summaryrefslogtreecommitdiff
path: root/drivers/gpu/drm/i915/gt/intel_engine.h
blob: d3c6993f4f46f8c71f89d14759ded4f10ca23e11 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
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
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
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
311
312
313
314
315
316
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
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
/* SPDX-License-Identifier: MIT */
#ifndef _INTEL_RINGBUFFER_H_
#define _INTEL_RINGBUFFER_H_

#include <drm/drm_util.h>

#include <linux/hashtable.h>
#include <linux/irq_work.h>
#include <linux/random.h>
#include <linux/seqlock.h>

#include "i915_pmu.h"
#include "i915_reg.h"
#include "i915_request.h"
#include "i915_selftest.h"
#include "gt/intel_timeline.h"
#include "intel_engine_types.h"
#include "intel_gpu_commands.h"
#include "intel_workarounds.h"

struct drm_printer;

/* 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
#define CACHELINE_DWORDS (CACHELINE_BYTES / sizeof(u32))

/*
 * The register defines to be used with the following macros need to accept a
 * base param, e.g:
 *
 * REG_FOO(base) _MMIO((base) + <relative offset>)
 * ENGINE_READ(engine, REG_FOO);
 *
 * register arrays are to be defined and accessed as follows:
 *
 * REG_BAR(base, i) _MMIO((base) + <relative offset> + (i) * <shift>)
 * ENGINE_READ_IDX(engine, REG_BAR, i)
 */

#define __ENGINE_REG_OP(op__, engine__, ...) \
	intel_uncore_##op__((engine__)->uncore, __VA_ARGS__)

#define __ENGINE_READ_OP(op__, engine__, reg__) \
	__ENGINE_REG_OP(op__, (engine__), reg__((engine__)->mmio_base))

#define ENGINE_READ16(...)	__ENGINE_READ_OP(read16, __VA_ARGS__)
#define ENGINE_READ(...)	__ENGINE_READ_OP(read, __VA_ARGS__)
#define ENGINE_READ_FW(...)	__ENGINE_READ_OP(read_fw, __VA_ARGS__)
#define ENGINE_POSTING_READ(...) __ENGINE_READ_OP(posting_read_fw, __VA_ARGS__)
#define ENGINE_POSTING_READ16(...) __ENGINE_READ_OP(posting_read16, __VA_ARGS__)

#define ENGINE_READ64(engine__, lower_reg__, upper_reg__) \
	__ENGINE_REG_OP(read64_2x32, (engine__), \
			lower_reg__((engine__)->mmio_base), \
			upper_reg__((engine__)->mmio_base))

#define ENGINE_READ_IDX(engine__, reg__, idx__) \
	__ENGINE_REG_OP(read, (engine__), reg__((engine__)->mmio_base, (idx__)))

#define __ENGINE_WRITE_OP(op__, engine__, reg__, val__) \
	__ENGINE_REG_OP(op__, (engine__), reg__((engine__)->mmio_base), (val__))

#define ENGINE_WRITE16(...)	__ENGINE_WRITE_OP(write16, __VA_ARGS__)
#define ENGINE_WRITE(...)	__ENGINE_WRITE_OP(write, __VA_ARGS__)
#define ENGINE_WRITE_FW(...)	__ENGINE_WRITE_OP(write_fw, __VA_ARGS__)

#define GEN6_RING_FAULT_REG_READ(engine__) \
	intel_uncore_read((engine__)->uncore, RING_FAULT_REG(engine__))

#define GEN6_RING_FAULT_REG_POSTING_READ(engine__) \
	intel_uncore_posting_read((engine__)->uncore, RING_FAULT_REG(engine__))

#define GEN6_RING_FAULT_REG_RMW(engine__, clear__, set__) \
({ \
	u32 __val; \
\
	__val = intel_uncore_read((engine__)->uncore, \
				  RING_FAULT_REG(engine__)); \
	__val &= ~(clear__); \
	__val |= (set__); \
	intel_uncore_write((engine__)->uncore, RING_FAULT_REG(engine__), \
			   __val); \
})

/* 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.
 */
enum intel_engine_hangcheck_action {
	ENGINE_IDLE = 0,
	ENGINE_WAIT,
	ENGINE_ACTIVE_SEQNO,
	ENGINE_ACTIVE_HEAD,
	ENGINE_ACTIVE_SUBUNITS,
	ENGINE_WAIT_KICK,
	ENGINE_DEAD,
};

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";
}

static inline unsigned int
execlists_num_ports(const struct intel_engine_execlists * const execlists)
{
	return execlists->port_mask + 1;
}

static inline struct i915_request *
execlists_active(const struct intel_engine_execlists *execlists)
{
	GEM_BUG_ON(execlists->active - execlists->inflight >
		   execlists_num_ports(execlists));
	return READ_ONCE(*execlists->active);
}

struct i915_request *
execlists_unwind_incomplete_requests(struct intel_engine_execlists *execlists);

static inline u32
intel_read_status_page(const struct intel_engine_cs *engine, int reg)
{
	/* Ensure that the compiler doesn't optimize away the load. */
	return READ_ONCE(engine->status_page.addr[reg]);
}

static inline void
intel_write_status_page(struct intel_engine_cs *engine, int reg, u32 value)
{
	/* 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.addr[reg]);
		engine->status_page.addr[reg] = value;
		clflush(&engine->status_page.addr[reg]);
		mb();
	} else {
		WRITE_ONCE(engine->status_page.addr[reg], value);
	}
}

/*
 * 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)
 * 0x20-0x2f: Reserved (Gen6+)
 *
 * The area from dword 0x30 to 0x3ff is available for driver usage.
 */
#define I915_GEM_HWS_PREEMPT		0x32
#define I915_GEM_HWS_PREEMPT_ADDR	(I915_GEM_HWS_PREEMPT * sizeof(u32))
#define I915_GEM_HWS_SEQNO		0x40
#define I915_GEM_HWS_SEQNO_ADDR		(I915_GEM_HWS_SEQNO * sizeof(u32))
#define I915_GEM_HWS_SCRATCH		0x80
#define I915_GEM_HWS_SCRATCH_ADDR	(I915_GEM_HWS_SCRATCH * sizeof(u32))

#define I915_HWS_CSB_BUF0_INDEX		0x10
#define I915_HWS_CSB_WRITE_INDEX	0x1f
#define CNL_HWS_CSB_WRITE_INDEX		0x2f

struct intel_ring *
intel_engine_create_ring(struct intel_engine_cs *engine, int size);
int intel_ring_pin(struct intel_ring *ring);
void intel_ring_reset(struct intel_ring *ring, u32 tail);
unsigned int intel_ring_update_space(struct intel_ring *ring);
void intel_ring_unpin(struct intel_ring *ring);
void intel_ring_free(struct kref *ref);

static inline struct intel_ring *intel_ring_get(struct intel_ring *ring)
{
	kref_get(&ring->ref);
	return ring;
}

static inline void intel_ring_put(struct intel_ring *ring)
{
	kref_put(&ring->ref, intel_ring_free);
}

void intel_engine_stop(struct intel_engine_cs *engine);
void intel_engine_cleanup(struct intel_engine_cs *engine);

int __must_check intel_ring_cacheline_align(struct i915_request *rq);

u32 __must_check *intel_ring_begin(struct i915_request *rq, unsigned int n);

static inline void intel_ring_advance(struct i915_request *rq, u32 *cs)
{
	/* 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()).
	 */
	GEM_BUG_ON((rq->ring->vaddr + rq->ring->emit) != cs);
}

static inline u32 intel_ring_wrap(const struct intel_ring *ring, u32 pos)
{
	return pos & (ring->size - 1);
}

static inline bool
intel_ring_offset_valid(const struct intel_ring *ring,
			unsigned int pos)
{
	if (pos & -ring->size) /* must be strictly within the ring */
		return false;

	if (!IS_ALIGNED(pos, 8)) /* must be qword aligned */
		return false;

	return true;
}

static inline u32 intel_ring_offset(const struct i915_request *rq, void *addr)
{
	/* Don't write ring->size (equivalent to 0) as that hangs some GPUs. */
	u32 offset = addr - rq->ring->vaddr;
	GEM_BUG_ON(offset > rq->ring->size);
	return intel_ring_wrap(rq->ring, offset);
}

static inline void
assert_ring_tail_valid(const struct intel_ring *ring, unsigned int tail)
{
	GEM_BUG_ON(!intel_ring_offset_valid(ring, tail));

	/*
	 * "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
}

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_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;
}

static inline unsigned int
__intel_ring_space(unsigned int head, unsigned int tail, unsigned int size)
{
	/*
	 * "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."
	 */
	GEM_BUG_ON(!is_power_of_2(size));
	return (head - tail - CACHELINE_BYTES) & (size - 1);
}

int intel_engines_init_mmio(struct drm_i915_private *i915);
int intel_engines_setup(struct drm_i915_private *i915);
int intel_engines_init(struct drm_i915_private *i915);
void intel_engines_cleanup(struct drm_i915_private *i915);

int intel_engine_init_common(struct intel_engine_cs *engine);
void intel_engine_cleanup_common(struct intel_engine_cs *engine);

int intel_ring_submission_setup(struct intel_engine_cs *engine);
int intel_ring_submission_init(struct intel_engine_cs *engine);

int intel_engine_stop_cs(struct intel_engine_cs *engine);
void intel_engine_cancel_stop_cs(struct intel_engine_cs *engine);

void intel_engine_set_hwsp_writemask(struct intel_engine_cs *engine, u32 mask);

u64 intel_engine_get_active_head(const struct intel_engine_cs *engine);
u64 intel_engine_get_last_batch_head(const struct intel_engine_cs *engine);

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

void intel_engine_init_execlists(struct intel_engine_cs *engine);

void intel_engine_init_breadcrumbs(struct intel_engine_cs *engine);
void intel_engine_fini_breadcrumbs(struct intel_engine_cs *engine);

void intel_engine_signal_breadcrumbs(struct intel_engine_cs *engine);
void intel_engine_disarm_breadcrumbs(struct intel_engine_cs *engine);

static inline void
intel_engine_queue_breadcrumbs(struct intel_engine_cs *engine)
{
	irq_work_queue(&engine->breadcrumbs.irq_work);
}

void intel_engine_breadcrumbs_irq(struct intel_engine_cs *engine);

void intel_engine_reset_breadcrumbs(struct intel_engine_cs *engine);
void intel_engine_fini_breadcrumbs(struct intel_engine_cs *engine);

void intel_engine_print_breadcrumbs(struct intel_engine_cs *engine,
				    struct drm_printer *p);

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;
}

static inline u32 *
gen8_emit_ggtt_write_rcs(u32 *cs, u32 value, u32 gtt_offset, u32 flags)
{
	/* We're using qword write, offset should be aligned to 8 bytes. */
	GEM_BUG_ON(!IS_ALIGNED(gtt_offset, 8));

	/* w/a for post sync ops following a GPGPU operation we
	 * need a prior CS_STALL, which is emitted by the flush
	 * following the batch.
	 */
	*cs++ = GFX_OP_PIPE_CONTROL(6);
	*cs++ = flags | PIPE_CONTROL_QW_WRITE | PIPE_CONTROL_GLOBAL_GTT_IVB;
	*cs++ = gtt_offset;
	*cs++ = 0;
	*cs++ = value;
	/* We're thrashing one dword of HWS. */
	*cs++ = 0;

	return cs;
}

static inline u32 *
gen8_emit_ggtt_write(u32 *cs, u32 value, u32 gtt_offset, u32 flags)
{
	/* w/a: bit 5 needs to be zero for MI_FLUSH_DW address. */
	GEM_BUG_ON(gtt_offset & (1 << 5));
	/* Offset should be aligned to 8 bytes for both (QW/DW) write types */
	GEM_BUG_ON(!IS_ALIGNED(gtt_offset, 8));

	*cs++ = (MI_FLUSH_DW + 1) | MI_FLUSH_DW_OP_STOREDW | flags;
	*cs++ = gtt_offset | MI_FLUSH_DW_USE_GTT;
	*cs++ = 0;
	*cs++ = value;

	return cs;
}

static inline void __intel_engine_reset(struct intel_engine_cs *engine,
					bool stalled)
{
	if (engine->reset.reset)
		engine->reset.reset(engine, stalled);
	engine->serial++; /* contexts lost */
}

bool intel_engine_is_idle(struct intel_engine_cs *engine);
bool intel_engines_are_idle(struct intel_gt *gt);

void intel_engines_reset_default_submission(struct intel_gt *gt);

bool intel_engine_can_store_dword(struct intel_engine_cs *engine);

__printf(3, 4)
void intel_engine_dump(struct intel_engine_cs *engine,
		       struct drm_printer *m,
		       const char *header, ...);

static inline void intel_engine_context_in(struct intel_engine_cs *engine)
{
	unsigned long flags;

	if (READ_ONCE(engine->stats.enabled) == 0)
		return;

	write_seqlock_irqsave(&engine->stats.lock, flags);

	if (engine->stats.enabled > 0) {
		if (engine->stats.active++ == 0)
			engine->stats.start = ktime_get();
		GEM_BUG_ON(engine->stats.active == 0);
	}

	write_sequnlock_irqrestore(&engine->stats.lock, flags);
}

static inline void intel_engine_context_out(struct intel_engine_cs *engine)
{
	unsigned long flags;

	if (READ_ONCE(engine->stats.enabled) == 0)
		return;

	write_seqlock_irqsave(&engine->stats.lock, flags);

	if (engine->stats.enabled > 0) {
		ktime_t last;

		if (engine->stats.active && --engine->stats.active == 0) {
			/*
			 * Decrement the active context count and in case GPU
			 * is now idle add up to the running total.
			 */
			last = ktime_sub(ktime_get(), engine->stats.start);

			engine->stats.total = ktime_add(engine->stats.total,
							last);
		} else if (engine->stats.active == 0) {
			/*
			 * After turning on engine stats, context out might be
			 * the first event in which case we account from the
			 * time stats gathering was turned on.
			 */
			last = ktime_sub(ktime_get(), engine->stats.enabled_at);

			engine->stats.total = ktime_add(engine->stats.total,
							last);
		}
	}

	write_sequnlock_irqrestore(&engine->stats.lock, flags);
}

int intel_enable_engine_stats(struct intel_engine_cs *engine);
void intel_disable_engine_stats(struct intel_engine_cs *engine);

ktime_t intel_engine_get_busy_time(struct intel_engine_cs *engine);

struct i915_request *
intel_engine_find_active_request(struct intel_engine_cs *engine);

u32 intel_engine_context_size(struct drm_i915_private *i915, u8 class);

#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)

static inline bool inject_preempt_hang(struct intel_engine_execlists *execlists)
{
	if (!execlists->preempt_hang.inject_hang)
		return false;

	complete(&execlists->preempt_hang.completion);
	return true;
}

#else

static inline bool inject_preempt_hang(struct intel_engine_execlists *execlists)
{
	return false;
}

#endif

void intel_engine_init_active(struct intel_engine_cs *engine,
			      unsigned int subclass);
#define ENGINE_PHYSICAL	0
#define ENGINE_MOCK	1
#define ENGINE_VIRTUAL	2

#endif /* _INTEL_RINGBUFFER_H_ */