/*      $NetBSD: in_offload.c,v 1.15 2024/07/05 04:31:54 rin Exp $      */

/*
* Copyright (c)2005, 2006 YAMAMOTO Takashi,
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
*    notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
*    notice, this list of conditions and the following disclaimer in the
*    documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*/

#include <sys/cdefs.h>
__KERNEL_RCSID(0, "$NetBSD: in_offload.c,v 1.15 2024/07/05 04:31:54 rin Exp $");

#include <sys/param.h>
#include <sys/mbuf.h>

#include <net/if.h>

#include <netinet/in.h>
#include <netinet/in_systm.h>
#include <netinet/ip.h>
#include <netinet/ip_var.h>
#include <netinet/tcp.h>
#include <netinet/in_offload.h>

/*
* Handle M_CSUM_TSOv4 in software. Split the TCP payload in chunks of
* size MSS, and return mbuf chain consists of them.
*/
struct mbuf *
tcp4_segment(struct mbuf *m, int off)
{
       int mss;
       int iphlen, thlen;
       int hlen, len;
       struct ip *ip;
       struct tcphdr *th;
       uint16_t ipid, phsum;
       uint32_t tcpseq;
       struct mbuf *hdr = NULL;
       struct mbuf *m0 = NULL;
       struct mbuf *prev = NULL;
       struct mbuf *n, *t;
       int nsegs;

       KASSERT((m->m_flags & M_PKTHDR) != 0);
       KASSERT((m->m_pkthdr.csum_flags & M_CSUM_TSOv4) != 0);

       m->m_pkthdr.csum_flags = 0;

       len = m->m_pkthdr.len;
       KASSERT(len >= off + sizeof(*ip) + sizeof(*th));

       hlen = off + sizeof(*ip);
       if (m->m_len < hlen) {
               m = m_pullup(m, hlen);
               if (m == NULL)
                       goto quit;
       }
       ip = (void *)(mtod(m, char *) + off);
       iphlen = ip->ip_hl * 4;
       KASSERT(ip->ip_v == IPVERSION);
       KASSERT(iphlen >= sizeof(*ip));
       KASSERT(ip->ip_p == IPPROTO_TCP);
       ipid = ntohs(ip->ip_id);

       hlen = off + iphlen + sizeof(*th);
       if (m->m_len < hlen) {
               m = m_pullup(m, hlen);
               if (m == NULL)
                       goto quit;
       }
       th = (void *)(mtod(m, char *) + off + iphlen);
       tcpseq = ntohl(th->th_seq);
       thlen = th->th_off * 4;
       hlen = off + iphlen + thlen;

       mss = m->m_pkthdr.segsz;
       KASSERT(mss != 0);
       KASSERT(len > hlen);

       t = m_split(m, hlen, M_NOWAIT);
       if (t == NULL)
               goto quit;
       hdr = m;
       m = t;

       len -= hlen;
       KASSERT(len % mss == 0);

       ip = (void *)(mtod(hdr, char *) + off);
       ip->ip_len = htons(iphlen + thlen + mss);
       phsum = in_cksum_phdr(ip->ip_src.s_addr, ip->ip_dst.s_addr,
           htons((uint16_t)(thlen + mss) + IPPROTO_TCP));

       for (nsegs = len / mss; nsegs > 0; nsegs--) {
               if (nsegs > 1) {
                       n = m_dup(hdr, 0, hlen, M_NOWAIT);
                       if (n == NULL)
                               goto quit;
               } else
                       n = hdr;
               KASSERT(n->m_len == hlen); /* XXX */

               if (nsegs > 1) {
                       t = m_split(m, mss, M_NOWAIT);
                       if (t == NULL) {
                               m_freem(n);
                               goto quit;
                       }
               } else
                       t = m;
               m_cat(n, m);
               m = t;

               KASSERT(n->m_len >= hlen); /* XXX */

               if (m0 == NULL)
                       m0 = n;

               if (prev != NULL)
                       prev->m_nextpkt = n;

               n->m_pkthdr.len = hlen + mss;
               n->m_nextpkt = NULL;    /* XXX */

               ip = (void *)(mtod(n, char *) + off);
               ip->ip_id = htons(ipid);
               ip->ip_sum = 0;
               ip->ip_sum = in4_cksum(n, 0, off, iphlen);

               th = (void *)(mtod(n, char *) + off + iphlen);
               th->th_seq = htonl(tcpseq);
               th->th_sum = phsum;
               th->th_sum = in4_cksum(n, 0, off + iphlen, thlen + mss);

               tcpseq += mss;
               ipid++;
               prev = n;
       }
       return m0;

quit:
       m_freem(hdr);
       m_freem(m);
       for (m = m0; m != NULL; m = n) {
               n = m->m_nextpkt;
               m_freem(m);
       }

       return NULL;
}

int
ip_tso_output(struct ifnet *ifp, struct mbuf *m, const struct sockaddr *sa,
   struct rtentry *rt)
{
       struct mbuf *n;
       int error = 0;

       m = tcp4_segment(m, 0);
       if (m == NULL)
               return ENOMEM;
       do {
               n = m->m_nextpkt;
               if (error == 0)
                       error = ip_if_output(ifp, m, sa, rt);
               else
                       m_freem(m);
               m = n;
       } while (m != NULL);
       return error;
}

/*
* Compute now in software the IP and TCP/UDP checksums. Cancel the
* hardware offloading.
*/
void
in_undefer_cksum(struct mbuf *mh, size_t hdrlen, int csum_flags)
{
       const size_t iphdrlen = M_CSUM_DATA_IPv4_IPHL(mh->m_pkthdr.csum_data);
       uint16_t csum;
       uint16_t ip_len;
       uint16_t *csump;
       struct mbuf *m = mh;

       KASSERT(mh->m_flags & M_PKTHDR);
       KASSERT(mh->m_pkthdr.len > hdrlen);
       KASSERT((mh->m_pkthdr.csum_flags & csum_flags) == csum_flags);

       /*
        * Deal with prepended frame header as done by e.g. ether_output().
        * If first mbuf in chain has just the header, use second mbuf
        * for the actual checksum. in4_csum() expects the passed mbuf
        * to have the whole (struct ip) area contiguous.
        */
       if (m->m_len <= hdrlen) {
               hdrlen -= m->m_len;
               m = m->m_next;
               KASSERT(m != NULL);
       }

       if (__predict_true(hdrlen + sizeof(struct ip) <= m->m_len)) {
               struct ip *ip = (struct ip *)(mtod(m, uint8_t *) + hdrlen);

               ip_len = ip->ip_len;
               csump = &ip->ip_sum;
       } else {
               const size_t ip_len_offset =
                   hdrlen + offsetof(struct ip, ip_len);

               m_copydata(m, ip_len_offset, sizeof(ip_len), &ip_len);
               csump = NULL;
       }
       ip_len = ntohs(ip_len);

       if (csum_flags & M_CSUM_IPv4) {
               csum = in4_cksum(m, 0, hdrlen, iphdrlen);
               if (csump != NULL) {
                       *csump = csum;
               } else {
                       const size_t offset = hdrlen +
                           offsetof(struct ip, ip_sum);

                       m_copyback(m, offset, sizeof(uint16_t), &csum);
               }
       }

       if (csum_flags & (M_CSUM_UDPv4|M_CSUM_TCPv4)) {
               size_t l4offset = hdrlen + iphdrlen;

               csum = in4_cksum(m, 0, l4offset, ip_len - iphdrlen);
               if (csum == 0 && (csum_flags & M_CSUM_UDPv4) != 0)
                       csum = 0xffff;

               l4offset += M_CSUM_DATA_IPv4_OFFSET(m->m_pkthdr.csum_data);

               if (__predict_true(l4offset + sizeof(uint16_t) <= m->m_len)) {
                       *(uint16_t *)(mtod(m, char *) + l4offset) = csum;
               } else {
                       m_copyback(m, l4offset, sizeof(csum), (void *)&csum);
               }
       }

       mh->m_pkthdr.csum_flags ^= csum_flags;
}

/*
* Compute now in software the TCP/UDP checksum. Cancel the hardware
* offloading.
*/
void
in_undefer_cksum_tcpudp(struct mbuf *m)
{
       struct ip *ip;
       uint16_t csum, offset;

       KASSERT((m->m_flags & M_PKTHDR) != 0);
       KASSERT((m->m_pkthdr.csum_flags & (M_CSUM_TCPv4|M_CSUM_UDPv4)) != 0);
       KASSERT((m->m_pkthdr.csum_flags & (M_CSUM_TCPv6|M_CSUM_UDPv6)) == 0);

       ip = mtod(m, struct ip *);
       offset = ip->ip_hl << 2;

       csum = in4_cksum(m, 0, offset, ntohs(ip->ip_len) - offset);
       if (csum == 0 && (m->m_pkthdr.csum_flags & M_CSUM_UDPv4) != 0)
               csum = 0xffff;

       offset += M_CSUM_DATA_IPv4_OFFSET(m->m_pkthdr.csum_data);

       if ((offset + sizeof(uint16_t)) <= m->m_len) {
               *(uint16_t *)(mtod(m, char *) + offset) = csum;
       } else {
               m_copyback(m, offset, sizeof(csum), (void *)&csum);
       }
}