From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mails.dpdk.org (mails.dpdk.org [217.70.189.124]) by inbox.dpdk.org (Postfix) with ESMTP id 7EE13A0C43; Wed, 20 Oct 2021 12:18:32 +0200 (CEST) Received: from [217.70.189.124] (localhost [127.0.0.1]) by mails.dpdk.org (Postfix) with ESMTP id F1A1540687; Wed, 20 Oct 2021 12:18:31 +0200 (CEST) Received: from mga17.intel.com (mga17.intel.com [192.55.52.151]) by mails.dpdk.org (Postfix) with ESMTP id B756340142; Wed, 20 Oct 2021 12:18:29 +0200 (CEST) X-IronPort-AV: E=McAfee;i="6200,9189,10142"; a="209541826" X-IronPort-AV: E=Sophos;i="5.87,166,1631602800"; d="scan'208";a="209541826" Received: from orsmga006.jf.intel.com ([10.7.209.51]) by fmsmga107.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 20 Oct 2021 03:18:28 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.87,166,1631602800"; d="scan'208";a="444307669" Received: from dpdk-xiaoyun3.sh.intel.com ([10.67.119.214]) by orsmga006.jf.intel.com with ESMTP; 20 Oct 2021 03:18:26 -0700 From: Xiaoyun Li To: konstantin.ananyev@intel.com, stephen@networkplumber.org, ferruh.yigit@intel.com Cc: dev@dpdk.org, Xiaoyun Li , stable@dpdk.org Date: Wed, 20 Oct 2021 18:10:58 +0800 Message-Id: <20211020101058.202994-1-xiaoyun.li@intel.com> X-Mailer: git-send-email 2.25.1 MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Subject: [dpdk-dev] [PATCH v3] app/testpmd: fix l4 sw csum over multi segments X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.29 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" In csum forwarding mode, software UDP/TCP csum calculation only takes the first segment into account while using the whole packet length so the calculation will read invalid memory region with multi-segments packets and will get wrong value. This patch fixes this issue. Fixes: af75078fece3 ("first public release") Cc: stable@dpdk.org Signed-off-by: Xiaoyun Li --- v3: * Use rte_raw_cksum() for multi-segs case instead of copying the whole * packet. v2: * Use static stack memory instead of dynamic allocating in datapath --- app/test-pmd/csumonly.c | 68 ++++++++++++++++++++++++++++++++--------- 1 file changed, 53 insertions(+), 15 deletions(-) diff --git a/app/test-pmd/csumonly.c b/app/test-pmd/csumonly.c index 090797318a..f3e60eb3c3 100644 --- a/app/test-pmd/csumonly.c +++ b/app/test-pmd/csumonly.c @@ -91,12 +91,41 @@ struct simple_gre_hdr { } __rte_packed; static uint16_t -get_udptcp_checksum(void *l3_hdr, void *l4_hdr, uint16_t ethertype) +get_udptcp_checksum(void *l3_hdr, struct rte_mbuf *m, uint16_t l4_off, + uint16_t ethertype) { + uint16_t off = l4_off; + uint32_t cksum = 0; + char *buf; + + while (m != NULL) { + buf = rte_pktmbuf_mtod_offset(m, char *, off); + cksum += rte_raw_cksum(buf, m->data_len - off); + off = 0; + m = m->next; + } if (ethertype == _htons(RTE_ETHER_TYPE_IPV4)) - return rte_ipv4_udptcp_cksum(l3_hdr, l4_hdr); + cksum += rte_ipv4_phdr_cksum(l3_hdr, 0); else /* assume ethertype == RTE_ETHER_TYPE_IPV6 */ - return rte_ipv6_udptcp_cksum(l3_hdr, l4_hdr); + cksum += rte_ipv6_phdr_cksum(l3_hdr, 0); + + cksum = ((cksum & 0xffff0000) >> 16) + (cksum & 0xffff); + cksum = (~cksum) & 0xffff; + + /* + * Per RFC 768:If the computed checksum is zero for UDP, + * it is transmitted as all ones + * (the equivalent in one's complement arithmetic). + */ + if (cksum == 0 && ethertype == _htons(RTE_ETHER_TYPE_IPV4) && + ((struct rte_ipv4_hdr *)l3_hdr)->next_proto_id == IPPROTO_UDP) + cksum = 0xffff; + + if (cksum == 0 && ethertype == _htons(RTE_ETHER_TYPE_IPV6) && + ((struct rte_ipv6_hdr *)l3_hdr)->proto == IPPROTO_UDP) + cksum = 0xffff; + + return (uint16_t)cksum; } /* Parse an IPv4 header to fill l3_len, l4_len, and l4_proto */ @@ -455,7 +484,7 @@ parse_encap_ip(void *encap_ip, struct testpmd_offload_info *info) * depending on the testpmd command line configuration */ static uint64_t process_inner_cksums(void *l3_hdr, const struct testpmd_offload_info *info, - uint64_t tx_offloads) + uint64_t tx_offloads, struct rte_mbuf *m) { struct rte_ipv4_hdr *ipv4_hdr = l3_hdr; struct rte_udp_hdr *udp_hdr; @@ -463,6 +492,7 @@ process_inner_cksums(void *l3_hdr, const struct testpmd_offload_info *info, struct rte_sctp_hdr *sctp_hdr; uint64_t ol_flags = 0; uint32_t max_pkt_len, tso_segsz = 0; + uint16_t l4_off; /* ensure packet is large enough to require tso */ if (!info->is_tunnel) { @@ -505,9 +535,15 @@ process_inner_cksums(void *l3_hdr, const struct testpmd_offload_info *info, if (tx_offloads & DEV_TX_OFFLOAD_UDP_CKSUM) { ol_flags |= PKT_TX_UDP_CKSUM; } else { + if (info->is_tunnel) + l4_off = info->l2_len + + info->outer_l3_len + + info->l2_len + info->l3_len; + else + l4_off = info->l2_len + info->l3_len; udp_hdr->dgram_cksum = 0; udp_hdr->dgram_cksum = - get_udptcp_checksum(l3_hdr, udp_hdr, + get_udptcp_checksum(l3_hdr, m, l4_off, info->ethertype); } } @@ -520,9 +556,14 @@ process_inner_cksums(void *l3_hdr, const struct testpmd_offload_info *info, else if (tx_offloads & DEV_TX_OFFLOAD_TCP_CKSUM) { ol_flags |= PKT_TX_TCP_CKSUM; } else { + if (info->is_tunnel) + l4_off = info->l2_len + info->outer_l3_len + + info->l2_len + info->l3_len; + else + l4_off = info->l2_len + info->l3_len; tcp_hdr->cksum = 0; tcp_hdr->cksum = - get_udptcp_checksum(l3_hdr, tcp_hdr, + get_udptcp_checksum(l3_hdr, m, l4_off, info->ethertype); } if (info->gso_enable) @@ -548,7 +589,7 @@ process_inner_cksums(void *l3_hdr, const struct testpmd_offload_info *info, /* Calculate the checksum of outer header */ static uint64_t process_outer_cksums(void *outer_l3_hdr, struct testpmd_offload_info *info, - uint64_t tx_offloads, int tso_enabled) + uint64_t tx_offloads, int tso_enabled, struct rte_mbuf *m) { struct rte_ipv4_hdr *ipv4_hdr = outer_l3_hdr; struct rte_ipv6_hdr *ipv6_hdr = outer_l3_hdr; @@ -602,12 +643,9 @@ process_outer_cksums(void *outer_l3_hdr, struct testpmd_offload_info *info, /* do not recalculate udp cksum if it was 0 */ if (udp_hdr->dgram_cksum != 0) { udp_hdr->dgram_cksum = 0; - if (info->outer_ethertype == _htons(RTE_ETHER_TYPE_IPV4)) - udp_hdr->dgram_cksum = - rte_ipv4_udptcp_cksum(ipv4_hdr, udp_hdr); - else - udp_hdr->dgram_cksum = - rte_ipv6_udptcp_cksum(ipv6_hdr, udp_hdr); + udp_hdr->dgram_cksum = get_udptcp_checksum(outer_l3_hdr, + m, info->l2_len + info->outer_l3_len, + info->outer_ethertype); } return ol_flags; @@ -942,7 +980,7 @@ pkt_burst_checksum_forward(struct fwd_stream *fs) /* process checksums of inner headers first */ tx_ol_flags |= process_inner_cksums(l3_hdr, &info, - tx_offloads); + tx_offloads, m); /* Then process outer headers if any. Note that the software * checksum will be wrong if one of the inner checksums is @@ -950,7 +988,7 @@ pkt_burst_checksum_forward(struct fwd_stream *fs) if (info.is_tunnel == 1) { tx_ol_flags |= process_outer_cksums(outer_l3_hdr, &info, tx_offloads, - !!(tx_ol_flags & PKT_TX_TCP_SEG)); + !!(tx_ol_flags & PKT_TX_TCP_SEG), m); } /* step 3: fill the mbuf meta data (flags and header lengths) */ -- 2.25.1