Merge tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4

Pull ext4 updates from Ted Ts'o:
 "New features for 3.12:

   - Added aggressive extent caching using the extent status tree.  This
     can actually decrease memory usage in read-mostly workloads since
     the information is much more compactly stored in the extent status
     tree than if we had to keep the extent tree metadata blocks in the
     buffer cache.  This also improves Asynchronous I/O since it is it
     makes much less likely that we need to do metadata I/O to lookup
     the extent tree information.

   - Improve the recovery after corrupted allocation bitmaps are found
     when running in errors=ignore mode.

  Also fixed some writeback vs truncate races when using a blocksize
  less than the page size"

* tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4: (25 commits)
  ext4: allow specifying external journal by pathname mount option
  ext4: mark group corrupt on group descriptor checksum
  ext4: mark block group as corrupt on inode bitmap error
  ext4: mark block group as corrupt on block bitmap error
  ext4: fix type declaration of ext4_validate_block_bitmap
  ext4: error out if verifying the block bitmap fails
  jbd2: Fix endian mixing problems in the checksumming code
  ext4: isolate ext4_extents.h file
  ext4: Fix misspellings using 'codespell' tool
  ext4: convert write_begin methods to stable_page_writes semantics
  ext4: fix use of potentially uninitialized variables in debugging code
  ext4: fix lost truncate due to race with writeback
  ext4: simplify truncation code in ext4_setattr()
  ext4: fix ext4_writepages() in presence of truncate
  ext4: move test whether extent to map can be extended to one place
  ext4: fix warning in ext4_da_update_reserve_space()
  quota: provide interface for readding allocated space into reserved space
  ext4: avoid reusing recently deleted inodes in no journal mode
  ext4: allocate delayed allocation blocks before rename
  ext4: start handle at least possible moment when renaming files
  ...

1 files changed, 42 insertions, 7 deletions

diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c
index 4bbbf13bd74..a41e3ba8cfa 100644
--- a/fs/ext4/mballoc.c
+++ b/fs/ext4/mballoc.c
@@ -751,13 +751,15 @@ void ext4_mb_generate_buddy(struct super_block *sb,
 
 	if (free != grp->bb_free) {
 		ext4_grp_locked_error(sb, group, 0, 0,
-				      "%u clusters in bitmap, %u in gd",
+				      "%u clusters in bitmap, %u in gd; "
+				      "block bitmap corrupt.",
 				      free, grp->bb_free);
 		/*
-		 * If we intent to continue, we consider group descritor
+		 * If we intend to continue, we consider group descriptor
 		 * corrupt and update bb_free using bitmap value
 		 */
 		grp->bb_free = free;
+		set_bit(EXT4_GROUP_INFO_BBITMAP_CORRUPT_BIT, &grp->bb_state);
 	}
 	mb_set_largest_free_order(sb, grp);
 
@@ -1398,6 +1400,10 @@ static void mb_free_blocks(struct inode *inode, struct ext4_buddy *e4b,
 
 	BUG_ON(last >= (sb->s_blocksize << 3));
 	assert_spin_locked(ext4_group_lock_ptr(sb, e4b->bd_group));
+	/* Don't bother if the block group is corrupt. */
+	if (unlikely(EXT4_MB_GRP_BBITMAP_CORRUPT(e4b->bd_info)))
+		return;
+
 	mb_check_buddy(e4b);
 	mb_free_blocks_double(inode, e4b, first, count);
 
@@ -1423,7 +1429,11 @@ static void mb_free_blocks(struct inode *inode, struct ext4_buddy *e4b,
 				      inode ? inode->i_ino : 0,
 				      blocknr,
 				      "freeing already freed block "
-				      "(bit %u)", block);
+				      "(bit %u); block bitmap corrupt.",
+				      block);
+		/* Mark the block group as corrupt. */
+		set_bit(EXT4_GROUP_INFO_BBITMAP_CORRUPT_BIT,
+			&e4b->bd_info->bb_state);
 		mb_regenerate_buddy(e4b);
 		goto done;
 	}
@@ -1790,6 +1800,11 @@ int ext4_mb_find_by_goal(struct ext4_allocation_context *ac,
 	if (err)
 		return err;
 
+	if (unlikely(EXT4_MB_GRP_BBITMAP_CORRUPT(e4b->bd_info))) {
+		ext4_mb_unload_buddy(e4b);
+		return 0;
+	}
+
 	ext4_lock_group(ac->ac_sb, group);
 	max = mb_find_extent(e4b, ac->ac_g_ex.fe_start,
 			     ac->ac_g_ex.fe_len, &ex);
@@ -1987,6 +2002,9 @@ static int ext4_mb_good_group(struct ext4_allocation_context *ac,
 	if (cr <= 2 && free < ac->ac_g_ex.fe_len)
 		return 0;
 
+	if (unlikely(EXT4_MB_GRP_BBITMAP_CORRUPT(grp)))
+		return 0;
+
 	/* We only do this if the grp has never been initialized */
 	if (unlikely(EXT4_MB_GRP_NEED_INIT(grp))) {
 		int ret = ext4_mb_init_group(ac->ac_sb, group);
@@ -4585,6 +4603,7 @@ void ext4_free_blocks(handle_t *handle, struct inode *inode,
 	struct buffer_head *gd_bh;
 	ext4_group_t block_group;
 	struct ext4_sb_info *sbi;
+	struct ext4_inode_info *ei = EXT4_I(inode);
 	struct ext4_buddy e4b;
 	unsigned int count_clusters;
 	int err = 0;
@@ -4673,6 +4692,10 @@ do_more:
 	overflow = 0;
 	ext4_get_group_no_and_offset(sb, block, &block_group, &bit);
 
+	if (unlikely(EXT4_MB_GRP_BBITMAP_CORRUPT(
+			ext4_get_group_info(sb, block_group))))
+		return;
+
 	/*
 	 * Check to see if we are freeing blocks across a group
 	 * boundary.
@@ -4784,7 +4807,6 @@ do_more:
 	ext4_block_bitmap_csum_set(sb, block_group, gdp, bitmap_bh);
 	ext4_group_desc_csum_set(sb, block_group, gdp);
 	ext4_unlock_group(sb, block_group);
-	percpu_counter_add(&sbi->s_freeclusters_counter, count_clusters);
 
 	if (sbi->s_log_groups_per_flex) {
 		ext4_group_t flex_group = ext4_flex_group(sbi, block_group);
@@ -4792,10 +4814,23 @@ do_more:
 			     &sbi->s_flex_groups[flex_group].free_clusters);
 	}
 
-	ext4_mb_unload_buddy(&e4b);
-
-	if (!(flags & EXT4_FREE_BLOCKS_NO_QUOT_UPDATE))
+	if (flags & EXT4_FREE_BLOCKS_RESERVE && ei->i_reserved_data_blocks) {
+		percpu_counter_add(&sbi->s_dirtyclusters_counter,
+				   count_clusters);
+		spin_lock(&ei->i_block_reservation_lock);
+		if (flags & EXT4_FREE_BLOCKS_METADATA)
+			ei->i_reserved_meta_blocks += count_clusters;
+		else
+			ei->i_reserved_data_blocks += count_clusters;
+		spin_unlock(&ei->i_block_reservation_lock);
+		if (!(flags & EXT4_FREE_BLOCKS_NO_QUOT_UPDATE))
+			dquot_reclaim_block(inode,
+					EXT4_C2B(sbi, count_clusters));
+	} else if (!(flags & EXT4_FREE_BLOCKS_NO_QUOT_UPDATE))
 		dquot_free_block(inode, EXT4_C2B(sbi, count_clusters));
+	percpu_counter_add(&sbi->s_freeclusters_counter, count_clusters);
+
+	ext4_mb_unload_buddy(&e4b);
 
 	/* We dirtied the bitmap block */
 	BUFFER_TRACE(bitmap_bh, "dirtied bitmap block");