Print this page
NEX-9200 Improve the scalability of attribute locking in zfs_zget
Reviewed by: Joyce McIntosh <joyce.mcintosh@nexenta.com>
Reviewed by: Sanjay Nadkarni <sanjay.nadkarni@nexenta.com>
NEX-9436 Rate limiting controls (was QoS) per ZFS dataset, updates from demo
Reviewed by: Gordon Ross <gordon.ross@nexenta.com>
Reviewed by: Rob Gittins <rob.gittins@nexenta.com>
NEX-8972 Async-delete side-effect that may cause unmount EBUSY
Reviewed by: Alek Pinchuk <alek@nexenta.com>
Reviewed by: Sanjay Nadkarni <sanjay.nadkarni@nexenta.com>
NEX-8852 Quality-of-Service (QoS) controls per NFS share
Reviewed by: Rob Gittins <rob.gittins@nexenta.com>
Reviewed by: Evan Layton <evan.layton@nexenta.com>
Reviewed by: Sanjay Nadkarni <sanjay.nadkarni@nexenta.com>
NEX-3762 Appliance crashes with a NULL pointer dereference during a zpool export when a zfs_vn_rele_taskq thread attempts to check a bogus rwlock from rw_write_held
Reviewed by: Josef 'Jeff' Sipek <josef.sipek@nexenta.com>
OS-80 support for vdev and CoS properties for the new I/O scheduler
OS-95 lint warning introduced by OS-61
Fixup merge results
re #13204 rb4280 zfs receive/rollback deadlock
re #6815 rb1758 need WORM in nza-kernel (4.0)
| Split |
Close |
| Expand all |
| Collapse all |
--- old/usr/src/uts/common/fs/zfs/sys/zfs_vfsops.h
+++ new/usr/src/uts/common/fs/zfs/sys/zfs_vfsops.h
1 1 /*
2 2 * CDDL HEADER START
3 3 *
4 4 * The contents of this file are subject to the terms of the
5 5 * Common Development and Distribution License (the "License").
6 6 * You may not use this file except in compliance with the License.
7 7 *
8 8 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
9 9 * or http://www.opensolaris.org/os/licensing.
10 10 * See the License for the specific language governing permissions
11 11 * and limitations under the License.
12 12 *
|
↓ open down ↓ |
12 lines elided |
↑ open up ↑ |
13 13 * When distributing Covered Code, include this CDDL HEADER in each
14 14 * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
15 15 * If applicable, add the following below this CDDL HEADER, with the
16 16 * fields enclosed by brackets "[]" replaced with your own identifying
17 17 * information: Portions Copyright [yyyy] [name of copyright owner]
18 18 *
19 19 * CDDL HEADER END
20 20 */
21 21 /*
22 22 * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
23 + * Copyright 2016 Nexenta Systems, Inc. All rights reserved.
23 24 */
24 25
25 26 #ifndef _SYS_FS_ZFS_VFSOPS_H
26 27 #define _SYS_FS_ZFS_VFSOPS_H
27 28
28 29 #include <sys/isa_defs.h>
29 30 #include <sys/types32.h>
30 31 #include <sys/list.h>
31 32 #include <sys/vfs.h>
32 33 #include <sys/zil.h>
33 34 #include <sys/sa.h>
|
↓ open down ↓ |
1 lines elided |
↑ open up ↑ |
34 35 #include <sys/rrwlock.h>
35 36 #include <sys/zfs_ioctl.h>
36 37
37 38 #ifdef __cplusplus
38 39 extern "C" {
39 40 #endif
40 41
41 42 typedef struct zfsvfs zfsvfs_t;
42 43 struct znode;
43 44
45 +/*
46 + * ZFS Quality of Service (QoS) I/O throttling state,
47 + * per file system. Limits the I/O rate in this FS.
48 + * See "Token Bucket" on Wikipedia
49 + */
50 +typedef struct zfs_rate_state {
51 + uint64_t rate_cap; /* zero means no cap */
52 + int64_t rate_token_bucket; /* bytes I/O allowed without waiting */
53 + hrtime_t rate_last_update;
54 + kmutex_t rate_lock;
55 + kcondvar_t rate_wait_cv;
56 + int rate_waiters;
57 +} zfs_rate_state_t;
58 +
59 +/*
60 + * Status of the zfs_unlinked_drain thread.
61 + */
62 +typedef enum drain_state {
63 + ZFS_DRAIN_SHUTDOWN = 0,
64 + ZFS_DRAIN_RUNNING,
65 + ZFS_DRAIN_SHUTDOWN_REQ
66 +} drain_state_t;
67 +
44 68 struct zfsvfs {
45 69 vfs_t *z_vfs; /* generic fs struct */
46 70 zfsvfs_t *z_parent; /* parent fs */
47 71 objset_t *z_os; /* objset reference */
48 72 uint64_t z_root; /* id of root znode */
49 73 uint64_t z_unlinkedobj; /* id of unlinked zapobj */
50 74 uint64_t z_max_blksz; /* maximum block size for files */
51 75 uint64_t z_fuid_obj; /* fuid table object number */
52 76 uint64_t z_fuid_size; /* fuid table size */
53 77 avl_tree_t z_fuid_idx; /* fuid tree keyed by index */
54 78 avl_tree_t z_fuid_domain; /* fuid tree keyed by domain */
55 79 krwlock_t z_fuid_lock; /* fuid lock */
56 80 boolean_t z_fuid_loaded; /* fuid tables are loaded */
57 81 boolean_t z_fuid_dirty; /* need to sync fuid table ? */
58 82 struct zfs_fuid_info *z_fuid_replay; /* fuid info for replay */
59 83 zilog_t *z_log; /* intent log pointer */
60 84 uint_t z_acl_mode; /* acl chmod/mode behavior */
|
↓ open down ↓ |
7 lines elided |
↑ open up ↑ |
61 85 uint_t z_acl_inherit; /* acl inheritance behavior */
62 86 zfs_case_t z_case; /* case-sense */
63 87 boolean_t z_utf8; /* utf8-only */
64 88 int z_norm; /* normalization flags */
65 89 boolean_t z_atime; /* enable atimes mount option */
66 90 boolean_t z_unmounted; /* unmounted */
67 91 rrmlock_t z_teardown_lock;
68 92 krwlock_t z_teardown_inactive_lock;
69 93 list_t z_all_znodes; /* all vnodes in the fs */
70 94 kmutex_t z_znodes_lock; /* lock for z_all_znodes */
95 + uint_t z_znodes_freeing_cnt; /* number of znodes to be freed */
71 96 vnode_t *z_ctldir; /* .zfs directory pointer */
72 97 boolean_t z_show_ctldir; /* expose .zfs in the root dir */
73 98 boolean_t z_issnap; /* true if this is a snapshot */
74 99 boolean_t z_vscan; /* virus scan on/off */
75 100 boolean_t z_use_fuids; /* version allows fuids */
76 101 boolean_t z_replay; /* set during ZIL replay */
77 102 boolean_t z_use_sa; /* version allow system attributes */
78 103 uint64_t z_version; /* ZPL version */
79 104 uint64_t z_shares_dir; /* hidden shares dir */
80 105 kmutex_t z_lock;
81 106 uint64_t z_userquota_obj;
82 107 uint64_t z_groupquota_obj;
83 108 uint64_t z_replay_eof; /* New end of file - replay only */
84 109 sa_attr_type_t *z_attr_table; /* SA attr mapping->id */
85 -#define ZFS_OBJ_MTX_SZ 64
86 - kmutex_t z_hold_mtx[ZFS_OBJ_MTX_SZ]; /* znode hold locks */
110 + boolean_t z_isworm; /* true if this is a WORM FS */
111 + /* true if suspend-resume cycle is in progress */
112 + boolean_t z_busy;
113 + int z_hold_mtx_sz; /* the size of z_hold_mtx array */
114 + kmutex_t *z_hold_mtx; /* znode hold locks */
115 + /* for controlling async zfs_unlinked_drain */
116 + kmutex_t z_drain_lock;
117 + kcondvar_t z_drain_cv;
118 + drain_state_t z_drain_state;
119 + zfs_rate_state_t z_rate;
87 120 };
88 121
89 122 /*
90 123 * Normal filesystems (those not under .zfs/snapshot) have a total
91 124 * file ID size limited to 12 bytes (including the length field) due to
92 125 * NFSv2 protocol's limitation of 32 bytes for a filehandle. For historical
93 126 * reasons, this same limit is being imposed by the Solaris NFSv3 implementation
94 127 * (although the NFSv3 protocol actually permits a maximum of 64 bytes). It
95 128 * is not possible to expand beyond 12 bytes without abandoning support
96 129 * of NFSv2.
97 130 *
98 131 * For normal filesystems, we partition up the available space as follows:
99 132 * 2 bytes fid length (required)
100 133 * 6 bytes object number (48 bits)
101 134 * 4 bytes generation number (32 bits)
102 135 *
103 136 * We reserve only 48 bits for the object number, as this is the limit
104 137 * currently defined and imposed by the DMU.
105 138 */
106 139 typedef struct zfid_short {
107 140 uint16_t zf_len;
108 141 uint8_t zf_object[6]; /* obj[i] = obj >> (8 * i) */
109 142 uint8_t zf_gen[4]; /* gen[i] = gen >> (8 * i) */
110 143 } zfid_short_t;
111 144
112 145 /*
113 146 * Filesystems under .zfs/snapshot have a total file ID size of 22 bytes
114 147 * (including the length field). This makes files under .zfs/snapshot
115 148 * accessible by NFSv3 and NFSv4, but not NFSv2.
116 149 *
117 150 * For files under .zfs/snapshot, we partition up the available space
118 151 * as follows:
119 152 * 2 bytes fid length (required)
120 153 * 6 bytes object number (48 bits)
121 154 * 4 bytes generation number (32 bits)
122 155 * 6 bytes objset id (48 bits)
123 156 * 4 bytes currently just zero (32 bits)
124 157 *
125 158 * We reserve only 48 bits for the object number and objset id, as these are
126 159 * the limits currently defined and imposed by the DMU.
127 160 */
128 161 typedef struct zfid_long {
129 162 zfid_short_t z_fid;
130 163 uint8_t zf_setid[6]; /* obj[i] = obj >> (8 * i) */
131 164 uint8_t zf_setgen[4]; /* gen[i] = gen >> (8 * i) */
132 165 } zfid_long_t;
133 166
134 167 #define SHORT_FID_LEN (sizeof (zfid_short_t) - sizeof (uint16_t))
135 168 #define LONG_FID_LEN (sizeof (zfid_long_t) - sizeof (uint16_t))
136 169
137 170 extern uint_t zfs_fsyncer_key;
138 171
139 172 extern int zfs_suspend_fs(zfsvfs_t *zfsvfs);
140 173 extern int zfs_resume_fs(zfsvfs_t *zfsvfs, struct dsl_dataset *ds);
141 174 extern int zfs_userspace_one(zfsvfs_t *zfsvfs, zfs_userquota_prop_t type,
142 175 const char *domain, uint64_t rid, uint64_t *valuep);
143 176 extern int zfs_userspace_many(zfsvfs_t *zfsvfs, zfs_userquota_prop_t type,
144 177 uint64_t *cookiep, void *vbuf, uint64_t *bufsizep);
145 178 extern int zfs_set_userquota(zfsvfs_t *zfsvfs, zfs_userquota_prop_t type,
146 179 const char *domain, uint64_t rid, uint64_t quota);
147 180 extern boolean_t zfs_owner_overquota(zfsvfs_t *zfsvfs, struct znode *,
148 181 boolean_t isgroup);
149 182 extern boolean_t zfs_fuid_overquota(zfsvfs_t *zfsvfs, boolean_t isgroup,
150 183 uint64_t fuid);
151 184 extern int zfs_set_version(zfsvfs_t *zfsvfs, uint64_t newvers);
152 185 extern int zfsvfs_create(const char *name, zfsvfs_t **zfvp);
153 186 extern int zfsvfs_create_impl(zfsvfs_t **zfvp, zfsvfs_t *zfsvfs, objset_t *os);
154 187 extern void zfsvfs_free(zfsvfs_t *zfsvfs);
155 188 extern int zfs_check_global_label(const char *dsname, const char *hexsl);
156 189
157 190 #ifdef __cplusplus
158 191 }
159 192 #endif
160 193
161 194 #endif /* _SYS_FS_ZFS_VFSOPS_H */
|
↓ open down ↓ |
65 lines elided |
↑ open up ↑ |
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX