file.c 18.4 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 22
#include <linux/ext2_fs.h>
#include <linux/crc32.h>
23
#include <linux/writeback.h>
D
David Teigland 已提交
24
#include <asm/uaccess.h>
25 26
#include <linux/dlm.h>
#include <linux/dlm_plock.h>
D
David Teigland 已提交
27 28

#include "gfs2.h"
29
#include "incore.h"
D
David Teigland 已提交
30 31 32 33 34 35 36 37 38 39
#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"
40
#include "util.h"
41
#include "eaops.h"
D
David Teigland 已提交
42 43 44 45 46 47 48 49 50 51 52 53 54 55 56

/**
 * 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)
{
57
	struct gfs2_inode *ip = GFS2_I(file->f_mapping->host);
D
David Teigland 已提交
58 59 60 61 62 63 64
	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
65
			error = generic_file_llseek_unlocked(file, offset, origin);
D
David Teigland 已提交
66 67 68
			gfs2_glock_dq_uninit(&i_gh);
		}
	} else
A
Andi Kleen 已提交
69
		error = generic_file_llseek_unlocked(file, offset, origin);
D
David Teigland 已提交
70 71 72 73 74

	return error;
}

/**
75
 * gfs2_readdir - Read directory entries from a directory
D
David Teigland 已提交
76 77 78 79 80 81 82
 * @file: The directory to read from
 * @dirent: Buffer for dirents
 * @filldir: Function used to do the copying
 *
 * Returns: errno
 */

83
static int gfs2_readdir(struct file *file, void *dirent, filldir_t filldir)
D
David Teigland 已提交
84
{
85
	struct inode *dir = file->f_mapping->host;
86
	struct gfs2_inode *dip = GFS2_I(dir);
D
David Teigland 已提交
87
	struct gfs2_holder d_gh;
88
	u64 offset = file->f_pos;
D
David Teigland 已提交
89 90
	int error;

91 92
	gfs2_holder_init(dip->i_gl, LM_ST_SHARED, 0, &d_gh);
	error = gfs2_glock_nq(&d_gh);
D
David Teigland 已提交
93 94 95 96 97
	if (error) {
		gfs2_holder_uninit(&d_gh);
		return error;
	}

98
	error = gfs2_dir_read(dir, &offset, dirent, filldir);
D
David Teigland 已提交
99 100 101 102 103 104 105 106

	gfs2_glock_dq_uninit(&d_gh);

	file->f_pos = offset;

	return error;
}

107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127
/**
 * 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
128

129 130 131 132 133 134
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,
135
	[14] = GFS2_DIF_INHERIT_JDATA,
136 137
};

138 139 140 141 142 143 144
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,
145
};
146

147
static int gfs2_get_flags(struct file *filp, u32 __user *ptr)
148
{
J
Josef Sipek 已提交
149
	struct inode *inode = filp->f_path.dentry->d_inode;
150
	struct gfs2_inode *ip = GFS2_I(inode);
151 152
	struct gfs2_holder gh;
	int error;
153
	u32 fsflags;
154

155 156
	gfs2_holder_init(ip->i_gl, LM_ST_SHARED, 0, &gh);
	error = gfs2_glock_nq(&gh);
157 158
	if (error)
		return error;
159

160 161
	fsflags = fsflags_cvt(gfs2_to_fsflags, ip->i_diskflags);
	if (!S_ISDIR(inode->i_mode) && ip->i_diskflags & GFS2_DIF_JDATA)
162
		fsflags |= FS_JOURNAL_DATA_FL;
163
	if (put_user(fsflags, ptr))
164 165
		error = -EFAULT;

166
	gfs2_glock_dq(&gh);
167 168 169 170
	gfs2_holder_uninit(&gh);
	return error;
}

171 172 173 174 175 176
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);
177
	if (ip->i_diskflags & GFS2_DIF_IMMUTABLE)
178
		flags |= S_IMMUTABLE;
179
	if (ip->i_diskflags & GFS2_DIF_APPENDONLY)
180
		flags |= S_APPEND;
181
	if (ip->i_diskflags & GFS2_DIF_NOATIME)
182
		flags |= S_NOATIME;
183
	if (ip->i_diskflags & GFS2_DIF_SYNC)
184 185 186 187
		flags |= S_SYNC;
	inode->i_flags = flags;
}

188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203
/* 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
 *
 */
204
static int do_gfs2_set_flags(struct file *filp, u32 reqflags, u32 mask)
205
{
J
Josef Sipek 已提交
206
	struct inode *inode = filp->f_path.dentry->d_inode;
207 208
	struct gfs2_inode *ip = GFS2_I(inode);
	struct gfs2_sbd *sdp = GFS2_SB(inode);
209 210 211
	struct buffer_head *bh;
	struct gfs2_holder gh;
	int error;
212
	u32 new_flags, flags;
213

M
Miklos Szeredi 已提交
214
	error = mnt_want_write(filp->f_path.mnt);
215
	if (error)
216 217
		return error;

M
Miklos Szeredi 已提交
218 219 220 221
	error = gfs2_glock_nq_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &gh);
	if (error)
		goto out_drop_write;

222
	flags = ip->i_diskflags;
223
	new_flags = (flags & ~mask) | (reqflags & mask);
224 225 226 227 228 229 230 231 232 233 234 235
	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;
236
	if (((new_flags ^ flags) & GFS2_DIF_IMMUTABLE) &&
237
	    !capable(CAP_LINUX_IMMUTABLE))
238
		goto out;
239
	if (!IS_IMMUTABLE(inode)) {
M
Miklos Szeredi 已提交
240
		error = gfs2_permission(inode, MAY_WRITE);
241 242 243
		if (error)
			goto out;
	}
244 245 246 247 248 249 250 251 252 253
	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;
	}
254
	error = gfs2_trans_begin(sdp, RES_DINODE, 0);
255 256
	if (error)
		goto out;
257 258 259
	error = gfs2_meta_inode_buffer(ip, &bh);
	if (error)
		goto out_trans_end;
260
	gfs2_trans_add_bh(ip->i_gl, bh, 1);
261
	ip->i_diskflags = new_flags;
262
	gfs2_dinode_out(ip, bh->b_data);
263
	brelse(bh);
264
	gfs2_set_inode_flags(inode);
265
	gfs2_set_aops(inode);
266 267
out_trans_end:
	gfs2_trans_end(sdp);
268 269
out:
	gfs2_glock_dq_uninit(&gh);
M
Miklos Szeredi 已提交
270 271
out_drop_write:
	mnt_drop_write(filp->f_path.mnt);
272 273 274
	return error;
}

275
static int gfs2_set_flags(struct file *filp, u32 __user *ptr)
276
{
277
	struct inode *inode = filp->f_path.dentry->d_inode;
278 279
	u32 fsflags, gfsflags;
	if (get_user(fsflags, ptr))
280
		return -EFAULT;
281
	gfsflags = fsflags_cvt(fsflags_to_gfs2, fsflags);
282 283 284 285 286 287
	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);
288 289
}

290
static long gfs2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
291 292
{
	switch(cmd) {
293
	case FS_IOC_GETFLAGS:
294
		return gfs2_get_flags(filp, (u32 __user *)arg);
295
	case FS_IOC_SETFLAGS:
296
		return gfs2_set_flags(filp, (u32 __user *)arg);
297 298 299 300
	}
	return -ENOTTY;
}

301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320
/**
 * 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;
321
		gfs2_block_map(inode, lblock, &bh, 1);
322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338
		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.
 */

339
static int gfs2_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf)
340
{
341
	struct page *page = vmf->page;
342 343 344 345
	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;
346
	u64 pos = page->index << PAGE_CACHE_SHIFT;
347 348 349 350 351 352
	unsigned int data_blocks, ind_blocks, rblocks;
	int alloc_required = 0;
	struct gfs2_holder gh;
	struct gfs2_alloc *al;
	int ret;

353 354
	gfs2_holder_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &gh);
	ret = gfs2_glock_nq(&gh);
355 356 357
	if (ret)
		goto out;

358 359 360
	set_bit(GLF_DIRTY, &ip->i_gl->gl_flags);
	set_bit(GIF_SW_PAGED, &ip->i_flags);

361 362 363
	ret = gfs2_write_alloc_required(ip, pos, PAGE_CACHE_SIZE, &alloc_required);
	if (ret || !alloc_required)
		goto out_unlock;
364
	ret = -ENOMEM;
365
	al = gfs2_alloc_get(ip);
366 367 368
	if (al == NULL)
		goto out_unlock;

369
	ret = gfs2_quota_lock_check(ip);
370 371
	if (ret)
		goto out_alloc_put;
372
	gfs2_write_calc_reserv(ip, PAGE_CACHE_SIZE, &data_blocks, &ind_blocks);
373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391
	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;
	if (ind_blocks || data_blocks)
		rblocks += RES_STATFS + RES_QUOTA;
	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;
392
	ret = 0;
393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414
	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);
415 416 417
	if (ret == -ENOMEM)
		ret = VM_FAULT_OOM;
	else if (ret)
418
		ret = VM_FAULT_SIGBUS;
419 420 421 422 423 424 425 426
	return ret;
}

static struct vm_operations_struct gfs2_vm_ops = {
	.fault = filemap_fault,
	.page_mkwrite = gfs2_page_mkwrite,
};

D
David Teigland 已提交
427 428 429 430 431
/**
 * gfs2_mmap -
 * @file: The file to map
 * @vma: The VMA which described the mapping
 *
432 433 434 435 436
 * 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
437 438 439 440
 */

static int gfs2_mmap(struct file *file, struct vm_area_struct *vma)
{
441
	struct gfs2_inode *ip = GFS2_I(file->f_mapping->host);
D
David Teigland 已提交
442

443 444 445
	if (!(file->f_flags & O_NOATIME)) {
		struct gfs2_holder i_gh;
		int error;
D
David Teigland 已提交
446

447 448 449 450 451 452
		gfs2_holder_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &i_gh);
		error = gfs2_glock_nq(&i_gh);
		file_accessed(file);
		if (error == 0)
			gfs2_glock_dq_uninit(&i_gh);
	}
453
	vma->vm_ops = &gfs2_vm_ops;
454
	vma->vm_flags |= VM_CAN_NONLINEAR;
D
David Teigland 已提交
455

456
	return 0;
D
David Teigland 已提交
457 458 459 460 461 462 463 464 465 466 467 468
}

/**
 * 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)
{
469
	struct gfs2_inode *ip = GFS2_I(inode);
D
David Teigland 已提交
470 471 472 473 474 475 476 477
	struct gfs2_holder i_gh;
	struct gfs2_file *fp;
	int error;

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

478
	mutex_init(&fp->f_fl_mutex);
D
David Teigland 已提交
479

480
	gfs2_assert_warn(GFS2_SB(inode), !file->private_data);
481
	file->private_data = fp;
D
David Teigland 已提交
482

483
	if (S_ISREG(ip->i_inode.i_mode)) {
D
David Teigland 已提交
484 485 486 487 488 489
		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) &&
490
		    ip->i_disksize > MAX_NON_LFS) {
491
			error = -EOVERFLOW;
D
David Teigland 已提交
492 493 494 495 496 497 498 499
			goto fail_gunlock;
		}

		gfs2_glock_dq_uninit(&i_gh);
	}

	return 0;

500
fail_gunlock:
D
David Teigland 已提交
501
	gfs2_glock_dq_uninit(&i_gh);
502
fail:
503
	file->private_data = NULL;
D
David Teigland 已提交
504 505 506 507 508 509 510 511 512 513 514 515 516 517
	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)
{
518
	struct gfs2_sbd *sdp = inode->i_sb->s_fs_info;
D
David Teigland 已提交
519 520
	struct gfs2_file *fp;

521 522
	fp = file->private_data;
	file->private_data = NULL;
D
David Teigland 已提交
523 524 525 526 527 528 529 530 531 532 533 534 535 536

	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
 *
537 538 539 540 541 542
 * 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.
 *
543 544 545 546 547
 * 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
548 549 550 551 552
 * Returns: errno
 */

static int gfs2_fsync(struct file *file, struct dentry *dentry, int datasync)
{
553 554 555 556 557 558 559 560
	struct inode *inode = dentry->d_inode;
	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
561

562 563
	if (sync_state != 0) {
		if (!datasync)
564
			ret = write_inode_now(inode, 0);
D
David Teigland 已提交
565

566 567 568 569 570
		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
571 572
}

573 574
#ifdef CONFIG_GFS2_FS_LOCKING_DLM

575 576 577 578 579 580
/**
 * gfs2_setlease - acquire/release a file lease
 * @file: the file pointer
 * @arg: lease type
 * @fl: file lock
 *
581 582 583 584
 * 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.
 *
585 586 587 588 589
 * Returns: errno
 */

static int gfs2_setlease(struct file *file, long arg, struct file_lock **fl)
{
590
	return -EINVAL;
591 592
}

D
David Teigland 已提交
593 594 595 596 597 598 599 600 601 602 603
/**
 * 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)
{
604 605
	struct gfs2_inode *ip = GFS2_I(file->f_mapping->host);
	struct gfs2_sbd *sdp = GFS2_SB(file->f_mapping->host);
606
	struct lm_lockstruct *ls = &sdp->sd_lockstruct;
D
David Teigland 已提交
607 608 609

	if (!(fl->fl_flags & FL_POSIX))
		return -ENOLCK;
610
	if (__mandatory_lock(&ip->i_inode))
D
David Teigland 已提交
611 612
		return -ENOLCK;

M
Marc Eshel 已提交
613 614 615 616 617
	if (cmd == F_CANCELLK) {
		/* Hack: */
		cmd = F_SETLK;
		fl->fl_type = F_UNLCK;
	}
618 619
	if (unlikely(test_bit(SDF_SHUTDOWN, &sdp->sd_flags)))
		return -EIO;
D
David Teigland 已提交
620
	if (IS_GETLK(cmd))
621
		return dlm_posix_get(ls->ls_dlm, ip->i_no_addr, file, fl);
D
David Teigland 已提交
622
	else if (fl->fl_type == F_UNLCK)
623
		return dlm_posix_unlock(ls->ls_dlm, ip->i_no_addr, file, fl);
D
David Teigland 已提交
624
	else
625
		return dlm_posix_lock(ls->ls_dlm, ip->i_no_addr, file, cmd, fl);
D
David Teigland 已提交
626 627 628 629
}

static int do_flock(struct file *file, int cmd, struct file_lock *fl)
{
630
	struct gfs2_file *fp = file->private_data;
D
David Teigland 已提交
631
	struct gfs2_holder *fl_gh = &fp->f_fl_gh;
J
Josef Sipek 已提交
632
	struct gfs2_inode *ip = GFS2_I(file->f_path.dentry->d_inode);
D
David Teigland 已提交
633 634 635 636 637 638
	struct gfs2_glock *gl;
	unsigned int state;
	int flags;
	int error = 0;

	state = (fl->fl_type == F_WRLCK) ? LM_ST_EXCLUSIVE : LM_ST_SHARED;
639
	flags = (IS_SETLKW(cmd) ? 0 : LM_FLAG_TRY) | GL_EXACT | GL_NOCACHE;
D
David Teigland 已提交
640

641
	mutex_lock(&fp->f_fl_mutex);
D
David Teigland 已提交
642 643 644 645 646 647

	gl = fl_gh->gh_gl;
	if (gl) {
		if (fl_gh->gh_state == state)
			goto out;
		flock_lock_file_wait(file,
648
				     &(struct file_lock){.fl_type = F_UNLCK});
649 650
		gfs2_glock_dq_wait(fl_gh);
		gfs2_holder_reinit(state, flags, fl_gh);
D
David Teigland 已提交
651
	} else {
652 653
		error = gfs2_glock_get(GFS2_SB(&ip->i_inode), ip->i_no_addr,
				       &gfs2_flock_glops, CREATE, &gl);
D
David Teigland 已提交
654 655
		if (error)
			goto out;
656 657
		gfs2_holder_init(gl, state, flags, fl_gh);
		gfs2_glock_put(gl);
D
David Teigland 已提交
658 659 660 661 662 663 664 665
	}
	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);
666
		gfs2_assert_warn(GFS2_SB(&ip->i_inode), !error);
D
David Teigland 已提交
667 668
	}

669
out:
670
	mutex_unlock(&fp->f_fl_mutex);
D
David Teigland 已提交
671 672 673 674 675
	return error;
}

static void do_unflock(struct file *file, struct file_lock *fl)
{
676
	struct gfs2_file *fp = file->private_data;
D
David Teigland 已提交
677 678
	struct gfs2_holder *fl_gh = &fp->f_fl_gh;

679
	mutex_lock(&fp->f_fl_mutex);
D
David Teigland 已提交
680 681 682
	flock_lock_file_wait(file, fl);
	if (fl_gh->gh_gl)
		gfs2_glock_dq_uninit(fl_gh);
683
	mutex_unlock(&fp->f_fl_mutex);
D
David Teigland 已提交
684 685 686 687 688 689 690 691 692 693 694 695 696
}

/**
 * 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)
{
697
	struct gfs2_inode *ip = GFS2_I(file->f_mapping->host);
D
David Teigland 已提交
698 699 700

	if (!(fl->fl_flags & FL_FLOCK))
		return -ENOLCK;
701
	if (__mandatory_lock(&ip->i_inode))
D
David Teigland 已提交
702 703 704 705 706
		return -ENOLCK;

	if (fl->fl_type == F_UNLCK) {
		do_unflock(file, fl);
		return 0;
707
	} else {
D
David Teigland 已提交
708
		return do_flock(file, cmd, fl);
709
	}
D
David Teigland 已提交
710 711
}

712
const struct file_operations gfs2_file_fops = {
713
	.llseek		= gfs2_llseek,
714
	.read		= do_sync_read,
715
	.aio_read	= generic_file_aio_read,
716
	.write		= do_sync_write,
717 718 719 720 721 722 723 724 725 726
	.aio_write	= generic_file_aio_write,
	.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,
727
	.setlease	= gfs2_setlease,
D
David Teigland 已提交
728 729
};

730
const struct file_operations gfs2_dir_fops = {
731 732 733 734 735 736 737
	.readdir	= gfs2_readdir,
	.unlocked_ioctl	= gfs2_ioctl,
	.open		= gfs2_open,
	.release	= gfs2_close,
	.fsync		= gfs2_fsync,
	.lock		= gfs2_lock,
	.flock		= gfs2_flock,
D
David Teigland 已提交
738 739
};

740 741
#endif /* CONFIG_GFS2_FS_LOCKING_DLM */

742
const struct file_operations gfs2_file_fops_nolock = {
743 744 745 746 747 748 749 750 751 752 753 754
	.llseek		= gfs2_llseek,
	.read		= do_sync_read,
	.aio_read	= generic_file_aio_read,
	.write		= do_sync_write,
	.aio_write	= generic_file_aio_write,
	.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,
755
	.setlease	= generic_setlease,
756 757
};

758
const struct file_operations gfs2_dir_fops_nolock = {
759 760 761 762 763 764 765
	.readdir	= gfs2_readdir,
	.unlocked_ioctl	= gfs2_ioctl,
	.open		= gfs2_open,
	.release	= gfs2_close,
	.fsync		= gfs2_fsync,
};