path: root/src/openvpn/mss.c

/*
 *  OpenVPN -- An application to securely tunnel IP networks
 *             over a single TCP/UDP port, with support for SSL/TLS-based
 *             session authentication and key exchange,
 *             packet encryption, packet authentication, and
 *             packet compression.
 *
 *  Copyright (C) 2002-2024 OpenVPN Inc <sales@openvpn.net>
 *
 *  This program is free software; you can redistribute it and/or modify
 *  it under the terms of the GNU General Public License version 2
 *  as published by the Free Software Foundation.
 *
 *  This program is distributed in the hope that it will be useful,
 *  but WITHOUT ANY WARRANTY; without even the implied warranty of
 *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 *  GNU General Public License for more details.
 *
 *  You should have received a copy of the GNU General Public License along
 *  with this program; if not, write to the Free Software Foundation, Inc.,
 *  51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
 */

#ifdef HAVE_CONFIG_H
#include "config.h"
#endif

#include "syshead.h"
#include "error.h"
#include "mss.h"
#include "crypto.h"
#include "ssl_common.h"
#include "memdbg.h"
#include "forward.h"

/*
 * Lower MSS on TCP SYN packets to fix MTU
 * problems which arise from protocol
 * encapsulation.
 */

/*
 * IPv4 packet: find TCP header, check flags for "SYN"
 *              if yes, hand to mss_fixup_dowork()
 */
void
mss_fixup_ipv4(struct buffer *buf, uint16_t maxmss)
{
    const struct openvpn_iphdr *pip;
    int hlen;

    if (BLEN(buf) < (int) sizeof(struct openvpn_iphdr))
    {
        return;
    }

    verify_align_4(buf);
    pip = (struct openvpn_iphdr *) BPTR(buf);

    hlen = OPENVPN_IPH_GET_LEN(pip->version_len);

    if (pip->protocol == OPENVPN_IPPROTO_TCP
        && ntohs(pip->tot_len) == BLEN(buf)
        && (ntohs(pip->frag_off) & OPENVPN_IP_OFFMASK) == 0
        && hlen <= BLEN(buf)
        && BLEN(buf) - hlen
        >= (int) sizeof(struct openvpn_tcphdr))
    {
        struct buffer newbuf = *buf;
        if (buf_advance(&newbuf, hlen))
        {
            struct openvpn_tcphdr *tc = (struct openvpn_tcphdr *) BPTR(&newbuf);
            if (tc->flags & OPENVPN_TCPH_SYN_MASK)
            {
                mss_fixup_dowork(&newbuf, maxmss);
            }
        }
    }
}

/*
 * IPv6 packet: find TCP header, check flags for "SYN"
 *              if yes, hand to mss_fixup_dowork()
 *              (IPv6 header structure is sufficiently different from IPv4...)
 */
void
mss_fixup_ipv6(struct buffer *buf, uint16_t maxmss)
{
    const struct openvpn_ipv6hdr *pip6;
    struct buffer newbuf;

    if (BLEN(buf) < (int) sizeof(struct openvpn_ipv6hdr))
    {
        return;
    }

    verify_align_4(buf);
    pip6 = (struct openvpn_ipv6hdr *) BPTR(buf);

    /* do we have the full IPv6 packet?
     * "payload_len" does not include IPv6 header (+40 bytes)
     */
    if (BLEN(buf) != (int) ntohs(pip6->payload_len)+40)
    {
        return;
    }

    /* follow header chain until we reach final header, then check for TCP
     *
     * An IPv6 packet could, theoretically, have a chain of multiple headers
     * before the final header (TCP, UDP, ...), so we'd need to walk that
     * chain (see RFC 2460 and RFC 6564 for details).
     *
     * In practice, "most typically used" extension headers (AH, routing,
     * fragment, mobility) are very unlikely to be seen inside an OpenVPN
     * tun, so for now, we only handle the case of "single next header = TCP"
     */
    if (pip6->nexthdr != OPENVPN_IPPROTO_TCP)
    {
        return;
    }

    /* skip IPv6 header (40 bytes),
     * verify remainder is large enough to contain a full TCP header
     */
    newbuf = *buf;
    if (buf_advance( &newbuf, 40 )
        && BLEN(&newbuf) >= (int) sizeof(struct openvpn_tcphdr))
    {
        struct openvpn_tcphdr *tc = (struct openvpn_tcphdr *) BPTR(&newbuf);
        if (tc->flags & OPENVPN_TCPH_SYN_MASK)
        {
            mss_fixup_dowork(&newbuf, maxmss-20);
        }
    }
}

/*
 * change TCP MSS option in SYN/SYN-ACK packets, if present
 * this is generic for IPv4 and IPv6, as the TCP header is the same
 */

void
mss_fixup_dowork(struct buffer *buf, uint16_t maxmss)
{
    int hlen, olen, optlen;
    uint8_t *opt;
    uint16_t mssval;
    int accumulate;
    struct openvpn_tcphdr *tc;

    if (BLEN(buf) < (int) sizeof(struct openvpn_tcphdr))
    {
        return;
    }

    verify_align_4(buf);
    tc = (struct openvpn_tcphdr *) BPTR(buf);
    hlen = OPENVPN_TCPH_GET_DOFF(tc->doff_res);

    /* Invalid header length or header without options. */
    if (hlen <= (int) sizeof(struct openvpn_tcphdr)
        || hlen > BLEN(buf))
    {
        return;
    }

    for (olen = hlen - sizeof(struct openvpn_tcphdr),
         opt = (uint8_t *)(tc + 1);
         olen > 1;
         olen -= optlen, opt += optlen)
    {
        if (*opt == OPENVPN_TCPOPT_EOL)
        {
            break;
        }
        else if (*opt == OPENVPN_TCPOPT_NOP)
        {
            optlen = 1;
        }
        else
        {
            optlen = *(opt + 1);
            if (optlen <= 0 || optlen > olen)
            {
                break;
            }
            if (*opt == OPENVPN_TCPOPT_MAXSEG)
            {
                if (optlen != OPENVPN_TCPOLEN_MAXSEG)
                {
                    continue;
                }
                mssval = opt[2] << 8;
                mssval += opt[3];
                if (mssval > maxmss)
                {
                    dmsg(D_MSS, "MSS: %" PRIu16 " -> %" PRIu16, mssval, maxmss);
                    accumulate = htons(mssval);
                    opt[2] = (uint8_t)((maxmss>>8)&0xff);
                    opt[3] = (uint8_t)(maxmss&0xff);
                    accumulate -= htons(maxmss);
                    ADJUST_CHECKSUM(accumulate, tc->check);
                }
            }
        }
    }
}

static inline size_t
adjust_payload_max_cbc(const struct key_type *kt, size_t target)
{
    if (!cipher_kt_mode_cbc(kt->cipher))
    {
        /* With stream ciphers (or block cipher in stream modes like CFB, AEAD)
         * we can just use the target as is */
        return target;
    }
    else
    {
        /* With CBC we need at least one extra byte for padding and then need
         * to ensure that the resulting CBC ciphertext length, which is always
         * a multiple of the block size, is not larger than the target value */
        size_t block_size = cipher_kt_block_size(kt->cipher);
        target = round_down_size(target, block_size);
        return target - 1;
    }
}

static size_t
get_ip_encap_overhead(const struct options *options,
                      const struct link_socket_info *lsi)
{
    /* Add the overhead of the encapsulating IP packets */
    sa_family_t af;
    int proto;

    if (lsi && lsi->lsa)
    {
        af = lsi->lsa->actual.dest.addr.sa.sa_family;
        proto = lsi->proto;
    }
    else
    {
        /* In the early init before the connection is established or we
         * are in listen mode we can only make an educated guess
         * from the af of the connection entry, in p2mp this will be
         * later updated */
        af = options->ce.af;
        proto = options->ce.proto;
    }
    return datagram_overhead(af, proto);
}

static void
frame_calculate_fragment(struct frame *frame, struct key_type *kt,
                         const struct options *options,
                         struct link_socket_info *lsi)
{
#if defined(ENABLE_FRAGMENT)
    size_t overhead;

    overhead = frame_calculate_protocol_header_size(kt, options, false);

    if (options->ce.fragment_encap)
    {
        overhead += get_ip_encap_overhead(options, lsi);
    }

    size_t target = options->ce.fragment - overhead;
    /* The 4 bytes of header that fragment adds itself. The other extra payload
     * bytes (Ethernet header/compression) are handled by the fragment code
     * just as part of the payload and therefore automatically taken into
     * account if the packet needs to fragmented */
    frame->max_fragment_size = clamp_size_to_int(adjust_payload_max_cbc(kt, target)) - 4;

    if (cipher_kt_mode_cbc(kt->cipher))
    {
        /* The packet id gets added to *each* fragment in CBC mode, so we need
         * to account for it */
        frame->max_fragment_size -= calc_packet_id_size_dc(options, kt);
    }
#endif
}

static void
frame_calculate_mssfix(struct frame *frame, struct key_type *kt,
                       const struct options *options,
                       struct link_socket_info *lsi)
{
    if (options->ce.mssfix_fixed)
    {
        /* we subtract IPv4 and TCP overhead here, mssfix method will add the
         * extra 20 for IPv6 */
        frame->mss_fix = (uint16_t)(options->ce.mssfix - (20 + 20));
        return;
    }

    size_t overhead, payload_overhead;

    overhead = frame_calculate_protocol_header_size(kt, options, false);

    /* Calculate the number of bytes that the payload differs from the payload
     * MTU. This are fragment/compression/ethernet headers */
    payload_overhead = frame_calculate_payload_overhead(frame->extra_tun, options, kt);

    /* We are in a "liberal" position with respect to MSS,
     * i.e. we assume that MSS can be calculated from MTU
     * by subtracting out only the IP and TCP header sizes
     * without options.
     *
     * (RFC 879, section 7). */

    if (options->ce.mssfix_encap)
    {
        /* Add the overhead of the encapsulating IP packets */
        overhead += get_ip_encap_overhead(options, lsi);
    }

    /* Add 20 bytes for the IPv4 header and 20 byte for the TCP header of the
     * payload, the mssfix method will add 20 extra if payload is IPv6 */
    payload_overhead += 20 + 20;

    /* Calculate the maximum MSS value from the max link layer size specified
     * by ce.mssfix */

    /* This is the target value our payload needs to be smaller */
    size_t target = options->ce.mssfix - overhead;
    frame->mss_fix = (uint16_t)(adjust_payload_max_cbc(kt, target) - payload_overhead);


}

void
frame_calculate_dynamic(struct frame *frame, struct key_type *kt,
                        const struct options *options,
                        struct link_socket_info *lsi)
{
    if (options->ce.fragment > 0)
    {
        frame_calculate_fragment(frame, kt, options, lsi);
    }

    if (options->ce.mssfix > 0)
    {
        frame_calculate_mssfix(frame, kt, options, lsi);
    }
}

/*
 * Adjust frame structure based on a Path MTU value given
 * to us by the OS.
 */
void
frame_adjust_path_mtu(struct context *c)
{
    struct link_socket_info *lsi = get_link_socket_info(c);
    struct options *o = &c->options;

    int pmtu = c->c2.link_socket->mtu;
    sa_family_t af = lsi->lsa->actual.dest.addr.sa.sa_family;
    int proto = lsi->proto;

    int encap_overhead = datagram_overhead(af, proto);

    /* check if mssfix and fragment need to be adjusted */
    if (pmtu < o->ce.mssfix
        || (o->ce.mssfix_encap && pmtu < o->ce.mssfix + encap_overhead))
    {
        const char *mtustr = o->ce.mssfix_encap ? " mtu" : "";
        msg(D_MTU_INFO, "Note adjusting 'mssfix %d%s' to 'mssfix %d mtu' "
            "according to path MTU discovery", o->ce.mssfix,
            mtustr, pmtu);
        o->ce.mssfix = pmtu;
        o->ce.mssfix_encap = true;
        frame_calculate_dynamic(&c->c2.frame, &c->c1.ks.key_type, o, lsi);
    }

#if defined(ENABLE_FRAGMENT)
    if (pmtu < o->ce.fragment
        || (o->ce.fragment_encap && pmtu < o->ce.fragment + encap_overhead))
    {
        const char *mtustr = o->ce.fragment_encap ? " mtu" : "";
        msg(D_MTU_INFO, "Note adjusting 'fragment %d%s' to 'fragment %d mtu' "
            "according to path MTU discovery", o->ce.fragment,
            mtustr, pmtu);
        o->ce.fragment = pmtu;
        o->ce.fragment_encap = true;
        frame_calculate_dynamic(&c->c2.frame_fragment, &c->c1.ks.key_type,
                                o, lsi);
    }
#endif
}