Btrfs: transaction ioctls

These ioctls let a user application hold a transaction open while it
performs a series of operations.  A final ioctl does a sync on the fs
(closing the current transaction).  This is the main requirement for
Ceph's OSD to be able to keep the data it's storing in a btrfs volume
consistent, and AFAICS it works just fine.  The application would do
something like

	fd = ::open("some/file", O_RDONLY);
	::ioctl(fd, BTRFS_IOC_TRANS_START);
	/* do a bunch of stuff */
	::ioctl(fd, BTRFS_IOC_TRANS_END);
or just
	::close(fd);

And to ensure it commits to disk,

	::ioctl(fd, BTRFS_IOC_SYNC);

When a transaction is held open, the trans_handle is attached to the
struct file (via private_data) so that it will get cleaned up if the
process dies unexpectedly.  A held transaction is also ended on fsync() to
avoid a deadlock.

A misbehaving application could also deliberately hold a transaction open,
effectively locking up the FS, so it may make sense to restrict something
like this to root or something.

Signed-off-by: Chris Mason <chris.mason@oracle.com>

1 files changed, 63 insertions, 0 deletions

diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 31aa4ba06fce..0f14697becef 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -3336,6 +3336,61 @@ out_fput:
         return ret;
 }
 
+/*
+ * there are many ways the trans_start and trans_end ioctls can lead
+ * to deadlocks.  They should only be used by applications that
+ * basically own the machine, and have a very in depth understanding
+ * of all the possible deadlocks and enospc problems.
+ */
+long btrfs_ioctl_trans_start(struct file *file)
+{
+        struct inode *inode = fdentry(file)->d_inode;
+        struct btrfs_root *root = BTRFS_I(inode)->root;
+        struct btrfs_trans_handle *trans;
+        int ret = 0;
+
+        mutex_lock(&root->fs_info->fs_mutex);
+        if (file->private_data) {
+                ret = -EINPROGRESS;
+                goto out;
+        }
+        trans = btrfs_start_transaction(root, 0);
+        if (trans)
+                file->private_data = trans;
+        else
+                ret = -ENOMEM;
+        /*printk(KERN_INFO "btrfs_ioctl_trans_start on %p\n", file);*/
+out:
+        mutex_unlock(&root->fs_info->fs_mutex);
+        return ret;
+}
+
+/*
+ * there are many ways the trans_start and trans_end ioctls can lead
+ * to deadlocks.  They should only be used by applications that
+ * basically own the machine, and have a very in depth understanding
+ * of all the possible deadlocks and enospc problems.
+ */
+long btrfs_ioctl_trans_end(struct file *file)
+{
+        struct inode *inode = fdentry(file)->d_inode;
+        struct btrfs_root *root = BTRFS_I(inode)->root;
+        struct btrfs_trans_handle *trans;
+        int ret = 0;
+
+        mutex_lock(&root->fs_info->fs_mutex);
+        trans = file->private_data;
+        if (!trans) {
+                ret = -EINVAL;
+                goto out;
+        }
+        btrfs_end_transaction(trans, root);
+        file->private_data = 0;
+out:
+        mutex_unlock(&root->fs_info->fs_mutex);
+        return ret;
+}
+
 long btrfs_ioctl(struct file *file, unsigned int
                 cmd, unsigned long arg)
 {
@@ -3356,6 +3411,13 @@ long btrfs_ioctl(struct file *file, unsigned int
                 return btrfs_balance(root->fs_info->dev_root);
         case BTRFS_IOC_CLONE:
                 return btrfs_ioctl_clone(file, arg);
+        case BTRFS_IOC_TRANS_START:
+                return btrfs_ioctl_trans_start(file);
+        case BTRFS_IOC_TRANS_END:
+                return btrfs_ioctl_trans_end(file);
+        case BTRFS_IOC_SYNC:
+                btrfs_sync_fs(file->f_dentry->d_sb, 1);
+                return 0;
         }
 
         return -ENOTTY;
@@ -3679,6 +3741,7 @@ static struct file_operations btrfs_dir_file_operations = {
 #ifdef CONFIG_COMPAT
         .compat_ioctl   = btrfs_ioctl,
 #endif
+        .release        = btrfs_release_file,
 };
 
 static struct extent_io_ops btrfs_extent_io_ops = {