intel_ringbuffer.h 20.7 KB
Newer Older
1 2 3
#ifndef _INTEL_RINGBUFFER_H_
#define _INTEL_RINGBUFFER_H_

4
#include <linux/hashtable.h>
5
#include "i915_gem_batch_pool.h"
6
#include "i915_gem_request.h"
7 8 9

#define I915_CMD_HASH_ORDER 9

10 11 12 13 14 15
/* Early gen2 devices have a cacheline of just 32 bytes, using 64 is overkill,
 * but keeps the logic simple. Indeed, the whole purpose of this macro is just
 * to give some inclination as to some of the magic values used in the various
 * workarounds!
 */
#define CACHELINE_BYTES 64
16
#define CACHELINE_DWORDS (CACHELINE_BYTES / sizeof(uint32_t))
17

18 19 20 21 22 23 24 25 26 27 28
/*
 * Gen2 BSpec "1. Programming Environment" / 1.4.4.6 "Ring Buffer Use"
 * Gen3 BSpec "vol1c Memory Interface Functions" / 2.3.4.5 "Ring Buffer Use"
 * Gen4+ BSpec "vol1c Memory Interface and Command Stream" / 5.3.4.5 "Ring Buffer Use"
 *
 * "If the Ring Buffer Head Pointer and the Tail Pointer are on the same
 * cacheline, the Head Pointer must not be greater than the Tail
 * Pointer."
 */
#define I915_RING_FREE_SPACE 64

29 30 31 32
struct intel_hw_status_page {
	struct i915_vma *vma;
	u32 *page_addr;
	u32 ggtt_offset;
33 34
};

35 36
#define I915_READ_TAIL(engine) I915_READ(RING_TAIL((engine)->mmio_base))
#define I915_WRITE_TAIL(engine, val) I915_WRITE(RING_TAIL((engine)->mmio_base), val)
37

38 39
#define I915_READ_START(engine) I915_READ(RING_START((engine)->mmio_base))
#define I915_WRITE_START(engine, val) I915_WRITE(RING_START((engine)->mmio_base), val)
40

41 42
#define I915_READ_HEAD(engine)  I915_READ(RING_HEAD((engine)->mmio_base))
#define I915_WRITE_HEAD(engine, val) I915_WRITE(RING_HEAD((engine)->mmio_base), val)
43

44 45
#define I915_READ_CTL(engine) I915_READ(RING_CTL((engine)->mmio_base))
#define I915_WRITE_CTL(engine, val) I915_WRITE(RING_CTL((engine)->mmio_base), val)
46

47 48
#define I915_READ_IMR(engine) I915_READ(RING_IMR((engine)->mmio_base))
#define I915_WRITE_IMR(engine, val) I915_WRITE(RING_IMR((engine)->mmio_base), val)
49

50 51
#define I915_READ_MODE(engine) I915_READ(RING_MI_MODE((engine)->mmio_base))
#define I915_WRITE_MODE(engine, val) I915_WRITE(RING_MI_MODE((engine)->mmio_base), val)
52

53 54 55
/* seqno size is actually only a uint32, but since we plan to use MI_FLUSH_DW to
 * do the writes, and that must have qw aligned offsets, simply pretend it's 8b.
 */
56 57 58
#define gen8_semaphore_seqno_size sizeof(uint64_t)
#define GEN8_SEMAPHORE_OFFSET(__from, __to)			     \
	(((__from) * I915_NUM_ENGINES  + (__to)) * gen8_semaphore_seqno_size)
59
#define GEN8_SIGNAL_OFFSET(__ring, to)			     \
60
	(dev_priv->semaphore->node.start + \
61
	 GEN8_SEMAPHORE_OFFSET((__ring)->id, (to)))
62
#define GEN8_WAIT_OFFSET(__ring, from)			     \
63
	(dev_priv->semaphore->node.start + \
64
	 GEN8_SEMAPHORE_OFFSET(from, (__ring)->id))
65

66
enum intel_engine_hangcheck_action {
67
	HANGCHECK_IDLE = 0,
68 69 70 71 72
	HANGCHECK_WAIT,
	HANGCHECK_ACTIVE,
	HANGCHECK_KICK,
	HANGCHECK_HUNG,
};
73

74 75
#define HANGCHECK_SCORE_RING_HUNG 31

76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94
#define I915_MAX_SLICES	3
#define I915_MAX_SUBSLICES 3

#define instdone_slice_mask(dev_priv__) \
	(INTEL_GEN(dev_priv__) == 7 ? \
	 1 : INTEL_INFO(dev_priv__)->sseu.slice_mask)

#define instdone_subslice_mask(dev_priv__) \
	(INTEL_GEN(dev_priv__) == 7 ? \
	 1 : INTEL_INFO(dev_priv__)->sseu.subslice_mask)

#define for_each_instdone_slice_subslice(dev_priv__, slice__, subslice__) \
	for ((slice__) = 0, (subslice__) = 0; \
	     (slice__) < I915_MAX_SLICES; \
	     (subslice__) = ((subslice__) + 1) < I915_MAX_SUBSLICES ? (subslice__) + 1 : 0, \
	       (slice__) += ((subslice__) == 0)) \
		for_each_if((BIT(slice__) & instdone_slice_mask(dev_priv__)) && \
			    (BIT(subslice__) & instdone_subslice_mask(dev_priv__)))

95 96 97 98
struct intel_instdone {
	u32 instdone;
	/* The following exist only in the RCS engine */
	u32 slice_common;
99 100
	u32 sampler[I915_MAX_SLICES][I915_MAX_SUBSLICES];
	u32 row[I915_MAX_SLICES][I915_MAX_SUBSLICES];
101 102
};

103
struct intel_engine_hangcheck {
104
	u64 acthd;
105
	u32 seqno;
106
	int score;
107
	enum intel_engine_hangcheck_action action;
108
	int deadlock;
109
	struct intel_instdone instdone;
110 111
};

112
struct intel_ring {
113
	struct i915_vma *vma;
114
	void *vaddr;
115

116
	struct intel_engine_cs *engine;
117

118 119
	struct list_head request_list;

120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136
	u32 head;
	u32 tail;
	int space;
	int size;
	int effective_size;

	/** We track the position of the requests in the ring buffer, and
	 * when each is retired we increment last_retired_head as the GPU
	 * must have finished processing the request and so we know we
	 * can advance the ringbuffer up to that position.
	 *
	 * last_retired_head is set to -1 after the value is consumed so
	 * we can detect new retirements.
	 */
	u32 last_retired_head;
};

137
struct i915_gem_context;
138
struct drm_i915_reg_table;
139

140 141 142 143 144 145 146 147 148 149 150
/*
 * we use a single page to load ctx workarounds so all of these
 * values are referred in terms of dwords
 *
 * struct i915_wa_ctx_bb:
 *  offset: specifies batch starting position, also helpful in case
 *    if we want to have multiple batches at different offsets based on
 *    some criteria. It is not a requirement at the moment but provides
 *    an option for future use.
 *  size: size of the batch in DWORDS
 */
151
struct i915_ctx_workarounds {
152 153 154 155
	struct i915_wa_ctx_bb {
		u32 offset;
		u32 size;
	} indirect_ctx, per_ctx;
156
	struct i915_vma *vma;
157 158
};

159 160
struct drm_i915_gem_request;

161 162
struct intel_engine_cs {
	struct drm_i915_private *i915;
163
	const char	*name;
164
	enum intel_engine_id {
165
		RCS = 0,
166
		BCS,
167 168 169
		VCS,
		VCS2,	/* Keep instances of the same type engine together. */
		VECS
170
	} id;
171
#define I915_NUM_ENGINES 5
172
#define _VCS(n) (VCS + (n))
173
	unsigned int exec_id;
174 175 176 177 178 179 180 181
	enum intel_engine_hw_id {
		RCS_HW = 0,
		VCS_HW,
		BCS_HW,
		VECS_HW,
		VCS2_HW
	} hw_id;
	enum intel_engine_hw_id guc_id; /* XXX same as hw_id? */
182
	u64 fence_context;
183
	u32		mmio_base;
184
	unsigned int irq_shift;
185
	struct intel_ring *buffer;
186

187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203
	/* Rather than have every client wait upon all user interrupts,
	 * with the herd waking after every interrupt and each doing the
	 * heavyweight seqno dance, we delegate the task (of being the
	 * bottom-half of the user interrupt) to the first client. After
	 * every interrupt, we wake up one client, who does the heavyweight
	 * coherent seqno read and either goes back to sleep (if incomplete),
	 * or wakes up all the completed clients in parallel, before then
	 * transferring the bottom-half status to the next client in the queue.
	 *
	 * Compared to walking the entire list of waiters in a single dedicated
	 * bottom-half, we reduce the latency of the first waiter by avoiding
	 * a context switch, but incur additional coherent seqno reads when
	 * following the chain of request breadcrumbs. Since it is most likely
	 * that we have a single client waiting on each seqno, then reducing
	 * the overhead of waking that client is much preferred.
	 */
	struct intel_breadcrumbs {
204
		struct task_struct __rcu *irq_seqno_bh; /* bh for interrupts */
205 206
		bool irq_posted;

207 208
		spinlock_t lock; /* protects the lists of requests */
		struct rb_root waiters; /* sorted by retirement, priority */
209
		struct rb_root signals; /* sorted by retirement */
210
		struct intel_wait *first_wait; /* oldest waiter by retirement */
211
		struct task_struct *signaler; /* used for fence signalling */
212
		struct drm_i915_gem_request *first_signal;
213
		struct timer_list fake_irq; /* used after a missed interrupt */
214 215 216
		struct timer_list hangcheck; /* detect missed interrupts */

		unsigned long timeout;
217 218 219

		bool irq_enabled : 1;
		bool rpm_wakelock : 1;
220 221
	} breadcrumbs;

222 223 224 225 226 227 228
	/*
	 * A pool of objects to use as shadow copies of client batch buffers
	 * when the command parser is enabled. Prevents the client from
	 * modifying the batch contents after software parsing.
	 */
	struct i915_gem_batch_pool batch_pool;

229
	struct intel_hw_status_page status_page;
230
	struct i915_ctx_workarounds wa_ctx;
231
	struct i915_vma *scratch;
232

233 234
	u32             irq_keep_mask; /* always keep these interrupts */
	u32		irq_enable_mask; /* bitmask to enable ring interrupt */
235 236
	void		(*irq_enable)(struct intel_engine_cs *engine);
	void		(*irq_disable)(struct intel_engine_cs *engine);
237

238
	int		(*init_hw)(struct intel_engine_cs *engine);
239 240
	void		(*reset_hw)(struct intel_engine_cs *engine,
				    struct drm_i915_gem_request *req);
241

242
	int		(*init_context)(struct drm_i915_gem_request *req);
243

244 245 246 247 248 249 250 251 252 253 254 255
	int		(*emit_flush)(struct drm_i915_gem_request *request,
				      u32 mode);
#define EMIT_INVALIDATE	BIT(0)
#define EMIT_FLUSH	BIT(1)
#define EMIT_BARRIER	(EMIT_INVALIDATE | EMIT_FLUSH)
	int		(*emit_bb_start)(struct drm_i915_gem_request *req,
					 u64 offset, u32 length,
					 unsigned int dispatch_flags);
#define I915_DISPATCH_SECURE BIT(0)
#define I915_DISPATCH_PINNED BIT(1)
#define I915_DISPATCH_RS     BIT(2)
	int		(*emit_request)(struct drm_i915_gem_request *req);
256 257 258 259 260 261 262

	/* Pass the request to the hardware queue (e.g. directly into
	 * the legacy ringbuffer or to the end of an execlist).
	 *
	 * This is called from an atomic context with irqs disabled; must
	 * be irq safe.
	 */
263
	void		(*submit_request)(struct drm_i915_gem_request *req);
264

265 266 267 268 269 270
	/* Some chipsets are not quite as coherent as advertised and need
	 * an expensive kick to force a true read of the up-to-date seqno.
	 * However, the up-to-date seqno is not always required and the last
	 * seen value is good enough. Note that the seqno will always be
	 * monotonic, even if not coherent.
	 */
271 272
	void		(*irq_seqno_barrier)(struct intel_engine_cs *engine);
	void		(*cleanup)(struct intel_engine_cs *engine);
273

274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310
	/* GEN8 signal/wait table - never trust comments!
	 *	  signal to	signal to    signal to   signal to      signal to
	 *	    RCS		   VCS          BCS        VECS		 VCS2
	 *      --------------------------------------------------------------------
	 *  RCS | NOP (0x00) | VCS (0x08) | BCS (0x10) | VECS (0x18) | VCS2 (0x20) |
	 *	|-------------------------------------------------------------------
	 *  VCS | RCS (0x28) | NOP (0x30) | BCS (0x38) | VECS (0x40) | VCS2 (0x48) |
	 *	|-------------------------------------------------------------------
	 *  BCS | RCS (0x50) | VCS (0x58) | NOP (0x60) | VECS (0x68) | VCS2 (0x70) |
	 *	|-------------------------------------------------------------------
	 * VECS | RCS (0x78) | VCS (0x80) | BCS (0x88) |  NOP (0x90) | VCS2 (0x98) |
	 *	|-------------------------------------------------------------------
	 * VCS2 | RCS (0xa0) | VCS (0xa8) | BCS (0xb0) | VECS (0xb8) | NOP  (0xc0) |
	 *	|-------------------------------------------------------------------
	 *
	 * Generalization:
	 *  f(x, y) := (x->id * NUM_RINGS * seqno_size) + (seqno_size * y->id)
	 *  ie. transpose of g(x, y)
	 *
	 *	 sync from	sync from    sync from    sync from	sync from
	 *	    RCS		   VCS          BCS        VECS		 VCS2
	 *      --------------------------------------------------------------------
	 *  RCS | NOP (0x00) | VCS (0x28) | BCS (0x50) | VECS (0x78) | VCS2 (0xa0) |
	 *	|-------------------------------------------------------------------
	 *  VCS | RCS (0x08) | NOP (0x30) | BCS (0x58) | VECS (0x80) | VCS2 (0xa8) |
	 *	|-------------------------------------------------------------------
	 *  BCS | RCS (0x10) | VCS (0x38) | NOP (0x60) | VECS (0x88) | VCS2 (0xb0) |
	 *	|-------------------------------------------------------------------
	 * VECS | RCS (0x18) | VCS (0x40) | BCS (0x68) |  NOP (0x90) | VCS2 (0xb8) |
	 *	|-------------------------------------------------------------------
	 * VCS2 | RCS (0x20) | VCS (0x48) | BCS (0x70) | VECS (0x98) |  NOP (0xc0) |
	 *	|-------------------------------------------------------------------
	 *
	 * Generalization:
	 *  g(x, y) := (y->id * NUM_RINGS * seqno_size) + (seqno_size * x->id)
	 *  ie. transpose of f(x, y)
	 */
311
	struct {
312
		u32	sync_seqno[I915_NUM_ENGINES-1];
313

314
		union {
315 316 317
#define GEN6_SEMAPHORE_LAST	VECS_HW
#define GEN6_NUM_SEMAPHORES	(GEN6_SEMAPHORE_LAST + 1)
#define GEN6_SEMAPHORES_MASK	GENMASK(GEN6_SEMAPHORE_LAST, 0)
318 319
			struct {
				/* our mbox written by others */
320
				u32		wait[GEN6_NUM_SEMAPHORES];
321
				/* mboxes this ring signals to */
322
				i915_reg_t	signal[GEN6_NUM_SEMAPHORES];
323
			} mbox;
324
			u64		signal_ggtt[I915_NUM_ENGINES];
325
		};
326 327

		/* AKA wait() */
328 329 330
		int	(*sync_to)(struct drm_i915_gem_request *req,
				   struct drm_i915_gem_request *signal);
		int	(*signal)(struct drm_i915_gem_request *req);
331
	} semaphore;
332

333
	/* Execlists */
334 335
	struct tasklet_struct irq_tasklet;
	spinlock_t execlist_lock; /* used inside tasklet, use spin_lock_bh */
336 337 338 339
	struct execlist_port {
		struct drm_i915_gem_request *request;
		unsigned int count;
	} execlist_port[2];
340
	struct list_head execlist_queue;
341
	unsigned int fw_domains;
342
	bool disable_lite_restore_wa;
343
	bool preempt_wa;
344
	u32 ctx_desc_template;
345

346 347 348 349 350 351
	/**
	 * List of breadcrumbs associated with GPU requests currently
	 * outstanding.
	 */
	struct list_head request_list;

352 353 354 355 356 357
	/**
	 * Seqno of request most recently submitted to request_list.
	 * Used exclusively by hang checker to avoid grabbing lock while
	 * inspecting request list.
	 */
	u32 last_submitted_seqno;
358
	u32 last_pending_seqno;
359

360 361
	/* An RCU guarded pointer to the last request. No reference is
	 * held to the request, users must carefully acquire a reference to
362
	 * the request using i915_gem_active_get_rcu(), or hold the
363 364 365 366
	 * struct_mutex.
	 */
	struct i915_gem_active last_request;

367
	struct i915_gem_context *last_context;
368

369
	struct intel_engine_hangcheck hangcheck;
370

371 372
	bool needs_cmd_parser;

373
	/*
374
	 * Table of commands the command parser needs to know about
375
	 * for this engine.
376
	 */
377
	DECLARE_HASHTABLE(cmd_hash, I915_CMD_HASH_ORDER);
378 379 380 381

	/*
	 * Table of registers allowed in commands that read/write registers.
	 */
382 383
	const struct drm_i915_reg_table *reg_tables;
	int reg_table_count;
384 385 386 387 388

	/*
	 * Returns the bitmask for the length field of the specified command.
	 * Return 0 for an unrecognized/invalid command.
	 *
389
	 * If the command parser finds an entry for a command in the engine's
390
	 * cmd_tables, it gets the command's length based on the table entry.
391 392 393
	 * If not, it calls this function to determine the per-engine length
	 * field encoding for the command (i.e. different opcode ranges use
	 * certain bits to encode the command length in the header).
394 395
	 */
	u32 (*get_cmd_length_mask)(u32 cmd_header);
396 397
};

398
static inline bool
399
intel_engine_initialized(const struct intel_engine_cs *engine)
400
{
401
	return engine->i915 != NULL;
402
}
403

404
static inline unsigned
405
intel_engine_flag(const struct intel_engine_cs *engine)
406
{
407
	return 1 << engine->id;
408 409
}

410
static inline u32
411 412
intel_engine_sync_index(struct intel_engine_cs *engine,
			struct intel_engine_cs *other)
413 414 415 416
{
	int idx;

	/*
R
Rodrigo Vivi 已提交
417 418 419 420 421
	 * rcs -> 0 = vcs, 1 = bcs, 2 = vecs, 3 = vcs2;
	 * vcs -> 0 = bcs, 1 = vecs, 2 = vcs2, 3 = rcs;
	 * bcs -> 0 = vecs, 1 = vcs2. 2 = rcs, 3 = vcs;
	 * vecs -> 0 = vcs2, 1 = rcs, 2 = vcs, 3 = bcs;
	 * vcs2 -> 0 = rcs, 1 = vcs, 2 = bcs, 3 = vecs;
422 423
	 */

424
	idx = (other - engine) - 1;
425
	if (idx < 0)
426
		idx += I915_NUM_ENGINES;
427 428 429 430

	return idx;
}

431
static inline void
432
intel_flush_status_page(struct intel_engine_cs *engine, int reg)
433
{
434 435 436
	mb();
	clflush(&engine->status_page.page_addr[reg]);
	mb();
437 438
}

439
static inline u32
440
intel_read_status_page(struct intel_engine_cs *engine, int reg)
441
{
442
	/* Ensure that the compiler doesn't optimize away the load. */
443
	return READ_ONCE(engine->status_page.page_addr[reg]);
444 445
}

M
Mika Kuoppala 已提交
446
static inline void
447
intel_write_status_page(struct intel_engine_cs *engine,
M
Mika Kuoppala 已提交
448 449
			int reg, u32 value)
{
450
	engine->status_page.page_addr[reg] = value;
M
Mika Kuoppala 已提交
451 452
}

453
/*
C
Chris Wilson 已提交
454 455 456 457 458 459 460 461 462 463 464
 * Reads a dword out of the status page, which is written to from the command
 * queue by automatic updates, MI_REPORT_HEAD, MI_STORE_DATA_INDEX, or
 * MI_STORE_DATA_IMM.
 *
 * The following dwords have a reserved meaning:
 * 0x00: ISR copy, updated when an ISR bit not set in the HWSTAM changes.
 * 0x04: ring 0 head pointer
 * 0x05: ring 1 head pointer (915-class)
 * 0x06: ring 2 head pointer (915-class)
 * 0x10-0x1b: Context status DWords (GM45)
 * 0x1f: Last written status offset. (GM45)
465
 * 0x20-0x2f: Reserved (Gen6+)
C
Chris Wilson 已提交
466
 *
467
 * The area from dword 0x30 to 0x3ff is available for driver usage.
C
Chris Wilson 已提交
468
 */
469
#define I915_GEM_HWS_INDEX		0x30
470
#define I915_GEM_HWS_INDEX_ADDR (I915_GEM_HWS_INDEX << MI_STORE_DWORD_INDEX_SHIFT)
471
#define I915_GEM_HWS_SCRATCH_INDEX	0x40
472
#define I915_GEM_HWS_SCRATCH_ADDR (I915_GEM_HWS_SCRATCH_INDEX << MI_STORE_DWORD_INDEX_SHIFT)
C
Chris Wilson 已提交
473

474 475
struct intel_ring *
intel_engine_create_ring(struct intel_engine_cs *engine, int size);
476 477
int intel_ring_pin(struct intel_ring *ring);
void intel_ring_unpin(struct intel_ring *ring);
478
void intel_ring_free(struct intel_ring *ring);
479

480 481
void intel_engine_stop(struct intel_engine_cs *engine);
void intel_engine_cleanup(struct intel_engine_cs *engine);
482

483 484
void intel_legacy_submission_resume(struct drm_i915_private *dev_priv);

485 486
int intel_ring_alloc_request_extras(struct drm_i915_gem_request *request);

487
int __must_check intel_ring_begin(struct drm_i915_gem_request *req, int n);
488
int __must_check intel_ring_cacheline_align(struct drm_i915_gem_request *req);
489

490
static inline void intel_ring_emit(struct intel_ring *ring, u32 data)
491
{
492 493
	*(uint32_t *)(ring->vaddr + ring->tail) = data;
	ring->tail += 4;
494 495
}

496
static inline void intel_ring_emit_reg(struct intel_ring *ring, i915_reg_t reg)
497
{
498
	intel_ring_emit(ring, i915_mmio_reg_offset(reg));
499
}
500

501
static inline void intel_ring_advance(struct intel_ring *ring)
502
{
503 504 505 506 507 508 509
	/* Dummy function.
	 *
	 * This serves as a placeholder in the code so that the reader
	 * can compare against the preceding intel_ring_begin() and
	 * check that the number of dwords emitted matches the space
	 * reserved for the command packet (i.e. the value passed to
	 * intel_ring_begin()).
510
	 */
511 512 513 514 515 516
}

static inline u32 intel_ring_offset(struct intel_ring *ring, u32 value)
{
	/* Don't write ring->size (equivalent to 0) as that hangs some GPUs. */
	return value & (ring->size - 1);
517
}
518

519
int __intel_ring_space(int head, int tail, int size);
520
void intel_ring_update_space(struct intel_ring *ring);
521

522
void intel_engine_init_seqno(struct intel_engine_cs *engine, u32 seqno);
523

524 525
void intel_engine_setup_common(struct intel_engine_cs *engine);
int intel_engine_init_common(struct intel_engine_cs *engine);
526
int intel_engine_create_scratch(struct intel_engine_cs *engine, int size);
527
void intel_engine_cleanup_common(struct intel_engine_cs *engine);
528

529
static inline int intel_engine_idle(struct intel_engine_cs *engine,
530
				    unsigned int flags)
531 532 533
{
	/* Wait upon the last request to be completed */
	return i915_gem_active_wait_unlocked(&engine->last_request,
534
					     flags, NULL, NULL);
535 536
}

537 538 539 540 541
int intel_init_render_ring_buffer(struct intel_engine_cs *engine);
int intel_init_bsd_ring_buffer(struct intel_engine_cs *engine);
int intel_init_bsd2_ring_buffer(struct intel_engine_cs *engine);
int intel_init_blt_ring_buffer(struct intel_engine_cs *engine);
int intel_init_vebox_ring_buffer(struct intel_engine_cs *engine);
542

543
u64 intel_engine_get_active_head(struct intel_engine_cs *engine);
544 545
u64 intel_engine_get_last_batch_head(struct intel_engine_cs *engine);

546 547 548 549
static inline u32 intel_engine_get_seqno(struct intel_engine_cs *engine)
{
	return intel_read_status_page(engine, I915_GEM_HWS_INDEX);
}
550

551
int init_workarounds_ring(struct intel_engine_cs *engine);
552

553 554 555
/*
 * Arbitrary size for largest possible 'add request' sequence. The code paths
 * are complex and variable. Empirical measurement shows that the worst case
556 557 558
 * is BDW at 192 bytes (6 + 6 + 36 dwords), then ILK at 136 bytes. However,
 * we need to allocate double the largest single packet within that emission
 * to account for tail wraparound (so 6 + 6 + 72 dwords for BDW).
559
 */
560
#define MIN_SPACE_FOR_ADD_REQUEST 336
561

562 563
static inline u32 intel_hws_seqno_address(struct intel_engine_cs *engine)
{
564
	return engine->status_page.ggtt_offset + I915_GEM_HWS_INDEX_ADDR;
565 566
}

567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584
/* intel_breadcrumbs.c -- user interrupt bottom-half for waiters */
int intel_engine_init_breadcrumbs(struct intel_engine_cs *engine);

static inline void intel_wait_init(struct intel_wait *wait, u32 seqno)
{
	wait->tsk = current;
	wait->seqno = seqno;
}

static inline bool intel_wait_complete(const struct intel_wait *wait)
{
	return RB_EMPTY_NODE(&wait->node);
}

bool intel_engine_add_wait(struct intel_engine_cs *engine,
			   struct intel_wait *wait);
void intel_engine_remove_wait(struct intel_engine_cs *engine,
			      struct intel_wait *wait);
585
void intel_engine_enable_signaling(struct drm_i915_gem_request *request);
586

587
static inline bool intel_engine_has_waiter(const struct intel_engine_cs *engine)
588
{
589
	return rcu_access_pointer(engine->breadcrumbs.irq_seqno_bh);
590 591
}

592
static inline bool intel_engine_wakeup(const struct intel_engine_cs *engine)
593 594
{
	bool wakeup = false;
595

596
	/* Note that for this not to dangerously chase a dangling pointer,
597
	 * we must hold the rcu_read_lock here.
598 599 600 601 602
	 *
	 * Also note that tsk is likely to be in !TASK_RUNNING state so an
	 * early test for tsk->state != TASK_RUNNING before wake_up_process()
	 * is unlikely to be beneficial.
	 */
603 604 605 606 607 608 609 610 611 612
	if (intel_engine_has_waiter(engine)) {
		struct task_struct *tsk;

		rcu_read_lock();
		tsk = rcu_dereference(engine->breadcrumbs.irq_seqno_bh);
		if (tsk)
			wakeup = wake_up_process(tsk);
		rcu_read_unlock();
	}

613 614 615
	return wakeup;
}

616
void intel_engine_reset_breadcrumbs(struct intel_engine_cs *engine);
617 618
void intel_engine_fini_breadcrumbs(struct intel_engine_cs *engine);
unsigned int intel_kick_waiters(struct drm_i915_private *i915);
619
unsigned int intel_kick_signalers(struct drm_i915_private *i915);
620

621 622 623 624 625
static inline bool intel_engine_is_active(struct intel_engine_cs *engine)
{
	return i915_gem_active_isset(&engine->last_request);
}

626
#endif /* _INTEL_RINGBUFFER_H_ */