Messages in this thread Patch in this message | | | From | Jakub Kicinski <> | Subject | [PATCH 2/3] locking/static_key: factor out the fast path of static_key_slow_dec() | Date | Fri, 29 Mar 2019 17:08:53 -0700 |
| |
static_key_slow_dec() checks if the atomic enable count is larger than 1, and if so there decrements it before taking the jump_label_lock. Move this logic into a helper for reuse in rate limitted keys.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com> Reviewed-by: Simon Horman <simon.horman@netronome.com> --- kernel/jump_label.c | 23 +++++++++++++++-------- 1 file changed, 15 insertions(+), 8 deletions(-)
diff --git a/kernel/jump_label.c b/kernel/jump_label.c index 73bbbaddbd9c..02c3d11264dd 100644 --- a/kernel/jump_label.c +++ b/kernel/jump_label.c @@ -202,13 +202,13 @@ void static_key_disable(struct static_key *key) } EXPORT_SYMBOL_GPL(static_key_disable); -static void __static_key_slow_dec_cpuslocked(struct static_key *key, - unsigned long rate_limit, - struct delayed_work *work) +static bool static_key_slow_try_dec(struct static_key *key) { int val; - lockdep_assert_cpus_held(); + val = atomic_fetch_add_unless(&key->enabled, -1, 1); + if (val == 1) + return false; /* * The negative count check is valid even when a negative @@ -217,11 +217,18 @@ static void __static_key_slow_dec_cpuslocked(struct static_key *key, * returns is unbalanced, because all other static_key_slow_inc() * instances block while the update is in progress. */ - val = atomic_fetch_add_unless(&key->enabled, -1, 1); - if (val != 1) { - WARN(val < 0, "jump label: negative count!\n"); + WARN(val < 0, "jump label: negative count!\n"); + return true; +} + +static void __static_key_slow_dec_cpuslocked(struct static_key *key, + unsigned long rate_limit, + struct delayed_work *work) +{ + lockdep_assert_cpus_held(); + + if (static_key_slow_try_dec(key)) return; - } jump_label_lock(); if (atomic_dec_and_test(&key->enabled)) { -- 2.21.0
| |