From 47b8ea7186aae7f474ec4c98f43eaa8da719cd83 Mon Sep 17 00:00:00 2001 From: Rik van Riel Date: Mon, 9 Mar 2015 12:12:08 -0400 Subject: cpusets, isolcpus: exclude isolcpus from load balancing in cpusets Ensure that cpus specified with the isolcpus= boot commandline option stay outside of the load balancing in the kernel scheduler. Operations like load balancing can introduce unwanted latencies, which is exactly what the isolcpus= commandline is there to prevent. Previously, simply creating a new cpuset, without even touching the cpuset.cpus field inside the new cpuset, would undo the effects of isolcpus=, by creating a scheduler domain spanning the whole system, and setting up load balancing inside that domain. The cpuset root cpuset.cpus file is read-only, so there was not even a way to undo that effect. This does not impact the majority of cpusets users, since isolcpus= is a fairly specialized feature used for realtime purposes. Cc: Peter Zijlstra Cc: Clark Williams Cc: Li Zefan Cc: Ingo Molnar Cc: Luiz Capitulino Cc: Mike Galbraith Cc: cgroups@vger.kernel.org Signed-off-by: Rik van Riel Tested-by: David Rientjes Acked-by: Peter Zijlstra (Intel) Acked-by: David Rientjes Acked-by: Zefan Li Signed-off-by: Tejun Heo --- kernel/cpuset.c | 13 +++++++++++-- 1 file changed, 11 insertions(+), 2 deletions(-) (limited to 'kernel/cpuset.c') diff --git a/kernel/cpuset.c b/kernel/cpuset.c index fc7f4748d34a..c68f0721df10 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -622,6 +622,7 @@ static int generate_sched_domains(cpumask_var_t **domains, int csn; /* how many cpuset ptrs in csa so far */ int i, j, k; /* indices for partition finding loops */ cpumask_var_t *doms; /* resulting partition; i.e. sched domains */ + cpumask_var_t non_isolated_cpus; /* load balanced CPUs */ struct sched_domain_attr *dattr; /* attributes for custom domains */ int ndoms = 0; /* number of sched domains in result */ int nslot; /* next empty doms[] struct cpumask slot */ @@ -631,6 +632,10 @@ static int generate_sched_domains(cpumask_var_t **domains, dattr = NULL; csa = NULL; + if (!alloc_cpumask_var(&non_isolated_cpus, GFP_KERNEL)) + goto done; + cpumask_andnot(non_isolated_cpus, cpu_possible_mask, cpu_isolated_map); + /* Special case for the 99% of systems with one, full, sched domain */ if (is_sched_load_balance(&top_cpuset)) { ndoms = 1; @@ -643,7 +648,8 @@ static int generate_sched_domains(cpumask_var_t **domains, *dattr = SD_ATTR_INIT; update_domain_attr_tree(dattr, &top_cpuset); } - cpumask_copy(doms[0], top_cpuset.effective_cpus); + cpumask_and(doms[0], top_cpuset.effective_cpus, + non_isolated_cpus); goto done; } @@ -666,7 +672,8 @@ static int generate_sched_domains(cpumask_var_t **domains, * the corresponding sched domain. */ if (!cpumask_empty(cp->cpus_allowed) && - !is_sched_load_balance(cp)) + !(is_sched_load_balance(cp) && + cpumask_intersects(cp->cpus_allowed, non_isolated_cpus))) continue; if (is_sched_load_balance(cp)) @@ -748,6 +755,7 @@ restart: if (apn == b->pn) { cpumask_or(dp, dp, b->effective_cpus); + cpumask_and(dp, dp, non_isolated_cpus); if (dattr) update_domain_attr_tree(dattr + nslot, b); @@ -760,6 +768,7 @@ restart: BUG_ON(nslot != ndoms); done: + free_cpumask_var(non_isolated_cpus); kfree(csa); /* -- cgit