ops_file.c 19.2 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 18 19
 */

#include <linux/sched.h>
#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>
#include <linux/smp_lock.h>
20
#include <linux/fs.h>
21
#include <linux/gfs2_ondisk.h>
22 23
#include <linux/ext2_fs.h>
#include <linux/crc32.h>
24
#include <linux/iflags.h>
D
David Teigland 已提交
25 26 27
#include <asm/uaccess.h>

#include "gfs2.h"
28 29
#include "lm_interface.h"
#include "incore.h"
D
David Teigland 已提交
30 31 32 33 34 35 36 37 38 39 40 41 42
#include "bmap.h"
#include "dir.h"
#include "glock.h"
#include "glops.h"
#include "inode.h"
#include "lm.h"
#include "log.h"
#include "meta_io.h"
#include "ops_file.h"
#include "ops_vm.h"
#include "quota.h"
#include "rgrp.h"
#include "trans.h"
43
#include "util.h"
44
#include "eaops.h"
D
David Teigland 已提交
45 46 47 48 49

/* "bad" is for NFS support */
struct filldir_bad_entry {
	char *fbe_name;
	unsigned int fbe_length;
50
	u64 fbe_offset;
D
David Teigland 已提交
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
	struct gfs2_inum fbe_inum;
	unsigned int fbe_type;
};

struct filldir_bad {
	struct gfs2_sbd *fdb_sbd;

	struct filldir_bad_entry *fdb_entry;
	unsigned int fdb_entry_num;
	unsigned int fdb_entry_off;

	char *fdb_name;
	unsigned int fdb_name_size;
	unsigned int fdb_name_off;
};

/* For regular, non-NFS */
struct filldir_reg {
	struct gfs2_sbd *fdr_sbd;
	int fdr_prefetch;

	filldir_t fdr_filldir;
	void *fdr_opaque;
};

76 77 78 79 80 81
/*
 * Most fields left uninitialised to catch anybody who tries to
 * use them. f_flags set to prevent file_accessed() from touching
 * any other part of this. Its use is purely as a flag so that we
 * know (in readpage()) whether or not do to locking.
 */
82
struct file gfs2_internal_file_sentinel = {
83 84 85
	.f_flags = O_NOATIME|O_RDONLY,
};

86 87 88 89 90 91 92 93 94 95 96
static int gfs2_read_actor(read_descriptor_t *desc, struct page *page,
			   unsigned long offset, unsigned long size)
{
	char *kaddr;
	unsigned long count = desc->count;

	if (size > count)
		size = count;

	kaddr = kmap(page);
	memcpy(desc->arg.buf, kaddr + offset, size);
97
	kunmap(page);
98

99 100 101 102
	desc->count = count - size;
	desc->written += size;
	desc->arg.buf += size;
	return size;
103 104 105 106 107
}

int gfs2_internal_read(struct gfs2_inode *ip, struct file_ra_state *ra_state,
		       char *buf, loff_t *pos, unsigned size)
{
108
	struct inode *inode = &ip->i_inode;
109 110 111 112 113
	read_descriptor_t desc;
	desc.written = 0;
	desc.arg.buf = buf;
	desc.count = size;
	desc.error = 0;
114
	do_generic_mapping_read(inode->i_mapping, ra_state,
115
				&gfs2_internal_file_sentinel, pos, &desc,
116
				gfs2_read_actor);
117 118
	return desc.written ? desc.written : desc.error;
}
D
David Teigland 已提交
119 120 121 122 123 124 125 126 127 128 129 130 131 132 133

/**
 * 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)
{
134
	struct gfs2_inode *ip = GFS2_I(file->f_mapping->host);
D
David Teigland 已提交
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
	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) {
			error = remote_llseek(file, offset, origin);
			gfs2_glock_dq_uninit(&i_gh);
		}
	} else
		error = remote_llseek(file, offset, origin);

	return error;
}

/**
 * filldir_reg_func - Report a directory entry to the caller of gfs2_dir_read()
 * @opaque: opaque data used by the function
 * @name: the name of the directory entry
 * @length: the length of the name
 * @offset: the entry's offset in the directory
 * @inum: the inode number the entry points to
 * @type: the type of inode the entry points to
 *
 * Returns: 0 on success, 1 if buffer full
 */

static int filldir_reg_func(void *opaque, const char *name, unsigned int length,
164
			    u64 offset, struct gfs2_inum *inum,
D
David Teigland 已提交
165 166 167 168 169 170 171
			    unsigned int type)
{
	struct filldir_reg *fdr = (struct filldir_reg *)opaque;
	struct gfs2_sbd *sdp = fdr->fdr_sbd;
	int error;

	error = fdr->fdr_filldir(fdr->fdr_opaque, name, length, offset,
S
Steven Whitehouse 已提交
172
				 inum->no_addr, type);
D
David Teigland 已提交
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
	if (error)
		return 1;

	if (fdr->fdr_prefetch && !(length == 1 && *name == '.')) {
		gfs2_glock_prefetch_num(sdp,
				       inum->no_addr, &gfs2_inode_glops,
				       LM_ST_SHARED, LM_FLAG_TRY | LM_FLAG_ANY);
		gfs2_glock_prefetch_num(sdp,
				       inum->no_addr, &gfs2_iopen_glops,
				       LM_ST_SHARED, LM_FLAG_TRY);
	}

	return 0;
}

/**
 * readdir_reg - Read directory entries from a directory
 * @file: The directory to read from
 * @dirent: Buffer for dirents
 * @filldir: Function used to do the copying
 *
 * Returns: errno
 */

static int readdir_reg(struct file *file, void *dirent, filldir_t filldir)
{
199
	struct inode *dir = file->f_mapping->host;
200
	struct gfs2_inode *dip = GFS2_I(dir);
D
David Teigland 已提交
201 202
	struct filldir_reg fdr;
	struct gfs2_holder d_gh;
203
	u64 offset = file->f_pos;
D
David Teigland 已提交
204 205
	int error;

206
	fdr.fdr_sbd = GFS2_SB(dir);
D
David Teigland 已提交
207 208 209 210 211 212 213 214 215 216 217
	fdr.fdr_prefetch = 1;
	fdr.fdr_filldir = filldir;
	fdr.fdr_opaque = dirent;

	gfs2_holder_init(dip->i_gl, LM_ST_SHARED, GL_ATIME, &d_gh);
	error = gfs2_glock_nq_atime(&d_gh);
	if (error) {
		gfs2_holder_uninit(&d_gh);
		return error;
	}

218
	error = gfs2_dir_read(dir, &offset, &fdr, filldir_reg_func);
D
David Teigland 已提交
219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241

	gfs2_glock_dq_uninit(&d_gh);

	file->f_pos = offset;

	return error;
}

/**
 * filldir_bad_func - Report a directory entry to the caller of gfs2_dir_read()
 * @opaque: opaque data used by the function
 * @name: the name of the directory entry
 * @length: the length of the name
 * @offset: the entry's offset in the directory
 * @inum: the inode number the entry points to
 * @type: the type of inode the entry points to
 *
 * For supporting NFS.
 *
 * Returns: 0 on success, 1 if buffer full
 */

static int filldir_bad_func(void *opaque, const char *name, unsigned int length,
242
			    u64 offset, struct gfs2_inum *inum,
D
David Teigland 已提交
243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288
			    unsigned int type)
{
	struct filldir_bad *fdb = (struct filldir_bad *)opaque;
	struct gfs2_sbd *sdp = fdb->fdb_sbd;
	struct filldir_bad_entry *fbe;

	if (fdb->fdb_entry_off == fdb->fdb_entry_num ||
	    fdb->fdb_name_off + length > fdb->fdb_name_size)
		return 1;

	fbe = &fdb->fdb_entry[fdb->fdb_entry_off];
	fbe->fbe_name = fdb->fdb_name + fdb->fdb_name_off;
	memcpy(fbe->fbe_name, name, length);
	fbe->fbe_length = length;
	fbe->fbe_offset = offset;
	fbe->fbe_inum = *inum;
	fbe->fbe_type = type;

	fdb->fdb_entry_off++;
	fdb->fdb_name_off += length;

	if (!(length == 1 && *name == '.')) {
		gfs2_glock_prefetch_num(sdp,
				       inum->no_addr, &gfs2_inode_glops,
				       LM_ST_SHARED, LM_FLAG_TRY | LM_FLAG_ANY);
		gfs2_glock_prefetch_num(sdp,
				       inum->no_addr, &gfs2_iopen_glops,
				       LM_ST_SHARED, LM_FLAG_TRY);
	}

	return 0;
}

/**
 * readdir_bad - Read directory entries from a directory
 * @file: The directory to read from
 * @dirent: Buffer for dirents
 * @filldir: Function used to do the copying
 *
 * For supporting NFS.
 *
 * Returns: errno
 */

static int readdir_bad(struct file *file, void *dirent, filldir_t filldir)
{
289
	struct inode *dir = file->f_mapping->host;
290 291
	struct gfs2_inode *dip = GFS2_I(dir);
	struct gfs2_sbd *sdp = GFS2_SB(dir);
D
David Teigland 已提交
292 293 294 295
	struct filldir_reg fdr;
	unsigned int entries, size;
	struct filldir_bad *fdb;
	struct gfs2_holder d_gh;
296
	u64 offset = file->f_pos;
D
David Teigland 已提交
297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322
	unsigned int x;
	struct filldir_bad_entry *fbe;
	int error;

	entries = gfs2_tune_get(sdp, gt_entries_per_readdir);
	size = sizeof(struct filldir_bad) +
	    entries * (sizeof(struct filldir_bad_entry) + GFS2_FAST_NAME_SIZE);

	fdb = kzalloc(size, GFP_KERNEL);
	if (!fdb)
		return -ENOMEM;

	fdb->fdb_sbd = sdp;
	fdb->fdb_entry = (struct filldir_bad_entry *)(fdb + 1);
	fdb->fdb_entry_num = entries;
	fdb->fdb_name = ((char *)fdb) + sizeof(struct filldir_bad) +
		entries * sizeof(struct filldir_bad_entry);
	fdb->fdb_name_size = entries * GFS2_FAST_NAME_SIZE;

	gfs2_holder_init(dip->i_gl, LM_ST_SHARED, GL_ATIME, &d_gh);
	error = gfs2_glock_nq_atime(&d_gh);
	if (error) {
		gfs2_holder_uninit(&d_gh);
		goto out;
	}

323
	error = gfs2_dir_read(dir, &offset, fdb, filldir_bad_func);
D
David Teigland 已提交
324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347

	gfs2_glock_dq_uninit(&d_gh);

	fdr.fdr_sbd = sdp;
	fdr.fdr_prefetch = 0;
	fdr.fdr_filldir = filldir;
	fdr.fdr_opaque = dirent;

	for (x = 0; x < fdb->fdb_entry_off; x++) {
		fbe = &fdb->fdb_entry[x];

		error = filldir_reg_func(&fdr,
					 fbe->fbe_name, fbe->fbe_length,
					 fbe->fbe_offset,
					 &fbe->fbe_inum, fbe->fbe_type);
		if (error) {
			file->f_pos = fbe->fbe_offset;
			error = 0;
			goto out;
		}
	}

	file->f_pos = offset;

348
out:
D
David Teigland 已提交
349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373
	kfree(fdb);
	return error;
}

/**
 * gfs2_readdir - Read directory entries from a directory
 * @file: The directory to read from
 * @dirent: Buffer for dirents
 * @filldir: Function used to do the copying
 *
 * Returns: errno
 */

static int gfs2_readdir(struct file *file, void *dirent, filldir_t filldir)
{
	int error;

	if (strcmp(current->comm, "nfsd") != 0)
		error = readdir_reg(file, dirent, filldir);
	else
		error = readdir_bad(file, dirent, filldir);

	return error;
}

374 375 376 377 378 379 380 381
static const u32 iflags_to_gfs2[32] = {
	[iflag_Sync] = GFS2_DIF_SYNC,
	[iflag_Immutable] = GFS2_DIF_IMMUTABLE,
	[iflag_Append] = GFS2_DIF_APPENDONLY,
	[iflag_NoAtime] = GFS2_DIF_NOATIME,
	[iflag_Index] = GFS2_DIF_EXHASH,
	[iflag_JournalData] = GFS2_DIF_JDATA,
	[iflag_DirectIO] = GFS2_DIF_DIRECTIO,
382 383
};

384 385 386 387 388 389 390 391
static const u32 gfs2_to_iflags[32] = {
	[gfs2fl_Sync] = IFLAG_SYNC,
	[gfs2fl_Immutable] = IFLAG_IMMUTABLE,
	[gfs2fl_AppendOnly] = IFLAG_APPEND,
	[gfs2fl_NoAtime] = IFLAG_NOATIME,
	[gfs2fl_ExHash] = IFLAG_INDEX,
	[gfs2fl_Jdata] = IFLAG_JOURNAL_DATA,
	[gfs2fl_Directio] = IFLAG_DIRECTIO,
392 393
	[gfs2fl_InheritDirectio] = IFLAG_DIRECTIO,
	[gfs2fl_InheritJdata] = IFLAG_JOURNAL_DATA,
394
};
395

396
static int gfs2_get_flags(struct file *filp, u32 __user *ptr)
397
{
398
	struct inode *inode = filp->f_dentry->d_inode;
399
	struct gfs2_inode *ip = GFS2_I(inode);
400 401
	struct gfs2_holder gh;
	int error;
402
	u32 iflags;
403 404 405 406 407

	gfs2_holder_init(ip->i_gl, LM_ST_SHARED, GL_ATIME, &gh);
	error = gfs2_glock_nq_m_atime(1, &gh);
	if (error)
		return error;
408
	
409 410
	iflags = iflags_cvt(gfs2_to_iflags, ip->i_di.di_flags);
	if (put_user(iflags, ptr))
411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435
		error = -EFAULT;

	gfs2_glock_dq_m(1, &gh);
	gfs2_holder_uninit(&gh);
	return error;
}

/* Flags that can be set by user space */
#define GFS2_FLAGS_USER_SET (GFS2_DIF_JDATA|			\
			     GFS2_DIF_DIRECTIO|			\
			     GFS2_DIF_IMMUTABLE|		\
			     GFS2_DIF_APPENDONLY|		\
			     GFS2_DIF_NOATIME|			\
			     GFS2_DIF_SYNC|			\
			     GFS2_DIF_SYSTEM|			\
			     GFS2_DIF_INHERIT_DIRECTIO|		\
			     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
 *
 */
436
static int do_gfs2_set_flags(struct file *filp, u32 reqflags, u32 mask)
437
{
438
	struct inode *inode = filp->f_dentry->d_inode;
439 440
	struct gfs2_inode *ip = GFS2_I(inode);
	struct gfs2_sbd *sdp = GFS2_SB(inode);
441 442 443
	struct buffer_head *bh;
	struct gfs2_holder gh;
	int error;
444
	u32 new_flags, flags;
445 446

	error = gfs2_glock_nq_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &gh);
447
	if (error)
448 449
		return error;

450 451
	flags = ip->i_di.di_flags;
	new_flags = (flags & ~mask) | (reqflags & mask);
452 453 454
	if ((new_flags ^ flags) == 0)
		goto out;

455 456 457 458 459 460 461
	if (S_ISDIR(inode->i_mode)) {
		if ((new_flags ^ flags) & GFS2_DIF_JDATA)
			new_flags ^= (GFS2_DIF_JDATA|GFS2_DIF_INHERIT_JDATA);
		if ((new_flags ^ flags) & GFS2_DIF_DIRECTIO)
			new_flags ^= (GFS2_DIF_DIRECTIO|GFS2_DIF_INHERIT_DIRECTIO);
	}

462 463 464 465 466 467 468 469 470
	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;
471 472
	if (((new_flags ^ flags) & GFS2_DIF_IMMUTABLE) && 
	    !capable(CAP_LINUX_IMMUTABLE))
473
		goto out;
474
	if (!IS_IMMUTABLE(inode)) {
475
		error = permission(inode, MAY_WRITE, NULL);
476 477 478
		if (error)
			goto out;
	}
479

480
	error = gfs2_trans_begin(sdp, RES_DINODE, 0);
481 482
	if (error)
		goto out;
483 484 485
	error = gfs2_meta_inode_buffer(ip, &bh);
	if (error)
		goto out_trans_end;
486 487 488 489
	gfs2_trans_add_bh(ip->i_gl, bh, 1);
	ip->i_di.di_flags = new_flags;
	gfs2_dinode_out(&ip->i_di, bh->b_data);
	brelse(bh);
490 491
out_trans_end:
	gfs2_trans_end(sdp);
492 493 494 495 496
out:
	gfs2_glock_dq_uninit(&gh);
	return error;
}

497
static int gfs2_set_flags(struct file *filp, u32 __user *ptr)
498
{
499 500
	u32 iflags, gfsflags;
	if (get_user(iflags, ptr))
501
		return -EFAULT;
502
	gfsflags = iflags_cvt(iflags_to_gfs2, iflags);
503
	return do_gfs2_set_flags(filp, gfsflags, ~0);
504 505
}

506
static long gfs2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
507 508
{
	switch(cmd) {
509
	case IFLAGS_GET_IOC:
510
		return gfs2_get_flags(filp, (u32 __user *)arg);
511
	case IFLAGS_SET_IOC:
512
		return gfs2_set_flags(filp, (u32 __user *)arg);
513 514 515 516 517
	}
	return -ENOTTY;
}


D
David Teigland 已提交
518 519 520 521 522 523 524 525 526 527
/**
 * gfs2_mmap -
 * @file: The file to map
 * @vma: The VMA which described the mapping
 *
 * Returns: 0 or error code
 */

static int gfs2_mmap(struct file *file, struct vm_area_struct *vma)
{
528
	struct gfs2_inode *ip = GFS2_I(file->f_mapping->host);
D
David Teigland 已提交
529 530 531 532 533 534 535 536 537 538
	struct gfs2_holder i_gh;
	int error;

	gfs2_holder_init(ip->i_gl, LM_ST_SHARED, GL_ATIME, &i_gh);
	error = gfs2_glock_nq_atime(&i_gh);
	if (error) {
		gfs2_holder_uninit(&i_gh);
		return error;
	}

539 540 541 542 543 544 545 546
	/* This is VM_MAYWRITE instead of VM_WRITE because a call
	   to mprotect() can turn on VM_WRITE later. */

	if ((vma->vm_flags & (VM_MAYSHARE | VM_MAYWRITE)) ==
	    (VM_MAYSHARE | VM_MAYWRITE))
		vma->vm_ops = &gfs2_vm_ops_sharewrite;
	else
		vma->vm_ops = &gfs2_vm_ops_private;
D
David Teigland 已提交
547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562

	gfs2_glock_dq_uninit(&i_gh);

	return error;
}

/**
 * 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)
{
563
	struct gfs2_inode *ip = GFS2_I(inode);
D
David Teigland 已提交
564 565 566 567 568 569 570 571
	struct gfs2_holder i_gh;
	struct gfs2_file *fp;
	int error;

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

572
	mutex_init(&fp->f_fl_mutex);
D
David Teigland 已提交
573

574
	gfs2_assert_warn(GFS2_SB(inode), !file->private_data);
575
	file->private_data = fp;
D
David Teigland 已提交
576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598

	if (S_ISREG(ip->i_di.di_mode)) {
		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) &&
		    ip->i_di.di_size > MAX_NON_LFS) {
			error = -EFBIG;
			goto fail_gunlock;
		}

		/* Listen to the Direct I/O flag */

		if (ip->i_di.di_flags & GFS2_DIF_DIRECTIO)
			file->f_flags |= O_DIRECT;

		gfs2_glock_dq_uninit(&i_gh);
	}

	return 0;

599
fail_gunlock:
D
David Teigland 已提交
600
	gfs2_glock_dq_uninit(&i_gh);
601
fail:
602
	file->private_data = NULL;
D
David Teigland 已提交
603 604 605 606 607 608 609 610 611 612 613 614 615 616
	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)
{
617
	struct gfs2_sbd *sdp = inode->i_sb->s_fs_info;
D
David Teigland 已提交
618 619
	struct gfs2_file *fp;

620 621
	fp = file->private_data;
	file->private_data = NULL;
D
David Teigland 已提交
622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640

	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
 *
 * Returns: errno
 */

static int gfs2_fsync(struct file *file, struct dentry *dentry, int datasync)
{
641
	struct gfs2_inode *ip = GFS2_I(dentry->d_inode);
D
David Teigland 已提交
642

643
	gfs2_log_flush(ip->i_gl->gl_sbd, ip->i_gl);
D
David Teigland 已提交
644 645 646 647 648 649 650 651 652 653 654 655 656 657 658

	return 0;
}

/**
 * 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)
{
659 660
	struct gfs2_inode *ip = GFS2_I(file->f_mapping->host);
	struct gfs2_sbd *sdp = GFS2_SB(file->f_mapping->host);
D
David Teigland 已提交
661 662 663 664 665 666 667 668 669 670 671
	struct lm_lockname name =
		{ .ln_number = ip->i_num.no_addr,
		  .ln_type = LM_TYPE_PLOCK };

	if (!(fl->fl_flags & FL_POSIX))
		return -ENOLCK;
	if ((ip->i_di.di_mode & (S_ISGID | S_IXGRP)) == S_ISGID)
		return -ENOLCK;

	if (sdp->sd_args.ar_localflocks) {
		if (IS_GETLK(cmd)) {
672 673 674
			struct file_lock tmp;
			int ret;
			ret = posix_test_lock(file, fl, &tmp);
D
David Teigland 已提交
675
			fl->fl_type = F_UNLCK;
676 677
			if (ret)
				memcpy(fl, &tmp, sizeof(struct file_lock));
D
David Teigland 已提交
678 679
			return 0;
		} else {
680
			return posix_lock_file_wait(file, fl);
D
David Teigland 已提交
681 682 683 684 685 686 687 688 689 690 691 692 693
		}
	}

	if (IS_GETLK(cmd))
		return gfs2_lm_plock_get(sdp, &name, file, fl);
	else if (fl->fl_type == F_UNLCK)
		return gfs2_lm_punlock(sdp, &name, file, fl);
	else
		return gfs2_lm_plock(sdp, &name, file, cmd, fl);
}

static int do_flock(struct file *file, int cmd, struct file_lock *fl)
{
694
	struct gfs2_file *fp = file->private_data;
D
David Teigland 已提交
695
	struct gfs2_holder *fl_gh = &fp->f_fl_gh;
696
	struct gfs2_inode *ip = GFS2_I(file->f_dentry->d_inode);
D
David Teigland 已提交
697 698 699 700 701 702
	struct gfs2_glock *gl;
	unsigned int state;
	int flags;
	int error = 0;

	state = (fl->fl_type == F_WRLCK) ? LM_ST_EXCLUSIVE : LM_ST_SHARED;
703
	flags = (IS_SETLKW(cmd) ? 0 : LM_FLAG_TRY) | GL_EXACT | GL_NOCACHE;
D
David Teigland 已提交
704

705
	mutex_lock(&fp->f_fl_mutex);
D
David Teigland 已提交
706 707 708 709 710 711 712 713 714 715

	gl = fl_gh->gh_gl;
	if (gl) {
		if (fl_gh->gh_state == state)
			goto out;
		gfs2_glock_hold(gl);
		flock_lock_file_wait(file,
				     &(struct file_lock){.fl_type = F_UNLCK});		
		gfs2_glock_dq_uninit(fl_gh);
	} else {
716
		error = gfs2_glock_get(GFS2_SB(&ip->i_inode),
D
David Teigland 已提交
717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732
				      ip->i_num.no_addr, &gfs2_flock_glops,
				      CREATE, &gl);
		if (error)
			goto out;
	}

	gfs2_holder_init(gl, state, flags, fl_gh);
	gfs2_glock_put(gl);

	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);
733
		gfs2_assert_warn(GFS2_SB(&ip->i_inode), !error);
D
David Teigland 已提交
734 735
	}

736
out:
737
	mutex_unlock(&fp->f_fl_mutex);
D
David Teigland 已提交
738 739 740 741 742
	return error;
}

static void do_unflock(struct file *file, struct file_lock *fl)
{
743
	struct gfs2_file *fp = file->private_data;
D
David Teigland 已提交
744 745
	struct gfs2_holder *fl_gh = &fp->f_fl_gh;

746
	mutex_lock(&fp->f_fl_mutex);
D
David Teigland 已提交
747 748 749
	flock_lock_file_wait(file, fl);
	if (fl_gh->gh_gl)
		gfs2_glock_dq_uninit(fl_gh);
750
	mutex_unlock(&fp->f_fl_mutex);
D
David Teigland 已提交
751 752 753 754 755 756 757 758 759 760 761 762 763
}

/**
 * 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)
{
764 765
	struct gfs2_inode *ip = GFS2_I(file->f_mapping->host);
	struct gfs2_sbd *sdp = GFS2_SB(file->f_mapping->host);
D
David Teigland 已提交
766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781

	if (!(fl->fl_flags & FL_FLOCK))
		return -ENOLCK;
	if ((ip->i_di.di_mode & (S_ISGID | S_IXGRP)) == S_ISGID)
		return -ENOLCK;

	if (sdp->sd_args.ar_localflocks)
		return flock_lock_file_wait(file, fl);

	if (fl->fl_type == F_UNLCK) {
		do_unflock(file, fl);
		return 0;
	} else
		return do_flock(file, cmd, fl);
}

782
const struct file_operations gfs2_file_fops = {
783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799
	.llseek		= gfs2_llseek,
	.read		= generic_file_read,
	.readv		= generic_file_readv,
	.aio_read	= generic_file_aio_read,
	.write		= generic_file_write,
	.writev		= generic_file_writev,
	.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,
	.sendfile	= generic_file_sendfile,
	.flock		= gfs2_flock,
	.splice_read	= generic_file_splice_read,
	.splice_write	= generic_file_splice_write,
D
David Teigland 已提交
800 801
};

802
const struct file_operations gfs2_dir_fops = {
803 804 805 806 807 808 809
	.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
810 811
};