WIP

Change-Id: Ia8d595095462c30a56b8a0b30c8d24632776e4cc

.gitignore

@@ -16,7 +16,6 @@ install-sh
 libtool
 ltmain.sh
 missing
-osmo-ggsn.spec
 stamp-h1
 INSTALL
 m4/
@@ -28,7 +27,7 @@ osmo-ggsn-*.tar*
 # debian
 debian/osmo-ggsn/
 debian/*.debhelper
-debian/libgtp1
+debian/libgtp*/
 debian/osmo-ggsn-dbg
 debian/*.log
 debian/autoreconf.*
@@ -69,3 +68,21 @@ tests/*/*_test
 tests/testsuite
 tests/testsuite.log
 tests/package.m4
+
+# manuals
+doc/manuals/*.html
+doc/manuals/*.svg
+doc/manuals/*.pdf
+doc/manuals/*__*.png
+doc/manuals/*.check
+doc/manuals/generated/
+doc/manuals/osmomsc-usermanual.xml
+doc/manuals/common
+doc/manuals/build
+doc/manuals/vty/ggsn_vty_reference.xml
+
+contrib/osmo-ggsn.spec
+/debian/gtp-echo-responder-dbg/
+/debian/gtp-echo-responder/
+/debian/osmo-ggsn-doc/
+/utils/gtp-echo-responder

Makefile.am

@@ -1,5 +1,14 @@
 ## Process this file with automake to produce Makefile.in
-SUBDIRS = lib gtp ggsn sgsnemu doc tests
+SUBDIRS = \
+	  lib \
+	  gtp \
+	  ggsn \
+	  sgsnemu \
+	  doc \
+	  contrib \
+	  utils \
+	  tests \
+	  $(NULL)
 
 pkgconfigdir = $(libdir)/pkgconfig
 pkgconfig_DATA = libgtp.pc
@@ -10,6 +19,17 @@ $(top_srcdir)/.version:
 dist-hook:
 	echo $(VERSION) > $(distdir)/.tarball-version
 
-EXTRA_DIST = git-version-gen .version README.md README.FreeBSD README.MacOSX
+EXTRA_DIST = \
+	     .version \
+	     README.FreeBSD \
+	     README.MacOSX \
+	     README.md \
+	     contrib/osmo-ggsn.spec.in \
+	     debian \
+	     git-version-gen \
+	     $(NULL)
+
+AM_DISTCHECK_CONFIGURE_FLAGS = \
+	--with-systemdsystemunitdir=$$dc_install_base/$(systemdsystemunitdir)
 
 @RELMAKE@

README.md

@@ -20,16 +20,17 @@ GIT Repository
 
 You can clone from the official osmo-ggsn.git repository using
 
-	git clone git://git.osmocom.org/osmo-ggsn.git
+	git clone https://gitea.osmocom.org/cellular-infrastructure/osmo-ggsn
 
-There is a cgit interface at http://git.osmocom.org/osmo-ggsn/
+There is a web interface at <https://gitea.osmocom.org/cellular-infrastructure/osmo-ggsn>
 
 Documentation
 -------------
 
-There currently is no other documentation other than the wiki on the
-homepage.  It would be great if somebody would work towards a user
-manual that can become part of the osmo-gsm-manuals project.
+The user manual and VTY reference are optionally built in PDF form
+as part of the build process. Find pre-rendered versions here:
+
+https://ftp.osmocom.org/docs/osmo-ggsn/master/
 
 Mailing List
 ------------
@@ -263,8 +264,7 @@ following:
 
 1. Install sgsnemu on a Linux Box. See under installation above.
 2. Connect your Linux box with sgsnemu installed to the GPRS core
-network. Use the same LAN switch as the one your SGSN is connected
-to. You also need a free IP address that can be used by sgsnemu.
+network. You also need a free IP address that can be used by sgsnemu.
 3. You need to configure networking in terms of interface address,
 subnet mask and default route. See the Linux Networking HOWTO for
 details.

TODO-RELEASE

@@ -7,3 +7,5 @@
 # If any interfaces have been added since the last public release: c:r:a + 1.
 # If any interfaces have been removed or changed since the last public release: c:r:0.
 #library	what		description / commit summary line
+libgtp		REMOVE		remove GTP cause defines of reserved values
+libgtpnl > 1.2.5		gtp_tunnel_set_family()

configure.ac

@@ -1,14 +1,15 @@
 # Process this file with autoconf to produce a configure script.
-AC_INIT(osmo-ggsn, m4_esyscmd([./git-version-gen .tarball-version]), osmocom-net-gprs@lists.osmocom.org)
+AC_INIT([osmo-ggsn],[m4_esyscmd(./git-version-gen .tarball-version)],[osmocom-net-gprs@lists.osmocom.org])
 AC_CONFIG_SRCDIR([gtp/gtp.c])
-AM_CONFIG_HEADER([config.h])
-#AC_CONFIG_HEADER([config.h])
+AC_CONFIG_HEADERS([config.h])
 
 dnl *This* is the root dir, even if an install-sh exists in ../ or ../../
 AC_CONFIG_AUX_DIR([.])
 
 AC_CONFIG_TESTDIR(tests)
-AC_CANONICAL_SYSTEM
+AC_CANONICAL_HOST
+
+CFLAGS="$CFLAGS -std=gnu11"
 
 dnl kernel style compile messages
 m4_ifdef([AM_SILENT_RULES], [AM_SILENT_RULES([yes])])
@@ -18,9 +19,13 @@ AC_PROG_CC
 AC_PROG_INSTALL
 AC_PROG_AWK
 AC_PROG_CPP
-AC_PROG_CXX
 LT_INIT
 
+dnl patching ${archive_cmds} to affect generation of file "libtool" to fix linking with clang
+AS_CASE(["$LD"],[*clang*],
+  [AS_CASE(["${host_os}"],
+     [*linux*],[archive_cmds='$CC -shared $pic_flag $libobjs $deplibs $compiler_flags $wl-soname $wl$soname -o $lib'])])
+
 dnl check for pkg-config (explained in detail in libosmocore/configure.ac)
 AC_PATH_PROG(PKG_CONFIG_INSTALLED, pkg-config, no)
 if test "x$PKG_CONFIG_INSTALLED" = "xno"; then
@@ -39,9 +44,9 @@ AC_SUBST(EXEC_LDFLAGS)
 
 
 case "${host}" in
-  i*86-*-linux-gnu*) 
+  i*86-*-linux-gnu*)
     EXEC_LDADD="" ;;
-  *solaris*) 
+  *solaris*)
     EXEC_LDADD="-lresolv -lsocket -lnsl" ;;
 esac
 
@@ -66,7 +71,7 @@ AC_ARG_ENABLE([gtp-linux],
 	[enable_gtp_linux="$enableval"], [enable_gtp_linux="no"])
 
 AS_IF([test "x$enable_gtp_linux" = "xyes"], [
-	PKG_CHECK_MODULES([LIBGTPNL], [libgtpnl >= 1.0.0])
+	PKG_CHECK_MODULES([LIBGTPNL], [libgtpnl >= 1.2.0])
 ])
 
 AM_CONDITIONAL([ENABLE_GTP_KERNEL], [test "$enable_gtp_linux" = "yes"])
@@ -76,8 +81,12 @@ AC_HEADER_STDC
 AC_HEADER_SYS_WAIT
 AC_CHECK_HEADERS([arpa/inet.h fcntl.h netdb.h netinet/in.h stdint.h stdlib.h string.h sys/ioctl.h sys/socket.h sys/time.h unistd.h])
 
-# Check for if header
-AC_CHECK_HEADERS([linux/if.h net/if.h])
+# Check for if header. Some versions of linux/if.h fail without sys/socket.h included beforehand:
+# see https://algorithmicallyrandom.blogspot.com/2012/07/error-on-including-include.html
+AC_CHECK_HEADERS([linux/if.h net/if.h], [], [], [#ifdef HAVE_SYS_SOCKET_H
+                                                # include <sys/socket.h>
+                                                # endif
+                                                ])
 
 # Check for tun header
 AC_CHECK_HEADERS([linux/if_tun.h net/if_tun.h])
@@ -124,10 +133,19 @@ AC_EGREP_HEADER(struct iphdr, netinet/ip.h,
  AC_DEFINE([HAVE_IPHDR])],
  AC_MSG_RESULT(no))
 
+# Address generation modes (enum) implemented in linux 3.17 (bc91b0f07ada5535427373a4e2050877bcc12218)
+# /proc/sys/net/ipv6/conf/${iface}/addr_gen_mode was added in linux 4.11 (d35a00b8e33dab7385f724e713ae71c8be0a49f4)
+AC_MSG_CHECKING(whether enum in6_addr_gen_mode.IN6_ADDR_GEN_MODE_NONE exists)
+AH_TEMPLATE(HAVE_IN6_ADDR_GEN_MODE_NONE)
+AC_EGREP_HEADER(IN6_ADDR_GEN_MODE_NONE, linux/if_link.h,
+[AC_MSG_RESULT(yes)
+ AC_DEFINE([HAVE_IN6_ADDR_GEN_MODE_NONE])],
+ AC_MSG_RESULT(no))
+
 # Checks for library functions.
 AC_PROG_GCC_TRADITIONAL
 # AC_FUNC_MALLOC
-# AC_FUNC_MEMCMP 
+# AC_FUNC_MEMCMP
 AC_CHECK_FUNCS([gethostbyname inet_ntoa memset select socket strdup strerror strtol])
 AC_CHECK_FUNCS(inet_aton inet_addr, break)
 
@@ -136,9 +154,101 @@ adl_FUNC_GETOPT_LONG
 
 AM_INIT_AUTOMAKE([foreign])
 
-PKG_CHECK_MODULES(LIBOSMOCORE, libosmocore >= 0.6.4)
-PKG_CHECK_MODULES(LIBOSMOVTY, libosmovty >= 0.3.0)
-PKG_CHECK_MODULES(LIBOSMOCTRL, libosmoctrl)
+PKG_CHECK_MODULES(LIBOSMOCORE, libosmocore >= 1.9.0)
+PKG_CHECK_MODULES(LIBOSMOGSM, libosmogsm >= 1.9.0)
+PKG_CHECK_MODULES(LIBOSMOVTY, libosmovty >= 1.9.0)
+PKG_CHECK_MODULES(LIBOSMOCTRL, libosmoctrl >= 1.9.0)
+
+AC_ARG_ENABLE(sanitize,
+	[AS_HELP_STRING(
+		[--enable-sanitize],
+		[Compile with address sanitizer enabled],
+	)],
+	[sanitize=$enableval], [sanitize="no"])
+if test x"$sanitize" = x"yes"
+then
+	CFLAGS="$CFLAGS -fsanitize=address -fsanitize=undefined"
+	CPPFLAGS="$CPPFLAGS -fsanitize=address -fsanitize=undefined"
+fi
+
+AC_ARG_ENABLE(werror,
+	[AS_HELP_STRING(
+		[--enable-werror],
+		[Turn all compiler warnings into errors, with exceptions:
+		 a) deprecation (allow upstream to mark deprecation without breaking builds);
+		 b) "#warning" pragmas (allow to remind ourselves of errors without breaking builds)
+		]
+	)],
+	[werror=$enableval], [werror="no"])
+if test x"$werror" = x"yes"
+then
+	WERROR_FLAGS="-Werror"
+	WERROR_FLAGS+=" -Wno-error=deprecated -Wno-error=deprecated-declarations"
+	WERROR_FLAGS+=" -Wno-error=cpp" # "#warning"
+	CFLAGS="$CFLAGS $WERROR_FLAGS"
+	CPPFLAGS="$CPPFLAGS $WERROR_FLAGS"
+fi
+
+# Generate manuals
+AC_ARG_ENABLE(manuals,
+	[AS_HELP_STRING(
+		[--enable-manuals],
+		[Generate manual PDFs [default=no]],
+	)],
+	[osmo_ac_build_manuals=$enableval], [osmo_ac_build_manuals="no"])
+AM_CONDITIONAL([BUILD_MANUALS], [test x"$osmo_ac_build_manuals" = x"yes"])
+AC_ARG_VAR(OSMO_GSM_MANUALS_DIR, [path to common osmo-gsm-manuals files, overriding pkg-config and "../osmo-gsm-manuals"
+	fallback])
+if test x"$osmo_ac_build_manuals" = x"yes"
+then
+	# Find OSMO_GSM_MANUALS_DIR (env, pkg-conf, fallback)
+	if test -n "$OSMO_GSM_MANUALS_DIR"; then
+		echo "checking for OSMO_GSM_MANUALS_DIR... $OSMO_GSM_MANUALS_DIR (from env)"
+	else
+		OSMO_GSM_MANUALS_DIR="$($PKG_CONFIG osmo-gsm-manuals --variable=osmogsmmanualsdir 2>/dev/null)"
+		if test -n "$OSMO_GSM_MANUALS_DIR"; then
+			echo "checking for OSMO_GSM_MANUALS_DIR... $OSMO_GSM_MANUALS_DIR (from pkg-conf)"
+		else
+			OSMO_GSM_MANUALS_DIR="../osmo-gsm-manuals"
+			echo "checking for OSMO_GSM_MANUALS_DIR... $OSMO_GSM_MANUALS_DIR (fallback)"
+		fi
+	fi
+	if ! test -d "$OSMO_GSM_MANUALS_DIR"; then
+		AC_MSG_ERROR("OSMO_GSM_MANUALS_DIR does not exist! Install osmo-gsm-manuals or set OSMO_GSM_MANUALS_DIR.")
+	fi
+
+	# Find and run check-depends
+	CHECK_DEPENDS="$OSMO_GSM_MANUALS_DIR/check-depends.sh"
+	if ! test -x "$CHECK_DEPENDS"; then
+		CHECK_DEPENDS="osmo-gsm-manuals-check-depends"
+	fi
+	if ! $CHECK_DEPENDS; then
+		AC_MSG_ERROR("missing dependencies for --enable-manuals")
+	fi
+
+	# Put in Makefile with absolute path
+	OSMO_GSM_MANUALS_DIR="$(realpath "$OSMO_GSM_MANUALS_DIR")"
+	AC_SUBST([OSMO_GSM_MANUALS_DIR])
+fi
+
+# https://www.freedesktop.org/software/systemd/man/daemon.html
+AC_ARG_WITH([systemdsystemunitdir],
+     [AS_HELP_STRING([--with-systemdsystemunitdir=DIR], [Directory for systemd service files])],,
+     [with_systemdsystemunitdir=auto])
+AS_IF([test "x$with_systemdsystemunitdir" = "xyes" -o "x$with_systemdsystemunitdir" = "xauto"], [
+     def_systemdsystemunitdir=$($PKG_CONFIG --variable=systemdsystemunitdir systemd)
+
+     AS_IF([test "x$def_systemdsystemunitdir" = "x"],
+   [AS_IF([test "x$with_systemdsystemunitdir" = "xyes"],
+    [AC_MSG_ERROR([systemd support requested but pkg-config unable to query systemd package])])
+    with_systemdsystemunitdir=no],
+   [with_systemdsystemunitdir="$def_systemdsystemunitdir"])])
+AS_IF([test "x$with_systemdsystemunitdir" != "xno"],
+      [AC_SUBST([systemdsystemunitdir], [$with_systemdsystemunitdir])])
+AM_CONDITIONAL([HAVE_SYSTEMD], [test "x$with_systemdsystemunitdir" != "xno"])
+
+AC_MSG_RESULT([CFLAGS="$CFLAGS"])
+AC_MSG_RESULT([CPPFLAGS="$CPPFLAGS"])
 
 AC_CONFIG_FILES([Makefile
                  doc/Makefile
@@ -148,12 +258,16 @@ AC_CONFIG_FILES([Makefile
                  lib/Makefile
                  intl/Makefile
                  po/Makefile
+                 utils/Makefile
                  sgsnemu/Makefile
+                 doc/manuals/Makefile
+                 contrib/Makefile
+                 contrib/systemd/Makefile
+                 contrib/osmo-ggsn.spec
                  tests/Makefile
                  tests/lib/Makefile
                  tests/gtp/Makefile
-                 libgtp.pc
-                 osmo-ggsn.spec])
+                 libgtp.pc])
 AC_OUTPUT
 
 echo "

contrib/Makefile.am

@@ -0,0 +1 @@
+SUBDIRS = systemd

contrib/jenkins.sh

@@ -1,5 +1,11 @@
 #!/usr/bin/env bash
 # jenkins build helper script for openbsc.  This is how we build on jenkins.osmocom.org
+#
+# environment variables:
+# * GTP: configure GTP tunneling Linux kernel (values: "--enable-gtp-linux" or "--disable-gtp-linux")
+# * WITH_MANUALS: build manual PDFs if set to "1"
+# * PUBLISH: upload manuals after building if set to "1" (ignored without WITH_MANUALS = "1")
+#
 
 if ! [ -x "$(command -v osmo-build-dep.sh)" ]; then
 	echo "Error: We need to have scripts/osmo-deps.sh from http://git.osmocom.org/osmo-ci/ in PATH !"
@@ -18,12 +24,22 @@ osmo-clean-workspace.sh
 
 mkdir "$deps" || true
 
+if [ "x$GTP" == "x--enable-gtp-linux" ]; then
+	osmo-build-dep.sh libgtpnl
+fi
 osmo-build-dep.sh libosmocore "" ac_cv_path_DOXYGEN=false
 
 verify_value_string_arrays_are_terminated.py $(find . -name "*.[hc]")
 
 export PKG_CONFIG_PATH="$inst/lib/pkgconfig:$PKG_CONFIG_PATH"
 export LD_LIBRARY_PATH="$inst/lib"
+export PATH="$inst/bin:$PATH"
+
+# Additional configure options and depends
+CONFIG=""
+if [ "$WITH_MANUALS" = "1" ]; then
+	CONFIG="--enable-manuals"
+fi
 
 set +x
 echo
@@ -35,8 +51,13 @@ set -x
 
 cd "$base"
 autoreconf --install --force
-./configure
+./configure --enable-sanitize --enable-werror $GTP $CONFIG
 $MAKE $PARALLEL_MAKE
-$MAKE distcheck
+DISTCHECK_CONFIGURE_FLAGS="$CONFIG" $MAKE $PARALLEL_MAKE distcheck
 
+if [ "$WITH_MANUALS" = "1" ] && [ "$PUBLISH" = "1" ]; then
+	make -C "$base/doc/manuals" publish
+fi
+
+$MAKE $PARALLEL_MAKE maintainer-clean
 osmo-clean-workspace.sh

contrib/osmo-ggsn.service

@@ -1,13 +0,0 @@
-[Unit]
-Description=OsmoGGSN
-After=networking.service
-
-[Service]
-Type=simple
-Restart=always
-ExecStart=/usr/bin/osmo-ggsn -c /etc/osmocom/osmo-ggsn.cfg -f
-RestartSec=2
-RestartPreventExitStatus=1
-
-[Install]
-WantedBy=multi-user.target

contrib/osmo-ggsn.spec.in

@@ -0,0 +1,137 @@
+#
+# spec file for package osmo-ggsn
+#
+# Copyright (c) 2018 SUSE LINUX GmbH, Nuernberg, Germany.
+#
+# All modifications and additions to the file contributed by third parties
+# remain the property of their copyright owners, unless otherwise agreed
+# upon. The license for this file, and modifications and additions to the
+# file, is the same license as for the pristine package itself (unless the
+# license for the pristine package is not an Open Source License, in which
+# case the license is the MIT License). An "Open Source License" is a
+# license that conforms to the Open Source Definition (Version 1.9)
+# published by the Open Source Initiative.
+
+## Disable LTO for now since it breaks compilation of the tests
+## https://osmocom.org/issues/4114
+%define _lto_cflags %{nil}
+
+Name:           osmo-ggsn
+Version:        @VERSION@
+Release:        0
+Summary:        GPRS Support Node
+License:        GPL-2.0-only AND LGPL-2.1-or-later
+Group:          Productivity/Telephony/Servers
+URL:            https://osmocom.org/projects/openggsn
+Source:         %{name}-%{version}.tar.xz
+BuildRequires:  libtool >= 2
+BuildRequires:  pkgconfig >= 0.20
+%if 0%{?suse_version}
+BuildRequires:  systemd-rpm-macros
+%endif
+BuildRequires:  pkgconfig(libgtpnl) >= 1.2.0
+BuildRequires:  pkgconfig(libosmocore) >= 1.9.0
+BuildRequires:  pkgconfig(libosmoctrl) >= 1.9.0
+BuildRequires:  pkgconfig(libosmovty) >= 1.9.0
+Obsoletes:      openggsn
+%{?systemd_requires}
+
+%description
+Osmo-GGSN is a C-language implementation of a GGSN (Gateway GPRS
+Support Node), a core network element of ETSI/3GPP cellular networks
+such as GPRS, EDGE, UMTS or HSPA.
+
+%package -n libgtp6
+Summary:        Library implementing GTP between SGSN and GGSN
+License:        GPL-2.0-only
+Group:          System/Libraries
+
+%description -n libgtp6
+libgtp implements the GPRS Tunneling Protocol between SGSN and GGSN.
+
+%package -n libgtp-devel
+Summary:        Development files for the GTP library
+License:        GPL-2.0-only
+Group:          Development/Libraries/C and C++
+Requires:       libgtp6 = %{version}
+
+%description -n libgtp-devel
+libgtp implements the GPRS Tunneling Protocol between SGSN and GGSN.
+
+This subpackage contains libraries and header files for developing
+applications that want to make use of libgtp.
+
+%package -n gtp-echo-responder
+Summary:        Small program answering GTP ECHO Request with GTP ECHO Response
+License:        MIT
+Group:          System/Libraries
+
+%description -n gtp-echo-responder
+Small program answering GTP ECHO Request with GTP ECHO Response for both GTPCv1
+and GTPCv2.
+
+%prep
+%setup -q
+
+%build
+echo "%{version}" >.tarball-version
+autoreconf -fi
+%configure \
+  --enable-gtp-linux \
+  --disable-static \
+  --docdir="%{_docdir}/%{name}" \
+  --with-systemdsystemunitdir=%{_unitdir} \
+  --includedir="%{_includedir}/%{name}"
+make %{?_smp_mflags} V=1
+
+%install
+%make_install
+find %{buildroot} -type f -name "*.la" -delete -print
+
+%check
+make %{?_smp_mflags} check || (find . -name testsuite.log -exec cat {} +)
+
+%if 0%{?suse_version}
+%pre
+%service_add_pre %{name}.service
+
+%post
+%service_add_post %{name}.service
+
+%preun
+%service_del_preun %{name}.service
+
+%postun
+%service_del_postun %{name}.service
+%endif
+
+%post   -n libgtp6 -p /sbin/ldconfig
+%postun -n libgtp6 -p /sbin/ldconfig
+
+%files
+%license COPYING
+%doc AUTHORS README.md
+%{_bindir}/osmo-ggsn
+%{_bindir}/sgsnemu
+%{_mandir}/man8/osmo-ggsn.8%{?ext_man}
+%{_mandir}/man8/sgsnemu.8%{?ext_man}
+%{_unitdir}/%{name}.service
+%dir %{_docdir}/%{name}/examples
+%{_docdir}/%{name}/examples/osmo-ggsn-kernel-gtp.cfg
+%{_docdir}/%{name}/examples/osmo-ggsn.cfg
+%{_docdir}/%{name}/examples/sgsnemu.conf
+%dir %{_sysconfdir}/osmocom
+%config(noreplace) %{_sysconfdir}/osmocom/osmo-ggsn.cfg
+
+%files -n libgtp6
+%{_libdir}/libgtp.so.6*
+
+%files -n libgtp-devel
+%{_includedir}/%{name}/
+%{_libdir}/libgtp.so
+%{_libdir}/pkgconfig/libgtp.pc
+
+%files -n gtp-echo-responder
+%{_bindir}/gtp-echo-responder
+
+%changelog

contrib/systemd/Makefile.am

@@ -0,0 +1,10 @@
+EXTRA_DIST = \
+	osmo-ggsn.service \
+	ggsn.network \
+	apn0.netdev \
+	$(NULL)
+
+if HAVE_SYSTEMD
+systemdsystemunit_DATA = \
+  osmo-ggsn.service
+endif

contrib/systemd/apn0.netdev

@@ -0,0 +1,7 @@
+[NetDev]
+Name=apn0
+Kind=tun
+
+[Tun]
+User=username
+Group=username

contrib/systemd/ggsn.network

@@ -0,0 +1,6 @@
+[Match]
+Name=apn0
+
+[Network]
+Address=192.168.7.1/24
+IPMasquerade=yes

contrib/systemd/osmo-ggsn.service

@@ -0,0 +1,16 @@
+[Unit]
+Description=OsmoGGSN
+After=network-online.target
+Wants=network-online.target
+
+[Service]
+Type=simple
+Restart=always
+StateDirectory=osmocom
+WorkingDirectory=%S/osmocom
+ExecStart=/usr/bin/osmo-ggsn -c /etc/osmocom/osmo-ggsn.cfg
+RestartSec=2
+RestartPreventExitStatus=1
+
+[Install]
+WantedBy=multi-user.target

debian/changelog

@@ -1,3 +1,507 @@
+osmo-ggsn (1.11.0) unstable; urgency=medium
+
+  [ Daniel Willmann ]
+  * gtp: Add net GTP cause values and a function to check for success
+
+ -- Oliver Smith <osmith@sysmocom.de>  Tue, 28 Nov 2023 13:38:29 +0100
+
+osmo-ggsn (1.10.2) unstable; urgency=medium
+
+  [ Vadim Yanitskiy ]
+  * Do not hard-code -g and -O2 in CFLAGS
+  * tests: use -no-install libtool flag to avoid ./lt-* scripts
+
+  [ Oliver Smith ]
+  * doc/manuals/chapters/configuration: fix typo
+  * debian: set compat level to 10
+  * systemd: depend on networking-online.target
+  * README: update documentation section
+  * doc: running: update kernel-gtp limitations
+  * lib/in46_addr: add in46a_from_gsna
+
+ -- Pau Espin Pedrol <pespin@sysmocom.de>  Tue, 12 Sep 2023 14:36:10 +0200
+
+osmo-ggsn (1.10.1) unstable; urgency=medium
+
+  [ Oliver Smith ]
+  * debian/libgtp6.shlibs: new file
+
+  [ Vadim Yanitskiy ]
+  * lib/icmpv6.h: fix struct icmpv6_{radv_hdr,opt_prefix}
+  * gtp/gsn.c: fix 'No newline at end of file'
+  * gtp: use OSMO_ASSERT() in gtp_new()
+
+ -- Vadim Yanitskiy <vyanitskiy@sysmocom.de>  Mon, 27 Feb 2023 22:35:47 +0700
+
+osmo-ggsn (1.10.0) unstable; urgency=medium
+
+  [ Max ]
+  * Set working directory in systemd service file
+  * Ignore .deb build byproducts
+  * ctrl: take both address and port from vty config
+
+  [ Pau Espin Pedrol ]
+  * cosmetic: gtp: Fix typo in comment
+  * Split gsn_t related APIs out of gtp.{c,h}
+  * Use rate_ctr for gsn_t available_counters
+  * ggsn: Introduce tdef and make it configurable over VTY
+  * gtp: Introduce VTY configurable GTP timer X3
+  * Fix typos in comments and VTY descriptions
+
+  [ arehbein ]
+  * osmo-ggsn: Transition to use of 'telnet_init_default'
+
+ -- Pau Espin Pedrol <pespin@sysmocom.de>  Tue, 07 Feb 2023 14:29:48 +0100
+
+osmo-ggsn (1.9.0) unstable; urgency=medium
+
+  [ Pau Espin Pedrol ]
+  * tests: in46a_test: Make coverity happy when calling in46a_from_eua
+  * vty: Fix cmd 'no echo-interval' doing nothing
+  * libgtp: Fix ggsn crash if pdp alloc array is full (PDP_MAX)
+  * libgtp: Define retransmit QUEUE_SIZE relative to PDP_MAX (increase)
+  * gtp: Use switch statement in gtp_create_pdp_ind()
+  * gtp: Log detection of rx duplicate
+  * gtp: Small log improvements in gtp_create_pdp_ind()
+  * gtp: Specify retrans queue name & seqnum in log lines
+  * gtp: Log retrans queue register&free entries
+  * gtp: Fix typo in comment
+  * pco.h: Fix typo in reference to spec
+
+  [ Vadim Yanitskiy ]
+  * tests: use 'check_PROGRAMS' instead of 'noinst_PROGRAMS'
+
+  [ Harald Welte ]
+  * update git URLs (git -> https; gitea)
+
+ -- Pau Espin Pedrol <pespin@sysmocom.de>  Tue, 28 Jun 2022 17:48:22 +0200
+
+osmo-ggsn (1.8.0) unstable; urgency=medium
+
+  [ Oliver Smith ]
+  * doc/examples/Makefile.am: add sgsnemu.conf
+  * doc/examples/osmo-ggsn-kernel-gtp.cfg: new file
+  * doc/manuals: describe GTP-U kernel module
+  * gitignore: add ggsn_vty_reference.xml
+
+  [ Harald Welte ]
+  * Don't install osmo-ggsn-kernel-gtp.cfg to /etc/osmocom/
+  * Don't install sgsnemu.conf to /etc/osmocom/
+  * ggsn: Reject PDP CTX ACT for static IP addresses
+  * vty: Inform user that static IP addresses are not supported
+
+  [ Pau Espin Pedrol ]
+  * gtp: Update teic_confirmed only on resp success
+  * gtp: Rework parsing logic of UpdatePdpCtxResponse
+  * ggsn: Improve logging on incoming DL data packets
+  * gtp: Improve logging of failing pdp ctx resolution from TEI/TID
+  * cosmetic: gtpie.c: Fix trailing whitespace
+  * gtp: constify pointer arg
+  * gtp: Support tx/rx RAN Information Relay message
+  * ggsn: Log tun fd write errors
+  * ggsn: Fix heap-use-after-free during Recovery without associated PDP
+  * cosmetic: configure.ac: Fix tabulation in line
+  * Introduce program gtp-echo-responder
+  * gtp_echo_responder: report invalid chars present in node-feautres cmdline arg as error
+
+ -- Pau Espin Pedrol <pespin@sysmocom.de>  Tue, 16 Nov 2021 13:49:16 +0100
+
+osmo-ggsn (1.7.1) unstable; urgency=medium
+
+  [ Harald Welte ]
+  * main: add --vty-ref-mode, use vty_dump_xml_ref_mode()
+  * manuals: generate vty reference xml at build time
+
+ -- Pau Espin Pedrol <pespin@sysmocom.de>  Tue, 23 Feb 2021 17:31:24 +0100
+
+osmo-ggsn (1.7.0) unstable; urgency=medium
+
+  [ Vadim Yanitskiy ]
+  * debian/control: change maintainer to the Osmocom team / mailing list
+
+  [ Pau Espin Pedrol ]
+  * configure.ac: Fix trailing whitespace
+  * doc: Update VTY reference xml file
+  * Support setting rt-prio and cpu-affinity mask through VTY
+  * contrib/jenkins: Enable parallel make in make distcheck
+  * ggsn: generate coredump and exit upon SIGABRT received
+  * tests: Explicitly drop category from log
+  * tests: Replace deprecated API log_set_print_filename
+
+  [ Keith ]
+  * Fix vty PDP lookups by IMSI
+  * Prevent Crash in show pdp-context from vty
+  * Minor: remove code duplication
+  * Use imsi_str2gtp() in sgsnemu
+  * sgsnemu: relax check on length of IMSI cmdline arg.
+  * GTP: Replace recently introduced imsi_str2gtp()
+
+  [ Harald Welte ]
+  * Use OSMO_FD_* instead of deprecated BSC_FD_*
+  * gtp-kernel: Remove duplicate #include section
+  * gtp-kernel: don't #include libmnl headers
+
+  [ Oliver Smith ]
+  * contrib/jenkins: don't build osmo-gsm-manuals
+  * configure.ac: set -std=gnu11
+  * apn_start: avoid segfault if missing tun-device
+  * .gitignore: ignore debian/libgtp*
+  * deb/rpm: build with --enable-gtp-linux
+
+ -- Pau Espin Pedrol <pespin@espeweb.net>  Tue, 23 Feb 2021 13:34:39 +0100
+
+osmo-ggsn (1.6.0) unstable; urgency=medium
+
+  [ Pau Espin Pedrol ]
+  * cosmetic: Fix comment typo
+  * netns: Improve error checking
+  * sgsnemu: cmdline: Drop unused function cmdline_parser_params_create()
+  * sgsnemu: Pass array of in64_addr to in46a_from_eua()
+  * sgsnemu: Rename sgsnemu's libgtp cb_conf
+  * sgsnemu: Set its default loglevel category to INFO
+  * Move icmpv6 and checksum files from ggsn/ dir to lib/
+  * netdev_addaddr6: Use prefixlen arg
+  * sgsnemu: Avoid adding extra autogenerated local link ipv6 addr to tun iface
+  * sgsnemu: Fix ping transmitted statistics output
+  * cosmetic: icmpv6.c: fix typo in comment
+  * icmpv6.c: Mark internal function as static
+  * sgsnemu: Get rid of duplicated options.destaddr
+  * sgsnemu: Get rid of duplicated options.net
+  * sgsnemu: tun_addaddr: Don't set local addr as dstaddr
+  * icmpv6.c: Move code generating ipv6 hdr to its own function
+  * Rename netdev_*route to end in route4
+  * sgsnemu: Fix build/run against linux < 4.11 (no sysctl addr_gen_mode support)
+  * sgsnemu: Handle IPv6 SLAAC in tun iface manually
+  * sgsnemu: Implement ping on IPv6 APNs
+  * sgsnemu: Fix assumption ipv6 Interface-Identifier of public addr == announced Prefix
+  * gtp: queue_test: Fix printf gcc warn under ARM
+
+  [ Andreas Schultz ]
+  * add Linux network namespace support for TUN device
+
+  [ Vadim Yanitskiy ]
+  * lib/netns: fix open_ns(): return fd from open()
+
+  [ Philipp Maier ]
+  * doc: do not use random ip address for dns in default conf
+  * doc: use 127.0.0.2 instead of 127.0.0.6 as bind ip.
+  * debug: use LOGL_NOTICE instead of LOGL_DEBUG
+
+  [ Eric ]
+  * configure.ac: fix libtool issue  with clang and sanitizer
+
+  [ Harald Welte ]
+  * lib/netns.c: Add comments to the code, including doxygen API docs
+  * lib/netns: OSMO_ASSERT() if user doesn't call init_netns()
+  * lib/netns: Fix up error paths
+  * example config: use RFC1918 addresses for GGSN pools
+
+  [ Dmitri Kalashnik ]
+  * sgsnemu: use real tun device name after the device is up.
+
+  [ Oliver Smith ]
+  * osmo-ggsn.spec.in: remove
+  * contrib: import RPM spec
+  * contrib: integrate RPM spec
+  * Makefile.am: EXTRA_DIST: debian, contrib/*.spec.in
+
+ -- Harald Welte <laforge@osmocom.org>  Thu, 13 Aug 2020 12:26:20 +0200
+
+osmo-ggsn (1.5.0) unstable; urgency=medium
+
+  [ Jan Engelhardt ]
+  * build: switch AC_CANONICAL_TARGET for AC_CANONICAL_HOST
+
+  [ Pau Espin Pedrol ]
+  * libgtp: Remove packets in tx queue belonging pdp being freed
+  * libgtp: announce pdp ctx deletion upon CreatePdpCtx being rejected
+  * Introduce in46a_is_v{4,6}() helpers
+  * ggsn: Move PCO handling code into its own file
+  * in46_addr: Improve in46a_ntop documentation
+  * ggsn_vty.c: Fix wrong use of in46a_from_eua, print IPv6 euas
+  * ggsn: Split application lifecycle related code into ggsn_main.c
+  * Move pdp_get_peer_ipv() to lib/util.*
+  * gtp-kernel.c: Fix wrong use of in46a_from_eua, print IPv6 euas
+  * Introduce LOGTUN log helper
+  * ggsn_vty.c: Avoid printing duplicates for pdp context with v4v6 EUAs
+  * pdp: constify param in pdp_count_secondary()
+  * ggsn_vty.c: Improve output of VTY show pdp-context
+  * doc: Update vty reference xml file
+  * libgtp: Introduce cb_recovery3
+  * ggsn: Implement echo req/resp and recovery
+  * cosmetic: fix formatting in if line
+  * gtp: Log msg retransmits and timeouts
+  * cosmetic: gtp: Drop commented out code calling pdp_freepdp()
+  * cosmetic: gtp: Improve documentation of gtp_delete_context_req2()
+  * ggsn: rx DeletePdpReq confirmation: Improve documentation and use gtp_freepdp()
+  * gtp: Manage queue timers internally
+  * ggsn, sgsnemu: Drop use of no-op deprecated gtp_retrans* APIs
+
+  [ Vadim Yanitskiy ]
+  * gtp_update_pdp_ind(): fix NULL-pointer dereference
+  * gtp_error_ind_conf(): fix: guard against an unknown GTP version
+  * gtp/gtp.c: cosmetic: use get_tid() where we need TID
+  * manuals/configuration.adoc: fix Network Address without prefix length
+  * manuals/configuration.adoc: fix IPv4 address mismatch in <<ggsn_no_root>>
+  * contrib/systemd: add systemd-networkd examples from manuals
+
+  [ Harald Welte ]
+  * sgsnemu: Fix null-pointer format string argument
+  * manual: Fix copy+paste error
+
+ -- Pau Espin Pedrol <pespin@sysmocom.de>  Thu, 02 Jan 2020 20:39:39 +0100
+
+osmo-ggsn (1.4.0) unstable; urgency=medium
+
+  [ Max ]
+  * Don't return error on normal shutdown
+
+  [ Harald Welte ]
+  * process_pco() const-ify 'apn' argument
+  * ggsn: Remove magic numbers from pco_contains_proto()
+  * ggsn: const-ify input / read-only arguments of PCO related functions
+  * ggsn: Remove magic numbers from ipcp_contains_option()
+  * ggsn: Fix build_ipcp_pco() in presence of invalid IPCP content
+  * ggsn.c: Refactor PCO processing during PDP activation
+  * ggsn: Add minimalistic PAP support
+  * ggsn: More logging from PCO handling (e.g. in case of malconfiguration)
+  * sgsnemu: Fix format string argument count
+
+  [ Vadim Yanitskiy ]
+  * osmo-ggsn: fix VTY command for getting PDP contexts by APN
+  * osmo-ggsn: add VTY command to show PDP context by IPv4
+  * osmo-ggsn: check result of osmo_apn_to_str()
+  * osmo-ggsn: print requested / actual APN in PDP info
+  * osmo-ggsn: properly show subscriber's MSISDN in the VTY
+
+  [ Pau Espin Pedrol ]
+  * ggsn: Drop unused param force in apn_stop()
+  * gtp: Document spec reasoning drop of Rx DeleteCtxReq
+  * ggsn: Start gtp retrans timer during startup
+  * gtp: Take queue_resp into account to schedule retrans timer
+  * gtp: Fix typo dublicate->duplicate
+  * pdp: Introduce new API pdp_count_secondary
+  * gtp_create_pdp_ind: simplify code by reordering and compacting parsing
+  * gtp: Refactor code to use gtp_freepdp(_teardown) APIs
+  * cosmetic: gtp: Document free pdp ctx in non-teardown scenario
+  * gtp: Re-arrange free pdp ctx code in non-teardown scenario
+  * pdp: Drop unused code for haship
+  * cosmetic: gtp.h: Remove trailing whitespaces
+  * ggsn: Fix undefined behaviour shifting beyond sign bit
+  * gtp: Introduce new pdp APIs (and deprecate old ones) to support multiple GSN
+  * gtp: Make use of new libgtp APIs with multi-gsn support
+  * ggsn_vty_reference.xml: Update from last code changes
+  * ggsn: vty: Require ggsn param in <show pdp-context> cmd
+  * sgsnemu: Replace use of deprecated libgtp API pdp_newpdp with new one
+  * cosmetic: gtp: queue: remove trailing whitespace
+  * gtp: Add missing headers
+  * gtp: queue.c: Document queue APIs
+  * gtp: queue: Add unit test queue_test
+  * ggsn: Avoid unaligned mem access reading PCO proto id
+  * ggsn: Use structures instead of raw arrays when parsing ipcp_hdr
+  * configure.ac: Replace obosolete macro AC_CANONICAL_SYSTEM
+  * configure.ac: Use brackets in AC_INIT params
+  * configure.ac: Use prefered AC_CONFIG_HEADERS over AM_CONFIG_HEADER
+  * configure.ac: some versions of linux/if.h require including sys/socket.h
+  * sgsnemu: Fix unaligned pointer access during ip/icmp checksum
+  * Remove undefined param passed to {logging,osmo_stats}_vty_add_cmds
+  * Require libosmocore 1.1.0
+
+  [ Oliver Smith ]
+  * debian: create -doc subpackage with pdf manuals
+  * ggsn: Use gtp_delete_context_req2() everywhere
+  * contrib/jenkins.sh: run "make maintainer-clean"
+
+  [ Daniel Willmann ]
+  * manuals: Add script to regenerate vty/counter documentation
+
+ -- Pau Espin Pedrol <pespin@sysmocom.de>  Wed, 07 Aug 2019 21:28:30 +0200
+
+osmo-ggsn (1.3.0) unstable; urgency=medium
+
+  [ Pau Espin Pedrol ]
+  * ggsn: ctrl iface: listen on IP configured by VTY
+  * gtp: Log type name of unexpected signalling message
+  * gtp: Allow recv DEL CTX REQ in sgsn and DEL CTX RSP in ggsn
+  * gtp: Log ignore CTX DEL REQ due to no teardown and only 1 ctx active
+  * gtp: Add new API to avoid freeing pdp contexts during DEL CTX REQ
+  * gtp: Add new replacement cb_recovery2 for cb_recovery
+  * Install systemd services with autotools
+  * Install sample cfg file to /etc/osmocom
+
+  [ Stefan Sperling ]
+  * fix unaligned access in build_ipcp_pco()
+  * fix support for multiple IPCP in PDP protocol configuration options
+  * check ioctl() call return value in tun_new()
+  * fix allocation of ippool's hash table
+  * replace bogus memcpy() call in ippool_newip()
+  * initialize local variable addr in ippool_new()
+  * fix format string error in ippool_printaddr()
+  * fix a format string directives in queue_seqset()
+  * properly store IPv6 addresses in struct tun_t
+
+  [ Harald Welte ]
+  * debian/rules: Don't overwrite .tarball-version
+  * osmo-ggsn.cfg: Ensure well-formed config file example
+  * sgsnemu: Fix printing of tun device name
+  * ippool.c: Use "%td" format string for ptrdiff_t
+  * initial version of OsmoGGSN user manual
+  * OsmoGGSN: Add VTY reference manual
+  * GGSN: Document how 'ip tuntap' is used for non-root; call netdev 'apn0'
+  * vty-ref: Update URI of docbook 5.0 schema
+
+  [ Alexander Couzens ]
+  * libgtp: implement gtp_clear_queues to clear req/resp queue
+
+  [ Neels Hofmeyr ]
+  * Importing history from osmo-gsm-manuals.git
+  * refactor Makefile build rules, don't use the FORCE
+  * GGSN: don't say 'NITB'
+  * OsmoGGSN: more info on non-root operation / tun creation
+  * OsmoGGSN: multiple instances: mention GTP port
+  * OsmoGGSN: add Routing section for IP forward and masquerading
+  * OsmoGGSN: typo: priveleges
+  * OsmoGGSN VTY ref: prep: convert newlines to unix
+  * OsmoGGSN vty: update VTY reference
+  * OsmoGGSN: fix VTY additions' node IDs
+  * OsmoGGSN: update vty reference
+  * ggsn: update vty reference
+
+  [ Max ]
+  * Expand OsmoGGSN manual
+
+  [ Oliver Smith ]
+  * build manuals moved here from osmo-gsm-manuals.git
+  * Fix DISTCHECK_CONFIGURE_FLAGS override
+  * contrib/jenkins.sh: build and publish manuals
+  * contrib: fix makedistcheck with disabled systemd
+
+ -- Harald Welte <laforge@gnumonks.org>  Sun, 20 Jan 2019 21:34:22 +0100
+
+osmo-ggsn (1.2.2) unstable; urgency=medium
+
+  [ Vadim Yanitskiy ]
+  * ggsn_vty.c: fix: use CONFIG_NODE as parent by default
+
+  [ Philipp Maier ]
+  * ggsn: fix misinterpreted length field in ipcp_contains_option()
+  * ggsn: make sure ipcp_option_hdr and and ipcp_hdr are packed
+
+ -- Pau Espin Pedrol <pespin@sysmocom.de>  Thu, 31 May 2018 12:44:54 +0200
+
+osmo-ggsn (1.2.1) unstable; urgency=medium
+
+  * debian/rules: Fix debian packaging after 1.2.0 release
+
+ -- Pau Espin Pedrol <pespin@sysmocom.de>  Fri, 04 May 2018 12:19:58 +0200
+
+osmo-ggsn (1.2.0) unstable; urgency=medium
+
+  [ Neels Hofmeyr ]
+  * fix compiler warnings: return 0 in main(), in 3 tests
+  * add --enable-sanitize config option
+  * sanitize build: ensure uint16/32 alignment in gtpie_test and in46a_test
+  * configure: add --enable-werror
+  * jenkins.sh: use --enable-werror configure flag, not CFLAGS
+
+  [ Harald Welte ]
+  * sgsnemu: Don't leak FILE handle in proc_read()
+  * sgsnemu: Fix format string in printing tun-device name
+  * sgsnemu: Make sure buffer has space for terminating-NUL
+  * sgsnemu: Free strings in error path
+  * gtp: Fix buffer overflow in imsi_gtp2str()
+  * gtp: Explicit OSMO_ASSERT to ensure pdp variable is set
+  * tun: Don't copy 16byte IPv6 address to 'struct in_addr'
+  * ippool: Correctly compute size of static pool
+  * remove unused argument to alloc_ippool_blacklist()
+  * factor out netdev_ip_local_get() from tun_ip_local_get()
+  * Properly NULL-out blacklist in alloc_ippool_blacklist()
+  * gtp_kernel: Change gtp_kernel_init() function signature
+  * gtp-kernel: Re-add support for kernel GTP-U acceleration
+  * gtp-kernel: Get rid of hard-coded kernel GTP device name
+  * gtp-kernel: shut down kernel GTP device in apn_down()
+  * gtp-kernel: Align logging for APN start in kernel-gtp case with that of TUN
+  * gtp-kernel: Avoid global state variable
+  * gtp-kernel: Make sure repeated calls to gtp_kernel_init() are safe
+  * gtp-kernel: proper cleanup in error path
+  * gtp-kernel: Get rid of SYS_ERR where not applicable
+  * gtp-kernel: Add function name to pdp_debug() function calls
+  * gtp-kernel: Add device nime in pdp_debug() log statements
+  * contrib/jenkins.sh: Allow jenkins job to specify if kernel GTP is used
+  * ggsn.c: Fix byte order of IPCP IPv4 DNS servers
+  * ggsn: Ignore PCO with length 0, don't abort processing
+  * README.md: Remove misleading sentence on sgsnemu
+  * Add talloc context introspection via VTY
+  * fix segfault in case of kernel gtp-u
+  * lib/tun.c: Generalize tun_sifflags() to netdev_sifflags
+  * lib/tun.c: generalize tun_*route() to netdev_*route()
+  * lib/tun.c: Generalize tun_{set,add}addr*() functions
+  * lib/tun: split generic network device related stuff to lib/netdev
+  * lib/netdev.c: Cosmetic changes (coding style / cleanups)
+  * ggsn: Don't explicitly use tun_setaddr() API anymore
+  * sgsnemu: Convert from tun_setaddr() to tun_addaddr()
+  * lib/tun: Remove tun_setaddr() API, as everyone is using tun_addaddr() now
+  * Move kernel GTP support from ggsn/ to lib/
+  * ggsn: don't use gtp_kernel_tunnel_{add,del}() for userspace tun
+
+  [ Pau Espin Pedrol ]
+  * ggsn_vty: Stop using deprecated API vty_install_default
+  * contrib/jenkins.sh: Enable Werror in C(PP)FLAGS
+  * examples: Add secondary ipv6 google DNS to osmo-ggsn.cfg
+  * tun_setaddr6: Fix log typo
+  * cosmetic: Reorder tun_addaddr to get rid of decl of tun_setaddr4
+  * ggsn.c: Print version of unhandled ip packet
+  * Remove unused empty src/Makefile.in
+  * tests: Split ipv6 specific tests into a new test group
+  * Add support for IPv4v6 End User Addresses
+  * contrib: jenkins.sh: Build libgtpnl as dep when building with gtp kernel support
+  * cosmetic: sgsnemu.c: Fix trailing whitespace
+  * ggsn.c: Improve logging info on link-local ipv6 addr not found
+  * tun.c: tun_addaddr: Fix segfault and wrong usage of tun_nlattr
+  * Set tun_addaddr ipv agnostic and add support for ipv6
+  * ggsn: Add 'ipv6 link-local' vty cmd
+  * ggsn_vty.c: Print ipv6 link-local cmd when writing config to file
+  * gtp.c: Fix trailing whitespace
+  * gtp.c: Determine GTP version from header
+  * gtp.c: Log unsupported GTP version number
+  * gtp/pdp: Fix trailing whitespace
+  * gtp/pdp: Remove unused APIs pdp_ntoeua pdp_euaton
+  * gtp.c: gtp_gpdu_ind: Convert ifelse to switch statement
+  * gtp.c: gtp_gpdu_ind: Early return to avoid use of uninitialized var
+  * gtp/gtp.c: Remove unused function char2ul_t
+  * gtp/gtp.c: Mark non exported functions as static
+  * gtp/gtp.c: Use uint8_t for version param in static functions
+  * ggsn: encaps_tun: Avoid forwarding packet if EUA is unassigned, fix crash
+  * ggsn: Validate packet src addr from MS
+  * ggsn: Parse PCO_IPCP
+  * ggsn: Parse PCO_IPCP for IPv4v6 pdp ctx
+  * ggsn: Print all addresses on successful pdp ctx creation
+  * ggsn.c: cb_tun_ind: Convert ifelse to switch statement
+  * ggsn.c: cb_tun_ind: log dst addr of packet without pdp ctx
+  * ggsn.c: cb_tun_ind: Don't drop packets targeting pdp ctx ll addr
+  * sgsnemu: Fix bad ptr during context deallocation
+  * sgsnemu: listen param is a host, not an interface
+  * use osmo_init_logging2
+
+  [ Max ]
+  * Log APN and tun names for packets
+  * Enable sanitize for CI tests
+  * Fix stow-enabled jenkins build failure
+  * Add GTP message names
+
+  [ Viktor Tsymbalyuk ]
+  * sgsnemu: sgsnemu stopped after recieving "Request accepted" from ggsn
+  * sgsnemu: created "pinghost" and "createif" modes for mutual exclusion
+  * sgsnemu: fix: no outgoing GTP-U in "createif" mode
+
+  [ Martin Hauke ]
+  * build: Remove AC_PROG_CXX, C++ is never used
+
+  [ Stefan Sperling ]
+  * remove the -f option from osmo-ggsn.service
+
+ -- Pau Espin Pedrol <pespin@sysmocom.de>  Thu, 03 May 2018 16:05:27 +0200
+
 osmo-ggsn (1.1.0) unstable; urgency=medium
 
   * libgtp: pdp.h: Addition of new tx_gpdu_seq struct member member

debian/compat

@@ -1 +1 @@
-9
+10

debian/control

@@ -1,16 +1,18 @@
 Source: osmo-ggsn
-Maintainer: Harald Welte <laforge@gnumonks.org>
+Maintainer: Osmocom team <openbsc@lists.osmocom.org>
 Section: net
 Priority: optional
-Build-Depends: debhelper (>= 9),
+Build-Depends: debhelper (>= 10),
                autotools-dev,
                pkg-config,
                libdpkg-perl, git,
                dh-autoreconf,
-               libosmocore-dev (>= 0.8.0)
+               libosmocore-dev (>= 1.9.0),
+               osmo-gsm-manuals-dev,
+               libgtpnl-dev (>= 1.2.0)
 Standards-Version: 3.9.6
-Vcs-Browser: http://git.osmocom.org/osmo-ggsn/
-Vcs-Git: git://git.osmocom.org/osmo-ggsn
+Vcs-Browser: https://gitea.osmocom.org/cellular-infrastructure/osmo-ggsn
+Vcs-Git: https://gitea.osmocom.org/cellular-infrastructure/osmo-ggsn
 Homepage: https://projects.osmocom.org/projects/openggsn
 
 Package: osmo-ggsn
@@ -22,7 +24,7 @@ Description: Osmocom Gateway GPRS Support Node (GGSN)
  operators as the interface between the Internet and the rest of the
  mobile network infrastructure.
 
-Package: libgtp2
+Package: libgtp6
 Architecture: any
 Multi-Arch: same
 Section: libs
@@ -36,12 +38,18 @@ Description: library implementing the GTP protocol between SGSN and GGSN
  This library is part of OsmoGGSN and implements the GTP protocol between
  SGSN (Serving GPRS support node) and GGSN.
 
+Package: gtp-echo-responder
+Architecture: any
+Depends: ${shlibs:Depends},
+         ${misc:Depends}
+Description: Small program answering GTP ECHO Request with GTP ECHO Response
+
 Package: libgtp-dev
 Architecture: any
 Multi-Arch: same
 Section: libdevel
 Depends: ${misc:Depends},
-         libgtp2 (= ${binary:Version})
+         libgtp6 (= ${binary:Version})
 Description: Development files for libgtp
  OsmoGGSN is a Gateway GPRS Support Node (GGSN). It is used by mobile
  operators as the interface between the Internet and the rest of the
@@ -54,18 +62,27 @@ Package: osmo-ggsn-dbg
 Section: debug
 Architecture: any
 Priority: extra
-Depends: ${shlibs:Depends}, ${misc:Depends}, libgtp2 (= ${binary:Version}), osmo-ggsn (= ${binary:Version})
+Depends: ${shlibs:Depends}, ${misc:Depends}, libgtp6 (= ${binary:Version}), osmo-ggsn (= ${binary:Version})
 Multi-Arch: same
 Description: Debug symbols for OsmoGGSN
  OsmoGGSN is a Gateway GPRS Support Node (GGSN). It is used by mobile
  operators as the interface between the Internet and the rest of the
  mobile network infrastructure.
 
+Package: gtp-echo-responder-dbg
+Section: debug
+Architecture: any
+Priority: extra
+Depends: ${shlibs:Depends}, ${misc:Depends}, gtp-echo-responder (= ${binary:Version})
+Multi-Arch: same
+Description: Debug symbols for gtp-echo-responder
+ Small program answering GTP ECHO Request with GTP ECHO Response.
+
 Package: libgtp-dbg
 Section: debug
 Architecture: any
 Priority: extra
-Depends: ${shlibs:Depends}, ${misc:Depends}, libgtp2 (= ${binary:Version})
+Depends: ${shlibs:Depends}, ${misc:Depends}, libgtp6 (= ${binary:Version})
 Multi-Arch: same
 Description: Debug symbols for OsmoGGSN
  OsmoGGSN is a Gateway GPRS Support Node (GGSN). It is used by mobile
@@ -74,3 +91,12 @@ Description: Debug symbols for OsmoGGSN
  .
  The library libgtp implements the GTP protocol between SGSN and GGSN
  and this package contains the development files for this library.
+
+Package: osmo-ggsn-doc
+Architecture: all
+Section: doc
+Priority: optional
+Depends: ${misc:Depends}
+Description: ${misc:Package} PDF documentation
+ Various manuals: user manual, VTY reference manual and/or
+ protocol/interface manuals.

debian/copyright

@@ -16,6 +16,11 @@ Files: lib/getopt.c
 Copyright: 1987-2001 Free Software Foundation, Inc.
 License: LGPL-2.1+
 
+Files: utils/gtp_echo_responder.c
+       utils/gtp_echo_responder_test.py
+Copyright: 2021 sysmocom - s.f.m.c. GmbH <info@sysmocom.de>
+License: MIT
+
 Files: debian/*
 Copyright: 2010-2017  Harald Welte <laforge@gnumonks.org>
            2016       Ruben Undheim <ruben.undheim@gmail.com>

debian/gtp-echo-responder.install

@@ -0,0 +1 @@
+/usr/bin/gtp-echo-responder

debian/libgtp6.shlibs

@@ -0,0 +1,2 @@
+# Most recent version of the package that added new symbols (OS#5318)
+libgtp 6 libgtp6 (>= 1.8.0)

debian/osmo-ggsn-doc.install

@@ -0,0 +1 @@
+usr/share/doc/osmo-ggsn-doc/*.pdf

debian/osmo-ggsn.examples

@@ -1,2 +1,3 @@
 doc/examples/osmo-ggsn.cfg
+doc/examples/osmo-ggsn-kernel-gtp.cfg
 doc/examples/sgsnemu.conf

debian/osmo-ggsn.install

@@ -1,3 +1,5 @@
+/etc/osmocom/osmo-ggsn.cfg
+/lib/systemd/system/osmo-ggsn.service
 /usr/bin/osmo-ggsn
 /usr/bin/sgsnemu
 /usr/share/man/man8/*

debian/osmo-ggsn.service

@@ -1 +0,0 @@
-../contrib/osmo-ggsn.service

debian/rules

@@ -16,8 +16,14 @@ export DEB_BUILD_MAINT_OPTIONS = hardening=+all
 
 override_dh_strip:
 	dh_strip -posmo-ggsn --dbg-package=osmo-ggsn-dbg
-	dh_strip -plibgtp2 --dbg-package=libgtp-dbg
+	dh_strip -plibgtp6 --dbg-package=libgtp-dbg
 
-override_dh_autoreconf:
-	echo $(VERSION) > .tarball-version
-	dh_autoreconf
+override_dh_auto_configure:
+	dh_auto_configure -- \
+		--enable-gtp-linux \
+		--with-systemdsystemunitdir=/lib/systemd/system \
+		--enable-manuals
+
+# Don't create .pdf.gz files (barely saves space and they can't be opened directly by most pdf readers)
+override_dh_compress:
+	dh_compress -X.pdf

doc/Makefile.am

@@ -4,4 +4,5 @@ EXTRA_DIST = $(man_MANS)
 
 SUBDIRS = \
 	examples \
+	manuals \
 	$(NULL)

doc/examples/Makefile.am

@@ -1,4 +1,13 @@
-CFG_FILES = find $(srcdir) -name '*.cfg*' | sed -e 's,^$(srcdir),,'
+OSMOCONF_FILES = \
+	osmo-ggsn.cfg \
+	$(NULL)
+
+osmoconfdir = $(sysconfdir)/osmocom
+osmoconf_DATA = $(OSMOCONF_FILES)
+
+EXTRA_DIST = $(OSMOCONF_FILES)
+
+CFG_FILES = find $(srcdir) -name '*.cfg' -o -name '*.conf' | sed -e 's,^$(srcdir),,'
 
 dist-hook:
 	for f in $$($(CFG_FILES)); do \

doc/examples/osmo-ggsn-kernel-gtp.cfg

@@ -0,0 +1,51 @@
+!
+! OpenGGSN (0.94.1-adac) configuration saved from vty
+!!
+!
+log stderr
+ logging filter all 1
+ logging color 1
+ logging print category 0
+ logging timestamp 0
+ logging level ip info
+ logging level tun info
+ logging level ggsn info
+ logging level sgsn notice
+ logging level icmp6 notice
+ logging level lglobal notice
+ logging level llapd notice
+ logging level linp notice
+ logging level lmux notice
+ logging level lmi notice
+ logging level lmib notice
+ logging level lsms notice
+ logging level lctrl notice
+ logging level lgtp info
+ logging level lstats notice
+ logging level lgsup notice
+ logging level loap notice
+ logging level lss7 notice
+ logging level lsccp notice
+ logging level lsua notice
+ logging level lm3ua notice
+ logging level lmgcp notice
+!
+stats interval 5
+!
+line vty
+ no login
+!
+ggsn ggsn0
+ gtp state-dir /tmp
+ gtp bind-ip 127.0.0.2
+ apn internet
+  gtpu-mode kernel-gtp
+  tun-device tun4
+  type-support v4
+  ip prefix dynamic 172.16.222.0/24
+  ip dns 0 8.8.8.8
+  ip dns 1 8.8.4.4
+  ip ifconfig 172.16.222.0/24
+  no shutdown
+ default-apn internet
+ no shutdown ggsn

doc/examples/osmo-ggsn.cfg

@@ -3,32 +3,32 @@
 !!
 !
 log stderr
-  logging filter all 1
-  logging color 1
-  logging print category 0
-  logging timestamp 0
-  logging level ip info
-  logging level tun info
-  logging level ggsn info
-  logging level sgsn notice
-  logging level icmp6 notice
-  logging level lglobal notice
-  logging level llapd notice
-  logging level linp notice
-  logging level lmux notice
-  logging level lmi notice
-  logging level lmib notice
-  logging level lsms notice
-  logging level lctrl notice
-  logging level lgtp info
-  logging level lstats notice
-  logging level lgsup notice
-  logging level loap notice
-  logging level lss7 notice
-  logging level lsccp notice
-  logging level lsua notice
-  logging level lm3ua notice
-  logging level lmgcp notice
+ logging filter all 1
+ logging color 1
+ logging print category 0
+ logging timestamp 0
+ logging level ip info
+ logging level tun info
+ logging level ggsn info
+ logging level sgsn notice
+ logging level icmp6 notice
+ logging level lglobal notice
+ logging level llapd notice
+ logging level linp notice
+ logging level lmux notice
+ logging level lmi notice
+ logging level lmib notice
+ logging level lsms notice
+ logging level lctrl notice
+ logging level lgtp info
+ logging level lstats notice
+ logging level lgsup notice
+ logging level loap notice
+ logging level lss7 notice
+ logging level lsccp notice
+ logging level lsua notice
+ logging level lm3ua notice
+ logging level lmgcp notice
 !
 stats interval 5
 !
@@ -37,15 +37,15 @@ line vty
 !
 ggsn ggsn0
  gtp state-dir /tmp
- gtp bind-ip 127.0.0.6
+ gtp bind-ip 127.0.0.2
  apn internet
   gtpu-mode tun
   tun-device tun4
   type-support v4
-  ip prefix dynamic 176.16.222.0/24
-  ip dns 0 192.168.100.1
-  ip dns 1 8.8.8.8
-  ip ifconfig 176.16.222.0/24
+  ip prefix dynamic 172.16.222.0/24
+  ip dns 0 8.8.8.8
+  ip dns 1 8.8.4.4
+  ip ifconfig 172.16.222.0/24
   no shutdown
  apn inet6
   gtpu-mode tun
@@ -53,18 +53,20 @@ ggsn ggsn0
   type-support v6
   ipv6 prefix dynamic 2001:780:44:2000:0:0:0:0/56
   ipv6 dns 0 2001:4860:4860::8888
+  ipv6 dns 1 2001:4860:4860::8844
   ipv6 ifconfig 2001:780:44:2000:0:0:0:0/56
   no shutdown
  apn inet46
   gtpu-mode tun
   tun-device tun46
   type-support v4v6
-  ip prefix dynamic 176.16.46.0/24
-  ip dns 0 192.168.100.1
-  ip dns 1 8.8.8.8
-  ip ifconfig 176.16.46.0/24
+  ip prefix dynamic 172.16.46.0/24
+  ip dns 0 8.8.8.8
+  ip dns 1 8.8.4.4
+  ip ifconfig 172.16.46.0/24
   ipv6 prefix dynamic 2001:780:44:2100:0:0:0:0/56
   ipv6 dns 0 2001:4860:4860::8888
+  ipv6 dns 1 2001:4860:4860::8844
   ipv6 ifconfig 2001:780:44:2100:0:0:0:0/56
   no shutdown
  default-apn internet

doc/manuals/Makefile.am

@@ -0,0 +1,24 @@
+EXTRA_DIST = osmoggsn-usermanual.adoc \
+    osmoggsn-usermanual-docinfo.xml \
+    osmoggsn-vty-reference.xml \
+    regen_doc.sh \
+    chapters \
+    vty
+
+if BUILD_MANUALS
+  ASCIIDOC = osmoggsn-usermanual.adoc
+  ASCIIDOC_DEPS = $(srcdir)/chapters/*.adoc
+  include $(OSMO_GSM_MANUALS_DIR)/build/Makefile.asciidoc.inc
+
+  VTY_REFERENCE = osmoggsn-vty-reference.xml
+
+  BUILT_REFERENCE_XML = $(builddir)/vty/ggsn_vty_reference.xml
+  $(builddir)/vty/ggsn_vty_reference.xml: $(top_builddir)/ggsn/osmo-ggsn
+	mkdir -p $(builddir)/vty
+	$(top_builddir)/ggsn/osmo-ggsn --vty-ref-xml > $@
+
+  include $(OSMO_GSM_MANUALS_DIR)/build/Makefile.vty-reference.inc
+
+  OSMO_REPOSITORY=osmo-ggsn
+  include $(OSMO_GSM_MANUALS_DIR)/build/Makefile.common.inc
+endif

doc/manuals/chapters/configuration.adoc

@@ -0,0 +1,335 @@
+== Configuring OsmoGGSN
+
+All configuration of OsmoGGSN is performed using the VTY. For more
+general information on the VTY interface, see <<vty>>.
+
+=== Configuring a virtual GGSN instance
+
+OsmoGGSN can run multiple GGSN instances inside one program/process.
+Each GGSN instance binds to its own transport-layer GTP IP address and
+has its own set of APNs and associated IP address pools + tun/gtp
+devices.
+
+In most usage cases, yo will only have a single GGSN instance inside
+your configuration file, like in below example:
+
+.Example: Single GGSN configuration section
+----
+ggsn ggsn0
+ gtp state-dir /tmp
+ gtp bind-ip 127.0.0.6
+ apn internet
+  gtpu-mode tun
+  tun-device tun4
+  type-support v4
+  ip prefix dynamic 176.16.222.0/24
+  ip dns 0 192.168.100.1
+  ip dns 1 8.8.8.8
+  ip ifconfig 176.16.222.0/24
+  no shutdown
+----
+
+
+==== Creating/Editing a GGSN instance
+
+Creating/Editing a GGSN instance can be done by the following sequence
+of VTY commands:
+
+----
+OsmoGGSN> enable <1>
+OsmoGGSN# configure terminal <2>
+OsmoGGSN(config)# ggsn ggsn0 <3>
+OsmoGGSN(config-ggsn)# <4>
+----
+<1> Change into privileged mode
+<2> Enter the interactive configuration mode
+<3> Create or edit the GGSN instance `ggsn0`. The name can be any ASCII
+    string, its significance is only to the local user.
+<4> Your prompt is now in the `ggsn` config node, where you can
+    configure the properties of this GGSN instance.
+
+NOTE:: After creating a new GGSN instance, it is in `shutdown` mode. See
+<<unshutdown_apn>> to take it out of shutdown, but make sure to configure it fully
+before taking it out of shutdown.
+
+==== Configuring a GGSN instance
+
+The following two mandatory configuration statements have to be given
+for every GGSN instance:
+
+----
+OsmoGGSN(config-ggsn)# gtp state-dir /var/lib/ggsn/ggsn0 <1>
+OsmoGGSN(config-ggsn)# gtp bind-ip 127.0.0.6 <2>
+----
+<1> Store the GSN restart state in the specified directory
+<2> Bind the GGSN instance to the specified local IPv4 address
+
+There are some further configuration statements that can be used at the
+GGSN node, some examples are given below.  For a full list, see the
+_OsmoGGSN VTY reference manual_ <<vty-ref-osmoggsn>>.
+
+----
+OsmoGGSN(config-ggsn)# default-apn foobar <1>
+----
+<1> Configure a default APN to be used if the user-requested APN is not
+    found.  The named APN must previously be configured
+
+
+==== Deleting a GGSN instance
+
+A GGSN instance can be removed like this
+
+.Example: Deleting a GGSN instance
+----
+OsmoGGSN> enable <1>
+OsmoGGSN# configure terminal <2>
+OsmoGGSN(config)# no ggsn ggsn0 <3>
+----
+<1> Change into privileged mode
+<2> Enter the interactive configuration mode
+<3> Delete the GGSN instance
+
+
+==== Taking a GGSN instance out of shutdown
+
+.Example: Taking a GGSN instance out of shutdown
+----
+OsmoGGSN> enable <1>
+OsmoGGSN# configure terminal <2>
+OsmoGGSN(config)# ggsn ggsn0 <3>
+OsmoGGSN(config-ggsn)# no shutdown ggsn <4>
+----
+<1> Change into privileged mode
+<2> Enter the interactive configuration mode
+<3> Enter the config node of the GGSN instance `ggsn0`
+<4> Take the GGSN instance out of shutdown
+
+
+==== Shutting a GGSN instance down
+
+If you would like to take a GGSN instance out of service, you can
+put it into shutdown mode.  This will make the entire GGSN unavailable
+to user traffic and permit you to e.g. reconfigure it before taking it
+out of shutdown again.
+
+.Example: Shutting down a GGSN instance
+----
+OsmoGGSN> enable <1>
+OsmoGGSN# configure terminal <2>
+OsmoGGSN(config)# ggsn ggsn0 <3>
+OsmoGGSN(config-ggsn)# shutdown ggsn <4>
+----
+<1> Change into privileged mode
+<2> Enter the interactive configuration mode
+<3> Enter the config node of the GGSN instance `ggsn0`
+<4> Shut down the GGSN instance
+
+
+=== Configuring an Access Point Name
+
+An Access Point Name (APN) represents a connection to an external packet
+data network, such as the public Internet or private corporate networsk.
+
+APNs are selected by terminals (MS/UE) when establishing PDP contexts.
+
+Each OsmoGGSN GGSN instance can have any number of APNs configured.
+Each APN is identified by a string name.
+
+==== Creating/Editing an APN
+
+.Example: Creating a new APN
+----
+OsmoGGSN> enable <1>
+OsmoGGSN# configure terminal <2>
+OsmoGGSN(config)# ggsn ggsn0 <3>
+OsmoGGSN(config-ggsn)# apn internet <4>
+OsmoGGSN(config-ggsn-apn)# <5>
+----
+<1> Change into privileged mode
+<2> Enter the interactive configuration mode
+<3> Enter the config node of the GGSN instance `ggsn0`
+<4> Create or Edit an APN called `internet`
+<5> Your prompt is now in the `ggsn` config node, where you can
+    configure the properties of this GGSN instance.
+
+NOTE:: The newly-create APN is created in `shutdown` mode. See <<unshutdown_apn>> to take it
+out of shutdown.
+
+
+==== Configuring an APN
+
+.Example: Configuring an APN
+----
+OsmoGGSN(config-ggsn-apn)# gtpu-mode tun <1>
+OsmoGGSN(config-ggsn-apn)# type-support v4 <2>
+OsmoGGSN(config-ggsn-apn)# ip prefix dynamic 176.16.222.0/24 <3>
+OsmoGGSN(config-ggsn-apn)# ip dns 0 192.168.100.1 <4>
+OsmoGGSN(config-ggsn-apn)# ip dns 1 8.8.8.8 <5>
+OsmoGGSN(config-ggsn-apn)# ip ifconfig 176.16.222.0/24 <6>
+----
+<1> Use the userspace GTP-U handling using a TUN device
+<2> Support (only) IPv4 Addresses
+<3> Specify the pool of dynamic IPv4 addresses to be allocated to PDP
+    contexts
+<4> Specify the primary DNS server to be provided using IPCP/PCO
+<5> Specify the secondary DNS server to be provided using IPCP/PCO
+<6> Request OsmoGGSN to configure the `tun4` device network/netmask
+
+NOTE:: If you use the optional `ip ifconfig` command to set the network
+device address/mask, OsmoGGSN must run with root or `CAP_NET_ADMIN`
+support.  It might be better to configure related tun devices at system
+startup and run OsmoGGSN as non-privileged user.  See <<ggsn_no_root>> for more
+details.
+
+
+==== Deleting an APN
+
+An APN configuration can be removed like this
+
+.Example: Deleting an APN
+----
+OsmoGGSN> enable <1>
+OsmoGGSN# configure terminal <2>
+OsmoGGSN(config)# ggsn ggsn0 <3>
+OsmoGGSN(config-ggsn)# no apn internet <4>
+----
+<1> Change into privileged mode
+<2> Enter the interactive configuration mode
+<3> Enter the config node of the GGSN instance `ggsn0`
+<4> Delete the APN `internet`
+
+[[unshutdown_apn]]
+==== Taking an APN out of shutdown
+
+In order to bring a deactived APN in `shutdown` state into active
+operation, use the `no shutdown` command at the APN node as explained in
+the following example:
+
+.Example: Taking an APN out of shutdown
+----
+OsmoGGSN> enable <1>
+OsmoGGSN# configure terminal <2>
+OsmoGGSN(config)# ggsn ggsn0 <3>
+OsmoGGSN(config-ggsn)# apn internet <4>
+OsmoGGSN(config-ggsn-apn)# no shutdown <5>
+----
+<1> Change into privileged mode
+<2> Enter the interactive configuration mode
+<3> Enter the config node of the GGSN instance `ggsn0`
+<4> Enter the config node of the APN `internet`
+<5> Take the APN out of shutdown
+
+
+==== Shutting an APN down
+
+If you would like to take an APN instance out of service, you can
+put it into shutdown mode.  This will make the APN unavailable
+to user traffic and permit you to e.g. reconfigure it before taking it
+out of shutdown again.
+
+.Example: Shutting down an APN
+----
+OsmoGGSN> enable <1>
+OsmoGGSN# configure terminal <2>
+OsmoGGSN(config)# ggsn ggsn0 <3>
+OsmoGGSN(config-ggsn)# apn internet <4>
+OsmoGGSN(config-ggsn-apn)# shutdown <5>
+----
+<1> Change into privileged mode
+<2> Enter the interactive configuration mode
+<3> Enter the config node of the GGSN instance `ggsn0`
+<4> Enter the config node of the APN `internet`
+<5> Shut down the APN
+
+[[ggsn_no_root]]
+=== Configuring for running without root privileges
+
+It's possible to run OsmoGGSN without root privileges if the tun devices are already configured.
+
+The interface creation + configuration must then happen before osmo-ggsn starting up.  This can be
+achieved by means such as
+
+* a custom shell script run as root before starting osmo-ggsn (e.g. as init script)
+* systemd .netdev and .network files, if your system is using systemd-networkd (see `networkctl status`).
+
+==== Manual TUN device creation / configuration
+
+If you chose to go for custom shell/init scripts, you may use the `ip` program which is the standard
+tool for network interface configuration on Linux, part of the `iproute2` package.  In order to
+create a tun device, you must call it like this:
+
+.Example: iproute2 command to create a tun device
+----
+# ip tuntap add dev apn0 mode tun user username group groupname
+----
+
+Where _username_ and _groupname_ correspond to the User and Group that will have ownership over the
+device, i.e. the privileges which you intend to run osmo-ggsn under, and _apn0_ will be the
+name of the network device created.  After creating the interface, you can configure its addresses
+using standard means like `ip addr add` or your distribution-specific utilities/tools
+to match the `ip prefix dynamic` config item, and activate the link, for example:
+
+----
+# ip addr add 192.168.7.1/24 dev apn0
+# ip link set apn0 up
+----
+
+==== systemd based TUN device creation+configuration
+
+If you want to have systemd take care of creating and configuring a tun device for you,
+you can use the below example config files.
+
+.Example: device config via systemd-networkd using apn0.netdev
+----
+[NetDev]
+Name=apn0 <1>
+Kind=tun
+
+[Tun]
+User=username <2>
+Group=username <3>
+----
+<1> The network interface name of the newly-created device
+<2> The username under which you will run OsmoGGSN
+<3> The group name under which you will run OsmoGGSN
+
+.Example: network settings via systemd-networkd using ggsn.network
+----
+[Match]
+Name=apn0 <1>
+
+[Network]
+Address=192.168.7.1/24 <2>
+IPMasquerade=yes <3>
+----
+<1> The network device name, which must match the one in the apn0.netdev unit file above
+<2> The local IP address configured on the device
+<3> Requesting systemd to configure IP masquerading for this interface.  Depending on your needs,
+    You may not want this if you have proper end-to-end routing set up, and want to have transparent
+    inbound IP access to your GPRS-attached devices.
+
+==== Config Changes
+
+With the tun device pre-configured in one of the ways outlined above, the main
+changes in your osmo-ggsn.cfg file are:
+
+* remove `ip ifconfig` directive,
+* make sure that `no shutdown` is present in the `apn` section as well as
+  `no shutdown ggsn` in the `ggsn` section.
+
+.Example: using externally configured tun device `apn0` as non-root
+----
+ggsn ggsn0
+ gtp state-dir /tmp
+ gtp bind-ip 127.0.0.6
+ apn internet
+  gtpu-mode tun
+  tun-device apn0
+  type-support v4
+  ip prefix dynamic 192.168.7.0/24
+  ip dns 0 192.168.100.1
+  ip dns 1 8.8.8.8
+  no shutdown
+ default-apn internet
+ no shutdown ggsn
+----

doc/manuals/chapters/overview.adoc

@@ -0,0 +1,145 @@
+[[chapter_introduction]]
+== Overview
+
+[[intro_overview]]
+=== About OsmoGGSN
+
+OsmoGGSN is a Free / Open Source Software implementation of the GPRS
+GGSN (Gateway GPRS support node) element in side the packet switched
+core network of 2G and 3G cellular networks.
+
+The GGSN function is the tunnel endpoint on the core network side,
+from where the external (IP) packet data network 
+
+=== Software Components
+
+==== GTP Implementation (libgtp)
+
+The OsmoGGSN source code includes a shared library implementation of
+the GTP protocol used on the GGSN-SGSN interface.  This library
+and associated header files are installed system-wide and are
+available to other programs/applications.
+
+In fact, libgtp is what the OsmoSGSN also uses for its use of GTP.
+
+==== sgsnemu
+
+In order to test OsmoGGSN without running a SGSN and other elements
+of a cellular network, there is a small command-line utility called
+*sgsnemu* which is able to simulate the customary operations of a SGSN
+towards the GGSN, such as a PDP Context Activation.
+
+*sgsnemu* can even be used for testing against other GGSNs, as the GTP
+protocol is standardized across implementations.
+
+==== osmo-ggsn
+
+*osmo-ggsn* is the actual name of the OsmoGGSN executable program.  It
+implements the GGSN functionality.  All parameters are set using the
+configuration file, by default located in *./osmo-ggsn.cfg*
+
+==== systemd service file
+
+In *contrib/osmo-ggsn.service* you can find a sample service file for
+OsmoGGSN which can be used with systemd.
+
+==== init script
+
+In *contrib/osmo-ggsn.init* you can find a sample init script to be used
+on systems with classic init process.
+
+=== Limitations
+
+OsmoGGSN supports both GTP0 (GSM 09.60) and GTP1 (3GPP 29.060). In the
+following tables the support of each individual message type is
+detailed. The numbers before each feature indicates the relevant
+section in the standard.
+
+==== GSM 09.60 (GTPv0)
+
+[options="header",cols="50%,15%,15%,15%,5%"]
+|===
+| Feature                | gtplib      | osmo-ggsn   | sgsnemu     | notes
+5+<|*7.4 Path Management Messages*
+|7.4.1 Echo Request      |Supported   |Supported   |Supported   |
+|7.4.2 Echo Response     |Supported   |Supported   |Supported   |
+|7.4.3 Version Not Supported      |Supported   |Supported   |Supported   |
+5+<| *7.5 Tunnel Management Messages*
+|7.5.1 Create PDP Context Request|Supported   |Supported   |Supported   |
+|7.5.2 Create PDP Context Response|Supported   |Supported   |Supported   |
+|7.5.3 Update PDP Context Request|Supported   |Supported   |Not         |
+|7.5.4 Update PDP Context Response|Supported   |Supported   |Not         |
+|7.5.5 Delete PDP Context Request|Supported   |Supported   |Supported   |
+|7.5.6 Delete PDP Context Response|Supported   |Supported   |Supported   |
+|7.5.7 Create AA PDP Context Request|Unsupported |Unsupported |Unsupported |
+|7.5.8 Create AA PDP Response|Unsupported |Unsupported |Unsupported |
+|7.5.9 Delete AA PDP Context Request|Unsupported |Unsupported |Unsupported |
+|7.5.10 Delete AA PDP Context Response|Unsupported |Unsupported |Unsupported |
+|7.5.11 Error Indication |Supported   |Supported   |Supported   |
+|7.5.12 PDU Notification Request|Unsupported |Unsupported |Unsupported |
+|7.5.13 PDU Notification Response|Unsupported |Unsupported |Unsupported |
+|7.5.14 PDU Notification Reject Request|Unsupported |Unsupported |Unsupported |
+|7.5.15 PDU Notification Reject Response|Unsupported |Unsupported |Unsupported |
+5+<| *7.6 Location Management Messages*
+|7.6.1 Send Routeing Information for GPRS Request|Unsupported |Unsupported |Not applicable  |
+|7.6.2 Send Routeing Information for GPRS Response|Unsupported |Unsupported |Not applicable  |
+|7.6.3 Failure Report Request|Unsupported |Unsupported |Not applicable |
+|7.6.3 Failure Report Response|Unsupported |Unsupported |Not applicable |
+|7.6.5 Note MS GPRS Present Request|Unsupported |Unsupported |Not applicable|
+|7.6.6 Note MS GPRS Present Response|Unsupported |Unsupported |Not applicable|
+5+<| *7.5 Mobility Management Messages*
+|7.5.1 Identification Request|Unsupported |Not applicable|Not applicable|
+|7.5.2 Identification Response|Unsupported |Not applicable|Not applicable |
+|7.5.3 SGSN Context Request|Unsupported |Not applicable|Not applicable|
+|7.5.4 SGSN Context Response|Unsupported |Not applicable|Not applicable|
+|7.5.5 SGSN Context Acknowledge|Unsupported |Not applicable|Not applicable|
+|===
+
+==== 3GPP 29.060 (GTPv1)
+
+[options="header",cols="50%,15%,15%,15%,5%"]
+|===
+|Feature                 |gtplib      |osmo-ggsn   |sgsnemu     |notes
+5+<|*7.2 Path Management Messages*
+|7.2.1 Echo Request      |Supported   |Supported   |Supported   |
+|7.2.2 Echo Response     |Supported   |Supported   |Supported   |
+|7.2.3 Version Not Supported|Supported   |Supported   |Supported   |
+|7.2.4 Extension Headers Notification|Supported   |Supported   |Supported   |
+5+<|*7.3 Tunnel Management Messages*
+|7.3.1 Create PDP Context Request|Supported   |Supported   |Supported   |1
+|7.3.2 Create PDP Context Response|Supported   |Supported   |Supported   |
+|7.3.3 Update PDP Context Request|Supported   |Supported   |Not applicable|1
+|7.3.4 Update PDP Context Response|Supported   |Supported   |Not applicable|
+|7.3.5 Delete PDP Context Request|Supported   |Supported   |Supported   |
+|7.3.6 Delete PDP Context Response|Supported   |Supported   |Supported   |
+|7.3.7 Error Indication  |Supported   |Supported   |Supported   |
+|7.3.8 PDU Notification Request|Unsupported |Unsupported |Unsupported |
+|7.3.9 PDU Notification Response|Unsupported |Unsupported |Unsupported |
+|7.3.10 PDU Notification Reject Request|Unsupported |Unsupported |Unsupported |
+|7.3.10 PDU Notification Reject Response|Unsupported |Unsupported |Unsupported |
+5+<|*7.4 Location Management Messages*
+|7.4.1 Send Routeing Information for GPRS Request|Unsupported |Unsupported |Not applicable  |
+|7.4.2 Send Routeing Information for GPRS Response|Unsupported |Unsupported |Not applicable  |
+|7.4.3 Failure Report Request|Unsupported |Unsupported |Not applicable|
+|7.4.3 Failure Report Response|Unsupported |Unsupported |Not applicable|
+|7.4.5 Note MS GPRS Present Request|Unsupported |Unsupported |Not applicable|
+|7.4.6 Note MS GPRS Present Response|Unsupported |Unsupported |Not applicable|
+5+<|*7.5 Mobility Management Messages*
+|7.5.1 Identification Request|Unsupported |Not applicable|Not applicable|
+|7.5.2 Identification Response|Unsupported |Not applicable |Not applicable|
+|7.5.3 SGSN Context Request|Unsupported |Not applicable|Not applicable|
+|7.5.4 SGSN Context Response|Unsupported |Not applicable |Not applicable|
+|7.5.5 SGSN Context Acknowledge|Unsupported |Not applicable|Not applicable|
+|7.5.6 Forward Relocation Request|Unsupported |Not applicable|Not applicable|
+|7.5.7 Forward Relocation Response|Unsupported |Not applicable|Not applicable|
+|7.5.8 Forward Relocation Complete|Unsupported |Not applicable|Not applicable|
+|7.5.9 Relocation Cancel Request|Unsupported |Not applicable|Not applicable|
+|7.5.10 Relocation Cancel Response|Unsupported |Not applicable|Not applicable|
+|7.5.11 Forward Relocation Complete     |Unsupported |Not applicable |Not applicable  |
+|7.5.12 Forward SRNS Context Acknowledge|Unsupported |Not applicable|Not applicable|
+|7.5.13 Forward SRNS Context|Unsupported |Not applicable|Not applicable|
+|===
+
+Notes
+
+1) The "Secondary PDP Context Activation Procedure" is not supported.

doc/manuals/chapters/running.adoc

@@ -0,0 +1,131 @@
+== Running OsmoGGSN
+
+The OsmoGGSN executable (`osmo-ggsn`) offers the following command-line
+arguments:
+
+=== SYNOPSIS
+
+*osmo-ggsn* [-h|-V] [-D] [-c 'CONFIGFILE']
+
+=== OPTIONS
+
+*-h, --help*::
+	Print a short help message about the supported options
+*-V, --version*::
+	Print the compile-time version number of the program
+*-D, --daemonize*::
+	Fork the process as a daemon into background.
+*-c, --config-file 'CONFIGFILE'*::
+	Specify the file and path name of the configuration file to be
+	used. If none is specified, use `osmo-ggsn.cfg` in the current
+	working directory.
+
+=== Routing
+
+Operating the OpenGGSN tun device naturally creates a network setup with
+multiple interfaces. Consider:
+
+* Typical Linux setups prevent forwarding of packets between separate
+  interfaces by default. To let subscribers reach the internet uplink from the
+  tun device, it may be required to enable IP forwarding.
+
+* Having a locally defined address range assigned to the tun device requires
+  either sensible routing for this address range, or that masquerading is
+  enabled to allow your single uplink IP address to "proxy" for the tun.
+
+These are decisions to be made on a network administration level.
+
+In a trivial case where you have a single box serving GPRS to few subscribers
+on an arbitrary IP address range not known in the larger network, the easiest
+way to enable GPRS uplink would be to enable IP forwarding and masquerading.
+
+To manually enable IPv4 forwarding and masquerading ad-hoc, you can do:
+
+----
+sh -c "echo 1 > /proc/sys/net/ipv4/ip_forward"
+iptables -t nat -A POSTROUTING -o '*' -j MASQUERADE
+----
+
+(You may want to replace `*` with the network device name, like `-o eth0`)
+
+There are various ways to enable these settings persistently, please refer to
+your distribution's documentation -- e.g. look for @net.ipv4.ip_forward=1@ in
+@/etc/sysctl.d/@, and https://wiki.debian.org/iptables for masquerading.
+
+=== Multiple instances
+
+Running multiple instances of `osmo-ggsn` is possible if all GGSN instances
+are binding to different local IP addresse and all other interfaces (VTY,
+OML) are separated using the appropriate configuration options. The IP based
+interfaces are binding to local host by default. In order to separate the
+processes, the user has to bind those services to specific but different
+IP addresses.
+
+The VTY and the control interface can be bound to IP addresses from the loopback
+address range.
+
+.Example: Binding VTY and control interface to a specific ip-address
+----
+line vty
+ bind 127.0.0.2
+ctrl
+ bind 127.0.0.2
+----
+
+Also make sure to place each instance's GTP bind on a separate IP address (GTP
+uses a port number that is fixed in the GTP specifications, so it will not be
+possible to pick differing ports on the same IP address), like:
+
+----
+ggsn ggsn0
+ gtp bind-ip 127.0.0.2
+----
+
+=== GTP-U kernel module
+
+WARNING: As of writing, IPv6 support for the kernel module has not been
+         upstreamed yet (OS#1952).
+
+WARNING: As of writing, it is not possible to configure multiple APNs with
+         gtpu-mode kernel-gpt. This is a limitation in OsmoGGSN, not in the
+         kernel module (OS#6106).
+
+OsmoGGSN has support to use the Linux kernel GTP-U tunnel driver to accelerate
+the data/user plane while still implementing the control plane (GTP-C) in
+userspace in OsmoGGSN. The kernel module is included in Linux 4.7.0 and higher.
+Notably the Debian GNU/Linux distribution has it enabled by default.
+
+In order to use this feature, make sure that your Linux kernel was configured
+to support it (`CONFIG_GTP=m` or `=y`). Furthermore, `osmo-ggsn` must have been
+built with `./configure` argument `--enable-gtp-linux` (which requires libgtpnl
+to be installed).
+
+Load the kernel module with:
+
+----
+$ sudo modprobe gtp
+----
+
+Then start OsmoGGSN with a configuration file that uses `gtpu-mode kernel-gtp`.
+
+A full example configuration is in `osmo-ggsn-kernel-gtp.cfg`.
+
+----
+$ sudo osmo-ggsn -c /usr/share/doc/osmo-ggsn/examples/osmo-ggsn-kernel-gtp.cfg
+----
+
+.Example: APN with kernel-gtp
+----
+ggsn ggsn0
+ gtp state-dir /tmp
+ gtp bind-ip 127.0.0.2
+ apn internet
+  gtpu-mode kernel-gtp
+  tun-device tun4
+  type-support v4
+  ip prefix dynamic 172.16.222.0/24
+  ip dns 0 8.8.8.8
+  ip dns 1 8.8.4.4
+  ip ifconfig 172.16.222.0/24
+  no shutdown
+----

doc/manuals/osmoggsn-usermanual-docinfo.xml

@@ -0,0 +1,46 @@
+<revhistory>
+  <revision>
+    <revnumber>1</revnumber>
+    <date>August 2017</date>
+    <authorinitials>HW</authorinitials>
+    <revremark>
+      Initial version.
+    </revremark>
+  </revision>
+ </revhistory>
+
+<authorgroup>
+  <author>
+    <firstname>Harald</firstname>
+    <surname>Welte</surname>
+    <email>hwelte@sysmocom.de</email>
+    <authorinitials>HW</authorinitials>
+    <affiliation>
+      <shortaffil>sysmocom</shortaffil>
+      <orgname>sysmocom - s.f.m.c. GmbH</orgname>
+      <jobtitle>Managing Director</jobtitle>
+    </affiliation>
+  </author>
+</authorgroup>
+
+<copyright>
+  <year>2013-2017</year>
+  <holder>sysmocom - s.f.m.c. GmbH</holder>
+</copyright>
+
+<legalnotice>
+  <para>
+	Permission is granted to copy, distribute and/or modify this
+	document under the terms of the GNU Free Documentation License,
+	Version 1.3 or any later version published by the Free Software
+	Foundation; with no Invariant Sections, no Front-Cover Texts,
+	and no Back-Cover Texts.  A copy of the license is included in
+	the section entitled "GNU Free Documentation License".
+  </para>
+  <para>
+	The Asciidoc source code of this manual can be found at
+	<ulink url="http://git.osmocom.org/osmo-gsm-manuals/">
+		http://git.osmocom.org/osmo-gsm-manuals/
+	</ulink>
+  </para>
+</legalnotice>

doc/manuals/osmoggsn-usermanual.adoc

@@ -0,0 +1,31 @@
+OsmoGGSN User Manual
+====================
+Harald Welte <hwelte@sysmocom.de>
+
+
+include::./common/chapters/preface.adoc[]
+
+include::{srcdir}/chapters/overview.adoc[]
+
+include::{srcdir}/chapters/running.adoc[]
+
+//include::{srcdir}/chapters/control.adoc[]
+
+include::./common/chapters/vty.adoc[]
+
+include::./common/chapters/logging.adoc[]
+
+
+include::{srcdir}/chapters/configuration.adoc[]
+
+include::./common/chapters/control_if.adoc[]
+
+include::./common/chapters/vty_cpu_sched.adoc[]
+
+include::./common/chapters/port_numbers.adoc[]
+
+include::./common/chapters/bibliography.adoc[]
+
+include::./common/chapters/glossary.adoc[]
+
+include::./common/chapters/gfdl.adoc[]

doc/manuals/osmoggsn-vty-reference.xml

@@ -0,0 +1,38 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!--
+  ex:ts=2:sw=42sts=2:et
+  -*- tab-width: 4; c-basic-offset: 4; indent-tabs-mode: nil -*-
+-->
+<!DOCTYPE book PUBLIC "-//OASIS//DTD DocBook XML 5.0//EN"
+"http://docbook.org/xml/5.0/dtd/docbook.dtd" [
+<!ENTITY chapter-vty      SYSTEM      "./common/chapters/vty.xml" >
+<!ENTITY sections-vty     SYSTEM      "generated/docbook_vty.xml"  >
+]>
+
+<book>
+  <info>
+    <revhistory>
+        <revision>
+            <revnumber>v1</revnumber>
+            <date>06th September 2017</date>
+            <authorinitials>hw</authorinitials>
+            <revremark>Initial version as of OsmoGGSN v1.0.0</revremark>
+        </revision>
+    </revhistory>
+
+    <title>OsmoGGSN VTY Reference</title>
+
+    <copyright>
+      <year>2017</year>
+    </copyright>
+
+    <legalnotice>
+      <para>This work is copyright by <orgname>sysmocom - s.f.m.c. GmbH</orgname>. All rights reserved.
+      </para>
+    </legalnotice>
+  </info>
+
+  <!-- Main chapters-->
+  &chapter-vty;
+</book>
+

doc/manuals/regen_doc.sh

@@ -0,0 +1,17 @@
+#!/bin/sh -x
+
+if [ -z "$DOCKER_PLAYGROUND" ]; then
+	echo "You need to set DOCKER_PLAYGROUND"
+	exit 1
+fi
+
+SCRIPT=$(realpath "$0")
+MANUAL_DIR=$(dirname "$SCRIPT")
+
+COMMIT=${COMMIT:-$(git log -1 --format=format:%H)}
+
+cd "$DOCKER_PLAYGROUND/scripts" || exit 1
+
+OSMO_GGSN_BRANCH=$COMMIT ./regen_doc.sh osmo-ggsn 4260 \
+	"$MANUAL_DIR/chapters/counters_generated.adoc" \
+	"$MANUAL_DIR/vty/ggsn_vty_reference.xml"

doc/manuals/vty/ggsn_vty_additions.xml

@@ -0,0 +1,30 @@
+<vtydoc xmlns='urn:osmocom:xml:libosmocore:vty:doc:1.0'>
+	<node id='config-line'>
+		<child_of nodeid='config' />
+		<name>Telnet/VTY Configuration Node</name>
+		<description>
+			Configure parameters of the Telnet/VTY Interface, such as to which IP address it should bind/listen to.
+		</description>
+	</node>
+	<node id='config-ctrl'>
+		<child_of nodeid='config' />
+		<name>CTRL Configuration Node</name>
+		<description>
+			Configure parameters of the CTRL Interface, such as to which IP address it should bind/listen to.
+		</description>
+	</node>
+	<node id='config-ggsn'>
+		<child_of nodeid='config' />
+		<name>GGSN Instance Configuration Node</name>
+		<description>
+			Configure an Instance of a (virtual) GGSN
+		</description>
+	</node>
+	<node id='config-ggsn-apn'>
+		<child_of nodeid='config-ggsn' />
+		<name>APN Configuration Node</name>
+		<description>
+			Configure an Access Point Name (APN) inside a GGSN Instance
+		</description>
+	</node>
+</vtydoc>

ggsn/Makefile.am

@@ -2,18 +2,22 @@ bin_PROGRAMS = osmo-ggsn
 
 AM_LDFLAGS = @EXEC_LDFLAGS@
 
-AM_CFLAGS = -O2 -D_GNU_SOURCE -fno-builtin -Wall -DSBINDIR='"$(sbindir)"' -ggdb $(LIBOSMOCORE_CFLAGS) $(LIBOSMOCTRL_CFLAGS) $(LIBOSMOVTY_CFLAGS)
+AM_CFLAGS = \
+	    -D_GNU_SOURCE \
+	    -fno-builtin \
+	    -Wall \
+	    -DSBINDIR='"$(sbindir)"' \
+	    $(LIBOSMOCORE_CFLAGS) \
+	    $(LIBOSMOCTRL_CFLAGS) \
+	    $(LIBOSMOVTY_CFLAGS) \
+	    $(NULL)
 
 osmo_ggsn_LDADD = @EXEC_LDADD@ -lgtp -L../gtp ../lib/libmisc.a $(LIBOSMOCORE_LIBS) $(LIBOSMOCTRL_LIBS) $(LIBOSMOVTY_LIBS)
 
 if ENABLE_GTP_KERNEL
-AM_CFLAGS += -DGTP_KERNEL
-osmo_ggsn_LDADD += -lgtpnl
+AM_CFLAGS += -DGTP_KERNEL $(LIBGTPNL_CFLAGS)
+osmo_ggsn_LDADD += $(LIBGTPNL_LIBS)
 endif
 
 osmo_ggsn_DEPENDENCIES = ../gtp/libgtp.la ../lib/libmisc.a
-osmo_ggsn_SOURCES = ggsn_vty.c ggsn.c ggsn.h gtp-kernel.h icmpv6.c icmpv6.h checksum.c checksum.h
-
-if ENABLE_GTP_KERNEL
-osmo_ggsn_SOURCES += gtp-kernel.c
-endif
+osmo_ggsn_SOURCES = ggsn_main.c ggsn_vty.c ggsn.c ggsn.h sgsn.c sgsn.h pco.c pco.h

ggsn/ggsn.h

@@ -6,6 +6,8 @@
 #include <osmocom/core/linuxlist.h>
 #include <osmocom/core/select.h>
 #include <osmocom/core/timer.h>
+#include <osmocom/core/tdef.h>
+#include <osmocom/ctrl/control_if.h>
 
 #include "../lib/tun.h"
 #include "../lib/ippool.h"
@@ -13,6 +15,8 @@
 #include "../lib/in46_addr.h"
 #include "../gtp/gtp.h"
 
+#include "sgsn.h"
+
 #define APN_TYPE_IPv4	0x01	/* v4-only */
 #define APN_TYPE_IPv6	0x02	/* v6-only */
 #define APN_TYPE_IPv4v6	0x04	/* v4v6 dual-stack */
@@ -22,6 +26,7 @@ struct ggsn_ctx;
 struct apn_ctx_ip {
 	struct {
 		struct in46_prefix ifconfig_prefix;
+		struct in46_prefix ll_prefix;
 		struct in46_prefix static_prefix;
 		struct in46_prefix dynamic_prefix;
 		/* v4 DNS server names */
@@ -61,9 +66,9 @@ struct apn_ctx {
 		uint32_t apn_type_mask;
 		/* GTP-U via TUN device or in Linux kernel */
 		enum apn_gtpu_mode gtpu_mode;
-		/* administratively shut-down (true) or not (false) */
+		/* administratively shut down (true) or not (false) */
 		bool shutdown;
-		/* transmit G-PDU sequeence numbers (true) or not (false) */
+		/* transmit G-PDU sequence numbers (true) or not (false) */
 		bool tx_gpdu_seq;
 	} cfg;
 
@@ -87,6 +92,14 @@ struct apn_ctx {
 	struct apn_ctx_ip v6;
 };
 
+struct pdp_priv_t {
+	struct pdp_t *lib; /* pointer to libgtp associated pdp_t instance */
+	struct sgsn_peer *sgsn;
+	struct apn_ctx *apn;
+	struct llist_head entry; /* to be included into sgsn_peer */
+	/* struct ggsn_ctx can be reached through lib->gsn->priv, or through sgsn->ggsn */
+};
+
 struct ggsn_ctx {
 	/* global list of GGSNs */
 	struct llist_head list;
@@ -94,6 +107,9 @@ struct ggsn_ctx {
 	/* list of APNs in this GGSN */
 	struct llist_head apn_list;
 
+	/* list of SGSN peers (struct sgsn_peer) in this GGSN. TODO: hash table with key <ip+port>? */
+	struct llist_head sgsn_list;
+
 	bool started;
 
 	struct {
@@ -110,7 +126,9 @@ struct ggsn_ctx {
 		struct in46_addr gtpu_addr;
 		/* directory for state file */
 		char *state_dir;
-		/* administratively shut-down (true) or not (false) */
+		/* Time between Echo requests on each SGSN */
+		unsigned int echo_interval;
+		/* administratively shut down (true) or not (false) */
 		bool shutdown;
 	} cfg;
 
@@ -121,8 +139,6 @@ struct ggsn_ctx {
 	struct osmo_fd gtp_fd0;
 	struct osmo_fd gtp_fd1c;
 	struct osmo_fd gtp_fd1u;
-
-	struct osmo_timer_list gtp_timer;
 };
 
 /* ggsn_vty.c */
@@ -134,9 +150,22 @@ struct ggsn_ctx *ggsn_find_or_create(void *ctx, const char *name);
 struct apn_ctx *ggsn_find_apn(struct ggsn_ctx *ggsn, const char *name);
 struct apn_ctx *ggsn_find_or_create_apn(struct ggsn_ctx *ggsn, const char *name);
 
-/* ggsn.c */
+/* ggsn_main.c */
+extern struct ctrl_handle *g_ctrlh;
 extern void *tall_ggsn_ctx;
+extern struct osmo_tdef_group ggsn_tdef_group[];
+
+/* ggsn.c */
 extern int ggsn_start(struct ggsn_ctx *ggsn);
 extern int ggsn_stop(struct ggsn_ctx *ggsn);
 extern int apn_start(struct apn_ctx *apn);
-extern int apn_stop(struct apn_ctx *apn, bool force);
+extern int apn_stop(struct apn_ctx *apn);
+void ggsn_close_one_pdp(struct pdp_t *pdp);
+
+#define LOGPAPN(level, apn, fmt, args...)			\
+	LOGP(DGGSN, level, "APN(%s): " fmt, (apn)->cfg.name, ## args)
+
+#define LOGPGGSN(level, ggsn, fmt, args...)			\
+	LOGP(DGGSN, level, "GGSN(%s): " fmt, (ggsn)->cfg.name, ## args)
+
+#define LOGPPDP(level, pdp, fmt, args...) LOGPDPX(DGGSN, level, pdp, fmt, ## args)

ggsn/ggsn_main.c

@@ -0,0 +1,261 @@
+/*
+ * OsmoGGSN - Gateway GPRS Support Node
+ * Copyright (C) 2002, 2003, 2004 Mondru AB.
+ * Copyright (C) 2017-2019 by Harald Welte <laforge@gnumonks.org>
+ * Copyright (C) 2019 sysmocom - s.f.m.c. GmbH <info@sysmocom.de>
+ *
+ * The contents of this file may be used under the terms of the GNU
+ * General Public License Version 2, provided that the above copyright
+ * notice and this permission notice is included in all copies or
+ * substantial portions of the software.
+ *
+ */
+
+#include "../config.h"
+
+#ifdef HAVE_STDINT_H
+#include <stdint.h>
+#endif
+
+#include <getopt.h>
+#include <ctype.h>
+#include <signal.h>
+#include <stdio.h>
+#include <string.h>
+#include <stdlib.h>
+#include <unistd.h>
+#include <inttypes.h>
+#include <errno.h>
+#include <sys/types.h>
+#include <sys/ioctl.h>
+
+#include <osmocom/core/application.h>
+#include <osmocom/core/select.h>
+#include <osmocom/core/stats.h>
+#include <osmocom/core/rate_ctr.h>
+#include <osmocom/core/timer.h>
+#include <osmocom/core/tdef.h>
+#include <osmocom/core/utils.h>
+#include <osmocom/ctrl/control_if.h>
+#include <osmocom/ctrl/control_cmd.h>
+#include <osmocom/ctrl/control_vty.h>
+#include <osmocom/ctrl/ports.h>
+#include <osmocom/vty/telnet_interface.h>
+#include <osmocom/vty/logging.h>
+#include <osmocom/vty/stats.h>
+#include <osmocom/vty/ports.h>
+#include <osmocom/vty/command.h>
+#include <osmocom/vty/misc.h>
+#include <osmocom/vty/cpu_sched_vty.h>
+
+#include "ggsn.h"
+
+void *tall_ggsn_ctx;
+
+static int end = 0;
+static int daemonize = 0;
+struct ctrl_handle *g_ctrlh;
+
+struct ul255_t qos;
+struct ul255_t apn;
+
+static char *config_file = "osmo-ggsn.cfg";
+
+struct osmo_tdef_group ggsn_tdef_group[] = {
+	{.name = "gtp", .tdefs = gtp_T_defs, .desc = "GTP (libgtp) timers" },
+	{ }
+};
+
+/* To exit gracefully. Used with GCC compilation flag -pg and gprof */
+static void signal_handler(int s)
+{
+	LOGP(DGGSN, LOGL_NOTICE, "signal %d received\n", s);
+	switch (s) {
+	case SIGINT:
+	case SIGTERM:
+		LOGP(DGGSN, LOGL_NOTICE, "SIGINT received, shutting down\n");
+		end = 1;
+		break;
+	case SIGABRT:
+		/* in case of abort, we want to obtain a talloc report and
+		 * then run default SIGABRT handler, who will generate coredump
+		 * and abort the process. abort() should do this for us after we
+		 * return, but program wouldn't exit if an external SIGABRT is
+		 * received.
+		 */
+		talloc_report(tall_vty_ctx, stderr);
+		talloc_report_full(tall_ggsn_ctx, stderr);
+		signal(SIGABRT, SIG_DFL);
+		raise(SIGABRT);
+		break;
+	case SIGUSR1:
+		talloc_report(tall_vty_ctx, stderr);
+		talloc_report_full(tall_ggsn_ctx, stderr);
+		break;
+	case SIGUSR2:
+		talloc_report_full(tall_vty_ctx, stderr);
+		break;
+	default:
+		break;
+	}
+}
+
+static void print_usage()
+{
+	printf("Usage: osmo-ggsn [-h] [-D] [-c configfile] [-V]\n");
+}
+
+static void print_help()
+{
+	printf(	"  Some useful help...\n"
+		"  -h --help		This help text\n"
+		"  -D --daemonize	Fork the process into a background daemon\n"
+		"  -c --config-file	filename The config file to use\n"
+		"  -V --version		Print the version of OsmoGGSN\n"
+		);
+
+	printf("\nVTY reference generation:\n");
+	printf("     --vty-ref-mode MODE        VTY reference generation mode (e.g. 'expert').\n");
+	printf("     --vty-ref-xml              Generate the VTY reference XML output and exit.\n");
+}
+
+static void handle_long_options(const char *prog_name, const int long_option)
+{
+	static int vty_ref_mode = VTY_REF_GEN_MODE_DEFAULT;
+
+	switch (long_option) {
+	case 1:
+		vty_ref_mode = get_string_value(vty_ref_gen_mode_names, optarg);
+		if (vty_ref_mode < 0) {
+			fprintf(stderr, "%s: Unknown VTY reference generation "
+				"mode '%s'\n", prog_name, optarg);
+			exit(2);
+		}
+		break;
+	case 2:
+		fprintf(stderr, "Generating the VTY reference in mode '%s' (%s)\n",
+			get_value_string(vty_ref_gen_mode_names, vty_ref_mode),
+			get_value_string(vty_ref_gen_mode_desc, vty_ref_mode));
+		vty_dump_xml_ref_mode(stdout, (enum vty_ref_gen_mode) vty_ref_mode);
+		exit(0);
+	default:
+		fprintf(stderr, "%s: error parsing cmdline options\n", prog_name);
+		exit(2);
+	}
+}
+
+static void handle_options(int argc, char **argv)
+{
+	while (1) {
+		int option_index = 0, c;
+		static int long_option = 0;
+		static struct option long_options[] = {
+			{ "help", 0, 0, 'h' },
+			{ "daemonize", 0, 0, 'D' },
+			{ "config-file", 1, 0, 'c' },
+			{ "version", 0, 0, 'V' },
+			{ "vty-ref-mode", 1, &long_option, 1 },
+			{ "vty-ref-xml", 0, &long_option, 2 },
+			{ 0, 0, 0, 0 }
+		};
+
+		c = getopt_long(argc, argv, "hdc:V", long_options, &option_index);
+		if (c == -1)
+			break;
+
+		switch (c) {
+		case 0:
+			handle_long_options(argv[0], long_option);
+			break;
+		case 'h':
+			print_usage();
+			print_help();
+			exit(0);
+		case 'D':
+			daemonize = 1;
+			break;
+		case 'c':
+			config_file = optarg;
+			break;
+		case 'V':
+			print_version(1);
+			exit(0);
+			break;
+		}
+	}
+}
+
+int main(int argc, char **argv)
+{
+	struct ggsn_ctx *ggsn;
+	int rc;
+
+	tall_ggsn_ctx = talloc_named_const(NULL, 0, "OsmoGGSN");
+	msgb_talloc_ctx_init(tall_ggsn_ctx, 0);
+	g_vty_info.tall_ctx = tall_ggsn_ctx;
+
+	/* Handle keyboard interrupt SIGINT */
+	signal(SIGINT, &signal_handler);
+	signal(SIGTERM, &signal_handler);
+	signal(SIGABRT, &signal_handler);
+	signal(SIGUSR1, &signal_handler);
+	signal(SIGUSR2, &signal_handler);
+
+	osmo_init_ignore_signals();
+	osmo_init_logging2(tall_ggsn_ctx, &log_info);
+	osmo_stats_init(tall_ggsn_ctx);
+
+	vty_init(&g_vty_info);
+	logging_vty_add_cmds();
+	osmo_talloc_vty_add_cmds();
+	osmo_stats_vty_add_cmds();
+	ggsn_vty_init();
+	ctrl_vty_init(tall_ggsn_ctx);
+	osmo_cpu_sched_vty_init(tall_ggsn_ctx);
+
+	handle_options(argc, argv);
+
+	rate_ctr_init(tall_ggsn_ctx);
+
+	rc = vty_read_config_file(config_file, NULL);
+	if (rc < 0) {
+		fprintf(stderr, "Failed to open config file: '%s'\n", config_file);
+		exit(2);
+	}
+
+	rc = telnet_init_default(tall_ggsn_ctx, NULL, OSMO_VTY_PORT_GGSN);
+	if (rc < 0)
+		exit(1);
+
+	g_ctrlh = ctrl_interface_setup(NULL, OSMO_CTRL_PORT_GGSN, NULL);
+	if (!g_ctrlh) {
+		LOGP(DGGSN, LOGL_ERROR, "Failed to create CTRL interface.\n");
+		exit(1);
+	}
+
+	if (daemonize) {
+		rc = osmo_daemonize();
+		if (rc < 0) {
+			perror("Error during daemonize");
+			exit(1);
+		}
+	}
+
+#if 0
+	/* qos                                                             */
+	qos.l = 3;
+	qos.v[2] = (args_info.qos_arg) & 0xff;
+	qos.v[1] = ((args_info.qos_arg) >> 8) & 0xff;
+	qos.v[0] = ((args_info.qos_arg) >> 16) & 0xff;
+#endif
+
+	/* Main select loop */
+	while (!end) {
+		osmo_select_main(0);
+	}
+
+	llist_for_each_entry(ggsn, &g_ggsn_list, list)
+		ggsn_stop(ggsn);
+
+	return 0;
+}

ggsn/ggsn_vty.c

@@ -26,16 +26,22 @@
 #include <osmocom/core/talloc.h>
 #include <osmocom/core/utils.h>
 #include <osmocom/core/rate_ctr.h>
+#include <osmocom/gsm/apn.h>
+#include <osmocom/gsm/gsm48_ie.h>
 #include <osmocom/gsm/protocol/gsm_04_08_gprs.h>
 
 #include <osmocom/vty/command.h>
 #include <osmocom/vty/vty.h>
 #include <osmocom/vty/misc.h>
+#include <osmocom/vty/tdef_vty.h>
 
 #include "../gtp/gtp.h"
 #include "../gtp/pdp.h"
 
+#include "../lib/util.h"
+
 #include "ggsn.h"
+#include "sgsn.h"
 
 #define PREFIX_STR	"Prefix (Network/Netmask)\n"
 #define IFCONFIG_STR	"GGSN-based interface configuration\n"
@@ -75,6 +81,7 @@ struct ggsn_ctx *ggsn_find_or_create(void *ctx, const char *name)
 	ggsn->cfg.state_dir = talloc_strdup(ggsn, "/tmp");
 	ggsn->cfg.shutdown = true;
 	INIT_LLIST_HEAD(&ggsn->apn_list);
+	INIT_LLIST_HEAD(&ggsn->sgsn_list);
 
 	llist_add_tail(&ggsn->list, &g_ggsn_list);
 	return ggsn;
@@ -292,7 +299,7 @@ DEFUN(cfg_ggsn_no_default_apn, cfg_ggsn_no_default_apn_cmd,
 
 DEFUN(cfg_ggsn_shutdown, cfg_ggsn_shutdown_cmd,
 	"shutdown ggsn",
-	"Put the GGSN in administrative shut-down\n" GGSN_STR)
+	"Put the GGSN in administrative shutdown\n" GGSN_STR)
 {
 	struct ggsn_ctx *ggsn = (struct ggsn_ctx *) vty->index;
 
@@ -309,7 +316,7 @@ DEFUN(cfg_ggsn_shutdown, cfg_ggsn_shutdown_cmd,
 
 DEFUN(cfg_ggsn_no_shutdown, cfg_ggsn_no_shutdown_cmd,
 	"no shutdown ggsn",
-	NO_STR GGSN_STR "Remove the GGSN from administrative shut-down\n")
+	NO_STR GGSN_STR "Remove the GGSN from administrative shutdown\n")
 {
 	struct ggsn_ctx *ggsn = (struct ggsn_ctx *) vty->index;
 
@@ -324,6 +331,79 @@ DEFUN(cfg_ggsn_no_shutdown, cfg_ggsn_no_shutdown_cmd,
 	return CMD_SUCCESS;
 }
 
+static void show_one_sgsn(struct vty *vty, const struct sgsn_peer *sgsn, const char* prefix)
+{
+	char buf[INET_ADDRSTRLEN];
+
+	inet_ntop(AF_INET, &sgsn->addr, buf, sizeof(buf));
+	vty_out(vty, "%s(S)GSN %s%s", prefix, buf, VTY_NEWLINE);
+	vty_out(vty, "%s Restart Counter: %d%s", prefix, sgsn->remote_restart_ctr, VTY_NEWLINE);
+	vty_out(vty, "%s PDP contexts: %d%s", prefix, llist_count(&sgsn->pdp_list), VTY_NEWLINE);
+	vty_out(vty, "%s Echo Requests in-flight: %u%s", prefix, sgsn->tx_msgs_queued, VTY_NEWLINE);
+}
+
+DEFUN(cfg_ggsn_show_sgsn, cfg_ggsn_show_sgsn_cmd,
+	"show sgsn",
+	NO_STR GGSN_STR "Remove the GGSN from administrative shutdown\n")
+{
+	struct ggsn_ctx *ggsn = (struct ggsn_ctx *) vty->index;
+	struct sgsn_peer *sgsn;
+
+	llist_for_each_entry(sgsn, &ggsn->sgsn_list, entry) {
+		show_one_sgsn(vty, sgsn, "");
+	}
+
+	return CMD_SUCCESS;
+}
+
+/* Seee 3GPP TS 29.060 section 7.2.1 */
+DEFUN(cfg_ggsn_echo_interval, cfg_ggsn_echo_interval_cmd,
+	"echo-interval <1-36000>",
+	GGSN_STR "GGSN Number\n"
+	"Send an echo request to this static GGSN every interval\n"
+	"Interval between echo requests in seconds\n")
+{
+	struct ggsn_ctx *ggsn = (struct ggsn_ctx *) vty->index;
+	int prev_interval = ggsn->cfg.echo_interval;
+	struct sgsn_peer *sgsn;
+
+	ggsn->cfg.echo_interval = atoi(argv[0]);
+
+	if (ggsn->cfg.echo_interval < 60)
+		vty_out(vty, "%% 3GPP TS 29.060 section states interval should " \
+			     "not be lower than 60 seconds, use this value for " \
+			     "testing purposes only!%s", VTY_NEWLINE);
+
+	if (prev_interval == ggsn->cfg.echo_interval)
+		return CMD_SUCCESS;
+
+	/* Re-enable echo timer for all sgsn */
+	llist_for_each_entry(sgsn, &ggsn->sgsn_list, entry)
+		sgsn_echo_timer_start(sgsn);
+
+	return CMD_SUCCESS;
+}
+
+DEFUN(cfg_ggsn_no_echo_interval, cfg_ggsn_no_echo_interval_cmd,
+	"no echo-interval",
+	GGSN_STR "GGSN Number\n"
+	NO_STR "Send an echo request to this static GGSN every interval.\n")
+{
+	struct ggsn_ctx *ggsn = (struct ggsn_ctx *) vty->index;
+	struct sgsn_peer *sgsn;
+
+	if (ggsn->cfg.echo_interval == 0)
+		return CMD_SUCCESS;
+
+	ggsn->cfg.echo_interval = 0;
+
+	/* Disable echo timer for all sgsn */
+	llist_for_each_entry(sgsn, &ggsn->sgsn_list, entry)
+		sgsn_echo_timer_stop(sgsn);
+
+	return CMD_SUCCESS;
+}
+
 /* APN Node */
 
 static struct cmd_node apn_node = {
@@ -453,9 +533,11 @@ DEFUN(cfg_apn_ip_prefix, cfg_apn_ip_prefix_cmd,
 	struct in46_prefix *pfx;
 
 	/* first update our parsed prefix */
-	if (!strcmp(argv[0], "static"))
+	if (!strcmp(argv[0], "static")) {
 		pfx = &apn->v4.cfg.static_prefix;
-	else
+		vty_out(vty, "%% static IP addresses currently not yet supported%s", VTY_NEWLINE);
+		return CMD_WARNING;
+	} else
 		pfx = &apn->v4.cfg.dynamic_prefix;
 	str2prefix(pfx, argv[1]);
 
@@ -487,9 +569,11 @@ DEFUN(cfg_apn_ipv6_prefix, cfg_apn_ipv6_prefix_cmd,
 	struct apn_ctx *apn = (struct apn_ctx *) vty->index;
 	struct in46_prefix *pfx;
 
-	if (!strcmp(argv[0], "static"))
+	if (!strcmp(argv[0], "static")) {
 		pfx = &apn->v6.cfg.static_prefix;
-	else
+		vty_out(vty, "%% static IP addresses currently not yet supported%s", VTY_NEWLINE);
+		return CMD_WARNING;
+	} else
 		pfx = &apn->v6.cfg.dynamic_prefix;
 	str2prefix(pfx, argv[1]);
 	return CMD_SUCCESS;
@@ -513,6 +597,24 @@ DEFUN(cfg_apn_no_ipv6_ifconfig, cfg_apn_no_ipv6_ifconfig_cmd,
 	return CMD_SUCCESS;
 }
 
+DEFUN(cfg_apn_ipv6_linklocal, cfg_apn_ipv6_linklocal_cmd,
+	"ipv6 link-local X:X::X:X/M",
+	IP6_STR IFCONFIG_STR "IPv6 Link-local Adress/Prefix-Length\n")
+{
+	struct apn_ctx *apn = (struct apn_ctx *) vty->index;
+	str2prefix(&apn->v6.cfg.ll_prefix, argv[0]);
+	return CMD_SUCCESS;
+}
+
+DEFUN(cfg_apn_no_ipv6_linklocal, cfg_apn_no_ipv6_linklocal_cmd,
+	"no ipv6 link-local",
+	NO_STR IP6_STR IFCONFIG_STR)
+{
+	struct apn_ctx *apn = (struct apn_ctx *) vty->index;
+	memset(&apn->v6.cfg.ll_prefix, 0, sizeof(apn->v6.cfg.ll_prefix));
+	return CMD_SUCCESS;
+}
+
 #define DNS_STRINGS "Configure DNS Server\n" "primary/secondary DNS\n" "IP address of DNS Sever\n"
 
 DEFUN(cfg_apn_ip_dns, cfg_apn_ip_dns_cmd,
@@ -579,12 +681,12 @@ DEFUN(cfg_apn_no_gpdu_seq, cfg_apn_no_gpdu_seq_cmd,
 
 DEFUN(cfg_apn_shutdown, cfg_apn_shutdown_cmd,
 	"shutdown",
-	"Put the APN in administrative shut-down\n")
+	"Put the APN in administrative shutdown\n")
 {
 	struct apn_ctx *apn = (struct apn_ctx *) vty->index;
 
 	if (!apn->cfg.shutdown) {
-		if (apn_stop(apn, false)) {
+		if (apn_stop(apn)) {
 			vty_out(vty, "%% Failed to Stop APN%s", VTY_NEWLINE);
 			return CMD_WARNING;
 		}
@@ -596,11 +698,15 @@ DEFUN(cfg_apn_shutdown, cfg_apn_shutdown_cmd,
 
 DEFUN(cfg_apn_no_shutdown, cfg_apn_no_shutdown_cmd,
 	"no shutdown",
-	NO_STR "Remove the APN from administrative shut-down\n")
+	NO_STR "Remove the APN from administrative shutdown\n")
 {
 	struct apn_ctx *apn = (struct apn_ctx *) vty->index;
 
 	if (apn->cfg.shutdown) {
+		if (!apn->tun.cfg.dev_name) {
+			vty_out(vty, "%% Failed to start APN, tun-device is not configured%s", VTY_NEWLINE);
+			return CMD_WARNING;
+		}
 		if (apn_start(apn) < 0) {
 			vty_out(vty, "%% Failed to start APN, check log for details%s", VTY_NEWLINE);
 			return CMD_WARNING;
@@ -634,9 +740,9 @@ static void config_write_apn(struct vty *vty, struct apn_ctx *apn)
 		vty_out(vty, "  ipdown-script %s%s", apn->tun.cfg.ipdown_script, VTY_NEWLINE);
 
 	for (i = 0; i < 32; i++) {
-		if (!(apn->cfg.apn_type_mask & (1 << i)))
+		if (!(apn->cfg.apn_type_mask & (UINT32_C(1) << i)))
 			continue;
-		vty_out(vty, "  type-support %s%s", get_value_string(pdp_type_names, (1 << i)),
+		vty_out(vty, "  type-support %s%s", get_value_string(pdp_type_names, (UINT32_C(1) << i)),
 			VTY_NEWLINE);
 	}
 
@@ -668,6 +774,8 @@ static void config_write_apn(struct vty *vty, struct apn_ctx *apn)
 	}
 	if (apn->v6.cfg.ifconfig_prefix.addr.len)
 		vty_dump_prefix(vty, "  ipv6 ifconfig", &apn->v6.cfg.ifconfig_prefix);
+	if (apn->v6.cfg.ll_prefix.addr.len)
+		vty_dump_prefix(vty, "  ipv6 link-local", &apn->v6.cfg.ll_prefix);
 
 	/* must be last */
 	vty_out(vty, "  %sshutdown%s", apn->cfg.shutdown ? "" : "no ", VTY_NEWLINE);
@@ -688,10 +796,13 @@ static int config_write_ggsn(struct vty *vty)
 			vty_out(vty, " gtp control-ip %s%s", in46a_ntoa(&ggsn->cfg.gtpc_addr), VTY_NEWLINE);
 		if (ggsn->cfg.gtpu_addr.v4.s_addr)
 			vty_out(vty, " gtp user-ip %s%s", in46a_ntoa(&ggsn->cfg.gtpu_addr), VTY_NEWLINE);
+		osmo_tdef_vty_groups_write(vty, " ");
 		llist_for_each_entry(apn, &ggsn->apn_list, list)
 			config_write_apn(vty, apn);
 		if (ggsn->cfg.default_apn)
 			vty_out(vty, " default-apn %s%s", ggsn->cfg.default_apn->cfg.name, VTY_NEWLINE);
+		if (ggsn->cfg.echo_interval)
+			vty_out(vty, " echo-interval %u%s", ggsn->cfg.echo_interval, VTY_NEWLINE);
 		/* must be last */
 		vty_out(vty, " %sshutdown ggsn%s", ggsn->cfg.shutdown ? "" : "no ", VTY_NEWLINE);
 	}
@@ -703,19 +814,46 @@ static const char *print_gsnaddr(const struct ul16_t *in)
 {
 	struct in46_addr in46;
 
-	in46.len = in->l;
-	OSMO_ASSERT(in->l <= sizeof(in46.v6));
-	memcpy(&in46.v6, in->v, in->l);
-
+	in46a_from_gsna(in, &in46);
 	return in46a_ntoa(&in46);
 }
 
-static void show_one_pdp(struct vty *vty, struct pdp_t *pdp)
+/* Useful for v4v6 APNs, where we first iterate over v4 pool and then over v6
+   pool. param v4only can be used to avoid printing duplicates for pdp context
+   containing both IPv4 and IPv6 addresses. */
+static void show_one_pdp_v4only(struct vty *vty, struct pdp_t *pdp, bool v4only)
 {
-	struct in46_addr eua46;
+	struct ippoolm_t *peer4, *peer6;
+	char name_buf[256];
+	char *apn_name;
+	int rc;
+
+	peer4 = pdp_get_peer_ipv(pdp, false);
+	peer6 = pdp_get_peer_ipv(pdp, true);
+
+	if (v4only && peer6)
+		return;
+
+	/* Attempt to decode MSISDN */
+	rc = gsm48_decode_bcd_number2(name_buf, sizeof(name_buf),
+				      pdp->msisdn.v, pdp->msisdn.l, 0);
 
 	vty_out(vty, "IMSI: %s, NSAPI: %u, MSISDN: %s%s", imsi_gtp2str(&pdp->imsi), pdp->nsapi,
-		osmo_hexdump_nospc(pdp->msisdn.v, pdp->msisdn.l), VTY_NEWLINE);
+		rc ? "(NONE)" : name_buf, VTY_NEWLINE);
+
+	vty_out(vty, " Version: %d", pdp->version);
+	if (pdp->version == 1) {
+		if (!pdp->secondary) {
+			vty_out(vty, ", Primary, Num Secondaries: %d%s%s",
+				pdp_count_secondary(pdp) - 1, /* primary included in count */
+				pdp->nodata ? ", No User Plane": "",
+				VTY_NEWLINE);
+		} else {
+			vty_out(vty, ", Secondary%s", VTY_NEWLINE);
+		}
+	} else {
+		vty_out(vty, "%s", VTY_NEWLINE);
+	}
 
 	vty_out(vty, " Control: %s:%08x ", print_gsnaddr(&pdp->gsnlc), pdp->teic_own);
 	vty_out(vty, "<-> %s:%08x%s", print_gsnaddr(&pdp->gsnrc), pdp->teic_gn, VTY_NEWLINE);
@@ -723,32 +861,61 @@ static void show_one_pdp(struct vty *vty, struct pdp_t *pdp)
 	vty_out(vty, " Data: %s:%08x ", print_gsnaddr(&pdp->gsnlu), pdp->teid_own);
 	vty_out(vty, "<-> %s:%08x%s", print_gsnaddr(&pdp->gsnru), pdp->teid_gn, VTY_NEWLINE);
 
-	in46a_from_eua(&pdp->eua, &eua46);
-	vty_out(vty, " End-User Address: %s%s", in46a_ntoa(&eua46), VTY_NEWLINE);
+	apn_name = osmo_apn_to_str(name_buf, pdp->apn_req.v, pdp->apn_req.l);
+	vty_out(vty, " APN requested: %s%s", apn_name ? name_buf : "(NONE)", VTY_NEWLINE);
+	apn_name = osmo_apn_to_str(name_buf, pdp->apn_use.v, pdp->apn_use.l);
+	vty_out(vty, " APN in use: %s%s", apn_name ? name_buf : "(NONE)", VTY_NEWLINE);
+
+	if (peer4)
+		vty_out(vty, " End-User Address (IPv4): %s%s",
+			in46a_ntop(&peer4->addr, name_buf, sizeof(name_buf)), VTY_NEWLINE);
+	if (peer6)
+		vty_out(vty, " End-User Address (IPv6): %s%s",
+			in46a_ntop(&peer6->addr, name_buf, sizeof(name_buf)), VTY_NEWLINE);
 	vty_out(vty, " Transmit GTP Sequence Number for G-PDU: %s%s",
 		pdp->tx_gpdu_seq ? "Yes" : "No", VTY_NEWLINE);
 }
 
+static void show_one_pdp(struct vty *vty, struct pdp_t *pdp)
+{
+	show_one_pdp_v4only(vty, pdp, false);
+}
+
 DEFUN(show_pdpctx_imsi, show_pdpctx_imsi_cmd,
-	"show pdp-context imsi IMSI [<0-15>]",
+	"show pdp-context ggsn NAME imsi IMSI [<0-15>]",
 	SHOW_STR "Display information on PDP Context\n"
+	GGSN_STR "GGSN Name\n"
 	"PDP contexts for given IMSI\n"
 	"PDP context for given NSAPI\n")
 {
-	uint64_t imsi = strtoull(argv[0], NULL, 10);
+	struct ggsn_ctx *ggsn;
+	uint64_t imsi;
 	unsigned int nsapi;
 	struct pdp_t *pdp;
 	int num_found = 0;
 
-	if (argc > 1) {
-		nsapi = atoi(argv[1]);
-		if (pdp_getimsi(&pdp, imsi, nsapi)) {
+	ggsn = ggsn_find(argv[0]);
+	if (!ggsn) {
+		vty_out(vty, "%% No such GGSN '%s'%s", argv[0], VTY_NEWLINE);
+		return CMD_WARNING;
+	}
+
+	if (strlen(argv[1]) < 6 || strlen(argv[1]) > 15) {
+		vty_out(vty, "%% Invalid IMSI '%s'%s", argv[1], VTY_NEWLINE);
+		return CMD_WARNING;
+	}
+
+	imsi = gtp_imsi_str2gtp(argv[1]);
+
+	if (argc > 2) {
+		nsapi = atoi(argv[2]);
+		if (!gtp_pdp_getimsi(ggsn->gsn, &pdp, imsi, nsapi)) {
 			show_one_pdp(vty, pdp);
 			num_found++;
 		}
 	} else {
 		for (nsapi = 0; nsapi < PDP_MAXNSAPI; nsapi++) {
-			if (pdp_getimsi(&pdp, imsi, nsapi))
+			if (gtp_pdp_getimsi(ggsn->gsn, &pdp, imsi, nsapi))
 				continue;
 			show_one_pdp(vty, pdp);
 			num_found++;
@@ -761,8 +928,49 @@ DEFUN(show_pdpctx_imsi, show_pdpctx_imsi_cmd,
 		return CMD_SUCCESS;
 }
 
+DEFUN(show_pdpctx_ip, show_pdpctx_ip_cmd,
+	"show pdp-context ggsn NAME ipv4 A.B.C.D",
+	SHOW_STR "Display information on PDP Context\n"
+	GGSN_STR "GGSN Name\n" "IPv4 address type\n" "IP address\n")
+{
+	struct ggsn_ctx *ggsn;
+	struct apn_ctx *apn;
+	unsigned int i;
+
+	ggsn = ggsn_find(argv[0]);
+	if (!ggsn) {
+		vty_out(vty, "%% No such GGSN '%s'%s", argv[0], VTY_NEWLINE);
+		return CMD_WARNING;
+	}
+
+	/* Iterate over all APNs of a given GGSN */
+	llist_for_each_entry(apn, &ggsn->apn_list, list) {
+		struct ippool_t *pool = apn->v4.pool;
+
+		/* In some rare cases, if GGSN fails to init TUN/TAP interfaces
+		 * (e.g. due to insufficient permissions), it will continue to
+		 * work in such broken state, and pool would be NULL. */
+		if (!pool)
+			continue;
+
+		/* Iterate over all IPv4 pool members */
+		for (i = 0; i < pool->listsize; i++) {
+			struct ippoolm_t *member = &pool->member[i];
+			if (member->inuse == 0)
+				continue;
+			if (strcmp(argv[1], in46a_ntoa(&member->addr)) == 0) {
+				show_one_pdp(vty, member->peer);
+				return CMD_SUCCESS;
+			}
+		}
+	}
+
+	vty_out(vty, "%% No PDP context found for IP '%s'%s", argv[1], VTY_NEWLINE);
+	return CMD_WARNING;
+}
+
 /* show all (active) PDP contexts within a pool */
-static void ippool_show_pdp_contexts(struct vty *vty, struct ippool_t *pool)
+static void ippool_show_pdp_contexts(struct vty *vty, struct ippool_t *pool, bool pdp_v4only)
 {
 	unsigned int i;
 
@@ -773,21 +981,21 @@ static void ippool_show_pdp_contexts(struct vty *vty, struct ippool_t *pool)
 		struct ippoolm_t *member = &pool->member[i];
 		if (member->inuse == 0)
 			continue;
-		show_one_pdp(vty, member->peer);
+		show_one_pdp_v4only(vty, member->peer, pdp_v4only);
 	}
 }
 
 /* show all (active) PDP contexts within an APN */
 static void apn_show_pdp_contexts(struct vty *vty, struct apn_ctx *apn)
 {
-	ippool_show_pdp_contexts(vty, apn->v4.pool);
-	ippool_show_pdp_contexts(vty, apn->v6.pool);
+	ippool_show_pdp_contexts(vty, apn->v4.pool, true);
+	ippool_show_pdp_contexts(vty, apn->v6.pool, false);
 }
 
 DEFUN(show_pdpctx, show_pdpctx_cmd,
-	"show pdp-context ggsn NAME [apn APN]",
+	"show pdp-context ggsn NAME",
 	SHOW_STR "Show PDP Context Information\n"
-	GGSN_STR "GGSN Name\n") // FIXME
+	GGSN_STR "GGSN Name\n")
 {
 	struct ggsn_ctx *ggsn;
 	struct apn_ctx *apn;
@@ -797,21 +1005,45 @@ DEFUN(show_pdpctx, show_pdpctx_cmd,
 		vty_out(vty, "%% No such GGSN '%s'%s", argv[0], VTY_NEWLINE);
 		return CMD_WARNING;
 	}
-	if (argc < 2) {
-		llist_for_each_entry(apn, &ggsn->apn_list, list)
-			apn_show_pdp_contexts(vty, apn);
-	} else {
-		apn = ggsn_find_apn(ggsn, argv[1]);
-		if (!apn) {
-			vty_out(vty, "%% No such APN '%s'%s", argv[1], VTY_NEWLINE);
-			return CMD_WARNING;
-		}
+
+	llist_for_each_entry(apn, &ggsn->apn_list, list)
 		apn_show_pdp_contexts(vty, apn);
-	}
 
 	return CMD_SUCCESS;
 }
 
+DEFUN(show_pdpctx_apn, show_pdpctx_apn_cmd,
+	"show pdp-context ggsn NAME apn APN",
+	SHOW_STR "Show PDP Context Information\n"
+	GGSN_STR "GGSN Name\n" "Filter by APN\n" "APN name\n")
+{
+	struct ggsn_ctx *ggsn;
+	struct apn_ctx *apn;
+
+	ggsn = ggsn_find(argv[0]);
+	if (!ggsn) {
+		vty_out(vty, "%% No such GGSN '%s'%s", argv[0], VTY_NEWLINE);
+		return CMD_WARNING;
+	}
+
+	apn = ggsn_find_apn(ggsn, argv[1]);
+	if (!apn) {
+		vty_out(vty, "%% No such APN '%s'%s", argv[1], VTY_NEWLINE);
+		return CMD_WARNING;
+	}
+
+	apn_show_pdp_contexts(vty, apn);
+	return CMD_SUCCESS;
+}
+
+/* Backwards compatibility: the VTY parser is (mis)interpreting
+ * "[apn APN]" as two separate elements: "[apn" and "APN]",
+ * but the first part somehow turns into command "ap". */
+ALIAS_DEPRECATED(show_pdpctx_apn, show_deprecated_pdpctx_apn_cmd,
+		 "show pdp-context ggsn NAME ap APN",
+		 SHOW_STR "Show PDP Context Information\n"
+		 GGSN_STR "GGSN Name\n" "Filter by APN\n" "APN name\n");
+
 static void show_apn(struct vty *vty, struct apn_ctx *apn)
 {
 	vty_out(vty, " APN: %s%s", apn->cfg.name, VTY_NEWLINE);
@@ -821,12 +1053,15 @@ static void show_apn(struct vty *vty, struct apn_ctx *apn)
 static void show_one_ggsn(struct vty *vty, struct ggsn_ctx *ggsn)
 {
 	struct apn_ctx *apn;
+	struct sgsn_peer *sgsn;
 	vty_out(vty, "GGSN %s: Bound to %s%s", ggsn->cfg.name, in46a_ntoa(&ggsn->cfg.listen_addr),
 		VTY_NEWLINE);
 	/* FIXME */
 
 	llist_for_each_entry(apn, &ggsn->apn_list, list)
 		show_apn(vty, apn);
+	llist_for_each_entry(sgsn, &ggsn->sgsn_list, entry)
+		show_one_sgsn(vty, sgsn, " ");
 }
 
 DEFUN(show_ggsn, show_ggsn_cmd,
@@ -851,13 +1086,16 @@ DEFUN(show_ggsn, show_ggsn_cmd,
 int ggsn_vty_init(void)
 {
 	install_element_ve(&show_pdpctx_cmd);
+	install_element_ve(&show_pdpctx_apn_cmd);
+	install_element_ve(&show_deprecated_pdpctx_apn_cmd);
 	install_element_ve(&show_pdpctx_imsi_cmd);
+	install_element_ve(&show_pdpctx_ip_cmd);
 	install_element_ve(&show_ggsn_cmd);
 
 	install_element(CONFIG_NODE, &cfg_ggsn_cmd);
 	install_element(CONFIG_NODE, &cfg_no_ggsn_cmd);
+
 	install_node(&ggsn_node, config_write_ggsn);
-	vty_install_default(GGSN_NODE);
 	install_element(GGSN_NODE, &cfg_description_cmd);
 	install_element(GGSN_NODE, &cfg_no_description_cmd);
 	install_element(GGSN_NODE, &cfg_ggsn_shutdown_cmd);
@@ -870,9 +1108,13 @@ int ggsn_vty_init(void)
 	install_element(GGSN_NODE, &cfg_ggsn_no_apn_cmd);
 	install_element(GGSN_NODE, &cfg_ggsn_default_apn_cmd);
 	install_element(GGSN_NODE, &cfg_ggsn_no_default_apn_cmd);
+	install_element(GGSN_NODE, &cfg_ggsn_show_sgsn_cmd);
+	install_element(GGSN_NODE, &cfg_ggsn_echo_interval_cmd);
+	install_element(GGSN_NODE, &cfg_ggsn_no_echo_interval_cmd);
+
+	osmo_tdef_vty_groups_init(GGSN_NODE, ggsn_tdef_group);
 
 	install_node(&apn_node, NULL);
-	vty_install_default(APN_NODE);
 	install_element(APN_NODE, &cfg_description_cmd);
 	install_element(APN_NODE, &cfg_no_description_cmd);
 	install_element(APN_NODE, &cfg_apn_shutdown_cmd);
@@ -894,6 +1136,8 @@ int ggsn_vty_init(void)
 	install_element(APN_NODE, &cfg_apn_no_ip_ifconfig_cmd);
 	install_element(APN_NODE, &cfg_apn_ipv6_ifconfig_cmd);
 	install_element(APN_NODE, &cfg_apn_no_ipv6_ifconfig_cmd);
+	install_element(APN_NODE, &cfg_apn_ipv6_linklocal_cmd);
+	install_element(APN_NODE, &cfg_apn_no_ipv6_linklocal_cmd);
 	install_element(APN_NODE, &cfg_apn_gpdu_seq_cmd);
 	install_element(APN_NODE, &cfg_apn_no_gpdu_seq_cmd);
 
@@ -927,6 +1171,10 @@ static int ggsn_vty_go_parent(struct vty *vty)
 			vty->index_sub = &apn->ggsn->cfg.description;
 		}
 		break;
+	default:
+		vty->node = CONFIG_NODE;
+		vty->index = NULL;
+		vty->index_sub = NULL;
 	}
 
 	return vty->node;

ggsn/gtp-kernel.c

@@ -1,223 +0,0 @@
-#ifdef __linux__
-#define _GNU_SOURCE 1		/* strdup() prototype, broken arpa/inet.h */
-#endif
-
-#include "../config.h"
-
-#ifdef HAVE_STDINT_H
-#include <stdint.h>
-#endif
-
-#include <stdio.h>
-#include <string.h>
-#include <stdlib.h>
-#include <sys/types.h>
-#include <arpa/inet.h>
-#include <net/if.h>
-
-#include <libgtpnl/gtp.h>
-#include <libgtpnl/gtpnl.h>
-#include <libmnl/libmnl.h>
-
-#include <errno.h>
-
-#include <time.h>
-
-#include "../lib/tun.h"
-#include "../lib/syserr.h"
-#include "../gtp/pdp.h"
-#include "../gtp/gtp.h"
-
-#include <libgtpnl/gtp.h>
-#include <libgtpnl/gtpnl.h>
-#include <libmnl/libmnl.h>
-
-#include "gtp-kernel.h"
-
-static void pdp_debug(struct pdp_t *pdp)
-{
-	int i;
-	uint64_t teid;
-
-	if (!debug)
-		return;
-
-	printf("version %u\n", pdp->version);
-	if (pdp->version == 0) {
-		teid = pdp_gettid(pdp->imsi, pdp->nsapi);
-		printf("flowid %u\n", pdp->flru);
-	} else {
-		teid = pdp->teid_gn; /* GTPIE_TEI_DI */
-	}
-
-	printf("teid %llx\n", teid);
-	printf("address (%u)\n", pdp->eua.l);
-
-	/* Byte 0: 0xf1 == IETF */
-	/* Byte 1: 0x21 == IPv4 */
-	/* Byte 2-6: IPv4 address */
-
-	for (i = 0; i < 6; i++)
-		printf("%x ", pdp->eua.v[i] & 0xff); /* GTPIE_EUA */
-
-	printf("\n");
-	printf("sgsn-addr (%u)\n", pdp->gsnrc.l);
-
-	for (i = 0; i < 4; i++)
-		printf("%x ", pdp->gsnrc.v[i] & 0xff); /* GTPIE_GSN_ADDR */
-
-	printf("\n");
-}
-
-static struct {
-	int			genl_id;
-	struct mnl_socket	*nl;
-	bool			enabled;
-} gtp_nl;
-
-/* Always forces the kernel to allocate gtp0. If it exists it hits EEXIST */
-#define GTP_DEVNAME	"gtp0"
-
-int gtp_kernel_init(struct gsn_t *gsn, struct in_addr *net,
-		    size_t prefixlen, const char *net_arg)
-{
-	if (gtp_dev_create(-1, GTP_DEVNAME, gsn->fd0, gsn->fd1u) < 0) {
-		SYS_ERR(DGGSN, LOGL_ERROR, 0,
-			"cannot create GTP tunnel device: %s\n",
-			strerror(errno));
-		return -1;
-	}
-	gtp_nl.enabled = true;
-
-	gtp_nl.nl = genl_socket_open();
-	if (gtp_nl.nl == NULL) {
-		SYS_ERR(DGGSN, LOGL_ERROR, 0,
-			"cannot create genetlink socket\n");
-		return -1;
-	}
-	gtp_nl.genl_id = genl_lookup_family(gtp_nl.nl, "gtp");
-	if (gtp_nl.genl_id < 0) {
-		SYS_ERR(DGGSN, LOGL_ERROR, 0,
-			"cannot lookup GTP genetlink ID\n");
-		return -1;
-	}
-	if (debug) {
-		SYS_ERR(DGGSN, LOGL_NOTICE, 0,
-			"Using the GTP kernel mode (genl ID is %d)\n",
-			gtp_nl.genl_id);
-	}
-
-	DEBUGP(DGGSN, "Setting route to reach %s via %s\n",
-	       net_arg, GTP_DEVNAME);
-
-	if (gtp_dev_config(GTP_DEVNAME, net, prefixlen) < 0) {
-		SYS_ERR(DGGSN, LOGL_ERROR, 0,
-			"Cannot add route to reach network %s\n",
-			net_arg);
-	}
-
-	/* launch script if it is set to bring up the route to reach
-	 * the MS, eg. ip ro add 10.0.0.0/8 dev gtp0. Better add this
-	 * using native rtnetlink interface given that we know the
-	 * MS network mask, later.
-	 */
-	if (ipup) {
-		char cmd[1024];
-		int err;
-
-		/* eg. /home/ggsn/ipup gtp0 10.0.0.0/8 */
-		snprintf(cmd, sizeof(cmd), "%s %s %s",
-			 ipup, GTP_DEVNAME, net_arg);
-		cmd[sizeof(cmd)-1] = '\0';
-
-		err = system(cmd);
-		if (err < 0) {
-			SYS_ERR(DGGSN, LOGL_ERROR, 0,
-				"Failed to launch script `%s'", ipup);
-			return -1;
-		}
-	}
-	SYS_ERR(DGGSN, LOGL_NOTICE, 0, "GTP kernel configured\n");
-
-	return 0;
-}
-
-void gtp_kernel_stop(void)
-{
-	if (!gtp_nl.enabled)
-		return;
-
-	gtp_dev_destroy(GTP_DEVNAME);
-}
-
-int gtp_kernel_tunnel_add(struct pdp_t *pdp)
-{
-	struct in_addr ms, sgsn;
-	struct gtp_tunnel *t;
-	int ret;
-
-	if (!gtp_nl.enabled)
-		return 0;
-
-	pdp_debug(pdp);
-
-	t = gtp_tunnel_alloc();
-	if (t == NULL)
-		return -1;
-
-	memcpy(&ms, &pdp->eua.v[2], sizeof(struct in_addr));
-	memcpy(&sgsn, &pdp->gsnrc.v[0], sizeof(struct in_addr));
-
-	gtp_tunnel_set_ifidx(t, if_nametoindex(GTP_DEVNAME));
-	gtp_tunnel_set_version(t, pdp->version);
-	gtp_tunnel_set_ms_ip4(t, &ms);
-	gtp_tunnel_set_sgsn_ip4(t, &sgsn);
-	if (pdp->version == 0) {
-		gtp_tunnel_set_tid(t, pdp_gettid(pdp->imsi, pdp->nsapi));
-		gtp_tunnel_set_flowid(t, pdp->flru);
-	} else {
-		gtp_tunnel_set_i_tei(t, pdp->teid_own);
-		/* use the TEI advertised by SGSN when sending packets
-		 * towards the SGSN */
-		gtp_tunnel_set_o_tei(t, pdp->teid_gn);
-	}
-
-	ret = gtp_add_tunnel(gtp_nl.genl_id, gtp_nl.nl, t);
-	gtp_tunnel_free(t);
-
-	return ret;
-}
-
-int gtp_kernel_tunnel_del(struct pdp_t *pdp)
-{
-	struct gtp_tunnel *t;
-	int ret;
-
-	if (!gtp_nl.enabled)
-		return 0;
-
-	pdp_debug(pdp);
-
-	t = gtp_tunnel_alloc();
-	if (t == NULL)
-		return -1;
-
-	gtp_tunnel_set_ifidx(t, if_nametoindex(GTP_DEVNAME));
-	gtp_tunnel_set_version(t, pdp->version);
-	if (pdp->version == 0) {
-		gtp_tunnel_set_tid(t, pdp_gettid(pdp->imsi, pdp->nsapi));
-		gtp_tunnel_set_flowid(t, pdp->flru);
-	} else {
-		gtp_tunnel_set_i_tei(t, pdp->teid_own);
-	}
-
-	ret = gtp_del_tunnel(gtp_nl.genl_id, gtp_nl.nl, t);
-	gtp_tunnel_free(t);
-
-	return ret;
-}
-
-int gtp_kernel_enabled(void)
-{
-	return gtp_nl.enabled;
-}

ggsn/gtp-kernel.h

@@ -1,45 +0,0 @@
-#ifndef _GTP_KERNEL_H_
-#define _GTP_KERNEL_H_
-
-struct gengetopt_args_info;
-
-extern int debug;
-extern char *ipup;
-
-#ifdef GTP_KERNEL
-int gtp_kernel_init(struct gsn_t *gsn, struct in_addr *net,
-		    size_t prefixlen, const char *net_arg);
-void gtp_kernel_stop(void);
-
-int gtp_kernel_tunnel_add(struct pdp_t *pdp);
-int gtp_kernel_tunnel_del(struct pdp_t *pdp);
-
-int gtp_kernel_enabled(void);
-
-#else
-static inline int gtp_kernel_init(struct gsn_t *gsn, struct in_addr *net,
-				  size_t prefixlen, const char *net_arg)
-{
-	SYS_ERR(DGGSN, LOGL_ERROR, 0, "ggsn compiled without GTP kernel support!\n");
-	return -1;
-}
-
-static inline void gtp_kernel_stop(void) {}
-
-static inline int gtp_kernel_tunnel_add(struct pdp_t *pdp)
-{
-	return 0;
-}
-
-static inline int gtp_kernel_tunnel_del(struct pdp_t *pdp)
-{
-	return 0;
-}
-
-static inline int gtp_kernel_enabled(void)
-{
-	return 0;
-}
-
-#endif
-#endif /* _GTP_KERNEL_H_ */

ggsn/icmpv6.h

@@ -1,7 +0,0 @@
-#pragma once
-
-#include "../gtp/gtp.h"
-#include "../gtp/pdp.h"
-
-int handle_router_mcast(struct gsn_t *gsn, struct pdp_t *pdp, const struct in6_addr *own_ll_addr,
-			const uint8_t *pack, unsigned len);

ggsn/pco.c

@@ -0,0 +1,253 @@
+/*
+ * PCO parsing related code
+ * Copyright (C) 2002, 2003, 2004 Mondru AB.
+ * Copyright (C) 2017-2019 by Harald Welte <laforge@gnumonks.org>
+ * Copyright (C) 2019 sysmocom - s.f.m.c. GmbH <info@sysmocom.de>
+ *
+ * The contents of this file may be used under the terms of the GNU
+ * General Public License Version 2, provided that the above copyright
+ * notice and this permission notice is included in all copies or
+ * substantial portions of the software.
+ *
+ */
+
+#include <unistd.h>
+#include <string.h>
+
+#include <osmocom/core/msgb.h>
+#include <osmocom/gsm/tlv.h>
+
+#include "../lib/util.h"
+
+#include "pco.h"
+#include "ggsn.h"
+
+/* determine if IPCP contains given option */
+static const uint8_t *ipcp_contains_option(const struct ipcp_hdr *ipcp, size_t ipcp_len,
+					   enum ipcp_options opt, size_t opt_minlen)
+{
+	const uint8_t *cur_opt = ipcp->options;
+
+	/* iterate over Options and check if protocol contained */
+	while (cur_opt + sizeof(struct ipcp_option_hdr) <= (uint8_t*)ipcp + ipcp_len) {
+		const struct ipcp_option_hdr *cur_opt_hdr = (const struct ipcp_option_hdr *)cur_opt;
+		/* length value includes 2 bytes type/length */
+		if (cur_opt_hdr->len < sizeof(struct ipcp_option_hdr))
+			return NULL;
+		if (cur_opt_hdr->type == opt &&
+		    cur_opt_hdr->len >= sizeof(struct ipcp_option_hdr) + opt_minlen)
+			return cur_opt;
+		cur_opt += cur_opt_hdr->len;
+	}
+	return NULL;
+}
+
+
+static const char *pap_welcome = "Welcome to OsmoGGSN " PACKAGE_VERSION;
+
+/* Handle PAP protocol according to RFC 1334 */
+static void process_pco_element_pap(const struct pco_element *pco_in, struct msgb *resp,
+				    const struct apn_ctx *apn, struct pdp_t *pdp)
+{
+	const struct pap_element *pap_in = (const struct pap_element *) pco_in->data;
+	uint16_t pap_in_len;
+	uint8_t peer_id_len;
+	const uint8_t *peer_id;
+	unsigned int pap_welcome_len;
+	uint8_t pap_out_size;
+	struct pap_element *pap_out;
+
+	if (pco_in->length < sizeof(struct pap_element))
+		goto ret_broken;
+
+	pap_in_len = osmo_load16be(&pap_in->len);
+	if (pco_in->length < pap_in_len)
+		goto ret_broken;
+	/* "pco_in->length > pap_in_len" is allowed: RFC1334 2.2 states:
+	   "Octets outside the range of the Length field should be treated as
+	   Data Link Layer padding and should be ignored on reception."
+	 */
+
+	switch (pap_in->code) {
+	case PAP_CODE_AUTH_REQ:
+		if (pap_in_len < sizeof(struct pap_element) + 1)
+			goto ret_broken_auth;
+		peer_id_len = pap_in->data[0];
+		if (pap_in_len < sizeof(struct pap_element) + 1 + peer_id_len)
+			goto ret_broken_auth;
+		peer_id = &pap_in->data[1];
+		LOGPPDP(LOGL_DEBUG, pdp, "PCO PAP PeerId = %s, ACKing\n",
+			osmo_quote_str((const char *)peer_id, peer_id_len));
+		/* Password-Length + Password following here, but we don't care */
+
+		/* Prepare response, we ACK all of them: */
+		pap_welcome_len = strlen(pap_welcome);
+		/* +1: Length field of pap_welcome Message */
+		pap_out_size = sizeof(struct pap_element) + 1 + pap_welcome_len;
+		pap_out = alloca(pap_out_size);
+		pap_out->code = PAP_CODE_AUTH_ACK;
+		pap_out->id = pap_in->id;
+		pap_out->len = htons(pap_out_size);
+		pap_out->data[0] = pap_welcome_len;
+		memcpy(pap_out->data+1, pap_welcome, pap_welcome_len);
+		msgb_t16lv_put(resp, PCO_P_PAP, pap_out_size, (uint8_t *) pap_out);
+		break;
+	case PAP_CODE_AUTH_ACK:
+	case PAP_CODE_AUTH_NAK:
+	default:
+		LOGPPDP(LOGL_NOTICE, pdp, "Unsupported PAP PCO Code %u, ignoring\n", pap_in->code);
+		break;
+	}
+	return;
+
+ret_broken_auth:
+	LOGPPDP(LOGL_NOTICE, pdp, "Invalid PAP AuthenticateReq: %s, ignoring\n",
+		osmo_hexdump_nospc((const uint8_t *)pco_in, pco_in->length));
+	return;
+
+ret_broken:
+	LOGPPDP(LOGL_NOTICE, pdp, "Invalid PAP PCO Length: %s, ignoring\n",
+		osmo_hexdump_nospc((const uint8_t *)pco_in, pco_in->length));
+}
+
+/* Handle IP Control Protocol, RFC 1332, extensions in RFC 1877 */
+static void process_pco_element_ipcp(const struct pco_element *pco_elem, struct msgb *resp,
+				     const struct apn_ctx *apn, struct pdp_t *pdp)
+{
+	struct ippoolm_t *peer_v4 = pdp_get_peer_ipv(pdp, false);
+	const struct in46_addr *dns1 = &apn->v4.cfg.dns[0];
+	const struct in46_addr *dns2 = &apn->v4.cfg.dns[1];
+	uint8_t *start = resp->tail;
+	const struct ipcp_hdr *ipcp;
+	uint16_t ipcp_len;
+	uint8_t *len1, *len2;
+	unsigned int len_appended;
+	ptrdiff_t consumed;
+	size_t remain;
+
+	if (!peer_v4) {
+		LOGPPDP(LOGL_ERROR, pdp, "IPCP but no IPv4 type ?!?\n");
+		return;
+	}
+
+	ipcp = (const struct ipcp_hdr *)pco_elem->data;
+	consumed = (pco_elem->data - &pdp->pco_req.v[0]);
+	remain = sizeof(pdp->pco_req.v) - consumed;
+	ipcp_len = osmo_load16be(&ipcp->len);
+	if (remain < 0 || remain < ipcp_len) {
+		LOGPPDP(LOGL_ERROR, pdp, "Malformed IPCP, ignoring\n");
+		return;
+	}
+
+	/* Three byte T16L header */
+	msgb_put_u16(resp, 0x8021);	/* IPCP */
+	len1 = msgb_put(resp, 1);	/* Length of contents: delay */
+
+	msgb_put_u8(resp, 0x02);	/* ACK */
+	msgb_put_u8(resp, ipcp->id);	/* ID: Needs to match request */
+	msgb_put_u8(resp, 0x00);	/* Length MSB */
+	len2 = msgb_put(resp, 1);	/* Length LSB: delay */
+
+	if (dns1->len == 4 && ipcp_contains_option(ipcp, ipcp_len, IPCP_OPT_PRIMARY_DNS, 4)) {
+		msgb_put_u8(resp, 0x81);		/* DNS1 Tag */
+		msgb_put_u8(resp, 2 + dns1->len);	/* DNS1 Length, incl. TL */
+		msgb_put_u32(resp, ntohl(dns1->v4.s_addr));
+	}
+
+	if (dns2->len == 4 && ipcp_contains_option(ipcp, ipcp_len, IPCP_OPT_SECONDARY_DNS, 4)) {
+		msgb_put_u8(resp, 0x83);		/* DNS2 Tag */
+		msgb_put_u8(resp, 2 + dns2->len);	/* DNS2 Length, incl. TL */
+		msgb_put_u32(resp, ntohl(dns2->v4.s_addr));
+	}
+
+	/* patch in length values */
+	len_appended = resp->tail - start;
+	*len1 = len_appended - 3;
+	*len2 = len_appended - 3;
+}
+
+static void process_pco_element_dns_ipv6(const struct pco_element *pco_elem, struct msgb *resp,
+					 const struct apn_ctx *apn, struct pdp_t *pdp)
+{
+	unsigned int i;
+	const uint8_t *tail = resp->tail;
+
+	for (i = 0; i < ARRAY_SIZE(apn->v6.cfg.dns); i++) {
+		const struct in46_addr *i46a = &apn->v6.cfg.dns[i];
+		if (i46a->len != 16)
+			continue;
+		msgb_t16lv_put(resp, PCO_P_DNS_IPv6_ADDR, i46a->len, i46a->v6.s6_addr);
+	}
+	if (resp->tail == tail)
+		LOGPPDP(LOGL_NOTICE, pdp, "MS requested IPv6 DNS, but APN has none configured\n");
+}
+
+static void process_pco_element_dns_ipv4(const struct pco_element *pco_elem, struct msgb *resp,
+					 const struct apn_ctx *apn, struct pdp_t *pdp)
+{
+	unsigned int i;
+	const uint8_t *tail = resp->tail;
+
+	for (i = 0; i < ARRAY_SIZE(apn->v4.cfg.dns); i++) {
+		const struct in46_addr *i46a = &apn->v4.cfg.dns[i];
+		if (i46a->len != 4)
+			continue;
+		msgb_t16lv_put(resp, PCO_P_DNS_IPv4_ADDR, i46a->len, (uint8_t *)&i46a->v4);
+	}
+	if (resp->tail == tail)
+		LOGPPDP(LOGL_NOTICE, pdp, "MS requested IPv4 DNS, but APN has none configured\n");
+}
+
+static void process_pco_element(const struct pco_element *pco_elem, struct msgb *resp,
+				const struct apn_ctx *apn, struct pdp_t *pdp)
+{
+	uint16_t protocol_id = osmo_load16be(&pco_elem->protocol_id);
+
+	LOGPPDP(LOGL_DEBUG, pdp, "PCO Protocol 0x%04x\n", protocol_id);
+	switch (protocol_id) {
+	case PCO_P_PAP:
+		process_pco_element_pap(pco_elem, resp, apn, pdp);
+		break;
+	case PCO_P_IPCP:
+		process_pco_element_ipcp(pco_elem, resp, apn, pdp);
+		break;
+	case PCO_P_DNS_IPv6_ADDR:
+		process_pco_element_dns_ipv6(pco_elem, resp, apn, pdp);
+		break;
+	case PCO_P_DNS_IPv4_ADDR:
+		process_pco_element_dns_ipv4(pco_elem, resp, apn, pdp);
+		break;
+	default:
+		LOGPPDP(LOGL_INFO, pdp, "Unknown/Unimplemented PCO Protocol 0x%04x: %s\n",
+			protocol_id, osmo_hexdump_nospc(pco_elem->data, pco_elem->length));
+		break;
+	}
+}
+
+/* process one PCO request from a MS/UE, putting together the proper responses */
+void process_pco(const struct apn_ctx *apn, struct pdp_t *pdp)
+{
+	struct msgb *resp = msgb_alloc(256, "PCO.resp");
+	const struct ul255_t *pco = &pdp->pco_req;
+	const struct pco_element *pco_elem;
+	const uint8_t *cur;
+
+	/* build the header of the PCO response */
+	OSMO_ASSERT(resp);
+	msgb_put_u8(resp, 0x80); /* ext-bit + configuration protocol byte */
+
+	/* iterate over the PCO elements in the request; call process_pco_element() for each */
+	for (cur = pco->v + 1, pco_elem = (const struct pco_element *) cur;
+	     cur + sizeof(struct pco_element) <= pco->v + pco->l;
+	     cur += pco_elem->length + sizeof(*pco_elem), pco_elem = (const struct pco_element *) cur) {
+		process_pco_element(pco_elem, resp, apn, pdp);
+	}
+
+	/* copy the PCO response msgb and copy its contents over to the PDP context */
+	if (msgb_length(resp) > 1) {
+		memcpy(pdp->pco_neg.v, msgb_data(resp), msgb_length(resp));
+		pdp->pco_neg.l = msgb_length(resp);
+	} else
+		pdp->pco_neg.l = 0;
+	msgb_free(resp);
+}

ggsn/pco.h

@@ -0,0 +1,80 @@
+#pragma once
+
+#include <stdint.h>
+
+#include "../gtp/pdp.h"
+
+/* 3GPP TS 24.008 10.5.6.3 */
+enum pco_protocols {
+	PCO_P_LCP		= 0xC021,
+	PCO_P_PAP		= 0xC023,
+	PCO_P_CHAP		= 0xC223,
+	PCO_P_IPCP		= 0x8021,
+	PCO_P_PCSCF_ADDR	= 0x0001,
+	PCO_P_IM_CN_SS_F	= 0x0002,
+	PCO_P_DNS_IPv6_ADDR	= 0x0003,
+	PCO_P_POLICY_CTRL_REJ	= 0x0004,	/* only in Network->MS */
+	PCO_P_MS_SUP_NETREQ_BCI	= 0x0005,
+	/* reserved */
+	PCO_P_DSMIPv6_HA_ADDR	= 0x0007,
+	PCO_P_DSMIPv6_HN_PREF	= 0x0008,
+	PCO_P_DSMIPv6_v4_HA_ADDR= 0x0009,
+	PCO_P_IP_ADDR_VIA_NAS	= 0x000a,	/* only MS->Network */
+	PCO_P_IPv4_ADDR_VIA_DHCP= 0x000b,	/* only MS->Netowrk */
+	PCO_P_PCSCF_IPv4_ADDR	= 0x000c,
+	PCO_P_DNS_IPv4_ADDR	= 0x000d,
+	PCO_P_MSISDN		= 0x000e,
+	PCO_P_IFOM_SUPPORT	= 0x000f,
+	PCO_P_IPv4_LINK_MTU	= 0x0010,
+	PCO_P_MS_SUPP_LOC_A_TFT	= 0x0011,
+	PCO_P_PCSCF_RESEL_SUP	= 0x0012,	/* only MS->Network */
+	PCO_P_NBIFOM_REQ	= 0x0013,
+	PCO_P_NBIFOM_MODE	= 0x0014,
+	PCO_P_NONIP_LINK_MTU	= 0x0015,
+	PCO_P_APN_RATE_CTRL_SUP	= 0x0016,
+	PCO_P_PS_DATA_OFF_UE	= 0x0017,
+	PCO_P_REL_DATA_SVC	= 0x0018,
+};
+
+struct pco_element {
+	uint16_t protocol_id;	/* network byte order */
+	uint8_t length;		/* length of data below */
+	uint8_t data[0];
+} __attribute__((packed));
+
+/* RFC 1332 IP Control Protocol options, extensions in RFC 1877 */
+enum ipcp_options {
+	IPCP_OPT_IPADDR = 3, /* RFC 1332 3.3 */
+	IPCP_OPT_PRIMARY_DNS = 129, /* RFC 1877 1.1 */
+	IPCP_OPT_SECONDARY_DNS = 131,  /* RFC 1877 1.2 */
+};
+
+struct ipcp_option_hdr {
+	uint8_t type;
+	uint8_t len;
+	uint8_t data[0];
+} __attribute__ ((packed));
+
+struct ipcp_hdr {
+	uint8_t code;
+	uint8_t id;
+	uint16_t len;
+	uint8_t options[0];
+} __attribute__ ((packed));
+
+/* RFC 1334, section 3.2. Packet Format */
+struct pap_element {
+	uint8_t code;
+	uint8_t id;
+	uint16_t len; /* length including header */
+	uint8_t data[0];
+} __attribute__((packed));
+
+enum pap_code {
+	PAP_CODE_AUTH_REQ = 1,
+	PAP_CODE_AUTH_ACK = 2,
+	PAP_CODE_AUTH_NAK = 3,
+};
+
+struct apn_ctx;
+void process_pco(const struct apn_ctx *apn, struct pdp_t *pdp);

ggsn/sgsn.c

@@ -0,0 +1,168 @@
+#include "sgsn.h"
+#include "ggsn.h"
+
+
+static bool sgsn_peer_attempt_free(struct sgsn_peer *sgsn)
+{
+	/* We have to be careful here, since if all pdp ctx for that sgsn were
+	   deactivated in-between we sent the Echo Req and receivied the timeout
+	   indication, the sgsn (cbp) may be already gone. We need to add some
+	   counter reference of echo requets in flight and only free sgsn
+	   structures when it goes to zero decreased for all Echo Resp. We do it
+	   this way because currently in libgtp there's no understanding of "gsn
+	   peer" for which messages are grouped and hence we cannot request
+	   libgtp to drop all queued messages for a specific peer. */
+	if (sgsn->tx_msgs_queued) {
+		LOGSGSN(LOGL_INFO, sgsn, "Delaying delete, still %u echo messages queued\n",
+			sgsn->tx_msgs_queued);
+		return false;
+	}
+	llist_del(&sgsn->entry);
+	LOGSGSN(LOGL_INFO, sgsn, "Deleting SGSN\n");
+	talloc_free(sgsn);
+	return true;
+}
+
+static void sgsn_peer_echo_req(struct sgsn_peer *sgsn)
+{
+	struct ggsn_ctx *ggsn = sgsn->ggsn;
+	LOGSGSN(LOGL_INFO, sgsn, "Tx Echo Request\n");
+	gtp_echo_req(ggsn->gsn, sgsn->gtp_version, sgsn, &sgsn->addr);
+	sgsn->tx_msgs_queued++;
+}
+
+void sgsn_peer_echo_resp(struct sgsn_peer *sgsn, bool timeout)
+{
+	if (timeout) {
+		LOGSGSN(LOGL_NOTICE, sgsn, "Rx Echo Request timed out!\n");
+		sgsn_peer_drop_all_pdp(sgsn);
+	} else {
+		LOGSGSN(LOGL_INFO, sgsn, "Rx Echo Response\n");
+	}
+
+	/* We decrement it here after dropping all pdps to make sure sgsn was
+	   not freed upon last pdp ctx deleted and is still alive now */
+	sgsn->tx_msgs_queued--;
+	if (llist_empty(&sgsn->pdp_list))
+		sgsn_peer_attempt_free(sgsn);
+}
+
+void sgsn_echo_timer_start(struct sgsn_peer *sgsn)
+{
+	if (sgsn->ggsn->cfg.echo_interval == 0)
+		return;
+	sgsn_peer_echo_req(sgsn);
+	osmo_timer_schedule(&sgsn->echo_timer, sgsn->ggsn->cfg.echo_interval, 0);
+}
+
+void sgsn_echo_timer_stop(struct sgsn_peer *sgsn)
+{
+	osmo_timer_del(&sgsn->echo_timer);
+}
+
+static void sgsn_echo_timer_cb(void *data)
+{
+	struct sgsn_peer *sgsn = (struct sgsn_peer *) data;
+	sgsn_echo_timer_start(sgsn);
+}
+
+struct sgsn_peer *sgsn_peer_allocate(struct ggsn_ctx *ggsn, struct in_addr *ia, unsigned int gtp_version)
+{
+	struct sgsn_peer *sgsn;
+
+	sgsn = talloc_zero_size(ggsn, sizeof(struct sgsn_peer));
+	sgsn->ggsn = ggsn;
+	sgsn->addr = *ia;
+	sgsn->gtp_version = gtp_version;
+	sgsn->remote_restart_ctr = -1;
+	INIT_LLIST_HEAD(&sgsn->pdp_list);
+	INIT_LLIST_HEAD(&sgsn->entry);
+
+	osmo_timer_setup(&sgsn->echo_timer, sgsn_echo_timer_cb, sgsn);
+
+	LOGSGSN(LOGL_INFO, sgsn, "Discovered\n");
+	return sgsn;
+}
+
+void sgsn_peer_add_pdp_priv(struct sgsn_peer *sgsn, struct pdp_priv_t *pdp_priv)
+{
+	bool was_empty = llist_empty(&sgsn->pdp_list);
+	pdp_priv->sgsn = sgsn;
+	llist_add(&pdp_priv->entry, &sgsn->pdp_list);
+	if (was_empty)
+		sgsn_echo_timer_start(sgsn);
+}
+
+void sgsn_peer_remove_pdp_priv(struct pdp_priv_t* pdp_priv)
+{
+	struct sgsn_peer *sgsn = pdp_priv->sgsn;
+	llist_del(&pdp_priv->entry);
+	if (sgsn && llist_empty(&sgsn->pdp_list)) {
+		/* No PDP contexts associated to this SGSN, no need to keep it */
+		sgsn_echo_timer_stop(sgsn);
+		/* sgsn may not be freed if there are some messages still queued
+		   in libgtp which could return a pointer to it */
+		sgsn_peer_attempt_free(sgsn);
+	}
+
+	pdp_priv->sgsn = NULL;
+}
+
+/* High-level function to be called in case a GGSN has disappeared or
+ * otherwise lost state (recovery procedure). It will detach all related pdp ctx
+ * from a ggsn and communicate deact to MS. Optionally (!NULL), one pdp ctx can
+ * be kept alive to allow handling later message which contained the Recovery IE. */
+static unsigned int sgsn_peer_drop_all_pdp_except(struct sgsn_peer *sgsn, struct pdp_priv_t *except)
+{
+	unsigned int num = 0;
+	char buf[INET_ADDRSTRLEN];
+	unsigned int count = llist_count(&sgsn->pdp_list);
+
+	inet_ntop(AF_INET, &sgsn->addr, buf, sizeof(buf));
+
+	struct pdp_priv_t *pdp, *pdp2;
+	llist_for_each_entry_safe(pdp, pdp2, &sgsn->pdp_list, entry) {
+		if (pdp == except)
+			continue;
+		ggsn_close_one_pdp(pdp->lib);
+		num++;
+		if (num == count) {
+			/* Note: if except is NULL, all pdp contexts are freed and sgsn
+			 * is most probably already freed at this point.
+			 * As a result, last access to sgsn->pdp_list before exiting
+			 * loop would access already freed memory. Avoid it by exiting
+			 * the loop without the last check, and make sure sgsn is not
+			 * accessed after this loop. */
+			 break;
+		}
+	}
+
+	LOGP(DGGSN, LOGL_INFO, "SGSN(%s) Dropped %u PDP contexts\n", buf, num);
+
+	return num;
+}
+
+unsigned int sgsn_peer_drop_all_pdp(struct sgsn_peer *sgsn)
+{
+	return sgsn_peer_drop_all_pdp_except(sgsn, NULL);
+}
+
+int sgsn_peer_handle_recovery(struct sgsn_peer *sgsn, struct pdp_t *pdp, uint8_t recovery)
+{
+	struct pdp_priv_t *pdp_priv = NULL;
+
+	if (sgsn->remote_restart_ctr == -1) {
+		/* First received ECHO RESPONSE, note the restart ctr */
+		sgsn->remote_restart_ctr = recovery;
+	} else if (sgsn->remote_restart_ctr != recovery) {
+		/* counter has changed (SGSN restart): release all PDP */
+		LOGSGSN(LOGL_NOTICE, sgsn, "SGSN recovery (%u->%u) pdp=%p, "
+		     "releasing all%s PDP contexts\n",
+		     sgsn->remote_restart_ctr, recovery, pdp, pdp ? " other" : "");
+		sgsn->remote_restart_ctr = recovery;
+		if (pdp)
+			pdp_priv = pdp->priv;
+		sgsn_peer_drop_all_pdp_except(sgsn, pdp_priv);
+	}
+	return 0;
+}

ggsn/sgsn.h

@@ -0,0 +1,46 @@
+#pragma once
+
+#include <stdint.h>
+#include <stdbool.h>
+#include <netinet/in.h>
+#include <arpa/inet.h>
+
+#include <osmocom/core/linuxlist.h>
+#include <osmocom/core/timer.h>
+
+#include "../gtp/pdp.h"
+
+struct ggsn_ctx;
+struct pdp_priv_t;
+
+struct sgsn_peer {
+	struct llist_head entry; /* to be included into ggsn_ctx */
+	struct ggsn_ctx *ggsn; /* backpointer to ggsn_ctx */
+	struct in_addr addr;	/* Addr of the sgsn peer */
+	unsigned int gtp_version; /* GTP version */
+	int remote_restart_ctr; /* Last received Restart Ctr from sgsn peer, -1 == unknown */
+	/* list of pdp contexts associated with this sgsn */
+	struct llist_head pdp_list;
+	/* Sends echo request towards SGSN on expiration. Echo Resp is received
+	   through cb_recovery2(), and echo Req timeout through
+	   cb_conf(GTP_ECHO_REQ, EOF, NULL, cbp); */
+	struct osmo_timer_list echo_timer;
+	/* Number of GTP messages in libgtp transmit queue */
+	unsigned int tx_msgs_queued;
+};
+
+struct sgsn_peer *sgsn_peer_allocate(struct ggsn_ctx *ggsn, struct in_addr *ia, unsigned int gtp_version);
+void sgsn_peer_add_pdp_priv(struct sgsn_peer *sgsn, struct pdp_priv_t *pdp_priv);
+void sgsn_peer_remove_pdp_priv(struct pdp_priv_t *pdp_priv);
+
+void sgsn_echo_timer_start(struct sgsn_peer *sgsn);
+void sgsn_echo_timer_stop(struct sgsn_peer *sgsn);
+
+void sgsn_peer_echo_resp(struct sgsn_peer *sgsn, bool timeout);
+unsigned int sgsn_peer_drop_all_pdp(struct sgsn_peer *sgsn);
+int sgsn_peer_handle_recovery(struct sgsn_peer *sgsn, struct pdp_t *pdp, uint8_t recovery);
+
+#define LOGSGSN(level, sgsn, fmt, args...) { \
+	char _buf[INET_ADDRSTRLEN]; \
+	LOGP(DGGSN, level, "SGSN(%s): " fmt, inet_ntop(AF_INET, &sgsn->addr, _buf, sizeof(_buf)), ## args); \
+	} while (0)

gtp/Makefile.am

@@ -1,17 +1,35 @@
 # This is _NOT_ the library release version, it's an API version.
 # Please read chapter "Library interface versions" of the libtool documentation
 # before making any modifications: https://www.gnu.org/software/libtool/manual/html_node/Versioning.html
-LIBVERSION=2:0:0
+# If major=current-age is increased, remember to update the dh_strip line in debian/rules!
+LIBVERSION=9:1:3
+
 lib_LTLIBRARIES = libgtp.la
 
-include_HEADERS = gtp.h pdp.h gtpie.h
+include_HEADERS = gtp.h gsn.h pdp.h gtpie.h
 
-AM_CFLAGS = -O2 -fno-builtin -Wall -DSBINDIR='"$(sbindir)"' -ggdb $(LIBOSMOCORE_CFLAGS)
+AM_CFLAGS = \
+	    -fno-builtin \
+	    -Wall \
+	    -DSBINDIR='"$(sbindir)"' \
+	    $(LIBOSMOCORE_CFLAGS) \
+		$(LIBOSMOGSM_CFLAGS) \
+	    $(NULL)
+
+libgtp_la_SOURCES = \
+		    gsn.c \
+		    gsn.h \
+		    gtp.c \
+		    gtp.h \
+		    gtpie.c \
+		    gtpie.h \
+		    lookupa.c \
+		    lookupa.h \
+		    pdp.c \
+		    pdp.h \
+		    queue.c \
+		    queue.h \
+		    $(NULL)
 
-libgtp_la_SOURCES = gtp.c gtp.h gtpie.c gtpie.h pdp.c pdp.h lookupa.c lookupa.h queue.c queue.h
 libgtp_la_LDFLAGS = -version-info $(LIBVERSION) -no-undefined
-libgtp_la_LIBADD = $(LIBOSMOCORE_LIBS)
-
-
-
-
+libgtp_la_LIBADD = $(LIBOSMOCORE_LIBS) $(LIBOSMOGSM_LIBS)

gtp/gsn.c

@@ -0,0 +1,582 @@
+/*
+ *  OsmoGGSN - Gateway GPRS Support Node
+ *  Copyright (C) 2002, 2003, 2004 Mondru AB.
+ *  Copyright (C) 2010-2011, 2016-2017 Harald Welte <laforge@gnumonks.org>
+ *  Copyright (C) 2015-2017 sysmocom - s.f.m.c. GmbH
+ *
+ *  The contents of this file may be used under the terms of the GNU
+ *  General Public License Version 2, provided that the above copyright
+ *  notice and this permission notice is included in all copies or
+ *  substantial portions of the software.
+ *
+ */
+
+/*
+ * gtp.c: Contains all GTP functionality. Should be able to handle multiple
+ * tunnels in the same program.
+ *
+ * TODO:
+ *  - Do we need to handle fragmentation?
+ */
+
+#ifdef __linux__
+#define _GNU_SOURCE 1
+#endif
+
+#include <osmocom/core/logging.h>
+#include <osmocom/core/utils.h>
+#include <osmocom/core/stats.h>
+#include <osmocom/core/rate_ctr.h>
+
+#if defined(__FreeBSD__)
+#include <sys/endian.h>
+#endif
+
+#include "../config.h"
+#ifdef HAVE_STDINT_H
+#include <stdint.h>
+#endif
+
+#include <stdio.h>
+#include <stdarg.h>
+#include <stdlib.h>
+#include <sys/time.h>
+#include <sys/types.h>
+#include <sys/socket.h>
+#include <netinet/in.h>
+#include <arpa/inet.h>
+#include <sys/stat.h>
+#include <time.h>
+#include <unistd.h>
+#include <string.h>
+#include <errno.h>
+#include <fcntl.h>
+#include <inttypes.h>
+
+#include <arpa/inet.h>
+
+/* #include <stdint.h>  ISO C99 types */
+
+#include "pdp.h"
+#include "gtp.h"
+#include "gtpie.h"
+#include "queue.h"
+
+/* Error reporting functions */
+
+#define LOGP_WITH_ADDR(ss, level, addr, fmt, args...)                    \
+		LOGP(ss, level, "addr(%s:%d) " fmt,                      \
+		     inet_ntoa((addr).sin_addr), htons((addr).sin_port), \
+		     ##args)
+
+static const struct rate_ctr_desc gsn_ctr_description[] = {
+	[GSN_CTR_ERR_SOCKET] = { "err:socket", "Socket error" },
+	[GSN_CTR_ERR_READFROM] = { "err:readfrom", "readfrom() errors" },
+	[GSN_CTR_ERR_SENDTO] = { "err:sendto", "sendto() errors" },
+	[GSN_CTR_ERR_QUEUEFULL] = { "err:queuefull", "Failed to queue message because queue is full" },
+	[GSN_CTR_ERR_SEQ] = { "err:seq", "Sequence number out of range" },
+	[GSN_CTR_ERR_ADDRESS] = { "err:address", "GSN address conversion failed" },
+	[GSN_CTR_ERR_UNKNOWN_PDP] = { "err:unknown_pdp", "Failed looking up PDP context" },
+	[GSN_CTR_ERR_UNEXPECTED_CAUSE] = { "err:unexpected_cause", "Unexpected cause value received" },
+	[GSN_CTR_ERR_OUT_OF_PDP] = { "err:out_of_pdp", "Out of storage for PDP contexts" },
+	[GSN_CTR_PKT_EMPTY] = { "pkt:empty", "Empty packet received" },
+	[GSN_CTR_PKT_UNSUP] = { "pkt:unsupported", "Unsupported GTP version received" },
+	[GSN_CTR_PKT_TOOSHORT] = { "pkt:too_short", "Packet too short received" },
+	[GSN_CTR_PKT_UNKNOWN] = { "pkt:unknown", "Unknown packet type received" },
+	[GSN_CTR_PKT_UNEXPECT] = { "pkt:unexpected", "Unexpected packet type received" },
+	[GSN_CTR_PKT_DUPLICATE] = { "pkt:duplicate", "Duplicate or unsolicited packet received" },
+	[GSN_CTR_PKT_MISSING] = { "pkt:missing", "Missing IE in packet received" },
+	[GSN_CTR_PKT_INCORRECT] = { "pkt:incorrect", "Incorrect IE in packet received" },
+	[GSN_CTR_PKT_INVALID] = { "pkt:invalid", "Invalid format in packet received" },
+};
+
+static const struct rate_ctr_group_desc gsn_ctrg_desc = {
+	"gsn",
+	"GSN Statistics",
+	OSMO_STATS_CLASS_PEER,
+	ARRAY_SIZE(gsn_ctr_description),
+	gsn_ctr_description,
+};
+static unsigned int gsn_ctr_next_idx = 0;
+
+/* Global timer definitions for GTP operation, provided for convenience. To make these user configurable, it is convenient to add
+ * gtp_gsn_tdefs as one of your program's osmo_tdef_group entries and call osmo_tdef_vty_init(). */
+struct osmo_tdef gtp_T_defs[] = {
+	{ .T = GTP_GSN_TIMER_T3_RESPONSE, .default_val = 5, .unit = OSMO_TDEF_S,
+	  .desc = "Timer T3-RESPONSE holds the maximum wait time for a response of a request message"
+	},
+	{ .T = GTP_GSN_TIMER_N3_REQUESTS, .default_val = 3, .unit = OSMO_TDEF_CUSTOM,
+	  .desc = "Counter N3-REQUESTS holds the maximum number of attempts made by GTP to send a request message"
+	},
+	{ .T = GTP_GSN_TIMER_T3_HOLD_RESPONSE, .default_val = 5 * 3 /* (GTP_GSN_TIMER_T3_RESPONSE * GTP_GSN_TIMER_N3_REQUESTS) */, .unit = OSMO_TDEF_S,
+	  .desc = "Time a GTP respoonse message is kept cached to re-transmit it when a duplicate request is received. Value is generally equal to (T3-RESPONSE * N3-REQUESTS) set at the peer"
+	},
+	{}
+};
+
+
+/* API Functions */
+
+/* Deprecated, use gtp_pdp_newpdp() instead */
+int gtp_newpdp(struct gsn_t *gsn, struct pdp_t **pdp,
+	       uint64_t imsi, uint8_t nsapi)
+{
+	int rc;
+	rc = gtp_pdp_newpdp(gsn, pdp, imsi, nsapi, NULL);
+	return rc;
+}
+
+int gtp_freepdp(struct gsn_t *gsn, struct pdp_t *pdp)
+{
+	if (gsn->cb_delete_context)
+		gsn->cb_delete_context(pdp);
+	return pdp_freepdp(pdp);
+}
+
+/* Free pdp and all its secondary PDP contexts. Must be called on the primary PDP context. */
+int gtp_freepdp_teardown(struct gsn_t *gsn, struct pdp_t *pdp)
+{
+	int n;
+	struct pdp_t *secondary_pdp;
+	OSMO_ASSERT(!pdp->secondary);
+
+	for (n = 0; n < PDP_MAXNSAPI; n++) {
+		if (pdp->secondary_tei[n]) {
+			if (gtp_pdp_getgtp1(gsn, &secondary_pdp,
+					     pdp->secondary_tei[n])) {
+				LOGP(DLGTP, LOGL_ERROR,
+					"Unknown secondary PDP context\n");
+				continue;
+			}
+			if (pdp != secondary_pdp) {
+				gtp_freepdp(gsn, secondary_pdp);
+			}
+		}
+	}
+
+	return gtp_freepdp(gsn, pdp);
+}
+
+/* gtp_gpdu */
+
+extern int gtp_fd(struct gsn_t *gsn)
+{
+	return gsn->fd0;
+}
+
+int gtp_set_cb_unsup_ind(struct gsn_t *gsn,
+			 int (*cb) (struct sockaddr_in * peer))
+{
+	gsn->cb_unsup_ind = cb;
+	return 0;
+}
+
+int gtp_set_cb_extheader_ind(struct gsn_t *gsn,
+			     int (*cb) (struct sockaddr_in * peer))
+{
+	gsn->cb_extheader_ind = cb;
+	return 0;
+}
+
+int gtp_set_cb_ran_info_relay_ind(struct gsn_t *gsn,
+			     int (*cb) (struct sockaddr_in * peer, union gtpie_member **ie))
+{
+	gsn->cb_ran_info_relay_ind = cb;
+	return 0;
+}
+
+/* API: Initialise delete context callback */
+/* Called whenever a pdp context is deleted for any reason */
+int gtp_set_cb_delete_context(struct gsn_t *gsn, int (*cb) (struct pdp_t * pdp))
+{
+	gsn->cb_delete_context = cb;
+	return 0;
+}
+
+int gtp_set_cb_conf(struct gsn_t *gsn,
+		    int (*cb) (int type, int cause,
+			       struct pdp_t * pdp, void *cbp))
+{
+	gsn->cb_conf = cb;
+	return 0;
+}
+
+int gtp_set_cb_recovery(struct gsn_t *gsn,
+			int (*cb) (struct sockaddr_in * peer, uint8_t recovery))
+{
+	gsn->cb_recovery = cb;
+	return 0;
+}
+
+/* cb_recovery()
+ * pdp may be NULL if Recovery IE was received from a message independent
+ * of any PDP ctx (such as Echo Response), or because pdp ctx is unknown to the
+ * local setup. In case pdp is known, caller may want to keep that pdp alive to
+ * handle subsequent msg cb as this specific pdp ctx is still valid according to
+ * specs.
+ */
+int gtp_set_cb_recovery2(struct gsn_t *gsn,
+			int (*cb_recovery2) (struct sockaddr_in * peer, struct pdp_t * pdp, uint8_t recovery))
+{
+	gsn->cb_recovery2 = cb_recovery2;
+	return 0;
+}
+
+/* cb_recovery()
+ * pdp may be NULL if Recovery IE was received from a message independent
+ * of any PDP ctx (such as Echo Response), or because pdp ctx is unknown to the
+ * local setup. In case pdp is known, caller may want to keep that pdp alive to
+ * handle subsequent msg cb as this specific pdp ctx is still valid according to
+ * specs.
+ */
+int gtp_set_cb_recovery3(struct gsn_t *gsn,
+			 int (*cb_recovery3) (struct gsn_t *gsn, struct sockaddr_in *peer,
+					      struct pdp_t *pdp, uint8_t recovery))
+{
+	gsn->cb_recovery3 = cb_recovery3;
+	return 0;
+}
+
+int gtp_set_cb_data_ind(struct gsn_t *gsn,
+			       int (*cb_data_ind) (struct pdp_t * pdp,
+						   void *pack, unsigned len))
+{
+	gsn->cb_data_ind = cb_data_ind;
+	return 0;
+}
+int gtp_set_cb_sgsn_context_request_ind(struct gsn_t *gsn,
+			     int (*cb) (struct gsn_t *gsn, struct sockaddr_in *peer, uint16_t seq, const struct osmo_routing_area_id *rai, uint32_t teic, struct osmo_mobile_identity *mi, union gtpie_member **ie))
+{
+	gsn->cb_sgsn_context_request_ind = cb;
+	return 0;
+}
+
+static int queue_timer_retrans(struct gsn_t *gsn)
+{
+	/* Retransmit any outstanding packets */
+	/* Remove from queue if maxretrans exceeded */
+	time_t now;
+	struct qmsg_t *qmsg;
+	unsigned int t3_response, n3_requests;
+
+	now = time(NULL);
+	t3_response = osmo_tdef_get(gsn->tdef, GTP_GSN_TIMER_T3_RESPONSE, OSMO_TDEF_S, -1);
+	n3_requests = osmo_tdef_get(gsn->tdef, GTP_GSN_TIMER_N3_REQUESTS, OSMO_TDEF_CUSTOM, -1);
+
+	/* get first element in queue, as long as the timeout of that
+	 * element has expired */
+	while ((!queue_getfirst(gsn->queue_req, &qmsg)) &&
+	       (qmsg->timeout <= now)) {
+		if (qmsg->retrans > n3_requests) {	/* Too many retrans */
+			LOGP(DLGTP, LOGL_NOTICE, "Retransmit req queue timeout of seq %" PRIu16 "\n",
+			     qmsg->seq);
+			if (gsn->cb_conf)
+				gsn->cb_conf(qmsg->type, EOF, NULL, qmsg->cbp);
+			queue_freemsg(gsn->queue_req, qmsg);
+		} else {
+			LOGP(DLGTP, LOGL_INFO, "Retransmit (%d) of seq %" PRIu16 "\n",
+			     qmsg->retrans, qmsg->seq);
+			if (sendto(qmsg->fd, &qmsg->p, qmsg->l, 0,
+				   (struct sockaddr *)&qmsg->peer,
+				   sizeof(struct sockaddr_in)) < 0) {
+				rate_ctr_inc2(gsn->ctrg, GSN_CTR_ERR_SENDTO);
+				LOGP(DLGTP, LOGL_ERROR,
+					"Sendto(fd0=%d, msg=%lx, len=%d) failed: Error = %s\n",
+					gsn->fd0, (unsigned long)&qmsg->p,
+					qmsg->l, strerror(errno));
+			}
+			queue_back(gsn->queue_req, qmsg);
+			qmsg->timeout = now + t3_response;
+			qmsg->retrans++;
+		}
+	}
+
+	/* Also clean up reply timeouts */
+	while ((!queue_getfirst(gsn->queue_resp, &qmsg)) &&
+	       (qmsg->timeout < now)) {
+		LOGP(DLGTP, LOGL_DEBUG, "Retransmit resp queue seq %"
+		     PRIu16 " expired, removing from queue\n", qmsg->seq);
+		queue_freemsg(gsn->queue_resp, qmsg);
+	}
+
+	return 0;
+}
+
+static int queue_timer_retranstimeout(struct gsn_t *gsn, struct timeval *timeout)
+{
+	time_t now, later, diff;
+	struct qmsg_t *qmsg;
+	timeout->tv_usec = 0;
+
+	if (queue_getfirst(gsn->queue_req, &qmsg)) {
+		timeout->tv_sec = 10;
+	} else {
+		now = time(NULL);
+		later = qmsg->timeout;
+		timeout->tv_sec = later - now;
+		if (timeout->tv_sec < 0)
+			timeout->tv_sec = 0;	/* No negative allowed */
+		if (timeout->tv_sec > 10)
+			timeout->tv_sec = 10;	/* Max sleep for 10 sec */
+	}
+
+	if (queue_getfirst(gsn->queue_resp, &qmsg)) {
+		/* already set by queue_req, do nothing */
+	} else { /* trigger faster if earlier timeout exists in queue_resp */
+		now = time(NULL);
+		later = qmsg->timeout;
+		diff = later - now;
+		if (diff < 0)
+			diff = 0;
+		if (diff < timeout->tv_sec)
+			timeout->tv_sec = diff;
+	}
+
+	return 0;
+}
+
+void gtp_queue_timer_start(struct gsn_t *gsn)
+{
+	struct timeval next;
+
+	/* Retrieve next retransmission as timeval */
+	queue_timer_retranstimeout(gsn, &next);
+
+	/* re-schedule the timer */
+	osmo_timer_schedule(&gsn->queue_timer, next.tv_sec, next.tv_usec/1000);
+}
+
+/* timer callback for libgtp retransmission and ping */
+static void queue_timer_cb(void *data)
+{
+	struct gsn_t *gsn = data;
+
+	/* do all the retransmissions as needed */
+	queue_timer_retrans(gsn);
+
+	gtp_queue_timer_start(gsn);
+}
+
+
+/**
+ * @brief clear the request and response queue. Useful for debugging to reset "some" state.
+ * @param gsn The GGSN instance
+ */
+void gtp_clear_queues(struct gsn_t *gsn)
+{
+	struct qmsg_t *qmsg;
+
+	LOGP(DLGTP, LOGL_INFO, "Clearing req & resp retransmit queues\n");
+	while (!queue_getfirst(gsn->queue_req, &qmsg)) {
+		queue_freemsg(gsn->queue_req, qmsg);
+	}
+
+	while (!queue_getfirst(gsn->queue_resp, &qmsg)) {
+		queue_freemsg(gsn->queue_resp, qmsg);
+	}
+}
+
+/* Perform restoration and recovery error handling as described in 29.060 */
+static void log_restart(struct gsn_t *gsn)
+{
+	FILE *f;
+	int i, rc;
+	int counter = 0;
+	char *filename;
+
+	filename = talloc_asprintf(NULL, "%s/%s", gsn->statedir, RESTART_FILE);
+	OSMO_ASSERT(filename);
+
+	/* We try to open file. On failure we will later try to create file */
+	if (!(f = fopen(filename, "r"))) {
+		LOGP(DLGTP, LOGL_NOTICE,
+			"State information file (%s) not found. Creating new file.\n",
+			filename);
+	} else {
+		rc = fscanf(f, "%d", &counter);
+		if (rc != 1) {
+			LOGP(DLGTP, LOGL_ERROR,
+				"fscanf failed to read counter value\n");
+			goto close_file;
+		}
+		if (fclose(f)) {
+			LOGP(DLGTP, LOGL_ERROR,
+				"fclose failed: Error = %s\n", strerror(errno));
+		}
+	}
+
+	gsn->restart_counter = (unsigned char)counter;
+	gsn->restart_counter++;
+
+	/* Keep the umask closely wrapped around our fopen() call in case the
+	 * log outputs cause file creation. */
+	i = umask(022);
+	f = fopen(filename, "w");
+	umask(i);
+	if (!f) {
+		LOGP(DLGTP, LOGL_ERROR,
+			"fopen(path=%s, mode=%s) failed: Error = %s\n", filename,
+			"w", strerror(errno));
+		goto free_filename;
+	}
+
+	fprintf(f, "%d\n", gsn->restart_counter);
+close_file:
+	if (fclose(f))
+		LOGP(DLGTP, LOGL_ERROR,
+			"fclose failed: Error = %s\n", strerror(errno));
+free_filename:
+	talloc_free(filename);
+}
+
+static int create_and_bind_socket(const char *name, struct gsn_t *gsn, int *fd, int domain,
+				  const struct in_addr *listen, int port)
+{
+	struct sockaddr_in addr;
+	int type = SOCK_DGRAM;
+	int protocol = 0;
+
+	*fd = socket(domain, type, protocol);
+
+	if (*fd < 0) {
+		rate_ctr_inc2(gsn->ctrg, GSN_CTR_ERR_SOCKET);
+		LOGP(DLGTP, LOGL_ERROR,
+		     "%s socket(domain=%d, type=%d, protocol=%d) failed: Error = %s\n",
+		     name, domain, type, protocol, strerror(errno));
+		return -errno;
+	}
+
+	memset(&addr, 0, sizeof(addr));
+	addr.sin_family = domain;
+	addr.sin_addr = *listen;
+	addr.sin_port = htons(port);
+#if defined(__FreeBSD__) || defined(__APPLE__)
+	addr.sin_len = sizeof(addr);
+#endif
+
+	if (bind(*fd, (struct sockaddr *)&addr, sizeof(addr)) < 0) {
+		rate_ctr_inc2(gsn->ctrg, GSN_CTR_ERR_SOCKET);
+		LOGP_WITH_ADDR(DLGTP, LOGL_ERROR, addr,
+			       "%s bind(fd=%d) failed: Error = %s\n",
+			       name, *fd, strerror(errno));
+		return -errno;
+	}
+
+	return 0;
+}
+
+int gtp_new(struct gsn_t **gsn, char *statedir, struct in_addr *listen,
+	    int mode)
+{
+	LOGP(DLGTP, LOGL_NOTICE, "GTP: gtp_newgsn() started at %s\n", inet_ntoa(*listen));
+
+	*gsn = calloc(sizeof(struct gsn_t), 1);	/* TODO */
+
+	(*gsn)->statedir = statedir;
+	log_restart(*gsn);
+
+	/* Initialise sequence number */
+	(*gsn)->seq_next = (*gsn)->restart_counter * 1024;
+
+	/* Initialize timers: */
+	(*gsn)->tdef = gtp_T_defs;
+	/* Small hack to properly reset tdef for old clients not using the tdef_group: */
+	OSMO_ASSERT(gtp_T_defs[0].default_val != 0);
+	if (gtp_T_defs[0].val == 0)
+		osmo_tdefs_reset((*gsn)->tdef);
+
+
+	/* Initialise request retransmit queue */
+	queue_new(&(*gsn)->queue_req);
+	queue_new(&(*gsn)->queue_resp);
+
+	/* Initialise pdp table */
+	pdp_init(*gsn);
+
+	/* Initialize internal queue timer */
+	osmo_timer_setup(&(*gsn)->queue_timer, queue_timer_cb, *gsn);
+
+	/* Initialize counter group: */
+	(*gsn)->ctrg = rate_ctr_group_alloc(NULL, &gsn_ctrg_desc, gsn_ctr_next_idx++);
+
+	/* Initialise call back functions */
+	(*gsn)->cb_create_context_ind = 0;
+	(*gsn)->cb_delete_context = 0;
+	(*gsn)->cb_unsup_ind = 0;
+	(*gsn)->cb_conf = 0;
+	(*gsn)->cb_data_ind = 0;
+
+	/* Store function parameters */
+	/* Same IP for user traffic and signalling */
+	(*gsn)->gsnc = *listen;
+	(*gsn)->gsnu = *listen;
+	(*gsn)->mode = mode;
+
+	(*gsn)->fd0 = -1;
+	(*gsn)->fd1c = -1;
+	(*gsn)->fd1u = -1;
+
+	/* Create GTP version 0 socket */
+	if (create_and_bind_socket("GTPv0", *gsn, &(*gsn)->fd0, AF_INET, listen, GTP0_PORT) < 0)
+		goto error;
+
+	/* Create GTP version 1 control plane socket */
+	if (create_and_bind_socket("GTPv1 control plane", *gsn, &(*gsn)->fd1c, AF_INET, listen, GTP1C_PORT) < 0)
+		goto error;
+
+	/* Create GTP version 1 user plane socket */
+	if (create_and_bind_socket("GTPv1 user plane", *gsn, &(*gsn)->fd1u, AF_INET, listen, GTP1U_PORT) < 0)
+		goto error;
+
+	/* Start internal queue timer */
+	gtp_queue_timer_start(*gsn);
+
+	return 0;
+error:
+	gtp_free(*gsn);
+	*gsn = NULL;
+	return -1;
+}
+
+int gtp_free(struct gsn_t *gsn)
+{
+
+	/* Cleanup internal queue timer */
+	osmo_timer_del(&gsn->queue_timer);
+
+	/* Clean up retransmit queues */
+	queue_free(gsn->queue_req);
+	queue_free(gsn->queue_resp);
+
+	close(gsn->fd0);
+	close(gsn->fd1c);
+	close(gsn->fd1u);
+
+	rate_ctr_group_free(gsn->ctrg);
+
+	free(gsn);
+	return 0;
+}
+
+/* API: Register create context indication callback */
+int gtp_set_cb_create_context_ind(struct gsn_t *gsn,
+				  int (*cb_create_context_ind) (struct pdp_t *
+								pdp))
+{
+	gsn->cb_create_context_ind = cb_create_context_ind;
+	return 0;
+}
+
+int gtp_retrans(struct gsn_t *gsn)
+{
+	/* dummy API, deprecated. */
+	return 0;
+}
+
+int gtp_retranstimeout(struct gsn_t *gsn, struct timeval *timeout)
+{
+	timeout->tv_sec = 24*60*60;
+	timeout->tv_usec = 0;
+	/* dummy API, deprecated. Return a huge timer to do nothing */
+	return 0;
+}

gtp/gsn.h

@@ -0,0 +1,189 @@
+/*
+ *  OsmoGGSN - Gateway GPRS Support Node
+ *  Copyright (C) 2002, 2003, 2004 Mondru AB.
+ *
+ *  The contents of this file may be used under the terms of the GNU
+ *  General Public License Version 2, provided that the above copyright
+ *  notice and this permission notice is included in all copies or
+ *  substantial portions of the software.
+ *
+ */
+
+#ifndef _GSN_H
+#define _GSN_H
+
+#include <osmocom/core/utils.h>
+#include <osmocom/core/defs.h>
+#include <osmocom/core/timer.h>
+#include <osmocom/core/tdef.h>
+#include <osmocom/core/rate_ctr.h>
+#include <osmocom/gsm/gsm23003.h>
+#include <osmocom/gsm/gsm48.h>
+
+#include "pdp.h"
+
+#define GTP_MODE_GGSN 1
+#define GTP_MODE_SGSN 2
+
+#define RESTART_FILE "gsn_restart"
+
+extern struct osmo_tdef gtp_T_defs[];
+
+/* ***********************************************************
+ * Information storage for each gsn instance
+ *
+ * Normally each instance of the application corresponds to
+ * one instance of a gsn.
+ *
+ * In order to avoid global variables in the application, and
+ * also in order to allow several instances of a gsn in the same
+ * application this struct is provided in order to store all
+ * relevant information related to the gsn.
+ *
+ * Note that this does not include information storage for '
+ * each pdp context. This is stored in another struct.
+ *************************************************************/
+
+enum gsn_rate_ctr_keys {
+	GSN_CTR_ERR_SOCKET,
+	GSN_CTR_ERR_READFROM,	/* Number of readfrom errors */
+	GSN_CTR_ERR_SENDTO,	/* Number of sendto errors */
+	GSN_CTR_ERR_QUEUEFULL,	/* Number of times queue was full */
+	GSN_CTR_ERR_SEQ,	/* Number of seq out of range */
+	GSN_CTR_ERR_ADDRESS,	/* GSN address conversion failed */
+	GSN_CTR_ERR_UNKNOWN_PDP,	/* GSN address conversion failed */
+	GSN_CTR_ERR_UNEXPECTED_CAUSE,	/* Unexpected cause value received */
+	GSN_CTR_ERR_OUT_OF_PDP,	/* Out of storage for PDP contexts */
+	GSN_CTR_PKT_EMPTY,		/* Number of empty packets */
+	GSN_CTR_PKT_UNSUP,		/* Number of unsupported version 29.60 11.1.1 */
+	GSN_CTR_PKT_TOOSHORT,	/* Number of too short headers 29.60 11.1.2 */
+	GSN_CTR_PKT_UNKNOWN,	/* Number of unknown messages 29.60 11.1.3 */
+	GSN_CTR_PKT_UNEXPECT,	/* Number of unexpected messages 29.60 11.1.4 */
+	GSN_CTR_PKT_DUPLICATE,	/* Number of duplicate or unsolicited replies */
+	GSN_CTR_PKT_MISSING,	/* Number of missing information field messages */
+	GSN_CTR_PKT_INCORRECT,	/* Number of incorrect information field messages */
+	GSN_CTR_PKT_INVALID,	/* Number of invalid message format messages */
+};
+
+/* 3GPP TS 29.006 14.1, 14,2 */
+enum gtp_gsn_timers {
+	GTP_GSN_TIMER_T3_RESPONSE = 3,
+	GTP_GSN_TIMER_N3_REQUESTS = 1003,
+	GTP_GSN_TIMER_T3_HOLD_RESPONSE = -3,
+};
+
+struct gsn_t {
+	/* Parameters related to the network interface */
+
+	int fd0;		/* GTP0 file descriptor */
+	int fd1c;		/* GTP1 control plane file descriptor */
+	int fd1u;		/* GTP0 user plane file descriptor */
+	int mode;		/* Mode of operation: GGSN or SGSN */
+	struct in_addr gsnc;	/* IP address of this gsn for signalling */
+	struct in_addr gsnu;	/* IP address of this gsn for user traffic */
+
+	/* Parameters related to signalling messages */
+	uint16_t seq_next;	/* Next sequence number to use */
+	int seq_first;		/* First packet in queue (oldest timeout) */
+	int seq_last;		/* Last packet in queue (youngest timeout) */
+
+	unsigned char restart_counter;	/* Increment on restart. Stored on disk */
+	char *statedir;		/* Disk location for permanent storage */
+	void *priv;		/* used by libgtp users to attach their own state) */
+	struct queue_t *queue_req;	/* Request queue */
+	struct queue_t *queue_resp;	/* Response queue */
+
+	struct pdp_t pdpa[PDP_MAX];	/* PDP storage */
+	struct pdp_t *hashtid[PDP_MAX];	/* Hash table for IMSI + NSAPI */
+
+	struct osmo_timer_list queue_timer; /* internal queue_{req,resp} timer */
+
+	/* Call back functions */
+	int (*cb_delete_context) (struct pdp_t *);
+	int (*cb_create_context_ind) (struct pdp_t *);
+	int (*cb_unsup_ind) (struct sockaddr_in * peer);
+	int (*cb_extheader_ind) (struct sockaddr_in * peer);
+	int (*cb_ran_info_relay_ind) (struct sockaddr_in *peer, union gtpie_member **ie);
+	int (*cb_conf) (int type, int cause, struct pdp_t * pdp, void *cbp);
+	int (*cb_data_ind) (struct pdp_t * pdp, void *pack, unsigned len);
+	int (*cb_recovery) (struct sockaddr_in * peer, uint8_t recovery);
+	int (*cb_recovery2) (struct sockaddr_in * peer, struct pdp_t * pdp, uint8_t recovery);
+	int (*cb_recovery3) (struct gsn_t *gsn, struct sockaddr_in *peer, struct pdp_t *pdp, uint8_t recovery);
+	int (*cb_sgsn_context_request_ind) (struct gsn_t *gsn, struct sockaddr_in *peer, uint16_t seq, const struct osmo_routing_area_id *rai, uint32_t teic, struct osmo_mobile_identity *mi, union gtpie_member **ie); /* Pass RAI and TLLI/TMSI/IMSI directly */
+
+	/* Counters */
+	struct rate_ctr_group *ctrg;
+
+	/* Timers: */
+	struct osmo_tdef *tdef;
+};
+
+/* External API functions */
+
+extern int gtp_new(struct gsn_t **gsn, char *statedir, struct in_addr *listen,
+		   int mode);
+
+extern int gtp_free(struct gsn_t *gsn);
+
+extern int gtp_newpdp(struct gsn_t *gsn, struct pdp_t **pdp,
+		      uint64_t imsi, uint8_t nsapi) OSMO_DEPRECATED("Use gtp_pdp_newpdp() instead");
+extern int gtp_freepdp(struct gsn_t *gsn, struct pdp_t *pdp);
+extern int gtp_freepdp_teardown(struct gsn_t *gsn, struct pdp_t *pdp);
+
+extern int gtp_create_context_req(struct gsn_t *gsn, struct pdp_t *pdp,
+				  void *cbp);
+
+extern int gtp_create_sgsn_context_req(struct gsn_t *gsn, struct pdp_t *pdp, void *cbp);
+
+extern int gtp_set_cb_create_context_ind(struct gsn_t *gsn,
+					 int (*cb_create_context_ind) (struct
+								       pdp_t *
+								       pdp));
+extern int gtp_set_cb_data_ind(struct gsn_t *gsn,
+			       int (*cb_data_ind) (struct pdp_t * pdp,
+						   void *pack, unsigned len));
+extern int gtp_set_cb_delete_context(struct gsn_t *gsn,
+				     int (*cb_delete_context) (struct pdp_t *
+							       pdp));
+/*extern int gtp_set_cb_create_context(struct gsn_t *gsn,
+  int (*cb_create_context) (struct pdp_t* pdp)); */
+
+extern int gtp_set_cb_unsup_ind(struct gsn_t *gsn,
+				int (*cb) (struct sockaddr_in * peer));
+
+extern int gtp_set_cb_extheader_ind(struct gsn_t *gsn,
+				    int (*cb) (struct sockaddr_in * peer));
+
+extern int gtp_set_cb_ran_info_relay_ind(struct gsn_t *gsn,
+				    int (*cb) (struct sockaddr_in * peer, union gtpie_member **ie));
+
+extern int gtp_set_cb_sgsn_context_request_ind(struct gsn_t *gsn,
+			     int (*cb) (struct gsn_t *gsn, struct sockaddr_in *peer, uint16_t seq, const struct osmo_routing_area_id *rai, uint32_t teic, struct osmo_mobile_identity *mi, union gtpie_member **ie));
+
+extern int gtp_set_cb_conf(struct gsn_t *gsn,
+			   int (*cb) (int type, int cause, struct pdp_t * pdp,
+				      void *cbp));
+
+int gtp_set_cb_recovery(struct gsn_t *gsn,
+			int (*cb) (struct sockaddr_in * peer,
+				   uint8_t recovery))
+	OSMO_DEPRECATED("Use gtp_set_cb_recovery2() instead, to obtain pdp ctx originating the recovery");
+int gtp_set_cb_recovery2(struct gsn_t *gsn,
+			int (*cb) (struct sockaddr_in * peer,
+				   struct pdp_t * pdp,
+				   uint8_t recovery))
+	OSMO_DEPRECATED("Use gtp_set_cb_recovery3() instead, to obtain gsn handling the recovery");
+int gtp_set_cb_recovery3(struct gsn_t *gsn,
+			int (*cb) (struct gsn_t * gsn, struct sockaddr_in * peer,
+				   struct pdp_t * pdp,
+				   uint8_t recovery));
+void gtp_clear_queues(struct gsn_t *gsn);
+extern int gtp_fd(struct gsn_t *gsn);
+
+extern int gtp_retrans(struct gsn_t *gsn) OSMO_DEPRECATED("This API is a no-op, libgtp already does the job internally");
+extern int gtp_retranstimeout(struct gsn_t *gsn, struct timeval *timeout) OSMO_DEPRECATED("This API is a no-op and will return a 1 day timeout");
+
+/* Internal APIs: */
+void gtp_queue_timer_start(struct gsn_t *gsn);
+
+#endif /* !_GSN_H */

gtp/gtp.h

@@ -1,19 +1,22 @@
-/* 
+/*
  *  OsmoGGSN - Gateway GPRS Support Node
  *  Copyright (C) 2002, 2003, 2004 Mondru AB.
- * 
+ *
  *  The contents of this file may be used under the terms of the GNU
  *  General Public License Version 2, provided that the above copyright
  *  notice and this permission notice is included in all copies or
  *  substantial portions of the software.
- * 
+ *
  */
 
 #ifndef _GTP_H
 #define _GTP_H
 
-#define GTP_MODE_GGSN 1
-#define GTP_MODE_SGSN 2
+#include <osmocom/core/utils.h>
+
+#include "gtpie.h"
+#include "pdp.h"
+#include "gsn.h"
 
 #define GTP0_PORT	3386
 #define GTP1C_PORT	2123
@@ -25,12 +28,10 @@
 #define GTP1_HEADER_SIZE_SHORT  8
 #define GTP1_HEADER_SIZE_LONG  12
 
+#define NAMESIZE 1024
 #define SYSLOG_PRINTSIZE 255
 #define ERRMSG_SIZE 255
 
-#define RESTART_FILE "gsn_restart"
-#define NAMESIZE 1024
-
 /* GTP version 1 extension header type definitions. */
 #define GTP_EXT_PDCP_PDU    0xC0	/* PDCP PDU Number */
 
@@ -52,7 +53,7 @@
 #define GTP_UPDATE_PDP_RSP   19	/* Update PDP Context Response */
 #define GTP_DELETE_PDP_REQ   20	/* Delete PDP Context Request */
 #define GTP_DELETE_PDP_RSP   21	/* Delete PDP Context Response */
-						       /* 22-25 For future use. *//* In version GTP 1 anonomous PDP context */
+/* 22-25 For future use. *//* In version GTP 1 anonomous PDP context */
 #define GTP_ERROR            26	/* Error Indication */
 #define GTP_PDU_NOT_REQ      27	/* PDU Notification Request */
 #define GTP_PDU_NOT_RSP      28	/* PDU Notification Response */
@@ -79,12 +80,17 @@
 #define GTP_FWD_SRNS         58	/* Forward SRNS Context */
 #define GTP_FWD_RELOC_ACK    59	/* Forward Relocation Complete Acknowledge */
 #define GTP_FWD_SRNS_ACK     60	/* Forward SRNS Context Acknowledge */
+#define GTP_RAN_INFO_RELAY   70	/* RAN Information Relay */
 /* 61-239 For future use. */
 #define GTP_DATA_TRAN_REQ   240	/* Data Record Transfer Request */
 #define GTP_DATA_TRAN_RSP   241	/* Data Record Transfer Response */
 /* 242-254 For future use. */
 #define GTP_GPDU            255	/* G-PDU */
 
+extern const struct value_string gtp_type_names[];
+static inline const char *gtp_type_name(uint8_t val)
+{ return get_value_string(gtp_type_names, val); }
+
 /* GTP information element cause codes from 29.060 v3.9.0 7.7 */
 /*                                                            */
 #define GTPCAUSE_REQ_IMSI                   0	/* Request IMSI */
@@ -93,19 +99,21 @@
 #define GTPCAUSE_NO_ID_NEEDED               3	/* No identity needed */
 #define GTPCAUSE_MS_REFUSES_X               4	/* MS refuses */
 #define GTPCAUSE_MS_NOT_RESP_X              5	/* MS is not GPRS responding */
-#define GTPCAUSE_006                        6	/* For future use 6-48 */
-#define GTPCAUSE_049                       49	/* Cause values reserved for GPRS charging protocol use (See GTP' in GSM 12.15) 49-63 */
-#define GTPCAUSE_064                       64	/* For future use 64-127 */
+/* 6-48 For future use */
+/* 49-63 Cause values reserved for GPRS charging protocol use (See GTP' in GSM 12.15) */
+/* 64-127 For future use */
 #define GTPCAUSE_ACC_REQ                  128	/* Request accepted */
-#define GTPCAUSE_129                      129	/* For future use 129-176 */
-#define GTPCAUSE_177                      177	/* Cause values reserved for GPRS charging protocol use (See GTP' In GSM 12.15) 177-191 */
+#define GTPCAUSE_NEW_PDP_NET_PREF         129	/* New PDP type due to network preference */
+#define GTPCAUSE_NEW_PDP_ADDR_BEAR        130	/* New PDP type due to single address bearer only */
+/* 131-176 For future use */
+/* 177-191 Cause values reserved for GPRS charging protocol use (See GTP' In GSM 12.15) */
 #define GTPCAUSE_NON_EXIST                192	/* Non-existent */
 #define GTPCAUSE_INVALID_MESSAGE          193	/* Invalid message format */
 #define GTPCAUSE_IMSI_NOT_KNOWN           194	/* IMSI not known */
 #define GTPCAUSE_MS_DETACHED              195	/* MS is GPRS detached */
 #define GTPCAUSE_MS_NOT_RESP              196	/* MS is not GPRS responding */
 #define GTPCAUSE_MS_REFUSES               197	/* MS refuses */
-#define GTPCAUSE_198                      198	/* For future use */
+/* 198 For future use */
 #define GTPCAUSE_NO_RESOURCES             199	/* No resources available */
 #define GTPCAUSE_NOT_SUPPORTED            200	/* Service not supported */
 #define GTPCAUSE_MAN_IE_INCORRECT         201	/* Mandatory IE incorrect */
@@ -128,14 +136,21 @@
 #define GTPCAUSE_SYN_ERR_FILTER           218	/* Syntactic errors in packet filter(s) */
 #define GTPCAUSE_MISSING_APN              219	/* Missing or unknown APN */
 #define GTPCAUSE_UNKNOWN_PDP              220	/* Unknown PDP address or PDP type */
-#define GTPCAUSE_221                      221	/* For Future Use 221-240 */
-#define GTPCAUSE_241                      241	/* Cause Values Reserved For Gprs Charging Protocol Use (See Gtp' In Gsm 12.15) 241-255 */
+/* 221-240 For future use */
+/* 241-255 Cause Values Reserved For Gprs Charging Protocol Use (See Gtp' In Gsm 12.15) */
+
+static inline bool gtp_cause_successful(uint8_t cause)
+{
+	return cause == GTPCAUSE_ACC_REQ ||
+		cause == GTPCAUSE_NEW_PDP_NET_PREF ||
+		cause == GTPCAUSE_NEW_PDP_ADDR_BEAR;
+}
 
 struct ul66_t;
 struct ul16_t;
 struct pdp_t;
 
-/* GTP 0 header. 
+/* GTP 0 header.
  * Explanation to some of the fields:
  * SNDCP NPDU Number flag = 0 except for inter SGSN handover situations
  * SNDCP N-PDU LCC Number 0 = 0xff except for inter SGSN handover situations
@@ -220,96 +235,13 @@ union gtp_packet {
 	struct gtp1_packet_long gtp1l;
 } __attribute__ ((packed));
 
-/* ***********************************************************
- * Information storage for each gsn instance
- *
- * Normally each instance of the application corresponds to
- * one instance of a gsn. 
- * 
- * In order to avoid global variables in the application, and
- * also in order to allow several instances of a gsn in the same
- * application this struct is provided in order to store all
- * relevant information related to the gsn.
- * 
- * Note that this does not include information storage for '
- * each pdp context. This is stored in another struct.
- *************************************************************/
-
-struct gsn_t {
-	/* Parameters related to the network interface */
-
-	int fd0;		/* GTP0 file descriptor */
-	int fd1c;		/* GTP1 control plane file descriptor */
-	int fd1u;		/* GTP0 user plane file descriptor */
-	int mode;		/* Mode of operation: GGSN or SGSN */
-	struct in_addr gsnc;	/* IP address of this gsn for signalling */
-	struct in_addr gsnu;	/* IP address of this gsn for user traffic */
-
-	/* Parameters related to signalling messages */
-	uint16_t seq_next;	/* Next sequence number to use */
-	int seq_first;		/* First packet in queue (oldest timeout) */
-	int seq_last;		/* Last packet in queue (youngest timeout) */
-
-	unsigned char restart_counter;	/* Increment on restart. Stored on disk */
-	char *statedir;		/* Disk location for permanent storage */
-	void *priv;		/* used by libgtp users to attach their own state) */
-	struct queue_t *queue_req;	/* Request queue */
-	struct queue_t *queue_resp;	/* Response queue */
-
-	/* Call back functions */
-	int (*cb_delete_context) (struct pdp_t *);
-	int (*cb_create_context_ind) (struct pdp_t *);
-	int (*cb_unsup_ind) (struct sockaddr_in * peer);
-	int (*cb_extheader_ind) (struct sockaddr_in * peer);
-	int (*cb_conf) (int type, int cause, struct pdp_t * pdp, void *cbp);
-	int (*cb_data_ind) (struct pdp_t * pdp, void *pack, unsigned len);
-	int (*cb_recovery) (struct sockaddr_in * peer, uint8_t recovery);
-
-	/* Counters */
-
-	uint64_t err_socket;	/* Number of socket errors */
-	uint64_t err_readfrom;	/* Number of readfrom errors */
-	uint64_t err_sendto;	/* Number of sendto errors */
-	uint64_t err_memcpy;	/* Number of memcpy */
-	uint64_t err_queuefull;	/* Number of times queue was full */
-	uint64_t err_seq;	/* Number of seq out of range */
-	uint64_t err_address;	/* GSN address conversion failed */
-	uint64_t err_unknownpdp;	/* GSN address conversion failed */
-	uint64_t err_unknowntid;	/* Application supplied unknown imsi+nsapi */
-	uint64_t err_cause;	/* Unexpected cause value received */
-	uint64_t err_outofpdp;	/* Out of storage for PDP contexts */
-
-	uint64_t empty;		/* Number of empty packets */
-	uint64_t unsup;		/* Number of unsupported version 29.60 11.1.1 */
-	uint64_t tooshort;	/* Number of too short headers 29.60 11.1.2 */
-	uint64_t unknown;	/* Number of unknown messages 29.60 11.1.3 */
-	uint64_t unexpect;	/* Number of unexpected messages 29.60 11.1.4 */
-	uint64_t duplicate;	/* Number of duplicate or unsolicited replies */
-	uint64_t missing;	/* Number of missing information field messages */
-	uint64_t incorrect;	/* Number of incorrect information field messages */
-	uint64_t invalid;	/* Number of invalid message format messages */
-};
-
 /* External API functions */
 
 extern const char *gtp_version();
-extern int gtp_new(struct gsn_t **gsn, char *statedir, struct in_addr *listen,
-		   int mode);
-
-extern int gtp_free(struct gsn_t *gsn);
-
-extern int gtp_newpdp(struct gsn_t *gsn, struct pdp_t **pdp,
-		      uint64_t imsi, uint8_t nsapi);
-extern int gtp_freepdp(struct gsn_t *gsn, struct pdp_t *pdp);
 
 extern int gtp_create_context_req(struct gsn_t *gsn, struct pdp_t *pdp,
 				  void *cbp);
 
-extern int gtp_set_cb_create_context_ind(struct gsn_t *gsn,
-					 int (*cb_create_context_ind) (struct
-								       pdp_t *
-								       pdp));
-
 extern int gtp_create_context_resp(struct gsn_t *gsn, struct pdp_t *pdp,
 				   int cause);
 
@@ -317,43 +249,29 @@ extern int gtp_update_context(struct gsn_t *gsn, struct pdp_t *pdp,
 			      void *cbp, struct in_addr *inetaddr);
 
 extern int gtp_delete_context_req(struct gsn_t *gsn, struct pdp_t *pdp,
-				  void *cbp, int teardown);
+				  void *cbp, int teardown)
+		OSMO_DEPRECATED("Use gtp_delete_context_req2() instead, to avoid freeing pdp ctx before reply");
+extern int gtp_delete_context_req2(struct gsn_t *gsn, struct pdp_t *pdp,
+				   void *cbp, int teardown);
 
 extern int gtp_data_req(struct gsn_t *gsn, struct pdp_t *pdp,
 			void *pack, unsigned len);
 
-extern int gtp_set_cb_data_ind(struct gsn_t *gsn,
-			       int (*cb_data_ind) (struct pdp_t * pdp,
-						   void *pack, unsigned len));
+extern int gtp_ran_info_relay_req(struct gsn_t *gsn, const struct sockaddr_in *peer,
+				  const uint8_t *ran_container, size_t ran_container_len,
+				  const uint8_t *rim_route_addr, size_t rim_route_addr_len,
+				  uint8_t rim_route_addr_discr);
+
+extern int gtp_sgsn_context_req(struct gsn_t *gsn, const struct in_addr *peer,
+			 const struct osmo_mobile_identity *mi, uint16_t tlli, uint32_t teic,
+			 const struct ul16_t *sgsn_addr, const struct ul255_t *rai, void *cbp);
+extern int gtp_sgsn_context_conf(struct gsn_t *gsn, struct sockaddr_in *peer, uint16_t seq,
+			 uint32_t teic, uint8_t cause, uint64_t imsi, const struct in_addr *sgsn_addr, struct pdp_t *pdpctx, uint16_t sapi, uint8_t *mmctx, int mm_len, void *cbp);
 
-extern int gtp_fd(struct gsn_t *gsn);
 extern int gtp_decaps0(struct gsn_t *gsn);
 extern int gtp_decaps1c(struct gsn_t *gsn);
 extern int gtp_decaps1u(struct gsn_t *gsn);
-extern int gtp_retrans(struct gsn_t *gsn);
-extern int gtp_retranstimeout(struct gsn_t *gsn, struct timeval *timeout);
-
-extern int gtp_set_cb_delete_context(struct gsn_t *gsn,
-				     int (*cb_delete_context) (struct pdp_t *
-							       pdp));
-/*extern int gtp_set_cb_create_context(struct gsn_t *gsn,
-  int (*cb_create_context) (struct pdp_t* pdp)); */
-
-extern int gtp_set_cb_unsup_ind(struct gsn_t *gsn,
-				int (*cb) (struct sockaddr_in * peer));
-
-extern int gtp_set_cb_extheader_ind(struct gsn_t *gsn,
-				    int (*cb) (struct sockaddr_in * peer));
-
-extern int gtp_set_cb_conf(struct gsn_t *gsn,
-			   int (*cb) (int type, int cause, struct pdp_t * pdp,
-				      void *cbp));
-
-int gtp_set_cb_recovery(struct gsn_t *gsn,
-			int (*cb) (struct sockaddr_in * peer,
-				   uint8_t recovery));
-
-/* Internal functions (not part of the API */
+/* Internal functions (not part of the API) */
 
 extern int gtp_echo_req(struct gsn_t *gsn, int version, void *cbp,
 			struct in_addr *inetaddrs);
@@ -408,5 +326,6 @@ extern int eua2ipv4(struct in_addr *dst, struct ul66_t *eua);
 extern int gsna2in_addr(struct in_addr *dst, struct ul16_t *gsna);
 extern int in_addr2gsna(struct ul16_t *gsna, struct in_addr *src);
 extern const char *imsi_gtp2str(const uint64_t *imsi);
+extern uint64_t gtp_imsi_str2gtp(const char *str);
 
 #endif /* !_GTP_H */

gtp/gtpie.c

@@ -1,17 +1,17 @@
-/* 
+/*
  *  OsmoGGSN - Gateway GPRS Support Node
  *  Copyright (C) 2002 Mondru AB.
- * 
+ *
  *  The contents of this file may be used under the terms of the GNU
  *  General Public License Version 2, provided that the above copyright
  *  notice and this permission notice is included in all copies or
  *  substantial portions of the software.
- * 
+ *
  */
 
 /*
- * gtpie.c: Contains functions to encapsulate and decapsulate GTP 
- * information elements 
+ * gtpie.c: Contains functions to encapsulate and decapsulate GTP
+ * information elements
  *
  *
  * Encapsulation

gtp/pdp.c

@@ -1,17 +1,17 @@
-/* 
+/*
  *  OsmoGGSN - Gateway GPRS Support Node
  *  Copyright (C) 2002, 2003, 2004 Mondru AB.
  *  Copyright (C) 2017 Harald Welte <laforge@gnumonks.org>
- * 
+ *
  *  The contents of this file may be used under the terms of the GNU
  *  General Public License Version 2, provided that the above copyright
  *  notice and this permission notice is included in all copies or
  *  substantial portions of the software.
- * 
+ *
  */
 
 /*
- * pdp.c: 
+ * pdp.c:
  *
  */
 
@@ -31,30 +31,23 @@
 #include "pdp.h"
 #include "gtp.h"
 #include "lookupa.h"
-
-/* ***********************************************************
- * Global variables TODO: most should be moved to gsn_t
- *************************************************************/
-
-static struct pdp_t pdpa[PDP_MAX];	/* PDP storage */
-static struct pdp_t *hashtid[PDP_MAX];	/* Hash table for IMSI + NSAPI */
-/* struct pdp_t* haship[PDP_MAX];  Hash table for IP and network interface */
+#include "queue.h"
 
 /* ***********************************************************
  * Functions related to PDP storage
  *
  * Lifecycle
- * For a GGSN pdp context life begins with the reception of a 
+ * For a GGSN pdp context life begins with the reception of a
  * create pdp context request. It normally ends with the reception
  * of a delete pdp context request, but will also end with the
- * reception of an error indication message. 
+ * reception of an error indication message.
  * Provisions should probably be made for terminating pdp contexts
- * based on either idle timeout, or by sending downlink probe 
+ * based on either idle timeout, or by sending downlink probe
  * messages (ping?) to see if the MS is still responding.
- * 
+ *
  * For an SGSN pdp context life begins with the application just
  * before sending off a create pdp context request. It normally
- * ends when a delete pdp context response message is received 
+ * ends when a delete pdp context response message is received
  * from the GGSN, but should also end when with the reception of
  * an error indication message.
  *
@@ -64,15 +57,15 @@ static struct pdp_t *hashtid[PDP_MAX];	/* Hash table for IMSI + NSAPI */
  * Downlink packets received in the GGSN are identified only by their
  * network interface together with their destination IP address (Two
  * network interfaces can use the same private IP address). Each IMSI
- * (mobile station) can have several PDP contexts using the same IP 
+ * (mobile station) can have several PDP contexts using the same IP
  * address. In this case the traffic flow template (TFT) is used to
- * determine the correct PDP context for a particular IMSI. Also it 
+ * determine the correct PDP context for a particular IMSI. Also it
  * should be possible for each PDP context to use several IP adresses
  * For fixed wireless access a mobile station might need a full class
  * C network. Even in the case of several IP adresses the PDP context
  * should be determined on the basis of the network IP address.
  * Thus we need a hash table based on network interface + IP address.
- * 
+ *
  * Uplink packets are for GTP0 identified by their IMSI and NSAPI, which
  * is collectively called the tunnel identifier. There is also a 16 bit
  * flow label that can be used for identification of uplink packets. This
@@ -85,7 +78,7 @@ static struct pdp_t *hashtid[PDP_MAX];	/* Hash table for IMSI + NSAPI */
  * Thus we need a hash table based on TID (IMSI and NSAPI). The TEID will
  * be used for directly addressing the PDP context.
 
- * pdp_newpdp 
+ * pdp_newpdp
  * Gives you a pdp context with no hash references In some way
  * this should have a limited lifetime.
  *
@@ -112,11 +105,16 @@ static struct pdp_t *hashtid[PDP_MAX];	/* Hash table for IMSI + NSAPI */
  *
  *************************************************************/
 
-int pdp_init()
+static struct gsn_t *g_gsn;
+
+int pdp_init(struct gsn_t *gsn)
 {
-	memset(&pdpa, 0, sizeof(pdpa));
-	memset(&hashtid, 0, sizeof(hashtid));
-	/*  memset(&haship, 0, sizeof(haship)); */
+	if (!g_gsn) {
+		g_gsn = gsn;
+	} else {
+		LOGP(DLGTP, LOGL_FATAL, "This interface is depreacted and doesn't support multiple GGSN!");
+		return -1;
+	}
 
 	return 0;
 }
@@ -124,6 +122,13 @@ int pdp_init()
 int pdp_newpdp(struct pdp_t **pdp, uint64_t imsi, uint8_t nsapi,
 	       struct pdp_t *pdp_old)
 {
+	return gtp_pdp_newpdp(g_gsn, pdp, imsi, nsapi, pdp_old);
+}
+
+int gtp_pdp_newpdp(struct gsn_t *gsn, struct pdp_t **pdp, uint64_t imsi, uint8_t nsapi,
+	       struct pdp_t *pdp_old)
+{
+	struct pdp_t *pdpa = gsn->pdpa;
 	int n;
 	for (n = 0; n < PDP_MAX; n++) {	/* TODO: Need to do better than linear search */
 		if (pdpa[n].inuse == 0) {
@@ -133,6 +138,7 @@ int pdp_newpdp(struct pdp_t **pdp, uint64_t imsi, uint8_t nsapi,
 			else
 				memset(*pdp, 0, sizeof(struct pdp_t));
 			(*pdp)->inuse = 1;
+			(*pdp)->gsn = gsn;
 			(*pdp)->imsi = imsi;
 			(*pdp)->nsapi = nsapi;
 			(*pdp)->fllc = (uint16_t) n + 1;
@@ -151,7 +157,7 @@ int pdp_newpdp(struct pdp_t **pdp, uint64_t imsi, uint8_t nsapi,
 			}
 			/* Default: Generate G-PDU sequence numbers on Tx */
 			(*pdp)->tx_gpdu_seq = true;
-
+			INIT_LLIST_HEAD(&(*pdp)->qmsg_list_req);
 			return 0;
 		}
 	}
@@ -160,6 +166,18 @@ int pdp_newpdp(struct pdp_t **pdp, uint64_t imsi, uint8_t nsapi,
 
 int pdp_freepdp(struct pdp_t *pdp)
 {
+	struct qmsg_t *qmsg, *qmsg2;
+	struct pdp_t *pdpa = pdp->gsn->pdpa;
+	int rc;
+
+	/* Remove all enqueued messages belonging to this pdp from req tx transmit
+	   queue. queue_freemsg will call llist_del(). */
+	llist_for_each_entry_safe(qmsg, qmsg2, &pdp->qmsg_list_req, entry) {
+		if ((rc = queue_freemsg(pdp->gsn->queue_req, qmsg)))
+			LOGP(DLGTP, LOGL_ERROR,
+			     "Failed freeing qmsg from qmsg_list_req during pdp_freepdp()! %d\n", rc);
+	}
+
 	pdp_tiddel(pdp);
 
 	/* Remove any references in primary context */
@@ -174,12 +192,20 @@ int pdp_freepdp(struct pdp_t *pdp)
 
 int pdp_getpdp(struct pdp_t **pdp)
 {
-	*pdp = &pdpa[0];
+	*pdp = &g_gsn->pdpa[0];
 	return 0;
 }
 
 int pdp_getgtp0(struct pdp_t **pdp, uint16_t fl)
 {
+	return gtp_pdp_getgtp0(g_gsn, pdp, fl);
+}
+
+
+int gtp_pdp_getgtp0(struct gsn_t *gsn, struct pdp_t **pdp, uint16_t fl)
+{
+	struct pdp_t *pdpa = gsn->pdpa;
+
 	if ((fl > PDP_MAX) || (fl < 1)) {
 		return EOF;	/* Not found */
 	} else {
@@ -194,6 +220,13 @@ int pdp_getgtp0(struct pdp_t **pdp, uint16_t fl)
 
 int pdp_getgtp1(struct pdp_t **pdp, uint32_t tei)
 {
+	return gtp_pdp_getgtp1(g_gsn, pdp, tei);
+}
+
+int gtp_pdp_getgtp1(struct gsn_t *gsn, struct pdp_t **pdp, uint32_t tei)
+{
+	struct pdp_t *pdpa = gsn->pdpa;
+
 	if ((tei > PDP_MAX) || (tei < 1)) {
 		return EOF;	/* Not found */
 	} else {
@@ -209,6 +242,12 @@ int pdp_getgtp1(struct pdp_t **pdp, uint32_t tei)
 /* get a PDP based on the *peer* address + TEI-Data.  Used for matching inbound Error Ind */
 int pdp_getgtp1_peer_d(struct pdp_t **pdp, const struct sockaddr_in *peer, uint32_t teid_gn)
 {
+	return gtp_pdp_getgtp1_peer_d(g_gsn, pdp, peer, teid_gn);
+}
+
+int gtp_pdp_getgtp1_peer_d(struct gsn_t *gsn, struct pdp_t **pdp, const struct sockaddr_in *peer, uint32_t teid_gn)
+{
+	struct pdp_t *pdpa = gsn->pdpa;
 	unsigned int i;
 
 	/* this is O(n) but we don't have (nor want) another hash... */
@@ -231,6 +270,7 @@ int pdp_tidhash(uint64_t tid)
 
 int pdp_tidset(struct pdp_t *pdp, uint64_t tid)
 {
+	struct pdp_t **hashtid = pdp->gsn->hashtid;
 	int hash = pdp_tidhash(tid);
 	struct pdp_t *pdp2;
 	struct pdp_t *pdp_prev = NULL;
@@ -249,6 +289,7 @@ int pdp_tidset(struct pdp_t *pdp, uint64_t tid)
 
 int pdp_tiddel(struct pdp_t *pdp)
 {
+	struct pdp_t **hashtid = pdp->gsn->hashtid;
 	int hash = pdp_tidhash(pdp->tid);
 	struct pdp_t *pdp2;
 	struct pdp_t *pdp_prev = NULL;
@@ -270,6 +311,12 @@ int pdp_tiddel(struct pdp_t *pdp)
 
 int pdp_tidget(struct pdp_t **pdp, uint64_t tid)
 {
+	return gtp_pdp_tidget(g_gsn, pdp, tid);
+}
+
+int gtp_pdp_tidget(struct gsn_t *gsn, struct pdp_t **pdp, uint64_t tid)
+{
+	struct pdp_t **hashtid = gsn->hashtid;
 	int hash = pdp_tidhash(tid);
 	struct pdp_t *pdp2;
 	DEBUGP(DLGTP, "Begin pdp_tidget tid = %"PRIx64"\n", tid);
@@ -286,102 +333,17 @@ int pdp_tidget(struct pdp_t **pdp, uint64_t tid)
 
 int pdp_getimsi(struct pdp_t **pdp, uint64_t imsi, uint8_t nsapi)
 {
-	return pdp_tidget(pdp,
-			  (imsi & 0x0fffffffffffffffull) +
-			  ((uint64_t) nsapi << 60));
+	return gtp_pdp_getimsi(g_gsn, pdp, imsi, nsapi);
 }
 
-/*
-int pdp_iphash(void* ipif, struct ul66_t *eua) {
-  /#printf("IPhash %ld\n", lookup(eua->v, eua->l, ipif) % PDP_MAX);#/
-  return (lookup(eua->v, eua->l, ipif) % PDP_MAX);
-}
-    
-int pdp_ipset(struct pdp_t *pdp, void* ipif, struct ul66_t *eua) {
-  int hash;
-  struct pdp_t *pdp2;
-  struct pdp_t *pdp_prev = NULL;
-
-  if (PDP_DEBUG) printf("Begin pdp_ipset %d %d %2x%2x%2x%2x\n",
-			(unsigned) ipif, eua->l,
-			eua->v[2], eua->v[3], 
-			eua->v[4], eua->v[5]);
-
-  pdp->ipnext = NULL;
-  pdp->ipif = ipif;
-  pdp->eua.l = eua->l;
-  memcpy(pdp->eua.v, eua->v, eua->l);
-
-  hash = pdp_iphash(pdp->ipif, &pdp->eua);
-
-  for (pdp2 = haship[hash]; pdp2; pdp2 = pdp2->ipnext)
-    pdp_prev = pdp2;
-  if (!pdp_prev) 
-    haship[hash] = pdp;
-  else 
-    pdp_prev->ipnext = pdp;
-  if (PDP_DEBUG) printf("End pdp_ipset\n");
-  return 0;
+int gtp_pdp_getimsi(struct gsn_t *gsn, struct pdp_t **pdp, uint64_t imsi, uint8_t nsapi)
+{
+	return gtp_pdp_tidget(gsn, pdp, pdp_gettid(imsi, nsapi));
 }
 
-int pdp_ipdel(struct pdp_t *pdp) {
-  int hash = pdp_iphash(pdp->ipif, &pdp->eua);
-  struct pdp_t *pdp2;
-  struct pdp_t *pdp_prev = NULL;
-  if (PDP_DEBUG) printf("Begin pdp_ipdel\n");
-  for (pdp2 = haship[hash]; pdp2; pdp2 = pdp2->ipnext) {
-    if (pdp2 == pdp) {
-      if (!pdp_prev) 
-	haship[hash] = pdp2->ipnext;
-      else 
-	pdp_prev->ipnext = pdp2->ipnext;
-      if (PDP_DEBUG) printf("End pdp_ipdel: PDP found\n");
-      return 0;
-    }
-    pdp_prev = pdp2;
-  }
-  if (PDP_DEBUG) printf("End pdp_ipdel: PDP not found\n");
-  return EOF; /# End of linked list and not found #/
-}
 
-int pdp_ipget(struct pdp_t **pdp, void* ipif, struct ul66_t *eua) {
-  int hash = pdp_iphash(ipif, eua);
-  struct pdp_t *pdp2;
-  /#printf("Begin pdp_ipget %d %d %2x%2x%2x%2x\n", (unsigned)ipif, eua->l, 
-    eua->v[2],eua->v[3],eua->v[4],eua->v[5]);#/
-  for (pdp2 = haship[hash]; pdp2; pdp2 = pdp2->ipnext) {
-    if ((pdp2->ipif == ipif) && (pdp2->eua.l == eua->l) && 
-	(memcmp(&pdp2->eua.v, &eua->v, eua->l) == 0)) {
-      *pdp = pdp2;
-      /#printf("End pdp_ipget. Found\n");#/
-      return 0;
-    }
-  }
-  if (PDP_DEBUG) printf("End pdp_ipget Notfound %d %d %2x%2x%2x%2x\n", 
-	 (unsigned)ipif, eua->l, eua->v[2],eua->v[3],eua->v[4],eua->v[5]);
-  return EOF; /# End of linked list and not found #/
-}
-*/
 /* Various conversion functions */
 
-int pdp_ntoeua(struct in_addr *src, struct ul66_t *eua)
-{
-	eua->l = 6;
-	eua->v[0] = PDP_EUA_ORG_IETF;
-	eua->v[1] = PDP_EUA_TYPE_v4;
-	memcpy(&eua->v[2], src, 4);	/* Copy a 4 byte address */
-	return 0;
-}
-
-int pdp_euaton(struct ul66_t *eua, struct in_addr *dst)
-{
-	if ((eua->l != 6) || (eua->v[0] != PDP_EUA_ORG_IETF) || (eua->v[1] != PDP_EUA_TYPE_v4)) {
-		return EOF;
-	}
-	memcpy(dst, &eua->v[2], 4);	/* Copy a 4 byte address */
-	return 0;
-}
-
 uint64_t pdp_gettid(uint64_t imsi, uint8_t nsapi)
 {
 	return (imsi & 0x0fffffffffffffffull) + ((uint64_t) nsapi << 60);
@@ -392,3 +354,17 @@ void pdp_set_imsi_nsapi(struct pdp_t *pdp, uint64_t teid)
 	pdp->imsi = teid & 0x0fffffffffffffffull;
 	pdp->nsapi = (teid & 0xf000000000000000ull) >> 60;
 }
+
+/* Count amount of secondary PDP contexts linked to this primary PDP context
+ * (itself included). Must be called on a primary PDP context. */
+unsigned int pdp_count_secondary(const struct pdp_t *pdp)
+{
+	unsigned int n;
+	unsigned int count = 0;
+	OSMO_ASSERT(!pdp->secondary);
+
+	for (n = 0; n < PDP_MAXNSAPI; n++)
+		if (pdp->secondary_tei[n])
+			count++;
+	return count;
+}

gtp/pdp.h

@@ -1,19 +1,23 @@
-/* 
+/*
  *  OsmoGGSN - Gateway GPRS Support Node
  *  Copyright (C) 2002, 2003 Mondru AB.
  *  Copyright (C) 2017 Harald Welte <laforge@gnumonks.org>
- * 
+ *
  *  The contents of this file may be used under the terms of the GNU
  *  General Public License Version 2, provided that the above copyright
  *  notice and this permission notice is included in all copies or
  *  substantial portions of the software.
- * 
+ *
  */
 
 #ifndef _PDP_H
 #define _PDP_H
 
 #include <stdbool.h>
+#include <netinet/in.h>
+
+#include <osmocom/core/defs.h>
+#include <osmocom/core/linuxlist.h>
 
 struct gsn_t;
 
@@ -26,6 +30,7 @@ struct gsn_t;
 #define PDP_EUA_ORG_IETF	0xF1
 #define PDP_EUA_TYPE_v4		0x21
 #define PDP_EUA_TYPE_v6		0x57
+#define PDP_EUA_TYPE_v4v6	0x8D
 
 /* GTP Information elements from 29.060 v3.9.0 7.7 Information Elements */
 /* Also covers version 0. Note that version 0 6: QOS Profile was superceded *
@@ -71,16 +76,16 @@ struct ul255_t {
  * and 09.60.
  * 31 * 4 + 15 structs +  = 120 + 15 structs ~ 2k / context
  * Structs: IP address 16+4 bytes (6), APN 255 bytes (2)
- * QOS: 255 bytes (3), msisdn 16 bytes (1), 
+ * QOS: 255 bytes (3), msisdn 16 bytes (1),
  *
  * TODO: We need to consider who manages the pdp_t hash tables
  * Is it gtp_lib, or is it the application?
- * I suppose that it will be gtp_lib. 
+ * I suppose that it will be gtp_lib.
  * SGSN will ask gtplib for new pdp_t. Fill out the fields,
  * and pass it on to gtp_create_pdp_req.
  * GGSN will receive gtp_create_pdp_ind, create new pdp_t and
  * send responce to SGSN.
- * SGSN will receive response and gtplib will find the 
+ * SGSN will receive response and gtplib will find the
  * original pdp_t corresponding to the request. This will be
  * passed on to the application.
  * Eventually the SGSN will close the connection, and the
@@ -88,10 +93,10 @@ struct ul255_t {
  * This means that gtplib need to have functions to
  * allocate, free, sort and find pdp_t
  * (newpdp, freepdp, getpdp)
- * Hash tables: TID, IMSI, IP etc.) 
+ * Hash tables: TID, IMSI, IP etc.)
  *
  *
- * Secondary PDP Context Activation Procedure 
+ * Secondary PDP Context Activation Procedure
  *
  * With GTP version 1 it is possible to establish multiple PDP
  * contexts with the same IP address. With this scheme the first
@@ -121,7 +126,7 @@ struct pdp_t {
 	/* Parameters shared by all PDP context belonging to the same MS */
 
 	void *ipif;		/* IP network interface */
-	void *peer;		/* Pointer to peer protocol */
+	void *peer[2];		/* Pointer to peer protocol */
 	void *asap;		/* Application specific service access point */
 
 	uint64_t imsi;		/* International Mobile Subscriber Identity. */
@@ -234,40 +239,42 @@ struct pdp_t {
 	/* to be used by libgtp callers/users (to attach their own private state) */
 	void *priv;
 
-	struct gsn_t *gsn;
+	struct gsn_t *gsn; /* Back pointer to GSN where this pdp ctx belongs to */
 
 	bool tx_gpdu_seq;		/* Transmit (true) or suppress G-PDU sequence numbers */
+
+	struct llist_head qmsg_list_req; /* list of req qmsg_t in retrans queue belonging this pdp ctx */
 };
 
 /* functions related to pdp_t management */
-int pdp_init();
-int pdp_newpdp(struct pdp_t **pdp, uint64_t imsi, uint8_t nsapi,
-	       struct pdp_t *pdp_old);
+int gtp_pdp_newpdp(struct gsn_t *gsn, struct pdp_t **pdp, uint64_t imsi,
+		   uint8_t nsapi, struct pdp_t *pdp_old);
 int pdp_freepdp(struct pdp_t *pdp);
-int pdp_getpdp(struct pdp_t **pdp);
-
-int pdp_getgtp0(struct pdp_t **pdp, uint16_t fl);
-int pdp_getgtp1(struct pdp_t **pdp, uint32_t tei);
-int pdp_getgtp1_peer_d(struct pdp_t **pdp, const struct sockaddr_in *peer, uint32_t teid_gn);
-
-int pdp_getimsi(struct pdp_t **pdp, uint64_t imsi, uint8_t nsapi);
+int gtp_pdp_getgtp0(struct gsn_t *gsn, struct pdp_t **pdp, uint16_t fl);
+int gtp_pdp_getgtp1(struct gsn_t *gsn, struct pdp_t **pdp, uint32_t tei);
+int gtp_pdp_getgtp1_peer_d(struct gsn_t *gsn, struct pdp_t **pdp, const struct sockaddr_in *peer, uint32_t teid_gn);
+int gtp_pdp_getimsi(struct gsn_t *gsn, struct pdp_t **pdp, uint64_t imsi, uint8_t nsapi);
+int gtp_pdp_tidget(struct gsn_t *gsn, struct pdp_t **pdp, uint64_t tid);
 
 int pdp_tidhash(uint64_t tid);
 int pdp_tidset(struct pdp_t *pdp, uint64_t tid);
 int pdp_tiddel(struct pdp_t *pdp);
-int pdp_tidget(struct pdp_t **pdp, uint64_t tid);
 
+uint64_t pdp_gettid(uint64_t imsi, uint8_t nsapi);
 void pdp_set_imsi_nsapi(struct pdp_t *pdp, uint64_t teid);
 
-/*
-int pdp_iphash(void* ipif, struct ul66_t *eua);
-int pdp_ipset(struct pdp_t *pdp, void* ipif, struct ul66_t *eua);
-int pdp_ipdel(struct pdp_t *pdp);
-int pdp_ipget(struct pdp_t **pdp, void* ipif, struct ul66_t *eua);
-*/
+unsigned int pdp_count_secondary(const struct pdp_t *pdp);
+
+/* Deprecated APIs (support for only 1 GSN per process). Must be used only after first call to gtp_new() and until it is freed. */
+int pdp_init(struct gsn_t *gsn); /* Use only allowed inside libgtp to keep compatiblity with deprecated APIs defined here. */
+int pdp_newpdp(struct pdp_t **pdp, uint64_t imsi, uint8_t nsapi,
+	       struct pdp_t *pdp_old) OSMO_DEPRECATED("Use gtp_pdp_newpdp() instead");
+int pdp_getpdp(struct pdp_t **pdp) OSMO_DEPRECATED("Use gsn_t->pdpa field instead");
+int pdp_getgtp0(struct pdp_t **pdp, uint16_t fl) OSMO_DEPRECATED("Use gtp_pdp_getgtp0() instead");
+int pdp_getgtp1(struct pdp_t **pdp, uint32_t tei) OSMO_DEPRECATED("Use gtp_pdp_getgtp1() instead");
+int pdp_getgtp1_peer_d(struct pdp_t **pdp, const struct sockaddr_in *peer, uint32_t teid_gn) OSMO_DEPRECATED("Use gtp_pdp_getgtp1_peer_d() instead");
+int pdp_getimsi(struct pdp_t **pdp, uint64_t imsi, uint8_t nsapi) OSMO_DEPRECATED("Use gtp_pdp_getimsi() instead");
+int pdp_tidget(struct pdp_t **pdp, uint64_t tid) OSMO_DEPRECATED("Use gtp_pdp_tidget() instead");
 
-int pdp_ntoeua(struct in_addr *src, struct ul66_t *eua);
-int pdp_euaton(struct ul66_t *eua, struct in_addr *dst);
-uint64_t pdp_gettid(uint64_t imsi, uint8_t nsapi);
 
 #endif /* !_PDP_H */

gtp/queue.c

@@ -1,14 +1,14 @@
-/* 
+/*
  *  OsmoGGSN - Gateway GPRS Support Node
  *  Copyright (C) 2002, 2003, 2004 Mondru AB.
  *  Copyright (C) 2011 Harald Welte <laforge@gnumonks.org>
  *  Copyright (C) 2016 sysmocom - s.f.m.c. GmbH
- * 
+ *
  *  The contents of this file may be used under the terms of the GNU
  *  General Public License Version 2, provided that the above copyright
  *  notice and this permission notice is included in all copies or
  *  substantial portions of the software.
- * 
+ *
  */
 
 /*
@@ -62,7 +62,7 @@ static int queue_seqhash(struct sockaddr_in *peer, uint16_t seq)
 	return seq % QUEUE_HASH_SIZE;
 }
 
-/*! \brief Insert a message with given sequence number into the hash
+/*! \brief Insert a message with given sequence number into the hash.
  *
  * This function sets the peer and the seq of the qmsg and then inserts
  * the qmsg into the queue hash.  To do so, it does a hashtable lookup
@@ -79,7 +79,7 @@ static int queue_seqset(struct queue_t *queue, struct qmsg_t *qmsg,
 	if (QUEUE_DEBUG)
 		printf("Begin queue_seqset seq = %d\n", (int)seq);
 	if (QUEUE_DEBUG)
-		printf("SIZEOF PEER %d, *PEER %d\n", sizeof(peer),
+		printf("SIZEOF PEER %zu, *PEER %zu\n", sizeof(peer),
 		       sizeof(*peer));
 
 	qmsg->seq = seq;
@@ -121,7 +121,10 @@ static int queue_seqdel(struct queue_t *queue, struct qmsg_t *qmsg)
 	return EOF;		/* End of linked list and not found */
 }
 
-/*! \brief Allocates and initialises new queue structure */
+/*! Allocates and initialises new queue structure.
+ *  \param[out] queue pointer where to store the allocated object. Must be freed with queue_free
+ *  \returns zero on success, non-zero on error
+ */
 int queue_new(struct queue_t **queue)
 {
 	if (QUEUE_DEBUG)
@@ -138,7 +141,10 @@ int queue_new(struct queue_t **queue)
 	return 0;
 }
 
-/*! \brief Deallocates queue structure */
+/*! Deallocates queue structure.
+ *  \param[in] queue pointer previously allocated by queue_new
+ *  \returns zero on success, non-zero on error.
+ */
 int queue_free(struct queue_t *queue)
 {
 	if (QUEUE_DEBUG)
@@ -149,7 +155,13 @@ int queue_free(struct queue_t *queue)
 	return 0;
 }
 
-/*! \brief Add a new message to the queue */
+/*! Add a new message to the queue.
+ *  \param[in] queue pointer previously allocated by queue_new
+ *  \param[out] qmsg first message from the queue (if succeeds)
+ *  \param[in] peer who sent the message to add
+ *  \param[in] seq sequence number of the message to add
+ *  \returns zero on success, non-zero on error.
+ */
 int queue_newmsg(struct queue_t *queue, struct qmsg_t **qmsg,
 		 struct sockaddr_in *peer, uint16_t seq)
 {
@@ -160,6 +172,7 @@ int queue_newmsg(struct queue_t *queue, struct qmsg_t **qmsg,
 	} else {
 		*qmsg = &queue->qmsga[queue->next];
 		queue_seqset(queue, *qmsg, peer, seq);
+		INIT_LLIST_HEAD(&(*qmsg)->entry);
 		(*qmsg)->state = 1;	/* Space taken */
 		(*qmsg)->this = queue->next;
 		(*qmsg)->next = -1;	/* End of the queue */
@@ -176,7 +189,11 @@ int queue_newmsg(struct queue_t *queue, struct qmsg_t **qmsg,
 	}
 }
 
-/*! \brief Simply remoev a given qmsg_t from the queue
+
+/*! Remove an element from the queue.
+ *  \param[in] queue pointer previously allocated by queue_new
+ *  \param[in] qmsg message to free
+ *  \returns zero on success, non-zero on error.
  *
  * Internally, we first delete the entry from the queue, and then update
  * up our global queue->first / queue->last pointers.  Finally,
@@ -190,6 +207,8 @@ int queue_freemsg(struct queue_t *queue, struct qmsg_t *qmsg)
 		return EOF;	/* Not in queue */
 	}
 
+	llist_del(&qmsg->entry);
+
 	queue_seqdel(queue, qmsg);
 
 	if (qmsg->next == -1)	/* Are we the last in queue? */
@@ -210,7 +229,11 @@ int queue_freemsg(struct queue_t *queue, struct qmsg_t *qmsg)
 	return 0;
 }
 
-/*! \brief Move a given qmsg_t to the end of the queue ?!? */
+/*! Move a given qmsg_t to the end of the queue.
+ *  \param[in] queue pointer previously allocated by queue_new
+ *  \param[in] qmsg message to move to the end of the queue
+ *  \returns zero on success, non-zero on error.
+ */
 int queue_back(struct queue_t *queue, struct qmsg_t *qmsg)
 {
 	if (QUEUE_DEBUG)
@@ -236,7 +259,11 @@ int queue_back(struct queue_t *queue, struct qmsg_t *qmsg)
 	return 0;
 }
 
-/*! \brief Get the first element in the entire queue */
+/*! Get the first element in the entire queue.
+ *  \param[in] queue pointer previously allocated by queue_new
+ *  \param[out] qmsg first message from the queue (if succeeds)
+ *  \returns zero on success, non-zero on error.
+ */
 int queue_getfirst(struct queue_t *queue, struct qmsg_t **qmsg)
 {
 	/*printf("queue_getfirst\n"); */
@@ -250,7 +277,13 @@ int queue_getfirst(struct queue_t *queue, struct qmsg_t **qmsg)
 	return 0;
 }
 
-/*! \brief Get a queue entry for a given peer + seq */
+/*! Get a queue entry for a given peer + seq.
+ *  \param[in] queue pointer previously allocated by queue_new
+ *  \param[out] qmsg first message from the queue (if succeeds)
+ *  \param[in] peer who sent the message to retrieve
+ *  \param[in] seq sequence number of the message to retrive
+ *  \returns zero on success, non-zero on error.
+ */
 int queue_seqget(struct queue_t *queue, struct qmsg_t **qmsg,
 		 struct sockaddr_in *peer, uint16_t seq)
 {
@@ -272,7 +305,14 @@ int queue_seqget(struct queue_t *queue, struct qmsg_t **qmsg,
 	return EOF;		/* End of linked list and not found */
 }
 
-/*! \brief look-up a given seq/peer, return cbp + type and free entry */
+/*! look-up a given seq/peer, return cbp + type and free entry.
+ *  \param[in] queue pointer previously allocated by queue_new
+ *  \param[in] peer who sent the message to retrieve
+ *  \param[in] seq sequence number of the message to retrive
+ *  \param[out] type GTP message type
+ *  \param[out] type callback pointer of the message
+ *  \returns zero on success, non-zero on error.
+ */
 int queue_freemsg_seq(struct queue_t *queue, struct sockaddr_in *peer,
 		      uint16_t seq, uint8_t * type, void **cbp)
 {

gtp/queue.h

@@ -1,12 +1,12 @@
-/* 
+/*
  *  OsmoGGSN - Gateway GPRS Support Node
  *  Copyright (C) 2002 Mondru AB.
- * 
+ *
  *  The contents of this file may be used under the terms of the GNU
  *  General Public License Version 2, provided that the above copyright
  *  notice and this permission notice is included in all copies or
  *  substantial portions of the software.
- * 
+ *
  */
 
 /*
@@ -17,9 +17,13 @@
 #ifndef _QUEUE_H
 #define _QUEUE_H
 
+#include <osmocom/core/linuxlist.h>
+
+#include "gtp.h"
+
 #define QUEUE_DEBUG 0		/* Print debug information */
 
-#define QUEUE_SIZE 1024		/* Size of retransmission queue */
+#define QUEUE_SIZE (PDP_MAX*2)	/* Size of retransmission queue */
 #define QUEUE_HASH_SIZE 65536	/* Size of hash table (2^16) */
 
 struct qmsg_t {			/* Holder for queued packets */
@@ -37,6 +41,7 @@ struct qmsg_t {			/* Holder for queued packets */
 	int this;		/* Pointer to myself */
 	time_t timeout;		/* When do we retransmit this packet? */
 	int retrans;		/* How many times did we retransmit this? */
+	struct llist_head entry; /* Listed with other qmsg_t belonging to a pdp_t->qmsg_list_req */
 };
 
 struct queue_t {

lib/Makefile.am

@@ -1,7 +1,43 @@
 noinst_LIBRARIES = libmisc.a
 
-noinst_HEADERS = gnugetopt.h ippool.h lookup.h syserr.h tun.h in46_addr.h
+noinst_HEADERS = \
+		 checksum.h \
+		 gnugetopt.h \
+		 gtp-kernel.h \
+		 icmpv6.h \
+		 in46_addr.h \
+		 ippool.h \
+		 lookup.h \
+		 netdev.h \
+		 netns.h \
+		 syserr.h \
+		 tun.h \
+		 util.h \
+		 $(NULL)
 
-AM_CFLAGS = -O2 -fno-builtin -Wall -DSBINDIR='"$(sbindir)"' -ggdb $(LIBOSMOCORE_CFLAGS)
+AM_CFLAGS = \
+	    -fno-builtin \
+	    -Wall \
+	    -DSBINDIR='"$(sbindir)"' \
+	    $(LIBOSMOCORE_CFLAGS) \
+	    $(NULL)
 
-libmisc_a_SOURCES = getopt1.c getopt.c ippool.c lookup.c tun.c debug.c in46_addr.c
+libmisc_a_SOURCES = \
+		    checksum.c \
+		    debug.c \
+		    getopt.c \
+		    getopt1.c \
+		    icmpv6.c \
+		    in46_addr.c \
+		    ippool.c \
+		    lookup.c \
+		    netdev.c \
+		    netns.c \
+		    tun.c \
+		    util.c \
+		    $(NULL)
+
+if ENABLE_GTP_KERNEL
+AM_CFLAGS += -DGTP_KERNEL $(LIBGTPNL_CFLAGS)
+libmisc_a_SOURCES += gtp-kernel.c
+endif

lib/debug.c

@@ -24,12 +24,12 @@ static const struct log_info_cat default_categories[] = {
 	[DSGSN] = {
 		.name = "DSGSN",
 		.description = "SGSN Emulator",
-		.enabled = 1, .loglevel = LOGL_NOTICE,
+		.enabled = 1, .loglevel = LOGL_INFO,
 	},
 	[DICMP6] = {
 		.name = "DICMP6",
 		.description = "ICMPv6",
-		.enabled = 1, .loglevel = LOGL_DEBUG,
+		.enabled = 1, .loglevel = LOGL_NOTICE,
 	},
 };
 

lib/gtp-kernel.c

@@ -0,0 +1,200 @@
+#ifdef __linux__
+#define _GNU_SOURCE 1		/* strdup() prototype, broken arpa/inet.h */
+#endif
+
+#include "../config.h"
+
+#ifdef HAVE_STDINT_H
+#include <stdint.h>
+#endif
+
+#include <stdio.h>
+#include <string.h>
+#include <stdlib.h>
+#include <inttypes.h>
+#include <sys/types.h>
+#include <arpa/inet.h>
+#include <net/if.h>
+
+#include <libgtpnl/gtp.h>
+#include <libgtpnl/gtpnl.h>
+
+#include <errno.h>
+
+#include <time.h>
+
+#include "../lib/tun.h"
+#include "../lib/syserr.h"
+#include "../lib/util.h"
+#include "../lib/ippool.h"
+#include "../gtp/pdp.h"
+#include "../gtp/gtp.h"
+
+#include "gtp-kernel.h"
+
+static void pdp_debug(const char *prefix, const char *devname, struct pdp_t *pdp)
+{
+	char buf4[INET_ADDRSTRLEN], buf6[INET6_ADDRSTRLEN];
+	struct ippoolm_t *peer;
+	struct in_addr ia;
+
+	buf4[0] = '\0';
+	if ((peer = pdp_get_peer_ipv(pdp, false)))
+		in46a_ntop(&peer->addr, buf4, sizeof(buf4));
+	buf6[0] = '\0';
+	if ((peer = pdp_get_peer_ipv(pdp, true)))
+		in46a_ntop(&peer->addr, buf6, sizeof(buf6));
+
+	gsna2in_addr(&ia, &pdp->gsnrc);
+
+	LOGPDPX(DGGSN, LOGL_DEBUG, pdp, "%s %s v%u TEID %"PRIx64" EUA=(%s,%s) SGSN=%s\n", prefix,
+		devname, pdp->version,
+		pdp->version == 0 ? pdp_gettid(pdp->imsi, pdp->nsapi) : pdp->teid_gn,
+		buf4, buf6, inet_ntoa(ia));
+}
+
+static struct {
+	int			genl_id;
+	struct mnl_socket	*nl;
+} gtp_nl;
+
+static int gtp_kernel_init_once(void)
+{
+	/* only initialize once */
+	if (gtp_nl.nl)
+		return 0;
+
+	gtp_nl.nl = genl_socket_open();
+	if (gtp_nl.nl == NULL) {
+		LOGP(DGGSN, LOGL_ERROR, "cannot create genetlink socket\n");
+		return -1;
+	}
+	gtp_nl.genl_id = genl_lookup_family(gtp_nl.nl, "gtp");
+	if (gtp_nl.genl_id < 0) {
+		LOGP(DGGSN, LOGL_ERROR, "cannot lookup GTP genetlink ID\n");
+		genl_socket_close(gtp_nl.nl);
+		gtp_nl.nl = NULL;
+		return -1;
+	}
+	LOGP(DGGSN, LOGL_NOTICE, "Initialized GTP kernel mode (genl ID is %d)\n", gtp_nl.genl_id);
+
+	return 0;
+}
+
+int gtp_kernel_create(int dest_ns, const char *devname, int fd0, int fd1u)
+{
+	if (gtp_kernel_init_once() < 0)
+		return -1;
+
+	return gtp_dev_create(dest_ns, devname, fd0, fd1u);
+}
+
+int gtp_kernel_create_sgsn(int dest_ns, const char *devname, int fd0, int fd1u)
+{
+	if (gtp_kernel_init_once() < 0)
+		return -1;
+
+	return gtp_dev_create_sgsn(dest_ns, devname, fd0, fd1u);
+}
+
+void gtp_kernel_stop(const char *devname)
+{
+	gtp_dev_destroy(devname);
+}
+
+int gtp_kernel_tunnel_add(struct pdp_t *pdp, const char *devname)
+{
+	int ms_addr_count;
+	struct in46_addr ms[2];
+	struct in46_addr sgsn;
+	struct gtp_tunnel *t;
+	int ret = 0;
+
+	pdp_debug(__func__, devname, pdp);
+
+	in46a_from_gsna(&pdp->gsnrc, &sgsn);
+
+	ms_addr_count = in46a_from_eua(&pdp->eua, ms);
+	if (ms_addr_count < 0)
+		return -1;
+
+	for (int i = 0; i < ms_addr_count; i++) {
+		t = gtp_tunnel_alloc();
+		if (t == NULL)
+			return -1;
+
+		gtp_tunnel_set_ifidx(t, if_nametoindex(devname));
+		gtp_tunnel_set_version(t, pdp->version);
+
+		if (in46a_to_af(&ms[i]) == AF_INET)
+			gtp_tunnel_set_ms_ip4(t, &ms[i].v4);
+		else {
+			/* In IPv6, EUA doesn't contain the actual IP
+			 * addr/prefix. Set higher bits to 0 to get the 64 bit
+			 * netmask. */
+			memset(((void *)&ms[i].v6) + 8, 0, 8);
+			gtp_tunnel_set_ms_ip6(t, &ms[i].v6);
+		}
+
+		if (in46a_to_af(&sgsn) == AF_INET)
+			gtp_tunnel_set_sgsn_ip4(t, &sgsn.v4);
+		else
+			gtp_tunnel_set_sgsn_ip6(t, &sgsn.v6);
+
+		if (pdp->version == 0) {
+			gtp_tunnel_set_tid(t, pdp_gettid(pdp->imsi, pdp->nsapi));
+			gtp_tunnel_set_flowid(t, pdp->flru);
+		} else {
+			gtp_tunnel_set_i_tei(t, pdp->teid_own);
+			/* use the TEI advertised by SGSN when sending packets
+			 * towards the SGSN */
+			gtp_tunnel_set_o_tei(t, pdp->teid_gn);
+		}
+
+		ret = gtp_add_tunnel(gtp_nl.genl_id, gtp_nl.nl, t);
+		gtp_tunnel_free(t);
+
+		if (ret != 0)
+			break;
+	}
+
+	return ret;
+}
+
+int gtp_kernel_tunnel_del(struct pdp_t *pdp, const char *devname)
+{
+	int ms_addr_count;
+	struct in46_addr ms[2];
+	struct gtp_tunnel *t;
+	int ret = 0;
+
+	pdp_debug(__func__, devname, pdp);
+
+	ms_addr_count = in46a_from_eua(&pdp->eua, ms);
+	if (ms_addr_count < 0)
+		return -1;
+
+	for (int i = 0; i < ms_addr_count; i++) {
+		t = gtp_tunnel_alloc();
+		if (t == NULL)
+			return -1;
+
+		gtp_tunnel_set_ifidx(t, if_nametoindex(devname));
+		gtp_tunnel_set_family(t, in46a_to_af(&ms[i]));
+		gtp_tunnel_set_version(t, pdp->version);
+		if (pdp->version == 0) {
+			gtp_tunnel_set_tid(t, pdp_gettid(pdp->imsi, pdp->nsapi));
+			gtp_tunnel_set_flowid(t, pdp->flru);
+		} else {
+			gtp_tunnel_set_i_tei(t, pdp->teid_own);
+		}
+
+		ret = gtp_del_tunnel(gtp_nl.genl_id, gtp_nl.nl, t);
+		gtp_tunnel_free(t);
+
+		if (ret != 0)
+			break;
+	}
+
+	return ret;
+}

lib/gtp-kernel.h

@@ -0,0 +1,38 @@
+#ifndef _GTP_KERNEL_H_
+#define _GTP_KERNEL_H_
+
+struct gengetopt_args_info;
+
+extern int debug;
+extern char *ipup;
+
+#ifdef GTP_KERNEL
+int gtp_kernel_create(int dest_ns, const char *devname, int fd0, int fd1u);
+int gtp_kernel_create_sgsn(int dest_ns, const char *devname, int fd0, int fd1u);
+void gtp_kernel_stop(const char *devname);
+
+int gtp_kernel_tunnel_add(struct pdp_t *pdp, const char *devname);
+int gtp_kernel_tunnel_del(struct pdp_t *pdp, const char *devname);
+
+#else
+static inline int gtp_kernel_create(int dest_ns, const char *devname, int fd0, int fd1u)
+{
+	SYS_ERR(DGGSN, LOGL_ERROR, 0, "ggsn compiled without GTP kernel support!\n");
+	return -1;
+}
+#define gtp_kernel_create_sgsn gtp_kernel_create
+
+static inline void gtp_kernel_stop(const char *devname) {}
+
+static inline int gtp_kernel_tunnel_add(struct pdp_t *pdp, const char *devname)
+{
+	return 0;
+}
+
+static inline int gtp_kernel_tunnel_del(struct pdp_t *pdp, const char *devname)
+{
+	return 0;
+}
+
+#endif
+#endif /* _GTP_KERNEL_H_ */

lib/icmpv6.c

@@ -1,7 +1,7 @@
 /* Minimal ICMPv6 code for generating router advertisements as required by
  * relevant 3GPP specs for a GGSN with IPv6 PDP contexts */
 
-/* (C) 2017 by Harald Welte <laforge@gnumonks.org> 
+/* (C) 2017 by Harald Welte <laforge@gnumonks.org>
  *
  * The contents of this file may be used under the terms of the GNU
  * General Public License Version 2, provided that the above copyright
@@ -25,8 +25,9 @@
 
 #include "../gtp/gtp.h"
 #include "../gtp/pdp.h"
-#include "../lib/ippool.h"
-#include "../lib/syserr.h"
+#include "ippool.h"
+#include "syserr.h"
+#include "icmpv6.h"
 #include "config.h"
 
 /* 29.061 11.2.1.3.4 IPv6 Router Configuration Variables in GGSN */
@@ -35,78 +36,68 @@
 #define GGSN_AdvValidLifetime	0xffffffff	/* infinite */
 #define GGSN_AdvPreferredLifetime 0xffffffff	/* infinite */
 
-struct icmpv6_hdr {
-	uint8_t type;
-	uint8_t code;
-	uint16_t csum;
-} __attribute__ ((packed));
+/* RFC3307 link-local scope multicast address */
+const struct in6_addr all_router_mcast_addr = {
+	.s6_addr = { 0xff,0x02,0,0,  0,0,0,0, 0,0,0,0,  0,0,0,2 }
+};
 
-/* RFC4861 Section 4.2 */
-struct icmpv6_radv_hdr {
-	struct icmpv6_hdr hdr;
-	uint8_t cur_ho_limit;
-#if BYTE_ORDER == LITTLE_ENDIAN
-	uint8_t res:6,
-		m:1,
-		o:1;
-#elif BYTE_ORDER == BIG_ENDIAN
-	uint8_t m:1,
-		o:1,
-		res:6;
-#else
-# error	"Please fix <bits/endian.h>"
-#endif
-	uint16_t router_lifetime;
-	uint32_t reachable_time;
-	uint32_t retrans_timer;
-	uint8_t options[0];
-} __attribute__ ((packed));
+/* Prepends the ipv6 header and returns checksum content */
+uint16_t icmpv6_prepend_ip6hdr(struct msgb *msg, const struct in6_addr *saddr,
+				  const struct in6_addr *daddr)
+{
+	uint32_t len;
+	uint16_t skb_csum;
+	struct ip6_hdr *i6h;
 
-/* RFC4861 Section 4.6 */
-struct icmpv6_opt_hdr {
-	uint8_t type;
-	/* length in units of 8 octets, including type+len! */
-	uint8_t len;
-	uint8_t data[0];
-} __attribute__ ((packed));
+	/* checksum */
+	skb_csum = csum_partial(msgb_data(msg), msgb_length(msg), 0);
+	len = msgb_length(msg);
+	skb_csum = csum_ipv6_magic(saddr, daddr, len, IPPROTO_ICMPV6, skb_csum);
 
-/* RFC4861 Section 4.6.2 */
-struct icmpv6_opt_prefix {
-	struct icmpv6_opt_hdr hdr;
-	uint8_t prefix_len;
-#if BYTE_ORDER == LITTLE_ENDIAN
-	uint8_t res:6,
-		a:1,
-		l:1;
-#elif BYTE_ORDER == BIG_ENDIAN
-	uint8_t l:1,
-		a:1,
-		res:6;
-#else
-# error	"Please fix <bits/endian.h>"
-#endif
-	uint32_t valid_lifetime;
-	uint32_t preferred_lifetime;
-	uint32_t res2;
-	uint8_t prefix[16];
-} __attribute__ ((packed));
+	/* Push IPv6 header in front of ICMPv6 packet */
+	i6h = (struct ip6_hdr *) msgb_push(msg, sizeof(*i6h));
+	/* 4 bits version, 8 bits TC, 20 bits flow-ID */
+	i6h->ip6_ctlun.ip6_un1.ip6_un1_flow = htonl(0x60000000);
+	i6h->ip6_ctlun.ip6_un1.ip6_un1_plen = htons(len);
+	i6h->ip6_ctlun.ip6_un1.ip6_un1_nxt = IPPROTO_ICMPV6;
+	i6h->ip6_ctlun.ip6_un1.ip6_un1_hlim = 255;
+	i6h->ip6_src = *saddr;
+	i6h->ip6_dst = *daddr;
+	return skb_csum;
+}
 
+/*! construct a RFC4861 compliant ICMPv6 router soliciation
+ *  \param[in] saddr Source IPv6 address for router advertisement
+ *  \param[in] daddr Destination IPv6 address for router advertisement IPv6 header
+ *  \param[in] prefix The single prefix to be advertised (/64 implied!)
+ *  \returns callee-allocated message buffer containing router advertisement */
+struct msgb *icmpv6_construct_rs(const struct in6_addr *saddr)
+{
+	struct msgb *msg = msgb_alloc_headroom(512,128, "IPv6 RS");
+	struct icmpv6_rsol_hdr *rs;
+	OSMO_ASSERT(msg);
+	rs = (struct icmpv6_rsol_hdr *) msgb_put(msg, sizeof(*rs));
+	rs->hdr.type = 133;	/* see RFC4861 4.1 */
+	rs->hdr.code = 0;	/* see RFC4861 4.1 */
+	rs->hdr.csum = 0;	/* updated below */
+	rs->reserved = 0;	/* see RFC4861 4.1 */
 
+	rs->hdr.csum = icmpv6_prepend_ip6hdr(msg, saddr, &all_router_mcast_addr);
+
+	return msg;
+}
 /*! construct a 3GPP 29.061 compliant router advertisement for a given prefix
  *  \param[in] saddr Source IPv6 address for router advertisement
  *  \param[in] daddr Destination IPv6 address for router advertisement IPv6 header
- *  \param[in] prefix The single prefix to be advertised (/64 implied!)i
+ *  \param[in] prefix The single prefix to be advertised (/64 implied!)
  *  \returns callee-allocated message buffer containing router advertisement */
-struct msgb *icmpv6_construct_ra(const struct in6_addr *saddr,
+static struct msgb *icmpv6_construct_ra(const struct in6_addr *saddr,
 				 const struct in6_addr *daddr,
 				 const struct in6_addr *prefix)
 {
 	struct msgb *msg = msgb_alloc_headroom(512,128, "IPv6 RA");
 	struct icmpv6_radv_hdr *ra;
 	struct icmpv6_opt_prefix *ra_opt_pref;
-	struct ip6_hdr *i6h;
-	uint32_t len;
-	uint16_t skb_csum;
 
 	OSMO_ASSERT(msg);
 
@@ -144,24 +135,12 @@ struct msgb *icmpv6_construct_ra(const struct in6_addr *saddr,
 	memcpy(ra_opt_pref->prefix, prefix, sizeof(ra_opt_pref->prefix));
 
 	/* checksum */
-	skb_csum = csum_partial(msgb_data(msg), msgb_length(msg), 0);
-	len = msgb_length(msg);
-	ra->hdr.csum = csum_ipv6_magic(saddr, daddr, len, IPPROTO_ICMPV6, skb_csum);
-
-	/* Push IPv6 header in front of ICMPv6 packet */
-	i6h = (struct ip6_hdr *) msgb_push(msg, sizeof(*i6h));
-	/* 4 bits version, 8 bits TC, 20 bits flow-ID */
-	i6h->ip6_ctlun.ip6_un1.ip6_un1_flow = htonl(0x60000000);
-	i6h->ip6_ctlun.ip6_un1.ip6_un1_plen = htons(len);
-	i6h->ip6_ctlun.ip6_un1.ip6_un1_nxt = IPPROTO_ICMPV6;
-	i6h->ip6_ctlun.ip6_un1.ip6_un1_hlim = 255;
-	i6h->ip6_src = *saddr;
-	i6h->ip6_dst = *daddr;
+	ra->hdr.csum = icmpv6_prepend_ip6hdr(msg, saddr, daddr);
 
 	return msg;
 }
 
-/* Walidate an ICMPv6 router solicitation according to RFC4861 6.1.1 */
+/* Validate an ICMPv6 router solicitation according to RFC4861 6.1.1 */
 static bool icmpv6_validate_router_solicit(const uint8_t *pack, unsigned len)
 {
 	const struct ip6_hdr *ip6h = (struct ip6_hdr *)pack;
@@ -179,18 +158,47 @@ static bool icmpv6_validate_router_solicit(const uint8_t *pack, unsigned len)
 	return true;
 }
 
+/* Validate an ICMPv6 router advertisement according to RFC4861 6.1.2.
+   Returns pointer packet header on success, NULL otherwise. */
+struct icmpv6_radv_hdr *icmpv6_validate_router_adv(const uint8_t *pack, unsigned len)
+{
+	const struct ip6_hdr *ip6h = (struct ip6_hdr *)pack;
+	const struct icmpv6_hdr *ic6h = (struct icmpv6_hdr *) (pack + sizeof(*ip6h));
+
+	/* ICMP length (derived from IP length) is 16 or more octets */
+	if (len < sizeof(*ip6h) + 16)
+		return NULL;
+
+	if (ic6h->type != 134) /* router advertismenet type */
+		return NULL;
+
+	/*Routers must use their link-local address */
+	if (!IN6_IS_ADDR_LINKLOCAL(&ip6h->ip6_src))
+		return NULL;
+	/* Hop limit field must have 255 */
+	if (ip6h->ip6_ctlun.ip6_un1.ip6_un1_hlim != 255)
+		return NULL;
+	/* ICMP Code is 0 */
+	if (ic6h->code != 0)
+		return NULL;
+	/* ICMP length (derived from IP length) is 16 or more octets */
+	if (ip6h->ip6_ctlun.ip6_un1.ip6_un1_plen < 16)
+		return NULL;
+	/* FIXME: All included options have a length > 0 */
+	/* FIXME: If IP source is unspecified, no source link-layer addr option */
+	return (struct icmpv6_radv_hdr *)ic6h;
+}
+
 /* handle incoming packets to the all-routers multicast address */
-int handle_router_mcast(struct gsn_t *gsn, struct pdp_t *pdp, const struct in6_addr *own_ll_addr,
+int handle_router_mcast(struct gsn_t *gsn, struct pdp_t *pdp,
+			const struct in6_addr *pdp_prefix,
+			const struct in6_addr *own_ll_addr,
 			const uint8_t *pack, unsigned len)
 {
-	struct ippoolm_t *member = pdp->peer;
 	const struct ip6_hdr *ip6h = (struct ip6_hdr *)pack;
 	const struct icmpv6_hdr *ic6h = (struct icmpv6_hdr *) (pack + sizeof(*ip6h));
 	struct msgb *msg;
 
-	OSMO_ASSERT(pdp);
-	OSMO_ASSERT(member);
-
 	if (len < sizeof(*ip6h)) {
 		LOGP(DICMP6, LOGL_NOTICE, "Packet too short: %u bytes\n", len);
 		return -1;
@@ -221,7 +229,7 @@ int handle_router_mcast(struct gsn_t *gsn, struct pdp_t *pdp, const struct in6_a
 		/* Send router advertisement from GGSN link-local
 		 * address to MS link-local address, including prefix
 		 * allocated to this PDP context */
-		msg = icmpv6_construct_ra(own_ll_addr, &ip6h->ip6_src, &member->addr.v6);
+		msg = icmpv6_construct_ra(own_ll_addr, &ip6h->ip6_src, pdp_prefix);
 		/* Send the constructed RA to the MS */
 		gtp_data_req(gsn, pdp, msgb_data(msg), msgb_length(msg));
 		msgb_free(msg);

lib/icmpv6.h

@@ -0,0 +1,98 @@
+#pragma once
+
+#include <stdbool.h>
+
+#include <osmocom/core/msgb.h>
+#include <osmocom/core/endian.h>
+
+#include "../gtp/gtp.h"
+#include "../gtp/pdp.h"
+
+#define ICMPv6_OPT_TYPE_PREFIX_INFO 0x03
+
+#define foreach_icmpv6_opt(icmpv6_pkt, icmpv6_len, opt_hdr) \
+		for (opt_hdr = (struct icmpv6_opt_hdr *)(icmpv6_pkt)->options; \
+		     (uint8_t*)(opt_hdr) + sizeof(struct icmpv6_opt_hdr) <= (((uint8_t*)(icmpv6_pkt)) + (icmpv6_len)); \
+		     opt_hdr = (struct icmpv6_opt_hdr*)((uint8_t*)(opt_hdr) + (opt_hdr)->len) \
+		    )
+
+struct icmpv6_hdr {
+	uint8_t type;
+	uint8_t code;
+	uint16_t csum;
+} __attribute__ ((packed));
+
+struct icmpv6_echo_hdr {
+	struct icmpv6_hdr hdr;
+	uint16_t ident;		/* Identifier */
+	uint16_t seq;		/* Sequence number */
+	uint8_t data[0];	/* Data */
+} __attribute__ ((packed));
+
+/* RFC4861 Section 4.1 */
+struct icmpv6_rsol_hdr {
+	struct icmpv6_hdr hdr;
+	uint32_t reserved;
+	uint8_t options[0];
+} __attribute__ ((packed));
+
+/* RFC4861 Section 4.2 */
+struct icmpv6_radv_hdr {
+	struct icmpv6_hdr hdr;
+	uint8_t cur_ho_limit;
+#if OSMO_IS_LITTLE_ENDIAN
+	uint8_t res:6,
+		m:1,
+		o:1;
+#elif OSMO_IS_BIG_ENDIAN
+/* auto-generated from the little endian part above (libosmocore/contrib/struct_endianness.py) */
+	uint8_t o:1, m:1, res:6;
+#endif
+	uint16_t router_lifetime;
+	uint32_t reachable_time;
+	uint32_t retrans_timer;
+	uint8_t options[0];
+} __attribute__ ((packed));
+
+
+/* RFC4861 Section 4.6 */
+struct icmpv6_opt_hdr {
+	uint8_t type;
+	/* length in units of 8 octets, including type+len! */
+	uint8_t len;
+	uint8_t data[0];
+} __attribute__ ((packed));
+
+/* RFC4861 Section 4.6.2 */
+struct icmpv6_opt_prefix {
+	struct icmpv6_opt_hdr hdr;
+	uint8_t prefix_len;
+#if OSMO_IS_LITTLE_ENDIAN
+	uint8_t res:6,
+		a:1,
+		l:1;
+#elif OSMO_IS_BIG_ENDIAN
+/* auto-generated from the little endian part above (libosmocore/contrib/struct_endianness.py) */
+	uint8_t l:1, a:1, res:6;
+#endif
+	uint32_t valid_lifetime;
+	uint32_t preferred_lifetime;
+	uint32_t res2;
+	uint8_t prefix[16];
+} __attribute__ ((packed));
+
+uint16_t icmpv6_prepend_ip6hdr(struct msgb *msg, const struct in6_addr *saddr,
+				  const struct in6_addr *daddr);
+
+struct msgb *icmpv6_construct_rs(const struct in6_addr *saddr);
+
+int handle_router_mcast(struct gsn_t *gsn, struct pdp_t *pdp,
+			const struct in6_addr *pdp_prefix,
+			const struct in6_addr *own_ll_addr,
+			const uint8_t *pack, unsigned len);
+
+struct icmpv6_radv_hdr *icmpv6_validate_router_adv(const uint8_t *pack, unsigned len);
+
+
+/* RFC3307 link-local scope multicast address */
+extern const struct in6_addr all_router_mcast_addr;

lib/in46_addr.c

@@ -60,7 +60,11 @@ int in46a_to_sas(struct sockaddr_storage *out, const struct in46_addr *in)
 	return 0;
 }
 
-/*! Convenience wrapper around inet_ntop() for \ref in46_addr */
+/*! Convenience wrapper around inet_ntop() for in46_addr.
+ *  \param[in] in the in46_addr to print
+ *  \param[out] dst destination buffer where string representation of the address is stored
+ *  \param[out] dst_size size dst. Usually it should be at least INET6_ADDRSTRLEN.
+ *  \return address of dst on success, NULL on error */
 const char *in46a_ntop(const struct in46_addr *in, char *dst, socklen_t dst_size)
 {
 	int af;
@@ -253,33 +257,66 @@ unsigned int in46a_netmasklen(const struct in46_addr *netmask)
 	}
 }
 
-/*! Convert given PDP End User Address to in46_addr
- *  \returns 0 on success; negative on error */
-int in46a_to_eua(const struct in46_addr *src, struct ul66_t *eua)
+/*! Convert given array of in46_addr to PDP End User Address
+ *  \param[in] src Array containing 1 or 2 in46_addr
+ *  \param[out] eua End User Address structure to fill
+ *  \returns 0 on success; negative on error
+ *
+ * In case size is 2, this function expects to find exactly one IPv4 and one
+ * IPv6 addresses in src. */
+int in46a_to_eua(const struct in46_addr *src, unsigned int size, struct ul66_t *eua)
 {
-	switch (src->len) {
-	case 4:
-		eua->l = 6;
-		eua->v[0] = PDP_EUA_ORG_IETF;
-		eua->v[1] = PDP_EUA_TYPE_v4;
-		memcpy(&eua->v[2], &src->v4, 4);	/* Copy a 4 byte address */
-		break;
-	case 8:
-	case 16:
-		eua->l = 18;
-		eua->v[0] = PDP_EUA_ORG_IETF;
-		eua->v[1] = PDP_EUA_TYPE_v6;
-		memcpy(&eua->v[2], &src->v6, 16);	/* Copy a 16 byte address */
-		break;
-	default:
-		OSMO_ASSERT(0);
-		return -1;
+	const struct in46_addr *src_v4, *src_v6;
+	if (size == 1) {
+		switch (src->len) {
+		case 4:
+			eua->l = 6;
+			eua->v[0] = PDP_EUA_ORG_IETF;
+			eua->v[1] = PDP_EUA_TYPE_v4;
+			memcpy(&eua->v[2], &src->v4, 4);	/* Copy a 4 byte address */
+			break;
+		case 8:
+		case 16:
+			eua->l = 18;
+			eua->v[0] = PDP_EUA_ORG_IETF;
+			eua->v[1] = PDP_EUA_TYPE_v6;
+			memcpy(&eua->v[2], &src->v6, 16);	/* Copy a 16 byte address */
+			break;
+		default:
+			OSMO_ASSERT(0);
+			return -1;
+		}
+		return 0;
 	}
+
+	if (src[0].len == src[1].len)
+		return -1; /* we should have a v4 and a v6 address */
+
+	src_v4 = (src[0].len == 4) ? &src[0] : &src[1];
+	src_v6 = (src[0].len == 4) ? &src[1] : &src[0];
+
+	eua->l = 22;
+	eua->v[0] = PDP_EUA_ORG_IETF;
+	eua->v[1] = PDP_EUA_TYPE_v4v6;
+	memcpy(&eua->v[2], &src_v4->v4, 4);
+	memcpy(&eua->v[6], &src_v6->v6, 16);
+
 	return 0;
 }
 
-/*! Convert given in46_addr to PDP End User Address
- *  \returns 0 on success; negative on error */
+/*! Convert given PDP End User Address to an array of in46_addr
+ *  \param[in] eua End User Address structure to parse
+ *  \param[out] dst Array containing 2 in46_addr
+ *  \returns number of parsed addresses (1 or 2) on success; negative on error
+ *
+ * This function expects to receive an End User Address struct together with an
+ * array of 2 zeroed in46_addr structs. The in46_addr structs are filled in
+ * order, hence if the function returns 1 the parsed address will be stored in
+ * the first struct and the second one will be left intact. If 2 is returned, it
+ * is guaranteed that one of them is an IPv4 and the other one is an IPv6, but
+ * the order in which they are presented is not specified and must be
+ * discovered for instance by checking the len field of each address.
+ */
 int in46a_from_eua(const struct ul66_t *eua, struct in46_addr *dst)
 {
 	if (eua->l < 2)
@@ -295,22 +332,53 @@ int in46a_from_eua(const struct ul66_t *eua, struct in46_addr *dst)
 			memcpy(&dst->v4, &eua->v[2], 4);	/* Copy a 4 byte address */
 		else
 			dst->v4.s_addr = 0;
-		break;
+		return 1;
 	case PDP_EUA_TYPE_v6:
 		dst->len = 16;
 		if (eua->l >= 18)
 			memcpy(&dst->v6, &eua->v[2], 16);	/* Copy a 16 byte address */
 		else
 			memset(&dst->v6, 0, 16);
-		break;
+		return 1;
+	case PDP_EUA_TYPE_v4v6:
+		/* 3GPP TS 29.060, section 7.7.27 */
+		switch (eua->l) {
+			case 2: /* v4 & v6 dynamic */
+				dst[0].v4.s_addr = 0;
+				memset(&dst[1].v6, 0, 16);
+				break;
+			case 6: /* v4 static, v6 dynamic */
+				memcpy(&dst[0].v4, &eua->v[2], 4);
+				memset(&dst[1].v6, 0, 16);
+				break;
+			case 18: /* v4 dynamic, v6 static */
+				dst[0].v4.s_addr = 0;
+				memcpy(&dst[1].v6, &eua->v[2], 16);
+				break;
+			case 22:  /* v4 & v6 static */
+				memcpy(&dst[0].v4, &eua->v[2], 4);
+				memcpy(&dst[1].v6, &eua->v[6], 16);
+				break;
+			default:
+				return -1;
+		}
+		dst[0].len = 4;
+		dst[1].len = 16;
+		return 2;
 	default:
 		return -1;
 	}
-	return 0;
 
 default_to_dyn_v4:
 	/* assume dynamic IPv4 by default */
 	dst->len = 4;
 	dst->v4.s_addr = 0;
-	return 0;
+	return 1;
+}
+
+void in46a_from_gsna(const struct ul16_t *in, struct in46_addr *dst)
+{
+	dst->len = in->l;
+	OSMO_ASSERT(in->l <= sizeof(dst->v6));
+	memcpy(&dst->v6, in->v, in->l);
 }

lib/in46_addr.h

@@ -29,5 +29,15 @@ extern int in46a_prefix_equal(const struct in46_addr *a, const struct in46_addr
 extern int in46a_within_mask(const struct in46_addr *addr, const struct in46_addr *net, size_t prefixlen);
 unsigned int in46a_netmasklen(const struct in46_addr *netmask);
 
-int in46a_to_eua(const struct in46_addr *src, struct ul66_t *eua);
+int in46a_to_eua(const struct in46_addr *src, unsigned int size, struct ul66_t *eua);
 int in46a_from_eua(const struct ul66_t *eua, struct in46_addr *dst);
+
+static inline bool in46a_is_v6(const struct in46_addr *addr) {
+	return addr->len == 8 || addr->len == 16;
+}
+
+static inline bool in46a_is_v4(const struct in46_addr *addr) {
+	return addr->len == sizeof(struct in_addr);
+}
+
+void in46a_from_gsna(const struct ul16_t *in, struct in46_addr *dst);

lib/ippool.c

@@ -26,16 +26,16 @@ int ippool_printaddr(struct ippool_t *this)
 {
 	unsigned int n;
 	printf("ippool_printaddr\n");
-	printf("Firstdyn %d\n", this->firstdyn - this->member);
-	printf("Lastdyn %d\n", this->lastdyn - this->member);
-	printf("Firststat %d\n", this->firststat - this->member);
-	printf("Laststat %d\n", this->laststat - this->member);
-	printf("Listsize %d\n", this->listsize);
+	printf("Firstdyn %td\n", this->firstdyn - this->member);
+	printf("Lastdyn %td\n", this->lastdyn - this->member);
+	printf("Firststat %td\n", this->firststat - this->member);
+	printf("Laststat %td\n", this->laststat - this->member);
+	printf("Listsize %u\n", this->listsize);
 
 	for (n = 0; n < this->listsize; n++) {
 		char s[256];
 		in46a_ntop(&this->member[n].addr, s, sizeof(s));
-		printf("Unit %d inuse %d prev %d next %d addr %s\n",
+		printf("Unit %d inuse %d prev %td next %td addr %s\n",
 		       n,
 		       this->member[n].inuse,
 		       this->member[n].prev - this->member,
@@ -202,7 +202,7 @@ int ippool_new(struct ippool_t **this, const struct in46_prefix *dyn, const stru
 	/* Parse only first instance of pool for now */
 
 	int i;
-	struct in46_addr addr;
+	struct in46_addr addr = { 0 };
 	size_t addrprefixlen;
 	struct in46_addr stataddr;
 	size_t stataddrprefixlen;
@@ -240,7 +240,7 @@ int ippool_new(struct ippool_t **this, const struct in46_prefix *dyn, const stru
 		stataddr = stat->addr;
 		stataddrprefixlen = stat->prefixlen;
 
-		statsize = (1 << (addr.len - stataddrprefixlen + 1)) -1;
+		statsize = (1 << (stataddr.len*8 - stataddrprefixlen));
 		if (statsize > IPPOOL_STATSIZE)
 			statsize = IPPOOL_STATSIZE;
 	}
@@ -276,9 +276,8 @@ int ippool_new(struct ippool_t **this, const struct in46_prefix *dyn, const stru
 	(*this)->hashmask = (*this)->hashsize - 1;
 
 	/* Allocate hash table */
-	if (!
-	    ((*this)->hash =
-	     calloc(sizeof(struct ippoolm_t), (*this)->hashsize))) {
+	(*this)->hash = calloc((*this)->hashsize, sizeof(struct ippoolm_t *));
+	if (!(*this)->hash) {
 		SYS_ERR(DIP, LOGL_ERROR, 0,
 			"Failed to allocate memory for hash members in ippool");
 		return -1;
@@ -513,7 +512,15 @@ int ippool_newip(struct ippool_t *this, struct ippoolm_t **member,
 		p2->next = NULL;
 		p2->prev = NULL;
 		p2->inuse = 2;	/* Static address in use */
-		memcpy(&p2->addr, addr, sizeof(addr));
+		/* p2->addr.len and addr->len already match (see above). */
+		if (p2->addr.len == sizeof(struct in_addr))
+			p2->addr.v4 = addr->v4;
+		else if (p2->addr.len == sizeof(struct in6_addr))
+			p2->addr.v6 = addr->v6;
+		else {
+			SYS_ERR(DIP, LOGL_ERROR, 0, "MS requested unsupported PDP context type");
+			return -GTPCAUSE_UNKNOWN_PDP;
+		}
 		*member = p2;
 		(void)ippool_hashadd(this, *member);
 		if (0)

lib/netdev.c

@@ -0,0 +1,792 @@
+/*
+ * TUN interface functions.
+ * Copyright (C) 2002, 2003, 2004 Mondru AB.
+ * Copyright (C) 2017-2018 by Harald Welte <laforge@gnumonks.org>
+ *
+ * The contents of this file may be used under the terms of the GNU
+ * General Public License Version 2, provided that the above copyright
+ * notice and this permission notice is included in all copies or
+ * substantial portions of the software.
+ *
+ */
+
+/*
+ * netdev.c: Contains generic network device related functionality.
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <sys/types.h>
+#include <sys/socket.h>
+#include <netinet/in.h>
+#include <arpa/inet.h>
+#include <sys/stat.h>
+#include <sys/time.h>
+#include <unistd.h>
+#include <string.h>
+#include <errno.h>
+#include <fcntl.h>
+
+#include <stdio.h>
+#include <fcntl.h>
+#include <unistd.h>
+#include <sys/time.h>
+#include <sys/ioctl.h>
+#include <sys/socket.h>
+#include <errno.h>
+#include <net/route.h>
+#include <net/if.h>
+
+#if defined(__linux__)
+#include <linux/netlink.h>
+#include <linux/rtnetlink.h>
+
+#elif defined (__FreeBSD__)
+#include <net/if_var.h>
+#include <netinet/in_var.h>
+
+#elif defined (__APPLE__)
+#include <net/if.h>
+
+#else
+#error  "Unknown platform!"
+#endif
+
+#include "netdev.h"
+#include "syserr.h"
+
+#if defined(__linux__)
+
+#include <linux/ipv6.h>
+
+static int netdev_nlattr(struct nlmsghdr *n, int nsize, int type, void *d, int dlen)
+{
+	int len = RTA_LENGTH(dlen);
+	int alen = NLMSG_ALIGN(n->nlmsg_len);
+	struct rtattr *rta = (struct rtattr *)(((void *)n) + alen);
+	if (alen + len > nsize)
+		return -1;
+	rta->rta_len = len;
+	rta->rta_type = type;
+	memcpy(RTA_DATA(rta), d, dlen);
+	n->nlmsg_len = alen + len;
+	return 0;
+}
+#endif
+
+static int netdev_sifflags(const char *devname, int flags)
+{
+	struct ifreq ifr;
+	int fd;
+
+	memset(&ifr, '\0', sizeof(ifr));
+	ifr.ifr_flags = flags;
+	strncpy(ifr.ifr_name, devname, IFNAMSIZ);
+	ifr.ifr_name[IFNAMSIZ - 1] = 0;	/* Make sure to terminate */
+	if ((fd = socket(AF_INET, SOCK_DGRAM, 0)) < 0) {
+		SYS_ERR(DTUN, LOGL_ERROR, errno, "socket() failed");
+		return -1;
+	}
+	if (ioctl(fd, SIOCSIFFLAGS, &ifr)) {
+		SYS_ERR(DTUN, LOGL_ERROR, errno,
+			"ioctl(SIOCSIFFLAGS) failed");
+		close(fd);
+		return -1;
+	}
+	close(fd);
+	return 0;
+}
+
+int netdev_setaddr4(const char *devname, struct in_addr *addr,
+		    struct in_addr *dstaddr, struct in_addr *netmask)
+{
+	struct ifreq ifr;
+	int fd;
+
+	memset(&ifr, '\0', sizeof(ifr));
+	ifr.ifr_addr.sa_family = AF_INET;
+	ifr.ifr_dstaddr.sa_family = AF_INET;
+
+#if defined(__linux__)
+	ifr.ifr_netmask.sa_family = AF_INET;
+#elif defined(__FreeBSD__) || defined (__APPLE__)
+	((struct sockaddr_in *)&ifr.ifr_addr)->sin_len =
+	    sizeof(struct sockaddr_in);
+	((struct sockaddr_in *)&ifr.ifr_dstaddr)->sin_len =
+	    sizeof(struct sockaddr_in);
+#endif
+
+	strncpy(ifr.ifr_name, devname, IFNAMSIZ);
+	ifr.ifr_name[IFNAMSIZ - 1] = 0;	/* Make sure to terminate */
+
+	/* Create a channel to the NET kernel. */
+	if ((fd = socket(AF_INET, SOCK_DGRAM, 0)) < 0) {
+		SYS_ERR(DTUN, LOGL_ERROR, errno, "socket() failed");
+		return -1;
+	}
+
+	if (addr) {		/* Set the interface address */
+		memcpy(&((struct sockaddr_in *)&ifr.ifr_addr)->sin_addr, addr,
+		       sizeof(*addr));
+		if (ioctl(fd, SIOCSIFADDR, (void *)&ifr) < 0) {
+			if (errno != EEXIST) {
+				SYS_ERR(DTUN, LOGL_ERROR, errno,
+					"ioctl(SIOCSIFADDR) failed");
+			} else {
+				SYS_ERR(DTUN, LOGL_NOTICE, errno,
+					"ioctl(SIOCSIFADDR): Address already exists");
+			}
+			close(fd);
+			return -1;
+		}
+	}
+
+	if (dstaddr) {		/* Set the destination address */
+		memcpy(&((struct sockaddr_in *)&ifr.ifr_dstaddr)->sin_addr,
+		       dstaddr, sizeof(*dstaddr));
+		if (ioctl(fd, SIOCSIFDSTADDR, (caddr_t) & ifr) < 0) {
+			SYS_ERR(DTUN, LOGL_ERROR, errno,
+				"ioctl(SIOCSIFDSTADDR) failed");
+			close(fd);
+			return -1;
+		}
+	}
+
+	if (netmask) {		/* Set the netmask */
+#if defined(__linux__)
+		memcpy(&((struct sockaddr_in *)&ifr.ifr_netmask)->sin_addr,
+		       netmask, sizeof(*netmask));
+#elif defined(__FreeBSD__) || defined (__APPLE__)
+		((struct sockaddr_in *)&ifr.ifr_addr)->sin_addr.s_addr =
+		    netmask->s_addr;
+#endif
+
+		if (ioctl(fd, SIOCSIFNETMASK, (void *)&ifr) < 0) {
+			SYS_ERR(DTUN, LOGL_ERROR, errno,
+				"ioctl(SIOCSIFNETMASK) failed");
+			close(fd);
+			return -1;
+		}
+	}
+
+	close(fd);
+
+	netdev_sifflags(devname, IFF_UP | IFF_RUNNING);
+
+	/* On linux the route to the interface is set automatically
+	   on FreeBSD we have to do this manually */
+#if defined(__FreeBSD__) || defined (__APPLE__)
+	netdev_addroute4(dstaddr, addr, &this->netmask);
+#endif
+
+	return 0;
+}
+
+int netdev_setaddr6(const char *devname, struct in6_addr *addr, struct in6_addr *dstaddr,
+		    size_t prefixlen)
+{
+	struct in6_ifreq ifr;
+	int fd;
+
+	memset(&ifr, 0, sizeof(ifr));
+
+#if defined(__linux__)
+	ifr.ifr6_prefixlen = prefixlen;
+	ifr.ifr6_ifindex = if_nametoindex(devname);
+	if (ifr.ifr6_ifindex == 0) {
+		SYS_ERR(DTUN, LOGL_ERROR, 0, "Error getting ifindex for %s\n", devname);
+		return -1;
+	}
+#elif defined(__FreeBSD__) || defined (__APPLE__)
+	strncpy(ifr.ifr_name, devname, IFNAMSIZ);
+#endif
+
+	/* Create a channel to the NET kernel */
+	if ((fd = socket(AF_INET6, SOCK_DGRAM, 0)) < 0) {
+		SYS_ERR(DTUN, LOGL_ERROR, 0, "socket() failed");
+		return -1;
+	}
+
+#if defined(__linux__)
+	if (addr) {
+		memcpy(&ifr.ifr6_addr, addr, sizeof(*addr));
+		if (ioctl(fd, SIOCSIFADDR, (void *) &ifr) < 0) {
+			if (errno != EEXIST) {
+				SYS_ERR(DTUN, LOGL_ERROR, 0, "ioctl(SIOCSIFADDR) failed");
+			} else {
+				SYS_ERR(DTUN, LOGL_NOTICE, 0, "ioctl(SIOCSIFADDR): Address already exists");
+			}
+			close(fd);
+			return -1;
+		}
+	}
+
+#if 0
+	/* FIXME: looks like this is not possible/necessary for IPv6? */
+	if (dstaddr) {
+		memcpy(&ifr.ifr6_addr, dstaddr, sizeof(*dstaddr));
+		if (ioctl(fd, SIOCSIFDSTADDR, (caddr_t *) &ifr) < 0) {
+			SYS_ERR(DTUN, LOGL_ERROR, "ioctl(SIOCSIFDSTADDR) failed");
+			close(fd);
+			return -1;
+		}
+	}
+#endif
+
+#elif defined(__FreeBSD__) || defined (__APPLE__)
+	if (addr)
+		memcpy(&ifr.ifr_ifru.ifru_addr, addr, sizeof(ifr.ifr_ifru.ifru_addr));
+	if (dstaddr)
+		memcpy(&ifr.ifr_ifru.ifru_dstaddr, dstaddr, sizeof(ifr.ifr_ifru.ifru_dstaddr));
+
+	if (ioctl(fd, SIOCSIFADDR_IN6, (struct ifreq *)&ifr) < 0) {
+		SYS_ERR(DTUN, LOGL_ERROR, 0, "ioctl(SIOCSIFADDR_IN6) failed");
+		close(fd);
+		return -1;
+	}
+#endif
+
+	close(fd);
+
+	netdev_sifflags(devname, IFF_UP | IFF_RUNNING);
+
+	/* On linux the route to the interface is set automatically
+	   on FreeBSD we have to do this manually */
+#if 0	/* FIXME */
+//#if defined(__FreeBSD__) || defined (__APPLE__)
+	netdev_addroute6(dstaddr, addr, prefixlen);
+#endif
+
+	return 0;
+}
+
+int netdev_addaddr4(const char *devname, struct in_addr *addr,
+		    struct in_addr *dstaddr, struct in_addr *netmask)
+{
+	int fd;
+#if defined(__linux__)
+	struct {
+		struct nlmsghdr n;
+		struct ifaddrmsg i;
+		char buf[TUN_NLBUFSIZE];
+	} req;
+
+	struct sockaddr_nl local;
+	socklen_t addr_len;
+	int status;
+
+	struct sockaddr_nl nladdr;
+	struct iovec iov;
+	struct msghdr msg;
+
+	memset(&req, 0, sizeof(req));
+	req.n.nlmsg_len = NLMSG_LENGTH(sizeof(struct ifaddrmsg));
+	req.n.nlmsg_flags = NLM_F_REQUEST | NLM_F_CREATE;
+	req.n.nlmsg_type = RTM_NEWADDR;
+	req.i.ifa_family = AF_INET;
+	req.i.ifa_prefixlen = 32;	/* 32 FOR IPv4 */
+	req.i.ifa_flags = 0;
+	req.i.ifa_scope = RT_SCOPE_HOST;	/* TODO or 0 */
+	req.i.ifa_index = if_nametoindex(devname);
+	if (!req.i.ifa_index) {
+		SYS_ERR(DTUN, LOGL_ERROR, errno, "Unable to get ifindex for %s", devname);
+		return -1;
+	}
+
+	netdev_nlattr(&req.n, sizeof(req), IFA_ADDRESS, addr, sizeof(*addr));
+	if (dstaddr)
+		netdev_nlattr(&req.n, sizeof(req), IFA_LOCAL, dstaddr, sizeof(*dstaddr));
+
+	if ((fd = socket(AF_NETLINK, SOCK_RAW, NETLINK_ROUTE)) < 0) {
+		SYS_ERR(DTUN, LOGL_ERROR, errno, "socket() failed");
+		return -1;
+	}
+
+	memset(&local, 0, sizeof(local));
+	local.nl_family = AF_NETLINK;
+	local.nl_groups = 0;
+
+	if (bind(fd, (struct sockaddr *)&local, sizeof(local)) < 0) {
+		SYS_ERR(DTUN, LOGL_ERROR, errno, "bind() failed");
+		close(fd);
+		return -1;
+	}
+
+	addr_len = sizeof(local);
+	if (getsockname(fd, (struct sockaddr *)&local, &addr_len) < 0) {
+		SYS_ERR(DTUN, LOGL_ERROR, errno,
+			"getsockname() failed");
+		close(fd);
+		return -1;
+	}
+
+	if (addr_len != sizeof(local)) {
+		SYS_ERR(DTUN, LOGL_ERROR, 0,
+			"Wrong address length %d", addr_len);
+		close(fd);
+		return -1;
+	}
+
+	if (local.nl_family != AF_NETLINK) {
+		SYS_ERR(DTUN, LOGL_ERROR, 0,
+			"Wrong address family %d", local.nl_family);
+		close(fd);
+		return -1;
+	}
+
+	iov.iov_base = (void *)&req.n;
+	iov.iov_len = req.n.nlmsg_len;
+
+	msg.msg_name = (void *)&nladdr;
+	msg.msg_namelen = sizeof(nladdr);
+	msg.msg_iov = &iov;
+	msg.msg_iovlen = 1;
+	msg.msg_control = NULL;
+	msg.msg_controllen = 0;
+	msg.msg_flags = 0;
+
+	memset(&nladdr, 0, sizeof(nladdr));
+	nladdr.nl_family = AF_NETLINK;
+	nladdr.nl_pid = 0;
+	nladdr.nl_groups = 0;
+
+	req.n.nlmsg_seq = 0;
+	req.n.nlmsg_flags |= NLM_F_ACK;
+
+	status = sendmsg(fd, &msg, 0);
+	if (status != req.n.nlmsg_len) {
+		SYS_ERR(DTUN, LOGL_ERROR, errno, "sendmsg() failed, returned %d", status);
+		close(fd);
+		return -1;
+	}
+
+	status = netdev_sifflags(devname, IFF_UP | IFF_RUNNING);
+	if (status == -1) {
+		close(fd);
+		return -1;
+	}
+#elif defined (__FreeBSD__) || defined (__APPLE__)
+	struct ifaliasreq areq;
+
+	memset(&areq, 0, sizeof(areq));
+
+	/* Set up interface name */
+	strncpy(areq.ifra_name, devname, IFNAMSIZ);
+	areq.ifra_name[IFNAMSIZ - 1] = 0;	/* Make sure to terminate */
+
+	((struct sockaddr_in *)&areq.ifra_addr)->sin_family = AF_INET;
+	((struct sockaddr_in *)&areq.ifra_addr)->sin_len =
+	    sizeof(areq.ifra_addr);
+	((struct sockaddr_in *)&areq.ifra_addr)->sin_addr.s_addr = addr->s_addr;
+
+	((struct sockaddr_in *)&areq.ifra_mask)->sin_family = AF_INET;
+	((struct sockaddr_in *)&areq.ifra_mask)->sin_len =
+	    sizeof(areq.ifra_mask);
+	((struct sockaddr_in *)&areq.ifra_mask)->sin_addr.s_addr =
+	    netmask->s_addr;
+
+	/* For some reason FreeBSD uses ifra_broadcast for specifying dstaddr */
+	((struct sockaddr_in *)&areq.ifra_broadaddr)->sin_family = AF_INET;
+	((struct sockaddr_in *)&areq.ifra_broadaddr)->sin_len =
+	    sizeof(areq.ifra_broadaddr);
+	((struct sockaddr_in *)&areq.ifra_broadaddr)->sin_addr.s_addr =
+	    dstaddr->s_addr;
+
+	/* Create a channel to the NET kernel. */
+	if ((fd = socket(AF_INET, SOCK_DGRAM, 0)) < 0) {
+		SYS_ERR(DTUN, LOGL_ERROR, errno, "socket() failed");
+		return -1;
+	}
+
+	if (ioctl(fd, SIOCAIFADDR, (void *)&areq) < 0) {
+		SYS_ERR(DTUN, LOGL_ERROR, errno,
+			"ioctl(SIOCAIFADDR) failed");
+		close(fd);
+		return -1;
+	}
+#endif
+	close(fd);
+	return 0;
+}
+
+int netdev_addaddr6(const char *devname, struct in6_addr *addr,
+		    struct in6_addr *dstaddr, int prefixlen)
+{
+	int fd;
+#if defined(__linux__)
+	struct {
+		struct nlmsghdr n;
+		struct ifaddrmsg i;
+		char buf[TUN_NLBUFSIZE];
+	} req;
+
+	struct sockaddr_nl local;
+	socklen_t addr_len;
+	int status;
+
+	struct sockaddr_nl nladdr;
+	struct iovec iov;
+	struct msghdr msg;
+
+	memset(&req, 0, sizeof(req));
+	req.n.nlmsg_len = NLMSG_LENGTH(sizeof(struct ifaddrmsg));
+	req.n.nlmsg_flags = NLM_F_REQUEST | NLM_F_CREATE;
+	req.n.nlmsg_type = RTM_NEWADDR;
+	req.i.ifa_family = AF_INET6;
+	req.i.ifa_prefixlen = prefixlen;	/* 64 FOR IPv6 */
+	req.i.ifa_flags = 0;
+	req.i.ifa_scope = RT_SCOPE_HOST;	/* TODO or 0 */
+	req.i.ifa_index = if_nametoindex(devname);
+	if (!req.i.ifa_index) {
+		SYS_ERR(DTUN, LOGL_ERROR, errno, "Unable to get ifindex for %s", devname);
+		return -1;
+	}
+
+	netdev_nlattr(&req.n, sizeof(req), IFA_ADDRESS, addr, sizeof(*addr));
+	if (dstaddr)
+		netdev_nlattr(&req.n, sizeof(req), IFA_LOCAL, dstaddr, sizeof(*dstaddr));
+
+	if ((fd = socket(AF_NETLINK, SOCK_RAW, NETLINK_ROUTE)) < 0) {
+		SYS_ERR(DTUN, LOGL_ERROR, errno, "socket() failed");
+		return -1;
+	}
+
+	memset(&local, 0, sizeof(local));
+	local.nl_family = AF_NETLINK;
+	local.nl_groups = 0;
+
+	if (bind(fd, (struct sockaddr *)&local, sizeof(local)) < 0) {
+		SYS_ERR(DTUN, LOGL_ERROR, errno, "bind() failed");
+		close(fd);
+		return -1;
+	}
+
+	addr_len = sizeof(local);
+	if (getsockname(fd, (struct sockaddr *)&local, &addr_len) < 0) {
+		SYS_ERR(DTUN, LOGL_ERROR, errno,
+			"getsockname() failed");
+		close(fd);
+		return -1;
+	}
+
+	if (addr_len != sizeof(local)) {
+		SYS_ERR(DTUN, LOGL_ERROR, 0,
+			"Wrong address length %d", addr_len);
+		close(fd);
+		return -1;
+	}
+
+	if (local.nl_family != AF_NETLINK) {
+		SYS_ERR(DTUN, LOGL_ERROR, 0,
+			"Wrong address family %d", local.nl_family);
+		close(fd);
+		return -1;
+	}
+
+	iov.iov_base = (void *)&req.n;
+	iov.iov_len = req.n.nlmsg_len;
+
+	msg.msg_name = (void *)&nladdr;
+	msg.msg_namelen = sizeof(nladdr);
+	msg.msg_iov = &iov;
+	msg.msg_iovlen = 1;
+	msg.msg_control = NULL;
+	msg.msg_controllen = 0;
+	msg.msg_flags = 0;
+
+	memset(&nladdr, 0, sizeof(nladdr));
+	nladdr.nl_family = AF_NETLINK;
+	nladdr.nl_pid = 0;
+	nladdr.nl_groups = 0;
+
+	req.n.nlmsg_seq = 0;
+	req.n.nlmsg_flags |= NLM_F_ACK;
+
+	status = sendmsg(fd, &msg, 0);
+	if (status != req.n.nlmsg_len) {
+		SYS_ERR(DTUN, LOGL_ERROR, errno, "sendmsg() failed, returned %d", status);
+		close(fd);
+		return -1;
+	}
+
+	status = netdev_sifflags(devname, IFF_UP | IFF_RUNNING);
+	if (status == -1) {
+		close(fd);
+		return -1;
+	}
+#elif defined (__FreeBSD__) || defined (__APPLE__)
+	struct ifaliasreq areq;
+
+	memset(&areq, 0, sizeof(areq));
+
+	/* Set up interface name */
+	strncpy(areq.ifra_name, devname, IFNAMSIZ);
+	areq.ifra_name[IFNAMSIZ - 1] = 0;	/* Make sure to terminate */
+
+	((struct sockaddr_in6 *)&areq.ifra_addr)->sin6_family = AF_INET6;
+	((struct sockaddr_in6 *)&areq.ifra_addr)->sin6_len = sizeof(areq.ifra_addr);
+	((struct sockaddr_in6 *)&areq.ifra_addr)->sin6_addr.s6_addr = addr->s6_addr;
+
+	((struct sockaddr_in6 *)&areq.ifra_mask)->sin6_family = AF_INET6;
+	((struct sockaddr_in6 *)&areq.ifra_mask)->sin6_len = sizeof(areq.ifra_mask);
+	((struct sockaddr_in6 *)&areq.ifra_mask)->sin6_addr.s6_addr = netmask->s6_addr;
+
+	/* For some reason FreeBSD uses ifra_broadcast for specifying dstaddr */
+	((struct sockaddr_in6 *)&areq.ifra_broadaddr)->sin6_family = AF_INET6;
+	((struct sockaddr_in6 *)&areq.ifra_broadaddr)->sin6_len = sizeof(areq.ifra_broadaddr);
+	((struct sockaddr_in6 *)&areq.ifra_broadaddr)->sin6_addr.s6_addr = dstaddr->s6_addr;
+
+	/* Create a channel to the NET kernel. */
+	if ((fd = socket(AF_INET, SOCK_DGRAM, 0)) < 0) {
+		SYS_ERR(DTUN, LOGL_ERROR, errno, "socket() failed");
+		return -1;
+	}
+
+	if (ioctl(fd, SIOCAIFADDR, (void *)&areq) < 0) {
+		SYS_ERR(DTUN, LOGL_ERROR, errno,
+			"ioctl(SIOCAIFADDR) failed");
+		close(fd);
+		return -1;
+	}
+#endif
+	close(fd);
+	return 0;
+}
+
+static int netdev_route4(struct in_addr *dst, struct in_addr *gateway, struct in_addr *mask, int delete)
+{
+	int fd;
+#if defined(__linux__)
+	struct rtentry r;
+
+	memset(&r, '\0', sizeof(r));
+	r.rt_flags = RTF_UP | RTF_GATEWAY;	/* RTF_HOST not set */
+
+	/* Create a channel to the NET kernel. */
+	if ((fd = socket(AF_INET, SOCK_DGRAM, 0)) < 0) {
+		SYS_ERR(DTUN, LOGL_ERROR, errno, "socket() failed");
+		return -1;
+	}
+
+	r.rt_dst.sa_family = AF_INET;
+	r.rt_gateway.sa_family = AF_INET;
+	r.rt_genmask.sa_family = AF_INET;
+	memcpy(&((struct sockaddr_in *)&r.rt_dst)->sin_addr, dst, sizeof(*dst));
+	memcpy(&((struct sockaddr_in *)&r.rt_gateway)->sin_addr, gateway,
+	       sizeof(*gateway));
+	memcpy(&((struct sockaddr_in *)&r.rt_genmask)->sin_addr, mask,
+	       sizeof(*mask));
+
+	if (delete) {
+		if (ioctl(fd, SIOCDELRT, (void *)&r) < 0) {
+			SYS_ERR(DTUN, LOGL_ERROR, errno,
+				"ioctl(SIOCDELRT) failed");
+			close(fd);
+			return -1;
+		}
+	} else {
+		if (ioctl(fd, SIOCADDRT, (void *)&r) < 0) {
+			SYS_ERR(DTUN, LOGL_ERROR, errno,
+				"ioctl(SIOCADDRT) failed");
+			close(fd);
+			return -1;
+		}
+	}
+#elif defined(__FreeBSD__) || defined (__APPLE__)
+	struct {
+		struct rt_msghdr rt;
+		struct sockaddr_in dst;
+		struct sockaddr_in gate;
+		struct sockaddr_in mask;
+	} req;
+	struct rt_msghdr *rtm;
+
+	if ((fd = socket(AF_ROUTE, SOCK_RAW, 0)) == -1) {
+		SYS_ERR(DTUN, LOGL_ERROR, errno, "socket() failed");
+		return -1;
+	}
+
+	memset(&req, 0x00, sizeof(req));
+
+	rtm = &req.rt;
+
+	rtm->rtm_msglen = sizeof(req);
+	rtm->rtm_version = RTM_VERSION;
+	if (delete) {
+		rtm->rtm_type = RTM_DELETE;
+	} else {
+		rtm->rtm_type = RTM_ADD;
+	}
+	rtm->rtm_flags = RTF_UP | RTF_GATEWAY | RTF_STATIC;	/* TODO */
+	rtm->rtm_addrs = RTA_DST | RTA_GATEWAY | RTA_NETMASK;
+	rtm->rtm_pid = getpid();
+	rtm->rtm_seq = 0044;	/* TODO */
+
+	req.dst.sin_family = AF_INET;
+	req.dst.sin_len = sizeof(req.dst);
+	req.mask.sin_family = AF_INET;
+	req.mask.sin_len = sizeof(req.mask);
+	req.gate.sin_family = AF_INET;
+	req.gate.sin_len = sizeof(req.gate);
+
+	req.dst.sin_addr.s_addr = dst->s_addr;
+	req.mask.sin_addr.s_addr = mask->s_addr;
+	req.gate.sin_addr.s_addr = gateway->s_addr;
+
+	if (write(fd, rtm, rtm->rtm_msglen) < 0) {
+		SYS_ERR(DTUN, LOGL_ERROR, errno, "write() failed");
+		close(fd);
+		return -1;
+	}
+#endif
+	close(fd);
+	return 0;
+}
+
+static int netdev_route6(struct in6_addr *dst, struct in6_addr *gateway, int prefixlen, const char *gw_iface, int delete)
+{
+	int fd;
+#if defined(__linux__)
+	struct in6_rtmsg r;
+	struct ifreq ifr;
+
+	memset(&r, 0, sizeof(r));
+	r.rtmsg_flags = RTF_UP | RTF_GATEWAY; /* RTF_HOST not set */
+	r.rtmsg_metric = 1;
+
+	/* Create a channel to the NET kernel. */
+	if ((fd = socket(AF_INET6, SOCK_DGRAM, 0)) < 0) {
+		SYS_ERR(DTUN, LOGL_ERROR, errno, "socket() failed");
+		return -1;
+	}
+
+	if (gw_iface) {
+		strncpy(ifr.ifr_name, gw_iface, IFNAMSIZ);
+		ifr.ifr_name[IFNAMSIZ - 1] = 0; /* Make sure to terminate */
+		if (ioctl(fd, SIOCGIFINDEX, &ifr) < 0) {
+			SYS_ERR(DTUN, LOGL_ERROR, errno,
+				"ioctl(SIOCGIFINDEX) failed");
+			close(fd);
+			return -1;
+		}
+		r.rtmsg_ifindex = ifr.ifr_ifindex;
+	}
+
+	memcpy(&r.rtmsg_dst, dst->s6_addr, sizeof(struct in6_addr));
+	memcpy(&r.rtmsg_gateway, gateway->s6_addr, sizeof(struct in6_addr));
+	r.rtmsg_dst_len = prefixlen;
+
+	if (delete) {
+		if (ioctl(fd, SIOCDELRT, (void *)&r) < 0) {
+			SYS_ERR(DTUN, LOGL_ERROR, errno,
+				"ioctl(SIOCDELRT) failed");
+			close(fd);
+			return -1;
+		}
+	} else {
+		if (ioctl(fd, SIOCADDRT, (void *)&r) < 0) {
+			SYS_ERR(DTUN, LOGL_ERROR, errno,
+				"ioctl(SIOCADDRT) failed");
+			close(fd);
+			return -1;
+		}
+	}
+	close(fd);
+#endif
+	return 0;
+}
+
+int netdev_addroute4(struct in_addr *dst, struct in_addr *gateway, struct in_addr *mask)
+{
+	return netdev_route4(dst, gateway, mask, 0);
+}
+
+int netdev_delroute4(struct in_addr *dst, struct in_addr *gateway, struct in_addr *mask)
+{
+	return netdev_route4(dst, gateway, mask, 1);
+}
+
+int netdev_addroute6(struct in6_addr *dst, struct in6_addr *gateway, int prefixlen, const char *gw_iface)
+{
+	return netdev_route6(dst, gateway, prefixlen, gw_iface, 0);
+}
+
+int netdev_delroute6(struct in6_addr *dst, struct in6_addr *gateway, int prefixlen, const char *gw_iface)
+{
+	return netdev_route6(dst, gateway, prefixlen, gw_iface, 1);
+}
+
+
+
+#include <ifaddrs.h>
+
+/*! Obtain the local address of a network device
+ *  \param[in] devname Target device owning the IP
+ *  \param[out] prefix_list List of prefix structures to fill with each IPv4/6 and prefix length found.
+ *  \param[in] prefix_size Amount of elements allowed to be fill in the prefix_list array.
+ *  \param[in] flags Specify which kind of IP to look for: IP_TYPE_IPv4, IP_TYPE_IPv6_LINK, IP_TYPE_IPv6_NONLINK
+ *  \returns The number of ips found following the criteria specified by flags, -1 on error.
+ *
+ * This function will fill prefix_list with up to prefix_size IPs following the
+ * criteria specified by flags parameter. It returns the number of IPs matching
+ * the criteria. As a result, the number returned can be bigger than
+ * prefix_size. It can be used with prefix_size=0 to get an estimate of the size
+ * needed for prefix_list.
+ */
+int netdev_ip_local_get(const char *devname, struct in46_prefix *prefix_list, size_t prefix_size, int flags)
+{
+	static const uint8_t ll_prefix[] = { 0xfe,0x80, 0,0, 0,0, 0,0 };
+	struct ifaddrs *ifaddr, *ifa;
+	struct in46_addr netmask;
+	size_t count = 0;
+	bool is_ipv6_ll;
+
+	if (getifaddrs(&ifaddr) == -1) {
+		return -1;
+	}
+
+	for (ifa = ifaddr; ifa != NULL; ifa = ifa->ifa_next) {
+		if (ifa->ifa_addr == NULL)
+			continue;
+
+		if (strcmp(ifa->ifa_name, devname))
+			continue;
+
+		if (ifa->ifa_addr->sa_family == AF_INET && (flags & IP_TYPE_IPv4)) {
+			struct sockaddr_in *sin4 = (struct sockaddr_in *) ifa->ifa_addr;
+			struct sockaddr_in *netmask4 = (struct sockaddr_in *) ifa->ifa_netmask;
+
+			if (count < prefix_size) {
+				netmask.len = sizeof(netmask4->sin_addr);
+				netmask.v4 = netmask4->sin_addr;
+				prefix_list[count].addr.len = sizeof(sin4->sin_addr);
+				prefix_list[count].addr.v4 = sin4->sin_addr;
+				prefix_list[count].prefixlen = in46a_netmasklen(&netmask);
+			}
+			count++;
+		}
+
+		if (ifa->ifa_addr->sa_family == AF_INET6 && (flags & IP_TYPE_IPv6)) {
+			struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *) ifa->ifa_addr;
+			struct sockaddr_in6 *netmask6 = (struct sockaddr_in6 *) ifa->ifa_netmask;
+
+			is_ipv6_ll = !memcmp(sin6->sin6_addr.s6_addr, ll_prefix, sizeof(ll_prefix));
+			if ((flags & IP_TYPE_IPv6_NONLINK) && is_ipv6_ll)
+				continue;
+			if ((flags & IP_TYPE_IPv6_LINK) && !is_ipv6_ll)
+				continue;
+
+			if (count < prefix_size) {
+				netmask.len = sizeof(netmask6->sin6_addr);
+				netmask.v6 = netmask6->sin6_addr;
+				prefix_list[count].addr.len = sizeof(sin6->sin6_addr);
+				prefix_list[count].addr.v6 = sin6->sin6_addr;
+				prefix_list[count].prefixlen = in46a_netmasklen(&netmask);
+			}
+			count++;
+		}
+	}
+
+	freeifaddrs(ifaddr);
+	return count;
+}

lib/netdev.h

@@ -0,0 +1,74 @@
+#pragma once
+/*
+ * TUN interface functions.
+ * Copyright (C) 2002, 2003 Mondru AB.
+ * Copyright (C) 2017-2018 by Harald Welte <laforge@gnumonks.org>
+ *
+ * The contents of this file may be used under the terms of the GNU
+ * General Public License Version 2, provided that the above copyright
+ * notice and this permission notice is included in all copies or
+ * substantial portions of the software.
+ *
+ */
+
+#include <net/if.h>
+
+#include "../lib/in46_addr.h"
+
+#define TUN_NLBUFSIZE   1024
+
+#include "config.h"
+
+/* ipv6 ip type flags for tun_ipv6_local_get() */
+enum {
+	IP_TYPE_IPv4 = 1,
+	IP_TYPE_IPv6_LINK = 2,
+	IP_TYPE_IPv6_NONLINK = 4,
+};
+#define IP_TYPE_IPv6 (IP_TYPE_IPv6_LINK | IP_TYPE_IPv6_NONLINK)
+
+
+#ifndef HAVE_IPHDR
+struct iphdr
+  {
+#if __BYTE_ORDER == __LITTLE_ENDIAN
+    unsigned int ihl:4;
+    unsigned int version:4;
+#elif __BYTE_ORDER == __BIG_ENDIAN
+    unsigned int version:4;
+    unsigned int ihl:4;
+#else
+# error	"Please fix <bits/endian.h>"
+#endif
+    u_int8_t tos;
+    u_int16_t tot_len;
+    u_int16_t id;
+    u_int16_t frag_off;
+    u_int8_t ttl;
+    u_int8_t protocol;
+    u_int16_t check;
+    u_int32_t saddr;
+    u_int32_t daddr;
+    /*The options start here. */
+  };
+#endif /* !HAVE_IPHDR */
+
+extern int netdev_setaddr4(const char *devname, struct in_addr *addr,
+			   struct in_addr *dstaddr, struct in_addr *netmask);
+
+extern int netdev_setaddr6(const char *devname, struct in6_addr *addr, struct in6_addr *dstaddr,
+			   size_t prefixlen);
+
+extern int netdev_addaddr4(const char *devname, struct in_addr *addr,
+			   struct in_addr *dstaddr, struct in_addr *netmask);
+
+extern int netdev_addaddr6(const char *devname, struct in6_addr *addr,
+			   struct in6_addr *dstaddr, int prefixlen);
+
+extern int netdev_addroute4(struct in_addr *dst, struct in_addr *gateway, struct in_addr *mask);
+extern int netdev_delroute4(struct in_addr *dst, struct in_addr *gateway, struct in_addr *mask);
+extern int netdev_addroute6(struct in6_addr *dst, struct in6_addr *gateway, int prefixlen, const char *gw_iface);
+extern int netdev_delroute6(struct in6_addr *dst, struct in6_addr *gateway, int prefixlen, const char *gw_iface);
+
+extern int netdev_ip_local_get(const char *devname, struct in46_prefix *prefix_list,
+				size_t prefix_size, int flags);

lib/netns.c

@@ -0,0 +1,272 @@
+/*
+ * Copyright (C) 2014-2017, Travelping GmbH <info@travelping.com>
+ * Copyright (C) 2020, Harald Welte <laforge@gnumonks.org>
+ *
+ * This program is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU Affero 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 Affero General Public License for more details.
+ *
+ * You should have received a copy of the GNU Affero General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+
+#if defined(__linux__)
+
+#ifdef HAVE_CONFIG_H
+# include "config.h"
+#endif
+
+#ifndef _GNU_SOURCE
+# define _GNU_SOURCE
+#endif
+
+#include <errno.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <unistd.h>
+#include <sched.h>
+#include <signal.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <sys/socket.h>
+#include <sys/mount.h>
+#include <sys/param.h>
+#include <fcntl.h>
+#include <errno.h>
+
+#include <osmocom/core/utils.h>
+
+#include "netns.h"
+
+#define NETNS_PATH "/var/run/netns"
+
+/*! default namespace of the GGSN process */
+static int default_nsfd = -1;
+
+/*! switch to a (non-default) namespace, store existing signal mask in oldmask.
+ *  \param[in] nsfd file descriptor representing the namespace to whch we shall switch
+ *  \param[out] oldmask caller-provided memory location to which old signal mask is stored
+ *  \ returns 0 on success or negative (errno) in case of error */
+int switch_ns(int nsfd, sigset_t *oldmask)
+{
+	sigset_t intmask;
+	int rc;
+
+	OSMO_ASSERT(default_nsfd >= 0);
+
+	if (sigfillset(&intmask) < 0)
+		return -errno;
+	if ((rc = sigprocmask(SIG_BLOCK, &intmask, oldmask)) != 0)
+		return -rc;
+
+	if (setns(nsfd, CLONE_NEWNET) < 0) {
+		/* restore old mask if we couldn't switch the netns */
+		sigprocmask(SIG_SETMASK, oldmask, NULL);
+		return -errno;
+	}
+	return 0;
+}
+
+/*! switch back to the default namespace, restoring signal mask.
+ *  \param[in] oldmask signal mask to restore after returning to default namespace
+ *  \returns 0 on successs; negative errno value in case of error */
+int restore_ns(sigset_t *oldmask)
+{
+	OSMO_ASSERT(default_nsfd >= 0);
+
+	int rc;
+	if (setns(default_nsfd, CLONE_NEWNET) < 0)
+		return -errno;
+
+	if ((rc = sigprocmask(SIG_SETMASK, oldmask, NULL)) != 0)
+		return -rc;
+	return 0;
+}
+
+/*! open a file from within specified network namespace */
+int open_ns(int nsfd, const char *pathname, int flags)
+{
+	sigset_t intmask, oldmask;
+	int ret;
+	int fd = -1;
+	int rc;
+
+	OSMO_ASSERT(default_nsfd >= 0);
+
+	/* mask off all signals, store old signal mask */
+	if (sigfillset(&intmask) < 0)
+		return -errno;
+	if ((rc = sigprocmask(SIG_BLOCK, &intmask, &oldmask)) != 0)
+		return -rc;
+
+	/* associate the calling thread with namespace file descriptor */
+	if (setns(nsfd, CLONE_NEWNET) < 0) {
+		ret = -errno;
+		goto restore_sigmask;
+	}
+	/* open the requested file/path */
+	if ((fd = open(pathname, flags)) < 0) {
+		ret = -errno;
+		goto restore_defaultns;
+	}
+	ret = fd;
+
+restore_defaultns:
+	/* return back to default namespace */
+	if (setns(default_nsfd, CLONE_NEWNET) < 0) {
+		if (fd >= 0)
+			close(fd);
+		return -errno;
+	}
+
+restore_sigmask:
+	/* restore process mask */
+	if ((rc = sigprocmask(SIG_SETMASK, &oldmask, NULL)) != 0) {
+		if (fd >= 0)
+			close(fd);
+		return -rc;
+	}
+
+	return ret;
+}
+
+/*! create a socket in another namespace.
+ *  Switches temporarily to namespace indicated by nsfd, creates a socket in
+ *  that namespace and then returns to the default namespace.
+ *  \param[in] nsfd File descriptor of the namspace in which to create socket
+ *  \param[in] domain Domain of the socket (AF_INET, ...)
+ *  \param[in] type Type of the socket (SOCK_STREAM, ...)
+ *  \param[in] protocol Protocol of the socket (IPPROTO_TCP, ...)
+ *  \returns 0 on success; negative errno in case of error */
+int socket_ns(int nsfd, int domain, int type, int protocol)
+{
+	sigset_t intmask, oldmask;
+	int ret;
+	int sk = -1;
+	int rc;
+
+	OSMO_ASSERT(default_nsfd >= 0);
+
+	/* mask off all signals, store old signal mask */
+	if (sigfillset(&intmask) < 0)
+		return -errno;
+	if ((rc = sigprocmask(SIG_BLOCK, &intmask, &oldmask)) != 0)
+		return -rc;
+
+	/* associate the calling thread with namespace file descriptor */
+	if (setns(nsfd, CLONE_NEWNET) < 0) {
+		ret = -errno;
+		goto restore_sigmask;
+	}
+
+	/* create socket of requested domain/type/proto */
+	if ((sk = socket(domain, type, protocol)) < 0) {
+		ret = -errno;
+		goto restore_defaultns;
+	}
+	ret = sk;
+
+restore_defaultns:
+	/* return back to default namespace */
+	if (setns(default_nsfd, CLONE_NEWNET) < 0) {
+		if (sk >= 0)
+			close(sk);
+		return -errno;
+	}
+
+restore_sigmask:
+	/* restore process mask */
+	if ((rc = sigprocmask(SIG_SETMASK, &oldmask, NULL)) != 0) {
+		if (sk >= 0)
+			close(sk);
+		return -rc;
+	}
+	return ret;
+}
+
+/*! initialize this network namespace helper module.
+ *  Must be called before using any other functions of this file.
+ *  \returns 0 on success; negative errno in case of error */
+int init_netns()
+{
+	/* store the default namespace for later reference */
+	if ((default_nsfd = open("/proc/self/ns/net", O_RDONLY)) < 0)
+		return -errno;
+	return 0;
+}
+
+/*! create obtain file descriptor for network namespace of give name.
+ *  Creates /var/run/netns  if it doesn't exist already.
+ *  \param[in] name Name of the network namespace (in /var/run/netns/)
+ *  \returns File descriptor of network namespace; negative errno in case of error */
+int get_nsfd(const char *name)
+{
+	int ret = 0;
+	int rc;
+	int fd;
+	sigset_t intmask, oldmask;
+	char path[MAXPATHLEN] = NETNS_PATH;
+
+	OSMO_ASSERT(default_nsfd >= 0);
+
+	/* create /var/run/netns, if it doesn't exist already */
+	rc = mkdir(path, S_IRWXU|S_IRGRP|S_IXGRP|S_IROTH|S_IXOTH);
+	if (rc < 0 && errno != EEXIST)
+		return rc;
+
+	/* create /var/run/netns/[name], if it doesn't exist already */
+	snprintf(path, sizeof(path), "%s/%s", NETNS_PATH, name);
+	fd = open(path, O_RDONLY|O_CREAT|O_EXCL, 0);
+	if (fd < 0) {
+		if (errno == EEXIST) {
+			if ((fd = open(path, O_RDONLY)) < 0)
+				return -errno;
+			return fd;
+		}
+		return -errno;
+	}
+	if (close(fd) < 0)
+		return -errno;
+
+	/* mask off all signals, store old signal mask */
+	if (sigfillset(&intmask) < 0)
+		return -errno;
+	if ((rc = sigprocmask(SIG_BLOCK, &intmask, &oldmask)) != 0)
+		return -rc;
+
+	/* create a new network namespace */
+	if (unshare(CLONE_NEWNET) < 0) {
+		ret = -errno;
+		goto restore_sigmask;
+	}
+	if (mount("/proc/self/ns/net", path, "none", MS_BIND, NULL) < 0)
+		ret = -errno;
+
+	/* switch back to default namespace */
+	if (setns(default_nsfd, CLONE_NEWNET) < 0)
+		return -errno;
+
+restore_sigmask:
+	/* restore process mask */
+	if ((rc = sigprocmask(SIG_SETMASK, &oldmask, NULL)) != 0)
+		return -rc;
+
+	/* might have been set above in case mount fails */
+	if (ret < 0)
+		return ret;
+
+	/* finally, open the created namespace file descriptor from default ns */
+	if ((fd = open(path, O_RDONLY)) < 0)
+		return -errno;
+
+	return fd;
+}
+
+#endif

lib/netns.h

@@ -0,0 +1,35 @@
+/*
+ * Copyright (C) 2014-2017, Travelping GmbH <info@travelping.com>
+ *
+ * This program is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU Affero 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 Affero General Public License for more details.
+ *
+ * You should have received a copy of the GNU Affero General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+
+#ifndef __NETNS_H
+#define __NETNS_H
+
+#if defined(__linux__)
+
+int init_netns(void);
+
+int switch_ns(int nsfd, sigset_t *oldmask);
+int restore_ns(sigset_t *oldmask);
+
+int open_ns(int nsfd, const char *pathname, int flags);
+int socket_ns(int nsfd, int domain, int type, int protocol);
+int get_nsfd(const char *name);
+
+#endif
+
+#endif

lib/tun.c

@@ -1,7 +1,7 @@
 /*
  * TUN interface functions.
  * Copyright (C) 2002, 2003, 2004 Mondru AB.
- * Copyright (C) 2017 by Harald Welte <laforge@gnumonks.org>
+ * Copyright (C) 2017-2018 by Harald Welte <laforge@gnumonks.org>
  *
  * The contents of this file may be used under the terms of the GNU
  * General Public License Version 2, provided that the above copyright
@@ -19,6 +19,7 @@
 
 #include <stdio.h>
 #include <stdlib.h>
+#include <stdbool.h>
 #include <sys/types.h>
 #include <sys/socket.h>
 #include <netinet/in.h>
@@ -42,8 +43,6 @@
 
 #if defined(__linux__)
 #include <linux/if_tun.h>
-#include <linux/netlink.h>
-#include <linux/rtnetlink.h>
 
 #elif defined (__FreeBSD__)
 #include <net/if_tun.h>
@@ -59,531 +58,104 @@
 
 #include "tun.h"
 #include "syserr.h"
-
-static int tun_setaddr4(struct tun_t *this, struct in_addr *addr,
-			struct in_addr *dstaddr, struct in_addr *netmask);
-
-#if defined(__linux__)
-
-#include <linux/ipv6.h>
-
-static int tun_nlattr(struct nlmsghdr *n, int nsize, int type, void *d, int dlen)
-{
-	int len = RTA_LENGTH(dlen);
-	int alen = NLMSG_ALIGN(n->nlmsg_len);
-	struct rtattr *rta = (struct rtattr *)(((void *)n) + alen);
-	if (alen + len > nsize)
-		return -1;
-	rta->rta_len = len;
-	rta->rta_type = type;
-	memcpy(RTA_DATA(rta), d, dlen);
-	n->nlmsg_len = alen + len;
-	return 0;
-}
-#endif
-
-static int tun_sifflags(struct tun_t *this, int flags)
-{
-	struct ifreq ifr;
-	int fd;
-
-	memset(&ifr, '\0', sizeof(ifr));
-	ifr.ifr_flags = flags;
-	strncpy(ifr.ifr_name, this->devname, IFNAMSIZ);
-	ifr.ifr_name[IFNAMSIZ - 1] = 0;	/* Make sure to terminate */
-	if ((fd = socket(AF_INET, SOCK_DGRAM, 0)) < 0) {
-		SYS_ERR(DTUN, LOGL_ERROR, errno, "socket() failed");
-		return -1;
-	}
-	if (ioctl(fd, SIOCSIFFLAGS, &ifr)) {
-		SYS_ERR(DTUN, LOGL_ERROR, errno,
-			"ioctl(SIOCSIFFLAGS) failed");
-		close(fd);
-		return -1;
-	}
-	close(fd);
-	return 0;
-}
-
-int tun_addaddr(struct tun_t *this,
-		struct in_addr *addr,
-		struct in_addr *dstaddr, struct in_addr *netmask)
-{
-
-#if defined(__linux__)
-	struct {
-		struct nlmsghdr n;
-		struct ifaddrmsg i;
-		char buf[TUN_NLBUFSIZE];
-	} req;
-
-	struct sockaddr_nl local;
-	socklen_t addr_len;
-	int fd;
-	int status;
-
-	struct sockaddr_nl nladdr;
-	struct iovec iov;
-	struct msghdr msg;
-
-	if (!this->addrs)	/* Use ioctl for first addr to make ping work */
-		return tun_setaddr4(this, addr, dstaddr, netmask);
-
-	memset(&req, 0, sizeof(req));
-	req.n.nlmsg_len = NLMSG_LENGTH(sizeof(struct ifaddrmsg));
-	req.n.nlmsg_flags = NLM_F_REQUEST | NLM_F_CREATE;
-	req.n.nlmsg_type = RTM_NEWADDR;
-	req.i.ifa_family = AF_INET;
-	req.i.ifa_prefixlen = 32;	/* 32 FOR IPv4 */
-	req.i.ifa_flags = 0;
-	req.i.ifa_scope = RT_SCOPE_HOST;	/* TODO or 0 */
-	req.i.ifa_index = if_nametoindex(this->devname);
-	if (!req.i.ifa_index) {
-		SYS_ERR(DTUN, LOGL_ERROR, errno, "Unable to get ifindex for %s", this->devname);
-		return -1;
-	}
-
-	tun_nlattr(&req.n, sizeof(req), IFA_ADDRESS, addr, sizeof(addr));
-	tun_nlattr(&req.n, sizeof(req), IFA_LOCAL, dstaddr, sizeof(dstaddr));
-
-	if ((fd = socket(AF_NETLINK, SOCK_RAW, NETLINK_ROUTE)) < 0) {
-		SYS_ERR(DTUN, LOGL_ERROR, errno, "socket() failed");
-		return -1;
-	}
-
-	memset(&local, 0, sizeof(local));
-	local.nl_family = AF_NETLINK;
-	local.nl_groups = 0;
-
-	if (bind(fd, (struct sockaddr *)&local, sizeof(local)) < 0) {
-		SYS_ERR(DTUN, LOGL_ERROR, errno, "bind() failed");
-		close(fd);
-		return -1;
-	}
-
-	addr_len = sizeof(local);
-	if (getsockname(fd, (struct sockaddr *)&local, &addr_len) < 0) {
-		SYS_ERR(DTUN, LOGL_ERROR, errno,
-			"getsockname() failed");
-		close(fd);
-		return -1;
-	}
-
-	if (addr_len != sizeof(local)) {
-		SYS_ERR(DTUN, LOGL_ERROR, 0,
-			"Wrong address length %d", addr_len);
-		close(fd);
-		return -1;
-	}
-
-	if (local.nl_family != AF_NETLINK) {
-		SYS_ERR(DTUN, LOGL_ERROR, 0,
-			"Wrong address family %d", local.nl_family);
-		close(fd);
-		return -1;
-	}
-
-	iov.iov_base = (void *)&req.n;
-	iov.iov_len = req.n.nlmsg_len;
-
-	msg.msg_name = (void *)&nladdr;
-	msg.msg_namelen = sizeof(nladdr);
-	msg.msg_iov = &iov;
-	msg.msg_iovlen = 1;
-	msg.msg_control = NULL;
-	msg.msg_controllen = 0;
-	msg.msg_flags = 0;
-
-	memset(&nladdr, 0, sizeof(nladdr));
-	nladdr.nl_family = AF_NETLINK;
-	nladdr.nl_pid = 0;
-	nladdr.nl_groups = 0;
-
-	req.n.nlmsg_seq = 0;
-	req.n.nlmsg_flags |= NLM_F_ACK;
-
-	status = sendmsg(fd, &msg, 0);
-	if (status != req.n.nlmsg_len) {
-		SYS_ERR(DTUN, LOGL_ERROR, errno, "sendmsg() failed, returned %d", status);
-		close(fd);
-		return -1;
-	}
-
-	status = tun_sifflags(this, IFF_UP | IFF_RUNNING);
-	if (status == -1) {
-		close(fd);
-		return -1;
-	}
-
-
-	close(fd);
-	this->addrs++;
-	return 0;
-
-#elif defined (__FreeBSD__) || defined (__APPLE__)
-
-	int fd;
-	struct ifaliasreq areq;
-
-	/* TODO: Is this needed on FreeBSD? */
-	if (!this->addrs)	/* Use ioctl for first addr to make ping work */
-		return tun_setaddr4(this, addr, dstaddr, netmask);	/* TODO dstaddr */
-
-	memset(&areq, 0, sizeof(areq));
-
-	/* Set up interface name */
-	strncpy(areq.ifra_name, this->devname, IFNAMSIZ);
-	areq.ifra_name[IFNAMSIZ - 1] = 0;	/* Make sure to terminate */
-
-	((struct sockaddr_in *)&areq.ifra_addr)->sin_family = AF_INET;
-	((struct sockaddr_in *)&areq.ifra_addr)->sin_len =
-	    sizeof(areq.ifra_addr);
-	((struct sockaddr_in *)&areq.ifra_addr)->sin_addr.s_addr = addr->s_addr;
-
-	((struct sockaddr_in *)&areq.ifra_mask)->sin_family = AF_INET;
-	((struct sockaddr_in *)&areq.ifra_mask)->sin_len =
-	    sizeof(areq.ifra_mask);
-	((struct sockaddr_in *)&areq.ifra_mask)->sin_addr.s_addr =
-	    netmask->s_addr;
-
-	/* For some reason FreeBSD uses ifra_broadcast for specifying dstaddr */
-	((struct sockaddr_in *)&areq.ifra_broadaddr)->sin_family = AF_INET;
-	((struct sockaddr_in *)&areq.ifra_broadaddr)->sin_len =
-	    sizeof(areq.ifra_broadaddr);
-	((struct sockaddr_in *)&areq.ifra_broadaddr)->sin_addr.s_addr =
-	    dstaddr->s_addr;
-
-	/* Create a channel to the NET kernel. */
-	if ((fd = socket(AF_INET, SOCK_DGRAM, 0)) < 0) {
-		SYS_ERR(DTUN, LOGL_ERROR, errno, "socket() failed");
-		return -1;
-	}
-
-	if (ioctl(fd, SIOCAIFADDR, (void *)&areq) < 0) {
-		SYS_ERR(DTUN, LOGL_ERROR, errno,
-			"ioctl(SIOCAIFADDR) failed");
-		close(fd);
-		return -1;
-	}
-
-	close(fd);
-	this->addrs++;
-	return 0;
-
-#endif
-
-}
+#include "gtp-kernel.h"
 
 static int tun_setaddr4(struct tun_t *this, struct in_addr *addr,
 			struct in_addr *dstaddr, struct in_addr *netmask)
 {
-	struct ifreq ifr;
-	int fd;
+	int rc;
+	rc = netdev_setaddr4(this->devname, addr, dstaddr, netmask);
+	if (rc < 0)
+		return rc;
 
-	memset(&ifr, '\0', sizeof(ifr));
-	ifr.ifr_addr.sa_family = AF_INET;
-	ifr.ifr_dstaddr.sa_family = AF_INET;
-
-#if defined(__linux__)
-	ifr.ifr_netmask.sa_family = AF_INET;
-
-#elif defined(__FreeBSD__) || defined (__APPLE__)
-	((struct sockaddr_in *)&ifr.ifr_addr)->sin_len =
-	    sizeof(struct sockaddr_in);
-	((struct sockaddr_in *)&ifr.ifr_dstaddr)->sin_len =
-	    sizeof(struct sockaddr_in);
-#endif
-
-	strncpy(ifr.ifr_name, this->devname, IFNAMSIZ);
-	ifr.ifr_name[IFNAMSIZ - 1] = 0;	/* Make sure to terminate */
-
-	/* Create a channel to the NET kernel. */
-	if ((fd = socket(AF_INET, SOCK_DGRAM, 0)) < 0) {
-		SYS_ERR(DTUN, LOGL_ERROR, errno, "socket() failed");
-		return -1;
+	if (addr) {
+		this->addr.len = sizeof(struct in_addr);
+		this->addr.v4.s_addr = addr->s_addr;
 	}
-
-	if (addr) {		/* Set the interface address */
-		this->addr.s_addr = addr->s_addr;
-		memcpy(&((struct sockaddr_in *)&ifr.ifr_addr)->sin_addr, addr,
-		       sizeof(*addr));
-		if (ioctl(fd, SIOCSIFADDR, (void *)&ifr) < 0) {
-			if (errno != EEXIST) {
-				SYS_ERR(DTUN, LOGL_ERROR, errno,
-					"ioctl(SIOCSIFADDR) failed");
-			} else {
-				SYS_ERR(DTUN, LOGL_NOTICE, errno,
-					"ioctl(SIOCSIFADDR): Address already exists");
-			}
-			close(fd);
-			return -1;
-		}
+	if (dstaddr) {
+		this->dstaddr.len = sizeof(struct in_addr);
+		this->dstaddr.v4.s_addr = dstaddr->s_addr;
 	}
-
-	if (dstaddr) {		/* Set the destination address */
-		this->dstaddr.s_addr = dstaddr->s_addr;
-		memcpy(&((struct sockaddr_in *)&ifr.ifr_dstaddr)->sin_addr,
-		       dstaddr, sizeof(*dstaddr));
-		if (ioctl(fd, SIOCSIFDSTADDR, (caddr_t) & ifr) < 0) {
-			SYS_ERR(DTUN, LOGL_ERROR, errno,
-				"ioctl(SIOCSIFDSTADDR) failed");
-			close(fd);
-			return -1;
-		}
-	}
-
-	if (netmask) {		/* Set the netmask */
+	if (netmask)
 		this->netmask.s_addr = netmask->s_addr;
-#if defined(__linux__)
-		memcpy(&((struct sockaddr_in *)&ifr.ifr_netmask)->sin_addr,
-		       netmask, sizeof(*netmask));
-
-#elif defined(__FreeBSD__) || defined (__APPLE__)
-		((struct sockaddr_in *)&ifr.ifr_addr)->sin_addr.s_addr =
-		    netmask->s_addr;
-#endif
-
-		if (ioctl(fd, SIOCSIFNETMASK, (void *)&ifr) < 0) {
-			SYS_ERR(DTUN, LOGL_ERROR, errno,
-				"ioctl(SIOCSIFNETMASK) failed");
-			close(fd);
-			return -1;
-		}
-	}
-
-	close(fd);
 	this->addrs++;
-
-	/* On linux the route to the interface is set automatically
-	   on FreeBSD we have to do this manually */
-
-	/* TODO: How does it work on Solaris? */
-
-	tun_sifflags(this, IFF_UP | IFF_RUNNING);
-
 #if defined(__FreeBSD__) || defined (__APPLE__)
-	tun_addroute(this, dstaddr, addr, &this->netmask);
 	this->routes = 1;
 #endif
 
-	return 0;
+	return rc;
 }
 
 static int tun_setaddr6(struct tun_t *this, struct in6_addr *addr, struct in6_addr *dstaddr,
 			size_t prefixlen)
 {
-	struct in6_ifreq ifr;
-	int fd;
-
-	memset(&ifr, 0, sizeof(ifr));
-
-#if defined(__linux__)
-	ifr.ifr6_prefixlen = prefixlen;
-	ifr.ifr6_ifindex = if_nametoindex(this->devname);
-	if (ifr.ifr6_ifindex == 0) {
-		SYS_ERR(DTUN, LOGL_ERROR, 0, "Error getting ifindex for %s\n", this->devname);
-		return -1;
-	}
-#elif defined(__FreeBSD__) || defined (__APPLE__)
-	strncpy(ifr.ifr_name, this->devname, IFNAMSIZ);
-#endif
-
-	/* Create a channel to the NET kernel */
-	if ((fd = socket(AF_INET6, SOCK_DGRAM, 0)) < 0) {
-		SYS_ERR(DTUN, LOGL_ERROR, 0, "socket() failed");
-		return -1;
-	}
-
-#if defined(__linux__)
-	if (addr) {
-		memcpy(&this->addr, addr, sizeof(*addr));
-		memcpy(&ifr.ifr6_addr, addr, sizeof(*addr));
-		if (ioctl(fd, SIOCSIFADDR, (void *) &ifr) < 0) {
-			if (errno != EEXIST) {
-				SYS_ERR(DTUN, LOGL_ERROR, 0, "ioctl(SIOCSIFADDR) failed");
-			} else {
-				SYS_ERR(DTUN, LOGL_NOTICE, 0, "ioctl(SIOCSIFADDR): Address alreadsy exists");
-			}
-			close(fd);
-			return -1;
-		}
-	}
-
-#if 0
-	/* FIXME: looks like this is not possible/necessary for IPv6? */
+	int rc;
+	rc = netdev_setaddr6(this->devname, addr, dstaddr, prefixlen);
+	if (rc < 0)
+		return rc;
 	if (dstaddr) {
-		memcpy(&this->dstaddr, dstaddr, sizeof(*dstaddr));
-		memcpy(&ifr.ifr6_addr, dstaddr, sizeof(*dstaddr));
-		if (ioctl(fd, SIOCSIFDSTADDR, (caddr_t *) &ifr) < 0) {
-			SYS_ERR(DTUN, LOGL_ERROR, "ioctl(SIOCSIFDSTADDR) failed");
-			close(fd);
-			return -1;
-		}
+		this->dstaddr.len = sizeof(*dstaddr);
+		memcpy(&this->dstaddr.v6, dstaddr, sizeof(*dstaddr));
 	}
-#endif
-
-#elif defined(__FreeBSD__) || defined (__APPLE__)
-	if (addr)
-		memcpy(&ifr.ifr_ifru.ifru_addr, addr, sizeof(ifr.ifr_ifru.ifru_addr));
-	if (dstaddr)
-		memcpy(&ifr.ifr_ifru.ifru_dstaddr, dstaddr, sizeof(ifr.ifr_ifru.ifru_dstaddr));
-
-	if (ioctl(fd, SIOCSIFADDR_IN6, (struct ifreq *)&ifr) < 0) {
-		SYS_ERR(DTUN, LOGL_ERROR, 0, "ioctl(SIOCSIFADDR_IN6) failed");
-		close(fd);
-		return -1;
-	}
-#endif
-
-	close(fd);
 	this->addrs++;
-
-	/* On linux the route to the interface is set automatically
-	   on FreeBSD we have to do this manually */
-
-	/* TODO: How does it work on Solaris? */
-
-	tun_sifflags(this, IFF_UP | IFF_RUNNING);
-
-#if 0	/* FIXME */
-//#if defined(__FreeBSD__) || defined (__APPLE__)
-	tun_addroute6(this, dstaddr, addr, prefixlen);
+#if defined(__FreeBSD__) || defined (__APPLE__)
 	this->routes = 1;
 #endif
 
-	return 0;
+	return rc;
 }
 
-int tun_setaddr(struct tun_t *this, struct in46_addr *addr, struct in46_addr *dstaddr, size_t prefixlen)
+static int tun_addaddr4(struct tun_t *this, struct in_addr *addr,
+			struct in_addr *dstaddr, struct in_addr *netmask)
+{
+	int rc;
+
+	/* TODO: Is this needed on FreeBSD? */
+	if (!this->addrs)	/* Use ioctl for first addr to make ping work */
+		return tun_setaddr4(this, addr, dstaddr, netmask);	/* TODO dstaddr */
+
+	rc = netdev_addaddr4(this->devname, addr, dstaddr, netmask);
+	if (rc < 0)
+		return rc;
+
+	this->addrs++;
+
+	return rc;
+}
+
+static int tun_addaddr6(struct tun_t *this,
+		struct in6_addr *addr,
+		struct in6_addr *dstaddr, int prefixlen)
+{
+	int rc;
+
+	if (!this->addrs)	/* Use ioctl for first addr to make ping work */
+		return tun_setaddr6(this, addr, dstaddr, prefixlen);
+
+	rc = netdev_addaddr6(this->devname, addr, dstaddr, prefixlen);
+	if (rc < 0)
+		return rc;
+
+	this->addrs++;
+
+	return rc;
+}
+
+int tun_addaddr(struct tun_t *this, struct in46_addr *addr, struct in46_addr *dstaddr, size_t prefixlen)
 {
 	struct in_addr netmask;
 	switch (addr->len) {
 	case 4:
 		netmask.s_addr = htonl(0xffffffff << (32 - prefixlen));
-		return tun_setaddr4(this, &addr->v4, dstaddr ? &dstaddr->v4 : NULL, &netmask);
+		return tun_addaddr4(this, &addr->v4, dstaddr ? &dstaddr->v4 : NULL, &netmask);
 	case 16:
-		return tun_setaddr6(this, &addr->v6, dstaddr ? &dstaddr->v6 : NULL, prefixlen);
+		return tun_addaddr6(this, &addr->v6, dstaddr ? &dstaddr->v6 : NULL, prefixlen);
 	default:
 		return -1;
 	}
 }
 
-static int tun_route(struct tun_t *this,
-	      struct in_addr *dst,
-	      struct in_addr *gateway, struct in_addr *mask, int delete)
-{
-
-#if defined(__linux__)
-
-	struct rtentry r;
-	int fd;
-
-	memset(&r, '\0', sizeof(r));
-	r.rt_flags = RTF_UP | RTF_GATEWAY;	/* RTF_HOST not set */
-
-	/* Create a channel to the NET kernel. */
-	if ((fd = socket(AF_INET, SOCK_DGRAM, 0)) < 0) {
-		SYS_ERR(DTUN, LOGL_ERROR, errno, "socket() failed");
-		return -1;
-	}
-
-	r.rt_dst.sa_family = AF_INET;
-	r.rt_gateway.sa_family = AF_INET;
-	r.rt_genmask.sa_family = AF_INET;
-	memcpy(&((struct sockaddr_in *)&r.rt_dst)->sin_addr, dst, sizeof(*dst));
-	memcpy(&((struct sockaddr_in *)&r.rt_gateway)->sin_addr, gateway,
-	       sizeof(*gateway));
-	memcpy(&((struct sockaddr_in *)&r.rt_genmask)->sin_addr, mask,
-	       sizeof(*mask));
-
-	if (delete) {
-		if (ioctl(fd, SIOCDELRT, (void *)&r) < 0) {
-			SYS_ERR(DTUN, LOGL_ERROR, errno,
-				"ioctl(SIOCDELRT) failed");
-			close(fd);
-			return -1;
-		}
-	} else {
-		if (ioctl(fd, SIOCADDRT, (void *)&r) < 0) {
-			SYS_ERR(DTUN, LOGL_ERROR, errno,
-				"ioctl(SIOCADDRT) failed");
-			close(fd);
-			return -1;
-		}
-	}
-	close(fd);
-	return 0;
-
-#elif defined(__FreeBSD__) || defined (__APPLE__)
-
-	struct {
-		struct rt_msghdr rt;
-		struct sockaddr_in dst;
-		struct sockaddr_in gate;
-		struct sockaddr_in mask;
-	} req;
-
-	int fd;
-	struct rt_msghdr *rtm;
-
-	if ((fd = socket(AF_ROUTE, SOCK_RAW, 0)) == -1) {
-		SYS_ERR(DTUN, LOGL_ERROR, errno, "socket() failed");
-		return -1;
-	}
-
-	memset(&req, 0x00, sizeof(req));
-
-	rtm = &req.rt;
-
-	rtm->rtm_msglen = sizeof(req);
-	rtm->rtm_version = RTM_VERSION;
-	if (delete) {
-		rtm->rtm_type = RTM_DELETE;
-	} else {
-		rtm->rtm_type = RTM_ADD;
-	}
-	rtm->rtm_flags = RTF_UP | RTF_GATEWAY | RTF_STATIC;	/* TODO */
-	rtm->rtm_addrs = RTA_DST | RTA_GATEWAY | RTA_NETMASK;
-	rtm->rtm_pid = getpid();
-	rtm->rtm_seq = 0044;	/* TODO */
-
-	req.dst.sin_family = AF_INET;
-	req.dst.sin_len = sizeof(req.dst);
-	req.mask.sin_family = AF_INET;
-	req.mask.sin_len = sizeof(req.mask);
-	req.gate.sin_family = AF_INET;
-	req.gate.sin_len = sizeof(req.gate);
-
-	req.dst.sin_addr.s_addr = dst->s_addr;
-	req.mask.sin_addr.s_addr = mask->s_addr;
-	req.gate.sin_addr.s_addr = gateway->s_addr;
-
-	if (write(fd, rtm, rtm->rtm_msglen) < 0) {
-		SYS_ERR(DTUN, LOGL_ERROR, errno, "write() failed");
-		close(fd);
-		return -1;
-	}
-	close(fd);
-	return 0;
-#endif
-
-}
-
-int tun_addroute(struct tun_t *this,
-		 struct in_addr *dst,
-		 struct in_addr *gateway, struct in_addr *mask)
-{
-	return tun_route(this, dst, gateway, mask, 0);
-}
-
-int tun_delroute(struct tun_t *this,
-		 struct in_addr *dst,
-		 struct in_addr *gateway, struct in_addr *mask)
-{
-	return tun_route(this, dst, gateway, mask, 1);
-}
-
-int tun_new(struct tun_t **tun, const char *dev_name)
+int tun_new(struct tun_t **tun, const char *dev_name, bool use_kernel, int fd0, int fd1u)
 {
 
 #if defined(__linux__)
@@ -606,31 +178,53 @@ int tun_new(struct tun_t **tun, const char *dev_name)
 	(*tun)->routes = 0;
 
 #if defined(__linux__)
-	/* Open the actual tun device */
-	if (((*tun)->fd = open("/dev/net/tun", O_RDWR)) < 0) {
-		SYS_ERR(DTUN, LOGL_ERROR, errno, "open() failed");
-		goto err_free;
+	if (!use_kernel) {
+		/* Open the actual tun device */
+		if (((*tun)->fd = open("/dev/net/tun", O_RDWR)) < 0) {
+			SYS_ERR(DTUN, LOGL_ERROR, errno, "open() failed");
+			goto err_free;
+		}
+
+		/* Set device flags. For some weird reason this is also the method
+		   used to obtain the network interface name */
+		memset(&ifr, 0, sizeof(ifr));
+		if (dev_name)
+			strcpy(ifr.ifr_name, dev_name);
+		ifr.ifr_flags = IFF_TUN | IFF_NO_PI;	/* Tun device, no packet info */
+		if (ioctl((*tun)->fd, TUNSETIFF, (void *)&ifr) < 0) {
+			SYS_ERR(DTUN, LOGL_ERROR, errno, "ioctl() failed");
+			goto err_close;
+		}
+
+		strncpy((*tun)->devname, ifr.ifr_name, IFNAMSIZ);
+		(*tun)->devname[IFNAMSIZ - 1] = 0;
+
+		/* Disable checksums */
+		if (ioctl((*tun)->fd, TUNSETNOCSUM, 1) < 0) {
+			SYS_ERR(DTUN, LOGL_NOTICE, errno, "could not disable checksum on %s", (*tun)->devname);
+		}
+		return 0;
+	} else {
+		strncpy((*tun)->devname, dev_name, IFNAMSIZ);
+		(*tun)->devname[IFNAMSIZ - 1] = 0;
+		(*tun)->fd = -1;
+
+		if (gtp_kernel_create(-1, dev_name, fd0, fd1u) < 0) {
+			LOGP(DTUN, LOGL_ERROR, "cannot create GTP tunnel device: %s\n",
+				strerror(errno));
+			return -1;
+		}
+		LOGP(DTUN, LOGL_NOTICE, "GTP kernel configured\n");
+		return 0;
 	}
 
-	/* Set device flags. For some weird reason this is also the method
-	   used to obtain the network interface name */
-	memset(&ifr, 0, sizeof(ifr));
-	if (dev_name)
-		strcpy(ifr.ifr_name, dev_name);
-	ifr.ifr_flags = IFF_TUN | IFF_NO_PI;	/* Tun device, no packet info */
-	if (ioctl((*tun)->fd, TUNSETIFF, (void *)&ifr) < 0) {
-		SYS_ERR(DTUN, LOGL_ERROR, errno, "ioctl() failed");
-		goto err_close;
-	}
-
-	strncpy((*tun)->devname, ifr.ifr_name, IFNAMSIZ);
-	(*tun)->devname[IFNAMSIZ - 1] = 0;
-
-	ioctl((*tun)->fd, TUNSETNOCSUM, 1);	/* Disable checksums */
-	return 0;
-
 #elif defined(__FreeBSD__) || defined (__APPLE__)
 
+	if (use_kernel) {
+		LOGP(DTUN, LOGL_ERROR, "No kernel GTP-U support in FreeBSD!\n");
+		return -1;
+	}
+
 	/* Find suitable device */
 	for (devnum = 0; devnum < 255; devnum++) {	/* TODO 255 */
 		snprintf(devname, sizeof(devname), "/dev/tun%d", devnum);
@@ -682,13 +276,17 @@ int tun_free(struct tun_t *tun)
 {
 
 	if (tun->routes) {
-		tun_delroute(tun, &tun->dstaddr, &tun->addr, &tun->netmask);
+		netdev_delroute4(&tun->dstaddr.v4, &tun->addr.v4, &tun->netmask);
 	}
 
-	if (close(tun->fd)) {
-		SYS_ERR(DTUN, LOGL_ERROR, errno, "close() failed");
+	if (tun->fd >= 0) {
+		if (close(tun->fd)) {
+			SYS_ERR(DTUN, LOGL_ERROR, errno, "close() failed");
+		}
 	}
 
+	gtp_kernel_stop(tun->devname);
+
 	/* TODO: For solaris we need to unlink streams */
 
 	free(tun);
@@ -720,7 +318,14 @@ int tun_decaps(struct tun_t *this)
 
 int tun_encaps(struct tun_t *tun, void *pack, unsigned len)
 {
-	return write(tun->fd, pack, len);
+	int rc;
+	rc = write(tun->fd, pack, len);
+	if (rc < 0) {
+		SYS_ERR(DTUN, LOGL_ERROR, errno, "TUN(%s): write() failed", tun->devname);
+	} else if (rc < len) {
+		LOGTUN(LOGL_ERROR, tun, "short write() %d < %u\n", rc, len);
+	}
+	return rc;
 }
 
 int tun_runscript(struct tun_t *tun, char *script)
@@ -731,7 +336,7 @@ int tun_runscript(struct tun_t *tun, char *script)
 	char smask[TUN_ADDRSIZE];
 	int rc;
 
-	strncpy(snet, inet_ntoa(tun->addr), sizeof(snet));
+	strncpy(snet, inet_ntoa(tun->addr.v4), sizeof(snet));
 	snet[sizeof(snet) - 1] = 0;
 	strncpy(smask, inet_ntoa(tun->netmask), sizeof(smask));
 	smask[sizeof(smask) - 1] = 0;
@@ -749,8 +354,6 @@ int tun_runscript(struct tun_t *tun, char *script)
 	return 0;
 }
 
-#include <ifaddrs.h>
-
 /*! Obtain the local address of the tun device.
  *  \param[in] tun Target device owning the IP
  *  \param[out] prefix_list List of prefix structures to fill with each IPv4/6 and prefix length found.
@@ -766,58 +369,5 @@ int tun_runscript(struct tun_t *tun, char *script)
  */
 int tun_ip_local_get(const struct tun_t *tun, struct in46_prefix *prefix_list, size_t prefix_size, int flags)
 {
-	static const uint8_t ll_prefix[] = { 0xfe,0x80, 0,0, 0,0, 0,0 };
-	struct ifaddrs *ifaddr, *ifa;
-	struct in46_addr netmask;
-	size_t count = 0;
-	bool is_ipv6_ll;
-
-	if (getifaddrs(&ifaddr) == -1) {
-		return -1;
-	}
-
-	for (ifa = ifaddr; ifa != NULL; ifa = ifa->ifa_next) {
-		if (ifa->ifa_addr == NULL)
-			continue;
-
-		if (strcmp(ifa->ifa_name, tun->devname))
-			continue;
-
-		if (ifa->ifa_addr->sa_family == AF_INET && (flags & IP_TYPE_IPv4)) {
-			struct sockaddr_in *sin4 = (struct sockaddr_in *) ifa->ifa_addr;
-			struct sockaddr_in *netmask4 = (struct sockaddr_in *) ifa->ifa_netmask;
-
-			if (count < prefix_size) {
-				netmask.len = sizeof(netmask4->sin_addr);
-				netmask.v4 = netmask4->sin_addr;
-				prefix_list[count].addr.len = sizeof(sin4->sin_addr);
-				prefix_list[count].addr.v4 = sin4->sin_addr;
-				prefix_list[count].prefixlen = in46a_netmasklen(&netmask);
-			}
-			count++;
-		}
-
-		if (ifa->ifa_addr->sa_family == AF_INET6 && (flags & IP_TYPE_IPv6)) {
-			struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *) ifa->ifa_addr;
-			struct sockaddr_in6 *netmask6 = (struct sockaddr_in6 *) ifa->ifa_netmask;
-
-			is_ipv6_ll = !memcmp(sin6->sin6_addr.s6_addr, ll_prefix, sizeof(ll_prefix));
-			if ((flags & IP_TYPE_IPv6_NONLINK) && is_ipv6_ll)
-				continue;
-			if ((flags & IP_TYPE_IPv6_LINK) && !is_ipv6_ll)
-				continue;
-
-			if (count < prefix_size) {
-				netmask.len = sizeof(netmask6->sin6_addr);
-				netmask.v6 = netmask6->sin6_addr;
-				prefix_list[count].addr.len = sizeof(sin6->sin6_addr);
-				prefix_list[count].addr.v6 = sin6->sin6_addr;
-				prefix_list[count].prefixlen = in46a_netmasklen(&netmask);
-			}
-			count++;
-		}
-	}
-
-	freeifaddrs(ifaddr);
-	return count;
+	return netdev_ip_local_get(tun->devname, prefix_list, prefix_size, flags);
 }

lib/tun.h

@@ -1,7 +1,7 @@
 /*
  * TUN interface functions.
  * Copyright (C) 2002, 2003 Mondru AB.
- * Copyright (C) 2017 by Harald Welte <laforge@gnumonks.org>
+ * Copyright (C) 2017-2018 by Harald Welte <laforge@gnumonks.org>
  *
  * The contents of this file may be used under the terms of the GNU
  * General Public License Version 2, provided that the above copyright
@@ -13,6 +13,7 @@
 #ifndef _TUN_H
 #define _TUN_H
 
+#include <stdbool.h>
 #include <net/if.h>
 
 #include "../lib/in46_addr.h"
@@ -20,43 +21,9 @@
 #define PACKET_MAX      8196	/* Maximum packet size we receive */
 #define TUN_SCRIPTSIZE   256
 #define TUN_ADDRSIZE     128
-#define TUN_NLBUFSIZE   1024
 
 #include "config.h"
-
-/* ipv6 ip type flags for tun_ipv6_local_get() */
-enum {
-	IP_TYPE_IPv4 = 1,
-	IP_TYPE_IPv6_LINK = 2,
-	IP_TYPE_IPv6_NONLINK = 4,
-};
-#define IP_TYPE_IPv6 (IP_TYPE_IPv6_LINK | IP_TYPE_IPv6_NONLINK)
-
-
-#ifndef HAVE_IPHDR
-struct iphdr
-  {
-#if __BYTE_ORDER == __LITTLE_ENDIAN
-    unsigned int ihl:4;
-    unsigned int version:4;
-#elif __BYTE_ORDER == __BIG_ENDIAN
-    unsigned int version:4;
-    unsigned int ihl:4;
-#else
-# error	"Please fix <bits/endian.h>"
-#endif
-    u_int8_t tos;
-    u_int16_t tot_len;
-    u_int16_t id;
-    u_int16_t frag_off;
-    u_int8_t ttl;
-    u_int8_t protocol;
-    u_int16_t check;
-    u_int32_t saddr;
-    u_int32_t daddr;
-    /*The options start here. */
-  };
-#endif /* !HAVE_IPHDR */
+#include "netdev.h"
 
 /* ***********************************************************
  * Information storage for each tun instance
@@ -64,8 +31,8 @@ struct iphdr
 
 struct tun_t {
 	int fd;			/* File descriptor to tun interface */
-	struct in_addr addr;
-	struct in_addr dstaddr;
+	struct in46_addr addr;
+	struct in46_addr dstaddr;
 	struct in_addr netmask;
 	int addrs;		/* Number of allocated IP addresses */
 	int routes;		/* One if we allocated an automatic route */
@@ -75,19 +42,13 @@ struct tun_t {
 	void *priv;
 };
 
-extern int tun_new(struct tun_t **tun, const char *dev_name);
+extern int tun_new(struct tun_t **tun, const char *dev_name, bool use_kernel, int fd0, int fd1u);
 extern int tun_free(struct tun_t *tun);
 extern int tun_decaps(struct tun_t *this);
 extern int tun_encaps(struct tun_t *tun, void *pack, unsigned len);
 
-extern int tun_addaddr(struct tun_t *this, struct in_addr *addr,
-		       struct in_addr *dstaddr, struct in_addr *netmask);
-
-extern int tun_setaddr(struct tun_t *this, struct in46_addr *our_adr,
-		       struct in46_addr *his_adr, size_t prefixlen);
-
-int tun_addroute(struct tun_t *this, struct in_addr *dst,
-		 struct in_addr *gateway, struct in_addr *mask);
+extern int tun_addaddr(struct tun_t *this, struct in46_addr *addr,
+		       struct in46_addr *dstaddr, size_t prefixlen);
 
 extern int tun_set_cb_ind(struct tun_t *this,
 			  int (*cb_ind) (struct tun_t * tun, void *pack,
@@ -98,4 +59,7 @@ extern int tun_runscript(struct tun_t *tun, char *script);
 int tun_ip_local_get(const struct tun_t *tun, struct in46_prefix *prefix_list,
 		     size_t prefix_size, int flags);
 
+#define LOGTUN(level, tun, fmt, args...) \
+	LOGP(DTUN, level, "TUN(%s): " fmt, (tun)->devname, ## args)
+
 #endif /* !_TUN_H */

lib/util.c

@@ -0,0 +1,35 @@
+/*
+ * misc helpers
+ * Copyright 2019 sysmocom - s.f.m.c. GmbH <info@sysmocom.de>
+ *
+ * The contents of this file may be used under the terms of the GNU
+ * General Public License Version 2, provided that the above copyright
+ * notice and this permission notice is included in all copies or
+ * substantial portions of the software.
+ *
+ */
+
+#include "../gtp/pdp.h"
+
+#include "ippool.h"
+#include "in46_addr.h"
+
+/*! Get the peer of pdp based on IP version used.
+*  \param[in] pdp PDP context to select the peer from.
+*  \param[in] v4v6 IP version to select. Valid values are 4 and 6.
+*  \returns The selected peer matching the given IP version. NULL if not present.
+*/
+struct ippoolm_t *pdp_get_peer_ipv(struct pdp_t *pdp, bool is_ipv6) {
+	uint8_t i;
+
+	for (i = 0; i < 2; i++) {
+		struct ippoolm_t * ippool = pdp->peer[i];
+		if (!ippool)
+			continue;
+		if (is_ipv6 && in46a_is_v6(&ippool->addr))
+			return ippool;
+		else if (!is_ipv6 && in46a_is_v4(&ippool->addr))
+			return ippool;
+	}
+	return NULL;
+}

lib/util.h

@@ -0,0 +1,18 @@
+#pragma once
+/*
+ * misc helpers
+ * Copyright 2019 sysmocom - s.f.m.c. GmbH <info@sysmocom.de>
+ *
+ * The contents of this file may be used under the terms of the GNU
+ * General Public License Version 2, provided that the above copyright
+ * notice and this permission notice is included in all copies or
+ * substantial portions of the software.
+ *
+ */
+
+#include <stdbool.h>
+
+struct ippoolm_t;
+struct pdp_t;
+
+struct ippoolm_t *pdp_get_peer_ipv(struct pdp_t *pdp, bool is_ipv6);

osmo-ggsn.spec.in

@@ -1,90 +0,0 @@
-Summary:   Osmocom Gateway GPRS Support Node (GGSN)
-Name:      @PACKAGE@
-Version:   @VERSION@
-Release:   1
-URL:       https://osmocom.org/projects/openggsn
-Source0:   http://prdownloads.sourceforge.net/ggsn/%{name}-%{version}.tar.gz
-License:   GPL
-Group:     System Environment/Daemons
-BuildRoot: %{_tmppath}/%{name}-root
-
-%description 
-OsmoGGSN is a Gateway GPRS Support Node (GGSN). It is used by mobile
-operators as the interface between the Internet and the rest of the
-mobile network infrastructure. The project also provides an SGSN
-emulator suitable for GPRS core network testing.
-
-%prep
-%setup -q
-
-%build
-
-./configure --prefix=/usr --enable-static-exec
-
-make
-
-%install
-
-make install prefix=$RPM_BUILD_ROOT/usr
-strip $RPM_BUILD_ROOT/usr/bin/osmo-ggsn
-strip $RPM_BUILD_ROOT/usr/bin/sgsnemu
-
-#Copy osmo-ggsn init script in place
-mkdir -p $RPM_BUILD_ROOT/etc/rc.d/init.d
-install -m755 examples/osmo-ggsn.init \
-	$RPM_BUILD_ROOT/etc/rc.d/init.d/osmo-ggsn
-
-#Copy osmo-ggsn.conf in place
-install -m755 examples/osmo-ggsn.cfg \
-	$RPM_BUILD_ROOT/etc/osmo-ggsn.cfg
-
-#Copy gsn_restart file in place
-mkdir -p $RPM_BUILD_ROOT/var/lib/osmo-ggsn
-echo "0" > $RPM_BUILD_ROOT/var/lib/osmo-ggsn/gsn_restart
-
-#Clean up unwanted library files
-rm -rf $RPM_BUILD_ROOT/usr/include/*
-rm -rf $RPM_BUILD_ROOT/usr/lib/*
-
-
-%clean
-rm -rf $RPM_BUILD_ROOT
-make clean
-
-%post
-/sbin/chkconfig --add osmo-ggsn
-
-%files
-%defattr(-,root,root)
-
-/usr/bin/osmo-ggsn
-/usr/bin/sgsnemu
-/etc/rc.d/init.d/osmo-ggsn
-%dir /var/lib/osmo-ggsn
-/var/lib/osmo-ggsn/gsn_restart
-
-%doc AUTHORS COPYING INSTALL NEWS README.md
-%doc examples/osmo-ggsn.conf
-%doc examples/sgsnemu.conf
-%doc examples/osmo-ggsn.init
-%doc examples/firewall
-%doc /usr/man/man8/osmo-ggsn.8.gz
-%doc /usr/man/man8/sgsnemu.8.gz
-
-%config /etc/osmo-ggsn.cfg
-
-
-#/usr/lib/libgtp.a
-#/usr/lib/libgtp.la
-#/usr/lib/libgtp.so
-#/usr/lib/libgtp.so.0
-#/usr/lib/libgtp.so.0.0.0
-
-
-
-%changelog
-* Mon Jun 30 2017  <laforge@gnumonks.org>
-- Update to OsmoGGSN
-
-* Mon Jun 30 2003  <jj@openggsn.org>
-- Initial build.

sgsnemu/Makefile.am

@@ -2,8 +2,23 @@ bin_PROGRAMS = sgsnemu
 
 AM_LDFLAGS = @EXEC_LDFLAGS@
 
-AM_CFLAGS = -O2 -D_GNU_SOURCE -fno-builtin -Wall -DSBINDIR='"$(sbindir)"' -ggdb $(LIBOSMOCORE_CFLAGS)
+AM_CFLAGS = \
+	    -D_GNU_SOURCE \
+	    -fno-builtin \
+	    -Wall \
+	    -DSBINDIR='"$(sbindir)"' \
+	    $(LIBOSMOCORE_CFLAGS) \
+	    $(NULL)
 
 sgsnemu_LDADD = @EXEC_LDADD@ -lgtp -L../gtp ../lib/libmisc.a $(LIBOSMOCORE_LIBS)
+
+if ENABLE_GTP_KERNEL
+AM_CFLAGS += \
+	     -DGTP_KERNEL \
+	     $(LIBGTPNL_CFLAGS) \
+	     $(NULL)
+sgsnemu_LDADD += $(LIBGTPNL_LIBS)
+endif
+
 sgsnemu_DEPENDENCIES = ../gtp/libgtp.la ../lib/libmisc.a
 sgsnemu_SOURCES = sgsnemu.c cmdline.c cmdline.h

sgsnemu/cmdline.c

@@ -41,7 +41,7 @@ const char *gengetopt_args_info_help[] = {
 	"      --pidfile=STRING          Filename of process id file\n                                  (default=`./sgsnemu.pid')",
 	"      --statedir=STRING         Directory of nonvolatile data  (default=`./')",
 	"      --dns=STRING              DNS Server to use",
-	"  -l, --listen=STRING           Local interface",
+	"  -l, --listen=STRING           Local host",
 	"  -r, --remote=STRING           Remote host",
 	"      --contexts=INT            Number of contexts  (default=`1')",
 	"      --timelimit=INT           Exit after timelimit seconds  (default=`0')",
@@ -65,12 +65,15 @@ const char *gengetopt_args_info_help[] = {
 	"      --charging=INT            Charging characteristics  (default=`0x0800')",
 	"  -u, --uid=STRING              Login user ID  (default=`mig')",
 	"  -p, --pwd=STRING              Login password  (default=`hemmelig')",
+	"\n Mode: createif\n  any option of this mode is related to tun interface, all payload going in and\n  out  via tunN interface",
 	"      --createif                Create local network interface  (default=off)",
 	"  -n, --net=STRING              Network address for local interface",
 	"      --defaultroute            Create default route  (default=off)",
 	"      --ipup=STRING             Script to run after link-up",
 	"      --ipdown=STRING           Script to run after link-down",
 	"      --tun-device=STRING       Name of the local network interface",
+	"      --netns=STRING            Network namespace to use",
+	"\n Mode: pinghost\n  generate ICMP payload inside G-PDU without setting up tun interface",
 	"      --pinghost=STRING         Ping remote host",
 	"      --pingrate=INT            Number of ping req per second  (default=`1')",
 	"      --pingsize=INT            Number of ping data bytes  (default=`56')",
@@ -161,6 +164,7 @@ void clear_given(struct gengetopt_args_info *args_info)
 	args_info->ipup_given = 0;
 	args_info->ipdown_given = 0;
 	args_info->tun_device_given = 0;
+	args_info->netns_given = 0;
 	args_info->pinghost_given = 0;
 	args_info->pingrate_given = 0;
 	args_info->pingsize_given = 0;
@@ -168,6 +172,8 @@ void clear_given(struct gengetopt_args_info *args_info)
 	args_info->pingquiet_given = 0;
 	args_info->no_tx_gpdu_seq_given = 0;
 	args_info->pdp_type_given = 0;
+	args_info->createif_mode_counter = 0;
+	args_info->pinghost_mode_counter = 0;
 }
 
 static
@@ -240,6 +246,8 @@ void clear_args(struct gengetopt_args_info *args_info)
 	args_info->ipdown_orig = NULL;
 	args_info->tun_device_arg = NULL;
 	args_info->tun_device_orig = NULL;
+	args_info->netns_arg = NULL;
+	args_info->netns_orig = NULL;
 	args_info->pinghost_arg = NULL;
 	args_info->pinghost_orig = NULL;
 	args_info->pingrate_arg = 1;
@@ -290,19 +298,20 @@ void init_args_info(struct gengetopt_args_info *args_info)
 	args_info->charging_help = gengetopt_args_info_help[28];
 	args_info->uid_help = gengetopt_args_info_help[29];
 	args_info->pwd_help = gengetopt_args_info_help[30];
-	args_info->createif_help = gengetopt_args_info_help[31];
-	args_info->net_help = gengetopt_args_info_help[32];
-	args_info->defaultroute_help = gengetopt_args_info_help[33];
-	args_info->ipup_help = gengetopt_args_info_help[34];
-	args_info->ipdown_help = gengetopt_args_info_help[35];
-	args_info->tun_device_help = gengetopt_args_info_help[36];
-	args_info->pinghost_help = gengetopt_args_info_help[37];
-	args_info->pingrate_help = gengetopt_args_info_help[38];
-	args_info->pingsize_help = gengetopt_args_info_help[39];
-	args_info->pingcount_help = gengetopt_args_info_help[40];
-	args_info->pingquiet_help = gengetopt_args_info_help[41];
-	args_info->no_tx_gpdu_seq_help = gengetopt_args_info_help[42];
-	args_info->pdp_type_help = gengetopt_args_info_help[43];
+	args_info->createif_help = gengetopt_args_info_help[32];
+	args_info->net_help = gengetopt_args_info_help[33];
+	args_info->defaultroute_help = gengetopt_args_info_help[34];
+	args_info->ipup_help = gengetopt_args_info_help[35];
+	args_info->ipdown_help = gengetopt_args_info_help[36];
+	args_info->tun_device_help = gengetopt_args_info_help[37];
+	args_info->netns_help = gengetopt_args_info_help[38];
+	args_info->pinghost_help = gengetopt_args_info_help[40];
+	args_info->pingrate_help = gengetopt_args_info_help[41];
+	args_info->pingsize_help = gengetopt_args_info_help[42];
+	args_info->pingcount_help = gengetopt_args_info_help[43];
+	args_info->pingquiet_help = gengetopt_args_info_help[44];
+	args_info->no_tx_gpdu_seq_help = gengetopt_args_info_help[45];
+	args_info->pdp_type_help = gengetopt_args_info_help[46];
 
 }
 
@@ -359,15 +368,6 @@ void cmdline_parser_params_init(struct cmdline_parser_params *params)
 	}
 }
 
-struct cmdline_parser_params *cmdline_parser_params_create(void)
-{
-	struct cmdline_parser_params *params =
-	    (struct cmdline_parser_params *)
-	    malloc(sizeof(struct cmdline_parser_params));
-	cmdline_parser_params_init(params);
-	return params;
-}
-
 static void free_string_field(char **s)
 {
 	if (*s) {
@@ -429,6 +429,8 @@ static void cmdline_parser_release(struct gengetopt_args_info *args_info)
 	free_string_field(&(args_info->ipdown_orig));
 	free_string_field(&(args_info->tun_device_arg));
 	free_string_field(&(args_info->tun_device_orig));
+	free_string_field(&(args_info->netns_arg));
+	free_string_field(&(args_info->netns_orig));
 	free_string_field(&(args_info->pinghost_arg));
 	free_string_field(&(args_info->pinghost_orig));
 	free_string_field(&(args_info->pingrate_orig));
@@ -542,6 +544,8 @@ int cmdline_parser_dump(FILE * outfile, struct gengetopt_args_info *args_info)
 	if (args_info->tun_device_given)
 		write_into_file(outfile, "tun-device",
 				args_info->tun_device_orig, 0);
+	if (args_info->netns_given)
+		write_into_file(outfile, "netns", args_info->netns_orig, 0);
 	if (args_info->pinghost_given)
 		write_into_file(outfile, "pinghost", args_info->pinghost_orig,
 				0);
@@ -706,6 +710,12 @@ cmdline_parser_required2(struct gengetopt_args_info *args_info,
 			prog_name, (additional_error ? additional_error : ""));
 		error_occurred = 1;
 	}
+	if (args_info->netns_given && !args_info->createif_given) {
+		fprintf(stderr,
+			"%s: '--netns' option depends on option 'createif'%s\n",
+			prog_name, (additional_error ? additional_error : ""));
+		error_occurred = 1;
+	}
 	if (args_info->pingrate_given && !args_info->pinghost_given) {
 		fprintf(stderr,
 			"%s: '--pingrate' option depends on option 'pinghost'%s\n",
@@ -853,6 +863,30 @@ int update_arg(void *field, char **orig_field,
 	return 0;		/* OK */
 }
 
+static int check_modes(int given1[], const char *options1[],
+		       int given2[], const char *options2[])
+{
+	int i = 0, j = 0, errors = 0;
+
+	while (given1[i] >= 0) {
+		if (given1[i]) {
+			while (given2[j] >= 0) {
+				if (given2[j]) {
+					++errors;
+					fprintf(stderr,
+						"%s: option %s conflicts with option %s\n",
+						package_name, options1[i],
+						options2[j]);
+				}
+				++j;
+			}
+		}
+		++i;
+	}
+
+	return errors;
+}
+
 int
 cmdline_parser_internal(int argc, char **argv,
 			struct gengetopt_args_info *args_info,
@@ -927,6 +961,7 @@ cmdline_parser_internal(int argc, char **argv,
 			{"ipup", 1, NULL, 0},
 			{"ipdown", 1, NULL, 0},
 			{"tun-device", 1, NULL, 0},
+			{"netns", 1, NULL, 0},
 			{"pinghost", 1, NULL, 0},
 			{"pingrate", 1, NULL, 0},
 			{"pingsize", 1, NULL, 0},
@@ -976,7 +1011,7 @@ cmdline_parser_internal(int argc, char **argv,
 				goto failure;
 
 			break;
-		case 'l':	/* Local interface.  */
+		case 'l':	/* Local host.  */
 
 			if (update_arg((void *)&(args_info->listen_arg),
 				       &(args_info->listen_orig),
@@ -1073,6 +1108,7 @@ cmdline_parser_internal(int argc, char **argv,
 
 			break;
 		case 'n':	/* Network address for local interface.  */
+			args_info->createif_mode_counter += 1;
 
 			if (update_arg((void *)&(args_info->net_arg),
 				       &(args_info->net_orig),
@@ -1391,6 +1427,7 @@ cmdline_parser_internal(int argc, char **argv,
 			else if (strcmp
 				 (long_options[option_index].name,
 				  "createif") == 0) {
+				args_info->createif_mode_counter += 1;
 
 				if (update_arg
 				    ((void *)&(args_info->createif_flag), 0,
@@ -1405,6 +1442,7 @@ cmdline_parser_internal(int argc, char **argv,
 			else if (strcmp
 				 (long_options[option_index].name,
 				  "defaultroute") == 0) {
+				args_info->createif_mode_counter += 1;
 
 				if (update_arg
 				    ((void *)&(args_info->defaultroute_flag), 0,
@@ -1419,6 +1457,7 @@ cmdline_parser_internal(int argc, char **argv,
 			/* Script to run after link-up.  */
 			else if (strcmp(long_options[option_index].name, "ipup")
 				 == 0) {
+				args_info->createif_mode_counter += 1;
 
 				if (update_arg((void *)&(args_info->ipup_arg),
 					       &(args_info->ipup_orig),
@@ -1434,6 +1473,7 @@ cmdline_parser_internal(int argc, char **argv,
 			else if (strcmp
 				 (long_options[option_index].name,
 				  "ipdown") == 0) {
+				args_info->createif_mode_counter += 1;
 
 				if (update_arg((void *)&(args_info->ipdown_arg),
 					       &(args_info->ipdown_orig),
@@ -1449,6 +1489,7 @@ cmdline_parser_internal(int argc, char **argv,
 			else if (strcmp
 				 (long_options[option_index].name,
 				  "tun-device") == 0) {
+				args_info->createif_mode_counter += 1;
 
 				if (update_arg
 				    ((void *)&(args_info->tun_device_arg),
@@ -1460,11 +1501,28 @@ cmdline_parser_internal(int argc, char **argv,
 				     additional_error))
 					goto failure;
 
+			}
+			/* Network namespace to use.  */
+			else if (strcmp
+				 (long_options[option_index].name,
+				  "netns") == 0) {
+				args_info->createif_mode_counter += 1;
+
+				if (update_arg((void *)&(args_info->netns_arg),
+					       &(args_info->netns_orig),
+					       &(args_info->netns_given),
+					       &(local_args_info.netns_given),
+					       optarg, 0, 0, ARG_STRING,
+					       check_ambiguity, override, 0, 0,
+					       "netns", '-', additional_error))
+					goto failure;
+
 			}
 			/* Ping remote host.  */
 			else if (strcmp
 				 (long_options[option_index].name,
 				  "pinghost") == 0) {
+				args_info->pinghost_mode_counter += 1;
 
 				if (update_arg
 				    ((void *)&(args_info->pinghost_arg),
@@ -1481,6 +1539,7 @@ cmdline_parser_internal(int argc, char **argv,
 			else if (strcmp
 				 (long_options[option_index].name,
 				  "pingrate") == 0) {
+				args_info->pinghost_mode_counter += 1;
 
 				if (update_arg
 				    ((void *)&(args_info->pingrate_arg),
@@ -1496,6 +1555,7 @@ cmdline_parser_internal(int argc, char **argv,
 			else if (strcmp
 				 (long_options[option_index].name,
 				  "pingsize") == 0) {
+				args_info->pinghost_mode_counter += 1;
 
 				if (update_arg
 				    ((void *)&(args_info->pingsize_arg),
@@ -1512,6 +1572,7 @@ cmdline_parser_internal(int argc, char **argv,
 			else if (strcmp
 				 (long_options[option_index].name,
 				  "pingcount") == 0) {
+				args_info->pinghost_mode_counter += 1;
 
 				if (update_arg
 				    ((void *)&(args_info->pingcount_arg),
@@ -1527,6 +1588,7 @@ cmdline_parser_internal(int argc, char **argv,
 			else if (strcmp
 				 (long_options[option_index].name,
 				  "pingquiet") == 0) {
+				args_info->pinghost_mode_counter += 1;
 
 				if (update_arg
 				    ((void *)&(args_info->pingquiet_flag), 0,
@@ -1566,6 +1628,32 @@ cmdline_parser_internal(int argc, char **argv,
 		}		/* switch */
 	}			/* while */
 
+	if (args_info->createif_mode_counter
+	    && args_info->pinghost_mode_counter) {
+		int createif_given[] =
+		    { args_info->createif_given, args_info->net_given,
+			args_info->defaultroute_given, args_info->ipup_given,
+			args_info->ipdown_given, args_info->tun_device_given,
+			args_info->netns_given, -1
+		};
+		const char *createif_desc[] =
+		    { "--createif", "--net", "--defaultroute", "--ipup",
+			"--ipdown", "--tun-device", "--netns", 0
+		};
+		int pinghost_given[] =
+		    { args_info->pinghost_given, args_info->pingrate_given,
+			args_info->pingsize_given, args_info->pingcount_given,
+			args_info->pingquiet_given, -1
+		};
+		const char *pinghost_desc[] =
+		    { "--pinghost", "--pingrate", "--pingsize", "--pingcount",
+			"--pingquiet", 0
+		};
+		error_occurred +=
+		    check_modes(createif_given, createif_desc, pinghost_given,
+				pinghost_desc);
+	}
+
 	if (check_required) {
 		error_occurred +=
 		    cmdline_parser_required2(args_info, argv[0],

sgsnemu/cmdline.ggo

@@ -15,6 +15,10 @@
 
 package "sgsnemu"
 
+defmode "createif"  modedesc="any option of this mode is related to tun interface, \
+all payload going in and out  via tunN interface"
+defmode "pinghost"  modedesc="generate ICMP payload inside G-PDU without setting up tun interface"
+
 option  "debug"        d "Run in debug mode"              flag   off
 
 option  "conf"         c "Read configuration file"        string no
@@ -22,7 +26,7 @@ option  "pidfile"      - "Filename of process id file"    string default="./sgsn
 option  "statedir"     - "Directory of nonvolatile data"  string default="./" no
 
 option  "dns"          - "DNS Server to use"              string no
-option  "listen"       l "Local interface"                string no
+option  "listen"       l "Local host"                     string no
 option  "remote"       r "Remote host"                    string no
 
 option  "contexts"     - "Number of contexts"             int    default="1" no
@@ -49,18 +53,19 @@ option  "charging"     - "Charging characteristics"       int    default="0x0800
 option  "uid"          u "Login user ID"                  string default="mig" no
 option  "pwd"          p "Login password"                 string default="hemmelig" no
 
-option  "createif"     - "Create local network interface" flag   off
-option  "net"          n "Network address for local interface" string dependon="createif" no
-option  "defaultroute" - "Create default route"           flag   dependon="createif" off
-option  "ipup"         - "Script to run after link-up"    string dependon="createif" no
-option  "ipdown"       - "Script to run after link-down"  string dependon="createif" no
-option  "tun-device"   - "Name of the local network interface" string dependon="createif" no
+modeoption   "createif"     - "Create local network interface" flag   off                           mode="createif"
+modeoption   "net"          n "Network address for local interface" string dependon="createif" no   mode="createif"
+modeoption   "defaultroute" - "Create default route"           flag   dependon="createif" off       mode="createif"
+modeoption   "ipup"         - "Script to run after link-up"    string dependon="createif" no        mode="createif"
+modeoption   "ipdown"       - "Script to run after link-down"  string dependon="createif" no        mode="createif"
+modeoption   "tun-device"   - "Name of the local network interface" string dependon="createif" no   mode="createif"
+modeoption   "netns"        - "Network namespace to use"       string dependon="createif" no        mode="createif"
 
-option  "pinghost"     - "Ping remote host"               string no
-option  "pingrate"     - "Number of ping req per second"  int default="1" dependon="pinghost" no
-option  "pingsize"     - "Number of ping data bytes"      int default="56" dependon="pinghost" no
-option  "pingcount"    - "Number of ping req to send"     int default="0" dependon="pinghost" no
-option  "pingquiet"    - "Do not print ping packet info"  flag dependon="pinghost" off
+modeoption   "pinghost"     - "Ping remote host"               string no                                mode="pinghost"
+modeoption   "pingrate"     - "Number of ping req per second"  int default="1" dependon="pinghost" no   mode="pinghost"
+modeoption   "pingsize"     - "Number of ping data bytes"      int default="56" dependon="pinghost" no  mode="pinghost"
+modeoption   "pingcount"    - "Number of ping req to send"     int default="0" dependon="pinghost" no   mode="pinghost"
+modeoption   "pingquiet"    - "Do not print ping packet info"  flag dependon="pinghost" off             mode="pinghost"
 
 option  "no-tx-gpdu-seq" - "Don't transmit G-PDU sequence nums"   flag    off
 option  "pdp-type"     t "PDP Type"                       string default="v4" no typestr="(v4|v6)"

sgsnemu/cmdline.h

@@ -69,11 +69,11 @@ extern "C" {
 		const char *dns_help;
 			/**< @brief DNS Server to use help description.  */
 		char *listen_arg;
-			/**< @brief Local interface.  */
+			/**< @brief Local host.  */
 		char *listen_orig;
-			/**< @brief Local interface original value given at command line.  */
+			/**< @brief Local host original value given at command line.  */
 		const char *listen_help;
-			   /**< @brief Local interface help description.  */
+			   /**< @brief Local host help description.  */
 		char *remote_arg;
 			/**< @brief Remote host.  */
 		char *remote_orig;
@@ -242,6 +242,12 @@ extern "C" {
 				/**< @brief Name of the local network interface original value given at command line.  */
 		const char *tun_device_help;
 			       /**< @brief Name of the local network interface help description.  */
+		char *netns_arg;
+			/**< @brief Network namespace to use.  */
+		char *netns_orig;
+			/**< @brief Network namespace to use original value given at command line.  */
+		const char *netns_help;
+			  /**< @brief Network namespace to use help description.  */
 		char *pinghost_arg;
 			/**< @brief Ping remote host.  */
 		char *pinghost_orig;
@@ -355,6 +361,8 @@ extern "C" {
 				/**< @brief Whether ipdown was given.  */
 		unsigned int tun_device_given;
 					/**< @brief Whether tun-device was given.  */
+		unsigned int netns_given;
+				/**< @brief Whether netns was given.  */
 		unsigned int pinghost_given;
 				/**< @brief Whether pinghost was given.  */
 		unsigned int pingrate_given;
@@ -370,6 +378,10 @@ extern "C" {
 		unsigned int pdp_type_given;
 				/**< @brief Whether pdp-type was given.  */
 
+		int createif_mode_counter;
+			     /**< @brief Counter for mode createif */
+		int pinghost_mode_counter;
+			     /**< @brief Counter for mode pinghost */
 	};
 
 /** @brief The additional parameters to pass to parser functions */
@@ -467,13 +479,6 @@ extern "C" {
  */
 	void cmdline_parser_params_init(struct cmdline_parser_params *params);
 
-/**
- * Allocates dynamically a cmdline_parser_params structure and initializes
- * all its fields to their default values
- * @return the created and initialized cmdline_parser_params structure
- */
-	struct cmdline_parser_params *cmdline_parser_params_create(void);
-
 /**
  * Initializes the passed gengetopt_args_info structure's fields
  * (also set default values for options that have a default)

tests/gtp/Makefile.am

@@ -1,19 +1,36 @@
-AM_CFLAGS = -Wall -I$(top_srcdir)/include $(LIBOSMOCORE_CFLAGS) -g
+AM_CFLAGS = \
+	    -Wall \
+	    -I$(top_srcdir)/include \
+	    $(LIBOSMOCORE_CFLAGS) \
+	    $(NULL)
+AM_LDFLAGS = -no-install
 
 EXTRA_DIST = \
 	gtpie_test.ok \
+	queue_test.ok \
 	$(NULL)
 
-noinst_PROGRAMS = \
+check_PROGRAMS = \
 	gtpie_test \
+	queue_test \
 	$(NULL)
 
 gtpie_test_SOURCES = \
 	gtpie_test.c \
 	$(NULL)
 
+queue_test_SOURCES = \
+	queue_test.c \
+	$(NULL)
+
 gtpie_test_LDADD = \
 	$(top_builddir)/lib/debug.o \
 	$(top_builddir)/gtp/libgtp.la \
 	$(LIBOSMOCORE_LIBS) \
 	$(NULL)
+
+queue_test_LDADD = \
+	$(top_builddir)/lib/debug.o \
+	$(top_builddir)/gtp/libgtp.la \
+	$(LIBOSMOCORE_LIBS) \
+	$(NULL)

tests/gtp/gtpie_test.c

@@ -7,6 +7,8 @@
 #include <osmocom/core/utils.h>
 #include <osmocom/core/application.h>
 #include <osmocom/core/logging.h>
+#include <osmocom/core/msgb.h>
+#include <osmocom/core/bits.h>
 
 #include "../../lib/syserr.h"
 #include "../../gtp/gtpie.h"
@@ -27,7 +29,7 @@ static void test_gtpie_tlv()
 	OSMO_ASSERT(rc == 0);
 	OSMO_ASSERT(len == sizeof(in) + 3);
 	OSMO_ASSERT(buf[0] == 23);
-	OSMO_ASSERT(ntohs(*(uint16_t *) &buf[1]) == sizeof(in));
+	OSMO_ASSERT(osmo_load16be(&buf[1]) == sizeof(in));
 	OSMO_ASSERT(!memcmp(buf+3, in, sizeof(in)));
 
 	/* overflow */
@@ -73,7 +75,7 @@ static void test_gtpie_tv2()
 	OSMO_ASSERT(rc == 0);
 	OSMO_ASSERT(len == 3);
 	OSMO_ASSERT(buf[0] == 42);
-	OSMO_ASSERT(ntohs(*(uint16_t *) &buf[1]) == 0xABCD);
+	OSMO_ASSERT(osmo_load16be(&buf[1]) == 0xABCD);
 }
 
 static void test_gtpie_tv4()
@@ -87,7 +89,7 @@ static void test_gtpie_tv4()
 	OSMO_ASSERT(rc == 0);
 	OSMO_ASSERT(len == 5);
 	OSMO_ASSERT(buf[0] == 42);
-	OSMO_ASSERT(ntohl(*(uint32_t *) &buf[1]) == 0xABCD0123);
+	OSMO_ASSERT(osmo_load32be(&buf[1]) == 0xABCD0123);
 }
 
 static void test_gtpie_tv8()
@@ -101,15 +103,17 @@ static void test_gtpie_tv8()
 	OSMO_ASSERT(rc == 0);
 	OSMO_ASSERT(len == 9);
 	OSMO_ASSERT(buf[0] == 42);
-	OSMO_ASSERT(ntohl(*(uint32_t *) &buf[1]) == 0x00010203);
-	OSMO_ASSERT(ntohl(*(uint32_t *) &buf[5]) == 0x04050607);
+	OSMO_ASSERT(osmo_load32be(&buf[1]) == 0x00010203);
+	OSMO_ASSERT(osmo_load32be(&buf[5]) == 0x04050607);
 }
 
 int main(int argc, char **argv)
 {
-	osmo_init_logging(&log_info);
+	void *tall_ctx = talloc_named_const(NULL, 1, "Root context");
+	msgb_talloc_ctx_init(tall_ctx, 0);
+	osmo_init_logging2(tall_ctx, &log_info);
 	log_set_use_color(osmo_stderr_target, 0);
-	log_set_print_filename(osmo_stderr_target, 0);
+	log_set_print_filename2(osmo_stderr_target, LOG_FILENAME_NONE);
 
 	srand(time(NULL));
 
@@ -124,4 +128,5 @@ int main(int argc, char **argv)
 	/* TODO: gtpie_encaps() */
 	/* TODO: gtpie_encaps2() */
 	/* TODO: gtpie_getie(), gtpie_exist(), gtpie_get*() */
+	return 0;
 }

tests/gtp/queue_test.c

@@ -0,0 +1,233 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <unistd.h>
+#include <inttypes.h>
+#include <arpa/inet.h>
+
+#include <osmocom/core/utils.h>
+#include <osmocom/core/application.h>
+#include <osmocom/core/logging.h>
+#include <osmocom/core/msgb.h>
+#include <osmocom/core/bits.h>
+
+#include "../../lib/syserr.h"
+#include "../../gtp/queue.h"
+
+static const struct qmsg_t qmsg_zero;
+
+static void queue_print(struct queue_t *queue, char* str)
+{
+	int n;
+	printf("=== [Queue %s] Next: %d First: %d Last: %d\n", str,
+		queue->next, queue->first, queue->last);
+	printf("#\tseq\tnext\tprev\ttimeout\tretrans\ttype\tcbp\n");
+	for (n = 0; n < QUEUE_SIZE; n++) {
+		if (queue->qmsga[n].state == 0) {
+			/* Nothing there, validate everything is zeroed */
+			OSMO_ASSERT(memcmp(&qmsg_zero, &queue->qmsga[n], sizeof(qmsg_zero)) == 0);
+			continue;
+		}
+		printf("%d\t%d\t%d\t%d\t%d\t%d\t%u\t%" PRIuPTR "\n",
+		       n,
+		       queue->qmsga[n].seq,
+		       queue->qmsga[n].next,
+		       queue->qmsga[n].prev,
+		       (int)queue->qmsga[n].timeout,
+		       queue->qmsga[n].retrans,
+		       queue->qmsga[n].type,
+		       (uintptr_t)queue->qmsga[n].cbp
+		);
+	}
+	printf("======================================================\n");
+}
+
+static void test_queue_empty()
+{
+	printf("***** Testing %s()\n", __func__);
+	struct queue_t *queue = NULL;
+	struct qmsg_t *qmsg = NULL;
+	uint16_t seq = 23;
+	uint8_t type = 0;
+	void *cbp = NULL;
+	struct sockaddr_in peer;
+	int rc;
+
+	rc = inet_pton(AF_INET, "127.0.0.1", &(peer.sin_addr));
+	OSMO_ASSERT(rc == 1);
+
+	rc = queue_new(&queue);
+	OSMO_ASSERT(rc == 0);
+	queue_print(queue, "created");
+
+	rc = queue_getfirst(queue, &qmsg);
+	OSMO_ASSERT(rc == EOF);
+	rc = queue_seqget(queue, &qmsg, &peer, seq);
+	OSMO_ASSERT(rc == EOF);
+	rc = queue_freemsg_seq(queue, &peer, seq, &type, &cbp);
+	OSMO_ASSERT(rc==EOF);
+
+	queue_print(queue, "pre-delete");
+	rc = queue_free(queue);
+	OSMO_ASSERT(rc == 0);
+}
+
+static void test_queue_one()
+{
+	printf("***** Testing %s()\n", __func__);
+	struct queue_t *queue = NULL;
+	struct qmsg_t *qmsg = NULL, *qmsg2 = NULL;;
+	uint16_t seq = 23;
+	uint8_t type = 0;
+	void *cbp = NULL;
+	struct sockaddr_in peer, peer2;
+	int rc;
+
+	rc = inet_pton(AF_INET, "127.0.0.1", &(peer.sin_addr));
+	OSMO_ASSERT(rc == 1);
+	rc = inet_pton(AF_INET, "127.0.0.2", &(peer2.sin_addr));
+	OSMO_ASSERT(rc == 1);
+
+	rc = queue_new(&queue);
+	OSMO_ASSERT(rc == 0);
+	queue_print(queue, "created");
+
+	rc = queue_newmsg(queue, &qmsg, &peer, seq);
+	OSMO_ASSERT(rc == 0);
+	queue_print(queue, "first added");
+	qmsg->type = GTP_ECHO_REQ;
+	qmsg->cbp = (void*) 0x13243546;
+	qmsg->seq = seq;
+
+	rc = queue_getfirst(queue, &qmsg2);
+	OSMO_ASSERT(rc == 0);
+	OSMO_ASSERT(qmsg == qmsg2);
+
+	rc = queue_seqget(queue, &qmsg2, &peer, seq);
+	OSMO_ASSERT(rc == 0);
+	OSMO_ASSERT(qmsg == qmsg2);
+	rc = queue_seqget(queue, &qmsg, &peer2, seq);
+	OSMO_ASSERT(rc == EOF);
+	rc = queue_seqget(queue, &qmsg, &peer, seq + 1);
+	OSMO_ASSERT(rc == EOF);
+	queue_print(queue, "after-get");
+
+	rc = queue_back(queue, qmsg);
+	OSMO_ASSERT(rc == 0);
+	queue_print(queue, "after-back");
+
+	rc = queue_freemsg_seq(queue, &peer2, seq, &type, &cbp);
+	OSMO_ASSERT(rc == EOF);
+	rc = queue_freemsg_seq(queue, &peer, seq + 1, &type, &cbp);
+	OSMO_ASSERT(rc == EOF);
+	queue_print(queue, "pree-freemsg");
+	rc = queue_freemsg_seq(queue, &peer, seq, &type, &cbp);
+	OSMO_ASSERT(rc == 0);
+	OSMO_ASSERT(type == GTP_ECHO_REQ);
+	OSMO_ASSERT(cbp == (void*)0x13243546);
+
+	queue_print(queue, "pre-delete");
+	rc = queue_free(queue);
+	OSMO_ASSERT(rc == 0);
+}
+
+#define newmsg_fill(queue, qmsg_ptr, peer_ptr, seq) \
+	do { \
+		int rc = queue_newmsg(queue, &(qmsg_ptr), peer_ptr, seq); \
+		OSMO_ASSERT(rc == 0); \
+		OSMO_ASSERT(qmsg_ptr); \
+		qmsg_ptr->type = GTP_CREATE_PDP_REQ; \
+		qmsg_ptr->cbp = (void*)(uintptr_t)seq; \
+	} while (0);
+
+#define freemsg_verify(seq, type, cbp) \
+	do { \
+		OSMO_ASSERT(type == GTP_CREATE_PDP_REQ); \
+		OSMO_ASSERT(cbp == (void*)(uintptr_t)seq); \
+	} while (0);
+
+static void test_queue_full()
+{
+	/* queue_newmsg until we receive EOF. Try moving back then. */
+	printf("***** Testing %s()\n", __func__);
+	struct queue_t *queue = NULL;
+	struct qmsg_t *qmsg = NULL;
+	uint8_t type = 0;
+	void *cbp = NULL;
+	struct sockaddr_in peer;
+	int rc;
+	int i;
+
+	rc = inet_pton(AF_INET, "127.0.0.1", &(peer.sin_addr));
+	OSMO_ASSERT(rc == 1);
+
+	rc = queue_new(&queue);
+	OSMO_ASSERT(rc == 0);
+	queue_print(queue, "created");
+
+	for (i = 0; i < QUEUE_SIZE - 1; i++) {
+		newmsg_fill(queue, qmsg, &peer, i);
+	}
+	queue_print(queue, "after-fill");
+
+	/* There's one slot left at the end, let's use first()->back() */
+	rc = queue_getfirst(queue, &qmsg);
+	OSMO_ASSERT(rc == 0);
+	rc = queue_back(queue, qmsg);
+	OSMO_ASSERT(rc == 0);
+	queue_print(queue, "after-back");
+
+	/* Now let's fill last empty slot */
+	newmsg_fill(queue, qmsg, &peer, QUEUE_SIZE - 1);
+	queue_print(queue, "after-full");
+
+	/* queue is now full, it should fail */
+	rc = queue_newmsg(queue, &qmsg, &peer, QUEUE_SIZE);
+	OSMO_ASSERT(rc == EOF);
+	queue_print(queue, "after-failing-full");
+
+	/* Remove 1before-last msg (the one moved back) and make sure we can
+	   re-add it at the end of the list */
+	rc = queue_seqget(queue, &qmsg, &peer, 0);
+	OSMO_ASSERT(rc == 0);
+	rc = queue_freemsg(queue, qmsg);
+	queue_print(queue, "after-freeing-0");
+	OSMO_ASSERT(rc == 0);
+	/* Now let's fill last empty slot which should be at the end */
+	newmsg_fill(queue, qmsg, &peer, 0);
+	queue_print(queue, "after-refilling-0");
+
+	/* Now free first half seq set in increasing order */
+	for (i = 0; i < QUEUE_SIZE / 2; i++) {
+		rc = queue_freemsg_seq(queue, &peer, i, &type, &cbp);
+		OSMO_ASSERT(rc == 0);
+		freemsg_verify(i, type, cbp);
+	}
+	queue_print(queue, "after-first-half-free");
+
+	/* Now free second half seq set in decreasing order */
+	for (i = QUEUE_SIZE - 1; i >= QUEUE_SIZE / 2; i--) {
+		rc = queue_freemsg_seq(queue, &peer, i, &type, &cbp);
+		OSMO_ASSERT(rc == 0);
+		freemsg_verify(i, type, cbp);
+	}
+	queue_print(queue, "after-second-half-free");
+
+	rc = queue_free(queue);
+	OSMO_ASSERT(rc == 0);
+}
+
+int main(int argc, char **argv)
+{
+	void *tall_ctx = talloc_named_const(NULL, 1, "Root context");
+	msgb_talloc_ctx_init(tall_ctx, 0);
+	osmo_init_logging2(tall_ctx, &log_info);
+	log_set_use_color(osmo_stderr_target, 0);
+	log_set_print_filename2(osmo_stderr_target, LOG_FILENAME_NONE);
+
+	test_queue_empty();
+	test_queue_one();
+	test_queue_full();
+
+	return 0;
+}

tests/lib/Makefile.am

@@ -1,9 +1,18 @@
-AM_CFLAGS = -Wall -I$(top_srcdir)/include $(LIBOSMOCORE_CFLAGS) -g
+AM_CFLAGS = \
+	    -Wall \
+	    -I$(top_srcdir)/include \
+	    $(LIBOSMOCORE_CFLAGS) \
+	    $(NULL)
+AM_LDFLAGS = -no-install
 
-EXTRA_DIST = ippool_test.ok ippool_test.err \
-	     in46a_test.ok
+EXTRA_DIST = ippool_test.ok \
+	     ippool_test.err \
+	     ippool_v6_test.ok \
+	     ippool_v6_test.err \
+	     in46a_test.ok \
+	     in46a_v6_test.ok
 
-noinst_PROGRAMS = ippool_test in46a_test
+check_PROGRAMS = ippool_test in46a_test
 
 ippool_test_SOURCES = \
 	ippool_test.c \

tests/lib/in46a_test.c

@@ -10,6 +10,8 @@
 #include <osmocom/core/utils.h>
 #include <osmocom/core/application.h>
 #include <osmocom/core/logging.h>
+#include <osmocom/core/msgb.h>
+#include <osmocom/core/bits.h>
 
 #include "../../lib/in46_addr.h"
 #include "../../lib/syserr.h"
@@ -19,41 +21,24 @@ static const struct in46_addr g_ia4 = {
 	.v4.s_addr = 0x0d0c0b0a,
 };
 
-static const struct in46_addr g_ia6 = {
-	.len = 16,
-	.v6.s6_addr = { 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16 },
-};
-
 static void test_in46a_to_af(void)
 {
-	struct in46_addr ia;
-
-	printf("Testing in46a_to_af()\n");
+	printf("Testing in46a_to_af() with IPv4 addresses\n");
 
 	OSMO_ASSERT(in46a_to_af(&g_ia4) == AF_INET);
-	OSMO_ASSERT(in46a_to_af(&g_ia6) == AF_INET6);
-
-	ia.len = 8;
-	OSMO_ASSERT(in46a_to_af(&ia) == AF_INET6);
 }
 
 static void test_in46a_to_sas(void)
 {
 	struct sockaddr_storage ss;
 	struct sockaddr_in *sin = (struct sockaddr_in *) &ss;
-	struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *) &ss;
 
-	printf("Testing in46a_to_sas()\n");
+	printf("Testing in46a_to_sas() with IPv4 addresses\n");
 
 	memset(&ss, 0, sizeof(ss));
 	OSMO_ASSERT(in46a_to_sas(&ss, &g_ia4) == 0);
 	OSMO_ASSERT(sin->sin_family == AF_INET);
 	OSMO_ASSERT(sin->sin_addr.s_addr == g_ia4.v4.s_addr);
-
-	memset(&ss, 0, sizeof(ss));
-	OSMO_ASSERT(in46a_to_sas(&ss, &g_ia6) == 0);
-	OSMO_ASSERT(sin6->sin6_family == AF_INET6);
-	OSMO_ASSERT(!memcmp(&sin6->sin6_addr, &g_ia6.v6, sizeof(sin6->sin6_addr)));
 }
 
 static void test_in46a_ntop(void)
@@ -62,7 +47,7 @@ static void test_in46a_ntop(void)
 	char buf[256];
 	const char *res;
 
-	printf("Testing in46a_ntop()\n");
+	printf("Testing in46a_ntop() with IPv4 addresses\n");
 
 	res = in46a_ntop(NULL, buf, sizeof(buf));
 	OSMO_ASSERT(res && !strcmp(res, "UNDEFINED"));
@@ -78,10 +63,6 @@ static void test_in46a_ntop(void)
 	res = in46a_ntop(&ia, buf, sizeof(buf));
 	OSMO_ASSERT(res && !strcmp(res, "1.2.3.4"));
 	printf("res = %s\n", res);
-
-	res = in46a_ntop(&g_ia6, buf, sizeof(buf));
-	OSMO_ASSERT(res && !strcmp(res, "102:304:506:708:90a:b0c:d0e:f10"));
-	printf("res = %s\n", res);
 }
 
 static void test_in46p_ntoa(void)
@@ -100,21 +81,14 @@ static void test_in46a_equal(void)
 {
 	struct in46_addr b;
 
-	printf("Testing in46a_equal()\n");
+	printf("Testing in46a_equal() with IPv4 addresses\n");
 
 	memset(&b, 0xff, sizeof(b));
 	b.len = g_ia4.len;
 	b.v4.s_addr = g_ia4.v4.s_addr;
 	OSMO_ASSERT(in46a_equal(&g_ia4, &b));
-
-	memset(&b, 0xff, sizeof(b));
-	b.len = g_ia6.len;
-	b.v6 = g_ia6.v6;
-	OSMO_ASSERT(in46a_equal(&g_ia6, &b));
-
 }
 
-
 static int log_in46a_within_mask(const struct in46_addr *addr, const struct in46_addr *net,
 				 size_t prefixlen)
 {
@@ -133,7 +107,7 @@ static void test_in46a_within_mask(void)
 {
 	struct in46_addr addr, mask;
 
-	printf("Testing in46a_within_mask()\n");
+	printf("Testing in46a_within_mask() with IPv4 addresses\n");
 
 	addr = g_ia4;
 	mask = g_ia4;
@@ -154,13 +128,9 @@ static void test_in46a_within_mask(void)
 
 static void test_in46a_to_eua(void)
 {
-	const struct in46_addr ia_v6_8 = {
-		.len = 8,
-		.v6.s6_addr = { 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16 },
-	};
 	struct ul66_t eua;
 
-	printf("testing in46a_to_eua()\n");
+	printf("testing in46a_to_eua() with IPv4 addresses\n");
 
 #if 0	/* triggers assert in current implementation */
 	const struct in46_addr ia_invalid = { .len = 3, };
@@ -168,82 +138,52 @@ static void test_in46a_to_eua(void)
 #endif
 
 	/* IPv4 address */
-	OSMO_ASSERT(in46a_to_eua(&g_ia4, &eua) == 0);
+	OSMO_ASSERT(in46a_to_eua(&g_ia4, 1, &eua) == 0);
 	OSMO_ASSERT(eua.v[0] == PDP_EUA_ORG_IETF);
 	OSMO_ASSERT(eua.v[1] == PDP_EUA_TYPE_v4);
-	OSMO_ASSERT(*(uint32_t *) &eua.v[2] == g_ia4.v4.s_addr);
-
-	/* IPv6 address */
-	OSMO_ASSERT(in46a_to_eua(&g_ia6, &eua) == 0);
-	OSMO_ASSERT(eua.v[0] == PDP_EUA_ORG_IETF);
-	OSMO_ASSERT(eua.v[1] == PDP_EUA_TYPE_v6);
-	OSMO_ASSERT(!memcmp(&eua.v[2], &g_ia6.v6, 16));
-
-	/* IPv6 address with prefix / length 8 */
-	OSMO_ASSERT(in46a_to_eua(&ia_v6_8, &eua) == 0);
-	OSMO_ASSERT(eua.v[0] == PDP_EUA_ORG_IETF);
-	OSMO_ASSERT(eua.v[1] == PDP_EUA_TYPE_v6);
-	OSMO_ASSERT(!memcmp(&eua.v[2], &ia_v6_8.v6, 16));
+	OSMO_ASSERT(osmo_load32le(&eua.v[2]) == g_ia4.v4.s_addr);
 }
 
 static void test_in46a_from_eua(void)
 {
-	struct in46_addr ia;
+	struct in46_addr ia[2];
 	struct ul66_t eua;
 	const uint8_t v4_unspec[] = { PDP_EUA_ORG_IETF, PDP_EUA_TYPE_v4 };
 	const uint8_t v4_spec[] = { PDP_EUA_ORG_IETF, PDP_EUA_TYPE_v4, 1,2,3,4 };
-	const uint8_t v6_unspec[] = { PDP_EUA_ORG_IETF, PDP_EUA_TYPE_v6 };
-	const uint8_t v6_spec[] = { PDP_EUA_ORG_IETF, PDP_EUA_TYPE_v6,
-				    1,2,3,4,5,6,7,8,9,0xa,0xb,0xc,0xd,0xe,0xf,0x10 };
-
 	memset(&eua, 0, sizeof(eua));
 
-	printf("Testing in46a_from_eua()\n");
+	printf("Testing in46a_from_eua() with IPv4 addresses\n");
 
 	/* default: v4 unspec */
-	OSMO_ASSERT(in46a_from_eua(&eua, &ia) == 0);
-	OSMO_ASSERT(ia.len == 4);
-	OSMO_ASSERT(ia.v4.s_addr == 0);
+	OSMO_ASSERT(in46a_from_eua(&eua, ia) == 1);
+	OSMO_ASSERT(ia[0].len == 4);
+	OSMO_ASSERT(ia[0].v4.s_addr == 0);
 
 	/* invalid */
 	eua.v[0] = 0x23;
 	eua.v[1] = PDP_EUA_TYPE_v4;
 	eua.l = 6;
-	OSMO_ASSERT(in46a_from_eua(&eua, &ia) < 0);
+	OSMO_ASSERT(in46a_from_eua(&eua, ia) < 0);
 
 	/* invalid */
 	eua.v[0] = PDP_EUA_ORG_IETF;
 	eua.v[1] = 0x23;
 	eua.l = 6;
-	OSMO_ASSERT(in46a_from_eua(&eua, &ia) < 0);
+	OSMO_ASSERT(in46a_from_eua(&eua, ia) < 0);
 
 	/* unspecified V4 */
 	memcpy(eua.v, v4_unspec, sizeof(v4_unspec));
 	eua.l = sizeof(v4_unspec);
-	OSMO_ASSERT(in46a_from_eua(&eua, &ia) == 0);
-	OSMO_ASSERT(ia.len == 4);
-	OSMO_ASSERT(ia.v4.s_addr == 0);
+	OSMO_ASSERT(in46a_from_eua(&eua, ia) == 1);
+	OSMO_ASSERT(ia[0].len == 4);
+	OSMO_ASSERT(ia[0].v4.s_addr == 0);
 
 	/* specified V4 */
 	memcpy(eua.v, v4_spec, sizeof(v4_spec));
 	eua.l = sizeof(v4_spec);
-	OSMO_ASSERT(in46a_from_eua(&eua, &ia) == 0);
-	OSMO_ASSERT(ia.len == 4);
-	OSMO_ASSERT(ia.v4.s_addr == htonl(0x01020304));
-
-	/* unspecified V6 */
-	memcpy(eua.v, v6_unspec, sizeof(v6_unspec));
-	eua.l = sizeof(v6_unspec);
-	OSMO_ASSERT(in46a_from_eua(&eua, &ia) == 0);
-	OSMO_ASSERT(ia.len == 16);
-	OSMO_ASSERT(IN6_IS_ADDR_UNSPECIFIED(&ia.v6));
-
-	/* specified V6 */
-	memcpy(eua.v, v6_spec, sizeof(v6_spec));
-	eua.l = sizeof(v6_spec);
-	OSMO_ASSERT(in46a_from_eua(&eua, &ia) == 0);
-	OSMO_ASSERT(ia.len == 16);
-	OSMO_ASSERT(!memcmp(&ia.v6, v6_spec+2, ia.len));
+	OSMO_ASSERT(in46a_from_eua(&eua, ia) == 1);
+	OSMO_ASSERT(ia[0].len == 4);
+	OSMO_ASSERT(ia[0].v4.s_addr == htonl(0x01020304));
 }
 
 static void test_in46a_netmasklen(void)
@@ -273,7 +213,188 @@ static void test_in46a_netmasklen(void)
 	netmask.v4.s_addr = 0x00000000;
 	len = in46a_netmasklen(&netmask);
 	OSMO_ASSERT(len == 0);
+}
 
+/* IPv6 specific tests */
+
+static const struct in46_addr g_ia6 = {
+	.len = 16,
+	.v6.s6_addr = { 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16 },
+};
+
+static void test_in46a_to_af_v6(void)
+{
+	struct in46_addr ia;
+
+	printf("Testing in46a_to_af() with IPv6 addresses\n");
+
+	OSMO_ASSERT(in46a_to_af(&g_ia6) == AF_INET6);
+
+	ia.len = 8;
+	OSMO_ASSERT(in46a_to_af(&ia) == AF_INET6);
+}
+
+static void test_in46a_to_sas_v6(void)
+{
+	struct sockaddr_storage ss;
+	struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *) &ss;
+
+	printf("Testing in46a_to_sas() with IPv6 addresses\n");
+
+	memset(&ss, 0, sizeof(ss));
+	OSMO_ASSERT(in46a_to_sas(&ss, &g_ia6) == 0);
+	OSMO_ASSERT(sin6->sin6_family == AF_INET6);
+	OSMO_ASSERT(!memcmp(&sin6->sin6_addr, &g_ia6.v6, sizeof(sin6->sin6_addr)));
+}
+
+static void test_in46a_ntop_v6(void)
+{
+	char buf[256];
+	const char *res;
+
+	printf("Testing in46a_ntop() with IPv6 addresses\n");
+
+	res = in46a_ntop(&g_ia6, buf, sizeof(buf));
+	OSMO_ASSERT(res && !strcmp(res, "102:304:506:708:90a:b0c:d0e:f10"));
+	printf("res = %s\n", res);
+}
+
+static void test_in46a_equal_v6(void)
+{
+	struct in46_addr b;
+
+	printf("Testing in46a_equal() with IPv6 addresses\n");
+
+	memset(&b, 0xff, sizeof(b));
+	b.len = g_ia6.len;
+	b.v6 = g_ia6.v6;
+	OSMO_ASSERT(in46a_equal(&g_ia6, &b));
+}
+
+static void test_in46a_to_eua_v6(void)
+{
+	const struct in46_addr ia_v6_8 = {
+		.len = 8,
+		.v6.s6_addr = { 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16 },
+	};
+	struct ul66_t eua;
+
+	printf("Testing in46a_to_eua() with IPv6 addresses\n");
+
+	/* IPv6 address */
+	OSMO_ASSERT(in46a_to_eua(&g_ia6, 1, &eua) == 0);
+	OSMO_ASSERT(eua.v[0] == PDP_EUA_ORG_IETF);
+	OSMO_ASSERT(eua.v[1] == PDP_EUA_TYPE_v6);
+	OSMO_ASSERT(!memcmp(&eua.v[2], &g_ia6.v6, 16));
+
+	/* IPv6 address with prefix / length 8 */
+	OSMO_ASSERT(in46a_to_eua(&ia_v6_8, 1, &eua) == 0);
+	OSMO_ASSERT(eua.v[0] == PDP_EUA_ORG_IETF);
+	OSMO_ASSERT(eua.v[1] == PDP_EUA_TYPE_v6);
+	OSMO_ASSERT(!memcmp(&eua.v[2], &ia_v6_8.v6, 16));
+}
+
+static void test_in46a_to_eua_v4v6() {
+	const struct in46_addr ia_v4v6[2] = {
+		{
+		.len = 16,
+		.v6.s6_addr = { 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16 },
+		},
+		{
+		.len = 4,
+		.v4.s_addr = 0x0d0c0b0a,
+		}
+	};
+	struct ul66_t eua;
+	printf("Testing in46a_to_eua() with IPv4v6 addresses\n");
+
+	/* IPv4 address */
+	OSMO_ASSERT(in46a_to_eua(ia_v4v6, 2, &eua) == 0);
+	OSMO_ASSERT(eua.v[0] == PDP_EUA_ORG_IETF);
+	OSMO_ASSERT(eua.v[1] == PDP_EUA_TYPE_v4v6);
+	OSMO_ASSERT(osmo_load32le(&eua.v[2]) == g_ia4.v4.s_addr);
+	OSMO_ASSERT(!memcmp(&eua.v[6], &g_ia6.v6, 16));
+}
+
+static void test_in46a_from_eua_v6(void)
+{
+	struct in46_addr ia[2];
+	struct ul66_t eua;
+	const uint8_t v6_unspec[] = { PDP_EUA_ORG_IETF, PDP_EUA_TYPE_v6 };
+	const uint8_t v6_spec[] = { PDP_EUA_ORG_IETF, PDP_EUA_TYPE_v6,
+				    1,2,3,4,5,6,7,8,9,0xa,0xb,0xc,0xd,0xe,0xf,0x10 };
+
+	memset(&eua, 0, sizeof(eua));
+
+	printf("Testing in46a_from_eua() with IPv6 addresses\n");
+
+	/* unspecified V6 */
+	memcpy(eua.v, v6_unspec, sizeof(v6_unspec));
+	eua.l = sizeof(v6_unspec);
+	OSMO_ASSERT(in46a_from_eua(&eua, ia) == 1);
+	OSMO_ASSERT(ia[0].len == 16);
+	OSMO_ASSERT(IN6_IS_ADDR_UNSPECIFIED(&ia[0].v6));
+
+	/* specified V6 */
+	memcpy(eua.v, v6_spec, sizeof(v6_spec));
+	eua.l = sizeof(v6_spec);
+	OSMO_ASSERT(in46a_from_eua(&eua, ia) == 1);
+	OSMO_ASSERT(ia[0].len == 16);
+	OSMO_ASSERT(!memcmp(&ia[0].v6, v6_spec+2, ia[0].len));
+}
+
+static void test_in46a_from_eua_v4v6(void) {
+	struct in46_addr ia[2];
+	struct ul66_t eua;
+	const uint8_t v4_unspec_v6_unspec[] = { PDP_EUA_ORG_IETF, PDP_EUA_TYPE_v4v6 };
+	const uint8_t v4_spec_v6_unspec[] = { PDP_EUA_ORG_IETF, PDP_EUA_TYPE_v4v6, 1,2,3,4 };
+	const uint8_t v4_unspec_v6_spec[] = { PDP_EUA_ORG_IETF, PDP_EUA_TYPE_v4v6, 1,2,3,4,5,6,7,8,9,0xa,0xb,0xc,0xd,0xe,0xf,0x10 };
+	const uint8_t v4_spec_v6_spec[] = { PDP_EUA_ORG_IETF, PDP_EUA_TYPE_v4v6, 1,2,3,4, 1,2,3,4,5,6,7,8,9,0xa,0xb,0xc,0xd,0xe,0xf,0x10 };
+
+	memset(&eua, 0, sizeof(eua));
+
+	printf("Testing in46a_from_eua() with IPv4v6 addresses\n");
+
+	/* unspecified V4 & V6 */
+	memcpy(eua.v, v4_unspec_v6_unspec, sizeof(v4_unspec_v6_unspec));
+	eua.l = sizeof(v4_unspec_v6_unspec);
+	OSMO_ASSERT(in46a_from_eua(&eua, ia) == 2);
+	OSMO_ASSERT(ia[0].len == 4);
+	OSMO_ASSERT(ia[1].len == 16);
+	OSMO_ASSERT(ia[0].v4.s_addr == 0);
+	OSMO_ASSERT(IN6_IS_ADDR_UNSPECIFIED(&ia[1].v6));
+
+	/* specified V4, unspecified V6 */
+	memcpy(eua.v, v4_spec_v6_unspec, sizeof(v4_spec_v6_unspec));
+	eua.l = sizeof(v4_spec_v6_unspec);
+	OSMO_ASSERT(in46a_from_eua(&eua, ia) == 2);
+	OSMO_ASSERT(ia[0].len == 4);
+	OSMO_ASSERT(ia[1].len == 16);
+	OSMO_ASSERT(ia[0].v4.s_addr == htonl(0x01020304));
+	OSMO_ASSERT(IN6_IS_ADDR_UNSPECIFIED(&ia[1].v6));
+
+	/* unspecified V4, specified V6 */
+	memcpy(eua.v, v4_unspec_v6_spec, sizeof(v4_unspec_v6_spec));
+	eua.l = sizeof(v4_unspec_v6_spec);
+	OSMO_ASSERT(in46a_from_eua(&eua, ia) == 2);
+	OSMO_ASSERT(ia[0].len == 4);
+	OSMO_ASSERT(ia[1].len == 16);
+	OSMO_ASSERT(ia[0].v4.s_addr == 0);
+	OSMO_ASSERT(!memcmp(&ia[1].v6, v4_unspec_v6_spec+2, ia[1].len));
+
+	/* specified V4, specified V6 */
+	memcpy(eua.v, v4_spec_v6_spec, sizeof(v4_spec_v6_spec));
+	eua.l = sizeof(v4_spec_v6_spec);
+	OSMO_ASSERT(in46a_from_eua(&eua, ia) == 2);
+	OSMO_ASSERT(ia[0].len == 4);
+	OSMO_ASSERT(ia[1].len == 16);
+	OSMO_ASSERT(ia[0].v4.s_addr == htonl(0x01020304));
+	OSMO_ASSERT(!memcmp(&ia[1].v6, v4_spec_v6_spec+6, ia[1].len));
+}
+
+static void test_in46a_netmasklen_v6(void)
+{
+	unsigned int len;
 	printf("Testing in46a_netmasklen() with IPv6 addresses\n");
 	const struct in46_addr netmaskA = {
 		.len = 16,
@@ -306,19 +427,34 @@ static void test_in46a_netmasklen(void)
 
 int main(int argc, char **argv)
 {
-	osmo_init_logging(&log_info);
+	void *tall_ctx = talloc_named_const(NULL, 1, "Root context");
+	msgb_talloc_ctx_init(tall_ctx, 0);
+	osmo_init_logging2(tall_ctx, &log_info);
 	log_set_use_color(osmo_stderr_target, 0);
-	log_set_print_filename(osmo_stderr_target, 0);
+	log_set_print_filename2(osmo_stderr_target, LOG_FILENAME_NONE);
 
 	srand(time(NULL));
 
-	test_in46a_to_af();
-	test_in46a_to_sas();
-	test_in46a_ntop();
-	test_in46p_ntoa();
-	test_in46a_equal();
-	test_in46a_within_mask();
-	test_in46a_to_eua();
-	test_in46a_from_eua();
-	test_in46a_netmasklen();
+	if (argc < 2 || strcmp(argv[1], "-v6")) {
+		test_in46a_to_af();
+		test_in46a_to_sas();
+		test_in46a_ntop();
+		test_in46p_ntoa();
+		test_in46a_equal();
+		test_in46a_within_mask();
+		test_in46a_to_eua();
+		test_in46a_from_eua();
+		test_in46a_netmasklen();
+	} else {
+		test_in46a_to_af_v6();
+		test_in46a_to_sas_v6();
+		test_in46a_ntop_v6();
+		test_in46a_equal_v6();
+		test_in46a_to_eua_v6();
+		test_in46a_from_eua_v6();
+		test_in46a_to_eua_v4v6();
+		test_in46a_from_eua_v4v6();
+		test_in46a_netmasklen_v6();
+	}
+	return 0;
 }

tests/lib/in46a_test.ok

@@ -1,19 +1,17 @@
-Testing in46a_to_af()
-Testing in46a_to_sas()
-Testing in46a_ntop()
+Testing in46a_to_af() with IPv4 addresses
+Testing in46a_to_sas() with IPv4 addresses
+Testing in46a_ntop() with IPv4 addresses
 res = UNDEFINED
 res = UNDEFINED
 res = 1.2.3.4
-res = 102:304:506:708:90a:b0c:d0e:f10
 in46p_ntoa() returns 16.32.48.0/24
-Testing in46a_equal()
-Testing in46a_within_mask()
+Testing in46a_equal() with IPv4 addresses
+Testing in46a_within_mask() with IPv4 addresses
 in46a_within_mask(10.11.12.13, 10.11.12.13, 32) = 1
 in46a_within_mask(10.11.12.13, 10.11.12.12, 30) = 1
 in46a_within_mask(10.11.12.13, 10.8.0.0, 13) = 1
 in46a_within_mask(10.11.12.14, 10.11.12.13, 32) = 0
 in46a_within_mask(10.11.12.14, 10.11.12.12, 30) = 1
-testing in46a_to_eua()
-Testing in46a_from_eua()
+testing in46a_to_eua() with IPv4 addresses
+Testing in46a_from_eua() with IPv4 addresses
 Testing in46a_netmasklen() with IPv4 addresses
-Testing in46a_netmasklen() with IPv6 addresses

tests/lib/in46a_v6_test.ok

@@ -0,0 +1,10 @@
+Testing in46a_to_af() with IPv6 addresses
+Testing in46a_to_sas() with IPv6 addresses
+Testing in46a_ntop() with IPv6 addresses
+res = 102:304:506:708:90a:b0c:d0e:f10
+Testing in46a_equal() with IPv6 addresses
+Testing in46a_to_eua() with IPv6 addresses
+Testing in46a_from_eua() with IPv6 addresses
+Testing in46a_to_eua() with IPv4v6 addresses
+Testing in46a_from_eua() with IPv4v6 addresses
+Testing in46a_netmasklen() with IPv6 addresses

tests/lib/ippool_test.c

@@ -7,6 +7,7 @@
 #include <osmocom/core/utils.h>
 #include <osmocom/core/application.h>
 #include <osmocom/core/logging.h>
+#include <osmocom/core/msgb.h>
 
 #include "../../lib/in46_addr.h"
 #include "../../lib/ippool.h"
@@ -113,21 +114,33 @@ static void test_pool_sizes(void)
 	/* 65534 addresses [0.1..255.254] */
 	test_pool_size("192.168.0.0/16", IPPOOL_NONETWORK | IPPOOL_NOBROADCAST, NULL, 0, 65534);
 
-	/* 256 prefixes of /64 each */
-	test_pool_size("2001:DB8::/56", 0, NULL, 0, 256);
-
 	/* 253 addresses [1..254] & exclude 192.168.23.1/24 */
 	char *blacklist[] = {"176.16.222.10/24", "192.168.23.1/24", "192.168.38.2/24"};
 	test_pool_size("192.168.23.0/24", IPPOOL_NONETWORK | IPPOOL_NOBROADCAST, blacklist, 3, 253);
 }
 
+static void test_pool_sizes_v6(void)
+{
+	/* 256 prefixes of /64 each */
+	test_pool_size("2001:DB8::/56", 0, NULL, 0, 256);
+}
+
 int main(int argc, char **argv)
 {
-	osmo_init_logging(&log_info);
+	void *tall_ctx = talloc_named_const(NULL, 1, "Root context");
+	msgb_talloc_ctx_init(tall_ctx, 0);
+	osmo_init_logging2(tall_ctx, &log_info);
 	log_set_use_color(osmo_stderr_target, 0);
-	log_set_print_filename(osmo_stderr_target, 0);
+	log_set_print_filename2(osmo_stderr_target, LOG_FILENAME_NONE);
+	log_set_print_category(osmo_stderr_target, 0);
+	log_set_print_category_hex(osmo_stderr_target, 0);
 
 	srand(time(NULL));
 
-	test_pool_sizes();
+	if (argc < 2 || strcmp(argv[1], "-v6")) {
+		test_pool_sizes();
+	} else {
+		test_pool_sizes_v6();
+	}
+	return 0;
 }

tests/lib/ippool_test.err

@@ -8,5 +8,3 @@ No more IP addresses available
 No more IP addresses available
 No more IP addresses available
 No more IP addresses available
-No more IP addresses available
-No more IP addresses available

tests/lib/ippool_test.ok

@@ -66301,263 +66301,6 @@ allocated address 192.168.255.251
 allocated address 192.168.255.252
 allocated address 192.168.255.253
 allocated address 192.168.255.254
-testing pool for prefix 2001:DB8::/56, flags=0x0, blacklist_size=0, expected_size=256
-allocated address 2001:db8::
-allocated address 2001:db8:0:1::
-allocated address 2001:db8:0:2::
-allocated address 2001:db8:0:3::
-allocated address 2001:db8:0:4::
-allocated address 2001:db8:0:5::
-allocated address 2001:db8:0:6::
-allocated address 2001:db8:0:7::
-allocated address 2001:db8:0:8::
-allocated address 2001:db8:0:9::
-allocated address 2001:db8:0:a::
-allocated address 2001:db8:0:b::
-allocated address 2001:db8:0:c::
-allocated address 2001:db8:0:d::
-allocated address 2001:db8:0:e::
-allocated address 2001:db8:0:f::
-allocated address 2001:db8:0:10::
-allocated address 2001:db8:0:11::
-allocated address 2001:db8:0:12::
-allocated address 2001:db8:0:13::
-allocated address 2001:db8:0:14::
-allocated address 2001:db8:0:15::
-allocated address 2001:db8:0:16::
-allocated address 2001:db8:0:17::
-allocated address 2001:db8:0:18::
-allocated address 2001:db8:0:19::
-allocated address 2001:db8:0:1a::
-allocated address 2001:db8:0:1b::
-allocated address 2001:db8:0:1c::
-allocated address 2001:db8:0:1d::
-allocated address 2001:db8:0:1e::
-allocated address 2001:db8:0:1f::
-allocated address 2001:db8:0:20::
-allocated address 2001:db8:0:21::
-allocated address 2001:db8:0:22::
-allocated address 2001:db8:0:23::
-allocated address 2001:db8:0:24::
-allocated address 2001:db8:0:25::
-allocated address 2001:db8:0:26::
-allocated address 2001:db8:0:27::
-allocated address 2001:db8:0:28::
-allocated address 2001:db8:0:29::
-allocated address 2001:db8:0:2a::
-allocated address 2001:db8:0:2b::
-allocated address 2001:db8:0:2c::
-allocated address 2001:db8:0:2d::
-allocated address 2001:db8:0:2e::
-allocated address 2001:db8:0:2f::
-allocated address 2001:db8:0:30::
-allocated address 2001:db8:0:31::
-allocated address 2001:db8:0:32::
-allocated address 2001:db8:0:33::
-allocated address 2001:db8:0:34::
-allocated address 2001:db8:0:35::
-allocated address 2001:db8:0:36::
-allocated address 2001:db8:0:37::
-allocated address 2001:db8:0:38::
-allocated address 2001:db8:0:39::
-allocated address 2001:db8:0:3a::
-allocated address 2001:db8:0:3b::
-allocated address 2001:db8:0:3c::
-allocated address 2001:db8:0:3d::
-allocated address 2001:db8:0:3e::
-allocated address 2001:db8:0:3f::
-allocated address 2001:db8:0:40::
-allocated address 2001:db8:0:41::
-allocated address 2001:db8:0:42::
-allocated address 2001:db8:0:43::
-allocated address 2001:db8:0:44::
-allocated address 2001:db8:0:45::
-allocated address 2001:db8:0:46::
-allocated address 2001:db8:0:47::
-allocated address 2001:db8:0:48::
-allocated address 2001:db8:0:49::
-allocated address 2001:db8:0:4a::
-allocated address 2001:db8:0:4b::
-allocated address 2001:db8:0:4c::
-allocated address 2001:db8:0:4d::
-allocated address 2001:db8:0:4e::
-allocated address 2001:db8:0:4f::
-allocated address 2001:db8:0:50::
-allocated address 2001:db8:0:51::
-allocated address 2001:db8:0:52::
-allocated address 2001:db8:0:53::
-allocated address 2001:db8:0:54::
-allocated address 2001:db8:0:55::
-allocated address 2001:db8:0:56::
-allocated address 2001:db8:0:57::
-allocated address 2001:db8:0:58::
-allocated address 2001:db8:0:59::
-allocated address 2001:db8:0:5a::
-allocated address 2001:db8:0:5b::
-allocated address 2001:db8:0:5c::
-allocated address 2001:db8:0:5d::
-allocated address 2001:db8:0:5e::
-allocated address 2001:db8:0:5f::
-allocated address 2001:db8:0:60::
-allocated address 2001:db8:0:61::
-allocated address 2001:db8:0:62::
-allocated address 2001:db8:0:63::
-allocated address 2001:db8:0:64::
-allocated address 2001:db8:0:65::
-allocated address 2001:db8:0:66::
-allocated address 2001:db8:0:67::
-allocated address 2001:db8:0:68::
-allocated address 2001:db8:0:69::
-allocated address 2001:db8:0:6a::
-allocated address 2001:db8:0:6b::
-allocated address 2001:db8:0:6c::
-allocated address 2001:db8:0:6d::
-allocated address 2001:db8:0:6e::
-allocated address 2001:db8:0:6f::
-allocated address 2001:db8:0:70::
-allocated address 2001:db8:0:71::
-allocated address 2001:db8:0:72::
-allocated address 2001:db8:0:73::
-allocated address 2001:db8:0:74::
-allocated address 2001:db8:0:75::
-allocated address 2001:db8:0:76::
-allocated address 2001:db8:0:77::
-allocated address 2001:db8:0:78::
-allocated address 2001:db8:0:79::
-allocated address 2001:db8:0:7a::
-allocated address 2001:db8:0:7b::
-allocated address 2001:db8:0:7c::
-allocated address 2001:db8:0:7d::
-allocated address 2001:db8:0:7e::
-allocated address 2001:db8:0:7f::
-allocated address 2001:db8:0:80::
-allocated address 2001:db8:0:81::
-allocated address 2001:db8:0:82::
-allocated address 2001:db8:0:83::
-allocated address 2001:db8:0:84::
-allocated address 2001:db8:0:85::
-allocated address 2001:db8:0:86::
-allocated address 2001:db8:0:87::
-allocated address 2001:db8:0:88::
-allocated address 2001:db8:0:89::
-allocated address 2001:db8:0:8a::
-allocated address 2001:db8:0:8b::
-allocated address 2001:db8:0:8c::
-allocated address 2001:db8:0:8d::
-allocated address 2001:db8:0:8e::
-allocated address 2001:db8:0:8f::
-allocated address 2001:db8:0:90::
-allocated address 2001:db8:0:91::
-allocated address 2001:db8:0:92::
-allocated address 2001:db8:0:93::
-allocated address 2001:db8:0:94::
-allocated address 2001:db8:0:95::
-allocated address 2001:db8:0:96::
-allocated address 2001:db8:0:97::
-allocated address 2001:db8:0:98::
-allocated address 2001:db8:0:99::
-allocated address 2001:db8:0:9a::
-allocated address 2001:db8:0:9b::
-allocated address 2001:db8:0:9c::
-allocated address 2001:db8:0:9d::
-allocated address 2001:db8:0:9e::
-allocated address 2001:db8:0:9f::
-allocated address 2001:db8:0:a0::
-allocated address 2001:db8:0:a1::
-allocated address 2001:db8:0:a2::
-allocated address 2001:db8:0:a3::
-allocated address 2001:db8:0:a4::
-allocated address 2001:db8:0:a5::
-allocated address 2001:db8:0:a6::
-allocated address 2001:db8:0:a7::
-allocated address 2001:db8:0:a8::
-allocated address 2001:db8:0:a9::
-allocated address 2001:db8:0:aa::
-allocated address 2001:db8:0:ab::
-allocated address 2001:db8:0:ac::
-allocated address 2001:db8:0:ad::
-allocated address 2001:db8:0:ae::
-allocated address 2001:db8:0:af::
-allocated address 2001:db8:0:b0::
-allocated address 2001:db8:0:b1::
-allocated address 2001:db8:0:b2::
-allocated address 2001:db8:0:b3::
-allocated address 2001:db8:0:b4::
-allocated address 2001:db8:0:b5::
-allocated address 2001:db8:0:b6::
-allocated address 2001:db8:0:b7::
-allocated address 2001:db8:0:b8::
-allocated address 2001:db8:0:b9::
-allocated address 2001:db8:0:ba::
-allocated address 2001:db8:0:bb::
-allocated address 2001:db8:0:bc::
-allocated address 2001:db8:0:bd::
-allocated address 2001:db8:0:be::
-allocated address 2001:db8:0:bf::
-allocated address 2001:db8:0:c0::
-allocated address 2001:db8:0:c1::
-allocated address 2001:db8:0:c2::
-allocated address 2001:db8:0:c3::
-allocated address 2001:db8:0:c4::
-allocated address 2001:db8:0:c5::
-allocated address 2001:db8:0:c6::
-allocated address 2001:db8:0:c7::
-allocated address 2001:db8:0:c8::
-allocated address 2001:db8:0:c9::
-allocated address 2001:db8:0:ca::
-allocated address 2001:db8:0:cb::
-allocated address 2001:db8:0:cc::
-allocated address 2001:db8:0:cd::
-allocated address 2001:db8:0:ce::
-allocated address 2001:db8:0:cf::
-allocated address 2001:db8:0:d0::
-allocated address 2001:db8:0:d1::
-allocated address 2001:db8:0:d2::
-allocated address 2001:db8:0:d3::
-allocated address 2001:db8:0:d4::
-allocated address 2001:db8:0:d5::
-allocated address 2001:db8:0:d6::
-allocated address 2001:db8:0:d7::
-allocated address 2001:db8:0:d8::
-allocated address 2001:db8:0:d9::
-allocated address 2001:db8:0:da::
-allocated address 2001:db8:0:db::
-allocated address 2001:db8:0:dc::
-allocated address 2001:db8:0:dd::
-allocated address 2001:db8:0:de::
-allocated address 2001:db8:0:df::
-allocated address 2001:db8:0:e0::
-allocated address 2001:db8:0:e1::
-allocated address 2001:db8:0:e2::
-allocated address 2001:db8:0:e3::
-allocated address 2001:db8:0:e4::
-allocated address 2001:db8:0:e5::
-allocated address 2001:db8:0:e6::
-allocated address 2001:db8:0:e7::
-allocated address 2001:db8:0:e8::
-allocated address 2001:db8:0:e9::
-allocated address 2001:db8:0:ea::
-allocated address 2001:db8:0:eb::
-allocated address 2001:db8:0:ec::
-allocated address 2001:db8:0:ed::
-allocated address 2001:db8:0:ee::
-allocated address 2001:db8:0:ef::
-allocated address 2001:db8:0:f0::
-allocated address 2001:db8:0:f1::
-allocated address 2001:db8:0:f2::
-allocated address 2001:db8:0:f3::
-allocated address 2001:db8:0:f4::
-allocated address 2001:db8:0:f5::
-allocated address 2001:db8:0:f6::
-allocated address 2001:db8:0:f7::
-allocated address 2001:db8:0:f8::
-allocated address 2001:db8:0:f9::
-allocated address 2001:db8:0:fa::
-allocated address 2001:db8:0:fb::
-allocated address 2001:db8:0:fc::
-allocated address 2001:db8:0:fd::
-allocated address 2001:db8:0:fe::
-allocated address 2001:db8:0:ff::
 testing pool for prefix 192.168.23.0/24, flags=0x3, blacklist_size=3, expected_size=253
 allocated address 192.168.23.2
 allocated address 192.168.23.3

tests/lib/ippool_v6_test.err

@@ -0,0 +1,2 @@
+No more IP addresses available
+No more IP addresses available

tests/lib/ippool_v6_test.ok

@@ -0,0 +1,257 @@
+testing pool for prefix 2001:DB8::/56, flags=0x0, blacklist_size=0, expected_size=256
+allocated address 2001:db8::
+allocated address 2001:db8:0:1::
+allocated address 2001:db8:0:2::
+allocated address 2001:db8:0:3::
+allocated address 2001:db8:0:4::
+allocated address 2001:db8:0:5::
+allocated address 2001:db8:0:6::
+allocated address 2001:db8:0:7::
+allocated address 2001:db8:0:8::
+allocated address 2001:db8:0:9::
+allocated address 2001:db8:0:a::
+allocated address 2001:db8:0:b::
+allocated address 2001:db8:0:c::
+allocated address 2001:db8:0:d::
+allocated address 2001:db8:0:e::
+allocated address 2001:db8:0:f::
+allocated address 2001:db8:0:10::
+allocated address 2001:db8:0:11::
+allocated address 2001:db8:0:12::
+allocated address 2001:db8:0:13::
+allocated address 2001:db8:0:14::
+allocated address 2001:db8:0:15::
+allocated address 2001:db8:0:16::
+allocated address 2001:db8:0:17::
+allocated address 2001:db8:0:18::
+allocated address 2001:db8:0:19::
+allocated address 2001:db8:0:1a::
+allocated address 2001:db8:0:1b::
+allocated address 2001:db8:0:1c::
+allocated address 2001:db8:0:1d::
+allocated address 2001:db8:0:1e::
+allocated address 2001:db8:0:1f::
+allocated address 2001:db8:0:20::
+allocated address 2001:db8:0:21::
+allocated address 2001:db8:0:22::
+allocated address 2001:db8:0:23::
+allocated address 2001:db8:0:24::
+allocated address 2001:db8:0:25::
+allocated address 2001:db8:0:26::
+allocated address 2001:db8:0:27::
+allocated address 2001:db8:0:28::
+allocated address 2001:db8:0:29::
+allocated address 2001:db8:0:2a::
+allocated address 2001:db8:0:2b::
+allocated address 2001:db8:0:2c::
+allocated address 2001:db8:0:2d::
+allocated address 2001:db8:0:2e::
+allocated address 2001:db8:0:2f::
+allocated address 2001:db8:0:30::
+allocated address 2001:db8:0:31::
+allocated address 2001:db8:0:32::
+allocated address 2001:db8:0:33::
+allocated address 2001:db8:0:34::
+allocated address 2001:db8:0:35::
+allocated address 2001:db8:0:36::
+allocated address 2001:db8:0:37::
+allocated address 2001:db8:0:38::
+allocated address 2001:db8:0:39::
+allocated address 2001:db8:0:3a::
+allocated address 2001:db8:0:3b::
+allocated address 2001:db8:0:3c::
+allocated address 2001:db8:0:3d::
+allocated address 2001:db8:0:3e::
+allocated address 2001:db8:0:3f::
+allocated address 2001:db8:0:40::
+allocated address 2001:db8:0:41::
+allocated address 2001:db8:0:42::
+allocated address 2001:db8:0:43::
+allocated address 2001:db8:0:44::
+allocated address 2001:db8:0:45::
+allocated address 2001:db8:0:46::
+allocated address 2001:db8:0:47::
+allocated address 2001:db8:0:48::
+allocated address 2001:db8:0:49::
+allocated address 2001:db8:0:4a::
+allocated address 2001:db8:0:4b::
+allocated address 2001:db8:0:4c::
+allocated address 2001:db8:0:4d::
+allocated address 2001:db8:0:4e::
+allocated address 2001:db8:0:4f::
+allocated address 2001:db8:0:50::
+allocated address 2001:db8:0:51::
+allocated address 2001:db8:0:52::
+allocated address 2001:db8:0:53::
+allocated address 2001:db8:0:54::
+allocated address 2001:db8:0:55::
+allocated address 2001:db8:0:56::
+allocated address 2001:db8:0:57::
+allocated address 2001:db8:0:58::
+allocated address 2001:db8:0:59::
+allocated address 2001:db8:0:5a::
+allocated address 2001:db8:0:5b::
+allocated address 2001:db8:0:5c::
+allocated address 2001:db8:0:5d::
+allocated address 2001:db8:0:5e::
+allocated address 2001:db8:0:5f::
+allocated address 2001:db8:0:60::
+allocated address 2001:db8:0:61::
+allocated address 2001:db8:0:62::
+allocated address 2001:db8:0:63::
+allocated address 2001:db8:0:64::
+allocated address 2001:db8:0:65::
+allocated address 2001:db8:0:66::
+allocated address 2001:db8:0:67::
+allocated address 2001:db8:0:68::
+allocated address 2001:db8:0:69::
+allocated address 2001:db8:0:6a::
+allocated address 2001:db8:0:6b::
+allocated address 2001:db8:0:6c::
+allocated address 2001:db8:0:6d::
+allocated address 2001:db8:0:6e::
+allocated address 2001:db8:0:6f::
+allocated address 2001:db8:0:70::
+allocated address 2001:db8:0:71::
+allocated address 2001:db8:0:72::
+allocated address 2001:db8:0:73::
+allocated address 2001:db8:0:74::
+allocated address 2001:db8:0:75::
+allocated address 2001:db8:0:76::
+allocated address 2001:db8:0:77::
+allocated address 2001:db8:0:78::
+allocated address 2001:db8:0:79::
+allocated address 2001:db8:0:7a::
+allocated address 2001:db8:0:7b::
+allocated address 2001:db8:0:7c::
+allocated address 2001:db8:0:7d::
+allocated address 2001:db8:0:7e::
+allocated address 2001:db8:0:7f::
+allocated address 2001:db8:0:80::
+allocated address 2001:db8:0:81::
+allocated address 2001:db8:0:82::
+allocated address 2001:db8:0:83::
+allocated address 2001:db8:0:84::
+allocated address 2001:db8:0:85::
+allocated address 2001:db8:0:86::
+allocated address 2001:db8:0:87::
+allocated address 2001:db8:0:88::
+allocated address 2001:db8:0:89::
+allocated address 2001:db8:0:8a::
+allocated address 2001:db8:0:8b::
+allocated address 2001:db8:0:8c::
+allocated address 2001:db8:0:8d::
+allocated address 2001:db8:0:8e::
+allocated address 2001:db8:0:8f::
+allocated address 2001:db8:0:90::
+allocated address 2001:db8:0:91::
+allocated address 2001:db8:0:92::
+allocated address 2001:db8:0:93::
+allocated address 2001:db8:0:94::
+allocated address 2001:db8:0:95::
+allocated address 2001:db8:0:96::
+allocated address 2001:db8:0:97::
+allocated address 2001:db8:0:98::
+allocated address 2001:db8:0:99::
+allocated address 2001:db8:0:9a::
+allocated address 2001:db8:0:9b::
+allocated address 2001:db8:0:9c::
+allocated address 2001:db8:0:9d::
+allocated address 2001:db8:0:9e::
+allocated address 2001:db8:0:9f::
+allocated address 2001:db8:0:a0::
+allocated address 2001:db8:0:a1::
+allocated address 2001:db8:0:a2::
+allocated address 2001:db8:0:a3::
+allocated address 2001:db8:0:a4::
+allocated address 2001:db8:0:a5::
+allocated address 2001:db8:0:a6::
+allocated address 2001:db8:0:a7::
+allocated address 2001:db8:0:a8::
+allocated address 2001:db8:0:a9::
+allocated address 2001:db8:0:aa::
+allocated address 2001:db8:0:ab::
+allocated address 2001:db8:0:ac::
+allocated address 2001:db8:0:ad::
+allocated address 2001:db8:0:ae::
+allocated address 2001:db8:0:af::
+allocated address 2001:db8:0:b0::
+allocated address 2001:db8:0:b1::
+allocated address 2001:db8:0:b2::
+allocated address 2001:db8:0:b3::
+allocated address 2001:db8:0:b4::
+allocated address 2001:db8:0:b5::
+allocated address 2001:db8:0:b6::
+allocated address 2001:db8:0:b7::
+allocated address 2001:db8:0:b8::
+allocated address 2001:db8:0:b9::
+allocated address 2001:db8:0:ba::
+allocated address 2001:db8:0:bb::
+allocated address 2001:db8:0:bc::
+allocated address 2001:db8:0:bd::
+allocated address 2001:db8:0:be::
+allocated address 2001:db8:0:bf::
+allocated address 2001:db8:0:c0::
+allocated address 2001:db8:0:c1::
+allocated address 2001:db8:0:c2::
+allocated address 2001:db8:0:c3::
+allocated address 2001:db8:0:c4::
+allocated address 2001:db8:0:c5::
+allocated address 2001:db8:0:c6::
+allocated address 2001:db8:0:c7::
+allocated address 2001:db8:0:c8::
+allocated address 2001:db8:0:c9::
+allocated address 2001:db8:0:ca::
+allocated address 2001:db8:0:cb::
+allocated address 2001:db8:0:cc::
+allocated address 2001:db8:0:cd::
+allocated address 2001:db8:0:ce::
+allocated address 2001:db8:0:cf::
+allocated address 2001:db8:0:d0::
+allocated address 2001:db8:0:d1::
+allocated address 2001:db8:0:d2::
+allocated address 2001:db8:0:d3::
+allocated address 2001:db8:0:d4::
+allocated address 2001:db8:0:d5::
+allocated address 2001:db8:0:d6::
+allocated address 2001:db8:0:d7::
+allocated address 2001:db8:0:d8::
+allocated address 2001:db8:0:d9::
+allocated address 2001:db8:0:da::
+allocated address 2001:db8:0:db::
+allocated address 2001:db8:0:dc::
+allocated address 2001:db8:0:dd::
+allocated address 2001:db8:0:de::
+allocated address 2001:db8:0:df::
+allocated address 2001:db8:0:e0::
+allocated address 2001:db8:0:e1::
+allocated address 2001:db8:0:e2::
+allocated address 2001:db8:0:e3::
+allocated address 2001:db8:0:e4::
+allocated address 2001:db8:0:e5::
+allocated address 2001:db8:0:e6::
+allocated address 2001:db8:0:e7::
+allocated address 2001:db8:0:e8::
+allocated address 2001:db8:0:e9::
+allocated address 2001:db8:0:ea::
+allocated address 2001:db8:0:eb::
+allocated address 2001:db8:0:ec::
+allocated address 2001:db8:0:ed::
+allocated address 2001:db8:0:ee::
+allocated address 2001:db8:0:ef::
+allocated address 2001:db8:0:f0::
+allocated address 2001:db8:0:f1::
+allocated address 2001:db8:0:f2::
+allocated address 2001:db8:0:f3::
+allocated address 2001:db8:0:f4::
+allocated address 2001:db8:0:f5::
+allocated address 2001:db8:0:f6::
+allocated address 2001:db8:0:f7::
+allocated address 2001:db8:0:f8::
+allocated address 2001:db8:0:f9::
+allocated address 2001:db8:0:fa::
+allocated address 2001:db8:0:fb::
+allocated address 2001:db8:0:fc::
+allocated address 2001:db8:0:fd::
+allocated address 2001:db8:0:fe::
+allocated address 2001:db8:0:ff::

tests/testsuite.at

@@ -8,14 +8,33 @@ cat $abs_srcdir/lib/ippool_test.err > experr
 AT_CHECK([$abs_top_builddir/tests/lib/ippool_test], [], [expout], [experr])
 AT_CLEANUP
 
+AT_SETUP([ippool_v6])
+AT_KEYWORDS([ippool_v6])
+cat $abs_srcdir/lib/ippool_v6_test.ok > expout
+cat $abs_srcdir/lib/ippool_v6_test.err > experr
+AT_CHECK([$abs_top_builddir/tests/lib/ippool_test -v6], [], [expout], [experr])
+AT_CLEANUP
+
 AT_SETUP([in46a])
 AT_KEYWORDS([in46a])
 cat $abs_srcdir/lib/in46a_test.ok > expout
 AT_CHECK([$abs_top_builddir/tests/lib/in46a_test], [], [expout], [])
 AT_CLEANUP
 
+AT_SETUP([in46a_v6])
+AT_KEYWORDS([in46a_v6])
+cat $abs_srcdir/lib/in46a_v6_test.ok > expout
+AT_CHECK([$abs_top_builddir/tests/lib/in46a_test -v6], [], [expout], [])
+AT_CLEANUP
+
 AT_SETUP([gtpie])
 AT_KEYWORDS([gtpie])
 cat $abs_srcdir/gtp/gtpie_test.ok > expout
 AT_CHECK([$abs_top_builddir/tests/gtp/gtpie_test], [], [expout], [])
 AT_CLEANUP
+
+AT_SETUP([queue])
+AT_KEYWORDS([queue])
+cat $abs_srcdir/gtp/queue_test.ok > expout
+AT_CHECK([$abs_top_builddir/tests/gtp/queue_test], [], [expout], [])
+AT_CLEANUP