DPDK patches and discussions
 help / color / mirror / Atom feed
From: Xiaoyun wang <cloud.wangxiaoyun@huawei.com>
To: <dev@dpdk.org>
Cc: <ferruh.yigit@intel.com>, <bluca@debian.org>,
	<luoxianjun@huawei.com>, <luoxingyu@huawei.com>,
	<zhouguoyang@huawei.com>, <yin.yinshi@huawei.com>,
	<david.yangxiaoliang@huawei.com>, <zhaohui8@huawei.com>,
	<zhengjingzhou@huawei.com>,
	Xiaoyun wang <cloud.wangxiaoyun@huawei.com>, <stable@dpdk.org>
Subject: [dpdk-dev] [PATCH v1 3/5] net/hinic: fix setting promiscuous mode problem
Date: Sat, 27 Jun 2020 11:55:46 +0800	[thread overview]
Message-ID: <7d949e59e08b619ba4820d59b57fe9e2efb9c12a.1593228634.git.cloud.wangxiaoyun@huawei.com> (raw)
In-Reply-To: <cover.1593228634.git.cloud.wangxiaoyun@huawei.com>

When setting promiscuous or allmulticast mode, increase
multi-thread resource protection, because the patch
"net/bonding: prefer allmulti to promiscuous for LACP"
adds trying to use allmulti when adding a slave, and
EVS bond driver also sets promisc with another thread,
which may lead to thread reentry and cause failure to
set promiscuous mode.

Fixes: cb7b6606ebff ("net/hinic: add RSS stats and promiscuous ops")

Cc: stable@dpdk.org
Signed-off-by: Xiaoyun wang <cloud.wangxiaoyun@huawei.com>
---
 drivers/net/hinic/hinic_pmd_ethdev.c | 34 ++++++++++++++++++++++++++++++----
 drivers/net/hinic/hinic_pmd_ethdev.h |  1 +
 2 files changed, 31 insertions(+), 4 deletions(-)

diff --git a/drivers/net/hinic/hinic_pmd_ethdev.c b/drivers/net/hinic/hinic_pmd_ethdev.c
index 0c3e1c0..8de9714 100644
--- a/drivers/net/hinic/hinic_pmd_ethdev.c
+++ b/drivers/net/hinic/hinic_pmd_ethdev.c
@@ -1271,14 +1271,25 @@ static void hinic_disable_interrupt(struct rte_eth_dev *dev)
 
 static int hinic_set_dev_promiscuous(struct hinic_nic_dev *nic_dev, bool enable)
 {
-	u32 rx_mode_ctrl = nic_dev->rx_mode_status;
+	u32 rx_mode_ctrl;
+	int err;
+
+	err = hinic_mutex_lock(&nic_dev->rx_mode_mutex);
+	if (err)
+		return err;
+
+	rx_mode_ctrl = nic_dev->rx_mode_status;
 
 	if (enable)
 		rx_mode_ctrl |= HINIC_RX_MODE_PROMISC;
 	else
 		rx_mode_ctrl &= (~HINIC_RX_MODE_PROMISC);
 
-	return hinic_config_rx_mode(nic_dev, rx_mode_ctrl);
+	err = hinic_config_rx_mode(nic_dev, rx_mode_ctrl);
+
+	(void)hinic_mutex_unlock(&nic_dev->rx_mode_mutex);
+
+	return err;
 }
 
 /**
@@ -1731,14 +1742,25 @@ static void hinic_remove_all_vlanid(struct rte_eth_dev *eth_dev)
 static int hinic_set_dev_allmulticast(struct hinic_nic_dev *nic_dev,
 				bool enable)
 {
-	u32 rx_mode_ctrl = nic_dev->rx_mode_status;
+	u32 rx_mode_ctrl;
+	int err;
+
+	err = hinic_mutex_lock(&nic_dev->rx_mode_mutex);
+	if (err)
+		return err;
+
+	rx_mode_ctrl = nic_dev->rx_mode_status;
 
 	if (enable)
 		rx_mode_ctrl |= HINIC_RX_MODE_MC_ALL;
 	else
 		rx_mode_ctrl &= (~HINIC_RX_MODE_MC_ALL);
 
-	return hinic_config_rx_mode(nic_dev, rx_mode_ctrl);
+	err = hinic_config_rx_mode(nic_dev, rx_mode_ctrl);
+
+	(void)hinic_mutex_unlock(&nic_dev->rx_mode_mutex);
+
+	return err;
 }
 
 /**
@@ -3130,6 +3152,8 @@ static int hinic_func_init(struct rte_eth_dev *eth_dev)
 	}
 	rte_bit_relaxed_set32(HINIC_DEV_INTR_EN, &nic_dev->dev_status);
 
+	hinic_mutex_init(&nic_dev->rx_mode_mutex, NULL);
+
 	/* initialize filter info */
 	filter_info = &nic_dev->filter;
 	tcam_info = &nic_dev->tcam;
@@ -3204,6 +3228,8 @@ static int hinic_dev_uninit(struct rte_eth_dev *dev)
 	if (rte_eal_process_type() != RTE_PROC_PRIMARY)
 		return 0;
 
+	hinic_mutex_destroy(&nic_dev->rx_mode_mutex);
+
 	hinic_dev_close(dev);
 
 	dev->dev_ops = NULL;
diff --git a/drivers/net/hinic/hinic_pmd_ethdev.h b/drivers/net/hinic/hinic_pmd_ethdev.h
index 77b4b91..3f2d51d 100644
--- a/drivers/net/hinic/hinic_pmd_ethdev.h
+++ b/drivers/net/hinic/hinic_pmd_ethdev.h
@@ -329,6 +329,7 @@ struct hinic_nic_dev {
 	unsigned int flags;
 	struct nic_service_cap nic_cap;
 	u32 rx_mode_status;	/* promisc or allmulticast */
+	pthread_mutex_t rx_mode_mutex;
 	u32 dev_status;
 
 	char proc_dev_name[HINIC_DEV_NAME_LEN];
-- 
1.8.3.1


  parent reply	other threads:[~2020-06-27  3:33 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-06-27  3:55 [dpdk-dev] [PATCH v1 0/5] fix promisc and tcam problem Xiaoyun wang
2020-06-27  3:55 ` [dpdk-dev] [PATCH v1 1/5] net/hinic/base: add out parameter for mgmt sync channel Xiaoyun wang
2020-06-27  3:55 ` [dpdk-dev] [PATCH v1 2/5] net/hinic/base: remove unused parameter Xiaoyun wang
2020-06-27  3:55 ` Xiaoyun wang [this message]
2020-06-27  3:55 ` [dpdk-dev] [PATCH v1 4/5] net/hinic: add tcam filter switch for FDIR Xiaoyun wang
2020-06-27  3:55 ` [dpdk-dev] [PATCH v1 5/5] net/hinic/base: modify return errors Xiaoyun wang
2020-07-07 19:42   ` Ferruh Yigit
2020-07-03 16:22 ` [dpdk-dev] [PATCH v1 0/5] fix promisc and tcam problem Ferruh Yigit

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=7d949e59e08b619ba4820d59b57fe9e2efb9c12a.1593228634.git.cloud.wangxiaoyun@huawei.com \
    --to=cloud.wangxiaoyun@huawei.com \
    --cc=bluca@debian.org \
    --cc=david.yangxiaoliang@huawei.com \
    --cc=dev@dpdk.org \
    --cc=ferruh.yigit@intel.com \
    --cc=luoxianjun@huawei.com \
    --cc=luoxingyu@huawei.com \
    --cc=stable@dpdk.org \
    --cc=yin.yinshi@huawei.com \
    --cc=zhaohui8@huawei.com \
    --cc=zhengjingzhou@huawei.com \
    --cc=zhouguoyang@huawei.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).