[dpdk-dev] [PATCH 3/3] app/testpmd: use of Rx/Tx in testpmd

[Ferruh Yigit]

On 3/23/2021 7:58 AM, Li, Xiaoyun wrote:
> 
> 
>> -----Original Message-----
>> From: oulijun <oulijun at huawei.com>
>> Sent: Tuesday, March 23, 2021 15:52
>> To: Tu, Lijuan <lijuan.tu at intel.com>; Andrew Rybchenko
>> <andrew.rybchenko at oktetlabs.ru>; Li, Xiaoyun <xiaoyun.li at intel.com>; Yigit,
>> Ferruh <ferruh.yigit at intel.com>
>> Cc: dev at dpdk.org; linuxarm at openeuler.org
>> Subject: Re: [dpdk-dev] [PATCH 3/3] app/testpmd: use of Rx/Tx in testpmd
>>
>>
>>
>> 在 2021/3/23 15:46, Tu, Lijuan 写道:
>>>
>>>
>>>> -----Original Message-----
>>>> From: Andrew Rybchenko <andrew.rybchenko at oktetlabs.ru>
>>>> Sent: 2021年3月23日 15:25
>>>> To: Li, Xiaoyun <xiaoyun.li at intel.com>; Lijun Ou
>>>> <oulijun at huawei.com>; Yigit, Ferruh <ferruh.yigit at intel.com>; Tu,
>>>> Lijuan <lijuan.tu at intel.com>
>>>> Cc: dev at dpdk.org; linuxarm at openeuler.org
>>>> Subject: Re: [dpdk-dev] [PATCH 3/3] app/testpmd: use of Rx/Tx in
>>>> testpmd
>>>>
>>>> On 3/23/21 6:17 AM, Li, Xiaoyun wrote:
>>>>> Hi
>>>>>
>>>>>> -----Original Message-----
>>>>>> From: Lijun Ou <oulijun at huawei.com>
>>>>>> Sent: Friday, March 5, 2021 18:22
>>>>>> To: Yigit, Ferruh <ferruh.yigit at intel.com>
>>>>>> Cc: Li, Xiaoyun <xiaoyun.li at intel.com>; dev at dpdk.org;
>>>>>> linuxarm at openeuler.org
>>>>>> Subject: [PATCH 3/3] app/testpmd: use of Rx/Tx in testpmd
>>>>>>
>>>>>> From: Hongbo Zheng <zhenghongbo3 at huawei.com>
>>>>>>
>>>>>> In testpmd, when we input "show config rxtx", we can see like this:
>>>>>>
>>>>>> 1: testpmd> show config rxtx
>>>>>> 2:   io packet forwarding packets/burst=32
>>>>>> 3:   nb forwarding cores=1 - nb forwarding ports=1
>>>>>> 4:   port 0: RX queue number: 1 Tx queue number: 1
>>>>>> 5:     Rx offloads=0x0 Tx offloads=0x10000
>>>>>> 6:     RX queue: 0
>>>>>> 7:       RX desc=1024 - RX free threshold=32
>>>>>> 8:       RX threshold registers: pthresh=0 hthresh=0  wthresh=0
>>>>>> 9:       RX Offloads=0x0
>>>>>> 10:    TX queue: 0
>>>>>> 11:      TX desc=1024 - TX free threshold=928
>>>>>> 12:      TX threshold registers: pthresh=0 hthresh=0  wthresh=0
>>>>>> 13:      TX offloads=0x10000 - TX RS bit threshold=32
>>>>>>
>>>>>> We can see RX/Rx/TX/Tx is mixed used. Also in other places in
>>>>>> testpmd, RX/Rx/TX/Tx is mixed used.
>>>>>>
>>>>>> This patch fix the mixed use of RX/Rx/TX/Tx in testpmd by change to
>>>>>> unified use Rx/Tx.
>>>>>
>>>>> The commit log is too redundant. The following is enough to explain
>>>>> what this
>>>> patch does:
>>>>> RX/TX and Rx/Tx are mixed used in testpmd print and comments. This
>>>>> patch
>>>> unifies them as Rx/Tx.
>>>>>
>>>>> Except this, the patch looks good to me.
>>>>>
>>>>> But one big concern, this patch will break all of the CI tests
>>>>> because the DTS
>>>> scripts check if the results are the same as expected and this patch
>>>> change a lot of the print.
>>>>> So I think the DTS maintainer needs to be aware of this.
>>>>
>>>> I think test dpdk-testpmd output is a part of API. Of course, it is
>>>> not an API, but such cosmetic changes in output will be much more painful
>> than API changes.
>>>> Output parsers will simply stop to work. I think such changes should
>>>> go through the deprecation process.
>>>>
>>>> Of course, it would be useful to change comments and may be even
>>>> error log messages right now, but not commands output.
>>>>
>>>>>
>>>>> +Lijuan DTS maintainer
>>>
>>> It has big impact with CI system, most of test cases depended on the testpmd
>> output, if changed, our CI system will be broken, Is it really worth it ?
>>>
>> I think it's necessary to keep the style consistent and to constrain everyone to do
>> so in the future.However, if the DTS changes greatly, I think it is possible to
>> require the new print to be consistent and the historical part to remain
>> unchanged.Because the uniform change is also Xiaoyun's suggestion, we think
>> the opinion is reasonable, but we are not sure how much it will affect DTS.
> 
> I think we should drop this patch. But we'll keep consistent for future print when reviewing.
> What do you think?
> 

+1 to have consistency on the new code, and as the old code changed by time. 
Instead of big syntax cleanup.

As well as output parsing problem, this also can cause trouble on backporting 
patches for stable trees because of conflicts it will create.

>>>>>
>>>>> And please separate this patch with the other twos. Don't mix them
>>>>> in one
>>>> patchset. It's not for the same purpose.
>>>>>
>>>>>>
>>>>>> Signed-off-by: Hongbo Zheng <zhenghongbo3 at huawei.com>
>>>>>> Signed-off-by: Lijun Ou <oulijun at huawei.com>
>>>>>> ---
>>>>>> V1->V2:
>>>>>> - fix all RX/TX
>>>>>> - rename patch title
>>>>>> ---
>>>>>>    app/test-pmd/cmdline.c     | 104 ++++++++++++++++++------------------
>>>>>>    app/test-pmd/config.c      | 128 ++++++++++++++++++++++------------------
>> ---
>>>> --
>>>>>>    app/test-pmd/csumonly.c    |  22 ++++----
>>>>>>    app/test-pmd/icmpecho.c    |   2 +-
>>>>>>    app/test-pmd/ieee1588fwd.c |  18 +++----
>>>>>> app/test-pmd/parameters.c
>>>>>> |  50 +++++++++---------
>>>>>>    app/test-pmd/testpmd.c     | 120 +++++++++++++++++++++------------------
>> ---
>>>>>>    app/test-pmd/testpmd.h     |  28 +++++-----
>>>>>>    app/test-pmd/txonly.c      |   2 +-
>>>>>>    9 files changed, 237 insertions(+), 237 deletions(-)
>>>>>> --
>>>>>> 2.7.4