r24302: add a module that overloads the file_id_create VFS function

and alters the device id depending on the configured algorithm.

The algorithm is configured via "fileid:algorithm":
- "fsname" (default) uses a uint64 hash over the mount point
- "fsid" uses the fsid returned from statfs()

This is needed for "clustering = yes" on some clusterfilesystems

metze
(This used to be commit 30f9171cca3e4f523cde7dfc96096c32e5af50be)

3 files changed, 326 insertions, 0 deletions

diff --git a/source3/Makefile.in b/source3/Makefile.in
index 4ffebdac35..36b5df1bed 100644
--- a/source3/Makefile.in
+++ b/source3/Makefile.in
@@ -479,6 +479,7 @@ VFS_COMMIT_OBJ = modules/vfs_commit.o
 VFS_GPFS_OBJ = modules/vfs_gpfs.o modules/gpfs.o modules/nfs4_acls.o
 VFS_NOTIFY_FAM_OBJ = modules/vfs_notify_fam.o
 VFS_READAHEAD_OBJ = modules/vfs_readahead.o
+VFS_FILEID_OBJ = modules/vfs_fileid.o
 
 PLAINTEXT_AUTH_OBJ = auth/pampass.o auth/pass_check.o
 
@@ -1681,6 +1682,10 @@ bin/readahead.@SHLIBEXT@: $(BINARY_PREREQS) $(VFS_READAHEAD_OBJ)
 	@echo "Building plugin $@"
 	@$(SHLD_MODULE) $(VFS_READAHEAD_OBJ)
 
+bin/fileid.@SHLIBEXT@: $(BINARY_PREREQS) $(VFS_FILEID_OBJ)
+	@echo "Building plugin $@"
+	@$(SHLD_MODULE) $(VFS_FILEID_OBJ)
+
 #########################################################
 ## IdMap NSS plugins
 
diff --git a/source3/configure.in b/source3/configure.in
index d1ef464202..4029a0a10c 100644
--- a/source3/configure.in
+++ b/source3/configure.in
@@ -6236,6 +6236,38 @@ AC_ARG_WITH(python,
   esac ])
 AC_SUBST(PYTHON)
 
+
+# Checks for the vfs_fileid module
+# Start
+AC_CHECK_FUNC(getmntent)
+
+AC_CHECK_HEADERS(sys/statfs.h)
+
+AC_MSG_CHECKING([vfs_fileid: checking for statfs() and struct statfs.f_fsid)])
+AC_CACHE_VAL(vfsfileid_cv_statfs,[
+	     AC_TRY_RUN([
+		#include <sys/types.h>
+		#include <sys/statfs.h>
+		int main(void)
+		{
+			struct statfs fsd;
+			fsid_t fsid = fsd.f_fsid;
+			return statfs (".", &fsd);
+		}],
+		vfsfileid_cv_statfs=yes,
+		vfsfileid_cv_statfs=no,
+		vfsfileid_cv_statfs=cross)
+])
+AC_MSG_RESULT($vfsfileid_cv_statfs)
+
+if test x"$ac_cv_func_getmntent" = x"yes" -a \
+	x"$vfsfileid_cv_statfs" = x"yes"; then
+	default_shared_modules="$default_shared_modules vfs_fileid"
+fi
+# End
+# Checks for the vfs_fileid module
+
+
 for i in `echo $default_static_modules | sed -e 's/,/ /g'`
 do
 	eval MODULE_DEFAULT_$i=STATIC
@@ -6356,6 +6388,7 @@ SMB_MODULE(vfs_commit, \$(VFS_COMMIT_OBJ), "bin/commit.$SHLIBEXT", VFS)
 SMB_MODULE(vfs_gpfs, \$(VFS_GPFS_OBJ), "bin/gpfs.$SHLIBEXT", VFS)
 SMB_MODULE(vfs_notify_fam, \$(VFS_NOTIFY_FAM_OBJ), "bin/notify_fam.$SHLIBEXT", VFS)
 SMB_MODULE(vfs_readahead, \$(VFS_READAHEAD_OBJ), "bin/readahead.$SHLIBEXT", VFS)
+SMB_MODULE(vfs_fileid, \$(VFS_FILEID_OBJ), "bin/fileid.$SHLIBEXT", VFS)
 
 SMB_SUBSYSTEM(VFS,smbd/vfs.o)
 
diff --git a/source3/modules/vfs_fileid.c b/source3/modules/vfs_fileid.c
new file mode 100644
index 0000000000..5954a10ddb
--- /dev/null
+++ b/source3/modules/vfs_fileid.c
@@ -0,0 +1,288 @@
+/*
+ * VFS module to alter the algorithm to calculate
+ * the struct file_id used as key for the share mode
+ * and byte range locking db's.
+ *
+ * Copyright (C) 2007, Stefan Metzmacher
+ *
+ * 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/>.
+ */
+
+#include "includes.h"
+
+static int vfs_fileid_debug_level = DBGC_VFS;
+
+#undef DBGC_CLASS
+#define DBGC_CLASS vfs_fileid_debug_level
+
+struct fileid_mount_entry {
+	SMB_DEV_T device;
+	const char *mnt_fsname;
+	fsid_t fsid;
+	uint64_t devid;
+};
+
+struct fileid_handle_data {
+	uint64_t (*device_mapping_fn)(struct fileid_handle_data *data,
+				      SMB_DEV_T dev);
+	unsigned num_mount_entries;
+	struct fileid_mount_entry *mount_entries;
+};
+
+/* load all the mount entries from the mtab */
+static void fileid_load_mount_entries(struct fileid_handle_data *data)
+{
+	FILE *f;
+	struct mntent *m;
+
+	data->num_mount_entries = 0;
+	TALLOC_FREE(data->mount_entries);
+
+	f = setmntent("/etc/mtab", "r");
+	if (!f) return;
+
+	while ((m = getmntent(f))) {
+		struct stat st;
+		struct statfs sfs;
+		struct fileid_mount_entry *cur;
+
+		if (stat(m->mnt_dir, &st) != 0) continue;
+		if (statfs(m->mnt_dir, &sfs) != 0) continue;
+
+		if (strncmp(m->mnt_fsname, "/dev/", 5) == 0) {
+			m->mnt_fsname += 5;
+		}
+
+		data->mount_entries = TALLOC_REALLOC_ARRAY(data,
+							   data->mount_entries,
+							   struct fileid_mount_entry,
+							   data->num_mount_entries+1);
+		if (data->mount_entries == NULL) {
+			goto nomem;
+		}
+
+		cur = &data->mount_entries[data->num_mount_entries];
+		cur->device	= st.st_dev;
+		cur->mnt_fsname = talloc_strdup(data->mount_entries,
+						m->mnt_fsname);
+		if (!cur->mnt_fsname) goto nomem;
+		cur->fsid	= sfs.f_fsid;
+		cur->devid	= (uint64_t)-1;
+
+		data->num_mount_entries++;
+	}
+	endmntent(f);
+	return;
+	
+nomem:
+	if (f) endmntent(f);
+
+	data->num_mount_entries = 0;
+	TALLOC_FREE(data->mount_entries);
+
+	return;
+}
+
+/* find a mount entry given a dev_t */
+static struct fileid_mount_entry *fileid_find_mount_entry(struct fileid_handle_data *data,
+							  SMB_DEV_T dev)
+{
+	int i;
+
+	if (data->num_mount_entries == 0) {
+		fileid_load_mount_entries(data);
+	}
+	for (i=0;i<data->num_mount_entries;i++) {
+		if (data->mount_entries[i].device == dev) {
+			return &data->mount_entries[i];
+		}
+	}
+	/* 2nd pass after reloading */
+	fileid_load_mount_entries(data);
+	for (i=0;i<data->num_mount_entries;i++) {
+		if (data->mount_entries[i].device == dev) {
+			return &data->mount_entries[i];
+		}
+	}	
+	return NULL;
+}
+
+
+/* a 64 bit hash, based on the one in tdb */
+static uint64_t fileid_uint64_hash(const uint8_t *s, size_t len)
+{
+	uint64_t value;	/* Used to compute the hash value.  */
+	uint32_t i;	/* Used to cycle through random values. */
+
+	/* Set the initial value from the key size. */
+	for (value = 0x238F13AFLL * len, i=0; i < len; i++)
+		value = (value + (s[i] << (i*5 % 24)));
+
+	return (1103515243LL * value + 12345LL);
+}
+
+/* a device mapping using a fsname */
+static uint64_t fileid_device_mapping_fsname(struct fileid_handle_data *data,
+					     SMB_DEV_T dev)
+{
+	struct fileid_mount_entry *m;
+
+	m = fileid_find_mount_entry(data, dev);
+	if (!m) return dev;
+
+	if (m->devid == (uint64_t)-1) {
+		m->devid = fileid_uint64_hash((uint8_t *)m->mnt_fsname,
+					      strlen(m->mnt_fsname));
+	}
+
+	return m->devid;
+}
+
+/* device mapping functions using a fsid */
+static uint64_t fileid_device_mapping_fsid(struct fileid_handle_data *data,
+					   SMB_DEV_T dev)
+{
+	struct fileid_mount_entry *m;
+
+	m = fileid_find_mount_entry(data, dev);
+	if (!m) return dev;
+
+	if (m->devid == (uint64_t)-1) {
+		if (sizeof(fsid_t) > sizeof(uint64_t)) {
+			m->devid = fileid_uint64_hash((uint8_t *)&m->fsid,
+						      sizeof(m->fsid));
+		} else {
+			union {
+				uint64_t ret;
+				fsid_t fsid;
+			} u;
+			ZERO_STRUCT(u);
+			u.fsid = m->fsid;
+			m->devid = u.ret;
+		}
+	}
+
+	return m->devid;
+}
+
+static int fileid_connect(struct vfs_handle_struct *handle,
+			  const char *service, const char *user)
+{
+	struct fileid_handle_data *data;
+	const char *algorithm;
+
+	data = talloc_zero(handle->conn->mem_ctx, struct fileid_handle_data);
+	if (!data) {
+		DEBUG(0, ("talloc_zero() failed\n"));
+		return -1;
+	}
+
+	data->device_mapping_fn	= fileid_device_mapping_fsid;
+	algorithm = lp_parm_const_string(SNUM(handle->conn),
+					 "fileid", "algorithm",
+					 "fsname");
+	if (strcmp("fsname", algorithm) == 0) {
+		data->device_mapping_fn	= fileid_device_mapping_fsname;
+	} else if (strcmp("fsid", algorithm) == 0) {
+		data->device_mapping_fn	= fileid_device_mapping_fsid;
+	} else {
+		DEBUG(0,("fileid_connect(): unknown algorithm[%s]\n", algorithm));
+		return -1;
+	}
+
+	SMB_VFS_HANDLE_SET_DATA(handle, data, NULL,
+				struct fileid_handle_data,
+				return -1);
+
+	DEBUG(10, ("fileid_connect(): connect to service[%s] with algorithm[%s]\n",
+		service, algorithm));
+
+	return SMB_VFS_NEXT_CONNECT(handle, service, user);
+}
+
+static void fileid_disconnect(struct vfs_handle_struct *handle)
+{
+	DEBUG(10,("fileid_disconnect() connect to service[%s].\n",
+		lp_servicename(SNUM(handle->conn))));
+
+	SMB_VFS_NEXT_DISCONNECT(handle);
+}
+
+static struct file_id fileid_file_id_create(struct vfs_handle_struct *handle,
+					    SMB_DEV_T dev, SMB_INO_T inode)
+{
+	struct fileid_handle_data *data;
+	struct file_id id;
+
+	ZERO_STRUCT(id);
+
+	SMB_VFS_HANDLE_GET_DATA(handle, data,
+				struct fileid_handle_data,
+				return id);
+
+	id.devid	= data->device_mapping_fn(data, dev);
+	id.inode	= inode;
+
+	return id;
+}
+
+static vfs_op_tuple fileid_ops[] = {
+
+	/* Disk operations */
+	{
+		SMB_VFS_OP(fileid_connect),
+		SMB_VFS_OP_CONNECT,
+		SMB_VFS_LAYER_TRANSPARENT
+	},
+	{
+		SMB_VFS_OP(fileid_disconnect),
+		SMB_VFS_OP_DISCONNECT,
+		SMB_VFS_LAYER_TRANSPARENT
+	},
+
+	/* File operations */
+	{
+		SMB_VFS_OP(fileid_file_id_create),
+		SMB_VFS_OP_FILE_ID_CREATE,
+		SMB_VFS_LAYER_OPAQUE
+	},
+
+	/* End marker */
+	{
+		SMB_VFS_OP(NULL),
+		SMB_VFS_OP_NOOP,
+		SMB_VFS_LAYER_NOOP
+	}
+};
+
+NTSTATUS vfs_fileid_init(void);
+NTSTATUS vfs_fileid_init(void)
+{
+	NTSTATUS ret;
+
+	ret = smb_register_vfs(SMB_VFS_INTERFACE_VERSION, "fileid", fileid_ops);
+	if (!NT_STATUS_IS_OK(ret)) {
+		return ret;
+	}
+
+	vfs_fileid_debug_level = debug_add_class("fileid");
+	if (vfs_fileid_debug_level == -1) {
+		vfs_fileid_debug_level = DBGC_VFS;
+		DEBUG(0, ("vfs_fileid: Couldn't register custom debugging class!\n"));
+	} else {
+		DEBUG(10, ("vfs_fileid: Debug class number of 'fileid': %d\n", vfs_fileid_debug_level));
+	}
+
+	return ret;
+}