Skip to contentSkip to conversation

mm, memcontrol: Implement memory.swap.events

PATCH New Tejun Heo 1 participant 3 messages Apr 16, 2018

Subsystems (6)
Memory management Documentation cgroup memcg +2 more subsystems
lore
Share & tools
lore

Tejun Heo

author maintainer
#0 lore raw

PATCH v2 1/1 mm, memcontrol: Implement memory.swap.events

Hello,

Rebased on top of e27be240df53 ("mm: memcg: make sure memory.events is uptodate when waking pollers").

This patchset implements memory.swap.events which contains max and fail events so that userland can monitor and respond to swap running out. It contains the following two patches.

0001-mm-memcontrol-Move-swap-charge-handling-into-get_swa.patch 0002-mm-memcontrol-Implement-memory.swap.events.patch

This patchset is on top of the current linus#master (a27fc14219f2e3c4a46ba9177b04d9b52c875532).

Thanks.

Signature · 1 line
tejun
author maintainer
#1 lore raw

PATCH 1/2 mm, memcontrol: Move swap charge handling into get_swap_page()

get_swap_page() is always followed by mem_cgroup_try_charge_swap(). This patch moves mem_cgroup_try_charge_swap() into get_swap_page() and makes get_swap_page() call the function even after swap allocation failure.

This simplifies the callers and consolidates memcg related logic and will ease adding swap related memcg events.

Signed-off-by Tejun Heo <[email protected]>
Cc Johannes Weiner <[email protected]>
Cc Michal Hocko <[email protected]>
Cc Vladimir Davydov <[email protected]>
Cc Roman Gushchin <[email protected]>
Cc Rik van Riel <[email protected]>
Cc Andrew Morton <[email protected]>
---
 mm/memcontrol.c |    3 +++
 mm/shmem.c      |    4 ----
 mm/swap_slots.c |   10 +++++++---
 mm/swap_state.c |    3 ---
 4 files changed, 10 insertions(+), 10 deletions(-)
mm/memcontrol.c +3 -0
@@ -6012,6 +6012,9 @@ int mem_cgroup_try_charge_swap(struct pa
 	if (!memcg)
 		return 0;
 
+	if (!entry.val)
+		return 0;
+
 	memcg = mem_cgroup_id_get_online(memcg);
 
 	if (!mem_cgroup_is_root(memcg) &&
mm/shmem.c +0 -4
@@ -1322,9 +1322,6 @@ static int shmem_writepage(struct page *
 	if (!swap.val)
 		goto redirty;
 
-	if (mem_cgroup_try_charge_swap(page, swap))
-		goto free_swap;
-
 	/*
 	 * Add inode to shmem_unuse()'s list of swapped-out inodes,
 	 * if it's not already there.  Do it now before the page is

@@ -1353,7 +1350,6 @@ static int shmem_writepage(struct page *
 	}
 
 	mutex_unlock(&shmem_swaplist_mutex);
-free_swap:
 	put_swap_page(page, swap);
 redirty:
 	set_page_dirty(page);
mm/swap_slots.c +7 -3
@@ -317,7 +317,7 @@ swp_entry_t get_swap_page(struct page *p
 	if (PageTransHuge(page)) {
 		if (IS_ENABLED(CONFIG_THP_SWAP))
 			get_swap_pages(1, true, &entry);
-		return entry;
+		goto out;
 	}
 
 	/*

@@ -347,10 +347,14 @@ repeat:
 		}
 		mutex_unlock(&cache->alloc_lock);
 		if (entry.val)
-			return entry;
+			goto out;
 	}
 
 	get_swap_pages(1, false, &entry);
-
+out:
+	if (mem_cgroup_try_charge_swap(page, entry)) {
+		put_swap_page(page, entry);
+		entry.val = 0;
+	}
 	return entry;
 }
mm/swap_state.c +0 -3
@@ -216,9 +216,6 @@ int add_to_swap(struct page *page)
 	if (!entry.val)
 		return 0;
 
-	if (mem_cgroup_try_charge_swap(page, entry))
-		goto fail;
-
 	/*
 	 * Radix-tree node allocations from PF_MEMALLOC contexts could
 	 * completely exhaust the page allocator. __GFP_NOMEMALLOC
author maintainer
#2 lore raw

PATCH 2/2 mm, memcontrol: Implement memory.swap.events

In reply to Tejun Heo

Add swap max and fail events so that userland can monitor and respond to running out of swap.

v2: Rebased on top of e27be240df53 ("mm: memcg: make sure
memory.events is uptodate when waking pollers")

Signed-off-by Tejun Heo <[email protected]>
Cc Johannes Weiner <[email protected]>
Cc Michal Hocko <[email protected]>
Cc Vladimir Davydov <[email protected]>
Cc Roman Gushchin <[email protected]>
Cc Rik van Riel <[email protected]>
Cc Andrew Morton <[email protected]>
Cc [email protected]

---
Hello,

I'm not too sure about the fail event. Right now, it's a bit confusing which stats / events are recursive and which aren't and also which ones reflect events which originate from a given cgroup and which targets the cgroup. No idea what the right long term solution is and it could just be that growing them organically is actually the only right thing to do.

Thanks.

 Documentation/cgroup-v2.txt |   16 ++++++++++++++++
 include/linux/memcontrol.h  |    5 +++++
 mm/memcontrol.c             |   24 +++++++++++++++++++++++-
 3 files changed, 44 insertions(+), 1 deletion(-)
Documentation/cgroup-v2.txt +16 -0
@@ -1199,6 +1199,22 @@ PAGE_SIZE multiple when read back.
 	Swap usage hard limit.  If a cgroup's swap usage reaches this
 	limit, anonymous memory of the cgroup will not be swapped out.
 
+  memory.swap.events
+	A read-only flat-keyed file which exists on non-root cgroups.
+	The following entries are defined.  Unless specified
+	otherwise, a value change in this file generates a file
+	modified event.
+
+	  max
+		The number of times the cgroup's swap usage was about
+		to go over the max boundary and swap allocation
+		failed.
+
+	  fail
+		The number of times swap allocation failed either
+		because of running out of swap system-wide or max
+		limit.
+
 
 Usage Guidelines
 ~~~~~~~~~~~~~~~~
include/linux/memcontrol.h +5 -0
@@ -53,6 +53,8 @@ enum memcg_memory_event {
 	MEMCG_HIGH,
 	MEMCG_MAX,
 	MEMCG_OOM,
+	MEMCG_SWAP_MAX,
+	MEMCG_SWAP_FAIL,
 	MEMCG_NR_MEMORY_EVENTS,
 };
 

@@ -208,6 +210,9 @@ struct mem_cgroup {
 	atomic_long_t memory_events[MEMCG_NR_MEMORY_EVENTS];
 	struct cgroup_file events_file;
 
+	/* handle for "memory.swap.events" */
+	struct cgroup_file swap_events_file;
+
 	/* protect arrays of thresholds */
 	struct mutex thresholds_lock;
mm/memcontrol.c +23 -1
@@ -6012,13 +6012,17 @@ int mem_cgroup_try_charge_swap(struct pa
 	if (!memcg)
 		return 0;
 
-	if (!entry.val)
+	if (!entry.val) {
+		memcg_memory_event(memcg, MEMCG_SWAP_FAIL);
 		return 0;
+	}
 
 	memcg = mem_cgroup_id_get_online(memcg);
 
 	if (!mem_cgroup_is_root(memcg) &&
 	    !page_counter_try_charge(&memcg->swap, nr_pages, &counter)) {
+		memcg_memory_event(memcg, MEMCG_SWAP_MAX);
+		memcg_memory_event(memcg, MEMCG_SWAP_FAIL);
 		mem_cgroup_id_put(memcg);
 		return -ENOMEM;
 	}

@@ -6156,6 +6160,18 @@ static ssize_t swap_max_write(struct ker
 	return nbytes;
 }
 
+static int swap_events_show(struct seq_file *m, void *v)
+{
+	struct mem_cgroup *memcg = mem_cgroup_from_css(seq_css(m));
+
+	seq_printf(m, "max %lu\n",
+		   atomic_long_read(&memcg->memory_events[MEMCG_SWAP_MAX]));
+	seq_printf(m, "fail %lu\n",
+		   atomic_long_read(&memcg->memory_events[MEMCG_SWAP_FAIL]));
+
+	return 0;
+}
+
 static struct cftype swap_files[] = {
 	{
 		.name = "swap.current",

@@ -6168,6 +6184,12 @@ static struct cftype swap_files[] = {
 		.seq_show = swap_max_show,
 		.write = swap_max_write,
 	},
+	{
+		.name = "swap.events",
+		.flags = CFTYPE_NOT_ON_ROOT,
+		.file_offset = offsetof(struct mem_cgroup, swap_events_file),
+		.seq_show = swap_events_show,
+	},
 	{ }	/* terminate */
 };