Re: [RFC PATCH 03/57] mm/collapse: add collapse.h for the shared collapse state

On Wed, Aug 19, 2026 at 04:19:41PM +0200, David Hildenbrand (Arm) wrote:
> On 8/17/26 00:45, Kiryl Shutsemau wrote:
> > From: "Kiryl Shutsemau (Meta)" <[email protected]>
> > 
> > Preparation for building the new collapse engine in its own file.  The
> > engine and khugepaged.c need to agree on what a collapse result is and
> > what state a scan carries.
> > 
> > Move enum scan_result and struct collapse_control into a new
> > mm/collapse.h.
> > 
> > No functional change intended.
> > 
> > Assisted-by: Claude-Code:claude-opus-5
> > Signed-off-by: Kiryl Shutsemau (Meta) <[email protected]>
> > ---
> 
> BTW, I'm wondering whether it would result in less churn (independent of the
> other changes) to
> 
> a) Rename mm/khugeapged.c to mm/collapse.c
> 
> b) Move stuff into mm/collapse.h that a new ...
> 
> c) mm/khugapegd.c will need.

I went for a clean collapse.c on purpose: the engine is new code, and a
new file keeps it readable while both mechanisms are in the tree, with 28
as a plain delete from khugepaged.c.

Your shape is less churn, though, and it keeps blame: 29-35 move about
1700 lines as add-and-delete today, and most of that would stay put, with
only the daemon -- mm_slot, the scan loop, sysfs -- extracted into a new
khugepaged.c.

If you prefer it that way, I can give it a try for v2.

The rename has to open the series, so it will collide with anything in
flight that touches khugepaged.c.

-- 
  Kiryl Shutsemau / Kirill A. Shutemov