Merge branch 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs-2.6

* 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs-2.6:
  jbd: change the field "b_cow_tid" of struct journal_head from type unsigned to tid_t
  ext3.txt: update the links in the section "useful links" to the latest ones
  ext3: Fix data corruption in inodes with journalled data
  ext2: check xattr name_len before acquiring xattr_sem in ext2_xattr_get
  ext3: Fix compilation with -DDX_DEBUG
  quota: Remove unused declaration
  jbd: Use WRITE_SYNC in journal checkpoint.
  jbd: Fix oops in journal_remove_journal_head()
  ext3: Return -EINVAL when start is beyond the end of fs in ext3_trim_fs()
  ext3/ioctl.c: silence sparse warnings about different address spaces
  ext3/ext4 Documentation: remove bh/nobh since it has been deprecated
  ext3: Improve truncate error handling
  ext3: use proper little-endian bitops
  ext2: include fs.h into ext2_fs.h
  ext3: Fix oops in ext3_try_to_allocate_with_rsv()
  jbd: fix a bug of leaking jh->b_jcount
  jbd: remove dependency on __GFP_NOFAIL
  ext3: Convert ext3 to new truncate calling convention
  jbd: Add fixed tracepoints
  ext3: Add fixed tracepoints

Resolve conflicts in fs/ext3/fsync.c due to fsync locking push-down and
new fixed tracepoints.

Documentation/filesystems/ext3.txt

@@ -147,15 +147,6 @@ grpjquota=<file>	during journal replay. They replace the above
 			package for more details
 			(http://sourceforge.net/projects/linuxquota).
 
-bh		(*)	ext3 associates buffer heads to data pages to
-nobh			(a) cache disk block mapping information
-			(b) link pages into transaction to provide
-			    ordering guarantees.
-			"bh" option forces use of buffer heads.
-			"nobh" option tries to avoid associating buffer
-			heads (supported only for "writeback" mode).
-
-
 Specification
 =============
 Ext3 shares all disk implementation with the ext2 filesystem, and adds
@@ -227,5 +218,5 @@ kernel source:	<file:fs/ext3/>
 programs: 	http://e2fsprogs.sourceforge.net/
 		http://ext2resize.sourceforge.net
 
-useful links:	http://www.ibm.com/developerworks/library/l-fs7.html
-		http://www.ibm.com/developerworks/library/l-fs8.html
+useful links:	http://www.ibm.com/developerworks/library/l-fs7/index.html
+        http://www.ibm.com/developerworks/library/l-fs8/index.html

Documentation/filesystems/ext4.txt

@@ -68,12 +68,12 @@ Note: More extensive information for getting started with ext4 can be
     '-o barriers=[0|1]' mount option for both ext3 and ext4 filesystems
     for a fair comparison.  When tuning ext3 for best benchmark numbers,
     it is often worthwhile to try changing the data journaling mode; '-o
-    data=writeback,nobh' can be faster for some workloads.  (Note
-    however that running mounted with data=writeback can potentially
-    leave stale data exposed in recently written files in case of an
-    unclean shutdown, which could be a security exposure in some
-    situations.)  Configuring the filesystem with a large journal can
-    also be helpful for metadata-intensive workloads.
+    data=writeback' can be faster for some workloads.  (Note however that
+    running mounted with data=writeback can potentially leave stale data
+    exposed in recently written files in case of an unclean shutdown,
+    which could be a security exposure in some situations.)  Configuring
+    the filesystem with a large journal can also be helpful for
+    metadata-intensive workloads.
 
 2. Features
 ===========
@@ -272,14 +272,6 @@ grpjquota=<file>	during journal replay. They replace the above
 			package for more details
 			(http://sourceforge.net/projects/linuxquota).
 
-bh		(*)	ext4 associates buffer heads to data pages to
-nobh			(a) cache disk block mapping information
-			(b) link pages into transaction to provide
-			    ordering guarantees.
-			"bh" option forces use of buffer heads.
-			"nobh" option tries to avoid associating buffer
-			heads (supported only for "writeback" mode).
-
 stripe=n		Number of filesystem blocks that mballoc will try
 			to use for allocation size and alignment. For RAID5/6
 			systems this should be the number of data
@@ -393,8 +385,7 @@ dioread_nolock		locking. If the dioread_nolock option is specified
 			write and convert the extent to initialized after IO
 			completes. This approach allows ext4 code to avoid
 			using inode mutex, which improves scalability on high
-			speed storages. However this does not work with nobh
-			option and the mount will fail. Nor does it work with
+			speed storages. However this does not work with
 			data journaling and dioread_nolock option will be
 			ignored with kernel warning. Note that dioread_nolock
 			code path is only used for extent-based files.

fs/ext2/xattr.c

@@ -161,6 +161,10 @@ ext2_xattr_get(struct inode *inode, int name_index, const char *name,
 
 	if (name == NULL)
 		return -EINVAL;
+	name_len = strlen(name);
+	if (name_len > 255)
+		return -ERANGE;
+
 	down_read(&EXT2_I(inode)->xattr_sem);
 	error = -ENODATA;
 	if (!EXT2_I(inode)->i_file_acl)
@@ -181,12 +185,8 @@ bad_block:	ext2_error(inode->i_sb, "ext2_xattr_get",
 		error = -EIO;
 		goto cleanup;
 	}
-	/* find named attribute */
-	name_len = strlen(name);
 
-	error = -ERANGE;
-	if (name_len > 255)
-		goto cleanup;
+	/* find named attribute */
 	entry = FIRST_ENTRY(bh);
 	while (!IS_LAST_ENTRY(entry)) {
 		struct ext2_xattr_entry *next =

fs/ext3/balloc.c

@@ -21,6 +21,7 @@
 #include <linux/quotaops.h>
 #include <linux/buffer_head.h>
 #include <linux/blkdev.h>
+#include <trace/events/ext3.h>
 
 /*
  * balloc.c contains the blocks allocation and deallocation routines
@@ -161,6 +162,7 @@ read_block_bitmap(struct super_block *sb, unsigned int block_group)
 	desc = ext3_get_group_desc(sb, block_group, NULL);
 	if (!desc)
 		return NULL;
+	trace_ext3_read_block_bitmap(sb, block_group);
 	bitmap_blk = le32_to_cpu(desc->bg_block_bitmap);
 	bh = sb_getblk(sb, bitmap_blk);
 	if (unlikely(!bh)) {
@@ -351,6 +353,7 @@ void ext3_rsv_window_add(struct super_block *sb,
 	struct rb_node * parent = NULL;
 	struct ext3_reserve_window_node *this;
 
+	trace_ext3_rsv_window_add(sb, rsv);
 	while (*p)
 	{
 		parent = *p;
@@ -476,8 +479,10 @@ void ext3_discard_reservation(struct inode *inode)
 	rsv = &block_i->rsv_window_node;
 	if (!rsv_is_empty(&rsv->rsv_window)) {
 		spin_lock(rsv_lock);
-		if (!rsv_is_empty(&rsv->rsv_window))
+		if (!rsv_is_empty(&rsv->rsv_window)) {
+			trace_ext3_discard_reservation(inode, rsv);
 			rsv_window_remove(inode->i_sb, rsv);
+		}
 		spin_unlock(rsv_lock);
 	}
 }
@@ -683,14 +688,10 @@ void ext3_free_blocks_sb(handle_t *handle, struct super_block *sb,
 void ext3_free_blocks(handle_t *handle, struct inode *inode,
 			ext3_fsblk_t block, unsigned long count)
 {
-	struct super_block * sb;
+	struct super_block *sb = inode->i_sb;
 	unsigned long dquot_freed_blocks;
 
-	sb = inode->i_sb;
-	if (!sb) {
-		printk ("ext3_free_blocks: nonexistent device");
-		return;
-	}
+	trace_ext3_free_blocks(inode, block, count);
 	ext3_free_blocks_sb(handle, sb, block, count, &dquot_freed_blocks);
 	if (dquot_freed_blocks)
 		dquot_free_block(inode, dquot_freed_blocks);
@@ -1136,6 +1137,7 @@ static int alloc_new_reservation(struct ext3_reserve_window_node *my_rsv,
 	else
 		start_block = grp_goal + group_first_block;
 
+	trace_ext3_alloc_new_reservation(sb, start_block);
 	size = my_rsv->rsv_goal_size;
 
 	if (!rsv_is_empty(&my_rsv->rsv_window)) {
@@ -1230,8 +1232,11 @@ static int alloc_new_reservation(struct ext3_reserve_window_node *my_rsv,
 	 * check if the first free block is within the
 	 * free space we just reserved
 	 */
-	if (start_block >= my_rsv->rsv_start && start_block <= my_rsv->rsv_end)
+	if (start_block >= my_rsv->rsv_start &&
+	    start_block <= my_rsv->rsv_end) {
+		trace_ext3_reserved(sb, start_block, my_rsv);
 		return 0;		/* success */
+	}
 	/*
 	 * if the first free bit we found is out of the reservable space
 	 * continue search for next reservable space,
@@ -1514,10 +1519,6 @@ ext3_fsblk_t ext3_new_blocks(handle_t *handle, struct inode *inode,
 
 	*errp = -ENOSPC;
 	sb = inode->i_sb;
-	if (!sb) {
-		printk("ext3_new_block: nonexistent device");
-		return 0;
-	}
 
 	/*
 	 * Check quota for allocation of this block.
@@ -1528,8 +1529,10 @@ ext3_fsblk_t ext3_new_blocks(handle_t *handle, struct inode *inode,
 		return 0;
 	}
 
+	trace_ext3_request_blocks(inode, goal, num);
+
 	sbi = EXT3_SB(sb);
-	es = EXT3_SB(sb)->s_es;
+	es = sbi->s_es;
 	ext3_debug("goal=%lu.\n", goal);
 	/*
 	 * Allocate a block from reservation only when
@@ -1742,6 +1745,10 @@ ext3_fsblk_t ext3_new_blocks(handle_t *handle, struct inode *inode,
 	brelse(bitmap_bh);
 	dquot_free_block(inode, *count-num);
 	*count = num;
+
+	trace_ext3_allocate_blocks(inode, goal, num,
+				   (unsigned long long)ret_block);
+
 	return ret_block;
 
 io_error:
@@ -1996,6 +2003,7 @@ ext3_grpblk_t ext3_trim_all_free(struct super_block *sb, unsigned int group,
 		if ((next - start) < minblocks)
 			goto free_extent;
 
+		trace_ext3_discard_blocks(sb, discard_block, next - start);
 		 /* Send the TRIM command down to the device */
 		err = sb_issue_discard(sb, discard_block, next - start,
 				       GFP_NOFS, 0);
@@ -2100,7 +2108,7 @@ int ext3_trim_fs(struct super_block *sb, struct fstrim_range *range)
 	if (unlikely(minlen > EXT3_BLOCKS_PER_GROUP(sb)))
 		return -EINVAL;
 	if (start >= max_blks)
-		goto out;
+		return -EINVAL;
 	if (start + len > max_blks)
 		len = max_blks - start;
 
@@ -2148,8 +2156,6 @@ int ext3_trim_fs(struct super_block *sb, struct fstrim_range *range)
 
 	if (ret >= 0)
 		ret = 0;
-
-out:
 	range->len = trimmed * sb->s_blocksize;
 
 	return ret;

fs/ext3/file.c

@@ -71,7 +71,6 @@ const struct file_operations ext3_file_operations = {
 };
 
 const struct inode_operations ext3_file_inode_operations = {
-	.truncate	= ext3_truncate,
 	.setattr	= ext3_setattr,
 #ifdef CONFIG_EXT3_FS_XATTR
 	.setxattr	= generic_setxattr,

fs/ext3/fsync.c

@@ -30,6 +30,7 @@
 #include <linux/jbd.h>
 #include <linux/ext3_fs.h>
 #include <linux/ext3_jbd.h>
+#include <trace/events/ext3.h>
 
 /*
  * akpm: A new design for ext3_sync_file().
@@ -51,12 +52,14 @@ int ext3_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 	int ret, needs_barrier = 0;
 	tid_t commit_tid;
 
+	trace_ext3_sync_file_enter(file, datasync);
+
 	if (inode->i_sb->s_flags & MS_RDONLY)
 		return 0;
 
 	ret = filemap_write_and_wait_range(inode->i_mapping, start, end);
 	if (ret)
-		return ret;
+		goto out;
 
 	/*
 	 * Taking the mutex here just to keep consistent with how fsync was
@@ -83,7 +86,8 @@ int ext3_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 	 */
 	if (ext3_should_journal_data(inode)) {
 		mutex_unlock(&inode->i_mutex);
-		return ext3_force_commit(inode->i_sb);
+		ret = ext3_force_commit(inode->i_sb);
+		goto out;
 	}
 
 	if (datasync)
@@ -104,6 +108,9 @@ int ext3_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 	 */
 	if (needs_barrier)
 		blkdev_issue_flush(inode->i_sb->s_bdev, GFP_KERNEL, NULL);
+
 	mutex_unlock(&inode->i_mutex);
+out:
+	trace_ext3_sync_file_exit(inode, ret);
 	return ret;
 }

fs/ext3/ialloc.c

@@ -23,6 +23,7 @@
 #include <linux/buffer_head.h>
 #include <linux/random.h>
 #include <linux/bitops.h>
+#include <trace/events/ext3.h>
 
 #include <asm/byteorder.h>
 
@@ -118,6 +119,7 @@ void ext3_free_inode (handle_t *handle, struct inode * inode)
 
 	ino = inode->i_ino;
 	ext3_debug ("freeing inode %lu\n", ino);
+	trace_ext3_free_inode(inode);
 
 	is_directory = S_ISDIR(inode->i_mode);
 
@@ -426,6 +428,7 @@ struct inode *ext3_new_inode(handle_t *handle, struct inode * dir,
 		return ERR_PTR(-EPERM);
 
 	sb = dir->i_sb;
+	trace_ext3_request_inode(dir, mode);
 	inode = new_inode(sb);
 	if (!inode)
 		return ERR_PTR(-ENOMEM);
@@ -601,6 +604,7 @@ struct inode *ext3_new_inode(handle_t *handle, struct inode * dir,
 	}
 
 	ext3_debug("allocating inode %lu\n", inode->i_ino);
+	trace_ext3_allocate_inode(inode, dir, mode);
 	goto really_out;
 fail:
 	ext3_std_error(sb, err);

fs/ext3/ioctl.c

@@ -285,7 +285,7 @@ long ext3_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 		if (!capable(CAP_SYS_ADMIN))
 			return -EPERM;
 
-		if (copy_from_user(&range, (struct fstrim_range *)arg,
+		if (copy_from_user(&range, (struct fstrim_range __user *)arg,
 				   sizeof(range)))
 			return -EFAULT;
 
@@ -293,7 +293,7 @@ long ext3_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 		if (ret < 0)
 			return ret;
 
-		if (copy_to_user((struct fstrim_range *)arg, &range,
+		if (copy_to_user((struct fstrim_range __user *)arg, &range,
 				 sizeof(range)))
 			return -EFAULT;
 

fs/ext3/namei.c

@@ -36,6 +36,7 @@
 #include <linux/quotaops.h>
 #include <linux/buffer_head.h>
 #include <linux/bio.h>
+#include <trace/events/ext3.h>
 
 #include "namei.h"
 #include "xattr.h"
@@ -287,7 +288,7 @@ static struct stats dx_show_leaf(struct dx_hash_info *hinfo, struct ext3_dir_ent
 				while (len--) printk("%c", *name++);
 				ext3fs_dirhash(de->name, de->name_len, &h);
 				printk(":%x.%u ", h.hash,
-				       ((char *) de - base));
+				       (unsigned) ((char *) de - base));
 			}
 			space += EXT3_DIR_REC_LEN(de->name_len);
 			names++;
@@ -1013,7 +1014,7 @@ static struct buffer_head * ext3_dx_find_entry(struct inode *dir,
 
 	*err = -ENOENT;
 errout:
-	dxtrace(printk("%s not found\n", name));
+	dxtrace(printk("%s not found\n", entry->name));
 	dx_release (frames);
 	return NULL;
 }
@@ -2140,6 +2141,7 @@ static int ext3_unlink(struct inode * dir, struct dentry *dentry)
 	struct ext3_dir_entry_2 * de;
 	handle_t *handle;
 
+	trace_ext3_unlink_enter(dir, dentry);
 	/* Initialize quotas before so that eventual writes go
 	 * in separate transaction */
 	dquot_initialize(dir);
@@ -2185,6 +2187,7 @@ static int ext3_unlink(struct inode * dir, struct dentry *dentry)
 end_unlink:
 	ext3_journal_stop(handle);
 	brelse (bh);
+	trace_ext3_unlink_exit(dentry, retval);
 	return retval;
 }
 

fs/ext3/super.c

@@ -44,6 +44,9 @@
 #include "acl.h"
 #include "namei.h"
 
+#define CREATE_TRACE_POINTS
+#include <trace/events/ext3.h>
+
 #ifdef CONFIG_EXT3_DEFAULTS_TO_ORDERED
   #define EXT3_MOUNT_DEFAULT_DATA_MODE EXT3_MOUNT_ORDERED_DATA
 #else
@@ -497,6 +500,14 @@ static struct inode *ext3_alloc_inode(struct super_block *sb)
 	return &ei->vfs_inode;
 }
 
+static int ext3_drop_inode(struct inode *inode)
+{
+	int drop = generic_drop_inode(inode);
+
+	trace_ext3_drop_inode(inode, drop);
+	return drop;
+}
+
 static void ext3_i_callback(struct rcu_head *head)
 {
 	struct inode *inode = container_of(head, struct inode, i_rcu);
@@ -788,6 +799,7 @@ static const struct super_operations ext3_sops = {
 	.destroy_inode	= ext3_destroy_inode,
 	.write_inode	= ext3_write_inode,
 	.dirty_inode	= ext3_dirty_inode,
+	.drop_inode	= ext3_drop_inode,
 	.evict_inode	= ext3_evict_inode,
 	.put_super	= ext3_put_super,
 	.sync_fs	= ext3_sync_fs,
@@ -2509,6 +2521,7 @@ static int ext3_sync_fs(struct super_block *sb, int wait)
 {
 	tid_t target;
 
+	trace_ext3_sync_fs(sb, wait);
 	if (journal_start_commit(EXT3_SB(sb)->s_journal, &target)) {
 		if (wait)
 			log_wait_commit(EXT3_SB(sb)->s_journal, target);

fs/ext3/xattr.c

@@ -803,8 +803,16 @@ ext3_xattr_block_set(handle_t *handle, struct inode *inode,
 			/* We need to allocate a new block */
 			ext3_fsblk_t goal = ext3_group_first_block_no(sb,
 						EXT3_I(inode)->i_block_group);
-			ext3_fsblk_t block = ext3_new_block(handle, inode,
-							goal, &error);
+			ext3_fsblk_t block;
+
+			/*
+			 * Protect us agaist concurrent allocations to the
+			 * same inode from ext3_..._writepage(). Reservation
+			 * code does not expect racing allocations.
+			 */
+			mutex_lock(&EXT3_I(inode)->truncate_mutex);
+			block = ext3_new_block(handle, inode, goal, &error);
+			mutex_unlock(&EXT3_I(inode)->truncate_mutex);
 			if (error)
 				goto cleanup;
 			ea_idebug(inode, "creating block %d", block);

fs/jbd/checkpoint.c

@@ -22,6 +22,8 @@
 #include <linux/jbd.h>
 #include <linux/errno.h>
 #include <linux/slab.h>
+#include <linux/blkdev.h>
+#include <trace/events/jbd.h>
 
 /*
  * Unlink a buffer from a transaction checkpoint list.
@@ -95,10 +97,14 @@ static int __try_to_free_cp_buf(struct journal_head *jh)
 
 	if (jh->b_jlist == BJ_None && !buffer_locked(bh) &&
 	    !buffer_dirty(bh) && !buffer_write_io_error(bh)) {
+		/*
+		 * Get our reference so that bh cannot be freed before
+		 * we unlock it
+		 */
+		get_bh(bh);
 		JBUFFER_TRACE(jh, "remove from checkpoint list");
 		ret = __journal_remove_checkpoint(jh) + 1;
 		jbd_unlock_bh_state(bh);
-		journal_remove_journal_head(bh);
 		BUFFER_TRACE(bh, "release");
 		__brelse(bh);
 	} else {
@@ -220,8 +226,8 @@ static int __wait_cp_io(journal_t *journal, transaction_t *transaction)
 			spin_lock(&journal->j_list_lock);
 			goto restart;
 		}
+		get_bh(bh);
 		if (buffer_locked(bh)) {
-			get_bh(bh);
 			spin_unlock(&journal->j_list_lock);
 			jbd_unlock_bh_state(bh);
 			wait_on_buffer(bh);
@@ -240,7 +246,6 @@ static int __wait_cp_io(journal_t *journal, transaction_t *transaction)
 		 */
 		released = __journal_remove_checkpoint(jh);
 		jbd_unlock_bh_state(bh);
-		journal_remove_journal_head(bh);
 		__brelse(bh);
 	}
 
@@ -253,9 +258,12 @@ static void
 __flush_batch(journal_t *journal, struct buffer_head **bhs, int *batch_count)
 {
 	int i;
+	struct blk_plug plug;
 
+	blk_start_plug(&plug);
 	for (i = 0; i < *batch_count; i++)
-		write_dirty_buffer(bhs[i], WRITE);
+		write_dirty_buffer(bhs[i], WRITE_SYNC);
+	blk_finish_plug(&plug);
 
 	for (i = 0; i < *batch_count; i++) {
 		struct buffer_head *bh = bhs[i];
@@ -304,12 +312,12 @@ static int __process_buffer(journal_t *journal, struct journal_head *jh,
 		ret = 1;
 		if (unlikely(buffer_write_io_error(bh)))
 			ret = -EIO;
+		get_bh(bh);
 		J_ASSERT_JH(jh, !buffer_jbddirty(bh));
 		BUFFER_TRACE(bh, "remove from checkpoint");
 		__journal_remove_checkpoint(jh);
 		spin_unlock(&journal->j_list_lock);
 		jbd_unlock_bh_state(bh);
-		journal_remove_journal_head(bh);
 		__brelse(bh);
 	} else {
 		/*
@@ -358,6 +366,7 @@ int log_do_checkpoint(journal_t *journal)
 	 * journal straight away.
 	 */
 	result = cleanup_journal_tail(journal);
+	trace_jbd_checkpoint(journal, result);
 	jbd_debug(1, "cleanup_journal_tail returned %d\n", result);
 	if (result <= 0)
 		return result;
@@ -503,6 +512,7 @@ int cleanup_journal_tail(journal_t *journal)
 	if (blocknr < journal->j_tail)
 		freed = freed + journal->j_last - journal->j_first;
 
+	trace_jbd_cleanup_journal_tail(journal, first_tid, blocknr, freed);
 	jbd_debug(1,
 		  "Cleaning journal tail from %d to %d (offset %u), "
 		  "freeing %u\n",
@@ -523,9 +533,9 @@ int cleanup_journal_tail(journal_t *journal)
 /*
  * journal_clean_one_cp_list
  *
- * Find all the written-back checkpoint buffers in the given list and release them.
+ * Find all the written-back checkpoint buffers in the given list and release
+ * them.
  *
- * Called with the journal locked.
  * Called with j_list_lock held.
  * Returns number of bufers reaped (for debug)
  */
@@ -632,8 +642,8 @@ int __journal_clean_checkpoint_list(journal_t *journal)
  * checkpoint lists.
  *
  * The function returns 1 if it frees the transaction, 0 otherwise.
+ * The function can free jh and bh.
  *
- * This function is called with the journal locked.
  * This function is called with j_list_lock held.
  * This function is called with jbd_lock_bh_state(jh2bh(jh))
  */
@@ -652,13 +662,14 @@ int __journal_remove_checkpoint(struct journal_head *jh)
 	}
 	journal = transaction->t_journal;
 
+	JBUFFER_TRACE(jh, "removing from transaction");
 	__buffer_unlink(jh);
 	jh->b_cp_transaction = NULL;
+	journal_put_journal_head(jh);
 
 	if (transaction->t_checkpoint_list != NULL ||
 	    transaction->t_checkpoint_io_list != NULL)
 		goto out;
-	JBUFFER_TRACE(jh, "transaction has no more buffers");
 
 	/*
 	 * There is one special case to worry about: if we have just pulled the
@@ -669,10 +680,8 @@ int __journal_remove_checkpoint(struct journal_head *jh)
 	 * The locking here around t_state is a bit sleazy.
 	 * See the comment at the end of journal_commit_transaction().
 	 */
-	if (transaction->t_state != T_FINISHED) {
-		JBUFFER_TRACE(jh, "belongs to running/committing transaction");
+	if (transaction->t_state != T_FINISHED)
 		goto out;
-	}
 
 	/* OK, that was the last buffer for the transaction: we can now
 	   safely remove this transaction from the log */
@@ -684,7 +693,6 @@ int __journal_remove_checkpoint(struct journal_head *jh)
 	wake_up(&journal->j_wait_logspace);
 	ret = 1;
 out:
-	JBUFFER_TRACE(jh, "exit");
 	return ret;
 }
 
@@ -703,6 +711,8 @@ void __journal_insert_checkpoint(struct journal_head *jh,
 	J_ASSERT_JH(jh, buffer_dirty(jh2bh(jh)) || buffer_jbddirty(jh2bh(jh)));
 	J_ASSERT_JH(jh, jh->b_cp_transaction == NULL);
 
+	/* Get reference for checkpointing transaction */
+	journal_grab_journal_head(jh2bh(jh));
 	jh->b_cp_transaction = transaction;
 
 	if (!transaction->t_checkpoint_list) {
@@ -752,6 +762,7 @@ void __journal_drop_transaction(journal_t *journal, transaction_t *transaction)
 	J_ASSERT(journal->j_committing_transaction != transaction);
 	J_ASSERT(journal->j_running_transaction != transaction);
 
+	trace_jbd_drop_transaction(journal, transaction);
 	jbd_debug(1, "Dropping transaction %d, all done\n", transaction->t_tid);
 	kfree(transaction);
 }

fs/jbd/commit.c

@@ -21,6 +21,7 @@
 #include <linux/pagemap.h>
 #include <linux/bio.h>
 #include <linux/blkdev.h>
+#include <trace/events/jbd.h>
 
 /*
  * Default IO end handler for temporary BJ_IO buffer_heads.
@@ -204,6 +205,8 @@ static int journal_submit_data_buffers(journal_t *journal,
 			if (!trylock_buffer(bh)) {
 				BUFFER_TRACE(bh, "needs blocking lock");
 				spin_unlock(&journal->j_list_lock);
+				trace_jbd_do_submit_data(journal,
+						     commit_transaction);
 				/* Write out all data to prevent deadlocks */
 				journal_do_submit_data(wbuf, bufs, write_op);
 				bufs = 0;
@@ -236,6 +239,8 @@ static int journal_submit_data_buffers(journal_t *journal,
 			jbd_unlock_bh_state(bh);
 			if (bufs == journal->j_wbufsize) {
 				spin_unlock(&journal->j_list_lock);
+				trace_jbd_do_submit_data(journal,
+						     commit_transaction);
 				journal_do_submit_data(wbuf, bufs, write_op);
 				bufs = 0;
 				goto write_out_data;
@@ -253,10 +258,6 @@ static int journal_submit_data_buffers(journal_t *journal,
 			jbd_unlock_bh_state(bh);
 			if (locked)
 				unlock_buffer(bh);
-			journal_remove_journal_head(bh);
-			/* One for our safety reference, other for
-			 * journal_remove_journal_head() */
-			put_bh(bh);
 			release_data_buffer(bh);
 		}
 
@@ -266,6 +267,7 @@ static int journal_submit_data_buffers(journal_t *journal,
 		}
 	}
 	spin_unlock(&journal->j_list_lock);
+	trace_jbd_do_submit_data(journal, commit_transaction);
 	journal_do_submit_data(wbuf, bufs, write_op);
 
 	return err;
@@ -316,12 +318,14 @@ void journal_commit_transaction(journal_t *journal)
 	commit_transaction = journal->j_running_transaction;
 	J_ASSERT(commit_transaction->t_state == T_RUNNING);
 
+	trace_jbd_start_commit(journal, commit_transaction);
 	jbd_debug(1, "JBD: starting commit of transaction %d\n",
 			commit_transaction->t_tid);
 
 	spin_lock(&journal->j_state_lock);
 	commit_transaction->t_state = T_LOCKED;
 
+	trace_jbd_commit_locking(journal, commit_transaction);
 	spin_lock(&commit_transaction->t_handle_lock);
 	while (commit_transaction->t_updates) {
 		DEFINE_WAIT(wait);
@@ -392,6 +396,7 @@ void journal_commit_transaction(journal_t *journal)
 	 */
 	journal_switch_revoke_table(journal);
 
+	trace_jbd_commit_flushing(journal, commit_transaction);
 	commit_transaction->t_state = T_FLUSH;
 	journal->j_committing_transaction = commit_transaction;
 	journal->j_running_transaction = NULL;
@@ -446,14 +451,9 @@ void journal_commit_transaction(journal_t *journal)
 		}
 		if (buffer_jbd(bh) && bh2jh(bh) == jh &&
 		    jh->b_transaction == commit_transaction &&
-		    jh->b_jlist == BJ_Locked) {
+		    jh->b_jlist == BJ_Locked)
 			__journal_unfile_buffer(jh);
-			jbd_unlock_bh_state(bh);
-			journal_remove_journal_head(bh);
-			put_bh(bh);
-		} else {
-			jbd_unlock_bh_state(bh);
-		}
+		jbd_unlock_bh_state(bh);
 		release_data_buffer(bh);
 		cond_resched_lock(&journal->j_list_lock);
 	}
@@ -493,6 +493,7 @@ void journal_commit_transaction(journal_t *journal)
 	commit_transaction->t_state = T_COMMIT;
 	spin_unlock(&journal->j_state_lock);
 
+	trace_jbd_commit_logging(journal, commit_transaction);
 	J_ASSERT(commit_transaction->t_nr_buffers <=
 		 commit_transaction->t_outstanding_credits);
 
@@ -797,10 +798,16 @@ void journal_commit_transaction(journal_t *journal)
 	while (commit_transaction->t_forget) {
 		transaction_t *cp_transaction;
 		struct buffer_head *bh;
+		int try_to_free = 0;
 
 		jh = commit_transaction->t_forget;
 		spin_unlock(&journal->j_list_lock);
 		bh = jh2bh(jh);
+		/*
+		 * Get a reference so that bh cannot be freed before we are
+		 * done with it.
+		 */
+		get_bh(bh);
 		jbd_lock_bh_state(bh);
 		J_ASSERT_JH(jh,	jh->b_transaction == commit_transaction ||
 			jh->b_transaction == journal->j_running_transaction);
@@ -858,28 +865,27 @@ void journal_commit_transaction(journal_t *journal)
 			__journal_insert_checkpoint(jh, commit_transaction);
 			if (is_journal_aborted(journal))
 				clear_buffer_jbddirty(bh);
-			JBUFFER_TRACE(jh, "refile for checkpoint writeback");
-			__journal_refile_buffer(jh);
-			jbd_unlock_bh_state(bh);
 		} else {
 			J_ASSERT_BH(bh, !buffer_dirty(bh));
-			/* The buffer on BJ_Forget list and not jbddirty means
+			/*
+			 * The buffer on BJ_Forget list and not jbddirty means
 			 * it has been freed by this transaction and hence it
 			 * could not have been reallocated until this
 			 * transaction has committed. *BUT* it could be
 			 * reallocated once we have written all the data to
 			 * disk and before we process the buffer on BJ_Forget
-			 * list. */
-			JBUFFER_TRACE(jh, "refile or unfile freed buffer");
-			__journal_refile_buffer(jh);
-			if (!jh->b_transaction) {
-				jbd_unlock_bh_state(bh);
-				 /* needs a brelse */
-				journal_remove_journal_head(bh);
-				release_buffer_page(bh);
-			} else
-				jbd_unlock_bh_state(bh);
+			 * list.
+			 */
+			if (!jh->b_next_transaction)
+				try_to_free = 1;
 		}
+		JBUFFER_TRACE(jh, "refile or unfile freed buffer");
+		__journal_refile_buffer(jh);
+		jbd_unlock_bh_state(bh);
+		if (try_to_free)
+			release_buffer_page(bh);
+		else
+			__brelse(bh);
 		cond_resched_lock(&journal->j_list_lock);
 	}
 	spin_unlock(&journal->j_list_lock);
@@ -946,6 +952,7 @@ void journal_commit_transaction(journal_t *journal)
 	}
 	spin_unlock(&journal->j_list_lock);
 
+	trace_jbd_end_commit(journal, commit_transaction);
 	jbd_debug(1, "JBD: commit %d complete, head %d\n",
 		  journal->j_commit_sequence, journal->j_tail_sequence);
 

fs/jbd/journal.c

@@ -38,6 +38,9 @@
 #include <linux/debugfs.h>
 #include <linux/ratelimit.h>
 
+#define CREATE_TRACE_POINTS
+#include <trace/events/jbd.h>
+
 #include <asm/uaccess.h>
 #include <asm/page.h>
 
@@ -1065,6 +1068,7 @@ void journal_update_superblock(journal_t *journal, int wait)
 	} else
 		write_dirty_buffer(bh, WRITE);
 
+	trace_jbd_update_superblock_end(journal, wait);
 out:
 	/* If we have just flushed the log (by marking s_start==0), then
 	 * any future commit will have to be careful to update the
@@ -1799,10 +1803,9 @@ static void journal_free_journal_head(struct journal_head *jh)
  * When a buffer has its BH_JBD bit set it is immune from being released by
  * core kernel code, mainly via ->b_count.
  *
- * A journal_head may be detached from its buffer_head when the journal_head's
- * b_transaction, b_cp_transaction and b_next_transaction pointers are NULL.
- * Various places in JBD call journal_remove_journal_head() to indicate that the
- * journal_head can be dropped if needed.
+ * A journal_head is detached from its buffer_head when the journal_head's
+ * b_jcount reaches zero. Running transaction (b_transaction) and checkpoint
+ * transaction (b_cp_transaction) hold their references to b_jcount.
  *
  * Various places in the kernel want to attach a journal_head to a buffer_head
  * _before_ attaching the journal_head to a transaction.  To protect the
@@ -1815,17 +1818,16 @@ static void journal_free_journal_head(struct journal_head *jh)
  *	(Attach a journal_head if needed.  Increments b_jcount)
  *	struct journal_head *jh = journal_add_journal_head(bh);
  *	...
- *	jh->b_transaction = xxx;
- *	journal_put_journal_head(jh);
- *
- * Now, the journal_head's b_jcount is zero, but it is safe from being released
- * because it has a non-zero b_transaction.
+ *      (Get another reference for transaction)
+ *      journal_grab_journal_head(bh);
+ *      jh->b_transaction = xxx;
+ *      (Put original reference)
+ *      journal_put_journal_head(jh);
  */
 
 /*
  * Give a buffer_head a journal_head.
  *
- * Doesn't need the journal lock.
  * May sleep.
  */
 struct journal_head *journal_add_journal_head(struct buffer_head *bh)
@@ -1889,61 +1891,29 @@ static void __journal_remove_journal_head(struct buffer_head *bh)
 	struct journal_head *jh = bh2jh(bh);
 
 	J_ASSERT_JH(jh, jh->b_jcount >= 0);
-
-	get_bh(bh);
-	if (jh->b_jcount == 0) {
-		if (jh->b_transaction == NULL &&
-				jh->b_next_transaction == NULL &&
-				jh->b_cp_transaction == NULL) {
-			J_ASSERT_JH(jh, jh->b_jlist == BJ_None);
-			J_ASSERT_BH(bh, buffer_jbd(bh));
-			J_ASSERT_BH(bh, jh2bh(jh) == bh);
-			BUFFER_TRACE(bh, "remove journal_head");
-			if (jh->b_frozen_data) {
-				printk(KERN_WARNING "%s: freeing "
-						"b_frozen_data\n",
-						__func__);
-				jbd_free(jh->b_frozen_data, bh->b_size);
-			}
-			if (jh->b_committed_data) {
-				printk(KERN_WARNING "%s: freeing "
-						"b_committed_data\n",
-						__func__);
-				jbd_free(jh->b_committed_data, bh->b_size);
-			}
-			bh->b_private = NULL;
-			jh->b_bh = NULL;	/* debug, really */
-			clear_buffer_jbd(bh);
-			__brelse(bh);
-			journal_free_journal_head(jh);
-		} else {
-			BUFFER_TRACE(bh, "journal_head was locked");
-		}
+	J_ASSERT_JH(jh, jh->b_transaction == NULL);
+	J_ASSERT_JH(jh, jh->b_next_transaction == NULL);
+	J_ASSERT_JH(jh, jh->b_cp_transaction == NULL);
+	J_ASSERT_JH(jh, jh->b_jlist == BJ_None);
+	J_ASSERT_BH(bh, buffer_jbd(bh));
+	J_ASSERT_BH(bh, jh2bh(jh) == bh);
+	BUFFER_TRACE(bh, "remove journal_head");
+	if (jh->b_frozen_data) {
+		printk(KERN_WARNING "%s: freeing b_frozen_data\n", __func__);
+		jbd_free(jh->b_frozen_data, bh->b_size);
 	}
+	if (jh->b_committed_data) {
+		printk(KERN_WARNING "%s: freeing b_committed_data\n", __func__);
+		jbd_free(jh->b_committed_data, bh->b_size);
+	}
+	bh->b_private = NULL;
+	jh->b_bh = NULL;	/* debug, really */
+	clear_buffer_jbd(bh);
+	journal_free_journal_head(jh);
 }
 
 /*
- * journal_remove_journal_head(): if the buffer isn't attached to a transaction
- * and has a zero b_jcount then remove and release its journal_head.   If we did
- * see that the buffer is not used by any transaction we also "logically"
- * decrement ->b_count.
- *
- * We in fact take an additional increment on ->b_count as a convenience,
- * because the caller usually wants to do additional things with the bh
- * after calling here.
- * The caller of journal_remove_journal_head() *must* run __brelse(bh) at some
- * time.  Once the caller has run __brelse(), the buffer is eligible for
- * reaping by try_to_free_buffers().
- */
-void journal_remove_journal_head(struct buffer_head *bh)
-{
-	jbd_lock_bh_journal_head(bh);
-	__journal_remove_journal_head(bh);
-	jbd_unlock_bh_journal_head(bh);
-}
-
-/*
- * Drop a reference on the passed journal_head.  If it fell to zero then try to
+ * Drop a reference on the passed journal_head.  If it fell to zero then
  * release the journal_head from the buffer_head.
  */
 void journal_put_journal_head(struct journal_head *jh)
@@ -1953,11 +1923,12 @@ void journal_put_journal_head(struct journal_head *jh)
 	jbd_lock_bh_journal_head(bh);
 	J_ASSERT_JH(jh, jh->b_jcount > 0);
 	--jh->b_jcount;
-	if (!jh->b_jcount && !jh->b_transaction) {
+	if (!jh->b_jcount) {
 		__journal_remove_journal_head(bh);
+		jbd_unlock_bh_journal_head(bh);
 		__brelse(bh);
-	}
-	jbd_unlock_bh_journal_head(bh);
+	} else
+		jbd_unlock_bh_journal_head(bh);
 }
 
 /*

fs/jbd/transaction.c

@@ -26,6 +26,7 @@
 #include <linux/mm.h>
 #include <linux/highmem.h>
 #include <linux/hrtimer.h>
+#include <linux/backing-dev.h>
 
 static void __journal_temp_unlink_buffer(struct journal_head *jh);
 
@@ -99,11 +100,10 @@ static int start_this_handle(journal_t *journal, handle_t *handle)
 
 alloc_transaction:
 	if (!journal->j_running_transaction) {
-		new_transaction = kzalloc(sizeof(*new_transaction),
-						GFP_NOFS|__GFP_NOFAIL);
+		new_transaction = kzalloc(sizeof(*new_transaction), GFP_NOFS);
 		if (!new_transaction) {
-			ret = -ENOMEM;
-			goto out;
+			congestion_wait(BLK_RW_ASYNC, HZ/50);
+			goto alloc_transaction;
 		}
 	}
 
@@ -696,7 +696,6 @@ do_get_write_access(handle_t *handle, struct journal_head *jh,
 	if (!jh->b_transaction) {
 		JBUFFER_TRACE(jh, "no transaction");
 		J_ASSERT_JH(jh, !jh->b_next_transaction);
-		jh->b_transaction = transaction;
 		JBUFFER_TRACE(jh, "file as BJ_Reserved");
 		spin_lock(&journal->j_list_lock);
 		__journal_file_buffer(jh, transaction, BJ_Reserved);
@@ -818,7 +817,6 @@ int journal_get_create_access(handle_t *handle, struct buffer_head *bh)
 		 * committed and so it's safe to clear the dirty bit.
 		 */
 		clear_buffer_dirty(jh2bh(jh));
-		jh->b_transaction = transaction;
 
 		/* first access by this transaction */
 		jh->b_modified = 0;
@@ -844,8 +842,8 @@ int journal_get_create_access(handle_t *handle, struct buffer_head *bh)
 	 */
 	JBUFFER_TRACE(jh, "cancelling revoke");
 	journal_cancel_revoke(handle, jh);
-	journal_put_journal_head(jh);
 out:
+	journal_put_journal_head(jh);
 	return err;
 }
 
@@ -1069,8 +1067,9 @@ int journal_dirty_data(handle_t *handle, struct buffer_head *bh)
 				ret = -EIO;
 				goto no_journal;
 			}
-
-			if (jh->b_transaction != NULL) {
+			/* We might have slept so buffer could be refiled now */
+			if (jh->b_transaction != NULL &&
+			    jh->b_transaction != handle->h_transaction) {
 				JBUFFER_TRACE(jh, "unfile from commit");
 				__journal_temp_unlink_buffer(jh);
 				/* It still points to the committing
@@ -1091,8 +1090,6 @@ int journal_dirty_data(handle_t *handle, struct buffer_head *bh)
 		if (jh->b_jlist != BJ_SyncData && jh->b_jlist != BJ_Locked) {
 			JBUFFER_TRACE(jh, "not on correct data list: unfile");
 			J_ASSERT_JH(jh, jh->b_jlist != BJ_Shadow);
-			__journal_temp_unlink_buffer(jh);
-			jh->b_transaction = handle->h_transaction;
 			JBUFFER_TRACE(jh, "file as data");
 			__journal_file_buffer(jh, handle->h_transaction,
 						BJ_SyncData);
@@ -1300,8 +1297,6 @@ int journal_forget (handle_t *handle, struct buffer_head *bh)
 			__journal_file_buffer(jh, transaction, BJ_Forget);
 		} else {
 			__journal_unfile_buffer(jh);
-			journal_remove_journal_head(bh);
-			__brelse(bh);
 			if (!buffer_jbd(bh)) {
 				spin_unlock(&journal->j_list_lock);
 				jbd_unlock_bh_state(bh);
@@ -1622,19 +1617,32 @@ static void __journal_temp_unlink_buffer(struct journal_head *jh)
 		mark_buffer_dirty(bh);	/* Expose it to the VM */
 }
 
+/*
+ * Remove buffer from all transactions.
+ *
+ * Called with bh_state lock and j_list_lock
+ *
+ * jh and bh may be already freed when this function returns.
+ */
 void __journal_unfile_buffer(struct journal_head *jh)
 {
 	__journal_temp_unlink_buffer(jh);
 	jh->b_transaction = NULL;
+	journal_put_journal_head(jh);
 }
 
 void journal_unfile_buffer(journal_t *journal, struct journal_head *jh)
 {
-	jbd_lock_bh_state(jh2bh(jh));
+	struct buffer_head *bh = jh2bh(jh);
+
+	/* Get reference so that buffer cannot be freed before we unlock it */
+	get_bh(bh);
+	jbd_lock_bh_state(bh);
 	spin_lock(&journal->j_list_lock);
 	__journal_unfile_buffer(jh);
 	spin_unlock(&journal->j_list_lock);
-	jbd_unlock_bh_state(jh2bh(jh));
+	jbd_unlock_bh_state(bh);
+	__brelse(bh);
 }
 
 /*
@@ -1661,16 +1669,12 @@ __journal_try_to_free_buffer(journal_t *journal, struct buffer_head *bh)
 			/* A written-back ordered data buffer */
 			JBUFFER_TRACE(jh, "release data");
 			__journal_unfile_buffer(jh);
-			journal_remove_journal_head(bh);
-			__brelse(bh);
 		}
 	} else if (jh->b_cp_transaction != NULL && jh->b_transaction == NULL) {
 		/* written-back checkpointed metadata buffer */
 		if (jh->b_jlist == BJ_None) {
 			JBUFFER_TRACE(jh, "remove from checkpoint list");
 			__journal_remove_checkpoint(jh);
-			journal_remove_journal_head(bh);
-			__brelse(bh);
 		}
 	}
 	spin_unlock(&journal->j_list_lock);
@@ -1733,7 +1737,7 @@ int journal_try_to_free_buffers(journal_t *journal,
 		/*
 		 * We take our own ref against the journal_head here to avoid
 		 * having to add tons of locking around each instance of
-		 * journal_remove_journal_head() and journal_put_journal_head().
+		 * journal_put_journal_head().
 		 */
 		jh = journal_grab_journal_head(bh);
 		if (!jh)
@@ -1770,10 +1774,9 @@ static int __dispose_buffer(struct journal_head *jh, transaction_t *transaction)
 	int may_free = 1;
 	struct buffer_head *bh = jh2bh(jh);
 
-	__journal_unfile_buffer(jh);
-
 	if (jh->b_cp_transaction) {
 		JBUFFER_TRACE(jh, "on running+cp transaction");
+		__journal_temp_unlink_buffer(jh);
 		/*
 		 * We don't want to write the buffer anymore, clear the
 		 * bit so that we don't confuse checks in
@@ -1784,8 +1787,7 @@ static int __dispose_buffer(struct journal_head *jh, transaction_t *transaction)
 		may_free = 0;
 	} else {
 		JBUFFER_TRACE(jh, "on running transaction");
-		journal_remove_journal_head(bh);
-		__brelse(bh);
+		__journal_unfile_buffer(jh);
 	}
 	return may_free;
 }
@@ -2070,6 +2072,8 @@ void __journal_file_buffer(struct journal_head *jh,
 
 	if (jh->b_transaction)
 		__journal_temp_unlink_buffer(jh);
+	else
+		journal_grab_journal_head(bh);
 	jh->b_transaction = transaction;
 
 	switch (jlist) {
@@ -2127,9 +2131,10 @@ void journal_file_buffer(struct journal_head *jh,
  * already started to be used by a subsequent transaction, refile the
  * buffer on that transaction's metadata list.
  *
- * Called under journal->j_list_lock
- *
+ * Called under j_list_lock
  * Called under jbd_lock_bh_state(jh2bh(jh))
+ *
+ * jh and bh may be already free when this function returns
  */
 void __journal_refile_buffer(struct journal_head *jh)
 {
@@ -2153,6 +2158,11 @@ void __journal_refile_buffer(struct journal_head *jh)
 
 	was_dirty = test_clear_buffer_jbddirty(bh);
 	__journal_temp_unlink_buffer(jh);
+	/*
+	 * We set b_transaction here because b_next_transaction will inherit
+	 * our jh reference and thus __journal_file_buffer() must not take a
+	 * new one.
+	 */
 	jh->b_transaction = jh->b_next_transaction;
 	jh->b_next_transaction = NULL;
 	if (buffer_freed(bh))
@@ -2169,30 +2179,21 @@ void __journal_refile_buffer(struct journal_head *jh)
 }
 
 /*
- * For the unlocked version of this call, also make sure that any
- * hanging journal_head is cleaned up if necessary.
- *
- * __journal_refile_buffer is usually called as part of a single locked
- * operation on a buffer_head, in which the caller is probably going to
- * be hooking the journal_head onto other lists.  In that case it is up
- * to the caller to remove the journal_head if necessary.  For the
- * unlocked journal_refile_buffer call, the caller isn't going to be
- * doing anything else to the buffer so we need to do the cleanup
- * ourselves to avoid a jh leak.
- *
- * *** The journal_head may be freed by this call! ***
+ * __journal_refile_buffer() with necessary locking added. We take our bh
+ * reference so that we can safely unlock bh.
+ *
+ * The jh and bh may be freed by this call.
  */
 void journal_refile_buffer(journal_t *journal, struct journal_head *jh)
 {
 	struct buffer_head *bh = jh2bh(jh);
 
+	/* Get reference so that buffer cannot be freed before we unlock it */
+	get_bh(bh);
 	jbd_lock_bh_state(bh);
 	spin_lock(&journal->j_list_lock);
-
 	__journal_refile_buffer(jh);
 	jbd_unlock_bh_state(bh);
-	journal_remove_journal_head(bh);
-
 	spin_unlock(&journal->j_list_lock);
 	__brelse(bh);
 }

include/linux/ext2_fs.h

@@ -18,6 +18,7 @@
 
 #include <linux/types.h>
 #include <linux/magic.h>
+#include <linux/fs.h>
 
 /*
  * The second extended filesystem constants/structures

include/linux/ext3_fs.h

@@ -418,12 +418,11 @@ struct ext3_inode {
 #define EXT2_MOUNT_DATA_FLAGS		EXT3_MOUNT_DATA_FLAGS
 #endif
 
-#define ext3_set_bit			__test_and_set_bit_le
+#define ext3_set_bit			__set_bit_le
 #define ext3_set_bit_atomic		ext2_set_bit_atomic
-#define ext3_clear_bit			__test_and_clear_bit_le
+#define ext3_clear_bit			__clear_bit_le
 #define ext3_clear_bit_atomic		ext2_clear_bit_atomic
 #define ext3_test_bit			test_bit_le
-#define ext3_find_first_zero_bit	find_first_zero_bit_le
 #define ext3_find_next_zero_bit		find_next_zero_bit_le
 
 /*
@@ -913,7 +912,7 @@ extern void ext3_dirty_inode(struct inode *, int);
 extern int ext3_change_inode_journal_flag(struct inode *, int);
 extern int ext3_get_inode_loc(struct inode *, struct ext3_iloc *);
 extern int ext3_can_truncate(struct inode *inode);
-extern void ext3_truncate (struct inode *);
+extern void ext3_truncate(struct inode *inode);
 extern void ext3_set_inode_flags(struct inode *);
 extern void ext3_get_inode_flags(struct ext3_inode_info *);
 extern void ext3_set_aops(struct inode *inode);

include/linux/jbd.h

@@ -940,7 +940,6 @@ extern int	   journal_force_commit(journal_t *);
  */
 struct journal_head *journal_add_journal_head(struct buffer_head *bh);
 struct journal_head *journal_grab_journal_head(struct buffer_head *bh);
-void journal_remove_journal_head(struct buffer_head *bh);
 void journal_put_journal_head(struct journal_head *jh);
 
 /*

include/linux/journal-head.h

@@ -45,7 +45,7 @@ struct journal_head {
 	 * has been cowed
 	 * [jbd_lock_bh_state()]
 	 */
-	unsigned b_cow_tid;
+	tid_t b_cow_tid;
 
 	/*
 	 * Copy of the buffer data frozen for writing to the log.

include/linux/quota.h

@@ -415,13 +415,5 @@ struct quota_module_name {
 	{QFMT_VFS_V0, "quota_v2"},\
 	{0, NULL}}
 
-#else
-
-# /* nodep */ include <sys/cdefs.h>
-
-__BEGIN_DECLS
-long quotactl __P ((unsigned int, const char *, int, caddr_t));
-__END_DECLS
-
 #endif /* __KERNEL__ */
 #endif /* _QUOTA_ */

include/trace/events/jbd.h

+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM jbd
+
+#if !defined(_TRACE_JBD_H) || defined(TRACE_HEADER_MULTI_READ)
+#define _TRACE_JBD_H
+
+#include <linux/jbd.h>
+#include <linux/tracepoint.h>
+
+TRACE_EVENT(jbd_checkpoint,
+
+	TP_PROTO(journal_t *journal, int result),
+
+	TP_ARGS(journal, result),
+
+	TP_STRUCT__entry(
+		__field(	dev_t,	dev			)
+		__field(	int,	result			)
+	),
+
+	TP_fast_assign(
+		__entry->dev		= journal->j_fs_dev->bd_dev;
+		__entry->result		= result;
+	),
+
+	TP_printk("dev %d,%d result %d",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->result)
+);
+
+DECLARE_EVENT_CLASS(jbd_commit,
+
+	TP_PROTO(journal_t *journal, transaction_t *commit_transaction),
+
+	TP_ARGS(journal, commit_transaction),
+
+	TP_STRUCT__entry(
+		__field(	dev_t,	dev			)
+		__field(	char,	sync_commit		)
+		__field(	int,	transaction		)
+	),
+
+	TP_fast_assign(
+		__entry->dev		= journal->j_fs_dev->bd_dev;
+		__entry->sync_commit = commit_transaction->t_synchronous_commit;
+		__entry->transaction	= commit_transaction->t_tid;
+	),
+
+	TP_printk("dev %d,%d transaction %d sync %d",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->transaction, __entry->sync_commit)
+);
+
+DEFINE_EVENT(jbd_commit, jbd_start_commit,
+
+	TP_PROTO(journal_t *journal, transaction_t *commit_transaction),
+
+	TP_ARGS(journal, commit_transaction)
+);
+
+DEFINE_EVENT(jbd_commit, jbd_commit_locking,
+
+	TP_PROTO(journal_t *journal, transaction_t *commit_transaction),
+
+	TP_ARGS(journal, commit_transaction)
+);
+
+DEFINE_EVENT(jbd_commit, jbd_commit_flushing,
+
+	TP_PROTO(journal_t *journal, transaction_t *commit_transaction),
+
+	TP_ARGS(journal, commit_transaction)
+);
+
+DEFINE_EVENT(jbd_commit, jbd_commit_logging,
+
+	TP_PROTO(journal_t *journal, transaction_t *commit_transaction),
+
+	TP_ARGS(journal, commit_transaction)
+);
+
+TRACE_EVENT(jbd_drop_transaction,
+
+	TP_PROTO(journal_t *journal, transaction_t *commit_transaction),
+
+	TP_ARGS(journal, commit_transaction),
+
+	TP_STRUCT__entry(
+		__field(	dev_t,	dev			)
+		__field(	char,	sync_commit		)
+		__field(	int,	transaction		)
+	),
+
+	TP_fast_assign(
+		__entry->dev		= journal->j_fs_dev->bd_dev;
+		__entry->sync_commit = commit_transaction->t_synchronous_commit;
+		__entry->transaction	= commit_transaction->t_tid;
+	),
+
+	TP_printk("dev %d,%d transaction %d sync %d",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->transaction, __entry->sync_commit)
+);
+
+TRACE_EVENT(jbd_end_commit,
+	TP_PROTO(journal_t *journal, transaction_t *commit_transaction),
+
+	TP_ARGS(journal, commit_transaction),
+
+	TP_STRUCT__entry(
+		__field(	dev_t,	dev			)
+		__field(	char,	sync_commit		)
+		__field(	int,	transaction		)
+		__field(	int,	head			)
+	),
+
+	TP_fast_assign(
+		__entry->dev		= journal->j_fs_dev->bd_dev;
+		__entry->sync_commit = commit_transaction->t_synchronous_commit;
+		__entry->transaction	= commit_transaction->t_tid;
+		__entry->head		= journal->j_tail_sequence;
+	),
+
+	TP_printk("dev %d,%d transaction %d sync %d head %d",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->transaction, __entry->sync_commit, __entry->head)
+);
+
+TRACE_EVENT(jbd_do_submit_data,
+	TP_PROTO(journal_t *journal, transaction_t *commit_transaction),
+
+	TP_ARGS(journal, commit_transaction),
+
+	TP_STRUCT__entry(
+		__field(	dev_t,	dev			)
+		__field(	char,	sync_commit		)
+		__field(	int,	transaction		)
+	),
+
+	TP_fast_assign(
+		__entry->dev		= journal->j_fs_dev->bd_dev;
+		__entry->sync_commit = commit_transaction->t_synchronous_commit;
+		__entry->transaction	= commit_transaction->t_tid;
+	),
+
+	TP_printk("dev %d,%d transaction %d sync %d",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		   __entry->transaction, __entry->sync_commit)
+);
+
+TRACE_EVENT(jbd_cleanup_journal_tail,
+
+	TP_PROTO(journal_t *journal, tid_t first_tid,
+		 unsigned long block_nr, unsigned long freed),
+
+	TP_ARGS(journal, first_tid, block_nr, freed),
+
+	TP_STRUCT__entry(
+		__field(	dev_t,	dev			)
+		__field(	tid_t,	tail_sequence		)
+		__field(	tid_t,	first_tid		)
+		__field(unsigned long,	block_nr		)
+		__field(unsigned long,	freed			)
+	),
+
+	TP_fast_assign(
+		__entry->dev		= journal->j_fs_dev->bd_dev;
+		__entry->tail_sequence	= journal->j_tail_sequence;
+		__entry->first_tid	= first_tid;
+		__entry->block_nr	= block_nr;
+		__entry->freed		= freed;
+	),
+
+	TP_printk("dev %d,%d from %u to %u offset %lu freed %lu",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->tail_sequence, __entry->first_tid,
+		  __entry->block_nr, __entry->freed)
+);
+
+TRACE_EVENT(jbd_update_superblock_end,
+	TP_PROTO(journal_t *journal, int wait),
+
+	TP_ARGS(journal, wait),
+
+	TP_STRUCT__entry(
+		__field(	dev_t,	dev			)
+		__field(	int,	wait			)
+	),
+
+	TP_fast_assign(
+		__entry->dev		= journal->j_fs_dev->bd_dev;
+		__entry->wait		= wait;
+	),
+
+	TP_printk("dev %d,%d wait %d",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		   __entry->wait)
+);
+
+#endif /* _TRACE_JBD_H */
+
+/* This part must be outside protection */
+#include <trace/define_trace.h>