summaryrefslogtreecommitdiffstats
path: root/include
diff options
context:
space:
mode:
authorJesper Nilsson <jesper.nilsson@axis.com>2008-01-28 18:07:58 +0100
committerJesper Nilsson <jesper.nilsson@axis.com>2008-02-08 11:06:39 +0100
commitde1c1419f9eb8f9d719aaaa3e3f3073069ecd657 (patch)
tree5f8e201b62ae6c197d383333b426870413a57a2c /include
parent4258fb19320ae083064aef21b042a02a509119b9 (diff)
downloadop-kernel-dev-de1c1419f9eb8f9d719aaaa3e3f3073069ecd657.zip
op-kernel-dev-de1c1419f9eb8f9d719aaaa3e3f3073069ecd657.tar.gz
CRIS v32: Completely rework spinlocks for ETRAX FS and ARTPEC-3
Diffstat (limited to 'include')
-rw-r--r--include/asm-cris/arch-v32/spinlock.h172
1 files changed, 67 insertions, 105 deletions
diff --git a/include/asm-cris/arch-v32/spinlock.h b/include/asm-cris/arch-v32/spinlock.h
index 5f43df0..0d5709b 100644
--- a/include/asm-cris/arch-v32/spinlock.h
+++ b/include/asm-cris/arch-v32/spinlock.h
@@ -1,40 +1,47 @@
#ifndef __ASM_ARCH_SPINLOCK_H
#define __ASM_ARCH_SPINLOCK_H
-#include <asm/system.h>
+#include <linux/spinlock_types.h>
#define RW_LOCK_BIAS 0x01000000
-#define SPIN_LOCK_UNLOCKED (spinlock_t) { 1 }
-#define spin_lock_init(x) do { *(x) = SPIN_LOCK_UNLOCKED; } while(0)
-
-#define spin_is_locked(x) (*(volatile signed char *)(&(x)->lock) <= 0)
-#define spin_unlock_wait(x) do { barrier(); } while(spin_is_locked(x))
extern void cris_spin_unlock(void *l, int val);
extern void cris_spin_lock(void *l);
-extern int cris_spin_trylock(void* l);
+extern int cris_spin_trylock(void *l);
-static inline void _raw_spin_unlock(spinlock_t *lock)
+static inline int __raw_spin_is_locked(raw_spinlock_t *x)
+{
+ return *(volatile signed char *)(&(x)->slock) <= 0;
+}
+
+static inline void __raw_spin_unlock(raw_spinlock_t *lock)
{
__asm__ volatile ("move.d %1,%0" \
- : "=m" (lock->lock) \
+ : "=m" (lock->slock) \
: "r" (1) \
: "memory");
}
-static inline int _raw_spin_trylock(spinlock_t *lock)
+static inline void __raw_spin_unlock_wait(raw_spinlock_t *lock)
+{
+ while (__raw_spin_is_locked(lock))
+ cpu_relax();
+}
+
+static inline int __raw_spin_trylock(raw_spinlock_t *lock)
{
- return cris_spin_trylock((void*)&lock->lock);
+ return cris_spin_trylock((void *)&lock->slock);
}
-static inline void _raw_spin_lock(spinlock_t *lock)
+static inline void __raw_spin_lock(raw_spinlock_t *lock)
{
- cris_spin_lock((void*)&lock->lock);
+ cris_spin_lock((void *)&lock->slock);
}
-static inline void _raw_spin_lock_flags (spinlock_t *lock, unsigned long flags)
+static inline void
+__raw_spin_lock_flags(raw_spinlock_t *lock, unsigned long flags)
{
- _raw_spin_lock(lock);
+ __raw_spin_lock(lock);
}
/*
@@ -46,120 +53,75 @@ static inline void _raw_spin_lock_flags (spinlock_t *lock, unsigned long flags)
* can "mix" irq-safe locks - any writer needs to get a
* irq-safe write-lock, but readers can get non-irqsafe
* read-locks.
+ *
*/
-typedef struct {
- spinlock_t lock;
- volatile int counter;
-#ifdef CONFIG_PREEMPT
- unsigned int break_lock;
-#endif
-} rwlock_t;
-
-#define RW_LOCK_UNLOCKED (rwlock_t) { {1}, 0 }
-
-#define rwlock_init(lp) do { *(lp) = RW_LOCK_UNLOCKED; } while (0)
-
-/**
- * read_can_lock - would read_trylock() succeed?
- * @lock: the rwlock in question.
- */
-#define read_can_lock(x) ((int)(x)->counter >= 0)
-
-/**
- * write_can_lock - would write_trylock() succeed?
- * @lock: the rwlock in question.
- */
-#define write_can_lock(x) ((x)->counter == 0)
-
-#define _raw_read_trylock(lock) generic_raw_read_trylock(lock)
-/* read_lock, read_unlock are pretty straightforward. Of course it somehow
- * sucks we end up saving/restoring flags twice for read_lock_irqsave aso. */
-
-static __inline__ void _raw_read_lock(rwlock_t *rw)
+static inline int __raw_read_can_lock(raw_rwlock_t *x)
{
- unsigned long flags;
- local_irq_save(flags);
- _raw_spin_lock(&rw->lock);
-
- rw->counter++;
-
- _raw_spin_unlock(&rw->lock);
- local_irq_restore(flags);
+ return (int)(x)->lock > 0;
}
-static __inline__ void _raw_read_unlock(rwlock_t *rw)
+static inline int __raw_write_can_lock(raw_rwlock_t *x)
{
- unsigned long flags;
- local_irq_save(flags);
- _raw_spin_lock(&rw->lock);
-
- rw->counter--;
-
- _raw_spin_unlock(&rw->lock);
- local_irq_restore(flags);
+ return (x)->lock == RW_LOCK_BIAS;
}
-/* write_lock is less trivial. We optimistically grab the lock and check
- * if we surprised any readers. If so we release the lock and wait till
- * they're all gone before trying again
- *
- * Also note that we don't use the _irqsave / _irqrestore suffixes here.
- * If we're called with interrupts enabled and we've got readers (or other
- * writers) in interrupt handlers someone fucked up and we'd dead-lock
- * sooner or later anyway. prumpf */
-
-static __inline__ void _raw_write_lock(rwlock_t *rw)
+static inline void __raw_read_lock(raw_rwlock_t *rw)
{
-retry:
- _raw_spin_lock(&rw->lock);
-
- if(rw->counter != 0) {
- /* this basically never happens */
- _raw_spin_unlock(&rw->lock);
-
- while(rw->counter != 0);
-
- goto retry;
- }
-
- /* got it. now leave without unlocking */
- rw->counter = -1; /* remember we are locked */
+ __raw_spin_lock(&rw->slock);
+ while (rw->lock == 0);
+ rw->lock--;
+ __raw_spin_unlock(&rw->slock);
}
-/* write_unlock is absolutely trivial - we don't have to wait for anything */
-
-static __inline__ void _raw_write_unlock(rwlock_t *rw)
+static inline void __raw_write_lock(raw_rwlock_t *rw)
{
- rw->counter = 0;
- _raw_spin_unlock(&rw->lock);
+ __raw_spin_lock(&rw->slock);
+ while (rw->lock != RW_LOCK_BIAS);
+ rw->lock == 0;
+ __raw_spin_unlock(&rw->slock);
}
-static __inline__ int _raw_write_trylock(rwlock_t *rw)
+static inline void __raw_read_unlock(raw_rwlock_t *rw)
{
- _raw_spin_lock(&rw->lock);
- if (rw->counter != 0) {
- /* this basically never happens */
- _raw_spin_unlock(&rw->lock);
-
- return 0;
- }
+ __raw_spin_lock(&rw->slock);
+ rw->lock++;
+ __raw_spin_unlock(&rw->slock);
+}
- /* got it. now leave without unlocking */
- rw->counter = -1; /* remember we are locked */
- return 1;
+static inline void __raw_write_unlock(raw_rwlock_t *rw)
+{
+ __raw_spin_lock(&rw->slock);
+ while (rw->lock != RW_LOCK_BIAS);
+ rw->lock == RW_LOCK_BIAS;
+ __raw_spin_unlock(&rw->slock);
}
-static __inline__ int is_read_locked(rwlock_t *rw)
+static inline int __raw_read_trylock(raw_rwlock_t *rw)
{
- return rw->counter > 0;
+ int ret = 0;
+ __raw_spin_lock(&rw->slock);
+ if (rw->lock != 0) {
+ rw->lock--;
+ ret = 1;
+ }
+ __raw_spin_unlock(&rw->slock);
+ return ret;
}
-static __inline__ int is_write_locked(rwlock_t *rw)
+static inline int __raw_write_trylock(raw_rwlock_t *rw)
{
- return rw->counter < 0;
+ int ret = 0;
+ __raw_spin_lock(&rw->slock);
+ if (rw->lock == RW_LOCK_BIAS) {
+ rw->lock == 0;
+ ret = 1;
+ }
+ __raw_spin_unlock(&rw->slock);
+ return 1;
}
+
#define _raw_spin_relax(lock) cpu_relax()
#define _raw_read_relax(lock) cpu_relax()
#define _raw_write_relax(lock) cpu_relax()
OpenPOWER on IntegriCloud