sched/core: Use try_cmpxchg in set_nr_{and_not,if}_polling

Use try_cmpxchg instead of cmpxchg (*ptr, old, new) != old in
set_nr_{and_not,if}_polling. x86 cmpxchg returns success in ZF flag,
so this change saves a compare after cmpxchg.

The definition of cmpxchg based fetch_or was changed in the
same way as atomic_fetch_##op definitions were changed
in 0bbeb7d6bfe7001fbe7829ca61cd3beb9e3a7479.

Also declare these two functions as inline to ensure inlining. In the
case of set_nr_and_not_polling, the compiler (gcc) tries to outsmart
itself by constructing the boolean return value with logic operations
on the fetched value, and these extra operations enlarge the function
over the inlining threshold value.

Signed-off-by: Uros Bizjak <ubizjak@gmail.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20220629151552.6015-1-ubizjak@gmail.com

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index dd69e85b78791a41cde67da9ecfcbc05f1ef79c3..c703d177f62db55306b97f1bc3b37cd7cc9e5473 100644 (file)
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -873,15 +873,11 @@ static inline void hrtick_rq_init(struct rq *rq)
        ({                                                              \
                typeof(ptr) _ptr = (ptr);                               \
                typeof(mask) _mask = (mask);                            \
-               typeof(*_ptr) _old, _val = *_ptr;                       \
+               typeof(*_ptr) _val = *_ptr;                             \
                                                                        \
-               for (;;) {                                              \
-                       _old = cmpxchg(_ptr, _val, _val | _mask);       \
-                       if (_old == _val)                               \
-                               break;                                  \
-                       _val = _old;                                    \
-               }                                                       \
-       _old;                                                           \
+               do {                                                    \
+               } while (!try_cmpxchg(_ptr, &_val, _val | _mask));      \
+       _val;                                                           \
 })
 
 #if defined(CONFIG_SMP) && defined(TIF_POLLING_NRFLAG)
@@ -890,7 +886,7 @@ static inline void hrtick_rq_init(struct rq *rq)
  * this avoids any races wrt polling state changes and thereby avoids
  * spurious IPIs.
  */
-static bool set_nr_and_not_polling(struct task_struct *p)
+static inline bool set_nr_and_not_polling(struct task_struct *p)
 {
        struct thread_info *ti = task_thread_info(p);
        return !(fetch_or(&ti->flags, _TIF_NEED_RESCHED) & _TIF_POLLING_NRFLAG);
@@ -905,30 +901,28 @@ static bool set_nr_and_not_polling(struct task_struct *p)
 static bool set_nr_if_polling(struct task_struct *p)
 {
        struct thread_info *ti = task_thread_info(p);
-       typeof(ti->flags) old, val = READ_ONCE(ti->flags);
+       typeof(ti->flags) val = READ_ONCE(ti->flags);
 
        for (;;) {
                if (!(val & _TIF_POLLING_NRFLAG))
                        return false;
                if (val & _TIF_NEED_RESCHED)
                        return true;
-               old = cmpxchg(&ti->flags, val, val | _TIF_NEED_RESCHED);
-               if (old == val)
+               if (try_cmpxchg(&ti->flags, &val, val | _TIF_NEED_RESCHED))
                        break;
-               val = old;
        }
        return true;
 }
 
 #else
-static bool set_nr_and_not_polling(struct task_struct *p)
+static inline bool set_nr_and_not_polling(struct task_struct *p)
 {
        set_tsk_need_resched(p);
        return true;
 }
 
 #ifdef CONFIG_SMP
-static bool set_nr_if_polling(struct task_struct *p)
+static inline bool set_nr_if_polling(struct task_struct *p)
 {
        return false;
 }