Question about Patch Submissions

Question about Patch Submissions

[rapier]

I've read the documentation I've been able to find on the patch 
submission process. I was hoping to clarify one thing though. When 
submitting a patch that spans multiple files and functions is it better 
to break the submission down by function or by file?

Thanks for your time,

Chris

Re: Question about Patch Submissions

[Cong Wang]

On Tue, Nov 18, 2014 at 11:13 AM, rapier <rapier@psc.edu> wrote:
> I've read the documentation I've been able to find on the patch submission
> process. I was hoping to clarify one thing though. When submitting a patch
> that spans multiple files and functions is it better to break the submission
> down by function or by file?
>

It depends on what you change, if your patch is trivial, say just cleaning up
whitespaces, or a big single patch doesn't harm review, it is okay not
to split it.
Otherwise, probably you need to split it logically, each of which needs to at
least compile.

Re: Question about Patch Submissions

[rapier]

Cong,

Thanks for the response.

It's a pretty extensive patch set and we're trying to make sure it makes 
sense in context. It touches quite a few files and they're mostly 
interdependent. That's why we're having problems trying to figure out 
how to break this down for presentation. We can make sure each section 
of the patch applies cleanly but the entirety of the patch set would 
have to be applied for it to compile.

Chris

On 11/18/14, 3:21 PM, Cong Wang wrote:
> On Tue, Nov 18, 2014 at 11:13 AM, rapier <rapier@psc.edu> wrote:
>> I've read the documentation I've been able to find on the patch submission
>> process. I was hoping to clarify one thing though. When submitting a patch
>> that spans multiple files and functions is it better to break the submission
>> down by function or by file?
>>
>
> It depends on what you change, if your patch is trivial, say just cleaning up
> whitespaces, or a big single patch doesn't harm review, it is okay not
> to split it.
> Otherwise, probably you need to split it logically, each of which needs to at
> least compile.
>

Re: Question about Patch Submissions

[Alexei Starovoitov]

On Tue, Nov 18, 2014 at 12:39 PM, rapier <rapier@psc.edu> wrote:
> break this down for presentation. We can make sure each section of the patch
> applies cleanly but the entirety of the patch set would have to be applied
> for it to compile.

all patches need to compile cleanly on all architectures,
otherwise bisect will be broken.

Re: Question about Patch Submissions

[rapier]

On 11/18/14, 8:38 PM, Alexei Starovoitov wrote:
> On Tue, Nov 18, 2014 at 12:39 PM, rapier <rapier@psc.edu> wrote:
>> break this down for presentation. We can make sure each section of the patch
>> applies cleanly but the entirety of the patch set would have to be applied
>> for it to compile.
>
> all patches need to compile cleanly on all architectures,
> otherwise bisect will be broken.

The issue is that this is large patch set. The diff comes in at around 
2200 lines and adds new functionality to the kernel by implementing RFC 
4898. What I'm really trying to figure out is how to split up the diff 
files so I'm not dropping all 2200 lines into a single email message. 
I'm assuming that's not really how people want to see submissions but if 
I am incorrect please let me know.

There are two major components that we should be able to submit as 
distinct patches so that should reduce the line count per submission.

Chris Rapier