Re: [External] [PATCH 2/3] mm: Charge active memcg when no mm is set
On Sat, Apr 3, 2021 at 3:17 AM Dan Schatzberg wrote: > > set_active_memcg() worked for kernel allocations but was silently > ignored for user pages. > > This patch establishes a precedence order for who gets charged: > > 1. If there is a memcg associated with the page already, that memcg is >charged. This happens during swapin. > > 2. If an explicit mm is passed, mm->memcg is charged. This happens >during page faults, which can be triggered in remote VMs (eg gup). > > 3. Otherwise consult the current process context. If there is an >active_memcg, use that. Otherwise, current->mm->memcg. > > Previously, if a NULL mm was passed to mem_cgroup_charge (case 3) it > would always charge the root cgroup. Now it looks up the active_memcg > first (falling back to charging the root cgroup if not set). > > Signed-off-by: Dan Schatzberg > Acked-by: Johannes Weiner > Acked-by: Tejun Heo > Acked-by: Chris Down > Reviewed-by: Shakeel Butt Reviewed-by: Muchun Song Thanks.
[PATCH 2/3] mm: Charge active memcg when no mm is set
set_active_memcg() worked for kernel allocations but was silently
ignored for user pages.
This patch establishes a precedence order for who gets charged:
1. If there is a memcg associated with the page already, that memcg is
charged. This happens during swapin.
2. If an explicit mm is passed, mm->memcg is charged. This happens
during page faults, which can be triggered in remote VMs (eg gup).
3. Otherwise consult the current process context. If there is an
active_memcg, use that. Otherwise, current->mm->memcg.
Previously, if a NULL mm was passed to mem_cgroup_charge (case 3) it
would always charge the root cgroup. Now it looks up the active_memcg
first (falling back to charging the root cgroup if not set).
Signed-off-by: Dan Schatzberg
Acked-by: Johannes Weiner
Acked-by: Tejun Heo
Acked-by: Chris Down
Reviewed-by: Shakeel Butt
---
mm/filemap.c| 2 +-
mm/memcontrol.c | 48 +++-
mm/shmem.c | 4 ++--
3 files changed, 34 insertions(+), 20 deletions(-)
diff --git a/mm/filemap.c b/mm/filemap.c
index c03463cb72d6..38648f7d2106 100644
--- a/mm/filemap.c
+++ b/mm/filemap.c
@@ -872,7 +872,7 @@ noinline int __add_to_page_cache_locked(struct page *page,
page->index = offset;
if (!huge) {
- error = mem_cgroup_charge(page, current->mm, gfp);
+ error = mem_cgroup_charge(page, NULL, gfp);
if (error)
goto error;
charged = true;
diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index c0b83a396299..d2939d6602b3 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -886,13 +886,24 @@ struct mem_cgroup *mem_cgroup_from_task(struct
task_struct *p)
}
EXPORT_SYMBOL(mem_cgroup_from_task);
+static __always_inline struct mem_cgroup *active_memcg(void)
+{
+ if (in_interrupt())
+ return this_cpu_read(int_active_memcg);
+ else
+ return current->active_memcg;
+}
+
/**
* get_mem_cgroup_from_mm: Obtain a reference on given mm_struct's memcg.
* @mm: mm from which memcg should be extracted. It can be NULL.
*
- * Obtain a reference on mm->memcg and returns it if successful. Otherwise
- * root_mem_cgroup is returned. However if mem_cgroup is disabled, NULL is
- * returned.
+ * Obtain a reference on mm->memcg and returns it if successful. If mm
+ * is NULL, then the memcg is chosen as follows:
+ * 1) The active memcg, if set.
+ * 2) current->mm->memcg, if available
+ * 3) root memcg
+ * If mem_cgroup is disabled, NULL is returned.
*/
struct mem_cgroup *get_mem_cgroup_from_mm(struct mm_struct *mm)
{
@@ -901,13 +912,23 @@ struct mem_cgroup *get_mem_cgroup_from_mm(struct
mm_struct *mm)
if (mem_cgroup_disabled())
return NULL;
+ /*
+* Page cache insertions can happen without an
+* actual mm context, e.g. during disk probing
+* on boot, loopback IO, acct() writes etc.
+*/
+ if (unlikely(!mm)) {
+ memcg = active_memcg();
+ if (unlikely(memcg)) {
+ /* remote memcg must hold a ref */
+ css_get(&memcg->css);
+ return memcg;
+ }
+ mm = current->mm;
+ }
+
rcu_read_lock();
do {
- /*
-* Page cache insertions can happen without an
-* actual mm context, e.g. during disk probing
-* on boot, loopback IO, acct() writes etc.
-*/
if (unlikely(!mm))
memcg = root_mem_cgroup;
else {
@@ -921,14 +942,6 @@ struct mem_cgroup *get_mem_cgroup_from_mm(struct mm_struct
*mm)
}
EXPORT_SYMBOL(get_mem_cgroup_from_mm);
-static __always_inline struct mem_cgroup *active_memcg(void)
-{
- if (in_interrupt())
- return this_cpu_read(int_active_memcg);
- else
- return current->active_memcg;
-}
-
static __always_inline bool memcg_kmem_bypass(void)
{
/* Allow remote memcg charging from any context. */
@@ -6537,7 +6550,8 @@ static int __mem_cgroup_charge(struct page *page, struct
mem_cgroup *memcg,
* @gfp_mask: reclaim mode
*
* Try to charge @page to the memcg that @mm belongs to, reclaiming
- * pages according to @gfp_mask if necessary.
+ * pages according to @gfp_mask if necessary. if @mm is NULL, try to
+ * charge to the active memcg.
*
* Do not use this for pages allocated for swapin.
*
diff --git a/mm/shmem.c b/mm/shmem.c
index 5cfd2fb6e52b..524fa5aa0459 100644
--- a/mm/shmem.c
+++ b/mm/shmem.c
@@ -1694,7 +1694,7 @@ static int shmem_swapin_page(struct inode *inode, pgoff_t
index,
{
struct address_space *mapping = inode->i_mapping;
struct shmem_inode_info *info = SHMEM_I(inode);
- struct mm_struct *charge_mm = vma ? vma->vm_mm : current->mm;
+ struct mm_struct *charge_mm = vma ? vma->vm_mm : NULL;
struct page *p
Re: [External] [PATCH 2/3] mm: Charge active memcg when no mm is set
On Mon, Mar 29, 2021 at 9:13 AM Muchun Song wrote:
>
> On Mon, Mar 29, 2021 at 10:49 PM Dan Schatzberg
> wrote:
[...]
>
> Since remote memcg must hold a reference, we do not
> need to do something like get_active_memcg() does.
> Just use css_get to obtain a ref, it is simpler. Just
> Like below.
>
> + if (unlikely(!mm)) {
> + memcg = active_memcg();
> + if (unlikely(memcg)) {
> + /* remote memcg must hold a ref. */
> + css_get(memcg);
> + return memcg;
> + }
>
I second Muchun's suggestion.
Re: [External] [PATCH 2/3] mm: Charge active memcg when no mm is set
On Mon, Mar 29, 2021 at 10:49 PM Dan Schatzberg
wrote:
>
> set_active_memcg() worked for kernel allocations but was silently
> ignored for user pages.
>
> This patch establishes a precedence order for who gets charged:
>
> 1. If there is a memcg associated with the page already, that memcg is
>charged. This happens during swapin.
>
> 2. If an explicit mm is passed, mm->memcg is charged. This happens
>during page faults, which can be triggered in remote VMs (eg gup).
>
> 3. Otherwise consult the current process context. If there is an
>active_memcg, use that. Otherwise, current->mm->memcg.
>
> Previously, if a NULL mm was passed to mem_cgroup_charge (case 3) it
> would always charge the root cgroup. Now it looks up the active_memcg
> first (falling back to charging the root cgroup if not set).
>
> Signed-off-by: Dan Schatzberg
> Acked-by: Johannes Weiner
> Acked-by: Tejun Heo
> Acked-by: Chris Down
> Reviewed-by: Shakeel Butt
> ---
> mm/filemap.c| 2 +-
> mm/memcontrol.c | 72 -
> mm/shmem.c | 4 +--
> 3 files changed, 44 insertions(+), 34 deletions(-)
>
> diff --git a/mm/filemap.c b/mm/filemap.c
> index eeeb8e2cc36a..63fd980e863a 100644
> --- a/mm/filemap.c
> +++ b/mm/filemap.c
> @@ -872,7 +872,7 @@ noinline int __add_to_page_cache_locked(struct page *page,
> page->index = offset;
>
> if (!huge) {
> - error = mem_cgroup_charge(page, current->mm, gfp);
> + error = mem_cgroup_charge(page, NULL, gfp);
> if (error)
> goto error;
> charged = true;
> diff --git a/mm/memcontrol.c b/mm/memcontrol.c
> index 668d1d7c2645..adc618814fd2 100644
> --- a/mm/memcontrol.c
> +++ b/mm/memcontrol.c
> @@ -884,13 +884,38 @@ struct mem_cgroup *mem_cgroup_from_task(struct
> task_struct *p)
> }
> EXPORT_SYMBOL(mem_cgroup_from_task);
>
> +static __always_inline struct mem_cgroup *active_memcg(void)
> +{
> + if (in_interrupt())
> + return this_cpu_read(int_active_memcg);
> + else
> + return current->active_memcg;
> +}
> +
> +static __always_inline struct mem_cgroup *get_active_memcg(void)
> +{
> + struct mem_cgroup *memcg;
> +
> + rcu_read_lock();
> + memcg = active_memcg();
> + /* remote memcg must hold a ref. */
> + if (memcg && WARN_ON_ONCE(!css_tryget(&memcg->css)))
> + memcg = root_mem_cgroup;
> + rcu_read_unlock();
> +
> + return memcg;
> +}
This function is already removed since the patchset below.
Use obj_cgroup APIs to charge kmem pages
https://lore.kernel.org/patchwork/cover/1399132/
I also suggest not reintroducing get_active_memcg.
There is only one user of it, just inline it into
get_mem_cgroup_from_mm(). Actually we don’t
need get_active_memcg() either.
> +
> /**
> * get_mem_cgroup_from_mm: Obtain a reference on given mm_struct's memcg.
> * @mm: mm from which memcg should be extracted. It can be NULL.
> *
> - * Obtain a reference on mm->memcg and returns it if successful. Otherwise
> - * root_mem_cgroup is returned. However if mem_cgroup is disabled, NULL is
> - * returned.
> + * Obtain a reference on mm->memcg and returns it if successful. If mm
> + * is NULL, then the memcg is chosen as follows:
> + * 1) The active memcg, if set.
> + * 2) current->mm->memcg, if available
> + * 3) root memcg
> + * If mem_cgroup is disabled, NULL is returned.
> */
> struct mem_cgroup *get_mem_cgroup_from_mm(struct mm_struct *mm)
> {
> @@ -899,13 +924,19 @@ struct mem_cgroup *get_mem_cgroup_from_mm(struct
> mm_struct *mm)
> if (mem_cgroup_disabled())
> return NULL;
>
> + /*
> +* Page cache insertions can happen without an
> +* actual mm context, e.g. during disk probing
> +* on boot, loopback IO, acct() writes etc.
> +*/
> + if (unlikely(!mm)) {
> + if (unlikely(active_memcg()))
> + return get_active_memcg();
Since remote memcg must hold a reference, we do not
need to do something like get_active_memcg() does.
Just use css_get to obtain a ref, it is simpler. Just
Like below.
+ if (unlikely(!mm)) {
+ memcg = active_memcg();
+ if (unlikely(memcg)) {
+ /* remote memcg must hold a ref. */
+ css_get(memcg);
+ return memcg;
+ }
Thanks.
> + mm = current->mm;
> + }
> +
> rcu_read_lock();
> do {
> - /*
> -* Page cache insertions can happen withou an
> -* actual mm context, e.g. during disk probing
> -* on boot, loopback IO, acct() writes etc.
> -*/
> if (unlikely(!mm))
> memcg = root_mem_cgroup;
> else {
> @@ -919,28 +950,6 @@ struct mem_cgroup *get_mem_cgroup_from_mm(s
[PATCH 2/3] mm: Charge active memcg when no mm is set
set_active_memcg() worked for kernel allocations but was silently
ignored for user pages.
This patch establishes a precedence order for who gets charged:
1. If there is a memcg associated with the page already, that memcg is
charged. This happens during swapin.
2. If an explicit mm is passed, mm->memcg is charged. This happens
during page faults, which can be triggered in remote VMs (eg gup).
3. Otherwise consult the current process context. If there is an
active_memcg, use that. Otherwise, current->mm->memcg.
Previously, if a NULL mm was passed to mem_cgroup_charge (case 3) it
would always charge the root cgroup. Now it looks up the active_memcg
first (falling back to charging the root cgroup if not set).
Signed-off-by: Dan Schatzberg
Acked-by: Johannes Weiner
Acked-by: Tejun Heo
Acked-by: Chris Down
Reviewed-by: Shakeel Butt
---
mm/filemap.c| 2 +-
mm/memcontrol.c | 72 -
mm/shmem.c | 4 +--
3 files changed, 44 insertions(+), 34 deletions(-)
diff --git a/mm/filemap.c b/mm/filemap.c
index eeeb8e2cc36a..63fd980e863a 100644
--- a/mm/filemap.c
+++ b/mm/filemap.c
@@ -872,7 +872,7 @@ noinline int __add_to_page_cache_locked(struct page *page,
page->index = offset;
if (!huge) {
- error = mem_cgroup_charge(page, current->mm, gfp);
+ error = mem_cgroup_charge(page, NULL, gfp);
if (error)
goto error;
charged = true;
diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index 668d1d7c2645..adc618814fd2 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -884,13 +884,38 @@ struct mem_cgroup *mem_cgroup_from_task(struct
task_struct *p)
}
EXPORT_SYMBOL(mem_cgroup_from_task);
+static __always_inline struct mem_cgroup *active_memcg(void)
+{
+ if (in_interrupt())
+ return this_cpu_read(int_active_memcg);
+ else
+ return current->active_memcg;
+}
+
+static __always_inline struct mem_cgroup *get_active_memcg(void)
+{
+ struct mem_cgroup *memcg;
+
+ rcu_read_lock();
+ memcg = active_memcg();
+ /* remote memcg must hold a ref. */
+ if (memcg && WARN_ON_ONCE(!css_tryget(&memcg->css)))
+ memcg = root_mem_cgroup;
+ rcu_read_unlock();
+
+ return memcg;
+}
+
/**
* get_mem_cgroup_from_mm: Obtain a reference on given mm_struct's memcg.
* @mm: mm from which memcg should be extracted. It can be NULL.
*
- * Obtain a reference on mm->memcg and returns it if successful. Otherwise
- * root_mem_cgroup is returned. However if mem_cgroup is disabled, NULL is
- * returned.
+ * Obtain a reference on mm->memcg and returns it if successful. If mm
+ * is NULL, then the memcg is chosen as follows:
+ * 1) The active memcg, if set.
+ * 2) current->mm->memcg, if available
+ * 3) root memcg
+ * If mem_cgroup is disabled, NULL is returned.
*/
struct mem_cgroup *get_mem_cgroup_from_mm(struct mm_struct *mm)
{
@@ -899,13 +924,19 @@ struct mem_cgroup *get_mem_cgroup_from_mm(struct
mm_struct *mm)
if (mem_cgroup_disabled())
return NULL;
+ /*
+* Page cache insertions can happen without an
+* actual mm context, e.g. during disk probing
+* on boot, loopback IO, acct() writes etc.
+*/
+ if (unlikely(!mm)) {
+ if (unlikely(active_memcg()))
+ return get_active_memcg();
+ mm = current->mm;
+ }
+
rcu_read_lock();
do {
- /*
-* Page cache insertions can happen withou an
-* actual mm context, e.g. during disk probing
-* on boot, loopback IO, acct() writes etc.
-*/
if (unlikely(!mm))
memcg = root_mem_cgroup;
else {
@@ -919,28 +950,6 @@ struct mem_cgroup *get_mem_cgroup_from_mm(struct mm_struct
*mm)
}
EXPORT_SYMBOL(get_mem_cgroup_from_mm);
-static __always_inline struct mem_cgroup *active_memcg(void)
-{
- if (in_interrupt())
- return this_cpu_read(int_active_memcg);
- else
- return current->active_memcg;
-}
-
-static __always_inline struct mem_cgroup *get_active_memcg(void)
-{
- struct mem_cgroup *memcg;
-
- rcu_read_lock();
- memcg = active_memcg();
- /* remote memcg must hold a ref. */
- if (memcg && WARN_ON_ONCE(!css_tryget(&memcg->css)))
- memcg = root_mem_cgroup;
- rcu_read_unlock();
-
- return memcg;
-}
-
static __always_inline bool memcg_kmem_bypass(void)
{
/* Allow remote memcg charging from any context. */
@@ -6549,7 +6558,8 @@ static int __mem_cgroup_charge(struct page *page, struct
mem_cgroup *memcg,
* @gfp_mask: reclaim mode
*
* Try to charge @page to the memcg that @mm belongs to, reclaiming
- * pages according to @gfp_mask if necessary.
+ * pages according to @gfp_mask if necessary. if
Re: [PATCH 2/3] mm: Charge active memcg when no mm is set
On Tue, Mar 16, 2021 at 08:50:16AM -0700, Shakeel Butt wrote: > You will need to rebase to the latest mm tree. This code has changed. Thanks for the feedback, I will address these comments in another rebase. I'll wait and see if there's any comments concerning the loop-related patches but it sounds like this will need to go through the mm-tree
Re: [PATCH 2/3] mm: Charge active memcg when no mm is set
On Tue, Mar 16, 2021 at 8:37 AM Dan Schatzberg wrote:
>
> memalloc_use_memcg() worked for kernel allocations but was silently
> ignored for user pages.
set_active_memcg()
>
> This patch establishes a precedence order for who gets charged:
>
> 1. If there is a memcg associated with the page already, that memcg is
>charged. This happens during swapin.
>
> 2. If an explicit mm is passed, mm->memcg is charged. This happens
>during page faults, which can be triggered in remote VMs (eg gup).
>
> 3. Otherwise consult the current process context. If it has configured
>a current->active_memcg, use that. Otherwise, current->mm->memcg.
It's a bit more sophisticated than current->active_memcg. It has been
extended to work in interrupt context as well.
>
> Previously, if a NULL mm was passed to mem_cgroup_try_charge (case 3) it
mem_cgroup_charge()
> would always charge the root cgroup. Now it looks up the current
> active_memcg first (falling back to charging the root cgroup if not
> set).
>
> Signed-off-by: Dan Schatzberg
> Acked-by: Johannes Weiner
> Acked-by: Tejun Heo
> Acked-by: Chris Down
> Reviewed-by: Shakeel Butt
> ---
> mm/filemap.c| 2 +-
> mm/memcontrol.c | 14 +++---
> mm/shmem.c | 4 ++--
> 3 files changed, 14 insertions(+), 6 deletions(-)
>
> diff --git a/mm/filemap.c b/mm/filemap.c
> index 43700480d897..5135f330f05c 100644
> --- a/mm/filemap.c
> +++ b/mm/filemap.c
> @@ -843,7 +843,7 @@ noinline int __add_to_page_cache_locked(struct page *page,
> page->index = offset;
>
> if (!huge) {
> - error = mem_cgroup_charge(page, current->mm, gfp);
> + error = mem_cgroup_charge(page, NULL, gfp);
> if (error)
> goto error;
> charged = true;
> diff --git a/mm/memcontrol.c b/mm/memcontrol.c
> index e064ac0d850a..9a1b23ed3412 100644
> --- a/mm/memcontrol.c
> +++ b/mm/memcontrol.c
> @@ -6690,7 +6690,8 @@ void mem_cgroup_calculate_protection(struct mem_cgroup
> *root,
> * @gfp_mask: reclaim mode
> *
> * Try to charge @page to the memcg that @mm belongs to, reclaiming
> - * pages according to @gfp_mask if necessary.
> + * pages according to @gfp_mask if necessary. if @mm is NULL, try to
> + * charge to the active memcg.
> *
> * Returns 0 on success. Otherwise, an error code is returned.
> */
> @@ -6726,8 +6727,15 @@ int mem_cgroup_charge(struct page *page, struct
> mm_struct *mm, gfp_t gfp_mask)
> rcu_read_unlock();
> }
>
> - if (!memcg)
> - memcg = get_mem_cgroup_from_mm(mm);
> + if (!memcg) {
> + if (!mm) {
> + memcg = get_mem_cgroup_from_current();
> + if (!memcg)
> + memcg = get_mem_cgroup_from_mm(current->mm);
> + } else {
> + memcg = get_mem_cgroup_from_mm(mm);
> + }
> + }
You will need to rebase to the latest mm tree. This code has changed.
[PATCH 2/3] mm: Charge active memcg when no mm is set
memalloc_use_memcg() worked for kernel allocations but was silently
ignored for user pages.
This patch establishes a precedence order for who gets charged:
1. If there is a memcg associated with the page already, that memcg is
charged. This happens during swapin.
2. If an explicit mm is passed, mm->memcg is charged. This happens
during page faults, which can be triggered in remote VMs (eg gup).
3. Otherwise consult the current process context. If it has configured
a current->active_memcg, use that. Otherwise, current->mm->memcg.
Previously, if a NULL mm was passed to mem_cgroup_try_charge (case 3) it
would always charge the root cgroup. Now it looks up the current
active_memcg first (falling back to charging the root cgroup if not
set).
Signed-off-by: Dan Schatzberg
Acked-by: Johannes Weiner
Acked-by: Tejun Heo
Acked-by: Chris Down
Reviewed-by: Shakeel Butt
---
mm/filemap.c| 2 +-
mm/memcontrol.c | 14 +++---
mm/shmem.c | 4 ++--
3 files changed, 14 insertions(+), 6 deletions(-)
diff --git a/mm/filemap.c b/mm/filemap.c
index 43700480d897..5135f330f05c 100644
--- a/mm/filemap.c
+++ b/mm/filemap.c
@@ -843,7 +843,7 @@ noinline int __add_to_page_cache_locked(struct page *page,
page->index = offset;
if (!huge) {
- error = mem_cgroup_charge(page, current->mm, gfp);
+ error = mem_cgroup_charge(page, NULL, gfp);
if (error)
goto error;
charged = true;
diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index e064ac0d850a..9a1b23ed3412 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -6690,7 +6690,8 @@ void mem_cgroup_calculate_protection(struct mem_cgroup
*root,
* @gfp_mask: reclaim mode
*
* Try to charge @page to the memcg that @mm belongs to, reclaiming
- * pages according to @gfp_mask if necessary.
+ * pages according to @gfp_mask if necessary. if @mm is NULL, try to
+ * charge to the active memcg.
*
* Returns 0 on success. Otherwise, an error code is returned.
*/
@@ -6726,8 +6727,15 @@ int mem_cgroup_charge(struct page *page, struct
mm_struct *mm, gfp_t gfp_mask)
rcu_read_unlock();
}
- if (!memcg)
- memcg = get_mem_cgroup_from_mm(mm);
+ if (!memcg) {
+ if (!mm) {
+ memcg = get_mem_cgroup_from_current();
+ if (!memcg)
+ memcg = get_mem_cgroup_from_mm(current->mm);
+ } else {
+ memcg = get_mem_cgroup_from_mm(mm);
+ }
+ }
ret = try_charge(memcg, gfp_mask, nr_pages);
if (ret)
diff --git a/mm/shmem.c b/mm/shmem.c
index b2db4ed0fbc7..353b362c370e 100644
--- a/mm/shmem.c
+++ b/mm/shmem.c
@@ -1695,7 +1695,7 @@ static int shmem_swapin_page(struct inode *inode, pgoff_t
index,
{
struct address_space *mapping = inode->i_mapping;
struct shmem_inode_info *info = SHMEM_I(inode);
- struct mm_struct *charge_mm = vma ? vma->vm_mm : current->mm;
+ struct mm_struct *charge_mm = vma ? vma->vm_mm : NULL;
struct page *page;
swp_entry_t swap;
int error;
@@ -1816,7 +1816,7 @@ static int shmem_getpage_gfp(struct inode *inode, pgoff_t
index,
}
sbinfo = SHMEM_SB(inode->i_sb);
- charge_mm = vma ? vma->vm_mm : current->mm;
+ charge_mm = vma ? vma->vm_mm : NULL;
page = pagecache_get_page(mapping, index,
FGP_ENTRY | FGP_HEAD | FGP_LOCK, 0);
--
2.30.2
[PATCH 2/3] mm: Charge active memcg when no mm is set
memalloc_use_memcg() worked for kernel allocations but was silently
ignored for user pages.
This patch establishes a precedence order for who gets charged:
1. If there is a memcg associated with the page already, that memcg is
charged. This happens during swapin.
2. If an explicit mm is passed, mm->memcg is charged. This happens
during page faults, which can be triggered in remote VMs (eg gup).
3. Otherwise consult the current process context. If it has configured
a current->active_memcg, use that. Otherwise, current->mm->memcg.
Previously, if a NULL mm was passed to mem_cgroup_try_charge (case 3) it
would always charge the root cgroup. Now it looks up the current
active_memcg first (falling back to charging the root cgroup if not
set).
Signed-off-by: Dan Schatzberg
Acked-by: Johannes Weiner
Acked-by: Tejun Heo
Acked-by: Chris Down
Reviewed-by: Shakeel Butt
---
mm/memcontrol.c | 11 ---
mm/shmem.c | 4 ++--
2 files changed, 10 insertions(+), 5 deletions(-)
diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index e27ac6d79a32..88b792da 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -6676,7 +6676,8 @@ void mem_cgroup_calculate_protection(struct mem_cgroup
*root,
* @gfp_mask: reclaim mode
*
* Try to charge @page to the memcg that @mm belongs to, reclaiming
- * pages according to @gfp_mask if necessary.
+ * pages according to @gfp_mask if necessary. if @mm is NULL, try to
+ * charge to the active memcg.
*
* Returns 0 on success. Otherwise, an error code is returned.
*/
@@ -6712,8 +6713,12 @@ int mem_cgroup_charge(struct page *page, struct
mm_struct *mm, gfp_t gfp_mask)
rcu_read_unlock();
}
- if (!memcg)
- memcg = get_mem_cgroup_from_mm(mm);
+ if (!memcg) {
+ if (!mm)
+ memcg = get_mem_cgroup_from_current();
+ else
+ memcg = get_mem_cgroup_from_mm(mm);
+ }
ret = try_charge(memcg, gfp_mask, nr_pages);
if (ret)
diff --git a/mm/shmem.c b/mm/shmem.c
index 271548ca20f3..1139f52ac4ee 100644
--- a/mm/shmem.c
+++ b/mm/shmem.c
@@ -1695,7 +1695,7 @@ static int shmem_swapin_page(struct inode *inode, pgoff_t
index,
{
struct address_space *mapping = inode->i_mapping;
struct shmem_inode_info *info = SHMEM_I(inode);
- struct mm_struct *charge_mm = vma ? vma->vm_mm : current->mm;
+ struct mm_struct *charge_mm = vma ? vma->vm_mm : NULL;
struct page *page;
swp_entry_t swap;
int error;
@@ -1809,7 +1809,7 @@ static int shmem_getpage_gfp(struct inode *inode, pgoff_t
index,
}
sbinfo = SHMEM_SB(inode->i_sb);
- charge_mm = vma ? vma->vm_mm : current->mm;
+ charge_mm = vma ? vma->vm_mm : NULL;
page = find_lock_entry(mapping, index);
if (xa_is_value(page)) {
--
2.24.1
