Skip to contentSkip to conversation

mm/slab: Fix drain freelist excessively

PATCH Superseded Wanpeng Li 2 participants 11 messages Jul 3, 2013

Subsystems (2)

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore

Wanpeng Li

author
#0 lore raw

PATCH v2 1/5 mm/slab: Fix drain freelist excessively

Changelog:
v1 -> v2:
* Fix the callers that pass # of objects. Make sure they pass # of slabs.

The drain_freelist is called to drain slabs_free lists for cache reap, cache shrink, memory hotplug callback etc. The tofree parameter should be the number of slab to free instead of the number of slab objects to free.

This patch fix the callers that pass # of objects. Make sure they pass # of slabs.

---
 mm/slab.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)
mm/slab.c +6 -3
index 17298c1..13f5ecc 100644
@@ -1241,7 +1241,8 @@ free_array_cache:
 		n = cachep->node[node];
 		if (!n)
 			continue;
-		drain_freelist(cachep, n, n->free_objects);
+		drain_freelist(cachep, n, (n->free_objects +
+						cachep->num - 1) / cachep->num);
 	}
 }
 
@@ -1408,7 +1409,8 @@ static int __meminit drain_cache_node_node(int node)
 		if (!n)
 			continue;
 
-		drain_freelist(cachep, n, n->free_objects);
+		drain_freelist(cachep, n, (n->free_objects +
+						cachep->num - 1) / cachep->num);
 
 		if (!list_empty(&n->slabs_full) ||
 		    !list_empty(&n->slabs_partial)) {
@@ -2532,7 +2534,8 @@ static int __cache_shrink(struct kmem_cache *cachep)
 		if (!n)
 			continue;
 
-		drain_freelist(cachep, n, n->free_objects);
+		drain_freelist(cachep, n, (n->free_objects +
+						cachep->num - 1) / cachep->num);
 
 		ret += !list_empty(&n->slabs_full) ||
 			!list_empty(&n->slabs_partial);
Signature · 1 line
1.8.1.2
author
#1 lore raw

PATCH v2 2/5 mm/slab: Sharing s_next and s_stop between slab and slub

This patch shares s_next and s_stop between slab and slub.

---
 mm/slab.c        | 10 ----------
 mm/slab.h        |  3 +++
 mm/slab_common.c |  4 ++--
 3 files changed, 5 insertions(+), 12 deletions(-)
mm/slab.c +0 -10
index 13f5ecc..bd14ea2 100644
@@ -4433,16 +4433,6 @@ static int leaks_show(struct seq_file *m, void *p)
 	return 0;
 }
 
-static void *s_next(struct seq_file *m, void *p, loff_t *pos)
-{
-	return seq_list_next(p, &slab_caches, pos);
-}
-
-static void s_stop(struct seq_file *m, void *p)
-{
-	mutex_unlock(&slab_mutex);
-}
-
 static const struct seq_operations slabstats_op = {
 	.start = leaks_start,
 	.next = s_next,
mm/slab.h +3 -0
index f96b49e..95c8860 100644
@@ -271,3 +271,6 @@ struct kmem_cache_node {
 #endif
 
 };
+
+void *s_next(struct seq_file *m, void *p, loff_t *pos);
+void s_stop(struct seq_file *m, void *p);
mm/slab_common.c +2 -2
index 2d41450..d161b81 100644
@@ -531,12 +531,12 @@ static void *s_start(struct seq_file *m, loff_t *pos)
 	return seq_list_start(&slab_caches, *pos);
 }
 
-static void *s_next(struct seq_file *m, void *p, loff_t *pos)
+void *s_next(struct seq_file *m, void *p, loff_t *pos)
 {
 	return seq_list_next(p, &slab_caches, pos);
 }
 
-static void s_stop(struct seq_file *m, void *p)
+void s_stop(struct seq_file *m, void *p)
 {
 	mutex_unlock(&slab_mutex);
 }
Signature · 7 lines
1.8.1.2

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to [email protected].  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
Acked-by on v2 2/5
#2 lore raw

Re: [PATCH v2 2/5] mm/slab: Sharing s_next and s_stop between slab and slub

Quoted 1 line · Wanpeng Li

This patch shares s_next and s_stop between slab and slub.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>

author
#3 lore raw

PATCH v2 3/5 mm/slab: Fix /proc/slabinfo unwriteable for slab

In reply to Wanpeng Li

Slab have some tunables like limit, batchcount, and sharedfactor can be tuned through function slabinfo_write. Commit (b7454ad3: mm/sl[au]b: Move slabinfo processing to slab_common.c) uncorrectly change /proc/slabinfo unwriteable for slab, this patch fix it by revert to original mode.

---
 mm/slab_common.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)
mm/slab_common.c +9 -1
index d161b81..6a5deda 100644
@@ -497,6 +497,13 @@ void __init create_kmalloc_caches(unsigned long flags)
 
 
 #ifdef CONFIG_SLABINFO
+
+#ifdef CONFIG_SLAB
+#define SLABINFO_RIGHTS (S_IWUSR | S_IRUSR)
+#else
+#define SLABINFO_RIGHTS S_IRUSR
+#endif
+
 void print_slabinfo_header(struct seq_file *m)
 {
 	/*
@@ -633,7 +640,8 @@ static const struct file_operations proc_slabinfo_operations = {
 
 static int __init slab_proc_init(void)
 {
-	proc_create("slabinfo", S_IRUSR, NULL, &proc_slabinfo_operations);
+	proc_create("slabinfo", SLABINFO_RIGHTS, NULL,
+						&proc_slabinfo_operations);
 	return 0;
 }
 module_init(slab_proc_init);
Signature · 7 lines
1.8.1.2

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to [email protected].  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
Acked-by on v2 3/5
#4 lore raw

Re: [PATCH v2 3/5] mm/slab: Fix /proc/slabinfo unwriteable for slab

Quoted 4 lines · Wanpeng Li

Slab have some tunables like limit, batchcount, and sharedfactor can be tuned through function slabinfo_write. Commit (b7454ad3: mm/sl[au]b: Move slabinfo processing to slab_common.c) uncorrectly change /proc/slabinfo unwriteable for slab, this patch fix it by revert to original mode.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>

author
#5 lore raw

PATCH v2 4/5 mm/slub: Drop unnecessary nr_partials

In reply to Wanpeng Li

This patch remove unused nr_partials variable.

---
 mm/slub.c | 2 --
 1 file changed, 2 deletions(-)
mm/slub.c +0 -2
index 4649ff0..84b84f4 100644
@@ -5269,7 +5269,6 @@ __initcall(slab_sysfs_init);
 #ifdef CONFIG_SLABINFO
 void get_slabinfo(struct kmem_cache *s, struct slabinfo *sinfo)
 {
-	unsigned long nr_partials = 0;
 	unsigned long nr_slabs = 0;
 	unsigned long nr_objs = 0;
 	unsigned long nr_free = 0;
@@ -5281,7 +5280,6 @@ void get_slabinfo(struct kmem_cache *s, struct slabinfo *sinfo)
 		if (!n)
 			continue;
 
-		nr_partials += n->nr_partial;
 		nr_slabs += atomic_long_read(&n->nr_slabs);
 		nr_objs += atomic_long_read(&n->total_objects);
 		nr_free += count_partial(n, count_free);
Signature · 7 lines
1.8.1.2

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to [email protected].  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
author
#7 lore raw

PATCH v2 5/5 mm/slub: Use node_nr_slabs and node_nr_objs in get_slabinfo

In reply to Wanpeng Li

Use existing interface node_nr_slabs and node_nr_objs to get nr_slabs and nr_objs.

---
 mm/slub.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
mm/slub.c +2 -2
index 84b84f4..d9135a8 100644
@@ -5280,8 +5280,8 @@ void get_slabinfo(struct kmem_cache *s, struct slabinfo *sinfo)
 		if (!n)
 			continue;
 
-		nr_slabs += atomic_long_read(&n->nr_slabs);
-		nr_objs += atomic_long_read(&n->total_objects);
+		nr_slabs += node_nr_slabs(n);
+		nr_objs += node_nr_objs(n);
 		nr_free += count_partial(n, count_free);
 	}
 
Signature · 7 lines
1.8.1.2

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to [email protected].  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
author
#9 lore raw
Quoted 6 lines · Christoph Lameter
Quoted 2 lines · Wanpeng Li

This patch fix the callers that pass # of objects. Make sure they pass # of slabs.

Hmm... These modifications are all the same. Create a new function?

Ok, I will introduce a helper function, thanks for your review, Christoph. ;-)

Regards,
Wanpeng Li

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>

author
#10 lore raw
In reply to Christoph Lameter
Quoted 6 lines · Christoph Lameter
Quoted 2 lines · Wanpeng Li

This patch fix the callers that pass # of objects. Make sure they pass # of slabs.

Hmm... These modifications are all the same. Create a new function?

Ok, I will introduce a helper function, thanks for your review, Christoph. ;-)

Regards,
Wanpeng Li

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>