Btrfs: Per file/directory controls for COW and compression

fs/btrfs/inode.c +0 -2

@@ -4581,8 +4583,6 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans,
 	location->offset = 0;
 	btrfs_set_key_type(location, BTRFS_INODE_ITEM_KEY);
 
-	btrfs_inherit_iflags(inode, dir);
-
 	if ((mode & S_IFREG)) {
 		if (btrfs_test_opt(root, NODATASUM))
 			BTRFS_I(inode)->flags |= BTRFS_INODE_NODATASUM;

Quoted 13 lines · liubo

fs/btrfs/inode.c +0 -2

@@ -4581,8 +4583,6 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans,
 	location->offset = 0;
 	btrfs_set_key_type(location, BTRFS_INODE_ITEM_KEY);
 
-	btrfs_inherit_iflags(inode, dir);
-
 	if ((mode & S_IFREG)) {
 		if (btrfs_test_opt(root, NODATASUM))
 			BTRFS_I(inode)->flags |= BTRFS_INODE_NODATASUM;

@@ -4590,6 +4590,8 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans,
 			BTRFS_I(inode)->flags |= BTRFS_INODE_NODATACOW;
 	}
 

fs/btrfs/inode.c static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans, · line 4593

Quoted 1 line

fs/btrfs/inode.c +1 -0

+	btrfs_inherit_iflags(inode, dir);

The problem is that btrfs_inherit_iflags() overwrites BTRFS_I(inode)->flags with the parent's flags, so you lose BTRFS_INODE_NODATA{SUM|COW}.

Johann

Data compression and data cow are controlled across the entire FS by =

options right now. ioctls are needed to set this on a per file or pe=

directory basis. This has been proposed previously, but VFS develope=

wanted us to use generic ioctls rather than btrfs-specific ones.

According to chris's comment, there should be just one true compressi=

method(probably LZO) stored in the super. However, before this, we w=

wait for that one method is stable enough to be adopted into the supe=

So I list it as a long term goal, and just store it in ram today.

After applying this patch, we can use the generic "FS_IOC_SETFLAGS" i=

control file and directory's datacow and compression attribute.

NOTE:
- The compression type is selected by such rules:
If we mount btrfs with compress options, ie, zlib/lzo, the type i=

Otherwise, we'll use the default compress type (zlib today).

v1->v2:
Rebase the patch with the latest btrfs.

Signed-off-by Liu Bo<[email protected]>

---
  fs/btrfs/ctree.h   |    1 +
  fs/btrfs/disk-io.c |    6 ++++++
  fs/btrfs/inode.c   |   32 ++++++++++++++++++++++++++++----
  fs/btrfs/ioctl.c   |   41 +++++++++++++++++++++++++++++++++++++----
  4 files changed, 72 insertions(+), 8 deletions(-)

fs/btrfs/ctree.h +0 -0

index 8b4b9d1..b77d1a5 100644

fs/btrfs/inode.c +2 -0

  			BTRFS_I(inode)->flags |=3D BTRFS_INODE_NODATACOW;
  	}
 
+	btrfs_inherit_iflags(inode, dir);
+
  	insert_inode_hash(inode);
  	inode_tree_add(inode);
  	return inode;

fs/btrfs/inode.c +6 -0

+{
+	struct btrfs_inode *b_dir =3D BTRFS_I(dir);
+	struct btrfs_inode *b_inode =3D BTRFS_I(inode);
+
+	if (b_dir->flags&  BTRFS_INODE_NODATACOW)
+		b_inode->flags |=3D BTRFS_INODE_NODATACOW;

fs/btrfs/ioctl.c +9 -0

  	return 0;
  }
 
+static int check_flags(unsigned int flags)
+{
+	if (flags&  ~(FS_IMMUTABLE_FL | FS_APPEND_FL | \
+		      FS_NOATIME_FL | FS_NODUMP_FL | \
+		      FS_SYNC_FL | FS_DIRSYNC_FL | \
+		      FS_NOCOMP_FL | FS_COMPR_FL | \
+		      FS_NOCOW_FL | FS_COW_FL))
+		return -EOPNOTSUPP;
+

fs/btrfs/ioctl.c +7 -0

+	 * flag may be changed automatically if compression code won't make
+	 * things smaller.
+	 */
+	if (flags&  FS_NOCOMP_FL) {
+		ip->flags&=3D ~BTRFS_INODE_COMPRESS;
+		ip->flags |=3D BTRFS_INODE_NOCOMPRESS;
+	} else if (flags&  FS_COMPR_FL) {

Hello,
I would like to ask about the status of this feature/patch, is it accepted into btrfs code, and how can I use it?
=20

I am interested in enabling compression in a specific folder(force-compress would be ideal) of a large btrfs volume, and disabling it for the rest.
=20

=20

Quoted 3 lines · liubo

Data compression and data cow are controlled across the entire FS by mount
options right now. ioctls are needed to set this on a per file or p=

Quoted 1 line

directory basis. This has been proposed previously, but VFS develop=

Quoted 3 lines

wanted us to use generic ioctls rather than btrfs-specific ones.

According to chris's comment, there should be just one true compress=

Quoted 1 line

method(probably LZO) stored in the super. However, before this, we =

Quoted 1 line

wait for that one method is stable enough to be adopted into the sup=

Quoted 9 lines

So I list it as a long term goal, and just store it in ram today.

After applying this patch, we can use the generic "FS_IOC_SETFLAGS" ioctl to
control file and directory's datacow and compression attribute.

NOTE:
- The compression type is selected by such rules:
If we mount btrfs with compress options, ie, zlib/lzo, the type =

Quoted 6 lines

it.
Otherwise, we'll use the default compress type (zlib today).

v1->v2:
Rebase the patch with the latest btrfs.

Quoted 6 lines

Signed-off-by Liu Bo<[email protected]>

---
  fs/btrfs/ctree.h   |    1 +
  fs/btrfs/disk-io.c |    6 ++++++
  fs/btrfs/inode.c   |   32 ++++++++++++++++++++++++++++----

fs/btrfs/ioctl.c | 41 +++++++++++++++++++++++++++++++++++++---=

Quoted 12 lines

  4 files changed, 72 insertions(+), 8 deletions(-)

fs/btrfs/ctree.h +1 -0

index 8b4b9d1..b77d1a5 100644
@@ -1283,6 +1283,7 @@ struct btrfs_root {
  #define BTRFS_INODE_NODUMP        (1<<  8)
  #define BTRFS_INODE_NOATIME        (1<<  9)
  #define BTRFS_INODE_DIRSYNC        (1<<  10)
+#define BTRFS_INODE_COMPRESS        (1<<  11)
 

fs/btrfs/ctree.h struct btrfs_root { · line 1288

Quoted 1 line

fs/btrfs/ctree.h +0 -0

  /* some macros to generate set/get funcs for the struct fields.  T=

Quoted 9 lines

fs/btrfs/ctree.h +0 -0

   * assumes there is a lefoo_to_cpu for every type, so lets make a

simple

fs/btrfs/disk-io.c +0 -0

index 3e1ea3e..a894c12 100644
@@ -1762,6 +1762,12 @@ struct btrfs_root *open_ctree(struct
 super_block *sb,
 

fs/btrfs/disk-io.c struct btrfs_root *open_ctree(struct super_block *sb,

Quoted 4 lines

fs/btrfs/disk-io.c +2 -0

      btrfs_check_super_valid(fs_info, sb->s_flags&  MS_RDONLY);
 
+    /*
+     * In the long term, we'll store the compression type in the su=

Quoted 12 lines

fs/btrfs/disk-io.c +4 -0

+     * block, and it'll be used for per file compression control.
+     */
+    fs_info->compress_type =3D BTRFS_COMPRESS_ZLIB;
+
      ret =3D btrfs_parse_options(tree_root, options);
      if (ret) {

          err =3D ret;

fs/btrfs/inode.c +0 -0

index db67821..e687bb9 100644
@@ -381,7 +381,8 @@ again:

fs/btrfs/inode.c again:

Quoted 8 lines

fs/btrfs/inode.c +2 -1

       */
      if (!(BTRFS_I(inode)->flags&  BTRFS_INODE_NOCOMPRESS)&&
      (btrfs_test_opt(root, COMPRESS) ||
-         (BTRFS_I(inode)->force_compress))) {
+         (BTRFS_I(inode)->force_compress) ||
+         (BTRFS_I(inode)->flags&  BTRFS_INODE_COMPRESS))) {
          WARN_ON(pages);
          pages =3D kzalloc(sizeof(struct page *) * nr_pages, GFP_NO=

Quoted 27 lines

fs/btrfs/inode.c +2 -1

@@ -1253,7 +1254,8 @@ static int run_delalloc_range(struct inode
 *inode, struct page *locked_page,
          ret =3D run_delalloc_nocow(inode, locked_page, start, end,
                       page_started, 0, nr_written);
      else if (!btrfs_test_opt(root, COMPRESS)&&
-         !(BTRFS_I(inode)->force_compress))
+         !(BTRFS_I(inode)->force_compress)&&
+         !(BTRFS_I(inode)->flags&  BTRFS_INODE_COMPRESS))
          ret =3D cow_file_range(inode, locked_page, start, end,
                        page_started, nr_written, 1);

else

fs/btrfs/inode.c +0 -2

@@ -4581,8 +4583,6 @@ static struct inode *btrfs_new_inode(struct
 btrfs_trans_handle *trans,
      location->offset =3D 0;
      btrfs_set_key_type(location, BTRFS_INODE_ITEM_KEY);
 
-    btrfs_inherit_iflags(inode, dir);
-
      if ((mode&  S_IFREG)) {
          if (btrfs_test_opt(root, NODATASUM))

              BTRFS_I(inode)->flags |=3D BTRFS_INODE_NODATASUM;

fs/btrfs/inode.c +0 -0

@@ -4590,6 +4590,8 @@ static struct inode *btrfs_new_inode(struct
 btrfs_trans_handle *trans,
              BTRFS_I(inode)->flags |=3D BTRFS_INODE_NODATACOW;
      }
 

fs/btrfs/inode.c static int btrfs_getattr(struct vfsmount *mnt=

Quoted 6 lines

fs/btrfs/inode.c +2 -0

+    btrfs_inherit_iflags(inode, dir);
+
      insert_inode_hash(inode);
      inode_tree_add(inode);

      return inode;

fs/btrfs/inode.c +0 -0

@@ -6803,6 +6805,26 @@ static int btrfs_getattr(struct vfsmount *mnt=

fs/btrfs/inode.c static int btrfs_getattr(struct vfsmount *mnt,

Quoted 5 lines

fs/btrfs/inode.c +2 -0

      return 0;
  }
 
+/*
+ * If a file is moved, it will inherit the cow and compression flag=

Quoted 4 lines

of the new
+ * directory.
+ */
+static void fixup_inode_flags(struct inode *dir, struct inode *inod=

Quoted 49 lines

fs/btrfs/inode.c +15 -0

+{
+    struct btrfs_inode *b_dir =3D BTRFS_I(dir);
+    struct btrfs_inode *b_inode =3D BTRFS_I(inode);
+
+    if (b_dir->flags&  BTRFS_INODE_NODATACOW)
+        b_inode->flags |=3D BTRFS_INODE_NODATACOW;
+    else
+        b_inode->flags&=3D ~BTRFS_INODE_NODATACOW;
+
+    if (b_dir->flags&  BTRFS_INODE_COMPRESS)
+        b_inode->flags |=3D BTRFS_INODE_COMPRESS;
+    else
+        b_inode->flags&=3D ~BTRFS_INODE_COMPRESS;
+}
+
  static int btrfs_rename(struct inode *old_dir, struct dentry

*old_dentry,
struct inode *new_dir, struct dentry *new_dentry)
{

fs/btrfs/inode.c +2 -0

@@ -6936,6 +6958,8 @@ static int btrfs_rename(struct inode *old_dir,
 struct dentry *old_dentry,
          }
      }
 
+    fixup_inode_flags(new_dir, old_inode);
+
      ret =3D btrfs_add_link(trans, new_dir, old_inode,
                   new_dentry->d_name.name,

                   new_dentry->d_name.len, 0, index);

fs/btrfs/ioctl.c +11 -0

index 5fdb2ab..93e7f6c 100644
@@ -138,6 +138,24 @@ static int btrfs_ioctl_getflags(struct file
 *file, void __user *arg)
      return 0;
  }
 
+static int check_flags(unsigned int flags)
+{
+    if (flags&  ~(FS_IMMUTABLE_FL | FS_APPEND_FL | \
+              FS_NOATIME_FL | FS_NODUMP_FL | \
+              FS_SYNC_FL | FS_DIRSYNC_FL | \
+              FS_NOCOMP_FL | FS_COMPR_FL | \
+              FS_NOCOW_FL | FS_COW_FL))
+        return -EOPNOTSUPP;
+
+    if ((flags&  FS_NOCOMP_FL)&&  (flags&  FS_COMPR_FL))
+        return -EINVAL;

fs/btrfs/ioctl.c static int btrfs_ioctl_getflags(struct file *file, void __user *arg)

Quoted 8 lines

fs/btrfs/ioctl.c +7 -0

+
+    if ((flags&  FS_NOCOW_FL)&&  (flags&  FS_COW_FL))
+        return -EINVAL;
+
+    return 0;
+}
+
  static int btrfs_ioctl_setflags(struct file *file, void __user *ar=

Quoted 22 lines

fs/btrfs/ioctl.c +0 -0

  {

      struct inode *inode =3D file->f_path.dentry->d_inode;

fs/btrfs/ioctl.c +3 -4

@@ -153,10 +171,9 @@ static int btrfs_ioctl_setflags(struct file
 *file, void __user *arg)
      if (copy_from_user(&flags, arg, sizeof(flags)))
          return -EFAULT;
 
-    if (flags&  ~(FS_IMMUTABLE_FL | FS_APPEND_FL | \
-              FS_NOATIME_FL | FS_NODUMP_FL | \
-              FS_SYNC_FL | FS_DIRSYNC_FL))
-        return -EOPNOTSUPP;
+    ret =3D check_flags(flags);
+    if (ret)
+        return ret;
 
      if (!is_owner_or_cap(inode))

          return -EACCES;

fs/btrfs/ioctl.c +0 -0

@@ -201,6 +218,22 @@ static int btrfs_ioctl_setflags(struct file
 *file, void __user *arg)
      else
          ip->flags&=3D ~BTRFS_INODE_DIRSYNC;
 

Quoted 4 lines

fs/btrfs/ioctl.c +2 -0

+    /*
+     * The COMPRESS flag can only be changed by users, while the

NOCOMPRESS
+ * flag may be changed automatically if compression code won't =

Quoted 14 lines

fs/btrfs/ioctl.c +13 -0

+     * things smaller.
+     */
+    if (flags&  FS_NOCOMP_FL) {
+        ip->flags&=3D ~BTRFS_INODE_COMPRESS;
+        ip->flags |=3D BTRFS_INODE_NOCOMPRESS;
+    } else if (flags&  FS_COMPR_FL) {
+        ip->flags |=3D BTRFS_INODE_COMPRESS;
+        ip->flags&=3D ~BTRFS_INODE_NOCOMPRESS;
+    }
+    if (flags&  FS_NOCOW_FL)
+        ip->flags |=3D BTRFS_INODE_NODATACOW;
+    else if (flags&  FS_COW_FL)
+        ip->flags&=3D ~BTRFS_INODE_NODATACOW;

Quoted 2 lines

      trans =3D btrfs_join_transaction(root, 1);
      BUG_ON(IS_ERR(trans));

=20
=20

Quoted 4 lines · Konstantinos Skarlatos

Hello,
I would like to ask about the status of this feature/patch, is it accepted into btrfs code, and how can I use it?

Yes, it is now in the latest 2.6.39-rc1.

Quoted 4 lines

I am interested in enabling compression in a specific folder(force-compress would be ideal) of a large btrfs volume, and disabling it for the rest.

hmm, I'm making the tool's patch, and will come soon. :)