perf_event.h 22.5 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
/*
 * Performance events x86 architecture header
 *
 *  Copyright (C) 2008 Thomas Gleixner <tglx@linutronix.de>
 *  Copyright (C) 2008-2009 Red Hat, Inc., Ingo Molnar
 *  Copyright (C) 2009 Jaswinder Singh Rajput
 *  Copyright (C) 2009 Advanced Micro Devices, Inc., Robert Richter
 *  Copyright (C) 2008-2009 Red Hat, Inc., Peter Zijlstra <pzijlstr@redhat.com>
 *  Copyright (C) 2009 Intel Corporation, <markus.t.metzger@intel.com>
 *  Copyright (C) 2009 Google, Inc., Stephane Eranian
 *
 *  For licencing details see kernel-base/COPYING
 */

#include <linux/perf_event.h>

17 18 19 20 21 22 23 24 25 26 27 28
#if 0
#undef wrmsrl
#define wrmsrl(msr, val) 						\
do {									\
	unsigned int _msr = (msr);					\
	u64 _val = (val);						\
	trace_printk("wrmsrl(%x, %Lx)\n", (unsigned int)(_msr),		\
			(unsigned long long)(_val));			\
	native_write_msr((_msr), (u32)(_val), (u32)(_val >> 32));	\
} while (0)
#endif

29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47
/*
 *          |   NHM/WSM    |      SNB     |
 * register -------------------------------
 *          |  HT  | no HT |  HT  | no HT |
 *-----------------------------------------
 * offcore  | core | core  | cpu  | core  |
 * lbr_sel  | core | core  | cpu  | core  |
 * ld_lat   | cpu  | core  | cpu  | core  |
 *-----------------------------------------
 *
 * Given that there is a small number of shared regs,
 * we can pre-allocate their slot in the per-cpu
 * per-core reg tables.
 */
enum extra_reg_type {
	EXTRA_REG_NONE  = -1,	/* not used */

	EXTRA_REG_RSP_0 = 0,	/* offcore_response_0 */
	EXTRA_REG_RSP_1 = 1,	/* offcore_response_1 */
48
	EXTRA_REG_LBR   = 2,	/* lbr_select */
49
	EXTRA_REG_LDLAT = 3,	/* ld_lat_threshold */
50 51 52 53 54 55 56 57 58 59 60 61

	EXTRA_REG_MAX		/* number of entries needed */
};

struct event_constraint {
	union {
		unsigned long	idxmsk[BITS_TO_LONGS(X86_PMC_IDX_MAX)];
		u64		idxmsk64;
	};
	u64	code;
	u64	cmask;
	int	weight;
62
	int	overlap;
63
	int	flags;
64
};
65
/*
66
 * struct hw_perf_event.flags flags
67 68
 */
#define PERF_X86_EVENT_PEBS_LDLAT	0x1 /* ld+ldlat data address sampling */
69
#define PERF_X86_EVENT_PEBS_ST		0x2 /* st data address sampling */
70
#define PERF_X86_EVENT_PEBS_ST_HSW	0x4 /* haswell style datala, store */
71
#define PERF_X86_EVENT_COMMITTED	0x8 /* event passed commit_txn */
72 73
#define PERF_X86_EVENT_PEBS_LD_HSW	0x10 /* haswell style datala, load */
#define PERF_X86_EVENT_PEBS_NA_HSW	0x20 /* haswell style datala, unknown */
74 75
#define PERF_X86_EVENT_RDPMC_ALLOWED	0x40 /* grant rdpmc permission */

76 77 78 79 80 81 82 83 84

struct amd_nb {
	int nb_id;  /* NorthBridge id */
	int refcnt; /* reference count */
	struct perf_event *owners[X86_PMC_IDX_MAX];
	struct event_constraint event_constraints[X86_PMC_IDX_MAX];
};

/* The maximal number of PEBS events: */
85
#define MAX_PEBS_EVENTS		8
86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127

/*
 * A debug store configuration.
 *
 * We only support architectures that use 64bit fields.
 */
struct debug_store {
	u64	bts_buffer_base;
	u64	bts_index;
	u64	bts_absolute_maximum;
	u64	bts_interrupt_threshold;
	u64	pebs_buffer_base;
	u64	pebs_index;
	u64	pebs_absolute_maximum;
	u64	pebs_interrupt_threshold;
	u64	pebs_event_reset[MAX_PEBS_EVENTS];
};

/*
 * Per register state.
 */
struct er_account {
	raw_spinlock_t		lock;	/* per-core: protect structure */
	u64                 config;	/* extra MSR config */
	u64                 reg;	/* extra MSR number */
	atomic_t            ref;	/* reference count */
};

/*
 * Per core/cpu state
 *
 * Used to coordinate shared registers between HT threads or
 * among events on a single PMU.
 */
struct intel_shared_regs {
	struct er_account       regs[EXTRA_REG_MAX];
	int                     refcnt;		/* per-core: #HT threads */
	unsigned                core_id;	/* per-core: core id */
};

#define MAX_LBR_ENTRIES		16

128 129 130 131 132 133
enum {
	X86_PERF_KFREE_SHARED = 0,
	X86_PERF_KFREE_EXCL   = 1,
	X86_PERF_KFREE_MAX
};

134 135 136 137 138 139 140 141 142
struct cpu_hw_events {
	/*
	 * Generic x86 PMC bits
	 */
	struct perf_event	*events[X86_PMC_IDX_MAX]; /* in counter order */
	unsigned long		active_mask[BITS_TO_LONGS(X86_PMC_IDX_MAX)];
	unsigned long		running[BITS_TO_LONGS(X86_PMC_IDX_MAX)];
	int			enabled;

143 144 145 146 147
	int			n_events; /* the # of events in the below arrays */
	int			n_added;  /* the # last events in the below arrays;
					     they've never been enabled yet */
	int			n_txn;    /* the # last events in the below arrays;
					     added in the current transaction */
148 149 150 151 152
	int			assign[X86_PMC_IDX_MAX]; /* event to counter assignment */
	u64			tags[X86_PMC_IDX_MAX];
	struct perf_event	*event_list[X86_PMC_IDX_MAX]; /* in enabled order */

	unsigned int		group_flag;
153
	int			is_fake;
154 155 156 157 158 159 160 161 162 163 164 165 166 167

	/*
	 * Intel DebugStore bits
	 */
	struct debug_store	*ds;
	u64			pebs_enabled;

	/*
	 * Intel LBR bits
	 */
	int				lbr_users;
	void				*lbr_context;
	struct perf_branch_stack	lbr_stack;
	struct perf_branch_entry	lbr_entries[MAX_LBR_ENTRIES];
168
	struct er_account		*lbr_sel;
169
	u64				br_sel;
170

171 172 173 174 175 176 177
	/*
	 * Intel host/guest exclude bits
	 */
	u64				intel_ctrl_guest_mask;
	u64				intel_ctrl_host_mask;
	struct perf_guest_switch_msr	guest_switch_msrs[X86_PMC_IDX_MAX];

178 179 180 181 182
	/*
	 * Intel checkpoint mask
	 */
	u64				intel_cp_status;

183 184 185 186 187 188 189 190 191
	/*
	 * manage shared (per-core, per-cpu) registers
	 * used on Intel NHM/WSM/SNB
	 */
	struct intel_shared_regs	*shared_regs;

	/*
	 * AMD specific bits
	 */
192 193 194
	struct amd_nb			*amd_nb;
	/* Inverted mask of bits to clear in the perf_ctr ctrl registers */
	u64				perf_ctr_virt_mask;
195

196
	void				*kfree_on_online[X86_PERF_KFREE_MAX];
197 198
};

199
#define __EVENT_CONSTRAINT(c, n, m, w, o, f) {\
200 201 202 203
	{ .idxmsk64 = (n) },		\
	.code = (c),			\
	.cmask = (m),			\
	.weight = (w),			\
204
	.overlap = (o),			\
205
	.flags = f,			\
206 207 208
}

#define EVENT_CONSTRAINT(c, n, m)	\
209
	__EVENT_CONSTRAINT(c, n, m, HWEIGHT(n), 0, 0)
210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232

/*
 * The overlap flag marks event constraints with overlapping counter
 * masks. This is the case if the counter mask of such an event is not
 * a subset of any other counter mask of a constraint with an equal or
 * higher weight, e.g.:
 *
 *  c_overlaps = EVENT_CONSTRAINT_OVERLAP(0, 0x09, 0);
 *  c_another1 = EVENT_CONSTRAINT(0, 0x07, 0);
 *  c_another2 = EVENT_CONSTRAINT(0, 0x38, 0);
 *
 * The event scheduler may not select the correct counter in the first
 * cycle because it needs to know which subsequent events will be
 * scheduled. It may fail to schedule the events then. So we set the
 * overlap flag for such constraints to give the scheduler a hint which
 * events to select for counter rescheduling.
 *
 * Care must be taken as the rescheduling algorithm is O(n!) which
 * will increase scheduling cycles for an over-commited system
 * dramatically.  The number of such EVENT_CONSTRAINT_OVERLAP() macros
 * and its counter masks must be kept at a minimum.
 */
#define EVENT_CONSTRAINT_OVERLAP(c, n, m)	\
233
	__EVENT_CONSTRAINT(c, n, m, HWEIGHT(n), 1, 0)
234 235 236 237 238 239 240 241 242 243 244 245 246 247 248

/*
 * Constraint on the Event code.
 */
#define INTEL_EVENT_CONSTRAINT(c, n)	\
	EVENT_CONSTRAINT(c, n, ARCH_PERFMON_EVENTSEL_EVENT)

/*
 * Constraint on the Event code + UMask + fixed-mask
 *
 * filter mask to validate fixed counter events.
 * the following filters disqualify for fixed counters:
 *  - inv
 *  - edge
 *  - cnt-mask
249 250
 *  - in_tx
 *  - in_tx_checkpointed
251 252 253
 *  The other filters are supported by fixed counters.
 *  The any-thread option is supported starting with v3.
 */
254
#define FIXED_EVENT_FLAGS (X86_RAW_EVENT_MASK|HSW_IN_TX|HSW_IN_TX_CHECKPOINTED)
255
#define FIXED_EVENT_CONSTRAINT(c, n)	\
256
	EVENT_CONSTRAINT(c, (1ULL << (32+n)), FIXED_EVENT_FLAGS)
257 258 259 260 261 262 263

/*
 * Constraint on the Event code + UMask
 */
#define INTEL_UEVENT_CONSTRAINT(c, n)	\
	EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVENT_MASK)

264 265 266 267
/* Like UEVENT_CONSTRAINT, but match flags too */
#define INTEL_FLAGS_UEVENT_CONSTRAINT(c, n)	\
	EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVENT_MASK|X86_ALL_EVENT_FLAGS)

268
#define INTEL_PLD_CONSTRAINT(c, n)	\
269
	__EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVENT_MASK|X86_ALL_EVENT_FLAGS, \
270 271
			   HWEIGHT(n), 0, PERF_X86_EVENT_PEBS_LDLAT)

272
#define INTEL_PST_CONSTRAINT(c, n)	\
273
	__EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVENT_MASK|X86_ALL_EVENT_FLAGS, \
274 275
			  HWEIGHT(n), 0, PERF_X86_EVENT_PEBS_ST)

276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299
/* Event constraint, but match on all event flags too. */
#define INTEL_FLAGS_EVENT_CONSTRAINT(c, n) \
	EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVENT_MASK|X86_ALL_EVENT_FLAGS)

/* Check only flags, but allow all event/umask */
#define INTEL_ALL_EVENT_CONSTRAINT(code, n)	\
	EVENT_CONSTRAINT(code, n, X86_ALL_EVENT_FLAGS)

/* Check flags and event code, and set the HSW store flag */
#define INTEL_FLAGS_EVENT_CONSTRAINT_DATALA_ST(code, n) \
	__EVENT_CONSTRAINT(code, n, 			\
			  ARCH_PERFMON_EVENTSEL_EVENT|X86_ALL_EVENT_FLAGS, \
			  HWEIGHT(n), 0, PERF_X86_EVENT_PEBS_ST_HSW)

/* Check flags and event code, and set the HSW load flag */
#define INTEL_FLAGS_EVENT_CONSTRAINT_DATALA_LD(code, n) \
	__EVENT_CONSTRAINT(code, n, 			\
			  ARCH_PERFMON_EVENTSEL_EVENT|X86_ALL_EVENT_FLAGS, \
			  HWEIGHT(n), 0, PERF_X86_EVENT_PEBS_LD_HSW)

/* Check flags and event code/umask, and set the HSW store flag */
#define INTEL_FLAGS_UEVENT_CONSTRAINT_DATALA_ST(code, n) \
	__EVENT_CONSTRAINT(code, n, 			\
			  INTEL_ARCH_EVENT_MASK|X86_ALL_EVENT_FLAGS, \
300 301
			  HWEIGHT(n), 0, PERF_X86_EVENT_PEBS_ST_HSW)

302 303 304 305 306 307 308 309 310 311 312 313 314
/* Check flags and event code/umask, and set the HSW load flag */
#define INTEL_FLAGS_UEVENT_CONSTRAINT_DATALA_LD(code, n) \
	__EVENT_CONSTRAINT(code, n, 			\
			  INTEL_ARCH_EVENT_MASK|X86_ALL_EVENT_FLAGS, \
			  HWEIGHT(n), 0, PERF_X86_EVENT_PEBS_LD_HSW)

/* Check flags and event code/umask, and set the HSW N/A flag */
#define INTEL_FLAGS_UEVENT_CONSTRAINT_DATALA_NA(code, n) \
	__EVENT_CONSTRAINT(code, n, 			\
			  INTEL_ARCH_EVENT_MASK|INTEL_ARCH_EVENT_MASK, \
			  HWEIGHT(n), 0, PERF_X86_EVENT_PEBS_NA_HSW)


315 316 317 318 319 320 321 322
/*
 * We define the end marker as having a weight of -1
 * to enable blacklisting of events using a counter bitmask
 * of zero and thus a weight of zero.
 * The end marker has a weight that cannot possibly be
 * obtained from counting the bits in the bitmask.
 */
#define EVENT_CONSTRAINT_END { .weight = -1 }
323

324 325 326
/*
 * Check for end marker with weight == -1
 */
327
#define for_each_event_constraint(e, c)	\
328
	for ((e) = (c); (e)->weight != -1; (e)++)
329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345

/*
 * Extra registers for specific events.
 *
 * Some events need large masks and require external MSRs.
 * Those extra MSRs end up being shared for all events on
 * a PMU and sometimes between PMU of sibling HT threads.
 * In either case, the kernel needs to handle conflicting
 * accesses to those extra, shared, regs. The data structure
 * to manage those registers is stored in cpu_hw_event.
 */
struct extra_reg {
	unsigned int		event;
	unsigned int		msr;
	u64			config_mask;
	u64			valid_mask;
	int			idx;  /* per_xxx->regs[] reg index */
346
	bool			extra_msr_access;
347 348 349
};

#define EVENT_EXTRA_REG(e, ms, m, vm, i) {	\
350 351 352 353 354 355
	.event = (e),			\
	.msr = (ms),			\
	.config_mask = (m),		\
	.valid_mask = (vm),		\
	.idx = EXTRA_REG_##i,		\
	.extra_msr_access = true,	\
356 357 358 359 360
	}

#define INTEL_EVENT_EXTRA_REG(event, msr, vm, idx)	\
	EVENT_EXTRA_REG(event, msr, ARCH_PERFMON_EVENTSEL_EVENT, vm, idx)

361 362 363 364 365 366 367 368 369 370
#define INTEL_UEVENT_EXTRA_REG(event, msr, vm, idx) \
	EVENT_EXTRA_REG(event, msr, ARCH_PERFMON_EVENTSEL_EVENT | \
			ARCH_PERFMON_EVENTSEL_UMASK, vm, idx)

#define INTEL_UEVENT_PEBS_LDLAT_EXTRA_REG(c) \
	INTEL_UEVENT_EXTRA_REG(c, \
			       MSR_PEBS_LD_LAT_THRESHOLD, \
			       0xffff, \
			       LDLAT)

371 372 373 374 375 376 377 378 379
#define EVENT_EXTRA_END EVENT_EXTRA_REG(0, 0, 0, 0, RSP_0)

union perf_capabilities {
	struct {
		u64	lbr_format:6;
		u64	pebs_trap:1;
		u64	pebs_arch_reg:1;
		u64	pebs_format:4;
		u64	smm_freeze:1;
380 381 382 383 384
		/*
		 * PMU supports separate counter range for writing
		 * values > 32bit.
		 */
		u64	full_width_write:1;
385 386 387 388
	};
	u64	capabilities;
};

389 390 391 392 393
struct x86_pmu_quirk {
	struct x86_pmu_quirk *next;
	void (*func)(void);
};

394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416
union x86_pmu_config {
	struct {
		u64 event:8,
		    umask:8,
		    usr:1,
		    os:1,
		    edge:1,
		    pc:1,
		    interrupt:1,
		    __reserved1:1,
		    en:1,
		    inv:1,
		    cmask:8,
		    event2:4,
		    __reserved2:4,
		    go:1,
		    ho:1;
	} bits;
	u64 value;
};

#define X86_CONFIG(args...) ((union x86_pmu_config){.bits = {args}}).value

417 418
enum {
	x86_lbr_exclusive_lbr,
419
	x86_lbr_exclusive_bts,
420 421 422 423
	x86_lbr_exclusive_pt,
	x86_lbr_exclusive_max,
};

424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441
/*
 * struct x86_pmu - generic x86 pmu
 */
struct x86_pmu {
	/*
	 * Generic x86 PMC bits
	 */
	const char	*name;
	int		version;
	int		(*handle_irq)(struct pt_regs *);
	void		(*disable_all)(void);
	void		(*enable_all)(int added);
	void		(*enable)(struct perf_event *);
	void		(*disable)(struct perf_event *);
	int		(*hw_config)(struct perf_event *event);
	int		(*schedule_events)(struct cpu_hw_events *cpuc, int n, int *assign);
	unsigned	eventsel;
	unsigned	perfctr;
442
	int		(*addr_offset)(int index, bool eventsel);
443
	int		(*rdpmc_index)(int index);
444 445 446 447 448 449
	u64		(*event_map)(int);
	int		max_events;
	int		num_counters;
	int		num_counters_fixed;
	int		cntval_bits;
	u64		cntval_mask;
450 451 452 453 454
	union {
			unsigned long events_maskl;
			unsigned long events_mask[BITS_TO_LONGS(ARCH_PERFMON_EVENTS_COUNT)];
	};
	int		events_mask_len;
455 456 457 458
	int		apic;
	u64		max_period;
	struct event_constraint *
			(*get_event_constraints)(struct cpu_hw_events *cpuc,
459
						 int idx,
460 461 462 463
						 struct perf_event *event);

	void		(*put_event_constraints)(struct cpu_hw_events *cpuc,
						 struct perf_event *event);
464 465 466 467 468 469 470 471 472

	void		(*commit_scheduling)(struct cpu_hw_events *cpuc,
					     struct perf_event *event,
					     int cntr);

	void		(*start_scheduling)(struct cpu_hw_events *cpuc);

	void		(*stop_scheduling)(struct cpu_hw_events *cpuc);

473
	struct event_constraint *event_constraints;
474
	struct x86_pmu_quirk *quirks;
475
	int		perfctr_second_write;
476
	bool		late_ack;
477
	unsigned	(*limit_period)(struct perf_event *event, unsigned l);
478

479 480 481
	/*
	 * sysfs attrs
	 */
482
	int		attr_rdpmc_broken;
483
	int		attr_rdpmc;
484
	struct attribute **format_attrs;
485
	struct attribute **event_attrs;
486

487
	ssize_t		(*events_sysfs_show)(char *page, u64 config);
488
	struct attribute **cpu_events;
489

490 491 492
	/*
	 * CPU Hotplug hooks
	 */
493 494 495 496
	int		(*cpu_prepare)(int cpu);
	void		(*cpu_starting)(int cpu);
	void		(*cpu_dying)(int cpu);
	void		(*cpu_dead)(int cpu);
497 498

	void		(*check_microcode)(void);
499 500
	void		(*sched_task)(struct perf_event_context *ctx,
				      bool sched_in);
501 502 503 504 505 506 507 508 509 510

	/*
	 * Intel Arch Perfmon v2+
	 */
	u64			intel_ctrl;
	union perf_capabilities intel_cap;

	/*
	 * Intel DebugStore bits
	 */
511
	unsigned int	bts		:1,
512 513 514 515
			bts_active	:1,
			pebs		:1,
			pebs_active	:1,
			pebs_broken	:1;
516 517 518
	int		pebs_record_size;
	void		(*drain_pebs)(struct pt_regs *regs);
	struct event_constraint *pebs_constraints;
519
	void		(*pebs_aliases)(struct perf_event *event);
520
	int 		max_pebs_events;
521 522 523 524 525 526

	/*
	 * Intel LBR
	 */
	unsigned long	lbr_tos, lbr_from, lbr_to; /* MSR base regs       */
	int		lbr_nr;			   /* hardware stack size */
527 528
	u64		lbr_sel_mask;		   /* LBR_SELECT valid bits */
	const int	*lbr_sel_map;		   /* lbr_select mappings */
529
	bool		lbr_double_abort;	   /* duplicated lbr aborts */
530

531 532 533 534 535
	/*
	 * Intel PT/LBR/BTS are exclusive
	 */
	atomic_t	lbr_exclusive[x86_lbr_exclusive_max];

536 537 538 539
	/*
	 * Extra registers for events
	 */
	struct extra_reg *extra_regs;
540
	unsigned int flags;
541 542 543 544 545

	/*
	 * Intel host/guest support (KVM)
	 */
	struct perf_guest_switch_msr *(*guest_get_msrs)(int *nr);
546 547
};

548 549 550 551 552 553 554
struct x86_perf_task_context {
	u64 lbr_from[MAX_LBR_ENTRIES];
	u64 lbr_to[MAX_LBR_ENTRIES];
	int lbr_callstack_users;
	int lbr_stack_state;
};

555 556 557 558 559 560 561 562 563
#define x86_add_quirk(func_)						\
do {									\
	static struct x86_pmu_quirk __quirk __initdata = {		\
		.func = func_,						\
	};								\
	__quirk.next = x86_pmu.quirks;					\
	x86_pmu.quirks = &__quirk;					\
} while (0)

564 565 566 567 568
/*
 * x86_pmu flags
 */
#define PMU_FL_NO_HT_SHARING	0x1 /* no hyper-threading resource sharing */
#define PMU_FL_HAS_RSP_1	0x2 /* has 2 equivalent offcore_rsp regs   */
569

570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586
#define EVENT_VAR(_id)  event_attr_##_id
#define EVENT_PTR(_id) &event_attr_##_id.attr.attr

#define EVENT_ATTR(_name, _id)						\
static struct perf_pmu_events_attr EVENT_VAR(_id) = {			\
	.attr		= __ATTR(_name, 0444, events_sysfs_show, NULL),	\
	.id		= PERF_COUNT_HW_##_id,				\
	.event_str	= NULL,						\
};

#define EVENT_ATTR_STR(_name, v, str)					\
static struct perf_pmu_events_attr event_attr_##v = {			\
	.attr		= __ATTR(_name, 0444, events_sysfs_show, NULL),	\
	.id		= 0,						\
	.event_str	= str,						\
};

587 588
extern struct x86_pmu x86_pmu __read_mostly;

589 590 591 592 593 594
static inline bool x86_pmu_has_lbr_callstack(void)
{
	return  x86_pmu.lbr_sel_map &&
		x86_pmu.lbr_sel_map[PERF_SAMPLE_BRANCH_CALL_STACK_SHIFT] > 0;
}

595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621
DECLARE_PER_CPU(struct cpu_hw_events, cpu_hw_events);

int x86_perf_event_set_period(struct perf_event *event);

/*
 * Generalized hw caching related hw_event table, filled
 * in on a per model basis. A value of 0 means
 * 'not supported', -1 means 'hw_event makes no sense on
 * this CPU', any other value means the raw hw_event
 * ID.
 */

#define C(x) PERF_COUNT_HW_CACHE_##x

extern u64 __read_mostly hw_cache_event_ids
				[PERF_COUNT_HW_CACHE_MAX]
				[PERF_COUNT_HW_CACHE_OP_MAX]
				[PERF_COUNT_HW_CACHE_RESULT_MAX];
extern u64 __read_mostly hw_cache_extra_regs
				[PERF_COUNT_HW_CACHE_MAX]
				[PERF_COUNT_HW_CACHE_OP_MAX]
				[PERF_COUNT_HW_CACHE_RESULT_MAX];

u64 x86_perf_event_update(struct perf_event *event);

static inline unsigned int x86_pmu_config_addr(int index)
{
622 623
	return x86_pmu.eventsel + (x86_pmu.addr_offset ?
				   x86_pmu.addr_offset(index, true) : index);
624 625 626 627
}

static inline unsigned int x86_pmu_event_addr(int index)
{
628 629
	return x86_pmu.perfctr + (x86_pmu.addr_offset ?
				  x86_pmu.addr_offset(index, false) : index);
630 631
}

632 633 634 635 636
static inline int x86_pmu_rdpmc_index(int index)
{
	return x86_pmu.rdpmc_index ? x86_pmu.rdpmc_index(index) : index;
}

637 638 639 640 641 642
int x86_add_exclusive(unsigned int what);

void x86_del_exclusive(unsigned int what);

void hw_perf_lbr_event_destroy(struct perf_event *event);

643 644 645 646 647 648 649 650 651
int x86_setup_perfctr(struct perf_event *event);

int x86_pmu_hw_config(struct perf_event *event);

void x86_pmu_disable_all(void);

static inline void __x86_pmu_enable_event(struct hw_perf_event *hwc,
					  u64 enable_mask)
{
652 653
	u64 disable_mask = __this_cpu_read(cpu_hw_events.perf_ctr_virt_mask);

654 655
	if (hwc->extra_reg.reg)
		wrmsrl(hwc->extra_reg.reg, hwc->extra_reg.config);
656
	wrmsrl(hwc->config_base, (hwc->config | enable_mask) & ~disable_mask);
657 658 659 660
}

void x86_pmu_enable_all(int added);

661
int perf_assign_events(struct perf_event **events, int n,
Yan, Zheng's avatar
Yan, Zheng committed
662
			int wmin, int wmax, int *assign);
663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681
int x86_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign);

void x86_pmu_stop(struct perf_event *event, int flags);

static inline void x86_pmu_disable_event(struct perf_event *event)
{
	struct hw_perf_event *hwc = &event->hw;

	wrmsrl(hwc->config_base, hwc->config);
}

void x86_pmu_enable_event(struct perf_event *event);

int x86_pmu_handle_irq(struct pt_regs *regs);

extern struct event_constraint emptyconstraint;

extern struct event_constraint unconstrained;

682 683 684 685 686 687 688 689 690
static inline bool kernel_ip(unsigned long ip)
{
#ifdef CONFIG_X86_32
	return ip > PAGE_OFFSET;
#else
	return (long)ip < 0;
#endif
}

691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710
/*
 * Not all PMUs provide the right context information to place the reported IP
 * into full context. Specifically segment registers are typically not
 * supplied.
 *
 * Assuming the address is a linear address (it is for IBS), we fake the CS and
 * vm86 mode using the known zero-based code segment and 'fix up' the registers
 * to reflect this.
 *
 * Intel PEBS/LBR appear to typically provide the effective address, nothing
 * much we can do about that but pray and treat it like a linear address.
 */
static inline void set_linear_ip(struct pt_regs *regs, unsigned long ip)
{
	regs->cs = kernel_ip(ip) ? __KERNEL_CS : __USER_CS;
	if (regs->flags & X86_VM_MASK)
		regs->flags ^= (PERF_EFLAGS_VM | X86_VM_MASK);
	regs->ip = ip;
}

711
ssize_t x86_event_sysfs_show(char *page, u64 config, u64 event);
712
ssize_t intel_event_sysfs_show(char *page, u64 config);
713

714 715 716 717 718 719 720 721 722 723 724 725 726 727 728
#ifdef CONFIG_CPU_SUP_AMD

int amd_pmu_init(void);

#else /* CONFIG_CPU_SUP_AMD */

static inline int amd_pmu_init(void)
{
	return 0;
}

#endif /* CONFIG_CPU_SUP_AMD */

#ifdef CONFIG_CPU_SUP_INTEL

729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751
static inline bool intel_pmu_needs_lbr_smpl(struct perf_event *event)
{
	/* user explicitly requested branch sampling */
	if (has_branch_stack(event))
		return true;

	/* implicit branch sampling to correct PEBS skid */
	if (x86_pmu.intel_cap.pebs_trap && event->attr.precise_ip > 1 &&
	    x86_pmu.intel_cap.pebs_format < 2)
		return true;

	return false;
}

static inline bool intel_pmu_has_bts(struct perf_event *event)
{
	if (event->attr.config == PERF_COUNT_HW_BRANCH_INSTRUCTIONS &&
	    !event->attr.freq && event->hw.sample_period == 1)
		return true;

	return false;
}

752 753 754
int intel_pmu_save_and_restart(struct perf_event *event);

struct event_constraint *
755 756
x86_get_event_constraints(struct cpu_hw_events *cpuc, int idx,
			  struct perf_event *event);
757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781

struct intel_shared_regs *allocate_shared_regs(int cpu);

int intel_pmu_init(void);

void init_debug_store_on_cpu(int cpu);

void fini_debug_store_on_cpu(int cpu);

void release_ds_buffers(void);

void reserve_ds_buffers(void);

extern struct event_constraint bts_constraint;

void intel_pmu_enable_bts(u64 config);

void intel_pmu_disable_bts(void);

int intel_pmu_drain_bts_buffer(void);

extern struct event_constraint intel_core2_pebs_event_constraints[];

extern struct event_constraint intel_atom_pebs_event_constraints[];

782 783
extern struct event_constraint intel_slm_pebs_event_constraints[];

784 785 786 787 788 789
extern struct event_constraint intel_nehalem_pebs_event_constraints[];

extern struct event_constraint intel_westmere_pebs_event_constraints[];

extern struct event_constraint intel_snb_pebs_event_constraints[];

790 791
extern struct event_constraint intel_ivb_pebs_event_constraints[];

792 793
extern struct event_constraint intel_hsw_pebs_event_constraints[];

794 795 796 797 798 799 800 801 802 803 804 805
struct event_constraint *intel_pebs_constraints(struct perf_event *event);

void intel_pmu_pebs_enable(struct perf_event *event);

void intel_pmu_pebs_disable(struct perf_event *event);

void intel_pmu_pebs_enable_all(void);

void intel_pmu_pebs_disable_all(void);

void intel_ds_init(void);

806 807
void intel_pmu_lbr_sched_task(struct perf_event_context *ctx, bool sched_in);

808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825
void intel_pmu_lbr_reset(void);

void intel_pmu_lbr_enable(struct perf_event *event);

void intel_pmu_lbr_disable(struct perf_event *event);

void intel_pmu_lbr_enable_all(void);

void intel_pmu_lbr_disable_all(void);

void intel_pmu_lbr_read(void);

void intel_pmu_lbr_init_core(void);

void intel_pmu_lbr_init_nhm(void);

void intel_pmu_lbr_init_atom(void);

826 827
void intel_pmu_lbr_init_snb(void);

828 829
void intel_pmu_lbr_init_hsw(void);

830 831
int intel_pmu_setup_lbr_filter(struct perf_event *event);

832 833
void intel_pt_interrupt(void);

834 835 836 837 838 839
int intel_bts_interrupt(void);

void intel_bts_enable_local(void);

void intel_bts_disable_local(void);

840 841 842 843
int p4_pmu_init(void);

int p6_pmu_init(void);

844 845
int knc_pmu_init(void);

846 847 848
ssize_t events_sysfs_show(struct device *dev, struct device_attribute *attr,
			  char *page);

849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869
#else /* CONFIG_CPU_SUP_INTEL */

static inline void reserve_ds_buffers(void)
{
}

static inline void release_ds_buffers(void)
{
}

static inline int intel_pmu_init(void)
{
	return 0;
}

static inline struct intel_shared_regs *allocate_shared_regs(int cpu)
{
	return NULL;
}

#endif /* CONFIG_CPU_SUP_INTEL */