINTFLIKE2*/
+void
+zfs_error_aux(libzfs_handle_t *hdl, const char *fmt, ...)
+{
+ va_list ap;
+
+ va_start(ap, fmt);
+
+ (void) vsnprintf(hdl->libzfs_desc, sizeof (hdl->libzfs_desc),
+ fmt, ap);
+ hdl->libzfs_desc_active = 1;
+
+ va_end(ap);
+}
+
+static void
+zfs_verror(libzfs_handle_t *hdl, int error, const char *fmt, va_list ap)
+{
+ (void) vsnprintf(hdl->libzfs_action, sizeof (hdl->libzfs_action),
+ fmt, ap);
+ hdl->libzfs_error = error;
+
+ if (hdl->libzfs_desc_active)
+ hdl->libzfs_desc_active = 0;
+ else
+ hdl->libzfs_desc[0] = '\0';
+
+ if (hdl->libzfs_printerr) {
+ if (error == EZFS_UNKNOWN) {
+ (void) fprintf(stderr, dgettext(TEXT_DOMAIN, "internal "
+ "error: %s\n"), libzfs_error_description(hdl));
+ abort();
+ }
+
+ (void) fprintf(stderr, "%s: %s\n", hdl->libzfs_action,
+ libzfs_error_description(hdl));
+ if (error == EZFS_NOMEM)
+ exit(1);
+ }
+}
+
+int
+zfs_error(libzfs_handle_t *hdl, int error, const char *msg)
+{
+ return (zfs_error_fmt(hdl, error, "%s", msg));
+}
+
+/*PRINTFLIKE3*/
+int
+zfs_error_fmt(libzfs_handle_t *hdl, int error, const char *fmt, ...)
+{
+ va_list ap;
+
+ va_start(ap, fmt);
+
+ zfs_verror(hdl, error, fmt, ap);
+
+ va_end(ap);
+
+ return (-1);
+}
+
+static int
+zfs_common_error(libzfs_handle_t *hdl, int error, const char *fmt,
+ va_list ap)
+{
+ switch (error) {
+ case EPERM:
+ case EACCES:
+ zfs_verror(hdl, EZFS_PERM, fmt, ap);
+ return (-1);
+
+ case EIO:
+ zfs_verror(hdl, EZFS_IO, fmt, ap);
+ return (-1);
+
+ case EINTR:
+ zfs_verror(hdl, EZFS_INTR, fmt, ap);
+ return (-1);
+ }
+
+ return (0);
+}
+
+int
+zfs_standard_error(libzfs_handle_t *hdl, int error, const char *msg)
+{
+ return (zfs_standard_error_fmt(hdl, error, "%s", msg));
+}
+
+/*PRINTFLIKE3*/
+int
+zfs_standard_error_fmt(libzfs_handle_t *hdl, int error, const char *fmt, ...)
+{
+ va_list ap;
+
+ va_start(ap, fmt);
+
+ if (zfs_common_error(hdl, error, fmt, ap) != 0) {
+ va_end(ap);
+ return (-1);
+ }
+
+
+ switch (error) {
+ case ENXIO:
+ zfs_verror(hdl, EZFS_IO, fmt, ap);
+ break;
+
+ case ENOENT:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "dataset does not exist"));
+ zfs_verror(hdl, EZFS_NOENT, fmt, ap);
+ break;
+
+ case ENOSPC:
+ case EDQUOT:
+ zfs_verror(hdl, EZFS_NOSPC, fmt, ap);
+ return (-1);
+
+ case EEXIST:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "dataset already exists"));
+ zfs_verror(hdl, EZFS_EXISTS, fmt, ap);
+ break;
+
+ case EBUSY:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "dataset is busy"));
+ zfs_verror(hdl, EZFS_BUSY, fmt, ap);
+ break;
+ default:
+ zfs_error_aux(hdl, strerror(errno));
+ zfs_verror(hdl, EZFS_UNKNOWN, fmt, ap);
+ break;
+ }
+
+ va_end(ap);
+ return (-1);
+}
+
+int
+zpool_standard_error(libzfs_handle_t *hdl, int error, const char *msg)
+{
+ return (zpool_standard_error_fmt(hdl, error, "%s", msg));
+}
+
+/*PRINTFLIKE3*/
+int
+zpool_standard_error_fmt(libzfs_handle_t *hdl, int error, const char *fmt, ...)
+{
+ va_list ap;
+
+ va_start(ap, fmt);
+
+ if (zfs_common_error(hdl, error, fmt, ap) != 0) {
+ va_end(ap);
+ return (-1);
+ }
+
+ switch (error) {
+ case ENODEV:
+ zfs_verror(hdl, EZFS_NODEVICE, fmt, ap);
+ break;
+
+ case ENOENT:
+ zfs_error_aux(hdl,
+ dgettext(TEXT_DOMAIN, "no such pool or dataset"));
+ zfs_verror(hdl, EZFS_NOENT, fmt, ap);
+ break;
+
+ case EEXIST:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "pool already exists"));
+ zfs_verror(hdl, EZFS_EXISTS, fmt, ap);
+ break;
+
+ case EBUSY:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "pool is busy"));
+ zfs_verror(hdl, EZFS_EXISTS, fmt, ap);
+ break;
+
+ case ENXIO:
+ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
+ "one or more devices is currently unavailable"));
+ zfs_verror(hdl, EZFS_BADDEV, fmt, ap);
+ break;
+
+ case ENAMETOOLONG:
+ zfs_verror(hdl, EZFS_DEVOVERFLOW, fmt, ap);
+ break;
+
+ case ENOTSUP:
+ zfs_verror(hdl, EZFS_POOL_NOTSUP, fmt, ap);
+ break;
+
+ case EINVAL:
+ zfs_verror(hdl, EZFS_POOL_INVALARG, fmt, ap);
+ break;
+
+ default:
+ zfs_error_aux(hdl, strerror(error));
+ zfs_verror(hdl, EZFS_UNKNOWN, fmt, ap);
+ }
+
+ va_end(ap);
+ return (-1);
+}
+
+/*
+ * Display an out of memory error message and abort the current program.
+ */
+int
+no_memory(libzfs_handle_t *hdl)
+{
+ return (zfs_error(hdl, EZFS_NOMEM, "internal error"));
+}
+
+/*
+ * A safe form of malloc() which will die if the allocation fails.
+ */
+void *
+zfs_alloc(libzfs_handle_t *hdl, size_t size)
+{
+ void *data;
+
+ if ((data = calloc(1, size)) == NULL)
+ (void) no_memory(hdl);
+
+ return (data);
+}
+
+/*
+ * A safe form of realloc(), which also zeroes newly allocated space.
+ */
+void *
+zfs_realloc(libzfs_handle_t *hdl, void *ptr, size_t oldsize, size_t newsize)
+{
+ void *ret;
+
+ if ((ret = realloc(ptr, newsize)) == NULL) {
+ (void) no_memory(hdl);
+ free(ptr);
+ return (NULL);
+ }
+
+ bzero((char *)ret + oldsize, (newsize - oldsize));
+ return (ret);
+}
+
+/*
+ * A safe form of strdup() which will die if the allocation fails.
+ */
+char *
+zfs_strdup(libzfs_handle_t *hdl, const char *str)
+{
+ char *ret;
+
+ if ((ret = strdup(str)) == NULL)
+ (void) no_memory(hdl);
+
+ return (ret);
+}
+
+/*
+ * Convert a number to an appropriately human-readable output.
+ */
+void
+zfs_nicenum(uint64_t num, char *buf, size_t buflen)
+{
+ uint64_t n = num;
+ int index = 0;
+ char u;
+
+ while (n >= 1024) {
+ n /= 1024;
+ index++;
+ }
+
+ u = " KMGTPE"[index];
+
+ if (index == 0) {
+ (void) snprintf(buf, buflen, "%llu", n);
+ } else if ((num & ((1ULL << 10 * index) - 1)) == 0) {
+ /*
+ * If this is an even multiple of the base, always display
+ * without any decimal precision.
+ */
+ (void) snprintf(buf, buflen, "%llu%c", n, u);
+ } else {
+ /*
+ * We want to choose a precision that reflects the best choice
+ * for fitting in 5 characters. This can get rather tricky when
+ * we have numbers that are very close to an order of magnitude.
+ * For example, when displaying 10239 (which is really 9.999K),
+ * we want only a single place of precision for 10.0K. We could
+ * develop some complex heuristics for this, but it's much
+ * easier just to try each combination in turn.
+ */
+ int i;
+ for (i = 2; i >= 0; i--) {
+ (void) snprintf(buf, buflen, "%.*f%c", i,
+ (double)num / (1ULL << 10 * index), u);
+ if (strlen(buf) <= 5)
+ break;
+ }
+ }
+}
+
+void
+libzfs_print_on_error(libzfs_handle_t *hdl, boolean_t printerr)
+{
+ hdl->libzfs_printerr = printerr;
+}
+
+libzfs_handle_t *
+libzfs_init(void)
+{
+ libzfs_handle_t *hdl;
+
+ if ((hdl = calloc(sizeof (libzfs_handle_t), 1)) == NULL) {
+ return (NULL);
+ }
+
+ if ((hdl->libzfs_fd = open(ZFS_DEV, O_RDWR)) < 0) {
+ free(hdl);
+ return (NULL);
+ }
+
+ if ((hdl->libzfs_mnttab = fopen(MNTTAB, "r")) == NULL) {
+ (void) close(hdl->libzfs_fd);
+ free(hdl);
+ return (NULL);
+ }
+
+ hdl->libzfs_sharetab = fopen(ZFS_EXPORTS_PATH, "r");
+
+ return (hdl);
+}
+
+void
+libzfs_fini(libzfs_handle_t *hdl)
+{
+ (void) close(hdl->libzfs_fd);
+ if (hdl->libzfs_mnttab)
+ (void) fclose(hdl->libzfs_mnttab);
+ if (hdl->libzfs_sharetab)
+ (void) fclose(hdl->libzfs_sharetab);
+ namespace_clear(hdl);
+ free(hdl);
+}
+
+libzfs_handle_t *
+zpool_get_handle(zpool_handle_t *zhp)
+{
+ return (zhp->zpool_hdl);
+}
+
+libzfs_handle_t *
+zfs_get_handle(zfs_handle_t *zhp)
+{
+ return (zhp->zfs_hdl);
+}
+
+/*
+ * Given a name, determine whether or not it's a valid path
+ * (starts with '/' or "./"). If so, walk the mnttab trying
+ * to match the device number. If not, treat the path as an
+ * fs/vol/snap name.
+ */
+zfs_handle_t *
+zfs_path_to_zhandle(libzfs_handle_t *hdl, char *path, zfs_type_t argtype)
+{
+ struct statfs statbuf;
+
+ if (path[0] != '/' && strncmp(path, "./", strlen("./")) != 0) {
+ /*
+ * It's not a valid path, assume it's a name of type 'argtype'.
+ */
+ return (zfs_open(hdl, path, argtype));
+ }
+
+ if (statfs(path, &statbuf) != 0) {
+ (void) fprintf(stderr, "%s: %s\n", path, strerror(errno));
+ return (NULL);
+ }
+
+ if (strcmp(statbuf.f_fstypename, MNTTYPE_ZFS) != 0) {
+ (void) fprintf(stderr, gettext("'%s': not a ZFS filesystem\n"),
+ path);
+ return (NULL);
+ }
+
+ return (zfs_open(hdl, statbuf.f_mntfromname, ZFS_TYPE_FILESYSTEM));
+}
+
+/*
+ * Initialize the zc_nvlist_dst member to prepare for receiving an nvlist from
+ * an ioctl().
+ */
+int
+zcmd_alloc_dst_nvlist(libzfs_handle_t *hdl, zfs_cmd_t *zc, size_t len)
+{
+ if (len == 0)
+ len = 2048;
+ zc->zc_nvlist_dst_size = len;
+ if ((zc->zc_nvlist_dst = (uint64_t)(uintptr_t)
+ zfs_alloc(hdl, zc->zc_nvlist_dst_size)) == 0)
+ return (-1);
+
+ return (0);
+}
+
+/*
+ * Called when an ioctl() which returns an nvlist fails with ENOMEM. This will
+ * expand the nvlist to the size specified in 'zc_nvlist_dst_size', which was
+ * filled in by the kernel to indicate the actual required size.
+ */
+int
+zcmd_expand_dst_nvlist(libzfs_handle_t *hdl, zfs_cmd_t *zc)
+{
+ free((void *)(uintptr_t)zc->zc_nvlist_dst);
+ if ((zc->zc_nvlist_dst = (uint64_t)(uintptr_t)
+ zfs_alloc(hdl, zc->zc_nvlist_dst_size))
+ == 0)
+ return (-1);
+
+ return (0);
+}
+
+/*
+ * Called to free the src and dst nvlists stored in the command structure.
+ */
+void
+zcmd_free_nvlists(zfs_cmd_t *zc)
+{
+ free((void *)(uintptr_t)zc->zc_nvlist_src);
+ free((void *)(uintptr_t)zc->zc_nvlist_dst);
+}
+
+int
+zcmd_write_src_nvlist(libzfs_handle_t *hdl, zfs_cmd_t *zc, nvlist_t *nvl,
+ size_t *size)
+{
+ char *packed;
+ size_t len;
+
+ verify(nvlist_size(nvl, &len, NV_ENCODE_NATIVE) == 0);
+
+ if ((packed = zfs_alloc(hdl, len)) == NULL)
+ return (-1);
+
+ verify(nvlist_pack(nvl, &packed, &len, NV_ENCODE_NATIVE, 0) == 0);
+
+ zc->zc_nvlist_src = (uint64_t)(uintptr_t)packed;
+ zc->zc_nvlist_src_size = len;
+
+ if (size)
+ *size = len;
+ return (0);
+}
+
+/*
+ * Unpacks an nvlist from the ZFS ioctl command structure.
+ */
+int
+zcmd_read_dst_nvlist(libzfs_handle_t *hdl, zfs_cmd_t *zc, nvlist_t **nvlp)
+{
+ if (nvlist_unpack((void *)(uintptr_t)zc->zc_nvlist_dst,
+ zc->zc_nvlist_dst_size, nvlp, 0) != 0)
+ return (no_memory(hdl));
+
+ return (0);
+}
+
+static void
+zfs_print_prop_headers(libzfs_get_cbdata_t *cbp)
+{
+ zfs_proplist_t *pl = cbp->cb_proplist;
+ int i;
+ char *title;
+ size_t len;
+
+ cbp->cb_first = B_FALSE;
+ if (cbp->cb_scripted)
+ return;
+
+ /*
+ * Start with the length of the column headers.
+ */
+ cbp->cb_colwidths[GET_COL_NAME] = strlen(dgettext(TEXT_DOMAIN, "NAME"));
+ cbp->cb_colwidths[GET_COL_PROPERTY] = strlen(dgettext(TEXT_DOMAIN,
+ "PROPERTY"));
+ cbp->cb_colwidths[GET_COL_VALUE] = strlen(dgettext(TEXT_DOMAIN,
+ "VALUE"));
+ cbp->cb_colwidths[GET_COL_SOURCE] = strlen(dgettext(TEXT_DOMAIN,
+ "SOURCE"));
+
+ /*
+ * Go through and calculate the widths for each column. For the
+ * 'source' column, we kludge it up by taking the worst-case scenario of
+ * inheriting from the longest name. This is acceptable because in the
+ * majority of cases 'SOURCE' is the last column displayed, and we don't
+ * use the width anyway. Note that the 'VALUE' column can be oversized,
+ * if the name of the property is much longer the any values we find.
+ */
+ for (pl = cbp->cb_proplist; pl != NULL; pl = pl->pl_next) {
+ /*
+ * 'PROPERTY' column
+ */
+ if (pl->pl_prop != ZFS_PROP_INVAL) {
+ len = strlen(zfs_prop_to_name(pl->pl_prop));
+ if (len > cbp->cb_colwidths[GET_COL_PROPERTY])
+ cbp->cb_colwidths[GET_COL_PROPERTY] = len;
+ } else {
+ len = strlen(pl->pl_user_prop);
+ if (len > cbp->cb_colwidths[GET_COL_PROPERTY])
+ cbp->cb_colwidths[GET_COL_PROPERTY] = len;
+ }
+
+ /*
+ * 'VALUE' column
+ */
+ if ((pl->pl_prop != ZFS_PROP_NAME || !pl->pl_all) &&
+ pl->pl_width > cbp->cb_colwidths[GET_COL_VALUE])
+ cbp->cb_colwidths[GET_COL_VALUE] = pl->pl_width;
+
+ /*
+ * 'NAME' and 'SOURCE' columns
+ */
+ if (pl->pl_prop == ZFS_PROP_NAME &&
+ pl->pl_width > cbp->cb_colwidths[GET_COL_NAME]) {
+ cbp->cb_colwidths[GET_COL_NAME] = pl->pl_width;
+ cbp->cb_colwidths[GET_COL_SOURCE] = pl->pl_width +
+ strlen(dgettext(TEXT_DOMAIN, "inherited from"));
+ }
+ }
+
+ /*
+ * Now go through and print the headers.
+ */
+ for (i = 0; i < 4; i++) {
+ switch (cbp->cb_columns[i]) {
+ case GET_COL_NAME:
+ title = dgettext(TEXT_DOMAIN, "NAME");
+ break;
+ case GET_COL_PROPERTY:
+ title = dgettext(TEXT_DOMAIN, "PROPERTY");
+ break;
+ case GET_COL_VALUE:
+ title = dgettext(TEXT_DOMAIN, "VALUE");
+ break;
+ case GET_COL_SOURCE:
+ title = dgettext(TEXT_DOMAIN, "SOURCE");
+ break;
+ default:
+ title = NULL;
+ }
+
+ if (title != NULL) {
+ if (i == 3 || cbp->cb_columns[i + 1] == 0)
+ (void) printf("%s", title);
+ else
+ (void) printf("%-*s ",
+ cbp->cb_colwidths[cbp->cb_columns[i]],
+ title);
+ }
+ }
+ (void) printf("\n");
+}
+
+/*
+ * Display a single line of output, according to the settings in the callback
+ * structure.
+ */
+void
+libzfs_print_one_property(const char *name, libzfs_get_cbdata_t *cbp,
+ const char *propname, const char *value, zfs_source_t sourcetype,
+ const char *source)
+{
+ int i;
+ const char *str;
+ char buf[128];
+
+ /*
+ * Ignore those source types that the user has chosen to ignore.
+ */
+ if ((sourcetype & cbp->cb_sources) == 0)
+ return;
+
+ if (cbp->cb_first)
+ zfs_print_prop_headers(cbp);
+
+ for (i = 0; i < 4; i++) {
+ switch (cbp->cb_columns[i]) {
+ case GET_COL_NAME:
+ str = name;
+ break;
+
+ case GET_COL_PROPERTY:
+ str = propname;
+ break;
+
+ case GET_COL_VALUE:
+ str = value;
+ break;
+
+ case GET_COL_SOURCE:
+ switch (sourcetype) {
+ case ZFS_SRC_NONE:
+ str = "-";
+ break;
+
+ case ZFS_SRC_DEFAULT:
+ str = "default";
+ break;
+
+ case ZFS_SRC_LOCAL:
+ str = "local";
+ break;
+
+ case ZFS_SRC_TEMPORARY:
+ str = "temporary";
+ break;
+
+ case ZFS_SRC_INHERITED:
+ (void) snprintf(buf, sizeof (buf),
+ "inherited from %s", source);
+ str = buf;
+ break;
+ }
+ break;
+
+ default:
+ continue;
+ }
+
+ if (cbp->cb_columns[i + 1] == 0)
+ (void) printf("%s", str);
+ else if (cbp->cb_scripted)
+ (void) printf("%s\t", str);
+ else
+ (void) printf("%-*s ",
+ cbp->cb_colwidths[cbp->cb_columns[i]],
+ str);
+
+ }
+
+ (void) printf("\n");
+}
diff --git a/contrib/opensolaris/lib/libzpool/common/kernel.c b/contrib/opensolaris/lib/libzpool/common/kernel.c
new file mode 100644
index 0000000..086f33e
--- /dev/null
+++ b/contrib/opensolaris/lib/libzpool/common/kernel.c
@@ -0,0 +1,830 @@
+/*
+ * 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"
+
+#include <assert.h>
+#include <fcntl.h>
+#include <poll.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <zlib.h>
+#include <sys/spa.h>
+#include <sys/stat.h>
+#include <sys/processor.h>
+#include <sys/zfs_context.h>
+#include <sys/zmod.h>
+
+/*
+ * Emulation of kernel services in userland.
+ */
+
+int hz = 119; /* frequency when using gethrtime() >> 23 for lbolt */
+uint64_t physmem;
+vnode_t *rootdir = (vnode_t *)0xabcd1234;
+
+/*
+ * =========================================================================
+ * threads
+ * =========================================================================
+ */
+/*ARGSUSED*/
+kthread_t *
+zk_thread_create(void (*func)(), void *arg)
+{
+ thread_t tid;
+
+ VERIFY(thr_create(0, 0, (void *(*)(void *))func, arg, THR_DETACHED,
+ &tid) == 0);
+
+ return ((void *)(uintptr_t)tid);
+}
+
+/*
+ * =========================================================================
+ * kstats
+ * =========================================================================
+ */
+/*ARGSUSED*/
+kstat_t *
+kstat_create(char *module, int instance, char *name, char *class,
+ uchar_t type, ulong_t ndata, uchar_t ks_flag)
+{
+ return (NULL);
+}
+
+/*ARGSUSED*/
+void
+kstat_install(kstat_t *ksp)
+{}
+
+/*ARGSUSED*/
+void
+kstat_delete(kstat_t *ksp)
+{}
+
+/*
+ * =========================================================================
+ * mutexes
+ * =========================================================================
+ */
+void
+zmutex_init(kmutex_t *mp)
+{
+ mp->m_owner = NULL;
+ (void) _mutex_init(&mp->m_lock, USYNC_THREAD, NULL);
+}
+
+void
+zmutex_destroy(kmutex_t *mp)
+{
+ ASSERT(mp->m_owner == NULL);
+ (void) _mutex_destroy(&(mp)->m_lock);
+ mp->m_owner = (void *)-1UL;
+}
+
+void
+mutex_enter(kmutex_t *mp)
+{
+ ASSERT(mp->m_owner != (void *)-1UL);
+ ASSERT(mp->m_owner != curthread);
+ VERIFY(mutex_lock(&mp->m_lock) == 0);
+ ASSERT(mp->m_owner == NULL);
+ mp->m_owner = curthread;
+}
+
+int
+mutex_tryenter(kmutex_t *mp)
+{
+ ASSERT(mp->m_owner != (void *)-1UL);
+ if (mutex_trylock(&mp->m_lock) == 0) {
+ ASSERT(mp->m_owner == NULL);
+ mp->m_owner = curthread;
+ return (1);
+ } else {
+ return (0);
+ }
+}
+
+void
+mutex_exit(kmutex_t *mp)
+{
+ ASSERT(mp->m_owner == curthread);
+ mp->m_owner = NULL;
+ VERIFY(mutex_unlock(&mp->m_lock) == 0);
+}
+
+void *
+mutex_owner(kmutex_t *mp)
+{
+ return (mp->m_owner);
+}
+
+/*
+ * =========================================================================
+ * rwlocks
+ * =========================================================================
+ */
+/*ARGSUSED*/
+void
+rw_init(krwlock_t *rwlp, char *name, int type, void *arg)
+{
+ rwlock_init(&rwlp->rw_lock, USYNC_THREAD, NULL);
+ rwlp->rw_owner = NULL;
+ rwlp->rw_count = 0;
+}
+
+void
+rw_destroy(krwlock_t *rwlp)
+{
+ rwlock_destroy(&rwlp->rw_lock);
+ rwlp->rw_owner = (void *)-1UL;
+ rwlp->rw_count = -2;
+}
+
+void
+rw_enter(krwlock_t *rwlp, krw_t rw)
+{
+ //ASSERT(!RW_LOCK_HELD(rwlp));
+ ASSERT(rwlp->rw_owner != (void *)-1UL);
+ ASSERT(rwlp->rw_owner != curthread);
+
+ if (rw == RW_READER) {
+ (void) rw_rdlock(&rwlp->rw_lock);
+ ASSERT(rwlp->rw_count >= 0);
+ atomic_add_int(&rwlp->rw_count, 1);
+ } else {
+ (void) rw_wrlock(&rwlp->rw_lock);
+ ASSERT(rwlp->rw_count == 0);
+ rwlp->rw_count = -1;
+ rwlp->rw_owner = curthread;
+ }
+}
+
+void
+rw_exit(krwlock_t *rwlp)
+{
+ ASSERT(rwlp->rw_owner != (void *)-1UL);
+
+ if (rwlp->rw_owner == curthread) {
+ /* Write locked. */
+ ASSERT(rwlp->rw_count == -1);
+ rwlp->rw_count = 0;
+ rwlp->rw_owner = NULL;
+ } else {
+ /* Read locked. */
+ ASSERT(rwlp->rw_count > 0);
+ atomic_add_int(&rwlp->rw_count, -1);
+ }
+ (void) rw_unlock(&rwlp->rw_lock);
+}
+
+int
+rw_tryenter(krwlock_t *rwlp, krw_t rw)
+{
+ int rv;
+
+ ASSERT(rwlp->rw_owner != (void *)-1UL);
+ ASSERT(rwlp->rw_owner != curthread);
+
+ if (rw == RW_READER)
+ rv = rw_tryrdlock(&rwlp->rw_lock);
+ else
+ rv = rw_trywrlock(&rwlp->rw_lock);
+
+ if (rv == 0) {
+ ASSERT(rwlp->rw_owner == NULL);
+ if (rw == RW_READER) {
+ ASSERT(rwlp->rw_count >= 0);
+ atomic_add_int(&rwlp->rw_count, 1);
+ } else {
+ ASSERT(rwlp->rw_count == 0);
+ rwlp->rw_count = -1;
+ rwlp->rw_owner = curthread;
+ }
+ return (1);
+ }
+
+ return (0);
+}
+
+/*ARGSUSED*/
+int
+rw_tryupgrade(krwlock_t *rwlp)
+{
+ ASSERT(rwlp->rw_owner != (void *)-1UL);
+
+ return (0);
+}
+
+int
+rw_lock_held(krwlock_t *rwlp)
+{
+
+ return (rwlp->rw_count != 0);
+}
+
+/*
+ * =========================================================================
+ * condition variables
+ * =========================================================================
+ */
+/*ARGSUSED*/
+void
+cv_init(kcondvar_t *cv, char *name, int type, void *arg)
+{
+ VERIFY(cond_init(cv, name, NULL) == 0);
+}
+
+void
+cv_destroy(kcondvar_t *cv)
+{
+ VERIFY(cond_destroy(cv) == 0);
+}
+
+void
+cv_wait(kcondvar_t *cv, kmutex_t *mp)
+{
+ ASSERT(mutex_owner(mp) == curthread);
+ mp->m_owner = NULL;
+ int ret = cond_wait(cv, &mp->m_lock);
+ VERIFY(ret == 0 || ret == EINTR);
+ mp->m_owner = curthread;
+}
+
+clock_t
+cv_timedwait(kcondvar_t *cv, kmutex_t *mp, clock_t abstime)
+{
+ int error;
+ struct timespec ts;
+ struct timeval tv;
+ clock_t delta;
+
+ ASSERT(abstime > 0);
+top:
+ delta = abstime - lbolt;
+ if (delta <= 0)
+ return (-1);
+
+ if (gettimeofday(&tv, NULL) != 0)
+ assert(!"gettimeofday() failed");
+
+ ts.tv_sec = tv.tv_sec + delta / hz;
+ ts.tv_nsec = tv.tv_usec * 1000 + (delta % hz) * (NANOSEC / hz);
+ ASSERT(ts.tv_nsec >= 0);
+
+ if(ts.tv_nsec >= NANOSEC) {
+ ts.tv_sec++;
+ ts.tv_nsec -= NANOSEC;
+ }
+
+ ASSERT(mutex_owner(mp) == curthread);
+ mp->m_owner = NULL;
+ error = pthread_cond_timedwait(cv, &mp->m_lock, &ts);
+ mp->m_owner = curthread;
+
+ if (error == EINTR)
+ goto top;
+
+ if (error == ETIMEDOUT)
+ return (-1);
+
+ ASSERT(error == 0);
+
+ return (1);
+}
+
+void
+cv_signal(kcondvar_t *cv)
+{
+ VERIFY(cond_signal(cv) == 0);
+}
+
+void
+cv_broadcast(kcondvar_t *cv)
+{
+ VERIFY(cond_broadcast(cv) == 0);
+}
+
+/*
+ * =========================================================================
+ * vnode operations
+ * =========================================================================
+ */
+/*
+ * Note: for the xxxat() versions of these functions, we assume that the
+ * starting vp is always rootdir (which is true for spa_directory.c, the only
+ * ZFS consumer of these interfaces). We assert this is true, and then emulate
+ * them by adding '/' in front of the path.
+ */
+
+/*ARGSUSED*/
+int
+vn_open(char *path, int x1, int flags, int mode, vnode_t **vpp, int x2, int x3)
+{
+ int fd;
+ vnode_t *vp;
+ int old_umask;
+ char realpath[MAXPATHLEN];
+ struct stat64 st;
+
+ /*
+ * If we're accessing a real disk from userland, we need to use
+ * the character interface to avoid caching. This is particularly
+ * important if we're trying to look at a real in-kernel storage
+ * pool from userland, e.g. via zdb, because otherwise we won't
+ * see the changes occurring under the segmap cache.
+ * On the other hand, the stupid character device returns zero
+ * for its size. So -- gag -- we open the block device to get
+ * its size, and remember it for subsequent VOP_GETATTR().
+ */
+ if (strncmp(path, "/dev/", 5) == 0) {
+ char *dsk;
+ fd = open64(path, O_RDONLY);
+ if (fd == -1)
+ return (errno);
+ if (fstat64(fd, &st) == -1) {
+ close(fd);
+ return (errno);
+ }
+ close(fd);
+ (void) sprintf(realpath, "%s", path);
+ dsk = strstr(path, "/dsk/");
+ if (dsk != NULL)
+ (void) sprintf(realpath + (dsk - path) + 1, "r%s",
+ dsk + 1);
+ } else {
+ (void) sprintf(realpath, "%s", path);
+ if (!(flags & FCREAT) && stat64(realpath, &st) == -1)
+ return (errno);
+ }
+
+ if (flags & FCREAT)
+ old_umask = umask(0);
+
+ /*
+ * The construct 'flags - FREAD' conveniently maps combinations of
+ * FREAD and FWRITE to the corresponding O_RDONLY, O_WRONLY, and O_RDWR.
+ */
+ fd = open64(realpath, flags - FREAD, mode);
+
+ if (flags & FCREAT)
+ (void) umask(old_umask);
+
+ if (fd == -1)
+ return (errno);
+
+ if (fstat64(fd, &st) == -1) {
+ close(fd);
+ return (errno);
+ }
+
+ (void) fcntl(fd, F_SETFD, FD_CLOEXEC);
+
+ *vpp = vp = umem_zalloc(sizeof (vnode_t), UMEM_NOFAIL);
+
+ vp->v_fd = fd;
+ vp->v_size = st.st_size;
+ vp->v_path = spa_strdup(path);
+
+ return (0);
+}
+
+int
+vn_openat(char *path, int x1, int flags, int mode, vnode_t **vpp, int x2,
+ int x3, vnode_t *startvp)
+{
+ char *realpath = umem_alloc(strlen(path) + 2, UMEM_NOFAIL);
+ int ret;
+
+ ASSERT(startvp == rootdir);
+ (void) sprintf(realpath, "/%s", path);
+
+ ret = vn_open(realpath, x1, flags, mode, vpp, x2, x3);
+
+ umem_free(realpath, strlen(path) + 2);
+
+ return (ret);
+}
+
+/*ARGSUSED*/
+int
+vn_rdwr(int uio, vnode_t *vp, void *addr, ssize_t len, offset_t offset,
+ int x1, int x2, rlim64_t x3, void *x4, ssize_t *residp)
+{
+ ssize_t iolen, split;
+
+ if (uio == UIO_READ) {
+ iolen = pread64(vp->v_fd, addr, len, offset);
+ } else {
+ /*
+ * To simulate partial disk writes, we split writes into two
+ * system calls so that the process can be killed in between.
+ */
+ split = (len > 0 ? rand() % len : 0);
+ iolen = pwrite64(vp->v_fd, addr, split, offset);
+ iolen += pwrite64(vp->v_fd, (char *)addr + split,
+ len - split, offset + split);
+ }
+
+ if (iolen == -1)
+ return (errno);
+ if (residp)
+ *residp = len - iolen;
+ else if (iolen != len)
+ return (EIO);
+ return (0);
+}
+
+void
+vn_close(vnode_t *vp)
+{
+ close(vp->v_fd);
+ spa_strfree(vp->v_path);
+ umem_free(vp, sizeof (vnode_t));
+}
+
+#ifdef ZFS_DEBUG
+
+/*
+ * =========================================================================
+ * Figure out which debugging statements to print
+ * =========================================================================
+ */
+
+static char *dprintf_string;
+static int dprintf_print_all;
+
+int
+dprintf_find_string(const char *string)
+{
+ char *tmp_str = dprintf_string;
+ int len = strlen(string);
+
+ /*
+ * Find out if this is a string we want to print.
+ * String format: file1.c,function_name1,file2.c,file3.c
+ */
+
+ while (tmp_str != NULL) {
+ if (strncmp(tmp_str, string, len) == 0 &&
+ (tmp_str[len] == ',' || tmp_str[len] == '\0'))
+ return (1);
+ tmp_str = strchr(tmp_str, ',');
+ if (tmp_str != NULL)
+ tmp_str++; /* Get rid of , */
+ }
+ return (0);
+}
+
+void
+dprintf_setup(int *argc, char **argv)
+{
+ int i, j;
+
+ /*
+ * Debugging can be specified two ways: by setting the
+ * environment variable ZFS_DEBUG, or by including a
+ * "debug=..." argument on the command line. The command
+ * line setting overrides the environment variable.
+ */
+
+ for (i = 1; i < *argc; i++) {
+ int len = strlen("debug=");
+ /* First look for a command line argument */
+ if (strncmp("debug=", argv[i], len) == 0) {
+ dprintf_string = argv[i] + len;
+ /* Remove from args */
+ for (j = i; j < *argc; j++)
+ argv[j] = argv[j+1];
+ argv[j] = NULL;
+ (*argc)--;
+ }
+ }
+
+ if (dprintf_string == NULL) {
+ /* Look for ZFS_DEBUG environment variable */
+ dprintf_string = getenv("ZFS_DEBUG");
+ }
+
+ /*
+ * Are we just turning on all debugging?
+ */
+ if (dprintf_find_string("on"))
+ dprintf_print_all = 1;
+}
+
+/*
+ * =========================================================================
+ * debug printfs
+ * =========================================================================
+ */
+void
+__dprintf(const char *file, const char *func, int line, const char *fmt, ...)
+{
+ const char *newfile;
+ va_list adx;
+
+ /*
+ * Get rid of annoying "../common/" prefix to filename.
+ */
+ newfile = strrchr(file, '/');
+ if (newfile != NULL) {
+ newfile = newfile + 1; /* Get rid of leading / */
+ } else {
+ newfile = file;
+ }
+
+ if (dprintf_print_all ||
+ dprintf_find_string(newfile) ||
+ dprintf_find_string(func)) {
+ /* Print out just the function name if requested */
+ flockfile(stdout);
+ if (dprintf_find_string("pid"))
+ (void) printf("%d ", getpid());
+ if (dprintf_find_string("tid"))
+ (void) printf("%u ", thr_self());
+#if 0
+ if (dprintf_find_string("cpu"))
+ (void) printf("%u ", getcpuid());
+#endif
+ if (dprintf_find_string("time"))
+ (void) printf("%llu ", gethrtime());
+ if (dprintf_find_string("long"))
+ (void) printf("%s, line %d: ", newfile, line);
+ (void) printf("%s: ", func);
+ va_start(adx, fmt);
+ (void) vprintf(fmt, adx);
+ va_end(adx);
+ funlockfile(stdout);
+ }
+}
+
+#endif /* ZFS_DEBUG */
+
+/*
+ * =========================================================================
+ * cmn_err() and panic()
+ * =========================================================================
+ */
+static char ce_prefix[CE_IGNORE][10] = { "", "NOTICE: ", "WARNING: ", "" };
+static char ce_suffix[CE_IGNORE][2] = { "", "\n", "\n", "" };
+
+void
+vpanic(const char *fmt, va_list adx)
+{
+ (void) fprintf(stderr, "error: ");
+ (void) vfprintf(stderr, fmt, adx);
+ (void) fprintf(stderr, "\n");
+
+ abort(); /* think of it as a "user-level crash dump" */
+}
+
+void
+panic(const char *fmt, ...)
+{
+ va_list adx;
+
+ va_start(adx, fmt);
+ vpanic(fmt, adx);
+ va_end(adx);
+}
+
+void
+vcmn_err(int ce, const char *fmt, va_list adx)
+{
+ if (ce == CE_PANIC)
+ vpanic(fmt, adx);
+ if (ce != CE_NOTE) { /* suppress noise in userland stress testing */
+ (void) fprintf(stderr, "%s", ce_prefix[ce]);
+ (void) vfprintf(stderr, fmt, adx);
+ (void) fprintf(stderr, "%s", ce_suffix[ce]);
+ }
+}
+
+/*PRINTFLIKE2*/
+void
+cmn_err(int ce, const char *fmt, ...)
+{
+ va_list adx;
+
+ va_start(adx, fmt);
+ vcmn_err(ce, fmt, adx);
+ va_end(adx);
+}
+
+/*
+ * =========================================================================
+ * kobj interfaces
+ * =========================================================================
+ */
+struct _buf *
+kobj_open_file(char *name)
+{
+ struct _buf *file;
+ vnode_t *vp;
+
+ /* set vp as the _fd field of the file */
+ if (vn_openat(name, UIO_SYSSPACE, FREAD, 0, &vp, 0, 0, rootdir) != 0)
+ return ((void *)-1UL);
+
+ file = umem_zalloc(sizeof (struct _buf), UMEM_NOFAIL);
+ file->_fd = (intptr_t)vp;
+ return (file);
+}
+
+int
+kobj_read_file(struct _buf *file, char *buf, unsigned size, unsigned off)
+{
+ ssize_t resid;
+
+ vn_rdwr(UIO_READ, (vnode_t *)file->_fd, buf, size, (offset_t)off,
+ UIO_SYSSPACE, 0, 0, 0, &resid);
+
+ return (size - resid);
+}
+
+void
+kobj_close_file(struct _buf *file)
+{
+ vn_close((vnode_t *)file->_fd);
+ umem_free(file, sizeof (struct _buf));
+}
+
+int
+kobj_get_filesize(struct _buf *file, uint64_t *size)
+{
+ struct stat64 st;
+ vnode_t *vp = (vnode_t *)file->_fd;
+
+ if (fstat64(vp->v_fd, &st) == -1) {
+ vn_close(vp);
+ return (errno);
+ }
+ *size = st.st_size;
+ return (0);
+}
+
+/*
+ * =========================================================================
+ * misc routines
+ * =========================================================================
+ */
+
+void
+delay(clock_t ticks)
+{
+ poll(0, 0, ticks * (1000 / hz));
+}
+
+#if 0
+/*
+ * Find highest one bit set.
+ * Returns bit number + 1 of highest bit that is set, otherwise returns 0.
+ * High order bit is 31 (or 63 in _LP64 kernel).
+ */
+int
+highbit(ulong_t i)
+{
+ register int h = 1;
+
+ if (i == 0)
+ return (0);
+#ifdef _LP64
+ if (i & 0xffffffff00000000ul) {
+ h += 32; i >>= 32;
+ }
+#endif
+ if (i & 0xffff0000) {
+ h += 16; i >>= 16;
+ }
+ if (i & 0xff00) {
+ h += 8; i >>= 8;
+ }
+ if (i & 0xf0) {
+ h += 4; i >>= 4;
+ }
+ if (i & 0xc) {
+ h += 2; i >>= 2;
+ }
+ if (i & 0x2) {
+ h += 1;
+ }
+ return (h);
+}
+#endif
+
+static int
+random_get_bytes_common(uint8_t *ptr, size_t len, char *devname)
+{
+ int fd = open(devname, O_RDONLY);
+ size_t resid = len;
+ ssize_t bytes;
+
+ ASSERT(fd != -1);
+
+ while (resid != 0) {
+ bytes = read(fd, ptr, resid);
+ ASSERT(bytes >= 0);
+ ptr += bytes;
+ resid -= bytes;
+ }
+
+ close(fd);
+
+ return (0);
+}
+
+int
+random_get_bytes(uint8_t *ptr, size_t len)
+{
+ return (random_get_bytes_common(ptr, len, "/dev/random"));
+}
+
+int
+random_get_pseudo_bytes(uint8_t *ptr, size_t len)
+{
+ return (random_get_bytes_common(ptr, len, "/dev/urandom"));
+}
+
+/*
+ * =========================================================================
+ * kernel emulation setup & teardown
+ * =========================================================================
+ */
+static int
+umem_out_of_memory(void)
+{
+ char errmsg[] = "out of memory -- generating core dump\n";
+
+ write(fileno(stderr), errmsg, sizeof (errmsg));
+ abort();
+ return (0);
+}
+
+void
+kernel_init(int mode)
+{
+ umem_nofail_callback(umem_out_of_memory);
+
+ physmem = sysconf(_SC_PHYS_PAGES);
+
+ dprintf("physmem = %llu pages (%.2f GB)\n", physmem,
+ (double)physmem * sysconf(_SC_PAGE_SIZE) / (1ULL << 30));
+
+ spa_init(mode);
+}
+
+void
+kernel_fini(void)
+{
+ spa_fini();
+}
+
+int
+z_uncompress(void *dst, size_t *dstlen, const void *src, size_t srclen)
+{
+ int ret;
+ uLongf len = *dstlen;
+
+ if ((ret = uncompress(dst, &len, src, srclen)) == Z_OK)
+ *dstlen = (size_t)len;
+
+ return (ret);
+}
+
+int
+z_compress_level(void *dst, size_t *dstlen, const void *src, size_t srclen,
+ int level)
+{
+ int ret;
+ uLongf len = *dstlen;
+
+ if ((ret = compress2(dst, &len, src, srclen, level)) == Z_OK)
+ *dstlen = (size_t)len;
+
+ return (ret);
+}
diff --git a/contrib/opensolaris/lib/libzpool/common/sys/zfs_context.h b/contrib/opensolaris/lib/libzpool/common/sys/zfs_context.h
new file mode 100644
index 0000000..6b9783f
--- /dev/null
+++ b/contrib/opensolaris/lib/libzpool/common/sys/zfs_context.h
@@ -0,0 +1,497 @@
+/*
+ * 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.
+ */
+
+#ifndef _SYS_ZFS_CONTEXT_H
+#define _SYS_ZFS_CONTEXT_H
+
+#pragma ident "%Z%%M% %I% %E% SMI"
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#define _SYS_MUTEX_H
+#define _SYS_RWLOCK_H
+#define _SYS_CONDVAR_H
+#define _SYS_SYSTM_H
+#define _SYS_DEBUG_H
+#define _SYS_T_LOCK_H
+#define _SYS_VNODE_H
+#define _SYS_VFS_H
+#define _SYS_SUNDDI_H
+#define _SYS_CALLB_H
+#define _SYS_SCHED_H_
+
+#include <solaris.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <stddef.h>
+#include <stdarg.h>
+#include <fcntl.h>
+#include <unistd.h>
+#include <errno.h>
+#include <string.h>
+#include <strings.h>
+#include <thread.h>
+#include <assert.h>
+#include <limits.h>
+#include <dirent.h>
+#include <time.h>
+#include <math.h>
+#include <umem.h>
+#include <vmem.h>
+#include <fsshare.h>
+#include <sys/note.h>
+#include <sys/types.h>
+#include <sys/atomic.h>
+#include <sys/sysmacros.h>
+#include <sys/bitmap.h>
+#include <sys/resource.h>
+#include <sys/byteorder.h>
+#include <sys/list.h>
+#include <sys/time.h>
+#include <sys/uio.h>
+#include <sys/mntent.h>
+#include <sys/mnttab.h>
+#include <sys/zfs_debug.h>
+#include <sys/debug.h>
+#include <sys/sdt.h>
+#include <sys/kstat.h>
+#include <sys/kernel.h>
+#include <machine/atomic.h>
+
+#define ZFS_EXPORTS_PATH "/etc/zfs/exports"
+
+/*
+ * Debugging
+ */
+
+/*
+ * Note that we are not using the debugging levels.
+ */
+
+#define CE_CONT 0 /* continuation */
+#define CE_NOTE 1 /* notice */
+#define CE_WARN 2 /* warning */
+#define CE_PANIC 3 /* panic */
+#define CE_IGNORE 4 /* print nothing */
+
+/*
+ * ZFS debugging
+ */
+
+#define ZFS_LOG(...) do { } while (0)
+
+typedef u_longlong_t rlim64_t;
+#define RLIM64_INFINITY ((rlim64_t)-3)
+
+#ifdef ZFS_DEBUG
+extern void dprintf_setup(int *argc, char **argv);
+#endif /* ZFS_DEBUG */
+
+extern void cmn_err(int, const char *, ...);
+extern void vcmn_err(int, const char *, __va_list);
+extern void panic(const char *, ...);
+extern void vpanic(const char *, __va_list);
+
+/* This definition is copied from assert.h. */
+#if defined(__STDC__)
+#if __STDC_VERSION__ - 0 >= 199901L
+#define verify(EX) (void)((EX) || \
+ (__assert_c99(#EX, __FILE__, __LINE__, __func__), 0))
+#else
+#define verify(EX) (void)((EX) || (__assert(#EX, __FILE__, __LINE__), 0))
+#endif /* __STDC_VERSION__ - 0 >= 199901L */
+#else
+#define verify(EX) (void)((EX) || (_assert("EX", __FILE__, __LINE__), 0))
+#endif /* __STDC__ */
+
+
+#define VERIFY verify
+#define ASSERT assert
+
+extern void __assert(const char *, const char *, int);
+
+#ifdef lint
+#define VERIFY3_IMPL(x, y, z, t) if (x == z) ((void)0)
+#else
+/* BEGIN CSTYLED */
+#define VERIFY3_IMPL(LEFT, OP, RIGHT, TYPE) do { \
+ const TYPE __left = (TYPE)(LEFT); \
+ const TYPE __right = (TYPE)(RIGHT); \
+ if (!(__left OP __right)) { \
+ char *__buf = alloca(256); \
+ (void) snprintf(__buf, 256, "%s %s %s (0x%llx %s 0x%llx)", \
+ #LEFT, #OP, #RIGHT, \
+ (u_longlong_t)__left, #OP, (u_longlong_t)__right); \
+ __assert(__buf, __FILE__, __LINE__); \
+ } \
+_NOTE(CONSTCOND) } while (0)
+/* END CSTYLED */
+#endif /* lint */
+
+#define VERIFY3S(x, y, z) VERIFY3_IMPL(x, y, z, int64_t)
+#define VERIFY3U(x, y, z) VERIFY3_IMPL(x, y, z, uint64_t)
+#define VERIFY3P(x, y, z) VERIFY3_IMPL(x, y, z, uintptr_t)
+
+#ifdef NDEBUG
+#define ASSERT3S(x, y, z) ((void)0)
+#define ASSERT3U(x, y, z) ((void)0)
+#define ASSERT3P(x, y, z) ((void)0)
+#else
+#define ASSERT3S(x, y, z) VERIFY3S(x, y, z)
+#define ASSERT3U(x, y, z) VERIFY3U(x, y, z)
+#define ASSERT3P(x, y, z) VERIFY3P(x, y, z)
+#endif
+
+/*
+ * Dtrace SDT probes have different signatures in userland than they do in
+ * kernel. If they're being used in kernel code, re-define them out of
+ * existence for their counterparts in libzpool.
+ */
+
+#ifdef DTRACE_PROBE1
+#undef DTRACE_PROBE1
+#define DTRACE_PROBE1(a, b, c) ((void)0)
+#endif /* DTRACE_PROBE1 */
+
+#ifdef DTRACE_PROBE2
+#undef DTRACE_PROBE2
+#define DTRACE_PROBE2(a, b, c, d, e) ((void)0)
+#endif /* DTRACE_PROBE2 */
+
+#ifdef DTRACE_PROBE3
+#undef DTRACE_PROBE3
+#define DTRACE_PROBE3(a, b, c, d, e, f, g) ((void)0)
+#endif /* DTRACE_PROBE3 */
+
+#ifdef DTRACE_PROBE4
+#undef DTRACE_PROBE4
+#define DTRACE_PROBE4(a, b, c, d, e, f, g, h, i) ((void)0)
+#endif /* DTRACE_PROBE4 */
+
+/*
+ * Threads
+ */
+#define curthread ((void *)(uintptr_t)thr_self())
+
+typedef struct kthread kthread_t;
+
+#define thread_create(stk, stksize, func, arg, len, pp, state, pri) \
+ zk_thread_create(func, arg)
+#define thread_exit() thr_exit(NULL)
+
+extern kthread_t *zk_thread_create(void (*func)(), void *arg);
+
+#define issig(why) (FALSE)
+#define ISSIG(thr, why) (FALSE)
+
+/*
+ * Mutexes
+ */
+typedef struct kmutex {
+ void *m_owner;
+ mutex_t m_lock;
+} kmutex_t;
+
+#define MUTEX_DEFAULT USYNC_THREAD
+#undef MUTEX_HELD
+#define MUTEX_HELD(m) ((m)->m_owner == curthread)
+
+/*
+ * Argh -- we have to get cheesy here because the kernel and userland
+ * have different signatures for the same routine.
+ */
+//extern int _mutex_init(mutex_t *mp, int type, void *arg);
+//extern int _mutex_destroy(mutex_t *mp);
+
+#define mutex_init(mp, b, c, d) zmutex_init((kmutex_t *)(mp))
+#define mutex_destroy(mp) zmutex_destroy((kmutex_t *)(mp))
+
+extern void zmutex_init(kmutex_t *mp);
+extern void zmutex_destroy(kmutex_t *mp);
+extern void mutex_enter(kmutex_t *mp);
+extern void mutex_exit(kmutex_t *mp);
+extern int mutex_tryenter(kmutex_t *mp);
+extern void *mutex_owner(kmutex_t *mp);
+
+/*
+ * RW locks
+ */
+typedef struct krwlock {
+ int rw_count;
+ void *rw_owner;
+ rwlock_t rw_lock;
+} krwlock_t;
+
+typedef int krw_t;
+
+#define RW_READER 0
+#define RW_WRITER 1
+#define RW_DEFAULT USYNC_THREAD
+
+#undef RW_READ_HELD
+
+#undef RW_WRITE_HELD
+#define RW_WRITE_HELD(x) ((x)->rw_owner == curthread)
+#define RW_LOCK_HELD(x) rw_lock_held(x)
+
+extern void rw_init(krwlock_t *rwlp, char *name, int type, void *arg);
+extern void rw_destroy(krwlock_t *rwlp);
+extern void rw_enter(krwlock_t *rwlp, krw_t rw);
+extern int rw_tryenter(krwlock_t *rwlp, krw_t rw);
+extern int rw_tryupgrade(krwlock_t *rwlp);
+extern void rw_exit(krwlock_t *rwlp);
+extern int rw_lock_held(krwlock_t *rwlp);
+#define rw_downgrade(rwlp) do { } while (0)
+
+/*
+ * Condition variables
+ */
+typedef cond_t kcondvar_t;
+
+#define CV_DEFAULT USYNC_THREAD
+
+extern void cv_init(kcondvar_t *cv, char *name, int type, void *arg);
+extern void cv_destroy(kcondvar_t *cv);
+extern void cv_wait(kcondvar_t *cv, kmutex_t *mp);
+extern clock_t cv_timedwait(kcondvar_t *cv, kmutex_t *mp, clock_t abstime);
+extern void cv_signal(kcondvar_t *cv);
+extern void cv_broadcast(kcondvar_t *cv);
+
+/*
+ * Kernel memory
+ */
+#define KM_SLEEP UMEM_NOFAIL
+#define KM_NOSLEEP UMEM_DEFAULT
+#define KMC_NODEBUG UMC_NODEBUG
+#define kmem_alloc(_s, _f) umem_alloc(_s, _f)
+#define kmem_zalloc(_s, _f) umem_zalloc(_s, _f)
+#define kmem_free(_b, _s) umem_free(_b, _s)
+#define kmem_cache_create(_a, _b, _c, _d, _e, _f, _g, _h, _i) \
+ umem_cache_create(_a, _b, _c, _d, _e, _f, _g, _h, _i)
+#define kmem_cache_destroy(_c) umem_cache_destroy(_c)
+#define kmem_cache_alloc(_c, _f) umem_cache_alloc(_c, _f)
+#define kmem_cache_free(_c, _b) umem_cache_free(_c, _b)
+#define kmem_debugging() 0
+#define kmem_cache_reap_now(c)
+
+typedef umem_cache_t kmem_cache_t;
+
+/*
+ * Task queues
+ */
+typedef struct taskq taskq_t;
+typedef uintptr_t taskqid_t;
+typedef void (task_func_t)(void *);
+
+#define TASKQ_PREPOPULATE 0x0001
+#define TASKQ_CPR_SAFE 0x0002 /* Use CPR safe protocol */
+#define TASKQ_DYNAMIC 0x0004 /* Use dynamic thread scheduling */
+
+#define TQ_SLEEP KM_SLEEP /* Can block for memory */
+#define TQ_NOSLEEP KM_NOSLEEP /* cannot block for memory; may fail */
+#define TQ_NOQUEUE 0x02 /* Do not enqueue if can't dispatch */
+
+extern taskq_t *taskq_create(const char *, int, pri_t, int, int, uint_t);
+extern taskqid_t taskq_dispatch(taskq_t *, task_func_t, void *, uint_t);
+extern void taskq_destroy(taskq_t *);
+extern void taskq_wait(taskq_t *);
+extern int taskq_member(taskq_t *, void *);
+
+/*
+ * vnodes
+ */
+typedef struct vnode {
+ uint64_t v_size;
+ int v_fd;
+ char *v_path;
+} vnode_t;
+
+typedef struct vattr {
+ uint_t va_mask; /* bit-mask of attributes */
+ u_offset_t va_size; /* file size in bytes */
+} vattr_t;
+
+#define AT_TYPE 0x0001
+#define AT_MODE 0x0002
+#define AT_UID 0x0004
+#define AT_GID 0x0008
+#define AT_FSID 0x0010
+#define AT_NODEID 0x0020
+#define AT_NLINK 0x0040
+#define AT_SIZE 0x0080
+#define AT_ATIME 0x0100
+#define AT_MTIME 0x0200
+#define AT_CTIME 0x0400
+#define AT_RDEV 0x0800
+#define AT_BLKSIZE 0x1000
+#define AT_NBLOCKS 0x2000
+#define AT_SEQ 0x8000
+
+#define CRCREAT 0
+
+#define VOP_CLOSE(vp, f, c, o, cr) 0
+#define VOP_PUTPAGE(vp, of, sz, fl, cr) 0
+#define VOP_GETATTR(vp, vap, fl, cr) ((vap)->va_size = (vp)->v_size, 0)
+
+#define VOP_FSYNC(vp, f, cr) fsync((vp)->v_fd)
+
+#define VN_RELE(vp) vn_close(vp)
+
+extern int vn_open(char *path, int x1, int oflags, int mode, vnode_t **vpp,
+ int x2, int x3);
+extern int vn_openat(char *path, int x1, int oflags, int mode, vnode_t **vpp,
+ int x2, int x3, vnode_t *vp);
+extern int vn_rdwr(int uio, vnode_t *vp, void *addr, ssize_t len,
+ offset_t offset, int x1, int x2, rlim64_t x3, void *x4, ssize_t *residp);
+extern void vn_close(vnode_t *vp);
+
+#define vn_remove(path, x1, x2) remove(path)
+#define vn_rename(from, to, seg) rename((from), (to))
+#define vn_is_readonly(vp) B_FALSE
+
+extern vnode_t *rootdir;
+
+#include <sys/file.h> /* for FREAD, FWRITE, etc */
+#define FTRUNC O_TRUNC
+
+/*
+ * Random stuff
+ */
+#define lbolt (gethrtime() >> 23)
+#define lbolt64 (gethrtime() >> 23)
+//#define hz 119 /* frequency when using gethrtime() >> 23 for lbolt */
+
+extern void delay(clock_t ticks);
+
+#define gethrestime_sec() time(NULL)
+
+#define max_ncpus 64
+
+#define minclsyspri 60
+#define maxclsyspri 99
+
+#define CPU_SEQID (thr_self() & (max_ncpus - 1))
+
+#define kcred NULL
+#define CRED() NULL
+
+extern uint64_t physmem;
+
+extern int highbit(ulong_t i);
+extern int random_get_bytes(uint8_t *ptr, size_t len);
+extern int random_get_pseudo_bytes(uint8_t *ptr, size_t len);
+
+extern void kernel_init(int);
+extern void kernel_fini(void);
+
+struct spa;
+extern void nicenum(uint64_t num, char *buf);
+extern void show_pool_stats(struct spa *);
+
+typedef struct callb_cpr {
+ kmutex_t *cc_lockp;
+} callb_cpr_t;
+
+#define CALLB_CPR_INIT(cp, lockp, func, name) { \
+ (cp)->cc_lockp = lockp; \
+}
+
+#define CALLB_CPR_SAFE_BEGIN(cp) { \
+ ASSERT(MUTEX_HELD((cp)->cc_lockp)); \
+}
+
+#define CALLB_CPR_SAFE_END(cp, lockp) { \
+ ASSERT(MUTEX_HELD((cp)->cc_lockp)); \
+}
+
+#define CALLB_CPR_EXIT(cp) { \
+ ASSERT(MUTEX_HELD((cp)->cc_lockp)); \
+ mutex_exit((cp)->cc_lockp); \
+}
+
+#define zone_dataset_visible(x, y) (1)
+#define INGLOBALZONE(z) (1)
+
+#ifdef __cplusplus
+}
+#endif
+
+/* ZFS Boot Related stuff. */
+
+struct _buf {
+ intptr_t _fd;
+};
+
+struct bootstat {
+ uint64_t st_size;
+};
+
+extern struct _buf *kobj_open_file(char *name);
+extern int kobj_read_file(struct _buf *file, char *buf, unsigned size,
+ unsigned off);
+extern void kobj_close_file(struct _buf *file);
+extern int kobj_get_filesize(struct _buf *file, uint64_t *size);
+/* Random compatibility stuff. */
+#define lbolt (gethrtime() >> 23)
+#define lbolt64 (gethrtime() >> 23)
+
+extern int hz;
+extern uint64_t physmem;
+
+#define gethrestime_sec() time(NULL)
+
+#define open64(...) open(__VA_ARGS__)
+#define pread64(d, p, n, o) pread(d, p, n, o)
+#define pwrite64(d, p, n, o) pwrite(d, p, n, o)
+#define readdir64(d) readdir(d)
+#define SIGPENDING(td) (0)
+#define root_mount_wait() do { } while (0)
+
+struct file {
+ void *dummy;
+};
+
+#define FCREAT O_CREAT
+#define FOFFMAX 0x0
+
+#define SX_SYSINIT(name, lock, desc)
+
+#define SYSCTL_DECL(...)
+#define SYSCTL_INT(...)
+#define SYSCTL_NODE(...)
+#ifdef TUNABLE_INT
+#undef TUNABLE_INT
+#endif
+#define TUNABLE_INT(...)
+
+/* Errors */
+
+#ifndef ERESTART
+#define ERESTART (-1)
+#endif
+
+#endif /* _SYS_ZFS_CONTEXT_H */
diff --git a/contrib/opensolaris/lib/libzpool/common/taskq.c b/contrib/opensolaris/lib/libzpool/common/taskq.c
new file mode 100644
index 0000000..f7b6571
--- /dev/null
+++ b/contrib/opensolaris/lib/libzpool/common/taskq.c
@@ -0,0 +1,250 @@
+/*
+ * CDDL HEADER START
+ *
+ * The contents of this file are subject to the terms of the
+ * Common Development and Distribution License, Version 1.0 only
+ * (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 2005 Sun Microsystems, Inc. All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#pragma ident "%Z%%M% %I% %E% SMI"
+
+#include <sys/zfs_context.h>
+
+int taskq_now;
+
+typedef struct task {
+ struct task *task_next;
+ struct task *task_prev;
+ task_func_t *task_func;
+ void *task_arg;
+} task_t;
+
+#define TASKQ_ACTIVE 0x00010000
+
+struct taskq {
+ kmutex_t tq_lock;
+ krwlock_t tq_threadlock;
+ kcondvar_t tq_dispatch_cv;
+ kcondvar_t tq_wait_cv;
+ thread_t *tq_threadlist;
+ int tq_flags;
+ int tq_active;
+ int tq_nthreads;
+ int tq_nalloc;
+ int tq_minalloc;
+ int tq_maxalloc;
+ task_t *tq_freelist;
+ task_t tq_task;
+};
+
+static task_t *
+task_alloc(taskq_t *tq, int tqflags)
+{
+ task_t *t;
+
+ if ((t = tq->tq_freelist) != NULL && tq->tq_nalloc >= tq->tq_minalloc) {
+ tq->tq_freelist = t->task_next;
+ } else {
+ mutex_exit(&tq->tq_lock);
+ if (tq->tq_nalloc >= tq->tq_maxalloc) {
+ if (!(tqflags & KM_SLEEP)) {
+ mutex_enter(&tq->tq_lock);
+ return (NULL);
+ }
+ /*
+ * We don't want to exceed tq_maxalloc, but we can't
+ * wait for other tasks to complete (and thus free up
+ * task structures) without risking deadlock with
+ * the caller. So, we just delay for one second
+ * to throttle the allocation rate.
+ */
+ delay(hz);
+ }
+ t = kmem_alloc(sizeof (task_t), tqflags);
+ mutex_enter(&tq->tq_lock);
+ if (t != NULL)
+ tq->tq_nalloc++;
+ }
+ return (t);
+}
+
+static void
+task_free(taskq_t *tq, task_t *t)
+{
+ if (tq->tq_nalloc <= tq->tq_minalloc) {
+ t->task_next = tq->tq_freelist;
+ tq->tq_freelist = t;
+ } else {
+ tq->tq_nalloc--;
+ mutex_exit(&tq->tq_lock);
+ kmem_free(t, sizeof (task_t));
+ mutex_enter(&tq->tq_lock);
+ }
+}
+
+taskqid_t
+taskq_dispatch(taskq_t *tq, task_func_t func, void *arg, uint_t tqflags)
+{
+ task_t *t;
+
+ if (taskq_now) {
+ func(arg);
+ return (1);
+ }
+
+ mutex_enter(&tq->tq_lock);
+ ASSERT(tq->tq_flags & TASKQ_ACTIVE);
+ if ((t = task_alloc(tq, tqflags)) == NULL) {
+ mutex_exit(&tq->tq_lock);
+ return (0);
+ }
+ t->task_next = &tq->tq_task;
+ t->task_prev = tq->tq_task.task_prev;
+ t->task_next->task_prev = t;
+ t->task_prev->task_next = t;
+ t->task_func = func;
+ t->task_arg = arg;
+ cv_signal(&tq->tq_dispatch_cv);
+ mutex_exit(&tq->tq_lock);
+ return (1);
+}
+
+void
+taskq_wait(taskq_t *tq)
+{
+ mutex_enter(&tq->tq_lock);
+ while (tq->tq_task.task_next != &tq->tq_task || tq->tq_active != 0)
+ cv_wait(&tq->tq_wait_cv, &tq->tq_lock);
+ mutex_exit(&tq->tq_lock);
+}
+
+static void *
+taskq_thread(void *arg)
+{
+ taskq_t *tq = arg;
+ task_t *t;
+
+ mutex_enter(&tq->tq_lock);
+ while (tq->tq_flags & TASKQ_ACTIVE) {
+ if ((t = tq->tq_task.task_next) == &tq->tq_task) {
+ if (--tq->tq_active == 0)
+ cv_broadcast(&tq->tq_wait_cv);
+ cv_wait(&tq->tq_dispatch_cv, &tq->tq_lock);
+ tq->tq_active++;
+ continue;
+ }
+ t->task_prev->task_next = t->task_next;
+ t->task_next->task_prev = t->task_prev;
+ mutex_exit(&tq->tq_lock);
+
+ rw_enter(&tq->tq_threadlock, RW_READER);
+ t->task_func(t->task_arg);
+ rw_exit(&tq->tq_threadlock);
+
+ mutex_enter(&tq->tq_lock);
+ task_free(tq, t);
+ }
+ tq->tq_nthreads--;
+ cv_broadcast(&tq->tq_wait_cv);
+ mutex_exit(&tq->tq_lock);
+ return (NULL);
+}
+
+/*ARGSUSED*/
+taskq_t *
+taskq_create(const char *name, int nthreads, pri_t pri,
+ int minalloc, int maxalloc, uint_t flags)
+{
+ taskq_t *tq = kmem_zalloc(sizeof (taskq_t), KM_SLEEP);
+ int t;
+
+ rw_init(&tq->tq_threadlock, NULL, RW_DEFAULT, NULL);
+ tq->tq_flags = flags | TASKQ_ACTIVE;
+ tq->tq_active = nthreads;
+ tq->tq_nthreads = nthreads;
+ tq->tq_minalloc = minalloc;
+ tq->tq_maxalloc = maxalloc;
+ tq->tq_task.task_next = &tq->tq_task;
+ tq->tq_task.task_prev = &tq->tq_task;
+ tq->tq_threadlist = kmem_alloc(nthreads * sizeof (thread_t), KM_SLEEP);
+
+ if (flags & TASKQ_PREPOPULATE) {
+ mutex_enter(&tq->tq_lock);
+ while (minalloc-- > 0)
+ task_free(tq, task_alloc(tq, KM_SLEEP));
+ mutex_exit(&tq->tq_lock);
+ }
+
+ for (t = 0; t < nthreads; t++)
+ (void) thr_create(0, 0, taskq_thread,
+ tq, THR_BOUND, &tq->tq_threadlist[t]);
+
+ return (tq);
+}
+
+void
+taskq_destroy(taskq_t *tq)
+{
+ int t;
+ int nthreads = tq->tq_nthreads;
+
+ taskq_wait(tq);
+
+ mutex_enter(&tq->tq_lock);
+
+ tq->tq_flags &= ~TASKQ_ACTIVE;
+ cv_broadcast(&tq->tq_dispatch_cv);
+
+ while (tq->tq_nthreads != 0)
+ cv_wait(&tq->tq_wait_cv, &tq->tq_lock);
+
+ tq->tq_minalloc = 0;
+ while (tq->tq_nalloc != 0) {
+ ASSERT(tq->tq_freelist != NULL);
+ task_free(tq, task_alloc(tq, KM_SLEEP));
+ }
+
+ mutex_exit(&tq->tq_lock);
+
+ for (t = 0; t < nthreads; t++)
+ (void) thr_join(tq->tq_threadlist[t], NULL, NULL);
+
+ kmem_free(tq->tq_threadlist, nthreads * sizeof (thread_t));
+
+ rw_destroy(&tq->tq_threadlock);
+
+ kmem_free(tq, sizeof (taskq_t));
+}
+
+int
+taskq_member(taskq_t *tq, void *t)
+{
+ int i;
+
+ if (taskq_now)
+ return (1);
+
+ for (i = 0; i < tq->tq_nthreads; i++)
+ if (tq->tq_threadlist[i] == (thread_t)(uintptr_t)t)
+ return (1);
+
+ return (0);
+}
diff --git a/contrib/opensolaris/lib/libzpool/common/util.c b/contrib/opensolaris/lib/libzpool/common/util.c
new file mode 100644
index 0000000..df49adb
--- /dev/null
+++ b/contrib/opensolaris/lib/libzpool/common/util.c
@@ -0,0 +1,144 @@
+/*
+ * 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 2006 Sun Microsystems, Inc. All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#pragma ident "%Z%%M% %I% %E% SMI"
+
+#include <assert.h>
+#include <sys/zfs_context.h>
+#include <sys/avl.h>
+#include <string.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <sys/spa.h>
+#include <sys/fs/zfs.h>
+#include <sys/refcount.h>
+
+/*
+ * Routines needed by more than one client of libzpool.
+ */
+
+void
+nicenum(uint64_t num, char *buf)
+{
+ uint64_t n = num;
+ int index = 0;
+ char u;
+
+ while (n >= 1024) {
+ n = (n + (1024 / 2)) / 1024; /* Round up or down */
+ index++;
+ }
+
+ u = " KMGTPE"[index];
+
+ if (index == 0) {
+ (void) sprintf(buf, "%llu", (u_longlong_t)n);
+ } else if (n < 10 && (num & (num - 1)) != 0) {
+ (void) sprintf(buf, "%.2f%c",
+ (double)num / (1ULL << 10 * index), u);
+ } else if (n < 100 && (num & (num - 1)) != 0) {
+ (void) sprintf(buf, "%.1f%c",
+ (double)num / (1ULL << 10 * index), u);
+ } else {
+ (void) sprintf(buf, "%llu%c", (u_longlong_t)n, u);
+ }
+}
+
+static void
+show_vdev_stats(const char *desc, nvlist_t *nv, int indent)
+{
+ nvlist_t **child;
+ uint_t c, children;
+ vdev_stat_t *vs;
+ uint64_t sec;
+ char used[6], avail[6];
+ char rops[6], wops[6], rbytes[6], wbytes[6], rerr[6], werr[6], cerr[6];
+
+ if (indent == 0) {
+ (void) printf(" "
+ " capacity operations bandwidth ---- errors ----\n");
+ (void) printf("description "
+ "used avail read write read write read write cksum\n");
+ }
+
+ VERIFY(nvlist_lookup_uint64_array(nv, ZPOOL_CONFIG_STATS,
+ (uint64_t **)&vs, &c) == 0);
+
+ sec = MAX(1, vs->vs_timestamp / NANOSEC);
+
+ nicenum(vs->vs_alloc, used);
+ nicenum(vs->vs_space - vs->vs_alloc, avail);
+ nicenum(vs->vs_ops[ZIO_TYPE_READ] / sec, rops);
+ nicenum(vs->vs_ops[ZIO_TYPE_WRITE] / sec, wops);
+ nicenum(vs->vs_bytes[ZIO_TYPE_READ] / sec, rbytes);
+ nicenum(vs->vs_bytes[ZIO_TYPE_WRITE] / sec, wbytes);
+ nicenum(vs->vs_read_errors, rerr);
+ nicenum(vs->vs_write_errors, werr);
+ nicenum(vs->vs_checksum_errors, cerr);
+
+ (void) printf("%*s%*s%*s%*s %5s %5s %5s %5s %5s %5s %5s\n",
+ indent, "",
+ indent - 19 - (vs->vs_space ? 0 : 12), desc,
+ vs->vs_space ? 6 : 0, vs->vs_space ? used : "",
+ vs->vs_space ? 6 : 0, vs->vs_space ? avail : "",
+ rops, wops, rbytes, wbytes, rerr, werr, cerr);
+
+ if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
+ &child, &children) != 0)
+ return;
+
+ for (c = 0; c < children; c++) {
+ nvlist_t *cnv = child[c];
+ char *cname, *tname;
+ uint64_t np;
+ if (nvlist_lookup_string(cnv, ZPOOL_CONFIG_PATH, &cname) &&
+ nvlist_lookup_string(cnv, ZPOOL_CONFIG_TYPE, &cname))
+ cname = "<unknown>";
+ tname = calloc(1, strlen(cname) + 2);
+ (void) strcpy(tname, cname);
+ if (nvlist_lookup_uint64(cnv, ZPOOL_CONFIG_NPARITY, &np) == 0)
+ tname[strlen(tname)] = '0' + np;
+ show_vdev_stats(tname, cnv, indent + 2);
+ free(tname);
+ }
+}
+
+void
+show_pool_stats(spa_t *spa)
+{
+ nvlist_t *config, *nvroot;
+ char *name;
+
+ spa_config_enter(spa, RW_READER, FTAG);
+ config = spa_config_generate(spa, NULL, -1ULL, B_TRUE);
+ spa_config_exit(spa, FTAG);
+
+ VERIFY(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE,
+ &nvroot) == 0);
+ VERIFY(nvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME,
+ &name) == 0);
+
+ show_vdev_stats(name, nvroot, 0);
+}
OpenPOWER on IntegriCloud