DPDK patches and discussions
 help / color / mirror / Atom feed
From: Xiaoyun Li <xiaoyun.li@intel.com>
To: ferruh.yigit@intel.com, qi.z.zhang@intel.com,
	beilei.xing@intel.com, dev@dpdk.org
Cc: Xiaoyun Li <xiaoyun.li@intel.com>
Subject: [dpdk-dev] [PATCH v4] net/i40e: config fdir automatically
Date: Sun, 30 Sep 2018 10:08:23 +0800	[thread overview]
Message-ID: <20180930020823.48273-1-xiaoyun.li@intel.com> (raw)
In-Reply-To: <20180928072453.6348-1-xiaoyun.li@intel.com>

I40e driver needed users to config exact fdir mode to create rte_flow
rules but it shouldn't. This patch allows users to create rte_flow rules
without configuring fdir mode and let the driver config fdir automatically.
And remove the workaround in flow filtering example.

Signed-off-by: Xiaoyun Li <xiaoyun.li@intel.com>
---
v4:
 * Polished commit log.
 * Added comments that the original fdir config codes should be removed
 * when the legacy filter API is deprecated.
v3:
 * Fixed fdir config doesn't automatically remove when last flow is
 * destroyed.
v2:
 * Added fdir teardown in i40e_flow_flush_fdir_filter.
 * Replace TAILQ_FIRST with TAILQ_EMPTY which is more intuitive.
 * Remove the workaround in flow_filtering example since the driver will
 * set the fdir config automatically.
---
 drivers/net/i40e/i40e_ethdev.c | 10 +++++++++-
 drivers/net/i40e/i40e_flow.c   | 35 +++++++++++++++++++++++++++++-----
 examples/flow_filtering/main.c | 16 ----------------
 3 files changed, 39 insertions(+), 22 deletions(-)

diff --git a/drivers/net/i40e/i40e_ethdev.c b/drivers/net/i40e/i40e_ethdev.c
index de25de650..3fd0d5dbf 100644
--- a/drivers/net/i40e/i40e_ethdev.c
+++ b/drivers/net/i40e/i40e_ethdev.c
@@ -1787,6 +1787,10 @@ i40e_dev_configure(struct rte_eth_dev *dev)
 	ad->tx_simple_allowed = true;
 	ad->tx_vec_allowed = true;
 
+	/* Only legacy filter API needs the following fdir config. So when the
+	 * legacy filter API is deprecated, the following codes should also be
+	 * removed.
+	 */
 	if (dev->data->dev_conf.fdir_conf.mode == RTE_FDIR_MODE_PERFECT) {
 		ret = i40e_fdir_setup(pf);
 		if (ret != I40E_SUCCESS) {
@@ -1844,7 +1848,11 @@ i40e_dev_configure(struct rte_eth_dev *dev)
 	rte_free(pf->vmdq);
 	pf->vmdq = NULL;
 err:
-	/* need to release fdir resource if exists */
+	/* Need to release fdir resource if exists.
+	 * Only legacy filter API needs the following fdir config. So when the
+	 * legacy filter API is deprecated, the following code should also be
+	 * removed.
+	 */
 	i40e_fdir_teardown(pf);
 	return ret;
 }
diff --git a/drivers/net/i40e/i40e_flow.c b/drivers/net/i40e/i40e_flow.c
index c67b264de..68ae00a27 100644
--- a/drivers/net/i40e/i40e_flow.c
+++ b/drivers/net/i40e/i40e_flow.c
@@ -3127,6 +3127,7 @@ i40e_flow_parse_fdir_filter(struct rte_eth_dev *dev,
 			    struct rte_flow_error *error,
 			    union i40e_filter_t *filter)
 {
+	struct i40e_pf *pf = I40E_DEV_PRIVATE_TO_PF(dev->data->dev_private);
 	struct i40e_fdir_filter_conf *fdir_filter =
 		&filter->fdir_filter;
 	int ret;
@@ -3148,14 +3149,29 @@ i40e_flow_parse_fdir_filter(struct rte_eth_dev *dev,
 
 	if (dev->data->dev_conf.fdir_conf.mode !=
 	    RTE_FDIR_MODE_PERFECT) {
-		rte_flow_error_set(error, ENOTSUP,
-				   RTE_FLOW_ERROR_TYPE_UNSPECIFIED,
-				   NULL,
-				   "Check the mode in fdir_conf.");
-		return -rte_errno;
+		/* Enable fdir when fdir flow is added at first time. */
+		ret = i40e_fdir_setup(pf);
+		if (ret != I40E_SUCCESS) {
+			rte_flow_error_set(error, ENOTSUP,
+					   RTE_FLOW_ERROR_TYPE_HANDLE,
+					   NULL, "Failed to setup fdir.");
+			return -rte_errno;
+		}
+		ret = i40e_fdir_configure(dev);
+		if (ret < 0) {
+			rte_flow_error_set(error, ENOTSUP,
+					   RTE_FLOW_ERROR_TYPE_HANDLE,
+					   NULL, "Failed to configure fdir.");
+			goto err;
+		}
+
+		dev->data->dev_conf.fdir_conf.mode = RTE_FDIR_MODE_PERFECT;
 	}
 
 	return 0;
+err:
+	i40e_fdir_teardown(pf);
+	return -rte_errno;
 }
 
 /* Parse to get the action info of a tunnel filter
@@ -4708,6 +4724,13 @@ i40e_flow_destroy(struct rte_eth_dev *dev,
 	case RTE_ETH_FILTER_FDIR:
 		ret = i40e_flow_add_del_fdir_filter(dev,
 		       &((struct i40e_fdir_filter *)flow->rule)->fdir, 0);
+
+		/* If the last flow is destroyed, disable fdir. */
+		if (!ret && !TAILQ_EMPTY(&pf->fdir.fdir_list)) {
+			i40e_fdir_teardown(pf);
+			dev->data->dev_conf.fdir_conf.mode =
+				   RTE_FDIR_MODE_NONE;
+		}
 		break;
 	case RTE_ETH_FILTER_HASH:
 		ret = i40e_config_rss_filter_del(dev,
@@ -4900,6 +4923,8 @@ i40e_flow_flush_fdir_filter(struct i40e_pf *pf)
 			pf->fdir.inset_flag[pctype] = 0;
 	}
 
+	i40e_fdir_teardown(pf);
+
 	return ret;
 }
 
diff --git a/examples/flow_filtering/main.c b/examples/flow_filtering/main.c
index b3f85b563..a73d120e3 100644
--- a/examples/flow_filtering/main.c
+++ b/examples/flow_filtering/main.c
@@ -131,22 +131,6 @@ init_port(void)
 				DEV_TX_OFFLOAD_SCTP_CKSUM  |
 				DEV_TX_OFFLOAD_TCP_TSO,
 		},
-		/*
-		 * Initialize fdir_conf of rte_eth_conf.
-		 * Fdir is used in flow filtering for I40e,
-		 * so rte_flow rules involve some fdir
-		 * configurations. In long term it's better
-		 * that drivers don't require any fdir
-		 * configuration for rte_flow, but we need to
-		 * get this workaround so that sample app can
-		 * run on I40e.
-		 */
-		.fdir_conf = {
-			.mode = RTE_FDIR_MODE_PERFECT,
-			.pballoc = RTE_FDIR_PBALLOC_64K,
-			.status = RTE_FDIR_REPORT_STATUS,
-			.drop_queue = 127,
-		},
 	};
 	struct rte_eth_txconf txq_conf;
 	struct rte_eth_rxconf rxq_conf;
-- 
2.17.1

  parent reply	other threads:[~2018-09-30  2:16 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-09-28  7:24 [dpdk-dev] [PATCH] net/i40e: select fdir config automatically Xiaoyun Li
2018-09-28  8:40 ` Xing, Beilei
2018-09-28 10:38   ` Li, Xiaoyun
2018-09-28 11:03 ` [dpdk-dev] [PATCH v2] " Xiaoyun Li
2018-09-29  8:36   ` Xing, Beilei
2018-09-29  8:40     ` Li, Xiaoyun
2018-09-29  8:44 ` [dpdk-dev] [PATCH v3] " Xiaoyun Li
2018-09-30  2:08 ` Xiaoyun Li [this message]
2018-09-30  2:29 ` [dpdk-dev] [PATCH v5] net/i40e: config fdir automatically Xiaoyun Li
2018-09-30  2:41   ` Xing, Beilei
2018-09-30  3:31     ` Zhang, Qi Z

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=20180930020823.48273-1-xiaoyun.li@intel.com \
    --to=xiaoyun.li@intel.com \
    --cc=beilei.xing@intel.com \
    --cc=dev@dpdk.org \
    --cc=ferruh.yigit@intel.com \
    --cc=qi.z.zhang@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).