VM: add may_swap flag to scan_control

Index: linux-2.6.12-rc5-mm1/mm/vmscan.c
===================================================================
@@ -74,6 +74,9 @@ struct scan_control {
 
 	int may_writepage;
 
+	/* Can pages be swapped as part of reclaim? */
+	int may_swap;
+
 	/* This context's SWAP_CLUSTER_MAX. If freeing memory for
 	 * suspend, we effectively ignore SWAP_CLUSTER_MAX.
 	 * In this context, it doesn't matter that we scan the

@@ -414,7 +417,7 @@ static int shrink_list(struct list_head 
 		 * Anonymous process memory has backing store?
 		 * Try to allocate it some swap space here.
 		 */
-		if (PageAnon(page) && !PageSwapCache(page)) {
+		if (PageAnon(page) && !PageSwapCache(page) && sc->may_swap) {
 			void *cookie = page->mapping;
 			pgoff_t index = page->index;
 

@@ -930,6 +933,7 @@ int try_to_free_pages(struct zone **zone
 
 	sc.gfp_mask = gfp_mask;
 	sc.may_writepage = 0;
+	sc.may_swap = 1;
 
 	inc_page_state(allocstall);
 

@@ -1030,6 +1034,7 @@ loop_again:
 	total_reclaimed = 0;
 	sc.gfp_mask = GFP_KERNEL;
 	sc.may_writepage = 0;
+	sc.may_swap = 1;
 	sc.nr_mapped = read_page_state(nr_mapped);
 
 	inc_page_state(pageoutrun);

Index: linux-2.6.12-rc5-mm1/include/linux/mmzone.h
===================================================================
@@ -145,6 +145,12 @@ struct zone {
 	int			all_unreclaimable; /* All pages pinned */
 
 	/*
+	 * Does the allocator try to reclaim pages from the zone as soon
+	 * as it fails a watermark_ok() in __alloc_pages?
+	 */
+	int			reclaim_pages;
+
+	/*
 	 * prev_priority holds the scanning priority for this zone.  It is
 	 * defined as the scanning priority at which we achieved our reclaim
 	 * target at the previous try_to_free_pages() or balance_pgdat()

Index: linux-2.6.12-rc5-mm1/include/linux/swap.h
===================================================================
@@ -173,6 +173,7 @@ extern void swap_setup(void);
 
 /* linux/mm/vmscan.c */
 extern int try_to_free_pages(struct zone **, unsigned int, unsigned int);
+extern int zone_reclaim(struct zone *, unsigned int, unsigned int);
 extern int shrink_all_memory(int);
 extern int vm_swappiness;
 

Index: linux-2.6.12-rc5-mm1/mm/page_alloc.c
===================================================================
@@ -724,6 +724,14 @@ int zone_watermark_ok(struct zone *z, in
 	return 1;
 }
 
+static inline int
+check_zone_reclaim(struct zone *z, unsigned int gfp_mask)
+{
+	if (!z->reclaim_pages)
+		return 0;
+	return 1;
+}
+
 /*
  * This is the 'heart' of the zoned buddy allocator.
  */

@@ -763,14 +771,29 @@ __alloc_pages(unsigned int __nocast gfp_
  restart:
 	/* Go through the zonelist once, looking for a zone with enough free */
 	for (i = 0; (z = zones[i]) != NULL; i++) {
-
-		if (!zone_watermark_ok(z, order, z->pages_low,
-				       classzone_idx, 0, 0))
-			continue;
+		int do_reclaim = check_zone_reclaim(z, gfp_mask);
 
 		if (!cpuset_zone_allowed(z))
 			continue;
 
+		/*
+		 * If the zone is to attempt early page reclaim then this loop
+		 * will try to reclaim pages and check the watermark a second
+		 * time before giving up and falling back to the next zone.
+		 */
+	zone_reclaim_retry:
+		if (!zone_watermark_ok(z, order, z->pages_low,
+				       classzone_idx, 0, 0)) {
+			if (!do_reclaim)
+				continue;
+			else {
+				zone_reclaim(z, gfp_mask, order);
+				/* Only try reclaim once */
+				do_reclaim = 0;
+				goto zone_reclaim_retry;
+			}
+		}
+
 		page = buffered_rmqueue(z, order, gfp_mask);
 		if (page)
 			goto got_pg;

67 lines not on this page · raw message

Index: linux-2.6.12-rc5-mm1/kernel/sys_ni.c
===================================================================
@@ -77,6 +77,7 @@ cond_syscall(sys_request_key);
 cond_syscall(sys_keyctl);
 cond_syscall(compat_sys_keyctl);
 cond_syscall(compat_sys_socketcall);
+cond_syscall(sys_set_zone_reclaim);
 
 /* arch-specific weak syscall entries */
 cond_syscall(sys_pciconfig_read);

Index: linux-2.6.12-rc5-mm1/include/asm-i386/unistd.h
===================================================================
@@ -256,7 +256,7 @@
 #define __NR_io_submit		248
 #define __NR_io_cancel		249
 #define __NR_fadvise64		250
-
+#define __NR_set_zone_reclaim	251
 #define __NR_exit_group		252
 #define __NR_lookup_dcookie	253
 #define __NR_epoll_create	254

Index: linux-2.6.12-rc5-mm1/include/asm-ia64/unistd.h
===================================================================
@@ -263,6 +263,7 @@
 #define __NR_add_key			1271
 #define __NR_request_key		1272
 #define __NR_keyctl			1273
+#define __NR_set_zone_reclaim		1276
 
 #ifdef __KERNEL__
 

Index: linux-2.6.12-rc5-mm1/include/linux/gfp.h
===================================================================
@@ -39,6 +39,7 @@ struct vm_area_struct;
 #define __GFP_COMP	0x4000u	/* Add compound page metadata */
 #define __GFP_ZERO	0x8000u	/* Return zeroed page on success */
 #define __GFP_NOMEMALLOC 0x10000u /* Don't use emergency reserves */
+#define __GFP_NORECLAIM  0x20000u /* No realy zone reclaim during allocation */
 
 #define __GFP_BITS_SHIFT 20	/* Room for 20 __GFP_FOO bits */
 #define __GFP_BITS_MASK ((1 << __GFP_BITS_SHIFT) - 1)

@@ -47,7 +48,7 @@ struct vm_area_struct;
 #define GFP_LEVEL_MASK (__GFP_WAIT|__GFP_HIGH|__GFP_IO|__GFP_FS| \
 			__GFP_COLD|__GFP_NOWARN|__GFP_REPEAT| \
 			__GFP_NOFAIL|__GFP_NORETRY|__GFP_NO_GROW|__GFP_COMP| \
-			__GFP_NOMEMALLOC)
+			__GFP_NOMEMALLOC|__GFP_NORECLAIM)
 
 #define GFP_ATOMIC	(__GFP_HIGH)
 #define GFP_NOIO	(__GFP_WAIT)

Index: linux-2.6.12-rc5-mm1/include/linux/pagemap.h
===================================================================
@@ -52,12 +52,12 @@ void release_pages(struct page **pages, 
 
 static inline struct page *page_cache_alloc(struct address_space *x)
 {
-	return alloc_pages(mapping_gfp_mask(x), 0);
+	return alloc_pages(mapping_gfp_mask(x)|__GFP_NORECLAIM, 0);
 }
 
 static inline struct page *page_cache_alloc_cold(struct address_space *x)
 {
-	return alloc_pages(mapping_gfp_mask(x)|__GFP_COLD, 0);
+	return alloc_pages(mapping_gfp_mask(x)|__GFP_COLD|__GFP_NORECLAIM, 0);
 }
 
 typedef int filler_t(void *, struct page *);

Index: linux-2.6.12-rc5-mm1/mm/page_alloc.c
===================================================================
@@ -729,6 +729,8 @@ check_zone_reclaim(struct zone *z, unsig
 {
 	if (!z->reclaim_pages)
 		return 0;
+	if (gfp_mask & __GFP_NORECLAIM)
+		return 0;
 	return 1;
 }
 

Index: linux-2.6.12-rc5-mm1/mm/vmscan.c
===================================================================
@@ -903,7 +903,9 @@ shrink_caches(struct zone **zones, struc
 		if (zone->all_unreclaimable && sc->priority != DEF_PRIORITY)
 			continue;	/* Let kswapd poll it */
 
+		atomic_inc(&zone->reclaim_in_progress);
 		shrink_zone(zone, sc);
+		atomic_dec(&zone->reclaim_in_progress);
 	}
 }
  

@@ -1114,7 +1116,9 @@ scan:
 			sc.nr_reclaimed = 0;
 			sc.priority = priority;
 			sc.swap_cluster_max = nr_pages? nr_pages : SWAP_CLUSTER_MAX;
+			atomic_inc(&zone->reclaim_in_progress);
 			shrink_zone(zone, &sc);
+			atomic_dec(&zone->reclaim_in_progress);
 			reclaim_state->reclaimed_slab = 0;
 			nr_slab = shrink_slab(sc.nr_scanned, GFP_KERNEL,
 						lru_pages);

@@ -1357,9 +1361,15 @@ int zone_reclaim(struct zone *zone, unsi
 	else
 		sc.swap_cluster_max = SWAP_CLUSTER_MAX;
 
+	/* Don't reclaim the zone if there are other reclaimers active */
+	if (!atomic_inc_and_test(&zone->reclaim_in_progress))
+		goto out;
+
 	shrink_zone(zone, &sc);
 	total_reclaimed = sc.nr_reclaimed;
 
+ out:
+	atomic_dec(&zone->reclaim_in_progress);
 	return total_reclaimed;
 }
 

Index: linux-2.6.12-rc5-mm1/include/linux/mmzone.h
===================================================================
@@ -149,6 +149,8 @@ struct zone {
 	 * as it fails a watermark_ok() in __alloc_pages?
 	 */
 	int			reclaim_pages;
+	/* A count of how many reclaimers are scanning this zone */
+	atomic_t		reclaim_in_progress;
 
 	/*
 	 * prev_priority holds the scanning priority for this zone.  It is

Index: linux-2.6.12-rc5-mm1/mm/page_alloc.c
===================================================================
@@ -1757,6 +1757,7 @@ static void __init free_area_init_core(s
 		zone->nr_scan_inactive = 0;
 		zone->nr_active = 0;
 		zone->nr_inactive = 0;
+		atomic_set(&zone->reclaim_in_progress, -1);
 		if (!size)
 			continue;