From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from dpdk.org (dpdk.org [92.243.14.124]) by inbox.dpdk.org (Postfix) with ESMTP id DDA30A09FE; Mon, 21 Dec 2020 04:51:20 +0100 (CET) Received: from [92.243.14.124] (localhost [127.0.0.1]) by dpdk.org (Postfix) with ESMTP id 45F3ACB7A; Mon, 21 Dec 2020 04:51:01 +0100 (CET) Received: from mail-m972.mail.163.com (mail-m972.mail.163.com [123.126.97.2]) by dpdk.org (Postfix) with ESMTP id AE2BCCACA for ; Mon, 21 Dec 2020 04:50:57 +0100 (CET) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=163.com; s=s110527; h=From:Subject:Date:Message-Id:MIME-Version; bh=KjCD6 LqiRvtO8tUdrIX6aoYWlLPBVmPlPI77PfNdknw=; b=JomZ5sArk+bybVyUjBax6 +wXZSHAmCEwAydXLR3nMUzdlNQH9iurhCe2obdH422S9hDmfrLjdR8JcEy4Fkt/Z Qzop7l6hrH9GwRczL+p385qoB8KynYvl1yWQv7bESsy3m74jyQrl5fE1f+5Ksu9f fahaC5ilR47v3tuBiIQ1I4= Received: from yangyi0100.home.langchao.com (unknown [111.207.123.53]) by smtp2 (Coremail) with SMTP id GtxpCgD3vb6cG+BfUOf7Fw--.755S3; Mon, 21 Dec 2020 11:50:54 +0800 (CST) From: yang_y_yi@163.com To: dev@dpdk.org Cc: jiayu.hu@intel.com, konstantin.ananyev@intel.com, thomas@monjalon.net, yangyi01@inspur.com, yang_y_yi@163.com Date: Mon, 21 Dec 2020 11:50:45 +0800 Message-Id: <20201221035052.128292-2-yang_y_yi@163.com> X-Mailer: git-send-email 2.19.2.windows.1 In-Reply-To: <20201221035052.128292-1-yang_y_yi@163.com> References: <20201221035052.128292-1-yang_y_yi@163.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-CM-TRANSID: GtxpCgD3vb6cG+BfUOf7Fw--.755S3 X-Coremail-Antispam: 1Uf129KBjvAXoWfuF4rGw4fZry8Aw4kAFy8Grg_yoW5WF1UKo WxZw15C34v9rnYvr40yr1vgF42g3W8W3WfZr4YgrWvqan2yryDG393Ww43Aay7Zr43tFnY v34xJayxXF4fArZ3n29KB7ZKAUJUUUUU529EdanIXcx71UUUUU7v73VFW2AGmfu7bjvjm3 AaLaJ3UbIYCTnIWIevJa73UjIFyTuYvjxU5NVkUUUUU X-Originating-IP: [111.207.123.53] X-CM-SenderInfo: 51dqwsp1b1xqqrwthudrp/1tbiEB4Ci18YEc33RAAAsv Subject: [dpdk-dev] [PATCH v1 1/8] gro: support TCP/IPv6 X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org Sender: "dev" From: Yi Yang TCP/IPv6 GRO can help improve TCP/IPv6 performance. With this enabled in DPDK, OVS DPDK can leverage it to improve VM-to-VM TCP/IPv6 performance, it will merge small adjacent TCP/IPv6 segments to a big TCP/IPv6 segment immediate after they are received from a physical NIC. It is very helpful in OVS DPDK VLAN use case. Signed-off-by: Yi Yang --- .../prog_guide/generic_receive_offload_lib.rst | 10 +- doc/guides/rel_notes/release_21_02.rst | 5 + lib/librte_gro/gro_tcp6.c | 368 +++++++++++++++++++++ lib/librte_gro/gro_tcp6.h | 314 ++++++++++++++++++ lib/librte_gro/meson.build | 2 +- lib/librte_gro/rte_gro.c | 74 ++++- lib/librte_gro/rte_gro.h | 3 + 7 files changed, 762 insertions(+), 14 deletions(-) create mode 100644 lib/librte_gro/gro_tcp6.c create mode 100644 lib/librte_gro/gro_tcp6.h diff --git a/doc/guides/prog_guide/generic_receive_offload_lib.rst b/doc/guides/prog_guide/generic_receive_offload_lib.rst index 98a5d29..fee856e 100644 --- a/doc/guides/prog_guide/generic_receive_offload_lib.rst +++ b/doc/guides/prog_guide/generic_receive_offload_lib.rst @@ -30,9 +30,9 @@ assumes the packets are complete (i.e., MF==0 && frag_off==0), when IP fragmentation is possible (i.e., DF==0). Additionally, it complies RFC 6864 to process the IPv4 ID field. -Currently, the GRO library provides GRO supports for TCP/IPv4 and UDP/IPv4 -packets as well as VxLAN packets which contain an outer IPv4 header and an -inner TCP/IPv4 or UDP/IPv4 packet. +Currently, the GRO library provides GRO supports for TCP/IPv4, UDP/IPv4, +and TCP/IPv6 packets as well as VxLAN packets which contain an outer IPv4 +header and an inner TCP/IPv4 or UDP/IPv4 packet. Two Sets of API --------------- @@ -83,8 +83,8 @@ Reassembly Algorithm The reassembly algorithm is used for reassembling packets. In the GRO library, different GRO types can use different algorithms. In this -section, we will introduce an algorithm, which is used by TCP/IPv4 GRO -and VxLAN GRO. +section, we will introduce an algorithm, which is used by TCP/IPv4 GRO, +TCP/IPv6 GRO and VxLAN GRO. Challenges ~~~~~~~~~~ diff --git a/doc/guides/rel_notes/release_21_02.rst b/doc/guides/rel_notes/release_21_02.rst index 638f981..1a1aeff 100644 --- a/doc/guides/rel_notes/release_21_02.rst +++ b/doc/guides/rel_notes/release_21_02.rst @@ -55,6 +55,11 @@ New Features Also, make sure to start the actual text at the margin. ======================================================= +* **Added TCP/IPv6 GRO support for non-VXLAN packets.** + + Added TCP/IPv6 GRO support for non-VXLAN packets, this enabled TCP/IPv6 + GRO support for VLAN use case. + Removed Items ------------- diff --git a/lib/librte_gro/gro_tcp6.c b/lib/librte_gro/gro_tcp6.c new file mode 100644 index 0000000..4a4df34 --- /dev/null +++ b/lib/librte_gro/gro_tcp6.c @@ -0,0 +1,368 @@ +/* SPDX-License-Identifier: BSD-3-Clause + * Copyright(c) 2021 Inspur Corporation + */ + +#include +#include +#include +#include + +#include "gro_tcp6.h" + +void * +gro_tcp6_tbl_create(uint16_t socket_id, + uint16_t max_flow_num, + uint16_t max_item_per_flow) +{ + struct gro_tcp6_tbl *tbl; + size_t size; + uint32_t entries_num, i; + + entries_num = max_flow_num * max_item_per_flow; + entries_num = RTE_MIN(entries_num, GRO_TCP6_TBL_MAX_ITEM_NUM); + + if (entries_num == 0) + return NULL; + + tbl = rte_zmalloc_socket(__func__, + sizeof(struct gro_tcp6_tbl), + RTE_CACHE_LINE_SIZE, + socket_id); + if (tbl == NULL) + return NULL; + + size = sizeof(struct gro_tcp6_item) * entries_num; + tbl->items = rte_zmalloc_socket(__func__, + size, + RTE_CACHE_LINE_SIZE, + socket_id); + if (tbl->items == NULL) { + rte_free(tbl); + return NULL; + } + tbl->max_item_num = entries_num; + + size = sizeof(struct gro_tcp6_flow) * entries_num; + tbl->flows = rte_zmalloc_socket(__func__, + size, + RTE_CACHE_LINE_SIZE, + socket_id); + if (tbl->flows == NULL) { + rte_free(tbl->items); + rte_free(tbl); + return NULL; + } + /* INVALID_ARRAY_INDEX indicates an empty flow */ + for (i = 0; i < entries_num; i++) + tbl->flows[i].start_index = INVALID_ARRAY_INDEX; + tbl->max_flow_num = entries_num; + + return tbl; +} + +void +gro_tcp6_tbl_destroy(void *tbl) +{ + struct gro_tcp6_tbl *tcp_tbl = tbl; + + if (tcp_tbl) { + rte_free(tcp_tbl->items); + rte_free(tcp_tbl->flows); + } + rte_free(tcp_tbl); +} + +static inline uint32_t +find_an_empty_item(struct gro_tcp6_tbl *tbl) +{ + uint32_t i; + uint32_t max_item_num = tbl->max_item_num; + + for (i = 0; i < max_item_num; i++) + if (tbl->items[i].firstseg == NULL) + return i; + return INVALID_ARRAY_INDEX; +} + +static inline uint32_t +find_an_empty_flow(struct gro_tcp6_tbl *tbl) +{ + uint32_t i; + uint32_t max_flow_num = tbl->max_flow_num; + + for (i = 0; i < max_flow_num; i++) + if (tbl->flows[i].start_index == INVALID_ARRAY_INDEX) + return i; + return INVALID_ARRAY_INDEX; +} + +static inline uint32_t +insert_new_item(struct gro_tcp6_tbl *tbl, + struct rte_mbuf *pkt, + uint64_t start_time, + uint32_t prev_idx, + uint32_t sent_seq) +{ + uint32_t item_idx; + + item_idx = find_an_empty_item(tbl); + if (item_idx == INVALID_ARRAY_INDEX) + return INVALID_ARRAY_INDEX; + + tbl->items[item_idx].firstseg = pkt; + tbl->items[item_idx].lastseg = rte_pktmbuf_lastseg(pkt); + tbl->items[item_idx].start_time = start_time; + tbl->items[item_idx].next_pkt_idx = INVALID_ARRAY_INDEX; + tbl->items[item_idx].sent_seq = sent_seq; + tbl->items[item_idx].nb_merged = 1; + tbl->item_num++; + + /* if the previous packet exists, chain them together. */ + if (prev_idx != INVALID_ARRAY_INDEX) { + tbl->items[item_idx].next_pkt_idx = + tbl->items[prev_idx].next_pkt_idx; + tbl->items[prev_idx].next_pkt_idx = item_idx; + } + + return item_idx; +} + +static inline uint32_t +delete_item(struct gro_tcp6_tbl *tbl, uint32_t item_idx, + uint32_t prev_item_idx) +{ + uint32_t next_idx = tbl->items[item_idx].next_pkt_idx; + + /* NULL indicates an empty item */ + tbl->items[item_idx].firstseg = NULL; + tbl->item_num--; + if (prev_item_idx != INVALID_ARRAY_INDEX) + tbl->items[prev_item_idx].next_pkt_idx = next_idx; + + return next_idx; +} + +static inline uint32_t +insert_new_flow(struct gro_tcp6_tbl *tbl, + struct tcp6_flow_key *src, + uint32_t item_idx) +{ + struct tcp6_flow_key *dst; + uint32_t flow_idx; + + flow_idx = find_an_empty_flow(tbl); + if (unlikely(flow_idx == INVALID_ARRAY_INDEX)) + return INVALID_ARRAY_INDEX; + + dst = &(tbl->flows[flow_idx].key); + + rte_ether_addr_copy(&(src->eth_saddr), &(dst->eth_saddr)); + rte_ether_addr_copy(&(src->eth_daddr), &(dst->eth_daddr)); + gro_ipv6_addr_copy(src->ip_saddr, dst->ip_saddr); + gro_ipv6_addr_copy(src->ip_daddr, dst->ip_daddr); + dst->recv_ack = src->recv_ack; + dst->src_port = src->src_port; + dst->dst_port = src->dst_port; + + tbl->flows[flow_idx].start_index = item_idx; + tbl->flow_num++; + + return flow_idx; +} + +/* + * update the packet length for the flushed packet. + */ +static inline void +update_header(struct gro_tcp6_item *item) +{ + struct rte_ipv6_hdr *ipv6_hdr; + struct rte_mbuf *pkt = item->firstseg; + + ipv6_hdr = (struct rte_ipv6_hdr *)(rte_pktmbuf_mtod(pkt, char *) + + pkt->l2_len); + ipv6_hdr->payload_len = rte_cpu_to_be_16(pkt->pkt_len - + pkt->l2_len - pkt->l3_len); +} + +int32_t +gro_tcp6_reassemble(struct rte_mbuf *pkt, + struct gro_tcp6_tbl *tbl, + uint64_t start_time) +{ + struct rte_ether_hdr *eth_hdr; + struct rte_ipv6_hdr *ipv6_hdr; + struct rte_tcp_hdr *tcp_hdr; + uint32_t sent_seq; + int32_t tcp_dl; + uint16_t hdr_len; + + struct tcp6_flow_key key; + uint32_t cur_idx, prev_idx, item_idx; + uint32_t i, max_flow_num, remaining_flow_num; + int cmp; + uint8_t find; + + /* + * Don't process the packet whose TCP header length is greater + * than 60 bytes or less than 20 bytes. + */ + if (unlikely(INVALID_TCP_HDRLEN(pkt->l4_len))) + return -1; + + eth_hdr = rte_pktmbuf_mtod(pkt, struct rte_ether_hdr *); + ipv6_hdr = (struct rte_ipv6_hdr *)((char *)eth_hdr + pkt->l2_len); + tcp_hdr = (struct rte_tcp_hdr *)((char *)ipv6_hdr + pkt->l3_len); + hdr_len = pkt->l2_len + pkt->l3_len + pkt->l4_len; + + /* + * Don't process the packet which has FIN, SYN, RST, PSH, URG, ECE + * or CWR set. + */ + if (tcp_hdr->tcp_flags != RTE_TCP_ACK_FLAG) + return -1; + /* + * Don't process the packet whose payload length is less than or + * equal to 0. + */ + if (pkt->pkt_len <= hdr_len) + return -1; + tcp_dl = pkt->pkt_len - hdr_len; + + /* + * Save IPv6 ID for the packet whose DF bit is 0. For the packet + * whose DF bit is 1, IPv6 ID is ignored. + */ + sent_seq = rte_be_to_cpu_32(tcp_hdr->sent_seq); + + rte_ether_addr_copy(&(eth_hdr->s_addr), &(key.eth_saddr)); + rte_ether_addr_copy(&(eth_hdr->d_addr), &(key.eth_daddr)); + gro_ipv6_addr_copy(ipv6_hdr->src_addr, key.ip_saddr); + gro_ipv6_addr_copy(ipv6_hdr->dst_addr, key.ip_daddr); + key.src_port = tcp_hdr->src_port; + key.dst_port = tcp_hdr->dst_port; + key.recv_ack = tcp_hdr->recv_ack; + + /* Search for a matched flow. */ + max_flow_num = tbl->max_flow_num; + remaining_flow_num = tbl->flow_num; + find = 0; + for (i = 0; i < max_flow_num && remaining_flow_num; i++) { + if (tbl->flows[i].start_index != INVALID_ARRAY_INDEX) { + if (is_same_tcp6_flow(tbl->flows[i].key, key)) { + find = 1; + break; + } + remaining_flow_num--; + } + } + + /* + * Fail to find a matched flow. Insert a new flow and store the + * packet into the flow. + */ + if (find == 0) { + item_idx = insert_new_item(tbl, pkt, start_time, + INVALID_ARRAY_INDEX, sent_seq); + if (item_idx == INVALID_ARRAY_INDEX) + return -1; + if (insert_new_flow(tbl, &key, item_idx) == + INVALID_ARRAY_INDEX) { + /* + * Fail to insert a new flow, so delete the + * stored packet. + */ + delete_item(tbl, item_idx, INVALID_ARRAY_INDEX); + return -1; + } + return 0; + } + + /* + * Check all packets in the flow and try to find a neighbor for + * the input packet. + */ + cur_idx = tbl->flows[i].start_index; + prev_idx = cur_idx; + do { + cmp = check_seq_option_v6(&(tbl->items[cur_idx]), tcp_hdr, + sent_seq, pkt->l4_len, tcp_dl, 0); + if (cmp) { + if (merge_two_tcp6_packets(&(tbl->items[cur_idx]), + pkt, cmp, sent_seq, 0)) + return 1; + /* + * Fail to merge the two packets, as the packet + * length is greater than the max value. Store + * the packet into the flow. + */ + if (insert_new_item(tbl, pkt, start_time, prev_idx, + sent_seq) == + INVALID_ARRAY_INDEX) + return -1; + return 0; + } + prev_idx = cur_idx; + cur_idx = tbl->items[cur_idx].next_pkt_idx; + } while (cur_idx != INVALID_ARRAY_INDEX); + + /* Fail to find a neighbor, so store the packet into the flow. */ + if (insert_new_item(tbl, pkt, start_time, prev_idx, sent_seq + ) == INVALID_ARRAY_INDEX) + return -1; + + return 0; +} + +uint16_t +gro_tcp6_tbl_timeout_flush(struct gro_tcp6_tbl *tbl, + uint64_t flush_timestamp, + struct rte_mbuf **out, + uint16_t nb_out) +{ + uint16_t k = 0; + uint32_t i, j; + uint32_t max_flow_num = tbl->max_flow_num; + + for (i = 0; i < max_flow_num; i++) { + if (unlikely(tbl->flow_num == 0)) + return k; + + j = tbl->flows[i].start_index; + while (j != INVALID_ARRAY_INDEX) { + if (tbl->items[j].start_time <= flush_timestamp) { + out[k++] = tbl->items[j].firstseg; + if (tbl->items[j].nb_merged > 1) + update_header(&(tbl->items[j])); + /* + * Delete the packet and get the next + * packet in the flow. + */ + j = delete_item(tbl, j, INVALID_ARRAY_INDEX); + tbl->flows[i].start_index = j; + if (j == INVALID_ARRAY_INDEX) + tbl->flow_num--; + + if (unlikely(k == nb_out)) + return k; + } else + /* + * The left packets in this flow won't be + * timeout. Go to check other flows. + */ + break; + } + } + return k; +} + +uint32_t +gro_tcp6_tbl_pkt_count(void *tbl) +{ + struct gro_tcp6_tbl *gro_tbl = tbl; + + if (gro_tbl) + return gro_tbl->item_num; + + return 0; +} diff --git a/lib/librte_gro/gro_tcp6.h b/lib/librte_gro/gro_tcp6.h new file mode 100644 index 0000000..a9dec7b --- /dev/null +++ b/lib/librte_gro/gro_tcp6.h @@ -0,0 +1,314 @@ +/* SPDX-License-Identifier: BSD-3-Clause + * Copyright(c) 2021 Inspur Corporation + */ + +#ifndef _GRO_TCP6_H_ +#define _GRO_TCP6_H_ + +#include +#include +#include + +#define INVALID_ARRAY_INDEX 0xffffffffUL +#define GRO_TCP6_TBL_MAX_ITEM_NUM (1024UL * 1024UL) + +/* + * The max length of a IPv6 packet, which includes the length of the L3 + * header, the L4 header and the data payload. + */ +#define MAX_IPV6_PKT_LENGTH UINT16_MAX + +/* The maximum TCP header length */ +#define MAX_TCP_HLEN 60 +#define INVALID_TCP_HDRLEN(len) \ + (((len) < sizeof(struct rte_tcp_hdr)) || ((len) > MAX_TCP_HLEN)) + +/* Header fields representing a TCP/IPv6 flow */ +struct tcp6_flow_key { + struct rte_ether_addr eth_saddr; + struct rte_ether_addr eth_daddr; + uint8_t ip_saddr[16]; /**< IPv6 address of source host. */ + uint8_t ip_daddr[16]; /**< IPv6 address of destination host(s). */ + + uint32_t recv_ack; + uint16_t src_port; + uint16_t dst_port; +}; + +struct gro_tcp6_flow { + struct tcp6_flow_key key; + /* + * The index of the first packet in the flow. + * INVALID_ARRAY_INDEX indicates an empty flow. + */ + uint32_t start_index; +}; + +struct gro_tcp6_item { + /* + * The first MBUF segment of the packet. If the value + * is NULL, it means the item is empty. + */ + struct rte_mbuf *firstseg; + /* The last MBUF segment of the packet */ + struct rte_mbuf *lastseg; + /* + * The time when the first packet is inserted into the table. + * This value won't be updated, even if the packet is merged + * with other packets. + */ + uint64_t start_time; + /* + * next_pkt_idx is used to chain the packets that + * are in the same flow but can't be merged together + * (e.g. caused by packet reordering). + */ + uint32_t next_pkt_idx; + /* TCP sequence number of the packet */ + uint32_t sent_seq; + /* the number of merged packets */ + uint16_t nb_merged; +}; + +/* + * TCP/IPv6 reassembly table structure. + */ +struct gro_tcp6_tbl { + /* item array */ + struct gro_tcp6_item *items; + /* flow array */ + struct gro_tcp6_flow *flows; + /* current item number */ + uint32_t item_num; + /* current flow num */ + uint32_t flow_num; + /* item array size */ + uint32_t max_item_num; + /* flow array size */ + uint32_t max_flow_num; +}; + +/** + * This function creates a TCP/IPv6 reassembly table. + * + * @param socket_id + * Socket index for allocating the TCP/IPv6 reassemble table + * @param max_flow_num + * The maximum number of flows in the TCP/IPv6 GRO table + * @param max_item_per_flow + * The maximum number of packets per flow + * + * @return + * - Return the table pointer on success. + * - Return NULL on failure. + */ +void *gro_tcp6_tbl_create(uint16_t socket_id, + uint16_t max_flow_num, + uint16_t max_item_per_flow); + +/** + * This function destroys a TCP/IPv6 reassembly table. + * + * @param tbl + * Pointer pointing to the TCP/IPv6 reassembly table. + */ +void gro_tcp6_tbl_destroy(void *tbl); + +/** + * This function merges a TCP/IPv6 packet. It does not process the packet, + * which has SYN, FIN, RST, PSH, CWR, ECE or URG set, or does not have + * payload. + * + * This function does not check if the packet has correct checksums and + * does not re-calculate checksums for the merged packet. Additionally, + * it assumes the packets are complete (i.e., MF==0 && frag_off==0), + * when IP fragmentation is possible (i.e., DF==0). It returns the + * packet, if the packet has invalid parameters (e.g. SYN bit is set) + * or there is no available space in the table. + * + * @param pkt + * Packet to reassemble + * @param tbl + * Pointer pointing to the TCP/IPv6 reassembly table + * @start_time + * The time when the packet is inserted into the table + * + * @return + * - Return a positive value if the packet is merged. + * - Return zero if the packet isn't merged but stored in the table. + * - Return a negative value for invalid parameters or no available + * space in the table. + */ +int32_t gro_tcp6_reassemble(struct rte_mbuf *pkt, + struct gro_tcp6_tbl *tbl, + uint64_t start_time); + +/** + * This function flushes timeout packets in a TCP/IPv6 reassembly table, + * and without updating checksums. + * + * @param tbl + * TCP/IPv6 reassembly table pointer + * @param flush_timestamp + * Flush packets which are inserted into the table before or at the + * flush_timestamp. + * @param out + * Pointer array used to keep flushed packets + * @param nb_out + * The element number in 'out'. It also determines the maximum number of + * packets that can be flushed finally. + * + * @return + * The number of flushed packets + */ +uint16_t gro_tcp6_tbl_timeout_flush(struct gro_tcp6_tbl *tbl, + uint64_t flush_timestamp, + struct rte_mbuf **out, + uint16_t nb_out); + +/** + * This function returns the number of the packets in a TCP/IPv6 + * reassembly table. + * + * @param tbl + * TCP/IPv6 reassembly table pointer + * + * @return + * The number of packets in the table + */ +uint32_t gro_tcp6_tbl_pkt_count(void *tbl); + +/* + * Check if two IPv6 address is same. + */ +static inline int rte_is_same_ipv6_addr(const uint8_t *ipv6_a1, + const uint8_t *ipv6_a2) +{ + const uint64_t *w1 = (const uint64_t *)ipv6_a1; + const uint64_t *w2 = (const uint64_t *)ipv6_a2; + + return ((w1[0] ^ w2[0]) | (w1[1] ^ w2[1])) == 0; +} + +/* Copy IPv6 addr */ +static inline void gro_ipv6_addr_copy(const uint8_t *ipv6_from, + uint8_t *ipv6_to) +{ + const uint64_t *from_words = (const uint64_t *)ipv6_from; + uint64_t *to_words = (uint64_t *)ipv6_to; + + to_words[0] = from_words[0]; + to_words[1] = from_words[1]; +} + +/* + * Check if two TCP/IPv6 packets belong to the same flow. + */ +static inline int +is_same_tcp6_flow(struct tcp6_flow_key k1, struct tcp6_flow_key k2) +{ + return (rte_is_same_ether_addr(&k1.eth_saddr, &k2.eth_saddr) && + rte_is_same_ether_addr(&k1.eth_daddr, &k2.eth_daddr) && + rte_is_same_ipv6_addr(k1.ip_saddr, k2.ip_saddr) && + rte_is_same_ipv6_addr(k1.ip_daddr, k2.ip_daddr) && + (k1.recv_ack == k2.recv_ack) && + (k1.src_port == k2.src_port) && + (k1.dst_port == k2.dst_port)); +} + +/* + * Merge two TCP/IPv6, packets without updating checksums. + * If cmp is larger than 0, append the new packet to the + * original packet. Otherwise, pre-pend the new packet to + * the original packet. + */ +static inline int +merge_two_tcp6_packets(struct gro_tcp6_item *item, + struct rte_mbuf *pkt, + int cmp, + uint32_t sent_seq, + uint16_t l2_offset) +{ + struct rte_mbuf *pkt_head, *pkt_tail, *lastseg; + uint16_t hdr_len, l2_len; + + if (cmp > 0) { + pkt_head = item->firstseg; + pkt_tail = pkt; + } else { + pkt_head = pkt; + pkt_tail = item->firstseg; + } + + /* check if the IPv6 packet length is greater than the max value */ + hdr_len = l2_offset + pkt_head->l2_len + pkt_head->l3_len + + pkt_head->l4_len; + l2_len = l2_offset > 0 ? pkt_head->outer_l2_len : pkt_head->l2_len; + if (unlikely(pkt_head->pkt_len - l2_len + pkt_tail->pkt_len - + hdr_len > MAX_IPV6_PKT_LENGTH)) + return 0; + + /* remove the packet header for the tail packet */ + rte_pktmbuf_adj(pkt_tail, hdr_len); + + /* chain two packets together */ + if (cmp > 0) { + item->lastseg->next = pkt; + item->lastseg = rte_pktmbuf_lastseg(pkt); + } else { + lastseg = rte_pktmbuf_lastseg(pkt); + lastseg->next = item->firstseg; + item->firstseg = pkt; + /* update sent_seq to the smaller value */ + item->sent_seq = sent_seq; + } + item->nb_merged++; + + /* update MBUF metadata for the merged packet */ + pkt_head->nb_segs += pkt_tail->nb_segs; + pkt_head->pkt_len += pkt_tail->pkt_len; + + return 1; +} + +/* + * Check if two TCP/IPv6 packets are neighbors. + */ +static inline int +check_seq_option_v6(struct gro_tcp6_item *item, + struct rte_tcp_hdr *tcph, + uint32_t sent_seq, + uint16_t tcp_hl, + uint16_t tcp_dl, + uint16_t l2_offset) +{ + struct rte_mbuf *pkt_orig = item->firstseg; + struct rte_ipv6_hdr *iph_orig; + struct rte_tcp_hdr *tcph_orig; + uint16_t len, tcp_hl_orig; + + iph_orig = (struct rte_ipv6_hdr *)(rte_pktmbuf_mtod(pkt_orig, char *) + + l2_offset + pkt_orig->l2_len); + tcph_orig = (struct rte_tcp_hdr *)((char *)iph_orig + pkt_orig->l3_len); + tcp_hl_orig = pkt_orig->l4_len; + + /* Check if TCP option fields equal */ + len = RTE_MAX(tcp_hl, tcp_hl_orig) - sizeof(struct rte_tcp_hdr); + if ((tcp_hl != tcp_hl_orig) || ((len > 0) && + (memcmp(tcph + 1, tcph_orig + 1, + len) != 0))) + return 0; + + /* check if the two packets are neighbors */ + len = pkt_orig->pkt_len - l2_offset - pkt_orig->l2_len - + pkt_orig->l3_len - tcp_hl_orig; + if (sent_seq == item->sent_seq + len) + /* append the new packet */ + return 1; + else if (sent_seq + tcp_dl == item->sent_seq) + /* pre-pend the new packet */ + return -1; + + return 0; +} +#endif diff --git a/lib/librte_gro/meson.build b/lib/librte_gro/meson.build index ea8b45c..6143dd0 100644 --- a/lib/librte_gro/meson.build +++ b/lib/librte_gro/meson.build @@ -1,6 +1,6 @@ # SPDX-License-Identifier: BSD-3-Clause # Copyright(c) 2017 Intel Corporation -sources = files('rte_gro.c', 'gro_tcp4.c', 'gro_udp4.c', 'gro_vxlan_tcp4.c', 'gro_vxlan_udp4.c') +sources = files('rte_gro.c', 'gro_tcp4.c', 'gro_udp4.c', 'gro_vxlan_tcp4.c', 'gro_vxlan_udp4.c', 'gro_tcp6.c') headers = files('rte_gro.h') deps += ['ethdev'] diff --git a/lib/librte_gro/rte_gro.c b/lib/librte_gro/rte_gro.c index 8ca4da6..86958d5 100644 --- a/lib/librte_gro/rte_gro.c +++ b/lib/librte_gro/rte_gro.c @@ -9,6 +9,7 @@ #include "rte_gro.h" #include "gro_tcp4.h" +#include "gro_tcp6.h" #include "gro_udp4.h" #include "gro_vxlan_tcp4.h" #include "gro_vxlan_udp4.h" @@ -21,15 +22,16 @@ static gro_tbl_create_fn tbl_create_fn[RTE_GRO_TYPE_MAX_NUM] = { gro_tcp4_tbl_create, gro_vxlan_tcp4_tbl_create, - gro_udp4_tbl_create, gro_vxlan_udp4_tbl_create, NULL}; + gro_udp4_tbl_create, gro_vxlan_udp4_tbl_create, + gro_tcp6_tbl_create, NULL}; static gro_tbl_destroy_fn tbl_destroy_fn[RTE_GRO_TYPE_MAX_NUM] = { gro_tcp4_tbl_destroy, gro_vxlan_tcp4_tbl_destroy, gro_udp4_tbl_destroy, gro_vxlan_udp4_tbl_destroy, - NULL}; + gro_tcp6_tbl_destroy, NULL}; static gro_tbl_pkt_count_fn tbl_pkt_count_fn[RTE_GRO_TYPE_MAX_NUM] = { gro_tcp4_tbl_pkt_count, gro_vxlan_tcp4_tbl_pkt_count, gro_udp4_tbl_pkt_count, gro_vxlan_udp4_tbl_pkt_count, - NULL}; + gro_tcp6_tbl_pkt_count, NULL}; #define IS_IPV4_TCP_PKT(ptype) (RTE_ETH_IS_IPV4_HDR(ptype) && \ ((ptype & RTE_PTYPE_L4_TCP) == RTE_PTYPE_L4_TCP) && \ @@ -65,6 +67,10 @@ ((ptype & RTE_PTYPE_INNER_L3_MASK) == \ RTE_PTYPE_INNER_L3_IPV4_EXT_UNKNOWN))) +#define IS_IPV6_TCP_PKT(ptype) (RTE_ETH_IS_IPV6_HDR(ptype) && \ + ((ptype & RTE_PTYPE_L4_TCP) == RTE_PTYPE_L4_TCP) && \ + (RTE_ETH_IS_TUNNEL_PKT(ptype) == 0)) + /* * GRO context structure. It keeps the table structures, which are * used to merge packets, for different GRO types. Before using @@ -165,17 +171,24 @@ struct gro_ctx { struct gro_vxlan_udp4_item vxlan_udp_items[RTE_GRO_MAX_BURST_ITEM_NUM] = {{{0}} }; + /* allocate a reassembly table for TCP/IPv6 GRO */ + struct gro_tcp6_tbl tcp6_tbl; + struct gro_tcp6_flow tcp6_flows[RTE_GRO_MAX_BURST_ITEM_NUM]; + struct gro_tcp6_item tcp6_items[RTE_GRO_MAX_BURST_ITEM_NUM] = {{0} }; + + struct rte_mbuf *unprocess_pkts[nb_pkts]; uint32_t item_num; int32_t ret; uint16_t i, unprocess_num = 0, nb_after_gro = nb_pkts; uint8_t do_tcp4_gro = 0, do_vxlan_tcp_gro = 0, do_udp4_gro = 0, - do_vxlan_udp_gro = 0; + do_vxlan_udp_gro = 0, do_tcp6_gro = 0; if (unlikely((param->gro_types & (RTE_GRO_IPV4_VXLAN_TCP_IPV4 | RTE_GRO_TCP_IPV4 | RTE_GRO_IPV4_VXLAN_UDP_IPV4 | - RTE_GRO_UDP_IPV4)) == 0)) + RTE_GRO_UDP_IPV4 | + RTE_GRO_TCP_IPV6)) == 0)) return nb_pkts; /* Get the maximum number of packets */ @@ -235,6 +248,20 @@ struct gro_ctx { do_udp4_gro = 1; } + if (param->gro_types & RTE_GRO_TCP_IPV6) { + for (i = 0; i < item_num; i++) + tcp6_flows[i].start_index = INVALID_ARRAY_INDEX; + + tcp6_tbl.flows = tcp6_flows; + tcp6_tbl.items = tcp6_items; + tcp6_tbl.flow_num = 0; + tcp6_tbl.item_num = 0; + tcp6_tbl.max_flow_num = item_num; + tcp6_tbl.max_item_num = item_num; + do_tcp6_gro = 1; + } + + for (i = 0; i < nb_pkts; i++) { /* @@ -275,6 +302,14 @@ struct gro_ctx { nb_after_gro--; else if (ret < 0) unprocess_pkts[unprocess_num++] = pkts[i]; + } else if (IS_IPV6_TCP_PKT(pkts[i]->packet_type) && + do_tcp6_gro) { + ret = gro_tcp6_reassemble(pkts[i], &tcp6_tbl, 0); + if (ret > 0) + /* merge successfully */ + nb_after_gro--; + else if (ret < 0) + unprocess_pkts[unprocess_num++] = pkts[i]; } else unprocess_pkts[unprocess_num++] = pkts[i]; } @@ -303,6 +338,12 @@ struct gro_ctx { i += gro_udp4_tbl_timeout_flush(&udp_tbl, 0, &pkts[i], nb_pkts - i); } + + if (do_tcp6_gro) { + i += gro_tcp6_tbl_timeout_flush(&tcp6_tbl, 0, + &pkts[i], nb_pkts - i); + } + /* Copy unprocessed packets */ if (unprocess_num > 0) { memcpy(&pkts[i], unprocess_pkts, @@ -322,21 +363,24 @@ struct gro_ctx { { struct rte_mbuf *unprocess_pkts[nb_pkts]; struct gro_ctx *gro_ctx = ctx; - void *tcp_tbl, *udp_tbl, *vxlan_tcp_tbl, *vxlan_udp_tbl; + void *tcp_tbl, *udp_tbl, *vxlan_tcp_tbl, *vxlan_udp_tbl, *tcp6_tbl; uint64_t current_time; uint16_t i, unprocess_num = 0; - uint8_t do_tcp4_gro, do_vxlan_tcp_gro, do_udp4_gro, do_vxlan_udp_gro; + uint8_t do_tcp4_gro, do_vxlan_tcp_gro, do_udp4_gro, do_vxlan_udp_gro, + do_tcp6_gro; if (unlikely((gro_ctx->gro_types & (RTE_GRO_IPV4_VXLAN_TCP_IPV4 | RTE_GRO_TCP_IPV4 | RTE_GRO_IPV4_VXLAN_UDP_IPV4 | - RTE_GRO_UDP_IPV4)) == 0)) + RTE_GRO_UDP_IPV4 | + RTE_GRO_TCP_IPV6)) == 0)) return nb_pkts; tcp_tbl = gro_ctx->tbls[RTE_GRO_TCP_IPV4_INDEX]; vxlan_tcp_tbl = gro_ctx->tbls[RTE_GRO_IPV4_VXLAN_TCP_IPV4_INDEX]; udp_tbl = gro_ctx->tbls[RTE_GRO_UDP_IPV4_INDEX]; vxlan_udp_tbl = gro_ctx->tbls[RTE_GRO_IPV4_VXLAN_UDP_IPV4_INDEX]; + tcp6_tbl = gro_ctx->tbls[RTE_GRO_TCP_IPV6_INDEX]; do_tcp4_gro = (gro_ctx->gro_types & RTE_GRO_TCP_IPV4) == RTE_GRO_TCP_IPV4; @@ -346,6 +390,8 @@ struct gro_ctx { RTE_GRO_UDP_IPV4; do_vxlan_udp_gro = (gro_ctx->gro_types & RTE_GRO_IPV4_VXLAN_UDP_IPV4) == RTE_GRO_IPV4_VXLAN_UDP_IPV4; + do_tcp6_gro = (gro_ctx->gro_types & RTE_GRO_TCP_IPV6) == + RTE_GRO_TCP_IPV6; current_time = rte_rdtsc(); @@ -370,6 +416,11 @@ struct gro_ctx { if (gro_udp4_reassemble(pkts[i], udp_tbl, current_time) < 0) unprocess_pkts[unprocess_num++] = pkts[i]; + } else if (IS_IPV6_TCP_PKT(pkts[i]->packet_type) && + do_tcp6_gro) { + if (gro_tcp6_reassemble(pkts[i], tcp6_tbl, + current_time) < 0) + unprocess_pkts[unprocess_num++] = pkts[i]; } else unprocess_pkts[unprocess_num++] = pkts[i]; } @@ -425,6 +476,13 @@ struct gro_ctx { gro_ctx->tbls[RTE_GRO_UDP_IPV4_INDEX], flush_timestamp, &out[num], left_nb_out); + left_nb_out = max_nb_out - num; + } + + if ((gro_types & RTE_GRO_TCP_IPV6) && left_nb_out > 0) { + num += gro_tcp6_tbl_timeout_flush(gro_ctx->tbls[ + RTE_GRO_TCP_IPV6_INDEX], + flush_timestamp, &out[num], left_nb_out); } return num; diff --git a/lib/librte_gro/rte_gro.h b/lib/librte_gro/rte_gro.h index 9f9ed49..39b57e8 100644 --- a/lib/librte_gro/rte_gro.h +++ b/lib/librte_gro/rte_gro.h @@ -38,6 +38,9 @@ #define RTE_GRO_IPV4_VXLAN_UDP_IPV4_INDEX 3 #define RTE_GRO_IPV4_VXLAN_UDP_IPV4 (1ULL << RTE_GRO_IPV4_VXLAN_UDP_IPV4_INDEX) /**< VxLAN UDP/IPv4 GRO flag. */ +#define RTE_GRO_TCP_IPV6_INDEX 4 +#define RTE_GRO_TCP_IPV6 (1ULL << RTE_GRO_TCP_IPV6_INDEX) +/**< TCP/IPv6 GRO flag */ /** * Structure used to create GRO context objects or used to pass -- 1.8.3.1