Messages in this thread Patch in this message | | | From | Paolo Bonzini <> | Subject | [PATCH 3/3] cpuset: make nr_cpusets private | Date | Tue, 1 Aug 2017 17:24:06 +0200 |
| |
Any use of key->enabled (that is static_key_enabled and static_key_count) outside jump_label_lock should handle its own serialization. In the case of cpusets_enabled_key, the key is always incremented/decremented under cpuset_mutex, and hence the same rule applies to nr_cpusets. The rule *is* respected currently, but the mutex is static so nr_cpusets should be static too.
Cc: Peter Zijlstra <peterz@infradead.org> Cc: Zefan Li <lizefan@huawei.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> --- include/linux/cpuset.h | 6 ------ kernel/cgroup/cpuset.c | 7 +++++++ 2 files changed, 7 insertions(+), 6 deletions(-)
diff --git a/include/linux/cpuset.h b/include/linux/cpuset.h index 119a3f9604b0..cedcc910b7a7 100644 --- a/include/linux/cpuset.h +++ b/include/linux/cpuset.h @@ -24,12 +24,6 @@ static inline bool cpusets_enabled(void) return static_branch_unlikely(&cpusets_enabled_key); } -static inline int nr_cpusets(void) -{ - /* jump label reference count + the top-level cpuset */ - return static_key_count(&cpusets_enabled_key.key) + 1; -} - static inline void cpuset_inc(void) { static_branch_inc(&cpusets_enabled_key); diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c index ae643412948a..f65db17e9e23 100644 --- a/kernel/cgroup/cpuset.c +++ b/kernel/cgroup/cpuset.c @@ -576,6 +576,13 @@ static void update_domain_attr_tree(struct sched_domain_attr *dattr, rcu_read_unlock(); } +/* Must be called with cpuset_mutex held. */ +static inline int nr_cpusets(void) +{ + /* jump label reference count + the top-level cpuset */ + return static_key_count(&cpusets_enabled_key.key) + 1; +} + /* * generate_sched_domains() * -- 1.8.3.1
| |