Re: [PATCH v4 01/25] mm: Introduce struct folio
From: Andrew Morton <akpm@linux-foundation.org>
Date: 2021-03-13 20:38:11
Also in:
linux-fsdevel, lkml
On Fri, 5 Mar 2021 04:18:37 +0000 "Matthew Wilcox (Oracle)" [off-list ref] wrote:
A struct folio refers to an entire (possibly compound) page. A function which takes a struct folio argument declares that it will operate on the entire compound page, not just PAGE_SIZE bytes. In return, the caller guarantees that the pointer it is passing does not point to a tail page. Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org> --- include/linux/mm.h | 30 ++++++++++++++++++++++++++++++ include/linux/mm_types.h | 17 +++++++++++++++++
Perhaps a new folio.h would be neater.
quoted hunk ↗ jump to hunk
@@ -1518,6 +1523,30 @@ static inline void set_page_links(struct page *page, enum zone_type zone, #endif } +static inline unsigned long folio_nr_pages(struct folio *folio) +{ + return compound_nr(&folio->page); +} + +static inline struct folio *next_folio(struct folio *folio) +{ +#if defined(CONFIG_SPARSEMEM) && !defined(CONFIG_SPARSEMEM_VMEMMAP) + return (struct folio *)nth_page(&folio->page, folio_nr_pages(folio)); +#else + return folio + folio_nr_pages(folio); +#endif +}
It's a shame this isn't called folio_something(), like the rest of the API. Unclear what this does. Some comments would help.
+static inline unsigned int folio_shift(struct folio *folio)
+{
+ return PAGE_SHIFT + folio_order(folio);
+}
+
+static inline size_t folio_size(struct folio *folio)
+{
+ return PAGE_SIZE << folio_order(folio);
+}Why size_t? That's pretty rare in this space and I'd have expected unsigned long.
quoted hunk ↗ jump to hunk
@@ -1623,6 +1652,7 @@ extern void pagefault_out_of_memory(void); #define offset_in_page(p) ((unsigned long)(p) & ~PAGE_MASK) #define offset_in_thp(page, p) ((unsigned long)(p) & (thp_size(page) - 1)) +#define offset_in_folio(folio, p) ((unsigned long)(p) & (folio_size(folio) - 1)) /* * Flags passed to show_mem() and show_free_areas() to suppress output indiff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index 0974ad501a47..a311cb48526f 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h@@ -223,6 +223,23 @@ struct page { #endif } _struct_page_alignment; +/* + * A struct folio is either a base (order-0) page or the head page of + * a compound page. + */ +struct folio { + struct page page; +}; + +static inline struct folio *page_folio(struct page *page) +{ + unsigned long head = READ_ONCE(page->compound_head); + + if (unlikely(head & 1)) + return (struct folio *)(head - 1); + return (struct folio *)page; +}
What purpose does the READ_ONCE() serve?