Skip to contentSkip to conversation

mips: micro-optimize calculate_cpu_foreign_map()

PATCH In review Yury Norov 2 participants 3 messages Jun 21, 2022 – Jul 5, 2022

Subsystems (2)
lore
Share & tools
lore

Yury Norov

author
#0 lore raw

PATCH 1/1 mips: micro-optimize calculate_cpu_foreign_map()

The inner loop in calculate_cpu_foreign_map() walks the whole cpumask to check if we have siblings for a given cpu.

We can just break after a 1st match and avoid useless traversing the rest of mask.

Loongarch has an identical function, so fix it here as well.

---

It looks like loongarch copied quite a lot from arch/mips/kernel/smp.c. For loongarch folks: Guys, can you consider moving shared code into a shared file(s)?

 arch/loongarch/kernel/smp.c | 7 +++----
 arch/mips/kernel/smp.c      | 7 +++----
 2 files changed, 6 insertions(+), 8 deletions(-)
arch/loongarch/kernel/smp.c +3 -4
index b8c53b755a25..c1c91df3c8ac 100644
@@ -463,17 +463,16 @@ static inline void set_cpu_core_map(int cpu)
  */
 void calculate_cpu_foreign_map(void)
 {
-	int i, k, core_present;
+	int i, k;
 	cpumask_t temp_foreign_map;
 
 	/* Re-calculate the mask */
 	cpumask_clear(&temp_foreign_map);
 	for_each_online_cpu(i) {
-		core_present = 0;
 		for_each_cpu(k, &temp_foreign_map)
 			if (cpus_are_siblings(i, k))
-				core_present = 1;
-		if (!core_present)
+				break;
+		if (k >= nr_cpu_ids)
 			cpumask_set_cpu(i, &temp_foreign_map);
 	}
 
arch/mips/kernel/smp.c +3 -4
index 1d93b85271ba..a2ce641f5f18 100644
@@ -115,17 +115,16 @@ static inline void set_cpu_core_map(int cpu)
  */
 void calculate_cpu_foreign_map(void)
 {
-	int i, k, core_present;
+	int i, k;
 	cpumask_t temp_foreign_map;
 
 	/* Re-calculate the mask */
 	cpumask_clear(&temp_foreign_map);
 	for_each_online_cpu(i) {
-		core_present = 0;
 		for_each_cpu(k, &temp_foreign_map)
 			if (cpus_are_siblings(i, k))
-				core_present = 1;
-		if (!core_present)
+				break;
+		if (k >= nr_cpu_ids)
 			cpumask_set_cpu(i, &temp_foreign_map);
 	}
 
Signature · 1 line
2.32.0
author
#1 lore raw
Quoted 9 lines · Yury Norov

The inner loop in calculate_cpu_foreign_map() walks the whole cpumask to check if we have siblings for a given cpu.

We can just break after a 1st match and avoid useless traversing the rest of mask.

Loongarch has an identical function, so fix it here as well.

Ping?

Quoted 62 lines

---

It looks like loongarch copied quite a lot from arch/mips/kernel/smp.c. For loongarch folks: Guys, can you consider moving shared code into a shared file(s)?

 arch/loongarch/kernel/smp.c | 7 +++----
 arch/mips/kernel/smp.c      | 7 +++----
 2 files changed, 6 insertions(+), 8 deletions(-)
arch/loongarch/kernel/smp.c +3 -4
index b8c53b755a25..c1c91df3c8ac 100644
@@ -463,17 +463,16 @@ static inline void set_cpu_core_map(int cpu)
  */
 void calculate_cpu_foreign_map(void)
 {
-	int i, k, core_present;
+	int i, k;
 	cpumask_t temp_foreign_map;
 
 	/* Re-calculate the mask */
 	cpumask_clear(&temp_foreign_map);
 	for_each_online_cpu(i) {
-		core_present = 0;
 		for_each_cpu(k, &temp_foreign_map)
 			if (cpus_are_siblings(i, k))
-				core_present = 1;
-		if (!core_present)
+				break;
+		if (k >= nr_cpu_ids)
 			cpumask_set_cpu(i, &temp_foreign_map);
 	}
 
arch/mips/kernel/smp.c +3 -4
index 1d93b85271ba..a2ce641f5f18 100644
@@ -115,17 +115,16 @@ static inline void set_cpu_core_map(int cpu)
  */
 void calculate_cpu_foreign_map(void)
 {
-	int i, k, core_present;
+	int i, k;
 	cpumask_t temp_foreign_map;
 
 	/* Re-calculate the mask */
 	cpumask_clear(&temp_foreign_map);
 	for_each_online_cpu(i) {
-		core_present = 0;
 		for_each_cpu(k, &temp_foreign_map)
 			if (cpus_are_siblings(i, k))
-				core_present = 1;
-		if (!core_present)
+				break;
+		if (k >= nr_cpu_ids)
 			cpumask_set_cpu(i, &temp_foreign_map);
 	}
 
Signature · 1 line
2.32.0
#2 lore raw
Quoted 4 lines · Yury Norov
Quoted 3 lines · Yury Norov

The inner loop in calculate_cpu_foreign_map() walks the whole cpumask to check if we have siblings for a given cpu.

Quoted 8 lines
Quoted 6 lines

We can just break after a 1st match and avoid useless traversing the rest of mask.

Loongarch has an identical function, so fix it here as well.

Ping?

as this is touching two different archs, it would have been easier to send this in two different patches.

Huacai, should I take the patch ?

Thomas.

Signature · 2 lines
Crap can work. Given enough thrust pigs will fly, but it's not necessarily a
good idea.                                                [ RFC1925, 2.3 ]