1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
|
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or http://www.opensolaris.org/os/licensing.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2013 by Delphix. All rights reserved.
*/
#include <sys/types.h>
#include <sys/param.h>
#include <sys/errno.h>
#include <sys/kmem.h>
#include <sys/sunddi.h>
#include <sys/zfs_ioctl.h>
#include <sys/zfs_onexit.h>
#include <sys/zvol.h>
/*
* ZFS kernel routines may add/delete callback routines to be invoked
* upon process exit (triggered via the close operation from the /dev/zfs
* driver).
*
* These cleanup callbacks are intended to allow for the accumulation
* of kernel state across multiple ioctls. User processes participate
* simply by opening ZFS_DEV. This causes the ZFS driver to do create
* some private data for the file descriptor and generating a unique
* minor number. The process then passes along that file descriptor to
* each ioctl that might have a cleanup operation.
*
* Consumers of the onexit routines should call zfs_onexit_fd_hold() early
* on to validate the given fd and add a reference to its file table entry.
* This allows the consumer to do its work and then add a callback, knowing
* that zfs_onexit_add_cb() won't fail with EBADF. When finished, consumers
* should call zfs_onexit_fd_rele().
*
* A simple example is zfs_ioc_recv(), where we might create an AVL tree
* with dataset/GUID mappings and then reuse that tree on subsequent
* zfs_ioc_recv() calls.
*
* On the first zfs_ioc_recv() call, dmu_recv_stream() will kmem_alloc()
* the AVL tree and pass it along with a callback function to
* zfs_onexit_add_cb(). The zfs_onexit_add_cb() routine will register the
* callback and return an action handle.
*
* The action handle is then passed from user space to subsequent
* zfs_ioc_recv() calls, so that dmu_recv_stream() can fetch its AVL tree
* by calling zfs_onexit_cb_data() with the device minor number and
* action handle.
*
* If the user process exits abnormally, the callback is invoked implicitly
* as part of the driver close operation. Once the user space process is
* finished with the accumulated kernel state, it can also just call close(2)
* on the cleanup fd to trigger the cleanup callback.
*/
void
zfs_onexit_init(zfs_onexit_t **zop)
{
zfs_onexit_t *zo;
zo = *zop = kmem_zalloc(sizeof (zfs_onexit_t), KM_SLEEP);
mutex_init(&zo->zo_lock, NULL, MUTEX_DEFAULT, NULL);
list_create(&zo->zo_actions, sizeof (zfs_onexit_action_node_t),
offsetof(zfs_onexit_action_node_t, za_link));
}
void
zfs_onexit_destroy(zfs_onexit_t *zo)
{
zfs_onexit_action_node_t *ap;
mutex_enter(&zo->zo_lock);
while ((ap = list_head(&zo->zo_actions)) != NULL) {
list_remove(&zo->zo_actions, ap);
mutex_exit(&zo->zo_lock);
ap->za_func(ap->za_data);
kmem_free(ap, sizeof (zfs_onexit_action_node_t));
mutex_enter(&zo->zo_lock);
}
mutex_exit(&zo->zo_lock);
list_destroy(&zo->zo_actions);
mutex_destroy(&zo->zo_lock);
kmem_free(zo, sizeof (zfs_onexit_t));
}
static int
zfs_onexit_minor_to_state(minor_t minor, zfs_onexit_t **zo)
{
*zo = zfsdev_get_state(minor, ZST_ONEXIT);
if (*zo == NULL)
return (SET_ERROR(EBADF));
return (0);
}
/*
* Add a callback to be invoked when the calling process exits.
*/
int
zfs_onexit_add_cb(minor_t minor, void (*func)(void *), void *data,
uint64_t *action_handle)
{
zfs_onexit_t *zo;
zfs_onexit_action_node_t *ap;
int error;
error = zfs_onexit_minor_to_state(minor, &zo);
if (error)
return (error);
ap = kmem_alloc(sizeof (zfs_onexit_action_node_t), KM_SLEEP);
list_link_init(&ap->za_link);
ap->za_func = func;
ap->za_data = data;
mutex_enter(&zo->zo_lock);
list_insert_tail(&zo->zo_actions, ap);
mutex_exit(&zo->zo_lock);
if (action_handle)
*action_handle = (uint64_t)(uintptr_t)ap;
return (0);
}
static zfs_onexit_action_node_t *
zfs_onexit_find_cb(zfs_onexit_t *zo, uint64_t action_handle)
{
zfs_onexit_action_node_t *match;
zfs_onexit_action_node_t *ap;
list_t *l;
ASSERT(MUTEX_HELD(&zo->zo_lock));
match = (zfs_onexit_action_node_t *)(uintptr_t)action_handle;
l = &zo->zo_actions;
for (ap = list_head(l); ap != NULL; ap = list_next(l, ap)) {
if (match == ap)
break;
}
return (ap);
}
/*
* Delete the callback, triggering it first if 'fire' is set.
*/
int
zfs_onexit_del_cb(minor_t minor, uint64_t action_handle, boolean_t fire)
{
zfs_onexit_t *zo;
zfs_onexit_action_node_t *ap;
int error;
error = zfs_onexit_minor_to_state(minor, &zo);
if (error)
return (error);
mutex_enter(&zo->zo_lock);
ap = zfs_onexit_find_cb(zo, action_handle);
if (ap != NULL) {
list_remove(&zo->zo_actions, ap);
mutex_exit(&zo->zo_lock);
if (fire)
ap->za_func(ap->za_data);
kmem_free(ap, sizeof (zfs_onexit_action_node_t));
} else {
mutex_exit(&zo->zo_lock);
error = SET_ERROR(ENOENT);
}
return (error);
}
/*
* Return the data associated with this callback. This allows consumers
* of the cleanup-on-exit interfaces to stash kernel data across system
* calls, knowing that it will be cleaned up if the calling process exits.
*/
int
zfs_onexit_cb_data(minor_t minor, uint64_t action_handle, void **data)
{
zfs_onexit_t *zo;
zfs_onexit_action_node_t *ap;
int error;
*data = NULL;
error = zfs_onexit_minor_to_state(minor, &zo);
if (error)
return (error);
mutex_enter(&zo->zo_lock);
ap = zfs_onexit_find_cb(zo, action_handle);
if (ap != NULL)
*data = ap->za_data;
else
error = SET_ERROR(ENOENT);
mutex_exit(&zo->zo_lock);
return (error);
}
|
