DPDK patches and discussions
 help / color / mirror / Atom feed
From: Xueming Li <xuemingl@nvidia.com>
To: <dev@dpdk.org>
Cc: <xuemingl@nvidia.com>, Matan Azrad <matan@nvidia.com>,
	"Viacheslav Ovsiienko" <viacheslavo@nvidia.com>,
	Dong Zhou <dongzhou@nvidia.com>
Subject: [dpdk-dev] [PATCH 7/8] net/mlx5: fix internal root table flow priroity
Date: Mon, 27 Sep 2021 16:32:55 +0800	[thread overview]
Message-ID: <20210927083256.337450-8-xuemingl@nvidia.com> (raw)
In-Reply-To: <20210927083256.337450-1-xuemingl@nvidia.com>

When creating internal transfer flow on root table with lowerest
priority, the flow was created with u32 priority. It was wrong since
the flow is created in kernel and  max priority supported is 16.

This patch fixes this by adding internal flow check.

Fixes: 5f8ae44dd454 ("net/mlx5: enlarge maximal flow priority")

Signed-off-by: Xueming Li <xuemingl@nvidia.com>
---
 drivers/net/mlx5/mlx5_flow.c    | 7 +++++--
 drivers/net/mlx5/mlx5_flow.h    | 4 ++--
 drivers/net/mlx5/mlx5_flow_dv.c | 3 ++-
 3 files changed, 9 insertions(+), 5 deletions(-)

diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index c914a7120cc..8dc79340f2d 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/net/mlx5/mlx5_flow.c
@@ -980,13 +980,15 @@ mlx5_get_lowest_priority(struct rte_eth_dev *dev,
  *   Pointer to device flow rule attributes.
  * @param[in] subpriority
  *   The priority based on the items.
+ * @param[in] external
+ *   Flow is user flow.
  * @return
  *   The matcher priority of the flow.
  */
 uint16_t
 mlx5_get_matcher_priority(struct rte_eth_dev *dev,
 			  const struct rte_flow_attr *attr,
-			  uint32_t subpriority)
+			  uint32_t subpriority, bool external)
 {
 	uint16_t priority = (uint16_t)attr->priority;
 	struct mlx5_priv *priv = dev->data->dev_private;
@@ -997,7 +999,8 @@ mlx5_get_matcher_priority(struct rte_eth_dev *dev,
 		return mlx5_os_flow_adjust_priority(dev, priority, subpriority);
 	}
 	if (attr->priority == MLX5_FLOW_LOWEST_PRIO_INDICATOR)
-		priority = MLX5_NON_ROOT_FLOW_MAX_PRIO;
+		priority = external ?
+			   MLX5_NON_ROOT_FLOW_MAX_PRIO : priv->config.flow_prio;
 	return priority * 3 + subpriority;
 }
 
diff --git a/drivers/net/mlx5/mlx5_flow.h b/drivers/net/mlx5/mlx5_flow.h
index 861e18fb3b1..1e31d25f319 100644
--- a/drivers/net/mlx5/mlx5_flow.h
+++ b/drivers/net/mlx5/mlx5_flow.h
@@ -1434,8 +1434,8 @@ uint32_t mlx5_flow_adjust_priority(struct rte_eth_dev *dev, int32_t priority,
 uint32_t mlx5_get_lowest_priority(struct rte_eth_dev *dev,
 					const struct rte_flow_attr *attr);
 uint16_t mlx5_get_matcher_priority(struct rte_eth_dev *dev,
-				     const struct rte_flow_attr *attr,
-				     uint32_t subpriority);
+				   const struct rte_flow_attr *attr,
+				   uint32_t subpriority, bool external);
 int mlx5_flow_get_reg_id(struct rte_eth_dev *dev,
 				     enum mlx5_feature_name feature,
 				     uint32_t id,
diff --git a/drivers/net/mlx5/mlx5_flow_dv.c b/drivers/net/mlx5/mlx5_flow_dv.c
index e388e2d5e10..3744f3e5917 100644
--- a/drivers/net/mlx5/mlx5_flow_dv.c
+++ b/drivers/net/mlx5/mlx5_flow_dv.c
@@ -13633,7 +13633,8 @@ flow_dv_translate(struct rte_eth_dev *dev,
 	matcher.crc = rte_raw_cksum((const void *)matcher.mask.buf,
 				    matcher.mask.size);
 	matcher.priority = mlx5_get_matcher_priority(dev, attr,
-					matcher.priority);
+						     matcher.priority,
+						     dev_flow->external);
 	/**
 	 * When creating meter drop flow in drop table, using original
 	 * 5-tuple match, the matcher priority should be lower than
-- 
2.33.0


  parent reply	other threads:[~2021-09-27  8:34 UTC|newest]

Thread overview: 47+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-09-27  8:32 [dpdk-dev] [PATCH 0/8] net/mlx5: support more than 255 representors Xueming Li
2021-09-27  8:32 ` [dpdk-dev] [PATCH 1/8] common/mlx5: add netlink API to get RDMA port state Xueming Li
2021-09-27  8:32 ` [dpdk-dev] [PATCH 2/8] net/mlx5: use netlink when IB port greater than 255 Xueming Li
2021-09-27  8:32 ` [dpdk-dev] [PATCH 3/8] net/mlx5: improve Verbs flow priority discover for scalable Xueming Li
2021-09-27  8:32 ` [dpdk-dev] [PATCH 4/8] net/mlx5: check DevX to support more Verb ports Xueming Li
2021-09-27  8:32 ` [dpdk-dev] [PATCH 5/8] net/mlx5: support flow item port of switch manager Xueming Li
2021-09-27  8:32 ` [dpdk-dev] [PATCH 6/8] net/mlx5: supports flow item of normal Tx queue Xueming Li
2021-09-27  8:32 ` Xueming Li [this message]
2021-09-27  8:32 ` [dpdk-dev] [PATCH 8/8] net/mlx5: enable DevX Tx queue creation Xueming Li
2021-10-16  8:07 ` [dpdk-dev] [PATCH v2 0/8] net/mlx5: support more than 255 representors Xueming Li
2021-10-16  8:07   ` [dpdk-dev] [PATCH v2 1/8] common/mlx5: add netlink API to get RDMA port state Xueming Li
2021-10-19  8:23     ` Slava Ovsiienko
2021-10-16  8:07   ` [dpdk-dev] [PATCH v2 2/8] net/mlx5: use netlink when IB port greater than 255 Xueming Li
2021-10-19  8:24     ` Slava Ovsiienko
2021-10-16  8:07   ` [dpdk-dev] [PATCH v2 3/8] net/mlx5: improve Verbs flow priority discover for scalable Xueming Li
2021-10-19  8:26     ` Slava Ovsiienko
2021-10-16  8:07   ` [dpdk-dev] [PATCH v2 4/8] net/mlx5: support E-Switch manager egress traffic match Xueming Li
2021-10-19  8:26     ` Slava Ovsiienko
2021-10-16  8:07   ` [dpdk-dev] [PATCH v2 5/8] net/mlx5: supports flow item of normal Tx queue Xueming Li
2021-10-19  8:27     ` Slava Ovsiienko
2021-10-16  8:07   ` [dpdk-dev] [PATCH v2 6/8] net/mlx5: fix internal root table flow priroity Xueming Li
2021-10-19  8:28     ` Slava Ovsiienko
2021-10-16  8:07   ` [dpdk-dev] [PATCH v2 7/8] net/mlx5: enable DevX Tx queue creation Xueming Li
2021-10-19  8:29     ` Slava Ovsiienko
2021-10-16  8:07   ` [dpdk-dev] [PATCH v2 8/8] net/mlx5: check DevX to support more Verbs ports Xueming Li
2021-10-19  8:30     ` Slava Ovsiienko
2021-10-19 10:34 ` [dpdk-dev] [PATCH v3 0/8] net/mlx5: support more than 255 representors Xueming Li
2021-10-19 10:34   ` [dpdk-dev] [PATCH v3 1/8] common/mlx5: add netlink API to get RDMA port state Xueming Li
2021-10-21 13:34     ` Ferruh Yigit
2021-10-19 10:34   ` [dpdk-dev] [PATCH v3 2/8] net/mlx5: use netlink when IB port greater than 255 Xueming Li
2021-10-19 10:34   ` [dpdk-dev] [PATCH v3 3/8] net/mlx5: improve Verbs flow priority discover for scalable Xueming Li
2021-10-19 10:34   ` [dpdk-dev] [PATCH v3 4/8] net/mlx5: support E-Switch manager egress traffic match Xueming Li
2021-10-19 10:34   ` [dpdk-dev] [PATCH v3 5/8] net/mlx5: supports flow item of normal Tx queue Xueming Li
2021-10-19 10:34   ` [dpdk-dev] [PATCH v3 6/8] net/mlx5: fix internal root table flow priroity Xueming Li
2021-10-19 10:35   ` [dpdk-dev] [PATCH v3 7/8] net/mlx5: enable DevX Tx queue creation Xueming Li
2021-10-19 10:35   ` [dpdk-dev] [PATCH v3 8/8] net/mlx5: check DevX to support more Verbs ports Xueming Li
2021-10-20 13:40   ` [dpdk-dev] [PATCH v3 0/8] net/mlx5: support more than 255 representors Raslan Darawsheh
2021-10-20 16:00     ` Xueming(Steven) Li
2021-10-22  9:11 ` [dpdk-dev] [PATCH v4 " Xueming Li
2021-10-22  9:11   ` [dpdk-dev] [PATCH v4 1/8] common/mlx5: add netlink API to get RDMA port state Xueming Li
2021-10-22  9:11   ` [dpdk-dev] [PATCH v4 2/8] net/mlx5: use netlink when IB port greater than 255 Xueming Li
2021-10-22  9:11   ` [dpdk-dev] [PATCH v4 3/8] net/mlx5: improve Verbs flow priority discover for scalable Xueming Li
2021-10-22  9:11   ` [dpdk-dev] [PATCH v4 4/8] net/mlx5: support E-Switch manager egress traffic match Xueming Li
2021-10-22  9:11   ` [dpdk-dev] [PATCH v4 5/8] net/mlx5: supports flow item of normal Tx queue Xueming Li
2021-10-22  9:11   ` [dpdk-dev] [PATCH v4 6/8] net/mlx5: fix internal root table flow priroity Xueming Li
2021-10-22  9:11   ` [dpdk-dev] [PATCH v4 7/8] net/mlx5: enable DevX Tx queue creation Xueming Li
2021-10-22  9:11   ` [dpdk-dev] [PATCH v4 8/8] net/mlx5: check DevX to support more Verbs ports Xueming Li

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=20210927083256.337450-8-xuemingl@nvidia.com \
    --to=xuemingl@nvidia.com \
    --cc=dev@dpdk.org \
    --cc=dongzhou@nvidia.com \
    --cc=matan@nvidia.com \
    --cc=viacheslavo@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).