processor.h

xen虚拟机源代码安装包

#ifndef _ASM_IA64_PROCESSOR_H
#define _ASM_IA64_PROCESSOR_H

/*
 * Copyright (C) 1998-2004 Hewlett-Packard Co
 *	David Mosberger-Tang <[e-mail]>
 *	Stephane Eranian <[e-mail]>
 * Copyright (C) 1999 Asit Mallick <[e-mail]>
 * Copyright (C) 1999 Don Dugger <[e-mail]>
 *
 * 11/24/98	S.Eranian	added ia64_set_iva()
 * 12/03/99	D. Mosberger	implement thread_saved_pc() via kernel unwind API
 * 06/16/00	A. Mallick	added csd/ssd/tssd for ia32 support
 */

#include <linux/config.h>

#include <asm/intrinsics.h>
#include <asm/kregs.h>
#include <asm/ptrace.h>
#include <asm/ustack.h>

/* Our arch specific arch_init_sched_domain is in arch/ia64/kernel/domain.c */
#define ARCH_HAS_SCHED_DOMAIN

#define IA64_NUM_DBG_REGS	8
/*
 * Limits for PMC and PMD are set to less than maximum architected values
 * but should be sufficient for a while
 */
#ifdef XEN
/*
 * These are increased in linux-2.6.16. Montecito requires 35PMDs.
 * This ifdef will become unnecessary when this header file is
 * upgraded to 2.6.16 or newer.
 */
#define IA64_NUM_PMC_REGS	64
#define IA64_NUM_PMD_REGS	64
#else
#define IA64_NUM_PMC_REGS	32
#define IA64_NUM_PMD_REGS	32
#endif

#define DEFAULT_MAP_BASE	__IA64_UL_CONST(0x2000000000000000)
#define DEFAULT_TASK_SIZE	__IA64_UL_CONST(0xa000000000000000)

/*
 * TASK_SIZE really is a mis-named.  It really is the maximum user
 * space address (plus one).  On IA-64, there are five regions of 2TB
 * each (assuming 8KB page size), for a total of 8TB of user virtual
 * address space.
 */
#define TASK_SIZE		(current->thread.task_size)

/*
 * This decides where the kernel will search for a free chunk of vm
 * space during mmap's.
 */
#define TASK_UNMAPPED_BASE	(current->thread.map_base)

#define IA64_THREAD_FPH_VALID	(__IA64_UL(1) << 0)	/* floating-point high state valid? */
#define IA64_THREAD_DBG_VALID	(__IA64_UL(1) << 1)	/* debug registers valid? */
#define IA64_THREAD_PM_VALID	(__IA64_UL(1) << 2)	/* performance registers valid? */
#define IA64_THREAD_UAC_NOPRINT	(__IA64_UL(1) << 3)	/* don't log unaligned accesses */
#define IA64_THREAD_UAC_SIGBUS	(__IA64_UL(1) << 4)	/* generate SIGBUS on unaligned acc. */
							/* bit 5 is currently unused */
#define IA64_THREAD_FPEMU_NOPRINT (__IA64_UL(1) << 6)	/* don't log any fpswa faults */
#define IA64_THREAD_FPEMU_SIGFPE  (__IA64_UL(1) << 7)	/* send a SIGFPE for fpswa faults */

#define IA64_THREAD_UAC_SHIFT	3
#define IA64_THREAD_UAC_MASK	(IA64_THREAD_UAC_NOPRINT | IA64_THREAD_UAC_SIGBUS)
#define IA64_THREAD_FPEMU_SHIFT	6
#define IA64_THREAD_FPEMU_MASK	(IA64_THREAD_FPEMU_NOPRINT | IA64_THREAD_FPEMU_SIGFPE)


/*
 * This shift should be large enough to be able to represent 1000000000/itc_freq with good
 * accuracy while being small enough to fit 10*1000000000<<IA64_NSEC_PER_CYC_SHIFT in 64 bits
 * (this will give enough slack to represent 10 seconds worth of time as a scaled number).
 */
#define IA64_NSEC_PER_CYC_SHIFT	30

#ifndef __ASSEMBLY__

#include <linux/cache.h>
#include <linux/compiler.h>
#include <linux/threads.h>
#include <linux/types.h>

#include <asm/fpu.h>
#include <asm/page.h>
#include <asm/percpu.h>
#include <asm/rse.h>
#include <asm/unwind.h>
#include <asm/atomic.h>
#ifdef CONFIG_NUMA
#include <asm/nodedata.h>
#endif

#ifdef XEN
#include <asm/xenprocessor.h>
#include <xen/bitops.h>
#else
/* like above but expressed as bitfields for more efficient access: */
struct ia64_psr {
	__u64 reserved0 : 1;
	__u64 be : 1;
	__u64 up : 1;
	__u64 ac : 1;
	__u64 mfl : 1;
	__u64 mfh : 1;
	__u64 reserved1 : 7;
	__u64 ic : 1;
	__u64 i : 1;
	__u64 pk : 1;
	__u64 reserved2 : 1;
	__u64 dt : 1;
	__u64 dfl : 1;
	__u64 dfh : 1;
	__u64 sp : 1;
	__u64 pp : 1;
	__u64 di : 1;
	__u64 si : 1;
	__u64 db : 1;
	__u64 lp : 1;
	__u64 tb : 1;
	__u64 rt : 1;
	__u64 reserved3 : 4;
	__u64 cpl : 2;
	__u64 is : 1;
	__u64 mc : 1;
	__u64 it : 1;
	__u64 id : 1;
	__u64 da : 1;
	__u64 dd : 1;
	__u64 ss : 1;
	__u64 ri : 2;
	__u64 ed : 1;
	__u64 bn : 1;
	__u64 reserved4 : 19;
};
#endif

/*
 * CPU type, hardware bug flags, and per-CPU state.  Frequently used
 * state comes earlier:
 */
struct cpuinfo_ia64 {
	__u32 softirq_pending;
	__u64 itm_delta;	/* # of clock cycles between clock ticks */
	__u64 itm_next;		/* interval timer mask value to use for next clock tick */
	__u64 nsec_per_cyc;	/* (1000000000<<IA64_NSEC_PER_CYC_SHIFT)/itc_freq */
	__u64 unimpl_va_mask;	/* mask of unimplemented virtual address bits (from PAL) */
	__u64 unimpl_pa_mask;	/* mask of unimplemented physical address bits (from PAL) */
	__u64 itc_freq;		/* frequency of ITC counter */
	__u64 proc_freq;	/* frequency of processor */
	__u64 cyc_per_usec;	/* itc_freq/1000000 */
	__u64 ptce_base;
	__u32 ptce_count[2];
	__u32 ptce_stride[2];
	struct task_struct *ksoftirqd;	/* kernel softirq daemon for this CPU */

#ifdef CONFIG_SMP
	__u64 loops_per_jiffy;
	int cpu;
	__u32 socket_id;	/* physical processor socket id */
	__u16 core_id;		/* core id */
	__u16 thread_id;	/* thread id */
	__u16 num_log;		/* Total number of logical processors on
				 * this socket that were successfully booted */
	__u8  cores_per_socket;	/* Cores per processor socket */
	__u8  threads_per_core;	/* Threads per core */
#endif

	/* CPUID-derived information: */
	__u64 ppn;
	__u64 features;
	__u8 number;
	__u8 revision;
	__u8 model;
	__u8 family;
	__u8 archrev;
	char vendor[16];

#ifdef CONFIG_NUMA
	struct ia64_node_data *node_data;
#endif
};

DECLARE_PER_CPU(struct cpuinfo_ia64, cpu_info);

/*
 * The "local" data variable.  It refers to the per-CPU data of the currently executing
 * CPU, much like "current" points to the per-task data of the currently executing task.
 * Do not use the address of local_cpu_data, since it will be different from
 * cpu_data(smp_processor_id())!
 */
#define local_cpu_data		(&__ia64_per_cpu_var(cpu_info))
#define cpu_data(cpu)		(&per_cpu(cpu_info, cpu))

extern void identify_cpu (struct cpuinfo_ia64 *);
extern void print_cpu_info (struct cpuinfo_ia64 *);

typedef struct {
	unsigned long seg;
} mm_segment_t;

#define SET_UNALIGN_CTL(task,value)								\
({												\
	(task)->thread.flags = (((task)->thread.flags & ~IA64_THREAD_UAC_MASK)			\
				| (((value) << IA64_THREAD_UAC_SHIFT) & IA64_THREAD_UAC_MASK));	\
	0;											\
})
#define GET_UNALIGN_CTL(task,addr)								\
({												\
	put_user(((task)->thread.flags & IA64_THREAD_UAC_MASK) >> IA64_THREAD_UAC_SHIFT,	\
		 (int __user *) (addr));							\
})

#define SET_FPEMU_CTL(task,value)								\
({												\
	(task)->thread.flags = (((task)->thread.flags & ~IA64_THREAD_FPEMU_MASK)		\
			  | (((value) << IA64_THREAD_FPEMU_SHIFT) & IA64_THREAD_FPEMU_MASK));	\
	0;											\
})
#define GET_FPEMU_CTL(task,addr)								\
({												\
	put_user(((task)->thread.flags & IA64_THREAD_FPEMU_MASK) >> IA64_THREAD_FPEMU_SHIFT,	\
		 (int __user *) (addr));							\
})

#ifdef CONFIG_IA32_SUPPORT
struct desc_struct {
	unsigned int a, b;
};

#define desc_empty(desc)		(!((desc)->a + (desc)->b))
#define desc_equal(desc1, desc2)	(((desc1)->a == (desc2)->a) && ((desc1)->b == (desc2)->b))

#define GDT_ENTRY_TLS_ENTRIES	3
#define GDT_ENTRY_TLS_MIN	6
#define GDT_ENTRY_TLS_MAX 	(GDT_ENTRY_TLS_MIN + GDT_ENTRY_TLS_ENTRIES - 1)

#define TLS_SIZE (GDT_ENTRY_TLS_ENTRIES * 8)

struct partial_page_list;
#endif

struct thread_struct {
	__u32 flags;			/* various thread flags (see IA64_THREAD_*) */
	/* writing on_ustack is performance-critical, so it's worth spending 8 bits on it... */
	__u8 on_ustack;			/* executing on user-stacks? */
	__u8 pad[3];
	__u64 ksp;			/* kernel stack pointer */
	__u64 map_base;			/* base address for get_unmapped_area() */
	__u64 task_size;		/* limit for task size */
	__u64 rbs_bot;			/* the base address for the RBS */
	int last_fph_cpu;		/* CPU that may hold the contents of f32-f127 */

#ifdef CONFIG_IA32_SUPPORT
	__u64 eflag;			/* IA32 EFLAGS reg */
	__u64 fsr;			/* IA32 floating pt status reg */
	__u64 fcr;			/* IA32 floating pt control reg */
	__u64 fir;			/* IA32 fp except. instr. reg */
	__u64 fdr;			/* IA32 fp except. data reg */
	__u64 old_k1;			/* old value of ar.k1 */
	__u64 old_iob;			/* old IOBase value */
	struct partial_page_list *ppl;	/* partial page list for 4K page size issue */
        /* cached TLS descriptors. */
	struct desc_struct tls_array[GDT_ENTRY_TLS_ENTRIES];

# define INIT_THREAD_IA32	.eflag =	0,			\
				.fsr =		0,			\
				.fcr =		0x17800000037fULL,	\
				.fir =		0,			\
				.fdr =		0,			\
				.old_k1 =	0,			\
				.old_iob =	0,			\
				.ppl =		NULL,
#else
# define INIT_THREAD_IA32
#endif /* CONFIG_IA32_SUPPORT */
#ifdef CONFIG_PERFMON
	__u64 pmcs[IA64_NUM_PMC_REGS];
	__u64 pmds[IA64_NUM_PMD_REGS];
	void *pfm_context;		     /* pointer to detailed PMU context */
	unsigned long pfm_needs_checking;    /* when >0, pending perfmon work on kernel exit */
# define INIT_THREAD_PM		.pmcs =			{0UL, },  \
				.pmds =			{0UL, },  \
				.pfm_context =		NULL,     \
				.pfm_needs_checking =	0UL,
#else
# define INIT_THREAD_PM
#endif
#ifndef XEN
	__u64 dbr[IA64_NUM_DBG_REGS];
	__u64 ibr[IA64_NUM_DBG_REGS];
#endif
	struct ia64_fpreg fph[96];	/* saved/loaded on demand */
};

#ifndef XEN
#define INIT_THREAD {						\
	.flags =	0,					\
	.on_ustack =	0,					\
	.ksp =		0,					\
	.map_base =	DEFAULT_MAP_BASE,			\
	.rbs_bot =	STACK_TOP - DEFAULT_USER_STACK_SIZE,	\
	.task_size =	DEFAULT_TASK_SIZE,			\
	.last_fph_cpu =  -1,					\
	INIT_THREAD_IA32					\
	INIT_THREAD_PM						\
	.dbr =		{0, },					\
	.ibr =		{0, },					\
	.fph =		{{{{0}}}, }				\
}

#define start_thread(regs,new_ip,new_sp) do {							\
	set_fs(USER_DS);									\
	regs->cr_ipsr = ((regs->cr_ipsr | (IA64_PSR_BITS_TO_SET | IA64_PSR_CPL))		\
			 & ~(IA64_PSR_BITS_TO_CLEAR | IA64_PSR_RI | IA64_PSR_IS));		\
	regs->cr_iip = new_ip;									\
	regs->ar_rsc = 0xf;		/* eager mode, privilege level 3 */			\
	regs->ar_rnat = 0;									\
	regs->ar_bspstore = current->thread.rbs_bot;						\
	regs->ar_fpsr = FPSR_DEFAULT;								\
	regs->loadrs = 0;									\
	regs->r8 = current->mm->dumpable;	/* set "don't zap registers" flag */		\
	regs->r12 = new_sp - 16;	/* allocate 16 byte scratch area */			\
	if (unlikely(!current->mm->dumpable)) {							\
		/*										\
		 * Zap scratch regs to avoid leaking bits between processes with different	\
		 * uid/privileges.								\
		 */										\
		regs->ar_pfs = 0; regs->b0 = 0; regs->pr = 0;					\
		regs->r1 = 0; regs->r9  = 0; regs->r11 = 0; regs->r13 = 0; regs->r15 = 0;	\
	}											\
} while (0)
#endif

/* Forward declarations, a strange C thing... */
struct mm_struct;
struct task_struct;

/*
 * Free all resources held by a thread. This is called after the
 * parent of DEAD_TASK has collected the exit status of the task via
 * wait().
 */
#define release_thread(dead_task)

/* Prepare to copy thread state - unlazy all lazy status */
#define prepare_to_copy(tsk)	do { } while (0)

/*
 * This is the mechanism for creating a new kernel thread.
 *
 * NOTE 1: Only a kernel-only process (ie the swapper or direct
 * descendants who haven't done an "execve()") should use this: it
 * will work within a system call from a "real" process, but the
 * process memory space will not be free'd until both the parent and
 * the child have exited.
 *
 * NOTE 2: This MUST NOT be an inlined function.  Otherwise, we get
 * into trouble in init/main.c when the child thread returns to
 * do_basic_setup() and the timing is such that free_initmem() has
 * been called already.
 */
extern pid_t kernel_thread (int (*fn)(void *), void *arg, unsigned long flags);

/* Get wait channel for task P.  */
extern unsigned long get_wchan (struct task_struct *p);

/* Return instruction pointer of blocked task TSK.  */
#define KSTK_EIP(tsk)					\
  ({							\
	struct pt_regs *_regs = ia64_task_regs(tsk);	\
	_regs->cr_iip + ia64_psr(_regs)->ri;		\
  })

/* Return stack pointer of blocked task TSK.  */
#define KSTK_ESP(tsk)  ((tsk)->thread.ksp)

extern void ia64_getreg_unknown_kr (void);
extern void ia64_setreg_unknown_kr (void);

#define ia64_get_kr(regnum)					\
({								\
	unsigned long r = 0;					\