Hi Rao -
On Wed, 21 Mar 2018, Rao Shoaib wrote:
There has been some misunderstanding on what an RFC patch is and what standards it needs
to meet. Following is the official response of David
Miller.
Based on those guidelines the RFC patch we submitted meets and exceeds the requirements.
So please review it. Replacement of function pointers
will be addressed later as a separate issue.
Given what David has said, I still think it's better to structure the
patch set according to the guidelines I pointed to before. Required? No,
but I think it's worth the effort. You might not agree, and that's ok.
I know at least one other community member is taking a look at the first
version of your patches. You mentioned that you are reworking your changes
to invoke MPTCP functions using macros instead of indirect calls, so my
personal plan is to take a closer look after you post v2.
In case there are other procedural issues that would prohibit technical discussion please
site the written rules or first ask David Miller.
There are a lot of technical issues to discuss, so if possible lets not get tied up on
process and bureaucracy. Looking forward to detailed
technical comments.
In my experience, a lot of the work involved in getting patches merged has
been the unglamourous stuff to get the patches set up for review and
upstreaming. My goal has been to help others and pass on that knowledge.
But, to be clear, I'm not waiting for review based on the patch set
structural changes, I want to look at everything in context with the
macro/indirect call changes.
Thanks,
Mat
-------- Forwarded Message --------
Subject:
Re: Few questions about submitting patches
Date:
Wed, 21 Mar 2018 12:46:17 -0400 (EDT)
From:
David Miller <davem(a)davemloft.net>
To:
rao.shoaib(a)oracle.com
CC:
netdev(a)vger.kernel.org, eric.dumazet(a)gmail.com
From: Rao Shoaib <rao.shoaib(a)oracle.com>
Date: Wed, 21 Mar 2018 09:41:13 -0700
> I am new to Linux. I would like to understand the rules and etiquettes
> of engaging with the community. I have read the materials that I could
> find. As I work with Linux I come across situations for which I can
> not seem to find any answers. Hopefully folks on the list can answer
> them.
>
> * Submitting an RFC Patch
>
> As I understand, an RFC patch is submitted to solicit comments and is
> not for inclusion. Is it sufficient for an RFC patch to have the
> correct coding style and compile, or does it need more? For example,
> If the patch consists of a series of patches, does each patch have to
> compile independently etc etc.
It should build and function, unless you explicitly state that the
patch is not build nor functionally tested and is intended to show
the design of the change.
> * #ifdef FOO
>
> In a regular patch consisting of a series of patches, can the above
> #ifdef be used in a patch before the patch that allows the selection
> of FOO. That patch is part of the series but comes later.
It is better to introduce them at the same time.
But if it is prohibitively difficult to do so, yet at the same
time properly split up your changes into manageable pieces, it
can be OK.
It is definitely determined on a case by case basis.
--
Mat Martineau
Intel OTC