file.c 25.7 KB
Newer Older
D
David Teigland 已提交
1 2
/*
 * Copyright (C) Sistina Software, Inc.  1997-2003 All rights reserved.
3
 * Copyright (C) 2004-2006 Red Hat, Inc.  All rights reserved.
D
David Teigland 已提交
4 5 6
 *
 * This copyrighted material is made available to anyone wishing to use,
 * modify, copy, or redistribute it subject to the terms and conditions
7
 * of the GNU General Public License version 2.
D
David Teigland 已提交
8 9 10 11 12 13 14 15 16 17
 */

#include <linux/slab.h>
#include <linux/spinlock.h>
#include <linux/completion.h>
#include <linux/buffer_head.h>
#include <linux/pagemap.h>
#include <linux/uio.h>
#include <linux/blkdev.h>
#include <linux/mm.h>
M
Miklos Szeredi 已提交
18
#include <linux/mount.h>
19
#include <linux/fs.h>
20
#include <linux/gfs2_ondisk.h>
21
#include <linux/ext2_fs.h>
22 23
#include <linux/falloc.h>
#include <linux/swap.h>
24
#include <linux/crc32.h>
25
#include <linux/writeback.h>
D
David Teigland 已提交
26
#include <asm/uaccess.h>
27 28
#include <linux/dlm.h>
#include <linux/dlm_plock.h>
D
David Teigland 已提交
29 30

#include "gfs2.h"
31
#include "incore.h"
D
David Teigland 已提交
32 33 34 35 36 37 38 39 40 41
#include "bmap.h"
#include "dir.h"
#include "glock.h"
#include "glops.h"
#include "inode.h"
#include "log.h"
#include "meta_io.h"
#include "quota.h"
#include "rgrp.h"
#include "trans.h"
42
#include "util.h"
D
David Teigland 已提交
43 44 45 46 47 48 49 50 51 52 53 54 55 56 57

/**
 * gfs2_llseek - seek to a location in a file
 * @file: the file
 * @offset: the offset
 * @origin: Where to seek from (SEEK_SET, SEEK_CUR, or SEEK_END)
 *
 * SEEK_END requires the glock for the file because it references the
 * file's size.
 *
 * Returns: The new offset, or errno
 */

static loff_t gfs2_llseek(struct file *file, loff_t offset, int origin)
{
58
	struct gfs2_inode *ip = GFS2_I(file->f_mapping->host);
D
David Teigland 已提交
59 60 61 62 63 64 65
	struct gfs2_holder i_gh;
	loff_t error;

	if (origin == 2) {
		error = gfs2_glock_nq_init(ip->i_gl, LM_ST_SHARED, LM_FLAG_ANY,
					   &i_gh);
		if (!error) {
A
Andi Kleen 已提交
66
			error = generic_file_llseek_unlocked(file, offset, origin);
D
David Teigland 已提交
67 68 69
			gfs2_glock_dq_uninit(&i_gh);
		}
	} else
A
Andi Kleen 已提交
70
		error = generic_file_llseek_unlocked(file, offset, origin);
D
David Teigland 已提交
71 72 73 74 75

	return error;
}

/**
76
 * gfs2_readdir - Read directory entries from a directory
D
David Teigland 已提交
77 78 79 80 81 82 83
 * @file: The directory to read from
 * @dirent: Buffer for dirents
 * @filldir: Function used to do the copying
 *
 * Returns: errno
 */

84
static int gfs2_readdir(struct file *file, void *dirent, filldir_t filldir)
D
David Teigland 已提交
85
{
86
	struct inode *dir = file->f_mapping->host;
87
	struct gfs2_inode *dip = GFS2_I(dir);
D
David Teigland 已提交
88
	struct gfs2_holder d_gh;
89
	u64 offset = file->f_pos;
D
David Teigland 已提交
90 91
	int error;

92 93
	gfs2_holder_init(dip->i_gl, LM_ST_SHARED, 0, &d_gh);
	error = gfs2_glock_nq(&d_gh);
D
David Teigland 已提交
94 95 96 97 98
	if (error) {
		gfs2_holder_uninit(&d_gh);
		return error;
	}

99
	error = gfs2_dir_read(dir, &offset, dirent, filldir);
D
David Teigland 已提交
100 101 102 103 104 105 106 107

	gfs2_glock_dq_uninit(&d_gh);

	file->f_pos = offset;

	return error;
}

108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128
/**
 * fsflags_cvt
 * @table: A table of 32 u32 flags
 * @val: a 32 bit value to convert
 *
 * This function can be used to convert between fsflags values and
 * GFS2's own flags values.
 *
 * Returns: the converted flags
 */
static u32 fsflags_cvt(const u32 *table, u32 val)
{
	u32 res = 0;
	while(val) {
		if (val & 1)
			res |= *table;
		table++;
		val >>= 1;
	}
	return res;
}
D
David Teigland 已提交
129

130 131 132 133 134 135
static const u32 fsflags_to_gfs2[32] = {
	[3] = GFS2_DIF_SYNC,
	[4] = GFS2_DIF_IMMUTABLE,
	[5] = GFS2_DIF_APPENDONLY,
	[7] = GFS2_DIF_NOATIME,
	[12] = GFS2_DIF_EXHASH,
136
	[14] = GFS2_DIF_INHERIT_JDATA,
137 138
};

139 140 141 142 143 144 145
static const u32 gfs2_to_fsflags[32] = {
	[gfs2fl_Sync] = FS_SYNC_FL,
	[gfs2fl_Immutable] = FS_IMMUTABLE_FL,
	[gfs2fl_AppendOnly] = FS_APPEND_FL,
	[gfs2fl_NoAtime] = FS_NOATIME_FL,
	[gfs2fl_ExHash] = FS_INDEX_FL,
	[gfs2fl_InheritJdata] = FS_JOURNAL_DATA_FL,
146
};
147

148
static int gfs2_get_flags(struct file *filp, u32 __user *ptr)
149
{
J
Josef Sipek 已提交
150
	struct inode *inode = filp->f_path.dentry->d_inode;
151
	struct gfs2_inode *ip = GFS2_I(inode);
152 153
	struct gfs2_holder gh;
	int error;
154
	u32 fsflags;
155

156 157
	gfs2_holder_init(ip->i_gl, LM_ST_SHARED, 0, &gh);
	error = gfs2_glock_nq(&gh);
158 159
	if (error)
		return error;
160

161 162
	fsflags = fsflags_cvt(gfs2_to_fsflags, ip->i_diskflags);
	if (!S_ISDIR(inode->i_mode) && ip->i_diskflags & GFS2_DIF_JDATA)
163
		fsflags |= FS_JOURNAL_DATA_FL;
164
	if (put_user(fsflags, ptr))
165 166
		error = -EFAULT;

167
	gfs2_glock_dq(&gh);
168 169 170 171
	gfs2_holder_uninit(&gh);
	return error;
}

172 173 174 175 176 177
void gfs2_set_inode_flags(struct inode *inode)
{
	struct gfs2_inode *ip = GFS2_I(inode);
	unsigned int flags = inode->i_flags;

	flags &= ~(S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC);
178
	if (ip->i_diskflags & GFS2_DIF_IMMUTABLE)
179
		flags |= S_IMMUTABLE;
180
	if (ip->i_diskflags & GFS2_DIF_APPENDONLY)
181
		flags |= S_APPEND;
182
	if (ip->i_diskflags & GFS2_DIF_NOATIME)
183
		flags |= S_NOATIME;
184
	if (ip->i_diskflags & GFS2_DIF_SYNC)
185 186 187 188
		flags |= S_SYNC;
	inode->i_flags = flags;
}

189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204
/* Flags that can be set by user space */
#define GFS2_FLAGS_USER_SET (GFS2_DIF_JDATA|			\
			     GFS2_DIF_IMMUTABLE|		\
			     GFS2_DIF_APPENDONLY|		\
			     GFS2_DIF_NOATIME|			\
			     GFS2_DIF_SYNC|			\
			     GFS2_DIF_SYSTEM|			\
			     GFS2_DIF_INHERIT_JDATA)

/**
 * gfs2_set_flags - set flags on an inode
 * @inode: The inode
 * @flags: The flags to set
 * @mask: Indicates which flags are valid
 *
 */
205
static int do_gfs2_set_flags(struct file *filp, u32 reqflags, u32 mask)
206
{
J
Josef Sipek 已提交
207
	struct inode *inode = filp->f_path.dentry->d_inode;
208 209
	struct gfs2_inode *ip = GFS2_I(inode);
	struct gfs2_sbd *sdp = GFS2_SB(inode);
210 211 212
	struct buffer_head *bh;
	struct gfs2_holder gh;
	int error;
213
	u32 new_flags, flags;
214

M
Miklos Szeredi 已提交
215
	error = mnt_want_write(filp->f_path.mnt);
216
	if (error)
217 218
		return error;

M
Miklos Szeredi 已提交
219 220 221 222
	error = gfs2_glock_nq_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &gh);
	if (error)
		goto out_drop_write;

223 224 225 226 227
	error = -EACCES;
	if (!is_owner_or_cap(inode))
		goto out;

	error = 0;
228
	flags = ip->i_diskflags;
229
	new_flags = (flags & ~mask) | (reqflags & mask);
230 231 232 233 234 235 236 237 238 239 240 241
	if ((new_flags ^ flags) == 0)
		goto out;

	error = -EINVAL;
	if ((new_flags ^ flags) & ~GFS2_FLAGS_USER_SET)
		goto out;

	error = -EPERM;
	if (IS_IMMUTABLE(inode) && (new_flags & GFS2_DIF_IMMUTABLE))
		goto out;
	if (IS_APPEND(inode) && (new_flags & GFS2_DIF_APPENDONLY))
		goto out;
242
	if (((new_flags ^ flags) & GFS2_DIF_IMMUTABLE) &&
243
	    !capable(CAP_LINUX_IMMUTABLE))
244
		goto out;
245
	if (!IS_IMMUTABLE(inode)) {
246
		error = gfs2_permission(inode, MAY_WRITE, 0);
247 248 249
		if (error)
			goto out;
	}
250 251 252 253 254 255 256 257 258 259
	if ((flags ^ new_flags) & GFS2_DIF_JDATA) {
		if (flags & GFS2_DIF_JDATA)
			gfs2_log_flush(sdp, ip->i_gl);
		error = filemap_fdatawrite(inode->i_mapping);
		if (error)
			goto out;
		error = filemap_fdatawait(inode->i_mapping);
		if (error)
			goto out;
	}
260
	error = gfs2_trans_begin(sdp, RES_DINODE, 0);
261 262
	if (error)
		goto out;
263 264 265
	error = gfs2_meta_inode_buffer(ip, &bh);
	if (error)
		goto out_trans_end;
266
	gfs2_trans_add_bh(ip->i_gl, bh, 1);
267
	ip->i_diskflags = new_flags;
268
	gfs2_dinode_out(ip, bh->b_data);
269
	brelse(bh);
270
	gfs2_set_inode_flags(inode);
271
	gfs2_set_aops(inode);
272 273
out_trans_end:
	gfs2_trans_end(sdp);
274 275
out:
	gfs2_glock_dq_uninit(&gh);
M
Miklos Szeredi 已提交
276 277
out_drop_write:
	mnt_drop_write(filp->f_path.mnt);
278 279 280
	return error;
}

281
static int gfs2_set_flags(struct file *filp, u32 __user *ptr)
282
{
283
	struct inode *inode = filp->f_path.dentry->d_inode;
284
	u32 fsflags, gfsflags;
285

286
	if (get_user(fsflags, ptr))
287
		return -EFAULT;
288

289
	gfsflags = fsflags_cvt(fsflags_to_gfs2, fsflags);
290 291 292 293 294 295
	if (!S_ISDIR(inode->i_mode)) {
		if (gfsflags & GFS2_DIF_INHERIT_JDATA)
			gfsflags ^= (GFS2_DIF_JDATA | GFS2_DIF_INHERIT_JDATA);
		return do_gfs2_set_flags(filp, gfsflags, ~0);
	}
	return do_gfs2_set_flags(filp, gfsflags, ~GFS2_DIF_JDATA);
296 297
}

298
static long gfs2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
299 300
{
	switch(cmd) {
301
	case FS_IOC_GETFLAGS:
302
		return gfs2_get_flags(filp, (u32 __user *)arg);
303
	case FS_IOC_SETFLAGS:
304
		return gfs2_set_flags(filp, (u32 __user *)arg);
305 306 307 308
	}
	return -ENOTTY;
}

309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328
/**
 * gfs2_allocate_page_backing - Use bmap to allocate blocks
 * @page: The (locked) page to allocate backing for
 *
 * We try to allocate all the blocks required for the page in
 * one go. This might fail for various reasons, so we keep
 * trying until all the blocks to back this page are allocated.
 * If some of the blocks are already allocated, thats ok too.
 */

static int gfs2_allocate_page_backing(struct page *page)
{
	struct inode *inode = page->mapping->host;
	struct buffer_head bh;
	unsigned long size = PAGE_CACHE_SIZE;
	u64 lblock = page->index << (PAGE_CACHE_SHIFT - inode->i_blkbits);

	do {
		bh.b_state = 0;
		bh.b_size = size;
329
		gfs2_block_map(inode, lblock, &bh, 1);
330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346
		if (!buffer_mapped(&bh))
			return -EIO;
		size -= bh.b_size;
		lblock += (bh.b_size >> inode->i_blkbits);
	} while(size > 0);
	return 0;
}

/**
 * gfs2_page_mkwrite - Make a shared, mmap()ed, page writable
 * @vma: The virtual memory area
 * @page: The page which is about to become writable
 *
 * When the page becomes writable, we need to ensure that we have
 * blocks allocated on disk to back that page.
 */

347
static int gfs2_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf)
348
{
349
	struct page *page = vmf->page;
350 351 352 353
	struct inode *inode = vma->vm_file->f_path.dentry->d_inode;
	struct gfs2_inode *ip = GFS2_I(inode);
	struct gfs2_sbd *sdp = GFS2_SB(inode);
	unsigned long last_index;
354
	u64 pos = page->index << PAGE_CACHE_SHIFT;
355 356 357 358 359
	unsigned int data_blocks, ind_blocks, rblocks;
	struct gfs2_holder gh;
	struct gfs2_alloc *al;
	int ret;

360 361
	gfs2_holder_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &gh);
	ret = gfs2_glock_nq(&gh);
362 363 364
	if (ret)
		goto out;

365 366 367
	set_bit(GLF_DIRTY, &ip->i_gl->gl_flags);
	set_bit(GIF_SW_PAGED, &ip->i_flags);

368
	if (!gfs2_write_alloc_required(ip, pos, PAGE_CACHE_SIZE))
369
		goto out_unlock;
370
	ret = -ENOMEM;
371
	al = gfs2_alloc_get(ip);
372 373 374
	if (al == NULL)
		goto out_unlock;

375
	ret = gfs2_quota_lock_check(ip);
376 377
	if (ret)
		goto out_alloc_put;
378
	gfs2_write_calc_reserv(ip, PAGE_CACHE_SIZE, &data_blocks, &ind_blocks);
379 380 381 382 383 384 385 386
	al->al_requested = data_blocks + ind_blocks;
	ret = gfs2_inplace_reserve(ip);
	if (ret)
		goto out_quota_unlock;

	rblocks = RES_DINODE + ind_blocks;
	if (gfs2_is_jdata(ip))
		rblocks += data_blocks ? data_blocks : 1;
387
	if (ind_blocks || data_blocks) {
388
		rblocks += RES_STATFS + RES_QUOTA;
389 390
		rblocks += gfs2_rg_blocks(al);
	}
391 392 393 394 395 396 397 398 399
	ret = gfs2_trans_begin(sdp, rblocks, 0);
	if (ret)
		goto out_trans_fail;

	lock_page(page);
	ret = -EINVAL;
	last_index = ip->i_inode.i_size >> PAGE_CACHE_SHIFT;
	if (page->index > last_index)
		goto out_unlock_page;
400
	ret = 0;
401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422
	if (!PageUptodate(page) || page->mapping != ip->i_inode.i_mapping)
		goto out_unlock_page;
	if (gfs2_is_stuffed(ip)) {
		ret = gfs2_unstuff_dinode(ip, page);
		if (ret)
			goto out_unlock_page;
	}
	ret = gfs2_allocate_page_backing(page);

out_unlock_page:
	unlock_page(page);
	gfs2_trans_end(sdp);
out_trans_fail:
	gfs2_inplace_release(ip);
out_quota_unlock:
	gfs2_quota_unlock(ip);
out_alloc_put:
	gfs2_alloc_put(ip);
out_unlock:
	gfs2_glock_dq(&gh);
out:
	gfs2_holder_uninit(&gh);
423 424 425
	if (ret == -ENOMEM)
		ret = VM_FAULT_OOM;
	else if (ret)
426
		ret = VM_FAULT_SIGBUS;
427 428 429
	return ret;
}

430
static const struct vm_operations_struct gfs2_vm_ops = {
431 432 433 434
	.fault = filemap_fault,
	.page_mkwrite = gfs2_page_mkwrite,
};

D
David Teigland 已提交
435 436 437 438 439
/**
 * gfs2_mmap -
 * @file: The file to map
 * @vma: The VMA which described the mapping
 *
440 441 442 443 444
 * There is no need to get a lock here unless we should be updating
 * atime. We ignore any locking errors since the only consequence is
 * a missed atime update (which will just be deferred until later).
 *
 * Returns: 0
D
David Teigland 已提交
445 446 447 448
 */

static int gfs2_mmap(struct file *file, struct vm_area_struct *vma)
{
449
	struct gfs2_inode *ip = GFS2_I(file->f_mapping->host);
D
David Teigland 已提交
450

451 452
	if (!(file->f_flags & O_NOATIME) &&
	    !IS_NOATIME(&ip->i_inode)) {
453 454
		struct gfs2_holder i_gh;
		int error;
D
David Teigland 已提交
455

456
		gfs2_holder_init(ip->i_gl, LM_ST_SHARED, LM_FLAG_ANY, &i_gh);
457
		error = gfs2_glock_nq(&i_gh);
458 459 460 461 462 463 464
		if (error == 0) {
			file_accessed(file);
			gfs2_glock_dq(&i_gh);
		}
		gfs2_holder_uninit(&i_gh);
		if (error)
			return error;
465
	}
466
	vma->vm_ops = &gfs2_vm_ops;
467
	vma->vm_flags |= VM_CAN_NONLINEAR;
D
David Teigland 已提交
468

469
	return 0;
D
David Teigland 已提交
470 471 472 473 474 475 476 477 478 479 480 481
}

/**
 * gfs2_open - open a file
 * @inode: the inode to open
 * @file: the struct file for this opening
 *
 * Returns: errno
 */

static int gfs2_open(struct inode *inode, struct file *file)
{
482
	struct gfs2_inode *ip = GFS2_I(inode);
D
David Teigland 已提交
483 484 485 486 487 488 489 490
	struct gfs2_holder i_gh;
	struct gfs2_file *fp;
	int error;

	fp = kzalloc(sizeof(struct gfs2_file), GFP_KERNEL);
	if (!fp)
		return -ENOMEM;

491
	mutex_init(&fp->f_fl_mutex);
D
David Teigland 已提交
492

493
	gfs2_assert_warn(GFS2_SB(inode), !file->private_data);
494
	file->private_data = fp;
D
David Teigland 已提交
495

496
	if (S_ISREG(ip->i_inode.i_mode)) {
D
David Teigland 已提交
497 498 499 500 501 502
		error = gfs2_glock_nq_init(ip->i_gl, LM_ST_SHARED, LM_FLAG_ANY,
					   &i_gh);
		if (error)
			goto fail;

		if (!(file->f_flags & O_LARGEFILE) &&
S
Steven Whitehouse 已提交
503
		    i_size_read(inode) > MAX_NON_LFS) {
504
			error = -EOVERFLOW;
D
David Teigland 已提交
505 506 507 508 509 510 511 512
			goto fail_gunlock;
		}

		gfs2_glock_dq_uninit(&i_gh);
	}

	return 0;

513
fail_gunlock:
D
David Teigland 已提交
514
	gfs2_glock_dq_uninit(&i_gh);
515
fail:
516
	file->private_data = NULL;
D
David Teigland 已提交
517 518 519 520 521 522 523 524 525 526 527 528 529 530
	kfree(fp);
	return error;
}

/**
 * gfs2_close - called to close a struct file
 * @inode: the inode the struct file belongs to
 * @file: the struct file being closed
 *
 * Returns: errno
 */

static int gfs2_close(struct inode *inode, struct file *file)
{
531
	struct gfs2_sbd *sdp = inode->i_sb->s_fs_info;
D
David Teigland 已提交
532 533
	struct gfs2_file *fp;

534 535
	fp = file->private_data;
	file->private_data = NULL;
D
David Teigland 已提交
536 537 538 539 540 541 542 543 544 545 546 547 548 549

	if (gfs2_assert_warn(sdp, fp))
		return -EIO;

	kfree(fp);

	return 0;
}

/**
 * gfs2_fsync - sync the dirty data for a file (across the cluster)
 * @file: the file that points to the dentry (we ignore this)
 * @dentry: the dentry that points to the inode to sync
 *
550 551 552 553 554 555
 * The VFS will flush "normal" data for us. We only need to worry
 * about metadata here. For journaled data, we just do a log flush
 * as we can't avoid it. Otherwise we can just bale out if datasync
 * is set. For stuffed inodes we must flush the log in order to
 * ensure that all data is on disk.
 *
556 557 558 559 560
 * The call to write_inode_now() is there to write back metadata and
 * the inode itself. It does also try and write the data, but thats
 * (hopefully) a no-op due to the VFS having already called filemap_fdatawrite()
 * for us.
 *
D
David Teigland 已提交
561 562 563
 * Returns: errno
 */

564
static int gfs2_fsync(struct file *file, int datasync)
D
David Teigland 已提交
565
{
566
	struct inode *inode = file->f_mapping->host;
567 568 569 570 571 572 573
	int sync_state = inode->i_state & (I_DIRTY_SYNC|I_DIRTY_DATASYNC);
	int ret = 0;

	if (gfs2_is_jdata(GFS2_I(inode))) {
		gfs2_log_flush(GFS2_SB(inode), GFS2_I(inode)->i_gl);
		return 0;
	}
D
David Teigland 已提交
574

575 576
	if (sync_state != 0) {
		if (!datasync)
577
			ret = write_inode_now(inode, 0);
D
David Teigland 已提交
578

579 580 581 582 583
		if (gfs2_is_stuffed(GFS2_I(inode)))
			gfs2_log_flush(GFS2_SB(inode), GFS2_I(inode)->i_gl);
	}

	return ret;
D
David Teigland 已提交
584 585
}

586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619
/**
 * gfs2_file_aio_write - Perform a write to a file
 * @iocb: The io context
 * @iov: The data to write
 * @nr_segs: Number of @iov segments
 * @pos: The file position
 *
 * We have to do a lock/unlock here to refresh the inode size for
 * O_APPEND writes, otherwise we can land up writing at the wrong
 * offset. There is still a race, but provided the app is using its
 * own file locking, this will make O_APPEND work as expected.
 *
 */

static ssize_t gfs2_file_aio_write(struct kiocb *iocb, const struct iovec *iov,
				   unsigned long nr_segs, loff_t pos)
{
	struct file *file = iocb->ki_filp;

	if (file->f_flags & O_APPEND) {
		struct dentry *dentry = file->f_dentry;
		struct gfs2_inode *ip = GFS2_I(dentry->d_inode);
		struct gfs2_holder gh;
		int ret;

		ret = gfs2_glock_nq_init(ip->i_gl, LM_ST_SHARED, 0, &gh);
		if (ret)
			return ret;
		gfs2_glock_dq_uninit(&gh);
	}

	return generic_file_aio_write(iocb, iov, nr_segs, pos);
}

620 621 622 623 624
static void empty_write_end(struct page *page, unsigned from,
			   unsigned to)
{
	struct gfs2_inode *ip = GFS2_I(page->mapping->host);

625
	zero_user(page, from, to-from);
626 627 628 629 630 631 632 633
	mark_page_accessed(page);

	if (!gfs2_is_writeback(ip))
		gfs2_page_add_databufs(ip, page, from, to);

	block_commit_write(page, from, to);
}

634
static int needs_empty_write(sector_t block, struct inode *inode)
635 636
{
	int error;
637
	struct buffer_head bh_map = { .b_state = 0, .b_blocknr = 0 };
638

639 640 641 642 643 644
	bh_map.b_size = 1 << inode->i_blkbits;
	error = gfs2_block_map(inode, block, &bh_map, 0);
	if (unlikely(error))
		return error;
	return !buffer_mapped(&bh_map);
}
645

646 647 648 649 650 651
static int write_empty_blocks(struct page *page, unsigned from, unsigned to)
{
	struct inode *inode = page->mapping->host;
	unsigned start, end, next, blksize;
	sector_t block = page->index << (PAGE_CACHE_SHIFT - inode->i_blkbits);
	int ret;
652

653
	blksize = 1 << inode->i_blkbits;
654 655
	next = end = 0;
	while (next < from) {
656 657
		next += blksize;
		block++;
658 659 660
	}
	start = next;
	do {
661 662 663 664 665
		next += blksize;
		ret = needs_empty_write(block, inode);
		if (unlikely(ret < 0))
			return ret;
		if (ret == 0) {
666
			if (end) {
667 668 669 670
				ret = __block_write_begin(page, start, end - start,
							  gfs2_block_map);
				if (unlikely(ret))
					return ret;
671 672 673 674 675 676 677
				empty_write_end(page, start, end);
				end = 0;
			}
			start = next;
		}
		else
			end = next;
678
		block++;
679 680 681
	} while (next < to);

	if (end) {
682 683 684
		ret = __block_write_begin(page, start, end - start, gfs2_block_map);
		if (unlikely(ret))
			return ret;
685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879
		empty_write_end(page, start, end);
	}

	return 0;
}

static int fallocate_chunk(struct inode *inode, loff_t offset, loff_t len,
			   int mode)
{
	struct gfs2_inode *ip = GFS2_I(inode);
	struct buffer_head *dibh;
	int error;
	u64 start = offset >> PAGE_CACHE_SHIFT;
	unsigned int start_offset = offset & ~PAGE_CACHE_MASK;
	u64 end = (offset + len - 1) >> PAGE_CACHE_SHIFT;
	pgoff_t curr;
	struct page *page;
	unsigned int end_offset = (offset + len) & ~PAGE_CACHE_MASK;
	unsigned int from, to;

	if (!end_offset)
		end_offset = PAGE_CACHE_SIZE;

	error = gfs2_meta_inode_buffer(ip, &dibh);
	if (unlikely(error))
		goto out;

	gfs2_trans_add_bh(ip->i_gl, dibh, 1);

	if (gfs2_is_stuffed(ip)) {
		error = gfs2_unstuff_dinode(ip, NULL);
		if (unlikely(error))
			goto out;
	}

	curr = start;
	offset = start << PAGE_CACHE_SHIFT;
	from = start_offset;
	to = PAGE_CACHE_SIZE;
	while (curr <= end) {
		page = grab_cache_page_write_begin(inode->i_mapping, curr,
						   AOP_FLAG_NOFS);
		if (unlikely(!page)) {
			error = -ENOMEM;
			goto out;
		}

		if (curr == end)
			to = end_offset;
		error = write_empty_blocks(page, from, to);
		if (!error && offset + to > inode->i_size &&
		    !(mode & FALLOC_FL_KEEP_SIZE)) {
			i_size_write(inode, offset + to);
		}
		unlock_page(page);
		page_cache_release(page);
		if (error)
			goto out;
		curr++;
		offset += PAGE_CACHE_SIZE;
		from = 0;
	}

	gfs2_dinode_out(ip, dibh->b_data);
	mark_inode_dirty(inode);

	brelse(dibh);

out:
	return error;
}

static void calc_max_reserv(struct gfs2_inode *ip, loff_t max, loff_t *len,
			    unsigned int *data_blocks, unsigned int *ind_blocks)
{
	const struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
	unsigned int max_blocks = ip->i_alloc->al_rgd->rd_free_clone;
	unsigned int tmp, max_data = max_blocks - 3 * (sdp->sd_max_height - 1);

	for (tmp = max_data; tmp > sdp->sd_diptrs;) {
		tmp = DIV_ROUND_UP(tmp, sdp->sd_inptrs);
		max_data -= tmp;
	}
	/* This calculation isn't the exact reverse of gfs2_write_calc_reserve,
	   so it might end up with fewer data blocks */
	if (max_data <= *data_blocks)
		return;
	*data_blocks = max_data;
	*ind_blocks = max_blocks - max_data;
	*len = ((loff_t)max_data - 3) << sdp->sd_sb.sb_bsize_shift;
	if (*len > max) {
		*len = max;
		gfs2_write_calc_reserv(ip, max, data_blocks, ind_blocks);
	}
}

static long gfs2_fallocate(struct file *file, int mode, loff_t offset,
			   loff_t len)
{
	struct inode *inode = file->f_path.dentry->d_inode;
	struct gfs2_sbd *sdp = GFS2_SB(inode);
	struct gfs2_inode *ip = GFS2_I(inode);
	unsigned int data_blocks = 0, ind_blocks = 0, rblocks;
	loff_t bytes, max_bytes;
	struct gfs2_alloc *al;
	int error;
	loff_t next = (offset + len - 1) >> sdp->sd_sb.sb_bsize_shift;
	next = (next + 1) << sdp->sd_sb.sb_bsize_shift;

	/* We only support the FALLOC_FL_KEEP_SIZE mode */
	if (mode & ~FALLOC_FL_KEEP_SIZE)
		return -EOPNOTSUPP;

	offset = (offset >> sdp->sd_sb.sb_bsize_shift) <<
		 sdp->sd_sb.sb_bsize_shift;

	len = next - offset;
	bytes = sdp->sd_max_rg_data * sdp->sd_sb.sb_bsize / 2;
	if (!bytes)
		bytes = UINT_MAX;

	gfs2_holder_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &ip->i_gh);
	error = gfs2_glock_nq(&ip->i_gh);
	if (unlikely(error))
		goto out_uninit;

	if (!gfs2_write_alloc_required(ip, offset, len))
		goto out_unlock;

	while (len > 0) {
		if (len < bytes)
			bytes = len;
		al = gfs2_alloc_get(ip);
		if (!al) {
			error = -ENOMEM;
			goto out_unlock;
		}

		error = gfs2_quota_lock_check(ip);
		if (error)
			goto out_alloc_put;

retry:
		gfs2_write_calc_reserv(ip, bytes, &data_blocks, &ind_blocks);

		al->al_requested = data_blocks + ind_blocks;
		error = gfs2_inplace_reserve(ip);
		if (error) {
			if (error == -ENOSPC && bytes > sdp->sd_sb.sb_bsize) {
				bytes >>= 1;
				goto retry;
			}
			goto out_qunlock;
		}
		max_bytes = bytes;
		calc_max_reserv(ip, len, &max_bytes, &data_blocks, &ind_blocks);
		al->al_requested = data_blocks + ind_blocks;

		rblocks = RES_DINODE + ind_blocks + RES_STATFS + RES_QUOTA +
			  RES_RG_HDR + gfs2_rg_blocks(al);
		if (gfs2_is_jdata(ip))
			rblocks += data_blocks ? data_blocks : 1;

		error = gfs2_trans_begin(sdp, rblocks,
					 PAGE_CACHE_SIZE/sdp->sd_sb.sb_bsize);
		if (error)
			goto out_trans_fail;

		error = fallocate_chunk(inode, offset, max_bytes, mode);
		gfs2_trans_end(sdp);

		if (error)
			goto out_trans_fail;

		len -= max_bytes;
		offset += max_bytes;
		gfs2_inplace_release(ip);
		gfs2_quota_unlock(ip);
		gfs2_alloc_put(ip);
	}
	goto out_unlock;

out_trans_fail:
	gfs2_inplace_release(ip);
out_qunlock:
	gfs2_quota_unlock(ip);
out_alloc_put:
	gfs2_alloc_put(ip);
out_unlock:
	gfs2_glock_dq(&ip->i_gh);
out_uninit:
	gfs2_holder_uninit(&ip->i_gh);
	return error;
}

880 881
#ifdef CONFIG_GFS2_FS_LOCKING_DLM

882 883 884 885 886 887
/**
 * gfs2_setlease - acquire/release a file lease
 * @file: the file pointer
 * @arg: lease type
 * @fl: file lock
 *
888 889 890 891
 * We don't currently have a way to enforce a lease across the whole
 * cluster; until we do, disable leases (by just returning -EINVAL),
 * unless the administrator has requested purely local locking.
 *
892 893
 * Locking: called under lock_flocks
 *
894 895 896 897 898
 * Returns: errno
 */

static int gfs2_setlease(struct file *file, long arg, struct file_lock **fl)
{
899
	return -EINVAL;
900 901
}

D
David Teigland 已提交
902 903 904 905 906 907 908 909 910 911 912
/**
 * gfs2_lock - acquire/release a posix lock on a file
 * @file: the file pointer
 * @cmd: either modify or retrieve lock state, possibly wait
 * @fl: type and range of lock
 *
 * Returns: errno
 */

static int gfs2_lock(struct file *file, int cmd, struct file_lock *fl)
{
913 914
	struct gfs2_inode *ip = GFS2_I(file->f_mapping->host);
	struct gfs2_sbd *sdp = GFS2_SB(file->f_mapping->host);
915
	struct lm_lockstruct *ls = &sdp->sd_lockstruct;
D
David Teigland 已提交
916 917 918

	if (!(fl->fl_flags & FL_POSIX))
		return -ENOLCK;
919
	if (__mandatory_lock(&ip->i_inode) && fl->fl_type != F_UNLCK)
D
David Teigland 已提交
920 921
		return -ENOLCK;

M
Marc Eshel 已提交
922 923 924 925 926
	if (cmd == F_CANCELLK) {
		/* Hack: */
		cmd = F_SETLK;
		fl->fl_type = F_UNLCK;
	}
927 928
	if (unlikely(test_bit(SDF_SHUTDOWN, &sdp->sd_flags)))
		return -EIO;
D
David Teigland 已提交
929
	if (IS_GETLK(cmd))
930
		return dlm_posix_get(ls->ls_dlm, ip->i_no_addr, file, fl);
D
David Teigland 已提交
931
	else if (fl->fl_type == F_UNLCK)
932
		return dlm_posix_unlock(ls->ls_dlm, ip->i_no_addr, file, fl);
D
David Teigland 已提交
933
	else
934
		return dlm_posix_lock(ls->ls_dlm, ip->i_no_addr, file, cmd, fl);
D
David Teigland 已提交
935 936 937 938
}

static int do_flock(struct file *file, int cmd, struct file_lock *fl)
{
939
	struct gfs2_file *fp = file->private_data;
D
David Teigland 已提交
940
	struct gfs2_holder *fl_gh = &fp->f_fl_gh;
J
Josef Sipek 已提交
941
	struct gfs2_inode *ip = GFS2_I(file->f_path.dentry->d_inode);
D
David Teigland 已提交
942 943 944 945 946 947
	struct gfs2_glock *gl;
	unsigned int state;
	int flags;
	int error = 0;

	state = (fl->fl_type == F_WRLCK) ? LM_ST_EXCLUSIVE : LM_ST_SHARED;
948
	flags = (IS_SETLKW(cmd) ? 0 : LM_FLAG_TRY) | GL_EXACT | GL_NOCACHE;
D
David Teigland 已提交
949

950
	mutex_lock(&fp->f_fl_mutex);
D
David Teigland 已提交
951 952 953 954 955 956

	gl = fl_gh->gh_gl;
	if (gl) {
		if (fl_gh->gh_state == state)
			goto out;
		flock_lock_file_wait(file,
957
				     &(struct file_lock){.fl_type = F_UNLCK});
958 959
		gfs2_glock_dq_wait(fl_gh);
		gfs2_holder_reinit(state, flags, fl_gh);
D
David Teigland 已提交
960
	} else {
961 962
		error = gfs2_glock_get(GFS2_SB(&ip->i_inode), ip->i_no_addr,
				       &gfs2_flock_glops, CREATE, &gl);
D
David Teigland 已提交
963 964
		if (error)
			goto out;
965 966
		gfs2_holder_init(gl, state, flags, fl_gh);
		gfs2_glock_put(gl);
D
David Teigland 已提交
967 968 969 970 971 972 973 974
	}
	error = gfs2_glock_nq(fl_gh);
	if (error) {
		gfs2_holder_uninit(fl_gh);
		if (error == GLR_TRYFAILED)
			error = -EAGAIN;
	} else {
		error = flock_lock_file_wait(file, fl);
975
		gfs2_assert_warn(GFS2_SB(&ip->i_inode), !error);
D
David Teigland 已提交
976 977
	}

978
out:
979
	mutex_unlock(&fp->f_fl_mutex);
D
David Teigland 已提交
980 981 982 983 984
	return error;
}

static void do_unflock(struct file *file, struct file_lock *fl)
{
985
	struct gfs2_file *fp = file->private_data;
D
David Teigland 已提交
986 987
	struct gfs2_holder *fl_gh = &fp->f_fl_gh;

988
	mutex_lock(&fp->f_fl_mutex);
D
David Teigland 已提交
989
	flock_lock_file_wait(file, fl);
990 991 992 993
	if (fl_gh->gh_gl) {
		gfs2_glock_dq_wait(fl_gh);
		gfs2_holder_uninit(fl_gh);
	}
994
	mutex_unlock(&fp->f_fl_mutex);
D
David Teigland 已提交
995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009
}

/**
 * gfs2_flock - acquire/release a flock lock on a file
 * @file: the file pointer
 * @cmd: either modify or retrieve lock state, possibly wait
 * @fl: type and range of lock
 *
 * Returns: errno
 */

static int gfs2_flock(struct file *file, int cmd, struct file_lock *fl)
{
	if (!(fl->fl_flags & FL_FLOCK))
		return -ENOLCK;
1010 1011
	if (fl->fl_type & LOCK_MAND)
		return -EOPNOTSUPP;
D
David Teigland 已提交
1012 1013 1014 1015

	if (fl->fl_type == F_UNLCK) {
		do_unflock(file, fl);
		return 0;
1016
	} else {
D
David Teigland 已提交
1017
		return do_flock(file, cmd, fl);
1018
	}
D
David Teigland 已提交
1019 1020
}

1021
const struct file_operations gfs2_file_fops = {
1022
	.llseek		= gfs2_llseek,
1023
	.read		= do_sync_read,
1024
	.aio_read	= generic_file_aio_read,
1025
	.write		= do_sync_write,
1026
	.aio_write	= gfs2_file_aio_write,
1027 1028 1029 1030 1031 1032 1033 1034 1035
	.unlocked_ioctl	= gfs2_ioctl,
	.mmap		= gfs2_mmap,
	.open		= gfs2_open,
	.release	= gfs2_close,
	.fsync		= gfs2_fsync,
	.lock		= gfs2_lock,
	.flock		= gfs2_flock,
	.splice_read	= generic_file_splice_read,
	.splice_write	= generic_file_splice_write,
1036
	.setlease	= gfs2_setlease,
1037
	.fallocate	= gfs2_fallocate,
D
David Teigland 已提交
1038 1039
};

1040
const struct file_operations gfs2_dir_fops = {
1041 1042 1043 1044 1045 1046 1047
	.readdir	= gfs2_readdir,
	.unlocked_ioctl	= gfs2_ioctl,
	.open		= gfs2_open,
	.release	= gfs2_close,
	.fsync		= gfs2_fsync,
	.lock		= gfs2_lock,
	.flock		= gfs2_flock,
1048
	.llseek		= default_llseek,
D
David Teigland 已提交
1049 1050
};

1051 1052
#endif /* CONFIG_GFS2_FS_LOCKING_DLM */

1053
const struct file_operations gfs2_file_fops_nolock = {
1054 1055 1056 1057
	.llseek		= gfs2_llseek,
	.read		= do_sync_read,
	.aio_read	= generic_file_aio_read,
	.write		= do_sync_write,
1058
	.aio_write	= gfs2_file_aio_write,
1059 1060 1061 1062 1063 1064 1065
	.unlocked_ioctl	= gfs2_ioctl,
	.mmap		= gfs2_mmap,
	.open		= gfs2_open,
	.release	= gfs2_close,
	.fsync		= gfs2_fsync,
	.splice_read	= generic_file_splice_read,
	.splice_write	= generic_file_splice_write,
1066
	.setlease	= generic_setlease,
1067
	.fallocate	= gfs2_fallocate,
1068 1069
};

1070
const struct file_operations gfs2_dir_fops_nolock = {
1071 1072 1073 1074 1075
	.readdir	= gfs2_readdir,
	.unlocked_ioctl	= gfs2_ioctl,
	.open		= gfs2_open,
	.release	= gfs2_close,
	.fsync		= gfs2_fsync,
1076
	.llseek		= default_llseek,
1077 1078
};