patches for DPDK stable branches
 help / color / mirror / Atom feed
From: Alvin Zhang <alvinx.zhang@intel.com>
To: xiaoyun.li@intel.com, konstantin.ananyev@intel.com
Cc: dev@dpdk.org, Alvin Zhang <alvinx.zhang@intel.com>, stable@dpdk.org
Subject: [dpdk-stable] [PATCH v2 2/2] app/testpmd: fix txonly forwording
Date: Fri, 17 Sep 2021 09:39:23 +0800	[thread overview]
Message-ID: <20210917013923.4004-2-alvinx.zhang@intel.com> (raw)
In-Reply-To: <20210917013923.4004-1-alvinx.zhang@intel.com>

When random number of Tx segments is enabled, because the actual
number of segments may be only one, the first segment of the Tx
packets must accommodate a complete being sending Eth/IP/UDP packet.

Besides, if multiple flow is enabled, the forwarding will update
the IP and UDP header, these headers shouldn't cross segments.
This also requires the first Tx segment can accommodate a complete
Eth/IP/UDP packet.

In addition, if time stamp is enabled, the forwarding need more
Tx segment space for time stamp information.

This patch adds checks in beginning of forward engine to make sure
all above conditions are meet.

Bugzilla ID: 797
Fixes: 79bec05b32b7 ("app/testpmd: add ability to split outgoing packets")
Cc: stable@dpdk.org

Signed-off-by: Alvin Zhang <alvinx.zhang@intel.com>
---
 app/test-pmd/txonly.c | 67 ++++++++++++++++++++++++++++++++++++++++++---------
 1 file changed, 55 insertions(+), 12 deletions(-)

diff --git a/app/test-pmd/txonly.c b/app/test-pmd/txonly.c
index 386a4ff..40f4075 100644
--- a/app/test-pmd/txonly.c
+++ b/app/test-pmd/txonly.c
@@ -40,6 +40,13 @@
 
 #include "testpmd.h"
 
+struct tx_timestamp {
+	rte_be32_t signature;
+	rte_be16_t pkt_idx;
+	rte_be16_t queue_idx;
+	rte_be64_t ts;
+};
+
 /* use RFC863 Discard Protocol */
 uint16_t tx_udp_src_port = 9;
 uint16_t tx_udp_dst_port = 9;
@@ -257,12 +264,7 @@
 
 	if (unlikely(timestamp_enable)) {
 		uint64_t skew = RTE_PER_LCORE(timestamp_qskew);
-		struct {
-			rte_be32_t signature;
-			rte_be16_t pkt_idx;
-			rte_be16_t queue_idx;
-			rte_be64_t ts;
-		} timestamp_mark;
+		struct tx_timestamp timestamp_mark;
 
 		if (unlikely(timestamp_init_req !=
 				RTE_PER_LCORE(timestamp_idone))) {
@@ -438,13 +440,23 @@
 static int
 tx_only_begin(portid_t pi)
 {
-	uint16_t pkt_data_len;
+	uint16_t pkt_hdr_len, pkt_data_len;
 	int dynf;
 
-	pkt_data_len = (uint16_t) (tx_pkt_length - (
-					sizeof(struct rte_ether_hdr) +
-					sizeof(struct rte_ipv4_hdr) +
-					sizeof(struct rte_udp_hdr)));
+	pkt_hdr_len = (uint16_t)(sizeof(struct rte_ether_hdr) +
+				 sizeof(struct rte_ipv4_hdr) +
+				 sizeof(struct rte_udp_hdr));
+	pkt_data_len = tx_pkt_length - pkt_hdr_len;
+
+	if ((tx_pkt_split == TX_PKT_SPLIT_RND || txonly_multi_flow) &&
+	    tx_pkt_seg_lengths[0] < pkt_hdr_len) {
+		TESTPMD_LOG(ERR,
+			    "Random segment number or multiple flow enabled,"
+			    " but tx_pkt_seg_lengths[0] %u < %u (needed)\n",
+			    tx_pkt_seg_lengths[0], pkt_hdr_len);
+		return -EINVAL;
+	}
+
 	setup_pkt_udp_ip_headers(&pkt_ip_hdr, &pkt_udp_hdr, pkt_data_len);
 
 	timestamp_enable = false;
@@ -463,8 +475,39 @@
 			   timestamp_mask &&
 			   timestamp_off >= 0 &&
 			   !rte_eth_read_clock(pi, &timestamp_initial[pi]);
-	if (timestamp_enable)
+
+	if (timestamp_enable) {
+		pkt_hdr_len += sizeof(struct timestamp);
+
+		if (tx_pkt_split == TX_PKT_SPLIT_RND) {
+			if (tx_pkt_seg_lengths[0] < pkt_hdr_len) {
+				TESTPMD_LOG(ERR,
+					    "Time stamp and random segment number enabled,"
+					    " but tx_pkt_seg_lengths[0] %u < %u (needed)\n",
+					    tx_pkt_seg_lengths[0], pkt_hdr_len);
+				return -EINVAL;
+			}
+		} else {
+			uint16_t total = 0;
+			uint8_t i;
+
+			for (i = 0; i < tx_pkt_nb_segs; i++) {
+				total += tx_pkt_seg_lengths[i];
+				if (total >= pkt_hdr_len)
+					break;
+			}
+
+			if (total < pkt_hdr_len) {
+				TESTPMD_LOG(ERR,
+					    "No enough Tx segment space for time stamp info."
+					    " total %u < %u (needed)\n",
+					    total, pkt_hdr_len);
+				return -EINVAL;
+			}
+		}
 		timestamp_init_req++;
+	}
+
 	/* Make sure all settings are visible on forwarding cores.*/
 	rte_wmb();
 	return 0;
-- 
1.8.3.1


  parent reply	other threads:[~2021-09-17  1:39 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-09-02  8:20 [dpdk-stable] [PATCH] app/testpmd: fix random number of Tx segments Alvin Zhang
2021-09-06  8:58 ` Li, Xiaoyun
2021-09-06 10:03   ` Zhang, AlvinX
2021-09-06 10:54     ` Li, Xiaoyun
2021-09-07  2:25       ` Zhang, AlvinX
2021-09-07  8:05         ` Li, Xiaoyun
     [not found] ` <20210917013923.4004-1-alvinx.zhang@intel.com>
2021-09-17  1:39   ` Alvin Zhang [this message]
     [not found]   ` <20210918030659.12448-1-alvinx.zhang@intel.com>
2021-09-18  3:06     ` [dpdk-stable] [PATCH v3 2/2] app/testpmd: fix txonly forwording Alvin Zhang
2021-09-18  8:20       ` Li, Xiaoyun
2021-09-22  2:49     ` [dpdk-stable] [PATCH v4 1/2] app/testpmd: update forward engine beginning Alvin Zhang
2021-09-22  2:49       ` [dpdk-stable] [PATCH v4 2/2] app/testpmd: fix txonly forwording Alvin Zhang
2021-09-22  5:58         ` Li, Xiaoyun
2021-09-22  5:59       ` [dpdk-stable] [PATCH v4 1/2] app/testpmd: update forward engine beginning Li, Xiaoyun
2021-09-23  1:49       ` [dpdk-stable] [PATCH v5 " Alvin Zhang
2021-09-23  1:49         ` [dpdk-stable] [PATCH v5 2/2] app/testpmd: fix txonly forwording Alvin Zhang
2021-09-23  4:25           ` Ivan Malov
2021-09-23  5:11             ` Zhang, AlvinX
2021-09-23  8:01         ` [dpdk-stable] [PATCH v6 1/2] app/testpmd: update forward engine beginning Alvin Zhang
2021-09-23  8:01           ` [dpdk-stable] [PATCH v6 2/2] app/testpmd: fix txonly forwarding Alvin Zhang
2021-10-08 17:01           ` [dpdk-stable] [PATCH v6 1/2] app/testpmd: update forward engine beginning Ferruh Yigit

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20210917013923.4004-2-alvinx.zhang@intel.com \
    --to=alvinx.zhang@intel.com \
    --cc=dev@dpdk.org \
    --cc=konstantin.ananyev@intel.com \
    --cc=stable@dpdk.org \
    --cc=xiaoyun.li@intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).