Import Debian changes 1.18.0-1ubuntu1

network-manager (1.18.0-1ubuntu1) eoan; urgency=medium

  * Update to 1.18, merge on Debian, new version includes nwe support for
    policy routing rules and for VLAN filtering for Linux bridge.
  * Remaining Ubuntu changes
    - Use systemd-resolved instead of dnsmasq
    - debian/control:
      + Depend on isc-dhcp-client instead of recommends
      + Recommend network-manager-pptp
      + Suggest avahi-autoipd for IPv4LL support
    - debian/rules, debian/network-manager.postinst:
      + Don't restart NetworkManager on upgrade but recommend restarting
        the computer
    - debian/rules, debian/network-manager.postinst:
      + Don't install sysvinit scripts or migrate from sysvinit
    - debian/network-manager.postinst:
      + Don't add the netdev group.
      + drop in an empty override file for NetworkManager to manage all
        devices for upgrade from any version, as long as there is no
        netplan configuration yet.
    - debian/default-wifi-powersave-on.conf, debian/rules:
      + Install a config file to enable WiFi powersave
    - Enable build tests
    - Add autopkgtests
    - debian/source_network-manager.py, debian/network-manager.install,
      debian/network-manager.links: Add apport hook
    - Add network-manager-config-connectivity-ubuntu package
    - NetworkManager.conf: disable MAC randomization feature. There is no
      easy way for desktop users to disable this feature yet. And there are
      reports that it doesn't work well with some systems.
    - Update Vcs links to point to Ubuntu branch
    - Add patches. See patch descriptions for more details:
      + Provide-access-to-some-of-NM-s-interfaces-to-whoopsie.patch
      + Update-dnsmasq-parameters.patch
      + Disable-general-with-expect.patch
      + libnm-Check-self-still-NMManager-or-not.patch
      + dns-manager-don-t-merge-split-DNS-search-domains.patch (but disabled)
      + Read-system-connections-from-run.patch
    - debian/tests/urfkill-integration - don't stop/start network manager
    - Revert "Add Conflicts to network-manager-dev against deprecated libraries"
      This reverts commit b4acc5e03e2b821e1cccc69529bb70826c741942.  We're still
      building libnm-glib for now, so these packages have a use in Ubuntu.
  * Removed delta, not needed anymore
    - debian/network-manager.maintscript
      + Remove /etc/dbus-1/system.d/nm-ofono.conf

1 files changed, 0 insertions, 46 deletions

diff --git a/shared/systemd/src/basic/stat-util.c b/shared/systemd/src/basic/stat-util.c
index 686adaf1..c9837fa1 100644
--- a/shared/systemd/src/basic/stat-util.c
+++ b/shared/systemd/src/basic/stat-util.c
@@ -226,52 +226,6 @@ int fd_is_network_fs(int fd) {
         return is_network_fs(&s);
 }
 
-int fd_is_network_ns(int fd) {
-        struct statfs s;
-        int r;
-
-        /* Checks whether the specified file descriptor refers to a network namespace. On old kernels there's no nice
-         * way to detect that, hence on those we'll return a recognizable error (EUCLEAN), so that callers can handle
-         * this somewhat nicely.
-         *
-         * This function returns > 0 if the fd definitely refers to a network namespace, 0 if it definitely does not
-         * refer to a network namespace, -EUCLEAN if we can't determine, and other negative error codes on error. */
-
-        if (fstatfs(fd, &s) < 0)
-                return -errno;
-
-        if (!is_fs_type(&s, NSFS_MAGIC)) {
-                /* On really old kernels, there was no "nsfs", and network namespace sockets belonged to procfs
-                 * instead. Handle that in a somewhat smart way. */
-
-                if (is_fs_type(&s, PROC_SUPER_MAGIC)) {
-                        struct statfs t;
-
-                        /* OK, so it is procfs. Let's see if our own network namespace is procfs, too. If so, then the
-                         * passed fd might refer to a network namespace, but we can't know for sure. In that case,
-                         * return a recognizable error. */
-
-                        if (statfs("/proc/self/ns/net", &t) < 0)
-                                return -errno;
-
-                        if (s.f_type == t.f_type)
-                                return -EUCLEAN; /* It's possible, we simply don't know */
-                }
-
-                return 0; /* No! */
-        }
-
-        r = ioctl(fd, NS_GET_NSTYPE);
-        if (r < 0) {
-                if (errno == ENOTTY) /* Old kernels didn't know this ioctl, let's also return a recognizable error in that case */
-                        return -EUCLEAN;
-
-                return -errno;
-        }
-
-        return r == CLONE_NEWNET;
-}
-
 int path_is_temporary_fs(const char *path) {
         _cleanup_close_ int fd = -1;