[PATCH v6] net/i40e: rework maximum frame size configuration

[Kevin Traynor]

On 22/03/2023 16:50, Kevin Traynor wrote:
> On 27/02/2023 00:35, Zhang, Qi Z wrote:
>>
> 
> Hi Simei/Qi/Yu
> 

Hi Yu,

>>
>>> -----Original Message-----
>>> From: Su, Simei <simei.su at intel.com>
>>> Sent: Monday, February 20, 2023 4:00 PM
>>> To: Xing, Beilei <beilei.xing at intel.com>; Zhang, Yuying
>>> <yuying.zhang at intel.com>; Zhang, Qi Z <qi.z.zhang at intel.com>;
>>> david.marchand at redhat.com
>>> Cc: dev at dpdk.org; Yang, Qiming <qiming.yang at intel.com>; Su, Simei
>>> <simei.su at intel.com>; stable at dpdk.org
>>> Subject: [PATCH v6] net/i40e: rework maximum frame size configuration
>>>
>>> One issue is reported by David Marchand that error occurs in OVS due to the
>>> fix patch in mentioned changes below. The detailed reproduce step and
>>> result are in https://patchwork.dpdk.org/project/dpdk/patch/
>>> 20211207085946.121032-1-dapengx.yu at intel.com/.
>>>
>>> This patch removes unnecessary link status check and directly sets mac config
>>> in dev_start. Also, it sets the parameter "wait to complete" true to wait for
>>> more time to make sure adminq command execute completed.
>>>
> 
>>> Fixes: a4ba77367923 ("net/i40e: enable maximum frame size at port level")
>>> Fixes: 2184f7cdeeaa ("net/i40e: fix max frame size config at port level")
>>> Fixes: 719469f13b11 ("net/i40e: fix jumbo frame Rx with X722")
>>> Cc: stable at dpdk.org
> 
> These patches caused an observable regression in multiple 20.11 and
> 21.11 LTS releases that was only caught a long time after releases.
> 
> Is there anything being added to LTS validation for regression testing
> this issue, so we don't get caught again?
> 

This is the issue I was talking about earlier during the release 
meeting. Not sure if we were talking about the same patch.

I was asking if there are some regression tests added/can be added to 
LTS validation that will be run during each LTS validation cycle so we 
don't have any more regressions on it.

thanks,
Kevin.

> After reverting the original patch and 2 fixes, I'm a bit reluctant to
> take more fixes without some form of regression testing in place.
> 
> thanks,
> Kevin.
> 
>>>
>>> Reported-by: David Marchand <david.marchand at redhat.com>
>>> Signed-off-by: Simei Su <simei.su at intel.com>
>>
>> Acked-by: Qi Zhang <qi.z.zhang at intel.com>
>>
>> Applied to dpdk-next-net-intel.
>>
>> Thanks
>> Qi
>>
>