diff mbox series

[v2] net/ixgbe: fix vf mac remains

Message ID 20210906023513.9608-1-chenqiming_huawei@163.com (mailing list archive)
State Not Applicable, archived
Delegated to: Qi Zhang
Headers
Series [v2] net/ixgbe: fix vf mac remains |

Checks

Context Check Description
ci/checkpatch success coding style OK
ci/github-robot: build success github build: passed
ci/iol-broadcom-Performance success Performance Testing PASS
ci/iol-broadcom-Functional success Functional Testing PASS
ci/iol-intel-Functional success Functional Testing PASS
ci/iol-intel-Performance success Performance Testing PASS
ci/iol-aarch64-compile-testing success Testing PASS
ci/iol-x86_64-unit-testing success Testing PASS
ci/iol-mellanox-Performance success Performance Testing PASS
ci/iol-x86_64-compile-testing success Testing PASS
ci/Intel-compilation success Compilation OK
ci/intel-Testing fail Testing issues

Commit Message

Qiming Chen Sept. 6, 2021, 2:35 a.m. UTC 
  In the following two scenarios, the mac address residual problem
will occur:
1) In the VF hard pass-through environment, after adding the mac
address, the process restarts, and the previously added mac is
still valid;
2) In the vf hard pass-through environment, after the mac address
is issued, the port will start/stop, and the previously added mac
is still valid;

The patch clears the mac address at the start and stop of the vf
port. After the start is cleared, the rte framework will restore
the mac addition to solve the problem of residual mac addresses.

Fixes: af75078fece3 ("first public release")
Cc: stable@dpdk.org

Signed-off-by: Qiming Chen <chenqiming_huawei@163.com>
---
v2:
  Modify fixes commit
---
 drivers/net/ixgbe/ixgbe_ethdev.c | 6 ++++++
 1 file changed, 6 insertions(+)

Comments

Wang, Haiyue Sept. 8, 2021, 8:35 a.m. UTC | #1 
> -----Original Message-----
> From: Qiming Chen <chenqiming_huawei@163.com>
> Sent: Monday, September 6, 2021 10:35
> To: dev@dpdk.org
> Cc: Wang, Haiyue <haiyue.wang@intel.com>; Qiming Chen <chenqiming_huawei@163.com>; stable@dpdk.org
> Subject: [PATCH v2] net/ixgbe: fix vf mac remains
> 
> In the following two scenarios, the mac address residual problem
> will occur:
> 1) In the VF hard pass-through environment, after adding the mac
> address, the process restarts, and the previously added mac is
> still valid;
> 2) In the vf hard pass-through environment, after the mac address
> is issued, the port will start/stop, and the previously added mac
> is still valid;

How did you test it ?

From the ixgbe PF, it will clean the MACs:

ixgbe_del_mac_filter(adapter, adapter->vfinfo[vf].vf_mac_addresses, vf);
ixgbe_set_vf_macvlan(adapter, vf, 0, NULL);

> 
> The patch clears the mac address at the start and stop of the vf
> port. After the start is cleared, the rte framework will restore
> the mac addition to solve the problem of residual mac addresses.
> 
> Fixes: af75078fece3 ("first public release")
> Cc: stable@dpdk.org
> 
> Signed-off-by: Qiming Chen <chenqiming_huawei@163.com>
> ---
> v2:
>   Modify fixes commit
> ---
>  drivers/net/ixgbe/ixgbe_ethdev.c | 6 ++++++
>  1 file changed, 6 insertions(+)
> 
> diff --git a/drivers/net/ixgbe/ixgbe_ethdev.c b/drivers/net/ixgbe/ixgbe_ethdev.c
> index 6a91f104e1..e40350d86e 100644
> --- a/drivers/net/ixgbe/ixgbe_ethdev.c
> +++ b/drivers/net/ixgbe/ixgbe_ethdev.c
> @@ -5410,6 +5410,9 @@ ixgbevf_dev_start(struct rte_eth_dev *dev)
>  	 */
>  	ixgbevf_dev_link_update(dev, 0);
> 
> +	/* Clear the mac address, the rte frame will be restored */
> +	ixgbevf_set_uc_addr_vf(hw, 0, NULL);
> +
>  	hw->adapter_stopped = false;
> 
>  	return 0;
> @@ -5454,6 +5457,9 @@ ixgbevf_dev_stop(struct rte_eth_dev *dev)
>  		intr_handle->intr_vec = NULL;
>  	}
> 
> +	/* Clear the mac address */
> +	ixgbevf_set_uc_addr_vf(hw, 0, NULL);
> +
>  	adapter->rss_reta_updated = 0;
> 
>  	return 0;
> --
> 2.30.1.windows.1
Qiming Chen Sept. 9, 2021, 1:41 a.m. UTC | #2 
thanks your reply.


kernal PF + dpdk VF mode, I developed a demo based on dpdk. I used igb_uio to take over the vf port, and then added a unicast MAC address to the port. As a result, the MAC packet was flowed to test ok, and then the demo was restarted, and the traffic still passed.
From this result, it can be seen that the mac remains, and there is no clear. This is vf driver quesion, not dpdk pf.


Also, the two interfaces you posted, I didn't find them in the dpdk code.
On 9/8/2021 16:35，Wang, Haiyue<haiyue.wang@intel.com> wrote：
-----Original Message-----
From: Qiming Chen <chenqiming_huawei@163.com>
Sent: Monday, September 6, 2021 10:35
To: dev@dpdk.org
Cc: Wang, Haiyue <haiyue.wang@intel.com>; Qiming Chen <chenqiming_huawei@163.com>; stable@dpdk.org
Subject: [PATCH v2] net/ixgbe: fix vf mac remains

In the following two scenarios, the mac address residual problem
will occur:
1) In the VF hard pass-through environment, after adding the mac
address, the process restarts, and the previously added mac is
still valid;
2) In the vf hard pass-through environment, after the mac address
is issued, the port will start/stop, and the previously added mac
is still valid;

How did you test it ?

From the ixgbe PF, it will clean the MACs:

ixgbe_del_mac_filter(adapter, adapter->vfinfo[vf].vf_mac_addresses, vf);
ixgbe_set_vf_macvlan(adapter, vf, 0, NULL);


The patch clears the mac address at the start and stop of the vf
port. After the start is cleared, the rte framework will restore
the mac addition to solve the problem of residual mac addresses.

Fixes: af75078fece3 ("first public release")
Cc: stable@dpdk.org

Signed-off-by: Qiming Chen <chenqiming_huawei@163.com>
---
v2:
Modify fixes commit
---
drivers/net/ixgbe/ixgbe_ethdev.c | 6 ++++++
1 file changed, 6 insertions(+)

diff --git a/drivers/net/ixgbe/ixgbe_ethdev.c b/drivers/net/ixgbe/ixgbe_ethdev.c
index 6a91f104e1..e40350d86e 100644
--- a/drivers/net/ixgbe/ixgbe_ethdev.c
+++ b/drivers/net/ixgbe/ixgbe_ethdev.c
@@ -5410,6 +5410,9 @@ ixgbevf_dev_start(struct rte_eth_dev *dev)
*/
ixgbevf_dev_link_update(dev, 0);

+  /* Clear the mac address, the rte frame will be restored */
+  ixgbevf_set_uc_addr_vf(hw, 0, NULL);
+
hw->adapter_stopped = false;

return 0;
@@ -5454,6 +5457,9 @@ ixgbevf_dev_stop(struct rte_eth_dev *dev)
intr_handle->intr_vec = NULL;
}

+  /* Clear the mac address */
+  ixgbevf_set_uc_addr_vf(hw, 0, NULL);
+
adapter->rss_reta_updated = 0;

return 0;
--
2.30.1.windows.1
Wang, Haiyue Sept. 9, 2021, 2:42 a.m. UTC | #3 
Please use the testpmd firstly to verify the PMD's issue!!!!!

And I tested in testpmd.

1. mac_addr set 0 00:11:22:33:44:88
2. quit testpmd.
3. Add debug message in PF: 

https://www.intel.com/content/www/us/en/download/14302/14687/intel-network-adapter-driver-for-pcie-intel-10-gigabit-ethernet-network-connections-under-linux.html

--- a/src/ixgbe_sriov.c
+++ b/src/ixgbe_sriov.c
@@ -776,6 +776,9 @@ static inline void ixgbe_vf_reset_event(struct ixgbe_adapter *adapter, u32 vf)
        /* Flush and reset the mta with the new values */
        ixgbe_set_rx_mode(adapter->netdev);

+       dev_info(&adapter->pdev->dev, "Deleting VF%u mac %pM \n", vf,
+                adapter->vfinfo[vf].vf_mac_addresses);
+
        ixgbe_del_mac_filter(adapter, adapter->vfinfo[vf].vf_mac_addresses, vf);
        ixgbe_set_vf_macvlan(adapter, vf, 0, NULL);

You can see PF will delete the MAC when VF exit.
  ixgbe 0000:86:00.0: Deleting VF0 mac 00:11:22:33:44:88


BR,
Haiyue

From: Qiming Chen <chenqiming_huawei@163.com> 
Sent: Thursday, September 9, 2021 09:41
To: Wang, Haiyue <haiyue.wang@intel.com>
Cc: dev@dpdk.org; stable@dpdk.org
Subject: Re: [PATCH v2] net/ixgbe: fix vf mac remains

thanks your reply.

kernal PF + dpdk VF mode, I developed a demo based on dpdk. I used igb_uio to take over the vf port, and then added a unicast MAC address to the port. As a result, the MAC packet was flowed to test ok, and then the demo was restarted, and the traffic still passed.
From this result, it can be seen that the mac remains, and there is no clear. This is vf driver quesion, not dpdk pf.

Also, the two interfaces you posted, I didn't find them in the dpdk code.
On 9/8/2021 16:35，mailto:haiyue.wang@intel.com wrote： 
-----Original Message-----
From: Qiming Chen <mailto:chenqiming_huawei@163.com>
Sent: Monday, September 6, 2021 10:35
To: mailto:dev@dpdk.org
Cc: Wang, Haiyue <mailto:haiyue.wang@intel.com>; Qiming Chen <mailto:chenqiming_huawei@163.com>; mailto:stable@dpdk.org
Subject: [PATCH v2] net/ixgbe: fix vf mac remains

In the following two scenarios, the mac address residual problem
will occur:
1) In the VF hard pass-through environment, after adding the mac
address, the process restarts, and the previously added mac is
still valid;
2) In the vf hard pass-through environment, after the mac address
is issued, the port will start/stop, and the previously added mac
is still valid;

How did you test it ?

From the ixgbe PF, it will clean the MACs:

ixgbe_del_mac_filter(adapter, adapter->vfinfo[vf].vf_mac_addresses, vf);
ixgbe_set_vf_macvlan(adapter, vf, 0, NULL);

The patch clears the mac address at the start and stop of the vf
port. After the start is cleared, the rte framework will restore
the mac addition to solve the problem of residual mac addresses.

Fixes: af75078fece3 ("first public release")
Cc: mailto:stable@dpdk.org

Signed-off-by: Qiming Chen <mailto:chenqiming_huawei@163.com>
---
v2:
Modify fixes commit
---
drivers/net/ixgbe/ixgbe_ethdev.c | 6 ++++++
1 file changed, 6 insertions(+)

diff --git a/drivers/net/ixgbe/ixgbe_ethdev.c b/drivers/net/ixgbe/ixgbe_ethdev.c
index 6a91f104e1..e40350d86e 100644
--- a/drivers/net/ixgbe/ixgbe_ethdev.c
+++ b/drivers/net/ixgbe/ixgbe_ethdev.c
@@ -5410,6 +5410,9 @@ ixgbevf_dev_start(struct rte_eth_dev *dev)
*/
ixgbevf_dev_link_update(dev, 0);

+  /* Clear the mac address, the rte frame will be restored */
+  ixgbevf_set_uc_addr_vf(hw, 0, NULL);
+
hw->adapter_stopped = false;

return 0;
@@ -5454,6 +5457,9 @@ ixgbevf_dev_stop(struct rte_eth_dev *dev)
intr_handle->intr_vec = NULL;
}

+  /* Clear the mac address */
+  ixgbevf_set_uc_addr_vf(hw, 0, NULL);
+
adapter->rss_reta_updated = 0;

return 0;
--
2.30.1.windows.1
Qiming Chen Sept. 9, 2021, 3:59 a.m. UTC | #4 
OK，Thanks your test and reply. 
It may be that the pf kernel version I tested before is older, and the code snippets you posted have not yet been implemented.


I will test it again based on the latest version.
Qiming Chen Sept. 9, 2021, 8:41 a.m. UTC | #5 
I have re-tested with kernel driver version 5.6.3 and found that this problem has been solved, thank you again for your reply.
diff mbox series

Patch

diff --git a/drivers/net/ixgbe/ixgbe_ethdev.c b/drivers/net/ixgbe/ixgbe_ethdev.c
index 6a91f104e1..e40350d86e 100644
--- a/drivers/net/ixgbe/ixgbe_ethdev.c
+++ b/drivers/net/ixgbe/ixgbe_ethdev.c
@@ -5410,6 +5410,9 @@  ixgbevf_dev_start(struct rte_eth_dev *dev)
 	 */
 	ixgbevf_dev_link_update(dev, 0);
 
+	/* Clear the mac address, the rte frame will be restored */
+	ixgbevf_set_uc_addr_vf(hw, 0, NULL);
+
 	hw->adapter_stopped = false;
 
 	return 0;
@@ -5454,6 +5457,9 @@  ixgbevf_dev_stop(struct rte_eth_dev *dev)
 		intr_handle->intr_vec = NULL;
 	}
 
+	/* Clear the mac address */
+	ixgbevf_set_uc_addr_vf(hw, 0, NULL);
+
 	adapter->rss_reta_updated = 0;
 
 	return 0;