diff options
author | Suren Baghdasaryan <surenb@google.com> | 2021-05-24 12:53:39 -0700 |
---|---|---|
committer | Tejun Heo <tj@kernel.org> | 2021-06-08 14:59:02 -0400 |
commit | 3958e2d0c34e18c41b60dc01832bd670a59ef70f (patch) | |
tree | 429f3434bd8e89fb688e9366c9087208b31fcda3 /kernel/sched | |
parent | 2ca11b0e043be6f5c2b188897e9a32275eaab046 (diff) |
cgroup: make per-cgroup pressure stall tracking configurable
PSI accounts stalls for each cgroup separately and aggregates it at each
level of the hierarchy. This causes additional overhead with psi_avgs_work
being called for each cgroup in the hierarchy. psi_avgs_work has been
highly optimized, however on systems with large number of cgroups the
overhead becomes noticeable.
Systems which use PSI only at the system level could avoid this overhead
if PSI can be configured to skip per-cgroup stall accounting.
Add "cgroup_disable=pressure" kernel command-line option to allow
requesting system-wide only pressure stall accounting. When set, it
keeps system-wide accounting under /proc/pressure/ but skips accounting
for individual cgroups and does not expose PSI nodes in cgroup hierarchy.
Signed-off-by: Suren Baghdasaryan <surenb@google.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Tejun Heo <tj@kernel.org>
Diffstat (limited to 'kernel/sched')
-rw-r--r-- | kernel/sched/psi.c | 30 |
1 files changed, 17 insertions, 13 deletions
diff --git a/kernel/sched/psi.c b/kernel/sched/psi.c index cc25a3cff41f..b773cae4c24b 100644 --- a/kernel/sched/psi.c +++ b/kernel/sched/psi.c @@ -148,6 +148,7 @@ static int psi_bug __read_mostly; DEFINE_STATIC_KEY_FALSE(psi_disabled); +DEFINE_STATIC_KEY_TRUE(psi_cgroups_enabled); #ifdef CONFIG_PSI_DEFAULT_DISABLED static bool psi_enable; @@ -211,6 +212,9 @@ void __init psi_init(void) return; } + if (!cgroup_psi_enabled()) + static_branch_disable(&psi_cgroups_enabled); + psi_period = jiffies_to_nsecs(PSI_FREQ); group_init(&psi_system); } @@ -744,23 +748,23 @@ static void psi_group_change(struct psi_group *group, int cpu, static struct psi_group *iterate_groups(struct task_struct *task, void **iter) { + if (*iter == &psi_system) + return NULL; + #ifdef CONFIG_CGROUPS - struct cgroup *cgroup = NULL; + if (static_branch_likely(&psi_cgroups_enabled)) { + struct cgroup *cgroup = NULL; - if (!*iter) - cgroup = task->cgroups->dfl_cgrp; - else if (*iter == &psi_system) - return NULL; - else - cgroup = cgroup_parent(*iter); + if (!*iter) + cgroup = task->cgroups->dfl_cgrp; + else + cgroup = cgroup_parent(*iter); - if (cgroup && cgroup_parent(cgroup)) { - *iter = cgroup; - return cgroup_psi(cgroup); + if (cgroup && cgroup_parent(cgroup)) { + *iter = cgroup; + return cgroup_psi(cgroup); + } } -#else - if (*iter) - return NULL; #endif *iter = &psi_system; return &psi_system; |