ext4: Fix bigalloc quota accounting and i_blocks value

With bigalloc changes, the i_blocks value was not correctly set (it was still
set to number of blocks being used, but in case of bigalloc, we want i_blocks
to represent the number of clusters being used). Since the quota subsystem sets
the i_blocks value, this patch fixes the quota accounting and makes sure that
the i_blocks value is set correctly.
Signed-off-by: Aditya Kali <adityakali@google.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>

fs/ext4/balloc.c

@@ -485,7 +485,7 @@ int ext4_should_retry_alloc(struct super_block *sb, int *retries)
  * @handle:             handle to this transaction
  * @inode:              file inode
  * @goal:               given target block(filesystem wide)
- * @count:		pointer to total number of blocks needed
+ * @count:		pointer to total number of clusters needed
  * @errp:               error code
  *
  * Return 1st allocated block number on success, *count stores total account
@@ -517,7 +517,8 @@ ext4_fsblk_t ext4_new_meta_blocks(handle_t *handle, struct inode *inode,
 		spin_lock(&EXT4_I(inode)->i_block_reservation_lock);
 		EXT4_I(inode)->i_allocated_meta_blocks += ar.len;
 		spin_unlock(&EXT4_I(inode)->i_block_reservation_lock);
-		dquot_alloc_block_nofail(inode, ar.len);
+		dquot_alloc_block_nofail(inode,
+				EXT4_C2B(EXT4_SB(inode->i_sb), ar.len));
 	}
 	return ret;
 }

fs/ext4/ext4.h

@@ -144,9 +144,17 @@ struct ext4_allocation_request {
 #define EXT4_MAP_UNWRITTEN	(1 << BH_Unwritten)
 #define EXT4_MAP_BOUNDARY	(1 << BH_Boundary)
 #define EXT4_MAP_UNINIT		(1 << BH_Uninit)
+/* Sometimes (in the bigalloc case, from ext4_da_get_block_prep) the caller of
+ * ext4_map_blocks wants to know whether or not the underlying cluster has
+ * already been accounted for. EXT4_MAP_FROM_CLUSTER conveys to the caller that
+ * the requested mapping was from previously mapped (or delayed allocated)
+ * cluster. We use BH_AllocFromCluster only for this flag. BH_AllocFromCluster
+ * should never appear on buffer_head's state flags.
+ */
+#define EXT4_MAP_FROM_CLUSTER	(1 << BH_AllocFromCluster)
 #define EXT4_MAP_FLAGS		(EXT4_MAP_NEW | EXT4_MAP_MAPPED |\
 				 EXT4_MAP_UNWRITTEN | EXT4_MAP_BOUNDARY |\
-				 EXT4_MAP_UNINIT)
+				 EXT4_MAP_UNINIT | EXT4_MAP_FROM_CLUSTER)
 
 struct ext4_map_blocks {
 	ext4_fsblk_t m_pblk;
@@ -1884,6 +1892,7 @@ extern int ext4_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf);
 extern qsize_t *ext4_get_reserved_space(struct inode *inode);
 extern void ext4_da_update_reserve_space(struct inode *inode,
 					int used, int quota_claim);
+extern int ext4_da_reserve_space(struct inode *inode, ext4_lblk_t lblock);
 
 /* indirect.c */
 extern int ext4_ind_map_blocks(handle_t *handle, struct inode *inode,
@@ -2284,6 +2293,11 @@ extern int ext4_multi_mount_protect(struct super_block *, ext4_fsblk_t);
 enum ext4_state_bits {
 	BH_Uninit	/* blocks are allocated but uninitialized on disk */
 	  = BH_JBDPrivateStart,
+	BH_AllocFromCluster,	/* allocated blocks were part of already
+				 * allocated cluster. Note that this flag will
+				 * never, ever appear in a buffer_head's state
+				 * flag. See EXT4_MAP_FROM_CLUSTER to see where
+				 * this is used. */
 };
 
 BUFFER_FNS(Uninit, uninit)

fs/ext4/ext4_extents.h

@@ -290,5 +290,7 @@ extern struct ext4_ext_path *ext4_ext_find_extent(struct inode *, ext4_lblk_t,
 							struct ext4_ext_path *);
 extern void ext4_ext_drop_refs(struct ext4_ext_path *);
 extern int ext4_ext_check_inode(struct inode *inode);
+extern int ext4_find_delalloc_cluster(struct inode *inode, ext4_lblk_t lblk,
+				      int search_hint_reverse);
 #endif /* _EXT4_EXTENTS */
 

fs/ext4/inode.c

@@ -300,14 +300,14 @@ void ext4_da_update_reserve_space(struct inode *inode,
 
 	/* Update quota subsystem for data blocks */
 	if (quota_claim)
-		dquot_claim_block(inode, used);
+		dquot_claim_block(inode, EXT4_C2B(sbi, used));
 	else {
 		/*
 		 * We did fallocate with an offset that is already delayed
 		 * allocated. So on delayed allocated writeback we should
 		 * not re-claim the quota for fallocated blocks.
 		 */
-		dquot_release_reservation_block(inode, used);
+		dquot_release_reservation_block(inode, EXT4_C2B(sbi, used));
 	}
 
 	/*
@@ -1037,14 +1037,14 @@ static int ext4_journalled_write_end(struct file *file,
 }
 
 /*
- * Reserve a single block located at lblock
+ * Reserve a single cluster located at lblock
  */
-static int ext4_da_reserve_space(struct inode *inode, ext4_lblk_t lblock)
+int ext4_da_reserve_space(struct inode *inode, ext4_lblk_t lblock)
 {
 	int retries = 0;
 	struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
 	struct ext4_inode_info *ei = EXT4_I(inode);
-	unsigned long md_needed;
+	unsigned int md_needed;
 	int ret;
 
 	/*
@@ -1054,7 +1054,8 @@ static int ext4_da_reserve_space(struct inode *inode, ext4_lblk_t lblock)
 	 */
 repeat:
 	spin_lock(&ei->i_block_reservation_lock);
-	md_needed = ext4_calc_metadata_amount(inode, lblock);
+	md_needed = EXT4_NUM_B2C(sbi,
+				 ext4_calc_metadata_amount(inode, lblock));
 	trace_ext4_da_reserve_space(inode, md_needed);
 	spin_unlock(&ei->i_block_reservation_lock);
 
@@ -1063,7 +1064,7 @@ static int ext4_da_reserve_space(struct inode *inode, ext4_lblk_t lblock)
 	 * us from metadata over-estimation, though we may go over by
 	 * a small amount in the end.  Here we just reserve for data.
 	 */
-	ret = dquot_reserve_block(inode, 1);
+	ret = dquot_reserve_block(inode, EXT4_C2B(sbi, 1));
 	if (ret)
 		return ret;
 	/*
@@ -1071,7 +1072,7 @@ static int ext4_da_reserve_space(struct inode *inode, ext4_lblk_t lblock)
 	 * we cannot afford to run out of free blocks.
 	 */
 	if (ext4_claim_free_blocks(sbi, md_needed + 1, 0)) {
-		dquot_release_reservation_block(inode, 1);
+		dquot_release_reservation_block(inode, EXT4_C2B(sbi, 1));
 		if (ext4_should_retry_alloc(inode->i_sb, &retries)) {
 			yield();
 			goto repeat;
@@ -1118,6 +1119,8 @@ static void ext4_da_release_space(struct inode *inode, int to_free)
 		 * We can release all of the reserved metadata blocks
 		 * only when we have written all of the delayed
 		 * allocation blocks.
+		 * Note that in case of bigalloc, i_reserved_meta_blocks,
+		 * i_reserved_data_blocks, etc. refer to number of clusters.
 		 */
 		percpu_counter_sub(&sbi->s_dirtyclusters_counter,
 				   ei->i_reserved_meta_blocks);
@@ -1130,7 +1133,7 @@ static void ext4_da_release_space(struct inode *inode, int to_free)
 
 	spin_unlock(&EXT4_I(inode)->i_block_reservation_lock);
 
-	dquot_release_reservation_block(inode, to_free);
+	dquot_release_reservation_block(inode, EXT4_C2B(sbi, to_free));
 }
 
 static void ext4_da_page_release_reservation(struct page *page,
@@ -1139,6 +1142,9 @@ static void ext4_da_page_release_reservation(struct page *page,
 	int to_release = 0;
 	struct buffer_head *head, *bh;
 	unsigned int curr_off = 0;
+	struct inode *inode = page->mapping->host;
+	struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
+	int num_clusters;
 
 	head = page_buffers(page);
 	bh = head;
@@ -1151,7 +1157,20 @@ static void ext4_da_page_release_reservation(struct page *page,
 		}
 		curr_off = next_off;
 	} while ((bh = bh->b_this_page) != head);
-	ext4_da_release_space(page->mapping->host, to_release);
+
+	/* If we have released all the blocks belonging to a cluster, then we
+	 * need to release the reserved space for that cluster. */
+	num_clusters = EXT4_NUM_B2C(sbi, to_release);
+	while (num_clusters > 0) {
+		ext4_fsblk_t lblk;
+		lblk = (page->index << (PAGE_CACHE_SHIFT - inode->i_blkbits)) +
+			((num_clusters - 1) << sbi->s_cluster_bits);
+		if (sbi->s_cluster_ratio == 1 ||
+		    !ext4_find_delalloc_cluster(inode, lblk, 1))
+			ext4_da_release_space(inode, 1);
+
+		num_clusters--;
+	}
 }
 
 /*
@@ -1352,7 +1371,8 @@ static void ext4_print_free_blocks(struct inode *inode)
 	       (long long) EXT4_C2B(EXT4_SB(inode->i_sb),
 		percpu_counter_sum(&sbi->s_freeclusters_counter)));
 	printk(KERN_CRIT "dirty_blocks=%lld\n",
-	       (long long) percpu_counter_sum(&sbi->s_dirtyclusters_counter));
+	       (long long) EXT4_C2B(EXT4_SB(inode->i_sb),
+		percpu_counter_sum(&sbi->s_dirtyclusters_counter)));
 	printk(KERN_CRIT "Block reservation details\n");
 	printk(KERN_CRIT "i_reserved_data_blocks=%u\n",
 	       EXT4_I(inode)->i_reserved_data_blocks);
@@ -1626,10 +1646,14 @@ static int ext4_da_get_block_prep(struct inode *inode, sector_t iblock,
 		/*
 		 * XXX: __block_write_begin() unmaps passed block, is it OK?
 		 */
-		ret = ext4_da_reserve_space(inode, iblock);
-		if (ret)
-			/* not enough space to reserve */
-			return ret;
+		/* If the block was allocated from previously allocated cluster,
+		 * then we dont need to reserve it again. */
+		if (!(map.m_flags & EXT4_MAP_FROM_CLUSTER)) {
+			ret = ext4_da_reserve_space(inode, iblock);
+			if (ret)
+				/* not enough space to reserve */
+				return ret;
+		}
 
 		map_bh(bh, inode->i_sb, invalid_block);
 		set_buffer_new(bh);

fs/ext4/mballoc.c

@@ -4718,6 +4718,9 @@ void ext4_free_blocks(handle_t *handle, struct inode *inode,
 
 	freed += count;
 
+	if (!(flags & EXT4_FREE_BLOCKS_NO_QUOT_UPDATE))
+		dquot_free_block(inode, EXT4_C2B(sbi, count_clusters));
+
 	/* We dirtied the bitmap block */
 	BUFFER_TRACE(bitmap_bh, "dirtied bitmap block");
 	err = ext4_handle_dirty_metadata(handle, NULL, bitmap_bh);
@@ -4736,8 +4739,6 @@ void ext4_free_blocks(handle_t *handle, struct inode *inode,
 	}
 	ext4_mark_super_dirty(sb);
 error_return:
-	if (freed && !(flags & EXT4_FREE_BLOCKS_NO_QUOT_UPDATE))
-		dquot_free_block(inode, freed);
 	brelse(bitmap_bh);
 	ext4_std_error(sb, err);
 	return;

fs/ext4/super.c

@@ -2473,7 +2473,8 @@ static ssize_t delayed_allocation_blocks_show(struct ext4_attr *a,
 					      char *buf)
 {
 	return snprintf(buf, PAGE_SIZE, "%llu\n",
-		(s64) percpu_counter_sum(&sbi->s_dirtyclusters_counter));
+		(s64) EXT4_C2B(sbi,
+			percpu_counter_sum(&sbi->s_dirtyclusters_counter)));
 }
 
 static ssize_t session_write_kbytes_show(struct ext4_attr *a,