vfs.c 47.6 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 18 19 20 21 22 23 24 25
#define MSNFS	/* HACK HACK */
/*
 * linux/fs/nfsd/vfs.c
 *
 * File operations used by nfsd. Some of these have been ripped from
 * other parts of the kernel because they weren't exported, others
 * are partial duplicates with added or changed functionality.
 *
 * Note that several functions dget() the dentry upon which they want
 * to act, most notably those that create directory entries. Response
 * dentry's are dput()'d if necessary in the release callback.
 * So if you notice code paths that apparently fail to dput() the
 * dentry, don't worry--they have been taken care of.
 *
 * Copyright (C) 1995-1999 Olaf Kirch <okir@monad.swb.de>
 * Zerocpy NFS support (C) 2002 Hirokazu Takahashi <taka@valinux.co.jp>
 */

#include <linux/string.h>
#include <linux/time.h>
#include <linux/errno.h>
#include <linux/fs.h>
#include <linux/file.h>
#include <linux/mount.h>
#include <linux/major.h>
26
#include <linux/splice.h>
L
Linus Torvalds 已提交
27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46
#include <linux/proc_fs.h>
#include <linux/stat.h>
#include <linux/fcntl.h>
#include <linux/net.h>
#include <linux/unistd.h>
#include <linux/slab.h>
#include <linux/pagemap.h>
#include <linux/in.h>
#include <linux/module.h>
#include <linux/namei.h>
#include <linux/vfs.h>
#include <linux/delay.h>
#include <linux/sunrpc/svc.h>
#include <linux/nfsd/nfsd.h>
#ifdef CONFIG_NFSD_V3
#include <linux/nfs3.h>
#include <linux/nfsd/xdr3.h>
#endif /* CONFIG_NFSD_V3 */
#include <linux/nfsd/nfsfh.h>
#include <linux/quotaops.h>
R
Robert Love 已提交
47
#include <linux/fsnotify.h>
L
Linus Torvalds 已提交
48 49 50
#include <linux/posix_acl.h>
#include <linux/posix_acl_xattr.h>
#include <linux/xattr.h>
51
#ifdef CONFIG_NFSD_V4
L
Linus Torvalds 已提交
52 53 54 55 56
#include <linux/nfs4.h>
#include <linux/nfs4_acl.h>
#include <linux/nfsd_idmap.h>
#include <linux/security.h>
#endif /* CONFIG_NFSD_V4 */
57
#include <linux/jhash.h>
L
Linus Torvalds 已提交
58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77

#include <asm/uaccess.h>

#define NFSDDBG_FACILITY		NFSDDBG_FILEOP


/*
 * This is a cache of readahead params that help us choose the proper
 * readahead strategy. Initially, we set all readahead parameters to 0
 * and let the VFS handle things.
 * If you increase the number of cached files very much, you'll need to
 * add a hash table here.
 */
struct raparms {
	struct raparms		*p_next;
	unsigned int		p_count;
	ino_t			p_ino;
	dev_t			p_dev;
	int			p_set;
	struct file_ra_state	p_ra;
78
	unsigned int		p_hindex;
L
Linus Torvalds 已提交
79 80
};

81 82 83 84 85
struct raparm_hbucket {
	struct raparms		*pb_head;
	spinlock_t		pb_lock;
} ____cacheline_aligned_in_smp;

L
Linus Torvalds 已提交
86
static struct raparms *		raparml;
87 88 89 90
#define RAPARM_HASH_BITS	4
#define RAPARM_HASH_SIZE	(1<<RAPARM_HASH_BITS)
#define RAPARM_HASH_MASK	(RAPARM_HASH_SIZE-1)
static struct raparm_hbucket	raparm_hash[RAPARM_HASH_SIZE];
L
Linus Torvalds 已提交
91 92 93 94

/* 
 * Called from nfsd_lookup and encode_dirent. Check if we have crossed 
 * a mount point.
95
 * Returns -EAGAIN or -ETIMEDOUT leaving *dpp and *expp unchanged,
L
Linus Torvalds 已提交
96 97 98 99 100 101 102 103 104 105
 *  or nfs_ok having possibly changed *dpp and *expp
 */
int
nfsd_cross_mnt(struct svc_rqst *rqstp, struct dentry **dpp, 
		        struct svc_export **expp)
{
	struct svc_export *exp = *expp, *exp2 = NULL;
	struct dentry *dentry = *dpp;
	struct vfsmount *mnt = mntget(exp->ex_mnt);
	struct dentry *mounts = dget(dentry);
106
	int err = 0;
L
Linus Torvalds 已提交
107 108 109

	while (follow_down(&mnt,&mounts)&&d_mountpoint(mounts));

110
	exp2 = rqst_exp_get_by_name(rqstp, mnt, mounts);
L
Linus Torvalds 已提交
111
	if (IS_ERR(exp2)) {
112 113
		if (PTR_ERR(exp2) != -ENOENT)
			err = PTR_ERR(exp2);
L
Linus Torvalds 已提交
114 115 116 117
		dput(mounts);
		mntput(mnt);
		goto out;
	}
118
	if ((exp->ex_flags & NFSEXP_CROSSMOUNT) || EX_NOHIDE(exp2)) {
L
Linus Torvalds 已提交
119 120 121 122 123 124
		/* successfully crossed mount point */
		exp_put(exp);
		*expp = exp2;
		dput(dentry);
		*dpp = mounts;
	} else {
125
		exp_put(exp2);
L
Linus Torvalds 已提交
126 127 128 129 130 131 132
		dput(mounts);
	}
	mntput(mnt);
out:
	return err;
}

133
__be32
134 135 136
nfsd_lookup_dentry(struct svc_rqst *rqstp, struct svc_fh *fhp,
		   const char *name, int len,
		   struct svc_export **exp_ret, struct dentry **dentry_ret)
L
Linus Torvalds 已提交
137 138 139 140
{
	struct svc_export	*exp;
	struct dentry		*dparent;
	struct dentry		*dentry;
141 142
	__be32			err;
	int			host_err;
L
Linus Torvalds 已提交
143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174

	dprintk("nfsd: nfsd_lookup(fh %s, %.*s)\n", SVCFH_fmt(fhp), len,name);

	/* Obtain dentry and export. */
	err = fh_verify(rqstp, fhp, S_IFDIR, MAY_EXEC);
	if (err)
		return err;

	dparent = fhp->fh_dentry;
	exp  = fhp->fh_export;
	exp_get(exp);

	/* Lookup the name, but don't follow links */
	if (isdotent(name, len)) {
		if (len==1)
			dentry = dget(dparent);
		else if (dparent != exp->ex_dentry) {
			dentry = dget_parent(dparent);
		} else  if (!EX_NOHIDE(exp))
			dentry = dget(dparent); /* .. == . just like at / */
		else {
			/* checking mountpoint crossing is very different when stepping up */
			struct svc_export *exp2 = NULL;
			struct dentry *dp;
			struct vfsmount *mnt = mntget(exp->ex_mnt);
			dentry = dget(dparent);
			while(dentry == mnt->mnt_root && follow_up(&mnt, &dentry))
				;
			dp = dget_parent(dentry);
			dput(dentry);
			dentry = dp;

175
			exp2 = rqst_exp_parent(rqstp, mnt, dentry);
176 177 178 179
			if (PTR_ERR(exp2) == -ENOENT) {
				dput(dentry);
				dentry = dget(dparent);
			} else if (IS_ERR(exp2)) {
180
				host_err = PTR_ERR(exp2);
L
Linus Torvalds 已提交
181 182 183 184 185 186 187 188 189 190 191 192
				dput(dentry);
				mntput(mnt);
				goto out_nfserr;
			} else {
				exp_put(exp);
				exp = exp2;
			}
			mntput(mnt);
		}
	} else {
		fh_lock(fhp);
		dentry = lookup_one_len(name, dparent, len);
193
		host_err = PTR_ERR(dentry);
L
Linus Torvalds 已提交
194 195 196 197 198 199
		if (IS_ERR(dentry))
			goto out_nfserr;
		/*
		 * check if we have crossed a mount point ...
		 */
		if (d_mountpoint(dentry)) {
200
			if ((host_err = nfsd_cross_mnt(rqstp, &dentry, &exp))) {
L
Linus Torvalds 已提交
201 202 203 204 205
				dput(dentry);
				goto out_nfserr;
			}
		}
	}
206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237
	*dentry_ret = dentry;
	*exp_ret = exp;
	return 0;

out_nfserr:
	exp_put(exp);
	return nfserrno(host_err);
}

/*
 * Look up one component of a pathname.
 * N.B. After this call _both_ fhp and resfh need an fh_put
 *
 * If the lookup would cross a mountpoint, and the mounted filesystem
 * is exported to the client with NFSEXP_NOHIDE, then the lookup is
 * accepted as it stands and the mounted directory is
 * returned. Otherwise the covered directory is returned.
 * NOTE: this mountpoint crossing is not supported properly by all
 *   clients and is explicitly disallowed for NFSv3
 *      NeilBrown <neilb@cse.unsw.edu.au>
 */
__be32
nfsd_lookup(struct svc_rqst *rqstp, struct svc_fh *fhp, const char *name,
					int len, struct svc_fh *resfh)
{
	struct svc_export	*exp;
	struct dentry		*dentry;
	__be32 err;

	err = nfsd_lookup_dentry(rqstp, fhp, name, len, &exp, &dentry);
	if (err)
		return err;
238 239 240
	err = check_nfsd_access(exp, rqstp);
	if (err)
		goto out;
L
Linus Torvalds 已提交
241 242 243 244 245 246 247
	/*
	 * Note: we compose the file handle now, but as the
	 * dentry may be negative, it may need to be updated.
	 */
	err = fh_compose(resfh, exp, dentry, fhp);
	if (!err && !dentry->d_inode)
		err = nfserr_noent;
248
out:
L
Linus Torvalds 已提交
249 250 251 252 253
	dput(dentry);
	exp_put(exp);
	return err;
}

254

L
Linus Torvalds 已提交
255 256 257 258
/*
 * Set various file attributes.
 * N.B. After this call fhp needs an fh_put
 */
259
__be32
L
Linus Torvalds 已提交
260 261 262 263 264 265 266 267
nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp, struct iattr *iap,
	     int check_guard, time_t guardtime)
{
	struct dentry	*dentry;
	struct inode	*inode;
	int		accmode = MAY_SATTR;
	int		ftype = 0;
	int		imode;
268 269
	__be32		err;
	int		host_err;
L
Linus Torvalds 已提交
270 271 272 273 274 275 276 277 278
	int		size_change = 0;

	if (iap->ia_valid & (ATTR_ATIME | ATTR_MTIME | ATTR_SIZE))
		accmode |= MAY_WRITE|MAY_OWNER_OVERRIDE;
	if (iap->ia_valid & ATTR_SIZE)
		ftype = S_IFREG;

	/* Get inode */
	err = fh_verify(rqstp, fhp, ftype, accmode);
279
	if (err)
L
Linus Torvalds 已提交
280 281 282 283 284
		goto out;

	dentry = fhp->fh_dentry;
	inode = dentry->d_inode;

285 286 287 288 289 290 291
	/* Ignore any mode updates on symlinks */
	if (S_ISLNK(inode->i_mode))
		iap->ia_valid &= ~ATTR_MODE;

	if (!iap->ia_valid)
		goto out;

292 293
	/*
	 * NFSv2 does not differentiate between "set-[ac]time-to-now"
L
Linus Torvalds 已提交
294 295 296 297 298 299 300 301 302 303 304 305
	 * which only requires access, and "set-[ac]time-to-X" which
	 * requires ownership.
	 * So if it looks like it might be "set both to the same time which
	 * is close to now", and if inode_change_ok fails, then we
	 * convert to "set to now" instead of "set to explicit time"
	 *
	 * We only call inode_change_ok as the last test as technically
	 * it is not an interface that we should be using.  It is only
	 * valid if the filesystem does not define it's own i_op->setattr.
	 */
#define BOTH_TIME_SET (ATTR_ATIME_SET | ATTR_MTIME_SET)
#define	MAX_TOUCH_TIME_ERROR (30*60)
306 307 308 309 310 311 312 313
	if ((iap->ia_valid & BOTH_TIME_SET) == BOTH_TIME_SET &&
	    iap->ia_mtime.tv_sec == iap->ia_atime.tv_sec) {
		/*
		 * Looks probable.
		 *
		 * Now just make sure time is in the right ballpark.
		 * Solaris, at least, doesn't seem to care what the time
		 * request is.  We require it be within 30 minutes of now.
L
Linus Torvalds 已提交
314
		 */
315 316 317 318 319 320 321 322 323 324 325 326
		time_t delta = iap->ia_atime.tv_sec - get_seconds();
		if (delta < 0)
			delta = -delta;
		if (delta < MAX_TOUCH_TIME_ERROR &&
		    inode_change_ok(inode, iap) != 0) {
			/*
			 * Turn off ATTR_[AM]TIME_SET but leave ATTR_[AM]TIME.
			 * This will cause notify_change to set these times
			 * to "now"
			 */
			iap->ia_valid &= ~BOTH_TIME_SET;
		}
L
Linus Torvalds 已提交
327 328
	}
	    
329 330 331 332
	/*
	 * The size case is special.
	 * It changes the file as well as the attributes.
	 */
L
Linus Torvalds 已提交
333 334
	if (iap->ia_valid & ATTR_SIZE) {
		if (iap->ia_size < inode->i_size) {
335
			err = nfsd_permission(rqstp, fhp->fh_export, dentry, MAY_TRUNC|MAY_OWNER_OVERRIDE);
L
Linus Torvalds 已提交
336 337 338 339 340 341 342 343
			if (err)
				goto out;
		}

		/*
		 * If we are changing the size of the file, then
		 * we need to break all leases.
		 */
344 345 346 347
		host_err = break_lease(inode, FMODE_WRITE | O_NONBLOCK);
		if (host_err == -EWOULDBLOCK)
			host_err = -ETIMEDOUT;
		if (host_err) /* ENOMEM or EWOULDBLOCK */
L
Linus Torvalds 已提交
348 349
			goto out_nfserr;

350 351
		host_err = get_write_access(inode);
		if (host_err)
L
Linus Torvalds 已提交
352 353 354
			goto out_nfserr;

		size_change = 1;
355 356
		host_err = locks_verify_truncate(inode, NULL, iap->ia_size);
		if (host_err) {
L
Linus Torvalds 已提交
357 358 359 360 361 362 363 364 365 366 367 368 369 370
			put_write_access(inode);
			goto out_nfserr;
		}
		DQUOT_INIT(inode);
	}

	imode = inode->i_mode;
	if (iap->ia_valid & ATTR_MODE) {
		iap->ia_mode &= S_IALLUGO;
		imode = iap->ia_mode |= (imode & ~S_IALLUGO);
	}

	/* Revoke setuid/setgid bit on chown/chgrp */
	if ((iap->ia_valid & ATTR_UID) && iap->ia_uid != inode->i_uid)
371
		iap->ia_valid |= ATTR_KILL_SUID | ATTR_KILL_PRIV;
L
Linus Torvalds 已提交
372 373 374 375 376 377 378 379 380 381
	if ((iap->ia_valid & ATTR_GID) && iap->ia_gid != inode->i_gid)
		iap->ia_valid |= ATTR_KILL_SGID;

	/* Change the attributes. */

	iap->ia_valid |= ATTR_CTIME;

	err = nfserr_notsync;
	if (!check_guard || guardtime == inode->i_ctime.tv_sec) {
		fh_lock(fhp);
382 383
		host_err = notify_change(dentry, iap);
		err = nfserrno(host_err);
L
Linus Torvalds 已提交
384 385 386 387 388 389 390 391 392 393 394
		fh_unlock(fhp);
	}
	if (size_change)
		put_write_access(inode);
	if (!err)
		if (EX_ISSYNC(fhp->fh_export))
			write_inode_now(inode, 1);
out:
	return err;

out_nfserr:
395
	err = nfserrno(host_err);
L
Linus Torvalds 已提交
396 397 398
	goto out;
}

399 400 401 402 403 404 405 406 407 408
#if defined(CONFIG_NFSD_V2_ACL) || \
    defined(CONFIG_NFSD_V3_ACL) || \
    defined(CONFIG_NFSD_V4)
static ssize_t nfsd_getxattr(struct dentry *dentry, char *key, void **buf)
{
	ssize_t buflen;

	buflen = vfs_getxattr(dentry, key, NULL, 0);
	if (buflen <= 0)
		return buflen;
L
Linus Torvalds 已提交
409

410 411 412 413
	*buf = kmalloc(buflen, GFP_KERNEL);
	if (!*buf)
		return -ENOMEM;

414
	return vfs_getxattr(dentry, key, *buf, buflen);
415 416 417 418
}
#endif

#if defined(CONFIG_NFSD_V4)
L
Linus Torvalds 已提交
419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438
static int
set_nfsv4_acl_one(struct dentry *dentry, struct posix_acl *pacl, char *key)
{
	int len;
	size_t buflen;
	char *buf = NULL;
	int error = 0;

	buflen = posix_acl_xattr_size(pacl->a_count);
	buf = kmalloc(buflen, GFP_KERNEL);
	error = -ENOMEM;
	if (buf == NULL)
		goto out;

	len = posix_acl_to_xattr(pacl, buf, buflen);
	if (len < 0) {
		error = len;
		goto out;
	}

439
	error = vfs_setxattr(dentry, key, buf, len, 0);
L
Linus Torvalds 已提交
440 441 442 443 444
out:
	kfree(buf);
	return error;
}

445
__be32
L
Linus Torvalds 已提交
446 447 448
nfsd4_set_nfs4_acl(struct svc_rqst *rqstp, struct svc_fh *fhp,
    struct nfs4_acl *acl)
{
449 450
	__be32 error;
	int host_error;
L
Linus Torvalds 已提交
451 452 453 454 455 456 457 458
	struct dentry *dentry;
	struct inode *inode;
	struct posix_acl *pacl = NULL, *dpacl = NULL;
	unsigned int flags = 0;

	/* Get inode */
	error = fh_verify(rqstp, fhp, 0 /* S_IFREG */, MAY_SATTR);
	if (error)
459
		return error;
L
Linus Torvalds 已提交
460 461 462 463 464 465

	dentry = fhp->fh_dentry;
	inode = dentry->d_inode;
	if (S_ISDIR(inode->i_mode))
		flags = NFS4_ACL_DIR;

466 467
	host_error = nfs4_acl_nfsv4_to_posix(acl, &pacl, &dpacl, flags);
	if (host_error == -EINVAL) {
468
		return nfserr_attrnotsupp;
469
	} else if (host_error < 0)
L
Linus Torvalds 已提交
470 471
		goto out_nfserr;

472 473
	host_error = set_nfsv4_acl_one(dentry, pacl, POSIX_ACL_XATTR_ACCESS);
	if (host_error < 0)
474
		goto out_release;
L
Linus Torvalds 已提交
475

476
	if (S_ISDIR(inode->i_mode))
477
		host_error = set_nfsv4_acl_one(dentry, dpacl, POSIX_ACL_XATTR_DEFAULT);
L
Linus Torvalds 已提交
478

479
out_release:
L
Linus Torvalds 已提交
480 481 482
	posix_acl_release(pacl);
	posix_acl_release(dpacl);
out_nfserr:
483
	if (host_error == -EOPNOTSUPP)
484
		return nfserr_attrnotsupp;
485
	else
486
		return nfserrno(host_error);
L
Linus Torvalds 已提交
487 488 489 490 491
}

static struct posix_acl *
_get_posix_acl(struct dentry *dentry, char *key)
{
492
	void *buf = NULL;
L
Linus Torvalds 已提交
493
	struct posix_acl *pacl = NULL;
494
	int buflen;
L
Linus Torvalds 已提交
495

496 497 498 499 500
	buflen = nfsd_getxattr(dentry, key, &buf);
	if (!buflen)
		buflen = -ENODATA;
	if (buflen <= 0)
		return ERR_PTR(buflen);
L
Linus Torvalds 已提交
501 502 503 504 505 506 507 508 509 510 511 512 513 514

	pacl = posix_acl_from_xattr(buf, buflen);
	kfree(buf);
	return pacl;
}

int
nfsd4_get_nfs4_acl(struct svc_rqst *rqstp, struct dentry *dentry, struct nfs4_acl **acl)
{
	struct inode *inode = dentry->d_inode;
	int error = 0;
	struct posix_acl *pacl = NULL, *dpacl = NULL;
	unsigned int flags = 0;

515
	pacl = _get_posix_acl(dentry, POSIX_ACL_XATTR_ACCESS);
L
Linus Torvalds 已提交
516 517 518 519 520 521 522 523 524
	if (IS_ERR(pacl) && PTR_ERR(pacl) == -ENODATA)
		pacl = posix_acl_from_mode(inode->i_mode, GFP_KERNEL);
	if (IS_ERR(pacl)) {
		error = PTR_ERR(pacl);
		pacl = NULL;
		goto out;
	}

	if (S_ISDIR(inode->i_mode)) {
525
		dpacl = _get_posix_acl(dentry, POSIX_ACL_XATTR_DEFAULT);
L
Linus Torvalds 已提交
526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591
		if (IS_ERR(dpacl) && PTR_ERR(dpacl) == -ENODATA)
			dpacl = NULL;
		else if (IS_ERR(dpacl)) {
			error = PTR_ERR(dpacl);
			dpacl = NULL;
			goto out;
		}
		flags = NFS4_ACL_DIR;
	}

	*acl = nfs4_acl_posix_to_nfsv4(pacl, dpacl, flags);
	if (IS_ERR(*acl)) {
		error = PTR_ERR(*acl);
		*acl = NULL;
	}
 out:
	posix_acl_release(pacl);
	posix_acl_release(dpacl);
	return error;
}

#endif /* defined(CONFIG_NFS_V4) */

#ifdef CONFIG_NFSD_V3
/*
 * Check server access rights to a file system object
 */
struct accessmap {
	u32		access;
	int		how;
};
static struct accessmap	nfs3_regaccess[] = {
    {	NFS3_ACCESS_READ,	MAY_READ			},
    {	NFS3_ACCESS_EXECUTE,	MAY_EXEC			},
    {	NFS3_ACCESS_MODIFY,	MAY_WRITE|MAY_TRUNC		},
    {	NFS3_ACCESS_EXTEND,	MAY_WRITE			},

    {	0,			0				}
};

static struct accessmap	nfs3_diraccess[] = {
    {	NFS3_ACCESS_READ,	MAY_READ			},
    {	NFS3_ACCESS_LOOKUP,	MAY_EXEC			},
    {	NFS3_ACCESS_MODIFY,	MAY_EXEC|MAY_WRITE|MAY_TRUNC	},
    {	NFS3_ACCESS_EXTEND,	MAY_EXEC|MAY_WRITE		},
    {	NFS3_ACCESS_DELETE,	MAY_REMOVE			},

    {	0,			0				}
};

static struct accessmap	nfs3_anyaccess[] = {
	/* Some clients - Solaris 2.6 at least, make an access call
	 * to the server to check for access for things like /dev/null
	 * (which really, the server doesn't care about).  So
	 * We provide simple access checking for them, looking
	 * mainly at mode bits, and we make sure to ignore read-only
	 * filesystem checks
	 */
    {	NFS3_ACCESS_READ,	MAY_READ			},
    {	NFS3_ACCESS_EXECUTE,	MAY_EXEC			},
    {	NFS3_ACCESS_MODIFY,	MAY_WRITE|MAY_LOCAL_ACCESS	},
    {	NFS3_ACCESS_EXTEND,	MAY_WRITE|MAY_LOCAL_ACCESS	},

    {	0,			0				}
};

592
__be32
L
Linus Torvalds 已提交
593 594 595 596 597 598
nfsd_access(struct svc_rqst *rqstp, struct svc_fh *fhp, u32 *access, u32 *supported)
{
	struct accessmap	*map;
	struct svc_export	*export;
	struct dentry		*dentry;
	u32			query, result = 0, sresult = 0;
599
	__be32			error;
L
Linus Torvalds 已提交
600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618

	error = fh_verify(rqstp, fhp, 0, MAY_NOP);
	if (error)
		goto out;

	export = fhp->fh_export;
	dentry = fhp->fh_dentry;

	if (S_ISREG(dentry->d_inode->i_mode))
		map = nfs3_regaccess;
	else if (S_ISDIR(dentry->d_inode->i_mode))
		map = nfs3_diraccess;
	else
		map = nfs3_anyaccess;


	query = *access;
	for  (; map->access; map++) {
		if (map->access & query) {
619
			__be32 err2;
L
Linus Torvalds 已提交
620 621 622

			sresult |= map->access;

623
			err2 = nfsd_permission(rqstp, export, dentry, map->how);
L
Linus Torvalds 已提交
624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657
			switch (err2) {
			case nfs_ok:
				result |= map->access;
				break;
				
			/* the following error codes just mean the access was not allowed,
			 * rather than an error occurred */
			case nfserr_rofs:
			case nfserr_acces:
			case nfserr_perm:
				/* simply don't "or" in the access bit. */
				break;
			default:
				error = err2;
				goto out;
			}
		}
	}
	*access = result;
	if (supported)
		*supported = sresult;

 out:
	return error;
}
#endif /* CONFIG_NFSD_V3 */



/*
 * Open an existing file or directory.
 * The access argument indicates the type of open (read/write/lock)
 * N.B. After this call fhp needs an fh_put
 */
658
__be32
L
Linus Torvalds 已提交
659 660 661 662 663
nfsd_open(struct svc_rqst *rqstp, struct svc_fh *fhp, int type,
			int access, struct file **filp)
{
	struct dentry	*dentry;
	struct inode	*inode;
664 665 666
	int		flags = O_RDONLY|O_LARGEFILE;
	__be32		err;
	int		host_err;
L
Linus Torvalds 已提交
667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685

	/*
	 * If we get here, then the client has already done an "open",
	 * and (hopefully) checked permission - so allow OWNER_OVERRIDE
	 * in case a chmod has now revoked permission.
	 */
	err = fh_verify(rqstp, fhp, type, access | MAY_OWNER_OVERRIDE);
	if (err)
		goto out;

	dentry = fhp->fh_dentry;
	inode = dentry->d_inode;

	/* Disallow write access to files with the append-only bit set
	 * or any access when mandatory locking enabled
	 */
	err = nfserr_perm;
	if (IS_APPEND(inode) && (access & MAY_WRITE))
		goto out;
J
J. Bruce Fields 已提交
686 687 688 689 690 691
	/*
	 * We must ignore files (but only files) which might have mandatory
	 * locks on them because there is no way to know if the accesser has
	 * the lock.
	 */
	if (S_ISREG((inode)->i_mode) && mandatory_lock(inode))
L
Linus Torvalds 已提交
692 693 694 695 696 697 698 699 700
		goto out;

	if (!inode->i_fop)
		goto out;

	/*
	 * Check to see if there are any leases on this file.
	 * This may block while leases are broken.
	 */
701 702 703 704
	host_err = break_lease(inode, O_NONBLOCK | ((access & MAY_WRITE) ? FMODE_WRITE : 0));
	if (host_err == -EWOULDBLOCK)
		host_err = -ETIMEDOUT;
	if (host_err) /* NOMEM or WOULDBLOCK */
L
Linus Torvalds 已提交
705 706 707
		goto out_nfserr;

	if (access & MAY_WRITE) {
708 709 710 711
		if (access & MAY_READ)
			flags = O_RDWR|O_LARGEFILE;
		else
			flags = O_WRONLY|O_LARGEFILE;
L
Linus Torvalds 已提交
712 713 714 715 716

		DQUOT_INIT(inode);
	}
	*filp = dentry_open(dget(dentry), mntget(fhp->fh_export->ex_mnt), flags);
	if (IS_ERR(*filp))
717
		host_err = PTR_ERR(*filp);
L
Linus Torvalds 已提交
718
out_nfserr:
719
	err = nfserrno(host_err);
L
Linus Torvalds 已提交
720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737
out:
	return err;
}

/*
 * Close a file.
 */
void
nfsd_close(struct file *filp)
{
	fput(filp);
}

/*
 * Sync a file
 * As this calls fsync (not fdatasync) there is no need for a write_inode
 * after it.
 */
738
static inline int nfsd_dosync(struct file *filp, struct dentry *dp,
739
			      const struct file_operations *fop)
L
Linus Torvalds 已提交
740 741 742
{
	struct inode *inode = dp->d_inode;
	int (*fsync) (struct file *, struct dentry *, int);
743
	int err;
L
Linus Torvalds 已提交
744

745 746 747 748 749
	err = filemap_fdatawrite(inode->i_mapping);
	if (err == 0 && fop && (fsync = fop->fsync))
		err = fsync(filp, dp, 0);
	if (err == 0)
		err = filemap_fdatawait(inode->i_mapping);
750

751
	return err;
L
Linus Torvalds 已提交
752 753 754
}
	

755
static int
L
Linus Torvalds 已提交
756 757
nfsd_sync(struct file *filp)
{
758
        int err;
759 760
	struct inode *inode = filp->f_path.dentry->d_inode;
	dprintk("nfsd: sync file %s\n", filp->f_path.dentry->d_name.name);
761
	mutex_lock(&inode->i_mutex);
762
	err=nfsd_dosync(filp, filp->f_path.dentry, filp->f_op);
763
	mutex_unlock(&inode->i_mutex);
764 765

	return err;
L
Linus Torvalds 已提交
766 767
}

768
int
L
Linus Torvalds 已提交
769 770
nfsd_sync_dir(struct dentry *dp)
{
771
	return nfsd_dosync(NULL, dp, dp->d_inode->i_fop);
L
Linus Torvalds 已提交
772 773 774 775 776 777 778 779 780 781 782 783
}

/*
 * Obtain the readahead parameters for the file
 * specified by (dev, ino).
 */

static inline struct raparms *
nfsd_get_raparms(dev_t dev, ino_t ino)
{
	struct raparms	*ra, **rap, **frap = NULL;
	int depth = 0;
784 785 786 787 788
	unsigned int hash;
	struct raparm_hbucket *rab;

	hash = jhash_2words(dev, ino, 0xfeedbeef) & RAPARM_HASH_MASK;
	rab = &raparm_hash[hash];
L
Linus Torvalds 已提交
789

790 791
	spin_lock(&rab->pb_lock);
	for (rap = &rab->pb_head; (ra = *rap); rap = &ra->p_next) {
L
Linus Torvalds 已提交
792 793 794 795 796 797 798 799
		if (ra->p_ino == ino && ra->p_dev == dev)
			goto found;
		depth++;
		if (ra->p_count == 0)
			frap = rap;
	}
	depth = nfsdstats.ra_size*11/10;
	if (!frap) {	
800
		spin_unlock(&rab->pb_lock);
L
Linus Torvalds 已提交
801 802 803 804 805 806 807
		return NULL;
	}
	rap = frap;
	ra = *frap;
	ra->p_dev = dev;
	ra->p_ino = ino;
	ra->p_set = 0;
808
	ra->p_hindex = hash;
L
Linus Torvalds 已提交
809
found:
810
	if (rap != &rab->pb_head) {
L
Linus Torvalds 已提交
811
		*rap = ra->p_next;
812 813
		ra->p_next   = rab->pb_head;
		rab->pb_head = ra;
L
Linus Torvalds 已提交
814 815 816
	}
	ra->p_count++;
	nfsdstats.ra_depth[depth*10/nfsdstats.ra_size]++;
817
	spin_unlock(&rab->pb_lock);
L
Linus Torvalds 已提交
818 819 820 821
	return ra;
}

/*
822 823 824
 * Grab and keep cached pages associated with a file in the svc_rqst
 * so that they can be passed to the network sendmsg/sendpage routines
 * directly. They will be released after the sending has completed.
L
Linus Torvalds 已提交
825 826
 */
static int
827 828
nfsd_splice_actor(struct pipe_inode_info *pipe, struct pipe_buffer *buf,
		  struct splice_desc *sd)
L
Linus Torvalds 已提交
829
{
830
	struct svc_rqst *rqstp = sd->u.data;
831
	struct page **pp = rqstp->rq_respages + rqstp->rq_resused;
832 833 834
	struct page *page = buf->page;
	size_t size;
	int ret;
L
Linus Torvalds 已提交
835

836
	ret = buf->ops->confirm(pipe, buf);
837 838 839 840
	if (unlikely(ret))
		return ret;

	size = sd->len;
L
Linus Torvalds 已提交
841 842 843

	if (rqstp->rq_res.page_len == 0) {
		get_page(page);
844 845 846
		put_page(*pp);
		*pp = page;
		rqstp->rq_resused++;
847
		rqstp->rq_res.page_base = buf->offset;
L
Linus Torvalds 已提交
848
		rqstp->rq_res.page_len = size;
849
	} else if (page != pp[-1]) {
L
Linus Torvalds 已提交
850
		get_page(page);
851 852
		if (*pp)
			put_page(*pp);
853 854
		*pp = page;
		rqstp->rq_resused++;
L
Linus Torvalds 已提交
855
		rqstp->rq_res.page_len += size;
856
	} else
L
Linus Torvalds 已提交
857 858 859 860 861
		rqstp->rq_res.page_len += size;

	return size;
}

862 863 864 865 866 867
static int nfsd_direct_splice_actor(struct pipe_inode_info *pipe,
				    struct splice_desc *sd)
{
	return __splice_from_pipe(pipe, sd, nfsd_splice_actor);
}

868 869 870 871 872 873 874 875 876
static inline int svc_msnfs(struct svc_fh *ffhp)
{
#ifdef MSNFS
	return (ffhp->fh_export->ex_flags & NFSEXP_MSNFS);
#else
	return 0;
#endif
}

877
static __be32
L
Linus Torvalds 已提交
878 879 880 881 882 883
nfsd_vfs_read(struct svc_rqst *rqstp, struct svc_fh *fhp, struct file *file,
              loff_t offset, struct kvec *vec, int vlen, unsigned long *count)
{
	struct inode *inode;
	struct raparms	*ra;
	mm_segment_t	oldfs;
884 885
	__be32		err;
	int		host_err;
L
Linus Torvalds 已提交
886 887

	err = nfserr_perm;
888
	inode = file->f_path.dentry->d_inode;
889 890

	if (svc_msnfs(fhp) && !lock_may_read(inode, offset, *count))
L
Linus Torvalds 已提交
891 892 893 894 895 896 897 898
		goto out;

	/* Get readahead parameters */
	ra = nfsd_get_raparms(inode->i_sb->s_dev, inode->i_ino);

	if (ra && ra->p_set)
		file->f_ra = ra->p_ra;

899 900 901 902 903 904 905 906
	if (file->f_op->splice_read && rqstp->rq_splice_ok) {
		struct splice_desc sd = {
			.len		= 0,
			.total_len	= *count,
			.pos		= offset,
			.u.data		= rqstp,
		};

907
		rqstp->rq_resused = 1;
908
		host_err = splice_direct_to_actor(file, &sd, nfsd_direct_splice_actor);
L
Linus Torvalds 已提交
909 910 911
	} else {
		oldfs = get_fs();
		set_fs(KERNEL_DS);
912
		host_err = vfs_readv(file, (struct iovec __user *)vec, vlen, &offset);
L
Linus Torvalds 已提交
913 914 915 916 917
		set_fs(oldfs);
	}

	/* Write back readahead params */
	if (ra) {
918 919
		struct raparm_hbucket *rab = &raparm_hash[ra->p_hindex];
		spin_lock(&rab->pb_lock);
L
Linus Torvalds 已提交
920 921 922
		ra->p_ra = file->f_ra;
		ra->p_set = 1;
		ra->p_count--;
923
		spin_unlock(&rab->pb_lock);
L
Linus Torvalds 已提交
924 925
	}

926 927 928
	if (host_err >= 0) {
		nfsdstats.io_read += host_err;
		*count = host_err;
L
Linus Torvalds 已提交
929
		err = 0;
930
		fsnotify_access(file->f_path.dentry);
L
Linus Torvalds 已提交
931
	} else 
932
		err = nfserrno(host_err);
L
Linus Torvalds 已提交
933 934 935 936
out:
	return err;
}

937 938 939
static void kill_suid(struct dentry *dentry)
{
	struct iattr	ia;
940
	ia.ia_valid = ATTR_KILL_SUID | ATTR_KILL_SGID | ATTR_KILL_PRIV;
941

942
	mutex_lock(&dentry->d_inode->i_mutex);
943
	notify_change(dentry, &ia);
944
	mutex_unlock(&dentry->d_inode->i_mutex);
945 946
}

947
static __be32
L
Linus Torvalds 已提交
948 949 950 951 952 953 954 955
nfsd_vfs_write(struct svc_rqst *rqstp, struct svc_fh *fhp, struct file *file,
				loff_t offset, struct kvec *vec, int vlen,
	   			unsigned long cnt, int *stablep)
{
	struct svc_export	*exp;
	struct dentry		*dentry;
	struct inode		*inode;
	mm_segment_t		oldfs;
956 957
	__be32			err = 0;
	int			host_err;
L
Linus Torvalds 已提交
958 959
	int			stable = *stablep;

960
#ifdef MSNFS
L
Linus Torvalds 已提交
961 962 963
	err = nfserr_perm;

	if ((fhp->fh_export->ex_flags & NFSEXP_MSNFS) &&
964
		(!lock_may_write(file->f_path.dentry->d_inode, offset, cnt)))
L
Linus Torvalds 已提交
965 966 967
		goto out;
#endif

968
	dentry = file->f_path.dentry;
L
Linus Torvalds 已提交
969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992
	inode = dentry->d_inode;
	exp   = fhp->fh_export;

	/*
	 * Request sync writes if
	 *  -	the sync export option has been set, or
	 *  -	the client requested O_SYNC behavior (NFSv3 feature).
	 *  -   The file system doesn't support fsync().
	 * When gathered writes have been configured for this volume,
	 * flushing the data to disk is handled separately below.
	 */

	if (file->f_op->fsync == 0) {/* COMMIT3 cannot work */
	       stable = 2;
	       *stablep = 2; /* FILE_SYNC */
	}

	if (!EX_ISSYNC(exp))
		stable = 0;
	if (stable && !EX_WGATHER(exp))
		file->f_flags |= O_SYNC;

	/* Write the data. */
	oldfs = get_fs(); set_fs(KERNEL_DS);
993
	host_err = vfs_writev(file, (struct iovec __user *)vec, vlen, &offset);
L
Linus Torvalds 已提交
994
	set_fs(oldfs);
995
	if (host_err >= 0) {
L
Linus Torvalds 已提交
996
		nfsdstats.io_write += cnt;
997
		fsnotify_modify(file->f_path.dentry);
L
Linus Torvalds 已提交
998 999 1000
	}

	/* clear setuid/setgid flag after write */
1001
	if (host_err >= 0 && (inode->i_mode & (S_ISUID | S_ISGID)))
1002
		kill_suid(dentry);
L
Linus Torvalds 已提交
1003

1004
	if (host_err >= 0 && stable) {
L
Linus Torvalds 已提交
1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029
		static ino_t	last_ino;
		static dev_t	last_dev;

		/*
		 * Gathered writes: If another process is currently
		 * writing to the file, there's a high chance
		 * this is another nfsd (triggered by a bulk write
		 * from a client's biod). Rather than syncing the
		 * file with each write request, we sleep for 10 msec.
		 *
		 * I don't know if this roughly approximates
		 * C. Juszak's idea of gathered writes, but it's a
		 * nice and simple solution (IMHO), and it seems to
		 * work:-)
		 */
		if (EX_WGATHER(exp)) {
			if (atomic_read(&inode->i_writecount) > 1
			    || (last_ino == inode->i_ino && last_dev == inode->i_sb->s_dev)) {
				dprintk("nfsd: write defer %d\n", current->pid);
				msleep(10);
				dprintk("nfsd: write resume %d\n", current->pid);
			}

			if (inode->i_state & I_DIRTY) {
				dprintk("nfsd: write sync %d\n", current->pid);
1030
				host_err=nfsd_sync(file);
L
Linus Torvalds 已提交
1031 1032 1033 1034 1035 1036 1037 1038 1039
			}
#if 0
			wake_up(&inode->i_wait);
#endif
		}
		last_ino = inode->i_ino;
		last_dev = inode->i_sb->s_dev;
	}

1040 1041
	dprintk("nfsd: write complete host_err=%d\n", host_err);
	if (host_err >= 0)
L
Linus Torvalds 已提交
1042 1043
		err = 0;
	else 
1044
		err = nfserrno(host_err);
L
Linus Torvalds 已提交
1045 1046 1047 1048 1049 1050 1051 1052 1053
out:
	return err;
}

/*
 * Read data from a file. count must contain the requested read count
 * on entry. On return, *count contains the number of bytes actually read.
 * N.B. After this call fhp needs an fh_put
 */
1054
__be32
L
Linus Torvalds 已提交
1055 1056 1057 1058
nfsd_read(struct svc_rqst *rqstp, struct svc_fh *fhp, struct file *file,
		loff_t offset, struct kvec *vec, int vlen,
		unsigned long *count)
{
1059
	__be32		err;
L
Linus Torvalds 已提交
1060 1061

	if (file) {
1062
		err = nfsd_permission(rqstp, fhp->fh_export, fhp->fh_dentry,
L
Linus Torvalds 已提交
1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082
				MAY_READ|MAY_OWNER_OVERRIDE);
		if (err)
			goto out;
		err = nfsd_vfs_read(rqstp, fhp, file, offset, vec, vlen, count);
	} else {
		err = nfsd_open(rqstp, fhp, S_IFREG, MAY_READ, &file);
		if (err)
			goto out;
		err = nfsd_vfs_read(rqstp, fhp, file, offset, vec, vlen, count);
		nfsd_close(file);
	}
out:
	return err;
}

/*
 * Write data to a file.
 * The stable flag requests synchronous writes.
 * N.B. After this call fhp needs an fh_put
 */
1083
__be32
L
Linus Torvalds 已提交
1084 1085 1086 1087
nfsd_write(struct svc_rqst *rqstp, struct svc_fh *fhp, struct file *file,
		loff_t offset, struct kvec *vec, int vlen, unsigned long cnt,
		int *stablep)
{
1088
	__be32			err = 0;
L
Linus Torvalds 已提交
1089 1090

	if (file) {
1091
		err = nfsd_permission(rqstp, fhp->fh_export, fhp->fh_dentry,
L
Linus Torvalds 已提交
1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119
				MAY_WRITE|MAY_OWNER_OVERRIDE);
		if (err)
			goto out;
		err = nfsd_vfs_write(rqstp, fhp, file, offset, vec, vlen, cnt,
				stablep);
	} else {
		err = nfsd_open(rqstp, fhp, S_IFREG, MAY_WRITE, &file);
		if (err)
			goto out;

		if (cnt)
			err = nfsd_vfs_write(rqstp, fhp, file, offset, vec, vlen,
					     cnt, stablep);
		nfsd_close(file);
	}
out:
	return err;
}

#ifdef CONFIG_NFSD_V3
/*
 * Commit all pending writes to stable storage.
 * Strictly speaking, we could sync just the indicated file region here,
 * but there's currently no way we can ask the VFS to do so.
 *
 * Unfortunately we cannot lock the file to make sure we return full WCC
 * data to the client, as locking happens lower down in the filesystem.
 */
1120
__be32
L
Linus Torvalds 已提交
1121 1122 1123 1124
nfsd_commit(struct svc_rqst *rqstp, struct svc_fh *fhp,
               loff_t offset, unsigned long count)
{
	struct file	*file;
1125
	__be32		err;
L
Linus Torvalds 已提交
1126 1127 1128 1129 1130 1131 1132 1133

	if ((u64)count > ~(u64)offset)
		return nfserr_inval;

	if ((err = nfsd_open(rqstp, fhp, S_IFREG, MAY_WRITE, &file)) != 0)
		return err;
	if (EX_ISSYNC(fhp->fh_export)) {
		if (file->f_op && file->f_op->fsync) {
1134
			err = nfserrno(nfsd_sync(file));
L
Linus Torvalds 已提交
1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152
		} else {
			err = nfserr_notsupp;
		}
	}

	nfsd_close(file);
	return err;
}
#endif /* CONFIG_NFSD_V3 */

/*
 * Create a file (regular, directory, device, fifo); UNIX sockets 
 * not yet implemented.
 * If the response fh has been verified, the parent directory should
 * already be locked. Note that the parent directory is left locked.
 *
 * N.B. Every call to nfsd_create needs an fh_put for _both_ fhp and resfhp
 */
1153
__be32
L
Linus Torvalds 已提交
1154 1155 1156 1157 1158 1159
nfsd_create(struct svc_rqst *rqstp, struct svc_fh *fhp,
		char *fname, int flen, struct iattr *iap,
		int type, dev_t rdev, struct svc_fh *resfhp)
{
	struct dentry	*dentry, *dchild = NULL;
	struct inode	*dirp;
1160 1161
	__be32		err;
	int		host_err;
L
Linus Torvalds 已提交
1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185

	err = nfserr_perm;
	if (!flen)
		goto out;
	err = nfserr_exist;
	if (isdotent(fname, flen))
		goto out;

	err = fh_verify(rqstp, fhp, S_IFDIR, MAY_CREATE);
	if (err)
		goto out;

	dentry = fhp->fh_dentry;
	dirp = dentry->d_inode;

	err = nfserr_notdir;
	if(!dirp->i_op || !dirp->i_op->lookup)
		goto out;
	/*
	 * Check whether the response file handle has been verified yet.
	 * If it has, the parent directory should already be locked.
	 */
	if (!resfhp->fh_dentry) {
		/* called from nfsd_proc_mkdir, or possibly nfsd3_proc_create */
P
Peter Zijlstra 已提交
1186
		fh_lock_nested(fhp, I_MUTEX_PARENT);
L
Linus Torvalds 已提交
1187
		dchild = lookup_one_len(fname, dentry, flen);
1188
		host_err = PTR_ERR(dchild);
L
Linus Torvalds 已提交
1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202
		if (IS_ERR(dchild))
			goto out_nfserr;
		err = fh_compose(resfhp, fhp->fh_export, dchild, fhp);
		if (err)
			goto out;
	} else {
		/* called from nfsd_proc_create */
		dchild = dget(resfhp->fh_dentry);
		if (!fhp->fh_locked) {
			/* not actually possible */
			printk(KERN_ERR
				"nfsd_create: parent %s/%s not locked!\n",
				dentry->d_parent->d_name.name,
				dentry->d_name.name);
A
Al Viro 已提交
1203
			err = nfserr_io;
L
Linus Torvalds 已提交
1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223
			goto out;
		}
	}
	/*
	 * Make sure the child dentry is still negative ...
	 */
	err = nfserr_exist;
	if (dchild->d_inode) {
		dprintk("nfsd_create: dentry %s/%s not negative!\n",
			dentry->d_name.name, dchild->d_name.name);
		goto out; 
	}

	if (!(iap->ia_valid & ATTR_MODE))
		iap->ia_mode = 0;
	iap->ia_mode = (iap->ia_mode & S_IALLUGO) | type;

	/*
	 * Get the dir op function pointer.
	 */
1224
	err = 0;
L
Linus Torvalds 已提交
1225 1226
	switch (type) {
	case S_IFREG:
1227
		host_err = vfs_create(dirp, dchild, iap->ia_mode, NULL);
L
Linus Torvalds 已提交
1228 1229
		break;
	case S_IFDIR:
1230
		host_err = vfs_mkdir(dirp, dchild, iap->ia_mode);
L
Linus Torvalds 已提交
1231 1232 1233 1234 1235
		break;
	case S_IFCHR:
	case S_IFBLK:
	case S_IFIFO:
	case S_IFSOCK:
1236
		host_err = vfs_mknod(dirp, dchild, iap->ia_mode, rdev);
L
Linus Torvalds 已提交
1237 1238 1239
		break;
	default:
	        printk("nfsd: bad file type %o in nfsd_create\n", type);
1240
		host_err = -EINVAL;
L
Linus Torvalds 已提交
1241
	}
1242
	if (host_err < 0)
L
Linus Torvalds 已提交
1243 1244 1245
		goto out_nfserr;

	if (EX_ISSYNC(fhp->fh_export)) {
1246
		err = nfserrno(nfsd_sync_dir(dentry));
L
Linus Torvalds 已提交
1247 1248 1249 1250 1251 1252 1253 1254 1255
		write_inode_now(dchild->d_inode, 1);
	}


	/* Set file attributes. Mode has already been set and
	 * setting uid/gid works only for root. Irix appears to
	 * send along the gid when it tries to implement setgid
	 * directories via NFS.
	 */
1256
	if ((iap->ia_valid &= ~(ATTR_UID|ATTR_GID|ATTR_MODE)) != 0) {
1257
		__be32 err2 = nfsd_setattr(rqstp, resfhp, iap, 0, (time_t)0);
1258 1259 1260
		if (err2)
			err = err2;
	}
L
Linus Torvalds 已提交
1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271
	/*
	 * Update the file handle to get the new inode info.
	 */
	if (!err)
		err = fh_update(resfhp);
out:
	if (dchild && !IS_ERR(dchild))
		dput(dchild);
	return err;

out_nfserr:
1272
	err = nfserrno(host_err);
L
Linus Torvalds 已提交
1273 1274 1275 1276 1277 1278 1279
	goto out;
}

#ifdef CONFIG_NFSD_V3
/*
 * NFSv3 version of nfsd_create
 */
1280
__be32
L
Linus Torvalds 已提交
1281 1282 1283
nfsd_create_v3(struct svc_rqst *rqstp, struct svc_fh *fhp,
		char *fname, int flen, struct iattr *iap,
		struct svc_fh *resfhp, int createmode, u32 *verifier,
1284
	        int *truncp, int *created)
L
Linus Torvalds 已提交
1285 1286 1287
{
	struct dentry	*dentry, *dchild = NULL;
	struct inode	*dirp;
1288 1289
	__be32		err;
	int		host_err;
L
Linus Torvalds 已提交
1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311
	__u32		v_mtime=0, v_atime=0;

	err = nfserr_perm;
	if (!flen)
		goto out;
	err = nfserr_exist;
	if (isdotent(fname, flen))
		goto out;
	if (!(iap->ia_valid & ATTR_MODE))
		iap->ia_mode = 0;
	err = fh_verify(rqstp, fhp, S_IFDIR, MAY_CREATE);
	if (err)
		goto out;

	dentry = fhp->fh_dentry;
	dirp = dentry->d_inode;

	/* Get all the sanity checks out of the way before
	 * we lock the parent. */
	err = nfserr_notdir;
	if(!dirp->i_op || !dirp->i_op->lookup)
		goto out;
P
Peter Zijlstra 已提交
1312
	fh_lock_nested(fhp, I_MUTEX_PARENT);
L
Linus Torvalds 已提交
1313 1314 1315 1316 1317

	/*
	 * Compose the response file handle.
	 */
	dchild = lookup_one_len(fname, dentry, flen);
1318
	host_err = PTR_ERR(dchild);
L
Linus Torvalds 已提交
1319 1320 1321 1322 1323 1324 1325 1326
	if (IS_ERR(dchild))
		goto out_nfserr;

	err = fh_compose(resfhp, fhp->fh_export, dchild, fhp);
	if (err)
		goto out;

	if (createmode == NFS3_CREATE_EXCLUSIVE) {
1327
		/* solaris7 gets confused (bugid 4218508) if these have
1328 1329 1330 1331
		 * the high bit set, so just clear the high bits. If this is
		 * ever changed to use different attrs for storing the
		 * verifier, then do_open_lookup() will also need to be fixed
		 * accordingly.
L
Linus Torvalds 已提交
1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370
		 */
		v_mtime = verifier[0]&0x7fffffff;
		v_atime = verifier[1]&0x7fffffff;
	}
	
	if (dchild->d_inode) {
		err = 0;

		switch (createmode) {
		case NFS3_CREATE_UNCHECKED:
			if (! S_ISREG(dchild->d_inode->i_mode))
				err = nfserr_exist;
			else if (truncp) {
				/* in nfsv4, we need to treat this case a little
				 * differently.  we don't want to truncate the
				 * file now; this would be wrong if the OPEN
				 * fails for some other reason.  furthermore,
				 * if the size is nonzero, we should ignore it
				 * according to spec!
				 */
				*truncp = (iap->ia_valid & ATTR_SIZE) && !iap->ia_size;
			}
			else {
				iap->ia_valid &= ATTR_SIZE;
				goto set_attr;
			}
			break;
		case NFS3_CREATE_EXCLUSIVE:
			if (   dchild->d_inode->i_mtime.tv_sec == v_mtime
			    && dchild->d_inode->i_atime.tv_sec == v_atime
			    && dchild->d_inode->i_size  == 0 )
				break;
			 /* fallthru */
		case NFS3_CREATE_GUARDED:
			err = nfserr_exist;
		}
		goto out;
	}

1371 1372
	host_err = vfs_create(dirp, dchild, iap->ia_mode, NULL);
	if (host_err < 0)
L
Linus Torvalds 已提交
1373
		goto out_nfserr;
1374 1375
	if (created)
		*created = 1;
L
Linus Torvalds 已提交
1376 1377

	if (EX_ISSYNC(fhp->fh_export)) {
1378
		err = nfserrno(nfsd_sync_dir(dentry));
L
Linus Torvalds 已提交
1379 1380 1381 1382
		/* setattr will sync the child (or not) */
	}

	if (createmode == NFS3_CREATE_EXCLUSIVE) {
1383
		/* Cram the verifier into atime/mtime */
L
Linus Torvalds 已提交
1384
		iap->ia_valid = ATTR_MTIME|ATTR_ATIME
1385
			| ATTR_MTIME_SET|ATTR_ATIME_SET;
L
Linus Torvalds 已提交
1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397
		/* XXX someone who knows this better please fix it for nsec */ 
		iap->ia_mtime.tv_sec = v_mtime;
		iap->ia_atime.tv_sec = v_atime;
		iap->ia_mtime.tv_nsec = 0;
		iap->ia_atime.tv_nsec = 0;
	}

	/* Set file attributes.
	 * Irix appears to send along the gid when it tries to
	 * implement setgid directories via NFS. Clear out all that cruft.
	 */
 set_attr:
1398
	if ((iap->ia_valid &= ~(ATTR_UID|ATTR_GID|ATTR_MODE)) != 0) {
1399
 		__be32 err2 = nfsd_setattr(rqstp, resfhp, iap, 0, (time_t)0);
1400
		if (err2)
1401
			err = err2;
1402 1403 1404 1405 1406 1407 1408
	}

	/*
	 * Update the filehandle to get the new inode info.
	 */
	if (!err)
		err = fh_update(resfhp);
L
Linus Torvalds 已提交
1409 1410 1411 1412 1413 1414 1415 1416

 out:
	fh_unlock(fhp);
	if (dchild && !IS_ERR(dchild))
		dput(dchild);
 	return err;
 
 out_nfserr:
1417
	err = nfserrno(host_err);
L
Linus Torvalds 已提交
1418 1419 1420 1421 1422 1423 1424 1425 1426
	goto out;
}
#endif /* CONFIG_NFSD_V3 */

/*
 * Read a symlink. On entry, *lenp must contain the maximum path length that
 * fits into the buffer. On return, it contains the true length.
 * N.B. After this call fhp needs an fh_put
 */
1427
__be32
L
Linus Torvalds 已提交
1428 1429 1430 1431 1432
nfsd_readlink(struct svc_rqst *rqstp, struct svc_fh *fhp, char *buf, int *lenp)
{
	struct dentry	*dentry;
	struct inode	*inode;
	mm_segment_t	oldfs;
1433 1434
	__be32		err;
	int		host_err;
L
Linus Torvalds 已提交
1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452

	err = fh_verify(rqstp, fhp, S_IFLNK, MAY_NOP);
	if (err)
		goto out;

	dentry = fhp->fh_dentry;
	inode = dentry->d_inode;

	err = nfserr_inval;
	if (!inode->i_op || !inode->i_op->readlink)
		goto out;

	touch_atime(fhp->fh_export->ex_mnt, dentry);
	/* N.B. Why does this call need a get_fs()??
	 * Remove the set_fs and watch the fireworks:-) --okir
	 */

	oldfs = get_fs(); set_fs(KERNEL_DS);
1453
	host_err = inode->i_op->readlink(dentry, buf, *lenp);
L
Linus Torvalds 已提交
1454 1455
	set_fs(oldfs);

1456
	if (host_err < 0)
L
Linus Torvalds 已提交
1457
		goto out_nfserr;
1458
	*lenp = host_err;
L
Linus Torvalds 已提交
1459 1460 1461 1462 1463
	err = 0;
out:
	return err;

out_nfserr:
1464
	err = nfserrno(host_err);
L
Linus Torvalds 已提交
1465 1466 1467 1468 1469 1470 1471
	goto out;
}

/*
 * Create a symlink and look up its inode
 * N.B. After this call _both_ fhp and resfhp need an fh_put
 */
1472
__be32
L
Linus Torvalds 已提交
1473 1474 1475 1476 1477 1478 1479
nfsd_symlink(struct svc_rqst *rqstp, struct svc_fh *fhp,
				char *fname, int flen,
				char *path,  int plen,
				struct svc_fh *resfhp,
				struct iattr *iap)
{
	struct dentry	*dentry, *dnew;
1480 1481
	__be32		err, cerr;
	int		host_err;
L
Linus Torvalds 已提交
1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496
	umode_t		mode;

	err = nfserr_noent;
	if (!flen || !plen)
		goto out;
	err = nfserr_exist;
	if (isdotent(fname, flen))
		goto out;

	err = fh_verify(rqstp, fhp, S_IFDIR, MAY_CREATE);
	if (err)
		goto out;
	fh_lock(fhp);
	dentry = fhp->fh_dentry;
	dnew = lookup_one_len(fname, dentry, flen);
1497
	host_err = PTR_ERR(dnew);
L
Linus Torvalds 已提交
1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508
	if (IS_ERR(dnew))
		goto out_nfserr;

	mode = S_IALLUGO;
	/* Only the MODE ATTRibute is even vaguely meaningful */
	if (iap && (iap->ia_valid & ATTR_MODE))
		mode = iap->ia_mode & S_IALLUGO;

	if (unlikely(path[plen] != 0)) {
		char *path_alloced = kmalloc(plen+1, GFP_KERNEL);
		if (path_alloced == NULL)
1509
			host_err = -ENOMEM;
L
Linus Torvalds 已提交
1510 1511 1512
		else {
			strncpy(path_alloced, path, plen);
			path_alloced[plen] = 0;
1513
			host_err = vfs_symlink(dentry->d_inode, dnew, path_alloced, mode);
L
Linus Torvalds 已提交
1514 1515 1516
			kfree(path_alloced);
		}
	} else
1517
		host_err = vfs_symlink(dentry->d_inode, dnew, path, mode);
L
Linus Torvalds 已提交
1518

1519
	if (!host_err) {
L
Linus Torvalds 已提交
1520
		if (EX_ISSYNC(fhp->fh_export))
1521 1522 1523
			host_err = nfsd_sync_dir(dentry);
	}
	err = nfserrno(host_err);
L
Linus Torvalds 已提交
1524 1525 1526 1527 1528 1529 1530 1531 1532
	fh_unlock(fhp);

	cerr = fh_compose(resfhp, fhp->fh_export, dnew, fhp);
	dput(dnew);
	if (err==0) err = cerr;
out:
	return err;

out_nfserr:
1533
	err = nfserrno(host_err);
L
Linus Torvalds 已提交
1534 1535 1536 1537 1538 1539 1540
	goto out;
}

/*
 * Create a hardlink
 * N.B. After this call _both_ ffhp and tfhp need an fh_put
 */
1541
__be32
L
Linus Torvalds 已提交
1542 1543 1544 1545 1546
nfsd_link(struct svc_rqst *rqstp, struct svc_fh *ffhp,
				char *name, int len, struct svc_fh *tfhp)
{
	struct dentry	*ddir, *dnew, *dold;
	struct inode	*dirp, *dest;
1547 1548
	__be32		err;
	int		host_err;
L
Linus Torvalds 已提交
1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563

	err = fh_verify(rqstp, ffhp, S_IFDIR, MAY_CREATE);
	if (err)
		goto out;
	err = fh_verify(rqstp, tfhp, -S_IFDIR, MAY_NOP);
	if (err)
		goto out;

	err = nfserr_perm;
	if (!len)
		goto out;
	err = nfserr_exist;
	if (isdotent(name, len))
		goto out;

P
Peter Zijlstra 已提交
1564
	fh_lock_nested(ffhp, I_MUTEX_PARENT);
L
Linus Torvalds 已提交
1565 1566 1567 1568
	ddir = ffhp->fh_dentry;
	dirp = ddir->d_inode;

	dnew = lookup_one_len(name, ddir, len);
1569
	host_err = PTR_ERR(dnew);
L
Linus Torvalds 已提交
1570 1571 1572 1573 1574 1575
	if (IS_ERR(dnew))
		goto out_nfserr;

	dold = tfhp->fh_dentry;
	dest = dold->d_inode;

1576 1577
	host_err = vfs_link(dold, dirp, dnew);
	if (!host_err) {
L
Linus Torvalds 已提交
1578
		if (EX_ISSYNC(ffhp->fh_export)) {
1579
			err = nfserrno(nfsd_sync_dir(ddir));
L
Linus Torvalds 已提交
1580 1581
			write_inode_now(dest, 1);
		}
1582
		err = 0;
L
Linus Torvalds 已提交
1583
	} else {
1584
		if (host_err == -EXDEV && rqstp->rq_vers == 2)
L
Linus Torvalds 已提交
1585 1586
			err = nfserr_acces;
		else
1587
			err = nfserrno(host_err);
L
Linus Torvalds 已提交
1588 1589 1590
	}

	dput(dnew);
1591 1592
out_unlock:
	fh_unlock(ffhp);
L
Linus Torvalds 已提交
1593 1594 1595 1596
out:
	return err;

out_nfserr:
1597
	err = nfserrno(host_err);
1598
	goto out_unlock;
L
Linus Torvalds 已提交
1599 1600 1601 1602 1603 1604
}

/*
 * Rename a file
 * N.B. After this call _both_ ffhp and tfhp need an fh_put
 */
1605
__be32
L
Linus Torvalds 已提交
1606 1607 1608 1609 1610
nfsd_rename(struct svc_rqst *rqstp, struct svc_fh *ffhp, char *fname, int flen,
			    struct svc_fh *tfhp, char *tname, int tlen)
{
	struct dentry	*fdentry, *tdentry, *odentry, *ndentry, *trap;
	struct inode	*fdir, *tdir;
1611 1612
	__be32		err;
	int		host_err;
L
Linus Torvalds 已提交
1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627

	err = fh_verify(rqstp, ffhp, S_IFDIR, MAY_REMOVE);
	if (err)
		goto out;
	err = fh_verify(rqstp, tfhp, S_IFDIR, MAY_CREATE);
	if (err)
		goto out;

	fdentry = ffhp->fh_dentry;
	fdir = fdentry->d_inode;

	tdentry = tfhp->fh_dentry;
	tdir = tdentry->d_inode;

	err = (rqstp->rq_vers == 2) ? nfserr_acces : nfserr_xdev;
1628
	if (ffhp->fh_export != tfhp->fh_export)
L
Linus Torvalds 已提交
1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642
		goto out;

	err = nfserr_perm;
	if (!flen || isdotent(fname, flen) || !tlen || isdotent(tname, tlen))
		goto out;

	/* cannot use fh_lock as we need deadlock protective ordering
	 * so do it by hand */
	trap = lock_rename(tdentry, fdentry);
	ffhp->fh_locked = tfhp->fh_locked = 1;
	fill_pre_wcc(ffhp);
	fill_pre_wcc(tfhp);

	odentry = lookup_one_len(fname, fdentry, flen);
1643
	host_err = PTR_ERR(odentry);
L
Linus Torvalds 已提交
1644 1645 1646
	if (IS_ERR(odentry))
		goto out_nfserr;

1647
	host_err = -ENOENT;
L
Linus Torvalds 已提交
1648 1649
	if (!odentry->d_inode)
		goto out_dput_old;
1650
	host_err = -EINVAL;
L
Linus Torvalds 已提交
1651 1652 1653 1654
	if (odentry == trap)
		goto out_dput_old;

	ndentry = lookup_one_len(tname, tdentry, tlen);
1655
	host_err = PTR_ERR(ndentry);
L
Linus Torvalds 已提交
1656 1657
	if (IS_ERR(ndentry))
		goto out_dput_old;
1658
	host_err = -ENOTEMPTY;
L
Linus Torvalds 已提交
1659 1660 1661 1662 1663 1664 1665
	if (ndentry == trap)
		goto out_dput_new;

#ifdef MSNFS
	if ((ffhp->fh_export->ex_flags & NFSEXP_MSNFS) &&
		((atomic_read(&odentry->d_count) > 1)
		 || (atomic_read(&ndentry->d_count) > 1))) {
1666
			host_err = -EPERM;
L
Linus Torvalds 已提交
1667 1668
	} else
#endif
1669 1670 1671 1672 1673
	host_err = vfs_rename(fdir, odentry, tdir, ndentry);
	if (!host_err && EX_ISSYNC(tfhp->fh_export)) {
		host_err = nfsd_sync_dir(tdentry);
		if (!host_err)
			host_err = nfsd_sync_dir(fdentry);
L
Linus Torvalds 已提交
1674 1675 1676 1677 1678 1679 1680
	}

 out_dput_new:
	dput(ndentry);
 out_dput_old:
	dput(odentry);
 out_nfserr:
1681
	err = nfserrno(host_err);
L
Linus Torvalds 已提交
1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699

	/* we cannot reply on fh_unlock on the two filehandles,
	 * as that would do the wrong thing if the two directories
	 * were the same, so again we do it by hand
	 */
	fill_post_wcc(ffhp);
	fill_post_wcc(tfhp);
	unlock_rename(tdentry, fdentry);
	ffhp->fh_locked = tfhp->fh_locked = 0;

out:
	return err;
}

/*
 * Unlink a file or directory
 * N.B. After this call fhp needs an fh_put
 */
1700
__be32
L
Linus Torvalds 已提交
1701 1702 1703 1704 1705
nfsd_unlink(struct svc_rqst *rqstp, struct svc_fh *fhp, int type,
				char *fname, int flen)
{
	struct dentry	*dentry, *rdentry;
	struct inode	*dirp;
1706 1707
	__be32		err;
	int		host_err;
L
Linus Torvalds 已提交
1708 1709 1710 1711 1712 1713 1714 1715

	err = nfserr_acces;
	if (!flen || isdotent(fname, flen))
		goto out;
	err = fh_verify(rqstp, fhp, S_IFDIR, MAY_REMOVE);
	if (err)
		goto out;

P
Peter Zijlstra 已提交
1716
	fh_lock_nested(fhp, I_MUTEX_PARENT);
L
Linus Torvalds 已提交
1717 1718 1719 1720
	dentry = fhp->fh_dentry;
	dirp = dentry->d_inode;

	rdentry = lookup_one_len(fname, dentry, flen);
1721
	host_err = PTR_ERR(rdentry);
L
Linus Torvalds 已提交
1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737
	if (IS_ERR(rdentry))
		goto out_nfserr;

	if (!rdentry->d_inode) {
		dput(rdentry);
		err = nfserr_noent;
		goto out;
	}

	if (!type)
		type = rdentry->d_inode->i_mode & S_IFMT;

	if (type != S_IFDIR) { /* It's UNLINK */
#ifdef MSNFS
		if ((fhp->fh_export->ex_flags & NFSEXP_MSNFS) &&
			(atomic_read(&rdentry->d_count) > 1)) {
1738
			host_err = -EPERM;
L
Linus Torvalds 已提交
1739 1740
		} else
#endif
1741
		host_err = vfs_unlink(dirp, rdentry);
L
Linus Torvalds 已提交
1742
	} else { /* It's RMDIR */
1743
		host_err = vfs_rmdir(dirp, rdentry);
L
Linus Torvalds 已提交
1744 1745 1746 1747
	}

	dput(rdentry);

1748 1749 1750 1751
	if (host_err)
		goto out_nfserr;
	if (EX_ISSYNC(fhp->fh_export))
		host_err = nfsd_sync_dir(dentry);
L
Linus Torvalds 已提交
1752 1753

out_nfserr:
1754
	err = nfserrno(host_err);
1755 1756
out:
	return err;
L
Linus Torvalds 已提交
1757 1758 1759 1760 1761 1762
}

/*
 * Read entries from a directory.
 * The  NFSv3/4 verifier we ignore for now.
 */
1763
__be32
L
Linus Torvalds 已提交
1764
nfsd_readdir(struct svc_rqst *rqstp, struct svc_fh *fhp, loff_t *offsetp, 
1765
	     struct readdir_cd *cdp, filldir_t func)
L
Linus Torvalds 已提交
1766
{
1767 1768
	__be32		err;
	int 		host_err;
L
Linus Torvalds 已提交
1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789
	struct file	*file;
	loff_t		offset = *offsetp;

	err = nfsd_open(rqstp, fhp, S_IFDIR, MAY_READ, &file);
	if (err)
		goto out;

	offset = vfs_llseek(file, offset, 0);
	if (offset < 0) {
		err = nfserrno((int)offset);
		goto out_close;
	}

	/*
	 * Read the directory entries. This silly loop is necessary because
	 * readdir() is not guaranteed to fill up the entire buffer, but
	 * may choose to do less.
	 */

	do {
		cdp->err = nfserr_eof; /* will be cleared on successful read */
1790
		host_err = vfs_readdir(file, func, cdp);
1791 1792 1793
	} while (host_err >=0 && cdp->err == nfs_ok);
	if (host_err)
		err = nfserrno(host_err);
L
Linus Torvalds 已提交
1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809
	else
		err = cdp->err;
	*offsetp = vfs_llseek(file, 0, 1);

	if (err == nfserr_eof || err == nfserr_toosmall)
		err = nfs_ok; /* can still be found in ->err */
out_close:
	nfsd_close(file);
out:
	return err;
}

/*
 * Get file system stats
 * N.B. After this call fhp needs an fh_put
 */
1810
__be32
L
Linus Torvalds 已提交
1811 1812
nfsd_statfs(struct svc_rqst *rqstp, struct svc_fh *fhp, struct kstatfs *stat)
{
1813
	__be32 err = fh_verify(rqstp, fhp, 0, MAY_NOP);
1814
	if (!err && vfs_statfs(fhp->fh_dentry,stat))
L
Linus Torvalds 已提交
1815 1816 1817 1818
		err = nfserr_io;
	return err;
}

J
J. Bruce Fields 已提交
1819
static int exp_rdonly(struct svc_rqst *rqstp, struct svc_export *exp)
1820
{
J
J. Bruce Fields 已提交
1821
	return nfsexp_flags(rqstp, exp) & NFSEXP_READONLY;
1822 1823
}

L
Linus Torvalds 已提交
1824 1825 1826
/*
 * Check for a user's access permissions to this inode.
 */
1827
__be32
1828 1829
nfsd_permission(struct svc_rqst *rqstp, struct svc_export *exp,
					struct dentry *dentry, int acc)
L
Linus Torvalds 已提交
1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859
{
	struct inode	*inode = dentry->d_inode;
	int		err;

	if (acc == MAY_NOP)
		return 0;
#if 0
	dprintk("nfsd: permission 0x%x%s%s%s%s%s%s%s mode 0%o%s%s%s\n",
		acc,
		(acc & MAY_READ)?	" read"  : "",
		(acc & MAY_WRITE)?	" write" : "",
		(acc & MAY_EXEC)?	" exec"  : "",
		(acc & MAY_SATTR)?	" sattr" : "",
		(acc & MAY_TRUNC)?	" trunc" : "",
		(acc & MAY_LOCK)?	" lock"  : "",
		(acc & MAY_OWNER_OVERRIDE)? " owneroverride" : "",
		inode->i_mode,
		IS_IMMUTABLE(inode)?	" immut" : "",
		IS_APPEND(inode)?	" append" : "",
		IS_RDONLY(inode)?	" ro" : "");
	dprintk("      owner %d/%d user %d/%d\n",
		inode->i_uid, inode->i_gid, current->fsuid, current->fsgid);
#endif

	/* Normally we reject any write/sattr etc access on a read-only file
	 * system.  But if it is IRIX doing check on write-access for a 
	 * device special file, we ignore rofs.
	 */
	if (!(acc & MAY_LOCAL_ACCESS))
		if (acc & (MAY_WRITE | MAY_SATTR | MAY_TRUNC)) {
J
J. Bruce Fields 已提交
1860
			if (exp_rdonly(rqstp, exp) || IS_RDONLY(inode))
L
Linus Torvalds 已提交
1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908
				return nfserr_rofs;
			if (/* (acc & MAY_WRITE) && */ IS_IMMUTABLE(inode))
				return nfserr_perm;
		}
	if ((acc & MAY_TRUNC) && IS_APPEND(inode))
		return nfserr_perm;

	if (acc & MAY_LOCK) {
		/* If we cannot rely on authentication in NLM requests,
		 * just allow locks, otherwise require read permission, or
		 * ownership
		 */
		if (exp->ex_flags & NFSEXP_NOAUTHNLM)
			return 0;
		else
			acc = MAY_READ | MAY_OWNER_OVERRIDE;
	}
	/*
	 * The file owner always gets access permission for accesses that
	 * would normally be checked at open time. This is to make
	 * file access work even when the client has done a fchmod(fd, 0).
	 *
	 * However, `cp foo bar' should fail nevertheless when bar is
	 * readonly. A sensible way to do this might be to reject all
	 * attempts to truncate a read-only file, because a creat() call
	 * always implies file truncation.
	 * ... but this isn't really fair.  A process may reasonably call
	 * ftruncate on an open file descriptor on a file with perm 000.
	 * We must trust the client to do permission checking - using "ACCESS"
	 * with NFSv3.
	 */
	if ((acc & MAY_OWNER_OVERRIDE) &&
	    inode->i_uid == current->fsuid)
		return 0;

	err = permission(inode, acc & (MAY_READ|MAY_WRITE|MAY_EXEC), NULL);

	/* Allow read access to binaries even when mode 111 */
	if (err == -EACCES && S_ISREG(inode->i_mode) &&
	    acc == (MAY_READ | MAY_OWNER_OVERRIDE))
		err = permission(inode, MAY_EXEC, NULL);

	return err? nfserrno(err) : 0;
}

void
nfsd_racache_shutdown(void)
{
1909
	if (!raparml)
L
Linus Torvalds 已提交
1910 1911 1912
		return;
	dprintk("nfsd: freeing readahead buffers.\n");
	kfree(raparml);
1913
	raparml = NULL;
L
Linus Torvalds 已提交
1914 1915 1916 1917 1918 1919 1920 1921
}
/*
 * Initialize readahead param cache
 */
int
nfsd_racache_init(int cache_size)
{
	int	i;
1922 1923
	int	j = 0;
	int	nperbucket;
L
Linus Torvalds 已提交
1924

1925 1926

	if (raparml)
L
Linus Torvalds 已提交
1927
		return 0;
1928 1929
	if (cache_size < 2*RAPARM_HASH_SIZE)
		cache_size = 2*RAPARM_HASH_SIZE;
1930 1931 1932
	raparml = kcalloc(cache_size, sizeof(struct raparms), GFP_KERNEL);

	if (!raparml) {
L
Linus Torvalds 已提交
1933
		printk(KERN_WARNING
1934
			"nfsd: Could not allocate memory read-ahead cache.\n");
L
Linus Torvalds 已提交
1935 1936
		return -ENOMEM;
	}
1937 1938 1939 1940 1941 1942

	dprintk("nfsd: allocating %d readahead buffers.\n", cache_size);
	for (i = 0 ; i < RAPARM_HASH_SIZE ; i++) {
		raparm_hash[i].pb_head = NULL;
		spin_lock_init(&raparm_hash[i].pb_lock);
	}
1943
	nperbucket = DIV_ROUND_UP(cache_size, RAPARM_HASH_SIZE);
1944 1945 1946 1947 1948 1949 1950
	for (i = 0; i < cache_size - 1; i++) {
		if (i % nperbucket == 0)
			raparm_hash[j++].pb_head = raparml + i;
		if (i % nperbucket < nperbucket-1)
			raparml[i].p_next = raparml + i + 1;
	}

L
Linus Torvalds 已提交
1951 1952 1953
	nfsdstats.ra_size = cache_size;
	return 0;
}
1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964

#if defined(CONFIG_NFSD_V2_ACL) || defined(CONFIG_NFSD_V3_ACL)
struct posix_acl *
nfsd_get_posix_acl(struct svc_fh *fhp, int type)
{
	struct inode *inode = fhp->fh_dentry->d_inode;
	char *name;
	void *value = NULL;
	ssize_t size;
	struct posix_acl *acl;

1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975
	if (!IS_POSIXACL(inode))
		return ERR_PTR(-EOPNOTSUPP);

	switch (type) {
	case ACL_TYPE_ACCESS:
		name = POSIX_ACL_XATTR_ACCESS;
		break;
	case ACL_TYPE_DEFAULT:
		name = POSIX_ACL_XATTR_DEFAULT;
		break;
	default:
1976 1977 1978
		return ERR_PTR(-EOPNOTSUPP);
	}

1979 1980 1981
	size = nfsd_getxattr(fhp->fh_dentry, name, &value);
	if (size < 0)
		return ERR_PTR(size);
1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001

	acl = posix_acl_from_xattr(value, size);
	kfree(value);
	return acl;
}

int
nfsd_set_posix_acl(struct svc_fh *fhp, int type, struct posix_acl *acl)
{
	struct inode *inode = fhp->fh_dentry->d_inode;
	char *name;
	void *value = NULL;
	size_t size;
	int error;

	if (!IS_POSIXACL(inode) || !inode->i_op ||
	    !inode->i_op->setxattr || !inode->i_op->removexattr)
		return -EOPNOTSUPP;
	switch(type) {
		case ACL_TYPE_ACCESS:
2002
			name = POSIX_ACL_XATTR_ACCESS;
2003 2004
			break;
		case ACL_TYPE_DEFAULT:
2005
			name = POSIX_ACL_XATTR_DEFAULT;
2006 2007 2008 2009 2010 2011
			break;
		default:
			return -EOPNOTSUPP;
	}

	if (acl && acl->a_count) {
2012
		size = posix_acl_xattr_size(acl->a_count);
2013 2014 2015
		value = kmalloc(size, GFP_KERNEL);
		if (!value)
			return -ENOMEM;
2016 2017
		error = posix_acl_to_xattr(acl, value, size);
		if (error < 0)
2018
			goto getout;
2019
		size = error;
2020 2021 2022 2023
	} else
		size = 0;

	if (size)
2024
		error = vfs_setxattr(fhp->fh_dentry, name, value, size, 0);
2025 2026 2027 2028
	else {
		if (!S_ISDIR(inode->i_mode) && type == ACL_TYPE_DEFAULT)
			error = 0;
		else {
2029
			error = vfs_removexattr(fhp->fh_dentry, name);
2030 2031 2032 2033 2034 2035 2036 2037 2038 2039
			if (error == -ENODATA)
				error = 0;
		}
	}

getout:
	kfree(value);
	return error;
}
#endif  /* defined(CONFIG_NFSD_V2_ACL) || defined(CONFIG_NFSD_V3_ACL) */