summaryrefslogtreecommitdiffstats
path: root/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c
blob: dc043c3d53249ec8899cd0b874971ce5eb6af203 (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
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
/**************************************************************************
 *
 * Copyright © 2009 VMware, Inc., Palo Alto, CA., USA
 * All Rights Reserved.
 *
 * 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, sub license, 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 NON-INFRINGEMENT. IN NO EVENT SHALL
 * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS 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 "vmwgfx_drv.h"
#include "drmP.h"
#include "ttm/ttm_placement.h"

bool vmw_fifo_have_3d(struct vmw_private *dev_priv)
{
	__le32 __iomem *fifo_mem = dev_priv->mmio_virt;
	uint32_t fifo_min, hwversion;

	if (!(dev_priv->capabilities & SVGA_CAP_EXTENDED_FIFO))
		return false;

	fifo_min = ioread32(fifo_mem  + SVGA_FIFO_MIN);
	if (fifo_min <= SVGA_FIFO_3D_HWVERSION * sizeof(unsigned int))
		return false;

	hwversion = ioread32(fifo_mem + SVGA_FIFO_3D_HWVERSION);
	if (hwversion == 0)
		return false;

	if (hwversion < SVGA3D_HWVERSION_WS65_B1)
		return false;

	return true;
}

bool vmw_fifo_have_pitchlock(struct vmw_private *dev_priv)
{
	__le32 __iomem *fifo_mem = dev_priv->mmio_virt;
	uint32_t caps;

	if (!(dev_priv->capabilities & SVGA_CAP_EXTENDED_FIFO))
		return false;

	caps = ioread32(fifo_mem + SVGA_FIFO_CAPABILITIES);
	if (caps & SVGA_FIFO_CAP_PITCHLOCK)
		return true;

	return false;
}

int vmw_fifo_init(struct vmw_private *dev_priv, struct vmw_fifo_state *fifo)
{
	__le32 __iomem *fifo_mem = dev_priv->mmio_virt;
	uint32_t max;
	uint32_t min;
	uint32_t dummy;
	int ret;

	fifo->static_buffer_size = VMWGFX_FIFO_STATIC_SIZE;
	fifo->static_buffer = vmalloc(fifo->static_buffer_size);
	if (unlikely(fifo->static_buffer == NULL))
		return -ENOMEM;

	fifo->last_buffer_size = VMWGFX_FIFO_STATIC_SIZE;
	fifo->last_data_size = 0;
	fifo->last_buffer_add = false;
	fifo->last_buffer = vmalloc(fifo->last_buffer_size);
	if (unlikely(fifo->last_buffer == NULL)) {
		ret = -ENOMEM;
		goto out_err;
	}

	fifo->dynamic_buffer = NULL;
	fifo->reserved_size = 0;
	fifo->using_bounce_buffer = false;

	mutex_init(&fifo->fifo_mutex);
	init_rwsem(&fifo->rwsem);

	/*
	 * Allow mapping the first page read-only to user-space.
	 */

	DRM_INFO("width %d\n", vmw_read(dev_priv, SVGA_REG_WIDTH));
	DRM_INFO("height %d\n", vmw_read(dev_priv, SVGA_REG_HEIGHT));
	DRM_INFO("bpp %d\n", vmw_read(dev_priv, SVGA_REG_BITS_PER_PIXEL));

	mutex_lock(&dev_priv->hw_mutex);
	dev_priv->enable_state = vmw_read(dev_priv, SVGA_REG_ENABLE);
	dev_priv->config_done_state = vmw_read(dev_priv, SVGA_REG_CONFIG_DONE);
	vmw_write(dev_priv, SVGA_REG_ENABLE, 1);

	min = 4;
	if (dev_priv->capabilities & SVGA_CAP_EXTENDED_FIFO)
		min = vmw_read(dev_priv, SVGA_REG_MEM_REGS);
	min <<= 2;

	if (min < PAGE_SIZE)
		min = PAGE_SIZE;

	iowrite32(min, fifo_mem + SVGA_FIFO_MIN);
	iowrite32(dev_priv->mmio_size, fifo_mem + SVGA_FIFO_MAX);
	wmb();
	iowrite32(min,  fifo_mem + SVGA_FIFO_NEXT_CMD);
	iowrite32(min,  fifo_mem + SVGA_FIFO_STOP);
	iowrite32(0, fifo_mem + SVGA_FIFO_BUSY);
	mb();

	vmw_write(dev_priv, SVGA_REG_CONFIG_DONE, 1);
	mutex_unlock(&dev_priv->hw_mutex);

	max = ioread32(fifo_mem + SVGA_FIFO_MAX);
	min = ioread32(fifo_mem  + SVGA_FIFO_MIN);
	fifo->capabilities = ioread32(fifo_mem + SVGA_FIFO_CAPABILITIES);

	DRM_INFO("Fifo max 0x%08x min 0x%08x cap 0x%08x\n",
		 (unsigned int) max,
		 (unsigned int) min,
		 (unsigned int) fifo->capabilities);

	atomic_set(&dev_priv->fence_seq, dev_priv->last_read_sequence);
	iowrite32(dev_priv->last_read_sequence, fifo_mem + SVGA_FIFO_FENCE);
	vmw_fence_queue_init(&fifo->fence_queue);
	return vmw_fifo_send_fence(dev_priv, &dummy);
out_err:
	vfree(fifo->static_buffer);
	fifo->static_buffer = NULL;
	return ret;
}

void vmw_fifo_ping_host(struct vmw_private *dev_priv, uint32_t reason)
{
	__le32 __iomem *fifo_mem = dev_priv->mmio_virt;

	mutex_lock(&dev_priv->hw_mutex);

	if (unlikely(ioread32(fifo_mem + SVGA_FIFO_BUSY) == 0)) {
		iowrite32(1, fifo_mem + SVGA_FIFO_BUSY);
		vmw_write(dev_priv, SVGA_REG_SYNC, reason);
	}

	mutex_unlock(&dev_priv->hw_mutex);
}

void vmw_fifo_release(struct vmw_private *dev_priv, struct vmw_fifo_state *fifo)
{
	__le32 __iomem *fifo_mem = dev_priv->mmio_virt;

	mutex_lock(&dev_priv->hw_mutex);

	while (vmw_read(dev_priv, SVGA_REG_BUSY) != 0)
		vmw_write(dev_priv, SVGA_REG_SYNC, SVGA_SYNC_GENERIC);

	dev_priv->last_read_sequence = ioread32(fifo_mem + SVGA_FIFO_FENCE);

	vmw_write(dev_priv, SVGA_REG_CONFIG_DONE,
		  dev_priv->config_done_state);
	vmw_write(dev_priv, SVGA_REG_ENABLE,
		  dev_priv->enable_state);

	mutex_unlock(&dev_priv->hw_mutex);
	vmw_fence_queue_takedown(&fifo->fence_queue);

	if (likely(fifo->last_buffer != NULL)) {
		vfree(fifo->last_buffer);
		fifo->last_buffer = NULL;
	}

	if (likely(fifo->static_buffer != NULL)) {
		vfree(fifo->static_buffer);
		fifo->static_buffer = NULL;
	}

	if (likely(fifo->dynamic_buffer != NULL)) {
		vfree(fifo->dynamic_buffer);
		fifo->dynamic_buffer = NULL;
	}
}

static bool vmw_fifo_is_full(struct vmw_private *dev_priv, uint32_t bytes)
{
	__le32 __iomem *fifo_mem = dev_priv->mmio_virt;
	uint32_t max = ioread32(fifo_mem + SVGA_FIFO_MAX);
	uint32_t next_cmd = ioread32(fifo_mem + SVGA_FIFO_NEXT_CMD);
	uint32_t min = ioread32(fifo_mem + SVGA_FIFO_MIN);
	uint32_t stop = ioread32(fifo_mem + SVGA_FIFO_STOP);

	return ((max - next_cmd) + (stop - min) <= bytes);
}

static int vmw_fifo_wait_noirq(struct vmw_private *dev_priv,
			       uint32_t bytes, bool interruptible,
			       unsigned long timeout)
{
	int ret = 0;
	unsigned long end_jiffies = jiffies + timeout;
	DEFINE_WAIT(__wait);

	DRM_INFO("Fifo wait noirq.\n");

	for (;;) {
		prepare_to_wait(&dev_priv->fifo_queue, &__wait,
				(interruptible) ?
				TASK_INTERRUPTIBLE : TASK_UNINTERRUPTIBLE);
		if (!vmw_fifo_is_full(dev_priv, bytes))
			break;
		if (time_after_eq(jiffies, end_jiffies)) {
			ret = -EBUSY;
			DRM_ERROR("SVGA device lockup.\n");
			break;
		}
		schedule_timeout(1);
		if (interruptible && signal_pending(current)) {
			ret = -ERESTARTSYS;
			break;
		}
	}
	finish_wait(&dev_priv->fifo_queue, &__wait);
	wake_up_all(&dev_priv->fifo_queue);
	DRM_INFO("Fifo noirq exit.\n");
	return ret;
}

static int vmw_fifo_wait(struct vmw_private *dev_priv,
			 uint32_t bytes, bool interruptible,
			 unsigned long timeout)
{
	long ret = 1L;
	unsigned long irq_flags;

	if (likely(!vmw_fifo_is_full(dev_priv, bytes)))
		return 0;

	vmw_fifo_ping_host(dev_priv, SVGA_SYNC_FIFOFULL);
	if (!(dev_priv->capabilities & SVGA_CAP_IRQMASK))
		return vmw_fifo_wait_noirq(dev_priv, bytes,
					   interruptible, timeout);

	mutex_lock(&dev_priv->hw_mutex);
	if (atomic_add_return(1, &dev_priv->fifo_queue_waiters) > 0) {
		spin_lock_irqsave(&dev_priv->irq_lock, irq_flags);
		outl(SVGA_IRQFLAG_FIFO_PROGRESS,
		     dev_priv->io_start + VMWGFX_IRQSTATUS_PORT);
		vmw_write(dev_priv, SVGA_REG_IRQMASK,
			  vmw_read(dev_priv, SVGA_REG_IRQMASK) |
			  SVGA_IRQFLAG_FIFO_PROGRESS);
		spin_unlock_irqrestore(&dev_priv->irq_lock, irq_flags);
	}
	mutex_unlock(&dev_priv->hw_mutex);

	if (interruptible)
		ret = wait_event_interruptible_timeout
		    (dev_priv->fifo_queue,
		     !vmw_fifo_is_full(dev_priv, bytes), timeout);
	else
		ret = wait_event_timeout
		    (dev_priv->fifo_queue,
		     !vmw_fifo_is_full(dev_priv, bytes), timeout);

	if (unlikely(ret == 0))
		ret = -EBUSY;
	else if (likely(ret > 0))
		ret = 0;

	mutex_lock(&dev_priv->hw_mutex);
	if (atomic_dec_and_test(&dev_priv->fifo_queue_waiters)) {
		spin_lock_irqsave(&dev_priv->irq_lock, irq_flags);
		vmw_write(dev_priv, SVGA_REG_IRQMASK,
			  vmw_read(dev_priv, SVGA_REG_IRQMASK) &
			  ~SVGA_IRQFLAG_FIFO_PROGRESS);
		spin_unlock_irqrestore(&dev_priv->irq_lock, irq_flags);
	}
	mutex_unlock(&dev_priv->hw_mutex);

	return ret;
}

void *vmw_fifo_reserve(struct vmw_private *dev_priv, uint32_t bytes)
{
	struct vmw_fifo_state *fifo_state = &dev_priv->fifo;
	__le32 __iomem *fifo_mem = dev_priv->mmio_virt;
	uint32_t max;
	uint32_t min;
	uint32_t next_cmd;
	uint32_t reserveable = fifo_state->capabilities & SVGA_FIFO_CAP_RESERVE;
	int ret;

	mutex_lock(&fifo_state->fifo_mutex);
	max = ioread32(fifo_mem + SVGA_FIFO_MAX);
	min = ioread32(fifo_mem + SVGA_FIFO_MIN);
	next_cmd = ioread32(fifo_mem + SVGA_FIFO_NEXT_CMD);

	if (unlikely(bytes >= (max - min)))
		goto out_err;

	BUG_ON(fifo_state->reserved_size != 0);
	BUG_ON(fifo_state->dynamic_buffer != NULL);

	fifo_state->reserved_size = bytes;

	while (1) {
		uint32_t stop = ioread32(fifo_mem + SVGA_FIFO_STOP);
		bool need_bounce = false;
		bool reserve_in_place = false;

		if (next_cmd >= stop) {
			if (likely((next_cmd + bytes < max ||
				    (next_cmd + bytes == max && stop > min))))
				reserve_in_place = true;

			else if (vmw_fifo_is_full(dev_priv, bytes)) {
				ret = vmw_fifo_wait(dev_priv, bytes,
						    false, 3 * HZ);
				if (unlikely(ret != 0))
					goto out_err;
			} else
				need_bounce = true;

		} else {

			if (likely((next_cmd + bytes < stop)))
				reserve_in_place = true;
			else {
				ret = vmw_fifo_wait(dev_priv, bytes,
						    false, 3 * HZ);
				if (unlikely(ret != 0))
					goto out_err;
			}
		}

		if (reserve_in_place) {
			if (reserveable || bytes <= sizeof(uint32_t)) {
				fifo_state->using_bounce_buffer = false;

				if (reserveable)
					iowrite32(bytes, fifo_mem +
						  SVGA_FIFO_RESERVED);
				return fifo_mem + (next_cmd >> 2);
			} else {
				need_bounce = true;
			}
		}

		if (need_bounce) {
			fifo_state->using_bounce_buffer = true;
			if (bytes < fifo_state->static_buffer_size)
				return fifo_state->static_buffer;
			else {
				fifo_state->dynamic_buffer = vmalloc(bytes);
				return fifo_state->dynamic_buffer;
			}
		}
	}
out_err:
	fifo_state->reserved_size = 0;
	mutex_unlock(&fifo_state->fifo_mutex);
	return NULL;
}

static void vmw_fifo_res_copy(struct vmw_fifo_state *fifo_state,
			      __le32 __iomem *fifo_mem,
			      uint32_t next_cmd,
			      uint32_t max, uint32_t min, uint32_t bytes)
{
	uint32_t chunk_size = max - next_cmd;
	uint32_t rest;
	uint32_t *buffer = (fifo_state->dynamic_buffer != NULL) ?
	    fifo_state->dynamic_buffer : fifo_state->static_buffer;

	if (bytes < chunk_size)
		chunk_size = bytes;

	iowrite32(bytes, fifo_mem + SVGA_FIFO_RESERVED);
	mb();
	memcpy_toio(fifo_mem + (next_cmd >> 2), buffer, chunk_size);
	rest = bytes - chunk_size;
	if (rest)
		memcpy_toio(fifo_mem + (min >> 2), buffer + (chunk_size >> 2),
			    rest);
}

static void vmw_fifo_slow_copy(struct vmw_fifo_state *fifo_state,
			       __le32 __iomem *fifo_mem,
			       uint32_t next_cmd,
			       uint32_t max, uint32_t min, uint32_t bytes)
{
	uint32_t *buffer = (fifo_state->dynamic_buffer != NULL) ?
	    fifo_state->dynamic_buffer : fifo_state->static_buffer;

	while (bytes > 0) {
		iowrite32(*buffer++, fifo_mem + (next_cmd >> 2));
		next_cmd += sizeof(uint32_t);
		if (unlikely(next_cmd == max))
			next_cmd = min;
		mb();
		iowrite32(next_cmd, fifo_mem + SVGA_FIFO_NEXT_CMD);
		mb();
		bytes -= sizeof(uint32_t);
	}
}

void vmw_fifo_commit(struct vmw_private *dev_priv, uint32_t bytes)
{
	struct vmw_fifo_state *fifo_state = &dev_priv->fifo;
	__le32 __iomem *fifo_mem = dev_priv->mmio_virt;
	uint32_t next_cmd = ioread32(fifo_mem + SVGA_FIFO_NEXT_CMD);
	uint32_t max = ioread32(fifo_mem + SVGA_FIFO_MAX);
	uint32_t min = ioread32(fifo_mem + SVGA_FIFO_MIN);
	bool reserveable = fifo_state->capabilities & SVGA_FIFO_CAP_RESERVE;

	BUG_ON((bytes & 3) != 0);
	BUG_ON(bytes > fifo_state->reserved_size);

	fifo_state->reserved_size = 0;

	if (fifo_state->using_bounce_buffer) {
		if (reserveable)
			vmw_fifo_res_copy(fifo_state, fifo_mem,
					  next_cmd, max, min, bytes);
		else
			vmw_fifo_slow_copy(fifo_state, fifo_mem,
					   next_cmd, max, min, bytes);

		if (fifo_state->dynamic_buffer) {
			vfree(fifo_state->dynamic_buffer);
			fifo_state->dynamic_buffer = NULL;
		}

	}

	down_write(&fifo_state->rwsem);
	if (fifo_state->using_bounce_buffer || reserveable) {
		next_cmd += bytes;
		if (next_cmd >= max)
			next_cmd -= max - min;
		mb();
		iowrite32(next_cmd, fifo_mem + SVGA_FIFO_NEXT_CMD);
	}

	if (reserveable)
		iowrite32(0, fifo_mem + SVGA_FIFO_RESERVED);
	mb();
	up_write(&fifo_state->rwsem);
	vmw_fifo_ping_host(dev_priv, SVGA_SYNC_GENERIC);
	mutex_unlock(&fifo_state->fifo_mutex);
}

int vmw_fifo_send_fence(struct vmw_private *dev_priv, uint32_t *sequence)
{
	struct vmw_fifo_state *fifo_state = &dev_priv->fifo;
	struct svga_fifo_cmd_fence *cmd_fence;
	void *fm;
	int ret = 0;
	uint32_t bytes = sizeof(__le32) + sizeof(*cmd_fence);

	fm = vmw_fifo_reserve(dev_priv, bytes);
	if (unlikely(fm == NULL)) {
		*sequence = atomic_read(&dev_priv->fence_seq);
		ret = -ENOMEM;
		(void)vmw_fallback_wait(dev_priv, false, true, *sequence,
					false, 3*HZ);
		goto out_err;
	}

	do {
		*sequence = atomic_add_return(1, &dev_priv->fence_seq);
	} while (*sequence == 0);

	if (!(fifo_state->capabilities & SVGA_FIFO_CAP_FENCE)) {

		/*
		 * Don't request hardware to send a fence. The
		 * waiting code in vmwgfx_irq.c will emulate this.
		 */

		vmw_fifo_commit(dev_priv, 0);
		return 0;
	}

	*(__le32 *) fm = cpu_to_le32(SVGA_CMD_FENCE);
	cmd_fence = (struct svga_fifo_cmd_fence *)
	    ((unsigned long)fm + sizeof(__le32));

	iowrite32(*sequence, &cmd_fence->fence);
	fifo_state->last_buffer_add = true;
	vmw_fifo_commit(dev_priv, bytes);
	fifo_state->last_buffer_add = false;
	(void) vmw_fence_push(&fifo_state->fence_queue, *sequence);
	vmw_update_sequence(dev_priv, fifo_state);

out_err:
	return ret;
}

/**
 * Map the first page of the FIFO read-only to user-space.
 */

static int vmw_fifo_vm_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
{
	int ret;
	unsigned long address = (unsigned long)vmf->virtual_address;

	if (address != vma->vm_start)
		return VM_FAULT_SIGBUS;

	ret = vm_insert_pfn(vma, address, vma->vm_pgoff);
	if (likely(ret == -EBUSY || ret == 0))
		return VM_FAULT_NOPAGE;
	else if (ret == -ENOMEM)
		return VM_FAULT_OOM;

	return VM_FAULT_SIGBUS;
}

static struct vm_operations_struct vmw_fifo_vm_ops = {
	.fault = vmw_fifo_vm_fault,
	.open = NULL,
	.close = NULL
};

int vmw_fifo_mmap(struct file *filp, struct vm_area_struct *vma)
{
	struct drm_file *file_priv;
	struct vmw_private *dev_priv;

	file_priv = filp->private_data;
	dev_priv = vmw_priv(file_priv->minor->dev);

	if (vma->vm_pgoff != (dev_priv->mmio_start >> PAGE_SHIFT) ||
	    (vma->vm_end - vma->vm_start) != PAGE_SIZE)
		return -EINVAL;

	vma->vm_flags &= ~(VM_WRITE | VM_MAYWRITE);
	vma->vm_flags |= VM_IO | VM_PFNMAP | VM_DONTEXPAND | VM_SHARED;
	vma->vm_page_prot = vm_get_page_prot(vma->vm_flags);
	vma->vm_page_prot = ttm_io_prot(TTM_PL_FLAG_UNCACHED,
					vma->vm_page_prot);
	vma->vm_ops = &vmw_fifo_vm_ops;
	return 0;
}