Re: [RFC PATCHv2 1/2] Add mm flag to control THP

[Alex Thorlton <athorlton@sgi.com>]

On Sun, Jan 19, 2014 at 01:41:34AM +0200, Kirill A. Shutemov wrote:
> On Thu, Jan 16, 2014 at 03:01:43PM -0600, Alex Thorlton wrote:
> > diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h
> > index 91672e2..475f59f 100644
> > --- a/include/linux/huge_mm.h
> > +++ b/include/linux/huge_mm.h
> > @@ -1,6 +1,8 @@
> >  #ifndef _LINUX_HUGE_MM_H
> >  #define _LINUX_HUGE_MM_H
> >  
> > +#include <linux/sched.h>
> > +
> 
> Hm, now <linux/mm.h> depends on <linux/sched.h>. It doesn't look as a good
> idea.
> 
> Why do we have MMF_* defines in <linux/sched.h>?
> Wouldn't it more appropriate to put them in <linux/mm_types.h>?

I'm not sure about that.  I just placed my definitions where everything
else was defined.  It might make more sense to have these in
mm_types.h, though that's a separate effort.  For now, I don't think I
have a choice but to include sched.h here, but it does seem to make
sense that I should be able to see the defined mm flags from huge_mm.h,
without having to include sched.h.  I'm sure there are other places where
this applies as well.

> >  extern int do_huge_pmd_anonymous_page(struct mm_struct *mm,
> >  				      struct vm_area_struct *vma,
> >  				      unsigned long address, pmd_t *pmd,
> > @@ -74,7 +76,8 @@ extern bool is_vma_temporary_stack(struct vm_area_struct *vma);
> >  	   (1<<TRANSPARENT_HUGEPAGE_REQ_MADV_FLAG) &&			\
> >  	   ((__vma)->vm_flags & VM_HUGEPAGE))) &&			\
> >  	 !((__vma)->vm_flags & VM_NOHUGEPAGE) &&			\
> > -	 !is_vma_temporary_stack(__vma))
> > +	 !is_vma_temporary_stack(__vma) &&				\
> > +	 !test_bit(MMF_THP_DISABLE, &(__vma)->vm_mm->flags))
> >  #define transparent_hugepage_defrag(__vma)				\
> >  	((transparent_hugepage_flags &					\
> >  	  (1<<TRANSPARENT_HUGEPAGE_DEFRAG_FLAG)) ||			\
> > @@ -227,7 +230,6 @@ static inline int do_huge_pmd_numa_page(struct mm_struct *mm, struct vm_area_str
> >  {
> >  	return 0;
> >  }
> > -
> >  #endif /* CONFIG_TRANSPARENT_HUGEPAGE */
> >  
> >  #endif /* _LINUX_HUGE_MM_H */
> 
> Why?

Are you asking about the line I removed there?  That was probably an
accident.  I'll fix that

- Alex