![Chen, Kenneth W](/assets/img/avatar_default.png)
For architecture like ia64, the switch stack structure is fairly large (currently 528 bytes). For context switch intensive application, we found that significant amount of cache misses occurs in switch_to() function. The following patch adds a hook in the schedule() function to prefetch switch stack structure as soon as 'next' task is determined. This allows maximum overlap in prefetch cache lines for that structure. Signed-off-by: Ken Chen <kenneth.w.chen@intel.com> Cc: Ingo Molnar <mingo@elte.hu> Cc: "Luck, Tony" <tony.luck@intel.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
290 lines
9.1 KiB
C
290 lines
9.1 KiB
C
#ifndef _ASM_IA64_SYSTEM_H
|
|
#define _ASM_IA64_SYSTEM_H
|
|
|
|
/*
|
|
* System defines. Note that this is included both from .c and .S
|
|
* files, so it does only defines, not any C code. This is based
|
|
* on information published in the Processor Abstraction Layer
|
|
* and the System Abstraction Layer manual.
|
|
*
|
|
* Copyright (C) 1998-2003 Hewlett-Packard Co
|
|
* David Mosberger-Tang <davidm@hpl.hp.com>
|
|
* Copyright (C) 1999 Asit Mallick <asit.k.mallick@intel.com>
|
|
* Copyright (C) 1999 Don Dugger <don.dugger@intel.com>
|
|
*/
|
|
#include <linux/config.h>
|
|
|
|
#include <asm/kregs.h>
|
|
#include <asm/page.h>
|
|
#include <asm/pal.h>
|
|
#include <asm/percpu.h>
|
|
|
|
#define GATE_ADDR RGN_BASE(RGN_GATE)
|
|
|
|
/*
|
|
* 0xa000000000000000+2*PERCPU_PAGE_SIZE
|
|
* - 0xa000000000000000+3*PERCPU_PAGE_SIZE remain unmapped (guard page)
|
|
*/
|
|
#define KERNEL_START (GATE_ADDR+0x100000000)
|
|
#define PERCPU_ADDR (-PERCPU_PAGE_SIZE)
|
|
|
|
#ifndef __ASSEMBLY__
|
|
|
|
#include <linux/kernel.h>
|
|
#include <linux/types.h>
|
|
|
|
struct pci_vector_struct {
|
|
__u16 segment; /* PCI Segment number */
|
|
__u16 bus; /* PCI Bus number */
|
|
__u32 pci_id; /* ACPI split 16 bits device, 16 bits function (see section 6.1.1) */
|
|
__u8 pin; /* PCI PIN (0 = A, 1 = B, 2 = C, 3 = D) */
|
|
__u32 irq; /* IRQ assigned */
|
|
};
|
|
|
|
extern struct ia64_boot_param {
|
|
__u64 command_line; /* physical address of command line arguments */
|
|
__u64 efi_systab; /* physical address of EFI system table */
|
|
__u64 efi_memmap; /* physical address of EFI memory map */
|
|
__u64 efi_memmap_size; /* size of EFI memory map */
|
|
__u64 efi_memdesc_size; /* size of an EFI memory map descriptor */
|
|
__u32 efi_memdesc_version; /* memory descriptor version */
|
|
struct {
|
|
__u16 num_cols; /* number of columns on console output device */
|
|
__u16 num_rows; /* number of rows on console output device */
|
|
__u16 orig_x; /* cursor's x position */
|
|
__u16 orig_y; /* cursor's y position */
|
|
} console_info;
|
|
__u64 fpswa; /* physical address of the fpswa interface */
|
|
__u64 initrd_start;
|
|
__u64 initrd_size;
|
|
} *ia64_boot_param;
|
|
|
|
/*
|
|
* Macros to force memory ordering. In these descriptions, "previous"
|
|
* and "subsequent" refer to program order; "visible" means that all
|
|
* architecturally visible effects of a memory access have occurred
|
|
* (at a minimum, this means the memory has been read or written).
|
|
*
|
|
* wmb(): Guarantees that all preceding stores to memory-
|
|
* like regions are visible before any subsequent
|
|
* stores and that all following stores will be
|
|
* visible only after all previous stores.
|
|
* rmb(): Like wmb(), but for reads.
|
|
* mb(): wmb()/rmb() combo, i.e., all previous memory
|
|
* accesses are visible before all subsequent
|
|
* accesses and vice versa. This is also known as
|
|
* a "fence."
|
|
*
|
|
* Note: "mb()" and its variants cannot be used as a fence to order
|
|
* accesses to memory mapped I/O registers. For that, mf.a needs to
|
|
* be used. However, we don't want to always use mf.a because (a)
|
|
* it's (presumably) much slower than mf and (b) mf.a is supported for
|
|
* sequential memory pages only.
|
|
*/
|
|
#define mb() ia64_mf()
|
|
#define rmb() mb()
|
|
#define wmb() mb()
|
|
#define read_barrier_depends() do { } while(0)
|
|
|
|
#ifdef CONFIG_SMP
|
|
# define smp_mb() mb()
|
|
# define smp_rmb() rmb()
|
|
# define smp_wmb() wmb()
|
|
# define smp_read_barrier_depends() read_barrier_depends()
|
|
#else
|
|
# define smp_mb() barrier()
|
|
# define smp_rmb() barrier()
|
|
# define smp_wmb() barrier()
|
|
# define smp_read_barrier_depends() do { } while(0)
|
|
#endif
|
|
|
|
/*
|
|
* XXX check on these---I suspect what Linus really wants here is
|
|
* acquire vs release semantics but we can't discuss this stuff with
|
|
* Linus just yet. Grrr...
|
|
*/
|
|
#define set_mb(var, value) do { (var) = (value); mb(); } while (0)
|
|
#define set_wmb(var, value) do { (var) = (value); mb(); } while (0)
|
|
|
|
#define safe_halt() ia64_pal_halt_light() /* PAL_HALT_LIGHT */
|
|
|
|
/*
|
|
* The group barrier in front of the rsm & ssm are necessary to ensure
|
|
* that none of the previous instructions in the same group are
|
|
* affected by the rsm/ssm.
|
|
*/
|
|
/* For spinlocks etc */
|
|
|
|
/*
|
|
* - clearing psr.i is implicitly serialized (visible by next insn)
|
|
* - setting psr.i requires data serialization
|
|
* - we need a stop-bit before reading PSR because we sometimes
|
|
* write a floating-point register right before reading the PSR
|
|
* and that writes to PSR.mfl
|
|
*/
|
|
#define __local_irq_save(x) \
|
|
do { \
|
|
ia64_stop(); \
|
|
(x) = ia64_getreg(_IA64_REG_PSR); \
|
|
ia64_stop(); \
|
|
ia64_rsm(IA64_PSR_I); \
|
|
} while (0)
|
|
|
|
#define __local_irq_disable() \
|
|
do { \
|
|
ia64_stop(); \
|
|
ia64_rsm(IA64_PSR_I); \
|
|
} while (0)
|
|
|
|
#define __local_irq_restore(x) ia64_intrin_local_irq_restore((x) & IA64_PSR_I)
|
|
|
|
#ifdef CONFIG_IA64_DEBUG_IRQ
|
|
|
|
extern unsigned long last_cli_ip;
|
|
|
|
# define __save_ip() last_cli_ip = ia64_getreg(_IA64_REG_IP)
|
|
|
|
# define local_irq_save(x) \
|
|
do { \
|
|
unsigned long psr; \
|
|
\
|
|
__local_irq_save(psr); \
|
|
if (psr & IA64_PSR_I) \
|
|
__save_ip(); \
|
|
(x) = psr; \
|
|
} while (0)
|
|
|
|
# define local_irq_disable() do { unsigned long x; local_irq_save(x); } while (0)
|
|
|
|
# define local_irq_restore(x) \
|
|
do { \
|
|
unsigned long old_psr, psr = (x); \
|
|
\
|
|
local_save_flags(old_psr); \
|
|
__local_irq_restore(psr); \
|
|
if ((old_psr & IA64_PSR_I) && !(psr & IA64_PSR_I)) \
|
|
__save_ip(); \
|
|
} while (0)
|
|
|
|
#else /* !CONFIG_IA64_DEBUG_IRQ */
|
|
# define local_irq_save(x) __local_irq_save(x)
|
|
# define local_irq_disable() __local_irq_disable()
|
|
# define local_irq_restore(x) __local_irq_restore(x)
|
|
#endif /* !CONFIG_IA64_DEBUG_IRQ */
|
|
|
|
#define local_irq_enable() ({ ia64_stop(); ia64_ssm(IA64_PSR_I); ia64_srlz_d(); })
|
|
#define local_save_flags(flags) ({ ia64_stop(); (flags) = ia64_getreg(_IA64_REG_PSR); })
|
|
|
|
#define irqs_disabled() \
|
|
({ \
|
|
unsigned long __ia64_id_flags; \
|
|
local_save_flags(__ia64_id_flags); \
|
|
(__ia64_id_flags & IA64_PSR_I) == 0; \
|
|
})
|
|
|
|
#ifdef __KERNEL__
|
|
|
|
#ifdef CONFIG_IA32_SUPPORT
|
|
# define IS_IA32_PROCESS(regs) (ia64_psr(regs)->is != 0)
|
|
#else
|
|
# define IS_IA32_PROCESS(regs) 0
|
|
struct task_struct;
|
|
static inline void ia32_save_state(struct task_struct *t __attribute__((unused))){}
|
|
static inline void ia32_load_state(struct task_struct *t __attribute__((unused))){}
|
|
#endif
|
|
|
|
/*
|
|
* Context switch from one thread to another. If the two threads have
|
|
* different address spaces, schedule() has already taken care of
|
|
* switching to the new address space by calling switch_mm().
|
|
*
|
|
* Disabling access to the fph partition and the debug-register
|
|
* context switch MUST be done before calling ia64_switch_to() since a
|
|
* newly created thread returns directly to
|
|
* ia64_ret_from_syscall_clear_r8.
|
|
*/
|
|
extern struct task_struct *ia64_switch_to (void *next_task);
|
|
|
|
struct task_struct;
|
|
|
|
extern void ia64_save_extra (struct task_struct *task);
|
|
extern void ia64_load_extra (struct task_struct *task);
|
|
|
|
#ifdef CONFIG_PERFMON
|
|
DECLARE_PER_CPU(unsigned long, pfm_syst_info);
|
|
# define PERFMON_IS_SYSWIDE() (__get_cpu_var(pfm_syst_info) & 0x1)
|
|
#else
|
|
# define PERFMON_IS_SYSWIDE() (0)
|
|
#endif
|
|
|
|
#define IA64_HAS_EXTRA_STATE(t) \
|
|
((t)->thread.flags & (IA64_THREAD_DBG_VALID|IA64_THREAD_PM_VALID) \
|
|
|| IS_IA32_PROCESS(ia64_task_regs(t)) || PERFMON_IS_SYSWIDE())
|
|
|
|
#define __switch_to(prev,next,last) do { \
|
|
if (IA64_HAS_EXTRA_STATE(prev)) \
|
|
ia64_save_extra(prev); \
|
|
if (IA64_HAS_EXTRA_STATE(next)) \
|
|
ia64_load_extra(next); \
|
|
ia64_psr(ia64_task_regs(next))->dfh = !ia64_is_local_fpu_owner(next); \
|
|
(last) = ia64_switch_to((next)); \
|
|
} while (0)
|
|
|
|
#ifdef CONFIG_SMP
|
|
/*
|
|
* In the SMP case, we save the fph state when context-switching away from a thread that
|
|
* modified fph. This way, when the thread gets scheduled on another CPU, the CPU can
|
|
* pick up the state from task->thread.fph, avoiding the complication of having to fetch
|
|
* the latest fph state from another CPU. In other words: eager save, lazy restore.
|
|
*/
|
|
# define switch_to(prev,next,last) do { \
|
|
if (ia64_psr(ia64_task_regs(prev))->mfh && ia64_is_local_fpu_owner(prev)) { \
|
|
ia64_psr(ia64_task_regs(prev))->mfh = 0; \
|
|
(prev)->thread.flags |= IA64_THREAD_FPH_VALID; \
|
|
__ia64_save_fpu((prev)->thread.fph); \
|
|
} \
|
|
__switch_to(prev, next, last); \
|
|
} while (0)
|
|
#else
|
|
# define switch_to(prev,next,last) __switch_to(prev, next, last)
|
|
#endif
|
|
|
|
/*
|
|
* On IA-64, we don't want to hold the runqueue's lock during the low-level context-switch,
|
|
* because that could cause a deadlock. Here is an example by Erich Focht:
|
|
*
|
|
* Example:
|
|
* CPU#0:
|
|
* schedule()
|
|
* -> spin_lock_irq(&rq->lock)
|
|
* -> context_switch()
|
|
* -> wrap_mmu_context()
|
|
* -> read_lock(&tasklist_lock)
|
|
*
|
|
* CPU#1:
|
|
* sys_wait4() or release_task() or forget_original_parent()
|
|
* -> write_lock(&tasklist_lock)
|
|
* -> do_notify_parent()
|
|
* -> wake_up_parent()
|
|
* -> try_to_wake_up()
|
|
* -> spin_lock_irq(&parent_rq->lock)
|
|
*
|
|
* If the parent's rq happens to be on CPU#0, we'll wait for the rq->lock
|
|
* of that CPU which will not be released, because there we wait for the
|
|
* tasklist_lock to become available.
|
|
*/
|
|
#define __ARCH_WANT_UNLOCKED_CTXSW
|
|
|
|
#define ARCH_HAS_PREFETCH_SWITCH_STACK
|
|
#define ia64_platform_is(x) (strcmp(x, platform_name) == 0)
|
|
|
|
void cpu_idle_wait(void);
|
|
|
|
#define arch_align_stack(x) (x)
|
|
|
|
#endif /* __KERNEL__ */
|
|
|
|
#endif /* __ASSEMBLY__ */
|
|
|
|
#endif /* _ASM_IA64_SYSTEM_H */
|