inode.c 9.5 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6
/*
 *   Copyright (C) International Business Machines Corp., 2000-2004
 *   Portions Copyright (C) Christoph Hellwig, 2001-2002
 *
 *   This program is free software;  you can redistribute it and/or modify
 *   it under the terms of the GNU General Public License as published by
D
Dave Kleikamp 已提交
7
 *   the Free Software Foundation; either version 2 of the License, or
L
Linus Torvalds 已提交
8
 *   (at your option) any later version.
D
Dave Kleikamp 已提交
9
 *
L
Linus Torvalds 已提交
10 11 12 13 14 15
 *   This program is distributed in the hope that it will be useful,
 *   but WITHOUT ANY WARRANTY;  without even the implied warranty of
 *   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See
 *   the GNU General Public License for more details.
 *
 *   You should have received a copy of the GNU General Public License
D
Dave Kleikamp 已提交
16
 *   along with this program;  if not, write to the Free Software
L
Linus Torvalds 已提交
17 18 19 20 21 22 23 24
 *   Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
 */

#include <linux/fs.h>
#include <linux/mpage.h>
#include <linux/buffer_head.h>
#include <linux/pagemap.h>
#include <linux/quotaops.h>
25
#include <linux/writeback.h>
L
Linus Torvalds 已提交
26
#include "jfs_incore.h"
27
#include "jfs_inode.h"
L
Linus Torvalds 已提交
28 29 30 31 32 33 34
#include "jfs_filsys.h"
#include "jfs_imap.h"
#include "jfs_extent.h"
#include "jfs_unicode.h"
#include "jfs_debug.h"


35
struct inode *jfs_iget(struct super_block *sb, unsigned long ino)
L
Linus Torvalds 已提交
36
{
37 38 39 40 41 42 43 44 45 46 47 48 49
	struct inode *inode;
	int ret;

	inode = iget_locked(sb, ino);
	if (!inode)
		return ERR_PTR(-ENOMEM);
	if (!(inode->i_state & I_NEW))
		return inode;

	ret = diRead(inode);
	if (ret < 0) {
		iget_failed(inode);
		return ERR_PTR(ret);
L
Linus Torvalds 已提交
50 51 52 53 54 55 56 57 58 59 60 61 62
	}

	if (S_ISREG(inode->i_mode)) {
		inode->i_op = &jfs_file_inode_operations;
		inode->i_fop = &jfs_file_operations;
		inode->i_mapping->a_ops = &jfs_aops;
	} else if (S_ISDIR(inode->i_mode)) {
		inode->i_op = &jfs_dir_inode_operations;
		inode->i_fop = &jfs_dir_operations;
	} else if (S_ISLNK(inode->i_mode)) {
		if (inode->i_size >= IDATASIZE) {
			inode->i_op = &page_symlink_inode_operations;
			inode->i_mapping->a_ops = &jfs_aops;
63
		} else {
64
			inode->i_op = &jfs_fast_symlink_inode_operations;
65 66 67 68 69 70
			/*
			 * The inline data should be null-terminated, but
			 * don't let on-disk corruption crash the kernel
			 */
			JFS_IP(inode)->i_inline[inode->i_size] = '\0';
		}
L
Linus Torvalds 已提交
71 72 73 74
	} else {
		inode->i_op = &jfs_file_inode_operations;
		init_special_inode(inode, inode->i_mode, inode->i_rdev);
	}
75 76
	unlock_new_inode(inode);
	return inode;
L
Linus Torvalds 已提交
77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110
}

/*
 * Workhorse of both fsync & write_inode
 */
int jfs_commit_inode(struct inode *inode, int wait)
{
	int rc = 0;
	tid_t tid;
	static int noisy = 5;

	jfs_info("In jfs_commit_inode, inode = 0x%p", inode);

	/*
	 * Don't commit if inode has been committed since last being
	 * marked dirty, or if it has been deleted.
	 */
	if (inode->i_nlink == 0 || !test_cflag(COMMIT_Dirty, inode))
		return 0;

	if (isReadOnly(inode)) {
		/* kernel allows writes to devices on read-only
		 * partitions and may think inode is dirty
		 */
		if (!special_file(inode->i_mode) && noisy) {
			jfs_err("jfs_commit_inode(0x%p) called on "
				   "read-only volume", inode);
			jfs_err("Is remount racy?");
			noisy--;
		}
		return 0;
	}

	tid = txBegin(inode->i_sb, COMMIT_INODE);
111
	mutex_lock(&JFS_IP(inode)->commit_mutex);
L
Linus Torvalds 已提交
112 113

	/*
114
	 * Retest inode state after taking commit_mutex
L
Linus Torvalds 已提交
115 116 117 118 119
	 */
	if (inode->i_nlink && test_cflag(COMMIT_Dirty, inode))
		rc = txCommit(tid, 1, &inode, wait ? COMMIT_SYNC : 0);

	txEnd(tid);
120
	mutex_unlock(&JFS_IP(inode)->commit_mutex);
L
Linus Torvalds 已提交
121 122 123
	return rc;
}

124
int jfs_write_inode(struct inode *inode, struct writeback_control *wbc)
L
Linus Torvalds 已提交
125
{
126 127
	int wait = wbc->sync_mode == WB_SYNC_ALL;

L
Linus Torvalds 已提交
128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151
	if (test_cflag(COMMIT_Nolink, inode))
		return 0;
	/*
	 * If COMMIT_DIRTY is not set, the inode isn't really dirty.
	 * It has been committed since the last change, but was still
	 * on the dirty inode list.
	 */
	 if (!test_cflag(COMMIT_Dirty, inode)) {
		/* Make sure committed changes hit the disk */
		jfs_flush_journal(JFS_SBI(inode->i_sb)->log, wait);
		return 0;
	 }

	if (jfs_commit_inode(inode, wait)) {
		jfs_err("jfs_write_inode: jfs_commit_inode failed!");
		return -EIO;
	} else
		return 0;
}

void jfs_delete_inode(struct inode *inode)
{
	jfs_info("In jfs_delete_inode, inode = 0x%p", inode);

152
	if (!is_bad_inode(inode))
153
		dquot_initialize(inode);
154

155
	if (!is_bad_inode(inode) &&
156
	    (JFS_IP(inode)->fileset == FILESYSTEM_I)) {
157
		truncate_inode_pages(&inode->i_data, 0);
158

159 160
		if (test_cflag(COMMIT_Freewmap, inode))
			jfs_free_zero_link(inode);
L
Linus Torvalds 已提交
161

162
		diFree(inode);
L
Linus Torvalds 已提交
163

164 165 166
		/*
		 * Free the inode from the quota allocation.
		 */
167
		dquot_initialize(inode);
168
		dquot_free_inode(inode);
169
		dquot_drop(inode);
170
	}
L
Linus Torvalds 已提交
171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193

	clear_inode(inode);
}

void jfs_dirty_inode(struct inode *inode)
{
	static int noisy = 5;

	if (isReadOnly(inode)) {
		if (!special_file(inode->i_mode) && noisy) {
			/* kernel allows writes to devices on read-only
			 * partitions and may try to mark inode dirty
			 */
			jfs_err("jfs_dirty_inode called on read-only volume");
			jfs_err("Is remount racy?");
			noisy--;
		}
		return;
	}

	set_cflag(COMMIT_Dirty, inode);
}

194 195
int jfs_get_block(struct inode *ip, sector_t lblock,
		  struct buffer_head *bh_result, int create)
L
Linus Torvalds 已提交
196 197 198 199 200 201
{
	s64 lblock64 = lblock;
	int rc = 0;
	xad_t xad;
	s64 xaddr;
	int xflag;
202
	s32 xlen = bh_result->b_size >> ip->i_blkbits;
L
Linus Torvalds 已提交
203 204 205 206

	/*
	 * Take appropriate lock on inode
	 */
207
	if (create)
D
Dave Kleikamp 已提交
208
		IWRITE_LOCK(ip, RDWRLOCK_NORMAL);
209
	else
D
Dave Kleikamp 已提交
210
		IREAD_LOCK(ip, RDWRLOCK_NORMAL);
L
Linus Torvalds 已提交
211 212

	if (((lblock64 << ip->i_sb->s_blocksize_bits) < ip->i_size) &&
213
	    (!xtLookup(ip, lblock64, xlen, &xflag, &xaddr, &xlen, 0)) &&
214
	    xaddr) {
L
Linus Torvalds 已提交
215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252
		if (xflag & XAD_NOTRECORDED) {
			if (!create)
				/*
				 * Allocated but not recorded, read treats
				 * this as a hole
				 */
				goto unlock;
#ifdef _JFS_4K
			XADoffset(&xad, lblock64);
			XADlength(&xad, xlen);
			XADaddress(&xad, xaddr);
#else				/* _JFS_4K */
			/*
			 * As long as block size = 4K, this isn't a problem.
			 * We should mark the whole page not ABNR, but how
			 * will we know to mark the other blocks BH_New?
			 */
			BUG();
#endif				/* _JFS_4K */
			rc = extRecord(ip, &xad);
			if (rc)
				goto unlock;
			set_buffer_new(bh_result);
		}

		map_bh(bh_result, ip->i_sb, xaddr);
		bh_result->b_size = xlen << ip->i_blkbits;
		goto unlock;
	}
	if (!create)
		goto unlock;

	/*
	 * Allocate a new block
	 */
#ifdef _JFS_4K
	if ((rc = extHint(ip, lblock64 << ip->i_sb->s_blocksize_bits, &xad)))
		goto unlock;
253
	rc = extAlloc(ip, xlen, lblock64, &xad, false);
L
Linus Torvalds 已提交
254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272
	if (rc)
		goto unlock;

	set_buffer_new(bh_result);
	map_bh(bh_result, ip->i_sb, addressXAD(&xad));
	bh_result->b_size = lengthXAD(&xad) << ip->i_blkbits;

#else				/* _JFS_4K */
	/*
	 * We need to do whatever it takes to keep all but the last buffers
	 * in 4K pages - see jfs_write.c
	 */
	BUG();
#endif				/* _JFS_4K */

      unlock:
	/*
	 * Release lock on inode
	 */
273 274 275 276
	if (create)
		IWRITE_UNLOCK(ip);
	else
		IREAD_UNLOCK(ip);
L
Linus Torvalds 已提交
277 278 279 280 281
	return rc;
}

static int jfs_writepage(struct page *page, struct writeback_control *wbc)
{
N
Nick Piggin 已提交
282
	return block_write_full_page(page, jfs_get_block, wbc);
L
Linus Torvalds 已提交
283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301
}

static int jfs_writepages(struct address_space *mapping,
			struct writeback_control *wbc)
{
	return mpage_writepages(mapping, wbc, jfs_get_block);
}

static int jfs_readpage(struct file *file, struct page *page)
{
	return mpage_readpage(page, jfs_get_block);
}

static int jfs_readpages(struct file *file, struct address_space *mapping,
		struct list_head *pages, unsigned nr_pages)
{
	return mpage_readpages(mapping, pages, nr_pages, jfs_get_block);
}

N
Nick Piggin 已提交
302 303 304
static int jfs_write_begin(struct file *file, struct address_space *mapping,
				loff_t pos, unsigned len, unsigned flags,
				struct page **pagep, void **fsdata)
L
Linus Torvalds 已提交
305
{
N
Nick Piggin 已提交
306
	return nobh_write_begin(file, mapping, pos, len, flags, pagep, fsdata,
N
Nick Piggin 已提交
307
				jfs_get_block);
L
Linus Torvalds 已提交
308 309 310 311 312 313 314 315 316 317 318 319 320 321
}

static sector_t jfs_bmap(struct address_space *mapping, sector_t block)
{
	return generic_block_bmap(mapping, block, jfs_get_block);
}

static ssize_t jfs_direct_IO(int rw, struct kiocb *iocb,
	const struct iovec *iov, loff_t offset, unsigned long nr_segs)
{
	struct file *file = iocb->ki_filp;
	struct inode *inode = file->f_mapping->host;

	return blockdev_direct_IO(rw, iocb, inode, inode->i_sb->s_bdev, iov,
322
				offset, nr_segs, jfs_get_block, NULL);
L
Linus Torvalds 已提交
323 324
}

325
const struct address_space_operations jfs_aops = {
L
Linus Torvalds 已提交
326 327 328 329 330
	.readpage	= jfs_readpage,
	.readpages	= jfs_readpages,
	.writepage	= jfs_writepage,
	.writepages	= jfs_writepages,
	.sync_page	= block_sync_page,
N
Nick Piggin 已提交
331
	.write_begin	= jfs_write_begin,
N
Nick Piggin 已提交
332
	.write_end	= nobh_write_end,
L
Linus Torvalds 已提交
333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356
	.bmap		= jfs_bmap,
	.direct_IO	= jfs_direct_IO,
};

/*
 * Guts of jfs_truncate.  Called with locks already held.  Can be called
 * with directory for truncating directory index table.
 */
void jfs_truncate_nolock(struct inode *ip, loff_t length)
{
	loff_t newsize;
	tid_t tid;

	ASSERT(length >= 0);

	if (test_cflag(COMMIT_Nolink, ip)) {
		xtTruncate(0, ip, length, COMMIT_WMAP);
		return;
	}

	do {
		tid = txBegin(ip->i_sb, 0);

		/*
357
		 * The commit_mutex cannot be taken before txBegin.
L
Linus Torvalds 已提交
358 359 360 361
		 * txBegin may block and there is a chance the inode
		 * could be marked dirty and need to be committed
		 * before txBegin unblocks
		 */
362
		mutex_lock(&JFS_IP(ip)->commit_mutex);
L
Linus Torvalds 已提交
363 364 365 366 367

		newsize = xtTruncate(tid, ip, length,
				     COMMIT_TRUNCATE | COMMIT_PWMAP);
		if (newsize < 0) {
			txEnd(tid);
368
			mutex_unlock(&JFS_IP(ip)->commit_mutex);
L
Linus Torvalds 已提交
369 370 371 372 373 374 375 376
			break;
		}

		ip->i_mtime = ip->i_ctime = CURRENT_TIME;
		mark_inode_dirty(ip);

		txCommit(tid, 1, &ip, 0);
		txEnd(tid);
377
		mutex_unlock(&JFS_IP(ip)->commit_mutex);
L
Linus Torvalds 已提交
378 379 380 381 382 383 384
	} while (newsize > length);	/* Truncate isn't always atomic */
}

void jfs_truncate(struct inode *ip)
{
	jfs_info("jfs_truncate: size = 0x%lx", (ulong) ip->i_size);

N
Nick Piggin 已提交
385
	nobh_truncate_page(ip->i_mapping, ip->i_size, jfs_get_block);
L
Linus Torvalds 已提交
386

D
Dave Kleikamp 已提交
387
	IWRITE_LOCK(ip, RDWRLOCK_NORMAL);
L
Linus Torvalds 已提交
388 389 390
	jfs_truncate_nolock(ip, ip->i_size);
	IWRITE_UNLOCK(ip);
}