inode.c 37.4 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
/*
 *  linux/fs/nfs/inode.c
 *
 *  Copyright (C) 1992  Rick Sladkey
 *
 *  nfs inode and superblock handling functions
 *
 *  Modularised by Alan Cox <Alan.Cox@linux.org>, while hacking some
 *  experimental NFS changes. Modularisation taken straight from SYS5 fs.
 *
 *  Change to nfs_read_super() to permit NFS mounts to multi-homed hosts.
 *  J.S.Peatfield@damtp.cam.ac.uk
 *
 */

#include <linux/module.h>
#include <linux/init.h>
A
Alexey Dobriyan 已提交
18
#include <linux/sched.h>
L
Linus Torvalds 已提交
19 20 21 22 23 24 25 26 27
#include <linux/time.h>
#include <linux/kernel.h>
#include <linux/mm.h>
#include <linux/string.h>
#include <linux/stat.h>
#include <linux/errno.h>
#include <linux/unistd.h>
#include <linux/sunrpc/clnt.h>
#include <linux/sunrpc/stats.h>
28
#include <linux/sunrpc/metrics.h>
L
Linus Torvalds 已提交
29 30 31 32 33 34 35 36 37
#include <linux/nfs_fs.h>
#include <linux/nfs_mount.h>
#include <linux/nfs4_mount.h>
#include <linux/lockd/bind.h>
#include <linux/smp_lock.h>
#include <linux/seq_file.h>
#include <linux/mount.h>
#include <linux/nfs_idmap.h>
#include <linux/vfs.h>
38 39
#include <linux/inet.h>
#include <linux/nfs_xdr.h>
L
Linus Torvalds 已提交
40 41 42 43

#include <asm/system.h>
#include <asm/uaccess.h>

44
#include "nfs4_fs.h"
45
#include "callback.h"
L
Linus Torvalds 已提交
46
#include "delegation.h"
47
#include "iostat.h"
D
David Howells 已提交
48
#include "internal.h"
L
Linus Torvalds 已提交
49 50 51

#define NFSDBG_FACILITY		NFSDBG_VFS

52 53 54 55 56
#define NFS_64_BIT_INODE_NUMBERS_ENABLED	1

/* Default is to see 64-bit inode numbers */
static int enable_ino64 = NFS_64_BIT_INODE_NUMBERS_ENABLED;

L
Linus Torvalds 已提交
57
static void nfs_invalidate_inode(struct inode *);
58
static int nfs_update_inode(struct inode *, struct nfs_fattr *);
L
Linus Torvalds 已提交
59

60
static struct kmem_cache * nfs_inode_cachep;
61

L
Linus Torvalds 已提交
62 63 64 65 66 67
static inline unsigned long
nfs_fattr_to_ino_t(struct nfs_fattr *fattr)
{
	return nfs_fileid_to_ino_t(fattr->fileid);
}

68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86
/**
 * nfs_compat_user_ino64 - returns the user-visible inode number
 * @fileid: 64-bit fileid
 *
 * This function returns a 32-bit inode number if the boot parameter
 * nfs.enable_ino64 is zero.
 */
u64 nfs_compat_user_ino64(u64 fileid)
{
	int ino;

	if (enable_ino64)
		return fileid;
	ino = fileid;
	if (sizeof(ino) < sizeof(fileid))
		ino ^= fileid >> (sizeof(fileid)-sizeof(ino)) * 8;
	return ino;
}

D
David Howells 已提交
87
int nfs_write_inode(struct inode *inode, int sync)
L
Linus Torvalds 已提交
88 89 90
{
	int ret;

T
Trond Myklebust 已提交
91 92 93 94 95 96 97 98 99 100
	if (sync) {
		ret = filemap_fdatawait(inode->i_mapping);
		if (ret == 0)
			ret = nfs_commit_inode(inode, FLUSH_SYNC);
	} else
		ret = nfs_commit_inode(inode, 0);
	if (ret >= 0)
		return 0;
	__mark_inode_dirty(inode, I_DIRTY_DATASYNC);
	return ret;
L
Linus Torvalds 已提交
101 102
}

D
David Howells 已提交
103
void nfs_clear_inode(struct inode *inode)
L
Linus Torvalds 已提交
104
{
T
Trond Myklebust 已提交
105 106 107 108
	/*
	 * The following should never happen...
	 */
	BUG_ON(nfs_have_writebacks(inode));
109
	BUG_ON(!list_empty(&NFS_I(inode)->open_files));
110
	nfs_zap_acl_cache(inode);
111
	nfs_access_zap_cache(inode);
L
Linus Torvalds 已提交
112 113
}

T
Trond Myklebust 已提交
114 115 116 117 118 119 120 121 122 123
/**
 * nfs_sync_mapping - helper to flush all mmapped dirty data to disk
 */
int nfs_sync_mapping(struct address_space *mapping)
{
	int ret;

	if (mapping->nrpages == 0)
		return 0;
	unmap_mapping_range(mapping, 0, 0, 0);
124
	ret = filemap_write_and_wait(mapping);
T
Trond Myklebust 已提交
125 126 127 128 129 130 131
	if (ret != 0)
		goto out;
	ret = nfs_wb_all(mapping->host);
out:
	return ret;
}

L
Linus Torvalds 已提交
132 133 134
/*
 * Invalidate the local caches
 */
135
static void nfs_zap_caches_locked(struct inode *inode)
L
Linus Torvalds 已提交
136 137 138 139
{
	struct nfs_inode *nfsi = NFS_I(inode);
	int mode = inode->i_mode;

C
Chuck Lever 已提交
140 141
	nfs_inc_stats(inode, NFSIOS_ATTRINVALIDATE);

142 143
	nfsi->attrtimeo = NFS_MINATTRTIMEO(inode);
	nfsi->attrtimeo_timestamp = jiffies;
L
Linus Torvalds 已提交
144 145 146

	memset(NFS_COOKIEVERF(inode), 0, sizeof(NFS_COOKIEVERF(inode)));
	if (S_ISREG(mode) || S_ISDIR(mode) || S_ISLNK(mode))
147
		nfsi->cache_validity |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_DATA|NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL|NFS_INO_REVAL_PAGECACHE;
L
Linus Torvalds 已提交
148
	else
149
		nfsi->cache_validity |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL|NFS_INO_REVAL_PAGECACHE;
150
}
151

152 153 154 155
void nfs_zap_caches(struct inode *inode)
{
	spin_lock(&inode->i_lock);
	nfs_zap_caches_locked(inode);
156
	spin_unlock(&inode->i_lock);
157 158
}

159 160 161 162 163 164 165 166 167
void nfs_zap_mapping(struct inode *inode, struct address_space *mapping)
{
	if (mapping->nrpages != 0) {
		spin_lock(&inode->i_lock);
		NFS_I(inode)->cache_validity |= NFS_INO_INVALID_DATA;
		spin_unlock(&inode->i_lock);
	}
}

168
void nfs_zap_acl_cache(struct inode *inode)
169 170 171 172 173 174
{
	void (*clear_acl_cache)(struct inode *);

	clear_acl_cache = NFS_PROTO(inode)->clear_acl_cache;
	if (clear_acl_cache != NULL)
		clear_acl_cache(inode);
175
	spin_lock(&inode->i_lock);
176
	NFS_I(inode)->cache_validity &= ~NFS_INO_INVALID_ACL;
177
	spin_unlock(&inode->i_lock);
L
Linus Torvalds 已提交
178 179
}

180 181 182 183 184 185 186
void nfs_invalidate_atime(struct inode *inode)
{
	spin_lock(&inode->i_lock);
	NFS_I(inode)->cache_validity |= NFS_INO_INVALID_ATIME;
	spin_unlock(&inode->i_lock);
}

L
Linus Torvalds 已提交
187
/*
188 189
 * Invalidate, but do not unhash, the inode.
 * NB: must be called with inode->i_lock held!
L
Linus Torvalds 已提交
190
 */
191
static void nfs_invalidate_inode(struct inode *inode)
L
Linus Torvalds 已提交
192
{
B
Benny Halevy 已提交
193
	set_bit(NFS_INO_STALE, &NFS_I(inode)->flags);
194
	nfs_zap_caches_locked(inode);
L
Linus Torvalds 已提交
195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229
}

struct nfs_find_desc {
	struct nfs_fh		*fh;
	struct nfs_fattr	*fattr;
};

/*
 * In NFSv3 we can have 64bit inode numbers. In order to support
 * this, and re-exported directories (also seen in NFSv2)
 * we are forced to allow 2 different inodes to have the same
 * i_ino.
 */
static int
nfs_find_actor(struct inode *inode, void *opaque)
{
	struct nfs_find_desc	*desc = (struct nfs_find_desc *)opaque;
	struct nfs_fh		*fh = desc->fh;
	struct nfs_fattr	*fattr = desc->fattr;

	if (NFS_FILEID(inode) != fattr->fileid)
		return 0;
	if (nfs_compare_fh(NFS_FH(inode), fh))
		return 0;
	if (is_bad_inode(inode) || NFS_STALE(inode))
		return 0;
	return 1;
}

static int
nfs_init_locked(struct inode *inode, void *opaque)
{
	struct nfs_find_desc	*desc = (struct nfs_find_desc *)opaque;
	struct nfs_fattr	*fattr = desc->fattr;

230
	set_nfs_fileid(inode, fattr->fileid);
L
Linus Torvalds 已提交
231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248
	nfs_copy_fh(NFS_FH(inode), desc->fh);
	return 0;
}

/* Don't use READDIRPLUS on directories that we believe are too large */
#define NFS_LIMIT_READDIRPLUS (8*PAGE_SIZE)

/*
 * This is our front-end to iget that looks up inodes by file handle
 * instead of inode number.
 */
struct inode *
nfs_fhget(struct super_block *sb, struct nfs_fh *fh, struct nfs_fattr *fattr)
{
	struct nfs_find_desc desc = {
		.fh	= fh,
		.fattr	= fattr
	};
249
	struct inode *inode = ERR_PTR(-ENOENT);
L
Linus Torvalds 已提交
250 251 252 253 254 255 256 257 258 259 260 261
	unsigned long hash;

	if ((fattr->valid & NFS_ATTR_FATTR) == 0)
		goto out_no_inode;

	if (!fattr->nlink) {
		printk("NFS: Buggy server - nlink == 0!\n");
		goto out_no_inode;
	}

	hash = nfs_fattr_to_ino_t(fattr);

262 263 264
	inode = iget5_locked(sb, hash, nfs_find_actor, nfs_init_locked, &desc);
	if (inode == NULL) {
		inode = ERR_PTR(-ENOMEM);
L
Linus Torvalds 已提交
265
		goto out_no_inode;
266
	}
L
Linus Torvalds 已提交
267 268 269

	if (inode->i_state & I_NEW) {
		struct nfs_inode *nfsi = NFS_I(inode);
270
		unsigned long now = jiffies;
L
Linus Torvalds 已提交
271 272 273 274 275 276 277 278 279 280 281

		/* We set i_ino for the few things that still rely on it,
		 * such as stat(2) */
		inode->i_ino = hash;

		/* We can't support update_atime(), since the server will reset it */
		inode->i_flags |= S_NOATIME|S_NOCMTIME;
		inode->i_mode = fattr->mode;
		/* Why so? Because we want revalidate for devices/FIFOs, and
		 * that's precisely what we have in nfs_file_inode_operations.
		 */
282
		inode->i_op = NFS_SB(sb)->nfs_client->rpc_ops->file_inode_ops;
L
Linus Torvalds 已提交
283 284 285 286 287
		if (S_ISREG(inode->i_mode)) {
			inode->i_fop = &nfs_file_operations;
			inode->i_data.a_ops = &nfs_file_aops;
			inode->i_data.backing_dev_info = &NFS_SB(sb)->backing_dev_info;
		} else if (S_ISDIR(inode->i_mode)) {
288
			inode->i_op = NFS_SB(sb)->nfs_client->rpc_ops->dir_inode_ops;
L
Linus Torvalds 已提交
289 290 291
			inode->i_fop = &nfs_dir_operations;
			if (nfs_server_capable(inode, NFS_CAP_READDIRPLUS)
			    && fattr->size <= NFS_LIMIT_READDIRPLUS)
B
Benny Halevy 已提交
292
				set_bit(NFS_INO_ADVISE_RDPLUS, &NFS_I(inode)->flags);
293 294
			/* Deal with crossing mountpoints */
			if (!nfs_fsid_equal(&NFS_SB(sb)->fsid, &fattr->fsid)) {
M
Manoj Naik 已提交
295 296 297 298
				if (fattr->valid & NFS_ATTR_FATTR_V4_REFERRAL)
					inode->i_op = &nfs_referral_inode_operations;
				else
					inode->i_op = &nfs_mountpoint_inode_operations;
299
				inode->i_fop = NULL;
300
				set_bit(NFS_INO_MOUNTPOINT, &nfsi->flags);
301
			}
L
Linus Torvalds 已提交
302 303 304 305 306
		} else if (S_ISLNK(inode->i_mode))
			inode->i_op = &nfs_symlink_inode_operations;
		else
			init_special_inode(inode, inode->i_mode, fattr->rdev);

307
		nfsi->read_cache_jiffies = fattr->time_start;
308 309
		nfsi->last_updated = now;
		nfsi->cache_change_attribute = now;
L
Linus Torvalds 已提交
310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327
		inode->i_atime = fattr->atime;
		inode->i_mtime = fattr->mtime;
		inode->i_ctime = fattr->ctime;
		if (fattr->valid & NFS_ATTR_FATTR_V4)
			nfsi->change_attr = fattr->change_attr;
		inode->i_size = nfs_size_to_loff_t(fattr->size);
		inode->i_nlink = fattr->nlink;
		inode->i_uid = fattr->uid;
		inode->i_gid = fattr->gid;
		if (fattr->valid & (NFS_ATTR_FATTR_V3 | NFS_ATTR_FATTR_V4)) {
			/*
			 * report the blocks in 512byte units
			 */
			inode->i_blocks = nfs_calc_block_size(fattr->du.nfs3.used);
		} else {
			inode->i_blocks = fattr->du.nfs2.blocks;
		}
		nfsi->attrtimeo = NFS_MINATTRTIMEO(inode);
328
		nfsi->attrtimeo_timestamp = now;
L
Linus Torvalds 已提交
329
		memset(nfsi->cookieverf, 0, sizeof(nfsi->cookieverf));
330
		nfsi->access_cache = RB_ROOT;
L
Linus Torvalds 已提交
331 332 333 334 335 336 337 338 339 340 341 342 343

		unlock_new_inode(inode);
	} else
		nfs_refresh_inode(inode, fattr);
	dprintk("NFS: nfs_fhget(%s/%Ld ct=%d)\n",
		inode->i_sb->s_id,
		(long long)NFS_FILEID(inode),
		atomic_read(&inode->i_count));

out:
	return inode;

out_no_inode:
344
	dprintk("nfs_fhget: iget failed with error %ld\n", PTR_ERR(inode));
L
Linus Torvalds 已提交
345 346 347
	goto out;
}

348
#define NFS_VALID_ATTRS (ATTR_MODE|ATTR_UID|ATTR_GID|ATTR_SIZE|ATTR_ATIME|ATTR_ATIME_SET|ATTR_MTIME|ATTR_MTIME_SET|ATTR_FILE)
L
Linus Torvalds 已提交
349 350 351 352 353 354 355 356

int
nfs_setattr(struct dentry *dentry, struct iattr *attr)
{
	struct inode *inode = dentry->d_inode;
	struct nfs_fattr fattr;
	int error;

C
Chuck Lever 已提交
357 358
	nfs_inc_stats(inode, NFSIOS_VFSSETATTR);

359 360 361 362
	/* skip mode change if it's just for clearing setuid/setgid */
	if (attr->ia_valid & (ATTR_KILL_SUID | ATTR_KILL_SGID))
		attr->ia_valid &= ~ATTR_MODE;

L
Linus Torvalds 已提交
363 364 365 366 367 368 369
	if (attr->ia_valid & ATTR_SIZE) {
		if (!S_ISREG(inode->i_mode) || attr->ia_size == i_size_read(inode))
			attr->ia_valid &= ~ATTR_SIZE;
	}

	/* Optimization: if the end result is no change, don't RPC */
	attr->ia_valid &= NFS_VALID_ATTRS;
370
	if ((attr->ia_valid & ~ATTR_FILE) == 0)
L
Linus Torvalds 已提交
371 372
		return 0;

373
	/* Write all dirty data */
374 375 376 377
	if (S_ISREG(inode->i_mode)) {
		filemap_write_and_wait(inode->i_mapping);
		nfs_wb_all(inode);
	}
378 379 380 381 382
	/*
	 * Return any delegations if we're going to change ACLs
	 */
	if ((attr->ia_valid & (ATTR_MODE|ATTR_UID|ATTR_GID)) != 0)
		nfs_inode_return_delegation(inode);
L
Linus Torvalds 已提交
383
	error = NFS_PROTO(inode)->setattr(dentry, &fattr, attr);
384
	if (error == 0)
L
Linus Torvalds 已提交
385
		nfs_refresh_inode(inode, &fattr);
386 387 388
	return error;
}

389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444
/**
 * nfs_vmtruncate - unmap mappings "freed" by truncate() syscall
 * @inode: inode of the file used
 * @offset: file offset to start truncating
 *
 * This is a copy of the common vmtruncate, but with the locking
 * corrected to take into account the fact that NFS requires
 * inode->i_size to be updated under the inode->i_lock.
 */
static int nfs_vmtruncate(struct inode * inode, loff_t offset)
{
	if (i_size_read(inode) < offset) {
		unsigned long limit;

		limit = current->signal->rlim[RLIMIT_FSIZE].rlim_cur;
		if (limit != RLIM_INFINITY && offset > limit)
			goto out_sig;
		if (offset > inode->i_sb->s_maxbytes)
			goto out_big;
		spin_lock(&inode->i_lock);
		i_size_write(inode, offset);
		spin_unlock(&inode->i_lock);
	} else {
		struct address_space *mapping = inode->i_mapping;

		/*
		 * truncation of in-use swapfiles is disallowed - it would
		 * cause subsequent swapout to scribble on the now-freed
		 * blocks.
		 */
		if (IS_SWAPFILE(inode))
			return -ETXTBSY;
		spin_lock(&inode->i_lock);
		i_size_write(inode, offset);
		spin_unlock(&inode->i_lock);

		/*
		 * unmap_mapping_range is called twice, first simply for
		 * efficiency so that truncate_inode_pages does fewer
		 * single-page unmaps.  However after this first call, and
		 * before truncate_inode_pages finishes, it is possible for
		 * private pages to be COWed, which remain after
		 * truncate_inode_pages finishes, hence the second
		 * unmap_mapping_range call must be made for correctness.
		 */
		unmap_mapping_range(mapping, offset + PAGE_SIZE - 1, 0, 1);
		truncate_inode_pages(mapping, offset);
		unmap_mapping_range(mapping, offset + PAGE_SIZE - 1, 0, 1);
	}
	return 0;
out_sig:
	send_sig(SIGXFSZ, current, 0);
out_big:
	return -EFBIG;
}

445 446 447 448 449 450 451 452 453 454 455
/**
 * nfs_setattr_update_inode - Update inode metadata after a setattr call.
 * @inode: pointer to struct inode
 * @attr: pointer to struct iattr
 *
 * Note: we do this in the *proc.c in order to ensure that
 *       it works for things like exclusive creates too.
 */
void nfs_setattr_update_inode(struct inode *inode, struct iattr *attr)
{
	if ((attr->ia_valid & (ATTR_MODE|ATTR_UID|ATTR_GID)) != 0) {
L
Linus Torvalds 已提交
456
		if ((attr->ia_valid & ATTR_MODE) != 0) {
457 458
			int mode = attr->ia_mode & S_IALLUGO;
			mode |= inode->i_mode & ~S_IALLUGO;
L
Linus Torvalds 已提交
459 460 461 462 463 464
			inode->i_mode = mode;
		}
		if ((attr->ia_valid & ATTR_UID) != 0)
			inode->i_uid = attr->ia_uid;
		if ((attr->ia_valid & ATTR_GID) != 0)
			inode->i_gid = attr->ia_gid;
465
		spin_lock(&inode->i_lock);
466
		NFS_I(inode)->cache_validity |= NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
467
		spin_unlock(&inode->i_lock);
468 469
	}
	if ((attr->ia_valid & ATTR_SIZE) != 0) {
C
Chuck Lever 已提交
470
		nfs_inc_stats(inode, NFSIOS_SETATTRTRUNC);
471
		nfs_vmtruncate(inode, attr->ia_size);
472
	}
L
Linus Torvalds 已提交
473 474
}

475 476 477 478 479 480 481 482
static int nfs_wait_schedule(void *word)
{
	if (signal_pending(current))
		return -ERESTARTSYS;
	schedule();
	return 0;
}

L
Linus Torvalds 已提交
483 484 485
/*
 * Wait for the inode to get unlocked.
 */
486
static int nfs_wait_on_inode(struct inode *inode)
L
Linus Torvalds 已提交
487 488 489
{
	struct nfs_inode *nfsi = NFS_I(inode);
	int error;
490 491

	error = wait_on_bit_lock(&nfsi->flags, NFS_INO_REVALIDATING,
492
					nfs_wait_schedule, TASK_KILLABLE);
493

L
Linus Torvalds 已提交
494 495 496
	return error;
}

497 498 499 500 501 502 503 504 505
static void nfs_wake_up_inode(struct inode *inode)
{
	struct nfs_inode *nfsi = NFS_I(inode);

	clear_bit(NFS_INO_REVALIDATING, &nfsi->flags);
	smp_mb__after_clear_bit();
	wake_up_bit(&nfsi->flags, NFS_INO_REVALIDATING);
}

L
Linus Torvalds 已提交
506 507 508
int nfs_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat)
{
	struct inode *inode = dentry->d_inode;
509
	int need_atime = NFS_I(inode)->cache_validity & NFS_INO_INVALID_ATIME;
L
Linus Torvalds 已提交
510 511
	int err;

512 513 514 515 516 517 518 519 520
	/*
	 * Flush out writes to the server in order to update c/mtime.
	 *
	 * Hold the i_mutex to suspend application writes temporarily;
	 * this prevents long-running writing applications from blocking
	 * nfs_wb_nocommit.
	 */
	if (S_ISREG(inode->i_mode)) {
		mutex_lock(&inode->i_mutex);
521
		nfs_wb_nocommit(inode);
522 523
		mutex_unlock(&inode->i_mutex);
	}
524 525 526 527 528 529 530 531 532 533 534 535

	/*
	 * We may force a getattr if the user cares about atime.
	 *
	 * Note that we only have to check the vfsmount flags here:
	 *  - NFS always sets S_NOATIME by so checking it would give a
	 *    bogus result
	 *  - NFS never sets MS_NOATIME or MS_NODIRATIME so there is
	 *    no point in checking those.
	 */
 	if ((mnt->mnt_flags & MNT_NOATIME) ||
 	    ((mnt->mnt_flags & MNT_NODIRATIME) && S_ISDIR(inode->i_mode)))
L
Linus Torvalds 已提交
536
		need_atime = 0;
537

L
Linus Torvalds 已提交
538 539 540 541
	if (need_atime)
		err = __nfs_revalidate_inode(NFS_SERVER(inode), inode);
	else
		err = nfs_revalidate_inode(NFS_SERVER(inode), inode);
P
Peter Staubach 已提交
542
	if (!err) {
L
Linus Torvalds 已提交
543
		generic_fillattr(inode, stat);
544
		stat->ino = nfs_compat_user_ino64(NFS_FILEID(inode));
P
Peter Staubach 已提交
545
	}
L
Linus Torvalds 已提交
546 547 548
	return err;
}

549
static struct nfs_open_context *alloc_nfs_open_context(struct vfsmount *mnt, struct dentry *dentry, struct rpc_cred *cred)
L
Linus Torvalds 已提交
550 551 552
{
	struct nfs_open_context *ctx;

553
	ctx = kmalloc(sizeof(*ctx), GFP_KERNEL);
L
Linus Torvalds 已提交
554
	if (ctx != NULL) {
555 556
		ctx->path.dentry = dget(dentry);
		ctx->path.mnt = mntget(mnt);
L
Linus Torvalds 已提交
557 558 559
		ctx->cred = get_rpccred(cred);
		ctx->state = NULL;
		ctx->lockowner = current->files;
560
		ctx->flags = 0;
L
Linus Torvalds 已提交
561
		ctx->error = 0;
562
		ctx->dir_cookie = 0;
T
Trond Myklebust 已提交
563
		atomic_set(&ctx->count, 1);
L
Linus Torvalds 已提交
564 565 566 567 568 569 570
	}
	return ctx;
}

struct nfs_open_context *get_nfs_open_context(struct nfs_open_context *ctx)
{
	if (ctx != NULL)
T
Trond Myklebust 已提交
571
		atomic_inc(&ctx->count);
L
Linus Torvalds 已提交
572 573 574
	return ctx;
}

575
static void __put_nfs_open_context(struct nfs_open_context *ctx, int wait)
L
Linus Torvalds 已提交
576
{
577
	struct inode *inode;
578

579 580 581 582
	if (ctx == NULL)
		return;

	inode = ctx->path.dentry->d_inode;
T
Trond Myklebust 已提交
583 584 585 586
	if (!atomic_dec_and_lock(&ctx->count, &inode->i_lock))
		return;
	list_del(&ctx->list);
	spin_unlock(&inode->i_lock);
587 588 589 590 591 592
	if (ctx->state != NULL) {
		if (wait)
			nfs4_close_sync(&ctx->path, ctx->state, ctx->mode);
		else
			nfs4_close_state(&ctx->path, ctx->state, ctx->mode);
	}
593 594
	if (ctx->cred != NULL)
		put_rpccred(ctx->cred);
J
Jan Blunck 已提交
595
	path_put(&ctx->path);
596 597 598
	kfree(ctx);
}

599 600 601 602 603 604 605 606 607 608
void put_nfs_open_context(struct nfs_open_context *ctx)
{
	__put_nfs_open_context(ctx, 0);
}

static void put_nfs_open_context_sync(struct nfs_open_context *ctx)
{
	__put_nfs_open_context(ctx, 1);
}

L
Linus Torvalds 已提交
609 610 611 612
/*
 * Ensure that mmap has a recent RPC credential for use when writing out
 * shared pages
 */
613
static void nfs_file_set_open_context(struct file *filp, struct nfs_open_context *ctx)
L
Linus Torvalds 已提交
614
{
615
	struct inode *inode = filp->f_path.dentry->d_inode;
L
Linus Torvalds 已提交
616 617 618 619 620 621 622 623
	struct nfs_inode *nfsi = NFS_I(inode);

	filp->private_data = get_nfs_open_context(ctx);
	spin_lock(&inode->i_lock);
	list_add(&ctx->list, &nfsi->open_files);
	spin_unlock(&inode->i_lock);
}

624 625 626 627
/*
 * Given an inode, search for an open context with the desired characteristics
 */
struct nfs_open_context *nfs_find_open_context(struct inode *inode, struct rpc_cred *cred, int mode)
L
Linus Torvalds 已提交
628 629 630 631 632 633
{
	struct nfs_inode *nfsi = NFS_I(inode);
	struct nfs_open_context *pos, *ctx = NULL;

	spin_lock(&inode->i_lock);
	list_for_each_entry(pos, &nfsi->open_files, list) {
634 635
		if (cred != NULL && pos->cred != cred)
			continue;
L
Linus Torvalds 已提交
636 637 638 639 640 641 642 643 644
		if ((pos->mode & mode) == mode) {
			ctx = get_nfs_open_context(pos);
			break;
		}
	}
	spin_unlock(&inode->i_lock);
	return ctx;
}

645
static void nfs_file_clear_open_context(struct file *filp)
L
Linus Torvalds 已提交
646
{
647
	struct inode *inode = filp->f_path.dentry->d_inode;
648
	struct nfs_open_context *ctx = nfs_file_open_context(filp);
L
Linus Torvalds 已提交
649 650 651 652 653 654

	if (ctx) {
		filp->private_data = NULL;
		spin_lock(&inode->i_lock);
		list_move_tail(&ctx->list, &NFS_I(inode)->open_files);
		spin_unlock(&inode->i_lock);
655
		put_nfs_open_context_sync(ctx);
L
Linus Torvalds 已提交
656 657 658 659 660 661 662 663 664 665 666
	}
}

/*
 * These allocate and release file read/write context information.
 */
int nfs_open(struct inode *inode, struct file *filp)
{
	struct nfs_open_context *ctx;
	struct rpc_cred *cred;

667
	cred = rpc_lookup_cred();
L
Linus Torvalds 已提交
668 669
	if (IS_ERR(cred))
		return PTR_ERR(cred);
670
	ctx = alloc_nfs_open_context(filp->f_path.mnt, filp->f_path.dentry, cred);
L
Linus Torvalds 已提交
671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699
	put_rpccred(cred);
	if (ctx == NULL)
		return -ENOMEM;
	ctx->mode = filp->f_mode;
	nfs_file_set_open_context(filp, ctx);
	put_nfs_open_context(ctx);
	return 0;
}

int nfs_release(struct inode *inode, struct file *filp)
{
	nfs_file_clear_open_context(filp);
	return 0;
}

/*
 * This function is called whenever some part of NFS notices that
 * the cached attributes have to be refreshed.
 */
int
__nfs_revalidate_inode(struct nfs_server *server, struct inode *inode)
{
	int		 status = -ESTALE;
	struct nfs_fattr fattr;
	struct nfs_inode *nfsi = NFS_I(inode);

	dfprintk(PAGECACHE, "NFS: revalidating (%s/%Ld)\n",
		inode->i_sb->s_id, (long long)NFS_FILEID(inode));

700
	nfs_inc_stats(inode, NFSIOS_INODEREVALIDATE);
701
	if (is_bad_inode(inode))
L
Linus Torvalds 已提交
702 703 704 705
 		goto out_nowait;
	if (NFS_STALE(inode))
 		goto out_nowait;

706 707 708
	status = nfs_wait_on_inode(inode);
	if (status < 0)
		goto out;
709 710 711 712

	status = -ESTALE;
	if (NFS_STALE(inode))
		goto out;
L
Linus Torvalds 已提交
713 714 715 716 717 718 719 720 721

	status = NFS_PROTO(inode)->getattr(server, NFS_FH(inode), &fattr);
	if (status != 0) {
		dfprintk(PAGECACHE, "nfs_revalidate_inode: (%s/%Ld) getattr failed, error=%d\n",
			 inode->i_sb->s_id,
			 (long long)NFS_FILEID(inode), status);
		if (status == -ESTALE) {
			nfs_zap_caches(inode);
			if (!S_ISDIR(inode->i_mode))
B
Benny Halevy 已提交
722
				set_bit(NFS_INO_STALE, &NFS_I(inode)->flags);
L
Linus Torvalds 已提交
723 724 725 726
		}
		goto out;
	}

727
	spin_lock(&inode->i_lock);
728
	status = nfs_update_inode(inode, &fattr);
L
Linus Torvalds 已提交
729
	if (status) {
730
		spin_unlock(&inode->i_lock);
L
Linus Torvalds 已提交
731 732 733 734 735
		dfprintk(PAGECACHE, "nfs_revalidate_inode: (%s/%Ld) refresh failed, error=%d\n",
			 inode->i_sb->s_id,
			 (long long)NFS_FILEID(inode), status);
		goto out;
	}
736
	spin_unlock(&inode->i_lock);
737

738
	if (nfsi->cache_validity & NFS_INO_INVALID_ACL)
739
		nfs_zap_acl_cache(inode);
740

L
Linus Torvalds 已提交
741 742 743 744
	dfprintk(PAGECACHE, "NFS: (%s/%Ld) revalidation complete\n",
		inode->i_sb->s_id,
		(long long)NFS_FILEID(inode));

745 746 747
 out:
	nfs_wake_up_inode(inode);

L
Linus Torvalds 已提交
748 749 750 751 752 753 754 755 756 757
 out_nowait:
	return status;
}

int nfs_attribute_timeout(struct inode *inode)
{
	struct nfs_inode *nfsi = NFS_I(inode);

	if (nfs_have_delegation(inode, FMODE_READ))
		return 0;
758 759 760 761 762 763 764
	/*
	 * Special case: if the attribute timeout is set to 0, then always
	 * 		 treat the cache as having expired (unless holding
	 * 		 a delegation).
	 */
	if (nfsi->attrtimeo == 0)
		return 1;
765
	return !time_in_range(jiffies, nfsi->read_cache_jiffies, nfsi->read_cache_jiffies + nfsi->attrtimeo);
L
Linus Torvalds 已提交
766 767 768 769 770 771 772 773 774 775 776
}

/**
 * nfs_revalidate_inode - Revalidate the inode attributes
 * @server - pointer to nfs_server struct
 * @inode - pointer to inode struct
 *
 * Updates inode attribute information by retrieving the data from the server.
 */
int nfs_revalidate_inode(struct nfs_server *server, struct inode *inode)
{
777
	if (!(NFS_I(inode)->cache_validity & NFS_INO_INVALID_ATTR)
L
Linus Torvalds 已提交
778 779 780 781 782
			&& !nfs_attribute_timeout(inode))
		return NFS_STALE(inode) ? -ESTALE : 0;
	return __nfs_revalidate_inode(server, inode);
}

783 784 785 786 787 788 789 790 791 792 793
static int nfs_invalidate_mapping_nolock(struct inode *inode, struct address_space *mapping)
{
	struct nfs_inode *nfsi = NFS_I(inode);
	
	if (mapping->nrpages != 0) {
		int ret = invalidate_inode_pages2(mapping);
		if (ret < 0)
			return ret;
	}
	spin_lock(&inode->i_lock);
	nfsi->cache_validity &= ~NFS_INO_INVALID_DATA;
794
	if (S_ISDIR(inode->i_mode))
795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816
		memset(nfsi->cookieverf, 0, sizeof(nfsi->cookieverf));
	spin_unlock(&inode->i_lock);
	nfs_inc_stats(inode, NFSIOS_DATAINVALIDATE);
	dfprintk(PAGECACHE, "NFS: (%s/%Ld) data cache invalidated\n",
			inode->i_sb->s_id, (long long)NFS_FILEID(inode));
	return 0;
}

static int nfs_invalidate_mapping(struct inode *inode, struct address_space *mapping)
{
	int ret = 0;

	mutex_lock(&inode->i_mutex);
	if (NFS_I(inode)->cache_validity & NFS_INO_INVALID_DATA) {
		ret = nfs_sync_mapping(mapping);
		if (ret == 0)
			ret = nfs_invalidate_mapping_nolock(inode, mapping);
	}
	mutex_unlock(&inode->i_mutex);
	return ret;
}

817
/**
818
 * nfs_revalidate_mapping_nolock - Revalidate the pagecache
819 820 821
 * @inode - pointer to host inode
 * @mapping - pointer to mapping
 */
822
int nfs_revalidate_mapping_nolock(struct inode *inode, struct address_space *mapping)
823 824
{
	struct nfs_inode *nfsi = NFS_I(inode);
825 826 827
	int ret = 0;

	if ((nfsi->cache_validity & NFS_INO_REVAL_PAGECACHE)
828
			|| nfs_attribute_timeout(inode) || NFS_STALE(inode)) {
829
		ret = __nfs_revalidate_inode(NFS_SERVER(inode), inode);
830 831 832 833 834 835 836 837
		if (ret < 0)
			goto out;
	}
	if (nfsi->cache_validity & NFS_INO_INVALID_DATA)
		ret = nfs_invalidate_mapping_nolock(inode, mapping);
out:
	return ret;
}
838

839 840 841 842 843 844 845 846 847 848 849 850
/**
 * nfs_revalidate_mapping - Revalidate the pagecache
 * @inode - pointer to host inode
 * @mapping - pointer to mapping
 *
 * This version of the function will take the inode->i_mutex and attempt to
 * flush out all dirty data if it needs to invalidate the page cache.
 */
int nfs_revalidate_mapping(struct inode *inode, struct address_space *mapping)
{
	struct nfs_inode *nfsi = NFS_I(inode);
	int ret = 0;
851

852 853 854 855 856
	if ((nfsi->cache_validity & NFS_INO_REVAL_PAGECACHE)
			|| nfs_attribute_timeout(inode) || NFS_STALE(inode)) {
		ret = __nfs_revalidate_inode(NFS_SERVER(inode), inode);
		if (ret < 0)
			goto out;
857
	}
858 859
	if (nfsi->cache_validity & NFS_INO_INVALID_DATA)
		ret = nfs_invalidate_mapping(inode, mapping);
860
out:
861
	return ret;
862 863
}

864 865 866 867
static void nfs_wcc_update_inode(struct inode *inode, struct nfs_fattr *fattr)
{
	struct nfs_inode *nfsi = NFS_I(inode);

868 869 870 871 872 873
	if ((fattr->valid & NFS_ATTR_WCC_V4) != 0 &&
			nfsi->change_attr == fattr->pre_change_attr) {
		nfsi->change_attr = fattr->change_attr;
		if (S_ISDIR(inode->i_mode))
			nfsi->cache_validity |= NFS_INO_INVALID_DATA;
	}
874 875
	/* If we have atomic WCC data, we may update some attributes */
	if ((fattr->valid & NFS_ATTR_WCC) != 0) {
876
		if (timespec_equal(&inode->i_ctime, &fattr->pre_ctime))
877 878 879
			memcpy(&inode->i_ctime, &fattr->ctime, sizeof(inode->i_ctime));
		if (timespec_equal(&inode->i_mtime, &fattr->pre_mtime)) {
			memcpy(&inode->i_mtime, &fattr->mtime, sizeof(inode->i_mtime));
880 881
			if (S_ISDIR(inode->i_mode))
				nfsi->cache_validity |= NFS_INO_INVALID_DATA;
882
		}
883
		if (i_size_read(inode) == nfs_size_to_loff_t(fattr->pre_size) &&
884
		    nfsi->npages == 0)
885
			i_size_write(inode, nfs_size_to_loff_t(fattr->size));
886 887 888
	}
}

L
Linus Torvalds 已提交
889
/**
890
 * nfs_check_inode_attributes - verify consistency of the inode attribute cache
L
Linus Torvalds 已提交
891 892 893 894 895 896 897
 * @inode - pointer to inode
 * @fattr - updated attributes
 *
 * Verifies the attribute cache. If we have just changed the attributes,
 * so that fattr carries weak cache consistency data, then it may
 * also update the ctime/mtime/change_attribute.
 */
898
static int nfs_check_inode_attributes(struct inode *inode, struct nfs_fattr *fattr)
L
Linus Torvalds 已提交
899 900 901
{
	struct nfs_inode *nfsi = NFS_I(inode);
	loff_t cur_size, new_isize;
902
	unsigned long invalid = 0;
L
Linus Torvalds 已提交
903

904

905 906 907 908 909 910
	/* Has the inode gone and changed behind our back? */
	if (nfsi->fileid != fattr->fileid
			|| (inode->i_mode & S_IFMT) != (fattr->mode & S_IFMT)) {
		return -EIO;
	}

911 912
	/* Do atomic weak cache consistency updates */
	nfs_wcc_update_inode(inode, fattr);
L
Linus Torvalds 已提交
913

914
	if ((fattr->valid & NFS_ATTR_FATTR_V4) != 0 &&
915
			nfsi->change_attr != fattr->change_attr)
916
		invalid |= NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE;
L
Linus Torvalds 已提交
917 918

	/* Verify a few of the more important attributes */
919
	if (!timespec_equal(&inode->i_mtime, &fattr->mtime))
920
		invalid |= NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE;
921 922 923

	cur_size = i_size_read(inode);
 	new_isize = nfs_size_to_loff_t(fattr->size);
924
	if (cur_size != new_isize && nfsi->npages == 0)
925
		invalid |= NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE;
L
Linus Torvalds 已提交
926 927 928 929 930

	/* Have any file permissions changed? */
	if ((inode->i_mode & S_IALLUGO) != (fattr->mode & S_IALLUGO)
			|| inode->i_uid != fattr->uid
			|| inode->i_gid != fattr->gid)
931
		invalid |= NFS_INO_INVALID_ATTR | NFS_INO_INVALID_ACCESS | NFS_INO_INVALID_ACL;
L
Linus Torvalds 已提交
932 933 934

	/* Has the link count changed? */
	if (inode->i_nlink != fattr->nlink)
935
		invalid |= NFS_INO_INVALID_ATTR;
L
Linus Torvalds 已提交
936 937

	if (!timespec_equal(&inode->i_atime, &fattr->atime))
938 939 940 941 942 943 944 945
		invalid |= NFS_INO_INVALID_ATIME;

	if (invalid != 0)
		nfsi->cache_validity |= invalid;
	else
		nfsi->cache_validity &= ~(NFS_INO_INVALID_ATTR
				| NFS_INO_INVALID_ATIME
				| NFS_INO_REVAL_PAGECACHE);
L
Linus Torvalds 已提交
946

947
	nfsi->read_cache_jiffies = fattr->time_start;
L
Linus Torvalds 已提交
948 949 950
	return 0;
}

951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969
/**
 * nfs_refresh_inode - try to update the inode attribute cache
 * @inode - pointer to inode
 * @fattr - updated attributes
 *
 * Check that an RPC call that returned attributes has not overlapped with
 * other recent updates of the inode metadata, then decide whether it is
 * safe to do a full update of the inode attributes, or whether just to
 * call nfs_check_inode_attributes.
 */
int nfs_refresh_inode(struct inode *inode, struct nfs_fattr *fattr)
{
	struct nfs_inode *nfsi = NFS_I(inode);
	int status;

	if ((fattr->valid & NFS_ATTR_FATTR) == 0)
		return 0;
	spin_lock(&inode->i_lock);
	if (time_after(fattr->time_start, nfsi->last_updated))
970
		status = nfs_update_inode(inode, fattr);
971 972 973 974 975 976 977
	else
		status = nfs_check_inode_attributes(inode, fattr);

	spin_unlock(&inode->i_lock);
	return status;
}

978 979 980 981 982 983 984
/**
 * nfs_post_op_update_inode - try to update the inode attribute cache
 * @inode - pointer to inode
 * @fattr - updated attributes
 *
 * After an operation that has changed the inode metadata, mark the
 * attribute cache as being invalid, then try to update it.
985 986 987 988 989 990
 *
 * NB: if the server didn't return any post op attributes, this
 * function will force the retrieval of attributes before the next
 * NFS request.  Thus it should be used only for operations that
 * are expected to change one or more attributes, to avoid
 * unnecessary NFS requests and trips through nfs_update_inode().
991 992 993 994 995
 */
int nfs_post_op_update_inode(struct inode *inode, struct nfs_fattr *fattr)
{
	struct nfs_inode *nfsi = NFS_I(inode);

996
	spin_lock(&inode->i_lock);
997
	nfsi->cache_validity |= NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE;
998 999
	if (S_ISDIR(inode->i_mode))
		nfsi->cache_validity |= NFS_INO_INVALID_DATA;
1000
	spin_unlock(&inode->i_lock);
1001
	return nfs_refresh_inode(inode, fattr);
1002 1003
}

1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025
/**
 * nfs_post_op_update_inode_force_wcc - try to update the inode attribute cache
 * @inode - pointer to inode
 * @fattr - updated attributes
 *
 * After an operation that has changed the inode metadata, mark the
 * attribute cache as being invalid, then try to update it. Fake up
 * weak cache consistency data, if none exist.
 *
 * This function is mainly designed to be used by the ->write_done() functions.
 */
int nfs_post_op_update_inode_force_wcc(struct inode *inode, struct nfs_fattr *fattr)
{
	if ((fattr->valid & NFS_ATTR_FATTR_V4) != 0 &&
			(fattr->valid & NFS_ATTR_WCC_V4) == 0) {
		fattr->pre_change_attr = NFS_I(inode)->change_attr;
		fattr->valid |= NFS_ATTR_WCC_V4;
	}
	if ((fattr->valid & NFS_ATTR_FATTR) != 0 &&
			(fattr->valid & NFS_ATTR_WCC) == 0) {
		memcpy(&fattr->pre_ctime, &inode->i_ctime, sizeof(fattr->pre_ctime));
		memcpy(&fattr->pre_mtime, &inode->i_mtime, sizeof(fattr->pre_mtime));
1026
		fattr->pre_size = i_size_read(inode);
1027 1028 1029 1030 1031
		fattr->valid |= NFS_ATTR_WCC;
	}
	return nfs_post_op_update_inode(inode, fattr);
}

L
Linus Torvalds 已提交
1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043
/*
 * Many nfs protocol calls return the new file attributes after
 * an operation.  Here we update the inode to reflect the state
 * of the server's inode.
 *
 * This is a bit tricky because we have to make sure all dirty pages
 * have been sent off to the server before calling invalidate_inode_pages.
 * To make sure no other process adds more write requests while we try
 * our best to flush them, we make them sleep during the attribute refresh.
 *
 * A very similar scenario holds for the dir cache.
 */
1044
static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
L
Linus Torvalds 已提交
1045
{
1046
	struct nfs_server *server;
L
Linus Torvalds 已提交
1047
	struct nfs_inode *nfsi = NFS_I(inode);
1048
	loff_t cur_isize, new_isize;
1049
	unsigned long invalid = 0;
1050
	unsigned long now = jiffies;
L
Linus Torvalds 已提交
1051 1052

	dfprintk(VFS, "NFS: %s(%s/%ld ct=%d info=0x%x)\n",
1053
			__func__, inode->i_sb->s_id, inode->i_ino,
L
Linus Torvalds 已提交
1054 1055
			atomic_read(&inode->i_count), fattr->valid);

1056 1057
	if (nfsi->fileid != fattr->fileid)
		goto out_fileid;
L
Linus Torvalds 已提交
1058 1059 1060 1061

	/*
	 * Make sure the inode's type hasn't changed.
	 */
1062
	if ((inode->i_mode & S_IFMT) != (fattr->mode & S_IFMT))
L
Linus Torvalds 已提交
1063 1064
		goto out_changed;

1065
	server = NFS_SERVER(inode);
T
Trond Myklebust 已提交
1066
	/* Update the fsid? */
1067 1068 1069
	if (S_ISDIR(inode->i_mode) &&
			!nfs_fsid_equal(&server->fsid, &fattr->fsid) &&
			!test_bit(NFS_INO_MOUNTPOINT, &nfsi->flags))
1070 1071
		server->fsid = fattr->fsid;

L
Linus Torvalds 已提交
1072 1073 1074
	/*
	 * Update the read time so we don't revalidate too often.
	 */
1075
	nfsi->read_cache_jiffies = fattr->time_start;
1076

1077 1078
	nfsi->cache_validity &= ~(NFS_INO_INVALID_ATTR | NFS_INO_INVALID_ATIME
			| NFS_INO_REVAL_PAGECACHE);
L
Linus Torvalds 已提交
1079

1080 1081 1082
	/* Do atomic weak cache consistency updates */
	nfs_wcc_update_inode(inode, fattr);

1083 1084 1085 1086 1087 1088 1089
	/* More cache consistency checks */
	if (!(fattr->valid & NFS_ATTR_FATTR_V4)) {
		/* NFSv2/v3: Check if the mtime agrees */
		if (!timespec_equal(&inode->i_mtime, &fattr->mtime)) {
			dprintk("NFS: mtime change on server for file %s/%ld\n",
					inode->i_sb->s_id, inode->i_ino);
			invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_DATA;
1090 1091
			if (S_ISDIR(inode->i_mode))
				nfs_force_lookup_revalidate(inode);
1092 1093
		}
		/* If ctime has changed we should definitely clear access+acl caches */
1094
		if (!timespec_equal(&inode->i_ctime, &fattr->ctime))
1095 1096 1097 1098 1099
			invalid |= NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
	} else if (nfsi->change_attr != fattr->change_attr) {
		dprintk("NFS: change_attr change on server for file %s/%ld\n",
				inode->i_sb->s_id, inode->i_ino);
		invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_DATA|NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
1100 1101
		if (S_ISDIR(inode->i_mode))
			nfs_force_lookup_revalidate(inode);
1102 1103
	}

1104
	/* Check if our cached file size is stale */
L
Linus Torvalds 已提交
1105 1106
 	new_isize = nfs_size_to_loff_t(fattr->size);
	cur_isize = i_size_read(inode);
1107
	if (new_isize != cur_isize) {
1108 1109 1110
		/* Do we perhaps have any outstanding writes, or has
		 * the file grown beyond our last write? */
		if (nfsi->npages == 0 || new_isize > cur_isize) {
1111
			i_size_write(inode, new_isize);
L
Linus Torvalds 已提交
1112 1113
			invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_DATA;
		}
1114 1115
		dprintk("NFS: isize change on server for file %s/%ld\n",
				inode->i_sb->s_id, inode->i_ino);
L
Linus Torvalds 已提交
1116 1117 1118
	}


1119 1120
	memcpy(&inode->i_mtime, &fattr->mtime, sizeof(inode->i_mtime));
	memcpy(&inode->i_ctime, &fattr->ctime, sizeof(inode->i_ctime));
L
Linus Torvalds 已提交
1121
	memcpy(&inode->i_atime, &fattr->atime, sizeof(inode->i_atime));
1122
	nfsi->change_attr = fattr->change_attr;
L
Linus Torvalds 已提交
1123 1124 1125 1126

	if ((inode->i_mode & S_IALLUGO) != (fattr->mode & S_IALLUGO) ||
	    inode->i_uid != fattr->uid ||
	    inode->i_gid != fattr->gid)
1127
		invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
L
Linus Torvalds 已提交
1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144

	inode->i_mode = fattr->mode;
	inode->i_nlink = fattr->nlink;
	inode->i_uid = fattr->uid;
	inode->i_gid = fattr->gid;

	if (fattr->valid & (NFS_ATTR_FATTR_V3 | NFS_ATTR_FATTR_V4)) {
		/*
		 * report the blocks in 512byte units
		 */
		inode->i_blocks = nfs_calc_block_size(fattr->du.nfs3.used);
 	} else {
 		inode->i_blocks = fattr->du.nfs2.blocks;
 	}

	/* Update attrtimeo value if we're out of the unstable period */
	if (invalid & NFS_INO_INVALID_ATTR) {
C
Chuck Lever 已提交
1145
		nfs_inc_stats(inode, NFSIOS_ATTRINVALIDATE);
L
Linus Torvalds 已提交
1146
		nfsi->attrtimeo = NFS_MINATTRTIMEO(inode);
1147
		nfsi->attrtimeo_timestamp = now;
1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159
		nfsi->last_updated = now;
	} else {
		if (!time_in_range(now, nfsi->attrtimeo_timestamp, nfsi->attrtimeo_timestamp + nfsi->attrtimeo)) {
			if ((nfsi->attrtimeo <<= 1) > NFS_MAXATTRTIMEO(inode))
				nfsi->attrtimeo = NFS_MAXATTRTIMEO(inode);
			nfsi->attrtimeo_timestamp = now;
		}
		/*
		 * Avoid jiffy wraparound issues with nfsi->last_updated
		 */
		if (!time_in_range(nfsi->last_updated, nfsi->read_cache_jiffies, now))
			nfsi->last_updated = nfsi->read_cache_jiffies;
L
Linus Torvalds 已提交
1160
	}
1161
	invalid &= ~NFS_INO_INVALID_ATTR;
L
Linus Torvalds 已提交
1162 1163 1164 1165
	/* Don't invalidate the data if we were to blame */
	if (!(S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode)
				|| S_ISLNK(inode->i_mode)))
		invalid &= ~NFS_INO_INVALID_DATA;
1166 1167
	if (!nfs_have_delegation(inode, FMODE_READ) ||
			(nfsi->cache_validity & NFS_INO_REVAL_FORCED))
1168
		nfsi->cache_validity |= invalid;
1169
	nfsi->cache_validity &= ~NFS_INO_REVAL_FORCED;
L
Linus Torvalds 已提交
1170 1171 1172 1173 1174 1175 1176

	return 0;
 out_changed:
	/*
	 * Big trouble! The inode has become a different object.
	 */
	printk(KERN_DEBUG "%s: inode %ld mode changed, %07o to %07o\n",
1177
			__func__, inode->i_ino, inode->i_mode, fattr->mode);
1178
 out_err:
L
Linus Torvalds 已提交
1179 1180 1181 1182 1183 1184 1185
	/*
	 * No need to worry about unhashing the dentry, as the
	 * lookup validation will know that the inode is bad.
	 * (But we fall through to invalidate the caches.)
	 */
	nfs_invalidate_inode(inode);
	return -ESTALE;
1186 1187 1188 1189

 out_fileid:
	printk(KERN_ERR "NFS: server %s error: fileid changed\n"
		"fsid %s: expected fileid 0x%Lx, got 0x%Lx\n",
1190
		NFS_SERVER(inode)->nfs_client->cl_hostname, inode->i_sb->s_id,
1191 1192
		(long long)nfsi->fileid, (long long)fattr->fileid);
	goto out_err;
L
Linus Torvalds 已提交
1193 1194
}

1195

L
Linus Torvalds 已提交
1196 1197 1198 1199 1200 1201 1202
#ifdef CONFIG_NFS_V4

/*
 * Clean out any remaining NFSv4 state that might be left over due
 * to open() calls that passed nfs_atomic_lookup, but failed to call
 * nfs_open().
 */
D
David Howells 已提交
1203
void nfs4_clear_inode(struct inode *inode)
L
Linus Torvalds 已提交
1204 1205
{
	/* If we are holding a delegation, return it! */
1206
	nfs_inode_return_delegation_noreclaim(inode);
L
Linus Torvalds 已提交
1207 1208 1209 1210 1211
	/* First call standard NFS clear_inode() code */
	nfs_clear_inode(inode);
}
#endif

D
David Howells 已提交
1212
struct inode *nfs_alloc_inode(struct super_block *sb)
L
Linus Torvalds 已提交
1213 1214
{
	struct nfs_inode *nfsi;
1215
	nfsi = (struct nfs_inode *)kmem_cache_alloc(nfs_inode_cachep, GFP_KERNEL);
L
Linus Torvalds 已提交
1216 1217
	if (!nfsi)
		return NULL;
1218 1219
	nfsi->flags = 0UL;
	nfsi->cache_validity = 0UL;
1220 1221 1222 1223
#ifdef CONFIG_NFS_V3_ACL
	nfsi->acl_access = ERR_PTR(-EAGAIN);
	nfsi->acl_default = ERR_PTR(-EAGAIN);
#endif
1224 1225 1226
#ifdef CONFIG_NFS_V4
	nfsi->nfs4_acl = NULL;
#endif /* CONFIG_NFS_V4 */
L
Linus Torvalds 已提交
1227 1228 1229
	return &nfsi->vfs_inode;
}

D
David Howells 已提交
1230
void nfs_destroy_inode(struct inode *inode)
L
Linus Torvalds 已提交
1231 1232 1233 1234
{
	kmem_cache_free(nfs_inode_cachep, NFS_I(inode));
}

A
Andrew Morton 已提交
1235 1236 1237 1238 1239 1240 1241 1242 1243
static inline void nfs4_init_once(struct nfs_inode *nfsi)
{
#ifdef CONFIG_NFS_V4
	INIT_LIST_HEAD(&nfsi->open_states);
	nfsi->delegation = NULL;
	nfsi->delegation_state = 0;
	init_rwsem(&nfsi->rwsem);
#endif
}
D
David Howells 已提交
1244

1245
static void init_once(void *foo)
L
Linus Torvalds 已提交
1246 1247 1248
{
	struct nfs_inode *nfsi = (struct nfs_inode *) foo;

C
Christoph Lameter 已提交
1249 1250 1251 1252 1253 1254 1255
	inode_init_once(&nfsi->vfs_inode);
	INIT_LIST_HEAD(&nfsi->open_files);
	INIT_LIST_HEAD(&nfsi->access_cache_entry_lru);
	INIT_LIST_HEAD(&nfsi->access_cache_inode_lru);
	INIT_RADIX_TREE(&nfsi->nfs_page_tree, GFP_ATOMIC);
	nfsi->ncommit = 0;
	nfsi->npages = 0;
T
Trond Myklebust 已提交
1256 1257 1258
	atomic_set(&nfsi->silly_count, 1);
	INIT_HLIST_HEAD(&nfsi->silly_list);
	init_waitqueue_head(&nfsi->waitqueue);
C
Christoph Lameter 已提交
1259
	nfs4_init_once(nfsi);
L
Linus Torvalds 已提交
1260
}
1261

D
David Howells 已提交
1262
static int __init nfs_init_inodecache(void)
L
Linus Torvalds 已提交
1263 1264 1265
{
	nfs_inode_cachep = kmem_cache_create("nfs_inode_cache",
					     sizeof(struct nfs_inode),
1266 1267
					     0, (SLAB_RECLAIM_ACCOUNT|
						SLAB_MEM_SPREAD),
1268
					     init_once);
L
Linus Torvalds 已提交
1269 1270 1271 1272 1273 1274
	if (nfs_inode_cachep == NULL)
		return -ENOMEM;

	return 0;
}

1275
static void nfs_destroy_inodecache(void)
L
Linus Torvalds 已提交
1276
{
1277
	kmem_cache_destroy(nfs_inode_cachep);
L
Linus Torvalds 已提交
1278 1279
}

T
Trond Myklebust 已提交
1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309
struct workqueue_struct *nfsiod_workqueue;

/*
 * start up the nfsiod workqueue
 */
static int nfsiod_start(void)
{
	struct workqueue_struct *wq;
	dprintk("RPC:       creating workqueue nfsiod\n");
	wq = create_singlethread_workqueue("nfsiod");
	if (wq == NULL)
		return -ENOMEM;
	nfsiod_workqueue = wq;
	return 0;
}

/*
 * Destroy the nfsiod workqueue
 */
static void nfsiod_stop(void)
{
	struct workqueue_struct *wq;

	wq = nfsiod_workqueue;
	if (wq == NULL)
		return;
	nfsiod_workqueue = NULL;
	destroy_workqueue(wq);
}

L
Linus Torvalds 已提交
1310 1311 1312 1313 1314 1315 1316
/*
 * Initialize NFS
 */
static int __init init_nfs_fs(void)
{
	int err;

T
Trond Myklebust 已提交
1317 1318 1319 1320
	err = nfsiod_start();
	if (err)
		goto out6;

1321 1322 1323 1324
	err = nfs_fs_proc_init();
	if (err)
		goto out5;

L
Linus Torvalds 已提交
1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347
	err = nfs_init_nfspagecache();
	if (err)
		goto out4;

	err = nfs_init_inodecache();
	if (err)
		goto out3;

	err = nfs_init_readpagecache();
	if (err)
		goto out2;

	err = nfs_init_writepagecache();
	if (err)
		goto out1;

	err = nfs_init_directcache();
	if (err)
		goto out0;

#ifdef CONFIG_PROC_FS
	rpc_proc_register(&nfs_rpcstat);
#endif
D
David Howells 已提交
1348
	if ((err = register_nfs_fs()) != 0)
L
Linus Torvalds 已提交
1349 1350 1351 1352 1353 1354 1355
		goto out;
	return 0;
out:
#ifdef CONFIG_PROC_FS
	rpc_proc_unregister("nfs");
#endif
	nfs_destroy_directcache();
1356 1357
out0:
	nfs_destroy_writepagecache();
L
Linus Torvalds 已提交
1358 1359 1360 1361 1362 1363 1364
out1:
	nfs_destroy_readpagecache();
out2:
	nfs_destroy_inodecache();
out3:
	nfs_destroy_nfspagecache();
out4:
1365 1366
	nfs_fs_proc_exit();
out5:
T
Trond Myklebust 已提交
1367 1368
	nfsiod_stop();
out6:
L
Linus Torvalds 已提交
1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381
	return err;
}

static void __exit exit_nfs_fs(void)
{
	nfs_destroy_directcache();
	nfs_destroy_writepagecache();
	nfs_destroy_readpagecache();
	nfs_destroy_inodecache();
	nfs_destroy_nfspagecache();
#ifdef CONFIG_PROC_FS
	rpc_proc_unregister("nfs");
#endif
D
David Howells 已提交
1382
	unregister_nfs_fs();
1383
	nfs_fs_proc_exit();
T
Trond Myklebust 已提交
1384
	nfsiod_stop();
L
Linus Torvalds 已提交
1385 1386 1387 1388 1389
}

/* Not quite true; I just maintain it */
MODULE_AUTHOR("Olaf Kirch <okir@monad.swb.de>");
MODULE_LICENSE("GPL");
1390
module_param(enable_ino64, bool, 0644);
L
Linus Torvalds 已提交
1391 1392 1393

module_init(init_nfs_fs)
module_exit(exit_nfs_fs)