I have a large patch in the 2000 lines. What to do?

I have a large patch in the 2000 lines. What to do?

[Ivan Safonov]

Hi!

I have a large patch in the 2000 lines, which replaces the macro BITn to 
BIT(n) in multiple files.
Is it worth to split this patch into several parts?
Is it sufficient to place the changes in each file in a separate part of 
patchset?

I have a large patch in the 2000 lines. What to do?

[Anish Kumar]

> On Oct 9, 2015, at 11:45 AM, Ivan Safonov <insafonov@gmail.com> wrote:
> 
> Hi!
> 
> I have a large patch in the 2000 lines, which replaces the macro BITn to 
> BIT(n) in multiple files.

I will recommend sending RFC patch first and 
Get the community comments.

If they like your patch then divide the patch based on logic rather than lines of codes.
> Is it worth to split this patch into several parts?
> Is it sufficient to place the changes in each file in a separate part of 
> patchset?
> 
> 
> 
> _______________________________________________
> Kernelnewbies mailing list
> Kernelnewbies at kernelnewbies.org
> http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies

I have a large patch in the 2000 lines. What to do?

[Greg KH]

On Fri, Oct 09, 2015 at 12:12:44PM -0700, Anish Kumar wrote:
> 
> 
> > On Oct 9, 2015, at 11:45 AM, Ivan Safonov <insafonov@gmail.com> wrote:
> > 
> > Hi!
> > 
> > I have a large patch in the 2000 lines, which replaces the macro BITn to 
> > BIT(n) in multiple files.
> 
> I will recommend sending RFC patch first and 
> Get the community comments.

No, we just ignore RFC patches, we have too many "real" patches to deal
with.

> If they like your patch then divide the patch based on logic rather than lines of codes.
> > Is it worth to split this patch into several parts?

Yes.

> > Is it sufficient to place the changes in each file in a separate part of 
> > patchset?

Make it one-patch-per-driver.

But note, cleanup patches like this are usually ignored by most
subsystems, be aware that this type of thing is only usually accepted in
the drivers/staging/ portion of the kernel.

good luck!

greg k-h

I have a large patch in the 2000 lines. What to do?

[Ivan Safonov]

On 10/10/2015 02:20 AM, Greg KH wrote:
> On Fri, Oct 09, 2015 at 12:12:44PM -0700, Anish Kumar wrote:
>>
>>> On Oct 9, 2015, at 11:45 AM, Ivan Safonov <insafonov@gmail.com> wrote:
>>>
>>> Hi!
>>>
>>> I have a large patch in the 2000 lines, which replaces the macro BITn to
>>> BIT(n) in multiple files.
>> I will recommend sending RFC patch first and
>> Get the community comments.
> No, we just ignore RFC patches, we have too many "real" patches to deal
> with.
>
>> If they like your patch then divide the patch based on logic rather than lines of codes.
>>> Is it worth to split this patch into several parts?
> Yes.
>
>>> Is it sufficient to place the changes in each file in a separate part of
>>> patchset?
> Make it one-patch-per-driver.
>
> But note, cleanup patches like this are usually ignored by most
> subsystems, be aware that this type of thing is only usually accepted in
> the drivers/staging/ portion of the kernel.
>
> good luck!
>
> greg k-h

The catch is that this patch is only to one driver.

I have a large patch in the 2000 lines. What to do?

[Greg KH]

On Sat, Oct 10, 2015 at 08:02:49AM +0700, Ivan Safonov wrote:
> On 10/10/2015 02:20 AM, Greg KH wrote:
> >On Fri, Oct 09, 2015 at 12:12:44PM -0700, Anish Kumar wrote:
> >>
> >>>On Oct 9, 2015, at 11:45 AM, Ivan Safonov <insafonov@gmail.com> wrote:
> >>>
> >>>Hi!
> >>>
> >>>I have a large patch in the 2000 lines, which replaces the macro BITn to
> >>>BIT(n) in multiple files.
> >>I will recommend sending RFC patch first and
> >>Get the community comments.
> >No, we just ignore RFC patches, we have too many "real" patches to deal
> >with.
> >
> >>If they like your patch then divide the patch based on logic rather than lines of codes.
> >>>Is it worth to split this patch into several parts?
> >Yes.
> >
> >>>Is it sufficient to place the changes in each file in a separate part of
> >>>patchset?
> >Make it one-patch-per-driver.
> >
> >But note, cleanup patches like this are usually ignored by most
> >subsystems, be aware that this type of thing is only usually accepted in
> >the drivers/staging/ portion of the kernel.
> >
> >good luck!
> >
> >greg k-h
> 
> The catch is that this patch is only to one driver.

You know, providing specifics is always nice when asking questions...

If it's just one driver, and you think the patch is correct, just send
it in as normal, in one patch, as it should be "easy" to understand.

But again, remember that these types of "cleanups" are not always
appreciated for some portions of the kernel (again, you were not
specific so I can't be specific back...)

greg k-h

I have a large patch in the 2000 lines. What to do?

[Julio Faracco]

I believe that it was answered.

What is the recommendation when I change a structure in a include file
and this change causes a lots of new changes, for example?

Julio Cesar Faracco

2015-10-10 11:14 GMT-03:00 Greg KH <greg@kroah.com>:
> On Sat, Oct 10, 2015 at 08:02:49AM +0700, Ivan Safonov wrote:
>> On 10/10/2015 02:20 AM, Greg KH wrote:
>> >On Fri, Oct 09, 2015 at 12:12:44PM -0700, Anish Kumar wrote:
>> >>
>> >>>On Oct 9, 2015, at 11:45 AM, Ivan Safonov <insafonov@gmail.com> wrote:
>> >>>
>> >>>Hi!
>> >>>
>> >>>I have a large patch in the 2000 lines, which replaces the macro BITn to
>> >>>BIT(n) in multiple files.
>> >>I will recommend sending RFC patch first and
>> >>Get the community comments.
>> >No, we just ignore RFC patches, we have too many "real" patches to deal
>> >with.
>> >
>> >>If they like your patch then divide the patch based on logic rather than lines of codes.
>> >>>Is it worth to split this patch into several parts?
>> >Yes.
>> >
>> >>>Is it sufficient to place the changes in each file in a separate part of
>> >>>patchset?
>> >Make it one-patch-per-driver.
>> >
>> >But note, cleanup patches like this are usually ignored by most
>> >subsystems, be aware that this type of thing is only usually accepted in
>> >the drivers/staging/ portion of the kernel.
>> >
>> >good luck!
>> >
>> >greg k-h
>>
>> The catch is that this patch is only to one driver.
>
> You know, providing specifics is always nice when asking questions...
>
> If it's just one driver, and you think the patch is correct, just send
> it in as normal, in one patch, as it should be "easy" to understand.
>
> But again, remember that these types of "cleanups" are not always
> appreciated for some portions of the kernel (again, you were not
> specific so I can't be specific back...)
>
> greg k-h
>
> _______________________________________________
> Kernelnewbies mailing list
> Kernelnewbies at kernelnewbies.org
> http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies

I have a large patch in the 2000 lines. What to do?

[Greg KH]

On Fri, Oct 16, 2015 at 06:14:01PM -0300, Julio Faracco wrote:
> 
> What is the recommendation when I change a structure in a include file
> and this change causes a lots of new changes, for example?

You do the change in a way that does not cause build failures and spread
the patches out over all subsystems to fix things up.

Do you have a specific example you were thinking of so that I can show
how it can be done in this way?

thanks,

greg k-h