[v5,1/4] memcg: introduce per-memcg reclaim interface

Commit Message

Yosry Ahmed April 25, 2022, 7 p.m. UTC

From: Shakeel Butt <[e-mail]>

Introduce a memcg interface to trigger memory reclaim on a memory cgroup.

Use case: Proactive Reclaim
---------------------------

A userspace proactive reclaimer can continuously probe the memcg to
reclaim a small amount of memory. This gives more accurate and
up-to-date workingset estimation as the LRUs are continuously
sorted and can potentially provide more deterministic memory
overcommit behavior. The memory overcommit controller can provide
more proactive response to the changing behavior of the running
applications instead of being reactive.

A userspace reclaimer's purpose in this case is not a complete replacement
for kswapd or direct reclaim, it is to proactively identify memory savings
opportunities and reclaim some amount of cold pages set by the policy
to free up the memory for more demanding jobs or scheduling new jobs.

A user space proactive reclaimer is used in Google data centers.
Additionally, Meta's TMO paper recently referenced a very similar
interface used for user space proactive reclaim:
https://dl.acm.org/doi/pdf/10.1145/3503222.3507731

Benefits of a user space reclaimer:
-----------------------------------

1) More flexible on who should be charged for the cpu of the memory
reclaim. For proactive reclaim, it makes more sense to be centralized.

2) More flexible on dedicating the resources (like cpu). The memory
overcommit controller can balance the cost between the cpu usage and
the memory reclaimed.

3) Provides a way to the applications to keep their LRUs sorted, so,
under memory pressure better reclaim candidates are selected. This also
gives more accurate and uptodate notion of working set for an
application.

Why memory.high is not enough?
------------------------------

- memory.high can be used to trigger reclaim in a memcg and can
  potentially be used for proactive reclaim.
  However there is a big downside in using memory.high. It can potentially
  introduce high reclaim stalls in the target application as the
  allocations from the processes or the threads of the application can hit
  the temporary memory.high limit.

- Userspace proactive reclaimers usually use feedback loops to decide
  how much memory to proactively reclaim from a workload. The metrics
  used for this are usually either refaults or PSI, and these metrics
  will become messy if the application gets throttled by hitting the
  high limit.

- memory.high is a stateful interface, if the userspace proactive
  reclaimer crashes for any reason while triggering reclaim it can leave
  the application in a bad state.

- If a workload is rapidly expanding, setting memory.high to proactively
  reclaim memory can result in actually reclaiming more memory than
  intended.

The benefits of such interface and shortcomings of existing interface
were further discussed in this RFC thread:
https://lore.kernel.org/linux-mm/[e-mail]/

Interface:
----------

Introducing a very simple memcg interface 'echo 10M > memory.reclaim' to
trigger reclaim in the target memory cgroup.

The interface is introduced as a nested-keyed file to allow for future
optional arguments to be easily added to configure the behavior of
reclaim.

Possible Extensions:
--------------------

- This interface can be extended with an additional parameter or flags
  to allow specifying one or more types of memory to reclaim from (e.g.
  file, anon, ..).

- The interface can also be extended with a node mask to reclaim from
  specific nodes. This has use cases for reclaim-based demotion in memory
  tiering systens.

- A similar per-node interface can also be added to support proactive
  reclaim and reclaim-based demotion in systems without memcg.

- Add a timeout parameter to make it easier for user space to call the
  interface without worrying about being blocked for an undefined amount
  of time.

For now, let's keep things simple by adding the basic functionality.

[[e-mail]: worked on versions v2 onwards, refreshed to
current master, updated commit message based on recent
discussions and use cases]

Signed-off-by: Shakeel Butt <[e-mail]>
Co-developed-by: Yosry Ahmed <[e-mail]>
Signed-off-by: Yosry Ahmed <[e-mail]>
Acked-by: Johannes Weiner <[e-mail]>
Acked-by: Michal Hocko <[e-mail]>
Acked-by: Wei Xu <[e-mail]>
Acked-by: Roman Gushchin <[e-mail]>
---
 Documentation/admin-guide/cgroup-v2.rst | 21 ++++++++++++
 mm/memcontrol.c                         | 45 +++++++++++++++++++++++++
 2 files changed, 66 insertions(+)

Comments

David Rientjes April 25, 2022, 7:15 p.m. UTC | #1

On Mon, 25 Apr 2022, Yosry Ahmed wrote:

> From: Shakeel Butt <[e-mail]>
> 
> Introduce a memcg interface to trigger memory reclaim on a memory cgroup.
> 
> Use case: Proactive Reclaim
> ---------------------------
> 
> A userspace proactive reclaimer can continuously probe the memcg to
> reclaim a small amount of memory. This gives more accurate and
> up-to-date workingset estimation as the LRUs are continuously
> sorted and can potentially provide more deterministic memory
> overcommit behavior. The memory overcommit controller can provide
> more proactive response to the changing behavior of the running
> applications instead of being reactive.
> 
> A userspace reclaimer's purpose in this case is not a complete replacement
> for kswapd or direct reclaim, it is to proactively identify memory savings
> opportunities and reclaim some amount of cold pages set by the policy
> to free up the memory for more demanding jobs or scheduling new jobs.
> 
> A user space proactive reclaimer is used in Google data centers.
> Additionally, Meta's TMO paper recently referenced a very similar
> interface used for user space proactive reclaim:
> https://dl.acm.org/doi/pdf/10.1145/3503222.3507731
> 
> Benefits of a user space reclaimer:
> -----------------------------------
> 
> 1) More flexible on who should be charged for the cpu of the memory
> reclaim. For proactive reclaim, it makes more sense to be centralized.
> 
> 2) More flexible on dedicating the resources (like cpu). The memory
> overcommit controller can balance the cost between the cpu usage and
> the memory reclaimed.
> 
> 3) Provides a way to the applications to keep their LRUs sorted, so,
> under memory pressure better reclaim candidates are selected. This also
> gives more accurate and uptodate notion of working set for an
> application.
> 
> Why memory.high is not enough?
> ------------------------------
> 
> - memory.high can be used to trigger reclaim in a memcg and can
>   potentially be used for proactive reclaim.
>   However there is a big downside in using memory.high. It can potentially
>   introduce high reclaim stalls in the target application as the
>   allocations from the processes or the threads of the application can hit
>   the temporary memory.high limit.
> 
> - Userspace proactive reclaimers usually use feedback loops to decide
>   how much memory to proactively reclaim from a workload. The metrics
>   used for this are usually either refaults or PSI, and these metrics
>   will become messy if the application gets throttled by hitting the
>   high limit.
> 
> - memory.high is a stateful interface, if the userspace proactive
>   reclaimer crashes for any reason while triggering reclaim it can leave
>   the application in a bad state.
> 
> - If a workload is rapidly expanding, setting memory.high to proactively
>   reclaim memory can result in actually reclaiming more memory than
>   intended.
> 
> The benefits of such interface and shortcomings of existing interface
> were further discussed in this RFC thread:
> https://lore.kernel.org/linux-mm/[e-mail]/
> 
> Interface:
> ----------
> 
> Introducing a very simple memcg interface 'echo 10M > memory.reclaim' to
> trigger reclaim in the target memory cgroup.
> 
> The interface is introduced as a nested-keyed file to allow for future
> optional arguments to be easily added to configure the behavior of
> reclaim.
> 
> Possible Extensions:
> --------------------
> 
> - This interface can be extended with an additional parameter or flags
>   to allow specifying one or more types of memory to reclaim from (e.g.
>   file, anon, ..).
> 
> - The interface can also be extended with a node mask to reclaim from
>   specific nodes. This has use cases for reclaim-based demotion in memory
>   tiering systens.
> 
> - A similar per-node interface can also be added to support proactive
>   reclaim and reclaim-based demotion in systems without memcg.
> 
> - Add a timeout parameter to make it easier for user space to call the
>   interface without worrying about being blocked for an undefined amount
>   of time.
> 
> For now, let's keep things simple by adding the basic functionality.
> 
> [[e-mail]: worked on versions v2 onwards, refreshed to
> current master, updated commit message based on recent
> discussions and use cases]
> 
> Signed-off-by: Shakeel Butt <[e-mail]>
> Co-developed-by: Yosry Ahmed <[e-mail]>
> Signed-off-by: Yosry Ahmed <[e-mail]>
> Acked-by: Johannes Weiner <[e-mail]>
> Acked-by: Michal Hocko <[e-mail]>
> Acked-by: Wei Xu <[e-mail]>
> Acked-by: Roman Gushchin <[e-mail]>

Acked-by: David Rientjes <[e-mail]>

"can over or under reclaim from the target cgroup" begs the question of 
how much more memory the kernel can decide to reclaim :)  I think it's 
assumed that it's minimal and that matches the current implementation that 
rounds up to SWAP_CLUSTER_MAX, though, so looks good.

Thanks Yosry!

Yosry Ahmed April 25, 2022, 7:31 p.m. UTC | #2

On Mon, Apr 25, 2022 at 12:15 PM David Rientjes <[e-mail]> wrote:
>
> On Mon, 25 Apr 2022, Yosry Ahmed wrote:
>
> > From: Shakeel Butt <[e-mail]>
> >
> > Introduce a memcg interface to trigger memory reclaim on a memory cgroup.
> >
> > Use case: Proactive Reclaim
> > ---------------------------
> >
> > A userspace proactive reclaimer can continuously probe the memcg to
> > reclaim a small amount of memory. This gives more accurate and
> > up-to-date workingset estimation as the LRUs are continuously
> > sorted and can potentially provide more deterministic memory
> > overcommit behavior. The memory overcommit controller can provide
> > more proactive response to the changing behavior of the running
> > applications instead of being reactive.
> >
> > A userspace reclaimer's purpose in this case is not a complete replacement
> > for kswapd or direct reclaim, it is to proactively identify memory savings
> > opportunities and reclaim some amount of cold pages set by the policy
> > to free up the memory for more demanding jobs or scheduling new jobs.
> >
> > A user space proactive reclaimer is used in Google data centers.
> > Additionally, Meta's TMO paper recently referenced a very similar
> > interface used for user space proactive reclaim:
> > https://dl.acm.org/doi/pdf/10.1145/3503222.3507731
> >
> > Benefits of a user space reclaimer:
> > -----------------------------------
> >
> > 1) More flexible on who should be charged for the cpu of the memory
> > reclaim. For proactive reclaim, it makes more sense to be centralized.
> >
> > 2) More flexible on dedicating the resources (like cpu). The memory
> > overcommit controller can balance the cost between the cpu usage and
> > the memory reclaimed.
> >
> > 3) Provides a way to the applications to keep their LRUs sorted, so,
> > under memory pressure better reclaim candidates are selected. This also
> > gives more accurate and uptodate notion of working set for an
> > application.
> >
> > Why memory.high is not enough?
> > ------------------------------
> >
> > - memory.high can be used to trigger reclaim in a memcg and can
> >   potentially be used for proactive reclaim.
> >   However there is a big downside in using memory.high. It can potentially
> >   introduce high reclaim stalls in the target application as the
> >   allocations from the processes or the threads of the application can hit
> >   the temporary memory.high limit.
> >
> > - Userspace proactive reclaimers usually use feedback loops to decide
> >   how much memory to proactively reclaim from a workload. The metrics
> >   used for this are usually either refaults or PSI, and these metrics
> >   will become messy if the application gets throttled by hitting the
> >   high limit.
> >
> > - memory.high is a stateful interface, if the userspace proactive
> >   reclaimer crashes for any reason while triggering reclaim it can leave
> >   the application in a bad state.
> >
> > - If a workload is rapidly expanding, setting memory.high to proactively
> >   reclaim memory can result in actually reclaiming more memory than
> >   intended.
> >
> > The benefits of such interface and shortcomings of existing interface
> > were further discussed in this RFC thread:
> > https://lore.kernel.org/linux-mm/[e-mail]/
> >
> > Interface:
> > ----------
> >
> > Introducing a very simple memcg interface 'echo 10M > memory.reclaim' to
> > trigger reclaim in the target memory cgroup.
> >
> > The interface is introduced as a nested-keyed file to allow for future
> > optional arguments to be easily added to configure the behavior of
> > reclaim.
> >
> > Possible Extensions:
> > --------------------
> >
> > - This interface can be extended with an additional parameter or flags
> >   to allow specifying one or more types of memory to reclaim from (e.g.
> >   file, anon, ..).
> >
> > - The interface can also be extended with a node mask to reclaim from
> >   specific nodes. This has use cases for reclaim-based demotion in memory
> >   tiering systens.
> >
> > - A similar per-node interface can also be added to support proactive
> >   reclaim and reclaim-based demotion in systems without memcg.
> >
> > - Add a timeout parameter to make it easier for user space to call the
> >   interface without worrying about being blocked for an undefined amount
> >   of time.
> >
> > For now, let's keep things simple by adding the basic functionality.
> >
> > [[e-mail]: worked on versions v2 onwards, refreshed to
> > current master, updated commit message based on recent
> > discussions and use cases]
> >
> > Signed-off-by: Shakeel Butt <[e-mail]>
> > Co-developed-by: Yosry Ahmed <[e-mail]>
> > Signed-off-by: Yosry Ahmed <[e-mail]>
> > Acked-by: Johannes Weiner <[e-mail]>
> > Acked-by: Michal Hocko <[e-mail]>
> > Acked-by: Wei Xu <[e-mail]>
> > Acked-by: Roman Gushchin <[e-mail]>
>
> Acked-by: David Rientjes <[e-mail]>
>
> "can over or under reclaim from the target cgroup" begs the question of
> how much more memory the kernel can decide to reclaim :)  I think it's
> assumed that it's minimal and that matches the current implementation that
> rounds up to SWAP_CLUSTER_MAX, though, so looks good.
>
> Thanks Yosry!

I think it could be more complex than this. Some functions that get
called during reclaim only use the nr_to_reclaim parameter to check if
they need one more iteration, but not to limit the actual reclaimed
pages per say. For example, nr_to_reclaim is not even passed to
shrink_slab() or mem_cgroup_soft_limit_reclaim(), so they have no way
to know that they should stop if nr_to_reclaim was already satisfied.
I think the general assumption is that each of these calls normally
does not reclaim a huge number of pages, so like you said, the kernel
should not over-reclaim too much. However, I don't think there are
guarantees about this.

Michal Hocko April 26, 2022, 8:22 a.m. UTC | #3

On Mon 25-04-22 12:31:51, Yosry Ahmed wrote:
> On Mon, Apr 25, 2022 at 12:15 PM David Rientjes <[e-mail]> wrote:
[...]
> > "can over or under reclaim from the target cgroup" begs the question of
> > how much more memory the kernel can decide to reclaim :)  I think it's
> > assumed that it's minimal and that matches the current implementation that
> > rounds up to SWAP_CLUSTER_MAX, though, so looks good.
> >
> > Thanks Yosry!
> 
> I think it could be more complex than this. Some functions that get
> called during reclaim only use the nr_to_reclaim parameter to check if
> they need one more iteration, but not to limit the actual reclaimed
> pages per say. For example, nr_to_reclaim is not even passed to
> shrink_slab() or mem_cgroup_soft_limit_reclaim(), so they have no way
> to know that they should stop if nr_to_reclaim was already satisfied.
> I think the general assumption is that each of these calls normally
> does not reclaim a huge number of pages, so like you said, the kernel
> should not over-reclaim too much. However, I don't think there are
> guarantees about this.

There are no guarantees indeed and it si definitely good to be explicit
about that so that userspace tools expect that and consider that in
the imeplementation. Sure we do not want to go overboard and huge excess
should be considered a bug. I am not sure we do agree on the notion of
"huge" so let's see.

Davidlohr Bueso May 4, 2022, 12:13 a.m. UTC | #4

On Mon, 25 Apr 2022, Yosry Ahmed wrote:

>From: Shakeel Butt <[e-mail]>
>
>Introduce a memcg interface to trigger memory reclaim on a memory cgroup.
>
>Use case: Proactive Reclaim
>---------------------------
>
>A userspace proactive reclaimer can continuously probe the memcg to
>reclaim a small amount of memory. This gives more accurate and
>up-to-date workingset estimation as the LRUs are continuously
>sorted and can potentially provide more deterministic memory
>overcommit behavior. The memory overcommit controller can provide
>more proactive response to the changing behavior of the running
>applications instead of being reactive.
>
>A userspace reclaimer's purpose in this case is not a complete replacement
>for kswapd or direct reclaim, it is to proactively identify memory savings
>opportunities and reclaim some amount of cold pages set by the policy
>to free up the memory for more demanding jobs or scheduling new jobs.
>
>A user space proactive reclaimer is used in Google data centers.
>Additionally, Meta's TMO paper recently referenced a very similar
>interface used for user space proactive reclaim:
>https://dl.acm.org/doi/pdf/10.1145/3503222.3507731
>
>Benefits of a user space reclaimer:
>-----------------------------------
>
>1) More flexible on who should be charged for the cpu of the memory
>reclaim. For proactive reclaim, it makes more sense to be centralized.
>
>2) More flexible on dedicating the resources (like cpu). The memory
>overcommit controller can balance the cost between the cpu usage and
>the memory reclaimed.
>
>3) Provides a way to the applications to keep their LRUs sorted, so,
>under memory pressure better reclaim candidates are selected. This also
>gives more accurate and uptodate notion of working set for an
>application.
>
>Why memory.high is not enough?
>------------------------------
>
>- memory.high can be used to trigger reclaim in a memcg and can
>  potentially be used for proactive reclaim.
>  However there is a big downside in using memory.high. It can potentially
>  introduce high reclaim stalls in the target application as the
>  allocations from the processes or the threads of the application can hit
>  the temporary memory.high limit.
>
>- Userspace proactive reclaimers usually use feedback loops to decide
>  how much memory to proactively reclaim from a workload. The metrics
>  used for this are usually either refaults or PSI, and these metrics
>  will become messy if the application gets throttled by hitting the
>  high limit.
>
>- memory.high is a stateful interface, if the userspace proactive
>  reclaimer crashes for any reason while triggering reclaim it can leave
>  the application in a bad state.
>
>- If a workload is rapidly expanding, setting memory.high to proactively
>  reclaim memory can result in actually reclaiming more memory than
>  intended.
>
>The benefits of such interface and shortcomings of existing interface
>were further discussed in this RFC thread:
>https://lore.kernel.org/linux-mm/[e-mail]/
>
>Interface:
>----------
>
>Introducing a very simple memcg interface 'echo 10M > memory.reclaim' to
>trigger reclaim in the target memory cgroup.
>
>The interface is introduced as a nested-keyed file to allow for future
>optional arguments to be easily added to configure the behavior of
>reclaim.
>
>Possible Extensions:
>--------------------
>
>- This interface can be extended with an additional parameter or flags
>  to allow specifying one or more types of memory to reclaim from (e.g.
>  file, anon, ..).
>
>- The interface can also be extended with a node mask to reclaim from
>  specific nodes. This has use cases for reclaim-based demotion in memory
>  tiering systens.
>
>- A similar per-node interface can also be added to support proactive
>  reclaim and reclaim-based demotion in systems without memcg.
>
>- Add a timeout parameter to make it easier for user space to call the
>  interface without worrying about being blocked for an undefined amount
>  of time.
>
>For now, let's keep things simple by adding the basic functionality.
>
>[[e-mail]: worked on versions v2 onwards, refreshed to
>current master, updated commit message based on recent
>discussions and use cases]
>
>Signed-off-by: Shakeel Butt <[e-mail]>
>Co-developed-by: Yosry Ahmed <[e-mail]>
>Signed-off-by: Yosry Ahmed <[e-mail]>
>Acked-by: Johannes Weiner <[e-mail]>
>Acked-by: Michal Hocko <[e-mail]>
>Acked-by: Wei Xu <[e-mail]>
>Acked-by: Roman Gushchin <[e-mail]>

Reviewed-by: Davidlohr Bueso <[e-mail]>

Patch

diff --git a/Documentation/admin-guide/cgroup-v2.rst b/Documentation/admin-guide/cgroup-v2.rst
index 69d7a6983f78..19bcd73cad03 100644
--- a/Documentation/admin-guide/cgroup-v2.rst
+++ b/Documentation/admin-guide/cgroup-v2.rst
@@ -1208,6 +1208,27 @@  PAGE_SIZE multiple when read back.
 	high limit is used and monitored properly, this limit's
 	utility is limited to providing the final safety net.
 
+  memory.reclaim
+	A write-only nested-keyed file which exists for all cgroups.
+
+	This is a simple interface to trigger memory reclaim in the
+	target cgroup.
+
+	This file accepts a single key, the number of bytes to reclaim.
+	No nested keys are currently supported.
+
+	Example::
+
+	  echo "1G" > memory.reclaim
+
+	The interface can be later extended with nested keys to
+	configure the reclaim behavior. For example, specify the
+	type of memory to reclaim from (anon, file, ..).
+
+	Please note that the kernel can over or under reclaim from
+	the target cgroup. If less bytes are reclaimed than the
+	specified amount, -EAGAIN is returned.
+
   memory.oom.group
 	A read-write single value file which exists on non-root
 	cgroups.  The default value is "0".
diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index 725f76723220..879f1716d6e9 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -6355,6 +6355,46 @@  static ssize_t memory_oom_group_write(struct kernfs_open_file *of,
 	return nbytes;
 }
 
+static ssize_t memory_reclaim(struct kernfs_open_file *of, char *buf,
+			      size_t nbytes, loff_t off)
+{
+	struct mem_cgroup *memcg = mem_cgroup_from_css(of_css(of));
+	unsigned int nr_retries = MAX_RECLAIM_RETRIES;
+	unsigned long nr_to_reclaim, nr_reclaimed = 0;
+	int err;
+
+	buf = strstrip(buf);
+	err = page_counter_memparse(buf, "", &nr_to_reclaim);
+	if (err)
+		return err;
+
+	while (nr_reclaimed < nr_to_reclaim) {
+		unsigned long reclaimed;
+
+		if (signal_pending(current))
+			return -EINTR;
+
+		/*
+		 * This is the final attempt, drain percpu lru caches in the
+		 * hope of introducing more evictable pages for
+		 * try_to_free_mem_cgroup_pages().
+		 */
+		if (!nr_retries)
+			lru_add_drain_all();
+
+		reclaimed = try_to_free_mem_cgroup_pages(memcg,
+						nr_to_reclaim - nr_reclaimed,
+						GFP_KERNEL, true);
+
+		if (!reclaimed && !nr_retries--)
+			return -EAGAIN;
+
+		nr_reclaimed += reclaimed;
+	}
+
+	return nbytes;
+}
+
 static struct cftype memory_files[] = {
 	{
 		.name = "current",
@@ -6413,6 +6453,11 @@  static struct cftype memory_files[] = {
 		.seq_show = memory_oom_group_show,
 		.write = memory_oom_group_write,
 	},
+	{
+		.name = "reclaim",
+		.flags = CFTYPE_NS_DELEGATABLE,
+		.write = memory_reclaim,
+	},
 	{ }	/* terminate */
 };