mm: a few small updates for radix-swap

Remove PageSwapBacked (!page_is_file_cache) cases from add_to_page_cache_locked() and add_to_page_cache_lru(): those pages now go through shmem_add_to_page_cache(). Remove a comment on maximum tmpfs size from fsstack_copy_inode_size(), and add a comment on swap entries to invalidate_mapping_pages(). And mincore_page() uses find_get_page() on what might be shmem or a tmpfs file: allow for a radix_tree_exceptional_entry(), and proceed to find_get_page() on swapper_space if so (oh, swapper_space needs #ifdef). Signed-off-by: Hugh Dickins <hughd@google.com> Acked-by: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm: a few small updates for radix-swap
Remove PageSwapBacked (!page_is_file_cache) cases from add_to_page_cache_locked() and add_to_page_cache_lru(): those pages now go through shmem_add_to_page_cache(). Remove a comment on maximum tmpfs size from fsstack_copy_inode_size(), and add a comment on swap entries to invalidate_mapping_pages(). And mincore_page() uses find_get_page() on what might be shmem or a tmpfs file: allow for a radix_tree_exceptional_entry(), and proceed to find_get_page() on swapper_space if so (oh, swapper_space needs #ifdef). Signed-off-by: Hugh Dickins <hughd@google.com> Acked-by: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Hugh Dickins · Linus Torvalds
1 parent 69f07ec938
Showing 4 changed files with 18 additions and 26 deletions Side-by-side Diff
fs/stack.c
mm/filemap.c
mm/mincore.c
mm/truncate.c
@@ -29,10 +29,7 @@
 	 *
 	 * We don't actually know what locking is used at the lower level;
 	 * but if it's a filesystem that supports quotas, it will be using
-	 * i_lock as in inode_add_bytes().  tmpfs uses other locking, and
-	 * its 32-bit is (just) able to exceed 2TB i_size with the aid of
-	 * holes; but its i_blocks cannot carry into the upper long without
-	 * almost 2TB swap - let's ignore that case.
+	 * i_lock as in inode_add_bytes().
 	 */
 	if (sizeof(i_blocks) > sizeof(long))
 		spin_lock(&src->i_lock);
@@ -33,7 +33,6 @@
 #include <linux/cpuset.h>
 #include <linux/hardirq.h> /* for BUG_ON(!in_atomic()) only */
 #include <linux/memcontrol.h>
-#include <linux/mm_inline.h> /* for page_is_file_cache() */
 #include <linux/cleancache.h>
 #include "internal.h"
  
@@ -462,6 +461,7 @@
 	int error;
  
 	VM_BUG_ON(!PageLocked(page));
+	VM_BUG_ON(PageSwapBacked(page));
  
 	error = mem_cgroup_cache_charge(page, current->mm,
 					gfp_mask & GFP_RECLAIM_MASK);
@@ -479,8 +479,6 @@
 		if (likely(!error)) {
 			mapping->nrpages++;
 			__inc_zone_page_state(page, NR_FILE_PAGES);
-			if (PageSwapBacked(page))
-				__inc_zone_page_state(page, NR_SHMEM);
 			spin_unlock_irq(&mapping->tree_lock);
 		} else {
 			page->mapping = NULL;
  
@@ -502,22 +500,9 @@
 {
 	int ret;
  
-	/*
-	 * Splice_read and readahead add shmem/tmpfs pages into the page cache
-	 * before shmem_readpage has a chance to mark them as SwapBacked: they
-	 * need to go on the anon lru below, and mem_cgroup_cache_charge
-	 * (called in add_to_page_cache) needs to know where they're going too.
-	 */
-	if (mapping_cap_swap_backed(mapping))
-		SetPageSwapBacked(page);
-
 	ret = add_to_page_cache(page, mapping, offset, gfp_mask);
-	if (ret == 0) {
-		if (page_is_file_cache(page))
-			lru_cache_add_file(page);
-		else
-			lru_cache_add_anon(page);
-	}
+	if (ret == 0)
+		lru_cache_add_file(page);
 	return ret;
 }
 EXPORT_SYMBOL_GPL(add_to_page_cache_lru);
@@ -69,12 +69,14 @@
 	 * file will not get a swp_entry_t in its pte, but rather it is like
 	 * any other file mapping (ie. marked !present and faulted in with
 	 * tmpfs's .fault). So swapped out tmpfs mappings are tested here.
-	 *
-	 * However when tmpfs moves the page from pagecache and into swapcache,
-	 * it is still in core, but the find_get_page below won't find it.
-	 * No big deal, but make a note of it.
 	 */
 	page = find_get_page(mapping, pgoff);
+#ifdef CONFIG_SWAP
+	if (radix_tree_exceptional_entry(page)) {
+		swp_entry_t swap = radix_to_swp_entry(page);
+		page = find_get_page(&swapper_space, swap.val);
+	}
+#endif
 	if (page) {
 		present = PageUptodate(page);
 		page_cache_release(page);
@@ -336,6 +336,14 @@
 	unsigned long count = 0;
 	int i;
  
+	/*
+	 * Note: this function may get called on a shmem/tmpfs mapping:
+	 * pagevec_lookup() might then return 0 prematurely (because it
+	 * got a gangful of swap entries); but it's hardly worth worrying
+	 * about - it can rarely have anything to free from such a mapping
+	 * (most pages are dirty), and already skips over any difficulties.
+	 */
+
 	pagevec_init(&pvec, 0);
 	while (index <= end && pagevec_lookup(&pvec, mapping, index,
 			min(end - index, (pgoff_t)PAGEVEC_SIZE - 1) + 1)) {
...	...	@@ -29,10 +29,7 @@
29	29	*
30	30	* We don't actually know what locking is used at the lower level;
31	31	* but if it's a filesystem that supports quotas, it will be using
32		- * i_lock as in inode_add_bytes(). tmpfs uses other locking, and
33		- * its 32-bit is (just) able to exceed 2TB i_size with the aid of
34		- * holes; but its i_blocks cannot carry into the upper long without
35		- * almost 2TB swap - let's ignore that case.
	32	+ * i_lock as in inode_add_bytes().
36	33	*/
37	34	if (sizeof(i_blocks) > sizeof(long))
38	35	spin_lock(&src->i_lock);
...	...	@@ -33,7 +33,6 @@
33	33	#include <linux/cpuset.h>
34	34	#include <linux/hardirq.h> /* for BUG_ON(!in_atomic()) only */
35	35	#include <linux/memcontrol.h>
36		-#include <linux/mm_inline.h> /* for page_is_file_cache() */
37	36	#include <linux/cleancache.h>
38	37	#include "internal.h"
39	38
...	...	@@ -462,6 +461,7 @@
462	461	int error;
463	462
464	463	VM_BUG_ON(!PageLocked(page));
	464	+ VM_BUG_ON(PageSwapBacked(page));
465	465
466	466	error = mem_cgroup_cache_charge(page, current->mm,
467	467	gfp_mask & GFP_RECLAIM_MASK);
...	...	@@ -479,8 +479,6 @@
479	479	if (likely(!error)) {
480	480	mapping->nrpages++;
481	481	__inc_zone_page_state(page, NR_FILE_PAGES);
482		- if (PageSwapBacked(page))
483		- __inc_zone_page_state(page, NR_SHMEM);
484	482	spin_unlock_irq(&mapping->tree_lock);
485	483	} else {
486	484	page->mapping = NULL;
487	485
...	...	@@ -502,22 +500,9 @@
502	500	{
503	501	int ret;
504	502
505		- /*
506		- * Splice_read and readahead add shmem/tmpfs pages into the page cache
507		- * before shmem_readpage has a chance to mark them as SwapBacked: they
508		- * need to go on the anon lru below, and mem_cgroup_cache_charge
509		- * (called in add_to_page_cache) needs to know where they're going too.
510		- */
511		- if (mapping_cap_swap_backed(mapping))
512		- SetPageSwapBacked(page);
513		-
514	503	ret = add_to_page_cache(page, mapping, offset, gfp_mask);
515		- if (ret == 0) {
516		- if (page_is_file_cache(page))
517		- lru_cache_add_file(page);
518		- else
519		- lru_cache_add_anon(page);
520		- }
	504	+ if (ret == 0)
	505	+ lru_cache_add_file(page);
521	506	return ret;
522	507	}
523	508	EXPORT_SYMBOL_GPL(add_to_page_cache_lru);
...	...	@@ -69,12 +69,14 @@
69	69	* file will not get a swp_entry_t in its pte, but rather it is like
70	70	* any other file mapping (ie. marked !present and faulted in with
71	71	* tmpfs's .fault). So swapped out tmpfs mappings are tested here.
72		- *
73		- * However when tmpfs moves the page from pagecache and into swapcache,
74		- * it is still in core, but the find_get_page below won't find it.
75		- * No big deal, but make a note of it.
76	72	*/
77	73	page = find_get_page(mapping, pgoff);
	74	+#ifdef CONFIG_SWAP
	75	+ if (radix_tree_exceptional_entry(page)) {
	76	+ swp_entry_t swap = radix_to_swp_entry(page);
	77	+ page = find_get_page(&swapper_space, swap.val);
	78	+ }
	79	+#endif
78	80	if (page) {
79	81	present = PageUptodate(page);
80	82	page_cache_release(page);
...	...	@@ -336,6 +336,14 @@
336	336	unsigned long count = 0;
337	337	int i;
338	338
	339	+ /*
	340	+ * Note: this function may get called on a shmem/tmpfs mapping:
	341	+ * pagevec_lookup() might then return 0 prematurely (because it
	342	+ * got a gangful of swap entries); but it's hardly worth worrying
	343	+ * about - it can rarely have anything to free from such a mapping
	344	+ * (most pages are dirty), and already skips over any difficulties.
	345	+ */
	346	+
339	347	pagevec_init(&pvec, 0);
340	348	while (index <= end && pagevec_lookup(&pvec, mapping, index,
341	349	min(end - index, (pgoff_t)PAGEVEC_SIZE - 1) + 1)) {