file.c 19.3 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"
D
David Teigland 已提交
41 42 43 44 45 46 47 48 49 50 51 52 53 54 55

/**
 * 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)
{
56
	struct gfs2_inode *ip = GFS2_I(file->f_mapping->host);
D
David Teigland 已提交
57 58 59 60 61 62 63
	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
64
			error = generic_file_llseek_unlocked(file, offset, origin);
D
David Teigland 已提交
65 66 67
			gfs2_glock_dq_uninit(&i_gh);
		}
	} else
A
Andi Kleen 已提交
68
		error = generic_file_llseek_unlocked(file, offset, origin);
D
David Teigland 已提交
69 70 71 72 73

	return error;
}

/**
74
 * gfs2_readdir - Read directory entries from a directory
D
David Teigland 已提交
75 76 77 78 79 80 81
 * @file: The directory to read from
 * @dirent: Buffer for dirents
 * @filldir: Function used to do the copying
 *
 * Returns: errno
 */

82
static int gfs2_readdir(struct file *file, void *dirent, filldir_t filldir)
D
David Teigland 已提交
83
{
84
	struct inode *dir = file->f_mapping->host;
85
	struct gfs2_inode *dip = GFS2_I(dir);
D
David Teigland 已提交
86
	struct gfs2_holder d_gh;
87
	u64 offset = file->f_pos;
D
David Teigland 已提交
88 89
	int error;

90 91
	gfs2_holder_init(dip->i_gl, LM_ST_SHARED, 0, &d_gh);
	error = gfs2_glock_nq(&d_gh);
D
David Teigland 已提交
92 93 94 95 96
	if (error) {
		gfs2_holder_uninit(&d_gh);
		return error;
	}

97
	error = gfs2_dir_read(dir, &offset, dirent, filldir);
D
David Teigland 已提交
98 99 100 101 102 103 104 105

	gfs2_glock_dq_uninit(&d_gh);

	file->f_pos = offset;

	return error;
}

106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126
/**
 * 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
127

128 129 130 131 132 133
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,
134
	[14] = GFS2_DIF_INHERIT_JDATA,
135 136
};

137 138 139 140 141 142 143
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,
144
};
145

146
static int gfs2_get_flags(struct file *filp, u32 __user *ptr)
147
{
J
Josef Sipek 已提交
148
	struct inode *inode = filp->f_path.dentry->d_inode;
149
	struct gfs2_inode *ip = GFS2_I(inode);
150 151
	struct gfs2_holder gh;
	int error;
152
	u32 fsflags;
153

154 155
	gfs2_holder_init(ip->i_gl, LM_ST_SHARED, 0, &gh);
	error = gfs2_glock_nq(&gh);
156 157
	if (error)
		return error;
158

159 160
	fsflags = fsflags_cvt(gfs2_to_fsflags, ip->i_diskflags);
	if (!S_ISDIR(inode->i_mode) && ip->i_diskflags & GFS2_DIF_JDATA)
161
		fsflags |= FS_JOURNAL_DATA_FL;
162
	if (put_user(fsflags, ptr))
163 164
		error = -EFAULT;

165
	gfs2_glock_dq(&gh);
166 167 168 169
	gfs2_holder_uninit(&gh);
	return error;
}

170 171 172 173 174 175
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);
176
	if (ip->i_diskflags & GFS2_DIF_IMMUTABLE)
177
		flags |= S_IMMUTABLE;
178
	if (ip->i_diskflags & GFS2_DIF_APPENDONLY)
179
		flags |= S_APPEND;
180
	if (ip->i_diskflags & GFS2_DIF_NOATIME)
181
		flags |= S_NOATIME;
182
	if (ip->i_diskflags & GFS2_DIF_SYNC)
183 184 185 186
		flags |= S_SYNC;
	inode->i_flags = flags;
}

187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202
/* 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
 *
 */
203
static int do_gfs2_set_flags(struct file *filp, u32 reqflags, u32 mask)
204
{
J
Josef Sipek 已提交
205
	struct inode *inode = filp->f_path.dentry->d_inode;
206 207
	struct gfs2_inode *ip = GFS2_I(inode);
	struct gfs2_sbd *sdp = GFS2_SB(inode);
208 209 210
	struct buffer_head *bh;
	struct gfs2_holder gh;
	int error;
211
	u32 new_flags, flags;
212

M
Miklos Szeredi 已提交
213
	error = mnt_want_write(filp->f_path.mnt);
214
	if (error)
215 216
		return error;

M
Miklos Szeredi 已提交
217 218 219 220
	error = gfs2_glock_nq_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &gh);
	if (error)
		goto out_drop_write;

221
	flags = ip->i_diskflags;
222
	new_flags = (flags & ~mask) | (reqflags & mask);
223 224 225 226 227 228 229 230 231 232 233 234
	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;
235
	if (((new_flags ^ flags) & GFS2_DIF_IMMUTABLE) &&
236
	    !capable(CAP_LINUX_IMMUTABLE))
237
		goto out;
238
	if (!IS_IMMUTABLE(inode)) {
M
Miklos Szeredi 已提交
239
		error = gfs2_permission(inode, MAY_WRITE);
240 241 242
		if (error)
			goto out;
	}
243 244 245 246 247 248 249 250 251 252
	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;
	}
253
	error = gfs2_trans_begin(sdp, RES_DINODE, 0);
254 255
	if (error)
		goto out;
256 257 258
	error = gfs2_meta_inode_buffer(ip, &bh);
	if (error)
		goto out_trans_end;
259
	gfs2_trans_add_bh(ip->i_gl, bh, 1);
260
	ip->i_diskflags = new_flags;
261
	gfs2_dinode_out(ip, bh->b_data);
262
	brelse(bh);
263
	gfs2_set_inode_flags(inode);
264
	gfs2_set_aops(inode);
265 266
out_trans_end:
	gfs2_trans_end(sdp);
267 268
out:
	gfs2_glock_dq_uninit(&gh);
M
Miklos Szeredi 已提交
269 270
out_drop_write:
	mnt_drop_write(filp->f_path.mnt);
271 272 273
	return error;
}

274
static int gfs2_set_flags(struct file *filp, u32 __user *ptr)
275
{
276
	struct inode *inode = filp->f_path.dentry->d_inode;
277 278
	u32 fsflags, gfsflags;
	if (get_user(fsflags, ptr))
279
		return -EFAULT;
280
	gfsflags = fsflags_cvt(fsflags_to_gfs2, fsflags);
281 282 283 284 285 286
	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);
287 288
}

289
static long gfs2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
290 291
{
	switch(cmd) {
292
	case FS_IOC_GETFLAGS:
293
		return gfs2_get_flags(filp, (u32 __user *)arg);
294
	case FS_IOC_SETFLAGS:
295
		return gfs2_set_flags(filp, (u32 __user *)arg);
296 297 298 299
	}
	return -ENOTTY;
}

300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319
/**
 * 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;
320
		gfs2_block_map(inode, lblock, &bh, 1);
321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337
		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.
 */

338
static int gfs2_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf)
339
{
340
	struct page *page = vmf->page;
341 342 343 344
	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;
345
	u64 pos = page->index << PAGE_CACHE_SHIFT;
346 347 348 349 350 351
	unsigned int data_blocks, ind_blocks, rblocks;
	int alloc_required = 0;
	struct gfs2_holder gh;
	struct gfs2_alloc *al;
	int ret;

352 353
	gfs2_holder_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &gh);
	ret = gfs2_glock_nq(&gh);
354 355 356
	if (ret)
		goto out;

357 358 359
	set_bit(GLF_DIRTY, &ip->i_gl->gl_flags);
	set_bit(GIF_SW_PAGED, &ip->i_flags);

360 361 362
	ret = gfs2_write_alloc_required(ip, pos, PAGE_CACHE_SIZE, &alloc_required);
	if (ret || !alloc_required)
		goto out_unlock;
363
	ret = -ENOMEM;
364
	al = gfs2_alloc_get(ip);
365 366 367
	if (al == NULL)
		goto out_unlock;

368
	ret = gfs2_quota_lock_check(ip);
369 370
	if (ret)
		goto out_alloc_put;
371
	gfs2_write_calc_reserv(ip, PAGE_CACHE_SIZE, &data_blocks, &ind_blocks);
372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390
	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;
391
	ret = 0;
392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413
	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);
414 415 416
	if (ret == -ENOMEM)
		ret = VM_FAULT_OOM;
	else if (ret)
417
		ret = VM_FAULT_SIGBUS;
418 419 420
	return ret;
}

421
static const struct vm_operations_struct gfs2_vm_ops = {
422 423 424 425
	.fault = filemap_fault,
	.page_mkwrite = gfs2_page_mkwrite,
};

D
David Teigland 已提交
426 427 428 429 430
/**
 * gfs2_mmap -
 * @file: The file to map
 * @vma: The VMA which described the mapping
 *
431 432 433 434 435
 * 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
436 437 438 439
 */

static int gfs2_mmap(struct file *file, struct vm_area_struct *vma)
{
440
	struct gfs2_inode *ip = GFS2_I(file->f_mapping->host);
D
David Teigland 已提交
441

442 443 444
	if (!(file->f_flags & O_NOATIME)) {
		struct gfs2_holder i_gh;
		int error;
D
David Teigland 已提交
445

446 447 448 449 450 451
		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);
	}
452
	vma->vm_ops = &gfs2_vm_ops;
453
	vma->vm_flags |= VM_CAN_NONLINEAR;
D
David Teigland 已提交
454

455
	return 0;
D
David Teigland 已提交
456 457 458 459 460 461 462 463 464 465 466 467
}

/**
 * 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)
{
468
	struct gfs2_inode *ip = GFS2_I(inode);
D
David Teigland 已提交
469 470 471 472 473 474 475 476
	struct gfs2_holder i_gh;
	struct gfs2_file *fp;
	int error;

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

477
	mutex_init(&fp->f_fl_mutex);
D
David Teigland 已提交
478

479
	gfs2_assert_warn(GFS2_SB(inode), !file->private_data);
480
	file->private_data = fp;
D
David Teigland 已提交
481

482
	if (S_ISREG(ip->i_inode.i_mode)) {
D
David Teigland 已提交
483 484 485 486 487 488
		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) &&
489
		    ip->i_disksize > MAX_NON_LFS) {
490
			error = -EOVERFLOW;
D
David Teigland 已提交
491 492 493 494 495 496 497 498
			goto fail_gunlock;
		}

		gfs2_glock_dq_uninit(&i_gh);
	}

	return 0;

499
fail_gunlock:
D
David Teigland 已提交
500
	gfs2_glock_dq_uninit(&i_gh);
501
fail:
502
	file->private_data = NULL;
D
David Teigland 已提交
503 504 505 506 507 508 509 510 511 512 513 514 515 516
	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)
{
517
	struct gfs2_sbd *sdp = inode->i_sb->s_fs_info;
D
David Teigland 已提交
518 519
	struct gfs2_file *fp;

520 521
	fp = file->private_data;
	file->private_data = NULL;
D
David Teigland 已提交
522 523 524 525 526 527 528 529 530 531 532 533 534 535

	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
 *
536 537 538 539 540 541
 * 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.
 *
542 543 544 545 546
 * 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
547 548 549 550 551
 * Returns: errno
 */

static int gfs2_fsync(struct file *file, struct dentry *dentry, int datasync)
{
552 553 554 555 556 557 558 559
	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
560

561 562
	if (sync_state != 0) {
		if (!datasync)
563
			ret = write_inode_now(inode, 0);
D
David Teigland 已提交
564

565 566 567 568 569
		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
570 571
}

572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605
/**
 * 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);
}

606 607
#ifdef CONFIG_GFS2_FS_LOCKING_DLM

608 609 610 611 612 613
/**
 * gfs2_setlease - acquire/release a file lease
 * @file: the file pointer
 * @arg: lease type
 * @fl: file lock
 *
614 615 616 617
 * 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.
 *
618 619 620 621 622
 * Returns: errno
 */

static int gfs2_setlease(struct file *file, long arg, struct file_lock **fl)
{
623
	return -EINVAL;
624 625
}

D
David Teigland 已提交
626 627 628 629 630 631 632 633 634 635 636
/**
 * 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)
{
637 638
	struct gfs2_inode *ip = GFS2_I(file->f_mapping->host);
	struct gfs2_sbd *sdp = GFS2_SB(file->f_mapping->host);
639
	struct lm_lockstruct *ls = &sdp->sd_lockstruct;
D
David Teigland 已提交
640 641 642

	if (!(fl->fl_flags & FL_POSIX))
		return -ENOLCK;
643
	if (__mandatory_lock(&ip->i_inode))
D
David Teigland 已提交
644 645
		return -ENOLCK;

M
Marc Eshel 已提交
646 647 648 649 650
	if (cmd == F_CANCELLK) {
		/* Hack: */
		cmd = F_SETLK;
		fl->fl_type = F_UNLCK;
	}
651 652
	if (unlikely(test_bit(SDF_SHUTDOWN, &sdp->sd_flags)))
		return -EIO;
D
David Teigland 已提交
653
	if (IS_GETLK(cmd))
654
		return dlm_posix_get(ls->ls_dlm, ip->i_no_addr, file, fl);
D
David Teigland 已提交
655
	else if (fl->fl_type == F_UNLCK)
656
		return dlm_posix_unlock(ls->ls_dlm, ip->i_no_addr, file, fl);
D
David Teigland 已提交
657
	else
658
		return dlm_posix_lock(ls->ls_dlm, ip->i_no_addr, file, cmd, fl);
D
David Teigland 已提交
659 660 661 662
}

static int do_flock(struct file *file, int cmd, struct file_lock *fl)
{
663
	struct gfs2_file *fp = file->private_data;
D
David Teigland 已提交
664
	struct gfs2_holder *fl_gh = &fp->f_fl_gh;
J
Josef Sipek 已提交
665
	struct gfs2_inode *ip = GFS2_I(file->f_path.dentry->d_inode);
D
David Teigland 已提交
666 667 668 669 670 671
	struct gfs2_glock *gl;
	unsigned int state;
	int flags;
	int error = 0;

	state = (fl->fl_type == F_WRLCK) ? LM_ST_EXCLUSIVE : LM_ST_SHARED;
672
	flags = (IS_SETLKW(cmd) ? 0 : LM_FLAG_TRY) | GL_EXACT | GL_NOCACHE;
D
David Teigland 已提交
673

674
	mutex_lock(&fp->f_fl_mutex);
D
David Teigland 已提交
675 676 677 678 679 680

	gl = fl_gh->gh_gl;
	if (gl) {
		if (fl_gh->gh_state == state)
			goto out;
		flock_lock_file_wait(file,
681
				     &(struct file_lock){.fl_type = F_UNLCK});
682 683
		gfs2_glock_dq_wait(fl_gh);
		gfs2_holder_reinit(state, flags, fl_gh);
D
David Teigland 已提交
684
	} else {
685 686
		error = gfs2_glock_get(GFS2_SB(&ip->i_inode), ip->i_no_addr,
				       &gfs2_flock_glops, CREATE, &gl);
D
David Teigland 已提交
687 688
		if (error)
			goto out;
689 690
		gfs2_holder_init(gl, state, flags, fl_gh);
		gfs2_glock_put(gl);
D
David Teigland 已提交
691 692 693 694 695 696 697 698
	}
	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);
699
		gfs2_assert_warn(GFS2_SB(&ip->i_inode), !error);
D
David Teigland 已提交
700 701
	}

702
out:
703
	mutex_unlock(&fp->f_fl_mutex);
D
David Teigland 已提交
704 705 706 707 708
	return error;
}

static void do_unflock(struct file *file, struct file_lock *fl)
{
709
	struct gfs2_file *fp = file->private_data;
D
David Teigland 已提交
710 711
	struct gfs2_holder *fl_gh = &fp->f_fl_gh;

712
	mutex_lock(&fp->f_fl_mutex);
D
David Teigland 已提交
713 714 715
	flock_lock_file_wait(file, fl);
	if (fl_gh->gh_gl)
		gfs2_glock_dq_uninit(fl_gh);
716
	mutex_unlock(&fp->f_fl_mutex);
D
David Teigland 已提交
717 718 719 720 721 722 723 724 725 726 727 728 729 730 731
}

/**
 * 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;
732 733
	if (fl->fl_type & LOCK_MAND)
		return -EOPNOTSUPP;
D
David Teigland 已提交
734 735 736 737

	if (fl->fl_type == F_UNLCK) {
		do_unflock(file, fl);
		return 0;
738
	} else {
D
David Teigland 已提交
739
		return do_flock(file, cmd, fl);
740
	}
D
David Teigland 已提交
741 742
}

743
const struct file_operations gfs2_file_fops = {
744
	.llseek		= gfs2_llseek,
745
	.read		= do_sync_read,
746
	.aio_read	= generic_file_aio_read,
747
	.write		= do_sync_write,
748
	.aio_write	= gfs2_file_aio_write,
749 750 751 752 753 754 755 756 757
	.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,
758
	.setlease	= gfs2_setlease,
D
David Teigland 已提交
759 760
};

761
const struct file_operations gfs2_dir_fops = {
762 763 764 765 766 767 768
	.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
769 770
};

771 772
#endif /* CONFIG_GFS2_FS_LOCKING_DLM */

773
const struct file_operations gfs2_file_fops_nolock = {
774 775 776 777
	.llseek		= gfs2_llseek,
	.read		= do_sync_read,
	.aio_read	= generic_file_aio_read,
	.write		= do_sync_write,
778
	.aio_write	= gfs2_file_aio_write,
779 780 781 782 783 784 785
	.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,
786
	.setlease	= generic_setlease,
787 788
};

789
const struct file_operations gfs2_dir_fops_nolock = {
790 791 792 793 794 795 796
	.readdir	= gfs2_readdir,
	.unlocked_ioctl	= gfs2_ioctl,
	.open		= gfs2_open,
	.release	= gfs2_close,
	.fsync		= gfs2_fsync,
};