[PATCH 0/2] space_info update/creation refactoring

[PATCH 0/2] space_info update/creation refactoring

[Nikolay Borisov]

Here are two patches which aim to disentangle and make more explicit 
the situation when a space_info has to be created VS when space_info values 
are being updated. 

It survived multiple xfstest runs with additional ASSERTs which I have removed
in this posting. One such assert which didn't trigger was in 
btrfs_make_block_group. Despite that I haven't changed the logic of the code
to try and allocate a space_info if one can't be found. However, I'd like 
someone more knowledgble of the internals of bg lifecycle to tell me if 
in fact btrfs_make_block_group can be called with flags for non-existant 
struct space_info. 



Nikolay Borisov (2):
  btrfs: Separate space_info create/update
  btrfs: Refactor update_space_info

 fs/btrfs/extent-tree.c | 172 +++++++++++++++++++++----------------------------
 1 file changed, 73 insertions(+), 99 deletions(-)

-- 
2.7.4

[PATCH 1/2] btrfs: Separate space_info create/update

[Nikolay Borisov]

Currently the struct space_info creation code is intermixed in the
udpate_space_info function. There are well-defined points at which the we
actually want to create brand-new space_info structs (e.g. during mount of
the filesystem as well as sometimes when adding/initialising new chunks). In
such cases udpate_space_info is called with 0 as the bytes parameter. All of
this makes for spaghetti code.

Fix it by factoring out the creation code in a separate create_space_info
structure. This also allows to simplify the internals. Furthermore it will
make the update_space_info function not fail, allowing to remove error
handling in callers. This will come in a follow up patch.

This bears no functional changes

Signed-off-by: Nikolay Borisov <nborisov@suse.com>
---
 fs/btrfs/extent-tree.c | 127 ++++++++++++++++++++++++-------------------------
 1 file changed, 62 insertions(+), 65 deletions(-)

diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index be5477676cc8..e71921d8c610 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -3914,15 +3914,57 @@ static const char *alloc_name(u64 flags)
 	};
 }
 
+static int create_space_info(struct btrfs_fs_info *info, u64 flags,
+			     struct btrfs_space_info **new) {
+
+	struct btrfs_space_info *space_info;
+	int i;
+	int ret;
+
+	space_info = kzalloc(sizeof(*space_info), GFP_NOFS);
+	if (!space_info)
+		return -ENOMEM;
+
+	ret = percpu_counter_init(&space_info->total_bytes_pinned, 0, GFP_KERNEL);
+	if (ret) {
+		kfree(space_info);
+		return ret;
+	}
+
+	for (i = 0; i < BTRFS_NR_RAID_TYPES; i++)
+		INIT_LIST_HEAD(&space_info->block_groups[i]);
+	init_rwsem(&space_info->groups_sem);
+	spin_lock_init(&space_info->lock);
+	space_info->flags = flags & BTRFS_BLOCK_GROUP_TYPE_MASK;
+	space_info->force_alloc = CHUNK_ALLOC_NO_FORCE;
+	init_waitqueue_head(&space_info->wait);
+	INIT_LIST_HEAD(&space_info->ro_bgs);
+	INIT_LIST_HEAD(&space_info->tickets);
+	INIT_LIST_HEAD(&space_info->priority_tickets);
+
+	ret = kobject_init_and_add(&space_info->kobj, &space_info_ktype,
+				    info->space_info_kobj, "%s",
+				    alloc_name(space_info->flags));
+	if (ret) {
+		kfree(space_info);
+		return ret;
+	}
+
+	*new = space_info;
+	list_add_rcu(&space_info->list, &info->space_info);
+	if (flags & BTRFS_BLOCK_GROUP_DATA)
+		info->data_sinfo = space_info;
+
+	return ret;
+}
+
 static int update_space_info(struct btrfs_fs_info *info, u64 flags,
 			     u64 total_bytes, u64 bytes_used,
 			     u64 bytes_readonly,
 			     struct btrfs_space_info **space_info)
 {
 	struct btrfs_space_info *found;
-	int i;
 	int factor;
-	int ret;
 
 	if (flags & (BTRFS_BLOCK_GROUP_DUP | BTRFS_BLOCK_GROUP_RAID1 |
 		     BTRFS_BLOCK_GROUP_RAID10))
@@ -3946,53 +3988,6 @@ static int update_space_info(struct btrfs_fs_info *info, u64 flags,
 		*space_info = found;
 		return 0;
 	}
-	found = kzalloc(sizeof(*found), GFP_NOFS);
-	if (!found)
-		return -ENOMEM;
-
-	ret = percpu_counter_init(&found->total_bytes_pinned, 0, GFP_KERNEL);
-	if (ret) {
-		kfree(found);
-		return ret;
-	}
-
-	for (i = 0; i < BTRFS_NR_RAID_TYPES; i++)
-		INIT_LIST_HEAD(&found->block_groups[i]);
-	init_rwsem(&found->groups_sem);
-	spin_lock_init(&found->lock);
-	found->flags = flags & BTRFS_BLOCK_GROUP_TYPE_MASK;
-	found->total_bytes = total_bytes;
-	found->disk_total = total_bytes * factor;
-	found->bytes_used = bytes_used;
-	found->disk_used = bytes_used * factor;
-	found->bytes_pinned = 0;
-	found->bytes_reserved = 0;
-	found->bytes_readonly = bytes_readonly;
-	found->bytes_may_use = 0;
-	found->full = 0;
-	found->max_extent_size = 0;
-	found->force_alloc = CHUNK_ALLOC_NO_FORCE;
-	found->chunk_alloc = 0;
-	found->flush = 0;
-	init_waitqueue_head(&found->wait);
-	INIT_LIST_HEAD(&found->ro_bgs);
-	INIT_LIST_HEAD(&found->tickets);
-	INIT_LIST_HEAD(&found->priority_tickets);
-
-	ret = kobject_init_and_add(&found->kobj, &space_info_ktype,
-				    info->space_info_kobj, "%s",
-				    alloc_name(found->flags));
-	if (ret) {
-		kfree(found);
-		return ret;
-	}
-
-	*space_info = found;
-	list_add_rcu(&found->list, &info->space_info);
-	if (flags & BTRFS_BLOCK_GROUP_DATA)
-		info->data_sinfo = found;
-
-	return ret;
 }
 
 static void set_avail_alloc_bits(struct btrfs_fs_info *fs_info, u64 flags)
@@ -4495,10 +4490,9 @@ static int do_chunk_alloc(struct btrfs_trans_handle *trans,
 
 	space_info = __find_space_info(fs_info, flags);
 	if (!space_info) {
-		ret = update_space_info(fs_info, flags, 0, 0, 0, &space_info);
+		ret = create_space_info(fs_info, flags, &space_info);
 		BUG_ON(ret); /* -ENOMEM */
 	}
-	BUG_ON(!space_info); /* Logic error */
 
 again:
 	spin_lock(&space_info->lock);
@@ -5763,7 +5757,7 @@ int btrfs_orphan_reserve_metadata(struct btrfs_trans_handle *trans,
 	 */
 	u64 num_bytes = btrfs_calc_trans_metadata_size(fs_info, 1);
 
-	trace_btrfs_space_reservation(fs_info, "orphan", btrfs_ino(inode), 
+	trace_btrfs_space_reservation(fs_info, "orphan", btrfs_ino(inode),
 			num_bytes, 1);
 	return btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes, 1);
 }
@@ -10153,6 +10147,7 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans,
 	struct btrfs_block_group_cache *cache;
 	int ret;
 
+
 	btrfs_set_log_full_commit(fs_info, trans);
 
 	cache = btrfs_create_block_group_cache(fs_info, chunk_offset, size);
@@ -10191,16 +10186,18 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans,
 	}
 #endif
 	/*
-	 * Call to ensure the corresponding space_info object is created and
-	 * assigned to our block group, but don't update its counters just yet.
-	 * We want our bg to be added to the rbtree with its ->space_info set.
+	 * Ensure the corresponding space_info object is created and
+	 * assigned to our block group. We want our bg to be added to the rbtree
+	 * with its ->space_info set.
 	 */
-	ret = update_space_info(fs_info, cache->flags, 0, 0, 0,
-				&cache->space_info);
-	if (ret) {
-		btrfs_remove_free_space_cache(cache);
-		btrfs_put_block_group(cache);
-		return ret;
+	cache->space_info = __find_space_info(fs_info, cache->flags);
+	if (!cache->space_info) {
+		ret = create_space_info(fs_info, cache->flags, &cache->space_info);
+		if (ret) {
+			btrfs_remove_free_space_cache(cache);
+			btrfs_put_block_group(cache);
+			return ret;
+		}
 	}
 
 	ret = btrfs_add_block_group_cache(fs_info, cache);
@@ -10774,21 +10771,21 @@ int btrfs_init_space_info(struct btrfs_fs_info *fs_info)
 		mixed = 1;
 
 	flags = BTRFS_BLOCK_GROUP_SYSTEM;
-	ret = update_space_info(fs_info, flags, 0, 0, 0, &space_info);
+	ret = create_space_info(fs_info, flags, &space_info);
 	if (ret)
 		goto out;
 
 	if (mixed) {
 		flags = BTRFS_BLOCK_GROUP_METADATA | BTRFS_BLOCK_GROUP_DATA;
-		ret = update_space_info(fs_info, flags, 0, 0, 0, &space_info);
+		ret = create_space_info(fs_info, flags, &space_info);
 	} else {
 		flags = BTRFS_BLOCK_GROUP_METADATA;
-		ret = update_space_info(fs_info, flags, 0, 0, 0, &space_info);
+		ret = create_space_info(fs_info, flags, &space_info);
 		if (ret)
 			goto out;
 
 		flags = BTRFS_BLOCK_GROUP_DATA;
-		ret = update_space_info(fs_info, flags, 0, 0, 0, &space_info);
+		ret = create_space_info(fs_info, flags, &space_info);
 	}
 out:
 	return ret;
-- 
2.7.4

[PATCH 2/2] btrfs: Refactor update_space_info

[Nikolay Borisov]

Following the factoring out of the creation code udpate_space_info can only
be called for already-existing space_info structs. As such it cannot fail.
Remove superfulous error handling and make the function return void.

Signed-off-by: Nikolay Borisov <nborisov@suse.com>
---
 fs/btrfs/extent-tree.c | 53 ++++++++++++++------------------------------------
 1 file changed, 15 insertions(+), 38 deletions(-)

diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index e71921d8c610..3ab1f88af038 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -3958,7 +3958,7 @@ static int create_space_info(struct btrfs_fs_info *info, u64 flags,
 	return ret;
 }
 
-static int update_space_info(struct btrfs_fs_info *info, u64 flags,
+static void update_space_info(struct btrfs_fs_info *info, u64 flags,
 			     u64 total_bytes, u64 bytes_used,
 			     u64 bytes_readonly,
 			     struct btrfs_space_info **space_info)
@@ -3973,21 +3973,18 @@ static int update_space_info(struct btrfs_fs_info *info, u64 flags,
 		factor = 1;
 
 	found = __find_space_info(info, flags);
-	if (found) {
-		spin_lock(&found->lock);
-		found->total_bytes += total_bytes;
-		found->disk_total += total_bytes * factor;
-		found->bytes_used += bytes_used;
-		found->disk_used += bytes_used * factor;
-		found->bytes_readonly += bytes_readonly;
-		if (total_bytes > 0)
-			found->full = 0;
-		space_info_add_new_bytes(info, found, total_bytes -
-					 bytes_used - bytes_readonly);
-		spin_unlock(&found->lock);
-		*space_info = found;
-		return 0;
-	}
+	spin_lock(&found->lock);
+	found->total_bytes += total_bytes;
+	found->disk_total += total_bytes * factor;
+	found->bytes_used += bytes_used;
+	found->disk_used += bytes_used * factor;
+	found->bytes_readonly += bytes_readonly;
+	if (total_bytes > 0)
+		found->full = 0;
+	space_info_add_new_bytes(info, found, total_bytes -
+				 bytes_used - bytes_readonly);
+	spin_unlock(&found->lock);
+	*space_info = found;
 }
 
 static void set_avail_alloc_bits(struct btrfs_fs_info *fs_info, u64 flags)
@@ -10041,19 +10038,9 @@ int btrfs_read_block_groups(struct btrfs_fs_info *info)
 		}
 
 		trace_btrfs_add_block_group(info, cache, 0);
-		ret = update_space_info(info, cache->flags, found_key.offset,
+		update_space_info(info, cache->flags, found_key.offset,
 					btrfs_block_group_used(&cache->item),
 					cache->bytes_super, &space_info);
-		if (ret) {
-			btrfs_remove_free_space_cache(cache);
-			spin_lock(&info->block_group_cache_lock);
-			rb_erase(&cache->cache_node,
-				 &info->block_group_cache_tree);
-			RB_CLEAR_NODE(&cache->cache_node);
-			spin_unlock(&info->block_group_cache_lock);
-			btrfs_put_block_group(cache);
-			goto error;
-		}
 
 		cache->space_info = space_info;
 
@@ -10212,18 +10199,8 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans,
 	 * the rbtree, update the space info's counters.
 	 */
 	trace_btrfs_add_block_group(fs_info, cache, 1);
-	ret = update_space_info(fs_info, cache->flags, size, bytes_used,
+	update_space_info(fs_info, cache->flags, size, bytes_used,
 				cache->bytes_super, &cache->space_info);
-	if (ret) {
-		btrfs_remove_free_space_cache(cache);
-		spin_lock(&fs_info->block_group_cache_lock);
-		rb_erase(&cache->cache_node,
-			 &fs_info->block_group_cache_tree);
-		RB_CLEAR_NODE(&cache->cache_node);
-		spin_unlock(&fs_info->block_group_cache_lock);
-		btrfs_put_block_group(cache);
-		return ret;
-	}
 	update_global_block_rsv(fs_info);
 
 	__link_block_group(cache->space_info, cache);
-- 
2.7.4

Re: [PATCH 1/2] btrfs: Separate space_info create/update

[Jeff Mahoney]

[-- Attachment #1.1: Type: text/plain, Size: 8177 bytes --]

On 5/9/17 4:34 AM, Nikolay Borisov wrote:
> Currently the struct space_info creation code is intermixed in the
> udpate_space_info function. There are well-defined points at which the we
> actually want to create brand-new space_info structs (e.g. during mount of
> the filesystem as well as sometimes when adding/initialising new chunks). In
> such cases udpate_space_info is called with 0 as the bytes parameter. All of
> this makes for spaghetti code.
> 
> Fix it by factoring out the creation code in a separate create_space_info
> structure. This also allows to simplify the internals. Furthermore it will
> make the update_space_info function not fail, allowing to remove error
> handling in callers. This will come in a follow up patch.
> 
> This bears no functional changes
> 
> Signed-off-by: Nikolay Borisov <nborisov@suse.com>
> ---
>  fs/btrfs/extent-tree.c | 127 ++++++++++++++++++++++++-------------------------
>  1 file changed, 62 insertions(+), 65 deletions(-)
> 
> diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
> index be5477676cc8..e71921d8c610 100644
> --- a/fs/btrfs/extent-tree.c
> +++ b/fs/btrfs/extent-tree.c
> @@ -3914,15 +3914,57 @@ static const char *alloc_name(u64 flags)
>  	};
>  }
>  
> +static int create_space_info(struct btrfs_fs_info *info, u64 flags,
> +			     struct btrfs_space_info **new) {
> +
> +	struct btrfs_space_info *space_info;
> +	int i;
> +	int ret;
> +
> +	space_info = kzalloc(sizeof(*space_info), GFP_NOFS);
> +	if (!space_info)
> +		return -ENOMEM;
> +
> +	ret = percpu_counter_init(&space_info->total_bytes_pinned, 0, GFP_KERNEL);
> +	if (ret) {
> +		kfree(space_info);
> +		return ret;
> +	}
> +
> +	for (i = 0; i < BTRFS_NR_RAID_TYPES; i++)
> +		INIT_LIST_HEAD(&space_info->block_groups[i]);
> +	init_rwsem(&space_info->groups_sem);
> +	spin_lock_init(&space_info->lock);
> +	space_info->flags = flags & BTRFS_BLOCK_GROUP_TYPE_MASK;
> +	space_info->force_alloc = CHUNK_ALLOC_NO_FORCE;
> +	init_waitqueue_head(&space_info->wait);
> +	INIT_LIST_HEAD(&space_info->ro_bgs);
> +	INIT_LIST_HEAD(&space_info->tickets);
> +	INIT_LIST_HEAD(&space_info->priority_tickets);
> +
> +	ret = kobject_init_and_add(&space_info->kobj, &space_info_ktype,
> +				    info->space_info_kobj, "%s",
> +				    alloc_name(space_info->flags));
> +	if (ret) {

Not your bug (it's mine from 6ab0a2029c), but we need
percpu_counter_destroy here.  I just sent a patch to fix the original
case so we can pull it into stable as well.

> +		kfree(space_info);
> +		return ret;
> +	}
> +
> +	*new = space_info;
> +	list_add_rcu(&space_info->list, &info->space_info);
> +	if (flags & BTRFS_BLOCK_GROUP_DATA)
> +		info->data_sinfo = space_info;
> +
> +	return ret;
> +}
> +
>  static int update_space_info(struct btrfs_fs_info *info, u64 flags,
>  			     u64 total_bytes, u64 bytes_used,
>  			     u64 bytes_readonly,
>  			     struct btrfs_space_info **space_info)
>  {
>  	struct btrfs_space_info *found;
> -	int i;
>  	int factor;
> -	int ret;
>  
>  	if (flags & (BTRFS_BLOCK_GROUP_DUP | BTRFS_BLOCK_GROUP_RAID1 |
>  		     BTRFS_BLOCK_GROUP_RAID10))
> @@ -3946,53 +3988,6 @@ static int update_space_info(struct btrfs_fs_info *info, u64 flags,
>  		*space_info = found;
>  		return 0;
>  	}
> -	found = kzalloc(sizeof(*found), GFP_NOFS);
> -	if (!found)
> -		return -ENOMEM;
> -
> -	ret = percpu_counter_init(&found->total_bytes_pinned, 0, GFP_KERNEL);
> -	if (ret) {
> -		kfree(found);
> -		return ret;
> -	}
> -
> -	for (i = 0; i < BTRFS_NR_RAID_TYPES; i++)
> -		INIT_LIST_HEAD(&found->block_groups[i]);
> -	init_rwsem(&found->groups_sem);
> -	spin_lock_init(&found->lock);
> -	found->flags = flags & BTRFS_BLOCK_GROUP_TYPE_MASK;
> -	found->total_bytes = total_bytes;
> -	found->disk_total = total_bytes * factor;
> -	found->bytes_used = bytes_used;
> -	found->disk_used = bytes_used * factor;
> -	found->bytes_pinned = 0;
> -	found->bytes_reserved = 0;
> -	found->bytes_readonly = bytes_readonly;
> -	found->bytes_may_use = 0;
> -	found->full = 0;
> -	found->max_extent_size = 0;
> -	found->force_alloc = CHUNK_ALLOC_NO_FORCE;
> -	found->chunk_alloc = 0;
> -	found->flush = 0;
> -	init_waitqueue_head(&found->wait);
> -	INIT_LIST_HEAD(&found->ro_bgs);
> -	INIT_LIST_HEAD(&found->tickets);
> -	INIT_LIST_HEAD(&found->priority_tickets);
> -
> -	ret = kobject_init_and_add(&found->kobj, &space_info_ktype,
> -				    info->space_info_kobj, "%s",
> -				    alloc_name(found->flags));
> -	if (ret) {
> -		kfree(found);
> -		return ret;
> -	}
> -
> -	*space_info = found;
> -	list_add_rcu(&found->list, &info->space_info);
> -	if (flags & BTRFS_BLOCK_GROUP_DATA)
> -		info->data_sinfo = found;
> -
> -	return ret;
>  }
>  
>  static void set_avail_alloc_bits(struct btrfs_fs_info *fs_info, u64 flags)
> @@ -4495,10 +4490,9 @@ static int do_chunk_alloc(struct btrfs_trans_handle *trans,
>  
>  	space_info = __find_space_info(fs_info, flags);
>  	if (!space_info) {
> -		ret = update_space_info(fs_info, flags, 0, 0, 0, &space_info);
> +		ret = create_space_info(fs_info, flags, &space_info);
>  		BUG_ON(ret); /* -ENOMEM */
>  	}
> -	BUG_ON(!space_info); /* Logic error */
>  
>  again:
>  	spin_lock(&space_info->lock);
> @@ -5763,7 +5757,7 @@ int btrfs_orphan_reserve_metadata(struct btrfs_trans_handle *trans,
>  	 */
>  	u64 num_bytes = btrfs_calc_trans_metadata_size(fs_info, 1);
>  
> -	trace_btrfs_space_reservation(fs_info, "orphan", btrfs_ino(inode), 
> +	trace_btrfs_space_reservation(fs_info, "orphan", btrfs_ino(inode),
>  			num_bytes, 1);
>  	return btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes, 1);
>  }
> @@ -10153,6 +10147,7 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans,
>  	struct btrfs_block_group_cache *cache;
>  	int ret;
>  
> +
>  	btrfs_set_log_full_commit(fs_info, trans);
>  
>  	cache = btrfs_create_block_group_cache(fs_info, chunk_offset, size);

Drop this chunk.

> @@ -10191,16 +10186,18 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans,
>  	}
>  #endif
>  	/*
> -	 * Call to ensure the corresponding space_info object is created and
> -	 * assigned to our block group, but don't update its counters just yet.
> -	 * We want our bg to be added to the rbtree with its ->space_info set.
> +	 * Ensure the corresponding space_info object is created and
> +	 * assigned to our block group. We want our bg to be added to the rbtree
> +	 * with its ->space_info set.
>  	 */
> -	ret = update_space_info(fs_info, cache->flags, 0, 0, 0,
> -				&cache->space_info);
> -	if (ret) {
> -		btrfs_remove_free_space_cache(cache);
> -		btrfs_put_block_group(cache);
> -		return ret;
> +	cache->space_info = __find_space_info(fs_info, cache->flags);
> +	if (!cache->space_info) {
> +		ret = create_space_info(fs_info, cache->flags, &cache->space_info);
> +		if (ret) {
> +			btrfs_remove_free_space_cache(cache);
> +			btrfs_put_block_group(cache);
> +			return ret;
> +		}
>  	}
>  
>  	ret = btrfs_add_block_group_cache(fs_info, cache);
> @@ -10774,21 +10771,21 @@ int btrfs_init_space_info(struct btrfs_fs_info *fs_info)
>  		mixed = 1;
>  
>  	flags = BTRFS_BLOCK_GROUP_SYSTEM;
> -	ret = update_space_info(fs_info, flags, 0, 0, 0, &space_info);
> +	ret = create_space_info(fs_info, flags, &space_info);
>  	if (ret)
>  		goto out;
>  
>  	if (mixed) {
>  		flags = BTRFS_BLOCK_GROUP_METADATA | BTRFS_BLOCK_GROUP_DATA;
> -		ret = update_space_info(fs_info, flags, 0, 0, 0, &space_info);
> +		ret = create_space_info(fs_info, flags, &space_info);
>  	} else {
>  		flags = BTRFS_BLOCK_GROUP_METADATA;
> -		ret = update_space_info(fs_info, flags, 0, 0, 0, &space_info);
> +		ret = create_space_info(fs_info, flags, &space_info);
>  		if (ret)
>  			goto out;
>  
>  		flags = BTRFS_BLOCK_GROUP_DATA;
> -		ret = update_space_info(fs_info, flags, 0, 0, 0, &space_info);
> +		ret = create_space_info(fs_info, flags, &space_info);
>  	}
>  out:
>  	return ret;
> 

Reviewed-by: Jeff Mahoney <jeffm@suse.com>

-Jeff

-- 
Jeff Mahoney
SUSE Labs


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 819 bytes --]

Re: [PATCH 2/2] btrfs: Refactor update_space_info

[Jeff Mahoney]

[-- Attachment #1.1: Type: text/plain, Size: 4200 bytes --]

On 5/9/17 4:34 AM, Nikolay Borisov wrote:
> Following the factoring out of the creation code udpate_space_info can only
> be called for already-existing space_info structs. As such it cannot fail.
> Remove superfulous error handling and make the function return void.
> 
> Signed-off-by: Nikolay Borisov <nborisov@suse.com>
> ---
>  fs/btrfs/extent-tree.c | 53 ++++++++++++++------------------------------------
>  1 file changed, 15 insertions(+), 38 deletions(-)
> 
> diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
> index e71921d8c610..3ab1f88af038 100644
> --- a/fs/btrfs/extent-tree.c
> +++ b/fs/btrfs/extent-tree.c
> @@ -3958,7 +3958,7 @@ static int create_space_info(struct btrfs_fs_info *info, u64 flags,
>  	return ret;
>  }
>  
> -static int update_space_info(struct btrfs_fs_info *info, u64 flags,
> +static void update_space_info(struct btrfs_fs_info *info, u64 flags,
>  			     u64 total_bytes, u64 bytes_used,
>  			     u64 bytes_readonly,
>  			     struct btrfs_space_info **space_info)

Since we don't have any errors to report anymore, let's just return the
space_info directly instead of passing it back via parameter.

> @@ -3973,21 +3973,18 @@ static int update_space_info(struct btrfs_fs_info *info, u64 flags,
>  		factor = 1;
>  
>  	found = __find_space_info(info, flags);

BUG_ON(!found);

> -	if (found) {
> -		spin_lock(&found->lock);
> -		found->total_bytes += total_bytes;
> -		found->disk_total += total_bytes * factor;
> -		found->bytes_used += bytes_used;
> -		found->disk_used += bytes_used * factor;
> -		found->bytes_readonly += bytes_readonly;
> -		if (total_bytes > 0)
> -			found->full = 0;
> -		space_info_add_new_bytes(info, found, total_bytes -
> -					 bytes_used - bytes_readonly);
> -		spin_unlock(&found->lock);
> -		*space_info = found;
> -		return 0;
> -	}
> +	spin_lock(&found->lock);
> +	found->total_bytes += total_bytes;
> +	found->disk_total += total_bytes * factor;
> +	found->bytes_used += bytes_used;
> +	found->disk_used += bytes_used * factor;
> +	found->bytes_readonly += bytes_readonly;
> +	if (total_bytes > 0)
> +		found->full = 0;
> +	space_info_add_new_bytes(info, found, total_bytes -
> +				 bytes_used - bytes_readonly);
> +	spin_unlock(&found->lock);
> +	*space_info = found;
>  }
>  
>  static void set_avail_alloc_bits(struct btrfs_fs_info *fs_info, u64 flags)
> @@ -10041,19 +10038,9 @@ int btrfs_read_block_groups(struct btrfs_fs_info *info)
>  		}
>  
>  		trace_btrfs_add_block_group(info, cache, 0);
> -		ret = update_space_info(info, cache->flags, found_key.offset,
> +		update_space_info(info, cache->flags, found_key.offset,
>  					btrfs_block_group_used(&cache->item),
>  					cache->bytes_super, &space_info);

Whitespace needs adjusting too.

> -		if (ret) {
> -			btrfs_remove_free_space_cache(cache);
> -			spin_lock(&info->block_group_cache_lock);
> -			rb_erase(&cache->cache_node,
> -				 &info->block_group_cache_tree);
> -			RB_CLEAR_NODE(&cache->cache_node);
> -			spin_unlock(&info->block_group_cache_lock);
> -			btrfs_put_block_group(cache);
> -			goto error;
> -		}
>  
>  		cache->space_info = space_info;
>  
> @@ -10212,18 +10199,8 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans,
>  	 * the rbtree, update the space info's counters.
>  	 */
>  	trace_btrfs_add_block_group(fs_info, cache, 1);
> -	ret = update_space_info(fs_info, cache->flags, size, bytes_used,
> +	update_space_info(fs_info, cache->flags, size, bytes_used,
>  				cache->bytes_super, &cache->space_info);

Same.

> -	if (ret) {
> -		btrfs_remove_free_space_cache(cache);
> -		spin_lock(&fs_info->block_group_cache_lock);
> -		rb_erase(&cache->cache_node,
> -			 &fs_info->block_group_cache_tree);
> -		RB_CLEAR_NODE(&cache->cache_node);
> -		spin_unlock(&fs_info->block_group_cache_lock);
> -		btrfs_put_block_group(cache);
> -		return ret;
> -	}
>  	update_global_block_rsv(fs_info);
>  
>  	__link_block_group(cache->space_info, cache);
> 

Otherwise, looks good.  With those changes, you can add:

Reviewed-by: Jeff Mahoney <jeffm@suse.com>

-Jeff

-- 
Jeff Mahoney
SUSE Labs


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 819 bytes --]

Re: [PATCH 0/2] space_info update/creation refactoring

[David Sterba]

On Tue, May 09, 2017 at 11:34:17AM +0300, Nikolay Borisov wrote:
> Here are two patches which aim to disentangle and make more explicit 
> the situation when a space_info has to be created VS when space_info values 
> are being updated. 
> 
> It survived multiple xfstest runs with additional ASSERTs which I have removed
> in this posting. One such assert which didn't trigger was in 
> btrfs_make_block_group. Despite that I haven't changed the logic of the code
> to try and allocate a space_info if one can't be found. However, I'd like 
> someone more knowledgble of the internals of bg lifecycle to tell me if 
> in fact btrfs_make_block_group can be called with flags for non-existant 
> struct space_info. 
> 
> 
> 
> Nikolay Borisov (2):
>   btrfs: Separate space_info create/update
>   btrfs: Refactor update_space_info

I don't see an updated version of this patchset, Jeff had some comments
that are beyond what I normally fix during commit. Please fix and
resend, thanks.

Re: [PATCH 0/2] space_info update/creation refactoring

[David Sterba]

On Thu, Jun 01, 2017 at 04:46:05PM +0200, David Sterba wrote:
> On Tue, May 09, 2017 at 11:34:17AM +0300, Nikolay Borisov wrote:
> > Here are two patches which aim to disentangle and make more explicit 
> > the situation when a space_info has to be created VS when space_info values 
> > are being updated. 
> > 
> > It survived multiple xfstest runs with additional ASSERTs which I have removed
> > in this posting. One such assert which didn't trigger was in 
> > btrfs_make_block_group. Despite that I haven't changed the logic of the code
> > to try and allocate a space_info if one can't be found. However, I'd like 
> > someone more knowledgble of the internals of bg lifecycle to tell me if 
> > in fact btrfs_make_block_group can be called with flags for non-existant 
> > struct space_info. 
> > 
> > 
> > 
> > Nikolay Borisov (2):
> >   btrfs: Separate space_info create/update
> >   btrfs: Refactor update_space_info
> 
> I don't see an updated version of this patchset, Jeff had some comments
> that are beyond what I normally fix during commit. Please fix and
> resend, thanks.

Never mind, I did not see it in the main inbox, because it's been
already added to the for-next queue.