aboutsummaryrefslogtreecommitdiffstats
path: root/fs/ocfs2/mmap.c
blob: a7f0ccc6fdd81e95c2cff63c331c34633b2d6bb8 (plain) (blame)
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
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
/* -*- mode: c; c-basic-offset: 8; -*-
 * vim: noexpandtab sw=8 ts=8 sts=0:
 *
 * mmap.c
 *
 * Code to deal with the mess that is clustered mmap.
 *
 * Copyright (C) 2002, 2004 Oracle.  All rights reserved.
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public
 * License as published by the Free Software Foundation; either
 * version 2 of the License, or (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * General Public License for more details.
 *
 * You should have received a copy of the GNU General Public
 * License along with this program; if not, write to the
 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
 * Boston, MA 021110-1307, USA.
 */

#include <linux/fs.h>
#include <linux/types.h>
#include <linux/slab.h>
#include <linux/highmem.h>
#include <linux/pagemap.h>
#include <linux/uio.h>
#include <linux/signal.h>
#include <linux/rbtree.h>

#define MLOG_MASK_PREFIX ML_FILE_IO
#include <cluster/masklog.h>

#include "ocfs2.h"

#include "aops.h"
#include "dlmglue.h"
#include "file.h"
#include "inode.h"
#include "mmap.h"

static inline int ocfs2_vm_op_block_sigs(sigset_t *blocked, sigset_t *oldset)
{
	/* The best way to deal with signals in the vm path is
	 * to block them upfront, rather than allowing the
	 * locking paths to return -ERESTARTSYS. */
	sigfillset(blocked);

	/* We should technically never get a bad return value
	 * from sigprocmask */
	return sigprocmask(SIG_BLOCK, blocked, oldset);
}

static inline int ocfs2_vm_op_unblock_sigs(sigset_t *oldset)
{
	return sigprocmask(SIG_SETMASK, oldset, NULL);
}

static int ocfs2_fault(struct vm_area_struct *area, struct vm_fault *vmf)
{
	sigset_t blocked, oldset;
	int error, ret;

	mlog_entry("(area=%p, page offset=%lu)\n", area, vmf->pgoff);

	error = ocfs2_vm_op_block_sigs(&blocked, &oldset);
	if (error < 0) {
		mlog_errno(error);
		ret = VM_FAULT_SIGBUS;
		goto out;
	}

	ret = filemap_fault(area, vmf);

	error = ocfs2_vm_op_unblock_sigs(&oldset);
	if (error < 0)
		mlog_errno(error);
out:
	mlog_exit_ptr(vmf->page);
	return ret;
}

static int __ocfs2_page_mkwrite(struct inode *inode, struct buffer_head *di_bh,
				struct page *page)
{
	int ret;
	struct address_space *mapping = inode->i_mapping;
	loff_t pos = page_offset(page);
	unsigned int len = PAGE_CACHE_SIZE;
	pgoff_t last_index;
	struct page *locked_page = NULL;
	void *fsdata;
	loff_t size = i_size_read(inode);

	/*
	 * Another node might have truncated while we were waiting on
	 * cluster locks.
	 */
	last_index = size >> PAGE_CACHE_SHIFT;
	if (page->index > last_index) {
		ret = -EINVAL;
		goto out;
	}

	/*
	 * The i_size check above doesn't catch the case where nodes
	 * truncated and then re-extended the file. We'll re-check the
	 * page mapping after taking the page lock inside of
	 * ocfs2_write_begin_nolock().
	 */
	if (!PageUptodate(page) || page->mapping != inode->i_mapping) {
		ret = -EINVAL;
		goto out;
	}

	/*
	 * Call ocfs2_write_begin() and ocfs2_write_end() to take
	 * advantage of the allocation code there. We pass a write
	 * length of the whole page (chopped to i_size) to make sure
	 * the whole thing is allocated.
	 *
	 * Since we know the page is up to date, we don't have to
	 * worry about ocfs2_write_begin() skipping some buffer reads
	 * because the "write" would invalidate their data.
	 */
	if (page->index == last_index)
		len = size & ~PAGE_CACHE_MASK;

	ret = ocfs2_write_begin_nolock(mapping, pos, len, 0, &locked_page,
				       &fsdata, di_bh, page);
	if (ret) {
		if (ret != -ENOSPC)
			mlog_errno(ret);
		goto out;
	}

	ret = ocfs2_write_end_nolock(mapping, pos, len, len, locked_page,
				     fsdata);
	if (ret < 0) {
		mlog_errno(ret);
		goto out;
	}
	BUG_ON(ret != len);
	ret = 0;
out:
	return ret;
}

static int ocfs2_page_mkwrite(struct vm_area_struct *vma, struct page *page)
{
	struct inode *inode = vma->vm_file->f_path.dentry->d_inode;
	struct buffer_head *di_bh = NULL;
	sigset_t blocked, oldset;
	int ret, ret2;

	ret = ocfs2_vm_op_block_sigs(&blocked, &oldset);
	if (ret < 0) {
		mlog_errno(ret);
		return ret;
	}

	/*
	 * The cluster locks taken will block a truncate from another
	 * node. Taking the data lock will also ensure that we don't
	 * attempt page truncation as part of a downconvert.
	 */
	ret = ocfs2_meta_lock(inode, &di_bh, 1);
	if (ret < 0) {
		mlog_errno(ret);
		goto out;
	}

	/*
	 * The alloc sem should be enough to serialize with
	 * ocfs2_truncate_file() changing i_size as well as any thread
	 * modifying the inode btree.
	 */
	down_write(&OCFS2_I(inode)->ip_alloc_sem);

	ret = __ocfs2_page_mkwrite(inode, di_bh, page);

	up_write(&OCFS2_I(inode)->ip_alloc_sem);

	brelse(di_bh);
	ocfs2_meta_unlock(inode, 1);

out:
	ret2 = ocfs2_vm_op_unblock_sigs(&oldset);
	if (ret2 < 0)
		mlog_errno(ret2);

	return ret;
}

static struct vm_operations_struct ocfs2_file_vm_ops = {
	.fault		= ocfs2_fault,
	.page_mkwrite	= ocfs2_page_mkwrite,
};

int ocfs2_mmap(struct file *file, struct vm_area_struct *vma)
{
	int ret = 0, lock_level = 0;

	ret = ocfs2_meta_lock_atime(file->f_dentry->d_inode,
				    file->f_vfsmnt, &lock_level);
	if (ret < 0) {
		mlog_errno(ret);
		goto out;
	}
	ocfs2_meta_unlock(file->f_dentry->d_inode, lock_level);
out:
	vma->vm_ops = &ocfs2_file_vm_ops;
	vma->vm_flags |= VM_CAN_NONLINEAR;
	return 0;
}