On 1/13/22 11:53, Van Haaren, Harry wrote:
>> -----Original Message-----
>> From: Ilya Maximets <[email protected]>
>> Sent: Wednesday, January 12, 2022 6:01 PM
>> To: Van Haaren, Harry <[email protected]>; Finn, Emma
>> <[email protected]>; [email protected]; Amber, Kumar
>> <[email protected]>
>> Cc: [email protected]; Stokes, Ian <[email protected]>; Flavio Leitner
>> <[email protected]>; Kevin Traynor <[email protected]>
>> Subject: Re: [PATCH v4 0/9] Actions Infrastructure + Optimizations
>>
>> On 1/6/22 14:11, Van Haaren, Harry wrote:
>>>> -----Original Message-----
>>>> From: Finn, Emma <[email protected]>
>>>> Sent: Wednesday, January 5, 2022 4:54 PM
>>>> To: [email protected]; Van Haaren, Harry <[email protected]>;
>>>> Amber, Kumar <[email protected]>
>>>> Cc: Finn, Emma <[email protected]>
>>>> Subject: [PATCH v4 0/9] Actions Infrastructure + Optimizations
>>>>
>>>> ---
>>>> v4:
>>>> - Rebase to master
>>>> - Add ISA implementation of push_vlan action
>>>
>>> Thanks for the updated patchset Emma & Amber.
>>>
>>> Overall, this is working as expected and I've only had some minor
>>> comments throughout the patchset. I've added my Acked-by to most
>>> patches, some small open questions remain to be addressed in a v5.
>>>
>>> +CC Ian/Ilya , I'd like to see the v5 get merged, so let's continue to work
>> towards that.
>>
>> Hi, Harry, Ian, others.
> 
> Hi Ilya,
> 
>> Following up from a brief conversation during today's upstream meeting.
>> It was brought to my attention that you're expecting this series and
>> the 'hash' one to be accepted into 2.17. Though there are few issues
>> with that:
>>
>> 1. This review for v4 was actually very first review of the patch set.
>>    The other one as of today doesn't have any reviews at all.
>>    Looking at the change log for this patch set it doesn't seem that
>>    internal reviews behind the closed doors (if there were any) requested
>>    any significant changes.  In any case, internal reviews is not the way
>>    how open-source projects work.
> 
> Actions & MFEX  were developed internally yes, and hence internal reviews and
> architecture was iterated on. Saying "not many large changes requested" is 
> not relevant,
> it means that internally the architecture was well aligned. If anything, it 
> means that
> reviewers did not have big concerns, hence we should have better confidence 
> to merge.

Me, who never seen these reviews and architecture discussions has exactly
zero confidence in these patch-set the same as the rest of the community.
Development behind closed doors is never taken into account while making
decisions in open-source projects, because this development and made
decisions are not visible and can not be peer-reviewed by anyone outside.

On a quick glance over the code and reviews done in a last couple of days,
I'd also question the quality of these public reviews as the patch set
at least contains a few issues in a copy-pasted code that was already
fixed on master branch (left alone that these code duplications should,
likley, not exist in a first place).

> 
> 
>> 2. The soft freeze for 2.17 began on Jan 3 in accordance with our
>>    release schedule (even a bit later), and as you know, during the soft
>>    freeze we're not normally accepting patches that wasn't already reviewed
>>    before the soft freeze begun.
>>    https://mail.openvswitch.org/pipermail/ovs-dev/2022-January/390487.html
> 
> Actions and MFEX Hashing were discussed and reviewed in public at OVS Conf;
> https://www.openvswitch.org/support/ovscon2021/#T32
> https://www.openvswitch.org/support/ovscon2021/#T33

They were presented, but not discussed or reviewed.

> 
> The closing "call to action" slide clearly states 2.17 upstream is intended,
> and welcome community review & comments, no concerns were raised.
> 
>> That's not the end of a world, but you need to request an exception in
>> reply to the email linked above.
> 
> As both the patches and intent to merge for OVS 2.17 were clearly discussed
> in public at OVS Conf, this "request exception" case does not apply.

Again, they were presented, not discussed.  It's the same as sending patches
to the mailing list.  We don't have a policy to accept everything that
wasn't explicitly rejected.  Lack of feedback / responses in most cases means
lack of interest. 

> 
> 
>> But I have a few high-level concerns regarding the patch set itself,
>> and that's a bigger problem for me:
>>
>> 1. What are the benefits of these patch sets?  A lot of infrastructure
>>    changes are made, but the benefits of them are unclear.  Why these
>>    changes are needed in the end?  I believe, that was the main reason
>>    why community had no interest in reviewing these patches.
>>    2.17 is supposed to be a new LTS, so infrastructure changes without
>>    clear benefits might not be a good fit taking into account time
>>    constraints and lack of reviews.
> 
> Customers workloads are accelerated, improving OVS datapath performance for 
> their
> workloads. The customers are not active in SW development themselves, hence 
> there
> is no reviews from them on OVS ML.

I don't see anything being accelerated.  I'm not asking your customers to
prove anything, I'm asking you and your team to provide a reason to review
the patches.  "Workloads are accelerated" is a blunt statement without
any numbers.  And even that you're saying for a very first time.   For what I
know it could be 0.00001% improvement that simply doesn't worth it.  I don't
see any claims about performance regression testing being made either.

> 
> If you have concerns over a patchset, these should be raised against a V1 of 
> a patchset,

I don't have time to look at every patch-set that doesn't even try to make
me or anyone else interested in reviewing it and you're re-spinning the
series without any reviews being made, so that is simply not practical.

> in appropriate time to be addressed. Asking these questions the week of the 
> integration
> deadline is unacceptable, and may not be used to block merge a patchset.

Again, there is no policy to accept everything that wasn't explicitly rejected.
You missed the review deadline for a soft freeze.  So, these patches was
not considered for integration at all at this point.  Hence questioning them
is justified.

> 
> 
>> 2. The same concern that I already brought to you attention in other
>>    conversations, e.g. on the ovs-security list about a month ago.
>>    It's related to all developments in that area: why this is tied to
>>    the userspace datapath?  i.e. why execution of actions depends on the
>>    datapath?  This seems artificial and complicates testing a lot.
>>    Like current autovalidator is not able to test most of the packet
>>    parsing cases, the same way autovalidator will not be able to test
>>    execution of actions.
> 
> Regarding testing, as per that same conversation, I replied that yes we would 
> investigate
> improving testing even further. There is ongoing work here around improving 
> our fuzz testing.

Didn't see any progress from your side on that front.  And still waiting for
your reply on this email:
  https://mail.openvswitch.org/pipermail/ovs-dev/2021-December/389945.html

> 
> Autovalidation is not complex, it validates all unit-tests, and has been 
> presented at OVS

'all unit tests' is a plain lie here.  You're perfectly aware that this is
not true at least for miniflow_extract part.

> conferences, and upstreamed for multiple components (DPCLS, MFEX, and now 
> Actions).
> If there is concern over approach, then this should not be held against 
> Actions patchset
> now in the week of merge, but addressed in general.

AFAICT, you're not planning to address concerns in general, because the issue
with miniflow_extract autovalidation not being usable for a big chunk of 
existing
tests was flagged several times including the email mentioned above, and you're
always avoiding to reply to these concerns.

> Not acceptable to block Actions
> or Hashing patchsets based on this, in the week of the integration deadline.

Continuing to re-use and build new things on top of broken abstractions
doesn't look good to me.

> 
> 
>> I have some more comments, but they are more related to the actual code
>> and above 2 are the most important for now.
>>
>> Best regards, Ilya Maximets.
> 
> Regards, -Harry
> 

_______________________________________________
dev mailing list
[email protected]
https://mail.openvswitch.org/mailman/listinfo/ovs-dev

Reply via email to