Re: [RFC PATCH 4/7] mm: move internal core VMA manipulation functions to own file

["Liam R. Howlett" <Liam.Howlett@oracle.com>]

* Lorenzo Stoakes <lstoakes@gmail.com> [240627 15:41]:
> On Thu, Jun 27, 2024 at 01:56:12PM -0400, Liam R. Howlett wrote:
> > * Lorenzo Stoakes <lstoakes@gmail.com> [240627 06:39]:
> > > This patch introduces vma.c and moves internal core VMA manipulation
> > > functions to this file from mmap.c.
> > >
> > > This allows us to isolate VMA functionality in a single place such that we
> > > can create userspace testing code that invokes this functionality in an
> > > environment where we can implement simple unit tests of core functionality.
> > >
> > > This patch ensures that core VMA functionality is explicitly marked as such
> > > by its presence in mm/vma.h.
> > >
> > > It also places the header includes required by vma.c in vma_internal.h,
> > > which is simply imported by vma.c. This makes the VMA functionality
> > > testable, as userland testing code can simply stub out functionality
> > > as required.
> >
> > My initial thought on vma_internal.h would be to contain the number of
> > 'helper' functions and internal structures while mm/vma.h would have the
> > interface.
> 
> This is what I've done though?
> 

Yes, for the most part.  I was just surprised to see all the #includes
move.  It makes sense though as it saves a huge number of stubs/#define
of header guards.

...

Thanks,
Liam