renegade-project/linux-next
2
0
Fork 0
mirror of https://github.com/edk2-porting/linux-next.git synced 2024-12-27 14:43:58 +08:00
Code
e50190a834
linux-next/include/asm-ia64/spinlock.h
Martin Schwidefsky ef6edc9746 [PATCH] Directed yield: cpu_relax variants for spinlocks and rw-locks 
On systems running with virtual cpus there is optimization potential in
regard to spinlocks and rw-locks.  If the virtual cpu that has taken a lock
is known to a cpu that wants to acquire the same lock it is beneficial to
yield the timeslice of the virtual cpu in favour of the cpu that has the
lock (directed yield).

With CONFIG_PREEMPT="n" this can be implemented by the architecture without
common code changes.  Powerpc already does this.

With CONFIG_PREEMPT="y" the lock loops are coded with _raw_spin_trylock,
_raw_read_trylock and _raw_write_trylock in kernel/spinlock.c.  If the lock
could not be taken cpu_relax is called.  A directed yield is not possible
because cpu_relax doesn't know anything about the lock.  To be able to
yield the lock in favour of the current lock holder variants of cpu_relax
for spinlocks and rw-locks are needed.  The new _raw_spin_relax,
_raw_read_relax and _raw_write_relax primitives differ from cpu_relax
insofar that they have an argument: a pointer to the lock structure.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Haavard Skinnemoen <hskinnemoen@atmel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-10-01 00:39:21 -07:00

221 lines
6.8 KiB
C
Raw Blame History

#ifndef _ASM_IA64_SPINLOCK_H
#define _ASM_IA64_SPINLOCK_H
/*
* Copyright (C) 1998-2003 Hewlett-Packard Co
* David Mosberger-Tang <davidm@hpl.hp.com>
* Copyright (C) 1999 Walt Drummond <drummond@valinux.com>
*
* This file is used for SMP configurations only.
*/
#include <linux/compiler.h>
#include <linux/kernel.h>
#include <asm/atomic.h>
#include <asm/bitops.h>
#include <asm/intrinsics.h>
#include <asm/system.h>
#define __raw_spin_lock_init(x) ((x)->lock = 0)
#ifdef ASM_SUPPORTED
/*
* Try to get the lock. If we fail to get the lock, make a non-standard call to
* ia64_spinlock_contention(). We do not use a normal call because that would force all
* callers of __raw_spin_lock() to be non-leaf routines. Instead, ia64_spinlock_contention() is
* carefully coded to touch only those registers that __raw_spin_lock() marks "clobbered".
*/
#define IA64_SPINLOCK_CLOBBERS "ar.ccv", "ar.pfs", "p14", "p15", "r27", "r28", "r29", "r30", "b6", "memory"
static inline void
__raw_spin_lock_flags (raw_spinlock_t *lock, unsigned long flags)
{
register volatile unsigned int *ptr asm ("r31") = &lock->lock;
#if (__GNUC__ == 3 && __GNUC_MINOR__ < 3)
# ifdef CONFIG_ITANIUM
/* don't use brl on Itanium... */
asm volatile ("{\n\t"
" mov ar.ccv = r0\n\t"
" mov r28 = ip\n\t"
" mov r30 = 1;;\n\t"
"}\n\t"
"cmpxchg4.acq r30 = [%1], r30, ar.ccv\n\t"
"movl r29 = ia64_spinlock_contention_pre3_4;;\n\t"
"cmp4.ne p14, p0 = r30, r0\n\t"
"mov b6 = r29;;\n\t"
"mov r27=%2\n\t"
"(p14) br.cond.spnt.many b6"
: "=r"(ptr) : "r"(ptr), "r" (flags) : IA64_SPINLOCK_CLOBBERS);
# else
asm volatile ("{\n\t"
" mov ar.ccv = r0\n\t"
" mov r28 = ip\n\t"
" mov r30 = 1;;\n\t"
"}\n\t"
"cmpxchg4.acq r30 = [%1], r30, ar.ccv;;\n\t"
"cmp4.ne p14, p0 = r30, r0\n\t"
"mov r27=%2\n\t"
"(p14) brl.cond.spnt.many ia64_spinlock_contention_pre3_4;;"
: "=r"(ptr) : "r"(ptr), "r" (flags) : IA64_SPINLOCK_CLOBBERS);
# endif /* CONFIG_MCKINLEY */
#else
# ifdef CONFIG_ITANIUM
/* don't use brl on Itanium... */
/* mis-declare, so we get the entry-point, not it's function descriptor: */
asm volatile ("mov r30 = 1\n\t"
"mov r27=%2\n\t"
"mov ar.ccv = r0;;\n\t"
"cmpxchg4.acq r30 = [%0], r30, ar.ccv\n\t"
"movl r29 = ia64_spinlock_contention;;\n\t"
"cmp4.ne p14, p0 = r30, r0\n\t"
"mov b6 = r29;;\n\t"
"(p14) br.call.spnt.many b6 = b6"
: "=r"(ptr) : "r"(ptr), "r" (flags) : IA64_SPINLOCK_CLOBBERS);
# else
asm volatile ("mov r30 = 1\n\t"
"mov r27=%2\n\t"
"mov ar.ccv = r0;;\n\t"
"cmpxchg4.acq r30 = [%0], r30, ar.ccv;;\n\t"
"cmp4.ne p14, p0 = r30, r0\n\t"
"(p14) brl.call.spnt.many b6=ia64_spinlock_contention;;"
: "=r"(ptr) : "r"(ptr), "r" (flags) : IA64_SPINLOCK_CLOBBERS);
# endif /* CONFIG_MCKINLEY */
#endif
}
#define __raw_spin_lock(lock) __raw_spin_lock_flags(lock, 0)
/* Unlock by doing an ordered store and releasing the cacheline with nta */
static inline void __raw_spin_unlock(raw_spinlock_t *x) {
barrier();
asm volatile ("st4.rel.nta [%0] = r0\n\t" :: "r"(x));
}
#else /* !ASM_SUPPORTED */
#define __raw_spin_lock_flags(lock, flags) __raw_spin_lock(lock)
# define __raw_spin_lock(x) \
do { \
__u32 *ia64_spinlock_ptr = (__u32 *) (x); \
__u64 ia64_spinlock_val; \
ia64_spinlock_val = ia64_cmpxchg4_acq(ia64_spinlock_ptr, 1, 0); \
if (unlikely(ia64_spinlock_val)) { \
do { \
while (*ia64_spinlock_ptr) \
ia64_barrier(); \
ia64_spinlock_val = ia64_cmpxchg4_acq(ia64_spinlock_ptr, 1, 0); \
} while (ia64_spinlock_val); \
} \
} while (0)
#define __raw_spin_unlock(x) do { barrier(); ((raw_spinlock_t *) x)->lock = 0; } while (0)
#endif /* !ASM_SUPPORTED */
#define __raw_spin_is_locked(x) ((x)->lock != 0)
#define __raw_spin_trylock(x) (cmpxchg_acq(&(x)->lock, 0, 1) == 0)
#define __raw_spin_unlock_wait(lock) \
do { while (__raw_spin_is_locked(lock)) cpu_relax(); } while (0)
#define __raw_read_can_lock(rw) (*(volatile int *)(rw) >= 0)
#define __raw_write_can_lock(rw) (*(volatile int *)(rw) == 0)
#define __raw_read_lock(rw) \
do { \
raw_rwlock_t *__read_lock_ptr = (rw); \
\
while (unlikely(ia64_fetchadd(1, (int *) __read_lock_ptr, acq) < 0)) { \
ia64_fetchadd(-1, (int *) __read_lock_ptr, rel); \
while (*(volatile int *)__read_lock_ptr < 0) \
cpu_relax(); \
} \
} while (0)
#define __raw_read_unlock(rw) \
do { \
raw_rwlock_t *__read_lock_ptr = (rw); \
ia64_fetchadd(-1, (int *) __read_lock_ptr, rel); \
} while (0)
#ifdef ASM_SUPPORTED
#define __raw_write_lock(rw) \
do { \
__asm__ __volatile__ ( \
"mov ar.ccv = r0\n" \
"dep r29 = -1, r0, 31, 1;;\n" \
"1:\n" \
"ld4 r2 = [%0];;\n" \
"cmp4.eq p0,p7 = r0,r2\n" \
"(p7) br.cond.spnt.few 1b \n" \
"cmpxchg4.acq r2 = [%0], r29, ar.ccv;;\n" \
"cmp4.eq p0,p7 = r0, r2\n" \
"(p7) br.cond.spnt.few 1b;;\n" \
:: "r"(rw) : "ar.ccv", "p7", "r2", "r29", "memory"); \
} while(0)
#define __raw_write_trylock(rw) \
({ \
register long result; \
\
__asm__ __volatile__ ( \
"mov ar.ccv = r0\n" \
"dep r29 = -1, r0, 31, 1;;\n" \
"cmpxchg4.acq %0 = [%1], r29, ar.ccv\n" \
: "=r"(result) : "r"(rw) : "ar.ccv", "r29", "memory"); \
(result == 0); \
})
static inline void __raw_write_unlock(raw_rwlock_t *x)
{
u8 *y = (u8 *)x;
barrier();
asm volatile ("st1.rel.nta [%0] = r0\n\t" :: "r"(y+3) : "memory" );
}
#else /* !ASM_SUPPORTED */
#define __raw_write_lock(l) \
({ \
__u64 ia64_val, ia64_set_val = ia64_dep_mi(-1, 0, 31, 1); \
__u32 *ia64_write_lock_ptr = (__u32 *) (l); \
do { \
while (*ia64_write_lock_ptr) \
ia64_barrier(); \
ia64_val = ia64_cmpxchg4_acq(ia64_write_lock_ptr, ia64_set_val, 0); \
} while (ia64_val); \
})
#define __raw_write_trylock(rw) \
({ \
__u64 ia64_val; \
__u64 ia64_set_val = ia64_dep_mi(-1, 0, 31,1); \
ia64_val = ia64_cmpxchg4_acq((__u32 *)(rw), ia64_set_val, 0); \
(ia64_val == 0); \
})
static inline void __raw_write_unlock(raw_rwlock_t *x)
{
barrier();
x->write_lock = 0;
}
#endif /* !ASM_SUPPORTED */
static inline int __raw_read_trylock(raw_rwlock_t *x)
{
union {
raw_rwlock_t lock;
__u32 word;
} old, new;
old.lock = new.lock = *x;
old.lock.write_lock = new.lock.write_lock = 0;
++new.lock.read_counter;
return (u32)ia64_cmpxchg4_acq((__u32 *)(x), new.word, old.word) == old.word;
}
#define _raw_spin_relax(lock) cpu_relax()
#define _raw_read_relax(lock) cpu_relax()
#define _raw_write_relax(lock) cpu_relax()
#endif /* _ASM_IA64_SPINLOCK_H */
View Git Blame Copy Permalink