X-Git-Url: http://git.onelab.eu/?a=blobdiff_plain;f=lib%2Fflow.c;h=13d0aa55933da9e55dee1656241b51d4b4223531;hb=1b84927316e090bc67cc6d923356857b52288670;hp=63c6ef80a87317159b9937f3366ef008e98cb294;hpb=2a4ca27c2659dbe5b07f7c1faa301bce17704ec3;p=sliver-openvswitch.git

diff --git a/lib/flow.c b/lib/flow.c
index 63c6ef80a..13d0aa559 100644
--- a/lib/flow.c
+++ b/lib/flow.c
@@ -41,6 +41,14 @@
 COVERAGE_DEFINE(flow_extract);
 COVERAGE_DEFINE(miniflow_malloc);
 
+/* U32 indices for segmented flow classification. */
+const uint8_t flow_segment_u32s[4] = {
+    FLOW_SEGMENT_1_ENDS_AT / 4,
+    FLOW_SEGMENT_2_ENDS_AT / 4,
+    FLOW_SEGMENT_3_ENDS_AT / 4,
+    FLOW_U32S
+};
+
 static struct arp_eth_header *
 pull_arp(struct ofpbuf *packet)
 {
@@ -511,11 +519,23 @@ flow_zero_wildcards(struct flow *flow, const struct flow_wildcards *wildcards)
     }
 }
 
+void
+flow_unwildcard_tp_ports(const struct flow *flow, struct flow_wildcards *wc)
+{
+    if (flow->nw_proto != IPPROTO_ICMP) {
+        memset(&wc->masks.tp_src, 0xff, sizeof wc->masks.tp_src);
+        memset(&wc->masks.tp_dst, 0xff, sizeof wc->masks.tp_dst);
+    } else {
+        wc->masks.tp_src = htons(0xff);
+        wc->masks.tp_dst = htons(0xff);
+    }
+}
+
 /* Initializes 'fmd' with the metadata found in 'flow'. */
 void
 flow_get_metadata(const struct flow *flow, struct flow_metadata *fmd)
 {
-    BUILD_ASSERT_DECL(FLOW_WC_SEQ == 22);
+    BUILD_ASSERT_DECL(FLOW_WC_SEQ == 23);
 
     fmd->tun_id = flow->tunnel.tun_id;
     fmd->tun_src = flow->tunnel.ip_src;
@@ -578,6 +598,24 @@ format_flags(struct ds *ds, const char *(*bit_to_string)(uint32_t),
     ds_chomp(ds, del);
 }
 
+void
+format_flags_masked(struct ds *ds, const char *name,
+                    const char *(*bit_to_string)(uint32_t), uint32_t flags,
+                    uint32_t mask)
+{
+    if (name) {
+        ds_put_format(ds, "%s=", name);
+    }
+    while (mask) {
+        uint32_t bit = rightmost_1bit(mask);
+        const char *s = bit_to_string(bit);
+
+        ds_put_format(ds, "%s%s", (flags & bit) ? "+" : "-",
+                      s ? s : "[Unknown]");
+        mask &= ~bit;
+    }
+}
+
 void
 flow_format(struct ds *ds, const struct flow *flow)
 {
@@ -604,6 +642,15 @@ flow_wildcards_init_catchall(struct flow_wildcards *wc)
     memset(&wc->masks, 0, sizeof wc->masks);
 }
 
+/* Clear the metadata and register wildcard masks. They are not packet
+ * header fields. */
+void
+flow_wildcards_clear_non_packet_fields(struct flow_wildcards *wc)
+{
+    memset(&wc->masks.metadata, 0, sizeof wc->masks.metadata);
+    memset(&wc->masks.regs, 0, sizeof wc->masks.regs);
+}
+
 /* Returns true if 'wc' matches every packet, false if 'wc' fixes any bits or
  * fields. */
 bool
@@ -662,11 +709,11 @@ static void
 flow_union_with_miniflow(struct flow *dst, const struct miniflow *src)
 {
     uint32_t *dst_u32 = (uint32_t *) dst;
-    int ofs = 0;
+    const uint32_t *p = src->values;
     uint64_t map;
 
     for (map = src->map; map; map = zero_rightmost_1bit(map)) {
-        dst_u32[raw_ctz(map)] |= src->values[ofs++];
+        dst_u32[raw_ctz(map)] |= *p++;
     }
 }
 
@@ -678,6 +725,43 @@ flow_wildcards_fold_minimask(struct flow_wildcards *wc,
     flow_union_with_miniflow(&wc->masks, &mask->masks);
 }
 
+inline uint64_t
+miniflow_get_map_in_range(const struct miniflow *miniflow,
+                          uint8_t start, uint8_t end, const uint32_t **data)
+{
+    uint64_t map = miniflow->map;
+    uint32_t *p = miniflow->values;
+
+    if (start > 0) {
+        uint64_t msk = (UINT64_C(1) << start) - 1; /* 'start' LSBs set */
+        p += count_1bits(map & msk);  /* Skip to start. */
+        map &= ~msk;
+    }
+    if (end < FLOW_U32S) {
+        uint64_t msk = (UINT64_C(1) << end) - 1; /* 'end' LSBs set */
+        map &= msk;
+    }
+
+    *data = p;
+    return map;
+}
+
+/* Fold minimask 'mask''s wildcard mask into 'wc's wildcard mask
+ * in range [start, end). */
+void
+flow_wildcards_fold_minimask_range(struct flow_wildcards *wc,
+                                   const struct minimask *mask,
+                                   uint8_t start, uint8_t end)
+{
+    uint32_t *dst_u32 = (uint32_t *)&wc->masks;
+    const uint32_t *p;
+    uint64_t map = miniflow_get_map_in_range(&mask->masks, start, end, &p);
+
+    for (; map; map = zero_rightmost_1bit(map)) {
+        dst_u32[raw_ctz(map)] |= *p++;
+    }
+}
+
 /* Returns a hash of the wildcards in 'wc'. */
 uint32_t
 flow_wildcards_hash(const struct flow_wildcards *wc, uint32_t basis)
@@ -850,14 +934,13 @@ flow_mask_hash_fields(const struct flow *flow, struct flow_wildcards *wc,
         }
         if (is_ip_any(flow)) {
             memset(&wc->masks.nw_proto, 0xff, sizeof wc->masks.nw_proto);
-            memset(&wc->masks.tp_src, 0xff, sizeof wc->masks.tp_src);
-            memset(&wc->masks.tp_dst, 0xff, sizeof wc->masks.tp_dst);
+            flow_unwildcard_tp_ports(flow, wc);
         }
         wc->masks.vlan_tci |= htons(VLAN_VID_MASK | VLAN_CFI);
         break;
 
     default:
-        NOT_REACHED();
+        OVS_NOT_REACHED();
     }
 }
 
@@ -875,7 +958,7 @@ flow_hash_fields(const struct flow *flow, enum nx_hash_fields fields,
         return flow_hash_symmetric_l4(flow, basis);
     }
 
-    NOT_REACHED();
+    OVS_NOT_REACHED();
 }
 
 /* Returns a string representation of 'fields'. */
@@ -995,6 +1078,78 @@ flow_set_mpls_bos(struct flow *flow, uint8_t bos)
     set_mpls_lse_bos(&flow->mpls_lse, bos);
 }
 
+
+static void
+flow_compose_l4(struct ofpbuf *b, const struct flow *flow)
+{
+    if (!(flow->nw_frag & FLOW_NW_FRAG_ANY)
+        || !(flow->nw_frag & FLOW_NW_FRAG_LATER)) {
+        if (flow->nw_proto == IPPROTO_TCP) {
+            struct tcp_header *tcp;
+
+            tcp = ofpbuf_put_zeros(b, sizeof *tcp);
+            tcp->tcp_src = flow->tp_src;
+            tcp->tcp_dst = flow->tp_dst;
+            tcp->tcp_ctl = TCP_CTL(ntohs(flow->tcp_flags), 5);
+            b->l7 = ofpbuf_tail(b);
+        } else if (flow->nw_proto == IPPROTO_UDP) {
+            struct udp_header *udp;
+
+            udp = ofpbuf_put_zeros(b, sizeof *udp);
+            udp->udp_src = flow->tp_src;
+            udp->udp_dst = flow->tp_dst;
+            b->l7 = ofpbuf_tail(b);
+        } else if (flow->nw_proto == IPPROTO_SCTP) {
+            struct sctp_header *sctp;
+
+            sctp = ofpbuf_put_zeros(b, sizeof *sctp);
+            sctp->sctp_src = flow->tp_src;
+            sctp->sctp_dst = flow->tp_dst;
+            b->l7 = ofpbuf_tail(b);
+        } else if (flow->nw_proto == IPPROTO_ICMP) {
+            struct icmp_header *icmp;
+
+            icmp = ofpbuf_put_zeros(b, sizeof *icmp);
+            icmp->icmp_type = ntohs(flow->tp_src);
+            icmp->icmp_code = ntohs(flow->tp_dst);
+            icmp->icmp_csum = csum(icmp, ICMP_HEADER_LEN);
+            b->l7 = ofpbuf_tail(b);
+        } else if (flow->nw_proto == IPPROTO_ICMPV6) {
+            struct icmp6_hdr *icmp;
+
+            icmp = ofpbuf_put_zeros(b, sizeof *icmp);
+            icmp->icmp6_type = ntohs(flow->tp_src);
+            icmp->icmp6_code = ntohs(flow->tp_dst);
+
+            if (icmp->icmp6_code == 0 &&
+                (icmp->icmp6_type == ND_NEIGHBOR_SOLICIT ||
+                 icmp->icmp6_type == ND_NEIGHBOR_ADVERT)) {
+                struct in6_addr *nd_target;
+                struct nd_opt_hdr *nd_opt;
+
+                nd_target = ofpbuf_put_zeros(b, sizeof *nd_target);
+                *nd_target = flow->nd_target;
+
+                if (!eth_addr_is_zero(flow->arp_sha)) {
+                    nd_opt = ofpbuf_put_zeros(b, 8);
+                    nd_opt->nd_opt_len = 1;
+                    nd_opt->nd_opt_type = ND_OPT_SOURCE_LINKADDR;
+                    memcpy(nd_opt + 1, flow->arp_sha, ETH_ADDR_LEN);
+                }
+                if (!eth_addr_is_zero(flow->arp_tha)) {
+                    nd_opt = ofpbuf_put_zeros(b, 8);
+                    nd_opt->nd_opt_len = 1;
+                    nd_opt->nd_opt_type = ND_OPT_TARGET_LINKADDR;
+                    memcpy(nd_opt + 1, flow->arp_tha, ETH_ADDR_LEN);
+                }
+            }
+            icmp->icmp6_cksum = (OVS_FORCE uint16_t)
+                csum(icmp, (char *)ofpbuf_tail(b) - (char *)icmp);
+            b->l7 = ofpbuf_tail(b);
+        }
+    }
+}
+
 /* Puts into 'b' a packet that flow_extract() would parse as having the given
  * 'flow'.
  *
@@ -1004,6 +1159,7 @@ flow_set_mpls_bos(struct flow *flow, uint8_t bos)
 void
 flow_compose(struct ofpbuf *b, const struct flow *flow)
 {
+    /* eth_compose() sets l3 pointer and makes sure it is 32-bit aligned. */
     eth_compose(b, flow->dl_dst, flow->dl_src, ntohs(flow->dl_type), 0);
     if (flow->dl_type == htons(FLOW_DL_TYPE_NONE)) {
         struct eth_header *eth = b->l2;
@@ -1018,7 +1174,7 @@ flow_compose(struct ofpbuf *b, const struct flow *flow)
     if (flow->dl_type == htons(ETH_TYPE_IP)) {
         struct ip_header *ip;
 
-        b->l3 = ip = ofpbuf_put_zeros(b, sizeof *ip);
+        ip = ofpbuf_put_zeros(b, sizeof *ip);
         ip->ip_ihl_ver = IP_IHL_VER(5, 4);
         ip->ip_tos = flow->nw_tos;
         ip->ip_ttl = flow->nw_ttl;
@@ -1032,44 +1188,32 @@ flow_compose(struct ofpbuf *b, const struct flow *flow)
                 ip->ip_frag_off |= htons(100);
             }
         }
-        if (!(flow->nw_frag & FLOW_NW_FRAG_ANY)
-            || !(flow->nw_frag & FLOW_NW_FRAG_LATER)) {
-            if (flow->nw_proto == IPPROTO_TCP) {
-                struct tcp_header *tcp;
-
-                b->l4 = tcp = ofpbuf_put_zeros(b, sizeof *tcp);
-                tcp->tcp_src = flow->tp_src;
-                tcp->tcp_dst = flow->tp_dst;
-                tcp->tcp_ctl = TCP_CTL(ntohs(flow->tcp_flags), 5);
-            } else if (flow->nw_proto == IPPROTO_UDP) {
-                struct udp_header *udp;
-
-                b->l4 = udp = ofpbuf_put_zeros(b, sizeof *udp);
-                udp->udp_src = flow->tp_src;
-                udp->udp_dst = flow->tp_dst;
-            } else if (flow->nw_proto == IPPROTO_SCTP) {
-                struct sctp_header *sctp;
-
-                b->l4 = sctp = ofpbuf_put_zeros(b, sizeof *sctp);
-                sctp->sctp_src = flow->tp_src;
-                sctp->sctp_dst = flow->tp_dst;
-            } else if (flow->nw_proto == IPPROTO_ICMP) {
-                struct icmp_header *icmp;
-
-                b->l4 = icmp = ofpbuf_put_zeros(b, sizeof *icmp);
-                icmp->icmp_type = ntohs(flow->tp_src);
-                icmp->icmp_code = ntohs(flow->tp_dst);
-                icmp->icmp_csum = csum(icmp, ICMP_HEADER_LEN);
-            }
-            b->l7 = ofpbuf_tail(b);
-        }
 
-        ip = b->l3;
+        b->l4 = ofpbuf_tail(b);
+
+        flow_compose_l4(b, flow);
+
         ip->ip_tot_len = htons((uint8_t *) b->data + b->size
                                - (uint8_t *) b->l3);
         ip->ip_csum = csum(ip, sizeof *ip);
     } else if (flow->dl_type == htons(ETH_TYPE_IPV6)) {
-        /* XXX */
+        struct ovs_16aligned_ip6_hdr *nh;
+
+        nh = ofpbuf_put_zeros(b, sizeof *nh);
+        put_16aligned_be32(&nh->ip6_flow, htonl(6 << 28) |
+                           htonl(flow->nw_tos << 20) | flow->ipv6_label);
+        nh->ip6_hlim = flow->nw_ttl;
+        nh->ip6_nxt = flow->nw_proto;
+
+        memcpy(&nh->ip6_src, &flow->ipv6_src, sizeof(nh->ip6_src));
+        memcpy(&nh->ip6_dst, &flow->ipv6_dst, sizeof(nh->ip6_dst));
+
+        b->l4 = ofpbuf_tail(b);
+
+        flow_compose_l4(b, flow);
+
+        nh->ip6_plen =
+            b->l7 ? htons((uint8_t *) b->l7 - (uint8_t *) b->l4) : htons(0);
     } else if (flow->dl_type == htons(ETH_TYPE_ARP) ||
                flow->dl_type == htons(ETH_TYPE_RARP)) {
         struct arp_eth_header *arp;
@@ -1101,7 +1245,7 @@ flow_compose(struct ofpbuf *b, const struct flow *flow)
 static int
 miniflow_n_values(const struct miniflow *flow)
 {
-    return popcount(flow->map);
+    return count_1bits(flow->map);
 }
 
 static uint32_t *
@@ -1117,12 +1261,16 @@ miniflow_alloc_values(struct miniflow *flow, int n)
 
 /* Completes an initialization of 'dst' as a miniflow copy of 'src' begun by
  * the caller.  The caller must have already initialized 'dst->map' properly
- * to indicate the nonzero uint32_t elements of 'src'.  'n' must be the number
- * of 1-bits in 'dst->map'.
+ * to indicate the significant uint32_t elements of 'src'.  'n' must be the
+ * number of 1-bits in 'dst->map'.
+ *
+ * Normally the significant elements are the ones that are non-zero.  However,
+ * when a miniflow is initialized from a (mini)mask, the values can be zeroes,
+ * so that the flow and mask always have the same maps.
  *
  * This function initializes 'dst->values' (either inline if possible or with
- * malloc() otherwise) and copies the nonzero uint32_t elements of 'src' into
- * it. */
+ * malloc() otherwise) and copies the uint32_t elements of 'src' indicated by
+ * 'dst->map' into it. */
 static void
 miniflow_init__(struct miniflow *dst, const struct flow *src, int n)
 {
@@ -1221,7 +1369,8 @@ miniflow_get__(const struct miniflow *flow, unsigned int u32_ofs)
         static const uint32_t zero = 0;
         return &zero;
     }
-    return flow->values + popcount(flow->map & ((UINT64_C(1) << u32_ofs) - 1));
+    return flow->values +
+           count_1bits(flow->map & ((UINT64_C(1) << u32_ofs) - 1));
 }
 
 /* Returns the uint32_t that would be at byte offset '4 * u32_ofs' if 'flow'
@@ -1401,6 +1550,33 @@ flow_hash_in_minimask(const struct flow *flow, const struct minimask *mask,
 
     return mhash_finish(hash, (p - mask->masks.values) * 4);
 }
+
+/* Returns a hash value for the bits of range [start, end) in 'flow',
+ * where there are 1-bits in 'mask', given 'hash'.
+ *
+ * The hash values returned by this function are the same as those returned by
+ * minimatch_hash_range(), only the form of the arguments differ. */
+uint32_t
+flow_hash_in_minimask_range(const struct flow *flow,
+                            const struct minimask *mask,
+                            uint8_t start, uint8_t end, uint32_t *basis)
+{
+    const uint32_t *flow_u32 = (const uint32_t *)flow;
+    const uint32_t *p;
+    uint64_t map = miniflow_get_map_in_range(&mask->masks, start, end, &p);
+    uint32_t hash = *basis;
+
+    for (; map; map = zero_rightmost_1bit(map)) {
+        if (*p) {
+            hash = mhash_add(hash, flow_u32[raw_ctz(map)] & *p);
+        }
+        p++;
+    }
+
+    *basis = hash; /* Allow continuation from the unfinished value. */
+    return mhash_finish(hash, (p - mask->masks.values) * 4);
+}
+
 
 /* Initializes 'dst' as a copy of 'src'.  The caller must eventually free 'dst'
  * with minimask_destroy(). */