+ mm-memcontrol-convert-nr_anon_thps-account-to-pages.patch added to -mm tree

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



The patch titled
     Subject: mm: memcontrol: convert NR_ANON_THPS account to pages
has been added to the -mm tree.  Its filename is
     mm-memcontrol-convert-nr_anon_thps-account-to-pages.patch

This patch should soon appear at
    https://ozlabs.org/~akpm/mmots/broken-out/mm-memcontrol-convert-nr_anon_thps-account-to-pages.patch
and later at
    https://ozlabs.org/~akpm/mmotm/broken-out/mm-memcontrol-convert-nr_anon_thps-account-to-pages.patch

Before you just go and hit "reply", please:
   a) Consider who else should be cc'ed
   b) Prefer to cc a suitable mailing list as well
   c) Ideally: find the original patch on the mailing list and do a
      reply-to-all to that, adding suitable additional cc's

*** Remember to use Documentation/process/submit-checklist.rst when testing your code ***

The -mm tree is included into linux-next and is updated
there every 3-4 working days

------------------------------------------------------
From: Muchun Song <songmuchun@xxxxxxxxxxxxx>
Subject: mm: memcontrol: convert NR_ANON_THPS account to pages

Currently we use struct per_cpu_nodestat to cache the vmstat counters,
which leads to inaccurate statistics expecially THP vmstat counters.  In
the systems with hundreads of processors it can be GBs of memory.  For
example, for a 96 CPUs system, the threshold is the maximum number of 125.
And the per cpu counters can cache 23.4375 GB in total.

The THP page is already a form of batched addition (it will add 512 worth
of memory in one go) so skipping the batching seems like sensible. 
Although every THP stats update overflows the per-cpu counter, resorting
to atomic global updates.  But it can make the statistics more accuracy
for the THP vmstat counters.

So we convert the NR_ANON_THPS account to pages.  This patch is consistent
with 8f182270dfec ("mm/swap.c: flush lru pvecs on compound page arrival").
Doing this also can make the unit of vmstat counters more unified. 
Finally, the unit of the vmstat counters are pages, kB and bytes.  The
B/KB suffix can tell us that the unit is bytes or kB.  The rest which is
without suffix are pages.

Link: https://lkml.kernel.org/r/20201217034356.4708-3-songmuchun@xxxxxxxxxxxxx
Signed-off-by: Muchun Song <songmuchun@xxxxxxxxxxxxx>
Cc: Alexey Dobriyan <adobriyan@xxxxxxxxx>
Cc: Feng Tang <feng.tang@xxxxxxxxx>
Cc: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
Cc: Hugh Dickins <hughd@xxxxxxxxxx>
Cc: Johannes Weiner <hannes@xxxxxxxxxxx>
Cc: Joonsoo Kim <iamjoonsoo.kim@xxxxxxx>
Cc: Michal Hocko <mhocko@xxxxxxxx>
Cc: NeilBrown <neilb@xxxxxxx>
Cc: Pankaj Gupta <pankaj.gupta@xxxxxxxxxxxxxxx>
Cc: Rafael. J. Wysocki <rafael@xxxxxxxxxx>
Cc: Randy Dunlap <rdunlap@xxxxxxxxxxxxx>
Cc: Roman Gushchin <guro@xxxxxx>
Cc: Sami Tolvanen <samitolvanen@xxxxxxxxxx>
Cc: Shakeel Butt <shakeelb@xxxxxxxxxx>
Cc: Vladimir Davydov <vdavydov.dev@xxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---

 drivers/base/node.c    |   15 +++++++++------
 fs/proc/meminfo.c      |    2 +-
 include/linux/mmzone.h |   10 ++++++++++
 mm/huge_memory.c       |    3 ++-
 mm/memcontrol.c        |   20 ++++++--------------
 mm/page_alloc.c        |    2 +-
 mm/rmap.c              |    6 +++---
 mm/vmstat.c            |   11 +++++++++--
 8 files changed, 41 insertions(+), 28 deletions(-)

--- a/drivers/base/node.c~mm-memcontrol-convert-nr_anon_thps-account-to-pages
+++ a/drivers/base/node.c
@@ -461,8 +461,7 @@ static ssize_t node_read_meminfo(struct
 			     nid, K(sunreclaimable)
 #ifdef CONFIG_TRANSPARENT_HUGEPAGE
 			     ,
-			     nid, K(node_page_state(pgdat, NR_ANON_THPS) *
-				    HPAGE_PMD_NR),
+			     nid, K(node_page_state(pgdat, NR_ANON_THPS)),
 			     nid, K(node_page_state(pgdat, NR_SHMEM_THPS) *
 				    HPAGE_PMD_NR),
 			     nid, K(node_page_state(pgdat, NR_SHMEM_PMDMAPPED) *
@@ -519,10 +518,14 @@ static ssize_t node_read_vmstat(struct d
 				     sum_zone_numa_state(nid, i));
 
 #endif
-	for (i = 0; i < NR_VM_NODE_STAT_ITEMS; i++)
-		len += sysfs_emit_at(buf, len, "%s %lu\n",
-				     node_stat_name(i),
-				     node_page_state_pages(pgdat, i));
+	for (i = 0; i < NR_VM_NODE_STAT_ITEMS; i++) {
+		unsigned long pages = node_page_state_pages(pgdat, i);
+
+		if (vmstat_item_print_in_thp(i))
+			pages /= HPAGE_PMD_NR;
+		len += sysfs_emit_at(buf, len, "%s %lu\n", node_stat_name(i),
+				     pages);
+	}
 
 	return len;
 }
--- a/fs/proc/meminfo.c~mm-memcontrol-convert-nr_anon_thps-account-to-pages
+++ a/fs/proc/meminfo.c
@@ -129,7 +129,7 @@ static int meminfo_proc_show(struct seq_
 
 #ifdef CONFIG_TRANSPARENT_HUGEPAGE
 	show_val_kb(m, "AnonHugePages:  ",
-		    global_node_page_state(NR_ANON_THPS) * HPAGE_PMD_NR);
+		    global_node_page_state(NR_ANON_THPS));
 	show_val_kb(m, "ShmemHugePages: ",
 		    global_node_page_state(NR_SHMEM_THPS) * HPAGE_PMD_NR);
 	show_val_kb(m, "ShmemPmdMapped: ",
--- a/include/linux/mmzone.h~mm-memcontrol-convert-nr_anon_thps-account-to-pages
+++ a/include/linux/mmzone.h
@@ -210,6 +210,16 @@ enum node_stat_item {
 };
 
 /*
+ * Returns true if the item should be printed in THPs (/proc/vmstat
+ * currently prints number of anon, file and shmem THPs. But the item
+ * is charged in pages).
+ */
+static __always_inline bool vmstat_item_print_in_thp(enum node_stat_item item)
+{
+	return item == NR_ANON_THPS;
+}
+
+/*
  * Returns true if the value is measured in bytes (most vmstat values are
  * measured in pages). This defines the API part, the internal representation
  * might be different.
--- a/mm/huge_memory.c~mm-memcontrol-convert-nr_anon_thps-account-to-pages
+++ a/mm/huge_memory.c
@@ -2176,7 +2176,8 @@ static void __split_huge_pmd_locked(stru
 		lock_page_memcg(page);
 		if (atomic_add_negative(-1, compound_mapcount_ptr(page))) {
 			/* Last compound_mapcount is gone. */
-			__dec_lruvec_page_state(page, NR_ANON_THPS);
+			__mod_lruvec_page_state(page, NR_ANON_THPS,
+						-HPAGE_PMD_NR);
 			if (TestClearPageDoubleMap(page)) {
 				/* No need in mapcount reference anymore */
 				for (i = 0; i < HPAGE_PMD_NR; i++)
--- a/mm/memcontrol.c~mm-memcontrol-convert-nr_anon_thps-account-to-pages
+++ a/mm/memcontrol.c
@@ -1533,7 +1533,7 @@ static struct memory_stat memory_stats[]
 	 * on some architectures, the macro of HPAGE_PMD_SIZE is not
 	 * constant(e.g. powerpc).
 	 */
-	{ "anon_thp", 0, NR_ANON_THPS },
+	{ "anon_thp", PAGE_SIZE, NR_ANON_THPS },
 	{ "file_thp", 0, NR_FILE_THPS },
 	{ "shmem_thp", 0, NR_SHMEM_THPS },
 #endif
@@ -1566,8 +1566,7 @@ static int __init memory_stats_init(void
 
 	for (i = 0; i < ARRAY_SIZE(memory_stats); i++) {
 #ifdef CONFIG_TRANSPARENT_HUGEPAGE
-		if (memory_stats[i].idx == NR_ANON_THPS ||
-		    memory_stats[i].idx == NR_FILE_THPS ||
+		if (memory_stats[i].idx == NR_FILE_THPS ||
 		    memory_stats[i].idx == NR_SHMEM_THPS)
 			memory_stats[i].ratio = HPAGE_PMD_SIZE;
 #endif
@@ -4089,10 +4088,6 @@ static int memcg_stat_show(struct seq_fi
 		if (memcg1_stats[i] == MEMCG_SWAP && !do_memsw_account())
 			continue;
 		nr = memcg_page_state_local(memcg, memcg1_stats[i]);
-#ifdef CONFIG_TRANSPARENT_HUGEPAGE
-		if (memcg1_stats[i] == NR_ANON_THPS)
-			nr *= HPAGE_PMD_NR;
-#endif
 		seq_printf(m, "%s %lu\n", memcg1_stat_names[i], nr * PAGE_SIZE);
 	}
 
@@ -4123,10 +4118,6 @@ static int memcg_stat_show(struct seq_fi
 		if (memcg1_stats[i] == MEMCG_SWAP && !do_memsw_account())
 			continue;
 		nr = memcg_page_state(memcg, memcg1_stats[i]);
-#ifdef CONFIG_TRANSPARENT_HUGEPAGE
-		if (memcg1_stats[i] == NR_ANON_THPS)
-			nr *= HPAGE_PMD_NR;
-#endif
 		seq_printf(m, "total_%s %llu\n", memcg1_stat_names[i],
 						(u64)nr * PAGE_SIZE);
 	}
@@ -5654,10 +5645,11 @@ static int mem_cgroup_move_account(struc
 			__mod_lruvec_state(from_vec, NR_ANON_MAPPED, -nr_pages);
 			__mod_lruvec_state(to_vec, NR_ANON_MAPPED, nr_pages);
 			if (PageTransHuge(page)) {
-				__dec_lruvec_state(from_vec, NR_ANON_THPS);
-				__inc_lruvec_state(to_vec, NR_ANON_THPS);
+				__mod_lruvec_state(from_vec, NR_ANON_THPS,
+						   -nr_pages);
+				__mod_lruvec_state(to_vec, NR_ANON_THPS,
+						   nr_pages);
 			}
-
 		}
 	} else {
 		__mod_lruvec_state(from_vec, NR_FILE_PAGES, -nr_pages);
--- a/mm/page_alloc.c~mm-memcontrol-convert-nr_anon_thps-account-to-pages
+++ a/mm/page_alloc.c
@@ -5580,7 +5580,7 @@ void show_free_areas(unsigned int filter
 			K(node_page_state(pgdat, NR_SHMEM_THPS) * HPAGE_PMD_NR),
 			K(node_page_state(pgdat, NR_SHMEM_PMDMAPPED)
 					* HPAGE_PMD_NR),
-			K(node_page_state(pgdat, NR_ANON_THPS) * HPAGE_PMD_NR),
+			K(node_page_state(pgdat, NR_ANON_THPS)),
 #endif
 			K(node_page_state(pgdat, NR_WRITEBACK_TEMP)),
 			node_page_state(pgdat, NR_KERNEL_STACK_KB),
--- a/mm/rmap.c~mm-memcontrol-convert-nr_anon_thps-account-to-pages
+++ a/mm/rmap.c
@@ -1144,7 +1144,7 @@ void do_page_add_anon_rmap(struct page *
 		 * disabled.
 		 */
 		if (compound)
-			__inc_lruvec_page_state(page, NR_ANON_THPS);
+			__mod_lruvec_page_state(page, NR_ANON_THPS, nr);
 		__mod_lruvec_page_state(page, NR_ANON_MAPPED, nr);
 	}
 
@@ -1186,7 +1186,7 @@ void page_add_new_anon_rmap(struct page
 		if (hpage_pincount_available(page))
 			atomic_set(compound_pincount_ptr(page), 0);
 
-		__inc_lruvec_page_state(page, NR_ANON_THPS);
+		__mod_lruvec_page_state(page, NR_ANON_THPS, nr);
 	} else {
 		/* Anon THP always mapped first with PMD */
 		VM_BUG_ON_PAGE(PageTransCompound(page), page);
@@ -1292,7 +1292,7 @@ static void page_remove_anon_compound_rm
 	if (!IS_ENABLED(CONFIG_TRANSPARENT_HUGEPAGE))
 		return;
 
-	__dec_lruvec_page_state(page, NR_ANON_THPS);
+	__mod_lruvec_page_state(page, NR_ANON_THPS, -thp_nr_pages(page));
 
 	if (TestClearPageDoubleMap(page)) {
 		/*
--- a/mm/vmstat.c~mm-memcontrol-convert-nr_anon_thps-account-to-pages
+++ a/mm/vmstat.c
@@ -1619,8 +1619,12 @@ static void zoneinfo_show_print(struct s
 	if (is_zone_first_populated(pgdat, zone)) {
 		seq_printf(m, "\n  per-node stats");
 		for (i = 0; i < NR_VM_NODE_STAT_ITEMS; i++) {
+			unsigned long pages = node_page_state_pages(pgdat, i);
+
+			if (vmstat_item_print_in_thp(i))
+				pages /= HPAGE_PMD_NR;
 			seq_printf(m, "\n      %-12s %lu", node_stat_name(i),
-				   node_page_state_pages(pgdat, i));
+				   pages);
 		}
 	}
 	seq_printf(m,
@@ -1740,8 +1744,11 @@ static void *vmstat_start(struct seq_fil
 	v += NR_VM_NUMA_STAT_ITEMS;
 #endif
 
-	for (i = 0; i < NR_VM_NODE_STAT_ITEMS; i++)
+	for (i = 0; i < NR_VM_NODE_STAT_ITEMS; i++) {
 		v[i] = global_node_page_state_pages(i);
+		if (vmstat_item_print_in_thp(i))
+			v[i] /= HPAGE_PMD_NR;
+	}
 	v += NR_VM_NODE_STAT_ITEMS;
 
 	global_dirty_limits(v + NR_DIRTY_BG_THRESHOLD,
_

Patches currently in -mm which might be from songmuchun@xxxxxxxxxxxxx are

mm-memcontrol-fix-nr_anon_thps-accounting-in-charge-moving.patch
mm-memcontrol-convert-nr_anon_thps-account-to-pages.patch
mm-memcontrol-convert-nr_file_thps-account-to-pages.patch
mm-memcontrol-convert-nr_shmem_thps-account-to-pages.patch
mm-memcontrol-convert-nr_shmem_pmdmapped-account-to-pages.patch
mm-memcontrol-convert-nr_file_pmdmapped-account-to-pages.patch
mm-memcontrol-make-the-slab-calculation-consistent.patch




[Index of Archives]     [Kernel Archive]     [IETF Annouce]     [DCCP]     [Netdev]     [Networking]     [Security]     [Bugtraq]     [Yosemite]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Linux SCSI]

  Powered by Linux