On Tue, Mar 08, 2022 at 03:14:32PM +0100, David Hildenbrand wrote:
> The basic question we would like to have a reliable and efficient answer
> to is: is this anonymous page exclusive to a single process or might it
> be shared?
Is this supposed to be for PAGE_SIZE pages as well, or is it only used
on pages > PAGE_SIZE?
> In an ideal world, we'd have a spare pageflag. Unfortunately, pageflags
> don't grow on trees, so we have to get a little creative for the time
> being.
This feels a little _too_ creative to me. There's now an implicit
requirement that SL[AOU]B doesn't use the bottom two bits of
->slab_cache, which is probably OK but would need to be documented.
I have plans to get rid of PageError and PagePrivate, but those are going
to be too late for you. I don't think mappedtodisk has meaning for anon
pages, even if they're in the swapcache. It would need PG_has_hwpoisoned
to shift to another bit ... but almost any bit will do for has_hwpoisoned.
Or have I overlooked something?
> @@ -920,6 +976,70 @@ extern bool is_free_buddy_page(struct page *page);
>
> __PAGEFLAG(Isolated, isolated, PF_ANY);
>
> +static __always_inline bool folio_test_slab(struct folio *folio)
> +{
> + return !folio_test_anon(folio) &&
> + test_bit(PG_slab, folio_flags(folio, FOLIO_PF_NO_TAIL));
> +}
> +
> +static __always_inline int PageSlab(struct page *page)
> +{
> + return !PageAnon(page) &&
> + test_bit(PG_slab, &PF_NO_TAIL(page, 0)->flags);
> +}
In case we do end up using this, this would be better implemented as
static __always_inline int PageSlab(struct page *page)
{
return folio_test_slab(page_folio(page));
}
since PageAnon already has a page_folio() call embedded in it.
> +static __always_inline void __SetPageSlab(struct page *page)
> +{
> + VM_BUG_ON_PGFLAGS(PageAnon(page), page);
> + __set_bit(PG_slab, &PF_NO_TAIL(page, 1)->flags);
> +}
There's only one caller of __SetPageSlab() left, in kfence. And that
code looks ... weird.
for (i = 0; i < KFENCE_POOL_SIZE / PAGE_SIZE; i++) {
if (!i || (i % 2))
continue;
/* Verify we do not have a compound head page. */
if (WARN_ON(compound_head(&pages[i]) != &pages[i]))
goto err;
__SetPageSlab(&pages[i]);
I think the author probably intended WARN_ON(PageCompound(page)) because
they're actually verifying that it's not a tail page, rather than head
page.
> +static __always_inline void __ClearPageSlab(struct page *page)
> +{
> + VM_BUG_ON_PGFLAGS(PageAnon(page), page);
> + __clear_bit(PG_slab, &PF_NO_TAIL(page, 1)->flags);
> +}
There are no remaining callers of __ClearPageSlab(). yay.
