dpdk/0401-app-testpmd-fix-tunnel-TSO-configuration.patch
Dengdui Huang ab5939fabf sync some bugfix from upstreaming about testpmd and doc
Sync some bugfix from upstreaming about testpmd and doc, modifies
are as follow:
- support set RSS hash algorithm
- ethdev: add new API to get RSS hash algorithm by name
- doc: fix description of RSS features
- doc: fix RSS flow description in hns3 guide
- doc: update features in hns3 guide
- doc: fix hns3 build option about max queue number
- app/testpmd: check port and queue Rx/Tx offloads
- app/testpmd: fix Tx offload command
- app/testpmd: allow offload config for all ports
- app/testpmd: fix tunnel TSO configuration
- app/testpmd: add explicit check for tunnel TSO
- app/testpmd: fix tunnel TSO capability check
- app/testpmd: remove useless check in TSO command

Signed-off-by: Dengdui Huang <huangdengdui@huawei.com>
(cherry picked from commit 7868d4e3ae469277d4b47241e84c77f53e09423b)
2023-12-09 10:35:25 +08:00

72 lines
2.6 KiB
Diff

From 27c78af1f290c342b6fb64c2f57a430628f5062f Mon Sep 17 00:00:00 2001
From: Huisong Li <lihuisong@huawei.com>
Date: Sat, 11 Nov 2023 12:59:43 +0800
Subject: [PATCH 401/410] app/testpmd: fix tunnel TSO configuration
[ upstream commit e43dc93803c4623840472c6109ef05e26286ec2f ]
Currently, there are two conditions to set tunnel TSO, like "parse
tunnel" and "outer IP checksum".
If these conditions are not satisfied, testpmd should not change their
configuration, like tx_offloads on port and per queue, and no need to
request "reconfig device".
Fixes: 597f9fafe13b ("app/testpmd: convert to new Tx offloads API")
Cc: stable@dpdk.org
Signed-off-by: Huisong Li <lihuisong@huawei.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@amd.com>
---
app/test-pmd/cmdline.c | 25 ++++++++++++++-----------
1 file changed, 14 insertions(+), 11 deletions(-)
diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c
index 444c7a79ff..fd897dc80c 100644
--- a/app/test-pmd/cmdline.c
+++ b/app/test-pmd/cmdline.c
@@ -5079,12 +5079,6 @@ cmd_tunnel_tso_set_parsed(void *parsed_result,
res->port_id);
return;
}
- check_tunnel_tso_nic_support(res->port_id, dev_info.tx_offload_capa);
-
- ports[res->port_id].dev_conf.txmode.offloads |=
- (all_tunnel_tso & dev_info.tx_offload_capa);
- printf("TSO segment size for tunneled packets is %d\n",
- ports[res->port_id].tunnel_tso_segsz);
/* Below conditions are needed to make it work:
* (1) tunnel TSO is supported by the NIC;
@@ -5097,14 +5091,23 @@ cmd_tunnel_tso_set_parsed(void *parsed_result,
* is not necessary for IPv6 tunneled pkts because there's no
* checksum in IP header anymore.
*/
-
- if (!ports[res->port_id].parse_tunnel)
+ if (!ports[res->port_id].parse_tunnel) {
fprintf(stderr,
- "Warning: csum parse_tunnel must be set so that tunneled packets are recognized\n");
+ "Error: csum parse_tunnel must be set so that tunneled packets are recognized\n");
+ return;
+ }
if (!(ports[res->port_id].dev_conf.txmode.offloads &
- RTE_ETH_TX_OFFLOAD_OUTER_IPV4_CKSUM))
+ RTE_ETH_TX_OFFLOAD_OUTER_IPV4_CKSUM)) {
fprintf(stderr,
- "Warning: csum set outer-ip must be set to hw if outer L3 is IPv4; not necessary for IPv6\n");
+ "Error: csum set outer-ip must be set to hw if outer L3 is IPv4; not necessary for IPv6\n");
+ return;
+ }
+
+ check_tunnel_tso_nic_support(res->port_id, dev_info.tx_offload_capa);
+ ports[res->port_id].dev_conf.txmode.offloads |=
+ (all_tunnel_tso & dev_info.tx_offload_capa);
+ printf("TSO segment size for tunneled packets is %d\n",
+ ports[res->port_id].tunnel_tso_segsz);
}
cmd_config_queue_tx_offloads(&ports[res->port_id]);
--
2.33.0