Re: [PATCH v12 3/6] mm, thp: introduce FOLL_SPLIT_PMD

[Oleg Nesterov <oleg@redhat.com>]

On 08/08, Song Liu wrote:
>
> > On Aug 8, 2019, at 9:37 AM, Oleg Nesterov <oleg@redhat.com> wrote:
> >
> > On 08/07, Song Liu wrote:
> >>
> >> @@ -399,7 +399,7 @@ static struct page *follow_pmd_mask(struct vm_area_struct *vma,
> >> 		spin_unlock(ptl);
> >> 		return follow_page_pte(vma, address, pmd, flags, &ctx->pgmap);
> >> 	}
> >> -	if (flags & FOLL_SPLIT) {
> >> +	if (flags & (FOLL_SPLIT | FOLL_SPLIT_PMD)) {
> >> 		int ret;
> >> 		page = pmd_page(*pmd);
> >> 		if (is_huge_zero_page(page)) {
> >> @@ -408,7 +408,7 @@ static struct page *follow_pmd_mask(struct vm_area_struct *vma,
> >> 			split_huge_pmd(vma, pmd, address);
> >> 			if (pmd_trans_unstable(pmd))
> >> 				ret = -EBUSY;
> >> -		} else {
> >> +		} else if (flags & FOLL_SPLIT) {
> >> 			if (unlikely(!try_get_page(page))) {
> >> 				spin_unlock(ptl);
> >> 				return ERR_PTR(-ENOMEM);
> >> @@ -420,6 +420,10 @@ static struct page *follow_pmd_mask(struct vm_area_struct *vma,
> >> 			put_page(page);
> >> 			if (pmd_none(*pmd))
> >> 				return no_page_table(vma, flags);
> >> +		} else {  /* flags & FOLL_SPLIT_PMD */
> >> +			spin_unlock(ptl);
> >> +			split_huge_pmd(vma, pmd, address);
> >> +			ret = pte_alloc(mm, pmd) ? -ENOMEM : 0;
> >> 		}
> >
> > Can't resist, let me repeat that I do not like this patch because imo
> > it complicates this code for no reason.
>
> Personally, I don't think this is more complicated than your version.

I do, but of course this is subjective.

> Also, if some code calls follow_pmd_mask() with flags contains both
> FOLL_SPLIT and FOLL_SPLIT_PMD, we should honor FOLL_SPLIT and split the
> huge page.

Heh. why not other way around?

> Of course, there is no code that sets both flags.

and of course, nobody should ever pass both FOLL_SPLIT and FOLL_SPLIT_PMD,
perhaps this deserves a warning.

Not to mention that it would be nice to kill FOLL_SPLIT which has a single
user, but this is another story.

Oleg.