[PATCH 1/2] btrfs: fix mixed block count of available space

[PATCH 1/2] btrfs: fix mixed block count of available space

[Luis de Bethencourt]

Metadata for mixed block is already accounted in total data and should not
be counted as part of the free metadata space.

Signed-off-by: Luis de Bethencourt <luisbg@osg.samsung.com>
Link: https://bugzilla.kernel.org/show_bug.cgi?id=114281
---
 fs/btrfs/super.c | 16 +++++++++++++---
 1 file changed, 13 insertions(+), 3 deletions(-)

diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c
index 00b8f37..bdca79c 100644
--- a/fs/btrfs/super.c
+++ b/fs/btrfs/super.c
@@ -2051,6 +2051,7 @@ static int btrfs_statfs(struct dentry *dentry, struct kstatfs *buf)
 	struct btrfs_block_rsv *block_rsv = &fs_info->global_block_rsv;
 	int ret;
 	u64 thresh = 0;
+	int mixed = 0;
 
 	/*
 	 * holding chunk_muext to avoid allocating new chunks, holding
@@ -2076,8 +2077,17 @@ static int btrfs_statfs(struct dentry *dentry, struct kstatfs *buf)
 				}
 			}
 		}
-		if (found->flags & BTRFS_BLOCK_GROUP_METADATA)
-			total_free_meta += found->disk_total - found->disk_used;
+
+		/*
+		 * Metadata in mixed block goup profiles are accounted in data
+		 */
+		if (!mixed && found->flags & BTRFS_BLOCK_GROUP_METADATA) {
+			if (found->flags & BTRFS_BLOCK_GROUP_DATA)
+				mixed = 1;
+			else
+				total_free_meta += found->disk_total -
+					found->disk_used;
+		}
 
 		total_used += found->disk_used;
 	}
@@ -2115,7 +2125,7 @@ static int btrfs_statfs(struct dentry *dentry, struct kstatfs *buf)
 	 */
 	thresh = 4 * 1024 * 1024;
 
-	if (total_free_meta - thresh < block_rsv->size)
+	if (!mixed && total_free_meta - thresh < block_rsv->size)
 		buf->f_bavail = 0;
 
 	buf->f_type = BTRFS_SUPER_MAGIC;
-- 
2.5.3

[PATCH 2/2] btrfs: avoid overflowing f_bfree

[Luis de Bethencourt]

Since mixed block groups accounting isn't byte-accurate and f_bree is an
unsigned integer, it could overflow. Avoid this.

Signed-off-by: Luis de Bethencourt <luisbg@osg.samsung.com>
Suggested-by: David Sterba <dsterba@suse.com>
---
 fs/btrfs/super.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c
index bdca79c..93376d0 100644
--- a/fs/btrfs/super.c
+++ b/fs/btrfs/super.c
@@ -2101,6 +2101,11 @@ static int btrfs_statfs(struct dentry *dentry, struct kstatfs *buf)
 	/* Account global block reserve as used, it's in logical size already */
 	spin_lock(&block_rsv->lock);
 	buf->f_bfree -= block_rsv->size >> bits;
+	/* Mixed block groups accounting is not byte-accurate, avoid overflow */
+	if (buf->f_bfree >= block_rsv->size >> bits)
+		buf->f_bfree -= block_rsv->size >> bits;
+	else
+		buf->f_bfree = 0;
 	spin_unlock(&block_rsv->lock);
 
 	buf->f_bavail = div_u64(total_free_data, factor);
-- 
2.5.3

Re: [PATCH 2/2] btrfs: avoid overflowing f_bfree

[Filipe Manana]

On Wed, Mar 30, 2016 at 9:53 PM, Luis de Bethencourt
<luisbg@osg.samsung.com> wrote:
> Since mixed block groups accounting isn't byte-accurate and f_bree is an
> unsigned integer, it could overflow. Avoid this.
>
> Signed-off-by: Luis de Bethencourt <luisbg@osg.samsung.com>
> Suggested-by: David Sterba <dsterba@suse.com>
> ---
>  fs/btrfs/super.c | 5 +++++
>  1 file changed, 5 insertions(+)
>
> diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c
> index bdca79c..93376d0 100644
> --- a/fs/btrfs/super.c
> +++ b/fs/btrfs/super.c
> @@ -2101,6 +2101,11 @@ static int btrfs_statfs(struct dentry *dentry, struct kstatfs *buf)
>         /* Account global block reserve as used, it's in logical size already */
>         spin_lock(&block_rsv->lock);
>         buf->f_bfree -= block_rsv->size >> bits;

You forgot to remove the line above, didn't you?

> +       /* Mixed block groups accounting is not byte-accurate, avoid overflow */
> +       if (buf->f_bfree >= block_rsv->size >> bits)
> +               buf->f_bfree -= block_rsv->size >> bits;
> +       else
> +               buf->f_bfree = 0;
>         spin_unlock(&block_rsv->lock);
>
>         buf->f_bavail = div_u64(total_free_data, factor);
> --
> 2.5.3
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html



-- 
Filipe David Manana,

"Reasonable men adapt themselves to the world.
 Unreasonable men adapt the world to themselves.
 That's why all progress depends on unreasonable men."

Re: [PATCH 2/2] btrfs: avoid overflowing f_bfree

[Luis de Bethencourt]

On 30/03/16 22:48, Filipe Manana wrote:
> On Wed, Mar 30, 2016 at 9:53 PM, Luis de Bethencourt
> <luisbg@osg.samsung.com> wrote:
>> Since mixed block groups accounting isn't byte-accurate and f_bree is an
>> unsigned integer, it could overflow. Avoid this.
>>
>> Signed-off-by: Luis de Bethencourt <luisbg@osg.samsung.com>
>> Suggested-by: David Sterba <dsterba@suse.com>
>> ---
>>  fs/btrfs/super.c | 5 +++++
>>  1 file changed, 5 insertions(+)
>>
>> diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c
>> index bdca79c..93376d0 100644
>> --- a/fs/btrfs/super.c
>> +++ b/fs/btrfs/super.c
>> @@ -2101,6 +2101,11 @@ static int btrfs_statfs(struct dentry *dentry, struct kstatfs *buf)
>>         /* Account global block reserve as used, it's in logical size already */
>>         spin_lock(&block_rsv->lock);
>>         buf->f_bfree -= block_rsv->size >> bits;
> 
> You forgot to remove the line above, didn't you?
> 

Shoot! Indeed I did, sorry. Thanks for noticing.

Sending version 2.

Luis

>> +       /* Mixed block groups accounting is not byte-accurate, avoid overflow */
>> +       if (buf->f_bfree >= block_rsv->size >> bits)
>> +               buf->f_bfree -= block_rsv->size >> bits;
>> +       else
>> +               buf->f_bfree = 0;
>>         spin_unlock(&block_rsv->lock);
>>
>>         buf->f_bavail = div_u64(total_free_data, factor);
>> --
>> 2.5.3
>>
>> --