DPDK patches and discussions
 help / color / mirror / Atom feed
From: Ferruh Yigit <ferruh.yigit@intel.com>
To: Matan Azrad <matan@mellanox.com>,
	Wenzhuo Lu <wenzhuo.lu@intel.com>,
	Jingjing Wu <jingjing.wu@intel.com>
Cc: "dev@dpdk.org" <dev@dpdk.org>, "stable@dpdk.org" <stable@dpdk.org>
Subject: Re: [dpdk-dev] [dpdk-stable] [PATCH 1/2] app/testpmd: fix scatter offload configuration
Date: Tue, 30 Jul 2019 19:55:49 +0100	[thread overview]
Message-ID: <dd4f22c6-a954-ee65-b825-2e95227cf9e9@intel.com> (raw)
In-Reply-To: <AM0PR0502MB4019BD72AE373B07FD760682D2DC0@AM0PR0502MB4019.eurprd05.prod.outlook.com>

On 7/30/2019 7:34 PM, Matan Azrad wrote:
> 
> 
> From: Ferruh Yigit 
>> On 7/30/2019 4:56 PM, Matan Azrad wrote:
>>> Hi Ferruh
>>>
>>>  From: Ferruh Yigit
>>>> Sent: Tuesday, July 30, 2019 6:22 PM
>>>> To: Matan Azrad <matan@mellanox.com>; Wenzhuo Lu
>>>> <wenzhuo.lu@intel.com>; Jingjing Wu <jingjing.wu@intel.com>
>>>> Cc: dev@dpdk.org; stable@dpdk.org
>>>> Subject: Re: [dpdk-dev] [PATCH 1/2] app/testpmd: fix scatter offload
>>>> configuration
>>>>
>>>> On 7/30/2019 2:17 PM, Matan Azrad wrote:
>>>>> Hi Ferruh
>>>>>
>>>>> From: Ferruh Yigit
>>>>>> Sent: Tuesday, July 30, 2019 4:09 PM
>>>>>> To: Matan Azrad <matan@mellanox.com>; Wenzhuo Lu
>>>>>> <wenzhuo.lu@intel.com>; Jingjing Wu <jingjing.wu@intel.com>
>>>>>> Cc: dev@dpdk.org; stable@dpdk.org
>>>>>> Subject: Re: [dpdk-dev] [PATCH 1/2] app/testpmd: fix scatter
>>>>>> offload configuration
>>>>>>
>>>>>> On 7/29/2019 1:36 PM, Matan Azrad wrote:
>>>>>>> When the mbuf data size cannot contain the maximum Rx packet
>>>>>>> length with the mbuf headroom, a packet should be scattered in
>>>>>>> more than one
>>>>>> mbuf.
>>>>>>>
>>>>>>> The application did not configure scatter offload in the above case.
>>>>>>>
>>>>>>> Enable the Rx scatter offload in the above case.
>>>>>>>
>>>>>>> Fixes: 33f9630fc23d ("app/testpmd: create mbuf based on max
>>>>>>> supported
>>>>>>> segments")
>>>>>>> Cc: stable@dpdk.org
>>>>>>>
>>>>>>> Signed-off-by: Matan Azrad <matan@mellanox.com>
>>>>>>
>>>>>> Deferring the patchset to next release, they were late anyway and
>>>>>> not actually fixing a defect, safer to defer than getting them in rc3.
>>>>>
>>>>> Yes this patch came late for RC3 but it is a fix.
>>>>>
>>>>> What are you concerns here?
>>>>> Why don't you think defect found?
>>>>
>>>> First patch changes the behavior, when mbuf size is larger than
>>>> configured size and user didn't provided the scatter offload, should
>>>> test application automatically enable it?
>>>
>>> No, only when the mbuf size is smaller than the max_rx_pkt_len with
>> headroom.
>>> If scatter is not enabled in the above case, how can the PMD provide a
>> packet with max_rx_pkt_len size?
>>>
> 
> Answer here?

Is it because drivers also "automatically" enable scattered Rx based on other
values? - which is also open to discussion I think.

> 
>>> I think not enabling scatter in this case it is a user conflict in configuration
>> and should raise an error in the PMD.  Maybe even in ethdev layer.
>>>
>>>> It may or not, but this is the change of the behavior, I think not a
>>>> fix.
>>>>
>>>> And second patch adds more detail into the statistics, so I believe
>>>> it is clear that it is not a fix.
>>>
>>>  Agree, this can wait.
>>>
>>>> The concern is getting changes very close to release, to balance
>>>> between risk and benefit of the feature. I don't see any reason why
>>>> these changes can't wait next release, so I don't see any reason to get the
>> risk.
>>>
>>> When  I changed the default max_rx_pkt_len and mbuf size in LRO testing I
>> met this issue.
>>>
>>> By default scatter will not be enabled.
>>
>> I think it is still arguable if scatter should be enabled by default,
> 
> I meant that with this patch it will not be enabled by default due to the default values of mbuf size and max_rx_pkt_len.

I mean the same thing indeed, still I believe arguable.

> 
>> but isn't there a way in testpmd to enable scatter explicitly? If so you have a way to test LRO.
> 
> Yes there is a way.
> 
> This patch is just the right way to do it.
> 

Good to know it is not blocking anyone, patch can be reviewed by its maintainers
and discussed more.

  reply	other threads:[~2019-07-30 18:55 UTC|newest]

Thread overview: 16+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-07-29 12:36 [dpdk-dev] " Matan Azrad
2019-07-29 12:36 ` [dpdk-dev] [PATCH 2/2] app/testpmd: add bits per second to statistics Matan Azrad
2019-07-30 11:41   ` Moti Haimovsky
2019-10-08 14:19     ` Yigit, Ferruh
2019-07-30  9:00 ` [dpdk-dev] [PATCH 1/2] app/testpmd: fix scatter offload configuration Matan Azrad
2019-07-30 11:36   ` Moti Haimovsky
2019-07-30 13:09 ` Ferruh Yigit
2019-07-30 13:17   ` Matan Azrad
2019-07-30 15:21     ` Ferruh Yigit
2019-07-30 15:56       ` Matan Azrad
2019-07-30 17:28         ` Ferruh Yigit
2019-07-30 18:34           ` Matan Azrad
2019-07-30 18:55             ` Ferruh Yigit [this message]
2019-07-31  6:11               ` [dpdk-dev] [dpdk-stable] " Matan Azrad
2019-10-08 14:18                 ` Yigit, Ferruh
2019-10-22  7:06                   ` Matan Azrad

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=dd4f22c6-a954-ee65-b825-2e95227cf9e9@intel.com \
    --to=ferruh.yigit@intel.com \
    --cc=dev@dpdk.org \
    --cc=jingjing.wu@intel.com \
    --cc=matan@mellanox.com \
    --cc=stable@dpdk.org \
    --cc=wenzhuo.lu@intel.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).