summaryrefslogtreecommitdiffstats
path: root/drivers/gpu/drm/i915/gt/selftest_ring_submission.c
blob: 3350e7c995bccff92751c8b448f81976d0c6a4c3 (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
// SPDX-License-Identifier: MIT
/*
 * Copyright © 2020 Intel Corporation
 */

#include "intel_engine_pm.h"
#include "selftests/igt_flush_test.h"

static struct i915_vma *create_wally(struct intel_engine_cs *engine)
{
	struct drm_i915_gem_object *obj;
	struct i915_vma *vma;
	u32 *cs;
	int err;

	obj = i915_gem_object_create_internal(engine->i915, 4096);
	if (IS_ERR(obj))
		return ERR_CAST(obj);

	vma = i915_vma_instance(obj, engine->gt->vm, NULL);
	if (IS_ERR(vma)) {
		i915_gem_object_put(obj);
		return vma;
	}

	err = i915_vma_pin(vma, 0, 0, PIN_USER | PIN_HIGH);
	if (err) {
		i915_gem_object_put(obj);
		return ERR_PTR(err);
	}

	err = i915_vma_sync(vma);
	if (err) {
		i915_gem_object_put(obj);
		return ERR_PTR(err);
	}

	cs = i915_gem_object_pin_map(obj, I915_MAP_WC);
	if (IS_ERR(cs)) {
		i915_gem_object_put(obj);
		return ERR_CAST(cs);
	}

	if (INTEL_GEN(engine->i915) >= 6) {
		*cs++ = MI_STORE_DWORD_IMM_GEN4;
		*cs++ = 0;
	} else if (INTEL_GEN(engine->i915) >= 4) {
		*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
		*cs++ = 0;
	} else {
		*cs++ = MI_STORE_DWORD_IMM | MI_MEM_VIRTUAL;
	}
	*cs++ = vma->node.start + 4000;
	*cs++ = STACK_MAGIC;

	*cs++ = MI_BATCH_BUFFER_END;

	i915_gem_object_flush_map(obj);
	i915_gem_object_unpin_map(obj);

	vma->private = intel_context_create(engine); /* dummy residuals */
	if (IS_ERR(vma->private)) {
		vma = ERR_CAST(vma->private);
		i915_gem_object_put(obj);
	}

	return vma;
}

static int context_sync(struct intel_context *ce)
{
	struct i915_request *rq;
	int err = 0;

	rq = intel_context_create_request(ce);
	if (IS_ERR(rq))
		return PTR_ERR(rq);

	i915_request_get(rq);
	i915_request_add(rq);

	if (i915_request_wait(rq, 0, HZ / 5) < 0)
		err = -ETIME;
	i915_request_put(rq);

	return err;
}

static int new_context_sync(struct intel_engine_cs *engine)
{
	struct intel_context *ce;
	int err;

	ce = intel_context_create(engine);
	if (IS_ERR(ce))
		return PTR_ERR(ce);

	err = context_sync(ce);
	intel_context_put(ce);

	return err;
}

static int mixed_contexts_sync(struct intel_engine_cs *engine, u32 *result)
{
	int pass;
	int err;

	for (pass = 0; pass < 2; pass++) {
		WRITE_ONCE(*result, 0);
		err = context_sync(engine->kernel_context);
		if (err || READ_ONCE(*result)) {
			if (!err) {
				pr_err("pass[%d] wa_bb emitted for the kernel context\n",
				       pass);
				err = -EINVAL;
			}
			return err;
		}

		WRITE_ONCE(*result, 0);
		err = new_context_sync(engine);
		if (READ_ONCE(*result) != STACK_MAGIC) {
			if (!err) {
				pr_err("pass[%d] wa_bb *NOT* emitted after the kernel context\n",
				       pass);
				err = -EINVAL;
			}
			return err;
		}

		WRITE_ONCE(*result, 0);
		err = new_context_sync(engine);
		if (READ_ONCE(*result) != STACK_MAGIC) {
			if (!err) {
				pr_err("pass[%d] wa_bb *NOT* emitted for the user context switch\n",
				       pass);
				err = -EINVAL;
			}
			return err;
		}
	}

	return 0;
}

static int double_context_sync_00(struct intel_engine_cs *engine, u32 *result)
{
	struct intel_context *ce;
	int err, i;

	ce = intel_context_create(engine);
	if (IS_ERR(ce))
		return PTR_ERR(ce);

	for (i = 0; i < 2; i++) {
		WRITE_ONCE(*result, 0);
		err = context_sync(ce);
		if (err)
			break;
	}
	intel_context_put(ce);
	if (err)
		return err;

	if (READ_ONCE(*result)) {
		pr_err("wa_bb emitted between the same user context\n");
		return -EINVAL;
	}

	return 0;
}

static int kernel_context_sync_00(struct intel_engine_cs *engine, u32 *result)
{
	struct intel_context *ce;
	int err, i;

	ce = intel_context_create(engine);
	if (IS_ERR(ce))
		return PTR_ERR(ce);

	for (i = 0; i < 2; i++) {
		WRITE_ONCE(*result, 0);
		err = context_sync(ce);
		if (err)
			break;

		err = context_sync(engine->kernel_context);
		if (err)
			break;
	}
	intel_context_put(ce);
	if (err)
		return err;

	if (READ_ONCE(*result)) {
		pr_err("wa_bb emitted between the same user context [with intervening kernel]\n");
		return -EINVAL;
	}

	return 0;
}

static int __live_ctx_switch_wa(struct intel_engine_cs *engine)
{
	struct i915_vma *bb;
	u32 *result;
	int err;

	bb = create_wally(engine);
	if (IS_ERR(bb))
		return PTR_ERR(bb);

	result = i915_gem_object_pin_map(bb->obj, I915_MAP_WC);
	if (IS_ERR(result)) {
		intel_context_put(bb->private);
		i915_vma_unpin_and_release(&bb, 0);
		return PTR_ERR(result);
	}
	result += 1000;

	engine->wa_ctx.vma = bb;

	err = mixed_contexts_sync(engine, result);
	if (err)
		goto out;

	err = double_context_sync_00(engine, result);
	if (err)
		goto out;

	err = kernel_context_sync_00(engine, result);
	if (err)
		goto out;

out:
	intel_context_put(engine->wa_ctx.vma->private);
	i915_vma_unpin_and_release(&engine->wa_ctx.vma, I915_VMA_RELEASE_MAP);
	return err;
}

static int live_ctx_switch_wa(void *arg)
{
	struct intel_gt *gt = arg;
	struct intel_engine_cs *engine;
	enum intel_engine_id id;

	/*
	 * Exercise the inter-context wa batch.
	 *
	 * Between each user context we run a wa batch, and since it may
	 * have implications for user visible state, we have to check that
	 * we do actually execute it.
	 *
	 * The trick we use is to replace the normal wa batch with a custom
	 * one that writes to a marker within it, and we can then look for
	 * that marker to confirm if the batch was run when we expect it,
	 * and equally important it was wasn't run when we don't!
	 */

	for_each_engine(engine, gt, id) {
		struct i915_vma *saved_wa;
		int err;

		if (!intel_engine_can_store_dword(engine))
			continue;

		if (IS_GEN_RANGE(gt->i915, 4, 5))
			continue; /* MI_STORE_DWORD is privileged! */

		saved_wa = fetch_and_zero(&engine->wa_ctx.vma);

		intel_engine_pm_get(engine);
		err = __live_ctx_switch_wa(engine);
		intel_engine_pm_put(engine);
		if (igt_flush_test(gt->i915))
			err = -EIO;

		engine->wa_ctx.vma = saved_wa;
		if (err)
			return err;
	}

	return 0;
}

int intel_ring_submission_live_selftests(struct drm_i915_private *i915)
{
	static const struct i915_subtest tests[] = {
		SUBTEST(live_ctx_switch_wa),
	};

	if (HAS_EXECLISTS(i915))
		return 0;

	return intel_gt_live_subtests(tests, &i915->gt);
}