[PATCH 0/4] vboxsf: Add support for the atomic_open directory-inode op

[PATCH 0/4] vboxsf: Add support for the atomic_open directory-inode op

[Hans de Goede]

Hi All,

Opening a new file is done in 2 steps on regular filesystems:

1. Call the create inode-op on the parent-dir to create an inode
to hold the meta-data related to the file.
2. Call the open file-op to get a handle for the file.

vboxsf however does not really use disk-backed inodes because it
is based on passing through file-related system-calls through to
the hypervisor. So both steps translate to an open(2) call being
passed through to the hypervisor. With the handle returned by
the first call immediately being closed again.

Making 2 open calls for a single open(..., O_CREATE, ...) calls
has 2 problems:

a) It is not really efficient.
b) It actually breaks some apps.

An example of b) is doing a git clone inside a vboxsf mount.
When git clone tries to create a tempfile to store the pak
files which is downloading the following happens:

1. vboxsf_dir_mkfile() gets called with a mode of 0444 and succeeds.
2. vboxsf_file_open() gets called with file->f_flags containing
O_RDWR. When the host is a Linux machine this fails because doing
a open(..., O_RDWR) on a file which exists and has mode 0444 results
in an -EPERM error.

This series fixes this by adding support for the atomic_open
directory-inode op.

Hans de Goede (4):
  vboxsf: Honor excl flag to the dir-inode create op
  vboxsf: Make vboxsf_dir_create() return the handle for the created file
  vboxsf: Add vboxsf_[create|release]_sf_handle() helpers
  vboxsf: Add support for the atomic_open directory-inode op

 fs/vboxsf/dir.c    | 76 +++++++++++++++++++++++++++++++++++++++-------
 fs/vboxsf/file.c   | 71 +++++++++++++++++++++++++++----------------
 fs/vboxsf/vfsmod.h |  7 +++++
 3 files changed, 116 insertions(+), 38 deletions(-)

Regards,

Hans

[PATCH 1/4] vboxsf: Honor excl flag to the dir-inode create op

[Hans de Goede]

Honor the excl flag to the dir-inode create op, instead of behaving
as if it is always set.

Note the old behavior still worked most of the time since a non-exclusive
open only calls the create op, if there is a race and the file is created
between the dentry lookup and the calling of the create call.

While at it change the type of the is_dir parameter to the
vboxsf_dir_create() helper from an int to a bool, to be consistent with
the use of bool for the excl parameter.

Fixes: 0fd169576648 ("fs: Add VirtualBox guest shared folder (vboxsf) support")
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
---
 fs/vboxsf/dir.c | 16 +++++++++-------
 1 file changed, 9 insertions(+), 7 deletions(-)

diff --git a/fs/vboxsf/dir.c b/fs/vboxsf/dir.c
index 4d569f14a8d8..c3e68ad6c0f4 100644
--- a/fs/vboxsf/dir.c
+++ b/fs/vboxsf/dir.c
@@ -253,7 +253,7 @@ static int vboxsf_dir_instantiate(struct inode *parent, struct dentry *dentry,
 }
 
 static int vboxsf_dir_create(struct inode *parent, struct dentry *dentry,
-			     umode_t mode, int is_dir)
+			     umode_t mode, bool is_dir, bool excl)
 {
 	struct vboxsf_inode *sf_parent_i = VBOXSF_I(parent);
 	struct vboxsf_sbi *sbi = VBOXSF_SBI(parent->i_sb);
@@ -261,10 +261,12 @@ static int vboxsf_dir_create(struct inode *parent, struct dentry *dentry,
 	int err;
 
 	params.handle = SHFL_HANDLE_NIL;
-	params.create_flags = SHFL_CF_ACT_CREATE_IF_NEW |
-			      SHFL_CF_ACT_FAIL_IF_EXISTS |
-			      SHFL_CF_ACCESS_READWRITE |
-			      (is_dir ? SHFL_CF_DIRECTORY : 0);
+	params.create_flags = SHFL_CF_ACT_CREATE_IF_NEW | SHFL_CF_ACCESS_READWRITE;
+	if (is_dir)
+		params.create_flags |= SHFL_CF_DIRECTORY;
+	if (excl)
+		params.create_flags |= SHFL_CF_ACT_FAIL_IF_EXISTS;
+
 	params.info.attr.mode = (mode & 0777) |
 				(is_dir ? SHFL_TYPE_DIRECTORY : SHFL_TYPE_FILE);
 	params.info.attr.additional = SHFLFSOBJATTRADD_NOTHING;
@@ -291,13 +293,13 @@ static int vboxsf_dir_create(struct inode *parent, struct dentry *dentry,
 static int vboxsf_dir_mkfile(struct inode *parent, struct dentry *dentry,
 			     umode_t mode, bool excl)
 {
-	return vboxsf_dir_create(parent, dentry, mode, 0);
+	return vboxsf_dir_create(parent, dentry, mode, false, excl);
 }
 
 static int vboxsf_dir_mkdir(struct inode *parent, struct dentry *dentry,
 			    umode_t mode)
 {
-	return vboxsf_dir_create(parent, dentry, mode, 1);
+	return vboxsf_dir_create(parent, dentry, mode, true, true);
 }
 
 static int vboxsf_dir_unlink(struct inode *parent, struct dentry *dentry)
-- 
2.28.0

[PATCH 2/4] vboxsf: Make vboxsf_dir_create() return the handle for the created file

[Hans de Goede]

Make vboxsf_dir_create() optionally return the vboxsf-handle for
the created file. This is a preparation patch for adding atomic_open
support.

Fixes: 0fd169576648 ("fs: Add VirtualBox guest shared folder (vboxsf) support")
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
---
 fs/vboxsf/dir.c | 18 +++++++++++-------
 1 file changed, 11 insertions(+), 7 deletions(-)

diff --git a/fs/vboxsf/dir.c b/fs/vboxsf/dir.c
index c3e68ad6c0f4..0664787f2b74 100644
--- a/fs/vboxsf/dir.c
+++ b/fs/vboxsf/dir.c
@@ -253,7 +253,7 @@ static int vboxsf_dir_instantiate(struct inode *parent, struct dentry *dentry,
 }
 
 static int vboxsf_dir_create(struct inode *parent, struct dentry *dentry,
-			     umode_t mode, bool is_dir, bool excl)
+			     umode_t mode, bool is_dir, bool excl, u64 *handle_ret)
 {
 	struct vboxsf_inode *sf_parent_i = VBOXSF_I(parent);
 	struct vboxsf_sbi *sbi = VBOXSF_SBI(parent->i_sb);
@@ -278,28 +278,32 @@ static int vboxsf_dir_create(struct inode *parent, struct dentry *dentry,
 	if (params.result != SHFL_FILE_CREATED)
 		return -EPERM;
 
-	vboxsf_close(sbi->root, params.handle);
-
 	err = vboxsf_dir_instantiate(parent, dentry, &params.info);
 	if (err)
-		return err;
+		goto out;
 
 	/* parent directory access/change time changed */
 	sf_parent_i->force_restat = 1;
 
-	return 0;
+out:
+	if (err == 0 && handle_ret)
+		*handle_ret = params.handle;
+	else
+		vboxsf_close(sbi->root, params.handle);
+
+	return err;
 }
 
 static int vboxsf_dir_mkfile(struct inode *parent, struct dentry *dentry,
 			     umode_t mode, bool excl)
 {
-	return vboxsf_dir_create(parent, dentry, mode, false, excl);
+	return vboxsf_dir_create(parent, dentry, mode, false, excl, NULL);
 }
 
 static int vboxsf_dir_mkdir(struct inode *parent, struct dentry *dentry,
 			    umode_t mode)
 {
-	return vboxsf_dir_create(parent, dentry, mode, true, true);
+	return vboxsf_dir_create(parent, dentry, mode, true, true, NULL);
 }
 
 static int vboxsf_dir_unlink(struct inode *parent, struct dentry *dentry)
-- 
2.28.0

[PATCH 3/4] vboxsf: Add vboxsf_[create|release]_sf_handle() helpers

[Hans de Goede]

Factor out the code to create / release a struct vboxsf_handle into
2 new helper functions.

This is a preparation patch for adding atomic_open support.

Fixes: 0fd169576648 ("fs: Add VirtualBox guest shared folder (vboxsf) support")
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
---
 fs/vboxsf/file.c   | 71 ++++++++++++++++++++++++++++------------------
 fs/vboxsf/vfsmod.h |  7 +++++
 2 files changed, 51 insertions(+), 27 deletions(-)

diff --git a/fs/vboxsf/file.c b/fs/vboxsf/file.c
index c4ab5996d97a..864c2fad23be 100644
--- a/fs/vboxsf/file.c
+++ b/fs/vboxsf/file.c
@@ -20,17 +20,39 @@ struct vboxsf_handle {
 	struct list_head head;
 };
 
-static int vboxsf_file_open(struct inode *inode, struct file *file)
+struct vboxsf_handle *vboxsf_create_sf_handle(struct inode *inode,
+					      u64 handle, u32 access_flags)
 {
 	struct vboxsf_inode *sf_i = VBOXSF_I(inode);
-	struct shfl_createparms params = {};
 	struct vboxsf_handle *sf_handle;
-	u32 access_flags = 0;
-	int err;
 
 	sf_handle = kmalloc(sizeof(*sf_handle), GFP_KERNEL);
 	if (!sf_handle)
-		return -ENOMEM;
+		return ERR_PTR(-ENOMEM);
+
+	/* the host may have given us different attr then requested */
+	sf_i->force_restat = 1;
+
+	/* init our handle struct and add it to the inode's handles list */
+	sf_handle->handle = handle;
+	sf_handle->root = VBOXSF_SBI(inode->i_sb)->root;
+	sf_handle->access_flags = access_flags;
+	kref_init(&sf_handle->refcount);
+
+	mutex_lock(&sf_i->handle_list_mutex);
+	list_add(&sf_handle->head, &sf_i->handle_list);
+	mutex_unlock(&sf_i->handle_list_mutex);
+
+	return sf_handle;
+}
+
+static int vboxsf_file_open(struct inode *inode, struct file *file)
+{
+	struct vboxsf_sbi *sbi = VBOXSF_SBI(inode->i_sb);
+	struct shfl_createparms params = {};
+	struct vboxsf_handle *sf_handle;
+	u32 access_flags = 0;
+	int err;
 
 	/*
 	 * We check the value of params.handle afterwards to find out if
@@ -83,23 +105,14 @@ static int vboxsf_file_open(struct inode *inode, struct file *file)
 	err = vboxsf_create_at_dentry(file_dentry(file), &params);
 	if (err == 0 && params.handle == SHFL_HANDLE_NIL)
 		err = (params.result == SHFL_FILE_EXISTS) ? -EEXIST : -ENOENT;
-	if (err) {
-		kfree(sf_handle);
+	if (err)
 		return err;
-	}
-
-	/* the host may have given us different attr then requested */
-	sf_i->force_restat = 1;
 
-	/* init our handle struct and add it to the inode's handles list */
-	sf_handle->handle = params.handle;
-	sf_handle->root = VBOXSF_SBI(inode->i_sb)->root;
-	sf_handle->access_flags = access_flags;
-	kref_init(&sf_handle->refcount);
-
-	mutex_lock(&sf_i->handle_list_mutex);
-	list_add(&sf_handle->head, &sf_i->handle_list);
-	mutex_unlock(&sf_i->handle_list_mutex);
+	sf_handle = vboxsf_create_sf_handle(inode, params.handle, access_flags);
+	if (IS_ERR(sf_handle)) {
+		vboxsf_close(sbi->root, params.handle);
+		return PTR_ERR(sf_handle);
+	}
 
 	file->private_data = sf_handle;
 	return 0;
@@ -114,22 +127,26 @@ static void vboxsf_handle_release(struct kref *refcount)
 	kfree(sf_handle);
 }
 
-static int vboxsf_file_release(struct inode *inode, struct file *file)
+void vboxsf_release_sf_handle(struct inode *inode, struct vboxsf_handle *sf_handle)
 {
 	struct vboxsf_inode *sf_i = VBOXSF_I(inode);
-	struct vboxsf_handle *sf_handle = file->private_data;
 
+	mutex_lock(&sf_i->handle_list_mutex);
+	list_del(&sf_handle->head);
+	mutex_unlock(&sf_i->handle_list_mutex);
+
+	kref_put(&sf_handle->refcount, vboxsf_handle_release);
+}
+
+static int vboxsf_file_release(struct inode *inode, struct file *file)
+{
 	/*
 	 * When a file is closed on our (the guest) side, we want any subsequent
 	 * accesses done on the host side to see all changes done from our side.
 	 */
 	filemap_write_and_wait(inode->i_mapping);
 
-	mutex_lock(&sf_i->handle_list_mutex);
-	list_del(&sf_handle->head);
-	mutex_unlock(&sf_i->handle_list_mutex);
-
-	kref_put(&sf_handle->refcount, vboxsf_handle_release);
+	vboxsf_release_sf_handle(inode, file->private_data);
 	return 0;
 }
 
diff --git a/fs/vboxsf/vfsmod.h b/fs/vboxsf/vfsmod.h
index 18f95b00fc33..a4050b166c99 100644
--- a/fs/vboxsf/vfsmod.h
+++ b/fs/vboxsf/vfsmod.h
@@ -18,6 +18,8 @@
 #define VBOXSF_SBI(sb)	((struct vboxsf_sbi *)(sb)->s_fs_info)
 #define VBOXSF_I(i)	container_of(i, struct vboxsf_inode, vfs_inode)
 
+struct vboxsf_handle;
+
 struct vboxsf_options {
 	unsigned long ttl;
 	kuid_t uid;
@@ -80,6 +82,11 @@ extern const struct file_operations vboxsf_reg_fops;
 extern const struct address_space_operations vboxsf_reg_aops;
 extern const struct dentry_operations vboxsf_dentry_ops;
 
+/* from file.c */
+struct vboxsf_handle *vboxsf_create_sf_handle(struct inode *inode,
+					      u64 handle, u32 access_flags);
+void vboxsf_release_sf_handle(struct inode *inode, struct vboxsf_handle *sf_handle);
+
 /* from utils.c */
 struct inode *vboxsf_new_inode(struct super_block *sb);
 void vboxsf_init_inode(struct vboxsf_sbi *sbi, struct inode *inode,
-- 
2.28.0

[PATCH 4/4] vboxsf: Add support for the atomic_open directory-inode op

[Hans de Goede]

Opening a new file is done in 2 steps on regular filesystems:

1. Call the create inode-op on the parent-dir to create an inode
to hold the meta-data related to the file.
2. Call the open file-op to get a handle for the file.

vboxsf however does not really use disk-backed inodes because it
is based on passing through file-related system-calls through to
the hypervisor. So both steps translate to an open(2) call being
passed through to the hypervisor. With the handle returned by
the first call immediately being closed again.

Making 2 open calls for a single open(..., O_CREATE, ...) calls
has 2 problems:

a) It is not really efficient.
b) It actually breaks some apps.

An example of b) is doing a git clone inside a vboxsf mount.
When git clone tries to create a tempfile to store the pak
files which is downloading the following happens:

1. vboxsf_dir_mkfile() gets called with a mode of 0444 and succeeds.
2. vboxsf_file_open() gets called with file->f_flags containing
O_RDWR. When the host is a Linux machine this fails because doing
a open(..., O_RDWR) on a file which exists and has mode 0444 results
in an -EPERM error.

Other network-filesystems and fuse avoid the problem of needing to
pass 2 open() calls to the other side by using the atomic_open
directory-inode op.

This commit fixes git clone not working inside a vboxsf mount,
by adding support for the atomic_open directory-inode op.
As an added bonus this should also make opening new files faster.

The atomic_open implementation is modelled after the atomic_open
implementations from the 9p and fuse code.

Fixes: 0fd169576648 ("fs: Add VirtualBox guest shared folder (vboxsf) support")
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
---
 fs/vboxsf/dir.c | 48 ++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 48 insertions(+)

diff --git a/fs/vboxsf/dir.c b/fs/vboxsf/dir.c
index 0664787f2b74..0d85959be0d5 100644
--- a/fs/vboxsf/dir.c
+++ b/fs/vboxsf/dir.c
@@ -306,6 +306,53 @@ static int vboxsf_dir_mkdir(struct inode *parent, struct dentry *dentry,
 	return vboxsf_dir_create(parent, dentry, mode, true, true, NULL);
 }
 
+static int vboxsf_dir_atomic_open(struct inode *parent, struct dentry *dentry,
+				  struct file *file, unsigned int flags, umode_t mode)
+{
+	struct vboxsf_sbi *sbi = VBOXSF_SBI(parent->i_sb);
+	struct vboxsf_handle *sf_handle;
+	struct dentry *res = NULL;
+	u64 handle;
+	int err;
+
+	if (d_in_lookup(dentry)) {
+		res = vboxsf_dir_lookup(parent, dentry, 0);
+		if (IS_ERR(res))
+			return PTR_ERR(res);
+
+		if (res)
+			dentry = res;
+	}
+
+	/* Only creates */
+	if (!(flags & O_CREAT) || d_really_is_positive(dentry))
+		return finish_no_open(file, res);
+
+	err = vboxsf_dir_create(parent, dentry, mode, false, flags & O_EXCL, &handle);
+	if (err)
+		goto out;
+
+	sf_handle = vboxsf_create_sf_handle(d_inode(dentry), handle, SHFL_CF_ACCESS_READWRITE);
+	if (IS_ERR(sf_handle)) {
+		vboxsf_close(sbi->root, handle);
+		err = PTR_ERR(sf_handle);
+		goto out;
+	}
+
+	err = finish_open(file, dentry, generic_file_open);
+	if (err) {
+		/* This also closes the handle passed to vboxsf_create_sf_handle() */
+		vboxsf_release_sf_handle(d_inode(dentry), sf_handle);
+		goto out;
+	}
+
+	file->private_data = sf_handle;
+	file->f_mode |= FMODE_CREATED;
+out:
+	dput(res);
+	return err;
+}
+
 static int vboxsf_dir_unlink(struct inode *parent, struct dentry *dentry)
 {
 	struct vboxsf_sbi *sbi = VBOXSF_SBI(parent->i_sb);
@@ -424,6 +471,7 @@ const struct inode_operations vboxsf_dir_iops = {
 	.lookup  = vboxsf_dir_lookup,
 	.create  = vboxsf_dir_mkfile,
 	.mkdir   = vboxsf_dir_mkdir,
+	.atomic_open = vboxsf_dir_atomic_open,
 	.rmdir   = vboxsf_dir_unlink,
 	.unlink  = vboxsf_dir_unlink,
 	.rename  = vboxsf_dir_rename,
-- 
2.28.0