linux_dsm_epyc7002/arch/blackfin/include/asm/processor.h
Christian Borntraeger 79ab11cdb9 locking/core: Introduce cpu_relax_yield()
For spinning loops people do often use barrier() or cpu_relax().
For most architectures cpu_relax and barrier are the same, but on
some architectures cpu_relax can add some latency.
For example on power,sparc64 and arc, cpu_relax can shift the CPU
towards other hardware threads in an SMT environment.
On s390 cpu_relax does even more, it uses an hypercall to the
hypervisor to give up the timeslice.
In contrast to the SMT yielding this can result in larger latencies.
In some places this latency is unwanted, so another variant
"cpu_relax_lowlatency" was introduced. Before this is used in more
and more places, lets revert the logic and provide a cpu_relax_yield
that can be called in places where yielding is more important than
latency. By default this is the same as cpu_relax on all architectures.

Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: Noam Camus <noamc@ezchip.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Russell King <linux@armlinux.org.uk>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Will Deacon <will.deacon@arm.com>
Cc: linuxppc-dev@lists.ozlabs.org
Cc: virtualization@lists.linux-foundation.org
Cc: xen-devel@lists.xenproject.org
Link: http://lkml.kernel.org/r/1477386195-32736-2-git-send-email-borntraeger@de.ibm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2016-11-16 10:15:09 +01:00

153 lines
3.4 KiB
C

/*
* Copyright 2004-2009 Analog Devices Inc.
*
* Licensed under the GPL-2 or later.
*/
#ifndef __ASM_BFIN_PROCESSOR_H
#define __ASM_BFIN_PROCESSOR_H
/*
* Default implementation of macro that returns current
* instruction pointer ("program counter").
*/
#define current_text_addr() ({ __label__ _l; _l: &&_l;})
#include <asm/ptrace.h>
#include <mach/blackfin.h>
static inline unsigned long rdusp(void)
{
unsigned long usp;
__asm__ __volatile__("%0 = usp;\n\t":"=da"(usp));
return usp;
}
static inline void wrusp(unsigned long usp)
{
__asm__ __volatile__("usp = %0;\n\t"::"da"(usp));
}
static inline unsigned long __get_SP(void)
{
unsigned long sp;
__asm__ __volatile__("%0 = sp;\n\t" : "=da"(sp));
return sp;
}
/*
* User space process size: 1st byte beyond user address space.
* Fairly meaningless on nommu. Parts of user programs can be scattered
* in a lot of places, so just disable this by setting it to 0xFFFFFFFF.
*/
#define TASK_SIZE 0xFFFFFFFF
#ifdef __KERNEL__
#define STACK_TOP TASK_SIZE
#endif
#define TASK_UNMAPPED_BASE 0
struct thread_struct {
unsigned long ksp; /* kernel stack pointer */
unsigned long usp; /* user stack pointer */
unsigned short seqstat; /* saved status register */
unsigned long esp0; /* points to SR of stack frame pt_regs */
unsigned long pc; /* instruction pointer */
void * debuggerinfo;
};
#define INIT_THREAD { \
sizeof(init_stack) + (unsigned long) init_stack, 0, \
PS_S, 0, 0 \
}
extern void start_thread(struct pt_regs *regs, unsigned long new_ip,
unsigned long new_sp);
/* Forward declaration, a strange C thing */
struct task_struct;
/* Free all resources held by a thread. */
static inline void release_thread(struct task_struct *dead_task)
{
}
/*
* Return saved PC of a blocked thread.
*/
#define thread_saved_pc(tsk) (tsk->thread.pc)
unsigned long get_wchan(struct task_struct *p);
#define KSTK_EIP(tsk) \
({ \
unsigned long eip = 0; \
if ((tsk)->thread.esp0 > PAGE_SIZE && \
MAP_NR((tsk)->thread.esp0) < max_mapnr) \
eip = ((struct pt_regs *) (tsk)->thread.esp0)->pc; \
eip; })
#define KSTK_ESP(tsk) ((tsk) == current ? rdusp() : (tsk)->thread.usp)
#define cpu_relax() smp_mb()
#define cpu_relax_yield() cpu_relax()
#define cpu_relax_lowlatency() cpu_relax()
/* Get the Silicon Revision of the chip */
static inline uint32_t __pure bfin_revid(void)
{
/* Always use CHIPID, to work around ANOMALY_05000234 */
uint32_t revid = (bfin_read_CHIPID() & CHIPID_VERSION) >> 28;
#ifdef _BOOTROM_GET_DXE_ADDRESS_TWI
/*
* ANOMALY_05000364
* Incorrect Revision Number in DSPID Register
*/
if (ANOMALY_05000364 &&
bfin_read16(_BOOTROM_GET_DXE_ADDRESS_TWI) == 0x2796)
revid = 1;
#endif
return revid;
}
static inline uint16_t __pure bfin_cpuid(void)
{
return (bfin_read_CHIPID() & CHIPID_FAMILY) >> 12;
}
static inline uint32_t __pure bfin_dspid(void)
{
return bfin_read_DSPID();
}
#define blackfin_core_id() (bfin_dspid() & 0xff)
static inline uint32_t __pure bfin_compiled_revid(void)
{
#if defined(CONFIG_BF_REV_0_0)
return 0;
#elif defined(CONFIG_BF_REV_0_1)
return 1;
#elif defined(CONFIG_BF_REV_0_2)
return 2;
#elif defined(CONFIG_BF_REV_0_3)
return 3;
#elif defined(CONFIG_BF_REV_0_4)
return 4;
#elif defined(CONFIG_BF_REV_0_5)
return 5;
#elif defined(CONFIG_BF_REV_0_6)
return 6;
#elif defined(CONFIG_BF_REV_ANY)
return 0xffff;
#else
return -1;
#endif
}
#endif