Re: [PATCH V3 5/5] mm/khugepaged: make khugepaged_collapse_control static

[Lorenzo Stoakes <lorenzo.stoakes@oracle.com>]

On Fri, Jan 23, 2026 at 07:02:13PM -0800, Andrew Morton wrote:
> On Fri, 23 Jan 2026 17:21:42 -0800 Andrew Morton <akpm@linux-foundation.org> wrote:
>
> >
> > lol, OK, I droppped $Subject and did this:
> >
> > --- a/mm/khugepaged.c~mm-khugepaged-make-khugepaged_collapse_control-a-local
> > +++ a/mm/khugepaged.c
> > @@ -827,10 +827,6 @@ static void khugepaged_alloc_sleep(void)
> >  	remove_wait_queue(&khugepaged_wait, &wait);
> >  }
> >
> > -struct collapse_control khugepaged_collapse_control = {
> > -	.is_khugepaged = true,
> > -};
> > -
> >  static bool hpage_collapse_scan_abort(int nid, struct collapse_control *cc)
> >  {
> >  	int i;
> > @@ -2618,13 +2614,16 @@ static void khugepaged_wait_work(void)
> >
> >  static int khugepaged(void *none)
> >  {
> > +	struct collapse_control cc = {
> > +		.is_khugepaged = true,
> > +	};
>
> nope, that thing's really big!
>
> mm/khugepaged.c: In function 'khugepaged':
> mm/khugepaged.c:2637:1: error: the frame size of 4696 bytes is larger than 2048

Yup, thanks Andrew, this change is completely crazy.

>
> We could kzalloc() it but I think I'll just restore the make-it-static
> patch.  Someone please add to the todo list?

And no to kzalloc(), let's just keep this static for now. It works fine as it is!

Thanks, Lorenzo