/*
 * Lock a (portion of) a file
 */
static int nfs_lock(struct file *filp, int cmd, struct file_lock *fl)
{
	struct inode *inode = filp->f_mapping->host;
	int ret = -ENOLCK;

	dprintk("NFS: lock(%s/%s, t=%x, fl=%x, r=%lld:%lld)\n",
			filp->f_path.dentry->d_parent->d_name.name,
			filp->f_path.dentry->d_name.name,
			fl->fl_type, fl->fl_flags,
			(long long)fl->fl_start, (long long)fl->fl_end);

	nfs_inc_stats(inode, NFSIOS_VFSLOCK);

	/* No mandatory locks over NFS */
	if (__mandatory_lock(inode) && fl->fl_type != F_UNLCK)
		goto out_err;

	if (NFS_PROTO(inode)->lock_check_bounds != NULL) {
		ret = NFS_PROTO(inode)->lock_check_bounds(fl);
		if (ret < 0)
			goto out_err;
	}

	if (IS_GETLK(cmd))
		ret = do_getlk(filp, cmd, fl);
	else if (fl->fl_type == F_UNLCK)
		ret = do_unlk(filp, cmd, fl);
	else
		ret = do_setlk(filp, cmd, fl);
out_err:
	return ret;
}

static int
do_getlk(struct file *filp, int cmd, struct file_lock *fl, int is_local)
{
	struct inode *inode = filp->f_mapping->host;
	int status = 0;
	unsigned int saved_type = fl->fl_type;

	/* Try local locking first */
	posix_test_lock(filp, fl);
	if (fl->fl_type != F_UNLCK) {
		/* found a conflict */
		goto out;
	}
	fl->fl_type = saved_type;

	if (NFS_PROTO(inode)->have_delegation(inode, FMODE_READ))
		goto out_noconflict;

	if (is_local)
		goto out_noconflict;

	status = NFS_PROTO(inode)->lock(filp, cmd, fl);
out:
	return status;
out_noconflict:
	fl->fl_type = F_UNLCK;
	goto out;
}

/*
 * Remove a file after making sure there are no pending writes,
 * and after checking that the file has only one user. 
 *
 * We invalidate the attribute cache and free the inode prior to the operation
 * to avoid possible races if the server reuses the inode.
 */
static int nfs_safe_remove(struct dentry *dentry)
{
	struct inode *dir = dentry->d_parent->d_inode;
	struct inode *inode = dentry->d_inode;
	int error = -EBUSY;
		
	dfprintk(VFS, "NFS: safe_remove(%s/%s)\n",
		dentry->d_parent->d_name.name, dentry->d_name.name);

	/* If the dentry was sillyrenamed, we simply call d_delete() */
	if (dentry->d_flags & DCACHE_NFSFS_RENAMED) {
		error = 0;
		goto out;
	}

	nfs_begin_data_update(dir);
	if (inode != NULL) {
		nfs_begin_data_update(inode);
		error = NFS_PROTO(dir)->remove(dir, &dentry->d_name);
		/* The VFS may want to delete this inode */
		if (error == 0)
			inode->i_nlink--;
		nfs_end_data_update(inode);
	} else
		error = NFS_PROTO(dir)->remove(dir, &dentry->d_name);
	nfs_end_data_update(dir);
out:
	return error;
}

/*
 * Lock a (portion of) a file
 */
int nfs_lock(struct file *filp, int cmd, struct file_lock *fl)
{
	struct inode *inode = filp->f_mapping->host;
	int ret = -ENOLCK;
	int is_local = 0;

	dprintk("NFS: lock(%pD2, t=%x, fl=%x, r=%lld:%lld)\n",
			filp, fl->fl_type, fl->fl_flags,
			(long long)fl->fl_start, (long long)fl->fl_end);

	nfs_inc_stats(inode, NFSIOS_VFSLOCK);

	/* No mandatory locks over NFS */
	if (__mandatory_lock(inode) && fl->fl_type != F_UNLCK)
		goto out_err;

	if (NFS_SERVER(inode)->flags & NFS_MOUNT_LOCAL_FCNTL)
		is_local = 1;

	if (NFS_PROTO(inode)->lock_check_bounds != NULL) {
		ret = NFS_PROTO(inode)->lock_check_bounds(fl);
		if (ret < 0)
			goto out_err;
	}

	if (IS_GETLK(cmd))
		ret = do_getlk(filp, cmd, fl, is_local);
	else if (fl->fl_type == F_UNLCK)
		ret = do_unlk(filp, cmd, fl, is_local);
	else
		ret = do_setlk(filp, cmd, fl, is_local);
out_err:
	return ret;
}

int
nfs_permission(struct inode *inode, int mask)
{
	int			error = vfs_permission(inode, mask);

	if (!NFS_PROTO(inode)->access)
		goto out;

	if (error == -EROFS)
		goto out;

	/*
	 * Trust UNIX mode bits except:
	 *
	 * 1) When override capabilities may have been invoked
	 * 2) When root squashing may be involved
	 * 3) When ACLs may overturn a negative answer */
	if (!capable(CAP_DAC_OVERRIDE) && !capable(CAP_DAC_READ_SEARCH)
	    && (current->fsuid != 0) && (current->fsgid != 0)
	    && error != -EACCES)
		goto out;

	error = NFS_PROTO(inode)->access(inode, mask, 0);

	if (error == -EACCES && NFS_CLIENT(inode)->cl_droppriv &&
	    current->uid != 0 && current->gid != 0 &&
	    (current->fsuid != current->uid || current->fsgid != current->gid))
		error = NFS_PROTO(inode)->access(inode, mask, 1);

 out:
	return error;
}

static int
nfs_symlink(struct inode *dir, struct dentry *dentry, const char *symname)
{
	struct iattr attr;
	struct nfs_fattr sym_attr;
	struct nfs_fh sym_fh;
	struct qstr qsymname;
	int error;

	dfprintk(VFS, "NFS: symlink(%s/%ld, %s, %s)\n", dir->i_sb->s_id,
		dir->i_ino, dentry->d_name.name, symname);

	error = -ENAMETOOLONG;
	switch (NFS_PROTO(dir)->version) {
		case 2:
			if (strlen(symname) > NFS2_MAXPATHLEN)
				goto out;
			break;
		case 3:
			if (strlen(symname) > NFS3_MAXPATHLEN)
				goto out;
		default:
			break;
	}

#ifdef NFS_PARANOIA
if (dentry->d_inode)
printk("nfs_proc_symlink: %s/%s not negative!\n",
dentry->d_parent->d_name.name, dentry->d_name.name);
#endif
	/*
	 * Fill in the sattr for the call.
 	 * Note: SunOS 4.1.2 crashes if the mode isn't initialized!
	 */
	attr.ia_valid = ATTR_MODE;
	attr.ia_mode = S_IFLNK | S_IRWXUGO;

	qsymname.name = symname;
	qsymname.len  = strlen(symname);

	lock_kernel();
	nfs_begin_data_update(dir);
	error = NFS_PROTO(dir)->symlink(dir, &dentry->d_name, &qsymname,
					  &attr, &sym_fh, &sym_attr);
	nfs_end_data_update(dir);
	if (!error) {
		error = nfs_instantiate(dentry, &sym_fh, &sym_attr);
	} else {
		if (error == -EEXIST)
			printk("nfs_proc_symlink: %s/%s already exists??\n",
			       dentry->d_parent->d_name.name, dentry->d_name.name);
		d_drop(dentry);
	}
	unlock_kernel();

out:
	return error;
}

static inline int
nfs_direct_write_rpc(struct file *file, struct nfs_writeargs *arg,
	struct nfs_writeverf *verf)
{
	int result;
	struct inode *inode = file->f_dentry->d_inode;
	struct nfs_fattr fattr;
        struct rpc_message msg;
        struct nfs_writeres res = { &fattr, verf, 0 };

#ifdef CONFIG_NFS_V3
	msg.rpc_proc = (NFS_PROTO(inode)->version == 3) ?
						NFS3PROC_WRITE : NFSPROC_WRITE;
#else
	msg.rpc_proc = NFSPROC_WRITE;
#endif
	msg.rpc_argp = arg;
        msg.rpc_resp = &res;

	lock_kernel();
	msg.rpc_cred = get_rpccred(nfs_file_cred(file));
	fattr.valid = 0;
        result = rpc_call_sync(NFS_CLIENT(inode), &msg, 0);
	nfs_write_attributes(inode, &fattr);
	put_rpccred(msg.rpc_cred);
	unlock_kernel();

#ifdef CONFIG_NFS_V3
	if (NFS_PROTO(inode)->version == 3) {
		if (result > 0) {
			if ((arg->stable == NFS_FILE_SYNC) &&
			    (verf->committed != NFS_FILE_SYNC)) {
				printk(KERN_ERR
				"%s: server didn't sync stable write request\n",
				__FUNCTION__);
				return -EIO;
			}

			if (result != arg->count) {
				printk(KERN_INFO
					"%s: short write, count=%u, result=%d\n",
					__FUNCTION__, arg->count, result);
			}
		}
		return result;
	} else {
#endif
        	verf->committed = NFS_FILE_SYNC; /* NFSv2 always syncs data */
		if (result == 0)
			return arg->count;
		return result;
#ifdef CONFIG_NFS_V3
	}
#endif
}

static struct dentry *nfs_lookup(struct inode *dir, struct dentry * dentry, struct nameidata *nd)
{
	struct inode *inode = NULL;
	int error;
	struct nfs_fh fhandle;
	struct nfs_fattr fattr;

	dfprintk(VFS, "NFS: lookup(%s/%s)\n",
		dentry->d_parent->d_name.name, dentry->d_name.name);

	error = -ENAMETOOLONG;
	if (dentry->d_name.len > NFS_SERVER(dir)->namelen)
		goto out;

	error = -ENOMEM;
	dentry->d_op = NFS_PROTO(dir)->dentry_ops;

	lock_kernel();
	/* Revalidate parent directory attribute cache */
	nfs_revalidate_inode(NFS_SERVER(dir), dir);

	/* If we're doing an exclusive create, optimize away the lookup */
	if (nfs_is_exclusive_create(dir, nd))
		goto no_entry;

	error = nfs_cached_lookup(dir, dentry, &fhandle, &fattr);
	if (error != 0) {
		error = NFS_PROTO(dir)->lookup(dir, &dentry->d_name,
				&fhandle, &fattr);
		if (error == -ENOENT)
			goto no_entry;
		if (error != 0)
			goto out_unlock;
	}
	error = -EACCES;
	inode = nfs_fhget(dentry->d_sb, &fhandle, &fattr);
	if (!inode)
		goto out_unlock;
no_entry:
	error = 0;
	d_add(dentry, inode);
	nfs_renew_times(dentry);
	nfs_set_verifier(dentry, nfs_save_change_attribute(dir));
out_unlock:
	unlock_kernel();
out:
	BUG_ON(error > 0);
	return ERR_PTR(error);
}

static int
nfs_symlink(struct inode *dir, struct dentry *dentry, const char *symname)
{
	struct iattr attr;
	struct nfs_fattr sym_attr;
	struct nfs_fh sym_fh;
	struct qstr qsymname;
	unsigned int maxlen;
	int error;

	dfprintk(VFS, "NFS: symlink(%x/%ld, %s, %s)\n",
		dir->i_dev, dir->i_ino, dentry->d_name.name, symname);

	error = -ENAMETOOLONG;
	maxlen = (NFS_PROTO(dir)->version==2) ? NFS2_MAXPATHLEN : NFS3_MAXPATHLEN;
	if (strlen(symname) > maxlen)
		goto out;

#ifdef NFS_PARANOIA
if (dentry->d_inode)
printk("nfs_proc_symlink: %s/%s not negative!\n",
dentry->d_parent->d_name.name, dentry->d_name.name);
#endif
	/*
	 * Fill in the sattr for the call.
 	 * Note: SunOS 4.1.2 crashes if the mode isn't initialized!
	 */
	attr.ia_valid = ATTR_MODE;
	attr.ia_mode = S_IFLNK | S_IRWXUGO;

	qsymname.name = symname;
	qsymname.len  = strlen(symname);

	nfs_zap_caches(dir);
	error = NFS_PROTO(dir)->symlink(dir, &dentry->d_name, &qsymname,
					  &attr, &sym_fh, &sym_attr);
	if (!error) {
		error = nfs_instantiate(dentry, &sym_fh, &sym_attr);
	} else {
		if (error == -EEXIST)
			printk("nfs_proc_symlink: %s/%s already exists??\n",
			       dentry->d_parent->d_name.name, dentry->d_name.name);
		d_drop(dentry);
	}

out:
	return error;
}

static inline int
nfs_direct_read_rpc(struct file *file, struct nfs_readargs *arg)
{
	int result;
	struct inode * inode = file->f_dentry->d_inode;
	struct nfs_fattr fattr;
        struct rpc_message msg;
        struct nfs_readres res = { &fattr, arg->count, 0 };

#ifdef CONFIG_NFS_V3
	msg.rpc_proc = (NFS_PROTO(inode)->version == 3) ?
						NFS3PROC_READ : NFSPROC_READ;
#else
	msg.rpc_proc = NFSPROC_READ;
#endif
	msg.rpc_argp = arg;
        msg.rpc_resp = &res;

	lock_kernel();
        msg.rpc_cred = nfs_file_cred(file);
        fattr.valid = 0;
        result = rpc_call_sync(NFS_CLIENT(inode), &msg, 0);
	nfs_refresh_inode(inode, &fattr);
	unlock_kernel();

	return result;
}

static int do_getlk(struct file *filp, int cmd, struct file_lock *fl)
{
	struct file_lock *cfl;
	struct inode *inode = filp->f_mapping->host;
	int status = 0;

	lock_kernel();
	/* Try local locking first */
	cfl = posix_test_lock(filp, fl);
	if (cfl != NULL) {
		locks_copy_lock(fl, cfl);
		goto out;
	}

	if (nfs_have_delegation(inode, FMODE_READ))
		goto out_noconflict;

	if (NFS_SERVER(inode)->flags & NFS_MOUNT_NONLM)
		goto out_noconflict;

	status = NFS_PROTO(inode)->lock(filp, cmd, fl);
out:
	unlock_kernel();
	return status;
out_noconflict:
	fl->fl_type = F_UNLCK;
	goto out;
}

/*
 * See comments for nfs_proc_create regarding failed operations.
 */
static int
nfs_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t rdev)
{
	struct iattr attr;
	struct nfs_fattr fattr;
	struct nfs_fh fhandle;
	int error;

	dfprintk(VFS, "NFS: mknod(%s/%ld, %s\n", dir->i_sb->s_id,
		dir->i_ino, dentry->d_name.name);

	if (!new_valid_dev(rdev))
		return -EINVAL;

	attr.ia_mode = mode;
	attr.ia_valid = ATTR_MODE;

	lock_kernel();
	nfs_begin_data_update(dir);
	error = NFS_PROTO(dir)->mknod(dir, &dentry->d_name, &attr, rdev,
					&fhandle, &fattr);
	nfs_end_data_update(dir);
	if (!error)
		error = nfs_instantiate(dentry, &fhandle, &fattr);
	else
		d_drop(dentry);
	unlock_kernel();
	return error;
}

/*
 * Code common to create, mkdir, and mknod.
 */
static int nfs_instantiate(struct dentry *dentry, struct nfs_fh *fhandle,
				struct nfs_fattr *fattr)
{
	struct inode *inode;
	int error = -EACCES;

	/* We may have been initialized further down */
	if (dentry->d_inode)
		return 0;
	if (fhandle->size == 0 || !(fattr->valid & NFS_ATTR_FATTR)) {
		struct inode *dir = dentry->d_parent->d_inode;
		error = NFS_PROTO(dir)->lookup(dir, &dentry->d_name, fhandle, fattr);
		if (error)
			goto out_err;
	}
	inode = nfs_fhget(dentry->d_sb, fhandle, fattr);
	if (inode) {
		d_instantiate(dentry, inode);
		nfs_renew_times(dentry);
		nfs_set_verifier(dentry, nfs_save_change_attribute(dentry->d_parent->d_inode));
		return 0;
	}
	error = -ENOMEM;
out_err:
	d_drop(dentry);
	return error;
}

/*
 * Following a failed create operation, we drop the dentry rather
 * than retain a negative dentry. This avoids a problem in the event
 * that the operation succeeded on the server, but an error in the
 * reply path made it appear to have failed.
 */
static int nfs_create(struct inode *dir, struct dentry *dentry, int mode)
{
	struct iattr attr;
	struct nfs_fattr fattr;
	struct nfs_fh fhandle;
	int error;

	dfprintk(VFS, "NFS: create(%x/%ld, %s\n",
		dir->i_dev, dir->i_ino, dentry->d_name.name);

	attr.ia_mode = mode;
	attr.ia_valid = ATTR_MODE;

	/*
	 * The 0 argument passed into the create function should one day
	 * contain the O_EXCL flag if requested. This allows NFSv3 to
	 * select the appropriate create strategy. Currently open_namei
	 * does not pass the create flags.
	 */
	nfs_zap_caches(dir);
	error = NFS_PROTO(dir)->create(dir, &dentry->d_name,
					 &attr, 0, &fhandle, &fattr);
	if (!error)
		error = nfs_instantiate(dentry, &fhandle, &fattr);
	else
		d_drop(dentry);
	return error;
}

/*
 * See comments for nfs_proc_create regarding failed operations.
 */
static int nfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
{
	struct iattr attr;
	struct nfs_fattr fattr;
	struct nfs_fh fhandle;
	int error;

	dfprintk(VFS, "NFS: mkdir(%x/%ld, %s\n",
		dir->i_dev, dir->i_ino, dentry->d_name.name);

	attr.ia_valid = ATTR_MODE;
	attr.ia_mode = mode | S_IFDIR;

#if 0
	/*
	 * Always drop the dentry, we can't always depend on
	 * the fattr returned by the server (AIX seems to be
	 * broken). We're better off doing another lookup than
	 * depending on potentially bogus information.
	 */
	d_drop(dentry);
#endif
	nfs_zap_caches(dir);
	error = NFS_PROTO(dir)->mkdir(dir, &dentry->d_name, &attr, &fhandle,
					&fattr);
	if (!error)
		error = nfs_instantiate(dentry, &fhandle, &fattr);
	else
		d_drop(dentry);
	return error;
}

/*
 * This is called every time the dcache has a lookup hit,
 * and we should check whether we can really trust that
 * lookup.
 *
 * NOTE! The hit can be a negative hit too, don't assume
 * we have an inode!
 *
 * If the parent directory is seen to have changed, we throw out the
 * cached dentry and do a new lookup.
 */
static int nfs_lookup_revalidate(struct dentry * dentry, int flags)
{
	struct inode *dir;
	struct inode *inode;
	int error;
	struct nfs_fh fhandle;
	struct nfs_fattr fattr;

	lock_kernel();
	dir = dentry->d_parent->d_inode;
	inode = dentry->d_inode;

	if (!inode) {
		if (nfs_neg_need_reval(dir, dentry))
			goto out_bad;
		goto out_valid;
	}

	if (is_bad_inode(inode)) {
		dfprintk(VFS, "nfs_lookup_validate: %s/%s has dud inode\n",
			dentry->d_parent->d_name.name, dentry->d_name.name);
		goto out_bad;
	}

	/* Force a full look up iff the parent directory has changed */
	if (nfs_check_verifier(dir, dentry)) {
		if (nfs_lookup_verify_inode(inode, flags))
			goto out_bad;
		goto out_valid;
	}

	if (NFS_STALE(inode))
		goto out_bad;

	error = NFS_PROTO(dir)->lookup(dir, &dentry->d_name, &fhandle, &fattr);
	if (error)
		goto out_bad;
	if (memcmp(NFS_FH(inode), &fhandle, sizeof(struct nfs_fh))!= 0)
		goto out_bad;
	if ((error = nfs_refresh_inode(inode, &fattr)) != 0)
		goto out_bad;

	nfs_renew_times(dentry);
 out_valid:
	unlock_kernel();
	return 1;
 out_bad:
	NFS_CACHEINV(dir);
	if (inode && S_ISDIR(inode->i_mode)) {
		/* Purge readdir caches. */
		nfs_zap_caches(inode);
		/* If we have submounts, don't unhash ! */
		if (have_submounts(dentry))
			goto out_valid;
		shrink_dcache_parent(dentry);
	}
	d_drop(dentry);
	unlock_kernel();
	return 0;
}

static struct dentry *nfs_lookup(struct inode *dir, struct dentry * dentry)
{
	struct inode *inode;
	int error;
	struct nfs_fh fhandle;
	struct nfs_fattr fattr;

	dfprintk(VFS, "NFS: lookup(%s/%s)\n",
		dentry->d_parent->d_name.name, dentry->d_name.name);

	error = -ENAMETOOLONG;
	if (dentry->d_name.len > NFS_SERVER(dir)->namelen)
		goto out;

	error = -ENOMEM;
	dentry->d_op = &nfs_dentry_operations;

	error = NFS_PROTO(dir)->lookup(dir, &dentry->d_name, &fhandle, &fattr);
	inode = NULL;
	if (error == -ENOENT)
		goto no_entry;
	if (!error) {
		error = -EACCES;
		inode = nfs_fhget(dentry, &fhandle, &fattr);
		if (inode) {
	    no_entry:
			d_add(dentry, inode);
			error = 0;
		}
		nfs_renew_times(dentry);
	}
out:
	return ERR_PTR(error);
}

static int do_getlk(struct file *filp, int cmd, struct file_lock *fl)
{
	struct inode *inode = filp->f_mapping->host;
	int status = 0;

	/* Try local locking first */
	posix_test_lock(filp, fl);
	if (fl->fl_type != F_UNLCK) {
		/* found a conflict */
		goto out;
	}

	if (nfs_have_delegation(inode, FMODE_READ))
		goto out_noconflict;

	if (NFS_SERVER(inode)->flags & NFS_MOUNT_NONLM)
		goto out_noconflict;

	status = NFS_PROTO(inode)->lock(filp, cmd, fl);
out:
	return status;
out_noconflict:
	fl->fl_type = F_UNLCK;
	goto out;
}

static int do_setlk(struct file *filp, int cmd, struct file_lock *fl)
{
	struct inode *inode = filp->f_mapping->host;
	int status;

	/*
	 * Flush all pending writes before doing anything
	 * with locks..
	 */
	status = nfs_sync_mapping(filp->f_mapping);
	if (status != 0)
		goto out;

	/* Use local locking if mounted with "-onolock" */
	if (!(NFS_SERVER(inode)->flags & NFS_MOUNT_NONLM))
		status = NFS_PROTO(inode)->lock(filp, cmd, fl);
	else
		status = do_vfs_lock(filp, fl);
	if (status < 0)
		goto out;
	/*
	 * Make sure we clear the cache whenever we try to get the lock.
	 * This makes locking act as a cache coherency point.
	 */
	nfs_sync_mapping(filp->f_mapping);
	if (!nfs_have_delegation(inode, FMODE_READ))
		nfs_zap_caches(inode);
out:
	return status;
}

static int
do_unlk(struct file *filp, int cmd, struct file_lock *fl, int is_local)
{
	struct inode *inode = filp->f_mapping->host;
	int status;

	/*
	 * Flush all pending writes before doing anything
	 * with locks..
	 */
	nfs_sync_mapping(filp->f_mapping);

	/* NOTE: special case
	 * 	If we're signalled while cleaning up locks on process exit, we
	 * 	still need to complete the unlock.
	 */
	/*
	 * Use local locking if mounted with "-onolock" or with appropriate
	 * "-olocal_lock="
	 */
	if (!is_local)
		status = NFS_PROTO(inode)->lock(filp, cmd, fl);
	else
		status = do_vfs_lock(filp, fl);
	return status;
}