| lh | 9ed821d | 2023-04-07 01:36:19 -0700 | [diff] [blame] | 1 | #include <linux/ceph/ceph_debug.h> | 
|  | 2 |  | 
|  | 3 | #include <linux/file.h> | 
|  | 4 | #include <linux/namei.h> | 
|  | 5 |  | 
|  | 6 | #include "super.h" | 
|  | 7 | #include "mds_client.h" | 
|  | 8 | #include <linux/ceph/pagelist.h> | 
|  | 9 |  | 
|  | 10 | /** | 
|  | 11 | * Implement fcntl and flock locking functions. | 
|  | 12 | */ | 
|  | 13 | static int ceph_lock_message(u8 lock_type, u16 operation, struct file *file, | 
|  | 14 | int cmd, u8 wait, struct file_lock *fl) | 
|  | 15 | { | 
|  | 16 | struct inode *inode = file->f_dentry->d_inode; | 
|  | 17 | struct ceph_mds_client *mdsc = | 
|  | 18 | ceph_sb_to_client(inode->i_sb)->mdsc; | 
|  | 19 | struct ceph_mds_request *req; | 
|  | 20 | int err; | 
|  | 21 | u64 length = 0; | 
|  | 22 |  | 
|  | 23 | req = ceph_mdsc_create_request(mdsc, operation, USE_AUTH_MDS); | 
|  | 24 | if (IS_ERR(req)) | 
|  | 25 | return PTR_ERR(req); | 
|  | 26 | req->r_inode = inode; | 
|  | 27 | ihold(inode); | 
|  | 28 |  | 
|  | 29 | /* mds requires start and length rather than start and end */ | 
|  | 30 | if (LLONG_MAX == fl->fl_end) | 
|  | 31 | length = 0; | 
|  | 32 | else | 
|  | 33 | length = fl->fl_end - fl->fl_start + 1; | 
|  | 34 |  | 
|  | 35 | dout("ceph_lock_message: rule: %d, op: %d, pid: %llu, start: %llu, " | 
|  | 36 | "length: %llu, wait: %d, type: %d", (int)lock_type, | 
|  | 37 | (int)operation, (u64)fl->fl_pid, fl->fl_start, | 
|  | 38 | length, wait, fl->fl_type); | 
|  | 39 |  | 
|  | 40 | req->r_args.filelock_change.rule = lock_type; | 
|  | 41 | req->r_args.filelock_change.type = cmd; | 
|  | 42 | req->r_args.filelock_change.pid = cpu_to_le64((u64)fl->fl_pid); | 
|  | 43 | /* This should be adjusted, but I'm not sure if | 
|  | 44 | namespaces actually get id numbers*/ | 
|  | 45 | req->r_args.filelock_change.pid_namespace = | 
|  | 46 | cpu_to_le64((u64)(unsigned long)fl->fl_nspid); | 
|  | 47 | req->r_args.filelock_change.start = cpu_to_le64(fl->fl_start); | 
|  | 48 | req->r_args.filelock_change.length = cpu_to_le64(length); | 
|  | 49 | req->r_args.filelock_change.wait = wait; | 
|  | 50 |  | 
|  | 51 | err = ceph_mdsc_do_request(mdsc, inode, req); | 
|  | 52 |  | 
|  | 53 | if ( operation == CEPH_MDS_OP_GETFILELOCK){ | 
|  | 54 | fl->fl_pid = le64_to_cpu(req->r_reply_info.filelock_reply->pid); | 
|  | 55 | if (CEPH_LOCK_SHARED == req->r_reply_info.filelock_reply->type) | 
|  | 56 | fl->fl_type = F_RDLCK; | 
|  | 57 | else if (CEPH_LOCK_EXCL == req->r_reply_info.filelock_reply->type) | 
|  | 58 | fl->fl_type = F_WRLCK; | 
|  | 59 | else | 
|  | 60 | fl->fl_type = F_UNLCK; | 
|  | 61 |  | 
|  | 62 | fl->fl_start = le64_to_cpu(req->r_reply_info.filelock_reply->start); | 
|  | 63 | length = le64_to_cpu(req->r_reply_info.filelock_reply->start) + | 
|  | 64 | le64_to_cpu(req->r_reply_info.filelock_reply->length); | 
|  | 65 | if (length >= 1) | 
|  | 66 | fl->fl_end = length -1; | 
|  | 67 | else | 
|  | 68 | fl->fl_end = 0; | 
|  | 69 |  | 
|  | 70 | } | 
|  | 71 | ceph_mdsc_put_request(req); | 
|  | 72 | dout("ceph_lock_message: rule: %d, op: %d, pid: %llu, start: %llu, " | 
|  | 73 | "length: %llu, wait: %d, type: %d, err code %d", (int)lock_type, | 
|  | 74 | (int)operation, (u64)fl->fl_pid, fl->fl_start, | 
|  | 75 | length, wait, fl->fl_type, err); | 
|  | 76 | return err; | 
|  | 77 | } | 
|  | 78 |  | 
|  | 79 | /** | 
|  | 80 | * Attempt to set an fcntl lock. | 
|  | 81 | * For now, this just goes away to the server. Later it may be more awesome. | 
|  | 82 | */ | 
|  | 83 | int ceph_lock(struct file *file, int cmd, struct file_lock *fl) | 
|  | 84 | { | 
|  | 85 | u8 lock_cmd; | 
|  | 86 | int err; | 
|  | 87 | u8 wait = 0; | 
|  | 88 | u16 op = CEPH_MDS_OP_SETFILELOCK; | 
|  | 89 |  | 
|  | 90 | fl->fl_nspid = get_pid(task_tgid(current)); | 
|  | 91 | dout("ceph_lock, fl_pid:%d", fl->fl_pid); | 
|  | 92 |  | 
|  | 93 | /* set wait bit as appropriate, then make command as Ceph expects it*/ | 
|  | 94 | if (F_SETLKW == cmd) | 
|  | 95 | wait = 1; | 
|  | 96 | if (F_GETLK == cmd) | 
|  | 97 | op = CEPH_MDS_OP_GETFILELOCK; | 
|  | 98 |  | 
|  | 99 | if (F_RDLCK == fl->fl_type) | 
|  | 100 | lock_cmd = CEPH_LOCK_SHARED; | 
|  | 101 | else if (F_WRLCK == fl->fl_type) | 
|  | 102 | lock_cmd = CEPH_LOCK_EXCL; | 
|  | 103 | else | 
|  | 104 | lock_cmd = CEPH_LOCK_UNLOCK; | 
|  | 105 |  | 
|  | 106 | err = ceph_lock_message(CEPH_LOCK_FCNTL, op, file, lock_cmd, wait, fl); | 
|  | 107 | if (!err) { | 
|  | 108 | if ( op != CEPH_MDS_OP_GETFILELOCK ){ | 
|  | 109 | dout("mds locked, locking locally"); | 
|  | 110 | err = posix_lock_file(file, fl, NULL); | 
|  | 111 | if (err && (CEPH_MDS_OP_SETFILELOCK == op)) { | 
|  | 112 | /* undo! This should only happen if | 
|  | 113 | * the kernel detects local | 
|  | 114 | * deadlock. */ | 
|  | 115 | ceph_lock_message(CEPH_LOCK_FCNTL, op, file, | 
|  | 116 | CEPH_LOCK_UNLOCK, 0, fl); | 
|  | 117 | dout("got %d on posix_lock_file, undid lock", | 
|  | 118 | err); | 
|  | 119 | } | 
|  | 120 | } | 
|  | 121 |  | 
|  | 122 | } else if (err == -ERESTARTSYS) { | 
|  | 123 | dout("undoing lock\n"); | 
|  | 124 | ceph_lock_message(CEPH_LOCK_FCNTL, op, file, | 
|  | 125 | CEPH_LOCK_UNLOCK, 0, fl); | 
|  | 126 | } | 
|  | 127 | return err; | 
|  | 128 | } | 
|  | 129 |  | 
|  | 130 | int ceph_flock(struct file *file, int cmd, struct file_lock *fl) | 
|  | 131 | { | 
|  | 132 | u8 lock_cmd; | 
|  | 133 | int err; | 
|  | 134 | u8 wait = 1; | 
|  | 135 |  | 
|  | 136 | fl->fl_nspid = get_pid(task_tgid(current)); | 
|  | 137 | dout("ceph_flock, fl_pid:%d", fl->fl_pid); | 
|  | 138 |  | 
|  | 139 | /* set wait bit, then clear it out of cmd*/ | 
|  | 140 | if (cmd & LOCK_NB) | 
|  | 141 | wait = 0; | 
|  | 142 | cmd = cmd & (LOCK_SH | LOCK_EX | LOCK_UN); | 
|  | 143 | /* set command sequence that Ceph wants to see: | 
|  | 144 | shared lock, exclusive lock, or unlock */ | 
|  | 145 | if (LOCK_SH == cmd) | 
|  | 146 | lock_cmd = CEPH_LOCK_SHARED; | 
|  | 147 | else if (LOCK_EX == cmd) | 
|  | 148 | lock_cmd = CEPH_LOCK_EXCL; | 
|  | 149 | else | 
|  | 150 | lock_cmd = CEPH_LOCK_UNLOCK; | 
|  | 151 |  | 
|  | 152 | err = ceph_lock_message(CEPH_LOCK_FLOCK, CEPH_MDS_OP_SETFILELOCK, | 
|  | 153 | file, lock_cmd, wait, fl); | 
|  | 154 | if (!err) { | 
|  | 155 | err = flock_lock_file_wait(file, fl); | 
|  | 156 | if (err) { | 
|  | 157 | ceph_lock_message(CEPH_LOCK_FLOCK, | 
|  | 158 | CEPH_MDS_OP_SETFILELOCK, | 
|  | 159 | file, CEPH_LOCK_UNLOCK, 0, fl); | 
|  | 160 | dout("got %d on flock_lock_file_wait, undid lock", err); | 
|  | 161 | } | 
|  | 162 | } else if (err == -ERESTARTSYS) { | 
|  | 163 | dout("undoing lock\n"); | 
|  | 164 | ceph_lock_message(CEPH_LOCK_FLOCK, | 
|  | 165 | CEPH_MDS_OP_SETFILELOCK, | 
|  | 166 | file, CEPH_LOCK_UNLOCK, 0, fl); | 
|  | 167 | } | 
|  | 168 | return err; | 
|  | 169 | } | 
|  | 170 |  | 
|  | 171 | /** | 
|  | 172 | * Must be called with BKL already held. Fills in the passed | 
|  | 173 | * counter variables, so you can prepare pagelist metadata before calling | 
|  | 174 | * ceph_encode_locks. | 
|  | 175 | */ | 
|  | 176 | void ceph_count_locks(struct inode *inode, int *fcntl_count, int *flock_count) | 
|  | 177 | { | 
|  | 178 | struct file_lock *lock; | 
|  | 179 |  | 
|  | 180 | *fcntl_count = 0; | 
|  | 181 | *flock_count = 0; | 
|  | 182 |  | 
|  | 183 | for (lock = inode->i_flock; lock != NULL; lock = lock->fl_next) { | 
|  | 184 | if (lock->fl_flags & FL_POSIX) | 
|  | 185 | ++(*fcntl_count); | 
|  | 186 | else if (lock->fl_flags & FL_FLOCK) | 
|  | 187 | ++(*flock_count); | 
|  | 188 | } | 
|  | 189 | dout("counted %d flock locks and %d fcntl locks", | 
|  | 190 | *flock_count, *fcntl_count); | 
|  | 191 | } | 
|  | 192 |  | 
|  | 193 | /** | 
|  | 194 | * Encode the flock and fcntl locks for the given inode into the ceph_filelock | 
|  | 195 | * array. Must be called with lock_flocks() already held. | 
|  | 196 | * If we encounter more of a specific lock type than expected, return -ENOSPC. | 
|  | 197 | */ | 
|  | 198 | int ceph_encode_locks_to_buffer(struct inode *inode, | 
|  | 199 | struct ceph_filelock *flocks, | 
|  | 200 | int num_fcntl_locks, int num_flock_locks) | 
|  | 201 | { | 
|  | 202 | struct file_lock *lock; | 
|  | 203 | int err = 0; | 
|  | 204 | int seen_fcntl = 0; | 
|  | 205 | int seen_flock = 0; | 
|  | 206 | int l = 0; | 
|  | 207 |  | 
|  | 208 | dout("encoding %d flock and %d fcntl locks", num_flock_locks, | 
|  | 209 | num_fcntl_locks); | 
|  | 210 |  | 
|  | 211 | for (lock = inode->i_flock; lock != NULL; lock = lock->fl_next) { | 
|  | 212 | if (lock->fl_flags & FL_POSIX) { | 
|  | 213 | ++seen_fcntl; | 
|  | 214 | if (seen_fcntl > num_fcntl_locks) { | 
|  | 215 | err = -ENOSPC; | 
|  | 216 | goto fail; | 
|  | 217 | } | 
|  | 218 | err = lock_to_ceph_filelock(lock, &flocks[l]); | 
|  | 219 | if (err) | 
|  | 220 | goto fail; | 
|  | 221 | ++l; | 
|  | 222 | } | 
|  | 223 | } | 
|  | 224 | for (lock = inode->i_flock; lock != NULL; lock = lock->fl_next) { | 
|  | 225 | if (lock->fl_flags & FL_FLOCK) { | 
|  | 226 | ++seen_flock; | 
|  | 227 | if (seen_flock > num_flock_locks) { | 
|  | 228 | err = -ENOSPC; | 
|  | 229 | goto fail; | 
|  | 230 | } | 
|  | 231 | err = lock_to_ceph_filelock(lock, &flocks[l]); | 
|  | 232 | if (err) | 
|  | 233 | goto fail; | 
|  | 234 | ++l; | 
|  | 235 | } | 
|  | 236 | } | 
|  | 237 | fail: | 
|  | 238 | return err; | 
|  | 239 | } | 
|  | 240 |  | 
|  | 241 | /** | 
|  | 242 | * Copy the encoded flock and fcntl locks into the pagelist. | 
|  | 243 | * Format is: #fcntl locks, sequential fcntl locks, #flock locks, | 
|  | 244 | * sequential flock locks. | 
|  | 245 | * Returns zero on success. | 
|  | 246 | */ | 
|  | 247 | int ceph_locks_to_pagelist(struct ceph_filelock *flocks, | 
|  | 248 | struct ceph_pagelist *pagelist, | 
|  | 249 | int num_fcntl_locks, int num_flock_locks) | 
|  | 250 | { | 
|  | 251 | int err = 0; | 
|  | 252 | __le32 nlocks; | 
|  | 253 |  | 
|  | 254 | nlocks = cpu_to_le32(num_fcntl_locks); | 
|  | 255 | err = ceph_pagelist_append(pagelist, &nlocks, sizeof(nlocks)); | 
|  | 256 | if (err) | 
|  | 257 | goto out_fail; | 
|  | 258 |  | 
|  | 259 | err = ceph_pagelist_append(pagelist, flocks, | 
|  | 260 | num_fcntl_locks * sizeof(*flocks)); | 
|  | 261 | if (err) | 
|  | 262 | goto out_fail; | 
|  | 263 |  | 
|  | 264 | nlocks = cpu_to_le32(num_flock_locks); | 
|  | 265 | err = ceph_pagelist_append(pagelist, &nlocks, sizeof(nlocks)); | 
|  | 266 | if (err) | 
|  | 267 | goto out_fail; | 
|  | 268 |  | 
|  | 269 | err = ceph_pagelist_append(pagelist, | 
|  | 270 | &flocks[num_fcntl_locks], | 
|  | 271 | num_flock_locks * sizeof(*flocks)); | 
|  | 272 | out_fail: | 
|  | 273 | return err; | 
|  | 274 | } | 
|  | 275 |  | 
|  | 276 | /* | 
|  | 277 | * Given a pointer to a lock, convert it to a ceph filelock | 
|  | 278 | */ | 
|  | 279 | int lock_to_ceph_filelock(struct file_lock *lock, | 
|  | 280 | struct ceph_filelock *cephlock) | 
|  | 281 | { | 
|  | 282 | int err = 0; | 
|  | 283 |  | 
|  | 284 | cephlock->start = cpu_to_le64(lock->fl_start); | 
|  | 285 | cephlock->length = cpu_to_le64(lock->fl_end - lock->fl_start + 1); | 
|  | 286 | cephlock->client = cpu_to_le64(0); | 
|  | 287 | cephlock->pid = cpu_to_le64(lock->fl_pid); | 
|  | 288 | cephlock->pid_namespace = | 
|  | 289 | cpu_to_le64((u64)(unsigned long)lock->fl_nspid); | 
|  | 290 |  | 
|  | 291 | switch (lock->fl_type) { | 
|  | 292 | case F_RDLCK: | 
|  | 293 | cephlock->type = CEPH_LOCK_SHARED; | 
|  | 294 | break; | 
|  | 295 | case F_WRLCK: | 
|  | 296 | cephlock->type = CEPH_LOCK_EXCL; | 
|  | 297 | break; | 
|  | 298 | case F_UNLCK: | 
|  | 299 | cephlock->type = CEPH_LOCK_UNLOCK; | 
|  | 300 | break; | 
|  | 301 | default: | 
|  | 302 | dout("Have unknown lock type %d", lock->fl_type); | 
|  | 303 | err = -EINVAL; | 
|  | 304 | } | 
|  | 305 |  | 
|  | 306 | return err; | 
|  | 307 | } |