*
* CDDL HEADER END
*/
+
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Portions Copyright 2011 Martin Matuska
+ * Portions Copyright 2012 Pawel Jakub Dawidek <pawel@dawidek.net>
+ * Copyright (c) 2012, Joyent, Inc. All rights reserved.
+ * Copyright 2015 Nexenta Systems, Inc. All rights reserved.
+ * Copyright (c) 2014, Joyent, Inc. All rights reserved.
+ * Copyright (c) 2011, 2015 by Delphix. All rights reserved.
+ * Copyright (c) 2013 by Saso Kiselkov. All rights reserved.
+ * Copyright (c) 2013 Steven Hartland. All rights reserved.
+ * Copyright (c) 2016 Actifio, Inc. All rights reserved.
+ */
+
+/*
+ * ZFS ioctls.
+ *
+ * This file handles the ioctls to /dev/zfs, used for configuring ZFS storage
+ * pools and filesystems, e.g. with /sbin/zfs and /sbin/zpool.
+ *
+ * There are two ways that we handle ioctls: the legacy way where almost
+ * all of the logic is in the ioctl callback, and the new way where most
+ * of the marshalling is handled in the common entry point, zfsdev_ioctl().
+ *
+ * Non-legacy ioctls should be registered by calling
+ * zfs_ioctl_register() from zfs_ioctl_init(). The ioctl is invoked
+ * from userland by lzc_ioctl().
+ *
+ * The registration arguments are as follows:
+ *
+ * const char *name
+ * The name of the ioctl. This is used for history logging. If the
+ * ioctl returns successfully (the callback returns 0), and allow_log
+ * is true, then a history log entry will be recorded with the input &
+ * output nvlists. The log entry can be printed with "zpool history -i".
+ *
+ * zfs_ioc_t ioc
+ * The ioctl request number, which userland will pass to ioctl(2).
+ * The ioctl numbers can change from release to release, because
+ * the caller (libzfs) must be matched to the kernel.
+ *
+ * zfs_secpolicy_func_t *secpolicy
+ * This function will be called before the zfs_ioc_func_t, to
+ * determine if this operation is permitted. It should return EPERM
+ * on failure, and 0 on success. Checks include determining if the
+ * dataset is visible in this zone, and if the user has either all
+ * zfs privileges in the zone (SYS_MOUNT), or has been granted permission
+ * to do this operation on this dataset with "zfs allow".
+ *
+ * zfs_ioc_namecheck_t namecheck
+ * This specifies what to expect in the zfs_cmd_t:zc_name -- a pool
+ * name, a dataset name, or nothing. If the name is not well-formed,
+ * the ioctl will fail and the callback will not be called.
+ * Therefore, the callback can assume that the name is well-formed
+ * (e.g. is null-terminated, doesn't have more than one '@' character,
+ * doesn't have invalid characters).
+ *
+ * zfs_ioc_poolcheck_t pool_check
+ * This specifies requirements on the pool state. If the pool does
+ * not meet them (is suspended or is readonly), the ioctl will fail
+ * and the callback will not be called. If any checks are specified
+ * (i.e. it is not POOL_CHECK_NONE), namecheck must not be NO_NAME.
+ * Multiple checks can be or-ed together (e.g. POOL_CHECK_SUSPENDED |
+ * POOL_CHECK_READONLY).
+ *
+ * boolean_t smush_outnvlist
+ * If smush_outnvlist is true, then the output is presumed to be a
+ * list of errors, and it will be "smushed" down to fit into the
+ * caller's buffer, by removing some entries and replacing them with a
+ * single "N_MORE_ERRORS" entry indicating how many were removed. See
+ * nvlist_smush() for details. If smush_outnvlist is false, and the
+ * outnvlist does not fit into the userland-provided buffer, then the
+ * ioctl will fail with ENOMEM.
+ *
+ * zfs_ioc_func_t *func
+ * The callback function that will perform the operation.
+ *
+ * The callback should return 0 on success, or an error number on
+ * failure. If the function fails, the userland ioctl will return -1,
+ * and errno will be set to the callback's return value. The callback
+ * will be called with the following arguments:
+ *
+ * const char *name
+ * The name of the pool or dataset to operate on, from
+ * zfs_cmd_t:zc_name. The 'namecheck' argument specifies the
+ * expected type (pool, dataset, or none).
+ *
+ * nvlist_t *innvl
+ * The input nvlist, deserialized from zfs_cmd_t:zc_nvlist_src. Or
+ * NULL if no input nvlist was provided. Changes to this nvlist are
+ * ignored. If the input nvlist could not be deserialized, the
+ * ioctl will fail and the callback will not be called.
+ *
+ * nvlist_t *outnvl
+ * The output nvlist, initially empty. The callback can fill it in,
+ * and it will be returned to userland by serializing it into
+ * zfs_cmd_t:zc_nvlist_dst. If it is non-empty, and serialization
+ * fails (e.g. because the caller didn't supply a large enough
+ * buffer), then the overall ioctl will fail. See the
+ * 'smush_nvlist' argument above for additional behaviors.
+ *
+ * There are two typical uses of the output nvlist:
+ * - To return state, e.g. property values. In this case,
+ * smush_outnvlist should be false. If the buffer was not large
+ * enough, the caller will reallocate a larger buffer and try
+ * the ioctl again.
+ *
+ * - To return multiple errors from an ioctl which makes on-disk
+ * changes. In this case, smush_outnvlist should be true.
+ * Ioctls which make on-disk modifications should generally not
+ * use the outnvl if they succeed, because the caller can not
+ * distinguish between the operation failing, and
+ * deserialization failing.
*/
#include <sys/types.h>
#include <sys/dsl_prop.h>
#include <sys/dsl_deleg.h>
#include <sys/dmu_objset.h>
+#include <sys/dmu_impl.h>
+#include <sys/dmu_tx.h>
#include <sys/ddi.h>
#include <sys/sunddi.h>
#include <sys/sunldi.h>
#include <sys/mount.h>
#include <sys/sdt.h>
#include <sys/fs/zfs.h>
+#include <sys/zfs_ctldir.h>
#include <sys/zfs_dir.h>
#include <sys/zfs_onexit.h>
#include <sys/zvol.h>
#include <sys/dsl_scan.h>
#include <sharefs/share.h>
-#include <sys/dmu_objset.h>
#include <sys/fm/util.h>
+#include <sys/dmu_send.h>
+#include <sys/dsl_destroy.h>
+#include <sys/dsl_bookmark.h>
+#include <sys/dsl_userhold.h>
+#include <sys/zfeature.h>
+
#include <linux/miscdevice.h>
+#include <linux/slab.h>
#include "zfs_namecheck.h"
#include "zfs_prop.h"
#include "zfs_deleg.h"
#include "zfs_comutil.h"
+/*
+ * Limit maximum nvlist size. We don't want users passing in insane values
+ * for zc->zc_nvlist_src_size, since we will need to allocate that much memory.
+ */
+#define MAX_NVLIST_SRC_SIZE KMALLOC_MAX_SIZE
+
kmutex_t zfsdev_state_lock;
-list_t zfsdev_state_list;
+zfsdev_state_t *zfsdev_state_list;
extern void zfs_init(void);
extern void zfs_fini(void);
-typedef int zfs_ioc_func_t(zfs_cmd_t *);
-typedef int zfs_secpolicy_func_t(zfs_cmd_t *, cred_t *);
+uint_t zfs_fsyncer_key;
+extern uint_t rrw_tsd_key;
+static uint_t zfs_allow_log_key;
+
+typedef int zfs_ioc_legacy_func_t(zfs_cmd_t *);
+typedef int zfs_ioc_func_t(const char *, nvlist_t *, nvlist_t *);
+typedef int zfs_secpolicy_func_t(zfs_cmd_t *, nvlist_t *, cred_t *);
typedef enum {
NO_NAME,
typedef enum {
POOL_CHECK_NONE = 1 << 0,
POOL_CHECK_SUSPENDED = 1 << 1,
- POOL_CHECK_READONLY = 1 << 2
+ POOL_CHECK_READONLY = 1 << 2,
} zfs_ioc_poolcheck_t;
typedef struct zfs_ioc_vec {
+ zfs_ioc_legacy_func_t *zvec_legacy_func;
zfs_ioc_func_t *zvec_func;
zfs_secpolicy_func_t *zvec_secpolicy;
zfs_ioc_namecheck_t zvec_namecheck;
- boolean_t zvec_his_log;
+ boolean_t zvec_allow_log;
zfs_ioc_poolcheck_t zvec_pool_check;
+ boolean_t zvec_smush_outnvlist;
+ const char *zvec_name;
} zfs_ioc_vec_t;
/* This array is indexed by zfs_userquota_prop_t */
nvlist_t **errors);
static int zfs_fill_zplprops_root(uint64_t, nvlist_t *, nvlist_t *,
boolean_t *);
-int zfs_set_prop_nvlist(const char *, zprop_source_t, nvlist_t *, nvlist_t **);
-
-/* _NOTE(PRINTFLIKE(4)) - this is printf-like, but lint is too whiney */
-void
-__dprintf(const char *file, const char *func, int line, const char *fmt, ...)
-{
- const char *newfile;
- char buf[512];
- 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;
- }
-
- va_start(adx, fmt);
- (void) vsnprintf(buf, sizeof (buf), fmt, adx);
- va_end(adx);
-
- /*
- * To get this data, use the zfs-dprintf probe as so:
- * dtrace -q -n 'zfs-dprintf \
- * /stringof(arg0) == "dbuf.c"/ \
- * {printf("%s: %s", stringof(arg1), stringof(arg3))}'
- * arg0 = file name
- * arg1 = function name
- * arg2 = line number
- * arg3 = message
- */
- DTRACE_PROBE4(zfs__dprintf,
- char *, newfile, char *, func, int, line, char *, buf);
-}
+int zfs_set_prop_nvlist(const char *, zprop_source_t, nvlist_t *, nvlist_t *);
+static int get_nvlist(uint64_t nvl, uint64_t size, int iflag, nvlist_t **nvp);
static void
history_str_free(char *buf)
if (zc->zc_history == 0)
return (NULL);
- buf = kmem_alloc(HIS_MAX_RECORD_LEN, KM_SLEEP | KM_NODEBUG);
+ buf = kmem_alloc(HIS_MAX_RECORD_LEN, KM_SLEEP);
if (copyinstr((void *)(uintptr_t)zc->zc_history,
buf, HIS_MAX_RECORD_LEN, NULL) != 0) {
history_str_free(buf);
}
/*
- * zfs_earlier_version
- *
- * Return non-zero if the spa version is less than requested version.
+ * Return non-zero if the spa version is less than requested version.
*/
static int
zfs_earlier_version(const char *name, int version)
}
/*
- * zpl_earlier_version
- *
* Return TRUE if the ZPL version is less than requested version.
*/
static boolean_t
if (spa_open(zc->zc_name, &spa, FTAG) == 0) {
if (spa_version(spa) >= SPA_VERSION_ZPOOL_HISTORY)
- (void) spa_history_log(spa, buf, LOG_CMD_NORMAL);
+ (void) spa_history_log(spa, buf);
spa_close(spa, FTAG);
}
history_str_free(buf);
*/
/* ARGSUSED */
static int
-zfs_secpolicy_none(zfs_cmd_t *zc, cred_t *cr)
+zfs_secpolicy_none(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr)
{
return (0);
}
*/
/* ARGSUSED */
static int
-zfs_secpolicy_read(zfs_cmd_t *zc, cred_t *cr)
+zfs_secpolicy_read(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr)
{
if (INGLOBALZONE(curproc) ||
zone_dataset_visible(zc->zc_name, NULL))
return (0);
- return (ENOENT);
+ return (SET_ERROR(ENOENT));
}
static int
*/
if (!INGLOBALZONE(curproc) &&
!zone_dataset_visible(dataset, &writable))
- return (ENOENT);
+ return (SET_ERROR(ENOENT));
if (INGLOBALZONE(curproc)) {
/*
* global zone.
*/
if (secpolicy_zfs(cr) && zoned)
- return (EPERM);
+ return (SET_ERROR(EPERM));
} else {
/*
* If we are in a local zone, the 'zoned' property must be set.
*/
if (!zoned)
- return (EPERM);
+ return (SET_ERROR(EPERM));
/* must be writable by this zone */
if (!writable)
- return (EPERM);
+ return (SET_ERROR(EPERM));
}
return (0);
}
uint64_t zoned;
if (dsl_prop_get_integer(dataset, "zoned", &zoned, NULL))
- return (ENOENT);
+ return (SET_ERROR(ENOENT));
return (zfs_dozonecheck_impl(dataset, zoned, cr));
}
{
uint64_t zoned;
- rw_enter(&ds->ds_dir->dd_pool->dp_config_rwlock, RW_READER);
- if (dsl_prop_get_ds(ds, "zoned", 8, 1, &zoned, NULL)) {
- rw_exit(&ds->ds_dir->dd_pool->dp_config_rwlock);
- return (ENOENT);
- }
- rw_exit(&ds->ds_dir->dd_pool->dp_config_rwlock);
+ if (dsl_prop_get_int_ds(ds, "zoned", &zoned))
+ return (SET_ERROR(ENOENT));
return (zfs_dozonecheck_impl(dataset, zoned, cr));
}
-int
-zfs_secpolicy_write_perms(const char *name, const char *perm, cred_t *cr)
+static int
+zfs_secpolicy_write_perms_ds(const char *name, dsl_dataset_t *ds,
+ const char *perm, cred_t *cr)
{
int error;
- error = zfs_dozonecheck(name, cr);
+ error = zfs_dozonecheck_ds(name, ds, cr);
if (error == 0) {
error = secpolicy_zfs(cr);
- if (error)
- error = dsl_deleg_access(name, perm, cr);
+ if (error != 0)
+ error = dsl_deleg_access_impl(ds, perm, cr);
}
return (error);
}
-int
-zfs_secpolicy_write_perms_ds(const char *name, dsl_dataset_t *ds,
- const char *perm, cred_t *cr)
+static int
+zfs_secpolicy_write_perms(const char *name, const char *perm, cred_t *cr)
{
int error;
+ dsl_dataset_t *ds;
+ dsl_pool_t *dp;
- error = zfs_dozonecheck_ds(name, ds, cr);
- if (error == 0) {
- error = secpolicy_zfs(cr);
- if (error)
- error = dsl_deleg_access_impl(ds, perm, cr);
+ error = dsl_pool_hold(name, FTAG, &dp);
+ if (error != 0)
+ return (error);
+
+ error = dsl_dataset_hold(dp, name, FTAG, &ds);
+ if (error != 0) {
+ dsl_pool_rele(dp, FTAG);
+ return (error);
}
+
+ error = zfs_secpolicy_write_perms_ds(name, ds, perm, cr);
+
+ dsl_dataset_rele(ds, FTAG);
+ dsl_pool_rele(dp, FTAG);
return (error);
}
/* First get the existing dataset label. */
error = dsl_prop_get(name, zfs_prop_to_name(ZFS_PROP_MLSLABEL),
1, sizeof (ds_hexsl), &ds_hexsl, NULL);
- if (error)
- return (EPERM);
+ if (error != 0)
+ return (SET_ERROR(EPERM));
if (strcasecmp(strval, ZFS_MLSLABEL_DEFAULT) == 0)
new_default = TRUE;
/* The label must be translatable */
if (!new_default && (hexstr_to_label(strval, &new_sl) != 0))
- return (EINVAL);
+ return (SET_ERROR(EINVAL));
/*
* In a non-global zone, disallow attempts to set a label that
*/
if (!INGLOBALZONE(curproc)) {
if (new_default || !blequal(&new_sl, CR_SL(CRED())))
- return (EPERM);
+ return (SET_ERROR(EPERM));
return (0);
}
*/
if (dsl_prop_get_integer(name,
zfs_prop_to_name(ZFS_PROP_ZONED), &zoned, NULL))
- return (EPERM);
+ return (SET_ERROR(EPERM));
if (!zoned) {
if (zfs_check_global_label(name, strval) != 0)
- return (EPERM);
+ return (SET_ERROR(EPERM));
}
/*
*/
error = dmu_objset_own(name, DMU_OST_ZFS, B_TRUE,
setsl_tag, &os);
- if (error)
- return (EPERM);
+ if (error != 0)
+ return (SET_ERROR(EPERM));
dmu_objset_disown(os, setsl_tag);
}
if (hexstr_to_label(strval, &new_sl) != 0)
- return (EPERM);
+ return (SET_ERROR(EPERM));
if (blstrictdom(&ds_sl, &new_sl))
needed_priv = PRIV_FILE_DOWNGRADE_SL;
return (PRIV_POLICY(cr, needed_priv, B_FALSE, EPERM, NULL));
return (0);
#else
- return ENOTSUP;
+ return (ENOTSUP);
#endif /* HAVE_MLSLABEL */
}
* Disallow setting of 'zoned' from within a local zone.
*/
if (!INGLOBALZONE(curproc))
- return (EPERM);
+ return (SET_ERROR(EPERM));
break;
case ZFS_PROP_QUOTA:
+ case ZFS_PROP_FILESYSTEM_LIMIT:
+ case ZFS_PROP_SNAPSHOT_LIMIT:
if (!INGLOBALZONE(curproc)) {
uint64_t zoned;
char setpoint[MAXNAMELEN];
/*
* Unprivileged users are allowed to modify the
- * quota on things *under* (ie. contained by)
+ * limit on things *under* (ie. contained by)
* the thing they own.
*/
if (dsl_prop_get_integer(dsname, "zoned", &zoned,
setpoint))
- return (EPERM);
+ return (SET_ERROR(EPERM));
if (!zoned || strlen(dsname) <= strlen(setpoint))
- return (EPERM);
+ return (SET_ERROR(EPERM));
}
break;
case ZFS_PROP_MLSLABEL:
if (!is_system_labeled())
- return (EPERM);
+ return (SET_ERROR(EPERM));
if (nvpair_value_string(propval, &strval) == 0) {
int err;
return (zfs_secpolicy_write_perms(dsname, zfs_prop_to_name(prop), cr));
}
-int
-zfs_secpolicy_fsacl(zfs_cmd_t *zc, cred_t *cr)
+/* ARGSUSED */
+static int
+zfs_secpolicy_set_fsacl(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr)
{
int error;
error = zfs_dozonecheck(zc->zc_name, cr);
- if (error)
+ if (error != 0)
return (error);
/*
return (0);
}
-int
-zfs_secpolicy_rollback(zfs_cmd_t *zc, cred_t *cr)
+/* ARGSUSED */
+static int
+zfs_secpolicy_rollback(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr)
{
return (zfs_secpolicy_write_perms(zc->zc_name,
ZFS_DELEG_PERM_ROLLBACK, cr));
}
-int
-zfs_secpolicy_send(zfs_cmd_t *zc, cred_t *cr)
+/* ARGSUSED */
+static int
+zfs_secpolicy_send(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr)
{
- spa_t *spa;
dsl_pool_t *dp;
dsl_dataset_t *ds;
char *cp;
*/
cp = strchr(zc->zc_name, '@');
if (cp == NULL)
- return (EINVAL);
- error = spa_open(zc->zc_name, &spa, FTAG);
- if (error)
+ return (SET_ERROR(EINVAL));
+ error = dsl_pool_hold(zc->zc_name, FTAG, &dp);
+ if (error != 0)
return (error);
- dp = spa_get_dsl(spa);
- rw_enter(&dp->dp_config_rwlock, RW_READER);
error = dsl_dataset_hold_obj(dp, zc->zc_sendobj, FTAG, &ds);
- rw_exit(&dp->dp_config_rwlock);
- spa_close(spa, FTAG);
- if (error)
+ if (error != 0) {
+ dsl_pool_rele(dp, FTAG);
return (error);
+ }
dsl_dataset_name(ds, zc->zc_name);
error = zfs_secpolicy_write_perms_ds(zc->zc_name, ds,
ZFS_DELEG_PERM_SEND, cr);
dsl_dataset_rele(ds, FTAG);
+ dsl_pool_rele(dp, FTAG);
return (error);
}
+/* ARGSUSED */
+static int
+zfs_secpolicy_send_new(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr)
+{
+ return (zfs_secpolicy_write_perms(zc->zc_name,
+ ZFS_DELEG_PERM_SEND, cr));
+}
+
#ifdef HAVE_SMB_SHARE
+/* ARGSUSED */
static int
-zfs_secpolicy_deleg_share(zfs_cmd_t *zc, cred_t *cr)
+zfs_secpolicy_deleg_share(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr)
{
vnode_t *vp;
int error;
(strcmp((char *)refstr_value(vp->v_vfsp->vfs_resource),
zc->zc_name) != 0)) {
VN_RELE(vp);
- return (EPERM);
+ return (SET_ERROR(EPERM));
}
VN_RELE(vp);
#endif /* HAVE_SMB_SHARE */
int
-zfs_secpolicy_share(zfs_cmd_t *zc, cred_t *cr)
+zfs_secpolicy_share(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr)
{
#ifdef HAVE_SMB_SHARE
if (!INGLOBALZONE(curproc))
- return (EPERM);
+ return (SET_ERROR(EPERM));
if (secpolicy_nfs(cr) == 0) {
return (0);
} else {
- return (zfs_secpolicy_deleg_share(zc, cr));
+ return (zfs_secpolicy_deleg_share(zc, innvl, cr));
}
#else
- return (ENOTSUP);
+ return (SET_ERROR(ENOTSUP));
#endif /* HAVE_SMB_SHARE */
}
int
-zfs_secpolicy_smb_acl(zfs_cmd_t *zc, cred_t *cr)
+zfs_secpolicy_smb_acl(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr)
{
#ifdef HAVE_SMB_SHARE
if (!INGLOBALZONE(curproc))
- return (EPERM);
+ return (SET_ERROR(EPERM));
if (secpolicy_smb(cr) == 0) {
return (0);
} else {
- return (zfs_secpolicy_deleg_share(zc, cr));
+ return (zfs_secpolicy_deleg_share(zc, innvl, cr));
}
#else
- return (ENOTSUP);
+ return (SET_ERROR(ENOTSUP));
#endif /* HAVE_SMB_SHARE */
}
} else {
cp = strrchr(parent, '/');
if (cp == NULL)
- return (ENOENT);
+ return (SET_ERROR(ENOENT));
cp[0] = '\0';
}
return (zfs_secpolicy_write_perms(name, ZFS_DELEG_PERM_DESTROY, cr));
}
+/* ARGSUSED */
static int
-zfs_secpolicy_destroy(zfs_cmd_t *zc, cred_t *cr)
+zfs_secpolicy_destroy(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr)
{
return (zfs_secpolicy_destroy_perms(zc->zc_name, cr));
}
/*
* Destroying snapshots with delegated permissions requires
- * descendent mount and destroy permissions.
- * Reassemble the full filesystem@snap name so dsl_deleg_access()
- * can do the correct permission check.
- *
- * Since this routine is used when doing a recursive destroy of snapshots
- * and destroying snapshots requires descendent permissions, a successfull
- * check of the top level snapshot applies to snapshots of all descendent
- * datasets as well.
- *
- * The target snapshot may not exist when doing a recursive destroy.
- * In this case fallback to permissions of the parent dataset.
+ * descendant mount and destroy permissions.
*/
+/* ARGSUSED */
static int
-zfs_secpolicy_destroy_snaps(zfs_cmd_t *zc, cred_t *cr)
+zfs_secpolicy_destroy_snaps(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr)
{
- int error;
- char *dsname;
-
- dsname = kmem_asprintf("%s@%s", zc->zc_name, zc->zc_value);
+ nvlist_t *snaps;
+ nvpair_t *pair, *nextpair;
+ int error = 0;
- error = zfs_secpolicy_destroy_perms(dsname, cr);
- if (error == ENOENT)
- error = zfs_secpolicy_destroy_perms(zc->zc_name, cr);
+ if (nvlist_lookup_nvlist(innvl, "snaps", &snaps) != 0)
+ return (SET_ERROR(EINVAL));
+ for (pair = nvlist_next_nvpair(snaps, NULL); pair != NULL;
+ pair = nextpair) {
+ nextpair = nvlist_next_nvpair(snaps, pair);
+ error = zfs_secpolicy_destroy_perms(nvpair_name(pair), cr);
+ if (error == ENOENT) {
+ /*
+ * Ignore any snapshots that don't exist (we consider
+ * them "already destroyed"). Remove the name from the
+ * nvl here in case the snapshot is created between
+ * now and when we try to destroy it (in which case
+ * we don't want to destroy it since we haven't
+ * checked for permission).
+ */
+ fnvlist_remove_nvpair(snaps, pair);
+ error = 0;
+ }
+ if (error != 0)
+ break;
+ }
- strfree(dsname);
return (error);
}
return (error);
}
+/* ARGSUSED */
static int
-zfs_secpolicy_rename(zfs_cmd_t *zc, cred_t *cr)
+zfs_secpolicy_rename(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr)
{
return (zfs_secpolicy_rename_perms(zc->zc_name, zc->zc_value, cr));
}
+/* ARGSUSED */
static int
-zfs_secpolicy_promote(zfs_cmd_t *zc, cred_t *cr)
+zfs_secpolicy_promote(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr)
{
- char parentname[MAXNAMELEN];
- objset_t *clone;
+ dsl_pool_t *dp;
+ dsl_dataset_t *clone;
int error;
error = zfs_secpolicy_write_perms(zc->zc_name,
ZFS_DELEG_PERM_PROMOTE, cr);
- if (error)
+ if (error != 0)
+ return (error);
+
+ error = dsl_pool_hold(zc->zc_name, FTAG, &dp);
+ if (error != 0)
return (error);
- error = dmu_objset_hold(zc->zc_name, FTAG, &clone);
+ error = dsl_dataset_hold(dp, zc->zc_name, FTAG, &clone);
if (error == 0) {
- dsl_dataset_t *pclone = NULL;
+ char parentname[MAXNAMELEN];
+ dsl_dataset_t *origin = NULL;
dsl_dir_t *dd;
- dd = clone->os_dsl_dataset->ds_dir;
+ dd = clone->ds_dir;
- rw_enter(&dd->dd_pool->dp_config_rwlock, RW_READER);
error = dsl_dataset_hold_obj(dd->dd_pool,
- dd->dd_phys->dd_origin_obj, FTAG, &pclone);
- rw_exit(&dd->dd_pool->dp_config_rwlock);
- if (error) {
- dmu_objset_rele(clone, FTAG);
+ dsl_dir_phys(dd)->dd_origin_obj, FTAG, &origin);
+ if (error != 0) {
+ dsl_dataset_rele(clone, FTAG);
+ dsl_pool_rele(dp, FTAG);
return (error);
}
- error = zfs_secpolicy_write_perms(zc->zc_name,
+ error = zfs_secpolicy_write_perms_ds(zc->zc_name, clone,
ZFS_DELEG_PERM_MOUNT, cr);
- dsl_dataset_name(pclone, parentname);
- dmu_objset_rele(clone, FTAG);
- dsl_dataset_rele(pclone, FTAG);
- if (error == 0)
- error = zfs_secpolicy_write_perms(parentname,
+ dsl_dataset_name(origin, parentname);
+ if (error == 0) {
+ error = zfs_secpolicy_write_perms_ds(parentname, origin,
ZFS_DELEG_PERM_PROMOTE, cr);
+ }
+ dsl_dataset_rele(clone, FTAG);
+ dsl_dataset_rele(origin, FTAG);
}
+ dsl_pool_rele(dp, FTAG);
return (error);
}
+/* ARGSUSED */
static int
-zfs_secpolicy_receive(zfs_cmd_t *zc, cred_t *cr)
+zfs_secpolicy_recv(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr)
{
int error;
ZFS_DELEG_PERM_SNAPSHOT, cr));
}
+/*
+ * Check for permission to create each snapshot in the nvlist.
+ */
+/* ARGSUSED */
+static int
+zfs_secpolicy_snapshot(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr)
+{
+ nvlist_t *snaps;
+ int error = 0;
+ nvpair_t *pair;
+
+ if (nvlist_lookup_nvlist(innvl, "snaps", &snaps) != 0)
+ return (SET_ERROR(EINVAL));
+ for (pair = nvlist_next_nvpair(snaps, NULL); pair != NULL;
+ pair = nvlist_next_nvpair(snaps, pair)) {
+ char *name = nvpair_name(pair);
+ char *atp = strchr(name, '@');
+
+ if (atp == NULL) {
+ error = SET_ERROR(EINVAL);
+ break;
+ }
+ *atp = '\0';
+ error = zfs_secpolicy_snapshot_perms(name, cr);
+ *atp = '@';
+ if (error != 0)
+ break;
+ }
+ return (error);
+}
+
+/*
+ * Check for permission to create each snapshot in the nvlist.
+ */
+/* ARGSUSED */
+static int
+zfs_secpolicy_bookmark(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr)
+{
+ int error = 0;
+ nvpair_t *pair;
+
+ for (pair = nvlist_next_nvpair(innvl, NULL);
+ pair != NULL; pair = nvlist_next_nvpair(innvl, pair)) {
+ char *name = nvpair_name(pair);
+ char *hashp = strchr(name, '#');
+
+ if (hashp == NULL) {
+ error = SET_ERROR(EINVAL);
+ break;
+ }
+ *hashp = '\0';
+ error = zfs_secpolicy_write_perms(name,
+ ZFS_DELEG_PERM_BOOKMARK, cr);
+ *hashp = '#';
+ if (error != 0)
+ break;
+ }
+ return (error);
+}
+
+/* ARGSUSED */
static int
-zfs_secpolicy_snapshot(zfs_cmd_t *zc, cred_t *cr)
+zfs_secpolicy_destroy_bookmarks(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr)
{
+ nvpair_t *pair, *nextpair;
+ int error = 0;
+
+ for (pair = nvlist_next_nvpair(innvl, NULL); pair != NULL;
+ pair = nextpair) {
+ char *name = nvpair_name(pair);
+ char *hashp = strchr(name, '#');
+ nextpair = nvlist_next_nvpair(innvl, pair);
+
+ if (hashp == NULL) {
+ error = SET_ERROR(EINVAL);
+ break;
+ }
+
+ *hashp = '\0';
+ error = zfs_secpolicy_write_perms(name,
+ ZFS_DELEG_PERM_DESTROY, cr);
+ *hashp = '#';
+ if (error == ENOENT) {
+ /*
+ * Ignore any filesystems that don't exist (we consider
+ * their bookmarks "already destroyed"). Remove
+ * the name from the nvl here in case the filesystem
+ * is created between now and when we try to destroy
+ * the bookmark (in which case we don't want to
+ * destroy it since we haven't checked for permission).
+ */
+ fnvlist_remove_nvpair(innvl, pair);
+ error = 0;
+ }
+ if (error != 0)
+ break;
+ }
+
+ return (error);
+}
- return (zfs_secpolicy_snapshot_perms(zc->zc_name, cr));
+/* ARGSUSED */
+static int
+zfs_secpolicy_log_history(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr)
+{
+ /*
+ * Even root must have a proper TSD so that we know what pool
+ * to log to.
+ */
+ if (tsd_get(zfs_allow_log_key) == NULL)
+ return (SET_ERROR(EPERM));
+ return (0);
}
static int
-zfs_secpolicy_create(zfs_cmd_t *zc, cred_t *cr)
+zfs_secpolicy_create_clone(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr)
{
char parentname[MAXNAMELEN];
int error;
+ char *origin;
if ((error = zfs_get_parent(zc->zc_name, parentname,
sizeof (parentname))) != 0)
return (error);
- if (zc->zc_value[0] != '\0') {
- if ((error = zfs_secpolicy_write_perms(zc->zc_value,
- ZFS_DELEG_PERM_CLONE, cr)) != 0)
- return (error);
- }
+ if (nvlist_lookup_string(innvl, "origin", &origin) == 0 &&
+ (error = zfs_secpolicy_write_perms(origin,
+ ZFS_DELEG_PERM_CLONE, cr)) != 0)
+ return (error);
if ((error = zfs_secpolicy_write_perms(parentname,
ZFS_DELEG_PERM_CREATE, cr)) != 0)
return (error);
- error = zfs_secpolicy_write_perms(parentname,
- ZFS_DELEG_PERM_MOUNT, cr);
-
- return (error);
+ return (zfs_secpolicy_write_perms(parentname,
+ ZFS_DELEG_PERM_MOUNT, cr));
}
/*
*/
/* ARGSUSED */
static int
-zfs_secpolicy_config(zfs_cmd_t *zc, cred_t *cr)
+zfs_secpolicy_config(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr)
{
if (secpolicy_sys_config(cr, B_FALSE) != 0)
- return (EPERM);
+ return (SET_ERROR(EPERM));
return (0);
}
*/
/* ARGSUSED */
static int
-zfs_secpolicy_diff(zfs_cmd_t *zc, cred_t *cr)
+zfs_secpolicy_diff(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr)
{
int error;
*/
/* ARGSUSED */
static int
-zfs_secpolicy_inject(zfs_cmd_t *zc, cred_t *cr)
+zfs_secpolicy_inject(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr)
{
return (secpolicy_zinject(cr));
}
+/* ARGSUSED */
static int
-zfs_secpolicy_inherit(zfs_cmd_t *zc, cred_t *cr)
+zfs_secpolicy_inherit_prop(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr)
{
zfs_prop_t prop = zfs_name_to_prop(zc->zc_value);
if (prop == ZPROP_INVAL) {
if (!zfs_prop_user(zc->zc_value))
- return (EINVAL);
+ return (SET_ERROR(EINVAL));
return (zfs_secpolicy_write_perms(zc->zc_name,
ZFS_DELEG_PERM_USERPROP, cr));
} else {
}
static int
-zfs_secpolicy_userspace_one(zfs_cmd_t *zc, cred_t *cr)
+zfs_secpolicy_userspace_one(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr)
{
- int err = zfs_secpolicy_read(zc, cr);
+ int err = zfs_secpolicy_read(zc, innvl, cr);
if (err)
return (err);
if (zc->zc_objset_type >= ZFS_NUM_USERQUOTA_PROPS)
- return (EINVAL);
+ return (SET_ERROR(EINVAL));
if (zc->zc_value[0] == 0) {
/*
}
static int
-zfs_secpolicy_userspace_many(zfs_cmd_t *zc, cred_t *cr)
+zfs_secpolicy_userspace_many(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr)
{
- int err = zfs_secpolicy_read(zc, cr);
+ int err = zfs_secpolicy_read(zc, innvl, cr);
if (err)
return (err);
if (zc->zc_objset_type >= ZFS_NUM_USERQUOTA_PROPS)
- return (EINVAL);
+ return (SET_ERROR(EINVAL));
return (zfs_secpolicy_write_perms(zc->zc_name,
userquota_perms[zc->zc_objset_type], cr));
}
+/* ARGSUSED */
static int
-zfs_secpolicy_userspace_upgrade(zfs_cmd_t *zc, cred_t *cr)
+zfs_secpolicy_userspace_upgrade(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr)
{
return (zfs_secpolicy_setprop(zc->zc_name, ZFS_PROP_VERSION,
NULL, cr));
}
+/* ARGSUSED */
static int
-zfs_secpolicy_hold(zfs_cmd_t *zc, cred_t *cr)
+zfs_secpolicy_hold(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr)
{
- return (zfs_secpolicy_write_perms(zc->zc_name,
- ZFS_DELEG_PERM_HOLD, cr));
+ nvpair_t *pair;
+ nvlist_t *holds;
+ int error;
+
+ error = nvlist_lookup_nvlist(innvl, "holds", &holds);
+ if (error != 0)
+ return (SET_ERROR(EINVAL));
+
+ for (pair = nvlist_next_nvpair(holds, NULL); pair != NULL;
+ pair = nvlist_next_nvpair(holds, pair)) {
+ char fsname[MAXNAMELEN];
+ error = dmu_fsname(nvpair_name(pair), fsname);
+ if (error != 0)
+ return (error);
+ error = zfs_secpolicy_write_perms(fsname,
+ ZFS_DELEG_PERM_HOLD, cr);
+ if (error != 0)
+ return (error);
+ }
+ return (0);
}
+/* ARGSUSED */
static int
-zfs_secpolicy_release(zfs_cmd_t *zc, cred_t *cr)
+zfs_secpolicy_release(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr)
{
- return (zfs_secpolicy_write_perms(zc->zc_name,
- ZFS_DELEG_PERM_RELEASE, cr));
+ nvpair_t *pair;
+ int error;
+
+ for (pair = nvlist_next_nvpair(innvl, NULL); pair != NULL;
+ pair = nvlist_next_nvpair(innvl, pair)) {
+ char fsname[MAXNAMELEN];
+ error = dmu_fsname(nvpair_name(pair), fsname);
+ if (error != 0)
+ return (error);
+ error = zfs_secpolicy_write_perms(fsname,
+ ZFS_DELEG_PERM_RELEASE, cr);
+ if (error != 0)
+ return (error);
+ }
+ return (0);
}
/*
* Policy for allowing temporary snapshots to be taken or released
*/
static int
-zfs_secpolicy_tmp_snapshot(zfs_cmd_t *zc, cred_t *cr)
+zfs_secpolicy_tmp_snapshot(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr)
{
/*
* A temporary snapshot is the same as a snapshot,
ZFS_DELEG_PERM_DIFF, cr)) == 0)
return (0);
- error = zfs_secpolicy_snapshot(zc, cr);
- if (!error)
- error = zfs_secpolicy_hold(zc, cr);
- if (!error)
- error = zfs_secpolicy_release(zc, cr);
- if (!error)
- error = zfs_secpolicy_destroy(zc, cr);
+ error = zfs_secpolicy_snapshot_perms(zc->zc_name, cr);
+ if (error == 0)
+ error = zfs_secpolicy_hold(zc, innvl, cr);
+ if (error == 0)
+ error = zfs_secpolicy_release(zc, innvl, cr);
+ if (error == 0)
+ error = zfs_secpolicy_destroy(zc, innvl, cr);
return (error);
}
* Read in and unpack the user-supplied nvlist.
*/
if (size == 0)
- return (EINVAL);
+ return (SET_ERROR(EINVAL));
- packed = kmem_alloc(size, KM_SLEEP | KM_NODEBUG);
+ packed = vmem_alloc(size, KM_SLEEP);
if ((error = ddi_copyin((void *)(uintptr_t)nvl, packed, size,
iflag)) != 0) {
- kmem_free(packed, size);
- return (error);
+ vmem_free(packed, size);
+ return (SET_ERROR(EFAULT));
}
if ((error = nvlist_unpack(packed, size, &list, 0)) != 0) {
- kmem_free(packed, size);
+ vmem_free(packed, size);
return (error);
}
- kmem_free(packed, size);
+ vmem_free(packed, size);
*nvp = list;
return (0);
}
+/*
+ * Reduce the size of this nvlist until it can be serialized in 'max' bytes.
+ * Entries will be removed from the end of the nvlist, and one int32 entry
+ * named "N_MORE_ERRORS" will be added indicating how many entries were
+ * removed.
+ */
static int
-fit_error_list(zfs_cmd_t *zc, nvlist_t **errors)
+nvlist_smush(nvlist_t *errors, size_t max)
{
size_t size;
- VERIFY(nvlist_size(*errors, &size, NV_ENCODE_NATIVE) == 0);
+ size = fnvlist_size(errors);
- if (size > zc->zc_nvlist_dst_size) {
+ if (size > max) {
nvpair_t *more_errors;
int n = 0;
- if (zc->zc_nvlist_dst_size < 1024)
- return (ENOMEM);
+ if (max < 1024)
+ return (SET_ERROR(ENOMEM));
- VERIFY(nvlist_add_int32(*errors, ZPROP_N_MORE_ERRORS, 0) == 0);
- more_errors = nvlist_prev_nvpair(*errors, NULL);
+ fnvlist_add_int32(errors, ZPROP_N_MORE_ERRORS, 0);
+ more_errors = nvlist_prev_nvpair(errors, NULL);
do {
- nvpair_t *pair = nvlist_prev_nvpair(*errors,
+ nvpair_t *pair = nvlist_prev_nvpair(errors,
more_errors);
- VERIFY(nvlist_remove_nvpair(*errors, pair) == 0);
+ fnvlist_remove_nvpair(errors, pair);
n++;
- VERIFY(nvlist_size(*errors, &size,
- NV_ENCODE_NATIVE) == 0);
- } while (size > zc->zc_nvlist_dst_size);
+ size = fnvlist_size(errors);
+ } while (size > max);
- VERIFY(nvlist_remove_nvpair(*errors, more_errors) == 0);
- VERIFY(nvlist_add_int32(*errors, ZPROP_N_MORE_ERRORS, n) == 0);
- ASSERT(nvlist_size(*errors, &size, NV_ENCODE_NATIVE) == 0);
- ASSERT(size <= zc->zc_nvlist_dst_size);
+ fnvlist_remove_nvpair(errors, more_errors);
+ fnvlist_add_int32(errors, ZPROP_N_MORE_ERRORS, n);
+ ASSERT3U(fnvlist_size(errors), <=, max);
}
return (0);
int error = 0;
size_t size;
- VERIFY(nvlist_size(nvl, &size, NV_ENCODE_NATIVE) == 0);
+ size = fnvlist_size(nvl);
if (size > zc->zc_nvlist_dst_size) {
- error = ENOMEM;
+ error = SET_ERROR(ENOMEM);
} else {
- packed = kmem_alloc(size, KM_SLEEP | KM_NODEBUG);
- VERIFY(nvlist_pack(nvl, &packed, &size, NV_ENCODE_NATIVE,
- KM_SLEEP) == 0);
+ packed = fnvlist_pack(nvl, &size);
if (ddi_copyout(packed, (void *)(uintptr_t)zc->zc_nvlist_dst,
size, zc->zc_iflags) != 0)
- error = EFAULT;
- kmem_free(packed, size);
+ error = SET_ERROR(EFAULT);
+ fnvlist_pack_free(packed, size);
}
zc->zc_nvlist_dst_size = size;
+ zc->zc_nvlist_dst_filled = B_TRUE;
return (error);
}
int error;
error = dmu_objset_hold(dsname, FTAG, &os);
- if (error)
+ if (error != 0)
return (error);
if (dmu_objset_type(os) != DMU_OST_ZFS) {
dmu_objset_rele(os, FTAG);
- return (EINVAL);
+ return (SET_ERROR(EINVAL));
}
mutex_enter(&os->os_user_ptr_lock);
if (*zsbp && (*zsbp)->z_sb) {
atomic_inc(&((*zsbp)->z_sb->s_active));
} else {
- error = ESRCH;
+ error = SET_ERROR(ESRCH);
}
mutex_exit(&os->os_user_ptr_lock);
dmu_objset_rele(os, FTAG);
/*
* Find a zfs_sb_t for a mounted filesystem, or create our own, in which
* case its z_sb will be NULL, and it will be opened as the owner.
+ * If 'writer' is set, the z_teardown_lock will be held for RW_WRITER,
+ * which prevents all inode ops from running.
*/
static int
zfs_sb_hold(const char *name, void *tag, zfs_sb_t **zsbp, boolean_t writer)
int error = 0;
if (get_zfs_sb(name, zsbp) != 0)
- error = zfs_sb_create(name, zsbp);
+ error = zfs_sb_create(name, NULL, zsbp);
if (error == 0) {
- rrw_enter(&(*zsbp)->z_teardown_lock, (writer) ? RW_WRITER :
+ rrm_enter(&(*zsbp)->z_teardown_lock, (writer) ? RW_WRITER :
RW_READER, tag);
if ((*zsbp)->z_unmounted) {
/*
* XXX we could probably try again, since the unmounting
* thread should be just about to disassociate the
- * objset from the zfsvfs.
+ * objset from the zsb.
*/
- rrw_exit(&(*zsbp)->z_teardown_lock, tag);
- return (EBUSY);
+ rrm_exit(&(*zsbp)->z_teardown_lock, tag);
+ return (SET_ERROR(EBUSY));
}
}
return (error);
static void
zfs_sb_rele(zfs_sb_t *zsb, void *tag)
{
- rrw_exit(&zsb->z_teardown_lock, tag);
+ rrm_exit(&zsb->z_teardown_lock, tag);
if (zsb->z_sb) {
deactivate_super(zsb->z_sb);
nvlist_t *config, *props = NULL;
nvlist_t *rootprops = NULL;
nvlist_t *zplprops = NULL;
- char *buf;
if ((error = get_nvlist(zc->zc_nvlist_conf, zc->zc_nvlist_conf_size,
zc->zc_iflags, &config)))
(void) nvlist_lookup_uint64(props,
zpool_prop_to_name(ZPOOL_PROP_VERSION), &version);
- if (version < SPA_VERSION_INITIAL || version > SPA_VERSION) {
- error = EINVAL;
+ if (!SPA_VERSION_IS_SUPPORTED(version)) {
+ error = SET_ERROR(EINVAL);
goto pool_props_bad;
}
(void) nvlist_lookup_nvlist(props, ZPOOL_ROOTFS_PROPS, &nvl);
VERIFY(nvlist_alloc(&zplprops, NV_UNIQUE_NAME, KM_SLEEP) == 0);
error = zfs_fill_zplprops_root(version, rootprops,
zplprops, NULL);
- if (error)
+ if (error != 0)
goto pool_props_bad;
}
- buf = history_str_get(zc);
-
- error = spa_create(zc->zc_name, config, props, buf, zplprops);
+ error = spa_create(zc->zc_name, config, props, zplprops);
/*
* Set the remaining root properties
ZPROP_SRC_LOCAL, rootprops, NULL)) != 0)
(void) spa_destroy(zc->zc_name);
- if (buf != NULL)
- history_str_free(buf);
-
pool_props_bad:
nvlist_free(rootprops);
nvlist_free(zplprops);
int error;
zfs_log_history(zc);
error = spa_destroy(zc->zc_name);
- if (error == 0)
- zvol_remove_minors(zc->zc_name);
+
return (error);
}
if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID, &guid) != 0 ||
guid != zc->zc_guid)
- error = EINVAL;
+ error = SET_ERROR(EINVAL);
else
error = spa_import(zc->zc_name, config, props, zc->zc_cookie);
error = err;
}
- if (error == 0)
- zvol_create_minors(zc->zc_name);
-
nvlist_free(config);
-
- if (props)
- nvlist_free(props);
+ nvlist_free(props);
return (error);
}
zfs_log_history(zc);
error = spa_export(zc->zc_name, NULL, force, hardforce);
- if (error == 0)
- zvol_remove_minors(zc->zc_name);
+
return (error);
}
int error;
if ((configs = spa_all_configs(&zc->zc_cookie)) == NULL)
- return (EEXIST);
+ return (SET_ERROR(EEXIST));
error = put_nvlist(zc, configs);
return (error);
}
+/*
+ * inputs:
+ * zc_name name of the pool
+ *
+ * outputs:
+ * zc_cookie real errno
+ * zc_nvlist_dst config nvlist
+ * zc_nvlist_dst_size size of config nvlist
+ */
static int
zfs_ioc_pool_stats(zfs_cmd_t *zc)
{
nvlist_free(tryconfig);
if (config == NULL)
- return (EINVAL);
+ return (SET_ERROR(EINVAL));
error = put_nvlist(zc, config);
nvlist_free(config);
if ((error = spa_open(zc->zc_name, &spa, FTAG)) != 0)
return (error);
- if (zc->zc_cookie < spa_version(spa) || zc->zc_cookie > SPA_VERSION) {
+ if (zc->zc_cookie < spa_version(spa) ||
+ !SPA_VERSION_IS_SUPPORTED(zc->zc_cookie)) {
spa_close(spa, FTAG);
- return (EINVAL);
+ return (SET_ERROR(EINVAL));
}
spa_upgrade(spa, zc->zc_cookie);
int error;
if ((size = zc->zc_history_len) == 0)
- return (EINVAL);
+ return (SET_ERROR(EINVAL));
if ((error = spa_open(zc->zc_name, &spa, FTAG)) != 0)
return (error);
if (spa_version(spa) < SPA_VERSION_ZPOOL_HISTORY) {
spa_close(spa, FTAG);
- return (ENOTSUP);
+ return (SET_ERROR(ENOTSUP));
}
hist_buf = vmem_alloc(size, KM_SLEEP);
}
static int
-zfs_ioc_dsobj_to_dsname(zfs_cmd_t *zc)
+zfs_ioc_pool_reguid(zfs_cmd_t *zc)
{
+ spa_t *spa;
int error;
- if ((error = dsl_dsobj_to_dsname(zc->zc_name,zc->zc_obj,zc->zc_value)))
- return (error);
+ error = spa_open(zc->zc_name, &spa, FTAG);
+ if (error == 0) {
+ error = spa_change_guid(spa);
+ spa_close(spa, FTAG);
+ }
+ return (error);
+}
- return (0);
+static int
+zfs_ioc_dsobj_to_dsname(zfs_cmd_t *zc)
+{
+ return (dsl_dsobj_to_dsname(zc->zc_name, zc->zc_obj, zc->zc_value));
}
/*
return (error);
if (dmu_objset_type(os) != DMU_OST_ZFS) {
dmu_objset_rele(os, FTAG);
- return (EINVAL);
+ return (SET_ERROR(EINVAL));
}
error = zfs_obj_to_path(os, zc->zc_obj, zc->zc_value,
sizeof (zc->zc_value));
return (error);
if (dmu_objset_type(os) != DMU_OST_ZFS) {
dmu_objset_rele(os, FTAG);
- return (EINVAL);
+ return (SET_ERROR(EINVAL));
}
error = zfs_obj_to_stats(os, zc->zc_obj, &zc->zc_stat, zc->zc_value,
sizeof (zc->zc_value));
{
spa_t *spa;
int error;
- nvlist_t *config, **l2cache, **spares;
- uint_t nl2cache = 0, nspares = 0;
+ nvlist_t *config;
error = spa_open(zc->zc_name, &spa, FTAG);
if (error != 0)
error = get_nvlist(zc->zc_nvlist_conf, zc->zc_nvlist_conf_size,
zc->zc_iflags, &config);
- (void) nvlist_lookup_nvlist_array(config, ZPOOL_CONFIG_L2CACHE,
- &l2cache, &nl2cache);
-
- (void) nvlist_lookup_nvlist_array(config, ZPOOL_CONFIG_SPARES,
- &spares, &nspares);
-
- /*
- * A root pool with concatenated devices is not supported.
- * Thus, can not add a device to a root pool.
- *
- * Intent log device can not be added to a rootpool because
- * during mountroot, zil is replayed, a seperated log device
- * can not be accessed during the mountroot time.
- *
- * l2cache and spare devices are ok to be added to a rootpool.
- */
- if (spa_bootfs(spa) != 0 && nl2cache == 0 && nspares == 0) {
- nvlist_free(config);
- spa_close(spa, FTAG);
- return (EDOM);
- }
-
if (error == 0) {
error = spa_vdev_add(spa, config);
nvlist_free(config);
break;
default:
- error = EINVAL;
+ error = SET_ERROR(EINVAL);
}
zc->zc_cookie = newstate;
spa_close(spa, FTAG);
* inconsistent. So this is a bit of a workaround...
* XXX reading with out owning
*/
- if (!zc->zc_objset_stats.dds_inconsistent) {
- if (dmu_objset_type(os) == DMU_OST_ZVOL)
- error = zvol_get_stats(os, nv);
+ if (!zc->zc_objset_stats.dds_inconsistent &&
+ dmu_objset_type(os) == DMU_OST_ZVOL) {
+ error = zvol_get_stats(os, nv);
+ if (error == EIO)
+ return (error);
+ VERIFY0(error);
}
if (error == 0)
error = put_nvlist(zc, nv);
static int
zfs_ioc_objset_stats(zfs_cmd_t *zc)
{
- objset_t *os = NULL;
+ objset_t *os;
int error;
- if ((error = dmu_objset_hold(zc->zc_name, FTAG, &os)))
- return (error);
-
- error = zfs_ioc_objset_stats_impl(zc, os);
-
- dmu_objset_rele(os, FTAG);
+ error = dmu_objset_hold(zc->zc_name, FTAG, &os);
+ if (error == 0) {
+ error = zfs_ioc_objset_stats_impl(zc, os);
+ dmu_objset_rele(os, FTAG);
+ }
return (error);
}
static int
zfs_ioc_objset_recvd_props(zfs_cmd_t *zc)
{
- objset_t *os = NULL;
- int error;
+ int error = 0;
nvlist_t *nv;
- if ((error = dmu_objset_hold(zc->zc_name, FTAG, &os)))
- return (error);
-
/*
* Without this check, we would return local property values if the
* caller has not already received properties on or after
* SPA_VERSION_RECVD_PROPS.
*/
- if (!dsl_prop_get_hasrecvd(os)) {
- dmu_objset_rele(os, FTAG);
- return (ENOTSUP);
- }
+ if (!dsl_prop_get_hasrecvd(zc->zc_name))
+ return (SET_ERROR(ENOTSUP));
if (zc->zc_nvlist_dst != 0 &&
- (error = dsl_prop_get_received(os, &nv)) == 0) {
+ (error = dsl_prop_get_received(zc->zc_name, &nv)) == 0) {
error = put_nvlist(zc, nv);
nvlist_free(nv);
}
- dmu_objset_rele(os, FTAG);
return (error);
}
err = put_nvlist(zc, nv);
nvlist_free(nv);
} else {
- err = ENOENT;
+ err = SET_ERROR(ENOENT);
}
dmu_objset_rele(os, FTAG);
return (err);
}
-static boolean_t
+boolean_t
dataset_name_hidden(const char *name)
{
/*
top:
if ((error = dmu_objset_hold(zc->zc_name, FTAG, &os))) {
if (error == ENOENT)
- error = ESRCH;
+ error = SET_ERROR(ESRCH);
return (error);
}
(void) strlcat(zc->zc_name, "/", sizeof (zc->zc_name));
p = zc->zc_name + strlen(zc->zc_name);
- /*
- * Pre-fetch the datasets. dmu_objset_prefetch() always returns 0
- * but is not declared void because its called by dmu_objset_find().
- */
- if (zc->zc_cookie == 0) {
- uint64_t cookie = 0;
- int len = sizeof (zc->zc_name) - (p - zc->zc_name);
-
- while (dmu_dir_list_next(os, len, p, NULL, &cookie) == 0)
- (void) dmu_objset_prefetch(p, NULL);
- }
-
do {
error = dmu_dir_list_next(os,
sizeof (zc->zc_name) - (p - zc->zc_name), p,
NULL, &zc->zc_cookie);
if (error == ENOENT)
- error = ESRCH;
- } while (error == 0 && dataset_name_hidden(zc->zc_name) &&
- !(zc->zc_iflags & FKIOCTL));
+ error = SET_ERROR(ESRCH);
+ } while (error == 0 && dataset_name_hidden(zc->zc_name));
dmu_objset_rele(os, FTAG);
/*
objset_t *os;
int error;
-top:
- if (zc->zc_cookie == 0)
- (void) dmu_objset_find(zc->zc_name, dmu_objset_prefetch,
- NULL, DS_FIND_SNAPSHOTS);
-
error = dmu_objset_hold(zc->zc_name, FTAG, &os);
- if (error)
+ if (error != 0) {
return (error == ENOENT ? ESRCH : error);
+ }
/*
* A dataset name of maximum length cannot have any snapshots,
*/
if (strlcat(zc->zc_name, "@", sizeof (zc->zc_name)) >= MAXNAMELEN) {
dmu_objset_rele(os, FTAG);
- return (ESRCH);
+ return (SET_ERROR(ESRCH));
}
error = dmu_snapshot_list_next(os,
zc->zc_name + strlen(zc->zc_name), &zc->zc_obj, &zc->zc_cookie,
NULL);
- if (error == 0) {
+ if (error == 0 && !zc->zc_simple) {
dsl_dataset_t *ds;
dsl_pool_t *dp = os->os_dsl_dataset->ds_dir->dd_pool;
- /*
- * Since we probably don't have a hold on this snapshot,
- * it's possible that the objsetid could have been destroyed
- * and reused for a new objset. It's OK if this happens during
- * a zfs send operation, since the new createtxg will be
- * beyond the range we're interested in.
- */
- rw_enter(&dp->dp_config_rwlock, RW_READER);
error = dsl_dataset_hold_obj(dp, zc->zc_obj, FTAG, &ds);
- rw_exit(&dp->dp_config_rwlock);
- if (error) {
- if (error == ENOENT) {
- /* Racing with destroy, get the next one. */
- *strchr(zc->zc_name, '@') = '\0';
- dmu_objset_rele(os, FTAG);
- goto top;
- }
- } else {
+ if (error == 0) {
objset_t *ossnap;
error = dmu_objset_from_ds(ds, &ossnap);
dsl_dataset_rele(ds, FTAG);
}
} else if (error == ENOENT) {
- error = ESRCH;
+ error = SET_ERROR(ESRCH);
}
dmu_objset_rele(os, FTAG);
/* if we failed, undo the @ that we tacked on to zc_name */
- if (error)
+ if (error != 0)
*strchr(zc->zc_name, '@') = '\0';
return (error);
}
VERIFY(nvpair_value_nvlist(pair, &attrs) == 0);
if (nvlist_lookup_nvpair(attrs, ZPROP_VALUE,
&pair) != 0)
- return (EINVAL);
+ return (SET_ERROR(EINVAL));
}
/*
if ((dash = strchr(propname, '-')) == NULL ||
nvpair_value_uint64_array(pair, &valary, &vallen) != 0 ||
vallen != 3)
- return (EINVAL);
+ return (SET_ERROR(EINVAL));
domain = dash + 1;
type = valary[0];
const char *propname = nvpair_name(pair);
zfs_prop_t prop = zfs_name_to_prop(propname);
uint64_t intval;
- int err;
+ int err = -1;
if (prop == ZPROP_INVAL) {
if (zfs_prop_userquota(propname))
err = dsl_dir_set_quota(dsname, source, intval);
break;
case ZFS_PROP_REFQUOTA:
- err = dsl_dataset_set_quota(dsname, source, intval);
+ err = dsl_dataset_set_refquota(dsname, source, intval);
+ break;
+ case ZFS_PROP_FILESYSTEM_LIMIT:
+ case ZFS_PROP_SNAPSHOT_LIMIT:
+ if (intval == UINT64_MAX) {
+ /* clearing the limit, just do it */
+ err = 0;
+ } else {
+ err = dsl_dir_activate_fs_ss_limit(dsname);
+ }
+ /*
+ * Set err to -1 to force the zfs_set_prop_nvlist code down the
+ * default path to set the value in the nvlist.
+ */
+ if (err == 0)
+ err = -1;
break;
case ZFS_PROP_RESERVATION:
err = dsl_dir_set_reservation(dsname, source, intval);
break;
case ZFS_PROP_REFRESERVATION:
- err = dsl_dataset_set_reservation(dsname, source, intval);
+ err = dsl_dataset_set_refreservation(dsname, source, intval);
break;
case ZFS_PROP_VOLSIZE:
err = zvol_set_volsize(dsname, intval);
break;
+ case ZFS_PROP_SNAPDEV:
+ err = zvol_set_snapdev(dsname, source, intval);
+ break;
case ZFS_PROP_VERSION:
{
zfs_sb_t *zsb;
if (err == 0 && intval >= ZPL_VERSION_USERSPACE) {
zfs_cmd_t *zc;
- zc = kmem_zalloc(sizeof (zfs_cmd_t),
- KM_SLEEP | KM_NODEBUG);
+ zc = kmem_zalloc(sizeof (zfs_cmd_t), KM_SLEEP);
(void) strcpy(zc->zc_name, dsname);
(void) zfs_ioc_userspace_upgrade(zc);
kmem_free(zc, sizeof (zfs_cmd_t));
}
break;
}
-
default:
err = -1;
}
/*
* This function is best effort. If it fails to set any of the given properties,
- * it continues to set as many as it can and returns the first error
- * encountered. If the caller provides a non-NULL errlist, it also gives the
- * complete list of names of all the properties it failed to set along with the
- * corresponding error numbers. The caller is responsible for freeing the
- * returned errlist.
+ * it continues to set as many as it can and returns the last error
+ * encountered. If the caller provides a non-NULL errlist, it will be filled in
+ * with the list of names of all the properties that failed along with the
+ * corresponding error numbers.
*
- * If every property is set successfully, zero is returned and the list pointed
- * at by errlist is NULL.
+ * If every property is set successfully, zero is returned and errlist is not
+ * modified.
*/
int
zfs_set_prop_nvlist(const char *dsname, zprop_source_t source, nvlist_t *nvl,
- nvlist_t **errlist)
+ nvlist_t *errlist)
{
nvpair_t *pair;
nvpair_t *propval;
int rv = 0;
uint64_t intval;
char *strval;
- nvlist_t *genericnvl;
- nvlist_t *errors;
- nvlist_t *retrynvl;
-
- VERIFY(nvlist_alloc(&genericnvl, NV_UNIQUE_NAME, KM_SLEEP) == 0);
- VERIFY(nvlist_alloc(&errors, NV_UNIQUE_NAME, KM_SLEEP) == 0);
- VERIFY(nvlist_alloc(&retrynvl, NV_UNIQUE_NAME, KM_SLEEP) == 0);
+ nvlist_t *genericnvl = fnvlist_alloc();
+ nvlist_t *retrynvl = fnvlist_alloc();
retry:
pair = NULL;
while ((pair = nvlist_next_nvpair(nvl, pair)) != NULL) {
propval = pair;
if (nvpair_type(pair) == DATA_TYPE_NVLIST) {
nvlist_t *attrs;
- VERIFY(nvpair_value_nvlist(pair, &attrs) == 0);
+ attrs = fnvpair_value_nvlist(pair);
if (nvlist_lookup_nvpair(attrs, ZPROP_VALUE,
&propval) != 0)
- err = EINVAL;
+ err = SET_ERROR(EINVAL);
}
/* Validate value type */
if (err == 0 && prop == ZPROP_INVAL) {
if (zfs_prop_user(propname)) {
if (nvpair_type(propval) != DATA_TYPE_STRING)
- err = EINVAL;
+ err = SET_ERROR(EINVAL);
} else if (zfs_prop_userquota(propname)) {
if (nvpair_type(propval) !=
DATA_TYPE_UINT64_ARRAY)
- err = EINVAL;
+ err = SET_ERROR(EINVAL);
+ } else {
+ err = SET_ERROR(EINVAL);
}
} else if (err == 0) {
if (nvpair_type(propval) == DATA_TYPE_STRING) {
if (zfs_prop_get_type(prop) != PROP_TYPE_STRING)
- err = EINVAL;
+ err = SET_ERROR(EINVAL);
} else if (nvpair_type(propval) == DATA_TYPE_UINT64) {
const char *unused;
- VERIFY(nvpair_value_uint64(propval,
- &intval) == 0);
+ intval = fnvpair_value_uint64(propval);
switch (zfs_prop_get_type(prop)) {
case PROP_TYPE_NUMBER:
break;
case PROP_TYPE_STRING:
- err = EINVAL;
+ err = SET_ERROR(EINVAL);
break;
case PROP_TYPE_INDEX:
if (zfs_prop_index_to_string(prop,
intval, &unused) != 0)
- err = EINVAL;
+ err = SET_ERROR(EINVAL);
break;
default:
cmn_err(CE_PANIC,
"unknown property type");
}
} else {
- err = EINVAL;
+ err = SET_ERROR(EINVAL);
}
}
}
}
- if (err != 0)
- VERIFY(nvlist_add_int32(errors, propname, err) == 0);
+ if (err != 0) {
+ if (errlist != NULL)
+ fnvlist_add_int32(errlist, propname, err);
+ rv = err;
+ }
}
if (nvl != retrynvl && !nvlist_empty(retrynvl)) {
propval = pair;
if (nvpair_type(pair) == DATA_TYPE_NVLIST) {
nvlist_t *attrs;
- VERIFY(nvpair_value_nvlist(pair, &attrs) == 0);
- VERIFY(nvlist_lookup_nvpair(attrs, ZPROP_VALUE,
- &propval) == 0);
+ attrs = fnvpair_value_nvlist(pair);
+ propval = fnvlist_lookup_nvpair(attrs,
+ ZPROP_VALUE);
}
if (nvpair_type(propval) == DATA_TYPE_STRING) {
- VERIFY(nvpair_value_string(propval,
- &strval) == 0);
- err = dsl_prop_set(dsname, propname, source, 1,
- strlen(strval) + 1, strval);
+ strval = fnvpair_value_string(propval);
+ err = dsl_prop_set_string(dsname, propname,
+ source, strval);
} else {
- VERIFY(nvpair_value_uint64(propval,
- &intval) == 0);
- err = dsl_prop_set(dsname, propname, source, 8,
- 1, &intval);
+ intval = fnvpair_value_uint64(propval);
+ err = dsl_prop_set_int(dsname, propname, source,
+ intval);
}
if (err != 0) {
- VERIFY(nvlist_add_int32(errors, propname,
- err) == 0);
+ if (errlist != NULL) {
+ fnvlist_add_int32(errlist, propname,
+ err);
+ }
+ rv = err;
}
}
}
nvlist_free(genericnvl);
nvlist_free(retrynvl);
- if ((pair = nvlist_next_nvpair(errors, NULL)) == NULL) {
- nvlist_free(errors);
- errors = NULL;
- } else {
- VERIFY(nvpair_value_int32(pair, &rv) == 0);
- }
-
- if (errlist == NULL)
- nvlist_free(errors);
- else
- *errlist = errors;
-
return (rv);
}
* Check that all the properties are valid user properties.
*/
static int
-zfs_check_userprops(char *fsname, nvlist_t *nvl)
+zfs_check_userprops(const char *fsname, nvlist_t *nvl)
{
nvpair_t *pair = NULL;
int error = 0;
while ((pair = nvlist_next_nvpair(nvl, pair)) != NULL) {
const char *propname = nvpair_name(pair);
- char *valstr;
if (!zfs_prop_user(propname) ||
nvpair_type(pair) != DATA_TYPE_STRING)
- return (EINVAL);
+ return (SET_ERROR(EINVAL));
if ((error = zfs_secpolicy_write_perms(fsname,
ZFS_DELEG_PERM_USERPROP, CRED())))
return (error);
if (strlen(propname) >= ZAP_MAXNAMELEN)
- return (ENAMETOOLONG);
+ return (SET_ERROR(ENAMETOOLONG));
- VERIFY(nvpair_value_string(pair, &valstr) == 0);
- if (strlen(valstr) >= ZAP_MAXVALUELEN)
- return (E2BIG);
+ if (strlen(fnvpair_value_string(pair)) >= ZAP_MAXVALUELEN)
+ return (SET_ERROR(E2BIG));
}
return (0);
}
}
static int
-clear_received_props(objset_t *os, const char *fs, nvlist_t *props,
+clear_received_props(const char *dsname, nvlist_t *props,
nvlist_t *skipped)
{
int err = 0;
* properties at least once on or after SPA_VERSION_RECVD_PROPS.
*/
zprop_source_t flags = (ZPROP_SRC_NONE |
- (dsl_prop_get_hasrecvd(os) ? ZPROP_SRC_RECEIVED : 0));
- err = zfs_set_prop_nvlist(fs, flags, cleared_props, NULL);
+ (dsl_prop_get_hasrecvd(dsname) ? ZPROP_SRC_RECEIVED : 0));
+ err = zfs_set_prop_nvlist(dsname, flags, cleared_props, NULL);
}
nvlist_free(cleared_props);
return (err);
boolean_t received = zc->zc_cookie;
zprop_source_t source = (received ? ZPROP_SRC_RECEIVED :
ZPROP_SRC_LOCAL);
- nvlist_t *errors = NULL;
+ nvlist_t *errors;
int error;
if ((error = get_nvlist(zc->zc_nvlist_src, zc->zc_nvlist_src_size,
if (received) {
nvlist_t *origprops;
- objset_t *os;
- if (dmu_objset_hold(zc->zc_name, FTAG, &os) == 0) {
- if (dsl_prop_get_received(os, &origprops) == 0) {
- (void) clear_received_props(os,
- zc->zc_name, origprops, nvl);
- nvlist_free(origprops);
- }
-
- dsl_prop_set_hasrecvd(os);
- dmu_objset_rele(os, FTAG);
+ if (dsl_prop_get_received(zc->zc_name, &origprops) == 0) {
+ (void) clear_received_props(zc->zc_name,
+ origprops, nvl);
+ nvlist_free(origprops);
}
+
+ error = dsl_prop_set_hasrecvd(zc->zc_name);
}
- error = zfs_set_prop_nvlist(zc->zc_name, source, nvl, &errors);
+ errors = fnvlist_alloc();
+ if (error == 0)
+ error = zfs_set_prop_nvlist(zc->zc_name, source, nvl, errors);
if (zc->zc_nvlist_dst != 0 && errors != NULL) {
(void) put_nvlist(zc, errors);
*/
if (prop == ZPROP_INVAL) {
if (!zfs_prop_user(propname))
- return (EINVAL);
+ return (SET_ERROR(EINVAL));
type = PROP_TYPE_STRING;
} else if (prop == ZFS_PROP_VOLSIZE ||
prop == ZFS_PROP_VERSION) {
- return (EINVAL);
+ return (SET_ERROR(EINVAL));
} else {
type = zfs_prop_get_type(prop);
}
break;
default:
nvlist_free(dummy);
- return (EINVAL);
+ return (SET_ERROR(EINVAL));
}
pair = nvlist_next_nvpair(dummy, NULL);
* they are not considered inheritable.
*/
if (prop != ZPROP_INVAL && !zfs_prop_inheritable(prop))
- return (EINVAL);
+ return (SET_ERROR(EINVAL));
}
- /* the property name has been validated by zfs_secpolicy_inherit() */
- return (dsl_prop_set(zc->zc_name, zc->zc_value, source, 0, 0, NULL));
+ /* property name has been validated by zfs_secpolicy_inherit_prop() */
+ return (dsl_prop_inherit(zc->zc_name, zc->zc_value, source));
}
static int
if (error == 0 && zc->zc_nvlist_dst != 0)
error = put_nvlist(zc, nvp);
else
- error = EFAULT;
+ error = SET_ERROR(EFAULT);
nvlist_free(nvp);
return (error);
}
-/*
- * inputs:
- * zc_name name of volume
- *
- * outputs: none
- */
-static int
-zfs_ioc_create_minor(zfs_cmd_t *zc)
-{
- return (zvol_create_minor(zc->zc_name));
-}
-
-/*
- * inputs:
- * zc_name name of volume
- *
- * outputs: none
- */
-static int
-zfs_ioc_remove_minor(zfs_cmd_t *zc)
-{
- return (zvol_remove_minor(zc->zc_name));
-}
-
/*
* inputs:
* zc_name name of filesystem
*/
if ((error = zfs_deleg_verify_nvlist(fsaclnv)) != 0) {
nvlist_free(fsaclnv);
- return (EINVAL);
+ return (SET_ERROR(EINVAL));
}
/*
*/
error = secpolicy_zfs(CRED());
- if (error) {
+ if (error != 0) {
if (zc->zc_perm_action == B_FALSE) {
error = dsl_deleg_can_allow(zc->zc_name,
fsaclnv, CRED());
return (error);
}
-#ifdef HAVE_SNAPSHOT
-/*
- * Search the vfs list for a specified resource. Returns a pointer to it
- * or NULL if no suitable entry is found. The caller of this routine
- * is responsible for releasing the returned vfs pointer.
- */
-static vfs_t *
-zfs_get_vfs(const char *resource)
-{
- struct vfs *vfsp;
- struct vfs *vfs_found = NULL;
-
- vfs_list_read_lock();
- vfsp = rootvfs;
- do {
- if (strcmp(refstr_value(vfsp->vfs_resource), resource) == 0) {
- mntget(vfsp);
- vfs_found = vfsp;
- break;
- }
- vfsp = vfsp->vfs_next;
- } while (vfsp != rootvfs);
- vfs_list_unlock();
- return (vfs_found);
-}
-#endif /* HAVE_SNAPSHOT */
-
/* ARGSUSED */
static void
zfs_create_cb(objset_t *os, void *arg, cred_t *cr, dmu_tx_t *tx)
/*
* inputs:
- * createprops list of properties requested by creator
- * default_zplver zpl version to use if unspecified in createprops
- * fuids_ok fuids allowed in this version of the spa?
* os parent objset pointer (NULL if root fs)
+ * fuids_ok fuids allowed in this version of the spa?
+ * sa_ok SAs allowed in this version of the spa?
+ * createprops list of properties requested by creator
*
* outputs:
* zplprops values for the zplprops we attach to the master node object
uint64_t sense = ZFS_PROP_UNDEFINED;
uint64_t norm = ZFS_PROP_UNDEFINED;
uint64_t u8 = ZFS_PROP_UNDEFINED;
+ int error;
ASSERT(zplprops != NULL);
(zplver < ZPL_VERSION_NORMALIZATION &&
(norm != ZFS_PROP_UNDEFINED || u8 != ZFS_PROP_UNDEFINED ||
sense != ZFS_PROP_UNDEFINED)))
- return (ENOTSUP);
+ return (SET_ERROR(ENOTSUP));
/*
* Put the version in the zplprops
VERIFY(nvlist_add_uint64(zplprops,
zfs_prop_to_name(ZFS_PROP_VERSION), zplver) == 0);
- if (norm == ZFS_PROP_UNDEFINED)
- VERIFY(zfs_get_zplprop(os, ZFS_PROP_NORMALIZE, &norm) == 0);
+ if (norm == ZFS_PROP_UNDEFINED &&
+ (error = zfs_get_zplprop(os, ZFS_PROP_NORMALIZE, &norm)) != 0)
+ return (error);
VERIFY(nvlist_add_uint64(zplprops,
zfs_prop_to_name(ZFS_PROP_NORMALIZE), norm) == 0);
*/
if (norm)
u8 = 1;
- if (u8 == ZFS_PROP_UNDEFINED)
- VERIFY(zfs_get_zplprop(os, ZFS_PROP_UTF8ONLY, &u8) == 0);
+ if (u8 == ZFS_PROP_UNDEFINED &&
+ (error = zfs_get_zplprop(os, ZFS_PROP_UTF8ONLY, &u8)) != 0)
+ return (error);
VERIFY(nvlist_add_uint64(zplprops,
zfs_prop_to_name(ZFS_PROP_UTF8ONLY), u8) == 0);
- if (sense == ZFS_PROP_UNDEFINED)
- VERIFY(zfs_get_zplprop(os, ZFS_PROP_CASE, &sense) == 0);
+ if (sense == ZFS_PROP_UNDEFINED &&
+ (error = zfs_get_zplprop(os, ZFS_PROP_CASE, &sense)) != 0)
+ return (error);
VERIFY(nvlist_add_uint64(zplprops,
zfs_prop_to_name(ZFS_PROP_CASE), sense) == 0);
}
/*
- * inputs:
- * zc_objset_type type of objset to create (fs vs zvol)
- * zc_name name of new objset
- * zc_value name of snapshot to clone from (may be empty)
- * zc_nvlist_src{_size} nvlist of properties to apply
+ * innvl: {
+ * "type" -> dmu_objset_type_t (int32)
+ * (optional) "props" -> { prop -> value }
+ * }
*
- * outputs: none
+ * outnvl: propname -> error code (int32)
*/
static int
-zfs_ioc_create(zfs_cmd_t *zc)
+zfs_ioc_create(const char *fsname, nvlist_t *innvl, nvlist_t *outnvl)
{
- objset_t *clone;
int error = 0;
- zfs_creat_t zct;
+ zfs_creat_t zct = { 0 };
nvlist_t *nvprops = NULL;
void (*cbfunc)(objset_t *os, void *arg, cred_t *cr, dmu_tx_t *tx);
- dmu_objset_type_t type = zc->zc_objset_type;
+ int32_t type32;
+ dmu_objset_type_t type;
+ boolean_t is_insensitive = B_FALSE;
- switch (type) {
+ if (nvlist_lookup_int32(innvl, "type", &type32) != 0)
+ return (SET_ERROR(EINVAL));
+ type = type32;
+ (void) nvlist_lookup_nvlist(innvl, "props", &nvprops);
+ switch (type) {
case DMU_OST_ZFS:
cbfunc = zfs_create_cb;
break;
cbfunc = NULL;
break;
}
- if (strchr(zc->zc_name, '@') ||
- strchr(zc->zc_name, '%'))
- return (EINVAL);
+ if (strchr(fsname, '@') ||
+ strchr(fsname, '%'))
+ return (SET_ERROR(EINVAL));
- if (zc->zc_nvlist_src != 0 &&
- (error = get_nvlist(zc->zc_nvlist_src, zc->zc_nvlist_src_size,
- zc->zc_iflags, &nvprops)) != 0)
- return (error);
-
- zct.zct_zplprops = NULL;
zct.zct_props = nvprops;
- if (zc->zc_value[0] != '\0') {
- /*
- * We're creating a clone of an existing snapshot.
- */
- zc->zc_value[sizeof (zc->zc_value) - 1] = '\0';
- if (dataset_namecheck(zc->zc_value, NULL, NULL) != 0) {
- nvlist_free(nvprops);
- return (EINVAL);
- }
-
- error = dmu_objset_hold(zc->zc_value, FTAG, &clone);
- if (error) {
- nvlist_free(nvprops);
- return (error);
- }
-
- error = dmu_objset_clone(zc->zc_name, dmu_objset_ds(clone), 0);
- dmu_objset_rele(clone, FTAG);
- if (error) {
- nvlist_free(nvprops);
- return (error);
- }
- } else {
- boolean_t is_insensitive = B_FALSE;
+ if (cbfunc == NULL)
+ return (SET_ERROR(EINVAL));
- if (cbfunc == NULL) {
- nvlist_free(nvprops);
- return (EINVAL);
- }
+ if (type == DMU_OST_ZVOL) {
+ uint64_t volsize, volblocksize;
- if (type == DMU_OST_ZVOL) {
- uint64_t volsize, volblocksize;
+ if (nvprops == NULL)
+ return (SET_ERROR(EINVAL));
+ if (nvlist_lookup_uint64(nvprops,
+ zfs_prop_to_name(ZFS_PROP_VOLSIZE), &volsize) != 0)
+ return (SET_ERROR(EINVAL));
- if (nvprops == NULL ||
- nvlist_lookup_uint64(nvprops,
- zfs_prop_to_name(ZFS_PROP_VOLSIZE),
- &volsize) != 0) {
- nvlist_free(nvprops);
- return (EINVAL);
- }
+ if ((error = nvlist_lookup_uint64(nvprops,
+ zfs_prop_to_name(ZFS_PROP_VOLBLOCKSIZE),
+ &volblocksize)) != 0 && error != ENOENT)
+ return (SET_ERROR(EINVAL));
- if ((error = nvlist_lookup_uint64(nvprops,
- zfs_prop_to_name(ZFS_PROP_VOLBLOCKSIZE),
- &volblocksize)) != 0 && error != ENOENT) {
- nvlist_free(nvprops);
- return (EINVAL);
- }
+ if (error != 0)
+ volblocksize = zfs_prop_default_numeric(
+ ZFS_PROP_VOLBLOCKSIZE);
- if (error != 0)
- volblocksize = zfs_prop_default_numeric(
- ZFS_PROP_VOLBLOCKSIZE);
-
- if ((error = zvol_check_volblocksize(
- volblocksize)) != 0 ||
- (error = zvol_check_volsize(volsize,
- volblocksize)) != 0) {
- nvlist_free(nvprops);
- return (error);
- }
- } else if (type == DMU_OST_ZFS) {
- int error;
+ if ((error = zvol_check_volblocksize(fsname,
+ volblocksize)) != 0 ||
+ (error = zvol_check_volsize(volsize,
+ volblocksize)) != 0)
+ return (error);
+ } else if (type == DMU_OST_ZFS) {
+ int error;
- /*
- * We have to have normalization and
- * case-folding flags correct when we do the
- * file system creation, so go figure them out
- * now.
- */
- VERIFY(nvlist_alloc(&zct.zct_zplprops,
- NV_UNIQUE_NAME, KM_SLEEP) == 0);
- error = zfs_fill_zplprops(zc->zc_name, nvprops,
- zct.zct_zplprops, &is_insensitive);
- if (error != 0) {
- nvlist_free(nvprops);
- nvlist_free(zct.zct_zplprops);
- return (error);
- }
+ /*
+ * We have to have normalization and
+ * case-folding flags correct when we do the
+ * file system creation, so go figure them out
+ * now.
+ */
+ VERIFY(nvlist_alloc(&zct.zct_zplprops,
+ NV_UNIQUE_NAME, KM_SLEEP) == 0);
+ error = zfs_fill_zplprops(fsname, nvprops,
+ zct.zct_zplprops, &is_insensitive);
+ if (error != 0) {
+ nvlist_free(zct.zct_zplprops);
+ return (error);
}
- error = dmu_objset_create(zc->zc_name, type,
- is_insensitive ? DS_FLAG_CI_DATASET : 0, cbfunc, &zct);
- nvlist_free(zct.zct_zplprops);
}
+ error = dmu_objset_create(fsname, type,
+ is_insensitive ? DS_FLAG_CI_DATASET : 0, cbfunc, &zct);
+ nvlist_free(zct.zct_zplprops);
+
/*
* It would be nice to do this atomically.
*/
if (error == 0) {
- error = zfs_set_prop_nvlist(zc->zc_name, ZPROP_SRC_LOCAL,
- nvprops, NULL);
- if (error != 0)
- (void) dmu_objset_destroy(zc->zc_name, B_FALSE);
+ error = zfs_set_prop_nvlist(fsname, ZPROP_SRC_LOCAL,
+ nvprops, outnvl);
+ if (error != 0) {
+ spa_t *spa;
+ int error2;
+
+ /*
+ * Volumes will return EBUSY and cannot be destroyed
+ * until all asynchronous minor handling has completed.
+ * Wait for the spa_zvol_taskq to drain then retry.
+ */
+ error2 = dsl_destroy_head(fsname);
+ while ((error2 == EBUSY) && (type == DMU_OST_ZVOL)) {
+ error2 = spa_open(fsname, &spa, FTAG);
+ if (error2 == 0) {
+ taskq_wait(spa->spa_zvol_taskq);
+ spa_close(spa, FTAG);
+ }
+ error2 = dsl_destroy_head(fsname);
+ }
+ }
}
- nvlist_free(nvprops);
return (error);
}
/*
- * inputs:
- * zc_name name of filesystem
- * zc_value short name of snapshot
- * zc_cookie recursive flag
- * zc_nvlist_src[_size] property list
+ * innvl: {
+ * "origin" -> name of origin snapshot
+ * (optional) "props" -> { prop -> value }
+ * }
*
* outputs:
- * zc_value short snapname (i.e. part after the '@')
+ * outnvl: propname -> error code (int32)
*/
static int
-zfs_ioc_snapshot(zfs_cmd_t *zc)
+zfs_ioc_clone(const char *fsname, nvlist_t *innvl, nvlist_t *outnvl)
{
+ int error = 0;
nvlist_t *nvprops = NULL;
- int error;
- boolean_t recursive = zc->zc_cookie;
+ char *origin_name;
- if (snapshot_namecheck(zc->zc_value, NULL, NULL) != 0)
- return (EINVAL);
+ if (nvlist_lookup_string(innvl, "origin", &origin_name) != 0)
+ return (SET_ERROR(EINVAL));
+ (void) nvlist_lookup_nvlist(innvl, "props", &nvprops);
- if (zc->zc_nvlist_src != 0 &&
- (error = get_nvlist(zc->zc_nvlist_src, zc->zc_nvlist_src_size,
- zc->zc_iflags, &nvprops)) != 0)
- return (error);
+ if (strchr(fsname, '@') ||
+ strchr(fsname, '%'))
+ return (SET_ERROR(EINVAL));
- error = zfs_check_userprops(zc->zc_name, nvprops);
- if (error)
- goto out;
+ if (dataset_namecheck(origin_name, NULL, NULL) != 0)
+ return (SET_ERROR(EINVAL));
+ error = dmu_objset_clone(fsname, origin_name);
+ if (error != 0)
+ return (error);
- if (!nvlist_empty(nvprops) &&
- zfs_earlier_version(zc->zc_name, SPA_VERSION_SNAP_PROPS)) {
- error = ENOTSUP;
- goto out;
+ /*
+ * It would be nice to do this atomically.
+ */
+ if (error == 0) {
+ error = zfs_set_prop_nvlist(fsname, ZPROP_SRC_LOCAL,
+ nvprops, outnvl);
+ if (error != 0)
+ (void) dsl_destroy_head(fsname);
}
-
- error = dmu_objset_snapshot(zc->zc_name, zc->zc_value, NULL,
- nvprops, recursive, B_FALSE, -1);
-
-out:
- nvlist_free(nvprops);
return (error);
}
-int
-zfs_unmount_snap(const char *name, void *arg)
+/*
+ * innvl: {
+ * "snaps" -> { snapshot1, snapshot2 }
+ * (optional) "props" -> { prop -> value (string) }
+ * }
+ *
+ * outnvl: snapshot -> error code (int32)
+ */
+static int
+zfs_ioc_snapshot(const char *poolname, nvlist_t *innvl, nvlist_t *outnvl)
{
-#ifdef HAVE_SNAPSHOT
- vfs_t *vfsp = NULL;
+ nvlist_t *snaps;
+ nvlist_t *props = NULL;
+ int error, poollen;
+ nvpair_t *pair, *pair2;
- if (arg) {
- char *snapname = arg;
- char *fullname = kmem_asprintf("%s@%s", name, snapname);
- vfsp = zfs_get_vfs(fullname);
- strfree(fullname);
- } else if (strchr(name, '@')) {
- vfsp = zfs_get_vfs(name);
- }
+ (void) nvlist_lookup_nvlist(innvl, "props", &props);
+ if ((error = zfs_check_userprops(poolname, props)) != 0)
+ return (error);
+
+ if (!nvlist_empty(props) &&
+ zfs_earlier_version(poolname, SPA_VERSION_SNAP_PROPS))
+ return (SET_ERROR(ENOTSUP));
+
+ if (nvlist_lookup_nvlist(innvl, "snaps", &snaps) != 0)
+ return (SET_ERROR(EINVAL));
+ poollen = strlen(poolname);
+ for (pair = nvlist_next_nvpair(snaps, NULL); pair != NULL;
+ pair = nvlist_next_nvpair(snaps, pair)) {
+ const char *name = nvpair_name(pair);
+ const char *cp = strchr(name, '@');
- if (vfsp) {
/*
- * Always force the unmount for snapshots.
+ * The snap name must contain an @, and the part after it must
+ * contain only valid characters.
*/
- int flag = MS_FORCE;
- int err;
+ if (cp == NULL ||
+ zfs_component_namecheck(cp + 1, NULL, NULL) != 0)
+ return (SET_ERROR(EINVAL));
- if ((err = vn_vfswlock(vfsp->vfs_vnodecovered)) != 0) {
- mntput(vfsp);
- return (err);
+ /*
+ * The snap must be in the specified pool.
+ */
+ if (strncmp(name, poolname, poollen) != 0 ||
+ (name[poollen] != '/' && name[poollen] != '@'))
+ return (SET_ERROR(EXDEV));
+
+ /* This must be the only snap of this fs. */
+ for (pair2 = nvlist_next_nvpair(snaps, pair);
+ pair2 != NULL; pair2 = nvlist_next_nvpair(snaps, pair2)) {
+ if (strncmp(name, nvpair_name(pair2), cp - name + 1)
+ == 0) {
+ return (SET_ERROR(EXDEV));
+ }
}
- mntput(vfsp);
- if ((err = dounmount(vfsp, flag, kcred)) != 0)
- return (err);
}
-#endif /* HAVE_SNAPSHOT */
- return (0);
+
+ error = dsl_dataset_snapshot(snaps, props, outnvl);
+
+ return (error);
}
/*
- * inputs:
- * zc_name name of filesystem
- * zc_value short name of snapshot
- * zc_defer_destroy mark for deferred destroy
- *
- * outputs: none
+ * innvl: "message" -> string
*/
+/* ARGSUSED */
static int
-zfs_ioc_destroy_snaps(zfs_cmd_t *zc)
+zfs_ioc_log_history(const char *unused, nvlist_t *innvl, nvlist_t *outnvl)
{
- int err;
+ char *message;
+ spa_t *spa;
+ int error;
+ char *poolname;
- if (snapshot_namecheck(zc->zc_value, NULL, NULL) != 0)
- return (EINVAL);
- err = dmu_objset_find(zc->zc_name,
- zfs_unmount_snap, zc->zc_value, DS_FIND_CHILDREN);
- if (err)
- return (err);
- return (dmu_snapshots_destroy(zc->zc_name, zc->zc_value,
- zc->zc_defer_destroy));
+ /*
+ * The poolname in the ioctl is not set, we get it from the TSD,
+ * which was set at the end of the last successful ioctl that allows
+ * logging. The secpolicy func already checked that it is set.
+ * Only one log ioctl is allowed after each successful ioctl, so
+ * we clear the TSD here.
+ */
+ poolname = tsd_get(zfs_allow_log_key);
+ (void) tsd_set(zfs_allow_log_key, NULL);
+ error = spa_open(poolname, &spa, FTAG);
+ strfree(poolname);
+ if (error != 0)
+ return (error);
+
+ if (nvlist_lookup_string(innvl, "message", &message) != 0) {
+ spa_close(spa, FTAG);
+ return (SET_ERROR(EINVAL));
+ }
+
+ if (spa_version(spa) < SPA_VERSION_ZPOOL_HISTORY) {
+ spa_close(spa, FTAG);
+ return (SET_ERROR(ENOTSUP));
+ }
+
+ error = spa_history_log(spa, message);
+ spa_close(spa, FTAG);
+ return (error);
}
/*
- * inputs:
- * zc_name name of dataset to destroy
- * zc_objset_type type of objset
- * zc_defer_destroy mark for deferred destroy
+ * The dp_config_rwlock must not be held when calling this, because the
+ * unmount may need to write out data.
*
- * outputs: none
+ * This function is best-effort. Callers must deal gracefully if it
+ * remains mounted (or is remounted after this call).
+ *
+ * Returns 0 if the argument is not a snapshot, or it is not currently a
+ * filesystem, or we were able to unmount it. Returns error code otherwise.
*/
-static int
-zfs_ioc_destroy(zfs_cmd_t *zc)
+int
+zfs_unmount_snap(const char *snapname)
{
int err;
- if (strchr(zc->zc_name, '@') && zc->zc_objset_type == DMU_OST_ZFS) {
- err = zfs_unmount_snap(zc->zc_name, NULL);
- if (err)
- return (err);
- }
- err = dmu_objset_destroy(zc->zc_name, zc->zc_defer_destroy);
- if (zc->zc_objset_type == DMU_OST_ZVOL && err == 0)
- (void) zvol_remove_minor(zc->zc_name);
- return (err);
+ if (strchr(snapname, '@') == NULL)
+ return (0);
+
+ err = zfsctl_snapshot_unmount((char *)snapname, MNT_FORCE);
+ if (err != 0 && err != ENOENT)
+ return (SET_ERROR(err));
+
+ return (0);
+}
+
+/* ARGSUSED */
+static int
+zfs_unmount_snap_cb(const char *snapname, void *arg)
+{
+ return (zfs_unmount_snap(snapname));
}
/*
- * inputs:
- * zc_name name of dataset to rollback (to most recent snapshot)
- *
- * outputs: none
+ * When a clone is destroyed, its origin may also need to be destroyed,
+ * in which case it must be unmounted. This routine will do that unmount
+ * if necessary.
*/
-static int
-zfs_ioc_rollback(zfs_cmd_t *zc)
+void
+zfs_destroy_unmount_origin(const char *fsname)
{
- dsl_dataset_t *ds, *clone;
int error;
- zfs_sb_t *zsb;
- char *clone_name;
-
- error = dsl_dataset_hold(zc->zc_name, FTAG, &ds);
- if (error)
- return (error);
-
- /* must not be a snapshot */
- if (dsl_dataset_is_snapshot(ds)) {
- dsl_dataset_rele(ds, FTAG);
- return (EINVAL);
- }
-
- /* must have a most recent snapshot */
- if (ds->ds_phys->ds_prev_snap_txg < TXG_INITIAL) {
- dsl_dataset_rele(ds, FTAG);
- return (EINVAL);
- }
+ objset_t *os;
+ dsl_dataset_t *ds;
- /*
- * Create clone of most recent snapshot.
- */
- clone_name = kmem_asprintf("%s/%%rollback", zc->zc_name);
- error = dmu_objset_clone(clone_name, ds->ds_prev, DS_FLAG_INCONSISTENT);
- if (error)
- goto out;
+ error = dmu_objset_hold(fsname, FTAG, &os);
+ if (error != 0)
+ return;
+ ds = dmu_objset_ds(os);
+ if (dsl_dir_is_clone(ds->ds_dir) && DS_IS_DEFER_DESTROY(ds->ds_prev)) {
+ char originname[MAXNAMELEN];
+ dsl_dataset_name(ds->ds_prev, originname);
+ dmu_objset_rele(os, FTAG);
+ (void) zfs_unmount_snap(originname);
+ } else {
+ dmu_objset_rele(os, FTAG);
+ }
+}
- error = dsl_dataset_own(clone_name, B_TRUE, FTAG, &clone);
- if (error)
- goto out;
+/*
+ * innvl: {
+ * "snaps" -> { snapshot1, snapshot2 }
+ * (optional boolean) "defer"
+ * }
+ *
+ * outnvl: snapshot -> error code (int32)
+ */
+/* ARGSUSED */
+static int
+zfs_ioc_destroy_snaps(const char *poolname, nvlist_t *innvl, nvlist_t *outnvl)
+{
+ nvlist_t *snaps;
+ nvpair_t *pair;
+ boolean_t defer;
- /*
- * Do clone swap.
- */
- if (get_zfs_sb(zc->zc_name, &zsb) == 0) {
+ if (nvlist_lookup_nvlist(innvl, "snaps", &snaps) != 0)
+ return (SET_ERROR(EINVAL));
+ defer = nvlist_exists(innvl, "defer");
+
+ for (pair = nvlist_next_nvpair(snaps, NULL); pair != NULL;
+ pair = nvlist_next_nvpair(snaps, pair)) {
+ (void) zfs_unmount_snap(nvpair_name(pair));
+ }
+
+ return (dsl_destroy_snapshots_nvl(snaps, defer, outnvl));
+}
+
+/*
+ * Create bookmarks. Bookmark names are of the form <fs>#<bmark>.
+ * All bookmarks must be in the same pool.
+ *
+ * innvl: {
+ * bookmark1 -> snapshot1, bookmark2 -> snapshot2
+ * }
+ *
+ * outnvl: bookmark -> error code (int32)
+ *
+ */
+/* ARGSUSED */
+static int
+zfs_ioc_bookmark(const char *poolname, nvlist_t *innvl, nvlist_t *outnvl)
+{
+ nvpair_t *pair, *pair2;
+
+ for (pair = nvlist_next_nvpair(innvl, NULL);
+ pair != NULL; pair = nvlist_next_nvpair(innvl, pair)) {
+ char *snap_name;
+
+ /*
+ * Verify the snapshot argument.
+ */
+ if (nvpair_value_string(pair, &snap_name) != 0)
+ return (SET_ERROR(EINVAL));
+
+
+ /* Verify that the keys (bookmarks) are unique */
+ for (pair2 = nvlist_next_nvpair(innvl, pair);
+ pair2 != NULL; pair2 = nvlist_next_nvpair(innvl, pair2)) {
+ if (strcmp(nvpair_name(pair), nvpair_name(pair2)) == 0)
+ return (SET_ERROR(EINVAL));
+ }
+ }
+
+ return (dsl_bookmark_create(innvl, outnvl));
+}
+
+/*
+ * innvl: {
+ * property 1, property 2, ...
+ * }
+ *
+ * outnvl: {
+ * bookmark name 1 -> { property 1, property 2, ... },
+ * bookmark name 2 -> { property 1, property 2, ... }
+ * }
+ *
+ */
+static int
+zfs_ioc_get_bookmarks(const char *fsname, nvlist_t *innvl, nvlist_t *outnvl)
+{
+ return (dsl_get_bookmarks(fsname, innvl, outnvl));
+}
+
+/*
+ * innvl: {
+ * bookmark name 1, bookmark name 2
+ * }
+ *
+ * outnvl: bookmark -> error code (int32)
+ *
+ */
+static int
+zfs_ioc_destroy_bookmarks(const char *poolname, nvlist_t *innvl,
+ nvlist_t *outnvl)
+{
+ int error, poollen;
+ nvpair_t *pair;
+
+ poollen = strlen(poolname);
+ for (pair = nvlist_next_nvpair(innvl, NULL);
+ pair != NULL; pair = nvlist_next_nvpair(innvl, pair)) {
+ const char *name = nvpair_name(pair);
+ const char *cp = strchr(name, '#');
+
+ /*
+ * The bookmark name must contain an #, and the part after it
+ * must contain only valid characters.
+ */
+ if (cp == NULL ||
+ zfs_component_namecheck(cp + 1, NULL, NULL) != 0)
+ return (SET_ERROR(EINVAL));
+
+ /*
+ * The bookmark must be in the specified pool.
+ */
+ if (strncmp(name, poolname, poollen) != 0 ||
+ (name[poollen] != '/' && name[poollen] != '#'))
+ return (SET_ERROR(EXDEV));
+ }
+
+ error = dsl_bookmark_destroy(innvl, outnvl);
+ return (error);
+}
+
+/*
+ * inputs:
+ * zc_name name of dataset to destroy
+ * zc_objset_type type of objset
+ * zc_defer_destroy mark for deferred destroy
+ *
+ * outputs: none
+ */
+static int
+zfs_ioc_destroy(zfs_cmd_t *zc)
+{
+ int err;
+
+ if (zc->zc_objset_type == DMU_OST_ZFS) {
+ err = zfs_unmount_snap(zc->zc_name);
+ if (err != 0)
+ return (err);
+ }
+
+ if (strchr(zc->zc_name, '@'))
+ err = dsl_destroy_snapshot(zc->zc_name, zc->zc_defer_destroy);
+ else
+ err = dsl_destroy_head(zc->zc_name);
+
+ return (err);
+}
+
+/*
+ * fsname is name of dataset to rollback (to most recent snapshot)
+ *
+ * innvl is not used.
+ *
+ * outnvl: "target" -> name of most recent snapshot
+ * }
+ */
+/* ARGSUSED */
+static int
+zfs_ioc_rollback(const char *fsname, nvlist_t *args, nvlist_t *outnvl)
+{
+ zfs_sb_t *zsb;
+ int error;
+
+ if (get_zfs_sb(fsname, &zsb) == 0) {
error = zfs_suspend_fs(zsb);
if (error == 0) {
int resume_err;
- if (dsl_dataset_tryown(ds, B_FALSE, FTAG)) {
- error = dsl_dataset_clone_swap(clone, ds,
- B_TRUE);
- dsl_dataset_disown(ds, FTAG);
- ds = NULL;
- } else {
- error = EBUSY;
- }
- resume_err = zfs_resume_fs(zsb, zc->zc_name);
+ error = dsl_dataset_rollback(fsname, zsb, outnvl);
+ resume_err = zfs_resume_fs(zsb, fsname);
error = error ? error : resume_err;
}
deactivate_super(zsb->z_sb);
} else {
- if (dsl_dataset_tryown(ds, B_FALSE, FTAG)) {
- error = dsl_dataset_clone_swap(clone, ds, B_TRUE);
- dsl_dataset_disown(ds, FTAG);
- ds = NULL;
- } else {
- error = EBUSY;
- }
+ error = dsl_dataset_rollback(fsname, NULL, outnvl);
}
+ return (error);
+}
- /*
- * Destroy clone (which also closes it).
- */
- (void) dsl_dataset_destroy(clone, FTAG, B_FALSE);
+static int
+recursive_unmount(const char *fsname, void *arg)
+{
+ const char *snapname = arg;
+ char *fullname;
+ int error;
+
+ fullname = kmem_asprintf("%s@%s", fsname, snapname);
+ error = zfs_unmount_snap(fullname);
+ strfree(fullname);
-out:
- strfree(clone_name);
- if (ds)
- dsl_dataset_rele(ds, FTAG);
return (error);
}
zfs_ioc_rename(zfs_cmd_t *zc)
{
boolean_t recursive = zc->zc_cookie & 1;
- int err;
+ char *at;
zc->zc_value[sizeof (zc->zc_value) - 1] = '\0';
if (dataset_namecheck(zc->zc_value, NULL, NULL) != 0 ||
strchr(zc->zc_value, '%'))
- return (EINVAL);
-
- /*
- * Unmount snapshot unless we're doing a recursive rename,
- * in which case the dataset code figures out which snapshots
- * to unmount.
- */
- if (!recursive && strchr(zc->zc_name, '@') != NULL &&
- zc->zc_objset_type == DMU_OST_ZFS) {
- err = zfs_unmount_snap(zc->zc_name, NULL);
- if (err)
- return (err);
- }
+ return (SET_ERROR(EINVAL));
+
+ at = strchr(zc->zc_name, '@');
+ if (at != NULL) {
+ /* snaps must be in same fs */
+ int error;
+
+ if (strncmp(zc->zc_name, zc->zc_value, at - zc->zc_name + 1))
+ return (SET_ERROR(EXDEV));
+ *at = '\0';
+ if (zc->zc_objset_type == DMU_OST_ZFS) {
+ error = dmu_objset_find(zc->zc_name,
+ recursive_unmount, at + 1,
+ recursive ? DS_FIND_CHILDREN : 0);
+ if (error != 0) {
+ *at = '@';
+ return (error);
+ }
+ }
+ error = dsl_dataset_rename_snapshot(zc->zc_name,
+ at + 1, strchr(zc->zc_value, '@') + 1, recursive);
+ *at = '@';
- err = dmu_objset_rename(zc->zc_name, zc->zc_value, recursive);
- if ((err == 0) && (zc->zc_objset_type == DMU_OST_ZVOL)) {
- (void) zvol_remove_minor(zc->zc_name);
- (void) zvol_create_minor(zc->zc_value);
+ return (error);
+ } else {
+ return (dsl_dir_rename(zc->zc_name, zc->zc_value));
}
-
- return (err);
}
static int
perm = ZFS_DELEG_PERM_GROUPQUOTA;
} else {
/* USERUSED and GROUPUSED are read-only */
- return (EINVAL);
+ return (SET_ERROR(EINVAL));
}
if ((err = zfs_secpolicy_write_perms(dsname, perm, cr)))
return (0);
}
- return (EINVAL);
+ return (SET_ERROR(EINVAL));
}
if (issnap)
- return (EINVAL);
+ return (SET_ERROR(EINVAL));
if (nvpair_type(pair) == DATA_TYPE_NVLIST) {
/*
* the SPA supports it. We ignore any errors here since
* we'll catch them later.
*/
- if (nvpair_type(pair) == DATA_TYPE_UINT64 &&
- nvpair_value_uint64(pair, &intval) == 0) {
+ if (nvpair_value_uint64(pair, &intval) == 0) {
if (intval >= ZIO_COMPRESS_GZIP_1 &&
intval <= ZIO_COMPRESS_GZIP_9 &&
zfs_earlier_version(dsname,
SPA_VERSION_GZIP_COMPRESSION)) {
- return (ENOTSUP);
+ return (SET_ERROR(ENOTSUP));
}
if (intval == ZIO_COMPRESS_ZLE &&
zfs_earlier_version(dsname,
SPA_VERSION_ZLE_COMPRESSION))
- return (ENOTSUP);
+ return (SET_ERROR(ENOTSUP));
+
+ if (intval == ZIO_COMPRESS_LZ4) {
+ spa_t *spa;
+
+ if ((err = spa_open(dsname, &spa, FTAG)) != 0)
+ return (err);
+
+ if (!spa_feature_is_enabled(spa,
+ SPA_FEATURE_LZ4_COMPRESS)) {
+ spa_close(spa, FTAG);
+ return (SET_ERROR(ENOTSUP));
+ }
+ spa_close(spa, FTAG);
+ }
/*
* If this is a bootable dataset then
*/
if (zfs_is_bootfs(dsname) &&
!BOOTFS_COMPRESS_VALID(intval)) {
- return (ERANGE);
+ return (SET_ERROR(ERANGE));
}
}
break;
case ZFS_PROP_COPIES:
if (zfs_earlier_version(dsname, SPA_VERSION_DITTO_BLOCKS))
- return (ENOTSUP);
+ return (SET_ERROR(ENOTSUP));
break;
case ZFS_PROP_DEDUP:
if (zfs_earlier_version(dsname, SPA_VERSION_DEDUP))
- return (ENOTSUP);
+ return (SET_ERROR(ENOTSUP));
+ break;
+
+ case ZFS_PROP_VOLBLOCKSIZE:
+ case ZFS_PROP_RECORDSIZE:
+ /* Record sizes above 128k need the feature to be enabled */
+ if (nvpair_value_uint64(pair, &intval) == 0 &&
+ intval > SPA_OLD_MAXBLOCKSIZE) {
+ spa_t *spa;
+
+ /*
+ * If this is a bootable dataset then
+ * the we don't allow large (>128K) blocks,
+ * because GRUB doesn't support them.
+ */
+ if (zfs_is_bootfs(dsname) &&
+ intval > SPA_OLD_MAXBLOCKSIZE) {
+ return (SET_ERROR(ERANGE));
+ }
+
+ /*
+ * We don't allow setting the property above 1MB,
+ * unless the tunable has been changed.
+ */
+ if (intval > zfs_max_recordsize ||
+ intval > SPA_MAXBLOCKSIZE)
+ return (SET_ERROR(ERANGE));
+
+ if ((err = spa_open(dsname, &spa, FTAG)) != 0)
+ return (err);
+
+ if (!spa_feature_is_enabled(spa,
+ SPA_FEATURE_LARGE_BLOCKS)) {
+ spa_close(spa, FTAG);
+ return (SET_ERROR(ENOTSUP));
+ }
+ spa_close(spa, FTAG);
+ }
break;
case ZFS_PROP_SHARESMB:
if (zpl_earlier_version(dsname, ZPL_VERSION_FUID))
- return (ENOTSUP);
+ return (SET_ERROR(ENOTSUP));
break;
case ZFS_PROP_ACLINHERIT:
if (intval == ZFS_ACL_PASSTHROUGH_X &&
zfs_earlier_version(dsname,
SPA_VERSION_PASSTHROUGH_X))
- return (ENOTSUP);
+ return (SET_ERROR(ENOTSUP));
}
break;
default:
VERIFY(nvlist_alloc(&errors, NV_UNIQUE_NAME, KM_SLEEP) == 0);
- zc = kmem_alloc(sizeof (zfs_cmd_t), KM_SLEEP | KM_NODEBUG);
+ zc = kmem_alloc(sizeof (zfs_cmd_t), KM_SLEEP);
(void) strcpy(zc->zc_name, dataset);
pair = nvlist_next_nvpair(props, NULL);
while (pair != NULL) {
(void) strcpy(zc->zc_value, nvpair_name(pair));
if ((err = zfs_check_settable(dataset, pair, CRED())) != 0 ||
- (err = zfs_secpolicy_inherit(zc, CRED())) != 0) {
+ (err = zfs_secpolicy_inherit_prop(zc, NULL, CRED())) != 0) {
VERIFY(nvlist_remove_nvpair(props, pair) == 0);
VERIFY(nvlist_add_int32(errors,
zc->zc_value, err) == 0);
zfs_ioc_recv(zfs_cmd_t *zc)
{
file_t *fp;
- objset_t *os;
dmu_recv_cookie_t drc;
boolean_t force = (boolean_t)zc->zc_guid;
int fd;
offset_t off;
nvlist_t *props = NULL; /* sent properties */
nvlist_t *origprops = NULL; /* existing properties */
- objset_t *origin = NULL;
+ char *origin = NULL;
char *tosnap;
char tofs[ZFS_MAXNAMELEN];
boolean_t first_recvd_props = B_FALSE;
if (dataset_namecheck(zc->zc_value, NULL, NULL) != 0 ||
strchr(zc->zc_value, '@') == NULL ||
strchr(zc->zc_value, '%'))
- return (EINVAL);
+ return (SET_ERROR(EINVAL));
(void) strcpy(tofs, zc->zc_value);
tosnap = strchr(tofs, '@');
fp = getf(fd);
if (fp == NULL) {
nvlist_free(props);
- return (EBADF);
+ return (SET_ERROR(EBADF));
}
VERIFY(nvlist_alloc(&errors, NV_UNIQUE_NAME, KM_SLEEP) == 0);
- if (props && dmu_objset_hold(tofs, FTAG, &os) == 0) {
- if ((spa_version(os->os_spa) >= SPA_VERSION_RECVD_PROPS) &&
- !dsl_prop_get_hasrecvd(os)) {
+ if (zc->zc_string[0])
+ origin = zc->zc_string;
+
+ error = dmu_recv_begin(tofs, tosnap,
+ &zc->zc_begin_record, force, origin, &drc);
+ if (error != 0)
+ goto out;
+
+ /*
+ * Set properties before we receive the stream so that they are applied
+ * to the new data. Note that we must call dmu_recv_stream() if
+ * dmu_recv_begin() succeeds.
+ */
+ if (props != NULL && !drc.drc_newfs) {
+ if (spa_version(dsl_dataset_get_spa(drc.drc_ds)) >=
+ SPA_VERSION_RECVD_PROPS &&
+ !dsl_prop_get_hasrecvd(tofs))
first_recvd_props = B_TRUE;
- }
/*
* If new received properties are supplied, they are to
* completely replace the existing received properties, so stash
* away the existing ones.
*/
- if (dsl_prop_get_received(os, &origprops) == 0) {
+ if (dsl_prop_get_received(tofs, &origprops) == 0) {
nvlist_t *errlist = NULL;
/*
* Don't bother writing a property if its value won't
*/
if (!first_recvd_props)
props_reduce(props, origprops);
- if (zfs_check_clearable(tofs, origprops,
- &errlist) != 0)
+ if (zfs_check_clearable(tofs, origprops, &errlist) != 0)
(void) nvlist_merge(errors, errlist, 0);
nvlist_free(errlist);
- }
-
- dmu_objset_rele(os, FTAG);
- }
-
- if (zc->zc_string[0]) {
- error = dmu_objset_hold(zc->zc_string, FTAG, &origin);
- if (error)
- goto out;
- }
-
- error = dmu_recv_begin(tofs, tosnap, zc->zc_top_ds,
- &zc->zc_begin_record, force, origin, &drc);
- if (origin)
- dmu_objset_rele(origin, FTAG);
- if (error)
- goto out;
- /*
- * Set properties before we receive the stream so that they are applied
- * to the new data. Note that we must call dmu_recv_stream() if
- * dmu_recv_begin() succeeds.
- */
- if (props) {
- nvlist_t *errlist;
-
- if (dmu_objset_from_ds(drc.drc_logical_ds, &os) == 0) {
- if (drc.drc_newfs) {
- if (spa_version(os->os_spa) >=
- SPA_VERSION_RECVD_PROPS)
- first_recvd_props = B_TRUE;
- } else if (origprops != NULL) {
- if (clear_received_props(os, tofs, origprops,
- first_recvd_props ? NULL : props) != 0)
- zc->zc_obj |= ZPROP_ERR_NOCLEAR;
- } else {
+ if (clear_received_props(tofs, origprops,
+ first_recvd_props ? NULL : props) != 0)
zc->zc_obj |= ZPROP_ERR_NOCLEAR;
- }
- dsl_prop_set_hasrecvd(os);
- } else if (!drc.drc_newfs) {
+ } else {
zc->zc_obj |= ZPROP_ERR_NOCLEAR;
}
+ }
- (void) zfs_set_prop_nvlist(tofs, ZPROP_SRC_RECEIVED,
- props, &errlist);
- (void) nvlist_merge(errors, errlist, 0);
- nvlist_free(errlist);
+ if (props != NULL) {
+ props_error = dsl_prop_set_hasrecvd(tofs);
+
+ if (props_error == 0) {
+ (void) zfs_set_prop_nvlist(tofs, ZPROP_SRC_RECEIVED,
+ props, errors);
+ }
}
- if (fit_error_list(zc, &errors) != 0 || put_nvlist(zc, errors) != 0) {
+ if (zc->zc_nvlist_dst_size != 0 &&
+ (nvlist_smush(errors, zc->zc_nvlist_dst_size) != 0 ||
+ put_nvlist(zc, errors) != 0)) {
/*
* Caller made zc->zc_nvlist_dst less than the minimum expected
* size or supplied an invalid address.
*/
- props_error = EINVAL;
+ props_error = SET_ERROR(EINVAL);
}
off = fp->f_offset;
* If the suspend fails, then the recv_end will
* likely also fail, and clean up after itself.
*/
- end_err = dmu_recv_end(&drc);
+ end_err = dmu_recv_end(&drc, zsb);
if (error == 0)
error = zfs_resume_fs(zsb, tofs);
error = error ? error : end_err;
deactivate_super(zsb->z_sb);
} else {
- error = dmu_recv_end(&drc);
+ error = dmu_recv_end(&drc, NULL);
}
}
error = 1;
}
#endif
+
/*
* On error, restore the original props.
*/
- if (error && props) {
- if (dmu_objset_hold(tofs, FTAG, &os) == 0) {
- if (clear_received_props(os, tofs, props, NULL) != 0) {
- /*
- * We failed to clear the received properties.
- * Since we may have left a $recvd value on the
- * system, we can't clear the $hasrecvd flag.
- */
- zc->zc_obj |= ZPROP_ERR_NORESTORE;
- } else if (first_recvd_props) {
- dsl_prop_unset_hasrecvd(os);
- }
- dmu_objset_rele(os, FTAG);
- } else if (!drc.drc_newfs) {
- /* We failed to clear the received properties. */
+ if (error != 0 && props != NULL && !drc.drc_newfs) {
+ if (clear_received_props(tofs, props, NULL) != 0) {
+ /*
+ * We failed to clear the received properties.
+ * Since we may have left a $recvd value on the
+ * system, we can't clear the $hasrecvd flag.
+ */
zc->zc_obj |= ZPROP_ERR_NORESTORE;
+ } else if (first_recvd_props) {
+ dsl_prop_unset_hasrecvd(tofs);
}
if (origprops == NULL && !drc.drc_newfs) {
* zc_obj fromorigin flag (mutually exclusive with zc_fromobj)
* zc_sendobj objsetid of snapshot to send
* zc_fromobj objsetid of incremental fromsnap (may be zero)
+ * zc_guid if set, estimate size of stream only. zc_cookie is ignored.
+ * output size in zc_objset_type.
+ * zc_flags lzc_send_flags
*
- * outputs: none
+ * outputs:
+ * zc_objset_type estimated size, if zc_guid is set
*/
static int
zfs_ioc_send(zfs_cmd_t *zc)
{
- objset_t *fromsnap = NULL;
- objset_t *tosnap;
- file_t *fp;
int error;
offset_t off;
- dsl_dataset_t *ds;
- dsl_dataset_t *dsfrom = NULL;
- spa_t *spa;
- dsl_pool_t *dp;
+ boolean_t estimate = (zc->zc_guid != 0);
+ boolean_t embedok = (zc->zc_flags & 0x1);
+ boolean_t large_block_ok = (zc->zc_flags & 0x2);
- error = spa_open(zc->zc_name, &spa, FTAG);
- if (error)
- return (error);
+ if (zc->zc_obj != 0) {
+ dsl_pool_t *dp;
+ dsl_dataset_t *tosnap;
- dp = spa_get_dsl(spa);
- rw_enter(&dp->dp_config_rwlock, RW_READER);
- error = dsl_dataset_hold_obj(dp, zc->zc_sendobj, FTAG, &ds);
- rw_exit(&dp->dp_config_rwlock);
- if (error) {
- spa_close(spa, FTAG);
- return (error);
- }
+ error = dsl_pool_hold(zc->zc_name, FTAG, &dp);
+ if (error != 0)
+ return (error);
- error = dmu_objset_from_ds(ds, &tosnap);
- if (error) {
- dsl_dataset_rele(ds, FTAG);
- spa_close(spa, FTAG);
- return (error);
+ error = dsl_dataset_hold_obj(dp, zc->zc_sendobj, FTAG, &tosnap);
+ if (error != 0) {
+ dsl_pool_rele(dp, FTAG);
+ return (error);
+ }
+
+ if (dsl_dir_is_clone(tosnap->ds_dir))
+ zc->zc_fromobj =
+ dsl_dir_phys(tosnap->ds_dir)->dd_origin_obj;
+ dsl_dataset_rele(tosnap, FTAG);
+ dsl_pool_rele(dp, FTAG);
}
- if (zc->zc_fromobj != 0) {
- rw_enter(&dp->dp_config_rwlock, RW_READER);
- error = dsl_dataset_hold_obj(dp, zc->zc_fromobj, FTAG, &dsfrom);
- rw_exit(&dp->dp_config_rwlock);
- spa_close(spa, FTAG);
- if (error) {
- dsl_dataset_rele(ds, FTAG);
+ if (estimate) {
+ dsl_pool_t *dp;
+ dsl_dataset_t *tosnap;
+ dsl_dataset_t *fromsnap = NULL;
+
+ error = dsl_pool_hold(zc->zc_name, FTAG, &dp);
+ if (error != 0)
return (error);
- }
- error = dmu_objset_from_ds(dsfrom, &fromsnap);
- if (error) {
- dsl_dataset_rele(dsfrom, FTAG);
- dsl_dataset_rele(ds, FTAG);
+
+ error = dsl_dataset_hold_obj(dp, zc->zc_sendobj, FTAG, &tosnap);
+ if (error != 0) {
+ dsl_pool_rele(dp, FTAG);
return (error);
}
- } else {
- spa_close(spa, FTAG);
- }
- fp = getf(zc->zc_cookie);
- if (fp == NULL) {
- dsl_dataset_rele(ds, FTAG);
- if (dsfrom)
- dsl_dataset_rele(dsfrom, FTAG);
- return (EBADF);
- }
+ if (zc->zc_fromobj != 0) {
+ error = dsl_dataset_hold_obj(dp, zc->zc_fromobj,
+ FTAG, &fromsnap);
+ if (error != 0) {
+ dsl_dataset_rele(tosnap, FTAG);
+ dsl_pool_rele(dp, FTAG);
+ return (error);
+ }
+ }
- off = fp->f_offset;
- error = dmu_sendbackup(tosnap, fromsnap, zc->zc_obj, fp->f_vnode, &off);
+ error = dmu_send_estimate(tosnap, fromsnap,
+ &zc->zc_objset_type);
- if (VOP_SEEK(fp->f_vnode, fp->f_offset, &off, NULL) == 0)
- fp->f_offset = off;
- releasef(zc->zc_cookie);
- if (dsfrom)
- dsl_dataset_rele(dsfrom, FTAG);
- dsl_dataset_rele(ds, FTAG);
+ if (fromsnap != NULL)
+ dsl_dataset_rele(fromsnap, FTAG);
+ dsl_dataset_rele(tosnap, FTAG);
+ dsl_pool_rele(dp, FTAG);
+ } else {
+ file_t *fp = getf(zc->zc_cookie);
+ if (fp == NULL)
+ return (SET_ERROR(EBADF));
+
+ off = fp->f_offset;
+ error = dmu_send_obj(zc->zc_name, zc->zc_sendobj,
+ zc->zc_fromobj, embedok, large_block_ok,
+ zc->zc_cookie, fp->f_vnode, &off);
+
+ if (VOP_SEEK(fp->f_vnode, fp->f_offset, &off, NULL) == 0)
+ fp->f_offset = off;
+ releasef(zc->zc_cookie);
+ }
return (error);
}
+/*
+ * inputs:
+ * zc_name name of snapshot on which to report progress
+ * zc_cookie file descriptor of send stream
+ *
+ * outputs:
+ * zc_cookie number of bytes written in send stream thus far
+ */
static int
-zfs_ioc_inject_fault(zfs_cmd_t *zc)
+zfs_ioc_send_progress(zfs_cmd_t *zc)
{
- int id, error;
+ dsl_pool_t *dp;
+ dsl_dataset_t *ds;
+ dmu_sendarg_t *dsp = NULL;
+ int error;
- error = zio_inject_fault(zc->zc_name, (int)zc->zc_guid, &id,
- &zc->zc_inject_record);
+ error = dsl_pool_hold(zc->zc_name, FTAG, &dp);
+ if (error != 0)
+ return (error);
+
+ error = dsl_dataset_hold(dp, zc->zc_name, FTAG, &ds);
+ if (error != 0) {
+ dsl_pool_rele(dp, FTAG);
+ return (error);
+ }
+
+ mutex_enter(&ds->ds_sendstream_lock);
+
+ /*
+ * Iterate over all the send streams currently active on this dataset.
+ * If there's one which matches the specified file descriptor _and_ the
+ * stream was started by the current process, return the progress of
+ * that stream.
+ */
+
+ for (dsp = list_head(&ds->ds_sendstreams); dsp != NULL;
+ dsp = list_next(&ds->ds_sendstreams, dsp)) {
+ if (dsp->dsa_outfd == zc->zc_cookie &&
+ dsp->dsa_proc->group_leader == curproc->group_leader)
+ break;
+ }
+
+ if (dsp != NULL)
+ zc->zc_cookie = *(dsp->dsa_off);
+ else
+ error = SET_ERROR(ENOENT);
+
+ mutex_exit(&ds->ds_sendstream_lock);
+ dsl_dataset_rele(ds, FTAG);
+ dsl_pool_rele(dp, FTAG);
+ return (error);
+}
+
+static int
+zfs_ioc_inject_fault(zfs_cmd_t *zc)
+{
+ int id, error;
+
+ error = zio_inject_fault(zc->zc_name, (int)zc->zc_guid, &id,
+ &zc->zc_inject_record);
if (error == 0)
zc->zc_guid = (uint64_t)id;
spa = spa_lookup(zc->zc_name);
if (spa == NULL) {
mutex_exit(&spa_namespace_lock);
- return (EIO);
+ return (SET_ERROR(EIO));
}
if (spa_get_log_state(spa) == SPA_LOG_MISSING) {
/* we need to let spa_open/spa_load clear the chains */
nvlist_t *config = NULL;
if (zc->zc_nvlist_src == 0)
- return (EINVAL);
+ return (SET_ERROR(EINVAL));
if ((error = get_nvlist(zc->zc_nvlist_src,
zc->zc_nvlist_src_size, zc->zc_iflags, &policy)) == 0) {
}
}
- if (error)
+ if (error != 0)
return (error);
spa_vdev_state_enter(spa, SCL_NONE);
if (vd == NULL) {
(void) spa_vdev_state_exit(spa, NULL, ENODEV);
spa_close(spa, FTAG);
- return (ENODEV);
+ return (SET_ERROR(ENODEV));
}
}
* Resume any suspended I/Os.
*/
if (zio_resume(spa) != 0)
- error = EIO;
+ error = SET_ERROR(EIO);
spa_close(spa, FTAG);
return (error);
}
+static int
+zfs_ioc_pool_reopen(zfs_cmd_t *zc)
+{
+ spa_t *spa;
+ int error;
+
+ error = spa_open(zc->zc_name, &spa, FTAG);
+ if (error != 0)
+ return (error);
+
+ spa_vdev_state_enter(spa, SCL_NONE);
+
+ /*
+ * If a resilver is already in progress then set the
+ * spa_scrub_reopen flag to B_TRUE so that we don't restart
+ * the scan as a side effect of the reopen. Otherwise, let
+ * vdev_open() decided if a resilver is required.
+ */
+ spa->spa_scrub_reopen = dsl_scan_resilvering(spa->spa_dsl_pool);
+ vdev_reopen(spa->spa_root_vdev);
+ spa->spa_scrub_reopen = B_FALSE;
+
+ (void) spa_vdev_state_exit(spa, NULL, 0);
+ spa_close(spa, FTAG);
+ return (0);
+}
/*
* inputs:
* zc_name name of filesystem
if (cp)
*cp = '\0';
(void) dmu_objset_find(zc->zc_value,
- zfs_unmount_snap, NULL, DS_FIND_SNAPSHOTS);
+ zfs_unmount_snap_cb, NULL, DS_FIND_SNAPSHOTS);
return (dsl_dataset_promote(zc->zc_name, zc->zc_string));
}
int error;
if (zc->zc_objset_type >= ZFS_NUM_USERQUOTA_PROPS)
- return (EINVAL);
+ return (SET_ERROR(EINVAL));
error = zfs_sb_hold(zc->zc_name, FTAG, &zsb, B_FALSE);
- if (error)
+ if (error != 0)
return (error);
error = zfs_userspace_one(zsb,
void *buf;
if (bufsize <= 0)
- return (ENOMEM);
+ return (SET_ERROR(ENOMEM));
error = zfs_sb_hold(zc->zc_name, FTAG, &zsb, B_FALSE);
- if (error)
+ if (error != 0)
return (error);
buf = vmem_alloc(bufsize, KM_SLEEP);
* objset_phys_t). Suspend/resume the fs will do that.
*/
error = zfs_suspend_fs(zsb);
- if (error == 0)
+ if (error == 0) {
+ dmu_objset_refresh_ownership(zsb->z_os,
+ zsb);
error = zfs_resume_fs(zsb, zc->zc_name);
+ }
}
if (error == 0)
error = dmu_objset_userspace_upgrade(zsb->z_os);
} else {
/* XXX kind of reading contents without owning */
error = dmu_objset_hold(zc->zc_name, FTAG, &os);
- if (error)
+ if (error != 0)
return (error);
error = dmu_objset_userspace_upgrade(os);
static int
zfs_ioc_share(zfs_cmd_t *zc)
{
- return (ENOSYS);
+ return (SET_ERROR(ENOSYS));
}
ace_t full_access[] = {
int error;
error = dmu_objset_hold(zc->zc_name, FTAG, &os);
- if (error)
+ if (error != 0)
return (error);
- error = dmu_object_next(os, &zc->zc_obj, B_FALSE,
- os->os_dsl_dataset->ds_phys->ds_prev_snap_txg);
+ error = dmu_object_next(os, &zc->zc_obj, B_FALSE, 0);
dmu_objset_rele(os, FTAG);
return (error);
* zc_cleanup_fd cleanup-on-exit file descriptor for calling process
*
* outputs:
+ * zc_value short name of new snapshot
*/
static int
zfs_ioc_tmp_snapshot(zfs_cmd_t *zc)
{
char *snap_name;
+ char *hold_name;
int error;
+ minor_t minor;
+
+ error = zfs_onexit_fd_hold(zc->zc_cleanup_fd, &minor);
+ if (error != 0)
+ return (error);
snap_name = kmem_asprintf("%s-%016llx", zc->zc_value,
(u_longlong_t)ddi_get_lbolt64());
+ hold_name = kmem_asprintf("%%%s", zc->zc_value);
- if (strlen(snap_name) >= MAXNAMELEN) {
- strfree(snap_name);
- return (E2BIG);
- }
-
- error = dmu_objset_snapshot(zc->zc_name, snap_name, snap_name,
- NULL, B_FALSE, B_TRUE, zc->zc_cleanup_fd);
- if (error != 0) {
- strfree(snap_name);
- return (error);
- }
-
- (void) strcpy(zc->zc_value, snap_name);
+ error = dsl_dataset_snapshot_tmp(zc->zc_name, snap_name, minor,
+ hold_name);
+ if (error == 0)
+ (void) strcpy(zc->zc_value, snap_name);
strfree(snap_name);
- return (0);
+ strfree(hold_name);
+ zfs_onexit_fd_rele(zc->zc_cleanup_fd);
+ return (error);
}
/*
static int
zfs_ioc_diff(zfs_cmd_t *zc)
{
- objset_t *fromsnap;
- objset_t *tosnap;
file_t *fp;
offset_t off;
int error;
- error = dmu_objset_hold(zc->zc_name, FTAG, &tosnap);
- if (error)
- return (error);
-
- error = dmu_objset_hold(zc->zc_value, FTAG, &fromsnap);
- if (error) {
- dmu_objset_rele(tosnap, FTAG);
- return (error);
- }
-
fp = getf(zc->zc_cookie);
- if (fp == NULL) {
- dmu_objset_rele(fromsnap, FTAG);
- dmu_objset_rele(tosnap, FTAG);
- return (EBADF);
- }
+ if (fp == NULL)
+ return (SET_ERROR(EBADF));
off = fp->f_offset;
- error = dmu_diff(tosnap, fromsnap, fp->f_vnode, &off);
+ error = dmu_diff(zc->zc_name, zc->zc_value, fp->f_vnode, &off);
if (VOP_SEEK(fp->f_vnode, fp->f_offset, &off, NULL) == 0)
fp->f_offset = off;
releasef(zc->zc_cookie);
- dmu_objset_rele(fromsnap, FTAG);
- dmu_objset_rele(tosnap, FTAG);
return (error);
}
(strcmp((char *)refstr_value(vp->v_vfsp->vfs_resource),
zc->zc_name) != 0)) {
VN_RELE(vp);
- return (EINVAL);
+ return (SET_ERROR(EINVAL));
}
dzp = VTOZ(vp);
ZFS_SHARES_DIR);
dmu_tx_hold_zap(tx, DMU_NEW_OBJECT, FALSE, NULL);
error = dmu_tx_assign(tx, TXG_WAIT);
- if (error) {
+ if (error != 0) {
dmu_tx_abort(tx);
} else {
error = zfs_create_share_dir(zsb, tx);
dmu_tx_commit(tx);
}
- if (error) {
+ if (error != 0) {
mutex_exit(&zsb->z_lock);
VN_RELE(vp);
ZFS_EXIT(zsb);
if ((error = get_nvlist(zc->zc_nvlist_src,
zc->zc_nvlist_src_size, zc->zc_iflags, &nvlist)) != 0) {
VN_RELE(vp);
+ VN_RELE(ZTOV(sharedir));
ZFS_EXIT(zsb);
return (error);
}
break;
default:
- error = EINVAL;
+ error = SET_ERROR(EINVAL);
break;
}
return (error);
#else
- return (ENOTSUP);
+ return (SET_ERROR(ENOTSUP));
#endif /* HAVE_SMB_SHARE */
}
/*
- * inputs:
- * zc_name name of filesystem
- * zc_value short name of snap
- * zc_string user-supplied tag for this hold
- * zc_cookie recursive flag
- * zc_temphold set if hold is temporary
- * zc_cleanup_fd cleanup-on-exit file descriptor for calling process
- * zc_sendobj if non-zero, the objid for zc_name@zc_value
- * zc_createtxg if zc_sendobj is non-zero, snap must have zc_createtxg
+ * innvl: {
+ * "holds" -> { snapname -> holdname (string), ... }
+ * (optional) "cleanup_fd" -> fd (int32)
+ * }
*
- * outputs: none
+ * outnvl: {
+ * snapname -> error value (int32)
+ * ...
+ * }
*/
+/* ARGSUSED */
static int
-zfs_ioc_hold(zfs_cmd_t *zc)
+zfs_ioc_hold(const char *pool, nvlist_t *args, nvlist_t *errlist)
{
- boolean_t recursive = zc->zc_cookie;
- spa_t *spa;
- dsl_pool_t *dp;
- dsl_dataset_t *ds;
+ nvpair_t *pair;
+ nvlist_t *holds;
+ int cleanup_fd = -1;
int error;
minor_t minor = 0;
- if (snapshot_namecheck(zc->zc_value, NULL, NULL) != 0)
- return (EINVAL);
-
- if (zc->zc_sendobj == 0) {
- return (dsl_dataset_user_hold(zc->zc_name, zc->zc_value,
- zc->zc_string, recursive, zc->zc_temphold,
- zc->zc_cleanup_fd));
- }
-
- if (recursive)
- return (EINVAL);
+ error = nvlist_lookup_nvlist(args, "holds", &holds);
+ if (error != 0)
+ return (SET_ERROR(EINVAL));
- error = spa_open(zc->zc_name, &spa, FTAG);
- if (error)
- return (error);
+ /* make sure the user didn't pass us any invalid (empty) tags */
+ for (pair = nvlist_next_nvpair(holds, NULL); pair != NULL;
+ pair = nvlist_next_nvpair(holds, pair)) {
+ char *htag;
- dp = spa_get_dsl(spa);
- rw_enter(&dp->dp_config_rwlock, RW_READER);
- error = dsl_dataset_hold_obj(dp, zc->zc_sendobj, FTAG, &ds);
- rw_exit(&dp->dp_config_rwlock);
- spa_close(spa, FTAG);
- if (error)
- return (error);
+ error = nvpair_value_string(pair, &htag);
+ if (error != 0)
+ return (SET_ERROR(error));
- /*
- * Until we have a hold on this snapshot, it's possible that
- * zc_sendobj could've been destroyed and reused as part
- * of a later txg. Make sure we're looking at the right object.
- */
- if (zc->zc_createtxg != ds->ds_phys->ds_creation_txg) {
- dsl_dataset_rele(ds, FTAG);
- return (ENOENT);
+ if (strlen(htag) == 0)
+ return (SET_ERROR(EINVAL));
}
- if (zc->zc_cleanup_fd != -1 && zc->zc_temphold) {
- error = zfs_onexit_fd_hold(zc->zc_cleanup_fd, &minor);
- if (error) {
- dsl_dataset_rele(ds, FTAG);
+ if (nvlist_lookup_int32(args, "cleanup_fd", &cleanup_fd) == 0) {
+ error = zfs_onexit_fd_hold(cleanup_fd, &minor);
+ if (error != 0)
return (error);
- }
}
- error = dsl_dataset_user_hold_for_send(ds, zc->zc_string,
- zc->zc_temphold);
- if (minor != 0) {
- if (error == 0) {
- dsl_register_onexit_hold_cleanup(ds, zc->zc_string,
- minor);
- }
- zfs_onexit_fd_rele(zc->zc_cleanup_fd);
- }
- dsl_dataset_rele(ds, FTAG);
-
+ error = dsl_dataset_user_hold(holds, minor, errlist);
+ if (minor != 0)
+ zfs_onexit_fd_rele(cleanup_fd);
return (error);
}
/*
- * inputs:
- * zc_name name of dataset from which we're releasing a user hold
- * zc_value short name of snap
- * zc_string user-supplied tag for this hold
- * zc_cookie recursive flag
+ * innvl is not used.
*
- * outputs: none
+ * outnvl: {
+ * holdname -> time added (uint64 seconds since epoch)
+ * ...
+ * }
*/
+/* ARGSUSED */
static int
-zfs_ioc_release(zfs_cmd_t *zc)
+zfs_ioc_get_holds(const char *snapname, nvlist_t *args, nvlist_t *outnvl)
{
- boolean_t recursive = zc->zc_cookie;
-
- if (snapshot_namecheck(zc->zc_value, NULL, NULL) != 0)
- return (EINVAL);
-
- return (dsl_dataset_user_release(zc->zc_name, zc->zc_value,
- zc->zc_string, recursive));
+ return (dsl_dataset_get_holds(snapname, outnvl));
}
/*
- * inputs:
- * zc_name name of filesystem
+ * innvl: {
+ * snapname -> { holdname, ... }
+ * ...
+ * }
*
- * outputs:
- * zc_nvlist_src{_size} nvlist of snapshot holds
+ * outnvl: {
+ * snapname -> error value (int32)
+ * ...
+ * }
*/
+/* ARGSUSED */
static int
-zfs_ioc_get_holds(zfs_cmd_t *zc)
+zfs_ioc_release(const char *pool, nvlist_t *holds, nvlist_t *errlist)
{
- nvlist_t *nvp;
- int error;
-
- if ((error = dsl_dataset_get_holds(zc->zc_name, &nvp)) == 0) {
- error = put_nvlist(zc, nvp);
- nvlist_free(nvp);
- }
-
- return (error);
+ return (dsl_dataset_user_release(holds, errlist));
}
/*
* inputs:
* zc_guid flags (ZEVENT_NONBLOCK)
+ * zc_cleanup_fd zevent file descriptor
*
* outputs:
* zc_nvlist_dst next nvlist event
* zc_cookie dropped events since last get
- * zc_cleanup_fd cleanup-on-exit file descriptor
*/
static int
zfs_ioc_events_next(zfs_cmd_t *zc)
break;
error = zfs_zevent_wait(ze);
- if (error)
+ if (error != 0)
break;
} while (1);
zfs_zevent_drain_all(&count);
zc->zc_cookie = count;
- return 0;
+ return (0);
}
/*
- * pool create, destroy, and export don't log the history as part of
- * zfsdev_ioctl, but rather zfs_ioc_pool_create, and zfs_ioc_pool_export
- * do the logging of those commands.
+ * inputs:
+ * zc_guid eid | ZEVENT_SEEK_START | ZEVENT_SEEK_END
+ * zc_cleanup zevent file descriptor
*/
-static zfs_ioc_vec_t zfs_ioc_vec[] = {
- { zfs_ioc_pool_create, zfs_secpolicy_config, POOL_NAME, B_FALSE,
- POOL_CHECK_NONE },
- { zfs_ioc_pool_destroy, zfs_secpolicy_config, POOL_NAME, B_FALSE,
- POOL_CHECK_NONE },
- { zfs_ioc_pool_import, zfs_secpolicy_config, POOL_NAME, B_TRUE,
- POOL_CHECK_NONE },
- { zfs_ioc_pool_export, zfs_secpolicy_config, POOL_NAME, B_FALSE,
- POOL_CHECK_NONE },
- { zfs_ioc_pool_configs, zfs_secpolicy_none, NO_NAME, B_FALSE,
- POOL_CHECK_NONE },
- { zfs_ioc_pool_stats, zfs_secpolicy_read, POOL_NAME, B_FALSE,
- POOL_CHECK_NONE },
- { zfs_ioc_pool_tryimport, zfs_secpolicy_config, NO_NAME, B_FALSE,
- POOL_CHECK_NONE },
- { zfs_ioc_pool_scan, zfs_secpolicy_config, POOL_NAME, B_TRUE,
- POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY },
- { zfs_ioc_pool_freeze, zfs_secpolicy_config, NO_NAME, B_FALSE,
- POOL_CHECK_READONLY },
- { zfs_ioc_pool_upgrade, zfs_secpolicy_config, POOL_NAME, B_TRUE,
- POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY },
- { zfs_ioc_pool_get_history, zfs_secpolicy_config, POOL_NAME, B_FALSE,
- POOL_CHECK_NONE },
- { zfs_ioc_vdev_add, zfs_secpolicy_config, POOL_NAME, B_TRUE,
- POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY },
- { zfs_ioc_vdev_remove, zfs_secpolicy_config, POOL_NAME, B_TRUE,
- POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY },
- { zfs_ioc_vdev_set_state, zfs_secpolicy_config, POOL_NAME, B_TRUE,
- POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY },
- { zfs_ioc_vdev_attach, zfs_secpolicy_config, POOL_NAME, B_TRUE,
- POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY },
- { zfs_ioc_vdev_detach, zfs_secpolicy_config, POOL_NAME, B_TRUE,
- POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY },
- { zfs_ioc_vdev_setpath, zfs_secpolicy_config, POOL_NAME, B_FALSE,
- POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY },
- { zfs_ioc_vdev_setfru, zfs_secpolicy_config, POOL_NAME, B_FALSE,
- POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY },
- { zfs_ioc_objset_stats, zfs_secpolicy_read, DATASET_NAME, B_FALSE,
- POOL_CHECK_SUSPENDED },
- { zfs_ioc_objset_zplprops, zfs_secpolicy_read, DATASET_NAME, B_FALSE,
- POOL_CHECK_NONE },
- { zfs_ioc_dataset_list_next, zfs_secpolicy_read, DATASET_NAME, B_FALSE,
- POOL_CHECK_SUSPENDED },
- { zfs_ioc_snapshot_list_next, zfs_secpolicy_read, DATASET_NAME, B_FALSE,
- POOL_CHECK_SUSPENDED },
- { zfs_ioc_set_prop, zfs_secpolicy_none, DATASET_NAME, B_TRUE,
- POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY },
- { zfs_ioc_create_minor, zfs_secpolicy_config, DATASET_NAME, B_FALSE,
- POOL_CHECK_NONE },
- { zfs_ioc_remove_minor, zfs_secpolicy_config, DATASET_NAME, B_FALSE,
- POOL_CHECK_NONE },
- { zfs_ioc_create, zfs_secpolicy_create, DATASET_NAME, B_TRUE,
- POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY },
- { zfs_ioc_destroy, zfs_secpolicy_destroy, DATASET_NAME, B_TRUE,
- POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY },
- { zfs_ioc_rollback, zfs_secpolicy_rollback, DATASET_NAME, B_TRUE,
- POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY },
- { zfs_ioc_rename, zfs_secpolicy_rename, DATASET_NAME, B_TRUE,
- POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY },
- { zfs_ioc_recv, zfs_secpolicy_receive, DATASET_NAME, B_TRUE,
- POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY },
- { zfs_ioc_send, zfs_secpolicy_send, DATASET_NAME, B_TRUE,
- POOL_CHECK_NONE },
- { zfs_ioc_inject_fault, zfs_secpolicy_inject, NO_NAME, B_FALSE,
- POOL_CHECK_NONE },
- { zfs_ioc_clear_fault, zfs_secpolicy_inject, NO_NAME, B_FALSE,
- POOL_CHECK_NONE },
- { zfs_ioc_inject_list_next, zfs_secpolicy_inject, NO_NAME, B_FALSE,
- POOL_CHECK_NONE },
- { zfs_ioc_error_log, zfs_secpolicy_inject, POOL_NAME, B_FALSE,
- POOL_CHECK_NONE },
- { zfs_ioc_clear, zfs_secpolicy_config, POOL_NAME, B_TRUE,
- POOL_CHECK_NONE },
- { zfs_ioc_promote, zfs_secpolicy_promote, DATASET_NAME, B_TRUE,
- POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY },
- { zfs_ioc_destroy_snaps, zfs_secpolicy_destroy_snaps, DATASET_NAME,
- B_TRUE, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY },
- { zfs_ioc_snapshot, zfs_secpolicy_snapshot, DATASET_NAME, B_TRUE,
- POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY },
- { zfs_ioc_dsobj_to_dsname, zfs_secpolicy_diff, POOL_NAME, B_FALSE,
- POOL_CHECK_NONE },
- { zfs_ioc_obj_to_path, zfs_secpolicy_diff, DATASET_NAME, B_FALSE,
- POOL_CHECK_SUSPENDED },
- { zfs_ioc_pool_set_props, zfs_secpolicy_config, POOL_NAME, B_TRUE,
- POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY },
- { zfs_ioc_pool_get_props, zfs_secpolicy_read, POOL_NAME, B_FALSE,
- POOL_CHECK_NONE },
- { zfs_ioc_set_fsacl, zfs_secpolicy_fsacl, DATASET_NAME, B_TRUE,
- POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY },
- { zfs_ioc_get_fsacl, zfs_secpolicy_read, DATASET_NAME, B_FALSE,
- POOL_CHECK_NONE },
- { zfs_ioc_share, zfs_secpolicy_share, DATASET_NAME, B_FALSE,
- POOL_CHECK_NONE },
- { zfs_ioc_inherit_prop, zfs_secpolicy_inherit, DATASET_NAME, B_TRUE,
- POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY },
- { zfs_ioc_smb_acl, zfs_secpolicy_smb_acl, DATASET_NAME, B_FALSE,
- POOL_CHECK_NONE },
- { zfs_ioc_userspace_one, zfs_secpolicy_userspace_one, DATASET_NAME,
- B_FALSE, POOL_CHECK_NONE },
- { zfs_ioc_userspace_many, zfs_secpolicy_userspace_many, DATASET_NAME,
- B_FALSE, POOL_CHECK_NONE },
- { zfs_ioc_userspace_upgrade, zfs_secpolicy_userspace_upgrade,
- DATASET_NAME, B_FALSE, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY },
- { zfs_ioc_hold, zfs_secpolicy_hold, DATASET_NAME, B_TRUE,
- POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY },
- { zfs_ioc_release, zfs_secpolicy_release, DATASET_NAME, B_TRUE,
- POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY },
- { zfs_ioc_get_holds, zfs_secpolicy_read, DATASET_NAME, B_FALSE,
- POOL_CHECK_SUSPENDED },
- { zfs_ioc_objset_recvd_props, zfs_secpolicy_read, DATASET_NAME, B_FALSE,
- POOL_CHECK_NONE },
- { zfs_ioc_vdev_split, zfs_secpolicy_config, POOL_NAME, B_TRUE,
- POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY },
- { zfs_ioc_next_obj, zfs_secpolicy_read, DATASET_NAME, B_FALSE,
- POOL_CHECK_NONE },
- { zfs_ioc_diff, zfs_secpolicy_diff, DATASET_NAME, B_FALSE,
- POOL_CHECK_NONE },
- { zfs_ioc_tmp_snapshot, zfs_secpolicy_tmp_snapshot, DATASET_NAME,
- B_FALSE, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY },
- { zfs_ioc_obj_to_stats, zfs_secpolicy_diff, DATASET_NAME, B_FALSE,
- POOL_CHECK_SUSPENDED },
- { zfs_ioc_events_next, zfs_secpolicy_config, NO_NAME, B_FALSE,
- POOL_CHECK_NONE },
- { zfs_ioc_events_clear, zfs_secpolicy_config, NO_NAME, B_FALSE,
- POOL_CHECK_NONE },
-};
+static int
+zfs_ioc_events_seek(zfs_cmd_t *zc)
+{
+ zfs_zevent_t *ze;
+ minor_t minor;
+ int error;
+
+ error = zfs_zevent_fd_hold(zc->zc_cleanup_fd, &minor, &ze);
+ if (error != 0)
+ return (error);
+
+ error = zfs_zevent_seek(ze, zc->zc_guid);
+ zfs_zevent_fd_rele(zc->zc_cleanup_fd);
+
+ return (error);
+}
+
+/*
+ * inputs:
+ * zc_name name of new filesystem or snapshot
+ * zc_value full name of old snapshot
+ *
+ * outputs:
+ * zc_cookie space in bytes
+ * zc_objset_type compressed space in bytes
+ * zc_perm_action uncompressed space in bytes
+ */
+static int
+zfs_ioc_space_written(zfs_cmd_t *zc)
+{
+ int error;
+ dsl_pool_t *dp;
+ dsl_dataset_t *new, *old;
+
+ error = dsl_pool_hold(zc->zc_name, FTAG, &dp);
+ if (error != 0)
+ return (error);
+ error = dsl_dataset_hold(dp, zc->zc_name, FTAG, &new);
+ if (error != 0) {
+ dsl_pool_rele(dp, FTAG);
+ return (error);
+ }
+ error = dsl_dataset_hold(dp, zc->zc_value, FTAG, &old);
+ if (error != 0) {
+ dsl_dataset_rele(new, FTAG);
+ dsl_pool_rele(dp, FTAG);
+ return (error);
+ }
+
+ error = dsl_dataset_space_written(old, new, &zc->zc_cookie,
+ &zc->zc_objset_type, &zc->zc_perm_action);
+ dsl_dataset_rele(old, FTAG);
+ dsl_dataset_rele(new, FTAG);
+ dsl_pool_rele(dp, FTAG);
+ return (error);
+}
+
+/*
+ * innvl: {
+ * "firstsnap" -> snapshot name
+ * }
+ *
+ * outnvl: {
+ * "used" -> space in bytes
+ * "compressed" -> compressed space in bytes
+ * "uncompressed" -> uncompressed space in bytes
+ * }
+ */
+static int
+zfs_ioc_space_snaps(const char *lastsnap, nvlist_t *innvl, nvlist_t *outnvl)
+{
+ int error;
+ dsl_pool_t *dp;
+ dsl_dataset_t *new, *old;
+ char *firstsnap;
+ uint64_t used, comp, uncomp;
+
+ if (nvlist_lookup_string(innvl, "firstsnap", &firstsnap) != 0)
+ return (SET_ERROR(EINVAL));
+
+ error = dsl_pool_hold(lastsnap, FTAG, &dp);
+ if (error != 0)
+ return (error);
+
+ error = dsl_dataset_hold(dp, lastsnap, FTAG, &new);
+ if (error == 0 && !new->ds_is_snapshot) {
+ dsl_dataset_rele(new, FTAG);
+ error = SET_ERROR(EINVAL);
+ }
+ if (error != 0) {
+ dsl_pool_rele(dp, FTAG);
+ return (error);
+ }
+ error = dsl_dataset_hold(dp, firstsnap, FTAG, &old);
+ if (error == 0 && !old->ds_is_snapshot) {
+ dsl_dataset_rele(old, FTAG);
+ error = SET_ERROR(EINVAL);
+ }
+ if (error != 0) {
+ dsl_dataset_rele(new, FTAG);
+ dsl_pool_rele(dp, FTAG);
+ return (error);
+ }
+
+ error = dsl_dataset_space_wouldfree(old, new, &used, &comp, &uncomp);
+ dsl_dataset_rele(old, FTAG);
+ dsl_dataset_rele(new, FTAG);
+ dsl_pool_rele(dp, FTAG);
+ fnvlist_add_uint64(outnvl, "used", used);
+ fnvlist_add_uint64(outnvl, "compressed", comp);
+ fnvlist_add_uint64(outnvl, "uncompressed", uncomp);
+ return (error);
+}
+
+/*
+ * innvl: {
+ * "fd" -> file descriptor to write stream to (int32)
+ * (optional) "fromsnap" -> full snap name to send an incremental from
+ * (optional) "largeblockok" -> (value ignored)
+ * indicates that blocks > 128KB are permitted
+ * (optional) "embedok" -> (value ignored)
+ * presence indicates DRR_WRITE_EMBEDDED records are permitted
+ * }
+ *
+ * outnvl is unused
+ */
+/* ARGSUSED */
+static int
+zfs_ioc_send_new(const char *snapname, nvlist_t *innvl, nvlist_t *outnvl)
+{
+ int error;
+ offset_t off;
+ char *fromname = NULL;
+ int fd;
+ file_t *fp;
+ boolean_t largeblockok;
+ boolean_t embedok;
+
+ error = nvlist_lookup_int32(innvl, "fd", &fd);
+ if (error != 0)
+ return (SET_ERROR(EINVAL));
+
+ (void) nvlist_lookup_string(innvl, "fromsnap", &fromname);
+
+ largeblockok = nvlist_exists(innvl, "largeblockok");
+ embedok = nvlist_exists(innvl, "embedok");
+
+ if ((fp = getf(fd)) == NULL)
+ return (SET_ERROR(EBADF));
+
+ off = fp->f_offset;
+ error = dmu_send(snapname, fromname, embedok, largeblockok,
+ fd, fp->f_vnode, &off);
+
+ if (VOP_SEEK(fp->f_vnode, fp->f_offset, &off, NULL) == 0)
+ fp->f_offset = off;
+
+ releasef(fd);
+ return (error);
+}
+
+/*
+ * Determine approximately how large a zfs send stream will be -- the number
+ * of bytes that will be written to the fd supplied to zfs_ioc_send_new().
+ *
+ * innvl: {
+ * (optional) "from" -> full snap or bookmark name to send an incremental
+ * from
+ * }
+ *
+ * outnvl: {
+ * "space" -> bytes of space (uint64)
+ * }
+ */
+static int
+zfs_ioc_send_space(const char *snapname, nvlist_t *innvl, nvlist_t *outnvl)
+{
+ dsl_pool_t *dp;
+ dsl_dataset_t *tosnap;
+ int error;
+ char *fromname;
+ uint64_t space;
+
+ error = dsl_pool_hold(snapname, FTAG, &dp);
+ if (error != 0)
+ return (error);
+
+ error = dsl_dataset_hold(dp, snapname, FTAG, &tosnap);
+ if (error != 0) {
+ dsl_pool_rele(dp, FTAG);
+ return (error);
+ }
+
+ error = nvlist_lookup_string(innvl, "from", &fromname);
+ if (error == 0) {
+ if (strchr(fromname, '@') != NULL) {
+ /*
+ * If from is a snapshot, hold it and use the more
+ * efficient dmu_send_estimate to estimate send space
+ * size using deadlists.
+ */
+ dsl_dataset_t *fromsnap;
+ error = dsl_dataset_hold(dp, fromname, FTAG, &fromsnap);
+ if (error != 0)
+ goto out;
+ error = dmu_send_estimate(tosnap, fromsnap, &space);
+ dsl_dataset_rele(fromsnap, FTAG);
+ } else if (strchr(fromname, '#') != NULL) {
+ /*
+ * If from is a bookmark, fetch the creation TXG of the
+ * snapshot it was created from and use that to find
+ * blocks that were born after it.
+ */
+ zfs_bookmark_phys_t frombm;
+
+ error = dsl_bookmark_lookup(dp, fromname, tosnap,
+ &frombm);
+ if (error != 0)
+ goto out;
+ error = dmu_send_estimate_from_txg(tosnap,
+ frombm.zbm_creation_txg, &space);
+ } else {
+ /*
+ * from is not properly formatted as a snapshot or
+ * bookmark
+ */
+ error = SET_ERROR(EINVAL);
+ goto out;
+ }
+ } else {
+ // If estimating the size of a full send, use dmu_send_estimate
+ error = dmu_send_estimate(tosnap, NULL, &space);
+ }
+
+ fnvlist_add_uint64(outnvl, "space", space);
+
+out:
+ dsl_dataset_rele(tosnap, FTAG);
+ dsl_pool_rele(dp, FTAG);
+ return (error);
+}
+
+static zfs_ioc_vec_t zfs_ioc_vec[ZFS_IOC_LAST - ZFS_IOC_FIRST];
+
+static void
+zfs_ioctl_register_legacy(zfs_ioc_t ioc, zfs_ioc_legacy_func_t *func,
+ zfs_secpolicy_func_t *secpolicy, zfs_ioc_namecheck_t namecheck,
+ boolean_t log_history, zfs_ioc_poolcheck_t pool_check)
+{
+ zfs_ioc_vec_t *vec = &zfs_ioc_vec[ioc - ZFS_IOC_FIRST];
+
+ ASSERT3U(ioc, >=, ZFS_IOC_FIRST);
+ ASSERT3U(ioc, <, ZFS_IOC_LAST);
+ ASSERT3P(vec->zvec_legacy_func, ==, NULL);
+ ASSERT3P(vec->zvec_func, ==, NULL);
+
+ vec->zvec_legacy_func = func;
+ vec->zvec_secpolicy = secpolicy;
+ vec->zvec_namecheck = namecheck;
+ vec->zvec_allow_log = log_history;
+ vec->zvec_pool_check = pool_check;
+}
+
+/*
+ * See the block comment at the beginning of this file for details on
+ * each argument to this function.
+ */
+static void
+zfs_ioctl_register(const char *name, zfs_ioc_t ioc, zfs_ioc_func_t *func,
+ zfs_secpolicy_func_t *secpolicy, zfs_ioc_namecheck_t namecheck,
+ zfs_ioc_poolcheck_t pool_check, boolean_t smush_outnvlist,
+ boolean_t allow_log)
+{
+ zfs_ioc_vec_t *vec = &zfs_ioc_vec[ioc - ZFS_IOC_FIRST];
+
+ ASSERT3U(ioc, >=, ZFS_IOC_FIRST);
+ ASSERT3U(ioc, <, ZFS_IOC_LAST);
+ ASSERT3P(vec->zvec_legacy_func, ==, NULL);
+ ASSERT3P(vec->zvec_func, ==, NULL);
+
+ /* if we are logging, the name must be valid */
+ ASSERT(!allow_log || namecheck != NO_NAME);
+
+ vec->zvec_name = name;
+ vec->zvec_func = func;
+ vec->zvec_secpolicy = secpolicy;
+ vec->zvec_namecheck = namecheck;
+ vec->zvec_pool_check = pool_check;
+ vec->zvec_smush_outnvlist = smush_outnvlist;
+ vec->zvec_allow_log = allow_log;
+}
+
+static void
+zfs_ioctl_register_pool(zfs_ioc_t ioc, zfs_ioc_legacy_func_t *func,
+ zfs_secpolicy_func_t *secpolicy, boolean_t log_history,
+ zfs_ioc_poolcheck_t pool_check)
+{
+ zfs_ioctl_register_legacy(ioc, func, secpolicy,
+ POOL_NAME, log_history, pool_check);
+}
+
+static void
+zfs_ioctl_register_dataset_nolog(zfs_ioc_t ioc, zfs_ioc_legacy_func_t *func,
+ zfs_secpolicy_func_t *secpolicy, zfs_ioc_poolcheck_t pool_check)
+{
+ zfs_ioctl_register_legacy(ioc, func, secpolicy,
+ DATASET_NAME, B_FALSE, pool_check);
+}
+
+static void
+zfs_ioctl_register_pool_modify(zfs_ioc_t ioc, zfs_ioc_legacy_func_t *func)
+{
+ zfs_ioctl_register_legacy(ioc, func, zfs_secpolicy_config,
+ POOL_NAME, B_TRUE, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY);
+}
+
+static void
+zfs_ioctl_register_pool_meta(zfs_ioc_t ioc, zfs_ioc_legacy_func_t *func,
+ zfs_secpolicy_func_t *secpolicy)
+{
+ zfs_ioctl_register_legacy(ioc, func, secpolicy,
+ NO_NAME, B_FALSE, POOL_CHECK_NONE);
+}
+
+static void
+zfs_ioctl_register_dataset_read_secpolicy(zfs_ioc_t ioc,
+ zfs_ioc_legacy_func_t *func, zfs_secpolicy_func_t *secpolicy)
+{
+ zfs_ioctl_register_legacy(ioc, func, secpolicy,
+ DATASET_NAME, B_FALSE, POOL_CHECK_SUSPENDED);
+}
+
+static void
+zfs_ioctl_register_dataset_read(zfs_ioc_t ioc, zfs_ioc_legacy_func_t *func)
+{
+ zfs_ioctl_register_dataset_read_secpolicy(ioc, func,
+ zfs_secpolicy_read);
+}
+
+static void
+zfs_ioctl_register_dataset_modify(zfs_ioc_t ioc, zfs_ioc_legacy_func_t *func,
+ zfs_secpolicy_func_t *secpolicy)
+{
+ zfs_ioctl_register_legacy(ioc, func, secpolicy,
+ DATASET_NAME, B_TRUE, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY);
+}
+
+static void
+zfs_ioctl_init(void)
+{
+ zfs_ioctl_register("snapshot", ZFS_IOC_SNAPSHOT,
+ zfs_ioc_snapshot, zfs_secpolicy_snapshot, POOL_NAME,
+ POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_TRUE, B_TRUE);
+
+ zfs_ioctl_register("log_history", ZFS_IOC_LOG_HISTORY,
+ zfs_ioc_log_history, zfs_secpolicy_log_history, NO_NAME,
+ POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_FALSE, B_FALSE);
+
+ zfs_ioctl_register("space_snaps", ZFS_IOC_SPACE_SNAPS,
+ zfs_ioc_space_snaps, zfs_secpolicy_read, DATASET_NAME,
+ POOL_CHECK_SUSPENDED, B_FALSE, B_FALSE);
+
+ zfs_ioctl_register("send", ZFS_IOC_SEND_NEW,
+ zfs_ioc_send_new, zfs_secpolicy_send_new, DATASET_NAME,
+ POOL_CHECK_SUSPENDED, B_FALSE, B_FALSE);
+
+ zfs_ioctl_register("send_space", ZFS_IOC_SEND_SPACE,
+ zfs_ioc_send_space, zfs_secpolicy_read, DATASET_NAME,
+ POOL_CHECK_SUSPENDED, B_FALSE, B_FALSE);
+
+ zfs_ioctl_register("create", ZFS_IOC_CREATE,
+ zfs_ioc_create, zfs_secpolicy_create_clone, DATASET_NAME,
+ POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_TRUE, B_TRUE);
+
+ zfs_ioctl_register("clone", ZFS_IOC_CLONE,
+ zfs_ioc_clone, zfs_secpolicy_create_clone, DATASET_NAME,
+ POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_TRUE, B_TRUE);
+
+ zfs_ioctl_register("destroy_snaps", ZFS_IOC_DESTROY_SNAPS,
+ zfs_ioc_destroy_snaps, zfs_secpolicy_destroy_snaps, POOL_NAME,
+ POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_TRUE, B_TRUE);
+
+ zfs_ioctl_register("hold", ZFS_IOC_HOLD,
+ zfs_ioc_hold, zfs_secpolicy_hold, POOL_NAME,
+ POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_TRUE, B_TRUE);
+ zfs_ioctl_register("release", ZFS_IOC_RELEASE,
+ zfs_ioc_release, zfs_secpolicy_release, POOL_NAME,
+ POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_TRUE, B_TRUE);
+
+ zfs_ioctl_register("get_holds", ZFS_IOC_GET_HOLDS,
+ zfs_ioc_get_holds, zfs_secpolicy_read, DATASET_NAME,
+ POOL_CHECK_SUSPENDED, B_FALSE, B_FALSE);
+
+ zfs_ioctl_register("rollback", ZFS_IOC_ROLLBACK,
+ zfs_ioc_rollback, zfs_secpolicy_rollback, DATASET_NAME,
+ POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_FALSE, B_TRUE);
+
+ zfs_ioctl_register("bookmark", ZFS_IOC_BOOKMARK,
+ zfs_ioc_bookmark, zfs_secpolicy_bookmark, POOL_NAME,
+ POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_TRUE, B_TRUE);
+
+ zfs_ioctl_register("get_bookmarks", ZFS_IOC_GET_BOOKMARKS,
+ zfs_ioc_get_bookmarks, zfs_secpolicy_read, DATASET_NAME,
+ POOL_CHECK_SUSPENDED, B_FALSE, B_FALSE);
+
+ zfs_ioctl_register("destroy_bookmarks", ZFS_IOC_DESTROY_BOOKMARKS,
+ zfs_ioc_destroy_bookmarks, zfs_secpolicy_destroy_bookmarks,
+ POOL_NAME,
+ POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_TRUE, B_TRUE);
+
+ /* IOCTLS that use the legacy function signature */
+
+ zfs_ioctl_register_legacy(ZFS_IOC_POOL_FREEZE, zfs_ioc_pool_freeze,
+ zfs_secpolicy_config, NO_NAME, B_FALSE, POOL_CHECK_READONLY);
+
+ zfs_ioctl_register_pool(ZFS_IOC_POOL_CREATE, zfs_ioc_pool_create,
+ zfs_secpolicy_config, B_TRUE, POOL_CHECK_NONE);
+ zfs_ioctl_register_pool_modify(ZFS_IOC_POOL_SCAN,
+ zfs_ioc_pool_scan);
+ zfs_ioctl_register_pool_modify(ZFS_IOC_POOL_UPGRADE,
+ zfs_ioc_pool_upgrade);
+ zfs_ioctl_register_pool_modify(ZFS_IOC_VDEV_ADD,
+ zfs_ioc_vdev_add);
+ zfs_ioctl_register_pool_modify(ZFS_IOC_VDEV_REMOVE,
+ zfs_ioc_vdev_remove);
+ zfs_ioctl_register_pool_modify(ZFS_IOC_VDEV_SET_STATE,
+ zfs_ioc_vdev_set_state);
+ zfs_ioctl_register_pool_modify(ZFS_IOC_VDEV_ATTACH,
+ zfs_ioc_vdev_attach);
+ zfs_ioctl_register_pool_modify(ZFS_IOC_VDEV_DETACH,
+ zfs_ioc_vdev_detach);
+ zfs_ioctl_register_pool_modify(ZFS_IOC_VDEV_SETPATH,
+ zfs_ioc_vdev_setpath);
+ zfs_ioctl_register_pool_modify(ZFS_IOC_VDEV_SETFRU,
+ zfs_ioc_vdev_setfru);
+ zfs_ioctl_register_pool_modify(ZFS_IOC_POOL_SET_PROPS,
+ zfs_ioc_pool_set_props);
+ zfs_ioctl_register_pool_modify(ZFS_IOC_VDEV_SPLIT,
+ zfs_ioc_vdev_split);
+ zfs_ioctl_register_pool_modify(ZFS_IOC_POOL_REGUID,
+ zfs_ioc_pool_reguid);
+
+ zfs_ioctl_register_pool_meta(ZFS_IOC_POOL_CONFIGS,
+ zfs_ioc_pool_configs, zfs_secpolicy_none);
+ zfs_ioctl_register_pool_meta(ZFS_IOC_POOL_TRYIMPORT,
+ zfs_ioc_pool_tryimport, zfs_secpolicy_config);
+ zfs_ioctl_register_pool_meta(ZFS_IOC_INJECT_FAULT,
+ zfs_ioc_inject_fault, zfs_secpolicy_inject);
+ zfs_ioctl_register_pool_meta(ZFS_IOC_CLEAR_FAULT,
+ zfs_ioc_clear_fault, zfs_secpolicy_inject);
+ zfs_ioctl_register_pool_meta(ZFS_IOC_INJECT_LIST_NEXT,
+ zfs_ioc_inject_list_next, zfs_secpolicy_inject);
+
+ /*
+ * pool destroy, and export don't log the history as part of
+ * zfsdev_ioctl, but rather zfs_ioc_pool_export
+ * does the logging of those commands.
+ */
+ zfs_ioctl_register_pool(ZFS_IOC_POOL_DESTROY, zfs_ioc_pool_destroy,
+ zfs_secpolicy_config, B_FALSE, POOL_CHECK_SUSPENDED);
+ zfs_ioctl_register_pool(ZFS_IOC_POOL_EXPORT, zfs_ioc_pool_export,
+ zfs_secpolicy_config, B_FALSE, POOL_CHECK_SUSPENDED);
+
+ zfs_ioctl_register_pool(ZFS_IOC_POOL_STATS, zfs_ioc_pool_stats,
+ zfs_secpolicy_read, B_FALSE, POOL_CHECK_NONE);
+ zfs_ioctl_register_pool(ZFS_IOC_POOL_GET_PROPS, zfs_ioc_pool_get_props,
+ zfs_secpolicy_read, B_FALSE, POOL_CHECK_NONE);
+
+ zfs_ioctl_register_pool(ZFS_IOC_ERROR_LOG, zfs_ioc_error_log,
+ zfs_secpolicy_inject, B_FALSE, POOL_CHECK_SUSPENDED);
+ zfs_ioctl_register_pool(ZFS_IOC_DSOBJ_TO_DSNAME,
+ zfs_ioc_dsobj_to_dsname,
+ zfs_secpolicy_diff, B_FALSE, POOL_CHECK_SUSPENDED);
+ zfs_ioctl_register_pool(ZFS_IOC_POOL_GET_HISTORY,
+ zfs_ioc_pool_get_history,
+ zfs_secpolicy_config, B_FALSE, POOL_CHECK_SUSPENDED);
+
+ zfs_ioctl_register_pool(ZFS_IOC_POOL_IMPORT, zfs_ioc_pool_import,
+ zfs_secpolicy_config, B_TRUE, POOL_CHECK_NONE);
+
+ zfs_ioctl_register_pool(ZFS_IOC_CLEAR, zfs_ioc_clear,
+ zfs_secpolicy_config, B_TRUE, POOL_CHECK_NONE);
+ zfs_ioctl_register_pool(ZFS_IOC_POOL_REOPEN, zfs_ioc_pool_reopen,
+ zfs_secpolicy_config, B_TRUE, POOL_CHECK_SUSPENDED);
+
+ zfs_ioctl_register_dataset_read(ZFS_IOC_SPACE_WRITTEN,
+ zfs_ioc_space_written);
+ zfs_ioctl_register_dataset_read(ZFS_IOC_OBJSET_RECVD_PROPS,
+ zfs_ioc_objset_recvd_props);
+ zfs_ioctl_register_dataset_read(ZFS_IOC_NEXT_OBJ,
+ zfs_ioc_next_obj);
+ zfs_ioctl_register_dataset_read(ZFS_IOC_GET_FSACL,
+ zfs_ioc_get_fsacl);
+ zfs_ioctl_register_dataset_read(ZFS_IOC_OBJSET_STATS,
+ zfs_ioc_objset_stats);
+ zfs_ioctl_register_dataset_read(ZFS_IOC_OBJSET_ZPLPROPS,
+ zfs_ioc_objset_zplprops);
+ zfs_ioctl_register_dataset_read(ZFS_IOC_DATASET_LIST_NEXT,
+ zfs_ioc_dataset_list_next);
+ zfs_ioctl_register_dataset_read(ZFS_IOC_SNAPSHOT_LIST_NEXT,
+ zfs_ioc_snapshot_list_next);
+ zfs_ioctl_register_dataset_read(ZFS_IOC_SEND_PROGRESS,
+ zfs_ioc_send_progress);
+
+ zfs_ioctl_register_dataset_read_secpolicy(ZFS_IOC_DIFF,
+ zfs_ioc_diff, zfs_secpolicy_diff);
+ zfs_ioctl_register_dataset_read_secpolicy(ZFS_IOC_OBJ_TO_STATS,
+ zfs_ioc_obj_to_stats, zfs_secpolicy_diff);
+ zfs_ioctl_register_dataset_read_secpolicy(ZFS_IOC_OBJ_TO_PATH,
+ zfs_ioc_obj_to_path, zfs_secpolicy_diff);
+ zfs_ioctl_register_dataset_read_secpolicy(ZFS_IOC_USERSPACE_ONE,
+ zfs_ioc_userspace_one, zfs_secpolicy_userspace_one);
+ zfs_ioctl_register_dataset_read_secpolicy(ZFS_IOC_USERSPACE_MANY,
+ zfs_ioc_userspace_many, zfs_secpolicy_userspace_many);
+ zfs_ioctl_register_dataset_read_secpolicy(ZFS_IOC_SEND,
+ zfs_ioc_send, zfs_secpolicy_send);
+
+ zfs_ioctl_register_dataset_modify(ZFS_IOC_SET_PROP, zfs_ioc_set_prop,
+ zfs_secpolicy_none);
+ zfs_ioctl_register_dataset_modify(ZFS_IOC_DESTROY, zfs_ioc_destroy,
+ zfs_secpolicy_destroy);
+ zfs_ioctl_register_dataset_modify(ZFS_IOC_RENAME, zfs_ioc_rename,
+ zfs_secpolicy_rename);
+ zfs_ioctl_register_dataset_modify(ZFS_IOC_RECV, zfs_ioc_recv,
+ zfs_secpolicy_recv);
+ zfs_ioctl_register_dataset_modify(ZFS_IOC_PROMOTE, zfs_ioc_promote,
+ zfs_secpolicy_promote);
+ zfs_ioctl_register_dataset_modify(ZFS_IOC_INHERIT_PROP,
+ zfs_ioc_inherit_prop, zfs_secpolicy_inherit_prop);
+ zfs_ioctl_register_dataset_modify(ZFS_IOC_SET_FSACL, zfs_ioc_set_fsacl,
+ zfs_secpolicy_set_fsacl);
+
+ zfs_ioctl_register_dataset_nolog(ZFS_IOC_SHARE, zfs_ioc_share,
+ zfs_secpolicy_share, POOL_CHECK_NONE);
+ zfs_ioctl_register_dataset_nolog(ZFS_IOC_SMB_ACL, zfs_ioc_smb_acl,
+ zfs_secpolicy_smb_acl, POOL_CHECK_NONE);
+ zfs_ioctl_register_dataset_nolog(ZFS_IOC_USERSPACE_UPGRADE,
+ zfs_ioc_userspace_upgrade, zfs_secpolicy_userspace_upgrade,
+ POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY);
+ zfs_ioctl_register_dataset_nolog(ZFS_IOC_TMP_SNAPSHOT,
+ zfs_ioc_tmp_snapshot, zfs_secpolicy_tmp_snapshot,
+ POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY);
+
+ /*
+ * ZoL functions
+ */
+ zfs_ioctl_register_legacy(ZFS_IOC_EVENTS_NEXT, zfs_ioc_events_next,
+ zfs_secpolicy_config, NO_NAME, B_FALSE, POOL_CHECK_NONE);
+ zfs_ioctl_register_legacy(ZFS_IOC_EVENTS_CLEAR, zfs_ioc_events_clear,
+ zfs_secpolicy_config, NO_NAME, B_FALSE, POOL_CHECK_NONE);
+ zfs_ioctl_register_legacy(ZFS_IOC_EVENTS_SEEK, zfs_ioc_events_seek,
+ zfs_secpolicy_config, NO_NAME, B_FALSE, POOL_CHECK_NONE);
+}
int
pool_status_check(const char *name, zfs_ioc_namecheck_t type,
error = spa_open(name, &spa, FTAG);
if (error == 0) {
if ((check & POOL_CHECK_SUSPENDED) && spa_suspended(spa))
- error = EAGAIN;
+ error = SET_ERROR(EAGAIN);
else if ((check & POOL_CHECK_READONLY) && !spa_writeable(spa))
- error = EROFS;
+ error = SET_ERROR(EROFS);
spa_close(spa, FTAG);
}
return (error);
{
zfsdev_state_t *zs;
- ASSERT(MUTEX_HELD(&zfsdev_state_lock));
-
- for (zs = list_head(&zfsdev_state_list); zs != NULL;
- zs = list_next(&zfsdev_state_list, zs)) {
+ for (zs = zfsdev_state_list; zs != NULL; zs = zs->zs_next) {
if (zs->zs_minor == minor) {
+ smp_rmb();
switch (which) {
- case ZST_ONEXIT: return (zs->zs_onexit);
- case ZST_ZEVENT: return (zs->zs_zevent);
- case ZST_ALL: return (zs);
+ case ZST_ONEXIT:
+ return (zs->zs_onexit);
+ case ZST_ZEVENT:
+ return (zs->zs_zevent);
+ case ZST_ALL:
+ return (zs);
}
}
}
- return NULL;
+ return (NULL);
}
void *
{
void *ptr;
- mutex_enter(&zfsdev_state_lock);
ptr = zfsdev_get_state_impl(minor, which);
- mutex_exit(&zfsdev_state_lock);
- return ptr;
+ return (ptr);
}
-minor_t
-zfsdev_getminor(struct file *filp)
+int
+zfsdev_getminor(struct file *filp, minor_t *minorp)
{
+ zfsdev_state_t *zs, *fpd;
+
ASSERT(filp != NULL);
- ASSERT(filp->private_data != NULL);
+ ASSERT(!MUTEX_HELD(&zfsdev_state_lock));
- return (((zfsdev_state_t *)filp->private_data)->zs_minor);
+ fpd = filp->private_data;
+ if (fpd == NULL)
+ return (EBADF);
+
+ mutex_enter(&zfsdev_state_lock);
+
+ for (zs = zfsdev_state_list; zs != NULL; zs = zs->zs_next) {
+
+ if (zs->zs_minor == -1)
+ continue;
+
+ if (fpd == zs) {
+ *minorp = fpd->zs_minor;
+ mutex_exit(&zfsdev_state_lock);
+ return (0);
+ }
+ }
+
+ mutex_exit(&zfsdev_state_lock);
+
+ return (EBADF);
}
/*
static int
zfsdev_state_init(struct file *filp)
{
- zfsdev_state_t *zs;
+ zfsdev_state_t *zs, *zsprev = NULL;
minor_t minor;
+ boolean_t newzs = B_FALSE;
ASSERT(MUTEX_HELD(&zfsdev_state_lock));
- minor = zfsdev_minor_alloc();
- if (minor == 0)
- return (ENXIO);
+ minor = zfsdev_minor_alloc();
+ if (minor == 0)
+ return (SET_ERROR(ENXIO));
- zs = kmem_zalloc( sizeof(zfsdev_state_t), KM_SLEEP);
- if (zs == NULL)
- return (ENOMEM);
+ for (zs = zfsdev_state_list; zs != NULL; zs = zs->zs_next) {
+ if (zs->zs_minor == -1)
+ break;
+ zsprev = zs;
+ }
+
+ if (!zs) {
+ zs = kmem_zalloc(sizeof (zfsdev_state_t), KM_SLEEP);
+ newzs = B_TRUE;
+ }
zs->zs_file = filp;
- zs->zs_minor = minor;
filp->private_data = zs;
zfs_onexit_init((zfs_onexit_t **)&zs->zs_onexit);
zfs_zevent_init((zfs_zevent_t **)&zs->zs_zevent);
- list_insert_tail(&zfsdev_state_list, zs);
+
+ /*
+ * In order to provide for lock-free concurrent read access
+ * to the minor list in zfsdev_get_state_impl(), new entries
+ * must be completely written before linking them into the
+ * list whereas existing entries are already linked; the last
+ * operation must be updating zs_minor (from -1 to the new
+ * value).
+ */
+ if (newzs) {
+ zs->zs_minor = minor;
+ smp_wmb();
+ zsprev->zs_next = zs;
+ } else {
+ smp_wmb();
+ zs->zs_minor = minor;
+ }
return (0);
}
ASSERT(filp->private_data != NULL);
zs = filp->private_data;
+ zs->zs_minor = -1;
zfs_onexit_destroy(zs->zs_onexit);
zfs_zevent_destroy(zs->zs_zevent);
- list_remove(&zfsdev_state_list, zs);
- kmem_free(zs, sizeof(zfsdev_state_t));
-
- return 0;
+ return (0);
}
static int
zfsdev_ioctl(struct file *filp, unsigned cmd, unsigned long arg)
{
zfs_cmd_t *zc;
- uint_t vec;
+ uint_t vecnum;
int error, rc, flag = 0;
+ const zfs_ioc_vec_t *vec;
+ char *saved_poolname = NULL;
+ nvlist_t *innvl = NULL;
+ fstrans_cookie_t cookie;
+
+ vecnum = cmd - ZFS_IOC_FIRST;
+ if (vecnum >= sizeof (zfs_ioc_vec) / sizeof (zfs_ioc_vec[0]))
+ return (-SET_ERROR(EINVAL));
+ vec = &zfs_ioc_vec[vecnum];
- vec = cmd - ZFS_IOC;
- if (vec >= sizeof (zfs_ioc_vec) / sizeof (zfs_ioc_vec[0]))
- return (-EINVAL);
+ /*
+ * The registered ioctl list may be sparse, verify that either
+ * a normal or legacy handler are registered.
+ */
+ if (vec->zvec_func == NULL && vec->zvec_legacy_func == NULL)
+ return (-SET_ERROR(EINVAL));
- zc = kmem_zalloc(sizeof (zfs_cmd_t), KM_SLEEP | KM_NODEBUG);
+ zc = kmem_zalloc(sizeof (zfs_cmd_t), KM_SLEEP);
error = ddi_copyin((void *)arg, zc, sizeof (zfs_cmd_t), flag);
- if (error != 0)
- error = EFAULT;
+ if (error != 0) {
+ error = SET_ERROR(EFAULT);
+ goto out;
+ }
- if ((error == 0) && !(flag & FKIOCTL))
- error = zfs_ioc_vec[vec].zvec_secpolicy(zc, CRED());
+ zc->zc_iflags = flag & FKIOCTL;
+ if (zc->zc_nvlist_src_size > MAX_NVLIST_SRC_SIZE) {
+ /*
+ * Make sure the user doesn't pass in an insane value for
+ * zc_nvlist_src_size. We have to check, since we will end
+ * up allocating that much memory inside of get_nvlist(). This
+ * prevents a nefarious user from allocating tons of kernel
+ * memory.
+ *
+ * Also, we return EINVAL instead of ENOMEM here. The reason
+ * being that returning ENOMEM from an ioctl() has a special
+ * connotation; that the user's size value is too small and
+ * needs to be expanded to hold the nvlist. See
+ * zcmd_expand_dst_nvlist() for details.
+ */
+ error = SET_ERROR(EINVAL); /* User's size too big */
+
+ } else if (zc->zc_nvlist_src_size != 0) {
+ error = get_nvlist(zc->zc_nvlist_src, zc->zc_nvlist_src_size,
+ zc->zc_iflags, &innvl);
+ if (error != 0)
+ goto out;
+ }
/*
* Ensure that all pool/dataset names are valid before we pass down to
* the lower layers.
*/
- if (error == 0) {
- zc->zc_name[sizeof (zc->zc_name) - 1] = '\0';
- zc->zc_iflags = flag & FKIOCTL;
- switch (zfs_ioc_vec[vec].zvec_namecheck) {
- case POOL_NAME:
- if (pool_namecheck(zc->zc_name, NULL, NULL) != 0)
- error = EINVAL;
+ zc->zc_name[sizeof (zc->zc_name) - 1] = '\0';
+ switch (vec->zvec_namecheck) {
+ case POOL_NAME:
+ if (pool_namecheck(zc->zc_name, NULL, NULL) != 0)
+ error = SET_ERROR(EINVAL);
+ else
error = pool_status_check(zc->zc_name,
- zfs_ioc_vec[vec].zvec_namecheck,
- zfs_ioc_vec[vec].zvec_pool_check);
- break;
+ vec->zvec_namecheck, vec->zvec_pool_check);
+ break;
- case DATASET_NAME:
- if (dataset_namecheck(zc->zc_name, NULL, NULL) != 0)
- error = EINVAL;
+ case DATASET_NAME:
+ if (dataset_namecheck(zc->zc_name, NULL, NULL) != 0)
+ error = SET_ERROR(EINVAL);
+ else
error = pool_status_check(zc->zc_name,
- zfs_ioc_vec[vec].zvec_namecheck,
- zfs_ioc_vec[vec].zvec_pool_check);
- break;
+ vec->zvec_namecheck, vec->zvec_pool_check);
+ break;
- case NO_NAME:
- break;
- }
+ case NO_NAME:
+ break;
}
- if (error == 0)
- error = zfs_ioc_vec[vec].zvec_func(zc);
+ if (error == 0 && !(flag & FKIOCTL)) {
+ cookie = spl_fstrans_mark();
+ error = vec->zvec_secpolicy(zc, innvl, CRED());
+ spl_fstrans_unmark(cookie);
+ }
+
+ if (error != 0)
+ goto out;
+
+ /* legacy ioctls can modify zc_name */
+ saved_poolname = strdup(zc->zc_name);
+ if (saved_poolname == NULL) {
+ error = SET_ERROR(ENOMEM);
+ goto out;
+ } else {
+ saved_poolname[strcspn(saved_poolname, "/@#")] = '\0';
+ }
+
+ if (vec->zvec_func != NULL) {
+ nvlist_t *outnvl;
+ int puterror = 0;
+ spa_t *spa;
+ nvlist_t *lognv = NULL;
+
+ ASSERT(vec->zvec_legacy_func == NULL);
+
+ /*
+ * Add the innvl to the lognv before calling the func,
+ * in case the func changes the innvl.
+ */
+ if (vec->zvec_allow_log) {
+ lognv = fnvlist_alloc();
+ fnvlist_add_string(lognv, ZPOOL_HIST_IOCTL,
+ vec->zvec_name);
+ if (!nvlist_empty(innvl)) {
+ fnvlist_add_nvlist(lognv, ZPOOL_HIST_INPUT_NVL,
+ innvl);
+ }
+ }
+
+ outnvl = fnvlist_alloc();
+ cookie = spl_fstrans_mark();
+ error = vec->zvec_func(zc->zc_name, innvl, outnvl);
+ spl_fstrans_unmark(cookie);
+
+ if (error == 0 && vec->zvec_allow_log &&
+ spa_open(zc->zc_name, &spa, FTAG) == 0) {
+ if (!nvlist_empty(outnvl)) {
+ fnvlist_add_nvlist(lognv, ZPOOL_HIST_OUTPUT_NVL,
+ outnvl);
+ }
+ (void) spa_history_log_nvl(spa, lognv);
+ spa_close(spa, FTAG);
+ }
+ fnvlist_free(lognv);
+
+ if (!nvlist_empty(outnvl) || zc->zc_nvlist_dst_size != 0) {
+ int smusherror = 0;
+ if (vec->zvec_smush_outnvlist) {
+ smusherror = nvlist_smush(outnvl,
+ zc->zc_nvlist_dst_size);
+ }
+ if (smusherror == 0)
+ puterror = put_nvlist(zc, outnvl);
+ }
+
+ if (puterror != 0)
+ error = puterror;
+
+ nvlist_free(outnvl);
+ } else {
+ cookie = spl_fstrans_mark();
+ error = vec->zvec_legacy_func(zc);
+ spl_fstrans_unmark(cookie);
+ }
+
+out:
+ nvlist_free(innvl);
rc = ddi_copyout(zc, (void *)arg, sizeof (zfs_cmd_t), flag);
- if (error == 0) {
- if (rc != 0)
- error = EFAULT;
- if (zfs_ioc_vec[vec].zvec_his_log)
- zfs_log_history(zc);
+ if (error == 0 && rc != 0)
+ error = SET_ERROR(EFAULT);
+ if (error == 0 && vec->zvec_allow_log) {
+ char *s = tsd_get(zfs_allow_log_key);
+ if (s != NULL)
+ strfree(s);
+ (void) tsd_set(zfs_allow_log_key, saved_poolname);
+ } else {
+ if (saved_poolname != NULL)
+ strfree(saved_poolname);
}
kmem_free(zc, sizeof (zfs_cmd_t));
static long
zfsdev_compat_ioctl(struct file *filp, unsigned cmd, unsigned long arg)
{
- return zfsdev_ioctl(filp, cmd, arg);
+ return (zfsdev_ioctl(filp, cmd, arg));
}
#else
-#define zfsdev_compat_ioctl NULL
+#define zfsdev_compat_ioctl NULL
#endif
static const struct file_operations zfsdev_fops = {
- .open = zfsdev_open,
- .release = zfsdev_release,
- .unlocked_ioctl = zfsdev_ioctl,
- .compat_ioctl = zfsdev_compat_ioctl,
- .owner = THIS_MODULE,
+ .open = zfsdev_open,
+ .release = zfsdev_release,
+ .unlocked_ioctl = zfsdev_ioctl,
+ .compat_ioctl = zfsdev_compat_ioctl,
+ .owner = THIS_MODULE,
};
static struct miscdevice zfs_misc = {
- .minor = MISC_DYNAMIC_MINOR,
- .name = ZFS_DRIVER,
- .fops = &zfsdev_fops,
+ .minor = MISC_DYNAMIC_MINOR,
+ .name = ZFS_DRIVER,
+ .fops = &zfsdev_fops,
};
static int
int error;
mutex_init(&zfsdev_state_lock, NULL, MUTEX_DEFAULT, NULL);
- list_create(&zfsdev_state_list, sizeof (zfsdev_state_t),
- offsetof(zfsdev_state_t, zs_next));
+ zfsdev_state_list = kmem_zalloc(sizeof (zfsdev_state_t), KM_SLEEP);
+ zfsdev_state_list->zs_minor = -1;
error = misc_register(&zfs_misc);
- if (error) {
+ if (error != 0) {
printk(KERN_INFO "ZFS: misc_register() failed %d\n", error);
return (error);
}
static void
zfs_detach(void)
{
- int error;
-
- error = misc_deregister(&zfs_misc);
- if (error)
- printk(KERN_INFO "ZFS: misc_deregister() failed %d\n", error);
+ zfsdev_state_t *zs, *zsprev = NULL;
+ misc_deregister(&zfs_misc);
mutex_destroy(&zfsdev_state_lock);
- list_destroy(&zfsdev_state_list);
+
+ for (zs = zfsdev_state_list; zs != NULL; zs = zs->zs_next) {
+ if (zsprev)
+ kmem_free(zsprev, sizeof (zfsdev_state_t));
+ zsprev = zs;
+ }
+ if (zsprev)
+ kmem_free(zsprev, sizeof (zfsdev_state_t));
}
-uint_t zfs_fsyncer_key;
-extern uint_t rrw_tsd_key;
+static void
+zfs_allow_log_destroy(void *arg)
+{
+ char *poolname = arg;
+ strfree(poolname);
+}
#ifdef DEBUG
-#define ZFS_DEBUG_STR " (DEBUG mode)"
+#define ZFS_DEBUG_STR " (DEBUG mode)"
#else
-#define ZFS_DEBUG_STR ""
+#define ZFS_DEBUG_STR ""
#endif
-int
+static int __init
_init(void)
{
int error;
+ error = -vn_set_pwd("/");
+ if (error) {
+ printk(KERN_NOTICE
+ "ZFS: Warning unable to set pwd to '/': %d\n", error);
+ return (error);
+ }
+
+ if ((error = -zvol_init()) != 0)
+ return (error);
+
spa_init(FREAD | FWRITE);
zfs_init();
- if ((error = zvol_init()) != 0)
- goto out1;
+ zfs_ioctl_init();
if ((error = zfs_attach()) != 0)
- goto out2;
+ goto out;
tsd_create(&zfs_fsyncer_key, NULL);
- tsd_create(&rrw_tsd_key, NULL);
+ tsd_create(&rrw_tsd_key, rrw_tsd_destroy);
+ tsd_create(&zfs_allow_log_key, zfs_allow_log_destroy);
- printk(KERN_NOTICE "ZFS: Loaded module v%s%s, "
- "ZFS pool version %s, ZFS filesystem version %s\n",
- ZFS_META_VERSION, ZFS_DEBUG_STR,
- SPA_VERSION_STRING, ZPL_VERSION_STRING);
+ printk(KERN_NOTICE "ZFS: Loaded module v%s-%s%s, "
+ "ZFS pool version %s, ZFS filesystem version %s\n",
+ ZFS_META_VERSION, ZFS_META_RELEASE, ZFS_DEBUG_STR,
+ SPA_VERSION_STRING, ZPL_VERSION_STRING);
+#ifndef CONFIG_FS_POSIX_ACL
+ printk(KERN_NOTICE "ZFS: Posix ACLs disabled by kernel\n");
+#endif /* CONFIG_FS_POSIX_ACL */
return (0);
-out2:
- (void) zvol_fini();
-out1:
+out:
zfs_fini();
spa_fini();
- printk(KERN_NOTICE "ZFS: Failed to Load ZFS Filesystem v%s%s"
- ", rc = %d\n", ZFS_META_VERSION, ZFS_DEBUG_STR, error);
+ (void) zvol_fini();
+ printk(KERN_NOTICE "ZFS: Failed to Load ZFS Filesystem v%s-%s%s"
+ ", rc = %d\n", ZFS_META_VERSION, ZFS_META_RELEASE,
+ ZFS_DEBUG_STR, error);
return (error);
}
-int
+static void __exit
_fini(void)
{
zfs_detach();
- zvol_fini();
zfs_fini();
spa_fini();
+ zvol_fini();
tsd_destroy(&zfs_fsyncer_key);
tsd_destroy(&rrw_tsd_key);
+ tsd_destroy(&zfs_allow_log_key);
- printk(KERN_NOTICE "ZFS: Unloaded module v%s%s\n",
- ZFS_META_VERSION, ZFS_DEBUG_STR);
-
- return (0);
+ printk(KERN_NOTICE "ZFS: Unloaded module v%s-%s%s\n",
+ ZFS_META_VERSION, ZFS_META_RELEASE, ZFS_DEBUG_STR);
}
#ifdef HAVE_SPL
-spl_module_init(_init);
-spl_module_exit(_fini);
+module_init(_init);
+module_exit(_fini);
MODULE_DESCRIPTION("ZFS");
MODULE_AUTHOR(ZFS_META_AUTHOR);
MODULE_LICENSE(ZFS_META_LICENSE);
+MODULE_VERSION(ZFS_META_VERSION "-" ZFS_META_RELEASE);
#endif /* HAVE_SPL */
projects / mirror_zfs.git / blobdiff