Re: [External] Re: [PATCH v9 09/11] mm/hugetlb: Introduce nr_free_vmemmap_pages in the struct hstate

From: Muchun Song
Date: Wed Dec 16 2020 - 08:58:05 EST


On Wed, Dec 16, 2020 at 9:44 PM Oscar Salvador <osalvador@xxxxxxx> wrote:
>
> On Sun, Dec 13, 2020 at 11:45:32PM +0800, Muchun Song wrote:
> > All the infrastructure is ready, so we introduce nr_free_vmemmap_pages
> > field in the hstate to indicate how many vmemmap pages associated with
> > a HugeTLB page that we can free to buddy allocator. And initialize it
> "can be freed to buddy allocator"
>
> > in the hugetlb_vmemmap_init(). This patch is actual enablement of the
> > feature.
> >
> > Signed-off-by: Muchun Song <songmuchun@xxxxxxxxxxxxx>
> > Acked-by: Mike Kravetz <mike.kravetz@xxxxxxxxxx>
>
> With below nits addressed you can add:
>
> Reviewed-by: Oscar Salvador <osalvador@xxxxxxx>

Thanks.

>
> > static int __init early_hugetlb_free_vmemmap_param(char *buf)
> > {
> > + /* We cannot optimize if a "struct page" crosses page boundaries. */
> > + if (!is_power_of_2(sizeof(struct page)))
> > + return 0;
> > +
>
> I wonder if we should report a warning in case someone wants to enable this
> feature and stuct page size it not power of 2.
> In case someone wonders why it does not work for him/her.
>
> > +void __init hugetlb_vmemmap_init(struct hstate *h)
> > +{
> > + unsigned int nr_pages = pages_per_huge_page(h);
> > + unsigned int vmemmap_pages;
> > +
> > + if (!hugetlb_free_vmemmap_enabled)
> > + return;
> > +
> > + vmemmap_pages = (nr_pages * sizeof(struct page)) >> PAGE_SHIFT;
> > + /*
> > + * The head page and the first tail page are not to be freed to buddy
> > + * system, the others page will map to the first tail page. So there
> > + * are the remaining pages that can be freed.
> "the other pages will map to the first tail page, so they can be freed."
> > + *
> > + * Could RESERVE_VMEMMAP_NR be greater than @vmemmap_pages? It is true
> > + * on some architectures (e.g. aarch64). See Documentation/arm64/
> > + * hugetlbpage.rst for more details.
> > + */
> > + if (likely(vmemmap_pages > RESERVE_VMEMMAP_NR))
> > + h->nr_free_vmemmap_pages = vmemmap_pages - RESERVE_VMEMMAP_NR;
> > +
> > + pr_info("can free %d vmemmap pages for %s\n", h->nr_free_vmemmap_pages,
> > + h->name);
>
> Maybe specify this is hugetlb code:
>
> pr_info("%s: blabla", __func__, ...)
> or
> pr_info("hugetlb: blalala", ...);
>
> although I am not sure whether we need that at all, or maybe just use
> pr_debug().

The pr_info can tell the user whether the feature is enabled. From this
point of view, it makes sense. Right?

Thanks.

>
> --
> Oscar Salvador
> SUSE L3



--
Yours,
Muchun