lops.c 20.5 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>
16
#include <linux/lm_interface.h>
D
David Teigland 已提交
17 18

#include "gfs2.h"
19
#include "incore.h"
D
David Teigland 已提交
20 21 22 23 24 25 26
#include "glock.h"
#include "log.h"
#include "lops.h"
#include "meta_io.h"
#include "recovery.h"
#include "rgrp.h"
#include "trans.h"
27
#include "util.h"
D
David Teigland 已提交
28 29 30 31

static void glock_lo_add(struct gfs2_sbd *sdp, struct gfs2_log_element *le)
{
	struct gfs2_glock *gl;
32
	struct gfs2_trans *tr = current->journal_info;
D
David Teigland 已提交
33

34
	tr->tr_touched = 1;
D
David Teigland 已提交
35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74

	if (!list_empty(&le->le_list))
		return;

	gl = container_of(le, struct gfs2_glock, gl_le);
	if (gfs2_assert_withdraw(sdp, gfs2_glock_is_held_excl(gl)))
		return;
	gfs2_glock_hold(gl);
	set_bit(GLF_DIRTY, &gl->gl_flags);

	gfs2_log_lock(sdp);
	sdp->sd_log_num_gl++;
	list_add(&le->le_list, &sdp->sd_log_le_gl);
	gfs2_log_unlock(sdp);
}

static void glock_lo_after_commit(struct gfs2_sbd *sdp, struct gfs2_ail *ai)
{
	struct list_head *head = &sdp->sd_log_le_gl;
	struct gfs2_glock *gl;

	while (!list_empty(head)) {
		gl = list_entry(head->next, struct gfs2_glock, gl_le.le_list);
		list_del_init(&gl->gl_le.le_list);
		sdp->sd_log_num_gl--;

		gfs2_assert_withdraw(sdp, gfs2_glock_is_held_excl(gl));
		gfs2_glock_put(gl);
	}
	gfs2_assert_warn(sdp, !sdp->sd_log_num_gl);
}

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);
	struct gfs2_trans *tr;

	if (!list_empty(&bd->bd_list_tr))
		return;

75
	tr = current->journal_info;
D
David Teigland 已提交
76 77 78 79 80 81 82 83 84 85
	tr->tr_touched = 1;
	tr->tr_num_buf++;
	list_add(&bd->bd_list_tr, &tr->tr_list_buf);

	if (!list_empty(&le->le_list))
		return;

	gfs2_trans_add_gl(bd->bd_gl);

	gfs2_meta_check(sdp, bd->bd_bh);
86
	gfs2_pin(sdp, bd->bd_bh);
D
David Teigland 已提交
87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120

	gfs2_log_lock(sdp);
	sdp->sd_log_num_buf++;
	list_add(&le->le_list, &sdp->sd_log_le_buf);
	gfs2_log_unlock(sdp);

	tr->tr_num_buf_new++;
}

static void buf_lo_incore_commit(struct gfs2_sbd *sdp, struct gfs2_trans *tr)
{
	struct list_head *head = &tr->tr_list_buf;
	struct gfs2_bufdata *bd;

	while (!list_empty(head)) {
		bd = list_entry(head->next, struct gfs2_bufdata, bd_list_tr);
		list_del_init(&bd->bd_list_tr);
		tr->tr_num_buf--;
	}
	gfs2_assert_warn(sdp, !tr->tr_num_buf);
}

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;
	unsigned int total = sdp->sd_log_num_buf;
	unsigned int offset = sizeof(struct gfs2_log_descriptor);
	unsigned int limit;
	unsigned int num;
	unsigned n;
	__be64 *ptr;

S
Steven Whitehouse 已提交
121
	offset += sizeof(__be64) - 1;
D
David Teigland 已提交
122 123 124 125 126 127 128 129 130 131
	offset &= ~(sizeof(__be64) - 1);
	limit = (sdp->sd_sb.sb_bsize - offset)/sizeof(__be64);
	/* for 4k blocks, limit = 503 */

	bd1 = bd2 = list_prepare_entry(bd1, &sdp->sd_log_le_buf, bd_le.le_list);
	while(total) {
		num = total;
		if (total > limit)
			num = limit;
		bh = gfs2_log_get_buf(sdp);
132
		sdp->sd_log_num_hdrs++;
D
David Teigland 已提交
133 134 135
		ld = (struct gfs2_log_descriptor *)bh->b_data;
		ptr = (__be64 *)(bh->b_data + offset);
		ld->ld_header.mh_magic = cpu_to_be32(GFS2_MAGIC);
136 137
		ld->ld_header.mh_type = cpu_to_be32(GFS2_METATYPE_LD);
		ld->ld_header.mh_format = cpu_to_be32(GFS2_FORMAT_LD);
D
David Teigland 已提交
138 139 140 141 142 143 144
		ld->ld_type = cpu_to_be32(GFS2_LOG_DESC_METADATA);
		ld->ld_length = cpu_to_be32(num + 1);
		ld->ld_data1 = cpu_to_be32(num);
		ld->ld_data2 = cpu_to_be32(0);
		memset(ld->ld_reserved, 0, sizeof(ld->ld_reserved));

		n = 0;
145 146
		list_for_each_entry_continue(bd1, &sdp->sd_log_le_buf,
					     bd_le.le_list) {
D
David Teigland 已提交
147 148 149 150 151 152 153 154 155
			*ptr++ = cpu_to_be64(bd1->bd_bh->b_blocknr);
			if (++n >= num)
				break;
		}

		set_buffer_dirty(bh);
		ll_rw_block(WRITE, 1, &bh);

		n = 0;
156 157
		list_for_each_entry_continue(bd2, &sdp->sd_log_le_buf,
					     bd_le.le_list) {
D
David Teigland 已提交
158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178
			bh = gfs2_log_fake_buf(sdp, bd2->bd_bh);
			set_buffer_dirty(bh);
			ll_rw_block(WRITE, 1, &bh);
			if (++n >= num)
				break;
		}

		total -= num;
	}
}

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--;

179
		gfs2_unpin(sdp, bd->bd_bh, ai);
D
David Teigland 已提交
180 181 182 183 184 185 186
	}
	gfs2_assert_warn(sdp, !sdp->sd_log_num_buf);
}

static void buf_lo_before_scan(struct gfs2_jdesc *jd,
			       struct gfs2_log_header *head, int pass)
{
187
	struct gfs2_sbd *sdp = GFS2_SB(jd->jd_inode);
D
David Teigland 已提交
188 189 190 191 192 193 194 195 196 197 198 199

	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)
{
200 201
	struct gfs2_inode *ip = GFS2_I(jd->jd_inode);
	struct gfs2_sbd *sdp = GFS2_SB(jd->jd_inode);
202
	struct gfs2_glock *gl = ip->i_gl;
D
David Teigland 已提交
203 204
	unsigned int blks = be32_to_cpu(ld->ld_data1);
	struct buffer_head *bh_log, *bh_ip;
205
	u64 blkno;
D
David Teigland 已提交
206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221
	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
222 223
		if (error)
			return error;
D
David Teigland 已提交
224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246

		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)
{
247 248
	struct gfs2_inode *ip = GFS2_I(jd->jd_inode);
	struct gfs2_sbd *sdp = GFS2_SB(jd->jd_inode);
D
David Teigland 已提交
249 250

	if (error) {
S
Steven Whitehouse 已提交
251
		gfs2_meta_sync(ip->i_gl, DIO_START | DIO_WAIT);
D
David Teigland 已提交
252 253 254 255 256
		return;
	}
	if (pass != 1)
		return;

257
	gfs2_meta_sync(ip->i_gl, DIO_START | DIO_WAIT);
D
David Teigland 已提交
258 259 260 261 262 263 264 265 266

	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;

267
	tr = current->journal_info;
D
David Teigland 已提交
268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291
	tr->tr_touched = 1;
	tr->tr_num_revoke++;

	gfs2_log_lock(sdp);
	sdp->sd_log_num_revoke++;
	list_add(&le->le_list, &sdp->sd_log_le_revoke);
	gfs2_log_unlock(sdp);
}

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;
	struct gfs2_revoke *rv;

	if (!sdp->sd_log_num_revoke)
		return;

	bh = gfs2_log_get_buf(sdp);
	ld = (struct gfs2_log_descriptor *)bh->b_data;
	ld->ld_header.mh_magic = cpu_to_be32(GFS2_MAGIC);
292 293
	ld->ld_header.mh_type = cpu_to_be32(GFS2_METATYPE_LD);
	ld->ld_header.mh_format = cpu_to_be32(GFS2_FORMAT_LD);
D
David Teigland 已提交
294
	ld->ld_type = cpu_to_be32(GFS2_LOG_DESC_REVOKE);
295
	ld->ld_length = cpu_to_be32(gfs2_struct2blk(sdp, sdp->sd_log_num_revoke,
296
						    sizeof(u64)));
D
David Teigland 已提交
297 298 299 300 301 302 303
	ld->ld_data1 = cpu_to_be32(sdp->sd_log_num_revoke);
	ld->ld_data2 = cpu_to_be32(0);
	memset(ld->ld_reserved, 0, sizeof(ld->ld_reserved));
	offset = sizeof(struct gfs2_log_descriptor);

	while (!list_empty(head)) {
		rv = list_entry(head->next, struct gfs2_revoke, rv_le.le_list);
304
		list_del_init(&rv->rv_le.le_list);
D
David Teigland 已提交
305 306
		sdp->sd_log_num_revoke--;

307
		if (offset + sizeof(u64) > sdp->sd_sb.sb_bsize) {
D
David Teigland 已提交
308 309 310 311 312 313
			set_buffer_dirty(bh);
			ll_rw_block(WRITE, 1, &bh);

			bh = gfs2_log_get_buf(sdp);
			mh = (struct gfs2_meta_header *)bh->b_data;
			mh->mh_magic = cpu_to_be32(GFS2_MAGIC);
314 315
			mh->mh_type = cpu_to_be32(GFS2_METATYPE_LB);
			mh->mh_format = cpu_to_be32(GFS2_FORMAT_LB);
D
David Teigland 已提交
316 317 318 319 320 321
			offset = sizeof(struct gfs2_meta_header);
		}

		*(__be64 *)(bh->b_data + offset) = cpu_to_be64(rv->rv_blkno);
		kfree(rv);

322
		offset += sizeof(u64);
D
David Teigland 已提交
323 324 325 326 327 328 329 330 331 332
	}
	gfs2_assert_withdraw(sdp, !sdp->sd_log_num_revoke);

	set_buffer_dirty(bh);
	ll_rw_block(WRITE, 1, &bh);
}

static void revoke_lo_before_scan(struct gfs2_jdesc *jd,
				  struct gfs2_log_header *head, int pass)
{
333
	struct gfs2_sbd *sdp = GFS2_SB(jd->jd_inode);
D
David Teigland 已提交
334 335 336 337 338 339 340 341 342 343 344 345

	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)
{
346
	struct gfs2_sbd *sdp = GFS2_SB(jd->jd_inode);
D
David Teigland 已提交
347 348 349 350
	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;
351
	u64 blkno;
D
David Teigland 已提交
352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367
	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);

368
		while (offset + sizeof(u64) <= sdp->sd_sb.sb_bsize) {
D
David Teigland 已提交
369 370 371 372 373 374 375 376 377 378
			blkno = be64_to_cpu(*(__be64 *)(bh->b_data + offset));

			error = gfs2_revoke_add(sdp, blkno, start);
			if (error < 0)
				return error;
			else if (error)
				sdp->sd_found_revokes++;

			if (!--revokes)
				break;
379
			offset += sizeof(u64);
D
David Teigland 已提交
380 381 382 383 384 385 386 387 388 389 390 391
		}

		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)
{
392
	struct gfs2_sbd *sdp = GFS2_SB(jd->jd_inode);
D
David Teigland 已提交
393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409

	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;
410
	struct gfs2_trans *tr = current->journal_info;
D
David Teigland 已提交
411

412
	tr->tr_touched = 1;
D
David Teigland 已提交
413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441

	if (!list_empty(&le->le_list))
		return;

	rgd = container_of(le, struct gfs2_rgrpd, rd_le);
	gfs2_rgrp_bh_hold(rgd);

	gfs2_log_lock(sdp);
	sdp->sd_log_num_rg++;
	list_add(&le->le_list, &sdp->sd_log_le_rg);
	gfs2_log_unlock(sdp);	
}

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);
}

442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457
/**
 * 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
458 459
static void databuf_lo_add(struct gfs2_sbd *sdp, struct gfs2_log_element *le)
{
460
	struct gfs2_bufdata *bd = container_of(le, struct gfs2_bufdata, bd_le);
461
	struct gfs2_trans *tr = current->journal_info;
462
	struct address_space *mapping = bd->bd_bh->b_page->mapping;
463
	struct gfs2_inode *ip = GFS2_I(mapping->host);
D
David Teigland 已提交
464

465
	tr->tr_touched = 1;
466
	if (list_empty(&bd->bd_list_tr) &&
467 468 469 470
	    (ip->i_di.di_flags & GFS2_DIF_JDATA)) {
		tr->tr_num_buf++;
		list_add(&bd->bd_list_tr, &tr->tr_list_buf);
		gfs2_pin(sdp, bd->bd_bh);
S
Steven Whitehouse 已提交
471
		tr->tr_num_buf_new++;
472
	}
473
	gfs2_trans_add_gl(bd->bd_gl);
D
David Teigland 已提交
474
	gfs2_log_lock(sdp);
475
	if (list_empty(&le->le_list)) {
476 477 478 479 480
		if (ip->i_di.di_flags & GFS2_DIF_JDATA)
			sdp->sd_log_num_jdata++;
		sdp->sd_log_num_databuf++;
		list_add(&le->le_list, &sdp->sd_log_le_databuf);
	}
D
David Teigland 已提交
481 482 483
	gfs2_log_unlock(sdp);
}

484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507
static int gfs2_check_magic(struct buffer_head *bh)
{
	struct page *page = bh->b_page;
	void *kaddr;
	__be32 *ptr;
	int rv = 0;

	kaddr = kmap_atomic(page, KM_USER0);
	ptr = kaddr + bh_offset(bh);
	if (*ptr == cpu_to_be32(GFS2_MAGIC))
		rv = 1;
	kunmap_atomic(page, KM_USER0);

	return rv;
}

/**
 * databuf_lo_before_commit - Scan the data buffers, writing as we go
 *
 * Here we scan through the lists of buffers and make the assumption
 * that any buffer thats been pinned is being journaled, and that
 * any unpinned buffer is an ordered write data buffer and therefore
 * will be written back rather than journaled.
 */
D
David Teigland 已提交
508 509 510
static void databuf_lo_before_commit(struct gfs2_sbd *sdp)
{
	LIST_HEAD(started);
511 512 513 514 515 516 517 518
	struct gfs2_bufdata *bd1 = NULL, *bd2, *bdt;
	struct buffer_head *bh = NULL;
	unsigned int offset = sizeof(struct gfs2_log_descriptor);
	struct gfs2_log_descriptor *ld;
	unsigned int limit;
	unsigned int total_dbuf = sdp->sd_log_num_databuf;
	unsigned int total_jdata = sdp->sd_log_num_jdata;
	unsigned int num, n;
519
	__be64 *ptr = NULL;
D
David Teigland 已提交
520

521
	offset += 2*sizeof(__be64) - 1;
522 523
	offset &= ~(2*sizeof(__be64) - 1);
	limit = (sdp->sd_sb.sb_bsize - offset)/sizeof(__be64);
D
David Teigland 已提交
524

525 526 527 528
	/*
	 * Start writing ordered buffers, write journaled buffers
	 * into the log along with a header
	 */
529
	gfs2_log_lock(sdp);
530 531
	bd2 = bd1 = list_prepare_entry(bd1, &sdp->sd_log_le_databuf,
				       bd_le.le_list);
532 533 534 535 536
	while(total_dbuf) {
		num = total_jdata;
		if (num > limit)
			num = limit;
		n = 0;
537 538 539
		list_for_each_entry_safe_continue(bd1, bdt,
						  &sdp->sd_log_le_databuf,
						  bd_le.le_list) {
540 541 542 543 544
			/* An ordered write buffer */
			if (bd1->bd_bh && !buffer_pinned(bd1->bd_bh)) {
				list_move(&bd1->bd_le.le_list, &started);
				if (bd1 == bd2) {
					bd2 = NULL;
545 546 547
					bd2 = list_prepare_entry(bd2,
							&sdp->sd_log_le_databuf,
							bd_le.le_list);
548 549 550 551 552
				}
				total_dbuf--;
				if (bd1->bd_bh) {
					get_bh(bd1->bd_bh);
					if (buffer_dirty(bd1->bd_bh)) {
553
						gfs2_log_unlock(sdp);
554
						wait_on_buffer(bd1->bd_bh);
555 556
						ll_rw_block(WRITE, 1,
							    &bd1->bd_bh);
557
						gfs2_log_lock(sdp);
558 559 560 561 562 563 564 565 566 567
					}
					brelse(bd1->bd_bh);
					continue;
				}
				continue;
			} else if (bd1->bd_bh) { /* A journaled buffer */
				int magic;
				gfs2_log_unlock(sdp);
				if (!bh) {
					bh = gfs2_log_get_buf(sdp);
568
					sdp->sd_log_num_hdrs++;
569 570
					ld = (struct gfs2_log_descriptor *)
					     bh->b_data;
571
					ptr = (__be64 *)(bh->b_data + offset);
572 573 574
					ld->ld_header.mh_magic =
						cpu_to_be32(GFS2_MAGIC);
					ld->ld_header.mh_type =
575
						cpu_to_be32(GFS2_METATYPE_LD);
576
					ld->ld_header.mh_format =
577
						cpu_to_be32(GFS2_FORMAT_LD);
578 579
					ld->ld_type =
						cpu_to_be32(GFS2_LOG_DESC_JDATA);
580 581 582 583 584 585 586 587 588 589 590
					ld->ld_length = cpu_to_be32(num + 1);
					ld->ld_data1 = cpu_to_be32(num);
					ld->ld_data2 = cpu_to_be32(0);
					memset(ld->ld_reserved, 0, sizeof(ld->ld_reserved));
				}
				magic = gfs2_check_magic(bd1->bd_bh);
				*ptr++ = cpu_to_be64(bd1->bd_bh->b_blocknr);
				*ptr++ = cpu_to_be64((__u64)magic);
				clear_buffer_escaped(bd1->bd_bh);
				if (unlikely(magic != 0))
					set_buffer_escaped(bd1->bd_bh);
591
				gfs2_log_lock(sdp);
592 593
				if (n++ > num)
					break;
594 595 596 597 598 599 600 601 602 603 604
			} else if (!bd1->bd_bh) {
				total_dbuf--;
				sdp->sd_log_num_databuf--;
				list_del_init(&bd1->bd_le.le_list);
				if (bd1 == bd2) {
					bd2 = NULL;
					bd2 = list_prepare_entry(bd2,
						&sdp->sd_log_le_databuf,
						bd_le.le_list);
                                }
				kmem_cache_free(gfs2_bufdata_cachep, bd1);
605 606
			}
		}
607
		gfs2_log_unlock(sdp);
D
David Teigland 已提交
608
		if (bh) {
609 610 611 612 613
			set_buffer_dirty(bh);
			ll_rw_block(WRITE, 1, &bh);
			bh = NULL;
		}
		n = 0;
614
		gfs2_log_lock(sdp);
615 616
		list_for_each_entry_continue(bd2, &sdp->sd_log_le_databuf,
					     bd_le.le_list) {
617 618 619
			if (!bd2->bd_bh)
				continue;
			/* copy buffer if it needs escaping */
620
			gfs2_log_unlock(sdp);
621 622 623 624 625
			if (unlikely(buffer_escaped(bd2->bd_bh))) {
				void *kaddr;
				struct page *page = bd2->bd_bh->b_page;
				bh = gfs2_log_get_buf(sdp);
				kaddr = kmap_atomic(page, KM_USER0);
626 627 628
				memcpy(bh->b_data,
				       kaddr + bh_offset(bd2->bd_bh),
				       sdp->sd_sb.sb_bsize);
629 630 631 632
				kunmap_atomic(page, KM_USER0);
				*(__be32 *)bh->b_data = 0;
			} else {
				bh = gfs2_log_fake_buf(sdp, bd2->bd_bh);
D
David Teigland 已提交
633
			}
634 635
			set_buffer_dirty(bh);
			ll_rw_block(WRITE, 1, &bh);
636
			gfs2_log_lock(sdp);
637 638 639 640 641 642
			if (++n >= num)
				break;
		}
		bh = NULL;
		total_dbuf -= num;
		total_jdata -= num;
D
David Teigland 已提交
643
	}
644 645
	gfs2_log_unlock(sdp);

646
	/* Wait on all ordered buffers */
D
David Teigland 已提交
647
	while (!list_empty(&started)) {
648
		gfs2_log_lock(sdp);
649 650
		bd1 = list_entry(started.next, struct gfs2_bufdata,
				 bd_le.le_list);
651
		list_del_init(&bd1->bd_le.le_list);
D
David Teigland 已提交
652
		sdp->sd_log_num_databuf--;
653
		bh = bd1->bd_bh;
D
David Teigland 已提交
654
		if (bh) {
655
			bh->b_private = NULL;
656
			get_bh(bh);
D
David Teigland 已提交
657 658 659 660 661 662
			gfs2_log_unlock(sdp);
			wait_on_buffer(bh);
			brelse(bh);
		} else
			gfs2_log_unlock(sdp);

663
		kmem_cache_free(gfs2_bufdata_cachep, bd1);
D
David Teigland 已提交
664 665
	}

666 667 668 669 670 671 672 673
	/* We've removed all the ordered write bufs here, so only jdata left */
	gfs2_assert_warn(sdp, sdp->sd_log_num_databuf == sdp->sd_log_num_jdata);
}

static int databuf_lo_scan_elements(struct gfs2_jdesc *jd, unsigned int start,
				    struct gfs2_log_descriptor *ld,
				    __be64 *ptr, int pass)
{
674 675
	struct gfs2_inode *ip = GFS2_I(jd->jd_inode);
	struct gfs2_sbd *sdp = GFS2_SB(jd->jd_inode);
676
	struct gfs2_glock *gl = ip->i_gl;
677 678
	unsigned int blks = be32_to_cpu(ld->ld_data1);
	struct buffer_head *bh_log, *bh_ip;
679 680
	u64 blkno;
	u64 esc;
681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724
	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)
{
725 726
	struct gfs2_inode *ip = GFS2_I(jd->jd_inode);
	struct gfs2_sbd *sdp = GFS2_SB(jd->jd_inode);
727 728

	if (error) {
729
		gfs2_meta_sync(ip->i_gl,
730
			       DIO_START | DIO_WAIT);
731 732 733 734 735 736
		return;
	}
	if (pass != 1)
		return;

	/* data sync? */
737
	gfs2_meta_sync(ip->i_gl, DIO_START | DIO_WAIT);
738 739 740 741 742 743 744 745 746 747 748 749

	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
750
		list_del_init(&bd->bd_le.le_list);
751 752 753 754
		sdp->sd_log_num_databuf--;
		sdp->sd_log_num_jdata--;
		gfs2_unpin(sdp, bd->bd_bh, ai);
	}
D
David Teigland 已提交
755
	gfs2_assert_warn(sdp, !sdp->sd_log_num_databuf);
756
	gfs2_assert_warn(sdp, !sdp->sd_log_num_jdata);
D
David Teigland 已提交
757 758
}

759

760
const struct gfs2_log_operations gfs2_glock_lops = {
D
David Teigland 已提交
761 762
	.lo_add = glock_lo_add,
	.lo_after_commit = glock_lo_after_commit,
763
	.lo_name = "glock",
D
David Teigland 已提交
764 765
};

766
const struct gfs2_log_operations gfs2_buf_lops = {
D
David Teigland 已提交
767 768 769 770 771 772 773
	.lo_add = buf_lo_add,
	.lo_incore_commit = buf_lo_incore_commit,
	.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,
774
	.lo_name = "buf",
D
David Teigland 已提交
775 776
};

777
const struct gfs2_log_operations gfs2_revoke_lops = {
D
David Teigland 已提交
778 779 780 781 782
	.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,
783
	.lo_name = "revoke",
D
David Teigland 已提交
784 785
};

786
const struct gfs2_log_operations gfs2_rg_lops = {
D
David Teigland 已提交
787 788
	.lo_add = rg_lo_add,
	.lo_after_commit = rg_lo_after_commit,
789
	.lo_name = "rg",
D
David Teigland 已提交
790 791
};

792
const struct gfs2_log_operations gfs2_databuf_lops = {
D
David Teigland 已提交
793
	.lo_add = databuf_lo_add,
794
	.lo_incore_commit = buf_lo_incore_commit,
D
David Teigland 已提交
795
	.lo_before_commit = databuf_lo_before_commit,
796 797 798
	.lo_after_commit = databuf_lo_after_commit,
	.lo_scan_elements = databuf_lo_scan_elements,
	.lo_after_scan = databuf_lo_after_scan,
799
	.lo_name = "databuf",
D
David Teigland 已提交
800 801
};

802
const struct gfs2_log_operations *gfs2_log_ops[] = {
D
David Teigland 已提交
803 804 805 806 807
	&gfs2_glock_lops,
	&gfs2_buf_lops,
	&gfs2_revoke_lops,
	&gfs2_rg_lops,
	&gfs2_databuf_lops,
808
	NULL,
D
David Teigland 已提交
809 810
};