btrfs: avoid overflowing f_bfree
authorLuis de Bethencourt <luisbg@osg.samsung.com>
Wed, 30 Mar 2016 22:18:14 +0000 (23:18 +0100)
committerDavid Sterba <dsterba@suse.com>
Fri, 6 May 2016 13:22:49 +0000 (15:22 +0200)
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>
Signed-off-by: David Sterba <dsterba@suse.com>
fs/btrfs/super.c

index bdca79ce45f1bb1d5097323a6c0136cf3dc58a45..fe03efb5bec0a218988f7d0cce6c1d96eb12056e 100644 (file)
@@ -2100,7 +2100,11 @@ 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);