Add 'zfs mount' support

By design the zfs utility is supposed to handle mounting and unmounting
a zfs filesystem.  We could allow zfs to do this directly.  There are
system calls available to mount/umount a filesystem.  And there are
library calls available to manipulate /etc/mtab.  But there are a
couple very good reasons not to take this appraoch... for now.

Instead of directly calling the system and library calls to (u)mount
the filesystem we fork and exec a (u)mount process.  The principle
reason for this is to delegate the responsibility for locking and
updating /etc/mtab to (u)mount(8).  This ensures maximum portability
and ensures the right locking scheme for your version of (u)mount
will be used.  If we didn't do this we would have to resort to an
autoconf test to determine what locking mechanism is used.

The downside to using mount(8) instead of mount(2) is that we lose
the exact errno which was returned by the kernel.  The return code
from mount(8) provides some insight in to what went wrong but it
not quite as good.  For the moment this is translated as a best
guess in to a errno for the higher layers of zfs.

In the long term a shared library called libmount is under development
which provides a common API to address the locking and errno issues.
Once the standard mount utility has been updated to use this library
we can then leverage it.  Until then this is the only safe solution.

  http://www.kernel.org/pub/linux/utils/util-linux/libmount-docs/index.html

1 files changed, 94 insertions, 16 deletions

diff --git a/lib/libzfs/libzfs_mount.c b/lib/libzfs/libzfs_mount.c
index 9950bf967..75ce3676f 100644
--- a/lib/libzfs/libzfs_mount.c
+++ b/lib/libzfs/libzfs_mount.c
@@ -259,6 +259,82 @@ zfs_is_mountable(zfs_handle_t *zhp, char *buf, size_t buflen,
 }
 
 /*
+ * The filesystem is mounted by invoking the system mount utility rather
+ * than by the system call mount(2).  This ensures that the /etc/mtab
+ * file is correctly locked for the update.  Performing our own locking
+ * and /etc/mtab update requires making an unsafe assumption about how
+ * the mount utility performs its locking.  Unfortunately, this also means
+ * in the case of a mount failure we do not have the exact errno.  We must
+ * make due with return value from the mount process.
+ *
+ * In the long term a shared library called libmount is under development
+ * which provides a common API to address the locking and errno issues.
+ * Once the standard mount utility has been updated to use this library
+ * we can add an autoconf check to conditionally use it.
+ *
+ * http://www.kernel.org/pub/linux/utils/util-linux/libmount-docs/index.html
+ */
+
+static int
+do_mount(const char *src, const char *mntpt, char *opts)
+{
+	char *argv[8] = {
+	    "/bin/mount",
+	    "-t", MNTTYPE_ZFS,
+	    "-o", opts,
+	    (char *)src,
+            (char *)mntpt,
+	    (char *)NULL };
+	int rc;
+
+	/* Return only the most critical mount error */
+	rc = libzfs_run_process(argv[0], argv);
+	if (rc) {
+		if (rc & MOUNT_FILEIO)
+			return EIO;
+		if (rc & MOUNT_USER)
+			return EINTR;
+		if (rc & MOUNT_SOFTWARE)
+			return EPIPE;
+		if (rc & MOUNT_SYSERR)
+			return EAGAIN;
+		if (rc & MOUNT_USAGE)
+			return EINVAL;
+
+		return ENXIO; /* Generic error */
+	}
+
+	return 0;
+}
+
+static int
+do_unmount(const char *mntpt, int flags)
+{
+	char force_opt[] = "-f";
+	char lazy_opt[] = "-l";
+	char *argv[7] = {
+	    "/bin/umount",
+	    "-t", MNTTYPE_ZFS,
+	    NULL, NULL, NULL, NULL };
+	int rc, count = 3;
+
+	if (flags & MS_FORCE) {
+		argv[count] = force_opt;
+		count++;
+	}
+
+	if (flags & MS_DETACH) {
+		argv[count] = lazy_opt;
+		count++;
+	}
+
+	argv[count] = (char *)mntpt;
+	rc = libzfs_run_process(argv[0], argv);
+
+	return (rc ? EINVAL : 0);
+}
+
+/*
  * Mount the given filesystem.
  */
 int
@@ -268,9 +344,10 @@ zfs_mount(zfs_handle_t *zhp, const char *options, int flags)
 	char mountpoint[ZFS_MAXPROPLEN];
 	char mntopts[MNT_LINE_MAX];
 	libzfs_handle_t *hdl = zhp->zfs_hdl;
+	int rc;
 
 	if (options == NULL)
-		mntopts[0] = '\0';
+		(void) strlcpy(mntopts, MNTOPT_DEFAULTS, sizeof (mntopts));
 	else
 		(void) strlcpy(mntopts, options, sizeof (mntopts));
 
@@ -278,7 +355,12 @@ zfs_mount(zfs_handle_t *zhp, const char *options, int flags)
 	 * If the pool is imported read-only then all mounts must be read-only
 	 */
 	if (zpool_get_prop_int(zhp->zpool_hdl, ZPOOL_PROP_READONLY, NULL))
-		flags |= MS_RDONLY;
+		(void) strlcat(mntopts, "," MNTOPT_RO, sizeof (mntopts));
+
+	/*
+	 * Append zfsutil option so the mount helper allow the mount
+	 */
+	strlcat(mntopts, "," MNTOPT_ZFSUTIL, sizeof (mntopts));
 
 #ifdef HAVE_LIBSELINUX
 	if (is_selinux_enabled())
@@ -302,12 +384,9 @@ zfs_mount(zfs_handle_t *zhp, const char *options, int flags)
 
 	/*
 	 * Determine if the mountpoint is empty.  If so, refuse to perform the
-	 * mount.  We don't perform this check if MS_OVERLAY is specified, which
-	 * would defeat the point.  We also avoid this check if 'remount' is
-	 * specified.
+	 * mount.  We don't perform this check if 'remount' is specified.
 	 */
-	if ((flags & MS_OVERLAY) == 0 &&
-	    strstr(mntopts, MNTOPT_REMOUNT) == NULL &&
+	if (strstr(mntopts, MNTOPT_REMOUNT) == NULL &&
 	    !dir_is_empty(mountpoint)) {
 		zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
 		    "directory is not empty"));
@@ -316,20 +395,20 @@ zfs_mount(zfs_handle_t *zhp, const char *options, int flags)
 	}
 
 	/* perform the mount */
-	if (mount(zfs_get_name(zhp), mountpoint, MS_OPTIONSTR | flags,
-	    MNTTYPE_ZFS, NULL, 0, mntopts, sizeof (mntopts)) != 0) {
+	rc = do_mount(zfs_get_name(zhp), mountpoint, mntopts);
+	if (rc) {
 		/*
 		 * Generic errors are nasty, but there are just way too many
 		 * from mount(), and they're well-understood.  We pick a few
 		 * common ones to improve upon.
 		 */
-		if (errno == EBUSY) {
+		if (rc == EBUSY) {
 			zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
 			    "mountpoint or dataset is busy"));
-		} else if (errno == EPERM) {
+		} else if (rc == EPERM) {
 			zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
 			    "Insufficient privileges"));
-		} else if (errno == ENOTSUP) {
+		} else if (rc == ENOTSUP) {
 			char buf[256];
 			int spa_version;
 
@@ -342,7 +421,7 @@ zfs_mount(zfs_handle_t *zhp, const char *options, int flags)
 			    ZFS_PROP_VERSION), spa_version);
 			zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, buf));
 		} else {
-			zfs_error_aux(hdl, strerror(errno));
+			zfs_error_aux(hdl, strerror(rc));
 		}
 		return (zfs_error_fmt(hdl, EZFS_MOUNTFAILED,
 		    dgettext(TEXT_DOMAIN, "cannot mount '%s'"),
@@ -350,8 +429,7 @@ zfs_mount(zfs_handle_t *zhp, const char *options, int flags)
 	}
 
 	/* add the mounted entry into our cache */
-	libzfs_mnttab_add(hdl, zfs_get_name(zhp), mountpoint,
-	    mntopts);
+	libzfs_mnttab_add(hdl, zfs_get_name(zhp), mountpoint, mntopts);
 	return (0);
 }
 
@@ -361,7 +439,7 @@ zfs_mount(zfs_handle_t *zhp, const char *options, int flags)
 static int
 unmount_one(libzfs_handle_t *hdl, const char *mountpoint, int flags)
 {
-	if (umount2(mountpoint, flags) != 0) {
+	if (do_unmount(mountpoint, flags) != 0) {
 		zfs_error_aux(hdl, strerror(errno));
 		return (zfs_error_fmt(hdl, EZFS_UMOUNTFAILED,
 		    dgettext(TEXT_DOMAIN, "cannot unmount '%s'"),