Blame view
fs/exofs/super.c
26.1 KB
ba9e5e98c exofs: super_oper... |
1 2 |
/* * Copyright (C) 2005, 2006 |
27d2e1491 exofs: Remove IBM... |
3 |
* Avishay Traeger (avishay@gmail.com) |
ba9e5e98c exofs: super_oper... |
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> |
143cb494c fs: add module.h ... |
38 |
#include <linux/module.h> |
8cf74b393 exofs: export_ope... |
39 |
#include <linux/exportfs.h> |
5a0e3ad6a include cleanup: ... |
40 |
#include <linux/slab.h> |
ba9e5e98c exofs: super_oper... |
41 42 |
#include "exofs.h" |
85e44df47 exofs: Move exofs... |
43 |
#define EXOFS_DBGMSG2(M...) do {} while (0) |
ba9e5e98c exofs: super_oper... |
44 45 46 47 48 49 50 51 |
/****************************************************************************** * MOUNT OPTIONS *****************************************************************************/ /* * struct to hold what we get from mount options */ struct exofs_mountopt { |
9ed964843 exofs: Add option... |
52 |
bool is_osdname; |
ba9e5e98c exofs: super_oper... |
53 54 55 56 57 58 59 60 |
const char *dev_name; uint64_t pid; int timeout; }; /* * exofs-specific mount-time options. */ |
9ed964843 exofs: Add option... |
61 |
enum { Opt_name, Opt_pid, Opt_to, Opt_err }; |
ba9e5e98c exofs: super_oper... |
62 63 64 65 66 67 68 |
/* * 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 exofs: Add option... |
69 |
{Opt_name, "osdname=%s"}, |
ba9e5e98c exofs: super_oper... |
70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 |
{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 exofs: Add option... |
101 102 103 104 105 106 107 108 |
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 exofs: super_oper... |
109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 |
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 fs: icache RCU fr... |
166 167 168 |
static void exofs_i_callback(struct rcu_head *head) { struct inode *inode = container_of(head, struct inode, i_rcu); |
fa0d7e3de fs: icache RCU fr... |
169 170 |
kmem_cache_free(exofs_inode_cachep, exofs_i(inode)); } |
ba9e5e98c exofs: super_oper... |
171 172 173 174 175 |
/* * Remove an inode from the cache */ static void exofs_destroy_inode(struct inode *inode) { |
fa0d7e3de fs: icache RCU fr... |
176 |
call_rcu(&inode->i_rcu, exofs_i_callback); |
ba9e5e98c exofs: super_oper... |
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 |
} /* * 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); } /****************************************************************************** |
85e44df47 exofs: Move exofs... |
212 |
* Some osd helpers |
ba9e5e98c exofs: super_oper... |
213 |
*****************************************************************************/ |
85e44df47 exofs: Move exofs... |
214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 |
void exofs_make_credential(u8 cred_a[OSD_CAP_LEN], const struct osd_obj_id *obj) { osd_sec_init_nosec_doall_caps(cred_a, obj, false, true); } static int exofs_read_kern(struct osd_dev *od, u8 *cred, struct osd_obj_id *obj, u64 offset, void *p, unsigned length) { struct osd_request *or = osd_start_request(od, GFP_KERNEL); /* struct osd_sense_info osi = {.key = 0};*/ int ret; if (unlikely(!or)) { EXOFS_DBGMSG("%s: osd_start_request failed. ", __func__); return -ENOMEM; } ret = osd_req_read_kern(or, obj, offset, p, length); if (unlikely(ret)) { EXOFS_DBGMSG("%s: osd_req_read_kern failed. ", __func__); goto out; } ret = osd_finalize_request(or, 0, cred, NULL); if (unlikely(ret)) { EXOFS_DBGMSG("Failed to osd_finalize_request() => %d ", ret); goto out; } ret = osd_execute_request(or); if (unlikely(ret)) EXOFS_DBGMSG("osd_execute_request() => %d ", ret); /* osd_req_decode_sense(or, ret); */ out: osd_end_request(or); EXOFS_DBGMSG2("read_kern(0x%llx) offset=0x%llx " "length=0x%llx dev=%p ret=>%d ", _LLU(obj->id), _LLU(offset), _LLU(length), od, ret); return ret; } |
1cea312ad exofs: Write sbi-... |
259 260 261 262 263 264 265 266 267 268 |
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, }; |
8ff660ab8 exofs: Rename rai... |
269 |
struct ore_io_state *ios; |
1cea312ad exofs: Write sbi-... |
270 |
int ret; |
5bf696dad exofs: Rename str... |
271 |
ret = ore_get_io_state(&sbi->layout, &sbi->oc, &ios); |
1cea312ad exofs: Write sbi-... |
272 |
if (unlikely(ret)) { |
8ff660ab8 exofs: Rename rai... |
273 274 |
EXOFS_ERR("%s: ore_get_io_state failed. ", __func__); |
1cea312ad exofs: Write sbi-... |
275 276 |
return ret; } |
1cea312ad exofs: Write sbi-... |
277 278 |
ios->in_attr = attrs; ios->in_attr_len = ARRAY_SIZE(attrs); |
8ff660ab8 exofs: Rename rai... |
279 |
ret = ore_read(ios); |
1cea312ad exofs: Write sbi-... |
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 |
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: |
8ff660ab8 exofs: Rename rai... |
309 |
ore_put_io_state(ios); |
1cea312ad exofs: Write sbi-... |
310 311 |
return ret; } |
8ff660ab8 exofs: Rename rai... |
312 |
static void stats_done(struct ore_io_state *ios, void *p) |
1cea312ad exofs: Write sbi-... |
313 |
{ |
8ff660ab8 exofs: Rename rai... |
314 |
ore_put_io_state(ios); |
1cea312ad exofs: Write sbi-... |
315 316 317 318 319 320 321 322 323 |
/* 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, }; |
8ff660ab8 exofs: Rename rai... |
324 |
struct ore_io_state *ios; |
1cea312ad exofs: Write sbi-... |
325 |
int ret; |
5bf696dad exofs: Rename str... |
326 |
ret = ore_get_io_state(&sbi->layout, &sbi->oc, &ios); |
1cea312ad exofs: Write sbi-... |
327 |
if (unlikely(ret)) { |
8ff660ab8 exofs: Rename rai... |
328 329 |
EXOFS_ERR("%s: ore_get_io_state failed. ", __func__); |
1cea312ad exofs: Write sbi-... |
330 331 332 333 334 335 |
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; |
9e9db4564 exofs: ios: Move ... |
336 |
|
1cea312ad exofs: Write sbi-... |
337 338 339 340 |
ios->done = stats_done; ios->private = sbi; ios->out_attr = attrs; ios->out_attr_len = ARRAY_SIZE(attrs); |
8ff660ab8 exofs: Rename rai... |
341 |
ret = ore_write(ios); |
1cea312ad exofs: Write sbi-... |
342 |
if (unlikely(ret)) { |
8ff660ab8 exofs: Rename rai... |
343 344 345 |
EXOFS_ERR("%s: ore_write failed. ", __func__); ore_put_io_state(ios); |
1cea312ad exofs: Write sbi-... |
346 347 348 349 |
} return ret; } |
85e44df47 exofs: Move exofs... |
350 351 352 353 354 |
/****************************************************************************** * SUPERBLOCK FUNCTIONS *****************************************************************************/ static const struct super_operations exofs_sops; static const struct export_operations exofs_export_ops; |
ba9e5e98c exofs: super_oper... |
355 356 357 |
/* * Write the superblock to the OSD */ |
de74b05ac exofs/super.c: lo... |
358 |
static int exofs_sync_fs(struct super_block *sb, int wait) |
ba9e5e98c exofs: super_oper... |
359 360 361 |
{ struct exofs_sb_info *sbi; struct exofs_fscb *fscb; |
8ff660ab8 exofs: Rename rai... |
362 |
struct ore_comp one_comp; |
5bf696dad exofs: Rename str... |
363 |
struct ore_components oc; |
8ff660ab8 exofs: Rename rai... |
364 |
struct ore_io_state *ios; |
80e09fb94 exofs: add ->sync_fs |
365 |
int ret = -ENOMEM; |
ba9e5e98c exofs: super_oper... |
366 |
|
1cea312ad exofs: Write sbi-... |
367 368 369 |
fscb = kmalloc(sizeof(*fscb), GFP_KERNEL); if (unlikely(!fscb)) return -ENOMEM; |
ba9e5e98c exofs: super_oper... |
370 |
sbi = sb->s_fs_info; |
06886a5a3 exofs: Move all o... |
371 |
|
1cea312ad exofs: Write sbi-... |
372 373 374 375 376 377 378 |
/* 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. */ |
9e9db4564 exofs: ios: Move ... |
379 |
|
5bf696dad exofs: Rename str... |
380 |
exofs_init_comps(&oc, &one_comp, sbi, EXOFS_SUPER_ID); |
9e9db4564 exofs: ios: Move ... |
381 |
|
5bf696dad exofs: Rename str... |
382 |
ret = ore_get_io_state(&sbi->layout, &oc, &ios); |
1cea312ad exofs: Write sbi-... |
383 |
if (unlikely(ret)) |
06886a5a3 exofs: Move all o... |
384 |
goto out; |
1cea312ad exofs: Write sbi-... |
385 |
lock_super(sb); |
04dc1e88a exofs: Multi-devi... |
386 |
ios->length = offsetof(struct exofs_fscb, s_dev_table_oid); |
06886a5a3 exofs: Move all o... |
387 |
memset(fscb, 0, ios->length); |
ba9e5e98c exofs: super_oper... |
388 389 390 391 |
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 exofs: Multi-devi... |
392 |
fscb->s_version = EXOFS_FSCB_VER; |
ba9e5e98c exofs: super_oper... |
393 |
|
06886a5a3 exofs: Move all o... |
394 395 |
ios->offset = 0; ios->kern_buff = fscb; |
ba9e5e98c exofs: super_oper... |
396 |
|
8ff660ab8 exofs: Rename rai... |
397 |
ret = ore_write(ios); |
1cea312ad exofs: Write sbi-... |
398 |
if (unlikely(ret)) |
8ff660ab8 exofs: Rename rai... |
399 400 |
EXOFS_ERR("%s: ore_write failed. ", __func__); |
1cea312ad exofs: Write sbi-... |
401 402 |
else sb->s_dirt = 0; |
ba9e5e98c exofs: super_oper... |
403 |
|
1cea312ad exofs: Write sbi-... |
404 |
unlock_super(sb); |
ba9e5e98c exofs: super_oper... |
405 |
out: |
06886a5a3 exofs: Move all o... |
406 407 |
EXOFS_DBGMSG("s_nextid=0x%llx ret=%d ", _LLU(sbi->s_nextid), ret); |
8ff660ab8 exofs: Rename rai... |
408 |
ore_put_io_state(ios); |
1cea312ad exofs: Write sbi-... |
409 |
kfree(fscb); |
80e09fb94 exofs: add ->sync_fs |
410 411 412 413 414 415 416 417 418 |
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 exofs: super_oper... |
419 |
} |
19fe294f2 exofs: Prints on ... |
420 421 422 423 424 425 426 427 428 |
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)); } |
de74b05ac exofs/super.c: lo... |
429 |
static void exofs_free_sbi(struct exofs_sb_info *sbi) |
04dc1e88a exofs: Multi-devi... |
430 |
{ |
d866d875f ore/exofs: Change... |
431 432 433 434 435 |
unsigned numdevs = sbi->oc.numdevs; while (numdevs) { unsigned i = --numdevs; struct osd_dev *od = ore_comp_dev(&sbi->oc, i); |
04dc1e88a exofs: Multi-devi... |
436 437 |
if (od) { |
d866d875f ore/exofs: Change... |
438 |
ore_comp_set_dev(&sbi->oc, i, NULL); |
04dc1e88a exofs: Multi-devi... |
439 440 441 |
osduld_put_device(od); } } |
d866d875f ore/exofs: Change... |
442 |
kfree(sbi->oc.ods); |
04dc1e88a exofs: Multi-devi... |
443 444 |
kfree(sbi); } |
ba9e5e98c exofs: super_oper... |
445 446 447 448 449 450 451 452 453 454 455 456 457 |
/* * 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 exofs: deprecate ... |
458 459 460 461 462 |
printk(KERN_NOTICE "%s: !!Pending operations in flight. " "This is a BUG. please report to osd-dev@open-osd.org ", __func__); |
ba9e5e98c exofs: super_oper... |
463 464 465 466 467 |
init_waitqueue_head(&wq); wait_event_timeout(wq, (atomic_read(&sbi->s_curr_pending) == 0), msecs_to_jiffies(100)); } |
d866d875f ore/exofs: Change... |
468 |
_exofs_print_device("Unmounting", NULL, ore_comp_dev(&sbi->oc, 0), |
9e9db4564 exofs: ios: Move ... |
469 |
sbi->one_comp.obj.partition); |
04dc1e88a exofs: Multi-devi... |
470 |
|
b3d0ab7e6 exofs: add bdi ba... |
471 |
bdi_destroy(&sbi->bdi); |
04dc1e88a exofs: Multi-devi... |
472 |
exofs_free_sbi(sbi); |
ba9e5e98c exofs: super_oper... |
473 474 |
sb->s_fs_info = NULL; } |
04dc1e88a exofs: Multi-devi... |
475 476 477 |
static int _read_and_match_data_map(struct exofs_sb_info *sbi, unsigned numdevs, struct exofs_device_table *dt) { |
5a51c0c7e ore/exofs: Define... |
478 |
int ret; |
5d952b839 exofs: RAID0 support |
479 |
|
8d2d83a83 exofs: Remove unu... |
480 |
sbi->layout.stripe_unit = |
04dc1e88a exofs: Multi-devi... |
481 |
le64_to_cpu(dt->dt_data_map.cb_stripe_unit); |
8d2d83a83 exofs: Remove unu... |
482 |
sbi->layout.group_width = |
04dc1e88a exofs: Multi-devi... |
483 |
le32_to_cpu(dt->dt_data_map.cb_group_width); |
8d2d83a83 exofs: Remove unu... |
484 |
sbi->layout.group_depth = |
04dc1e88a exofs: Multi-devi... |
485 |
le32_to_cpu(dt->dt_data_map.cb_group_depth); |
8d2d83a83 exofs: Remove unu... |
486 487 488 |
sbi->layout.mirrors_p1 = le32_to_cpu(dt->dt_data_map.cb_mirror_cnt) + 1; sbi->layout.raid_algorithm = |
04dc1e88a exofs: Multi-devi... |
489 |
le32_to_cpu(dt->dt_data_map.cb_raid_algorithm); |
5a51c0c7e ore/exofs: Define... |
490 |
ret = ore_verify_layout(numdevs, &sbi->layout); |
5d952b839 exofs: RAID0 support |
491 |
|
6d4073e88 exofs: BUG: Avoid... |
492 493 494 495 496 497 498 499 500 |
EXOFS_DBGMSG("exofs: layout: " "num_comps=%u stripe_unit=0x%x group_width=%u " "group_depth=0x%llx mirrors_p1=%u raid_algorithm=%u ", numdevs, sbi->layout.stripe_unit, sbi->layout.group_width, _LLU(sbi->layout.group_depth), sbi->layout.mirrors_p1, |
8d2d83a83 exofs: Remove unu... |
501 |
sbi->layout.raid_algorithm); |
5a51c0c7e ore/exofs: Define... |
502 |
return ret; |
04dc1e88a exofs: Multi-devi... |
503 |
} |
8ff660ab8 exofs: Rename rai... |
504 |
static unsigned __ra_pages(struct ore_layout *layout) |
66cd6cad4 exofs: Override r... |
505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 |
{ 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 exofs: Multi-devi... |
520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 |
/* @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); } |
d866d875f ore/exofs: Change... |
545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 |
int __alloc_dev_table(struct exofs_sb_info *sbi, unsigned numdevs, struct exofs_dev **peds) { struct __alloc_ore_devs_and_exofs_devs { /* Twice bigger table: See exofs_init_comps() and comment at * exofs_read_lookup_dev_table() */ struct ore_dev *oreds[numdevs * 2 - 1]; struct exofs_dev eds[numdevs]; } *aoded; struct exofs_dev *eds; unsigned i; aoded = kzalloc(sizeof(*aoded), GFP_KERNEL); if (unlikely(!aoded)) { EXOFS_ERR("ERROR: faild allocating Device array[%d] ", numdevs); return -ENOMEM; } sbi->oc.ods = aoded->oreds; *peds = eds = aoded->eds; for (i = 0; i < numdevs; ++i) aoded->oreds[i] = &eds[i].ored; return 0; } |
6d4073e88 exofs: BUG: Avoid... |
572 573 |
static int exofs_read_lookup_dev_table(struct exofs_sb_info *sbi, struct osd_dev *fscb_od, |
04dc1e88a exofs: Multi-devi... |
574 575 |
unsigned table_count) { |
8ff660ab8 exofs: Rename rai... |
576 |
struct ore_comp comp; |
04dc1e88a exofs: Multi-devi... |
577 |
struct exofs_device_table *dt; |
d866d875f ore/exofs: Change... |
578 |
struct exofs_dev *eds; |
04dc1e88a exofs: Multi-devi... |
579 580 581 582 583 584 585 586 587 588 589 590 |
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; } |
5bf696dad exofs: Rename str... |
591 |
sbi->oc.numdevs = 0; |
9e9db4564 exofs: ios: Move ... |
592 593 594 595 596 597 598 |
comp.obj.partition = sbi->one_comp.obj.partition; comp.obj.id = EXOFS_DEVTABLE_ID; exofs_make_credential(comp.cred, &comp.obj); ret = exofs_read_kern(fscb_od, comp.cred, &comp.obj, 0, dt, table_bytes); |
04dc1e88a exofs: Multi-devi... |
599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 |
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; |
d866d875f ore/exofs: Change... |
615 616 617 618 619 620 621 622 623 624 |
ret = __alloc_dev_table(sbi, numdevs, &eds); if (unlikely(ret)) goto out; /* exofs round-robins the device table view according to inode * number. We hold a: twice bigger table hence inodes can point * to any device and have a sequential view of the table * starting at this device. See exofs_init_comps() */ memcpy(&sbi->oc.ods[numdevs], &sbi->oc.ods[0], (numdevs - 1) * sizeof(sbi->oc.ods[0])); |
04dc1e88a exofs: Multi-devi... |
625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 |
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); |
d866d875f ore/exofs: Change... |
641 642 |
/* the exofs id is currently the table index */ eds[i].did = i; |
04dc1e88a exofs: Multi-devi... |
643 644 645 646 647 |
/* 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)) { |
d866d875f ore/exofs: Change... |
648 |
eds[i].ored.od = fscb_od; |
5bf696dad exofs: Rename str... |
649 |
++sbi->oc.numdevs; |
04dc1e88a exofs: Multi-devi... |
650 651 652 653 654 |
fscb_od = NULL; continue; } od = osduld_info_lookup(&odi); |
2c722c9a4 exofs: Remove red... |
655 |
if (IS_ERR(od)) { |
04dc1e88a exofs: Multi-devi... |
656 657 658 659 660 661 |
ret = PTR_ERR(od); EXOFS_ERR("ERROR: device requested is not found " "osd_name-%s =>%d ", odi.osdname, ret); goto out; } |
d866d875f ore/exofs: Change... |
662 |
eds[i].ored.od = od; |
5bf696dad exofs: Rename str... |
663 |
++sbi->oc.numdevs; |
04dc1e88a exofs: Multi-devi... |
664 665 666 667 |
/* Read the fscb of the other devices to make sure the FS * partition is there. */ |
9e9db4564 exofs: ios: Move ... |
668 |
ret = exofs_read_kern(od, comp.cred, &comp.obj, 0, &fscb, |
04dc1e88a exofs: Multi-devi... |
669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 |
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); |
d866d875f ore/exofs: Change... |
686 |
if (unlikely(fscb_od && !ret)) { |
9e9db4564 exofs: ios: Move ... |
687 688 689 690 |
EXOFS_ERR("ERROR: Bad device-table container device not present "); osduld_put_device(fscb_od); return -EINVAL; |
9e9db4564 exofs: ios: Move ... |
691 |
} |
04dc1e88a exofs: Multi-devi... |
692 693 |
return ret; } |
ba9e5e98c exofs: super_oper... |
694 695 696 697 698 699 700 701 |
/* * 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 exofs: Move all o... |
702 |
struct osd_dev *od; /* Master device */ |
ba9e5e98c exofs: super_oper... |
703 |
struct exofs_fscb fscb; /*on-disk superblock info */ |
8ff660ab8 exofs: Rename rai... |
704 |
struct ore_comp comp; |
04dc1e88a exofs: Multi-devi... |
705 |
unsigned table_count; |
ba9e5e98c exofs: super_oper... |
706 707 708 709 710 |
int ret; sbi = kzalloc(sizeof(*sbi), GFP_KERNEL); if (!sbi) return -ENOMEM; |
ba9e5e98c exofs: super_oper... |
711 712 |
/* use mount options to fill superblock */ |
9ed964843 exofs: Add option... |
713 714 715 716 717 718 |
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); |
9ce730475 exofs: Small clea... |
719 720 |
kfree(opts->dev_name); opts->dev_name = NULL; |
9ed964843 exofs: Add option... |
721 722 723 |
} else { od = osduld_path_lookup(opts->dev_name); } |
06886a5a3 exofs: Move all o... |
724 |
if (IS_ERR(od)) { |
9ed964843 exofs: Add option... |
725 |
ret = -EINVAL; |
ba9e5e98c exofs: super_oper... |
726 727 |
goto free_sbi; } |
45d3abcb1 exofs: Move layou... |
728 |
/* Default layout in case we do not have a device-table */ |
5d952b839 exofs: RAID0 support |
729 730 731 |
sbi->layout.stripe_unit = PAGE_SIZE; sbi->layout.mirrors_p1 = 1; sbi->layout.group_width = 1; |
50a76fd3c exofs: groups sup... |
732 733 |
sbi->layout.group_depth = -1; sbi->layout.group_count = 1; |
ba9e5e98c exofs: super_oper... |
734 |
sbi->s_timeout = opts->timeout; |
9e9db4564 exofs: ios: Move ... |
735 736 737 |
sbi->one_comp.obj.partition = opts->pid; sbi->one_comp.obj.id = 0; exofs_make_credential(sbi->one_comp.cred, &sbi->one_comp.obj); |
5bf696dad exofs: Rename str... |
738 739 740 |
sbi->oc.numdevs = 1; sbi->oc.single_comp = EC_SINGLE_COMP; sbi->oc.comps = &sbi->one_comp; |
9e9db4564 exofs: ios: Move ... |
741 |
|
ba9e5e98c exofs: super_oper... |
742 743 744 745 746 747 748 749 750 |
/* 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; |
9e9db4564 exofs: ios: Move ... |
751 752 753 |
comp.obj.partition = sbi->one_comp.obj.partition; comp.obj.id = EXOFS_SUPER_ID; exofs_make_credential(comp.cred, &comp.obj); |
ba9e5e98c exofs: super_oper... |
754 |
|
9e9db4564 exofs: ios: Move ... |
755 |
ret = exofs_read_kern(od, comp.cred, &comp.obj, 0, &fscb, sizeof(fscb)); |
06886a5a3 exofs: Move all o... |
756 |
if (unlikely(ret)) |
ba9e5e98c exofs: super_oper... |
757 |
goto free_sbi; |
ba9e5e98c exofs: super_oper... |
758 759 |
sb->s_magic = le16_to_cpu(fscb.s_magic); |
1cea312ad exofs: Write sbi-... |
760 |
/* NOTE: we read below to be backward compatible with old versions */ |
ba9e5e98c exofs: super_oper... |
761 762 763 764 765 766 767 768 769 770 771 |
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 exofs: Write sbi-... |
772 |
if (le32_to_cpu(fscb.s_version) > EXOFS_FSCB_VER) { |
04dc1e88a exofs: Multi-devi... |
773 774 775 776 777 778 |
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 exofs: super_oper... |
779 780 781 782 |
/* 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 exofs: Multi-devi... |
783 784 |
table_count = le64_to_cpu(fscb.s_dev_table_count); if (table_count) { |
6d4073e88 exofs: BUG: Avoid... |
785 |
ret = exofs_read_lookup_dev_table(sbi, od, table_count); |
04dc1e88a exofs: Multi-devi... |
786 787 |
if (unlikely(ret)) goto free_sbi; |
6d4073e88 exofs: BUG: Avoid... |
788 |
} else { |
d866d875f ore/exofs: Change... |
789 790 791 792 793 794 795 |
struct exofs_dev *eds; ret = __alloc_dev_table(sbi, 1, &eds); if (unlikely(ret)) goto free_sbi; ore_comp_set_dev(&sbi->oc, 0, od); |
04dc1e88a exofs: Multi-devi... |
796 |
} |
1cea312ad exofs: Write sbi-... |
797 |
__sbi_read_stats(sbi); |
ba9e5e98c exofs: super_oper... |
798 |
/* set up operation vectors */ |
66cd6cad4 exofs: Override r... |
799 |
sbi->bdi.ra_pages = __ra_pages(&sbi->layout); |
b3d0ab7e6 exofs: add bdi ba... |
800 |
sb->s_bdi = &sbi->bdi; |
06886a5a3 exofs: Move all o... |
801 |
sb->s_fs_info = sbi; |
ba9e5e98c exofs: super_oper... |
802 |
sb->s_op = &exofs_sops; |
8cf74b393 exofs: export_ope... |
803 |
sb->s_export_op = &exofs_export_ops; |
ba9e5e98c exofs: super_oper... |
804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 |
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; } |
6d4073e88 exofs: BUG: Avoid... |
829 830 831 832 |
ret = bdi_setup_and_register(&sbi->bdi, "exofs", BDI_CAP_MAP_COPY); if (ret) { EXOFS_DBGMSG("Failed to bdi_setup_and_register "); |
da01636a6 exofs: oops after... |
833 834 |
dput(sb->s_root); sb->s_root = NULL; |
6d4073e88 exofs: BUG: Avoid... |
835 836 |
goto free_sbi; } |
d866d875f ore/exofs: Change... |
837 838 |
_exofs_print_device("Mounting", opts->dev_name, ore_comp_dev(&sbi->oc, 0), |
9e9db4564 exofs: ios: Move ... |
839 |
sbi->one_comp.obj.partition); |
06886a5a3 exofs: Move all o... |
840 |
return 0; |
ba9e5e98c exofs: super_oper... |
841 842 |
free_sbi: |
06886a5a3 exofs: Move all o... |
843 844 |
EXOFS_ERR("Unable to mount exofs on %s pid=0x%llx err=%d ", |
9e9db4564 exofs: ios: Move ... |
845 |
opts->dev_name, sbi->one_comp.obj.partition, ret); |
04dc1e88a exofs: Multi-devi... |
846 |
exofs_free_sbi(sbi); |
06886a5a3 exofs: Move all o... |
847 |
return ret; |
ba9e5e98c exofs: super_oper... |
848 849 850 851 852 |
} /* * Set up the superblock (calls exofs_fill_super eventually) */ |
3c26ff6e4 convert get_sb_no... |
853 |
static struct dentry *exofs_mount(struct file_system_type *type, |
ba9e5e98c exofs: super_oper... |
854 |
int flags, const char *dev_name, |
3c26ff6e4 convert get_sb_no... |
855 |
void *data) |
ba9e5e98c exofs: super_oper... |
856 857 858 859 860 861 |
{ struct exofs_mountopt opts; int ret; ret = parse_options(data, &opts); if (ret) |
3c26ff6e4 convert get_sb_no... |
862 |
return ERR_PTR(ret); |
ba9e5e98c exofs: super_oper... |
863 |
|
9ed964843 exofs: Add option... |
864 865 |
if (!opts.dev_name) opts.dev_name = dev_name; |
3c26ff6e4 convert get_sb_no... |
866 |
return mount_nodev(type, flags, &opts, exofs_fill_super); |
ba9e5e98c exofs: super_oper... |
867 868 869 870 871 872 873 874 875 876 |
} /* * 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; |
8ff660ab8 exofs: Rename rai... |
877 |
struct ore_io_state *ios; |
ba9e5e98c exofs: super_oper... |
878 879 880 881 882 883 884 885 |
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 exofs: super_oper... |
886 |
int ret; |
5bf696dad exofs: Rename str... |
887 |
ret = ore_get_io_state(&sbi->layout, &sbi->oc, &ios); |
06886a5a3 exofs: Move all o... |
888 |
if (ret) { |
8ff660ab8 exofs: Rename rai... |
889 890 |
EXOFS_DBGMSG("ore_get_io_state failed. "); |
06886a5a3 exofs: Move all o... |
891 |
return ret; |
ba9e5e98c exofs: super_oper... |
892 |
} |
06886a5a3 exofs: Move all o... |
893 894 |
ios->in_attr = attrs; ios->in_attr_len = ARRAY_SIZE(attrs); |
8ff660ab8 exofs: Rename rai... |
895 |
ret = ore_read(ios); |
ba9e5e98c exofs: super_oper... |
896 897 |
if (unlikely(ret)) goto out; |
06886a5a3 exofs: Move all o... |
898 |
ret = extract_attr_from_ios(ios, &attrs[0]); |
cae012d85 exofs: statfs blo... |
899 |
if (likely(!ret)) { |
ba9e5e98c exofs: super_oper... |
900 |
capacity = get_unaligned_be64(attrs[0].val_ptr); |
cae012d85 exofs: statfs blo... |
901 902 903 |
if (unlikely(!capacity)) capacity = ULLONG_MAX; } else |
ba9e5e98c exofs: super_oper... |
904 905 |
EXOFS_DBGMSG("exofs_statfs: get capacity failed. "); |
06886a5a3 exofs: Move all o... |
906 |
ret = extract_attr_from_ios(ios, &attrs[1]); |
ba9e5e98c exofs: super_oper... |
907 908 909 910 911 912 913 914 915 |
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 exofs: statfs blo... |
916 917 |
buf->f_blocks = capacity >> 9; buf->f_bfree = (capacity - used) >> 9; |
ba9e5e98c exofs: super_oper... |
918 919 920 921 922 923 |
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: |
8ff660ab8 exofs: Rename rai... |
924 |
ore_put_io_state(ios); |
ba9e5e98c exofs: super_oper... |
925 926 927 928 929 930 931 |
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 convert exofs to ... |
932 |
.evict_inode = exofs_evict_inode, |
ba9e5e98c exofs: super_oper... |
933 934 |
.put_super = exofs_put_super, .write_super = exofs_write_super, |
80e09fb94 exofs: add ->sync_fs |
935 |
.sync_fs = exofs_sync_fs, |
ba9e5e98c exofs: super_oper... |
936 937 938 939 |
.statfs = exofs_statfs, }; /****************************************************************************** |
8cf74b393 exofs: export_ope... |
940 941 |
* EXPORT OPERATIONS *****************************************************************************/ |
de74b05ac exofs/super.c: lo... |
942 |
static struct dentry *exofs_get_parent(struct dentry *child) |
8cf74b393 exofs: export_ope... |
943 944 945 946 |
{ unsigned long ino = exofs_parent_ino(child); if (!ino) |
a803b8067 fix exofs ->get_p... |
947 |
return ERR_PTR(-ESTALE); |
8cf74b393 exofs: export_ope... |
948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 |
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 exofs: super_oper... |
989 990 991 992 993 994 995 996 997 |
* INSMOD/RMMOD *****************************************************************************/ /* * struct that describes this file system */ static struct file_system_type exofs_type = { .owner = THIS_MODULE, .name = "exofs", |
3c26ff6e4 convert get_sb_no... |
998 |
.mount = exofs_mount, |
ba9e5e98c exofs: super_oper... |
999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 |
.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) |