Memory controller: memory accounting
Add the accounting hooks. The accounting is carried out for RSS and Page
Cache (unmapped) pages. There is now a common limit and accounting for both.
The RSS accounting is accounted at page_add_*_rmap() and page_remove_rmap()
time. Page cache is accounted at add_to_page_cache(),
__delete_from_page_cache(). Swap cache is also accounted for.
Each page's page_cgroup is protected with the last bit of the
page_cgroup pointer, this makes handling of race conditions involving
simultaneous mappings of a page easier. A reference count is kept in the
page_cgroup to deal with cases where a page might be unmapped from the RSS
of all tasks, but still lives in the page cache.
Credits go to Vaidyanathan Srinivasan for helping with reference counting work
of the page cgroup. Almost all of the page cache accounting code has help
from Vaidyanathan Srinivasan.
[hugh@veritas.com: fix swapoff breakage]
[akpm@linux-foundation.org: fix locking]
Signed-off-by: Vaidyanathan Srinivasan <svaidy@linux.vnet.ibm.com>
Signed-off-by: Balbir Singh <balbir@linux.vnet.ibm.com>
Cc: Pavel Emelianov <xemul@openvz.org>
Cc: Paul Menage <menage@google.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Nick Piggin <nickpiggin@yahoo.com.au>
Cc: Kirill Korotaev <dev@sw.ru>
Cc: Herbert Poetzl <herbert@13thfloor.at>
Cc: David Rientjes <rientjes@google.com>
Cc: <Valdis.Kletnieks@vt.edu>
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
diff --git a/mm/filemap.c b/mm/filemap.c
index 81fb9bf..b7a01e9 100644
--- a/mm/filemap.c
+++ b/mm/filemap.c
@@ -33,6 +33,7 @@
#include <linux/syscalls.h>
#include <linux/cpuset.h>
#include <linux/hardirq.h> /* for BUG_ON(!in_atomic()) only */
+#include <linux/memcontrol.h>
#include "internal.h"
/*
@@ -118,6 +119,7 @@
{
struct address_space *mapping = page->mapping;
+ mem_cgroup_uncharge_page(page);
radix_tree_delete(&mapping->page_tree, page->index);
page->mapping = NULL;
mapping->nrpages--;
@@ -461,6 +463,11 @@
int error = radix_tree_preload(gfp_mask & ~__GFP_HIGHMEM);
if (error == 0) {
+
+ error = mem_cgroup_charge(page, current->mm);
+ if (error)
+ goto out;
+
write_lock_irq(&mapping->tree_lock);
error = radix_tree_insert(&mapping->page_tree, offset, page);
if (!error) {
@@ -470,10 +477,13 @@
page->index = offset;
mapping->nrpages++;
__inc_zone_page_state(page, NR_FILE_PAGES);
- }
+ } else
+ mem_cgroup_uncharge_page(page);
+
write_unlock_irq(&mapping->tree_lock);
radix_tree_preload_end();
}
+out:
return error;
}
EXPORT_SYMBOL(add_to_page_cache);
diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index 4d4805e..ebca767 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -21,6 +21,9 @@
#include <linux/memcontrol.h>
#include <linux/cgroup.h>
#include <linux/mm.h>
+#include <linux/page-flags.h>
+#include <linux/bit_spinlock.h>
+#include <linux/rcupdate.h>
struct cgroup_subsys mem_cgroup_subsys;
@@ -31,7 +34,9 @@
* to help the administrator determine what knobs to tune.
*
* TODO: Add a water mark for the memory controller. Reclaim will begin when
- * we hit the water mark.
+ * we hit the water mark. May be even add a low water mark, such that
+ * no reclaim occurs from a cgroup at it's low water mark, this is
+ * a feature that will be implemented much later in the future.
*/
struct mem_cgroup {
struct cgroup_subsys_state css;
@@ -49,6 +54,14 @@
};
/*
+ * We use the lower bit of the page->page_cgroup pointer as a bit spin
+ * lock. We need to ensure that page->page_cgroup is atleast two
+ * byte aligned (based on comments from Nick Piggin)
+ */
+#define PAGE_CGROUP_LOCK_BIT 0x0
+#define PAGE_CGROUP_LOCK (1 << PAGE_CGROUP_LOCK_BIT)
+
+/*
* A page_cgroup page is associated with every page descriptor. The
* page_cgroup helps us identify information about the cgroup
*/
@@ -56,6 +69,8 @@
struct list_head lru; /* per cgroup LRU list */
struct page *page;
struct mem_cgroup *mem_cgroup;
+ atomic_t ref_cnt; /* Helpful when pages move b/w */
+ /* mapped and cached states */
};
@@ -88,14 +103,157 @@
css_put(&mm->mem_cgroup->css);
}
+static inline int page_cgroup_locked(struct page *page)
+{
+ return bit_spin_is_locked(PAGE_CGROUP_LOCK_BIT,
+ &page->page_cgroup);
+}
+
void page_assign_page_cgroup(struct page *page, struct page_cgroup *pc)
{
- page->page_cgroup = (unsigned long)pc;
+ int locked;
+
+ /*
+ * While resetting the page_cgroup we might not hold the
+ * page_cgroup lock. free_hot_cold_page() is an example
+ * of such a scenario
+ */
+ if (pc)
+ VM_BUG_ON(!page_cgroup_locked(page));
+ locked = (page->page_cgroup & PAGE_CGROUP_LOCK);
+ page->page_cgroup = ((unsigned long)pc | locked);
}
struct page_cgroup *page_get_page_cgroup(struct page *page)
{
- return page->page_cgroup;
+ return (struct page_cgroup *)
+ (page->page_cgroup & ~PAGE_CGROUP_LOCK);
+}
+
+void __always_inline lock_page_cgroup(struct page *page)
+{
+ bit_spin_lock(PAGE_CGROUP_LOCK_BIT, &page->page_cgroup);
+ VM_BUG_ON(!page_cgroup_locked(page));
+}
+
+void __always_inline unlock_page_cgroup(struct page *page)
+{
+ bit_spin_unlock(PAGE_CGROUP_LOCK_BIT, &page->page_cgroup);
+}
+
+/*
+ * Charge the memory controller for page usage.
+ * Return
+ * 0 if the charge was successful
+ * < 0 if the cgroup is over its limit
+ */
+int mem_cgroup_charge(struct page *page, struct mm_struct *mm)
+{
+ struct mem_cgroup *mem;
+ struct page_cgroup *pc, *race_pc;
+
+ /*
+ * Should page_cgroup's go to their own slab?
+ * One could optimize the performance of the charging routine
+ * by saving a bit in the page_flags and using it as a lock
+ * to see if the cgroup page already has a page_cgroup associated
+ * with it
+ */
+ lock_page_cgroup(page);
+ pc = page_get_page_cgroup(page);
+ /*
+ * The page_cgroup exists and the page has already been accounted
+ */
+ if (pc) {
+ atomic_inc(&pc->ref_cnt);
+ goto done;
+ }
+
+ unlock_page_cgroup(page);
+
+ pc = kzalloc(sizeof(struct page_cgroup), GFP_KERNEL);
+ if (pc == NULL)
+ goto err;
+
+ rcu_read_lock();
+ /*
+ * We always charge the cgroup the mm_struct belongs to
+ * the mm_struct's mem_cgroup changes on task migration if the
+ * thread group leader migrates. It's possible that mm is not
+ * set, if so charge the init_mm (happens for pagecache usage).
+ */
+ if (!mm)
+ mm = &init_mm;
+
+ mem = rcu_dereference(mm->mem_cgroup);
+ /*
+ * For every charge from the cgroup, increment reference
+ * count
+ */
+ css_get(&mem->css);
+ rcu_read_unlock();
+
+ /*
+ * If we created the page_cgroup, we should free it on exceeding
+ * the cgroup limit.
+ */
+ if (res_counter_charge(&mem->res, 1)) {
+ css_put(&mem->css);
+ goto free_pc;
+ }
+
+ lock_page_cgroup(page);
+ /*
+ * Check if somebody else beat us to allocating the page_cgroup
+ */
+ race_pc = page_get_page_cgroup(page);
+ if (race_pc) {
+ kfree(pc);
+ pc = race_pc;
+ atomic_inc(&pc->ref_cnt);
+ res_counter_uncharge(&mem->res, 1);
+ css_put(&mem->css);
+ goto done;
+ }
+
+ atomic_set(&pc->ref_cnt, 1);
+ pc->mem_cgroup = mem;
+ pc->page = page;
+ page_assign_page_cgroup(page, pc);
+
+done:
+ unlock_page_cgroup(page);
+ return 0;
+free_pc:
+ kfree(pc);
+ return -ENOMEM;
+err:
+ unlock_page_cgroup(page);
+ return -ENOMEM;
+}
+
+/*
+ * Uncharging is always a welcome operation, we never complain, simply
+ * uncharge.
+ */
+void mem_cgroup_uncharge(struct page_cgroup *pc)
+{
+ struct mem_cgroup *mem;
+ struct page *page;
+
+ if (!pc)
+ return;
+
+ if (atomic_dec_and_test(&pc->ref_cnt)) {
+ page = pc->page;
+ lock_page_cgroup(page);
+ mem = pc->mem_cgroup;
+ css_put(&mem->css);
+ page_assign_page_cgroup(page, NULL);
+ unlock_page_cgroup(page);
+ res_counter_uncharge(&mem->res, 1);
+ kfree(pc);
+ }
}
static ssize_t mem_cgroup_read(struct cgroup *cont, struct cftype *cft,
@@ -150,6 +308,8 @@
return NULL;
res_counter_init(&mem->res);
+ INIT_LIST_HEAD(&mem->active_list);
+ INIT_LIST_HEAD(&mem->inactive_list);
return &mem->css;
}
diff --git a/mm/memory.c b/mm/memory.c
index 9d073fa..0ba224e 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -50,6 +50,7 @@
#include <linux/delayacct.h>
#include <linux/init.h>
#include <linux/writeback.h>
+#include <linux/memcontrol.h>
#include <asm/pgalloc.h>
#include <asm/uaccess.h>
@@ -1144,16 +1145,20 @@
{
int retval;
pte_t *pte;
- spinlock_t *ptl;
+ spinlock_t *ptl;
+
+ retval = mem_cgroup_charge(page, mm);
+ if (retval)
+ goto out;
retval = -EINVAL;
if (PageAnon(page))
- goto out;
+ goto out_uncharge;
retval = -ENOMEM;
flush_dcache_page(page);
pte = get_locked_pte(mm, addr, &ptl);
if (!pte)
- goto out;
+ goto out_uncharge;
retval = -EBUSY;
if (!pte_none(*pte))
goto out_unlock;
@@ -1165,8 +1170,12 @@
set_pte_at(mm, addr, pte, mk_pte(page, prot));
retval = 0;
+ pte_unmap_unlock(pte, ptl);
+ return retval;
out_unlock:
pte_unmap_unlock(pte, ptl);
+out_uncharge:
+ mem_cgroup_uncharge_page(page);
out:
return retval;
}
@@ -1641,6 +1650,9 @@
cow_user_page(new_page, old_page, address, vma);
__SetPageUptodate(new_page);
+ if (mem_cgroup_charge(new_page, mm))
+ goto oom_free_new;
+
/*
* Re-check the pte - we dropped the lock
*/
@@ -1672,7 +1684,9 @@
/* Free the old page.. */
new_page = old_page;
ret |= VM_FAULT_WRITE;
- }
+ } else
+ mem_cgroup_uncharge_page(new_page);
+
if (new_page)
page_cache_release(new_page);
if (old_page)
@@ -1696,6 +1710,8 @@
put_page(dirty_page);
}
return ret;
+oom_free_new:
+ __free_page(new_page);
oom:
if (old_page)
page_cache_release(old_page);
@@ -2036,6 +2052,12 @@
count_vm_event(PGMAJFAULT);
}
+ if (mem_cgroup_charge(page, mm)) {
+ delayacct_clear_flag(DELAYACCT_PF_SWAPIN);
+ ret = VM_FAULT_OOM;
+ goto out;
+ }
+
mark_page_accessed(page);
lock_page(page);
delayacct_clear_flag(DELAYACCT_PF_SWAPIN);
@@ -2073,8 +2095,10 @@
if (write_access) {
/* XXX: We could OR the do_wp_page code with this one? */
if (do_wp_page(mm, vma, address,
- page_table, pmd, ptl, pte) & VM_FAULT_OOM)
+ page_table, pmd, ptl, pte) & VM_FAULT_OOM) {
+ mem_cgroup_uncharge_page(page);
ret = VM_FAULT_OOM;
+ }
goto out;
}
@@ -2085,6 +2109,7 @@
out:
return ret;
out_nomap:
+ mem_cgroup_uncharge_page(page);
pte_unmap_unlock(page_table, ptl);
unlock_page(page);
page_cache_release(page);
@@ -2114,6 +2139,9 @@
goto oom;
__SetPageUptodate(page);
+ if (mem_cgroup_charge(page, mm))
+ goto oom_free_page;
+
entry = mk_pte(page, vma->vm_page_prot);
entry = maybe_mkwrite(pte_mkdirty(entry), vma);
@@ -2131,8 +2159,11 @@
pte_unmap_unlock(page_table, ptl);
return 0;
release:
+ mem_cgroup_uncharge_page(page);
page_cache_release(page);
goto unlock;
+oom_free_page:
+ __free_page(page);
oom:
return VM_FAULT_OOM;
}
@@ -2246,6 +2277,11 @@
}
+ if (mem_cgroup_charge(page, mm)) {
+ ret = VM_FAULT_OOM;
+ goto out;
+ }
+
page_table = pte_offset_map_lock(mm, pmd, address, &ptl);
/*
@@ -2281,6 +2317,7 @@
/* no need to invalidate: a not-present page won't be cached */
update_mmu_cache(vma, address, entry);
} else {
+ mem_cgroup_uncharge_page(page);
if (anon)
page_cache_release(page);
else
diff --git a/mm/migrate.c b/mm/migrate.c
index 857a987..417bbda 100644
--- a/mm/migrate.c
+++ b/mm/migrate.c
@@ -29,6 +29,7 @@
#include <linux/mempolicy.h>
#include <linux/vmalloc.h>
#include <linux/security.h>
+#include <linux/memcontrol.h>
#include "internal.h"
@@ -152,6 +153,11 @@
return;
}
+ if (mem_cgroup_charge(new, mm)) {
+ pte_unmap(ptep);
+ return;
+ }
+
ptl = pte_lockptr(mm, pmd);
spin_lock(ptl);
pte = *ptep;
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 37576b8..26a54a1 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -43,6 +43,7 @@
#include <linux/backing-dev.h>
#include <linux/fault-inject.h>
#include <linux/page-isolation.h>
+#include <linux/memcontrol.h>
#include <asm/tlbflush.h>
#include <asm/div64.h>
@@ -987,6 +988,7 @@
if (!PageHighMem(page))
debug_check_no_locks_freed(page_address(page), PAGE_SIZE);
+ VM_BUG_ON(page_get_page_cgroup(page));
arch_free_page(page, 0);
kernel_map_pages(page, 1, 0);
@@ -2525,6 +2527,7 @@
set_page_links(page, zone, nid, pfn);
init_page_count(page);
reset_page_mapcount(page);
+ page_assign_page_cgroup(page, NULL);
SetPageReserved(page);
/*
diff --git a/mm/rmap.c b/mm/rmap.c
index 57ad276..4a34879 100644
--- a/mm/rmap.c
+++ b/mm/rmap.c
@@ -48,6 +48,7 @@
#include <linux/rcupdate.h>
#include <linux/module.h>
#include <linux/kallsyms.h>
+#include <linux/memcontrol.h>
#include <asm/tlbflush.h>
@@ -554,8 +555,14 @@
VM_BUG_ON(address < vma->vm_start || address >= vma->vm_end);
if (atomic_inc_and_test(&page->_mapcount))
__page_set_anon_rmap(page, vma, address);
- else
+ else {
__page_check_anon_rmap(page, vma, address);
+ /*
+ * We unconditionally charged during prepare, we uncharge here
+ * This takes care of balancing the reference counts
+ */
+ mem_cgroup_uncharge_page(page);
+ }
}
/*
@@ -586,6 +593,12 @@
{
if (atomic_inc_and_test(&page->_mapcount))
__inc_zone_page_state(page, NR_FILE_MAPPED);
+ else
+ /*
+ * We unconditionally charged during prepare, we uncharge here
+ * This takes care of balancing the reference counts
+ */
+ mem_cgroup_uncharge_page(page);
}
#ifdef CONFIG_DEBUG_VM
@@ -646,6 +659,8 @@
page_clear_dirty(page);
set_page_dirty(page);
}
+ mem_cgroup_uncharge_page(page);
+
__dec_zone_page_state(page,
PageAnon(page) ? NR_ANON_PAGES : NR_FILE_MAPPED);
}
diff --git a/mm/swap_state.c b/mm/swap_state.c
index ec42f01..f96e3ff 100644
--- a/mm/swap_state.c
+++ b/mm/swap_state.c
@@ -17,6 +17,7 @@
#include <linux/backing-dev.h>
#include <linux/pagevec.h>
#include <linux/migrate.h>
+#include <linux/memcontrol.h>
#include <asm/pgtable.h>
@@ -76,6 +77,11 @@
BUG_ON(PagePrivate(page));
error = radix_tree_preload(gfp_mask);
if (!error) {
+
+ error = mem_cgroup_charge(page, current->mm);
+ if (error)
+ goto out;
+
write_lock_irq(&swapper_space.tree_lock);
error = radix_tree_insert(&swapper_space.page_tree,
entry.val, page);
@@ -86,10 +92,13 @@
total_swapcache_pages++;
__inc_zone_page_state(page, NR_FILE_PAGES);
INC_CACHE_INFO(add_total);
+ } else {
+ mem_cgroup_uncharge_page(page);
}
write_unlock_irq(&swapper_space.tree_lock);
radix_tree_preload_end();
}
+out:
return error;
}
@@ -104,6 +113,7 @@
BUG_ON(PageWriteback(page));
BUG_ON(PagePrivate(page));
+ mem_cgroup_uncharge_page(page);
radix_tree_delete(&swapper_space.page_tree, page_private(page));
set_page_private(page, 0);
ClearPageSwapCache(page);
diff --git a/mm/swapfile.c b/mm/swapfile.c
index afae7b1..fddc4cc 100644
--- a/mm/swapfile.c
+++ b/mm/swapfile.c
@@ -27,6 +27,7 @@
#include <linux/mutex.h>
#include <linux/capability.h>
#include <linux/syscalls.h>
+#include <linux/memcontrol.h>
#include <asm/pgtable.h>
#include <asm/tlbflush.h>
@@ -506,9 +507,12 @@
* just let do_wp_page work it out if a write is requested later - to
* force COW, vm_page_prot omits write permission from any private vma.
*/
-static void unuse_pte(struct vm_area_struct *vma, pte_t *pte,
+static int unuse_pte(struct vm_area_struct *vma, pte_t *pte,
unsigned long addr, swp_entry_t entry, struct page *page)
{
+ if (mem_cgroup_charge(page, vma->vm_mm))
+ return -ENOMEM;
+
inc_mm_counter(vma->vm_mm, anon_rss);
get_page(page);
set_pte_at(vma->vm_mm, addr, pte,
@@ -520,6 +524,7 @@
* immediately swapped out again after swapon.
*/
activate_page(page);
+ return 1;
}
static int unuse_pte_range(struct vm_area_struct *vma, pmd_t *pmd,
@@ -529,7 +534,7 @@
pte_t swp_pte = swp_entry_to_pte(entry);
pte_t *pte;
spinlock_t *ptl;
- int found = 0;
+ int ret = 0;
pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl);
do {
@@ -538,13 +543,12 @@
* Test inline before going to call unuse_pte.
*/
if (unlikely(pte_same(*pte, swp_pte))) {
- unuse_pte(vma, pte++, addr, entry, page);
- found = 1;
+ ret = unuse_pte(vma, pte++, addr, entry, page);
break;
}
} while (pte++, addr += PAGE_SIZE, addr != end);
pte_unmap_unlock(pte - 1, ptl);
- return found;
+ return ret;
}
static inline int unuse_pmd_range(struct vm_area_struct *vma, pud_t *pud,
@@ -553,14 +557,16 @@
{
pmd_t *pmd;
unsigned long next;
+ int ret;
pmd = pmd_offset(pud, addr);
do {
next = pmd_addr_end(addr, end);
if (pmd_none_or_clear_bad(pmd))
continue;
- if (unuse_pte_range(vma, pmd, addr, next, entry, page))
- return 1;
+ ret = unuse_pte_range(vma, pmd, addr, next, entry, page);
+ if (ret)
+ return ret;
} while (pmd++, addr = next, addr != end);
return 0;
}
@@ -571,14 +577,16 @@
{
pud_t *pud;
unsigned long next;
+ int ret;
pud = pud_offset(pgd, addr);
do {
next = pud_addr_end(addr, end);
if (pud_none_or_clear_bad(pud))
continue;
- if (unuse_pmd_range(vma, pud, addr, next, entry, page))
- return 1;
+ ret = unuse_pmd_range(vma, pud, addr, next, entry, page);
+ if (ret)
+ return ret;
} while (pud++, addr = next, addr != end);
return 0;
}
@@ -588,6 +596,7 @@
{
pgd_t *pgd;
unsigned long addr, end, next;
+ int ret;
if (page->mapping) {
addr = page_address_in_vma(page, vma);
@@ -605,8 +614,9 @@
next = pgd_addr_end(addr, end);
if (pgd_none_or_clear_bad(pgd))
continue;
- if (unuse_pud_range(vma, pgd, addr, next, entry, page))
- return 1;
+ ret = unuse_pud_range(vma, pgd, addr, next, entry, page);
+ if (ret)
+ return ret;
} while (pgd++, addr = next, addr != end);
return 0;
}
@@ -615,6 +625,7 @@
swp_entry_t entry, struct page *page)
{
struct vm_area_struct *vma;
+ int ret = 0;
if (!down_read_trylock(&mm->mmap_sem)) {
/*
@@ -627,15 +638,11 @@
lock_page(page);
}
for (vma = mm->mmap; vma; vma = vma->vm_next) {
- if (vma->anon_vma && unuse_vma(vma, entry, page))
+ if (vma->anon_vma && (ret = unuse_vma(vma, entry, page)))
break;
}
up_read(&mm->mmap_sem);
- /*
- * Currently unuse_mm cannot fail, but leave error handling
- * at call sites for now, since we change it from time to time.
- */
- return 0;
+ return (ret < 0)? ret: 0;
}
/*