inode.c 38.7 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7
/*
 *  linux/fs/nfs/inode.c
 *
 *  Copyright (C) 1992  Rick Sladkey
 *
 *  nfs inode and superblock handling functions
 *
A
Alan Cox 已提交
8
 *  Modularised by Alan Cox <alan@lxorguk.ukuu.org.uk>, while hacking some
L
Linus Torvalds 已提交
9 10 11 12 13 14 15 16 17
 *  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
		nfsi->attr_gencount = fattr->gencount;
L
Linus Torvalds 已提交
309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326
		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);
327
		nfsi->attrtimeo_timestamp = now;
L
Linus Torvalds 已提交
328
		memset(nfsi->cookieverf, 0, sizeof(nfsi->cookieverf));
329
		nfsi->access_cache = RB_ROOT;
L
Linus Torvalds 已提交
330 331 332 333 334 335 336 337 338 339 340 341 342

		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:
343
	dprintk("nfs_fhget: iget failed with error %ld\n", PTR_ERR(inode));
L
Linus Torvalds 已提交
344 345 346
	goto out;
}

347
#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
348 349 350 351 352 353 354 355

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
356 357
	nfs_inc_stats(inode, NFSIOS_VFSSETATTR);

358 359 360 361
	/* 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
362 363 364 365 366 367 368
	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;
369
	if ((attr->ia_valid & ~ATTR_FILE) == 0)
L
Linus Torvalds 已提交
370 371
		return 0;

372
	/* Write all dirty data */
373 374 375 376
	if (S_ISREG(inode->i_mode)) {
		filemap_write_and_wait(inode->i_mapping);
		nfs_wb_all(inode);
	}
377 378 379 380 381
	/*
	 * 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
382
	error = NFS_PROTO(inode)->setattr(dentry, &fattr, attr);
383
	if (error == 0)
L
Linus Torvalds 已提交
384
		nfs_refresh_inode(inode, &fattr);
385 386 387
	return error;
}

388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443
/**
 * 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;
}

444 445 446 447 448 449 450 451 452 453 454
/**
 * 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) {
455
		spin_lock(&inode->i_lock);
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
		NFS_I(inode)->cache_validity |= NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
466
		spin_unlock(&inode->i_lock);
467 468
	}
	if ((attr->ia_valid & ATTR_SIZE) != 0) {
C
Chuck Lever 已提交
469
		nfs_inc_stats(inode, NFSIOS_SETATTRTRUNC);
470
		nfs_vmtruncate(inode, attr->ia_size);
471
	}
L
Linus Torvalds 已提交
472 473 474 475 476
}

int nfs_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat)
{
	struct inode *inode = dentry->d_inode;
477
	int need_atime = NFS_I(inode)->cache_validity & NFS_INO_INVALID_ATIME;
L
Linus Torvalds 已提交
478 479
	int err;

480 481 482 483 484 485 486 487 488
	/*
	 * 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);
489
		nfs_wb_nocommit(inode);
490 491
		mutex_unlock(&inode->i_mutex);
	}
492 493 494 495 496 497 498 499 500 501 502 503

	/*
	 * 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
504
		need_atime = 0;
505

L
Linus Torvalds 已提交
506 507 508 509
	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
510
	if (!err) {
L
Linus Torvalds 已提交
511
		generic_fillattr(inode, stat);
512
		stat->ino = nfs_compat_user_ino64(NFS_FILEID(inode));
P
Peter Staubach 已提交
513
	}
L
Linus Torvalds 已提交
514 515 516
	return err;
}

517
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
518 519 520
{
	struct nfs_open_context *ctx;

521
	ctx = kmalloc(sizeof(*ctx), GFP_KERNEL);
L
Linus Torvalds 已提交
522
	if (ctx != NULL) {
523 524
		ctx->path.dentry = dget(dentry);
		ctx->path.mnt = mntget(mnt);
L
Linus Torvalds 已提交
525 526 527
		ctx->cred = get_rpccred(cred);
		ctx->state = NULL;
		ctx->lockowner = current->files;
528
		ctx->flags = 0;
L
Linus Torvalds 已提交
529
		ctx->error = 0;
530
		ctx->dir_cookie = 0;
T
Trond Myklebust 已提交
531
		atomic_set(&ctx->count, 1);
L
Linus Torvalds 已提交
532 533 534 535 536 537 538
	}
	return ctx;
}

struct nfs_open_context *get_nfs_open_context(struct nfs_open_context *ctx)
{
	if (ctx != NULL)
T
Trond Myklebust 已提交
539
		atomic_inc(&ctx->count);
L
Linus Torvalds 已提交
540 541 542
	return ctx;
}

543
static void __put_nfs_open_context(struct nfs_open_context *ctx, int wait)
L
Linus Torvalds 已提交
544
{
545
	struct inode *inode;
546

547 548 549 550
	if (ctx == NULL)
		return;

	inode = ctx->path.dentry->d_inode;
T
Trond Myklebust 已提交
551 552 553 554
	if (!atomic_dec_and_lock(&ctx->count, &inode->i_lock))
		return;
	list_del(&ctx->list);
	spin_unlock(&inode->i_lock);
555 556 557 558 559 560
	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);
	}
561 562
	if (ctx->cred != NULL)
		put_rpccred(ctx->cred);
J
Jan Blunck 已提交
563
	path_put(&ctx->path);
564 565 566
	kfree(ctx);
}

567 568 569 570 571 572 573 574 575 576
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
577 578 579 580
/*
 * Ensure that mmap has a recent RPC credential for use when writing out
 * shared pages
 */
581
static void nfs_file_set_open_context(struct file *filp, struct nfs_open_context *ctx)
L
Linus Torvalds 已提交
582
{
583
	struct inode *inode = filp->f_path.dentry->d_inode;
L
Linus Torvalds 已提交
584 585 586 587 588 589 590 591
	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);
}

592 593 594
/*
 * Given an inode, search for an open context with the desired characteristics
 */
595
struct nfs_open_context *nfs_find_open_context(struct inode *inode, struct rpc_cred *cred, fmode_t mode)
L
Linus Torvalds 已提交
596 597 598 599 600 601
{
	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) {
602 603
		if (cred != NULL && pos->cred != cred)
			continue;
L
Linus Torvalds 已提交
604 605 606 607 608 609 610 611 612
		if ((pos->mode & mode) == mode) {
			ctx = get_nfs_open_context(pos);
			break;
		}
	}
	spin_unlock(&inode->i_lock);
	return ctx;
}

613
static void nfs_file_clear_open_context(struct file *filp)
L
Linus Torvalds 已提交
614
{
615
	struct inode *inode = filp->f_path.dentry->d_inode;
616
	struct nfs_open_context *ctx = nfs_file_open_context(filp);
L
Linus Torvalds 已提交
617 618 619 620 621 622

	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);
623
		put_nfs_open_context_sync(ctx);
L
Linus Torvalds 已提交
624 625 626 627 628 629 630 631 632 633 634
	}
}

/*
 * 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;

635
	cred = rpc_lookup_cred();
L
Linus Torvalds 已提交
636 637
	if (IS_ERR(cred))
		return PTR_ERR(cred);
638
	ctx = alloc_nfs_open_context(filp->f_path.mnt, filp->f_path.dentry, cred);
L
Linus Torvalds 已提交
639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667
	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));

668
	if (is_bad_inode(inode))
669
		goto out;
L
Linus Torvalds 已提交
670
	if (NFS_STALE(inode))
671
		goto out;
672

673
	nfs_inc_stats(inode, NFSIOS_INODEREVALIDATE);
L
Linus Torvalds 已提交
674 675 676 677 678 679 680 681
	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
682
				set_bit(NFS_INO_STALE, &NFS_I(inode)->flags);
L
Linus Torvalds 已提交
683 684 685 686
		}
		goto out;
	}

687
	status = nfs_refresh_inode(inode, &fattr);
L
Linus Torvalds 已提交
688 689 690 691 692 693
	if (status) {
		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;
	}
694

695
	if (nfsi->cache_validity & NFS_INO_INVALID_ACL)
696
		nfs_zap_acl_cache(inode);
697

L
Linus Torvalds 已提交
698 699 700 701
	dfprintk(PAGECACHE, "NFS: (%s/%Ld) revalidation complete\n",
		inode->i_sb->s_id,
		(long long)NFS_FILEID(inode));

702
 out:
L
Linus Torvalds 已提交
703 704 705 706 707 708 709 710 711
	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;
712
	return !time_in_range_open(jiffies, nfsi->read_cache_jiffies, nfsi->read_cache_jiffies + nfsi->attrtimeo);
L
Linus Torvalds 已提交
713 714 715 716 717 718 719 720 721 722 723
}

/**
 * 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)
{
724
	if (!(NFS_I(inode)->cache_validity & NFS_INO_INVALID_ATTR)
L
Linus Torvalds 已提交
725 726 727 728 729
			&& !nfs_attribute_timeout(inode))
		return NFS_STALE(inode) ? -ESTALE : 0;
	return __nfs_revalidate_inode(server, inode);
}

730 731 732 733 734 735 736 737 738 739 740
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;
741
	if (S_ISDIR(inode->i_mode))
742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763
		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;
}

764
/**
765
 * nfs_revalidate_mapping_nolock - Revalidate the pagecache
766 767 768
 * @inode - pointer to host inode
 * @mapping - pointer to mapping
 */
769
int nfs_revalidate_mapping_nolock(struct inode *inode, struct address_space *mapping)
770 771
{
	struct nfs_inode *nfsi = NFS_I(inode);
772 773 774
	int ret = 0;

	if ((nfsi->cache_validity & NFS_INO_REVAL_PAGECACHE)
775
			|| nfs_attribute_timeout(inode) || NFS_STALE(inode)) {
776
		ret = __nfs_revalidate_inode(NFS_SERVER(inode), inode);
777 778 779 780 781 782 783 784
		if (ret < 0)
			goto out;
	}
	if (nfsi->cache_validity & NFS_INO_INVALID_DATA)
		ret = nfs_invalidate_mapping_nolock(inode, mapping);
out:
	return ret;
}
785

786 787 788 789 790 791 792 793 794 795 796 797
/**
 * 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;
798

799 800 801 802 803
	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;
804
	}
805 806
	if (nfsi->cache_validity & NFS_INO_INVALID_DATA)
		ret = nfs_invalidate_mapping(inode, mapping);
807
out:
808
	return ret;
809 810
}

811 812 813 814
static void nfs_wcc_update_inode(struct inode *inode, struct nfs_fattr *fattr)
{
	struct nfs_inode *nfsi = NFS_I(inode);

815 816 817 818 819 820
	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;
	}
821 822
	/* If we have atomic WCC data, we may update some attributes */
	if ((fattr->valid & NFS_ATTR_WCC) != 0) {
823
		if (timespec_equal(&inode->i_ctime, &fattr->pre_ctime))
824 825 826
			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));
827 828
			if (S_ISDIR(inode->i_mode))
				nfsi->cache_validity |= NFS_INO_INVALID_DATA;
829
		}
830
		if (i_size_read(inode) == nfs_size_to_loff_t(fattr->pre_size) &&
831
		    nfsi->npages == 0)
832
			i_size_write(inode, nfs_size_to_loff_t(fattr->size));
833 834 835
	}
}

L
Linus Torvalds 已提交
836
/**
837
 * nfs_check_inode_attributes - verify consistency of the inode attribute cache
L
Linus Torvalds 已提交
838 839 840 841 842 843 844
 * @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.
 */
845
static int nfs_check_inode_attributes(struct inode *inode, struct nfs_fattr *fattr)
L
Linus Torvalds 已提交
846 847 848
{
	struct nfs_inode *nfsi = NFS_I(inode);
	loff_t cur_size, new_isize;
849
	unsigned long invalid = 0;
L
Linus Torvalds 已提交
850

851

852 853 854 855 856 857
	/* 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;
	}

858
	if ((fattr->valid & NFS_ATTR_FATTR_V4) != 0 &&
859
			nfsi->change_attr != fattr->change_attr)
860
		invalid |= NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE;
L
Linus Torvalds 已提交
861 862

	/* Verify a few of the more important attributes */
863
	if (!timespec_equal(&inode->i_mtime, &fattr->mtime))
864
		invalid |= NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE;
865 866 867

	cur_size = i_size_read(inode);
 	new_isize = nfs_size_to_loff_t(fattr->size);
868
	if (cur_size != new_isize && nfsi->npages == 0)
869
		invalid |= NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE;
L
Linus Torvalds 已提交
870 871 872 873 874

	/* 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)
875
		invalid |= NFS_INO_INVALID_ATTR | NFS_INO_INVALID_ACCESS | NFS_INO_INVALID_ACL;
L
Linus Torvalds 已提交
876 877 878

	/* Has the link count changed? */
	if (inode->i_nlink != fattr->nlink)
879
		invalid |= NFS_INO_INVALID_ATTR;
L
Linus Torvalds 已提交
880 881

	if (!timespec_equal(&inode->i_atime, &fattr->atime))
882 883 884 885
		invalid |= NFS_INO_INVALID_ATIME;

	if (invalid != 0)
		nfsi->cache_validity |= invalid;
L
Linus Torvalds 已提交
886

887
	nfsi->read_cache_jiffies = fattr->time_start;
L
Linus Torvalds 已提交
888 889 890
	return 0;
}

891
static int nfs_ctime_need_update(const struct inode *inode, const struct nfs_fattr *fattr)
892
{
893 894 895 896 897 898 899
	return timespec_compare(&fattr->ctime, &inode->i_ctime) > 0;
}

static int nfs_size_need_update(const struct inode *inode, const struct nfs_fattr *fattr)
{
	return nfs_size_to_loff_t(fattr->size) > i_size_read(inode);
}
900

901
static atomic_long_t nfs_attr_generation_counter;
902 903 904

static unsigned long nfs_read_attr_generation_counter(void)
{
905
	return atomic_long_read(&nfs_attr_generation_counter);
906 907 908 909
}

unsigned long nfs_inc_attr_generation_counter(void)
{
910
	return atomic_long_inc_return(&nfs_attr_generation_counter);
911 912 913 914 915 916 917 918 919
}

void nfs_fattr_init(struct nfs_fattr *fattr)
{
	fattr->valid = 0;
	fattr->time_start = jiffies;
	fattr->gencount = nfs_inc_attr_generation_counter();
}

920 921 922 923 924 925 926 927 928 929 930 931 932
/**
 * nfs_inode_attrs_need_update - check if the inode attributes need updating
 * @inode - pointer to inode
 * @fattr - attributes
 *
 * Attempt to divine whether or not an RPC call reply carrying stale
 * attributes got scheduled after another call carrying updated ones.
 *
 * To do so, the function first assumes that a more recent ctime means
 * that the attributes in fattr are newer, however it also attempt to
 * catch the case where ctime either didn't change, or went backwards
 * (if someone reset the clock on the server) by looking at whether
 * or not this RPC call was started after the inode was last updated.
933
 * Note also the check for wraparound of 'attr_gencount'
934 935 936 937 938 939 940 941 942
 *
 * The function returns 'true' if it thinks the attributes in 'fattr' are
 * more recent than the ones cached in the inode.
 *
 */
static int nfs_inode_attrs_need_update(const struct inode *inode, const struct nfs_fattr *fattr)
{
	const struct nfs_inode *nfsi = NFS_I(inode);

943
	return ((long)fattr->gencount - (long)nfsi->attr_gencount) > 0 ||
T
Trond Myklebust 已提交
944 945
		nfs_ctime_need_update(inode, fattr) ||
		nfs_size_need_update(inode, fattr) ||
946
		((long)nfsi->attr_gencount - (long)nfs_read_attr_generation_counter() > 0);
947 948 949 950 951
}

static int nfs_refresh_inode_locked(struct inode *inode, struct nfs_fattr *fattr)
{
	if (nfs_inode_attrs_need_update(inode, fattr))
952 953 954 955
		return nfs_update_inode(inode, fattr);
	return nfs_check_inode_attributes(inode, fattr);
}

956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972
/**
 * 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)
{
	int status;

	if ((fattr->valid & NFS_ATTR_FATTR) == 0)
		return 0;
	spin_lock(&inode->i_lock);
973
	status = nfs_refresh_inode_locked(inode, fattr);
974 975 976 977
	spin_unlock(&inode->i_lock);
	return status;
}

978 979 980 981 982 983 984 985 986 987 988 989
static int nfs_post_op_update_inode_locked(struct inode *inode, struct nfs_fattr *fattr)
{
	struct nfs_inode *nfsi = NFS_I(inode);

	nfsi->cache_validity |= NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE;
	if (S_ISDIR(inode->i_mode))
		nfsi->cache_validity |= NFS_INO_INVALID_DATA;
	if ((fattr->valid & NFS_ATTR_FATTR) == 0)
		return 0;
	return nfs_refresh_inode_locked(inode, fattr);
}

990 991 992 993 994 995 996
/**
 * 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.
997 998 999 1000 1001 1002
 *
 * 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().
1003 1004 1005
 */
int nfs_post_op_update_inode(struct inode *inode, struct nfs_fattr *fattr)
{
1006
	int status;
1007

1008
	spin_lock(&inode->i_lock);
1009
	status = nfs_post_op_update_inode_locked(inode, fattr);
1010
	spin_unlock(&inode->i_lock);
1011
	return status;
1012 1013
}

1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026
/**
 * 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)
{
1027 1028 1029 1030 1031 1032 1033 1034 1035
	int status;

	spin_lock(&inode->i_lock);
	/* Don't do a WCC update if these attributes are already stale */
	if ((fattr->valid & NFS_ATTR_FATTR) == 0 ||
			!nfs_inode_attrs_need_update(inode, fattr)) {
		fattr->valid &= ~(NFS_ATTR_WCC_V4|NFS_ATTR_WCC);
		goto out_noforce;
	}
1036 1037 1038 1039 1040 1041 1042 1043 1044
	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));
1045
		fattr->pre_size = i_size_read(inode);
1046 1047
		fattr->valid |= NFS_ATTR_WCC;
	}
1048 1049 1050 1051
out_noforce:
	status = nfs_post_op_update_inode_locked(inode, fattr);
	spin_unlock(&inode->i_lock);
	return status;
1052 1053
}

L
Linus Torvalds 已提交
1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065
/*
 * 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.
 */
1066
static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
L
Linus Torvalds 已提交
1067
{
1068
	struct nfs_server *server;
L
Linus Torvalds 已提交
1069
	struct nfs_inode *nfsi = NFS_I(inode);
1070
	loff_t cur_isize, new_isize;
1071
	unsigned long invalid = 0;
1072
	unsigned long now = jiffies;
L
Linus Torvalds 已提交
1073 1074

	dfprintk(VFS, "NFS: %s(%s/%ld ct=%d info=0x%x)\n",
1075
			__func__, inode->i_sb->s_id, inode->i_ino,
L
Linus Torvalds 已提交
1076 1077
			atomic_read(&inode->i_count), fattr->valid);

1078 1079
	if (nfsi->fileid != fattr->fileid)
		goto out_fileid;
L
Linus Torvalds 已提交
1080 1081 1082 1083

	/*
	 * Make sure the inode's type hasn't changed.
	 */
1084
	if ((inode->i_mode & S_IFMT) != (fattr->mode & S_IFMT))
L
Linus Torvalds 已提交
1085 1086
		goto out_changed;

1087
	server = NFS_SERVER(inode);
T
Trond Myklebust 已提交
1088
	/* Update the fsid? */
1089 1090 1091
	if (S_ISDIR(inode->i_mode) &&
			!nfs_fsid_equal(&server->fsid, &fattr->fsid) &&
			!test_bit(NFS_INO_MOUNTPOINT, &nfsi->flags))
1092 1093
		server->fsid = fattr->fsid;

L
Linus Torvalds 已提交
1094 1095 1096
	/*
	 * Update the read time so we don't revalidate too often.
	 */
1097
	nfsi->read_cache_jiffies = fattr->time_start;
1098

1099 1100
	nfsi->cache_validity &= ~(NFS_INO_INVALID_ATTR | NFS_INO_INVALID_ATIME
			| NFS_INO_REVAL_PAGECACHE);
L
Linus Torvalds 已提交
1101

1102 1103 1104
	/* Do atomic weak cache consistency updates */
	nfs_wcc_update_inode(inode, fattr);

1105 1106 1107 1108 1109 1110 1111
	/* 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;
1112 1113
			if (S_ISDIR(inode->i_mode))
				nfs_force_lookup_revalidate(inode);
1114 1115
		}
		/* If ctime has changed we should definitely clear access+acl caches */
1116
		if (!timespec_equal(&inode->i_ctime, &fattr->ctime))
1117
			invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
1118 1119 1120 1121
	} 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;
1122 1123
		if (S_ISDIR(inode->i_mode))
			nfs_force_lookup_revalidate(inode);
1124 1125
	}

1126
	/* Check if our cached file size is stale */
L
Linus Torvalds 已提交
1127 1128
 	new_isize = nfs_size_to_loff_t(fattr->size);
	cur_isize = i_size_read(inode);
1129
	if (new_isize != cur_isize) {
1130 1131 1132
		/* Do we perhaps have any outstanding writes, or has
		 * the file grown beyond our last write? */
		if (nfsi->npages == 0 || new_isize > cur_isize) {
1133
			i_size_write(inode, new_isize);
L
Linus Torvalds 已提交
1134 1135
			invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_DATA;
		}
1136 1137
		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
1138 1139 1140
	}


1141 1142
	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
1143
	memcpy(&inode->i_atime, &fattr->atime, sizeof(inode->i_atime));
1144
	nfsi->change_attr = fattr->change_attr;
L
Linus Torvalds 已提交
1145 1146 1147 1148

	if ((inode->i_mode & S_IALLUGO) != (fattr->mode & S_IALLUGO) ||
	    inode->i_uid != fattr->uid ||
	    inode->i_gid != fattr->gid)
1149
		invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
L
Linus Torvalds 已提交
1150

1151 1152 1153
	if (inode->i_nlink != fattr->nlink)
		invalid |= NFS_INO_INVALID_ATTR;

L
Linus Torvalds 已提交
1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169
	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
1170
		nfs_inc_stats(inode, NFSIOS_ATTRINVALIDATE);
L
Linus Torvalds 已提交
1171
		nfsi->attrtimeo = NFS_MINATTRTIMEO(inode);
1172
		nfsi->attrtimeo_timestamp = now;
1173
		nfsi->attr_gencount = nfs_inc_attr_generation_counter();
1174
	} else {
1175
		if (!time_in_range_open(now, nfsi->attrtimeo_timestamp, nfsi->attrtimeo_timestamp + nfsi->attrtimeo)) {
1176 1177 1178 1179
			if ((nfsi->attrtimeo <<= 1) > NFS_MAXATTRTIMEO(inode))
				nfsi->attrtimeo = NFS_MAXATTRTIMEO(inode);
			nfsi->attrtimeo_timestamp = now;
		}
L
Linus Torvalds 已提交
1180
	}
1181
	invalid &= ~NFS_INO_INVALID_ATTR;
L
Linus Torvalds 已提交
1182 1183 1184 1185
	/* 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;
1186 1187
	if (!nfs_have_delegation(inode, FMODE_READ) ||
			(nfsi->cache_validity & NFS_INO_REVAL_FORCED))
1188
		nfsi->cache_validity |= invalid;
1189
	nfsi->cache_validity &= ~NFS_INO_REVAL_FORCED;
L
Linus Torvalds 已提交
1190 1191 1192 1193 1194 1195 1196

	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",
1197
			__func__, inode->i_ino, inode->i_mode, fattr->mode);
1198
 out_err:
L
Linus Torvalds 已提交
1199 1200 1201 1202 1203 1204 1205
	/*
	 * 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;
1206 1207 1208 1209

 out_fileid:
	printk(KERN_ERR "NFS: server %s error: fileid changed\n"
		"fsid %s: expected fileid 0x%Lx, got 0x%Lx\n",
1210
		NFS_SERVER(inode)->nfs_client->cl_hostname, inode->i_sb->s_id,
1211 1212
		(long long)nfsi->fileid, (long long)fattr->fileid);
	goto out_err;
L
Linus Torvalds 已提交
1213 1214
}

1215

L
Linus Torvalds 已提交
1216 1217 1218 1219 1220 1221 1222
#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
1223
void nfs4_clear_inode(struct inode *inode)
L
Linus Torvalds 已提交
1224 1225
{
	/* If we are holding a delegation, return it! */
1226
	nfs_inode_return_delegation_noreclaim(inode);
L
Linus Torvalds 已提交
1227 1228 1229 1230 1231
	/* First call standard NFS clear_inode() code */
	nfs_clear_inode(inode);
}
#endif

D
David Howells 已提交
1232
struct inode *nfs_alloc_inode(struct super_block *sb)
L
Linus Torvalds 已提交
1233 1234
{
	struct nfs_inode *nfsi;
1235
	nfsi = (struct nfs_inode *)kmem_cache_alloc(nfs_inode_cachep, GFP_KERNEL);
L
Linus Torvalds 已提交
1236 1237
	if (!nfsi)
		return NULL;
1238 1239
	nfsi->flags = 0UL;
	nfsi->cache_validity = 0UL;
1240 1241 1242 1243
#ifdef CONFIG_NFS_V3_ACL
	nfsi->acl_access = ERR_PTR(-EAGAIN);
	nfsi->acl_default = ERR_PTR(-EAGAIN);
#endif
1244 1245 1246
#ifdef CONFIG_NFS_V4
	nfsi->nfs4_acl = NULL;
#endif /* CONFIG_NFS_V4 */
L
Linus Torvalds 已提交
1247 1248 1249
	return &nfsi->vfs_inode;
}

D
David Howells 已提交
1250
void nfs_destroy_inode(struct inode *inode)
L
Linus Torvalds 已提交
1251 1252 1253 1254
{
	kmem_cache_free(nfs_inode_cachep, NFS_I(inode));
}

A
Andrew Morton 已提交
1255 1256 1257 1258 1259 1260 1261 1262 1263
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
1264

1265
static void init_once(void *foo)
L
Linus Torvalds 已提交
1266 1267 1268
{
	struct nfs_inode *nfsi = (struct nfs_inode *) foo;

C
Christoph Lameter 已提交
1269 1270 1271 1272 1273 1274 1275
	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
1276 1277 1278
	atomic_set(&nfsi->silly_count, 1);
	INIT_HLIST_HEAD(&nfsi->silly_list);
	init_waitqueue_head(&nfsi->waitqueue);
C
Christoph Lameter 已提交
1279
	nfs4_init_once(nfsi);
L
Linus Torvalds 已提交
1280
}
1281

D
David Howells 已提交
1282
static int __init nfs_init_inodecache(void)
L
Linus Torvalds 已提交
1283 1284 1285
{
	nfs_inode_cachep = kmem_cache_create("nfs_inode_cache",
					     sizeof(struct nfs_inode),
1286 1287
					     0, (SLAB_RECLAIM_ACCOUNT|
						SLAB_MEM_SPREAD),
1288
					     init_once);
L
Linus Torvalds 已提交
1289 1290 1291 1292 1293 1294
	if (nfs_inode_cachep == NULL)
		return -ENOMEM;

	return 0;
}

1295
static void nfs_destroy_inodecache(void)
L
Linus Torvalds 已提交
1296
{
1297
	kmem_cache_destroy(nfs_inode_cachep);
L
Linus Torvalds 已提交
1298 1299
}

T
Trond Myklebust 已提交
1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329
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
1330 1331 1332 1333 1334 1335 1336
/*
 * Initialize NFS
 */
static int __init init_nfs_fs(void)
{
	int err;

T
Trond Myklebust 已提交
1337 1338 1339 1340
	err = nfsiod_start();
	if (err)
		goto out6;

1341 1342 1343 1344
	err = nfs_fs_proc_init();
	if (err)
		goto out5;

L
Linus Torvalds 已提交
1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367
	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
1368
	if ((err = register_nfs_fs()) != 0)
L
Linus Torvalds 已提交
1369 1370 1371 1372 1373 1374 1375
		goto out;
	return 0;
out:
#ifdef CONFIG_PROC_FS
	rpc_proc_unregister("nfs");
#endif
	nfs_destroy_directcache();
1376 1377
out0:
	nfs_destroy_writepagecache();
L
Linus Torvalds 已提交
1378 1379 1380 1381 1382 1383 1384
out1:
	nfs_destroy_readpagecache();
out2:
	nfs_destroy_inodecache();
out3:
	nfs_destroy_nfspagecache();
out4:
1385 1386
	nfs_fs_proc_exit();
out5:
T
Trond Myklebust 已提交
1387 1388
	nfsiod_stop();
out6:
L
Linus Torvalds 已提交
1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401
	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
1402
	unregister_nfs_fs();
1403
	nfs_fs_proc_exit();
T
Trond Myklebust 已提交
1404
	nfsiod_stop();
L
Linus Torvalds 已提交
1405 1406 1407 1408 1409
}

/* Not quite true; I just maintain it */
MODULE_AUTHOR("Olaf Kirch <okir@monad.swb.de>");
MODULE_LICENSE("GPL");
1410
module_param(enable_ino64, bool, 0644);
L
Linus Torvalds 已提交
1411 1412 1413

module_init(init_nfs_fs)
module_exit(exit_nfs_fs)