Blame view
fs/exofs/super.c
25.2 KB
ba9e5e98c
|
1 2 |
/* * Copyright (C) 2005, 2006 |
27d2e1491
|
3 |
* Avishay Traeger (avishay@gmail.com) |
ba9e5e98c
|
4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 |
* Copyright (C) 2008, 2009 * Boaz Harrosh <bharrosh@panasas.com> * * Copyrights for code taken from ext2: * Copyright (C) 1992, 1993, 1994, 1995 * Remy Card (card@masi.ibp.fr) * Laboratoire MASI - Institut Blaise Pascal * Universite Pierre et Marie Curie (Paris VI) * from * linux/fs/minix/inode.c * Copyright (C) 1991, 1992 Linus Torvalds * * This file is part of exofs. * * exofs is free software; you can redistribute it and/or modify * it under the terms of the GNU General Public License as published by * the Free Software Foundation. Since it is based on ext2, and the only * valid version of GPL for the Linux kernel is version 2, the only valid * version of GPL for exofs is version 2. * * exofs is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * * You should have received a copy of the GNU General Public License * along with exofs; if not, write to the Free Software * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA */ #include <linux/string.h> #include <linux/parser.h> #include <linux/vfs.h> #include <linux/random.h> |
8cf74b393
|
38 |
#include <linux/exportfs.h> |
5a0e3ad6a
|
39 |
#include <linux/slab.h> |
ba9e5e98c
|
40 41 42 43 44 45 46 47 48 49 50 |
#include "exofs.h" /****************************************************************************** * MOUNT OPTIONS *****************************************************************************/ /* * struct to hold what we get from mount options */ struct exofs_mountopt { |
9ed964843
|
51 |
bool is_osdname; |
ba9e5e98c
|
52 53 54 55 56 57 58 59 |
const char *dev_name; uint64_t pid; int timeout; }; /* * exofs-specific mount-time options. */ |
9ed964843
|
60 |
enum { Opt_name, Opt_pid, Opt_to, Opt_err }; |
ba9e5e98c
|
61 62 63 64 65 66 67 |
/* * Our mount-time options. These should ideally be 64-bit unsigned, but the * kernel's parsing functions do not currently support that. 32-bit should be * sufficient for most applications now. */ static match_table_t tokens = { |
9ed964843
|
68 |
{Opt_name, "osdname=%s"}, |
ba9e5e98c
|
69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 |
{Opt_pid, "pid=%u"}, {Opt_to, "to=%u"}, {Opt_err, NULL} }; /* * The main option parsing method. Also makes sure that all of the mandatory * mount options were set. */ static int parse_options(char *options, struct exofs_mountopt *opts) { char *p; substring_t args[MAX_OPT_ARGS]; int option; bool s_pid = false; EXOFS_DBGMSG("parse_options %s ", options); /* defaults */ memset(opts, 0, sizeof(*opts)); opts->timeout = BLK_DEFAULT_SG_TIMEOUT; while ((p = strsep(&options, ",")) != NULL) { int token; char str[32]; if (!*p) continue; token = match_token(p, tokens, args); switch (token) { |
9ed964843
|
100 101 102 103 104 105 106 107 |
case Opt_name: opts->dev_name = match_strdup(&args[0]); if (unlikely(!opts->dev_name)) { EXOFS_ERR("Error allocating dev_name"); return -ENOMEM; } opts->is_osdname = true; break; |
ba9e5e98c
|
108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 |
case Opt_pid: if (0 == match_strlcpy(str, &args[0], sizeof(str))) return -EINVAL; opts->pid = simple_strtoull(str, NULL, 0); if (opts->pid < EXOFS_MIN_PID) { EXOFS_ERR("Partition ID must be >= %u", EXOFS_MIN_PID); return -EINVAL; } s_pid = 1; break; case Opt_to: if (match_int(&args[0], &option)) return -EINVAL; if (option <= 0) { EXOFS_ERR("Timout must be > 0"); return -EINVAL; } opts->timeout = option * HZ; break; } } if (!s_pid) { EXOFS_ERR("Need to specify the following options: "); EXOFS_ERR(" -o pid=pid_no_to_use "); return -EINVAL; } return 0; } /****************************************************************************** * INODE CACHE *****************************************************************************/ /* * Our inode cache. Isn't it pretty? */ static struct kmem_cache *exofs_inode_cachep; /* * Allocate an inode in the cache */ static struct inode *exofs_alloc_inode(struct super_block *sb) { struct exofs_i_info *oi; oi = kmem_cache_alloc(exofs_inode_cachep, GFP_KERNEL); if (!oi) return NULL; oi->vfs_inode.i_version = 1; return &oi->vfs_inode; } |
fa0d7e3de
|
165 166 167 168 169 170 |
static void exofs_i_callback(struct rcu_head *head) { struct inode *inode = container_of(head, struct inode, i_rcu); INIT_LIST_HEAD(&inode->i_dentry); kmem_cache_free(exofs_inode_cachep, exofs_i(inode)); } |
ba9e5e98c
|
171 172 173 174 175 |
/* * Remove an inode from the cache */ static void exofs_destroy_inode(struct inode *inode) { |
fa0d7e3de
|
176 |
call_rcu(&inode->i_rcu, exofs_i_callback); |
ba9e5e98c
|
177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 |
} /* * Initialize the inode */ static void exofs_init_once(void *foo) { struct exofs_i_info *oi = foo; inode_init_once(&oi->vfs_inode); } /* * Create and initialize the inode cache */ static int init_inodecache(void) { exofs_inode_cachep = kmem_cache_create("exofs_inode_cache", sizeof(struct exofs_i_info), 0, SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, exofs_init_once); if (exofs_inode_cachep == NULL) return -ENOMEM; return 0; } /* * Destroy the inode cache */ static void destroy_inodecache(void) { kmem_cache_destroy(exofs_inode_cachep); } /****************************************************************************** * SUPERBLOCK FUNCTIONS *****************************************************************************/ static const struct super_operations exofs_sops; |
8cf74b393
|
215 |
static const struct export_operations exofs_export_ops; |
ba9e5e98c
|
216 |
|
1cea312ad
|
217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 |
static const struct osd_attr g_attr_sb_stats = ATTR_DEF( EXOFS_APAGE_SB_DATA, EXOFS_ATTR_SB_STATS, sizeof(struct exofs_sb_stats)); static int __sbi_read_stats(struct exofs_sb_info *sbi) { struct osd_attr attrs[] = { [0] = g_attr_sb_stats, }; struct exofs_io_state *ios; int ret; ret = exofs_get_io_state(&sbi->layout, &ios); if (unlikely(ret)) { EXOFS_ERR("%s: exofs_get_io_state failed. ", __func__); return ret; } ios->cred = sbi->s_cred; ios->in_attr = attrs; ios->in_attr_len = ARRAY_SIZE(attrs); ret = exofs_sbi_read(ios); if (unlikely(ret)) { EXOFS_ERR("Error reading super_block stats => %d ", ret); goto out; } ret = extract_attr_from_ios(ios, &attrs[0]); if (ret) { EXOFS_ERR("%s: extract_attr of sb_stats failed ", __func__); goto out; } if (attrs[0].len) { struct exofs_sb_stats *ess; if (unlikely(attrs[0].len != sizeof(*ess))) { EXOFS_ERR("%s: Wrong version of exofs_sb_stats " "size(%d) != expected(%zd) ", __func__, attrs[0].len, sizeof(*ess)); goto out; } ess = attrs[0].val_ptr; sbi->s_nextid = le64_to_cpu(ess->s_nextid); sbi->s_numfiles = le32_to_cpu(ess->s_numfiles); } out: exofs_put_io_state(ios); return ret; } static void stats_done(struct exofs_io_state *ios, void *p) { exofs_put_io_state(ios); /* Good thanks nothing to do anymore */ } /* Asynchronously write the stats attribute */ int exofs_sbi_write_stats(struct exofs_sb_info *sbi) { struct osd_attr attrs[] = { [0] = g_attr_sb_stats, }; struct exofs_io_state *ios; int ret; ret = exofs_get_io_state(&sbi->layout, &ios); if (unlikely(ret)) { EXOFS_ERR("%s: exofs_get_io_state failed. ", __func__); return ret; } sbi->s_ess.s_nextid = cpu_to_le64(sbi->s_nextid); sbi->s_ess.s_numfiles = cpu_to_le64(sbi->s_numfiles); attrs[0].val_ptr = &sbi->s_ess; ios->cred = sbi->s_cred; ios->done = stats_done; ios->private = sbi; ios->out_attr = attrs; ios->out_attr_len = ARRAY_SIZE(attrs); ret = exofs_sbi_write(ios); if (unlikely(ret)) { EXOFS_ERR("%s: exofs_sbi_write failed. ", __func__); exofs_put_io_state(ios); } return ret; } |
ba9e5e98c
|
317 318 319 |
/* * Write the superblock to the OSD */ |
baaf94cdc
|
320 |
int exofs_sync_fs(struct super_block *sb, int wait) |
ba9e5e98c
|
321 322 323 |
{ struct exofs_sb_info *sbi; struct exofs_fscb *fscb; |
06886a5a3
|
324 |
struct exofs_io_state *ios; |
80e09fb94
|
325 |
int ret = -ENOMEM; |
ba9e5e98c
|
326 |
|
1cea312ad
|
327 328 329 |
fscb = kmalloc(sizeof(*fscb), GFP_KERNEL); if (unlikely(!fscb)) return -ENOMEM; |
ba9e5e98c
|
330 |
sbi = sb->s_fs_info; |
06886a5a3
|
331 |
|
1cea312ad
|
332 333 334 335 336 337 338 |
/* NOTE: We no longer dirty the super_block anywhere in exofs. The * reason we write the fscb here on unmount is so we can stay backwards * compatible with fscb->s_version == 1. (What we are not compatible * with is if a new version FS crashed and then we try to mount an old * version). Otherwise the exofs_fscb is read-only from mkfs time. All * the writeable info is set in exofs_sbi_write_stats() above. */ |
45d3abcb1
|
339 |
ret = exofs_get_io_state(&sbi->layout, &ios); |
1cea312ad
|
340 |
if (unlikely(ret)) |
06886a5a3
|
341 |
goto out; |
1cea312ad
|
342 |
lock_super(sb); |
04dc1e88a
|
343 |
ios->length = offsetof(struct exofs_fscb, s_dev_table_oid); |
06886a5a3
|
344 |
memset(fscb, 0, ios->length); |
ba9e5e98c
|
345 346 347 348 |
fscb->s_nextid = cpu_to_le64(sbi->s_nextid); fscb->s_numfiles = cpu_to_le32(sbi->s_numfiles); fscb->s_magic = cpu_to_le16(sb->s_magic); fscb->s_newfs = 0; |
04dc1e88a
|
349 |
fscb->s_version = EXOFS_FSCB_VER; |
ba9e5e98c
|
350 |
|
06886a5a3
|
351 352 353 354 |
ios->obj.id = EXOFS_SUPER_ID; ios->offset = 0; ios->kern_buff = fscb; ios->cred = sbi->s_cred; |
ba9e5e98c
|
355 |
|
06886a5a3
|
356 |
ret = exofs_sbi_write(ios); |
1cea312ad
|
357 |
if (unlikely(ret)) |
06886a5a3
|
358 359 |
EXOFS_ERR("%s: exofs_sbi_write failed. ", __func__); |
1cea312ad
|
360 361 |
else sb->s_dirt = 0; |
ba9e5e98c
|
362 |
|
1cea312ad
|
363 |
unlock_super(sb); |
ba9e5e98c
|
364 |
out: |
06886a5a3
|
365 366 367 |
EXOFS_DBGMSG("s_nextid=0x%llx ret=%d ", _LLU(sbi->s_nextid), ret); exofs_put_io_state(ios); |
1cea312ad
|
368 |
kfree(fscb); |
80e09fb94
|
369 370 371 372 373 374 375 376 377 |
return ret; } static void exofs_write_super(struct super_block *sb) { if (!(sb->s_flags & MS_RDONLY)) exofs_sync_fs(sb, 1); else sb->s_dirt = 0; |
ba9e5e98c
|
378 |
} |
19fe294f2
|
379 380 381 382 383 384 385 386 387 |
static void _exofs_print_device(const char *msg, const char *dev_path, struct osd_dev *od, u64 pid) { const struct osd_dev_info *odi = osduld_device_info(od); printk(KERN_NOTICE "exofs: %s %s osd_name-%s pid-0x%llx ", msg, dev_path ?: "", odi->osdname, _LLU(pid)); } |
04dc1e88a
|
388 389 |
void exofs_free_sbi(struct exofs_sb_info *sbi) { |
45d3abcb1
|
390 391 392 |
while (sbi->layout.s_numdevs) { int i = --sbi->layout.s_numdevs; struct osd_dev *od = sbi->layout.s_ods[i]; |
04dc1e88a
|
393 394 |
if (od) { |
45d3abcb1
|
395 |
sbi->layout.s_ods[i] = NULL; |
04dc1e88a
|
396 397 398 399 400 |
osduld_put_device(od); } } kfree(sbi); } |
ba9e5e98c
|
401 402 403 404 405 406 407 408 409 410 411 412 413 |
/* * This function is called when the vfs is freeing the superblock. We just * need to free our own part. */ static void exofs_put_super(struct super_block *sb) { int num_pend; struct exofs_sb_info *sbi = sb->s_fs_info; /* make sure there are no pending commands */ for (num_pend = atomic_read(&sbi->s_curr_pending); num_pend > 0; num_pend = atomic_read(&sbi->s_curr_pending)) { wait_queue_head_t wq; |
a49fb4c3d
|
414 415 416 417 418 |
printk(KERN_NOTICE "%s: !!Pending operations in flight. " "This is a BUG. please report to osd-dev@open-osd.org ", __func__); |
ba9e5e98c
|
419 420 421 422 423 |
init_waitqueue_head(&wq); wait_event_timeout(wq, (atomic_read(&sbi->s_curr_pending) == 0), msecs_to_jiffies(100)); } |
45d3abcb1
|
424 425 |
_exofs_print_device("Unmounting", NULL, sbi->layout.s_ods[0], sbi->layout.s_pid); |
04dc1e88a
|
426 |
|
b3d0ab7e6
|
427 |
bdi_destroy(&sbi->bdi); |
04dc1e88a
|
428 |
exofs_free_sbi(sbi); |
ba9e5e98c
|
429 430 |
sb->s_fs_info = NULL; } |
04dc1e88a
|
431 432 433 |
static int _read_and_match_data_map(struct exofs_sb_info *sbi, unsigned numdevs, struct exofs_device_table *dt) { |
5d952b839
|
434 |
u64 stripe_length; |
04dc1e88a
|
435 436 437 438 439 440 441 442 443 444 445 446 |
sbi->data_map.odm_num_comps = le32_to_cpu(dt->dt_data_map.cb_num_comps); sbi->data_map.odm_stripe_unit = le64_to_cpu(dt->dt_data_map.cb_stripe_unit); sbi->data_map.odm_group_width = le32_to_cpu(dt->dt_data_map.cb_group_width); sbi->data_map.odm_group_depth = le32_to_cpu(dt->dt_data_map.cb_group_depth); sbi->data_map.odm_mirror_cnt = le32_to_cpu(dt->dt_data_map.cb_mirror_cnt); sbi->data_map.odm_raid_algorithm = le32_to_cpu(dt->dt_data_map.cb_raid_algorithm); |
50a76fd3c
|
447 |
/* FIXME: Only raid0 for now. if not so, do not mount */ |
5d952b839
|
448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 |
if (sbi->data_map.odm_num_comps != numdevs) { EXOFS_ERR("odm_num_comps(%u) != numdevs(%u) ", sbi->data_map.odm_num_comps, numdevs); return -EINVAL; } if (sbi->data_map.odm_raid_algorithm != PNFS_OSD_RAID_0) { EXOFS_ERR("Only RAID_0 for now "); return -EINVAL; } if (0 != (numdevs % (sbi->data_map.odm_mirror_cnt + 1))) { EXOFS_ERR("Data Map wrong, numdevs=%d mirrors=%d ", numdevs, sbi->data_map.odm_mirror_cnt); return -EINVAL; } |
5d952b839
|
465 466 467 468 469 470 471 472 473 474 |
if (0 != (sbi->data_map.odm_stripe_unit & ~PAGE_MASK)) { EXOFS_ERR("Stripe Unit(0x%llx)" " must be Multples of PAGE_SIZE(0x%lx) ", _LLU(sbi->data_map.odm_stripe_unit), PAGE_SIZE); return -EINVAL; } sbi->layout.stripe_unit = sbi->data_map.odm_stripe_unit; sbi->layout.mirrors_p1 = sbi->data_map.odm_mirror_cnt + 1; |
50a76fd3c
|
475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 |
if (sbi->data_map.odm_group_width) { sbi->layout.group_width = sbi->data_map.odm_group_width; sbi->layout.group_depth = sbi->data_map.odm_group_depth; if (!sbi->layout.group_depth) { EXOFS_ERR("group_depth == 0 && group_width != 0 "); return -EINVAL; } sbi->layout.group_count = sbi->data_map.odm_num_comps / sbi->layout.mirrors_p1 / sbi->data_map.odm_group_width; } else { if (sbi->data_map.odm_group_depth) { printk(KERN_NOTICE "Warning: group_depth ignored " "group_width == 0 && group_depth == %d ", sbi->data_map.odm_group_depth); sbi->data_map.odm_group_depth = 0; } sbi->layout.group_width = sbi->data_map.odm_num_comps / |
5d952b839
|
496 |
sbi->layout.mirrors_p1; |
50a76fd3c
|
497 498 499 500 501 502 503 504 505 506 507 |
sbi->layout.group_depth = -1; sbi->layout.group_count = 1; } stripe_length = (u64)sbi->layout.group_width * sbi->layout.stripe_unit; if (stripe_length >= (1ULL << 32)) { EXOFS_ERR("Total Stripe length(0x%llx)" " >= 32bit is not supported ", _LLU(stripe_length)); return -EINVAL; } |
5d952b839
|
508 509 |
return 0; |
04dc1e88a
|
510 |
} |
66cd6cad4
|
511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 |
static unsigned __ra_pages(struct exofs_layout *layout) { const unsigned _MIN_RA = 32; /* min 128K read-ahead */ unsigned ra_pages = layout->group_width * layout->stripe_unit / PAGE_SIZE; unsigned max_io_pages = exofs_max_io_pages(layout, ~0); ra_pages *= 2; /* two stripes */ if (ra_pages < _MIN_RA) ra_pages = roundup(_MIN_RA, ra_pages / 2); if (ra_pages > max_io_pages) ra_pages = max_io_pages; return ra_pages; } |
04dc1e88a
|
527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 |
/* @odi is valid only as long as @fscb_dev is valid */ static int exofs_devs_2_odi(struct exofs_dt_device_info *dt_dev, struct osd_dev_info *odi) { odi->systemid_len = le32_to_cpu(dt_dev->systemid_len); memcpy(odi->systemid, dt_dev->systemid, odi->systemid_len); odi->osdname_len = le32_to_cpu(dt_dev->osdname_len); odi->osdname = dt_dev->osdname; /* FIXME support long names. Will need a _put function */ if (dt_dev->long_name_offset) return -EINVAL; /* Make sure osdname is printable! * mkexofs should give us space for a null-terminator else the * device-table is invalid. */ if (unlikely(odi->osdname_len >= sizeof(dt_dev->osdname))) odi->osdname_len = sizeof(dt_dev->osdname) - 1; dt_dev->osdname[odi->osdname_len] = 0; /* If it's all zeros something is bad we read past end-of-obj */ return !(odi->systemid_len || odi->osdname_len); } static int exofs_read_lookup_dev_table(struct exofs_sb_info **psbi, unsigned table_count) { struct exofs_sb_info *sbi = *psbi; struct osd_dev *fscb_od; |
45d3abcb1
|
558 |
struct osd_obj_id obj = {.partition = sbi->layout.s_pid, |
04dc1e88a
|
559 560 561 562 563 564 565 566 567 568 569 570 571 572 |
.id = EXOFS_DEVTABLE_ID}; struct exofs_device_table *dt; unsigned table_bytes = table_count * sizeof(dt->dt_dev_table[0]) + sizeof(*dt); unsigned numdevs, i; int ret; dt = kmalloc(table_bytes, GFP_KERNEL); if (unlikely(!dt)) { EXOFS_ERR("ERROR: allocating %x bytes for device table ", table_bytes); return -ENOMEM; } |
45d3abcb1
|
573 574 575 |
fscb_od = sbi->layout.s_ods[0]; sbi->layout.s_ods[0] = NULL; sbi->layout.s_numdevs = 0; |
04dc1e88a
|
576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 |
ret = exofs_read_kern(fscb_od, sbi->s_cred, &obj, 0, dt, table_bytes); if (unlikely(ret)) { EXOFS_ERR("ERROR: reading device table "); goto out; } numdevs = le64_to_cpu(dt->dt_num_devices); if (unlikely(!numdevs)) { ret = -EINVAL; goto out; } WARN_ON(table_count != numdevs); ret = _read_and_match_data_map(sbi, numdevs, dt); if (unlikely(ret)) goto out; if (likely(numdevs > 1)) { |
45d3abcb1
|
595 |
unsigned size = numdevs * sizeof(sbi->layout.s_ods[0]); |
04dc1e88a
|
596 597 598 599 600 601 |
sbi = krealloc(sbi, sizeof(*sbi) + size, GFP_KERNEL); if (unlikely(!sbi)) { ret = -ENOMEM; goto out; } |
45d3abcb1
|
602 603 |
memset(&sbi->layout.s_ods[1], 0, size - sizeof(sbi->layout.s_ods[0])); |
04dc1e88a
|
604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 |
*psbi = sbi; } for (i = 0; i < numdevs; i++) { struct exofs_fscb fscb; struct osd_dev_info odi; struct osd_dev *od; if (exofs_devs_2_odi(&dt->dt_dev_table[i], &odi)) { EXOFS_ERR("ERROR: Read all-zeros device entry "); ret = -EINVAL; goto out; } printk(KERN_NOTICE "Add device[%d]: osd_name-%s ", i, odi.osdname); /* On all devices the device table is identical. The user can * specify any one of the participating devices on the command * line. We always keep them in device-table order. */ if (fscb_od && osduld_device_same(fscb_od, &odi)) { |
45d3abcb1
|
628 629 |
sbi->layout.s_ods[i] = fscb_od; ++sbi->layout.s_numdevs; |
04dc1e88a
|
630 631 632 633 634 |
fscb_od = NULL; continue; } od = osduld_info_lookup(&odi); |
2c722c9a4
|
635 |
if (IS_ERR(od)) { |
04dc1e88a
|
636 637 638 639 640 641 |
ret = PTR_ERR(od); EXOFS_ERR("ERROR: device requested is not found " "osd_name-%s =>%d ", odi.osdname, ret); goto out; } |
45d3abcb1
|
642 643 |
sbi->layout.s_ods[i] = od; ++sbi->layout.s_numdevs; |
04dc1e88a
|
644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 |
/* Read the fscb of the other devices to make sure the FS * partition is there. */ ret = exofs_read_kern(od, sbi->s_cred, &obj, 0, &fscb, sizeof(fscb)); if (unlikely(ret)) { EXOFS_ERR("ERROR: Malformed participating device " "error reading fscb osd_name-%s ", odi.osdname); goto out; } /* TODO: verify other information is correct and FS-uuid * matches. Benny what did you say about device table * generation and old devices? */ } out: kfree(dt); if (unlikely(!ret && fscb_od)) { EXOFS_ERR( "ERROR: Bad device-table container device not present "); osduld_put_device(fscb_od); ret = -EINVAL; } return ret; } |
ba9e5e98c
|
676 677 678 679 680 681 682 683 |
/* * Read the superblock from the OSD and fill in the fields */ static int exofs_fill_super(struct super_block *sb, void *data, int silent) { struct inode *root; struct exofs_mountopt *opts = data; struct exofs_sb_info *sbi; /*extended info */ |
06886a5a3
|
684 |
struct osd_dev *od; /* Master device */ |
ba9e5e98c
|
685 |
struct exofs_fscb fscb; /*on-disk superblock info */ |
ba9e5e98c
|
686 |
struct osd_obj_id obj; |
04dc1e88a
|
687 |
unsigned table_count; |
ba9e5e98c
|
688 689 690 691 692 |
int ret; sbi = kzalloc(sizeof(*sbi), GFP_KERNEL); if (!sbi) return -ENOMEM; |
ba9e5e98c
|
693 |
|
b3d0ab7e6
|
694 695 696 |
ret = bdi_setup_and_register(&sbi->bdi, "exofs", BDI_CAP_MAP_COPY); if (ret) goto free_bdi; |
ba9e5e98c
|
697 |
/* use mount options to fill superblock */ |
9ed964843
|
698 699 700 701 702 703 704 705 706 |
if (opts->is_osdname) { struct osd_dev_info odi = {.systemid_len = 0}; odi.osdname_len = strlen(opts->dev_name); odi.osdname = (u8 *)opts->dev_name; od = osduld_info_lookup(&odi); } else { od = osduld_path_lookup(opts->dev_name); } |
06886a5a3
|
707 |
if (IS_ERR(od)) { |
9ed964843
|
708 |
ret = -EINVAL; |
ba9e5e98c
|
709 710 |
goto free_sbi; } |
45d3abcb1
|
711 |
/* Default layout in case we do not have a device-table */ |
5d952b839
|
712 713 714 |
sbi->layout.stripe_unit = PAGE_SIZE; sbi->layout.mirrors_p1 = 1; sbi->layout.group_width = 1; |
50a76fd3c
|
715 716 |
sbi->layout.group_depth = -1; sbi->layout.group_count = 1; |
45d3abcb1
|
717 718 719 |
sbi->layout.s_ods[0] = od; sbi->layout.s_numdevs = 1; sbi->layout.s_pid = opts->pid; |
ba9e5e98c
|
720 721 722 723 724 725 726 727 728 729 730 |
sbi->s_timeout = opts->timeout; /* fill in some other data by hand */ memset(sb->s_id, 0, sizeof(sb->s_id)); strcpy(sb->s_id, "exofs"); sb->s_blocksize = EXOFS_BLKSIZE; sb->s_blocksize_bits = EXOFS_BLKSHIFT; sb->s_maxbytes = MAX_LFS_FILESIZE; atomic_set(&sbi->s_curr_pending, 0); sb->s_bdev = NULL; sb->s_dev = 0; |
45d3abcb1
|
731 |
obj.partition = sbi->layout.s_pid; |
ba9e5e98c
|
732 733 |
obj.id = EXOFS_SUPER_ID; exofs_make_credential(sbi->s_cred, &obj); |
06886a5a3
|
734 735 |
ret = exofs_read_kern(od, sbi->s_cred, &obj, 0, &fscb, sizeof(fscb)); if (unlikely(ret)) |
ba9e5e98c
|
736 |
goto free_sbi; |
ba9e5e98c
|
737 738 |
sb->s_magic = le16_to_cpu(fscb.s_magic); |
1cea312ad
|
739 |
/* NOTE: we read below to be backward compatible with old versions */ |
ba9e5e98c
|
740 741 742 743 744 745 746 747 748 749 750 |
sbi->s_nextid = le64_to_cpu(fscb.s_nextid); sbi->s_numfiles = le32_to_cpu(fscb.s_numfiles); /* make sure what we read from the object store is correct */ if (sb->s_magic != EXOFS_SUPER_MAGIC) { if (!silent) EXOFS_ERR("ERROR: Bad magic value "); ret = -EINVAL; goto free_sbi; } |
1cea312ad
|
751 |
if (le32_to_cpu(fscb.s_version) > EXOFS_FSCB_VER) { |
04dc1e88a
|
752 753 754 755 756 757 |
EXOFS_ERR("ERROR: Bad FSCB version expected-%d got-%d ", EXOFS_FSCB_VER, le32_to_cpu(fscb.s_version)); ret = -EINVAL; goto free_sbi; } |
ba9e5e98c
|
758 759 760 761 |
/* start generation numbers from a random point */ get_random_bytes(&sbi->s_next_generation, sizeof(u32)); spin_lock_init(&sbi->s_next_gen_lock); |
04dc1e88a
|
762 763 764 765 766 767 |
table_count = le64_to_cpu(fscb.s_dev_table_count); if (table_count) { ret = exofs_read_lookup_dev_table(&sbi, table_count); if (unlikely(ret)) goto free_sbi; } |
1cea312ad
|
768 |
__sbi_read_stats(sbi); |
ba9e5e98c
|
769 |
/* set up operation vectors */ |
66cd6cad4
|
770 |
sbi->bdi.ra_pages = __ra_pages(&sbi->layout); |
b3d0ab7e6
|
771 |
sb->s_bdi = &sbi->bdi; |
06886a5a3
|
772 |
sb->s_fs_info = sbi; |
ba9e5e98c
|
773 |
sb->s_op = &exofs_sops; |
8cf74b393
|
774 |
sb->s_export_op = &exofs_export_ops; |
ba9e5e98c
|
775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 |
root = exofs_iget(sb, EXOFS_ROOT_ID - EXOFS_OBJ_OFF); if (IS_ERR(root)) { EXOFS_ERR("ERROR: exofs_iget failed "); ret = PTR_ERR(root); goto free_sbi; } sb->s_root = d_alloc_root(root); if (!sb->s_root) { iput(root); EXOFS_ERR("ERROR: get root inode failed "); ret = -ENOMEM; goto free_sbi; } if (!S_ISDIR(root->i_mode)) { dput(sb->s_root); sb->s_root = NULL; EXOFS_ERR("ERROR: corrupt root inode (mode = %hd) ", root->i_mode); ret = -EINVAL; goto free_sbi; } |
45d3abcb1
|
800 801 |
_exofs_print_device("Mounting", opts->dev_name, sbi->layout.s_ods[0], sbi->layout.s_pid); |
9ed964843
|
802 803 |
if (opts->is_osdname) kfree(opts->dev_name); |
06886a5a3
|
804 |
return 0; |
ba9e5e98c
|
805 806 |
free_sbi: |
b3d0ab7e6
|
807 808 |
bdi_destroy(&sbi->bdi); free_bdi: |
06886a5a3
|
809 810 |
EXOFS_ERR("Unable to mount exofs on %s pid=0x%llx err=%d ", |
45d3abcb1
|
811 |
opts->dev_name, sbi->layout.s_pid, ret); |
04dc1e88a
|
812 |
exofs_free_sbi(sbi); |
9ed964843
|
813 814 |
if (opts->is_osdname) kfree(opts->dev_name); |
06886a5a3
|
815 |
return ret; |
ba9e5e98c
|
816 817 818 819 820 |
} /* * Set up the superblock (calls exofs_fill_super eventually) */ |
3c26ff6e4
|
821 |
static struct dentry *exofs_mount(struct file_system_type *type, |
ba9e5e98c
|
822 |
int flags, const char *dev_name, |
3c26ff6e4
|
823 |
void *data) |
ba9e5e98c
|
824 825 826 827 828 829 |
{ struct exofs_mountopt opts; int ret; ret = parse_options(data, &opts); if (ret) |
3c26ff6e4
|
830 |
return ERR_PTR(ret); |
ba9e5e98c
|
831 |
|
9ed964843
|
832 833 |
if (!opts.dev_name) opts.dev_name = dev_name; |
3c26ff6e4
|
834 |
return mount_nodev(type, flags, &opts, exofs_fill_super); |
ba9e5e98c
|
835 836 837 838 839 840 841 842 843 844 |
} /* * Return information about the file system state in the buffer. This is used * by the 'df' command, for example. */ static int exofs_statfs(struct dentry *dentry, struct kstatfs *buf) { struct super_block *sb = dentry->d_sb; struct exofs_sb_info *sbi = sb->s_fs_info; |
06886a5a3
|
845 |
struct exofs_io_state *ios; |
ba9e5e98c
|
846 847 848 849 850 851 852 853 |
struct osd_attr attrs[] = { ATTR_DEF(OSD_APAGE_PARTITION_QUOTAS, OSD_ATTR_PQ_CAPACITY_QUOTA, sizeof(__be64)), ATTR_DEF(OSD_APAGE_PARTITION_INFORMATION, OSD_ATTR_PI_USED_CAPACITY, sizeof(__be64)), }; uint64_t capacity = ULLONG_MAX; uint64_t used = ULLONG_MAX; |
ba9e5e98c
|
854 855 |
uint8_t cred_a[OSD_CAP_LEN]; int ret; |
45d3abcb1
|
856 |
ret = exofs_get_io_state(&sbi->layout, &ios); |
06886a5a3
|
857 858 859 860 |
if (ret) { EXOFS_DBGMSG("exofs_get_io_state failed. "); return ret; |
ba9e5e98c
|
861 |
} |
06886a5a3
|
862 863 864 865 866 867 |
exofs_make_credential(cred_a, &ios->obj); ios->cred = sbi->s_cred; ios->in_attr = attrs; ios->in_attr_len = ARRAY_SIZE(attrs); ret = exofs_sbi_read(ios); |
ba9e5e98c
|
868 869 |
if (unlikely(ret)) goto out; |
06886a5a3
|
870 |
ret = extract_attr_from_ios(ios, &attrs[0]); |
cae012d85
|
871 |
if (likely(!ret)) { |
ba9e5e98c
|
872 |
capacity = get_unaligned_be64(attrs[0].val_ptr); |
cae012d85
|
873 874 875 |
if (unlikely(!capacity)) capacity = ULLONG_MAX; } else |
ba9e5e98c
|
876 877 |
EXOFS_DBGMSG("exofs_statfs: get capacity failed. "); |
06886a5a3
|
878 |
ret = extract_attr_from_ios(ios, &attrs[1]); |
ba9e5e98c
|
879 880 881 882 883 884 885 886 887 |
if (likely(!ret)) used = get_unaligned_be64(attrs[1].val_ptr); else EXOFS_DBGMSG("exofs_statfs: get used-space failed. "); /* fill in the stats buffer */ buf->f_type = EXOFS_SUPER_MAGIC; buf->f_bsize = EXOFS_BLKSIZE; |
cae012d85
|
888 889 |
buf->f_blocks = capacity >> 9; buf->f_bfree = (capacity - used) >> 9; |
ba9e5e98c
|
890 891 892 893 894 895 |
buf->f_bavail = buf->f_bfree; buf->f_files = sbi->s_numfiles; buf->f_ffree = EXOFS_MAX_ID - sbi->s_numfiles; buf->f_namelen = EXOFS_NAME_LEN; out: |
06886a5a3
|
896 |
exofs_put_io_state(ios); |
ba9e5e98c
|
897 898 899 900 901 902 903 |
return ret; } static const struct super_operations exofs_sops = { .alloc_inode = exofs_alloc_inode, .destroy_inode = exofs_destroy_inode, .write_inode = exofs_write_inode, |
4ec70c9b4
|
904 |
.evict_inode = exofs_evict_inode, |
ba9e5e98c
|
905 906 |
.put_super = exofs_put_super, .write_super = exofs_write_super, |
80e09fb94
|
907 |
.sync_fs = exofs_sync_fs, |
ba9e5e98c
|
908 909 910 911 |
.statfs = exofs_statfs, }; /****************************************************************************** |
8cf74b393
|
912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 |
* EXPORT OPERATIONS *****************************************************************************/ struct dentry *exofs_get_parent(struct dentry *child) { unsigned long ino = exofs_parent_ino(child); if (!ino) return NULL; return d_obtain_alias(exofs_iget(child->d_inode->i_sb, ino)); } static struct inode *exofs_nfs_get_inode(struct super_block *sb, u64 ino, u32 generation) { struct inode *inode; inode = exofs_iget(sb, ino); if (IS_ERR(inode)) return ERR_CAST(inode); if (generation && inode->i_generation != generation) { /* we didn't find the right inode.. */ iput(inode); return ERR_PTR(-ESTALE); } return inode; } static struct dentry *exofs_fh_to_dentry(struct super_block *sb, struct fid *fid, int fh_len, int fh_type) { return generic_fh_to_dentry(sb, fid, fh_len, fh_type, exofs_nfs_get_inode); } static struct dentry *exofs_fh_to_parent(struct super_block *sb, struct fid *fid, int fh_len, int fh_type) { return generic_fh_to_parent(sb, fid, fh_len, fh_type, exofs_nfs_get_inode); } static const struct export_operations exofs_export_ops = { .fh_to_dentry = exofs_fh_to_dentry, .fh_to_parent = exofs_fh_to_parent, .get_parent = exofs_get_parent, }; /****************************************************************************** |
ba9e5e98c
|
962 963 964 965 966 967 968 969 970 |
* INSMOD/RMMOD *****************************************************************************/ /* * struct that describes this file system */ static struct file_system_type exofs_type = { .owner = THIS_MODULE, .name = "exofs", |
3c26ff6e4
|
971 |
.mount = exofs_mount, |
ba9e5e98c
|
972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 |
.kill_sb = generic_shutdown_super, }; static int __init init_exofs(void) { int err; err = init_inodecache(); if (err) goto out; err = register_filesystem(&exofs_type); if (err) goto out_d; return 0; out_d: destroy_inodecache(); out: return err; } static void __exit exit_exofs(void) { unregister_filesystem(&exofs_type); destroy_inodecache(); } MODULE_AUTHOR("Avishay Traeger <avishay@gmail.com>"); MODULE_DESCRIPTION("exofs"); MODULE_LICENSE("GPL"); module_init(init_exofs) module_exit(exit_exofs) |