netdev.vger.kernel.org archive mirror
 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@nvidia.com>,
	Saeed Mahameed <saeedm@nvidia.com>
Subject: [net-next  15/16] net/mlx5: DR, Ignore modify TTL if device doesn't support it
Date: Tue, 28 Dec 2021 22:25:01 -0800	[thread overview]
Message-ID: <20211229062502.24111-16-saeed@kernel.org> (raw)
In-Reply-To: <20211229062502.24111-1-saeed@kernel.org>

From: Yevgeny Kliteynik <kliteyn@nvidia.com>

When modifying TTL, packet's csum has to be recalculated.
Due to HW issue in ConnectX-5, csum recalculation for modify TTL
is supported through a work-around that is specifically enabled
by configuration.
If the work-around isn't enabled, ignore the modify TTL action
rather than adding an unsupported action.

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_action.c   | 21 ++++++++++++++++---
 include/linux/mlx5/mlx5_ifc.h                 |  2 +-
 2 files changed, 19 insertions(+), 4 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_action.c b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_action.c
index f0faf04536d3..c61a5e83c78c 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_action.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_action.c
@@ -1560,6 +1560,12 @@ dr_action_modify_check_is_ttl_modify(const void *sw_action)
 	return sw_field == MLX5_ACTION_IN_FIELD_OUT_IP_TTL;
 }
 
+static bool dr_action_modify_ttl_ignore(struct mlx5dr_domain *dmn)
+{
+	return !mlx5dr_ste_supp_ttl_cs_recalc(&dmn->info.caps) &&
+	       !MLX5_CAP_ESW_FLOWTABLE(dmn->mdev, fdb_ipv4_ttl_modify);
+}
+
 static int dr_actions_convert_modify_header(struct mlx5dr_action *action,
 					    u32 max_hw_actions,
 					    u32 num_sw_actions,
@@ -1591,8 +1597,13 @@ static int dr_actions_convert_modify_header(struct mlx5dr_action *action,
 		if (ret)
 			return ret;
 
-		if (!(*modify_ttl))
-			*modify_ttl = dr_action_modify_check_is_ttl_modify(sw_action);
+		if (!(*modify_ttl) &&
+		    dr_action_modify_check_is_ttl_modify(sw_action)) {
+			if (dr_action_modify_ttl_ignore(dmn))
+				continue;
+
+			*modify_ttl = true;
+		}
 
 		/* Convert SW action to HW action */
 		ret = dr_action_modify_sw_to_hw(dmn,
@@ -1631,7 +1642,7 @@ static int dr_actions_convert_modify_header(struct mlx5dr_action *action,
 			 * modify actions doesn't exceeds the limit
 			 */
 			hw_idx++;
-			if ((num_sw_actions + hw_idx - i) >= max_hw_actions) {
+			if (hw_idx >= max_hw_actions) {
 				mlx5dr_dbg(dmn, "Modify header action number exceeds HW limit\n");
 				return -EINVAL;
 			}
@@ -1642,6 +1653,10 @@ static int dr_actions_convert_modify_header(struct mlx5dr_action *action,
 		hw_idx++;
 	}
 
+	/* if the resulting HW actions list is empty, add NOP action */
+	if (!hw_idx)
+		hw_idx++;
+
 	*num_hw_actions = hw_idx;
 
 	return 0;
diff --git a/include/linux/mlx5/mlx5_ifc.h b/include/linux/mlx5/mlx5_ifc.h
index deaa0f71213f..598ac3bcc901 100644
--- a/include/linux/mlx5/mlx5_ifc.h
+++ b/include/linux/mlx5/mlx5_ifc.h
@@ -833,7 +833,7 @@ struct mlx5_ifc_flow_table_eswitch_cap_bits {
 	u8      fdb_to_vport_reg_c_id[0x8];
 	u8      reserved_at_8[0xd];
 	u8      fdb_modify_header_fwd_to_table[0x1];
-	u8      reserved_at_16[0x1];
+	u8      fdb_ipv4_ttl_modify[0x1];
 	u8      flow_source[0x1];
 	u8      reserved_at_18[0x2];
 	u8      multi_fdb_encap[0x1];
-- 
2.33.1


  parent reply	other threads:[~2021-12-29  6:25 UTC|newest]

Thread overview: 21+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-12-29  6:24 [pull request][net-next 00/16] mlx5 updates 2021-12-28 Saeed Mahameed
2021-12-29  6:24 ` [net-next 01/16] net/mlx5: DR, Fix error flow in creating matcher Saeed Mahameed
2021-12-29  6:24 ` [net-next 02/16] net/mlx5: DR, Fix lower case macro prefix "mlx5_" to "MLX5_" Saeed Mahameed
2021-12-29  6:24 ` [net-next 03/16] net/mlx5: DR, Remove unused struct member in matcher Saeed Mahameed
2021-12-29  6:24 ` [net-next 04/16] net/mlx5: DR, Rename list field in matcher struct to list_node Saeed Mahameed
2021-12-29  6:24 ` [net-next 05/16] net/mlx5: DR, Add check for flex parser ID value Saeed Mahameed
2021-12-29  6:24 ` [net-next 06/16] net/mlx5: DR, Add missing reserved fields to dr_match_param Saeed Mahameed
2021-12-29  6:24 ` [net-next 07/16] net/mlx5: DR, Add support for dumping steering info Saeed Mahameed
2021-12-30  2:16   ` Jakub Kicinski
2021-12-31  1:49     ` Yevgeny Kliteynik
2021-12-31  2:09       ` Jakub Kicinski
2022-01-02 10:18         ` Yevgeny Kliteynik
2021-12-29  6:24 ` [net-next 08/16] net/mlx5: DR, Add support for UPLINK destination type Saeed Mahameed
2021-12-29  6:24 ` [net-next 09/16] net/mlx5: DR, Warn on failure to destroy objects due to refcount Saeed Mahameed
2021-12-29  6:24 ` [net-next 10/16] net/mlx5: Add misc5 flow table match parameters Saeed Mahameed
2021-12-29  6:24 ` [net-next 11/16] net/mlx5: DR, Add misc5 to match_param structs Saeed Mahameed
2021-12-29  6:24 ` [net-next 12/16] net/mlx5: DR, Support matching on tunnel headers 0 and 1 Saeed Mahameed
2021-12-29  6:24 ` [net-next 13/16] net/mlx5: DR, Add support for matching on geneve_tlv_option_0_exist field Saeed Mahameed
2021-12-29  6:25 ` [net-next 14/16] net/mlx5: DR, Improve steering for empty or RX/TX-only matchers Saeed Mahameed
2021-12-29  6:25 ` Saeed Mahameed [this message]
2021-12-29  6:25 ` [net-next 16/16] net/mlx5: Set SMFS as a default steering mode if device supports it 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=20211229062502.24111-16-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@nvidia.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).