RSS controller based on process containers (v2)

diff -upr linux-2.6.20.orig/include/linux/res_counter.h linux-2.6.20-2/include/linux/res_counter.h
@@ -0,0 +1,83 @@
+#ifndef __RES_COUNTER_H__
+#define __RES_COUNTER_H__
+/*
+ * resource counters
+ *
+ * Copyright 2007 OpenVZ SWsoft Inc
+ *
+ * Author: Pavel Emelianov <[email protected]>
+ *
+ */
+
+#include <linux/container.h>
+
+struct res_counter {
+	unsigned long usage;
+	unsigned long limit;
+	unsigned long failcnt;
+	spinlock_t lock;
+};
+
+enum {
+	RES_USAGE,
+	RES_LIMIT,
+	RES_FAILCNT,
+};
+
+ssize_t res_counter_read(struct res_counter *cnt, int member,
+		const char __user *buf, size_t nbytes, loff_t *pos);
+ssize_t res_counter_write(struct res_counter *cnt, int member,
+		const char __user *buf, size_t nbytes, loff_t *pos);
+
+static inline void res_counter_init(struct res_counter *cnt)
+{
+	spin_lock_init(&cnt->lock);
+	cnt->limit = (unsigned long)LONG_MAX;
+}
+
+static inline int res_counter_charge_locked(struct res_counter *cnt,
+		unsigned long val)
+{
+	if (cnt->usage <= cnt->limit - val) {
+		cnt->usage += val;
+		return 0;
+	}
+
+	cnt->failcnt++;
+	return -ENOMEM;
+}
+
+static inline int res_counter_charge(struct res_counter *cnt,
+		unsigned long val)
+{
+	int ret;
+	unsigned long flags;
+
+	spin_lock_irqsave(&cnt->lock, flags);
+	ret = res_counter_charge_locked(cnt, val);
+	spin_unlock_irqrestore(&cnt->lock, flags);
+	return ret;
+}
+
+static inline void res_counter_uncharge_locked(struct res_counter *cnt,
+		unsigned long val)
+{
+	if (unlikely(cnt->usage < val)) {
+		WARN_ON(1);
+		val = cnt->usage;
+	}
+
+	cnt->usage -= val;
+}
+
+static inline void res_counter_uncharge(struct res_counter *cnt,
+		unsigned long val)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&cnt->lock, flags);
+	res_counter_uncharge_locked(cnt, val);
+	spin_unlock_irqrestore(&cnt->lock, flags);
+}
+
+#endif

72 lines not on this page · raw message

diff -upr linux-2.6.20.orig/include/linux/mm.h linux-2.6.20-2/include/linux/mm.h
@@ -220,6 +220,12 @@ struct vm_operations_struct {
 struct mmu_gather;
 struct inode;
 
+#ifdef CONFIG_RSS_CONTAINER
+#define page_container(page)	(page->rss_container)
+#else
+#define page_container(page)	(NULL)
+#endif
+
 #define page_private(page)		((page)->private)
 #define set_page_private(page, v)	((page)->private = (v))
 

diff -upr linux-2.6.20.orig/include/linux/mm_types.h linux-2.6.20-2/include/linux/mm_types.h
@@ -62,6 +62,9 @@ struct page {
 	void *virtual;			/* Kernel virtual address (NULL if
 					   not kmapped, ie. highmem) */
 #endif /* WANT_PAGE_VIRTUAL */
+#ifdef CONFIG_RSS_CONTAINER
+	struct page_container *rss_container;
+#endif
 };
 
 #endif /* _LINUX_MM_TYPES_H */

diff -upr linux-2.6.20.orig/include/linux/sched.h linux-2.6.20-2/include/linux/sched.h
@@ -373,6 +373,9 @@ struct mm_struct {
 	/* aio bits */
 	rwlock_t		ioctx_list_lock;
 	struct kioctx		*ioctx_list;
+#ifdef CONFIG_RSS_CONTAINER
+	struct rss_container	*rss_container;
+#endif
 };
 
 struct sighand_struct {

diff -upr linux-2.6.20.orig/kernel/fork.c linux-2.6.20-2/kernel/fork.c
@@ -57,6 +57,8 @@
 #include <asm/cacheflush.h>
 #include <asm/tlbflush.h>
 
+#include <linux/rss_container.h>
+
 /*
  * Protected counters by write_lock_irq(&tasklist_lock)
  */

@@ -325,7 +327,7 @@ static inline void mm_free_pgd(struct mm
 
 #include <linux/init_task.h>
 
-static struct mm_struct * mm_init(struct mm_struct * mm)
+static struct mm_struct * mm_init(struct mm_struct *mm, struct task_struct *tsk)
 {
 	atomic_set(&mm->mm_users, 1);
 	atomic_set(&mm->mm_count, 1);

@@ -340,11 +342,14 @@ static struct mm_struct * mm_init(struct
 	mm->ioctx_list = NULL;
 	mm->free_area_cache = TASK_UNMAPPED_BASE;
 	mm->cached_hole_size = ~0UL;
+	mm_init_container(mm, tsk);
 
 	if (likely(!mm_alloc_pgd(mm))) {
 		mm->def_flags = 0;
 		return mm;
 	}
+
+	mm_free_container(mm);
 	free_mm(mm);
 	return NULL;
 }

@@ -359,7 +364,7 @@ struct mm_struct * mm_alloc(void)
 	mm = allocate_mm();
 	if (mm) {
 		memset(mm, 0, sizeof(*mm));
-		mm = mm_init(mm);
+		mm = mm_init(mm, current);
 	}
 	return mm;
 }

@@ -373,6 +378,7 @@ void fastcall __mmdrop(struct mm_struct 
 {
 	BUG_ON(mm == &init_mm);
 	mm_free_pgd(mm);
+	mm_free_container(mm);
 	destroy_context(mm);
 	free_mm(mm);
 }

@@ -493,7 +499,7 @@ static struct mm_struct *dup_mm(struct t
 	mm->token_priority = 0;
 	mm->last_interval = 0;
 
-	if (!mm_init(mm))
+	if (!mm_init(mm, tsk))
 		goto fail_nomem;
 
 	if (init_new_context(tsk, mm))

@@ -520,6 +526,7 @@ fail_nocontext:
 	 * because it calls destroy_context()
 	 */
 	mm_free_pgd(mm);
+	mm_free_container(mm);
 	free_mm(mm);
 	return NULL;
 }

diff -upr linux-2.6.20.orig/include/linux/container_subsys.h linux-2.6.20-2/include/linux/container_subsys.h
@@ -9,6 +9,10 @@
 SUBSYS(cpuset)
 #endif
 
+#ifdef CONFIG_RSS_CONTAINER
+SUBSYS(rss)
+#endif
+
 /* */
 
 /* */

55 lines not on this page · raw message

274 lines not on this page · raw message

diff -upr linux-2.6.20.orig/fs/exec.c linux-2.6.20-2/fs/exec.c
@@ -58,6 +58,8 @@
 #include <linux/kmod.h>
 #endif
 
+#include <linux/rss_container.h>
+
 int core_uses_pid;
 char core_pattern[128] = "core";
 int suid_dumpable = 0;

@@ -309,27 +311,34 @@ void install_arg_page(struct vm_area_str
 	struct mm_struct *mm = vma->vm_mm;
 	pte_t * pte;
 	spinlock_t *ptl;
+	struct page_container *pcont;
 
 	if (unlikely(anon_vma_prepare(vma)))
 		goto out;
 
+	if (container_rss_prepare(page, vma, &pcont))
+		goto out;
+
 	flush_dcache_page(page);
 	pte = get_locked_pte(mm, address, &ptl);
 	if (!pte)
-		goto out;
+		goto out_release;
 	if (!pte_none(*pte)) {
 		pte_unmap_unlock(pte, ptl);
-		goto out;
+		goto out_release;
 	}
 	inc_mm_counter(mm, anon_rss);
 	lru_cache_add_active(page);
 	set_pte_at(mm, address, pte, pte_mkdirty(pte_mkwrite(mk_pte(
 					page, vma->vm_page_prot))));
-	page_add_new_anon_rmap(page, vma, address);
+	page_add_new_anon_rmap(page, vma, address, pcont);
 	pte_unmap_unlock(pte, ptl);
 
 	/* no need for flush_tlb */
 	return;
+
+out_release:
+	container_rss_release(pcont);
 out:
 	__free_page(page);
 	force_sig(SIGKILL, current);

diff -upr linux-2.6.20.orig/include/linux/rmap.h linux-2.6.20-2/include/linux/rmap.h
@@ -69,9 +69,13 @@ void __anon_vma_link(struct vm_area_stru
 /*
  * rmap interfaces called when adding or removing pte of page
  */
-void page_add_anon_rmap(struct page *, struct vm_area_struct *, unsigned long);
-void page_add_new_anon_rmap(struct page *, struct vm_area_struct *, unsigned long);
-void page_add_file_rmap(struct page *);
+struct page_container;
+
+void page_add_anon_rmap(struct page *, struct vm_area_struct *,
+		unsigned long, struct page_container *);
+void page_add_new_anon_rmap(struct page *, struct vm_area_struct *,
+		unsigned long, struct page_container *);
+void page_add_file_rmap(struct page *, struct page_container *);
 void page_remove_rmap(struct page *, struct vm_area_struct *);
 
 /**

diff -upr linux-2.6.20.orig/mm/fremap.c linux-2.6.20-2/mm/fremap.c
@@ -20,6 +20,8 @@
 #include <asm/cacheflush.h>
 #include <asm/tlbflush.h>
 
+#include <linux/rss_container.h>
+
 static int zap_pte(struct mm_struct *mm, struct vm_area_struct *vma,
 			unsigned long addr, pte_t *ptep)
 {

@@ -57,6 +59,10 @@ int install_page(struct mm_struct *mm, s
 	pte_t *pte;
 	pte_t pte_val;
 	spinlock_t *ptl;
+	struct page_container *pcont;
+
+	if (container_rss_prepare(page, vma, &pcont))
+		goto out_release;
 
 	pte = get_locked_pte(mm, addr, &ptl);
 	if (!pte)

@@ -81,13 +87,16 @@ int install_page(struct mm_struct *mm, s
 	flush_icache_page(vma, page);
 	pte_val = mk_pte(page, prot);
 	set_pte_at(mm, addr, pte, pte_val);
-	page_add_file_rmap(page);
+	page_add_file_rmap(page, pcont);
 	update_mmu_cache(vma, addr, pte_val);
 	lazy_mmu_prot_update(pte_val);
 	err = 0;
 unlock:
 	pte_unmap_unlock(pte, ptl);
 out:
+	if (err != 0)
+		container_rss_release(pcont);
+out_release:
 	return err;
 }
 EXPORT_SYMBOL(install_page);

188 lines not on this page · raw message

diff -upr linux-2.6.20.orig/mm/migrate.c linux-2.6.20-2/mm/migrate.c
@@ -28,6 +28,7 @@
 #include <linux/mempolicy.h>
 #include <linux/vmalloc.h>
 #include <linux/security.h>
+#include <linux/rss_container.h>
 
 #include "internal.h"
 

@@ -134,6 +135,7 @@ static void remove_migration_pte(struct 
 	pte_t *ptep, pte;
  	spinlock_t *ptl;
 	unsigned long addr = page_address_in_vma(new, vma);
+	struct page_container *pcont;
 
 	if (addr == -EFAULT)
 		return;

@@ -157,6 +159,11 @@ static void remove_migration_pte(struct 
  		return;
  	}
 
+	if (container_rss_prepare(new, vma, &pcont)) {
+		pte_unmap(ptep);
+		return;
+	}
+
  	ptl = pte_lockptr(mm, pmd);
  	spin_lock(ptl);
 	pte = *ptep;

@@ -175,16 +182,19 @@ static void remove_migration_pte(struct 
 	set_pte_at(mm, addr, ptep, pte);
 
 	if (PageAnon(new))
-		page_add_anon_rmap(new, vma, addr);
+		page_add_anon_rmap(new, vma, addr, pcont);
 	else
-		page_add_file_rmap(new);
+		page_add_file_rmap(new, pcont);
 
 	/* No need to invalidate - it was non-present before */
 	update_mmu_cache(vma, addr, pte);
 	lazy_mmu_prot_update(pte);
+	pte_unmap_unlock(ptep, ptl);
+	return;
 
 out:
 	pte_unmap_unlock(ptep, ptl);
+	container_rss_release(pcont);
 }
 
 /*

diff -upr linux-2.6.20.orig/mm/rmap.c linux-2.6.20-2/mm/rmap.c
@@ -51,6 +51,8 @@
 
 #include <asm/tlbflush.h>
 
+#include <linux/rss_container.h>
+
 struct kmem_cache *anon_vma_cachep;
 
 static inline void validate_anon_vma(struct vm_area_struct *find_vma)

@@ -526,14 +528,19 @@ static void __page_set_anon_rmap(struct 
  * @page:	the page to add the mapping to
  * @vma:	the vm area in which the mapping is added
  * @address:	the user virtual address mapped
+ * @pcont:	the page beancounter to charge page with
  *
  * The caller needs to hold the pte lock.
  */
 void page_add_anon_rmap(struct page *page,
-	struct vm_area_struct *vma, unsigned long address)
+	struct vm_area_struct *vma, unsigned long address,
+	struct page_container *pcont)
 {
-	if (atomic_inc_and_test(&page->_mapcount))
+	if (atomic_inc_and_test(&page->_mapcount)) {
+		container_rss_add(pcont);
 		__page_set_anon_rmap(page, vma, address);
+	} else
+		container_rss_release(pcont);
 	/* else checking page index and mapping is racy */
 }
 

@@ -542,27 +549,35 @@ void page_add_anon_rmap(struct page *pag
  * @page:	the page to add the mapping to
  * @vma:	the vm area in which the mapping is added
  * @address:	the user virtual address mapped
+ * @pcont:	the page beancounter to charge page with
  *
  * Same as page_add_anon_rmap but must only be called on *new* pages.
  * This means the inc-and-test can be bypassed.
  */
 void page_add_new_anon_rmap(struct page *page,
-	struct vm_area_struct *vma, unsigned long address)
+	struct vm_area_struct *vma, unsigned long address,
+	struct page_container *pcont)
 {
 	atomic_set(&page->_mapcount, 0); /* elevate count by 1 (starts at -1) */
+	container_rss_add(pcont);
 	__page_set_anon_rmap(page, vma, address);
 }
 
 /**
  * page_add_file_rmap - add pte mapping to a file page
- * @page: the page to add the mapping to
+ * @page:	the page to add the mapping to
+ * @pcont:	the page beancounter to charge page with
  *
  * The caller needs to hold the pte lock.
  */
-void page_add_file_rmap(struct page *page)
+void page_add_file_rmap(struct page *page, struct page_container *pcont)
 {
-	if (atomic_inc_and_test(&page->_mapcount))
+	if (atomic_inc_and_test(&page->_mapcount)) {
+		if (pcont)
+			container_rss_add(pcont);
 		__inc_zone_page_state(page, NR_FILE_MAPPED);
+	} else if (pcont)
+		container_rss_release(pcont);
 }
 
 /**

@@ -573,6 +588,9 @@ void page_add_file_rmap(struct page *pag
  */
 void page_remove_rmap(struct page *page, struct vm_area_struct *vma)
 {
+	struct page_container *pcont;
+
+	pcont = page_container(page);
 	if (atomic_add_negative(-1, &page->_mapcount)) {
 		if (unlikely(page_mapcount(page) < 0)) {
 			printk (KERN_EMERG "Eeek! page_mapcount(page) went negative! (%d)\n", page_mapcount(page));

@@ -588,6 +606,8 @@ void page_remove_rmap(struct page *page,
 			BUG();
 		}
 
+		if (pcont)
+			container_rss_del(pcont);
 		/*
 		 * It would be tidy to reset the PageAnon mapping here,
 		 * but that might overwrite a racing page_add_anon_rmap

diff -upr linux-2.6.20.orig/mm/swapfile.c linux-2.6.20-2/mm/swapfile.c
@@ -32,6 +32,8 @@
 #include <asm/tlbflush.h>
 #include <linux/swapops.h>
 
+#include <linux/rss_container.h>
+
 DEFINE_SPINLOCK(swap_lock);
 unsigned int nr_swapfiles;
 long total_swap_pages;

@@ -507,13 +509,14 @@ unsigned int count_swap_pages(int type, 
  * force COW, vm_page_prot omits write permission from any private vma.
  */
 static void unuse_pte(struct vm_area_struct *vma, pte_t *pte,
-		unsigned long addr, swp_entry_t entry, struct page *page)
+		unsigned long addr, swp_entry_t entry, struct page *page,
+		struct page_container *pcont)
 {
 	inc_mm_counter(vma->vm_mm, anon_rss);
 	get_page(page);
 	set_pte_at(vma->vm_mm, addr, pte,
 		   pte_mkold(mk_pte(page, vma->vm_page_prot)));
-	page_add_anon_rmap(page, vma, addr);
+	page_add_anon_rmap(page, vma, addr, pcont);
 	swap_free(entry);
 	/*
 	 * Move the page to the active list so it is not

@@ -530,6 +533,10 @@ static int unuse_pte_range(struct vm_are
 	pte_t *pte;
 	spinlock_t *ptl;
 	int found = 0;
+	struct page_container *pcont;
+
+	if (container_rss_prepare(page, vma, &pcont))
+		return 0;
 
 	pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl);
 	do {

@@ -538,12 +545,14 @@ static int unuse_pte_range(struct vm_are
 		 * Test inline before going to call unuse_pte.
 		 */
 		if (unlikely(pte_same(*pte, swp_pte))) {
-			unuse_pte(vma, pte++, addr, entry, page);
+			unuse_pte(vma, pte++, addr, entry, page, pcont);
 			found = 1;
 			break;
 		}
 	} while (pte++, addr += PAGE_SIZE, addr != end);
 	pte_unmap_unlock(pte - 1, ptl);
+	if (!found)
+		container_rss_release(pcont);
 	return found;
 }
 

diff -upr linux-2.6.20.orig/include/linux/rss_container.h linux-2.6.20-2/include/linux/rss_container.h
@@ -19,6 +19,7 @@
 void container_rss_add(struct page_container *);
 void container_rss_del(struct page_container *);
 void container_rss_release(struct page_container *);
+void container_out_of_memory(struct rss_container *);
 
 void mm_init_container(struct mm_struct *mm, struct task_struct *tsk);
 void mm_free_container(struct mm_struct *mm);

81 lines not on this page · raw message

diff -upr linux-2.6.20.orig/mm/vmscan.c linux-2.6.20-2/mm/vmscan.c
@@ -45,6 +45,8 @@
 
 #include "internal.h"
 
+#include <linux/rss_container.h>
+
 struct scan_control {
 	/* Incremented by the number of inactive pages that were scanned */
 	unsigned long nr_scanned;

@@ -66,6 +68,10 @@ struct scan_control {
 	int swappiness;
 
 	int all_unreclaimable;
+	struct rss_container *cnt;
+	unsigned long (*isolate_pages)(unsigned long nr, struct list_head *dst,
+			unsigned long *scanned, struct zone *zone,
+			struct rss_container *cont, int active);
 };
 
 /*

@@ -654,6 +660,17 @@ static unsigned long isolate_lru_pages(u
 	return nr_taken;
 }
 
+static unsigned long isolate_pages_global(unsigned long nr,
+		struct list_head *dst, unsigned long *scanned,
+		struct zone *z, struct rss_container *cont,
+		int active)
+{
+	if (active)
+		return isolate_lru_pages(nr, &z->active_list, dst, scanned);
+	else
+		return isolate_lru_pages(nr, &z->inactive_list, dst, scanned);
+}
+
 /*
  * shrink_inactive_list() is a helper for shrink_zone().  It returns the number
  * of reclaimed pages

@@ -676,9 +693,9 @@ static unsigned long shrink_inactive_lis
 		unsigned long nr_scan;
 		unsigned long nr_freed;
 
-		nr_taken = isolate_lru_pages(sc->swap_cluster_max,
-					     &zone->inactive_list,
-					     &page_list, &nr_scan);
+		nr_taken = sc->isolate_pages(sc->swap_cluster_max, &page_list,
+				&nr_scan, zone, sc->cnt, 0);
+
 		zone->nr_inactive -= nr_taken;
 		zone->pages_scanned += nr_scan;
 		spin_unlock_irq(&zone->lru_lock);

@@ -822,8 +839,8 @@ force_reclaim_mapped:
 
 	lru_add_drain();
 	spin_lock_irq(&zone->lru_lock);
-	pgmoved = isolate_lru_pages(nr_pages, &zone->active_list,
-				    &l_hold, &pgscanned);
+	pgmoved = sc->isolate_pages(nr_pages, &l_hold, &pgscanned, zone,
+			sc->cnt, 1);
 	zone->pages_scanned += pgscanned;
 	zone->nr_active -= pgmoved;
 	spin_unlock_irq(&zone->lru_lock);

@@ -1012,7 +1031,9 @@ static unsigned long shrink_zones(int pr
  * holds filesystem locks which prevent writeout this might not work, and the
  * allocation attempt will fail.
  */
-unsigned long try_to_free_pages(struct zone **zones, gfp_t gfp_mask)
+
+static unsigned long do_try_to_free_pages(struct zone **zones, gfp_t gfp_mask,
+		struct scan_control *sc)
 {
 	int priority;
 	int ret = 0;

@@ -1021,13 +1042,6 @@ unsigned long try_to_free_pages(struct z
 	struct reclaim_state *reclaim_state = current->reclaim_state;
 	unsigned long lru_pages = 0;
 	int i;
-	struct scan_control sc = {
-		.gfp_mask = gfp_mask,
-		.may_writepage = !laptop_mode,
-		.swap_cluster_max = SWAP_CLUSTER_MAX,
-		.may_swap = 1,
-		.swappiness = vm_swappiness,
-	};
 
 	count_vm_event(ALLOCSTALL);
 

@@ -1041,17 +1055,18 @@ unsigned long try_to_free_pages(struct z
 	}
 
 	for (priority = DEF_PRIORITY; priority >= 0; priority--) {
-		sc.nr_scanned = 0;
+		sc->nr_scanned = 0;
 		if (!priority)
 			disable_swap_token();
-		nr_reclaimed += shrink_zones(priority, zones, &sc);
-		shrink_slab(sc.nr_scanned, gfp_mask, lru_pages);
+		nr_reclaimed += shrink_zones(priority, zones, sc);
+		if (sc->cnt == NULL)
+			shrink_slab(sc->nr_scanned, gfp_mask, lru_pages);
 		if (reclaim_state) {
 			nr_reclaimed += reclaim_state->reclaimed_slab;
 			reclaim_state->reclaimed_slab = 0;
 		}
-		total_scanned += sc.nr_scanned;
-		if (nr_reclaimed >= sc.swap_cluster_max) {
+		total_scanned += sc->nr_scanned;
+		if (nr_reclaimed >= sc->swap_cluster_max) {
 			ret = 1;
 			goto out;
 		}

@@ -1063,18 +1078,18 @@ unsigned long try_to_free_pages(struct z
 		 * that's undesirable in laptop mode, where we *want* lumpy
 		 * writeout.  So in laptop mode, write out the whole world.
 		 */
-		if (total_scanned > sc.swap_cluster_max +
-					sc.swap_cluster_max / 2) {
+		if (total_scanned > sc->swap_cluster_max +
+					sc->swap_cluster_max / 2) {
 			wakeup_pdflush(laptop_mode ? 0 : total_scanned);
-			sc.may_writepage = 1;
+			sc->may_writepage = 1;
 		}
 
 		/* Take a nap, wait for some writeback to complete */
-		if (sc.nr_scanned && priority < DEF_PRIORITY - 2)
+		if (sc->nr_scanned && priority < DEF_PRIORITY - 2)
 			congestion_wait(WRITE, HZ/10);
 	}
 	/* top priority shrink_caches still had more to do? don't OOM, then */
-	if (!sc.all_unreclaimable)
+	if (!sc->all_unreclaimable)
 		ret = 1;
 out:
 	/*

@@ -1097,6 +1112,21 @@ out:
 	return ret;
 }
 
+unsigned long try_to_free_pages(struct zone **zones, gfp_t gfp_mask)
+{
+	struct scan_control sc = {
+		.gfp_mask = gfp_mask,
+		.may_writepage = !laptop_mode,
+		.swap_cluster_max = SWAP_CLUSTER_MAX,
+		.may_swap = 1,
+		.swappiness = vm_swappiness,
+		.cnt = NULL,
+		.isolate_pages = isolate_pages_global,
+	};
+
+	return do_try_to_free_pages(zones, gfp_mask, &sc);
+}
+
 /*
  * For kswapd, balance_pgdat() will work across all this node's zones until
  * they are all at pages_high.

@@ -1131,6 +1190,8 @@ static unsigned long balance_pgdat(pg_da
 		.may_swap = 1,
 		.swap_cluster_max = SWAP_CLUSTER_MAX,
 		.swappiness = vm_swappiness,
+		.cnt = NULL,
+		.isolate_pages = isolate_pages_global,
 	};
 	/*
 	 * temp_priority is used to remember the scanning priority at which

@@ -1436,6 +1497,8 @@ unsigned long shrink_all_memory(unsigned
 		.swap_cluster_max = nr_pages,
 		.may_writepage = 1,
 		.swappiness = vm_swappiness,
+		.cnt = NULL,
+		.isolate_pages = isolate_pages_global,
 	};
 
 	current->reclaim_state = &reclaim_state;

@@ -1619,6 +1682,8 @@ static int __zone_reclaim(struct zone *z
 					SWAP_CLUSTER_MAX),
 		.gfp_mask = gfp_mask,
 		.swappiness = vm_swappiness,
+		.cnt = NULL,
+		.isolate_pages = isolate_pages_global,
 	};
 	unsigned long slab_reclaimable;
 

diff -upr linux-2.6.20.orig/include/linux/swap.h linux-2.6.20-2/include/linux/swap.h
@@ -187,6 +187,9 @@ extern void swap_setup(void);
 
 /* linux/mm/vmscan.c */
 extern unsigned long try_to_free_pages(struct zone **, gfp_t);
+
+struct rss_container;
+extern unsigned long try_to_free_pages_in_container(struct rss_container *);
 extern unsigned long shrink_all_memory(unsigned long nr_pages);
 extern int vm_swappiness;
 extern int remove_mapping(struct address_space *mapping, struct page *page);

diff -upr linux-2.6.20.orig/mm/vmscan.c linux-2.6.20-2/mm/vmscan.c
@@ -872,6 +872,7 @@ force_reclaim_mapped:
 		ClearPageActive(page);
 
 		list_move(&page->lru, &zone->inactive_list);
+		container_rss_move_lists(page, 0);
 		pgmoved++;
 		if (!pagevec_add(&pvec, page)) {
 			zone->nr_inactive += pgmoved;

@@ -900,6 +901,7 @@ force_reclaim_mapped:
 		SetPageLRU(page);
 		VM_BUG_ON(!PageActive(page));
 		list_move(&page->lru, &zone->active_list);
+		container_rss_move_lists(page, 1);
 		pgmoved++;
 		if (!pagevec_add(&pvec, page)) {
 			zone->nr_active += pgmoved;

@@ -1110,6 +1112,35 @@ out:
 	return ret;
 }
 
+#ifdef CONFIG_RSS_CONTAINER
+unsigned long try_to_free_pages_in_container(struct rss_container *cnt)
+{
+	struct scan_control sc = {
+		.gfp_mask = GFP_KERNEL,
+		.may_writepage = 1,
+		.swap_cluster_max = 1,
+		.may_swap = 1,
+		.swappiness = vm_swappiness,
+		.cnt = cnt,
+		.isolate_pages = isolate_pages_in_container,
+	};
+	int node;
+	struct zone **zones;
+
+	for_each_node(node) {
+#ifdef CONFIG_HIGHMEM
+		zones = NODE_DATA(node)->node_zonelists[ZONE_HIGHMEM].zones;
+		if (do_try_to_free_pages(zones, sc.gfp_mask, &sc))
+			return 1;
+#endif
+		zones = NODE_DATA(node)->node_zonelists[ZONE_NORMAL].zones;
+		if (do_try_to_free_pages(zones, sc.gfp_mask, &sc))
+			return 1;
+	}
+	return 0;
+}
+#endif
+
 unsigned long try_to_free_pages(struct zone **zones, gfp_t gfp_mask)
 {
 	struct scan_control sc = {

diff -upr linux-2.6.20.orig/mm/swap.c linux-2.6.20-2/mm/swap.c
@@ -30,6 +30,7 @@
 #include <linux/cpu.h>
 #include <linux/notifier.h>
 #include <linux/init.h>
+#include <linux/rss_container.h>
 
 /* How many pages do we try to swap or page in/out together? */
 int page_cluster;

@@ -147,6 +148,7 @@ void fastcall activate_page(struct page 
 		SetPageActive(page);
 		add_page_to_active_list(zone, page);
 		__count_vm_event(PGACTIVATE);
+		container_rss_move_lists(page, 1);
 	}
 	spin_unlock_irq(&zone->lru_lock);
 }

78 lines not on this page · raw message

diff -upr linux-2.6.20.orig/include/linux/rss_container.h linux-2.6.20-2/include/linux/rss_container.h
@@ -24,6 +24,10 @@
 void mm_init_container(struct mm_struct *mm, struct task_struct *tsk);
 void mm_free_container(struct mm_struct *mm);
 
+void container_rss_move_lists(struct page *pg, bool active);
+unsigned long isolate_pages_in_container(unsigned long nr_to_scan,
+		struct list_head *dst, unsigned long *scanned,
+		struct zone *zone, struct rss_container *, int active);
 #else
 static inline int container_rss_prepare(struct page *pg,
 		struct vm_area_struct *vma, struct page_container **pc)

@@ -52,5 +56,7 @@
 {
 }
 
+#define isolate_container_pages(nr, dst, scanned, rss, act, zone) ({ BUG(); 0;})
+#define container_rss_move_lists(pg, active) do { } while (0)
 #endif
 #endif

Index: linux-2.6.20/mm/rss_container.c
===================================================================
@@ -212,6 +212,7 @@ void container_rss_del(struct page_conta
 
 	css_put(&rss->css);
 	kfree(pc);
+	init_page_container(page);
 }
 
 static void rss_move_task(struct container_subsys *ss,

Index: linux-2.6.20/mm/page_alloc.c
===================================================================
@@ -41,6 +41,7 @@
 #include <linux/pfn.h>
 #include <linux/backing-dev.h>
 #include <linux/fault-inject.h>
+#include <linux/rss_container.h>
 
 #include <asm/tlbflush.h>
 #include <asm/div64.h>

@@ -1977,6 +1978,7 @@ void __meminit memmap_init_zone(unsigned
 		set_page_links(page, zone, nid, pfn);
 		init_page_count(page);
 		reset_page_mapcount(page);
+		init_page_container(page);
 		SetPageReserved(page);
 		INIT_LIST_HEAD(&page->lru);
 #ifdef WANT_PAGE_VIRTUAL

Index: linux-2.6.20/include/linux/rss_container.h
===================================================================
@@ -28,6 +28,11 @@ void container_rss_move_lists(struct pag
 unsigned long isolate_pages_in_container(unsigned long nr_to_scan,
 		struct list_head *dst, unsigned long *scanned,
 		struct zone *zone, struct rss_container *, int active);
+static inline void init_page_container(struct page *page)
+{
+	page_container(page) = NULL;
+}
+
 #else
 static inline int container_rss_prepare(struct page *pg,
 		struct vm_area_struct *vma, struct page_container **pc)

@@ -56,6 +61,10 @@ static inline void mm_free_container(str
 {
 }
 
+static inline void init_page_container(struct page *page)
+{
+}
+
 #define isolate_container_pages(nr, dst, scanned, rss, act, zone) ({ BUG(); 0;})
 #define container_rss_move_lists(pg, active) do { } while (0)
 #endif

Index: linux-2.6.20/mm/page_alloc.c
===================================================================
@@ -41,6 +41,7 @@
 #include <linux/pfn.h>
 #include <linux/backing-dev.h>
 #include <linux/fault-inject.h>
+#include <linux/rss_container.h>
 
 #include <asm/tlbflush.h>
 #include <asm/div64.h>

@@ -791,6 +792,7 @@ static void fastcall free_hot_cold_page(
 
 	if (!PageHighMem(page))
 		debug_check_no_locks_freed(page_address(page), PAGE_SIZE);
+	init_page_container(page);
 	arch_free_page(page, 0);
 	kernel_map_pages(page, 1, 0);
 

@@ -1977,6 +1979,7 @@ void __meminit memmap_init_zone(unsigned
 		set_page_links(page, zone, nid, pfn);
 		init_page_count(page);
 		reset_page_mapcount(page);
+		init_page_container(page);
 		SetPageReserved(page);
 		INIT_LIST_HEAD(&page->lru);
 #ifdef WANT_PAGE_VIRTUAL

Index: linux-2.6.20/include/linux/rss_container.h
===================================================================
@@ -28,6 +28,11 @@ void container_rss_move_lists(struct pag
 unsigned long isolate_pages_in_container(unsigned long nr_to_scan,
 		struct list_head *dst, unsigned long *scanned,
 		struct zone *zone, struct rss_container *, int active);
+static inline void init_page_container(struct page *page)
+{
+	page_container(page) = NULL;
+}
+
 #else
 static inline int container_rss_prepare(struct page *pg,
 		struct vm_area_struct *vma, struct page_container **pc)

@@ -56,6 +61,10 @@ static inline void mm_free_container(str
 {
 }
 
+static inline void init_page_container(struct page *page)
+{
+}
+
 #define isolate_container_pages(nr, dst, scanned, rss, act, zone) ({ BUG(); 0;})
 #define container_rss_move_lists(pg, active) do { } while (0)
 #endif

Index: linux-2.6.20/mm/rss_container.c
===================================================================
@@ -199,12 +199,9 @@ void container_rss_add(struct page_conta
 
 void container_rss_del(struct page_container *pc)
 {
-	struct page *page;
 	struct rss_container *rss;
 
-	page = pc->page;
 	rss = pc->cnt;
-
 	spin_lock_irq(&rss->res.lock);
 	list_del(&pc->list);
 	res_counter_uncharge_locked(&rss->res, 1);