Shrinking struct page progress

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



At the bottom of this email is a patch which is all it currently takes
to make an allnoconfig build without page->index and page->mapping.
Assuming you first apply
https://lore.kernel.org/linux-mm/20231108182809.602073-1-willy@xxxxxxxxxxxxx/

diffstat:
 14 files changed, 27 insertions(+), 120 deletions(-)

There are a few "cheats" in it.  The obvious one is, of course,
allnoconfig.  That lets me handwave over every filesystem.  Progress is
being made in that area, but there's a lot left to do.

Similarly, using allnoconfig lets me delete some of the helper functions
which still use page->index and page->mapping like page_index().
That one's down to just ten callers, so we are making progress (we've
removed nine in the last year).

So, what do I think this shows?

First, there's something odd going on with perf_mmap_fault().  Why does
it need to set ->mapping and ->index?

Second, percpu needs to stop abusing page->index ;-)

Third, it's getting to be time to turn the page allocator into a
folio allocator.  I'll have some patches to move us in that direction
soon.

diff --git a/include/linux/migrate.h b/include/linux/migrate.h
index 2ce13e8a309b..2e8af7bd483d 100644
--- a/include/linux/migrate.h
+++ b/include/linux/migrate.h
@@ -138,7 +138,7 @@ const struct movable_operations *page_movable_ops(struct page *page)
 	VM_BUG_ON(!__PageMovable(page));
 
 	return (const struct movable_operations *)
-		((unsigned long)page->mapping - PAGE_MAPPING_MOVABLE);
+		((unsigned long)page->___mapping - PAGE_MAPPING_MOVABLE);
 }
 
 #ifdef CONFIG_NUMA_BALANCING
diff --git a/include/linux/mm.h b/include/linux/mm.h
index 418d26608ece..36fcd4421fb9 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -2216,17 +2216,6 @@ static inline void *folio_address(const struct folio *folio)
 
 extern pgoff_t __page_file_index(struct page *page);
 
-/*
- * Return the pagecache index of the passed page.  Regular pagecache pages
- * use ->index whereas swapcache pages use swp_offset(->private)
- */
-static inline pgoff_t page_index(struct page *page)
-{
-	if (unlikely(PageSwapCache(page)))
-		return __page_file_index(page);
-	return page->index;
-}
-
 /*
  * Return true only if the page has been allocated with
  * ALLOC_NO_WATERMARKS and the low watermark was not
diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h
index 957ce38768b2..bc490ba9362f 100644
--- a/include/linux/mm_types.h
+++ b/include/linux/mm_types.h
@@ -103,9 +103,9 @@ struct page {
 				struct list_head pcp_list;
 			};
 			/* See page-flags.h for PAGE_MAPPING_FLAGS */
-			struct address_space *mapping;
+			struct address_space *___mapping;
 			union {
-				pgoff_t index;		/* Our offset within mapping. */
+				pgoff_t ___index;		/* Our offset within mapping. */
 				unsigned long share;	/* share count for fsdax */
 			};
 			/**
@@ -361,9 +361,9 @@ struct folio {
 	static_assert(offsetof(struct page, pg) == offsetof(struct folio, fl))
 FOLIO_MATCH(flags, flags);
 FOLIO_MATCH(lru, lru);
-FOLIO_MATCH(mapping, mapping);
+FOLIO_MATCH(___mapping, mapping);
 FOLIO_MATCH(compound_head, lru);
-FOLIO_MATCH(index, index);
+FOLIO_MATCH(___index, index);
 FOLIO_MATCH(private, private);
 FOLIO_MATCH(_mapcount, _mapcount);
 FOLIO_MATCH(_refcount, _refcount);
@@ -449,7 +449,7 @@ struct ptdesc {
 TABLE_MATCH(flags, __page_flags);
 TABLE_MATCH(compound_head, pt_list);
 TABLE_MATCH(compound_head, _pt_pad_1);
-TABLE_MATCH(mapping, __page_mapping);
+TABLE_MATCH(___mapping, __page_mapping);
 TABLE_MATCH(rcu_head, pt_rcu_head);
 TABLE_MATCH(page_type, __page_type);
 TABLE_MATCH(_refcount, _refcount);
diff --git a/include/linux/page-flags.h b/include/linux/page-flags.h
index a88e64acebfe..a7dbc86cb6af 100644
--- a/include/linux/page-flags.h
+++ b/include/linux/page-flags.h
@@ -642,11 +642,6 @@ static __always_inline bool folio_mapping_flags(struct folio *folio)
 	return ((unsigned long)folio->mapping & PAGE_MAPPING_FLAGS) != 0;
 }
 
-static __always_inline int PageMappingFlags(struct page *page)
-{
-	return ((unsigned long)page->mapping & PAGE_MAPPING_FLAGS) != 0;
-}
-
 static __always_inline bool folio_test_anon(struct folio *folio)
 {
 	return ((unsigned long)folio->mapping & PAGE_MAPPING_ANON) != 0;
@@ -665,7 +660,7 @@ static __always_inline bool __folio_test_movable(const struct folio *folio)
 
 static __always_inline int __PageMovable(struct page *page)
 {
-	return ((unsigned long)page->mapping & PAGE_MAPPING_FLAGS) ==
+	return ((unsigned long)page->___mapping & PAGE_MAPPING_FLAGS) ==
 				PAGE_MAPPING_MOVABLE;
 }
 
diff --git a/include/linux/pagemap.h b/include/linux/pagemap.h
index bcc1ea44b4e8..34bddb7204c9 100644
--- a/include/linux/pagemap.h
+++ b/include/linux/pagemap.h
@@ -865,30 +865,10 @@ static inline struct folio *read_mapping_folio(struct address_space *mapping,
  */
 static inline pgoff_t page_to_pgoff(struct page *page)
 {
-	struct page *head;
-
-	if (likely(!PageTransTail(page)))
-		return page->index;
-
-	head = compound_head(page);
-	/*
-	 *  We don't initialize ->index for tail pages: calculate based on
-	 *  head page
-	 */
-	return head->index + page - head;
-}
-
-/*
- * Return byte-offset into filesystem object for page.
- */
-static inline loff_t page_offset(struct page *page)
-{
-	return ((loff_t)page->index) << PAGE_SHIFT;
-}
+	struct folio *folio;
 
-static inline loff_t page_file_offset(struct page *page)
-{
-	return ((loff_t)page_index(page)) << PAGE_SHIFT;
+	folio = page_folio(page);
+	return folio->index + folio_page_idx(folio, page);
 }
 
 /**
@@ -897,7 +877,7 @@ static inline loff_t page_file_offset(struct page *page)
  */
 static inline loff_t folio_pos(struct folio *folio)
 {
-	return page_offset(&folio->page);
+	return (loff_t)folio->index * PAGE_SIZE;
 }
 
 /**
@@ -909,7 +889,7 @@ static inline loff_t folio_pos(struct folio *folio)
  */
 static inline loff_t folio_file_pos(struct folio *folio)
 {
-	return page_file_offset(&folio->page);
+	return (loff_t)folio_index(folio) * PAGE_SIZE;
 }
 
 /*
@@ -1464,34 +1444,6 @@ static inline ssize_t folio_mkwrite_check_truncate(struct folio *folio,
 	return offset;
 }
 
-/**
- * page_mkwrite_check_truncate - check if page was truncated
- * @page: the page to check
- * @inode: the inode to check the page against
- *
- * Returns the number of bytes in the page up to EOF,
- * or -EFAULT if the page was truncated.
- */
-static inline int page_mkwrite_check_truncate(struct page *page,
-					      struct inode *inode)
-{
-	loff_t size = i_size_read(inode);
-	pgoff_t index = size >> PAGE_SHIFT;
-	int offset = offset_in_page(size);
-
-	if (page->mapping != inode->i_mapping)
-		return -EFAULT;
-
-	/* page is wholly inside EOF */
-	if (page->index < index)
-		return PAGE_SIZE;
-	/* page is wholly past EOF */
-	if (page->index > index || !offset)
-		return -EFAULT;
-	/* page is partially inside EOF */
-	return offset;
-}
-
 /**
  * i_blocks_per_folio - How many blocks fit in this folio.
  * @inode: The inode which contains the blocks.
diff --git a/include/linux/poison.h b/include/linux/poison.h
index 851a855d3868..0d99a5265fba 100644
--- a/include/linux/poison.h
+++ b/include/linux/poison.h
@@ -29,10 +29,6 @@
 /********** mm/page_poison.c **********/
 #define PAGE_POISON 0xaa
 
-/********** mm/page_alloc.c ************/
-
-#define TAIL_MAPPING	((void *) 0x400 + POISON_POINTER_DELTA)
-
 /********** mm/slab.c **********/
 /*
  * Magic nums for obj red zoning.
diff --git a/kernel/events/core.c b/kernel/events/core.c
index 683dc086ef10..13acb878da17 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -6147,8 +6147,6 @@ static vm_fault_t perf_mmap_fault(struct vm_fault *vmf)
 		goto unlock;
 
 	get_page(vmf->page);
-	vmf->page->mapping = vmf->vma->vm_file->f_mapping;
-	vmf->page->index   = vmf->pgoff;
 
 	ret = 0;
 unlock:
diff --git a/kernel/events/ring_buffer.c b/kernel/events/ring_buffer.c
index e8d82c2f07d0..5bae0c228f3b 100644
--- a/kernel/events/ring_buffer.c
+++ b/kernel/events/ring_buffer.c
@@ -637,7 +637,6 @@ static void rb_free_aux_page(struct perf_buffer *rb, int idx)
 	struct page *page = virt_to_page(rb->aux_pages[idx]);
 
 	ClearPagePrivate(page);
-	page->mapping = NULL;
 	__free_page(page);
 }
 
@@ -808,7 +807,6 @@ static void perf_mmap_free_page(void *addr)
 {
 	struct page *page = virt_to_page(addr);
 
-	page->mapping = NULL;
 	__free_page(page);
 }
 
diff --git a/mm/debug.c b/mm/debug.c
index ee533a5ceb79..de9b1a53cbf9 100644
--- a/mm/debug.c
+++ b/mm/debug.c
@@ -75,7 +75,7 @@ static void __dump_page(struct page *page)
 		 * and potentially other situations. (See the page_mapping()
 		 * implementation for what's missing here.)
 		 */
-		unsigned long tmp = (unsigned long)page->mapping;
+		unsigned long tmp = (unsigned long)page->___mapping;
 
 		if (tmp & PAGE_MAPPING_ANON)
 			mapping = NULL;
diff --git a/mm/huge_memory.c b/mm/huge_memory.c
index f31f02472396..da89280d10e4 100644
--- a/mm/huge_memory.c
+++ b/mm/huge_memory.c
@@ -2478,11 +2478,8 @@ static void __split_huge_page_tail(struct folio *folio, int tail,
 			 (1L << PG_dirty) |
 			 LRU_GEN_MASK | LRU_REFS_MASK));
 
-	/* ->mapping in first and second tail page is replaced by other uses */
-	VM_BUG_ON_PAGE(tail > 2 && page_tail->mapping != TAIL_MAPPING,
-			page_tail);
-	page_tail->mapping = head->mapping;
-	page_tail->index = head->index + tail;
+	new_folio->mapping = folio->mapping;
+	new_folio->index = folio->index + tail;
 
 	/*
 	 * page->private should not be set in tail pages. Fix up and warn once
diff --git a/mm/internal.h b/mm/internal.h
index 7e84ec0219b1..8d92d8f15793 100644
--- a/mm/internal.h
+++ b/mm/internal.h
@@ -438,7 +438,6 @@ static inline void prep_compound_tail(struct page *head, int tail_idx)
 {
 	struct page *p = head + tail_idx;
 
-	p->mapping = TAIL_MAPPING;
 	set_compound_head(p, head);
 	set_page_private(p, 0);
 }
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 733732e7e0ba..8c75f703de0a 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -215,12 +215,12 @@ gfp_t gfp_allowed_mask __read_mostly = GFP_BOOT_MASK;
  */
 static inline int get_pcppage_migratetype(struct page *page)
 {
-	return page->index;
+	return page->___index;
 }
 
 static inline void set_pcppage_migratetype(struct page *page, int migratetype)
 {
-	page->index = migratetype;
+	page->___index = migratetype;
 }
 
 #ifdef CONFIG_HUGETLB_PAGE_SIZE_VARIABLE
@@ -911,7 +911,7 @@ static inline bool page_expected_state(struct page *page,
 	if (unlikely(atomic_read(&page->_mapcount) != -1))
 		return false;
 
-	if (unlikely((unsigned long)page->mapping |
+	if (unlikely((unsigned long)page->___mapping |
 			page_ref_count(page) |
 #ifdef CONFIG_MEMCG
 			page->memcg_data |
@@ -928,7 +928,7 @@ static const char *page_bad_reason(struct page *page, unsigned long flags)
 
 	if (unlikely(atomic_read(&page->_mapcount) != -1))
 		bad_reason = "nonzero mapcount";
-	if (unlikely(page->mapping != NULL))
+	if (unlikely(page->___mapping != NULL))
 		bad_reason = "non-NULL mapping";
 	if (unlikely(page_ref_count(page) != 0))
 		bad_reason = "nonzero _refcount";
@@ -981,9 +981,7 @@ static int free_tail_page_prepare(struct page *head_page, struct page *page)
 		ret = 0;
 		goto out;
 	}
-	switch (page - head_page) {
-	case 1:
-		/* the first tail page: these may be in place of ->mapping */
+	if (page - head_page == 1) {
 		if (unlikely(folio_entire_mapcount(folio))) {
 			bad_page(page, "nonzero entire_mapcount");
 			goto out;
@@ -996,19 +994,6 @@ static int free_tail_page_prepare(struct page *head_page, struct page *page)
 			bad_page(page, "nonzero pincount");
 			goto out;
 		}
-		break;
-	case 2:
-		/*
-		 * the second tail page: ->mapping is
-		 * deferred_list.next -- ignore value.
-		 */
-		break;
-	default:
-		if (page->mapping != TAIL_MAPPING) {
-			bad_page(page, "corrupted mapping in tail page");
-			goto out;
-		}
-		break;
 	}
 	if (unlikely(!PageTail(page))) {
 		bad_page(page, "PageTail not set");
@@ -1020,7 +1005,7 @@ static int free_tail_page_prepare(struct page *head_page, struct page *page)
 	}
 	ret = 0;
 out:
-	page->mapping = NULL;
+	page->___mapping = NULL;
 	clear_compound_head(page);
 	return ret;
 }
@@ -1080,8 +1065,10 @@ static __always_inline bool free_pages_prepare(struct page *page,
 	bool skip_kasan_poison = should_skip_kasan_poison(page, fpi_flags);
 	bool init = want_init_on_free();
 	bool compound = PageCompound(page);
+	struct folio *folio;
 
 	VM_BUG_ON_PAGE(PageTail(page), page);
+	folio = (struct folio *)page;
 
 	trace_mm_page_free(page, order);
 	kmsan_free_page(page, order);
@@ -1121,8 +1108,8 @@ static __always_inline bool free_pages_prepare(struct page *page,
 			(page + i)->flags &= ~PAGE_FLAGS_CHECK_AT_PREP;
 		}
 	}
-	if (PageMappingFlags(page))
-		page->mapping = NULL;
+	if (folio_mapping_flags(folio))
+		folio->mapping = NULL;
 	if (memcg_kmem_online() && PageMemcgKmem(page))
 		__memcg_kmem_uncharge_page(page, order);
 	if (is_check_pages_enabled()) {
diff --git a/mm/percpu.c b/mm/percpu.c
index 7b97d31df767..01fb98220d6e 100644
--- a/mm/percpu.c
+++ b/mm/percpu.c
@@ -253,13 +253,13 @@ static int pcpu_chunk_slot(const struct pcpu_chunk *chunk)
 /* set the pointer to a chunk in a page struct */
 static void pcpu_set_page_chunk(struct page *page, struct pcpu_chunk *pcpu)
 {
-	page->index = (unsigned long)pcpu;
+	page->___index = (unsigned long)pcpu;
 }
 
 /* obtain pointer to a chunk from a page struct */
 static struct pcpu_chunk *pcpu_get_page_chunk(struct page *page)
 {
-	return (struct pcpu_chunk *)page->index;
+	return (struct pcpu_chunk *)page->___index;
 }
 
 static int __maybe_unused pcpu_page_idx(unsigned int cpu, int page_idx)
diff --git a/tools/include/linux/poison.h b/tools/include/linux/poison.h
index 2e6338ac5eed..76eca9f9a68e 100644
--- a/tools/include/linux/poison.h
+++ b/tools/include/linux/poison.h
@@ -38,10 +38,6 @@
 /********** mm/page_poison.c **********/
 #define PAGE_POISON 0xaa
 
-/********** mm/page_alloc.c ************/
-
-#define TAIL_MAPPING	((void *) 0x400 + POISON_POINTER_DELTA)
-
 /********** mm/slab.c **********/
 /*
  * Magic nums for obj red zoning.






[Index of Archives]     [Linux ARM Kernel]     [Linux ARM]     [Linux Omap]     [Fedora ARM]     [IETF Annouce]     [Bugtraq]     [Linux OMAP]     [Linux MIPS]     [eCos]     [Asterisk Internet PBX]     [Linux API]

  Powered by Linux