Adding upstream version 2:4.20.0+dfsg.upstream/2%4.20.0+dfsg

Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>

1 files changed, 1960 insertions, 0 deletions

diff --git a/source3/modules/vfs_ceph.c b/source3/modules/vfs_ceph.c
new file mode 100644
index 0000000..c9ee541
--- /dev/null
+++ b/source3/modules/vfs_ceph.c
@@ -0,0 +1,1960 @@
+/*
+   Unix SMB/CIFS implementation.
+   Wrap disk only vfs functions to sidestep dodgy compilers.
+   Copyright (C) Tim Potter 1998
+   Copyright (C) Jeremy Allison 2007
+   Copyright (C) Brian Chrisman 2011 <bchrisman@gmail.com>
+   Copyright (C) Richard Sharpe 2011 <realrichardsharpe@gmail.com>
+
+   This program is free software; you can redistribute it and/or modify
+   it under the terms of the GNU General Public License as published by
+   the Free Software Foundation; either version 3 of the License, or
+   (at your option) any later version.
+
+   This program is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+   GNU General Public License for more details.
+
+   You should have received a copy of the GNU General Public License
+   along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+/*
+ * This VFS only works with the libcephfs.so user-space client. It is not needed
+ * if you are using the kernel client or the FUSE client.
+ *
+ * Add the following smb.conf parameter to each share that will be hosted on
+ * Ceph:
+ *
+ *   vfs objects = [any others you need go here] ceph
+ */
+
+#include "includes.h"
+#include "smbd/smbd.h"
+#include "system/filesys.h"
+#include <dirent.h>
+#include <sys/statvfs.h>
+#include "cephfs/libcephfs.h"
+#include "smbprofile.h"
+#include "modules/posixacl_xattr.h"
+#include "lib/util/tevent_unix.h"
+
+#undef DBGC_CLASS
+#define DBGC_CLASS DBGC_VFS
+
+#ifndef LIBCEPHFS_VERSION
+#define LIBCEPHFS_VERSION(maj, min, extra) ((maj << 16) + (min << 8) + extra)
+#define LIBCEPHFS_VERSION_CODE LIBCEPHFS_VERSION(0, 0, 0)
+#endif
+
+/*
+ * Use %llu whenever we have a 64bit unsigned int, and cast to (long long unsigned)
+ */
+#define llu(_var) ((long long unsigned)_var)
+
+/*
+ * Note, libcephfs's return code model is to return -errno! So we have to
+ * convert to what Samba expects, which is to set errno to -return and return -1
+ */
+#define WRAP_RETURN(_res) \
+	errno = 0; \
+	if (_res < 0) { \
+		errno = -_res; \
+		return -1; \
+	} \
+	return _res \
+
+/*
+ * Track unique connections, as virtual mounts, to cephfs file systems.
+ * Individual mounts will be set on the handle->data attribute, but
+ * the mounts themselves will be shared so as not to spawn extra mounts
+ * to the same cephfs.
+ *
+ * Individual mounts are IDed by a 'cookie' value that is a string built
+ * from identifying parameters found in smb.conf.
+ */
+
+static struct cephmount_cached {
+	char *cookie;
+	uint32_t count;
+	struct ceph_mount_info *mount;
+	struct cephmount_cached *next, *prev;
+} *cephmount_cached;
+
+static int cephmount_cache_add(const char *cookie,
+			       struct ceph_mount_info *mount)
+{
+	struct cephmount_cached *entry = NULL;
+
+	entry = talloc_zero(NULL, struct cephmount_cached);
+	if (entry == NULL) {
+		errno = ENOMEM;
+		return -1;
+	}
+
+	entry->cookie = talloc_strdup(entry, cookie);
+	if (entry->cookie == NULL) {
+		talloc_free(entry);
+		errno = ENOMEM;
+		return -1;
+	}
+
+	entry->mount = mount;
+	entry->count = 1;
+
+	DBG_DEBUG("adding mount cache entry for %s\n", entry->cookie);
+	DLIST_ADD(cephmount_cached, entry);
+	return 0;
+}
+
+static struct ceph_mount_info *cephmount_cache_update(const char *cookie)
+{
+	struct cephmount_cached *entry = NULL;
+
+	for (entry = cephmount_cached; entry; entry = entry->next) {
+		if (strcmp(entry->cookie, cookie) == 0) {
+			entry->count++;
+			DBG_DEBUG("updated mount cache: count is [%"
+				  PRIu32 "]\n", entry->count);
+			return entry->mount;
+		}
+	}
+
+	errno = ENOENT;
+	return NULL;
+}
+
+static int cephmount_cache_remove(struct ceph_mount_info *mount)
+{
+	struct cephmount_cached *entry = NULL;
+
+	for (entry = cephmount_cached; entry; entry = entry->next) {
+		if (entry->mount == mount) {
+			if (--entry->count) {
+				DBG_DEBUG("updated mount cache: count is [%"
+					  PRIu32 "]\n", entry->count);
+				return entry->count;
+			}
+
+			DBG_DEBUG("removing mount cache entry for %s\n",
+				  entry->cookie);
+			DLIST_REMOVE(cephmount_cached, entry);
+			talloc_free(entry);
+			return 0;
+		}
+	}
+	errno = ENOENT;
+	return -1;
+}
+
+static char *cephmount_get_cookie(TALLOC_CTX * mem_ctx, const int snum)
+{
+	const char *conf_file =
+	    lp_parm_const_string(snum, "ceph", "config_file", ".");
+	const char *user_id = lp_parm_const_string(snum, "ceph", "user_id", "");
+	const char *fsname =
+	    lp_parm_const_string(snum, "ceph", "filesystem", "");
+	return talloc_asprintf(mem_ctx, "(%s/%s/%s)", conf_file, user_id,
+			       fsname);
+}
+
+static int cephmount_select_fs(struct ceph_mount_info *mnt, const char *fsname)
+{
+	/*
+	 * ceph_select_filesystem was added in ceph 'nautilus' (v14).
+	 * Earlier versions of libcephfs will lack that API function.
+	 * At the time of this writing (Feb 2023) all versions of ceph
+	 * supported by ceph upstream have this function.
+	 */
+#if defined(HAVE_CEPH_SELECT_FILESYSTEM)
+	DBG_DEBUG("[CEPH] calling: ceph_select_filesystem with %s\n", fsname);
+	return ceph_select_filesystem(mnt, fsname);
+#else
+	DBG_ERR("[CEPH] ceph_select_filesystem not available\n");
+	return -ENOTSUP;
+#endif
+}
+
+static struct ceph_mount_info *cephmount_mount_fs(const int snum)
+{
+	int ret;
+	char buf[256];
+	struct ceph_mount_info *mnt = NULL;
+	/* if config_file and/or user_id are NULL, ceph will use defaults */
+	const char *conf_file =
+	    lp_parm_const_string(snum, "ceph", "config_file", NULL);
+	const char *user_id =
+	    lp_parm_const_string(snum, "ceph", "user_id", NULL);
+	const char *fsname =
+	    lp_parm_const_string(snum, "ceph", "filesystem", NULL);
+
+	DBG_DEBUG("[CEPH] calling: ceph_create\n");
+	ret = ceph_create(&mnt, user_id);
+	if (ret) {
+		errno = -ret;
+		return NULL;
+	}
+
+	DBG_DEBUG("[CEPH] calling: ceph_conf_read_file with %s\n",
+		  (conf_file == NULL ? "default path" : conf_file));
+	ret = ceph_conf_read_file(mnt, conf_file);
+	if (ret) {
+		goto err_cm_release;
+	}
+
+	DBG_DEBUG("[CEPH] calling: ceph_conf_get\n");
+	ret = ceph_conf_get(mnt, "log file", buf, sizeof(buf));
+	if (ret < 0) {
+		goto err_cm_release;
+	}
+
+	/* libcephfs disables POSIX ACL support by default, enable it... */
+	ret = ceph_conf_set(mnt, "client_acl_type", "posix_acl");
+	if (ret < 0) {
+		goto err_cm_release;
+	}
+	/* tell libcephfs to perform local permission checks */
+	ret = ceph_conf_set(mnt, "fuse_default_permissions", "false");
+	if (ret < 0) {
+		goto err_cm_release;
+	}
+	/*
+	 * select a cephfs file system to use:
+	 * In ceph, multiple file system support has been stable since 'pacific'.
+	 * Permit different shares to access different file systems.
+	 */
+	if (fsname != NULL) {
+		ret = cephmount_select_fs(mnt, fsname);
+		if (ret < 0) {
+			goto err_cm_release;
+		}
+	}
+
+	DBG_DEBUG("[CEPH] calling: ceph_mount\n");
+	ret = ceph_mount(mnt, NULL);
+	if (ret >= 0) {
+		goto cm_done;
+	}
+
+      err_cm_release:
+	ceph_release(mnt);
+	mnt = NULL;
+	DBG_DEBUG("[CEPH] Error mounting fs: %s\n", strerror(-ret));
+      cm_done:
+	/*
+	 * Handle the error correctly. Ceph returns -errno.
+	 */
+	if (ret) {
+		errno = -ret;
+	}
+	return mnt;
+}
+
+/* Check for NULL pointer parameters in cephwrap_* functions */
+
+/* We don't want to have NULL function pointers lying around.  Someone
+   is sure to try and execute them.  These stubs are used to prevent
+   this possibility. */
+
+static int cephwrap_connect(struct vfs_handle_struct *handle,
+			    const char *service, const char *user)
+{
+	int ret = 0;
+	struct ceph_mount_info *cmount = NULL;
+	int snum = SNUM(handle->conn);
+	char *cookie = cephmount_get_cookie(handle, snum);
+	if (cookie == NULL) {
+		return -1;
+	}
+
+	cmount = cephmount_cache_update(cookie);
+	if (cmount != NULL) {
+		goto connect_ok;
+	}
+
+	cmount = cephmount_mount_fs(snum);
+	if (cmount == NULL) {
+		ret = -1;
+		goto connect_fail;
+	}
+	ret = cephmount_cache_add(cookie, cmount);
+	if (ret) {
+		goto connect_fail;
+	}
+
+      connect_ok:
+	handle->data = cmount;
+	DBG_WARNING("Connection established with the server: %s\n", cookie);
+	/*
+	 * Unless we have an async implementation of getxattrat turn this off.
+	 */
+	lp_do_parameter(SNUM(handle->conn), "smbd async dosmode", "false");
+      connect_fail:
+	talloc_free(cookie);
+	return ret;
+}
+
+static void cephwrap_disconnect(struct vfs_handle_struct *handle)
+{
+	int ret = cephmount_cache_remove(handle->data);
+	if (ret < 0) {
+		DBG_ERR("failed to remove ceph mount from cache: %s\n",
+			strerror(errno));
+		return;
+	}
+	if (ret > 0) {
+		DBG_DEBUG("mount cache entry still in use\n");
+		return;
+	}
+
+	ret = ceph_unmount(handle->data);
+	if (ret < 0) {
+		DBG_ERR("[CEPH] failed to unmount: %s\n", strerror(-ret));
+	}
+
+	ret = ceph_release(handle->data);
+	if (ret < 0) {
+		DBG_ERR("[CEPH] failed to release: %s\n", strerror(-ret));
+	}
+	handle->data = NULL;
+}
+
+/* Disk operations */
+
+static uint64_t cephwrap_disk_free(struct vfs_handle_struct *handle,
+				const struct smb_filename *smb_fname,
+				uint64_t *bsize,
+				uint64_t *dfree,
+				uint64_t *dsize)
+{
+	struct statvfs statvfs_buf = { 0 };
+	int ret;
+
+	if (!(ret = ceph_statfs(handle->data, smb_fname->base_name,
+			&statvfs_buf))) {
+		/*
+		 * Provide all the correct values.
+		 */
+		*bsize = statvfs_buf.f_bsize;
+		*dfree = statvfs_buf.f_bavail;
+		*dsize = statvfs_buf.f_blocks;
+		DBG_DEBUG("[CEPH] bsize: %llu, dfree: %llu, dsize: %llu\n",
+			llu(*bsize), llu(*dfree), llu(*dsize));
+		return *dfree;
+	} else {
+		DBG_DEBUG("[CEPH] ceph_statfs returned %d\n", ret);
+		WRAP_RETURN(ret);
+	}
+}
+
+static int cephwrap_get_quota(struct vfs_handle_struct *handle,
+				const struct smb_filename *smb_fname,
+				enum SMB_QUOTA_TYPE qtype,
+				unid_t id,
+				SMB_DISK_QUOTA *qt)
+{
+	/* libcephfs: Ceph does not implement this */
+#if 0
+/* was ifdef HAVE_SYS_QUOTAS */
+	int ret;
+
+	ret = ceph_get_quota(handle->conn->connectpath, qtype, id, qt);
+
+	if (ret) {
+		errno = -ret;
+		ret = -1;
+	}
+
+	return ret;
+#else
+	errno = ENOSYS;
+	return -1;
+#endif
+}
+
+static int cephwrap_set_quota(struct vfs_handle_struct *handle,  enum SMB_QUOTA_TYPE qtype, unid_t id, SMB_DISK_QUOTA *qt)
+{
+	/* libcephfs: Ceph does not implement this */
+#if 0
+/* was ifdef HAVE_SYS_QUOTAS */
+	int ret;
+
+	ret = ceph_set_quota(handle->conn->connectpath, qtype, id, qt);
+	if (ret) {
+		errno = -ret;
+		ret = -1;
+	}
+
+	return ret;
+#else
+	WRAP_RETURN(-ENOSYS);
+#endif
+}
+
+static int cephwrap_statvfs(struct vfs_handle_struct *handle,
+			    const struct smb_filename *smb_fname,
+			    struct vfs_statvfs_struct *statbuf)
+{
+	struct statvfs statvfs_buf = { 0 };
+	int ret;
+
+	ret = ceph_statfs(handle->data, smb_fname->base_name, &statvfs_buf);
+	if (ret < 0) {
+		WRAP_RETURN(ret);
+	}
+
+	statbuf->OptimalTransferSize = statvfs_buf.f_frsize;
+	statbuf->BlockSize = statvfs_buf.f_bsize;
+	statbuf->TotalBlocks = statvfs_buf.f_blocks;
+	statbuf->BlocksAvail = statvfs_buf.f_bfree;
+	statbuf->UserBlocksAvail = statvfs_buf.f_bavail;
+	statbuf->TotalFileNodes = statvfs_buf.f_files;
+	statbuf->FreeFileNodes = statvfs_buf.f_ffree;
+	statbuf->FsIdentifier = statvfs_buf.f_fsid;
+	DBG_DEBUG("[CEPH] f_bsize: %ld, f_blocks: %ld, f_bfree: %ld, f_bavail: %ld\n",
+		(long int)statvfs_buf.f_bsize, (long int)statvfs_buf.f_blocks,
+		(long int)statvfs_buf.f_bfree, (long int)statvfs_buf.f_bavail);
+
+	return ret;
+}
+
+static uint32_t cephwrap_fs_capabilities(struct vfs_handle_struct *handle,
+					 enum timestamp_set_resolution *p_ts_res)
+{
+	uint32_t caps = FILE_CASE_SENSITIVE_SEARCH | FILE_CASE_PRESERVED_NAMES;
+
+	*p_ts_res = TIMESTAMP_SET_NT_OR_BETTER;
+
+	return caps;
+}
+
+/* Directory operations */
+
+static DIR *cephwrap_fdopendir(struct vfs_handle_struct *handle,
+			       struct files_struct *fsp,
+			       const char *mask,
+			       uint32_t attributes)
+{
+	int ret = 0;
+	struct ceph_dir_result *result = NULL;
+
+#ifdef HAVE_CEPH_FDOPENDIR
+	int dirfd = fsp_get_io_fd(fsp);
+	DBG_DEBUG("[CEPH] fdopendir(%p, %d)\n", handle, dirfd);
+	ret = ceph_fdopendir(handle->data, dirfd, &result);
+#else
+	DBG_DEBUG("[CEPH] fdopendir(%p, %p)\n", handle, fsp);
+	ret = ceph_opendir(handle->data, fsp->fsp_name->base_name, &result);
+#endif
+	if (ret < 0) {
+		result = NULL;
+		errno = -ret; /* We return result which is NULL in this case */
+	}
+
+	DBG_DEBUG("[CEPH] fdopendir(...) = %d\n", ret);
+	return (DIR *) result;
+}
+
+static struct dirent *cephwrap_readdir(struct vfs_handle_struct *handle,
+				       struct files_struct *dirfsp,
+				       DIR *dirp)
+{
+	struct dirent *result = NULL;
+
+	DBG_DEBUG("[CEPH] readdir(%p, %p)\n", handle, dirp);
+	result = ceph_readdir(handle->data, (struct ceph_dir_result *) dirp);
+	DBG_DEBUG("[CEPH] readdir(...) = %p\n", result);
+
+	return result;
+}
+
+static void cephwrap_rewinddir(struct vfs_handle_struct *handle, DIR *dirp)
+{
+	DBG_DEBUG("[CEPH] rewinddir(%p, %p)\n", handle, dirp);
+	ceph_rewinddir(handle->data, (struct ceph_dir_result *) dirp);
+}
+
+static int cephwrap_mkdirat(struct vfs_handle_struct *handle,
+			files_struct *dirfsp,
+			const struct smb_filename *smb_fname,
+			mode_t mode)
+{
+	int result = -1;
+#ifdef HAVE_CEPH_MKDIRAT
+	int dirfd = fsp_get_pathref_fd(dirfsp);
+
+	DBG_DEBUG("[CEPH] mkdirat(%p, %d, %s)\n",
+		  handle,
+		  dirfd,
+		  smb_fname->base_name);
+
+	result = ceph_mkdirat(handle->data, dirfd, smb_fname->base_name, mode);
+
+	DBG_DEBUG("[CEPH] mkdirat(...) = %d\n", result);
+
+	WRAP_RETURN(result);
+#else
+	struct smb_filename *full_fname = NULL;
+
+	full_fname = full_path_from_dirfsp_atname(talloc_tos(),
+						dirfsp,
+						smb_fname);
+	if (full_fname == NULL) {
+		return -1;
+	}
+
+	DBG_DEBUG("[CEPH] mkdir(%p, %s)\n",
+		  handle, smb_fname_str_dbg(full_fname));
+
+	result = ceph_mkdir(handle->data, full_fname->base_name, mode);
+
+	TALLOC_FREE(full_fname);
+
+	WRAP_RETURN(result);
+#endif
+}
+
+static int cephwrap_closedir(struct vfs_handle_struct *handle, DIR *dirp)
+{
+	int result;
+
+	DBG_DEBUG("[CEPH] closedir(%p, %p)\n", handle, dirp);
+	result = ceph_closedir(handle->data, (struct ceph_dir_result *) dirp);
+	DBG_DEBUG("[CEPH] closedir(...) = %d\n", result);
+	WRAP_RETURN(result);
+}
+
+/* File operations */
+
+static int cephwrap_openat(struct vfs_handle_struct *handle,
+			   const struct files_struct *dirfsp,
+			   const struct smb_filename *smb_fname,
+			   files_struct *fsp,
+			   const struct vfs_open_how *how)
+{
+	int flags = how->flags;
+	mode_t mode = how->mode;
+	struct smb_filename *name = NULL;
+	bool have_opath = false;
+	bool became_root = false;
+	int result = -ENOENT;
+#ifdef HAVE_CEPH_OPENAT
+	int dirfd = -1;
+#endif
+
+	if (how->resolve != 0) {
+		errno = ENOSYS;
+		return -1;
+	}
+
+	if (smb_fname->stream_name) {
+		goto out;
+	}
+
+#ifdef O_PATH
+	have_opath = true;
+	if (fsp->fsp_flags.is_pathref) {
+		flags |= O_PATH;
+	}
+#endif
+
+#ifdef HAVE_CEPH_OPENAT
+	dirfd = fsp_get_pathref_fd(dirfsp);
+
+	DBG_DEBUG("[CEPH] openat(%p, %d, %p, %d, %d)\n",
+		  handle, dirfd, fsp, flags, mode);
+
+	if (fsp->fsp_flags.is_pathref && !have_opath) {
+		become_root();
+		became_root = true;
+	}
+
+	result = ceph_openat(handle->data,
+			     dirfd,
+			     smb_fname->base_name,
+			     flags,
+			     mode);
+
+#else
+	if (fsp_get_pathref_fd(dirfsp) != AT_FDCWD) {
+		name = full_path_from_dirfsp_atname(talloc_tos(),
+						    dirfsp,
+						    smb_fname);
+		if (name == NULL) {
+			return -1;
+		}
+		smb_fname = name;
+	}
+
+	DBG_DEBUG("[CEPH] openat(%p, %s, %p, %d, %d)\n", handle,
+		  smb_fname_str_dbg(smb_fname), fsp, flags, mode);
+
+	if (fsp->fsp_flags.is_pathref && !have_opath) {
+		become_root();
+		became_root = true;
+	}
+
+	result = ceph_open(handle->data, smb_fname->base_name, flags, mode);
+#endif
+	if (became_root) {
+		unbecome_root();
+	}
+out:
+	TALLOC_FREE(name);
+	fsp->fsp_flags.have_proc_fds = false;
+	DBG_DEBUG("[CEPH] open(...) = %d\n", result);
+	WRAP_RETURN(result);
+}
+
+static int cephwrap_close(struct vfs_handle_struct *handle, files_struct *fsp)
+{
+	int result;
+
+	DBG_DEBUG("[CEPH] close(%p, %p)\n", handle, fsp);
+	result = ceph_close(handle->data, fsp_get_pathref_fd(fsp));
+	DBG_DEBUG("[CEPH] close(...) = %d\n", result);
+
+	WRAP_RETURN(result);
+}
+
+static ssize_t cephwrap_pread(struct vfs_handle_struct *handle, files_struct *fsp, void *data,
+			size_t n, off_t offset)
+{
+	ssize_t result;
+
+	DBG_DEBUG("[CEPH] pread(%p, %p, %p, %llu, %llu)\n", handle, fsp, data, llu(n), llu(offset));
+
+	result = ceph_read(handle->data, fsp_get_io_fd(fsp), data, n, offset);
+	DBG_DEBUG("[CEPH] pread(...) = %llu\n", llu(result));
+	WRAP_RETURN(result);
+}
+
+struct cephwrap_pread_state {
+	ssize_t bytes_read;
+	struct vfs_aio_state vfs_aio_state;
+};
+
+/*
+ * Fake up an async ceph read by calling the synchronous API.
+ */
+static struct tevent_req *cephwrap_pread_send(struct vfs_handle_struct *handle,
+					      TALLOC_CTX *mem_ctx,
+					      struct tevent_context *ev,
+					      struct files_struct *fsp,
+					      void *data,
+					      size_t n, off_t offset)
+{
+	struct tevent_req *req = NULL;
+	struct cephwrap_pread_state *state = NULL;
+	int ret = -1;
+
+	DBG_DEBUG("[CEPH] %s\n", __func__);
+	req = tevent_req_create(mem_ctx, &state, struct cephwrap_pread_state);
+	if (req == NULL) {
+		return NULL;
+	}
+
+	ret = ceph_read(handle->data, fsp_get_io_fd(fsp), data, n, offset);
+	if (ret < 0) {
+		/* ceph returns -errno on error. */
+		tevent_req_error(req, -ret);
+		return tevent_req_post(req, ev);
+	}
+
+	state->bytes_read = ret;
+	tevent_req_done(req);
+	/* Return and schedule the completion of the call. */
+	return tevent_req_post(req, ev);
+}
+
+static ssize_t cephwrap_pread_recv(struct tevent_req *req,
+				   struct vfs_aio_state *vfs_aio_state)
+{
+	struct cephwrap_pread_state *state =
+		tevent_req_data(req, struct cephwrap_pread_state);
+
+	DBG_DEBUG("[CEPH] %s\n", __func__);
+	if (tevent_req_is_unix_error(req, &vfs_aio_state->error)) {
+		return -1;
+	}
+	*vfs_aio_state = state->vfs_aio_state;
+	return state->bytes_read;
+}
+
+static ssize_t cephwrap_pwrite(struct vfs_handle_struct *handle, files_struct *fsp, const void *data,
+			size_t n, off_t offset)
+{
+	ssize_t result;
+
+	DBG_DEBUG("[CEPH] pwrite(%p, %p, %p, %llu, %llu)\n", handle, fsp, data, llu(n), llu(offset));
+	result = ceph_write(handle->data, fsp_get_io_fd(fsp), data, n, offset);
+	DBG_DEBUG("[CEPH] pwrite(...) = %llu\n", llu(result));
+	WRAP_RETURN(result);
+}
+
+struct cephwrap_pwrite_state {
+	ssize_t bytes_written;
+	struct vfs_aio_state vfs_aio_state;
+};
+
+/*
+ * Fake up an async ceph write by calling the synchronous API.
+ */
+static struct tevent_req *cephwrap_pwrite_send(struct vfs_handle_struct *handle,
+					       TALLOC_CTX *mem_ctx,
+					       struct tevent_context *ev,
+					       struct files_struct *fsp,
+					       const void *data,
+					       size_t n, off_t offset)
+{
+	struct tevent_req *req = NULL;
+	struct cephwrap_pwrite_state *state = NULL;
+	int ret = -1;
+
+	DBG_DEBUG("[CEPH] %s\n", __func__);
+	req = tevent_req_create(mem_ctx, &state, struct cephwrap_pwrite_state);
+	if (req == NULL) {
+		return NULL;
+	}
+
+	ret = ceph_write(handle->data, fsp_get_io_fd(fsp), data, n, offset);
+	if (ret < 0) {
+		/* ceph returns -errno on error. */
+		tevent_req_error(req, -ret);
+		return tevent_req_post(req, ev);
+	}
+
+	state->bytes_written = ret;
+	tevent_req_done(req);
+	/* Return and schedule the completion of the call. */
+	return tevent_req_post(req, ev);
+}
+
+static ssize_t cephwrap_pwrite_recv(struct tevent_req *req,
+				    struct vfs_aio_state *vfs_aio_state)
+{
+	struct cephwrap_pwrite_state *state =
+		tevent_req_data(req, struct cephwrap_pwrite_state);
+
+	DBG_DEBUG("[CEPH] %s\n", __func__);
+	if (tevent_req_is_unix_error(req, &vfs_aio_state->error)) {
+		return -1;
+	}
+	*vfs_aio_state = state->vfs_aio_state;
+	return state->bytes_written;
+}
+
+static off_t cephwrap_lseek(struct vfs_handle_struct *handle, files_struct *fsp, off_t offset, int whence)
+{
+	off_t result = 0;
+
+	DBG_DEBUG("[CEPH] cephwrap_lseek\n");
+	result = ceph_lseek(handle->data, fsp_get_io_fd(fsp), offset, whence);
+	WRAP_RETURN(result);
+}
+
+static ssize_t cephwrap_sendfile(struct vfs_handle_struct *handle, int tofd, files_struct *fromfsp, const DATA_BLOB *hdr,
+			off_t offset, size_t n)
+{
+	/*
+	 * We cannot support sendfile because libcephfs is in user space.
+	 */
+	DBG_DEBUG("[CEPH] cephwrap_sendfile\n");
+	errno = ENOTSUP;
+	return -1;
+}
+
+static ssize_t cephwrap_recvfile(struct vfs_handle_struct *handle,
+			int fromfd,
+			files_struct *tofsp,
+			off_t offset,
+			size_t n)
+{
+	/*
+	 * We cannot support recvfile because libcephfs is in user space.
+	 */
+	DBG_DEBUG("[CEPH] cephwrap_recvfile\n");
+	errno=ENOTSUP;
+	return -1;
+}
+
+static int cephwrap_renameat(struct vfs_handle_struct *handle,
+			files_struct *srcfsp,
+			const struct smb_filename *smb_fname_src,
+			files_struct *dstfsp,
+			const struct smb_filename *smb_fname_dst)
+{
+	struct smb_filename *full_fname_src = NULL;
+	struct smb_filename *full_fname_dst = NULL;
+	int result = -1;
+
+	DBG_DEBUG("[CEPH] cephwrap_renameat\n");
+	if (smb_fname_src->stream_name || smb_fname_dst->stream_name) {
+		errno = ENOENT;
+		return result;
+	}
+
+	full_fname_src = full_path_from_dirfsp_atname(talloc_tos(),
+						  srcfsp,
+						  smb_fname_src);
+	if (full_fname_src == NULL) {
+		errno = ENOMEM;
+		return -1;
+	}
+	full_fname_dst = full_path_from_dirfsp_atname(talloc_tos(),
+						  dstfsp,
+						  smb_fname_dst);
+	if (full_fname_dst == NULL) {
+		TALLOC_FREE(full_fname_src);
+		errno = ENOMEM;
+		return -1;
+	}
+
+	result = ceph_rename(handle->data,
+			     full_fname_src->base_name,
+			     full_fname_dst->base_name);
+
+	TALLOC_FREE(full_fname_src);
+	TALLOC_FREE(full_fname_dst);
+
+	WRAP_RETURN(result);
+}
+
+/*
+ * Fake up an async ceph fsync by calling the synchronous API.
+ */
+
+static struct tevent_req *cephwrap_fsync_send(struct vfs_handle_struct *handle,
+					TALLOC_CTX *mem_ctx,
+					struct tevent_context *ev,
+					files_struct *fsp)
+{
+	struct tevent_req *req = NULL;
+	struct vfs_aio_state *state = NULL;
+	int ret = -1;
+
+	DBG_DEBUG("[CEPH] cephwrap_fsync_send\n");
+
+	req = tevent_req_create(mem_ctx, &state, struct vfs_aio_state);
+	if (req == NULL) {
+		return NULL;
+	}
+
+	/* Make sync call. */
+	ret = ceph_fsync(handle->data, fsp_get_io_fd(fsp), false);
+
+	if (ret != 0) {
+		/* ceph_fsync returns -errno on error. */
+		tevent_req_error(req, -ret);
+		return tevent_req_post(req, ev);
+	}
+
+	/* Mark it as done. */
+	tevent_req_done(req);
+	/* Return and schedule the completion of the call. */
+	return tevent_req_post(req, ev);
+}
+
+static int cephwrap_fsync_recv(struct tevent_req *req,
+				struct vfs_aio_state *vfs_aio_state)
+{
+	struct vfs_aio_state *state =
+		tevent_req_data(req, struct vfs_aio_state);
+
+	DBG_DEBUG("[CEPH] cephwrap_fsync_recv\n");
+
+	if (tevent_req_is_unix_error(req, &vfs_aio_state->error)) {
+		return -1;
+	}
+	*vfs_aio_state = *state;
+	return 0;
+}
+
+#define SAMBA_STATX_ATTR_MASK	(CEPH_STATX_BASIC_STATS|CEPH_STATX_BTIME)
+
+static void init_stat_ex_from_ceph_statx(struct stat_ex *dst, const struct ceph_statx *stx)
+{
+	DBG_DEBUG("[CEPH]\tstx = {dev = %llx, ino = %llu, mode = 0x%x, "
+		  "nlink = %llu, uid = %d, gid = %d, rdev = %llx, size = %llu, "
+		  "blksize = %llu, blocks = %llu, atime = %llu, mtime = %llu, "
+		  "ctime = %llu, btime = %llu}\n",
+		  llu(stx->stx_dev), llu(stx->stx_ino), stx->stx_mode,
+		  llu(stx->stx_nlink), stx->stx_uid, stx->stx_gid,
+		  llu(stx->stx_rdev), llu(stx->stx_size), llu(stx->stx_blksize),
+		  llu(stx->stx_blocks), llu(stx->stx_atime.tv_sec),
+		  llu(stx->stx_mtime.tv_sec), llu(stx->stx_ctime.tv_sec),
+		  llu(stx->stx_btime.tv_sec));
+
+	if ((stx->stx_mask & SAMBA_STATX_ATTR_MASK) != SAMBA_STATX_ATTR_MASK) {
+		DBG_WARNING("%s: stx->stx_mask is incorrect (wanted %x, got %x)\n",
+				__func__, SAMBA_STATX_ATTR_MASK, stx->stx_mask);
+	}
+
+	dst->st_ex_dev = stx->stx_dev;
+	dst->st_ex_rdev = stx->stx_rdev;
+	dst->st_ex_ino = stx->stx_ino;
+	dst->st_ex_mode = stx->stx_mode;
+	dst->st_ex_uid = stx->stx_uid;
+	dst->st_ex_gid = stx->stx_gid;
+	dst->st_ex_size = stx->stx_size;
+	dst->st_ex_nlink = stx->stx_nlink;
+	dst->st_ex_atime = stx->stx_atime;
+	dst->st_ex_btime = stx->stx_btime;
+	dst->st_ex_ctime = stx->stx_ctime;
+	dst->st_ex_mtime = stx->stx_mtime;
+	dst->st_ex_blksize = stx->stx_blksize;
+	dst->st_ex_blocks = stx->stx_blocks;
+}
+
+static int cephwrap_stat(struct vfs_handle_struct *handle,
+			struct smb_filename *smb_fname)
+{
+	int result = -1;
+	struct ceph_statx stx = { 0 };
+
+	DBG_DEBUG("[CEPH] stat(%p, %s)\n", handle, smb_fname_str_dbg(smb_fname));
+
+	if (smb_fname->stream_name) {
+		errno = ENOENT;
+		return result;
+	}
+
+	result = ceph_statx(handle->data, smb_fname->base_name, &stx,
+				SAMBA_STATX_ATTR_MASK, 0);
+	DBG_DEBUG("[CEPH] statx(...) = %d\n", result);
+	if (result < 0) {
+		WRAP_RETURN(result);
+	}
+
+	init_stat_ex_from_ceph_statx(&smb_fname->st, &stx);
+	DBG_DEBUG("[CEPH] mode = 0x%x\n", smb_fname->st.st_ex_mode);
+	return result;
+}
+
+static int cephwrap_fstat(struct vfs_handle_struct *handle, files_struct *fsp, SMB_STRUCT_STAT *sbuf)
+{
+	int result = -1;
+	struct ceph_statx stx = { 0 };
+	int fd = fsp_get_pathref_fd(fsp);
+
+	DBG_DEBUG("[CEPH] fstat(%p, %d)\n", handle, fd);
+	result = ceph_fstatx(handle->data, fd, &stx,
+				SAMBA_STATX_ATTR_MASK, 0);
+	DBG_DEBUG("[CEPH] fstat(...) = %d\n", result);
+	if (result < 0) {
+		WRAP_RETURN(result);
+	}
+
+	init_stat_ex_from_ceph_statx(sbuf, &stx);
+	DBG_DEBUG("[CEPH] mode = 0x%x\n", sbuf->st_ex_mode);
+	return result;
+}
+
+static int cephwrap_fstatat(struct vfs_handle_struct *handle,
+			    const struct files_struct *dirfsp,
+			    const struct smb_filename *smb_fname,
+			    SMB_STRUCT_STAT *sbuf,
+			    int flags)
+{
+	int result = -1;
+	struct ceph_statx stx = { 0 };
+#ifdef HAVE_CEPH_STATXAT
+	int dirfd = fsp_get_pathref_fd(dirfsp);
+
+	DBG_DEBUG("[CEPH] fstatat(%p, %d, %s)\n",
+		  handle, dirfd, smb_fname->base_name);
+	result = ceph_statxat(handle->data, dirfd, smb_fname->base_name,
+			      &stx, SAMBA_STATX_ATTR_MASK, 0);
+#else
+	struct smb_filename *full_fname = NULL;
+
+	full_fname = full_path_from_dirfsp_atname(talloc_tos(),
+						  dirfsp,
+						  smb_fname);
+	if (full_fname == NULL) {
+		errno = ENOMEM;
+		return -1;
+	}
+
+	DBG_DEBUG("[CEPH] fstatat(%p, %s)\n",
+		  handle, smb_fname_str_dbg(full_fname));
+	result = ceph_statx(handle->data, full_fname->base_name,
+			    &stx, SAMBA_STATX_ATTR_MASK, 0);
+
+	TALLOC_FREE(full_fname);
+#endif
+
+	DBG_DEBUG("[CEPH] fstatat(...) = %d\n", result);
+	if (result < 0) {
+		WRAP_RETURN(result);
+	}
+
+	init_stat_ex_from_ceph_statx(sbuf, &stx);
+	DBG_DEBUG("[CEPH] mode = 0x%x\n", sbuf->st_ex_mode);
+
+	return 0;
+}
+
+static int cephwrap_lstat(struct vfs_handle_struct *handle,
+			 struct smb_filename *smb_fname)
+{
+	int result = -1;
+	struct ceph_statx stx = { 0 };
+
+	DBG_DEBUG("[CEPH] lstat(%p, %s)\n", handle, smb_fname_str_dbg(smb_fname));
+
+	if (smb_fname->stream_name) {
+		errno = ENOENT;
+		return result;
+	}
+
+	result = ceph_statx(handle->data, smb_fname->base_name, &stx,
+				SAMBA_STATX_ATTR_MASK, AT_SYMLINK_NOFOLLOW);
+	DBG_DEBUG("[CEPH] lstat(...) = %d\n", result);
+	if (result < 0) {
+		WRAP_RETURN(result);
+	}
+
+	init_stat_ex_from_ceph_statx(&smb_fname->st, &stx);
+	return result;
+}
+
+static int cephwrap_fntimes(struct vfs_handle_struct *handle,
+			    files_struct *fsp,
+			    struct smb_file_time *ft)
+{
+	struct ceph_statx stx = { 0 };
+	int result;
+	int mask = 0;
+
+	if (!is_omit_timespec(&ft->atime)) {
+		stx.stx_atime = ft->atime;
+		mask |= CEPH_SETATTR_ATIME;
+	}
+	if (!is_omit_timespec(&ft->mtime)) {
+		stx.stx_mtime = ft->mtime;
+		mask |= CEPH_SETATTR_MTIME;
+	}
+	if (!is_omit_timespec(&ft->create_time)) {
+		stx.stx_btime = ft->create_time;
+		mask |= CEPH_SETATTR_BTIME;
+	}
+
+	if (!mask) {
+		return 0;
+	}
+
+	if (!fsp->fsp_flags.is_pathref) {
+		/*
+		 * We can use an io_fd to set xattrs.
+		 */
+		result = ceph_fsetattrx(handle->data,
+					fsp_get_io_fd(fsp),
+					&stx,
+					mask);
+	} else {
+		/*
+		 * This is no longer a handle based call.
+		 */
+		result = ceph_setattrx(handle->data,
+				       fsp->fsp_name->base_name,
+				       &stx,
+				       mask,
+				       0);
+	}
+
+	DBG_DEBUG("[CEPH] ntimes(%p, %s, {%ld, %ld, %ld, %ld}) = %d\n",
+		  handle, fsp_str_dbg(fsp), ft->mtime.tv_sec, ft->atime.tv_sec,
+		  ft->ctime.tv_sec, ft->create_time.tv_sec, result);
+
+	return result;
+}
+
+static int cephwrap_unlinkat(struct vfs_handle_struct *handle,
+			struct files_struct *dirfsp,
+			const struct smb_filename *smb_fname,
+			int flags)
+{
+	int result = -1;
+#ifdef HAVE_CEPH_UNLINKAT
+	int dirfd = fsp_get_pathref_fd(dirfsp);
+
+	DBG_DEBUG("[CEPH] unlinkat(%p, %d, %s)\n",
+		  handle,
+		  dirfd,
+		  smb_fname_str_dbg(smb_fname));
+
+	if (smb_fname->stream_name) {
+		errno = ENOENT;
+		return result;
+	}
+
+	result = ceph_unlinkat(handle->data,
+			       dirfd,
+			       smb_fname->base_name,
+			       flags);
+	DBG_DEBUG("[CEPH] unlinkat(...) = %d\n", result);
+	WRAP_RETURN(result);
+#else
+	struct smb_filename *full_fname = NULL;
+
+	DBG_DEBUG("[CEPH] unlink(%p, %s)\n",
+		handle,
+		smb_fname_str_dbg(smb_fname));
+
+	if (smb_fname->stream_name) {
+		errno = ENOENT;
+		return result;
+	}
+
+	full_fname = full_path_from_dirfsp_atname(talloc_tos(),
+						  dirfsp,
+						  smb_fname);
+	if (full_fname == NULL) {
+		return -1;
+	}
+
+	if (flags & AT_REMOVEDIR) {
+		result = ceph_rmdir(handle->data, full_fname->base_name);
+	} else {
+		result = ceph_unlink(handle->data, full_fname->base_name);
+	}
+	TALLOC_FREE(full_fname);
+	DBG_DEBUG("[CEPH] unlink(...) = %d\n", result);
+	WRAP_RETURN(result);
+#endif
+}
+
+static int cephwrap_fchmod(struct vfs_handle_struct *handle, files_struct *fsp, mode_t mode)
+{
+	int result;
+
+	DBG_DEBUG("[CEPH] fchmod(%p, %p, %d)\n", handle, fsp, mode);
+	if (!fsp->fsp_flags.is_pathref) {
+		/*
+		 * We can use an io_fd to change permissions.
+		 */
+		result = ceph_fchmod(handle->data, fsp_get_io_fd(fsp), mode);
+	} else {
+		/*
+		 * This is no longer a handle based call.
+		 */
+		result = ceph_chmod(handle->data,
+				    fsp->fsp_name->base_name,
+				    mode);
+	}
+	DBG_DEBUG("[CEPH] fchmod(...) = %d\n", result);
+	WRAP_RETURN(result);
+}
+
+static int cephwrap_fchown(struct vfs_handle_struct *handle, files_struct *fsp, uid_t uid, gid_t gid)
+{
+	int result;
+
+	DBG_DEBUG("[CEPH] fchown(%p, %p, %d, %d)\n", handle, fsp, uid, gid);
+	if (!fsp->fsp_flags.is_pathref) {
+		/*
+		 * We can use an io_fd to change ownership.
+		 */
+		result = ceph_fchown(handle->data,
+				     fsp_get_io_fd(fsp),
+				     uid,
+				     gid);
+	} else {
+		/*
+		 * This is no longer a handle based call.
+		 */
+		result = ceph_chown(handle->data,
+				    fsp->fsp_name->base_name,
+				    uid,
+				    gid);
+	}
+
+	DBG_DEBUG("[CEPH] fchown(...) = %d\n", result);
+	WRAP_RETURN(result);
+}
+
+static int cephwrap_lchown(struct vfs_handle_struct *handle,
+			const struct smb_filename *smb_fname,
+			uid_t uid,
+			gid_t gid)
+{
+	int result;
+	DBG_DEBUG("[CEPH] lchown(%p, %s, %d, %d)\n", handle, smb_fname->base_name, uid, gid);
+	result = ceph_lchown(handle->data, smb_fname->base_name, uid, gid);
+	DBG_DEBUG("[CEPH] lchown(...) = %d\n", result);
+	WRAP_RETURN(result);
+}
+
+static int cephwrap_chdir(struct vfs_handle_struct *handle,
+			const struct smb_filename *smb_fname)
+{
+	int result = -1;
+	DBG_DEBUG("[CEPH] chdir(%p, %s)\n", handle, smb_fname->base_name);
+	result = ceph_chdir(handle->data, smb_fname->base_name);
+	DBG_DEBUG("[CEPH] chdir(...) = %d\n", result);
+	WRAP_RETURN(result);
+}
+
+static struct smb_filename *cephwrap_getwd(struct vfs_handle_struct *handle,
+			TALLOC_CTX *ctx)
+{
+	const char *cwd = ceph_getcwd(handle->data);
+	DBG_DEBUG("[CEPH] getwd(%p) = %s\n", handle, cwd);
+	return synthetic_smb_fname(ctx,
+				cwd,
+				NULL,
+				NULL,
+				0,
+				0);
+}
+
+static int strict_allocate_ftruncate(struct vfs_handle_struct *handle, files_struct *fsp, off_t len)
+{
+	off_t space_to_write;
+	int result;
+	NTSTATUS status;
+	SMB_STRUCT_STAT *pst;
+
+	status = vfs_stat_fsp(fsp);
+	if (!NT_STATUS_IS_OK(status)) {
+		return -1;
+	}
+	pst = &fsp->fsp_name->st;
+
+#ifdef S_ISFIFO
+	if (S_ISFIFO(pst->st_ex_mode))
+		return 0;
+#endif
+
+	if (pst->st_ex_size == len)
+		return 0;
+
+	/* Shrink - just ftruncate. */
+	if (pst->st_ex_size > len) {
+		result = ceph_ftruncate(handle->data, fsp_get_io_fd(fsp), len);
+		WRAP_RETURN(result);
+	}
+
+	space_to_write = len - pst->st_ex_size;
+	result = ceph_fallocate(handle->data, fsp_get_io_fd(fsp), 0, pst->st_ex_size,
+				space_to_write);
+	WRAP_RETURN(result);
+}
+
+static int cephwrap_ftruncate(struct vfs_handle_struct *handle, files_struct *fsp, off_t len)
+{
+	int result = -1;
+
+	DBG_DEBUG("[CEPH] ftruncate(%p, %p, %llu\n", handle, fsp, llu(len));
+
+	if (lp_strict_allocate(SNUM(fsp->conn))) {
+		return strict_allocate_ftruncate(handle, fsp, len);
+	}
+
+	result = ceph_ftruncate(handle->data, fsp_get_io_fd(fsp), len);
+	WRAP_RETURN(result);
+}
+
+static int cephwrap_fallocate(struct vfs_handle_struct *handle,
+			      struct files_struct *fsp,
+			      uint32_t mode,
+			      off_t offset,
+			      off_t len)
+{
+	int result;
+
+	DBG_DEBUG("[CEPH] fallocate(%p, %p, %u, %llu, %llu\n",
+		  handle, fsp, mode, llu(offset), llu(len));
+	/* unsupported mode flags are rejected by libcephfs */
+	result = ceph_fallocate(handle->data, fsp_get_io_fd(fsp), mode, offset, len);
+	DBG_DEBUG("[CEPH] fallocate(...) = %d\n", result);
+	WRAP_RETURN(result);
+}
+
+static bool cephwrap_lock(struct vfs_handle_struct *handle, files_struct *fsp, int op, off_t offset, off_t count, int type)
+{
+	DBG_DEBUG("[CEPH] lock\n");
+	return true;
+}
+
+static int cephwrap_filesystem_sharemode(struct vfs_handle_struct *handle,
+					 files_struct *fsp,
+					 uint32_t share_access,
+					 uint32_t access_mask)
+{
+	DBG_ERR("[CEPH] filesystem sharemodes unsupported! Consider setting "
+		"\"kernel share modes = no\"\n");
+
+	errno = ENOSYS;
+	return -1;
+}
+
+static int cephwrap_fcntl(vfs_handle_struct *handle,
+			  files_struct *fsp, int cmd, va_list cmd_arg)
+{
+	/*
+	 * SMB_VFS_FCNTL() is currently only called by vfs_set_blocking() to
+	 * clear O_NONBLOCK, etc for LOCK_MAND and FIFOs. Ignore it.
+	 */
+	if (cmd == F_GETFL) {
+		return 0;
+	} else if (cmd == F_SETFL) {
+		va_list dup_cmd_arg;
+		int opt;
+
+		va_copy(dup_cmd_arg, cmd_arg);
+		opt = va_arg(dup_cmd_arg, int);
+		va_end(dup_cmd_arg);
+		if (opt == 0) {
+			return 0;
+		}
+		DBG_ERR("unexpected fcntl SETFL(%d)\n", opt);
+		goto err_out;
+	}
+	DBG_ERR("unexpected fcntl: %d\n", cmd);
+err_out:
+	errno = EINVAL;
+	return -1;
+}
+
+static bool cephwrap_getlock(struct vfs_handle_struct *handle, files_struct *fsp, off_t *poffset, off_t *pcount, int *ptype, pid_t *ppid)
+{
+	DBG_DEBUG("[CEPH] getlock returning false and errno=0\n");
+
+	errno = 0;
+	return false;
+}
+
+/*
+ * We cannot let this fall through to the default, because the file might only
+ * be accessible from libcephfs (which is a user-space client) but the fd might
+ * be for some file the kernel knows about.
+ */
+static int cephwrap_linux_setlease(struct vfs_handle_struct *handle, files_struct *fsp,
+				int leasetype)
+{
+	int result = -1;
+
+	DBG_DEBUG("[CEPH] linux_setlease\n");
+	errno = ENOSYS;
+	return result;
+}
+
+static int cephwrap_symlinkat(struct vfs_handle_struct *handle,
+		const struct smb_filename *link_target,
+		struct files_struct *dirfsp,
+		const struct smb_filename *new_smb_fname)
+{
+	int result = -1;
+#ifdef HAVE_CEPH_SYMLINKAT
+	int dirfd = fsp_get_pathref_fd(dirfsp);
+
+	DBG_DEBUG("[CEPH] symlinkat(%p, %s, %d, %s)\n",
+		  handle,
+		  link_target->base_name,
+		  dirfd,
+		  new_smb_fname->base_name);
+
+	result = ceph_symlinkat(handle->data,
+				link_target->base_name,
+				dirfd,
+				new_smb_fname->base_name);
+	DBG_DEBUG("[CEPH] symlinkat(...) = %d\n", result);
+	WRAP_RETURN(result);
+#else
+	struct smb_filename *full_fname = NULL;
+
+	full_fname = full_path_from_dirfsp_atname(talloc_tos(),
+						dirfsp,
+						new_smb_fname);
+	if (full_fname == NULL) {
+		return -1;
+	}
+
+	DBG_DEBUG("[CEPH] symlink(%p, %s, %s)\n", handle,
+			link_target->base_name,
+			full_fname->base_name);
+
+	result = ceph_symlink(handle->data,
+			link_target->base_name,
+			full_fname->base_name);
+	TALLOC_FREE(full_fname);
+	DBG_DEBUG("[CEPH] symlink(...) = %d\n", result);
+	WRAP_RETURN(result);
+#endif
+}
+
+static int cephwrap_readlinkat(struct vfs_handle_struct *handle,
+		const struct files_struct *dirfsp,
+		const struct smb_filename *smb_fname,
+		char *buf,
+		size_t bufsiz)
+{
+	int result = -1;
+#ifdef HAVE_CEPH_READLINKAT
+	int dirfd = fsp_get_pathref_fd(dirfsp);
+
+	DBG_DEBUG("[CEPH] readlinkat(%p, %d, %s, %p, %llu)\n",
+		  handle,
+		  dirfd,
+		  smb_fname->base_name,
+		  buf,
+		  llu(bufsiz));
+
+	result = ceph_readlinkat(handle->data,
+				 dirfd,
+				 smb_fname->base_name,
+				 buf,
+				 bufsiz);
+
+	DBG_DEBUG("[CEPH] readlinkat(...) = %d\n", result);
+	WRAP_RETURN(result);
+#else
+	struct smb_filename *full_fname = NULL;
+
+	full_fname = full_path_from_dirfsp_atname(talloc_tos(),
+						dirfsp,
+						smb_fname);
+	if (full_fname == NULL) {
+		return -1;
+	}
+
+	DBG_DEBUG("[CEPH] readlink(%p, %s, %p, %llu)\n", handle,
+			full_fname->base_name, buf, llu(bufsiz));
+
+	result = ceph_readlink(handle->data, full_fname->base_name, buf, bufsiz);
+	TALLOC_FREE(full_fname);
+	DBG_DEBUG("[CEPH] readlink(...) = %d\n", result);
+	WRAP_RETURN(result);
+#endif
+}
+
+static int cephwrap_linkat(struct vfs_handle_struct *handle,
+		files_struct *srcfsp,
+		const struct smb_filename *old_smb_fname,
+		files_struct *dstfsp,
+		const struct smb_filename *new_smb_fname,
+		int flags)
+{
+	struct smb_filename *full_fname_old = NULL;
+	struct smb_filename *full_fname_new = NULL;
+	int result = -1;
+
+	full_fname_old = full_path_from_dirfsp_atname(talloc_tos(),
+					srcfsp,
+					old_smb_fname);
+	if (full_fname_old == NULL) {
+		return -1;
+	}
+	full_fname_new = full_path_from_dirfsp_atname(talloc_tos(),
+					dstfsp,
+					new_smb_fname);
+	if (full_fname_new == NULL) {
+		TALLOC_FREE(full_fname_old);
+		return -1;
+	}
+
+	DBG_DEBUG("[CEPH] link(%p, %s, %s)\n", handle,
+			full_fname_old->base_name,
+			full_fname_new->base_name);
+
+	result = ceph_link(handle->data,
+				full_fname_old->base_name,
+				full_fname_new->base_name);
+	DBG_DEBUG("[CEPH] link(...) = %d\n", result);
+	TALLOC_FREE(full_fname_old);
+	TALLOC_FREE(full_fname_new);
+	WRAP_RETURN(result);
+}
+
+static int cephwrap_mknodat(struct vfs_handle_struct *handle,
+		files_struct *dirfsp,
+		const struct smb_filename *smb_fname,
+		mode_t mode,
+		SMB_DEV_T dev)
+{
+	struct smb_filename *full_fname = NULL;
+	int result = -1;
+
+	full_fname = full_path_from_dirfsp_atname(talloc_tos(),
+						dirfsp,
+						smb_fname);
+	if (full_fname == NULL) {
+		return -1;
+	}
+
+	DBG_DEBUG("[CEPH] mknodat(%p, %s)\n", handle, full_fname->base_name);
+	result = ceph_mknod(handle->data, full_fname->base_name, mode, dev);
+	DBG_DEBUG("[CEPH] mknodat(...) = %d\n", result);
+
+	TALLOC_FREE(full_fname);
+
+	WRAP_RETURN(result);
+}
+
+/*
+ * This is a simple version of real-path ... a better version is needed to
+ * ask libcephfs about symbolic links.
+ */
+static struct smb_filename *cephwrap_realpath(struct vfs_handle_struct *handle,
+				TALLOC_CTX *ctx,
+				const struct smb_filename *smb_fname)
+{
+	char *result = NULL;
+	const char *path = smb_fname->base_name;
+	size_t len = strlen(path);
+	struct smb_filename *result_fname = NULL;
+	int r = -1;
+
+	if (len && (path[0] == '/')) {
+		r = asprintf(&result, "%s", path);
+	} else if ((len >= 2) && (path[0] == '.') && (path[1] == '/')) {
+		if (len == 2) {
+			r = asprintf(&result, "%s",
+					handle->conn->cwd_fsp->fsp_name->base_name);
+		} else {
+			r = asprintf(&result, "%s/%s",
+					handle->conn->cwd_fsp->fsp_name->base_name, &path[2]);
+		}
+	} else {
+		r = asprintf(&result, "%s/%s",
+				handle->conn->cwd_fsp->fsp_name->base_name, path);
+	}
+
+	if (r < 0) {
+		return NULL;
+	}
+
+	DBG_DEBUG("[CEPH] realpath(%p, %s) = %s\n", handle, path, result);
+	result_fname = synthetic_smb_fname(ctx,
+				result,
+				NULL,
+				NULL,
+				0,
+				0);
+	SAFE_FREE(result);
+	return result_fname;
+}
+
+
+static int cephwrap_fchflags(struct vfs_handle_struct *handle,
+			struct files_struct *fsp,
+			unsigned int flags)
+{
+	errno = ENOSYS;
+	return -1;
+}
+
+static NTSTATUS cephwrap_get_real_filename_at(
+	struct vfs_handle_struct *handle,
+	struct files_struct *dirfsp,
+	const char *name,
+	TALLOC_CTX *mem_ctx,
+	char **found_name)
+{
+	/*
+	 * Don't fall back to get_real_filename so callers can differentiate
+	 * between a full directory scan and an actual case-insensitive stat.
+	 */
+	return NT_STATUS_NOT_SUPPORTED;
+}
+
+static const char *cephwrap_connectpath(
+	struct vfs_handle_struct *handle,
+	const struct files_struct *dirfsp,
+	const struct smb_filename *smb_fname)
+{
+	return handle->conn->connectpath;
+}
+
+/****************************************************************
+ Extended attribute operations.
+*****************************************************************/
+
+static ssize_t cephwrap_fgetxattr(struct vfs_handle_struct *handle,
+				  struct files_struct *fsp,
+				  const char *name,
+				  void *value,
+				  size_t size)
+{
+	int ret;
+	DBG_DEBUG("[CEPH] fgetxattr(%p, %p, %s, %p, %llu)\n",
+		  handle,
+		  fsp,
+		  name,
+		  value,
+		  llu(size));
+	if (!fsp->fsp_flags.is_pathref) {
+		ret = ceph_fgetxattr(handle->data,
+				     fsp_get_io_fd(fsp),
+				     name,
+				     value,
+				     size);
+	} else {
+		ret = ceph_getxattr(handle->data,
+				    fsp->fsp_name->base_name,
+				    name,
+				    value,
+				    size);
+	}
+	DBG_DEBUG("[CEPH] fgetxattr(...) = %d\n", ret);
+	if (ret < 0) {
+		WRAP_RETURN(ret);
+	}
+	return (ssize_t)ret;
+}
+
+static ssize_t cephwrap_flistxattr(struct vfs_handle_struct *handle, struct files_struct *fsp, char *list, size_t size)
+{
+	int ret;
+	DBG_DEBUG("[CEPH] flistxattr(%p, %p, %p, %llu)\n",
+		  handle, fsp, list, llu(size));
+	if (!fsp->fsp_flags.is_pathref) {
+		/*
+		 * We can use an io_fd to list xattrs.
+		 */
+		ret = ceph_flistxattr(handle->data,
+					fsp_get_io_fd(fsp),
+					list,
+					size);
+	} else {
+		/*
+		 * This is no longer a handle based call.
+		 */
+		ret = ceph_listxattr(handle->data,
+					fsp->fsp_name->base_name,
+					list,
+					size);
+	}
+	DBG_DEBUG("[CEPH] flistxattr(...) = %d\n", ret);
+	if (ret < 0) {
+		WRAP_RETURN(ret);
+	}
+	return (ssize_t)ret;
+}
+
+static int cephwrap_fremovexattr(struct vfs_handle_struct *handle, struct files_struct *fsp, const char *name)
+{
+	int ret;
+	DBG_DEBUG("[CEPH] fremovexattr(%p, %p, %s)\n", handle, fsp, name);
+	if (!fsp->fsp_flags.is_pathref) {
+		/*
+		 * We can use an io_fd to remove xattrs.
+		 */
+		ret = ceph_fremovexattr(handle->data, fsp_get_io_fd(fsp), name);
+	} else {
+		/*
+		 * This is no longer a handle based call.
+		 */
+		ret = ceph_removexattr(handle->data,
+					fsp->fsp_name->base_name,
+					name);
+	}
+	DBG_DEBUG("[CEPH] fremovexattr(...) = %d\n", ret);
+	WRAP_RETURN(ret);
+}
+
+static int cephwrap_fsetxattr(struct vfs_handle_struct *handle, struct files_struct *fsp, const char *name, const void *value, size_t size, int flags)
+{
+	int ret;
+	DBG_DEBUG("[CEPH] fsetxattr(%p, %p, %s, %p, %llu, %d)\n", handle, fsp, name, value, llu(size), flags);
+	if (!fsp->fsp_flags.is_pathref) {
+		/*
+		 * We can use an io_fd to set xattrs.
+		 */
+		ret = ceph_fsetxattr(handle->data,
+				fsp_get_io_fd(fsp),
+				name,
+				value,
+				size,
+				flags);
+	} else {
+		/*
+		 * This is no longer a handle based call.
+		 */
+		ret = ceph_setxattr(handle->data,
+				fsp->fsp_name->base_name,
+				name,
+				value,
+				size,
+				flags);
+	}
+	DBG_DEBUG("[CEPH] fsetxattr(...) = %d\n", ret);
+	WRAP_RETURN(ret);
+}
+
+static bool cephwrap_aio_force(struct vfs_handle_struct *handle, struct files_struct *fsp)
+{
+
+	/*
+	 * We do not support AIO yet.
+	 */
+
+	DBG_DEBUG("[CEPH] cephwrap_aio_force(%p, %p) = false (errno = ENOTSUP)\n", handle, fsp);
+	errno = ENOTSUP;
+	return false;
+}
+
+static NTSTATUS cephwrap_create_dfs_pathat(struct vfs_handle_struct *handle,
+				struct files_struct *dirfsp,
+				const struct smb_filename *smb_fname,
+				const struct referral *reflist,
+				size_t referral_count)
+{
+	TALLOC_CTX *frame = talloc_stackframe();
+	NTSTATUS status = NT_STATUS_NO_MEMORY;
+	int ret;
+	char *msdfs_link = NULL;
+	struct smb_filename *full_fname = NULL;
+
+	full_fname = full_path_from_dirfsp_atname(talloc_tos(),
+						dirfsp,
+						smb_fname);
+	if (full_fname == NULL) {
+		goto out;
+	}
+
+	/* Form the msdfs_link contents */
+	msdfs_link = msdfs_link_string(frame,
+					reflist,
+					referral_count);
+	if (msdfs_link == NULL) {
+		goto out;
+	}
+
+	ret = ceph_symlink(handle->data,
+			msdfs_link,
+			full_fname->base_name);
+	if (ret == 0) {
+		status = NT_STATUS_OK;
+	} else {
+		status = map_nt_error_from_unix(-ret);
+        }
+
+  out:
+
+	DBG_DEBUG("[CEPH] create_dfs_pathat(%s) = %s\n",
+			full_fname != NULL ? full_fname->base_name : "",
+			nt_errstr(status));
+
+	TALLOC_FREE(frame);
+	return status;
+}
+
+/*
+ * Read and return the contents of a DFS redirect given a
+ * pathname. A caller can pass in NULL for ppreflist and
+ * preferral_count but still determine if this was a
+ * DFS redirect point by getting NT_STATUS_OK back
+ * without incurring the overhead of reading and parsing
+ * the referral contents.
+ */
+
+static NTSTATUS cephwrap_read_dfs_pathat(struct vfs_handle_struct *handle,
+				TALLOC_CTX *mem_ctx,
+				struct files_struct *dirfsp,
+				struct smb_filename *smb_fname,
+				struct referral **ppreflist,
+				size_t *preferral_count)
+{
+	NTSTATUS status = NT_STATUS_NO_MEMORY;
+	size_t bufsize;
+	char *link_target = NULL;
+	int referral_len;
+	bool ok;
+#if defined(HAVE_BROKEN_READLINK)
+	char link_target_buf[PATH_MAX];
+#else
+	char link_target_buf[7];
+#endif
+	struct ceph_statx stx = { 0 };
+	struct smb_filename *full_fname = NULL;
+	int ret;
+
+	if (is_named_stream(smb_fname)) {
+		status = NT_STATUS_OBJECT_NAME_NOT_FOUND;
+		goto err;
+	}
+
+	if (ppreflist == NULL && preferral_count == NULL) {
+		/*
+		 * We're only checking if this is a DFS
+		 * redirect. We don't need to return data.
+		 */
+		bufsize = sizeof(link_target_buf);
+		link_target = link_target_buf;
+	} else {
+		bufsize = PATH_MAX;
+		link_target = talloc_array(mem_ctx, char, bufsize);
+		if (!link_target) {
+			goto err;
+		}
+	}
+
+	full_fname = full_path_from_dirfsp_atname(talloc_tos(),
+						  dirfsp,
+						  smb_fname);
+	if (full_fname == NULL) {
+		status = NT_STATUS_NO_MEMORY;
+		goto err;
+	}
+
+	ret = ceph_statx(handle->data,
+			 full_fname->base_name,
+			 &stx,
+			 SAMBA_STATX_ATTR_MASK,
+			 AT_SYMLINK_NOFOLLOW);
+	if (ret < 0) {
+		status = map_nt_error_from_unix(-ret);
+		goto err;
+	}
+
+        referral_len = ceph_readlink(handle->data,
+                                full_fname->base_name,
+                                link_target,
+                                bufsize - 1);
+        if (referral_len < 0) {
+		/* ceph errors are -errno. */
+		if (-referral_len == EINVAL) {
+			DBG_INFO("%s is not a link.\n",
+				full_fname->base_name);
+			status = NT_STATUS_OBJECT_TYPE_MISMATCH;
+		} else {
+	                status = map_nt_error_from_unix(-referral_len);
+			DBG_ERR("Error reading "
+				"msdfs link %s: %s\n",
+				full_fname->base_name,
+			strerror(errno));
+		}
+                goto err;
+        }
+        link_target[referral_len] = '\0';
+
+        DBG_INFO("%s -> %s\n",
+                        full_fname->base_name,
+                        link_target);
+
+        if (!strnequal(link_target, "msdfs:", 6)) {
+                status = NT_STATUS_OBJECT_TYPE_MISMATCH;
+                goto err;
+        }
+
+        if (ppreflist == NULL && preferral_count == NULL) {
+                /* Early return for checking if this is a DFS link. */
+		TALLOC_FREE(full_fname);
+		init_stat_ex_from_ceph_statx(&smb_fname->st, &stx);
+                return NT_STATUS_OK;
+        }
+
+        ok = parse_msdfs_symlink(mem_ctx,
+                        lp_msdfs_shuffle_referrals(SNUM(handle->conn)),
+                        link_target,
+                        ppreflist,
+                        preferral_count);
+
+        if (ok) {
+		init_stat_ex_from_ceph_statx(&smb_fname->st, &stx);
+                status = NT_STATUS_OK;
+        } else {
+                status = NT_STATUS_NO_MEMORY;
+        }
+
+  err:
+
+        if (link_target != link_target_buf) {
+                TALLOC_FREE(link_target);
+        }
+	TALLOC_FREE(full_fname);
+        return status;
+}
+
+static struct vfs_fn_pointers ceph_fns = {
+	/* Disk operations */
+
+	.connect_fn = cephwrap_connect,
+	.disconnect_fn = cephwrap_disconnect,
+	.disk_free_fn = cephwrap_disk_free,
+	.get_quota_fn = cephwrap_get_quota,
+	.set_quota_fn = cephwrap_set_quota,
+	.statvfs_fn = cephwrap_statvfs,
+	.fs_capabilities_fn = cephwrap_fs_capabilities,
+
+	/* Directory operations */
+
+	.fdopendir_fn = cephwrap_fdopendir,
+	.readdir_fn = cephwrap_readdir,
+	.rewind_dir_fn = cephwrap_rewinddir,
+	.mkdirat_fn = cephwrap_mkdirat,
+	.closedir_fn = cephwrap_closedir,
+
+	/* File operations */
+
+	.create_dfs_pathat_fn = cephwrap_create_dfs_pathat,
+	.read_dfs_pathat_fn = cephwrap_read_dfs_pathat,
+	.openat_fn = cephwrap_openat,
+	.close_fn = cephwrap_close,
+	.pread_fn = cephwrap_pread,
+	.pread_send_fn = cephwrap_pread_send,
+	.pread_recv_fn = cephwrap_pread_recv,
+	.pwrite_fn = cephwrap_pwrite,
+	.pwrite_send_fn = cephwrap_pwrite_send,
+	.pwrite_recv_fn = cephwrap_pwrite_recv,
+	.lseek_fn = cephwrap_lseek,
+	.sendfile_fn = cephwrap_sendfile,
+	.recvfile_fn = cephwrap_recvfile,
+	.renameat_fn = cephwrap_renameat,
+	.fsync_send_fn = cephwrap_fsync_send,
+	.fsync_recv_fn = cephwrap_fsync_recv,
+	.stat_fn = cephwrap_stat,
+	.fstat_fn = cephwrap_fstat,
+	.lstat_fn = cephwrap_lstat,
+	.fstatat_fn = cephwrap_fstatat,
+	.unlinkat_fn = cephwrap_unlinkat,
+	.fchmod_fn = cephwrap_fchmod,
+	.fchown_fn = cephwrap_fchown,
+	.lchown_fn = cephwrap_lchown,
+	.chdir_fn = cephwrap_chdir,
+	.getwd_fn = cephwrap_getwd,
+	.fntimes_fn = cephwrap_fntimes,
+	.ftruncate_fn = cephwrap_ftruncate,
+	.fallocate_fn = cephwrap_fallocate,
+	.lock_fn = cephwrap_lock,
+	.filesystem_sharemode_fn = cephwrap_filesystem_sharemode,
+	.fcntl_fn = cephwrap_fcntl,
+	.linux_setlease_fn = cephwrap_linux_setlease,
+	.getlock_fn = cephwrap_getlock,
+	.symlinkat_fn = cephwrap_symlinkat,
+	.readlinkat_fn = cephwrap_readlinkat,
+	.linkat_fn = cephwrap_linkat,
+	.mknodat_fn = cephwrap_mknodat,
+	.realpath_fn = cephwrap_realpath,
+	.fchflags_fn = cephwrap_fchflags,
+	.get_real_filename_at_fn = cephwrap_get_real_filename_at,
+	.connectpath_fn = cephwrap_connectpath,
+
+	/* EA operations. */
+	.getxattrat_send_fn = vfs_not_implemented_getxattrat_send,
+	.getxattrat_recv_fn = vfs_not_implemented_getxattrat_recv,
+	.fgetxattr_fn = cephwrap_fgetxattr,
+	.flistxattr_fn = cephwrap_flistxattr,
+	.fremovexattr_fn = cephwrap_fremovexattr,
+	.fsetxattr_fn = cephwrap_fsetxattr,
+
+	/* Posix ACL Operations */
+	.sys_acl_get_fd_fn = posixacl_xattr_acl_get_fd,
+	.sys_acl_blob_get_fd_fn = posix_sys_acl_blob_get_fd,
+	.sys_acl_set_fd_fn = posixacl_xattr_acl_set_fd,
+	.sys_acl_delete_def_fd_fn = posixacl_xattr_acl_delete_def_fd,
+
+	/* aio operations */
+	.aio_force_fn = cephwrap_aio_force,
+};
+
+static_decl_vfs;
+NTSTATUS vfs_ceph_init(TALLOC_CTX *ctx)
+{
+	return smb_register_vfs(SMB_VFS_INTERFACE_VERSION,
+				"ceph", &ceph_fns);
+}