Contributors: 14
| Author |
Tokens |
Token Proportion |
Commits |
Commit Proportion |
| Peter Zijlstra |
137 |
43.22% |
8 |
30.77% |
| Waiman Long |
101 |
31.86% |
5 |
19.23% |
| Uros Bizjak |
22 |
6.94% |
1 |
3.85% |
| Pan Xinhui |
14 |
4.42% |
1 |
3.85% |
| Juergen Gross |
10 |
3.15% |
1 |
3.85% |
| Will Deacon |
7 |
2.21% |
2 |
7.69% |
| Jeremy Fitzhardinge |
5 |
1.58% |
1 |
3.85% |
| Pavel Machek |
4 |
1.26% |
1 |
3.85% |
| Zwane Mwaikambo |
4 |
1.26% |
1 |
3.85% |
| Duan Zhenzhong |
4 |
1.26% |
1 |
3.85% |
| Glauber de Oliveira Costa |
3 |
0.95% |
1 |
3.85% |
| Marco Elver |
3 |
0.95% |
1 |
3.85% |
| Chen Yu |
2 |
0.63% |
1 |
3.85% |
| Greg Kroah-Hartman |
1 |
0.32% |
1 |
3.85% |
| Total |
317 |
|
26 |
|
/* SPDX-License-Identifier: GPL-2.0 */
#ifndef _ASM_X86_QSPINLOCK_H
#define _ASM_X86_QSPINLOCK_H
#include <linux/jump_label.h>
#include <asm/cpufeature.h>
#include <asm-generic/qspinlock_types.h>
#include <asm/paravirt.h>
#include <asm/rmwcc.h>
#define _Q_PENDING_LOOPS (1 << 9)
#define queued_fetch_set_pending_acquire queued_fetch_set_pending_acquire
static __always_inline u32 queued_fetch_set_pending_acquire(struct qspinlock *lock)
{
u32 val;
/*
* We can't use GEN_BINARY_RMWcc() inside an if() stmt because asm goto
* and CONFIG_PROFILE_ALL_BRANCHES=y results in a label inside a
* statement expression, which GCC doesn't like.
*/
val = GEN_BINARY_RMWcc(LOCK_PREFIX "btsl", lock->val.counter, c,
"I", _Q_PENDING_OFFSET) * _Q_PENDING_VAL;
val |= atomic_read(&lock->val) & ~_Q_PENDING_MASK;
return val;
}
#ifdef CONFIG_PARAVIRT_SPINLOCKS
extern void native_queued_spin_lock_slowpath(struct qspinlock *lock, u32 val);
extern void __pv_init_lock_hash(void);
extern void __pv_queued_spin_lock_slowpath(struct qspinlock *lock, u32 val);
extern void __raw_callee_save___pv_queued_spin_unlock(struct qspinlock *lock);
extern bool nopvspin;
#define queued_spin_unlock queued_spin_unlock
/**
* queued_spin_unlock - release a queued spinlock
* @lock : Pointer to queued spinlock structure
*
* A smp_store_release() on the least-significant byte.
*/
static inline void native_queued_spin_unlock(struct qspinlock *lock)
{
smp_store_release(&lock->locked, 0);
}
static inline void queued_spin_lock_slowpath(struct qspinlock *lock, u32 val)
{
pv_queued_spin_lock_slowpath(lock, val);
}
static inline void queued_spin_unlock(struct qspinlock *lock)
{
kcsan_release();
pv_queued_spin_unlock(lock);
}
#define vcpu_is_preempted vcpu_is_preempted
static inline bool vcpu_is_preempted(long cpu)
{
return pv_vcpu_is_preempted(cpu);
}
#endif
#ifdef CONFIG_PARAVIRT
/*
* virt_spin_lock_key - disables by default the virt_spin_lock() hijack.
*
* Native (and PV wanting native due to vCPU pinning) should keep this key
* disabled. Native does not touch the key.
*
* When in a guest then native_pv_lock_init() enables the key first and
* KVM/XEN might conditionally disable it later in the boot process again.
*/
DECLARE_STATIC_KEY_FALSE(virt_spin_lock_key);
/*
* Shortcut for the queued_spin_lock_slowpath() function that allows
* virt to hijack it.
*
* Returns:
* true - lock has been negotiated, all done;
* false - queued_spin_lock_slowpath() will do its thing.
*/
#define virt_spin_lock virt_spin_lock
static inline bool virt_spin_lock(struct qspinlock *lock)
{
int val;
if (!static_branch_likely(&virt_spin_lock_key))
return false;
/*
* On hypervisors without PARAVIRT_SPINLOCKS support we fall
* back to a Test-and-Set spinlock, because fair locks have
* horrible lock 'holder' preemption issues.
*/
__retry:
val = atomic_read(&lock->val);
if (val || !atomic_try_cmpxchg(&lock->val, &val, _Q_LOCKED_VAL)) {
cpu_relax();
goto __retry;
}
return true;
}
#endif /* CONFIG_PARAVIRT */
#include <asm-generic/qspinlock.h>
#endif /* _ASM_X86_QSPINLOCK_H */