intel_breadcrumbs.c 10.6 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
/*
 * Copyright © 2015 Intel Corporation
 *
 * Permission is hereby granted, free of charge, to any person obtaining a
 * copy of this software and associated documentation files (the "Software"),
 * to deal in the Software without restriction, including without limitation
 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
 * and/or sell copies of the Software, and to permit persons to whom the
 * Software is furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice (including the next
 * paragraph) shall be included in all copies or substantial portions of the
 * Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
 * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
 * IN THE SOFTWARE.
 *
 */

25
#include <linux/kthread.h>
26
#include <trace/events/dma_fence.h>
27
#include <uapi/linux/sched/types.h>
28

29 30
#include "i915_drv.h"

31 32 33 34 35 36 37 38 39 40 41 42
static void irq_enable(struct intel_engine_cs *engine)
{
	if (!engine->irq_enable)
		return;

	/* Caller disables interrupts */
	spin_lock(&engine->i915->irq_lock);
	engine->irq_enable(engine);
	spin_unlock(&engine->i915->irq_lock);
}

static void irq_disable(struct intel_engine_cs *engine)
43
{
44 45 46 47 48 49 50 51
	if (!engine->irq_disable)
		return;

	/* Caller disables interrupts */
	spin_lock(&engine->i915->irq_lock);
	engine->irq_disable(engine);
	spin_unlock(&engine->i915->irq_lock);
}
52

53 54
static void __intel_breadcrumbs_disarm_irq(struct intel_breadcrumbs *b)
{
55 56
	lockdep_assert_held(&b->irq_lock);

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
	GEM_BUG_ON(!b->irq_enabled);
	if (!--b->irq_enabled)
		irq_disable(container_of(b,
					 struct intel_engine_cs,
					 breadcrumbs));

	b->irq_armed = false;
}

void intel_engine_disarm_breadcrumbs(struct intel_engine_cs *engine)
{
	struct intel_breadcrumbs *b = &engine->breadcrumbs;

	if (!b->irq_armed)
		return;

	spin_lock_irq(&b->irq_lock);
	if (b->irq_armed)
		__intel_breadcrumbs_disarm_irq(b);
	spin_unlock_irq(&b->irq_lock);
}

static inline bool __request_completed(const struct i915_request *rq)
{
	return i915_seqno_passed(__hwsp_seqno(rq), rq->fence.seqno);
}

84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99
__maybe_unused static bool
check_signal_order(struct intel_context *ce, struct i915_request *rq)
{
	if (!list_is_last(&rq->signal_link, &ce->signals) &&
	    i915_seqno_passed(rq->fence.seqno,
			      list_next_entry(rq, signal_link)->fence.seqno))
		return false;

	if (!list_is_first(&rq->signal_link, &ce->signals) &&
	    i915_seqno_passed(list_prev_entry(rq, signal_link)->fence.seqno,
			      rq->fence.seqno))
		return false;

	return true;
}

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
static bool
__dma_fence_signal(struct dma_fence *fence)
{
	return !test_and_set_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &fence->flags);
}

static void
__dma_fence_signal__timestamp(struct dma_fence *fence, ktime_t timestamp)
{
	fence->timestamp = timestamp;
	set_bit(DMA_FENCE_FLAG_TIMESTAMP_BIT, &fence->flags);
	trace_dma_fence_signaled(fence);
}

static void
__dma_fence_signal__notify(struct dma_fence *fence)
{
	struct dma_fence_cb *cur, *tmp;

	lockdep_assert_held(fence->lock);
	lockdep_assert_irqs_disabled();

	list_for_each_entry_safe(cur, tmp, &fence->cb_list, node) {
		INIT_LIST_HEAD(&cur->node);
		cur->func(fence, cur);
	}
	INIT_LIST_HEAD(&fence->cb_list);
}

129
void intel_engine_breadcrumbs_irq(struct intel_engine_cs *engine)
130 131
{
	struct intel_breadcrumbs *b = &engine->breadcrumbs;
132
	const ktime_t timestamp = ktime_get();
133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148
	struct intel_context *ce, *cn;
	struct list_head *pos, *next;
	LIST_HEAD(signal);

	spin_lock(&b->irq_lock);

	if (b->irq_armed && list_empty(&b->signalers))
		__intel_breadcrumbs_disarm_irq(b);

	list_for_each_entry_safe(ce, cn, &b->signalers, signal_link) {
		GEM_BUG_ON(list_empty(&ce->signals));

		list_for_each_safe(pos, next, &ce->signals) {
			struct i915_request *rq =
				list_entry(pos, typeof(*rq), signal_link);

149 150
			GEM_BUG_ON(!check_signal_order(ce, rq));

151 152 153 154 155
			if (!__request_completed(rq))
				break;

			GEM_BUG_ON(!test_bit(I915_FENCE_FLAG_SIGNAL,
					     &rq->fence.flags));
156 157 158 159
			clear_bit(I915_FENCE_FLAG_SIGNAL, &rq->fence.flags);

			if (!__dma_fence_signal(&rq->fence))
				continue;
160 161 162 163 164 165 166 167 168 169

			/*
			 * Queue for execution after dropping the signaling
			 * spinlock as the callback chain may end up adding
			 * more signalers to the same context or engine.
			 */
			i915_request_get(rq);
			list_add_tail(&rq->signal_link, &signal);
		}

170
		/*
171 172 173
		 * We process the list deletion in bulk, only using a list_add
		 * (not list_move) above but keeping the status of
		 * rq->signal_link known with the I915_FENCE_FLAG_SIGNAL bit.
174
		 */
175 176 177 178 179 180 181 182 183 184 185 186 187
		if (!list_is_first(pos, &ce->signals)) {
			/* Advance the list to the first incomplete request */
			__list_del_many(&ce->signals, pos);
			if (&ce->signals == pos) /* now empty */
				list_del_init(&ce->signal_link);
		}
	}

	spin_unlock(&b->irq_lock);

	list_for_each_safe(pos, next, &signal) {
		struct i915_request *rq =
			list_entry(pos, typeof(*rq), signal_link);
188

189 190 191 192 193 194
		__dma_fence_signal__timestamp(&rq->fence, timestamp);

		spin_lock(&rq->lock);
		__dma_fence_signal__notify(&rq->fence);
		spin_unlock(&rq->lock);

195
		i915_request_put(rq);
196
	}
197 198
}

199
void intel_engine_signal_breadcrumbs(struct intel_engine_cs *engine)
200
{
201
	local_irq_disable();
202
	intel_engine_breadcrumbs_irq(engine);
203
	local_irq_enable();
204 205
}

206 207 208 209 210 211 212 213
static void signal_irq_work(struct irq_work *work)
{
	struct intel_engine_cs *engine =
		container_of(work, typeof(*engine), breadcrumbs.irq_work);

	intel_engine_breadcrumbs_irq(engine);
}

214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235
void intel_engine_pin_breadcrumbs_irq(struct intel_engine_cs *engine)
{
	struct intel_breadcrumbs *b = &engine->breadcrumbs;

	spin_lock_irq(&b->irq_lock);
	if (!b->irq_enabled++)
		irq_enable(engine);
	GEM_BUG_ON(!b->irq_enabled); /* no overflow! */
	spin_unlock_irq(&b->irq_lock);
}

void intel_engine_unpin_breadcrumbs_irq(struct intel_engine_cs *engine)
{
	struct intel_breadcrumbs *b = &engine->breadcrumbs;

	spin_lock_irq(&b->irq_lock);
	GEM_BUG_ON(!b->irq_enabled); /* no underflow! */
	if (!--b->irq_enabled)
		irq_disable(engine);
	spin_unlock_irq(&b->irq_lock);
}

236
static void __intel_breadcrumbs_arm_irq(struct intel_breadcrumbs *b)
237 238 239
{
	struct intel_engine_cs *engine =
		container_of(b, struct intel_engine_cs, breadcrumbs);
240

241
	lockdep_assert_held(&b->irq_lock);
242
	if (b->irq_armed)
243
		return;
244

245 246
	/*
	 * The breadcrumb irq will be disarmed on the interrupt after the
247 248 249 250 251 252
	 * waiters are signaled. This gives us a single interrupt window in
	 * which we can add a new waiter and avoid the cost of re-enabling
	 * the irq.
	 */
	b->irq_armed = true;

253 254
	/*
	 * Since we are waiting on a request, the GPU should be busy
255 256 257 258
	 * and should have its own rpm reference. This is tracked
	 * by i915->gt.awake, we can forgo holding our own wakref
	 * for the interrupt as before i915->gt.awake is released (when
	 * the driver is idle) we disarm the breadcrumbs.
259 260
	 */

261
	if (!b->irq_enabled++)
262
		irq_enable(engine);
263 264
}

265
void intel_engine_init_breadcrumbs(struct intel_engine_cs *engine)
266
{
267
	struct intel_breadcrumbs *b = &engine->breadcrumbs;
268

269 270
	spin_lock_init(&b->irq_lock);
	INIT_LIST_HEAD(&b->signalers);
271

272
	init_irq_work(&b->irq_work, signal_irq_work);
273 274
}

275
void intel_engine_reset_breadcrumbs(struct intel_engine_cs *engine)
276 277
{
	struct intel_breadcrumbs *b = &engine->breadcrumbs;
278
	unsigned long flags;
279

280
	spin_lock_irqsave(&b->irq_lock, flags);
281

282 283
	if (b->irq_enabled)
		irq_enable(engine);
284
	else
285
		irq_disable(engine);
286

287
	spin_unlock_irqrestore(&b->irq_lock, flags);
288 289
}

290
void intel_engine_fini_breadcrumbs(struct intel_engine_cs *engine)
291
{
292 293
}

294
bool i915_request_enable_breadcrumb(struct i915_request *rq)
295
{
296 297
	lockdep_assert_held(&rq->lock);
	lockdep_assert_irqs_disabled();
298

299 300
	if (test_bit(I915_FENCE_FLAG_ACTIVE, &rq->fence.flags)) {
		struct intel_breadcrumbs *b = &rq->engine->breadcrumbs;
301 302
		struct intel_context *ce = rq->hw_context;
		struct list_head *pos;
303

304 305 306
		spin_lock(&b->irq_lock);
		GEM_BUG_ON(test_bit(I915_FENCE_FLAG_SIGNAL, &rq->fence.flags));

307
		__intel_breadcrumbs_arm_irq(b);
308

309
		/*
310 311 312 313 314 315 316 317 318 319 320 321
		 * We keep the seqno in retirement order, so we can break
		 * inside intel_engine_breadcrumbs_irq as soon as we've passed
		 * the last completed request (or seen a request that hasn't
		 * event started). We could iterate the timeline->requests list,
		 * but keeping a separate signalers_list has the advantage of
		 * hopefully being much smaller than the full list and so
		 * provides faster iteration and detection when there are no
		 * more interrupts required for this context.
		 *
		 * We typically expect to add new signalers in order, so we
		 * start looking for our insertion point from the tail of
		 * the list.
322
		 */
323 324 325
		list_for_each_prev(pos, &ce->signals) {
			struct i915_request *it =
				list_entry(pos, typeof(*it), signal_link);
326

327
			if (i915_seqno_passed(rq->fence.seqno, it->fence.seqno))
328 329
				break;
		}
330 331 332
		list_add(&rq->signal_link, pos);
		if (pos == &ce->signals) /* catch transitions from empty list */
			list_move_tail(&ce->signal_link, &b->signalers);
333
		GEM_BUG_ON(!check_signal_order(ce, rq));
334

335
		set_bit(I915_FENCE_FLAG_SIGNAL, &rq->fence.flags);
336
		spin_unlock(&b->irq_lock);
337 338
	}

339
	return !__request_completed(rq);
340 341
}

342
void i915_request_cancel_breadcrumb(struct i915_request *rq)
343
{
344
	struct intel_breadcrumbs *b = &rq->engine->breadcrumbs;
345

346 347
	lockdep_assert_held(&rq->lock);
	lockdep_assert_irqs_disabled();
348

349 350 351 352 353 354
	/*
	 * We must wait for b->irq_lock so that we know the interrupt handler
	 * has released its reference to the intel_context and has completed
	 * the DMA_FENCE_FLAG_SIGNALED_BIT/I915_FENCE_FLAG_SIGNAL dance (if
	 * required).
	 */
355 356 357
	spin_lock(&b->irq_lock);
	if (test_bit(I915_FENCE_FLAG_SIGNAL, &rq->fence.flags)) {
		struct intel_context *ce = rq->hw_context;
358

359 360 361
		list_del(&rq->signal_link);
		if (list_empty(&ce->signals))
			list_del_init(&ce->signal_link);
362

363 364 365
		clear_bit(I915_FENCE_FLAG_SIGNAL, &rq->fence.flags);
	}
	spin_unlock(&b->irq_lock);
366 367
}

368 369
void intel_engine_print_breadcrumbs(struct intel_engine_cs *engine,
				    struct drm_printer *p)
370 371
{
	struct intel_breadcrumbs *b = &engine->breadcrumbs;
372 373
	struct intel_context *ce;
	struct i915_request *rq;
374

375 376
	if (list_empty(&b->signalers))
		return;
377

378
	drm_printf(p, "Signals:\n");
379

380 381 382 383 384 385 386 387 388 389 390 391
	spin_lock_irq(&b->irq_lock);
	list_for_each_entry(ce, &b->signalers, signal_link) {
		list_for_each_entry(rq, &ce->signals, signal_link) {
			drm_printf(p, "\t[%llx:%llx%s] @ %dms\n",
				   rq->fence.context, rq->fence.seqno,
				   i915_request_completed(rq) ? "!" :
				   i915_request_started(rq) ? "*" :
				   "",
				   jiffies_to_msecs(jiffies - rq->emitted_jiffies));
		}
	}
	spin_unlock_irq(&b->irq_lock);
392
}