aboutsummaryrefslogtreecommitdiffstats
path: root/include
diff options
context:
space:
mode:
authorLinus Torvalds <torvalds@linux-foundation.org>2012-12-17 16:39:11 -0500
committerLinus Torvalds <torvalds@linux-foundation.org>2012-12-17 16:39:11 -0500
commit9228ff90387e276ad67b10c0eb525c9d6a57d5e9 (patch)
treee7c87b68daba7cf7ca4c342c6b52165bd78fbe16 /include
parent9360b53661a2c7754517b2925580055bacc8ec38 (diff)
parentd2ec180c23a5a1bfe34d8638b0342a47c00cf70f (diff)
Merge branch 'for-3.8/drivers' of git://git.kernel.dk/linux-block
Pull block driver update from Jens Axboe: "Now that the core bits are in, here are the driver bits for 3.8. The branch contains: - A huge pile of drbd bits that were dumped from the 3.7 merge window. Following that, it was both made perfectly clear that there is going to be no more over-the-wall pulls and how the situation on individual pulls can be improved. - A few cleanups from Akinobu Mita for drbd and cciss. - Queue improvement for loop from Lukas. This grew into adding a generic interface for waiting/checking an even with a specific lock, allowing this to be pulled out of md and now loop and drbd is also using it. - A few fixes for xen back/front block driver from Roger Pau Monne. - Partition improvements from Stephen Warren, allowing partiion UUID to be used as an identifier." * 'for-3.8/drivers' of git://git.kernel.dk/linux-block: (609 commits) drbd: update Kconfig to match current dependencies drbd: Fix drbdsetup wait-connect, wait-sync etc... commands drbd: close race between drbd_set_role and drbd_connect drbd: respect no-md-barriers setting also when changed online via disk-options drbd: Remove obsolete check drbd: fixup after wait_even_lock_irq() addition to generic code loop: Limit the number of requests in the bio list wait: add wait_event_lock_irq() interface xen-blkfront: free allocated page xen-blkback: move free persistent grants code block: partition: msdos: provide UUIDs for partitions init: reduce PARTUUID min length to 1 from 36 block: store partition_meta_info.uuid as a string cciss: use check_signature() cciss: cleanup bitops usage drbd: use copy_highpage drbd: if the replication link breaks during handshake, keep retrying drbd: check return of kmalloc in receive_uuids drbd: Broadcast sync progress no more often than once per second drbd: don't try to clear bits once the disk has failed ...
Diffstat (limited to 'include')
-rw-r--r--include/linux/drbd.h81
-rw-r--r--include/linux/drbd_genl.h378
-rw-r--r--include/linux/drbd_genl_api.h55
-rw-r--r--include/linux/drbd_limits.h90
-rw-r--r--include/linux/drbd_nl.h163
-rw-r--r--include/linux/drbd_tag_magic.h84
-rw-r--r--include/linux/genhd.h8
-rw-r--r--include/linux/genl_magic_func.h422
-rw-r--r--include/linux/genl_magic_struct.h277
-rw-r--r--include/linux/idr.h11
-rw-r--r--include/linux/loop.h3
-rw-r--r--include/linux/lru_cache.h67
-rw-r--r--include/linux/wait.h164
13 files changed, 1473 insertions, 330 deletions
diff --git a/include/linux/drbd.h b/include/linux/drbd.h
index 47e3d4850584..0c5a18ec322c 100644
--- a/include/linux/drbd.h
+++ b/include/linux/drbd.h
@@ -51,12 +51,11 @@
51 51
52#endif 52#endif
53 53
54
55extern const char *drbd_buildtag(void); 54extern const char *drbd_buildtag(void);
56#define REL_VERSION "8.3.13" 55#define REL_VERSION "8.4.2"
57#define API_VERSION 88 56#define API_VERSION 1
58#define PRO_VERSION_MIN 86 57#define PRO_VERSION_MIN 86
59#define PRO_VERSION_MAX 96 58#define PRO_VERSION_MAX 101
60 59
61 60
62enum drbd_io_error_p { 61enum drbd_io_error_p {
@@ -66,7 +65,8 @@ enum drbd_io_error_p {
66}; 65};
67 66
68enum drbd_fencing_p { 67enum drbd_fencing_p {
69 FP_DONT_CARE, 68 FP_NOT_AVAIL = -1, /* Not a policy */
69 FP_DONT_CARE = 0,
70 FP_RESOURCE, 70 FP_RESOURCE,
71 FP_STONITH 71 FP_STONITH
72}; 72};
@@ -102,6 +102,20 @@ enum drbd_on_congestion {
102 OC_DISCONNECT, 102 OC_DISCONNECT,
103}; 103};
104 104
105enum drbd_read_balancing {
106 RB_PREFER_LOCAL,
107 RB_PREFER_REMOTE,
108 RB_ROUND_ROBIN,
109 RB_LEAST_PENDING,
110 RB_CONGESTED_REMOTE,
111 RB_32K_STRIPING,
112 RB_64K_STRIPING,
113 RB_128K_STRIPING,
114 RB_256K_STRIPING,
115 RB_512K_STRIPING,
116 RB_1M_STRIPING,
117};
118
105/* KEEP the order, do not delete or insert. Only append. */ 119/* KEEP the order, do not delete or insert. Only append. */
106enum drbd_ret_code { 120enum drbd_ret_code {
107 ERR_CODE_BASE = 100, 121 ERR_CODE_BASE = 100,
@@ -122,7 +136,7 @@ enum drbd_ret_code {
122 ERR_AUTH_ALG = 120, 136 ERR_AUTH_ALG = 120,
123 ERR_AUTH_ALG_ND = 121, 137 ERR_AUTH_ALG_ND = 121,
124 ERR_NOMEM = 122, 138 ERR_NOMEM = 122,
125 ERR_DISCARD = 123, 139 ERR_DISCARD_IMPOSSIBLE = 123,
126 ERR_DISK_CONFIGURED = 124, 140 ERR_DISK_CONFIGURED = 124,
127 ERR_NET_CONFIGURED = 125, 141 ERR_NET_CONFIGURED = 125,
128 ERR_MANDATORY_TAG = 126, 142 ERR_MANDATORY_TAG = 126,
@@ -130,8 +144,8 @@ enum drbd_ret_code {
130 ERR_INTR = 129, /* EINTR */ 144 ERR_INTR = 129, /* EINTR */
131 ERR_RESIZE_RESYNC = 130, 145 ERR_RESIZE_RESYNC = 130,
132 ERR_NO_PRIMARY = 131, 146 ERR_NO_PRIMARY = 131,
133 ERR_SYNC_AFTER = 132, 147 ERR_RESYNC_AFTER = 132,
134 ERR_SYNC_AFTER_CYCLE = 133, 148 ERR_RESYNC_AFTER_CYCLE = 133,
135 ERR_PAUSE_IS_SET = 134, 149 ERR_PAUSE_IS_SET = 134,
136 ERR_PAUSE_IS_CLEAR = 135, 150 ERR_PAUSE_IS_CLEAR = 135,
137 ERR_PACKET_NR = 137, 151 ERR_PACKET_NR = 137,
@@ -155,6 +169,14 @@ enum drbd_ret_code {
155 ERR_CONG_NOT_PROTO_A = 155, 169 ERR_CONG_NOT_PROTO_A = 155,
156 ERR_PIC_AFTER_DEP = 156, 170 ERR_PIC_AFTER_DEP = 156,
157 ERR_PIC_PEER_DEP = 157, 171 ERR_PIC_PEER_DEP = 157,
172 ERR_RES_NOT_KNOWN = 158,
173 ERR_RES_IN_USE = 159,
174 ERR_MINOR_CONFIGURED = 160,
175 ERR_MINOR_EXISTS = 161,
176 ERR_INVALID_REQUEST = 162,
177 ERR_NEED_APV_100 = 163,
178 ERR_NEED_ALLOW_TWO_PRI = 164,
179 ERR_MD_UNCLEAN = 165,
158 180
159 /* insert new ones above this line */ 181 /* insert new ones above this line */
160 AFTER_LAST_ERR_CODE 182 AFTER_LAST_ERR_CODE
@@ -296,7 +318,8 @@ enum drbd_state_rv {
296 SS_NOT_SUPPORTED = -17, /* drbd-8.2 only */ 318 SS_NOT_SUPPORTED = -17, /* drbd-8.2 only */
297 SS_IN_TRANSIENT_STATE = -18, /* Retry after the next state change */ 319 SS_IN_TRANSIENT_STATE = -18, /* Retry after the next state change */
298 SS_CONCURRENT_ST_CHG = -19, /* Concurrent cluster side state change! */ 320 SS_CONCURRENT_ST_CHG = -19, /* Concurrent cluster side state change! */
299 SS_AFTER_LAST_ERROR = -20, /* Keep this at bottom */ 321 SS_O_VOL_PEER_PRI = -20,
322 SS_AFTER_LAST_ERROR = -21, /* Keep this at bottom */
300}; 323};
301 324
302/* from drbd_strings.c */ 325/* from drbd_strings.c */
@@ -313,7 +336,9 @@ extern const char *drbd_set_st_err_str(enum drbd_state_rv);
313#define MDF_FULL_SYNC (1 << 3) 336#define MDF_FULL_SYNC (1 << 3)
314#define MDF_WAS_UP_TO_DATE (1 << 4) 337#define MDF_WAS_UP_TO_DATE (1 << 4)
315#define MDF_PEER_OUT_DATED (1 << 5) 338#define MDF_PEER_OUT_DATED (1 << 5)
316#define MDF_CRASHED_PRIMARY (1 << 6) 339#define MDF_CRASHED_PRIMARY (1 << 6)
340#define MDF_AL_CLEAN (1 << 7)
341#define MDF_AL_DISABLED (1 << 8)
317 342
318enum drbd_uuid_index { 343enum drbd_uuid_index {
319 UI_CURRENT, 344 UI_CURRENT,
@@ -333,37 +358,23 @@ enum drbd_timeout_flag {
333 358
334#define UUID_JUST_CREATED ((__u64)4) 359#define UUID_JUST_CREATED ((__u64)4)
335 360
361/* magic numbers used in meta data and network packets */
336#define DRBD_MAGIC 0x83740267 362#define DRBD_MAGIC 0x83740267
337#define BE_DRBD_MAGIC __constant_cpu_to_be32(DRBD_MAGIC)
338#define DRBD_MAGIC_BIG 0x835a 363#define DRBD_MAGIC_BIG 0x835a
339#define BE_DRBD_MAGIC_BIG __constant_cpu_to_be16(DRBD_MAGIC_BIG) 364#define DRBD_MAGIC_100 0x8620ec20
365
366#define DRBD_MD_MAGIC_07 (DRBD_MAGIC+3)
367#define DRBD_MD_MAGIC_08 (DRBD_MAGIC+4)
368#define DRBD_MD_MAGIC_84_UNCLEAN (DRBD_MAGIC+5)
369
370
371/* how I came up with this magic?
372 * base64 decode "actlog==" ;) */
373#define DRBD_AL_MAGIC 0x69cb65a2
340 374
341/* these are of type "int" */ 375/* these are of type "int" */
342#define DRBD_MD_INDEX_INTERNAL -1 376#define DRBD_MD_INDEX_INTERNAL -1
343#define DRBD_MD_INDEX_FLEX_EXT -2 377#define DRBD_MD_INDEX_FLEX_EXT -2
344#define DRBD_MD_INDEX_FLEX_INT -3 378#define DRBD_MD_INDEX_FLEX_INT -3
345 379
346/* Start of the new netlink/connector stuff */
347
348#define DRBD_NL_CREATE_DEVICE 0x01
349#define DRBD_NL_SET_DEFAULTS 0x02
350
351
352/* For searching a vacant cn_idx value */
353#define CN_IDX_STEP 6977
354
355struct drbd_nl_cfg_req {
356 int packet_type;
357 unsigned int drbd_minor;
358 int flags;
359 unsigned short tag_list[];
360};
361
362struct drbd_nl_cfg_reply {
363 int packet_type;
364 unsigned int minor;
365 int ret_code; /* enum ret_code or set_st_err_t */
366 unsigned short tag_list[]; /* only used with get_* calls */
367};
368
369#endif 380#endif
diff --git a/include/linux/drbd_genl.h b/include/linux/drbd_genl.h
new file mode 100644
index 000000000000..d0d8fac8a6e4
--- /dev/null
+++ b/include/linux/drbd_genl.h
@@ -0,0 +1,378 @@
1/*
2 * General overview:
3 * full generic netlink message:
4 * |nlmsghdr|genlmsghdr|<payload>
5 *
6 * payload:
7 * |optional fixed size family header|<sequence of netlink attributes>
8 *
9 * sequence of netlink attributes:
10 * I chose to have all "top level" attributes NLA_NESTED,
11 * corresponding to some real struct.
12 * So we have a sequence of |tla, len|<nested nla sequence>
13 *
14 * nested nla sequence:
15 * may be empty, or contain a sequence of netlink attributes
16 * representing the struct fields.
17 *
18 * The tag number of any field (regardless of containing struct)
19 * will be available as T_ ## field_name,
20 * so you cannot have the same field name in two differnt structs.
21 *
22 * The tag numbers themselves are per struct, though,
23 * so should always begin at 1 (not 0, that is the special "NLA_UNSPEC" type,
24 * which we won't use here).
25 * The tag numbers are used as index in the respective nla_policy array.
26 *
27 * GENL_struct(tag_name, tag_number, struct name, struct fields) - struct and policy
28 * genl_magic_struct.h
29 * generates the struct declaration,
30 * generates an entry in the tla enum,
31 * genl_magic_func.h
32 * generates an entry in the static tla policy
33 * with .type = NLA_NESTED
34 * generates the static <struct_name>_nl_policy definition,
35 * and static conversion functions
36 *
37 * genl_magic_func.h
38 *
39 * GENL_mc_group(group)
40 * genl_magic_struct.h
41 * does nothing
42 * genl_magic_func.h
43 * defines and registers the mcast group,
44 * and provides a send helper
45 *
46 * GENL_notification(op_name, op_num, mcast_group, tla list)
47 * These are notifications to userspace.
48 *
49 * genl_magic_struct.h
50 * generates an entry in the genl_ops enum,
51 * genl_magic_func.h
52 * does nothing
53 *
54 * mcast group: the name of the mcast group this notification should be
55 * expected on
56 * tla list: the list of expected top level attributes,
57 * for documentation and sanity checking.
58 *
59 * GENL_op(op_name, op_num, flags and handler, tla list) - "genl operations"
60 * These are requests from userspace.
61 *
62 * _op and _notification share the same "number space",
63 * op_nr will be assigned to "genlmsghdr->cmd"
64 *
65 * genl_magic_struct.h
66 * generates an entry in the genl_ops enum,
67 * genl_magic_func.h
68 * generates an entry in the static genl_ops array,
69 * and static register/unregister functions to
70 * genl_register_family_with_ops().
71 *
72 * flags and handler:
73 * GENL_op_init( .doit = x, .dumpit = y, .flags = something)
74 * GENL_doit(x) => .dumpit = NULL, .flags = GENL_ADMIN_PERM
75 * tla list: the list of expected top level attributes,
76 * for documentation and sanity checking.
77 */
78
79/*
80 * STRUCTS
81 */
82
83/* this is sent kernel -> userland on various error conditions, and contains
84 * informational textual info, which is supposedly human readable.
85 * The computer relevant return code is in the drbd_genlmsghdr.
86 */
87GENL_struct(DRBD_NLA_CFG_REPLY, 1, drbd_cfg_reply,
88 /* "arbitrary" size strings, nla_policy.len = 0 */
89 __str_field(1, DRBD_GENLA_F_MANDATORY, info_text, 0)
90)
91
92/* Configuration requests typically need a context to operate on.
93 * Possible keys are device minor (fits in the drbd_genlmsghdr),
94 * the replication link (aka connection) name,
95 * and/or the replication group (aka resource) name,
96 * and the volume id within the resource. */
97GENL_struct(DRBD_NLA_CFG_CONTEXT, 2, drbd_cfg_context,
98 __u32_field(1, DRBD_GENLA_F_MANDATORY, ctx_volume)
99 __str_field(2, DRBD_GENLA_F_MANDATORY, ctx_resource_name, 128)
100 __bin_field(3, DRBD_GENLA_F_MANDATORY, ctx_my_addr, 128)
101 __bin_field(4, DRBD_GENLA_F_MANDATORY, ctx_peer_addr, 128)
102)
103
104GENL_struct(DRBD_NLA_DISK_CONF, 3, disk_conf,
105 __str_field(1, DRBD_F_REQUIRED | DRBD_F_INVARIANT, backing_dev, 128)
106 __str_field(2, DRBD_F_REQUIRED | DRBD_F_INVARIANT, meta_dev, 128)
107 __s32_field(3, DRBD_F_REQUIRED | DRBD_F_INVARIANT, meta_dev_idx)
108
109 /* use the resize command to try and change the disk_size */
110 __u64_field(4, DRBD_GENLA_F_MANDATORY | DRBD_F_INVARIANT, disk_size)
111 /* we could change the max_bio_bvecs,
112 * but it won't propagate through the stack */
113 __u32_field(5, DRBD_GENLA_F_MANDATORY | DRBD_F_INVARIANT, max_bio_bvecs)
114
115 __u32_field_def(6, DRBD_GENLA_F_MANDATORY, on_io_error, DRBD_ON_IO_ERROR_DEF)
116 __u32_field_def(7, DRBD_GENLA_F_MANDATORY, fencing, DRBD_FENCING_DEF)
117
118 __u32_field_def(8, DRBD_GENLA_F_MANDATORY, resync_rate, DRBD_RESYNC_RATE_DEF)
119 __s32_field_def(9, DRBD_GENLA_F_MANDATORY, resync_after, DRBD_MINOR_NUMBER_DEF)
120 __u32_field_def(10, DRBD_GENLA_F_MANDATORY, al_extents, DRBD_AL_EXTENTS_DEF)
121 __u32_field_def(11, DRBD_GENLA_F_MANDATORY, c_plan_ahead, DRBD_C_PLAN_AHEAD_DEF)
122 __u32_field_def(12, DRBD_GENLA_F_MANDATORY, c_delay_target, DRBD_C_DELAY_TARGET_DEF)
123 __u32_field_def(13, DRBD_GENLA_F_MANDATORY, c_fill_target, DRBD_C_FILL_TARGET_DEF)
124 __u32_field_def(14, DRBD_GENLA_F_MANDATORY, c_max_rate, DRBD_C_MAX_RATE_DEF)
125 __u32_field_def(15, DRBD_GENLA_F_MANDATORY, c_min_rate, DRBD_C_MIN_RATE_DEF)
126
127 __flg_field_def(16, DRBD_GENLA_F_MANDATORY, disk_barrier, DRBD_DISK_BARRIER_DEF)
128 __flg_field_def(17, DRBD_GENLA_F_MANDATORY, disk_flushes, DRBD_DISK_FLUSHES_DEF)
129 __flg_field_def(18, DRBD_GENLA_F_MANDATORY, disk_drain, DRBD_DISK_DRAIN_DEF)
130 __flg_field_def(19, DRBD_GENLA_F_MANDATORY, md_flushes, DRBD_MD_FLUSHES_DEF)
131 __u32_field_def(20, DRBD_GENLA_F_MANDATORY, disk_timeout, DRBD_DISK_TIMEOUT_DEF)
132 __u32_field_def(21, 0 /* OPTIONAL */, read_balancing, DRBD_READ_BALANCING_DEF)
133 /* 9: __u32_field_def(22, DRBD_GENLA_F_MANDATORY, unplug_watermark, DRBD_UNPLUG_WATERMARK_DEF) */
134 __flg_field_def(23, 0 /* OPTIONAL */, al_updates, DRBD_AL_UPDATES_DEF)
135)
136
137GENL_struct(DRBD_NLA_RESOURCE_OPTS, 4, res_opts,
138 __str_field_def(1, DRBD_GENLA_F_MANDATORY, cpu_mask, 32)
139 __u32_field_def(2, DRBD_GENLA_F_MANDATORY, on_no_data, DRBD_ON_NO_DATA_DEF)
140)
141
142GENL_struct(DRBD_NLA_NET_CONF, 5, net_conf,
143 __str_field_def(1, DRBD_GENLA_F_MANDATORY | DRBD_F_SENSITIVE,
144 shared_secret, SHARED_SECRET_MAX)
145 __str_field_def(2, DRBD_GENLA_F_MANDATORY, cram_hmac_alg, SHARED_SECRET_MAX)
146 __str_field_def(3, DRBD_GENLA_F_MANDATORY, integrity_alg, SHARED_SECRET_MAX)
147 __str_field_def(4, DRBD_GENLA_F_MANDATORY, verify_alg, SHARED_SECRET_MAX)
148 __str_field_def(5, DRBD_GENLA_F_MANDATORY, csums_alg, SHARED_SECRET_MAX)
149 __u32_field_def(6, DRBD_GENLA_F_MANDATORY, wire_protocol, DRBD_PROTOCOL_DEF)
150 __u32_field_def(7, DRBD_GENLA_F_MANDATORY, connect_int, DRBD_CONNECT_INT_DEF)
151 __u32_field_def(8, DRBD_GENLA_F_MANDATORY, timeout, DRBD_TIMEOUT_DEF)
152 __u32_field_def(9, DRBD_GENLA_F_MANDATORY, ping_int, DRBD_PING_INT_DEF)
153 __u32_field_def(10, DRBD_GENLA_F_MANDATORY, ping_timeo, DRBD_PING_TIMEO_DEF)
154 __u32_field_def(11, DRBD_GENLA_F_MANDATORY, sndbuf_size, DRBD_SNDBUF_SIZE_DEF)
155 __u32_field_def(12, DRBD_GENLA_F_MANDATORY, rcvbuf_size, DRBD_RCVBUF_SIZE_DEF)
156 __u32_field_def(13, DRBD_GENLA_F_MANDATORY, ko_count, DRBD_KO_COUNT_DEF)
157 __u32_field_def(14, DRBD_GENLA_F_MANDATORY, max_buffers, DRBD_MAX_BUFFERS_DEF)
158 __u32_field_def(15, DRBD_GENLA_F_MANDATORY, max_epoch_size, DRBD_MAX_EPOCH_SIZE_DEF)
159 __u32_field_def(16, DRBD_GENLA_F_MANDATORY, unplug_watermark, DRBD_UNPLUG_WATERMARK_DEF)
160 __u32_field_def(17, DRBD_GENLA_F_MANDATORY, after_sb_0p, DRBD_AFTER_SB_0P_DEF)
161 __u32_field_def(18, DRBD_GENLA_F_MANDATORY, after_sb_1p, DRBD_AFTER_SB_1P_DEF)
162 __u32_field_def(19, DRBD_GENLA_F_MANDATORY, after_sb_2p, DRBD_AFTER_SB_2P_DEF)
163 __u32_field_def(20, DRBD_GENLA_F_MANDATORY, rr_conflict, DRBD_RR_CONFLICT_DEF)
164 __u32_field_def(21, DRBD_GENLA_F_MANDATORY, on_congestion, DRBD_ON_CONGESTION_DEF)
165 __u32_field_def(22, DRBD_GENLA_F_MANDATORY, cong_fill, DRBD_CONG_FILL_DEF)
166 __u32_field_def(23, DRBD_GENLA_F_MANDATORY, cong_extents, DRBD_CONG_EXTENTS_DEF)
167 __flg_field_def(24, DRBD_GENLA_F_MANDATORY, two_primaries, DRBD_ALLOW_TWO_PRIMARIES_DEF)
168 __flg_field(25, DRBD_GENLA_F_MANDATORY | DRBD_F_INVARIANT, discard_my_data)
169 __flg_field_def(26, DRBD_GENLA_F_MANDATORY, tcp_cork, DRBD_TCP_CORK_DEF)
170 __flg_field_def(27, DRBD_GENLA_F_MANDATORY, always_asbp, DRBD_ALWAYS_ASBP_DEF)
171 __flg_field(28, DRBD_GENLA_F_MANDATORY | DRBD_F_INVARIANT, tentative)
172 __flg_field_def(29, DRBD_GENLA_F_MANDATORY, use_rle, DRBD_USE_RLE_DEF)
173 /* 9: __u32_field_def(30, DRBD_GENLA_F_MANDATORY, fencing_policy, DRBD_FENCING_DEF) */
174)
175
176GENL_struct(DRBD_NLA_SET_ROLE_PARMS, 6, set_role_parms,
177 __flg_field(1, DRBD_GENLA_F_MANDATORY, assume_uptodate)
178)
179
180GENL_struct(DRBD_NLA_RESIZE_PARMS, 7, resize_parms,
181 __u64_field(1, DRBD_GENLA_F_MANDATORY, resize_size)
182 __flg_field(2, DRBD_GENLA_F_MANDATORY, resize_force)
183 __flg_field(3, DRBD_GENLA_F_MANDATORY, no_resync)
184)
185
186GENL_struct(DRBD_NLA_STATE_INFO, 8, state_info,
187 /* the reason of the broadcast,
188 * if this is an event triggered broadcast. */
189 __u32_field(1, DRBD_GENLA_F_MANDATORY, sib_reason)
190 __u32_field(2, DRBD_F_REQUIRED, current_state)
191 __u64_field(3, DRBD_GENLA_F_MANDATORY, capacity)
192 __u64_field(4, DRBD_GENLA_F_MANDATORY, ed_uuid)
193
194 /* These are for broadcast from after state change work.
195 * prev_state and new_state are from the moment the state change took
196 * place, new_state is not neccessarily the same as current_state,
197 * there may have been more state changes since. Which will be
198 * broadcasted soon, in their respective after state change work. */
199 __u32_field(5, DRBD_GENLA_F_MANDATORY, prev_state)
200 __u32_field(6, DRBD_GENLA_F_MANDATORY, new_state)
201
202 /* if we have a local disk: */
203 __bin_field(7, DRBD_GENLA_F_MANDATORY, uuids, (UI_SIZE*sizeof(__u64)))
204 __u32_field(8, DRBD_GENLA_F_MANDATORY, disk_flags)
205 __u64_field(9, DRBD_GENLA_F_MANDATORY, bits_total)
206 __u64_field(10, DRBD_GENLA_F_MANDATORY, bits_oos)
207 /* and in case resync or online verify is active */
208 __u64_field(11, DRBD_GENLA_F_MANDATORY, bits_rs_total)
209 __u64_field(12, DRBD_GENLA_F_MANDATORY, bits_rs_failed)
210
211 /* for pre and post notifications of helper execution */
212 __str_field(13, DRBD_GENLA_F_MANDATORY, helper, 32)
213 __u32_field(14, DRBD_GENLA_F_MANDATORY, helper_exit_code)
214
215 __u64_field(15, 0, send_cnt)
216 __u64_field(16, 0, recv_cnt)
217 __u64_field(17, 0, read_cnt)
218 __u64_field(18, 0, writ_cnt)
219 __u64_field(19, 0, al_writ_cnt)
220 __u64_field(20, 0, bm_writ_cnt)
221 __u32_field(21, 0, ap_bio_cnt)
222 __u32_field(22, 0, ap_pending_cnt)
223 __u32_field(23, 0, rs_pending_cnt)
224)
225
226GENL_struct(DRBD_NLA_START_OV_PARMS, 9, start_ov_parms,
227 __u64_field(1, DRBD_GENLA_F_MANDATORY, ov_start_sector)
228 __u64_field(2, DRBD_GENLA_F_MANDATORY, ov_stop_sector)
229)
230
231GENL_struct(DRBD_NLA_NEW_C_UUID_PARMS, 10, new_c_uuid_parms,
232 __flg_field(1, DRBD_GENLA_F_MANDATORY, clear_bm)
233)
234
235GENL_struct(DRBD_NLA_TIMEOUT_PARMS, 11, timeout_parms,
236 __u32_field(1, DRBD_F_REQUIRED, timeout_type)
237)
238
239GENL_struct(DRBD_NLA_DISCONNECT_PARMS, 12, disconnect_parms,
240 __flg_field(1, DRBD_GENLA_F_MANDATORY, force_disconnect)
241)
242
243GENL_struct(DRBD_NLA_DETACH_PARMS, 13, detach_parms,
244 __flg_field(1, DRBD_GENLA_F_MANDATORY, force_detach)
245)
246
247/*
248 * Notifications and commands (genlmsghdr->cmd)
249 */
250GENL_mc_group(events)
251
252 /* kernel -> userspace announcement of changes */
253GENL_notification(
254 DRBD_EVENT, 1, events,
255 GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED)
256 GENL_tla_expected(DRBD_NLA_STATE_INFO, DRBD_F_REQUIRED)
257 GENL_tla_expected(DRBD_NLA_NET_CONF, DRBD_GENLA_F_MANDATORY)
258 GENL_tla_expected(DRBD_NLA_DISK_CONF, DRBD_GENLA_F_MANDATORY)
259 GENL_tla_expected(DRBD_NLA_SYNCER_CONF, DRBD_GENLA_F_MANDATORY)
260)
261
262 /* query kernel for specific or all info */
263GENL_op(
264 DRBD_ADM_GET_STATUS, 2,
265 GENL_op_init(
266 .doit = drbd_adm_get_status,
267 .dumpit = drbd_adm_get_status_all,
268 /* anyone may ask for the status,
269 * it is broadcasted anyways */
270 ),
271 /* To select the object .doit.
272 * Or a subset of objects in .dumpit. */
273 GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_GENLA_F_MANDATORY)
274)
275
276 /* add DRBD minor devices as volumes to resources */
277GENL_op(DRBD_ADM_NEW_MINOR, 5, GENL_doit(drbd_adm_add_minor),
278 GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED))
279GENL_op(DRBD_ADM_DEL_MINOR, 6, GENL_doit(drbd_adm_delete_minor),
280 GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED))
281
282 /* add or delete resources */
283GENL_op(DRBD_ADM_NEW_RESOURCE, 7, GENL_doit(drbd_adm_new_resource),
284 GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED))
285GENL_op(DRBD_ADM_DEL_RESOURCE, 8, GENL_doit(drbd_adm_del_resource),
286 GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED))
287
288GENL_op(DRBD_ADM_RESOURCE_OPTS, 9,
289 GENL_doit(drbd_adm_resource_opts),
290 GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED)
291 GENL_tla_expected(DRBD_NLA_RESOURCE_OPTS, DRBD_GENLA_F_MANDATORY)
292)
293
294GENL_op(
295 DRBD_ADM_CONNECT, 10,
296 GENL_doit(drbd_adm_connect),
297 GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED)
298 GENL_tla_expected(DRBD_NLA_NET_CONF, DRBD_F_REQUIRED)
299)
300
301GENL_op(
302 DRBD_ADM_CHG_NET_OPTS, 29,
303 GENL_doit(drbd_adm_net_opts),
304 GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED)
305 GENL_tla_expected(DRBD_NLA_NET_CONF, DRBD_F_REQUIRED)
306)
307
308GENL_op(DRBD_ADM_DISCONNECT, 11, GENL_doit(drbd_adm_disconnect),
309 GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED))
310
311GENL_op(DRBD_ADM_ATTACH, 12,
312 GENL_doit(drbd_adm_attach),
313 GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED)
314 GENL_tla_expected(DRBD_NLA_DISK_CONF, DRBD_F_REQUIRED)
315)
316
317GENL_op(DRBD_ADM_CHG_DISK_OPTS, 28,
318 GENL_doit(drbd_adm_disk_opts),
319 GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED)
320 GENL_tla_expected(DRBD_NLA_DISK_OPTS, DRBD_F_REQUIRED)
321)
322
323GENL_op(
324 DRBD_ADM_RESIZE, 13,
325 GENL_doit(drbd_adm_resize),
326 GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED)
327 GENL_tla_expected(DRBD_NLA_RESIZE_PARMS, DRBD_GENLA_F_MANDATORY)
328)
329
330GENL_op(
331 DRBD_ADM_PRIMARY, 14,
332 GENL_doit(drbd_adm_set_role),
333 GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED)
334 GENL_tla_expected(DRBD_NLA_SET_ROLE_PARMS, DRBD_F_REQUIRED)
335)
336
337GENL_op(
338 DRBD_ADM_SECONDARY, 15,
339 GENL_doit(drbd_adm_set_role),
340 GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED)
341 GENL_tla_expected(DRBD_NLA_SET_ROLE_PARMS, DRBD_F_REQUIRED)
342)
343
344GENL_op(
345 DRBD_ADM_NEW_C_UUID, 16,
346 GENL_doit(drbd_adm_new_c_uuid),
347 GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED)
348 GENL_tla_expected(DRBD_NLA_NEW_C_UUID_PARMS, DRBD_GENLA_F_MANDATORY)
349)
350
351GENL_op(
352 DRBD_ADM_START_OV, 17,
353 GENL_doit(drbd_adm_start_ov),
354 GENL_tla_expected(DRBD_NLA_START_OV_PARMS, DRBD_GENLA_F_MANDATORY)
355)
356
357GENL_op(DRBD_ADM_DETACH, 18, GENL_doit(drbd_adm_detach),
358 GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED)
359 GENL_tla_expected(DRBD_NLA_DETACH_PARMS, DRBD_GENLA_F_MANDATORY))
360
361GENL_op(DRBD_ADM_INVALIDATE, 19, GENL_doit(drbd_adm_invalidate),
362 GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED))
363GENL_op(DRBD_ADM_INVAL_PEER, 20, GENL_doit(drbd_adm_invalidate_peer),
364 GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED))
365GENL_op(DRBD_ADM_PAUSE_SYNC, 21, GENL_doit(drbd_adm_pause_sync),
366 GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED))
367GENL_op(DRBD_ADM_RESUME_SYNC, 22, GENL_doit(drbd_adm_resume_sync),
368 GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED))
369GENL_op(DRBD_ADM_SUSPEND_IO, 23, GENL_doit(drbd_adm_suspend_io),
370 GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED))
371GENL_op(DRBD_ADM_RESUME_IO, 24, GENL_doit(drbd_adm_resume_io),
372 GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED))
373GENL_op(DRBD_ADM_OUTDATE, 25, GENL_doit(drbd_adm_outdate),
374 GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED))
375GENL_op(DRBD_ADM_GET_TIMEOUT_TYPE, 26, GENL_doit(drbd_adm_get_timeout_type),
376 GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED))
377GENL_op(DRBD_ADM_DOWN, 27, GENL_doit(drbd_adm_down),
378 GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED))
diff --git a/include/linux/drbd_genl_api.h b/include/linux/drbd_genl_api.h
new file mode 100644
index 000000000000..9ef50d51e34e
--- /dev/null
+++ b/include/linux/drbd_genl_api.h
@@ -0,0 +1,55 @@
1#ifndef DRBD_GENL_STRUCT_H
2#define DRBD_GENL_STRUCT_H
3
4/**
5 * struct drbd_genlmsghdr - DRBD specific header used in NETLINK_GENERIC requests
6 * @minor:
7 * For admin requests (user -> kernel): which minor device to operate on.
8 * For (unicast) replies or informational (broadcast) messages
9 * (kernel -> user): which minor device the information is about.
10 * If we do not operate on minors, but on connections or resources,
11 * the minor value shall be (~0), and the attribute DRBD_NLA_CFG_CONTEXT
12 * is used instead.
13 * @flags: possible operation modifiers (relevant only for user->kernel):
14 * DRBD_GENL_F_SET_DEFAULTS
15 * @volume:
16 * When creating a new minor (adding it to a resource), the resource needs
17 * to know which volume number within the resource this is supposed to be.
18 * The volume number corresponds to the same volume number on the remote side,
19 * whereas the minor number on the remote side may be different
20 * (union with flags).
21 * @ret_code: kernel->userland unicast cfg reply return code (union with flags);
22 */
23struct drbd_genlmsghdr {
24 __u32 minor;
25 union {
26 __u32 flags;
27 __s32 ret_code;
28 };
29};
30
31/* To be used in drbd_genlmsghdr.flags */
32enum {
33 DRBD_GENL_F_SET_DEFAULTS = 1,
34};
35
36enum drbd_state_info_bcast_reason {
37 SIB_GET_STATUS_REPLY = 1,
38 SIB_STATE_CHANGE = 2,
39 SIB_HELPER_PRE = 3,
40 SIB_HELPER_POST = 4,
41 SIB_SYNC_PROGRESS = 5,
42};
43
44/* hack around predefined gcc/cpp "linux=1",
45 * we cannot possibly include <1/drbd_genl.h> */
46#undef linux
47
48#include <linux/drbd.h>
49#define GENL_MAGIC_VERSION API_VERSION
50#define GENL_MAGIC_FAMILY drbd
51#define GENL_MAGIC_FAMILY_HDRSZ sizeof(struct drbd_genlmsghdr)
52#define GENL_MAGIC_INCLUDE_FILE <linux/drbd_genl.h>
53#include <linux/genl_magic_struct.h>
54
55#endif
diff --git a/include/linux/drbd_limits.h b/include/linux/drbd_limits.h
index fb670bf603f7..1fa19c5f5e64 100644
--- a/include/linux/drbd_limits.h
+++ b/include/linux/drbd_limits.h
@@ -16,29 +16,37 @@
16#define DEBUG_RANGE_CHECK 0 16#define DEBUG_RANGE_CHECK 0
17 17
18#define DRBD_MINOR_COUNT_MIN 1 18#define DRBD_MINOR_COUNT_MIN 1
19#define DRBD_MINOR_COUNT_MAX 256 19#define DRBD_MINOR_COUNT_MAX 255
20#define DRBD_MINOR_COUNT_DEF 32 20#define DRBD_MINOR_COUNT_DEF 32
21#define DRBD_MINOR_COUNT_SCALE '1'
22
23#define DRBD_VOLUME_MAX 65535
21 24
22#define DRBD_DIALOG_REFRESH_MIN 0 25#define DRBD_DIALOG_REFRESH_MIN 0
23#define DRBD_DIALOG_REFRESH_MAX 600 26#define DRBD_DIALOG_REFRESH_MAX 600
27#define DRBD_DIALOG_REFRESH_SCALE '1'
24 28
25/* valid port number */ 29/* valid port number */
26#define DRBD_PORT_MIN 1 30#define DRBD_PORT_MIN 1
27#define DRBD_PORT_MAX 0xffff 31#define DRBD_PORT_MAX 0xffff
32#define DRBD_PORT_SCALE '1'
28 33
29/* startup { */ 34/* startup { */
30 /* if you want more than 3.4 days, disable */ 35 /* if you want more than 3.4 days, disable */
31#define DRBD_WFC_TIMEOUT_MIN 0 36#define DRBD_WFC_TIMEOUT_MIN 0
32#define DRBD_WFC_TIMEOUT_MAX 300000 37#define DRBD_WFC_TIMEOUT_MAX 300000
33#define DRBD_WFC_TIMEOUT_DEF 0 38#define DRBD_WFC_TIMEOUT_DEF 0
39#define DRBD_WFC_TIMEOUT_SCALE '1'
34 40
35#define DRBD_DEGR_WFC_TIMEOUT_MIN 0 41#define DRBD_DEGR_WFC_TIMEOUT_MIN 0
36#define DRBD_DEGR_WFC_TIMEOUT_MAX 300000 42#define DRBD_DEGR_WFC_TIMEOUT_MAX 300000
37#define DRBD_DEGR_WFC_TIMEOUT_DEF 0 43#define DRBD_DEGR_WFC_TIMEOUT_DEF 0
44#define DRBD_DEGR_WFC_TIMEOUT_SCALE '1'
38 45
39#define DRBD_OUTDATED_WFC_TIMEOUT_MIN 0 46#define DRBD_OUTDATED_WFC_TIMEOUT_MIN 0
40#define DRBD_OUTDATED_WFC_TIMEOUT_MAX 300000 47#define DRBD_OUTDATED_WFC_TIMEOUT_MAX 300000
41#define DRBD_OUTDATED_WFC_TIMEOUT_DEF 0 48#define DRBD_OUTDATED_WFC_TIMEOUT_DEF 0
49#define DRBD_OUTDATED_WFC_TIMEOUT_SCALE '1'
42/* }*/ 50/* }*/
43 51
44/* net { */ 52/* net { */
@@ -47,75 +55,91 @@
47#define DRBD_TIMEOUT_MIN 1 55#define DRBD_TIMEOUT_MIN 1
48#define DRBD_TIMEOUT_MAX 600 56#define DRBD_TIMEOUT_MAX 600
49#define DRBD_TIMEOUT_DEF 60 /* 6 seconds */ 57#define DRBD_TIMEOUT_DEF 60 /* 6 seconds */
58#define DRBD_TIMEOUT_SCALE '1'
50 59
51 /* If backing disk takes longer than disk_timeout, mark the disk as failed */ 60 /* If backing disk takes longer than disk_timeout, mark the disk as failed */
52#define DRBD_DISK_TIMEOUT_MIN 0 /* 0 = disabled */ 61#define DRBD_DISK_TIMEOUT_MIN 0 /* 0 = disabled */
53#define DRBD_DISK_TIMEOUT_MAX 6000 /* 10 Minutes */ 62#define DRBD_DISK_TIMEOUT_MAX 6000 /* 10 Minutes */
54#define DRBD_DISK_TIMEOUT_DEF 0 /* disabled */ 63#define DRBD_DISK_TIMEOUT_DEF 0 /* disabled */
64#define DRBD_DISK_TIMEOUT_SCALE '1'
55 65
56 /* active connection retries when C_WF_CONNECTION */ 66 /* active connection retries when C_WF_CONNECTION */
57#define DRBD_CONNECT_INT_MIN 1 67#define DRBD_CONNECT_INT_MIN 1
58#define DRBD_CONNECT_INT_MAX 120 68#define DRBD_CONNECT_INT_MAX 120
59#define DRBD_CONNECT_INT_DEF 10 /* seconds */ 69#define DRBD_CONNECT_INT_DEF 10 /* seconds */
70#define DRBD_CONNECT_INT_SCALE '1'
60 71
61 /* keep-alive probes when idle */ 72 /* keep-alive probes when idle */
62#define DRBD_PING_INT_MIN 1 73#define DRBD_PING_INT_MIN 1
63#define DRBD_PING_INT_MAX 120 74#define DRBD_PING_INT_MAX 120
64#define DRBD_PING_INT_DEF 10 75#define DRBD_PING_INT_DEF 10
76#define DRBD_PING_INT_SCALE '1'
65 77
66 /* timeout for the ping packets.*/ 78 /* timeout for the ping packets.*/
67#define DRBD_PING_TIMEO_MIN 1 79#define DRBD_PING_TIMEO_MIN 1
68#define DRBD_PING_TIMEO_MAX 300 80#define DRBD_PING_TIMEO_MAX 300
69#define DRBD_PING_TIMEO_DEF 5 81#define DRBD_PING_TIMEO_DEF 5
82#define DRBD_PING_TIMEO_SCALE '1'
70 83
71 /* max number of write requests between write barriers */ 84 /* max number of write requests between write barriers */
72#define DRBD_MAX_EPOCH_SIZE_MIN 1 85#define DRBD_MAX_EPOCH_SIZE_MIN 1
73#define DRBD_MAX_EPOCH_SIZE_MAX 20000 86#define DRBD_MAX_EPOCH_SIZE_MAX 20000
74#define DRBD_MAX_EPOCH_SIZE_DEF 2048 87#define DRBD_MAX_EPOCH_SIZE_DEF 2048
88#define DRBD_MAX_EPOCH_SIZE_SCALE '1'
75 89
76 /* I don't think that a tcp send buffer of more than 10M is useful */ 90 /* I don't think that a tcp send buffer of more than 10M is useful */
77#define DRBD_SNDBUF_SIZE_MIN 0 91#define DRBD_SNDBUF_SIZE_MIN 0
78#define DRBD_SNDBUF_SIZE_MAX (10<<20) 92#define DRBD_SNDBUF_SIZE_MAX (10<<20)
79#define DRBD_SNDBUF_SIZE_DEF 0 93#define DRBD_SNDBUF_SIZE_DEF 0
94#define DRBD_SNDBUF_SIZE_SCALE '1'
80 95
81#define DRBD_RCVBUF_SIZE_MIN 0 96#define DRBD_RCVBUF_SIZE_MIN 0
82#define DRBD_RCVBUF_SIZE_MAX (10<<20) 97#define DRBD_RCVBUF_SIZE_MAX (10<<20)
83#define DRBD_RCVBUF_SIZE_DEF 0 98#define DRBD_RCVBUF_SIZE_DEF 0
99#define DRBD_RCVBUF_SIZE_SCALE '1'
84 100
85 /* @4k PageSize -> 128kB - 512MB */ 101 /* @4k PageSize -> 128kB - 512MB */
86#define DRBD_MAX_BUFFERS_MIN 32 102#define DRBD_MAX_BUFFERS_MIN 32
87#define DRBD_MAX_BUFFERS_MAX 131072 103#define DRBD_MAX_BUFFERS_MAX 131072
88#define DRBD_MAX_BUFFERS_DEF 2048 104#define DRBD_MAX_BUFFERS_DEF 2048
105#define DRBD_MAX_BUFFERS_SCALE '1'
89 106
90 /* @4k PageSize -> 4kB - 512MB */ 107 /* @4k PageSize -> 4kB - 512MB */
91#define DRBD_UNPLUG_WATERMARK_MIN 1 108#define DRBD_UNPLUG_WATERMARK_MIN 1
92#define DRBD_UNPLUG_WATERMARK_MAX 131072 109#define DRBD_UNPLUG_WATERMARK_MAX 131072
93#define DRBD_UNPLUG_WATERMARK_DEF (DRBD_MAX_BUFFERS_DEF/16) 110#define DRBD_UNPLUG_WATERMARK_DEF (DRBD_MAX_BUFFERS_DEF/16)
111#define DRBD_UNPLUG_WATERMARK_SCALE '1'
94 112
95 /* 0 is disabled. 113 /* 0 is disabled.
96 * 200 should be more than enough even for very short timeouts */ 114 * 200 should be more than enough even for very short timeouts */
97#define DRBD_KO_COUNT_MIN 0 115#define DRBD_KO_COUNT_MIN 0
98#define DRBD_KO_COUNT_MAX 200 116#define DRBD_KO_COUNT_MAX 200
99#define DRBD_KO_COUNT_DEF 0 117#define DRBD_KO_COUNT_DEF 7
118#define DRBD_KO_COUNT_SCALE '1'
100/* } */ 119/* } */
101 120
102/* syncer { */ 121/* syncer { */
103 /* FIXME allow rate to be zero? */ 122 /* FIXME allow rate to be zero? */
104#define DRBD_RATE_MIN 1 123#define DRBD_RESYNC_RATE_MIN 1
105/* channel bonding 10 GbE, or other hardware */ 124/* channel bonding 10 GbE, or other hardware */
106#define DRBD_RATE_MAX (4 << 20) 125#define DRBD_RESYNC_RATE_MAX (4 << 20)
107#define DRBD_RATE_DEF 250 /* kb/second */ 126#define DRBD_RESYNC_RATE_DEF 250
127#define DRBD_RESYNC_RATE_SCALE 'k' /* kilobytes */
108 128
109 /* less than 7 would hit performance unnecessarily. 129 /* less than 7 would hit performance unnecessarily.
110 * 3833 is the largest prime that still does fit 130 * 919 slots context information per transaction,
111 * into 64 sectors of activity log */ 131 * 32k activity log, 4k transaction size,
132 * one transaction in flight:
133 * 919 * 7 = 6433 */
112#define DRBD_AL_EXTENTS_MIN 7 134#define DRBD_AL_EXTENTS_MIN 7
113#define DRBD_AL_EXTENTS_MAX 3833 135#define DRBD_AL_EXTENTS_MAX 6433
114#define DRBD_AL_EXTENTS_DEF 127 136#define DRBD_AL_EXTENTS_DEF 1237
137#define DRBD_AL_EXTENTS_SCALE '1'
115 138
116#define DRBD_AFTER_MIN -1 139#define DRBD_MINOR_NUMBER_MIN -1
117#define DRBD_AFTER_MAX 255 140#define DRBD_MINOR_NUMBER_MAX ((1 << 20) - 1)
118#define DRBD_AFTER_DEF -1 141#define DRBD_MINOR_NUMBER_DEF -1
142#define DRBD_MINOR_NUMBER_SCALE '1'
119 143
120/* } */ 144/* } */
121 145
@@ -124,11 +148,12 @@
124 * the upper limit with 64bit kernel, enough ram and flexible meta data 148 * the upper limit with 64bit kernel, enough ram and flexible meta data
125 * is 1 PiB, currently. */ 149 * is 1 PiB, currently. */
126/* DRBD_MAX_SECTORS */ 150/* DRBD_MAX_SECTORS */
127#define DRBD_DISK_SIZE_SECT_MIN 0 151#define DRBD_DISK_SIZE_MIN 0
128#define DRBD_DISK_SIZE_SECT_MAX (1 * (2LLU << 40)) 152#define DRBD_DISK_SIZE_MAX (1 * (2LLU << 40))
129#define DRBD_DISK_SIZE_SECT_DEF 0 /* = disabled = no user size... */ 153#define DRBD_DISK_SIZE_DEF 0 /* = disabled = no user size... */
154#define DRBD_DISK_SIZE_SCALE 's' /* sectors */
130 155
131#define DRBD_ON_IO_ERROR_DEF EP_PASS_ON 156#define DRBD_ON_IO_ERROR_DEF EP_DETACH
132#define DRBD_FENCING_DEF FP_DONT_CARE 157#define DRBD_FENCING_DEF FP_DONT_CARE
133#define DRBD_AFTER_SB_0P_DEF ASB_DISCONNECT 158#define DRBD_AFTER_SB_0P_DEF ASB_DISCONNECT
134#define DRBD_AFTER_SB_1P_DEF ASB_DISCONNECT 159#define DRBD_AFTER_SB_1P_DEF ASB_DISCONNECT
@@ -136,38 +161,59 @@
136#define DRBD_RR_CONFLICT_DEF ASB_DISCONNECT 161#define DRBD_RR_CONFLICT_DEF ASB_DISCONNECT
137#define DRBD_ON_NO_DATA_DEF OND_IO_ERROR 162#define DRBD_ON_NO_DATA_DEF OND_IO_ERROR
138#define DRBD_ON_CONGESTION_DEF OC_BLOCK 163#define DRBD_ON_CONGESTION_DEF OC_BLOCK
164#define DRBD_READ_BALANCING_DEF RB_PREFER_LOCAL
139 165
140#define DRBD_MAX_BIO_BVECS_MIN 0 166#define DRBD_MAX_BIO_BVECS_MIN 0
141#define DRBD_MAX_BIO_BVECS_MAX 128 167#define DRBD_MAX_BIO_BVECS_MAX 128
142#define DRBD_MAX_BIO_BVECS_DEF 0 168#define DRBD_MAX_BIO_BVECS_DEF 0
169#define DRBD_MAX_BIO_BVECS_SCALE '1'
143 170
144#define DRBD_C_PLAN_AHEAD_MIN 0 171#define DRBD_C_PLAN_AHEAD_MIN 0
145#define DRBD_C_PLAN_AHEAD_MAX 300 172#define DRBD_C_PLAN_AHEAD_MAX 300
146#define DRBD_C_PLAN_AHEAD_DEF 0 /* RS rate controller disabled by default */ 173#define DRBD_C_PLAN_AHEAD_DEF 20
174#define DRBD_C_PLAN_AHEAD_SCALE '1'
147 175
148#define DRBD_C_DELAY_TARGET_MIN 1 176#define DRBD_C_DELAY_TARGET_MIN 1
149#define DRBD_C_DELAY_TARGET_MAX 100 177#define DRBD_C_DELAY_TARGET_MAX 100
150#define DRBD_C_DELAY_TARGET_DEF 10 178#define DRBD_C_DELAY_TARGET_DEF 10
179#define DRBD_C_DELAY_TARGET_SCALE '1'
151 180
152#define DRBD_C_FILL_TARGET_MIN 0 181#define DRBD_C_FILL_TARGET_MIN 0
153#define DRBD_C_FILL_TARGET_MAX (1<<20) /* 500MByte in sec */ 182#define DRBD_C_FILL_TARGET_MAX (1<<20) /* 500MByte in sec */
154#define DRBD_C_FILL_TARGET_DEF 0 /* By default disabled -> controlled by delay_target */ 183#define DRBD_C_FILL_TARGET_DEF 100 /* Try to place 50KiB in socket send buffer during resync */
184#define DRBD_C_FILL_TARGET_SCALE 's' /* sectors */
155 185
156#define DRBD_C_MAX_RATE_MIN 250 /* kByte/sec */ 186#define DRBD_C_MAX_RATE_MIN 250
157#define DRBD_C_MAX_RATE_MAX (4 << 20) 187#define DRBD_C_MAX_RATE_MAX (4 << 20)
158#define DRBD_C_MAX_RATE_DEF 102400 188#define DRBD_C_MAX_RATE_DEF 102400
189#define DRBD_C_MAX_RATE_SCALE 'k' /* kilobytes */
159 190
160#define DRBD_C_MIN_RATE_MIN 0 /* kByte/sec */ 191#define DRBD_C_MIN_RATE_MIN 0
161#define DRBD_C_MIN_RATE_MAX (4 << 20) 192#define DRBD_C_MIN_RATE_MAX (4 << 20)
162#define DRBD_C_MIN_RATE_DEF 4096 193#define DRBD_C_MIN_RATE_DEF 250
194#define DRBD_C_MIN_RATE_SCALE 'k' /* kilobytes */
163 195
164#define DRBD_CONG_FILL_MIN 0 196#define DRBD_CONG_FILL_MIN 0
165#define DRBD_CONG_FILL_MAX (10<<21) /* 10GByte in sectors */ 197#define DRBD_CONG_FILL_MAX (10<<21) /* 10GByte in sectors */
166#define DRBD_CONG_FILL_DEF 0 198#define DRBD_CONG_FILL_DEF 0
199#define DRBD_CONG_FILL_SCALE 's' /* sectors */
167 200
168#define DRBD_CONG_EXTENTS_MIN DRBD_AL_EXTENTS_MIN 201#define DRBD_CONG_EXTENTS_MIN DRBD_AL_EXTENTS_MIN
169#define DRBD_CONG_EXTENTS_MAX DRBD_AL_EXTENTS_MAX 202#define DRBD_CONG_EXTENTS_MAX DRBD_AL_EXTENTS_MAX
170#define DRBD_CONG_EXTENTS_DEF DRBD_AL_EXTENTS_DEF 203#define DRBD_CONG_EXTENTS_DEF DRBD_AL_EXTENTS_DEF
204#define DRBD_CONG_EXTENTS_SCALE DRBD_AL_EXTENTS_SCALE
205
206#define DRBD_PROTOCOL_DEF DRBD_PROT_C
207
208#define DRBD_DISK_BARRIER_DEF 0
209#define DRBD_DISK_FLUSHES_DEF 1
210#define DRBD_DISK_DRAIN_DEF 1
211#define DRBD_MD_FLUSHES_DEF 1
212#define DRBD_TCP_CORK_DEF 1
213#define DRBD_AL_UPDATES_DEF 1
214
215#define DRBD_ALLOW_TWO_PRIMARIES_DEF 0
216#define DRBD_ALWAYS_ASBP_DEF 0
217#define DRBD_USE_RLE_DEF 1
171 218
172#undef RANGE
173#endif 219#endif
diff --git a/include/linux/drbd_nl.h b/include/linux/drbd_nl.h
deleted file mode 100644
index a8706f08ab36..000000000000
--- a/include/linux/drbd_nl.h
+++ /dev/null
@@ -1,163 +0,0 @@
1/*
2 PAKET( name,
3 TYPE ( pn, pr, member )
4 ...
5 )
6
7 You may never reissue one of the pn arguments
8*/
9
10#if !defined(NL_PACKET) || !defined(NL_STRING) || !defined(NL_INTEGER) || !defined(NL_BIT) || !defined(NL_INT64)
11#error "The macros NL_PACKET, NL_STRING, NL_INTEGER, NL_INT64 and NL_BIT needs to be defined"
12#endif
13
14NL_PACKET(primary, 1,
15 NL_BIT( 1, T_MAY_IGNORE, primary_force)
16)
17
18NL_PACKET(secondary, 2, )
19
20NL_PACKET(disk_conf, 3,
21 NL_INT64( 2, T_MAY_IGNORE, disk_size)
22 NL_STRING( 3, T_MANDATORY, backing_dev, 128)
23 NL_STRING( 4, T_MANDATORY, meta_dev, 128)
24 NL_INTEGER( 5, T_MANDATORY, meta_dev_idx)
25 NL_INTEGER( 6, T_MAY_IGNORE, on_io_error)
26 NL_INTEGER( 7, T_MAY_IGNORE, fencing)
27 NL_BIT( 37, T_MAY_IGNORE, use_bmbv)
28 NL_BIT( 53, T_MAY_IGNORE, no_disk_flush)
29 NL_BIT( 54, T_MAY_IGNORE, no_md_flush)
30 /* 55 max_bio_size was available in 8.2.6rc2 */
31 NL_INTEGER( 56, T_MAY_IGNORE, max_bio_bvecs)
32 NL_BIT( 57, T_MAY_IGNORE, no_disk_barrier)
33 NL_BIT( 58, T_MAY_IGNORE, no_disk_drain)
34 NL_INTEGER( 89, T_MAY_IGNORE, disk_timeout)
35)
36
37NL_PACKET(detach, 4,
38 NL_BIT( 88, T_MANDATORY, detach_force)
39)
40
41NL_PACKET(net_conf, 5,
42 NL_STRING( 8, T_MANDATORY, my_addr, 128)
43 NL_STRING( 9, T_MANDATORY, peer_addr, 128)
44 NL_STRING( 10, T_MAY_IGNORE, shared_secret, SHARED_SECRET_MAX)
45 NL_STRING( 11, T_MAY_IGNORE, cram_hmac_alg, SHARED_SECRET_MAX)
46 NL_STRING( 44, T_MAY_IGNORE, integrity_alg, SHARED_SECRET_MAX)
47 NL_INTEGER( 14, T_MAY_IGNORE, timeout)
48 NL_INTEGER( 15, T_MANDATORY, wire_protocol)
49 NL_INTEGER( 16, T_MAY_IGNORE, try_connect_int)
50 NL_INTEGER( 17, T_MAY_IGNORE, ping_int)
51 NL_INTEGER( 18, T_MAY_IGNORE, max_epoch_size)
52 NL_INTEGER( 19, T_MAY_IGNORE, max_buffers)
53 NL_INTEGER( 20, T_MAY_IGNORE, unplug_watermark)
54 NL_INTEGER( 21, T_MAY_IGNORE, sndbuf_size)
55 NL_INTEGER( 22, T_MAY_IGNORE, ko_count)
56 NL_INTEGER( 24, T_MAY_IGNORE, after_sb_0p)
57 NL_INTEGER( 25, T_MAY_IGNORE, after_sb_1p)
58 NL_INTEGER( 26, T_MAY_IGNORE, after_sb_2p)
59 NL_INTEGER( 39, T_MAY_IGNORE, rr_conflict)
60 NL_INTEGER( 40, T_MAY_IGNORE, ping_timeo)
61 NL_INTEGER( 67, T_MAY_IGNORE, rcvbuf_size)
62 NL_INTEGER( 81, T_MAY_IGNORE, on_congestion)
63 NL_INTEGER( 82, T_MAY_IGNORE, cong_fill)
64 NL_INTEGER( 83, T_MAY_IGNORE, cong_extents)
65 /* 59 addr_family was available in GIT, never released */
66 NL_BIT( 60, T_MANDATORY, mind_af)
67 NL_BIT( 27, T_MAY_IGNORE, want_lose)
68 NL_BIT( 28, T_MAY_IGNORE, two_primaries)
69 NL_BIT( 41, T_MAY_IGNORE, always_asbp)
70 NL_BIT( 61, T_MAY_IGNORE, no_cork)
71 NL_BIT( 62, T_MANDATORY, auto_sndbuf_size)
72 NL_BIT( 70, T_MANDATORY, dry_run)
73)
74
75NL_PACKET(disconnect, 6,
76 NL_BIT( 84, T_MAY_IGNORE, force)
77)
78
79NL_PACKET(resize, 7,
80 NL_INT64( 29, T_MAY_IGNORE, resize_size)
81 NL_BIT( 68, T_MAY_IGNORE, resize_force)
82 NL_BIT( 69, T_MANDATORY, no_resync)
83)
84
85NL_PACKET(syncer_conf, 8,
86 NL_INTEGER( 30, T_MAY_IGNORE, rate)
87 NL_INTEGER( 31, T_MAY_IGNORE, after)
88 NL_INTEGER( 32, T_MAY_IGNORE, al_extents)
89/* NL_INTEGER( 71, T_MAY_IGNORE, dp_volume)
90 * NL_INTEGER( 72, T_MAY_IGNORE, dp_interval)
91 * NL_INTEGER( 73, T_MAY_IGNORE, throttle_th)
92 * NL_INTEGER( 74, T_MAY_IGNORE, hold_off_th)
93 * feature will be reimplemented differently with 8.3.9 */
94 NL_STRING( 52, T_MAY_IGNORE, verify_alg, SHARED_SECRET_MAX)
95 NL_STRING( 51, T_MAY_IGNORE, cpu_mask, 32)
96 NL_STRING( 64, T_MAY_IGNORE, csums_alg, SHARED_SECRET_MAX)
97 NL_BIT( 65, T_MAY_IGNORE, use_rle)
98 NL_INTEGER( 75, T_MAY_IGNORE, on_no_data)
99 NL_INTEGER( 76, T_MAY_IGNORE, c_plan_ahead)
100 NL_INTEGER( 77, T_MAY_IGNORE, c_delay_target)
101 NL_INTEGER( 78, T_MAY_IGNORE, c_fill_target)
102 NL_INTEGER( 79, T_MAY_IGNORE, c_max_rate)
103 NL_INTEGER( 80, T_MAY_IGNORE, c_min_rate)
104)
105
106NL_PACKET(invalidate, 9, )
107NL_PACKET(invalidate_peer, 10, )
108NL_PACKET(pause_sync, 11, )
109NL_PACKET(resume_sync, 12, )
110NL_PACKET(suspend_io, 13, )
111NL_PACKET(resume_io, 14, )
112NL_PACKET(outdate, 15, )
113NL_PACKET(get_config, 16, )
114NL_PACKET(get_state, 17,
115 NL_INTEGER( 33, T_MAY_IGNORE, state_i)
116)
117
118NL_PACKET(get_uuids, 18,
119 NL_STRING( 34, T_MAY_IGNORE, uuids, (UI_SIZE*sizeof(__u64)))
120 NL_INTEGER( 35, T_MAY_IGNORE, uuids_flags)
121)
122
123NL_PACKET(get_timeout_flag, 19,
124 NL_BIT( 36, T_MAY_IGNORE, use_degraded)
125)
126
127NL_PACKET(call_helper, 20,
128 NL_STRING( 38, T_MAY_IGNORE, helper, 32)
129)
130
131/* Tag nr 42 already allocated in drbd-8.1 development. */
132
133NL_PACKET(sync_progress, 23,
134 NL_INTEGER( 43, T_MAY_IGNORE, sync_progress)
135)
136
137NL_PACKET(dump_ee, 24,
138 NL_STRING( 45, T_MAY_IGNORE, dump_ee_reason, 32)
139 NL_STRING( 46, T_MAY_IGNORE, seen_digest, SHARED_SECRET_MAX)
140 NL_STRING( 47, T_MAY_IGNORE, calc_digest, SHARED_SECRET_MAX)
141 NL_INT64( 48, T_MAY_IGNORE, ee_sector)
142 NL_INT64( 49, T_MAY_IGNORE, ee_block_id)
143 NL_STRING( 50, T_MAY_IGNORE, ee_data, 32 << 10)
144)
145
146NL_PACKET(start_ov, 25,
147 NL_INT64( 66, T_MAY_IGNORE, start_sector)
148)
149
150NL_PACKET(new_c_uuid, 26,
151 NL_BIT( 63, T_MANDATORY, clear_bm)
152)
153
154#ifdef NL_RESPONSE
155NL_RESPONSE(return_code_only, 27)
156#endif
157
158#undef NL_PACKET
159#undef NL_INTEGER
160#undef NL_INT64
161#undef NL_BIT
162#undef NL_STRING
163#undef NL_RESPONSE
diff --git a/include/linux/drbd_tag_magic.h b/include/linux/drbd_tag_magic.h
deleted file mode 100644
index 82de1f9e48b1..000000000000
--- a/include/linux/drbd_tag_magic.h
+++ /dev/null
@@ -1,84 +0,0 @@
1#ifndef DRBD_TAG_MAGIC_H
2#define DRBD_TAG_MAGIC_H
3
4#define TT_END 0
5#define TT_REMOVED 0xE000
6
7/* declare packet_type enums */
8enum packet_types {
9#define NL_PACKET(name, number, fields) P_ ## name = number,
10#define NL_RESPONSE(name, number) P_ ## name = number,
11#define NL_INTEGER(pn, pr, member)
12#define NL_INT64(pn, pr, member)
13#define NL_BIT(pn, pr, member)
14#define NL_STRING(pn, pr, member, len)
15#include <linux/drbd_nl.h>
16 P_nl_after_last_packet,
17};
18
19/* These struct are used to deduce the size of the tag lists: */
20#define NL_PACKET(name, number, fields) \
21 struct name ## _tag_len_struct { fields };
22#define NL_INTEGER(pn, pr, member) \
23 int member; int tag_and_len ## member;
24#define NL_INT64(pn, pr, member) \
25 __u64 member; int tag_and_len ## member;
26#define NL_BIT(pn, pr, member) \
27 unsigned char member:1; int tag_and_len ## member;
28#define NL_STRING(pn, pr, member, len) \
29 unsigned char member[len]; int member ## _len; \
30 int tag_and_len ## member;
31#include <linux/drbd_nl.h>
32
33/* declare tag-list-sizes */
34static const int tag_list_sizes[] = {
35#define NL_PACKET(name, number, fields) 2 fields ,
36#define NL_INTEGER(pn, pr, member) + 4 + 4
37#define NL_INT64(pn, pr, member) + 4 + 8
38#define NL_BIT(pn, pr, member) + 4 + 1
39#define NL_STRING(pn, pr, member, len) + 4 + (len)
40#include <linux/drbd_nl.h>
41};
42
43/* The two highest bits are used for the tag type */
44#define TT_MASK 0xC000
45#define TT_INTEGER 0x0000
46#define TT_INT64 0x4000
47#define TT_BIT 0x8000
48#define TT_STRING 0xC000
49/* The next bit indicates if processing of the tag is mandatory */
50#define T_MANDATORY 0x2000
51#define T_MAY_IGNORE 0x0000
52#define TN_MASK 0x1fff
53/* The remaining 13 bits are used to enumerate the tags */
54
55#define tag_type(T) ((T) & TT_MASK)
56#define tag_number(T) ((T) & TN_MASK)
57
58/* declare tag enums */
59#define NL_PACKET(name, number, fields) fields
60enum drbd_tags {
61#define NL_INTEGER(pn, pr, member) T_ ## member = pn | TT_INTEGER | pr ,
62#define NL_INT64(pn, pr, member) T_ ## member = pn | TT_INT64 | pr ,
63#define NL_BIT(pn, pr, member) T_ ## member = pn | TT_BIT | pr ,
64#define NL_STRING(pn, pr, member, len) T_ ## member = pn | TT_STRING | pr ,
65#include <linux/drbd_nl.h>
66};
67
68struct tag {
69 const char *name;
70 int type_n_flags;
71 int max_len;
72};
73
74/* declare tag names */
75#define NL_PACKET(name, number, fields) fields
76static const struct tag tag_descriptions[] = {
77#define NL_INTEGER(pn, pr, member) [ pn ] = { #member, TT_INTEGER | pr, sizeof(int) },
78#define NL_INT64(pn, pr, member) [ pn ] = { #member, TT_INT64 | pr, sizeof(__u64) },
79#define NL_BIT(pn, pr, member) [ pn ] = { #member, TT_BIT | pr, sizeof(int) },
80#define NL_STRING(pn, pr, member, len) [ pn ] = { #member, TT_STRING | pr, (len) },
81#include <linux/drbd_nl.h>
82};
83
84#endif
diff --git a/include/linux/genhd.h b/include/linux/genhd.h
index 4f440b3e89fe..79b8bba19363 100644
--- a/include/linux/genhd.h
+++ b/include/linux/genhd.h
@@ -88,10 +88,14 @@ struct disk_stats {
88}; 88};
89 89
90#define PARTITION_META_INFO_VOLNAMELTH 64 90#define PARTITION_META_INFO_VOLNAMELTH 64
91#define PARTITION_META_INFO_UUIDLTH 16 91/*
92 * Enough for the string representation of any kind of UUID plus NULL.
93 * EFI UUID is 36 characters. MSDOS UUID is 11 characters.
94 */
95#define PARTITION_META_INFO_UUIDLTH 37
92 96
93struct partition_meta_info { 97struct partition_meta_info {
94 u8 uuid[PARTITION_META_INFO_UUIDLTH]; /* always big endian */ 98 char uuid[PARTITION_META_INFO_UUIDLTH];
95 u8 volname[PARTITION_META_INFO_VOLNAMELTH]; 99 u8 volname[PARTITION_META_INFO_VOLNAMELTH];
96}; 100};
97 101
diff --git a/include/linux/genl_magic_func.h b/include/linux/genl_magic_func.h
new file mode 100644
index 000000000000..023bc346b877
--- /dev/null
+++ b/include/linux/genl_magic_func.h
@@ -0,0 +1,422 @@
1#ifndef GENL_MAGIC_FUNC_H
2#define GENL_MAGIC_FUNC_H
3
4#include <linux/genl_magic_struct.h>
5
6/*
7 * Magic: declare tla policy {{{1
8 * Magic: declare nested policies
9 * {{{2
10 */
11#undef GENL_mc_group
12#define GENL_mc_group(group)
13
14#undef GENL_notification
15#define GENL_notification(op_name, op_num, mcast_group, tla_list)
16
17#undef GENL_op
18#define GENL_op(op_name, op_num, handler, tla_list)
19
20#undef GENL_struct
21#define GENL_struct(tag_name, tag_number, s_name, s_fields) \
22 [tag_name] = { .type = NLA_NESTED },
23
24static struct nla_policy CONCAT_(GENL_MAGIC_FAMILY, _tla_nl_policy)[] = {
25#include GENL_MAGIC_INCLUDE_FILE
26};
27
28#undef GENL_struct
29#define GENL_struct(tag_name, tag_number, s_name, s_fields) \
30static struct nla_policy s_name ## _nl_policy[] __read_mostly = \
31{ s_fields };
32
33#undef __field
34#define __field(attr_nr, attr_flag, name, nla_type, _type, __get, \
35 __put, __is_signed) \
36 [attr_nr] = { .type = nla_type },
37
38#undef __array
39#define __array(attr_nr, attr_flag, name, nla_type, _type, maxlen, \
40 __get, __put, __is_signed) \
41 [attr_nr] = { .type = nla_type, \
42 .len = maxlen - (nla_type == NLA_NUL_STRING) },
43
44#include GENL_MAGIC_INCLUDE_FILE
45
46#ifndef __KERNEL__
47#ifndef pr_info
48#define pr_info(args...) fprintf(stderr, args);
49#endif
50#endif
51
52#ifdef GENL_MAGIC_DEBUG
53static void dprint_field(const char *dir, int nla_type,
54 const char *name, void *valp)
55{
56 __u64 val = valp ? *(__u32 *)valp : 1;
57 switch (nla_type) {
58 case NLA_U8: val = (__u8)val;
59 case NLA_U16: val = (__u16)val;
60 case NLA_U32: val = (__u32)val;
61 pr_info("%s attr %s: %d 0x%08x\n", dir,
62 name, (int)val, (unsigned)val);
63 break;
64 case NLA_U64:
65 val = *(__u64*)valp;
66 pr_info("%s attr %s: %lld 0x%08llx\n", dir,
67 name, (long long)val, (unsigned long long)val);
68 break;
69 case NLA_FLAG:
70 if (val)
71 pr_info("%s attr %s: set\n", dir, name);
72 break;
73 }
74}
75
76static void dprint_array(const char *dir, int nla_type,
77 const char *name, const char *val, unsigned len)
78{
79 switch (nla_type) {
80 case NLA_NUL_STRING:
81 if (len && val[len-1] == '\0')
82 len--;
83 pr_info("%s attr %s: [len:%u] '%s'\n", dir, name, len, val);
84 break;
85 default:
86 /* we can always show 4 byte,
87 * thats what nlattr are aligned to. */
88 pr_info("%s attr %s: [len:%u] %02x%02x%02x%02x ...\n",
89 dir, name, len, val[0], val[1], val[2], val[3]);
90 }
91}
92
93#define DPRINT_TLA(a, op, b) pr_info("%s %s %s\n", a, op, b);
94
95/* Name is a member field name of the struct s.
96 * If s is NULL (only parsing, no copy requested in *_from_attrs()),
97 * nla is supposed to point to the attribute containing the information
98 * corresponding to that struct member. */
99#define DPRINT_FIELD(dir, nla_type, name, s, nla) \
100 do { \
101 if (s) \
102 dprint_field(dir, nla_type, #name, &s->name); \
103 else if (nla) \
104 dprint_field(dir, nla_type, #name, \
105 (nla_type == NLA_FLAG) ? NULL \
106 : nla_data(nla)); \
107 } while (0)
108
109#define DPRINT_ARRAY(dir, nla_type, name, s, nla) \
110 do { \
111 if (s) \
112 dprint_array(dir, nla_type, #name, \
113 s->name, s->name ## _len); \
114 else if (nla) \
115 dprint_array(dir, nla_type, #name, \
116 nla_data(nla), nla_len(nla)); \
117 } while (0)
118#else
119#define DPRINT_TLA(a, op, b) do {} while (0)
120#define DPRINT_FIELD(dir, nla_type, name, s, nla) do {} while (0)
121#define DPRINT_ARRAY(dir, nla_type, name, s, nla) do {} while (0)
122#endif
123
124/*
125 * Magic: provide conversion functions {{{1
126 * populate struct from attribute table:
127 * {{{2
128 */
129
130/* processing of generic netlink messages is serialized.
131 * use one static buffer for parsing of nested attributes */
132static struct nlattr *nested_attr_tb[128];
133
134#ifndef BUILD_BUG_ON
135/* Force a compilation error if condition is true */
136#define BUILD_BUG_ON(condition) ((void)BUILD_BUG_ON_ZERO(condition))
137/* Force a compilation error if condition is true, but also produce a
138 result (of value 0 and type size_t), so the expression can be used
139 e.g. in a structure initializer (or where-ever else comma expressions
140 aren't permitted). */
141#define BUILD_BUG_ON_ZERO(e) (sizeof(struct { int:-!!(e); }))
142#define BUILD_BUG_ON_NULL(e) ((void *)sizeof(struct { int:-!!(e); }))
143#endif
144
145#undef GENL_struct
146#define GENL_struct(tag_name, tag_number, s_name, s_fields) \
147/* *_from_attrs functions are static, but potentially unused */ \
148static int __ ## s_name ## _from_attrs(struct s_name *s, \
149 struct genl_info *info, bool exclude_invariants) \
150{ \
151 const int maxtype = ARRAY_SIZE(s_name ## _nl_policy)-1; \
152 struct nlattr *tla = info->attrs[tag_number]; \
153 struct nlattr **ntb = nested_attr_tb; \
154 struct nlattr *nla; \
155 int err; \
156 BUILD_BUG_ON(ARRAY_SIZE(s_name ## _nl_policy) > ARRAY_SIZE(nested_attr_tb)); \
157 if (!tla) \
158 return -ENOMSG; \
159 DPRINT_TLA(#s_name, "<=-", #tag_name); \
160 err = drbd_nla_parse_nested(ntb, maxtype, tla, s_name ## _nl_policy); \
161 if (err) \
162 return err; \
163 \
164 s_fields \
165 return 0; \
166} __attribute__((unused)) \
167static int s_name ## _from_attrs(struct s_name *s, \
168 struct genl_info *info) \
169{ \
170 return __ ## s_name ## _from_attrs(s, info, false); \
171} __attribute__((unused)) \
172static int s_name ## _from_attrs_for_change(struct s_name *s, \
173 struct genl_info *info) \
174{ \
175 return __ ## s_name ## _from_attrs(s, info, true); \
176} __attribute__((unused)) \
177
178#define __assign(attr_nr, attr_flag, name, nla_type, type, assignment...) \
179 nla = ntb[attr_nr]; \
180 if (nla) { \
181 if (exclude_invariants && ((attr_flag) & DRBD_F_INVARIANT)) { \
182 pr_info("<< must not change invariant attr: %s\n", #name); \
183 return -EEXIST; \
184 } \
185 assignment; \
186 } else if (exclude_invariants && ((attr_flag) & DRBD_F_INVARIANT)) { \
187 /* attribute missing from payload, */ \
188 /* which was expected */ \
189 } else if ((attr_flag) & DRBD_F_REQUIRED) { \
190 pr_info("<< missing attr: %s\n", #name); \
191 return -ENOMSG; \
192 }
193
194#undef __field
195#define __field(attr_nr, attr_flag, name, nla_type, type, __get, __put, \
196 __is_signed) \
197 __assign(attr_nr, attr_flag, name, nla_type, type, \
198 if (s) \
199 s->name = __get(nla); \
200 DPRINT_FIELD("<<", nla_type, name, s, nla))
201
202/* validate_nla() already checked nla_len <= maxlen appropriately. */
203#undef __array
204#define __array(attr_nr, attr_flag, name, nla_type, type, maxlen, \
205 __get, __put, __is_signed) \
206 __assign(attr_nr, attr_flag, name, nla_type, type, \
207 if (s) \
208 s->name ## _len = \
209 __get(s->name, nla, maxlen); \
210 DPRINT_ARRAY("<<", nla_type, name, s, nla))
211
212#include GENL_MAGIC_INCLUDE_FILE
213
214#undef GENL_struct
215#define GENL_struct(tag_name, tag_number, s_name, s_fields)
216
217/*
218 * Magic: define op number to op name mapping {{{1
219 * {{{2
220 */
221const char *CONCAT_(GENL_MAGIC_FAMILY, _genl_cmd_to_str)(__u8 cmd)
222{
223 switch (cmd) {
224#undef GENL_op
225#define GENL_op(op_name, op_num, handler, tla_list) \
226 case op_num: return #op_name;
227#include GENL_MAGIC_INCLUDE_FILE
228 default:
229 return "unknown";
230 }
231}
232
233#ifdef __KERNEL__
234#include <linux/stringify.h>
235/*
236 * Magic: define genl_ops {{{1
237 * {{{2
238 */
239
240#undef GENL_op
241#define GENL_op(op_name, op_num, handler, tla_list) \
242{ \
243 handler \
244 .cmd = op_name, \
245 .policy = CONCAT_(GENL_MAGIC_FAMILY, _tla_nl_policy), \
246},
247
248#define ZZZ_genl_ops CONCAT_(GENL_MAGIC_FAMILY, _genl_ops)
249static struct genl_ops ZZZ_genl_ops[] __read_mostly = {
250#include GENL_MAGIC_INCLUDE_FILE
251};
252
253#undef GENL_op
254#define GENL_op(op_name, op_num, handler, tla_list)
255
256/*
257 * Define the genl_family, multicast groups, {{{1
258 * and provide register/unregister functions.
259 * {{{2
260 */
261#define ZZZ_genl_family CONCAT_(GENL_MAGIC_FAMILY, _genl_family)
262static struct genl_family ZZZ_genl_family __read_mostly = {
263 .id = GENL_ID_GENERATE,
264 .name = __stringify(GENL_MAGIC_FAMILY),
265 .version = GENL_MAGIC_VERSION,
266#ifdef GENL_MAGIC_FAMILY_HDRSZ
267 .hdrsize = NLA_ALIGN(GENL_MAGIC_FAMILY_HDRSZ),
268#endif
269 .maxattr = ARRAY_SIZE(drbd_tla_nl_policy)-1,
270};
271
272/*
273 * Magic: define multicast groups
274 * Magic: define multicast group registration helper
275 */
276#undef GENL_mc_group
277#define GENL_mc_group(group) \
278static struct genl_multicast_group \
279CONCAT_(GENL_MAGIC_FAMILY, _mcg_ ## group) __read_mostly = { \
280 .name = #group, \
281}; \
282static int CONCAT_(GENL_MAGIC_FAMILY, _genl_multicast_ ## group)( \
283 struct sk_buff *skb, gfp_t flags) \
284{ \
285 unsigned int group_id = \
286 CONCAT_(GENL_MAGIC_FAMILY, _mcg_ ## group).id; \
287 if (!group_id) \
288 return -EINVAL; \
289 return genlmsg_multicast(skb, 0, group_id, flags); \
290}
291
292#include GENL_MAGIC_INCLUDE_FILE
293
294int CONCAT_(GENL_MAGIC_FAMILY, _genl_register)(void)
295{
296 int err = genl_register_family_with_ops(&ZZZ_genl_family,
297 ZZZ_genl_ops, ARRAY_SIZE(ZZZ_genl_ops));
298 if (err)
299 return err;
300#undef GENL_mc_group
301#define GENL_mc_group(group) \
302 err = genl_register_mc_group(&ZZZ_genl_family, \
303 &CONCAT_(GENL_MAGIC_FAMILY, _mcg_ ## group)); \
304 if (err) \
305 goto fail; \
306 else \
307 pr_info("%s: mcg %s: %u\n", #group, \
308 __stringify(GENL_MAGIC_FAMILY), \
309 CONCAT_(GENL_MAGIC_FAMILY, _mcg_ ## group).id);
310
311#include GENL_MAGIC_INCLUDE_FILE
312
313#undef GENL_mc_group
314#define GENL_mc_group(group)
315 return 0;
316fail:
317 genl_unregister_family(&ZZZ_genl_family);
318 return err;
319}
320
321void CONCAT_(GENL_MAGIC_FAMILY, _genl_unregister)(void)
322{
323 genl_unregister_family(&ZZZ_genl_family);
324}
325
326/*
327 * Magic: provide conversion functions {{{1
328 * populate skb from struct.
329 * {{{2
330 */
331
332#undef GENL_op
333#define GENL_op(op_name, op_num, handler, tla_list)
334
335#undef GENL_struct
336#define GENL_struct(tag_name, tag_number, s_name, s_fields) \
337static int s_name ## _to_skb(struct sk_buff *skb, struct s_name *s, \
338 const bool exclude_sensitive) \
339{ \
340 struct nlattr *tla = nla_nest_start(skb, tag_number); \
341 if (!tla) \
342 goto nla_put_failure; \
343 DPRINT_TLA(#s_name, "-=>", #tag_name); \
344 s_fields \
345 nla_nest_end(skb, tla); \
346 return 0; \
347 \
348nla_put_failure: \
349 if (tla) \
350 nla_nest_cancel(skb, tla); \
351 return -EMSGSIZE; \
352} \
353static inline int s_name ## _to_priv_skb(struct sk_buff *skb, \
354 struct s_name *s) \
355{ \
356 return s_name ## _to_skb(skb, s, 0); \
357} \
358static inline int s_name ## _to_unpriv_skb(struct sk_buff *skb, \
359 struct s_name *s) \
360{ \
361 return s_name ## _to_skb(skb, s, 1); \
362}
363
364
365#undef __field
366#define __field(attr_nr, attr_flag, name, nla_type, type, __get, __put, \
367 __is_signed) \
368 if (!exclude_sensitive || !((attr_flag) & DRBD_F_SENSITIVE)) { \
369 DPRINT_FIELD(">>", nla_type, name, s, NULL); \
370 if (__put(skb, attr_nr, s->name)) \
371 goto nla_put_failure; \
372 }
373
374#undef __array
375#define __array(attr_nr, attr_flag, name, nla_type, type, maxlen, \
376 __get, __put, __is_signed) \
377 if (!exclude_sensitive || !((attr_flag) & DRBD_F_SENSITIVE)) { \
378 DPRINT_ARRAY(">>",nla_type, name, s, NULL); \
379 if (__put(skb, attr_nr, min_t(int, maxlen, \
380 s->name ## _len + (nla_type == NLA_NUL_STRING)),\
381 s->name)) \
382 goto nla_put_failure; \
383 }
384
385#include GENL_MAGIC_INCLUDE_FILE
386
387
388/* Functions for initializing structs to default values. */
389
390#undef __field
391#define __field(attr_nr, attr_flag, name, nla_type, type, __get, __put, \
392 __is_signed)
393#undef __array
394#define __array(attr_nr, attr_flag, name, nla_type, type, maxlen, \
395 __get, __put, __is_signed)
396#undef __u32_field_def
397#define __u32_field_def(attr_nr, attr_flag, name, default) \
398 x->name = default;
399#undef __s32_field_def
400#define __s32_field_def(attr_nr, attr_flag, name, default) \
401 x->name = default;
402#undef __flg_field_def
403#define __flg_field_def(attr_nr, attr_flag, name, default) \
404 x->name = default;
405#undef __str_field_def
406#define __str_field_def(attr_nr, attr_flag, name, maxlen) \
407 memset(x->name, 0, sizeof(x->name)); \
408 x->name ## _len = 0;
409#undef GENL_struct
410#define GENL_struct(tag_name, tag_number, s_name, s_fields) \
411static void set_ ## s_name ## _defaults(struct s_name *x) __attribute__((unused)); \
412static void set_ ## s_name ## _defaults(struct s_name *x) { \
413s_fields \
414}
415
416#include GENL_MAGIC_INCLUDE_FILE
417
418#endif /* __KERNEL__ */
419
420/* }}}1 */
421#endif /* GENL_MAGIC_FUNC_H */
422/* vim: set foldmethod=marker foldlevel=1 nofoldenable : */
diff --git a/include/linux/genl_magic_struct.h b/include/linux/genl_magic_struct.h
new file mode 100644
index 000000000000..eecd19b37001
--- /dev/null
+++ b/include/linux/genl_magic_struct.h
@@ -0,0 +1,277 @@
1#ifndef GENL_MAGIC_STRUCT_H
2#define GENL_MAGIC_STRUCT_H
3
4#ifndef GENL_MAGIC_FAMILY
5# error "you need to define GENL_MAGIC_FAMILY before inclusion"
6#endif
7
8#ifndef GENL_MAGIC_VERSION
9# error "you need to define GENL_MAGIC_VERSION before inclusion"
10#endif
11
12#ifndef GENL_MAGIC_INCLUDE_FILE
13# error "you need to define GENL_MAGIC_INCLUDE_FILE before inclusion"
14#endif
15
16#include <linux/genetlink.h>
17#include <linux/types.h>
18
19#define CONCAT__(a,b) a ## b
20#define CONCAT_(a,b) CONCAT__(a,b)
21
22extern int CONCAT_(GENL_MAGIC_FAMILY, _genl_register)(void);
23extern void CONCAT_(GENL_MAGIC_FAMILY, _genl_unregister)(void);
24
25/*
26 * Extension of genl attribute validation policies {{{2
27 */
28
29/*
30 * @DRBD_GENLA_F_MANDATORY: By default, netlink ignores attributes it does not
31 * know about. This flag can be set in nlattr->nla_type to indicate that this
32 * attribute must not be ignored.
33 *
34 * We check and remove this flag in drbd_nla_check_mandatory() before
35 * validating the attribute types and lengths via nla_parse_nested().
36 */
37#define DRBD_GENLA_F_MANDATORY (1 << 14)
38
39/*
40 * Flags specific to drbd and not visible at the netlink layer, used in
41 * <struct>_from_attrs and <struct>_to_skb:
42 *
43 * @DRBD_F_REQUIRED: Attribute is required; a request without this attribute is
44 * invalid.
45 *
46 * @DRBD_F_SENSITIVE: Attribute includes sensitive information and must not be
47 * included in unpriviledged get requests or broadcasts.
48 *
49 * @DRBD_F_INVARIANT: Attribute is set when an object is initially created, but
50 * cannot subsequently be changed.
51 */
52#define DRBD_F_REQUIRED (1 << 0)
53#define DRBD_F_SENSITIVE (1 << 1)
54#define DRBD_F_INVARIANT (1 << 2)
55
56#define __nla_type(x) ((__u16)((x) & NLA_TYPE_MASK & ~DRBD_GENLA_F_MANDATORY))
57
58/* }}}1
59 * MAGIC
60 * multi-include macro expansion magic starts here
61 */
62
63/* MAGIC helpers {{{2 */
64
65/* possible field types */
66#define __flg_field(attr_nr, attr_flag, name) \
67 __field(attr_nr, attr_flag, name, NLA_U8, char, \
68 nla_get_u8, nla_put_u8, false)
69#define __u8_field(attr_nr, attr_flag, name) \
70 __field(attr_nr, attr_flag, name, NLA_U8, unsigned char, \
71 nla_get_u8, nla_put_u8, false)
72#define __u16_field(attr_nr, attr_flag, name) \
73 __field(attr_nr, attr_flag, name, NLA_U16, __u16, \
74 nla_get_u16, nla_put_u16, false)
75#define __u32_field(attr_nr, attr_flag, name) \
76 __field(attr_nr, attr_flag, name, NLA_U32, __u32, \
77 nla_get_u32, nla_put_u32, false)
78#define __s32_field(attr_nr, attr_flag, name) \
79 __field(attr_nr, attr_flag, name, NLA_U32, __s32, \
80 nla_get_u32, nla_put_u32, true)
81#define __u64_field(attr_nr, attr_flag, name) \
82 __field(attr_nr, attr_flag, name, NLA_U64, __u64, \
83 nla_get_u64, nla_put_u64, false)
84#define __str_field(attr_nr, attr_flag, name, maxlen) \
85 __array(attr_nr, attr_flag, name, NLA_NUL_STRING, char, maxlen, \
86 nla_strlcpy, nla_put, false)
87#define __bin_field(attr_nr, attr_flag, name, maxlen) \
88 __array(attr_nr, attr_flag, name, NLA_BINARY, char, maxlen, \
89 nla_memcpy, nla_put, false)
90
91/* fields with default values */
92#define __flg_field_def(attr_nr, attr_flag, name, default) \
93 __flg_field(attr_nr, attr_flag, name)
94#define __u32_field_def(attr_nr, attr_flag, name, default) \
95 __u32_field(attr_nr, attr_flag, name)
96#define __s32_field_def(attr_nr, attr_flag, name, default) \
97 __s32_field(attr_nr, attr_flag, name)
98#define __str_field_def(attr_nr, attr_flag, name, maxlen) \
99 __str_field(attr_nr, attr_flag, name, maxlen)
100
101#define GENL_op_init(args...) args
102#define GENL_doit(handler) \
103 .doit = handler, \
104 .flags = GENL_ADMIN_PERM,
105#define GENL_dumpit(handler) \
106 .dumpit = handler, \
107 .flags = GENL_ADMIN_PERM,
108
109/* }}}1
110 * Magic: define the enum symbols for genl_ops
111 * Magic: define the enum symbols for top level attributes
112 * Magic: define the enum symbols for nested attributes
113 * {{{2
114 */
115
116#undef GENL_struct
117#define GENL_struct(tag_name, tag_number, s_name, s_fields)
118
119#undef GENL_mc_group
120#define GENL_mc_group(group)
121
122#undef GENL_notification
123#define GENL_notification(op_name, op_num, mcast_group, tla_list) \
124 op_name = op_num,
125
126#undef GENL_op
127#define GENL_op(op_name, op_num, handler, tla_list) \
128 op_name = op_num,
129
130enum {
131#include GENL_MAGIC_INCLUDE_FILE
132};
133
134#undef GENL_notification
135#define GENL_notification(op_name, op_num, mcast_group, tla_list)
136
137#undef GENL_op
138#define GENL_op(op_name, op_num, handler, attr_list)
139
140#undef GENL_struct
141#define GENL_struct(tag_name, tag_number, s_name, s_fields) \
142 tag_name = tag_number,
143
144enum {
145#include GENL_MAGIC_INCLUDE_FILE
146};
147
148#undef GENL_struct
149#define GENL_struct(tag_name, tag_number, s_name, s_fields) \
150enum { \
151 s_fields \
152};
153
154#undef __field
155#define __field(attr_nr, attr_flag, name, nla_type, type, \
156 __get, __put, __is_signed) \
157 T_ ## name = (__u16)(attr_nr | ((attr_flag) & DRBD_GENLA_F_MANDATORY)),
158
159#undef __array
160#define __array(attr_nr, attr_flag, name, nla_type, type, \
161 maxlen, __get, __put, __is_signed) \
162 T_ ## name = (__u16)(attr_nr | ((attr_flag) & DRBD_GENLA_F_MANDATORY)),
163
164#include GENL_MAGIC_INCLUDE_FILE
165
166/* }}}1
167 * Magic: compile time assert unique numbers for operations
168 * Magic: -"- unique numbers for top level attributes
169 * Magic: -"- unique numbers for nested attributes
170 * {{{2
171 */
172
173#undef GENL_struct
174#define GENL_struct(tag_name, tag_number, s_name, s_fields)
175
176#undef GENL_op
177#define GENL_op(op_name, op_num, handler, attr_list) \
178 case op_name:
179
180#undef GENL_notification
181#define GENL_notification(op_name, op_num, mcast_group, tla_list) \
182 case op_name:
183
184static inline void ct_assert_unique_operations(void)
185{
186 switch (0) {
187#include GENL_MAGIC_INCLUDE_FILE
188 ;
189 }
190}
191
192#undef GENL_op
193#define GENL_op(op_name, op_num, handler, attr_list)
194
195#undef GENL_notification
196#define GENL_notification(op_name, op_num, mcast_group, tla_list)
197
198#undef GENL_struct
199#define GENL_struct(tag_name, tag_number, s_name, s_fields) \
200 case tag_number:
201
202static inline void ct_assert_unique_top_level_attributes(void)
203{
204 switch (0) {
205#include GENL_MAGIC_INCLUDE_FILE
206 ;
207 }
208}
209
210#undef GENL_struct
211#define GENL_struct(tag_name, tag_number, s_name, s_fields) \
212static inline void ct_assert_unique_ ## s_name ## _attributes(void) \
213{ \
214 switch (0) { \
215 s_fields \
216 ; \
217 } \
218}
219
220#undef __field
221#define __field(attr_nr, attr_flag, name, nla_type, type, __get, __put, \
222 __is_signed) \
223 case attr_nr:
224
225#undef __array
226#define __array(attr_nr, attr_flag, name, nla_type, type, maxlen, \
227 __get, __put, __is_signed) \
228 case attr_nr:
229
230#include GENL_MAGIC_INCLUDE_FILE
231
232/* }}}1
233 * Magic: declare structs
234 * struct <name> {
235 * fields
236 * };
237 * {{{2
238 */
239
240#undef GENL_struct
241#define GENL_struct(tag_name, tag_number, s_name, s_fields) \
242struct s_name { s_fields };
243
244#undef __field
245#define __field(attr_nr, attr_flag, name, nla_type, type, __get, __put, \
246 __is_signed) \
247 type name;
248
249#undef __array
250#define __array(attr_nr, attr_flag, name, nla_type, type, maxlen, \
251 __get, __put, __is_signed) \
252 type name[maxlen]; \
253 __u32 name ## _len;
254
255#include GENL_MAGIC_INCLUDE_FILE
256
257#undef GENL_struct
258#define GENL_struct(tag_name, tag_number, s_name, s_fields) \
259enum { \
260 s_fields \
261};
262
263#undef __field
264#define __field(attr_nr, attr_flag, name, nla_type, type, __get, __put, \
265 is_signed) \
266 F_ ## name ## _IS_SIGNED = is_signed,
267
268#undef __array
269#define __array(attr_nr, attr_flag, name, nla_type, type, maxlen, \
270 __get, __put, is_signed) \
271 F_ ## name ## _IS_SIGNED = is_signed,
272
273#include GENL_MAGIC_INCLUDE_FILE
274
275/* }}}1 */
276#endif /* GENL_MAGIC_STRUCT_H */
277/* vim: set foldmethod=marker nofoldenable : */
diff --git a/include/linux/idr.h b/include/linux/idr.h
index 87259a44c251..de7e190f1af4 100644
--- a/include/linux/idr.h
+++ b/include/linux/idr.h
@@ -152,4 +152,15 @@ void ida_simple_remove(struct ida *ida, unsigned int id);
152 152
153void __init idr_init_cache(void); 153void __init idr_init_cache(void);
154 154
155/**
156 * idr_for_each_entry - iterate over an idr's elements of a given type
157 * @idp: idr handle
158 * @entry: the type * to use as cursor
159 * @id: id entry's key
160 */
161#define idr_for_each_entry(idp, entry, id) \
162 for (id = 0, entry = (typeof(entry))idr_get_next((idp), &(id)); \
163 entry != NULL; \
164 ++id, entry = (typeof(entry))idr_get_next((idp), &(id)))
165
155#endif /* __IDR_H__ */ 166#endif /* __IDR_H__ */
diff --git a/include/linux/loop.h b/include/linux/loop.h
index 6492181bcb1d..460b60fa7adf 100644
--- a/include/linux/loop.h
+++ b/include/linux/loop.h
@@ -53,10 +53,13 @@ struct loop_device {
53 53
54 spinlock_t lo_lock; 54 spinlock_t lo_lock;
55 struct bio_list lo_bio_list; 55 struct bio_list lo_bio_list;
56 unsigned int lo_bio_count;
56 int lo_state; 57 int lo_state;
57 struct mutex lo_ctl_mutex; 58 struct mutex lo_ctl_mutex;
58 struct task_struct *lo_thread; 59 struct task_struct *lo_thread;
59 wait_queue_head_t lo_event; 60 wait_queue_head_t lo_event;
61 /* wait queue for incoming requests */
62 wait_queue_head_t lo_req_wait;
60 63
61 struct request_queue *lo_queue; 64 struct request_queue *lo_queue;
62 struct gendisk *lo_disk; 65 struct gendisk *lo_disk;
diff --git a/include/linux/lru_cache.h b/include/linux/lru_cache.h
index cafc7f99e124..4019013c6593 100644
--- a/include/linux/lru_cache.h
+++ b/include/linux/lru_cache.h
@@ -166,9 +166,11 @@ struct lc_element {
166 /* if we want to track a larger set of objects, 166 /* if we want to track a larger set of objects,
167 * it needs to become arch independend u64 */ 167 * it needs to become arch independend u64 */
168 unsigned lc_number; 168 unsigned lc_number;
169
170 /* special label when on free list */ 169 /* special label when on free list */
171#define LC_FREE (~0U) 170#define LC_FREE (~0U)
171
172 /* for pending changes */
173 unsigned lc_new_number;
172}; 174};
173 175
174struct lru_cache { 176struct lru_cache {
@@ -176,6 +178,7 @@ struct lru_cache {
176 struct list_head lru; 178 struct list_head lru;
177 struct list_head free; 179 struct list_head free;
178 struct list_head in_use; 180 struct list_head in_use;
181 struct list_head to_be_changed;
179 182
180 /* the pre-created kmem cache to allocate the objects from */ 183 /* the pre-created kmem cache to allocate the objects from */
181 struct kmem_cache *lc_cache; 184 struct kmem_cache *lc_cache;
@@ -186,7 +189,7 @@ struct lru_cache {
186 size_t element_off; 189 size_t element_off;
187 190
188 /* number of elements (indices) */ 191 /* number of elements (indices) */
189 unsigned int nr_elements; 192 unsigned int nr_elements;
190 /* Arbitrary limit on maximum tracked objects. Practical limit is much 193 /* Arbitrary limit on maximum tracked objects. Practical limit is much
191 * lower due to allocation failures, probably. For typical use cases, 194 * lower due to allocation failures, probably. For typical use cases,
192 * nr_elements should be a few thousand at most. 195 * nr_elements should be a few thousand at most.
@@ -194,18 +197,19 @@ struct lru_cache {
194 * 8 high bits of .lc_index to be overloaded with flags in the future. */ 197 * 8 high bits of .lc_index to be overloaded with flags in the future. */
195#define LC_MAX_ACTIVE (1<<24) 198#define LC_MAX_ACTIVE (1<<24)
196 199
200 /* allow to accumulate a few (index:label) changes,
201 * but no more than max_pending_changes */
202 unsigned int max_pending_changes;
203 /* number of elements currently on to_be_changed list */
204 unsigned int pending_changes;
205
197 /* statistics */ 206 /* statistics */
198 unsigned used; /* number of lelements currently on in_use list */ 207 unsigned used; /* number of elements currently on in_use list */
199 unsigned long hits, misses, starving, dirty, changed; 208 unsigned long hits, misses, starving, locked, changed;
200 209
201 /* see below: flag-bits for lru_cache */ 210 /* see below: flag-bits for lru_cache */
202 unsigned long flags; 211 unsigned long flags;
203 212
204 /* when changing the label of an index element */
205 unsigned int new_number;
206
207 /* for paranoia when changing the label of an index element */
208 struct lc_element *changing_element;
209 213
210 void *lc_private; 214 void *lc_private;
211 const char *name; 215 const char *name;
@@ -221,10 +225,15 @@ enum {
221 /* debugging aid, to catch concurrent access early. 225 /* debugging aid, to catch concurrent access early.
222 * user needs to guarantee exclusive access by proper locking! */ 226 * user needs to guarantee exclusive access by proper locking! */
223 __LC_PARANOIA, 227 __LC_PARANOIA,
224 /* if we need to change the set, but currently there is a changing 228
225 * transaction pending, we are "dirty", and must deferr further 229 /* annotate that the set is "dirty", possibly accumulating further
226 * changing requests */ 230 * changes, until a transaction is finally triggered */
227 __LC_DIRTY, 231 __LC_DIRTY,
232
233 /* Locked, no further changes allowed.
234 * Also used to serialize changing transactions. */
235 __LC_LOCKED,
236
228 /* if we need to change the set, but currently there is no free nor 237 /* if we need to change the set, but currently there is no free nor
229 * unused element available, we are "starving", and must not give out 238 * unused element available, we are "starving", and must not give out
230 * further references, to guarantee that eventually some refcnt will 239 * further references, to guarantee that eventually some refcnt will
@@ -236,9 +245,11 @@ enum {
236}; 245};
237#define LC_PARANOIA (1<<__LC_PARANOIA) 246#define LC_PARANOIA (1<<__LC_PARANOIA)
238#define LC_DIRTY (1<<__LC_DIRTY) 247#define LC_DIRTY (1<<__LC_DIRTY)
248#define LC_LOCKED (1<<__LC_LOCKED)
239#define LC_STARVING (1<<__LC_STARVING) 249#define LC_STARVING (1<<__LC_STARVING)
240 250
241extern struct lru_cache *lc_create(const char *name, struct kmem_cache *cache, 251extern struct lru_cache *lc_create(const char *name, struct kmem_cache *cache,
252 unsigned max_pending_changes,
242 unsigned e_count, size_t e_size, size_t e_off); 253 unsigned e_count, size_t e_size, size_t e_off);
243extern void lc_reset(struct lru_cache *lc); 254extern void lc_reset(struct lru_cache *lc);
244extern void lc_destroy(struct lru_cache *lc); 255extern void lc_destroy(struct lru_cache *lc);
@@ -249,7 +260,7 @@ extern struct lc_element *lc_try_get(struct lru_cache *lc, unsigned int enr);
249extern struct lc_element *lc_find(struct lru_cache *lc, unsigned int enr); 260extern struct lc_element *lc_find(struct lru_cache *lc, unsigned int enr);
250extern struct lc_element *lc_get(struct lru_cache *lc, unsigned int enr); 261extern struct lc_element *lc_get(struct lru_cache *lc, unsigned int enr);
251extern unsigned int lc_put(struct lru_cache *lc, struct lc_element *e); 262extern unsigned int lc_put(struct lru_cache *lc, struct lc_element *e);
252extern void lc_changed(struct lru_cache *lc, struct lc_element *e); 263extern void lc_committed(struct lru_cache *lc);
253 264
254struct seq_file; 265struct seq_file;
255extern size_t lc_seq_printf_stats(struct seq_file *seq, struct lru_cache *lc); 266extern size_t lc_seq_printf_stats(struct seq_file *seq, struct lru_cache *lc);
@@ -258,32 +269,40 @@ extern void lc_seq_dump_details(struct seq_file *seq, struct lru_cache *lc, char
258 void (*detail) (struct seq_file *, struct lc_element *)); 269 void (*detail) (struct seq_file *, struct lc_element *));
259 270
260/** 271/**
261 * lc_try_lock - can be used to stop lc_get() from changing the tracked set 272 * lc_try_lock_for_transaction - can be used to stop lc_get() from changing the tracked set
262 * @lc: the lru cache to operate on 273 * @lc: the lru cache to operate on
263 * 274 *
264 * Note that the reference counts and order on the active and lru lists may 275 * Allows (expects) the set to be "dirty". Note that the reference counts and
265 * still change. Returns true if we acquired the lock. 276 * order on the active and lru lists may still change. Used to serialize
277 * changing transactions. Returns true if we aquired the lock.
266 */ 278 */
267static inline int lc_try_lock(struct lru_cache *lc) 279static inline int lc_try_lock_for_transaction(struct lru_cache *lc)
268{ 280{
269 return !test_and_set_bit(__LC_DIRTY, &lc->flags); 281 return !test_and_set_bit(__LC_LOCKED, &lc->flags);
270} 282}
271 283
272/** 284/**
285 * lc_try_lock - variant to stop lc_get() from changing the tracked set
286 * @lc: the lru cache to operate on
287 *
288 * Note that the reference counts and order on the active and lru lists may
289 * still change. Only works on a "clean" set. Returns true if we aquired the
290 * lock, which means there are no pending changes, and any further attempt to
291 * change the set will not succeed until the next lc_unlock().
292 */
293extern int lc_try_lock(struct lru_cache *lc);
294
295/**
273 * lc_unlock - unlock @lc, allow lc_get() to change the set again 296 * lc_unlock - unlock @lc, allow lc_get() to change the set again
274 * @lc: the lru cache to operate on 297 * @lc: the lru cache to operate on
275 */ 298 */
276static inline void lc_unlock(struct lru_cache *lc) 299static inline void lc_unlock(struct lru_cache *lc)
277{ 300{
278 clear_bit(__LC_DIRTY, &lc->flags); 301 clear_bit(__LC_DIRTY, &lc->flags);
279 smp_mb__after_clear_bit(); 302 clear_bit_unlock(__LC_LOCKED, &lc->flags);
280} 303}
281 304
282static inline int lc_is_used(struct lru_cache *lc, unsigned int enr) 305extern bool lc_is_used(struct lru_cache *lc, unsigned int enr);
283{
284 struct lc_element *e = lc_find(lc, enr);
285 return e && e->refcnt;
286}
287 306
288#define lc_entry(ptr, type, member) \ 307#define lc_entry(ptr, type, member) \
289 container_of(ptr, type, member) 308 container_of(ptr, type, member)
diff --git a/include/linux/wait.h b/include/linux/wait.h
index 168dfe122dd3..7cb64d4b499d 100644
--- a/include/linux/wait.h
+++ b/include/linux/wait.h
@@ -550,6 +550,170 @@ do { \
550 __ret; \ 550 __ret; \
551}) 551})
552 552
553
554#define __wait_event_lock_irq(wq, condition, lock, cmd) \
555do { \
556 DEFINE_WAIT(__wait); \
557 \
558 for (;;) { \
559 prepare_to_wait(&wq, &__wait, TASK_UNINTERRUPTIBLE); \
560 if (condition) \
561 break; \
562 spin_unlock_irq(&lock); \
563 cmd; \
564 schedule(); \
565 spin_lock_irq(&lock); \
566 } \
567 finish_wait(&wq, &__wait); \
568} while (0)
569
570/**
571 * wait_event_lock_irq_cmd - sleep until a condition gets true. The
572 * condition is checked under the lock. This
573 * is expected to be called with the lock
574 * taken.
575 * @wq: the waitqueue to wait on
576 * @condition: a C expression for the event to wait for
577 * @lock: a locked spinlock_t, which will be released before cmd
578 * and schedule() and reacquired afterwards.
579 * @cmd: a command which is invoked outside the critical section before
580 * sleep
581 *
582 * The process is put to sleep (TASK_UNINTERRUPTIBLE) until the
583 * @condition evaluates to true. The @condition is checked each time
584 * the waitqueue @wq is woken up.
585 *
586 * wake_up() has to be called after changing any variable that could
587 * change the result of the wait condition.
588 *
589 * This is supposed to be called while holding the lock. The lock is
590 * dropped before invoking the cmd and going to sleep and is reacquired
591 * afterwards.
592 */
593#define wait_event_lock_irq_cmd(wq, condition, lock, cmd) \
594do { \
595 if (condition) \
596 break; \
597 __wait_event_lock_irq(wq, condition, lock, cmd); \
598} while (0)
599
600/**
601 * wait_event_lock_irq - sleep until a condition gets true. The
602 * condition is checked under the lock. This
603 * is expected to be called with the lock
604 * taken.
605 * @wq: the waitqueue to wait on
606 * @condition: a C expression for the event to wait for
607 * @lock: a locked spinlock_t, which will be released before schedule()
608 * and reacquired afterwards.
609 *
610 * The process is put to sleep (TASK_UNINTERRUPTIBLE) until the
611 * @condition evaluates to true. The @condition is checked each time
612 * the waitqueue @wq is woken up.
613 *
614 * wake_up() has to be called after changing any variable that could
615 * change the result of the wait condition.
616 *
617 * This is supposed to be called while holding the lock. The lock is
618 * dropped before going to sleep and is reacquired afterwards.
619 */
620#define wait_event_lock_irq(wq, condition, lock) \
621do { \
622 if (condition) \
623 break; \
624 __wait_event_lock_irq(wq, condition, lock, ); \
625} while (0)
626
627
628#define __wait_event_interruptible_lock_irq(wq, condition, \
629 lock, ret, cmd) \
630do { \
631 DEFINE_WAIT(__wait); \
632 \
633 for (;;) { \
634 prepare_to_wait(&wq, &__wait, TASK_INTERRUPTIBLE); \
635 if (condition) \
636 break; \
637 if (signal_pending(current)) { \
638 ret = -ERESTARTSYS; \
639 break; \
640 } \
641 spin_unlock_irq(&lock); \
642 cmd; \
643 schedule(); \
644 spin_lock_irq(&lock); \
645 } \
646 finish_wait(&wq, &__wait); \
647} while (0)
648
649/**
650 * wait_event_interruptible_lock_irq_cmd - sleep until a condition gets true.
651 * The condition is checked under the lock. This is expected to
652 * be called with the lock taken.
653 * @wq: the waitqueue to wait on
654 * @condition: a C expression for the event to wait for
655 * @lock: a locked spinlock_t, which will be released before cmd and
656 * schedule() and reacquired afterwards.
657 * @cmd: a command which is invoked outside the critical section before
658 * sleep
659 *
660 * The process is put to sleep (TASK_INTERRUPTIBLE) until the
661 * @condition evaluates to true or a signal is received. The @condition is
662 * checked each time the waitqueue @wq is woken up.
663 *
664 * wake_up() has to be called after changing any variable that could
665 * change the result of the wait condition.
666 *
667 * This is supposed to be called while holding the lock. The lock is
668 * dropped before invoking the cmd and going to sleep and is reacquired
669 * afterwards.
670 *
671 * The macro will return -ERESTARTSYS if it was interrupted by a signal
672 * and 0 if @condition evaluated to true.
673 */
674#define wait_event_interruptible_lock_irq_cmd(wq, condition, lock, cmd) \
675({ \
676 int __ret = 0; \
677 \
678 if (!(condition)) \
679 __wait_event_interruptible_lock_irq(wq, condition, \
680 lock, __ret, cmd); \
681 __ret; \
682})
683
684/**
685 * wait_event_interruptible_lock_irq - sleep until a condition gets true.
686 * The condition is checked under the lock. This is expected
687 * to be called with the lock taken.
688 * @wq: the waitqueue to wait on
689 * @condition: a C expression for the event to wait for
690 * @lock: a locked spinlock_t, which will be released before schedule()
691 * and reacquired afterwards.
692 *
693 * The process is put to sleep (TASK_INTERRUPTIBLE) until the
694 * @condition evaluates to true or signal is received. The @condition is
695 * checked each time the waitqueue @wq is woken up.
696 *
697 * wake_up() has to be called after changing any variable that could
698 * change the result of the wait condition.
699 *
700 * This is supposed to be called while holding the lock. The lock is
701 * dropped before going to sleep and is reacquired afterwards.
702 *
703 * The macro will return -ERESTARTSYS if it was interrupted by a signal
704 * and 0 if @condition evaluated to true.
705 */
706#define wait_event_interruptible_lock_irq(wq, condition, lock) \
707({ \
708 int __ret = 0; \
709 \
710 if (!(condition)) \
711 __wait_event_interruptible_lock_irq(wq, condition, \
712 lock, __ret, ); \
713 __ret; \
714})
715
716
553/* 717/*
554 * These are the old interfaces to sleep waiting for an event. 718 * These are the old interfaces to sleep waiting for an event.
555 * They are racy. DO NOT use them, use the wait_event* interfaces above. 719 * They are racy. DO NOT use them, use the wait_event* interfaces above.
generated by cgit v1.2.2 (git 2.25.0) at 2025-01-07 12:16:10 -0500