2011-02-17 17:15:18

by Peter Zijlstra

[permalink] [raw]
Subject: [PATCH 3/3] mm: Simplify anon_vma refcounts

This patch changes the anon_vma refcount to be 0 when the object is
free. It does this by adding 1 ref to being in use in the anon_vma
structure (iow. the anon_vma->head list is not empty).

This allows a simpler release scheme without having to check both the
refcount and the list as well as avoids taking a ref for each entry
on the list.

Reviewed-by: KAMEZAWA Hiroyuki <[email protected]>
Acked-by: Hugh Dickins <[email protected]>
Signed-off-by: Peter Zijlstra <[email protected]>
---
include/linux/rmap.h | 11 +++++--
mm/rmap.c | 79 ++++++++++++++++++---------------------------------
2 files changed, 37 insertions(+), 53 deletions(-)

Index: linux-2.6/include/linux/rmap.h
===================================================================
--- linux-2.6.orig/include/linux/rmap.h
+++ linux-2.6/include/linux/rmap.h
@@ -73,7 +73,13 @@ static inline void get_anon_vma(struct a
atomic_inc(&anon_vma->refcount);
}

-void put_anon_vma(struct anon_vma *);
+void __put_anon_vma(struct anon_vma *anon_vma);
+
+static inline void put_anon_vma(struct anon_vma *anon_vma)
+{
+ if (atomic_dec_and_test(&anon_vma->refcount))
+ __put_anon_vma(anon_vma);
+}

static inline struct anon_vma *page_anon_vma(struct page *page)
{
@@ -116,7 +122,6 @@ void unlink_anon_vmas(struct vm_area_str
int anon_vma_clone(struct vm_area_struct *, struct vm_area_struct *);
int anon_vma_fork(struct vm_area_struct *, struct vm_area_struct *);
void __anon_vma_link(struct vm_area_struct *);
-void anon_vma_free(struct anon_vma *);

static inline void anon_vma_merge(struct vm_area_struct *vma,
struct vm_area_struct *next)
@@ -125,6 +130,8 @@ static inline void anon_vma_merge(struct
unlink_anon_vmas(next);
}

+struct anon_vma *page_get_anon_vma(struct page *page);
+
/*
* rmap interfaces called when adding or removing pte of page
*/
Index: linux-2.6/mm/rmap.c
===================================================================
--- linux-2.6.orig/mm/rmap.c
+++ linux-2.6/mm/rmap.c
@@ -67,11 +67,24 @@ static struct kmem_cache *anon_vma_chain

static inline struct anon_vma *anon_vma_alloc(void)
{
- return kmem_cache_alloc(anon_vma_cachep, GFP_KERNEL);
+ struct anon_vma *anon_vma;
+
+ anon_vma = kmem_cache_alloc(anon_vma_cachep, GFP_KERNEL);
+ if (anon_vma) {
+ atomic_set(&anon_vma->refcount, 1);
+ /*
+ * Initialise the anon_vma root to point to itself. If called
+ * from fork, the root will be reset to the parents anon_vma.
+ */
+ anon_vma->root = anon_vma;
+ }
+
+ return anon_vma;
}

-void anon_vma_free(struct anon_vma *anon_vma)
+static inline void anon_vma_free(struct anon_vma *anon_vma)
{
+ VM_BUG_ON(atomic_read(&anon_vma->refcount));
kmem_cache_free(anon_vma_cachep, anon_vma);
}

@@ -133,11 +146,6 @@ int anon_vma_prepare(struct vm_area_stru
if (unlikely(!anon_vma))
goto out_enomem_free_avc;
allocated = anon_vma;
- /*
- * This VMA had no anon_vma yet. This anon_vma is
- * the root of any anon_vma tree that might form.
- */
- anon_vma->root = anon_vma;
}

anon_vma_lock(anon_vma);
@@ -156,7 +164,7 @@ int anon_vma_prepare(struct vm_area_stru
anon_vma_unlock(anon_vma);

if (unlikely(allocated))
- anon_vma_free(allocated);
+ put_anon_vma(allocated);
if (unlikely(avc))
anon_vma_chain_free(avc);
}
@@ -241,9 +249,9 @@ int anon_vma_fork(struct vm_area_struct
*/
anon_vma->root = pvma->anon_vma->root;
/*
- * With KSM refcounts, an anon_vma can stay around longer than the
- * process it belongs to. The root anon_vma needs to be pinned
- * until this anon_vma is freed, because the lock lives in the root.
+ * With refcounts, an anon_vma can stay around longer than the
+ * process it belongs to. The root anon_vma needs to be pinned until
+ * this anon_vma is freed, because the lock lives in the root.
*/
get_anon_vma(anon_vma->root);
/* Mark this anon_vma as the one where our new (COWed) pages go. */
@@ -253,7 +261,7 @@ int anon_vma_fork(struct vm_area_struct
return 0;

out_error_free_anon_vma:
- anon_vma_free(anon_vma);
+ put_anon_vma(anon_vma);
out_error:
unlink_anon_vmas(vma);
return -ENOMEM;
@@ -272,15 +280,11 @@ static void anon_vma_unlink(struct anon_
list_del(&anon_vma_chain->same_anon_vma);

/* We must garbage collect the anon_vma if it's empty */
- empty = list_empty(&anon_vma->head) && !atomic_read(&anon_vma->refcount);
+ empty = list_empty(&anon_vma->head);
anon_vma_unlock(anon_vma);

- if (empty) {
- /* We no longer need the root anon_vma */
- if (anon_vma->root != anon_vma)
- put_anon_vma(anon_vma->root);
- anon_vma_free(anon_vma);
- }
+ if (empty)
+ put_anon_vma(anon_vma);
}

void unlink_anon_vmas(struct vm_area_struct *vma)
@@ -1482,38 +1486,11 @@ int try_to_munlock(struct page *page)
return try_to_unmap_file(page, TTU_MUNLOCK);
}

-/*
- * Drop an anon_vma refcount, freeing the anon_vma and anon_vma->root
- * if necessary. Be careful to do all the tests under the lock. Once
- * we know we are the last user, nobody else can get a reference and we
- * can do the freeing without the lock.
- */
-void put_anon_vma(struct anon_vma *anon_vma)
-{
- BUG_ON(atomic_read(&anon_vma->refcount) <= 0);
- if (atomic_dec_and_lock(&anon_vma->refcount, &anon_vma->root->lock)) {
- struct anon_vma *root = anon_vma->root;
- int empty = list_empty(&anon_vma->head);
- int last_root_user = 0;
- int root_empty = 0;
-
- /*
- * The refcount on a non-root anon_vma got dropped. Drop
- * the refcount on the root and check if we need to free it.
- */
- if (empty && anon_vma != root) {
- BUG_ON(atomic_read(&root->refcount) <= 0);
- last_root_user = atomic_dec_and_test(&root->refcount);
- root_empty = list_empty(&root->head);
- }
- anon_vma_unlock(anon_vma);
-
- if (empty) {
- anon_vma_free(anon_vma);
- if (root_empty && last_root_user)
- anon_vma_free(root);
- }
- }
+void __put_anon_vma(struct anon_vma *anon_vma)
+{
+ if (anon_vma->root != anon_vma)
+ put_anon_vma(anon_vma->root);
+ anon_vma_free(anon_vma);
}

#ifdef CONFIG_MIGRATION


2011-02-17 17:48:06

by Rik van Riel

[permalink] [raw]
Subject: Re: [PATCH 3/3] mm: Simplify anon_vma refcounts

On 02/17/2011 11:19 AM, Peter Zijlstra wrote:

Acked-by: Rik van Riel <[email protected]>

2011-02-17 18:31:00

by Linus Torvalds

[permalink] [raw]
Subject: Re: [PATCH 3/3] mm: Simplify anon_vma refcounts

On Thu, Feb 17, 2011 at 8:19 AM, Peter Zijlstra <[email protected]> wrote:
>
> +void __put_anon_vma(struct anon_vma *anon_vma)
> +{
> + ? ? ? if (anon_vma->root != anon_vma)
> + ? ? ? ? ? ? ? put_anon_vma(anon_vma->root);
> + ? ? ? anon_vma_free(anon_vma);
> ?}

So this makes me nervous. It looks like recursion.

Now, I don't think we can ever get a chain of these things (because
the root should be the root of everything), but I still preferred the
older code that made that "one-level root" case explicit, and didn't
have recursion.

IOW, even though it should be entirely equivalent, I think I'd really
prefer something like

void __put_anon_vma(struct anon_vma *anon_vma)
{
struct anon_vma *root = anon_vma->root;

if (root != anon_vma && atomic_dec_and_test(&root->refcount))
anon_vma_free(root);
anon_vma_free(anon_vma);
}

instead. Exactly because it makes it very clear that the "root" is a
root, and we're not doing some possibly arbitrarily deep list like the
dentry tree (which avoids recursion by open-coding its freeing as a
loop).

Hmm? (The above is obviously untested, maybe it has some stupid bug)

Linus

2011-02-18 11:29:30

by Peter Zijlstra

[permalink] [raw]
Subject: Re: [PATCH 3/3] mm: Simplify anon_vma refcounts

On Thu, 2011-02-17 at 10:30 -0800, Linus Torvalds wrote:
> On Thu, Feb 17, 2011 at 8:19 AM, Peter Zijlstra <[email protected]> wrote:
> >
> > +void __put_anon_vma(struct anon_vma *anon_vma)
> > +{
> > + if (anon_vma->root != anon_vma)
> > + put_anon_vma(anon_vma->root);
> > + anon_vma_free(anon_vma);
> > }
>
> So this makes me nervous. It looks like recursion.
>
> Now, I don't think we can ever get a chain of these things (because
> the root should be the root of everything),

Exactly.

> but I still preferred the
> older code that made that "one-level root" case explicit, and didn't
> have recursion.
>
> IOW, even though it should be entirely equivalent, I think I'd really
> prefer something like
>
> void __put_anon_vma(struct anon_vma *anon_vma)
> {
> struct anon_vma *root = anon_vma->root;
>
> if (root != anon_vma && atomic_dec_and_test(&root->refcount))
> anon_vma_free(root);
> anon_vma_free(anon_vma);
> }
>
> instead. Exactly because it makes it very clear that the "root" is a
> root, and we're not doing some possibly arbitrarily deep list like the
> dentry tree (which avoids recursion by open-coding its freeing as a
> loop).
>
> Hmm? (The above is obviously untested, maybe it has some stupid bug)

Looks about right, I'll give it a spin.

2011-02-18 13:43:49

by Peter Zijlstra

[permalink] [raw]
Subject: Re: [PATCH 3/3] mm: Simplify anon_vma refcounts

Subject: mm: Simplify anon_vma refcounts
From: Peter Zijlstra <[email protected]>
Date: Fri, 26 Nov 2010 15:38:49 +0100

This patch changes the anon_vma refcount to be 0 when the object is
free. It does this by adding 1 ref to being in use in the anon_vma
structure (iow. the anon_vma->head list is not empty).

This allows a simpler release scheme without having to check both the
refcount and the list as well as avoids taking a ref for each entry
on the list.

Reviewed-by: KAMEZAWA Hiroyuki <[email protected]>
Acked-by: Hugh Dickins <[email protected]>
Signed-off-by: Peter Zijlstra <[email protected]>
---
include/linux/rmap.h | 11 +++++--
mm/rmap.c | 78 ++++++++++++++++++---------------------------------
2 files changed, 38 insertions(+), 51 deletions(-)

Index: linux-2.6/include/linux/rmap.h
===================================================================
--- linux-2.6.orig/include/linux/rmap.h
+++ linux-2.6/include/linux/rmap.h
@@ -73,7 +73,13 @@ static inline void get_anon_vma(struct a
atomic_inc(&anon_vma->refcount);
}

-void put_anon_vma(struct anon_vma *);
+void __put_anon_vma(struct anon_vma *anon_vma);
+
+static inline void put_anon_vma(struct anon_vma *anon_vma)
+{
+ if (atomic_dec_and_test(&anon_vma->refcount))
+ __put_anon_vma(anon_vma);
+}

static inline struct anon_vma *page_anon_vma(struct page *page)
{
@@ -116,7 +122,6 @@ void unlink_anon_vmas(struct vm_area_str
int anon_vma_clone(struct vm_area_struct *, struct vm_area_struct *);
int anon_vma_fork(struct vm_area_struct *, struct vm_area_struct *);
void __anon_vma_link(struct vm_area_struct *);
-void anon_vma_free(struct anon_vma *);

static inline void anon_vma_merge(struct vm_area_struct *vma,
struct vm_area_struct *next)
@@ -125,6 +130,8 @@ static inline void anon_vma_merge(struct
unlink_anon_vmas(next);
}

+struct anon_vma *page_get_anon_vma(struct page *page);
+
/*
* rmap interfaces called when adding or removing pte of page
*/
Index: linux-2.6/mm/rmap.c
===================================================================
--- linux-2.6.orig/mm/rmap.c
+++ linux-2.6/mm/rmap.c
@@ -67,11 +67,24 @@ static struct kmem_cache *anon_vma_chain

static inline struct anon_vma *anon_vma_alloc(void)
{
- return kmem_cache_alloc(anon_vma_cachep, GFP_KERNEL);
+ struct anon_vma *anon_vma;
+
+ anon_vma = kmem_cache_alloc(anon_vma_cachep, GFP_KERNEL);
+ if (anon_vma) {
+ atomic_set(&anon_vma->refcount, 1);
+ /*
+ * Initialise the anon_vma root to point to itself. If called
+ * from fork, the root will be reset to the parents anon_vma.
+ */
+ anon_vma->root = anon_vma;
+ }
+
+ return anon_vma;
}

-void anon_vma_free(struct anon_vma *anon_vma)
+static inline void anon_vma_free(struct anon_vma *anon_vma)
{
+ VM_BUG_ON(atomic_read(&anon_vma->refcount));
kmem_cache_free(anon_vma_cachep, anon_vma);
}

@@ -133,11 +146,6 @@ int anon_vma_prepare(struct vm_area_stru
if (unlikely(!anon_vma))
goto out_enomem_free_avc;
allocated = anon_vma;
- /*
- * This VMA had no anon_vma yet. This anon_vma is
- * the root of any anon_vma tree that might form.
- */
- anon_vma->root = anon_vma;
}

anon_vma_lock(anon_vma);
@@ -156,7 +164,7 @@ int anon_vma_prepare(struct vm_area_stru
anon_vma_unlock(anon_vma);

if (unlikely(allocated))
- anon_vma_free(allocated);
+ put_anon_vma(allocated);
if (unlikely(avc))
anon_vma_chain_free(avc);
}
@@ -241,9 +249,9 @@ int anon_vma_fork(struct vm_area_struct
*/
anon_vma->root = pvma->anon_vma->root;
/*
- * With KSM refcounts, an anon_vma can stay around longer than the
- * process it belongs to. The root anon_vma needs to be pinned
- * until this anon_vma is freed, because the lock lives in the root.
+ * With refcounts, an anon_vma can stay around longer than the
+ * process it belongs to. The root anon_vma needs to be pinned until
+ * this anon_vma is freed, because the lock lives in the root.
*/
get_anon_vma(anon_vma->root);
/* Mark this anon_vma as the one where our new (COWed) pages go. */
@@ -253,7 +261,7 @@ int anon_vma_fork(struct vm_area_struct
return 0;

out_error_free_anon_vma:
- anon_vma_free(anon_vma);
+ put_anon_vma(anon_vma);
out_error:
unlink_anon_vmas(vma);
return -ENOMEM;
@@ -272,15 +280,11 @@ static void anon_vma_unlink(struct anon_
list_del(&anon_vma_chain->same_anon_vma);

/* We must garbage collect the anon_vma if it's empty */
- empty = list_empty(&anon_vma->head) && !atomic_read(&anon_vma->refcount);
+ empty = list_empty(&anon_vma->head);
anon_vma_unlock(anon_vma);

- if (empty) {
- /* We no longer need the root anon_vma */
- if (anon_vma->root != anon_vma)
- put_anon_vma(anon_vma->root);
- anon_vma_free(anon_vma);
- }
+ if (empty)
+ put_anon_vma(anon_vma);
}

void unlink_anon_vmas(struct vm_area_struct *vma)
@@ -1470,38 +1474,14 @@ int try_to_munlock(struct page *page)
return try_to_unmap_file(page, TTU_MUNLOCK);
}

-/*
- * Drop an anon_vma refcount, freeing the anon_vma and anon_vma->root
- * if necessary. Be careful to do all the tests under the lock. Once
- * we know we are the last user, nobody else can get a reference and we
- * can do the freeing without the lock.
- */
-void put_anon_vma(struct anon_vma *anon_vma)
-{
- BUG_ON(atomic_read(&anon_vma->refcount) <= 0);
- if (atomic_dec_and_lock(&anon_vma->refcount, &anon_vma->root->lock)) {
- struct anon_vma *root = anon_vma->root;
- int empty = list_empty(&anon_vma->head);
- int last_root_user = 0;
- int root_empty = 0;
+void __put_anon_vma(struct anon_vma *anon_vma)
+{
+ struct anon_vma *root = anon_vma->root;

- /*
- * The refcount on a non-root anon_vma got dropped. Drop
- * the refcount on the root and check if we need to free it.
- */
- if (empty && anon_vma != root) {
- BUG_ON(atomic_read(&root->refcount) <= 0);
- last_root_user = atomic_dec_and_test(&root->refcount);
- root_empty = list_empty(&root->head);
- }
- anon_vma_unlock(anon_vma);
+ if (root != anon_vma && atomic_dec_and_test(&root->refcount))
+ anon_vma_free(root);

- if (empty) {
- anon_vma_free(anon_vma);
- if (root_empty && last_root_user)
- anon_vma_free(root);
- }
- }
+ anon_vma_free(anon_vma);
}

#ifdef CONFIG_MIGRATION

2011-02-18 14:48:46

by Rik van Riel

[permalink] [raw]
Subject: Re: [PATCH 3/3] mm: Simplify anon_vma refcounts

On 02/18/2011 08:44 AM, Peter Zijlstra wrote:
> Subject: mm: Simplify anon_vma refcounts
> From: Peter Zijlstra<[email protected]>
> Date: Fri, 26 Nov 2010 15:38:49 +0100
>
> This patch changes the anon_vma refcount to be 0 when the object is
> free. It does this by adding 1 ref to being in use in the anon_vma
> structure (iow. the anon_vma->head list is not empty).
>
> This allows a simpler release scheme without having to check both the
> refcount and the list as well as avoids taking a ref for each entry
> on the list.
>
> Reviewed-by: KAMEZAWA Hiroyuki<[email protected]>
> Acked-by: Hugh Dickins<[email protected]>
> Signed-off-by: Peter Zijlstra<[email protected]>

Acked-by: Rik van Riel <[email protected]>

2011-03-09 16:38:37

by Mel Gorman

[permalink] [raw]
Subject: Re: [PATCH 3/3] mm: Simplify anon_vma refcounts

On Fri, Feb 18, 2011 at 02:44:43PM +0100, Peter Zijlstra wrote:
> Subject: mm: Simplify anon_vma refcounts
> From: Peter Zijlstra <[email protected]>
> Date: Fri, 26 Nov 2010 15:38:49 +0100
>
> This patch changes the anon_vma refcount to be 0 when the object is
> free. It does this by adding 1 ref to being in use in the anon_vma
> structure (iow. the anon_vma->head list is not empty).
>
> This allows a simpler release scheme without having to check both the
> refcount and the list as well as avoids taking a ref for each entry
> on the list.
>
> Reviewed-by: KAMEZAWA Hiroyuki <[email protected]>
> Acked-by: Hugh Dickins <[email protected]>
> Signed-off-by: Peter Zijlstra <[email protected]>

Acked-by: Mel Gorman <[email protected]>

--
Mel Gorman
SUSE Labs