[PATCH v4 1/2] app/testpmd: fix vlan offload of rxq

[lihuisong (C)]

在 2022/11/4 16:21, Ye, MingjinX 写道:
>
>> -----Original Message-----
>> From: lihuisong (C) <lihuisong at huawei.com>
>> Sent: 2022年11月3日 15:01
>> To: Ye, MingjinX <mingjinx.ye at intel.com>; dev at dpdk.org
>> Cc: stable at dpdk.org; Zhou, YidingX <yidingx.zhou at intel.com>; Singh, Aman
>> Deep <aman.deep.singh at intel.com>; Zhang, Yuying
>> <yuying.zhang at intel.com>
>> Subject: Re: [PATCH v4 1/2] app/testpmd: fix vlan offload of rxq
>>
>>
>> 在 2022/11/3 9:28, Ye, MingjinX 写道:
>>>> -----Original Message-----
>>>> From: lihuisong (C) <lihuisong at huawei.com>
>>>> Sent: 2022年10月28日 10:09
>>>> To: Ye, MingjinX <mingjinx.ye at intel.com>; dev at dpdk.org
>>>> Cc: stable at dpdk.org; Zhou, YidingX <yidingx.zhou at intel.com>; Singh,
>>>> Aman Deep <aman.deep.singh at intel.com>; Zhang, Yuying
>>>> <yuying.zhang at intel.com>
>>>> Subject: Re: [PATCH v4 1/2] app/testpmd: fix vlan offload of rxq
>>>>
>>>>
>>>> 在 2022/10/27 19:02, Ye, MingjinX 写道:
>>>>> Hi lihuisong,
>>>>>
>>>>> This means that queue offloads need to update by recalling
>>>>> dev_configure and setup target queues.
>>>> Why not update queue offloads in PMD?
>>>>> Can you tell me, where is the limitation?
>>>> According to other Rx/Tx offload configurations, this may not be a
>> limitation.
>>>> But it seems to create a dependency on user usage.
>>>>
>>>> Port VLAN releated offloads are set by ethdev ops. There is no
>>>> requirement in ehedev layer that this port needs to stopped when set
>> these offloads.
>>>> Now it depends on user does recall dev_configure and setup queues to
>>>> update queue offloads because of setting these offloads.
>>>>> Thanks,
>>>>> Mingjin
>>>>>
>>>>>> -----Original Message-----
>>>>>> From: lihuisong (C) <lihuisong at huawei.com>
>>>>>> Sent: 2022年10月26日 17:53
>>>>>> To: Ye, MingjinX <mingjinx.ye at intel.com>; dev at dpdk.org
>>>>>> Cc: stable at dpdk.org; Zhou, YidingX <yidingx.zhou at intel.com>; Singh,
>>>>>> Aman Deep <aman.deep.singh at intel.com>; Zhang, Yuying
>>>>>> <yuying.zhang at intel.com>
>>>>>> Subject: Re: [PATCH v4 1/2] app/testpmd: fix vlan offload of rxq
>>>>>>
>>>>>>
>>>>>> 在 2022/10/27 1:10, Mingjin Ye 写道:
>>>>>>> After setting vlan offload in testpmd, the result is not updated
>>>>>>> to rxq. Therefore, the queue needs to be reconfigured after
>>>>>>> executing the "vlan offload" related commands.
>>>>>>>
>>>>>>> Fixes: a47aa8b97afe ("app/testpmd: add vlan offload support")
>>>>>>> Cc: stable at dpdk.org
>>>>>>>
>>>>>>> Signed-off-by: Mingjin Ye <mingjinx.ye at intel.com>
>>>>>>> ---
>>>>>>>      app/test-pmd/cmdline.c | 1 +
>>>>>>>      1 file changed, 1 insertion(+)
>>>>>>>
>>>>>>> diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c index
>>>>>>> 17be2de402..ce125f549f 100644
>>>>>>> --- a/app/test-pmd/cmdline.c
>>>>>>> +++ b/app/test-pmd/cmdline.c
>>>>>>> @@ -4133,6 +4133,7 @@ cmd_vlan_offload_parsed(void
>> *parsed_result,
>>>>>>>      	else
>>>>>>>      		vlan_extend_set(port_id, on);
>>>>>>>
>>>>>>> +	cmd_reconfig_device_queue(port_id, 1, 1);
>>>> In addition, I have some comments:
>>>> 1) Normally, the parsed function of testpmd command needed to
>>>> re-config port and queue needs to check if port status is STOPED. Why
>>>> don't you add this check?
>>> The check is exist.
>> Where is the check? Currently, it seems that this check does not exist in the
>> this command parsed function.
> Check if the port is forwarded, in the source file test-pmd.c:2835.
I don't understand why you mention the check in start_port(). It should 
be done in command parsed link other command.
>>>> If the check is not exist, queue offloads are not updated until the
>>>> next port stop/start command is executed. Right?
>>> yes
>>>> 2) Why is the queue-based VLAN offload API not used?
>>> VLAN offload is a port-related configuration. If a single port is
>>> changed, the associated queue needs to be all updated in
>>> configuration. Therefore, there will be no additional api to configure.
>>>>       Like, rte_eth_dev_set_vlan_strip_on_queue. It seems that this
>>>> kind of API is
>>>>       dedicated to do this.
>>>>>> This means that queue offloads need to upadte by re-calling
>>>>>> dev_configure and setup all queues, right?
>>>>>> If it is, this adds a usage limitation.
>>>>>>>      	return;
>>>>>>>      }
>>>>>>>