Skip to contentSkip to conversation

sched/fair: Clear SMT siblings after determining the core is not idle

PATCH Applied Mel Gorman 5 participants 6 messages Nov 30, 2020 – Dec 11, 2020 Scheduler

lore
Share & tools
lore

Mel Gorman

author
#0 lore raw

PATCH 1/1 sched/fair: Clear SMT siblings after determining the core is not idle

The clearing of SMT siblings from the SIS mask before checking for an idle core is a small but unnecessary cost. Defer the clearing of the siblings until the scan moves to the next potential target. The cost of this was not measured as it is borderline noise but it should be self-evident.

---
 kernel/sched/fair.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
kernel/sched/fair.c +2 -1
index 0d54d69ba1a5..d9acd55d309b 100644
@@ -6087,10 +6087,11 @@ static int select_idle_core(struct task_struct *p, struct sched_domain *sd, int
 				break;
 			}
 		}
-		cpumask_andnot(cpus, cpus, cpu_smt_mask(core));
 
 		if (idle)
 			return core;
+
+		cpumask_andnot(cpus, cpus, cpu_smt_mask(core));
 	}
 
 	/*
maintainer Reviewed-by on v1 1/1
#1 lore raw
Quoted 5 lines · Mel Gorman

The clearing of SMT siblings from the SIS mask before checking for an idle core is a small but unnecessary cost. Defer the clearing of the siblings until the scan moves to the next potential target. The cost of this was not measured as it is borderline noise but it should be self-evident.

Good point

Quoted 2 lines
Quoted 21 lines
---
 kernel/sched/fair.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
kernel/sched/fair.c +2 -1
index 0d54d69ba1a5..d9acd55d309b 100644
@@ -6087,10 +6087,11 @@ static int select_idle_core(struct task_struct *p, struct sched_domain *sd, int
                                break;
                        }
                }
-               cpumask_andnot(cpus, cpus, cpu_smt_mask(core));
 
                if (idle)
                        return core;
+
+               cpumask_andnot(cpus, cpus, cpu_smt_mask(core));
        }
 
        /*
#2 lore raw
Quoted 8 lines · Vincent Guittot
Quoted 5 lines · Mel Gorman

The clearing of SMT siblings from the SIS mask before checking for an idle core is a small but unnecessary cost. Defer the clearing of the siblings until the scan moves to the next potential target. The cost of this was not measured as it is borderline noise but it should be self-evident.

Good point

This is more reasonable, thanks Mel.

Quoted 27 lines
Quoted 3 lines
Quoted 21 lines
---
 kernel/sched/fair.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
kernel/sched/fair.c +2 -1
index 0d54d69ba1a5..d9acd55d309b 100644
@@ -6087,10 +6087,11 @@ static int select_idle_core(struct task_struct *p, struct sched_domain *sd, int
                                break;
                        }
                }
-               cpumask_andnot(cpus, cpus, cpu_smt_mask(core));
 
                if (idle)
                        return core;
+
+               cpumask_andnot(cpus, cpus, cpu_smt_mask(core));
        }
 
        /*
maintainer
#3 lore raw
In reply to Vincent Guittot
Quoted 7 lines · Vincent Guittot
Quoted 5 lines · Mel Gorman

The clearing of SMT siblings from the SIS mask before checking for an idle core is a small but unnecessary cost. Defer the clearing of the siblings until the scan moves to the next potential target. The cost of this was not measured as it is borderline noise but it should be self-evident.

Quoted 6 lines

Good point

Quoted 2 lines

Thanks!

merge bot Applied
#4 lore raw
In reply to Mel Gorman
The following commit has been merged into the sched/core branch of tip:

The following commit has been merged into the sched/core branch of tip:

Commit-ID:     82b738de57d571cd366d89e75b5fd60f3060852b
Gitweb:        https://git.kernel.org/tip/82b738de57d571cd366d89e75b5fd60f3060852b
Author:        Mel Gorman <[email protected]>
AuthorDate:    Mon, 30 Nov 2020 14:40:20 
Committer:     Peter Zijlstra <[email protected]>
CommitterDate: Thu, 03 Dec 2020 10:00:36 +01:00

sched/fair: Clear SMT siblings after determining the core is not idle

The clearing of SMT siblings from the SIS mask before checking for an idle core is a small but unnecessary cost. Defer the clearing of the siblings until the scan moves to the next potential target. The cost of this was not measured as it is borderline noise but it should be self-evident.

---
 kernel/sched/fair.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
kernel/sched/fair.c +2 -1
index f5dceda..efac224 100644
@@ -6086,10 +6086,11 @@ static int select_idle_core(struct task_struct *p, struct sched_domain *sd, int 
 				break;
 			}
 		}
-		cpumask_andnot(cpus, cpus, cpu_smt_mask(core));
 
 		if (idle)
 			return core;
+
+		cpumask_andnot(cpus, cpus, cpu_smt_mask(core));
 	}
 
 	/*
merge bot Applied
#5 lore raw
In reply to Mel Gorman
The following commit has been merged into the sched/core branch of tip:

The following commit has been merged into the sched/core branch of tip:

Commit-ID:     13d5a5e9f9b8515da3c04305ae1bb03ab91be7a7
Gitweb:        https://git.kernel.org/tip/13d5a5e9f9b8515da3c04305ae1bb03ab91be7a7
Author:        Mel Gorman <[email protected]>
AuthorDate:    Mon, 30 Nov 2020 14:40:20 
Committer:     Ingo Molnar <[email protected]>
CommitterDate: Fri, 11 Dec 2020 10:30:38 +01:00

sched/fair: Clear SMT siblings after determining the core is not idle

The clearing of SMT siblings from the SIS mask before checking for an idle core is a small but unnecessary cost. Defer the clearing of the siblings until the scan moves to the next potential target. The cost of this was not measured as it is borderline noise but it should be self-evident.

---
 kernel/sched/fair.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
kernel/sched/fair.c +2 -1
index f5dceda..efac224 100644
@@ -6086,10 +6086,11 @@ static int select_idle_core(struct task_struct *p, struct sched_domain *sd, int 
 				break;
 			}
 		}
-		cpumask_andnot(cpus, cpus, cpu_smt_mask(core));
 
 		if (idle)
 			return core;
+
+		cpumask_andnot(cpus, cpus, cpu_smt_mask(core));
 	}
 
 	/*