Skip to contentSkip to conversation

memcg: add pagetable comsumption to memory.stat

PATCH Superseded Shakeel Butt 3 participants 10 messages Nov 26, 2020 – Nov 30, 2020

Subsystems (9)

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

cover memcg: add pagetable comsumption to memory.stat

Many workloads consumes significant amount of memory in pagetables. This patch series exposes the pagetable comsumption for each memory cgroup.

Shakeel Butt (2):
mm: move lruvec stats update functions to vmstat.h
mm: memcontrol: account pagetables per node

 Documentation/admin-guide/cgroup-v2.rst |   3 +
 arch/nds32/mm/mm-nds32.c                |   6 +-
 drivers/base/node.c                     |   2 +-
 fs/proc/meminfo.c                       |   2 +-
 include/linux/memcontrol.h              | 112 ------------------------
 include/linux/mm.h                      |  11 ++-
 include/linux/mmzone.h                  |   2 +-
 include/linux/vmstat.h                  | 104 ++++++++++++++++++++++
 mm/memcontrol.c                         |  19 ++++
 mm/page_alloc.c                         |   6 +-
 10 files changed, 142 insertions(+), 125 deletions(-)
Signature · 1 line
2.29.2.454.gaff20da3a2-goog
#1 lore raw

PATCH 1/2 mm: move lruvec stats update functions to vmstat.h

This patch does not change any functionality and only move the functions which update the lruvec stats to vmstat.h from memcontrol.h. The main reason for this patch is to be able to use these functions in the page table contructor function which is defined in mm.h and we can not include the memcontrol.h in that file.

---
 include/linux/memcontrol.h | 112 -------------------------------------
 include/linux/vmstat.h     | 104 ++++++++++++++++++++++++++++++++++
 mm/memcontrol.c            |  18 ++++++
 3 files changed, 122 insertions(+), 112 deletions(-)
include/linux/memcontrol.h +0 -112
index 87ed56dc75f9..cd7b9136fb39 100644
@@ -947,8 +947,6 @@ static inline unsigned long lruvec_page_state_local(struct lruvec *lruvec,
 
 void __mod_memcg_lruvec_state(struct lruvec *lruvec, enum node_stat_item idx,
 			      int val);
-void __mod_lruvec_state(struct lruvec *lruvec, enum node_stat_item idx,
-			int val);
 void __mod_lruvec_kmem_state(void *p, enum node_stat_item idx, int val);
 
 static inline void mod_lruvec_kmem_state(void *p, enum node_stat_item idx,
@@ -971,44 +969,6 @@ static inline void mod_memcg_lruvec_state(struct lruvec *lruvec,
 	local_irq_restore(flags);
 }
 
-static inline void mod_lruvec_state(struct lruvec *lruvec,
-				    enum node_stat_item idx, int val)
-{
-	unsigned long flags;
-
-	local_irq_save(flags);
-	__mod_lruvec_state(lruvec, idx, val);
-	local_irq_restore(flags);
-}
-
-static inline void __mod_lruvec_page_state(struct page *page,
-					   enum node_stat_item idx, int val)
-{
-	struct page *head = compound_head(page); /* rmap on tail pages */
-	struct mem_cgroup *memcg = page_memcg(head);
-	pg_data_t *pgdat = page_pgdat(page);
-	struct lruvec *lruvec;
-
-	/* Untracked pages have no memcg, no lruvec. Update only the node */
-	if (!memcg) {
-		__mod_node_page_state(pgdat, idx, val);
-		return;
-	}
-
-	lruvec = mem_cgroup_lruvec(memcg, pgdat);
-	__mod_lruvec_state(lruvec, idx, val);
-}
-
-static inline void mod_lruvec_page_state(struct page *page,
-					 enum node_stat_item idx, int val)
-{
-	unsigned long flags;
-
-	local_irq_save(flags);
-	__mod_lruvec_page_state(page, idx, val);
-	local_irq_restore(flags);
-}
-
 unsigned long mem_cgroup_soft_limit_reclaim(pg_data_t *pgdat, int order,
 						gfp_t gfp_mask,
 						unsigned long *total_scanned);
@@ -1411,30 +1371,6 @@ static inline void __mod_memcg_lruvec_state(struct lruvec *lruvec,
 {
 }
 
-static inline void __mod_lruvec_state(struct lruvec *lruvec,
-				      enum node_stat_item idx, int val)
-{
-	__mod_node_page_state(lruvec_pgdat(lruvec), idx, val);
-}
-
-static inline void mod_lruvec_state(struct lruvec *lruvec,
-				    enum node_stat_item idx, int val)
-{
-	mod_node_page_state(lruvec_pgdat(lruvec), idx, val);
-}
-
-static inline void __mod_lruvec_page_state(struct page *page,
-					   enum node_stat_item idx, int val)
-{
-	__mod_node_page_state(page_pgdat(page), idx, val);
-}
-
-static inline void mod_lruvec_page_state(struct page *page,
-					 enum node_stat_item idx, int val)
-{
-	mod_node_page_state(page_pgdat(page), idx, val);
-}
-
 static inline void __mod_lruvec_kmem_state(void *p, enum node_stat_item idx,
 					   int val)
 {
@@ -1490,30 +1426,6 @@ static inline void lruvec_memcg_debug(struct lruvec *lruvec, struct page *page)
 }
 #endif /* CONFIG_MEMCG */
 
-static inline void __inc_lruvec_state(struct lruvec *lruvec,
-				      enum node_stat_item idx)
-{
-	__mod_lruvec_state(lruvec, idx, 1);
-}
-
-static inline void __dec_lruvec_state(struct lruvec *lruvec,
-				      enum node_stat_item idx)
-{
-	__mod_lruvec_state(lruvec, idx, -1);
-}
-
-static inline void __inc_lruvec_page_state(struct page *page,
-					   enum node_stat_item idx)
-{
-	__mod_lruvec_page_state(page, idx, 1);
-}
-
-static inline void __dec_lruvec_page_state(struct page *page,
-					   enum node_stat_item idx)
-{
-	__mod_lruvec_page_state(page, idx, -1);
-}
-
 static inline void __inc_lruvec_kmem_state(void *p, enum node_stat_item idx)
 {
 	__mod_lruvec_kmem_state(p, idx, 1);
@@ -1524,30 +1436,6 @@ static inline void __dec_lruvec_kmem_state(void *p, enum node_stat_item idx)
 	__mod_lruvec_kmem_state(p, idx, -1);
 }
 
-static inline void inc_lruvec_state(struct lruvec *lruvec,
-				    enum node_stat_item idx)
-{
-	mod_lruvec_state(lruvec, idx, 1);
-}
-
-static inline void dec_lruvec_state(struct lruvec *lruvec,
-				    enum node_stat_item idx)
-{
-	mod_lruvec_state(lruvec, idx, -1);
-}
-
-static inline void inc_lruvec_page_state(struct page *page,
-					 enum node_stat_item idx)
-{
-	mod_lruvec_page_state(page, idx, 1);
-}
-
-static inline void dec_lruvec_page_state(struct page *page,
-					 enum node_stat_item idx)
-{
-	mod_lruvec_page_state(page, idx, -1);
-}
-
 static inline struct lruvec *parent_lruvec(struct lruvec *lruvec)
 {
 	struct mem_cgroup *memcg;
include/linux/vmstat.h +104 -0

108 lines not on this page · raw message

mm/memcontrol.c +18 -0
index 9922f1510956..8b9352ddff0b 100644
@@ -847,6 +847,24 @@ void __mod_lruvec_state(struct lruvec *lruvec, enum node_stat_item idx,
 		__mod_memcg_lruvec_state(lruvec, idx, val);
 }
 
+void __mod_lruvec_page_state(struct page *page, enum node_stat_item idx,
+			     int val)
+{
+	struct page *head = compound_head(page); /* rmap on tail pages */
+	struct mem_cgroup *memcg = page_memcg(head);
+	pg_data_t *pgdat = page_pgdat(page);
+	struct lruvec *lruvec;
+
+	/* Untracked pages have no memcg, no lruvec. Update only the node */
+	if (!memcg) {
+		__mod_node_page_state(pgdat, idx, val);
+		return;
+	}
+
+	lruvec = mem_cgroup_lruvec(memcg, pgdat);
+	__mod_lruvec_state(lruvec, idx, val);
+}
+
 void __mod_lruvec_kmem_state(void *p, enum node_stat_item idx, int val)
 {
 	pg_data_t *pgdat = page_pgdat(virt_to_page(p));
Signature · 1 line
2.29.2.454.gaff20da3a2-goog
maintainer Acked-by on v1 1/2
#2 lore raw

Re: [PATCH 1/2] mm: move lruvec stats update functions to vmstat.h

Quoted 7 lines · Shakeel Butt

This patch does not change any functionality and only move the functions which update the lruvec stats to vmstat.h from memcontrol.h. The main reason for this patch is to be able to use these functions in the page table contructor function which is defined in mm.h and we can not include the memcontrol.h in that file.

This is a better place for this interface in general. The lruvec abstraction, while invented for memcg, isn't specific to memcg at all.

#3 lore raw

PATCH 2/2 mm: memcontrol: account pagetables per node

In reply to Shakeel Butt

For many workloads, pagetable consumption is significant and it makes sense to expose it in the memory.stat for the memory cgroups. However at the moment, the pagetables are accounted per-zone. Converting them to per-node and using the right interface will correctly account for the memory cgroups as well.

---
This patch was posted at [1] and [2] but more work was needed to make it build for all archs.

[1] http://lkml.kernel.org/r/[email protected]
[2] http://lkml.kernel.org/r/[email protected]

 Documentation/admin-guide/cgroup-v2.rst | 3 +++
 arch/nds32/mm/mm-nds32.c                | 6 +++---
 drivers/base/node.c                     | 2 +-
 fs/proc/meminfo.c                       | 2 +-
 include/linux/mm.h                      | 8 ++++----
 include/linux/mmzone.h                  | 2 +-
 mm/memcontrol.c                         | 1 +
 mm/page_alloc.c                         | 6 +++---
 8 files changed, 17 insertions(+), 13 deletions(-)
Documentation/admin-guide/cgroup-v2.rst +3 -0
index 515bb13084a0..63521cd36ce5 100644
@@ -1274,6 +1274,9 @@ PAGE_SIZE multiple when read back.
 	  kernel_stack
 		Amount of memory allocated to kernel stacks.
 
+	  pagetables
+                Amount of memory allocated for page tables.
+
 	  percpu(npn)
 		Amount of memory used for storing per-cpu kernel
 		data structures.
arch/nds32/mm/mm-nds32.c +3 -3
index 55bec50ccc03..f2778f2b39f6 100644
@@ -34,8 +34,8 @@ pgd_t *pgd_alloc(struct mm_struct *mm)
 	cpu_dcache_wb_range((unsigned long)new_pgd,
 			    (unsigned long)new_pgd +
 			    PTRS_PER_PGD * sizeof(pgd_t));
-	inc_zone_page_state(virt_to_page((unsigned long *)new_pgd),
-			    NR_PAGETABLE);
+	inc_lruvec_page_state(virt_to_page((unsigned long *)new_pgd),
+			      NR_PAGETABLE);
 
 	return new_pgd;
 }
@@ -59,7 +59,7 @@ void pgd_free(struct mm_struct *mm, pgd_t * pgd)
 
 	pte = pmd_page(*pmd);
 	pmd_clear(pmd);
-	dec_zone_page_state(virt_to_page((unsigned long *)pgd), NR_PAGETABLE);
+	dec_lruvec_page_state(virt_to_page((unsigned long *)pgd), NR_PAGETABLE);
 	pte_free(mm, pte);
 	mm_dec_nr_ptes(mm);
 	pmd_free(mm, pmd);
drivers/base/node.c +1 -1
index 6ffa470e2984..04f71c7bc3f8 100644
@@ -450,7 +450,7 @@ static ssize_t node_read_meminfo(struct device *dev,
 #ifdef CONFIG_SHADOW_CALL_STACK
 			     nid, node_page_state(pgdat, NR_KERNEL_SCS_KB),
 #endif
-			     nid, K(sum_zone_node_page_state(nid, NR_PAGETABLE)),
+			     nid, K(node_page_state(pgdat, NR_PAGETABLE)),
 			     nid, 0UL,
 			     nid, K(sum_zone_node_page_state(nid, NR_BOUNCE)),
 			     nid, K(node_page_state(pgdat, NR_WRITEBACK_TEMP)),
fs/proc/meminfo.c +1 -1
index 887a5532e449..d6fc74619625 100644
@@ -107,7 +107,7 @@ static int meminfo_proc_show(struct seq_file *m, void *v)
 		   global_node_page_state(NR_KERNEL_SCS_KB));
 #endif
 	show_val_kb(m, "PageTables:     ",
-		    global_zone_page_state(NR_PAGETABLE));
+		    global_node_page_state(NR_PAGETABLE));
 
 	show_val_kb(m, "NFS_Unstable:   ", 0);
 	show_val_kb(m, "Bounce:         ",
include/linux/mm.h +4 -4
index eabe7d9f80d8..d1f64744ace2 100644
@@ -2199,7 +2199,7 @@ static inline bool pgtable_pte_page_ctor(struct page *page)
 	if (!ptlock_init(page))
 		return false;
 	__SetPageTable(page);
-	inc_zone_page_state(page, NR_PAGETABLE);
+	inc_lruvec_page_state(page, NR_PAGETABLE);
 	return true;
 }
 
@@ -2207,7 +2207,7 @@ static inline void pgtable_pte_page_dtor(struct page *page)
 {
 	ptlock_free(page);
 	__ClearPageTable(page);
-	dec_zone_page_state(page, NR_PAGETABLE);
+	dec_lruvec_page_state(page, NR_PAGETABLE);
 }
 
 #define pte_offset_map_lock(mm, pmd, address, ptlp)	\
@@ -2294,7 +2294,7 @@ static inline bool pgtable_pmd_page_ctor(struct page *page)
 	if (!pmd_ptlock_init(page))
 		return false;
 	__SetPageTable(page);
-	inc_zone_page_state(page, NR_PAGETABLE);
+	inc_lruvec_page_state(page, NR_PAGETABLE);
 	return true;
 }
 
@@ -2302,7 +2302,7 @@ static inline void pgtable_pmd_page_dtor(struct page *page)
 {
 	pmd_ptlock_free(page);
 	__ClearPageTable(page);
-	dec_zone_page_state(page, NR_PAGETABLE);
+	dec_lruvec_page_state(page, NR_PAGETABLE);
 }
 
 /*
include/linux/mmzone.h +1 -1
index 15132adaa233..b593316bff3d 100644
@@ -151,7 +151,6 @@ enum zone_stat_item {
 	NR_ZONE_UNEVICTABLE,
 	NR_ZONE_WRITE_PENDING,	/* Count of dirty, writeback and unstable pages */
 	NR_MLOCK,		/* mlock()ed pages found and moved off LRU */
-	NR_PAGETABLE,		/* used for pagetables */
 	/* Second 128 byte cacheline */
 	NR_BOUNCE,
 #if IS_ENABLED(CONFIG_ZSMALLOC)
@@ -206,6 +205,7 @@ enum node_stat_item {
 #if IS_ENABLED(CONFIG_SHADOW_CALL_STACK)
 	NR_KERNEL_SCS_KB,	/* measured in KiB */
 #endif
+	NR_PAGETABLE,		/* used for pagetables */
 	NR_VM_NODE_STAT_ITEMS
 };
 
mm/memcontrol.c +1 -0
index 8b9352ddff0b..b80328f52fb4 100644
@@ -1519,6 +1519,7 @@ static struct memory_stat memory_stats[] = {
 	{ "anon", PAGE_SIZE, NR_ANON_MAPPED },
 	{ "file", PAGE_SIZE, NR_FILE_PAGES },
 	{ "kernel_stack", 1024, NR_KERNEL_STACK_KB },
+	{ "pagetables", PAGE_SIZE, NR_PAGETABLE },
 	{ "percpu", 1, MEMCG_PERCPU_B },
 	{ "sock", PAGE_SIZE, MEMCG_SOCK },
 	{ "shmem", PAGE_SIZE, NR_SHMEM },
mm/page_alloc.c +3 -3
index f91df593bf71..df749013f539 100644
@@ -5520,7 +5520,7 @@ void show_free_areas(unsigned int filter, nodemask_t *nodemask)
 		global_node_page_state_pages(NR_SLAB_UNRECLAIMABLE_B),
 		global_node_page_state(NR_FILE_MAPPED),
 		global_node_page_state(NR_SHMEM),
-		global_zone_page_state(NR_PAGETABLE),
+		global_node_page_state(NR_PAGETABLE),
 		global_zone_page_state(NR_BOUNCE),
 		global_zone_page_state(NR_FREE_PAGES),
 		free_pcp,
@@ -5552,6 +5552,7 @@ void show_free_areas(unsigned int filter, nodemask_t *nodemask)
 #ifdef CONFIG_SHADOW_CALL_STACK
 			" shadow_call_stack:%lukB"
 #endif
+			" pagetables:%lukB"
 			" all_unreclaimable? %s"
 			"\n",
 			pgdat->node_id,
@@ -5577,6 +5578,7 @@ void show_free_areas(unsigned int filter, nodemask_t *nodemask)
 #ifdef CONFIG_SHADOW_CALL_STACK
 			node_page_state(pgdat, NR_KERNEL_SCS_KB),
 #endif
+			K(node_page_state(pgdat, NR_PAGETABLE)),
 			pgdat->kswapd_failures >= MAX_RECLAIM_RETRIES ?
 				"yes" : "no");
 	}
@@ -5608,7 +5610,6 @@ void show_free_areas(unsigned int filter, nodemask_t *nodemask)
 			" present:%lukB"
 			" managed:%lukB"
 			" mlocked:%lukB"
-			" pagetables:%lukB"
 			" bounce:%lukB"
 			" free_pcp:%lukB"
 			" local_pcp:%ukB"
@@ -5629,7 +5630,6 @@ void show_free_areas(unsigned int filter, nodemask_t *nodemask)
 			K(zone->present_pages),
 			K(zone_managed_pages(zone)),
 			K(zone_page_state(zone, NR_MLOCK)),
-			K(zone_page_state(zone, NR_PAGETABLE)),
 			K(zone_page_state(zone, NR_BOUNCE)),
 			K(free_pcp),
 			K(this_cpu_read(zone->pageset->pcp.count)),
Signature · 1 line
2.29.2.454.gaff20da3a2-goog
maintainer Acked-by on v1 2/2
#4 lore raw

Re: [PATCH 2/2] mm: memcontrol: account pagetables per node

Quoted 7 lines · Shakeel Butt

For many workloads, pagetable consumption is significant and it makes sense to expose it in the memory.stat for the memory cgroups. However at the moment, the pagetables are accounted per-zone. Converting them to per-node and using the right interface will correctly account for the memory cgroups as well.

Agreed, this is a useful stat item to have.

Just one trivial issue:

Quoted 11 lines
include/linux/mmzone.h +0 -1
@@ -151,7 +151,6 @@ enum zone_stat_item {
 	NR_ZONE_UNEVICTABLE,
 	NR_ZONE_WRITE_PENDING,	/* Count of dirty, writeback and unstable pages */
 	NR_MLOCK,		/* mlock()ed pages found and moved off LRU */
-	NR_PAGETABLE,		/* used for pagetables */
 	/* Second 128 byte cacheline */
 	NR_BOUNCE,
 #if IS_ENABLED(CONFIG_ZSMALLOC)

include/linux/mmzone.h enum node_stat_item { · line 210

Quoted 6 lines
include/linux/mmzone.h +1 -0
@@ -206,6 +205,7 @@ enum node_stat_item {
 #if IS_ENABLED(CONFIG_SHADOW_CALL_STACK)
 	NR_KERNEL_SCS_KB,	/* measured in KiB */
 #endif
+	NR_PAGETABLE,		/* used for pagetables */
 	NR_VM_NODE_STAT_ITEMS
 };

You need to update mm/vmstat.c::vmstat_text accordingly or /proc/vmstat output will be bogus.

With that fixed, please feel free to add:

#5 lore raw

Re: [PATCH 2/2] mm: memcontrol: account pagetables per node

Quoted 26 lines · Johannes Weiner
Quoted 7 lines · Shakeel Butt

For many workloads, pagetable consumption is significant and it makes sense to expose it in the memory.stat for the memory cgroups. However at the moment, the pagetables are accounted per-zone. Converting them to per-node and using the right interface will correctly account for the memory cgroups as well.

Agreed, this is a useful stat item to have.

Just one trivial issue:

Quoted 12 lines
include/linux/mmzone.h +0 -1
@@ -151,7 +151,6 @@ enum zone_stat_item {
      NR_ZONE_UNEVICTABLE,
      NR_ZONE_WRITE_PENDING,  /* Count of dirty, writeback and unstable pages */
      NR_MLOCK,               /* mlock()ed pages found and moved off LRU */
-     NR_PAGETABLE,           /* used for pagetables */
      /* Second 128 byte cacheline */
      NR_BOUNCE,
 #if IS_ENABLED(CONFIG_ZSMALLOC)
@@ -206,6 +205,7 @@ enum node_stat_item {
 #if IS_ENABLED(CONFIG_SHADOW_CALL_STACK)
Quoted 8 lines

include/linux/mmzone.h enum node_stat_item { · line 210

Quoted 5 lines
include/linux/mmzone.h +1 -0
      NR_KERNEL_SCS_KB,       /* measured in KiB */
 #endif
+     NR_PAGETABLE,           /* used for pagetables */
      NR_VM_NODE_STAT_ITEMS
 };

You need to update mm/vmstat.c::vmstat_text accordingly or /proc/vmstat output will be bogus.

Oh I missed that. Thanks for catching.

Quoted 3 lines

With that fixed, please feel free to add:

Thanks.

#6 lore raw
In reply to Shakeel Butt
Quoted 2 lines · Shakeel Butt

Many workloads consumes significant amount of memory in pagetables. This patch series exposes the pagetable comsumption for each memory cgroup.

Hi Shakeel!

The code looks good to me. However I'm not sure I understand what's the use case for the new statistics? Can you, please, elaborate a bit more here?

From a very first glance, the size of pagetables should be _roughly_ equal to the size_of(pte)/PAGE_SIZE*(size of a cgroup) and should not exceed 1% of the cgroup size. So for all but very large cgroups the value will be in the noise of per-cpu counters. Perhaps I'm missing some important cases?

Thanks!

Quoted 20 lines

Shakeel Butt (2):
mm: move lruvec stats update functions to vmstat.h
mm: memcontrol: account pagetables per node

 Documentation/admin-guide/cgroup-v2.rst |   3 +
 arch/nds32/mm/mm-nds32.c                |   6 +-
 drivers/base/node.c                     |   2 +-
 fs/proc/meminfo.c                       |   2 +-
 include/linux/memcontrol.h              | 112 ------------------------
 include/linux/mm.h                      |  11 ++-
 include/linux/mmzone.h                  |   2 +-
 include/linux/vmstat.h                  | 104 ++++++++++++++++++++++
 mm/memcontrol.c                         |  19 ++++
 mm/page_alloc.c                         |   6 +-
 10 files changed, 142 insertions(+), 125 deletions(-)
Signature · 1 line
2.29.2.454.gaff20da3a2-goog
#7 lore raw
Quoted 7 lines · Roman Gushchin
Quoted 2 lines · Shakeel Butt

Many workloads consumes significant amount of memory in pagetables. This patch series exposes the pagetable comsumption for each memory cgroup.

Hi Shakeel!

Quoted 8 lines

The code looks good to me. However I'm not sure I understand what's the use case for the new statistics? Can you, please, elaborate a bit more here?

From a very first glance, the size of pagetables should be _roughly_ equal to the size_of(pte)/PAGE_SIZE*(size of a cgroup) and should not exceed 1% of the cgroup size. So for all but very large cgroups the value will be in the noise of per-cpu counters. Perhaps I'm missing some important cases?

I think this is in general a good metric to have but one specific use-case we have is the user space network driver which mmaps the memory of the applications for zero copy data transfers. This driver can consume a large amount of memory in page tables. So, this metric becomes really useful here.

Shakeel

#8 lore raw
Quoted 17 lines · Shakeel Butt
Quoted 7 lines · Roman Gushchin
Quoted 2 lines · Shakeel Butt

Many workloads consumes significant amount of memory in pagetables. This patch series exposes the pagetable comsumption for each memory cgroup.

Hi Shakeel!

Quoted 8 lines

The code looks good to me. However I'm not sure I understand what's the use case for the new statistics? Can you, please, elaborate a bit more here?

From a very first glance, the size of pagetables should be _roughly_ equal to the size_of(pte)/PAGE_SIZE*(size of a cgroup) and should not exceed 1% of the cgroup size. So for all but very large cgroups the value will be in the noise of per-cpu counters. Perhaps I'm missing some important cases?

Quoted 5 lines

I think this is in general a good metric to have but one specific use-case we have is the user space network driver which mmaps the memory of the applications for zero copy data transfers. This driver can consume a large amount of memory in page tables. So, this metric becomes really useful here.

Got it, thank you for the explanation!
Would you mind to add this text to the cover letter as an example?

for the series.

Thanks!

#9 lore raw
Quoted 19 lines · Roman Gushchin
Quoted 17 lines · Shakeel Butt
Quoted 15 lines · Roman Gushchin
Quoted 2 lines · Shakeel Butt

Many workloads consumes significant amount of memory in pagetables. This patch series exposes the pagetable comsumption for each memory cgroup.

Hi Shakeel!

The code looks good to me. However I'm not sure I understand what's the use case for the new statistics? Can you, please, elaborate a bit more here?

From a very first glance, the size of pagetables should be _roughly_ equal to the size_of(pte)/PAGE_SIZE*(size of a cgroup) and should not exceed 1% of the cgroup size. So for all but very large cgroups the value will be in the noise of per-cpu counters. Perhaps I'm missing some important cases?

Quoted 8 lines
Quoted 5 lines

I think this is in general a good metric to have but one specific use-case we have is the user space network driver which mmaps the memory of the applications for zero copy data transfers. This driver can consume a large amount of memory in page tables. So, this metric becomes really useful here.

Got it, thank you for the explanation!
Would you mind to add this text to the cover letter as an example?

Sure.

Quoted 3 lines

for the series.

Thanks for the review.