Re: [PATCH v3 5/5] hugetlb: add hugetlb demote page support

[Oscar Salvador <osalvador@suse.de>]

On Fri, Oct 01, 2021 at 10:52:10AM -0700, Mike Kravetz wrote:
> Demote page functionality will split a huge page into a number of huge
> pages of a smaller size.  For example, on x86 a 1GB huge page can be
> demoted into 512 2M huge pages.  Demotion is done 'in place' by simply
> splitting the huge page.
> 
> Added '*_for_demote' wrappers for remove_hugetlb_page,
> destroy_compound_gigantic_page and prep_compound_gigantic_page for use
> by demote code.
> 
> Signed-off-by: Mike Kravetz <mike.kravetz@oracle.com>
> ---
>  mm/hugetlb.c | 82 +++++++++++++++++++++++++++++++++++++++++++++++-----
>  1 file changed, 74 insertions(+), 8 deletions(-)
> 
...  
> +static int demote_free_huge_page(struct hstate *h, struct page *page)
> +{
> +	int i, nid = page_to_nid(page);
> +	struct hstate *target_hstate;
> +	int rc = 0;
> +
> +	target_hstate = size_to_hstate(PAGE_SIZE << h->demote_order);
> +
> +	remove_hugetlb_page_for_demote(h, page, false);
> +	spin_unlock_irq(&hugetlb_lock);
> +
> +	rc = alloc_huge_page_vmemmap(h, page);
> +	if (rc) {
> +		/* Allocation of vmemmmap failed, we can not demote page */
> +		spin_lock_irq(&hugetlb_lock);
> +		set_page_refcounted(page);
> +		add_hugetlb_page(h, page, false);
> +		return rc;
> +	}

Question: You keep the original error code returned from alloc_huge_page_vmemmap()
here, but then you lose it on demote_pool_huge_page() when doing the
!demote_free_huge_page. Would not make more sense to keep it all the way down to 
demote_store() in case you want to return the actual error code?

> +
> +	/*
> +	 * Use destroy_compound_gigantic_page_for_demote for all huge page
> +	 * sizes as it will not ref count pages.
> +	 */
> +	destroy_compound_gigantic_page_for_demote(page, huge_page_order(h));

It seems that for now we only allow gigantic pages to be demoted, but
destroy_compound_gigantic_page_for_demote feels kind of wrong, even
if it is only a wrapper that ends up calling _*gigantic_ functions.

We want a routine that destroy a hugetlb to be demoted into smaller hugetlb
pages, so the name gigantic makes little sense to appear in my opinion.

>  static int demote_pool_huge_page(struct hstate *h, nodemask_t *nodes_allowed)
>  	__must_hold(&hugetlb_lock)
>  {
> +	int nr_nodes, node;
> +	struct page *page;
>  	int rc = 0;
>  
>  	lockdep_assert_held(&hugetlb_lock);
> @@ -3313,9 +3377,15 @@ static int demote_pool_huge_page(struct hstate *h, nodemask_t *nodes_allowed)
>  	if (!h->demote_order)
>  		return rc;
>  
> -	/*
> -	 * TODO - demote fucntionality will be added in subsequent patch
> -	 */
> +	for_each_node_mask_to_free(h, nr_nodes, node, nodes_allowed) {
> +		if (!list_empty(&h->hugepage_freelists[node])) {
> +			page = list_entry(h->hugepage_freelists[node].next,
> +					struct page, lru);
> +			rc = !demote_free_huge_page(h, page);

I kinda dislike this as I pointed out.


-- 
Oscar Salvador
SUSE Labs