intel_ringbuffer.h 24.9 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
#include "i915_gem_timeline.h"
8
#include "i915_selftest.h"
9 10 11

#define I915_CMD_HASH_ORDER 9

12 13 14 15 16 17
/* 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
18
#define CACHELINE_DWORDS (CACHELINE_BYTES / sizeof(uint32_t))
19

20 21 22 23
struct intel_hw_status_page {
	struct i915_vma *vma;
	u32 *page_addr;
	u32 ggtt_offset;
24 25
};

26 27
#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)
28

29 30
#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)
31

32 33
#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)
34

35 36
#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)
37

38 39
#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)
40

41 42
#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)
43

44 45 46
/* 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.
 */
47 48 49
#define gen8_semaphore_seqno_size sizeof(uint64_t)
#define GEN8_SEMAPHORE_OFFSET(__from, __to)			     \
	(((__from) * I915_NUM_ENGINES  + (__to)) * gen8_semaphore_seqno_size)
50
#define GEN8_SIGNAL_OFFSET(__ring, to)			     \
51
	(dev_priv->semaphore->node.start + \
52
	 GEN8_SEMAPHORE_OFFSET((__ring)->id, (to)))
53
#define GEN8_WAIT_OFFSET(__ring, from)			     \
54
	(dev_priv->semaphore->node.start + \
55
	 GEN8_SEMAPHORE_OFFSET(from, (__ring)->id))
56

57
enum intel_engine_hangcheck_action {
58 59 60 61 62 63 64
	ENGINE_IDLE = 0,
	ENGINE_WAIT,
	ENGINE_ACTIVE_SEQNO,
	ENGINE_ACTIVE_HEAD,
	ENGINE_ACTIVE_SUBUNITS,
	ENGINE_WAIT_KICK,
	ENGINE_DEAD,
65
};
66

67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88
static inline const char *
hangcheck_action_to_str(const enum intel_engine_hangcheck_action a)
{
	switch (a) {
	case ENGINE_IDLE:
		return "idle";
	case ENGINE_WAIT:
		return "wait";
	case ENGINE_ACTIVE_SEQNO:
		return "active seqno";
	case ENGINE_ACTIVE_HEAD:
		return "active head";
	case ENGINE_ACTIVE_SUBUNITS:
		return "active subunits";
	case ENGINE_WAIT_KICK:
		return "wait kick";
	case ENGINE_DEAD:
		return "dead";
	}

	return "unknown";
}
89

90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108
#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__)))

109 110 111 112
struct intel_instdone {
	u32 instdone;
	/* The following exist only in the RCS engine */
	u32 slice_common;
113 114
	u32 sampler[I915_MAX_SLICES][I915_MAX_SUBSLICES];
	u32 row[I915_MAX_SLICES][I915_MAX_SUBSLICES];
115 116
};

117
struct intel_engine_hangcheck {
118
	u64 acthd;
119
	u32 seqno;
120
	enum intel_engine_hangcheck_action action;
121
	unsigned long action_timestamp;
122
	int deadlock;
123
	struct intel_instdone instdone;
124
	struct drm_i915_gem_request *active_request;
125
	bool stalled;
126 127
};

128
struct intel_ring {
129
	struct i915_vma *vma;
130
	void *vaddr;
131

132 133
	struct list_head request_list;

134 135
	u32 head;
	u32 tail;
136
	u32 emit;
137

138 139 140
	u32 space;
	u32 size;
	u32 effective_size;
141 142
};

143
struct i915_gem_context;
144
struct drm_i915_reg_table;
145

146 147 148 149 150 151 152 153 154 155 156
/*
 * 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
 */
157
struct i915_ctx_workarounds {
158 159 160 161
	struct i915_wa_ctx_bb {
		u32 offset;
		u32 size;
	} indirect_ctx, per_ctx;
162
	struct i915_vma *vma;
163 164
};

165
struct drm_i915_gem_request;
166
struct intel_render_state;
167

168 169 170 171 172 173 174 175 176 177 178 179 180
/*
 * Engine IDs definitions.
 * Keep instances of the same type engine together.
 */
enum intel_engine_id {
	RCS = 0,
	BCS,
	VCS,
	VCS2,
#define _VCS(n) (VCS + (n))
	VECS
};

181 182 183 184 185 186
struct i915_priolist {
	struct rb_node node;
	struct list_head requests;
	int priority;
};

187 188
#define INTEL_ENGINE_CS_MAX_NAME 8

189 190
struct intel_engine_cs {
	struct drm_i915_private *i915;
191
	char name[INTEL_ENGINE_CS_MAX_NAME];
192
	enum intel_engine_id id;
193
	unsigned int uabi_id;
194
	unsigned int hw_id;
195
	unsigned int guc_id;
196 197 198

	u8 class;
	u8 instance;
199 200
	u32 context_size;
	u32 mmio_base;
201
	unsigned int irq_shift;
202

203
	struct intel_ring *buffer;
204
	struct intel_timeline *timeline;
205

206 207
	struct intel_render_state *render_state;

208
	atomic_t irq_count;
209 210
	unsigned long irq_posted;
#define ENGINE_IRQ_BREADCRUMB 0
211
#define ENGINE_IRQ_EXECLIST 1
212

213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229
	/* 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 {
230 231 232 233
		spinlock_t irq_lock; /* protects irq_*; irqsafe */
		struct intel_wait *irq_wait; /* oldest waiter by retirement */

		spinlock_t rb_lock; /* protects the rb and wraps irq_lock */
234
		struct rb_root waiters; /* sorted by retirement, priority */
235 236
		struct rb_root signals; /* sorted by retirement */
		struct task_struct *signaler; /* used for fence signalling */
237
		struct drm_i915_gem_request __rcu *first_signal;
238
		struct timer_list fake_irq; /* used after a missed interrupt */
239 240
		struct timer_list hangcheck; /* detect missed interrupts */

241
		unsigned int hangcheck_interrupts;
242

243
		bool irq_armed : 1;
244
		bool irq_enabled : 1;
245
		I915_SELFTEST_DECLARE(bool mock : 1);
246 247
	} breadcrumbs;

248 249 250 251 252 253 254
	/*
	 * 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;

255
	struct intel_hw_status_page status_page;
256
	struct i915_ctx_workarounds wa_ctx;
257
	struct i915_vma *scratch;
258

259 260
	u32             irq_keep_mask; /* always keep these interrupts */
	u32		irq_enable_mask; /* bitmask to enable ring interrupt */
261 262
	void		(*irq_enable)(struct intel_engine_cs *engine);
	void		(*irq_disable)(struct intel_engine_cs *engine);
263

264
	int		(*init_hw)(struct intel_engine_cs *engine);
265 266
	void		(*reset_hw)(struct intel_engine_cs *engine,
				    struct drm_i915_gem_request *req);
267

268 269
	void		(*set_default_submission)(struct intel_engine_cs *engine);

270 271
	struct intel_ring *(*context_pin)(struct intel_engine_cs *engine,
					  struct i915_gem_context *ctx);
272 273
	void		(*context_unpin)(struct intel_engine_cs *engine,
					 struct i915_gem_context *ctx);
274
	int		(*request_alloc)(struct drm_i915_gem_request *req);
275
	int		(*init_context)(struct drm_i915_gem_request *req);
276

277 278 279 280 281 282 283 284 285 286 287
	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)
C
Chris Wilson 已提交
288
	void		(*emit_breadcrumb)(struct drm_i915_gem_request *req,
289
					   u32 *cs);
290
	int		emit_breadcrumb_sz;
291 292 293 294 295 296 297

	/* 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.
	 */
298
	void		(*submit_request)(struct drm_i915_gem_request *req);
299

300 301 302 303 304 305 306 307 308
	/* Call when the priority on a request has changed and it and its
	 * dependencies may need rescheduling. Note the request itself may
	 * not be ready to run!
	 *
	 * Called under the struct_mutex.
	 */
	void		(*schedule)(struct drm_i915_gem_request *request,
				    int priority);

309 310 311 312 313 314
	/* 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.
	 */
315 316
	void		(*irq_seqno_barrier)(struct intel_engine_cs *engine);
	void		(*cleanup)(struct intel_engine_cs *engine);
317

318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354
	/* 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)
	 */
355
	struct {
356
		union {
357 358 359
#define GEN6_SEMAPHORE_LAST	VECS_HW
#define GEN6_NUM_SEMAPHORES	(GEN6_SEMAPHORE_LAST + 1)
#define GEN6_SEMAPHORES_MASK	GENMASK(GEN6_SEMAPHORE_LAST, 0)
360 361
			struct {
				/* our mbox written by others */
362
				u32		wait[GEN6_NUM_SEMAPHORES];
363
				/* mboxes this ring signals to */
364
				i915_reg_t	signal[GEN6_NUM_SEMAPHORES];
365
			} mbox;
366
			u64		signal_ggtt[I915_NUM_ENGINES];
367
		};
368 369

		/* AKA wait() */
370 371
		int	(*sync_to)(struct drm_i915_gem_request *req,
				   struct drm_i915_gem_request *signal);
372
		u32	*(*signal)(struct drm_i915_gem_request *req, u32 *cs);
373
	} semaphore;
374

375
	/* Execlists */
376
	struct tasklet_struct irq_tasklet;
377 378
	struct i915_priolist default_priolist;
	bool no_priolist;
379
	struct execlist_port {
380 381 382 383 384 385 386 387 388
		struct drm_i915_gem_request *request_count;
#define EXECLIST_COUNT_BITS 2
#define port_request(p) ptr_mask_bits((p)->request_count, EXECLIST_COUNT_BITS)
#define port_count(p) ptr_unmask_bits((p)->request_count, EXECLIST_COUNT_BITS)
#define port_pack(rq, count) ptr_pack_bits(rq, count, EXECLIST_COUNT_BITS)
#define port_unpack(p, count) ptr_unpack_bits((p)->request_count, count, EXECLIST_COUNT_BITS)
#define port_set(p, packed) ((p)->request_count = (packed))
#define port_isset(p) ((p)->request_count)
#define port_index(p, e) ((p) - (e)->execlist_port)
389
		GEM_DEBUG_DECL(u32 context_id);
390
	} execlist_port[2];
391 392
	struct rb_root execlist_queue;
	struct rb_node *execlist_first;
393
	unsigned int fw_domains;
394

395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412
	/* Contexts are pinned whilst they are active on the GPU. The last
	 * context executed remains active whilst the GPU is idle - the
	 * switch away and write to the context object only occurs on the
	 * next execution.  Contexts are only unpinned on retirement of the
	 * following request ensuring that we can always write to the object
	 * on the context switch even after idling. Across suspend, we switch
	 * to the kernel context and trash it as the save may not happen
	 * before the hardware is powered down.
	 */
	struct i915_gem_context *last_retired_context;

	/* We track the current MI_SET_CONTEXT in order to eliminate
	 * redudant context switches. This presumes that requests are not
	 * reordered! Or when they are the tracking is updated along with
	 * the emission of individual requests into the legacy command
	 * stream (ring).
	 */
	struct i915_gem_context *legacy_active_context;
413

414 415 416
	/* status_notifier: list of callbacks for context-switch changes */
	struct atomic_notifier_head context_status_notifier;

417
	struct intel_engine_hangcheck hangcheck;
418

419 420
	bool needs_cmd_parser;

421
	/*
422
	 * Table of commands the command parser needs to know about
423
	 * for this engine.
424
	 */
425
	DECLARE_HASHTABLE(cmd_hash, I915_CMD_HASH_ORDER);
426 427 428 429

	/*
	 * Table of registers allowed in commands that read/write registers.
	 */
430 431
	const struct drm_i915_reg_table *reg_tables;
	int reg_table_count;
432 433 434 435 436

	/*
	 * Returns the bitmask for the length field of the specified command.
	 * Return 0 for an unrecognized/invalid command.
	 *
437
	 * If the command parser finds an entry for a command in the engine's
438
	 * cmd_tables, it gets the command's length based on the table entry.
439 440 441
	 * 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).
442 443
	 */
	u32 (*get_cmd_length_mask)(u32 cmd_header);
444 445
};

446
static inline unsigned int
447
intel_engine_flag(const struct intel_engine_cs *engine)
448
{
449
	return BIT(engine->id);
450 451
}

452
static inline u32
453
intel_read_status_page(struct intel_engine_cs *engine, int reg)
454
{
455
	/* Ensure that the compiler doesn't optimize away the load. */
456
	return READ_ONCE(engine->status_page.page_addr[reg]);
457 458
}

M
Mika Kuoppala 已提交
459
static inline void
460
intel_write_status_page(struct intel_engine_cs *engine, int reg, u32 value)
M
Mika Kuoppala 已提交
461
{
462 463 464 465 466 467 468 469 470 471 472 473 474 475
	/* Writing into the status page should be done sparingly. Since
	 * we do when we are uncertain of the device state, we take a bit
	 * of extra paranoia to try and ensure that the HWS takes the value
	 * we give and that it doesn't end up trapped inside the CPU!
	 */
	if (static_cpu_has(X86_FEATURE_CLFLUSH)) {
		mb();
		clflush(&engine->status_page.page_addr[reg]);
		engine->status_page.page_addr[reg] = value;
		clflush(&engine->status_page.page_addr[reg]);
		mb();
	} else {
		WRITE_ONCE(engine->status_page.page_addr[reg], value);
	}
M
Mika Kuoppala 已提交
476 477
}

478
/*
C
Chris Wilson 已提交
479 480 481 482 483 484 485 486 487 488 489
 * 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)
490
 * 0x20-0x2f: Reserved (Gen6+)
C
Chris Wilson 已提交
491
 *
492
 * The area from dword 0x30 to 0x3ff is available for driver usage.
C
Chris Wilson 已提交
493
 */
494
#define I915_GEM_HWS_INDEX		0x30
495
#define I915_GEM_HWS_INDEX_ADDR (I915_GEM_HWS_INDEX << MI_STORE_DWORD_INDEX_SHIFT)
496
#define I915_GEM_HWS_SCRATCH_INDEX	0x40
497
#define I915_GEM_HWS_SCRATCH_ADDR (I915_GEM_HWS_SCRATCH_INDEX << MI_STORE_DWORD_INDEX_SHIFT)
C
Chris Wilson 已提交
498

499 500
struct intel_ring *
intel_engine_create_ring(struct intel_engine_cs *engine, int size);
501 502 503
int intel_ring_pin(struct intel_ring *ring,
		   struct drm_i915_private *i915,
		   unsigned int offset_bias);
504
void intel_ring_reset(struct intel_ring *ring, u32 tail);
505
unsigned int intel_ring_update_space(struct intel_ring *ring);
506
void intel_ring_unpin(struct intel_ring *ring);
507
void intel_ring_free(struct intel_ring *ring);
508

509 510
void intel_engine_stop(struct intel_engine_cs *engine);
void intel_engine_cleanup(struct intel_engine_cs *engine);
511

512 513
void intel_legacy_submission_resume(struct drm_i915_private *dev_priv);

514
int __must_check intel_ring_cacheline_align(struct drm_i915_gem_request *req);
515

516 517
u32 __must_check *intel_ring_begin(struct drm_i915_gem_request *req,
				   unsigned int n);
518

519 520
static inline void
intel_ring_advance(struct drm_i915_gem_request *req, u32 *cs)
521
{
522 523 524 525 526 527 528
	/* 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()).
529
	 */
530
	GEM_BUG_ON((req->ring->vaddr + req->ring->emit) != cs);
531 532
}

533
static inline u32
534 535 536 537 538 539 540
intel_ring_wrap(const struct intel_ring *ring, u32 pos)
{
	return pos & (ring->size - 1);
}

static inline u32
intel_ring_offset(const struct drm_i915_gem_request *req, void *addr)
541 542
{
	/* Don't write ring->size (equivalent to 0) as that hangs some GPUs. */
543 544
	u32 offset = addr - req->ring->vaddr;
	GEM_BUG_ON(offset > req->ring->size);
545
	return intel_ring_wrap(req->ring, offset);
546
}
547

548 549 550 551 552 553 554 555 556
static inline void
assert_ring_tail_valid(const struct intel_ring *ring, unsigned int tail)
{
	/* We could combine these into a single tail operation, but keeping
	 * them as seperate tests will help identify the cause should one
	 * ever fire.
	 */
	GEM_BUG_ON(!IS_ALIGNED(tail, 8));
	GEM_BUG_ON(tail >= ring->size);
557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575

	/*
	 * "Ring Buffer Use"
	 *	Gen2 BSpec "1. Programming Environment" / 1.4.4.6
	 *	Gen3 BSpec "1c Memory Interface Functions" / 2.3.4.5
	 *	Gen4+ BSpec "1c Memory Interface and Command Stream" / 5.3.4.5
	 * "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."
	 *
	 * We use ring->head as the last known location of the actual RING_HEAD,
	 * it may have advanced but in the worst case it is equally the same
	 * as ring->head and so we should never program RING_TAIL to advance
	 * into the same cacheline as ring->head.
	 */
#define cacheline(a) round_down(a, CACHELINE_BYTES)
	GEM_BUG_ON(cacheline(tail) == cacheline(ring->head) &&
		   tail < ring->head);
#undef cacheline
576 577
}

578 579 580 581 582 583 584 585 586 587 588 589 590
static inline unsigned int
intel_ring_set_tail(struct intel_ring *ring, unsigned int tail)
{
	/* Whilst writes to the tail are strictly order, there is no
	 * serialisation between readers and the writers. The tail may be
	 * read by i915_gem_request_retire() just as it is being updated
	 * by execlists, as although the breadcrumb is complete, the context
	 * switch hasn't been seen.
	 */
	assert_ring_tail_valid(ring, tail);
	ring->tail = tail;
	return tail;
}
591

592
void intel_engine_init_global_seqno(struct intel_engine_cs *engine, u32 seqno);
593

594 595
void intel_engine_setup_common(struct intel_engine_cs *engine);
int intel_engine_init_common(struct intel_engine_cs *engine);
596
int intel_engine_create_scratch(struct intel_engine_cs *engine, int size);
597
void intel_engine_cleanup_common(struct intel_engine_cs *engine);
598

599 600 601 602
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_blt_ring_buffer(struct intel_engine_cs *engine);
int intel_init_vebox_ring_buffer(struct intel_engine_cs *engine);
603

604
u64 intel_engine_get_active_head(struct intel_engine_cs *engine);
605 606
u64 intel_engine_get_last_batch_head(struct intel_engine_cs *engine);

607 608 609 610
static inline u32 intel_engine_get_seqno(struct intel_engine_cs *engine)
{
	return intel_read_status_page(engine, I915_GEM_HWS_INDEX);
}
611

612 613 614 615 616 617 618 619 620
static inline u32 intel_engine_last_submit(struct intel_engine_cs *engine)
{
	/* We are only peeking at the tail of the submit queue (and not the
	 * queue itself) in order to gain a hint as to the current active
	 * state of the engine. Callers are not expected to be taking
	 * engine->timeline->lock, nor are they expected to be concerned
	 * wtih serialising this hint with anything, so document it as
	 * a hint and nothing more.
	 */
621
	return READ_ONCE(engine->timeline->seqno);
622 623
}

624
int init_workarounds_ring(struct intel_engine_cs *engine);
625
int intel_ring_workarounds_emit(struct drm_i915_gem_request *req);
626

627 628 629
void intel_engine_get_instdone(struct intel_engine_cs *engine,
			       struct intel_instdone *instdone);

630 631 632
/*
 * Arbitrary size for largest possible 'add request' sequence. The code paths
 * are complex and variable. Empirical measurement shows that the worst case
633 634 635
 * 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).
636
 */
637
#define MIN_SPACE_FOR_ADD_REQUEST 336
638

639 640
static inline u32 intel_hws_seqno_address(struct intel_engine_cs *engine)
{
641
	return engine->status_page.ggtt_offset + I915_GEM_HWS_INDEX_ADDR;
642 643
}

644 645 646
/* intel_breadcrumbs.c -- user interrupt bottom-half for waiters */
int intel_engine_init_breadcrumbs(struct intel_engine_cs *engine);

647 648
static inline void intel_wait_init(struct intel_wait *wait,
				   struct drm_i915_gem_request *rq)
649 650
{
	wait->tsk = current;
651
	wait->request = rq;
652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667
}

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

static inline bool intel_wait_has_seqno(const struct intel_wait *wait)
{
	return wait->seqno;
}

static inline bool
intel_wait_update_seqno(struct intel_wait *wait, u32 seqno)
{
668
	wait->seqno = seqno;
669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689
	return intel_wait_has_seqno(wait);
}

static inline bool
intel_wait_update_request(struct intel_wait *wait,
			  const struct drm_i915_gem_request *rq)
{
	return intel_wait_update_seqno(wait, i915_gem_request_global_seqno(rq));
}

static inline bool
intel_wait_check_seqno(const struct intel_wait *wait, u32 seqno)
{
	return wait->seqno == seqno;
}

static inline bool
intel_wait_check_request(const struct intel_wait *wait,
			 const struct drm_i915_gem_request *rq)
{
	return intel_wait_check_seqno(wait, i915_gem_request_global_seqno(rq));
690 691 692 693 694 695 696 697 698 699 700
}

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);
701 702
void intel_engine_enable_signaling(struct drm_i915_gem_request *request,
				   bool wakeup);
703
void intel_engine_cancel_signaling(struct drm_i915_gem_request *request);
704

705
static inline bool intel_engine_has_waiter(const struct intel_engine_cs *engine)
706
{
707
	return READ_ONCE(engine->breadcrumbs.irq_wait);
708 709
}

710 711
unsigned int intel_engine_wakeup(struct intel_engine_cs *engine);
#define ENGINE_WAKEUP_WAITER BIT(0)
712 713 714 715
#define ENGINE_WAKEUP_ASLEEP BIT(1)

void __intel_engine_disarm_breadcrumbs(struct intel_engine_cs *engine);
void intel_engine_disarm_breadcrumbs(struct intel_engine_cs *engine);
716

717
void intel_engine_reset_breadcrumbs(struct intel_engine_cs *engine);
718
void intel_engine_fini_breadcrumbs(struct intel_engine_cs *engine);
719
bool intel_breadcrumbs_busy(struct intel_engine_cs *engine);
720

721 722 723 724 725 726 727 728 729 730 731
static inline u32 *gen8_emit_pipe_control(u32 *batch, u32 flags, u32 offset)
{
	memset(batch, 0, 6 * sizeof(u32));

	batch[0] = GFX_OP_PIPE_CONTROL(6);
	batch[1] = flags;
	batch[2] = offset;

	return batch + 6;
}

732
bool intel_engine_is_idle(struct intel_engine_cs *engine);
733
bool intel_engines_are_idle(struct drm_i915_private *dev_priv);
734

735
void intel_engines_mark_idle(struct drm_i915_private *i915);
736 737
void intel_engines_reset_default_submission(struct drm_i915_private *i915);

738
bool intel_engine_can_store_dword(struct intel_engine_cs *engine);
739

740
#endif /* _INTEL_RINGBUFFER_H_ */