diff mbox series

[v2,7/8] net/hns3: add restriction on setting VF MTU

Message ID	20200909092339.31488-8-huwei013@chinasoftinc.com (mailing list archive)
State	Accepted, archived
Delegated to:	Ferruh Yigit
Headers	From: "Wei Hu (Xavier)" <huwei013@chinasoftinc.com> To: <dev@dpdk.org> CC: <xavier.huwei@huawei.com> Date: Wed, 9 Sep 2020 17:23:38 +0800 Message-ID: <20200909092339.31488-8-huwei013@chinasoftinc.com> In-Reply-To: <20200909092339.31488-1-huwei013@chinasoftinc.com> References: <20200907090825.1761-1-huwei013@chinasoftinc.com> <20200909092339.31488-1-huwei013@chinasoftinc.com> MIME-Version: 1.0 Content-Type: text/plain Subject: [dpdk-dev] [PATCH v2 7/8] net/hns3: add restriction on setting VF MTU Precedence: list Errors-To: dev-bounces@dpdk.org Sender: "dev" <dev-bounces@dpdk.org>
Series	net/hns3: updates for Rx Tx \| [v2,0/8] net/hns3: updates for Rx Tx [v2,1/8] net/hns3: report Rx free threshold [v2,2/8] net/hns3: reduce address calculation in Rx [v2,3/8] net/hns3: add simple Rx process function [v2,4/8] net/hns3: add simple Tx process function [v2,5/8] net/hns3: add vector Tx burst with NEON instructions [v2,6/8] net/hns3: add vector Rx burst with NEON instructions [v2,7/8] net/hns3: add restriction on setting VF MTU [v2,8/8] net/hns3: fix segfault when Tx multiple buffer packets

Checks

Context	Check	Description
ci/checkpatch	success	coding style OK

Commit Message

Wei Hu (Xavier) Sept. 9, 2020, 9:23 a.m. UTC

  From: "Wei Hu (Xavier)" <xavier.huwei@huawei.com>

when Rx of scattered packets is off, we have some possibility of using
vector Rx process function or simple Rx functions in hns3 PMD driver.
If the input MTU is increased and the maximum length of received packets
is greater than the length of a buffer for Rx packets, the hardware network
engine needs to use multiple BDs and buffers to store these packets. This
will cause problems when still using vector Rx process function or simple
Rx function to receiving packets. So, when Rx of scattered packets is off
and device is started, it is not permitted to increase MTU so that the
maximum length of Rx packets is greater than Rx buffer length.

Signed-off-by: Chengwen Feng <fengchengwen@huawei.com>
Signed-off-by: Wei Hu (Xavier) <xavier.huwei@huawei.com>
---
 drivers/net/hns3/hns3_ethdev_vf.c | 19 +++++++++++++++++++
 1 file changed, 19 insertions(+)

diff mbox series

Patch

diff --git a/drivers/net/hns3/hns3_ethdev_vf.c b/drivers/net/hns3/hns3_ethdev_vf.c
index 93f2c93..44e51b5 100644
--- a/drivers/net/hns3/hns3_ethdev_vf.c
+++ b/drivers/net/hns3/hns3_ethdev_vf.c
@@ -871,6 +871,25 @@  hns3vf_dev_mtu_set(struct rte_eth_dev *dev, uint16_t mtu)
 		return -EIO;
 	}
 
+	/*
+	 * when Rx of scattered packets is off, we have some possibility of
+	 * using vector Rx process function or simple Rx functions in hns3 PMD
+	 * driver. If the input MTU is increased and the maximum length of
+	 * received packets is greater than the length of a buffer for Rx
+	 * packet, the hardware network engine needs to use multiple BDs and
+	 * buffers to store these packets. This will cause problems when still
+	 * using vector Rx process function or simple Rx function to receiving
+	 * packets. So, when Rx of scattered packets is off and device is
+	 * started, it is not permitted to increase MTU so that the maximum
+	 * length of Rx packets is greater than Rx buffer length.
+	 */
+	if (dev->data->dev_started && !dev->data->scattered_rx &&
+	    frame_size > hw->rx_buf_len) {
+		hns3_err(hw, "failed to set mtu because current is "
+			"not scattered rx mode");
+		return -EOPNOTSUPP;
+	}
+
 	rte_spinlock_lock(&hw->lock);
 	ret = hns3vf_config_mtu(hw, mtu);
 	if (ret) {