Netdev Archive on lore.kernel.org
help / color / mirror / Atom feed
From: Saeed Mahameed <saeed@kernel.org>
To: "David S. Miller" <davem@davemloft.net>,
Jakub Kicinski <kuba@kernel.org>
Cc: netdev@vger.kernel.org, Yevgeny Kliteynik <kliteyn@nvidia.com>,
Alex Vesker <valex@mellanox.com>, Alex Vesker <valex@nvidia.com>,
Saeed Mahameed <saeedm@nvidia.com>
Subject: [net-next 12/17] net/mlx5: DR, Skip source port matching on FDB RX domain
Date: Thu, 26 Aug 2021 17:57:57 -0700 [thread overview]
Message-ID: <20210827005802.236119-13-saeed@kernel.org> (raw)
In-Reply-To: <20210827005802.236119-1-saeed@kernel.org>
From: Yevgeny Kliteynik <kliteyn@nvidia.com>
The FDB RX pipe is connected to the wire and the source port for all
incoming packets equals to wire, single uplink port per PF, this means
there is no point of matching on the source port in such case.
Once we recognize such case, we will optimize the RX steering rule.
Note that in such case we clean both source_eswitch_owner_vhca_id and
source_port.
Signed-off-by: Alex Vesker <valex@mellanox.com>
Signed-off-by: Yevgeny Kliteynik <kliteyn@nvidia.com>
Reviewed-by: Alex Vesker <valex@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
.../mellanox/mlx5/core/steering/dr_matcher.c | 14 +++++++++++++-
1 file changed, 13 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_matcher.c b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_matcher.c
index 6f6191d1d5a6..f0d9f941acfd 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_matcher.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_matcher.c
@@ -396,6 +396,7 @@ static int dr_matcher_set_ste_builders(struct mlx5dr_matcher *matcher,
struct mlx5dr_domain *dmn = matcher->tbl->dmn;
struct mlx5dr_ste_ctx *ste_ctx = dmn->ste_ctx;
struct mlx5dr_match_param mask = {};
+ bool allow_empty_match = false;
struct mlx5dr_ste_build *sb;
bool inner, rx;
int idx = 0;
@@ -428,6 +429,16 @@ static int dr_matcher_set_ste_builders(struct mlx5dr_matcher *matcher,
if (ret)
return ret;
+ /* Optimize RX pipe by reducing source port match, since
+ * the FDB RX part is connected only to the wire.
+ */
+ if (dmn->type == MLX5DR_DOMAIN_TYPE_FDB &&
+ rx && mask.misc.source_port) {
+ mask.misc.source_port = 0;
+ mask.misc.source_eswitch_owner_vhca_id = 0;
+ allow_empty_match = true;
+ }
+
/* Outer */
if (matcher->match_criteria & (DR_MATCHER_CRITERIA_OUTER |
DR_MATCHER_CRITERIA_MISC |
@@ -619,7 +630,8 @@ static int dr_matcher_set_ste_builders(struct mlx5dr_matcher *matcher,
}
/* Empty matcher, takes all */
- if (matcher->match_criteria == DR_MATCHER_CRITERIA_EMPTY)
+ if ((!idx && allow_empty_match) ||
+ matcher->match_criteria == DR_MATCHER_CRITERIA_EMPTY)
mlx5dr_ste_build_empty_always_hit(&sb[idx++], rx);
if (idx == 0) {
--
2.31.1
next prev parent reply other threads:[~2021-08-27 0:59 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-08-27 0:57 [pull request][net-next 00/17] mlx5 updates 2021-08-26 Saeed Mahameed
2021-08-27 0:57 ` [net-next 01/17] net/mlx5: DR, Added support for REMOVE_HEADER packet reformat Saeed Mahameed
2021-08-27 9:10 ` patchwork-bot+netdevbpf
2021-08-27 0:57 ` [net-next 02/17] net/mlx5: DR, Split modify VLAN state to separate pop/push states Saeed Mahameed
2021-08-27 0:57 ` [net-next 03/17] net/mlx5: DR, Enable VLAN pop on TX and VLAN push on RX Saeed Mahameed
2021-08-27 0:57 ` [net-next 04/17] net/mlx5: DR, Enable QP retransmission Saeed Mahameed
2021-08-27 0:57 ` [net-next 05/17] net/mlx5: DR, Improve error flow in actions_build_ste_arr Saeed Mahameed
2021-08-27 0:57 ` [net-next 06/17] net/mlx5: DR, Warn and ignore SW steering rule insertion on QP err Saeed Mahameed
2021-08-27 0:57 ` [net-next 07/17] net/mlx5: DR, Reduce print level for FT chaining level check Saeed Mahameed
2021-08-27 0:57 ` [net-next 08/17] net/mlx5: DR, Support IPv6 matching on flow label for STEv0 Saeed Mahameed
2021-08-27 0:57 ` [net-next 09/17] net/mlx5: DR, replace uintN_t with kernel-style types Saeed Mahameed
2021-08-27 0:57 ` [net-next 10/17] net/mlx5: DR, Use FW API when updating FW-owned flow table Saeed Mahameed
2021-08-27 0:57 ` [net-next 11/17] net/mlx5: DR, Add ignore_flow_level support for multi-dest flow tables Saeed Mahameed
2021-08-27 0:57 ` Saeed Mahameed [this message]
2021-08-27 0:57 ` [net-next 13/17] net/mlx5: DR, Merge DR_STE_SIZE enums Saeed Mahameed
2021-08-27 0:57 ` [net-next 14/17] net/mlx5: DR, Remove HW specific STE type from nic domain Saeed Mahameed
2021-08-27 0:58 ` [net-next 15/17] net/mlx5: DR, Remove rehash ctrl struct from dr_htbl Saeed Mahameed
2021-08-27 0:58 ` [net-next 16/17] net/mlx5: DR, Improve rule tracking memory consumption Saeed Mahameed
2021-08-27 0:58 ` [net-next 17/17] net/mlx5: DR, Add support for update FTE Saeed Mahameed
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=20210827005802.236119-13-saeed@kernel.org \
--to=saeed@kernel.org \
--cc=davem@davemloft.net \
--cc=kliteyn@nvidia.com \
--cc=kuba@kernel.org \
--cc=netdev@vger.kernel.org \
--cc=saeedm@nvidia.com \
--cc=valex@mellanox.com \
--cc=valex@nvidia.com \
--subject='Re: [net-next 12/17] net/mlx5: DR, Skip source port matching on FDB RX domain' \
/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
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).