net_pkt: Extend packet abstraction as required by e1000e functionality

This patch extends the TX/RX packet abstractions with features that will
be used by the e1000e device implementation.

Changes are:

  1. Support iovec lists for RX buffers
  2. Deeper RX packets parsing
  3. Loopback option for TX packets
  4. Extended VLAN headers handling
  5. RSS processing for RX packets

Signed-off-by: Dmitry Fleytman <dmitry.fleytman@ravellosystems.com>
Signed-off-by: Leonid Bloch <leonid.bloch@ravellosystems.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>

2 files changed, 373 insertions, 44 deletions

diff --git a/net/checksum.c b/net/checksum.c
index d0fa424cc1..196aaa3459 100644
--- a/net/checksum.c
+++ b/net/checksum.c
@@ -95,12 +95,11 @@ void net_checksum_calculate(uint8_t *data, int length)
 
 uint32_t
 net_checksum_add_iov(const struct iovec *iov, const unsigned int iov_cnt,
-                     uint32_t iov_off, uint32_t size)
+                     uint32_t iov_off, uint32_t size, uint32_t csum_offset)
 {
     size_t iovec_off, buf_off;
     unsigned int i;
     uint32_t res = 0;
-    uint32_t seq = 0;
 
     iovec_off = 0;
     buf_off = 0;
@@ -109,8 +108,8 @@ net_checksum_add_iov(const struct iovec *iov, const unsigned int iov_cnt,
             size_t len = MIN((iovec_off + iov[i].iov_len) - iov_off , size);
             void *chunk_buf = iov[i].iov_base + (iov_off - iovec_off);
 
-            res += net_checksum_add_cont(len, chunk_buf, seq);
-            seq += len;
+            res += net_checksum_add_cont(len, chunk_buf, csum_offset);
+            csum_offset += len;
 
             buf_off += len;
             iov_off += len;
diff --git a/net/eth.c b/net/eth.c
index 7e32d274c7..95fe15c23f 100644
--- a/net/eth.c
+++ b/net/eth.c
@@ -21,8 +21,8 @@
 #include "qemu-common.h"
 #include "net/tap.h"
 
-void eth_setup_vlan_headers(struct eth_header *ehdr, uint16_t vlan_tag,
-    bool *is_new)
+void eth_setup_vlan_headers_ex(struct eth_header *ehdr, uint16_t vlan_tag,
+    uint16_t vlan_ethtype, bool *is_new)
 {
     struct vlan_header *vhdr = PKT_GET_VLAN_HDR(ehdr);
 
@@ -36,7 +36,7 @@ void eth_setup_vlan_headers(struct eth_header *ehdr, uint16_t vlan_tag,
     default:
         /* No VLAN header, put a new one */
         vhdr->h_proto = ehdr->h_proto;
-        ehdr->h_proto = cpu_to_be16(ETH_P_VLAN);
+        ehdr->h_proto = cpu_to_be16(vlan_ethtype);
         *is_new = true;
         break;
     }
@@ -79,26 +79,100 @@ eth_get_gso_type(uint16_t l3_proto, uint8_t *l3_hdr, uint8_t l4proto)
     return VIRTIO_NET_HDR_GSO_NONE | ecn_state;
 }
 
-void eth_get_protocols(const uint8_t *headers,
-                       uint32_t hdr_length,
+uint16_t
+eth_get_l3_proto(const struct iovec *l2hdr_iov, int iovcnt, size_t l2hdr_len)
+{
+    uint16_t proto;
+    size_t copied;
+    size_t size = iov_size(l2hdr_iov, iovcnt);
+    size_t proto_offset = l2hdr_len - sizeof(proto);
+
+    if (size < proto_offset) {
+        return ETH_P_UNKNOWN;
+    }
+
+    copied = iov_to_buf(l2hdr_iov, iovcnt, proto_offset,
+                        &proto, sizeof(proto));
+
+    return (copied == sizeof(proto)) ? be16_to_cpu(proto) : ETH_P_UNKNOWN;
+}
+
+static bool
+_eth_copy_chunk(size_t input_size,
+                const struct iovec *iov, int iovcnt,
+                size_t offset, size_t length,
+                void *buffer)
+{
+    size_t copied;
+
+    if (input_size < offset) {
+        return false;
+    }
+
+    copied = iov_to_buf(iov, iovcnt, offset, buffer, length);
+
+    if (copied < length) {
+        return false;
+    }
+
+    return true;
+}
+
+static bool
+_eth_tcp_has_data(bool is_ip4,
+                  const struct ip_header  *ip4_hdr,
+                  const struct ip6_header *ip6_hdr,
+                  size_t full_ip6hdr_len,
+                  const struct tcp_header *tcp)
+{
+    uint32_t l4len;
+
+    if (is_ip4) {
+        l4len = be16_to_cpu(ip4_hdr->ip_len) - IP_HDR_GET_LEN(ip4_hdr);
+    } else {
+        size_t opts_len = full_ip6hdr_len - sizeof(struct ip6_header);
+        l4len = be16_to_cpu(ip6_hdr->ip6_ctlun.ip6_un1.ip6_un1_plen) - opts_len;
+    }
+
+    return l4len > TCP_HEADER_DATA_OFFSET(tcp);
+}
+
+void eth_get_protocols(const struct iovec *iov, int iovcnt,
                        bool *isip4, bool *isip6,
-                       bool *isudp, bool *istcp)
+                       bool *isudp, bool *istcp,
+                       size_t *l3hdr_off,
+                       size_t *l4hdr_off,
+                       size_t *l5hdr_off,
+                       eth_ip6_hdr_info *ip6hdr_info,
+                       eth_ip4_hdr_info *ip4hdr_info,
+                       eth_l4_hdr_info  *l4hdr_info)
 {
     int proto;
-    size_t l2hdr_len = eth_get_l2_hdr_length(headers);
-    assert(hdr_length >= eth_get_l2_hdr_length(headers));
+    bool fragment = false;
+    size_t l2hdr_len = eth_get_l2_hdr_length_iov(iov, iovcnt);
+    size_t input_size = iov_size(iov, iovcnt);
+    size_t copied;
+
     *isip4 = *isip6 = *isudp = *istcp = false;
 
-    proto = eth_get_l3_proto(headers, l2hdr_len);
+    proto = eth_get_l3_proto(iov, iovcnt, l2hdr_len);
+
+    *l3hdr_off = l2hdr_len;
+
     if (proto == ETH_P_IP) {
-        *isip4 = true;
+        struct ip_header *iphdr = &ip4hdr_info->ip4_hdr;
 
-        struct ip_header *iphdr;
+        if (input_size < l2hdr_len) {
+            return;
+        }
+
+        copied = iov_to_buf(iov, iovcnt, l2hdr_len, iphdr, sizeof(*iphdr));
 
-        assert(hdr_length >=
-            eth_get_l2_hdr_length(headers) + sizeof(struct ip_header));
+        *isip4 = true;
 
-        iphdr = PKT_GET_IP_HDR(headers);
+        if (copied < sizeof(*iphdr)) {
+            return;
+        }
 
         if (IP_HEADER_VERSION(iphdr) == IP_HEADER_VERSION_4) {
             if (iphdr->ip_p == IP_PROTO_TCP) {
@@ -107,24 +181,135 @@ void eth_get_protocols(const uint8_t *headers,
                 *isudp = true;
             }
         }
-    } else if (proto == ETH_P_IPV6) {
-        uint8_t l4proto;
-        size_t full_ip6hdr_len;
 
-        struct iovec hdr_vec;
-        hdr_vec.iov_base = (void *) headers;
-        hdr_vec.iov_len = hdr_length;
+        ip4hdr_info->fragment = IP4_IS_FRAGMENT(iphdr);
+        *l4hdr_off = l2hdr_len + IP_HDR_GET_LEN(iphdr);
+
+        fragment = ip4hdr_info->fragment;
+    } else if (proto == ETH_P_IPV6) {
 
         *isip6 = true;
-        if (eth_parse_ipv6_hdr(&hdr_vec, 1, l2hdr_len,
-                              &l4proto, &full_ip6hdr_len)) {
-            if (l4proto == IP_PROTO_TCP) {
+        if (eth_parse_ipv6_hdr(iov, iovcnt, l2hdr_len,
+                               ip6hdr_info)) {
+            if (ip6hdr_info->l4proto == IP_PROTO_TCP) {
                 *istcp = true;
-            } else if (l4proto == IP_PROTO_UDP) {
+            } else if (ip6hdr_info->l4proto == IP_PROTO_UDP) {
                 *isudp = true;
             }
+        } else {
+            return;
+        }
+
+        *l4hdr_off = l2hdr_len + ip6hdr_info->full_hdr_len;
+        fragment = ip6hdr_info->fragment;
+    }
+
+    if (!fragment) {
+        if (*istcp) {
+            *istcp = _eth_copy_chunk(input_size,
+                                     iov, iovcnt,
+                                     *l4hdr_off, sizeof(l4hdr_info->hdr.tcp),
+                                     &l4hdr_info->hdr.tcp);
+
+            if (istcp) {
+                *l5hdr_off = *l4hdr_off +
+                    TCP_HEADER_DATA_OFFSET(&l4hdr_info->hdr.tcp);
+
+                l4hdr_info->has_tcp_data =
+                    _eth_tcp_has_data(proto == ETH_P_IP,
+                                      &ip4hdr_info->ip4_hdr,
+                                      &ip6hdr_info->ip6_hdr,
+                                      *l4hdr_off - *l3hdr_off,
+                                      &l4hdr_info->hdr.tcp);
+            }
+        } else if (*isudp) {
+            *isudp = _eth_copy_chunk(input_size,
+                                     iov, iovcnt,
+                                     *l4hdr_off, sizeof(l4hdr_info->hdr.udp),
+                                     &l4hdr_info->hdr.udp);
+            *l5hdr_off = *l4hdr_off + sizeof(l4hdr_info->hdr.udp);
+        }
+    }
+}
+
+bool
+eth_strip_vlan(const struct iovec *iov, int iovcnt, size_t iovoff,
+               uint8_t *new_ehdr_buf,
+               uint16_t *payload_offset, uint16_t *tci)
+{
+    struct vlan_header vlan_hdr;
+    struct eth_header *new_ehdr = (struct eth_header *) new_ehdr_buf;
+
+    size_t copied = iov_to_buf(iov, iovcnt, iovoff,
+                               new_ehdr, sizeof(*new_ehdr));
+
+    if (copied < sizeof(*new_ehdr)) {
+        return false;
+    }
+
+    switch (be16_to_cpu(new_ehdr->h_proto)) {
+    case ETH_P_VLAN:
+    case ETH_P_DVLAN:
+        copied = iov_to_buf(iov, iovcnt, iovoff + sizeof(*new_ehdr),
+                            &vlan_hdr, sizeof(vlan_hdr));
+
+        if (copied < sizeof(vlan_hdr)) {
+            return false;
+        }
+
+        new_ehdr->h_proto = vlan_hdr.h_proto;
+
+        *tci = be16_to_cpu(vlan_hdr.h_tci);
+        *payload_offset = iovoff + sizeof(*new_ehdr) + sizeof(vlan_hdr);
+
+        if (be16_to_cpu(new_ehdr->h_proto) == ETH_P_VLAN) {
+
+            copied = iov_to_buf(iov, iovcnt, *payload_offset,
+                                PKT_GET_VLAN_HDR(new_ehdr), sizeof(vlan_hdr));
+
+            if (copied < sizeof(vlan_hdr)) {
+                return false;
+            }
+
+            *payload_offset += sizeof(vlan_hdr);
+        }
+        return true;
+    default:
+        return false;
+    }
+}
+
+bool
+eth_strip_vlan_ex(const struct iovec *iov, int iovcnt, size_t iovoff,
+                  uint16_t vet, uint8_t *new_ehdr_buf,
+                  uint16_t *payload_offset, uint16_t *tci)
+{
+    struct vlan_header vlan_hdr;
+    struct eth_header *new_ehdr = (struct eth_header *) new_ehdr_buf;
+
+    size_t copied = iov_to_buf(iov, iovcnt, iovoff,
+                               new_ehdr, sizeof(*new_ehdr));
+
+    if (copied < sizeof(*new_ehdr)) {
+        return false;
+    }
+
+    if (be16_to_cpu(new_ehdr->h_proto) == vet) {
+        copied = iov_to_buf(iov, iovcnt, iovoff + sizeof(*new_ehdr),
+                            &vlan_hdr, sizeof(vlan_hdr));
+
+        if (copied < sizeof(vlan_hdr)) {
+            return false;
         }
+
+        new_ehdr->h_proto = vlan_hdr.h_proto;
+
+        *tci = be16_to_cpu(vlan_hdr.h_tci);
+        *payload_offset = iovoff + sizeof(*new_ehdr) + sizeof(vlan_hdr);
+        return true;
     }
+
+    return false;
 }
 
 void
@@ -133,7 +318,12 @@ eth_setup_ip4_fragmentation(const void *l2hdr, size_t l2hdr_len,
                             size_t l3payload_len,
                             size_t frag_offset, bool more_frags)
 {
-    if (eth_get_l3_proto(l2hdr, l2hdr_len) == ETH_P_IP) {
+    const struct iovec l2vec = {
+        .iov_base = (void *) l2hdr,
+        .iov_len = l2hdr_len
+    };
+
+    if (eth_get_l3_proto(&l2vec, 1, l2hdr_len) == ETH_P_IP) {
         uint16_t orig_flags;
         struct ip_header *iphdr = (struct ip_header *) l3hdr;
         uint16_t frag_off_units = frag_offset / IP_FRAG_UNIT_SIZE;
@@ -158,7 +348,9 @@ eth_fix_ip4_checksum(void *l3hdr, size_t l3hdr_len)
 }
 
 uint32_t
-eth_calc_pseudo_hdr_csum(struct ip_header *iphdr, uint16_t csl)
+eth_calc_ip4_pseudo_hdr_csum(struct ip_header *iphdr,
+                             uint16_t csl,
+                             uint32_t *cso)
 {
     struct ip_pseudo_header ipph;
     ipph.ip_src = iphdr->ip_src;
@@ -166,7 +358,26 @@ eth_calc_pseudo_hdr_csum(struct ip_header *iphdr, uint16_t csl)
     ipph.ip_payload = cpu_to_be16(csl);
     ipph.ip_proto = iphdr->ip_p;
     ipph.zeros = 0;
-    return net_checksum_add(sizeof(ipph), (uint8_t *) &ipph);
+    *cso = sizeof(ipph);
+    return net_checksum_add(*cso, (uint8_t *) &ipph);
+}
+
+uint32_t
+eth_calc_ip6_pseudo_hdr_csum(struct ip6_header *iphdr,
+                             uint16_t csl,
+                             uint8_t l4_proto,
+                             uint32_t *cso)
+{
+    struct ip6_pseudo_header ipph;
+    ipph.ip6_src = iphdr->ip6_src;
+    ipph.ip6_dst = iphdr->ip6_dst;
+    ipph.len = cpu_to_be16(csl);
+    ipph.zero[0] = 0;
+    ipph.zero[1] = 0;
+    ipph.zero[2] = 0;
+    ipph.next_hdr = l4_proto;
+    *cso = sizeof(ipph);
+    return net_checksum_add(*cso, (uint8_t *)&ipph);
 }
 
 static bool
@@ -186,33 +397,152 @@ eth_is_ip6_extension_header_type(uint8_t hdr_type)
     }
 }
 
-bool eth_parse_ipv6_hdr(struct iovec *pkt, int pkt_frags,
-                        size_t ip6hdr_off, uint8_t *l4proto,
-                        size_t *full_hdr_len)
+static bool
+_eth_get_rss_ex_dst_addr(const struct iovec *pkt, int pkt_frags,
+                        size_t rthdr_offset,
+                        struct ip6_ext_hdr *ext_hdr,
+                        struct in6_address *dst_addr)
+{
+    struct ip6_ext_hdr_routing *rthdr = (struct ip6_ext_hdr_routing *) ext_hdr;
+
+    if ((rthdr->rtype == 2) &&
+        (rthdr->len == sizeof(struct in6_address) / 8) &&
+        (rthdr->segleft == 1)) {
+
+        size_t input_size = iov_size(pkt, pkt_frags);
+        size_t bytes_read;
+
+        if (input_size < rthdr_offset + sizeof(*ext_hdr)) {
+            return false;
+        }
+
+        bytes_read = iov_to_buf(pkt, pkt_frags,
+                                rthdr_offset + sizeof(*ext_hdr),
+                                dst_addr, sizeof(dst_addr));
+
+        return bytes_read == sizeof(dst_addr);
+    }
+
+    return false;
+}
+
+static bool
+_eth_get_rss_ex_src_addr(const struct iovec *pkt, int pkt_frags,
+                        size_t dsthdr_offset,
+                        struct ip6_ext_hdr *ext_hdr,
+                        struct in6_address *src_addr)
+{
+    size_t bytes_left = (ext_hdr->ip6r_len + 1) * 8 - sizeof(*ext_hdr);
+    struct ip6_option_hdr opthdr;
+    size_t opt_offset = dsthdr_offset + sizeof(*ext_hdr);
+
+    while (bytes_left > sizeof(opthdr)) {
+        size_t input_size = iov_size(pkt, pkt_frags);
+        size_t bytes_read, optlen;
+
+        if (input_size < opt_offset) {
+            return false;
+        }
+
+        bytes_read = iov_to_buf(pkt, pkt_frags, opt_offset,
+                                &opthdr, sizeof(opthdr));
+
+        if (bytes_read != sizeof(opthdr)) {
+            return false;
+        }
+
+        optlen = (opthdr.type == IP6_OPT_PAD1) ? 1
+                                               : (opthdr.len + sizeof(opthdr));
+
+        if (optlen > bytes_left) {
+            return false;
+        }
+
+        if (opthdr.type == IP6_OPT_HOME) {
+            size_t input_size = iov_size(pkt, pkt_frags);
+
+            if (input_size < opt_offset + sizeof(opthdr)) {
+                return false;
+            }
+
+            bytes_read = iov_to_buf(pkt, pkt_frags,
+                                    opt_offset + sizeof(opthdr),
+                                    src_addr, sizeof(src_addr));
+
+            return bytes_read == sizeof(src_addr);
+        }
+
+        opt_offset += optlen;
+        bytes_left -= optlen;
+    }
+
+    return false;
+}
+
+bool eth_parse_ipv6_hdr(const struct iovec *pkt, int pkt_frags,
+                        size_t ip6hdr_off, eth_ip6_hdr_info *info)
 {
-    struct ip6_header ip6_hdr;
     struct ip6_ext_hdr ext_hdr;
     size_t bytes_read;
+    uint8_t curr_ext_hdr_type;
+    size_t input_size = iov_size(pkt, pkt_frags);
+
+    info->rss_ex_dst_valid = false;
+    info->rss_ex_src_valid = false;
+    info->fragment = false;
+
+    if (input_size < ip6hdr_off) {
+        return false;
+    }
 
     bytes_read = iov_to_buf(pkt, pkt_frags, ip6hdr_off,
-                            &ip6_hdr, sizeof(ip6_hdr));
-    if (bytes_read < sizeof(ip6_hdr)) {
+                            &info->ip6_hdr, sizeof(info->ip6_hdr));
+    if (bytes_read < sizeof(info->ip6_hdr)) {
         return false;
     }
 
-    *full_hdr_len = sizeof(struct ip6_header);
+    info->full_hdr_len = sizeof(struct ip6_header);
+
+    curr_ext_hdr_type = info->ip6_hdr.ip6_nxt;
 
-    if (!eth_is_ip6_extension_header_type(ip6_hdr.ip6_nxt)) {
-        *l4proto = ip6_hdr.ip6_nxt;
+    if (!eth_is_ip6_extension_header_type(curr_ext_hdr_type)) {
+        info->l4proto = info->ip6_hdr.ip6_nxt;
+        info->has_ext_hdrs = false;
         return true;
     }
 
+    info->has_ext_hdrs = true;
+
     do {
-        bytes_read = iov_to_buf(pkt, pkt_frags, ip6hdr_off + *full_hdr_len,
+        if (input_size < ip6hdr_off + info->full_hdr_len) {
+            return false;
+        }
+
+        bytes_read = iov_to_buf(pkt, pkt_frags, ip6hdr_off + info->full_hdr_len,
                                 &ext_hdr, sizeof(ext_hdr));
-        *full_hdr_len += (ext_hdr.ip6r_len + 1) * IP6_EXT_GRANULARITY;
-    } while (eth_is_ip6_extension_header_type(ext_hdr.ip6r_nxt));
 
-    *l4proto = ext_hdr.ip6r_nxt;
+        if (bytes_read < sizeof(ext_hdr)) {
+            return false;
+        }
+
+        if (curr_ext_hdr_type == IP6_ROUTING) {
+            info->rss_ex_dst_valid =
+                _eth_get_rss_ex_dst_addr(pkt, pkt_frags,
+                                         ip6hdr_off + info->full_hdr_len,
+                                         &ext_hdr, &info->rss_ex_dst);
+        } else if (curr_ext_hdr_type == IP6_DESTINATON) {
+            info->rss_ex_src_valid =
+                _eth_get_rss_ex_src_addr(pkt, pkt_frags,
+                                         ip6hdr_off + info->full_hdr_len,
+                                         &ext_hdr, &info->rss_ex_src);
+        } else if (curr_ext_hdr_type == IP6_FRAGMENT) {
+            info->fragment = true;
+        }
+
+        info->full_hdr_len += (ext_hdr.ip6r_len + 1) * IP6_EXT_GRANULARITY;
+        curr_ext_hdr_type = ext_hdr.ip6r_nxt;
+    } while (eth_is_ip6_extension_header_type(curr_ext_hdr_type));
+
+    info->l4proto = ext_hdr.ip6r_nxt;
     return true;
 }