lops.c 19.0 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
 */

#include <linux/sched.h>
#include <linux/slab.h>
#include <linux/spinlock.h>
#include <linux/completion.h>
#include <linux/buffer_head.h>
15
#include <linux/gfs2_ondisk.h>
S
Steven Whitehouse 已提交
16 17
#include <linux/bio.h>
#include <linux/fs.h>
D
David Teigland 已提交
18 19

#include "gfs2.h"
20
#include "incore.h"
21
#include "inode.h"
D
David Teigland 已提交
22 23 24 25 26 27 28
#include "glock.h"
#include "log.h"
#include "lops.h"
#include "meta_io.h"
#include "recovery.h"
#include "rgrp.h"
#include "trans.h"
29
#include "util.h"
S
Steven Whitehouse 已提交
30
#include "trace_gfs2.h"
D
David Teigland 已提交
31

32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56
/**
 * gfs2_pin - Pin a buffer in memory
 * @sdp: The superblock
 * @bh: The buffer to be pinned
 *
 * The log lock must be held when calling this function
 */
static void gfs2_pin(struct gfs2_sbd *sdp, struct buffer_head *bh)
{
	struct gfs2_bufdata *bd;

	gfs2_assert_withdraw(sdp, test_bit(SDF_JOURNAL_LIVE, &sdp->sd_flags));

	clear_buffer_dirty(bh);
	if (test_set_buffer_pinned(bh))
		gfs2_assert_withdraw(sdp, 0);
	if (!buffer_uptodate(bh))
		gfs2_io_error_bh(sdp, bh);
	bd = bh->b_private;
	/* If this buffer is in the AIL and it has already been written
	 * to in-place disk block, remove it from the AIL.
	 */
	if (bd->bd_ail)
		list_move(&bd->bd_ail_st_list, &bd->bd_ail->ai_ail2_list);
	get_bh(bh);
57
	atomic_inc(&sdp->sd_log_pinned);
S
Steven Whitehouse 已提交
58
	trace_gfs2_pin(bd, 1);
59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82
}

/**
 * gfs2_unpin - Unpin a buffer
 * @sdp: the filesystem the buffer belongs to
 * @bh: The buffer to unpin
 * @ai:
 *
 */

static void gfs2_unpin(struct gfs2_sbd *sdp, struct buffer_head *bh,
		       struct gfs2_ail *ai)
{
	struct gfs2_bufdata *bd = bh->b_private;

	gfs2_assert_withdraw(sdp, buffer_uptodate(bh));

	if (!buffer_pinned(bh))
		gfs2_assert_withdraw(sdp, 0);

	lock_buffer(bh);
	mark_buffer_dirty(bh);
	clear_buffer_pinned(bh);

D
Dave Chinner 已提交
83
	spin_lock(&sdp->sd_ail_lock);
84 85 86 87 88 89 90 91 92 93
	if (bd->bd_ail) {
		list_del(&bd->bd_ail_st_list);
		brelse(bh);
	} else {
		struct gfs2_glock *gl = bd->bd_gl;
		list_add(&bd->bd_ail_gl_list, &gl->gl_ail_list);
		atomic_inc(&gl->gl_ail_count);
	}
	bd->bd_ail = ai;
	list_add(&bd->bd_ail_st_list, &ai->ai_ail1_list);
D
Dave Chinner 已提交
94 95
	spin_unlock(&sdp->sd_ail_lock);

96 97
	if (test_and_clear_bit(GLF_LFLUSH, &bd->bd_gl->gl_flags))
		gfs2_glock_schedule_for_reclaim(bd->bd_gl);
S
Steven Whitehouse 已提交
98
	trace_gfs2_pin(bd, 0);
99
	unlock_buffer(bh);
100
	atomic_dec(&sdp->sd_log_pinned);
101 102
}

103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135

static inline struct gfs2_log_descriptor *bh_log_desc(struct buffer_head *bh)
{
	return (struct gfs2_log_descriptor *)bh->b_data;
}

static inline __be64 *bh_log_ptr(struct buffer_head *bh)
{
	struct gfs2_log_descriptor *ld = bh_log_desc(bh);
	return (__force __be64 *)(ld + 1);
}

static inline __be64 *bh_ptr_end(struct buffer_head *bh)
{
	return (__force __be64 *)(bh->b_data + bh->b_size);
}


static struct buffer_head *gfs2_get_log_desc(struct gfs2_sbd *sdp, u32 ld_type)
{
	struct buffer_head *bh = gfs2_log_get_buf(sdp);
	struct gfs2_log_descriptor *ld = bh_log_desc(bh);
	ld->ld_header.mh_magic = cpu_to_be32(GFS2_MAGIC);
	ld->ld_header.mh_type = cpu_to_be32(GFS2_METATYPE_LD);
	ld->ld_header.mh_format = cpu_to_be32(GFS2_FORMAT_LD);
	ld->ld_type = cpu_to_be32(ld_type);
	ld->ld_length = 0;
	ld->ld_data1 = 0;
	ld->ld_data2 = 0;
	memset(ld->ld_reserved, 0, sizeof(ld->ld_reserved));
	return bh;
}

D
David Teigland 已提交
136 137 138
static void buf_lo_add(struct gfs2_sbd *sdp, struct gfs2_log_element *le)
{
	struct gfs2_bufdata *bd = container_of(le, struct gfs2_bufdata, bd_le);
139
	struct gfs2_meta_header *mh;
D
David Teigland 已提交
140 141
	struct gfs2_trans *tr;

142
	lock_buffer(bd->bd_bh);
143
	gfs2_log_lock(sdp);
144 145
	if (!list_empty(&bd->bd_list_tr))
		goto out;
146
	tr = current->journal_info;
D
David Teigland 已提交
147 148 149 150
	tr->tr_touched = 1;
	tr->tr_num_buf++;
	list_add(&bd->bd_list_tr, &tr->tr_list_buf);
	if (!list_empty(&le->le_list))
151
		goto out;
152 153
	set_bit(GLF_LFLUSH, &bd->bd_gl->gl_flags);
	set_bit(GLF_DIRTY, &bd->bd_gl->gl_flags);
D
David Teigland 已提交
154
	gfs2_meta_check(sdp, bd->bd_bh);
155
	gfs2_pin(sdp, bd->bd_bh);
156 157 158
	mh = (struct gfs2_meta_header *)bd->bd_bh->b_data;
	mh->__pad0 = cpu_to_be64(0);
	mh->mh_jid = cpu_to_be32(sdp->sd_jdesc->jd_jid);
D
David Teigland 已提交
159 160 161
	sdp->sd_log_num_buf++;
	list_add(&le->le_list, &sdp->sd_log_le_buf);
	tr->tr_num_buf_new++;
162 163 164
out:
	gfs2_log_unlock(sdp);
	unlock_buffer(bd->bd_bh);
D
David Teigland 已提交
165 166 167 168 169 170 171
}

static void buf_lo_before_commit(struct gfs2_sbd *sdp)
{
	struct buffer_head *bh;
	struct gfs2_log_descriptor *ld;
	struct gfs2_bufdata *bd1 = NULL, *bd2;
172
	unsigned int total;
D
David Teigland 已提交
173 174 175 176 177
	unsigned int limit;
	unsigned int num;
	unsigned n;
	__be64 *ptr;

178
	limit = buf_limit(sdp);
D
David Teigland 已提交
179 180
	/* for 4k blocks, limit = 503 */

181 182
	gfs2_log_lock(sdp);
	total = sdp->sd_log_num_buf;
D
David Teigland 已提交
183 184 185 186 187
	bd1 = bd2 = list_prepare_entry(bd1, &sdp->sd_log_le_buf, bd_le.le_list);
	while(total) {
		num = total;
		if (total > limit)
			num = limit;
188
		gfs2_log_unlock(sdp);
189
		bh = gfs2_get_log_desc(sdp, GFS2_LOG_DESC_METADATA);
190
		gfs2_log_lock(sdp);
191 192
		ld = bh_log_desc(bh);
		ptr = bh_log_ptr(bh);
D
David Teigland 已提交
193 194 195 196
		ld->ld_length = cpu_to_be32(num + 1);
		ld->ld_data1 = cpu_to_be32(num);

		n = 0;
197 198
		list_for_each_entry_continue(bd1, &sdp->sd_log_le_buf,
					     bd_le.le_list) {
D
David Teigland 已提交
199 200 201 202 203
			*ptr++ = cpu_to_be64(bd1->bd_bh->b_blocknr);
			if (++n >= num)
				break;
		}

204
		gfs2_log_unlock(sdp);
S
Steven Whitehouse 已提交
205
		submit_bh(WRITE_SYNC_PLUG, bh);
206
		gfs2_log_lock(sdp);
D
David Teigland 已提交
207 208

		n = 0;
209 210
		list_for_each_entry_continue(bd2, &sdp->sd_log_le_buf,
					     bd_le.le_list) {
211
			get_bh(bd2->bd_bh);
212
			gfs2_log_unlock(sdp);
213
			lock_buffer(bd2->bd_bh);
D
David Teigland 已提交
214
			bh = gfs2_log_fake_buf(sdp, bd2->bd_bh);
S
Steven Whitehouse 已提交
215
			submit_bh(WRITE_SYNC_PLUG, bh);
216
			gfs2_log_lock(sdp);
D
David Teigland 已提交
217 218 219 220
			if (++n >= num)
				break;
		}

221
		BUG_ON(total < num);
D
David Teigland 已提交
222 223
		total -= num;
	}
224
	gfs2_log_unlock(sdp);
D
David Teigland 已提交
225 226 227 228 229 230 231 232 233 234 235 236
}

static void buf_lo_after_commit(struct gfs2_sbd *sdp, struct gfs2_ail *ai)
{
	struct list_head *head = &sdp->sd_log_le_buf;
	struct gfs2_bufdata *bd;

	while (!list_empty(head)) {
		bd = list_entry(head->next, struct gfs2_bufdata, bd_le.le_list);
		list_del_init(&bd->bd_le.le_list);
		sdp->sd_log_num_buf--;

237
		gfs2_unpin(sdp, bd->bd_bh, ai);
D
David Teigland 已提交
238 239 240 241 242
	}
	gfs2_assert_warn(sdp, !sdp->sd_log_num_buf);
}

static void buf_lo_before_scan(struct gfs2_jdesc *jd,
A
Al Viro 已提交
243
			       struct gfs2_log_header_host *head, int pass)
D
David Teigland 已提交
244
{
245
	struct gfs2_sbd *sdp = GFS2_SB(jd->jd_inode);
D
David Teigland 已提交
246 247 248 249 250 251 252 253 254 255 256 257

	if (pass != 0)
		return;

	sdp->sd_found_blocks = 0;
	sdp->sd_replayed_blocks = 0;
}

static int buf_lo_scan_elements(struct gfs2_jdesc *jd, unsigned int start,
				struct gfs2_log_descriptor *ld, __be64 *ptr,
				int pass)
{
258 259
	struct gfs2_inode *ip = GFS2_I(jd->jd_inode);
	struct gfs2_sbd *sdp = GFS2_SB(jd->jd_inode);
260
	struct gfs2_glock *gl = ip->i_gl;
D
David Teigland 已提交
261 262
	unsigned int blks = be32_to_cpu(ld->ld_data1);
	struct buffer_head *bh_log, *bh_ip;
263
	u64 blkno;
D
David Teigland 已提交
264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279
	int error = 0;

	if (pass != 1 || be32_to_cpu(ld->ld_type) != GFS2_LOG_DESC_METADATA)
		return 0;

	gfs2_replay_incr_blk(sdp, &start);

	for (; blks; gfs2_replay_incr_blk(sdp, &start), blks--) {
		blkno = be64_to_cpu(*ptr++);

		sdp->sd_found_blocks++;

		if (gfs2_revoke_check(sdp, blkno, start))
			continue;

		error = gfs2_replay_read_block(jd, start, &bh_log);
S
Steven Whitehouse 已提交
280 281
		if (error)
			return error;
D
David Teigland 已提交
282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304

		bh_ip = gfs2_meta_new(gl, blkno);
		memcpy(bh_ip->b_data, bh_log->b_data, bh_log->b_size);

		if (gfs2_meta_check(sdp, bh_ip))
			error = -EIO;
		else
			mark_buffer_dirty(bh_ip);

		brelse(bh_log);
		brelse(bh_ip);

		if (error)
			break;

		sdp->sd_replayed_blocks++;
	}

	return error;
}

static void buf_lo_after_scan(struct gfs2_jdesc *jd, int error, int pass)
{
305 306
	struct gfs2_inode *ip = GFS2_I(jd->jd_inode);
	struct gfs2_sbd *sdp = GFS2_SB(jd->jd_inode);
D
David Teigland 已提交
307 308

	if (error) {
S
Steven Whitehouse 已提交
309
		gfs2_meta_sync(ip->i_gl);
D
David Teigland 已提交
310 311 312 313 314
		return;
	}
	if (pass != 1)
		return;

S
Steven Whitehouse 已提交
315
	gfs2_meta_sync(ip->i_gl);
D
David Teigland 已提交
316 317 318 319 320 321 322 323 324

	fs_info(sdp, "jid=%u: Replayed %u of %u blocks\n",
	        jd->jd_jid, sdp->sd_replayed_blocks, sdp->sd_found_blocks);
}

static void revoke_lo_add(struct gfs2_sbd *sdp, struct gfs2_log_element *le)
{
	struct gfs2_trans *tr;

325
	tr = current->journal_info;
D
David Teigland 已提交
326 327 328 329 330 331 332 333 334 335 336 337 338
	tr->tr_touched = 1;
	tr->tr_num_revoke++;
	sdp->sd_log_num_revoke++;
	list_add(&le->le_list, &sdp->sd_log_le_revoke);
}

static void revoke_lo_before_commit(struct gfs2_sbd *sdp)
{
	struct gfs2_log_descriptor *ld;
	struct gfs2_meta_header *mh;
	struct buffer_head *bh;
	unsigned int offset;
	struct list_head *head = &sdp->sd_log_le_revoke;
339
	struct gfs2_bufdata *bd;
D
David Teigland 已提交
340 341 342 343

	if (!sdp->sd_log_num_revoke)
		return;

344 345
	bh = gfs2_get_log_desc(sdp, GFS2_LOG_DESC_REVOKE);
	ld = bh_log_desc(bh);
346
	ld->ld_length = cpu_to_be32(gfs2_struct2blk(sdp, sdp->sd_log_num_revoke,
347
						    sizeof(u64)));
D
David Teigland 已提交
348 349 350 351
	ld->ld_data1 = cpu_to_be32(sdp->sd_log_num_revoke);
	offset = sizeof(struct gfs2_log_descriptor);

	while (!list_empty(head)) {
352 353
		bd = list_entry(head->next, struct gfs2_bufdata, bd_le.le_list);
		list_del_init(&bd->bd_le.le_list);
D
David Teigland 已提交
354 355
		sdp->sd_log_num_revoke--;

356
		if (offset + sizeof(u64) > sdp->sd_sb.sb_bsize) {
S
Steven Whitehouse 已提交
357
			submit_bh(WRITE_SYNC_PLUG, bh);
D
David Teigland 已提交
358 359 360 361

			bh = gfs2_log_get_buf(sdp);
			mh = (struct gfs2_meta_header *)bh->b_data;
			mh->mh_magic = cpu_to_be32(GFS2_MAGIC);
362 363
			mh->mh_type = cpu_to_be32(GFS2_METATYPE_LB);
			mh->mh_format = cpu_to_be32(GFS2_FORMAT_LB);
D
David Teigland 已提交
364 365 366
			offset = sizeof(struct gfs2_meta_header);
		}

367
		*(__be64 *)(bh->b_data + offset) = cpu_to_be64(bd->bd_blkno);
368
		kmem_cache_free(gfs2_bufdata_cachep, bd);
D
David Teigland 已提交
369

370
		offset += sizeof(u64);
D
David Teigland 已提交
371 372 373
	}
	gfs2_assert_withdraw(sdp, !sdp->sd_log_num_revoke);

S
Steven Whitehouse 已提交
374
	submit_bh(WRITE_SYNC_PLUG, bh);
D
David Teigland 已提交
375 376 377
}

static void revoke_lo_before_scan(struct gfs2_jdesc *jd,
A
Al Viro 已提交
378
				  struct gfs2_log_header_host *head, int pass)
D
David Teigland 已提交
379
{
380
	struct gfs2_sbd *sdp = GFS2_SB(jd->jd_inode);
D
David Teigland 已提交
381 382 383 384 385 386 387 388 389 390 391 392

	if (pass != 0)
		return;

	sdp->sd_found_revokes = 0;
	sdp->sd_replay_tail = head->lh_tail;
}

static int revoke_lo_scan_elements(struct gfs2_jdesc *jd, unsigned int start,
				   struct gfs2_log_descriptor *ld, __be64 *ptr,
				   int pass)
{
393
	struct gfs2_sbd *sdp = GFS2_SB(jd->jd_inode);
D
David Teigland 已提交
394 395 396 397
	unsigned int blks = be32_to_cpu(ld->ld_length);
	unsigned int revokes = be32_to_cpu(ld->ld_data1);
	struct buffer_head *bh;
	unsigned int offset;
398
	u64 blkno;
D
David Teigland 已提交
399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414
	int first = 1;
	int error;

	if (pass != 0 || be32_to_cpu(ld->ld_type) != GFS2_LOG_DESC_REVOKE)
		return 0;

	offset = sizeof(struct gfs2_log_descriptor);

	for (; blks; gfs2_replay_incr_blk(sdp, &start), blks--) {
		error = gfs2_replay_read_block(jd, start, &bh);
		if (error)
			return error;

		if (!first)
			gfs2_metatype_check(sdp, bh, GFS2_METATYPE_LB);

415
		while (offset + sizeof(u64) <= sdp->sd_sb.sb_bsize) {
D
David Teigland 已提交
416 417 418
			blkno = be64_to_cpu(*(__be64 *)(bh->b_data + offset));

			error = gfs2_revoke_add(sdp, blkno, start);
B
Bob Peterson 已提交
419 420
			if (error < 0) {
				brelse(bh);
D
David Teigland 已提交
421
				return error;
B
Bob Peterson 已提交
422
			}
D
David Teigland 已提交
423 424 425 426 427
			else if (error)
				sdp->sd_found_revokes++;

			if (!--revokes)
				break;
428
			offset += sizeof(u64);
D
David Teigland 已提交
429 430 431 432 433 434 435 436 437 438 439 440
		}

		brelse(bh);
		offset = sizeof(struct gfs2_meta_header);
		first = 0;
	}

	return 0;
}

static void revoke_lo_after_scan(struct gfs2_jdesc *jd, int error, int pass)
{
441
	struct gfs2_sbd *sdp = GFS2_SB(jd->jd_inode);
D
David Teigland 已提交
442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458

	if (error) {
		gfs2_revoke_clean(sdp);
		return;
	}
	if (pass != 1)
		return;

	fs_info(sdp, "jid=%u: Found %u revoke tags\n",
	        jd->jd_jid, sdp->sd_found_revokes);

	gfs2_revoke_clean(sdp);
}

static void rg_lo_add(struct gfs2_sbd *sdp, struct gfs2_log_element *le)
{
	struct gfs2_rgrpd *rgd;
459
	struct gfs2_trans *tr = current->journal_info;
D
David Teigland 已提交
460

461
	tr->tr_touched = 1;
D
David Teigland 已提交
462 463 464 465

	rgd = container_of(le, struct gfs2_rgrpd, rd_le);

	gfs2_log_lock(sdp);
466 467 468 469 470
	if (!list_empty(&le->le_list)){
		gfs2_log_unlock(sdp);
		return;
	}
	gfs2_rgrp_bh_hold(rgd);
D
David Teigland 已提交
471 472
	sdp->sd_log_num_rg++;
	list_add(&le->le_list, &sdp->sd_log_le_rg);
473
	gfs2_log_unlock(sdp);
D
David Teigland 已提交
474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491
}

static void rg_lo_after_commit(struct gfs2_sbd *sdp, struct gfs2_ail *ai)
{
	struct list_head *head = &sdp->sd_log_le_rg;
	struct gfs2_rgrpd *rgd;

	while (!list_empty(head)) {
		rgd = list_entry(head->next, struct gfs2_rgrpd, rd_le.le_list);
		list_del_init(&rgd->rd_le.le_list);
		sdp->sd_log_num_rg--;

		gfs2_rgrp_repolish_clones(rgd);
		gfs2_rgrp_bh_put(rgd);
	}
	gfs2_assert_warn(sdp, !sdp->sd_log_num_rg);
}

492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507
/**
 * databuf_lo_add - Add a databuf to the transaction.
 *
 * This is used in two distinct cases:
 * i) In ordered write mode
 *    We put the data buffer on a list so that we can ensure that its
 *    synced to disk at the right time
 * ii) In journaled data mode
 *    We need to journal the data block in the same way as metadata in
 *    the functions above. The difference is that here we have a tag
 *    which is two __be64's being the block number (as per meta data)
 *    and a flag which says whether the data block needs escaping or
 *    not. This means we need a new log entry for each 251 or so data
 *    blocks, which isn't an enormous overhead but twice as much as
 *    for normal metadata blocks.
 */
D
David Teigland 已提交
508 509
static void databuf_lo_add(struct gfs2_sbd *sdp, struct gfs2_log_element *le)
{
510
	struct gfs2_bufdata *bd = container_of(le, struct gfs2_bufdata, bd_le);
511
	struct gfs2_trans *tr = current->journal_info;
512
	struct address_space *mapping = bd->bd_bh->b_page->mapping;
513
	struct gfs2_inode *ip = GFS2_I(mapping->host);
D
David Teigland 已提交
514

515
	lock_buffer(bd->bd_bh);
516
	gfs2_log_lock(sdp);
517 518 519 520 521 522 523 524
	if (tr) {
		if (!list_empty(&bd->bd_list_tr))
			goto out;
		tr->tr_touched = 1;
		if (gfs2_is_jdata(ip)) {
			tr->tr_num_buf++;
			list_add(&bd->bd_list_tr, &tr->tr_list_buf);
		}
525
	}
526
	if (!list_empty(&le->le_list))
527
		goto out;
528

529 530
	set_bit(GLF_LFLUSH, &bd->bd_gl->gl_flags);
	set_bit(GLF_DIRTY, &bd->bd_gl->gl_flags);
531 532 533
	if (gfs2_is_jdata(ip)) {
		gfs2_pin(sdp, bd->bd_bh);
		tr->tr_num_databuf_new++;
534
		sdp->sd_log_num_databuf++;
535
		list_add_tail(&le->le_list, &sdp->sd_log_le_databuf);
536
	} else {
537
		list_add_tail(&le->le_list, &sdp->sd_log_le_ordered);
538 539
	}
out:
D
David Teigland 已提交
540
	gfs2_log_unlock(sdp);
541
	unlock_buffer(bd->bd_bh);
D
David Teigland 已提交
542 543
}

544
static void gfs2_check_magic(struct buffer_head *bh)
545 546 547 548
{
	void *kaddr;
	__be32 *ptr;

549 550
	clear_buffer_escaped(bh);
	kaddr = kmap_atomic(bh->b_page, KM_USER0);
551 552
	ptr = kaddr + bh_offset(bh);
	if (*ptr == cpu_to_be32(GFS2_MAGIC))
553
		set_buffer_escaped(bh);
554
	kunmap_atomic(kaddr, KM_USER0);
555 556
}

557 558 559
static void gfs2_write_blocks(struct gfs2_sbd *sdp, struct buffer_head *bh,
			      struct list_head *list, struct list_head *done,
			      unsigned int n)
D
David Teigland 已提交
560
{
561
	struct buffer_head *bh1;
562
	struct gfs2_log_descriptor *ld;
563 564
	struct gfs2_bufdata *bd;
	__be64 *ptr;
565

566 567
	if (!bh)
		return;
D
David Teigland 已提交
568

569 570 571
	ld = bh_log_desc(bh);
	ld->ld_length = cpu_to_be32(n + 1);
	ld->ld_data1 = cpu_to_be32(n);
D
David Teigland 已提交
572

573 574 575
	ptr = bh_log_ptr(bh);
	
	get_bh(bh);
S
Steven Whitehouse 已提交
576
	submit_bh(WRITE_SYNC_PLUG, bh);
577
	gfs2_log_lock(sdp);
578 579 580 581 582 583 584
	while(!list_empty(list)) {
		bd = list_entry(list->next, struct gfs2_bufdata, bd_le.le_list);
		list_move_tail(&bd->bd_le.le_list, done);
		get_bh(bd->bd_bh);
		while (be64_to_cpu(*ptr) != bd->bd_bh->b_blocknr) {
			gfs2_log_incr_head(sdp);
			ptr += 2;
585
		}
586
		gfs2_log_unlock(sdp);
587 588 589 590 591 592 593 594 595 596 597 598 599 600
		lock_buffer(bd->bd_bh);
		if (buffer_escaped(bd->bd_bh)) {
			void *kaddr;
			bh1 = gfs2_log_get_buf(sdp);
			kaddr = kmap_atomic(bd->bd_bh->b_page, KM_USER0);
			memcpy(bh1->b_data, kaddr + bh_offset(bd->bd_bh),
			       bh1->b_size);
			kunmap_atomic(kaddr, KM_USER0);
			*(__be32 *)bh1->b_data = 0;
			clear_buffer_escaped(bd->bd_bh);
			unlock_buffer(bd->bd_bh);
			brelse(bd->bd_bh);
		} else {
			bh1 = gfs2_log_fake_buf(sdp, bd->bd_bh);
601
		}
S
Steven Whitehouse 已提交
602
		submit_bh(WRITE_SYNC_PLUG, bh1);
603
		gfs2_log_lock(sdp);
604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626
		ptr += 2;
	}
	gfs2_log_unlock(sdp);
	brelse(bh);
}

/**
 * databuf_lo_before_commit - Scan the data buffers, writing as we go
 *
 */

static void databuf_lo_before_commit(struct gfs2_sbd *sdp)
{
	struct gfs2_bufdata *bd = NULL;
	struct buffer_head *bh = NULL;
	unsigned int n = 0;
	__be64 *ptr = NULL, *end = NULL;
	LIST_HEAD(processed);
	LIST_HEAD(in_progress);

	gfs2_log_lock(sdp);
	while (!list_empty(&sdp->sd_log_le_databuf)) {
		if (ptr == end) {
627
			gfs2_log_unlock(sdp);
628 629 630 631 632
			gfs2_write_blocks(sdp, bh, &in_progress, &processed, n);
			n = 0;
			bh = gfs2_get_log_desc(sdp, GFS2_LOG_DESC_JDATA);
			ptr = bh_log_ptr(bh);
			end = bh_ptr_end(bh) - 1;
633
			gfs2_log_lock(sdp);
634
			continue;
635
		}
636 637 638 639 640 641
		bd = list_entry(sdp->sd_log_le_databuf.next, struct gfs2_bufdata, bd_le.le_list);
		list_move_tail(&bd->bd_le.le_list, &in_progress);
		gfs2_check_magic(bd->bd_bh);
		*ptr++ = cpu_to_be64(bd->bd_bh->b_blocknr);
		*ptr++ = cpu_to_be64(buffer_escaped(bh) ? 1 : 0);
		n++;
D
David Teigland 已提交
642
	}
643
	gfs2_log_unlock(sdp);
644 645 646 647
	gfs2_write_blocks(sdp, bh, &in_progress, &processed, n);
	gfs2_log_lock(sdp);
	list_splice(&processed, &sdp->sd_log_le_databuf);
	gfs2_log_unlock(sdp);
648 649 650 651 652 653
}

static int databuf_lo_scan_elements(struct gfs2_jdesc *jd, unsigned int start,
				    struct gfs2_log_descriptor *ld,
				    __be64 *ptr, int pass)
{
654 655
	struct gfs2_inode *ip = GFS2_I(jd->jd_inode);
	struct gfs2_sbd *sdp = GFS2_SB(jd->jd_inode);
656
	struct gfs2_glock *gl = ip->i_gl;
657 658
	unsigned int blks = be32_to_cpu(ld->ld_data1);
	struct buffer_head *bh_log, *bh_ip;
659 660
	u64 blkno;
	u64 esc;
661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704
	int error = 0;

	if (pass != 1 || be32_to_cpu(ld->ld_type) != GFS2_LOG_DESC_JDATA)
		return 0;

	gfs2_replay_incr_blk(sdp, &start);
	for (; blks; gfs2_replay_incr_blk(sdp, &start), blks--) {
		blkno = be64_to_cpu(*ptr++);
		esc = be64_to_cpu(*ptr++);

		sdp->sd_found_blocks++;

		if (gfs2_revoke_check(sdp, blkno, start))
			continue;

		error = gfs2_replay_read_block(jd, start, &bh_log);
		if (error)
			return error;

		bh_ip = gfs2_meta_new(gl, blkno);
		memcpy(bh_ip->b_data, bh_log->b_data, bh_log->b_size);

		/* Unescape */
		if (esc) {
			__be32 *eptr = (__be32 *)bh_ip->b_data;
			*eptr = cpu_to_be32(GFS2_MAGIC);
		}
		mark_buffer_dirty(bh_ip);

		brelse(bh_log);
		brelse(bh_ip);
		if (error)
			break;

		sdp->sd_replayed_blocks++;
	}

	return error;
}

/* FIXME: sort out accounting for log blocks etc. */

static void databuf_lo_after_scan(struct gfs2_jdesc *jd, int error, int pass)
{
705 706
	struct gfs2_inode *ip = GFS2_I(jd->jd_inode);
	struct gfs2_sbd *sdp = GFS2_SB(jd->jd_inode);
707 708

	if (error) {
S
Steven Whitehouse 已提交
709
		gfs2_meta_sync(ip->i_gl);
710 711 712 713 714 715
		return;
	}
	if (pass != 1)
		return;

	/* data sync? */
S
Steven Whitehouse 已提交
716
	gfs2_meta_sync(ip->i_gl);
717 718 719 720 721 722 723 724 725 726 727 728

	fs_info(sdp, "jid=%u: Replayed %u of %u data blocks\n",
		jd->jd_jid, sdp->sd_replayed_blocks, sdp->sd_found_blocks);
}

static void databuf_lo_after_commit(struct gfs2_sbd *sdp, struct gfs2_ail *ai)
{
	struct list_head *head = &sdp->sd_log_le_databuf;
	struct gfs2_bufdata *bd;

	while (!list_empty(head)) {
		bd = list_entry(head->next, struct gfs2_bufdata, bd_le.le_list);
S
Steven Whitehouse 已提交
729
		list_del_init(&bd->bd_le.le_list);
730 731 732
		sdp->sd_log_num_databuf--;
		gfs2_unpin(sdp, bd->bd_bh, ai);
	}
D
David Teigland 已提交
733 734 735
	gfs2_assert_warn(sdp, !sdp->sd_log_num_databuf);
}

736

737
const struct gfs2_log_operations gfs2_buf_lops = {
D
David Teigland 已提交
738 739 740 741 742 743
	.lo_add = buf_lo_add,
	.lo_before_commit = buf_lo_before_commit,
	.lo_after_commit = buf_lo_after_commit,
	.lo_before_scan = buf_lo_before_scan,
	.lo_scan_elements = buf_lo_scan_elements,
	.lo_after_scan = buf_lo_after_scan,
744
	.lo_name = "buf",
D
David Teigland 已提交
745 746
};

747
const struct gfs2_log_operations gfs2_revoke_lops = {
D
David Teigland 已提交
748 749 750 751 752
	.lo_add = revoke_lo_add,
	.lo_before_commit = revoke_lo_before_commit,
	.lo_before_scan = revoke_lo_before_scan,
	.lo_scan_elements = revoke_lo_scan_elements,
	.lo_after_scan = revoke_lo_after_scan,
753
	.lo_name = "revoke",
D
David Teigland 已提交
754 755
};

756
const struct gfs2_log_operations gfs2_rg_lops = {
D
David Teigland 已提交
757 758
	.lo_add = rg_lo_add,
	.lo_after_commit = rg_lo_after_commit,
759
	.lo_name = "rg",
D
David Teigland 已提交
760 761
};

762
const struct gfs2_log_operations gfs2_databuf_lops = {
D
David Teigland 已提交
763 764
	.lo_add = databuf_lo_add,
	.lo_before_commit = databuf_lo_before_commit,
765 766 767
	.lo_after_commit = databuf_lo_after_commit,
	.lo_scan_elements = databuf_lo_scan_elements,
	.lo_after_scan = databuf_lo_after_scan,
768
	.lo_name = "databuf",
D
David Teigland 已提交
769 770
};

771
const struct gfs2_log_operations *gfs2_log_ops[] = {
772
	&gfs2_databuf_lops,
D
David Teigland 已提交
773 774
	&gfs2_buf_lops,
	&gfs2_rg_lops,
775
	&gfs2_revoke_lops,
776
	NULL,
D
David Teigland 已提交
777 778
};