summaryrefslogtreecommitdiffstats
path: root/drivers/gpu/drm/amd/amdkfd/kfd_process.c
blob: 3c76ef05cbcf798483b482ac09b7bc62223d92c5 (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
/*
 * Copyright 2014 Advanced Micro Devices, Inc.
 *
 * 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 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 COPYRIGHT HOLDER(S) OR AUTHOR(S) 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.
 */

#include <linux/mutex.h>
#include <linux/log2.h>
#include <linux/sched.h>
#include <linux/slab.h>
#include <linux/amd-iommu.h>
#include <linux/notifier.h>
#include <linux/compat.h>

struct mm_struct;

#include "kfd_priv.h"

/*
 * Initial size for the array of queues.
 * The allocated size is doubled each time
 * it is exceeded up to MAX_PROCESS_QUEUES.
 */
#define INITIAL_QUEUE_ARRAY_SIZE 16

/*
 * List of struct kfd_process (field kfd_process).
 * Unique/indexed by mm_struct*
 */
#define KFD_PROCESS_TABLE_SIZE 5 /* bits: 32 entries */
static DEFINE_HASHTABLE(kfd_processes_table, KFD_PROCESS_TABLE_SIZE);
static DEFINE_MUTEX(kfd_processes_mutex);

DEFINE_STATIC_SRCU(kfd_processes_srcu);

static struct workqueue_struct *kfd_process_wq;

struct kfd_process_release_work {
	struct work_struct kfd_work;
	struct kfd_process *p;
};

static struct kfd_process *find_process(const struct task_struct *thread);
static struct kfd_process *create_process(const struct task_struct *thread);

void kfd_process_create_wq(void)
{
	if (!kfd_process_wq)
		kfd_process_wq = create_workqueue("kfd_process_wq");
}

void kfd_process_destroy_wq(void)
{
	if (kfd_process_wq) {
		flush_workqueue(kfd_process_wq);
		destroy_workqueue(kfd_process_wq);
		kfd_process_wq = NULL;
	}
}

struct kfd_process *kfd_create_process(const struct task_struct *thread)
{
	struct kfd_process *process;

	BUG_ON(!kfd_process_wq);

	if (thread->mm == NULL)
		return ERR_PTR(-EINVAL);

	/* Only the pthreads threading model is supported. */
	if (thread->group_leader->mm != thread->mm)
		return ERR_PTR(-EINVAL);

	/* Take mmap_sem because we call __mmu_notifier_register inside */
	down_write(&thread->mm->mmap_sem);

	/*
	 * take kfd processes mutex before starting of process creation
	 * so there won't be a case where two threads of the same process
	 * create two kfd_process structures
	 */
	mutex_lock(&kfd_processes_mutex);

	/* A prior open of /dev/kfd could have already created the process. */
	process = find_process(thread);
	if (process)
		pr_debug("kfd: process already found\n");

	if (!process)
		process = create_process(thread);

	mutex_unlock(&kfd_processes_mutex);

	up_write(&thread->mm->mmap_sem);

	return process;
}

struct kfd_process *kfd_get_process(const struct task_struct *thread)
{
	struct kfd_process *process;

	if (thread->mm == NULL)
		return ERR_PTR(-EINVAL);

	/* Only the pthreads threading model is supported. */
	if (thread->group_leader->mm != thread->mm)
		return ERR_PTR(-EINVAL);

	process = find_process(thread);

	return process;
}

static struct kfd_process *find_process_by_mm(const struct mm_struct *mm)
{
	struct kfd_process *process;

	hash_for_each_possible_rcu(kfd_processes_table, process,
					kfd_processes, (uintptr_t)mm)
		if (process->mm == mm)
			return process;

	return NULL;
}

static struct kfd_process *find_process(const struct task_struct *thread)
{
	struct kfd_process *p;
	int idx;

	idx = srcu_read_lock(&kfd_processes_srcu);
	p = find_process_by_mm(thread->mm);
	srcu_read_unlock(&kfd_processes_srcu, idx);

	return p;
}

static void kfd_process_wq_release(struct work_struct *work)
{
	struct kfd_process_release_work *my_work;
	struct kfd_process_device *pdd, *temp;
	struct kfd_process *p;

	my_work = (struct kfd_process_release_work *) work;

	p = my_work->p;

	mutex_lock(&p->mutex);

	list_for_each_entry_safe(pdd, temp, &p->per_device_data,
							per_device_list) {
		amd_iommu_unbind_pasid(pdd->dev->pdev, p->pasid);
		list_del(&pdd->per_device_list);

		kfree(pdd);
	}

	kfd_pasid_free(p->pasid);

	mutex_unlock(&p->mutex);

	mutex_destroy(&p->mutex);

	kfree(p->queues);

	kfree(p);

	kfree((void *)work);
}

static void kfd_process_destroy_delayed(struct rcu_head *rcu)
{
	struct kfd_process_release_work *work;
	struct kfd_process *p;

	BUG_ON(!kfd_process_wq);

	p = container_of(rcu, struct kfd_process, rcu);
	BUG_ON(atomic_read(&p->mm->mm_count) <= 0);

	mmdrop(p->mm);

	work = (struct kfd_process_release_work *)
		kmalloc(sizeof(struct kfd_process_release_work), GFP_ATOMIC);

	if (work) {
		INIT_WORK((struct work_struct *) work, kfd_process_wq_release);
		work->p = p;
		queue_work(kfd_process_wq, (struct work_struct *) work);
	}
}

static void kfd_process_notifier_release(struct mmu_notifier *mn,
					struct mm_struct *mm)
{
	struct kfd_process *p;

	/*
	 * The kfd_process structure can not be free because the
	 * mmu_notifier srcu is read locked
	 */
	p = container_of(mn, struct kfd_process, mmu_notifier);
	BUG_ON(p->mm != mm);

	mutex_lock(&kfd_processes_mutex);
	hash_del_rcu(&p->kfd_processes);
	mutex_unlock(&kfd_processes_mutex);
	synchronize_srcu(&kfd_processes_srcu);

	mutex_lock(&p->mutex);

	/* In case our notifier is called before IOMMU notifier */
	pqm_uninit(&p->pqm);

	mutex_unlock(&p->mutex);

	/*
	 * Because we drop mm_count inside kfd_process_destroy_delayed
	 * and because the mmu_notifier_unregister function also drop
	 * mm_count we need to take an extra count here.
	 */
	atomic_inc(&p->mm->mm_count);
	mmu_notifier_unregister_no_release(&p->mmu_notifier, p->mm);
	mmu_notifier_call_srcu(&p->rcu, &kfd_process_destroy_delayed);
}

static const struct mmu_notifier_ops kfd_process_mmu_notifier_ops = {
	.release = kfd_process_notifier_release,
};

static struct kfd_process *create_process(const struct task_struct *thread)
{
	struct kfd_process *process;
	int err = -ENOMEM;

	process = kzalloc(sizeof(*process), GFP_KERNEL);

	if (!process)
		goto err_alloc_process;

	process->queues = kmalloc_array(INITIAL_QUEUE_ARRAY_SIZE,
					sizeof(process->queues[0]), GFP_KERNEL);
	if (!process->queues)
		goto err_alloc_queues;

	process->pasid = kfd_pasid_alloc();
	if (process->pasid == 0)
		goto err_alloc_pasid;

	mutex_init(&process->mutex);

	process->mm = thread->mm;

	/* register notifier */
	process->mmu_notifier.ops = &kfd_process_mmu_notifier_ops;
	err = __mmu_notifier_register(&process->mmu_notifier, process->mm);
	if (err)
		goto err_mmu_notifier;

	hash_add_rcu(kfd_processes_table, &process->kfd_processes,
			(uintptr_t)process->mm);

	process->lead_thread = thread->group_leader;

	process->queue_array_size = INITIAL_QUEUE_ARRAY_SIZE;

	INIT_LIST_HEAD(&process->per_device_data);

	err = pqm_init(&process->pqm, process);
	if (err != 0)
		goto err_process_pqm_init;

	/* init process apertures*/
	process->is_32bit_user_mode = is_compat_task();
	if (kfd_init_apertures(process) != 0)
		goto err_init_apretures;

	return process;

err_init_apretures:
	pqm_uninit(&process->pqm);
err_process_pqm_init:
	hash_del_rcu(&process->kfd_processes);
	synchronize_rcu();
	mmu_notifier_unregister_no_release(&process->mmu_notifier, process->mm);
err_mmu_notifier:
	kfd_pasid_free(process->pasid);
err_alloc_pasid:
	kfree(process->queues);
err_alloc_queues:
	kfree(process);
err_alloc_process:
	return ERR_PTR(err);
}

struct kfd_process_device *kfd_get_process_device_data(struct kfd_dev *dev,
							struct kfd_process *p,
							int create_pdd)
{
	struct kfd_process_device *pdd = NULL;

	list_for_each_entry(pdd, &p->per_device_data, per_device_list)
		if (pdd->dev == dev)
			return pdd;

	if (create_pdd) {
		pdd = kzalloc(sizeof(*pdd), GFP_KERNEL);
		if (pdd != NULL) {
			pdd->dev = dev;
			INIT_LIST_HEAD(&pdd->qpd.queues_list);
			INIT_LIST_HEAD(&pdd->qpd.priv_queue_list);
			pdd->qpd.dqm = dev->dqm;
			list_add(&pdd->per_device_list, &p->per_device_data);
		}
	}

	return pdd;
}

/*
 * Direct the IOMMU to bind the process (specifically the pasid->mm)
 * to the device.
 * Unbinding occurs when the process dies or the device is removed.
 *
 * Assumes that the process lock is held.
 */
struct kfd_process_device *kfd_bind_process_to_device(struct kfd_dev *dev,
							struct kfd_process *p)
{
	struct kfd_process_device *pdd = kfd_get_process_device_data(dev, p, 1);
	int err;

	if (pdd == NULL)
		return ERR_PTR(-ENOMEM);

	if (pdd->bound)
		return pdd;

	err = amd_iommu_bind_pasid(dev->pdev, p->pasid, p->lead_thread);
	if (err < 0)
		return ERR_PTR(err);

	pdd->bound = true;

	return pdd;
}

void kfd_unbind_process_from_device(struct kfd_dev *dev, unsigned int pasid)
{
	struct kfd_process *p;
	struct kfd_process_device *pdd;
	int idx, i;

	BUG_ON(dev == NULL);

	idx = srcu_read_lock(&kfd_processes_srcu);

	hash_for_each_rcu(kfd_processes_table, i, p, kfd_processes)
		if (p->pasid == pasid)
			break;

	srcu_read_unlock(&kfd_processes_srcu, idx);

	BUG_ON(p->pasid != pasid);

	mutex_lock(&p->mutex);

	pqm_uninit(&p->pqm);

	pdd = kfd_get_process_device_data(dev, p, 0);

	/*
	 * Just mark pdd as unbound, because we still need it to call
	 * amd_iommu_unbind_pasid() in when the process exits.
	 * We don't call amd_iommu_unbind_pasid() here
	 * because the IOMMU called us.
	 */
	if (pdd)
		pdd->bound = false;

	mutex_unlock(&p->mutex);
}

struct kfd_process_device *kfd_get_first_process_device_data(struct kfd_process *p)
{
	return list_first_entry(&p->per_device_data,
				struct kfd_process_device,
				per_device_list);
}

struct kfd_process_device *kfd_get_next_process_device_data(struct kfd_process *p,
						struct kfd_process_device *pdd)
{
	if (list_is_last(&pdd->per_device_list, &p->per_device_data))
		return NULL;
	return list_next_entry(pdd, per_device_list);
}

bool kfd_has_process_device_data(struct kfd_process *p)
{
	return !(list_empty(&p->per_device_data));
}