michaelh/zephyr
1
0
Fork 0
Code Releases Activity

kernel/spinlock: Force inlining

Browse source 
Something is going wrong with code generation here, potentially the
inline assembly generated by _arch_irq_un/lock(), and these calls are
not being inlined by gcc.  So what should be a ~3 instruction sequence
on most uniprocessor architectures is turning into 8-20 cycles worth
of work to implement the API as written.

Use an ALWAYS_INLINE, which is sort of ugly semantically but produces
much better code.

Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
This commit is contained in:
Andy Ross 2019-01-28 09:35:37 -08:00 • committed by Anas Nashif
commit 4ff2dfce09
1 changed files with 3 additions and 3 deletions
Download patch file Download diff file Expand all files Collapse all files

6
include/spinlock.h
View file

@ -32,8 +32,7 @@ struct k_spinlock {
#endif
};
static inline k_spinlock_key_t k_spin_lock(struct k_spinlock *l)
static ALWAYS_INLINE k_spinlock_key_t k_spin_lock(struct k_spinlock *l)
{
k_spinlock_key_t k;
@ -59,7 +58,8 @@ static inline k_spinlock_key_t k_spin_lock(struct k_spinlock *l)
return k;
}
static inline void k_spin_unlock(struct k_spinlock *l, k_spinlock_key_t key)
static ALWAYS_INLINE void k_spin_unlock(struct k_spinlock *l,
k_spinlock_key_t key)
{
#ifdef SPIN_VALIDATE
__ASSERT(l->thread_cpu == (_current_cpu->id | (u32_t)_current),