Re: [PATCH v1] net/i40e: remove redundant number of packets check

[Ferruh Yigit]

On 3/4/2022 1:23 AM, Zhang, Qi Z wrote:

-----Original Message-----
From: Yigit, Ferruh <ferruh.yi...@intel.com>
Sent: Friday, March 4, 2022 12:44 AM
To: Zhang, Qi Z <qi.z.zh...@intel.com>; Feifei Wang <feifei.wa...@arm.com>;
Xing, Beilei <beilei.x...@intel.com>; David Christensen
<d...@linux.vnet.ibm.com>; Richardson, Bruce <bruce.richard...@intel.com>;
Ananyev, Konstantin <konstantin.anan...@intel.com>; Ruifeng Wang
<ruifeng.w...@arm.com>
Cc: dev@dpdk.org; n...@arm.com; Honnappa Nagarahalli
<honnappa.nagaraha...@arm.com>; Thomas Monjalon
<tho...@monjalon.net>; David Marchand <david.march...@redhat.com>
Subject: Re: [PATCH v1] net/i40e: remove redundant number of packets check

On 3/3/2022 2:28 AM, Zhang, Qi Z wrote:

-----Original Message-----
From: Feifei Wang <feifei.wa...@arm.com>
Sent: Thursday, March 3, 2022 9:54 AM
To: Xing, Beilei <beilei.x...@intel.com>; David Christensen
<d...@linux.vnet.ibm.com>; Richardson, Bruce
<bruce.richard...@intel.com>; Ananyev, Konstantin
<konstantin.anan...@intel.com>; Ruifeng Wang <ruifeng.w...@arm.com>
Cc: dev@dpdk.org; n...@arm.com; Feifei Wang <feifei.wa...@arm.com>;
Honnappa Nagarahalli <honnappa.nagaraha...@arm.com>
Subject: [PATCH v1] net/i40e: remove redundant number of packets
check

For i40e_xmit_pkts_vec_xx function, it checks nb_pkts to ensure
nb_pkts does not cross rs_thresh.

However, in i40e_xmit_fixed_burst_vec_xx function, this check will be
performed again. To improve code, delete this redundant check.

Suggested-by: Honnappa Nagarahalli <honnappa.nagaraha...@arm.com>
Signed-off-by: Feifei Wang <feifei.wa...@arm.com>
Reviewed-by: Ruifeng Wang <ruifeng.w...@arm.com>

Applied to dpdk-next-net-intel.

Hi Qi,

This patch is not acked by the i40e maintainers.

And this is changing the datapath for the -rc3, two weeks before the release. Is
it tested enough?

What is the gain with this patch, I don't see any numbers in the commit log.
If the gain is small, can we postpone this patch to next release instead of 
getting
it for -rc3?

The patch applied the same thing as below which I have reviewed.
https://patchwork.dpdk.org/project/dpdk/patch/20220222224818.8612-1-kathleen.cape...@arm.com/

I didn't see the risk of having it, and I will add a "reviewed-by" to avoid 
confusion, but if you think it's risky, we can still defer it to next-net.

Code change looks simple, but risk is more theoretical
to get a datapath code optimization close to the release,
and Feifei reported that there is no visible performance
gain.
So I think safer to postpone to next release if you don't
object.